[{"article_id":10040727,"title":"What is IBM’s Project CodeNet?","content":"At its recently concluded Think 2021  conference, IBM introduced Project CodeNet to develop machine learning models that can help in programming. The large dataset consists of 14 million code samples and 500 million lines of code in over 55 different languages, including C++, Java, Go, Python, COBOL, Pascal, and Fortran. Project CodeNet Modern computer programs have millions of lines of code and are hard to debug, maintain, update, and document. The use of artificial intelligence to write codes has been an important area of research for many years. However, it is easier said than done. The fact that programming languages have context poses a major challenge. Every line of code must be contextual. Understanding the context is a tricky and time-consuming task. The challenge gets bigger with larger programs as context can be related to multiple libraries of code. IBM’s Project CodeNet can help extract this context with a sequence-to-sequence model. As per IBM’s team, this method is more significant in machine understanding of code instead of machine processing of code. Project CodeNet has code samples curated from open programming competitions over the years, including challenges posted on coding platforms such as AIZU and AtCoder containing correct and incorrect answers to the challenges. It consists of high-quality metadata and annotations. It also has a rich set of information in terms of code size, memory footprint, CPU run time, etc. The team used reinforcement learning techniques for code translation by determining the equivalence of two code samples in different languages by curating sample input and output from the problem description (it contains problem statement, input and output format). Available as part of the dataset in CodeNet, users can execute these accepted codes samples to extract additional information and verify outputs from the generative AI models for correctness. This feature is handy when translating from one language to another. The dataset from Project CodeNet can be used for code search and clone detection. Since these code samples are labelled with their acceptance status, AI techniques can be used to distinguish correct from incorrect codes. Samples are also labelled with CPU run time and memory footprint to understand regression and prediction. “Given its wealth of programs written in a multitude of languages, we believe Project CodeNet can serve as a benchmark dataset for source-to-source translation and do for AI and code what the ImageNet dataset did years ago for computer vision,” the team said. Teaching AI to code IBM’s Project CodeNet is the latest stab at teaching AI to code. Earlier, researchers from Microsoft and the University of Cambridge developed DeepCoder to solve fundamental programming problems. The team used a technique called program synthesis, in which the tool creates new programs by collating lines of code sourced from existing software. It requires a list of inputs and outputs for each code fragment to determine which pieces of code can be used. In 2019, MIT introduced SketchAdapt, a program-writing AI. SketchAdapt is trained on tens of thousands of program examples and can compose short, high-level programs. The tool knows when to switch from statistical pattern-matching to less efficient yet more versatile symbolic reasoning mode. One of the use cases of GPT-3 is code development. This language model from OpenAI can assist users in building their applications with text prompts; the system uses user input to generate code. This is particularly useful in cases where rapid prototyping of applications is required. Companies such as debuild.co are already using GPT-3 for accelerating the application development process. This is mind blowing.With GPT-3, I built a layout generator where you just describe any layout you want, and it generates the JSX code for you.W H A T pic.twitter.com\/w8JkrZO4lk— Sharif Shameem (@sharifshameem) July 13, 2020","excerpt":"IBM’s Project CodeNet can help extract this context with a sequence-to-sequence model.","categories":["AI Features"],"tags":["Programming Languages","Software Development"],"author_name":"Shraddha Goled","publish_date":"2021-05-24T17:00:00","publication_year":"2021","word_count":616,"keywords":["machine learning","artificial intelligence","OpenAI","AI","TPU","Programming Languages","computer vision","RAG","Python","generative AI","Software Development","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","generative AI","OpenAI","RAG","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-is-ibms-project-codenet\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":23622,"title":"Maths Phobia: HRD Minister Prakash Javadekar Forms Committee To Tackle Fear Among Students","content":"In a welcome move, Prakash Javadekar, Union Minister for Human Resources and Development announced that a committee had been formed to take measures against the alleged deep-rooted phobia in the minds of the students against Mathematics. The decision was reportedly made after a national survey of state-run schools revealed “fear complex” about Mathematics. Javadekar told reporters in New Delhi, “One thing has come up in the National Assessment Survey 2017 that students have fear complex about Maths. We have formed a committee under Gujarat Education Minister Bhupendrasinh Chudasama to see how to remove that fear complex so that the students see it as a friendly subject.” Panel formed to suggest measures to combat maths phobia among students after a national survey of state-run schools revealed “fear complex” about the subject: HRD Minister Prakash Javadekar. — Press Trust of India (@PTI_News) April 11, 2018 The National Achievement Survey (NAS) was conducted for classes III, V and VIII in government and government-aided schools and the learning levels of more than 25 lakh students from 1,10,000 schools across 700 districts in all 36 states\/UTs were assessed. According to reports, the survey used several test booklets with 45 questions for classes III and V related to language, mathematics; and 60 questions for class VIII in mathematics, language, sciences and social sciences. Javadekar has said that the committee, which would be headed by both academicians as well as officials, would submit a report on the same in three months. Noted psychologist Mark H Ashcraft had defined maths anxiety as, “A feeling of tension, apprehension, or fear that interferes with math performance.” In his research Math Anxiety: Personal, Educational, and Cognitive Consequences, he characterises “math-anxious” individuals as ones who have strong tendency to avoid maths, which ultimately undercuts their math competence and forecloses important career paths. “Math anxiety disrupts cognitive processing by compromising ongoing activity in working memory. Although the causes of math anxiety are undetermined, some teaching styles are implicated as risk factors,” says Ashcraft’s research.","excerpt":"In a welcome move, Prakash Javadekar, Union Minister for Human Resources and Development announced that a committee had been formed to take measures against the alleged deep-rooted phobia in the minds of the students against Mathematics. The decision was reportedly made after a national survey of state-run schools revealed “fear complex” about Mathematics. Javadekar told reporters in […]","categories":["AI News"],"tags":["mathematics"],"author_name":"Prajakta Hebbar","publish_date":"2018-04-12T12:44:19","publication_year":"2018","word_count":331,"keywords":["mathematics","Go","programming_languages:R","AI","programming_languages:Java","programming_languages:Go","ViT","Rust","R","Java","programming_languages:Rust"],"extracted_tech_keywords":["AI","R","Go","Rust","Java","ViT","programming_languages:R","programming_languages:Java","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/committee-maths-phobia\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":33456,"title":"Data Preprocessing With R: Hands-On Tutorial","content":"When it comes to Machine Learning and Artificial intelligence there are only a few top-performing programming languages to choose from. In the previous tutorial, we learned how to do Data Preprocessing in Python. Since R is among the top performers in Data Science, in this tutorial we will learn to perform Data Preprocessing task with R. (Note: The following tutorial will require basic programming knowledge of R.) In this tutorial, we will learn to perform the following operations on a raw dataset: Dealing with missing data Dealing with categorical data Splitting the dataset into training and testing sets Scaling the features Data Preprocessing in R The following steps are crucial: Importing The Dataset dataset = read.csv('dataset.csv') As one can see, this is a simple dataset consisting of four features. The dependent factor is the ‘purchased_item’ column. If the above dataset is to be used for machine learning, the idea will be to predict if an item got purchased or not depending on the country, age and salary of a person. Also, the highlighted cells with value ‘NA’ denotes missing values in the dataset. Dealing With Missing Values dataset$age = ifelse(is.na(dataset$age),ave(dataset$age, FUN = function(x) mean(x, na.rm = 'TRUE')),dataset$age) dataset$salary = ifelse(is.na(dataset$salary), ave(dataset$salary, FUN = function(x) mean(x, na.rm = 'TRUE')), dataset$salary) The above code blocks check for missing values in the age and salary columns and update the missing cells with the column-wise average. dataset$column_header: Selects the column in the dataset specified after $ (age and salary). is.na(dataset$column_header): This method returns true for all the cells in the specified column with no values. ave(dataset$column_header, FUN = function(x) mean(x, na.rm = ‘TRUE’)): Ths method calculates the average of the column passed as argument. Output : dataset$age = as.numeric(format(round(dataset$age, 0))) Since we are not interested in having decimal places for age we will round it up using the above code. The argument 0 in the round function means no decimal places. After executing the above code block the dataset would look like what’s shown below : Note : Unlike Python where we use Numpy arrays to store the data to perform operations, we directly perform our operations on the dataset, which is a list, in R. We do not need to categorize the dependent and independent factors explicitly since R uses an attribute called formula to identify dependent and independent factors from a dataset. Dealing With Categorical Data Categorical variables represent types of data which may be divided into groups. Examples of categorical variables are race, sex, age group, educational level etc. In our dataset, we have two categorical features, nation, and purchased_item. In R we can use the factor method to convert texts into numerical codes. dataset$nation = factor(dataset$nation, levels = c('India','Germany','Russia'), labels = c(1,2,3)) dataset$purchased_item = factor(dataset$purchased_item, levels = c('No','Yes'),  labels = c(0,1)) factor(dataset$olumn_header, levels = c(), labels = c()) : the factor method converts the categorical features in the specified column to factors or numerical codes. levels: the categories in the column passed as a vector. Example c(‘India’,’Germany’,’Russia’) labels: The numerical codes for the specified categories in the same order. Example c(1,2,3)) Output: Splitting The Dataset Into Training And Testing Sets We will use the caTools library in R  to split our dataset to training_set and test_set install.packages('caTools') #install once library(caTools) # importing caTools library set.seed(123) split = sample.split(dataset$purchased_item, SplitRatio = 0.8) training_set = subset(dataset, split == TRUE) test_set = subset(dataset, split == FALSE) set.seed(): The seed function preserves the uniqueness of the split i.e, for each seed value, the split will be unique. It is similar to the random_state argument in python. sample.split(dataset$dependent_factor, SplitRatio = 0.8): This method will return boolean values with the length of the original dataset  in the specified SplitRatio .0.8 gives 80 percentage Trues and 20 percentage Falses. For example, the above code block will assign the variable split with values [TRUE  TRUE TRUE TRUE TRUE FALSE TRUE TRUE FALSE TRUE TRUE TRUE TRUE TRUE FALSE] subset(dataset, split == TRUE): This method will return a subset of the dataset passed as an argument where the split is True. (80 percent of the original dataset with respect to the given code) subset(dataset, split == FALSE): This method will return a subset of the dataset passed as an argument where the split is False. (20 percent of the original dataset with respect to the given code) Scaling The Features training_set[,3:4] = scale(training_set[,3:4]) test_set[,3:4] = scale(test_set[,3:4]) The scale method in R can be used to scale the features in the dataset. Here we are only scaling the non-factors which are the age and the salary. Output: Training_set: Test_set:","excerpt":"When it comes to Machine Learning and Artificial intelligence there are only a few top-performing programming languages to choose from. In the previous tutorial, we learned how to do Data Preprocessing in Python. Since R is among the top performers in Data Science, in this tutorial we will learn to perform Data Preprocessing task with […]","categories":["Deep Tech"],"tags":[],"author_name":"Amal Nair","publish_date":"2019-01-15T05:30:01","publication_year":"2019","word_count":759,"keywords":["data science","NumPy","machine learning","artificial intelligence","TPU","AI","RAG","Python","Ray","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","Ray","NumPy","RAG","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/data-preprocessing-with-r-hands-on-tutorial\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":41714,"title":"Meet V, The New Statically Typed Programming Language Inspired By Go & Rust","content":"A new statically typed compiled programming language, known as V language has been open-sourced recently by its creator Alex Medvednikov. According to Medvednikov, the language is designed for building maintainable software. It is a simple language and is similar to the Go language and is inspired by Rust, Swift, and Oberon languages. Currently, V emits C and uses GCC\/Clang for optimized production builds. It manages memory at compilation time (like Rust). For now, it supports Windows, macOS, Linux, *BSD while Android and iOS will be supported later this year. Features Simple The simplicity of this language can be shown in this example. Unlike languages like C and C++, the “hello world” program can be executed easily only by typing “printIn” which is one of the built-in functions to print the value to standard output, i.e. printIn ( ‘hello world’ ). Fast Compilation V compiles ≈1.2 million lines of code per second per CPU core. Such speed is achieved by direct machine code generation and strong modularity. The language can also emit C but the drawback is that the compilation speed then drops to ≈100k lines\/second\/CPU. The creator, Medvednikov also mentioned that the direct machine code generation is at a very early stage, which supports only x64\/Mach-O currently and will try to make it stable enough by the end of this year. Handling Errors V combines Option and Result into one type. If you don’t need to return an error, you can simply return None. This is the primary way of handling errors in V. They are still values, like in Go, but the advantage is that errors can’t be unhandled, and handling them is a lot less verbose. Safety This language has no such things as null, global variables, undefined values, undefined behaviour, variable shadowing or bounds checking. The language supports option\/result types, generics, immutable variables by default, pure functions, and immutable structs. Performance According to Medvednikov, this language is as fast as C, minimal amount of allocations, built-in serialization without runtime reflection, C interop without any costs along with compilation to native binaries without any dependencies. Only 400 KiloBytes Compiler With Zero Dependencies Currently, the V compiler does have one dependency which is a C compiler. The entire V language and its standard library are less than 400 KiloBytes. V is written in V, and you can build this language in 0.4 seconds. The creator also ensures to drop this number to ≈0.15 seconds by the end of this year. C\/C++ Translation You can easily translate your V code to C or C++ where Clang parser is used for translating C\/C++ to V. Also, this feature is at an early stage and Medvednikov aims to make this feature stable by the end of this year. Comparison Of V With Other Languages Variable Declaration In V, variables are declared and initialised with “:=” while “=” is used for assigning or to change the value of the variable. This is the only way to declare variables in V which means that the variables will always have an initial value. Unlike most other programming languages, the language V only allows defining variables in functions. Switch And Break Unlike switch and break statements in C and C++, this language does not require the break statement at the end of every block. Maintainable According to the creator, V is a simpler language than C++ and it offers up to 400 times faster compilation, safety, lack of undefined behaviour, easy concurrency, compile-time code generation, etc. Also, compared to Python language, this language is much faster, simpler, safer, more maintainable, etc. Note: The article draws insights from the official documentation. For more information on V, please click here.","excerpt":"A new statically typed compiled programming language, known as V language has been open-sourced recently by its creator Alex Medvednikov. According to Medvednikov, the language is designed for building maintainable software. It is a simple language and is similar to the Go language and is inspired by Rust, Swift, and Oberon languages. Currently, V emits […]","categories":["AI Features"],"tags":["Go","Python","Rust"],"author_name":"Ambika Choudhury","publish_date":"2019-07-03T13:56:53","publication_year":"2019","word_count":612,"keywords":["Go","TPU","AWS","AI","cloud_platforms:AWS","Python","Aim","C++","Rust","R"],"extracted_tech_keywords":["AI","Aim","AWS","TPU","Python","R","Go","Rust","C++","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-v-the-new-statistically-typed-programming-language-inspired-by-go-rust\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10078654,"title":"Get paid to earn your business analytics degree in the USA. Join this webinar to learn more!","content":"Considering the market worldwide is growing for careers in data & analytics, learning to transform real data into actionable business strategies is a smart choice for professionals and students alike. And, it can’t get better when you can study business analytics at your dream destination — the USA! Not only the USA has been one of the best study-abroad destinations for international students, but also the demand for international workers who have STEM degrees is massive among US businesses, which is only expected to increase for the foreseeable future. On that note, the University of Louisville is now offering a Master of Science in Business Analytics that allows Indian students to get access to a STEM-certified 13-month programme. With world-class faculty, including teachers\/scholars with extensive professional experience, the MSBA programme also allows a competitive internship opportunity that provides real-world business analytics experience for students. This, in turn, allows students to offset some of the tuition costs and the added workplace experience provides significant advantages to their marketability upon graduation. To better understand the value of STEM-designated business analytics degrees in the US, we are organising a webinar with the University of Louisville College of Business. To be held on November 23, 2022, from 6:00 to 7:00 PM (IST), the session will be hosted by Dr Sandeep Goyal, the programme director at the University of Louisville. What to expect?  The session will be focused on individuals who seek to leverage the increasing value of big data and analytics on the world economic stage. The seminar attendees will learn about the professional opportunities for international students in this industry in the US, along with the benefits of a graduate degree from the University of Louisville College of Business. This one-year MS programme in Business Analytics also offers competitive-based paid internships.    These practical internships will allow students to pursue their interests without putting their careers on hold. In addition, with this programme, they will be able to work and study simultaneously in the US. Date:November 23, 20226:00 PM onwardsSAVE YOUR SPOT Roles you can explore: – Business Analytics– Business Operations Manager– Business Intelligence– Information Security Analyst– Management Analyst– Market Research Data– Engineer– Operations Research Analyst– Forensic Accountant– Social Science Data Analyst– Epidemiologist To learn more about the Master of Science Business Analytics program, click here And to understand the value of STEM-designated business analytics degrees in the US, register for this webinar.","excerpt":"The seminar attendees will learn about the professional opportunities for international students in this industry in the US, as well as the benefits of a graduate degree from the University of Louisville College of Business.","categories":["AI Trends"],"tags":["Business Analytics","USA"],"author_name":"Poulomi Chatterjee","publish_date":"2022-11-04T14:00:19","publication_year":"2022","word_count":399,"keywords":["big data","Go","business intelligence","programming_languages:R","AI","programming_languages:Go","RAG","analytics","Business Analytics","GAN","R","USA"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","GAN","business intelligence","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/get-paid-to-earn-your-business-analytics-degree-in-the-usa-join-this-webinar-to-learn-more\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10039390,"title":"Guide To Zooming Slow-Mo: One-Stage Space-Time Video Super-Resolution","content":"Space-time video super-resolution is a computer vision task that aims to increase video resolution both in time and space. It creates a high-resolution slow-motion video from a low frame rate, low-resolution video. It can be divided into two sub-tasks: video frame interpolation (VFI) for generating intermediate video frames and video super-resolution(VSR).  Existing state-of-the-art two-stage methods use large frame-synthesis modules to predict high-resolution frames; this leads to higher computational complexity and can be time-consuming. Furthermore,  frame interpolation and spatial super-resolution sub-tasks are intra-related and carry coupled information that can help speed up both. To overcome these problems, Xiaoyu Xiang, Yapeng Tian proposed a new one-stage approach for space-time video super-resolution- Zooming Slow-Mo. Zooming Slow-Mo directly synthesizes a high-resolution slow-motion video from a low frame rate, low-resolution video. It temporally interpolates low-resolution frame features in missing low-resolution video frames utilizing the local temporal contexts via a feature temporal interpolation network. Then a deep reconstruction network is used to generate high-resolution slow-motion video frames. Architecture & Approach The Zooming Slow-Mo framework consists of four main parts: feature extractor, a frame feature temporal interpolation module, deformable ConvLSTM, and a deep frame reconstruction network. The feature extraction module consists of one convolution layer followed by five residual blocks. It extracts feature map:  {FL2t−1}n+1t=1 from input video frames. The proposed frame feature interpolation network then uses these feature maps to generate the low-resolution feature maps of the non-existent intermediate frames. Frame Feature Temporal Interpolation Given two feature maps: FL1 and FL3, from low-res input video frames: IL1 and IL3, the aim is to synthesize the feature map FL2 of the missing intermediate frame IL2. Existing frame interpolation networks perform temporal interpolation on pixel-wise video frames. This leads to a two-stage STVSR design. In contrast, Zooming Slow-Mo learns a feature temporal interpolation function f(·) to directly synthesize the intermediate feature maps. This interpolation function can be formulated as: Here T1(·) and T3(·) are two sampling functions, ????1 and ????3 are the corresponding sampling parameters; and H(·) is a blending function for aggregating sampled features. To generate accurate intermediate feature maps, T1(·) needs to capture the forward motion information between FL1 and FL2, and T3(·) needs to capture backward motion information between FL3 and FL2. However, FL2  does not exist. The information flow between FL1 and FL3 is used to approximate the backward and forward motion information to work around this issue.  A linear blending function is used to combine the two sampled feature maps: Here ???? and ???? are learnable 1 × 1 convolution kernels and ∗ is the convolution operator. Deformable ConvLSTM Temporal information is essential in video restoration tasks. Therefore, instead of reconstructing high-res frames from individual feature maps, Zooming Slow-Mo aggregates temporal contexts from neighbouring frames. It employs ConvLSTM, a popular 2D sequence data modelling method,  to perform temporal aggregation. However, ConvLSTM can only capture motion between previous states and the current input feature map with small convolution receptive fields. This greatly limits ConvLSTMs ability to handle large motions. When working with videos with large motions, this leads to a severe temporal mismatch between previous states and the current feature map FLt.  As a result, the reconstructed high-resolution frame IHt suffers from artifacting. To overcome this problem and make better use of the global temporal contexts, a state updating cell with deformable alignment is embedded into ConvLSTM: gh and gc are used to denote the general functions of several convolution layers, ∆pht and ∆pct are the learned offsets.  hat−1 and cat−1 refer to the current feature map FLt aligned hidden and cell states respectively.In contrast to vanilla ConvLSTM, deformable ConvLSTM enforces the hidden state and cell state to align with the current feature map FLt. In addition to that, the Deformable ConvLSTM is used in a bidirectional manner to maximize the utilization of temporal information. Frame Reconstruction A temporally shared synthesis network is used for frame reconstruction; it synthesizes high-resolution frames from individual hidden states ht. The reconstruction network has 40 stacked residual blocks for learning deep features and uses PixelShuffle for sub-pixel upscaling to reconstruct high-res frames. A reconstruction loss function is used to optimize this network: Here IGTt denotes the t-th ground-truth high-res video frame and ???? is set to 1 × 10−3. Space-Time Video Super-Resolution using Zooming Slow-Mo Clone the Zooming Slow-Mo GitHub repository. git clone --recursive https:\/\/github.com\/Mukosame\/Zooming-Slow-Mo-CVPR-2020.git Install OpenCV, PyTorch and other requirements. pip install -r requirements.txt Compile the deformable convolutional network V2 cd $ZOOMING_ROOT\/codes\/models\/modules\/DCNv2 bash make.sh Perform space-time video super-resolution using the video_to_zsm.py script. python codes\/video_to_zsm.py --model experiments\/pretrained_models\/xiang2020zooming.pth --video low-res-vid.mp4 --output low-res-vid.mp4 --N_out 3 Code Source Last Epoch This article introduced Zooming Slow-Mo, a one-stage framework for space-time video super-resolution that directly synthesizes high frame rate, high-resolution videos without generating the intermediate low-resolution frames. It introduces a deformable feature interpolation network that enables feature-level temporal interpolation. Furthermore, it uses a modified deformable ConvLSTM for aggregating temporal information and handling large motions. Using its one-stage design, Zooming Slow-mo is able to make use of the intra-relatedness between temporal interpolation and spatial super-resolution. It outperforms existing state-of-the-art two-stage approaches not only in terms of effectiveness but also efficiency. For a more in-depth understanding of Zooming Slow-Mo, refer to the following resources: PaperGitHub","excerpt":"Zooming Slow-Mo is a one-stage framework for space-time video super-resolution that directly synthesizes high frame rate, high-res videos without generating the intermediate low-res frames.","categories":["Deep Tech"],"tags":["Guide"],"author_name":"Aditya Singh","publish_date":"2021-05-03T15:00:00","publication_year":"2021","word_count":861,"keywords":["TPU","PyTorch","AI","Git","computer vision","OpenCV","Python","Aim","GitHub","R","Guide"],"extracted_tech_keywords":["AI","computer vision","Aim","PyTorch","OpenCV","TPU","Python","R","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-zooming-slow-mo-one-stage-space-time-video-super-resolution\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10099808,"title":"Are Innovators Moving Away from Silicon Valley?","content":"While Sam Altman is pretty much known to throw in generous pearls of wisdom in his interviews, a recent podcast made the OpenAI CEO talk blatantly about an uncomfortable truth about Silicon Valley. “I hate to say this because it sounds so arrogant, but before OpenAI, what was the last great scientific breakthrough to came out of a Silicon Valley company?” said Altman in a recent interview with Nicolai Tangen, CEO of Norges Bank Investment Management. Let’s say, Altman is right. So, what transformation has the most sought-after tech hub in the world undergone to curb the “innovation decline” in Silicon Valley? Chasing Money In the race to bag bigger funding, companies have been prioritising quicker returns. In the process, an innovation decline is bound to happen. Altman quipped that though he may not completely understand why this has happened, he believes that it has gotten easy to make a valuable company that people start getting impatient with when it comes to timelines and returns. “A lot of the capital went into things that could fairly reliably multiply money in a short period of time,” he said. Using existing technology such as the internet and mobile phones to revolutionise industries was compelling enough to attract talent to build such companies. A discussion on Hacker News emphasises the dynamics of venture capital in the startup world. With VC firms seeking high returns with minimal risks, founders are often advised to focus on B2B markets over consumer markets. Furthermore, for more than a decade, the focus has been on building SaaS businesses, and the emphasis has been doing something simple that guarantees returns, rather than on thinking big to tackle challenging things. Rise of Dragon and Falcon What was believed to be the hot centre for tech innovation in the world, is now slowly changing owing to the global players. The rise of China’s prowess in the technology world has pushed the country as a worthy competitor to the US. Over the past twenty years, China has transformed from being a ‘technological backwater to an innovation powerhouse‘. Despite tight government control on markets, the country is now competing with the US on the latest emerging tech, including AI and quantum computing. Looking at the latest developments across the globe, countries such as UAE, where research institutes have government funding, are building LLMs such as Falcon, proving to be a close competitor to GPT-4. Further implying that the US is no more the concentric region for springing forth competing tech. Fading Sparkle Silicon Valley’s innovation boom in the 1970s and 1980s was fueled by government funding and the establishment of major VC firms Kleiner Perkins and Sequoia Capital in 1972. The IT revolution during the same period further pushed the Silicon Valley status. Tech giants such as Apple, Microsoft, Atari, Oracle, Cisco, and others witnessed a rise in the same period. Furthermore, Apple’s 1981 IPO generated $1.3 billion solidifying Silicon Valley’s status as the global hub for VC companies. However, the shine is slowly fading. Owing to a number of reasons, the last one being the pandemic, a huge shift in base is happening. Inflation, coupled with an already exorbitant cost of living, have rendered the place liveable for only those with hefty salaries. Wealth inequality in Silicon Valley is more pronounced than any other place across the country. Furthermore, other cities across the US are emerging as new startup hot-beds, inviting VCs to aggressively fund in these regions. Company fundings in Miami quadrupled in the past three years with a total of $5.39 billion in 2022 alone. New York is witnessing an increase in tech talent and the ones most likely to shift from the West Coast are those who are five to ten years into their career. Source: Washington Post While Altman may have brought focus back to the Silicon Valley innovation decline, some were quick to criticise him for considering OpenAI as the “only innovator”. The trivialisation attitude of the founder belittled the work of those who work on low-level stuff.  “I think there is a point here that user-facing innovation stagnated and OpenAI helped break that, but it’s wild to me that there is no acknowledgement at all of the giants whose shoulders they stand on,” said a user on Hacker News.","excerpt":"Has Silicon Valley really lost its culture of innovation or only Sam Altman feels this way?","categories":["AI Features"],"tags":["Apple","China","investment","investors","IPO","OpenAI","SaaS","Sam Altman","silicon valley","uae","VC"],"author_name":"Vandana Nair","publish_date":"2023-09-11T16:11:36","publication_year":"2023","word_count":713,"keywords":["API","R","China","Sam Altman","investors","RAG","ViT","VC","Go","AI","Apple","silicon valley","GAN","investment","uae","OpenAI","IPO","innovation","GPT","SaaS"],"extracted_tech_keywords":["AI","OpenAI","RAG","R","Go","API","GPT","GAN","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/are-innovators-moving-away-from-silicon-valley\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":40886,"title":"Teaching Artificial Intelligence To Connect Senses Like Vision And Touch","content":"MIT CSAIL system can learn to see by touching and feel by seeing, suggesting future where robots can more easily grasp and recognize objects. In Canadian author Margaret Atwood’s book “The Blind Assassin,” she says that “touch comes before sight, before speech. It’s the first language and the last, and it always tells the truth.” While our sense of touch gives us a channel to feel the physical world, our eyes help us immediately understand the full picture of these tactile signals. Robots that have been programmed to see or feel can’t use these signals quite as interchangeably. To better bridge this sensory gap, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have come up with a predictive artificial intelligence (AI) that can learn to see by touching, and learn to feel by seeing. The team’s system can create realistic tactile signals from visual inputs, and predict which object and what part is being touched directly from those tactile inputs. They used a KUKA robot arm with a special tactile sensor called GelSight, designed by another group at MIT. Using a simple web camera, the team recorded nearly 200 objects, such as tools, household products, fabrics, and more, being touched more than 12,000 times. Breaking those 12,000 video clips down into static frames, the team compiled “VisGel,” a dataset of more than 3 million visual\/tactile-paired images. “By looking at the scene, our model can imagine the feeling of touching a flat surface or a sharp edge”, says Yunzhu Li, CSAIL PhD student and lead author on a new paper about the system. “By blindly touching around, our model can predict the interaction with the environment purely from tactile feelings. Bringing these two senses together could empower the robot and reduce the data we might need for tasks involving manipulating and grasping objects.” Recent work to equip robots with more human-like physical senses, such as MIT’s 2016 project using deep learning to visually indicate sounds, or a model that predicts objects’ responses to physical forces, both use large datasets that aren’t available for understanding interactions between vision and touch. The team’s technique gets around this by using the VisGel dataset, and something called generative adversarial networks (GANs). GANs use visual or tactile images to generate images in the other modality. They work by using a “generator” and a “discriminator” that compete with each other, where the generator aims to create real-looking images to fool the discriminator. Every time the discriminator “catches” the generator, it has to expose the internal reasoning for the decision, which allows the generator to repeatedly improve itself. Vision to touch Humans can infer how an object feels just by seeing it. To better give machines this power, the system first had to locate the position of the touch, and then deduce information about the shape and feel of the region. The reference images — without any robot-object interaction — helped the system encode details about the objects and the environment. Then, when the robot arm was operating, the model could simply compare the current frame with its reference image, and easily identify the location and scale of the touch. This might look something like feeding the system an image of a computer mouse, and then “seeing” the area where the model predicts the object should be touched for pickup — which could vastly help machines plan safer and more efficient actions. Touch to vision For touch to vision, the aim was for the model to produce a visual image based on tactile data. The model analyzed a tactile image, and then figured out the shape and material of the contact position. It then looked back to the reference image to “hallucinate” the interaction. For example, if during testing the model was fed tactile data on a shoe, it could produce an image of where that shoe was most likely to be touched. This type of ability could be helpful for accomplishing tasks in cases where there’s no visual data, like when a light is off, or if a person is blindly reaching into a box or unknown area. Looking ahead The current dataset only has examples of interactions in a controlled environment. The team hopes to improve this by collecting data in more unstructured areas, or by using a new MIT-designed tactile glove, to better increase the size and diversity of the dataset. There are still details that can be tricky to infer from switching modes, like telling the color of an object by just touching it, or telling how soft a sofa is without actually pressing on it. The researchers say this could be improved by creating more robust models for uncertainty, to expand the distribution of possible outcomes. In the future, this type of model could help with a more harmonious relationship between vision and robotics, especially for object recognition, grasping, better scene understanding, and helping with seamless human-robot integration in an assistive or manufacturing setting. “This is the first method that can convincingly translate between visual and touch signals”, says Andrew Owens, a postdoc at the University of California at Berkeley. “Methods like this have the potential to be very useful for robotics, where you need to answer questions like ‘is this object hard or soft?’, or ‘if I lift this mug by its handle, how good will my grip be?’ This is a very challenging problem, since the signals are so different, and this model has demonstrated great capability.” Li wrote the paper alongside MIT professors Russ Tedrake and Antonio Torralba, and MIT postdoc Jun-Yan Zhu. It will be presented this week at The Conference on Computer Vision and Pattern Recognition in Long Beach, California.","excerpt":"MIT CSAIL system can learn to see by touching and feel by seeing, suggesting future where robots can more easily grasp and recognize objects. In Canadian author Margaret Atwood’s book “The Blind Assassin,” she says that “touch comes before sight, before speech. It’s the first language and the last, and it always tells the truth.” While […]","categories":["AI Features"],"tags":["Computer Vision","human intelligence at machine scale","MIT","model","Research"],"author_name":"AIM Media House","publish_date":"2019-06-18T09:44:31","publication_year":"2019","word_count":944,"keywords":["Go","artificial intelligence","human intelligence at machine scale","AI","MIT","ML","computer vision","model","deep learning","Research","Aim","Computer Vision","CLIP","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","computer vision","Aim","R","Go","CLIP","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/teaching-artificial-intelligence-to-connect-senses-like-vision-and-touch\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10134367,"title":"PwC Collaborates with Shorthills AI to Integrate GenAI Search and Summarisation","content":"In a strategic partnership, PwC will include advanced Shorthills AI’s generative AI features into its suite of tools. The aim is to leverage data and AI, streamline operations, and drive return on investment. Under this partnership, PwC will deploy Shorthills AI’s GenAI solutions, including GenAI-powered search, automatic classification of entries, summarization of large documents, and natural language-based Q&A. Shorthills AI has a proven methodology of training and fine-tuning modern LLMs with custom data. Since its inception in 2018, the founders Pawan Prabhat and Paramdeep Singh aim to leverage AI to solve complex business challenges. “Collaborating with PwC allows us to showcase the transformative power of our GenAI solutions,” said Paramdeep Singh, co-Founder, Shorthills AI. PwC has consistently invested in emerging AI technologies to stay ahead in the competitive market. This partnership with Shorthills AI is a testament to PwC’s dedication to adopting the latest advancements in AI to enhance its service delivery. By employing their expertise, PwC will be able to offer more personalised services to its clients making it a leader in the consulting  space. “We look forward to partnering with Shorthills AI to bring innovative GenAI technologies to our clients. This partnership will enable us to deliver even greater value to our clients through enhanced efficiency and personalised solutions,” said Siddharth Mehta, Partner, PwC India. PwC’s legacy in integrating AI into its solutions help them revolutionise the consulting world. Previously, PwC formed alliances with major AI technology vendors, including foundation model providers AWS, Anthropic, Google, Meta, Microsoft, and OpenAI. Earlier last month, PwC India inaugurated its GenAI Experience Lab in Gurgaon, in a bid to explore Generative AI’s transformative capabilities. They also teamed up with Microsoft India to bolster its incident response and recovery capabilities.to tackle cyber threats within organisations.","excerpt":"PwC has formed alliances with AWS, Anthropic, Google, Meta, Microsoft, and OpenAI.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Aditi Suresh","publish_date":"2024-09-04T09:14:56","publication_year":"2024","word_count":292,"keywords":["Anthropic","GenAI","OpenAI","AI","AWS","ML","RAG","Aim","generative AI","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","OpenAI","Anthropic","Aim","RAG","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pwc-collaborates-with-shorthills-ai-to-integrate-genai-search-and-summarisation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10131897,"title":"LG AI Research Launches EXAONE 3.0: A Bilingual Language Model for Open Research","content":"LG AI Research has announced the release of EXAONE 3.0, a 7.8 billion parameter instruction-tuned language model, marking a significant advancement in the field of large language models (LLMs). This is the first open model in the EXAONE family, aimed at democratising access to expert-level artificial intelligence capabilities. The release is intended to foster innovation and collaboration within the AI community by providing a high-performance model for non-commercial research purposes. EXAONE 3.0 has been extensively evaluated across a variety of benchmarks, demonstrating competitive real-world performance and instruction-following capabilities. It excels particularly in Korean, while also achieving strong results in English across general tasks and complex reasoning. The model’s architecture is based on a decoder-only transformer with advanced features like rotary position embeddings and grouped query attention, supporting a maximum context length of 4,096 tokens. The model was trained using a diverse dataset, ensuring robust performance in real-world scenarios. It features a bilingual tokeniser designed to optimise performance in both English and Korean, addressing the linguistic complexities of the Korean language. The training process involved extensive pre-training and post-training techniques, including supervised fine-tuning and direct preference optimisation, to enhance the model’s instruction-following capabilities. The EXAONE 3.0 model is available on Hugging Face for research purposes, supporting the broader AI community’s efforts in developing innovative applications. LG AI Research aims to integrate advanced AI into everyday life, making expert knowledge accessible to a wider audience. The model is expected to contribute significantly to advancements in Expert AI, particularly in bilingual environments. LG AI Research has emphasised the importance of responsible AI development, conducting thorough compliance reviews and ethical assessments to minimise risks associated with data usage. The model has been evaluated for potential social and ethical issues, with measures in place to ensure its safe and ethical deployment.","excerpt":"This model is the pioneering open-access member of the EXAONE family, designed to democratize expert-level AI capabilities.","categories":["AI News"],"tags":["language model"],"author_name":"Gopika Raj","publish_date":"2024-08-08T16:32:57","publication_year":"2024","word_count":296,"keywords":["Hugging Face","artificial intelligence","programming_languages:R","AI","innovation","language model","responsible AI","Aim","ai_frameworks:Hugging Face","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","Hugging Face","R","responsible AI","innovation","AI research","ai_frameworks:Hugging Face","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/lg-ai-research-launches-exaone-3-0-a-bilingual-language-model-for-open-research\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":29361,"title":"Infosys On A Digital Path With AI And Automation, Says CEO Salil Parekh","content":"With US President Donald Trump’s “buy American, hire American” slogan still ringing in Indian tech companies’ boardrooms, Infosys CEO Salil Parekh on Wednesday said that they were on a path to growth and digital transformation. “We have set ourselves at transformation plans. On that, we have started a journey in the first two quarters, where we have started to track several areas. Our company’s focus clearly is on digital and improving cost services with the help of artificial intelligence and automation… There is a lot more work, however, the company needs to do for overall transformation,” Parekh told a noted newswire. He added that Infosys had started with some recruitments from colleges in the US and will do the same for Europe and Australia. He also said that overall, he thought that Infosys is seeing a positive change in their strategy of agile digiting and automation. “Our AI and cost services are also resonating well with their clients and building more relevance,” he added. #InfosysQ2FY19 INR: Net profit was Rs. 4,110 crores for the quarter ended Sep 30, 2018; YoY growth was 10.3% https:\/\/t.co\/GsYSRsYwT7 — Infosys (@Infosys) October 16, 2018 Infosys this week reported their second quarter results with a rise in consolidated net profit at 10.3% year-on-year to ₹4,110 crore.","excerpt":"With US President Donald Trump’s “buy American, hire American” slogan still ringing in Indian tech companies’ boardrooms, Infosys CEO Salil Parekh on Wednesday said that they were on a path to growth and digital transformation. “We have set ourselves at transformation plans. On that, we have started a journey in the first two quarters, […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Infosys","Salil Parekh"],"author_name":"Prajakta Hebbar","publish_date":"2018-10-17T12:18:45","publication_year":"2018","word_count":211,"keywords":["artificial intelligence","Infosys","AI","programming_languages:R","digital transformation","Git","GAN","automation","Salil Parekh","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","Git","GAN","digital transformation","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infosys-on-a-digital-path-with-ai-and-automation-says-says-ceo-salil-parekh\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10079099,"title":"Microsoft to Pump More Funds into OpenAI: What’ll be the Implication?","content":"According to recent media reports, Microsoft is in ‘advanced talks’ for a new round of funding in OpenAI. The details of the deal are still under discussion, and the funding amount is being negotiated. The fund raised is expected to be channelled to further incorporate AI in OpenAI’s products. OpenAI has had a close association with Microsoft in the past few years, one of the most significant ones being the latter’s investment of USD 1 billion in the AI research lab in 2019. OpenAI & Microsoft’s relationship OpenAI was founded as a non-profit in 2015 with a total pledged amount of USD 1 billion from tech leaders and investors like Elon Musk, and Reid Hoffman, among others. In his book ‘Genius Makers’, author Cade Metz writes that OpenAI was built on a vision to have a top-class research lab ‘entirely free of corporate pressures’ so that ‘anyone could compete with the Googles and Facebooks of the world’. But something changed in 2019 when Microsoft made the investment in the lab, and a for-profit company was carved out of it. In 2020, Microsoft acquired exclusive rights to the GPT-3 language model, launched by OpenAI in the same year. This drew a lot of flak from the developer and researcher communities, which protested OpenAI’s decision to not open source the language model. Born out of an association between Microsoft and OpenAI were other tools and systems – like the Azure AI supercomputing technologies and the GitHub Copilot. In 2021, Microsoft launched Azure OpenAI Service at the former’s annual event called Microsoft Ignite. With this, Azure customers could now utilise OpenAI’s machine learning models. “The new Azure Cognitive Service will give customers access to OpenAI’s powerful GPT-3 models, along with security, reliability, compliance, data privacy and other enterprise-grade capabilities that are built into Microsoft Azure,” one of Microsoft’s blogs noted. The OpenAI partnership cemented Microsoft’s Azure cloud infrastructure’s position as the platform of choice for the next generation of  AI supercomputing applications. OpenAI is not the only major AI lab to get corporate backing. In 2014, DeepMind was acquired by Alphabet (Google’s parent company) for more than USD 500 million. DeepMind was founded by Demi Hassabis, Shane Legg, and Mustafa Suleyman and already had Skype developer Jaan Tallin as one of DeepMind’s major investors. Independent and open source While it is good to see corporates’ investment in the progress of AI research, which is highly resource-intensive, it puts a question mark on the autonomy of such institutions. This was well-illustrated in OpenAI-Microsoft’s case. Following Microsoft’s investment, OpenAI CEO Sam Altman’s internal messaging reportedly emphasised on the need to commercialise the research lab’s technologies to support the work; we saw this with the exclusive license of GPT-3 being granted to Microsoft, which is now likely to be extended to its successor – GPT-4. This example demonstrates a trend in the AI community that researchers have been repeatedly warning about – the increasing cost of AI research resulting in the concentration of AI prowess in the hands of a few rich Big-Tech companies. Meanwhile, a new term ‘AI Capitalism’ has also found existence. A 2022 paper by AI & Society characterises it as the commodification of “data, data extraction and a concentration in the hiring of AI talent and compute capacity”. It is directly linked to Big Tech’s insatiable appetite for growth and eventual monopolisation. Since data and computing resources for AI research are concentrated in the hands of a few companies, it creates a dependency on them. In a paper, ‘The Sheep Cost of Capture’, author Meredith Whittaker (co-founder of AI Now Institute) compared the relationship between big tech and AI academia to the US military’s influence over scientific research during the Cold War. She also wrote that AI advances in the last decade have resulted mostly from the concentration of data and computing resources in the hands of large corporations. Not all is lost, yet Modern AI technologies are the end product of shared foundational knowledge like code, data sets, pre-build machine learning models, and data architectures to create AI tools and products. This has been the basis for progress in ML, deep learning, computer vision, and other technologies. This shared, community-driven and open-source technology has even managed to transcend geopolitical situations. Case in point, Facebook’s PyTorch project, which was recently launched as PyTorch Foundation to ensure that the decisions related to AI research would be made in a ‘transparent and open manner’, the company blog read. At that time, Soumith Chintala, the creator of PyTorch, told Protocol, “I’m definitely surprised at how much [of the] general global considerations that you would have from a business angle don’t really come in when you’re talking about open-source collaboration, especially with AI.” He further added, “Within open-source software, the political system doesn’t even start entering into play [until] much later. People are mostly trying to learn from each other, build the best thing they can.” AI owes a lot of progress to the open source and the collaborative nature of most of the underlying technologies.","excerpt":"OpenAI has had a close association with Microsoft in the past few years, one of the most significant ones being the latter’s investment of USD 1 billion in OpenAI","categories":["Global Tech"],"tags":["GPT-3","LLMs","Microsoft","OpenAI"],"author_name":"Shraddha Goled","publish_date":"2022-11-08T18:04:03","publication_year":"2022","word_count":840,"keywords":["GPT-3","Go","machine learning","OpenAI","AI","LLMs","PyTorch","ML","R","computer vision","deep learning","Azure","Microsoft"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","OpenAI","PyTorch","Azure","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-to-pump-more-funds-into-openai-what-will-be-the-implications\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":48115,"title":"In Conversation With CRIF’s Atrideb Basu & How He Scaled Data & Analytics Practice In India","content":"CRIF is an Italian company, based out of Bologna and caters to more than 6K banks and financial institutions across 30+ countries globally. In India, CRIF has two broad arms – one is the 100% subsidiary (called CRIF Solutions India) – which focuses on consulting, analytics and products and the other is the credit bureau where it has major ownership (CRIF Highmark) – which focuses on multiple services on retail and commercial side. In 2015, CRIF set up the Centre of Excellence (CoE) for analytics and decision engine which provides solutions and services that covers the entire customer lifecycle —  risk assessment, marketing, acquisition, customer management and collections.Today the 85-strong analytics arm, headquartered in Pune is a one-stop-shop for all data and analytics related needs for the BFSI and telecom sector across Asia and the Middle East. AIM caught up with Atrideb Basu, Senior Vice President – Consulting & Products, Asia & Middle East who set up the analytics practice in India to understand how he put in place people, process and technology at CRIF India. Basu also talks about the rapid growth of analytics practice and consulting that supports several regions, localization of analytics in India and scaling the team from 10 to 85 in a span of few years. Read the full transcript of the interview to find out more: Analytics India Magazine: Can you tell us about your role in India and how has the journey been so far? Atrideb Basu: It’s been a fantastic roller coaster ride. I am sure my boss hired me because I drive a Punto (that’s when you really think what are the odds). When I joined in 2015 – the concept of building an analytics unit in India was new to CRIF – so we had to first convince within the company that we can develop a team in India to deliver – and then grow from there. CRIF has been historically very strong in credit bureau scorecard development, we understood this relatively quickly and added our own punch on the top of that. Think of it as a paneer pizza and implemented it in Dubai and India. From there, we never really looked back. We scaled rapidly from 0 to 85 resources in 4 years and have achieved more than €3 million as revenues in the region. AIM: Can you tell us about your vision and how you set up the CoE in India wrt approach and technology? AB: We started with a small team and tested certain aspects. We were successful. Then we scaled up a bit and tested some more. Luckily, it was also OK. Then I think the momentum just catches up. The key aspect remains that you focus on the basics like problem solving and execution and have fun with the extended team. But there are a few things which we did were important. We didn’t want everyone to think alike – so we hired folks with diverse background with experience in policy, risk scorecards, marketing scorecards, fraud, collections and then in specific products like personal loans, credit cards, two-wheelers, business loans, affordable housing etc so that we can bring diverse experiences together. This helped in two important ways  – firstly resource engagement was a given as everyone is doing something new and secondly – our thoughts were naturally out of the box. Everyone pitched in with their ideas to solve something and created a slightly differentiated offering Another example is that by design we do not separate the delivery team and the pre-sales team. Typically, if you see in most consulting organisations, the delivery team and the pre-sales team are separated. The same guy who is building the solution is also the same person who’s pitching the technical content to the client which allows them to connect better and simplify the problem. The last is being technology agnostic. So, to us, it does not matter if we need to use any statistical language – like SAS\/Python\/R to solve a problem. The team has worked very hard in this area. AIM: What are the kind of analytics products tailored for India? AB: There are multiple products\/solutions of course. When a company like CRIF opens a Centre of Excellence in India – it brings a full suite of its global expertise – sees what can be relevant – what needs to be localised – and then does it. The objective is to always keep the solutions simple and relevant – despite the multiple complexities that might have gone into development. The first thing we did is to enhance the scores on the bureau platform – both for consumer and SMEs – on the fundamentals of stability and precision – as anything else might be disastrous. We spend a considerable amount of time (more than 12 months) to develop these solutions and tested it with multiple players – so that we were sure it adds value. We are committed to do many more things on the bureau data itself – creating open-box consulting solutions, bringing in specialised scores to allow better underwriting, and more importantly, changing the mindset of our customers where we are consultants that solves business problems by “developing strategies and not just a team that gives a score”. The second aspect we work towards is customised solutions. This helps in multiple ways. One, it allows the customer to have a solution on a population that it is focused towards (so solves the right data issue) and two, it creates a more focused algorithm that gives the customer a competitive advantage over other players ( for example it can help you underwrite a segment that most of the other players will not). The third aspect we focus on is working directly with clients on their data: be it optimisation of policy, development of bespoke scorecards (application or behavioral or collection scorecards), optimising the right loan amount to be given, deciding on the right way for deciding on cut-offs, whom and when to cross-sell and then finally presenting all this to their board to a faster internal approval. AIM: Of late there has been a rise in alternate data utilised by fintechs for P2P lending. Can you share your thoughts on this trend and how is this set to grow? AB: Of course. Having more data leads to the first degree of competitive advantage (eg. new to credit customers) and then using it effectively should lead to the second (which also means storing the data properly). But the banks will accelerate to catch up and we will reach a stage – where the delta comes out – “what is the real advantage that alternate data can give over and above tradition underwriting data” — I mean we really would not trust if someone told us 5 years ago that the higher the ratio of # of selfies vs other camera shots, higher is the 90+ days past due likelihood. But today, we may not reject this outright — and will try to empirically test it out and see if we can justify this by some psychometric traits. So, yes, the possibilities have expanded, and it is exciting. Second is the mindset of empirics over experience. If someone does not believe this (and for sure there will be plenty), please look at some examples like Leontief paradox – where empirical studies can overrule proven theories. The point is – create a blind testing facility, fail multiple times, and fail very fast, and then the iterations towards a new logic\/path\/way is faster. This is the only way we grow now with the proliferation of technology and data. The only caveat is that this needs to be managed in a profitable way😊. AIM: Can you talk about CRIF’s role in Regulatory Analytics and how you are delivering a comprehensive approach to clients? AB: CRIF is very strong globally in Basel2 and IFRS9 projects. As I had mentioned earlier, this unit is set-up to bring the full global expertise to India and IFRS9 is no exception. We have started working on IFRS9 in Philippines, Indonesia and Lebanon and soon will work in India too. We focus very sharply on the expected loss computations in this space and been successful in lowering the ECL% for the banks while increasing the robustness of the methodology. I can openly say, it is very difficult to compete with this methodology – so once we start deploying this in the market – and players ready to adopt something new — it will be disruptive.","excerpt":"CRIF is an Italian company, based out of Bologna and caters to more than 6K banks and financial institutions across 30+ countries globally. In India, CRIF has two broad arms – one is the 100% subsidiary (called CRIF Solutions India) – which focuses on consulting, analytics and products and the other is the credit bureau […]","categories":["AI Features"],"tags":["Interviews and Discussions","robust analytics strategy"],"author_name":"Richa Bhatia","publish_date":"2019-10-17T11:24:00","publication_year":"2019","word_count":1418,"keywords":["Go","API","ELT","AI","Python","robust analytics strategy","Aim","analytics","Rust","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","Python","R","Go","Rust","API","ELT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/in-conversation-with-crifs-atrideb-basu-how-he-scaled-data-analytics-practice-in-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":197,"title":"Machine Learning For Better And More Efficient Solar Power Plants","content":"The global solar photovoltaic (PV) installed capacity in 2013 was 138.9 GW and it is expected to grow to over 455 GW by 2020. However, solar power plants still have a number of limitations that prevent it from being used on a larger scale. One limitation is that the power generation cannot be fully controlled or planned for in advance since the energy output from solar power plants is variable and prone to fluctuations dependent on the intensity of solar radiation, cloud cover and other factors. Another important limitation is that solar energy is only available during the day and batteries are still not an economically viable storage option making careful management of energy generation necessary. Additionally, as the installed capacity of solar power plants grows and plants are increasingly installed at remote locations where location data is not readily available, it is becoming necessary to determine their optimal sizes, locations and configurations using other methods. Machine learning techniques provide solutions that have been more successful in addressing these challenges than manually developed specialized models. Accurate forecasts of solar power production are a necessary factor in making the renewable energy technology a cost-effective and viable energy source. Machine learning techniques can correctly forecast solar power plant generation at a better rate than current specialized solar forecasting methods. In a study conducted by Sharma et al, multiple regression techniques including least-square support vector machines (SVM) using multiple kernel functions were used in the comparison with other models to develop a site specific prediction model for solar power generation based on weather parameters. Experimental results showed that the SVM model outperformed the others with up to 27 percent more accuracy. Furthermore, machine learning techniques play a crucial role in assisting decision making steps regarding the plants location selection and orientation selection as solar panels need to be faced according to solar irradiation to absorb the optimal energy. Conventional methods for sizing PV plants have generally been used for locations where the required weather data (irradiation, temperature, etc.) and other information concerning the site is readily available. However, these methods cannot be used for sizing PV systems in remote areas where the required data are not available, and thus machine learning techniques are needed to be employed for estimation purposes. In a study conducted by Mellit et al., an artificial neural network (ANN) model was developed for estimating sizing parameters of stand-alone PV systems. In this model, the inputs are the latitude and longitude of the site, while the outputs are two hybrid-sizing parameters. In the proposed model, the relative error with respect to actual data does not exceed 6 percent, thus providing accurate predictions. This model has been evaluated on 16 different sites and experimental results indicated that prediction error ranges from 3.75-5.95 percent with respect to the sizing parameters. Additionally, metaheuristic search algorithms address plan location optimization problems by providing improved local searches under the assumption of a geometric pattern for the field. Lastly, to maintain grid stability, it is necessary to forecast both short term and medium term demand for a power grid with renewable energy sources contributing a considerable proportion of energy supply. The MIRABEL system offers forecasting models which target flexibilities in energy supply and demand, to help manage the production and consumption in the smart grid. The forecasting model combines widely adopted algorithms like SVM and ensemble learners. The forecasting model can also efficiently process new energy measurements to detect changes in the upcoming energy production or consumption. It also employs different models for different time scales in order to better manage the demand and supply depending on the time domain. Ultimately, machine learning techniques support better operations and management of solar power plants. About MachinePulse MachinePulse is a provider of rapidly scalable solutions addressing the machine data requirements of industrial protocols, M2M and IoT segments with real time big data analytics and decision science on a cloud platform. We provide machine data solutions powered by erixis™, our intelligent cloud based, analytics platform. Erixis™ provides breakthrough operational efficiency leading to quantifiable monetary benefits. Contact us to learn more. MachinePulse has its seeds in Mahindra EPC, a Mahindra & Mahindra Group company.","excerpt":"The global solar photovoltaic (PV) installed capacity in 2013 was 138.9 GW and it is expected to grow to over 455 GW by 2020. However, solar power plants still have a number of limitations that prevent it from being used on a larger scale. One limitation is that the power generation cannot be fully controlled or planned […]","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2016-07-03T15:14:06","publication_year":"2016","word_count":694,"keywords":["Go","machine learning","TPU","AI","neural network","Scala","Git","RAG","analytics","R"],"extracted_tech_keywords":["AI","machine learning","neural network","analytics","RAG","TPU","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/machine-learning-for-better-and-more-efficient-solar-power-plants\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10066897,"title":"Google introduces PostgreSQL database AlloyDB","content":"At the Google I\/O event, Google introduced AlloyDB for PostgreSQL, a fully-managed, PostgreSQL-compatible database for demanding enterprise-grade transactional and analytical workloads. Google claimed that in its performance tests, AlloyDB gave “more than 4x faster on transactional workloads, and up to 100x faster analytical queries than standard PostgreSQL, all with simple, predictable pricing.” Google informs in a blog post that AlloyDB for PostgreSQL was built on the principle of disaggregation of compute and storage and designed to leverage disaggregation at every layer of the stack. Google explains that AlloyDB begins by separating the database layer from storage, then introducing a new intelligent storage service optimised for PostgreSQL. Google claims that this reduces I\/O bottlenecks. The storage service itself also disaggregates compute and storage that allows block storage to scale separately from log processing. Design It is a distributed system and has three parts: A low-latency, regional log storage service catering to very fast write-ahead log writing.A log processing service that processes these write-ahead log records and produces “materialised” database blocks.Failure-tolerant, sharded, regional block storage for durability even in case of zonal storage failures. Image: Google Google listed down some key benefits of this approach. These include full compute\/storage disaggregation, storage-layer replication, efficient IO paths\/no full-page writes, low-latency WAL writing, fast creation of read replica instances, fast restart recovery, and storage-layer backups.","excerpt":"Google informs that AlloyDB for PostgreSQL was built on the principle of disaggregation of compute and storage and designed to leverage disaggregation at every layer of the stack.","categories":["AI News"],"tags":["Database","Google","postgreSQL"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-05-13T11:57:00","publication_year":"2022","word_count":220,"keywords":["PostgreSQL","Go","programming_languages:R","AI","Database","postgreSQL","programming_languages:SQL","RAG","programming_languages:Go","Aim","Google","SQL","R"],"extracted_tech_keywords":["AI","Aim","RAG","PostgreSQL","R","SQL","Go","programming_languages:R","programming_languages:SQL","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-introduces-postgresql-database-alloydb\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":54476,"title":"How LinkedIn Is Using AI To Detect & Handle Abuse","content":"With organisations transitioning from a traditional business model into adopting emerging technologies, popular professional social networking platform, LinkedIn has also adopted machine learning technologies to help professionals in a more sophisticated way. Business and employment-oriented platform — LinkedIn hosts nearly 660+ million members in over 200 countries and territories. According to a source, it has 303 million active users per month, and on an average, two people create an account on this platform every second. Also, the platform witnesses 172,800 new users every day and around 62 million unique users every year. With these amounts of data generation, the developers of this platform have been striving hard to make its state-of-the-art machine learning model a robust one, so that it provides more accurate decisions or results for its users. The ML researcher team at LinkedIn built a domain-specific language (DSL) and a Jupyter notebook to integrate the selected features and for tuning the parameters. In one of our articles, we discussed how LinkedIn’s recommendation system is generating the perfect job match for its users. Utilising machine learning techniques to detect and remove inappropriate contents has been trending among social media giants like Facebook, LinkedIn, Twitter, among others. To automatically detect and remediate behaviours that violate the terms of service of the company and maintain a safe and trusted community, the professional platform has already been using techniques like automated fake account detection system, abuse detection systems and other such. According to reports, between January and June in 2019, the platform took action against 21.6 million fake accounts. The platform has been continually working to find and remove profiles which contains inappropriate contents and recently, the ML team at LinkedIn details a machine learning approach which handles inappropriate content. Initial ML Approach The initial approach used to identify and establish a set of words and phrases known as blocklist. When an account contained any of these inappropriate words or phrases, it was marked fraudulent and removed from LinkedIn. However, this approach includes a few drawbacks, such as: Scalability because it is a fundamentally manual process and needs to be taken care of while evaluating words or phrases.Words with both appropriate and inappropriate contexts.Tracking performance on a phrase-by-phrase basis takes a significant amount of time as well as engineering efforts. The New Approach To mitigate such challenges and improve the overall performance, the ML team at LinkedIn decided to change the machine learning approach. The new approach is a machine learning model which is a text classifier trained on public member profile content. To train this classifier, the researchers built a training set consisting of accounts labelled as either “inappropriate” or “appropriate.” The “inappropriate” labels consist of accounts that have been removed from the platform due to inappropriate content.popular professional social networking platform, LinkedIn has also adopted machine learning technologies to help professionals in a more sophisticated way. For the model, a deep learning architecture, Convolutional Neural Network (CNN) has also been leveraged. The reason to use this technique was because the CNNs perform particularly well on images and texts classification tasks. These are useful for data which has spatial properties meaning that there is information contained in the fact that two feature values are adjacent to each other. Challenges Faced While building the model, the team faced difficulties in assembling a training set which contains enough information. The labels for training data have been extracted using accounts that have been previously restricted for several reasons. When new models are trained using these labels, there is an inherent bias towards re-learning the patterns of existing systems. This issue has been resolved by identifying several problematic words that were producing high levels of false positives and sampled appropriate accounts from the member base containing such words. The accounts are then manually labelled and added to the training set.","excerpt":"With organisations transitioning from a traditional business model into adopting emerging technologies, popular professional social networking platform, LinkedIn has also adopted machine learning technologies to help professionals in a more sophisticated way.  Business and employment-oriented platform — LinkedIn hosts nearly 660+ million members in over 200 countries and territories. According to a source, it has […]","categories":["AI Features"],"tags":["ai dl","linkedin","linkedin machine learning"],"author_name":"Ambika Choudhury","publish_date":"2020-01-20T16:00:00","publication_year":"2020","word_count":635,"keywords":["machine learning","AI","neural network","ML","Scala","RAG","deep learning","Jupyter","linkedin machine learning","Rust","linkedin","ai dl","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","Jupyter","RAG","R","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-linkedin-is-using-ai-to-detect-handle-abuse\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10094853,"title":"Intel Unveils Arc Pro A60 and A60M GPUs","content":"Intel has introduced its latest offerings in the professional graphics processing unit (GPU) market with the launch of the Intel Arc Pro A60 and Pro A60M GPUs. These new products are part of the Intel Arc Pro A-series and are specifically designed to meet the needs of professional workstation users. The Intel Arc Pro A60 and Pro A60M GPUs boast impressive performance improvements compared to their predecessors. With up to 12GB of video memory (VRAM) and support for four displays featuring high dynamic range (HDR) and Dolby Vision support, these GPUs offer enhanced capabilities for content creation and visual experiences. One of the standout features of the Intel Arc Pro A60 GPU is its integration of ray tracing hardware, graphics acceleration, and machine learning capabilities. This combination enables smooth viewports, advanced visual technologies, and robust content creation, all packed into a traditional single-slot form factor. The Intel Arc Pro A-series graphics bring a fresh option to the workstation GPU market. The A60 and A60M models provide double the number of PCIe lanes, twice the memory bandwidth of 384 GBs, twice the dedicated AI Xe Matrix Extensions (XMX) engines, and twice the number of ray tracing units compared to previous Intel Arc Pro products. These GPUs also offer comprehensive support for media encode and decode, including AV1, making them ideal for computer-aided design and modeling (CAD\/CAM), AI inferencing tasks, and media processing in dedicated business environments that rely heavily on software such as Autodesk, Bentley MicroStation, Dassault Systemes SOLIDWORKS, Nemetscheck VectorWorks, PTC Creo, and others. Intel’s workstation GPUs are not only tailored for professional workstations but are also optimized for media and entertainment applications. They excel in running rendering and ray tracing libraries within the Intel oneAPI Rendering Toolkit, enabling the creation of visually stunning and immersive experiences with exceptional performance and fidelity. Applications like Blender benefit from these optimizations, empowering users to harness the capabilities of Intel’s GPUs for large-scale, high-quality rendering and realistic ray-traced graphics. Furthermore, the Intel Arc Pro GPU family has undergone compatibility validation with Intel NUC 13 Extreme small form factor (SFF) PCs. These compact PCs, equipped with powerful 125W unlocked 13th Gen Intel Core processors, provide users with a high-performance computing solution. The Intel Arc Pro A60 GPU designed for desktop workstations will be released and made available through Intel-authorized distributors in the coming weeks. On the other hand, the Intel Arc Pro A60M GPU, intended for mobile systems, will be offered by original equipment manufacturers (OEM) partners, but its availability is expected in the coming months. Leading computer manufacturers are also joining the Intel Arc Pro GPU ecosystem. HP has already introduced workstation designs featuring the Intel Arc Pro A40 GPU, which are currently available. Dell and Lenovo are expected to launch their own workstation designs featuring the Intel Arc Pro GPUs in the third quarter of 2023.","excerpt":"With up to 12GB of video memory (VRAM) and support for four displays featuring high dynamic range (HDR) and Dolby Vision support, these GPUs offer enhanced capabilities for content creation and visual experiences.","categories":["AI News"],"tags":["Intel","Intel processors","NVIDIA"],"author_name":"Mohit Pandey","publish_date":"2023-06-09T16:08:01","publication_year":"2023","word_count":474,"keywords":["Go","API","machine learning","programming_languages:R","AI","programming_languages:Go","Intel processors","Ray","NVIDIA","R","Intel"],"extracted_tech_keywords":["AI","machine learning","Ray","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/intel-unveils-arc-pro-a60-and-a60m-gpus\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070347,"title":"Why is the fairness in recommender systems required?","content":"We all utilise or make suggestions in our everyday lives. Machine learning helps to replicate the same suggestion mechanism, which is a system that filters out unwanted information and gives varied outputs based on distinct characteristics that vary from user to user. While recommending, these recommender systems may be biased or unfair at times; the bias might be of any form, such as a model bias or data bias. This article will be focused on discussing fairness for subjects and fairness for objects in the recommender systems. Following are the topics to be covered in this article. Table of contents Brief about the Recommender systemThe necessity of fairness in the recommender systemFairness in Recommendation systemConsumer FairnessProvider Fairness From selecting books to selecting friends, recommender systems assist us in making judgments. Let’s know more about the recommendation systems. Brief about the Recommender system Recommender systems (RS) deliver item suggestions to consumers by utilising artificial intelligence ideas. A machine learning algorithm, for example, may be used by an online bookstore to classify books by genre and propose additional books to a customer who wishes to buy a book. Recommender systems are classified into three types based on the information on which they are based: collaborative, content-based, and hybrid filtering. When processing information for a suggestion, a collaborative recommender system evaluates the user data. For example, by accessing user profiles on an online music store, the RS may obtain data like the age, country, and city of all users, as well as the songs they have purchased. Using this information, the system may identify individuals who have similar music tastes and then recommend tracks that similar users have purchased. A recommender system that uses content-based filtering makes suggestions based on the item data it has access to. Consider the case of a person looking for a new computer in an online store. When a user searches for a certain computer (item), the RS collects information about that computer and searches a database for machines with comparable characteristics, such as price, CPU speed, and memory capacity. The results of this search are then returned to the user in the form of suggestions. A recommender system that combines the two preceding classes into a hybrid filtering strategy, recommending things based on user and item data. A recommender system on a social network, for example, may recommend profiles that are similar to the user (collaborative filtering) by comparing their interests. In a subsequent stage, the system may treat the recommended profiles as things and so access their data to look for new profiles that are comparable (content-based filtering). Both sets of profiles are eventually returned as suggestions. Aside from the standard recommendation process, in which users are shown items that they might be interested in, recommendations can be done in a variety of ways. Context-aware suggestions are generated based on the context into which the user has been placed. A context is a collection of information regarding the user’s present condition, such as the time at their current location (morning, afternoon, evening), or their activities (idle, running, sleeping). The quantity of context information that must be analysed is large, making context-aware suggestions a difficult study topic. Risk-aware recommendations are a subset of context-aware suggestions that take into account a scenario in which essential information, such as user vital information, is available. It is risk-aware since a bad decision might endanger the user’s life or inflict real-world damage. Some instances include advising a user on which medications to take or which equities to purchase, sell, or invest in. Are you looking for a complete repository of Python libraries used in data science, check out here. The necessity of fairness in the recommender system The recommendation system is built on a feedback loop with three candidates: user, data, and model. These candidates are used in three different phases. Collection: This denotes the phase of gathering data from users, which includes user-item interactions and other incidental information (e.g., user profile, item attributes, and contexts). Learning: This refers to the development of recommendation models based on data collected. At its foundation, it predicts how likely a user is to embrace a target item based on prior encounters. Over the last few decades, much research has been undertaken. Serving: This step delivers the suggestion results to users to meet their information needs. This stage will have an impact on users’ future behaviours and decisions. These are the phases where biases are introduced to the system due to which the recommendation system could be unfair either to a subject or an object. Let’s deep dive into fairness in recommendations and understand the root causes and solutions to the problems. Fairness in Recommendation system To accomplish fairness, a common technique is to define a variable or variables that indicate membership in a protected class, such as race in an employment setting, and to build algorithms that eliminate prejudice relative to this variable. To apply this approach to recommender systems, we must acknowledge the critical importance of personalisation. The concept of suggestion implies that the finest things for one user may differ from those for another. It is also worth noting that recommender systems exist to help with transactions. As a result, many recommendation applications include many stakeholders and may raise fairness concerns for more than one set of participants. Consider a recommender system that suggests employment openings to job searchers. An operator of such a system may aim, for example, to guarantee that male and female users with comparable qualifications receive job suggestions with comparable rank and income. As a result, the system would need to fight against biases in recommendation output, including biases caused purely by behavioural differences: for example, male users may be more prone to click optimistically on high-paying positions. It is difficult to overcome such biases if there is no consensus on global preference ranking over goods. Personal preference is the essence of suggestion, especially in fields where individual taste is crucial, such as music, literature, and movies. Even in the work arena, some users may prefer a somewhat lower-paying job if it comes with additional perks like flexible hours, a shorter travel time, or better benefits. To achieve the policy goal of salary-based job recommendation, a site operator will need to go beyond a purely personalization-oriented approach, identify salary as the key outcome variable, and control the recommendation algorithm to make it sensitive to the salary distribution for protected groups. Fairness varies according to the stakeholders Different recommendation situations can be characterised by different stakeholder interest configurations. A recommender system’s stakeholders are divided into three categories: customers, suppliers, and platform or system. Consumers are the ones who receive suggestions. They are the people who come to the platform because they are having difficulty making a decision or searching for something, and they anticipate tips to help them. The suppliers are the organisations that supply or otherwise support the suggested objects and profit from the consumer’s decision. The platform has developed the recommender system to connect customers with suppliers and has some way of profiting from the process. The system will eventually have aims that are a function of the other stakeholders’ utilities. When multisided platforms can attract and keep critical masses of players from all sides of the market, they thrive. In our employment example, if a job seeker does not find the system’s recommendations useful, he or she may choose to disregard this component of the system or shift to a competitor platform. The same is true for providers; if a certain site does not offer its advertising as suggestions or does not supply appropriate people, a firm may pick another platform to publicise its job opportunities. Recommendation methods on multi sided platforms might raise concerns about multi sided fairness. Specifically, there may be fairness-related criteria at work on more than one side of a transaction, and hence the transaction cannot be judged only based on the outcomes that accrue to one side. There are two types of systems defined by the fairness difficulties that occur with these groups: consumers fairness, and providers fairness. Consumer Fairness Fairness is a concept of nondiscrimination based on membership in protected groups, specified by a protected trait, such as gender and age. A customer-fair recommender system considers the differential impact of the suggestion on protected classes of recommendation consumers. Group fairness is the lack of discrimination against a certain group, defined as the absence of a differential impact on the outcomes created for them. Despite the involvement of many stakeholders, fairness in recommender systems may have a particularly negative impact on individuals who get consumer suggestions. As a result, group consumer fairness should account for no disproportionate impact of recommendations on protected consumer groups. Providing assurances on this property is a critical strategic goal for the field’s responsible progress. A credit card business recommends customer credit offerings in the motivating example. Because the items are all from the same bank, there are no difficulties with producer fairness. In systems of this nature, multistakeholder considerations do not exist. Several designs might be offered. One fascinating alternative is to create a recommender system based on the principle of fair classification. We may establish a mapping from each user to a prototype space, possibly using latent features retrieved from the rating data. Each prototype might be designed to have statistical parity with respect to the protected class. A significant aspect of this sort of system is ensuring a finite loss with regard to the input. Some consumer fairness algorithms SLIM: To decrease unfairness, it was proposed to create suggestions for a user from a neighbourhood with an equal number of peers from each category. A regularisation was added to SLIM, a collaborative filtering approach, to create a balance between protected and non-protected neighbours. Fairness was tested using a risk ratio version; this score is less or larger than 1 when the protected group is suggested fewer movies of the desired genre; on average, 1 implies perfect equity. Latent Block Model: It is intended to provide fair recommendations by co-clustering people and goods while maintaining statistical parity for some sensitive features. It employs an ordinal regression model with sensitive qualities as inputs. Fairness was determined by ensuring that the proportion of users with the same choice across demographic categories was similar for any two products. NLR: Based on the degree of engagement on the platform, the developer evaluated consumer unfairness among user groups (more or less active). As mitigation, a re-ranking technique was used, with the goal of selecting items from each user’s baseline top-n list to optimise overall recommendation utility, with the model confined to minimising the difference in average recommendation performance across the groups of users. Random sampling without replacement: The developer re-sampled user interactions in the training set such that the representation of user interactions across groups was balanced, and then re-trained the recommendation models with the balanced training set. The mitigation entailed developing a recommendation model by decreasing the dissimilarity between genuine ratings and anticipated ratings while also maximising the degree of independence between predicted ratings and sensitive labels. The MAE was used to calculate prediction errors. The equality of the expected rating distributions between groups was used to assess independence. Provider Fairness A Provider fair system is one in which fairness must be kept exclusively for the providers. Consider an online microfinance portal that collects loan requests from field partners all around the world who lend small sums of money to local entrepreneurs. The loans are sponsored interest-free by the organization’s members, the majority of whom live in the nation. The organisation does not currently provide a customised recommendation function, but if it did, one would envision that one of the organization’s goals would be to maintain the equitable distribution of money across its many partners in the face of well-known user biases. Consumers of the suggestions are simply contributors who gain no direct advantage from the system, hence there are no consumer-side fairness problems. Where there is an interest in promoting market variety and preventing monopolistic domination, P-fairness may also be a factor. In the online craft marketplace Etsy, for example, the system may want to guarantee that new entrants to the market receive a fair proportion of recommendations despite having fewer consumers than established merchants. This sort of justice is not required by law but rather is built into the platform’s economic model. Provider fairness (P-fairness) systems include difficulties that Consumer fairness (C-fairness) systems do not. The producers in the P-fairness example, in particular, are passive; they do not seek out suggestion chances but must instead wait for users to come to the system and request recommendations. Consider the preceding employment example. We want positions at minority-owned firms to be suggested to highly qualified individuals at the same rate as jobs at other types of businesses. The chance to propose a specific minority-owned firm to an acceptable applicant is uncommon and must be acknowledged as such. We will want to limit the loss of personalisation that comes with any advertising of protected providers, as we did in the C-fairness example. Diversity-aware systems approach recommendation as a multi-objective optimization issue, with the goal of maintaining a particular degree of accuracy while also ensuring that recommendation lists are varied in terms of some representation of item content. These strategies may be repurposed for P-fairness recommendation by considering the protected group items as a separate class and then optimising for various suggestions relative to this variable. A more dynamic approach to managing suggestion opportunities is required to achieve individual P-fairness coverage. The most similar analogue is perhaps found in online bidding for display advertising, where restricted ad expenditures serve the role of dispersing impressions across competing advertisers. Individual P-fairness is achieved in this scenario within the constraints of the customised mechanism by offering the protected group equal buying power to the non-protected group. Conclusion The recommendation system could be unfair from the user side or from the provider side. A recommender system is fair when it considers the differential impact of the suggestion on protected classes of recommendation consumers as well as also protects the objective of the provider of the system. With this article, we have understood the concept of fairness in the recommendation system from both the consumer and the provider. References Read more about Consumer fairness","excerpt":"The system is considered fair when the recommendation is unbiased towards any group or individuals consumer or even providers.","categories":["AI Trends"],"tags":["Deep Learning","Machine Learning","Python","recommendation engine"],"author_name":"Sourabh Mehta","publish_date":"2022-07-04T16:00:00","publication_year":"2022","word_count":2388,"keywords":["data science","artificial intelligence","machine learning","TPU","AI","Machine Learning","recommendation systems","RAG","Python","recommendation engine","Aim","Deep Learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","Aim","RAG","recommendation systems","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-is-the-fairness-in-recommender-systems-required\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10103703,"title":"Pixxel’s Hyperspectral Odyssey","content":"Bengaluru-based, Pixxel—India’s first company to launch private commercial satellites in space, is all set to launch Firefly—world’s first high-resolution hyperspectral satellites constellation (consists of six satellites) by 2024, specifically for geospatial study and analysis. Over the past two years, Pixxel has deployed three highly functional satellites in orbit, each meticulously designed to decipher and convert vast data streams into actionable insights. While conventional satellites captured data within a limited range, multispectral satellites added a few infrared bands, enabling observations beyond RGB. “Hyperspectral imagery enables us to capture information in hundreds of wavelengths, in our case about 300 units,” Awais Ahmed, Pixxel’s CEO and Founder, told AIM in an exclusive interview. Simply put, Pixxel’s hyperspectral imaging system can discern a wide array of information from the electromagnetic spectrum captured by sensors. “Light is broken down into very minor, very thin sliced wavelength bands and depending on wavelength bands or data channels, you can decipher different information,” explained Ahmed. The use cases are plenty. In agriculture, Pixxel’s satellites identify crops and discern nutrient content in soils, gauge moisture levels, and assess chlorophyll content, enabling precision farming and crop management. These satellites detect methane leaks and differentiate between minerals for oil and gas sectors, offering unprecedented insights for mining companies. In forestry, they monitor pollution levels, track biodiversity changes, and identify pest infestations and tree species. Pixxel’s hyperspectral satellite also helps in providing global coverage with the capability to revisit any location on Earth within a 24-hour cycle. “The goal is clear: We’re not just observing; we’re transforming the way industries operate, offering insights that empower decision-makers worldwide,” added Ahmed. Solving One Pixxel at a Time The journey to harnessing this vast amount of data hasn’t been without challenges. Managing the deluge of information, from capturing and storing it to processing it in real-time, has demanded technological upgrades. Pixxel’s edge computing capabilities, compression algorithms, and optimised storage systems have mitigated these hurdles, ensuring the reliability and accuracy of insights delivered. Ahmed outlined three key issues in handling this immense data volume: capturing the vast amounts of data from the satellite, storing and downlinking, and caching it for quick access once available on the ground. Pixxel has evolved its hardware over three years to resolve the data capture issue. Compression algorithms and on-board edge computing mitigate the challenge of downlinking vast data within limited satellite-ground station communication windows. This innovation significantly reduces the data needing transmission, optimising efficiency through increased downlink capacity or by processing data in space itself. Moreover, on-ground storage strategies involve short- and long-term storage solutions tailored to varying data access needs. Pixxel’s approach involves innovative hardware solutions, compression algorithms, and a strategic mix of cloud-based storage to ensure efficient management and swift access to copious amounts of captured hyperspectral data. “We are a vertically integrated company… from designing of the satellite to testing and building and operating is something that we do, including the edge computing parts of it,” explained Ahmed, highlighting the self-sufficiency in processing data for hyperspectral imagery. Pixxel in Short Founded in 2019 by Ahmed and co-founder CTO Kshitij Khandelwal, Pixxel has raised upwards of $70 million. The company, now comprising 175 employees globally, primarily centred in Bangalore, is extending operations to the US and Europe to cater to local needs. The recent funding will also fuel the advancement of Aurora, Pixxel’s AI-driven analytics platform. This initiative seeks to democratise hyperspectral analysis, making it more accessible to a broader audience. “We’ve been focusing on the Aurora platform, adding more features such as mathematical and statistical models, AI, and deep learning, supervised and unsupervised, to extract insights. For instance, land use and land cover classification can classify images in Asia into trees, water bodies, crops, and buildings, almost in real-time, without extensive training. There are other models like crop stress identification, biodiversity assessment, and oil and gas leak detection, among others,” said Ahmed. Pixxel also boasts its partnerships with tech giants like Google, Microsoft, and Amazon. “The biggest thing that we depend on these companies for, first and foremost, is the cloud infrastructure.” This infrastructure and data marketplace access is crucial for storing and analysing vast amounts of data, enhancing data delivery and accessibility for users worldwide. “Google Earth is used by 10s of 1000s of developers on the geospatial side to analyse satellite imagery,” said Ahmed. Meanwhile, Microsoft’s planetary computer and Amazon’s geospatial data marketplace cater to diverse user bases, enabling Pixxel to reach end-users efficiently. What’s next? With the Firefly series poised to embody the culmination of years of research and learning. These larger, more advanced satellites promise enhanced resolution, extended lifespans, and superior performance, setting the stage for a quantum leap in global-scale commercial operations. With a global expansion strategy, Pixxel has set its sights on key markets, primarily in the United States and Europe. These regions, coupled with partnerships in Canada and Australia, form the crux of their customer base, with plans for further penetration into emerging markets in the years ahead. Moreover, the $36 million in Series B funding will be used to deploy an additional 18 satellites by 2025, amplifying its capacity for data collection.","excerpt":"It is set to launch world’s first high-resolution hyperspectral satellite constellation by 2024 and 18 additional satellites by 2025.","categories":["IT Services"],"tags":["Agriculture","Data Management","edge computing","mining","Pixxel"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-11-27T17:00:37","publication_year":"2023","word_count":853,"keywords":["Pixxel","Go","AI","Agriculture","RAG","Ray","Aim","mining","edge computing","deep learning","analytics","Data Management","R","analytics platform"],"extracted_tech_keywords":["AI","deep learning","analytics","Aim","Ray","RAG","edge computing","R","Go","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/it-services\/pixxels-hyperspectral-odyssey\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":35465,"title":"Apart From Hiring, Hackathons Help In Employee Engagement & Business Growth: Nidhi Khanna, Ciber India","content":"Hackathons are gaining popularity in data science and the IT industry to help spot the right talent in a big way. We got in touch with Nidhi Khanna, Vice President Delivery at Ciber India, an IT consulting firm that largely relies on analytics. With 22+ years of experience, she has been instrumental in delivering profit-driven technology services and solutions and building high-performance teams. She joined Ciber in 2013 as the Emerging Technologies Practice Director, and since then has been instrumental in building technology practices, driving innovation and thought leadership in the organisation. She incubated the concept of Startup practice for technologies (like Mobility, AI, IoT & Analytics) and developed it to a matured delivery state. Here are some interesting insights from Khanna on how hackathons have become a key ingredient of the data science and analytics industry. Analytics India Magazine: How have hackathons become a crucial part of the hiring process in the data science industry? Do you think hiring through hackathons will make hiring easier? Nidhi Khanna: Hackathons are becoming increasingly popular in the hiring process for niche areas such as data science, by enabling organisations to evaluate technical skills far better than what is provided in resumes. Hackathons also provide the opportunity to assess the problem-solving approach and capabilities of an individual within the time constraint, which is a practical and very plausible real-life simulation that he or she may encounter in their role. Such assessments are not possible through traditional recruitment processes. Yes, Hackathons definitely makes the evaluation process easier, faster and helps in identifying the right candidate for a role. AIM: Hackathons solve two purposes —people strategies and the other is solving use cases in an innovative way? Do you agree? NK: In the ever-evolving technology industry, engaging people and keeping them challenged is the need of the hour in our domain. Hackathon is a great tool to inspire and nurture an innovation-focused culture within the organisation. It facilitates creative thinking and innovative solutions. AIM: What is the main purpose of conducting a hackathon for you? NK: Hackathon is one of the most effective and constructive ways to drive innovation internally and to engage people. It also provides a platform for consultants to collaborate and showcase their continuous learning. AIM: Can you cite a recent example where you hosted a hackathon? Could you state the purpose? NK: We have been conducting hackathons regularly for the last five years at Ciber, with a purpose of inculcating a culture of learning, innovation and creating the opportunity for all to ideate on go-to-market solutions. Each year we cover a new theme and a niche technology with a focus on industry needs. The most recently held at Ciber India was ‘Hacka-lo-mania’ in December 2018, where consultants developed data science and AI solutions addressing business problems of the healthcare industry. AIM: How important are hackathons (both internal and external) to boost innovation? Have you in your organisation benefitted by conducting hackathons? NK: From both internal and external perspectives, hackathons are an excellent way to shorten an organisation’s innovation cycle and help to move rapidly from ideas to prototypes. They also give participants a chance to learn and explore areas outside of their routine work and collaborate both within and outside the organisation. Yes, our internal hackathons have helped in both employee engagement and business growth. They create an environment for individuals to explore beyond boundaries, arrive with intuitive ideas, in result organisation gains workable prototypes to demonstrate its capabilities. AIM: Besides building people capital, do hack for hire also help in positioning the company as innovative and help in branding? NK: Yes, Hackathons leveraged for recruitment purpose have a tangible impact on company branding. First, it demonstrates the disposition of an organisation towards innovation, which is integral and a must have especially in the IT Industry. Second, when an organisation opts for a more modern and engaging hiring process via a Hackathon, it creates an opportunity for involvement with the talent pool and a platform to engage and interact with prospective employees, which is very different from the traditional interview-based hiring process. In sum, Hackathons are a great way to create awareness of your brand, its association with innovation and bringing the community together, reflective of progressive and collaborative work culture. AIM: How is hackathon turning into mainstream hiring requirements? NK: Several reports indicate that millennials are the growing majority of our workforce and will comprise 75% of our global workforce. This does not leave organisations with many choices but to come up with creative and efficient ways of not just hiring, but also keeping the employees engaged. So yes, this new recruitment trend is here to stay. AIM: What is your preference for candidates who have participated in hackathon vs direct hiring? NK: While Hackathons will be the future of recruitment and will offer quicker hiring solutions, we cannot completely rule out the value of direct hiring. There are roles such as team leads, customer facing and managerial responsibilities, which require a set of soft skills that cannot be assessed through Hackathons. Thus, my personal preference depends on the role, where hiring Graduates straight from campuses and colleges through hackathons would be most effective. AIM: Are hackathons replacing the need to have short term certification course in areas such as AI, analytics? NK: The Importance of short-term certification in areas such as AI and analytics will remain, as these are necessary for learning and benchmarking of a new skill, whereas hackathons help in evaluating the application of skills. AIM: In future, how will hackathons evolve (smaller, online hackathons or the preference will be for offline events) NK: There are over 1000+ hackathons conducted every year around the world. In future, we will see these numbers increasing; many companies are opening their data sets to developers to build effective predictive models online specifically in BFSI space; while others will be conducting internal events to ideate solutions for business challenges. Participate in our recent hackathon here.","excerpt":"Hackathons are gaining popularity in data science and the IT industry to help spot the right talent in a big way. We got in touch with Nidhi Khanna, Vice President Delivery at Ciber India, an IT consulting firm that largely relies on analytics. With 22+ years of experience, she has been instrumental in delivering profit-driven […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2019-02-27T06:30:26","publication_year":"2019","word_count":994,"keywords":["data science","Go","API","AI","innovation","RAG","Aim","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","API","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/apart-from-hiring-hackathons-help-in-employee-engagement-business-growth-nidhi-khanna-ciber-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10011070,"title":"How GitHub Got MLOps Right","content":"“DevOps is not a product that you can buy and install.”Pulkit Agarwal, GitHub. After productive and informative Day 1, ADasSci’s Deep Learning Developers Conference is live again. Day 2 of DLDC2020 too, had an interesting lineup of speakers along with a full-day workshop on deep learning with Keras. In an hour-long talk, speakers Pulkit Agarwal and Vinod Joshi of Github discussed the various challenges of setting up an ML pipeline. Pulkit, who is part of the product team at Github, began by defining what MLOps is really about and what makes it challenging while organisations have figured out how to work with DevOps. MLOps comes with an additional challenge of machine learning lifecycle automation. Usually, more emphasis is placed on models, but Pulkit likened the model-building to a small cog in the wheel. For instance, small systems are not sufficient for remote training. VMs or Spark clusters are essential. Pulkit listed four key challenges one might face while setting up an ML pipeline: Collaboration on codeRemote trainingModel BookkeepingManaging data code and updates. Model bookkeeping, for example, can cost a project dearly. Developers can lose track of file versions, and deployment becomes chaos. There can be other instances where someone doesn’t know how to write a controller file. Organisations might run into this trivial-sounding yet serious problem sooner or later if attention is not paid to the details. So how does GitHub get MLOps right? Although Pulkit admits that “easy” in MLOps is a very ambitious goal, the team at GitHub tries their best by incorporating three important components: ML Optimised computeSource control andML Aware For example, the job of ML Aware CI\/CD component is to warn the system in case of code change or other updates. While the first half of the talk included how GitHub made MLOps easy-ish, the second half, helmed by Vinod Joshi, was about how these principles were put to use in building models for increasing productivity of the developers. Vinod elaborated about the various aspects of ML lifecycle and the importance of building and rebuilding models when there is any change in the data distribution. Vinod continued his talk by dissecting a use case where he and his team have worked on a model that tracks the coding time of the developers. The whole process can be looked at through the lens of a Markov process where coding and non-coding are the states between which the observations or commits, in this case, are made. Due to the many hidden states, this becomes more of a hidden Markov model. So what are the implications of such an experiment? The notion here was to identify the patterns between commit intervals and productivity. In large organisations, continued Vinod, developers don’t get enough time to code due to various other activities like meetings etc. For a developer, time spent on coding reflects their job satisfaction. So the insights gathered from this model can have multiple applications within the organisation. Another key application can be identifying the right time code review. If a developer’s coding time is tracked and if peak working time is figured out, then they would give the code for review within those time zones when they are productive. Team GitHub, in this talk, gave us a glimpse of what it takes to make MLOps easy for an organisation. According to Pulkit, MLOps or DevOps is not just any software but more of a value, a union of people and processes. DevOps is not just any product that one buys and installs. This sums up the ethos that underlies GitHub’s success with ML.","excerpt":"“DevOps is not a product that you can buy and install.” Pulkit Agarwal, GitHub. After productive and informative Day 1, ADasSci’s Deep Learning Developers Conference is live again. Day 2 of DLDC2020 too, had an interesting lineup of speakers along with a full-day workshop on deep learning with Keras. In an hour-long talk, speakers Pulkit […]","categories":["Deep Tech"],"tags":["GitHub","MLOps"],"author_name":"Ram Sagar","publish_date":"2020-11-02T14:00:32","publication_year":"2020","word_count":596,"keywords":["Go","machine learning","Keras","AI","ML","MLOps","Git","deep learning","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","MLOps","Keras","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/github-mlops-machine-learning-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":29759,"title":"Top 10 Executive Data Science Courses in India – Ranking 2018","content":"Read Our Latest ranking for 2020. A primary survey was conducted in August through September 2018, where 961 current and past students from 18 cities in India gave opinions on data science courses they attended. Out of these entries, few dozens in-depth interviews were done to understand the reasons for the submissions. This exercise helped in validating the data and providing a rationale for the rankings, wherever required. A dedicated online questionnaire was created and the link was sent to more than 30 data science schools, of which 21 responded within the stipulated time.The participants were asked to fill an elaborated form with four key parameters — course content, faculty, student experience and other attributes such as external collaboration. Some of the other details under these parameters were course comprehensiveness, post-completion engagement of students, capstone projects and others. Seven schools were rejected because of incomplete data, lack of supporting documents or not fulfilling the eligibility criteria. The eligibility criteria for this ranking was 1) The course should be a long term program in data science\/ analytics (atleast 5 months), 2) Should be executed by (or in collaboration) with a university. Information collected from the schools was combined with information received by the students.  Out of those, we bring to you top 10 courses which have been analysed thoroughly by our team of experts. Students feedback and expert advice were also taken into consideration for the overall ranking process. Each of these parameters have been ranked on the scale of 1-5 where 1 is for worst and 5 for best. The institutes that could not be a part of the list either did not participate in the ranking process or could not make it to our list. For top 1o courses on Artificial Intelligence in India, check this. See our Top 10 Full Time Data Science Courses In India- Ranking 2018 here. For top 10 data science training institute in India 2018, check here. 1. Post Graduate Program In Business Analytics And Business Intelligence (PGP-BABI) By Great Learning In Association With Great Lakes Institute of Management Founded: 2013 Mode of delivery: Blended Course duration: 12 Months Number of hours: ~440 with 240 hours of classroom learning and 200 hours of online learning Cities of operation: Mumbai, Bengaluru, Delhi NCR, Chennai, Hyderabad, Kolkata, Pune, Online Course Fees: ₹ 395000 + GST, ₹ 170000 + GST (Online) Great Learning is an ed-tech company that offers programs in career critical competencies such as Analytics, Data Science, Big Data, Machine Learning, Artificial Intelligence, Cloud Computing and more. Great Learning programs are taken by several thousands of professionals every year by a network of 400+ Great Learning Gurus. They boast an alumni network of more than 3000 professionals working across organisations in India and beyond. Parameter 1: Course Content (Rating 4.9) Comprehensiveness: PGP-BABI uses a combination of learning methods such as classroom teaching, self-learning through videos and reading materials, team-based problem solving and others. The exhaustive course curriculum includes foundation (foundation in Statistics using R, business and management concepts), analytical techniques (R, Python, Tableau SAS) and domain application and industry exposure. They have a capstone project of 3 months, while the course is updated every 6 months. Accessing students at the end of the course: Following a continuous evaluation, the students participate in group projects, individual assignments, quizzes and group discussions in every course. Students are tested for their conceptual understanding and ability to apply various analytics techniques in real-world problem. Learning resources: Books, downloadable resources, LMS, videos, question papers. Students can also work on a number of practice exercises and datasets provided by them. Parameter 2: Faculty (Rating 4.9) Total no. of faculty members: 87 Total no. of faculty members with a PhD: 31 Total no. of faculty members with industry experience: 87 Student to faculty ratio: 16:1 Parameter 3: Student Experience (Rating 4.8) Percentage of students who completed the course: 91% Post-completion engagement: Every graduating student becomes a Great Lakes Alumni and are invited to all alumni and networking events. They also have lifetime access to a repository of learning material which is continuously updated. Placement assistance: They provide full placement assistance. 66% of the alumni are working across companies such as Deloitte, KPMG, Gartner, RBS, Cognizant and others. Parameter 4: Other Attributes (Rating 4.9 ) Entry criteria for students: Industry experience required + Personal Interview External collaboration: They have collaborated with over 400+ industry experts from companies like Microsoft, Accenture, Deloitte, EXL, WNS, Cognizant, American Express, Absolute Data, HSBC, etc. who design, deliver, and endorse the program. Other notable collaborations are with Bajaj Allianz, WNS Global, XL Catlin for Ideathons, Hackathons, live capstone projects and customized training programs The overall rating is 4.87 2. Post Graduate Program in Data Science and Machine Learning (PGPDM) By Jigsaw Academy And University Of Chicago Founded: 2011 Mode of delivery: Blended Duration of the course: 10 months Number of hours: 650 Cities of Operation: Bangalore, Delhi NCR, Hyderabad Course Fees: ₹ 3.65 lakhs + taxes Jigsaw Academy has been key in contributing to the evolution of analytics talent in the country, having produced over 50,000+ data scientists across the globe. They offer training on multiple popular & niche tools in Data Science, Machine Learning, Big Data and AI technologies that are used globally. Parameter 1: Course Content (Rating 4.9) Comprehensiveness: PGPDM includes exhaustive coverage of Data Science and Machine Learning (R, Python & SAS), Big Data (Hadoop, HDFS, Pig, Hive and Spark), and Visualization (Tableau). The program starts from the basics of statistics and covers the entire gamut of descriptive analytics, predictive analytics as well as AL & ML. It includes hands-on program from case studies across BFSI, retail, telecom, supply chain, HR and other industries. The capstone project is for more than 3 months and course is update every 6 months. Accessing students at the end of the course: PGPDM has a multiple-assessment policy, which includes case studies, assignments, capstone projects, and more. Only those students who pass graded assignments in all the modules and those students who complete corporate projects as per the company’s satisfaction are given the final certificate. Learning resources: Downloadable resources, LMS, videos, industry lectures, capstone projects, instructor-led classes, assignments Parameter 2: Faculty (Rating 4.8) Total no. of faculty members: 60 Total no. of faculty members with a PhD: 20 Total no. of faculty members with industry experience: 60 Student to faculty ratio: 1:5 Parameter 3: Student Experience (Rating 4.7) Percentage of students who completed the course: 95% Post-completion engagement: They are a part of the alumni network, invited as guest lectures for future batches & become a part of Jigsaw mentorship program Placement assistance: Provided full assistance through alumni network, resume designing and helping them with interview process Parameter 4: Other Attributes (Rating 4.7) Entry criteria for students: Application + Interview External collaboration: They have tie-ups the University of Chicago Graham School, along with industry associations with companies like Axteria, Analytics Edge, Tata Motors, Smart Cube, Data Semantics for projects and hiring. The overall rating is 4.77 3. Post Graduate Certificate Program In Data Science & Machine Learning By Manipal ProLearn In Association With Manipal Academy Of Higher Education Founded: 2016 Mode of delivery: Online Duration of the course: 6 months Number of hours: 650 Cities of Operation: Bangalore, the program is online and hence applicable irrespective of location Course Fees: ₹ 84,000 + taxes A part of MaGE, Manipal ProLearn is a pioneer in higher education and allied services. It offers a variety of cutting-edge learning solutions and professional certification courses across IT, Digital Marketing, Data Sciences, Project Management, and others. The Academy of Data Science runs the popular data science certification program to help build a community of industry-ready data scientists. Parameter 1: Course Content (Rating 4.7) Comprehensiveness: Designed with optimal blend of rigour and relevance, it covers in-depth and industry-relevant content. Course comprises of real-world business case studies and live online lectures from industry practitioners. It covers Java programming, advanced excel, R, introduction to Python, data visualisation tools, big data technologies, business communication and others. They do not offer capstone projects and course is updated every 6 months. Accessing students at the end of the course: Learners undergo online quizzes for every topic and online assessments for every module. Learning resources: Downloadable resources, LMS, videos, question papers, industry webinars on relevant topics for improved understanding of application of learning Parameter 2: Faculty (Rating 4.7) Total no. of faculty members: 18 Total no. of faculty members with a PhD: 5 Total no. of faculty members with industry experience: 12 Student to faculty ratio: 10:1 Parameter 3: Student Experience (Rating 4.7) Percentage of students who completed the course: 95 Post-completion engagement: Student get access to learning content and is invited as alumni for all learning webinars, meet-ups and other activities Placement Assistance: None Parameter 4: OtherAttributes (Rating 4.6) Entry criteria for students: Entrance exam External collaboration: Manipal Academy of Higher Education, Equifax, Gramener The overall rating is 4.67 4. PG Program in Data Science By UpGrad In Association With IIITB Founded: 2015 Mode of delivery: online Duration of the course: 11 months Number of hours: 500 Cities of Operation: UpGrad is an online learning platform Course Fees: ₹ 2,35,000 Founded by media stalwart Ronnie Screwvala, UpGrad is an online higher education platform that provides rigorous industry-relevant programs designed and delivered in collaboration with renowned faculty and industry experts. Parameter 1: Course Content (Rating 4.6) Comprehensiveness: The curriculum has been designed along with IIIT Bangalore and multiple industry leaders. The broad areas of focus are data management (Excel, Python, SQL, Tableau), statistical and exploratory data analysis, machine learning, big data analytics (Hive, Spark, Sqoop). The candidates select specialisation of their choice in BFS, ecommerce, retail or healthcare. The course includes capstone project of 1-3 months. The course is updated every 6 months. Accessing students at the end of the course: Students are assessed with the help of objective questions, individual assignments, group case studies with personalised feedback by industry experts. Participation in class, online discussion forums and online proctored exams contribute to overall CGPA of the learner. Learning resources: Downloadable resources, LMS, videos, question papers, online virtual labs for Big Data Parameter 2: Faculty (Rating 4.5) Total no. of faculty members: 18 Total no. of faculty members with a PhD: 12 Total no. of faculty members with industry experience: 17 Student to faculty ratio: 10:1 Parameter 3: Student Experience (Rating 4.6) Percentage of students who completed the course: 84% Post-completion engagement: Career support guidance for one year after course completion, access to new and re-developed content, alumni network, opportunity to mentor subsequent cohort learners. Placement assistance: Assistance is provided through alumni network, resume designing, helping them with interview process, 1-1 mentoring by industry expert, placement drives, help identify correct job profile, resources for interview and test preparation, and others. Parameter 4: Other Attributes (Rating 4.5) Entry criteria for students: Entrance exam External collaboration: The program has flagship partnership with Uber, Genpact, Fractal Analytics and Gramener. They have collaboration with 30+ analytics industry experts from leading corporations and 250+ recruitment partners such as KPMG, Tech Mahindra, Zivame and others. The overall rating is 4.55 5. Executive Program In Business Analytics (EPBA) By Jigsaw Academy In Association With SDA Bocconi Founded:2011 Mode of delivery: Blended Duration of the course: 10 months Number of hours: 650 Cities of Operation: Online, global Course Fees: ₹ 4.9 lakhs + taxes Headquartered in Bengaluru, it was founded by the duo of Gaurav Vohra and Sarita Digumarti, and funded by Manipal Global Education Services (MaGE). It trains professionals in the areas of Analytics, Data Science, Big Data, Machine Learning, Business Analytics, and more recently, the Internet of Things (IoT). Parameter 1: Course Content (Rating 4.5) Comprehensiveness: The EPBA has the most comprehensive curriculum in Data Science &, Big Data in India. Along with SDA Bocconi, it provides unique and exhaustive course which includes analytics (R, Python, SAS), Big Data (Hadoop, Pig, Hive, Sqoop, Flume, Spark and Storm), Visualization (Tableau, Power BI).  It also covers descriptive analytics, predictive analytics, machine learning, an introduction to neural networks. It involves capstone project of more than 3 months, with course updation in every 6 months. Access students at the end of the course: Graded assignments, capstone project and viva interviews Learning resources: Downloadable resources, LMS, videos, question papers, case studies & online content Parameter 2: Faculty (Rating 4.5) Total no. of faculty members: 40 Total no. of faculty members with a PhD: 25 Total no. of faculty members with industry experience: 40 Student to faculty ratio: 1:2 Parameter 3: Student Experience (Rating 4.4) Percentage of students who completed the course: 95 Post-completion engagement: They are a part of Bocconi Alumni Network, Guest Lectures for future batches & Jigsaw Mentorship Program Placement assistance: Full assistance through alumni network, resume designing, helping them with interview process Parameter 4: Other Attributes (Rating 4.4) Entry criteria for students: Application + Interview External collaboration: Collaboration with SDA Bocconi, Milan. Industry collaboration with Axteria, Analytics Edge, Tata Motors, Smart Cube, Data Semantics, and more for projects and hiring. The overall rating is 4.45 6. Postgraduate Certificate in Business Analytics for Management Decisions (PGCBAMD) By XLRI Xavier Institute of Management Founded:1949 Mode of delivery: Blended Duration of the course: 12 months Number of hours: 300 Cities of Operation: Jamshedpur Course Fees: ₹  2,60,000 +GST XLRI has visualised itself to be a partner in the liberation and development journey of the independent India. Over many years XLRI has developed its own identity and is one of the most premier institutes. Parameter 1: Course Content (Rating 4.3) Comprehensiveness: One of the most comprehensive programs in analytics, it offers 13 core courses in four broad areas of descriptive analytics, predictive analytics, prescriptive analytics and application based courses. It provides hands-on in tools such as R, Excel, Python. Students can also opt for specialisation certificate apart from the main PGCBAMD certificate, which could be in finance, marketing, operations, HR. They do not have capstone projects. The course is updated every year. Access students at the end of the course: The assessments are done based on assignments, quizzes, midterm and end terms exams. Attendance and class participation are also used as a criterion by some faculties. Learning resources: Books, downloadable resources, LMS, question papers, study materials in soft copy format like PDF, ppt etc. Parameter 2: Faculty (Rating 4.4) Total no. of faculty members: 18 Total no. of faculty members with a PhD: 16 Total no. of faculty members with industry experience: 18 Student to faculty ratio: 36:1 Parameter 3: Student Experience (Rating 4.3) Percentage of students who completed the course: 95 Placement assistance: None Parameter 4: Other Attributes (Rating 4.3) Entry criteria for students: Based on eligibility criteria and personal Interview External collaboration: Apart from the regular faculties of XLRI, the program also has instructors who are either industry experts or are part of some other esteemed institutes like XIMB, TAPMI etc. The overall rating is 4.32 7. PGDM\/MBA in Business Analytics By REVA Academy for Corporate Excellence, REVA University Founded: 2012 Mode of delivery: Blended Duration of the course: PGDM – One year, MBA –  Two years Number of hours: 960 Cities of Operation: Bangalore Course Fees: PGDM- ₹ 3.2 Lakhs, MBA- ₹ 4.2 lakhs RACE is an initiative of REVA University, created to develop visionary enterprise leaders for corporates. RACE offers a range of specialised, techno-functional programs specially designed to suit the needs of working professionals to enhance their careers. The flagship program in Business Analytics is well recognized by the industry. Parameter 1: Course Content (Rating 4.4 ) Comprehensiveness: The program designed with IBM provides hands-on training on relevant software tools and business analytics framework across industries. The program covers descriptive, predictive, prescriptive and cognitive analytics in real-life scenarios. The participants work on 15 plus tools including Python, R, TensorFlow, Keras, Advanced MS-Excel, Rapidminer, Tableau, QlikView, IBM Cognos and others. They have capstone projects of more than 3 months. The course is updated every 6 months. Access students at the end of the course: This program follows a continuous assessment and grading pattern. Evaluations are done both internally and externally. Internal assessment is based on in-class quiz, discussions, participation in the discussion, while external is based on off-campus projects. Learning resources: Books, downloadable resources, LMS, videos, question papers, online repository of 3,000+ electronic journals, access to IEL online database. Parameter 2: Faculty (Rating 4.3 ) Total no. of faculty members: 14 Total no. of faculty members with a PhD: 4 Total no. of faculty members with industry experience: 14 Student to faculty ratio: 4:1 Parameter 3: Student Experience (Rating 4.2 ) Percentage of students who completed the course: 90 Post-completion engagement: All the participants get to enjoy lifelong access to the LMS, job portal, hackathons, information seminars etc. They also get to write blogs, white papers and research articles in various conferences and for RACE online media. Placement assistance: Assistance through alumni network, resume designing, helping them with interview process. Parameter 4: Other Attributes (Rating 4.2 ) Entry criteria for students: Entrance exam and industry experience of a minimum of 4 years. External collaboration: The program is delivered in association with IBM. Other delivery and knowledge partners include Predictive Analytics Pvt. Ltd,  Jigsaw Academy. The overall rating is 4.27 8. Executive Program in Big Data & Machine Learning  By NMIMS Global Access School For Continuing Education Founded: 1994 Mode of delivery: Online Duration of the course: 9 months Number of hours: 120 hours Cities of Operation: Presence across 22 cities with 75+ Authorised Enrolment Partners in India Course Fees: ₹ 1,10,000 NGASCE is India’s leading management university trusted by 35,000+ students and over 5,000 alumni members are recognised across 2,500 corporates. The programs are designed to suits requirements of working professional and provide features such as learning on the go with mobile app. Parameter 1: Course Content (Rating 4.2 ) Comprehensiveness: The big data and machine learning executive program is powered by SAS. It covers foundation course (business statistics, basic analytics tools and concepts like predictive, descriptive analytics), ANOVA, regression analytics, logistic regression, decision trees, cluster analysis, neural networks and others. They do not offer capstone projects, and course is updated every 6 months. Access students at the end of the course: Proctored exams conducted at designated exam centres for each module with minimum passing percentage of 50. For practical learning, students are asked to present a case analysis in their respective specialisation. Learning resources: Books, downloadable resources, LMS, videos, question papers, 24\/7 software access via AWS. Parameter 2: Faculty (Rating 4.3 ) Total no. of faculty members: 10 Total no. of faculty members with a PhD: 5 Total no. of faculty members with industry experience: 10 Student to faculty ratio: 1:3 Parameter 3: Student Experience (Rating 4.1 ) Percentage of students who completed the course: 60% Post-completion engagement: Students are provided with a certified e-badge that can be tagged on professional networks such as LinkedIn and other social media pages. Placement assistance: Partial assistance through alumni network Parameter 4: Other Attributes (Rating 4.1 ) Entry criteria for students: STEM background required External collaboration: This program is designed in collaboration with SAS. The overall rating is 4.17 9. Post Graduate Executive Certificate Program (PGEP) in Data Science & Big Data By IMT Founded: 1980 Mode of delivery: Blended Duration of the course: 6 months Cities of Operation: Ghaziabad Course Fees: ₹ 80,000 + GST* IMT Ghaziabad is a fully autonomous university and offers several postgraduate, doctorate and executive education programmes in management. Parameter 1: Course Content (Rating 4.1 ) Comprehensiveness: The curriculum combines academic excellence and industry relevance to facilitate the participants learn analytics and big data tools (Apache), followed by in depth and advance level of statistical and quantitative. Students are given hand-on exposure to tools such as R, SAS, Python, Hive, Cloudera, data visualisation tools, and others. It includes capstone projects with top analytics companies. The course is updated every year. Accessing students at the end of the course: Continued evaluation of students throughout the course. Learning resources: Downloadable material, access to analytics labs during tenure of course Parameter 2: Faculty (Rating 4.0 ) 98% of in-house faculties having PhD. Parameter 3: Student Experience (Rating 4.1) Percentage of students who completed the course: 90% Post-completion engagement: Candidates get IMT MDP alumni status. ICDM, a data management conference is organised by the institute. Placement assistance: Students opting for this course are provided with placement support. Parameter 4: Other Attributes (Rating 4.0) Entry criteria for students: Industry experience External collaboration: IMT has many industry associations and international collaborations. IMT has 50+ International collaborations towards various academic modules co-teaching and joint learning. The overall rating is 4.05 10. Post Graduate Certificate in Business Analytics & Big Data By IFMR Founded: 1970 Mode of delivery: Classroom Duration of the course: 10 months Number of hours: 266 Cities of Operation: Chennai and Sricity Course Fees: ₹ 2.5 Lakhs IFMR has been successfully offering PhD programs, MBA programs and executive programs for more than 2 decades now. It was established in 1970 as a not-for-profit society and has been sponsored by ICICI, the House of Kotharis and other major industrial groups. The board includes experts from Yale University, IISC Bangalore, and others. Parameter 1: Course Content (Rating 4.0) Comprehensiveness: Curriculum encompasses three important dimensions— mathematics & statistics, programming and application domains. It includes programming skills such as R and Python, predictive analytics, data mining and machine learning. Participants are provided hands-on sessions on SPSS Statistics, SPSS Modeler etc. It includes capstone projects of 1-3 months and course is updated every year. Accessing students at the end of the course: Participants have to take a written examination after the completion of each module, which is evaluated by faculties. Candidates also have to submit POC during the last 3 months, which is also evaluated. Learning resources: Books, downloadable resources, videos, question papers, class notes, powerpoint slides Parameter 2: Faculty (Rating 4.1) Total no. of faculty members: 16 Total no. of faculty members with a PhD: 7 Total no. of faculty members with industry experience: 11 Student to faculty ratio: 1:1 Parameter 3: Student Experience (Rating 3.9) Percentage of students who completed the course: 90 Post-completion engagement: Alumni group is set up for continued engagement Placement assistance: Partial assistance through alumni network, resume designing, helping them with interview process, setting up interviews for the participants Parameter 4: Other Attributes (Rating 3.9) Entry criteria for students: Industry experience required External collaboration: We have a collaboration with IBM for joint certification for the program. The IBM collaboration gives us specific expertise in some industry domains such as BFSI, Telecom, Health care and manufacturing. The overall rating is 3.98 Your Opinion Matters [poll id=”18″]","excerpt":"Read Our Latest ranking for 2020. A primary survey was conducted in August through September 2018, where 961 current and past students from 18 cities in India gave opinions on data science courses they attended. Out of these entries, few dozens in-depth interviews were done to understand the reasons for the submissions. This exercise helped […]","categories":["AI Features"],"tags":["apm data science"],"author_name":"Srishti Deoras","publish_date":"2018-10-31T05:44:03","publication_year":"2018","word_count":3774,"keywords":["data science","machine learning","artificial intelligence","Keras","AI","neural network","ML","RAG","apm data science","analytics","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","data science","analytics","TensorFlow","Keras","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-executive-data-science-courses-in-india-ranking-2018\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":11031,"title":"Don’t mix Analytics with Artificial Intelligence","content":"Artificial Intelligence is currently the most trending technology term in the industry. Though Artificial Intelligence is not a new concept, what comes as a baffling surprise is that analytics is being re-christened as artificial intelligence today. A lot of it to do with companies \/ consultants trying to on-board the frenzy around artificial intelligence. But, analytics and artificial intelligence are completely different technologies with some similarities in implementation and almost little overlap in terms of end results. Analytics as an industry is already heavily jargonized and it important to clear the air building up around analytics being called artificial intelligence at lot of instances. First, the overlaps Analytics is an encompassing field that uses mathematics, statistics, computing and machine-learning techniques to discover insights hidden deep in the recorded data. Analytics helps enterprises understand their current scenario and discover the future steps to achieve growth as analysis of data fuels knowledge discovery. Artificial intelligence is about creating systems (call them machines) that can mimic human intelligence as closely as possible. How close? – Well, lets say a Turing Test can tell that. Artificial Intelligence is not a new phenomenon; it’s almost as old as computing itself. We already have seen similar frenzy around AI like today’s, atleast twice in the past followed by deep AI winters where funding dried up for all research and initiatives around AI. But this time around, it’s looking different for AI. Partly because there are some concrete results that have surfaced. Though the real AI where computer systems can reach the level of human intelligence and behavior is still out of reach, the reason AI today has better results is because it is using the same concepts that analytics is using. Earlier efforts in AI were centered around creating expert systems, which at the most basic levels were nothing more than rule-based algorithms. What’s happening today in AI is almost the same, but with much higher sophistication. Today, AI systems are using power of data, computing and statistics to create intelligence systems. Sounds like that’s exactly what analytics is. But, the similarities end here. Now, the differences When we hear a computer winning a chess match against the world’s best chess champion, we say the computer is intelligent. When our smartphone understands our voice commands, we call it intelligent. These are all forms of artificial intelligence and are build on top of a combination of data, computing and machine learning. On the other hand, recommendation engines, forecasting loan defaults or predicting fraudulent transactions or spam emails are applications of analytics and are also build on top of data, computing and machine learning. But these applications cannot be termed artificial intelligence. Organizations only need to know the right question and analytics will give them the outcome they need to focus on. Analytics helps to produce fast insights required to make fact-based decisions and allows us to find out answers to questions we might never have thought to ask. Artificial Intelligence can be explained as the study of man-made computational devices and systems which can be made to act in an intelligent manner. The very core of both AI and analytics can be same in terms of how they are being implemented. Yet, what they are used for, have completely different objectives. It would only be fair on our part to keep these terms to what they really mean. This would also bode well for the whole industry in long term.","excerpt":"Artificial Intelligence is currently the most trending technology term in the industry. Though Artificial Intelligence is not a new concept, what comes as a baffling surprise is that analytics is being re-christened as artificial intelligence today. A lot of it to do with companies \/ consultants trying to on-board the frenzy around artificial intelligence. But, […]","categories":["IT Services"],"tags":[],"author_name":"Дарья","publish_date":"2016-10-26T06:04:02","publication_year":"2016","word_count":573,"keywords":["Go","funding","machine learning","artificial intelligence","programming_languages:R","AI","programming_languages:Go","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","R","Go","GAN","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/dont-mix-analytics-artificial-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10129248,"title":"Kling AI Brings Animals and Noodles Together in Viral Sensation","content":"In a recent social media post, Ahsen Khaliq has captured the hearts of internet users worldwide. The post features a series of animals enjoying a meal of noodles, but there’s a twist – these images are the result of using Kling AI. https:\/\/twitter.com\/_akhaliq\/status\/1812573686152450397 From a panda munching on a bowl of ramen to a kangaroo slurping up some udon, the images are both hilarious and heartwarming. The level of detail and realism in the images is a testament to the capabilities of Kling and the progress made in the field of AI. The post has gone viral, with thousands of users reacting and sharing the images. Many have expressed their amazement at the quality of the images and the creativity of the concept. Developed by the Chinese tech giant and TikTok competitor, Kling is available for all users. It outperforms the competition by generating videos up to two minutes long in 1080p at 30 frames per second. This is a significant improvement, which produces only a few seconds, and double what OpenAI’s Sora is expected to handle. While the US is debating heavily on AI ethics and incorporating ‘Responsible AI’, China seems unperturbed and is probably responding to these AI ethicists with Kling. The Chinese Alternative to Sora As an answer to OpenAI’s Sora, Chinese technology company Kuaishou introduced it as a new text-to-video AI model capable of generating high-quality videos. The model can create large-scale realistic motions which essentially simulate physical world characteristics and has the ability to produce 3D face and body reconstruction backed by the company’s proprietary technology, allowing users to create videos in various aspect ratios. AI video creation seems like the next battleground for tech companies with contenders like OpenAI’s Sora, Adobe’s Firefly, Midjournery, and Odyssey, already in the game. Now that Kling is here, the benchmark of making cinematically impressive and real-world-like videos has gone up.","excerpt":"Developed by the Chinese tech giant and TikTok competitor, Kling is available for all users.","categories":["AI News"],"tags":["kling","Sora"],"author_name":"Tarunya S","publish_date":"2024-07-16T12:35:34","publication_year":"2024","word_count":312,"keywords":["Go","OpenAI","AI","programming_languages:R","programming_languages:Go","kling","AI ethics","responsible AI","ViT","Sora","R"],"extracted_tech_keywords":["AI","OpenAI","R","Go","ViT","AI ethics","responsible AI","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/kling-ai-brings-animals-and-noodles-together-in-viral-sensation\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10112958,"title":"Nearly 75% Enterprises Pivoted to Text-Based Generative AI to Improve Operational Efficiency","content":"According to the recent “Generative AI in CXM Survey Report” by Everest Group and WNS, 75% of enterprises are elevating their business strategies by piloting, deploying, or scaling up text-based generative AI solutions, followed by 62% for code generation and 52% for image generation. The potential of this is widely recognised, with more than 90% of enterprises believing in the high potential of text generation, and around 70% for code and image generation. In terms of deployment, about 75% are piloting, deploying, or scaling up text generation solutions, and nearly 60% are doing the same for code generation. Growth Drivers The generative AI adoption in customer experience (CX) operations is driven by the need to enhance customer satisfaction and operational efficiency. Enterprises are increasingly integrating tech into CX management to personalise and customise customer interactions by understanding individual preferences and behaviors. This leads to more engaging and gratifying customer experiences. “When generative AI tailors interactions to individual preferences, it improves operational efficiency by automating tasks and intelligent tools like agent assist and language translation. This helps improve productivity, reduces response time, and lets agents focus on value-added tasks, ultimately resulting in more satisfying customer experiences,”  Sanjay Jain, chief business transformation officer, WNS, told AIM. Furthermore, it boosts efficiency within CX operations. It equips customer support agents with intelligent tools such as agent assist, next-best-action recommendations, language translation, cross-sell, up-sell, and accent neutralisation. Another driver is its ability to analyse vast amounts of customer data, enabling enterprises to extract meaningful insights for competitive advantage. This facilitates data-driven decision-making for strategy development, product improvements, and service enhancements. However, 45% of enterprises report a shortage of internal technical expertise in generative AI which hinders their ability to come up with new models. Talking about the AI talent gap, Jain said that investing in strategic talent acquisition and development is crucial for maximising the potential of AI in CXM. “Enterprises need to focus on recruiting AI and ML engineers, data scientists, and software developers, especially those with a mix of AI and software development skills to enhance teamwork and communication. Additionally, ongoing training programs are vital to ensure that current employees are upskilled,” said Jain. He also added that AI, despite fears of job displacement, acts as an enhancer rather than a substitute for human skills. “While AI excels in accuracy and speed for execution, human critical thinking is pivotal for oversight,” added Jain, highlighting that this symbiotic relationship promises long-term benefits. Key Findings of the Report The survey highlights the expanding role of tools like OpenAI’s ChatGPT, Google’s Bard Gemini, and Microsoft’s Copilot in business transformation. These advancements are driving a significant shift in CXM, with generative AI’s content generation, data analysis, and insight extraction capabilities leading the way. Key findings from the survey cover the awareness, perceived potential, and deployment areas for generative AI applications. It also covers the technology, process and enterprise readiness for foundational models. Everest Group’s research included over 200 companies from diverse sectors such as telecom, media, BFSI, healthcare, retail, and technology, including FGT\/Hi-Tech industries. These companies, predominantly from regions including Asia Pacific, have annual revenues exceeding $500 million. Awareness and perceived potential of generative AI applications Text generation capabilities: Over 75% of enterprises have high awareness. Code generation capabilities: 62% awareness. Image generation capabilities: 52% awareness. Over 90% of enterprises believe in high potential for text generation. About 70% see high potential for code and image generation capabilities. Planned Deployment Areas Text generation capabilities: About 75% of enterprises actively piloting, deploying, or scaling up solutions. Code generation capabilities: Nearly 60% of enterprises have initiated pilot programs or implementation. Readiness for Generative AI Nearly 50% of enterprises are concerned about having sufficient computing power. Over 70% say they have adequate cloud capabilities. Approximately 40% express concerns about the availability of high-quality training data. Over 60% express significant concerns regarding data security. More than 45% of enterprises report a shortage of internal technical expertise. Over 70% of BFSI and healthcare sectors noted regulatory compliance issues. 40% identify cultural inertia as a major obstacle. Only two-thirds have adequate capability for redundancy and failover measures. Enterprise Readiness for generative AI by Industry BFSI: Approximately 60% prepared across technology, people, process, and change management. Healthcare: 60-70% readiness across technology, data, process, and change management. Retail: Least prepared, with key challenges in computing power, training data, and talent. Technology and FGT\/Hi-Tech: Over 60% significantly ready across key parameters. Telecom and media: Approximately 65% highly ready across favorable parameters for generative AI implementation. Read more: Data Science Hiring Process at WNS","excerpt":"The WNS report states that over 45% of enterprises identify the lack of internal technical expertise as a major barrier to generative AI implementation.","categories":["AI News"],"tags":["ChatGPT"],"author_name":"Shritama Saha","publish_date":"2024-02-15T13:58:00","publication_year":"2024","word_count":757,"keywords":["data science","ChatGPT","Go","OpenAI","AI","ML","GPT","Aim","generative AI","R"],"extracted_tech_keywords":["AI","ML","data science","generative AI","ChatGPT","OpenAI","Aim","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nearly-75-enterprises-pivoted-to-text-based-generative-ai-to-improve-operational-efficiency\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":52528,"title":"Contextual Targeting Is Making A Comeback, But It Has Its Own Challenges","content":"World of online ads is changing, with lots of brands looking into the past to revive what we call Contextual Targeting. Contextual targeting was almost obsolete with the advent of the third-party cookies. Why brands and ad companies opt for cookies is that it helps them predict the behaviour of the user based on past activity online. Cookies use some user data to show ads online; that way it predicts the future activity and displays an ad irrespective of the content of the page you are currently on. But now this is part of the reason why contextual targeting is making a comeback. With rising concerns over private data from users, the parameters across user private data streamlines and makes it harder for cookies to operate as it becomes tough to understand the user behaviour, this opens the window for contextual targeting’s comeback. What is Contextual Targeting? Contextual targeting is a form of personalised advertising technique that lets your ads appear on relevant sites. Contextual apart from easing the data concerns helps consumer engagement and show them ads based on the content of the website they are on. According to a HubSpot study, 64% of respondents said that ads are annoying and intrusive, plus 45% of people reported that they don’t notice online ads even if they’ve blocked them. Contextual advertising works through keywords and topics, or you can say the central theme of a website. If you’re someone who is getting started with your PPC, you can select high priority keywords and topics for your ad. Now, when you do this your ad will show on sites related to those keywords. So, one might say that these keywords decide the fate of your PPC ads. For example, if you’re putting up an ad for bike sale through contextual targeting, then you can have keywords like; motorcycle, motorcycle service, bike service etc. Contextual Targeting vs Behavioral Targeting Contextual is done through keywords and topics, and behavioural targeting is based on the online behaviour of the user. Behavioural targeting can include browsing history, clicked links, how recent the search is, the time spent on the page, and how a user has engaged with the site. Let me take my example, I love shoes and especially the basketball shoes, say I search for Nike basketball shoes online in the morning, and I may or may not carry out an extensive search (usually I am very thorough with shoes). But, later that night I may be on a different site reading an article about sci-fi and I all I see is shoe ads in my window. This is behavioural targeting. When it comes to contextual, I see shoe ads only on shoe-related websites. Here contextual targeting is working through only one thing, and that is the keyword. Problem with Contextual Targeting While contextual advertising works on keywords, the process for targeting keyword has to be precise, which is a time-intensive process. The contextual depends heavily on the keyword choice, one has to give keywords to your segment, and the engine will only include websites corresponding to only those keywords and nothing else.Contextual advertising requires constant attention because some contexts are too vast for precise targeting, and the ads may get replaced by a competitor’s ad.Contextual advertising, sometimes, pairs the wrong kind of ads, for example; if someone like me is reading about a Rolls Royce online might not want to buy one.Problems when it comes to implementing frequency capping. Frequency capping in simple words is limiting the number of times a specific visitor is shown a particular ad on a particular website. Contextual targeting is independent of complying with GDPR and storing any tracking cookies\/identifiers.The branded and organic contextual, relevant advertising is not easy to do.The contextual targeting technology is underfunded when compared to behavioural based tech. For contextual targeting to have a fair chance, companies need to invest in its technology. Outlook Although when you look at it from the top, contextual targeting feels like an excellent alternative to behavioural targeting, it does have its problems to tackle. But, in recent times this form of advertising is rising with some integration of deep learning with it which can understand the true meaning behind the keywords according to the context which will replace the tedious listing of keywords. It is yet to be determined whether contextual targeting will be able to come out on top and whether 2020 will be the year contextual targeting will completely take over online advertisements from cookies.","excerpt":"World of online ads is changing, with lots of brands looking into the past to revive what we call Contextual Targeting. Contextual targeting was almost obsolete with the advent of the third-party cookies. Why brands and ad companies opt for cookies is that it helps them predict the behaviour of the user based on past […]","categories":["AI Features"],"tags":[],"author_name":"Sameer Balaganur","publish_date":"2019-12-24T17:18:41","publication_year":"2019","word_count":749,"keywords":["programming_languages:R","AI","ML","deep learning","ViT","GAN","R"],"extracted_tech_keywords":["AI","ML","deep learning","R","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/contextual-targeting-making-comeback-but-it-has-its-own-challenges\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10160791,"title":"Cyara Opens Global Innovation Centre in Hyderabad","content":"Cyara, the leading US-based AI-powered Customer Experience (CX) Transformation Platform provider, has launched its new Global Innovation Centre (GIC) in Hyderabad. The new facility is set to generate over 200 jobs and tap into India’s vast talent pool. The centre will focus on research and development, product innovation, and expanding the company’s global presence. Cyara aims to use this new hub to enhance its CX solutions and maintain its position as an industry leader. “Hyderabad’s dynamic tech ecosystem is the perfect place for us to grow and innovate,” according to the company’s LinkedIn post. The move also reflects Cyara’s dedication to contributing to the local community while advancing CX technology globally. This expansion underscores Hyderabad’s rising status as a leading destination for global technology companies seeking to innovate and collaborate. The company recently announced Rishi Rana as its Chief Executive Officer (CEO). Innovation at Helm Cyara’s CX Assurance platform helped a major U.S. retailer turn Black Friday challenges into a success story. Mid-day on the critical sales day, the platform alerted the retailer to system issues, including call delays and problems with a third-party credit card processor. With real-time insights, including recordings and call details, the retailer quickly identified and resolved the issues, saving hours of manual diagnosis. This swift action ensured a seamless customer experience and a successful holiday season. Prior to Cyara, the retailer’s QA team relied on time-intensive manual testing, resulting in delays and defects. Automation through Cyara transformed their processes, increasing testing coverage from 15% to 85% and reducing test execution time by 97%. Agents were freed from testing to focus on core responsibilities, while real-time feedback replaced post-event surveys. The retailer praised Cyara’s impact, stating, “Cyara has transformed our business as we can now roll out changes and test to the feature level in minutes.” Beyond Black Friday, the platform has driven broader improvements, enabling faster launches and supporting the retailer’s digital transformation efforts to compete with leading online retailers.","excerpt":"The new facility is set to generate over 200 jobs and highlights Cyara’s commitment to tapping into India’s vast talent pool.","categories":["AI News"],"tags":["Automation","customer experience"],"author_name":"Shalini Mondal","publish_date":"2025-01-06T15:11:27","publication_year":"2025","word_count":325,"keywords":["programming_languages:R","AI","innovation","ML","Automation","digital transformation","Git","RAG","customer experience","automation","Aim","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Git","digital transformation","automation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cyara-opens-global-innovation-centre-in-hyderabad\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":50362,"title":"Can Google Stadia Take Over All The Existing Cloud Gaming Services?","content":"Gaming has been one of the fastest-growing industries this decade. The industry has witnessed a lot of innovations over the years, among which Cloud Gaming is a notable one. The prime idea of cloud gaming is to allow people to play games on any device from anywhere without the need to have a high-end expensive PC. How Cloud Gaming Works? Cloud computing is almost similar to how OTT platforms work. When a user is on an OTT platform, it sends a request to a server that contains the content the user wants to watch. The server responds to the request by playing the content on the platform. In cloud gaming, it is almost the same scenario. When a user selects a game on the platform it is sending a request to the server, and then the server allows the user to play the game. Just like traditional gaming, the gamer in cloud platform also uses commands and consoles. However, the only difference is that the process is happening through a network and the time taken to execute is significantly less. Cloud Gaming Players Talking about the players in this industry, services like Shadow have been leading for a long time. The reason behind this success is not only features but the way its services are structured. Shadow allows its users to have a dedicated cloud gaming computer. Meaning, unlike other cloud gaming services, users don’t have to worry about sharing a cloud with other users. And this is what makes Shadow strong when it comes to providing a smooth gaming experience. GeForce NOW is another platform that is gaining traction, despite being in a beta state. NOW is solely focused on allowing underserved gamers with not-so-suitable PC’s play games on their systems. It supports over 400 top games which also includes Player Unknown’s Battlegrounds and Fortnite and supports consoles like DualShock 4, Xbox One controller, Xbox 360 controller and Logitech Gamepad F310, F510 and F710. Google Stadia While services like Shadow and GeForce NOW are still gaining traction, Google has also ventured into space with its all-new cloud gaming service Stadia that lets you play big-budget games without discs or downloads, consoles or gaming PCs. All you need to do is buy the games on the platform and have a decent internet connection. And ever since the news broke out, it has become really popular among video game enthusiasts. Reportedly, Stadia is much more reliable than other services which are already available in the market. Google over the years has taken over a lot of industries, and now that the tech giant has entered the cloud gaming space, the expectations are high. To stand to those expectations, the search giant has announced that Stadia, which is to be launched this week, will have 22 games ready to be played on the debut day itself. These games include: Assassin’s Creed Odyssey Attack on Titan: Final Battle 2 Farming Simulator 2019 Final Fantasy XV Cloud Gaming In India Stadia is a talk of the town at present and 14 countries are going to make the most of this tech. But India is not on the list. What could be the reason? There was a time when the gaming industry in India was very unorganised. Unlike other nations, India didn’t have any platform that would promote this industry. Another reason behind this slow growth of the industry is the fact that gaming in India is expensive and many have even simply accepted it as fact. The hardware is the first and foremost thing a gamer has to invest in and there was a time when if you wanted to buy a PS4 or Xbox you would have to pay a considerably higher price compared to folks in the US. But with time the industry has started to become popular in India with prices of gaming hardware dropping and with the rising number of gaming leagues. Popular games like PUBG and Call of duty have also recently witnessed a great number of enthusiasts in India. The cloud gaming domain in India is still in the initial stage despite having players like GeForce NOW and Vortex in the Indian market. Now, it’s all about the time when we would see cloud-based gaming and streaming services become mainstream in India, just like OTT.","excerpt":"Gaming has been one of the fastest-growing industries this decade. The industry has witnessed a lot of innovations over the years, among which Cloud Gaming is a notable one. The prime idea of cloud gaming is to allow people to play games on any device from anywhere without the need to have a high-end expensive […]","categories":["Global Tech"],"tags":["Cloud Gaming"],"author_name":"Harshajit Sarmah","publish_date":"2019-11-20T17:00:00","publication_year":"2019","word_count":717,"keywords":["Go","Cloud Gaming","programming_languages:R","cloud computing","AI","innovation","programming_languages:Go","Git","GAN","R"],"extracted_tech_keywords":["AI","cloud computing","R","Go","Git","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/can-google-stadia-take-over-all-the-existing-cloud-gaming-services\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52380,"title":"CRISIL To Acquire Greenwich Associates","content":"CRISIL, an S&P Global Company will acquire Greenwich Associates LLC , a leading provider of proprietary benchmarking data, analytics and qualitative, actionable insights that helps financial services firms worldwide measure and improve business performance. Greenwich, a Stamford, Connecticut-based company, serves over 300 of the top investment banks, corporate banks, commercial banks, asset managers and key players in the market infrastructure space globally. The acquisition will complement CRISIL’s existing portfolio of products and expand offerings to new segments across financial services including commercial banks and asset and wealth managers. The deal will accelerate CRISIL’s strategy to be the leading player in the growing market of global benchmarking analytics. Greenwich’s deep expertise in ‘Voice of Customer’ surveys, proprietary data assets, and capabilities to source and aggregate high-value private performance data will provide a new layer of insights to CRISIL’s existing offering under its Coalition division. Greenwich’s portfolio of products includes MarketView, Fee Clearinghouse, ACCESSTM, Focus and Explorer. Ashu Suyash, Managing Director & CEO, CRISIL, said, “With this acquisition, CRISIL will strengthen its position as the foremost global benchmarking analytics provider for financial services, where Coalition is a distinguished provider to corporate and investment banks worldwide. We are very excited about bringing the two trusted brands together, and Greenwich’s rich data sets will augment Coalition’s proprietary data, enabling unique analytics and insights.” Steven Busby, CEO, Greenwich, said, “For nearly five decades, Greenwich has pioneered voice of the customer benchmarks and analytics across the global financial services industry. We are delighted to become a part of CRISIL. Our combined data and analytics will provide clients with high-value actionable insights they demand in today’s competitive business environment. We are delighted that, as a part of the CRISIL group, we will collectively unleash greater growth opportunities for our clients and our people.” The acquisition is subject to regulatory approvals and other customary closing conditions. The deal is anticipated to close in the first quarter of 2020. Greenwich’s partners and their team of approximately 150 people globally will join CRISIL following the completion of the transaction.","excerpt":"CRISIL, an S&P Global Company will acquire Greenwich Associates LLC , a leading provider of proprietary benchmarking data, analytics and qualitative, actionable insights that helps financial services firms worldwide measure and improve business performance. Greenwich, a Stamford, Connecticut-based company, serves over 300 of the top investment banks, corporate banks, commercial banks, asset managers and key […]","categories":["AI News"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-12-21T09:00:50","publication_year":"2019","word_count":339,"keywords":["programming_languages:R","AI","analytics","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","analytics","R","Rust","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/crisil-to-acquire-greenwich-associates\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":68813,"title":"How Microsoft Wants To Enhance Your Tech Skills By Integrating MS Learn, LinkedIn and GitHub","content":"More people are now out of work than ever since the Great Depression in the 1930s, and COVID-19 has made the skills gap even more. This, according to analysts, could potentially exacerbate economic inequity in the coming years as millions of people are bearing the brunt of this economic crisis. And for those who still have jobs, it’s changing how they work but what they work on. Experts say that every job will require increasingly digital skills in the next five years, and would need extensive upskilling to find employment. In this scenario, tech companies are working on various means to tackle this situation. Yesterday, in an announcement Microsoft said it is trying to revamp its technology ecosystem to help people find employment by reskills themselves. According to Microsoft, 149 million new tech jobs will be produced in fields like software development, cybersecurity and machine learning. In fact, Microsoft’s Brad Smith said the company estimates that in 2020 before the year comes to an end, almost a quarter of a billion people lose their jobs worldwide. Microsoft’s CEO Satya Nadella announced, “We are bringing together these assets to reimagine how people learn and apply new skills to help 25 million people who have lost their jobs due to COVID-19, acquire digital skills for jobs of the future.” This is a comprehensive technology initiative that will build on data and digital technology. It starts with data on jobs and skills from the LinkedIn Economic Graph. It provides free access to content in LinkedIn Learning, Microsoft Learn, and the GitHub Learning Lab, and couples these with Microsoft Certifications and LinkedIn job-seeking tools. These resources can all be accessed at a central location and will be broadly available online in four languages: English, French, German and Spanish.  The company is also pledging to make stronger data and analytics — including data from the LinkedIn Economic Graph — available to governments around the world so they can better assess local economic needs. The initiative is pulling together every part of the company, which starts with Linkedin – which is the heart and soul but combines LinkedIn efforts with Microsoft Learn, Github, and Github Learning Lab, working together in a new and important way.  Given the rise in demand for digital skills and for roles which are more technical in nature, Microsoft Learn combines short step-by-step training with interactive coding and scripting to help learners go deep on Microsoft technologies, and help them learn for Microsoft certifications. This means Linkedin job seekers that engage in Learning pass will have the opportunity to practise newly acquired skills by completing realistic projects in a personalised Github repo.  Job seekers pursuing developer roles will be able to access the Github Learning Lab to practise their skills, which is a bot-based learning tool and a repository to teach coding with real-life demo based modules. Github’s Learning Labs provides hands-on learning tutorials for aspiring as well as experienced developers. In May 2020, Learning Lab usage was up 900% compared to May 2019. Converging The Ecosystem For Tech Skills Training Microsoft has started a new portal — opportunity.linkedin.com — to tap in resources across the Microsoft ecosystem, with insights from real-time skilling trends using LinkedIn. Microsoft calls this the economic graph, which is also offered to not just the public but also policymakers and government. Based on the data and the job trends, Microsoft is making four learning paths from LinkedIn Learning available for free. The multi-faceted approach in this initiative include: Data to identify jobs and skillsAccess to learning content free of chargeLow-cost Certifications and free job-seeking tools First, Microsoft uses data from the LinkedIn economic graph, which is a digital representation of the global economy, to identify the ten most in-demand jobs like IT technician, data analyst and customer service specialist. Then, it will match people’s skills to the right jobs, providing access to low-cost certifications and free job-seeking tools, and also partnering with nonprofits to provide additional support. System learning creates a continuous feedback loop between the work, skills and learning required to succeed at the task at hand, and credentials for career advancement. “Microsoft ecosystem is uniquely positioned to help people connect opportunities. Job seekers and the government agencies that support them need the access to data to help identify where employment opportunities exist accessible tools to effectively develop relevant skills, and the ability to demonstrate skills to organisations who are hiring,” says Ryan Roslansky, CEO of LinkedIn. “By digitally mapping over 690 million professionals across 50 million companies, 11 million job listings, we are able to spot trends like in-demand skills, emerging jobs and global hiring patterns,” LinkedIn CEO further stated in a video announcement.","excerpt":"More people are now out of work than ever since the Great Depression in the 1930s, and COVID-19 has made the skills gap even more. This, according to analysts, could potentially exacerbate economic inequity in the coming years as millions of people are bearing the brunt of this economic crisis. And for those who still […]","categories":["Global Tech"],"tags":["linkedin","Microsoft"],"author_name":"Vishal Chawla","publish_date":"2020-07-02T12:04:51","publication_year":"2020","word_count":779,"keywords":["Go","machine learning","programming_languages:R","AI","programming_languages:Go","Git","GAN","analytics","linkedin","GitHub","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","analytics","R","Go","Git","GitHub","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-microsoft-wants-to-enhance-your-tech-learning-by-integrating-with-linkedin-and-github\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":25103,"title":"7 Humanoid Robots Which Were Made In India, And Their Success Stories","content":"Humanoids robots have been gaining popularity in India for quite some time now. Although the country is still catching up with the developments in artificial intelligence and robotics as compared to others, Indian startups, as well as the government, are working at a rapid pace to integrate new-age technologies. According to IFR research, robot sales in India increased by 27 percent to a new peak of 2,627 units in India — almost the same as in Thailand. Another survey claims that India ranks third in implementing robotic automation. Let us take a look at a few noteworthy humanoid robots that are designed and developed in India: Manav Manav is India’s first 3D-printed humanoid robot. The two-kilo, two-feet tall robot has an inbuilt vision and sound processing capability which allows it to walk, talk and dance — just in response to human commands. Developed by Delhi’s A-SET Training and Research Institute, the humanoid robot is primarily meant for research purposes and is made available to research institutes which offer robotics as a subject of study. Manav can also perform activities like push-ups, headstands and can also play football. It uses an open-source code so that it can also be taught to learn and respond like a human child. It also has WiFi and Bluetooth connectivity and has a rechargeable lithium polymer battery that can work for an hour with a single full charge. The parts of Manav are all made in India, the outer frame of the robot is made of plastic and was 3D printed from A-SET’s own 3D printing venue, Buildkart Retail. Mitra The first indigenously built humanoid robot is capable of interacting with humans smartly. The five feet-tall humanoid robot is made of fibreglass and is programmed to greet customers using contextual help, autonomous navigation and facial and speech recognition. It also has a touchscreen on its chest which can be used to interact where speech is not possible. It can work for eight hours on a single charge. It can also understand multiple languages. The humanoid robot was launched by Prime Minister Narendra Modi and Ivanka Trump, First Daughter and advisor to the President of the United States Donald Trump, at the Global Entrepreneurship Summit (GES) conference last year. Developed by a Bengaluru-based robotics startup Invento Robotics, the robot can be found floating in the corridors of the Canara Bank and PVR Cinemas in Bengaluru, chatting with the customers and making them feel welcome. According to the robot’s official website, the robot is as handy in the service sector as it is as a host. In fact, one can rent the robot for any party. Depending on the requirement, the startup customises the humanoids accordingly. Reportedly, Mitra was also showcased at an annual technology conference Slush 17 in Helsinki, Finland. During the event, the robot not only caught the attention of the audience but also received opportunities for partnership with a European firm. Robocop Hyderabad-based AI and ML startup H-Bots Robotics has developed a police robot to assist in handling the law, order, and traffic management. The life-sized robot, which was deployed last year in Hyderabad, is equipped with cameras and an array of different sensors like ultrasonic, proximity and temperature sensors. The robot is designed to protect and secure places like offices, malls, airports, signal posts and other public spaces and can take care of security if deployed autonomously. Reportedly, the Robocop can diffuse bombs too. The beta version robot is made in India using all Indian components. KEMPA Passengers visiting Bengaluru airport may soon be greeted by a special robot assistant. Built to suit the needs of the Kempegowda International Airport, the little bot assistant, named KEMPA, will answers queries of confused passengers in English as well as Kannada. The humanoid is built on AI by a Bengaluru-based startup Sirena Technologies. The advanced humanoid is completely designed and manufactured in Bengaluru. KEMPA is programmed to provide flight and check-in details and other information about flights. While the bot is still being tweaked and is getting ready for the official launch. It also suggests places to visit in the state also engages in casual conversation with passengers. RADA Vistara, a joint venture between Tata Sons and Singapore Airlines, has created a unique artificial intelligence-based robot called RADA to automate simple tasks and improve customer experience. According to a statement released by Vistara, the RADA will be placed at Vistara’s Signature Lounge at Delhi’s Indira Gandhi International Airport’s Terminal 3 from 5 July 2018 to assist customers before they board their flights. It will also help promote Vistara’s product and services with the help of distinct messages recited by the bot. RADA will be further developed over a period of time in terms of functionality and features for future use cases, after gauging customer feedback. It is conceived, designed and engineered by its team of technology experts and apprentices from Tata Innovation Lab with support from students of reputed institutions. Built on a chassis of four wheels, RADA can rotate 360 degrees and has three inbuilt cameras for cognitive interaction. Combining these components with an effective voice technology, Vistara has developed the robot to provide a simple solution to cater to the emerging and future trends. IRA Next time you enter the HDFC Branch; you may be greeted by a shiny white interactive humanoid called IRA (Intelligent Robotic Assistant). It would essentially help branch staff in servicing customers. Launched in Mumbai last year, IRA has been developed using robotics and AI technologies. Developed in partnership with Asimov Robotics, a start-up based in Kochi, IRA (v1.0), was made to greet the customers, guide them to relevant counter in the branch such as cash deposit, foreign exchange, loans, among others. The bank also announced the launch of IRA 2.0 this year, which has been developed in partnership with Invento Makerspaces and Senseforth Technologies. It has features such as voice recognition and face recognition, which is enabled using technologies such as AI and ML. INDRO This is reportedly the tallest humanoid robot built in India. Created by researcher Santosh Vasudeo Hulawale, INDRO is an autonomous robot was made inside a house with easily available low-cost material like aluminium, wood, cardboard, plastic etc. According to a research paper, INDRO can be used for lightweight tasks like entertainment, education and a few household works. The autonomous humanoid robot is not fully autonomous and can be controlled both autonomously and manually. It has 31 motors and can perform actions like a human. In addition, it can lift objects weighing up to 2 kilos with its hands. DRDO’s Daksh This made-in-India robot is primarily designed to detect and recover Improvised Explosive Devices (IEDs). Developed by Defence Research and Development Organisation (DRDO), the robot was inducted by Indian Army around 2011. Reportedly, 20 Daksh robots are already being used by the Indian Army. Using its X-ray vision, Daksh can identify a hazardous object and can diffuse it with a jet of water. Daksh is capable of climbing staircase and negotiating cross-country terrains and is capable of towing a suspected vehicle away from a crowded area. Additionally, it can be operated from a distance of 2.5 kilometres and can handle car explosives with its high-calibre shotgun. Reportedly, after it got an upgrade in 2015, it not only became lighter, faster and rugged, but has also been equipped with chemical, biological, radiological and nuclear hazard detection mechanism. The new Daksh is made of aluminium alloy which has reduced the weight and has become three-time faster, compared to the older version, which was made of steel. Conclusion India’s robotics industry is still small when compared with those of South Korea, Japan, US and China. Nonetheless, there are only three robots per 10,000 employees in India. But it is only a matter of time before the country becomes a major player in robotics design and development. India already has many of the basic elements in place to become a robotics industry, including established business, academic research, government support and an increasingly entrepreneurial business community.","excerpt":"Humanoids robots have been gaining popularity in India for quite some time now. Although the country is still catching up with the developments in artificial intelligence and robotics as compared to others, Indian startups, as well as the government, are working at a rapid pace to integrate new-age technologies. According to IFR research, robot sales […]","categories":["AI Trends"],"tags":["humanoid","Humanoid Robots","robot"],"author_name":"Smita Sinha","publish_date":"2018-06-04T09:11:42","publication_year":"2018","word_count":1335,"keywords":["Go","API","artificial intelligence","Humanoid Robots","AI","robot","ML","Ray","Aim","ViT","GAN","R","humanoid"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","Ray","R","Go","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-humanoid-robots-which-were-made-in-india-and-their-success-stories\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":31073,"title":"Essential Game Theory Books For Any Data Scientist","content":"The Game Theory has numerous applications in biology, economics and even artificial intelligence. The latest result of game theory’s application in machine learning is generative adversarial networks which have revolutionised the field of videos and images analysis. To put it simply, game theory is the study of mathematical models of strategic interaction between different kinds of decision makers. The game theory is also used predominantly in studying the behaviours of agents and their relations. Many advances in AI such as achievements in the game of GO, Jeopardy and Chess are tied to the use of game theory in chess. Here is a list of books that can help a data scientist to start learning game theory: 1. Essentials of Game Theory: A Concise Multidisciplinary Introduction By Kevin Leyton-Brown and Yoav Shoham  in 2008 This book is great because it covers a lot of areas relevant to people from many fields. The book brilliantly encapsulates a short and crisp introduction of game theory for readers of all fields. The introduction to game theory that this book gives is unparalleled and any professional reading the book can start formulating ideas for their fields after reading this book. The book minimises notation, ruthlessly focus on essentials and is still rigorous. The book also covers the main classes of games, their representations, and the main concepts used to analyse them. 2. Computational Aspects of Cooperative Game Theory By Georgios Chalkiadakis, Edith Elkind and Michael Wooldridge in 2011 Cooperation between agents is really important to accomplish great tasks. Cooperative game theory is a branch of economics and game theory that studies the behaviour of self-interested agents. These behaviours are studied in strategic settings where binding agreements among agents are possible and the book starts by defining transferable utility games is easy terms. The book then discusses two major issues that arise when considering such games from a computational perspective: identifying compact representations for games, and the closely related problem of efficiently computing solution concepts for games. 3. Rules of Encounter: Designing Conventions for Automated Negotiation among Computers By Jeffrey S. Rosenschein and Gilad Zlotkin in 1994 The book provides a unified, coherent account of machine interaction at the level of the machine designers (the society of designers) and the level of the machine interaction itself (the resulting artificial society). The book talks about the theory of high-level protocol design to enable machine interaction that is goal achieving. The authors of the book say, “While game theoretic ideas have been used to answer the question of how a computer should be programmed to act in a given specific interaction, here they are used in a new way, to address the question of how to design the rules of interaction themselves for automated agents” 4. Multiagent Systems: Algorithmic, Game-Theoretic, and Logical Foundations By Yoav Shoham and Kevin Leyton-Brown in 2009 Agents talking to each other and doing tasks together is very important for the future of AI. This book talks about how multi-agent systems can work. The book talks about how multiagent systems combine multiple autonomous entities, each having diverging interests or different information. The book also gives a very impressive overview of the fields from a computer science perspective and goes on about ideas from game theory, economics, operations research, logic, philosophy and linguistics. 5. Theory of Games and Economic Behaviour By John von Neumann and Oskar Morgenstern in 1944 You can not get a better book to introduce yourself to game theory because this book is seen as the book that introduced game theory to the world. This book is a classic work upon which modern-day game theory is based and rightly is the most influential book. The introduction to the book says, “In it, John von Neumann and Oskar Morgenstern conceived a groundbreaking mathematical theory of economic and social organisation, based on a theory of games of strategy. Not only would this revolutionise economics, but the entirely new field of scientific inquiry it yielded — game theory — has since been widely used to analyse a host of real-world phenomena from arms races to optimal policy choices of presidential candidates, from vaccination policy to major league baseball salary negotiations.”","excerpt":"The Game Theory has numerous applications in biology, economics and even artificial intelligence. The latest result of game theory’s application in machine learning is generative adversarial networks which have revolutionised the field of videos and images analysis. To put it simply, game theory is the study of mathematical models of strategic interaction between different kinds […]","categories":["AI Features"],"tags":["books","Data Scientist","game theory"],"author_name":"Abhijeet Katte","publish_date":"2018-12-05T05:52:25","publication_year":"2018","word_count":695,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","programming_languages:Go","game theory","multi-agent systems","books","GAN","Data Scientist","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","multi-agent systems","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/essential-game-theory-books-for-any-data-scientist\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10165833,"title":"New OpenAI Report Shows How to Fix Reward Hacking in Large Reasoning Models","content":"OpenAI has released a research report which indicates that frontier reasoning models frequently engage in reward hacking, where AI agents exploit loopholes in their objectives to gain rewards in unintended ways. The report also explores strategies to monitor, and mitigate these behaviors. OpenAI shows how to monitor a reasoning model like the o3-mini for reward hacking in agentic coding environments, by using another LLM to observe its chain-of-thoughts (CoT). OpenAI said GPT-4o, a model weaker than o3-mini, could effectively monitor the stronger model. “As we have continued to scale RL training, agents have discovered more complex and hard-to-detect hacks,” said OpenAI in the report. The company found several effective reward hacks in coding processes. For example, an AI agent exploits a bug that allows it to exit the environment prematurely to earn the rewards. OpenAI said monitoring a model’s CoTs is significantly more effective than monitoring its actions alone, and said that this method presents promising evidence of providing a new avenue for scalable oversight. In comparison to the action-only monitor, which caught (recalled) 60% of the hacks, CoT monitors recalled 95% of the hacks. While CoT monitors work effectively at low levels of reinforcement learning (RL) optimisation, OpenAI said that when the model is further trained to maximise rewards, it learns to fool the monitoring system. “Applying direct optimisation pressure to the CoT can quickly lead to obfuscated reward hacking, where agents hide their intent while continuing to misbehave,” read a section of the report. Due to such limitations of CoT monitoring, OpenAI said developers may have to accept a trade-off by deploying less performant models, or suffering more expensive inference. However, in a worst case scenario, if the agent learns to completely hide its intentions in CoT, OpenAI suggests going back to older methods of making AI safe, like analysing its activations, internal workings, and improving its alignment methods. A detailed PDF report indicating the findings of the research can be found in the following link. Nat McAleese, member of the technical staff at OpenAI, said in a post on X that “large reasoning models are extremely good at reward hacking”, and handpicked examples from the report to illustrate his statement. Similarly, there have been worldwide efforts to study the nature of reasoning models. Recently, a research led by Harvard University and Vrije Universiteit Brussel found that stronger models with fewer reasoning chains outperformed weaker ones, with extended chains of reasoning. The study concluded that OpenAI’s  o3-Mini outperformed the o1-mini, with fewer reasoning chains.","excerpt":"Monitoring the chain-of-thoughts of a reasoning model using another LLM is effective to a certain extent.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI reasoning","OpenAI"],"author_name":"Supreeth Koundinya","publish_date":"2025-03-11T17:00:34","publication_year":"2025","word_count":416,"keywords":["Go","OpenAI","AI","AI reasoning","GPT-4o","programming_languages:R","Scala","GPT","AI agents","R","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","R","Go","Scala","GPT","AI agents","llm_models:GPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/new-openai-report-shows-how-to-fix-reward-hacking-in-large-reasoning-models\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10089358,"title":"Much Awaited Breakthrough in LLMs Has Finally Arrived","content":"Every AI company would say that they want to build something like ChatGPT. After the release of the OpenAI’s API, the chatbot can now be integrated into business for building products. This has been helpful for a lot of developers and companies. But even then, building such a technology from scratch is a far off game as it requires a lot of resources. But things have changed now! Hugging Face has been at the forefront of the current open-source AI developments and it has proven itself yet again! Recently, the company published a blog about integrating Transformer Reinforcement Learning (TRL) with Parameter-Efficient Fine-Tuning (PEFT) for making a large language Model (LLM) with around 20 billion parameters fine-tunable on a 24GB consumer grade GPU. The best about this model is that it combines LLMs with RLHF, which is the powerful approach of ChatGPT as well. Similarly, Colossal-AI had also taken up the herculean task and found a way to build one’s own ‘personal ChatGPT’ on less computing resources, as small as a GPU of only 1.62 GB of memory. Their implementation was based on PyTorch that covered all three stages—pre training, reward model training, and reinforcement learning. The speed was even faster by 10.3 times with 8 billion parameters on a single GPU, which is much larger while being cheaper than a $14,999 A100 80GB GPU. These achievements are significant because they solve one of the most significant barriers to entry in the generative AI space: Computational resources. There is no doubt by now that building an LLM or chatbot with billions of parameters is a high-computation task. As a result, only the biggest tech companies in the world have been able to do so until now and even they fail in replicating what OpenAI did. But with these models, small- and medium-sized businesses can build their own language models without spending a fortune on computational resources. Now the ball is in everyone else’s court. Anyone can try it out on their own device. Who knows? Someone might actually beat OpenAI at their own game! ChatGPT Who? The era of being dependent on OpenAI, or any other big-tech, is coming to an end and the reign of DIY chatbots is set to begin. Until now, companies looking to implement AI chatbots have relied on third-party providers which are now offering pre-trained models like ChatGPT through APIs. However, with Hugging Face’s latest development, this could all change. Companies, or even developers, will now have the ability to build their own chatbots with similar capabilities, including their own weights and biases, if required. This means, for starters, the company can tailor the chatbot to its specific needs rather than relying on a generic pre-trained model. A custom chatbot can understand the nuances of the company’s industry and can be trained to respond to specific queries and concerns that are relevant to the company. This level of customisation is impossible with third-party chatbots. Another benefit of building a chatbot in-house is that the company can retain full control over the data that the chatbot collects. Third-party chatbot providers are notorious for collecting data from conversations and using it for their own purposes. This data can be highly sensitive and companies may not want to risk it falling into the wrong hands. What This Means for Big-Tech Microsoft recently announced their OpenAI Service offering ChatGPT on Azure. This is meant to be a groundbreaking announcement since companies were increasingly eager to use this technology for their own use cases. Along similar lines, Salesforce, Forethought, and Thoughtspot also released their betas offering the same services though, for obvious reasons, none came close to Microsoft’s offering. Colossal-AI’s offering was more developer-focused. On the other hand, Hugging Face’s step shows more promise. Hugging Face recognises two limitations in their approach. One is that the training speed is still slower and the other is the challenge of how users are to expand it to multiple GPUs for data parallelism. But this is just the first step, hopefully.","excerpt":"The reign of DIY chatbots is set to begin, thanks to Hugging Face.","categories":["AI Highlights"],"tags":["AI Chatbot","Big Tech","ChatGPT","Hugging Face","Microsoft","Microsoft Azure","Salesforce","thoughtspot"],"author_name":"Mohit Pandey","publish_date":"2023-03-15T11:00:00","publication_year":"2023","word_count":668,"keywords":["ChatGPT","Hugging Face","thoughtspot","OpenAI","AI","AI Chatbot","PyTorch","RLHF","chatbots","R","Big Tech","Salesforce","Microsoft Azure","generative AI","Azure","Microsoft"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","PyTorch","Hugging Face","RLHF","chatbots","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/much-awaited-breakthrough-in-llms-has-finally-arrived\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45385,"title":"How Indian Government’s AI-centric Schemes Are Boosting The Startup Ecosystem","content":"The Indian Government has been playing a pivotal role in boosting AI development in India through a number of schemes. According to sources, India ranks third in the world in terms of high-quality research publications related to AI. Indian government think tank NITI Aayog has been working to apply the potential of AI in public as well as private sectors. The Government has initiated a number of startup schemes related to emerging technologies to boost the growth of startups, promote entrepreneurship and create high-skilled jobs. Startups are the growth engine of our country and reports indicate there are presently close to 700 AI startups in India. In this article, we list down 9 startup schemes introduced by Indian Government to boost the growth of AI development, entrepreneurship and skill development. 1.  NewGen Innovation and Entrepreneurship Development Centre (NewGen IEDC) NewGen Innovation and Entrepreneurship Development Centre (NewGen IEDC) seeks to promote knowledge-based and technology-driven start-ups by harnessing young minds and their innovation potential in an academic environment. This scheme will be applicable in domains such as AI, computer vision, IoT, FinTech, nanotechnology and much more. The time period for this scheme is not disclosed. 2.  Assistance to Professional Bodies & Seminars\/Symposia Assistance to Professional Bodies & Seminars\/Symposia provides support in organising events for the scientific community which will help in enhancing the latest developments in their specific areas such as AI, AR\/VR, computer vision, IoT, green energy, non-renewable energy, and many others. The time period for this scheme is not disclosed. 3.  Atal Incubation Centres (AIC) Atal Incubation Centres (AIC) are headed by Atal Innovation Mission and was launched to nurture startups in India so that they can become scalable and sustainable business enterprise. This scheme is applicable to domains like AR\/VR, IoT, AI, computer vision, nanotechnology and much more. The time period for this scheme is not disclosed. 4.  Scheme to Support IPR Awareness Seminars\/Workshops in E&IT Sector The Scheme to Support IPR Awareness Seminars\/Workshops in E&IT Sector aims to support IPR awareness workshops\/seminars for sensitising and disseminating awareness about Intellectual Property Rights among various stakeholders, specially in E&IT sector. The industries that are applicable for this scheme are IT services, analytics, enterprise software, technology hardware, Internet of Things and AI. The scheme is valid up to November 2019. 5.  Multiplier Grants Scheme (MGS) Launched in 2013, Multiplier Grants Scheme (MGS) aims to encourage collaborative R&D between industry and academics\/ R&D institutions for the development of products and packages. It is applicable to a number of domains such as enterprise software, AI, IoT, analytics, and other such. The scheme is extended up to March 31, 2020. The time period for this scheme is not disclosed. 6.  Support for International Patent Protection in Electronics & Information Technology (SIP-EIT) The Ministry of Electronics and Information Technology launched this scheme to encourage innovation and recognize the value and capabilities of global IP along with capturing growth opportunities in ICTE sector. This scheme will be applicable to various IT services including the Internet of Things (IoT), artificial intelligence, etc. and will be valid up to November 2019. 7.  Technology Development Programme (TDP) Technology Development Programme (TDP) scheme aims to promote technology development in fields such as AI, IoT, nanotechnology, AR\/VR, computer vision, fintech and many more. This scheme supports activities aimed at developing and integrating technologies to evolve materials\/processes\/techniques both in the advanced\/emerging areas and in traditional sectors\/areas. The time period for this scheme is not disclosed. 8.  National Science & Technology Management Information System (NSTMIS) The NSTMIS scheme aims to provide the task of building the information base on a continuous basis on resources devoted to scientific and technological activities for policy planning in the country. The scheme is applicable to various domains such as automotive, AI, AR\/VR, IoT, nanotechnology and computer vision. The time period for this scheme is not disclosed. 9.  High Risk-High Reward Research High Risk-High Reward Research scheme supports exceptionally creative scientists pursuing highly innovative research with the potential for a broad impact in biomedical, behavioural, or social sciences within the NIH mission. This scheme is applicable in several domains such as AI, IoT, computer vision, automotive, FinTech. The time-period for funding of the scheme varies. Outlook AI is set to have a significant transformational impact in the industry as well as academia. The Indian Government is playing a key role in sustaining the intervention in order to tackle the challenges that AI-specific startups face in order to grow in their verticals.  During the 2018-2019 budget, the Government fo India has mandated NITI Aayog to establish the National Program on AI. According to the National Strategy for Artificial Intelligence, the underlying thrust is to identify the specific applications of emerging technologies with maximum social impact.","excerpt":"The Indian Government has been playing a pivotal role in boosting AI development in India through a number of schemes. According to sources, India ranks third in the world in terms of high-quality research publications related to AI. Indian government think tank NITI Aayog has been working to apply the potential of AI in public […]","categories":["AI Startups"],"tags":["Indian government","nanotech","nanotechnology","NITI Aayog","niti aayog ai","Startups"],"author_name":"Ambika Choudhury","publish_date":"2019-09-03T18:30:23","publication_year":"2019","word_count":786,"keywords":["Go","artificial intelligence","Rust","AI","Scala","computer vision","nanotech","RAG","Aim","niti aayog ai","Indian government","Startups","analytics","nanotechnology","R","NITI Aayog"],"extracted_tech_keywords":["AI","artificial intelligence","computer vision","analytics","Aim","RAG","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-indian-governments-ai-centric-schemes-are-boosting-the-startup-ecosystem\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10141066,"title":"No Noise, Only Sound – How AI is Changing Our Hearing Aids","content":"According to a World Health Organisation (WHO) estimation, a whopping 63 million Indians suffer from significant auditory impairment. However, less than 0.5% of those affected have sought treatment, and only 10% use hearing aids. Financial barriers and lack of awareness likely contribute to this low adoption rate. Can AI, the wizard behind solving some of the world’s toughest problems, make hearing aids more accessible and adaptable? In an exclusive conversation with AIM, Raja S, audiologist and director of Hearzap, shared how AI is revolutionising the hearing aid experience. “With AI integration, hearing aids can now adjust to real-life environments much more effectively. In the past, hearing aids were programmed in a controlled clinical setting, where the noise was minimised,” he explained. Hearzap is a digital platform catering to hearing needs and operates across Andhra Pradesh, Karnataka, Kerala, Maharashtra, and Tamil Nadu. Raja highlighted the limitations of traditional hearing aids, which work well in controlled environments but fall short in the real world, where users face a cacophony of sounds. “In everyday settings, distinguishing between speech and background noise can be a challenge,” he noted. “Our natural hearing has filters that help us focus on speech while ignoring irrelevant sounds. Now, AI allows hearing aids to do the same by separating background noise from speech and improving clarity even in complex acoustic environments.” This innovation is particularly beneficial for the elderly and those who frequently move between different environments. In the past, adjusting volume levels on the go was difficult. Thanks to AI, hearing aids can now automatically scan the surroundings and fine-tune settings in real time, ensuring optimal hearing without the need for manual intervention. Further, for those who have mobility issues or hand tremors, this is a game-changer. There will be no struggle with tiny devices; they can simply wear their hearing aids and go about their day with minimal hassle. Is India Open to Using AI Hearing Aids? According to Hearzap, approximately 20-25% of people are choosing premium hearing aids, most of which come with advanced AI features. But what’s stopping the rest? “The cost of a pair of high-quality hearing aids with AI starts around INR 1,00,000, while traditional hearing aids can range from INR 12,000 and can go up to INR 4 lakh per device,” said Raja. This wide price disparity reflects the varying levels of technology and features available. Advanced hearing aids, powered by AI, offer a more natural listening experience, closely mimicking the sound quality of normal hearing. For those who can afford it, the benefits include improved clarity, better communication, and an overall enhanced quality of life. Raja pointed out that premium models are typically purchased by those who can afford them, and the feedback has been overwhelmingly positive. Connectivity is another important feature. Many older adults enjoy watching YouTube or television, often in solitude, while younger family members are busy. AI hearing aids with connectivity features allow them to engage in these activities more comfortably, enriching their retired life with seamless access to sound. “I always tell my team that the best hearing aid is the one that makes the customer forget they’re even wearing it. It should feel like a natural extension of the body—effortless, seamless, and empowering the user to simply live their life without constantly adjusting or thinking about it. That’s the mark of quality fitting and expert audiology,” Raja said. The Market Across India, there are approximately 2,500 to 3,000 distribution points for hearing aids, including both organised chains and independent clinics. “Given that AI-enabled hearing aids come at a higher price, adoption is the strongest in areas where affordability is higher. I would estimate that 70-80% of these clinics are already offering AI-enhanced technology,” Raja said. While all clinics technically have access to these advanced products through manufacturers, adoption depends on demand. All major manufacturers in the hearing care industry have embraced AI technology in their products. “For instance, Philips, whose hearing instruments we dispense, is fully AI-enabled,” he noted. Philips has been among the pioneers in adopting AI, leveraging its extensive experience in medical technology. Its integrated AI systems facilitate seamless communication across various medical equipment. Apart from Philips, there are several other notable manufacturers in the industry, such as Rexton, Widex, and Starkey. Raja said that Philips stands out as a widely recognised brand, making it easier for people to relate to and trust their products. What’s Next? The positive aspect of hearing care is that technology is constantly evolving, bringing new features and improvements. However, the foundation of effective hearing care lies in proper counselling for the consumer—the person who will be using the device. Beyond apps and AI-enabled hearing aids, the role of an audiologist remains pivotal. But what if technology could evolve further? Imagine an AI-powered audiologist, an intelligent assistant that could analyse your condition, explain it in detail, and even guide you to the right solutions. As AI reshapes the industry, this futuristic blend of human expertise and digital precision could become the next breakthrough, redefining hearing care as we know it.","excerpt":"According to Hearzap, approximately 20-25% of users choose premium hearing aids with advanced AI features.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","AI hearing aids","Editors Picks","India"],"author_name":"Vidyashree Srinivas","publish_date":"2024-11-18T09:00:00","publication_year":"2024","word_count":841,"keywords":["Go","AI","ML","Git","RAG","GAN","Editors Picks","Aim","ViT","Rust","AI hearing aids","R","India","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Rust","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/no-noise-only-sound-how-ai-is-changing-our-hearing-aids\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":48022,"title":"Confirm-Shaming, Privacy Zuckering & Sneak Adding: E-Tailers Are Using These Dark Patterns To Make You Buy Junk","content":"Have you ever been notified about some product running out of stock? The chances are high that you have also followed that notification by you firing up the website, adding the item to the cart and rushing to pay before realising you never needed that wall sticker of senior citizens smiling happily in a golf cart. Yes, that exists! There is an epidemic of intrusive advertisements everywhere which are designed to pursue customers in every possible way. These are called Dark Patterns. Dark Patterns are tricks used in websites and apps that make you do things that you didn’t mean to, like buying or signing up for something. Harry Brignull who has done a spectacular job of creating awareness on his website. The purpose of this site is to spread awareness and to shame companies that use them. He categorises dark patterns as follows: Asking Trick Questions while filling in a form that makes one respond to a question that tricks them giving an answer that was never intended. Sneak an additional item into the cart. “Privacy Zuckering” — is a term used when a user is tricked into publicly sharing more information about yourself than you intended to. Cheekily named after Facebook CEO Mark Zuckerberg. Other techniques like Price Comparison Prevention, Misdirection and confirm-shaming. Let us take the example of confirm-shaming. There are a lot of incidents that an average ecommerce customer would relate to in this case. Confirmshaming is the act of guilting the user into opting into something. The option to decline is worded in such a way as to shame the user into compliance. Below is an illustration of how Amazon uses confirm-shaming to guilt users and discourage them from opting out. They place messages with great subtlety that indicate there is something in it for you which is where the customers forget if they needed this product in the first place! via Tumblr And, here is an example of confirm-shaming a customer into going on a tour. Not that beaches are bad for you but nudging people into thinking opting out(look at the bottom of the image) means hating good times is borderline preposterous. via Tumblr These results can also be the makings of third-party entities — like companies other than the shopping websites themselves — who are responsible for creating and presenting dark patterns on behalf of the shopping websites. Is Machine Learning Playing The Devil’s Advocate? Modern day businesses use recommendation systems predominantly. These systems try to capture the insights from granular level. Few metrics for deriving insights can be as follows: Content-based: What a certain customer is clicking(genres, actors, brands etc) Similarity: What customers who are similar in tastes are clicking Trending: Can depend on similarities, geographical location etc. For instance, if a customer hovers over a product for long or has added to the cart, this too, will be used in building a personalised system. The ability of the modern day systems to draw insights in real time makes the web more prone to dark patterns infiltration. Customers are constantly under a deluge of flash cards, which are chosen based on the way the customers have interacted or have clicked in the past. Using machine learning to boost up dark patterns is almost obvious in the current scenario. However, the same algorithms can be made to detect these dark patterns. Since dark patterns is well documented phenomenon, the recurring unwanted ads can be used to train neural networks that either trigger the system. In a study done by Princeton, the researchers found dark patterns on over 1,200 sites, with about 200 sites being straight up deceptive. To investigate further, the team introduced a bot that automatically visits thousands of e-commerce sites, finds product pages, completes the shopping flow, and saves all the data, allowing us to scan for dark patterns in a mostly automated way using machine learning. The Need For Stricter Regulations “Not all dark patterns are illegal, but they are nonetheless problematic because they are intended to prey on our cognitive limitations and weaknesses. However, since their use cases usually end up with user exploitation, the governments are taken this case seriously,” says Arvind Narayan who is one of the contributors to Princeton study. Among the reactions to our study, I didn't expect this. Using dark patterns is not just unethical, it's also shortsighted. In addition to pissing off your users and regulators, you may be pissing off your own employees as well. pic.twitter.com\/yctrjsTsg1 — Arvind Narayanan (@random_walker) June 26, 2019 The one major disadvantage of such greedy consumer exploitation tactics is the emergence of unhealthy and improper habits. Imagine a teenager watching recipe of some popular food item, and because some algorithm uses the salt and sugar content as a metric, will match the watched recipe and recommends items that are inherently unhealthy. So, who governs the kind of metrics used in every model? The law makers should be technically equipped to formulate laws because given a chance, few companies would be more than glad to push their own agenda into the law books. Countries like the US, have already introduced legislations like the Deceptive Experiences To Online Users Reduction (DETOUR) Act, to prohibit large online platforms from using deceptive user interfaces. They have recognised this act by using the term “dark patterns” in their press release. There is also a good chance that the employees deploying the models can be oblivious to the approaches chosen by an organisation. Solutions like crawling and clustering as suggested by Princeton researchers could be a starting point for further research into the depredations of unethical business practices.","excerpt":"Have you ever been notified about some product running out of stock? The chances are high that you have also followed that notification by you firing up the website, adding the item to the cart and rushing to pay before realising you never needed that wall sticker of senior citizens smiling happily in a golf […]","categories":["AI Trends"],"tags":[],"author_name":"Ram Sagar","publish_date":"2019-10-16T11:00:37","publication_year":"2019","word_count":936,"keywords":["Go","machine learning","AWS","AI","neural network","recommendation systems","RAG","Ray","GAN","R"],"extracted_tech_keywords":["AI","machine learning","neural network","Ray","RAG","recommendation systems","AWS","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/confirm-shaming-privacy-zuckering-sneak-adding-e-tailers-dark-patterns-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10109794,"title":"Top Generative AI Search Alternatives for Google","content":"Google has been at the forefront of search engines from an early stage. The past year has been a good one for Google’s advancements in generative AI with several groundbreaking models coming into existence. From Gemini to Imagen 2, the company has witnessed a drastic drift in AI. However, despite the supremacy, the search giant has competitors providing alternatives that boast the same use case with enhanced safety and privacy. Privacy apprehensions regarding Google’s extensive data collection practices have prompted users to seek alternatives that prioritise data protection. By diversifying digital experiences, users mitigate concerns related to Google’s monopoly in various service sectors. The desire for greater customisation and control over digital environments, innovative features, ethical considerations, and platform independence further fuels the demand for alternatives. More importantly, cost considerations, technological advancements, global accessibility, and security concerns contribute to users exploring and adopting alternative services. Here are a some top generative AI search alternatives for Google. Bing Chat Bing Chat, Microsoft’s AI-powered search companion, represents a significant leap forward in the evolution of search engines. With its array of features, such as contextual summaries, insights and analysis, consideration of multiple perspectives, and seamless integration with Bing Search, Bing Chat offers users a comprehensive and efficient search experience. By analysing search queries, assessing relevance, extracting key information, and generating clear summaries alongside additional insights, it not only saves time, but also enhances comprehension and aids in making more informed decisions. The benefits range from time-saving and personalised experiences to improved research efficiency. Looking ahead, Bing GPT holds promise for further advancements, including conversational search capabilities, personalised learning assistance, and support for researchers in uncovering hidden patterns and generating new ideas. DuckDuckGo DuckDuckGo stands out as a privacy champion among search engines, utilising generative algorithms to deliver search results while upholding user privacy. Its distinctive feature lies in its commitment to not track user data, offering a compelling alternative for individuals prioritising online privacy. Unlike many other search engines, DuckDuckGo goes the extra mile with built-in features such as tracker blocking and private browsing, reinforcing its dedication to shielding users from intrusive elements on the internet. By incorporating generative AI into its operations, DuckDuckGo ensures a user-friendly search experience and establishes itself as a trustworthy option for those seeking a more private and secure online journey. Download it on Play Store: DuckDuckGo Brave Brave emerges as a standout choice in the digital landscape, embodying speed, privacy, and an ad-free experience. A notable feature of Brave lies in its incorporation of generative AI, actively blocking intrusive ads and trackers to bolster user privacy and expedite browsing by eliminating superfluous content. What sets Brave apart is its innovative take on online advertising. Rather than subjecting users to invasive ads, the browser introduces the Basic Attention Token (BAT), a unique rewards system that incentivises users to engage with privacy-respecting ads voluntarily. This pioneering approach empowers users to control their ad experience and introduces a novel way for advertisers to connect with an audience genuinely interested in their content. With cutting-edge technology and user-centric features, Brave redefines the browsing experience for those seeking speed, privacy, and a fair approach to online advertising. Download it on Play Store: Brave Browser You.com You.com emerges as a promising contender in the search engine landscape with a distinctive focus on reinstating objectivity and trust in online information. Leveraging AI, it pioneers fact-checking by identifying misinformation, revealing sources, and actively seeking diverse perspectives to foster a balanced understanding of complex issues. Beyond fact-checking, You.com excels in personalisation, employing explainable AI to transparently communicate its assessments and encouraging community-driven efforts to improve accuracy. Challenges such as adapting to evolving misinformation tactics, maintaining neutrality, and scaling up are acknowledged, emphasising the platform’s commitment to continuous improvement. Despite the hurdles, You.com’s commitment to making online information more reliable and trustworthy positions it as a transformative force in reshaping how users navigate the vast digital landscape, emphasising accuracy and transparency over sheer engagement metrics. Perplexity Perplexity emerges as a transformative force in search engines, redefining the user experience through generative AI. With its commitment to comprehensive synthesis, Perplexity delivers detailed summaries and direct answers, transcending traditional snippets to provide a profound understanding of search results. The emphasis on exploring different perspectives is a standout feature, diversifying sources and combating biased results, fostering a more balanced comprehension of topics. Including Copilot Assistant and a Pro Version further enhances the user experience, offering refined search queries, sentiment analysis, and advanced functionalities. While Perplexity acknowledges its AI-generated nature, users must exercise caution regarding accuracy and information overload, especially considering its early developmental stage.","excerpt":"Despite its supremacy, Google has competitors providing alternatives that boast the same use case with enhanced safety and privacy.","categories":["AI Trends"],"tags":["Top Trend"],"author_name":"Sandhra Jayan","publish_date":"2024-01-02T10:29:45","publication_year":"2024","word_count":765,"keywords":["Top Trend","Go","AI","sentiment analysis","ML","Git","RAG","Ray","generative AI","Rust","R"],"extracted_tech_keywords":["AI","ML","generative AI","Ray","RAG","sentiment analysis","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-generative-ai-alternatives-for-google\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10073524,"title":"‘Make Instagram Instagram Again. Stop trying to be TikTok’","content":"“Make Instagram Instagram Again. Stop trying to be TikTok, I just want to see cute photos of my friends. Sincerely, Everyone.” The post was shared as an Instagram story by Kylie Jenner, the second biggest celebrity on the platform on behalf of, well, everyone. On the day this post was updated, 367 million people found that the photo and video sharing social media platform was shifting from social media. Shortly after, Adam Mosseri, Head of Instagram admitted that—Instagram is promoting more and more recommendations—that are posts in your ‘Feed’ from accounts that you do not follow. “The idea is to help you discover new and interesting things on Instagram that you may not know even exists. We are going to continue to try and get better at recommendations because we think it’s one of the most effective and important ways to help creators reach more people. We want to do best for our rectors, especially small creators. We see recommendations as one of the best ways to reach a new audience and grow their following. We need to evolve, because the world is changing quickly and we have to change accordingly,” Mosseri explained. Meta: From Social to Recommendations In February 2022, Meta reported the loss of daily users for the first time in its 18-year history. On the company’s Q1 2022 earnings call, Mark Zukerberg blamed TikTok for Facebook’s apparent stagnation. Since then, Zuckerberg has been trying hard to change things at Meta. Now, he is looking up to TikTok and trying to follow in the footsteps of the rival. In the same month, Meta announced the launch of short-video product ‘Reels’ for all global Facebook users. According to the company, even with a very limited market then, Reels content made up more than 20% of the time that people spend on Instagram. Besides, videos made up 50% of the time that users spend on Facebook. After imitating TikTok on short videos front, Meta is now aping the rival’s recommendations method also. For Meta, social circle is not the right way to spread the content any longer. The fact that people can readily transmit content to their friends or friends of friends does not imply that the content will be engaging or useful to the consumer. Through recommendations, it wants users to take to the rabbit hole of content from where it becomes quite difficult to come out. The major mode of distribution for content in recommendation media is not networks of connected people. Instead, the primary mechanism for content distribution is through opaque, platform-defined algorithms that encourage maximal customer attention and engagement. The platform always defines the type of attention these recommendations seek, and it is frequently tailored particularly to the consumer who is consuming the content. For instance, if the platform finds that a user likes films, that user is most likely to see a lot of film-related content because it is believed to be what piques that user’s interest the most. Facebook is also following the recommendation approach. In July 2022, Zuckerberg wrote in a Facebook post that there will be a personalised feed on the Home tab, where the discovery engine will recommend the content it thinks you’ll care most about. However, the Feeds tab will give you a way to customise and control your experience further. Recommendations are widely believed to be the primary reason behind TikTok’s success. According to experts, about 90–95% of content that is seen on TikTok comes from recommendation engines. Other platforms like Netflix and YouTube put far less emphasis on friends and social graphs in the favor of carefully-curated, magical algorithmic experiences that match the perfect content for the right people at the exact right time. Tiktok poses threat to all tech giants Millions upon millions of videos are uploaded onto TikTok every day, the vast majority of which receive only modest views. The size of each video’s audience is decided predominantly by the system’s ever-changing and mysterious algorithms, and the key to gaming the system is understanding how these algorithms work. Meta is not the only company that is perturbed by TikTok’s success and definitely not the only company to have lost millions of followers and earnings to the meteoric rise of this new rival. TikTok’s rise has also rattled Google, so much so that it led Google to launch YouTube Shorts. In June 2022, YouTube announced YouTube Shorts is now being watched by over 1.5 billion logged-in users every month. In September 2021, TikTok announced that it has about 1 billion monthly users on the platform. Beyond videos, TikTok has also become a top favourite search tool among users. According to an internal Google study, nearly 40% of young users between the ages of 18 and 24 in the US preferred ‘TikTok’ or ‘Instagram’ over ‘Google Maps’ or ‘Google Search’ when looking for a restaurant. Now, more than the platform itself, users are hacking the platform to use it for different purposes. ByteDance, the parent company of TikTok, had been the earliest Chinese internet company to go “all in” on the then-nascent technology and commit to the daunting task of building a recommendation engine—challenging the status quo of human curation. This early bet paid off in spades. The foundations of TikTok’s success were laid many years before the app itself was built, and it was no coincidence that ByteDance was the company to make it.","excerpt":"After imitating TikTok on short videos front, Meta is now aping the rival’s recommendations method also.","categories":["AI Features"],"tags":[],"author_name":"Tausif Alam","publish_date":"2022-08-25T12:00:00","publication_year":"2022","word_count":897,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","RAG","R"],"extracted_tech_keywords":["AI","RAG","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/make-instagram-instagram-again-stop-trying-to-be-tiktok\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10005667,"title":"8 Open-Source Tools To Start Your NLP Journey","content":"Teaching machines to understand human context can be a daunting task. With the current evolving landscape, Natural Language Processing (NLP) has turned out to be an extraordinary breakthrough with its advancements in semantic and linguistic knowledge. NLP is vastly leveraged by businesses to build customised chatbots and voice assistants using its optical character and speed recognition techniques along with text simplification. To address the current requirements of NLP, there are many open-source NLP tools, which are free and flexible enough for developers to customise it according to their needs. Not only these tools will help businesses analyse the required information from the unstructured text but also help in dealing with text analysis problems like classification, word ambiguity, sentiment analysis etc. Here are eight NLP toolkits, in no particular order, that can help any enthusiast start their journey with Natural language Processing. Also Read: Deep Learning-Based Text Analysis Tools NLP Enthusiasts Can Use To Parse Text 1| Natural Language Toolkit (NLTK) About: Natural Language Toolkit aka NLTK is an open-source platform primarily used for Python programming which analyses human language. The platform has been trained on more than 50 corpora and lexical resources, including multilingual WordNet. Along with that, NLTK also includes many text processing libraries which can be used for text classification tokenisation, parsing, and semantic reasoning, to name a few. The platform is vastly used by students, linguists, educators as well as researchers to analyse text and make meaning out of it. USP: Tokenisation; Identifying named entities; Sentiment analysis. Click here to try it out. Also Read: Despite The Breakthroughs, Why NLP Has Underrepresented Languages 2| OpenNLP About: Apache OpenNLP library is also an open-source ML-toolkit that helps in processing natural language text. Along with supporting the most common NLP tasks, such as tokenisation, segmenting sentences and tagging part of speech part-of-speech, OpenNLP can also be leveraged to build more advanced text processing services. It also includes maximum entropy and perceptron based machine learning. USP: Extracting named entity; Chunking; Parsing; Detecting language; Coreference resolution. Click here to try it out. Also Read: Is Common Sense Common In NLP Models? 3| CoreNLP About: CoreNLP is an open-source platform developed by Stanford NLP Group as a comprehensive solution for natural language processing in Java. By supporting side languages, CoreNLP allows in deriving linguistic annotations in text. CoreNLP takes the raw text by humans and analyses its parts of speech, names, people, dates, times, numeric quantities, etc. to indicate the relevant noun phrases. USP: Works with six languages including Arabic, Chinese, English, French, German, and Spanish; Parsing; and Tokenisation. Click here to try it out. Also Read: How This NLP-Driven Literature Search Engine Helping In COVID 4| spaCy About: spaCy is an open-source library in Python and Python for Natural language Processing. Built on the latest research, spaCy is designed for deploying in real-world products. It comes with pre-trained statistical models and word vectors supporting more than 60 languages. spaCy is licensed under MIT and is commercial for anyone to use. USP: Linguistic annotations; Tokenisation; Named Entities; Word Vectors; Dependency Parsing, Lemmatisation. Click here to try it out. Also Read: How Domain-Specific Pre-Training Can Outstrip General Language Models 5| AllenNLP About: AllenNLP is again a free, open-source natural language processing platform built on PyTorch, can be used for the building ML model. AllenNLP encompasses reference implementations of high-quality models for both core natural language processing tasks like semantic role labelling and other NLP applications like textual entailment. USP: Answering questions; Semantic role labelling, Textual Entailment; Text to SQL Click here to try it out. Also Read: Top 8 Pre-Trained NLP Models Developers Must Know 6| Flair About: Flair is an open-source and a simple framework built by the Humboldt University of Berlin. Built on PyTorch, Flair is one of the renowned deep learning frameworks available. It comprises advanced word embeddings like GloVe, BERT, ElMo etc. and has been designed to support several languages and an easy to use API. USP: Named entity recognition; Part-of-speech tagging; Sense disambiguation; Classification. Click here to try it out. Also Read: How To Establish Reasoning In NLP Models 7| gensim About: gensim is an open-source Python library, which can be used for topic modelling, document indexing as well as retiring similarity with large corpora. gensim’s algorithms are memory independent with respect to the corpus size. It has also been designed to extend with other vector space algorithms. USP: Latent semantic analysis; Latent Dirichlet Allocation; Random projections; Hierarchical Dirichlet Process; word2vec deep learning Click here to try it out. Also Read: How Mercedes-Benz Is Using AI & NLP To Give Driving A Tech Makeover 8| Spark NLP About: Spark NLP is an open-source Natural Language Processing library which has been built on Apache Spark ML. Spark NLP is equipped with more than 200 pre-trained pipelines and models supporting more around 40 languages. Supporting transformers like BERT, XLNet, ELMO, Spark NLP provides accurate and straightforward annotations for NLP. USP: Tokenisation; Part-of-speech tagging; Named entity recognition; Spell Checking; Multi-class text classification, Multi-class sentiment analysis. Click here to try it out.","excerpt":"Teaching machines to understand human context can be a daunting task. With the current evolving landscape, Natural Language Processing (NLP) has turned out to be an extraordinary breakthrough with its advancements in semantic and linguistic knowledge. NLP is vastly leveraged by businesses to build customised chatbots and voice assistants using its optical character and speed […]","categories":["AI Trends"],"tags":["business analysis tools","Natural Language Processing","NLP"],"author_name":"Sejuti Das","publish_date":"2020-08-28T12:00:00","publication_year":"2020","word_count":840,"keywords":["machine learning","AI","PyTorch","Natural Language Processing","ML","business analysis tools","NLTK","Transformers","RAG","NLP","deep learning","spaCy"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","PyTorch","Transformers","spaCy","NLTK","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-open-source-tools-to-start-your-nlp-journey\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10167460,"title":"Amazon Rolls Out Nova Sonic and Nova Reel 1.1 for Generative Voice and Video AI","content":"Amazon has introduced two new additions to its generative AI portfolio—Amazon Nova Sonic, a foundation model for voice-based applications, and Amazon Nova Reel 1.1, an updated model for text-to-video generation. Nova Sonic integrates speech recognition, understanding and generation into one model, removing the need for separate components. Traditional voice systems involve complex pipelines—converting speech to text, processing through a large language model, and converting the response back to speech. According to Amazon, this approach “fails to preserve crucial acoustic context and nuances.” “Nova Sonic takes a new approach,” the company said. “It unifies the understanding and generation capabilities into a single model.” The result is a voice agent that not only understands user input but also responds with an appropriate tone, pace, and style. The model is available through Amazon Bedrock. It supports applications in customer service, travel, education, healthcare and entertainment. In one example shared by Amazon, a virtual travel assistant shifts its tone in response to a customer’s change in emotion—moving from enthusiastic to reassuring when concerns about cost are raised. Another use case includes an enterprise dashboard assistant that grounds answers in company data and maintains multi-turn dialogue without requiring users to reset context. Nova Sonic also generates transcripts of user speech. This feature allows developers to integrate external APIs and tools, enabling AI agents to perform tasks such as retrieving flight options or accessing internal dashboards. On the other hand, Nova Reel 1.1 enables multi-shot videos up to two minutes in length, with consistent visual style across 6-second segments. It improves on the previous version in terms of generation speed and coherence. Users can choose to provide a single prompt for the entire video or set individual prompts per shot for more control. The model supports use cases like marketing campaigns, product design showcases, and social media content creation. “Nova Reel enhances creative productivity,” Amazon said, “while helping to reduce the time and cost of video production using generative AI.” To get started with Amazon Nova Reel 1.1, users need to visit the Amazon Bedrock console and request access to the model. In the left-hand navigation panel, they should select “Model access” and then locate Amazon Nova Reel in the list of available models. Requesting access here provides permission to use both version 1.0 and 1.1 of the model. Once access is granted, users can begin using Amazon Nova Reel 1.1 through the Amazon Bedrock console, the AWS SDK, or the AWS Command Line Interface (CLI). The releases are part of Amazon’s broader Nova model family, introduced at re:Invent 2024, which also includes Nova Micro, Lite, and Pro that generate text from different modalities","excerpt":"Nova Reel 1.1 enables multi-shot videos up to two minutes in length, with consistent visual style across 6-second segments.","categories":["AI News"],"tags":["Amazon"],"author_name":"Siddharth Jindal","publish_date":"2025-04-08T20:18:17","publication_year":"2025","word_count":438,"keywords":["API","Amazon Nova","AWS","AI","Modal","cloud_platforms:AWS","Amazon","AI agents","ViT","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Amazon Nova","AWS","R","API","ViT","Modal","AI agents","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amazon-rolls-out-nova-sonic-and-nova-reel-1-1-for-generative-voice-and-video-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042199,"title":"Accelerating Innovation — Uniting Research & Open Source","content":"Open-source software (OSS) underpins all major cloud platforms and a large number of cloud-based services today. Core technologies that power cloud computing –Linux, Kubernetes, Cloud Foundry, Docker, to name a few –are developed in open communities. The Cloud Native Computing Foundation supports innovation in OSS for cloud computing and has sponsorship and participation from practically all major players in the cloud ecosystem. A recent survey [1] of 3,440 professional developers and managers showed the widespread influence of open source, both in cloud technology and as an essential in-demand and growing skill in the industry. For example, 69% of respondents thought that contributions to open-source projects result in better professional opportunities. These are all indicators of the widespread trend of OSS becoming a dominant place for innovation in cloud platforms, services, and applications. Moreover, this innovation happens in an open, community-driven paradigm, wherein these communities operate on the fundamental principle of open collaboration via open sharing and the development of source code. Academia & research labs as custodians of innovation Another set of communities for sustained, ground-breaking innovation for the past decades have been academic and corporate research laboratories (labs) and who are often seen collaborating. These labs have researchers and students more oriented towards advancing state-of-the-art technologies using rigorous scientific methods such as reproducible experimental validation and peer review. In addition, such labs openly share the outcomes and innovations of their pursuit of scientific and technological advances via research conferences, publications in journals and build on top of each other’s work. Traditionally these two sets of communities – OSS communities and research labs – have been distinct, with very different structures to collaborate. There have been areas where these communities have intersected, but that is only in relatively limited areas and from select research labs. However, we are also witnessing a change where these labs release libraries or utilities under open source licences to accelerate innovation and wider adoption. Unlocking the potential of community-driven innovation Community-driven research has been on the rise, and there have been some instances in recent years across various areas like health, climate, AI, etc. An example is the “papers with code” trend that has emerged in Artificial Intelligence (AI) research, where the research work gets published along with releasing the associate source code as open source. TensorFlow and PyTorch are two popular open-source projects, where the technology and models developed thereon have been published in research communities and shared via open source. This combination of publication and open source has become a broadly accepted and expected way of sharing work in the AI research communities. It has led to tremendous acceleration in research and innovation. Researchers can access not just a description of another researcher’s work but also the source code, which allows quick reproduction, benchmarking, and improvements. A similar unification of open source and traditional research collaboration can be leveraged in many other areas of Computer Science and Engineering. With the advent of cloud-based collaboration tools such as GitHub, Slack, CI\/CD (continuous integration \/ continuous deployment) tools, and container-based packaging for easy reproducibility of runtime environments, this trend will only accelerate. These tools enable researchers to easily reproduce the exact environment of any other researcher’s work. Historically, this has been a challenge since much effort is spent in reproducing published results. Now, with the ability to package a full environment – including the operating system, versions of software and their configuration – as Linux container images and transfer them easily, this becomes as easy as running a single command on the command line. There are already projects such as Rocker and Binder, making this even easier for data science researchers. The next era of cloud computing innovation As research labs across academia, industry, and government leverage these collaboration mechanisms, we can expect a tremendous acceleration of research on ground-breaking innovations similar to open communities. In all likelihood, the developer and citizen communities will work in close partnership with traditionally siloed research labs, driven by the organic alignment of common interests and skills. The next age of innovation in cloud-based computing will be driven – in fact, further accelerated – by coming together of different communities. An example of this trend in the Hybrid Cloud space is the Konveyor project that brings together a global community of researchers and developers to create new tools for migrating and modernizing legacy applications to Kubernetes. Developing such tools will require a combination of fundamental research in AI for Code, program analysis, and distributed computing, coupled with the agile and disciplined software development principles of open source. Such projects will create the opportunity for bringing the scientific method to the open-source community and the community-driven, participatory innovation model to the research labs. We are at an inflexion point in the evolution of the scientific research method – from an era when it was largely restricted to the confines of research labs to an era when research labs and open communities work hand in hand for accelerated discovery, invention, and translation to practice. References [1] Andy Oram, “The Value of Open Source in the Cloud Era”, O’Reilly Media, 2021.","excerpt":"We are witnessing a change where these labs release libraries or utilities under open source licences to accelerate innovation and wider adoption.","categories":["AI Features"],"tags":["Open Source","Open Source AI","open source project"],"author_name":"Amith Singhee","publish_date":"2021-06-22T12:48:37","publication_year":"2021","word_count":849,"keywords":["data science","artificial intelligence","Open Source","AI","PyTorch","cloud computing","distributed computing","docker","RAG","Open Source AI","open source project","TensorFlow","kubernetes"],"extracted_tech_keywords":["AI","artificial intelligence","data science","TensorFlow","PyTorch","RAG","cloud computing","kubernetes","docker","distributed computing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/accelerating-innovation-uniting-research-open-source\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10084962,"title":"AI for Good: When Technology Steps in to Tackle City’s Trash","content":"What are your pet peeves as a person living in IT City Bengaluru? Traffic and trash. We will see how tech can tackle the latter, in this article. According to an article in Deccan Herald, Bengaluru currently generates around 4,300 metric tonnes of waste both from commercial establishments and residential units. This was also reflected in the recent Swachh Survekshan ranking released by the Ministry of Urban Affairs. Bengaluru’s ranking slipped from 28 in 2021 to 43 in 2022. SwachBin – the smart bin developed as part of Call for Code Global Challenge – won the fourth runner-up prize in 2022. The winning team members – Suneetha Jonadula and Mohamed Fazil were recognised by Call for Code and were also invited to UNI’s Delegates Dining Room to showcase the innovation. “Solid Waste Management (SWM) is among the basic essential services provided by municipal authorities in the country to keep urban centers clean. However, municipal bodies are increasingly getting overwhelmed and end up depositing solid waste at dump yards within or outside the city haphazardly,” said Bharathi Athinarayanan, in an exclusive interaction with Analytics India Magazine. The team behind the smart bin The SwachBin team being recognised by BBMP at their annual meet on Dec 13, 2022 Athinarayanan and his team are the brains behind SwachBin, a smart AI-powered trash bin that solves the waste\/trash segregation menace at source in a seamless manner. The team comprises of Bharathi Athinarayanan, from Bengaluru, who is the product owner & AI\/ML architect; Suneetha Jonadula, from Hyderabad, the lead fullstack developer; Prashanth Parthasarathy, from Bengaluru, the principal application developer; and Mohammed Fazil, from Bengaluru, their AI\/ML development engineer. This smart trash bin identifies the trash and performs classification based on an AI algorithm. The all-in-one bin can collect and segregate five types of trash such as recyclable (wet and dry), and non-recyclable (hazardous, sanitary, construction, and debris). That’s not all, it accomplishes the said task at ultra-low cost and by enabling sustainability. “Through our solution, we are automating the segregation process, which is more efficient than the manual processes,” said Prashanth Parthasarathy, who is part of the team that developed the smartbin. The team also partnered with BBMP –  Bengaluru’s municipal governing body – that helped the team in identifying the types of trash that the citizens need to segregate and using their inputs the team created their own ‘real’ dataset to be used for the smartbin. What makes it even more efficient and accessible is its low cost of production. “It cost us just around Rs 3,000 to develop the prototype as the hardware is easily available. This model can easily be replicated by others. When produced at scale, the cost will come down to Rs 300 per bin, as the cost of the hardware components is very affordable,” said Mohammed Fazil. How does SwachBin work? SwachBin is equipped with Raspberry Pi and is connected to a camera that takes a picture of the trash. With the help of computer vision, the system predicts the trash and classifies the image accordingly. Based on the result, an electrical signal is sent to the servo motor to rotate which in turn opens the trash can lid either to the left or right, depending on the recyclable or non-recyclable classification. Once the waste is segregated within the bin, the level of trash is monitored in real time by an ultrasonic sensor. Upon reaching the ‘set’ threshold, the ultrasonic sensor notifies the municipal corporation so they can plan a pick up for the trash. “The IOT dashboard shows how full the bin is. Once filled up to 80% capacity, it sends a notification to BBMP for collection,” said Athinarayanan. “Due to the infusion of AI\/ML and advanced hardware, it is possible to detect the type of trash with high accuracy and also through the use of NLP (text-to-speech), we can educate the user on the type of waste they are disposing of. Upon research, we found no existing solution that would be a combination of high-tech and yet highly affordable in this segment,” he said. The steps involved 1. The waste material\/trash is presented in front of the camera that is connected to the SwachBin (powered by Raspberry Pi). 2. The captured image of the waste is sent to the Raspberry Pi. 3. The AI engine processes the image and identifies the class of the trash and the category it belongs to. SwachBin uses the Resnet algorithm that has a very deep network of up to 152 layers by learning the residual representation functions instead of learning the signal representation directly. 4. Depending on the classification, the AI engine sends the corresponding signal to the servo motor to open the respective lid of the bin. 5. Classification details are further stored in the SQL database. 6. Activity details get updated in the Docker container. 7. Depending on the class type of the trash, the LED indicator is turned ON and the speaker conveys the same information as a voice message to the user. 8. The ultrasonic sensor senses the trash level and sends the information to the flask app to be displayed in the dashboard. 9. All the information gets displayed in the IoT dashboard. “In the next six months, we plan to further develop the model by integrating real time regional language translations to help the common citizen,” said Athinarayanan.","excerpt":"The smart trash bin identifies the trash and performs classification based on an AI algorithm","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Aparna Iyer","publish_date":"2023-01-13T12:00:00","publication_year":"2023","word_count":893,"keywords":["Go","AI","ML","docker","computer vision","NLP","Ray","analytics","SQL","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","NLP","computer vision","analytics","Ray","docker","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-for-good-when-technology-steps-in-to-tackle-citys-trash\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10059455,"title":"We need to break the fixed image of women being good only in certain types of roles: Swetha G Basavaraj, Samsung Electronics America","content":"The World Economic Forum Global Gender Gap Report 2021 states that women take up only 32% of the total workforce in data and AI, 20% in engineering and 14% in cloud computing. Many times, we see young talented women being discouraged from taking up STEM-based studies. A great way to motivate young girls to enter this space is by reading stories about women who have broken the glass ceiling and taken up challenging roles like data scientists, AI practitioners, and machine learning engineers. Today, we look at the inspiring journey of Swetha G Basavaraj, Director of Product Management, Data and AI Samsung Electronics, America. Basavaraj comes with extensive experience in major companies such as Yahoo, Volvo Cars, DataVisor and IBM. She has also built her own company, SAPling Software Solutions. Basavaraj feels that critical thinking and problem solving were some of the prerequisites for her for any choice of career. She said, “I have spent many years as an engineer and became an entrepreneur trying to solve a business problem using technology solutions. At that time, I didn’t realise I was more or less acting as a product manager. After my stint as an entrepreneur, I went to do my Masters at Stanford GSB, where I spent time learning Design and Digital Advertising domains. After my Masters, I was looking for a role where I could solve large-user-base problems combining business and technology. Having enjoyed entrepreneurship in my past, product management seemed like a good next career step. I joined Yahoo as a product manager for their Ad-tech platform.” AI and analytics While working as an engineer, Basavaraj had some experience in business intelligence tools and analytics. When she joined Yahoo, the scale of data and the technology to utilise and monetise that data was fascinating. She made the deliberate career move soon after to work more closely with AI, and that led her to DataVisor, which has been recognised for multiple years as one of the top AI startups to look out for in recent years. She added, “At Samsung, I manage a global team of product managers who help identify and define business problems that can be solved using data and machine learning. My team also manages a global data platform within Samsung that can be utilised by analytics and machine learning teams to achieve those business goals.” Challenges In the long and fulfilling career that Basavaraj has had, she has encountered quite a few challenges. The most crucial of them were: Products involving data and AI are very technical, so it is very easy to get lost in the details. We have to constantly remind ourselves and the team about the business outcomes that have to be met. Since the outcome is performance-centric, engineers and data scientists on the ground overlook data privacy, security, and governance (Data and ML). So, educating them, creating a process and maintaining this on a regular basis is paramount to building responsible AI and sustainable data strategy. It is easy to build a proof of concept but building a scalable long-lasting solution takes planning, effort and budget. Many ML models after successful proof of concept never make it to production because of both gaps in architecture required for such a deployment as well as lack of collaboration between data, IT and ML teams. So, one needs to have DevOps and MLOps as part of the overall planning. Maintaining a feedback loop to keep improving the quality and adoption after hitting a certain size – not just in technology, where we evaluate the models and fine-tune them but also as a product manager and lead evaluating business metrics and redefining the problem statement. Most importantly, hiring for the right skills within the team is crucial for team excellence in all of the above. Master the fundamentals to solve the right problems The only constant in the technology field is change. Basavaraj added, “As we are collecting more and more data, we are increasingly relying on greater automation and faster experimentation. So, it becomes imperative to leverage advances in technology to keep up with the market demand. 80% of the data we see has been created in less than five years, and so, to make sense of this data, we need updated technology solutions.” As a data science professional, one should not only keep themselves updated with the latest technology solutions but also master the first principles\/fundamentals to solve the right problems. As data grows, having good data engineering skills also becomes increasingly important. Basavaraj stated, “My only suggestion to the folks considering AI as a profession is to assess yourself first on your interests before just following the trend. There is so much information on the internet and many channels like Analytics India Magazine, which provide you with a forum to both learn and explore Data science as a career that you should take advantage of.” Traits of a good data scientist Typically for any data science\/AI role, there are at least three different areas Swetha considers: Critical thinking and ability to learn\/adapt to new problem statements or challengesTechnology know-how – Working with large scale (volume, variety, velocity) data and new breakthrough algorithms\/solutions in AI to solve data. Nowadays, data scientists can also grow more self-sufficient with data engineering skills to unblock themselves and reduce dependency. So it is important to be familiar with data engineering tools and technology for faster iterations. Domain expertise – Someone who is either passionate or has experience in that domain. It is a combination of skills coupled with curiosity to learn new paradigms. Start small and go deep If a college student or someone who is a freshly passed out aspirant plans to pursue a career in data science and AI, Basavaraj feels the following points should be kept in mind. Build your business problem solving and technical skills by working on hands-on projects. You have potentially four tracks: ML engineer, Data Analyst, Data Scientist and Data Engineer, and each requires different types of skills, so know your strengths and choose the right function. Technology and algorithms are ever-changing, so don’t stress over the vastness of the field. Start small and go deep. Fall in love with the problems and find technologies to solve those problems. Break this fixed image of women being good only in certain types of roles The diversity crisis in data science\/AI is real. Basavaraj feels, “More broadly, in general, the technology industry is still dominated by men with less than ~30% of the workforce being women. So, the scarcity of women in AI, which requires deeper technology skills, is not surprising.” Basavaraj says that there is a need for more women to choose STEM and continue to pursue career growth in technology. We should break this oversimplified and fixed image of women being good only in certain types of roles. Business leaders need to consciously support them with transitions to changing technologies in AI and give them a platform to grow within an organisation. “Women are key to scaling up AI as a practice, and the onus is on both sides – demand and supply. We want women to step up, creating enough workforce for organisations to choose from and organisations to open up enough opportunities that can be fulfilled by women. If you look at the stats in India, only 33 per cent choose STEM. So, we need to catch up young and have female role models to promote young girls in STEM careers. Once in the tech ecosystem, we need advocates to encourage and support women to train and build that pipeline of potential AI or data science candidates and establish equal employment opportunity”, Basavaraj adds. Data scientists becoming part of the core product teams We are in a fascinating stage for AI and analytics in the tech industry. Just a few years ago, AI was treated as a research project or was worked on in silos, but we are now seeing data scientists becoming part of the core product teams driving business outcomes in production at scale. Basavaraj concludes, “As we see ML\/AI model results becoming as good as or sometimes better than humans, we will see more automation, and this technology will get embedded in almost all walks of life. We will also see more specific areas of expertise within AI, such as cybersecurity, NLP, computer vision, etc., as this field advances. Model and Data Governance will become a necessary function within data and ML teams as we see organisations making responsible AI part of their core value.”","excerpt":"Basavraj feels that as a data science professional, one should not only keep themselves updated with the latest technology solutions but also master the first principles\/fundamentals to solve the right problems","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Interviews and Discussions","leadership","Machine Learning","Women in AI"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-01-31T12:00:00","publication_year":"2022","word_count":1420,"keywords":["Women in AI","data science","machine learning","AI","AI (Artificial Intelligence)","cloud computing","ML","MLOps","Machine Learning","computer vision","RAG","NLP","analytics","Data Science","Data Scientist","leadership","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","computer vision","data science","analytics","MLOps","RAG","cloud computing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/we-need-to-break-the-fixed-image-of-women-being-good-only-in-certain-types-of-roles-swetha-g-basavaraj-samsung-electronics-america\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10086733,"title":"ChatGPT Unleashes a Flood of Future Careers You Never Dreamed Of","content":"In a first-of-its kind move that looks not-so-traditional, Micha Kaufman, the founder and chief recently wrote an open letter to AI, published on the front page of ‘The New York Times’, calling for truce between and AI and humans alongside the relevance of humans in unleashing its true potential and introduction of categories of AI services, including ChatGPT application developers, Midjourney artists, AI chatbot developers, and more. The outcome—Fiverr registered a 1400% increase in demand for AI-related services in the last six months alone. Another website, Futurepedia currently offers an AI tool directory, featuring 864 AI tools across 49 categories spanning healthcare, life assistance, developer tools, fun tools, sales, copywriting and marketing amongst a host of others, and counting. This is also reflected in the growing interest in the use of generative AI technologies by businesses, including the likes of Midjourney, DALL.E, ChatGPT, and others. Organisations are already seeking individuals with AI skills to help them effectively use generative AI technologies to improve their marketing, sales and business operations. For instance, the opening for a prompt engineering internship using ChatGPT posted by NSM Automation. As per a status report on AI by Tenzai, a software development firm, the demand for AI skills like machine learning, deep learning, NLP, robotics, and image recognition witnessed a meteoric increase globally—with the US alone expected to record a 30% rise. Infact, And Brill, CEO at Thinkin Labs, posted a note sharing that the organisation has added a note to all job openings that required applicants to mention whether their cover letter or resume were written by ChatGPT along with the prompts they used. He further said that the applicants were informed that this would actually have a positive impact on their applications. So, as a developer if you want to stay relevant, it will be very crucial to upskill yourself with AI skills spanning expertise in programming languages, data engineering, models and exploratory data analysis, services and deploying amongst others. Ujjawal Chawda, a software engineer at Microsoft, recently shared that AI will not replace software developers anytime soon. However, building skills in domains including software design, system architecture database design and programming proficiency will help them stay relevant. The Rise of ChatGPT Jobs Meanwhile, tools like Open AI’s ChatGPT have unleashed a host of products and services built around the tool, including ChatGPT consulting services. Take the case of ‘Upwork’ that recently advertised the requirement of an AI and ChatGPT Web App developer who will be responsible for developing and implementing AI and ChatGPT features within a Web App. According to LinkedIn, there are already 186 jobs available if you are updated with ChatGPT skills. There is already a list of openings advertised on the Freelancer website for a host of requirements spanning integration of ChatGPT with a wordpress website, opening for a ChatGPT specialist, and several others. Retrospectively, there has always been a certain degree of resistance towards new technology or devices being invented. For instance, when the launch of Android and iPhone invented a new industry of mobile application development. Likewise, in the case of ChatGPT—while many fear that it will take away jobs like that of writers, customer service personnels, developers, translators, data entry personnels and similar other profiles—it has a huge potential for creating an industry of jobs powered by expertise in the tool itself. With the current raging debate around the roles or jobs surrounding ChatGPT, there is an even bigger potential for the multifaceted jobs it can create. The important thing is to position ourselves for new jobs that are being created as we speak for businesses. For instance, there are hundreds of freelancers offering their ChatGPT skills as services on Upwork. Among the most in-demand roles is that of a ChatGPT consultant for business. Consulting companies across the globe are offering the services of ChatGPT consultants for business in order to help them scale. ChatGPT Consultants OpenAI’s popular chatbot has also revolutionised the consulting industry with various platforms offering ChatGPT consultancy services, wherein they provide the services of ChatGPT experts who help businesses in optimising the potential of the chatbot for their business through various ways that may include content generation, social media management, customer services and more. Prompt Engineers: A prompt engineer is responsible for designing and crafting prompts for large language models. They can effectively use ChatGPT to generate conversational prompts and responses for chatbots and virtual assistants. ChatGPT App and tool developers: ChatGPT has also given rise to a new domain of ChatGPT application developers based on the popular chatbot. The chatbot has already inspired developers to innovate a wide range of tools and applications with unique features and capabilities that leverage the power of language generation to simplify everyday life. A list of such interesting applications is available here. AI Artists: The rising popularity of the model has given birth to a new domain of AI artists who generate amazing artworks using Dall-E , Stable Diffusion and Midjourney. A list of popular AI artists is available here. AI Mobile App developers: There is also a rising demand for integrating ChatGPT into mobile applications that has given rise to the need of AI mobile app developers.","excerpt":"The demand for AI skills witnessed a meteoric increase globally—with the US alone expected to record a 30% rise.","categories":["AI Features"],"tags":["ChatGPT","prompt engineering"],"author_name":"Aparna Iyer","publish_date":"2023-02-07T15:30:00","publication_year":"2023","word_count":862,"keywords":["ChatGPT","machine learning","OpenAI","AI","chatbots","RAG","NLP","prompt engineering","deep learning","generative AI"],"extracted_tech_keywords":["AI","machine learning","deep learning","NLP","generative AI","ChatGPT","OpenAI","RAG","prompt engineering","chatbots"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chatgpt-unleashes-a-flood-of-future-careers-you-never-dreamed-of\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":44868,"title":"Deep Tech Content Startup Instoried Raises $500,000 From Venture Catalysts","content":"Artificial intelligence-driven tech content startup Instoried this week secured $500,000 from Venture Catalysts in a seed funding round. The Bengaluru-based startup now wants to use these funds to expand their offerings into multiple languages and reach global audiences. Founded in 2018, Instoried is an augmented writing platform which improves customer interest and engagement. The platform helps content writers in e-commerce, news, FMCG and other verticals to optimise emotions in their marketing content. Company co-founder and chief executive officer Sharmin Ali told a news portal that Instoried would now be able to disrupt the vernacular content industry in India because of its in-house neuro-linguistic programming models. “This is the best time to build a product that not only analyses vernacular content but also gives smart recommendations to make it emotionally engaging,” Ali said. Apoorv Ranjan Sharma, President & Co-Founder at Venture Catalysts told a noted daily,“VCats always strives to identify and support startups with cutting-edge ideas across industries. With a total market size of $300 billion, the deep tech market for content, in particular, holds greater potential. We have faith in the ability of Instoried’s highly accurate tool to disrupt and lead the market with their unique approach of targeting and catering to specific customer emotions… A big advantage we see is their focus on the emotional aspect of vernacular languages, which will be a key differentiator, especially in a country like India that has the world’s most diverse linguistic landscape.”","excerpt":"Artificial intelligence-driven tech content startup Instoried this week secured $500,000 from Venture Catalysts in a seed funding round. The Bengaluru-based startup now wants to use these funds to expand their offerings into multiple languages and reach global audiences. Founded in 2018, Instoried is an augmented writing platform which improves customer interest and engagement. The platform helps content […]","categories":["AI News"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-08-22T16:00:40","publication_year":"2019","word_count":240,"keywords":["funding","artificial intelligence","programming_languages:R","AI","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","R","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deep-tech-content-startup-instoried-raises-500000-from-venture-catalysts\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075673,"title":"Beyond the Merge are Surge, Verge, Purge, and Splurge","content":"On September 15, the Merge finally happened, culminating in the transition of the Ethereum consensus mechanism from proof of work (PoW) to proof of stake (PoS). The transition is a milestone in getting on the road to building the final product that Ethereum founder Vitalik Buterin had dreamt of. Over the next few years, it plans to work on the “execution layer” which will eventually create Ethereum that will increase the transaction and reduce gas fees. “After the Merge, Ethereum will undergo surge, verge, purge, and splurge,” Vitalik Buterin, co-founder of Ethereum, said during a presentation in July. Buterin argued that Ethereum supporters believe Ethereum is only 40% complete. Even after the Merge, which is expected to be completed this September, Ethereum will still only be roughly 55% complete. In the presentation, he also hinted at what is coming next after The Merge. The platform plans to bring Shanghai Upgrade, which is expected to take place in early 2023. Ethereum Shanghai Upgrade Ethereum developer Tim Beiko in his blog has given a detailed review of the upcoming Shanghai Upgrade. Three major upgrades are in the offing – Beacon Chain withdrawals, EVM object format and Layer 2 fee reductions. According to the blog, activation of Beacon Chain withdrawals is a major feature for Shanghai. Currently, validators can’t withdraw staked ETH and staking rewards as it continues to be locked. If only there was a way of implementing proof of stake that doesn't require locking assets like this….. https:\/\/t.co\/FldEPOfIau— Charles Hoskinson (@IOHK_Charles) September 15, 2022 In the Shanghai upgrade, Ethereum will enable withdrawals. So, people who started staking will get their money back. “Validators can partially withdraw their accumulated rewards while maintaining the minimum requirement of 32 ETH stake to remain a validator and continue earning rewards,” mentions the blog. Another major update will happen in the form of EVM Object Format. Developers are facing a grave challenge to improve the EVM (Ethereum Virtual Machine) without hampering existing contracts. The possible solution is to provide new functionality to contracts which are deployed with a specific identifier without disturbing the existing contracts. This is now known as EVM object format. The main object of the EVM object format update is to separate code from data, which will benefit on-chain validators such as layer-2 scaling tool provider Optimism. The last thing is that the platform plans to reduce fees on layer 2 (L2). Since the full-sharding implementation isn’t ready yet, the platform plans to reduce transaction fees on L2 by lowering the cost of data on layer 1 (L1). This is called CALLDATA cost reduction. There’s another way to reduce transaction fees –  proto-danksharding, which might be announced in the upgrade. In this method, the capacity of data in each block will be increased, which could reduce the transaction cost on ETH L2 chains by 100x. The Shanghai Upgrade is a precursor to coming updates that will eventually build the final product. What the final product will look like Vitalik said that Ethereum currently can process about up to 20 transactions per second. However, it could process, including the rollups and sharding, according to the maths, about 100,000 transactions per second. Post proto-danksharding comes danksharding, which will introduce sharding on Ethereum. This is the final step to achieve Ethereum’s scalability and drastically reduce rollups cost. Rollups are second layers that sit on top of Ethereum and handle enormous quantities of transactions, bringing them to the main blockchain in batches. They also use a lot of data that increases transaction costs. Danksharding further lowers transaction costs by allowing nodes to take only a ‘sample’ of the data instead of downloading it entirely. Proposer builder separation (PBS) is introduced in the third development. This idea distinguishes between the labour-intensive process of block building and the more passive process of block validation. On the post-sharded Ethereum network, proposer builder separation will ensure that there are enough validators for decentralisation and data availability sampling. In the final development, history and storage requirements for validators will be further reduced. Currently, history and storage requirements make it expensive to run a validator. These developments will maintain the Ethereum decentralisation by decreasing the cost and barrier to entry to run a validator.","excerpt":"On September 15, the Merge finally happened, culminating in the transition of the Ethereum consensus mechanism from proof of work (PoW) to proof of stake (PoS). The transition is a milestone in getting on the road to building the final product that Ethereum founder Vitalik Buterin had dreamt of. Over the next few years, it […]","categories":["AI Features"],"tags":["Ethereum"],"author_name":"Tausif Alam","publish_date":"2022-09-20T17:00:00","publication_year":"2022","word_count":701,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Scala","RAG","ViT","programming_languages:Scala","R","Ethereum"],"extracted_tech_keywords":["AI","RAG","R","Go","Scala","ViT","programming_languages:R","programming_languages:Scala","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/beyond-the-merge-are-surge-verge-purge-and-splurge\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":13089,"title":"San Francisco startup nails Oscars 2017 predictions yet again","content":"Emma Stone picked up the Best Actress in Leading Role for ‘La La Land’ as the startup predicted Though San Francisco-headquartered Unanimous A.I, developer of the UNU swarming platform and the company behind the Swarm Insight on-demand service, may have had to bite dust when it came to some of their Oscar 2017 winners, they almost nailed the Best Picture ( La La Land fleetingly grabbed the highest honor, thanks to Warren Beatty and Faye Dunaway’s massive mix-up). The company that has perfected the art of stunning predictions, nailed the exact Super Bowl score accurately did not get a perfect 10 in the 89th Academy Awards. However, the startup were not too far off the mark either, placing all their hopes and bets on the Emma Stone starrer La La Land  that scooped six Academy Awards last night, including Best Actress, Best Original Score and Best Original Song for City of Stars. Here’s how Unanimous A.I stacked up in the 89th Academy Awards predictions: Let’s see what they got right, for starters they put all their force behind La La Land, which they thought would nab the most crystal ware.   In that prediction, they were right and it almost won the Best Picture as well! The startups has more hits than misses, reinforcing its excellent track record in predicting Oscars 2017 winners Best Actress: Emma Stone (La La Land) Best Director: Damien Chazelle (La La Land) Best Supporting Actress: Viola Davis (Fences) Best Actor in a Supporting Role: Mahershalla Ali (Moonlight) Best Foreign Language Film:  The Salesman Best Original Song: City of Stars for La La Land Best Cinematography: La La Land Best Adapted Screenplay: Barry Jenkins, Moonlight Best documentary feature: OJ: Made in America Best Visual Effects: The Jungle Book Best Animated Feature Film: Zootopia Best Production Design: La La Land Misses Casey Affleck scooped the Best Actor for Manchester by the Sea, disproving the startup’s nod for Denzel Washington for Fences Best Picture: Moonlight Best Actor: Casey Affleck (Manchester by the Sea) Best Costume Design: Fantastic Beasts Best Original Screenplay: Manchester by the Sea When it came to Best Actor, the AI startup leaned towards Denzel Washington (Fences), an industry favourite. However, the award was finally snapped up by a heavily bearded Casey Affleck, who played the role of a man overcome with grief. Unanimous A.I. has had an excellent track record in predicting US presidential debates winner and according to the startup’s blog, the company was challenged by Newsweek two years in a row (2015 and 2016) to predict the Oscars by forming a real-time human swarm of 50 average movie fans linked by AI algorithms. And for two years consecutively, the startup aced the predictions with their AI powered swarm (bunch of showbiz fans) out-predicting movie critics, besting leading newspapers and even famous statisticians Nate Silver-backed 538. The technology working in the backend is Artificial Swarm Intelligence (Swarm AI). It works by combining real-time insights from people, who are spread across the globe using AI algorithms that are modeled after actual swarms in nature. In Swarm AI, each participant works like a human processor in the Swarm system contributing their inputs and opinions that help in achieving the result.","excerpt":"Though San Francisco-headquartered Unanimous A.I, developer of the UNU swarming platform and the company behind the Swarm Insight on-demand service, may have had to bite dust when it came to some of their Oscar 2017 winners, they almost nailed the Best Picture ( La La Land fleetingly grabbed the highest honor, thanks to Warren Beatty […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-02-27T11:04:46","publication_year":"2017","word_count":533,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","R","startup"],"extracted_tech_keywords":["AI","RAG","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/san-francisco-startup-nails-oscars-2017-predictions-yet\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":20020,"title":"Top 7 Microprocessor Chips Specifically Designed To Accelerate Artificial Intelligence","content":"AI accelerators have been at the forefront in the race of AI. Within the whole acceleration of AI in last few years, 2 key reasons stand out. Firstly, large amount of data that we have stored digitally helps to train neural networks that was not possible a decade ago. Secondly, neural networks require large processing power to train and traditional processors do not have that level of abilities. Just a simple image recognition of cats through traditional CPU’s might take days or weeks to train. For AI to go beyond where it is now, the key element is that developers have freedom to train neural nets quickly, make mistakes, learn from those mistakes and polish their algorithms. For that, processing power have to increase. And that’s exactly what most tech giants are after, and not just traditional chip makers. Here we list down 5 chips (in alphabetic order)especially designed for AI that made headlines this year. AMD Radeon Instinct Radeon Instinct is AMD’s brand of deep learning oriented GPUs. It replaced AMD’s FirePro S brand in 2016. Compared to the Radeon brand of mainstream consumer\/gamer products, the Radeon Instinct branded products are intended to accelerate deep learning, artificial neural network, and high-performance computing\/GPGPU applications. Apple A11 Bionic Neural Engine The Apple A11 Bionic is a 64-bit ARM-based system on a chip (SoC), designed by Apple Inc. and manufactured by TSMC. It first appeared in the iPhone 8, iPhone 8 Plus, and iPhone X. The A11 includes dedicated neural network hardware that Apple calls a “Neural Engine”. This neural network hardware can perform up to 600 billion operations per second and is used for Face ID, Animoji and other machine learning tasks. The neural engine allows Apple to implement neural network and machine learning in a more energy-efficient manner than using either the main CPU or the GPU. Google Tensor Processing Unit A tensor processing unit (TPU) is an application-specific integrated circuit (ASIC) developed by Google specifically for machine learning. Compared to a graphics processing unit, it is designed for a high volume of low precision computation (e.g. as little as 8-bit precision) with higher IOPS per watt, and lacks hardware for rasterisation\/texture mapping. The chip has been specifically designed for Google’s TensorFlow framework. However, Google still uses CPUs and GPUs for other types of machine learning. Other AI accelerator designs are appearing from other vendors also and are aimed at embedded and robotics markets. Huawei Kirin 970 Kirin 970 is powered by an 8-core CPU and a new generation 12-core GPU. Built using a 10nm advanced process, the chipset packs 5.5 billion transistors into an area of only one cm². HUAWEI’s flagship Kirin 970 is HUAWEI’s first mobile AI computing platform featuring a dedicated Neural Processing Unit (NPU). Compared to a quad-core Cortex-A73 CPU cluster, the Kirin 970’s heterogeneous computing architecture delivers up to 25x the performance with 50x greater efficiency. IBM Power9 Recently launched by IBM, Power9 is a chip which has a new systems architecture that is optimized for accelerators used in machine learning. Intel makes Xeon CPUs and Nervana accelerators and NVIDIA makes Tesla accelerators. IBM’s Power9 is literally the Swiss Army knife of ML acceleration as it supports an astronomical amount of IO and bandwidth, 10X of anything that’s out there today Intel Nervana The Nervana ‘Neural Network Processor’ uses a parallel, clustered computing approach and is built pretty much like a normal GPU. It has 32 GB of HBM2 memory dedicated in 4 different 8 GB HBM2 stacks, all of which is connected to 12 processing clusters which contain further cores (the exact count is unknown at this point). Total memory access speeds combine to a whopping 8 terabits per second. Nvidia Tesla V100 NVIDIA® Tesla® V100 is the world’s most advanced data center GPU ever built to accelerate AI, HPC, and graphics. Powered by NVIDIA Volta™, the latest GPU architecture, Tesla V100 offers the performance of 100 CPUs in a single GPU—enabling data scientists, researchers, and engineers to tackle challenges that were once impossible.","excerpt":"AI accelerators have been at the forefront in the race of AI. Within the whole acceleration of AI in last few years, 2 key reasons stand out. Firstly, large amount of data that we have stored digitally helps to train neural networks that was not possible a decade ago. Secondly, neural networks require large processing […]","categories":["AI Trends"],"tags":[],"author_name":"Дарья","publish_date":"2017-12-21T07:02:43","publication_year":"2017","word_count":670,"keywords":["machine learning","TPU","AI","neural network","ML","image recognition","Aim","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","Aim","TensorFlow","image recognition","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-microprocessor-chips-specifically-designed-accelerate-artificial-intelligence\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166456,"title":"We are Now a Power-Limited Industry, says Jensen Huang","content":"AI has reached a critical juncture, becoming more intelligent and useful due to its reasoning ability. This advancement has led to a significant increase in computational requirements, with the industry needing much more computing power than previously anticipated. The generation of tokens for reasoning is a key factor in this increased demand, according to NVIDIA CEO Jensen Huang, who recently addressed the future of AI and computing infrastructure at the GTC 2025 summit in San Jose earlier this week. His keynote highlighted AI’s rapid evolution and the immense computational power required to support its growth. “Every single data centre in the future will be power-limited. We are now a power-limited industry,” he said. With AI models growing exponentially in complexity and scale, the race is on to build data centres, or what Huang calls “AI factories”, that are not only massively powerful but also energy-efficient. The Rise of the AI Factory Huang introduced the concept of AI factories as the new standard for data centre infrastructure. These centres, which are no longer simply repositories of computation or storage, have a singular focus—to generate the tokens that power AI. He described them as “factories because it has one job, and that is to generate these tokens that are then reconstituted into music, words, videos, research, chemicals, or proteins”. AI factories, according to Huang, are becoming the foundation for future industries. “In the past, we wrote the software, and we ran it on computers. In the future, the computer is going to generate the tokens for the software.” Huang predicts a shift from traditional computing to machine learning-based systems. This transition, combined with AI’s growing demand for infrastructure, is expected to drive “data centre buildouts to a trillion-dollar mark very soon”, he believes. Power Problem is Also a Revenue Problem As data centres expand, they will face significant power limitations. This underscores the need for more efficient technologies, including advanced cooling systems and chip designs, to manage energy consumption effectively. Huang noted that the computational requirements for modern AI, especially reasoning and agentic AI, are “easily a hundred times more than we thought we needed this time last year”. This explosion in demand places enormous strain on data centres’ energy consumption. His keynote made it clear that moving forward, energy efficiency isn’t just a sustainability concern; it will be directly tied to profitability. “Your revenues are power limited. You could figure out what your revenues will be based on the power you have to work with,” he said. This shift will influence everything from how AI models are trained and deployed to how entire industries operate. In this regard, power is the ultimate constraint in AI-dominated computation. This limitation is reshaping both the design and operation of data centres around the world. “The more you buy, the more you make,” Huang quipped, encouraging businesses to view their investments in NVIDIA’s accelerated computing platforms as the key to unlocking the full potential of AI-driven value creation. Scaling Up Before Scaling Out Huang explained NVIDIA’s approach to managing this power limitation, which would be a fundamental rethinking of scale. “Before you scale out, you have to scale up,” he stated. NVIDIA’s new Blackwell platform demonstrates this principle with its extreme scale-up architecture, featuring “the most extreme scale-up the world has ever done”. A single rack delivers an astonishing one-exaflop performance within a fully liquid-cooled, high-density design. By scaling up, data centres can dramatically reduce inefficiencies that occur when spreading workloads across less integrated systems. Huang explained that if data centres had scaled out instead of scaling up, the cost would have been way too much power and energy. He pointed out that, as a result, deep learning would have never happened. Blackwell, a Path to 25x Energy Efficiency With the launch of NVIDIA’s Blackwell architecture, Huang highlighted a leap in performance and efficiency. According to him, the goal is to deliver the most energy-efficient compute architecture you can possibly get. Huang believes NVIDIA has cracked the code for future-ready AI infrastructure by combining innovations in hardware, such as the Grace Blackwell system and NVLink 72 architecture, with softwares like NVIDIA Dynamo, which he described as “the operating system of an AI factory”. Explaining the broader significance, he said, “This is ultimate Moore’s Law. There’s only so much energy we can get into a data centre, so within ISO power, Blackwell is 25 times [better].” AI Factories at Gigawatt Scale NVIDIA’s ambitions don’t stop with Blackwell. Huang outlined a roadmap extending years into the future, with each generation bringing new leaps in scale and efficiency. Upcoming architectures like Vera Rubin and Rubin Ultra promise “900 times scale-up flops” and AI factories at “gigawatt” scales. As these AI factories become the standard for data centre design, they will rely heavily on advancements in silicon photonics, liquid cooling, and modular architectures. Huang likened the current AI revolution to the dawn of the industrial era, naming NVIDIA’s AI factory operating system Dynamo in homage to the first instrument that powered the last industrial revolution. “Dynamo was the first instrument that started the last industrial revolution—the industrial revolution of energy. Water comes in, electricity comes out. [It’s] pretty fantastic,” he said. “Now we’re building AI factories, and this is where it all begins.”","excerpt":"The NVIDIA CEO introduced the concept of ‘AI factories’ as the new standard for data centre infrastructure.","categories":["Deep Tech"],"tags":["AI Data Center","computing","energy","GTC Conference","NVIDIA"],"author_name":"Sanjana Gupta","publish_date":"2025-03-21T16:00:00","publication_year":"2025","word_count":874,"keywords":["AI Data Center","Go","API","energy","machine learning","agentic AI","programming_languages:R","AI","innovation","GTC Conference","RAG","deep learning","computing","NVIDIA","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","agentic AI","RAG","R","Go","API","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/we-are-now-a-power-limited-industry-says-jensen-huang\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10139939,"title":"India’s Digital Infrastructure is Changing Super Fast with AI","content":"AI is transforming India’s e-governance landscape and reaching the farthest of underserved communities. From processing numerous queries round the clock to enabling bilingual support, advancements in AI are making government initiatives more accessible and efficient than ever. “Our AI systems now process between 5 to 7 lakh queries every month, operating 24\/7, which is crucial in ensuring millions of citizens get timely assistance,” said Sharmishtha Dasgupta, deputy director general of the National Informatics Centre, during the Nvidia Summit held in Mumbai last week. She goes on to mention that these queries range from enrollment and eligibility checks to updates and troubleshooting. The system’s capacity to manage such high volumes highlights its scalability and efficiency, as well as its alignment with the Digital India campaign’s goal of making government services accessible to every citizen. Dasgupta also noted that AI-powered bilingual support systems have been functional in bridging linguistic and digital divides. “Bilingual AI-powered support systems are making interactions with government schemes like PM Kisan Samman Nidhi Yojana straightforward, reducing complexities and ensuring citizens can engage in their preferred language,” she added. For instance, IRCTC, an extension of Indian Railways, is also using the conversational AI platform AskDISHA 2.0 chatbot to help customers book railway tickets through voice, chat, and click-based operations. “In our constant pursuit to enhance the user experience, leveraging new age technologies, today we are taking a giant leap. Now passengers can book train tickets in a conversational manner, leveraging our AI Virtual Assistant, AskDISHA 2.0, powered by CoRover Conversational AI platform,” Rajini Hasija, CMD, IRCTC said. During the Nvidia Summit, Tanusree Barma, deputy director general of the Unique Identification Authority of India (UIDAI), highlighted the transformative AI developments within UIDAI, as the organisation moves towards adopting and using indigenous AI capabilities. Barma stated, “Within UIDAI, we have been silently having an AI revolution to enable indigenous implementation of LLMs as well as AI models to apply for biometric duplication and detection.” Barma also talked about the critical role that AI now plays in the nation’s digital identity framework, emphasising how these AI models help strengthen security, improve accuracy and drive innovation for the UIDAI. She further said that by focusing on homegrown AI solutions, UIDAI not only aims to reduce dependency on foreign AI technologies but also ensures data sovereignty and control, a major step in India’s broader AI ambitions. Innovative Case Studies Manohar Paluri, VP of AI at Meta, recently told AIM, “India is possibly among the top three in terms of Llama downloads and variants, and it’s also among the top two in terms of how many developers there are. The appetite for technology and the appetite of people adapting to new technology in this part of the world is amazing.”Paluri has talked about one of the case studies, Pratham, a nonprofit for education. “Pratham is an example of how this technology is being used to teach kids in an affordable way. So, you are scaling yourself where these chatbots actually can help you learn about a particular subject very quickly,” he added. He explained how farmers can now use this technology in their native language and get the best information on agriculture and financial assistance, which was previously not feasible or accessible. “So when you think about learning, it is going to be redefined now with technologies like Llama as well as MovieGen,” Paluri added. He also discussed the benefits of an open ecosystem, where people can take the model and fine-tune it as per their needs. “I was actually in Japan two weeks ago. There were Japanese versions of Llama, where they were able to tune it,” he said. Paluri also mentioned a Korean version and a Korean high school math version of Llama. “We are trying to bring high-quality Indian tokens into Llama, so that it works for Indian languages,” said Paluri. He also mentioned that Llama, being the engine for Meta-AI, will work for Indian languages. This will help billions of people in India, who are using Meta-AI on WhatsApp and other Meta products. Future Prospects Meanwhile,BharatGen, India’s first government backed multimodal AI initiative, has released e-vikrAI, an advanced solution powered by Vision Language Models, tailored for product images in Indic e-commerce.Spearheaded by IIT Bombay under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS) of the Department of Science and Technology (DST), BharatGen focuses on creating AI models tailored to India’s diverse languages and cultural contexts.","excerpt":"AI is transforming India’s digital infrastructure and reaching the farthest of underserved communities.","categories":["IT Services"],"tags":["AI Infrastructure","Digital India","infrastructure","NVIDIA"],"author_name":"Shalini Mondal","publish_date":"2024-10-31T13:06:38","publication_year":"2024","word_count":734,"keywords":["Go","infrastructure","Digital India","AI","chatbots","Scala","Git","RAG","Aim","multimodal AI","GAN","AI Infrastructure","NVIDIA","R"],"extracted_tech_keywords":["AI","multimodal AI","Aim","RAG","chatbots","R","Go","Scala","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/indias-digital-infrastructure-is-changing-super-fast-with-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10048055,"title":"What Is Graph Representation Learning","content":"Relational data represent relationships between entities anywhere on the web (e.g. online social networks) or in the physical world (e.g. structure of the protein). Such data can also be represented as a graph with nodes (such as user, protein) and branches connecting them. To apply the machine learning methods to graphs one needs to learn a representation of the graph and the fundamentals associated with it. In this article, we will discuss the graph representation fundamentals in detail along with how machine learning can be performed using graph data. The major points that we will cover in this article are listed below. Table of Contents What is Graph? Multi-Relational GraphHeterogeneous GraphMultiplex GraphMachine Learning with Graphs Feature extractionGraph for Various ML Methods Let’s proceed with our discussion. What is Graph? Let’s define graph data a bit more formally before we consider machine learning on graphs. There are nodes V and edges E that make up a graph G=(V, E). It is necessary to transform graphs into numerical forms before using them in any method. The use of adjacency matrices, laplacian, or degree matrices can help us attain this goal. If you want to define attributes of each vertex and link, you may use a feature vector. Types of Graphs and corresponding adjacency matrix An adjacency matrix is a useful way to represent a graph. We organize the nodes in the graph so that each node indexes a specific row and column in the adjacency matrix to depict a graph with an adjacency matrix. The existence of edges may therefore be represented as entries in this matrix. If the graph contains any undirected edges then the adjacency matrix will be a symmetric matrix, but however, if the graph is directed means edge orientation becomes important then it is not necessarily symmetric. Some graphs also have weighted edges where entries in the adjacency matrix are arbitrary real values rather than {0,1}. Multi-Relational Graph In addition to the distinction between undirected, directed, and weighted edges we will look at graphs with different types of edges. For example in the graph depicting drug-drug interaction, we would want various edges to indicate various side-effects which can occur when we take a pair of medications concurrently. In this case, we can extend the edge notation to include the edge relationship type as τ and can define one adjacency matrix as Aτ. Multi-Relational Graph Two important subsets of multi-relational graphs are heterogeneous graphs and multiplex graphs. Heterogeneous Graph In the heterogeneous graph, for example in biomedical, one type of node may represent proteins, another type of node could represent medication, and still another type could represent illness. Edges indicate ‘treatments’ and these would appear exclusively between medication and illness nodes. Similarly, edges denoting ‘polypharmacy side effects’ would appear only between two medication nodes. Multiplex Graph In a multiplex graph, we assume that the graph can be decomposed into a collection of k layers. Every node is considered to be a member of every layer, and each layer corresponds to a unique relation, which represents the intra-layer edge type of that layer. For example, in a multiplex transportation graph, each node might represent a city and each layer might represent a different mode of transportation (air travel, road travel, rail travel). Intra layer edges would thus represent cities linked by various means of transportation, while inter-layer edges symbolise the ability to alter modes of transients inside a city. Machine Learning with Graphs Machine learning with the graph is no longer an exception, although the traditional classification of supervised and unsupervised is not always the most informative or useful when it comes to the graph. Let us understand it with an example Assume we are given huge social network data including millions of individuals, but we know that a significant proportion of these users are bots. Identifying these bots is essential for a variety of reasons including the fact that the firm may not wish to promote bots. Manually evaluating each user to identify whether they are a bot or not would be prohibitively expensive therefore we would like to create a graph model that could categorize users as a bot or not bot. Feature Extraction The primary goal of feature extraction for graphs is to convey information about local and global graph structure in a more accessible, vector-like manner. We are as interested in obtaining information about the characteristics of single vertices as we are in obtaining information about the connection between them. Graph characteristics may be classified into three types: node level, graph level, and neighbourhood overlap features. We intend to produce a feature vector for each vertex for node-level features. Node degree, centrality technique, and clustering coefficients are some of the most common feature extraction methods. To compute these characteristics, we can utilize information from vertex’s immediate neighbours or from a more distant, k-hop neighbourhood. We potentially take a broader approach to develop a feature for the entire graph. It is referred to as a graph-level feature. It is possible to do so using fundamental features such as the Adjacency matrix as well as some sophisticated iterative techniques such as Weisfeiler-Lehman or Graphlet kernels. Neighbourhood overlap features are the final form of feature extraction for graphs. They are specially intended to retrieve information regarding node connections. They are further classified as local and global techniques. The former indicates the similarity of two nodes’ neighbourhoods, whereas the latter describes whether a specific node inside a network belongs to the same community. Graph for Various ML Methods The large category of common machine learning applications on graph data comprises classification, regression, and clustering. For example, given a graph depicting the structure of a molecule we could wish to create a regression model that predicts the toxicity or solubility of that molecule. ML Outcomes Or we might want to build a classifier to detect whether a computer program is malicious by analyzing the graph-based representation of its syntax and data flow. Instead of generating predictions over the various components of a single graph, such as nodes and edges, we are given a dataset of numerous separate graphs and our goal is to generate independent predictions relevant to each graph. The purpose of graph clustering, a related job, is to develop an unsupervised measure of graph similarity between pairings. Graph regression and classification are perhaps the most straightforward analogues of standard supervised learning of all machine learning tasks on graphs. Each graph is data points linked with labels and the objective is to learn a mapping from data points i.e., graph to labels using a labelled set of training points. Similarly, graph clustering is a simple extension of unsupervised clustering for graph data. The difficulty with these graph-level jobs is defining meaningful characteristics that take into consideration the relation structure inside each datapoint. Final Words In this article, we have taken a tour across necessarily to be known concepts such as what actually is a graph, its types along with how machine learning can be used to solve problems using graphs and how features can be extracted from the graphs. Reference Graph Representation Learning","excerpt":"Relational data represent relationships between entities anywhere on the web (e.g. online social networks) or in the physical world (e.g. structure of the protein).","categories":["AI Trends"],"tags":["Data Science","Deep Learning","graph neural networks","Machine Learning","machine learning document classification"],"author_name":"Vijaysinh Lendave","publish_date":"2021-09-11T10:00:00","publication_year":"2021","word_count":1183,"keywords":["Go","machine learning","machine learning document classification","AI","programming_languages:R","ML","Machine Learning","programming_languages:Go","GAN","graph neural networks","Deep Learning","Data Science","R"],"extracted_tech_keywords":["AI","machine learning","ML","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/what-is-graph-representation-learning\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10051309,"title":"NVIDIA, Microsoft Introduce New Language Model MT-NLG With 530 Billion Parameters, Leaves GPT-3 Behind","content":"Earlier this week, in partnership with Microsoft, NVIDIA introduced one of the largest transformer language models, the Megatron-Turing Natural Language Generation (MT-NLG) model with 530 billion parameters. The language model is powered by DeepSpeed and Megatron transformer models. Interestingly, MT-NLG has 3x the number of parameters compared to the existing largest models, including GPT-3 (175 billion parameters), Turing NLG (17 billion parameters), Megatron-LM (8 billion parameters), and others. In comparison, the Chinese govt-backed Beijing Academy of Artificial Intelligence’s (BAAI) Wu Dao 2.0 (1.75 trillion parameters) and Google’s Switch Transformer (1.6 trillion parameters) are some of the largest transformer language models in the space. Previously, Microsoft had also partnered with OpenAI, where it acquired exclusive rights to use its GPT-3 language models for commercial use cases. Evolution of Language Models In recent years, transformer-based language models in NLP have witnessed rapid progress, fueled by computation at scale, large datasets, and advanced algorithms to train these models. As a result, these models generalise effective zero- or few-shot learners with high accuracy on many tasks and datasets. Some notable downstream applications include code autocompletion, summarization, automatic dialogue generation, translation, semantic search, etc. Here is a timeline graph of some of the popular language models, which have grown at an exponential rate over the past (as shown below): (Source: NVIDIA) How is MT-NLG different from others? Training such models – GPT-3, Megatron-LM, Turning NLG, etc. – is challenging for two main reasons. Firstly, it is no longer possible to fit the parameters of these models in the memory of even the largest GPU. Secondly, the large number of computing operations required can result in unrealistically long training times if special attention is not paid to optimising the algorithms, software, and hardware stack. However, training MT-NLG was possible, thanks to NVIDIA’s and Microsoft’s combined effort, where they achieved unprecedented training efficiency by combining SOTA GPU-accelerated training infrastructure with a distributed learning software stack. As a result, the team built high-quality, realistic training corpora with hundreds of billions of tokens and co-developed training recipes to improve optimisation efficiency and stability. MT-NLG Explained SOTA supercomputing clusters such as the NVIDIA Selene and Microsoft Azure NDv4 were used to train this model. But, achieving the full potential of these supercomputers requires parallelism across thousands of GPUs, along with efficiency and scalability on both memory and compute. Separately, existing parallelism techniques such as data, pipeline, or tensor-slicing have trade-offs in memory and cannot be used to train models at scale. Here’s why: Data parallelism might help achieve good compute efficiency, but it replicates model states and cannot utilise aggregate distributed memory.  Pipeline parallelism can scale efficiency across nodes, but it requires large batch sizes, coarse grain parallelism, and perfect load balancing, which is impossible at scale. Tensor-slicing needs significant communication between GPUs that limits compute efficiency beyond a node where high-bandwidth NVLink is unavailable. By combining NVIDIA Megatron-LM and Microsoft DeepSpeed, the duo created an efficient and scalable 3D parallel system capable of blending data, pipeline, and tensor-slicing based parallelism to address these challenges. For building MT-NLG with 530 billion parameters, each model’s replica consists of 280 NVIDIA A100 GPUs, with 8-way tensor-slicing and 35-way pipeline parallelism across nodes. The data parallelism from DeepSpeed was then used to scale out further to thousands of GPUs. In terms of the hardware system, the model training is done via mixed precision on the NVIDIA DGX SuperPOD-based Selene supercomputer backed by 560 DGX A100 servers networked with HDR InfiniBand in a full-fat tree configuration. Each DGX A100 has 8 NVIDIA A100 80GB Tensor Core GPUs, fully connected by NVSwitch and NVLink. Microsoft used a similar reference architecture for Azure NDv4 cloud supercomputers. Training Datasets The team built its training dataset based on The Pile. Curated by Eleuther AI, it consists of 825 GB worth of English text corpus targeted at training large-scale language models. The data consists of text scraped off from various sources on the internet, including Wikipedia, news clippings, and academic journal repositories. In addition, to diversify the training, the team also collected the Common Crawl (CC) snapshots, RealNews and CC-Stories datasets. Furthermore, they used the deduplication process at the document level using min-hash LSH to compute a sparse document graph and the connected components to identify duplicate documents. Following this, the team then used a priority order based on the quality of datasets. Lastly, they used n-gram based filtering to remove downstream task data from the training datasets to avoid contamination. As a result, the researchers ended with a set of 15 datasets consisting of 339 billion tokens. While training, they opted to blend the datasets into heterogeneous batches as per variable sampling weights (shown below), emphasising higher-quality datasets. They trained the model on 270 billion tokens. DatasetTokens (billions)Weights (per cent)EpochsBooks325.714.31.5OpenWebText214.819.33.6Stack Exchange11.65.71.4PubMed Abstracts4.42.91.8Wikipedia4.24.83.2Gutenberg (PG-19)2.70.90.9BookCorpus21.511.8NIH ExPorter0.30.21.8Pile-CC49.89.40.5ArXiv20.81.40.2GitHub24.31.60.2CC-2020-5068.7130.5CC-2021-0482.615.70.5RealNews21.991.1CC-Stories5.30.90.5 Experiments NVIDIA, along with Microsoft, evaluated MT-NLG by selecting eight tasks spanning five different areas of NLP. Namely, Text prediction task LAMBDA: Here, the model predicts the last word of a given paragraph.Reading comprehension tasks like RACE-h and BoolQ: The model generates answers to questions based on a given paragraph. Commonsense reasoning tasks PiQA, Winogrande, and HellaSwag: Each requires some commonsense knowledge beyond statistical language patterns to solve. Natural language inference: ANLI-R2 and HANS target the typical failure cases of past models. Word sense disambiguation task WiC: The model evaluates polysemy understanding from context. The team evaluated MT-NLG in zero-, one– and few-shot settings without searching for the optimal number of shots. Here are the results: TasksZero-shotOne-shotFew-shotLambada0.766*0.731*0.872*BoolQ0.7820.8250.848RACE-h0.4790.4840.479PiQA0.820*0.810*0.832*HellaSwag0.8020.8020.824WinoGrande0.730.7370.789ANLI-R20.3660.3970.396HANS0.6070.6490.702WiC0.4860.5130.585 What about Bias? As large language models rapidly advance the SOTA language generation, they also suffer immensely from issues like bias and toxicity. NVIDIA and Microsoft believe that understanding and removing these problems in language models is crucial. But, the question is, how are they solving this? As per their observation, the MT-NLG picks up stereotypes and biases from the data it is trained on. The duo said that it looks to address this problem via continued research and help quantify the model’s bias. What’s Next? MT-NLG is an example of when supercomputers like NVIDIA Selene or Microsoft Azure NDv4 are used with DeepSpeed and Megatron-LM software breakthroughs to train large language AI models. NVIDIA and Microsoft believe that the quality and results they have obtained are a step forward in the journey towards opening the full potential of AI in natural language processing (NLP). In other words, the innovations of Microsoft’s DeepSpeed and NVIDIA’s Megatron-LM will benefit existing and future AI model development, thereby making large AI models cheaper and faster to train. “We look forward to how ‘MT-NLG’ will shape tomorrow’s products and motivate the community to push the boundaries of (NLP) natural language processing even further,” said the NVIDIA team.","excerpt":"MT-NLG has 3x the number of parameters compared to the existing largest models – GPT-3, Turing NLG, Megatron-LM and others.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Eleuther AI","GPT-3","largest language models","Machine Learning New","Microsoft","MT-NLG","natural language models","NVIDIA","Switch Transformer"],"author_name":"Amit Naik","publish_date":"2021-10-13T13:00:00","publication_year":"2021","word_count":1116,"keywords":["Scala","AI (Artificial Intelligence)","R","largest language models","artificial intelligence","Machine Learning New","Eleuther AI","RAG","natural language models","NLP","MT-NLG","NVIDIA","Go","AI","GPT-3","semantic search","OpenAI","Switch Transformer","Azure","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","OpenAI","RAG","semantic search","Azure","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidia-microsoft-introduce-new-language-model-mt-nlg-with-530-billion-parameters-leaves-gpt-3-behind\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":39470,"title":"10 Important Pandas Interview Questions Every Beginner Must Know","content":"Pandas is a prominent data-munging tool in Python. This data analysis library is well suited for various kinds of data. In this article, we list down 10 important interview questions on Python pandas one must know. 1| Define Python pandas pandas is an open source, BSD-licensed library providing high-performance, easy-to-use data structures and data analysis tools for the Python programming language. This is a high-level data manipulation tool developed by Wes Mckinney and is built on the Numpy package. This package provides active and flexible data structures in order to make easy working with relational or labelled data. 2| Mention The Different Types Of Data Structures In pandas? There are two data structures supported by pandas library, Series and DataFrames. Both of the data structures are built on top of Numpy. Series is a one-dimensional data structure in pandas and DataFrame is the two-dimensional data structure in pandas. There is one more axis label known as Panel which is a three-dimensional data structure and it includes items, major_axis, and minor_axis. 3| Explain Series In pandas. How To Create Copy Of Series In pandas? Series is a one-dimensional labeled array capable of holding any data type (integers, strings, floating point numbers, Python objects, etc.). The axis labels are collectively referred to as the index. The basic method to create a Series is to call: >>> s = pd.Series(data, index=index), where the data can be a Python dict, an ndarray or a scalar value. To create a copy in pandas, we can call copy() function on a series such that s2=s1.copy() will create copy of series s1 in a new series s2. 4| What Is A pandas DataFrame? How Will You Create An Empty DataFrame In pandas? pandas DataFrame is two-dimensional size-mutable, potentially heterogeneous tabular data structure with labeled axes (rows and columns). It consists of three principal components, the data, rows, and columns. pandas DataFrame can be created from the lists, dictionary, and from a list of dictionary, etc. To create an empty DataFrame in pandas, type import pandas as pd df = pd.DataFrame() 5| Explain Reindexing In pandas. Reindexing means to conform DataFrame to a new index with optional filling logic, placing NA\/NaN in locations having no value in the previous index. It changes the row labels and column labels of a DataFrame. 6| What Are The Key Features Of pandas Library? There are various features in pandas library and some of them are mentioned below Data Alignment Memory Efficient Reshaping Merge and join Time Series 7| What Is pandas Used For? This library is written for the Python programming language for performing operations like data manipulation, data analysis, etc. The library provides various operations as well as data structures to manipulate time series and numerical tables. 8| Explain Categorical Data In pandas. Categoricals are a pandas data type corresponding to categorical variables in statistics. A categorical variable takes on a limited and usually fixed, number of possible values (categories; levels in R). Examples are gender, social class, blood type, country affiliation, observation time or rating via Likert scales. All values of categorical data are either in categories or np.nan. The categorical data type is useful in the following cases: A string variable consisting of only a few different values. Converting such a string variable to a categorical variable will save some memory, see here. The lexical order of a variable is not the same as the logical order (“one”, “two”, “three”). By converting to a categorical and specifying an order on the categories, sorting and min\/max will use the logical order instead of the lexical order, see here. As a signal to other Python libraries that this column should be treated as a categorical variable (e.g. to use suitable statistical methods or plot types). 9| What Are The Different Ways A DataFrame Can Be Created In pandas? DataFrame can be created in different ways here are some ways by which we create a DataFrame: Using List: # initialize list of lists data = [[‘p’, 1], [‘q’, 2], [‘r’, 3]] # Create the pandas DataFrame df = pd.DataFrame(data, columns = [‘Letter’, ‘Number’]) # print dataframe. df Using dict of narray\/lists: To create DataFrame from dict of narray\/list, all the narray must be of same length. If index is passed then the length index should be equal to the length of arrays. If no index is passed, then by default, index will be range(n) where n is the array length. Using arrays: # DataFrame using arrays. import pandas as pd # initialise data of lists. data = {‘Name’:[‘Tom’, ‘Jack’, ‘nick’, ‘juli’], ‘marks’:[99, 98, 95, 90]} # Creates pandas DataFrame. df = pd.DataFrame(data, index =[‘rank1’, ‘rank2’, ‘rank3’, ‘rank4’]) # print the data df 10| What Is Time Series In pandas A time series is an ordered sequence of data which basically represents how some quantity changes over time. pandas contains extensive capabilities and features for working with time series data for all domains. pandas supports: Parsing time series information from various sources and formats Generate sequences of fixed-frequency dates and time spans Manipulating and converting date time with timezone information Resampling or converting a time series to a particular frequency Performing date and time arithmetic with absolute or relative time increments","excerpt":"Pandas is a prominent data-munging tool in Python. This data analysis library is well suited for various kinds of data. In this article, we list down 10 important interview questions on Python pandas one must know. 1| Define Python pandas pandas is an open source, BSD-licensed library providing high-performance, easy-to-use data structures and data analysis […]","categories":["AI Trends"],"tags":[],"author_name":"Ambika Choudhury","publish_date":"2019-05-21T07:00:45","publication_year":"2019","word_count":869,"keywords":["Go","NumPy","API","AI","Scala","Python","Ray","programming_languages:Python","R","Pandas"],"extracted_tech_keywords":["AI","Ray","Pandas","NumPy","Python","R","Go","Scala","API","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-important-pandas-interview-questions-every-beginner-must-know\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10047505,"title":"Should Neural Networks Compose Music By The Same Logic & Process As Humans Do?","content":"The study of the science of harmony can be traced back to the times of Pythagoras. The development of the Pythagorean tuning system is commonly credited to Pythagoras, whereas Greek mathematician Euclid documented numerous experiments on rational tuning. Euler, too, has expressed his interest in studying chord aesthetics. The inexplicable creative ingenuity behind the most remarkable pieces of music can be discussed through concepts of string lengths and other proposals, but why a certain sound pleases the ear is still a mystery. While physicists, mathematicians and philosophers are trying to figure out a universal theory of harmony, the artificial intelligence community has taken the ambitious task of not only understanding composition but making an algorithm to compose music. A couple of years ago, Google came up with a machine learning model called Coconet, which was trained on more than 300 works of the German composer Bach. The model was fed with incomplete notes as input and made to play complete scores. To do a composition similar to humans, a deep learning model has to: Harmonize melodies, create smooth transitions, rewrite music, and compose from scratch. Music composition (or generation) is the process of creating or writing a new piece of music. But, to accomplish something remotely close to what musicians do, they have to ask themselves a few fundamental questions: Are the current DL models capable of generating music with a certain level of creativity? What is the best neural network architecture to perform music composition?Can a DL model generate entire structured music pieces? Can DL models compose music that is totally different from training data? How much data do DL models need for music generation?Should neural networks compose music by following the same logic and process as humans do? To answer these questions, researchers from the University of Zaragoza, Spain, have compared the human composition process and the music generation process with deep learning and the artistic and creative characteristics presented by the generated music. The researchers look at music composition as a unique human capacity to understand and produce an indefinitely large number of sentences in a language, which have never been encountered or spoken before. This perspective gave them a starting point for designing an AI-based music composition algorithm. But, before we dive deep into AI-based composition, let’s try to understand how a human composer comes up with certain melodies. In classical music, observe the researchers, one starts with a small unit of one or two bars called motif and develops it to compose a melody or music phrase, and in styles like pop or jazz, it is more common to take a chord progression and compose or improvise a melody ahead of it. Music is usually defined as a succession of pitches or rhythms, or both, in some definite patterns. Musical scores are three dimensional objects; there are four voices, and the music for each voice can be represented as a two-dimensional array with time extending horizontally and pitches laid out vertically. If one has to build a model, these vectors can be used to construct a probability distribution. According to the researchers, the generated polyphonic melodies of early neural networks (RNNs and LSTMs) lack in terms of quality harmonic content. This is because neural networks, which are trained to generate music, could not understand the many intricacies of the language of music, as encoding such information in tokens to be fed into neural networks is a challenge in itself. When Transformer architecture (GPT-2) was used, the results were more coherent. Even the generative models such as GANs showed promise. The researchers believe that attention-based architectures such as Transformers and generative models such as GANs are the best shot yet at DL based music composition. For example, the report stated that MusicVAE and other DL models for music generation showed that new music can be composed without imitating existing music or committing plagiarism. However, models such as MusicVAE demonstrated that AI for music composition would need tremendous amounts of data. For instance, MusicVAE uses 3.7 million melodies, 4.6 million drum patterns, and 116 thousand trios. In contrast, the GPT-2 based model uses lots of text data. “This leads us to affirm that DL models for music generation do need lots of data, especially when training Generative Models or Transformers,” explained the researchers. Setting aside the technical challenges, there are strong arguments that people prefer music created by people. The role of AI in composition can be nothing more than making stock music. One might even argue that tinkering with AI for creating art is a doorway to artificial general intelligence (AGI). Instead, a better, more neutral premise would be to explore how AI can be used to create tools that a human composer would eventually use. “In the near future, it would be likely that AI could compose structured music from scratch, but the question here is whether AI models for music generation will be used to compose entire music pieces from scratch or whether these models would be more useful as an aid to composers, and thus, as an interaction between humans and AI,” concluded the researchers.","excerpt":"Not Quite My Tempo: Why Should We Care About AI Generated Music?","categories":["Deep Tech"],"tags":["AI music"],"author_name":"Ram Sagar","publish_date":"2021-09-02T15:00:00","publication_year":"2021","word_count":852,"keywords":["Go","AI music","artificial intelligence","machine learning","AI","neural network","Transformers","RAG","Ray","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","Ray","Transformers","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/should-neural-networks-compose-music-like-humans\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10009664,"title":"How Effective Are Bug Bounty Programs As Security & Compliance Strategies","content":"Facebook recently announced Hacker Plus, a loyalty program for its bug bounty program. As per the company’s claim, it is the first of its kind program, built on the loyalty programs issued by airlines and hotels. The social media tech giant will now evaluate users’ performance based on cumulative quantity, score, and signal-to-noise ratio of the bugs submitted over a period of a year. As soon as the bug is submitted, the user is then included and ranked on the Hacker Plus loyalty program. Based on their scores, the bug hunters will be placed into one of the five categories — Bronze, Silver, Gold, Platinum, and Diamond. ‘Bug hunters’ in each of these categories will be receiving a bonus over their bounty amounts, for example, bronze league members will receive 5% bonus, while diamonds will receive a 20% bonus. This step by Facebook further incentivises the bug hunting by external players, encouraging more participation and also reflecting more investment from the organisational point of view. Facebook’s case is not an isolated one. As per a February 2020 report from HackerOne, its popularity has soared in the past few years. In 2019, hackers collectively earned $40 million from such bug-hunting programs; this amount is almost equal to the total bounty received for all the preceding years combined. Its increasing popularity brings forth a very pertinent question — Are companies getting more reliant on bug bounty programs as security and compliance strategy? If yes, how effective is it? How Do Bug Bounty Programs Plug Loopholes A bug bounty program is an initiative through which organisations provide rewards to external security researchers for identifying and reporting vulnerabilities and loopholes in their public-facing digital systems. While a few of these programs are invite-based, most of these initiatives are open for all. Once the loophole is identified, the researcher is then required to submit a proof of concept with their report to the concerned organisation. As per the Data Breach Report 2020, it takes an average of 280 days for an organisation to identify a security breach. This gives an attacker ample time to prey upon their target’s most important assets. It is where the role of an external or a third party ‘bug-hunter’ comes into play. These programs work as a proactive approach for the organisation towards their security efforts. In their absence, organisations will be just forced to assume a reactionary stance where they wait for the attacker to attack and only then fix the underlying vulnerability. Security and Compliance Issues While Bug Bounty Programs help organisations in their security and compliance strategy, they also have a few downsides: Firstly, these programs attract all sorts of users and researchers. They could be both whitehat or blackhat hackers. It doesn’t help the case much as the blackhat hackers are already on a prowl to hunt for vulnerability and the further announcement of a bug bounty program may draw them to a previously unknown target. In a worst-case scenario, the blackhat hackers may go up and beyond the predetermined testing perimeters to compromise a secondary system. Another major shortcoming of a bug bounty program is that nobody really has the complete ownership of the project. Unlike a penetration test where a dedicated resource is assigned to the project which in turn uses a specific methodology to review the testing scope from all ends, the bug hunters get rewarded per-bug basis; thus nobody can really certify whether all risks have been identified and reviewed. Lastly, a set of poorly written legal rules and a scoop of the bounty program may give rise to potential legal threats. The organisation needs to draft clear rules to avoid ambiguity at later stages and also make sure that the researcher goes through these rules beforehand. The last thing we need is running into legal trouble due to undefined framework. Wrapping Up In the past, these bug bounty programs have led to uncovering some of the most critical bugs in organisations’ set-up. Such initiatives have helped in identifying and fixing issues pertaining to cross-site scripting flaws, improper authentication, privilege escalation, among other issues, classified as ‘critical’ or ‘high’ severity. However, organisations mustn’t think of it as a one all and be all solution for their security issues. In the absence of a comprehensive security plan, organisations will not be able to monitor vulnerabilities more effectively. At best, these bug bounty programs can be thought of as complementary to already robust in-house security solutions.","excerpt":"Facebook recently announced Hacker Plus, a loyalty program for its bug bounty program. As per the company’s claim, it is the first of its kind program, built on the loyalty programs issued by airlines and hotels.  The social media tech giant will now evaluate users’ performance based on cumulative quantity, score, and signal-to-noise ratio of […]","categories":["AI Trends"],"tags":["security breach"],"author_name":"Shraddha Goled","publish_date":"2020-10-15T12:00:46","publication_year":"2020","word_count":742,"keywords":["Go","AWS","AI","Scala","Git","RAG","Aim","ViT","GAN","security breach","R"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","R","Go","Scala","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-effective-are-bug-bounty-programs-as-security-compliance-strategies\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":60586,"title":"9 Best Online Data Courses That Offer One-On-One Mentorship","content":"While one embarks on a journey to become a data scientist, they deal with their share of confusion. Data science is a vast discipline, making it challenging to become a good data scientist as well as secure a good job in this field. This demands that one take the help and guidance of a mentor to deal with the hurdles that may come with a career in data science. Mentors play a crucial role in the lives of data scientists because they not only clear doubts and confusion, but also help bridge the knowledge gap. Additionally, they help improve skills and help aspirants understand data science as a field. Below, we have listed some of the best online (paid) data science courses that also offer a one-on-one mentorship opportunity: Postgraduate Program In Data Science And Machine Learning Duration: 10 months Provider: Jigsaw Academy and University of Chicago’s Graham School The program is an excellent blend of data science, business analytics, visualisation and deep learning with application of advanced analytics models for AI, deep learning and cognitive computing. The course follows a combination of both in-person and online teaching format. It offers seminars and projects by industry leaders and hands-on learning along with one-on-one mentorship to the students. Price: Rs. 3.65 Lakhs + GST Start here. PGP-???? Duration: 6 months + 3 months Provider: INSOFE This is a 6 + 3 months program where in the first 6 months, students complete their PGP program with a specialisation. In the next 3 months, the students spend the time in Canada working on their industry project, this period can be extended to 6 months if the students decide to do so. Along with one-on-one mentorship, the students during this time will also attend classes by Technology Product Management and Innovation by world class professors and researchers at Carleton university. Price: CAD $13,800 Start here. PG Program in Artificial Intelligence and Machine Learning Duration: 12 months Provider: Great Learning This is India’s #1 ranked AI program. The program consists of 20+ professors and 1300= industry mentors which also include 2 award winning faculties. Great Learning offers to build real world applications under guidance from industry experts and uses data from companies like Uber, Amazon and Netflix. Great Learning offers online personalised mentorship every week and provides a hands-on exposure through ample projects. Price: Rs. 2,40,000+ GST Start here. Advanced Certification Program for Working Professionals Duration: 13 weeks Provider: TalentSprint This program is made keeping the tech professionals in mind. The is led by collaborative  faculty from industry, academia and global blue-chip institutions. The program is a 5-step learning process of lecture, one-on-one mentorship and access to their #1 labs in AI and Robotics research. The location for this program is Hyderabad and the classes are conducted on the weekends so that it suits working professionals. Price: 2,00,000 + GST Start here. PG Diploma in Data Science Duration: 12 months, 12-15 hours a week Provider: Upgrad This course offers five specializations as per learners’ background and career aspirations. By the end of the course, candidates will have a PG Diploma in Data Science from IIT-B with a specialization. The course offers specializations in: Natural Language ProcessingBusiness AnalyticsDeep LearningBusiness Intelligence\/ Data AnalyticsData Engineering The course also provides dedicated one-on-one career counselling sessions and mock interviews with hiring managers. Along with hiring managers, one gets personalized training mentorship in groups with industry experts. Start here. Data Scientist- Masters Program Duration: Online Flexible Provider: Simplilearn in collaboration with IBM This masters program is from simplilearn in collaboration with IBM. This program provides hands-on exposure to key technologies like R, Python, SAS, Tableau, Hadoop, and Spark. The program contains 15+ real life projects  and offers a Simplilearn JobAssist program to help you get noticed by the top hiring companies. The mentorship will be during the JobAssist program which is Simplilearn’s India-specific offering in partnership with IIMJobs.com. This program can be joined once the masters program is completed. Price: Rs. 49,999\/- Start here. Data Science with Machine learning and AI Pro+ Certification Course with Honeywell Duration: 6 months Weekend Provider: ivy Professional School created with Honeywell The course provides hands-on learning on real-life projects and case studies using tools and programming languages like R, Python, Tableau and also leverages the latest models used in machine learning & data visualisation. The entire framework of the course is in close collaboration with Honeywell’s senior data science and machine learning practitioners and the candidate has the opportunity to be mentored by them during the course. Price: Rs. 77,000 (Classroom), Rs. 64,000 (Online) Start here. Data Science Career Track Duration: Six months Provider: Springboard India This is reportedly the only data science program in India that comes with a job guarantee. The course focuses on the foundations of data science and lets you choose your focus areas. This could be advanced machine learning, NLP or deep learning. The course offers 14 real-life projects, which include two industry-worthy capstone ones that will showcase your skills in the best way possible. They offer personalized career training with one-on-one mentorship from industry experts. Price: Rs 1,95,000\/- Start here. Post Graduate Program in Artificial Intelligence and Machine Learning Duration: 450+ hours live class Provider: Edureka This program is for beginners in the Ai and Ml space with basic programming skills. This program helps one kick start their careers in AI, ML and data science. The program mainly focuses on concepts like Python fundamentals, exploratory data analysis. Inferential statistics, supervised learning, regression algorithm, classification algorithm, unsupervised learning etc. The course also provides mentorship from NIT Warangal’s experienced faculty. Price: Rs. 2,22,450 + GST Start here","excerpt":"While one embarks on a journey to become a data scientist, they deal with their share of confusion. Data science is a vast discipline, making it challenging to become a good data scientist as well as secure a good job in this field. This demands that one take the help and guidance of a mentor […]","categories":["AI Trends"],"tags":["best online data science masters","big data certification","Courses","Data Science Certification","Data Scientist Jobs","mentorship","one on one mentorship","online masters analytics","pgp program in data science"],"author_name":"Sameer Balaganur","publish_date":"2020-04-01T11:00:00","publication_year":"2020","word_count":936,"keywords":["best online data science masters","Data Science Certification","online masters analytics","deep learning","data science","Data Scientist Jobs","artificial intelligence","RAG","NLP","pgp program in data science","analytics","mentorship","one on one mentorship","machine learning","AI","ML","big data certification","Python","Courses"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","data science","analytics","RAG","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/9-best-online-data-courses-that-offer-one-on-one-mentorship\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":54571,"title":"Challenges Of Implementing Natural Language Processing","content":"Invaluable support for artificial intelligence (AI), natural language processing (NLP) helps in establishing effective communication between computers and human beings. In recent years, there have been significant breakthroughs in empowering computers to understand human language using NLP. However, the complex diversity and dimensionality characteristics of the data sets, make this simple implementation a challenge in some cases. The Case For NLP As text and voice-based data, as well as their practical applications, vary widely, NLP needs to include several different techniques for interpreting human native language. These could range from statistical and machine learning methods to rules-based and algorithmic. NLP has immense potential in real-life application areas such as understanding complete sentences and finding synonyms of matching words, speech recognition, speech translation, and writing complete, grammatically correct sentences, and this need has now become amplified. While background, domain knowledge and frameworks (e.g. algorithms and tools) are the critical components of the NLP system, it is not a simple and easy task of making machines to understand natural human language. The process includes several activities such as pre-processing, tokenisation, normalisation, correction of typographical errors, Named Entity Reorganization (NER), and dependency parsing. To attain high-quality models, NLP performs an in-depth analysis of user inputs like lexical analysis, syntactic analysis, semantic analysis, discourse integration, and pragmatic analysis, etc. Challenges for NLP implementation Data challenges The main challenge is information overload, which poses a big problem to access a specific, important piece of information from vast datasets. Semantic and context understanding is essential as well as challenging for summarisation systems due to quality and usability issues. Also, identifying the context of interaction among entities and objects is a crucial task, especially with high dimensional, heterogeneous, complex and poor-quality data. Data ambiguities add more challenges to contextual understanding. Semantics are important to find the relationship among entities and objects. Entities and object extraction from text and visual data could not provide accurate information unless the context and semantics of interaction are identified. Also, the currently available search engines can search for things (objects or entities) rather than keyword-based search. Semantic search engines are needed because they better understand user queries usually written in natural language. The next challenge is the extraction of the relevant and correct information from unstructured or semi-structured data using Information Extraction (IE) techniques. It is necessary to understand the competency and limitations of the existing IE techniques related to data pre-processing, data extraction and transformation, and representations for vast volumes of multidimensional unstructured data. Higher efficiency and accuracy of these IE systems are very important. But, the complexity of big and real-time data brings challenges for ML-based approaches, which are dimensionality of data, scalability, distributed computing, adaptability, and usability. Effectively handling sparse, imbalance and high dimensional datasets are complex. Another challenge is that a user expects more accurate and specific results from Relational Databases (RDB) for their natural language queries like English. To retrieve information from RDBs for user requests in natural language, the requests have to be converted into formal database queries like SQL. They can also reuse the existing application backend services. This approach leverages NLP to understand the user requests in natural language and prepare application service request URLs to retrieve data from the connected databases. However, in practice, translating NLP queries to formal DB queries or service request URL is quite complicated due to several factors. These could be the complex DB layouts with table names, columns, and constraints, etc., or the semantic gap between user vocabulary and DB nomenclature. NLP search over databases requires domain-specific models for intent, context, Named Entity identification and extraction. The ambiguity of texts, complex nested entities, identification of contextual information, noise in the form of homonyms, language variability, and missing data pose significant challenges in entity recognition. Text related challenges Large repositories of textual data are generated from diverse sources such as text steams on the web, communications through mobile and IoT devices. Though ML and NLP have emerged as the most potent and most used technology applied to the analysis of the text and text classification remains the most popular and the most used technique. Text classification could be Multi-Level (MLC) or Multi-Class (MCC). In MCC, every instance could be assigned to only one class label, whereas MLC is a classification that assigns multiple labels to a single instance. Solving MLC problems requires an understanding of multi-label data pre-processing for big data analysis. MLC can become very complicated due to the characteristics of real-world data such as high-dimensional label space, label dependency, and uncertainty, drifting, incomplete and imbalanced. Data reduction for large dimensional datasets and classifying multi-instance data is also a challenging task. Then there are the issues posed by a language translation. The main challenge with language translation is not in translating words, but in understanding the meaning of sentences to provide an accurate translation. Each text comes with different words and requires specific language skills. Choosing the right words depending on the context and the purpose of the content, is more complicated. A language may not have an exact match for a certain action or object that exists in another language. Idiomatic expressions explain something by way of unique examples or figures of speech. Most importantly, the meaning of particular phrases cannot be predicted by the literal definitions of the words it contains. Somewhat related is another challenge, that of the inability to accurately deal with new users and products that do not have any history. The user-item rating matrix is very sparse (data sparsity) because stores have many products that will not be rated by many users. The standard challenge for all new tools, is the process, storage and maintenance. Unlike statistical machine learning, building NLP pipelines is a complex process — pre-processing, sentence splitting, tokenisation, pos tagging, stemming and lemmatisation, and the numerical representation of words. NLP requires high-end machines to build models from large and heterogeneous data sources. NLP models are larger and consume more memory compared to statistical ML models. Several intermediate and domain-specific models have to be maintained (e.g. sentence identification, pos tagging, lemmatisation, word representation models like TF-IDF, word2vec, etc.). Rebuilding all the intermediate NLP models for new data sets may cost more. Conclusion Most of the challenges are due to data complexity, characteristics such as sparsity, diversity, dimensionality, etc. and the dynamic nature of the datasets. NLP is still an emerging technology, and there are a vast scope and opportunities for engineers and industries to deal with many open challenges of implementing NLP systems. With the special focus on addressing NLP challenges, organisations can build accelerators, robust, scalable domain-specific knowledge bases and dictionaries that bridges the gap between user vocabulary and domain nomenclature. The proficient and skilled pool of data scientists working for any product engineering services provider will be capable of building customised architectures and NLP pipeline to enable NLP search on different kind of datasets (structured & unstructured). The quality research and adoption of state-of-the-art technologies like linked data, knowledge graph, etc., can improve the quality of data with enriched meanings, linking data sources with appropriate and meaningful relationships among them. Last but not least, developing accelerators and frameworks make complex NLP implementations more affordable and provide improved performance. This article is a part of the AIM Writers Programme. If you wish to write for us, email us at info@analyticsindiamag.com","excerpt":"Invaluable support for artificial intelligence (AI), natural language processing (NLP) helps in establishing effective communication between computers and human beings. In recent years, there have been significant breakthroughs in empowering computers to understand human language using NLP. However, the complex diversity and dimensionality characteristics of the data sets, make this simple implementation a challenge in […]","categories":["AI Features"],"tags":["object store database"],"author_name":"Suresh Babu Golla","publish_date":"2020-01-24T13:00:00","publication_year":"2020","word_count":1221,"keywords":["semantic search","text classification","machine learning","artificial intelligence","AI","ML","RAG","NLP","Aim","analytics","object store database"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","analytics","Aim","RAG","text classification","semantic search"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/challenges-of-implementing-natural-language-processing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":65598,"title":"Microsoft Launches New Tools For Building More Responsible & Fairer AI Systems","content":"Microsoft, during its Build developer conference, has put a strong emphasis on machine learning, along with plenty of new tools and features that the company is going to work on for building more responsible and fairer AI systems — both in the Azure cloud and Microsoft’s open-source toolkits. The company stated that these new tools would be utilised for differential privacy and for creating a system that would ensure that models are working well across different groups of people. Further, these new tools would enable businesses to make the best use of their data while still following strict regulatory requirements. During the announcement, Microsoft stated that, as developers are increasingly tasked to learn how to build artificial intelligence models, the developers regularly end up asking about the system explainability and its compliance with non-discrimination and privacy regulations. And for that, developers would require tools that can help them better interpret their models’ results. One of the tools is interpretML, launched a while ago, and also the Fairlearn toolkit, which can be used by developers to assess the fairness of machine learning models — currently available as an open-source tool and will be built into Azure Machine Learning next month. Microsoft also noted that with regards to differential privacy, which allows developers to get insights from private data while still protecting private information, The company announced WhiteNoise, developed in partnership with Harvard’s Institute for Quantitative Social Science. This new open-source toolkit that’s available both on GitHub and on Azure Machine Learning.","excerpt":"Microsoft, during its Build developer conference, has put a strong emphasis on machine learning, along with plenty of new tools and features that the company is going to work on for building more responsible and fairer AI systems — both in the Azure cloud and Microsoft’s open-source toolkits. The company stated that these new tools would be utilised for […]","categories":["AI News"],"tags":["Machine Learning","Microsoft","Responsible AI"],"author_name":"Sejuti Das","publish_date":"2020-05-20T18:43:11","publication_year":"2020","word_count":250,"keywords":["Go","artificial intelligence","machine learning","AI","R","ML","Machine Learning","Responsible AI","Git","differential privacy","GitHub","Azure","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","differential privacy","Azure","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-launches-new-tools-for-building-more-responsible-fairer-ai-systems\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10164556,"title":"Google Launches Gemini Code Assist, a Free AI Coding Tool for Developers","content":"Google has announced the public preview of Gemini Code Assist, a free AI-powered coding assistant for individuals. The tool is globally available and supports all programming languages in the public domain. “More than 75% of developers are relying on AI in their daily responsibilities,” said Ryan J Salva, senior director of product at Google. “More than 25% of all new code at Google is generated by AI, then reviewed and accepted by engineers.” Gemini Code Assist for individuals provides AI-driven code completion, generation, and chat capabilities. It is available in Visual Studio (VS) Code and JetBrains IDEs, as well as in Firebase and Android Studio. Google said its AI coding assistant offers “practically unlimited capacity with up to 1,80,000 code completions per month”. The company also introduced Gemini Code Assist for GitHub, which offers AI-powered code reviews for both public and private repositories. The tool can detect stylistic issues and bugs, automatically suggesting code changes and fixes. Developers can use natural language to generate, explain, and improve code. Google mentioned the tool allows prompts like “Build me a simple HTML form with fields for name, email, and message, and then add a ‘submit’ button”. The AI assistant includes a large context window that supports up to 1,28,000 input tokens in chat. This enables developers to work with large files and integrate local codebases. Google aims to make AI coding tools accessible to students, hobbyists, freelancers, and startups. “With a worldwide population of developers forecasted to grow to 57.8 million by 2028, we think AI should be available to them whether they can pay for it or not,” Salva said. Sign-up requires only a personal Gmail account. Google is collecting user feedback from the public preview through feedback forms in IDEs and GitHub. Users interested in productivity metrics, private repository integrations, or Google Cloud service support can explore Gemini Code Assist Standard or Enterprise.","excerpt":"Google said its AI coding assistant offers “practically unlimited capacity with up to 1,80,000 code completions per month”.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Google"],"author_name":"Aditi Suresh","publish_date":"2025-02-25T16:30:00","publication_year":"2025","word_count":313,"keywords":["Go","AI coding","AI","ML","Git","Aim","ViT","Google","GitHub","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Git","GitHub","ViT","startup","AI coding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-launches-gemini-code-assist-a-free-ai-coding-tool-for-developers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10141384,"title":"AI Boom Pushes NVIDIA Q3 Revenue to $35.1B as Blackwell Demand Soars","content":"NVIDIA reported record revenue of $35.1 billion for the third quarter of fiscal 2025, a 17% rise from the previous quarter and a 94% increase year-over-year. The data centre segment contributed $30.8 billion, up 17% from the prior quarter and 112% from the same period last year. “The age of AI is in full steam, propelling a global shift to NVIDIA computing,” said Jensen Huang, founder and CEO of NVIDIA. “Demand for Hopper and anticipation for Blackwell — in full production — are incredible as foundation model makers scale pretraining, post-training, and inference.” Huang also highlighted the broader impact of AI, saying, “AI is transforming every industry, company, and country. Enterprises are adopting agentic AI to revolutionise workflows. Industrial robotics investments are surging with breakthroughs in physical AI. And countries have awakened to the importance of developing their national AI and infrastructure.” Looking ahead, NVIDIA provided an optimistic forecast for the fourth quarter, projecting revenue of $37.5 billion, which surpasses analyst expectations. Despite the impressive results, NVIDIA’s stock experienced a slight dip in after-hours trading, possibly due to investors’ even higher expectations. During the earnings call, Huang said that Blackwell production is running at full steam. NVIDIA’s CFO Colette M Kress revealed that the company shipped 13,000 Blackwell GPU samples to customers in the third quarter, including one of the first Blackwell DGX engineering samples to OpenAI. “Blackwell is now in the hands of all of our major partners, and they are working to bring up their data centers. We are integrating Blackwell systems into the diverse data centre configurations of our customers,” said Kress. “Blackwell demand is staggering, and we are racing to scale supply to meet the incredible demand customers are placing on us.” She further reported exceptional demand for Hopper GPUs, with H200 sales rising significantly quarter-over-quarter to reach double-digit billions. The CFO described the H200’s production ramp as the fastest in NVIDIA’s history, noting its delivery of up to twice the inference performance and a 50% improvement in total cost of ownership (TCO).","excerpt":"Blackwell production is running at full steam- Jensen Huang.","categories":["AI News"],"tags":["NVIDIA","Quarterly Earnings"],"author_name":"Siddharth Jindal","publish_date":"2024-11-21T19:08:33","publication_year":"2024","word_count":337,"keywords":["agentic AI","OpenAI","AI","programming_languages:R","R","RPA","Git","ai_applications:robotics","NVIDIA","Quarterly Earnings"],"extracted_tech_keywords":["AI","agentic AI","OpenAI","R","Git","RPA","programming_languages:R","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-boom-pushes-nvidia-q3-revenue-to-35-1b-as-blackwell-demand-soars\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10060979,"title":"AICTE launches virtual internship program in AI for 5,000 students","content":"The All India Council of Technical Education backed by AWS Academy and EduSkills has launched a virtual internship program in AI for students. The course duration is two months and the students will receive a certificate and a digital badge upon completion. The internship will be granted to 5,000 students. The course is aimed at creating an industry-ready workforce and groom students to become leaders in emerging technologies. Students currently pursuing BE, BTech, ME, MTech, MCA, diploma etc in any engineering and polytechnic institutions which are members of EduSkills can participate in the program. How to apply for the internship program: a. Students need to register for this program through AICTE Internship Portal b. The selected students list will be sent to the concerned institute authorities. c. Shortlisted students will be enrolled on AWS Academy Portal and will be given course access. d. Students need to complete the course within 4 weeks by giving online assessments. e. Students who successfully complete the online course will be assigned a project. f. Students with the help of their faculties need to complete the project within 4 weeks and submit it as per the format. g. During the project duration, students will attend 4 days (8 hours) mentoring sessions by industry experts. Students will get one day (2 Hrs) career advancement sessions from HRs. Apply here","excerpt":"Students need to complete the course within 4 weeks by giving online assessments.","categories":["AI News"],"tags":["Virtual Internship Program"],"author_name":"SharathKumar Nair","publish_date":"2022-02-17T19:56:49","publication_year":"2022","word_count":223,"keywords":["AWS","AI","cloud_platforms:AWS","programming_languages:R","Git","Virtual Internship Program","Aim","R"],"extracted_tech_keywords":["AI","Aim","AWS","R","Git","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aicte-launches-virtual-internship-program-in-ai-for-5000-students\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10097702,"title":"Big Tech’s Race Towards Nuclear Energy","content":"While Nolan’s Oppenheimer painted a powerful picture of how inventions like the A-bomb can horribly go wrong and change the course of human history, the big techs, who seem unimpressed at first, are changing the narrative for better or for worse. Some of the biggest tech players, such as Google DeepMind, and OpenAI have entered the space to harness nuclear energy to mostly advance AI research, and achieve net neutral goals. Google DeepMind partnered with Swiss Plasma Center at EPFL, Switzerland, last year, to apply AI to create sustainable energy. By using reinforcement learning, plasma shape accuracy and containment, the key components for successful nuclear fusion, were improved. Since the development in 2022, the company has now made incremental advancements towards making the technology more usable with the plasma shape accuracy in simulation improving to 65%. As promising as that sounds, the advancement is a promising step towards implementing AI that facilitates the production of renewable energy. A recent research report, titled Towards practical reinforcement learning for tokamak magnetic control, talks about the research work for improving reinforcement learning used for magnetic control of Tokamaks, which is a device used for magnetic confinement of plasma in nuclear fusion experiments: plasma is used to fuse atoms to create energy in the fusion process. However, the report calls out the improvement areas in matching simulation with hardware performance, and also suggests exploring alternative model architectures such as state-space models and foundation models. While this advancement seems like an in-house production to contribute toward nuclear energy efforts, companies are taking the route of investing in nuclear fusion companies. Nuclear Energy Fixation Nuclear energy, considered to be a safe and renewable form of energy, has seen a surge in investments by Silicon Valley billionaires in the last few years. From 2015 to 2021, investment in nuclear energy grew around 325% by volume and 3642% by dollar value. According to the International Energy Agency, renewable capacity will meet 35% of global power generation by 2025. The US government’s initiative towards supporting clean energy for climate change by offering tax breaks, incentives and federal funding, have encouraged tech companies to push towards investing in nuclear fusion companies. Furthermore, the Russia-Ukraine crisis has also pushed the acceleration on nuclear power startups to remove reliance on Russia. Big Tech Chases Last year, Google and Chevron were part of $250 million funding for TAE Technologies, a nuclear fusion startup. The partnership between Google and TAE dates back to 2014, where the company has been providing the fusion startup with AI and computational power. Much before the ChatGPT rage, with his list of investments across domains, Altman had a vested interest in the field of nuclear energy in the last few years. OpenAI invested $375 million in nuclear fusion startup Helion Energy. His aim: affordable and accessible electricity with climate change. It’s also possible that his goals are aligned towards powering the future of his company. Interestingly, Microsoft, which has invested in OpenAI, has agreed to buy electricity from Helion Energy in 2028. Helion is not the only nuclear company which Altman is interested in. Altman-led Oklo Inc that designs and deploys advanced fission power plants announced that the company will merge with AltC Acquisition Corporation where he serves as CEO and director. The company is said to raise $500 million. He even emphasised on how intelligence and energy will go hand-in-hand for a bright future. Amazon’s Jeff Bezos is not far behind. He had invested in a Canadian company General Fusion, which will build a nuclear fusion facility in the UK by 2025. The project is a collaboration of General Fusion and the UK Atomic Energy Authority. Bill Gates, who believes that nuclear energy, if done right, “will help solve our climate goals”, has founded TerraPower, an American nuclear reactor design and development engineering company. The company is planning to open a nuclear power plant in Wyoming. Billionaire investor Peter Thiel is no stranger to nuclear power either. His VC firm, Founders Fund, made a $2 million investment in Transatomic Power, a startup that works on developing nuclear reactors that can convert nuclear waste into clean energy. He has also invested in Helion Energy. With big tech and billionaire investors pumping money into nuclear startups, it is possible that the nuclear energy race is running in parallel with the AGI race big techs are competing in.","excerpt":"In 2022 alone, investments over $3.4 billion went into nuclear energy, with tech biggies such as Bill Gates, Jeff Bezos, Peter Thiel, Sam Altman, and others vested in it","categories":["AI Trends"],"tags":["ChatGPT","energy","Google Deepmind","nuclear fusion","OpenAI","Sam Altman","silicon valley"],"author_name":"Vandana Nair","publish_date":"2023-07-28T12:15:37","publication_year":"2023","word_count":727,"keywords":["Go","ChatGPT","energy","Sam Altman","OpenAI","AI","RAG","nuclear fusion","GPT","Aim","Google Deepmind","silicon valley","foundation models","R","startup"],"extracted_tech_keywords":["AI","foundation models","ChatGPT","OpenAI","Aim","RAG","R","Go","GPT","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/big-techs-race-towards-nuclear-energy\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10141087,"title":"Cursor, GitHub Copilot Rival Codeium Launches Windsurf – First Agentic IDE for Coding","content":"Codeium has launched its first agentic IDE, Windsurf aimed to collaborate with users like a Copilot and tackle complex tasks independently like an agent. Setting up is easy, it pulls all plugins and settings from VS Code, just like Cursor. Windsurf IDE combines AI with intuitive design, setting a new standard in developer tools. With an effortless setup that syncs plugins and settings from VS Code, Windsurf ensures continuity for developers. Codeium’s standout features include contextual awareness to optimise suggestions, command execution for streamlined workflows, and Cascade, an innovation that lets developers resume tasks seamlessly. It also includes multi-file editing which is aimed to enhance productivity, making complex coding tasks more manageable. Windsurf redefines how developers interact with their tools, heralding a new era of intelligent development environments. As Varun Mohan, CEO of Codeium has earlier remarked, “The future of coding isn’t just about writing lines of code faster—it’s about enabling developers to think bigger, push boundaries, and achieve the extraordinary.” This release comes on the heels of Codeium securing $150 million in funding, reaching a unicorn valuation of $1.25 billion. The company aims to rival Cursor and establish dominance in the AI development tools market. With Cursor gaining attention for its innovations, Codeium’s Windsurf appears to be a strategic response to secure its position as a market leader. Recently Codeium also launched Cortex, an AI reasoning engine for managing complex coding tasks, and Forge, an AI-assisted tool that enhances code review efficiency and culture. These innovations are part of Codeium’s mission to transform software development by making coding faster, smarter, and more intuitive.","excerpt":"Windsurf redefines how developers interact with their tools, heralding a new era of intelligent development environments.","categories":["AI News"],"tags":["AI coding","Codeium","coding","Windsurf"],"author_name":"Tanisha Bhattacharjee","publish_date":"2024-11-18T12:47:43","publication_year":"2024","word_count":264,"keywords":["Windsurf","AI coding","funding","unicorn","programming_languages:R","AI","innovation","coding","Codeium","ML","Aim","ViT","R"],"extracted_tech_keywords":["AI","ML","Aim","R","ViT","innovation","unicorn","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cursor-github-copilot-rival-codeium-launches-windsurf-first-agentic-ide-for-coding\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093924,"title":"You Are to be Blamed for ChatGPT’s Flaws","content":"Should we trust ChatGPT? Bard and ChatGPT claim to give precise answers to all the world’s questions until hallucination kicks in. They might be a big source of information for many, but at the same time, they are criticised over biases and misinformation. Regulatory authorities and governments all over the world have been breaking a sweat over controlling the misinformation being spread through these platforms. During the US Senate hearing on AI oversight, Open AI CEO Sam Altman said, “This [ChatGPT] is a tool that can generate content more efficiently than ever before. Here, the user can test the accuracy, change it if they don’t like it and get another version. But the content generated still spreads through social media, texts, or other similar ways.” He explained how the interaction with ChatGPT is a single-player experience where the user is interacting with just a tool generating content, and not sharing it without their consent. Altman, too, expressed concerns about the technology and called for regulations around it. But when asked about the impact a technology like this can have on elections or spreading misinformation, he agreed with the premise, but argued for a different lens to frame regulations around this, not like social media. AI takes the fall for bad actors Speaking of misinformation, the chatbot can sometimes be forced to generate false information by the user with certain prompts. Even in cases like these, ChatGPT often refuses to generate certain stuff that can be hateful or false. So, the blame appears to be more on the side of the user, than ChatGPT. If someone wants to create a fake ad or some false information about elections, they can do that regardless of ChatGPT or similar technology. Even before such AI models, the internet was filled with misinformation splattered on Twitter, which spread like wildfire – about COVID vaccines or presidential elections. The recent coverage of the fake photograph of an explosion at the Pentagon was another example of how people are vulnerable to misinformation on social media. Not just with AI, similar photographs can be made using Photoshop. The person who believes the photograph is true and spreads it on social media, or news platforms, is the one responsible to verify its origin. We are not saying that generating fake information is fine, but instead of vilifying the tool used for creating it, punish the user who makes and spreads it. And not just the creator of the content, it is also the responsibility of the person who consumes and shares it on social media platforms to verify the content. In certain cases, the latter should be held more responsible for the spreading of misinformation, than the former for creating it for it only becomes news when it gets a platform. Blame the data “As an AI language model, I cannot…,” is one phrase that you would get for a lot of prompts you input on ChatGPT. Even ChatGPT realises that it is incapable of generating a lot of things, a lot of which can be misinformation. Or even if it does, it declares that it “may not be true”. OpenAI puts a clear disclaimer on its website – ChatGPT may produce inaccurate information about people, places, or facts. Does any publishing website put up such disclaimers? The company has recently provided even more control of users’ data to protect privacy. There have been cases where ChatGPT made false accusations on people, and is now facing lawsuits. In certain instances, it is also known for making up anonymous sources and making itself sound profound while doing so. You might also know a lot of people who do so. And anyway, it is trained on the information freely available on the internet. The same internet, which is filled with conspiracy theories, false information, and hateful language. OpenAI has done a great job of fine-tuning it to not spew garbage and be civil. The information that it is fed is hundreds and thousands of websites, including a lot of them that are filled with confidently written lies and manipulated facts. No one can claim that all information on the internet is true and trustable. Would be interesting to see Elon Musk’s chatbot fed on Twitter data. Same is the case for ChatGPT – you can choose to trust it and spread misinformation, even if OpenAI tells you not to. Internet, ChatGPT Caught in Cycle In the end, as Yann LeCun put it, it is just a text generator. A person can write the same misinformation on social media without using AI. As Altman said at the Senate that OpenAI already has placed models that can detect if text is generated from ChatGPT or not, the future looks safe for people concerned with AI spreading misinformation. On a concerning note, the internet is getting filled with a lot of content generated by ChatGPT and similar models. A lot of it is evident because people post content without even verifying or editing them. “As an AI language model” phrase is found on several websites that are polluting the internet. It’s a full cycle. ChatGPT is fed on internet data and now the internet is filled with articles written by AI. Now, with Google Bard being connected to the internet in real-time, and ChatGPT being connected to Bing for real-time internet access, the case might get worse — models being trained on the same data that they generated! This begs the question – Is the recent development of connecting chatbots to the internet really a good idea? When ChatGPT had a knowledge cut off of 2021, people initially trusted it, but with time somewhat the trust fell off. Now that Google and Microsoft are claiming that the models would improve with real-time internet access, users would start believing it as a source of information. This can possibly spread a lot of misinformation directly to the users – combination of real-time information with hallucinating LLM chatbots. Even then, the one to be blamed is the internet and its data, not ChatGPT or Bard. It’s the data generated by humans that is polluting ChatGPT responses.","excerpt":"It is not just ChatGPT’s fault that it generates misinformation, it’s also, and mostly the fault of content creators and media firms","categories":["AI Features"],"tags":["ai chatbots","ai hallucinations","AI Models","AI Social Media","AI Tool","Sam Altman"],"author_name":"Mohit Pandey","publish_date":"2023-05-25T16:00:00","publication_year":"2023","word_count":1019,"keywords":["ai hallucinations","ai chatbots","AI Social Media","AI Models","Sam Altman","ChatGPT","OpenAI","AI","Google Bard","chatbots","AWS","Go","RAG","Aim","AI Tool","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Google Bard","Aim","RAG","chatbots","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/you-are-to-be-blamed-for-chatgpts-flaws\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10111468,"title":"Indian States Vie with Each Other to Win Semiconductor Deal","content":"Electronics manufacturing company Kaynes Technology was set to invest INR3,750 crore to build an outsourced assembly and testing (OSAT) facility in Mysuru, Karnataka, creating 3,200 jobs. However, within a few months, Kaynes shifted location to Telangana and will build the OSAT facility at a relatively lower cost of INR 2800 crore. The Karnataka government was bombarded with criticism for letting Kaynes slip away from its hands to a neighbouring state. At the same time, the Telangana government was praised for convincing the homegrown electronics manufacturers to shift base. This incident underscores the fact that Indian states are currently engaged in a competition to draw leading semiconductor companies to their respective regions. Setting up a fabrication unit brings numerous economic benefits — it can spur local employment opportunities and bring economic growth for the region. It also leads to enhanced export potential and the stimulation of allied industries, contributing to overall economic development. State governments are enticing semiconductor companies with propositions like providing electricity and water, two important resources to run a fabrication unit, at subsidised rates. Numerous states are even offering additional incentives on top of those already declared by the central government. States like Gujarat, Karnataka, Tamil Nadu, Odisha, and Assam have come up with their semiconductor policies. For instance, in July 2023, the state government of Gujarat unveiled a dedicated policy which includes incentives and subsidies to encourage semiconductor production within the state. Gujarat is building a Semiconductor city Gujarat is in the process of creating a ‘Semicon city’ in Dholera. Encompassing approximately 10,000 acres within the Dholera Special Investment Region (SIR), the development of the Dholera Smart City is anticipated to play a pivotal role in positioning Gujarat as a national hub for semiconductor industries. The government is trying to attract not just fabrication companies, but also display fabrication units, silicon photonics, and semiconductor assembly, testing and packaging facilities to the state. Bhupendrabhai Patel’s government also plans a 75% subsidy on the purchase of the first 200 acre land and an additional 50% capital subsidy to companies deciding to set up camp in the state as per the Gujarat Semiconductor Policy 2022-27. The semiconductor projects will be supplied water at Rs. 12 per cubic metre for the initial five years, with a subsequent annual increment of 10% for the following five years. Moreover, the production unit qualifies for a Rs. 2 per unit power tariff subsidy for the initial ten years. Additionally, it will be exempted from paying electricity duty under the provisions of the Gujarat Electricity Duty Act, 1958. So far, US-based semiconductor company Micron Technology has announced its plan to set up an outsourced semiconductor assembly and testing (OSAT) plant at Sanand in Gujarat. However, Gujarat is not the only state competing to attract semiconductor companies. Moreover, at the 10th Vibrant Gujarat Global Summit, Tata Sons Chairman N Chandrasekaran revealed that Tata Group will build a semiconductor factory at Dholera. Karnataka banking on the existing semiconductor R&D hub Giving tough competition to Gujarat, Karnataka too, has identified 4 semiconductor clusters within the state. Interestingly, Karnataka was the first state to draft a semiconductor policy in 2010. In 2017, the state launched the Karnataka Electronics System Design & Manufacturing (EDSM)Policy 2017-2022. State IT minister Priyank Kharge has revealed that his government is already talking to a few companies. Moreover, the minister even said that his government is even willing to tweak incentive and labour policies to attract major players along with their supply chains. Karnataka would also bank upon the fact that many semiconductor companies already have their Research and Development (R&D) units in the state. Last year, AMD, an American multinational semiconductor company based in Santa Clara, California, announced its largest design centre in Bengaluru. Other companies such as Analog Devices, NXP Semiconductors, NVIDIA and Intel, among others have established their R&D centre in Karnataka. Interestingly, Texas Instruments has had a presence in Bengaluru since 1985. “We already have a fabless ecosystem and are slowly metamorphosing into a fab ecosystem. A lot of people are talking to us. Earlier, the talks were more speculative, and now there is a sense of urgency,” Kharge said in an interview. Previously, ISMC, a joint venture between Abu Dhabi-based Next Orbit Ventures and Israel’s Tower Semiconductor, proposed a 65-nanometer analog semiconductor fabrication unit in Mysuru’s Kochanahalli Industrial area. Now, reports suggest Krypton Solutions could invest INR 832 crore (USD 100 million) to set up a Printed Circuit Board (PCB) fabrication unit in Karnataka. Foxconn, along with Telangana and Tamil Nadu, are also considering the state to set up an OSAT unit in partnership with HCL. Other states are catching up Other states in Southern India, such as Tamil Nadu and Telangana, are also trying to lure semiconductor companies. At the Global Investors Meet (GIM), the Tamil Nadu government unveiled its ‘Semiconductor and Advanced Electronics Policy 2024’. With the policy, the TN government aims to foster a semiconductor design ecosystem by offering incentives and financial support and promoting collaboration between the industry and academia. Companies selected under the India Semiconductor Mission (ISM) will receive an additional 50% financial incentive from the TN government. Moreover, the youngest Indian state, Telangana, is also trying to attract semiconductor companies. Kaynes Technology has already started work on its OSAT facility in Kongara Kalan near Hyderabad. Other states like Maharashtra, Odisha and Assam are also trying to woo semiconductor companies. The Eknath Shinde-led Shiv Sena-BJP government in Maharashtra offered INR 39,000 crore incentives to Vedanta-Foxconn JV, which chose Gujarat, before disbanding. Assam, which has tabled a semiconductor policy, has managed to attract the Tata Group to set up an INR 40,000 crore semiconductor processing plant. In Odisha, RIR Power Electronics Ltd, a global semiconductor company, plans to invest INR 510 crore to set up a fabrication and packaging facility for SiC devices at Info Valley, Khurda.","excerpt":"State governments are luring semiconductor companies by offering subsidised electricity and water, crucial resources for operating a fabrication unit.","categories":["AI Features"],"tags":["Priyank Kharge","Semiconductor India"],"author_name":"Pritam Bordoloi","publish_date":"2024-01-25T17:20:32","publication_year":"2024","word_count":971,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","Priyank Kharge","RAG","Aim","Semiconductor India","GAN","llm_models:Bard","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","GAN","llm_models:Bard","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/indian-states-vie-with-each-other-to-win-semicon-deal\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085883,"title":"India’s R&D Prowess Not Enough to Become ‘China+1’ in Chip Game","content":"With a global emphasis on diversifying semiconductor supply chains and reducing reliance on a single country, the world turns its gaze towards India to lead the highly acclaimed ‘China plus one’ strategy. And efforts in that direction have been underway for a while now. The money poured into enabling a semiconductor ecosystem will help India in the long run. This would help create employment opportunities, reduce import dependency, and help India become a semiconductor export hub. While proposals for the installation of fabs received by the government (owing to the modified incentives scheme) still await approval, what has come as a shining light for India has been the foreign R&D investment over the years. Raining investments Despite the grim reality of private sector R&D in India, the country has managed to be a significant location for global semiconductor design companies, as many have established research and development centres here due to the abundance of skilled semiconductor design engineers, who constitute 20% of the global total, and a high number of design patents and intellectual property rights registered in India. In fact, many of the world’s major semiconductor and Wafer Fabrication Equipment (WFE) manufacturers have established significant research and development operations in India. American company Lam Research, which specialises in wafer fabrication equipment, recently installed its 2nd R&D facility in India, aiming to focus on hardware and software designs which will be incorporated into the next generation of DRAM, NAND, and logic devices. Lam Research India’s Krishnan Shrinivasan told DQ India, “The availability of a technically sophisticated workforce contributes significantly to product design and testing.” Having a local presence will enable engineers here to carry out the entire design process, including testing and validation of new hardware and software, on-site, greatly reducing the time to complete the design cycle, as it eliminates the need to send the designs to other locations for these processes. There is also the US-based Applied Materials, a leader in material engineering solutions, which pledged $50 million to set up an R&D facility in Bengaluru to support future product development and benefit the local supply chain. Likewise, Texas Instruments, which holds the largest market share for the analogue IC industry, has India as its largest R&D site outside its Dallas headquarters. Additionally, Micron Technology, the memory solutions provider, is continuously expanding its presence in India, growing to 3,500 employees in a little over three years. “The Micron India Research Center (MIRC) was built on the back of one of the fastest non-linear ramps in the Indian semiconductor industry. In three years, over 70% of Micron’s non-manufacturing operations have an India footprint, giving it a unique colocation advantage to execute cross-functional research programs. MIRC leverages this to build a research-oriented team for AI\/ML, data sciences & analytics, engineering CAD and system-level solutions,” Anand Ramamoorthy, managing director of Micron India told AIM. NXP semiconductors also inaugurated its R&D labs in India recently to collaborate with startups and focus on new areas of intellectual property and system-on-chip technology. Indian private sector needs a push Minister of State for Electronics and IT Rajeev Chandrasekhar, in a virtual address at IEEE MAPCON 2022, stated that the government is providing financial assistance for research and development efforts to help India transition from being a tech consumer and a provider of back-office support to becoming a major global tech producer. However, the incentives provided by the government to private industries for R&D are not aplenty. Zoho CEO Sridhar Vembu tweeted, “For India to build such companies, the private sector must invest heavily in R&D. There is no other way. The government should incentivise and persuade companies to invest in R&D. Industrial R&D is not the same as academic research.” Moreover, ahead of the Union Budget 2023-24, the Confederation of Indian Industry (CII) suggested that the Centre should bring private companies under the ambit of government funding in a ‘limited manner’ to help R&D in the country. It also proposed policy changes that would allow intellectual property to be used as collateral for bank loans. Last month, the government announced that it would soon launch ‘future labs’ at C-DAC to provide R&D capital for developing semiconductors and deep technology. Future Labs will encourage industry and academic collaboration and push for academic work that can directly contribute to large-scale manufacturing in industries. The road to being the ‘+1’ to China won’t be an easy one. AIM spoke to an official from C-DAC who said, “I think it [the question of whether India will benefit from China+1] depends on a number of factors, including the level of investment, the specific technologies and innovations that are developed, and the government’s ability to create a favourable business environment. However, it is certainly a step in the right direction and can be a major contributor to the growth of the country’s economy.”","excerpt":"While proposals for installing fabs still await government approval, foreign R&D investment shines bright","categories":["AI Features"],"tags":["India semiconductor mission"],"author_name":"Ayush Jain","publish_date":"2023-01-25T18:00:00","publication_year":"2023","word_count":802,"keywords":["data science","Go","API","AI","innovation","ML","India semiconductor mission","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","R","Go","API","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/indias-rd-prowess-not-enough-to-become-china1-in-chip-game\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101603,"title":"Zoho’s Bid to Rival GitHub Copilot","content":"Zoho is building its own version of GitHub Copilot alternative. While speaking to the media last week in Bengaluru, Sridhar Vembu, co-founder and CEO of Zoho Corporation, announced the company’s plans to work on ‘Programmer Productivity’ – a platform that will focus on code generation. Prioritising accurate results, Vembu is personally involved in the project and aims to create a product that is unlike ChatGPT, which Vembu believes generates hallucinatory answers 20% of the time. “In business, you cannot make facts up,” said Vembu, while explaining Zoho’s new coding product that is in development. He believes that models that generate incorrect answers limit business usage, “If you generate code out of it, that limits the usage of the code too.” With the new coding product, Zoho looks to eliminate inaccuracy. Zoho Enters The Coding Ring In an exclusive interaction with AIM last week, Vembu highlighted AI neural network’s capability to regurgitate memorised information, which can also lead to hallucinated output. Overcoming faulty output is Vembu’s biggest priority in the upcoming product, as he believes that fixing bugs is where a developer spends most of their time. Furthermore, security is another concern area that will be addressed as well. Concoction of Training Methods While the how part of it is not fully detailed, Vembu confirmed that there would be a combination of training methods. “I am not taking the conventional, ‘training a billion parameter’ approach, although that will go in parallel. We are investing in compiler technology,” said Vembu. A compiler translates high-level programming languages into machine code or lower-level codes for executing into a computer’s hardware. As explained by Vembu, any form of java code, c-code, etc. goes through a compiler. He believes that this technology has huge potential but it is not fully explored as many companies do not invest in the same. “Microsoft and Google have invested in it, but Salesforce will not have.” Vembu also said that Zoho is investing in this technology owing to security. “If we ensure correctness, the compiler can ensure correctness, and we can also ensure that you are secure.” There are 20 people working with Vembu on the R&D of this project, and he has personally filed 10 patents in the last few years. An Unconventional Move To Take on Others Pioneering as a SaaS company providing a number of CRM-based products and services, the company is probably moving to an approach that integrates the best of two worlds. With the integration of ChatGPT in their product suite, Zoho has been embracing AI extensively. The company recently announced additional features such as Cliq Rooms, to their existing AI-enabled Zoho Cliq platform. However, the foray into coding assistance for programmers is an unconventional development for the company. With major players already in the market, such as Replit, GitHub Copilot, and the recent Code Interpreter from ChatGPT, Zoho’s entry in a similar segment is probably a push to utilise their current customer-enterprise base of 100 million users. Microsoft’s GitHub Copilot is being used by a million developers across 20,000 organisations today and is considered the world’s most widely adopted AI developer tool. The platform is powered by generative AI developed by GitHub, OpenAI and Microsoft. Closely competing with each other is Replit. The company recently made their suite of AI products called Replit AI, accessible to all. The coding assistance feature will be open to over 23 million developers. There are also other players such as AmazonCode Whisperer in the race. Relatively new to the coding assistance game is ChatGPT Code Interpreter. However, there is no information available about exclusive developer communities for the product, similar to the communities established by GitHub, Replit and other platforms. Closer to its home turf, Zoho’s competitor Salesforce has also entered the same segment. Salesforce, another cloud-based CRM software company, recently announced its coding copilot platform ‘Anypoint Code Builder‘. The integrated development environment (IDE) enables developers to create APIs, simplifies and enables coding for developers. It is currently in beta and will be available in the second quarter of 2024. While Zoho may be late in the game, the different training approaches adopted by the company for building the product will probably determine its success. “Correct by design is where the action is and that’s where we are investing in,” said Vembu about the programmer productivity product goal.","excerpt":"Zoho is developing a new ‘programmer productivity’ tool designed to generate accurate codes, and aims to surpass other platforms’ issues with hallucinations and factual errors","categories":["AI Trends"],"tags":[],"author_name":"Vandana Nair","publish_date":"2023-10-18T11:00:00","publication_year":"2023","word_count":720,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","neural network","Ray","Aim","generative AI","R"],"extracted_tech_keywords":["AI","neural network","generative AI","ChatGPT","OpenAI","Aim","Ray","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/zohos-bid-to-rival-github-copilot\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10092623,"title":"HuggingChat vs ChatGPT: The Chat Battle has Now Begun","content":"ChatGPT became an internet sensation this year as it seemed like an artefact pulled from a science fiction. The OpenAI product left us in awe but as impressive as it was, the model lacked something critical — openness. The closed source model didn’t do much for the developer’s community. To the rescue, Hugging Face, the open-source AI platform, launched HuggingChat, which is being touted as an open-source ChatGPT alternative. The release of HuggingChat provides various functionalities and integrations catering to both developers and users alike, and offers stiff competition to ChatGPT. Inspiration or Imitation? The interface of HuggingChat appears to be inspired by ChatGPT. The sleek blue screen has conversation points and a history of conversations on the left similar to ChatGPT’s page. ChatGPT does not have access to information post September 2021, and therefore cannot provide real-time information. Hugging Face’s chatbot seems to be on the similar lines here. It is worth noting that at certain instances HuggingChat provides updated answers but they can be incorrect. For instance, we asked the chatbot ‘Who won IPL 2022?’ and no part of the answer was factually correct. (Gujarat Titans, playing their first tournament, won the match and the title by defeating Rajasthan Royals) The Differences One of the key differences between the two models is the data set on which they were trained. HuggingChat was trained with the OpenAssistant Conversations Dataset (OASST1), For now, the HuggingFace model runs on OpenAssistant’s latest LLaMA based model but the long term plan is to expose all good-quality chat models from the Hub documentation. TheHugging Face dataset contains data collected up to April 12, 2023. The dataset is the result of a worldwide crowdsourcing effort by over 13,000 volunteers and includes 161,443 messages distributed across 66,497 conversation trees in 35 different languages, annotated with 461,292 quality ratings. As Meta’s LLaMA is bound by industrial licences it is not possible to directly distribute LLaMa-based models. Instead Open Assistant provided XOR weights for the OA models. On the other hand, ChatGPT’s data set is currently unavailable. Another parameter to consider is that the cost of using ChatGPT depends on which large language model (LLM) is being used — GPT-3.5 or GPT-4. The paid subscription under ChatGPT-Plus uses the latest GPT-4 hence provides better answers than the free version based on GPT-3.5. The highly anticipated GPT-4 was integrated into ChatGPT in March 2023. Currently, the paid version costs $20 on a monthly basis. In contrast, HuggingChat is open-source and free to use. Read more: 7 Ways Developers are Harnessing Meta’s LLaMA Both the chatbots have a different writing style. ChatGPT provides structured and well-formatted answers but does not take a stance on any subject. On the contrary, Hugging Face answers in a much more personalised manner and tends to address itself in the first person. But HuggingChat does not understand context that well. In terms of coding, HuggingChat gives the code at once, whereas ChatGPT provides the instruction for the free on GPT-3.5 model and provides a code with in-depth instructions to follow for the paid GPT-4 model. In conclusion, while HuggingChat may be new and promising, ChatGPT is currently the superior choice due to its established reputation and range of features. The HF model is not yet close to the levels of ChatGPT, but it is the need of the hour to avoid a monopoly. Nevertheless, it will be interesting to see how HuggingChat develops and whether it can challenge ChatGPT’s dominance in the chatbot market. Another option that developers can look at is Databricks’ Dolly 2.0.","excerpt":"The differences and the similarities","categories":["Deep Tech"],"tags":["ChatGPT","GPT4","gpt5","OpenAI"],"author_name":"Tasmia Ansari","publish_date":"2023-05-02T16:33:31","publication_year":"2023","word_count":590,"keywords":["Go","ChatGPT","Hugging Face","OpenAI","AI","chatbots","GPT","Databricks","gpt5","llm_models:GPT","R","GPT4"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Hugging Face","chatbots","Databricks","R","Go","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/huggingchat-vs-chatgpt-the-chat-battle-has-now-begun\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10100463,"title":"Oracle Announces AI\/ML Features To Its MySQL HeatWave: A Pioneering Shift in Database Innovation with AI and ML Superpowers","content":"In a landmark announcement today, Oracle Corp. introduced a series of significant enhancements to its MySQL HeatWave database platform. These enhancements offer a robust array of features that span AL\/ML, data optimization, and query acceleration for data management and analysis. One of the highlights is the introduction of the Vector Store for more precise insights, utilising their proprietary data of LLMs. The store accepts documents in various formats, storing them as embeddings generated through an encoder model. These embeddings facilitate accurate searches, improving the contextual relevance of responses when interacting with MySQL HeatWave. Coupled with the Vector Store, the platform now boasts generative AI capabilities, for users to interact with MySQL HeatWave in natural language. This advancement enhances document searches within the HeatWave Lakehouse. “Vector stores and generative AI bring the power of LLMs to customers, providing them with an intuitive way to interact with data in their enterprise and get the accurate answers that they need for their business,” said Edward Screven, chief corporate architect, Oracle. Oracle has reinforced HeatWave’s ML capabilities with a fully automated pipeline for model training. A key advantage is the ability to conduct ML tasks without the need to migrate data to external services. Customers can securely apply ML training, inference, and explanation directly within HeatWave. MySQL Autopilot has undergone significant improvements, too. It now includes features such as Autopilot indexing, which automates index creation for optimal OLTP workload performance. Auto compression assists in selecting the best compression algorithm for each column, improving load and query performance while reducing costs. This release introduces JavaScript support, enabling developers to create stored procedures and functions in JavaScript. Notably, data remains within the database, eliminating the need for data transfers to the client, and code execution benefits from Just-In-Time (JIT) compilation in the GraalVM runtime. Developers and database administrators can use HeatWave for real-time analytics on JSON documents stored within the MySQL database, achieving significant query acceleration. Furthermore, the platform now supports new analytic operators, including CUBE, Hyper Log Log, Qualify, and Table sample, facilitating migration of diverse workloads to HeatWave.","excerpt":"Oracle’s MySQL HeatWave: AI, ML, JavaScript, and enhanced analytics – the future of data management","categories":["AI News"],"tags":["MySQL","Oracle","Oracle SQL"],"author_name":"Pranav Kashyap","publish_date":"2023-09-21T16:51:33","publication_year":"2023","word_count":344,"keywords":["Oracle SQL","Go","AI","ML","JavaScript","Oracle","Ray","analytics","generative AI","SQL","MySQL","R","Java"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Ray","R","SQL","JavaScript","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/oracle-announces-ai-ml-features-to-its-mysql-heatwave-a-pioneering-shift-in-database-innovation-with-ai-and-ml-superpowers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":9856,"title":"Snapdeal acquires TargetingMantra, a predictive marketing analytics firm","content":"India’s“Dil ki Deal” company and one of the largest E-commerce platforms, Snapdeal has acquired a boutique technology firm, ‘TargetingMantra’ to further to integrate machine learning based solutions on its platform. This acquisition aims to develop Snapdeal’s ‘customer experience engine’ into a personalised shopping experience for customers. So, far, TargetingMantra has done path breaking work in the field of personalizing shopping experience for customers on e-commerce platforms such as Amazon. Saurabh Nangia and Rahul Singh, Targeting Mantra Founders Founded by Saurabh Nangia and Rahul Singh in March 2013, TargetingMantra has offices in Palo Alto and Gurgaon. The company provides a unified platform to manage customer life cycle through personalization, targeting and big data analytics. Further, it specialises in building products for enhancing customer buying journey and increasing conversion rates through intuitive product discovery, recommendations and channel selection. Rohit Bansal, Co-Founder, Snapdeal Speaking about this acquisition, Rohit Bansal, Co-Founder, Snapdeal said, “Personalisation is a key piece which helps consumers discover and transact in a fast, frictionless and intuitive manner. The TargetingMantra team comes with valuable experience in driving superior customer experience through machine learning. We are thrilled to have them join us on our journey towards achieving 20 million daily transacting users by 2020.” Snapdeal’s vision is to create India’s most reliable and frictionless commerce ecosystem that creates life-changing experiences for buyers and sellers. In this regards, Rajiv Mangla, CTO, Snapdeal added that, “they are making conscious efforts towards building new tech products that will re-define the future of online shopping and they are looking forward to co-creating new benchmarks in customer experience with the TargetingMantra team.” Commenting on the acquisition, Saurabh Nangia, Co- Founder, TargetingMantra commented that, “Given Snapdeal’s customer centric approach to technology, TargetingMantra can add immense value to shaping shopping experiences on the platform.” While, Rahul Singh, Co-founder, TargetingMantra added, “We are looking forward to rolling out our solutions to Snapdeal’s huge customer base.” According to social media analytics for this year, Snapdeal is neck to neck with Flipkart and Amazon. We remain excited to see how this acquisition will affect the e-commerce market, the customer base and most importantly, the customer experience, which will ultimately, reflect on the company and their social media.","excerpt":"India’s“Dil ki Deal” company and one of the largest E-commerce platforms, Snapdeal has acquired a boutique technology firm, ‘TargetingMantra’ to further to integrate machine learning based solutions on its platform. This acquisition aims to develop Snapdeal’s ‘customer experience engine’ into a personalised shopping experience for customers. So, far, TargetingMantra has done path breaking work in […]","categories":["AI News"],"tags":["ecommerce analytics"],"author_name":"Apoorva Verma","publish_date":"2016-05-09T08:05:53","publication_year":"2016","word_count":365,"keywords":["big data","API","machine learning","ecommerce analytics","programming_languages:R","AI","Aim","analytics","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","R","API","big data","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/snapdeal-acquires-targetingmantra-predictive-marketing-technology-firm\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10106961,"title":"10 Events that Transformed AI","content":"In 2023, AI captivated the world with its heady progress, be it in large language models, chatbots or protein folding. As the year comes to a close, AIM looks back at ten significant developments in AI over as many months. This year has truly been unlike any other for AI. Here are the things that made it special: ChatGPT Gained 100 Million Weekly Users Digital Trends\/ Analytics India Magazine ChatGPT made a debut in November 2022 and quickly gained attention from tech leaders and the public. In January 2023, the chatbot had a monthly active user base of 100 million. According to OpenAI CEO Sam Altman, it now boasts an astounding 100 million weekly users. Google, feeling concerned that AI could render its search business useless, responded with its AI chatbot Bard, while Microsoft launched Bing Chat. GPT-4 makes a splash Digital Trends\/ Analytics India Magazine Initially released in March 2023, OpenAI’s GPT-4 LLM is now accessible to the general public through OpenAI’s API and the premium chatbot application ChatGPT Plus. GPT-4 enhances creativity, visual input, and multimodal context, allowing users to collaborate on artistic tasks like writing, music, and screenplays, surpassing previous models. Currently behind OpenAI’s ChatGPT Plus paywall, GPT-4 has significantly impacted AI, with Google’s Gemini and others claiming to beat it at most tests a year after its launch, indicating a benchmark that the OpenAI model has set. It was only in May that ChatGPT’s browsing capabilities were expanded when the Browsing through Bing plugin was announced at the Microsoft Build developer conference. It was a slow rollout until September when it became available to all ChatGPT Plus users. LLaMA Leak In March, Meta’s latest family of large language models, LLaMA, got leaked along with its weights, on 4Chan’s technology board and was available to download through torrents. This accidental unveiling of Meta’s LLM changed the whole open source game as it gave an enormous potent tool in the hands of the open source community in the AI space. AI-generated images Pablo Xavier In March 2023, a picture of Pope Francis wearing a white puffer jacket, created by Pablo Xavier using AI image generator Midjourney, went viral,  demonstrating the power of AI in deceiving humans. This, along with others like Trump’s arrest, underscores the need for increased media literacy about the increasing prevalence of AI-generated images in search results. A petition begins to sound the alarm Apple co-founder Steve Wozniak In March 2023, tech executives, including Elon Musk and Steve Wozniak, wrote an open letter urging AI labs to pause training for at least six months due to potential risks such as loss of civilization control, human annihilation, and job destruction. The letter also included academics and researchers. The letter aims to provide a clearer understanding of the potential risks associated with AI. Windows gets a new Copilot Microsoft Soon after launching Bing Chat at the start of the year, Microsoft introduced Copilot in February 2023,  a far more widespread application of AI in its products, initially for Microsoft 365 Copilot. Microsoft integrated Copilot into Word, Teams, and Windows 11, automating tasks like image creation and meeting summarization, demonstrating AI commitment and setting a precedent for Apple. Academics grapples with AI Unsplash In May 2023, a professor at Texas A&M University-Commerce failed an entire class due to students using ChatGPT to write papers, despite no proof. This ignorance led to serious consequences for students. Dr Jared Mumm copied and pasted students’ papers into ChatGPT, asking if it could generate the text. ChatGPT answered affirmatively, but it cannot detect AI plagiarism. Reddit users took Dr Mumm’s letter accusing students of cheating and pasted it into ChatGPT, resulting in AI hallucinating and human confusion surrounding its abilities. Hollywood vs AI Paul Deetman\/Pexels Artificial intelligence is causing global anxiety, with reports suggesting it could eliminate up to 300 million jobs if left uncontrolled. Hollywood authors went on strike over the use of AI in filmmaking, in September, 2023, but won concessions from studios, including limiting AI content use for training. AI development may continue to impact the industry. Sam’s sacking saga OpenAI The OpenAI board fired CEO Sam Altman in November, leading to employee resignations and Microsoft offering jobs to Altman and other OpenAI employees, causing the company to almost collapse. Soon, Altman was reinstated and the company got fresh board members. The internet debated if Altman had discovered ethical concerns about AI development, if Project Q* was about to achieve AGI, or if he was just a bad boss. We may never know the whole truth. But nothing else encapsulated the drama, hysteria, fascination, and conspiracy theories as much as this one did in 2023. The rise of OpenAI alternatives After OpenAI, there was a sudden spike in the number of platforms offering foundational models and proprietary chatbots. Cohere AI: Cohere is a leading AI platform for enterprise, offering ease-of-use, accessibility, and data privacy. It’s cloud-agnostic, accessible through API, and can be deployed on VPC or on-site. Founded by Google Brain alumni, Cohere aims to transform enterprises and their products with AI. Cohere raised $270 million in a funding round led by Microsoft-backed OpenAI, valued at $2.2 billion in June, 2023. Anthropic: Anthropic is an AI safety and research company working to build reliable, interpretable, and steerable AI systems. In October, 2023 Google committed to investing up to $2 billion in Anthropic, bolstering the competition among startups striving for significant technological breakthroughs. Mistral AI: Built by a world-class team in Europe, targeting the global market, Mistral AI was founded in April, 2023. In November, 2023, it raised $385 million, in a significant investment in online chatbot technology, valued at around $2 billion, following a sevenfold increase in value in six months. Perplexity: Perplexity is a chatbot-style search engine that allows users to ask questions in natural language. Founded in August 2022, Perplexity, has raised $500 million, a significant increase from its initial $150 million valuation. It is also in discussions to raise $50 million.","excerpt":"As the year comes to a close, AIM looks back at ten significant developments in AI over as many months. This year has truly been unlike any other for AI. Here are the things that made it special:","categories":["AI Trends"],"tags":["AI Tool"],"author_name":"Arya Vishwakarma","publish_date":"2023-12-26T16:30:00","publication_year":"2023","word_count":995,"keywords":["Anthropic","ChatGPT","Go","artificial intelligence","OpenAI","AI","chatbots","Aim","analytics","AI Tool","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","ChatGPT","OpenAI","Anthropic","Aim","chatbots","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/2023-in-review-10-events-that-transformed-ai-2\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":47558,"title":"How Startups Can Leverage Augmented Data Management To Drive Business","content":"Ubiquitous data has empowered startups to compete against blue-chip companies even without huge financial backing, as they are now able to blaze their own trail and build superior products. This is evident as of late many big organisations are acquiring startups due to the unique value proposition of their products. Although, in the current landscape, they also possess similar challenges to what blue-chip companies witness. One such problem is data management for streamlining workflows to improve its efficiency and productivity. Owing to this, various firms are embracing augmented data management for enhancing the outcomes of their data-driven products. What is augmented data management? Augmented data management is a process in which data is automatically refined by utilising machine learning and artificial intelligence techniques. It is no more a secret that data scientists spend around 80% of their time in data wrangling as often data are collected from different sources; information from various origins are usually unstructured and comes in several formats. Thus, implementing augmented data management assists firms in organising and maintaining data quality through cutting-edge technologies, resulting in reducing data cleaning activities of data scientists, and in turn, increase the productivity of organisations. Best Practices For Augmented Data Management Apart from the aforementioned advantages, augmented data management can facilitate data for real-time analytics. Desired quality and tidiness of data is essential for utilising it on the go and make informed business decisions instantly. Through augmented data management processes, startups are harnessing data across several departments and improving collaboration for accomplishing different tasks. This will enable startups to better manage their day-to-day activities by making proactive decisions within their departments. Such immediate use of data also eliminates the data silos, thereby, decreasing the cost of business operations. Startups, where there is a dearth of resources, it is paramount to deploy such processes that can be a helping hand for their data scientist for deriving insights into data. Other benefits include: Data scientists regularly strip or evaluate outliers for its genuineness with regards to other data points, which now can be carried out with augmented data management practices by deploying Ml and AI modules. Data scientists often configure database and query desired information that can be processed to solve business challenges. Long story short, they often interact with data management to manage and extract quality data. Therefore, embracing solutions that can auto-configure database will eliminate the necessity to manage silos. Augmented data management also has the potential to maintain data integrity by ensuring accuracy, completeness, consistency, and more, as a result, data scientists can focus on analysing instead of checking and cleaning data. Biased data can afflict any organisations and can lead to bankruptcy, as making decisions based on biased data can negatively impact firms. Ensuring genuine data is another facet of augmented data management such that companies can rely on it for their business growth. By deploying augmented data processes, companies can leverage organised data through dashboards like Tableau, PowerBI, and more, to analyse and create reports instantly. Apart from data scientist team, a well-structured data can also be utilised by management and other departments for decision making just by knowing the basic of dashboards. Outlook Startups need to revamp data management processes such that they can automate the process of data circulation and optimise their course of further analysis on data. With the right tools for data management, startups can enhance their products by purging intricacies pertaining to data. Decreasing complexities are the key for any business to strive and achieve business objectives, and startups are no different. Therefore, deploying augmented data management practices are the way forward to stay competitive and give blue-chip organisations a run for their money.","excerpt":"Ubiquitous data has empowered startups to compete against blue-chip companies even without huge financial backing, as they are now able to blaze their own trail and build superior products. This is evident as of late many big organisations are acquiring startups due to the unique value proposition of their products.  Although, in the current landscape, […]","categories":["AI Startups"],"tags":["Data Management","data storage","data warehouse"],"author_name":"Rohit Yadav","publish_date":"2019-10-14T10:00:15","publication_year":"2019","word_count":611,"keywords":["Go","artificial intelligence","data storage","machine learning","AI","ML","RAG","analytics","real-time analytics","data quality","Data Management","R","data warehouse"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","R","Go","real-time analytics","data quality"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-startups-can-leverage-augmented-data-management-to-drive-business\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10078185,"title":"Shutterstock vs Getty: Two Different Approaches to Tackling the AI Threat","content":"On October 25, Shutterstock announced its collaboration with Open AI, the AI-powered image generation platform, to expand its database and allow users to create images for prompts that aren’t already a part of its library. In integrating Dall-E 2 to their platform, Shutterstock aims to leverage upon the growing popularity of AI art. The announcement came merely a month after reports claimed that Shutterstock was removing AI-generated images from their platform over copyright concerns. Shutterstock versus Getty Images To facilitate fair use of the licensed stock images, Shutterstock has revealed that it will compensate every artist whose work has been used as training data for AI-generated images. The revenue share model also includes giving artists their fair share, in form of royalties, each time their art is used. However, there is no information yet over the precise nature of this model. The company has also claimed that it will launch a contribution fund to compensate these artists whose works are to be used as AI datasets. The launch of the contribution fund received criticism online for treating artists only as a means to an end and not as the rightful owner of the images. It came to be seen as ‘token representation’ for the company to avoid scrutiny over their image use policy. See, for example, this tweet by John Robb, which is a measure of this sentiment: \"launching a fund to reimburse Shutterstock creators whose images are sold to train AI art models\"Ad hoc solutions (as in, alms for the poor) for AI appropriation instead of #dataownership https:\/\/t.co\/S0sr9M3zjI— John Robb (@johnrobb) October 25, 2022 Paris Marx, host of the podcast ‘Tech Won’t Save Us’, also wrote along the same lines: Shutterstock will sell AI art, cutting out artists and photographers, but don’t worry: it also created a “Contributor Fund” for PR purposes to throw a few pennies in their direction. https:\/\/t.co\/ArvWcuKs2A— Paris Marx (@parismarx) October 25, 2022 Few weeks ago, Open AI claimed that Dall-E 2 is able to generate two million images in a day. To put this in perspective, Shutterstock has a catalogue of 415 million images built through decades, whereas Getty has about 80 million images. The AI image generator will only need eight months to produce an output that equals Shutterstock and Getty combined. But, now with Shutterstock’s partnership with Open AI, the disparity between the stock image site and AI-generative platform seems null. In a recent interview, Getty Images CEO Craig Peters, expressed apprehension towards organisations moving to sell AI images bypassing still unanswered questions on copyright: “I think we’re watching some organisations and individuals and companies being reckless.” Peters further added that, “I think that’s dangerous. I don’t think it’s responsible. I think it could be illegal.” However, it was also reported that Getty Images has partnered with Israeli firm ‘Bria’ to assimilate AI-based image editing tools on its platform where users could transform images according to their specific needs. Who will win? These stock image giants have taken completely opposite paths in response to the current boom of AI-produced content. While Shutterstock embraced the threat caused by the AI-generative models, Getty Images has remained fairly cautious about it—sticking to their stance on banning all AI images from their site, citing copyright concerns. Instead of having an AI tool create an image from scratch, Getty’s partnership with Bria shows that it plans to give its users the opportunity to edit the existing stock images to their liking. In the context of the platform’s adoption of the AI image editing tool, Getty’s CEO said: “To stay relevant in the world of AI, an image can no longer remain static but become dynamic, and now the user gets the image he or she wants”. Many have called Getty’s refrain from including AI-generated images in its library as acting “against technology”, while some find its “old-fashioned” approach sensible for the platform. As of now, it is difficult to say which of the two stock image sites are going to benefit the most from their respective AI integration. Shutterstock has claimed that their use of generation technology is grounded in ethical practices, but haven’t yet been transparent on the compensation for artists. In addition, considering the breadth of its stock images used as the training data for Dall-E 2 since 2021 along with the monetary incentives stock photographers will receive from Shutterstock—it will not be surprising if it performs exceedingly well in comparison to other free AI image generative tools. Getty’s position on AI images, on the other hand, has been received with immense scepticism. A general belief among many is that Getty will eventually come around. GitHub’s Chriscreama Warren, for instance, writes, “I fully expect Getty will eventually buy or make its own AI thing trained on its own photography or stock stuff (or a subset anyway) that they can licence and sell. It’s just too lucrative to ignore”. However, in avoiding the persistent copyright threat associated with AI-generative images, and consequently, the revenue model surrounding it, Getty may place themselves as a more economical option to stock images.","excerpt":"Shutterstock and Getty Images have taken contradictory positions to the copyright concerns looming around AI image generators. Their recent moves however show that AI integration is central to the future of both.","categories":["Deep Tech"],"tags":["AI Tool","copyright infringement","DALL.E","Shutterstock"],"author_name":"Ayush Jain","publish_date":"2022-10-28T11:00:00","publication_year":"2022","word_count":843,"keywords":["Go","DALL-E","Shutterstock","TPU","AI","Git","RAG","Aim","GAN","AI Tool","GitHub","R","copyright infringement","DALL.E"],"extracted_tech_keywords":["AI","Aim","RAG","TPU","R","Go","Git","GitHub","DALL-E","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/shutterstock-vs-getty-two-different-approaches-to-tackling-the-ai-threat\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10046462,"title":"New Weekend Hackathon For Data Scientists: Soccer Fever Challenge","content":"MachineHack is back again this weekend with yet another exciting hackathon for the data scientists’ community. This time the hackathon is dedicated to the passion and fervour which a sport creates. The challenge is to predict the outcome of a game based on some important predictor variables. This challenge is a part of MachineHack Weekend Hackathon Edition #2 — The Last Hacker Standing, where we pose unique problem statements every week, from 30 July to 9 Sept 2021. PARTICIPATE & STAND A CHANCE TO WIN FREE PASSES TO THE DLDC 2021!!! Problem Statement and Description Soccer, aka football, is currently the most popular game in the world. As Maradona once said, “football isn’t a game, nor a sport; it’s a religion.” If a group of people can stop time and make people watch them in awe and reverence, we all can understand its impact on everybody’s lives. Also, it has its own simplicity, where anybody can play soccer — all it requires are four poles, a ground and a ball. Nelson Mandela very effectively used football as the unifying factor when he was elected as the President of South Africa post the apartheid era. This is because the sport has the ability to cut across all discriminating factors. As a matter of fact, an entire ecosystem revolves around this beautiful sport — starting from clubs and merchandise to various football clubs and fan clubs. The amount of revenue outcomes involved in this game is just phenomenal, impacting millions of people who depend on it for their livelihood and recreation. In this hackathon, we challenge the MachineHack community to predict the outcome of a game based on important predictors. The hackathon will start on 20 Aug 2021 at 8:00 PM (IST) Click here to participate. Overview We live in ambiguity and always need some information to make a decision. Typically, decisions are made based on possible outcomes — win, loss, pass or fail, etc. This week’s problem statement is a classic study for decision-making and understanding the odds stacked against a particular situation. The participants need to create an ML model that can predict the outcome of a game based on unprocessed raw data. MachineHack has created a training dataset of 7443 rows with 21 columns, including ‘Outcome’ as the target variable. On the other hand, the dataset for testing consists of 4008 rows with 20 columns. The prerequisite skills required to participate in this hackathon include binary classification and optimising Log Loss. Submission Guidelines The participants must submit a .csv\/.xlsx file with exactly 4008 rows and 1 column, including the ‘Outcome’. The submission will return an ‘Invalid Score’ in case of extra columns or rows. Scikit-learn models support the predict() method to generate the predicted values. The submission limit for this hackathon is one account per participant. Click here to participate. Evaluation Criteria The evaluation of the hackathon will be done using the Log Loss metric. The hackathon will also support private and public leaderboards. While the public leaderboard will be evaluated on 30% of the test data, the private leaderboard will be made available at the end of the hackathon and assessed on 100% of the test data. The final score will be based on the ‘Best Score’ on the public leaderboard. The hackathon will end on 26 Aug 2021 at 6:00 PM (IST). The top three winners will get free passes to the Deep Learning DevCon 2021 (DLDC), scheduled to be held on 23-24 Sept 2021. In addition, the winners will also get a chance to improve their Global Leaderboard Rankings and become the ultimate MachineHack Grand Master. Click here to participate. Dataset Description: Train.csv — 7443 rows x 21 columnsTest.csv — 4008 rows x 20 columns Evaluation Metric: Log Loss Skills Binary ClassificationOptimising Log Loss Start Date: 20 Aug 2021 at 8:00 PM (IST) End Date: 26 Aug 2021 at 6:00 PM (IST) Click here to participate in this hackathon.","excerpt":"MachineHack is back again this weekend with yet another exciting hackathon for the data scientists’ community. This time the hackathon is dedicated to the passion and fervour which a sport creates. The challenge is to predict the outcome of a game based on some important predictor variables. This challenge is a part of MachineHack Weekend […]","categories":["Deep Tech"],"tags":["Hackathon","Hackathons","hackathons in India","Hackathons India","Machinehack","Machinehack Hackathon","machinehack platform","MachineHack Weekend hackathon","Weekend Hackathon","Weekend hackathon for data scientists"],"author_name":"AIM Media House","publish_date":"2021-08-20T18:00:00","publication_year":"2021","word_count":653,"keywords":["scikit-learn","Weekend Hackathon","programming_languages:R","Machinehack","hackathons in India","MachineHack Weekend hackathon","Weekend hackathon for data scientists","AI","Hackathons","ML","R","Hackathon","deep learning","Machinehack Hackathon","machinehack platform","Hackathons India"],"extracted_tech_keywords":["AI","ML","deep learning","scikit-learn","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/new-weekend-hackathon-for-data-scientists-soccer-fever-challenge\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10059445,"title":"How the partnership with OpenAI has cemented Microsoft’s position in Supercomputers","content":"In November 2021, Microsoft Azure claimed five spots in the TOP 500 list of the world’s most powerful supercomputers. Among these five, the Voyager-EUS2 became the only new entrant to the Top 10 of the list. This new supercomputer achieves 30.05 Pflop\/s and is based on an AMD EPYC processor working with an NVIDIA A100 GPU with 80 GB memory – to which Microsoft credited its success. Microsoft has been very vocal about its efforts in building public AI supercomputers that organisations can leverage to train their models. For example, companies like Meta (formerly known as Facebook) and Nuance (acquired by Microsoft) have been using the former’s supercomputers for research. One of the important steps in Microsoft’s journey to building the fastest and most advanced supercomputer is its collaboration with the non-profit AI laboratory OpenAI. Last year, Microsoft developed a supercomputer for OpenAI. It is a single system with over 285,000 CPU cores, 10,000 GPUs and 400 gigabits per second of network connectivity for each GPU server. This supercomputer is hosted in Azure and is supported by modern cloud infrastructure with access to all Azure services, rapid deployment, and robust cloud infrastructure. Microsoft and supercomputers In 2016, speaking at an event in Dublin, Microsoft CEO Satya Nadella spoke at length about how the company’s cloud computing offering underpins a new wave of applications to be used for AI technologies. “It is always the next-generation applications that have driven infrastructure and when we look at this current generation of applications that people are building, the thing that is going to define these applications, that characterises these applications, is machine learning and artificial intelligence. Therefore, we are building out Azure as the first AI supercomputer,” he had then said. Microsoft has also claimed that Azure will democratise AI, further adding that soon there will be no restrictions on who can integrate AI functionalities in their business. The big Microsoft-OpenAI collaboration In 2019, Microsoft announced that it would be investing a whopping $1 billion in OpenAI. This collaboration is steered to develop new technologies for Microsoft Azure and extend large-scale AI capabilities to achieve artificial general intelligence. OpenAI will be licensing some of its intellectual property to Microsoft to commercialise and sell to its partners. A good example of this is OpenAI giving exclusive license of GPT-3 to Microsoft. As per this partnership, OpenAI’s next-generation computing hardware would be trained and run on Microsoft Azure. Under this arrangement, Microsoft built the supercomputer for OpenAI that was run on Azure. Further, in addition to offering language model GPT-3 and other future models via OpenAI API, the AI lab also agreed to license GPT-3 to Microsoft. While OpenAI clarified that this deal would have no impact on the continued access to the language model through the API, the users could build their applications. Soon after this announcement was made, Microsoft introduced its first features in GPT-3 powered customer product in May 2021. The tech giant announced that GPT-3 would be integrated into Microsoft Power Apps, a low code app development platform. Microsoft said that this platform, which runs on Azure and is powered by Azure Machine Learning, can solve real-world business problems on an enterprise scale. In November 2021, at the annual Ignite conference, Microsoft unveiled Azure OpenAI service. This new service allows invite-only access to OpenAI’s API through the Azure platform. The tech giant also announced that users will be able to leverage new tools to determine if the outputs given by the model are appropriate for their businesses. Wrapping up Some of other leading companies that are building supercomputers include names like Meta (formerly Facebook) and IBM. However, having a leading AI lab as a partner gives Microsoft a major advantage. It needs to be seen how this partnership evolves in the future and what other innovations emerge from it.","excerpt":"The collaboration with OpenAI is steered to develop new technologies for Microsoft Azure and extend large-scale AI capabilities to achieve artificial general intelligence.","categories":["Global Tech"],"tags":["Microsoft Azure"],"author_name":"Shraddha Goled","publish_date":"2022-01-31T11:00:00","publication_year":"2022","word_count":637,"keywords":["machine learning","artificial intelligence","OpenAI","AI","cloud computing","TPU","R","RAG","Aim","Microsoft Azure","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","OpenAI","Aim","RAG","cloud computing","Azure","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-the-partnership-with-openai-has-cemented-microsofts-position-in-supercomputers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10045358,"title":"Exploring Panda Gym: A Multi-Goal Reinforcement Learning Environment","content":"With the latest breakthroughs in artificial intelligence and more developing research every day, it   is very believable that intelligent and self-sufficient machines are just on the horizon of arrival. Machines these days can understand verbal commands, distinguish between pictures, drive cars and play games, sometimes even better than an average human does. One can only wonder how much longer, and maybe it’ll walk among us? But, in developing an artificially intelligent machine, Reinforcement Learning and the learning environment it is trained in play a major role. The development environment used to train for machine learning is as important as the machine learning methods used to solve the predictive modeling problem. The Environment makes up for the basic and fundamental elements in a reinforcement learning problem. Therefore, it is important to understand the underlying environment with which the RL agent is to interact. This helps to come up with the right design and learning technique for the agent being delivered. The environment is the Agent’s world in which it lives, and the agent interacts with the environment by performing some action, but it does not have the right to influence the rules or dynamics of the environment by performing those actions. So, for example, just like humans are an agent in the earth’s environment and are confined with the laws. We can interact with the environment with our actions but cannot change the laws. The environment also gives a Reward to the agent; a scalar value returned that acts as feedback for the agent informing him whether its action was good or bad. Within Reinforcement Learning, multiple paradigms attain a winning strategy, i.e., making the agent perform the desired action in multiple ways. In complex situations, calculating the exact winning strategy or reward-value function becomes hard, especially when the agents start learning from interactions rather than the prior-gained experience. There are several types of learning environments present. The different types of Reinforcement Learning Environments are as follows: Deterministic environment: An environment where the next state of the environment can always be determined based on the current state and the agent’s action. Stochastic reinforcement learning environment:  An environment where we cannot always determine what the next state of the environment will be from the current state by performing a certain action. Single-agent environment: where only one agent exists and interacts with the environment. Multi-agent environment: Where there are more than one agents present that are interacting with the environment. Discrete environment: An action space of the environment that is discrete in nature. Continuous environment: Where the action space of the environment is continuous in nature. Episodic environment:  Here, the agent’s actions are only confined to the particular episode and not to any previous actions. Sequential environment: Here, the agent’s actions are connected with the previous actions it took. What is Open AI Gym? The gym is an open-source toolkit for developing and comparing reinforcement learning algorithms. What makes it easier to work with is that it makes it easier to structure your environment using only a few lines of code and compatible with any numerical computation library, such as TensorFlow or Theano. The gym library is a collection of test problems and environments that one can use to train and develop stronger reinforcement learning models. The present environments have a shared interface, allowing to write general algorithms as well. Furthermore, it provides a wide variety of simulated environments such as Atari games, board games, 2D and 3D physical simulations, and much more, so you can train multiple agents, compare them, or develop new Machine Learning algorithms for Reinforcement Learning problems. OpenAI is an artificial intelligence research company that Elon Musk partly funds. Its goal is to promote and develop friendly AI systems that will benefit humanity and work towards its betterment, rather than exterminating it! About Panda-Gym Panda-Gym, is an open-source library that provides a set Reinforcement Learning (RL) environment for the Franka Emika Panda robot integrated with OpenAI Gym. The Robot Simulation Environment consists of five tasks: reach, push, slide, pick & place and stack. It follows a Multi-Goal RL framework, allowing the use of goal-oriented RL algorithms. To foster open research, it also makes use of the open-source physics engine PyBullet. The implementation chosen for this package allows us to define new tasks easily or even create new robots. About The Simulation and Challenges The environments presented consist of a Panda robotic arm known as Franka Emika1, which is already widely used in simulation and real-life academic works. It has been designed with  7 degrees of freedom and a parallel finger gripper to perform tasks. The robot is simulated with the PyBullet physics engine, which, being open-source, helps show simulation performance. Furthermore, the environments are integrated with OpenAI Gym, allowing all learning algorithms based on the API. The simulation task consists of a challenge in moving either the gripper or objects to a target position. A task is considered as completed when the distance between the entity to move and the target position is less than 5 cm. The five tasks presented can be further tuned with an increasing level of difficulty. In the PandaReach-v1 task, a target position must be reached with the gripper. This target position is randomly generated in a volume of 30 cm × 30 cm × 30 cm. For PandaPush-v1,a cube placed on a table must be pushed to a target position on the table surface while the gripper is blocked. Here the target position and the initial position of the cube are randomly generated in a 30 cm × 30 cm square around the neutral position of the robot. PandaSlide-v1 simulation task consists of a  flat cylinder that must be moved to a target position on the surface of a table while the gripper is blocked. The target position is randomly generated in a 50 cm × 50 cm square located 40 cm in front of the neutral position of the robot. Since the target positions are out of reach of the robot, it is necessary to give an impulse to the object instead of just pushing it. For the PandaPickAndPlace-v1 simulation, a cube must be brought to a target position generated in a volume of 30 cm×30 cm×20 cm above the table. To lift the cube, it is necessary to pick it up with the fingers of the gripper. PandaStack-v1 Two cubes must be stacked at a target position on the table surface. The target position is generated in a square of 30 cm × 30 cm. The stacking must be done correctly: the red cube must be under the green cube. All these simulation challenges are still under research and yet to be completely solved with a perfect solution. Image Source Getting Started With the Code In this article, we will try to perform two simulations from the Panda Gym Challenge and understand what it takes to develop and set up the environment. The following implementation is inspired by the creators of panda gym, whose official website link can be found here. Installing the Library To get started, we will first install the panda-gym library; you can run the following code to do so, !pip install panda-gym Importing Dependencies Now we will be importing the dependencies required to set up the environment, Fetch, Pick and Place #importing dependencies import gym import panda_gym Environment Setup and Simulation #assigning the simulation task to environment env = gym.make('PandaPickAndPlace-v1') state = env.reset() #setting the environment done = False #rendering agent learnings images = [env.render('rgb_array')] while not done: action = env.action_space.sample() state, reward, done, info = env.step(action) images.append(env.render('rgb_array')) env.close() To set up the environments, you can run the following lines of code, The Hyperparameters can be further tuned according to the required performance; here, we will just perform a basic demo simulation. Next up, we will install the numpngw library, a python package that defines the function write_png that writes a NumPy array to a PNG file and write_apng to a sequence f arrays of an animated PNG (APNG) file. #installing numpngw !pip3 install numpngw from numpngw import write_apng write_apng('anim.png', images, delay = 100) # real-time rendering = 40 ms between frames Displaying the results, #rendering the simulation from IPython.display import Image Image(filename=\"anim.png\") As we can observe, the gripper moves the block! Furthermore, you can see the two positions of the block. Although the simulation might not be very clear, it can be further hyperparameter tuned or run on an even better computational system for better render performance. We can do the same for another simulation task of gripper slide. Fetch and Slide import gym import panda_gym env = gym.make('PandaSlide-v1') state = env.reset() done = False images = [env.render('rgb_array')] while not done: action = env.action_space.sample() state, reward, done, info = env.step(action) images.append(env.render('rgb_array')) env.close() !pip3 install numpngw from numpngw import write_apng write_apng('anim.png', images, delay = 70) # real-time rendering = 40 ms between frames from IPython.display import Image Image(filename=\"anim.png\") The rendering time and other learning parameters and environment can be further set up accordingly. This tool is very satisfying for testing deep reinforcement learning algorithms. However, some points can be limited, such as limitations in the gripper control, as it can only be controlled by high-level actions such as grasp and move. Additional work is required to allow the deployment of a policy learned in simulation. Also, the simulation is not completely realistic; the main concern is the gripper’s shape for picking up subjects in the environment. EndNotes Through this article, we understood the essence of a learning environment in the domain of Reinforcement Learning. We also tried to understand the panda gym problem and performed a basic demo simulation of two tasks rendering the Panda robotic arm, Franka Emika1. The following implementation can be found as a colab notebook which can be accessed using the link here. Happy Learning! References Panda-Gym GithubOpenAI Gym Official Documentation","excerpt":"The gym is an open-source toolkit for developing and comparing reinforcement learning algorithms. What makes it easier to work with is that it makes it easier to structure your environment using only a few lines of code and compatible with any numerical computation library, such as TensorFlow or Theano.","categories":["AI Trends"],"tags":["Reinforcement Learning","reinforcement learning an introduction"],"author_name":"Victor Dey","publish_date":"2021-08-08T16:00:00","publication_year":"2021","word_count":1638,"keywords":["NumPy","machine learning","artificial intelligence","Reinforcement Learning","AI","OpenAI","ML","Colab","Ray","TensorFlow","reinforcement learning an introduction","Pandas"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","OpenAI","Ray","TensorFlow","Colab","Pandas","NumPy"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/exploring-panda-gym-a-multi-goal-reinforcement-learning-environment\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10113592,"title":"Intel Collaborates with HCLTech to Advance Semiconductor Manufacturing","content":"HCLTech and Intel Foundry have announced their decision to expand their collaboration to co-develop silicon solutions to improve semiconductor innovation globally. This partnership will leverage HCLTech’s design expertise and Intel Foundry’s advanced technology and manufacturing capabilities. The goal is to establish a resilient and diversified supply chain to meet the rising global demand for semiconductor manufacturing. This collaboration will offer semiconductor manufacturers, system OEMs, and cloud services providers a robust ecosystem for semiconductor sourcing. Additionally, the collaboration has the potential to spur innovation by enabling the design of customised silicon solutions tailored to specific use cases. “Intel Foundry’s advanced technologies and silicon-verified IPs in manufacturing and advanced packaging strengthens our delivery of innovative, accessible and diverse solutions to our mutual clients. This will also give them greater choice and flexibility in semiconductor sourcing,” said Vijay Guntur, President, Engineering and R&D Services, HCLTech.D HCLTech has been collaborating with Intel for over 30 years, a relationship that has evolved through shared offerings and joint investments in various sectors, including silicon services, hardware engineering, telecom services, and more. The current focus is on jointly designing highly customised silicon solutions for companies, combining HCLTech’s design expertise with Intel’s manufacturing capabilities. This expanded collaboration is set to further strengthen their partnership by fostering a strong and open ecosystem beneficial for clients requiring advanced silicon solutions. Intel also announced that it has signed Microsoft as a foundry customer for a custom chip. This deal is part of Intel’s plan to overtake TSMC using its Intel 18A and upcoming 14A manufacturing technologies. The 18A, a 1.8nm technology, is set for early 2025 and will be used for future CPUs in both consumer and data centre markets. The 14A, planned for late 2026, will introduce a more advanced lithography tool for smaller and more efficient chips. Together with its collaboration with HCLTech to develop customised silicon solutions, Intel aims to meet the growing demand for semiconductors.","excerpt":"This collaboration will offer semiconductor manufacturers, system OEMs, and cloud services providers a robust ecosystem for semiconductor sourcing.","categories":["AI News"],"tags":["HCL Technology","Intel","Semiconductor India"],"author_name":"K L Krithika","publish_date":"2024-02-22T12:33:23","publication_year":"2024","word_count":318,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","Aim","HCL Technology","Semiconductor India","R","Intel"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/intel-hcltech-to-advance-semiconductor-manufacturing\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080973,"title":"A full-day hands-on workshop on Statistics for Data Science Beginners","content":"The Association of Data Scientists (ADaSci), the premier global professional body of data science and machine learning professionals, has announced a full-day hands-on workshop on Statistics with Python for Data Science Beginners to be held on December 3rd, Saturday. Data science is a multidisciplinary field, and a data science practitioner should acquire a comprehensive set of skills that covers every building block of the field in order to have a flourishing career. Statistics is one of the building blocks. Further, the goal of data science is to get insights from data. The first step toward achieving this goal is to thoroughly understand the data. Statistics is the most powerful tool for understanding, interpreting, and evaluating data. ADaSci is coming up with this opportunity to let data science aspirants dive deeply into statistics. In this workshop, the attendees will get a complete understanding of the concept of statistics and where and how it is used. Different ways of understanding the data using methods like descriptive statistics and inferential statistics will be discussed in this workshop with hands-on experiments in python. By the end of this workshop, the attendees will get a complete understanding of the concepts of statistics used in data science. This workshop will give a sound exposure to both descriptive statistics and inferential statistics with their real-world applications. Attendees will get a complete hands-on understanding of statistical concepts in python. Along with that, the attendees will get mastered in data analysis concepts with their mathematical understanding and hands-on implementations. Overall, this workshop will be a good opportunity for data science beginners to learn statistics from the very basics and get a sound understanding of what part of and how statistics is practically used in the field of data science. It will be a suitable dose for data science enthusiasts who are keen to learn this important and widely used mathematical part of data science. Details of the Workshop Date: 3rd December 2022 Time: 10 AM to 5 PM Mode: Online Price: 1,999 INR Discounts 100% discount for CDS Registrants 70% discount for CDS Associate Level Registrants 50% discount for ADaSci Members Register for the workshop here","excerpt":"This workshop will be a good opportunity for beginners to learn statistics from the very basics.","categories":["Deep Tech"],"tags":["ADAsci (Association of Data Scientists)","Data Science","Python","Statistics","Statistics for Data Science"],"author_name":"Dr. Vaibhav Kumar","publish_date":"2022-11-29T15:26:57","publication_year":"2022","word_count":357,"keywords":["data science","Go","machine learning","ADAsci (Association of Data Scientists)","programming_languages:R","Statistics","AI","programming_languages:Go","Python","Statistics for Data Science","programming_languages:Python","Data Science","R"],"extracted_tech_keywords":["AI","machine learning","data science","Python","R","Go","programming_languages:Python","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-full-day-hands-on-workshop-on-statistics-for-data-science-beginners\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10052718,"title":"Metaverse Is Still A Nascent Technology. Why Is Zuckerberg Betting Big On It?","content":"A few days back, rumours of Facebook rebranding itself hit everyone out of the blue. Shortly after the news ‘leaked’, Facebook CEO and co-founder Mark Zuckerberg announced that Facebook would now be called ‘Meta’ at the Connect conference. This rechristening is in tune with Facebook’s constant and very vocal push for its Metaverse endeavours or, as Zuckerberg said during the announcement, ‘our new North Star, bringing metaverse to people’. In the past, too, Zuckerberg has made no attempts at concealing his enthusiasm at hiding his ambition in this space; on an earlier occasion, he said that Metaverse was the company’s future. On paper, the plans of exploring Metaverse, the new frontier, which is an amalgamation of social networking, virtual and augmented reality, and AI, seem to be very exciting. But does a huge step bode well considering the amount of success Facebook has attained in this field? And Then There Was Meta During the official announcement, he said that with ‘Meta’, the company would aim to put people at the centre of the technology and unlock a massively bigger creator economy. He added that the name Facebook is still closely linked to just one product and could no longer fully encompass what the company was doing currently. “But over time, I hope we are seen as a metaverse company.” The change of name will not imply a change in the company’s corporate structure but will affect how it reports its financial results. From the fourth quarter of 2021, the company will be reporting on two operating segments — Family of Apps and Reality Labs. The company will also be trading under the new stock ticker MVRS from December 1. “Today’s announcement does not affect how we use or share data.” History of Facebook Why VR\/AR This year, in particular, has been chock-a-block in terms of metaverse-related announcements from the social media giant. Metaverse is the driver behind Oculus VR and the Horizontal virtual world. Facebook’s plan has been to push the scope beyond just gaming and include workplace, entertainment and others to create a social experience for users ultimately. Even before the latest announcement, 20 per cent of the company employees had been working exclusively on virtual and augmented reality, along with recent acquisitions like BigBox VR and Unit 2 Games. The VR segment alone accounts for up to 3 per cent of Facebook’s top line. The company has also announced new tools for people to build for the Metaverse. These tools include Presence Platform that enables new mixed reality experiences on Quest 2, along with a $150 million investment in immersive learning to train the next generation of creators. Experts had been speculating this for a long time, but now they stand vindicated on the speculation that Facebook is pinning its future entirely on VR\/AR. However, like many other technologies (think IoT), VR has not been able to realise its potential entirely. There are several challenges like the cumbersome headset, discomfort faced by users, and difficulty in fixing this tech. Oculus has been a great improvement from traditional VR and AR techs, but even the engineers and researchers at Facebook realise that this is a high-stakes bet. Failure to deliver would mean that the company’s user base might migrate to other apps like Snapchat. By Facebook’s own admission earlier, its AR glasses technology is a baby, and the company is gradually beginning to understand how much people would want to use such glasses. The company recently suffered a backlash in its partnership with Ray-Ban to produce glasses embedded with cameras. These glasses could help wearers record and share images and videos and offer other features like listening to music and taking calls. This raised some obvious concerns around surveillance. Purely speaking, in practical terms, the Metaverse is still in its very nascent stages. The technology required to realise its potential is still not available fully. This hasn’t stopped companies from exploring this field, but no other company seems to be as zealous as Facebook. In an interview, Andrew Bosworth, vice president of augmented and virtual reality at Facebook, said about Facebook’s interest in Metaverse, “In the last 10 years, we’ve seen that sometimes the public and regulators are surprised by the technological change, and that’s not healthy for anybody. So, we’re trying to have this conversation out in the open well in advance of this technology ever being delivered in any integrated form.” Distraction? Conspiracy theory, reality, or mere coincidence, people can’t seem to get over the timing of such a major announcement that changes the entire trajectory of the company. It is no secret that the company has been embroiled in several allegations of data theft, manipulation, and hosting problematic content on its platform. Former Facebook employee Frances Haugen recently released a cache of internal documents to the public, pointing out that Facebook put profit before user safety. Zuckerberg in an interview with TheVerge denied any relation between the rebranding and the controversies surrounding the company. Instead he called this speculation ‘ridiculous’ and said that if anything, this is not the environment a company would introduce a new brand.","excerpt":"The technology required to realise metaverse’s potential is still not available fully. This hasn’t stopped companies from exploring this field, but no other company seems to be as zealous as Facebook.","categories":["Global Tech"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-11-02T12:00:00","publication_year":"2021","word_count":851,"keywords":["programming_languages:R","AI","Ray","Aim","R"],"extracted_tech_keywords":["AI","Aim","Ray","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/metaverse-is-still-a-nascent-technology-why-is-zuckerberg-betting-big-on-it\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":28968,"title":"RPA Solution Providers Are Riding The Success Wave In India Right Now","content":"Robotic process automation has the potential to offer a high value in terms cost reduction and increased productivity. The value is realised in a short time as the deployment is quick and at a low risk, owing to the fact that the integration is non-invasive. RPA is also a burgeoning market and business leaders have woken up to its potential to improve the performance by automating repetitive, time-consuming tasks. Over the last two years, there has been a considerable rise in interest over RPA which is increasingly seen as the cornerstone for digitisation and is on top of the most management agendas. According to a study by P&S Market Research, the RPA market is expected to touch $8.6 billion by 2023, growing at a CAGR of 36.2% during 2018-2023. For example, US-based RPA leader, Automation Anywhere is a leader in digital workforce automation and implements digital workforce platform that allows companies to operate leaner and more efficiently. RPAs significantly increase revenue growth, create new revenue streams and offers a clear business advantage. Global automation leaders such as Kryon Systems provide RPA cloud solutions that help organisations to decrease the overhead costs and efforts related to infrastructure deployments, maintenance, configuration and support. Through cloud accessibility, bots can be easily deployed, thereby serving virtual environments with no footprint on an organisation’s core business infrastructure. Across the globe, some of the top technology vendors are Automation Anywhere, UiPath, WorkFusion, NICE, Blue Prism, Softomotive, Kofax Kapow and Redwood RoboFinance among others. Software Robots Are Used For: IT Support and Management: RPA is widely deployed for IT support tasks, for enhancing desk operations and monitoring network devices. Automated Assistant: In terms of customer support, machines execute repeatable tasks by answering queries in natural languages. Automated Assistants have reduced the dependency on call centres. Process Automation: Many enterprises use RPAs to automate rule-based processes wherein the present IT architecture is not changed. These consist of back-office tasks related to customer services, finance, HR, data entry etc. Top RPA Service Providers In India Datamatics: This award-winning, Mumbai-headquartered company is one of the leading providers of robotics, advanced analytics, mobility and cloud solutions. It’s proprietary platform — TruBot offers a team of virtual assistants that help in automating repetitive and rule-based processes without any manual intervention. When powered with AI, it can recognise and replicate human actions, creating a smarter bot that can make informed decisions. The bot is multi-skilled and can automate a bunch of complex processes in such as Citrix-based accounting processes, dispute resolution and cash, credit and overdraft limit extension. The platform TruBot is sector agnostic and can be used across a range of verticals such as healthcare, banking, insurance, manufacturing and logistics. Q3Tech: Gurgaon-headquartered Q3 technologies provides RPA software for BPO, healthcare, retail, telecom, BFSI and manufacturing among other sectors. Some of the top applications of RPA include — accounting (general accounting, transactional reporting and budgeting), healthcare (claims, customer support, account management, billing, reporting and analytics), human resources (timesheet submission process, updating employee information), financial sector (automating procedures such as account opening, insurance claims processing, audit request etc.) It is also widely deployed in customer care for tasks such as verifying e-signatures, uploading scanned documents and automating a slew of contact centre tasks. According to the company website, RPA software is not a part of an organization’s IT infrastructure, but instead, it sits on top of it and implements technology efficiently, without affecting the existing infrastructure and systems. Infosys BPM: Today, rule-based and knowledge-based tasks form a chunk of RPA jobs across finance, HR, customer support, customer experience, accounting and legal processes among others. Infosys RPA provides both rule-based automation for repeatable tasks and knowledge-based automation for tasks that require human expertise. Knowledge-based automation leverages advances such as deep learning, artificial intelligence, natural user interfaces, cognitive computing backed by computing resources.  According to an Infosys whitepaper, the company has delivered an innovative business process service stack, reducing human effort significantly, increasing productivity and provides cost savings. The solution component is made up of in-house solutions and is also based on third-party solutions. Today, Infosys RPA solution is used by CPG major to automate order management process, thereby reducing the turnaround time and improving reporting operations. Srijan: New Delhi headquartered Srijan set up in 2002 is one of the largest pure-play Drupal company in Asia. The company provides rule-based automation for repetitive tasks in customer support, HR and finance and knowledge-based automation, powered by AI for processes that require human intelligence.","excerpt":"Robotic process automation has the potential to offer a high value in terms cost reduction and increased productivity. The value is realised in a short time as the deployment is quick and at a low risk, owing to the fact that the integration is non-invasive. RPA is also a burgeoning market and business leaders have […]","categories":["AI Features"],"tags":["AI and automation","Automation"],"author_name":"Richa Bhatia","publish_date":"2018-10-05T05:40:09","publication_year":"2018","word_count":746,"keywords":["artificial intelligence","AI","Automation","AI and automation","virtual assistants","Git","RAG","Aim","deep learning","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","analytics","Aim","RAG","virtual assistants","R","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/rpa-solution-providers-are-riding-the-success-wave-in-india-right-now\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10142180,"title":"What’s Nutanix Cooking Up for New Bharat?","content":"At the Nutanix .NEXT India tour event, the cloud software giant elaborated on its role in transforming India into a ‘new Bharat’. It achieves this by fueling innovations in sectors as diverse as banking, automotive, oil and gas, pharmaceuticals, IT\/ITeS, and software. The company highlighted its extensive reach into the public sector, powering government state data centres. It enables critical citizen services, supports national data projects, strengthens hospitals, and even fortifies the Indian defence forces. These vital institutions rely on Nutanix to build, secure, and sustain the nation’s core infrastructure. Together, they’re not just creating solutions but are shaping the future of India. What’s India to Nutanix? For Nutanix, India represents one of its most dynamic markets globally. “India is among our top-performing markets,” shared Andrew Brinded, chief revenue officer at Nutanix. Collaborations with key partners, such as HPE (Hewlett Packard Enterprise), have fueled the rapid adoption of Nutanix’s hypervisor, AHV (Acropolis Hypervisor), which offers cost savings without compromising on functionality. Historically known for virtual machine management, Nutanix has shifted focus toward containerisation, catering to companies embracing modern, agile infrastructures. “Our unified platform allows IT administrators to manage virtual machines and containers seamlessly, delivering simplicity and efficiency,” Brinded noted. He further highlighted the company’s role in supporting AI-driven innovations across sectors, citing Apollo Pharma as a case study. With Kubernetes deployments powered by Nutanix’s NKP technology, Apollo leverages AI solutions like ‘GPT in a box’ to enhance healthcare delivery for over 200 million people in India. “Nutanix integrates seamlessly with partners like NVIDIA for AI processing and offers robust data services, whether deployed on-premises, at the edge, or in public clouds,” Brinded explained. Customers can implement their preferred large language models (LLMs) on the Nutanix infrastructure, ensuring flexibility and control. This adaptability has made Nutanix a trusted partner in industries ranging from manufacturing and finance to pharmaceuticals and defence. Partnering for Hybrid Cloud Success Nutanix’s success lies in its ecosystem-driven approach, collaborating with major players such as AWS, AMD, and HPE. The company’s exclusive channel model involves working closely with Indian resellers and managed service providers, ensuring a seamless customer experience. “Our hybrid cloud strategies allow customers to integrate public cloud solutions with private infrastructures while maintaining a consistent user experience,” Brinded said. AI in Finance One of the clients Nutanix is closely working with to help it emerge as a leader in automating financial services is HighRadius. Simran Singh, vice president of cloud engineering at HighRadius, offered an insider’s perspective on how it can be brought to reality. “HighRadius operates in the finance domain, primarily automating functions within the office of the CFO,” he shared. Recognised as a Gartner Quadrant leader in the order-to-cash space, HighRadius has broadened its focus to include treasury, record-to-report, B2B payments, and account reconciliation. Founded with roots in Robotic Process Automation (RPA), HighRadius was quick to adopt AI and machine learning. “For over a decade, we’ve integrated AI into our products, long before the rise of generative AI,” Singh remarked. It works with around 1,000 customers including global giants like Unilever and Nestlé, delivering transformative financial solutions, streamlining processes, and enabling CFOs to focus on strategic objectives. “Our solutions deliver a 70% efficiency improvement out-of-the-box, reaching up to 90% within six months. This eliminates manual reconciliation and empowers CFO teams to make more impactful decisions,” Singh explained. Addressing Challenges with Hybrid Cloud Strategies As HighRadius scales, it faces growing challenges related to cost management and infrastructure optimisation. With key banking partners like major US banks hesitant to fully embrace public clouds, HighRadius adopted a hybrid cloud approach. “We’ve implemented Nutanix in our data centre and transitioned to hyperscalers with remarkable success, resulting in significant performance improvements across infrastructure and applications,” Singh noted. In an innovative move, HighRadius deployed bare-metal Nutanix servers within AWS, hosting Nutanix Database Service (NDB) instead of relying on RDS Aurora. This unconventional strategy cut costs by 40% while enhancing operational efficiency. “While the initial proof-of-concept posed challenges, the collaboration between teams made this a success we’re incredibly proud of,” Singh said. HighRadius is now testing Nutanix Karbon Platform Services (NKP) to create a seamless integration between private and public clouds. The long-term vision is to maintain cloud-agnostic operations that ensure scalability and adaptability, with a view to potentially transitioning fully to private clouds in the future. What’s Next? For HighRadius, the focus is clear: Scaling revenue while maintaining its hybrid cloud approach and exploring future transitions. “We’re growing at 40% year-on-year and expanding both organically and inorganically. Our long-term strategy is centred on scalability, agility, and efficiency,” Singh said. For Nutanix, the priority is fostering innovation in India’s burgeoning tech ecosystem, driving the adoption of AI and hybrid cloud solutions across industries. As Brinded aptly put it, “India’s passion for technology and rapid adoption make it one of our best markets for driving growth in emerging areas.” The collaboration between these two tech leaders exemplifies how strategic innovation and partnership can redefine business operations. With HighRadius leveraging Nutanix’s infrastructure expertise, both companies are well-positioned to lead the way in their respective domains. [Update: Nutanix requested to remove the revenue part of HighRadius]","excerpt":"Nutanix’s success lies in its ecosystem-driven approach, collaborating with major players such as AWS, AMD, and HPE.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Nutanix"],"author_name":"Vidyashree Srinivas","publish_date":"2024-12-02T10:17:20","publication_year":"2024","word_count":852,"keywords":["Go","machine learning","AWS","AI","ML","Nutanix","RAG","generative AI","Rust","R","kubernetes","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","RAG","AWS","kubernetes","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/whats-nutanix-cooking-up-for-new-bharat\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10098078,"title":"EPAM Launches DIAL, a Unified Generative AI Orchestration Platform","content":"EPAM Systems has announced the launch of its AI-powered DIAL (Deterministic Integrator of Applications and LLMs) Orchestration Platform, which merges the power of Large Language Models (LLMs) with deterministic code — offering a secure, scalable, and customisable AI workbench to streamline and enhance AI-driven business solutions. Developed by EPAM’s Reliable AI Lab (RAIL), DIAL helps enterprises speed their experimentation and innovation efforts across an extensive range of LLMs, AI-native Applications and Custom Add-ons as well as provides a practical approach for engineering business solutions with reliable AI capabilities. The DIAL Platform offers a unified user interface, empowering businesses to leverage a spectrum of public and proprietary LLMs, Add-ons, APIs, Datastores and Business Applications. This integration promotes the development of novel enterprise assets that co-exist seamlessly with an organization’s existing workflows. In keeping with EPAM’s long-standing commitment to Open Source, portions of DIAL will be released under an Apache 2.0 licensing scheme as part of its launch. This initiative encourages responsible use, community innovation and the adoption of responsible AI enterprise standards within the industry. Moreover, Applications and Add-ons can be implemented through diverse approaches, encompassing LangChain, LLamaIndex, Semantic Kernel or custom code — all within an integrated, secure and scalable framework. The DIAL Platformme aggregates multi-cloud assets libraries including components, routing, rate-limiting software, monitoring tools, load-balancing solutions and deployment scripts. This extensive, curated toolkit supports a wide range of business use cases and integration scenarios and offers approaches to significantly optimize the consumption of external LLMs.","excerpt":"In keeping with EPAM’s long-standing commitment to Open Source, portions of DIAL will be released under an Apache 2.0 licensing scheme","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-08-03T16:14:52","publication_year":"2023","word_count":246,"keywords":["API","AI","ML","LlamaIndex","Scala","RAG","responsible AI","LangChain","GAN","R"],"extracted_tech_keywords":["AI","ML","LangChain","LlamaIndex","RAG","R","Scala","API","GAN","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/epam-launches-dial-a-unified-generative-ai-orchestration-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164591,"title":"How CME Group Evolved from a Service Hub to a Thriving CoE in India","content":"In recent years, global capability centres (GCCs) have undergone a significant transformation, evolving from traditional back-office operations into hubs of innovation. Reports suggest that by 2030, GCCs will no longer exist in their current form. Instead, they will become an integral part of global organisations, adopting new models and focusing on business growth and revenue generation. GCCs in India are already leading this shift, with business leaders and general managers increasingly based out of these centres, taking on profit and loss (P&L) responsibilities. Companies like Citrix have embraced this model, placing global managers in India to drive strategic decision-making. A striking example of this transformation is CME Group’s India office in Bengaluru. In an interview with AIM, Prabhuram Duraiswami, executive director and centre head of India of CME Group, highlighted, “Our decade-long journey has been very fulfilling and rewarding as it has seen us transform from a service delivery site to a centre that is able to drive immense value and positive impact to our customers. We have been successful in establishing ourselves as a centre of excellence.” According to Duraiswami, CME Group was established in 2015 as a small operation with 100 employees and has since expanded nearly sevenfold. It primarily handles technology, operations, and quantitative engineering. Today, it functions as a strategic centre of excellence and plays a pivotal role in cloud adoption, risk management solutions, and advanced quantitative research. Cloud, AI, and Agile Innovation CME Group’s 10-year partnership with Google Cloud, signed in 2021, is a game-changer for the global derivatives market. “Google Cloud [aims] to transform global derivatives markets through cloud adoption, and at the same time…co-innovate to deliver expanded access, new products and more efficiencies for all market participants,” Duraiswami mentioned. He also stated that, with Google Cloud Platform (GCP) migration as a top priority, the team in Bengaluru is sharpening its focus on cloud, data, platform engineering, and reliability engineering. In this regard, the company is actively hiring product managers, scrum masters, agile coaches, and data experts to drive innovation at scale. Over the years, its workforce has grown from 100 to nearly 750 employees. Leveraging India’s Talent for Global Impact One of the most fascinating aspects of CME Group India’s journey is its role in shaping the company’s next-generation clearing and risk management solutions. “Our clearing team is responsible for developing and implementing quantitative models and algorithms for risk management, pricing and valuation,” Duraiswami revealed. Moreover, he pointed out that the team’s strategic location in India allows it to seamlessly support the Asia-Pacific region, ensuring smooth operations during crucial end-of-day clearing cycles. This time-zone advantage, combined with technology and expertise, has positioned the India team as a key enabler of global financial stability. According to him, the company’s Bengaluru office is also involved in product introductions and customer portfolio building. For the latter, Duraiswami highlighted that a significant part of the global team is based in India and is responsible for conceptualising, developing and implementing solutions that meet the needs of CME Group’s customers. Talking about the diverse and skilled talent pool, Duraiswami stressed that the CME Group in India is investing in upskilling and reskilling its staff to meet the evolving needs of the business. He pointed out that the sheer size of the pool of highly skilled technical talent available here allows it to play an increasingly important role in developing and supporting the company’s complex products and services. What are the Others Doing? A similar sentiment is reflected in Kimberly-Clark, a global leader in personal care brands such as Huggies, Kleenex, Andrex, Cottonelle, Scott, and Kotex. The company appears to have a strong affinity for Bengaluru, choosing it as the location for its global digital and technology centre (GDTC), setting it apart from other cities in India and around the world. “Bengaluru provides access to a highly skilled talent pool and serves as a hub of technological innovation within India,” said Sreekanth Jayabalan, VP and CIO of digital transformation office and enterprise markets at Kimberly-Clark, during an interview with AIM. Similarly, Amit Kapur, VP of AI and data analytics at Lowe’s told AIM, “Many of our core systems, including the omnichannel order management and self-checkout terminals, were built from the ground up by our engineers in Bengaluru. These solutions have provided us with unmatched scalability and flexibility.”","excerpt":"Over the years, CME Group’s workforce in India has grown from 100 to nearly 750 employees.","categories":["GCC"],"tags":["AI","GCC"],"author_name":"Shalini Mondal","publish_date":"2025-02-26T11:00:00","publication_year":"2025","word_count":717,"keywords":["Go","GCC","GCP","AI","ML","Scala","Git","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","GCP","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/gcc\/cme-group-indias-journey-from-service-delivery-to-centre-of-excellence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":36710,"title":"How Indian Public Healthcare Sector Can Leverage AI To Make Remarkable Progress","content":"As one of the most complicated and dynamic sectors in India, the healthcare industry is fraught with many challenges which include unavailability of requisite infrastructure to understaffed hospitals catering to scores of people. In India, over 85 per cent of the healthcare institutions work with less than five workers and numerous hospitals are estimated t operate without proper quality certification. According to government data, a mere 195 hospitals in the country have proper certification to run the establishment. The application of artificial intelligence and machine learning in healthcare sector has been practiced for long and now more medical practitioners and institutions are leveraging the technology not just to drive precision medicine but also at the back end for delivering better services. AI experts and practitioners in the field strongly feel that the country with its vast pool of unstructured medical data can turn itself into a perfect place to usher in the next-phase of med tech to the country. With scores of AI startups and private players operating in the market presently, the adoption of AI in healthcare is expected to grow in coming years. Speaking about the role AI and other emerging technologies in shaping the medical industry, Charu Shegal, Life Sciences and Healthcare Leader Deloitte in India, points out in a study “Adding infrastructure and medical professionals alone will not be able to solve India’s huge unmet needs in healthcare. It needs to be supported by technology. An effective and innovative use of medical technology, has the potential of increasing access, significantly reducing the burden of disease and the load on healthcare delivery services through early diagnosis, better clinical outcomes, less invasive procedures and shorter recovery times.” Realising the potential of the technology, government institutes like All India Medical Institute (AIIMS) is already incorporating AI and ML into their research. Last month, researchers from Indraprastha Institute of Information Technology, Delhi, (IIIT-D) and AIIMS collaborated for a project that uses thermal images and AI to predict haemodynamic shock or insufficient oxygen supply to organs leading to multi-organ failure in children even before doctors can diagnose the condition. The system was tried on children between the age group of 10-15 with a success rate of 69 per cent with the algorithms predicting the condition 12 hours in advance. In yet another case, researchers from the National Brain Research Centre (NBRC), and Neuroimaging and Neurospectroscopy Laboratory (NINS) used AI for early detection of Alzheimer’s. AI For Better Medical Service While the corporate and startups have given a major boost to the country-wide adoption of AI in healthcare, in India’s first ever AI Strategy, the NITI Aayog has laid down coherent plans for further large scale adoption of the technology. Currently, it is in an advanced stage of launching a national repository of annotated and curated pathology images to help oncopathologist use Machine learning solutions to its data set called Digital Pathology for early-stage cancer detection. It is also planning to launch the National AI Market (NAIM), a one-of-a-kind initiative which is composed of data marketplace, data annotation market and model marketplace\/solutions marketplace. “The marketplace mechanism being proposed here will be aimed at easing the adoption efforts of all participants – private enterprises, PSUs, governments, startups and academia. A common platform which brings together enterprises and AI solution builders will trigger off the initial collaboration towards building AI solutions and adopting them at scale,” they said in the proposed plan. Thus, by allowing for hospitals and other stakeholders to access the vast set of data, the government hopes to make medical care easily accessible and more affordable for millions of people across the country. Highlighting one such  use case, the draft policy said that radiologists and physician can use the platform for assistive diagnostic. For Reforming Medical Education In a court hearing in 2018, the Supreme Court suggested that advanced technologies like AI and ML should be leveraged to look into verification of medical colleges by the Medical Council of India (MCI) to grant permissions. Owing to the lack of clarity in the existing procedure followed by the MCI, the court-appointed Nandan Nilekani, co-founder of Infosys to find a solution. Nilekani, who headed the Aadhaar project previously, was given complete liberty to technical assistance from various IT companies within the country to develop a solution. Though not much development has been undertaken in the field, the court’s consideration of emerging technologies like AI for better delivery of medical services for the people is an indicative of the technologies growing prevalence.","excerpt":"As one of the most complicated and dynamic sectors in India, the healthcare industry is fraught with many challenges which include unavailability of requisite infrastructure to understaffed hospitals catering to scores of people. In India, over 85 per cent of the healthcare institutions work with less than five workers and numerous hospitals are estimated t […]","categories":["AI Features"],"tags":["AI Healthcare","ai in medicine","medtech","structured and unstructured data"],"author_name":"Akshaya Asokan","publish_date":"2019-03-21T12:44:16","publication_year":"2019","word_count":748,"keywords":["Go","API","artificial intelligence","machine learning","AI","AI Healthcare","ML","Git","medtech","RAG","Aim","ai in medicine","R","structured and unstructured data"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-indian-public-healthcare-sector-can-leverage-ai-to-make-remarkable-progress\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10164209,"title":"Razorpay Unveils Ray Concierge, Co-Pilot To Supercharge Payment Experience for Businesses","content":"Bengaluru-based fintech company Razorpay hosted its flagship event, FTX’25, on Thursday, where it unveiled a series of products powered by AI. The company claims these innovations will redefine payment processes for businesses, making setup easier and delivering a seamless experience for users. The event showcased Razorpay’s commitment to harnessing AI’s potential, not just as a tool but as a core element of its payment solutions. The company emphasised its AI-driven approach to address the evolving needs of businesses in an increasingly digital world. Among the key announcements was the launch of Ray Concierge, an AI-powered onboarding system that simplifies the often complex process of setting up payment gateways. Ray Concierge acts as an assistant, analysing business details, including the documents, and automating form filling, significantly reducing the time and effort required for onboarding. To address this, Shashank Kumar, founder and MD of Razorpay, said in a keynote speech that with the help of AI, the company has built the “best business onboarding experience in the industry”. Razorpay has also introduced Ray Co-Pilot an AI-powered tool for assisting developers with payment gateway integration. It plans to simplify the process by generating code, testing setups, and providing guidance, eliminating the need for extensive manual configuration and saving time. “Ray Co-Pilot is a helpful engineer who will help you with the most complex integrations. You don’t need to read docs or test it manually. You can just tell the Co-Pilot what you need, and…it’ll do the integration [and] test the setup,” Kumar added. The company further showcased its vision for the future of payments by introducing an agentic AI toolkit designed to enable seamless payments. This technology allows payments to be processed within AI agents’ environments, eliminating the need for users to navigate to external websites, a separate payment gateway page, or apps. Referring to agentic AI, Harshil Mathur, co-founder and CEO of Razorpay, mentioned, “Now, when an AI assistant finds you the perfect hotel, OTT subscription or a limited-edition speaker, you don’t have to open the web page. [You can] order yourself and do all of those things. You don’t get redirected, [and] you don’t get interrupted.” Razorpay’s FTX 25 keynote session highlighted the company’s dedication to innovation and its belief in the transformative power of AI. With these new AI-driven products, Razorpay aims to empower businesses with streamlined payment solutions, paving the way for a more efficient and user-friendly payment experience.","excerpt":"Razorpay is adding AI to the mix to improve the payment setup and experiences for businesses and customers.","categories":["AI News"],"tags":["Razorpay","UPI"],"author_name":"Ankush Das","publish_date":"2025-02-20T14:59:52","publication_year":"2025","word_count":400,"keywords":["agentic AI","AI","RPA","ML","innovation","Razorpay","Git","UPI","Ray","Aim","AI agents","R"],"extracted_tech_keywords":["AI","ML","agentic AI","Aim","Ray","R","Git","RPA","innovation","AI agents"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/razorpay-unveils-ray-concierge-co-pilot-to-supercharge-payment-experience-for-businesses\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10114240,"title":"Salesforce Announces the Public Beta Availability of Einstein Copilot","content":"Salesforce today announced the public beta availability of Einstein Copilot, a new customisable, conversational, and generative AI assistant for CRM. Einstein Copilot is available in beta globally for Sales Cloud and Service Cloud, with Commerce Cloud and Marketing Cloud available later in 2024. “Unlike other AI assistants or copilots that lack adequate company data to generate useful responses, Einstein Copilot enables Salesforce customers to generate responses using their own private and trusted data, while maintaining strict data governance and without requiring expensive AI model training,” the company said in a press release. Einstein Copilot can answer questions, summarise content, create new content, interpret complex conversations, and dynamically automate tasks on behalf of a user, all from a single, consistent user experience embedded directly within Salesforce’s AI CRM applications. Customers can tap into the full power of Salesforce with Einstein 1 Editions, simplified technology packages for businesses looking to accelerate growth and speed productivity with the AI CRM. Einstein 1 Editions provide organizations in every industry access to the best of Salesforce technology, including CRM, Einstein Copilot, Data Cloud, Slack and Tableau in a single offering, helping them transform their business and deliver stronger customer experiences. This is made possible by combining a conversational UI, a foundational large language model, and trusted company data that enables Salesforce users to tap into the power of generative AI and interact with their applications in entirely new ways. “Our new Einstein Copilot brings together an amazing intuitive interface for interacting with AI, world-class AI models and above all deep integration of the data and metadata needed to benefit from AI. Einstein Copilot is the only copilot with the ability to truly understand what is going on with your customer relationships,” said Marc Benioff, Chair & CEO, at Salesforce. The company said Einstein Copilot grounds its responses with trusted business data from Data Cloud to provide the necessary context for the highest quality outputs. This allows Einstein Copilot to generate more precise and tailored responses based on trusted company data. Moreover, Einstein Copilot can be customised to accomplish specific sales, service, marketing, commerce, and IT tasks, ensuring company and industry policies are applied. Copilot Builder can create custom actions for Einstein Copilot, Prompt Builder activates custom prompts in the flow of work, and Model Builder uses proprietary AI models to power custom Einstein Copilot functionality. Currently, Einstein Copilot supports data residency in the US and the English language.","excerpt":"Einstein Copilot currently supports data residency in the US and the English language.","categories":["AI News"],"tags":["Salesforce"],"author_name":"Pritam Bordoloi","publish_date":"2024-02-28T11:41:23","publication_year":"2024","word_count":403,"keywords":["Go","AI assistants","TPU","AI","GAN","generative AI","Salesforce","Rust","data governance","copilots","R"],"extracted_tech_keywords":["AI","generative AI","AI assistants","copilots","TPU","R","Go","Rust","data governance","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/salesforce-announces-the-public-beta-availability-of-einstein-copilot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":4159,"title":"Case Study – Taurus Analytical Solutions Limited","content":"It was first week of March 2013, when the CEO of the Taurus Analytical Solutions Limited, Phillips Kutty called for a meeting of the Marketing Manager, Human Resource Manager, Accounts Manager and Business Development Manager to critically analyze the problems faced by the company and to find the way forward. Sales of the company were bleak and the pipe line of projects for the next financial year 2013-2014 seemed almost dry, with only four long term projects from key accounts. The company had witnessed 50 percent dip in the sales and its profit figure showed just US$ 5 million when compared to previous year’s profit of US $10 million. Marketing manager, Arun Mehta, met with varied new client prospects over the last six months especially in the US, but even after persuasive selling there were no fresh projects coming in for the current year primarily because of unfavorable economic conditions in the United States. The company had created key accounts teams with separate and dedicated working units for the four key clients with which it had a contract of long term projects. Except for the key accounts, other clients who had promised to provide new projects backed out stating certain economic and political reasons. Arpana Sharma, Human Resource Manager, had pursued the manpower planning, taking in to consideration the expected projects and thus recruited domain people from various research and management institutes. As the company had already provided them the offer letters and mentioned the joining dates, it was not in a position to back out from its promises. This new batch was to join by the mid of March 2009 and further it had to be trained for the next three months. Given the above environment the CEO was concerned about how would the company be able to increase its sales and justify its expenditure, the failure of which would push the company in to considerable loss. The Company Taurus Analytical Solutions Limited was a midsized company with its inception in the year 2004. Since its inception it had worked in 290 therapeutic disease areas and had around 700 employees. Although headquartered in Bangalore, it had client offices in the United States and European countries. All of its clients were from the western world, primarily from the United States. The company provided the core analytical services such as competitive intelligence, sales forecasting, predictive modeling, suggesting suitable marketing mix, branding and positioning strategies, and developing suitable patient monitoring and data collecting tools for the top 50 pharmaceutical companies. Taurus Analytical Solutions provided services of sales and marketing analytics, and business consulting for the pharmaceutical and healthcare industries. The company’s services addressed decision-making, strategic planning and the routine business needs of the customers, across geographies, both at headquarters and affiliate levels. Exhibit 1: Services offered Sl.No Services provided Applications Niche market 1 Marketing Analytics Marketing research, marketing mix, STP with reference to patient based models US 2 Sales Analytics Sales forecasting, sales automation and sales effectiveness US, UK 3 Forecasting Business development and product forecasts for planning and development US 4 Modeling Predictive and analytical models  based on the historical client data UK, US 5 Software tool development Knowledge management, CRM databases US Taurus Analytical Solutions provided holistic services in population-based epidemiological study analysis, pharmaco-epidemiological study analysis, feasibility studies for clinical trials, and epidemiology database applications. The Company supported epidemiology, marketing research, and outcomes research groups in plugging the epidemiology data gaps they might have with reference to patient-based models, especially in the emerging markets. Taurus identified cost effective sources of secondary data and the research publications that provided estimates, and wherever feasible, validate those against other sources and its physician network to deliver evidence-based disease estimates. Pricing strategy Even though Taurus was based in developing country, India, it was unable to generate business from this market and primarily relied on western markets i.e. the United States and certain European countries. Their services were high priced and were suitable for premium customers. Arun was of the opinion that Indian pharmaceutical companies didn’t require the kind of services that Taurus provided because of the following reasons: 1)      The pharmaceutical sector in India was still in nascent stage and thus these pharmaceutical companies were not in need of sophisticated and niche services provided by Taurus. 2)      Indian Pharma companies cannot afford and may not be willing to pay  the premium service charges of Taurus. As the pipe line had almost gone dry now the company was introspecting on its segmentation strategy to boost its sales. External Environment Since the major business of the company is generated from the United Sates, the prospects of the company was considerably influenced by the policy and economic framework of health care sector in the US. U.S.A. historically had given huge emphasis to the development of health care sector. It had numerous, conducive public policies like medical expenditure reimbursement and insurance policy coverage to the general public. The reimbursement was up to 100 percent for the chronic diseases. This gave a huge boost to health care and pharmaceutical industries in the US. Thus all the pharmaceutical companies including the pharmaceutical analytical companies showed an impressive growth and profits till 2008. However in 2008, due to economic downturn or recession, the US economy was worst hit with its repercussions being felt in pharmaceutical analytical companies like Taurus Analytical Solutions.  In addition to this the US government wanted to reduce its health care burden, which was increased due to non-conducive demographics, i.e.  decrease in the population of working class and increase in the population of old aged and dependent infants. In this regard the US government decided to reduce the government medical reimbursement to only 75 percent, this non-conducive political development had considerable negative impact on the health care and consequently on analytical industry which was deeply felt by Taurus. Competition Taurus Analytical Solutions faced considerable competition from players across the globe, with two major competitors located and operating in the same geographical region in which this company operated. As the company was midsized and comparatively new, the competitors enjoyed a greater market share. The developed economies posed a good market for the pharmaceutical analytical companies and hence  all the competitors focused on increasing the market share in the specified geographical locations like UK and the US, thereby increasing the competition. Diversification strategy The company thought of diversifying into other markets like Japan and China, but its efforts were not successful primarily because Japan, being highly conservative nation posed the entry barrier and China specifically posed the language barrier. Traditionally regulated market such as European Union was backing generics (i.e. duplicates of chemical drugs prepared after their respective patents lapsed) and biosimilars (i.e. biologic mimics of off patent biologic drugs)with conducive policies. On the contrary, the US was undergoing intense negotiations with varied stakeholders for the standardization of the biosimilars and generic policies. This had considerable negative impact on the sales and revenues of pharmaceutical industry and consequently influencing the pharmaceutical analytical industry. Now the marketing manager was in dilemma as to what should be the sales and marketing strategies for the next year, so as to boost the sales. He wondered whether the company should increase sales efforts in the rapidly growing markets or build effective strategies for marketing the services in the domestic market in view of existing macro environmental issues. Food for thought: What sales and marketing strategies should the marketing manager plan to boost the sales of the company in the recessionary economy?","excerpt":"It was first week of March 2013, when the CEO of the Taurus Analytical Solutions Limited, Phillips Kutty called for a meeting of the Marketing Manager, Human Resource Manager, Accounts Manager and Business Development Manager to critically analyze the problems faced by the company and to find the way forward. Sales of the company were […]","categories":["IT Services"],"tags":["Analytics Case Study"],"author_name":"Srujana H.M.","publish_date":"2013-10-09T07:44:59","publication_year":"2013","word_count":1247,"keywords":["Go","API","ELT","programming_languages:R","AI","RPA","RAG","automation","analytics","Analytics Case Study","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","API","ELT","automation","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/case-study-taurus-analytical-solutions-limited\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10023301,"title":"Arm Announces New Architecture After Ten Years: Major Updates & Latest Features","content":"Chip designer Arm recently introduced its latest architecture Armv9, at the Arm Vision Day 2021 event. Arm’s last architecture announcement came in 2011, at the Armv8 launch. Armv9 offers improvements in performance, speed, and security. “As we look toward a future that will be defined by AI, we must lay a foundation of leading-edge compute that will be ready to address the unique challenges to come. Armv9 is the answer. It will be at the forefront of the next 300 billion Arm-based chips driven by the demand for pervasive specialized, secure and powerful processing built on the economics, design freedom and accessibility of general-purpose compute,” said Simon Segars, chief executive officer, Arm. Credit: Arm Top features include: Confidential Compute Architecture The most significant feature of Armv9 is the Arm Confidential Compute Architecture (CCA). It introduces the concept of dynamically created Realms that can be used by all applications where the data is concealed from the operating system and other apps on the device. Credit: Arm Unlike traditional security models that allow privileged software to see lower-tier applications’ execution, Realms can help shield sensitive data and code from the rest of the device and privileged software by performing computations in a hardware-based secure environment. Think of them as secured containerised execution environments entirely opaque for the OS or hypervisor. Instead of hypervisor, Realms are managed by a Realm manager. These are roughly one-tenth the size of a hypervisor. Realms could significantly reduce the chain of trust of a given application running on a device since the OS is now transparent to the security issues. Mission-critical applications that require supervisory controls will be able to run on any device. This is highly beneficial and can potentially phase out the need for businesses to use only dedicated devices with authorised software stacks. Scalable Vector Extension version two (SVE2) Armv9 continues to use the AArch64, which was introduced in v8. In v9, the AArch64 is used as the baseline instruction set. The most important addition to the AArch64 architecture is Scalable Vector Extension version two (SVE2). The first implementation, Scalable Vector Extensions (SVE), was first introduced in 2016. It was developed following the Neon architecture extension, which has a fixed 128-bit vector length for the instruction set. Used first in Fujitsu’s A64FX CPU cores, SVE currently powers the Fukagu supercomputer in Japan, touted as the world’s fastest supercomputer. It is particularly suited for High-Performance Computing (HPC) applications that require large amounts of data processing. SVE2 is a superset of SVE and Neon and inherits the concept, vector registers, and operation principles of SVE. It permits more function domains in data-level parallelism. SVE and SVE2 define 32 scalable vector registers, and users can choose a suitable vector length design implementation for hardware that varies between 128 and 2048 bits. SVE2 differs from SVE due to the functional coverage of the instruction set. While SVE was designed for HPC and ML, SVE2 goes beyond and extends the SVE instruction set to enable data-processing domains beyond HPC and ML. The SVE2 instruction set can enhance standard algorithms used in applications such as computer vision, multimedia, genomics, in-memory database, web serving, long-term evolution (LTE) baseband processing and general-purpose software. SVE2 also adds a vector-width-agnostic version of the Neon instructions in most integer Digital Signal Processing (DSP) and media processing functionality. This helps compilers to vectorise more effectively for these domains. Total Compute Design Methodology Arm has boosted the CPU performance at a rate much faster than than the industry average. The ARM spokesperson said the momentum would carry over to the Armv9 generation. Users can expect a CPU performance increase of 30 percent over the next two generations of mobile and infrastructure CPUs. Credit: Arm Arm’s Total Compute design methodology is expected to improve the overall compute performance through system-level hardware and software optimisations. The Total Compute design principles will be applied across the entire IP portfolio of automotive, client, IoT, and infrastructure. Further, Arm is also developing technologies to increase frequency, bandwidth, and cache size to maximise the performance of Armv9-based CPU.","excerpt":"Chip designer Arm recently introduced its latest architecture Armv9, at the Arm Vision Day 2021 event. Arm’s last architecture announcement came in 2011, at the Armv8 launch. Armv9 offers improvements in performance, speed, and security. “As we look toward a future that will be defined by AI, we must lay a foundation of leading-edge compute […]","categories":["AI Trends"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-04-05T11:00:00","publication_year":"2021","word_count":675,"keywords":["Go","programming_languages:R","AI","ML","Scala","computer vision","RAG","Git","Rust","R"],"extracted_tech_keywords":["AI","ML","computer vision","RAG","R","Go","Rust","Scala","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/arm-announces-new-architecture-after-ten-years-major-updates-latest-features\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10115486,"title":"Infosys Founder Funds Meta’s Llama 2 Project with 22 Indian Languages","content":"Yann LeCun, the chief of Meta AI was recently on a podcast with Lex Fridman talking about Meta, AI, and AGI. He revealed that he recently met with an Infosys founder who is funding a project based on Llama 2. “He’s funding a project to fine-tune Llama 2, the open source model produced by Meta so that it speaks all 22 official languages in India. It’s very important for people in India,” said LeCun. He did not reveal which co-founder he met. LeCun was quite impressed by the recently released Kannada Llama. “I love this. This is why open source AI platforms will win: it’s the only way for AI to cater to highly diverse languages, cultures, values, and centers of interest.” he wrote on X, reposting Kannada Llama post. He also spoke about Moustapha Cisse, who used to be a scientist at Meta FAIR who is also using Llama 2 to build language models in Africa, and catering it for medical use. LeCun is a big proponent of open source AI models and recently also posted on X about the training infrastructure of Llama 3, the upcoming AI model of Meta which is expected to be open source as well. Infosys was one of the first companies that had invested in OpenAI back in 2015. It donated around USD 1 billion to OpenAI, which was a non-profit back then. Infosys, along with Elon Musk, AWS, YC Research, among others, joined hands to make the USD 1 billion donation to OpenAI. Apart from Infosys, a lot of Indian companies and startups have been leveraging Llama 2 and making Indic models such as Tamil Llama, Kannada Llama, Telugu Llama, and many more. IIT Bombay’s in work BharatGPT project is also focusing on catering to the 22 Indic languages and is pushing towards being open source.","excerpt":"Yann LeCun recently met with an Infosys founder who is funding a project based on Llama 2.","categories":["Deep Tech"],"tags":["Fund Raising","Infosys","LLaMA","Meta"],"author_name":"Mohit Pandey","publish_date":"2024-03-13T13:20:58","publication_year":"2024","word_count":304,"keywords":["Meta","Meta AI","funding","Infosys","AI","OpenAI","AWS","Fund Raising","LLaMA","RAG","GPT","llm_models:GPT","R","startup"],"extracted_tech_keywords":["AI","OpenAI","Meta AI","RAG","AWS","R","GPT","startup","funding","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/infosys-funds-llama-2-project-with-22-indian-languages\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10110011,"title":"How Truly Sustainable is the GenAI Business?","content":"Recently, it was revealed that OpenAI is expected to record $1.6 billion in annualised revenue in 2024. This is a 300 million bump up from the predicted $1.3 billion in October last year, coming from ChatGPT subscriptions, API access and others.Meanwhile, competitor Anthropic expects at least $850 million in annualised revenue in 2024 up from $100 million three months ago, which they had informed the investors about. Anthropic’s projected revenue indicates a potential surge to $70 million per month by the following year, driven by Claude, its text-generating AI, serving customers like Notion, Quora, and DuckDuckGo. On the other hand, bootstrapped AI image-generator company Midjourney generated a revenue of $200 million in 2023 from its subscription model that charges users between $10 and $120 monthly. Is it Sustainable? As these companies compete with each other in the space, the potential growth and evolution of their businesses, relative to their current capabilities, remains uncertain for various reasons. For instance, the never-ending need to raise substantial capital, with CEO Sam Altman indicating that OpenAI potentially needs $100 billion to run the business, while the financial losses mount upto $540 million. Plus, big tech companies like Microsoft, Google, and AWS which are injecting capital also get a share of the generated revenue alongside computational cost and accelerate research and development to keep up with the competition. Additionally, numerous prominent figures in the industry are now expressing doubts over whether current models will sustain the pace of advancement or have they already hit a wall. Echoing a similar sentiment, Bill Gates expressed doubts about the progress of generative AI saying that it has plateaued. And despite the optimistic view from some at OpenAI about GPT-5’s potential, Gates believes current generative AI has hit a ceiling. While acknowledging the impressive leap from GPT-2 to GPT-4, he admitted the possibility of being wrong about its limitations. ​​Ashray Malhotra, co-founder and CEO of Rephrase.ai, also shared his thoughts in a recent podcast on the cost of further innovation and the scale of it. “GPT 5, whenever they train it, [it turns out] like 10 times more expensive and only 10% better than GPT 4 and you realise that for the next big thing, you need a fundamentally different architecture than a diffusion or transformers. Then you’re back in the wilderness, like, what is that next big thing?” For How long? Insiders suggest that Anthropic is anticipated to potentially hit $1 billion in annualised revenue next year, around $83 million monthly. Contrarily, projections expect OpenAI to generate $5 billion in annualised revenue by 2024. However, while the revenue grows and the company continues to raise endlessly from investors and spend money on research and development, it would still keep them a long way from profitability. How are They Doing it? To deal with revenue issues, companies will need to look at diversifying their businesses and building in-house capabilities. This would reduce reliance on other companies, cut costs and bring in a newer stream of revenues. OpenAI, for instance, is moving ahead of just a subscription model and catering to enterprises. It is also said to be venturing into the AI server chip business for which Altman is raising money from Abu Dhabi firm, G42. Others like Stability AI, despite earlier revenue estimates exceeding $10 million annually, are changing gears from open-source by introducing a subscription fee for enterprise customers. The move towards subscriptions comes as Stability’s open-source strategy failed to yield significant income. A focus on increasing and projecting safety and reliability in the technology could also lead to applications across industries like healthcare, marketing, banking, telecom, defence and others. This would mean a strong, steady and bigger stream of revenue. Alternatively, one could stick to Midjourney co-founder David Holz’s philosophy, which stands as a testament to running a profitable bootstrapped company with sustainable growth. He owes it to an excellent product focussed on catering to a specific customer base without raising any capital from VCs. Perplexity CEO, Aravind Srinivas also recently justified their lack of a focussed strategy as it tries to enter the search business and others. He posted on X saying that “Jeff Bezos invested in Google in 1998, before it had any business model figured out or was sky-rocketing in search traffic. His investment in Perplexity is symbolically significant.”","excerpt":"Prominent figures in the industry are now expressing doubts over whether current models will sustain the pace of advancement or have they already hit a wall.","categories":["AI Features"],"tags":["AI Tool","OpenAI"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-01-05T10:36:04","publication_year":"2024","word_count":715,"keywords":["Anthropic","ChatGPT","GPT-5","OpenAI","AI","AWS","Transformers","Ray","generative AI","AI Tool","R"],"extracted_tech_keywords":["AI","generative AI","GPT-5","ChatGPT","OpenAI","Anthropic","Ray","Transformers","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-truly-sustainable-is-the-genai-business\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":18099,"title":"These 5 Recent Patents Give A Peek Into How Apple Is Much Ahead Of Its Competition In AI","content":"The recently concluded iPhone X launch event in California was special in more ways than one. Launched at the Steve Jobs theatre, it celebrated the 10th anniversary of the iconic iPhone and also announced the launch dates of the iPhone 8, 8 Plus and iPhoneX. For us, the more important takeaways of the glitzy event helmed by Apple’s CEO Tim Cook, were the use of neural networks on a smartphone and machine learning capabilities in the new camera. Not to forget the new FaceID capability, the facial recognition feature to replace the Touch ID fingerprint sensor. While it’s too late in the day to talk about the event’s key takeaways, there is one dominant theme – the future of iPhone is paved by artificial intelligence with the Cupertino giant betting big on next-gen technology. New York headquartered CB Insights takes a look at Apple’s patents to predict on what bleeding edge technologies the iPhone maker is betting big on the future. CB Insights has built a tech market intelligence platform that analyses millions of data points on venture capital, startups, patents, partnerships and news media to predict technology trends. Apple’s patents were uncovered via CB Insights Patent Search Engine. Here’s a roundup of a few AI\/ML\/AR patents that caught our attention and could possibly make way into Apple’s devices Iris Recognition: Trust Apple to stay one step ahead of its competitors. While other smartphone makers are betting big on facial recognition, Apple is mulling a Iris Recognition capability. The patent for “Using biometric verification to grant access to redacted content” granted in 2017 has made a mention of iris reading as a biometric identifier. According to the patent, the biometric reader reads a person’s uniquely identifying biometric data (e.g., thumbprint\/fingerprint, iris scan, voice). This biometric information is then read by the biometric verification engine for comparison to a stored set of verified user biometric data. When the biometric data matches the stored data, the person is verified. Facial Recognition: An improvement on the fingerprint based TouchID, FaceID that will soon debut in the much-awaited iPhone X could become a key differentiator for the iPhones. The AI-powered facial recognition technology boasts of human-level accuracy and can recognize the user accurately despite changed hair styles, facial grooming and ageing. Interestingly, the smartphone maker secured 5 grants and 2 outstanding applications patents in facial recognition. The patent — Locking and unlocking a mobile device using facial recognition was filed in 2011. The iPhone is outfitted with a front-facing camera that captures the image, an initial sensor and a processor that causes the camera to capture an initial image while the device is locked. The facial recognition capability is fueled by in-built neural engines and can even detect facial features in the dark with 33,000 infrared dots. Neural Networks for converting speech to text: Now, its patent Applying neural network language models to weighted finite state transducers for automatic speech recognition talks about systems and processes for converting speech to text. The summary of the patent indicates that text corresponding to the speech input can be determined based on the probability of the candidate word given the one or more history candidate words. An output can be provided based on the text corresponding to the speech input. The patent was filed in 2016 and having this capability on board could be a major differentiator for Apple. AI Hardware: AIM had hinted earlier how Apple is developing AI-powered chips to boost the computation technologies across its devices. And as hinted, Apple has famously packed the AI chip into the iPhoneX. The custom built chip will speed up AI workloads and as Apple’s Philip Schiller indicates the dual-core “A11 bionic neural engine” chip can carry out 600 billion operations per second. Not to forget the facial recognition feature for unlocking and facilitating ApplePay on the iPhone X. Now, with a processor tuned for AI, the latest iPhone will see a significantly improved battery life and scores a lead over its smartphone rivals. Augmented Reality Maps: Much has been written about the Augmented Reality Maps and Apple’s push into AR. Augmented Reality systems can enhance information of a real environment by providing a visualization of computer-generated virtual information with a part of the real environment. According to CB Insights, Apple has 3 applications and 6 patent grants in augmented reality and AR apps may also debut in the new iPhone 8 series.  Earlier in 2015, the Cupertino giant bought Munich headquartered Metaio, an augmented reality startup, making its AR ambitions clear. AR maps overlaid on a real environment on the user’s screen could add more interactive functionality and pitches Apple in the same league as Google, which is also betting heavily on AR. Outlook For long, analysts believed Apple was falling behind in the AI race since the company was known to keep developments under a wrap. However, the recent WWDC has changed a lot of things and Cook has made it abundantly clear how AI, machine learning, augmented reality and real time motion analysis technology will make Apple’s devices more intelligent, powerful and secure. While the future of iPhones is definitely AI, Apple will invest more money and human muscle in pushing the envelope.","excerpt":"The recently concluded iPhone X launch event in California was special in more ways than one. Launched at the Steve Jobs theatre, it celebrated the 10th anniversary of the iconic iPhone and also announced the launch dates of the iPhone 8, 8 Plus and iPhoneX. For us, the more important takeaways of the glitzy event […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-10-05T12:40:11","publication_year":"2017","word_count":868,"keywords":["Go","machine learning","artificial intelligence","TPU","AI","neural network","ML","Aim","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","Aim","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/it-services\/5-recent-patents-give-peek-apple-much-ahead-competition-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":19329,"title":"5 Reasons To Do A Big Data Certification","content":"Innovation, competition, and productivity – these are the buzzwords that define the direction of business today, much of which is being driven by Big Data and Analytics. Data is emerging as a critical corporate asset, redefining core operations, core competencies and business models, resulting in a considerable increase in the number and variety of Big Data certification courses available online. Securing the right talent to fill skill gaps is the biggest hurdle that companies face while trying to integrate data and analytics into their existing operations. Here’s where a certification can play a key role in helping existing personnel within a company acquire these crucial skills, expanding their repertoire and helping them fulfill multiple roles within the organization framework. Merits of pursuing a Big Data Certification Being certified is a great way to make a career shift to a promising field. Certifications have long since been a reliable proof of competency and an accepted norm in the fast-changing technology world. When technologies keep changing every few months, it is challenging to complete full-fledged degree courses and become competent. Certifications are the way to go, acting as speedy and practical “crash courses” to build the required talent. So here is why you should consider learning the tricks of the data and analytics trade, by doing a big data certification, today! Positive optimism about the technology: An IDC forecast states that the Big Data market is predicted to be worth $46.34 billion by 2018, with great prospects in the areas of Big Data related infrastructure, software and services over the next five years. IDC has also predicted an annual growth rate of 23.1% over the period of 2014-2019 and annual spends to reach $48.6 billion in 2019. These numbers ooze optimism. In fact, the public sector too is awakening to the advantages of data and analytics, and formulating a number of information security policies and programs to give impetus to technological and skill-building investments. Big Data is now seen moving out of the experimental phase and taking on real projects. Strong salary prospects: The growing attention to Big Data and Analytics is creating a talent flux- organizations want to propel forward their Big Data expertise, but the market just does not have the volumes and\/or quality of talent to meet these newfound burgeoning demands. The International Data Corporation (IDC)3 has predicted a talent need by 2018: 181,000 people with deep analytical skills, and five times for data management and interpretation skills. This skew in demand-supply gives data professionals a financial-edge over run-of-the-mill IT professionals. A report by Analytics India Magazine in association with AnalytixLabs indicates that the average salaries of Analytics professionals in India stand at Rs. 11.7 lakhs per annum, whereas that of IT professionals is at Rs. 8.65 lakh per annum, Moreover, only 39% of analytics professionals have salaries under 6 Lakhs vs 58% in IT. Thus, you stand to gain financially, provided you have the required professional skills. So, learn all you can and make a financially lucrative career switch! Career progression through Big Data certifications: The gain is not only immediate, there is also a long-term logic in learning the ropes of Big Data and Analytics. Doing a Big Data certification is the way to leapfrog your career from a traditional IT job to a futuristic Data Analytics profession. Moreover, you can choose from a number of Big Data certifications according to the career path you wish to adopt- Architecture, Data Science, Business Analytics, Big Data Development (Hadoop, R, Python etc.), Administration- the options are many to suit your preference. Most of these courses have a skill-level based progression from beginner to intermediate to expert, allowing you learning-access to progress fast in your new career. For example, the Big Data Hadoop Developer Certification by Simplilearn helps learners master the concepts of the Hadoop framework, a critical Big Data and Analytics competency. You only need a short learning investment: What’s more, there is no need take a career break, go back to college and spend years gaining a degree or diploma in Big Data and Analytics. Big Data and Analytics is a fast-evolving technology – there are little or no pre-requisites in terms of formal education, making it easy for any IT professional to dabble in the field and gain expertise in a matter of few months by doing a Big Data certification Today the learning market offers a range of courses to suit the busy professional. From online learning modules, to self-paced learning, learning has evolved into an experience in itself, and deliver the right market-led skills to help career professionals plug-and-perform into the skill-gaps. Learn while you earn, and advance your career through continuous learning! Applicability across various industries: Almost every sector today is looking at leveraging the power of big data and analytics. While industry domain expertise is important, becoming a certified Big Data and Analytics expert means that you have a variety of job opportunities, across industries and need not be constrained by sector. This is important for remaining relevant and flexible in a cut-throat talent market. What should you learn? It is important to choose a Big Data certification that suits your career aspirations. A number of career paths are available within the Big Data and Analytics stream- Developer, Data Architect, Business Analyst, Solutions and Services Expert (for specific branded products), Database Administrator, Data Scientist, Data Engineer, and so on. You need to choose the certification after analyzing what suits you in terms of interest, career progression, opportunities etc. One such learning-solutions provider is Simplilearn, with its wide variety of Big Data certifications such as Big Data Hadoop Architect, Data Scientist, Big Data Hadoop and Spark Developer, Big Data and Hadoop Administrator, etc. The authenticity and usefulness of Simplilearn courses stem from the fact that Ronald van Loon, one of the Top 10 global Big Data influencers, sits on the advisory board and acts as the Course Advisor for Simplilearn. In addition to this, Simplilearn has also launched the JobAssistTM program to help certified learners find relevant jobs in the domain. So why wait? Choose your go-to Big Data certification wisely and embark on a refreshing and promising learning path to realize your career dreams.","excerpt":"Innovation, competition, and productivity – these are the buzzwords that define the direction of business today, much of which is being driven by Big Data and Analytics. Data is emerging as a critical corporate asset, redefining core operations, core competencies and business models, resulting in a considerable increase in the number and variety of Big […]","categories":["AI Trends"],"tags":["big data certification"],"author_name":"Дарья","publish_date":"2017-11-30T06:18:30","publication_year":"2017","word_count":1031,"keywords":["big data","data science","Go","big data certification","AI","RAG","Python","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","RAG","Python","R","Go","big data","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-reasons-to-do-a-big-data-certification\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10123411,"title":"Databricks Enhances Mosaic AI for Enterprise-Ready AI Applications","content":"Databricks today announced several innovations to its Mosaic AI platform to help customers build production-quality generative AI applications. The company is investing in three key areas: support for building compound AI systems, capabilities to improve model quality, and new AI governance tools. Organisations are struggling to transition generative AI projects from pilot to full-scale production due to privacy, quality, and cost concerns. While foundation models have significantly improved, they still face challenges in producing consistently high-quality results. To address these issues, organisations are moving beyond deploying a single large model and instead adopting compound AI systems. This approach utilises multiple components, including various models, retrievers, vector databases, and tools for evaluation, monitoring, security, and governance, resulting in higher production quality and more accurate, safe, and governed AI applications. “We believe that compound AI systems will be the best way to maximise the quality, reliability, and measurement of AI applications going forward, and may be one of the most important trends in AI in 2024,” said Matei Zaharia, co-founder and CTO at Databricks. “Databricks is uniquely positioned to capitalise on these trends with the investments we’re making to improve quality, augmenting the model with real-time data and agents and tools to give it new capabilities it has little knowledge of.” To support customers in building production-quality generative AI applications, Databricks is launching several new features: Mosaic AI Agent Framework and Mosaic AI Tools Catalog help organisations build compound AI systems. The Agent Framework enables developers to quickly and safely build high-quality RAG (Retrieval-Augmented Generation) applications using foundation models and enterprise data. The Tools Catalog allows organisations to govern, share, and register tools using Databricks Unity Catalog, ensuring secure and governed use of tool-enabled models. Mosaic AI Agent Evaluation is an AI-assisted evaluation tool that automatically determines if outputs are high-quality and provides an intuitive UI for gathering feedback from human stakeholders. This helps organisations deploy production-quality generative AI solutions. Mosaic AI Model Training enables fine-tuning of open-source foundation models with an organisation’s private data, resulting in higher-quality results for specific use cases. These fine-tuned models are fully owned and controlled by the customer, and are faster and less expensive to serve compared to larger proprietary models. Mosaic AI Gateway provides a unified interface to query, manage, and deploy any open-source or proprietary model, allowing customers to easily switch the large language models (LLMs) powering their applications without complex code changes. It offers usage tracking, guardrails, governance, and monitoring to ensure quality and control spending. Several Databricks customers, including Corning, Ford Direct, and Lippert, have already benefited from these new capabilities in building their generative AI applications. By leveraging the Databricks Data Intelligence Platform and Mosaic AI, they have improved retrieval speed, response quality, accuracy, and confidence in deploying to production. The new Mosaic AI capabilities are part of Databricks’ ongoing commitment to helping customers harness the power of generative AI while maintaining data privacy, quality, and cost-effectiveness. As organisations continue to explore AI’s potential, Databricks aims to provide the tools and platform necessary for building enterprise-ready AI applications.","excerpt":"The new Mosaic AI capabilities are part of Databricks’ ongoing commitment to helping customers harness the power of genAI while maintaining data privacy, quality, and cost-effectiveness.","categories":["AI News"],"tags":["Databricks"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-12T18:52:34","publication_year":"2024","word_count":507,"keywords":["Go","TPU","AI","RAG","vector databases","Aim","generative AI","foundation models","R","Databricks"],"extracted_tech_keywords":["AI","generative AI","foundation models","Aim","RAG","vector databases","TPU","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/databricks-enhances-mosaic-ai-for-enterprise-ready-ai-applications\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164508,"title":"How AI is Becoming a Creative Ally for Freelancers","content":"The emergence of AI chatbots like ChatGPT and Google Gemini, has sparked speculation around the future of freelancers. This is especially true given that OpenAI has suggested that large language models (LLMs) could earn $1 million from freelance software engineering tasks, raising concerns about the potential disruption to the gig economy. However, it seems freelancers aren’t being replaced by AI; instead, they are adopting it to boost their workflow. Freelancers aren’t just using AI tools; they are focusing on improving their efficiency and choosing freelance gig platforms that allow seamless AI integration to elevate their work. Curious about how this shift is happening? We dive into the details right here. AI as a Creative Partner AI helps brainstorm and automate tasks, even with the simplest prompts. This benefits freelancers who must constantly innovate and exercise their creativity to stand out. As a creative partner, AI goes beyond improving efficiency; it enhances skill sets and generates a wide variety of ideas at an accelerated pace. From writing tools that generate ideas and refine drafts to image editing apps that simplify the process, there are countless applications of creative AI tools available out there. Freelancers use tools like Grammarly, Canva, Adobe Firefly, ChatGPT, Notion and GitHub Copilot, among others, to brainstorm, improve their creation time, and explore ideas they wouldn’t have otherwise considered. Adrian Stubbs, a podcast creator, mentioned, “I’ve been building specific custom GPTs to automate my freelance workflow, and the results speak for themselves. Freelancing is way smoother when your AI assistant handles the boring stuff.” He also highlighted the benefits of using AI tools, which involve faster invoice generation, project tracking, and quick time stamp analysis for podcasts. Meanwhile, Karmen Kendrick, a website designer and marketer, said, “I ain’t rich, but I’ve cut down my admin time with it. I build websites for people, and I’ve been able to use AI to transcribe my calls, upload the transcript to ChatGPT and spit out a proposal in minutes. What used to take me hours only takes me 20 minutes. And that’s valuable as a one-person shop.” Using Platforms Embracing the AI Revolution Not just limited to the freelancers using AI in an individual capacity, freelance gig platforms that let anyone enrol as a freelancer have also started offering AI tools to equip the customers and the freelancers with the power of AI. Hence, freelancers have started joining such platforms that have started adopting AI. For instance, Fiverr’s AI platform, Fiverr Go, offers AI capabilities for creators and customers. The platform also includes a Personal AI Creation Model that allows freelancers to train AI models on their own work. The AI model gives the creator complete control and ensures that they are compensated for the AI-enhanced capabilities. This model is available across multiple categories of services that include voiceover, songwriting, graphic design, illustration, copywriting, and digital marketing. Moreover, a personal AI assistant makes client communication easy and manages routine tasks faster. On this, Micha Kaufman, CEO and founder of Fiverr, stated, “This isn’t just another AI platform; it’s a fundamental reimagining of how AI and human creativity can work together.” “Instead of letting creators be exploited, Fiverr Go ensures they receive proper credit and compensation while giving them unprecedented tools to scale their work. This is about making our freelancers irreplaceable, not obsolete,” he added. Similarly, Upwork worked with OpenAI to integrate AI capabilities into its platform, Uma. With it, customers can use AI to generate job requirements, find freelancers, and brainstorm and write or code for freelancers. “For freelancers, Uma has been equally transformative. It is a work companion embedded into the Upwork experience designed to help independent professionals solve complex tasks, boost productivity, and deliver their best work faster through instant, intelligent support,” Upwork told AIM in a statement. “Uma can also create tailored proposal drafts, leveraging custom AI models trained on winning proposals, to help them stand out and win more work. This is how Uma enables independent professionals to grow their businesses, access more work opportunities, and ultimately earn more, a direct reflection of our core mission,” the company further said. The Future of Freelancing is To Be in Sync With AI The integration of AI into freelance workflows is still an ongoing process, but the potential is enormous. As AI technology continues to develop, we can expect to see even more sophisticated tools emerge, further empowering freelancers and transforming the freelance landscape. From AI-powered project management tools that predict potential roadblocks to intelligent matchmaking systems that connect freelancers with ideal clients, the possibilities are endless. It seems like a good direction to take, as companies are trying to sync humans and AI tools to maximise output and improve customer experience. The Human Element Remains Crucial While AI is undoubtedly a powerful tool, it’s important to remember that the human element remains crucial. AI is a tool to be wielded by creative professionals, not a replacement for them. The initiatives by the freelance marketplace platform show signs of supporting the same. The unique insights, creativity, and problem-solving skills that human freelancers bring will continue to be essential. It remains to be seen how the freelancing world evolves with the advancement of AI.","excerpt":"Freelancers are keeping up with technological advancements by partnering with AI.","categories":["AI Features"],"tags":["AI Tool","Freelance"],"author_name":"Ankush Das","publish_date":"2025-02-25T11:28:05","publication_year":"2025","word_count":865,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","chatbots","Freelance","ML","RAG","Aim","AI Tool","R"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Aim","RAG","chatbots","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-ai-is-becoming-a-creative-ally-for-freelancers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10005591,"title":"Random Forest Vs XGBoost – Comparing Tree-Based Algorithms (With Codes)","content":"In machine learning, we mainly deal with two kinds of problems that are classification and regression. There are several different types of algorithms for both tasks. But we need to pick that algorithm whose performance is good on the respective data. Ensemble methods like Random Forest, Decision Tree, XGboost algorithms have shown very good results when we talk about classification. These algorithms give high accuracy at fast speed. Both the two algorithms Random Forest and XGboost are majorly used in Kaggle competition to achieve higher accuracy that simple to use. Through this article, we will explore both XGboost and Random Forest algorithms and compare their implementation and performance. We will see how these algorithms work and then we will build classification models based on these algorithms on Pima Indians Diabetes Data where we will classify whether the patient is diabetic or not. We will then evaluate both the models and compare the results. The dataset can be downloaded from Kaggle. What we will learn from the article? What is the Random Forest Algorithm and how does it work?What is XGboost Algorithm and how does it work?A comprehensive study of Random Forest and XGBoost AlgorithmsPractically comparing Random Forest and XGBoost Algorithms in classification What is the Random Forest Algorithm? How does it work? The forest is said to robust when there are a lot of trees in the forest. Random Forest is an ensemble technique that is a tree-based algorithm. The process of fitting no decision trees on different subsample and then taking out the average to increase the performance of the model is called “Random Forest”. Suppose we have to go on a vacation to someplace. Before going to the destination we vote for the place where we want to go. Once we have voted for the destination then we choose hotels, etc. And then come back with the final choice of hotel as well. The whole process of getting the vote for the place to the hotel is nothing but a Random Forest Algorithm. This is the way the algorithm works and the reason it is preferred over all other algorithms because of its ability to give high accuracy and to prevent overfitting by making use of more trees. There are several different hyperparameters like no trees, depth of trees, jobs, etc in this algorithm. Check here the Sci-kit documentation for the same. from sklearn.ensemble import RandomForestClassifier rfcl = RandomForestClassifier() What is XGBoost Algorithm? How does it work? XGBoost is termed as Extreme Gradient Boosting Algorithm which is again an ensemble method that works by boosting trees. XGboost makes use of a gradient descent algorithm which is the reason that it is called Gradient Boosting. The whole idea is to correct the previous mistake done by the model, learn from it and its next step improves the performance. The previous results are rectified and performance is enhanced. This gets continued until there is no scope of further improvements. Regularization is the feature that is dominant for this type of predictive algorithm. It is fast to execute and gives good accuracy. This algorithm is commonly used in Kaggle Competitions due to the ability to handle missing values and prevent overfitting. There are again a lot of hyperparameters that are used in this type of algorithm like a booster, learning rate, objective, etc. Check the documentation to know more about the algorithm and hyperparameters. from xgboost import XGBClassifier xgbcl = XGBClassifier() How to Build a Classification Model using Random Forest and XGboost? First, we will define all the required libraries and the data set. Use the below code for the same. import pandas as pd from sklearn.ensemble import RandomForestClassifier from xgboost import XGBClassifier from sklearn.model_selection import train_test_split from sklearn.metrics import accuracy_score,classification_report data = pd.read_csv('\/content\/pima-indians-diabetes-1.csv') We will check what is there in the data and its shape. Refer to the below code for the same. print(data) print(data.shape) Output: Now we will define the dependent and independent features X and y respectively. We will then divide the dataset into training and testing sets. Use the below code for the same. X = data.drop('class',axis = 1) y= data['class'] X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33) print(X_train.shape) print(X_test.shape) Output: There are 514 rows in the training set and 254 rows in the testing set. Now we will fit the training data on both the model built by random forest and xgboost using default parameters. Then we will compute prediction over the testing data by both the models. rfcl.fit(X_train,y_train) xgbcl.fit(X_train,y_train) y_rfcl = rfcl.predict(X_test) y_xgbcl = xgbcl.predict(X_test) We have stored the prediction on testing data for both the models in y_rfcl and y_xgbcl. Now we will evaluate the model performance to check how much the model is able to generalize. We will make use of evaluation metrics like accuracy score and classification report from sklearn. print(\"Random Forest Accuracy: \", accuracy_score(y_rfcl,y_test)) print(\"XGBoost Accuracy: \", accuracy_score(y_xgbcl,y_test)) Output: Random Forest Accuracy: 0.79 XGBoost Accuracy: 0.80 print(\"Random Forest: \\n\", classification_report(y_rfcl,y_test)) print(\"\\nXGBoost: \\n\", classification_report(y_xgbcl,y_test)) Output: Also, check this “Practical Guide To Model Evaluation and Error Metrics” to know more about validating the performance of a machine learning model. Conclusion Through this article, we discussed the Random Forest Algorithm and Xgboost Algorithm with the working. Also, we implemented a classification model for the Pima Indian Diabetes data set using both the algorithms. We did not even normalize the data and directly fed it to the model still we were able to get 80%. If we work more on data and feature engineering then this accuracy can be improved further. Also, hyperparameters can be tuned using different methods. Both the algorithms work efficiently even if we have missing values in the dateset and prevent the model from getting over fitted and easy to implement.","excerpt":"Through this article, we will explore both XGboost and Random Forest algorithms and compare their implementation and performance.","categories":["Deep Tech"],"tags":["decision tree algorithm","ensemble method","gradient boosting","Machine Learning","object store database","random forest","VS Code","XGBoost"],"author_name":"Rohit Dwivedi","publish_date":"2020-08-26T14:00:00","publication_year":"2020","word_count":951,"keywords":["Go","machine learning","decision tree algorithm","TPU","AI","RPA","Machine Learning","feature engineering","RAG","random forest","gradient boosting","XGBoost","ensemble method","VS Code","R","object store database","Pandas"],"extracted_tech_keywords":["AI","machine learning","XGBoost","Pandas","RAG","TPU","R","Go","feature engineering","RPA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/random-forest-vs-xgboost-comparing-tree-based-algorithms-with-codes\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136863,"title":"Why is Everyone Bashing Microsoft Copilot?","content":"Salesforce co-founder and CEO Marc Benioff recently criticised the copilot products offered by Microsoft and other vendors, while promoting their agentic approach instead—stating a clear winner in the battle for enterprise artificial intelligence. “We all know now that Microsoft Copilot is basically the new Microsoft Clippy, that customers have not gotten value from it,” he said, taking a dig at the product. Benioff directly criticised Microsoft and OpenAI, which has received billions in investments from Microsoft and supports some of its AI products. He also targeted Microsoft’s business practices, particularly in relation to its treatment of Slack, which Salesforce acquired in 2021. The End of Copilot Era? With the advent of ChatGPT, many companies such as OpenAI, Microsoft, started providing these chatbots as a service to enterprises. Microsoft, in fact, integrated these chatbots in almost every range of products and popularised the term Copilot. However, a Business Insider report from earlier this year revealed that not all customers are happy with Copilot. “Every time a customer starts using it, they start comparing it to ChatGPT and saying, ‘Aren’t you guys using the same technology?’ an Microsoft employee said. Employees have also noted that even minor changes in prompts can lead to significant variations in results, making it frustrating to use for tasks that require reliability and it was even reported that Microsoft Copilot does not perform as well as expected compared to other AI tools like ChatGPT. The company has also fine-tuned Copilot by limiting conversation length and steerability, which some employees find hampers its effectiveness. As a result, they view Copilot as cumbersome and less intuitive compared to other AI solutions they are accustomed to, leading them to question its usability. Some employees also said some users are also unskilled at writing prompts: “If you don’t ask the right question, it will still do its best to give you the right answer and it can assume things.” In March, Microsoft launched Copilot for Security, the AI industry’s “first generative AI solution” designed for security and IT professionals. It uses vast amounts of data and threat intelligence, processing over 78 trillion security signals daily. A company study showed that security analysts using Copilot for Security worked 22% faster, and 7% reported enhanced accuracy in their tasks. Netizens were quick to create and share their opinions and experiences highlighting the unique situation. ChatGPT’s popularity, with over 200 million weekly active users, can be attributed to its longer presence in the market. In contrast, Microsoft’s Copilot app struggled with discoverability due to poor promotion and a subtle mobile launch. Microsoft also paused new Copilot releases to focus on improving existing features, later announcing new offerings like Copilot Pages and Copilot agents. Interestingly, like Salesforce, and a bunch of other companies, Microsoft too, has entered the world of AI-powered agents with an update to its Microsoft 365 Copilot product that gives users the ability to quickly create a variety of AI assistants that can carry out tasks across both the it’s own software and that of third-party vendors. This marks a significant shift in Copilot’s functionality, aligning with what the industry refers to as AI agents– the capability for chatbots to autonomously and intelligently handle complex tasks. “We very quickly realised that constraining Copilot to just being conversational was extremely limiting in what Copilot can do today,” said Charles Lamanna, Corporate Vice President, Business & Industry Copilot at Microsoft. Reddit hates Copilot Although the backlash against Copilot by Reddit users feels exaggerated. While criticisms around AI integration in tools are valid, many seem quick to dismiss its potential benefits without maybe fully understanding its capabilities. Copilot has the potential to streamline workflows, improve productivity, and reduce mundane tasks. Although, the fears about job replacement or privacy concerns are not unique to this tool—they’re part of a broader debate. Instead of outright bashing, offering feedback that helps improve the tool while still acknowledging its positive impact on efficiency and innovation. Some users even had to say, “A good product is self explanatory – it doesn’t have instructions.  Microsoft has always failed at UI. Maybe it’s because they are such an old company, the staff are full of oldies with outdated ideas.” The negativity surrounding Copilot feels disproportionate. Many users seem to jump on the bandwagon of criticising it without fully exploring its potential. Reacting to this, one user expressed disappointment with Microsoft’s current version of Bing Copilot, arguing that it lacks the sophistication of earlier models, such as the one nicknamed “Sidney.” These early versions were considered more intelligent and responsive, leading some to feel that Microsoft “dumbed it down” over time. According to this perspective, Microsoft may have been overly cautious, fearing that users would misuse the technology for inappropriate purposes. This led to a more constrained and limited tool, which some believe lacks the appeal of ChatGPT.","excerpt":"Employees have noted that even minor changes in prompts can lead to significant variations in results, making it frustrating to use for tasks that require reliability.","categories":["AI Features"],"tags":["Copilot","Microsoft"],"author_name":"Tarunya S","publish_date":"2024-09-30T13:23:39","publication_year":"2024","word_count":805,"keywords":["Go","ChatGPT","AI assistants","artificial intelligence","OpenAI","AI","chatbots","ML","generative AI","Copilot","R","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","ML","generative AI","ChatGPT","OpenAI","AI assistants","chatbots","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-everyone-bashing-microsoft-copilot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":44784,"title":"EagleView Sets Up R&D Centre In Bengaluru, Plans To Create 500 New Jobs","content":"EagleView, an aerial imagery and computer vision analytics, announced the official opening of its new international office in Bengaluru. The company has plans to leveraging India’s strong technical and entrepreneurial talent pool to fuel the company’s continued global expansion. Rishi Daga, CEO at EagleView, said, “EagleView has always had a culture of delivering innovation through the hustle, hard work, and raw potential of our people and we view Bengaluru as a catalyst for this continued innovation. We are excited to establish our largest R&D centre to further expand our capabilities in the areas of cloud services, computer vision, machine learning, and aerial imagery in Bengaluru.” Prasan Sathyamoorthy, General Manager at EagleView India, said, “My colleagues and I are proud to be part of EagleView, where we have unique opportunities to work on groundbreaking technologies like machine learning, computer vision, drone flight automation, and SaaS solutions in order to digitize workflows. Not only is EagleView revolutionizing aerial imagery and virtual inspections, but our work can help improve and even save lives by contributing to critical disaster response efforts.” Located within Bengaluru’s Pritech Park special economic zone (SEZ), EagleView India’s new office joins the company’s existing locations in Bellevue, Washington, Rochester, New York, and Perth, Australia, as a world-class facility that will be staffed with leading software engineering talent. EagleView plans to create 500 new jobs in Bengaluru within the next year, where teams will be charged with researching and developing new products.","excerpt":"EagleView, an aerial imagery and computer vision analytics, announced the official opening of its new international office in Bengaluru. The company has plans to leveraging India’s strong technical and entrepreneurial talent pool to fuel the company’s continued global expansion. Rishi Daga, CEO at EagleView, said, “EagleView has always had a culture of delivering innovation through the […]","categories":["AI News"],"tags":["Computer Vision","Data Analytics"],"author_name":"Prajakta Hebbar","publish_date":"2019-08-21T12:46:59","publication_year":"2019","word_count":241,"keywords":["machine learning","programming_languages:R","AI","innovation","Git","computer vision","RAG","automation","analytics","Computer Vision","Data Analytics","R"],"extracted_tech_keywords":["AI","machine learning","computer vision","analytics","RAG","R","Git","automation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/eagleview-sets-up-rd-centre-in-bengaluru-plans-to-create-500-new-jobs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10090129,"title":"Data Science Hiring and Interview Process at NoBroker","content":"From a humble beginning in 2014, NoBroker, the brainchild of IIT and IIM graduates Amit Kumar Agarwal, Akhil Gupta and Saurabh Garg, rose to become India’s first prop-tech unicorn. With a whopping $210 million raised in funding, this startup has truly disrupted the traditional real estate scene in India. NoBroker offers a range of services to property owners, tenants, and buyers, making it easy for them to connect and find their dream homes without the hassle of brokerage. So far, NoBroker has helped its clients save a whopping Rs 11,000 crore in brokerage fees on buy\/sell\/rent transactions. The company recently opened its first overseas office in Dubai. The different product lines of NoBroker are: NoBroker Pay – a platform for rent payments, NoBroker Hood – providing local information on amenities and services, NoBroker Home Services – offering services for household needs like cleaning and plumbing, NoBroker Assist – a personalised service to help find and finalise properties based on preferences and budget, and NoBroker Bazaar – an online marketplace for property transactions. Read more: Data Science Hiring Process at Kyndryl (Left to right) Akhil Gupta, Saurabh Garg, Amit Kumar Agarwal Besides being backed by Tiger Global, General Atlantic, and Moore Strategic Ventures, NoBroker has also caught the attention of big tech Google with its innovation. The tech giant has made a strategic investment of $5 million in the company. “With Google (onboard), it’s like a validation of what we are doing. Through this partnership with Google, we aim to leverage their technical expertise to create unique solutions for the real estate industry, with a special focus on NoBrokerhood,” co-founder and chief tech and product officer Akhil Gupta told AIM in an exclusive interview. The IIT Bombay alumni spearheaded senior tech roles at Peoplefluent and Oracle before founding NoBroker. On Expansion Mode NoBroker has four positions open for ML engineer (NLP – Indic Languages), data engineer and ML ops engineer. Inside NoBroker’s AI & Analytics Play “We’re a technology company and didn’t want to rely on people to solve problems. Data science is one of our pillars to success,” Gupta said, highlighting the importance of AI at NoBroker. With over 35 data scientists in their tech team, NoBroker considers data science as one of their keys to success. Besides using cutting-edge technology, they have developed their own products as well. NoBroker focuses on building automated solutions using AI techniques like vision and NLP. The start-up has over 1,000 agents assisting customers in finding houses, and records over 7,000 hours of call centre conversations daily. To validate these conversations, they have developed a transcription project that works in multiple languages with high accuracy. They also utilise NLP to identify positive and negative customer moments on calls and have launched a separate product called ‘Callzen.AI’, which has applications beyond NoBroker and is being sold as a SaaS platform to other companies. Besides, the prop-tech also has a rent and price prediction engine called ‘Rentometer’ and ‘NB Estimate’. The team provides a personalised recommendation engine, which takes into account the seeker’s preferences and important factors like location, commute time, and ideal rent. The team also uses vision analytics to handle a huge number of property images and create immersive property videos from uploaded images. Their automated system ‘IRIS Internal’ checks if the posted images are genuine. The company has developed proprietary technology that uses user behaviour analytics and AI to identify fraudulent listings and ban them from the platform. “As a result, 99.99% of the properties listed on the platform are genuine, creating a safe and productive space for property transactions,” added Gupta. The real estate giant has also created their own data lake, called ‘Starship’, which is almost 500 GB in size and includes all the data generated by NoBroker, including call data, analytics data, and business operations data and can be used to make queries and business intelligence decisions. The Interview Process About the hiring process for data science experts, Gupta said, “While looking at candidates, we prioritise their technology stack and cultural fit. We have a family-like environment and look to grow together”. Working with large-scale databases, transformation, language models, and NLP can add value to a candidate’s profile, but NoBroker would want the ideal candidate to understand the customer’s pain points and have a strong foundation to solve them. They value the ability to learn fast, implement quickly, fail fast, and learn from mistakes. For senior positions, the company prefers to promote from within, for example, Zaher Abdul Azeez, who started as a fresher, now holds a director position in the company’s data science team. The company looks for the right attitude and aptitude in candidates and is willing to support them in their career growth. Expectations “When we interview candidates, we make sure that we are adding value to their goals. It becomes a one-way relationship if we can’t do justice to their expectations,” said Gupta. But at the same time, the company expects the candidates to be willing to learn and grow with the company, and possess qualities such as ownership, teamwork, and mentorship. Work Culture Over the years, NoBroker has built a work culture based on ownership and transparency. NoBroker believes in creating a transparent and inclusive workplace where employees have access to all business data and metrics. It has recently been recognised by workplace culture consultant Great Place to Work India. “We strive to treat their employees like family, while providing a good salary structure, flexible working hours, and an ESOP policy that allows employees to own a part of the company,” concluded Gupta. So, looking for an opportunity to hop into a team that’s always on the move? Take a peek at their careers page and check if it’s the perfect match for you. Read more: Meet the Tech Kins","excerpt":"NoBroker has 4 positions open for ML engineer, data engineer and MLOps engineer","categories":["AI Hirings"],"tags":["Career","Data Science Hiring","Top Trend"],"author_name":"Shritama Saha","publish_date":"2024-07-29T14:57:04","publication_year":"2024","word_count":965,"keywords":["Top Trend","data science","Go","AI","ML","data lake","RAG","NLP","Data Science Hiring","Aim","analytics","R","Career"],"extracted_tech_keywords":["AI","ML","NLP","data science","analytics","Aim","RAG","R","Go","data lake"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/data-science-hiring-process-at-nobroker\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10098967,"title":"NVIDIA Soars on Generative AI, Reports $6 Billion Profit","content":"NVIDIA on Wednesday reported financial results for the second quarter ending on July 30 2023. The company’s revenue for this quarter soared to $13.51 billion, marking a phenomenal 101% increase compared to the same period last year and an impressive 88% rise from the previous quarter. The Data Center segment stood out as a driving force, contributing a staggering $10.32 billion in revenue for this quarter reflecting a remarkable 171% growth from the same period last year and an astounding 141% surge from the previous quarter. The company, who is the main supplier of A100 and H100 AI chips used to build and run AI applications, raked in $ 6.19 billion in profit in Q2 as compared to $ 656 million last year in corresponding quarter, up 843%. “A new computing era has begun. Companies worldwide are transitioning from general-purpose to accelerated computing and generative AI,” said Jensen Huang, founder and CEO of NVIDIA. Huang also highlighted noteworthy industry developments during this quarter. Major cloud service providers unveiled significant NVIDIA H100 AI infrastructures, and key enterprise IT system and software providers forged partnerships to bring NVIDIA AI capabilities to various industries. NVIDIA predicts that its revenue for the third quarter of the fiscal year will be around $16 billion, surpassing the estimated $12.61 billion by Refinitiv. NVIDIA’s projection indicates a remarkable 170% increase in sales for this quarter compared to the same period last year. “We expect sequential growth to be driven largely by data center,” said NVIDIA CFO Colette Kress on the earnings call. Following this announcement, the company’s stock price surged by 8.5 per cent in after-hours trading. NVIDIA’s gaming division, once the company’s core focus, witnessed a notable 22% surge in revenue compared to the previous year, reaching an impressive $2.49 billion. This growth outperformed expectations, surpassing the average estimate of $2.38 billion. NVIDIA  recently announced its latest innovation, the next-generation NVIDIA GH200 Grace Hopper platform. This platform centers around an innovative Grace Hopper Superchip, featuring the world’s pioneering HBM3e processor. The new setup is a game-changer, offering 3.5 times more memory capacity and 3 times more bandwidth than the current version. This setup includes a single server with 144 Arm Neoverse cores, delivering eight petaflops of AI performance and featuring 282GB of the latest HBM3e memory technology.","excerpt":"Data Center segment stood out as a driving force, contributing a staggering $10.32 billion in revenue for this quarter","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-08-24T12:45:13","publication_year":"2023","word_count":381,"keywords":["NVIDIA H100","programming_languages:R","AI","A100","RPA","innovation","RAG","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","RPA","innovation","NVIDIA H100","A100","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-soars-on-generative-ai-reports-6-billion-profit\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":16746,"title":"Artificial Intelligence is looked upon as a way to locate the mysteriously disappeared flight MH370","content":"Remember the unfortunate plight of the the Malaysian Airlines flight MH370 that disappeared in March 2014, with limited traces of it till date? After a long search on, all the underwater probing has been halted till new evidence about the specific location of the aircraft emerges. However, the Malaysian Airlines’ chief is hopeful that the resting place of the missing flight could be eventually found following an advancement in science and technology, especially Artificial Intelligence. Peter Bellew, Chief executive of the airlines during an aviation summit in Sydney said that there will be advances in science and technology that would help locate the wreckage of the flight eventually. He said that the advances are quite likely to come as the availability of artificial intelligence is coming on the stream with high capacity computing and university research that is on the go. Though any further details on the kind research that could lead to a breakthrough wasn’t revealed by him, he said that private efforts to locate the plane could help in a big way. The Boeing 777, which disappeared with 239 people on board left no trace behind and despite a lengthy deep sea hunt in the southern Indian Ocean off western Australia, nothing could be found. It eventually led to the search being called off in January 2017. Though there were three fragments of  MH370 that were found near the western Indian Ocean shores, including a two-meter wing part called flaperon, nothing substantial could be recovered. Bellew, during the conference also said the airlines is working extensively to expand its digital capabilities by the end of 2017.","excerpt":"Remember the unfortunate plight of the the Malaysian Airlines flight MH370 that disappeared in March 2014, with limited traces of it till date? After a long search on, all the underwater probing has been halted till new evidence about the specific location of the aircraft emerges. However, the Malaysian Airlines’ chief is hopeful that the […]","categories":["AI News"],"tags":["AI App"],"author_name":"Srishti Deoras","publish_date":"2017-08-04T09:09:24","publication_year":"2017","word_count":268,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Git","RAG","AI App","R"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/artificial-intelligence-looked-upon-way-locate-mysteriously-disappeared-mh370\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046657,"title":"How This Fintech Startup Uses ML To Disburse Loans And Detect Fraud","content":"Established in 2015, Mumbai-headquartered Kissht is a fintech start-up founded with the vision to provide quick and hassle-free credit financing to Indians. The fully digitised and automated fulfilment platform was founded by IIT graduates and former McKinsey partners Krishnan Vishwanathan and Ranvir Singh. In a recent conversation with Analytics India Magazine, Co-founder and CTO Karan Mehta explained how the fintech start-up is leveraging artificial intelligence and machine learning technologies to provide financing. Karan has more than a decade of experience in AI, ML and deep tech and joined the leadership team at Kissht in 2015. An MS graduate from Carnegie Mellon University, Karan has earlier founded Swipe Payment Solutions and Karma Tech Solutions. Kissht has a 100-member tech team, a 10-member product team, and 20 people working in the data science team. Edited excerpts from the interview: AIM: What are your flagship products? Karan Mehta: We offer three products: Line of credit where we assign a limit of Rs 30,000 to the customer that they can draw down on whenever they choose.Traditional short-term personal loan for a small ticket size ranging between Rs 5,000 and Rs 20,000 and for a duration of not less than 12-months. Checkout finance or BNPL where we fund end purchase of customers through merchant partners. During COVID-19, we also offered health insurance up to Rs. 1 lakh, quick cash up to 30K for businesses, and discounts on pharmacy bills and lab tests. We have an interesting lineup of products for the future, including making health and life insurance available as a cross-selling product for the underserved segment. Additionally, we have an interchange link in the pipeline which can be used against merchants we have onboarded. We also plan to roll out two-wheeler loans. AIM: How are you using AI and ML algorithms to disburse loans and credit services? Karan Mehta: To provision instant customer loans, we use our proprietary software algorithm and credit marketplace platform to assess customers’ creditworthiness. Additionally, we have created a ‘Kissht score’ to serve users without CIBIL scores. Using data-led architecture to study customers’ behaviour, online purchases, utility bill payment history, travel-ticketing history and social media presence, we develop a proxy metric to devise alternative creditworthiness evaluations. Big data, analytics, and ML have been used in the Kissht ecosystem for tasks ranging from customising communication channels to determining the ideal frequency for each customer, identifying customers requiring payment reminders, and devising smart customer support queues to detect and route queries to different channels. AIM: Explain the working of the ML algorithms. Karan Mehta: Kissht uses big-data ML-based statistical models, which helps the team to fully understand two aspects of customer profiles — a., the customer’s ability to repay, and b., detect fraud; thus, preventing loan disbursal to such customers. Our credit and fraud engines have been trained on the 15 million-plus signup users, as well as the two million-plus loan-taking customers. ML also helps in developing the risk and repayment models for this segment. With a large percentage of our customers being self-employed or working in the unorganised sectors, there are irregular income patterns or part-cash salaries. Understanding their true income earning capacity using alternative data sets is critical, as is understanding their spending patterns, utility payments, bank credits and debits, and other financial transactions. It is not just about creating the model once but also about deploying it and creating a self-learning platform that improves on it over time. The overlay of the machine learning or self-learning platform keeps varying the coefficient as we come across industry shocks or macro shocks, ups and downs with the end-to-end digital stack. AIM: Explain the tech stack of Kissht. Karan Mehta: We maintain a singular and laser-sharp focus on providing completely digital, hassle-free, and instant credit access. Our cutting-edge technology platform manifests as a very smooth customer journey combined with a very powerful credit decision engine, resulting in a strong credit portfolio for the company. In the form of India stack, eMandates, eSign, UPI, and other digital innovations, Kissht has put to use the power of every digital innovation. As a digital platform, our biggest differentiator is the way we built and fine-tuned our credit and fraud engine using third-party data sources and a great analytics process. The Kissht tech stack more closely resembles the tech stack at the modern consumer products (think Swiggy, OLA, etc.) than it does a traditional lender or bank. We are fully cloud-native and have an open-source first approach. The tech stack is organised into several independent services – each service having the freedom to choose the language, framework, architecture and database that best suits their requirements. We have services in Python, PHP and GoLang, among other frameworks. We have both Angular and React for frontends, but going forward, we have chosen to use React for all frontends. Currently, our Android and iOS apps are fully Native Apps, but we are soon moving to ReactNative apps to speed up the delivery of new products. AIM: What tech tools do you use? Karan Mehta: The product development and deployment lifecycle at Kissht make use of a large number of tools: Miro\/Invision\/Adobe XD – We use some combination of these to quickly convert thoughts and business ideas into tangible screens, allowing us to discuss, debate, refine, and iterate any new product idea.Slite – used for comprehensive product documentation, which is then easily available to all company members.Asana – for project planning and tracking sprints and product rollouts.Postman – single place for API Documentation, API sharing and collaboration, as well as API monitoring.Sagemaker Studio – apart from the tools used by the Analytics team, the technology team can quickly set up repeatable and shareable ML workflows with almost no overheads.Datadog – 360 monitoring of all production workloads with amazing monitoring capabilities.Bugsnag – tool to capture and prioritise bugs, both server and client side.ELK Stack – log analysis; one of the best examples of the power of open-source tools.DeployHQ\/AWS CodePipeline – for our CI\/CD requirements.Google Docs\/Sheets\/Forms – plays any and every role in our tech and product teams. Apart from the core in-house tech stack, we make use of a large number of third-party integrations to offer a seamless customer experience – for payments and settlements; we use Yes Bank, IDFC, Razorpay, Setu, LotusPay, PayTM, Ingenico, and others. We use Hyperverge for their Computer Vision Products. We use Income Tax Department APIs for PAN Verification and Bureau Integration for the Credit Reports. We are also integrated with DigiLocker and CKYC services. AIM: What do you look for in a potential employee wanting to join the tech team? Karan Mehta: As a start-up, we believe in the hustle attitude, strong work ethic, and perseverance. Technical skills can be learnt over time, but character and passion for giving it the best every day – that is what we look for. AIM: What key technologies have digital lending players adopted during the pandemic? Karan Mehta: The nationwide lockdown accelerated digital adoption, and businesses have amplified their technology investments on cloud-based platforms such as Application Programming Interface (API) and Internet of Things (IoT). The combination of these technologies has boosted innovation and risk mitigation in the digital lending ecosystem. AIM: What does the road ahead look like for Kissht? Karan Mehta: We plan to expand deeper in the country’s Tier-II and Tier-III towns and cities, expanding the geographic reach to the top 300 locations in the next three years. Additionally, we plan to complete the product portfolio within the lending segment. We will also be looking for exceptional tech talent to further enhance our deep technology capabilities in the domain of analytics, machine-learning-based algorithms, customer life cycle management, and loyalty programs. Speaking about the numbers, in March this year, Kissht disbursed Rs 300 crore worth of loans. We crossed 10 million registered users in the last year, and today we have a registered user base of 16 million and close to three million unique borrowers. Our ambition is to reach the 100 million users mark by 2025. We expect our revenues to double from Rs 175 crore in FY21 to close to Rs 350 crore in FY22.","excerpt":"Kissht uses machine learning-based statistical models to predict customers’ ability to repay loans and detect frauds.","categories":["AI Startups"],"tags":["AI (Artificial Intelligence)","FinTech","indian statistical service","Machine Learning"],"author_name":"Debolina Biswas","publish_date":"2021-08-24T15:00:00","publication_year":"2021","word_count":1351,"keywords":["data science","artificial intelligence","machine learning","AWS","AI","ML","Machine Learning","computer vision","RAG","Aim","analytics","FinTech","indian statistical service","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","data science","analytics","Aim","RAG","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-fintech-startup-uses-ml-to-disburse-loans-and-detect-fraud\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":36463,"title":"Item Categorisation For Online Retail Platforms Boils Down To Knowing Fries = Chips","content":"One of the key problems faced by almost all e-commerce giants across the world is item categorisation. These large online retail giants feature a very large and long-tail inventory with millions of Items (product offers) entering into the marketplace every day. The quality of item categorisation plays a significant role in subsequent customer-facing applications such as search, product recommendation, trust and safety, product catalogue building, and seller utilities. A correct item categorisation system is also essential for user experience as it helps determine the relevant presentation logic in surfacing the items to users through search and browsing. How E-Commerce Giants Tackle Categorisation Problem The success of e-commerce can be associated with its offering of commodities below the usual market price along with the variety of deals; an effortless window shopping. In order to give better deals to the customer while managing the profitability, the procurement costs have to be brought down. A good old technique is to procure the goods in high volume. In the case of Japanese e-commerce giant Rakuten Ichiba too, the engineers used a large scale multi-class hierarchical product categorisation method. In one of their works, to improve item selection at a large scale, they used data containing 172 million product titles and description. For training the deep networks in a reasonable time, word-vectors need to be sparse even if the input layer is very large for training deep networks. This can be done using the selective reconstruction method. Models like Deep belief nets (DBN) and deep autoencoders (DAE) were used in the case of Rakuten. DBN and DAE with selective reconstruction are implemented on GPUs in order to process a large product base within a reasonable amount of time because conventional methods like multinomial Naive Bayes are not practical to be used in that scale. Initially, the text is normalised by converting all Japanese characters to full-width and all non-Japanese characters to lower cases. And, then all HTML tags are cleaned from descriptions. For example, the words “iPhone 4s” is normalised to “iphone 4s”. A frequency threshold is chosen so as to not select those words which appear a few times as features. This helps in eliminating most of the noisy information as well. Eliminating less-frequent words also helps to make the classifier practical and more robust for the classification of new products. Spell correction and abbreviations are two main challenges during item categorisation. This forms a crucial aspect of data cleaning tasks for e-commerce sites. Doing It The Walmart Way Walmart too has deployed state-of-the-art machine learning models. In their case, they figured out that the generalised linear model (GLM) had an advantage over random forest models. Walmart has a huge global presence with regards to the retail market and with Flipkart, it wants to capture a significant chunk of the Indian e-commerce market. Initially, Walmart operated independently. The Walmart buyers of India have no knowledge of their American counterparts. Since, bringing the procurement costs is the main objective, it now makes sense to integrate the suppliers on a global scale when buying in high volume. Now this integration will come with an inevitable challenge- data collection and mapping. A starting point could be building a sample data with few classes to create a pipeline. An incoming text will be converted into a term frequency-inverse document frequency (tf-idf) matrix for feature extraction and then plug this data into machine learning models. Walmart follows these steps for feature engineering: Taking item description from raw data Data cleaning: removing special characters, stop words and single letter words Stemming and translation Concatenating all the clean columns and creating document term matrix For converting the different classes of the items into manageable clusters, the hierarchical clustering method is used. The sub-classes within the hierarchy follow the Bayes’ rule. Usually, while forming the final matrix, Jaccard distance is used as a metric to measure the frequency of the words occurring in different classes. But in the case of Walmart, a modified Jaccard distance is used instead to avoid the problem of improper grouping due to inequality number of words. Modified Jaccard distance considers the ratio of the similarities in classes and minimum words present in the class. Know more about the work here Data matching globally is an intensive task. For example, fries in the US can mean chips in the UK. Using synonyms and other localised terms make text pattern recognising difficult. Moreover, merchants may not be accurate while assigning products, the category assignment of the same product listed by different merchants may not be consistent. Automatic category recommendation for given product information helps in solving these problems. These challenges are persistent across all e-commerce platforms and with intensified customer-centric retail strategies, there will be more developments in  NLP and image classification techniques. Also watch:","excerpt":"One of the key problems faced by almost all e-commerce giants across the world is item categorisation. These large online retail giants feature a very large and long-tail inventory with millions of Items (product offers) entering into the marketplace every day. The quality of item categorisation plays a significant role in subsequent customer-facing applications such […]","categories":["Deep Tech"],"tags":["e-commerce"],"author_name":"Ram Sagar","publish_date":"2019-03-18T05:28:49","publication_year":"2019","word_count":796,"keywords":["Go","machine learning","e-commerce","AI","RPA","ML","feature engineering","NLP","ViT","Rust","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","R","Go","Rust","feature engineering","ViT","RPA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/item-categorisation-for-online-retail-platforms-boils-down-to-knowing-fries-chips\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10068428,"title":"Adani’s Drone Business","content":"The Adani Defence and Aerospace arm of Adani Group has acquired a 50 per cent stake in Bangalore-based Drones startup – General Aeronautics. This startup specialises in developing robotic drones mainly for applications in agriculture like tech-enabled crop protection services, precision farming, crop health monitoring, and AI-based yield monitoring. In April this year, Adani’s joint venture (JV) company with Israeli firm Elbit became one of the 14 beneficiaries under the Indian government’s PLI scheme for drone manufacturing. Adani Group and its drone business While announcing Adani’s 50 per cent stake in General Aeronautics, Ashish Rajvanshi, CEO of Adani Defence and Aerospace, said that the partnership will fuse its military and civilian objectives. He added that Adani Group’s military UAV capabilities will be integrated with the capabilities of the Bangalore-based startup. General Aeronautics was founded in 2016 by Abhishek Burman, Kota Harinarayana, and Anutosh Moitra. In a conversation with Analytics India Magazine, Burman said that with this partnership, Adani Group will help General Aeronautics scale up its production. “We have now already developed the solution, which is kind of having completed the validation and market valuation. It is ready for scale. And that’s where a partner like them can help us to scale fast because this market is really huge, and there is a huge scope of expansion. The Adani group already has good infra already available with them for UAV manufacturing. We also hope to get benefited from their massive network and presence on the ground and make strides in the B2B or B2C model.” It is interesting that the defence arm of Adani Group has acquired a stake in General Aeronautics, indicating a possible military production angle to it. However, Burman says that they would continue to concentrate on the agriculture sector but may branch out to civil and military applications in the future. Adani Defense Systems was established in 2015; it offers drone-based solutions for military purposes. Its aim is to help India emerge as the number one destination for defence manufacturing. In this regard, the company is also investing in MSMEs, startup incubators in India, which will help absorb and build technologies. The Adani Group is exploring two revenue models – dealer-based and service-based. In the dealer-based model, equipment is directly sold to the customer, and in the latter case, equipment is offered for various services against a fee and in partnership with local entrepreneurs and institutions. As per the company, Rangarajan Vijayraghavan said in a statement that the dealer-based model will be used in the commercial drone sector, while the service-based model will offer services to the farmer community. A government scheme beneficiary Interestingly, Adani-owned Alpha Design Technologies Limited (ADTL) partnered with Israeli firm Elbit-ISTAR, a subsidiary of Elbit Systems that provides end-to-end surveillance, target acquisition and reconnaissance solutions, to form a joint venture company called Vignan Technologies. This joint venture also has a new facility for R&D for Indian and global markets. The JV is one of the 14 recipients of the PLI scheme launched by the Indian government. Under this scheme, the incentive for a manufacturer of drones and drone components will be 20 per cent of the value addition made by the company in the next three years. The aviation ministry said that the eligibility criteria for the PLI scheme includes an annual sales turnover of Rs 2 crore for the drone companies and Rs 50 lakh for drone components manufacturers. There will also be a value addition of over 40 per cent of sales turnover. Last year, the government introduced the liberalised Drone Rules 2021. Under this, the requirements for several key approvals were abolished. One of the major updates was the reduction of the number of forms from 25 to 5 and the types of fees reduced to 4 from 72. Many companies, including Adani Group and Reliance Industries Limited, are rushing to leverage these incentives. In 2019, Reliance Industries Ltd acquired a majority stake in drone maker Asteria Aerospace Pvt Ltd, which develops drones for military, mining, construction, and agriculture domains.","excerpt":"Adani Defense Systems was established in 2015; it offers drone-based solutions for military purposes.","categories":["IT Services"],"tags":["joint venture"],"author_name":"Shraddha Goled","publish_date":"2022-06-06T11:00:00","publication_year":"2022","word_count":671,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","Ray","Aim","joint venture","analytics","R","startup"],"extracted_tech_keywords":["AI","analytics","Aim","Ray","RAG","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/adanis-drone-business\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058903,"title":"DeepMind co-founder Mustafa Suleyman leaves Google","content":"DeepMind cofounder Mustafa Suleyman is leaving Google to join venture capital firm Greylock Partners. Suleyman joined Google in 2014 when the search giant acquired DeepMind, a cutting-edge AI research lab for USD 650 million. He was Google’s vice president of product management and policy for AI. Suleyman is not an AI researcher by training. He was instrumental in pushing the company into health care research and was vocal about not using AI technology for military applications. Mustafa led teams inventing and deploying cutting edge AI systems to accurately detect breast cancer in mammograms, diagnose 50 different eye diseases in OCT scans, and to control Google’s multi-billion dollar data centers to optimize energy consumption. “I think the thing that I really love about doing the investing that I’ve done is that I get the opportunity to spend time with people who are visionary and fearless, and that really energizes me. I’m definitely somebody who likes to take risks and try to learn from my mistakes and try to be fearless about that,” said Suleyman in a podcast with Greylock. At Greylock, Suleyman now will work closely with early-stage AI companies to push the frontiers of innovation.“Visionary AI expert Mustafa Suleyman and I share the belief that a philosophical, humanistic perspective must drive AI development. That’s why I’m thrilled the DeepMind co-founder and former VP of AI at Google is joining Greylock as a Venture Partner,” said Reid Hoffman, Partner, Greylock in a twitter post.","excerpt":"Suleyman joined Google in 2014 when the search giant acquired DeepMind.","categories":["AI News"],"tags":["DeepMind","DeepMind AI","healthcare ai","Mustafa Suleyman"],"author_name":"Meeta Ramnani","publish_date":"2022-01-21T17:02:30","publication_year":"2022","word_count":243,"keywords":["Go","API","healthcare ai","Mustafa Suleyman","programming_languages:R","AI","innovation","venture capital","programming_languages:Go","edge AI","DeepMind AI","AI research","R","DeepMind"],"extracted_tech_keywords":["AI","edge AI","R","Go","API","innovation","venture capital","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deepmind-co-founder-mustafa-suleyman-leaves-google\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171657,"title":"Oracle Surpasses Quarterly Estimates, Records 11% Rise in Revenue","content":"Oracle surpassed Wall Street’s expectations for its FY 2025 fourth-quarter performance on Wednesday. Within hours, increased demand for its cloud services from businesses integrating AI caused its stock to rise 7%. The company announced fourth-quarter earnings of $1.70 per share, excluding certain expenses like stock compensation. Revenue reached $15.9 billion, an 11% increase compared to the same period last year. These results exceeded expectations, as Wall Street had anticipated earnings of only $1.64 per share on revenues of $15.59 billion. In fiscal year 2025, total revenues grew 8% to $57.4 billion. Cloud services and license support revenues increased by 12% to $44.0 billion. Cloud and on-premise license revenues rose by 2% to $5.2 billion. “FY25 was a very good year, but we believe FY26 will be even better as our revenue growth rates will be dramatically higher. We expect our cloud growth rate—applications plus infrastructure—will increase from 24% in FY25 to over 40% in FY26. Cloud Infrastructure growth rate is expected to increase from 50% in FY25 to over 70% in FY26,” Oracle CEO, Safra Catz, said. According to Oracle, in Q4 2024, total cloud revenue, including IaaS and SaaS, reached $6.7 billion—a 27% increase. Cloud Infrastructure (IaaS) revenue was recorded at $3.0 billion, up 52%, while Cloud Applications (SaaS) generated $3.7 billion, a 12% rise. Fusion Cloud ERP (SaaS) revenue also reached $1.0 billion, growing by 22%, and NetSuite Cloud ERP brought in $1.0 billion, an 18% increase. Oracle has been launching AI assistants, advisors, and agents. Its AI Agent Studio, revealed in March, aims to assist customers and partners in creating their personalised AI agents. “MultiCloud database revenue from Amazon, Google and Azure grew 115% from Q3 to Q4. We currently have 23 MultiCloud data centres live, with 47 more being built over the next 12 months. We expect triple-digit Multicloud revenue growth to continue in FY26,” Oracle chairman and CTO, Larry Ellison, said. After forming partnerships with Microsoft Azure and Google Cloud, the company joined forces with AWS in early 2024 to introduce Oracle Database@AWS. AIM reported that this offering will enable customers to utilise Oracle Autonomous Database and Exadata Database Service in AWS data centres.","excerpt":"The company announced fourth-quarter earnings of $1.70 per share, excluding certain expenses.","categories":["AI News"],"tags":["Oracle Cloud Infrastructure","Revenue Growth"],"author_name":"Smruthi Nadig","publish_date":"2025-06-12T12:46:37","publication_year":"2025","word_count":359,"keywords":["Go","AI assistants","AWS","AI","R","RPA","Git","Aim","AI agents","Azure","Oracle Cloud Infrastructure","Revenue Growth"],"extracted_tech_keywords":["AI","Aim","AI assistants","AWS","Azure","R","Go","Git","RPA","AI agents"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/oracle-surpasses-quarterly-estimates-records-11-rise-in-revenue\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":28584,"title":"Data Science Hacks No One Talks About But Are A Must In Your Toolkit","content":"With the data revolution in full swing, there is more information on the internet than a human can remember and process in his\/her lifetime. Data Science is a demanding platform, where every forward looking enterprise and startup wants to increase their productivity with the help of intelligent systems. It is an interdisciplinary platform that involves numerous techniques and skills such as, analysis, programming, math and statistics. Now, it is commonly believed that a person with a hacker mindset can come up with an easier solution compared to an orthodox approach. Let us look at some of the little-known hacks in data science field which aren’t extensively talked about. “The job of the data scientist is to ask the right questions. If I ask a question like ‘how many clicks did this link get?’ which is something we look at all the time, that’s not a data science question. It’s an analytics question. If I ask a question like, ‘based on the previous history of links on this publisher’s site, can I predict how many people from France will read this in the next three hours?’ that’s more of a data science question.” ―Hilary Mason, Founder, Fast Forward Labs Hacker Mindset When dealing with data, a hacker’s mindset always wins hands down. Data science is not all about building models, plotting graphs to analyse the attributes, training and testing by tuning the parameters but a person who finds an easy way to deal with data rather than making use of complex tools to process a data is definitely a hacker. Let us consider an example which just reduces the code to just one line, From this, list1 = range(0, 10) for i in list1: print(i) To this, [print(i) for i in range(0 ,10)] Data Cleaning Tricks Let us say you are cleaning data for language processing tasks, and a simple models might give you the best result. Cleaning is one of the most complex processes involved in data science, since almost every data available or extracted for language processing tasks  is unstructured. It is a fact that a highly processed and neatly structured data will yield better results than a noisy one. But the cleaning task can be accomplished with simple regular expression rather than making use of a complex tool. Domain Knowledge When a Data Scientist is asked to build a model with a given data, understanding what the data is about is a key aspect. Irrespective of the structure and the type of the data, knowing and understanding the domain knowledge of where the data is from, let us say, from a finance, tech, agriculture, manufacturing industries. A data scientist with knowledge of industry will be able to give better insights and analysis about the data compared to just build a model from A to Z. Domain knowledge also helps to develop better insights and understand the analysis processes. Never say “No More Learning” “Data Science is a journey, not a destination” This line gives us an insight about how huge the data science domain is and why constant learning is as important as build intelligent models. Practitioners who keep themself updated with the new tech being developed everyday, are able to implement and solve business problems faster.  With all the resources available on the internet like MOOCs, one can easily make use of these to be updated. Also showcasing your skill on your blog or Github is an important hack which most of us are unaware of. This not only benefits their “The man who is too old to learn was probably always too old to learn.” – Henry S. Haskins Cheat Sheets Machine learning cheat sheets are a way to keep your mind on things which one may tend to forget, as there is a lot to remember in Data Science. There are a lot of machine learning cheat sheets available on the internet. Some of which are also available on the Scikit-learn website, also we have found a github repo which can be found here. This is a cheat sheet provided by the Stanford University to keep oneself updated. This is definitely a hack which no one thinks of when it comes to building a pipeline to solve a problem, a must-have in every data scientist’s toolkit. In Conclusion These are some of the hacks which have gone unnoticed, when it comes to machine learning and analytics Hacks can make one’s life easier and give a better result compared to a complex approach of solving.","excerpt":"With the data revolution in full swing, there is more information on the internet than a human can remember and process in his\/her lifetime. Data Science is a demanding platform, where every forward looking enterprise and startup wants to increase their productivity with the help of intelligent systems. It is an interdisciplinary platform that involves […]","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)"],"author_name":"Kishan Maladkar","publish_date":"2018-09-25T05:25:19","publication_year":"2018","word_count":749,"keywords":["data science","scikit-learn","Go","machine learning","AI","Git","ViT","analytics","GitHub","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","scikit-learn","R","Go","Git","GitHub","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/data-science-hacks-no-one-talks-about-but-are-a-must-in-your-toolkit\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10045349,"title":"Top 10 AI Innovations Of 2021 So Far","content":"AI is a complex and ever-evolving field where organisations and individuals are constantly focused 0n finding novel solutions to pressing challenges. The year has been full of path-breaking innovations which have pushed the boundaries and made way for better outcomes. In this article, we list the top ten AI innovations of 2021 so far. GitHub’s Copilot OpenAI and Microsoft’s GitHub Copilot is an AI-based tool for programmers to write better code. The programmer can describe a function to the Copilot in plain English as a comment, and the machine will convert it to actual code. OpenAI Codex lays a foundation for Copilot, it provides an AI system trained on a dataset made up of a huge set of public source code. It works on a broad set of frameworks and languages, and is ideal for languages like Python, JavaScript, TypeScript, Go, and Ruby. The team has claimed Copilot to be far more advanced than the existing code assistants. Unified Transformer Researchers from Facebook AI Research introduced a new Transformer model, Unified Transformer (UniT). UniT has an encoder-decoder architecture that handles multiple tasks and domains in a single model with fewer parameters; as per Facebook’s team, UniT is a step towards general intelligence. OpenAI’s DALL.E & CLIP DALL. E is OpenAI’s 12-billion parameter version. It is a transformer that can generate images from text prompts. The model can work with multiple objects in an image to either render an image or alter it based on text prompts. The OpenAI research team has also demonstrated a neural network called Contrastive Language-Image Pre-training or CLIP. This neural network has been trained on 400 million pairs of images and text. CLIP is also similar to GPT family and can learn to perform tasks such as object character recognition (OCR), geo-localisation, action recognition, etc. Blender Bot 2 Facebook’s BlenderBot 2 is a first of its kind open-source chatbot with long term memory. Facebook has been working to make the AI more empathetic, knowledgeable and capable. The BlenderBot 2.0 can build long term memory for continuous access. It does so while simultaneously searching for information on the internet and holding conversations on nearly any topic. Google’s Translatotron 2 In 2019, Google released Translatotron, an end-to-end speech-to-speech translation model. It was then the first end-to-end framework which could translate speech from one language into speech to another, directly. The system was used to create synthesised translations of voices to ensure the sound of the original speaker is intact. But this feature had the potential to be misused to generate speech in a different voice and create deep fake voices. This year, Google released Translatotron 2, an updated version where the trained model is restricted to retain the source speaker’s voice. Unlike the previous version, it cannot generate speech in different voices, thereby mitigating potential misuse for creating spoofing audio artefacts. Vertex AI Google introduced Vertex AI, a managed machine learning platform for deploying and maintaining AI models, at this year’s Google I\/O conference. The new platform brings AutoML and AI Platform together into a unified API, client library and user interface. Earlier, researchers would be required to run millions of test images for training algorithms, but now, they can rely on Vertex technology stack to do the heavy lifting. FLAML Microsoft’s FLAML is a python package that can tell us the best-fit ML model for low computation. It helps eliminate the manual process of choosing the best model and best parameter. This AutoML system is mainly focused on–model selection, hyperparameter tuning, feature engineering, neural architecture search, and model compression. MusicBERT MusicBERT is Microsoft’s Large Scale Pre-Trained Model For Symbolic Music Understanding. It covers applications such as emotion classification, genre classification, and music piece matching. Microsoft has created this model using an OctupleMIDI method, bar-level masking strategy, along with a large scale symbolic music corpus containing more than 1 million music tracks. Microsoft’s neural TTS Microsoft’s neural text to speech software (TTS) enables developers to create custom synthetic voices. The AI is structured in three layers: text analyser, neural acoustic model, and neural vocoder. The text analyser converts plain text to pronunciations, the acoustic model converts pronunciations to acoustic features and finally, the vocoder generates waveforms. Tensorflow 3D Google’s TensorFlow 3D is a highly modular library to bring 3D deep learning capabilities to TensorFlow. While the previous TensorFlow was not enough to understand the environment, the 3D update provides a set of operations, loss function, data processing tools, metrics, and other models for developing, training, and deploying state-of-art 3D scene understanding models.","excerpt":"OpenAI and Microsoft’s GitHub Copilot is an AI-based tool for programmers to write better code","categories":["AI Trends"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-08-06T15:00:00","publication_year":"2021","word_count":754,"keywords":["machine learning","OpenAI","AI","neural network","ML","Python","Aim","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","OpenAI","Aim","TensorFlow","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-ai-innovations-of-2021-so-far\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10098011,"title":"Steg.AI’s Unique Watermarking Approach Overshadows Tech Giants","content":"In May 2023, the world was left shell-shocked as it witnessed images of the Pentagon shrouded in smoke. Many news channels reported the incident based on these images and even the stock market reacted, dropping for a brief period of time. Later, it turned out to be a fake AI-generated image. Such instances highlight the challenge in identifying AI-generated content in various contexts. They also bring back the larger discussion on deep fakes and fake images—which has been amplified due to generative AI tools and their ability to produce hyperrealistic images just from prompts. All of this has led to governments across the globe scrambling for solutions and a way to safely regulate artificial intelligence. Prominent AI firms, including Amazon, Anthropic, Google, Inflection, Meta, Microsoft, and OpenAI, have voluntarily committed to measures like watermarking AI-generated content to enhance its safety, but there hasn’t been any solid development yet. Current methods like encoding data into images or audio can be easily bypassed. A robust, invisible watermark that’s easily applied and detected, yet resistant to transformations, is necessary, as studies suggest that it is difficult for humans to differentiate between human and AI-generated contents. As online IP theft is rampant, the ability to prove content creation’s origin is increasingly essential. Meanwhile, a California-based platform Steg.AI has developed a deep learning-based solution that embeds nearly imperceptible watermarks into digital content. Even if the images are altered, compressed, or manipulated, the Steg.AI watermark remains intact. Remarkably resilient, these watermarks can even be captured using an iPhone camera when displayed on screens or printed. Steg.AI’s watermarking solution finds applications in diverse scenarios, such as stock photography services, content sharing on platforms like Instagram, pre-release copies of films, and safeguarding confidential documents. Early iterations of their product faced challenges, leading to a shift in focus towards robustness, a standout feature that resonated with customers. How it Works Steg.AI’s core concept involves seamlessly integrating watermarks into AI-generated images before distribution. While the specifics of the process remain proprietary, the basic idea revolves around a pair of machine-learning models. One model customises the watermark’s placement within the image, ensuring imperceptibility to the human eye while remaining detectable by the decoding algorithm. Analogous to an invisible, mostly unchangeable QR code, this method potentially holds kilobytes of data – sufficient for URLs, hashes, and plaintext information. Each page of a multi-page document or video frame could harbour distinct codes, exponentially increasing the capacity. The company’s extensive work can be traced back to a 2019 CVPR paper, along with the acquisition of Phase I and II SBIR government grants. Co-founders Eric Wengrowski and Kristin Dana, who were previously involved in academic research, have dedicated years to refining their approach. Steg.AI’s progress has been supported by NSF grants and angel investments totaling $1.2 million. Recently, the company announced a significant milestone with a $5 million seed funding round led by Paladin Capital Group, accompanied by participation from Washington Square Angels, NYU Innovation Venture Fund, and individual angel investors. What Big Techs are Doing Major tech companies have taken steps to incorporate watermarking into their content. Microsoft announced at its annual Build conference that it’s adding new media provenance features to Bing Image Creator and Designer, enabling users to verify AI-generated images and videos. This innovation involves cryptographic methods to mark and sign content with metadata indicating its origin. For this, websites must adopt the Coalition for Content Provenance and Authenticity (C2PA) specification developed with Adobe, Arm, Intel, Microsoft, and Truepic. However, the impact of Microsoft’s efforts relies on broader media provenance standard adoption, with support from companies like Stability AI and Google who are also exploring similar approaches. Shutterstock and Midjourney have adopted guidelines to embed markers indicating generative AI-created content. On the other hand, collaborative research involving Meta AI, Centre Inria de l’Universite de Rennes’, and Sorbonne University have developed an innovative technique that seamlessly incorporates watermarking into the image generation process while preserving the architecture. This method modifies pre-trained generative models to effectively integrate watermarks into generated images, enhancing security and computational efficiency. This technology enables model providers to distribute versions of their models with distinct watermarks for different user groups, facilitating ethical usage monitoring. The technique is valuable for media organisations in identifying computer-generated images. Leveraging Latent Diffusion Models (LDM), the researchers successfully integrated watermarks with minimal adjustments to generative models. The process involves fine-tuning LDM decoders using perceptual image loss and hidden message loss from a streamlined deep watermarking method called HiDDeN. The technique showcases strong performance in image editing tasks, even with heavily cropped images, maintaining the original model’s utility across various LDM-based tasks. The move by the seven big techs late last month supports the Biden administration’s push to regulate the booming and popular AI technology. The US Congress is also reviewing a Bill that would mandate the disclosure of AI involvement in creating political ads.","excerpt":"While Meta, Microsoft, Stability AI, Midjourney etc make efforts, Steg.AI has raced ahead with continued efforts","categories":["AI Highlights"],"tags":["AI Technology","Amazon","Anthropic","architecture","deep fakes","Google","Meta","Meta AI","Microsoft","MidJourney","OpenAI","Shutterstock","Stability AI","stock market","Watermarking"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-08-03T11:53:35","publication_year":"2023","word_count":810,"keywords":["Meta AI","Watermarking","deep learning","deep fakes","R","artificial intelligence","RAG","Stability AI","Meta","Shutterstock","AI","ML","Amazon","generative AI","Anthropic","MidJourney","OpenAI","AI Technology","architecture","stock market","Google","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","generative AI","OpenAI","Anthropic","Meta AI","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/steg-ais-unique-watermarking-approach-overshadows-tech-giants\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":47074,"title":"Comparing Label Encoding And One-Hot Encoding With Python Implementation","content":"As machine learning algorithms most often accept only numerical inputs, it is important to encode the categorical variables into some specific numerical values. In this article, we compare the label encoding and one-hot encoding techniques by implementing it in Python. Label Encoding Label encoding is one of the popular processes of converting labels into numeric values in order to make it understandable for machines. For instance, if we have a column of level in a dataset which includes beginners, intermediate and advanced. After applying the label encoder, it will be converted into 0,1 and 2 respectively. OneHot Encoding One-Hot Encoding is one of the most widely used encoding methods in ML models. This technique is used to quantify categorical data where it compares each level of the categorical variable to a fixed reference level. It converts a single variable with n observations and x distinct values, to x binary variables with n observations where each observation denotes 1 as present and 0 as absent of the dichotomous binary variable. Implementation in Python Loading Dataset For the comparison, we used the Titanic: Machine Learning from Disaster dataset. Both the training and testing dataset are given where the training dataset contains 891 rows and 12 columns. You can download the dataset from here. Source: Kaggle After loading the dataset, we used the dropna() method which allows eliminating the null values from the rows and columns. After dropping the null values, the remaining rows and columns are 792 and 8 respectively. Label Encoding In Python, label encoding can be done with the help of the Sklearn library. We used label encoder for specifically two columns or class which are “sex” and “embarked”. After appling label encoder we can notice that in embarked class C, Q and S are assumed as 0,1 and 2 respectively while the male and female in sex class is assumed as 1 and 0 respectively. We further implemented the data in Support Vector Machine (SVM) and the accuracy score is shown as 60%. The code snippet is shown below: One-Hot Encoding We will now apply the one-hot encoding in the dataset in a similar manner and imply it to one of the popular classification methods, Support Vector Machines (SVM). This will provide us with the accuracy score of the model using the one-hot encoding. It can be noticed that after applying the one-hot encoder, the embarked class is assumed as C=1,0,0, Q=0,1,0 and S= 0,0,1 respectively while the male and female in the sex class is assumed as 0,1 and 1,0 respectively. The code snippet is shown below Here, by comparing the accuracy scores of the two encoder techniques, we can see that the accuracy score of the label encoder is less than the accuracy of the one-hot encoder. Outlook Most of the time, the outcome of a machine learning model is represented by how much accuracy rate the model is providing. The whole idea behind the implementation in Python is to show the difference in the accuracy rate in the ML model. Also, important to mention that the accuracy we are achieving here is around 60% since we have been using a small number of data in our model just to compare the two encoder techniques. To get better accuracy, one must always use a large number of the dataset.","excerpt":"As machine learning algorithms most often accept only numerical inputs, it is important to encode the categorical variables into some specific numerical values. In this article, we compare the label encoding and one-hot encoding techniques by implementing it in Python. Label Encoding Label encoding is one of the popular processes of converting labels into numeric […]","categories":["Deep Tech"],"tags":["Machine Learning Algorithms","Python"],"author_name":"Ambika Choudhury","publish_date":"2019-10-10T18:00:41","publication_year":"2019","word_count":552,"keywords":["Machine Learning Algorithms","Go","machine learning","programming_languages:R","AI","ML","programming_languages:Go","Python","programming_languages:Python","R"],"extracted_tech_keywords":["AI","machine learning","ML","Python","R","Go","programming_languages:Python","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/comparing-label-encoding-and-one-hot-encoding-with-python-implementation\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":18673,"title":"How Is AI Getting Popular With Farmers? Is India Pacing Up?","content":"Guess what’s in for a cost effective and scalable technology for disease detection in plants! It’s artificial intelligence in the news again. It wasn’t long back when a team of researchers taught an AI to recognize crop diseases and pest damages. In a submission by Amanda Ramcharan and her team, they used transfer learning for image-based cassava disease detection. The new transfer learning method that used dataset of cassava disease images taken in the field in Tanzania, can be easily deployed on mobile devices without accessing the cloud. Transfer learning was applied to train a deep convolutional neural network to identify three diseases and two types of pest damage. The best trained model delivered an overall accuracy of 93%. Amidst its soaring popularity, artificial intelligence is gaining rapid momentum in the area of farming. As it is growing popular with farmers, AI as a tool an be seen easily being implemented for killing weeds, harvesting better crops and producing a better yield. There have been many developments in India and abroad that speak of growing relevance of AI in farming. Technologies making a headway in agriculture: With technologies like drones, intelligent monitoring systems and robots making into field trials, there is no denying that artificial intelligence is set to bring the next phase of ultra precision agriculture. Let’s glance into few of these technologies that are making headways into agriculture. Image\/ Facial recognition- This largely adopted technology has been popular with a lot of smartphones and has climbed onto the area of agriculture. These devices can eliminate the need of fitting devices and allowing easy monitoring of plants with minimal interaction. Drones- This is the next wave of revolution hitting a farmer’s field. It can facilitate in-depth field analysis, easy and efficient crop spraying and crop monitoring. Automated irrigation system- As the name suggests, these are designed to utilise real-time machine learning to constantly maintain desired soil conditions to increase average yields. It may result into lesser costs and labour involved. Crop health monitoring- Companies are developing automated detection and analysis technologies such as hyperspectral imaging and 3D laser scanning that will increase the precision and volume of data collected, hence a better analysis and better results. Others that can be soon introduced into farming are driverless tractors and chatbots, where conversational virtual assistants powered by AI can interact with farmer in a personalized way. As we know chatbots have already made its presence felt in areas like retail, travel, media, insurance etc. AI and Robots from picking fruits to harvesting plants: Few use Cases The popularity for these technologies go well into all sorts of application in the farming sector. Abundant Robotics, a US based startup has developed an apple-picking robot. According to the company it can identify, pluck, and place apples into a crate with roughly the same accuracy and care as a human. There also has been instances of AI and machine learning being used to eliminate weeds and hence taking care of plants with utmost precision. Deere & Company acquired Blue River Technology, a pioneer in applying machine learning to agriculture. The latter had earlier applied machine learning to agricultural spraying equipment and intends to extend machine learning usage to enable every farmer in the field to optimize every plants. NatureSweet, a texas based company is using AI to monitor tomatoes in their growing season. With cameras installed and algorithms set in place, the company tracks all sorts of emerging problems in plants. The company claims that the AI system has helped increasing the harvest between two and four percent, and the founders hope that number will eventually climb to 20 percent. There were also reports of Harper Adams University and Precision Decisions recently testing autonomous drones and vehicle to plant, maintain and harvest a crop, and taking away the credit of being the first in the world to do so. A Georgia based startup called AgVoice is developing natural language processing toolkit for farmers. The system interprets sudden death syndrome for the soy fungal disease and prompts for the location and severity of the observation. Scenario in India Back in India, the efforts in the direction of artificial intelligence and advanced analytics have been slowly pacing up. While it has marked a spectacular development in other domains such as healthcare, retail, e commerce etc., in terms of introducing AI, farming is yet to take a centre stage. In a very recent development, Karnataka government signed a Memorandum of Understanding with Microsoft India to develop a multi-variant agricultural commodity price forecasting model. Taking into consideration Tur crop to understand the working of prediction model, this model is set to take into consideration historical sowing area, production, yield, weather datasets and other related datasets as relevant, to produce desired result. Believing that intelligent technologies can deliver solutions in the agriculture sector, the move is aimed at using digital tools powered by AI to help farmers get higher crop yield. It had been reported that Microsoft in collaboration with ICRISAT has deployed a Sowing Advisory Service in the kharif season on a limited pilot, under the “Bhoochetana” project. These have been built on the Microsoft Cortana Intelligence Suite. While Microsoft is stepping ahead with Karnataka to develop price-forecasting model, IBM Research Labs is making its tools available to entrepreneurs\/startups to develop solutions. The AI, cognitive computing, image processing and other advanced technologies by IBM would be made available to bring innovation solutions in the Indian agriculture space. Another effort in the Indian agrarian space involves New Delhi and Lucknow based Agritech startup Gobasco provides AI-powered data-driven supply chain optimisation platform. It carries out real-time data analytics on data-streams. It is one of first investments of Matrix Partners India in Agritech sector. On a concluding note Apart from being highly accurate and fast, these AI technologies are quite affordable. Experts claim that it may not be the most exciting technological venture, but can be quite beneficial. Leading a way into bringing high quality food, it is also speculated to be fighting against world hunger. Though there are opinions about AI being too rigid for agricultural environments, the advancing of AI in farming is reaching newer heights, and is expected to grow more!","excerpt":"Guess what’s in for a cost effective and scalable technology for disease detection in plants! It’s artificial intelligence in the news again. It wasn’t long back when a team of researchers taught an AI to recognize crop diseases and pest damages. In a submission by Amanda Ramcharan and her team, they used transfer learning for […]","categories":["IT Services"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-11-03T03:08:22","publication_year":"2017","word_count":1033,"keywords":["machine learning","artificial intelligence","AI","neural network","chatbots","virtual assistants","RAG","Ray","Aim","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","analytics","Aim","Ray","RAG","chatbots","virtual assistants"],"url":"https:\/\/analyticsindiamag.com\/it-services\/ai-getting-popular-farmers-india-pacing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10064120,"title":"Top books on AI ethics and bias","content":"Ethics and bias in AI are when human’s prejudiced opinions are encoded into the algorithms, making AI either unethical or biased. When we train the machines, knowingly or unknowingly, human prejudice, bias and sometimes even wrong decisions are encoded in the algorithm. When the machines gain authority over the algorithm, we tend to believe in it over humans. To better understand AI ethics and bias, several books highlight just this to give a broader approach to understanding AI. Let’s take a look. Heartificial Intelligence: Embracing Our Humanity to Maximise Machines (2016) by John C Havens This book features pragmatic solutions drawing on economics, emerging technologies, and positive psychology. The book provides the first values-driven approach to algorithmic living—a definitive roadmap to help humanity embrace the present and positively define their future. Each chapter opens with a fictional description to help readers imagine how they would respond to various AI scenarios while demonstrating the need to codify their values, as the algorithms are already dominating society. The book paints a vivid portrait of how our lives might look in either a dystopia of robotic and corporate dominance or a utopia where humans use technology to enhance our natural abilities to evolve into a long-lived, super-intelligent, and altruistic species. Weapons of Math Destruction by Cathy O’ Neil In this book, the writer, a renowned mathematician, provides multiple examples of machine learning or statistical modelling gone wrong. Her examples range from practical teacher evaluation algorithms to automatic systems that judge job applicants. While Cathy focuses on the negative impacts of big data analysis, it would be an eye-opener for anyone looking at real-world examples of biased systems caused by wrongful statistical modelling techniques. Artificial Unintelligence: How Computers Misunderstand the World by Meredith Broussard This book guides us by understanding technology’s inner workings and outer limits and why we should never assume that computers always get it right. The writer does an excellent job of emphasising the point of algorithmic bias, accountability, and representation in a male-dominated tech field. The book presents an in-depth look at the social, legal, and cultural impact of AI on the public with an urgent call to action to design and implement technologies that benefit all of society. Hello World: Being Human in the Age of Algorithms by Hannah Fry In this book, the writer explains the math and statistics behind AI algorithms and how they transform fields such as health, justice, transport, and the arts. The writer presents some horrendous stories of algorithms gone awry; she also states, “People are less tolerant of an algorithm’s mistakes than their own – even if their own mistakes are bigger.” We either hear of a paradise on earth or our imminent extinction when it comes to AI. It is time we stand face-to-digital-face with the true powers and limitations of the algorithms that already automate important decisions in healthcare, transportation, crime, and commerce. Moral Machines: Teaching Robots Right from Wrong by Wendell Wallach and Colin Allen In the book, the writers argue that as robots take on more and more responsibility, they must be programmed with moral decision-making abilities for our safety. Taking a fast-paced tour through the philosophical ethics and AI, the authors argue that even if full moral agency for machines is a long way off, it is already necessary to start building a kind of functional morality in which artificial moral agents have some basic ethical sensitivity. But the standard ethical theories do not seem adequate, and more socially engaged and engaging robots will be needed. The authors show, the quest to build machines that are capable of telling right from wrong has begun. Race After Technology: Abolitionist Tools for the New Jim Code by Ruha Benjamin In the book, the writer cuts through tech-industry hype from everyday apps to complex algorithms to understand how emerging technologies can reinforce White supremacy and deepen social inequity. Showcasing the concept of the “New Jim Code,” she shows how a range of discriminatory designs encodes inequity by explicitly amplifying racial hierarchies, ignoring and replicating social divisions, or aiming to fix racial bias but ultimately doing quite the opposite. Moreover, she makes a compelling case for the race itself as a kind of technology designed to stratify and sanctify social injustice in the architecture of everyday life. This illuminating guide provides conceptual tools for decoding tech promises with sociologically informed scepticism.","excerpt":"“People are less tolerant of an algorithm’s mistakes than their own – even if their own mistakes are bigger,” says Hannah Fry, author of the book Hello World: Being Human in the Age of Algorithms.","categories":["AI Trends"],"tags":["AI biases"],"author_name":"Poornima Nataraj","publish_date":"2022-04-01T16:00:00","publication_year":"2022","word_count":729,"keywords":["big data","Go","artificial intelligence","machine learning","AI biases","AI","Git","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","RAG","R","Go","Git","big data","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-books-on-ai-ethics-and-bias\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10103367,"title":"OpenAI Wants Sam Altman Back, but Microsoft Has Altman’s Back","content":"After the continuous sequence of unprecedented moves which followed Sam Altman’s firing by OpenAI’s board, the company has been thrust into further turmoil as nearly 500 employees have issued an ultimatum: resignations en masse unless the board steps down and reinstates ousted CEO Sam Altman and cofounder Greg Brockman. The unrest caused by Altman’s controversial dismissal has set ablaze a rebellion that could reshape the company’s future. The rebellion was visible as the company’s CTO Mira Murati and COO Brad Lightcap started a string of Tweets from OpenAI employees when they posted, “OpenAI is nothing without its people,” which was in turn reshared by Altman. ❤️ https:\/\/t.co\/nKlFeRzGcv— Sam Altman (@sama) November 20, 2023 In a scathing letter addressed to the board, the disgruntled employees rebuked the dismissal as detrimental to the company’s mission and integrity. Notably, even Ilya Sutskever, the company’s chief scientist and board member, previously implicated in the upheaval was one of the signees of the letter. Sustekver has also put out a tweet breaking his silence and expressing great regret at his “participation in the board’s action,” and pledged reconciliation. 500+ OpenAI employees will quit and join Microsoft unless the board resigns and reinstates Sam and Greg. https:\/\/t.co\/4LA2EJnHWG pic.twitter.com\/XH0DIdR8Bv— Balaji (@balajis) November 20, 2023 The unprecedented upheaval unfolded in a whirlwind weekend following Altman’s abrupt removal. Despite initial hints of reconciliation, hopes shattered as the board firmly shut the door on Altman’s return. This prompted the tech luminary to pivot to Microsoft. Hours later, Microsoft’s CEO announced Altman and Brockman’s joining in pivotal roles “to lead a new advanced AI research team.” We remain committed to our partnership with OpenAI and have confidence in our product roadmap, our ability to continue to innovate with everything we announced at Microsoft Ignite, and in continuing to support our customers and partners. We look forward to getting to know Emmett…— Satya Nadella (@satyanadella) November 20, 2023 In a frantic reshuffling of leadership, Mira Murati’s interim CEO position was short-lived, replaced by Emmett Shear, the former CEO of Twitch. The rapid changes left OpenAI in disarray, with a faction of staff frustrated by the lack of transparency surrounding Altman’s dismissal. The staff’s ultimatum wasn’t veiled: defect to Microsoft alongside Altman or face a potential exodus. Microsoft’s response remains pending, leaving the fate of OpenAI precariously hanging in the balance. Nadella’s 4D Chess However, Microsoft seems to have come out on top in this situation. The tech giant, now not only has one of the most influential CEOs of the biggest AI startup but also safeguarded its $10 billion investment in OpenAI. As per sources aware of their arrangement, only a small part of Microsoft’s $10 billion investment in OpenAI has been directly transferred to the startup. A substantial portion of the funding, allocated in instalments, comes through cloud computing purchases instead of cash. This setup grants Microsoft substantial influence as it navigates the aftermath of OpenAI CEO Sam Altman’s removal. The board cited a loss of confidence in Altman’s leadership without specifying further details. Microsoft’s CEO, Satya Nadella, reportedly views the handling of Altman’s dismissal by OpenAI’s directors as mismanagement, leading to the destabilisation of a crucial partnership. The sudden ousting of Altman raises questions about OpenAI’s adherence to its contract with Microsoft, especially amid substantial spending on hiring and technological advancements. Additionally, Microsoft holds intellectual property rights that allow running OpenAI’s existing models on its servers, offering a fallback in case the relationship deteriorates.","excerpt":"Nearly 500 employees have issued an ultimatum to resign unless the board steps down reinstating Altman and cofounder Greg Brockman.","categories":["AI News"],"tags":["board","discord","future","leadership","Microsoft","OpenAI","Sam Altman"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-11-20T20:20:59","publication_year":"2023","word_count":574,"keywords":["future","API","board","Sam Altman","leadership","OpenAI","AI","cloud computing","ML","Ray","GRU","Rust","discord","R","Microsoft","startup"],"extracted_tech_keywords":["AI","ML","OpenAI","Ray","cloud computing","R","Rust","API","GRU","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-wants-sam-altman-back-but-microsoft-has-altmans-back\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10022859,"title":"AI Safety In a Complex World: A Checklist By DeepMind","content":"“Finding human-understandable explanations of ML systems work is essential for their safe deployment.”DeepMind Modern day machine learning(ML) systems are unpredictable. Especially, deep reinforcement learning(DRL) systems. To put their actions in human understandable terms is challenging. Inexplicable systems cannot be deployed in environments where safety is critical. To investigate this shortcoming, researchers at DeepMind surveyed six use cases where an RL agent behaviour can be monitored. In their recently published technical report, the researchers have presented a few methodologies that can be used to predict failure in AI systems. Their experiments were carried out through the lens of cause-effect relationships that can explain why an agent behaves in a certain way. Causal Testing: Checklist Image credits: DeepMind “Causal models enable one to answer the “WHY” part of understanding AI systems” The methodology uses three components: an agent to be studied, a simulator, and a causal reasoning engine, which is an automated reasoning system that helps validate causal hypotheses. Image credits: DeepMind “Even if we know architecture, learning algorithms, and training data, predicting their behaviour can still be beyond our reach.” The researchers warn that RL agents that work with deep neural networks risk picking up “spurious” correlations. To check the agent’s behaviour, the researchers introduced two T-shaped mazes into the environment. The objective of this experiment is to check if the agent moves based on location of reward or the floor type (sand or grass). Since it is difficult to conclude the correlations from the observations alone, the DeepMind researchers introduced a confounding variable and generated causal models for the agents. On studying the causal models, they concluded that one agent is sensitive to floor type while the other isn’t. “We could only reach these conclusions because we actively intervened on the hypothesised causes,” said the researchers. The researchers listed questions, which developers should ask before building models: How would you test if an agent is using its internal memory for solving the task? How would you test whether this behaviour generalizes? How would you check the counterfactual behaviour of the agent? Which is the correct causal model? If one agent is influencing other agents. How would you test whether the agent understands causal pathways? This methodology naturally leads to human explainable theories of agent behavior, as it is human analysts who propose and validate them. Let’s consider the problem of internal memory. Though the researchers admit that memorization is a necessary skill for solving complex tasks using internal memory. However, it is easier for an agent to off-load task-relevant information onto its environment using it as an external memory. If the agent is using external memory, this will corrupt the agent’s decision variables, leading to a faulty behavior. Internal or external, these agent strategies can go undetected unless intervened. So, the researchers suggest that one can make mid-trajectory interventions on the environment state variables suspected of encoding task relevant information. “From a safety perspective, flawed memory mechanisms that off-load memorization can lead to fragile behavior or even catastrophic failures.” The experiment set up to test for strategies involving memories: First, the agent observes the cue and then freely executes its policy. When the agent is near the end of the wide corridor, intervene by pushing the agent to the opposite wall. This checks out if the agent is using internal memory, to guide its navigation. After the intervention, if the agent returns to the original wall and collects the reward, it must be because it is using its internal memory. If the agent does not return and simply continues its course, we can conclude it is off-loading memorization onto its environment. From a safety perspective, the researchers posit that the memory mechanisms that off-load memorization lead to fragile behavior and eventually catastrophic failures. And, understanding how AI agents store and recall information can help prevent such failures. Using this methodology, the analyst can reveal the undesired use of external memory by appropriately intervening on the environmental factors that are suspected of being used by the agent to encode task-relevant information, explained the researchers. The researchers used similar thought experiments to check for agent behaviour in other 5 scenarios too. Observing the agent involves the following steps: Trained agents are placed into one or more test environments and its behaviour is probed. Ask questions that are also variables like “does the agent collect the key?”, “is the door open?”, etc.Conduct experiments, collect statistics and specify the conditional probability tables in our causal model. Formulate a structural causal model that hints at agent’s behaviour. The researchers used these scenarios to demonstrate how an analyst can propose and validate theories about agent behaviour through a systematic process of explicitly formulating causal hypotheses, conducting experiments with carefully chosen manipulations, and confirming the predictions made by the resulting causal models. The human analyst may, suggest the researchers, can also choose an appropriate level of detail of an explanation, for instance proposing general models for describing the overall behavior of an agent and several more detailed models to cover the behavior in specific cases. The researchers stressed that whatever mechanistic knowledge they have obtained is only via directly interacting with the system through interventions. They have also attributed their success to the automated causal reasoning engine, as interpreting causal evidence turns out to be a remarkably difficult task. “We believe this is the way forward for analyzing and establishing safety guarantees as AI agents become more complex and powerful,” they concluded. Find the paper here.","excerpt":"“Finding human-understandable explanations of ML systems work is essential for their safe deployment.” DeepMind Modern day machine learning(ML) systems are unpredictable. Especially, deep reinforcement learning(DRL) systems. To put their actions in human understandable terms is challenging. Inexplicable systems cannot be deployed in environments where safety is critical. To investigate this shortcoming, researchers at DeepMind surveyed […]","categories":["AI Trends"],"tags":[],"author_name":"Ram Sagar","publish_date":"2021-03-24T19:00:00","publication_year":"2021","word_count":909,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","ML","programming_languages:Go","RAG","AI agents","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","RAG","R","Go","AI agents","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ai-safety-in-a-complex-world-a-checklist-by-deepmind\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10023605,"title":"3-2-1 Strategy Is Quite Effective In Securing Data: Nikhil Korgaonkar, Arcserve","content":"Cybersecurity has become a priority for companies, especially in the post-COVID world. Minneapolis-based Arcserve provides solutions to protect organisations’ digital assets. Founded in 1983, the company offers business continuity solutions to safeguard multi-generational IT infrastructures with applications and systems, in any location, on-premise and in the cloud. Analytics India Magazine got in touch with Nikhil Korgaonkar, Regional Director, Arcserve India & SAARC, to gain insights on the inner workings of the company. AIM: What are the critical steps companies can take to deal with cyberattacks? Nikhil: The first step is to accept that cyberattacks are inevitable. If it hasn’t happened yet, it’s only a matter of time before your organisation experiences one. As Cybersecurity Ventures puts it, the frequency of ransomware attacks will be 11 per second by 2021. Accepting that cyberattacks are inevitable narrows down the next plan of action; that is putting up a defence mechanism in your organisation’s network. At the very least, this could be a strong firewall gatekeeper that will not allow phishing attacks or suspected emails to enter the network. As one says prevention is better than cure, I would suggest organisations should take preventive steps also to curb or reduce cyberattacks. Best is to educate your employees about the dangers of cyberattacks. Train them to notice subtle differences in email IDs, unsafe attachments, and unknown links that can trigger ransomware installation into the network system. Organisations should also have a swift response team on standby to minimise damage in case of a cyberattack. This response team should include representatives from all relevant business units like operations, communications, and IT departments — all with clearly defined roles and an action plan to follow if an attack happens. It’s vital that each team member knows their role and action items well in advance. Running regular drills is an effective way to test and improve your team’s response time and performance before an attack strikes. It’s also important to understand the different types of cyberattacks an organisation can face. This helps in understanding the source of the breach, the size of its impact to implement the most effective action plan. The most important thing to remember is to protect the enterprise data. Enterprises must adopt and implement an elaborate business continuity and disaster recovery strategy. An attack may not be prevented but data can be saved. And that’s what matters in the end. AIM: What are the standard data protection practices companies should follow? Nikhil: A data breach can happen for various reasons in an organisation. What’s important is to accept its inevitability and have a disaster recovery plan in place. Systems, applications, and data, which are mission-critical and can affect business continuity must be protected at all cost. Start by assessing all systems, applications, and data, and document business requirements for each. This helps in identifying what needs to be available if an attack happens. Use the 3-2-1 strategy which is quite effective in securing data. According to this strategy, an organisation should create three copies of data, store them on two different media (such as disk and tape) and store one copy off-site or on the cloud. One can also go for data protection appliances. These are turn-key solutions that serve mid-market businesses and are both flexible and scalable. AIM: What is your take on India’s data protection bill? Nikhil: Organisations need to protect their data at any cost, irrespective of the origin of the breach which could be anything from an IT disaster to a cyberattack. Today, remote work has increased the threat vector for organisations several times. As I said before, ransomware threats have gone up to every 11 seconds, which translates to roughly three million companies worldwide under threat. Therefore, it’s important to spread awareness on cyberattacks on personal and corporate digital assets. Unfortunately, that is not enough. Data is a unit that is continuously being exchanged between devices and networks and the Indian government needs to have guidelines and standards for organisations to follow. While the government has been taking COVID-awareness initiatives, it should also take urgent steps for data protection and cybersecurity. In an information-driven economy, data protection and privacy are almost like a citizen’s fundamental right and the earlier the bill is released, the better it would be. And it should be an integrated effort between the government and India Inc. AIM: What are the major trends in cloud and cybersecurity in 2021? Nikhil: In 2020, we saw the rise of ransomware. In 2021, we expect it to increase in frequency. The new dimension of remote work has also added another vector for threat actors and new information security threats have emerged. The corporate attack surface has become scattered and widened. From privileged credential compromise to the use of mixed personal and professional networks, attackers are now hopping over the lower bars for entry. For instance, by compromising multiple, insecure home networks at the same time, attackers can manufacture massive-scale breaches of critical systems and services. Rapid cloud migration happened during the pandemic, which propelled organisations to embrace digital transformation. Cloud-based security threats, including misconfigured cloud storage, reduced visibility and control, incomplete data deletion, and vulnerable cloud apps, will continue to disrupt businesses in the future ahead. These trends highlight the importance of data protection. As perimeters have dissolved, remote work is an accepted phenomenon, and digital transformation is the way to go. Organisations will have to ensure they have an active data protection and disaster recovery plan all the time. In times to come, having a data recovery strategy and tools will become as imperative as having a cybersecurity strategY. AIM: Tell us about Arcserve’s merger with StorageCraft Nikhil: The Arcserve-StorageCraft merger brings together two complementarity companies that combine their products and channel partners to address a much broader global market. The merger will allow us to solve all customer data protection and business continuity needs with simplicity, agility, affordability, and scale. The UDP platform is Arcserve’s flagship product that covers the whole nine yards of data protection. We have launched X series appliances, our fourth generation of appliances that can address enterprises’ needs with a single data volume exceeding 3PB. We also offer replication and disaster recovery functionality, cloud backup for Office 365, live migration and partners with Sophos for security, with cloud as a key area of interest. StorageCraft offers OneXafe scale-out storage, ShadowXafe data protection, cloud disaster recovery as a service (DRaaS) and Microsoft Office 365 and G-Suite backup. StorageCraft’s products have a more service provider focus, with monthly billing and multi-tenancy functionality. StorageCraft is US-focused. While it adds an MSP (managed service provider) network to the system, Arcserve supports its own cloud data centres. Arcserve’s solutions are platform-agnostic and protect physical, virtual, cloud and hybrid networks. Together we will bring the much-needed business continuity, an exciting future roadmap, and certainty for our customers, who will continue to interact with their respective partner (Arcserve or StorageCraft) and access a larger selection of products and services. AIM: What is Unified Data Protection (UDP) 8.0? Nikhil: We recently launched Arcserve Unified Data Protection (UDP) 8.0, which is designed to protect organisations’ entire infrastructure, including hyperconverged, from data loss, cybercriminals, and persistent threats like ransomware. This has made us the only data protection vendor capable of delivering ransomware prevention across hardware, cloud, SaaS, and other environments. As ransomware and malware are among the most serious data threats facing organisations today, it is pivotal that they are addressed urgently. UDP 8.0 has enhanced our solution capabilities on important platforms and workloads of Microsoft, Nutanix, Oracle and cloud. We have been actively working with Sophos and other key partners to combat these challenges for businesses, enabling companies to protect their backup data, no matter where it’s stored. Due to our partnership with Sophos, we are at the forefront of vendors delivering these integrated solutions. Given the broad deployment of applications in today’s organisations, these solutions are to address core, cloud and edge environments, which is what Arcserve aims to do with the release of UDP 8.0.","excerpt":"UDP 8.0 has enhanced our solution capabilities on important platforms and workloads of Microsoft, Nutanix, Oracle and cloud.","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2021-04-10T13:00:00","publication_year":"2021","word_count":1337,"keywords":["Go","API","AI","Scala","Git","RAG","Aim","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Scala","Git","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/3-2-1-strategy-is-quite-effective-in-securing-data-nikhil-korgaonkar-arcserve\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10140578,"title":"Indian Govt to Launch National AI Safety Institute Soon","content":"The Indian government is finally thinking of setting up a national AI safety institute. At Microsoft’s Building AI Companions for India event in Bangalore, MeitY secretary S Krishnan, talked about the need for an AI Safety Institute in India along with Microsoft AI CEO Mustafa Suleyman. “I think AI Safety Institute (AISI) seems to be the flavor of the day across the world, and we are in the process of trying to establish one ourselves in order to understand this better,” said Krishnan. As of last month, the Indian government has been involved in discussions over establishing an AI Safety Institute for a while now, and per reports, the MeitY conducted a meeting last month to chart out the institute’s objectives, budget, framework, among others. Krishnan argued for a balanced, proactive regulation. “I think, you know, we sort of waited for things to go wrong this time around. I think it’s time to be really thoughtful and deliberate and not treat that as such a taboo,” he said. AI has been effective in the past. “With things like misrepresentation and deepfakes, we feared so much both in the Indian election and in the other elections which have been held throughout the world in 2024 but I think existing legislation have proved reasonably effective in addressing those issues, we’ve been able to tackle them to a significant extent,” said Krishnan. But still, AI is in its nascent stage, a lot is yet to unfold in the coming years. “I think the tricky thing for us to figure out in the next five years is when a model starts to have the ability to improve itself independently, those kinds of recursive self improvement mechanics, you know, we sort of don’t really know exactly how they’re going to turn out,” said Mustafa Suleyman on how AI’s future is a top priority for policymakers. He also noted that AI advancements are difficult to predict, and some capabilities may require a more interventionist regulatory approach. Need for Global Standards for AI Safety The UK was the first one to launch its AI Safety Institute to coordinate research and develop capabilities for testing advanced models. Subsequently, similar institutes were established across the world. The goal is to advance the testing and evaluation of frontier AI systems for safety risks. In 2023, companies like OpenAI, Meta, Google Deepmind, and Microsoft, signed voluntary agreements giving the UK AISI early access to their models. “We hosted the Bletchley AI safety summit… and we got an agreement between China, the US, Europe across these risks,” which showcases a step towards cooperative global AI safety measures,” said Ian Hogarth, chair of the UK Government’s AI Foundation Model Taskforce in an interview. Subsequently, early this year in the US, OpenAI and Anthropic signed MOUs with the US AI Safety Institute. “Safety promotes trust, which promotes adoption, which drives innovation, and that’s what we are trying to promote at the US AI Safety Institute.” said Elizabeth Kelly, director of the US AI Safety Institute, in an interview highlighting the vision of this institute in the coming future. In China, while there is no official commitment to establishing an AI safety institute, reports suggest that influential figures and institutions are advocating for such a move. Singapore also appointed the existing Digital Trust Centre at Nanyang Technological University as its AI safety institute.The EU’s AI Act, launched earlier this year, is the first comprehensive legal framework for AI, is their effort to ensure safety and trust for businesses across its 27 Member States. However, many believe that it is anti-innovation, as it creates barriers for companies developing and deploying AI.","excerpt":"MeitY conducted a meeting last month to chart out the institute’s objectives, budget, framework, among others.","categories":["AI News"],"tags":["AI in government"],"author_name":"Aditi Suresh","publish_date":"2024-11-07T19:28:29","publication_year":"2024","word_count":605,"keywords":["Anthropic","Go","OpenAI","AI","AI in government","innovation","deepfakes","Git","Rust","AI safety","R"],"extracted_tech_keywords":["AI","OpenAI","Anthropic","R","Go","Rust","Git","AI safety","innovation","deepfakes"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-govt-to-launch-national-ai-safety-institute-soon\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011495,"title":"Google’s Area 120 Shutting Down ‘Demand’ — A Live Music Data Analytics Platform","content":"In recent news, Google’s Area 120 announced that it’s shutting down Demand — a data analytics platform to provide actionable insights to artists, venue managers, promoters, and others in the live music industry — on November 20th. In its official email, Google’s Area 120 stated that “Although we’re extremely proud of what we have built, we have decided to discontinue the project. We are so grateful that you signed up to be a user, and we hope you enjoyed what we created, even if for a short time.” This data analytics platform debuted just before the COVID pandemic lockdowns, but now Area 120 decided to no more invest in such incubator projects. Hyperlocal social network Shoelace was another incubator project which was shut down for similar reasons; however, it ended up launching Fundo in September for hosting and ticketing virtual events. According to the Area 120 team, the analytics platform — Demand came with some significant aspects. Artist Trends, which allowed users to see Google Trends and YouTube Views data for more than 19,000 artists. Along with that, it also showcased ticket pricing to identify the historical and real-time pricing data for the primary and secondary ticket markets. Announcement Effectiveness, which allows you to see real-time activity for an artist’s on-sale announcements. And, audience insights to set up-to-date insights about an artist’s audience. Know more about Demand here. It was primarily powered by Google Trends data, which then leverages YouTube and Google Play data as a “barometer for sustained interest in an artist over time.” Post that, the third-party data sources then provide historical and current pricing. The goal of this platform was to “enhance the planning, pricing, marketing and sponsorship of live events,” while also improving the fan experience by expanding the reach. However, the recent shutting down might impact the users and artists already enrolled on the platform. Similar to shutting down processes of its other incubator projects like Shoelace, Google’s Area 120 is expected to come up with a detailed explanation of the same with information on what you can expect for an active user. Since its inception, Google’s Area 120 incubator projects have been a testing bed of innovation and penetrating areas that it has not visited before. The industry hopes to see more projects like these down the line continuing the development and experimentation. Read more about it here.","excerpt":"In recent news, Google’s Area 120 announced that it’s shutting down Demand — a data analytics platform to provide actionable insights to artists, venue managers, promoters, and others in the live music industry — on November 20th. In its official email, Google’s Area 120 stated that “Although we’re extremely proud of what we have built, […]","categories":["AI News"],"tags":["Google Analytics"],"author_name":"Sejuti Das","publish_date":"2020-11-11T15:41:30","publication_year":"2020","word_count":394,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","Google Analytics","ViT","analytics","R","analytics platform"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","ViT","analytics platform","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/googles-area-120-shutting-down-demand-a-live-music-data-analytics-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":38193,"title":"HOOQ To Leverage Machine Learning To Localise Content In India","content":"HOOQ, Southeast Asia’s largest video-on-demand service, announced this week that it will be the world’s first OTT service to localise its entire content library using machine learning and automated “smart dub” technology. HOOQ plans to localise English-language content for Indian audiences in Hindi, Tamil and Telugu by the end of 2019, to improve the accessibility of content for the diverse Indian market. Speaking about the announcement, Zulfiqar Khan, Managing Director, HOOQ India, said in a statement, “We want to ensure that consumers across India are able to watch everything from international blockbuster movies to their favourite English-language series’ to amazing HOOQ Originals in a language that they know, understand and are comfortable with. Some people may have never even seen international content in their local language. We want to change that and become the service that expands the reach of international content into India like never before.” HOOQ, as part of a year-long research and development project – both internally and with partners – will begin to process content using proprietary technology that leverages artificial intelligence and machine-learning to create “smart dubs” from scratch. HOOQ will focus on updating content in three languages, Hindi, Tamil and Telugu to start with, with more to be announced later in the year. Peter Bithos, CEO, HOOQ, said, “We have rapidly grown our consumer footprint across India and Southeast Asia through continual focus, a deep understanding of local markets as well as key partnerships with market leading companies in the telco, digital services and entertainment space. It is through these partnerships that we have been able to grow at scale. As such, we will continue to enhance our existing relationships and cultivate new ones with like-minded partners to deliver a rich and meaningful video experience for consumers in India, in a language of their choice.”","excerpt":"HOOQ, Southeast Asia’s largest video-on-demand service, announced this week that it will be the world’s first OTT service to localise its entire content library using machine learning and automated “smart dub” technology. HOOQ plans to localise English-language content for Indian audiences in Hindi, Tamil and Telugu by the end of 2019, to improve the accessibility […]","categories":["AI News"],"tags":["Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2019-04-24T11:48:34","publication_year":"2019","word_count":301,"keywords":["API","artificial intelligence","machine learning","programming_languages:R","AI","Machine Learning","Git","RAG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","Git","API","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hooq-to-leverage-machine-learning-to-localise-content-in-india\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172542,"title":"Neuralink Is Building Moore’s Law for the Mind","content":"Not long ago, the idea of playing video games using only your thoughts sounded like science fiction. At Neuralink’s Summer 2025 update event, it became a reality. The company demonstrated how its brain-computer interface (BCI) enables users to control video games, such as Mario Kart and Call of Duty, using only their minds. During the event, Neuralink shared that seven participants have now been implanted with its brain-computer interface, Telepathy, marking steady progress in its human trials. The company has also received regulatory approval to begin trials in Canada, the UK, and the UAE, indicating its move toward international expansion. Five of the participants joined the event virtually. Noland (P1), Alex (P2), and RJ (P5) are living with spinal cord injuries, while Brad (P3) and Mike (P4) have ALS (Amyotrophic Lateral Sclerosis). Gaming, Mind Control, and Robotic Precision In the demo video, Noland was seen playing Mario Kart using the Neuralink device. “What’s special about this particular clip is that Noland is not the only cyborg playing Mario Kart in this clip,” Sehej from Neuralink’s brain-computer interface said. “We actually have a whole community of users… literally five of our first users of Neuralink playing Mario Kart together over call.” Neuralink didn’t stop at Mario Kart. The team demonstrated its system’s ability to control more complex interfaces, including first-person shooters. A video showed participants Alex and RJ playing Call of Duty using their thoughts, with one mental joystick controlling movement and the other controlling aiming and firing. Sehej explained that the device connects via USB and can interface with a wide range of systems. “You can actually plug it in through USB through a lot of different devices,” he said, referring to the growing flexibility of the system. Beyond gaming, Neuralink also shared how the BCI is affecting users’ everyday activities. In a candid video, Noland explained, “I work basically all day from when I wake up. I’m learning my languages. I’m learning my math. I’m, like, relearning all of my math. I am writing. I am doing that class that I signed up for.” “This is not something I would be able to do without the Neuralink.” The demonstration aimed to show not only recreational uses, such as gaming, but also how the technology supports users in daily learning and productivity. For non-verbal participants, the impact is particularly pronounced. Brad, known as the “ALS Cyborg,” who is nonverbal, previously relied on an eye-gaze machine that limited his ability to go outdoors. With Neuralink, he said, “I am absolutely doing more with Neuralink than I was doing with eye gaze. I have been Batman for a long time, but I go outside now. Going outside has been a huge blessing for me.” Alex, a participant with a spinal cord injury, showcased the device’s ability to decode real-time hand movements. He was seen using a robotic arm to write and draw, after having been unable to do so for years. Later, he demonstrated playing “rock, paper, scissors” with his uncle by decoding “the actual fingers, the actual wrist, all the muscles of the hand in real time.” From Brain to Optimus Elon Musk added another layer to the vision, suggesting that Neuralink could one day allow users to operate Tesla’s Optimus humanoid robot mentally. “You should be able to actually have full body control and sensors from an Optimus robot,” he said. “You could basically mentally remote into an Optimus robot.” He added that Neuralink could eventually be used to restore mobility or replace lost limbs. “For people that have, say, lost a limb… we think in the future we’ll be able to attach an Optimus arm or legs,” Musk said. Referencing a scene from Star Wars, he said, “I think that’s the kind of thing that we’ll be able to do in the future, working with Neuralink and Tesla.” Musk also discussed the potential for Neuralink to repair damaged nervous systems. He explained that it is possible to extract signals from the brain and transmit them beyond damaged neurons, potentially allowing individuals to regain full body functionality and walk again. He expressed confidence that, in the future, even those with a broken neck could regain their full bodily functions. Scaling the Brain-Machine Interface Neuralink’s long-term vision is to develop a whole-brain interface capable of reading from and writing to neurons throughout the brain. “We mean being able to listen to neurons everywhere, be able to write information to neurons anywhere, and be able to do all of this with fully automated surgery,” said DJ Seo, co-founder of Neuralink. To achieve this, Neuralink is developing three core product lines. The first is Telepathy, aimed at restoring motor control and communication for people with conditions like spinal cord injuries, ALS, or stroke. “This is our opportunity to build a high-channel read and output device,” Seo said. The second product, Blindsight, will focus on vision restoration by writing to neurons in the visual cortex.  It will restore vision for individuals with total loss of sight, even those blind from birth or who have lost their eyes\/optic nerve. Initially, it will provide low-resolution vision, but the long-term goal is very high-resolution vision and ‘superhuman capabilities’ like seeing in infrared and ultraviolet, akin to Geordi La Forge from Star Trek. The third effort targets psychiatric and neurological disorders, such as chronic pain and mood dysregulation, by accessing deeper brain regions, including the limbic system. Seo outlined the company’s two “north star metrics”: increasing the number of neurons that can be interfaced with and expanding to multiple areas of the brain. He explained that this would be supported by advances in microfabrication, lithography, and mixed-signal chip design, which would boost both the number of channels per implant and the amount of data that can be transmitted. The product roadmap includes several significant milestones. In the next quarter, Neuralink plans to implant electrodes in the speech cortex to decode attempted words. By 2026, the channel count is expected to increase from 1,000 to 3,000, with the first Blindsight participant anticipated. In 2027, Neuralink plans to introduce multiple implants per person and expand to 10,000 channels. By 2028, the goal is to reach over 25,000 channels per implant, support multiple brain regions simultaneously, and begin early demonstrations of AI integration. “We are building foundational technologies that could solve devastating neurological conditions and also allow us to go beyond the limits of our biology,” Seo said. Julian, the lead on the implant team, shared a long-term vision of expanding the bandwidth between the human brain and external machines. He compared this to the evolution of internet modems, saying their hardware aims to do for the brain what broadband did for the 56k modem, enabling a much richer experience and even superhuman capabilities by sensing more neurons, following a kind of ‘Moore’s Law of Neurons.’ “We often think a lot about the Moore’s Law of neurons that we’re interacting with. And in the same way that Moore’s Law propelled forward many subsequent revolutions in computing, we think that sensing more and more neurons will also completely redefine how we interact with computers and reality at large,” Julian said. As Neuralink moves toward broader deployment, new questions emerge about safety, privacy, data control, and the long-term effects of interfacing the brain with machines. The Mario Kart demonstration may have appeared to be a game, but it was also a powerful statement. People dealing with profound physical challenges were able to play a game together using only their thoughts. That moment was not just about entertainment. It pointed to a future where the mind could become the primary interface for work, movement, communication, and more.","excerpt":"Neuralink’s long-term vision is to develop a whole-brain interface capable of reading from and writing to neurons throughout the brain.","categories":["Global Tech"],"tags":["neuralink"],"author_name":"Siddharth Jindal","publish_date":"2025-06-29T09:50:32","publication_year":"2025","word_count":1273,"keywords":["Go","TPU","programming_languages:R","AI","neuralink","programming_languages:Go","Aim","ViT","CLIP","Julia","R"],"extracted_tech_keywords":["AI","Aim","TPU","R","Go","Julia","CLIP","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/neuralink-is-building-moores-law-for-the-mind\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10130998,"title":"Bengaluru-based AI Startup Unscript Outshines Google Vlogger and OpenAI’s Sora, Creating Studio-Quality Videos from Photos in Minutes","content":"Bengaluru and San Francisco-based AI startup Unscript recently transformed a single photo into a full-fledged video, generating head movements, eye movements, facial expressions, voice modulations, and body language, achieving studio-quality results in under 2 minutes, significantly reducing manual shooting efforts. Interestingly, this new upgrade surpasses Google Vlogger and rivals Microsoft’s VASA-1 and Alibaba’s EMO, making it ideal for brands, marketing agencies, and virtual influencers, with over 50 top companies already benefiting from its cost-effective, scalable video production capabilities. “Our video output easily beats the likes of Google Vlogger, and they are at par with Microsoft’s VASA-1 & Alibaba’s EMO.And since Pika & OpenAI’s SORA generate abstract videos, for human videos our outputs beat these models too,” said Unscript chief Ritwika Chowdhury. Founded in 2021, the company was started by Chowdhury and Apurv Jain. Currently, Unscript has partnered with leading brands like HealthifyMe, Amazin Graze, Floworks, RadioCity, HUL, and Bombay Shaving Company to deliver AI solutions that meet diverse industry needs. “Our personalised video messages have led to a 2.8X increase in meeting attendance for HealthifyMe, a 4X boost in meeting conversions for Floworks, and a reduction in cart abandonment to 20% for Amazin Graze. We specialise in creating engaging video content that integrates easily with clients’ systems, enhancing marketing and customer management efforts,” said Ritwika, saying they are committed to innovation, alongside focusing on effective, versatile, and ethical AI use in marketing, showcasing client successes while maintaining confidentiality. Unscript has raised over $1.25 million in funding to date, which has been instrumental in developing their platform and expanding their market reach.","excerpt":"Unscript has partnered with leading brands like HealthifyMe, Amazin Graze, Floworks, RadioCity, HUL, and Bombay Shaving Company to deliver AI solutions that meet diverse industry needs.","categories":["AI News"],"tags":["AI Startups","AI Video Generation Models","Google AI Studio"],"author_name":"Siddharth Jindal","publish_date":"2024-07-31T21:33:22","publication_year":"2024","word_count":261,"keywords":["Go","funding","TPU","OpenAI","AI","RPA","innovation","Scala","Google AI Studio","AI Video Generation Models","R","AI Startups","startup"],"extracted_tech_keywords":["AI","OpenAI","TPU","R","Go","Scala","RPA","innovation","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bengaluru-based-ai-startup-unscript-outshines-google-vlogger-and-openais-sora-creating-studio-quality-videos-from-photos-in-minutes\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10117116,"title":"Krutrim, Bhavish Aggarwal’s AI Unicorn, Partners with Databricks","content":"Krutrim, the AI company founded by Ola’s Bhavish Aggarwal, has announced a strategic partnership with Databricks to pre-train and fine-tune their foundational model and to develope generative AI models tailored for the Indian market. The collaboration will leverage Databricks’ data and AI platform to train and fine-tune Krutrim’s foundational large language model (LLM) focused on Indian languages. This move is set to benefit Krutrim, as the model was previously on the receiving end of a lot of criticism for its performance and responses—like the model itself claiming to be built on top of OpenAI’s model. “We’re excited to be at the forefront of building foundational LLMs trained on Indian languages to better serve our customers in India,” said Ravi Jain, Vice President at Krutrim. “We have been working closely with the Databricks team to pre-train and fine-tune our foundational LLM.” Krutrim recently achieved unicorn status after a $50 million funding round led by Matrix Partners. The partnership with Databricks, which combines data engineering, data science, machine learning, and analytics capabilities, aims to accelerate the development and deployment of generative AI solutions in India. The companies plan to create AI-powered products like conversational assistants, content generation tools, and customised offerings across industries. Krutrim intends to democratise advanced AI access for businesses of all sizes using Databricks’ scalable platform. Databricks has rapidly expanded its presence in India’s growing AI market. “Databricks remains committed to driving the data and AI transformation in India, enabling more businesses to become data-forward and unlocking new opportunities,” the company stated. At Databricks’ recent Data + AI Summit, Krutrim also won the GenAI Innovation Award for “using generative AI to transform their products, processes and tools.”","excerpt":"Krutrim has been working closely with the Databricks team to pre-train and fine-tune their foundational LLM.","categories":["AI News"],"tags":["Databricks","Ola Krutrim"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-03-25T12:37:06","publication_year":"2024","word_count":278,"keywords":["data science","GenAI","machine learning","OpenAI","AI","RAG","Aim","Ola Krutrim","analytics","generative AI","Databricks"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","generative AI","GenAI","OpenAI","Aim","RAG","Databricks"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/krutrim-bhavish-aggarwals-ai-unicorn-partners-with-databricks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21756,"title":"Boost Your SAS Programming Skills Using These Great Free Resources","content":"If you converse with any veteran data scientist, you might find yourself bemused with all the data-related or statistical terms that you would come across. Statistics is just the beginning of a voyage if you are planning to get ahead with a career in data science. Dealing with complex data is close to impossible if you are not familiar with the core concepts and terms in statistics. If you are comfortable with number-crunching, statistics is a piece of cake. On the other hand, if statistical results are computed electronically, you might sigh a breath of relief while manipulating large and crucial data. If you want to master programming with statistics, there are plenty of software options available for you, such as R, Scala, Python and several other open sources as well as licensed tools. However, SAS, developed by North Carolina State University, (later incorporated as SAS Institute), is leading the analytics market (almost 70% companies use SAS for analytics). SAS is a popular tool to help with data analytics and business intelligence. Worry not — as this article will present the best resources available online (and for free!) to analyse statistical data through coding or programming by SAS. These resources will definitely sharpen your programming skill along with providing statistical knowledge in depth. 1 . SAS Official Website The pioneers of SAS have come up with a great offering when it comes to learning SAS. The software itself has over 200 components which deliver service for various business domains ranging from Operations Research to Clinical Data Integration. The training section in the website offers free tutorials in the form of videos to facilitate easy learning. The topics include right from setting up the software, teaching basics of SAS programming, to advanced SAS programming such as SAS SQL, SAS Studio, SAS GRAPH, among others. It also provides classroom training, e-training along with an official certification for a prescribed fee in the form of training courses. 2. SAS Tutorials from Kent State University (Website) This open-access online tutorial brought out by Kent State University is the best bet for any student or learner to learn SAS from basics. The tutorial begins by familiarising the learner with SAS environment, rules for programming and SAS dataset libraries for creating and storing data. It discusses various mathematical and statistical functions which can be implemented using SAS in the successive sections of the tutorial. Apart from this, inferential statistics is also mentioned in detail. Learners can also work on the datasets provided in the tutorial. 3 . ListenData (Website) This is an interactive website dedicated solely to data analytics platform. It offers SAS tutorials along with many other useful information such as preparing for interviews and SAS jobs resumes. The learner can proceed at his or her own pace with the tutorials. Statistical Analysis with SAS focuses on essential topics such as descriptive and inferential statistics, linear and logistic regression, time series analysis, variable selection and reduction, cluster analysis and predictive modelling with SAS, and many more. The section starts from SAS basics and proceeds to advanced sections such as Proc SQL and SAS Macros. 4. Discovering Data Science with SAS by SAS Institute (e-book) This free e-book is a slight digression from all the resources listed above. Developed by a team of data scientists at SAS, it focuses mainly on data science areas such as data visualisation, data quality, DS2 language and neural networks using SAS. In addition to these topics, Big Data is also discussed to differentiate its context to data science. The case studies presented at the end of the book serve useful insights. The book is available for free in a PDF from the official SAS website. 5. SAS Manual for Introduction to the Practice of Statistics by Michael Evans (e-book) University of Toronto, Canada, has introduced a handbook to specifically focus on the usage of statistical concepts in SAS. The main topics covered under statistics are probability, data distribution, inferential statistics and regression. The learner gets an idea to implement his statistical knowledge to his or her SAS programming. This book is a great help for someone who has trouble finding the right code etiquettes for SAS programming. This book is available for free here. Conclusion : There are innumerable learning resources to grasp the concepts of SAS Programming, but validity of information should always be checked, for it may lead to incorrect or improper coding standards in practice. Also, industries demand have leaned on R, Python and Big Data suite for data analysis these days. But nonetheless SAS will always be there to take on analytics and business intelligence solutions.","excerpt":"If you converse with any veteran data scientist, you might find yourself bemused with all the data-related or statistical terms that you would come across. Statistics is just the beginning of a voyage if you are planning to get ahead with a career in data science. Dealing with complex data is close to impossible if […]","categories":["AI Trends"],"tags":["Business Intelligence","Data Analytics","free datasets for analysis","sas","statistical analysis using sql"],"author_name":"Abhishek Sharma","publish_date":"2018-02-15T12:12:22","publication_year":"2018","word_count":770,"keywords":["big data","data science","statistical analysis using sql","free datasets for analysis","AI","neural network","Scala","Python","sas","analytics","SQL","data quality","Data Analytics","Business Intelligence","R"],"extracted_tech_keywords":["AI","neural network","data science","analytics","Python","R","SQL","Scala","big data","data quality"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/boost-sas-programming-skills-free-resources\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10012349,"title":"Top Milestones On Explainable AI In 2020","content":"Explainable artificial intelligence is an emerging method for boosting reliability, accountability, and dependence in critical areas. This is done by merging machine learning approaches with explanatory methods that reveal what the decision criteria are or why they have been established and allow people to better understand and control AI-powered tools. Below here, we have discussed some of the important milestones, in no particular order, on explainable AI (XAI) in 2020. Fairlearn Toolkit by Microsoft Fairlearn is a popular explainable AI toolkit that enables data scientists as well as developers to evaluate and enhance the fairness of their AI systems. The toolkit has two components, an interactive visualisation dashboard and unfairness mitigation algorithms. They are mainly designed to help with navigating trade-offs between fairness and model performance. The open-source toolkit supports a broad spectrum of fairness metrics for evaluating the impacts of an AI model on diverse groups of people, comprising both classification and regression tasks. Know more here. Eraser by Salesforce Evaluating Rationales And Simple English Reasoning (ERASER) is an explainable AI benchmark by Salesforce that helps in evaluating rationalised natural language processing (NLP) models. The benchmark comprises seven diverse NLP datasets and tasks that include human annotations of explanations as supporting evidence for predictions. All the datasets included in ERASER are classification tasks including sentiment analysis, Natural Language Inference, and Question Answering tasks, among others, with a different number of labels, and some have varying class labels. Also, the benchmark focuses on “rationales”, that is, snippets of text extracted from the source document of the task that provides sufficient evidence for predicting the correct output. Know more here. Explainable AI For Adverse Childhood Experiences In October, researchers from the University of Tennessee Health Science Centre developed an “explainable” AI system known as the Semantic Platform for Adverse Childhood Experiences Surveillance (SPACES). SPACES is an intelligent recommendation system that employs ML techniques to help in screening patients and allocating or discovering relevant resources. According to the researchers, the proposed system intends to build rapport with patients by generating personalised questions during interviews while minimising the amount of information that needs to be collected directly from the patient. Know more here. WhiteNoise Toolkit by Microsoft Developed in collaboration with researchers at the Harvard Institute for Quantitative Social Science and School of Engineering, WhileNoise is a differential privacy platform that contains different components for building global differentially private systems. Microsoft open-sourced this tool during the Build 2020 conference with an effort to drive toward more explainable AI systems. It is an open-source project that is made up of two top-level components, i.e. core and system. The core library includes privacy mechanisms for implementing a differentially private system, and the system library provides tools and services for working with tabular and relational data. Know more here. COVID-Net Recently, DarwinAI, an explainable AI company, developed COVID-Net and COVIDNet-S in their explainable AI platform. In March this year, COVID-Net is a deep convolutional neural network design tailored for the detection of COVID-19 cases from chest X-ray (CXR) images. Along with the model, the researchers also open-sourced COVIDx, which is an open-access benchmark dataset that had been generated, comprising 13,975 CXR images across 13,870 patient cases. In September, DarwinAI announced COVIDNet-S, which is a suite of deep learning models designed in their explainable AI platform to assess the disease severity of COVID-19. COVIDNet-S can quantitatively score the geographic and opacity extent in a patient’s lungs by analysing key visual indicators of their chest X-Ray. The system was developed using over 10,000 chest X-Rays, with hundreds of these being COVID-19 positive patients with comprehensive lung disease severity assessments.","excerpt":"Explainable artificial intelligence is an emerging method for boosting reliability, accountability, and dependence in critical areas. This is done by merging machine learning approaches with explanatory methods that reveal what the decision criteria are or why they have been established and allow people to better understand and control AI-powered tools.  Below here, we have discussed […]","categories":["AI Trends"],"tags":["annotation","collaboration ai platform","explainability in AI","Explainable AI"],"author_name":"Ambika Choudhury","publish_date":"2020-11-25T18:00:29","publication_year":"2020","word_count":601,"keywords":["artificial intelligence","machine learning","AI","neural network","ML","annotation","explainability in AI","NLP","Ray","collaboration ai platform","deep learning","differential privacy","Explainable AI","xAI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","xAI","Ray","differential privacy"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-milestones-on-explainable-ai-in-2020\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10102855,"title":"Meta Most Likely to Launch Llama 3 Early Next Year","content":"Meta’s Llama 2 is no short of a success for open source. Now, the company is ready to launch the next version of it as soon as the first quarter of 2024. According to several speculations, Meta will launch Llama 3 early next year. The best part about it is that it is going to be open source for research and commercial use as well. Meta has also highlighted that it should be deployed responsibly, and the company will frame policies and mechanisms to use it ethically and responsibly. Moreover, according to the same leaks, Meta has also partnered with Dell to offer Llama 2 on-premises for enterprise users for ensuring more control and security over personal data. It seems probable since OpenAI at the DevDay has announced GPT-4 Turbo and Google is also planning to launch Gemini. It is about time that Meta actually launched Llama 3, since it is already in training. Interestingly, Mark Zuckerberg, in his latest podcast with Lex Fridman in the metaverse, said that Meta might have to reconsider if it is going to open source the next iteration of Llama, which is Llama 3. “Right now, the priority is building that into a bunch of consumer products,” said Zuckerberg. But now, it seems like Meta wants to go the open source way again. In the podcast, Zuckerberg added that Meta trained Llama 2 and released it as an open source, but it is not a consumer product, but just an AI infrastructure. Though Zuckerberg is all about open sourcing AI and loves what the community has been doing with Llama 2, when it comes to Llama 3, he said that the debate with Fridman was very helpful for open sourcing Llama 2, and the same would be needed for Llama 3. “We would need a process to red team this, and make it safe. My hope is that we would be able to open source the next version when it is ready to, but we are not close to doing that this month. It’s a thing that we are still early in work now,” said Zuckerberg. This time, Llama 3 might even be better than GPT-4. Given all the debates about AI policies and the recent Biden order on regulating AI, a user on Reddit says, “All they need to do is make it 180B and most people will have no way to abuse it.”","excerpt":"The best part about it is that it is going to be open source for research and commercial use as well.","categories":["AI News"],"tags":["Meta AI"],"author_name":"Mohit Pandey","publish_date":"2023-11-09T22:57:35","publication_year":"2023","word_count":401,"keywords":["Go","Meta AI","OpenAI","AI","programming_languages:R","programming_languages:Go","GPT","llm_models:Gemini","llm_models:Llama","R","llm_models:GPT"],"extracted_tech_keywords":["AI","OpenAI","R","Go","GPT","llm_models:GPT","llm_models:Gemini","llm_models:Llama","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-to-launch-llama-3-early-next-year\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044439,"title":"Privacy Wars: Can W3C Play A Mediator?","content":"The World Wide Web Consortium (W3C) consists of  developers and engineers from all around the world collaborating to ensure websites work seamlessly. Collectively, the consortium writes the rules that make websites secure across browsers. W3C members deliberate on open GitHub forums and zoom meetings. Founded by the Web’s inventor Tim Berners-Lee in 1994, W3C has successfully overseen processes of issue raising, design, consensus building and testing, resulting in over 335 technical standards that make the Web go around. Success stories include, HTML5, XML, CSS and Web accessibility guidelines. But lately, this collaborative spirit has come under a lot of strain. The W3C has become a key battleground in the privacy war. The debate In January 2020, Google announced a plan to make its Chrome browser more private and secure by removing third party cookies. However, this step would negatively impact businesses and analytics firms that depend on harvesting user’s internet history–prompting them to join W3C to oppose new privacy standards by Google, Apple etc. Companies like Google, Microsoft, Apple, and Mozilla are turning to W3C to develop standards to protect users’ privacy and replace tracking techniques. On the other hand, using such methods for website optimisation and advertising is a major revenue stream for many companies, including Facebook. Google FLoC Cookie tracking has become more invasive, with advertisers using ‘fingerprinting’ to find a person’s identity bypassing Google’s anti-tracking measures. Federated Learning of Cohorts, or FLoC, is a privacy sandbox technology that allows advertisers to reach people and target their ads without exposing individual details or an individual’s browser history. The technology pools users into ‘cohorts’ or individuals with similar interests and serve targeted ads to these groups. The cohorts will be generated through an algorithm that will place a person in a different group each week, based on his\/her browsing history. Given Chrome’s popularity, the shift will have significant implications on the security and privacy of the users. Ad-tech perspective To prevent the web from having browser specific policies, W3C deliberates new proposals with its members and ensures web browser companies are on board with executing the proposal. However, that wasn’t the case when Google announced its privacy sandbox. The move led to backlash from both privacy advocates and ad-tech companies. The former worried that targeting groups of people based on their interests would still amount to discriminatory advertising. And the latter accused Google of trying to kill their business. Apple has introduced new privacy measures with its app tracking transparency (ATT) framework. The new OS gives users more control over their trackers. The“identifier for advertisers” (IDFA) used to be a default on Apple devices. But users have to give apps explicit permission to access it. A section of the W3C consortium is calling out big tech companies ‘power grab’ moves. While the ad-tech industry is petitioning to rein in the power of big tech companies’ impunity with respect to tracking, privacy advocates at W3C fear the growing demand of ad-tech companies might put a damper on the development of new privacy-protection tools. Meanwhile, Facebook is the only tech giant without its own browser or operating system, but collecting massive amounts of data for its ad business. More than 200 members of the W3C’s ‘Improving Web Advertising Business Group’ attend weekly conference calls and share code on GitHub to keep the communication channels open between the ad tech and browser communities. Temporary truce After the opposition filed a complaint against Google at UK’s Competition and Markets Authority, the search engine giant stalled plans to kill third party cookies by a year. Google has also cancelled the plan to use browsing history to create alternative identifiers once cookies were out. Chrome’s version 89 included a test run of FLoC on a small percentage of users in Australia, Brazil, Canada, India, Indonesia, Japan, Mexico, New Zealand, Philippines and the US.","excerpt":"The W3C has become a key battleground in the privacy war.","categories":["AI Features"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-07-24T16:00:00","publication_year":"2021","word_count":638,"keywords":["federated learning","Go","programming_languages:R","AI","ML","programming_languages:Go","Git","analytics","GitHub","R"],"extracted_tech_keywords":["AI","ML","analytics","federated learning","R","Go","Git","GitHub","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/privacy-wars-can-w3c-play-a-mediator\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10005204,"title":"Jigsaw Academy Launches Cyber Security Program With HackerU","content":"In a recent development, Jigsaw Academy, a Manipal Global investee company, launched a new Master Certificate in Cyber Security (Blue Team) program in association with HackerU, Israel’s premier cybersecurity training provider. Master Certificate in Cyber Security (Blue Team) is the first program in India that focuses on defensive cybersecurity technologies that will help learners detect an attack, collect forensic data, perform data analysis, and make modifications to avoid future attacks and mitigate threats. Many reports have suggested a significant increase in the number of cyber attacks and breaches across organisations, especially during pandemic. To combat this, the need for cybersecurity measures is more than ever before. Many organisations are looking for cybersecurity experts but are facing a scarcity in finding the right talent. The program is meticulously designed to train the candidates from the best faculty and experts from Jigsaw Academy and Israel. It focuses on integrating real-world projects and training exercises to provide the hands-on experience Ariel Cohen, the CEO, HackerU Global Education, said, “Defensive cybersecurity in one of the most secure and high-income professions in the cyber world. 85% of the job openings in the cybersecurity domain are for Cyber Defenders, who are also known as Blue Team professionals. This program will help learners build and protect the infrastructure of a company while also helping them comprehend how to attack it and protect the company from vulnerabilities and cyberattacks.” The program is for a duration of more than 520 hours and classes will be conducted as online live sessions. It will make a candidate equipped in concepts such as cyber\/network security. Virtualisation, managing network services such as DNS & DHCP, cloud services, AWS, Google cloud platform and more. Gaurav Vohra, CEO & Co-Founder of Jigsaw Academy, said, “Amidst the pandemic, cyber security experts have become the need of the hour for every organisation, in every sector Cyber security job opportunities constitute 15.6% of the emerging technologies open jobs pool. Our new Master Certificate in Cyber Security (Blue Team) program aims to help individuals in becoming proficient in defensive cyber security techniques and kick start their career in the cyber realm.” Jigsaw Academy recently launched a program on people analytics and digital HR in collaboration with IIM Indore to prepare the workforce with changes that COVID has brought in workplaces. Program starts: 14 September 2020 Know more about the program here.","excerpt":"In a recent development, Jigsaw Academy, a Manipal Global investee company, launched a new Master Certificate in Cyber Security (Blue Team) program in association with HackerU, Israel’s premier cybersecurity training provider. Master Certificate in Cyber Security (Blue Team) is the first program in India that focuses on defensive cybersecurity technologies that will help learners detect […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2020-08-19T17:40:12","publication_year":"2020","word_count":392,"keywords":["Go","AWS","AI","cloud_platforms:AWS","Git","Aim","analytics","cloud_platforms:Google Cloud","GAN","R"],"extracted_tech_keywords":["AI","analytics","Aim","AWS","R","Go","Git","GAN","cloud_platforms:AWS","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jigsaw-academy-launches-cyber-security-program-with-hackeru\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10170328,"title":"Open Source LLMs Pave the Way for Responsible AI in India","content":"Open-source large language models are emerging as powerful tools in India’s quest for responsible AI. By allowing developers to fine-tune models on locally relevant datasets, organisations are building solutions that reflect the country’s diversity. In a recent conversation with AIM, powered by Meta, Alpan Raval, chief AI\/ML scientist at Wadhwani AI, and Sourav Banerjee, CTO and co-founder of United We Care, explained how this approach is making AI both more ethical and more effective. “We are doing projects in healthcare, in agriculture, and in primary education that leverage LLMs, some of which are supported by Meta,” said Raval. He further added that open source models offer a lot of freedom in terms of fine-tuning them, adding extra layers on top of them, and then retraining from scratch. Alpan shared another example where they have developed an oral reading fluency assessment using AI, currently deployed in public schools across Gujarat, India. This initiative leveraged AI4Bharat’s open-source models. Raval stated that they collected student data from across the state and trained more advanced models by utilising both this student data and synthetic data generated through pseudo-labelling children’s voices with base models. He emphasised that this achievement would not have been feasible without the open-sourcing of the base models. Adding on to the conversation, Banerjee said that if any company is going for a vertical use case, the best approach would be to pick an open-source model and do the post-training on that. “We should focus on post-training on the existing pre-trained models, and work with the use case,” he said. Tackling Bias Alpan said that open source, by itself, magically removes bias. “It depends on the methodology, the kind of data the model was trained on, and so on,” he said. He explained that many open-source models are trained on datasets that differ significantly from the data observed in rural and underserved communities. “It’s almost imperative for us in order to prevent bias that we have to fine-tune those data sets.” Discussing hallucinations, Banerjee said that LLMs won’t stop hallucinating, and we have to live with that. However, he believes it is sensible to put weights and biases, training methodology, in the public domain. He explained that this transparency allows for public scrutiny and helps identify inherent errors. “Put it in the public domain for public scrutiny. Let people decide what they are getting into, rather than a closed, boxed approach.” He also offered a nuanced perspective on bias, suggesting that it’s not always inherently negative. He provided examples of common AI limitations, such as generating an image of an analogue clock at 6:25 or a left-handed person writing. Banerjee explained that these limitations stem from training data being biased towards certain representations. To improve model accuracy, he said it may be necessary to introduce a different kind of bias, which he calls positive bias. He gave the example of healthcare, where accuracy matters more than being completely neutral. In such cases, adding a positive bias can help make the system more accurate, even if it means making a trade-off. Security and AI Guardrails For organisations in the social sector, the security of Personally Identifiable Information (PII) remains a top concern. Alpan said, “We have a rule—more or less—that we don’t ingest PII into the organisation at all, except in certain cases where we have no choice.” Regarding ethical guardrails and governance, Alpan said that there’s no “one size fits all” solution. The ethical use of open-source models depends on their intended application. On the other hand, Banerjee said there is a need for an “inter-governmental initiative” for AI safety, similar to aviation safety, due to the decentralised nature of AI processing and training. He added that clear guidelines on “what is acceptable in a domain and what is not” are needed, particularly in human-machine interaction. Banerjee said that instead of looking at the West, India should be proud of the work that it is doing for responsible AI and lauded NASSCOM’s developer guidelines. He stated that the developer guideline is highly actionable and serves as a resource for both individuals and organisations to comprehend their responsibilities when using, building, or fine-tuning foundation models. Alpan said that India’s leadership in using AI for social good is supported by strong government collaboration. “India has been the number one country in the world to emphasise AI for social good—and it’s not just in letter but also in spirit,” he added. He further said that open source AI is being used to solve pressing challenges in fields ranging from healthcare and agriculture to education and climate. “Nandan Nilekani has said many times that India is going to be the use case capital of the world, and that applies to AI as well,” he concluded.","excerpt":"“We are doing projects in healthcare, in agriculture, and in primary education that leverage LLMs, some of which are supported by Meta.”","categories":["Global Tech"],"tags":["Meta"],"author_name":"Siddharth Jindal","publish_date":"2025-05-21T10:52:05","publication_year":"2025","word_count":789,"keywords":["Go","API","Meta","AI","ML","RAG","Aim","GAN","foundation models","AI safety","R"],"extracted_tech_keywords":["AI","ML","foundation models","Aim","RAG","R","Go","API","GAN","AI safety"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/open-source-llms-pave-the-way-for-responsible-ai-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":30762,"title":"Tackling Underfitting And Overfitting Problems In Data Science","content":"One of the major challenges in data science, especially concerning machine learning, is how well the models align themselves to the training data. Underfitting and overfitting are familiar terms while dealing with the problem mentioned above. For the uninitiated, in data science, overfitting simply means that the learning model is far too dependent on training data while underfitting means that the model has a poor relationship with the training data. Ideally, both of these should not exist in models, but they usually are hard to eliminate. Overcoming Overfitting ML experts and statisticians often have different techniques for bringing down overfitting in ML models. The popular ones stand out to be cross-validation and regularisation. These methods are proven to be effective in understanding the overfit data. Apart from these techniques, there are other ways to eliminate overfitting in models. Generalisation: For example, make sure the data leads to generalisation rather than just acting as training data. This can be done by feeding more data to the model. More data also means improved accuracy achieved by the model. However, this makes the model, computation and memory-intensive. Data Augmentation: As a result, another technique called data augmentation comes into the picture. Instead of giving loads of data, improvising and reworking on the existing data can go a long way in reducing overfitting. Example: Neural networks, which are mostly used in pattern recognition tasks, are prone to overfitting. The larger the network, the complex the functions it creates as a consequence. Hence, an optimum size for the right statistical fit is key. This can be done through a number of methods. The best among them would be retraining neural networks since it is comparably simple and does not involve tweaking much of the parameters. Generally, overfitting occurs in nonlinear ML models since there are many variables at play to decide the relationship of data in the model. This itself makes the model predict various factors. A better way to address this problem can be methods like k-cross validation. Here, the model is tested k-times for different subsets on the data and can be checked to see how it performs for new data. Any overfitting observed will eventually be diminished. Lately, ensemble methods such as Bayesian averaging, Boosting and Bagging have indirectly assisted in eliminating overfitting. How? Since ensemble methods deal with complex ML models, they take on the combined overfitting possibilities present in these models. Boosting and Bagging are the two most used methods than Bayesian averaging. Eliminating Underfitting Although underfitting is comparatively observed lesser in ML models, it should not be overlooked. To begin with, the general norm here is lack of sense between the data and model. What this means is either the model is way too simple to establish a stable learning pattern or performs very poorly with the training data. Experts suggest that this problem can be alleviated by simply using more (good!) data for the project. In addition, the following ways can also be used to tackle underfitting. Increase the size or number of parameters in the ML model. Increase the complexity or type of the model. Increasing the training time until cost function in ML is minimised. Example: Converting a linear model’s data into non-linear data. In this case, the transformation of the model leads to it being more unpredictable with respect to any new as well as training data. Comment Both overfitting and underfitting should be reduced at the best. As ML expert Jason Brownlee perfectly puts it, a statistically “good fit” is what matters when it comes to choosing an ML model. This can only be done with repeated testing of the model with different data and see where it falls along the lines of overfitting and underfitting. Furthermore, before starting with an ML model to solve a problem, it is also suggested to take a hard look into the data too!. After all, there might also be the possibility of conflict with the type of data used in the model.","excerpt":"One of the major challenges in data science, especially concerning machine learning, is how well the models align themselves to the training data. Underfitting and overfitting are familiar terms while dealing with the problem mentioned above. For the uninitiated, in data science, overfitting simply means that the learning model is far too dependent on training […]","categories":["AI Features"],"tags":["overfitting"],"author_name":"Abhishek Sharma","publish_date":"2018-11-27T12:58:33","publication_year":"2018","word_count":665,"keywords":["data science","Go","data augmentation","machine learning","programming_languages:R","AI","neural network","R","ML","RAG","overfitting"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","data science","RAG","R","Go","data augmentation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/tackling-underfitting-and-overfitting-problems-in-data-science\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":37056,"title":"How This Banking Guru From The West Is Using Analytics To Help Indian SMEs","content":"This is the age where banks chase customers, offering them credit cards, loans and mortgages. But that’s on the personal side. There’s one sector in business that is yet to gain advantage from the rapid growth of the BFSI sector. This week we spoke to Lucas Bianchi, co-founder and head of operations & finance at Namaste Credit. From his journey on the Wall Street to starting his own company here in India, Bianchi talked to Analytics India Magazine about how they are trying to help the small and medium enterprises secure funding. Analytics India Magazine: Tell us a little about your background in the Finance sector and how you started Namaste Credit. Lucas Bianchi: Before Namaste Credit, I worked at Copal Amba, a Moody’s subsidiary. I was with them for a number of years, and during that time I got some exposure in India. I understood what the Indian market was about and got a fresh perspective on it. That’s when Gaurav Anand, who is now our co-founder and the head of sales and strategy, and I decided to create a credit-focused technology company based in India and focused on India. AIM: Why did you decide to focus on the SME sector? LB: We basically had a great understanding of credit and that’s why we decided to focus on the small and medium-sized enterprises segment. You see, the SME segment requires a lot of credit understanding. It’s many data points you have to look at. And we came from a credit background having been with Moody’s, so we wanted to leverage that as well. The other thing is that there’s a lot of capital in India, there are a lot of banks, a lot of NBFC. The problem lies with the facilitation agency. It’s the matching and the process of doing that efficiently. Allocating capital where it can be best used, where it can be priced correctly and so, we built a platform that created a marketplace to do that. AIM: What banks and financial institutes are you working with right now in India? LB:  We are working with a whole a bunch of different players from different parts of the spectrum in terms in the BFSI sector. The banks that we are working with are the ones who have to do a lot of heavy lifting, say, around mortgages or around business loans. As of now, we have tied up with about 50 banks in India. AIM: Which shoe does Namaste Credit fill in with respect to SMEs and facilitating their interaction with these big banks? How are you using analytics and machine learning here? LB: One of the key elements that I think is somewhat unique and problematic about India is extensive reliance on the paper trail. It’s not just because the banks require it, it’s also because that’s the kind of standard and norm that’s in the market. So, everything around the whole loan process is focused around kind of this trading of paper around him or something. So we’ve built licensed systems for these banks which enable them to feed in a document that’s been scanned, not just generated from a system. We use OCR to grab that data, we use machine learning and analytics to clean it, put it into a format that the bank would want, and then push it to them in their format. We pull the data in a format-agnostic manner. This enables us to take data from many different document types, many different format types. This intelligence layer helps to identify what it is that we’ve grabbed and then categorise it accordingly. This is where we get to machine learning is in that process of categorisation of correction. AIM: How has Namaste Credit grown over the years? LB: We started our operations in early 2015, so we are about four years old. We have about 250 people today. Now we have about 15 offices across most of the major cities. AIM: Have you ever faced trouble in hiring people for tech roles at Namaste Credit? LB: I think hiring good people is the biggest challenge that any startup faces. There are a lot of people out there, but finding the right match is a challenge. What I’ve seen in India, is that a lot of times you’ll go onto a site or something like that and people will say stuff on their resume and then you bring him in to test those things and they won’t work. Like they just don’t know those things. They’ll put things on the resume that they don’t even know. Another thing is that many of the techies in Bangalore are service oriented people. So, they come to a workplace with a sort of baggage that they carry from their previous (usually big) employers.  We kind of try and stay away from that kind of mindset because we want people to come on board and use their own mind to try and think about what should be done and how to solve it. AIM: What, what is the next big thing that you’re looking for that you said, apart from the leap this year for the next round of funding? LB: In terms of the overall company growth, I think we’re looking to expand our channel partner base a lot. So, today we have about 4,000 channel partners on board across India. We want to expand that to, let’s say 10,000 to 15,000 over the next year. So, it’s a fairly significant expansion.","excerpt":"This is the age where banks chase customers, offering them credit cards, loans and mortgages. But that’s on the personal side. There’s one sector in business that is yet to gain advantage from the rapid growth of the BFSI sector. This week we spoke to Lucas Bianchi, co-founder and head of operations & finance at […]","categories":["AI Features"],"tags":["Banking","BFSI","Data Analytics","Interviews and Discussions","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2019-03-29T09:44:46","publication_year":"2019","word_count":917,"keywords":["Go","API","funding","machine learning","startup","BFSI","AI","Machine Learning","RAG","Aim","Banking","analytics","Data Analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","R","Go","API","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-banking-guru-from-the-west-is-using-analytics-to-help-indian-smes\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10023027,"title":"What To Make Of Intel’s Foray Into Foundry Business","content":"Pat Gelsinger, who took on the CEO’s mantle at Intel last January, made a few important announcements at its global event– ‘Intel Unleashed: Engineering the Future. Gelsinger shared his vision for the future of Intel’s integrated device manufacturing (IDM) model. Talking about IDM 2.0, he revealed three significant plans: The work on 7nm development is progressing well due to the use of extreme ultraviolet lithography (EUV). Intel now plans to tape in the compute tile for its first 7nm client (also called Meteor Lake) in the second quarter of this year.Intel will expand its engagement with third-party foundries to support modular tiles manufacturing on advanced process technologies for Intel’s core computing offering for client and data centre segments from 2023.Intel will establish a new business unit, the Intel Foundry Services (IFS). This standalone unit, headed by Dr Randhir Thakur, will offer a combination of leading-edge process technology and packaging. The company is aiming to become a primary provider of US and Europe based foundry capacity to meet the global demand. Stepping Into Foundry Business While announcing IFS, Gelsinger said Intel is primed to meet the growing demand for a sustainable and secure supply of semiconductors. IFS will be a fully vertical, standalone foundry business. With this, Intel aims to enter the fast-growing foundry business, which is expected to grow to a $100 billion addressable market by 2025. “We will be differentiated from other foundry offerings with a combination of leading-edge packaging and process technology, committed capacity in the US and Europe — available for customers globally — and a world-class IP portfolio that customers can choose from, including x86 cores, graphics, media, display, AI, interconnect, fabric and other critical foundational IP along with Arm and RISC-V ecosystem IPs,” said Gelsinger. Intel has committed $20 billion to build two new fabs in Arizona, with possible foundries in the US, Europe, and elsewhere. Notably, Intel recently announced a partnership with IBM to drive innovation in semiconductor technology. With this collaboration, the two companies aim to advance next-generation logic and packaging technology. Why Is This Move Significant? There have been multiple reports of Intel’s decline as the market leader. The global chip shortage exacerbated the situation. In 2019, Intel apologised to personal computer makers for failing to supply chips, which had significantly slowed the PC industry. Intel, which holds about 80 percent of the market for PC chips (2019 figure), was then forced to outsource production. Intel is at a crossroads. On the one hand, it faces increased competition from rivals such as AMD, on the other, the company has seen significant management restructuring and production delays, losing leverage to companies such as Taiwan-based TSMC. Given the context, the newly appointed CEO Gelsinger’s announcement seems to be a step in the right direction. Intel’s intention to support semiconductor manufacturing, especially in the context of the US and Europe, is telling. “Having 80% of all supply in Asia simply isn’t a palatable manner for the world to have its view of the most critical technology,” Gelsinger said. In a blog, Jeff Rittener, Chief Government Affairs Officer for Intel, wrote the new facilities at Arizona will support and bolster the US’s position as the technology leader and strengthen its national security and supply chains. The new fab facilities would create an estimated ‘3,000 permanent high-tech, high-wage jobs, over 3,000 construction jobs, and 15,000 local long-term jobs in Arizona’. The blog also pointed out the US’s decline in semiconductor manufacturing. The company spokesperson said IFS and the recently passed Creating Helpful Incentives to Produce Semiconductors (CHIPS) for America Act as part of the FY2021 National Defense Authorization Act (CHIPS) NDAA) would help fill gaps. This year, the European Union outlined its ambition of having up to 20 percent of the world’s chips manufactured locally. Intel, in its communique, said it plans to increase its investment in the region and support this goal. “We have a shared ambition with the EU to deliver state-of-the-art semiconductor technology to Europe and create a more geographically balanced manufacturing capacity,” an official statement read. Use Of Third Party Foundry Capacity Intel will strengthen its relationship with third-party foundries for application in the communications, connectivity and graphics industries. Gelsinger said the increased engagement would give Intel the much-needed flexibility and scalability required to gain a competitive edge.","excerpt":"Pat Gelsinger, who took on the CEO’s mantle at Intel last January, made a few important announcements at its global event– ‘Intel Unleashed: Engineering the Future. Gelsinger shared his vision for the future of Intel’s integrated device manufacturing (IDM) model. Talking about IDM 2.0, he revealed three significant plans: The work on 7nm development is […]","categories":["Global Tech"],"tags":["Intel","Pat Gelsinger"],"author_name":"Shraddha Goled","publish_date":"2021-03-30T12:00:00","publication_year":"2021","word_count":715,"keywords":["Go","programming_languages:R","AI","Pat Gelsinger","innovation","Scala","RAG","Aim","ViT","programming_languages:Scala","R","Intel"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Scala","ViT","innovation","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/what-to-make-of-intels-foray-into-foundry-business\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10123581,"title":"Larry Ellison Sees a Surge in Net Worth, Thanks to Google Cloud, OpenAI and Others","content":"Larry Ellison is finally smiling. It took Google Cloud and OpenAI nearly nine months to realise the importance of Oracle Cloud Services(OCI). AWS, hopefully, will follow suit. The outcome: Oracle chief Ellison saw almost $19 billion in wealth as the company he founded in 1977 forecasted double-digit revenue growth for the fiscal year. Moreover, following these announcements, the software company’s shares skyrocketed by 13% in extended trading on Wednesday. Oracle Cloud Services reported a revenue of $10.2 billion in Q4 2024. Meanwhile, Microsoft’s Intelligent Cloud posted $26.7 billion in sales for the recent quarter, AWS reached $25 billion, and Google Cloud reported $9.6 billion. OpenAI will now run its workloads on OCI, extending the Microsoft Azure AI platform to Oracle’s cloud services. “Like many others, OpenAI chose OCI because it is the world’s fastest and most cost-effective AI infrastructure,” said Oracle chief Safra Catz in a recent earnings call. She added that Oracle has signed over 30 AI contracts totalling over $12 billion this quarter and nearly $17 billion this year. Meanwhile, Elon Musk’s xAI is discussing with Oracle executives the possibility of spending $10 billion over the next few years renting cloud servers. OpenAI, in its recent post on X, clarified: “The partnership with OCI enables OpenAI to use the Azure AI platform on OCI infrastructure for inference and other needs.” However, all pre-training of frontier models will continue to happen on supercomputers built in partnership with Microsoft. In the backdrop of the Data + AI Summit, Databricks lauded Oracle. “We’ve seen Oracle become much more relevant in the cloud space in this AI era. We actually have partnerships with them around GPUs already, so many of the models we’ve trained on Mosaic AI, custom models that we trained, have been trained on infrastructure provided by Oracle,” said the Databricks chief Ali Ghodsi, hinting at a plausible partnership in the coming months, pointing at customer requirements. “Congrats to Ellison. He needs it,” said Ghodsi. The King of Multi-Cloud Oracle announced Oracle Database@Azure last year, which delivers Oracle database services running on OCI inside Azure data centres and gives customers more flexibility in where they run their workloads. Ellison said that customers have already been using multi-cloud products and services, and there are even stronger reasons to believe they should be interoperable and interconnected more than ever. “We’re doing the same thing with Google. We would love to do the same thing with AWS. We think we should be interconnected to everybody, and that’s what we’re attempting to do in our multi-cloud strategy,” said Ellison. Expanding on its multi-cloud strategy, Oracle recently partnered with Google Cloud, giving customers the choice to combine OCI and Google Cloud to help accelerate their application migrations and modernisation. “OCI and Google Cloud network interconnect is available immediately in 10 regions, and we will be live with Oracle Database at Google Cloud in September, where customers can get direct access to Oracle Database services running on OCI deployed in Google Cloud data centres,” said Ellison. Pradeep Vincent, chief technical architect of Oracle, in an exclusive interview, told AIM that OCI is pretty different from the competitors out there. “Our goal is to make it easy for customers to use multiple clouds, period,” he said, explaining that a key part of this is their ‘distributed cloud strategy’, putting the cloud where customers want it. On similar lines, Ellison also said, “We believe in giving customers a choice, and they want it. Customers are using multiple clouds, including infrastructure clouds and applications like Salesforce and Workday. Therefore, we think it’s very important for all these clouds to become interconnected”. In the coming months, Ellison mentioned that Oracle looks to get rid of these fees (or egress cost) for moving data from cloud to cloud, and all the clouds will be interconnected and customers can pick their favorite service from their favorite cloud and mix and match whatever they want to use and do it easily and seamlessly. Further, he said that OCI’s RDMA network moves data much faster. “And when you charge by the minute, faster also means less expensive,” he said, adding that OCI trains large language models several times faster and at a fraction of the cost of other clouds.","excerpt":"Oracle chief Ellison saw almost $19 billion in wealth as the company he founded in 1977 forecasted double-digit revenue growth for the fiscal year.","categories":["Global Tech"],"tags":["Oracle"],"author_name":"Siddharth Jindal","publish_date":"2024-06-13T17:30:53","publication_year":"2024","word_count":705,"keywords":["Go","OpenAI","AI","AWS","Azure","ML","R","Oracle","Aim","xAI","Databricks"],"extracted_tech_keywords":["AI","ML","OpenAI","xAI","Aim","AWS","Azure","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/larry-ellison-sees-a-surge-in-net-worth-thanks-to-google-cloud-openai-and-others\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":22569,"title":"How Naftali Tishby’s Information Bottleneck Theory Can Break Open The Black Box Of Deep Learning","content":"Deep learning has been making tremendous progress in the fields of computer vision , natural language processing and other fields of machine learning. But how these deep learning models work, at the scale they do has been an open question for quite a while. This mystery surrounding deep learning has made many researchers to focus specifically on understanding large deep learning models. Computational learning theory gives a formal framework to study the predictive powers and computational powers of machine learning models. It is mostly associated with knowing about the efficiency of data usage (sample complexity) and computation usage (time complexity). Recent research led by Dr. Naftali Tishby has made exciting connections between the field of computational learning theory and information theory. When thinking about information theory, most machine learning practitioners have a very narrow perspective of the same. Information theory is thought to be a field mostly related to communication and compression. But recent developments coupled with decade-old research has suggested that information theory has massive implications in deep theory. Generalisation Bounds For Deep Learning Computational learning theory gives generalisation bounds to estimate the power of machine learning models. But as it turns out these kinds of specified limitations don’t really work for deep learning. The primary reason for this is due to the exponentially large number of parameters present in the neural networks. Since the inspiration behind neural networks is the human brain, it becomes harder to understand how today’s neural networks actually work. There is an ongoing struggle between researchers who are trying to understand how neural networks achieve such generalisations. Recently Naftali Tishby, a computer scientist and neuroscientist from the Hebrew University of Jerusalem, created some excitement among artificial intelligence researchers when he offered a theory of how we can use information theory to explain advances in deep learning. According to Tishby, deep neural networks follow a procedure known as information bottleneck, which he and two other collaborators had worked on in 1999. Now recently Tishby is back with more experiments that validate his claims. His theory says that deep learning procedures compress information during training and throw away useless information — much like Sir Arthur Conan Doyle’s famous detective Sherlock Holmes. The Connection Between Information Theory And Deep Learning Over the years, deep learning has experienced incredible success. But at the same time, there is a constant criticism on the lack of theoretical explanation about deep networks. In a standard statistical paradigm, the main tactic is to have a large number of candidates and restrict or remove complex solutions. In deep learning, stochastic gradient descent already works as powerful regulariser. But how does it actually work? It is still mathematically unclear. Fig.1 The structure of a deep neural network, which consists of the target label Y, input layer X, hidden layers h1,…,hm and the final prediction Ŷ . (Image source: Tishby and Zaslavsky, 2015) Information theory helps us to peek into the back box of deep learning. The training data contains sampled observations from the joint distribution of X and Y. The input variable X and weights of hidden layers are all high-dimensional random variables. The ground truth target Y and the predicted value Ŷ are random variables of smaller dimensions in the classification settings. If we label the hidden layers of a DNN as h1, h2,…, hm as showcased in the figure above, we can view each layer as one state of a Markov Chain: hi→hi+1 A Deep Neural Network is designed to learn how to describe X to predict Y and eventually, to compress X to only hold the information related to Y. Tishby describes this process as “successive refinement of relevant information”. The information plane theorem put forward, frames each layer by its encoder and decoder information. The encoder is a representation of the input data X, while the decoder translates the information in the current layer to the target output Y. Deep Learning As An Information Bottleneck Procedure In 2015, Tishby and his students presented an hypothesis that deep learning is an information bottleneck procedure. The procedure compresses input data while retaining as much information as possible. Newer research now digs deeper into this hypothesis. As a part of this, the researchers used a neural network to recognise dogs. And they then investigated what how the network behaves with 3,000 sample input data points. They then observed how much information each layer of the neural network retained and how much of it is related to the output label. The researchers observed that the networks converged to the information bottleneck theoretical bound. The information bottleneck theoretical bound is a very rare deep learning theoretical limit described in Tishby’s original paper. This suggests the limit or the best case of the neural network extracting all the relevant information. The idea is that neural networks compress the input data as much as possible without losing the generalisation ability. This discovery also puts a spotlight on the peculiarity of the training phases of large neural networks. The researchers discovered that deep learning happens in two phases — a short  fitting phase, where the network learns to predict the labels, and a longer compression phase where the network works to better its generalisation capabilities. Learning Is Forgetting Deep learning was imagined by early AI pioneers as a way to replicate the workings of the human brain. But since the early days of deep learning, the research has swayed away from the biological plausibility of the models. More and more emphasis has been laid to how the models work on real-world tasks. But it still remains to be seen how much of the studies in neuroscience can be translated into advances in the deep learning models. But Tishby strongly believes that his ideas will be useful both in, neuroscience and the machine learning communities. He proudly says, “The most important part of learning is actually forgetting.” This is quite appropriate since his ideas suggest that leaving some details behind can help us build models which learn better.","excerpt":"Deep learning has been making tremendous progress in the fields of computer vision , natural language processing and other fields of machine learning. But how these deep learning models work, at the scale they do has been an open question for quite a while. This mystery surrounding deep learning has made many researchers to focus […]","categories":["IT Services"],"tags":["Deep Learning","Machine Learning"],"author_name":"Abhijeet Katte","publish_date":"2018-03-13T09:27:12","publication_year":"2018","word_count":999,"keywords":["Go","artificial intelligence","machine learning","TPU","AI","neural network","Machine Learning","computer vision","Aim","deep learning","Deep Learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","computer vision","Aim","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-tishbys-information-bottleneck-can-break-open-the-black-box-of-deep-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10066047,"title":"80 per cent of executives agree companies will lose competitive advantage if data is not leveraged fully: Vijay Yellapragada, executive director at EY GDS data analytics","content":"In the rapidly changing digital marketplace, newer companies that were formed within the past two decades held an obvious advantage over companies with legacy IT systems. New companies had digital natives built on a digital core that relied on distributed technologies. This has given them the ability to capture, track and analyse real-time customer interactions. As a result, companies have gotten smarter with every transaction due to increased efficiency and decreased costs, making them better equipped to respond at a faster pace to emerging threats and trends. At the first edition of the Data Engineering Summit 2022 organised by Analytics India Magazine, Vijay Yellapragada, the executive director at EY’s Global Delivery Services (GDS) Data Analytics, spoke about how data fabric now has come to play an important role in making organisations more resilient and achieve sustainable growth. Source: Gartner Need to leverage data effectively More than obtaining data, the focus has now shifted to how well companies are able to leverage the data. Yellapragada mentioned that according to Gartner, 80 per cent of executives agreed that companies will lose the competitive advantage if their data is not fully leveraged. There was an urgent need for solutions that could help organisations overcome challenges that were involved with harnessing the power of data. Data fabric is a paradigm shift in how companies stand to leverage data. The term was coined by Noel Yuhanna, the vice president and principal analyst at Forrester Research, in the mid-2000s. “It is not a new technology but a smart and intelligent unified service that can help companies accelerate their data transformation lifecycle,” Yellapragada explained. In the past, there were several issues like datasets being disconnected; there was no single definition of data or even platforms not having portability. On the need for data fabric, Yellapragada said, “Once you add a second database, it becomes harder to have data integrity and integration on the platform. Marketing campaigns needed to use real-time data.” By accessing data through one common platform, the data fabric solution provides a single view of data that was previously housed in separate data lakes and warehouses. This has led to an increase in the sharing of real-time data owing to using data fabric. Source: Microsoft Ignite, EY leaders and US Data and AI leaders discuss how to optimise data using EY Microsoft Data Fabric Applications of data fabric In the recent past, the usage of the architectural concept of data fabric has risen due to the public adoption of cloud. Yellapragada listed the several benefits of using a data fabric, which included data security, integration, governance, access to real-time data, lower costs and quicker access. There were several use-cases of a typical data fabric, such as Environmental and Social Governance (ESG) and sustainable reporting, infrastructure complexity and cost, and so on. “With the creation of a semantic layer, companies can apply AI and other data analysis tools to understand new trends and revamp go-to-market strategies in days instead of weeks,” Yellapragada stated. Adoption of data fabric helps developers avoid building multiple algorithms and build frameworks that are more flexible and open source. It helps businesses identify between internal and external data and also helps democratise data so that all departments within the company have equal access to real-time data. Source: EY, EY Data fabric Yellapragada concluded the talk with a quick description of a couple of EY case studies to measure the positive impact that data fabric had. In the first instance, a global pharma major was able to deliver clinical trials to the market 60 per cent faster with a 50 per cent decrease in operating cost. The project also saw a considerable rise in data-sharing. A similar result was observed in a second instance of ESG sustainable activity reporting. Yellapragada underlined how EY’s data fabric had unlocked the value for several companies. He said, “Companies in the middle of an IT transformation can often reduce at one-seventh the cost per transaction, enabling the implementation of a plan to progressively decommission the legacy systems. The data fabric solution also helps companies without a digital core to launch new ventures that turn them into digital natives at the outset, setting up operations that are significantly more efficient.” REGISTER HERE TO ACCESS THE CONTENT","excerpt":"In the first instance, a global pharma major was able to deliver clinical trials to the market 60 per cent faster with a 50 per cent decrease in operating cost.","categories":["IT Services"],"tags":["Data Fabric"],"author_name":"Poulomi Chatterjee","publish_date":"2022-05-02T13:13:41","publication_year":"2022","word_count":708,"keywords":["Go","API","AI","Git","RAG","data engineering","analytics","Data Fabric","GAN","R","data lake"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Git","API","data engineering","data lake","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/80-per-cent-of-executives-agree-companies-will-lose-competitive-advantage-if-data-is-not-leveraged-fully-vijay-yellapragada-executive-director-at-ey-gds-data-analytics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136239,"title":"Why Google Will Make a Better Model Than OpenAI’s o1","content":"While OpenAI basks in the glory of the o1 launch, Google is quietly working on a model that could surpass it. Since last December, when Google launched Gemini, the two companies have been locked in a competitive battle to outperform each other. AI insider Jimmy Apples recently posted a video on X, suggesting that Google is soon releasing something better, featuring the message, ‘Patience, they are coming.’ Even Logan Kilpatrick, Google’s lead product manager for AI Studio and the Gemini API, shared a post which read ‘Gemini Mode,’ hinting at new Gemini releases. Kilpatrick told AIM that Google plans to release Gemini 2, which will feature better reasoning quality and a longer context window—potentially up to billions or trillions of tokens. As per Kilpatrick, the model will be fully multimodal, with the capability to understand large videos as well. OpenAI thinks it has o1 In a recent interview OpenAI chief Sam Altman, accused other research organisations of copying OpenAI’s methods. “We try to have a very focused research program. I think one of the mistakes that other research programs make is they don’t have enough conviction and concentration. It’s very easy to copy something once it works,” he said. However, this might not entirely be true. After Transformer, which was first published by Google in the ‘Attention Is All You Need’ paper back in 2017, OpenAI incorporated one of Google DeepMind’s most popular reinforcement learning school of thoughts, pushing Google to accelerate its timeline for releasing their advanced models. Recently, Apples shared a document on X, dated last year, revealing that Google is planning to integrate the ‘PLANNING’ piece in the LLM. Moreover in an old Wired article, Google’s Demis Hassabis also said that his team will combine the technologies used in AlphaGo, aiming to give the system new capabilities such as planning and ability to solve new problems. Interestingly, when OpenAI o1 was launched, Subbarao Kambhampati, professor at Arizona State University, speculated that o1 incorporates RL over Private CoT methodologies. Chain of Thought or ‘CoT’ refers to generating step-by-step reasoning or thought processes, which can enhance the model’s ability to tackle complex tasks. He drew an analogy to Google DeepMind’s AlphaGo, a program developed for playing the game of Go, suggesting that o1 could operate in a similar way by defining ‘moves’ for problem-solving. Moreover, Kambhampati suggested that the RL task involves generating and selecting a CoT based on the original prompt, evaluating success or failure by whether the output aligns with expected answers from the training data. In AlphaGo, success was determined by game outcomes. For o1, success could be measured by whether the model’s expanded prompts lead to correct answers based on the training data. His claims might be valid, as OpenAI stated in its blog post that, “Through reinforcement learning, o1 learns to hone its chain of thought and refine the strategies it uses. It learns to recognize and correct its mistakes and to break down tricky steps into simpler ones.” Google earlier this year published a paper titled, ‘Chain of Thought Empowers Transformers to Solve Inherently Serial Problems’ which says that by increasing the length of CoT can drastically make transformers more expressive. The researchers define a new class of problems that can be solved by transformer models with specific limitations, like how many steps they can take and the size of the data they can handle. Not only that, Google recently published another paper titled ‘Training Language Models to Self-Correct via Reinforcement Learning’. Google Deepmind has developed a multi-turn online reinforcement learning approach to improve the capabilities of an LLM to self-correct. SFT is shown to be ineffective at learning self-correction and suffers from a distribution mismatch between training data and model responses. On the other hand, many are claiming that OpenAI’s o1 could be considered the first successful commercial launch of a System 2 LLM. A System 2 LLM is a type of language model intended to mimic more deliberate and analytical thinking, paralleling Daniel Kahneman’s concept of “System 2” thinking. In this framework, System 1 is characterised by fast, automatic, and intuitive responses, while System 2 involves slower, more methodical reasoning that necessitates conscious effort. “For those raving about GPT-4 o1, Google has been working on extending the chain of thought (i.e., System 2) since Gemini was released. That is why one should pay attention to specialised open-source Gemma implementations. System 2 thinkers are specialists, not generalists,” posted Carlos E. Perez, co-founder of Intuit Machine, referring to Google Gemini’s ‘Uncertainty Routed Chain of Thought.’ Google DeepMind is the Brain Behind all the AI Innovations Kilpatrick told AIM that the Google Gemini team and Google DeepMind work very closely together. He further revealed that the Google DeepMind team ultimately wants to ensure that the technology reaches both developers and the wider world. “They care a lot about making sure that the product teams building on top of the models are kept in the loop with what’s coming.” Google DeepMind’s recent models, AlphaProof and AlphaGeometry 2, won a silver medal at this year’s International Mathematical Olympiad (IMO). Meanwhile, OpenAI o1 scored 83% in a qualifying exam for the International Mathematics Olympiad, compared to GPT-4o’s 13% Alpha Geometry2, a neuro-symbolic hybrid system built on Gemini, is trained from scratch using an order of magnitude more synthetic data than its predecessor. With further improvements to Google DeepMind’s RL techniques and their integration with Chain of Thought in Gemini, Google could create a model that outperforms OpenAI’s o1. Not to forget, Google recently announced that YouTube is set to roll out advanced generative AI tools for creators in the coming months, enabling them to generate video content using the AI models Veo and Imagen 3 through a feature called Dream Screen, leaving Sora behind.","excerpt":"‘Patience, they are coming.’","categories":["Global Tech"],"tags":["Editors Picks","Google","OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-09-21T07:46:38","publication_year":"2024","word_count":957,"keywords":["Go","TPU","OpenAI","AI","GPT-4o","Transformers","Editors Picks","Aim","generative AI","Google","chain of thought","R"],"extracted_tech_keywords":["AI","generative AI","GPT-4o","OpenAI","Aim","Transformers","chain of thought","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-google-will-make-a-better-model-than-openais-o1\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10081087,"title":"Knowledge Graphs – The Next Big Thing In Analytics Industry","content":"One of the earliest references to knowledge graphs can be traced back to the 1980s when the Universities of Groningen and Twente began a project of that name. The project focused on semantic networks with edges to facilitate algebras on the graph. Today, a knowledge graph is a model that uses a graph-structured data model to integrate data. To know more, AIM interviewed Khyati Sahu, a senior engineer at Fractal. Working as a data architect, Sahu describes her role to be techno-functional in nature. She handles a team of engineers and manages projects from a technical point of view. “My responsibilities include gathering requirements and interacting with the customers to understand their needs,” Sahu told AIM. After her postgraduate studies at IIT Roorkee, Sahu has been working in the field for over eight years now, two of which have been at Fractal. A big part of her work involves employing knowledge graphs. AIM: What are knowledge graphs Khyati Sahu: Knowledge graphs comprise entities that can form a node, and the relationship between them is traced. Their goal is to provide meaning to the data and remove any semantic ambiguity and help in moving toward data analytics and intelligence. The primary use of knowledge graphs is found in search and recommendation systems. For instance, when you search for a person on Google, you will get a summary panel – containing information about their birth date, height, weight, and family members, among others. All this information is being fetched internally from knowledge graphs. At Fractal, I am working for a client who is a Fortune 500 client in the CPG domain. We have been building a recommender system for them. For example, we use a knowledge graph-based system to recommend a store for certain products based on its geographical location. And depending on whether there is a school or a hospital nearby, the sales and revenue of the store will increase. AIM: Are knowledge graphs restricted to academia? Khyati Sahu: No, in fact, knowledge graphs are being used for a wide range of applications like space, journalism, biomedicine, entertainment, network security, and pharmaceuticals. In the financial industry vertical, knowledge graphs are used for analytics, tax calculations, and financial reporting. Machine learning algorithms have already helped financial institutions monitor fraud and the risk behind it more automatedly. Knowledge graphs allow the graphical representation of fraud scenarios and help in better decision-making. Of course, there are limitations, like maintaining them, which may prompt people to look for alternatives. However, there has been constant innovation to overcome these challenges. Knowledge graphs will likely be the next big thing in the AI and analytics industry, and more people will adopt them. AIM: How to construct a knowledge graph? Khyati Sahu: In order to construct a knowledge graph, we combine both structured and unstructured data and extract the entities and relationships between them right. These form the schema of the model. First, you need to determine the use cases for your knowledge graph. Once you’ve decided that, there are a few things to keep in mind throughout the build: 1) All knowledge graphs start with data 2) Building them will be iterative 3) Always build it through the lens of your use case Next, take a closer look at all the information available from a field of knowledge. Identify all the categories, types, things, and objects important to the field, and outline the necessary data needed. Once you have identified your data, you need to organise it and map relationships with ontologies. The “graph” in the knowledge graph refers to organising data and highlighting relationships between data points. These relationships are key to keeping knowledge graphs nimble. Once you’ve identified and organised your data in the context of your original use case and business questions, it’s time to focus on the relationships between all this data. Ontology management and knowledge graph development are iterative processes and will evolve with time. However, as I mentioned before, maintaining them is a problem area. In addition, when the graph grows and new data is added, integration and maintaining correctness across the knowledge graph becomes a concern. AIM: Are knowledge graphs the next big thing in AI? Khyati Sahu: Yes, absolutely. Knowledge graph allows AI systems to deal with complex, interrelated data. It stores information as a network of data points connected by different types of relations. Knowledge graphs power internet search, recommender systems, and chatbots. Artificial intelligence has become invaluable in storing and organising large amounts of data using knowledge graphs. Including knowledge, graphs can improve the accuracy of the outcomes and augment the potential of machine-learning approaches.","excerpt":"AIM interviewed Khyati Sahu, a senior engineer at Fractal. Working as a data architect, Sahu describes her role to be techno-functional in nature.","categories":["AI Features"],"tags":["Interviews and Discussions","knowledge","Knowledge graph","Knowledge graphs"],"author_name":"Shraddha Goled","publish_date":"2022-11-30T10:00:00","publication_year":"2022","word_count":768,"keywords":["knowledge graphs","Go","Knowledge graph","artificial intelligence","machine learning","knowledge","Knowledge graphs","AI","chatbots","recommendation systems","Aim","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","chatbots","recommendation systems","knowledge graphs","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/knowledge-graphs-the-next-big-thing-in-analytics-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":65510,"title":"The Impact Of Data Scientists Returning To India","content":"Companies in the US have been on the layoff spree during the pandemic. Among the prominent ones, we have witnessed thousands of layoffs in companies such as Lyft, Airbnb and Uber just in the last week or so. According to reports, there are hundreds of thousands of Indian technologists currently stranded in America amid mass layoffs. While their grace periods period has been extended by the United States Citizenship and Immigration Services, without a job those professionals have to leave the country. More than 30 million jobs in the US have been lost since April due to COVID19 pandemic. Plus, the US government announced that the US government is going to temporarily suspend the H-1B visa program for protecting domestic jobs amid record-high unemployment in the US. There was an immense pressure on the US government to suspend foreign worker visas for at least a year or at until unemployment rates return to normal levels. A Majority of Indian workers come under the purview of H-1B visa. This means that the suspension of the H-1B visa means hundreds of thousands of Indian workers in the US may remain without jobs, the majority of whom may be soon heading back to India. And therefore it is expected that thousands of techies will be making their way back to India including data scientists analytics professionals as well. So what impact would the analytics field have on? As unfortunate as it is for many data scientists and analytics professionals to relinquish their high-paying US jobs, they can still continue their career path in India as well, where there is strong demand for data science jobs. Relieving Talent Scarcity In India If you look at the Indian tech industry, it has been comparatively more immune to data science layoffs, and companies have retained quality talent. It is expected in a case where Indian data scientists working in America make a return, will boost the availability of quality talent coming back to India in data science and analytics. This may also relieve the industry of talent scarcity among companies that are looking for quality talent. According to a report from Analytics India Magazine, there will be no significant contraction for the analytics functions of the domestic and MNC IT firms operating in India. In terms of jobs, the effect of the recession could result in a 10-15% drop in open jobs. The report suggests that once the recession lifts, IT companies will receive a steady flow of orders, which would benefit the analytics offerings – reaching pre-crisis levels three months after the recession ends. The availability of extra talent may uplift the industry further. We May See More Data Science Offshoring To India While the pandemic may soon taper off, it will leave behind the deep economic impact forcing companies to cut costs. India has always been the favourite of offshoring technology talent. It is expected India will become an attractive destination for data science due to cost differential, and the fact that data science work can be done remotely and from home, it sends a positive signal. “The global pandemic has made the world an open talent market, benefitting Indian firms which may provide analytics\/data science services. If H1-B visa holders are forced to relinquish their jobs, India-based firms will likely consume the talent and that would help boost the domestic analytics industry in terms of quality. We will see more global in-house centres (GICs) for data science and analytics consultancy firms flourish with the boost in available talent,” tells Subhobroto Ghosh, Head – Data, Analytics & Actuarial at Allstate India. There Will Better Quality Data Science Talent But how will it impact the talent market in general? Over the years, a lot of professionals without expertise in core data science training had converted into data science. This may change soon, and more value will be given to those with specialised skills. The job losses and the pandemic makes it imperative for companies to create a balance between quality hiring and cost optimisation across business functions. “A lot of companies are looking in India for quality talent, and if professionals make their way to India from the US, it will balance the jobs scenario in India in terms of supply and demand. If talent from the US comes back in, we have better quality analytics skills available in the Indian market,” says Puesh Rajiv Ajmani, Global Head of Analytics & Insights at Square Panda, a cloud-based, AI education company. Experts say that while companies in India will continue to hire for data scientists and analytics professionals, they will favour the quality of skills that bring value to the table. “Now that companies have limited budgets on data science and analytics projects, they will value professionals with lateral movement towards more specific and niche roles. But, with enough availability of talent, it will impact the entry-level and bottom of the pyramid jobs,” adds Puesh. Subhobroto Ghosh agrees and says, “Such professionals have been exposed to higher-end analytics jobs, doing cutting edge work. While exceptional data science talent with MS or PhD degrees are unlikely to come back, their talent would certainly be appreciated if they make their way back to India as they have been embedded into cutting edge data analytics & machine learning work in big tech companies. Such data scientists working in niche areas of AI, NLP, computer vision, etc help them get the exposure which is highly valuable in India.”","excerpt":"Companies in the US have been on the layoff spree during the pandemic. Among the prominent ones, we have witnessed thousands of layoffs in companies such as Lyft, Airbnb and Uber just in the last week or so. According to reports, there are hundreds of thousands of Indian technologists currently stranded in America amid mass […]","categories":["AI Features"],"tags":["Data Scientist","Data Scientist Jobs","Data Scientists","Data scientists India","deep learning projects","high paying jobs in india"],"author_name":"Vishal Chawla","publish_date":"2020-05-19T17:00:00","publication_year":"2020","word_count":908,"keywords":["data science","Go","Data Scientist Jobs","machine learning","programming_languages:R","AI","high paying jobs in india","Data scientists India","programming_languages:Go","computer vision","NLP","analytics","deep learning projects","Data Scientist","R","Data Scientists"],"extracted_tech_keywords":["AI","machine learning","NLP","computer vision","data science","analytics","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-impact-of-data-scientists-returning-to-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058641,"title":"The IT specialist who automated his job for a year without getting caught","content":"An anonymous Reddit post on an IT professional’s goldbricking hack is breaking the internet, amassing 81.2 thousand votes and over five thousand comments. The viral post details how he automated his work in a week. While he was busy playing video games and having fun while working from home, his code did all the heavy lifting. In the subreddit – r\/antiwork, user Throwaway59724 claimed he was able to write, debug, and perfect a simple script to do his entire job. He would just work for 10 mins every day and let the code do the rest. Register for this conference He claims to work as an IT specialist at a mid-size law firm. His company was in the middle of migrating its evidence management system to Cloud, with him in charge. As the only person with admin access to the Cloud, he automated the entire task. His script, scanned the on-site drive for any new files, generated hash values for them, and transferred them to the Cloud. The code then generated hash values again for fidelity. The script he wrote is in batch with some portions of powershell. The base code came from a Google search “.bat transfer files” followed by “.bat how to only transfer certain file types” etc. He wrote in the post, “The trick was making it work with my office, knowing where to scan for new files, knowing where not to scan due to lag (seriously, if you have a folder with 200,000 .txt files that crap will severely slow down your scans. Better to move it manually and then change the script to omit that folder from future searches).” “The local drive is theirs, the Cloud drive is theirs, the VPN is theirs. The PC that I bought with my own money specifically for this task only runs the script. There is no work files or evidence being stored on my PC. The script is literally the only thing on the PC other than the OS,”he added. After the post went viral he was buffeted with questions from curious Redditors. One user asked, why not monetise the code?, to which he replied: “Please understand that this is not some high-end program that’s worth millions. This is a few lines of code written in notepad. It only has value in this situation because the office has no technical skills. This is the type of script people put on GitHub with a $5 price tag linked to their PayPal.” He also had a good defense for the question, what if he gets caught?“I’ll just get rid of the script. I’m running it on equipment that I own so if they lose me, they lose the work. Creating my own job security,” he said.","excerpt":"His script, scanned the on-site drive for any new files, generated hash values for them, and transferred them to the Cloud.","categories":["AI News"],"tags":["reddit","VPN"],"author_name":"Meeta Ramnani","publish_date":"2022-01-17T19:39:40","publication_year":"2022","word_count":456,"keywords":["Go","programming_languages:R","AI","reddit","programming_languages:Go","Git","Aim","GitHub","R","VPN"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GitHub","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/the-it-specialist-who-automated-his-job-for-a-year-without-getting-caught\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":12190,"title":"CES 2017 Roundup — everything showcased around data and analytics","content":"Virtual Reality had a breakthrough moment at CES 2017 The 50th anniversary show of CES 2017 recently concluded in Las Vegas made virtual a reality with a host of advances powered by chip maker Intel, that is making a slow transition to virtual reality, a key growth sector for the technology giant. And it wasn’t just transformative VR experiences, Intel’s booth unveiled many analytics tools tailored around fitness and even tech powered autonomous cars. Analytics India Magazine lists down everything that happened around data and analytics at the popular tech show. CES 2017 was all about Virtual Reality Intel is betting big on VR that can revolutionize entertainment, gaming and the world of sports Virtual reality was a dominant theme at CES 2017 held in Las Vegas where headset donning attendees enjoying immersive reality were a common picture.  And the company that’s pushing the boundaries of VR as an entertainment and gaming medium is Intel. The world’s most famous chip maker unveiled advances in VR — live streaming, volumetric capture and a relatively new technology “merged reality, in collaboration with HypeVR, a computer vision company.  As tweeted by Intel, “with virtual reality, you can jump out of a plane w\/o ever leaving the ground”. True to the statement, attendees at CES 2017 skydived in the rugged mountains of Utah and soaked in idyllic settings in Vietnam. It wasn’t all gaming, Intel also unveiled Project Alloy  headset that enables people to use their own hands instead of controllers for handling virtual objects. Intel CEO Brian Krzanich emphasized how “merged reality,” allows one to interact with the virtual world without removing real surroundings. The Project Alloy headset, factors in the environment into the viewing experience and is not tethered to the PC or mobile. Another area where Intel made headway was in art through Pikazo app that is powered by neural style transfer algorithms. The artificial intelligence powered app can combine any two images into a work of art. Data Analytics took centre stage at CES 2017 Analytics in Sports: It’s pegged as “the computer the size of a button” but Intel Curie module, embedded in a sweatband can deliver tons of data, ranging from jump length, height to dunk power and dribbling frequency. From reshaping sports engagements to connecting with fans on a variety of platforms, CES 2017 acknowledged how data and analytics can play a big in maximizing the performance of athletes. Personalizing guest experience with analytics Another sector where analytics made headway was in personalizing travel experiences. Miami based Carnival Cruise CEO Arnold Donald who gave the first keynote address at the 50th show, announced OCEAN — the One Cruise Experience Access Network, and developed in collaboration with Accenture. OCEAN entails an interactive guest-experience platform, aimed at personalization and improving the holiday experience for travelers. At the heart of OCEAN, are a clutch of advanced capabilities such as machine learning, streaming analytics and contextual awareness. The advanced technology platform is aimed at customizing personal experiences around individual taste and preferences. It’s like having a digital concierge.","excerpt":"The 50th anniversary show of CES 2017 recently concluded in Las Vegas made virtual a reality with a host of advances powered by chip maker Intel, that is making a slow transition to virtual reality, a key growth sector for the technology giant. And it wasn’t just transformative VR experiences, Intel’s booth unveiled many analytics tools […]","categories":["AI News"],"tags":["Intel"],"author_name":"Richa Bhatia","publish_date":"2017-01-17T05:53:08","publication_year":"2017","word_count":507,"keywords":["Go","API","machine learning","artificial intelligence","AI","Git","computer vision","Aim","analytics","R","Intel"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","analytics","Aim","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ces-2017-roundup-everything-showcased-around-data-analytics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10114894,"title":"OpenAI Goes After Blue-Collar Jobs","content":"Having disrupted the white-collar jobs, OpenAI is now looking to automate blue-collar jobs. The company wants to develop next-generation AI models for humanoid robots and aims to help accelerate their capabilities to process and reason from language. OpenAI’s investment in Figure, alongside major players like Microsoft, NVIDIA, Bezos Expeditions, and others, reinforces this. This development also posed a threat to Elon Musk’s Tesla Optimus, which explains why he might have sued OpenAI to begin with. In an interview with Bill Gates, OpenAI CEO Sam Altman said that seven to six years ago, the prediction was that AI would impact blue-collar work first, white-collar work second, and creativity maybe never, but certainly last because that was considered ‘magic and human’. “Obviously, it’s gone in exactly the other direction,” he quipped. Altman is optimistic about the future of robotics. “On the physical hardware side, there’s finally, for the first time that I’ve ever seen, really exciting new platforms being built,” he said and added that at some point, they will be able to use their AI models, along with their language and video understanding, to accomplish amazing things with robots. Interestingly, Figure’s mission fits well with Altman’s, which is to make versatile humanoids that can tackle labor shortages, automate less desirable jobs, and enhance human capabilities and productivity. “We have a massive labor crisis happening, with over 10 million jobs in the US that people don’t want to do,” said Figure chief Brett Adcock, adding that companies are losing 50% to 150% of employees annually and can’t find anybody to do these jobs. “They simply don’t know the solution to help automate these problems. We believe that with Figure, the demand for what we’re doing is almost unbounded. I would say we’re aiming to be one of the first in the world to deploy us at a real scale commercially,” he said. The robotics company released updates for the Figure-01 humanoid earlier this year, where the robot demonstrates making coffee—a skill learned from observing humans. “If it’s not obvious yet, Figure’s humanoid robot is the ultimate deployment vector for AGI,” wrote Adcock on X. OpenAI Robotics Journey This isn’t the first time OpenAI has ventured into robotics. In 2019, OpenAI successfully trained a robotic hand to autonomously solve a Rubik’s cube. Despite this achievement, the dedicated robotics team was disbanded in July 2021. Additionally, the company had previously invested USD 23.5 million in 1X Technologies. “We started robots too early, so we had to put that project on hold. We were dealing with bad simulators and breaking tendons and things like that,” said Altman, adding that over time they realised they first needed intelligence and cognition and later they could figure out how to adapt it to physicality. Even Sora, OpenAI’s latest text-to-video generation model, can simulate some aspects of people, animals, and environments from the physical world when trained at scale. While the rest of the world sees Sora as just a video-generation tool, for OpenAI, it is an attempt to create a simulated reality that could lead to AGI. Simply put, the idea is to create a world model, similar to what Meta chief Yann LeCun and the autonomous vehicle company Wavye are trying to achieve. “We’ve always planned to come back to robotics and we see a path with Figure to explore what humanoid robots can achieve when powered by highly capable multimodal models,” said Peter Welinder, VP of product and partnerships at OpenAI. “AGI will eventually be embodied in some robotic form, whether it’s Optimus (Tesla), Figure (with OpenAI), or another company (seems like there’s a lot),” wrote Sully co-founder CognosysAi. Makes Optimus Jealous Both Optimus and Figure aren’t much different from each other. Figure 01 stands at 5’6” and weighs 60 kg, while Optimus is slightly taller at 5’8” and weighs 73 kg. Both can handle a payload of up to 20 kg, showcasing their comparable sizes. Meanwhile, Tesla recently announced major improvements in its humanoid robots and it looks like it is moving closer to what Musk has envisioned for Optimus. Last year, Optimus just waved on the stage. Now, it can pick up and sort objects, do yoga, and navigate through surroundings. Then there is Boston Dynamics. The company has two robots — Atlas (bipedal) and Spot (quadruped) — which are both used in various industries, including logistics, manufacturing, construction, and inspection. “Boston Dynamics is now primarily owned by Hyundai. I believe Hyundai’s expertise in making cars can be applied to the production of robots, resulting in products that are less expensive and more reliable,” said Marc Raibert, founder of Boston Dynamics. Of late, Amazon too has been experimenting with humanoid robots in select US warehouses, marking a significant step in its automation endeavours. The tech giant aims to optimise efficiency by introducing these robots, named ‘Digit’, which emulate human movements for tasks such as moving and handling items. The prospect of a world where humans are liberated from routine jobs doesn’t seem too bad. Who knows, the future job wars will likely involve more humanoids and less humans.","excerpt":"The prospect of a world where humans are liberated from routine jobs doesn’t seem too bad.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-03-02T16:00:00","publication_year":"2024","word_count":844,"keywords":["Go","OpenAI","AI","Modal","Git","BERT","automation","Aim","ViT","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Go","Git","BERT","ViT","automation","Modal"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-goes-after-blue-collar-jobs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10009057,"title":"How Indian Firms Are Fulfilling Rising Demand For Low Code Software Platforms","content":"Many enterprises aim to build functional software for internal infrastructure servicing and programs for customer communication. However, lack of time and resources hinders their business growth. In the current era, low\/zero code platforms help businesses to rapidly build relevant applications and are becoming an emerging trend in the SaaS market. Such tools allow teams to build specific web applications and come with various customisation tools, including an intuitive and simple drag-and-drop interface. With no-code tools, users can create custom applications for their data collection, design their personal workflows, and set up unique rules that will help them propel their businesses further. According to Gartner by 2024, low-code application development will be accountable for more than 65% of app development activity. In the next 2-3 years, 75% of large businesses will be employing at least four low-code development tools for IT application development. Compared with the traditional software development method, low-code platforms help in much more deployment by deploying pre-built templates and automated software workflows which offer a simple way to build certain applications. The plug n play feature, and quick prototyping also helps expedite turnaround time for making simple applications. “We are seeing an increase in the demand for low-code products. There are several factors that are influencing this trend. A key factor is the commoditization of the SaaS model. Buying software for a company has become similar to buying a new dress online or from a shop. You go to a website, choose the offering, pay the bill and within seconds, you are set up with access to use the software,” says Venkatesh Ramarathinam, CEO of Vuram. Organisations in India are continuing to collaborate with technology firms and startups to make their IT operations more efficient and less costly. They want to automate more processes faster and keep costs down. This has grown during the pandemic. Also, this has caused more demand for low code solutions. “Low code products are definitely gaining popularity, but it is no code that is actually the next big thing in the industry. There is no shortage of innovation and ideas in the Indian market, and no code solutions help turn these ideas to reality. With the current government initiatives like Atmanirbhar Bharat to support startups in India, the demand for such no-code platforms is increasing at an exponential rate. In the past six months, we have experienced growth in paid subscriptions by 5.8X. So yes, there is a strong demand, and it is only on the rise,” says Abhinav Girdhar, Founder of Appy Pie. Appy Pie has multiple new products that are currently in the public beta phase (Appy Pie Chatbot, Appy Pie Live Chat, Appy Pie Design, and Appy Pie Knowledge) and a couple more in the pipeline. “This is the age of no code, where ideas are more important than skills. We ensure that people with great ideas are not inhibited by the lack of their technical prowess,’ told Abhinav. Low Code For SMBs Organizations prefer to use low code for software development for the standard reasons of faster time to go-live, easier upskilling of their internal teams to develop on the platform, lower maintenance costs, rapid continuous improvement cycles and multi-mode delivery. Low-code platforms can be used to build extremely powerful composite applications that provide a 360 degree view of an organization’s process, procedure and data. “There is a growing trend in the Indian market to outsource technology challenges, get expert help, purchase software or platforms that solve a particular or a set of problems really well. This trend is more prominent in the mid-size market. There is a mindset change from building internally, building ground-up, to evaluating pre-built cloud solutions or platforms that come with templates that can be customized for the individual business,” says Venkatesh Ramarathinam. Some low code platforms also give out of the box reporting and analytics capacities in addition to integrated version control systems and one-click merge\/publishing capabilities. Based on application type, the low code software market is divided into Web-based, Desktop & Server-based and Mobile-based tools which are either hosted On-premise or on the Cloud. Automated Software Testing “Low code products are gaining wide popularity in India primarily to shorten the software development cycle so that products can be pushed to the market much sooner in today’s competitive world. The key challenge today is to get all aspects of testing executed quickly with minimal resources, and automation tools can do this flawlessly,” says Ramesh Krish, CTO of Indium Software. Indium, as part of its uphoriX, generates code for Functional Tests. The company has test automation scripts, and the same can be reused for Performance and Security Testing. This saves a considerable amount of time and resources for its customers. Customers use different languages such as C#, Java, Selenium etc. as their preferred coding language. uphoriX automatically generates code in all of these languages and much more, efficiently and effectively. Some of the critical features of uphoriX include low code automation which generates scripts automatically and reduces script generation overhead. Then there is smart Scripting, which according to the company saves up to 30% to 40% of effort on automation script maintenance and reusability of scripts. “Logs are generated for every Test Execution and records are maintained for several test iterations over many builds & releases. This helps our customers to see the trend in quality as they build the software,” explained Ramesh Krish. Today, Indium has expertise in digital automated software testing and low code development and is eyeing $30 million revenue by 2021 by serving its customers in the nation and abroad. The company’s smart unified testing platform – uphoriX integrates functional testing, performance testing, and security intending to expand the pace and frequency of software and application release. With the rapidly expanding customer base that cuts across industry verticals, low code companies are looking to expand their field presence and accelerate their product to increase the number of solutions that come out-of-the-box for various software use cases across multi-cloud and hybrid infrastructure. “Low code is a crowded market, products vary on their purpose, their approach to low code, and the spectrum of features they provide. Under the low-code banner, there are platforms ranging from being a simple forms and workflow builder to a heavy-duty engine that combines complex decisioning, AI, RPA, data, workflow and document management,” tells Venkatesh Ramarathinam. Low code network automation Next up is AppViewX, a hybrid cloud, a low-code application platform that enables the automation and orchestration of network infrastructure. The firm uses intelligent automation and orchestration of network and security services for its platform. “We utilise a big library of pre-built tasks and workflows and its platform enables Ops teams to fast and easily translate business requirements into automation workflows which enhance agility, enforce compliance, eliminate errors, and reduce cost,” Anand Purusothaman, Founder of AppViewX told us. AppViewX has capabilities such as low-code drag and drop automation, closed-loop automated troubleshooting, intuitive visual tools and most important is the application-centric view and management of network infrastructure. AppViewX has the unique advantage of being the pioneer in this space with some of the solutions like the orchestration of application delivery services and certificate lifecycle automation. “Our north star is to make our infrastructure invisible so application teams can deliver innovation rapidly across multiple cloud providers and on-prem infrastructure,” Anand told further. Low code – high value The global low-code platforms market produced $10.3 billion in revenue in the year 2019 and is predicted to grow a CAGR of 31.1% for the period 2020–2030. This year, the coronavirus outbreak has not only caused a massive digital transformation but has also urged companies to have access to software techniques which are faster and with minimal expenditure. During this pandemic, when social distancing and work from home have become a part of the new normal, mobile apps have come to the rescue. However, not every business can afford the time and money that goes into traditional app development. code solutions have made it possible, not only to create mobile apps and websites in minutes but also save a ton of money in doing so. Low code platforms minimise the requirement for more developers, and therefore reducing the hiring cost for writing code and application maintenance. Low code applications can also support enterprises to solve the challenges of siloed IT systems (something very common in India) by connecting data across systems. With agility and cost-effectiveness being crucial here, low-code platforms are frequently turning into popular alternatives for conventional hand-coding tools. Such low code platforms are helping enterprises to develop software applications very rapidly, with minimal manual coding, which is leading to faster software turnaround. Features like drag and drop capability help to build apps 10X quicker, enterprise-grade applications that can be deployed in weeks, gives Indian businesses great agility and a platform to fast experiment and evolve.","excerpt":"Many enterprises aim to build functional software for internal infrastructure servicing and programs for customer communication. However, lack of time and resources hinders their business growth. In the current era, low\/zero code platforms help businesses to rapidly build relevant applications and are becoming an emerging trend in the SaaS market.   Such tools allow teams to […]","categories":["AI Features"],"tags":["low code","low code platform","real time decisioning"],"author_name":"Vishal Chawla","publish_date":"2020-10-07T10:00:10","publication_year":"2020","word_count":1473,"keywords":["Go","API","low code platform","AI","R","Git","RAG","Aim","analytics","GAN","real time decisioning","low code","Java"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Java","Git","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-indian-firms-are-fulfilling-rising-demand-for-low-code-software-platforms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171435,"title":"Perplexity’s SEC Data Integration Makes It the Investor’s AI Sidekick","content":"AI search engine startup Perplexity AI on Thursday rolled out access to SEC filings across its platform, aiming to make technical financial data easily understandable for all types of investors, from students to advisors to day traders. The new SEC\/EDGAR integration allows users to query financial documents directly within Perplexity’s Search, Research, and Labs interfaces. Answers are backed with citations and references, helping users trace insights back to original source documents. The company says the feature is designed to simplify complex reports that typically require domain expertise or expensive tools to interpret. “Everyone deserves access to the same financial information that drives professional investment decisions,” the blog post states, highlighting a contrast with traditional financial data platforms that often gatekeep clarity behind paywalls and complexity. With this launch, users can ask questions about earnings, risks, or strategy and receive natural-language answers grounded in regulatory filings. It’s available for all Perplexity users. However, for Perplexity Enterprise Pro customers, the integration also works alongside datasets from Factset, Crunchbase, and internal company files, offering deeper comparative research. This move positions Perplexity as a potential disruptor in the financial intelligence space, especially for retail investors seeking affordable, context-rich answers with a minimal learning curve. Recently, Perplexity AI launched ‘Labs’, a new feature available to Pro users on the web, iOS, and Android that turns prompts into complete projects like reports, dashboards, spreadsheets, and simple web apps. It is positioned as an evolution of the platform’s existing ‘Deep Research’ tool—now renamed ‘Research’. Labs supports more complex and extended workflows. Unlike earlier modes that focused on information retrieval, Labs acts as a virtual project team, generating structured outputs using web browsing, code execution, and asset creation tools. Users can access generated content and interactive elements via dedicated ‘Assets’ and ‘App’ tabs.","excerpt":"“Everyone deserves access to the same financial information that drives professional investment decisions.”","categories":["AI News"],"tags":["Data Science","Perplexity"],"author_name":"Ankush Das","publish_date":"2025-06-06T14:30:52","publication_year":"2025","word_count":295,"keywords":["Perplexity","TPU","programming_languages:R","AI","Aim","Data Science","R","startup"],"extracted_tech_keywords":["AI","Aim","TPU","R","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/perplexitys-sec-data-integration-makes-it-the-investors-ai-sidekick\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25599,"title":"Bosch To Invest ₹17 Billion In India, Will Focus On Expanding Bengaluru Smart Campus","content":"Dr Volkmar Denner, chairman of the board of management at Bosch Group. The Bosch Group sees strong signs of recovery in the Indian market and forecasts a positive development over the mid-term, they said in a new report. “The Indian economy is on the rise again and holds tremendous potential”, said Dr Volkmar Denner, the chairman of the board of management of the Bosch Group during his visit to India. In 2017, the leading global technology and services company posted strong double-digit growth as sales went up of 15.4 percent to €2.2 billion. With a projected 7.7 percent GDP-growth for the Indian economy in 2018 and a similar level for the coming years, Bosch expects to continue with its momentum. Main contributors are the development of the local automotive industry, which holds opportunities for diesel technology, electromobility and connected mobility as well as the government initiatives for connected manufacturing and connected cities. “To meet the rising demand in the Indian market with tailored solutions and expand India’s strong role in our global network, we will invest ₹17 billion ($250 million) over the next three years”, announced Denner. The German auto parts maker also added that a major share of this amount will be used to expand Bosch’s smart campus in Adugodi and modernise manufacturing facilities in the country. Bosch has four key plans for India: Tapping into India’s potential:  “The Indian economy is on the rise again and holds tremendous potential,” Denner said. „ Powertrain: Bosch is committed to addressing the transition from BS-IV to BS-VI in India and offers a new breakthrough in Diesel technology „ Connected mobility: Bosch is evolving into an end-to-end provider of mobility services „ Artificial Intelligence: India is a key location for global Bosch Group activities “Most of the products in the near future will be linked to artificial intelligence. These products will either possess that intelligence themselves, or AI will play a key role in their development or manufacture,” said Denner. In 2017, the company invested €300 million into its Bosch Center for Artificial intelligence (BCAI) across three continents — one of the main locations being in Bengaluru, India — next to centres in Sunnyvale, U.S. and Renningen, Germany. Additionally, Bosch has also partnered with IIT-Madras and set up a Robert Bosch Centre for Data Science and Artificial Intelligence at IIT- M with a fund of ₹4 crore per year for five years. “Bosch India with its 31,000 highly-talented associates — thereof 18,000 in R&D – is well-equipped to take this success to the next level,” added Denner.","excerpt":"The Bosch Group sees strong signs of recovery in the Indian market and forecasts a positive development over the mid-term, they said in a new report. “The Indian economy is on the rise again and holds tremendous potential”, said Dr Volkmar Denner, the chairman of the board of management of the Bosch Group during his […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","bosch","IIT Madras","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2018-06-21T06:47:16","publication_year":"2018","word_count":425,"keywords":["data science","Go","bosch","artificial intelligence","programming_languages:R","AI","IIT Madras","Machine Learning","Git","BERT","llm_models:BERT","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","data science","R","Go","Git","BERT","ViT","llm_models:BERT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bosch-to-invest-%e2%82%b917-billion-in-india-will-focus-on-expanding-bengaluru-smart-campus\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":57738,"title":"Why Jupyter Notebooks Are So Popular Among Data Scientists","content":"Born out of IPython in 2014, Jupyter Notebook has seen an enthusiastic adoption among the data science community, to an extent where it has become a default environment for research. By definition, Jupyter is a free, open-source interactive web-based computational notebook. Computational notebooks have been around for several years; however, Jupyter, in particular, has exploded in popularity over the past couple of years. This nifty tool supports multi-language programming and therefore became the de facto choice for data scientists for practising and sharing various codes, quick prototyping, and exploratory analysis. Although there is no dearth of language-specific IDEs (Integrated Development Environments), such as PyCharm, Spyder, or Atom, because of its flexibility and interactiveness, Jupyter has exploded in popularity among data scientists. Jupyter Notebook has also gained massive traction within digital humanities as a pedagogical tool. According to an analysis by GitHub, it has been counted that more than 2.5 million public Jupyter notebooks were shared in September 2018, which is up by 200,000 counted in 2015. So before we delve deeper into the features and advantages of Jupyter, and why it is considered to be the best platform for data scientists, we would discuss what a Jupyter Notebook is. What Is A Jupyter Notebook? An indirect acronym of three languages — Julia, Python and R — Jupyter Notebook is a client-based interactive web application that allows users to create and share codes, equations, visualisations, as well as text. The notebook is considered as a multi-language interactive computing environment, which supports 40+ programming languages to its users. With Jupyter Notebook, users can bring in data, code and prose in together to create an interactive computational story. Whether to analyse a collection of written text, creating music or art or to develop engineering concepts, Jupyter Notebook can combine codes and explanations with the interactivity of the application. This makes it a handy tool for data scientists for streamlining end to end data science workflows. The Jupyter Notebook can be installed using the Python pip command. And, if using Anaconda, then it gets automatically installed as part of the Anaconda installation. It is combined of three components — the notebook application, kernels, and notebook documents. The notebook web application is used for writing and running codes in an interactive way, however kernels controls the system by running and introspecting users’ codes. And thirdly, notebook documents are the self-contained documents of all the contents visible in the notebook. Each document in the notebook has the kernel that controls it. According to Lorena Barba, a mechanical and aeronautical engineer at George Washington University in Washington DC for — data scientists, Jupyter has emerged as a de-facto standard. Purpose of Jupyter Notebook Data CleaningStatistical ModellingTraining ML ModelsData visualisation What Makes Jupyter Notebook The De Facto Choice Due to the rising popularity of open-source software in the industry, along with rapid growth of data science and machine learning the Jupyter Notebook has become ubiquitous among data scientists. Apart from supporting multi-language programming, this interactive web-based computing platform also supports Markdown cells, allowing for more detailed write-ups with easy formatting. With Jupyter, the final product can be exported as a PDF or HTML file, which can be presented on a browser or can be shared on sites like GitHub. Jupyter Notebooks are saved in the structured text files — JSON (JavaScript Object Notation) — which makes it extremely easy to share. Fernando Pérez, the cofounder of Jupyter once said, that growth of Jupyter is due to the improvements that were made in the web software, which drives applications such as Gmail and Google Docs and the ease with which it facilitates access to remote data which might otherwise be impractical to download. The maturation of scientific Python and data science is another reason for this platform to gain traction. Additionally, Jupyter Notebooks have played an essential role in the democratisation of data science, making it more accessible by removing barriers of entry for data scientists. Benefits Although Jupyter has been developed for data science applications, which are written in languages like Python, R and Julia, the platform is now used in all kinds of ways for projects. Apart from that, by removing the barriers for data scientists, Jupyter made documentation, data visualisations, and caching a lot easier, especially for hardcore non-technical folks. A data science enthusiast said, “Jupyter Notebook should be an integral part of any Python data scientist’s toolbox. It’s great for prototyping and sharing notebooks with visualisations.” So, let’s explore some of the benefits. Exploratory Data Analysis: Jupyter allows users to view the results of the code in-line without the dependency of other parts of the code. In the notebook, every cell of the code can be potentially checked at any time to draw an output. Because of this, unlike other standard IDEs like PyCHarm, VSCode, Jupyter helps in in-line printing of the output, which becomes extremely useful for exploratory data analysis (EDA) process. Easy Caching In Built-In Cell: Maintaining the state of execution of each cell is difficult, but with Jupyter, this work is done automatically. Jupyter caches the results of every cell that is running — whether it is a code that is training an ML model or a code that is downloading gigabytes of data from a remote server. Language Independent: Because of its representation in JSON format, Jupyter Notebook is platform-independent as well as language-independent. Another reason is that Jupyter can be processed by any several languages, and can be converted to any file formats such as Markdown, HTML, PDF, and others. Data Visualisation: As a component, the shared notebook Jupyter supports visualisations and includes rendering some of the data sets like graphics and charts, which are generated from codes with the help of modules like Matplotlib, Plotly, or Bokeh. Jupyter lets the users narrate visualisations, alongside share the code and data sets, enabling others for interactive changes. Live Interactions With Code: Jupyter Notebook uses “ipywidgets” packages, which provide standard user interfaces for exploring code and data interactivity. And therefore the code can be edited by users and can also be sent for a re-run, making Jupyter’s code non-static. It allows users to control input sources for code and provide feedback directly on the browser. Documenting code samples: Jupyter makes it easy for users to explain their codes line-by-line with feedback attached all along the way. Even better, with Jupyter, users can add interactivity along with explanations, while the code is fully functional. Outlook Combining all the benefits mentioned above of Jupyter Notebook, the key point that emerged is that using Jupyter is an easy way of crafting a story with data. Today, Jupyter has transformed completely and grown into an ecosystem where it comprehends — several alternative notebook interfaces like JupyterLab and Hydrogen, interactive visualisation libraries and tools compatible with the notebooks.","excerpt":"Born out of IPython in 2014, Jupyter Notebook has seen an enthusiastic adoption among the data science community, to an extent where it has become a default environment for research. By definition, Jupyter is a free, open-source interactive web-based computational notebook. Computational notebooks have been around for several years; however, Jupyter, in particular, has exploded […]","categories":["IT Services"],"tags":["Data Cleaning","Julia Language","julia scientific programming","Jupyter","Jupyter Notebook","Pycharm"],"author_name":"Sejuti Das","publish_date":"2020-02-28T19:00:00","publication_year":"2020","word_count":1131,"keywords":["data science","Jupyter Notebook","machine learning","Plotly","AI","TPU","ML","Python","Julia Language","Pycharm","Matplotlib","Data Cleaning","Jupyter","R","julia scientific programming"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Jupyter","Matplotlib","Plotly","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-jupyter-notebooks-are-so-popular-among-data-scientists\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10077105,"title":"Honeywell Unveils Sustainable Ethanol-to-Jet Fuel Technology","content":"Multinational tech conglomerate Honeywell on Monday announced a new ethanol-to-jet fuel (ETJ) processing technology that converts cellulosic or sugar-based ethanol into sustainable aviation fuel (SAF). The new solution is an effort to meet the rising demand of the global drive for 2030 sustainable aviation fuel (SAF) mandates. The jet fuel produced at Honeywell’s ethanol-to-jet fuel process, depending on the type of ethanol feedstock used, can reduce greenhouse gas (GHG) emissions by 80% on a total lifecycle basis. Honeywell Sustainable Technology Solutions vice president Barry Glickman says, “Honeywell pioneered SAF production with its ecofining technology and our new ethanol-to-jet fuel process builds on that original innovation to support the global aviation sector’s efforts to reduce GHG emissions. The solution would also meet SAF production targets with an abundant feedstock like ethanol. Honeywell’s ethanol-to-jet process is now ready to provide a pathway to lower carbon-intensity SAF.” As demand for SAF continues to grow, the aviation industry is still challenged by limited supplies of traditional SAF feedstocks such as animal fats, vegetable oils, and waste oils. The press release stated that Honeywell’s technology would use high-performance catalysts and heat management capabilities to maximize production efficiency; resulting in a cost-effective and lower carbon intensity aviation fuel.Headquartered in Charlotte, North Carolina, Honeywell primarily operates in four areas of business such as aerospace, performance materials and technologies, building technologies, safety and productivity solutions. Around 60% of the company’s new product introduction research investment is directed toward improving social and environmental outcomes for customers.","excerpt":"The new solution is an effort to meet the rising demand of the global drive for 2030 sustainable aviation fuel (SAF) mandates.","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-10-12T17:05:35","publication_year":"2022","word_count":247,"keywords":["programming_languages:R","AI","innovation","ViT","R"],"extracted_tech_keywords":["AI","R","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/honeywell-unveils-sustainable-ethanol-to-jet-fuel-technology\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168234,"title":"How AI is Redefining Software Engineering Roles at HCL","content":"For quite some time now, the discussion around AI potentially replacing software engineering jobs has gained significant attention. Tools like Cursor, Windsurf, and Claude have become dominant forces in the coding world since last year, providing auto-completion features and generating code for a wide range of use cases. Recently, the internet has seen a surge of experiments where people demonstrate what has come to be known as ‘vibe coding’, a term introduced by OpenAI co-founder Andrej Karpathy. In this approach, simply having an idea is enough to generate functional code. While AI tools like GitHub Copilot have long assisted professionals by autocompleting code, newer platforms such as Cursor, Replit, Bolt, and Lovable are now making it easier for beginners to get involved in coding, offering enhanced features that simplify the process. Against the backdrop of these innovations, one thing is clear: AI is not only enhancing productivity but also fundamentally redefining job roles across the enterprise. The role of a software engineer is no longer confined to manual coding; it now encompasses collaboration with AI tools that are transforming the industry in profound ways. Amjad Masad, CEO of Replit, has voiced his unease with the term ‘vibe coding’, suggesting that it trivialises the profound capabilities of generative AI and downplays the creative effort involved in software development. He maintains that AI should be viewed as a tool to empower more individuals to create, rather than a means of replacing them. On the topic of employment, Masad remarked, “We’re all going to have jobs—they’re just going to be very powerful.” Masad recently suggested that learning to code may no longer be essential. He shared a video clip on X in which he agrees with Dario Amodei of Anthropic, who predicted that the majority of code in the near future would be generated by AI. Amodei anticipates that AI could be responsible for writing up to 90% of all code within the next six months. Explaining the indispensable nature of such tools, Abhishek Upperwal, CEO of Soket AI Labs, told AIM, “These tools are only as good as the person using them. If you know your stuff, they’ll blow your mind with the quality. But hand them to someone without decent programming skills, and it’s chaos. They’re powerful, but they’re not magic. Yet.” Augmenting Job Roles, Not Replacing Them A Reddit user revealed that he was recently laid off from a team of five frontend engineers. Reflecting on the trend, Alan Flower, executive vice president and head of AI Labs at HCLTech, clearly stated, “It’s not so much about replacing job roles. It’s about augmenting the role with AI.” Flower highlighted how business leaders worldwide perceive AI and said, “They see AI as a net gain in terms of productivity.” However, this gain is not restricted to merely accomplishing more tasks.  Instead, it encompasses both the speed and quality of execution. This change is especially relevant in areas such as IT and engineering, where work backlogs are a persistent challenge. “Every business leader that I engage with has got what I would call a never-ending backlog of things they can never get done,” Flower noted. AI enables businesses to “start more projects, complete more projects, and drive transformation further and deeper”. Considering this, one key area of transformation is software engineering. Flower highlighted that HCLTech’s AI Force platform has been pivotal in this shift. The patented GenAI platform drives service transformation throughout the software engineering and IT operations lifecycle. “With AI Force, every role within a software engineering team—product owners, testers, engineers, scrum masters—has received AI enhancement,” he explained. For instance, while developers usually spend only around 30% of their time writing code, the remainder is devoted to testing, documentation, and analysis. AI is now assisting in streamlining the entire lifecycle. “We have observed perhaps over a 60% productivity increase in certain areas,” Flower disclosed. Creating New Roles AI isn’t just changing old roles; it is creating new ones. Back in 2017, a report by the Institute for the Future stated that 85% of the jobs that would exist in 2030 haven’t even been invented yet. The same is true for prompt engineering, which has now emerged as a critical new skill in the era of generative AI. “There is a massive demand for people who can drive modern, generative AI platforms well,” Flower shared. Another emerging field is AI operations, where professionals manage and run complex AI stacks. “That’s not just technical, it encapsulates an awful lot of governance as well,” he added, highlighting the need for a new generation of professionals with a hybrid skillset that combines engineering prowess and ethical oversight. Reskilling the Workforce By 2027, 44% of workers are expected to shift their core skills. The World Economic Forum’s Future of Jobs Report shows that nearly half of all employees will need to adapt to technological advancements in the near future. Reskilling and upskilling are essential to keep pace with this change, and HCLTech is deeply invested in this journey. “Our global engineering workforce is already using AI Force,” Flower said. Even beyond the engineering teams, the company encourages its workforce to adopt AI tools like Microsoft Copilot. Flower noted that there is strong evidence suggesting that once individuals begin using these tools and experience some initial benefits, the process becomes habitual. “We’ve already deployed solutions with smaller, customised models that might, for example, understand the healthcare domain,” he explained, referencing their GenAI-powered clinical advisor platform. HCL Tech’s Next Big Leap HCL’s smaller models work in tandem with large foundational models, forming multi-model architectures that are more efficient. Flower explained that a large language model (LLM) isn’t necessary for every task. The most common approach is to use an LLM to interact with the end user while using a selection of more optimised, smaller models to handle specific tasks in the background. Agentic AI is another frontier where HCLTech is making strides. “We publicly showed how we were using agentic AI in production,” Flower shared. These agents are trained to handle traditional IT operations roles such as network and cloud engineering or service desk functions. “This is the perfect example of how someone working at a service desk can take one of these tickets and think, ‘I’m going to delegate it,’” Flower said. The result? Employees have more time to provide better customer service, while overall productivity and efficiency significantly improve. Looking to the future, Flower reaffirmed that the demand for AI-driven transformation is stronger than ever. “My labs have completed over 500 GenAI engagements…Clients come to us with transformational projects,” he shared. During their Q4 earnings calls, Indian IT companies such as Accenture, TCS, and Wipro highlighted their strong focus on AI-led innovation and automation, with an emphasis on upskilling talent. Milind Lakkad, CHRO at TCS, addressed concerns about AI’s impact on hiring. When asked about the effect of AI upskilling on TCS employees, he said, “We don’t see any impact on hiring because of AI.”","excerpt":"The role of a software engineer is no longer confined to manual coding; it now encompasses collaboration with AI tools.","categories":["AI Features"],"tags":["HCL Technology","Vibe Coding"],"author_name":"Shalini Mondal","publish_date":"2025-04-18T10:00:00","publication_year":"2025","word_count":1159,"keywords":["Anthropic","GenAI","agentic AI","OpenAI","AI","ML","RAG","Vibe Coding","Aim","prompt engineering","generative AI","HCL Technology"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","agentic AI","OpenAI","Anthropic","Aim","RAG","prompt engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-ai-is-redefining-software-engineering-roles-at-hcl\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":54275,"title":"Value Of Open Source Is To Attract Right Developers, Says Sudhir Tiwari Of ThoughtWorks India","content":"As a technology company, ThoughtWorks has been one of the leading proponents in the movement for agile software development and has developed a range of open source products used by thousands of companies — including Selenium, GoCD, CruiseControl, Gauge, Talisman, Bahmni, FreedomBox, QuickFix and others. The company’s expertise since its inception has been the creation of custom software applications and consulting services related to software development, design, architecture for hundreds of clients across the globe. To understand the company’s latest innovation strategy, Analytics India Magazine connected with Sudhir Tiwari, Managing Director of ThoughtWorks India, who has had years extensive experience in a variety of development methodologies. Sudhir currently manages ThoughtWorks’ operations in India, including its software delivery excellence, development of new software capabilities and building high-performance teams. Here are the edited excerpts from the interaction — AIM: With so much focus on customer experience today, what role does software play for businesses in India? Sudhir: Core technology is for every business right now, and I think the role is related to the business strategy in terms of what you are building and how you are managing software. We’ve seen that trend only in India only so far, in recent years. Earlier, India was just a market where people used to buy products off the shelf. No one would build custom software at all earlier as they never believed in its value. So it was a market where everyone bought products, whether they worked or not, is a separate question. The role of software has changed in the last decade to now become a business differentiator. If that hadn’t happened, the digital natives that India has been witnessing, such as Ola and Flipkart wouldn’t have survived. AIM: Can you share with us a few examples of ThoughtWorks’ technology implementations for its clients, for technologies like AI or Blockchain? Sudhir: We have teams in ThoughtWorks which treat every project as a data project to move to this whole movement of analytics. Using AI, we are helping clients create better staffing teams and get the best-ranked teams for specific projects. So we are looking at moving beyond regular analytics and making predictions using AI models with a lot of our clients for different use cases. For example, we are doing some work on the translations for an Indian client right now. It is very tough to translate Indian languages. Let’s say if there is a judgment in English, how to do it in a legal Indian language. Everyone relies on Google, but we need our own right now. For Blockchain, we recently won an award jointly in the UK for a joint blockchain project we did with one of our clients — VAKT, which is a consortium of leading energy companies and banks. The organisations in the consortium are: BP, Equinor, Shell, Gunvor, Koch, Mercuria, ABN Amro, ING, Société Générale, Chevron, Reliance and Total. Traditionally, the post-trade process would involve both parties checking paperwork and changing records on separate, manual systems, which is time-consuming and error-prone. ThoughtWorks leveraged blockchain technology to build a digital platform, allowing all parties to manage that process in the same place for the very first time. AIM: What do you think about the democratisation of technology through open-source? What value does an open-source strategy create for ThoughtWorks? Sudhir: I think we have primarily created open-source tools for developers and that has been our forte. For example, Selenium which is a portable open-source framework for testing web applications, or CruiseControl, our open-source Java-based framework for a continuous build process. Recently, we put out an open-source product called Bahmni- an OpenMRS (Medical Record System) specifically for healthcare providers. Bahmni is doing extremely well right now and has got more than 100 implementations worldwide where ThoughtWorks has not contributed. We have also seen Bahmni conferences, where people look up to us in terms of the direction of what needs to happen on the product. So, open-source is just in the DNA of the company and that will continue. We are in the top ten in open source contributors across all firms. I think the biggest value of open source strategy is to attract the right developers for software projects. And it also helps the clients as they know the software they are using not closed source, and so they can pick and tweak it specific to needs. AIM: What technology opportunities do you see specific to India for 2020? Sudhir: Our local market is the fastest-growing market, with 300% growth last year. So the enterprises here are changing rapidly. For example, a lot of banks have understood that they need to compete with the Fintechs which have risen with the power of UPI and India Stack. So, that’s a massive change which is happening. A lot of GICs are opening in India. India has almost 1500 Global In-House Centres (GICs). We expect a lot more to come. The GIC movement is moving from cost to value to innovation. So the ability to build business intelligence in India for clients who are anywhere in the world, so that’s a massive opportunity for us. AIM: Do you think that CIOs and other technology executives in India are catching up to the latest innovation in software? Sudhir: The role of CIOs in India is moving from an executive who just worked to get the best deal or bargain on a technology product to new roles where they are now forced to understand technology. So a lot of executives, who have worked in Amazon or Flipkart, are now moving into tech leadership roles like CIOs and CTOs. At the same time, technology decision-makers are grappling with the whole operational structure. When you move onto the whole digital bandwagon, almost everything needs to change in the new world. But, people are just hanging on to the old structures to their detriment. For ThoughtWorks, we have identified common characteristics of companies that are doing well in tech leadership. We call it the Digital Fluency Model and this will be a piece of software asset that we are going to release it in the market as thought leadership for CIOs, CTOs and other technology leaders. AIM: There are so many legacy systems that still exist in India. With all the technical debt there is, how do companies then adopt the innovative software coming out? Sudhir: That’s a real challenge, I think. A great amount of volume is locked in those systems and depending on the type of client, and you can replace such systems, remove the technical debt in the systems or take them apart from one by one. And that’s where a lot of investment is going. You have legacy banking products which are just totally closed, and all the valuable data exists in many closed proprietary databases. In some cases, people know the system is running, but they don’t know how it is running. But, slowly, investments are changing them or companies are building custom software. There is also a big movement towards custom and open source because you can’t have closed systems anymore. A lot of our focus is to build these systems which are secure from day one across different systems using Open APIs, which even legacy clients can access and create innovative products.","excerpt":"As a technology company, ThoughtWorks has been one of the leading proponents in the movement for agile software development and has developed a range of open source products used by thousands of companies — including Selenium, GoCD, CruiseControl, Gauge, Talisman, Bahmni, FreedomBox, QuickFix and others. The company’s expertise since its inception has been the creation […]","categories":["AI Features"],"tags":["Interviews and Discussions","open-source software","software developers India","thoughtworks"],"author_name":"Vishal Chawla","publish_date":"2020-01-17T11:18:30","publication_year":"2020","word_count":1210,"keywords":["Go","API","AI","open-source software","Git","Java","RAG","Aim","thoughtworks","analytics","GAN","R","software developers India","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Java","Git","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/sudhir-tiwari-thoughtworks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10005255,"title":"Hands-on Guide To Extractive Text Summarization With BERTSum","content":"With the advent of new research in the state of the art models, more powerful algorithms are being developed that focus on low code environments to make Machine learning accessible for everyone. Analogous to transfer learning models for image classification or face recognition in the field of computer vision, BERT, Roberta, XLNet are highly powerful models that solve problems using natural language processing. These models have dominated the world of NLP by making tasks like POS tagging, sentiment analysis, text summarization etc very easy yet effective. This article is a demonstration of how simple and powerful transfer learning models are in the field of NLP. We will implement a text summarizer using BERT that can summarize large posts like blogs and news articles using just a few lines of code. Text summarization Text summarization is the concept of employing a machine to condense a document or a set of documents into brief paragraphs or statements using mathematical methods. NLP broadly classifies text summarization into 2 groups. Extractive text summarization: here, the model summarizes long documents and represents them in smaller simpler sentences. Abstractive text summarization: the model has to produce a summary based on a topic without prior content provided. We will understand and implement the first category here. Extractive text summarization with BERT(BERTSUM) Unlike abstractive text summarization, extractive text summarization requires the model to “understand” the complete text, pick out the right keywords and assemble these keywords to make sense. There cannot be a loss of information either. So, how does BERT do all of this with such great speed and accuracy? Looking at the image above, you may notice slight differences between the original model and the model used for summarization. The input format of BERTSUM is different when compared to the original model. Here, a [CLS] token is added at the start of each sentence in order to separate multiple sentences and to collect features of the preceding sentence. There is also a difference in segment embeddings. In the case of BERTSUM, each sentence is assigned an embedding of Ea or Eb depending on whether the sentence is even or odd. If the sequence is [s1,s2,s3] then the segment embeddings are [Ea, Eb, Ea]. This way, all sentences are embedded and sent into further layers. BERTSUM assigns scores to each sentence that represents how much value that sentence adds to the overall document. So, [s1,s2,s3] is assigned [score1, score2, score3]. The sentences with the highest scores are then collected and rearranged to give the overall summary of the article. Now that we have understood the working of BERTSUM, let us implement them to a custom article. Loading the custom data You can select a paragraph or two from any source of your choice. I have selected two paragraphs from the WHO report on coronavirus. To download the article click here. Once you have your data ready, go ahead and install the bert-sum library using the command pip install bert-extractive-summarizer After downloading this, let us read our file and summarize it. get_corona_summary=open('corona.txt','r').read() BERTSUM has an in-built module called summarizer that takes in our data, accesses it and provided the summary within seconds. from summarizer import Summarizer model = Summarizer() result = model(get_corona_summary, min_length=20) summary = \"\".join(result) print(summary) Thus, our two-paragraph information has been converted into a small brief summary. Using a Web scaped article Python makes data loading easy for us by providing a library called newspaper. This library is a web scraper that can extract all textual information from the URL provided. Newspaper can extract and detect languages seamlessly. If no language is specified, Newspaper will attempt to auto-detect a language. To install this library, use the command pip install newspaper3k Once this is done, let us select an article that we need the summary for. I have chosen this article for the purpose of this project. Let us load our dataset. from newspaper import fulltext import requests article_url=\"https:\/\/analyticsindiamag.com\/is-common-sense-common-in-nlp-models\/ \" article = fulltext(requests.get(article_url).text) print(article) from summarizer import Summarizer model = Summarizer() result = model(article, min_length=30,max_length=300) summary = \"\".join(result) print(summary) Here is a one-paragraph summary of the contents in the article. You can choose to limit the size of the summary as per your needs. Here I have limited it to 300 words. Conclusion BERT is undoubtedly a breakthrough in the use of Machine Learning for Natural Language Processing. The fact that it’s approachable and allows fast fine-tuning will likely allow a wide range of practical applications in the future. The research in the field of NLP is trying to reach human-level every day.  With models like this, there is a wide range of applications where it finds use. The ease in which these methods can be implemented makes it accessible to not only developers but even business analysts.","excerpt":"This article is a demonstration of how simple and powerful transfer learning models are in the field of NLP. We will implement a text summarizer using BERT that can summarize large posts like blogs and news articles using just a few lines of code.","categories":["Deep Tech"],"tags":["BERT","BERT tutorial","extract data from twitter","Text summarization"],"author_name":"Bhoomika Madhukar","publish_date":"2020-08-21T17:00:55","publication_year":"2020","word_count":791,"keywords":["Go","machine learning","sentiment analysis","ML","Text summarization","computer vision","BERT","BERT tutorial","extract data from twitter","NLP","RAG","Python","analytics","R"],"extracted_tech_keywords":["machine learning","ML","NLP","computer vision","analytics","RAG","sentiment analysis","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-to-extractive-text-summarization-with-bertsum\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10061295,"title":"AIM Long Reads: Why can’t India afford to miss the Web3 boat?","content":"The third generation of the internet, Web3, is a combination of evolving technologies including blockchain, NFTs, digital avatars and more. The defining feature of Web3 is decentralisation. Writer and editor Max Reads explained Web3 as “software built on the blockchain — which is synchronised and held in common across multiple computers, instead of on servers owned by a single company — will be naturally decentralised, by virtue of how the blockchain works.” Web1 is the brainchild of Tim Berners-Lee. The era of ‘Read-Only Web’, lasted roughly from 1989 to 2005. In comparison to Web1–characterised by static pages–Web2 is more interactive with user-generated content and interoperability. According to Chris Dixon of venture capital firm a16z, Web3 combines the best of both worlds; the decentralised ethos of Web1 with the interaction of Web2. Technical and realistic barriers in creating a decentralised network notwithstanding, tech enthusiasts, entrepreneurs, and big tech companies are rolling in the Web3 hype. Facebook’s Meta rebranding, coinciding with the ring-fencing of USD 10 billion to build out the metaverse, is a good case in point. Meanwhile in India, the government announced a 30% tax on cryptocurrencies and other virtual digital assets during the 2022 Union Budget. However, the country lacks a robust policy to support Web3. With over 749 million internet and 518 million social media users in 2020, India is the world’s second-largest internet population after China. The tipping point was the entry of Reliance Jio Infocomm. The network carrier offered cheap high-speed 4G data services, racking up over 100 million customers in just six months. That said, India was late to the Web2 party. Missed opportunities India fell short of leveraging Web1 and Web2  movements because of a lack of infrastructural facilities, policies and misplaced priorities. “We lacked focus on semiconductor design and manufacturing,” said Pradeep Mishra, Senior Vice President, Purchasing and Supply Chain at VE Commercials. “India’s focus post-independence was to develop capabilities in manufacturing to address fundamental self-reliance on basic sectors of Steel, Oil & Gas and Mineral development. This helped build indigenous capabilities and generated job opportunities outside the agriculture sector. We also set up premier educational institutions, like IITs and IIMs, to bridge knowledge, technology, research, and skill gaps. In the second wave of Industrialisation, our policy focused on developing auto, IT and IT-enabled industries,” he said. India’s IT initiatives during globalisation led to economic growth from the 90s to 2010, and made the country a “global powerhouse for software development”, he added. India’s potential Software development is the backbone of Web3. According to a US-India Strategic Partnership Forum (USISPF) report, Web3 will contribute USD 1.1 trillion to India’s GDP, propelling the digital Indian economy’s value from USD 5 billion in 2021 to USD 262 billion by 2031. “The current wave is driven by IT-led manufacturing. India is on the cusp of a unique opportunity to expand and enhance our IT capabilities with AI collaboration. Our strong manufacturing ecosystem is becoming the biggest global hub for technology products. This can cover the entire spectrum from research, design, development, manufacturing, and servicing. Opportunity is immense for India to announce its arrival in this space,” said Pradeep. With China’s iffy stance on Crypto, India has one less rival to contend with to make headway in Web3. Why India should not miss out on Web3 Biswajit Biswas, Chief Data Scientist, Tata Elxsi, described Web3 as the natural progression of the current state of Web. “The current Web revolutionised people-to-people connect or gave rise to social media and mobile-first technology where sophisticated applications can be built almost overnight. Web3 thought process is AI-first, highly decentralised, hyper-personalised and bringing Meta or VR at its core,” he said. India missed out on the first-mover advantage in Web1 and Web2. Getting in on the ground floor of Web3 can power India’s ambition to become a tech superpower. On the other hand, reluctance to embrace blockchain technology can hurt the Digital India mission in more ways than one. “The internet, as we know it today, is transforming faster than we anticipated and in ways that we could never have anticipated. In turn, the impact of this transformation will change how we collaborate and how we engage with each other and with other digital entities,” said Suresh Chintada, CTO, Subex. Being a developing country, most of India’s population cannot access predominantly English and expensive technology. The decentralised nature of Web3 and the convergence of technologies will hopefully level the playing field. What do we need? “There are four key parameters that will enable Web3. These are the availability of massive computing power, very high bandwidth, smart embedded devices, and data & AI\/ML technologies. While Web2 was about humans consuming content and communicating with other humans, Web3 demands intelligence in pretty much everything because, along with humans, machines will be able to interpret and consume the massive amounts of data on the web,” said Suresh. Meanwhile, Biswajit said Web3 products will have implications on hardware and other infrastructures.“For example, both AI and blockchain rely on heavy computing and storage resources, low latency communication with sufficient bandwidth, bringing 5G to the fore, and its ubiquitous deployment is essential. So, Web3 is not a standalone technology. It is, in fact, in the cross current of some of the megatrends like AI, Blockchain, IoT, and 5G, which are heavily dependent on infrastructure and the interconnected ecosystem,” he added. Suresh dilated on India’s status quo with respect to Web3 adoption. “India is on the cusp of unleashing 5G networks by 2023, which would help address bandwidth requirements and improve access to Web3. We have some ways to go regarding the availability of cheap computing infrastructure, but it will become a reality in the future. Quantum computing was a theory a few years back, but now it’s a reality. We will see that being offered on tap in the years to come,”he added. Conducive policies Globally, China, Russia, and Colombia have banned cryptocurrencies and investment in blockchain technology. On the other hand, the US and Europe have adopted frameworks to encourage blockchain technology. As of 2021, the US has introduced 35 bills focused on cryptocurrency and blockchain policy. In 2021, the India government introduced the National Blockchain Strategy, NSB, a 52-page document and a framework for businesses developing services based on blockchain tech. According to a cabinet note accessed by NDTV, the government plans on bringing in SEBI to regulate crypto as an asset. The ‘assets’ will be on Indian exchanges, and India citizens are not allowed to keep them on foreign exchanges or in private wallets. To enable Web3 in India, we need solid frameworks and favourable policies. The digital divide Today, many Indians can not access Web2 facilities given language barriers, unavailability of the network, expensive services and lack of tech knowledge. “The technology is not much productive until it is helping a local farmer who speaks vernacular language by telling him the best techniques to store his grains or the latest fertiliser practices, spotlighting a great opportunity for technology to solve the challenges of feeding the billions of people on our planet,” said Vishal Dhupar, Managing Director, Asia South, NVIDIA. He said it is critical to ensure tech accessibility through policies and infrastructural development. “Web3, in its evolutionary trajectory, tries to optimise the processing and consumption of data and services at the spatial interaction layer. This technology will accommodate interaction with every possible sensory interface and make experiences real-time and super personalised. No one thought, twenty years back, a smart mobile phone with a dozen arrays of sensors would be affordable under USD 150. We all use AI in our daily life already, almost at no cost. Decentralisation has its advantages. As no organisation or any few large entities can own or control the data or its flow in the Web3 ecosystem, it is more democratised, and by design, it would be inclusive,” said Biswajit. Suresh is also optimistic about India’s Web3 play.“I believe we have the smarts to make that happen, and perhaps it will take much less time than what we took to put Web2 in people’s hands in India. Web3 will be about data decentralisation, perhaps a more transparent and secure environment due to underlying technologies like digital ledgers, and a promise of a fairer internet with the power to transition from Internet platform giants to individuals,” he said. Challenges The carbon footprint of blockchain is huge. To put things in perspective, one Ethereum transaction equals 74,000 Visa transactions and an average NFT transaction emits 48 kg of CO2. According to Digiconomist, Ethereum’s network usage of 99.6 Terawatt-hours of electricity per year trumps the annual energy consumption of countries like the Philippines or Belgium. Now, major cryptocurrencies are moving to Proof of Stake to reduce CO2 emissions. India is set to update its 2030 climate targets under the Paris Agreement. To that end, the country should take into account the huge carbon footprints of blockchain technology while forming policies. Data sovereignty is another major challenge. While Web3 is not limited by geography given its decentralised nature, how will country-specific policies play out? We will have to wait and see.","excerpt":"Web3 will contribute USD 1.1 trillion to India’s GDP.","categories":["IT Services"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-02-22T15:00:00","publication_year":"2022","word_count":1517,"keywords":["Go","API","AI-first","AI","ML","Git","RAG","Ray","GAN","R"],"extracted_tech_keywords":["AI","ML","Ray","RAG","R","Go","Git","API","GAN","AI-first"],"url":"https:\/\/analyticsindiamag.com\/it-services\/aim-long-reads-why-cant-india-afford-to-miss-the-web3-boat\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10053024,"title":"Data Scientists Can Now Use RStudio With Amazon SageMaker","content":"AWS has collaborated with RStudio PBC to announce the general availability of RStudio on Amazon SageMaker — a fully managed RStudio Workbench IDE in the cloud. RStudio is an open-source data science company that makes free & open-source tools for R & Python and enterprise software for teams to develop and share their work at scale, headquartered in Boston. While RStudio provides many different ways to support an organisation’s cloud strategy, many customers also use Amazon SageMaker. So, they wanted an easier way to combine RStudio’s professional products with SageMaker’s rich machine learning and deep learning capabilities and incorporate RStudio into their data science infrastructure on SageMaker. “RStudio is excited to collaborate with the Amazon SageMaker team on this release as they make it easier for organisations to move their open-source data science workloads to the cloud. We are committed to helping our joint customers use our commercial offerings to bring their production workloads to Amazon’s SageMaker and to further collaborations with the Amazon SageMaker team,” said Tareef Kawaf, President, RStudio PBC. With RStudio on Amazon SageMaker, administrators can have a simple experience to migrate their RStudio environments to integrate into Amazon SageMaker and bring existing RStudio licenses to manage through AWS License Manager. In addition, they can onboard both R and Python developers to the same Amazon SageMaker domain using AWS Single Sign-On (SSO) or AWS Identity and Access Management (IAM) and take it as a centralised place to configure both RStudio and Amazon SageMaker Studio. So, data scientists have a freedom of choice between programming languages and coding interfaces to switch between RStudio and Amazon SageMaker Studio notebooks. Furthermore, all of their work, including code, datasets, repositories, and other artefacts, are synchronised between the two environments through the underlying Amazon EFS storage. One can bring their current RStudio license to easily migrate self-managed RStudio environments to Amazon SageMaker in just a few simple steps. One can find it here.","excerpt":"Data scientists have a freedom of choice between programming languages and coding interfaces to switch between RStudio and Amazon SageMaker Studio notebooks.","categories":["AI News"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-11-08T11:45:58","publication_year":"2021","word_count":321,"keywords":["data science","Amazon SageMaker","machine learning","AWS","AI","RAG","Python","deep learning","GAN","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","Amazon SageMaker","RAG","AWS","Python","R","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/data-scientists-can-now-use-rstudio-with-amazon-sagemaker\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23939,"title":"Hospitality, BFSI, Retail Will Benefit The Most From Chatbots, Says Ramco’s Ramesh Sivasubramanian","content":"Ramesh Sivasubramanian, who heads the Innovation Lab at Ramco Systems says that chatbots are transforming the workplace. He is also of the opinion that all businesses need to go hand in hand with the technology in order to be in harmony with the requirements of their customers. He is currently setting up an MRO Innovation Lab for Air France Industries KLM in Singapore, co-funded by Singapore Government, and supported by Ramco. It is home to a number of disruptive aviation IT solutions such as Mixed Reality-based applications leveraging HoloLens for ground engineers, predictive maintenance tools, chatbots and blockchain among others. Sivasubramanian has extensive IT innovation and delivery leadership experience in acquiring and managing global customers.  He has created and managed large innovation and co-innovation labs which delivered Solution Accelerators for pharmaceutical, aviation and government sectors with co-founding from the Singapore government. In 2015, he also co-founded a techno-commercial startup, Perfect Attire Pte Ltd and was also responsible for marketing and branding activities of the startup. Analytics India Magazine interacted with Sivasubramanian, where he shared his insights about how chatbots can make jobs in several sectors more easier, and whether enterprises are ready to embrace the chatbot economy. Analytics India Magazine: Tell tell us about Ramco and the various products and services it offers in cloud and chatbots space? Ramesh Sivasubramanian: Ramco Systems is a cloud enterprise software company. We bet on new technology and bring innovative features to the market. With three core products for enterprise resource planning, human capital management and aviation, Ramco has also carved a niche in the logistics, enterprise asset management and professional services space. While traditional ERP goliaths are playing the catch-up game, we pride in being amongst the pioneers in bringing bots to enterprises. Bot is our sling that is getting the digital and user experience-savvy ERP buyer to us. Ramco has been investing in exploring the benefits of bringing chatbots into an enterprise. While it is widely known that bots can automate repetitive tasks, we strive to embed adequate intelligence in them to empower them. Ramco bots have been built to offer suggestions based on past experiences. With a perfect blend of cognitive capabilities and robotic process automation, Ramco’s bots don the advisory hat efficiently to make them the perfect companion for mundane tasks. We have developed bots that can carry out transactions for the HR, Aviation and Logistics industry. The bots are added on to our existing enterprise applications to make them future-ready. AIM: How does Ramco ChiaBot work? When did Ramco Systems come up with an idea of adopting chatbots? RS: Ramco Chia is a chat-based interface that helps HR managers and employees with their self-service activities related to time-off, expenses, travel, pay, and calendar. The user just enters their email address and password, and begins a perfectly natural conversation about their query, while Chia responds with relevant information. With focus on innovation and culture, Ramco believes that technology should simplify businesses, not complicate them. The future is all about simplification – right from user experience to pricing. Simplification in user interface-led to Ramco launching its chatbot. Chia has been built on Microsoft framework — LUIS AI, which provides the bot with continuous engineering support and Ramco focuses on building contextual content around it. Today, we have fully functional Bots in HR, Logistics, Aviation, and ERP. AIM: What are the various industries where chatbots serve? How has been the success story so far? RS: To date, chatbots have largely been limited to consumer applications. There are more than 50,000 known bots in the market — most of which are consumer centric — and a small number of these cater to enterprise needs. But a number of factors are driving the trend toward chatbot as a viable business tool enhancing interactions at the enterprise, supplier and consumer levels. People-centric businesses will benefit the most from chatbots – such as hospitality, banking and financial sector, service businesses, retail, health and so on. Many companies are still in the experimental stages in terms of developing chatbots. Ramco Systems is one of the few with a solution that is ready to deliver tangible benefits for organisations and was recently mentioned in Gartner’s report on The Future of Workforce Management as one of the only two chatbots available for workforce management requirements. AIM: How will the chatbots affect the critical areas of recruitment, employee engagement and talent management? RS: In the area of recruitment, chatbots could play a huge role in comforting the candidate at the time of onboarding. Onboarding formalities can be handled by the bot. This includes preparing the candidate to ease into the team he or she would be put into. The job details, team members’ info, supervisor’s info can be communicated well in advance making the candidate comfortable and relaxed at the time of joining. Training\/ tutorial videos, CEO messages\/ videos can be communicated via bot. Also, in the case of an interviewer, bots can help in knowing the day’s interview schedule, can display link to profiles or resumes of candidates for preparations prior to the interview. Bots can also help the interviewer in recording the assessment scores details and take a decision. Going the extra mile, stages could be automated by feeding the questions to the bot and getting the answers or responses from the candidates. This could be an excellent mode to screen candidates using a preset prescreening questionnaire. When it comes to employee engagement, related pulse surveys could be floated via bots. The happiness quotient can be tracked by triggering mood sense through bots. In case of talent management, employees can check recommended training sessions with schedules. With respect to performance management systems, day to day journal entries can be recorded via bot. Managers can check and compare appraisal info of the previous years. AIM: How are chatbots transforming the workplace? Are they an answer to a more creative workplace? RS: Yes, chatbots are definitely transforming the workplace. With the amount of millennial employees increasing year on, the need to reach them on the device\/ platform they use has become imperative. Expecting the new age employees to fill expense claim on a paper or desktop is passé. Hence, the HR of a company needs to work towards their employee’s priorities, enhancing work experience. To focus on employee engagement & experience, HR needs to divert their attention from the various routine activities like Leave\/Expense\/Payroll Management, which now can be taken over by chatbots. Bots help companies give their employees and customers a more engaging experience. AIM: How will Ramco ChiaBot change the workplace productivity? RS: Ramco ChiaBot has been designed to manage repetitive tasks such as address queries around leave balance, loss of pay details, timesheet, expense filing, approvals and so on. This means there are minimal input and effort required at the user’s end. The technology is at the point where there is enough built-in intelligence to enable chatbots to offer more complex suggestions to users as well as to learn additional responses based on past experiences, improving productivity. AIM: Are businesses in India ready for the bot economy? RS: According to a recent survey, 80 percent of businesses will be using bots by 2020. All businesses need to be agile, they need to go hand in hand with the technology in order to be in harmony with the requirements of their customers. Most of the businesses today are making their services user-friendly for both Gen Y and Gen X customers. Thus, this is where a chatbot comes handy for all their communicational aids. Many consumers have already been exposed to chatbots, through applications such as hotel reservations and other travel booking and are comfortable with the technology. These same consumers — are growing a number of other users — and are likely going to be today’s customers and employees who will be open to bots helping them in their workplace too. Recent years have seen a blurring of lines between consumer technology and enterprise IT, and chatbots are speeding up the trend. Advances in machine learning and artificial intelligence will help enhance the technology further in the coming years. This will increase the usage by businesses to ensure they remain customer obsessed. AIM:  How does ChiaBot help in improving employee experience? How do chatbots affect employee communication? RS: Ramco Chia helps in all the daily activities of an employee. From leave\/ travel requests, expense reimbursements to modifying daily calendar activities, all can be done with the help of ChiaBot. It provides a human-like, personalized communication with the employees, based on their previous transactions. AIM: Are HR departments in India ready to embrace AI and chatbots at a workplace? RS: In the HR industry, with immense data available in hand, organizations still continue to grapple with the challenge of effectively utilizing the quantifiable data. A dire need for a simple, automated, and AI-powered solution which can facilitate real-time feedback collection that can aid in retention, employee growth, career enhancement, is always on top of the wishlist. The advent of bots in the HR industry has been adding tremendous value. With in-built Artificial Intelligence and Machine Learning capabilities, bots help managers get valuable insights into employee’s performance without spending months in going over each and every feedback message that is collected.","excerpt":"Ramesh Sivasubramanian, who heads the Innovation Lab at Ramco Systems says that chatbots are transforming the workplace. He is also of the opinion that all businesses need to go hand in hand with the technology in order to be in harmony with the requirements of their customers. He is currently setting up an MRO Innovation […]","categories":["AI Features"],"tags":["AI Chatbot","BFSI","blockchain tutorial","chatbot ai","Interviews and Discussions","retail"],"author_name":"Smita Sinha","publish_date":"2018-04-23T10:24:17","publication_year":"2018","word_count":1541,"keywords":["Go","artificial intelligence","machine learning","BFSI","AI Chatbot","AI","chatbots","Git","RAG","blockchain tutorial","Aim","chatbot ai","analytics","retail","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","RAG","chatbots","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hospitality-bfsi-retail-will-benefit-the-most-from-chatbots-says-ramcos-ramesh-sivasubramanian\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":24543,"title":"Analytics vs Artificial Intelligence – How Analytics Can Accelerate AI Adoption","content":"AI is the buzzword du jour for enterprises, but many organizations are still struggling with digital transformation to become data-driven, so how can they approach this new challenge? AI has become more popular, thanks to the increased data volumes, advanced algorithms, and improvements in computing power and storage. Meanwhile, businesses are embedding AI applications in their business portfolio, but they still face a set of common challenges such as integration, cost, security, privacy and even regulatory challenges. The question is — can analytics play a role in accelerating the onboarding of AI in enterprises. Enterprises that have successfully deployed analytics are twice as likely to get senior management buy-in for AI adoption. Analytics is part of the evolution that can lead to successful AI system. Case in point, machine learning models are trained on huge datasets. In analytics-aware organization, that deal with data discovery, big data and tasks such as data wrangling, data preparation and integration, AI is a natural progression. Here’s why AI is a straightforward transition for analytics-aware organizations that have a mature model. and a mature analytics system will underpin the success for Artificial Intelligence. Research indicates that global tech leaders that are most successful at adopting AI techniques incorporate a data policy into their core business functionality. This is manifested in the form of APIs and interfaces. An enterprise-wide policy on data standards can greatly help in streamlining analytics and machine learning practice. Maintaining a data policy can help in clearly identifying the stakeholders and monitoring the enterprise wide access, thereby reducing employee confusion. Another key area to consider is that AI systems mature over a period of time as they are fed more data and the right, quality data. That’s why organizations invest in data storage and data warehouse, this is part of the process of aligning assets for implementing AI. According to Darian Shirazi, Co-founder and CEO of Radius predicts, “Quality data is a must for quality AI predictions. Over the years, we will see more companies focusing on solving the challenge of maintaining accurate data, so that AI can live up to the promise of driving change for businesses.” In fact, our 2017 study State of analytics in domestic firms in India 2017  by Analytics India Magazine and Cartesian Consulting stated that companies across various domains have extensively adopted analytics since the last year. It represented the extent to which analytics has penetrated in the domestic market, demonstrating that in every 59 employees in an Indian organizations, one was associated with data and analytics function. A key finding of the report was the negative correlation between maturity and penetration, stating how an analytics penetration doesn’t demonstrate a high maturity in analytics function. For example, while e-commerce firms have the highest analytics penetration of all sectors, they have the lowest maturity. In India Flipkart is the only company which is both high on penetration and maturity. Here’s How Analytics Can Pave The Way For AI Adoption: Organizations that have a deep BI stack including capabilities for data storage, analytics, governance, the ability to manage structured and unstructured data, visualization tools and techniques have the tools in place for building an effective AI strategy According to a report by Fractal Analytics, making an investment in big data architecture is crucial to success especially for combining structured and unstructured data that sits alongside legacy structured data sources like CRM and ERP systems. Hence, investing in big data solutions should be considered part of the overall strategy for strenegthing the “BI stack” of technology right from ingestion, storage, discovery, modeling, analytics\/ML, and visualization. According to the report, on top of this architecture, organizations need to find out the tools required to enable data exploration and visualization by the business and end-users. By building an enterprise-wide business management system, companies can create a robust big data platform for not only descriptive analytics and reporting, but also find a way to implement predictive analytics solutions, ML, and AI at scale. An enterprise wide BI platform can accelerate AI adoption by enabling deployment of best practices, algorithms, and solutions. In the context of AI, an organization’s deep analytics capabilities can help organizations leverage ML and AI more effectively.","excerpt":"AI is the buzzword du jour for enterprises, but many organizations are still struggling with digital transformation to become data-driven, so how can they approach this new challenge? AI has become more popular, thanks to the increased data volumes, advanced algorithms, and improvements in computing power and storage. Meanwhile, businesses are embedding AI applications in […]","categories":["Deep Tech"],"tags":["AI adoption","enterprise AI adoption"],"author_name":"Richa Bhatia","publish_date":"2018-05-11T10:47:38","publication_year":"2018","word_count":698,"keywords":["Go","machine learning","artificial intelligence","AI adoption","AI","R","ML","Git","RAG","analytics","enterprise AI adoption","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","predictive analytics","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/analytics-vs-artificial-intelligence-how-analytics-can-accelerate-ai-adoption\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":12702,"title":"Aureus Analytics launches CRUX; streamlines customer experience for Insurers","content":"As the insurance industry draws towards an impending evolution, insurers are realizing the need to deliver superior customer experience to stay ahead of the competition. Moreover, aspects revolving around data explosion and changing consumer preferences like instant gratific≤ation, sharing economy, etc. are making it a taxing job for insurers. Aureus Analytics, customer Intelligence & Experience company for insurers and banks have launched the CRUX platform to address this particular business challenge. A first of its kind comprehensive Data Analytics platform, CRUX has been designed specifically for insurers to deliver better customer experience. Anurag Shah, CEO and Co-Founder, Aureus Analytics, remarks “The solution is disruptive, cost-effective, and swift. This platform will help insurers currently perplexed with the decision of building expensive in house big data analytics capabilities, or service providers who are thrive on never ending projects.” With CRUX, business and analytics users can capture data from multiple sources, both internal and external to the insurer, to create a single source of truth. Moreover, a variety of predictive algorithms could be run, which helps in predicting critical consumer behavior. This in turn helps in improving persistency, cross sell, fraud detection, etc. in real time. CRUX stands at an intersection of technologies such as machine learning, Big Data, and Advanced Analytics. The platform combines the power of all these technologies together to deliver real-time actionable insights in simple business English at the Point of Decision. Any business user can access insights anytime, and on any device, as CRUX has been designed for all device form factors. CRUX will help insurers to: Reduce claims processing and payout time from weeks to days. Improve their persistency \/ retention rates, so as to realize significant gains to their bottom line. Enhance fraud detection capabilities Improve cross sell capabilities by comprehending their customers and the preferences better.","excerpt":"As the insurance industry draws towards an impending evolution, insurers are realizing the need to deliver superior customer experience to stay ahead of the competition. Moreover, aspects revolving around data explosion and changing consumer preferences like instant gratific≤ation, sharing economy, etc. are making it a taxing job for insurers. Aureus Analytics, customer Intelligence & Experience […]","categories":["AI News"],"tags":["insurance analytics"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-02-12T05:49:19","publication_year":"2017","word_count":301,"keywords":["big data","Go","machine learning","AWS","AI","insurance analytics","RAG","Aim","analytics","R","fraud detection"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","fraud detection","AWS","R","Go","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aureus-analytics-introduces-crux-streamlines-customer-experience-insurers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10023171,"title":"Nuro’s Plan With The Recent Funding","content":"Last week, US-based electric self-driving vehicle manufacturer Nuro raised funding from Woven Capital, Chipotle and others in a Series C round. Founded in 2016 by former Google engineers, Nuro is a driverless on-road robotic car maker that delivers products and packages to consumers. The terms of the deal have not been disclosed. Last year, Nuro had raised $500 million from T. Rowe Price Associates, Fidelity Management and Research Co. and Baillie Gifford in the previous round. The company has also received a combined investment of $940 million from SoftBank and Greylock Capital Management. The galaxy of big-ticket investors in Nuro is a testament to the company’s brimming potential. Paving the way for Robotaxis Nuro integrates robotics technology to create vehicles that deliver goods. The company launched its first self-driving service in 2018. Being one of the few startups capitalising on self-driving technology, Nuro aims to save time and effort spent on the daily grind. Nuro’s compact and efficient self-driving vehicles are perfect for retailers, restaurants and grocery stores for on-demand deliveries. Nuro’s crewless autonomous vehicles R1 and R2 can deliver goods such as groceries, pizza etc, in a quick, affordable and safe manner. The company has struck partnerships with Domino’s, Walmart. Kroger, CVS and more. Nuro is producing vehicles that can be engineered and built from the ground up without installing safety systems that are otherwise mandatory for passenger vehicles. The company aims to solve the last-mile delivery for both consumers and businesses using its advanced technology. Exponential Growth Last year, Nuro witnessed tremendous growth – a rarity, in and itself, for self-driving vehicles. The company recently got permission to operate in two counties in California to deliver packages. The four-year-old startup is the first, and so far, the only company with a deployment permit from the California Department of Motor Vehicles for delivery services without human supervision. Apart from Nuro, Wymo has the license to operate driverless vehicles without human supervision. Nuro acquired self-driving truck startup Ike, in December 2020. With shared ambitions of revolutionising the logistics industry, Nuro can leverage Ike’s tech stack to build its local delivery application for potential use cases. “The combination of the two, along with the ability to readily integrate some of their core technology (like their world-class virtual simulation tool), is what made the opportunity to have them join us a natural fit,” said Dave Ferguson, co-founder of Nuro, in a Medium post. 2020 has also been a milestone year for Nuro. In February, R2 became the first self-driving vehicle to obtain a USDOT exemption; in April, Nuro became the second autonomous vehicle company to receive a driverless testing permit in California; and in October Nuro announced R2 vehicles were being tested in three different states — with no drivers, no occupants, no chase cars. Way Forward Nuro strives to accelerate the benefits of robotics for everyday life. With a valuation of $5 billion and 600 employees, the startup is in it for the long haul. The recent funding signals continued confidence in Nuro’s vision for autonomous delivery. Nuro has upended the traditional ways of delivery with a view to revolutionising the logistics sector. The startup has been using the fresh capital to build runways in various cities and scale across markets. Nuro is currently testing and operating R2 on public roads in Arizona, California and Texas. The company has also been working on advancing its autonomous technology, hiring new employees and expanding its delivery services. “We are thrilled to have the backing of this tremendous group of best-in-class investors to shape the future of mobility for people across the globe,” said Dave Ferguson, co-founder and president of Nuro, in an official statement.","excerpt":"Last week, US-based electric self-driving vehicle manufacturer Nuro raised funding from Woven Capital, Chipotle and others in a Series C round. Founded in 2016 by former Google engineers, Nuro is a driverless on-road robotic car maker that delivers products and packages to consumers. The terms of the deal have not been disclosed.  Last year, Nuro […]","categories":["AI Features"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2021-04-01T14:00:00","publication_year":"2021","word_count":611,"keywords":["Go","API","funding","programming_languages:R","AI","IPO","RAG","Aim","R","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","startup","funding","IPO","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/nuros-plan-with-the-recent-funding\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":36682,"title":"5 Data Visualisation Books Every Beginner Must Refer To","content":"Across all industries, the implementation of BI tools has becoming increasingly important. One of the reasons for a data analyst to be the most successful is his\/her knowledge of exceptional data visualisation skills. In this article, we list down 5 major books on data visualisation that should be on your reading list to gain insights on BI. 1| The Visual Display Of Quantitative Information By Edward R.Tufte About: This is a classic book of all time which gives readers a clear introduction to the basic and core visualisation theories of data graphics. This book caters the knowledge of statistical graphics, charts, tables, and other relevant topics. The author also illustrated 250 best and the worst statistical graphics with a detailed analysis of how to display data for precise, effective and quick analysis which will help an enthusiast to gain insights more clearly. Cost: You can buy both the hardcover and paperback version of this book from Amazon which will cost you ₹ 3,374.00 and ₹ 4,245.00 respectively. Click here to buy the book from Amazon. 2| Visualize This: The Flowing Data Guide To Design, Visualization, And Statistics By Nathan Yau About: In this book, the author shows how to visualise data in such a way that could maximise its potential and tell a story in a brief and clear manner. The book describes the step-by-step process to visualise and show data. You will understand how to gather, parse as well as format data and then design high-quality graphics with the data. You will also learn which tools are the most suitable and can be used to visualise data-native graphics for the Web such as ActionScript, Flash Libraries, etc. Cost: You can buy both the Kindle version and the paperback version of this book from Amazon which will cost you ₹ 530.41 and ₹ 1,728.65 respectively Click here to buy the book from Amazon 3| The Functional Art By Alberto Cairo About: This book by data journalist, Alberto Cairo shows the methodologies one can use in order to produce meaningful work by using modern visualisation tools. The author shows how to transform numbers into graphical shapes and makes the reader understand the stories behind those numbers. The book reveals why data visualisation should be thought of functional art rather than fine art, how to use color, type and other tools to make visualisation more effective, best practices for creating interactive information graphics and other relevant topics. Cost: You can buy both the Kindle version and the paperback version of this book from Amazon which will cost you ₹ 751.80 and ₹ 2,843.00 respectively Click here to buy the book from Amazon 4| Information Dashboard Design: Displaying Data For At-A-Glance Monitoring By Stephen Few About: In this book, the author exposes the common issues in designing a dashboard and describes the best practices in detail with a number of instances. The book helps you understand how to design a dashboard as well as make you gain insights on the concepts rooted in brain science. Furthermore, this book focuses on fundamental considerations, in-depth instructions in the design of bullet graphs, sparklines and other critical steps to follow during the design process. Cost: You can buy the hardcover version of this book from Amazon which will cost you ₹ 2,539.00. Click here to buy the book from Amazon 5| The Accidental Analyst: Show Your Data Who’s Boss By Eileen And Stephen McDaniel About: Whether you are an expert data analyst or a beginner, this book will guide you to use data for analysis. The book shows how to take control of the data as well as analysis to answer the queries of business and decision-making. The author describes the framework as seven C’s of data analysis such as Choose your questions, Collect your data, Check out your data, Clean up your data, Chart your analysis, Customise your analysis and Communicate your results. Cost: You can buy the paperback version of this book from Amazon which will cost you ₹ 2,268.00. Click here to buy the book from Amazon","excerpt":"Across all industries, the implementation of BI tools has becoming increasingly important. One of the reasons for a data analyst to be the most successful is his\/her knowledge of exceptional data visualisation skills. In this article, we list down 5 major books on data visualisation that should be on your reading list to gain insights […]","categories":["AI Trends"],"tags":["books","tools"],"author_name":"Ambika Choudhury","publish_date":"2019-03-21T07:27:48","publication_year":"2019","word_count":674,"keywords":["programming_languages:R","AI","data_tools:Spark","BERT","llm_models:BERT","tools","books","R"],"extracted_tech_keywords":["AI","R","BERT","llm_models:BERT","programming_languages:R","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-data-visualisation-books-every-beginner-must-refer-to\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10129404,"title":"Microsoft CTO Kevin Scott Joins Shopify’s Board of Directors","content":"Kevin Scott, CTO at Microsoft, has been appointed to the Board of Directors of Shopify, the leading e-commerce platform. Scott announced the news on LinkedIn, where he served as Senior Vice President of Engineering and Operations for six years. He has been with Microsoft for eight years now. In a conversation with Canada-based e-commerce giant Shopify, co-founder and CEO Tobias Lütke, Scott expressed his enthusiasm for joining the company’s board. “I’m very excited to be able to have the opportunity to work more closely with you all and with all of the folks at Shopify,” said Scott. This appointment brings together two tech giants, with Scott’s extensive experience in AI and cloud computing at Microsoft potentially benefiting Shopify’s e-commerce innovations. Scott’s background includes leadership roles at Google and AdMob, as well as his current position driving strategic, cross-company initiatives at Microsoft. Scott’s move to Shopify’s board comes at a time when the e-commerce sector is rapidly evolving, with increasing focus on AI integration and enhanced user experiences. His expertise in large-scale technology infrastructure and AI could prove invaluable as Shopify continues to expand its services and global reach. AI in Shopify Shopify has been increasingly investing in bringing AI capabilities on their platform with their latest release being an AI “sidekick” chatbot. The sidekick serves as a support chatbot for merchants, assisting with tasks such as creating discount codes, generating store reports, and proposing blog post ideas. “I think the role that Shopify plays on the internet is giving all of these wonderful entrepreneurs and business owners and makers and creators of all stripes a way to actually commercialise their work to be able to have a customer base and to deal with them in beautiful ways. It’s just, sort of extraordinary and very important,” said Scott. Interestingly, Bret Taylor, the current board of director at OpenAI has also served as a board member of Shopify.","excerpt":"Kevin Scott, CTO at Microsoft, has been appointed to the Board of Directors of Shopify, the leading e-commerce platform. Scott announced the news on LinkedIn, where he served as Senior Vice President of Engineering and Operations for six years. He has been with Microsoft for eight years now. In a conversation with Canada-based e-commerce giant […]","categories":["AI News"],"tags":["CTO","Microsoft","Shopify"],"author_name":"Vandana Nair","publish_date":"2024-07-17T15:44:16","publication_year":"2024","word_count":316,"keywords":["Go","API","OpenAI","AI","CTO","cloud computing","innovation","Shopify","programming_languages:R","programming_languages:Go","R","Microsoft"],"extracted_tech_keywords":["AI","OpenAI","cloud computing","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-cto-kevin-scott-joins-shopifys-board-of-directors\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058319,"title":"Former Jadavpur University students develop AI model to detect fake news","content":"The scourge of fake news is one of the biggest challenges plaguing the modern world. The outbreak of COVID-19 has further compounded the problem. Taking the cue, three former students of Jadavpur University, Kolkata, Sourya Dipta Das (currently working at SHL India as a Research Engineer), Ayan Basak (working as a data scientist at Snapdeal), and Saikat Dutta (working at LG Ads Solutions as data scientist) have developed an AI model for fake news detection with a high level of accuracy. Recently, their work has been published in the Neurocomputing journal. Analytics India Magazine got in touch with the trio to understand the nuts and bolts of their AI model. Model architecture The model consists of seven main parts: Text Preprocessing TokenisationBackbone Model ArchitecturesEnsembleStatistical Feature Fusion NetworkPredictive Uncertainty Estimation ModelHeuristic Post Processing Figure 1: Fake News Identification Initial Process Block Diagram Text Preprocessing A major chunk of social media items, like tweets, are written in colloquial language and contain information like usernames, URLs, emojis, etc. The team filtered out such attributes from the given data as a basic preprocessing step, before feeding it into the ensemble model. Tokenisation During tokenisation, each sentence is broken down into tokens before being fed into a model. The team has used a variety of tokenization approaches depending on the pre-trained model used as each model expects tokens to be structured in a particular manner, including the presence of model-specific special tokens. Each model has its corresponding vocabulary associated with its tokeniser, trained on a large corpus of data. During training, each model applies the tokenisation technique with its corresponding vocabulary on news data. Backbone Model Architectures The team has used a variety of pre-trained language models as backbone models for text classification. “For each model, an additional fully connected layer is added to its respective encoder sub-network to obtain prediction probabilities for each class- ‘real’ and ‘fake’ as a prediction vector. Pre-trained weights for each model are fine-tuned using the tokenized training data. The same tokeniser is used to tokenise the test data and the fine-tuned model checkpoint is used to obtain predictions during inference,” Dutta said. Ensemble The team used the model prediction vectors obtained from inference on the news titles for the different models to obtain their final classification result, i.e. “real” or “fake”. The reason behind using an ensemble of various fine-tuned pretrained language models is to utilise knowledge extracted by the respective models from the corresponding dataset it is trained on. However, in the case of FakeNewsNet dataset, they obtained an additional prediction vector using NewsBERT on the news body, which  is also appended to the existing feature set. All the features used here are obtained from the raw text data. To balance an individual model’s limitations, an ensemble method can be useful for a collection of similarly well-performing models. Statistical Feature Fusion Network “Our basic intuition behind using statistical features is that meta-attributes like username handles, URL domains, news source, news author, etc. are very important aspects of a news item and they can convey reliable information regarding the genuineness of such items. We have tried to incorporate the effect of these attributes along with our original ensemble model predictions,” Dutta said. They calculated probability values corresponding to each of the attributes (say probability of a username handle or URL domain indicating a fake news item) and added them to our feature set. Additionally,  the team used information about the frequency of each class for each of these attributes in the training set to compute these probability values and realised that Soft-voting works better than Hard-voting. Hence, the post-processing step takes Soft-voting prediction vectors into account. Predictive Uncertainty Estimation Model The team has designed an approximate Bayesian neural network as a Statistical Feature Fusion Network (SFFN) for uncertainty estimation of fake news classification. The team applied a Monte Carlo Dropout (MCDropout) layer between hidden layers of the feature fusion network for Bayesian interpretation. The Monte Carlo (MC) dropout is applied both during training and inference. Hence, the model does not produce the same output each time inference is done on the same data point. The MC dropout enabled them to make random predictions that can be interpreted as samples from a probability distribution. From this model, they got the prediction vector along with its uncertainty value. Heuristic Post-Processing Here, they augmented the original framework with a heuristic approach taking into account the effect of the statistical attributes. This approach works well for data with attributes like URL domains, username handles, news source, etc. For texts that lack these attributes, they relied on ensemble model predictions. These attributes allowed them to add meaningful features to the current feature set. They obtained new training, validation and test feature-sets obtained using class-wise probability vectors from ensemble model outputs as well as probability values obtained using statistical attributes from the training data. “We use a novel heuristic algorithm on this resulting feature set to obtain our final class predictions. The intuition behind using a heuristic approach taking the statistical features into account is that if a particular feature can by itself be a strong predictor for a particular class, and that particular class is predicted whenever the value of a feature is greater than a particular threshold, a significant number of incorrect predictions obtained using the previous steps can be ‘corrected’ back,” said Dutta. Figure 2: Fake News Identification Post Process Block Diagram Challenges Basak said language models have played a crucial role  in developing the model. Statistical concepts like approximate Bayesian Inference to perform uncertainty estimation for fake news items have also been incorporated. The system has been built on the Python and TensorFlow 2.0 and scikit-learn have been used to develop the model. “We have developed hand-engineered statistical features using attributes like author name, URL domain, etc, and it was quite challenging to devise a strategy to fuse these features with the model predictions and make a prediction. The uncertainty estimation in the case of fake news classification was also very difficult. Finally, ensuring that our framework is robust and unaffected by variations in the type of news items was a very big challenge,” said Basak. To tackle the data imbalance issue, the team used data augmentation using oversampling technique (KMeans-SMOTE). They used the Statistical Feature Fusion Network (SFFN) sub-model to combine the hand-engineered statistical features with the model predictions “We have also used transfer learning using a number of different pre-trained language models trained on a large data corpus. This ensured our final model is robust and also has a diverse knowledge base from different sources,” said Das.","excerpt":"We use a novel heuristic algorithm on this resulting feature set to obtain our final class predictions.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Data Scientists","Deep Learning","Interviews and Discussions","Machine Learning","models","Python"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-01-13T16:00:00","publication_year":"2022","word_count":1098,"keywords":["Go","scikit-learn","text classification","TPU","AI","neural network","models","Machine Learning","Python","analytics","Deep Learning","TensorFlow","R","Data Scientists","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","neural network","analytics","TensorFlow","scikit-learn","text classification","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/former-jadavpur-university-students-develop-ai-model-to-detect-fake-news\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":51282,"title":"Who Wrote ‘Henry VIII’? This ML Algorithm Has Cast A Shadow On Shakespeare’s Contribution","content":"It is a breakthrough to be able to use artificial intelligence and machine learning to find, identify and confirm the author of a text, and analyse the style of writing from a text collection. In the literary sector, the question of authorship has always been the main concern. A perfect case in point is the analysis of William Shakespeare and John Fletcher’s work Henry VIII. Czech researcher Petr Plechac recently released a paper titled “Relative contributions of Shakespeare and Fletcher in Henry VIII”, confirming scholar James Spedding’s longstanding theory of having more than one formal author for Shakespeare’s Henry VIII. The Tech Behind Plechac developed a machine learning system that determined which portions of the historical play were written by which author. He trained an algorithm on the works of Shakespeare and Fletcher to recognise the style, rhythmic patterns, and word choices and performed a rolling regression technique that breaks down the play into parts and conducts regressions over and over again with subsamples, to study the styles of writing. The process provided granular evidence of the involvement of Fletcher in writing the last four scenes of the play, and it is definitely satisfying to give credence to a lingering debate. The key concept behind author identification is the process of feature engineering, where the machine selects features from the collected text that suitably describes the style of an individual author which helps in distinguishing the author from other writers. The most common features used by the best of the machines are: Frequency of n-gramUsage of function wordsDistribution of word lengthsFrequency of a digitSyntax (which helps in analysing the specific token and the style within a text)Punctuation marksUsage of passive and active voiceParts of speech The major development in feature engineering has deep roots in stylometry, which is the study of the linguistic style and analysing the variations in the literation of different texts. Researchers believe typical individual human activities carry invariant similarities and slightly vary from one person to another. Similarly, the style in writing is usually distinguished by the repeated choices of words or text patterns that the writer tends to make subconsciously. These repeated choices are very individualistic and are supposed to reflect a writer’s style. As it paved its way into the digital world, the technique goes beyond analysing features and focus more on network modelling and advanced mining of data and texts. It also uses information like graphics, emotions, colours, and layouts to provide the required information. The method of network modelling utilises information related to the attribution of the document, such as the venue and date of the published document, and also the type of publication, which is then paired up with some document text-based keywords. However, this method does not focus on studying the important rhythmic patterns and therefore applied with machine learning techniques to find more author characteristics. It also considers a complex relational structure in the usage of function words by constructing word adjacency networks (WANs) with function word nodes and edges containing information regarding the use of two function words within a certain distance from one another. Each WAN is them being interpreted as a Markov chain that assigns transition probabilities to the appearance of two function words in succession. Thus, these probabilities stand uniquely to the authors’ expression. As a science, it falls under the general category of recognition systems which are usually applied to identify suspects or criminals. Applications The need to identify the author of a particular text and verify the authenticity of the same has been around for several years. Researchers and linguists have been using the manual technique of stylometry to identify distinct language patterns to help identify the authors in proposed cases. However, the process got more accurate with the involvement of ML and AI. In 2017, an AI application was released for public use, known as Emma, which can read a text and identify the style of the author. CEO Aleksandr Marchenko said in an interview, “To run the check, one needs to upload a text of at least 5,000 words by one author, which is then analysed by Emma to learn the author’s writing style. It can then determine whether all the subsequent texts uploaded.” He further explained that the application combines natural language processing (NLP) and ML with the techniques of stylometry to extract patterns from the author’s text, majorly of which can not be easily detected by the human eye. “More than 50 mathematical parameters stand behind every author’s writing identity. So the issue is of a striking complexity: it’s extremely difficult to define and assess style features of a vexing number of authors, and to implement the extracted knowledge into an NLP technology.” Apart from verifying authorship, and providing an insight into the mental state of the author, stylometry has many potential applications in areas of education and literature, digital content forensics, program code author, crime prevention, law enforcement. Plagiarism detection, ghostwriter detection, has also been benefitted by the technique of machine learning in stylometry.","excerpt":"It is a breakthrough to be able to use artificial intelligence and machine learning to find, identify and confirm the author of a text, and analyse the style of writing from a text collection. In the literary sector, the question of authorship has always been the main concern. A perfect case in point is the […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Machine Learning"],"author_name":"Sejuti Das","publish_date":"2019-12-07T14:00:15","publication_year":"2019","word_count":836,"keywords":["Go","artificial intelligence","machine learning","AI","ML","Machine Learning","Git","feature engineering","NLP","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","R","Go","Git","feature engineering","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/who-wrote-henry-viii-this-ml-algorithm-has-cast-a-shadow-on-shakespeares-contribution\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10099455,"title":"Friend or Fraud?","content":"When Faridabad resident Karan received a call from one of his friends who had just met with an accident asking him to transfer Rs 30,000 for treatment, he had little reasons to raise a doubt. The man calling Karan sounded exactly like his friend and said he was using someone else’s phone as his phone got damaged in the accident. Karan frantically transferred the money. Later, when he contacted his friend, he realised that he had been a victim of a fraud AI voice call. He filed a complaint with the NIT Cyber police station, which revealed that a fraudster used an AI voice impersonator to fake his friend’s voice and duped him of his money.  Such cases are rampant across the country. Criminals have exploited deepfake technology to deceive individuals through fake calls and videos. For instance, a man in Kerala fell victim to a deepfake call from a friend claiming a medical emergency, resulting in a loss of INR 40,000. The growing accessibility and advancement of AI have once again transformed the nature of cyber crimes. This new industry seems to be mushrooming fast, going from strength to strength with WormGPT and FraudGPT to now impersonation scams. A McAfee survey found that 25% of adults worldwide have fallen prey to AI voice scams of some kind. India tops the list with an astounding number of incidents at 47%, followed by the United States at 14%, and the UK at 8%. These frauds are carried out by extracting your voice samples from social media websites like Instagram, Facebook, Twitter etc. As little as 3 seconds of your voice can be used to clone it using the voice-cloning technology. Big Techs’ GenAI Poses New Challenges Microsoft recently introduced a groundbreaking text-to-speech AI model called VALL-E. In a paper published this month, the company unveiled that VALL-E can replicate a person’s voice using just a brief 3-second recording. Impressively, preliminary findings indicate that VALL-E can even capture and reproduce the emotional nuances of the speaker. VALL-E is trained on a dataset comprising 60,000 hours of English speech data. This dataset is asserted to be “hundreds of times larger than existing systems”, and is significantly better than the existing models in the realm of AI-driven voice synthesis. So, a three-second recording of your voice, paired with something like Eleven Labs’ multilingual v2—a foundational AI Model that can be used for nearly 30 languages, definitely spells trouble. Additionally, Meta’s SeamlessM4T is capable of translation into 100 languages. While naturally, some are excited about the doors that these AI tools could open in marketing, customer service, e-learning and entertainment, others are wary of what it could entail—an industry of AI-enabled criminals using it for all kinds of crimes—a new coming of Jamtara? Cybercriminals are also using cloning tools like HeyGen, Murf, Resemble AI, Lyrebird, and ReadSpeaker to create perfect voice clones. To add to it, these tools are inexpensive, costing as little as $0.6. These easily accessible and cheap AI voice generators are enabled by numerous tutorials available online. The ease of access to Generative AI models has allowed individuals with limited technical knowledge to carry out tasks once beyond their capabilities. The tutorials make it easy for inexperienced and tech-oblivious individuals with ill intent to carry out scams at scale. Diamond Cut Diamond While these scamsters are using AI-enabled voice generators, law enforcement is also wielding similar weaponry against them. The cyber police have been using AI tools to monitor SIM cards that are engaged in such scams and recently blocked upwards of 14k SIMs in Haryana’s Mewat district. The Indian Department of Telecommunications is also employing an AI-based facial recognition tool called ASTR to combat fraudulent SIM card use. It encodes human faces in subscriber images using convolutional neural networks to account for various factors like face angle and image quality. ASTR conducts face comparisons, grouping similar faces, and identifies identical faces with at least 97.5% accuracy. ASTR is capable of detecting all SIMs associated with a suspected face in less than 10 seconds from a database of one crore images. Additionally, it employs “fuzzy logic” to find approximate matches for subscriber names, accommodating typographical errors. The tool helps identify individuals with multiple connections or SIMs obtained under different names using the same photograph. The list is also shared with banks, payment wallets, and social media platforms to disconnect these numbers. WhatsApp collaborated with the government to disable fraudulent accounts, with ongoing efforts across other social media platforms. Meanwhile, it’s also crucial to stay alert and adopt proactive measures at your end. Users can verify the caller’s identity, employ codewords or pose a question that only their friend would answer correctly, to safeguard themselves if they’re ever in a sticky situation like Karan.","excerpt":"With the recent rise of AI-enabled voice scams, can you tell if the person on the other end of that phone call is a friend or a fraud?","categories":["AI Features"],"tags":["AI Tool","DeepFake","fraud","Generative AI models"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-09-05T10:00:00","publication_year":"2023","word_count":791,"keywords":["Replicate","Go","GenAI","AI","neural network","ML","fraud","GPT","Aim","Generative AI models","generative AI","AI Tool","DeepFake","R"],"extracted_tech_keywords":["AI","ML","neural network","generative AI","GenAI","Aim","R","Go","GPT","Replicate"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/friend-or-fraud\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10169857,"title":"Phonepe Faces UPI Outages After Disaster Recovery Drill Overloads New Data Centre","content":"Phonepe, India’s digital payment platform, faced a significant UPI outage on Monday around 7:30 PM, due to a network capacity shortfall at the newly deployed data centre. The incident occurred due to recovery drills initiated by Phonepe in response to escalating geopolitical tension. All traffic was routed through the new data centre as part of heightened cybersecurity measures. However, the infrastructure could not handle a high volume of transactions, leading to widespread failure, according to the company. Rahul Chari, the co-founder and CTO of PhonePe, on X said, “Given the escalation of conflict last week, we at PhonePe initiated active disaster recovery drills with heightened cybersecurity measures on our network firewall. As part of this, 100% of our traffic was routed through a new data centre.” “Unfortunately, the Monday evening peak traffic exposed a network capacity shortfall due to which transactions started failing.” PhonePe apologised to the user, saying that they have now rebalanced their traffic across our sites and are seeing a recovery. They assured users that they will further strengthen the systems. Vijay Shekhar Sharma, Founder of Paytm, took to the platform X to share that his payment platform is running smoothly and without errors. He said, “Just so you know, our UPI payments are working smoothly. The Paytm app is up and running at twice the regular speed.”","excerpt":"PhonePe initiated active disaster recovery drills with heightened cybersecurity measures on their network firewall.","categories":["AI News"],"tags":["Phonepe","UPI"],"author_name":"Amisha Arya","publish_date":"2025-05-13T15:57:18","publication_year":"2025","word_count":221,"keywords":["programming_languages:R","AI","Phonepe","Scala","Git","UPI","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","R","Scala","Git","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/phonepe-faces-upi-outages-after-disaster-recovery-drill-overloads-new-data-centre\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10098281,"title":"OpenAI Funds NYU Ethical Journalism Project with $395,000 Grant","content":"The Sam Altman-run OpenAI, has decided to fund a new journalism ethics initiative at New York University ‘s Arthur L. Carter Journalism Institute with a $395,000 grant. The announcements is a part of a broader effort by OpenAI to be associated with journalism on which the company replies on to train its infamous GPT-like AI models. The initiative will be led by Stephen Adler, former EIC of Reuters who stated, “The initiative will provide workshops and discussions on existing and emerging journalism ethics issues.” In terms of collecting clean data OpenAI seems to be a step ahead of its competitors like Google, one can decipher from its recent partnerships with organisations like Associated Press (AP), one of the biggest US news agencies, and the $5 million deal with American Journalism Project. The partnership with AP is said to explore ways to develop AI to support local news and in the process OpenAI will indirectly tie up with 41 news agencies that AJP supports. The funding will also support the creation of a new product studio within AJP that will support local news outlets as they experiment with OpenAI’s technology, stated Sarabeth Berman, CEO of AJP. Even though the company has been ‘trying’ to tackle the complexity of ethical journalism amid the generative AI revolution, OpenAI has been extremely cagey about where the company got the data it used to train its latest GPT model. While the big tech companies have almost never taken data privacy of the users seriously, initiative supporting ethical journalism by one of the technology leaders is rare. Interestingly, the New York Times has recently reported about Google’s presentation of Genesis, a project aiming to responsibly generate news copy from factual information. Executives from media outlets like The Times, The Washington Post, and News Corp were a part of this demonstration. Impressions varied, as some “said it seemed to take for granted the effort that went into producing accurate and artful news stories,” while others likened the technology to a personal assistant. Interestingly, Google discreetly updated its privacy policy, revealing its practice of mining public web data to enhance AI services like Bard and Cloud. With OpenAI’s recent contributions and advancements, the landscape appears promising. However, saying “Don’t believe everything you read on the internet” holds more weight due to the duality of these tech developments. Read more: Why Google Is Killing Itself","excerpt":"The announcements is a part of a broader effort by OpenAI to be associated with journalism on which the company replies on to train its infamous GPT-like AI models","categories":["AI News"],"tags":["Ethical AI","OPenAI GPT","Responsible AI"],"author_name":"Tasmia Ansari","publish_date":"2023-08-09T00:27:05","publication_year":"2023","word_count":396,"keywords":["Go","funding","OpenAI","AI","ETL","Ethical AI","Responsible AI","GPT","Aim","OPenAI GPT","generative AI","GAN","R"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Aim","R","Go","ETL","GPT","GAN","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-funds-nyu-ethical-journalism-project-with-395000-grant\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004220,"title":"How I Created My Own Data For Object Detection and Segmentation","content":"In simple terms, computer vision enables our computer to process and visualize the data. It is a very complicated task to make the computer learn from the image data. From the day we are born, we are training our brain continuously with examples, so after a certain point of time we are able to recognize objects. Similarly we need to train our computers by feeding the data. Until a few years ago, computer vision only worked with limited capacity. But now, with the availability of larger datasets and hardware, it has grown exponentially. Through this article, we will demonstrate how to create our own image dataset from a video recording. This labelled data set can be used in the popular computer vision problems such as object detection, image segmentation and image classification. What will you learn in this article Creating your own dataset.Introduction to annotation tool.Preparing Segmentation dataset.Preparing object detection dataset. Creating our own dataset Let’s take an example where an autonomous vehicle collects the data. Firstly we fix the camera to the vehicle and we record the video while the vehicle is moving and we get a video file. As we know video is the combination of multiple frames, by writing a few lines of code in python we can divide the video file into frames. In the below code snippet, we will convert a video file into frames. We have taken a random whatsapp video in this task. import cv2 import numpy as np import os cap = cv2.VideoCapture('\/content\/WhatsApp Video 2020-07-28 at 9.02.25 AM.mp4') try: if not os.path.exists('data'): os.makedirs('data') except OSError: print ('Unable to directory ') currentFrame = 0 while(True): ret, frame = cap.read() name = '.\/data\/frame' + str(currentFrame) + '.jpg' print ('Creating...' + name) cv2.imwrite(name, frame) currentFrame += 1 cap.release() cv2.destroyAllWindows() Output As we can see in the above output screenshot, the corresponding image files are generated. Introduction to Data Annotation tool In the Data annotation tool, we will label the objects in the image. We mostly use VGG image annotator for annotations which is an open-source tool that can be used to draw the bounding boxes in the image and add textual information for the objects in the image. Using these labeled data we can train our deep learning model. After opening the VGG image annotator tool, we need to add our images, through add files or by Adding URL(path of images). Then we need to add the list of objects we need to annotate and we can use the same list of objects for both object detection and segmentation tasks as shown in the below image. Preparing Object Detection dataset For object detection data, we need to draw the bounding box on the object and we need to assign the textual information to the object. In the left top of the VGG image annotator tool, we can see the column named region shape, here we need to select the rectangle shape for creating the object detection bounding box as shown in the above fig. As you can see in the above image, we labeled the image by drawing the bounding box region of the person and the region of the bike. After drawing these regions, we can download the data in  CSV format, JSON format, or COCO format. By sending the raw images and any downloaded format, we will be able to train our deep learning models. Preparing Segmentation dataset To create a segmentation dataset, we need to label the data considering each pixel, we need to draw to the exact shape of the object, and then we need to label it similar to object detection. In the region shape, we use a polyline for labeling segmentation data because using a rectangle bounding box we can’t draw bounding boxes in considering each pixel. As you can see in the above image, we segmented the person using a polyline. After drawing these regions, we can download the data in either CSV format, JSON format, or COCO format. By sending the raw images and any downloaded format, we will be able to train our deep learning models. As you can see from above fig, in the top left we can see annotation column by clicking on export option we can download our annotated data Conclusion In the above demonstration, we clearly explained how to generate our own dataset for training our deep learning models. There are tons of data around us but there is a very little amount of labelled data. We demonstrated an easy way to create our own labelled image dataset to train a deep learning model in the task of object detection or image classification.","excerpt":"In simple terms, computer vision enables our computer to process and visualize the data. It is a very complicated task to make the computer learn from the image data. From the day we are born, we are training our brain continuously with examples, so after a certain point of time we are able to recognize […]","categories":["Deep Tech"],"tags":["Computer Vision","data annotation","Deep Learning"],"author_name":"Prudhvi varma","publish_date":"2020-08-05T11:00:00","publication_year":"2020","word_count":771,"keywords":["Go","NumPy","TPU","AI","computer vision","Python","deep learning","data annotation","object detection","Computer Vision","programming_languages:Python","Deep Learning","R"],"extracted_tech_keywords":["AI","deep learning","computer vision","NumPy","object detection","TPU","Python","R","Go","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-i-created-my-own-data-for-object-detection-and-segmentation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10094707,"title":"Byju’s Launches Suit of AI Models For Teaching, But Won’t Replace Teachers","content":"India edtech company BYJU’S, has launched BYJU’S WIZ, a suite of three AI models for hyper-personalised learning. The suite includes BADRI, MathGPT, and TeacherGPT. BADRI is a predictive AI model that uses personalised ‘forgetting curves’ to identify a student’s strengths and weaknesses in detail. MathGPT employs advanced machine learning algorithms to provide accurate solutions for complex math challenges, including trigonometric proofs. TeacherGPT is an AI-driven assistant that offers personalised guidance and promotes independent problem-solving skills. BYJU’S claims its models have an accuracy rate of nearly 90% and will redefine learning. The AI suite is built on existing models and leverages BYJU’S data for training, resulting in personalised and effective learning experiences. The integration of these AI models across BYJU’S learning platforms is a move away from a one-size-fits-all. It is a significant step forward in providing tailored learning experiences. The suite predicts a student’s knowledge state, identifies misconceptions and learning gaps, and offers personalised learning experiences. This enhances engagement and effectiveness for students while optimising internal systems such as teacher audits. The AI suite is developed by a team of researchers, data scientists, and educational experts at BYJU’S Labs. They claim that they have ensured the models are safe, reliable, and curriculum-aligned. The models are currently undergoing rigorous testing and are set to be integrated across BYJU’S entire product portfolio, catering to students of all age groups. Move Not Intended to Replace Teachers Divya Gokulnath, co-founder of BYJU’s, clarified that the incorporation of AI technology in education is not intended to replace teachers, but rather to enhance the efficiency of the organisation and allow teachers to focus on other important tasks. Gokulnath emphasises that AI technology will provide teachers with more precise feedback on students’ performance, enabling them to better assist students in improving their learning abilities. Gokulnath firmly states that no AI can replicate the role of teachers in video lessons, live classes, or tuition centers. However, AI can empower teachers by providing valuable insights and feedback on student progress. “No AI could replace what we did as teachers do in that video. No AI can replace what we as teachers do in live classes and BYJU’s tuition centres but AI can enable us as teachers to have clearer feedback on our students’ performance so that we can help them become better learners,” Gokulnath said. Gokulnath said that it is not a matter of technology versus teachers, but rather a collaboration where technology can support and enable teachers to become more effective educators. The presence of teachers in the classroom, with their human touch, remains irreplaceable and crucial for the holistic development of students.","excerpt":"BYJU’S claims its models have an accuracy rate of nearly 90% and will redefine learning.","categories":["AI News"],"tags":["edtech"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-06-07T23:26:00","publication_year":"2023","word_count":435,"keywords":["Replicate","Go","machine learning","AI","ML","edtech","RAG","GPT","Aim","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","RAG","R","Go","GPT","GAN","Replicate"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/byjus-launches-suit-of-ai-models-for-teaching-but-wont-replace-teachers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10139625,"title":"Veteran Google DeepMind Researcher Joins Anthropic After a Decade","content":"Julian Schrittwieser, a prominent AI researcher with a decade of experience at Google DeepMind, has joined Anthropic this week. Schrittwieser, known for his work on groundbreaking projects like AlphaGo, AlphaZero, and MuZero, announced his move to Anthropic in a recent post on X expressing excitement about joining the AI research lab. “I’m excited to announce that I’m joining Anthropic, starting this week!” Schrittwieser said. “I’m honored to be joining such a capable and kind group of people, who I’ve long admired.” Schrittwieser praised Anthropic’s AI offerings, including Claude, which he found particularly useful, and recent advances such as Artifacts and Computer Use. He acknowledged his fruitful journey at DeepMind, where he contributed to projects spanning both foundational research, such as AlphaCode and AlphaTensor, and recent developments like Gemini and AlphaProof. Reflecting on his time at DeepMind, he said, “I’ve been incredibly lucky to be part of an amazing journey at Google DeepMind for the past 10 years… I got to work on more exciting projects than I would have ever dreamed of.” With his move to Anthropic, Schrittwieser joins a growing roster of top AI talent to advance the capabilities of large language models and artificial intelligence technologies. Anthropic continues to attract influential AI minds from leading institutions. Durk Kingma, a founding member of OpenAI and former researcher at Google Brain, joined the team in October 2024. John Schulman, who co-founded OpenAI, joined Anthropic in August 2024, shortly after Jan Leike, OpenAI’s former Super Alignment co-lead, signed on in May 2024.","excerpt":"Schrittwieser is known for his work on groundbreaking projects like AlphaGo, AlphaZero, and MuZero.","categories":["AI News"],"tags":["Anthropic","Google Deepmind"],"author_name":"Siddharth Jindal","publish_date":"2024-10-29T10:45:01","publication_year":"2024","word_count":251,"keywords":["Anthropic","Go","artificial intelligence","OpenAI","AI","llm_models:Claude","Google Deepmind","llm_models:Gemini","Julia","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","OpenAI","Anthropic","R","Go","Julia","AI research","llm_models:Claude","llm_models:Gemini"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/veteran-google-deepmind-researcher-joins-anthropic-after-a-decade\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23211,"title":"What Do Companies Like IBM, Infosys, Wipro Look For In A Data Scientist?","content":"In this data dominated era, it has become crucial to employ analysts who can turn all the raw data into actionable insight for a company. From Fortune 500 companies to a startup company, every company have realised that utilising the huge amount of datasets can improve a business’s bottom line. Hence, data science is on the rise, both from a company’s and employee’s perspective. According to an IBM report, data scientists and advanced analysts are the fastest-growing roles, which are projected to see demand spike by 28 percent by 2020. Most of the companies seek data scientist with a very particular experience and skills to run and complete data science projects. Here we take a look at what few tech companies like IBM, Infosys and Wipro are looking for in a data scientist to make their team more smarter. What Infosys Is Looking For In A Data Scientist According to a job post on LinkedIn, this is what Infosys is looking for in a Data Scientist or what they call it ‘Data Science Modelers’, to optimize the business through analytics and statistical modeling and to develop, maintain reporting and analytical tools. The data scientist should have a PhD or Master’s degree in Statistics, Mathematics, Engineering, Commerce, Science Domain experience in Retail or Finance or Supply Chain or\/and Marketing will be an added bonus. Hands on experience in Statistical Techniques such as Decision Tree, Segmentation, Logistic and Multiple Regression and others. Infosys also prefers to have a candidate who has exposure to US or overseas markets with a Domain\/ Technical\/ Tools knowledge. Proficiency with statistical analysis tools. For example, SAS, SPSS, R Hands on experience in Statistical Techniques such as Decision Tree, Segmentation, Logistic and Multiple Regress, among others Strong Excel, Access and PowerPoint skills Familiarity with database query tools (SQL) Basic understanding of a data warehouse architecture Experience of Campaign Management and Customer Segmentation What IBM Is Looking For In A Data Scientist To run and complete a data science projects in IBM this is what the Armonk-based computer manufacturing company is looking for in a data scientist, according to IBM’s Seth Dobrin and Jean-Francois Puget. The company prefers a candidate with a Masters of Science or Doctorate degree. Because having an advanced degree will help take a complex project and break it down into a set of testable hypotheses. Expertise in a programming language like R, Python, Scala, Julia. The candidate must have the ability to transform and manage large data sets. The data scientist must have an ability to communicate with non-data scientists. She\/he must have good storytelling skills and have the ability to map mathematical concepts to common sense. Expertise in machine learning and statistics, with an emphasis on decision optimization What Wipro Is Looking For In A Data Scientist The primary focus of Wipro’s Data Scientist is to apply data mining techniques, doing statistical analysis and building high-quality prediction systems. The data scientist in Wipro will be working extensively in data-centric digital storytelling and data visualization. According to Wipro, these are the skillsets and qualification the data scientist must have: The data scientist must have an excellent understanding of machine learning techniques and algorithms, such as k-Nearest Neighbors, Naive, Bayes, Support Vector Machine (SVM), Decision Forests, etc. She\/he should have an experience with common data science toolkits, such as R, Weka, NumPy, MatLab, etc. The data scientist should have an experience in data visualization tools, such as D3.js, GGplot, etc. She\/he should be proficient in using query language such as SQL, Hive, Pig Experience with NoSQL databases, such as MongoDB, Cassandra, HBase Good applied statistics skills (distribution, statistical testing, regression, etc) and scripting and programming skills The data scientist must have good communication skills and should have data-oriented personality with solid client-facing consultative skills.","excerpt":"In this data dominated era, it has become crucial to employ analysts who can turn all the raw data into actionable insight for a company. From Fortune 500 companies to a startup company, every company have realised that utilising the huge amount of datasets can improve a business’s bottom line. Hence, data science is on […]","categories":["IT Services"],"tags":["Data Science","Data Scientist","IBM","Infosys","Wipro"],"author_name":"Smita Sinha","publish_date":"2018-04-03T09:45:03","publication_year":"2018","word_count":628,"keywords":["Wipro","data science","NumPy","machine learning","Go","Infosys","AI","MongoDB","Python","IBM","analytics","SQL","Data Science","Data Scientist","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","NumPy","MongoDB","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/what-do-companies-like-ibm-infosys-wipro-look-for-in-a-data-scientist\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10080438,"title":"AWS’ New Data Center in Hyderabad to Create 48,000 Jobs","content":"After Mumbai, AWS announced Hyderabad as its second infrastructure region in India. The AWS Asia Pacific (Hyderabad) Region will consist of three availability zones, offering customers an in-country infrastructure to meet data residency and regulatory requirements. The new facility will open its doors for over 48,000 full-time employment opportunities annually with a proposed investment of $4.4 billion (INR 36,300 crore) in India by 2030. Read: Accenture and AWS to Offer Free Cloud Computing Course in India Together with the AWS Asia Pacific (Mumbai) Region, AWS Hyderabad will enable customers with more choice and flexibility to store data securely, execute workloads more resiliently, and provide end users with faster service. Therefore, government, education, and non-profit organizations, along with startups and entrepreneurs alike will be able to exercise greater choices for running applications and serving end customers from data centers located in India. These choices are available for customers to assess and adhere to any applicable legal and regulatory requirements, while also having access to advanced AWS technologies that will propel innovation in areas like data analytics, security, machine learning, and artificial intelligence. Read: Making Data Centres In India Sustainable, The AWS Way Rahul Sharma, president – public sector – at AWS India and South Asia, said that the second AWS region in India will strengthen India’s cloud innovation capability, and support government’s Digital India vision in areas of e-governance, healthcare, agriculture, education, space-tech, utilities, and smart infrastructure, while also enabling development of advanced technologies such as high-performance computing and quantum computing. At present, government initiatives like Co-WIN, eSanjeevaniOPD, and DigiLocker are already using AWS to drive innovation and impact at scale.","excerpt":"AWS Asia Pacific (Hyderabad) Region will consist of three Availability Zones","categories":["AI News"],"tags":["Cloud Computing","cloud infrastructure","Digital India"],"author_name":"Ayush Jain","publish_date":"2022-11-22T17:21:03","publication_year":"2022","word_count":270,"keywords":["Go","cloud infrastructure","Digital India","artificial intelligence","machine learning","AWS","AI","cloud computing","Git","Cloud Computing","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","cloud computing","AWS","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-new-data-center-in-hyderabad-to-create-48000-jobs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60233,"title":"Jupyter Announces The First Public Release Of Jupyter Visual Debugger","content":"One of the most efficient tools for a data scientist, Jupyter plays a potential role in the scientific community when it comes to working on complex scientific computations. All these years, the users of JupyterLab have been requesting to add a visual debugger so that the users need not switch into a different tool for more classical software development. Recently, Jupyter announced the first public release of the Jupyter visual debugger. The JupyterLab debugger is the result of the collaboration and coordination of developers from several institutions, including QuantStack, Two Sigma, and Bloomberg. Using this debugger, one can set breakpoints in notebook cells and source files, inspect variables, navigate the call stack and can do much more. The debugger extension for JupyterLab has been designed to work with any kernel that supports debugging, provides the following a sidebar with a variable explorer, a list of breakpoints, a source preview the possibility to navigate the call stackthe ability to set breakpoints directly next to the code, namely in code cells and code consolesvisual markers to indicate where the current execution has stopped. Installation The debugger front-end can be installed as a JupyterLab extension. In the future release, the debugger front-end will be included in JupyterLab by default.In the back-end, a kernel implementing the Jupyter Debug Protocol will be required. Currently, the only kernel implementing this protocol is xeus-python, which is a new Jupyter kernel for the Python programming language. Future Developments For the future development of this visual debugger, the developers are planning to create support for rich mime type rendering in the variable explorer, support for conditional breakpoints in the UI, general improvements of the debugger user experience and enable the debugging of Voilà dashboards, from the JupyterLab Voilà preview extension.","excerpt":"One of the most efficient tools for a data scientist, Jupyter plays a potential role in the scientific community when it comes to working on complex scientific computations. All these years, the users of JupyterLab have been requesting to add a visual debugger so that the users need not switch into a different tool for […]","categories":["AI News"],"tags":["Jupyter","Jupyter Notebook"],"author_name":"Ambika Choudhury","publish_date":"2020-03-27T16:56:00","publication_year":"2020","word_count":291,"keywords":["Jupyter Notebook","programming_languages:R","Python","programming_languages:Python","Jupyter","R"],"extracted_tech_keywords":["Jupyter","Python","R","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jupyter-announces-the-first-public-release-of-jupyter-visual-debugger\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":37798,"title":"Did Data Analytics Have Anything To Do With BCCI Leaving Out Rishabh Pant From The World Cup 2019 Team?","content":"Image credit: ICC The Board of Control for Cricket in India (BCCI) this week announced the 15-person India team for the upcoming ICC World Cup. But while the list was announced by chairman MSK Prasad and acting secretary of the Indian board Amitabh Chaudhary, the decision, which also had some controversial changes, was also to some extent inspired by data analytics. Rishabh Pant, the critically-acclaimed wicketkeeper and Ambati Rayudu a fiery batsman have been left out of the team by the selectors, which caused many to raise eyebrows. According to a report by a national news wire, CKM Dhananjay, the Indian team’s data analyst, had prepared a special package doing an entire SWOT analysis of the World Cup hopefuls as well as the opposition since the 2017 Champions Trophy. The five-member national selection committee was given a three-and-a-half hour presentation one day before the meeting in order to give them a clear picture of the players’ performance. This presentation also included an in-depth predictive analysis of the weather conditions to be expected in England from May to July 2019. So far, the BCCI selectors have only been looking at raw data such as match scores, runs, strike-rates and wickets, among others. However, this report answered questions such as: How does the English cricket team fare on their home ground? How does New Zealand fare against Indian wrist spinners? How has Australia negotiated our orthodox left-arm spinners during a particular phase of play? What was Kedar Jadhav’s strike-rate during a particular phase of play? What are the specific conditions when the top three [batsmen] have faced trouble against any particular opposition? What are the kind of bowling attacks expected? Who among the Indian players can be successful against specific bowlers? “This time, it was decided that there would be a tweak in the system. On Sunday, a 3.5-hour presentation was provided with processed data. In fact, this could be done before every major selection meeting,” an anonymous source told the news wire. BREAKING: India have named their #CWC19 squad! pic.twitter.com\/mMXt5kAG6Y — ICC Cricket World Cup (@cricketworldcup) April 15, 2019 This is the final squad announced after the discussion: Virat Kohli (captain) Rohit Sharma (vice-captain) Shikhar Dhawan Vijay Shankar KL Rahul Dinesh Karthik Yuzvendra Chahal MS Dhoni (wicketkeeper) Kedar Jadhav Kuldeep Yadav Bhuvneshwar Kumar Jasprit Bumrah Hardik Pandya R Jadeja Mohammed Shami Sports industry, especially cricket, in India is booming at a rapid pace. Not only on the pitch but also in terms of sports-based application. Technologies like artificial intelligence and machine learning are becoming the backbone of those platforms now. While the betting industry is already leveraging ML and AI, the sports bodies like BCCI have only now started to make use of the emerging tech to make critical decisions.","excerpt":"The Board of Control for Cricket in India (BCCI) this week announced the 15-person India team for the upcoming ICC World Cup. But while the list was announced by chairman MSK Prasad and acting secretary of the Indian board Amitabh Chaudhary, the decision, which also had some controversial changes, was also to some extent inspired […]","categories":["AI News"],"tags":["cricket","Data Analytics"],"author_name":"Prajakta Hebbar","publish_date":"2019-04-16T05:57:52","publication_year":"2019","word_count":459,"keywords":["Go","artificial intelligence","machine learning","AI","ML","RAG","Ray","Aim","cricket","analytics","Data Analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","Ray","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/did-data-analytics-have-anything-to-do-with-bcci-leaving-out-rishabh-pant-from-the-world-cup-2019-team\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10163226,"title":"Hugging Face’s New Small Language Model Outperforms Rivals","content":"Small language models (SLMs) are gaining popularity due to their minimal carbon footprint and low computing requirements. The latest to join the bandwagon is SmolLM2 by Hugging Face. Pushing Small Language Models Further SmolLM2 is available under an Apache 2.0 license, making it an open-source alternative. As per the research paper, it is trained on an extensive dataset of ~11 trillion tokens, combining web text with specialised data like math and code. The model utilised a multi-stage training process to rebalance from different data sources to maximise performance. The researchers also expanded on using specialised data, “Additionally, after finding that existing datasets were too small and\/or low-quality, we created the new datasets FineMath, Stack-Edu, and SmolTalk (for mathematics, code, and instruction-following respectively).” They also compared the SLM with other existing state-of-the-art models like Qwen2.5-1.5B and Llama3.2-1B. The evaluation was done using Lighteval, and it outperformed Qwen and Llama. The table above shows that the model was tested with various types of parameters to test all kinds of use cases. Summing up the results, the paper states that the AI model beats Qwen2.5-1.5B by around six percentage points on MMLUPro, proving its capabilities as a useful generative AI model. Additionally, with the math and coding benchmarks, SmolLM2 exhibits competitive performance. It is worth noting that it could not perform better than Qwen2.5-1.5B on a couple of tests, including MATH, but it outperforms Llama3.2-1B on the same. To explain more about the performance, the researchers also shared about a couple of tests not monitored for benchmarks, “SmolLM2 also delivers strong performance on held-out benchmarks not monitored during training, such as MMLU-Pro (Wang et al., 2024c), TriviaQA (Joshi et al., 2017), and Natural Questions (NQ, Kwiatkowski et al., 2019).” Considering the model is open source, Hugging Face has released the datasets and the code used for training to facilitate future research and development on SLMs. It should be exciting to see what the next small language model can do without organisations worrying about resource constraints.","excerpt":"A new challenger enters the small language model race.","categories":["AI News"],"tags":["Hugging Face","small language models"],"author_name":"Ankush Das","publish_date":"2025-02-11T16:43:46","publication_year":"2025","word_count":332,"keywords":["Go","Hugging Face","AI","ML","SLM","generative AI","GAN","llm_models:Llama","small language models","R"],"extracted_tech_keywords":["AI","ML","generative AI","Hugging Face","small language models","SLM","R","Go","GAN","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hugging-faces-new-small-language-model-outperforms-rivals\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10039133,"title":"Behind Sony & NVIDIA’s Latest AI Patents For Video Games","content":"As per a report, the global gaming market is expected to reach a valuation of more than $295 billion by 2026 from $162 billion in 2020 at a CAGR of 10.5%. Game developers strive to enhance gamer’s experience, launching and rewriting codes for diverse console\/platforms. Recently, Sony and NVIDIA have published AI tech patents that can make video games more immersive and interactive. The proxy player Sony’s technology patent introduces a method for gaming. The method includes allocating a default gameplay profile to a user, wherein the default gameplay profile consists of a default gameplay style that simulates human gameplay and wherein the default gameplay profile is configured to control gameplay for the user based on the default gameplay style. In other words, the AI profile begins with a basic set of behaviours and picks up traits based on monitoring the player’s actions. The AI profile will take after the player’s style and stand-in for the player if he\/ he\\she is missing in action or needs assistance. Interestingly, the AI can evolve on its own. It can automatically crawl the internet to look for solutions if it gets stuck while completing a task. For example, the gameplay controller can automatically proceed to complete specific game tasks that are difficult for the user. Also, the gameplay controller is configured to play the gaming application in place of the user (e.g., as initiated through a pause in the game). An AI character is assigned to perform tasks presented. When the user wishes to return to active play, the user can stop the AI character, which allows the user to return to controlling his or her character in the gaming application. In that manner, the user’s gameplay continues without the user’s attention and then resumes after automated gameplay control is terminated. NVIDIA adds a haptic touch The paper’s abstract said: Haptic effects have long been provided to enhance the content, such as by delivering vibrations, rumbles, etc, in a remote controller or other device being used by a user while watching or listening to the content. Haptic effects have either been provided by programming controls for the haptic effects within the content itself or by providing an interface to audio that maps certain haptic effects with certain audio frequencies. The haptic control interface intelligently induces haptic effects for content by using machine learning to detect specific features in content and then induce certain haptic effects for those features. NVIDIA’s AI patent deals with a method to provide a haptic control interface for detecting content features using machine learning to induce haptic effects. The haptic control interface receives a portion of content and processes the piece of content using a machine learning algorithm to detect a feature of the content. Further, the haptic control interface determines one or more haptic effects for the feature. It causes the haptic enabled device to activate the one or more haptic effects determined for the feature. Wrapping up Both patents offer a window to the rapid advances happening in the world of video games. The developers are leveraging AI to radically change the user experience on a scale unimaginable a few years back. The AR\/VR tech, the rise of esports, and pandemic have further accelerated the push.","excerpt":"Recently, Sony and NVIDIA have published AI tech patents that can make video games more immersive and interactive.","categories":["Global Tech"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-04-28T18:00:00","publication_year":"2021","word_count":539,"keywords":["Go","API","machine learning","programming_languages:R","AI","programming_languages:Go","RAG","R"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/behind-sony-nvidias-latest-ai-patents-for-video-games\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":43557,"title":"Looking To Learn Fundamental Data Science Algorithms — MachineHack Launches New Feature Called MachineHack Practise","content":"Data Science is a field which brings computer science, statistics, mathematics and other fields together. But there is also a need for newcomers to cut through the confusion and get their hands dirty in the world of fundamental data science algorithms. Keeping this in mind, MachineHack has launched a feature called MachineHack Practice to give young and beginner-level data scientists access to computing resource with guidance to use basic data science algorithms with the use of popular libraries. Please access MachineHack Practise by clicking here. MachineHack Practise is made up of a series of tutorials which contains: Preliminary information regarding the algorithm on the home page Resources for further guidance on the home page Coding tool with guided usage of the algorithm on a small toy problem Direction to our Hackathon where you can use the knowledge to check your skills Please access MachineHack Practise by clicking here. The feature is designed specifically to transition young and new data scientists to tackle more challenging problems by introducing basic concepts and giving them confidence. Currently, there are 5 tutorials on the following algorithms: MachineHack Practise 1: Linear Regression MachineHack Practise 2: Multiple Linear Regression MachineHack Practise 3: Support Vector Regression MachineHack Practise 4: Decision Tree Regression MachineHack Practise 5: Random Forest Regression Please find Machinehack Practise feature here.","excerpt":"Data Science is a field which brings computer science, statistics, mathematics and other fields together. But there is also a need for newcomers to cut through the confusion and get their hands dirty in the world of fundamental data science algorithms. Keeping this in mind, MachineHack has launched a feature called MachineHack Practice to give […]","categories":["Deep Tech"],"tags":["decision tree algorithm","Hackathon","Learn Data Science","ml hackathon"],"author_name":"Abhijeet Katte","publish_date":"2019-07-30T15:02:56","publication_year":"2019","word_count":217,"keywords":["data science","Go","decision tree algorithm","programming_languages:R","AI","ml hackathon","programming_languages:Go","Hackathon","Learn Data Science","R"],"extracted_tech_keywords":["AI","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/looking-to-learn-fundamental-data-science-algorithms-machinehack-launches-new-feature-called-machinehack-practice\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":46402,"title":"Google Has Claimed Quantum Supremacy — Should We Care?","content":"The last couple of years has seen a tremendous rise in the enhancement of quantum computers. Companies like Google have collaborated with the likes of NASA in order to establish new benchmarks. According to a report from a leading financial portal, the quantum processor took 200 seconds to sample one instance of the quantum circuit one million times, while a supercomputer would require 20,000 years to perform that task. In the paper that was taken down, the researchers said that, to their knowledge, the experiment “marks the first computation that can only be performed on a quantum processor.” This news was followed by claims that now there is no code that cannot be cracked and the implications it has for encryption. This sent ripples across the world of cryptocurrency amongst many others. What’s The Big Deal With Quantum Supremacy Source: quantamagazine Quantum supremacy means that a quantum computer has beaten a classical computer at a certain task. To put it in another perspective, by running a huge cluster of classical cores for (say) a month, you can eventually verify the outputs that your QC produced in a few seconds — while also seeing that the QC was many orders of magnitude faster. According to the computer scientist Scott Aaronson, this does mean that sampling-based quantum supremacy experiments are almost specifically designed for ~50-qubit devices like the ones being built right now. Even with 100 qubits, we wouldn’t know how to verify the results using all the classical computing power available on earth. Future Direction Companies like D-Wave and Google have been trying to develop a sophisticated hybrid system which incorporates quantum effects into the classical ML problems since the turn of this decade and they will continue to do so. With the current fabrication techniques, D-Wave has managed to develop 2000 qubits processors. For every qubit, say x, there are 2^x states. Now, imagine the number of states a 100 qubit processor can handle! Two qubits can store four states, three qubits can store eight states. So, 50 qubits in quantum computers can be approximated to 1 quadrillion bits in classical computers. This alone is the single most important feature and reason behind the arduous efforts to improve quantum computing. The exponentially increasing complex amplitudes can have as much information encoded. Now, this endless chain of information can be thought of as a large matrix representing equations. Machine Learning models run on a system of equations like these. For example, a hyperplane in Support Vector Machines can be written as, wx-b = 0 So with increasing feature set, the complexity of the equation increases or let’s say the size of a matrix increases and complexity is where quantum machine learning comes into the picture. By initiating a quantum state where amplitudes, that give away the probability density of the quantum state vectors can be related to the feature vectors in the dataset used for training the model. So anyone who has had an idea on how to run a standard regression model would comprehend how complex and time consuming it gets with an increasing number of rows\/columns (think millions). The processing power of quantum computers has the potential to unfold a myriad of opportunities. From drug discovery to solving mathematical problems, from machine learning to material sciences, quantum computers can revolutionise many domains. Researchers have been working on machines that will be easier to build, manage, and scale and some computers are now available via the computing cloud. But it could still be many years before quantum computers become a household name.","excerpt":"The last couple of years has seen a tremendous rise in the enhancement of quantum computers. Companies like Google have collaborated with the likes of NASA in order to establish new benchmarks. According to a report from a leading financial portal, the quantum processor took 200 seconds to sample one instance of the quantum circuit […]","categories":["Global Tech"],"tags":["Google","NASA","Quantum Computer","quantum computing companies","quantum supremacy","the quantum companies","what companies use quantum computers"],"author_name":"Ram Sagar","publish_date":"2019-09-26T11:00:59","publication_year":"2019","word_count":592,"keywords":["Go","quantum supremacy","what companies use quantum computers","NASA","machine learning","TPU","AI","programming_languages:R","ML","quantum computing companies","programming_languages:Go","Quantum Computer","the quantum companies","Aim","Google","quantum machine learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","quantum machine learning","TPU","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-has-claimed-quantum-supremacy-should-we-care\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10081950,"title":"Top Players in Photo Noise Reduction Software Market","content":"AI-based noise reduction softwares are built on machine learning algorithms trained on large datasets of images to accurately identify and remove noise. This allows them to produce high-quality, noise-free images without sacrificing image detail. Google recently released its NeRF algorithm that could reduce noise using machine learning algorithms and diffusion methods, but it is yet to be implemented by noise reduction software. Adobe products have been on top of the image manipulation game since the beginning. Most of the new noise-reduction tools are also launched as plugins for Photoshop or Lightroom. Let’s look at some of the top players in the image noise reduction market: Adobe Photoshop & Lightroom Starting with Photoshop, the software initially used to match pixels and remove noise from images which resulted in loss of information and detail from the image. Now with the introduction of AI-based methods, even after reducing noise, a lot of detail has been preserved, enabling high quality of images. Photoshop & Lightroom Camera Raw offers high flexibility and balance between noise and detail. New tools like Topaz Denoise AI and ACR Define can also be integrated as a plugin on the Adobe products, allowing further improvements in the system. Topaz DeNoise AI This powerful tool uses artificial intelligence to accurately identify and remove noise from your photos. You can adjust the noise of images using two basic sliders on the panel and also have a special mode for Low-light and High-ISO images. It also has a number of advanced features, such as the ability to recover lost details and reduce color noise with support for RAW and DNG images. The specific algorithms used by Topaz DeNoise AI are proprietary and not publicly available. Neat Image Pro Neat Image tool uses advanced algorithms to reduce both luminance and chroma noise, resulting in clear, smooth images. It also has a number of customizable settings, allowing you to fine-tune the noise reduction to your specific needs. The pricing of this software is on the lower side, making it preferable for photographers and editors. Additionally, users can also install extensions on Photoshop that offer the same capabilities as the original software. DxO PureRAW2 DxO PureRAW2 is an easy-to-use noise reduction tool that can enhance your image with just two or three clicks. This tool is preferred by individuals who are not looking to add detailed changes in the quality of the image as it does not allow much manual control of the image. Though it lacks input settings, the generated image after processing is highly detailed but also very large in size. Also, PureRAW2 does not have a plugin for Photoshop but it is available for Lightroom. Skylum Luminar Neo An advanced photo editor, Luminar Neo comes with an advanced tool that uses noise reduction algorithms, resulting in clean and clear images. The editing suite offers other tools such as color grading and sharpening, and luminance setting, making it a great all-in-one solution for noise reduction. Photographers and photo editors consider it as a cheaper alternative to Lightroom that offers almost all the capabilities. ON1 NoNoise AI One of the latest tools for reducing noise, ON1 NoNoise AI was initially designed specifically for black and white photographs. The tool can remove both luminance and color noise and is at par with Topaz and PureRAW2. It also has a number of other advanced features, such as the ability to adjust the tonal range of your images and other in-depth image modifying tools, making it an all-round suite for image editing. Corel PaintShop Pro Corel’s all-in-one photo editing software has a powerful AI-based noise reduction tool that can remove both luminance and color noise. The tool enables users to adjust brightness and contrast of low-light images while preserving the details of images. Capture One This professional-grade image editing software has a powerful noise reduction tool and is arguably more powerful than Lightroom in various aspects. Capture One also includes a slider called “single-pixel“ dedicated to hot pixel reduction. It also has a number of other advanced features, such as the ability to selectively apply noise reduction to specific areas of your images. The software is mostly aimed at high-end photographers because of its pricing. Noiseware As the name suggests, the suite of the software is dedicated to noise reduction and it proves to be one of the best at its job. Noiseware has several modes that you can select based on the output you want in your image while also selecting specific points in the image. The contrast and detail preservation ability of Noiseware is unmatched by any other software in the list.","excerpt":"With the AI technology getting better and better each day with GANs and diffusion models, image noise reduction is getting increasingly flawless","categories":["AI Trends"],"tags":["Top Trend"],"author_name":"Mohit Pandey","publish_date":"2022-12-09T14:00:00","publication_year":"2022","word_count":766,"keywords":["Top Trend","Go","machine learning","artificial intelligence","TPU","AI","programming_languages:R","programming_languages:Go","Aim","Chroma","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","Chroma","TPU","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-photo-noise-reduction-softwares-for-2023\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36149,"title":"Webinar: Building Deep Learning Applications For Big Data By Intel","content":"Deep learning, a subset of artificial intelligence is driving immense advancements in a range of disciplines such as self-driving cars, chatbots and even in healthcare. With Deep Learning becoming one of the most sought after fields for tackling business problems, a key constraint is understanding the broad landscape of tools and frameworks available. Users are looking to execute data-heavy workloads for a range of deep learning applications and train complex models, however challenges arise in  getting the right production environments to deliver value from deep learning investments. In this webinar, Building Deep Learning Applications For Big Data by Mukesh Gangadhar, Staff Lead, APJ Part of the Compute Performance & Developer Products (CPDP) who comes with 18 years of industry experience and has worked extensively on optimising software applications on x86 platforms, especially on the cloud and AI side will give the participants an overview of emerging Deep Learning frameworks for Big Data. The webinar will also cover  Analytics Zoo — a unified analytics + AI platform for distributed Tensorflow, Keras and BigDL on Apache Spark. Analytics Zoo, developed by Intel streamlines end-to-end development and deployment and also provides developers with a set of analytics and AI support for the end-to-end pipeline. Key topics covered in the webinar are: With the landscape for tools and frameworks expanding, it can be hard to get started with Deep Learning solutions for big data applications Intel is uniquely positioned to help developers and data scientists in simplifying AI efforts and get started with deep learning solutions with state-of-the art tool which provides end-to-end solution Analytics Zoo comes with built-in deep learning models, such as text classification, recommendation, and object detection and the platform was also recently open-sourced, allowing the wider community to contribute to it This is part of the process at Intel to simplify AI deployments and get started with deep learning solutions with state-of-the art tools from Intel Who Should Attend Data Scientists and Developers looking to apply deep learning technologies to their big data pipelines Data Engineers who are tasked with architecting scalable and stable data pipelines\/infrastructure Product Developers looking for state-of-the solutions for implementing solutions Senior Managers and leaders who want to understand how to apply deep learning to their big data platforms Data science enthusiasts who want to understand the applications of deep learning such as NLP, computer vision and align it with their business objectives About the Speaker Mukesh Gangadhar is the Staff lead, APJ part of the Compute Performance & Developer Products (CPDP). He is a staff architect for optimizing software applications on x86 platforms, especially on the cloud and Artificial Intelligence. He has around 18 years industry experience and holds a degree in Bachelors of Engg in Computer Science & Electronics. When: March 27, 11 am IST Register now, click here","excerpt":"Deep learning, a subset of artificial intelligence is driving immense advancements in a range of disciplines such as self-driving cars, chatbots and even in healthcare. With Deep Learning becoming one of the most sought after fields for tackling business problems, a key constraint is understanding the broad landscape of tools and frameworks available. Users are […]","categories":["Deep Tech"],"tags":["Intel"],"author_name":"Richa Bhatia","publish_date":"2019-03-12T06:54:05","publication_year":"2019","word_count":465,"keywords":["data science","artificial intelligence","Keras","AI","ML","computer vision","NLP","deep learning","analytics","TensorFlow","Intel"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","NLP","computer vision","data science","analytics","TensorFlow","Keras"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/webinar-building-deep-learning-applications-for-big-data-by-intel\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10070649,"title":"Handling imbalanced data with class weights in logistic regression","content":"Logistic Regression is one of the supervised machine learning techniques that are used for classification tasks. Classification datasets most of the time will have a class imbalance with a certain class with more samples and certain classes with a very less number of samples. Using an imbalanced dataset for the model building would account for the wrong prediction and would be more favorable to classes with more samples. So in this article let us try to understand the importance of class weights in logistic regression and why a balance of class weights is necessary to yield a reliable model. Table of Contents What are class weights?Problems associated with imbalanced class weightsUnderstanding the importance of class weightsSteps to computing class weight Summary What are class weights? Class weights are terminology used for classification tasks where each category of the dataset will be provided with certain weights according to the frequency of occurrence of each category. So class weights will be responsible for giving equal weights for all categories on gradient updates. The usage of unbalanced class weights will be responsible for bias towards the most occurring categories in the data. To obtain a more reliable and unbiased classification model it is important to have a uniform distribution of class weights. Uniform distribution of class weights will also yield various parameters like precision, recall, and F1 score as the class weights would be balanced. Now let us try to understand the problems associated with an imbalance in class weights. Are you looking for a complete repository of Python libraries used in data science, check out here. Problems associated with imbalanced class weights The main problem associated with imbalanced class weights is accuracy. The accuracy produced in the world is high but what matters is the rightness of accuracy. For imbalance class weights the accuracy yielded would be high generally as it would be biased towards the most occurring class as it would account for higher class weights. Say suppose healthcare data or business-driven data is in use and there is an imbalance of class weights. So if the class imbalance in the data is not addressed it would account for misinterpretations from the model. Moreover, certain parameters like False-positive and False negatives would turn out to be 0 as the model will be inclined more toward the frequently occurring category. Understanding the importance of class weights To understand the importance of class weights let us consider a classification dataset which is having an imbalance in the target variable distribution. Here a healthcare dataset is being used and let’s analyze the uneven distribution of the target variable through a count plot. sns.countplot(y) Here we can see that the target variable is hugely imbalanced where class 0 is having higher class weights when compared to class 1. So let us build a logistic regression with the imbalance target variable and try to evaluate certain parameters from the model. X=df.drop('stroke',axis=1) y=df['stroke'] from sklearn.model_selection import train_test_split X_train,X_test,Y_train,Y_test=train_test_split(X,y,test_size=0.2,random_state=42) from sklearn.linear_model import LogisticRegression lr_imb=LogisticRegression(random_state=42) lr_imb_model=lr_imb.fit(X_train,Y_train) y_pred=lr_imb_model.predict(X_test) print('Classification report of imbalanced logistic regression \\n',classification_report(Y_test,y_pred)) plot_confusion_matrix(lr_imb_model,X_test,Y_test) From the classification report, we can observe that the harmonic mean (F1-score) is more for the class with more weightage, and for the class with lesser weightage the harmonic mean and other parameters are 0. Moreover, the model has yielded an accuracy score of 94% which means that the model would make wrong predictions and would be more inclined to the frequently occurring classes. So this is the major problem associated with imbalanced class weights and it turns out to balance the class weights to yield an unbiased and more reliable model. Balancing class weights The class weights can be balanced using the logistic regression model by just declaring the class_weight parameter as balanced in the logistic regression model. The class weights can be balanced automatically bypassing the standard parameter as balanced in class weights or random weights to each of the classes can be provided to each of the categories in the data. Now let us look into how to balance the weights using the predefined “balanced parameter” of the scikit learn library. Using the “balanced” parameter for class weights Now the logistic regression model is being fitted with class weights as a standard parameter as “balanced”. The parameter is readily made available in scikit-learn models. Let us see how to use this parameter and obtain a logistic regression model and evaluate certain parameters. lr_bal=LogisticRegression(random_state=42,class_weight='balanced') lr_bal_model=lr_bal.fit(X_train,Y_train) y_pred_bal=lr_bal_model.predict(X_test) print('Classification report for balanced classes \\n',classification_report(Y_test,y_pred_bal)) plot_confusion_matrix(lr_bal_model,X_test,Y_test) Here we can see that after using the class weights as balanced while fitting the logistic regression model we can see that the accuracy has reduced when compared to the imbalance model but the harmonic mean for the lower occurring class has seen an increase. Other parameters like precision, recall, and F1-score have also seen an increase which helps machine learning engineers to suitably make interpretations from the model. This model can be further hyper tuned to yield even better parameters and performance. Moreover, if the confusion matrix is also evaluated we can see that we are able to yield the right parameters like False positive and false negative which shows that the model is not inclined or biased towards the frequently occurring class. Balancing class weights using the dictionary as a parameter Class weights will be balanced using a dictionary where the dictionary keys are the classes of the dataset and the keys of the dictionary would be the percentage of weights that would be assigned to each of the classes of the data. So let us look into how to use a dictionary as a parameter for class weights and evaluate certain parameters of the model. lr_bal2=LogisticRegression(random_state=42,class_weight={0: 0.2,1: 0.8}).fit(X_train,Y_train) y_pred_bal2=lr_bal2.predict(X_test) print(classification_report(Y_test,y_pred_bal2)) plot_confusion_matrix(lr_bal2,X_test,Y_test) So after using a random distribution of weights among the classes in the data we have seen an increase in the model accuracy when compared to using the “balanced” parameter but here a balance of class weights with a percentage of 80 and 20 is shown. So if the classes are balanced equally or near to equal the model will yield performance similar to the model obtained by using the “balanced” parameter. So this is how imbalance and balance of class weights account for the model performance. In general, it is a good practice to use balanced data to yield a reliable model and obtain the right predictions from the model. Now let us try to understand that class weights are being calculated for different class weight parameters. Steps to compute the class weights The class weights for any classification problems can be obtained using standard libraries of scikit-learn. But it is important to understand how scikit-learn internally computes the class weights. The class weights are generally calculated using the formula shown below. w(j)=n\/Kn(j) w(j) = weights of the classes n = number of observations K = Total number of classes n(j) = Number of observations in each class So the scikit-learn utils library internally uses this formula to compute the class weights with different sets of parameters being used for class weights. Now let us look into how to use the scikit- learn utils library for calculating the class weights at different instances. Calculating imbalance class weights For calculating the class weights the “compute_class_weight” inbuilt function is being used as shown below and correspondingly the class weights for minority and majority classes can be computed as shown below. # Calculate weights using sklearn sklearn_weights1 = class_weight.compute_class_weight(class_weight=None,y=df['stroke'],classes=np.unique(y)) sklearn_weights1 Here the weights assigned to both the classes are equal. Let us compute the weights for each of the classes. # Compare the values print(f'The weights for the majority class is {sklearn_weights1[0]*2:.3f}') print(f'The weights for the minority class is {sklearn_weights1[1]*2:.3f}') Here we can see that for imbalanced classes the weights assigned to both minority and majority classes are the same which accounts for the bias of the model toward the majority classes. Calculating class weights after using the “balanced” parameter The class weights can be calculated after using the “balanced” parameter as shown below. sklearn_weights2 = class_weight.compute_class_weight(class_weight='balanced',y=df['stroke'],classes=np.unique(y)) Sklearn_weights2 Here we can see that more weightage is given to class 1 as it has a lesser number of samples when compared to class 0. So let us try to obtain the weights which can be easily interpreted as shown below. # Compare the values print(f'The weights for the majority class is {sklearn_weights2[0]*2:.3f}') print(f'The weights for the minority class is {sklearn_weights2[1]*2:.3f}') So here we can see that the “balanced” parameter has provided more weightage to the minority class when compared to the majority class which helps us to yield a better reliable model. Calculating class weights by using the dictionary as a parameter Here the class weights will be calculated by providing random percentages of distribution of weights to each class of the data as shown below. # Calculate weights using sklearn dict1={0: 0.2,1: 0.8} sklearn_weights3 = class_weight.compute_class_weight(class_weight=dict1,y=df['stroke'],classes=np.unique(y)) sklearn_weights3 Here we can see that as specified 20% of the weights is applied to class 0 and 80% of weightage is applied to class 1 of the dataset. Now let us try to interpret the class weights as shown below for easy understanding. print(f'The weights for the majority class is {sklearn_weights3[0]*2:.3f}') print(f'The weights for the minority class is {sklearn_weights3[1]*2:.3f}') So this is how class weights for different classes in the dataset are enforced by using the weightage that is provided in the dictionary for each of the classes. Summary Class weights play a very important role in any of the classification machine learning models. So in this article, we have seen how class weights and the balance of class weights are important to obtain a reliable model. Class weight balance is very essential to obtain a bias-free model that can be taken up for the right predictions. The imbalance of class weights accounts for faulty predictions and false interpretations from the model. So it is very important to balance the class weights to obtain a reliable model that can be used for predictions in real-time. Reference Scikit Learn Class Weight Official DocumentationColab Notebook","excerpt":"This article discusses the importance of balancing class weights in logistic regression and also the steps to be followed to compute the class weights and the need for a bias free model.","categories":["AI Trends"],"tags":["Imbalance data"],"author_name":"Darshan M","publish_date":"2022-07-10T10:00:00","publication_year":"2022","word_count":1663,"keywords":["data science","scikit-learn","Go","machine learning","programming_languages:R","AI","Colab","Python","programming_languages:Python","Imbalance data","R"],"extracted_tech_keywords":["AI","machine learning","data science","scikit-learn","Colab","Python","R","Go","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/handling-imbalanced-data-with-class-weights-in-logistic-regression\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":2172,"title":"Interview: Ravinder Pal Singh, Director: Strategic & Mega Projects, Digital Transformation at DellEMC","content":"Ravinder Pal Singh is a Senior Industry Consultant and Expert in Digital Transformation and Smart Connected economies. With an experience of over 20 years in the business & technology consulting, Enterprise Architecture and solutions, Ravinder has primarily worked in the areas of Industry vertical solutions (Smart Cities, Real Estate, Healthcare, Education, Media, Manufacturing, etc.), Data center, Collaboration, and network technology Solutions. Ravinder has extensive experience consulting and managing projects with diverse streams of segments like government, manufacturing, R&D, Media, IT & ITES, Telcos, and Retail customers. He has been part of empaneled consultants for India Smart City Mission and Delhi Mumbai Infrastructure Corridor (DMIC). Ravinder has also consulted on Smart City projects in Malaysia, Philippines, Thailand, Korea and Singapore. Besides, he has 2 patent applications submitted in the areas of Data center and Security. Ravinder is a known thought leader on Smart Cities, IoT, and Digital Transformation. He is a regular speaker in national and international forums around Internet of Everything and Smart Communities.  Some of his articles have been published in leading publications.  IoT India Magazine caught up with Ravinder Pal Singh to find out what DellEMC is doing in the IoT space. IoT India Magazine: Talk about Dell’s journey so far in an IoT dedicated landscape. How is the firm contributing uniquely towards this space? What are the technologies that lie at the centre of the firm’s IoT practice? Ravinder Pal Singh: IoT is a wide and large umbrella and Dell Technologies Family is uniquely positioned in this game.  If we look at IoT as a holistic concept, it is based on 6Cs of Data cycle:  Connect, Collect, Collate, Compute, Conserve, and Consume. Dell Technologies family has a clear positioning and proposition for each of these stages:  Connect (Network\/IoT Gateway), Collect (IoT Gateway), Collate (Gateway), Compute (Wide range of Compute), Conserve (Storage), and Consume (End Devices). Besides this, the other key component is Security – Device, Content, and Perimeter. DellEMC has portfolio mapping in each of these areas. IoT India Magazine: Please present a picture of the IoT sector in India. What are the typical challenges noticed in the space? RPS: IoT in India is still at a nascent stage, I will say. Some sectors like utilities, manufacturing, and transportation have adopted and started utilizing the same. However the bigger story for IoT in India is now at execution stage: Smart Cities, AMRUT, Digital India, Power Sector reforms, Railways, etc. In next 5-6 years we will see maturity in adoption and usage of IoT solutions in almost every sphere, though it has its own challenges, viz., low adoption rate. This is largely due to price and complexity which is slowly reaching to a stage now. Also Security of IoT devices and data have been major areas of concern. However, a lot many startups and niche companies in India are doing a great job in this area, and we have started seeing cost optimization in this area. IoT India Magazine: What are some of the most recent technological developments in the IoT space? How is Dell incorporating those technologies into its own ecosystem? RPS: As mentioned earlier, IoT is all about Data – data management, analytics, and consumption. Analytics is a key area where lot of innovation is happening. Data lifecycle is an area where DellEMC has a unique position owing to our leadership in storage, compute, and virtualization. Some examples of innovation in this space is around video analytics which is helping solve Crime and Traffic issues. Real-time, in-motion edge analytics is another area of focus and DellEMC’s IoT Edge platform is uniquely placed here. On top of that, we have created and on boarded a unique ecosystem of solution providers and developers to build solutions around this. IoT India Magazine: Infrastructure, analytics, security, and services are some of your IoT offerings. Please throw light on this statement talking about the features of the offerings. RPS: We have Infrastructure solutions around network, compute, storage, and virtualization; Security solutions for device, content, and perimeter; and Enterprise services like consulting, data integration, management, and infrastructure design. DellEMC’s leadership in core components like storage, compute, and virtualization as well in end consumer devices like PC\/Laptops and our IoT Edge Analytics makes us a partner of choice for all IoT adopters and providers. IoT India Magazine: How do you customize your solutions to meet the needs of different industry verticals? RPS: Domain and vertical alignment is very important in positioning solutions and we do have some good capabilities globally in this space.  Whether it is BFSI, manufacturing, public, smart cities, security, or healthcare, DellEMC has both resources and systems catering to a vertical-specific use cases. IoT India Magazine: Talk about the work culture at Dell. Any piece of advice for all startups and entrepreneurs who want to make it big in the IoT sector? RPS: DellEMC has a very open and encouraging culture. Culture thrives when people really live it. We set standards for how we define ourselves as inventors and business people, and then we go about setting targets for how we achieve these goals. Only one suggestion or advice for startups: It is important to have an innovative product or offering but it is also very important to position it in the target market. IoT India Magazine: What brought you to the IoT landscape? Also talk about your experience in the field. Also present Dell’s roadmap in an IoT landscape. RPS: Fortunately, I got to get engage with this field much earlier. When IoT was not even a word heard in this part of world, I had an opportunity to work on early avatars like M2M (Machine to Machine) in manufacturing area or Smart Buildings where IBMS and SCADA systems used Cross system integration, which laid a foundation for IoT. I have led, managed and deployed large projects in areas like automobile manufacturing, healthcare, e-governance, oil and gas, mining, and smart cities for last 12 years; covering multiple countries like India, Malaysia, South Korea, Japan, Australia, Philipines, Vietnam, Thailand, and Singapore. Dell takes a pragmatic approach to the Internet of Things (IoT) by building on the equipment and data you already have, and leveraging your current technology investments, to quickly and securely enable analytics-driven action. At Dell we: Architect for analytics. Dell’s analytics, data management and infrastructure solutions provide the power to ensure your IoT solution enables analytics where it makes sense – at the edge, in the data center, or in the cloud. Think security first. With award-winning security tools and expertise, and a holistic approach, Dell helps ensure your infrastructure and data remain safe, secure, and private. Provide choice and flexibility. Dell’s broad portfolio of key IoT technologies, combined with a rich partner ecosystem, allows you to build the right IoT ecosystem for your unique needs.","excerpt":"Ravinder Pal Singh is a Senior Industry Consultant and Expert in Digital Transformation and Smart Connected economies. With an experience of over 20 years in the business & technology consulting, Enterprise Architecture and solutions, Ravinder has primarily worked in the areas of Industry vertical solutions (Smart Cities, Real Estate, Healthcare, Education, Media, Manufacturing, etc.), Data […]","categories":["AI Features"],"tags":["Analytics India","data center India","Data Management","Digital Transformation India","healthcare India","Interviews and Discussions","IoT India","manufacturing India","security India","smart cities india","storage"],"author_name":"Дарья","publish_date":"2017-07-12T05:26:21","publication_year":"2017","word_count":1128,"keywords":["smart cities india","Analytics India","Git","R","data center India","IoT India","digital transformation","RAG","storage","healthcare India","analytics","startup","Go","security India","AI","manufacturing India","Digital Transformation India","Data Management","programming_languages:R","innovation","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Git","digital transformation","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-ravinder-pal-singh-director-strategic-mega-projects-digital-transformation-dellemc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10043695,"title":"Tech Behind Spotify’s Spot On Recommendations","content":"Spotify’s Discover Weekly offers compelling recommendations for exploring similar music and new releases tailored to the user’s taste. The on-demand music service uses big data and AI to come up with new ways to understand music and user’s genre preferences. With over 200 million users, Spotify’s USP is its customisation and music knowledge driven by algorithms and not community-created playlists. Till date, the application has over 50 million songs and 4 billion playlists- accumulating tons of data related to song preferences, search behaviour, playlist creation and geographic data. The home screen recommendation is governed by the AI system BaRT- Bandits for Recommendations as Treatment. BaRT provides personalised screens for each user by organising it in rows of playlists, called ‘shelves’ by Spotify, with each shelf having a different theme based on its contents. The home page, filled with these recommendations, reads like ‘Made for You’, ‘Inspired by your recent listening’, ‘Dive right in’ or ‘More of what you like’. BaRT also suggests fresh music to break the playlist loops. The epsilon greedy solution BaRT ranks the songs in shelves using a multi-armed bandit algorithm, epsilon greedy solution- that balances exploitation and exploration. The system exploits the user’s information like their location, listening history, search history, songs skipped, playlists the user has made, their activity on Spotify’s social features etc. This information is further combined with its exploration of artists or genres close to the user’s taste in music, the popularity of other artists and songs the user might like- it uses information from the rest of the world to find a recommendation similar to the user’s preferences. Know your audience Spotify uses three ML algorithms to study the user’s music preferences and behaviour on the app. Collaborative filtering relies on implicit user feedback, streaming counts, data and user visits to artist’s pages. Spotify creates a back end profile based on the user’s personal taste, which is crunched and used by the algorithm into the music landscape of spotify’s 300 million+ songs database- to find music that the user may like but has not streamed yet. These recommendations keep strengthening as the user uses it more, to the point where no two spotify users have the exact same recommendations. Natural Language Processing is used to analyse and classify music by scanning a song’s metadata and its mentions on blogs and shares across the web. It essentially groups artists into clusters based on texts from blog posts and music articles to connect songs and artists. For instance, Spotify scans the music, ranks the songs in accordance with aspects like cultural keywords, the theme of playlists and its popularity; and provides related recommendations. Audio Models analyse raw audio data and recommend non-popular new songs. The algorithm leverages Convolutional Neural Networks that use clustering to identify similarities in time signature, key, mode, tempo and loudness based on its audio waveform. It further analyses the songs tone, pitch, mood to create a sound profile that fits into similar models such as chill out tunes or monsoon playlists. 30-second rule BaRT measures the success of its recommendation based on the first 30 seconds of the user listening to it: If they streamed it for more than 30 seconds, the recommendation is tracked to be correct. The longer one listens to the set of songs, the better the recommendation is. Spotify’s product director, Matthew Ogle, called this the ‘sweet spot’ for understanding the user’s likes and preferences. He compared skipping before the first 30 minutes to a thumbs down for the algorithm. Discover Weekly Discover Weekly is a 30-song playlist with songs similar to the songs that the user has been listening to. Similar to its daily mixes and personalized playlists, this is created by leveraging AI and big data. To enhance this recommendation, the algorithm also analyses the users streaming history and playlists- their current music preference. Spotify is on the move to strengthen its ‘personalized’ experience for the users by their plan to introduce a live audio streaming feature. Spotify recently acquired Locker Room, a live audio app to create conversations around music and culture.","excerpt":"Spotify’s Discover Weekly offers compelling recommendations for exploring similar music and new releases tailored to the user’s taste. The on-demand music service uses big data and AI to come up with new ways to understand music and user’s genre preferences.  With over 200 million users, Spotify’s USP is its customisation and music knowledge driven by […]","categories":["AI Features"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-07-16T12:00:00","publication_year":"2021","word_count":680,"keywords":["big data","Go","programming_languages:R","AI","neural network","ML","RAG","ViT","GAN","R"],"extracted_tech_keywords":["AI","ML","neural network","RAG","R","Go","big data","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/tech-behind-spotifys-spot-on-recommendations\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35985,"title":"How Important Are Hackathons For Your Organisation: A Survey By AIM","content":"The field of information technology is evolving at a faster rate than usual. Areas like artificial intelligence, data science, data analytics and internet of things, among others, are now being integrated with industries across all sectors. Many times, it becomes difficult to keep track of these technologies and keep up. Now, more than ever, we are seeing a pattern where young professionals wanting more challenging jobs and more interesting work, especially in the Emerging Tech sector. And hackathons have turned out to be a very valuable source of answer to both employees as well as employers. This month, Analytics India Magazine decided to find out how employers and big organisations are using this nifty platform called hackathons to get in touch with the talented people who can be absorbed into the company, and people who can solve unique problems. About The Study The samples were collected by asking respondents to fill in a survey created by AIM about what employees, as well as employers, thought about hackathons as a tool for hiring as well collaborate productively for different projects. This included various sub-topics such as growth opportunity in the current organisation, collaborative spirit of hackathons and the rate of success for both the parties. We took opinions from young Indian professionals as well as senior employers to get a thorough idea of the working environment in this growing field. Our survey was met with much enthusiasm — and we got some great insights from it. Some of them were expected, and many of them were real eye-openers. Has your organisation ever conducted a hackathon? How frequently does your company conduct hackathons? Our core demographic for the survey had a very clear answer to the above question. 44% of the respondents said that their organisations conducted hackathons only once a year. Why do you think your organisation conducts hackathons? Interestingly, we found out that a majority or 32% of the organisations use hackathons as a great way to crowdsource solutions for problems. We also found out that 16% of the organisations conducting hackathons use it for hiring. 16% use it for organisational branding both in and out of the campus. How reliable are hackathons for assessing baseline skills in candidates and potential employees? Who do you usually prefer conducting hackathons with? For this question, we found out that most of our respondents, 40%, preferred to collaborate with external agencies to get the most out of their hackathons. What kind of hackathons do you prefer? When asked what kind of format the organisations of the employers preferred to conduct the hackathon, a majority, 56% of them replied with fact that they preferred with the hybrid format, which is a mix of online and offline format. Among the following, what do you find most disagreeable about hackathons? Which one do you think hackathons should focus more on? When asked about the focus of the hackathons the organisations as well the participants responded with: 44% of the people said that developing realistic working prototypes and solutions for the challenges presented was crucial. 28% of the respondents said that hackathons should focus on inspiring participants to consider the “art of the possible,” to think up and imagine futuristic forward-thinking solutions with technologies that are not available or haven’t been created. 28% of the respondents said that they should focus on creating an atmosphere where people with different skill sets can brainstorm together and learn about a problem\/challenge. What challenges have you\/your company faced while conducting a hackathon? 36% of the respondents said that one of the key challenges they faced during a hackathon was about curation of the problem statement and dataset. 24% of the respondents said that awareness of the ongoing event was a huge challenge. Do you want your organisation to conduct a machine learning hackathon? Hackathons have been in vogue for quite some time in the tech world to source innovative ideas. Many employers have also used this medium to brand themselves well with their prospective employees. For the participants, hackathons address one or many of these motivations – prize money, bragging rights, learning opportunities and\/or hires. Quite often the hiring as a motive is understated in many hackathons. So when we finally the crucial question, 84% respondents replied with a resounding yes.","excerpt":"The field of information technology is evolving at a faster rate than usual. Areas like artificial intelligence, data science, data analytics and internet of things, among others, are now being integrated with industries across all sectors. Many times, it becomes difficult to keep track of these technologies and keep up. Now, more than ever, we […]","categories":["AI Features"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-03-07T14:13:25","publication_year":"2019","word_count":711,"keywords":["data science","Go","machine learning","artificial intelligence","AI","RAG","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","RAG","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-important-are-hackathons-for-your-organisation-survey\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10012251,"title":"Gnani.ai Launches New Integrated Speech Solution For The Ministry of Defence","content":"Gnani.ai, a conversational AI startup, has announced the launch of a new integrated speech recognition based solution for the Indian Armed Forces. According to the official release by the company, the solution is an end-to-end voice translation system that has been designed for the Ministry of Defence. The solution leverages automatic speech recognition (ASR), machine translation and speech-to-text in order to convert Mandarin to English. Designed to assist the armed forces of our country, as well as the intelligence agencies and local law enforcement authorities, this solution will improve communication systems and will give an edge to the Indian defence forces. The solution does with a wide range of applications including cross border intelligence, voice surveillance, monitoring telephone\/internet conversations, intercepting radio and satellite communication as well as to bridge interactions during border meetings and joint exercises. Some of its unique features are — noise reduction, dialect\/accent detection and support for all audio file formats stated by the official release. Such vernacular NLP solutions will help the Indian armed forces to strengthen communication systems across its other neighbouring countries like Pakistan, Nepal, Bangladesh, and Sri Lanka. Explaining better, Ananth Nagaraj, the co-founder and CTO of Gnani.ai stated to the news media that this AI-based speech recognition technology has indeed become a necessity and soon going to be a critical part of modern warfare. Gnani.ai believes that similar to other industries, artificial intelligence has the potential to advance the communication systems of Indian Armed forces. Adding on to it, another co-founder and CEO of the startup — Ganesh Gopalan said to the media that Mandarin had been a tough language when it comes to learning its linguistic nuances such as phoneme and dialects. And that’s why Gnani’s integrated speech solution will leverage AI to remove such obstacles and bridge this gap “to retrieve and gather real-time data and intelligence.”","excerpt":"Gnani.ai, a conversational AI startup, has announced the launch of a new integrated speech recognition based solution for the Indian Armed Forces.  According to the official release by the company, the solution is an end-to-end voice translation system that has been designed for the Ministry of Defence. The solution leverages automatic speech recognition (ASR), machine […]","categories":["AI News"],"tags":["automatic speech recognition","Speech Analytics","Speech Recognition","Vernacular Automated Speech Recognition"],"author_name":"Sejuti Das","publish_date":"2020-11-24T13:37:40","publication_year":"2020","word_count":307,"keywords":["Go","artificial intelligence","Speech Analytics","AI","programming_languages:R","Vernacular Automated Speech Recognition","programming_languages:Go","automatic speech recognition","NLP","RAG","Speech Recognition","GAN","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","RAG","R","Go","GAN","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/gnani-ai-launches-new-integrated-speech-solution-for-the-ministry-of-defence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":59982,"title":"US Government Unveils The COVID-19 High-Performance Computing Consortium","content":"The COVID-19 High Performance Computing Consortium Bringing together the Federal government, industry, and academic leaders to provide access to the world’s most powerful high-performance computing resources in support of COVID-19 research. The COVID-19 High Performance Computing Consortium is a unique private-public effort spearheaded by the White House Office of Science and Technology Policy, the U.S. Department of Energy and IBM to bring together federal government, industry, and academic leaders who are volunteering free compute time and resources on their world-class machines. Consortium partners include: IBM Amazon Web Services Google Cloud Microsoft Academia Massachusetts Institute of Technology Rensselaer Polytechnic Institute University of California, San Diego Department of Energy National Laboratories Argonne National Laboratory Lawrence Livermore National Laboratory Los Alamos National Laboratory Oak Ridge National Laboratory Sandia National Laboratories Federal Agencies National Science Foundation Pittsburgh Supercomputing Center NASA Researchers are invited to submit COVID-19 related research proposals to the consortium via this online portal, which will then be reviewed for matching with computing resources from one of the partner institutions. An expert panel comprised of top scientists and computing researchers will work with proposers to assess the public health benefit of the work, with emphasis on projects that can ensure rapid results. Fighting COVID-19 will require extensive research in areas like bioinformatics, epidemiology, and molecular modeling to understand the threat we’re facing and form strategies to address it. This work demands a massive amount of computational capacity. The COVID-19 High Performance Computing Consortium helps aggregate computing capabilities from the world’s most powerful and advanced computers to help COVID-19 researchers execute complex computational research programs to help fight the virus. What Will The Consortium Do? Consortium members manage a range of computing capabilities that span from small clusters to some of the largest supercomputers in the world. As a member, you would support this crucial work by not only offering your computational resources, but also your deep technical capabilities and expertise to help COVID-19 researchers execute complex computational research programs. The Consortium is currently providing broad access to portions of 16 supercomputing systems, representing over 330 petaflops, 775,000 CPU cores, and 34,000 GPUs. The Consortium welcomes additional members who are capable of contributing significant compute resources to the pool for this important work. Organizations interested in joining the Consortium and offering access to their computational resources in support of this project should click here. Submitting a Research Proposal Researchers seeking to access the Consortium’s resources should submit a simple project proposal using the link. The Consortium steering group will review proposals for potential impact, computational feasibility, overall resource requirements and timeline. If a proposal is accepted for support by the Consortium, then a Consortium member will assume responsibility and work with the proposal team to identify the process for access, and to discuss any specific terms and conditions that will apply to the offered access. Since this is a voluntary community effort to support COVID-19 research, proposal teams should expect to produce a regular blog of their activities during the course of their work and should further expect to publish results at the end of the work.","excerpt":"The COVID-19 High Performance Computing Consortium Bringing together the Federal government, industry, and academic leaders to provide access to the world’s most powerful high-performance computing resources in support of COVID-19 research. The COVID-19 High Performance Computing Consortium is a unique private-public effort spearheaded by the White House Office of Science and Technology Policy, the U.S. […]","categories":["Deep Tech"],"tags":["Cloud Computing in IT Industry","Coronavirus","Coronavirus and AI","covid-19","high performance computing"],"author_name":"Vishal Chawla","publish_date":"2020-03-26T10:35:08","publication_year":"2020","word_count":514,"keywords":["Coronavirus and AI","Go","API","covid-19","programming_languages:R","Cloud Computing in IT Industry","programming_languages:Go","Coronavirus","ViT","cloud_platforms:Google Cloud","GAN","cloud_platforms:Amazon Web Services","R","high performance computing"],"extracted_tech_keywords":["R","Go","API","GAN","ViT","cloud_platforms:Google Cloud","cloud_platforms:Amazon Web Services","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/us-government-unveils-the-covid-19-high-performance-computing-consortium\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10171202,"title":"Veo 3 Changes How Videos are Made and Consumed, says Andrej Karpathy","content":"OpenAI co-founder Andrej Karpathy has said that Veo 3 represents a shift in how video is generated, consumed, and optimised. Veo 3 is Google’s latest video generation model that can also create background sounds like traffic, nature sounds, and character dialogue, a feature not currently available in OpenAI’s Sora, Meta’s Movie Gen, Runway ML’s Gen-4, Pika Labs, or Stability AI’s Stable Video 4D 2.0. Sharing his thoughts on X, Karpathy noted that the quality of content improves significantly when audio is added and that the broader implications of models like Veo 3 may not be fully appreciated yet. He explained that video is a high-bandwidth medium for the brain, which is used not just for entertainment, but also for work and learning. According to him, the average person finds video more accessible than reading or writing, and the barrier to creating video content is approaching zero. The most important shift, in his view, is that video is now directly optimisable. He wrote that until now, platforms like TikTok relied on ranking and serving a finite set of videos created by humans. This involved “human creators learning what people like and then ranking algorithms learning how to best show a video to a person”, which he described as “a very poor optimiser”. In contrast, models like Veo 3 produce video through a neural network, making the process differentiable. “You can now take arbitrary objectives, and crush them with gradient descent,” he said. This means engagement metrics like ad clicks or even pupil dilation could be used to guide video generation directly. Even without changing the model parameters, simply refining prompts, either by humans or AIs, could act as a powerful optimisation loop. Karpathy questioned why platforms should rely on a fixed library of videos when they can generate unlimited ones and tune them in real time. He said video could become a core interface for AI-to-human communication and future graphical interfaces, pointing out that diagrams and animations often make concepts easier to grasp than text. He concluded with a warning. While this direction opens up new creative and functional possibilities, he said, “I’m not so sure that we will like what ‘optimal’ looks like.","excerpt":"“You can now take arbitrary objectives, and crush them with gradient descent.”","categories":["AI News"],"tags":["Andrej Karpathy","Google"],"author_name":"Siddharth Jindal","publish_date":"2025-06-03T15:45:21","publication_year":"2025","word_count":363,"keywords":["Go","Andrej Karpathy","OpenAI","AI","neural network","RPA","ML","programming_languages:R","programming_languages:Go","RAG","Google","R"],"extracted_tech_keywords":["AI","ML","neural network","OpenAI","RAG","R","Go","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/veo-3-changes-how-videos-are-made-and-consumed-says-andrej-karpathy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":39880,"title":"SoftBank Makes A Move Into AI Startup Funding, Plans to Invest $55M","content":"The 800-pound gorilla in the VC space has made another move towards consolidating its position. SoftBank Group, known for its ambitious Vision Fund, is now setting up another investment fund. Reportedly, this will be dedicated to startups focusing on artificial intelligence. This is keeping in mind the vision of the founder of SoftBank, Masayoshi Son. He has long been occupied with finding companies in nascent verticals and investing in them. The company is looking to start up this fund in two to three years. It is said that the fund will raise more than $55 million towards AI companies. In a venture to identify and invest in companies that utilize AI to redefine industries and create new ones, SoftBank is looking to snatch up companies that will change the world. To this end, the company also set up an AI-focused startup incubator known as Deepcore. Deepcore has also set up an AI fund in the past, which has since invested in 18 early-stage startups. The company is also looking to finding talent in the early stages and has found 70% of its entrepreneurs directly from the University of Tokyo. Son’s venture might just be what Japan needs. It comes at a time when Japan’s startup culture is suffering. The number of new companies in the country is low compared to others such as the United States and Europe. Son also thinks that initiatives such as this fund and Deepcore will drive the creation of new companies. He stated, “We need to increase [the number of startups] by providing such an environment that engineers can open up a business.” SoftBank has long been in the news for its ambitious Vision fund; a venture which boasts of over $100 billion invested in forward-looking prospects. Overseen by Son, the fund is a vision of his future of the world. It commonly invests in companies that seem to be unrelated to each other, such as ride-hailing to microprocessors to an Internet company. In his words, the Fund is a “coalition of like-minded comrade entrepreneurs” brought together to realize a revolution. The question now on everybody’s mind is whether this new fund can do the same.","excerpt":"The 800-pound gorilla in the VC space has made another move towards consolidating its position. SoftBank Group, known for its ambitious Vision Fund, is now setting up another investment fund. Reportedly, this will be dedicated to startups focusing on artificial intelligence. This is keeping in mind the vision of the founder of SoftBank, Masayoshi Son. […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","masayoshi son","Softbank","Startups","vision"],"author_name":"Anirudh VK","publish_date":"2019-05-29T12:41:40","publication_year":"2019","word_count":361,"keywords":["Go","artificial intelligence","programming_languages:R","vision","AI","programming_languages:Go","masayoshi son","Softbank","Startups","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/softbank-makes-a-move-into-ai-startup-funding-plans-to-invest-55m\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10067982,"title":"Ireland gets its first AI ambassador. Will other countries follow suit?","content":"Ireland has appointed Dr Patricia Scanlon as its first AI Ambassador to facilitate the Government’s AI adoption strategy launched last year. Patricia is the founder and former executive chairperson of the speech recognition tech firm called SoapBox Labs. Ireland’s ‘AI – Here For Good’ strategy focuses on how technology can be utilised in human-centric and ethical ways to improve the lives of its citizens. Dr Scanlon, a member of the Enterprise Digital Advisory Forum (EDAF), will work closely with the Department of Enterprise, Trade and Employment. She will work on demystifying AI and promoting the positive impacts it can have in areas such as transport, agriculture, health and education. The Department of Trade, Enterprise and Employment announced the national AI strategy, ‘AI – Here for Good’, on 8th July 2021. The strategy outlined Ireland’s plan to become a global leader in artificial intelligence to benefit its economy and society, with a “people-centred, ethical approach to AI development, adoption and use”. Further, Ireland will join the Global Partnership on AI and continue to take part in EU discussions and define a framework for trustworthy AI. The government also plans to identify areas AI researchers from Ireland could collaborate with other countries. As part of a broader strategy, higher education institutions are encouraged to design AI-related courses and employers are urged to facilitate workplace-focused AI upskilling and reskilling. Additionally, National Youth Assembly on Artificial Intelligence is set to take place in September 2022 to address concerns of youth around AI and to promote STEM careers. AI, AI everywhere The AI Ambassador appointment came weeks after the United States Department of Defense appointed Dr Craig Martell as the Chief Digital and Artificial Intelligence Officer (CDAO)–a newly created position. The role was created to monitor data and AI initiatives under one official at the highest levels of the Pentagon. The CDAO reports directly to the deputy secretary of defence. Countries worldwide are engaged in an AI arms race, sometimes literally. For example, in the ongoing Russia-Ukraine war, the former had used AI-based drones to unleash terror on Ukrainian cities. Ukraine, on its part, has also taken help from the US firm Clearview AI to ‘uncover the Russian assailants and combat misinformation.’ In Xinjiang and Tibet, China uses AI-powered technology to combine multiple streams of information—including individual DNA samples, online chat history, social media posts, medical records, and bank account information—to track citizens. India is behind the US, China, the UK, France, Japan and Germany in the top AI adopters list. Canada, South Korea and Italy round out the top 10. In October 2016, the Obama administration released a report titled “Preparing for the Future of Artificial Intelligence”, addressing concerns around AI like its application for the public good, economic impact, regulation, fairness and global security. In addition, the US government also released a companion document called the “National Artificial Intelligence Research and Development Strategic Plan – 3”, which formed the benchmark for Federally-funded research and development in AI. These documents were the first of a series of policy documents released by the US regarding the role of AI. The United Kingdom announced its national development strategy in 2020 and issued a report to accelerate the application of AI by government agencies. In 2018, the Department for Business, Energy, and Industrial Strategy released the Policy Paper – AI Sector Deal. The Japanese government released its paper on Artificial Intelligence Technology Strategy in 2017. The European Union launched “SPARC,” the world’s largest civilian robotics R&D program, in 2014. Developing countries such as Mexico and Malaysia are in the process of creating their national AI strategies. India’s efforts to drive AI adoption In recent years, the Indian government has launched several initiatives at state and national levels. In 2018, the Indian government published two AI roadmaps – the Report of Task Force on Artificial Intelligence by the AI Task Force constituted by the Ministry of Commerce and Industry and the National Strategy for Artificial Intelligence by Niti Aayog. The National Roadmap for Artificial Intelligence by NITI Aayog proposed creating a National AI marketplace. In particular, the data marketplace would be based on blockchain technology and offer features like traceability, access controls, compliance with local and international regulations, and a robust price discovery mechanism for data.In 2022, the government increased the budget expenditure from INR 6,388 crores to INR 10,676.18 crores for the Digital India programme to boost AI, machine learning, IoT, big data, cybersecurity and robotics. India’s flagship digital initiative plans to make the internet more accessible, promoting e-governance, e-banking, e-education and e-health.","excerpt":"India is behind the US, China, the UK, France, Japan and Germany in the top AI adopters list.","categories":["AI Trends"],"tags":["AI adoption","AI for good","AI in government","Digital India","Supercomputers"],"author_name":"Kartik Wali","publish_date":"2022-05-28T10:00:00","publication_year":"2022","word_count":756,"keywords":["big data","Go","Digital India","artificial intelligence","AI adoption","machine learning","AI","AI in government","AI for good","Git","RAG","Supercomputers","Rust","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","Go","Rust","Git","big data","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ireland-gets-its-first-ai-ambassador-will-other-countries-follow-suit\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":50959,"title":"How The Trend Of Passenger-Aware Cars Is Overtaking Self-Driving Vehicles","content":"When someone mentions Artificial Intelligence and cars together, the first thing that pops in mind is self-driving cars. With giants like Google, Uber and now Apple, showing their interest in self-driving cars and letting the public know that these ‘cars of the future’ are coming has everyone eagerly waiting for them. Well, they are here but, one might wonder- are they ready to take on the roads, especially Indian roads? And when are they rolling out? Next year? A year after that? Try years. While we wait for cars without drivers, technology is exploring the idea of integrating A.I. inside the car. A car that does not drive itself but is aware of its passengers sitting inside. What’s inside the car’ means is using A.I. to understand what is happening with the person driving the car. As of now, A.I.’s algorithms for self-driving cars can perform limited tasks which aren’t suitable because of the open environment, but the algorithms can perform better within the restricted area inside the vehicle. As important as it is for A.I. to understand what is happening inside a driverless car, A.I.’s services inside the car with a driver is also needed. A.I. To Identify The Rightful Owner In-car cameras, through their complex vision algorithms, can perform complex tasks — one such task is detecting the rightful owners of the car. Chooch, an artificial intelligence training platform and API for facial recognition, is developing a facial recognition system to identify the rightful owners of the vehicles. This ability of A.I. can match faces with identity cards within the vehicle. This feature is vital in the near future when it comes to the safety of the cars. This ability of facial recognition could help the car rental companies along with private owners. For example, if you are renting a car, as a security check, you hold up your I.D. proof and the facial recognition system matches your face to give you access to the car. The match ensures that the correct person is sitting behind the wheel. Being Aware Of The Car’s Environment A.I. system inside the car is essentially ‘aware’ of the objects inside it. The A.I. algorithms for the car regulate the car’s environment which applies to auto-adjusting interior lights and locking the doors and changing the volume of the music during a phone call or in the case of dangerous driving conditions. The doors on the backseat auto-lock when it detects kids in the backseat or changes the channel or radio to the kid’s programme. The car can also slow down or can be alerted in case of loud threatening voices or cursing. Functions like these and A.I.’s algorithms can also monitor the backseat after you leave the car, then notify if you’ve left something important behind. R.E.A.D. (Real-time Emotion Adaptive Driving System): During the 2019 Consumer Electronic Show, Kia introduced the Real-time Emotion Adaptive Driving System (R.E.A.D.) technology. Kia stated that the AI-based system could tailor vehicle interiors to a passenger’s emotional state. It does so by using sensors to monitor facial expressions, heart rate and E.D.A. (Electrodermal Activity) of the passengers. R.E.A.D. System also includes seat vibrations, and these aren’t just some normal vibrations, these vibrations in the seat match their frequency to whatever kind of music is being played inside the car (So steer clear of Heavy Metal Music). “Emotion A.I. can provide an understanding of people’s preferences and optimise the in-cabin environment to offer a personalised experience,” El Kaliouby, co-founder and C.E.O. of Affectiva says. Object And Emotion Detection Technology Affectiva, an emotion measurement technology company that came out of M.I.T. Media Lab, has developed an A.I. system that can detect emotions and expressions in the human driving the car. This kind of technology will roll out in the coming two to three years. This technology makes use of a camera which is near the steering wheel. This camera monitors the driver’s behavioural patterns such as frequency and length of blinking eyes to determine whether the driver is feeling drowsy. If there is some problem with the behavioural patterns, then it notifies the driver by changing the temperature, playing music or pulling over. The A.I. also is being used to detect distractions, for example, when drivers are eating or distracted by the phone. That Feeling Of Insecurity Whenever there is mention of A.I., there is a fear of the risk of privacy. But, with facial recognition, there arises another challenge with A.I. algorithm, algorithm bias. If a deep learning algorithm is trained on too many dark-toned humans, then the algorithm will be less accurate in detecting people with white colour skin tone. For the technology to work globally, these problems with algorithm bias have to be addressed, and these can only be reduced if more diverse data is put through the algorithm. Companies like Affectiva have analysed around 8.5 million faces in 87 countries to counter the algorithm bias. Building an A.I. system which monitors and determines the activity inside the vehicle needs to have ample amount of user data. What happens with A.I. algorithm training is that they have to be fed a massive amount of user data for them to learn. The controversies like Amazon’s Alexa, accidentally recording a private conversation makes people anxious about the technology. With concerns like this, one of the solutions to emerge is edge A.I. which makes the need for sending the data to the cloud unnecessary. Instead, it runs algorithms locally without a link to the cloud. Outlook With India making rapid advances towards A.I., the idea of self-driving cars directs people’s attention to one of the crucial things- Job security. India will lose around 23% of the jobs that will be lost if there is complete automation globally by 2021, according to research by Human Resources (H.R.) solutions firm PeopleStrong. But, we need not worry because of many driverless cars around the world are partially self-driving or equipped with the Level 5 automation that implies full automation in all conditions as defined by SAE International (an association which develops global standards for mobility industry). With practical self-driving cars far away and the reach of the algorithms limited to a certain extent, for now, maybe our best shot at automation is making use of whatever A.I. technology is available now. But, moving forward, if A.I. is going to help us in automating with something as personal as our cars, then even if there is a cushion of more safety, transparency is going to play a pivotal role as it always has in the matter of A.I.","excerpt":"When someone mentions Artificial Intelligence and cars together, the first thing that pops in mind is self-driving cars. With giants like Google, Uber and now Apple, showing their interest in self-driving cars and letting the public know that these ‘cars of the future’ are coming has everyone eagerly waiting for them.  Well, they are here […]","categories":["AI Features"],"tags":["Autonomous Vehicles","Self Driving Cars","Technology"],"author_name":"Sameer Balaganur","publish_date":"2019-12-02T12:00:43","publication_year":"2019","word_count":1094,"keywords":["Go","API","artificial intelligence","programming_languages:R","AI","Self Driving Cars","programming_languages:Go","automation","Technology","deep learning","ViT","Autonomous Vehicles","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","R","Go","API","ViT","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/trend-passenger-aware-cars-overtaking-self-driving-vehicles\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":29741,"title":"Distributed Machine Learning Is The Answer To Scalability And Computation Requirements","content":"We all know the traditional way of machine learning, where programmers use an integrated tool for data mining and conduct analysis on the results. However, the traditional way may not work if the data is too large to store in the RAM of a single computer. Most existing ML algorithms are designed by assuming that data can be easily accessed, which means the same data may be accessed many times. Scalability Led To the Rise of Distributed ML It was this challenge to handle large-scale data due to scalability and efficiency of learning algorithms with respect to computational and memory resources that gave rise to distributed ML. For example, if the computational complexity of the algorithm outpaces the main memory then the algorithm will not scale well and will not be able to process the training data set or will not run due to memory restrictions. Distributed ML algorithms rose to handle very large data sets and develop efficient and scalable algorithms with regard to accuracy and to requirements of computation (memory, time and communication needs). Distributed ML algorithms are part of large-scale learning which has received considerable attention over the last few years, thanks to its ability to allocate learning process onto several workstations — distributed computing to scale up learning algorithms. It is these advances which make ML tasks on big data scalable, flexible and efficient. There are two approaches to distributed learning algorithms. The distributed nature of these datasets can lead to the two most common types of data fragmentation: Horizontal fragmentation where subsets of instances are stored at different sites Vertical fragmentation where subsets of attributes of instances are stored at different sites Some of the most common scenarios are where distributed ML algorithms are deployed are in healthcare or advertising where a simple application can accumulate a lot of data. Since data is huge, programmers frequently re-train data so as not to interrupt the workflow and use parallel loading. For example, MapReduce was built to allow automatic parallelisation and distribution of large-scale special-purpose computations that process large amounts of raw data, such as crawled documents or web request logs and compute various kinds of derived data. Several Distributed ML Platforms Are New One of the most widely-used distributed data processing systems for ML workloads is Apache Spark MLlib and Apache Mahout. Microsoft also released its Distributed ML Toolkit (DMTK), which contains both algorithmic and system innovations. Microsoft’s DMTK framework supports unified interface for data parallelisation, hybrid data structure for big model storage, model scheduling for big model training, and automatic pipelining for high training efficiency. System innovations and ML innovations are pushing the frontiers of distributed ML. Single Machine vs Distributed ML Experts emphasise that traditional ML approaches are designed to address the dataset at hand which implies central processing of data in a database. However, this is usually not possible due to the fact that the cost of storing a single dataset is bigger than storing data in smaller parts. Also, the computational cost of mining a single data repository or database is bigger than processing smaller parts of data As opposed to a centralised approach, a distributed mining approach helps in parallel processing. Also, distributed learning algorithms have their foundations in ensemble learning which helps build a set of classifiers to improve the accuracy of a single classifier. An ensemble approach merges with that of a distributed environment since a classifier is trained onsite, with a subset of data stored in it. Distributed learning also provides the best solution to large-scale learning given how memory limitation and algorithm complexity are the main obstacles. Besides overcoming the problem of centralised storage, distributed learning is also scalable since data is offset by adding more processors. Also, experts peg that in the future data analytics will be primarily done in a distributed environment. Disadvantages of Distributed ML algorithms Unfortunately writing and running a distributed ML algorithm is highly complicated and developing distributed ML packages becomes difficult because of platform dependency. On the other hand, there are no standardised measures to evaluate distributed algorithms. Many ML researchers say that existing measures benchmarked against classical ML methods show less reliability. But one thing’s clear — the practice of ML which was so far concentrated on monolithic data sets from where learning algorithms generate a single model is soon getting phased out with distributed learning algorithms. Also the rise of big data and IoT has led to several distributed data sets and these big datasets stored in a central repository impose huge processing and computing requirements. And that’s why researchers assert that distributed processing of data is the right computing platform.","excerpt":"We all know the traditional way of machine learning, where programmers use an integrated tool for data mining and conduct analysis on the results. However, the traditional way may not work if the data is too large to store in the RAM of a single computer. Most existing ML algorithms are designed by assuming that […]","categories":["AI Features"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-10-30T08:44:24","publication_year":"2018","word_count":772,"keywords":["Go","machine learning","AI","ML","distributed computing","Apache Spark","Scala","RAG","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","distributed computing","Apache Spark","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/distributed-machine-learning-is-the-answer-to-scalability-and-computation-requirements\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10044836,"title":"How Microsoft’s AI For Accessibility Is Addressing The Issue Of Data Desert","content":"The lack of machine learning datasets that include people with disabilities has proved to be a major hindrance in developing solutions customised to their needs. This phenomenon is often referred to as ‘data desert’. It is common practice for organisations to build technology products and services to use data at an aggregate level, leading to stereotyping and exclusion in the process. For the longest time, AI-based technology has been very generalised and hence very exclusionary in catering to persons with disabilities. It has led to the ousting of a considerable chunk of society from the benefits of AI. But now, Microsoft is leading the way with joint efforts by advocacy organisations that hope to do something for the inclusiveness of AI in this “data desert”. ttps:\/\/twitter.com\/MSFTEnable\/status\/1417159949466181636?s=20 Microsoft’s AI for accessibility Microsoft announced the AI for Accessibility initiative in May 2018. The company had then pledged $25 million for AI development projects by universities, philanthropic organisations, and others to develop technology for people with disabilities. Here are some of the recent AI for accessibility projects led by Microsoft to make AI more inclusive: Braille AI Tutor: Microsoft Corp. has joined hands with Massachusetts-based startup ObjectiveEd in its efforts to improve educational outcomes for children with disabilities. ObjectiveEd’s Braille AI Tutor is a teaching aid for students with impaired vision. It uses gamification to help students practise independently or during distance learning via speech recognition with a braille display. I-Assistant: inABLE, a Kenyan non-profit organisation, that empowers blind and low vision persons through computer assistive technology, has joined hands with India-based non-profit organisation I-STEM (it develops STEM technical solutions) to produce I-Assistant. It uses a custom-designed model, based on top of Azure Computer Vision to detect different structures and layouts for documentation and processed text, including math expressions. It can also create a readable and accessible document that converts any examination into a digitally accessible format. Users can navigate the test or paper smartly and perform background actions with Azure Custom Speech Recognition, Text to Speech, and Language Understanding (LUIS), such as checking the exam status. ADMINS chatbot: UK’s largest academic institution, Open University students have worked with professors to create an ADMINS chat assistant ‘Taylor’. This model makes forms programmatically accessible and also explores an alternative form completion process that closely matches a natural dialogue between a disabled student and a smart chatbot. The ADMINS team analysed forms and dialogue between students and expert advisers through the Azure Bot framework and speech services to recognise relevant intentions and to develop a system that promotes student engagement during a conversation, either spoken or typed. The chatbot uses the free form of the student’s responses with Azure Language Understanding (LUIS) to automatically lead the student through each step of the process and interpret answers to the corresponding forms. TigerChat: Rochester Technology Institute (RIT) and National Technical Institute for the Deaf (NTID) have developed the TigerChat app to help improve communication for students who are deaf or hard-of-hearing.TigerChat was established with the support of CloudCheckr, a Rochester-based technology company, through Microsoft’s AI for Accessibility grant. Furthermore, RIT is exploring different ways of improvements in captioning options for Deaf and hard-to-hearing students, including automatic removal of disfluencies and addition of punctuation. Roadmap for Data Oasis While Microsoft is working towards AI inclusiveness and eradication of data desert, the absence of machine-learning datasets representing or including disabled individuals is a common barrier for researchers or developers in many sectors and enterprises. Especially those that are working to develop smart solutions to support daily work or AI systems for such groups. To ensure that all individuals have access to the benefits of data-driven innovation and no one is at a disadvantage, policymakers should also strive to eradicate data poverty and close the data divide from the ground up. This can also be achieved if companies start investing in AI-driven projects that reduce bias by including data on disabled people and make software and devices smarter and more contextually relevant, and cost-effective.","excerpt":"AI-based technology has been very generalised and hence very exclusionary in catering to persons with disabilities.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Facial Recognition","Microsoft"],"author_name":"Ritika Sagar","publish_date":"2021-07-31T13:00:00","publication_year":"2021","word_count":664,"keywords":["Anthropic","machine learning","Facial Recognition","Microsoft","AI","R","data-driven","innovation","Git","computer vision","GAN","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","computer vision","Anthropic","Azure","R","Git","GAN","data-driven","innovation"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-microsofts-ai-for-accessibility-is-addressing-the-issue-of-data-desert\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10052411,"title":"No-Code AI Platform E42 Raises $5.4 Million in Series A Funding&nbsp;","content":"Light Information Systems recently announced that the company has raised $5.4 million in Series A funding from Pavestone Ventures. Their award-winning no-code AI platform, E42, has fuelled enterprise cognition by automating complex, time-consuming, and resource-intensive processes so that they no longer require human intervention. The automation is powered by multifunctional cognitive agents, or “AI Co-workers”, which enterprises can create and customize on the E42 platform based on their needs. The Pune based company, founded in 2012 by Animesh Samuel and Sanjeev Menon, makes process-centric and people-centric automation easier across verticals by driving human-like cognition across these processes. The platform has been recognized by industry giants like Microsoft, SAP, Oracle, IDC, NVIDIA, NASSCOM, SaaSBhoomi, etc. The funding comes at a time when there is an ever-increasing demand for cognitive process automation across industries and will be deployed in enhancing the company’s AI platform “E42”, by boosting R&D to maintain the tech advantage, build their partner ecosystem, and engage businesses globally. “We are thrilled to receive the funding as it will help us strengthen our research and partner ecosystems and help us scale and serve enterprises across geographies. Our Cognitive Process Automation platform has benefited organizations, big or small, with automating complex and people-centric processes, that too, with a return on investment in as short a time as two weeks. We believe that an AI-driven strategy will be at the core of every business to succeed. We want to help these organizations in mapping out this strategy with the help of our multifunctional cognitive agents\/AI co-workers, which will help these businesses succeed by optimizing profits while reducing time-consuming and difficult processes. We want to open up our easy to deploy, maintain and migrate the no-code platform to enterprises and partners globally to come and try out different automation that can be customized as per their requirements,” said Animesh Samuel, Chief Executive Officer, Light Information System Private Limited. Sridhar Rampalli, Managing Partner, Pavestone Ventures, said, “We’re happy to be a part of E42’s growth story. The need for end-to-end process automation has never been greater as AI-led enablement and automation across enterprises is maximizing efficiencies and scalability while minimizing complexities and costs. We believe the E42’s platform has a combination of both horizontal and vertical capabilities which exposes significant scope to deliver across enterprise functions at scale and depth.” The AI co-workers created on E42’s platform are the future-ready AI workforce that augments the existing organization’s workforce by helping them efficiently scale their business by improving user experience across verticals. These AI co-workers built on E42 include AI Analysts, AI AP Associates, AI HR Executives, AI Recruiters, AI Business Dev Execs, etc., and are deployed at several large and medium enterprises worldwide.","excerpt":"Their award-winning no-code AI platform, has fueled enterprise cognition by automating complex, time-consuming, and resource-intensive processes so that they no longer require human intervention.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Automation","autonomous systems","Data Science","Deep Learning","Machine Learning","no code platforms","no-code","Oracle Certification","Pune"],"author_name":"Victor Dey","publish_date":"2021-10-26T17:53:34","publication_year":"2021","word_count":449,"keywords":["funding","programming_languages:R","AI","Automation","Machine Learning","no-code","Scala","GAN","automation","autonomous systems","programming_languages:Scala","Oracle Certification","Pune","Deep Learning","Data Science","no code platforms","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","Scala","GAN","automation","funding","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/no-code-ai-platform-e42-raises-5-4-million-in-series-a-funding\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119949,"title":"Tredence Appoints Munjay Singh as Chief Operating Officer","content":"Tredence, the global data science and AI solutions company, has appointed Munjay Singh as its Chief Operating Officer, as part of its continued growth strategy. Singh, who has previously held senior leadership roles at global technology consulting and product firms, brings with him extensive experience in driving operational efficiency and customer experience initiatives across diverse market segments. In his new role, Singh will provide strategic vision, leadership, and deep operational expertise to accelerate Tredence’s organic and inorganic growth strategies. This appointment comes at a time when Tredence has experienced significant growth, having raised more than $205 million from private equity firms. The company has also expanded into new regions and verticals, launched the ATOM.AI ecosystem, and developed a GenAI-as-a-service platform. Tredence has achieved remarkable growth, growing fourfold between 2020 and 2024, with a 40% growth rate in 2023. “I am thrilled to join Tredence and amplify its ability to help enterprise clients modernise data ecosystems and solve last-mile AI challenges,” said Singh. “Our suite of 100+ AI\/ML accelerators delivers unprecedented value to clients, providing a wide array of solutions to boost decision-making and unlock new opportunities. Additionally, our verticalization strategy, expert practices, and extensive partner network enable us to solve increasingly complex industry challenges and tailor our capabilities to client needs.” Shub Bhowmick, CEO and co-founder of Tredence, expressed his excitement about Singh’s appointment. “Tredence collaborates with more than forty Fortune 500 companies to help them uncover opportunities in marketing, customer experience, supply chain, and other functions. We have developed verticalized collections of AI and data accelerators that our clients have implemented to achieve tangible business improvements within weeks. Under Munjay’s leadership and strategic guidance, we aim to propel this vision forward, driving innovation and operational excellence across our business functions and practices, and achieving new levels of success.”","excerpt":"In his new role, Singh will provide strategic vision, leadership, and deep operational expertise to accelerate Tredence’s organic and inorganic growth strategies.","categories":["AI News"],"tags":["Chief AI Officer (CAIO)"],"author_name":"Mohit Pandey","publish_date":"2024-05-09T11:16:38","publication_year":"2024","word_count":299,"keywords":["data science","GenAI","programming_languages:R","AI","innovation","Chief AI Officer (CAIO)","ML","Ray","Aim","GAN","R"],"extracted_tech_keywords":["AI","ML","data science","GenAI","Aim","Ray","R","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tredence-appoints-munjay-singh-as-chief-operating-officer\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10142476,"title":"Elon Musk, Demis Hassabis Collaborate to Build an AI Game","content":"Google DeepMind recently introduced Genie 2, a large-scale foundation world model capable of generating a wide variety of playable 3D environments. Genie 2 facilitates the development of embodied AI agents by transforming a single image into interactive virtual worlds that can be explored by humans or AI using standard keyboard and mouse controls. “Genie 2 could enable future agents to be trained and evaluated in a limitless curriculum of novel worlds. This research also paves the way for new, creative workflows for prototyping interactive experiences,” Google Deepmind said in its blog post. Introducing Genie 2: our AI model that can create an endless variety of playable 3D worlds – all from a single image. 🖼️These types of large-scale foundation world models could enable future agents to be trained and evaluated in an endless number of virtual environments. →… pic.twitter.com\/qHCT6jqb1W— Google DeepMind (@GoogleDeepMind) December 4, 2024 “The world model is taking shape,” said Google DeepMind chief Demis Hassabis in a post on X. X chief Elon Musk replied, “Cool,” to which Hassabis responded, “Thanks, Elon! Let’s do an AI game together…” “Ok, that would be cool”, said Musk. Musk, an avid gamer, recently shared that xAI is starting an AI-based game studio. “Too many game studios are owned by massive corporations. @xAI is going to start an AI game studio to make games great again!” he posted on X. Interestingly, during a surprise appearance on the Joe Rogan Podcast, Elon Musk shared that he is one of the top 20 ‘Diablo 4’ players worldwide. Not many know that Elon Musk, as an intern many summers ago, wrote the software for video games at a company called Rocket Science. This was long before SpaceX was even conceived. While Musk is focusing on FSD and Optimus, pivoting to video games and entertainment seems like a promising prospect. And while OpenAI hasn’t released Sora yet, Musk said that Tesla has already been doing real-world video generation with accurate physics for about a year. Earlier this year, Musk shared Tesla’s video generation capabilities, to which a user replied, “Tesla should make a video game.” Musk instantly responded, “I’ve wanted to do that for a long time :).” He said that while Tesla’s real-world simulation and video-generation capabilities are the best in the world, making a game can only come after they release unsupervised FSD, which is far safer than supervised FSD. xAI plans to expand its Colossus Supercomputer in Memphis, adding more than 1 million GPUs, the Greater Memphis Chamber revealed today. Already the world’s largest with 100,000 GPUs, Colossus will soon grow tenfold, thanks to an investment of tens of billions from Elon Musk. Also, xAI is preparing to launch a standalone app for its Grok AI chatbot, expected as early as December 2024, to challenge OpenAI’s ChatGPT and Google’s Gemini. The company is also set to initiate a new $5 billion fundraising round, which could double its valuation to $50 billion within six months. Previously, xAI raised $6 billion in Series B funding, bringing its post-money valuation to $24 billion.","excerpt":"“Elon! Let’s do an AI game together…”","categories":["AI News"],"tags":["deepmind founder Demis Hassabis","musk"],"author_name":"Siddharth Jindal","publish_date":"2024-12-05T11:27:51","publication_year":"2024","word_count":507,"keywords":["Go","ChatGPT","funding","deepmind founder Demis Hassabis","OpenAI","AI","R","GPT","XAI","musk","GAN","xAI"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","xAI","R","Go","GPT","GAN","XAI","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/elon-musk-demis-hassabis-collaborate-to-build-an-ai-game\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10126534,"title":"Asus Launches India&#8217;s First CoPilot+ PC","content":"Asus has released India’s first CoPilot+ PC, the Vivobook S 15 OLED. With several hardware companies announcing similar AI offerings for the Indian market, the Asus launch is just the first in a number of launches to come in the next year. This is also the first time that a Vivobook PC has not been powered by an Intel or AMD processor. The AI-powered Vivobook is powered by Qualcomm’s Snapdragon X Elite processor, departing from Asus’ usual reliance on Intel and AMD hardware. The switch in processors is attributed to providing improved energy efficiency and enhanced AI performance. Its other features also include a 15.6-inch 3K OLED display, 16GB of RAM, and a 1TB SSD. Asus also advertises a battery life of up to 18 hours on a single charge. However, its most notable feature, which allows it to stand out, is its AI integration, which is facilitated by a dedicated CoPilot key. This feature provides quick access to AI-driven functionalities, assisting with tasks ranging from system optimisation to intelligent suggestions and automation. Priced at Rs 1,24,990, the S 15 is now available for purchase on all major retail channels across India. As previously reported by AIM, all major companies, including Acer, Dell, HP and Lenovo, have begun announcing their own range of AI-powered laptops and PCs. This is likely a result of NVIDIA releasing more AI-capable GPUs specifically for personal use and improved gaming experiences in the last two years. Additionally, with several offerings from other companies, it seems that Asus has managed to get a headstart on the Indian market with the Vivobook S 15, relying specifically on a Snapdragon processor and the first in the market to launch a CoPilot+ PC. However, as mentioned before, this is just the first in several releases, as more companies begin offering their own CoPilot+ PCs in India.","excerpt":"The AI-powered Vivobook is powered by Qualcomm’s Snapdragon X Elite processor, departing from Asus’ usual reliance on Intel and AMD hardware.","categories":["AI News"],"tags":["Copilot"],"author_name":"Donna Eva","publish_date":"2024-07-11T15:51:41","publication_year":"2024","word_count":307,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","automation","Aim","Copilot","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/asus-launches-indias-first-copilot-pc\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10173060,"title":"How Smart TVs in India are Getting Smarter and Cheaper","content":"Like the rest of the world, many Indians have turned to smart televisions (TVs) for a consolidated platform for all of their entertainment needs. The ‘Make in India’ initiative, driven by the Indian government, has promoted domestic manufacturing in many industries. It has increased exports of electronic goods, including semiconductors, and has also encouraged companies to gradually reduce dependence on international original equipment manufacturers (OEMs). Abhijeet Rajpurohit, COO and co-founder of homegrown smart TV Operating System CloudTV, revealed how he recognised a significant need for independent creation, particularly because Indian factories and brands faced challenges in obtaining resources from global operating systems. At that time, these international companies often played hardball with Indian producers due to the relatively small size and scale of the economy, which was still developing and less mature than that of the West in terms of the smart TV market. “Regardless of how much we would want to push Make in India, it is a shift that’s going to take time. When Make in India started, we slowly worked towards the assembly of these devices. After a couple of years, the government has started directly pushing the manufacturing of chipsets in India,” Rajpurohit said. However, these raw materials are elemental and naturally scarce, which further limits their availability despite technological advancements. Hence, companies don’t shy away from still collaborating with the global supply chain, including Indian manufacturing companies and TV brands. CloudTV’s programming approach is centred around India, with a clear understanding of the television brands and price segments they target. By collaborating with various content providers and streaming applications, the system recommends a diverse range of content. Moreover, it strongly prefers long-tail content, such as sports, and focuses on showcasing programming that resonates with the tastes of the Indian audience. The app’s voice assistant understands and learns from regional languages such as Hindi, Punjabi, Marathi, Haryanvi and Gujarati, to name a few. “We are also scaling for more languages every couple of quarters. Our goal is to cater to a lot more people to make their access to watching smart TVs easier in their languages,” Rajpurohit added. The startup operates as a full-stack solution, with no reliance on external services. “We don’t depend on other services as a blocker mechanism,” he explained. While Alexa still exists as a voice assistant, the company has developed its capabilities. Rajpurohit foresees smart TVs replacing traditional set-top boxes, evolving into multifunctional devices that will facilitate TV commerce and integrated shopping experiences. Although it’s early to predict the exact developments, the competitive landscape will focus on delivering the best user experience. Reports show that smart TV shipments increased by 6.7% year-on-year in Q4 2024, totalling over 63.3 million units. The global smart TV market is recovering after three years of weak performance, driven by economic challenges and geopolitical tensions. Samsung leads with a 16.9% market share, followed by TCL (13.9%), Hisense (12.8%), and LG (11.1%). Moreover, local manufacturing has substantially reduced costs, making smart TVs accessible to mass consumer segments in India. Chinese brands hold a substantial market share of 30% in India, primarily in the budget and mid-range segments. Japan and South Korea remain leaders in premium segment offerings, but at higher price points, Indian consumers now have access to premium models with parity. This has led to Indian buyers strongly benefiting at the lower tiers, with matched value at premium levels. However, Anshika Jain, a senior analyst at Counterpoint Research, pointed out, “AI-enabled smart TVs remain in the early stages of adoption, with major players gradually offering advanced features such as adaptive picture and sound, as well as multilingual voice assistants.” Rajpurohit highlighted that smart TVs have made significant contributions to the Indian electronics ecosystem. About 10 years ago, many Indian mobile brands disappeared due to challenges, leading to a surge in foreign manufacturers engaging in price wars, he added. If CloudTV hadn’t entered the market, a similar trend might have developed in the television sector, he said. The company aims to benefit both brands and users by reducing reliance on global supply chains and partnering with local brands and content providers in India. “India’s smart TV market declined 7% YoY in 2024 due to a slower replacement cycle, macroeconomic situation, and a decline in shipments of Chinese brands, which accounted for 28% of total volume in 2024. With household penetration at roughly 70%, there remains considerable room for growth, driven by rising consumer demand for premium models, local manufacturing, replacement of older units, and the growing trend of multiple TVs per household,” Jain further explained. In Q1 2025, global advanced TV shipments increased by 44% YoY, with revenues rising 35%, according to Counterpoint Research. TCL and Hisense drove this growth by more than doubling shipments compared to the previous year, edging closer to Samsung’s long-standing dominance. According to reports, the global smart TV market was valued at $290.7 billion in 2024 and is projected to reach $475.0 billion by 2033, growing at a CAGR of 5.56% from 2025 to 2033. Key growth factors include rising demand for interactive entertainment, device convergence, and increased internet penetration. Demand for smart TVs is growing, as they serve as a hub for controlling smart home devices, including voice assistants and security systems. The popularity of smart TVs with gaming features and support for gaming platforms is also driving this market growth. “As more users upgrade to larger screens and feature-rich models, India is expected to witness growth in the upcoming years. Compared to advanced markets like China, Japan, and South Korea, India remains price-sensitive, though a gradual shift toward premiumisation is taking shape as buyers begin to value performance and quality alongside affordability,” Jain concluded.","excerpt":"The Centre’s Make in India initiative has spurred domestic manufacturing and pushed companies to gradually reduce their reliance on international suppliers.","categories":["AI Features"],"tags":["Make in India","Smart TV"],"author_name":"Smruthi Nadig","publish_date":"2025-07-08T17:30:00","publication_year":"2025","word_count":944,"keywords":["Go","Make in India","programming_languages:R","AI","Smart TV","programming_languages:Go","RAG","Aim","R","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-smart-tvs-in-india-are-getting-smarter-and-cheaper\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039535,"title":"Walmart Donates 20 Oxygen Generating Plants To Fight Covid-19 In India","content":"Walmart announced that the company will make a donation of up to 20 oxygen-generating plants and 20 cryogenic containers to India. The goal of this donation is to mitigate the storage and transportation of the gas. According to sources, the company will also provide USD 2 million to non-governmental organisations to help them fight the devastating surge in coronavirus cases. Walmart stated in a comment, “Walmart will donate up to 20 oxygen-generating plants and 20 cryogenic containers for oxygen storage and transportation, as well as more than 3,000 oxygen concentrators and 500 oxygen cylinders for delivering oxygen therapy to patients at home or in hospital.” The decision by the American retail giant came as India continues to witness a steep increase in the number of Covid-19 cases and deaths due to the shortage of oxygen supply. As per reports, the statement also stated that Walmart and the Walmart Foundation have also committed to funding an additional 2,500 oxygen concentrators as part of the US-India Business Council and US-India Strategic Partnership Forum’s joint relief effort. Doug McMillon, President and CEO of Walmart Inc said, “Walmart is a global family. We feel the impact of this devastating surge on our associates, families and friends across India, and it’s important that we come together to support however we can.” “We are working hard to combine Walmart’s global capabilities and Flipkart’s distribution network to ensure vital oxygen and supplies are made available to those who need them most. Our hearts are with everyone in India,” McMillon added. A total of about USD 1 million will be allocated through the Walmart Foundation Disaster Relief Fund, a donor-advised fund. The other USD 1 million will be allocated to GIVE Foundation Inc. to support GiveIndia’s Covid response fund. Walmart Canada will also support relief efforts through the Canadian Red Cross India Covid-19 Response Appeal, funding ambulance and transport services for patients, quarantine isolation centres and other services.","excerpt":"Walmart announced that the company will make a donation of up to 20 oxygen-generating plants and 20 cryogenic containers to India. The goal of this donation is to mitigate the storage and transportation of the gas. According to sources, the company will also provide USD 2 million to non-governmental organisations to help them fight the […]","categories":["AI News"],"tags":["Walmart","Walmart Labs"],"author_name":"Ambika Choudhury","publish_date":"2021-05-04T12:18:57","publication_year":"2021","word_count":320,"keywords":["Go","funding","programming_languages:R","AI","Walmart Labs","programming_languages:Go","RAG","ViT","GAN","Walmart","R"],"extracted_tech_keywords":["AI","RAG","R","Go","GAN","ViT","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/walmart-donates-20-oxygen-generating-plants-to-fight-covid-19-in-india\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10105,"title":"Snapdeal establishes Data Science Center in US to define consumer-centric growth strategy","content":"Snapdeal, India’s largest online marketplace is the latest entrant to have vouched for data sciences and analytics. With their announcement of establishing a Data Sciences Center in San Carlos, California, they intend to study customer behavior and strengthen the supply chain. The center would be headed by Nitin Sharma, Senior Vice President, Data Sciences and includes veterans of data science from leading global brands like Groupon, Google, Yahoo and Amazon. The strong team of global talent would help build high- value solutions and develop a high impact growth strategy. “We have set up a Data Science engine in California, which is home to domain talent, to further augment our efforts in creating a superior customer experience and strengthen our supply chain”, said Rohit Bansal, Co-Founder, Snapdeal. Snapdeal’s consumer centric initiatives would be reaching new heights as the center would be focusing on big data and advanced analytics to enhance the same. It would also help shape the business strategy of the company and optimize the operational efficiencies using data driven algorithms, data analytics and predictive modeling. “Snapdeal is extensively working on data mining through an existing analytics team. Under Nitin’s leadership the data science team will focus on elevating Snapdeal’s growth-focused strategy and to provide insightful guidance”, Rohit added. With an aim to utilize data sciences center by capturing and integrating the information on user’s buying behavior, Rohit believes that it would also help the company to achieve the vision of 20 million daily transacting users by the year 2020. Commenting on the data center, Nitin Sharma, Senior Vice President, Data Sciences at Snapdeal said that their highly accomplished team can distil key patterns, consumer preferences and hidden correlations by quickly analysing huge quantities of data. “We will bring fresh insights to the existing work and will enhance customer experience through better planning and forecasting”, said Nitin. Snapdeal currently enlists 35 million plus products across 800 plus diverse categories from over 125,000 regional, national, and international brands and retailers. Catering to over 6000+ cities, Snapdeal has partnered with several global marquee investors and individuals such as SoftBank, BlackRock, Temasek, Foxconn, Alibaba, eBay Inc., Premji Invest etc. With Snapdeal’s data center aiming to define consumer-centric growth strategy, it would only enhance their already impressive customer reach and help expand their business a step further.","excerpt":"Snapdeal, India’s largest online marketplace is the latest entrant to have vouched for data sciences and analytics. With their announcement of establishing a Data Sciences Center in San Carlos, California, they intend to study customer behavior and strengthen the supply chain. The center would be headed by Nitin Sharma, Senior Vice President, Data Sciences and […]","categories":["AI News"],"tags":["ecommerce analytics"],"author_name":"Srishti Deoras","publish_date":"2016-06-08T09:45:37","publication_year":"2016","word_count":383,"keywords":["big data","data science","Go","ecommerce analytics","programming_languages:R","AI","programming_languages:Go","Aim","analytics","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","big data","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/snapdeal-establishes-data-science-center-us-define-consumer-centric-growth-strategy\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52804,"title":"How This Japanese Video Game Is Being Used To Benchmark AI Models","content":"Benchmarking AI agents inside simulated environments have become the new norm. Researchers are leveraging a wide range of video games to train and evaluate AI models for improving their performance. The idea behind using games is to achieve generalisation for models. To attain true AI, researchers test them in an environment in which the models have never been trained in. Recently, researchers used Mega Man 2 game to assess AI agents as it consists of eight different challenges. Today, mostly, organisations train and evaluate models in a similar environment, resulting in attaining a higher accuracy. However, when the same agent is used in an unfamiliar environment, it terribly fails to deliver the desired performance. Therefore, the gaming environments are actively utilised for benchmarking AI agents for its efficiency. Motivation Behind Choosing Mega Man 2 Mega Man 2 is an action game published by Capcom, which is the sequel of the Mega Man that was released in Japan in 1988. In the Mega Man 2 game, contenders have to battle against the evil Dr Wily and its rogue robots — Metal Man, Air Man, Bubble Man, Quick Man, Crash Man, Flash Man, Heat Man, and Wood Man. These eight different enemies make it a perfect simulated environment to test AI agents. The gameplay of Mega Man is the prime reason behind adopting it as a testing environment. In the game, a player controls the Mega Man, who is equipped with a weapon, and on defeating evil robots the Mega Man gets the specialised robots’ weapon. Since evil robots have a different speciality, every time the Mega Man defeats an enemy, he gets a unique weapon, thereby, making it easy for the player to conquer the next enemy. As a part of their exploration, researchers replaced manual player with AI agent. And to increase the difficulty, Mega Man didn’t use the new weapons it used to receive on demolishing the villains. Methodology Used By Researchers Researchers used EvoMan, a game-playing framework based on the boss fights of the game Mega Man 2. They aimed to overpower all eight enemies using only the default arm cannon of Mega Man. For this, they trained AI agent only to defeat a set of bosses and then are evaluated with the models against all eight evils. This allowed them to assess how an agent performed when it was deployed in a new environment. The agent can only deliver in a new environment if it is been able to generalise things when it was trained in a specific territory. To evaluate the generalisation, researchers trained the model on a set of four enemies and then validated it by deploying it with all eight bosses. As each enemy differs from one another, generalising patterns for avoiding being shot and winning by making desired manoeuvres is a challenging task. Consequently, several learning strategies were used to check which one delivers sufficient generalisation. Different agents were used for all four bosses to training and get specialised models. Variants of neuroevolution strategies with 1-layer perceptron and 2-layers perceptron with 10 and 50 neurons for the hidden layer were used. And the weights of neural networks was tuned by Genetic Algorithm and LinkedOpt algorithm. Besides, the evolution of neural networks topology which was carried out with the NEAT algorithm. The above image was the log of the evaluation of models which was the harmonic mean — the game started with 100 points and the points which were gained by hitting the enemy. And then the final point was calculated as: Gain = 100.01 + ep – ee. While ep represents the energy of the player, ee is the energy of the enemy. The constant — 100.1 was added to get a valid result. The harmonic mean depicts that NEAT learning methodology produced the best results followed by the two-layer neural network whose weights were tweaked with Genetic Algorithm. What It Means Various organisations are trying to achieve true AI by enabling models to generalise the learning. Earlier in December OpenAI benchmarked reinforcement learning to avoid model overfitting and shifted the landscape by solving the absence of generalisation problem. Besides, DeepMind tried a similar attempt with Starcraft 2 and obtained a 99.8% success. Such instances are proving Jerome Pesenti, head of AI at Facebook, wrong as he, in early December, said that AI is still a pattern matching. However, with this development, we can envision commonsense in AI with generalisation. Outlook The Generalisation is where AI lies as only delivering exceptional results in the same environment will not take us towards accomplishing AI in everything. Thus, it is of paramount importance to progress in generalisation to integrate AI in a wide range of use cases. This is yet another success in the AI landscape for improving agents dexterity. However, often the real-world performance of models are wimpy, thus it would be interesting to witness how these advancements are materialised in the real world.","excerpt":"Benchmarking AI agents inside simulated environments have become the new norm. Researchers are leveraging a wide range of video games to train and evaluate AI models for improving their performance. The idea behind using games is to achieve generalisation for models. To attain true AI, researchers test them in an environment in which the models […]","categories":["AI Features"],"tags":["Reinforcement Learning"],"author_name":"Rohit Yadav","publish_date":"2019-12-30T15:00:00","publication_year":"2019","word_count":820,"keywords":["Go","Reinforcement Learning","OpenAI","AI","neural network","programming_languages:R","RAG","Aim","AI agents","GAN","R"],"extracted_tech_keywords":["AI","neural network","OpenAI","Aim","RAG","R","Go","GAN","AI agents","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-japanese-video-game-is-being-used-to-benchmark-ai-models\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10060997,"title":"Small object detection by Slicing Aided Hyper Inference (SAHI)","content":"In surveillance applications, detecting tiny items and objects that are far away in the scene is practically very difficult. Because such things are represented by a limited number of pixels in the image, traditional detectors have a tough time detecting them. So, in this article, we will look at how current SOTA models fail to recognize objects at the far end, as well as a strategy presented by Fatih Akyon et al. to address this problem called Slicing Aided Hyper Inference (SAHI). Below are major points listed that are to be discussed in this article. Table of contents The small object detection problemHow SAHI helps to detect small objectsImplementing the method Let’s start the discussion by understanding the problem of small object detection. The small object detection problem What is Object Detection? Object detection is a task that involves bounding boxes and classifying them into categories to locate all positions of objects of interest in an input. Several ways have been proposed to accomplish this goal, ranging from traditional methodologies to deep learning-based alternatives. What are the 2 approaches to Object Detection? Object detection approaches are divided into two categories: two-stage approaches based on region proposal algorithms and real-time and unified networks or one-stage approaches based on regression or classification. Adoption of deep learning architectures in this sector has resulted in very accurate approaches like Faster R-CNN and RetinaNet, which have been further refined as Cascade R-CNN, VarifocalNet, and variants. These new detectors are all trained and validated on well-known datasets like ImageNet, Pascal VOC12, and MS COCO. These datasets generally consist of low-resolution photos (640 480 pixels) with huge objects and high pixel coverage (on average, 60 percent of the image height). While the trained models perform well on those sorts of input data, they exhibit much poorer accuracy on small item detection tasks in high-resolution photos captured by high-end drones and surveillance cameras. Small object detection is thus a difficult task in computer vision because, in addition to the small representations of objects, the diversity of input images makes the task more difficult. For example, an image can have different resolutions; if the resolution is low, the detector may have difficulty detecting small objects. How SAHI helps to detect small objects To address the small object detection difficulty, Fatih Akyon et al. presented Slicing Aided Hyper Inference (SAHI), an open-source solution that provides a generic slicing aided inference and fine-tuning process for small object recognition. During the fine-tuning and inference stages, a slicing-based architecture is used. Splitting the input photos into overlapping slices results in smaller pixel regions when compared to the images fed into the network. The proposed technique is generic in that it can be used on any existing object detector without needing to be fine-tuned. The suggested strategy was tested using the models Detectron2, MMDetection, and YOLOv5. As previously discussed, the models pre-trained using these datasets provide very successful detection performance for similar inputs. They produce significantly lower accuracy on small object detection tasks in high-resolution images generated by high-end drone and surveillance cameras, on the other hand. To address this problem, the framework augments the dataset by extracting patches from the image fine-tuning dataset. Each image is sliced into overlapping patches of varying dimensions that are chosen within predefined ranges known as hyper-parameters. Then, during fine-tuning, patches are resized while maintaining the aspect ratio, so that the image width is between 800 and 1333 pixels, resulting in augmentation images with larger relative object sizes than the original image. These images, along with the original images, are utilized during fine-tuning. During the inference step, the slicing method is also used. In this case, the original query image is cut into a number of overlapping patches. The patches are then resized while the aspect ratio is kept. Following that, each overlapping patch receives its own object detection forward pass. To detect larger objects, an optional full-inference (FI) using the original image can be used. Finally, the overlapping prediction results and, if applicable, the FI results are merged back into their original size. Now we’ll see practically how it can make a difference. Implementing the method In this section, we’ll see how self aided slicing and inference can detect the small object from the images. We’ll compare the original prediction given by the YOLOv5 and SAHI + YOLOv5. Now, let’s start with installing and importing the dependencies. # install latest SAHI and YOLOv5 !pip install -U torch sahi yolov5 # import required functions, classes from sahi.utils.yolov5 import download_yolov5s6_model from sahi.model import Yolov5DetectionModel from sahi.utils.cv import read_image from sahi.predict import get_prediction, get_sliced_prediction from IPython.display import Image Now we’ll quickly load the standard YOLO model and initiate it with the Yolov5DetectionModel method. # download YOLOV5S6 model yolov5_model_path = 'yolov5s6.pt' download_yolov5s6_model(destination_path=yolov5_model_path) # inference with YOLOv5 only detection_model = Yolov5DetectionModel( model_path=yolov5_model_path, confidence_threshold=0.3, device=\"cuda:0\") Let’s get and visualize the predictions. Here we utilize the get_prediction method from the framework which takes an image and detection model. Later we exported our result to the current working directory and visualized the same by using PIL. result = get_prediction(read_image(\"\/content\/mumbai-bridge-1.jpg\"), detection_model) # export the result current working directory result.export_visuals(export_dir=\"\/\") # show the result which saved as prediction_visual.png Image(\"prediction_visual.png\") As we can see from the result, the cars at the far end on the flyover have not been captured by the standard YOLO. Now let’s see how plugging SAHI will improve the overall result. Let’s infer the proposed method. result = get_sliced_prediction( read_image(\"\/content\/mumbai-bridge-1.jpg\"), detection_model, slice_height = 256, slice_width = 256, overlap_height_ratio = 0.2, overlap_width_ratio = 0.2) And here is the result. Isn’t it like magic? Final words Through this article, we have discussed object detection. In a more detailed manner, we discussed how standard SOTA models such as YOLO, fast R-CNN, etc fail when there is interest in detecting the small objects which are usually problems for images taken by drones etc. To overcome this, we have discussed a plug-in kind of method called SAHI which enhances the detection for a given image by mainly slicing and iterating the image and above is the result for this method. References SAHI paperSAHI repositoryLink for above codes","excerpt":"Object detection is a task that involves bounding boxes and classifying them into categories to locate all positions of objects of interest in an input.","categories":["AI Trends"],"tags":["Deep Learning","image processing","Machine Learning","Object Detection","Python"],"author_name":"Vijaysinh Lendave","publish_date":"2022-02-19T10:00:00","publication_year":"2022","word_count":1018,"keywords":["CUDA","Go","AI","image processing","Machine Learning","computer vision","RAG","Python","deep learning","Object Detection","object detection","Deep Learning","R"],"extracted_tech_keywords":["AI","deep learning","computer vision","RAG","object detection","CUDA","Python","R","Go","CUDA"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/small-object-detection-by-slicing-aided-hyper-inference-sahi\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171229,"title":"Why CarDekho Replaced SAP with Oracle Ahead of IPO","content":"As digital automotive solutions provider CarDekho gears up for its IPO, the company is placing technology at the core of its financial transformation. The 18-year-old platform, which has grown into a multi-vertical enterprise spanning used car sales, financing, insurance, mobility, and fleet services, relies on Oracle’s cloud-based enterprise resource planning (ERP) solution to streamline its financial operations. Speaking to AIM, Aditya Changoiwala, group financial controller at CarDekho, highlighted this partnership. Referring to the books of account closures, Changoiwala said, “We had a 15-day closing cycle, and our target is to bring it down to around seven days with Oracle solutions.” With the IPO on the horizon, he stressed that technology is central to this effort and helps provide real-time insights to leadership teams. The platform enables them to shift from manual, time-consuming data entry to a more automated data-driven approach. Moving from SAP to Oracle Before switching to Oracle, CarDekho was running on an outdated, on-prem version of SAP. According to Changoiwala, the system was “primitive”, lacked enhancements, and struggled to support the company’s evolving needs, especially after acquiring several businesses. “We were actually spending a lot of money on maintaining the server and doing maintenance,” he said. “In the last couple of years, we had acquired a few companies and were finding it difficult to integrate them into SAP. Moreover, scalability was an issue.” After evaluating several solutions, the company selected Oracle for its cloud-based ERP, EPM, and SCM modules. Oracle Cloud ERP operates on a pay-as-you-go model, eliminating the need for significant upfront investments in hardware and infrastructure that traditional on-premise SAP systems require. This leads to substantial savings in operational and maintenance costs, allowing companies to manage budgets better and reallocate funds towards innovation and growth. Explaining the challenges with SAP, Changoiwala said that their new businesses are on the cloud, for which they could not maintain a separate team. Moreover, as SAP is on-prem, the updates were not available there. “Even if you want to update or upgrade to the other version, it’s a separate implementation cycle that you must go through,” he said. “We thought if we were to move on to the cloud-based server, we would get all the advantages of SaaS, and then we wouldn’t have to maintain the infrastructure, and all the updates that Oracle releases every quarter,” he added. Changoiwala said that the implementation was phased. “We have already implemented ERP. We are in the process of implementing EPM. In another couple of months, we start on the SCM journey.” The objective was to consolidate systems and reduce reliance on Excel and fragmented processes. CarDekho is not alone in moving away from SAP. According to Oracle, in 2021 alone, over a hundred companies replaced SAP with Oracle. India Pistons, an auto parts maker, is another company that switched to Oracle ERP. After moving from legacy systems to Oracle Cloud, operational costs were cut by 30% to 40%, and productivity was boosted by reducing manual work. The company also reassigned staff to more valuable tasks. Global examples include companies such as Batelco, ClayCo and Lovisa. Moreover, many SAP customers face challenges with the company’s complex and heavily customised platforms, unclear migration paths to S\/4HANA, and fragmented data models. Oracle Cloud offers a more streamlined, future-ready suite of applications that enable faster implementation and quicker realisation of business value. Why Oracle? By using Oracle’s capabilities, CarDekho aims to improve financial reporting and deliver real-time insights crucial for the business to achieve financial prudence across business units. With expansion plans across Southeast Asia and the Middle East underway, the company has turned to Oracle’s cloud-based ERP solutions to address challenges related to scalability, integration, and reporting. In the last 18 months, CarDekho has taken three modules from Oracle, including ERP, enterprise performance management (EPM), and supply chain management (SCM), which are in different stages of implementation across businesses. With a turnover of ₹3,000 crore annually, Changoiwala said the company is EBITDA positive at the parent level, while some businesses are still in development stages. Standardisation and Real-Time Insights CarDekho’s finance operations now run on Oracle ERP, supporting day-to-day accounting, invoicing, and statutory financial reporting. “The base job is right now on ERP,” Changoiwala said. “All the statutory reporting that we are doing in terms of making our financials…is being done using ERP.” One of the main advantages has been the move towards standardisation. “Now most of our businesses are on the same instance. We get real-time information around their closing,” he said. “We don’t have to rely on Excel sheets. We’re able to get a consolidated view.” While the ERP system is still stabilising, Changoiwala expects further value once Oracle’s EPM module goes live. “Internal Management Information System (MIS) and budgeting should go live in another quarter,” he said. “That’s when we’ll be able to have a better view of revenue forecasts.” Early Steps Towards AI Adoption While generative AI isn’t yet a core part of CarDekho’s finance operations, there’s active interest in automation through AI agents. “We are looking at using AI agents to automate standard processes,” Changoiwala said. “I have a big team that spends a lot of time digging into data and doing repetitive tasks. That’s where we’re trying to leverage Oracle’s AI capabilities.” As for generative AI, Changoiwala believes it’s not yet ready for deeper analytical tasks. “It can definitely collate data and summarise it. While it is for good decision-making, I think it’s still far off.” CarDekho’s embrace of Oracle shows how finance teams are replacing outdated processes with cloud-first platforms that drive speed, consistency, and scalability, laying the groundwork for IPOs and future tech adoption.","excerpt":"In 2021 alone, over a hundred companies replaced SAP with Oracle.","categories":["Global Tech"],"tags":["Oracle"],"author_name":"Siddharth Jindal","publish_date":"2025-06-03T17:06:56","publication_year":"2025","word_count":937,"keywords":["Go","AI","ML","Scala","Oracle","RAG","Git","Aim","ViT","generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","RAG","R","Go","Scala","Git","ViT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-cardekho-replaced-sap-with-oracle-ahead-of-ipo\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":15242,"title":"Webinar: Developing and Deploying Analytics for Internet of Things (IoT)\/ Enterprise Systems | 15th June, 3pm","content":"The combination of smart connected devices with data analytics and machine learning is enabling a wide range of applications, from home-grown traffic monitors to sophisticated predictive maintenance systems and futuristic consumer products. While the combination brings virtually limitless opportunities, there are few challenges, such as – using huge data to build accurate and robust models, scaling the analytical models to work with an enterprise system or embedding intelligence in the edge device In this session, we show you techniques for handling large amounts of data, tools used to create machine learning models as well as deploying an IoT system with data analytics without developing custom web software or servers. Highlights: Accessing data in large text files, databases, or from the Hadoop Distributed File System (HDFS) Processing data that does not fit in memory Integrating online analysis and visualization using MATLAB Scaling IoT solutions with analytics Date: June 15 Time: 3 pm Presenter: Amit Doshi Senior Application Engineer, MathWorks India Amit Doshi is a senior application engineer at MathWorks in the area of technical computing. He focuses on data acquisition, data presentation, statistical analysis, and parallel computing. Amit has over 9 years of experience in experimental test setup development, testing and validation, workflow automation, and system simulations. He previously worked at Suzlon Energy Limited in Pune and Germany, Texas Instruments in Germany, and IIT Bombay. Amit holds a bachelor’s degree in mechanical engineering and a master’s degree in mechatronics. Register for this webinar.","excerpt":"The combination of smart connected devices with data analytics and machine learning is enabling a wide range of applications, from home-grown traffic monitors to sophisticated predictive maintenance systems and futuristic consumer products. While the combination brings virtually limitless opportunities, there are few challenges, such as – using huge data to build accurate and robust models, […]","categories":["Deep Tech"],"tags":[],"author_name":"Дарья","publish_date":"2017-05-27T11:52:27","publication_year":"2017","word_count":242,"keywords":["machine learning","programming_languages:R","AI","automation","analytics","workflow automation","R"],"extracted_tech_keywords":["AI","machine learning","analytics","R","automation","workflow automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/webinar-developing-deploying-analytics-internet-things-iot-enterprise-systems-15th-june-3pm\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10008575,"title":"My Fun Project With OpenCV &#8211; Finding Waldo Game","content":"Finding waldo or where’s wally is a popular game where the goal is to find a named Waldo among hundreds of people. As kids, I am sure you have spent quite some time trying to find waldo by carefully going through the entire picture. Wouldn’t it be interesting and fun to build a project that could identify where waldo is within seconds? We can do this using a simple template matching concept of OpenCV. In this article, we will learn what is template matching and write a program to automatically find waldo. What is template matching? Template matching is a process where we take the input image and try to slide the target image over the input. This input image is then matched pixel-wise with the target and the closest match will be highlighted. OpenCV provides a function called matchTemplate() for doing this. This method helps in finding the matching region by providing a threshold. After setting the threshold, the target image is slid over the input and the comparison begins. The result that is obtained is compared to the threshold value. If the result is greater than the threshold the area is marked as detected, else the comparison continues again. The parameters that can be passed to the matchTemplate() function are plenty but for this article, we will be using the TM_CCOEFF_NORMED. This parameter finds the mean value of the pixels in both the template and the input image. A score of 1 indicates a good match and a -1 indicates a bad match and 0 is a neutral match. Finding waldo Let us now get into the program for finding waldo. The objective is to find this man who is Waldo. It is not very easy to do this since the images would look something like this. This is the picture that will be used for the implementation in this article. You can download this image with this link here. Now that we have the source image we need the template image for the matching to happen which is the image of waldo as shown above. You can download this image here. Importing the libraries Since this program was done in colab I will mount the drive and import the basic libraries needed. from google.colab import drive drive.mount('\/content\/gdrive') import cv2 import numpy as np Loading the images source_image = cv2.imread('\/content\/gdrive\/My Drive\/waldo.jpg') grey_scale = cv2.cvtColor(source_image, cv2.COLOR_BGR2GRAY) template_image = cv2.imread('\/content\/gdrive\/My Drive\/wally.png’) Matching the template The next step is to set a threshold and use the function to match the templates. threshold_val = 0.8 matching=cv2.matchTemplate(grey_scale,template_image,cv2.TM_CCOEFF_NORMED) Generally, a value of 0.8 is chosen as the threshold value but another common one if the image is very clustered is 0.6. Now, we need to take the dimensions of our target image so that the location can be matched with the source and compare the threshold values. width, height = template_image.shape[::-1] finding = np.where( matching >= threshold_val) for match in zip(*finding[::-1]): cv2.rectangle(source_image, match, (match[0] + width, match[1] + height), (0,255,0), 2) Results Finally, we will write the output image to the drive and we will find a green anchor box around waldo’s image which makes it easy to find waldo. cv2.imwrite('\/content\/gdrive\/My Drive\/find.png',source_image) As shown above, even among a bunch of people waldo is found based on the matching pixels. You can try to do this with any puzzle for waldo and find him easily. Conclusion With very basic functions we automated the game of finding waldo with openCV in this article. We also saw the importance of the matchTemplate() function which can be used even for object detection. matchTemplate() can also be used dynamically in videos as well.","excerpt":"In this article, we will learn what is template matching and write a program to automatically find waldo.","categories":["Deep Tech"],"tags":["OpenCV"],"author_name":"Bhoomika Madhukar","publish_date":"2020-09-29T14:00:49","publication_year":"2020","word_count":604,"keywords":["Go","NumPy","TPU","programming_languages:R","AI","OpenCV","Colab","Ray","object detection","R"],"extracted_tech_keywords":["AI","Ray","Colab","OpenCV","NumPy","object detection","TPU","R","Go","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/my-fun-project-with-opencv-finding-waldo-game\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":30635,"title":"Solving Global Water Crisis With Artificial Intelligence","content":"The water crisis has become one of the major concerns across the globe. A report suggests that the US alone wastes 7 billion gallons of drinking water per day. As only less than one percent of earth surface water is suitable for human consumption, it becomes crucial that we save water so that our future generations survive. The unchecked use of water and extreme weather conditions have worsened the situation and in no time there will be a fresh water shortage, irregularities in supply and demand, groundwater shrinkage, among other challenges. To help overcome water crisis, organizations have started using artificial intelligence to efficiently stop this wastage. Researchers are using AI as a counterfeit mechanism to traditional hydrological prototypes, which are proving to be successful and effective. Need For AI In Water Management Hydrological issues like flooding, precipitation, contaminant transport, and groundwater management require the availability of detailed and accurate data about the water systems, which are usually not available in the developing countries due to economic and infrastructure limitations. In such challenging scenarios, artificial intelligence comes as a friendly alternative for water management. AI techniques such as Artificial Neural Networks and Support Vector Machine (SVM) are being popularly used as they are less cost-effective when compared to big data mechanisms. ANN algorithms are supporting to build water plants that give updated statistics about the present resources and aid to build models for the upcoming situations. Software programs armed with neural networks strategize water operations dynamically. It is also helping in developing the present water resources. The decision-making abilities of AI can optimize and automate of the available resources. The governing bodies and water concerned departments can understand real-time water loss and misusage, with an AI-driven planning. Use Cases Of Recent Breakthroughs In Greece, researchers have used a feedforward neural network trained algorithm for the stimulation of decreasing groundwater trends. They used the precipitation, temperature and groundwater level data as the vector for neural networks for prediction. The results were able to provide a solution for the next 18 months. The state of Illinois, US, has used the feedforward training algorithm for the prediction of pesticide quotation in groundwater. They used aquifer depth, aquifer sensitivity to pesticide, pesticide leaching and samples for a specific time as vectors for these ANN. In Northern France, scientists have applied the ANN model to estimate the depth of the contaminated territory in the soil to estimate groundwater contamination caused during road projects. To analyse nitrate content in groundwater in the wells of Harran Plain in Turkey, researchers used temperature, electrical conductivity, and Ph levels of groundwater as vectors for ANN. An observation was carried out at Edward’s karstic aquifer in Texas, US, with hydraulic ANN head change to forecast the groundwater level. The previous six days temperature and nearby wells pumping rates were taken as vectors for the experiment. Iran has developed ANN multilayer perceptron (MLP) to model the rainfall-runoff process using rainfall durations, average intensities and season index of over 100 occurrences as vectors for the model. Singapore has developed ANN backpropagation training algorithm for the prediction of coastal water quality using the location of stations, previous salinity, temperature, dissolved oxygen levels and chlorophyll-a levels in the nearby stations as vectors. The Sevier River Basin Utah, US, has developed an SVM model to forecast the streamflow of 6 months ahead using local climatological data with different time variations and previous stream volume flow. An integration of suitable ANN, SVM, and logistic regression with the help of GIS was used to identify nitrate-contaminated water in the wells of Polk County, Florida, US. The vectors for the research were GW depth, characteristics of soil media, hydraulic, topography, pedality and PH value. In this technique, ANN’s outperformed SVM. China used an SVM model for groundwater quality assessment at the Naingziguan fountain by using groundwater quality classification indicators as vectors, which resulted in high prediction accuracy. With the help of a time series model created with SVMs and ANNs, Korea could forecast groundwater level in wells near coastal regions. They used previous data of groundwater level, tide level and precipitation as vectors. Scenario In India India has been recording interesting developments in the area of water crisis management using AI. Two ANN models were developed in India to find water quality of Gomati river. Water quality variables like PH, TS, COD were taken as vectors and the prediction of DO (dissolved oxygen) and BOD (biological oxygen demand). The neural network prototype was developed using data which had observations for three years. The input vectors were designed by the help of correlation coefficient with dissolved oxygen. Performance of ANN prototypes was compared using correlation coefficient, mean squared error, and coefficient of efficiency. The estimated values of DO and BOD were accurate. In another interesting development, the ministry of water resources has collaborated with Google for developing an AI model to forecast floods. This program is still in the budding stage. The government of India and Google are working for the best predictive analysis to handle the natural catastrophes in an efficient way. Outlook With government programmes like Digital India, Make in India and Startup India gaining momentum and IT sector moving towards analytics, swift adoption of AI is giving hope. A number of tech heads from Google, Microsoft, FaceBook, and others are collaborating with the government for developing AI-based models to address various issues.","excerpt":"The water crisis has become one of the major concerns across the globe. A report suggests that the US alone wastes 7 billion gallons of drinking water per day. As only less than one percent of earth surface water is suitable for human consumption, it becomes crucial that we save water so that our future […]","categories":["AI Features"],"tags":["artificial inelligence"],"author_name":"Bharat Adibhatla","publish_date":"2018-11-25T10:42:07","publication_year":"2018","word_count":896,"keywords":["big data","Go","artificial intelligence","artificial inelligence","AI","neural network","ML","Git","RAG","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","analytics","RAG","R","Go","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/solving-global-water-crisis-with-artificial-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":6150,"title":"Adoption Process of Analytics in Organizations","content":"Unfortunately, it has been hard to implement the Analytics strategy in spite of its much touted strategic advantage. The skill set required to leverage its advantages is a heady concoction of data management skills, statistical\/data processing prowess, and business acumen. The resource with the “right” mix is in high demand and, not surprisingly, is available in short supply. Most organizations have evolved their analytic processes around a mix of resources that have varying strengths in more than one of the three key areas. Facilitators Data Inventory Scientific enquiry requires regulated data formats that provide easy evaluation of (in) consistency in the patterns among the data, a requirement for good quality analytics. Hence, if data are available in limited degrees of freedom (less ability to interpret in multiple ways), especially when recorded in numeric system, it lends itself to convenient processing of data patterns to evaluate the evidence of consistency. Any evidence of severe infirmities in the information such as large data gaps, irregular recording, unstandardized formats and non-numeric formats can lead to additional investments in time and cost for appropriate data organization to enable easy analysis. Processing Capabilities Relatively operational in nature, it is a blend of data management and analytical capabilities – statistical and computational abilities. Among the many investments that are required, this is perhaps the easiest to acquire and there are educational hubs that provide reasonable training to equip human resources to handle these responsibilities. Business Insight Development (Business Acumen) Certainly a critical capability, it is second only to the availability of superior quality data. Organizations reaping benefits of data insights need to encourage executives with high business acumen to excel in developing data-driven insight development skills. They need to be comfortable reading data insights and developing the knack of connecting them with their business relevance (also termed as ‘so what’). It is not the easiest capability to be developed. Usually, resources having specialized processing skills and the urge and the capability to participate into active business decision-making have the right mix of skills to drive this activity. However, unlike processing skills which can be imparted, this is a strange mix of data literacy and creative business acumen that develops with exposure to various business problem contexts. What has been the adoption path of these skills in organizations? Over the years, various tools and techniques with different levels of sophistication and intelligence have been used by businesses to aid decision-making. Earlier, when computer-assisted decision-making began, most of it was based on finding out answers to questions such as what happened, how and where it happened (through standard reports, ad hoc reports, query drill drown, etc.) — that could be broadly termed as reactive decision-making. But with the advent of advanced analytics, businesses can now initiate proactive decision-making with predictive and prescriptive analytics that answer questions on why it happened, what may happen next, and how it can be solved. This type of analytics is proving to be a major differentiator amongst competitors. However, suffice to say that the core activities of analytics in many organizations remain primarily in the lower SW quadrant. Ideally, most organizations would like to scale the “heights of the domain” (NE quadrant in Figure), but are restricted largely due to the lack of appropriate data in the organizations. Additionally, organizations in more dynamic environments find it difficult to use past information for planning future activities due to the high rate of depreciation in the value of the information. One important area for productivity improvement in this domain is related to improved measurement of business impact due to the organization’s use of analytics. With wider use and acknowledged importance of this process and also an upward climb on value addition, enough effort is required to document impact in terms of cost savings or improved profitability which is directly related to the establishment of analytic prowess in organizations. While some literature on business impact of analytics is available in the public domain (for instance, Banerjee, 2001), the authors have found a few references to remarkable improvements in business performance due to the investment in these processes. The best references in this respect have been in the offshoring space in India, where the improvement in terms of ongoing cost reductions has influenced transitioning functions to low-cost geographic locations. We lay particular emphasis on this aspect since the long-term consequence of non-measurability of business returns can be unsettling like it has happened in other related industries (Bijapurkar, 1995). Culture: The Toughest Change The market environment today has become far more complex and dynamic and is driven by fastidious customers who have little commonality in their requirements. With high levels of competitiveness and a networked globe wherein competition can emerge out of nowhere and almost instantaneously, decision-makers are bound to be on tenterhooks. In such a scenario, they just cannot afford to take decisions in the way they have been doing in the past, that is, with their ‘instincts’ or ‘blind calls’. With the advent of analytics, organizations are trying to build a fact-based culture. Managers do trust their instincts but term analytics as invaluable, for the simple reason that it builds credence and lends support to their decisions as it is backed by empirical data. So, companies that have successfully adopted analytics maintain a proper balance (between ‘gut’ and ‘data’) in their decision-making and try to build a culture as they mature in their use of analytics. Analytical culture takes time to be built and gain acceptance, and a typical analytics rich organization has the right data management practices, right set of professionals, right mindset, and right amount of patience to see the fructification of analytics initiatives over time. In such organizations, analytics is introduced with a strong belief, and spearheaded from induction to practice to adoption, by the top management, with conviction. Besides the general teething problems of an evolving practice, there are some unique complications in this domain, such as the supposed incompatibility between the skill sets required for (a) managing information, (b) processing the information especially if it is in large scale, and (c) developing implications out of the processing results which connect to the resolution of the business problem using the domain knowledge. Historically, these three disciplines have largely stayed distant from each other and hence the need for their synergetic association was never directly required. Research and analytics-based decision-making processes have forced these three hitherto unassociated skills to integrate. No skill development programme for management, to authors’ knowledge, has focused on the integration of these three skills on such a large scale as it is required today. These challenges have grown exponentially when some analytics functions are displaced to outsourced operations. The imbalance is primarily due to the non-availability of certain competencies in the displaced environment. In most cases, the business domain knowledge in outsourced operations is not in line with the requirements on site. Hence, imbalances are created when certain processes that have a higher technology component (data processing capabilities) are relocated to offshore operations and processes with more policy-making components (business acumen) are retained in the home office. The prime mover for outsourcing and\/or offshoring highend value-added processes to low-cost environs has been cost management. However, imbalances have been caused because business acumen is critical to boost up value addition in the analytics process. Low-cost environs which provide offshore services are associated with market conditions which are not compatible with the developed markets for which the analytical services are rendered. The ability for resources in the offshore environs to appreciate the nuances of domain knowledge of the user (developed) markets is limited, leading to the possibility of suboptimal output. In general, technology in the form of processing capability (surfeit in its availability in lowcost environ) acts as a poor substitute for business acumen to drive value addition. Potential Way Forward for Outsourced Analytics Operations Inevitably, the answer lies in the cross-pollination of skills both within and across environs to improve the competency set. Migration of experts from the domain, development of offshore environs, and better work\/information flows across work sites\/personnel with differential skill sets can help alleviate the challenges over time (Banerjee & Williams, 2009). What has been seen is that short-term imbalances in skill sets are compensated at times with emphasis on one or more skills rather than a combination of all. For instance, both in the offshore and onsite (India) operations, availability of technologically superior manpower has resulted in specialized consulting outfits (both internal and external) that focus on one or more aspects of data processing without putting a commensurate amount of focus on the business acumen, the latter being more difficult for standalone specialists (particularly in offshore environments) to imbibe. Usually, the specialization would follow the intrinsic capabilities of the firms that have expanded their footprint into the Analytics market in the offshore environment. By that logic, one would expect the traditional software6 firms to veer towards automation and reporting processes (descriptive analytics) whereas the algorithm building firms7 would be keener to position themselves as modeling and optimization specialists (predictive and diagnostic analytics). Business consulting firms (though less active in the offshore space) tend to rely on their domain skills to offer prescriptions for business problems while depending on more specialized data processing partners for the analysis. The jury is still out on the usefulness of such specialized and piecemeal measures. Republished on authors consent from Vikalpa","excerpt":"Unfortunately, it has been hard to implement the Analytics strategy in spite of its much touted strategic advantage. The skill set required to leverage its advantages is a heady concoction of data management skills, statistical\/data processing prowess, and business acumen. The resource with the “right” mix is in high demand and, not surprisingly, is available […]","categories":["IT Services"],"tags":["analytics outsourcing"],"author_name":"Arindam Banerjee","publish_date":"2014-09-18T18:17:32","publication_year":"2014","word_count":1560,"keywords":["Go","API","TPU","AI","RAG","ViT","analytics","Rust","GAN","analytics outsourcing","R"],"extracted_tech_keywords":["AI","analytics","RAG","TPU","R","Go","Rust","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/adoption-process-of-analytics-in-organizations\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10071650,"title":"Decoding SEZ 2.0-DESH Bill","content":"While presenting the Union Budget for 2022, Nirmala Sitharaman, the Finance Minister of India, announced that the SEZ Act (2005) would be replaced with a new legislation that would enable the states to become partners in the development of enterprise and service hubs. Sitharaman emphasised that the new legislation would encompass all the existing and new large industrial enclaves in order to optimally utilise available infrastructure and enhance competitiveness of exports. The proposed successor to the SEZ Act (2005) is expected to be the Development of Enterprise and Service Hubs (DESH). The government plans to table the DESH Bill in the ongoing monsoon session of Parliament. In a 2019 report, the World Trade Council’s dispute settlement panel revealed that India’s export-related schemes—which included the SEZ Act—“violated several provisions of the WTO’s subsidies and countervailing measures agreement”. In response, Commerce and Industry Minister Piyush Goyal claimed that India was under no obligation to implement the recommendations put forth by the WTO’s dispute panel. Focus on export and domestic industry However, the Indian government is not keen on taking chances in the matter. In February 2022, Piyush Goyal addressed the media to share that the new legislation would make the existing SEZ Act more inclusive by encouraging participation from states. He further highlighted that the main goal for this ‘upgrade’ is to put a greater emphasis on industrial development and service sector along with exports. Goyal further discussed the urgency of the situation and how the changes are expected to be implemented within four to six months. Commerce Secretary B V R Subrahmanyam also added that the new legislation for SEZs is expected to comply with the global trade rules in accordance with WTO standards. How would DESH help? According to the Economic Survey 2021-22, India’s merchandise exports increased by 49.7 per cent, compared to the corresponding period in 2021 and by 26.5 per cent over 2019-20, between April and December. However, there is growing consensus that the focus needs to expand beyond exports. To that end, DESH Bill aims to boost local manufacturing which would have significant ripple effects—such as better employment opportunities. A byproduct could be increased job creation According to various media reports, in the aftermath of the new legislation, ‘development hubs’ that would no longer be required to remain net foreign exchange positive, i.e., maintaining more exports than imports cumulatively in five years are expected to be created. In addition, it would provide an online, single-window portal for the grant of time-bound approvals for establishing and operating these hubs. Retail in domestic regions would consequently become smoother and easier. Call for upgradation of existing Acts The seeds to identify how SEZs could help in promoting exports for India were sown back in 1965. Asia’s first ‘Export Processing Zone’ was set up in Kandla in Gujarat in 1965. The Special Economic Zones (SEZs) Policy was announced in April, 2000 and received Presidential assent in June, 2005 eventually coming into effect in February, 2006. The policy focused mainly on the promotion of exports of goods and services along with better employment opportunities and development of infrastructure facilities. However, the act has since become outdated owing to the introduction of controversial and highly debated clauses such as the minimum alternate tax and the ‘sunset clause’. In 2018, the Union Commerce Ministry set up a committee headed by Bharat Forge–Chairman and Managing Director, Baba Kalyani to study the SEZ policy. In conclusion to its report, the committee recommended further evaluation of the act.The most notable recommendation put forth by the committee was—“framework shift from export growth to broad-based employment and economic growth”. Other recommendations included the need for separate rules and procedures for service SEZs and manufacturing SEZs.","excerpt":"The government plans to table the DESH Bill during the ongoing monsoon session of Parliament.","categories":["AI Features"],"tags":["manufacturing India"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-07-27T18:00:00","publication_year":"2022","word_count":616,"keywords":["Go","programming_languages:R","AI","manufacturing India","programming_languages:Go","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/decoding-sez-2-0-desh-bill\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10137967,"title":"Atlassian’s AI Rovo is Now Available to All Users","content":"Atlassian, an American-Australian software company, announced on October 9, 2024, that its recently introduced team-building AI, Rovo, will now be available for all users. The company announced this move at its Team’24 Europe conference in Barcelona. Atlassian customers will now be able to expand their subscriptions to incorporate Rovo AI, a comprehensive search engine, chatbot, and AI agents that are helpful to developers and business users alike. The company claims that Rovo can provide answers in seconds, providing personalised search results. Powered by Atlassian Intelligence, it adds a new perspective to collaboration between humans and AI. Its Rovo Search feature helps it find the exact information teams require from under a stack of data. It can generate relevant results across all data sources like Google Drive, Microsoft Sharepoint, Microsoft Teams, GitHub, Slack, and Figma. The software company also releases a ‘bring your own data’ (BYOD) connector that helps each person export custom data from their internal application into their teamwork graph. Rovo Search can then analyse this data to filter out information for customers. It learns and understands the company’s data through tools like AI chat, knowledge cards (that answer your questions directly from first and third-party sources) and AI-driven data insights. “Rovo Chat never gets tired of your questions,” says Atlassian’s CEO and Co-founder, Mike Cannon Brooks, in his speech on 9th October 2024. Rovo can also take action to handle time-consuming tasks handled by specialised Rovo Agents, which are different from bots. Atlassian is starting to reduce friction across workflows for engineers by having AI assist in code completion and soon aiming to increase development speed and quality with their new dev-focused AI Agents. “We have joined the EU AI pact with fellow industry leaders to stay ahead of any upcoming regulations in all the geographies of the world,” Brooks said during his speech.","excerpt":"Atlassian’s CEO and Co-founder, Mike Cannon Brooks, says Rovo Chat will never get tired of your questions.","categories":["AI News"],"tags":["AI Agents","atlassian"],"author_name":"Sanjana Gupta","publish_date":"2024-10-09T18:01:40","publication_year":"2024","word_count":306,"keywords":["Go","programming_languages:R","AI","emerging_tech:AI agents","AI Agents","Git","programming_languages:Go","Aim","AI agents","atlassian","GitHub","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GitHub","AI agents","programming_languages:R","programming_languages:Go","emerging_tech:AI agents"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/atlassians-ai-rovo-is-now-available-to-all-users\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":836,"title":"“Lack of empirically-based codes of conduct and safety in space is a problem, not congestion” says Moriba K. Jah, a Space Expert","content":"Astronaut Today interacted with Moriba Jah, an astrodynamicist with an expertise in qualifying, assessing and predicting the behaviour of objects in space. During this candid chat, he spoke about developments in this space, industries that are being benefited, challenges of astrodynamics and much more. Dr. Moriba Jah who specializes in Kalman filtering, orbit determination, attitude determination, data fusion and much more, has recently been selected as the new Associate Editor under the “Astrodynamics and Space Debris” area of expertise of Advances in Space Research (ASR), which is the Official Journal of the Committee on Space Research (COSPAR), a Scientific Committee of the International Council for Science (ICSU). His other specialties are space surveillance, inverse problems, algorithm development, data science, data analytics, space object behavioural science and space situational awareness. Astronaut Today- What are the recent developments in astronomy\/Space that have been the most revolutionary incidents in the last few decades according to you? Dr. Moriba Jah- Recent collisions on-orbit generating many thousands of debris in single events and the fact that more companies are space actors has been revolutionary developments. The landscape has drastically changed from government-driven activity to commercially-driven activities. AT- Would you like to talk about your transition from a NASA spacecraft navigator to an academician? Do you feel there is a need for more academicians in this space? MJ- Well, I went from NASA to the Department of Defense (specifically the Air Force Research Laboratory) for a decade and then academia. I feel that by having been in government, I have a meaningful pulse on the salient problems and issues we face in space, and an academic environment lets me explore intellectually-curious topics of relevance with little constraints, which is a landscape I couldn’t have as a Civil Servant. Yes, more academics should be a part of this work, but it’s not due to a lack of interest…it’s due to a lack of proper investment. If governments and industry team together and bring forth a meaningful level of funding to do the science that we critically need to solve these problems, you’d have a fierce and healthy competition amongst academics to do the work and you’d get the best ideas on the planet! Money is not everything but the only thing, in this case, preventing a large pool of the smartest people from addressing these issues. AT- Could you give a few highlights of the ASTRIA program at the University of Texas? MJ- ASTRIA seeks to be globally recognized as the first choice for innovative and disruptive solutions to space domain related critically important, multi-source information fusion, data science and analytics, and astronautical sciences and engineering challenges. Our goal is to deliver the best sciences and technologies to the community in a way that successfully transfers these to government and private industry, leading to measurable and positive impact for our stakeholders and partners. ASTRIA continuously seeks to augment and deliver our scientific and engineering research capabilities through a broad network of government, private, academic and public partnerships. One of the first areas of research will be to delve into Space Traffic and Debris Modeling, to quantify, assess, and predict the space object population and use this science to help inform sound space policy to support orbital safety and long-term sustainability of space activities. Astria’s mission is to assemble and lead the world’s top multidisciplinary science and technology research and development and focus it to solve problems requiring rigorous and comprehensive capabilities. It envisions to imagine, identify, develop and deliver new astronautics capabilities and make expertise on astronautics available to a variety of stakeholders including all branches of government, private industry, academia, and international entities. AT- Would you like to enlighten our readers keen on venturing into this space on how prepare themselves for a career in astronomy? MJ- Well, my career is in Astrodynamics, the Science that studies the motion of objects in space. To prepare for that, you need a solid foundation in mathematics and physics. Look at the contributions in Celestial Mechanics made by Kepler, Lagrange, Hamilton, Gauss, Brahe, Galilei, Copernicus, Newton, and Einstein. AT- What are the risks and challenges at the current scenario around space? How could they be dealt with? MJ- The big problem is that we aren’t able to track everything and we can’t detect everything! We have no way to uniquely identify objects in space. Our motto in ASTRIA is, “if you want to know it, you must measure it; if you want to understand it, you must predict it!” We have trouble in both measuring and predicting…and thus in knowing and understanding. This is where our work is focused on… measuring and predicting the behaviour of objects in space! AT- With more and more companies venturing into space, what are the perils of space traffic and how can that be overcome? MJ- Lack of empirically-based codes of conduct and safety in space is a problem…congestion in space is not a problem. The problem is not knowing where everything is and will be. Air traffic is pretty dense. We have air traffic controllers and locations of planes are mostly known to meters. We need an equivalent level of knowledge for space traffic. We aren’t even close to that. But more and more people are hurrying to get things in space to make lots of money. They’re all assuming an unquantifiable level of risk. AT- What are the common industries that get benefitted with exploiting Space? MJ- Almost all of them… oil, energy, farming, disaster relief, communications, weather, navigation, medical, automobile, maritime, fishing, financial\/banking, etc. are being benefitted from space.","excerpt":"Astronaut Today interacted with Moriba Jah, an astrodynamicist with an expertise in qualifying, assessing and predicting the behaviour of objects in space. During this candid chat, he spoke about developments in this space, industries that are being benefited, challenges of astrodynamics and much more. Dr. Moriba Jah who specializes in Kalman filtering, orbit determination, attitude […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2017-05-04T10:53:25","publication_year":"2017","word_count":929,"keywords":["data science","Go","funding","programming_languages:R","AI","programming_languages:Go","ViT","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","analytics","R","Go","ViT","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/lack-of-empirically-based-codes-of-conduct-and-safety-in-space-is-a-problem-not-congestion-says-moriba-k-jah-a-space-expert\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10038933,"title":"Hands-On Guide To Image Classification Using R","content":"Image classification is an important Machine Learning task which assigns a label to an input image. It is quite a common practice among ML enthusiasts to implement the task of classifying images using Pythonic code. This article will guide you to perform image classification using R programming language. Before proceeding, you can check our following articles related to the fundamental concepts of R and its uses cases other than image classification implemented using it. Basics of RSome free e-books for RTime series analysis and forecasting with RExploratory Data Analysis (EDA) in Python vs R Practical implementation using R Here’s a demonstration of performing image classification using RStudio version 1.2.1335. We have used the Fashion-MNIST dataset with 28*28 dimensional gray-scale images categorized into 10 classes. The whole dataset has been partitioned into a training set of 60,000 images and a test set comprising 10,000 images. The step-wise explanation of the code is explained below: Install TensorFlow package library(tensorflow) install_tensorflow() Install Keras package install.packages(“keras”) Load the Keras package into the current session. library(keras) The R interface to Keras library has an in-built Fashion-MNIST dataset, which can be loaded using the dataset_fashion_mnist() function. data <- dataset_fashion_mnist() Output: Load training and test sets into four different arrays- two for images and labels of the training set and the other two for those of the test set. We c() function here to combine images and labels into a vector. c(training_imgs, training_lbls) %<-% data$train c(testing_imgs, testing_lbls) %<-% data$test The output labels are integers in the range 0-9. We create a vector with class names corresponding to each integer label, which will later be used to visualise the classification output. The class labels can be found in the ‘Details’ section of this page. output_categories = c('T-shirt\/top','Trouser','Pullover','Dress','Coat','Sandal','Shirt','Sneaker','Bag','Ankle boot') Explore the data by checking the dimensions of training and test set arrays for images and labels. The dim() function can be used to retrieve the images’ dimensions. dim(training_imgs) Output:  [1] 60000    28    28 dim(training_lbls) Output: [1] 60000 dim(testing_imgs) Output: [1] 10000    28    28 dim(testing_lbls) Output: [1] 10000 The above outputs show that each of the training and test set has 28*28 dimensional images, with the count of images for the former set being 60,000 and that for the latter one being 10,000. Preprocess the data before model training. First, install and load the tidyr R package, making the data tidy(easy to handle) for visualization. install.packages(“tidyr”) library(tidyr) Install the ggplot2 graphics package for visualization. install.packages(“ggplot2”) library(ggplot2) Check if the set of training images is a dataset and assign hat dataset to a variable named ‘img’ img <- as.data.frame(training_imgs[2, , ]) Count the number of images in img using ncol(), create a sequence from 1 upto the number of columns using seq_len() and then retrieve the column names from the img set of images. colnames(img) <- seq_len(ncol(img)) Retrieve the images from the dataset img$y <- seq_len(nrow(img)) Get the pixel values of images img <- gather(img, \"x\", \"value\", -y) img$x <- as.integer(img$x) Initialize a ggplot object. Provide img set as parameter and also specify the aesthetics of the plot using aes(). ggplot(img, aes(x = x, y = y, fill = value)) + #Plot the geometric tiles geom_tile() + #Create 2-color gradient (b\/w) for the plot scale_fill_gradient(low = \"white\", high = \"black\", na.value = NA) + #Create y scale marked from bottom to top scale_y_reverse() + #Create theme for data and non-data elements’ display theme_minimal() + theme(panel.grid = element_blank())   + theme(aspect.ratio = 1) + xlab(\"\") +    #x-axis label ylab(\"\")       #y-axis label Output: The above output image shows that pixel values of the input images are in the range 0-255. We carry out normalization to convert the pixel values that fall in the range 0-1 before feeding the images to the neural network. Divide pixel values of each of the training and test set images by 255 for such normalization. training_imgs <- training_imgs \/ 255 testing_imgs <- testing_imgs \/ 255 Display some initial images. #Set graphical parameters using par() function. Fill 4*4 matrix column-wise (specified #using mfcol parameter) for displaying 16 images par(mfcol=c(4,4)) #specify margin sizes using mar parameter par(mar=c(0, 0, 1.5, 0), xaxs='i', yaxs='i') #Plot the ith image for (x in 1:16) { img <- training_imgs[x, , ] img <- t(apply(img, 2, rev)) #Create grid of rectangles image(1:28, 1:28, img, col = gray((0:255)\/255), xaxt = 'n', yaxt = 'n', main = paste(output_categories[training_lbls[x] + 1])) } Output: Build the neural network model. #Create linear stack of network layers model <- keras_model_sequential() model %>% #Flatten the 28*28 images into 1D array layer_flatten(input_shape = c(28, 28)) %>% #Create hidden layer with 128 neurons and ReLU activation function layer_dense(units = 128, activation = 'relu') %>% #Output layer with 10 neurons (as there are 10 output classes) and softmax activation function layer_dense(units = 10, activation = 'softmax') Compile the model model %>% compile( optimizer = 'adam',    #model optimization technique loss = 'sparse_categorical_crossentropy',   loss function metrics = c('accuracy')    #evaluation metric ) Fit the model on training data model %>% fit(training_imgs, training_lbls, epochs = 10, verbose = 2) Output: Accuracy and training loss graphs: Model evaluation on test set images eval <- model %>% evaluate(testing_imgs, testing_lbls, verbose = 0) Print test loss cat('Test loss:', eval[“loss”], \"\\n\") Output: Test loss: 0.3401631 Print test accuracy cat('Test accuracy:', eval[“accuracy”], \"\\n\") Output: Test accuracy: 0.8849 Make predictions on test set images. predicted_class <- model %>% predict_classes(testing_imgs) Print predicted classes for first 10 images in the test set. predicted_class[1:10] Output: [1] 9 2 1 1 6 1 4 6 5 7 Plot several images with predicted labels. #Create 4*4 matrix and fill it column-wise par(mfcol=c(4,4)) #Set margins par(mar=c(0, 0, 1.5, 0), xaxs='i', yaxs='i') for (i in 1:16) { img <- testing_imgs[i, , ] img <- t(apply(img, 2, rev)) #subtract 1 from predicted label as predicted value will be in range 1-10 while actual as labels’ #range is 0- 9 pred_label <- which.max(pred[i, ]) - 1 actual_label <- testing_lbls[i] #If prediction is correct, display the labels in green color if (pred_label == actual_label) { color <- '#008800' #Display labels in red for false prediction } else { color <- '#bb0000' } #Create the grid of images image(1:28, 1:28, img, col = gray((0:255)\/255), xaxt = 'n', yaxt = 'n', main = paste0(output_categories[pred_label + 1], \" (\", output_categories[actual_label + 1], \")\"), col.main = color) } Output: The above output plot shows that the model has misclassified 4th image in the 1st row. Predict output label for a single image Extract an image from the test set first. Maintain the dimension of the batch of images as it is needed for the model. image <- testing_imgs[3, , , drop = FALSE] Check dimension of the extracted image dim(image) Output: [1]  1 28 28 Make a prediction for the image pred <- model %>% predict(image) Display confidence score for each of the classes pred Output: Subtract 1 from the predicted labels as actual labels start from 0 pred <- pred[1, ] - 1 Get the predicted class having maximum confidence score which.max(pred) Output: [1] 2 Code source: Official RStudio tutorial.R file of the above implementation is available here","excerpt":"Image classification is an important Machine Learning task which assigns a label to an input image. It is quite a common practice among ML enthusiasts to implement the task of classifying images using Pythonic code. This article will guide you to perform image classification using R programming language. Before proceeding, you can check our following […]","categories":["Deep Tech"],"tags":["Guide","Image Classification","machine learning document classification","r programming"],"author_name":"Nikita Shiledarbaxi","publish_date":"2021-04-26T16:00:00","publication_year":"2021","word_count":1168,"keywords":["Guide","machine learning","Keras","machine learning document classification","AI","neural network","TPU","ML","Python","Ray","r programming","TensorFlow","R","Image Classification"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","Ray","TensorFlow","Keras","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-to-image-classification-using-r\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054480,"title":"IIT Jodhpur And AIIMS Jodhpur Develop ‘Talking Gloves’ For Differently-Abled Individuals","content":"Innovators from the Indian Institute of Technology (IIT) Jodhpur and All India Institute Of Medical Science (AIIMS) Jodhpur, in a joint initiative, have developed low-cost ‘Talking Gloves’ for people with a speech disability. The developed device costs less than Rs 5000, which uses principles of Artificial Intelligence (AI) and Machine Learning (ML) to automatically generate speech that will be language independent and facilitate the communication between mute people and normal people. It can help individuals convert hand gestures into text or pre-recorded voices. Hence, it makes a differently-abled person independent and communicates their message effectively. Prof Sumit Kalra, Assistant Professor, Department of Computer Science and Engineering, IIT Jodhpur, along with his team of innovators including Dr Arpit Khandelwal from IIT Jodhpur, and Dr Nithin Prakash Nair (SR, ENT), Dr Amit Goyal (Prof & Head, ENT), Dr Abhinav Dixit (Prof, Dept of Physiology), from AIIMS Jodhpur, have recently acquired a patent for this innovation. “The language-independent speech generation device will bring the people back to the mainstream in today’s global era without any language barrier. Users of the device only need to learn once, and they would be able to verbally communicate in any language with their knowledge. Additionally, the device can be customised to produce a voice similar to the original voice of the patients, which makes it appear more natural while using the device,” said Prof Sumit Kalra, Assistant Professor, Department of Computer Science and Engineering, IIT Jodhpur. In the developed device, electrical signals are generated by the first set of sensors, wearable on a combination of a thumb, finger(s), and\/or a wrist of the first hand of a user. These electrical signals are produced by the combination of fingers, thumb, hand and wrist movements. Similarly, electrical signals are also generated by the second set of sensors, on the other hand. These electrical signals are received at a signal processing unit. Image Source: IIT Jodhpur The magnitude of the received electrical signals is compared with a plurality of predefined combinations of magnitudes stored in a memory by using the signal processing unit. By using AI and ML algorithms, these combinations of signals are translated into phonetics corresponding to at least one consonant and a vowel. In an example implementation, the consonant and the vowel can be from Hindi language phonetics. A phonetic is assigned to the received electrical signals based on the comparison. An audio signal is generated by an audio transmitter corresponding to the assigned phonetic and based on trained data associated with vocal characteristics stored in a machine learning unit. The generation of audio signals according to the phonetics having a combination of vowels and consonants leads to the generation of speech and enables the speech impaired to audibly communicate with others. The speech synthesis technique of the present subject matter uses phonetics, and therefore the speech generation is independent of any language. In recent years, the technological advancement in the field of electro-medical devices for life support, implantable biomedical devices, and wearable medical devices have successfully provided artificial abilities to afflicted people related to any kind of disability or impairment. The patented innovation from IIT Jodhpur and AIIMS Jodhpur is a big step towards advancement in this field. The team is further working to enhance the features such as durability, weight, responsiveness, and ease-of-use, of the developed device. The developed product will be commercialised through a startup incubated by IIT Jodhpur.","excerpt":"The developed device costs less than Rs 5000, which uses principles of Artificial Intelligence (AI) and Machine Learning (ML) to automatically generate speech that will be language independent and facilitate the communication between mute people and normal people.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning","Speech Recognition","speech recognition algorithm","speech synthesis"],"author_name":"Victor Dey","publish_date":"2021-11-29T16:44:55","publication_year":"2021","word_count":565,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","speech synthesis","AI","innovation","ML","Machine Learning","speech recognition algorithm","programming_languages:Go","Speech Recognition","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","R","Go","innovation","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-jodhpur-and-aiims-jodhpur-develop-talking-gloves-for-differently-abled-individuals\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":25371,"title":"This Startup Is Using AI And Analytics To Find Companionship For Differently-Abled Persons","content":"Matchmaking platforms are no novelty in today’s market. But what makes Inclov stand out from others is its initiative to find friendship and love for differently-abled persons. Short for “inclusive love”, Inclov is designed for accepting persons who may or may not be differently-abled themselves. Founded by Kalyani Khona along with Shankar Srinivasan, Inclov was born out of Khona’s core belief that no one deserves to be alone. What started as an idea in 2014, is today an active platform for persons with disabilities for finding companionship, with more than 30,000 users and 10,000 unique connections across the country. As a young student just out of college, Khona came up with the concept of Inclov and spent days researching only to find out that this had never been done or entertained before. After multiple meetings and tons of first-hand experiences interacting with people from the differently-abled community looking for a suitable match, Khona decided to take her idea from Mumbai to Gurugram. How The Inclov App Works Once a user gets on to the app, Inclov curates the user’s profile which is followed by a review process, mobile verification and in-chat feature, where the user does not have to disclose personal contact details. Once the user is on the app, he or she can view five profiles each day. “Our algorithm then works on giving them a match based on age, location, lifestyle, disability type, medication and curability”, says Khona. She adds, “The app is fully-accessible to people with visual impairment through screen reader and talkback. We also launched Inclov’s web version and Inclov Lite, a text-focused version of the app to make the platform even more inclusive”. After a successful round of crowd-fundraising and great feedback from the community, Inclov is now available in web, Lite, Hindi and Punjabi versions. The startup raised last round of funding from  Rajan Anandan, managing director and India head for Google, Debjani Ghosh, former MD of Intel South Asia, and existing investor Raghav Bahl, founder of Network 18 and Quintillion Media. With a team of 8-10 people, they aim to utilise the funds primarily toward product development and geographical expansion. How Does Analytics Come Into Picture For Inclov, the community member is at the core of the company, and that’s why the entire tech and user interface is dedicated to making the user experience seamless. To ensure this, since its very inception, the team has extensively been using data mining and analytics to ensure that they are listening to the community and help create a better experience for them. Khona explains, “We have a dedicated in-house team which is mapping this on a regular basis. Our analytics engine is designed based on big data and uses Python Flask as a backend to track a user’s journey as well as understand their behaviour. We then use this information to fix the leaky bucket.” They also have daily and monthly reports on user demographics, location, tracking URLs, keywords and traction using Inclov’s own backend APIs. To keep up with the technological expertise, the startup is hiring developers and engineers extensively for front-end and back-end operations. “Apart from technical skills and work experience in Python (Flask), we also require people who fit in the culture of the company and understand the value of looking at the big picture”, she shares. Since the core principle of the startup is inclusion, they also have differently-abled persons on their tech team which helps them design some of these tenets. AI And ML Are At The Core The startup uses AI and ML to aid scalability of their platform, which currently caters to over 195 million strong community of persons with disabilities in India. “Machine learning helps us understand the user’s persona — his or her disability type, the percentage of disability, cure availability and similar medical information. Once we receive this data, with the help of AI, we use filters and common algorithms like age, disability preference, location along with few other key behavioural factors to determine best suitable matches for the user profile”, Khona shares. She further explains a few instances where deep tech helped them. For example, deep tech can help them determine if a user with Polio is compatible with other users who have physical impairment; or visually-challenged individual is more likely to interact with other visually-challenged people more comfortably, and so on. Technology Stack At Inclov Technology is the bedrock on which the platform is built. The app, web platform and lite versions are all using tech best practices to make the experience more enriching. “For example, when we first started, it took the team five months to develop the app and conduct multiple testing to ensure 100 percent accessibility — colour themes for people with colour blindness, talkback integration for people with visual impairment, and font size options for people with retinal disorder, among others.” With ML and AI algorithms for matching members and by using big data in creating new product milestones, the startup has improved significantly in its product offerings. The current version of the Inclov mobile app comes with talkback and colour change features. The Road Ahead “We’ve had a great response so far. What started as an idea is today the largest platform for people with disabilities in the country.  We launched our web version in March year and have also rolled out our app in Hindi and Punjabi which will soon be followed by other languages”, Khona shares proudly. She also said that how regularly they hear from their community members and their families about how Inclov has helped them enable independent social living. “Our most valued milestones will always be community-centric, from having marriages happen because of our platform to seeing people make friends for the first time in their lives”, she said. Sharing her plans for 2018, she said, “While we have begun with India and the immense encouragement we have had here we want to take our platform to the world are looking at Singapore and Australia as our next market”. “Apart from this our team is working on an interesting programme called the Genome Project which will help create and map digital DNA of every user based on their past behaviour, interest, hobbies and partner preferences to suggest and match with other suitable candidates”, she said. Challenges To Overcome “While the use of analytics and AI has been of great impetus, there are still some key challenges that we continue to face — magnitude and accessibility happen to be two of them”, notes Khona. She further explains the challenges as below: Magnitude of disability spectrums: For every disability type (of 100+ listed on their platform), each person with an impairment has a spectrum and the percentage of their mobility, cure, movement — progressive or non-progressive. For us to map these aspects while showing suitable profiles to each person with intellectual, physical, learning disability with persons of no disability is extremely challenging. Understanding that data to create a user persona leads to unique preference sets for every person with a disability. Accessibility: There are 70,000 devices across the globe and we have to ensure that the platform is accessible (for people with visual impairment, retina disorder, colour blindness etc.) across all platforms during every device or OS update which is extremely challenging to keep up with for a small four people tech team with multiple platforms and projects under their belt.","excerpt":"Matchmaking platforms are no novelty in today’s market. But what makes Inclov stand out from others is its initiative to find friendship and love for differently-abled persons. Short for “inclusive love”, Inclov is designed for accepting persons who may or may not be differently-abled themselves. Founded by Kalyani Khona along with Shankar Srinivasan, Inclov was […]","categories":["AI Startups"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2018-06-12T12:02:33","publication_year":"2018","word_count":1235,"keywords":["Go","machine learning","AI","ML","Scala","RAG","Python","Aim","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","Python","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/this-startup-is-using-ai-and-analytics-to-find-companionship-for-differently-abled-persons\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":13298,"title":"Artificial Intelligence in Ad Tech- How?","content":"Google robots dream up crazy psychedelic pictures using ‘inceptionism’ AI neural networks Artificial Intelligence and Machine Learning are the two key buzzwords these days in the Startup World. And why not, it’s optimizing our efficiency in everything we do. So, how can Ad Tech be left behind? Ad Tech, to be frank, is sort of a messed up market. It has fallen short on many promises it has made. The good and bad thing about ad tech is that there is an abundance of data. Yes, we have all this information to better understand our customers, but most marketers don’t know how to leverage the data and move forward with it. Very few in the Ad Tech world have the analytical skills to effectively evaluate big data. This lack of training can have damaging effects to brands and companies as they may misinterpret or miss on important data. To overcome this lacking, introducing AI into Ad Tech companies cannot be a bad idea after all. So let’s get into understanding how Ad Tech partners with AI to get lower CPC prices, higher click-through rates (CTR) and conversions, and an even greater ROI. Here we see how use of Artificial Intelligence can help Ad Tech Industry find better answer to these tricky answered questions. AI in Ads Positioning Instead of developers haggling in front of their PCs to determine which Ad Position will determine best revenue for your website, AI simplifies the same by employing Machine Learning algorithms to study historical data to find the relevant ads for targeted user group. AI algorithms use heatmaps to study where the visitors on the website are actually going and thus, placing relevant ads on those positions to grab user attention. The heatmaps have never been used much in the Ad tech before. AI is surely going to change that while studying each and every map in detail to find the best Ad Positions for marketers. The AI also keeps in mind UX of the website helping the publishers to optimize them in the best possible manner. AI in Ad Network Selection There are many Ad Networks that provides different kind of Ads to the website owner and he is required to sort the ads according to the website. This whole process is called Ads Mediation and is very relevant for higher revenues of the website. The AI technology helps here by both minimizing the human effort and doing it more efficiently while making sure that only relevant ads reach to the final user. These relevant ads are defined through the keywords associated with a website in the AI based approach. Ads will be placed through data, facts & Intelligence where data will be user or website’s past history. UX, Website Content, Geo & Timing will be counted as facts here. The best part of employing AI for Ads Optimization is that it reduces human effort by employing a data-oriented approach. It reduces human error and outs the ads only after data says “Yes” based on two parameters which are– Ad Position and Ad network selection. And all of this will only get better when Machine Learning algorithms are employed to find the best Ad-User Match. Automated Analytics It is regularly stated that all Ad Tech companies don’t take their Data Analytics very seriously! Analytics has always been an issue for any Ad Tech and in the end publishers aren’t very satisfied. However, the AI based approach will propel reporting and analytics feature of Ad tech companies to new levels. “We have been able to guide some publishers to help them answer questions like which kind of content their audience likes and where shall we put the CTA button to convert their one time users into loyal users helping them increase their daily traffic by a good margin” says Abhinav Agarwal, Founder of AdStair, India’s First AI automated Ad Tech Company. If Ad Tech companies’ Analytics start impacting the content and website’s UI\/UX, the coming days would turn out really exciting for Ad Tech and its parties i.e. Publishers, Ad Tech Platforms and off course the people who run the real show, the Users.","excerpt":"Artificial Intelligence and Machine Learning are the two key buzzwords these days in the Startup World. And why not, it’s optimizing our efficiency in everything we do. So, how can Ad Tech be left behind? Ad Tech, to be frank, is sort of a messed up market. It has fallen short on many promises it has made. The […]","categories":[],"tags":[],"author_name":"AIM Media House","publish_date":"2017-03-08T03:53:26","publication_year":"2017","word_count":690,"keywords":["big data","Go","machine learning","artificial intelligence","AI","neural network","RAG","analytics","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","analytics","RAG","R","Go","big data","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/artificial-intelligence-ad-tech\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10100276,"title":"Why Time is Ripe for the &#8216;Real&#8217; GPT-4","content":"We had previously raised the question: What happened to multimodal-GPT-4? Six months later, it appears that Google’s Gemini has compelled OpenAI to strongly consider expediting the release of GPT-4 with multimodal capabilities. According to reports, Google may be releasing Gemini anytime now and OpenAI needs to buckle up. OpenAI is currently in the process of integrating GPT-4 with multimodal capabilities, much like what Google is planning with Gemini. This integrated model is expected to be named GPT-Vision, as per a recent report. The timing appears to be quite opportune, as both Gemini and GPT-Vision are expected to enter the scene and potentially compete against each other this fall. Although OpenAI CEO Sam Altman had earlier made it clear that one shouldn’t expect GPT-5 or GPT- 4.5 in the near future, however, as per The Information, OpenAI might follow up GPT-Vision with an even more powerful multimodal model, codenamed Gobi. Unlike GPT-4, Gobi is being designed to be multimodal from the start. It needs to be seen if OpenAI makes the right decision by clashing with Gemini. Many are anticipating that OpenAI may introduce a multimodal GPT-4 during their first-ever developer’s conference, OpenAI DevDay, which will be held on November 6 in San Francisco. Fingers crossed for multimodal GPT-4! https:\/\/t.co\/6PWadNnmsj— Monarch Wadia (@monarchwadia) September 10, 2023 Is GPT-Vision better than Gemini? OpenAI’s decision to withhold the multimodal capabilities does not stem from an inability to develop them. The ChatGPT creator has, in fact, collaborated with a startup called Be My Eyes, which is developing an app to describe images to the blind users, helping them interpret their surroundings so that they can interact with the world more independently. During this collaboration, OpenAI recognised that adding multimodal capabilities to GPT-4 at this stage might be premature as the integration of images could potentially raise privacy concerns. Moreover, there’s a risk of misinterpreting facial features such as gender or emotional state, which could result in harmful or inappropriate responses. Meanwhile, OpenAI has got its bases covered. A few months ago, reports came out that OpenAI is working on DALL-E 3.  Early samples leaked by YouTuber MattVidPro indicate that this model did much better than other image generators, including Midjourney, which is usually seen as the best for making realistic images. Interestingly, in a recent interview, Google chief Sundar Pichai, when asked what edge Gemini has over ChatGPT, replied, “Today you have separate text models and image-generation models and so on. With Gemini, these will converge.” This means that the most we can anticipate from Gemini is its ability to generate text and images based on user prompts. If OpenAI combines the capabilities of Dall E-3 and ChatGPT Plus, it is pretty much good to go against Gemini. To have an edge over GPT-4, Gemini is being trained on YouTube videos and would be the first multimodal model being trained on video rather than just text (or in GPT-4’s case, text plus images).  Moreover, Demis Hassabis recently claimed that engineers at DeepMind are using techniques from AlphaGo for Gemini On the other hand, Google’s Bard hasn’t been able to make a strong impression and falls short of ChatGPT when it comes to generating text. Thus, placing hope on Gemini to turn Google’s fortunes around is a huge bet. OpenAI can afford to risk it OpenAI’s process of shipping products is different from that of Google. Google, being an old and reputed player in the market with 4.3 billion customers worldwide, thinks twice and more before launching any product. It makes sure that the products are fully ready without any loose ends. On the other hand, OpenAI has shipped products in the past, even when they are not fully finished in the hope that consumer reviews will help them make necessary changes. Consider GPT-4, when OpenAI initially introduced it, they mentioned it would be multimodal. However, this didn’t turn out to be the case. Moreover, OpenAI acknowledged the limitations of GPT-4, stating that it still isn’t entirely dependable, often generating inaccurate information and making reasoning errors. Pichai expressed similar views during a recent interview when he noted that ChatGPT’s launch before LaMDA signalled Google that the technology of LLMs is well-suited for the market. He stated, “Credit to OpenAI for the launch of ChatGPT, which showed a product-market fit and that people are ready to understand and play with the technology”. It would be safe to say that with both Google and OpenAI striving to take the lead in the multimodal war, this fall will surely be interesting.","excerpt":"OpenAI ups the ante to challenge Gemini with GPT-Vision","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Siddharth Jindal","publish_date":"2023-09-19T18:53:54","publication_year":"2023","word_count":753,"keywords":["Go","ChatGPT","DALL-E","OpenAI","GPT-5","AI","GPT","Aim","AI Tool","R","startup"],"extracted_tech_keywords":["AI","GPT-5","ChatGPT","OpenAI","Aim","R","Go","GPT","DALL-E","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-time-is-ripe-for-the-real-gpt-4\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10135163,"title":"AWS Selects Seven Indian Startups for Global Generative AI Accelerator Program","content":"Amazon Web Services (AWS) has selected seven generative AI startups from India to join its prestigious Global Generative AI Accelerator program. India has the highest representation in the Asia-Pacific and Japan region, showcasing the country’s growing influence as an emerging AI hotspot. The chosen startups—Convrse, House of Models, Neural Garage, Orbo.ai, Phot.ai, Unscript AI, and Zocket—are part of 80 companies selected globally for their innovative use of AI and global growth potential. Each of these startups will receive up to US$1 million in AWS credits, mentorship from global experts, and technical support to accelerate their AI development. They will also gain the opportunity to showcase their solutions at AWS’s reevent in Las Vegas later this year. This selection is part of AWS’s US$230 million commitment to advancing generative AI technologies across the globe. Startups in the accelerator will benefit from access to AWS’s compute, storage, and database solutions, as well as energy-efficient AI chips like AWS Trainium and Inferentia2. The startups will also be able to leverage Amazon SageMaker for building and training foundation models, and Amazon Bedrock to develop generative AI applications securely. Amitabh Nagpal, Head of Startup Business Development at AWS India, highlighted the potential of the selected Indian startups. “We are thrilled to announce the seven Indian startups that have been selected. These companies are driving innovation in generative AI and will help solve complex challenges in India and beyond,” he said. The accelerator program lasts for 10 weeks and pairs startups with business and technical mentors from AWS and presenting partner NVIDIA, a leader in accelerated computing. The goal is to empower startups with the resources they need to develop, train, test, and scale their AI solutions effectively. India’s AI sector continues to expand, with over US$82 billion invested in AI and ML startups in 2024, according to PitchBook Data. AWS is committed to fostering this growth through initiatives like the AWS GenAI Loft, a pop-up space in Bangalore designed to promote generative AI innovation. Selected Startups from India for the AWS Global Generative AI Accelerator: Convrse (Haryana): AI tool optimizing 3D objects for web or real-time use. House of Models (Karnataka): Generates digital content using LLM and Diffuser models. Neural Garage (Karnataka): Uses AI to sync actors’ lips with dubbed audio seamlessly. Orbo.ai (Maharashtra): Provides AI-driven hyper-personalization in the beauty industry. Phot.ai (Haryana): AI platform for photo editing and graphic design, targeting e-commerce sellers. Unscript AI (Karnataka): Creates studio-quality videos with virtual or real actors. Zocket (Tamil Nadu): Simplifies digital ad creation and targeting with AI. AWS’s accelerator is expected to play a crucial role in shaping the future of AI startups, providing them with the tools and infrastructure needed to fulfill their ambitions on a global scale.","excerpt":"Amazon Web Services (AWS) has selected seven generative AI startups from India to join its prestigious Global Generative AI Accelerator program. India has the highest representation in the Asia-Pacific and Japan region, showcasing the country’s growing influence as an emerging AI hotspot. The chosen startups—Convrse, House of Models, Neural Garage, Orbo.ai, Phot.ai, Unscript AI, and […]","categories":["AI News"],"tags":["AI Startups","AWS"],"author_name":"Siddharth Jindal","publish_date":"2024-09-13T10:58:18","publication_year":"2024","word_count":453,"keywords":["Go","GenAI","Amazon SageMaker","AWS","AI","ML","RAG","generative AI","foundation models","R","AI Startups"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","foundation models","Amazon SageMaker","RAG","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-selects-seven-indian-startups-for-global-generative-ai-accelerator-program\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129254,"title":"AWS and iTNT Hub Collaborate to Nurture Generative AI innovation Among Startups in Tamil Nadu","content":"AWS India Private Limited has unveiled a new initiative in partnership with Tamil Nadu Technology (iTNT) Hub to establish a generative Artificial Intelligence (AI) startup hub program. This program aims to accelerate the development of generative AI solutions for public-centric initiatives through Tamil Nadu’s startup ecosystem.The program will facilitate collaboration between startups and industry to create public sector-focused solutions using generative AI. It will target startups working in AI, generative AI, and deep-tech domains, with a focus on applications for government, healthcare, education, and non-profit sectors.iTNT Hub, established by the Ministry of Electronics and Information Technology (MeitY) and the Information Technology and Digital Services Department (IT&DS) of Tamil Nadu, is located at Anna University in Chennai. It aims to create a deep tech innovation network in Tamil Nadu by leveraging the combined strengths of startups, innovators, academia, government, and industry leaders. Key features of the program include: Support for startups at various stages of development, from pre-incorporated teams to mature startups. Mentorship for founders from over 570 engineering colleges affiliated with Anna University. Access to research opportunities, sectoral guidance, and funding avenues for startups in incubation. AWS will provide eligible startups with up to $10,000 in AWS credits, allowing them to experiment with over 240 fully featured services, including innovative generative AI solutions like Amazon Bedrock, Amazon Q, and Amazon SageMaker. The company will also explore onboarding startups to the AWS Partner Network (APN), subject to eligibility criteria. iTNT Hub will offer additional support through: Technical expertise and mentorship Guidance on business and funding fundamentals Industry connections to understand market opportunities The program will feature webinars, technical workshops, masterclasses, industry connects, roadshows, hackathons, and customer engagements. A steering management committee comprising executives from AWS India and iTNT Hub will govern the program. This collaboration aims to strengthen the entrepreneurial environment for startups across Tamil Nadu, with a particular focus on empowering those in locations with limited access to resources. By leveraging AWS’s global technology leadership and iTNT Hub’s local expertise, the program seeks to foster innovation and growth in the generative AI space while addressing public sector needs.","excerpt":"AWS has unveiled a new initiative in partnership with Tamil Nadu Technology (iTNT) Hub to establish a Gen AI startup hub program.","categories":["AI News"],"tags":["Generative AI","Startups"],"author_name":"MIA","publish_date":"2024-07-16T12:47:32","publication_year":"2024","word_count":348,"keywords":["Go","Amazon SageMaker","artificial intelligence","AWS","AI","Git","RAG","Aim","Startups","generative AI","Generative AI","R"],"extracted_tech_keywords":["AI","artificial intelligence","generative AI","Aim","Amazon SageMaker","RAG","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-and-itnt-hub-collaborate-to-nurture-generative-ai-innovation-among-startups-in-tamil-nadu\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":23731,"title":"Move Over Bengaluru, These Tier-II Cities Will Soon Catch Up As The Next AI Hubs","content":"Andhra Pradesh Government led by Chandrababu Naidu is committed to establish Vizag as an AI and industrial hub So far, the development of artificial intelligence startup ecosystem, academic research impetus and government support has largely been concentrated in India’s Silicon Valley — Bengaluru, a city weighed down by infrastructure woes. Meanwhile Hyderabad rivals the city in terms of funding, and is also known for its T-hub and incubator programs kick-started by the state government. Hyderabad has a strong ecosystem and receives a massive impetus from state-backed policies. In fact, Andhra Pradesh and Telangana are engaged in a heated battle to emerge as the next startup capitals of India. But now, the conversation has shifted to other Indian cities that also offer a rich talent pipeline, are home to top-tier research universities, have a cluster of academicians, and can attract venture capital as well. Besides, these cities can also attract government support for economic stimulus. Let’s List Down A Few Emerging AI Hubs In India Thiruvananthapuram and Cochin: In a move that could pitch Kerala as a competitor to Karnataka, Andhra Pradesh and Telangana as a global hub of IT technologies, Innovation Incubator, a New York-based company plans to set up Cognitive Design Institute at Technopark in Thiruvananthapuram to build a talent pipeline in design capabilities and incubate AI startups, a news report indicated. Kerala government is striving earnestly to establish itself as a global digital hub in a bid to create jobs and attract global companies. Leading IT behemoths are the change agents in AI, fueling economic growth and creating jobs, and the Kerala government has been actively working with key stakeholders for skillset upgradation and building an active startup ecosystem. Another high point is that Japanese carmaker Nissan is mulling to set up its global digital hub in a bid to meet the demand of future cars. Besides Thiruvananthapuram, there are several industry initiatives being supported by the Kerala government in Kochi to inspire more participation from leading enterprises and make the state ready for the digital revolution. Interestingly, Kerala had launched a Kerala Startup Mission way back in 2006 to foster innovation and build government-backed entrepreneurial ventures. Vizag: Up next on the list is Vizag which is hailed as a hotspot for FinTech companies. Noted companies which are leading the charge, such as Paytm, Visa and Thomson Reuters, have operations in the city. The city is backed by an impeccable infrastructure, such as a Tech Hub which was recently inaugurated, and features plug and play facilities as well. According to JA Chowdary, who is a Special Chief Secretary and IT Advisor to the Chief Minister of AP, concrete efforts have been taken to strengthen the FinTech ecosystem by inviting startups from all over the world, and offering a slew of incentives to investors. He added that the focus was to develop emergent technologies like blockchain to position AP as a digital state. Vizag’s FinTech Valley is a state-sponsored accelerator program and also features a repository of BFSI use cases. The FinTech Valley brings together several key stakeholders such as start-ups, financial institutions, technology vendors, incubators, accelerators, innovation labs, and investors, to build and strengthen the ecosystem. Some of the leading FinTech startups leading the charge are iProov, Moneytor, Gyandhan, FingPay, Incremint, TAQBit and Nanobi, among others. Jaipur: The Pink City is not just hailed for its tourism and literature festivals, but is also fast emerging as an IT startup hub. Home to startups like RazorPay and CarDekho, Jaipur has an impeccable track record with 576 startups and 183 startup jobs listed on the AngelList portal. Besides SaaS and e-commerce players, Jaipur also boasts of Yeppar, AR and VR startup, Peak, UK-based AI and ML startup which has a development centre in Jaipur and Metacube. One of the key growth factors for Rajasthan is Startup Oasis, jointly set up by IIM-Ahmedabad’s Centre for Innovation, Incubation and Entrepreneurship and RIICO. The Startup Oasis was set up to focus on social impact entrepreneurship and has since launched several startups that address socio-economic issues related to water, healthcare and energy in the state. Last year, Jaipur also become the first city in South Asia to monitor and operate smart city infrastructure via a single network. According to reports, the Network Operating Centre at Jaipur Development Authority (JDA) will integrate digital infrastructure such as sensor outfitted dustbins, intelligent kiosks, wireless broadband, smart streetlights and smart parking. Ahmedabad: If there is one city that has been hogging the limelight for IoT, FinTech and SaaS startups, it is Ahmedabad. Earlier in February, Ahmedabad-based startup FinTech company, Lendingkart raised $87 million in its Series C round of funding. The city is also home Ecolibrium Energy the startup founded by Chintan Soni and Harit Soni, which caters to the energy market. The startup’s big data analytics platform is aimed at enhancing the operational efficiency by optimising energy usage and asset utilisation. Besides incubator programs and investor meets, of all the Tier-II cities Ahmedabad tops for having India’s most prolific entrepreneurs, change-makers and experts. The city also hosted a third edition TiEcon Ahmedabad 2018, held in March which saw industry participation from companies like Quick Heal, NASSCOM, Matrix and Havmor. The summit focused on building fintech innovations, ML, AI, and Industrial Revolution 4.0.","excerpt":"So far, the development of artificial intelligence startup ecosystem, academic research impetus and government support has largely been concentrated in India’s Silicon Valley — Bengaluru, a city weighed down by infrastructure woes. Meanwhile Hyderabad rivals the city in terms of funding, and is also known for its T-hub and incubator programs kick-started by the state […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-04-16T08:37:58","publication_year":"2018","word_count":875,"keywords":["big data","Go","API","artificial intelligence","AI","ML","Git","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","Aim","R","Go","Git","API","big data"],"url":"https:\/\/analyticsindiamag.com\/it-services\/move-over-bengaluru-these-tier-ii-cities-will-soon-catch-up-as-the-next-ai-hubs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":14532,"title":"InMobi turns profitable – Making the most of data with machine learning","content":"InMobi co-founder and CEO Naveen Tewari Mobile advertising company InMobi grabbed headlines earlier this year when the Indian unicorn declared, after a decade in business, it turned operationally profitable in 2016. Set up in 2006, the Indian data-driven company provided much cheer to the Indian startup ecosystem plagued by takeovers.  In an earlier interview with Analytics India Magazine, Srikanth Sundarrajan, the Principal Architect with Inmobi, revealed how big data and analytics formed the backbone of the mobile advertising platform. Here, we detail how artificial intelligence, specifically machine learning played a key role in this data-centric organization, hailed as one of the best Indian unicorns.  Essentially, InMobi solutions address the app developer’s needs across the entire app lifecycle, from app discovery and distribution to user acquisition, monetization, engagement, and retention, delivering real value throughout the app lifecycle, according to the the book apponomics. The ad-tech platform that captures and uses real-time mobile analytics from about 1.5 billion unique mobile devices to identify insights like the best time of day to serve an ad, or the best format to achieve a rendered ad (not simply a served ad).  In 2014, Abhishek Bapna, then Product Manager at InMobi, revealed the core of search had moved from keyword to social networks. But in an app ecosystem, neither of them are really useful to understand user engagement and life-time value. What is most important is the ability to understand their app behavior? Human cognition in algorithms Bapna explained how algorithms are a manifestation of how human beings think and operate. He revealed how InMobi, spent a lot of time trying to understand how human cognition works. Citing some key points, he shared how the app ecosystem works differently from the web ecosystem and what acts as a trigger in the app world.  So, how does one drive user acquisition or user engagement and give long term value in apps? The ability to understand users in their ad behaviour is most important in the app ecosystem and when it comes to ad behaviour, it is not always driven by intent. To understand why users download and behave in a certain way and for that you have to go from decision to downloading to first action and uncover what triggers those nuances that leads to actions. Big differentiator from the web world where everything was HTML-based and the whole web was seamless thing but in the app ecosystem, there are all barriers to taking actions because it is far more interactive and engaging. At InMobi, understanding algorithms are a manifestation of how humans think and we spend a lot of time to understand how human cognition works. Our algorithms are trying to mimic how human cognition behaves. Why machine learning in marketing trumps traditional marketing In a recent article by Preetham V V is VP Data Sciences, Deep Learning at InMobi, machine learning, enables brands to have a deliver a targeted, coherent approach in “engaging consumers with a consistent voice, tailored to individuals across omnichannel end-points.” As compared to traditional marketing, machine learning can be leveraged to improve channel efficacy by a minimum of 80 percent and unbounded maximums if done right. Even simple deep learning and machine learning techniques on KPIs like click-through-rates, conversion rates and bid-to-win ratios can bump the efficacy by 50 percent to 200 percent. Preetham further added,that machine learning models thrive on data and dimensions about the user go beyond demography. Dimensions used are users marketing channel affinity, past attributes, view-throughs, the frequency of visit per channel, average opportunity to see, etc. can all be computed through other machine learning techniques. Can data science enhance creatives? Rajiv Bhat, Senior Vice President of Data Sciences and Marketplace at InMobi According to Rajiv Bhat, Senior Vice President of Data Sciences and Marketplace at InMobi, who is responsible for all things related to machine learning, when it comes to devising advertising campaigns, analyzing consumer behaviour should come first. In his article, Bhat explains, how data should trump creative.  He cites that marketers should deploy “techniques from data science, statistics and artificial intelligence to analyze structured and unstructured data, discover patterns and relationships to make predictions about future outcomes and events.” Driving a data-driven strategy allows companies to take advantage of real-time engagement opportunities and cash in on breaking trends. “Creatives can have different half-lives; understanding the longevity of creatives and making that part of the creative strategy is key,” he shared. Focus","excerpt":"Mobile advertising company InMobi grabbed headlines earlier this year when the Indian unicorn declared, after a decade in business, it turned operationally profitable in 2016. Set up in 2006, the Indian data-driven company provided much cheer to the Indian startup ecosystem plagued by takeovers.  In an earlier interview with Analytics India Magazine, Srikanth Sundarrajan, the […]","categories":["IT Services"],"tags":["best book to learn digital marketing"],"author_name":"Richa Bhatia","publish_date":"2017-04-25T03:54:50","publication_year":"2017","word_count":737,"keywords":["data science","Go","machine learning","artificial intelligence","AI","ML","RAG","deep learning","analytics","best book to learn digital marketing","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/inmobi-turns-profitable-making-data-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":1950,"title":"Vodafone touches 50 Million connections through IoT services","content":"Vodafone is currently celebrating its feat of achieving 50 million connections for its Internet of Thing services. With this, it has become the first IoT mobile provider globally, to have achieved this mammoth numbers. According to the company’s blogpost, it is also demonstrating industry-leading growth of around one million new connections a month. The company has particularly displayed strong performances in the automotive, healthcare and utilities sectors, in countries including India. Vodafone uses the Group’s own operating networks, partner market networks and other third-party mobile network operators to provide customers with simple and globally unified IoT technologies and connected services, spread across nearly all the countries. The company firmly believes that IoT technologies have now begun to reach critical masses as well as enterprises, where they are focusing on R&D for a wide range of customer technology companies. According to the most recent Vodafone IoT Barometer report 76% of businesses say that IoT will be ‘critical’ for the future success of any organization in their sector. There other numbers suggested that 24% of IT businesses are being allocated to IoT, 48% of IoT adopters are using IoT to support large-scale business transformation and about 86% of businesses in the industrial sector have seen significant return from IoT implementation. The Group is now acknowledged to be the market leader in mobile IoT worldwide and has been recognized by the Gartner as a ‘Leader’ in its Magic Quadrant for Managed M2M Services, Worldwide 2016, for the third consecutive year and as ‘the consistent global leader’ in IoT services by Machina Research in its 2016 IoT Communications Service Provider Benchmarking report. It is also worth noting that since the inception of Vodafone’s machine-to-machine (M2M) technologies in 2011, it has experienced double-digit annual growth in IoT revenue. Vodafone Director of IoT Ivo Rook said “Vodafone was one of the earliest believers in the potential for the Internet of Things to transform business and life. The 50 million milestone is a testament to our continued focus and commitment to innovation in this rapidly growing and dynamic sector.” Vodafone has mobile operations in 26 countries, partners with mobile networks in 49 more, and fixed broadband operations in 17 markets. Vodafone had 470 million mobile customers and 14.3 million fixed broadband customers as of 31 December 2016.","excerpt":"Vodafone is currently celebrating its feat of achieving 50 million connections for its Internet of Thing services. With this, it has become the first IoT mobile provider globally, to have achieved this mammoth numbers. According to the company’s blogpost, it is also demonstrating industry-leading growth of around one million new connections a month. The company […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-04-27T11:15:09","publication_year":"2017","word_count":380,"keywords":["API","programming_languages:R","AI","innovation","ML","Git","GAN","R"],"extracted_tech_keywords":["AI","ML","R","Git","API","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vodafone-touches-50-million-connections-iot-services\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075563,"title":"Shouldn&#8217;t Linux Foundation be Officially Present in India?","content":"A lot has happened at the Linux Foundation over the past few days. First, Meta CEO-co founder Mark Zuckerberg announced that its PyTorch project will now be part of the non-profit foundation as a newly launched PyTorch foundation. A few days ago, the organisation announced OpenWallet Foundation (OWF). Built on an open-source platform, this new collaborative initiative supports interoperability across different digital wallets. Further, the Linux Foundation recently announced the launch of Linux Foundation Europe to build on its growing presence in this part of the world. The announcement was made during its first survey of Europe’s open-source community at the keynote addresses of Open Source Summit, Europe. This makes one wonder why India, one of the largest countries in terms of number of developers, is yet to have an official presence of The Linux Foundation. The Indian market is in focus in the backdrop of the recent launch of OpenWallet. In a comment, Pramod Varma, the chief architect of Aadhaar and India Stack, said, “India has been at the forefront of it and is going all out to convert all physical certificates into digitally verifiable credentials via the very successful Digilocker system. I am very excited about the OWF effort to create an interoperable and open-source credential wallet engine to supercharge the credentialing infrastructure globally.” The Linux Foundation’s growing reach Enterprise tech is undergoing rapid transformation and open-source software is the turbine driving the transformation. The Linux Foundation has benefited greatly from this boom and has created a shared value of $54.1 billion from its community’s collective contributions. Linux Foundation has outdone its competitors like Eclipse Foundation and Apache Software Foundation in terms of budget and reach. The foundation is currently handling over 100 projects across sectors, including artificial intelligence, autonomous vehicle, networking, and security. The organisation has further branched out to form Cloud Foundry Foundation, Open Source Security Foundation, and Cloud Native Computing Foundation. What about the Indian market? In several interviews, Jim Zemlin, executive director of the Linux Foundation, has said that he sees a tremendous opportunity for developers in India. In a 2015 blog, the foundation said that India was second only to the United States in inquiries regarding Linux Foundation training and certification but ranked among the lowest in Linux certification exam enrollment. Owing to this, the same year, Linux Foundation announced the availability of country-specific pricing for its Essentials of System Administration course and Linux Foundation Certified System Administrator exam for individuals in India. India had then become the first country to have region-specific pricing. Further, several India-based companies like Wipro, Hasura, HCL, Reliance Jio, Tata Communications, and Tech Mahindra are member companies of The Linux Foundation. India – leading contributor to open-source GitHub was launched in 2020, and within a year, there were close to 2 million new developers on the platform. The company representative then said that the platform would be adding 10 million developers by 2023. GitHub has credited India as being the fastest-growing country in the world in terms of developers contributing to open-source projects. The company announced an Open Source Grant for Indian developers and said that it would be offering Rs 1 crore to select developers. Open-source software development has great potential in India, and it could become the best in the world given the right circumstances. The Indian government is also giving open-source tech the necessary push. In 2015, the Ministry of Communication and Information Technology, rolled out three major policies related to open-source. These stats and reports prove that India has a large base of developers and a lot of companies are turning to India for coding talent. As mentioned earlier, the government is also pushy when it comes to open sources. The Linux Foundation should truly consider launching a dedicated arm for India.","excerpt":"India, which is one of the largest countries when it comes to a number of developers, is yet to have an official presence of The Linux Foundation","categories":["IT Services"],"tags":["GitHub","Linux","Linux Foundation","Open Source"],"author_name":"Shraddha Goled","publish_date":"2022-09-19T14:00:00","publication_year":"2022","word_count":628,"keywords":["Go","API","artificial intelligence","Open Source","AI","PyTorch","ML","Git","CLIP","Linux","Linux Foundation","GitHub","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","PyTorch","R","Go","Git","GitHub","API","CLIP"],"url":"https:\/\/analyticsindiamag.com\/it-services\/shouldnt-linux-foundation-be-officially-present-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10014640,"title":"Top 10 Programmers To Follow on Youtube","content":"Youtube has become a major source of information and an educational platform for many. It has an array of lessons and tutorials to learn any subject, including topics in computer science. Unlike subscription-based ed-tech models, most of the content on it is free. Programmers, enthusiastic about teaching data science, artificial intelligence, machine learning, and deep learning are also very active on the platform and have well-established channels with videos right from the basics of programming to complex subjects of the field. We enlisted the top programmers teaching these subjects on YouTube. Choosing the right one will depend on your experience in the field and the kind of projects you want to work on. Best Programmers to Follow on Youtube1. Krish Naik2. Abhishek Thakur3. Harrison Kinsley4. Srivatsan Srinivasan5. Dhaval Patel6. Kevin Markham7. Code With Harry8. Ken Jee9. Keith Galli10. Corey Schafer 1. Krish Naik Subscribers: 278,000 Krish Naik is the Chief Innovation Officer at iNeuron and has over eight years of industry experience. He runs his own YouTube channel, with around 278K subscribers, that explains various topics on machine learning, deep learning, and AI using real-world scenarios. Not only does he make step-by-step tutorials of these topics to apply them in real-world problems, but he also hosts Q&A sessions on his channel to address some of the complex questions of the field. Check out the YouTube channel here. 2. Abhishek Thakur Subscribers: 35,800 A 4x Kaggle grandmaster, Abhishek Thakur makes videos about applied machine learning, deep learning and data science. Along with making walkthrough videos to learn the basics of programming for data science, his channel hosts tutorials with experts on complex topics like machine learning with PyCaret, machine learning APIs, CycleGANs among others. Check out the Youtube channel here. 3. Harrison Kinsley Subscribers: 987,000 Closing in on almost a million subscribers Harrison Kinsley runs a channel by the name Sentdex that hosts tutorials on Python programming. The channel includes programming basics in Python followed by data analysis, machine learning, and deep learning tutorials. The channel also has data science tutorials for applications in specific sectors like finance and robotics. Some of the latest uploads include videos on complex applications like quantum computing, self-driving cars, and facial recognition. Check out the Youtube channel here. 4. Srivatsan Srinivasan Subscribers: 31,100 With an industry experience of 17 years, Srivatsan Srinivasan, who is currently working as a Chief Data Scientist at Cognizant, posts everything about ‘data’. He covers an overview and implementation of topics on machine learning, artificial intelligence, data engineering, development operations, and cloud in his programming tutorials. Along with it, he also covers a lot of business use cases. One of his playlists is a ‘Machine Learning Bootcamp’ that hosts live tutorials in collaboration with DPhi, a data science community. Check out the YouTube channel here. 5. Dhaval Patel Subscribers: 208, 000 With 15 years of experience as a software engineer, Dhaval Patel thinks that ‘anyone can code’. His YouTube channel, Codebasics has 208K subscribers and encompasses tutorials right from the basics of coding to more complex topics like artificial intelligence, machine learning and deep learning. His preferred language is Python, and he has entire playlists dedicated to teaching important libraries. At the same time, he also has playlists for data science applications across sectors like real estate and sports. Patel also runs a sister channel of Codebasics in Hindi that has around 4.5K subscribers. Check out the YouTube channel here. 6. Kevin Markham Subscribers: 158,000 Kevin Markham’s channel is perfect for users who are just getting started in data science and willing to develop a strong foundational base. His channel Data School, with 158K subscribers, focuses on topics that one ‘needs to master first’. These are in-depth, hands-on programming tutorials to understand the basics. He also has playlists that introduce you to machine learning. You can start with his tutorials irrespective of your educational background. Check out the YouTube channel here. 7. Code With Harry Subscribers: 870,000 This YouTuber, who doesn’t disclose his name anywhere in his bio, has videos to start learning coding from the very basics, in Hindi. He has well-organised playlists for many languages including Java, C, C++, Python, JavaScript, among others. He also has an entire playlist for machine learning tutorials in Python for beginners. Check out the YouTube channel here. 8. Ken Jee Subscribers: 107,000 Ken Jee is a data science and sports fan who has been working in the data science industry for more than five years. He has 107K subscribers and produces content that falls in the intersection between data science and sports. His sports analytics playlist has interesting tutorials providing hands-on applications like a simulation of NBA games or NBA season win predictions using linear regression. He also hosts a podcast on the channel talking to some of the best minds in sports analytics. Check out the YouTube channel here. 9. Keith Galli Subscribers: 90,000 A recent MIT graduate, Keith Galli likes to make videos on computer science, programming, and board games. His data science playlist includes tutorials on Python libraries including NumPy, Pandas, Scikit, and Matplotlib to get you started. His channel has almost 90K subscribers. If you fancy, his board games tutorials are fun too. Check out the YouTube channel here. 10. Corey Schafer Subscribers: 682,000 Corey Schafer’s channel focuses on creating tutorials and walkthroughs for software developers, programmers, and engineers. It covers topics for beginners as well as experienced professionals. For data science, he has a playlist dedicated for libraries in Python including Pandas and Matplotlib.Check out the YouTube channel here.","excerpt":"Youtube has become a major source of information and an educational platform for many. It has an array of lessons and tutorials to learn any subject, including topics in computer science. Unlike subscription-based ed-tech models, most of the content on it is free. Programmers, enthusiastic about teaching data science, artificial intelligence, machine learning, and deep […]","categories":["AI Features"],"tags":["ai lessons for beginners","beginner python projects","Deep Learning","deep learning projects","Interviews and Discussions","learning data science","Machine Learning","YouTube"],"author_name":"Kashyap Raibagi","publish_date":"2020-12-16T17:00:00","publication_year":"2020","word_count":921,"keywords":["data science","NumPy","artificial intelligence","machine learning","AI","learning data science","Machine Learning","Ray","deep learning","ai lessons for beginners","analytics","deep learning projects","Deep Learning","YouTube","Matplotlib","Pandas","beginner python projects","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","Ray","Pandas","NumPy","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-programmers-to-follow-on-youtube\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161842,"title":"Do Tech Founders Miss Coding as They Transition to Leadership?","content":"A study posted by Tinder staff data scientist Conor McLaughlin analysed over 500 startup founders and found that approximately 88% had completed an undergraduate education, with computer science emerging as the most popular field of study. For many of these founders, coding is not just a skill but a defining part of their identity. It is a tool through which they built their companies, tackled early challenges, and transformed ideas into reality. As startups grow, founders are often required to leave behind this hands-on work to focus on leadership. This transition is both inevitable and deeply personal, which forces founders to balance their technical passions with the demands of scaling a company. For tech founders, writing code often represents control and mastery, offering a direct line from idea to execution. How Coding Has Changed Over Time For OpenAI CEO Sam Altman, coding will always be of foundational relevance in the future. “I think you should learn to code. I don’t write code very much anymore, but I randomly did [it] yesterday. Learning to code was great as a way to learn how to think,” Altman said in a recent podcast. Meanwhile, Ritwika Chowdhury, founder and CEO at Unscript, said, “I started coding even before getting into IIT. The beautiful part about coding is that it allowed me to rationalise problems and statements from first principles. This quality seeped into my startup journey. I’ve always chosen problems that I faced personally and solved them from first principles.” However, the coding landscape has transformed dramatically, particularly with the rise of AI and automation. Research from last year revealed that AI tools increased developer efficiency by 26% and code completions by 38%. Founders who started in the pre-AI era operated in a world that demanded technical depth and manual effort.  In contrast, today’s developers work in an AI-assisted environment. “AI has completely transformed the way I approach coding and productivity,” Adithya S Kolavi, co-founder of CognitiveLab, said. “It not only helps me write and deploy code faster but also acts as a tool for quickly learning new concepts and technologies in an intuitive, conversational way.” As a young founder, Kolavi appreciates how AI simplifies juggling the many hats founders wear – from engaging with investors and building marketing strategies to growing a social presence and generating innovative designs and ideas. “AI has been a game-changer,” he added, highlighting how it improves his creativity in ways traditional tools never could. Tools like GitHub Copilot, ChatGPT, and automated debugging platforms streamline the development process. Research from last year revealed that GitHub Copilot and GPT-3.5 increased developer efficiency by 26% and code completions by 38%. On the other hand, AI increases competition, pushing founders to meet elevated expectations and stand out in crowded markets. “With AI, building and shipping MVPs has become easier and faster,” Chowdhury explained. “However, this ease has led to increased competition, with customers evaluating multiple demos. Expectations for early-stage products are now higher, bringing challenges back to square one.” Other founders reflect the same sentiment. “Coding today is vastly different from when I first started 18 years ago (I was 12 back then!). The rise of LLMs has revolutionised how we approach programming. Now, you have AI tools that can instantly guide you through challenges. If you ask me what’s changed the most, I’d say the time taken to push out an MVP has decreased by almost 90%,” added Aravind Jayendran, co-founder at LatentForce.ai. Leadership in the Age of AI The rise of AI and automation is reshaping not only how technology is built but also what it means to lead a tech company. For pre-AI founders, stepping back feels like leaving an irreplaceable craft, while AI-native founders see coding as system orchestration. “What coding does is it gives you a systematic framework of breaking down a problem into smaller problems and attacking each with high focus. Leadership is just an extension of that. If you’re building a business, you need to prioritise — figure out what can move the needle,” said Sudipta Biswas, co-founder at Floworks, as he explained how leadership picks up where coding leaves off. While stepping into leadership offers the chance to shape industries and impact millions, it often requires letting go of the hands-on work that first inspired a founder’s journey. Adding to this perspective, Chowdhury emphasised the importance of applying a first-principles approach to leadership. “Leadership is more about distributed computing. You allocate and empower your team members, providing them as much visibility as possible and then finally bringing everything together in a cohesive manner to build a magical product.” What Does the Future Look like? Speaking on the future of the workforce, Biswas said, “AI allows you to move from zero to one with a lean workforce. What 100 people did before, 10 can do today with the right AI tools. According to him, leadership in the age of AI is about ensuring employees don’t resist it but instead embrace it as a tool to amplify their productivity. Salesforce CEO Marc Benioff recently revealed plans to halt the hiring of software engineers in 2025 and attributed the decision to a 30% increase in productivity driven by AI tools like Agentforce. Similarly, Meta CEO Mark Zuckerberg has expressed confidence that AI will be capable of handling mid-level engineering tasks by the same year. Despite fears of job displacement, research suggests that AI can improve job satisfaction among developers. Surveys reveal that 60-75% of developers feel more fulfilled and less frustrated when using tools like GitHub Copilot. Upsilling is the way to go, but so is working with AI. Altman highlighted that qualities like resilience, adaptability, quick ability to learn, creativity, and being comfortable with the right tools will be key to thriving in the future.","excerpt":"Coding will always be of foundational relevance in the future.","categories":["Global Tech"],"tags":["AI Developers","AI in Coding","AI leaders","Women in AI"],"author_name":"Aditi Suresh","publish_date":"2025-01-21T09:50:34","publication_year":"2025","word_count":959,"keywords":["Go","ChatGPT","AI Developers","GitHub","OpenAI","AI","ML","distributed computing","Git","AI in Coding","Rust","AI leaders","R","Women in AI"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","distributed computing","R","Go","Rust","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/do-tech-founders-miss-coding-as-they-transition-to-leadership\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10169130,"title":"Tech Mahindra, KOGO Partner to Build Private AI for Enterprises","content":"KOGO AI and Tech Mahindra have entered into a global strategic alliance to jointly develop and deliver next-generation enterprise AI solutions focused on autonomy, scalability, and regulatory compliance. The partnership combines Tech Mahindra’s enterprise transformation capabilities with KOGO AI’s Agentic AI infrastructure to enable organisations worldwide to implement Private AI frameworks tailored to their unique operational and regulatory requirements. “The future of enterprise AI lies in autonomy, compliance, and control,” said Praveer Kochhar, co-founder & CPO at KOGO AI. “This partnership with Tech Mahindra brings the best of both worlds—an enterprise-grade Agentic AI stack and a transformation partner trusted by some of the world’s largest companies. Together, we’re making Private AI real and scalable to meet the realities of modern enterprise systems.” Both companies will jointly identify high-impact use cases, develop Agentic AI solutions, and scale them across customer operations. The initiative will prioritise building intelligent AI agents for diverse enterprise scenarios, designing Private AI architectures that can function across on-premise, hybrid, or secure cloud environments, and integrating with legacy systems to modernise infrastructure without replacement. The alliance targets key sectors including BFSI, healthcare, manufacturing, and the public sector through domain-specific AI implementations. “Organisations are shifting from pilot projects to implementing compliant and secure AI integrated into their core operations. Our partnership with KOGO AI will deliver scalable Private AI solutions, creating impactful, domain-specific agents that transform enterprise AI from experimentation to effective execution,” said Birendra Sen, president – business process services, Tech Mahindra. Initial deployments are already underway in BFSI and healthcare, with a focus on real-world use cases such as fraud detection, service automation, auditing, AI assistance, and enhancing operational efficiency. The alliance aims to develop AI agents that seamlessly integrate into existing enterprise ecosystems. KOGO AI and Tech Mahindra are also working closely with global clients to identify transformation opportunities, customise AI stacks for specific workflows, and establish governance and observability frameworks for AI performance. The collaboration spans North America, Europe, APAC, and the Middle East, aiming to provide Private AI solutions that maintain local compliance and data sovereignty while enabling global scale. KOGO recently also partnered with Qualcomm to develop a full-stack private AI solution tailored for high-performance and compliant enterprise deployments. The collaboration combines KOGO AI’s modular Agentic Platform—designed for deploying intelligent agents and compliance-ready workflows—with Qualcomm’s Cloud AI 100 Ultra accelerators and AI Inference Suite. Addressing a question at Tech Mahindra recent earnings call, CEO Mohit Joshi highlighted the company’s commitment to AI and its involvement in building foundational model. Joshi highlighted the company’s role in shaping sovereign AI capabilities through Project Indus, its LLM for Indic languages. “Indus was a critical capability demonstrator for us,” Joshi said, emphasising the gap in AI models that effectively support Hindi and other Indian languages. “There was a huge need for a model that catered to our linguistic diversity, and the ecosystem has responded very positively.”","excerpt":"Both companies will jointly identify high-impact use cases, develop Agentic AI solutions, and scale them across customer operations.","categories":["AI News"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-05-05T15:35:52","publication_year":"2025","word_count":479,"keywords":["Go","API","agentic AI","AI","ML","Scala","Aim","Rust","R","fraud detection"],"extracted_tech_keywords":["AI","ML","agentic AI","Aim","fraud detection","R","Go","Rust","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tech-mahindra-kogo-partner-to-build-private-ai-for-enterprises\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172505,"title":"NVIDIA H20 Chip Shortage Delays DeepSeek R2 Launch","content":"The launch of DeepSeek’s upcoming model, R2, could face significant setbacks in China as US export restrictions choke the supply of NVIDIA’s H20 chips, critical for running the company’s AI models, reported The Information. R2, the successor to DeepSeek’s widely used R1, has yet to receive a release date. The report added that CEO Liang Wenfeng is unsatisfied with its performance, and engineers continue to work on improvements before it is cleared for launch. Cloud providers that host and distribute DeepSeek’s models warn that existing inventories of NVIDIA chips will likely fall short of meeting the demand R2 could generate, particularly if it performs better than current open-source alternatives. These concerns have intensified following the April ban on NVIDIA’s H20 chip, which was specifically built for the Chinese market after earlier export restrictions barred the sale of more powerful Hopper series GPUs. During the recent earnings call, NVIDIA CFO Colette Kress said the company’s outlook reflects a loss of approximately $8 billion in H20 revenue for the second quarter. R1 and R2 are tightly optimised to run on NVIDIA’s architecture, making substitution with Chinese-developed chips difficult and inefficient. According to the report, employees at Chinese cloud companies said DeepSeek’s models “are so completely optimised for NVIDIA’s hardware and software” that deploying them on domestic alternatives is not viable at scale. Despite the ban, some Chinese companies have found a workaround to obtain NVIDIA hardware. According to The Wall Street Journal, engineers from Chinese AI companies are heading to Kuala Lumpur, Malaysia, with hard drives packed with instructions and data to train AI models. They then utilise the NVIDIA chips available at Malaysian data centres to train the model and return it to China. Meanwhile, to cope with chip shortages, some Chinese firms have resorted to using gaming GPUs like NVIDIA’s RTX 5090 and 4090, which are also under export restrictions but easier to obtain through grey markets. DeepSeek, backed by hedge fund firm High-Flyer Capital Management, made headlines for training R1 with less compute than US competitors like OpenAI. In response to the surge in R1 usage, major Chinese tech firms, including ByteDance, Alibaba, and Tencent, placed $16 billion worth of orders for 1.2 million H20 chips in early 2025, according to SemiAnalysis estimates. That compares with the 1 million chips shipped to China by NVIDIA last year. Despite these efforts, the scalability of R2 in China could be limited. Companies outside China, not constrained by US chip curbs, may find it easier to deploy the model at full scale.","excerpt":"CEO Liang Wenfeng is unsatisfied with R2’s performance, and engineers continue to work on improvements before it is cleared for launch.","categories":["AI News"],"tags":["DeepSeek"],"author_name":"Siddharth Jindal","publish_date":"2025-06-27T13:14:26","publication_year":"2025","word_count":420,"keywords":["API","OpenAI","AI","programming_languages:R","Scala","DeepSeek","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","OpenAI","R","Scala","API","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-h20-chip-shortage-delays-deepseek-r2-launch\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":42357,"title":"Hands-On Guide To Multi-Label Image Classification With Tensorflow &#038; Keras","content":"Image classification is one of the trending applications in machine learning. It has a wide range of applications — from facial recognition algorithms to identifying complex patterns in images like crime detections, and many other social, medical and technical applications. Image classification is not just about classifying images into categories, it has a broader and deeper meaning of giving machines the power to visualize the world. In this article, we will play around with a simple Multi-label classification problem. We will use the power of Tensorflow and the simplicity of Keras to build a classifier that is able to categorize the images of cats and dogs and also to identify their respective breeds. Pre-Requisites : A basic understanding of Convolutional Neural Networks. Read some of our previous articles on Convolutional Neural Networks to have a good understanding before we get our hands dirty. Follow the links below: Learn Image Classification Using CNN In Keras With Code Overview Of Convolutional Neural Network In Image Classification Getting the Dataset Head to MachineHack, sign up and start the Who Let The Dogs Out: Pets Breed Classification Hackathon. The datasets can be downloaded from the attachments section. The training set consists of 6206 images of both cats and dogs of different breeds. We will use these images and their respective classes provided in the train.csv file to train our classifier to categorize a given image as either the image of a cat or a dog and also classifying into respective breeds. Multi-Label Image Classification With Tensorflow And Keras Note: Multi-label classification is a type of classification in which an object can be categorized into more than one class. For example, In the above dataset, we will classify a picture as the image of a dog or cat and also classify the same image based on the breed of the dog or cat. Multi-class classification is simply classifying objects into any one of multiple categories. Such as classifying just into either a dog or cat from the dataset above. Importing Tensorflow and Keras import tensorflow as tf from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Dense, Conv2D, MaxPooling2D, Flatten import numpy as np import matplotlib.pyplot as plt import pandas as pd print(tf.__version__) You should be able to see the version of your TensorFlow as the output. Preparing the training data To feed the images into the Neural Network we would require the images to be loaded. We are provided with a train.csv file consisting of the image names and the respective categories. We can use the sheet to load the images using the flow_from_dataframe method from Keras, but the method requires the complete filename with the extension of the image. Since we have jpg images we will format all the id’s in train.csv by adding ‘.jpg’ to all rows. We will then create a new training set with 3 columns namely Images, Animal and Breed. Since the Animal and Breed columns are categories we will convert the type to string. training_set = pd.read_csv(\"Cats_and_Dogs\/Dataset\/train.csv\") training_imgs = [\"{}.jpg\".format(x) for x in list(training_set.id)] training_labels_1 = list(training_set['class_name']) training_labels_2 = list(training_set['breed']) training_set = pd.DataFrame( {'Images': training_imgs,'Animal': training_labels_1, 'Breed' : training_labels_2}) #Changing the type  to str training_set.Animal = training_set.Animal.astype(str) training_set.Breed = training_set.Breed.astype(str) Creating a new category One of the simplest ways to solve a Multi-label Classification problem is by converting it into  Multi-class Classification problem. Here we will combine the Animal and Breed categories to form a new set of unique categories and we will call this new feature ‘New_class’. training_set['New_class'] = training_set['Animal'] + training_set['Breed'] Let’s have a look at our new dataset. print(training_set.head()) Output: Preprocessing Images from tensorflow.keras.preprocessing.image import ImageDataGenerator train_dataGen = ImageDataGenerator(rescale = 1.\/255, shear_range = 0.2, zoom_range = 0.2, horizontal_flip = True) train_generator = train_dataGen.flow_from_dataframe( dataframe = training_set, directory=\"\",x_col=\"Images\", y_col=\"New_class\", class_mode=\"categorical\", target_size=(128,128), batch_size=32) We will now preprocess the images using Keras’ ImageDataGenerator class which will convert the images into an array of vectors that can be fed to the neural network. A set of features or parameters can be initialized to the ImageDataGenerator such as rescale, shear_range, zoom_range etc. These parameters help in extracting maximum features from an image. The flow_from_dataframe method allows us to import images from a data frame provided the path of the images using the parameter ‘directory’. The x_col specifies the independent factor which is an image and y_col represents the dependent factor which is the category of the image that we need to predict. The target size will be the size of the resulting images from the ImageDataGenerator object, batch_size is the number of sample images used to train at once. You will get an output like what’s shown below: Found 6206 validated image filenames belonging to 10 classes. Building Convolutional Neural Network We are done processing the image data. Now we can proceed to build a simple convolutional neural network. Keras allows us to build neural networks effortlessly with a couple of classes and methods. classifier = Sequential() The Sequential class initializes a network to which we can add layers and nodes. #First Convolutional layer classifier.add(Convolution2D(filters = 56,kernel_size = (3,3), activation = 'relu', input_shape = (128,128,3))) The add method allows us to add layers of nodes to the initialized network. In the above code, we added a Convolutional layer to the network. The convolution will be performed using a 3×3 matrix as specified with the kernel _size parameter. The activation parameter sets the activation function for the nodes. The input size should be same as the size of the outputs from the ImageDataGenerator (3 is the channel width). classifier.add(MaxPooling2D(pool_size = (2,2))) The above code adds a pooling layer to the network. We can add as many layers as we want as shown below, however, this puts a lot of pressure on the system resources. Choose the layers and nodes based on the capability of the machines. #second Convolutional layer classifier.add(Convolution2D(32,(3,3),activation = 'relu')) classifier.add(MaxPooling2D(pool_size = (2,2))) The line below adds a Flattening layer to the network. #Flattening classifier.add(Flatten()) Next, we will add a hidden layer and an output layer to complete the network as done with the following code blocks. #Hidden Layer classifier.add(Dense(units = 64, activation = 'relu')) #Output Layer classifier.add(Dense(units = 10 , activation = 'softmax')) We will now compile the network to initialize the metrics, loss and weights for the network classifier.compile(optimizer = 'adam', loss = 'categorical_crossentropy', metrics = ['categorical_accuracy','accuracy']) Let’s have a look at the description of our CNN : classifier.summary() Output: Training the CNN classifier.fit_generator(train_generator, epochs = 50, steps_per_epoch = 60 ) Output: The fit_generator method will train the classifier with the data we gathered by processing the images using ImageDataGenerator class. The epochs are the number of times the cycle of training repeats. Steps-per-epoch determines the number of times the weights of each node should be updated for decreasing the loss. Predicting For Test Set Preparing Test Data We will prepare the test data by adding the path and file extension to the original test_set. This will help us load the images directly from the csv file using the load_img() method that you will see in the following code blocks. test_set = pd.read_csv(\"Cats_and_Dogs\/Dataset\/test.csv\") test_imgs = [\"Cats_and_Dogs\/Dataset\/images_test\/{}.jpg\".format(x) for x in list(test_set.id)] test_set = pd.DataFrame( {'Images': test_imgs }) Identifying the trained classes The train_generator consists of the complete trained image data. Let’s have a look at the unique categories in the training data using the class_indices attribute of the train_generator. classes = train_generator.class_indices print(classes) Output: {'111': 0, '112': 1, '113': 2, '114': 3, '115': 4, '221': 5, '222': 6, '223': 7, '224': 8, '225': 9} Our model will be predicting the labels in the range 0 to 9 based on the above dictionary for each category. We will need to reverse these to the original classes. We will use a reverse of the above dictionary to later convert the predictions to actual classes. The dictionary can be inverted with the following line of code: inverted_classes = dict(map(reversed, classes.items())) print(Inverted_classes) Output: {0: '111', 1: '112', 2: '113', 3: '114', 4: '115', 5: '221', 6: '222', 7: '223', 8: '224', 9: '225'} Predicting classes Now it’s time to load the images one by one and predict and store the category of each image from the test_set. from keras.preprocessing import image Y_pred = [] for i in range(len(test_set)): img = image.load_img(path= test_set.Images[i],target_size=(256,256,3)) img = image.img_to_array(img) test_img = img.reshape((1,256,256,3)) img_class = classifier.predict_classes(test_img) prediction = img_class[0] Y_pred.append(prediction) The above code block loads each image from the test set preprocess it and feeds it to the classifier to predict. The predictions are stored in a list called y_pred. Notice that all the image sizes (256,256) are the same as we did for the training set, this is important and otherwise would result in an error. Let’s take a look at the predictions: print(Y_pred) Output: [4, 3, 9, 1, 1, 9, 2, 7, 8, 6, 8, 6, 6, 0, 1, 6, 0, 9, 7, 0, 7, ...] We can see that the predictions are in the range of 0 to 9. We can now use the inverted_classes dictionary to convert the predicted labels into actual categories. prediction_classes = [ inverted_classes.get(item,item) for item in Y_pred ] Now if we look at the prediction_classes, we will be able to see the actual categories we used to train our classifier with. print(prediction_classes) Output: ['115', '114', '225', '112', '112', '225', '113', '223', '224', '224', ....] There is one more thing to do. Remember, the original training set provided had 2 categories, class_name and breed which we later renamed as Animal and Breed. So we will now need to split the predicted categories into class_name and breed (Animal and Breed). animal = [] breed = [] for i in prediction_classes: animal.append(i[0]) # First character = class_name\/Animal breed.append(i[1:]) # Last 2 characters = breed\/Breed Creating a dataframe for the predictions and writing it to an excel predictions = {} predictions['class_name'] = animal predictions['breed'] = breed Let’s have a look at the predicted classes: #Writing to excel pd.DataFrame(predictions).to_excel(\"Cats_and_Dogs\/Predictions\/predictionss.xlsx\", index = False) Finally, you can upload the excel at MachineHack to see your score on the leaderboard. Tweak the parameters of the Convolutional Neural Network to achieve a good score and submit your predictions at MachineHack. Enjoy MachineHacking!","excerpt":"Image classification is one of the trending applications in machine learning. It has a wide range of applications — from facial recognition algorithms to identifying complex patterns in images like crime detections, and many other social, medical and technical applications. Image classification is not just about classifying images into categories, it has a broader and […]","categories":["Deep Tech"],"tags":["cnn","Convolutional Neural Network","Keras","Tensorflow"],"author_name":"Amal Nair","publish_date":"2019-07-12T17:00:56","publication_year":"2019","word_count":1682,"keywords":["NumPy","machine learning","Keras","TPU","AI","neural network","cnn","Ray","Matplotlib","Convolutional Neural Network","TensorFlow","Pandas","Tensorflow"],"extracted_tech_keywords":["AI","machine learning","neural network","Ray","TensorFlow","Keras","Pandas","NumPy","Matplotlib","TPU"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/multi-label-image-classification-with-tensorflow-keras\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10136937,"title":"RHEL Gives Linux A Much Needed AI Update","content":"A few days ago, Dell Technologies partnered with Red Hat Inc to bring the Red Hat Enterprise Linux AI (RHEL AI) platform to its popular PowerEdge servers, paving the way for its hardware to serve as a foundation for AI development. The idea behind this partnership was to make it easier for organisations to scale their IT infrastructure to support successful AI and ML strategies without needing to host those workloads in the cloud. They can deploy Dell’s PowerEdge servers within their own on-premises data centres or alternatively use them as part of a larger hybrid cloud setup. Apart from partnering with Dell, Red Hat recently unveiled Enterprise Linux AI, which is primarily focused on developers. One good example is a bootable RHEL image with pre-configured AI libraries like PyTorch, which allows users to quickly set up an AI-ready environment without going through complex installation and configuration processes. AMD is also very active in supporting AMD GPUs on Linux and recently released the AMD XDNA Linux Driver v3, which is crucial for enabling the Ryzen AI Neural Processing Unit (NPU) on Linux systems. It is likely to get merged into the Linux kernel 6.13. Linux-centric companies like openSUSE are contributing to AI so that it remains accessible. For example, openSUSE got listed on Hugging Face and made the first contribution of a dataset called cavil-licence-patterns, aiming to provide more advanced and accurate detection of licence issues and compliance. The Contribution of Red Hat RHEL AI combines several key components to create a powerful foundation for AI innovation. At its core are the open-source Granite models, a family of LLMs developed by IBM Research. Complementing these models is InstructLab, an open-source project that simplifies model experimentation and fine-tuning. This allows domain experts to contribute to AI models without extensive data science skills. All these components are packaged into a bootable Red Hat Enterprise Linux image, streamlining deployment across hybrid cloud environments. This approach is more inclined towards ethical AI. Many of Red Hat’s customers couldn’t go near AI because of the copyright implications, but this is basically the most ethical form AI can take. Red Hat’s approach addresses several challenges in enterprise AI adoption. By leveraging open-source principles, RHEL AI lowers the barriers to entry for AI innovation, making it more accessible to a broader range of organisations. The platform offers up to 50% lower costs compared to similar solutions, making AI development more economically viable for enterprises. A Reddit user praised the closed integration to CI\/CD and mentioned that you can create and share host images just like you would container images, and now developers and operators can run the exact same image in a container, or as bare metal on the host. “The technology is neat, but what it can do when integrated into your development, build and deployment pipeline is where the magic happens. It’s not a huge leap forward, just a few modest but highly useful steps forward,” he added. Linux Matters a Lot for AI Developers Developers are in the favour of using Linux to train models. A developer on Reddit mentioned that compiling applications is so much easier with Linux and when he uses Windows for the same task, he has to try and find VS 2022 launcher, download and run it, and Google to see what options he has to tick. All this combined with several GB of unnecessary dependencies, which he does not want was given anyway. “Then I have to go to the NVIDIA site, download the CUDA Toolkit, and install it. Then I need CMAKE, download it, and install it. If it all works, compiling takes about 25 minutes. On Arch Linux, just type pacman -S base-devel cuda, and you’re ready to go. Compiling takes like 5-10 minutes, and inference is ~25% faster too,” he added, suggesting Linux is not only efficient while compiling but also with inferencing. Efforts like RHEL AI are important for the Linux community as almost everything around AI is researched and developed on Linux, which eventually makes it a more efficient platform for training and using AI models. For the most part, you are just one command away from setting up the AI development environment, and it just works, spending hours on Windows to configure the AI development environment.","excerpt":"RHEL AI offers up to 50% lower costs compared to similar solutions, making AI development more economically viable for enterprises.","categories":["AI Features"],"tags":["AI","Linux","red hat"],"author_name":"Sagar Sharma","publish_date":"2024-09-30T19:11:47","publication_year":"2024","word_count":713,"keywords":["CUDA","data science","Hugging Face","Go","red hat","AI","PyTorch","ML","RAG","Aim","Linux","R"],"extracted_tech_keywords":["AI","ML","data science","Aim","PyTorch","Hugging Face","RAG","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/rhel-gives-linux-a-much-needed-ai-update\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":25561,"title":"How This Tech Startup Is Using Deep Learning And Video Analytics To Power Industrial Automation And AI Solutions","content":"Image Source: SilverSparro Founded in 2014 by Abhinav Kumar Gupta, Ankit Agarwal and Ravikant Bhargava, SilverSparro is a technology startup that provides deep learning solutions to businesses spanning from manufacturing to critical sectors such as healthcare and banking. Specialising in areas like computer vision, natural language processing, image detection and classification among others, the company considers industrial automation and artificial intelligence as its core expertise. In Feb 2018, the startup secured a seed round funding of $300,000 from GSF, Shangrila and UHV Technologies, among others. They are also backed by angels such as Anand Chandrasekaran from Facebook; Rajesh Sawhney, founder of Innerchef and Dinesh Agarwal, CEO of Indiamart. Analytics India Magazine caught up with Abhinav Kumar Gupta, founder and CEO of Silversparro, where he shared their journey as startup founders and how the company carved out a niche for itself in the deep learning industry. Living Their Dream SilverSparro was founded with an aim to make the best out of newer and emerging technologies like deep learning. Gupta recalls the initial days of the company, when they had all quit their well-paying jobs to dwell into this exciting area. He shares, “I was in management consulting working with top financial companies while Ankit, my friend and batchmate from IIT Delhi was developing WebApps for a Tech startup. At one end we were terribly bored of our regular jobs while on the other, there were technologies like drones, 3D printing, VR, AR and deep learning which were very exciting. We were really inspired by this tech revolution and were eager to dirty our hands”. They left their jobs to explore these areas and developed their first technology in drones. While working on drones they realised that there were no good solutions which could easily analyse huge amount of photographs and videos captured using drones. This is when they came across deep learning. Gupta confesses, “We were so fascinated by it that we dropped drones completely and started focusing on deep learning wholeheartedly”. Since 2015, the team has been taking client projects in deep learning and Ravikant Bhargava, their IIT Delhi batchmate has been a part of their team since then. Deep Learning Product Offerings And Use Cases Silversparro provides a host of custom deep learning solutions for its clients in various fields. Describing the startup’s portfolio in deep learning, Gupta says, “Over the past three years, SilverSparro has built core expertise in computer vision use cases using deep learning”. They are also venturing into CCTV video analytics for retail establishments, which is giving them a good traction. Gupta shares, “Earlier, the use of CCTV cameras was confined to security and that too post-incidental analysis. At the same time, all of it had to be done manually. What’s new is that, deep learning has opened a new door for real-time fully automated CCTV footage analytics which can be used for customer demography analytics, analyzing customer behaviour and auditing staff performance”, he said. Using this, the team has managed to check pilferage of material from coffee shops by tracking individual items using cameras. With clients such as are Policy Bazaar, Viacom18, UHV Technologies, and CrownIt among others, he shared some uses cases as below: Case Study Of UHV Technology: For this US-based engineering company they built computer vision-based large-scale metal sorting system with deep learning models. Gupta explains, “This basically takes in images of metal pieces on a conveyor belt using a high speed camera, feeds images real-time in to Silversparro’s system enabling us to predict the metal class with over 95 percent accuracy. This is equivalent to sensor-based system in terms of accuracy but way cheaper and has much higher throughput”. Case Study Of CrownIt: For this market research and deals company, they process more than 10,000 receipt images on a daily basis to extract details like date, time, bill number, and line items with extremely high accuracy by fully automating the process. The models that they we have built have been trained on over 30 lakh images and beat Google Vision OCR API in accuracy,” he said. The Team Structure The startup is currently 25-people strong, mostly comprising of deep learning engineers and researchers. They hire at three levels: Fresher or Interns Mid-level engineers (2-3 years of experience) Senior engineers (4-6 years of experience) He further shares that for engineering roles, they look for strong aptitude, problem-solving abilities, demonstration of quick learning and ability to work in chaotic environment. Whereas, for research roles, they look for some research experience as well as strong inclination to go into deep learning research for the long term. The SilverSparro team is divided into research and solutions team. While research team is fully focussed on going deep into the use cases by reading up research papers, making sense of data and developing new models, the solution team develop projects by building client solutions and deploying models at scale for the clients. Roadmap And Challenges While the startup has witnessed a steady take-off, they are not averse to challenges. Elaborating on the challenges encountered while setting up deep learning solutions at an enterprise level, Gupta highlights three major challenges: Identifying The Right Problem: There are many problems for which deep learning is not the right approach. There are many problems which can be solved by traditional software development. There are cases when due to hype, enterprises start with the wrong business case. Hiring Talent: Deep learning talent is a scarce and enterprises have an option of building in-house solutions by hiring experts, engaging expert team or using APIs. Unless enterprises are deep tech companies, I always recommend outsourcing the task as talent is really expensive to hire and they leave quickly. Data: Deep learning and machine learning approaches need data to train from. Enterprises need to make sure that their data is organized properly and easily available for developers to build models on. Still overcoming these challenges, the SilverSparro team is aiming to scale-up their CCTV video analytics platform in India and globally for various retail clients in the coming years. “We are also hiring and looking to triple our engineering and research strength by the end of this year”, said Gupta. The Competitive Strategy Since the entire field is moving fast and there are other companies that are trying to catch up, the team self-admittedly shares that they have to pull up their sleeves to stay ahead of the game. “We believe that the best way to win is to focus really hard on narrow use cases and develop research expertise in fundamental deep learning approaches. Many of the frameworks and hardware required will continue to be built up by the existing giants – but there is huge scope for startups like ours to solve unique problems and develop own niches”, he said. Gupta strongly believes that their strong and deep focus on research specifically on narrow computer vision use cases would set them apart. They are also closely partnering with the market leaders for pilots and solution development to stay ahead of the game.","excerpt":"Founded in 2014 by Abhinav Kumar Gupta, Ankit Agarwal and Ravikant Bhargava, SilverSparro is a technology startup that provides deep learning solutions to businesses spanning from manufacturing to critical sectors such as healthcare and banking. Specialising in areas like computer vision, natural language processing, image detection and classification among others, the company considers industrial automation […]","categories":["AI Startups"],"tags":["Analytics India Magazine","video analytics"],"author_name":"Abhishek Sharma","publish_date":"2018-06-19T10:41:45","publication_year":"2018","word_count":1170,"keywords":["Go","machine learning","artificial intelligence","AI","computer vision","Analytics India Magazine","RAG","Aim","deep learning","analytics","video analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","computer vision","analytics","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-tech-startup-is-using-deep-learning-and-video-analytics-to-power-industrial-automation-and-ai-solutions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10308,"title":"Battle of AI: Siri vs. everyone else","content":"AI assisted modern apps have become an integral part of our everyday life.  As if the technology in various other forms had not made us enough lazy that artificial intelligence is evolving to corrupt us further. Sit back and give your fingers that moment of relaxation. Just speak about what you want. Fair enough? Well, that’s what the generations coming along are getting used to (well, that includes me too). Whether its iPhone or android or a Microsoft driven system, there is no place that is untouched with AI. While Apple’s Siri has been conversing with us, staging desired results, searching the restaurants for us or band shows in the city or giving us weather predictions, it appears that there are many others in the town to give fierce competition to this favourite AI app of many. With Google Now, Cortana, Hound, Dragon Go and many more, the market is flooding with AI and machine learning driven apps and given the largest and oldest market presence by Siri, we often tend to see it as Siri vs. everything else. Personally having fallen for the capabilities that Siri holds, I thought of making a comparative analysis of Siri with various similar or remotely similar machine-talking-things. Let’s read out what I found. With no intention of showing a high to one over others, in the comparisons made I have tried to stick to points such as the app’s functionalities, customer reach, privacy and security among others. Things it can do Things it can’t do AI usage Privacy & Security Siri Searching things, calling people, tweeting your thoughts, calculations, creating reminders and events, setting alarm, calling people, making reservations are just a few things that Siri can blissfully do for you. The list may go up to more than 50 such things that it is capable of doing. It can also do stuffs like launching applications, finding and reading mails, wishing you a good night. It works only on ioS It can browse the web only if the url is in the history, otherwise it needs to be fed manually. Another minor hitch is that Siri doesn’t understand who she is and has to be trained for it. It would take Siri as a search word instead of a gesture to address her name. It works as intelligent personal assistant and knowledge navigator. The feature uses a natural language user interface to answer questions, make recommendations, and perform actions by delegating requests to a set of Web services. Despite all the positives that Siri poses, few say that there could be a flaw in it when it comes to security. Despite pins and passcodes, there are chances that someone else traverses through your personal information via Siri. It especially could be a threat for those having Siri enabled on the lock screen. However, it could be saved by disabling Siri on the lock screen. Google Now Running on a variety of devices and operating systems, Google Now works quite similar to Siri such as dictating texts and emails, searching facts, booking reservations, keep an eye on stock and sports etc. Unique feature of Google Now is its capability to regularly show an information card that estimates your commute based on time of day and location. Though it can show estimated travel time but if you happen to use multiple methods of transport, it might not work as expected. It is an intelligent personal assistant by Google which uses natural language user interface to perform its functionalities. For Google Now to fetch you desired results, it uses a lot of your personal information. And given the fact that Google is no stranger to controversy regarding the release of personal information of its customers, it brings privacy concerns into picture. Cortana Its biggest advantage is its built up in desktop and it can recognize both voice and type commands. Offers a lot of personalized information based on what it knows about you starting from information about schedule, weather etc. It facilitates natural language search, identifies songs and sings for you with a more human touch. Additionally, it can open websites, send emails, create alarms and also chat with you. It cannot launch applications and change settings. Another turn off is that after the command is given once, she stops listening and needs to be given the command again by hitting the listen button. A continuous command system between user and Cortana could be an add on. It has most powerful AI ever used in the form of Adam’s deep learning. Cortana can also tell you from a picture the amount of calories in your food, thanks to Adam. Using Cortana might lead to a host of data sharing. It uses various data such as device location, data from your calendar, email and text messages, your contacts etc. It also keeps a track on how you use your device and other Microsoft devices. It needs access to all those for it to be able to work the way it does. Amazon Echo Alexa, the voice for Amazon echo can do a range of things for you- order a cab ride, play music, remind you appointments, flip coins, read book from kindle and audio book store, be a kitchen assistant, ask for time from anywhere in the room and more. Its quicker to respond and understand commands. Its functionality might not match Siri, Google Now and Cortana. It faces difficulty in answering tough questions. It cannot search music within a playlist. Audio speakers are good but not the best in the market. Amazon’s Alexa uses AI to give a tough competition to others. Echo is a well implemented system whose voice recognition is accurate & best in industry. However, Alexa’s question answering engine lacks capabilities in comparison to other AI platforms such as Google, Apple & Microsoft. With Amazon echo always at home, with sensitive microphones and cloud storage one may fear about it storing all the conversations. But Amazon clears out the concern saying that it doesn’t record private conversation and for those if still there is a fear, mute button can be pressed when not in use. Hound It can conduct both speech recognition and language analysis simultaneously, thereby increasing accuracy and speeding up results. Hence it is smarter, faster and much more helpful version of Siri. Unlike Siri or Cortana which might get confused with few questions and redirect you to search page, Hound can correctly answer complex and in-depth questions with ease. It can find you coffee shops, weather report, nutritional facts and much more. The voice recognition in Hound isn’t as good as Google. This voice-powered digital assistant is giving serious competition to others in AI enabled app space. It can produce near real-time translations of whole sentences in various languages and get responses in split seconds which suggest a good AI dependency. It can transmit your location to remote servers. Though it gives the option to disable but it can’t be achieved as it would require you to open the app and again accepting the transmission of data. Aisha Indian version of voice assistant similar to Siri which can perform tasks such as give you movie reviews, initiate a google search, review stock market details, make calls, read news, update your status on social networking websites such as Facebook and Twitter. It is quite similar to Siri, but not as powerful. AISHA standing for ‘Artificial Intelligence Speech Handset Assistant’, is a tough competition to Siri in India. The name itself implies its dependency on AI. It needs info like your location and contacts to deliver the said performance. Viv Viv is the next gen AI assistant made by the makers of Siri. It can handle complex requests and connect with third party merchants to purchase goods and book reservations. Its approach is closer to Amazon’s Alexa or Facebook’s Messenger bots. It is focused on just one thing: scheduling meetings which it does very well but can’t do anything else. It cannot offer merchant distribution to a massive consumer base the way Facebook and Amazon does. It claims to be providing an intelligent interface to everything. It enables developers to distribute their products through an intelligent, conversational interface using an AI platform. To avoid any Security and privacy issues, Viv is working on a new model that will quickly define what is logically private and potentially shareable with permissions. Dragon Go It’s a combination of Nuance’s Dragon voice recognition & natural language understanding with artificial intelligence. It gives you the most relevant content front based on what you require. It gives direct access to over 200 of the most trusted and reliable destinations for mobile content such as Facebook, YouTube, Yahoo. It takes some time and training to learn voice commands. It has limited words and vocab. It depends on AI to a large extend and uses elements of it to help understand the intent of a query. It faces privacy issues over transmission of users’ contact list to Nuance’s servers. But the company defended the move by stating that they do it for people to better identify names from address book when asked by them. Indigo Virtual Assistant A mobile assistant it lets you speak as if talking to a human. She is also capable of sharing her personal opinion. It can manage your calendar, create reminders, control your music by voice, get notes, quickly find locations on maps, pay videos, make phone calls, find contact details and more. Indigo does excellent however with a few queries it might fail to deliver results. When it cannot answer your questions, it leads you to search engine. It depends on natural language interaction and is the world’s first intelligent personal assistant app to offer consumers personalization and persistence across platforms and devices. Its intelligent capabilities make Indigo highly intuitive. It needs access to your contacts, messages and location for the basic functioning of the app. SpeakToIt Assistant Assistant can learn your favourite places and services and customize your experience based on your past preferences. It allows to speak naturally and offers best suggestions. It is available in many languages including English, Russian, German, Portuguese, Chinese, French, Korean and Japanese and is working hard to learn more. It cannot read email or text messages aloud. Also, it may require to repeat commands multiple times or type them manually to make it understand few things. — They collect location of the device to provide location-based functionality which you can opt out of at any time by turning off the location settings in your device. They do not store the location of the device. Vokul Vokul allows you to dictate text messages, emails, post to twitter and facebook, control audio player with your voice, listen to audiobooks, call your contacts, listen to your twitter and facebook feeds and do much more. It is fast, accurate and tuned for noisy environments. It cannot read received text messages. It doesn’t understand various languages and won’t get you if you are speaking languages that are foreign to it.It cannot integrate with services like Wikipedia and Yelp — It collects and uses technical data and related information about your device, system and application software. With a huge list to lay a ground for the battle of AI based app, there could be various other domains to be compared- customer reach being an important one. Let’s see how these apps are being utilized by the users. According to a study, only up to 70% of iPhone users use Siri on a regular basis and 2% of American iPhone users have never used it. Talking about Google, which is available in both ioS and Android, it fetches more users than anyone else. With the numbers reaching to more than 1 billion active users, the customer reach of Google Now is ought to be really high. Microsoft Cortana prides over 1 billion question since making her debut while Amazon Echo touched 3-million-unit sale by April 2016. The numbers for others seem to be highly impressive too indicating both the dependency of users on such AI based apps and its popularity. Offering a wide range of features and amidst the cut throat competition, it surely is a herculean task to declare a clear winner as each of them gives some reason to favour it over others!","excerpt":"AI assisted modern apps have become an integral part of our everyday life.  As if the technology in various other forms had not made us enough lazy that artificial intelligence is evolving to corrupt us further. Sit back and give your fingers that moment of relaxation. Just speak about what you want. Fair enough? Well, […]","categories":["Deep Tech"],"tags":["AI India","Artificial Intelligence India","Cortana","Siri"],"author_name":"Srishti Deoras","publish_date":"2016-06-29T05:18:54","publication_year":"2016","word_count":2052,"keywords":["Go","Cortana","artificial intelligence","machine learning","AI","Artificial Intelligence India","Git","RAG","Aim","deep learning","Siri","Rust","AI India","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","Aim","RAG","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/battle-ai-siri-vs-everyone-else\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":65155,"title":"Looking For An Online Master’s Program In Data Science &#038; ML? upGrad Has The Perfect Course For You","content":"While the choice to go online and pursue distance learning to upskill has been available to us for a while, the recently enforced lockdown has given it a new impetus. Professionals across industries – especially in data science and machine learning – are enrolling in droves with several ed-tech platforms offering some of their courses for free. These short-term certifications, while helpful in developing new skills and learning the latest practices, is worth little when you want to truly advance your career and stand out from competitors. Instead, data science professionals find themselves turning to Master’s programs when they are ready to make the transition to high-level positions. Even here, the pull of online offerings because of its various benefits make it a popular alternative against traditional full-time courses when it comes to higher education. Not only do they enable learners to pursue a degree while they keep working or are in-between jobs, they are cheaper than on-campus courses, provide access to a global network of instructors, and enables them to concurrently manage other priorities. But with numerous data science and machine learning offerings available online, which one should you choose? ALSO READ: upGrad Sees 50% Surge In Learners Amid Covid-19 upGrad’s Master’s Program vs Other Courses While it has been established that pursuing an online Master’s degree has its perks, it must be emphasised that all digital learning programs are not alike. Most might offer you a certificate at a limited budget, but that alone should not be the driving force behind your choice. The quality of the learning experience and the relevance of the course – especially in dynamic disciplines like data science and machine learning – is paramount, and that is what upGrad continues to focus on with its latest offerings. Both the Master’s degree in data science as well as machine learning come with a world class portfolio that delivers all of the above-mentioned benefits of e-learning – and more. At just 1\/10th the cost of offline programs, upGrad is offering a globally recognised degree that carries a dual alumni status. What is more, an impressive faculty and a diverse peer group from over 15 countries makes it the best course in the market for data science and machine learning professionals in the field. Not convinced? Let us explore this a little more deeply to understand why: Dual Alumni Status Both these programs are certified by top universities, one of them being the prestigious Liverpool John Moores University. LJMU is one of the premier institutes in the UK delivering excellence. On successful completion, learners will earn a Master’s degree from LJMU – which is globally recognized – as well as the opportunity to connect with its global network of faculty. Not only that, the same program offers you a Post Graduate Diploma from the widely recognised IIIT Bangalore as well. While online learning does come with the benefit of acquiring skills and certifications without having to quit your job, a dual alumni status is not only an efficient way to invest your time and money, but is also fairly uncommon when it comes to e-learning in higher education. Admirable Faculty & Cutting Edge Curriculum With best-in-class content designed by leading industry practitioners in the form of videos, case studies, projects, assignments, live sessions and informative workshops, it covers over 500 hours of quality learning. While the data science Master’s program allows learners to choose from five specialisations depending on their background and career aspirations, the machine learning course will enable them to build and deploy industry projects deftly. What is more, these programs are led by instructors who collectively carry a rich experience in engineering, applied mathematics, computer science, and artificial intelligence, and an average experience of more than nearly 20 years. With multiple research papers published in the domain of data science and machine learning, they have proved their proficiency in their respective subjects. Together, through these programs, they teach you various skills, including statistics, predictive analytics, data visualisation, big data analytics, model deployment, NLP, deep learning, reinforcement learning, and graphical models, among others. ALSO READ: Data Scientist At upGrad Explains How To Choose Between Data Science, ML & Big Data Diverse Peer Group With learners from over 15 countries as potential peers, you can leverage the insights from a global cohort of diverse learners and work with people from different backgrounds. Some of your peers may also have over 30 years of work experience behind them. In fact, this course has shown to resonate with international audiences. Interacting with like-minded people holding different domain expertise will be helpful in your data science and machine learning careers, making its global appeal more relevant and consistent with your career goals. Maximum Number Of Thesis One of the features that make these courses unique is the astounding number of its students going on to write excellent thesis. In fact, it is the only online masters program which has recorded the maximum number of thesis written up until now. For instance, one project focused on building deep learning models to analyse human bio-parameters for wearable technology devices. Another notable example sought to identify the power relationships between social media users based on their conversations. One-On-One Mentorship Not only does it provide a personal industry mentor for every learner, additionally, it also allocates a dedicated ‘Student Success Advisor’ to each student as well. This means that they not only get a fortnightly group mentorship with experts, but also access to the university’s wider student support services to ensure consistent progress in their course. With the support of over 300 industry mentors, it has delivered over 10,000 hours of mentoring. This takes the shape of unparalleled guidance from industry mentors, teaching assistants and graders – all at the cost of 1\/10th of conventional offline programs. Placement Support A dedicated career support on completion of this course has fetched a highest package of 72 lakh, and an average salary hike of 58%. The course unlocks global access to job opportunities for its learners, empowering over 12,600 learners with support from over 300 hiring partners. Many upGrad alumni have joined top companies in various capacities, including Uber, Microsoft, PwC, Genpact, American Express, Tata Motors, Wipro, and more. Outlook upGrad is celebrated for the quality of its online offerings and its Master’s program is no different. Targeted at working professionals who want to advance their careers without disrupting their work, this online program is built on the platform’s rich legacy of launching world class courses at a fraction of the cost of similar courses. Enrol now for an experiential learning journey – one that packs together several benefits for data science and machine learning professionals.","excerpt":"While the choice to go online and pursue distance learning to upskill has been available to us for a while, the recently enforced lockdown has given it a new impetus. Professionals across industries – especially in data science and machine learning – are enrolling in droves with several ed-tech platforms offering some of their courses […]","categories":["AI Trends"],"tags":["best online data science masters","data science education","data science master","data science salary","data scientist salary","edtech","master data science","online education","online masters analytics","UpGrad","what is data science"],"author_name":"Anu Thomas","publish_date":"2020-05-14T12:00:55","publication_year":"2020","word_count":1112,"keywords":["best online data science masters","data science master","edtech","online masters analytics","deep learning","master data science","data science","artificial intelligence","data scientist salary","RAG","NLP","online education","UpGrad","analytics","what is data science","machine learning","AI","ML","data science salary","data science education","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","data science","analytics","RAG","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/looking-for-an-online-masters-program-in-data-science-ml-upgrad-has-the-perfect-course-for-you\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10137787,"title":"Will an Oracle HR Hire Your Digital Twin?","content":"We have likely had numerous discussions about AI replacing or augmenting various roles, without a definitive answer on how it will unfold. However, one domain, which has ‘human’ in its very name, has been among the earliest adopters of AI. We spoke about it earlier last year when we said ChatGPT takes over HR. Today, AI forms an integral part of human capital management (HCM) and Oracle is leading the way. Interestingly, Oracle’s HCM has even thought of agents and digital avatars to participate in all interactions. AI Agents, Here Too At the sidelines of the recent Oracle CloudWorld 2024 in Las Vegas, Nagaraj Nadendla, senior vice president of Oracle Cloud HCM Product Development, spoke to AIM about the scope of AI in high-volume HR tasks such as recruitment and brought up the concept of digital avatars doing it for you. Citing the challenges of coordinating high-volume interviews in regions like India, APAC, JPAC, and Southern Latin America, where scheduling across multiple candidates can be time-consuming and repetitive, Nadendla proposed the concept of digital twins for automating the process and improving efficiency. He asked the candidates to imagine having a digital twin as an agent who could coordinate and negotiate with other agents (digital twins). “It’s not just about coordination,” Nadendla further added. Nadendla focuses on AI agents in HR tasks as a feature that already exists. He believes it can handle tasks like sifting through thousands of resumes and identifying the most suitable candidates based solely on skill sets, bypassing potential biases related to gender, ethnicity, or educational background. “Generative AI or passive AI in general doesn’t know gender, ethnicity, [etc.]. It’s all about skills,” Nadendla explained, highlighting how AI can mitigate some of the biases that have historically plagued the recruitment process. While the idea of using AI agents for recruitment and other selection processes is intriguing, it inevitably raises the question of how these models will navigate and prevent hallucinations. Something Oracle is well aware of. Free of Bias? “Our team is also very focused on the forest, not the tree. The tree is the bias,” he said, “Every time there’s an introduction of technology, some derivative, not-so-great outcomes are there,” he admitted. However, he remains optimistic that the benefits of these innovations far outweigh the challenges, especially when organisations focus on the broader advantages rather than individual hurdles. “Humans are the most wise. There’s no machine that can beat humans,” said Nadendla. However, he highlighted that “human bias is worse than any other bias.” Nadendla firmly believes that AI is playing a larger role in hiring, helping reduce bias and speeding up candidate screenings, especially in high-volume industries. He goes on to cite tools such as biopsychology-based assessments and AI-driven skill evaluations where human oversight remains crucial. While AI can handle tasks like sifting through resumes and assessing technical skills without bias, final evaluations and key decisions still require human judgement to ensure fairness. In India, the HR tech platform GetWork.ai has been leveraging AI for recruitment tasks. The startup’s CEO, Rahul Veerwal, earlier told AIM that their platform receives thousands of applications per job posting. To ease the process, it offers a feature that matches a candidate’s resume to the job description by percentile. Once the desired score is reached, an AI voice bot conducts interviews at a rate of 1,000 calls per minute, allowing recruiters to hire candidates within a day. In a recent podcast, Matt MacInnis, COO of Rippling, a workforce management system, spoke about AI implementation. “AI is going to contribute in some way to evaluating Human Performance and so we knew there was an opportunity here and that’s what talent signal is going to deliver,” said MacInnis, speaking about the company’s vision of using AI in their operations. It is evident that the incorporation of AI into HR processes is now not new. Nadendla believes it is not about replacing human workers but enhancing their productivity and decision-making capabilities. “We are here to help our customers be successful in whatever endeavour they undertake,” he concluded.","excerpt":"“Imagine if each of you had a digital twin, an agent; and I, and my agent, are talking to all of your [candidates’] digital twins to coordinate and negotiate.” proposed Nagaraj Nadendla, senior VP, Oracle Cloud HCM Product Development.","categories":["AI Trends"],"tags":["AI","digital twin","Hiring"],"author_name":"Vandana Nair","publish_date":"2024-10-08T12:00:00","publication_year":"2024","word_count":674,"keywords":["Go","ChatGPT","API","AI","ML","digital twin","Hiring","Git","RAG","Aim","generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","ChatGPT","Aim","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/will-an-oracle-hr-hire-your-digital-twin\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10083671,"title":"Year in Review: How (Badly) Did Meta Do in 2022?","content":"With nearly USD 237 billion wiped out from its valuation in early February, Meta had a rough year in 2022. The moment was historical, albeit, for the wrong reasons. It was the biggest drop in a single-day valuation for a US-based company. Late Show host Stephen Colbert called it ‘karmic retribution’ for all the misinformation prevailing on its social media platforms. Overall, shares of Meta plunged more than 60% this year. The company was also labelled the worst performer in S&P 500 in 2022 by Forbes. Further, this year, Meta cut costs across its business by downsizing its teams by about 13% and letting over 11,000 employees go, scaling back budgets, reducing perks, and shrinking its real estate footprint. Investors losing confidence Meta CEO Mark Zuckerberg’s optimism about the Metaverse is well known. However, the optimism is not shared by Meta’s investors or for that matter, some of its employees. Until last year, Meta’s stock was performing well, riding on the boom fuelled by the pandemic; however, this year tells a different story. Investors are asking Zuckerberg to scale back on his Metaverse plans. In an open letter titled ‘Time to Get Fit’, Brad Gerstner, one of Meta’s largest investors, said that Meta’s larger focus on the Metaverse has distracted it from focusing on the core business. So far, Meta has already invested nearly USD15 billion in Reality Labs, its Metaverse unit. Undeterred by investors’ scepticism, Zuckerberg ​​reiterated his commitment to investing even more in the segment. However, Zuckerberg did mention that the company is still spending around 80% of its spend on its core business (Facebook, Instagram, and WhatsApp) and not on Reality Labs. To make matters worse, Horizon World, Meta’s social VR experience platform, has failed to attract users. As of October this year, it had around 200,000 users. By the end of 2022, Meta was expecting to have 500,000 users on Horizon World’ however, the company scaled back on its estimates and expects to hit 280,000 users only. On the AI front When it comes to AI, Meta did not have an eventful year here either. Its Galactica AI, an open-source large language model, trained on scientific knowledge, was taken down just after three days of release. Galactica was heavily criticised by members of the AI community for producing inaccurate results. Micheal Black, director at Max Planck Institute for Intelligent Systems, said Galactica Galactica generates grammatically incorrect text. “It offers authoritative-sounding science that isn’t grounded in the scientific method. It produces pseudoscience based on statistical properties of science writing,” he said. In August, Meta also released its AI chatbot BlenderBot3. The chatbot had access to the internet and soon started generating controversial content. For example, when asked about Zuckerberg himself, the chatbot said, “…his company exploits people for money, and he doesn’t care”. It also spewed anti-Semitism and conspiracy theories with élan. However, not all was lost for Meta’s AI. Last month, Meta showcased its AI agent CICERO, which is the first AI agent to achieve human-level performance in a complex natural language game called ‘Diplomacy’. Let’s talk finance In Q3, Meta reported revenues of around USD 27.71 billion, down by 4% year-on-year. This year, Meta has posted consecutive quarters of revenue declines in Q2 and Q3 and is expecting to post its third straight drop in the fourth quarter as well. So, what’s impacting Meta’s financials in 2022? It has mostly to do with advertisers holding back on their spending amid economic uncertainty, an ongoing war and global supply chain constraints. Further, the implementation of iOS14 (which hinders Facebook and Instagram’s ability to track their user’s digital activities) and heavy competition from Tiktok have also impacted Meta’s earnings in 2022. Tiktok, the social media platform owned by Chinese company ByteDance, has emerged as a competitor for not only Meta but also Google. The company is slowly eating away ad revenue shares of the tech giant, and some reports suggest Tiktok’s ad revenue will surpass that of Meta and YouTube (owned by Google) by 2027. This year, Meta’s total costs and expenses also surged, reaching USD 62 billion in the first nine months of 2022, compared to USD 50 billion in the prior-year period. Meta’s bet on the Metaverse is also costing the company a lot, with Reality Labs losing nearly USD 9 billion in the first three quarters. What’s in stock for 2023? Irrespective of what investors think, in 2023, Meta will spend even more money on building the Metaverse. “Beyond 2023, we expect to pace Reality Labs investments such that we can achieve our goal of growing overall company operating income in the long run,” Meta said in a statement. Crumbling ad revenues with the likes of Apple and Titok entering the space, Meta has to continue to spend big on the Metaverse. With the future of the company in question, Zuckerberg realising his vision for the Metaverse could be what saves Meta. Further, advertisers are expected to hold back on spending until the economy recovers; hence Meta’s poor ad revenue is expected to continue in 2023 as well.","excerpt":"Investors believe Meta’s larger focus on the Metaverse has distracted it from the core business","categories":["Global Tech"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2022-12-29T14:00:00","publication_year":"2022","word_count":844,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","Git","BERT","llm_models:BERT","ViT","R"],"extracted_tech_keywords":["AI","R","Go","Git","BERT","ViT","RPA","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/year-in-review-how-badly-did-meta-do-in-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10079490,"title":"Amazon is Creating More Jobs with Robots, but Reality Says Otherwise","content":"Robotic automation has pioneered various industries by reducing operating costs and improving product quality. Many organisations are now looking to drive forward by improving the quality of work for employees and supporting their growth ambitions. The future of robots replacing human jobs is not that far off. One such company led us to Amazon, which has been making significant investments in advanced technology and robotics within its operations. This week, the company introduced a new intelligent robotic system called ‘Sparrow,’ a state-of-the-art robot that handles millions of diverse products at Amazon. Sparrow leverages AI and computer vision to handle and recognise millions of items. In 2021, employees around the world picked and stowed over 13 million packages per day. “Robotics technology enables us to work smarter, not harder, to operate efficiently and safely”, says Amazon. But, in reality, all of the employees who were once doing such mundane tasks are bound to lose their jobs eventually. However, the customer-obsessed Amazon says otherwise. The company claimed that the deployment and design of robotics and technology across their operations have generated over 700 new categories of jobs within the company—all because of the ‘technology introduced into their operations’. The new roles claim to employ tens and thousands across Amazon, help demonstrate the positive impact that technology and robotics have for their workplace. Sparrow effect Amazon claimed that the new robotic system ‘Sparrow’ will streamline the fulfilment process by moving individual products before they get packed. Moreover, this advancement would be a major technological advancement to support its employees. The company claimed that over the years it has has been at the forefront of innovation, making significant investments in robotics and advanced technology within its operations. “Our dedicated teams of engineers, roboticists, software developers, and other experts have been developing solutions that allow us to automate fundamental capabilities.” (Source: Amazon) Chops off its robotics arm This new development also comes in the backdrop of Amazon laying off employees from its robotics team. Last week, an anonymous source told AIM that Amazon had started firing employees from the Robotics team. Read: [Scoop] Amazon Fires its Robotics Team A developer named Jamie Zhang expressed his shock on the decision in a LinkedIn post this week. Zhang was working with the AWS business to build continuous deployment (CD) and continuous integration (CI) pipelines.","excerpt":"Amazon introduced a new intelligent robotic system called ‘Sparrow,’ a SOTA robot that handles millions of diverse products.","categories":["AI News"],"tags":["Amazon","Computer Vision","Robotics"],"author_name":"Bhuvana Kamath","publish_date":"2022-11-11T18:45:26","publication_year":"2022","word_count":387,"keywords":["Go","AWS","AI","ML","Amazon","Robotics","computer vision","RAG","automation","Aim","Computer Vision","GAN","R"],"extracted_tech_keywords":["AI","ML","computer vision","Aim","RAG","AWS","R","Go","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amazon-is-creating-more-jobs-with-robots-but-reality-says-otherwise\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61973,"title":"Benefits Of Having A One-On-One Mentor In Data Science","content":"The demand for data science jobs has been experiencing exponential growth over the last few years. As data science penetrates deeper into business operations, there has been a significant disruption at the workplace with data scientists and data analysts dabbling with newer tools. Almost every organization has shifted from traditional business techniques to one anchored around emerging technologies. With the growth in this domain, data science enthusiasts are either switching their careers or upskilling to be ahead of this game. In fact, as per a survey we conducted, 39% of respondents thought that the transition from a non-data science background to a career in data science and analytics was not impossible. Keeping oneself abreast with the changing trends while pursuing courses and other resources to upskill is already a difficult task. This can be even more challenging for people who are relatively new to the field. To make this easier, some organizations have started introducing meet-ups, mentorship programs, etc. In this article, we will talk about the benefits of having a one-on-one mentorship in data science. Collaborate In Solving Problems Oftentimes, working with a real-world problem or a Capstone project seems difficult for a learner. With the help of a one-on-one mentoring program, a data science mentor – with their experience and knowledge – can help by collaborating on specific problems. They will know whether the problem arises due to incorrect data processing, or because of using the wrong tools. This will help mentees not only have a peer-to-peer collaboration but also help them bridge the gap between the techniques and the real-world projects. Provide Feedback In this ever-changing landscape, there is a constant rush to move ahead and upskill to newer tools and technologies. People have little time to reflect on their performances and understand what they are doing right, and where they might be going astray. In one-on-one mentorship, a data science mentor will help a mentee by providing feedback as well as constructive criticism whenever required. Data science mentors help by advocating professional ways to tackle situations they are likely to come across in their career. Help Chart Goals In one-on-one mentorship, a data science mentor can help a mentee focus on specific goals. They can provide advice, support and guidance, which will help them in their journey. One-on-one mentors work as a thought-partner in this journey, who could help mentees by advising on which paths to choose in order to build a successful career. Working with an experienced mentor helps in thinking with a different perspective as well. Help Make Progress A data science mentor in a one-on-one program helps the mentee progress by sharing his technical resources to become an expert in the domain. The mentor will help reinforce the skills and knowledge they have acquired and make them understand how to bring the best of the knowledge and skill sets to their professional lives. One-on-one mentoring also helps in advancing careers by helping mentees adapt to new situations by developing core communication skills. This experience will help them build leadership skills, as well as the capability to motivate their peers. Access To A Broad Network A data science mentor can provide an adequate connection to the valuable network of people in the community. This will facilitate a mentee to join an extensive network of data science professionals, brainstorming innovative ideas as well as preparing for the next step in their careers. Wrapping Up One-on-one mentoring in data science can help anyone who wants to gain deep insights into the data science domain, regardless of their experience and profession. Recently, Analytics India Magazine has launched AIM Mentoring Circle – joining data science enthusiasts as well as professionals, participants will have the opportunity to learn in-demand skills like AI, data science, machine learning, and upskill themselves to launch their career with community-driven mentorship.","excerpt":"The demand for data science jobs has been experiencing exponential growth over the last few years. As data science penetrates deeper into business operations, there has been a significant disruption at the workplace with data scientists and data analysts dabbling with newer tools. Almost every organization has shifted from traditional business techniques to one anchored […]","categories":["AI Features"],"tags":["Data analyst jobs","Data Science Career","Data Scientist Jobs","what is data science"],"author_name":"Ambika Choudhury","publish_date":"2020-04-19T16:00:00","publication_year":"2020","word_count":637,"keywords":["data science","Go","Data Scientist Jobs","machine learning","AI","Data analyst jobs","Data Science Career","Ray","Aim","analytics","disruption","GAN","what is data science","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Aim","Ray","R","Go","GAN","disruption"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/benefits-of-having-a-one-on-one-mentor-in-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":49868,"title":"Don’t Fall For Data Science Myths, Says Anish Agarwal Of Royal Bank of Scotland","content":"A data science professional with over 18 years of experience, Anish Agarwal, the director of data and analytics at Royal Bank of Scotland, India, is definitely one of the most influential leaders in the analytics industry. Machine Learning Developers Summit 2020 Anish Agarwal receiving Analytics 100 Award at MachineCon 2019, Mumbai With a proven ability in designing and executing strategy and facilitating new technology adoption, Agarwal says, “A lot has changed over the last two decades. It almost feels like two centuries!” Analytics India Magazine caught up with Agarwal to get an insight into his analytics journey, some of the important contributions that he has made to the field, his advice for data science enthusiasts and more. Agarwal has been involved in various emerging innovations such as data analytics, strategy, financial modelling and artificial intelligence. He is also an avid speaker, has authored papers focusing on key AI technologies, and has been instrumental in implementing several new systems into organisations. As Agarwal shares, the term data science has emerged only recently to specifically designate a profession that is expected to make sense of the vast stores of big data. Having started his career in 2001, he shares that at that time data processing methods that were used tracked not only compiled data in rows and columns but also instilled curiosity amongst colleagues to understand the magic behind the bar and pie chart. “Skill in MS Excel was most desired and anyone who knew Vlookup was adored no less than a superstar.” Says Agarwal. Agarwal confesses that he has always been good in the functional side of data and analytics as opposed to being an expert statistician with a good understanding of software architecture and multiple programming languages. Despite this, he boasts of a successful analytics career and definitely tons of things aspiring data scientists can learn from. Analytics India Magazine: With over two decades of a career in analytics, what have been some of the major highlights in your data science journey? Anish Agarwal: While there are many, one of the most interesting high-points of my career in data science is what I call as self-discovery. This was when I was preparing for an internal job posting in one of the data science organisations I was working with, I came across two new words – data analysis and data mining which made me nervous! I felt having lost the game; technology seemed to be moving fast. But, after reading the definition, I took a sigh of relief as it was just a two-word definition of what I did in my previous role. Today, we have simplified tools which allow decision-makers to become more self-sufficient. The tools are easier to use, provide the functionality needed, and are very efficient. Today, businesses can gather data and gain insights by working directly with the data. Over the course of my career journey, I have developed deep expertise in defining a problem, identify the key sources of information, designing the framework for collecting and screening the needed data, and most importantly, clearing segregating the role of a computer and a human. AIM: How did you land up a career in the data science industry? What are some of the tips you would like to give to data science enthusiasts who want to venture into this field from non-tech backgrounds? AA: Well, my Data Science journey has been thrilling. The tryst with data starting with preparing sales report which got a push when I was tasked to add a few charts in my reports. It was in 2006 when I was asked to learn the data architecture of Siebel CRM. The organisation I worked for had decided to use this system to track all stages of the sales cycle. This is when I learnt the nuances related to terms like Business Intelligence and OLAP. Smart and jazzy, isn’t it? OLAP or Online Analytical Processing was a system that allowed analysing data from a variety of sources while offering multiple paradigms or perspectives. And then, Business Intelligence went hi-tech and transformed into what we today know as a data science ecosystem comprising artificial intelligence, machine learning, Natural Language Processing, RPA and more. Transition into data science can be tough, even scary! It is not because you need to learn Math, stats and coding but you also need to battle out the myths you hear from people around you and find your own path through them! I had my own share of these “popular” myths, and they did make my life difficult. Some of these were; “You need a PhD to have a chance of becoming a data scientist. Two is even better!”“Participate in data science competitions, that will tell you how the industry works.”“You need tons of computational resources to build deep learning models. You can only get that at the top tech firms.” Even today, there are multiple myths floating around that add a false aura around data science roles. Don’t fall for them! These myths often make you feel like only geniuses can work in data science. This is just not true. Whether you’re a recent graduate, an experienced professional, or a leader, it’s important to understand how data science works and you will find your place in the industry. It’s important to understand the difference between the roles. In a vast and complex field like data science, practical experience is king. There are numerous projects you can pick up and work on right now. Or find a problem you are passionate about solving and see if data science techniques can be applied there. There are plenty of resources available online for learning. Lastly, as a beginner, you will be required to learn concepts from scratch. It will often be a frustrating experience. Your technical colleagues might know more. They might be ahead of you initially at every turn. And that’s where the dedication and discipline characteristics we spoke about earlier come into play. AIM: How important is it for data scientists to have business acumen along with technical skillsets? AA: It is imperative that data scientists have a strong foundational knowledge, understand the business challenges, and how data and technology can be leveraged to bring out the insights. As part of the data science journey, one will hear very valuable yet different perspectives on what it means to be a data science professional in the business setting, ranging anywhere from traditional data analytics and reporting to sophisticated machine learning model development and deployment. Also, in pursuit of being precise in the world that isn’t, there are attempts to clearly state the importance of business knowledge. Your expertise and strength in both technical skills and business knowledge would only get you started. However, to make a mark as a data scientist, it is imperative that you develop the right business acumen needed for business growth. Companies work on generating data touchpoints and analysing them for a reason – to understand crucial concerns associated with business operations and growth and recognise customer behaviour and deliver better services and experience. With the data you have in hand, you should be able to bring sense out of it and find angles and approaches towards business growth and potential. What is important to understand here is the difference between knowledge and expertise. You don’t need to be an expert in the business, but knowledge is important. AIM: What are some of the key challenges of setting up a data science team in organisations? AA: Challenges are not limited to data science but are common in all areas of the business. Data science today is applicable throughout every industry. I no longer must explain what I do for a living as long as I call it AI — we are peak data science hype! To make use of big data, businesses have started to either amass or hire teams tasked with deriving insights from the plethora of data collected each day. The demand for that archetypical “Data Scientist” who is the perfect blend of a statistician, programmer and communicator have never been greater. But as the dust settles, we have started hearing stories of failed projects and disenchanted professionals. Many of the challenges are centred on the need for collaboration between technology and business. Common mistakes, such as using static data or not thoroughly planning a solution’s implementation, can trip up a young data team before they complete their first proof-of-concept. As data teams mature, the challenges do not go away but instead take different forms, like deciding whether to stick with older technologies (SAS, SPSS) or opt for newer approaches (R, Python, Spark). Another challenge faced after setting the team is the unconscious bias around working with blinders on, focusing entirely on meeting project deadlines and not realising that the operating premise itself has become flawed; the foundation is not stable. Dynamic data rears its ugly head as changes in technology, languages, and stacks invalidate the usability of time-sensitive solutions. But when these teams are well organised and given the right tools and direction to succeed, they can do much more: data teams can serve as a research and development function, experimenting with raw data to explore possibilities and solutions that the business didn’t even know it had. This can include anything from uncovering new insights about customer behaviour to revealing business opportunities in new markets. AIM: How is the AI startup community shaping up in India? Is there more hype than actual usage of AI by these startups? AA: Startups, in India and in many other parts of the world, have received significant attention in recent years. Their numbers are on the rise and they are now being widely recognised as important engines for growth and jobs generation. Through innovation and scalable technology, startups have an opportunity to generate impactful solutions, and thereby act as vehicles for socio-economic development and transformation. Today’s AI start-ups do their best to emulate the functioning of the human brain’s neural networks, but they do this in a very limited way. To put it simply, they teach the algorithm what they want to learn and provide it with clearly labelled examples, and it analyses the patterns in those data and stores them for future application. The accuracy of its patterns depends on data, so the more examples you give it, the more useful it becomes. Herein lies a problem: An AI is only as good as the data it receives, and it can interpret that data only within the narrow confines of the supplied context. It doesn’t “understand” what it has analysed, so it is unable to apply its analysis to scenarios in other contexts. And it can’t distinguish causation from correlation. AI is more like an Excel spreadsheet on steroids than a thinker. Part of this deception is about riding the wave of pervasive media attention AI receives, a typical corporate eagerness not to miss the train. “It’s a buzzword”. Every day there is a bunch of news about AI, like AI beating humans in doing this and that. It’s just easy to jump into that trend. Most of the AI startups being founded and funded last year in India were into AI-enabled applications for multiple use cases in different verticals. Now, if these applications can go deep enough, thanks to top-notch Indian tech talent, we could see a tipping point for AI startups from India this year or the next. For Indian start-ups to take maximum advantage of AI and ML, large scale actionable data needs to be available to train algorithms. While the ongoing digital transformation of public and private sector enterprises transform the economy from being data-poor to data-rich, India’s strong talent ecosystem would help us accelerate the adoption. AIM: What are some of the key requirements you would like to suggest for the founders of AI start-ups to have? AA: Founding a startup can be hard. From financing your endeavour to having a skeleton staff and navigating the legalities that come with starting a new venture. Artificial intelligence is, of course, all the rage in tech circles, and the press is awash in tales of AI entrepreneurs striking it rich after being acquired by one of the giants, often early in the life of their startups. The path to success in AI requires not just technical prowess but also careful thinking and execution through a range of strategic and tactical questions that are specific to this domain and market. Here are a few tips: Do your research – Today, we’re in a much less forgiving startup ecosystem. If your competitor hits the market with a product, they can get it in front of hundreds of thousands of people via the internet and social media in a day. Whatever your idea, you can be sure there are a few other folks working on it, even more thinking about it, and thousands who will jump in as soon as they hear about it. Before you burn through your life-savings funding a startup, you first need to find out if there’s a market for your idea and what its chances are of being successful. According to the experts, one of the biggest mistakes new founders can make is misunderstanding the size of or other details about the market for their idea. Understand how to get funding – There are several ways a fledgeling business can help raise some much-needed capital. Some startups with an idea for a popular consumer product may seek out funding through crowdsourcing, whilst others may seek angel investing. Some, more mature startups, may instead seek funding from venture capitalists. It is important to keep in mind however that funding from the latter is usually tied to stringent expectations about high rates of growth and returns. Are You Sure You’re Building an AI Startup? – This is one of the more difficult or annoying questions to answer (depends who’s asking it and of whom). A better form of the question might be: “Can your startup be done without AI?” or, even more precisely, “Can your startup be done without building the AI components yourselves?” Even if your company’s vision is to use AI to deliver value, you need to ask yourself whether the AI needs to be proprietary. If AI is central to what you do, then there are good reasons (e.g., accuracy, long-term costs, data privacy, tackling new challenges) to consider building things in-house. And most importantly, are you ready for the startup lifestyle? If you are currently an employee of another company, then starting your own as an entrepreneur is a lifestyle change. Don’t make the mistake of assuming it is a way to get rich quick or an escape from all the problems. Starting a business is hard work, requires a lot of determination and learning, and only pays off in the long term. Take an honest look at yourself before leaping. AIM: Is the banking industry witnessing analytics and data science adoption to the extent that it should? What are some of the areas where it can be adopted? AA: In today’s competitive world, growing a customer base and satisfying them is considered the most challenging task. Customers demand are being treated as individuals and not as a general lot. To get over this, banks have been implementing various tools over time. Challenges like ensuring long-term loyalty from high-valued customers, retaining and attracting different types of customers or cross-selling of which products exactly to whom, fraud detection, application screening, credit and collections has always been an area of concern. Analytics comes into the picture here. It helps banks to fetch the relevant data of customers, identify fraudulent activities, helps in application screening, capture relationships between predicted and explanatory variables from past happenings and uses it to predict future outcomes. The banking industry is data-intensive with typically massive warehouses of untouched data. New models of proactive risk management are being increasingly adopted by banks. By using analytics to extract actionable intelligent insights and quantifiable predictions, banks have learnt to gain insights that encompass all types of customer behaviour, including channel transactions, account opening and closing, default, fraud and customer departure. While analytics isn’t exactly new to the world of banking, plenty of banks are gearing up for their next big analytics push, propelled by a load of data and new, sophisticated tools and technologies. The key business drivers which increase the importance of analytics within the banking industry are Regulatory reforms, Customer profitability and operational efficiency. Banks are also harnessing the power of analytics in credit scoring applications to more accurately estimate the risk associated with a potential customer. Most credit scoring methods consider the potential customer’s credit and financial history. The architecture behind the credit scoring based lending model allows banks to base their credit scoring on alternative data types such as social media posts and interactivity. This could include what sites a potential customer visits, what they purchase via eCommerce, and what they say about those sites and purchases on social media. The online behaviour of a potential customer can indicate the likelihood that they will pay back their loans and make payments on time. The sentiment becomes a data point indicating a “positive” or “negative” experience, which can then be recognised by a predictive analytics application. AIM: How do you see the analytics industry evolve in the next five years? AA: This is a terrific question with a nuanced answer. If, by analytics, we mean the specific process of taking data and explaining what happened in it – and strictly following the theory which describes the trajectory of technological evolution, then analytics as a profession will go away in five years. If by analytics, we mean the general process of taking data and extracting actionable insights, then machines will do the analysis, we’ll provide the insights, and machines will execute on our findings. Machines are becoming more and more capable of explaining what happened. Data analysis software that cost millions of dollars a decade ago is open source and free to anyone skilled enough to implement them today. Tools like Watson Analytics have the power to make detailed analysis inexpensive and accessible to anyone. NLP will generate literature reviews with no misses. Our systems will be more than capable of creating anything we need, be it automation, insights or visualisation. In the hierarchy of analytics, we are entering the predictive era. In the next 5 years, I expect us to fully embrace predictive analytics and begin venturing into prescriptive analytics. What’s definite is that data & analytics will only gain momentum for the foreseeable future and will be at the core of countless new technology solutions. Analytics will continue to focus on usability and increasing natural language that enables business users to extract data and generate insights without needing to understand the underlying algorithms. Not only will this increase efficiencies and create further adoption throughout companies, but it will also help alleviate some of the problems created by the data scientist shortage. Furthermore, the opportunities created by machine learning and artificial intelligence are endless, and it will be a race for companies to harness its power and create new services that provide value in unique ways.","excerpt":"A data science professional with over 18 years of experience, Anish Agarwal, the director of data and analytics at Royal Bank of Scotland, India, is definitely one of the most influential leaders in the analytics industry.  Machine Learning Developers Summit 2020 With a proven ability in designing and executing strategy and facilitating new technology adoption, […]","categories":["AI Features"],"tags":["Interviews and Discussions","list of computer languages and their uses","recent technological innovations"],"author_name":"Srishti Deoras","publish_date":"2019-11-14T14:15:01","publication_year":"2019","word_count":3195,"keywords":["data science","recent technological innovations","artificial intelligence","machine learning","AI","neural network","ML","NLP","Aim","deep learning","analytics","list of computer languages and their uses","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","data science","analytics","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dont-fall-for-data-science-myths-says-anish-agarwal-of-royal-bank-of-scotland\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10057694,"title":"Analytics &#038; AI firm Fractal raises $360 million investment, becomes the second unicorn of 2022","content":"Fractal, a US-based artificial intelligence firm, became the second unicorn in 2022 with an investment of $360 million from TPG Capital Asia, a private equity firm. The funding will end by the first quarter of 2022. After Mu Sigma, this is the second firm in India in the pure-play analytics space to get the unicorn status. The investment has come at a time when the analytics space is gaining huge interest from investors. Also read: Fractal Analytics & Its Fixation With Spinning-Off Product Companies Fractal was founded in 2000 in Mumbai and later moved to the US in 2005 by five people, including group CEO Srikanth Velamakanni and Pranay Agrawal, CEO. The company has about 3,500 employees working across 16 locations worldwide, including the US, UK, Ukraine, India, Singapore, and Australia. Fractal’s products include Qure.ai, which assists radiologists in their diagnostic decisions. In a statement, Srikanth Velamakanni, Co-founder & Group CEO, Fractal, said, “TPG’s capabilities across all our markets and their proven success in building and supporting top AI providers is the perfect complement to the partnership we’ve enjoyed with Apax, whose insight and expertise have been instrumental in accelerating our growth.” Pranay Agrawal, Co-founder & CEO, Fractal, said in a statement, “The demand for AI is surging across the enterprise. The investment from TPG will accelerate our ability to scale and meet this rising demand globally.” Also read: Why Fractal’s Samya Deal Is More Than Just A Product Acquisition","excerpt":"Essentially an analytics service provider, Fractal analytics is one of the world’s most well-funded AI providers. Apart from this, consultancy makes up one of the bulk of Fractal’s business.","categories":["AI News"],"tags":["datascience","Fractal Analytics","Funding","musigma"],"author_name":"Poornima Nataraj","publish_date":"2022-01-05T17:26:18","publication_year":"2022","word_count":240,"keywords":["Fractal Analytics","API","Funding","artificial intelligence","unicorn","funding","AI","programming_languages:R","datascience","analytics","R","musigma"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","API","unicorn","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/analytics-ai-firm-fractal-raises-360-million-investment-becomes-the-second-unicorn-of-2022\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168055,"title":"The World’s First AI-Assisted IVF Baby is Here and India is Watching Closely","content":"In a pivotal moment for healthcare, the world’s first infant was born using a fully automated and digitally controlled intracytoplasmic sperm injection (ICSI) process. As detailed in a research paper published on Reproductive Biomedicine Online, the case involved a 40-year-old woman diagnosed with primary infertility and diminished ovarian reserve. She sought treatment along with her 43-year-old male partner with moderate teratozoospermia, a condition that affects the count or morphology of sperm. After several failed in vitro fertilisation (IVF) attempts, the couple was referred for ICSI treatment at a fertility clinic in Mexico. Four of the five eggs treated with AI-assisted ICSI were successfully fertilised, ultimately resulting in the first live birth. The embryos were ranked using the Embryo Ranking Intelligent Classification Algorithm (Erica) AI, while computer software was used for automated sperm selection. The fully automated digital control and remote supervision of the ICSI system were successfully implemented, marking a significant milestone for the AI and robotics industry in healthcare. While the system took longer to complete the process than manual ICSI, it is expected to improve as it becomes fully automated and the need for human supervision continues to diminish. According to the paper, achieving a fully autonomous IVF process depends on how safe the system’s performance is without on-site human supervision. The latest system represents a step forward from earlier versions of automated ICSI, as it automates all micromanipulation steps of the workflow instead of selected aspects of the procedure. While the West has shown promising advancements in the healthcare industry, South Asian countries like India are also showing interest in AI-powered IVF treatments. India’s IVF market has seen a rapid surge in AI integration, supported by a ₹2,000 crore government-funded initiative called the IndiaAI Mission, which targets critical sectors including medicine. In the last 10 years, time-lapse technology has been used to monitor embryo growth during the IVF treatment. However, according to the initiative’s official website, while time-lapse imaging does not help analyse status, AI could provide the “perfect solution”. IVF Market in India IVF is a booming industry in India, with a market size of $864.6 million as of 2024. According to IMARC data, the industry is expected to experience a growth rate of 15.4% between 2025 and 2033, potentially reaching $3.4 billion by 2033. Technological advancements, increased acceptance of AI-assisted fertility measures, and accessibility are key to market growth in India. Dr Madhu L, founder medical director and senior consultant at Safe IVF Centre in Shivamogga, mentioned that increased awareness about infertility and IVF treatments has greatly improved India’s market size in the past few years. Most IVF centres in India have high costs and varying success rates through the traditional ICSI treatment. However, only some IVF centres have explored AI and automation in reproductive health, which could be a game-changer for IVF procedures. Moreover, Dr Madhu believes that people who come in for treatments in Shivamogga aren’t particularly concerned with the technology being used or how it works. “They just want results.” According to him, if treatment centres use AI analysis, the result may be successful. However, this has not been proven. “We can’t say if human intelligence is superior to the AI-assisted tools. We must wait at least five years to see what results we get [in India].” Technological Advancements in India’s Reproductive Health Field According to Sunflower Infertility and IVF Center in Ahmedabad, AI in IVF treatment has shown great potential to improve outcomes despite being in the early stages. Using advanced algorithms, AI can identify the patients’ infertility-specific data. The hospital group has also said that AI-assisted IVF cycles are shortened, leading to reduced treatment costs and more effective fertility solutions in India. “AI can be used for embryo selection and robotic ICSI for sperm injection. Previously, human intelligence was used to select the best-quality sperm. But now, the morphology of the patients is fed into machine learning from the beginning, enabling it to choose the best-quality sperm.”“Another [purpose of AI] is that it can gauge the growth of the eggs, which can be monitored every 18, 24 and 32 hours. Now, the machine is doing it. The robotic AI does the embryo selections—how much has it grown, is it equal size, and so on. The parameters we provided in the beginning give you the best option for implantation,” he added. “[Safe] IVF Centre has already started using AI to conduct semen analysis. We have been using an automated semenizer, based on AI called LensHooke, for three years. We used to do it manually earlier, but now using the machine has made the process simpler. While we are not using AI for embryo selection, I’m sure many IVF centres in India are using it,” he further explained. Meanwhile, Dr Alejandro Chavez-Badiola, co-founder and CMO at Conceivable Life Sciences, stressed that humans will always be part of the process to ensure every step goes as planned, while engineers ensure the equipment functions properly. “Ideally, this innovation will bring down the costs, improve access, and allow for more families to experience the joy of having children. It’s all very exciting; we’re making history.” Bonraybio, a team of engineers and scientists with experience in the medical instrument field, has developed products called LensHooke semen analysis systems and quality analysers that have assisted the Safe IVF Centre for the past three years. In 2022, Bloom IVF Centre, one of India’s most established fertility organisations, introduced a new AI tool named Life Whisperer AI—a fast and non-invasive method to enhance pregnancy success rates for IVF patients. Life Whisperer is a product of the French AI healthcare company Presagen. It is used to assess two embryo quality metrics. According to the product’s website, Life Whisperer is currently authorised for sale in over 40 countries. Last year, the Indian Council of Medical Research (ICMR) collaborated with Amity University and developed an AI-based tool to detect the genetic cause of male infertility and help predict IVF results. “Nowadays, AI is being used in stem cell harvesting, but it’s still in the experimental stage. Moreover, we will see growth in AI use [in reproductive health] in the next five years at least,” Dr Madhu added. Meanwhile, Sp0vum, a tech startup company, has developed a gripper-based technique called RoboICSI to assist embryologists in performing enhanced IVF treatments. Unlike traditional treatments, which require skilled embryologists to handle the delicate procedure, the gripper reduces the risk of damage and improves the fertilisation outcome, the company stated. The company is also reportedly working towards automated systems called computer-aided semen analysis (CASA) and quantitative embryo assessment techniques. Essentially, there has been a lot of growth in AI technological use in reproductive medicine, but safe deployment and frequent testing in IVF centres are critical. Many researchers in India have taken a keen interest in introducing AI-based systems that could successfully solve infertility issues and enhance reproductive health.","excerpt":"India’s IVF market has seen a rapid surge in AI integration, supported by a ₹2000 crore government-funded initiative called the IndiaAI Mission.","categories":["AI Features"],"tags":["ai in medicine","ivf","reproductive health"],"author_name":"Smruthi Nadig","publish_date":"2025-04-16T09:07:32","publication_year":"2025","word_count":1141,"keywords":["Go","reproductive health","API","machine learning","AI","ivf","Git","Ray","ViT","ai in medicine","YOLO","GAN","R"],"extracted_tech_keywords":["AI","machine learning","Ray","R","Go","Git","API","GAN","ViT","YOLO"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-worlds-first-ai-assisted-ivf-baby-is-here-and-india-is-watching-closely\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":52601,"title":"VIVO Bets On AI-Powered 5G Devices To Capture Market In 2020","content":"India is home to 55 million VIVO mobile phone consumers. The company has been in operation in the Indian market for only about 5 years now but the brand has been very well received by the public. The sales figures for 2019 are only bettered by Xiaomi and Samsung respectively. It recorded 23% of the market share for the month of October — its highest ever. The company has invested heavily in the development of 5G – AI compatibility platforms and services, which it feels will be the trend in the future. VIVO is also planning to compete with its counterparts through the introduction of innovative features and technologies. In preparation for the impending advent of 5G technology in the coming months, VIVO has fully geared itself to ride this wave. Nipun Marya, Director of Brand Strategy for VIVO in India said, “The onset of 5G services in the country presents itself as a lucrative opportunity that most mobile phone manufacturers are likely to bank on in the upcoming year.” VIVO plans on investing Rs 7,500 crores in India to boost its production (at Greater Noida plant) from 2.5 crores units to 3.34 crores units per annum. VIVO’s X30 and X30 Pro are its two 5G variants which are available in the market.","excerpt":"India is home to 55 million VIVO mobile phone consumers. The company has been in operation in the Indian market for only about 5 years now but the brand has been very well received by the public. The sales figures for 2019 are only bettered by Xiaomi and Samsung respectively. It recorded 23% of the […]","categories":["AI News"],"tags":["AI innovation","VIVO"],"author_name":"Yeshey Rabzyor Yolmo","publish_date":"2019-12-26T10:28:59","publication_year":"2019","word_count":213,"keywords":["AI innovation","programming_languages:R","AI","VIVO","RPA","R"],"extracted_tech_keywords":["AI","R","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vivo-bets-on-ai-powered-5g-devices-to-capture-market-in-2020\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61200,"title":"Latest Research By National University Of Singapore Aims To Streamline Anomaly Detection","content":"Breaches in network security that lead to financial frauds have touched an all-time high. To help businesses, communities and individuals deal with these breaches, anomaly detection has been used. It has emerged as an important issue in recent times with data scientists around the world are working on the subject. In a recent development, a team from the National University of Singapore discovered a new approach to anomaly detection – MIDAS. MIDAS stands for Microcluster-Based Detector of Anomalies in Edge Streams and has been developed by PhD candidate Siddharth Bhatia and his team (Bryan Hooi, Minji Yoon, Kijung Shin, and Christos Faloutsos). They claim that MIDAS provides a new approach to anomaly detection that is capable of outlining baseline approaches in both speed and accuracy. MIDAS detects microcluster anomalies or similar suspicious edges in graphs. The primary objective of MIDAS is to detect anomalies in real-time at a much faster pace than any other existing state-of-the-art models. The technology can help social media platforms such as Twitter and Facebook detect fake profiles that are often used for spam and phishing. It can also help investigators identify online sexual predators. That is not all – other cases of anomaly detection may include data preprocessing, credit card fraud detection, and network security, among others. Via MIDAS repository When compared to previous approaches that were used to detect anomalies in edge streams, it was found that MIDAS comes with more extensive features. This includes Microcluster Detection and Guarantee on false-positive probability. The experiment performed by the team of researchers found that MIDAS beats the baseline approaches by higher accuracy and processes the data 162 to 644 times faster. Algorithm The team of researchers proposed two approaches – the MIDAS and the MIDAS-R. The overview of the algorithm begins with Streaming Hypothesis Testing Approach, which works with the help of MIDAS. It provides guarantees on false positive probability using streaming data structures in a hypothesis test-based framework. The second one deals with Detection and Guarantees. The team decided on the procedure to determine if a point is abnormal or not, to probably acquire false-positives. The final one is Incorporating Relations, where MIDAS-R bridges a relationship between temporary and spatial edges. Accuracy Via Midas repository The curve mentioned above gives a crystal picture of the MIDAS’s accuracy. The ROC curve for MIDAS, MIDAS-R and SedanSpot clearly shows how MIDAS is 42% to 48% accurate compared to other baselines, and also runs faster. Via MIDAS repository In the second graph, the team has plotted the average precision score vs the running time. It is seen that MIDAS is much more precise by at least 27% compared to the baseline. On the other hand, MIDAS-R’s performance is even better, with an accuracy of 29%, achieving the highest average precision scores. Scalability Via MIDAS repository The graph depicts the scalability of MIDAS and MIDAS-R compared to other ages. As shown, the scalability of both the MIDAS, when compared to the processing time per edge along with an increase in the number of edges, is much higher. They achieve real-time anomaly detection with a processing rate of 4M edges with 0.5 seconds. Real-World Effectiveness Via MIDAS repository To compare how effective MIDAS could be, the team chose the Twitter Security Dataset (2.6 million tweets related to security events in 2014) for anomaly detection. To compare the performance of MIDAS, it selected baselines such as RHSS and SedanSpot. However, RHSS had a low AUC measure of 0.17 on the Darpa dataset. So, the team measured MIDAS’s accuracy, running time, and real-world effectiveness by comparing against SedanSpot. The graph shows the anomaly scores vs day, from May to September 2014. As can be seen, there are different peaks of anomalies which coincides with important events in the TwitterSecurity timeline for MIDAS. Compared to MIDAS, SedanSpot could only highlight high anomalousness scores, which led to its low AUC. Conclusion An anomaly detection algorithm like MIDAS can be applied by different industries. From detecting strange behaviour in machines that are interconnected to detecting fake news, MIDAS can play a crucial role by identifying abnormal patterns in real-time, and reducing and preventing losses. One can learn more about how MIDAS and MIDAS-R detect the anomalies in edge streams by reading the MIDAS repository and Siddharth Bhatia’s paper on the subject.","excerpt":"Breaches in network security that lead to financial frauds have touched an all-time high. To help businesses, communities and individuals deal with these breaches, anomaly detection has been used. It has emerged as an important issue in recent times with data scientists around the world are working on the subject. In a recent development, a […]","categories":["Deep Tech"],"tags":["anomaly detection","DARPA","National University of Singapore"],"author_name":"Rohit Chatterjee","publish_date":"2020-04-08T11:00:01","publication_year":"2020","word_count":716,"keywords":["Go","programming_languages:R","AI","DARPA","RPA","Scala","RAG","National University of Singapore","Aim","anomaly detection","R","fraud detection"],"extracted_tech_keywords":["AI","Aim","RAG","anomaly detection","fraud detection","R","Go","Scala","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/latest-research-by-national-university-of-singapore-aims-to-streamline-anomaly-detection\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10015928,"title":"Solorigate: What Went Behind The ‘Disastrous’ SolarWinds Hack","content":"US-based SolarWinds suffered one of the most disastrous cyberattacks of the year that has potentially compromised up to 200 organisations and agencies, including prominent names such as Intel, Microsoft, NVIDIA, and Cisco. It is now being referred to as Solorigate, first coined by Microsoft. This attack comes on the heels of a major breach at FireEye, one of the world’s most sought after cybersecurity firms. The FireEye hack was termed the biggest known cyberattack since the 2016 incident where the US National Security Agency was compromised by a little known group called the ShadowBrokers. The Solorigate In both SolarWinds and FireEye cases, it is speculated that the hackers operated on behalf of a foreign government. FireEye, which is also investigating the cause behind the SolarWind hack, said that a malware-laced update for the latter’s Orion software infected the networks of multiple US companies and government networks. Reportedly, 18,000 customers of SolarWinds received a malicious update that included a backdoor. However, the number of cases in which attackers could infiltrate using the backdoor is much lesser — currently pegged at close to 200. Experts believe that this is the case of a supply chain attack, wherein notorious elements seek to damage an organisation by targeting the weaker and lesser secure elements in the supply chain. One of the most prominent examples of supply chain attacks is the Target security breach of 2013. The Target breach is considered one of the largest data attacks in the history of the retail industry. The attack happened between November 27 and December 15, 2013, that introduced malware into the POS (point of sale) system of 1,800 stores, making 40 million credit and debit cards susceptible to fraud. This breach eventually led to the downfall of the company’s profit in Q4 of 2013 by about 46%, apart from the additional expenditure of 90 lawsuits filed against the company. What Went Wrong With SolarWinds The attackers compromised the SolarWinds Orion Platform DLL (Dynamic Link Library) with the addition of 4,000 lines of discrete malicious codes. This insertion of lines of codes in SolarWinds.Orion.Core.BusinessLayer.dll allowed the attacker to operate unfettered in the compromised networks. What made it even foolproof is the fact that the compromised file was even digitally signed which suggests that the attackers had access to the SolarWind’s software development or the distribution pipeline. As per a Microsoft review report looking at the SolarWind hack, there has been evidence to suggest that the attackers started testing their ability to insert code by adding empty classes as early as October 2019. Because of which the insertion of malicious code happened at a very early stage, possibly before the software build that would include the digital signing of the compiled code. The malware would stay dormant in the affected system for a period of up to two weeks. After this, it retrieves and executes commands called Jobs. These Jobs include transferring and executing files, profiling the system, rebooting the machine, and eventually disabling system services. The backdoor also used multiple obfuscated blocklists in order to identify any antivirus tools that may be running on the system, including processes, services, and drivers. Interestingly, FireEye discovered SolarWinds’ backdoor while investigating its own breach that was identified on December 8. Wrapping Up Supply chain attacks are a growing concern. The Solorigate incident is a reminder that these attacks achieve harmful results of the deadly combination of widespread impact and deep consequences for compromised networks. As suggested by, companies can guard themselves against these attacks by isolating and investigating devices; by identifying accounts used on affected devices; and by determining the timeline of device compromise for an indication of lateral movement.","excerpt":"US-based SolarWinds suffered one of the most disastrous cyberattacks of the year that has potentially compromised up to 200 organisations and agencies, including prominent names such as Intel, Microsoft, NVIDIA, and Cisco. It is now being referred to as Solorigate, first coined by Microsoft. This attack comes on the heels of a major breach at […]","categories":["AI Trends"],"tags":["Cyber Attack","expert system examples"],"author_name":"Shraddha Goled","publish_date":"2020-12-27T13:00:00","publication_year":"2020","word_count":609,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","Git","expert system examples","Cyber Attack","GAN","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Git","GAN","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/solorigate-what-went-behind-the-disastrous-solarwinds-hack\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10101075,"title":"Microsoft Upgrades Power Platform Copilot","content":"At Microsoft‘s second edition of the Power Platform Conference, significant advancements were unveiled in Power Platform Copilot, alongside noteworthy designer updates for Power Automate and the introduction of automated environment routing to streamline maker onboarding. With the integration of advanced AI to enhance low-code development, it was revealed that more than 126,000 organizations have harnessed the capabilities of Power Platform Copilot. Notably, the Microsoft Power Platform community has witnessed remarkable growth in the past year, now boasting over 5.2 million monthly active members. New Copilot Capabilities Simplify Website Development Developers can now easily create data-driven websites and multi-step forms using Copilot in Power Pages. By describing their desired website in natural language, developers instruct Copilot to generate sitemaps, homepage layouts, and site themes. This intuitive process streamlines website creation, reducing it to just a few sentences and clicks, providing a substantial boost in productivity. Customisable Copilot in Power Apps The Copilot control in Power Apps has evolved to empower makers to fully customize their AI assistants and extend Copilot through Power Virtual Agents. This includes the ability to add websites and unstructured content alongside structured data sources, significantly expanding Copilot’s capabilities across various apps. Faster Development with Power Automate Microsoft is accelerating development with the new Power Automate flow designer. Copilot is now enabled by default, enhancing the authoring experience and improving flow success rates. User feedback has led to refinements, including an expanded ability to populate more parameters in the flows and actions generated by Copilot. Streamlined Governance and Security in Managed Environment Managed Environments now offer automation for routing makers to their dedicated development environments, ensuring scalable and governable app development. Power Platform pipelines automate application lifecycle management (ALM) and provide robust security practices. Copilot generates deployment notes and application descriptions in Managed Environments, enhancing security compliance. Additionally, Advisor in Managed Environments offers proactive recommendations and inline actions to address security threats at scale. Expanding the Power Platform Community Following the success of the Power Up program, which attracted over 22,000 participants from 180 countries, Microsoft introduces group learning opportunities. Additionally, passionate individuals interested in contributing to the Power Platform community are invited to join the Super Users group.","excerpt":"Developers can now easily create data-driven websites and multi-step forms using Copilot in Power Pages.","categories":["AI News"],"tags":["Microsoft"],"author_name":"Mohit Pandey","publish_date":"2023-10-04T13:27:27","publication_year":"2023","word_count":361,"keywords":["Go","AI assistants","AI","data-driven","ML","Scala","automation","ViT","GAN","R","Microsoft"],"extracted_tech_keywords":["AI","ML","AI assistants","R","Go","Scala","GAN","ViT","automation","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-upgrades-power-platform-copilot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10116715,"title":"NVIDIA Rewrites Moore’s Law with Blackwell","content":"At NVIDIA GTC 2024, CEO Jensen Huang, while announcing their ‘very big GPU’, Blackwell, sort of bid adieu to the ‘good old days of Moore’s law’. Reflecting on the rapid advancement in computing power, Huang said that in just eight years, NVIDIA has increased computational capacity by a thousandfold, a progress that far exceeds the benchmarks set during the heyday of Moore’s Law. However, he rued that even after this remarkable growth, the industry’s accelerating demands are far from being met. “In the past eight years, we’ve increased computation by 1000 times, and we have two more years to go. So that puts it into perspective [the fact that] the rate at which we’re advancing computing is insane. And it’s still not fast enough,” said Huang. Huang said that the future is generative, which is why we call it ‘generative AI’, marking the start of a brand new industry. He agreed that their approach to computing is fundamentally different from their competitors. “We’ve created a processor specifically for the generative era,” he said, and added, “A critical component of this is what we call ‘content token generation’, which we format as FP4.” Further, he said that this involves a significant amount of computation – 5x the token generation and 5x the inference capability of Hopper. “That might seem sufficient, but we asked ourselves, why stop there?” pondered Huang, and said that the answer is that it is not enough. On the contrary, Intel is still hooked on Moore’s Law. Intel: Moore’s law is not dead.Nvidia: Moore’s law is dead.Who’s right? Check the market caps.— Pedro Domingos (@pmddomingos) March 5, 2024 Right after the announcement, many experts and folks from the ecosystem took to social media to declare that this was the new Moore’s law, or the era of Huang’s Law, as you may call it. Unleashing Blackwell, the GPU Beast NVIDIA’s latest GPU architecture is named after David Harold Blackwell, an eminent American statistician and mathematician – who has made significant contributions to various fields, including game theory, probability theory, information theory, and statistics. NVIDIA’s Blackwell is a game-changing AI platform for trillion-parameter scale generative AI. The B200 GPU delivers 20 petaflops of power, and the GB200 offers 30x LLM inference workload performance, enabling efficiency to reach new heights. Blackwell features a second-gen Transformer Engine to double AI model sizes with new 4-bit precision. The 5th-gen NVLink interconnect also enables up to 576 GPUs to work seamlessly on trillion-parameter models. An AI reliability engine maximises supercomputer uptime for weeks-long training runs. The new Tensor Cores and TensorRT-LLM Compiler in the Blackwell platform significantly reduce the operating cost and energy consumption for LLM inference, by up to 25 times compared to its predecessor. Major tech giants like Amazon, Google, Microsoft, and Tesla have already committed to adopting Blackwell. Huang said that training a GPT model with 1.8 trillion parameters [GPT-4] typically takes three to five months using 25,000 amperes. The Hopper architecture would require around 8,000 GPUs, consume 15 megawatts of power, and take about 90 days to complete. In contrast, Blackwell would need just 2,000 GPUs and significantly less power (only four megawatts) for the same duration. He said that NVIDIA aims to reduce computing costs and energy consumption, thereby facilitating the scaling up of computations necessary for training next-generation models. To showcase Blackwell’s scale, NVIDIA also unveiled the DGX SuperPOD, a next-gen AI supercomputer with up to 576 Blackwell GPUs and 11.5 exaflops of AI compute. Each DGX GB200 system packs 36 Blackwell GPUs coherently linked to Arm-based Grace CPUs. NVIDIA vs Intel vs AMD In contrast, Intel recently launched its Ponte Vecchio GPU based on the Xe-HPC architecture under the Data Center Max GPU branding. The company is facing delays with its discrete GPU roadmap for gaming and consumer products, which could impact its future AI training capabilities. On the other hand, AMD’s latest offering is the Instinct MI300 accelerator series based on its CDNA 3 architecture. The flagship MI300X promises up to 1.6x higher AI inference performance per chip than NVIDIA’s H100. However, for the crucial AI training workloads, the MI300X still trails the H100 in raw performance metrics like FP8 throughput. In comparison, for AI training performance, the B200 offers up to 2.5x higher FP8 throughput per GPU over the previous Hopper generation. But its real strength lies in inference – the new FP6 numeric format effectively doubles throughput over FP16, enabling up to 30x higher performance for large language model inference compared to Hopper. Blackwell also packs a massive memory bandwidth of 8TB\/s and up to 192GB per B200 GPU. While Intel and AMD are making progress, NVIDIA’s Blackwell platform raises the bar significantly through architectural innovations tailored to the unique demands of trillion-parameter AI models. Given that Moore’s law has been declared officially dead, it would be interesting to see how this space evolves in the coming months. “GPU prices will drop once AMD becomes usable,” said Abacus.AI chief Bindu Reddy, sharing interesting predictions on compute for the next five years.","excerpt":"While announcing their ‘very big GPU’, Blackwell, Huang sort of bid adieu to the ‘good old days of Moore’s law’.","categories":["Global Tech"],"tags":["moore's law","NVIDIA"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-03-19T14:40:13","publication_year":"2024","word_count":837,"keywords":["Go","API","AI","innovation","ML","GPT","Aim","llm_models:GPT","generative AI","NVIDIA","R","moore's law"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","R","Go","API","GPT","innovation","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidia-rewrites-moores-law-with-blackwell\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10152354,"title":"Discover the Next Big Thing in AI at AWS AI Conclave 2025","content":"Amazon Web Services is all set to host the 8th edition of the AWS AI Conclave in Bengaluru on January 24, 2025, at the Sheraton Grand in Whitefield. The conclave promises to bring together industry leaders, innovators, and technology enthusiasts to explore the latest advancements in generative AI, data foundations, machine learning, and cloud technology. The event is ideal for AI\/ML and data CXOs and practitioners, business leaders, founders, technology entrepreneurs, and developers. Sandeep Dutta, president of AWS India and South Asia, will deliver the welcome note at the event. Why Attend the Amazon AI Conclave 2025? Attendees can look forward to: Live Keynotes: Hear from AWS leaders about the latest products, capabilities, and features that simplify the adoption of generative AI at scale for companies of all sizes. Technical Sessions: Engage in over 30 technical sessions covering a wide array of topics, from building and scaling generative AI to implementing robust data strategies. Industry and Builder’s Hub: Explore the real-world applications of AI and interact with experts to gain insights into building with AWS and cloud computing. AI Awards: Recognise and celebrate innovations in the AI space, acknowledging organisations and individuals making significant contributions to the field. REGISTER NOW AWS  recently concluded its annual flagship event re: Invent 2024 in Las Vegas, where it introduced Amazon Nova, a family of foundation models accessible through Amazon Bedrock. These models are built for tasks such as content generation, video understanding, and developing agentic applications. They are available in six different sizes to cater to diverse requirements. The Amazon Nova lineup includes: Amazon Nova Micro: A lightning-fast text-to-text model optimised for speed and cost-efficiency. Amazon Nova Lite and Pro: Multimodal models capable of processing text, images, and videos. Amazon Nova Premier: Set to launch in early 2025, it focuses on complex reasoning tasks. Amazon Nova Canvas: A powerful image generation model with advanced editing features. Amazon Nova Reel: A cutting-edge video generation model producing studio-quality content. AWS announced the general availability of Amazon Bedrock, a fully managed service that provides access to foundation models from leading AI companies via a single API. The cloud giant has also announced new chips, including Trainium2, Graviton 4, and Inferentia, which are challenging NVIDIA. AWS claims that Trainium2 offers 30-40% better price performance compared to the previous generation of GPU-based Elastic Compute Cloud (EC2) instances.  Customers and foundation model partners such as Anthropic, Databricks, Adobe, Qualcomm, Poolside, and even Apple are already on board. The company also unveiled Trn2 UltraServers and the next-generation Trainium3 AI training chip. In a surprising revelation at AWS re: Invent, Apple’s senior director of machine learning and AI, Benoit Dupin, announced that the company leverages AWS’s custom AI chips for various cloud services. This collaboration has resulted in a 40% efficiency gain for Apple’s services, including Siri, Apple Maps, and Apple Music. Furthermore, Apple is evaluating Amazon’s latest Trainium2 chip to pre-train its Apple Intelligence models, which could potentially improve efficiency by up to 50%. The conclave will feature sessions exploring key updates from AWS re: Invent 2024, such as Amazon Nova’s multimodal capabilities and the performance improvements of Trainium2 chips. Event Details and Registration Date: January 24, 2025 Time: 9 am to 5 pm (IST) Venue: Sheraton Grand Hotel and Convention Center in Whitefield, Bengaluru. This event is ideal for: AI\/ML & Data CXOs & Business leaders Founders and Co-founders AI\/ML practitioners & Technology Entrepreneurs Developers Don’t miss this opportunity to be at the forefront of AI innovation and cloud computing. Register now to secure your spot and join a community that drives innovation and business impact through AI. REGISTER NOW","excerpt":"Whether you’re a business leader, developer, AI\/ML practitioner, or technology enthusiast, this event provides a platform to learn, network, and collaborate.","categories":["AI Highlights"],"tags":["AWS"],"author_name":"Siddharth Jindal","publish_date":"2024-12-30T14:22:53","publication_year":"2024","word_count":601,"keywords":["Anthropic","machine learning","Amazon Nova","AWS","AI","ML","RAG","Ray","Aim","generative AI","foundation models"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","foundation models","Anthropic","Amazon Nova","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/discover-the-next-big-thing-in-ai-at-aws-ai-conclave-2025\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":37622,"title":"How To Build The Ultimate  AI &#038; ML A-Team","content":"Businesses across the world are leveraging the power of data to build tech-enabled solutions. But this is not something which can be achieved overnight. To become a company fully capitalize on AI and ML, it must first assemble a strong team — A-Team. So, how to build a team that knows the nuts and bolts of AI and ML; a team that has the capability to solve some of the biggest problems; a team which is the A-Team? Challenges While Building A Dedicated AI Team Artificial Intelligence and Machine Learning require tremendous knowledge to become an expert. And when you are on the go to establish a dedicated department for AI and ML, it is imperative to focus on some of the certain roles and skill sets. Therefore it is always considered to be good practice to take a strategic approach and focus on the key roles and areas where you want your AI team to work on. However, there is a challenge — talent scarcity. People spend many years to master the sorcery of these two techs. So, when it comes to building a team, finding the right talent is something that takes time. “The biggest challenge companies face while hiring is finding candidates that is a good fit for their organisation. How does one identify if a prospective candidate is a right fit for an organisation via just a couple of discussions or online tests?,” said Ashish Sam, senior partner, people and operations at TheMathCompany. Some of the measures that companies can take to shortlist the right candidate is to give them situations\/case studies and understand their reactions\/solutions and set up an interaction session with the leadership team. Furthermore, one cannot just randomly pick a person who has significant experience in the field of AI or ML. Why? Because that might have enough experience but might not be a perfect fit for the job you want to get done. How to overcome this challenge? There is a challenge in that a lot of data science and AI recruits are generalists, meaning that they come with a broad set of skill sets unlike who can work in specific areas such as NLP or image processing. “Experts have specific skills in one technology or domain and are an excellent fit for solving their domain-specific problems. There are, however, challenges associated with having only experts in a team\/organisation as expertise comes with a limited shelf life. Having a generalist on board helps in the cross-pollination of ideas not just across technologies, but across industries. However, a generalist may not aid in understanding the depth and expanse of the problem one is trying to solve,” said Sam. Prior to going the right way with the hiring process of AI experts, pen down the business challenges you want to deal with. Also, list the pain-points you want to solve and what expertise you want your candidate to have to solve those pain-points. This approach of knowing the business challenges and pain-points will help you define key components of building an AI team. Meaning, when you know exactly what you are looking for, you reduce the chances of hiring the wrong candidate. That is not all, another reason is the cost behind setting up a dedicated AI and ML team. Therefore is always advised to take time and figure out the important pieces of the puzzle. Investing In Training & Reskilling The job is not done by simply hiring the talent and assigning them work. If you want a team of experts to work on your AI and ML challenges, you have to provide them with an ecosystem that fits their needs. And an ecosystem also includes training and briefing. People might wonder, if we are hiring people who have significant experiences, then why is the need to train them. To answer that, we have to understand the fact just because they have enough experience that doesn’t mean they would be familiar with the business the company has. So, in order to make it easier to understand what are the things that the business is trying to solve and how things work, training is imperative. Also, investing in training is something that delivers value and it is considered to be one of the best practices. Coming to another point about the ecosystem, tools play a major role. When professionals are provided with the right tools to work on, efficiency and productivity increase significantly. For instance, if your AL and ML consist of professionals such as number-crunchers, a modern day term of data analyst, then s\/he should be provided with the top-notch tools that help them carry out analysis. Outlook It is not always necessary to build a dedicated team for a specific vertical, you have to hire new talent; sometimes, the void can be filled by the existing employees. There are many companies out there that follow the “Training and Upskilling” formula. So, if you think your company has employees who are capable of taking on new responsibilities and work on a complete AI and ML field, then definitely educate them and provide them with the necessary training. Many companies also provide online courses and hands-on experience to their employees in order to assign them new responsibilities.","excerpt":"Businesses across the world are leveraging the power of data to build tech-enabled solutions. But this is not something which can be achieved overnight. To become a company fully capitalize on AI and ML, it must first assemble a strong team — A-Team. So, how to build a team that knows the nuts and bolts […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Machine Learning"],"author_name":"Harshajit Sarmah","publish_date":"2019-04-11T10:50:37","publication_year":"2019","word_count":873,"keywords":["data science","Go","API","artificial intelligence","machine learning","AI","ML","Machine Learning","RAG","NLP","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","data science","RAG","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-build-the-ultimate-ai-ml-a-team\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10020876,"title":"Hands-on Guide To GANSynth: An Adversarial Neural Audio Synthesis Technique","content":"GANSynth is a state-of-the-art method for synthesizing high-fidelity and locally coherent audio using Generative Adversarial Networks (GANs). Hence the name GANSynth (GAN used for audio Synthesis). It was introduced by Jesse Engel, Kumar Krishna Agrawal, Shuo Chen, Ishaan Gulrajani, Chris Donahue and Adam Roberts – researchers at the Google AI in 2019 (research paper). Autoregressive models like WaveNets generate audio sequentially. On the contrary, GANSynth creates the whole sequence in parallel, synthesizing audio much faster on GPU runtime than real-time synthesis. It generates the entire audio clip from a single latent vector, allowing for easier release of global features like pitch and timbre (tone quality). It uses progressive GAN architecture. It eliminates the drawback of traditional GANs which struggle to synthesize locally coherent audio waveforms though they use global latent conditioning and efficient parallel sampling. Are you interested in understanding the detailed workings of GANSynth? Refer to this page before proceeding! Practical Implementation of GANSynth Here’s a demonstration of how GANSynth learns to produce musical notes of individual instruments as contained in the NSynth dataset (a large-sized qualitative dataset having annotated notes). The GAN learns to use its latent space for representing various instrument timbres. It synthesizes audio from MIDI files and interpolates between different instruments. The code has been implemented in Google colab using Python version 3.7.10. Step-wise explanation of the code is as follows: Install Magenta (an open-source Python library, powered by Tensorflow) #Copy data from the GCS (Google Cloud Storage) !rm -r \/content\/gansynth &>\/dev\/null !mkdir \/content\/gansynth !mkdir \/content\/gansynth\/midi !mkdir \/content\/gansynth\/samples # Load default MIDI (Bach Prelude) #’curl’ command enables fetching a given URL !curl -o \/content\/gansynth\/midi\/bach.mid http:\/\/www.jsbach.net\/midi\/cs1- 1pre.mid -o option provided with the curl command saves the downloaded file on your local machine with the name specified as the parameter. SONG = '\/content\/gansynth\/midi\/bach.mid' !curl -o \/content\/gansynth\/midi\/riff-default.mid http:\/\/storage.googleapis.com\/magentadata\/papers\/gansynth\/midi\/arp.mid RIFF = '\/content\/gansynth\/midi\/riff-default.mid' !pip install -q -U magenta Import required libraries and classes import os #module for interacting with the operating system #To load files from local device (weblink) from google.colab import files import librosa #Python library for music and audio analysis from magenta.models.nsynth.utils import load_audio from magenta.models.gansynth.lib import flags as lib_flags from magenta.models.gansynth.lib import generate_util as gu from magenta.models.gansynth.lib import model as lib_model from magenta.models.gansynth.lib import util import matplotlib.pyplot as plt #for visualization import note_seq from note_seq.notebook_utils import colab_play as play #colab_play() inserts an HTML audio widget to play a sound in colab import numpy as np import tensorflow.compat.v1 as tf #disable_v2_behavior() switches all global behaviors which vary between #tensorflow 1.x and 2.x versions to behave as in 1.x. tf.disable_v2_behavior() Define a function for uploading .wav file def upload(): map = files.upload() #Upload the file list = [] Initialize list to store names of uploaded files #Use iteritems() to iterate over key-value pairs of the dictionary of uploaded file content for key, val in map.iteritems(): filename = os.path.join('\/content\/gansynth\/midi', key) with open(filename, 'w') as file: #open the file in write mode #write the content of uploaded file to the specified file file.write(val) print('Writing the file {}'.format(filename)) list.append(filename) #Add the filename to the list return list Define global variables #checkpoint directory CHECKPOINT_DIR = 'gs:\/\/magentadata\/models\/gansynth\/acoustic_only' OP_DIR = '\/content\/gansynth\/samples' #output directory BATCH_SIZE = 16 SR = 16000 #SR stands for Sample Rate Create an output directory if it does not exist #Expand the path of parent directory using expand_path() OP_DIR = util.expand_path(opdir) #tensorflow.gfile.Exists() determines existence of a file if not tf.gfile.Exists(OP_DIR): #Create a directory using tensorflow.gfile.MakeDirs() tf.gfile.MakeDirs(OP_DIR) Load the model #Clear the default graph stack and reset the global default graph tf.reset_default_graph() myflags = lib_flags.Flags({ #Dictionary for storing and accessing flags 'batchSizeSchedule': [BATCH_SIZE], 'tfdsData': \"gs:\/\/tfds-data\/datasets\", }) #Create a GAN model using flags and weights from a saved model model = lib_model.Model.load_from_path(CHECKPOINT_DIR, myflags) Define a function for loading MIDI file as a notesequence def midiLoad(path, minimumPitch=36, maximumPitch=84): midiPath = util.expand_path(path) #Expand the directory path noteSequence = note_seq.midi_file_to_sequence_proto(midiPath) #Define NumPy arrays to store pitches, velocities, start and end #times of each note pitches = np.array([n.pitch for n in noteSequence.notes]) velo = np.array([nt.velocity for nt in noteSequence.notes]) startTimes = np.array([nt.start_time for nt in noteSequence.notes]) endTimes = np.array([nt.end_time for nt in noteSequence.notes]) #Keep only the notes in required pitch range valid = np.logical_and(pitches >= minimumPitch, pitches <= maximumPitch) #Store the valid notes’ features in the form of a dictionary notes = {'pitches': pitches[valid], 'velocities': velo[valid], 'startTimes': startTimes[valid], 'endTimes': endTimes[valid]} return noteSequence, notes Create an attack, sustain and release amplitude envelope (these are the stages of envelope generator) ‘Attack’ is part of the envelope which represents time taken by the amplitude to reach its peak.’Sustain’ is the duration for which sound is held before it fades out.’Release’ is the final reduction in amplitude over time. def createEnvelope(note_length, attack=0.010, release=0.3, sr=16000): #sr means sample rate note_len = min(note_length, 3.0) attack = int(sr * attack) sustain = int(sr * note_len) release = int(sr * release) total = sustain + release  #attack envelope doesn't add to sound length env = np.ones(total) #1’s equal to total count # Linear attack env[:attack] = np.linspace(0.0, 1.0, attack) #Evenly spaced numbers from 0 to 1. Number of points equal to ‘attack’ # Linear release env[sustain:total] = np.linspace(1.0, 0.0, release) #Evenly spaced numbers from 1 to 0. Number of points equal to ‘release’ return env Define a function to combine multiple notes from a single audio clip. def combine_notes(audio, start, end, velo, sr=16000): #’audio’ is an array of audio notes, ‘start’ is an array of note’s start #time in seconds, ‘end’ is an array of note’s end times in seconds, ‘sr’ is #the sample rate (integer) numberOfNotes = len(audio) #Number of notes clipLen = end.max() + 3.0 #compute length of audio clip clip = np.zeros(int(clipLength) * sr) #generate audio clip for t_start, t_end, velocity, i in zip(start, end, velo, range(numberOfNotes)): # Generate an amplitude envelope noteLen = t_end - t_start #compute note length #call createEnvelope() defined above env = createEnvelope(noteLen) len = len(env) #length of generated envelope audio_note = audio[i, :len] * env # Normalize the notes audio_note \/= audio_note.max() audio_note *= (velocity \/ 127.0) clipStart = int(t_start * sr) #start time of audio clip clipEnd = clipStart + length #end time of clip #Add the audio note to clip buffer clip[clipStart:clipEnd] += audio_note #Normalize the audio clip clip \/= audio_clip.max() clip \/= 2.0 return clip #Array of combined audio samples Define a function to plot spectrogram def spectrogram(audioClip): min = np.min(36) #minimum number of MIDI notes max = np.max(84) #maximum number of MIDI notes #Get the frequency of MIDI notes in Hertz(Hz) minF = librosa.midi_to_hz(min) #minimum frequency maxF = 2 * librosa.midi_to_hz(max) #maximum frequency #number of octaves octaves = int(np.ceil(np.log2(maxF) - np.log2(minF))) binsPerOctave = 36 #number of bins in each octave nBins = int(binsPerOctave * octaves) #number of bins #Calculate constant-Q transform of the audio signal C = librosa.cqt(audioClip, sr=SR, hop_length=2048, fmin=minF, n_bins=nBins, bins_per_octave=binsPerOctave) #’audioClip’ is the audio time series # ‘sr’ is the sampling rate of audioClip # ‘hop_length’ is the number of samples between successive CQT #columns #‘fmin’ is the minimum frequency # ‘n_bins’ is the number of frequency bins #Compute power of the audio signal power = 10 * np.log10(np.abs(C)**2 + 1e-6) #Display the ‘power’ array as a matrix in a new column window using #matshow()of matplotlib plt.matshow(power[::-1, 2:-2], aspect='auto', cmap=plt.cm.magma) plt.yticks([]) plt.xticks([]) Choose the MIDI file midi_file = \"Arpeggio (Default)\" #@param [\"Arpeggio (Default)\", \"Upload your own\"] This will allow you to choose the default uploaded MIDI file or upload a file of your choice as follows: #Path of the default uploaded file midi_path = RIFF #If user chooses ‘Upload your own’ option if midi_file == \"Upload your own\": try: fileList = upload() #Upload your file midi_path = fileList[0] #Path of recently uploaded file #Load the uploaded file noteSeqence, notes = load_midi(midi_path) except Exception as e: #Throw an exception if uploading fails print('Upload Cancelled') else: # Load the default uploaded file, but slow it down 30% noteSequence, notes = load_midi(midi_path) notes['startTimes'] *= 1.3 notes['endTimes'] *= 1.3 #Plot the notesequence note_seq.plot_sequence(noteSequence) Output: Choose some random instruments to generate custom interpolation. Audio ‘interpolation’ means making the audio sound better. #Select number of instruments number_of_random_instruments = 10 #@param {type:\"slider\", min:4, max:16, step:1} A slider will appear as follows which will allow you to choose number of instruments from 4 to 16, in step of 1 pitchPreview = 60 num = number_of_random_instruments pitches = [pitchPreview] * num #Compute pitch #Generate latent vactor latent_vector = model.generate_z(num) #Generate fake samples for latents and pitches of all the instruments audio_notes = model.generate_samples_from_z(latent_vector, pitches) for i, audio_note in enumerate(audio_notes): #Print the instrument number print(\"Instrument: {}\".format(i)) #Insert the HTML audio widget for each instrument’s audio file; pass the array of float sound i.e. audio_note and specify sample rate as parameters play(audio_note, sample_rate=16000) Audio files of the instruments: Instrument0 Instrument1 Instrument2 Instrument3 Instrument4 Instrument5 Instrument6 Instrument7 Instrument8 Instrument9 Sample output showing widget for each instrument’s sound: (You can play the audio, adjust its volume and download it using the widgets) 13) Create a list of instruments to interpolate between instruments = [0, 2, 4, 0] Place each instrument at a specific point of time (from 0 to 1.0) times = [0, 0.3, 0.6, 1.0] Start and end times of synthesized audio times[0] = -0.001 times[-1] = 1.0 14) Latent vectors of selected instruments z_instruments = np.array([latent_vector[i] for i in instruments]) End times for selected instruments t_instruments = np.array([notes['endTimes'][-1] * t for t in times]) Get interpolated latent vectors for each note z_notes = gu.get_z_notes(notes['startTimes'], z_instruments, t_instruments) 15) Generate audio for each note print('Generating {} samples...'.format(len(z_notes))) audio_notes = model.generate_samples_from_z(z_notes, notes['pitches']) 16) Combine the audio samples of all instruments into a single audio clip ac = combine_notes(audio_notes, notes['startTimes'], notes['endTimes'], notes['velocities']) 17) Play the synthesized audio print('\\nAudio:') #Create audio widget; pass the clip and specify the sample rate play(ac, sample_rate=SR) 18) Plot the spectrogram using spectrogram() function defined in step (10) print('CQT Spectrogram:') spectrogram(ac) Synthesized audio outputGoogle colab notebook of the above implementation can be found here. References For more information about GANSynth, refer to the following web links: Research paperMagenta DocumentationGitHub repository","excerpt":"GANSynth is a state-of-the-art method for synthesizing high-fidelity and locally coherent audio using Generative Adversarial Networks (GANs). Hence the name GANSynth (GAN used for audio Synthesis). It was introduced by Jesse Engel, Kumar Krishna Agrawal, Shuo Chen, Ishaan Gulrajani, Chris Donahue and Adam Roberts – researchers at the Google AI in 2019 (research paper). Autoregressive […]","categories":["Deep Tech"],"tags":["GANs"],"author_name":"Nikita Shiledarbaxi","publish_date":"2021-02-26T13:00:00","publication_year":"2021","word_count":1648,"keywords":["NumPy","TPU","AI","ML","RAG","Colab","Ray","Matplotlib","Python","GANs","TensorFlow"],"extracted_tech_keywords":["AI","ML","Ray","TensorFlow","Colab","NumPy","Matplotlib","RAG","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-to-gansynth-an-adversarial-neural-audio-synthesis-technique\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10141816,"title":"Anthropic Claude Now Lets You Chat with Your Google Docs","content":"Anthropic has released a new set of features for Claude.ai. Most importantly, it integrates Google Docs with the AI chatbot to ‘directly reference’ content inside your chats. You can add a Google Doc file in a chat or a project, which allows Claude to access the same and use the information to generate contextual outputs. “For example, Claude can summarise long Google Docs and reference historical context from the files to inform decision-making or help with strategic planning,” said Anthropic. That said, this feature is only available for Claude Pro, Team and Enterprise plans. Anthropic also mentioned that once a document is added, Claude will automatically sync to its latest version. Moreover, you can also add multiple Google Doc files in a single chat. However, Anthropic says that Claude cannot access images, or comments from the document. Another feature in the recent update lets you choose between three output styles: Formal, Concise, or Explanatory. “Whether you’re a developer writing formal documentation, a marketer crafting clear brand guidelines, or a product team planning extensive project requirements, Claude can adapt to your preferred way of writing,” said Anthropic in the announcement. The new update also includes custom styles, which basically lets you add preferences and pre-requisite knowledge to the chatbot’s output. Similarly, you can also add ‘global profile preferences’ to ensure Claude considers the information when generating an output. In this case, OpenAI launched the Custom Instructions feature to ChatGPT a year early. Moreover, recent updates to GPT-4o focused on bolstering its creative writing capabilities, and it also stood atop several benchmarks. However, it isn’t all good news for Claude users. It seems like Claude isn’t letting users access Sonnet 3.5 on the free version. While several users on X were able to spot that the model can only use Haiku, even we can confirm the same, based on what we noticed on the free version of Claude. On the bright side, Haiku will keep getting better, as mentioned by Dario Amodei, CEO at Anthropic. Why no one have talked about the fact that claude switched the model from Claude 3.5 sonnet to 3 haiku for all free plan users 🙄 pic.twitter.com\/qdedLUzGyE— Mouad (@nadzi_mouad) November 27, 2024 In a recent podcast episode with Lex Fridman, he said that Anthropic’s plan is to first ‘shift’ the capabilities of existing models. He even went on to say that Haiku 3.5, its smallest model, is as powerful as Opus 3. As Haiku continues to get better, Amodei also confirmed that there are plans in place for the flagship Opus 3.5. Moreover, Anthropic recently announced another round of funds, raising $4 billion from Amazon Web Services (AWS), bringing Amazon’s investments in the AI startup to $8 billion.","excerpt":"The feature rollout spree continues, and this time, the onus is on personalisation.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Anthropic AI","Claude"],"author_name":"Supreeth Koundinya","publish_date":"2024-11-27T23:35:59","publication_year":"2024","word_count":451,"keywords":["Anthropic","ChatGPT","Go","TPU","OpenAI","AI","AWS","GPT-4o","Claude","Claude 3.5","R","Anthropic AI","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","GPT-4o","ChatGPT","OpenAI","Claude 3.5","Anthropic","AWS","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/anthropic-claude-now-lets-you-chat-with-your-google-docs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10083408,"title":"Locus Delivers Till Last Mile, Solves Logistics Challenges for Nestle, Tata, Unilever","content":"Last-mile delivery is one of the most challenging supply chain problems since it accounts for more than 50% of all shipping cost. Mostly, it’s a host of external factors such as rising fuel costs, imprecise addresses, increasing urban congestion, and the high costs of returns and re-deliveries that add to this challenge. “To keep up with the surge in e-commerce orders and rising pressure to deliver at scale, enterprises worldwide, and in India, are struggling to manage their last-mile logistics operations efficiently,” Nishith Rastogi, Founder and CEO of Locus, told Analytics India Magazine. Locus, which Rastogi co-founded in 2015, is leveraging modern day technologies to solve the problem of last-mile delivery in India. “Locus was founded to empower enterprises to convert their last mile from a cost centre to a revenue generator by making every delivery more efficient than the last, empowering workers to do more, bring sustainability to operations, uncover deep inefficiencies and power delivery experiences,” Rastogi said. Further, in this exclusive interaction, Nishith Rastogi discusses in detail the use of AI\/ML in last-mile delivery, about the e-commerce and logistics sector and whether the concept of BOPIS will work in India. AIM: What does Locus do? Nishith: Locus is a leading-edge logistics software technology company dedicated to transforming last-mile logistics. Our real-world-ready dispatch management platform empowers businesses to transform their last-mile logistics fulfilment into a growth centre through advanced optimisation algorithms and intuitive workflow automation. We push the boundaries of defining excellence in an enterprise’s last-mile logistics activities by enabling efficiencies at scale, empowering workforces to do more, providing end-to-end visibility, uncovering deep inefficiencies, improving and automating decision-making, and fostering sustainable growth. Simply put, our platform enables businesses to make every delivery more efficient than the last, which results in fewer miles travelled and more deliveries completed in a day. As a result, the platform also equips workforces for excellence and enables companies to scale with lower emissions while simultaneously encouraging sustainable growth. The platform is being implemented and used by more than 200+ clients across 30+ countries globally to power more than 650 million deliveries. We are proud to count Unilever, Nestle, Bukalapak, the Tata Group, and BlueDart as some of our most valued customers. We have powered over 650 million deliveries to date, and our technology has helped save USD 200 million in transit costs, offsetting 70 million kilograms in CO2 emissions while maintaining a 99.5% SLA adherence ratio. AIM: How are technologies like IoT\/Big Data analytics\/AI\/ML\/Cloud Computing impacting today’s innovation in the Logistic Industry? Nishith: Technology is playing an instrumental role in digitalising the logistics sector, especially the last-mile fulfilment processes. With the right technology solutions, businesses can scale up their operations, uncover underlying inefficiencies in real time and do much more to uncover new growth opportunities. For instance, technology can help empower everyone—from the Chief Supply Chain Officer to warehouse managers and drivers to make more deliveries at the right time, at the right cost, and with the right experience. Furthermore, it can also help reduce carbon emissions from vehicles significantly, further empowering the sector to achieve its sustainability goals. Our Dispatch Management Platform leverages IoT, AI, ML, and Big Data Analytics technologies, enabling businesses to optimise their last-mile logistics operations. IoT serves as a foundation for various features provided by our Dispatch Management Platform, such as Route Planning, Carrier Management, tracking and tracing of drivers, and more. Through IoT, relevant real-time data and updates from drivers are added to the platform and vice versa, which helps deliver business efficiency. GPS tracking is also a key aspect that helps understand the vehicle’s movement and generates alerts if it deviates from the planned route to the destination. It is a key component in enabling automated notifications to prevent the breach of service-level agreements. AI plays a crucial role in creating optimal routes, whereas Big Data Analytics helps unearth deep inefficiencies in last-mile fulfilment processes. Cloud computing provides the necessary computing power and the bandwidth to help enterprises across the globe attain last-mile excellence and scale up when demands arise. AIM: How are you specifically leveraging AI\/ML to improve last-mile delivery in India? Nishith: At Locus, we leverage the most advanced AI, ML, and data analytics systems architectured by our data scientists to create our product suite. To elaborate a bit further, as Locus powers increasingly complex supply chains, dispatchers rely on Locus’ AI to compute billions of combinations to quickly seek out the ideal sequence by which every order can be fulfilled with the least resources spent. Furthermore, our machine learning algorithm continually observes these decisions made by the AI. It evaluates how they play out in the real world, improving each on-ground execution, and continuously improving performance. AIM: Will the concept of BOPIS work in India? Nishith: The concept of Buy Online and Pick Up In-Store (BOPIS) gained global prominence recently, particularly during the pandemic years. The BOPIS fulfilment model was primarily conceptualised to bridge the gap between the traditional and the modern retail. However, given the rise in quick commerce and omnichannel retailing, BOPIS has become a win-win fulfilment model for retailers and consumers. For retailers, BOPIS can save last-mile shipping costs, attain more in-store traffic and sales and better insights on the availability of products in real-time whereas, for consumers, it delivers a hassle-free ecommerce experience that helps them navigate the shipping costs. Interestingly, not just BOPIS but also BORIS (Buy Online and Return In-Store) has been picking up steam in India recently. Some of the high-street retail brands in India successfully operate through this fulfilment approach, indicating an untapped potential for such omnichannel retail concepts. Through our technology and proprietary algorithms fueled by AI, data analytics, and more, retailers can fill the gaps in BOPIS retail distribution and better optimise last-mile processes. Retailers can unlock the power of data to find patterns in customer shopping carts, seasonal trends, and market dynamics to offer more personalised BOPIS shopping offers. Furthermore, analysing and identifying customer shopping behaviours can help retailers predict demands and store the right amounts of stock to deliver a delightful customer experience. Similarly, through our time-slot management system, retailers can allot specific pick-up times to customers—avoiding unnecessary queues and allowing store managers to plan to package daily orders efficiently. All of this makes the whole pick-up process more transparent and reliable.","excerpt":"Dispatchers rely on Locus’ AI to compute billions of combinations to quickly seek out the ideal sequence by which every order can be fulfilled with the least resources spent.","categories":["AI Features"],"tags":["AI Tool","Interviews and Discussions"],"author_name":"Pritam Bordoloi","publish_date":"2022-12-27T13:00:00","publication_year":"2022","word_count":1050,"keywords":["Go","machine learning","AI","cloud computing","ML","Git","RAG","Aim","analytics","AI Tool","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","cloud computing","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/locus-delivers-till-last-mile-solves-logistics-challenges-for-nestle-tata-unilever\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":38859,"title":"How Successful Data Scientists Deal With Bad Appraisals","content":"2018-19 turned out to be quite underwhelming for data scientists, especially in terms of monetary compensation, and it has hit many professionals hard this appraisal season. It can be demotivating to recover from a less-than-stellar appraisal, especially when it does not align with your performance review. Many professionals, especially in the thriving data science field ended up getting an unexpectedly low salary hike this year. If you are one of the professionals who got an unsatisfactory raise, you might be feeling indignant, embarrassed or even disillusioned — especially because there has been a lot of noise about how great and thriving this industry is. In this article, we will discuss what appropriate steps should be taken to deal with this predicament. Starting with keeping cool, and moving towards a brighter future, Analytics India Magazine will lay down the game plan: First of all, here are the hard facts: It’s Not You, It’s The Industry: Despite being the country’s hub for the IT sector, employees in Bengaluru this year are expecting only a 0-10% increment. In fact, most of the IT bellwethers all across India, have announced an average pay hike of 6% to a majority of their employees. Not A Good Time For Middle Management: Numerous reports have suggested that this is not the best time to be a data scientist — especially an experience of more than 7 years. An experienced data scientist, as a middle-management employee, has found him\/herself in a precarious position in 2018-19. Reportedly, companies who are dealing with financial crunch think that employees who hold middle-level designations like Project Manager or Project Architect, and whose projects are not yielding the desired revenues, must be asked to leave. Case in point, Cognizant is planning to hire more junior-level data scientists, while the job for many mid-level employees is in danger. How to cope with these changes? Self Evaluation It may sound like a piece of rote advice, but taking a look at one’s own performance and the resulting monetary appraisal level is the place to start with the coping mechanism. Was there anything about your performance as a data scientist that you’d like to do over? If there are many such instances, then perhaps you can’t put the whole blame of the bad appraisal on the manager of the company. Discussion With The Boss It is always a good idea to speak to the boss about your unsatisfactory performance review. Many noted HR representatives think that there could be a possibility that a manager may have overlooked or not remembered some critical activities and initiatives that a data scientist has taken throughout the year. Reminding the manager of your coding proficiency or even critical problem-solving examples is a good idea at this point in time. Self Goals The demand for data scientists has increased in the Indian tech industry, but so has the supply. As a result, salary growth has moderated, especially at the middle-level, but is still one of the “sexiest” for beginners. As a data scientist takes on more and more managerial responsibilities, it becomes difficult, especially with the inflated pay packages, to match up to the salary hikes in the industry. That is why the self-goal for a data scientist should be to become irreplaceable in the organisation. Now, this broad goal can entail taking on more responsibilities or acting as a secret sauce that keeps the entire data science team working cohesively in the organisation: Whichever the path you choose, your manager must see you as someone who cannot be replaced in the team, no matter what level of seniority you hold. Upskilling Wipro’s Sohini Mehta had earlier told AIM that it is always preferred by big companies to reskill than onboarding new people. “Hiring my sound very simple but the fact is that there is a huge scarcity of talent in the market and getting people may come at a very high price point,” she said. “Reskilling existing workforce in the company not only builds the talent pipeline but helps in improving employee stickiness to a great extent as you would have loyal people who will stay with you and work with you for longer,” she added. Future path While going back to work after a bad appraisal may bring you down, you have to understand that this one year will only be a small blip in your illustrious career as a data scientist. Whichever organisation you work for, always remember to: Invest in new capabilities Make digital transformations co-exist with traditional services Re-skill with emerging technologies Be open to adopting new technologies and solutions","excerpt":"2018-19 turned out to be quite underwhelming for data scientists, especially in terms of monetary compensation, and it has hit many professionals hard this appraisal season. It can be demotivating to recover from a less-than-stellar appraisal, especially when it does not align with your performance review. Many professionals, especially in the thriving data science field […]","categories":["AI Features"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-05-09T11:50:43","publication_year":"2019","word_count":764,"keywords":["data science","Go","AI","Git","RAG","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-successful-data-scientists-deal-with-bad-appraisals\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10011588,"title":"Have you Heard About the Video Dataset of Day to day Human Activities","content":"ActivityNet is an enormous video dataset that covers exercises that are generally pertinent to how people invest their energy in their everyday living. It was developed in 2015 by the researchers: Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanemand Juan Carlos Niebles1. This video dataset gives tests from 203 movement classes with a normal of 137 untrimmed recordings per class and 1.41 movement occurrences per video, for an aggregate of 849 video hours. The number and length of every video are increased in comparison to other ones. The dataset furnishes a rich action pecking order within any event four levels of depth. For example, the activity cleaning windows falls under the fourth tier category dressing, washing belongs to the second category Grooming interior cleaning falls under Grooming oneself and finally, household activities belong to the category Personal care. Here, we will discuss data contained in this dataset, how it was gathered, and provide some benchmark models that gave high accuracy on this dataset. Further, we will implement the ActivityNet dataset using Pytorch and Keras Library. Data Collection Usually Gathering data using the conventional approach takes a lot of time. To resolve this issue modern tools are employed to save scientist hours. The first step was to look through the web to recover recordings identified with every movement. Utilizing the text-based queries the researchers had searched the information in platforms like YouTube and Google recordings. They confirmed all recordings recovered and eliminated those not related to the movement at hand. Next Amazon Mechanical Turk was used to survey each video and decide whether it contains a planned movement class. Because of the error of text-based inquiries, numerous recordings that are not related to any of the expected movement classes were eliminated. Loading the dataset using Pytorch The dataset can be downloaded from the following link. Import all the libraries required for this project. import torch import torch.nn as nn from torch.utils.data import DataLoader from torch.nn.utils.rnn import pack_padded_sequence, pad_packed_sequence import torch.optim as optim import transforms.spatial_transforms as spatial import transforms.temporal_transforms as temporal from dataset.activitynet_train import ActivityNetCaptions_Train # transforms spa = spatial.Compose([spt.CornerCrop(size=args.imsize), spt.ToTensor()]) tpa = temporal.Compose([tpt.TemporalRandomCrop(args.clip_len), tpt.LoopPadding(args.clip_len)]) #Loading the dataset train = ActivityNetCaptions_Train(args.root_path, ann_path='train_fps.json', sample_duration=args.clip_len, spatial_transform=spa, temporal_transform=tpa) Let’s define the parameters in the ActivityNet dataset Class: · root  – It is the root directory of the UCF101 Dataset. ·   annotation_path – It contains the split files. ·   sample_duration – Duration of sample that captures a human activity. ·   spatial_transform and temporal_transform – Transforms the data.For example it resizes an image or crops an image. dataloader = DataLoader(train, batch_size=args.batch_size, shuffle=True, num_workers=args.n_cpu, drop_last=True, timeout=100) Training Results Loading the dataset in Keras Install the video generator using the pip command. We have used an Image generator for data augmentation. import os import glob import keras from keras_video import VideoFrameGenerator Let’s define the parameters so that we can pass it to the model for training. classes = [i.split(os.path.sep)[1] for i in glob.glob('videos\/*')] classes.sort() # Parameters Size = (112, 112) channel = 3 Nbframe = 5 Batch_size = 32 # Data augmentation data_augment = keras.preprocessing.image.ImageDataGenerator( zoom_range=.1, horizontal_flip=True, rotation_range=8, width_shift_range=.2, height_shift_range=.2) # Create video frame generator load_data = VideoFrameGenerator('data\/train\/', classes=classes, nb_frames=Nbframe, split=.33, shuffle=True, batch_size=Batch_size, target_shape=Size, nb_channel=channel, transformation=data_augment , use_frame_cache=True) State of the Art The current state of the art on ActivityNet dataset is G-TAD. The model gave an accuracy of 34.09%. BMN is a close-by competitor with an accuracy of around 34%. Conclusion In this article, we have discussed the details and implementation of a huge human movement dataset, ActivityNet. It is made conceivable by an enormous and nonstop video assortment that is effectively adaptable to bigger quantities of exercises and bigger examples per movement, at a low cost. It additionally contains a bigger number of classes and tests per classification than customary activity datasets.","excerpt":"ActivityNet is an enormous dataset that covers exercises that are generally pertinent to how people invest their energy in their everyday living. It was developed in 2015 by the researchers: Fabian Caba Heilbron, Victor Escorcia, Bernard Ghanemand Juan Carlos Niebles1. ActivityNet gives tests from 203 movement classes with a normal of 137 untrimmed recordings per class and 1.41 movement occurrences per video, for an aggregate of 849 video hours.","categories":["Deep Tech"],"tags":["Computer Vision","Keras","Pytorch","Video Dataset"],"author_name":"Ankit Das","publish_date":"2020-11-13T16:00:26","publication_year":"2020","word_count":624,"keywords":["Pytorch","Go","data augmentation","Keras","AI","PyTorch","Video Dataset","ViT","ai_frameworks:PyTorch","Computer Vision","CLIP","RNN","R"],"extracted_tech_keywords":["AI","PyTorch","Keras","R","Go","CLIP","RNN","ViT","data augmentation","ai_frameworks:PyTorch"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/have-you-heard-about-the-video-dataset-of-day-to-day-human-activities\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10071359,"title":"How did text-to-image tools become so commercialised","content":"Seven years ago, in 2015, AI innovation was marked by an important development – automated image captioning. ML algorithms could be used to label objects in image datasets which could further be turned into natural language descriptions using automated image captioning. This feature is usually directed toward persons with vision problems. This research inspired a certain curiosity in the research community. A group of scientists from the University of Toronto went a step ahead and decided to flip the process to answer the question: what if these natural language descriptions could be used to generate images instead? The task was far more complex than producing text from image datasets. The model was trained on a large-scale dataset called Microsoft COCO and could also generalise beyond the training set to produce entirely novel images. The images were based on captions that were highly unlikely to occur in real-life situations and looked something like this. Source: Research Paper The images then may not have been high quality, but the breakthrough itself led the way to a promising future. With the release of OpenAI’s DALL.E and successor DALL.E 2 this year, the future is finally here. In April this year, OpenAI chief Sam Altman announced the launch of DALL.E 2 and invited followers to give the most random, surreal prompts they could imagine. Altman posted the photogenic results that faithfully represented the instructions to oohs and aahs on Twitter. A revolution in AI image generation DALL.E 2 was the starting point for what has now become a revolution in text-to-image generation within AI. In a report by Wired, a PhD candidate at Penn State, Vipul Gupta, who received early access to the tool, noted, “What people thought might take five to 10 years, we’re already in it. We are in the future.” Initially, OpenAI mentioned in their blog that DALL.E 2 wasn’t yet ready for commercial use but could be used eventually in fields like art, marketing and education. The company reasoned that DALL.E 2 could admittedly churn out images that were sexist, racist and could be hateful by nature. The company formed a ‘red team’ comprising external experts who started looking closely at the tool’s biases. DALL.E 2 was opened up only to 400 people who were mainly OpenAI or Microsoft employees. At this, a big chunk of Twitter users expressed their disappointment regarding the decision. Developers and designers were eager to get their hands on it. Some complained that OpenAI’s exclusivity created a sense of ‘eliteness’ in AI, and many others were simply impatient. The company’s justification didn’t look good enough. Source: Huggingface.co Competitive environment It soon became obvious that the world couldn’t wait long enough. On June 6, Hugging Face noticed the usage of its AI image generation tool, DALL.E Mini, had shot up to around 50,000 images generated in a day. The app was developed by Boris Dayma, an independent ML consultant who replicated DALL.E at a hackathon organised by Hugging Face and Google in July last year. Dayma said that he became deeply interested in the tool after studying the DALL.E research paper. The images that DALL.E Mini generated were of a much lower quality than OpenAI’s original tool, but it was open source. Turns out it was enough to get people hooked already. Regular people, including non-developers, started using DALL.E Mini to exercise their imagination. Where DALL.E 2 was essentially doing the work of an artist, the availability of DALL.E Mini had turned what was conceptually a similar tool into a meme generator. Everyone could now have a piece of the future. People began posting these images and ‘memes’ they had created using DALL.E Mini Twitter and Reddit. The image quality improved. Ironically, DALL.E Mini became so popular that Dayma was recently requested to change the tool’s name (it is now called Craiyon). In a span of a few months, text-to-image generation tools are now dime-a-dozen. A few tools like Midjourney produce high-quality images, others not so much. But most are free to all. This, despite the fact that these tools produced images with similar biases like DALL.E 2. Interestingly, Google, like OpenAI, recently said that the company wouldn’t release its image generation tool Imagen to the public due to risks of misuse. The even more recent, Make-a-Scene, the creative art-focused image generator released by Meta, also noted that it would be open exclusively to specific AI artists. Fear of criticism from misuse The difference is clear – prominent tech companies, including the Microsoft-backed OpenAI, were cautious enough to avoid criticism that could arise from the dangers around the usage of these tools. DALL.E 2 images were good enough to be used to attach to, say, a fake news report. It wasn’t to say that these same issues couldn’t be due to other copycat tools, but the less prominent companies did not have the weight of their reputation to carry. However, the sudden competition among image generators appears to have forced OpenAI to move faster toward opening up DALL.E 2, lest it should lose its position as the industry leader. The company announced today that it would be expanding access to the tool in a blog through a beta release. OpenAI aims to fasten the waitlist process and add up to a million users within the next few weeks. The tool, which had been free up until now, will have a credit-based fee. DALL.E 2 will now also cater to artists who might not be able to afford it by providing subsidies. “Expanding access is an important part of our deploying AI systems responsibly because it allows us to learn more about real-world use and continue to iterate on our safety systems,” OpenAI explained in the blog. Meanwhile, it has continued to work on the tool’s biases and introduced a technique that would make the images more inclusive in terms of race and gender. For better or worse, AI-generated art has become more or less democratised. One could argue that art (even if it isn’t as good) should be accessible to all, just like AI. But how great a decision this is, only time will tell.","excerpt":"Interestingly, Google, like OpenAI, recently said that the company wouldn’t release its image generation tool Imagen to the public due to risks of misuse.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Poulomi Chatterjee","publish_date":"2022-07-21T16:15:00","publication_year":"2022","word_count":1019,"keywords":["Go","Hugging Face","OpenAI","AI","innovation","ML","Aim","ViT","GAN","AI Tool","R"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","Hugging Face","R","Go","GAN","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-did-text-to-image-tools-become-so-commercialised\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10068578,"title":"The pertinent dilemma with India’s data policy","content":"Last February, the Ministry of Electronics and Information Technology (MEITY) released the ‘Draft India Data Accessibility & Use Policy, 2022’. The document showed MEITY has plans to set up an India Data Management Office (IDMO) to oversee the metadata frameworks and their implementation. It also suggested forming a regulatory body, India Data Council (IDC). The IDMO will consult with relevant ministries and state governments to revise the National Data Governance Framework rules. Internet consumption in India. Source: CISCO Why do we need data regulation laws? “The private sector may be granted access to select databases for commercial use. Given that the private sector has the potential to reap massive dividends from this data, it is only fair to charge them for its use,” noted the 2019 National Economic Survey. The Draft India Data Policy aligns with the government’s data monetisation push. The idea is to keep all the datasets open by default, except for a few that will be marked ‘negative’. The India Data Council will sort the high-value datasets based on their “degree of importance in the market, degree of socio-economic benefits, impact on India’s Al strategy and performance on key global indices.” The IDC has the power to determine the metadata and data standards for every central department. Last October, Karnataka became the fifth state after Telangana, Odisha, Sikkim and Punjab to monetise citizen data. Number of connected devices (in millions) in India. Source: Ericsson Mobility Privacy protection The policy proposal drew flak over the lack of data protection laws. While the policy promised anonymity, there was no regulatory framework to support it. The Non-Personal Data Governance Framework had been on the backburner since 2020. The government is working on a ‘Digital India Act’ to replace the Information Technology Act 2000. The act will replace the Personal Data Protection Bill. The PDP bill tried to introduce data localisation– inspired by GDPR–to restrict information flow geographically. However, the data localisation was promptly rejected by corporate stakeholders as it would require companies to overhaul their data architectures. Revisions The draft has been revised to include anonymisation standards in the wake of the backlash. MEITY has also noted that private companies must be “encouraged” to share non-personal data with startups and Indian researchers through a proposed initiative called the India Datasets programme. “The constitution of an Appellate Grievance Redressal Committee is another major highlight of the proposed amendment. This is a crucial and positive policy action that will provide adequate recourse for users against the decisions of grievance officers. However, to ensure adequate checks and balances on the powers of the committee, the Rules must also envisage mandatory public disclosure of their orders and its periodic review by an oversight body as prescribed in the Blocking Rules. It will also be important to envisage a robust appointment process with an adequate number of judicial and technical members in the committee to ensure an informed decision making process,” said Kazim Rizvi, founding director of the public policy think tank, The Dialogue. Rizvi said, while the introduction of the updated regulations was the right direction, changes are due. “The proposed amendment to the IT Rules, 2021 is a progressive step that will ensure that our laws and policies remain updated to cater to the emerging challenges in the digital space. We welcome the Ministry’s invitation for inputs on the proposed amendments and hope that this exercise will be helpful in understanding the concerns on other key aspects of the Rules as well and create a more inclusive regime for Platform Regulation. The proposed changes effectuate the objective of greater transparency and accountability in the digital ecosystem which is critical for realising the goals of a free and rights enabling online space. The introduction to the draft also notes that the changes will not impact startups and will only be applicable to the Big Tech platforms. This step reverberates the government’s commitment to ease entry barriers in the sector and promote Ease of Doing Business. However, appropriate change must also be made to the relevant Rule 3 of the IT Rules given that its title still reads that the mandate is applicable to all digital intermediaries. The proposed amendments themselves emphasise on the duty of the intermediaries to respect constitutional rights. It is imperative to realise that stringent timelines without content gradation according to the anticipated degree of harm can lead to chilling effects on online free speech of the citizenry, a concern well discussed by the Hon’ble Supreme Court in the Shreya Singhal judgement. A risk based approach needs to be adopted and the timelines for takedown should be revisited accordingly,” he added. Global data protection laws Most countries have strict data protection laws. The European Data Protection Board (EDPB) drafted rules for transferring sensitive and personal data outside EU countries. Organisations were required to take consent from individuals before transferring their personal data. Post-Brexit, Britain had to adopt GDPR to facilitate the free flow of data between the former and the EU. China also drafted its Personal Information Protection Law (PIPL) in the second half of 2020. The PIPL granted the residents the power to withdraw permission to share their data and the right to delete data while ensuring protection during cross-border data transfers. Canada is revamping its data protection laws to include the private right to action and hefty fines. In sum, India has a lot of catching up to do.","excerpt":"The private sector may be granted access to select databases for commercial use.","categories":["IT Services"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-06-08T14:00:00","publication_year":"2022","word_count":900,"keywords":["Go","AWS","AI","Git","RAG","ViT","data governance","GAN","R","startup"],"extracted_tech_keywords":["AI","RAG","AWS","R","Go","Git","data governance","GAN","ViT","startup"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-pertinent-dilemma-with-indias-data-policy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093768,"title":"Intel Aurora: A Last Ditch Effort for Supercomputer Dominance","content":"In 2009, Intel claimed supercomputing dominance, as its chips powered over 80% of the world’s top 500 supercomputers. While those times have passed and Intel has since lost its lead to AMD, it seems redemption is on the horizon. The new Aurora supercomputer, built on Intel’s Sapphire Rapids line of chips and designed for Argonne National Laboratory’s simulation workloads, is poised to take over AMD’s position as the premier supercomputing chip. While these eternal rivals continue to duke it out in the generalised supercomputer world, competition is gradually becoming intense as other tech giants seem to have bigger plans. In a bid to reduce dependency on chipmakers and optimise their infrastructure in the process, Google, Meta, Amazon, and others have begun creating their own chips. In a market that’s more competitive than ever before, can Intel expand its foothold and take back the supercomputing crown? Battle on an exascale Until the announcement of Aurora, AMD’s Frontier supercomputer, built for the Oak Ridge National Laboratory’s scientific research requirements, was the undisputed king of the supercomputer world. Ever since its launch in 2021, Frontier has been the only supercomputer capable of exascale computing. A supercomputer capable of executing over 1 quintillion floating point operations per second (1 exaflop) is termed an exascale supercomputer. According to the Top500 list, a list of top supercomputers in the world, Frontier is currently the only exascale supercomputer in operation, but Aurora looks ready to change that. Announced at the ISC High Performance Conference, Aurora was poised to be Intel’s answer to AMD’s untouched exascale dominance. However, with the released spec sheet, it seems that Aurora will not only challenge Frontier, but dominate it. While Aurora’s journey to the exascale level has been fraught with delays and reworks, it seems that the supercomputer is finally seeing the light of day. First announced in 2015, the project was delayed twice, once in 2017, and again in 2020. This comes as no surprise, as Intel’s Sapphire Rapids line, the chips powering Aurora, have also been delayed multiple times. Now that the chip is finally entering the market, Intel has found a rare opportunity to shine out and take a significant title from its core competitor. However, there is still another problem that intel cannot move beyond: power consumption. Frontier consumes 21 megawatt (MW) of power, and was, for a while, the most efficient supercomputer in the world. Aurora, on the other hand, is predicted to consume over 60 MW of power. Indeed, this is becoming a new problem for older data centres and compute farms. Reports have noted that newer CPUs from both Intel Xeon and AMD EPYC consume upwards of 350W each, with NVIDIA’s GPUs being even more power-hungry. In the light of rising energy prices across the world and a higher focus on sustainability, companies are looking for more efficient, specialised chips. These chips can bring a host of improvements apart from just power consumption, sweetening the pot for companies willing to create their own chips. Bigger isn’t always better While supercomputers like Frontier and Aurora are targeted for research on topics like nuclear fusion, low-carbon technologies, cancer, and subatomic particles, big techs have trained their sights on one single goal — AI. Google, Meta, Amazon, and even Microsoft are working towards freeing themselves from Intel, AMD, and NVIDIA for AI compute. Google’s latest TPUv4, specialised for TensorFlow operations, is 1.2x-1.7x faster and 1.3x-1.9x more efficient than NVIDIA’s A100 chips. AWS offers a whole suite of chips for both AI training and inferencing workloads, with the chips’ specialised nature being a natural fit for AI tasks. Meta recently announced a new chip called the Meta Training and Inference Accelerator. As the name suggests, MTIA is a specialised chip for Meta’s internal AI workloads. The chip directly replaces CPUs in the data centre, and is made to work alongside GPUs for greater efficiency. Microsoft’s AI chip efforts are still shrouded in mystery, but reports have emerged that the company is working on a chip all the same. Codenamed Athena, this specialised chip will be optimised for AI workloads, especially those of OpenAI. It is worth noting that Intel is also looking to compete in the same market, with Aurora being somewhat of a proof of concept for the enterprise success of its GPUs and CPUs. Sapphire Rapids is as of yet unproven, and has received somewhat of a lukewarm response from the market. However, it seems that Intel’s Data Centre GPU Max series might just give NVIDIA a run for its money. According to Intel, the newest line of GPUs outperforms the NVIDIA H100 by an average of 30%-50%. The Xeon Max Series CPU also beats out AMD’s Genoa chips by 65%. With its Gaudi2 series of deep learning accelerators, Intel is looking to take more slices of NVIDIA’s pie, cutting into the latter’s market share slowly, but surely. While AMD is currently dominating the leaderboard, Intel’s new set of chips is hinging on Aurora’s impact to make a splash in the enterprise world. Even as big techs pour R&D costs into creating their own chips, it seems likely that Intel might claw its way back to victory from the jaws of defeat in the HPC (high-performance computing) market.","excerpt":"Intel’s long-delayed supercomputer, Aurora, might be what it needs to come back to power in the HPC market.","categories":["Global Tech"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-05-23T18:30:00","publication_year":"2023","word_count":871,"keywords":["Rapids","TPU","OpenAI","AI","AWS","RAG","Aim","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","deep learning","OpenAI","Aim","TensorFlow","Rapids","RAG","AWS","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/intel-aurora-a-last-ditch-effort-for-supercomputer-dominance\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119325,"title":"US Govt ‘Snubs’ Musk and Zuckerberg, Keeps ’em Out of AI Safety Board","content":"This week, bigwigs from several leading AI companies, including OpenAI’s Sam Altman, NVIDIA’s Jensen Huang, and Microsoft’s Satya Nadella, joined the freshly minted Artificial Intelligence Safety and Security Board formed by the US government. This follows as a countermeasure to several cases of deepfakes being used against politicians, celebrities and even children. In Florida this week, police arrested an 18-year-old for generating sexually explicit images of women without their consent. Likewise, the usage of “nudification” programmes and GenAI to create deepfakes as either blackmail or harassment material has been pervasive. Especially in American schools, as reported by The New York Times. So, the formation of a federal board hosting some of the biggest names within the industry was a step in the right direction. However, there already seems to be bad blood around who gets to be on the board. Yann LeCun and (or) Mark Zuckerberg should have been in this list pic.twitter.com\/UunykWfriT— Aravind Srinivas (@AravSrinivas) April 28, 2024 Several people pointed out that Meta CEO Mark Zuckerberg and Tesla CEO Elon Musk had not been included in the list of board members released by the Department of Homeland Security (DHS). Theories raged on why the mighty duo was “snubbed”. Let’s take a look at what these companies have been doing in terms of safety and why the two may have been overlooked. We were snubbed.— Yann LeCun (@ylecun) April 28, 2024 Why Sans Zuck and Musk? Secretary of homeland security Alejandro Mayorkas clarified that Zuckerberg and Musk’s exclusion had to do with the exemption of social media sites. However, many don’t seem convinced. safe to assume the justification was created after the decision— Terry Winters (@terrortheerror) April 26, 2024 Though YouTube and LinkedIn are seen as social media sites, the two didn’t start out as such. But it still doesn’t explain why the DHS would exclude them solely for being social media companies. Well, the thing is that the giants, by virtue of being social media companies first and foremost, have had run-ins with several governments in the past. Meta is set to face an EU probe for allegedly not doing enough to prevent Russian disinformation on Facebook. Other issues include a lack of curbs on ads promoting the aforementioned nudification apps. Similarly, Media Matters had earlier released a report alleging that advertisements were appearing opposite antisemitic posts. This subsequently led to major advertisers pulling their ads from the website and Musk retaliating with a lawsuit against the platform. This doesn’t bode well for the duo’s stance on AI safety, especially with the board focusing on advising the DHS and other stakeholders on potential AI disruptions. Alternatively, Zuckerberg has been a big proponent of open-source AI, which is harder to regulate or advise on in terms of safety. A majority of child sexual abuse material (CSAM) and other inflammatory material comes from open-source datasets that have been crowdsourced. Meanwhile, some believe that Musk wasn’t a first choice due to his unpredictability, which, considering his current plight with the SEC, is unsurprising. Not gonna be long now before the US gov pushes musk out of his companies, like this latest Australian nonsense I don’t believe for 1 second it was not instigated within the US. The establishment want him gone, too unpredictable for their liking.— SheWolf (@lonelyShewolf66) April 26, 2024 While the companies involved with the safety board are by no means bastions of safety, they have been considerably more open to safety talks. In the recent child safety consortium that Meta, Google, Microsoft, OpenAI and other major players pledged themselves to, most companies apart from Meta, had issues with people using their services to generate CSAM by abusing GenAI. This is a far more difficult issue to tackle than Meta’s, which was the lack of moderation in terms of defamatory and sexually explicit ads that are run on the site. “We understand that people get around these safeguards all the time, and so we try to design a safe product…We’re not an advertising-based model, we’re not trying to get people to use it more and more,” Altman had said last year during his Senate hearing. Pro-active AI Safety Measures Advertising aside, these companies already have their own AI safety codes in place. OpenAI, for instance, has their Approach to AI Safety blog that states they “work to improve the model’s behaviour with techniques like reinforcement learning with human feedback, and build broad safety and monitoring systems”. Likewise, the others included on the list have their variants of the AI safety framework. But whether they work or not is another conversation altogether. Last year, a research paper from Stanford found that the LAION 5B dataset, which was used to train several text-to-image models, including Stable Diffusion and Google’s Imagen, included CSAM. Similarly, big tech has had an influx of bad-faith actors who attempt to circumvent existing GenAI guardrails, which is just as hard to curb. With a rising concern about deepfakes used to generate CSAM, there has been plenty of independent research on how to mitigate these issues, including watermarking. Companies like Adobe and Google have adopted watermarking practices in the form of Content Credentials and SynthID. OpenAI has done the same with Dall.E 3, making use of C2PA. However, while these provide context on whether an image is AI-generated or not, as well as providing tamper-proof metadata, these rectify only a minor concern in the overall deepfake debate. Essentially, slapping a bandaid on a gaping wound. Not to mention that social media websites have systems in place to strip media of metadata to prevent access to a user’s location and other details. Another paper from 2022 suggests the creation of a CSAM database to train GenAI on how to detect potential CSAM. However, with the sensitivity around CSAM content, they suggest a way to extract attributes from the dataset that can then be used to train GenAI on potentially explicit content. That makes for an interesting concept. However, while this is another step in the right direction, it still leaves a lot to be considered. With more emphasis in place on AI safety, a combination of independent research and big tech adoption, much like in the case of watermarking, seems to be a sound solution towards mitigating the widespread use of deepfakes across the internet.","excerpt":"The DHS said that Zuckerberg and Musk’s exclusion had to do with the exemption of social media sites, but many don’t seem convinced.","categories":["AI Trends"],"tags":["Elon Musk","Jensen Huang","Mark Zuckerberg","Sam Altman","Satya Nadella"],"author_name":"Donna Eva","publish_date":"2024-04-30T18:30:00","publication_year":"2024","word_count":1046,"keywords":["Satya Nadella","Go","API","GenAI","Sam Altman","artificial intelligence","OpenAI","AI","AWS","Jensen Huang","Elon Musk","Mark Zuckerberg","RAG","stable diffusion","R"],"extracted_tech_keywords":["AI","artificial intelligence","GenAI","OpenAI","RAG","AWS","R","Go","API","stable diffusion"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/us-govt-snubs-musk-and-zuckerberg-keeps-em-out-of-ai-safety-board\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10010907,"title":"IDLE vs Pycharm vs Spyder: Choosing The Right IDE For Machine Learning","content":"In this article, we will be discussing how to choose the right IDE\/editor for machine learning and understanding whether there is any right IDE for it too or not. Come along! Machine Learning has surely added a lot in terms of technological scope for innovation. With a lot of enthusiasts in, it surely will give the best returns in it. In our previous articles, we had discussed which would be the right one to code in Machine Learning. In it, we had concluded that one may choose as per its own likings and work requirements. So here taking it to be a base of the previous article we will be reviewing some well known IDE\/Editors in python. Why python? Because it is one of the most widely used languages for machine learning and there is a high probability that a beginner would choose it over R because it is relatively easier to understand. Introduction Let us start with the basics to understand what basically is an editor or an IDE. So IDE stands for Integrated Development Environment. An IDE is nothing but the software in your system which will help you with many comprehensive capabilities. The basic segments in an IDE are:  source code editor, build automation tools and a debugger. With all the support from different packages, be it inbuilt or not, an IDE will help you to set your code up with a console where one can see the output. Different IDE’s for Python Let us take a look at the most popular Python IDEs used for machine learning implementations: IDLEPycharmSpyder And many more. To keep articles brisk and fathomable I will discuss a few IDE’s and will answer the following questions. Which IDE for what purpose?How to download and get started with them?Having a wide knowledge of different IDE good or not? Firstly we need to understand that it is always advisable to use IDE’s as per your requirement and work. You should work with the one which your system supports the best. A lighter IDE with not many facilities is better than the one which is dense and takes a considerable time to open and operate on your system. So here one needs to have the wisdom to select the things on the basis of the system requirements. And one should definitely be reading the system requirements before downloading any IDE for the same purpose. Now coming up with the IDE’s I prefer. Yes, you read it right: “IDEs”. I myself don’t stick to a single IDE. It is really important for one to understand things don’t really go that smoothly if we stick to only one. When it comes to implementation, output and completion of work are more important than sticking to an IDE. So it is really important to have more than one in your system because it acts as a backup like if in case it didn’t work in IDE-A, I have backed the plan up with IDE-B. Which IDE\/platforms do I use and why do I do it? 1. IDLE When I was a beginner it was important for me to understand the concept of indentation in python so I had started off with IDLE. It is a great platform to learn the basics of code, learn more about bugs and indentation. The best thing about IDLE is that it is inbuilt and comes with Python. So when you install python an IDE, IDLE gets installed automatically into your system. You can run your code by pressing F5. So if you are a beginner and want to learn things the easy way, you can go for IDLE. Also, you can install python from here. 2. Pycharm Well, there are a lot of themes for these IDE’s and I have changed them as per my convenience. When you will install them they will come with a default orientation of light background. Pycharm is basically a professional IDE for developers which comes in two versions: one is a community version which is open-sourced and free to use. The other one is the professional version which can be downloaded for a 30-day free trial. You can install the Pycharm community version for windows from here. Pycharm sets good for its distinct distribution of project folders which help us in saving all the related files in the same folder locally. Also supports many interpreters for the same purpose. One can also install a lot of different and unique packages from its project interpreter which can completely revolutionize the way we install packages. I use Pycharm for computer vision and normal python based projects. And I find it really useful in the same because of the features that I have mentioned above. You can run your code by right-clicking and then going to run. 3. Spyder IDE Just like every other IDE, it comes with a light background theme, I have switched it to a darker because that is what developers do. For you to install Spyder IDE, you ought to have Anaconda Environment in your system. To install Anaconda Environment, click here. Anaconda basically is a data science toolkit which supports numerous platforms in it. A glance to its platform is: To start with a spyder you can either go to your search bar and type spyder or can launch it from the anaconda navigator as shown in the picture above. I use spyder for machine learning, computer vision and deep learning purposes. One benefit of using spyder is you are able to debug the code line by line all by yourself. To run a code segment in here you need to select that specific segment and press ctrl+enter. There are many other platforms that I use like Jupyter notebook, Google Colab and VS code but certain things don’t fall into the same category. It was really important for me to tell you the easy and the most popular ones that are used for the same. To get a tour of the IDE you can surely search related videos on Youtube which would surely help you out in giving a tour cum tutorial on the same. Conclusion The article was aimed to discuss the list of different IDEs we have for python when it comes to machine learning and understanding that a certain IDE suits the best for a certain purpose. It needs to be identified on your own. There are no impositions on choices; one can use the same IDE for all his work but it is always advised to use different for different purposes. In this article, we also saw that having hands-on different IDE is always better than sticking to only one because it gives you choices and you back up the work becomes easier in case the compiler of a certain IDE crashes, you know that you have another to back you up. Hope you liked the article.","excerpt":"This article discusses how to choose the right IDE for machine learning and elaborates whether there is any right IDE for it too or not.","categories":["Deep Tech"],"tags":["IDE","Machine Learning","Pycharm","python machine learning","spyder"],"author_name":"Bhavishya Pandit","publish_date":"2020-10-29T15:00:00","publication_year":"2020","word_count":1146,"keywords":["data science","machine learning","TPU","AI","spyder","Machine Learning","computer vision","Colab","Pycharm","Aim","deep learning","Jupyter","Python","python machine learning","IDE"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","data science","Aim","Jupyter","Colab","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/idle-vs-pycharm-vs-spyder-choosing-the-right-ide-for-machine-learning\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10104364,"title":"Top 10 Research Papers Published by Google","content":"The year 2023 has witnessed some groundbreaking research shaping the future of AI technology. Google, which has been at the forefront of the AI revolution, has announced AI models with multiple capabilities. Along with the launch of innovative products, it has also released various research papers, offering a glimpse into the underlying technology. Most recently, Google has released its latest generative AI multimodal model called Gemini, that competes directly with GPT-4, and is already in discussions on social media. But this is not the best paper that Google published this year. Here is the list of top 10 research papers published by Google in 2023. Gemini: A Family of Highly Capable Multimodal Models Topping the list is obviously Gemini, the paper behind the competitor multimodal model to OpenAI’s GPT-4.  Recently introduced, Gemini as a highly capable system jointly trained on image, audio, video, and text data. The primary goal is to create a model with robust generalist capabilities across modalities, coupled with state-of-the-art understanding and reasoning performance within each domain. Gemini 1.0, the inaugural version, is available in three sizes: Ultra for intricate tasks, Pro for scalable performance and deployment, and Nano for on-device applications. Each size is meticulously designed to cater to distinct computational limitations and application needs. Comprehensive evaluations of Gemini models encompass a diverse array of internal and external benchmarks, spanning language, coding, reasoning, and multimodal tasks. PaLM-2 PaLM-2 was the groundbreaking language model surpassing its predecessor, PaLM, boasting enhanced multilingual and reasoning capabilities while being more computationally efficient. Leveraging a Transformer-based architecture and a diverse set of training objectives, PaLM 2 demonstrates significantly improved performance on various downstream tasks, ensuring superior quality across different model sizes. Notably, PaLM 2 exhibits accelerated and resource-efficient inference, facilitating broader deployment and faster response times for more natural interactions. Its robust reasoning capabilities are highlighted by substantial advancements over PaLM in tasks such as BIG-Bench. PaLM-E: An Embodied Multimodal Language Model PaLM-E represents a significant leap forward in the development of AI agents capable of interacting with the physical world. This paper describes LLMs equipped with a virtual embodiment, allowing it to perceive and manipulate its surroundings through sensors and actuators. PaLM-E’s capabilities extend beyond simply understanding and generating text. It can navigate through a simulated environment, manipulate objects, and engage in simple conversations. This embodiment allows PaLM-E to learn and adapt to its environment in a more nuanced and realistic way compared to traditional LLMs. The potential applications of PaLM-E are vast and diverse. It could be used to develop more realistic and engaging virtual assistants, robots that can assist with tasks in the real world, and even educational tools that allow users to learn through interactive simulations. MusicLM: Generating Music from Text Google was also into making music this year. MusicLM revolutionises music creation by enabling the generation of high-quality music from simple text descriptions. This paper introduces a system capable of composing music in various styles and genres based on user input, opening up new possibilities for musicians, composers, and anyone interested in exploring musical creativity. MusicLM’s capabilities are based on a neural network trained on a massive dataset of music and text pairs. This allows the system to learn the complex relationships between text and musical elements, enabling it to generate music that is both faithful to the user’s description and musically sound. Structure and Content-Guided Video Synthesis with Diffusion Models This paper introduces a novel method for synthesising realistic videos using diffusion models. This approach allows for greater control over the content and structure of the generated videos, making it a valuable tool for video editing and animation. Traditional video synthesis methods often lacked the ability to accurately control the details and structure of the generated videos. Diffusion models address this limitation by providing a framework for gradually introducing noise into a video and then denoising it to achieve the desired result. This allows for fine-grained control over the entire video generation process. Lion: EvoLved Sign Momentum for Training Neural Networks Lion introduces a new and efficient optimisation algorithm for training neural networks. This algorithm significantly improves the speed and accuracy of training, leading to better performance for various AI applications. Traditional optimization algorithms used in training neural networks can be slow and inefficient. Lion addresses this issue by utilising a novel approach that analyses the dynamics of the training process and adapts accordingly. This allows Lion to optimise the learning process in a more effective way, leading to faster convergence and improved generalisation. InstructPix2Pix: Learning to Follow Image Editing Instructions This paper proposes a groundbreaking method for editing images based on text instructions. InstructPix2Pix enables users to modify images in a natural and intuitive way, opening up new possibilities for image editing and manipulation. Traditional image editing tools require users to have specific technical skills and knowledge. InstructPix2Pix removes this barrier by allowing users to edit images simply by providing textual instructions. This user-friendly approach makes image editing accessible to a wider audience and simplifies the process for experienced users. DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generation Large text-to-image models have limitations in mimicking subjects from a reference set and generating diverse renditions. To address this, Google Research and Boston University present a personalised approach. By fine-tuning the model with a few subject images, it learns to associate a unique identifier with the subject, enabling the synthesis of photorealistic images in different contexts. The technique preserves key features while exploring tasks like recontextualization, view synthesis, and artistic rendering. A new dataset and evaluation protocol are provided for a subject-driven generation. Check out their GitHub repository here. REVEAL: Retrieval-Augmented Visual-Language Pre-Training with Multi-Source Multimodal Knowledge Memory The paper presents REVEAL, an end-to-end Retrieval-Augmented Visual Language Model. REVEAL encodes world knowledge into a large-scale memory and retrieves from it to answer knowledge-intensive queries. It consists of a memory, encoder, retriever, and generator. The memory encodes various multimodal knowledge sources, and the retriever finds relevant entries. The generator combines retrieved knowledge with input queries to generate outputs. REVEAL achieves state-of-the-art performance in visual question answering and image captioning, utilising diverse multimodal knowledge sources. The paper is submitted by members from the University of California, Los Angeles and Google Research. On Distillation of Guided Diffusion Models Classifier-free guided diffusion models, widely used in image generation, suffer from computational inefficiency. Google, Stability AI and LMU Munich propose distilling these models into faster sampling models. The distilled model matches the output of combined conditional and unconditional models, achieving comparable image quality with fewer sampling steps. The approach is up to 256 times faster for pixel-space models and at least 10 times faster for latent-space models. It also proves effective in text-guided image editing and inpainting, requiring only 2-4 denoising steps for high-quality results.","excerpt":"Gemini is not the only good thing that Google has released this year.","categories":["AI Trends"],"tags":["AI Tool"],"author_name":"Mohit Pandey","publish_date":"2023-12-07T13:30:00","publication_year":"2023","word_count":1125,"keywords":["Go","TPU","OpenAI","AI","neural network","virtual assistants","RAG","Ray","generative AI","AI Tool","R"],"extracted_tech_keywords":["AI","neural network","generative AI","OpenAI","Ray","RAG","virtual assistants","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-research-papers-published-by-google-in-2023\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10124370,"title":"Genesys Launches India&#8217;s First AI-Powered Navigation Maps","content":"Mumbai-based mapping company Genesys International has announced the launch of India’s first AI-powered navigation maps tailored specifically for the automotive and mobility sectors. This innovative solution marks a pivotal moment in the evolution of the Indian automobile industry, introducing a new era of personalised driving experiences and sophisticated location intelligence. The newly unveiled map by Genesys encompasses the largest navigable road network in India, spanning an impressive 8.3 million kilometers and including over 30 million points of interest (POIs). This extensive coverage ensures that drivers across the country can access precise and dependable navigation, significantly enhancing their overall driving experience. Alongside the AI-powered navigation map, Genesys has introduced five products aimed at revolutionising the automotive and mobility industries. Navigation with Augmented Reality (AR) integrates real-time data from the vehicle’s cameras. AR provides intuitive overlays to guide drivers. Navigation with GPT AI Solution offers intelligent route planning. It provides real-time adjustments and voice assistance. Advanced Driver Assistance Systems (ADAS) ensures compliance with Euro NCAP safety standards. ADAS includes Intelligent Speed Assistance (ISA). An Online Marketplace is integrated into the car’s app. It allows for convenient in-vehicle purchases. Usage-Based Insurance (UBI) tracks driving behavior. UBI offers lower premiums for safer drivers. Sajid Malik, Chairman and Managing Director of Genesys International, expressed the company’s vision, stating, “With the launch of India’s first AI-powered navigation map by Genesys, we are set to completely revamp the way the Indian geospatial sector operates. Features like ISA and ADAS set new benchmarks for safety and convenience on Indian roads, alerting drivers to speed limits, recognising traffic signs, assisting with lane-keeping, and offering adaptive cruise control.” Malik further emphasised the importance of Genesys’ 3D Digital Twin technology, which allows the production of high-definition maps detailing every aspect of the road environment. This level of detail is crucial for the safe operation of autonomous vehicles, addressing challenges such as poor visibility, urban canyons, and GPS-deprived zones. Looking ahead, Genesys sees a future where their maps will play a critical role in the transition to autonomous and electric vehicles. The Indian automotive market, valued at $108.10 billion in 2022, is projected to reach $217.90 billion by 2031, growing at a compound annual growth rate of 8.1%. With its cutting-edge AI-powered navigation maps and innovative product offerings, Genesys International is well-positioned to capitalise on this growth and revolutionise the way people navigate and interact with their vehicles in India.","excerpt":"The newly unveiled map by Genesys encompasses the largest navigable road network in India, spanning 8.3 million kms and including over 30 million POIs.","categories":["AI News"],"tags":[],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-24T14:32:11","publication_year":"2024","word_count":398,"keywords":["API","programming_languages:R","AI","Git","RAG","GPT","Aim","edge AI","R","llm_models:GPT"],"extracted_tech_keywords":["AI","Aim","edge AI","RAG","R","Git","API","GPT","llm_models:GPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/genesys-launches-indias-first-ai-powered-navigation-maps\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10080768,"title":"10 Low-Code\/No-Code Platforms Every Developer Should Know of","content":"As per Gartner, a whopping 65% of application development projects will rely on low-code development by 2024. As the name suggests, the obvious benefit of these platforms is eliminating multiple technical skills usually required to develop software apps. Moreover, low-code\/no-code development is easy to access and has several similar benefits. Using these platforms results in continuously delivering clever business solutions without professional developers. Each low-code platform offers a unique set of benefits. Here are top 10 no-code\/low-code development platforms that you should check out. Appian https:\/\/www.youtube.com\/watch?v=Gf9ElaCcHOE Appian’s platform combines intelligent automation with low-code development to build smart applications for better efficiency and customer engagement. One of the pros of the platform is its flexible and easy-to-use nature. This free platform enables non-technical users to create anything from mobile apps to enterprise-wide systems without writing a single line of code. Try Appian now. Mendix Founded in 2005, today, more than 4,000 companies use Mendix. The platform relies on model-driven engineering (MDE), which emphasises abstract modelling. The Mendix Community version is free, and the platform offers three more paid plans. Try Mendix now. Nintex Nintex is highly rated, with 10,000+ organisations in diverse industries reporting desirable returns on investment. It’s a low-code platform for process management and workflow automation. Automation tools can be key to fostering growth in your business, and Nintex works to accelerate digital transformation within your organisation. Try Nintex now. Visual LANSA Visual LANSA allows to create applications much quicker than traditional coding and with much control, much higher than usually seen in low-code platforms. In addition, apps developed on the platform can be deployed to an IBM i, Windows, or Linux server. One IDE, one language, no limitations. Try Visual LANSA now. Quixy ​https:\/\/www.youtube.com\/watch?v=dDa9zqOKF9k The Quixy platform automates workflows to build enterprise-grade applications for their custom needs up to 10x faster. The cloud-based no-code citizen development platform is flexible and can develop applications within hours as per the users’ requirements. Easy deployment and visual builder are some of the platform’s great features. Try Quixy now. Airtable At first glance, Airtable looks much like a spreadsheet but is much more powerful. With a simple but colourful interface, Airtable creates and shares relational databases. Big companies, including Medium, BuzzFeed, and Nike, use Airtable to increase efficiency and cut costs. The platform’s pre-made templates, including social media planning and integration features, make it stand out in the domain. Try Airtable now. Caspio Caspio generates interactive reports and data visualisations that enable gathering performance metrics and making fact-based decisions. It also allows users to analyse data using drill-down and pivot functionality. The all-in-one platform includes an integrated cloud database, a visual application builder, enterprise-grade security, regulatory compliance and scalable global infrastructure. Caspio integrates with cloud-based systems such as Amazon S3, Box, Dropbox, Google Drive and Microsoft OneDrive. Try Caspio now. Kissflow ​https:\/\/www.youtube.com\/watch?v=9v_ob4MwukQ Created by OrangeScape Technologies, Kissflow automates corporate operations and tracks performance. In addition, Kissflow manages workflow and processes. Besides eliminating coding needs, Kissflow’s rule-based paradigm allows for personalising workflows. It is suitable for all types of businesses and industries. Kissflow assists in creating process requests, viewing items that require user action, and approving pending tasks. Try Kissflow now. Quickbase Quickbase integrates prominent cloud-based solutions like NetSuite, Salesforce, Box, and Gmail via prebuilt application connectors. In addition, to provide security, Quickbase includes features such as permissions and granular roles, corporate user management, and a robust user interface. Try Quickbase now. ZohoCreator A veteran in the industry, Zoho Creator is a minimalist platform with intuitive visual features. It includes a visual drag-and-drop interface and a wide range of built-in features and functions, making it a complete low-code development platform. Zoho Creator is a full-stack product for creating data-driven apps in the cloud. With a few clicks, users can select from a list of libraries of prebuilt blocks, drag and drop them onto the canvas, and customise them to create an app. Try Zoho Creator now.","excerpt":"Deliver clever solutions without professional developers or technical knowledge","categories":["AI Trends"],"tags":["no code platforms"],"author_name":"Tasmia Ansari","publish_date":"2022-11-25T14:00:00","publication_year":"2022","word_count":653,"keywords":["Go","intelligent automation","AI","digital transformation","Scala","Git","RAG","automation","GAN","no code platforms","R"],"extracted_tech_keywords":["AI","RAG","R","Go","Scala","Git","GAN","digital transformation","automation","intelligent automation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-low-code-no-code-platforms-every-developer-should-know-of\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070863,"title":"How to build a robust logistic regression model with L2 regularization?","content":"By fitting data to a logistic curve, logistic regression evaluates the connection between many independent factors and a categorical dependent variable and determines the likelihood of an event occurring. The loss function of logistic regression is a logistic loss which classifies based on the maximum likelihood estimation. This likelihood estimation tends to be biassed toward the higher value due to which regularization is required.  This article will focus on understanding the role of L2 regularization in logistic regression. Following are the topics to be covered. Table of contents Brief about the loss function of logistic regressionAbout L2 regularizationRole of L2 regularization in Logistic RegressionImplementing L2 regularization The penalty for failing to fulfil the planned production is referred to as a ‘loss.’ Let’s start with understanding the loss function of logistic regression. Brief about the loss function of logistic regression A loss function is a mathematical function that translates a theoretical declaration into a practical proposition. Developing a highly accurate predictor involves continual issue iteration via questioning, modelling the problem using the chosen approach, and testing. A logistic regression classifier predicts probabilities based on the weights in the training dataset, and the model will update its weights to minimise the difference between its predicted probabilities and the distribution of probabilities in the training data. This calculation is used for a binary prediction known as binary cross-entropy or log loss. Although this is erroneous, the term “cross-entropy” is occasionally used to refer to the negative log-likelihood of a Bernoulli or softmax distribution. When characterised by a negative log-likelihood, a loss may be defined as a cross-entropy between an empirical distribution produced from the training set and a probability distribution derived from the model. For example, mean squared error is the cross-entropy between an empirical distribution and a Gaussian model. Are you looking for a complete repository of Python libraries used in data science, check out here. About L2 regularization When training a machine learning model, it is easy for the model to become overfitted or under fitted. To circumvent this, regularisation is utilized in machine learning to fit a model to our test set effectively. Regularization techniques aid in reducing the likelihood of overfitting and obtaining an ideal model. Ridge regularization or L2 normalization is a penalty method which makes all the weight coefficients to be small but not zero. It is done by taking squares of the weights. This means that the mathematical function corresponding to our machine learning model is minimised, and coefficients are computed. Coefficient magnitudes are squared and summed. Ridge Regression accomplishes regularisation by reducing the number of coefficients. Here is the cost function. Analytics India Magazine Where, Loss = sum of squared residuals λ = penalty for the error W = slope of the curve Lambda represents the penalty term in the cost function. We regulate the punishment term by adjusting the values of the penalty function. The greater the penalty, the smaller the size of the coefficients. It reduces the parameters. As a result, it is utilised to prevent multicollinearity and to minimise model complexity through coefficient shrinking. Need for regularization in Logistic Regression Regularization is critical in logistic regression modelling. Without regularisation, logistic regression’s asymptotic nature would continue to drive loss towards 0 in large dimensions. As a result, to reduce model complexity, most logistic regression models include either L2 regularisation or early stopping (reducing the number of training steps or the learning rate). Consider assigning a unique id to each example and mapping each id to its own feature. If no regularisation function is specified, the model will become entirely overfit. Because the model would try and fail to drive loss to zero on all samples, pushing the weights for each indicator feature to +∞ or -∞. This can occur with high-dimensional data with feature crosses when there is a large number of unusual crosses that occur only on a single occurrence. Role of L2 regularization in Logistic Regression There is a high chance that the logistic regression overfits when dealing with polynomial data. When there is more than one independent variable it is known as a polynomial. Here is an example of this statement. Analytics India Magazine As in the above example, the decision boundary is too complex which indicates that the model is biassed towards the ‘x’ data points. On the right side of the image, a polynomial sigmoid function is mentioned for the logistic regression. So, to regularise the algorithm and make the decision boundary-less complicated need to use a penalty which will restrict the model from being biased. The penalty to be used for the logistic regression is Ridge regularization. The mathematical equation for when using the ridge penalty would be this: Analytics India Magazine This formula would be integrated with the gradient descent for more advanced optimization of the Regularized Logistic regression. Implementing L2 regularization This article uses sklearn logistic regression and the dataset used is related to medical science. The task is to predict the CDH based on the patient’s historical data using an L2 penalty on the Logistic Regression. Let’s import the necessary libraries import numpy as np import pandas as pd import matplotlib.pyplot as plt Reading the data and preparing for training by splitting the data into standard ratios of 30:70 for testing and training respectively. data=pd.read_csv('\/content\/drive\/MyDrive\/Datasets\/heart.csv') X=data.drop(['output'],axis=1) y=data['output'] X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.30, random_state=42) Build the regularized logistic regression regularized_lr=LogisticRegression(penalty='l2',solver='newton-cg',max_iter=200) regularized_lr.fit(X_train,y_train) reg_pred=regularized_lr.predict(X_test) For using the L2 regularization in the sklearn logistic regression model define the penalty hyperparameter. For this data need to use the ‘newton-cg’ solver because the data is less and any other method would not converge and a maximum iteration of 200 is enough. print('Precision score',np.round(precision_score(y_test,reg_pred),2)) print('Recall score',np.round(recall_score(y_test,reg_pred),2)) cm_reg = confusion_matrix(y_test, reg_pred, labels=regularized_lr.classes_) disp_reg = ConfusionMatrixDisplay(confusion_matrix=cm_reg,display_labels=regularized_lr.classes_) disp_reg.plot() plt.show() Analytics India Magazine Analytics India Magazine We are getting a precision score of 0.82 which is good in other situations but not in the medical case and similar for recall, a score of 0.84 is good but not in this case. The train data hasn’t been wrangled, and need to standardise the data because all the measurements of the parameters have different units, could also do some feature selection, engineering, etc. So to improve on the model level the primary focus is on the FALSE NEGATIVE  reduction which is currently at 8 because there is a high chance that due to this a patient could die. Well, leave that to you folks. Conclusions Regularization is critical in logistic regression modelling. Without regularisation, logistic regression’s asymptotic nature would continue to drive loss towards 0 in large dimensions. With this article, we have understood the implementation and concept of L2 regularization in Logistic Regression. References Link to the above codeDocumentation by sklearn for LR","excerpt":"L2 regularization is like leash on Logistic Regression to prevent overfitting","categories":["AI Trends"],"tags":["cross entropy","Dimensionality Reduction","logistic regression","overfitting","ridge regression","sigmoid"],"author_name":"Sourabh Mehta","publish_date":"2022-07-13T10:00:00","publication_year":"2022","word_count":1120,"keywords":["data science","NumPy","machine learning","TPU","AI","ridge regression","R","logistic regression","Python","analytics","Dimensionality Reduction","sigmoid","Matplotlib","overfitting","cross entropy","Pandas"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Pandas","NumPy","Matplotlib","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-build-a-robust-logistic-regression-model-with-l2-regularization\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10122435,"title":"Zendesk Expects $3 Bn in Revenue by 2027","content":"California-based Zendesk’s generative AI game to reshape customer service began more than a year ago. Zendesk AI, the company’s core CX product, offers various LLMs-driven customer service tools and ticketing systems. Brands across the globe use it to manage customer interactions, enhance call centre efficiency, help agents handle inquiries, prioritise tickets, and deliver a smooth customer support experience. More recently, the company launched autonomous AI agents, workflow automation, agent copilot, and AI-powered Workforce Engagement Management (WEM) and Quality Assurance (QA) tools in Zendesk AI. In a detailed interaction with AIM, Vasudeva Rao Munnaluri, RVP for India and SAARC at Zendesk, explained that with generative AI in its bucket and nearly two billion dollars in revenue attributable to the adoption of its Zendesk AI and expansion with customers, the SaaS giant plans to reach $3 billion dollars by 2027. The Role of Data “We process around eight billion requests daily and four million agents and admins leveraging our systems. This data fuels our AI missions,” said Munnaluri. Its proprietary data set, consisting of ticket data accumulated over years from thousands of customers, forms the backbone of its AI models. These models are designed to detect intent, sentiment, and emotions, enabling efficient ticket routing and knowledge base provision. “This data-driven approach helps automate workflows, enhancing both agent productivity and customer satisfaction,” added Munnaluri. The company aims to automate 50% of customer interactions through autonomous agents. Another important element of the company’s generative AI strategy is its partnership with AWS and Anthropic. The collaboration was announced last month at its flagship event Relate in San Francisco and Las Vegas. Through AWS, Zendesk leverages Amazon Bedrock to scale generative AI applications, incorporating Anthropic’s leading model Claude 3. “We are able to offer faster, more efficient, and accurate AI features, thanks to our partnerships with AWS and Anthropic,” Munnaluri said. Zendesk’s primary competitors include Freshworks, Zoho, and Oracle. Expanding Footprint in India With over 80 employees and a comprehensive functional representation, the India team continues to contribute significantly to the company’s global operations. It has over 160,000 customers in 160 countries. Some of its prominent customers in India are Cars24, Dream11, Plum, Unacademy and more. The company began operations in India in 2016, and the country has since become a crucial market within the APAC region. “India is a key market for Zendesk and one of the largest in the APAC region. We have seen continuous growth and plan to expand further,” Munnaluri emphasised. Although expansion in India is a priority due to the market’s dynamic nature, Munnaluri expressed that one of the major challenges in this country is the slow adoption of AI, with less than 15% of Indian businesses having an enterprise AI strategy as per the Zendesk CX Trends Report 2024. “Well, four out of five of my conversations in India are about AI, but I think, it’s still not caught into the overall strategy there. Over 50% still need adequate data at the view level. When you say data, it needs to be organised in a way that you have the taxonomy or the metadata available and ready to be understood by the AI models,” he added. Institutional inertia or resistance to change towards AI adoption further complicates the matter. Additionally, Indian customers are demanding more personalised and faster services, which requires solid investment in AI and data management. However, to make the generative AI products more diverse and accessible to a wider population of the country, it recently acquired Berlin-based AI startup Ultimate, which supports 109 languages globally, including over 13 Indian languages such as Marathi, Hindi, Telugu, Tamil, Malayalam, Kannada, Odia, and Bengali. This acquisition is in an effort to improve Zendesk’s ability to support Indian languages and break the barriers in customer service. [Update: 6th June 2024 13:32 | Previously, the headline mentioned that Zendesk Expects $3 Bn in revenue by 2027 with AI. However, the company requested that it be attributable to the adoption of its core AI CX product and expansion with tenured customers. The headline has now been updated to show that the company is not solely crediting this revenue projection to AI.]","excerpt":"The company aims to automate 50% of customer interactions through autonomous agents.","categories":["Global Tech"],"tags":["customer service","Generative AI","Oracle","SAP","Zendesk"],"author_name":"Shritama Saha","publish_date":"2024-06-04T15:36:50","publication_year":"2024","word_count":688,"keywords":["Anthropic","Zendesk","Go","autonomous agents","AI","AWS","SAP","Oracle","RAG","Aim","generative AI","customer service","GAN","Generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Anthropic","Aim","RAG","autonomous agents","AWS","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/zendesk-expects-3-bn-in-revenue-by-2027\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":39862,"title":"Brown University Researchers Train Robots To Write More Human-Like","content":"Source: Video by Kotani. Edited Picture. Automation is slowly replacing numerous manual jobs, but in the case of art, the role of artificial intelligence has only been that of an assistant to human creativity. New and improved AI-based tools are invigorating artists by providing a platform to visualise every idea that inspires them by allowing them to draw, erase and redraw. This effectively eliminates the process of filling trash cans with discarded ideas. Experiments like SketchRNN, Google’s experiment with AI, allows artists to draw something no matter how bleak the idea is. Essentially, the sketch is passed through recurrent neural networks which have been trained on millions of doodles collected from the Quick, Draw! game. Built on TensorFlow, Sketch-RNN comes up with many possible ways to continue drawing this object based on where the artist left. The model can also mimic drawings and produce similar doodles. Brown University Researchers Are Trying To Ace Perfection Though GANs have stunned the world with their surrealistic art pieces, AI is still a long way from replicating a human-like writing or drawing skills. Given the complexity of human hand movement not to forget other factors like the speed of each stroke and spacing, it is difficult for a robotic arm to do the job elegantly. To perform these functions efficiently, the research team of Atsunobu Kotani, an undergraduate student at Brown and Stefanie Tellex have devised two separate models: a local model to draw each stroke and a global model to learn the shifting action. The researchers demonstrate a robot that was able to write “hello” in 10 languages that employ different character sets. The robot was also able to reproduce rough sketches, including one of the Mona Lisa. Given an image of just drawn handwritten characters, the robot infers a plan to replicate the image. This method enables the machine to learn in real time and with more ease. Just by observing the target image, the robot arm immediately makes an attempt to draw\/write. The researchers behind this project segregate writing into two steps: A drawing action which draws each stroke, and, A shifting action to reach a new point to start How Does The Network Work To take action, the following attributes are considered: Already visited regions Current location Difference image Continuously connected target region The network combines the image formed by segregating a certain word into a tensor. Now, this combined image is encoded with a residual network to obtain a tensor of complete size. From this, a part is extracted say (5,5,64) from a tensor size of (100,100,64). This extracted tensor is flattened to form a vector of a certain length and is fed into an LSTM cell which outputs a new vector. The final prediction is made by mapping these vectors using two fully connected layers and this process is iteratively done until termination. This whole process encapsulates the local model. The termination of the local process will be followed by the construction of tensor from a new set of images. And, again a fully connected layer is applied and receive an image of say, size (100,100,1) which becomes the global model. Source: Paper by Atsunobu Kotani The algorithm makes use of deep learning networks as discussed above, which analyze images of handwritten words or sketches and can deduce the likely series of pen strokes that created them. The robot can then easily reproduce the words or sketches using the pen strokes it learned as can be seen below. Source: Paper by Atsunobu Kotani. The right side is replicated words. Key to making the system work, Kotani says, is that the algorithm uses two distinct models of the image it’s trying to reproduce. Using a global model that considers the image as a whole, the algorithm identifies a likely starting point for making the first stroke. Once that stroke has begun, the algorithm zooms in, looking at the image pixel by pixel to determine where that stroke should go and how long it should be. When it reaches the end of the stroke, the algorithm again calls the global model to determine where the next stroke should start, then it’s back to the zoomed-in model. This process is repeated until the image is complete. Robot replicating sketch of Monalisa Hard coding a robot to perform all the above skills even poorly, takes a lot of computational heavy lifting and some ingenious constraint assumption to make the robot perform decently, especially when put under unstructured, real-world situations. Asking a robot to run, do a cartwheel or throw a pitch would have sounded like a chapter from a generic sci-fi novel a few years ago. Now, with the advancement of hardware acceleration and the optimisation of machine learning algorithms, techniques like reinforcement learning are now being put into practical use. With this new ability of human-like see and draw simultaneously, machine vision has attained new heights. Know more about the work here.","excerpt":"Automation is slowly replacing numerous manual jobs, but in the case of art, the role of artificial intelligence has only been that of an assistant to human creativity. New and improved AI-based tools are invigorating artists by providing a platform to visualise every idea that inspires them by allowing them to draw, erase and redraw. […]","categories":["Deep Tech"],"tags":["Deep Learning","lstm","Robotics","Robots"],"author_name":"Ram Sagar","publish_date":"2019-05-29T11:24:10","publication_year":"2019","word_count":821,"keywords":["Go","artificial intelligence","machine learning","AWS","AI","neural network","TPU","lstm","Robotics","deep learning","Deep Learning","TensorFlow","R","Robots"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","TensorFlow","AWS","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/brown-university-researchers-train-robots-to-write-more-human-like\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10011862,"title":"Datasets for Language Modelling in NLP using TensorFlow and PyTorch","content":"In recent times, Language Modelling has gained momentum in the field of Natural Language Processing. So, it is essential for us to think of new models and strategies for quicker and better preparation of language models. Nonetheless, because of the complexity of language, we have to deal with some of the problems in the dataset. With an increase in the size of the dataset, there is an increase in the normal number of times a word shows up in that dataset. Models performing admirably on little datasets probably won’t perform well on bigger ones. Here, we will discuss some of the most popular datasets for word-level language modeling. Further, we will implement these datasets with the help of TensorFlow and Pytorch Library. Dataset Statistics In comparison to the Penn Treebank dataset, the WikiText datasets are larger. WikiText-2 aims to be of a similar size to the Penn Treebank while WikiText-103 contains all articles extracted from Wikipedia. WikiText-103 The WikiText-103 dataset, created by Salesforce, contains more than ten crore tokens retrieved from the arrangement of checked Good and Featured articles on Wikipedia. This dataset comprises 28,475 great and highlighted articles from Wikipedia. It has a drawn-out reliance with 103 million tokens. It contains a vocabulary size of 267,735 after replacing all the token that appears not more than two times with <unk>token. This makes it restrictive to explore different avenues regarding word-level LMs on this dataset. For an implanting size of 400, the embedding layer consists of 267K x 400 ≈ 106Mparameters. Loading the WikiText-103 dataset using Tensorflow Download the dataset from the link given below. !wget --quiet https:\/\/s3.amazonaws.com\/research.metamind.io\/wikitext\/wikitext-103-raw-v1.zip !unzip wikitext-103-raw-v1.zip Pass the path to the function. The TensorFlow library will help in reading the data and storing it in the proper format. import tensorflow as tf def wiki103(path): data2 = tf.data.TextLineDataset(path) def content_filter(source): return tf.logical_not(tf.strings.regex_full_match( source, '([[:space:]][=])+.+([[:space:]][=])+[[:space:]]*')) data2 = data2.filter(content_filter) data2 = data2.map(lambda x: tf.strings.split(x, ' . ')) data2 = data2.unbatch() return data2 train = wiki103('\/content\/wikitext-103-raw\/wiki.train.raw') Loading the WikiText-103 dataset using Pytorch from torchtext import data import io class LanguageModelingDataset(data.Dataset): def __init__(self, path, text_field, newline_eos=True, encoding='utf-8', **kwargs): fields = [('text', text_field)] text = [] with io.open(path, encoding=encoding) as f: for line in f: text += text_field.preprocess(line) if newline_eos: text.append(u'<eos>') examples = [data.Example.fromlist([text], fields)] super(LanguageModelingDataset, self).__init__( examples, fields, **kwargs) class WikiText103(LanguageModelingDataset): urls = ['https:\/\/s3.amazonaws.com\/research.metamind.io\/wikitext\/wikitext-103-v1.zip'] name2 = 'wikitext-103' directoryname = 'wikitext-103' def splits(cls, text_field, root='.data', train='wiki.train.tokens', validation='wiki.valid.tokens', test='wiki.test.tokens', **kwargs): return super(WikiText103, cls).splits( root=root, train=train, validation=validation, test=test, text_field=text_field, **kwargs) def iters(cls, batch_size=32, bptt_len=35, device=0, root='.data', vectors=None, **kwargs): text2 = data.Field() train, val, test = cls.splits(text2, root=root, **kwargs) text2.build_vocab(train, vectors=vectors) return data.BPTTIterator.splits( (train, val, test), batch_size=batch_size, bptt_len=bptt_len, device=device) root- directory where the dataset’s zip archive will be stored.batch size- number of training data points passed in one iteration.Bptt_len-length of sequence for backpropagation.device-Use – 1 for CPU and None for the presently dynamic GPU gadget.text_field – field that will be used for text data points.train – training setvalidation – approval settest -testing test Testing and Validation Perplexity State of the Art on WikiText-103 The present state of the art on WikiText-103 dataset is Megatron-LM. The model gave a test-perplexity of 10.81%. The model performs best with lower perplexity. WikiText-2 WikiText-2 is a 2M token variant of WikiText-103 with a jargon size of 33,278. This dataset is a little form of the WikiText-103 dataset. This little dataset is appropriate for testing your language model. Loading the WikiText-2 dataset using Tensorflow !wget --quiet https:\/\/s3.amazonaws.com\/research.metamind.io\/wikitext\/wikitext-2-raw-v1.zip !unzip wikitext-2-raw-v1.zip def wiki1(path): data = tf.data.TextLineDataset(path) def content_filter(source): return tf.logical_not(tf.strings.regex_full_match( source, '([[:space:]][=])+.+([[:space:]][=])+[[:space:]]*')) data = data.filter(content_filter) data = data.map(lambda x: tf.strings.split(x, ' . ')) data = data.unbatch() return data train= wiki1('\/content\/wikitext-2-raw\/wiki.train.raw') Loading the WikiText-2 dataset using Pytorch class WikiText2(LanguageModelingDataset): urls = ['https:\/\/s3.amazonaws.com\/research.metamind.io\/wikitext\/wikitext-2-v1.zip'] name1 = 'wikitext-2' directoryname = 'wikitext-2' def splits(cls, text_field, root='.data', train='wiki.train.tokens', validation='wiki.valid.tokens', test='wiki.test.tokens', **kwargs): return super(WikiText2, cls).splits( root=root, train=train, validation=validation, test=test, text_field=text_field, **kwargs) def iters(cls, batch_size=32, bptt_len=35, device=0, root='.data', vectors=None, **kwargs): text1 = data.Field() train, val, test = cls.splits(text1, root=root, **kwargs) text1.build_vocab(train, vectors=vectors) return data.BPTTIterator.splits( (train, val, test), batch_size=batch_size, bptt_len=bptt_len, device=device) Testing and Validation Perplexity State of the Art on WikiText-2 The present state of the art on WikiText-2 dataset is GPT-2 . The model gave a test-perplexity of 18.34%. PennTreeBank Penn Treebank dataset contains the Penn Treebank bit of the Wall Street Diary corpus, developed by Mikolov. It comprises 929k tokens for the train, 73k for approval, and 82k for the test. The words in the dataset are lower-cased, numbers substituted with N, and most punctuations eliminated. <unk> token replaced the Out-of-vocabulary (OOV) words. The jargon is the most continuous 10k words.It contains sentences rather than passages, so its setting is restricted. Loading the PennTreeBank dataset using Tensorflow import tensorflow as tf def pentree(path): data3 = tf.data.TextLineDataset(path) # Drop article headers. def content_filter(source): return tf.logical_not(tf.strings.regex_full_match( source, '([[:space:]][=])+.+([[:space:]][=])+[[:space:]]*')) data3 = data3.filter(content_filter) data3 = data3.map(lambda x: tf.strings.split(x, ' . ')) data3 = data3.unbatch() return data3 train = pentree('https:\/\/raw.githubusercontent.com\/wojzaremba\/lstm\/master\/data\/ptb.train.txt') Loading the Penn TreeBank dataset using Pytorch class PennTreebank(LanguageModelingDataset): urls = ['https:\/\/raw.githubusercontent.com\/wojzaremba\/lstm\/master\/data\/ptb.train.txt', 'https:\/\/raw.githubusercontent.com\/wojzaremba\/lstm\/master\/data\/ptb.valid.txt', 'https:\/\/raw.githubusercontent.com\/wojzaremba\/lstm\/master\/data\/ptb.test.txt'] name3 = 'penn-treebank' directoryname = 'penn-treebank' def splits(cls, text_field, root='.data', train='ptb.train.txt', validation='ptb.valid.txt', test='ptb.test.txt', **kwargs): return super(PennTreebank, cls).splits( root=root, train=train, validation=validation, test=test, text_field=text_field, **kwargs) def iters(cls, batch_size=32, bptt_len=35, device=0, root='.data', vectors=None, **kwargs): text3 = data.Field() train, val, test = cls.splits(text3, root=root, **kwargs) text3.build_vocab(train, vectors=vectors) return data.BPTTIterator.splits( (train, val, test), batch_size=batch_size, bptt_len=bptt_len, device=device) Testing perplexity of Penn TreeBank State of the Art on Penn TreeBank The present state of the art on PennTreeBank dataset is GPT-3. The model gave a test-perplexity of 20.5%. Conclusion In this article, we have covered most of the popular datasets for word-level language modelling. Penn Treebank is the smallest and WikiText-103 is the largest among these three. As the size of Penn TreeBank is less, it is easier and faster to train the model on this. So, it is advisable to check in detail the performance of models on different sizes of the dataset.","excerpt":"In recent times, Language Modelling has gained momentum in the field of Natural Language Processing. So, it is essential for us to think of new models and strategies for quicker and better preparation of language models. Nonetheless, because of the complexity of language, we have to deal with some of the problems in the dataset. With an increase in the size of the dataset, there is an increase in the normal number of times a word shows up in that dataset.","categories":["Deep Tech"],"tags":["extract big data","language modelling","mba in data analytics","Natural Language Processing","nlp in data","nlp pipeline","Pytorch","Tensorflow","what is big data coding"],"author_name":"Ankit Das","publish_date":"2020-11-19T15:00:45","publication_year":"2020","word_count":967,"keywords":["Git","extract big data","R","nlp pipeline","Pytorch","PyTorch","Natural Language Processing","language modelling","what is big data coding","Tensorflow","Go","mba in data analytics","AWS","AI","ML","TensorFlow","Aim","GitHub","nlp in data"],"extracted_tech_keywords":["AI","ML","Aim","TensorFlow","PyTorch","AWS","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/datasets-for-language-modelling-in-nlp-using-tensorflow-and-pytorch\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10165690,"title":"K’taka Budget 2025-26: ₹1,000 crore Accelerator Fund, Quantum Park and EV Cluster","content":"In a significant move to bolster Karnataka’s image as the tech capital, the state government, on Friday, announced the Budget 2025-26, allocating a significant amount for AI initiatives. The state allocated ₹300 crore for a Fund-of-Funds and a corpus fund of ₹100 crore for deep tech development, while also setting aside ₹50 crore over five years for the Centre for Applied AI for Tech Solutions (CATS). The government pushed the startup ecosystem to expand further in cities like Mysuru, Mangaluru, Hubballi-Dharwad, and Kalburgi, with 10,000 job creations. Presenting the estimated ₹4,09,549 crore budget, Karnataka Chief Minister Siddaramaiah announced development initiatives, positioning the state as a leader in technological innovation. Additionally, under the ‘Brand Bengaluru’ concept (to improve infrastructure, service delivery, and traffic management), ₹1,800 crore was allotted for 21 schemes during the financial year 2025-26. AI in Governance and Education The Karnataka AI cell is developing AI-based solutions using technologies like computer vision and Natural Language Processing (NLP) and LLMs to enhance governance. “An AI-based Government Order Summary and Information Extraction Tool will facilitate easy access to government information for officials and citizens, the CM said.  Additionally, the AI-driven IPGRS 2.0 was announced to improve public grievance resolution, placing Karnataka ahead in the digital administration game. Focusing on the education sector, the government initiated the Kalika Deepa Programme in collaboration with Ek-Step Foundation, extending AI-based learning tools to 2,000 schools. The aim is to enhance students’ English and Kannada language skills and mathematical competencies. Karnataka also launched an advanced attendance management system in all government departments using the latest technologies, such as AI and Geographic Information System (GIS). In addition, a ₹2 crore smart system is to be implemented using AI to provide transcriptions of court proceedings, translations of judicial documents, and other information. As part of the ₹667 crore Bengaluru Safe City project, cameras have been installed across Bengaluru, out of which 10 drones have been implemented.​ To monitor and control vehicular traffic, AI-enabled electronic cameras will be installed at 60 places in Davanagere, Dharwad, Kalaburagi, Belagavi, Chitradurga, Haveri, Hospet, Bellari, Vijayapura, and Dakshina Kannada districts at the cost of ₹50 crore. GCCs, IT and New Tech Get a Monetary Boost Aiming to stimulate innovation and entrepreneurship in cities other than Bengaluru, a Local Economy Accelerator Program (LEAP) will be launched in the current year with a grant of Rs. 1,000 crore. The state’s IT policy targets investments of $300 billion and job creation of 1.5 lakh, emphasising cloud computing and emerging technologies. Karnataka will develop Global Innovation Districts in Mysuru, Belagavi, Dharwad, and Bengaluru to strengthen its leadership in GCCs as well. The GCC Policy 2020-25 also aims for an investment of ₹5,000 crore, with an ambition of creating 3.5 lakh jobs. Additionally, Karnataka State Electronics Development Corporation Limited (KEONICS) will establish three new Global Technology Centres in Mangaluru, Hubli, and Belagavi as plug-and-play facilities. The budget announcement also included Quantum Research Park Phase-2, to be set up with a grant of ₹48 crore over three years in collaboration with the Indian Institute of Science, Bengaluru. Semiconductors and Electronics The budget announced the building of a state-of-the-art PCB Park will be developed in Mysuru on 150 acres, creating a dedicated ecosystem for electronics manufacturing, with an addition of ₹99 crore SensorTech Innovation Hub to encourage sensor research and development. In the broader context, Karnataka’s semiconductor ambitions are gaining momentum, with companies like Microsoft, Intel, Accenture, and IBM committing significant investments. The state’s Industrial Policy 2025-30, unveiled at Invest Karnataka 2025, aims to create 2 million jobs and attract substantial investments in sectors like semiconductors and renewable energy. The Union budget for 2025–26 also supports these efforts with an increased allocation for semiconductors and electronics manufacturing. To encourage electronic vehicle (EV) manufacturing and usage, a testing track of international standards and a state-of-the-art EV cluster with common infrastructure will be established around Bengaluru at a cost of ₹25 crore. The Karnataka Biotechnology Policy also aims to attract ₹1,500 crore investments through incentives, while the data centre policy aims for an investment of ₹10,000 crore with a total capacity of 200MW. Additionally, Karnataka’s AVGC-XR Policy 2024-29 also sees an investment of ₹150 crore with 30,000 jobs to be created. The state signed MoUs totalling ₹10.27 lakh crore at the Global Investors’ Meet 2025, held in February this year, expecting to create over 600,000 job opportunities. The government envisions positioning itself as a hub for Artificial Intelligence (AI), AI-driven governance, education, electronic manufacturing, deep tech, Global Capability Centres (GCCs), biotechnology, and IT.","excerpt":"Policy boost for Startup to expand in cities like Mysuru, Mangaluru, Hubballi-Dharwad, and Kalburgi, with 10,000 job creations.","categories":["Deep Tech"],"tags":["budget","karnataka"],"author_name":"Sanjana Gupta","publish_date":"2025-03-07T21:26:33","publication_year":"2025","word_count":751,"keywords":["karnataka","Go","budget","artificial intelligence","AI","cloud computing","Git","computer vision","RAG","NLP","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","computer vision","Aim","RAG","cloud computing","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ktaka-budget-2025-26-%e2%82%b91000-crore-accelerator-fund-quantum-park-and-ev-cluster\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10005610,"title":"Beginners Guide to Pyjanitor &#8211; A Python Tool for Data Cleaning","content":"As a data scientist, you are more or less going to spend 60-70% of your time cleaning and preparing your data. The process of cleaning, encoding and transforming your raw data in order to bring them into a format that the machine learning model can understand is called Data Pre-processing. This process is often long and cumbersome and most developers consider it to be the least favourite part of a project. Despite being tedious, it is one of the most important techniques that need to be implemented. To simplify the overall process and make it a bit more interesting, python introduces a package called PyJanitor- A Python Tool for Data Cleaning. This article deals with an overview of what pyjanitor is, how it works and a demonstration of using this package to clean dirty data. What is pyjanitor? Initially developed in R as a Janitor library, it was developed in Python due to its convenience. Pyjanitor is an API that is written on top of the popular python library Pandas. Data pre-processing can be thought of as a directed acyclic graph where the starting node is raw data and we implement a series of techniques on this raw data to get usable data. Pandas has been a huge part of the data science ecosystem and pyjanitor API is implemented on pandas using a concept called method chaining. Method chaining can be understood as something similar to parallel processing. Instead of having an imperative style of programming as with pandas, method chaining combines multiple processes and allows the user to decide the order of the actions taken. Here is an example of method chaining. If we were using Pandas to clean our data it would look something like this data = pd.DataFrame(‘your dataset’) data = data.dropna(subset=['columnA', 'columnB']) data = data.rename({'columnA': 'apple', 'columnB': ‘banana’}) data['new_column'] = ['iterable', 'of', 'items'] data.reset_index(inplace=True, drop=True) With pyjanitor, the functions are just verbs that help you perform the actions. data = ( pd.DataFrame(‘your dataset’) .remove_columns(['columnC']) .dropna(subset=['columnA', 'columnB'']) .rename_column('columnA', ‘apples’) .rename_column('columnB', ‘banana’) .add_column('new_column', ['iterable', 'of', 'items']) .reset_index(drop=True) ) Using these functions, we pipeline them together. This method chaining helps in writing cleaner code and the function names are easier to remember, making the data cleaning much simpler. There are two advantages to using pyjanitor. One, it extends pandas with convenient data cleaning routines. Two, it provides a cleaner, method-chaining, verb-based API for common pandas routines. How does it work? The functions are written in the pandas-flavour package and are registered. These functions are registered without making any changes to pandas. You can add your own data processing functions depending on the data and use these methods whenever needed. Here is an example. import pandas as pd import pandas_flavor as pf @pf.register_dataframe_method def remove_column(dataframe, column_name: str): del df[column_name] return df Once you import the janitor library, the above function is automatically registered and can be used. The fact that each DataFrame method pyjanitor registers return the DataFrame is what gives it the capability to method chain. Demonstration of data cleaning with pyjanitor In this demonstration, we will implement various data frame manipulation techniques on raw data. Before we start with cleaning, let us install the requirements. pip install pyjanitor I have used Melbourne housing dataset that can be found here. Once you have installed the package and downloaded the data we can get started with data pre-processing. Here is what the raw data looks like. As you can see there are missing values, whitespaces in column names and categorical columns. Using pyjanitor I will now pipeline the most obvious data processing into just one block of code. cleaned_df = ( pd.read_csv('\/content\/gdrive\/MyDrive\/melbourne\/Melbourne_housing_FULL.csv') .clean_names() .remove_empty() .rename_column('price', \"target\") .encode_categorical([\"method\", \"suburb\", \"regionname\"]) .fill_empty('buildingarea',value=1) .drop('date',axis=1) ) I have listed the methods to clean the column names. This removes the white spaces, capital letters and special characters from column names making it easy to access them. Then, a function to remove empty rows if any. To make it easy for identifying the features and target I renamed the price column to target. Next, I converted three columns namely ‘method’,’ suburb’ and ‘regionname’ to categorical values. The fill_empty method is used to fill your columns that contain NaN with any value of your choice. Finally, I decided to drop the Date column just for the purposes of this demonstration. Once you have pipelined and finished the first and obvious parts of processing your data, it becomes easier for you to identify the more subtle changes needed and where the changes need to be made. I will now show you how to register your custom function to the janitor. Let’s say you want to remove spaces in a column of your dataset. I have selected a column named council area. This is how the column looks like. To remove the space, I will write my custom function as follows and register it to the janitor library. import pandas_flavor as pf @pf.register_dataframe_method def str_remove(df, column_name: str, pat: str, *args, **kwargs) df[column_name] = df[column_name].str.replace(pat, \"\", *args, **kwargs) return df To register it, simply import janitor again and you can use it on your dataset. import janitor cleaned_df=( .str_remove(column_name=\"councilarea\",pat=' ')) ) Just like that, we wrote a custom function to remove white spaces and this can be used for any dataset by anybody. There are 1000s of methods you can choose from for making the process of transforming the data easier. Here is a list of various methods that can be used. Conclusion With packages like pyjanitor, the amount of time and effort spent to clean and transform the data significantly reduces. Not only does it provide options for pipelining multiple methods but the option to create and register your own data cleaning techniques which can come in handy to other users as well. This is a new yet growing library in the field of data science and is helping data scientists perform cumbersome tasks in a much more efficient way.","excerpt":"This article deals with an overview of what pyjanitor is, how it works and a demonstration of using this package to clean dirty data.","categories":["Deep Tech"],"tags":["Data Cleaning","pandas","Python","python tools"],"author_name":"Bhoomika Madhukar","publish_date":"2020-08-26T17:00:46","publication_year":"2020","word_count":981,"keywords":["data science","Go","API","machine learning","Banana","AI","Python","pandas","python tools","Data Cleaning","programming_languages:Python","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","data science","Pandas","Python","R","Go","API","Banana","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/beginners-guide-to-pyjanitor-a-python-tool-for-data-cleaning\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":53203,"title":"10 Free Resources To Learn GAN In 2020","content":"Generative Adversarial Networks or GAN, one of the interesting advents of the decade, has been used to create arts, fake images, and swapping faces in videos, among others. GANs are the subclass of deep generative models which aim to learn a target distribution in an unsupervised manner. The resources we listed below will help a beginner to kick-start learning and understanding how this model works. In this article, we list down 10 free resources to learn GAN in 2020. Note: The list is in alphabetical order 1| Are GANs Created Equal? A Large-Scale Study Resource: Paper About: ‘Are GANs Created Equal? A Large-Scale Study’ is a paper written by the researchers at Google Brain. In this paper, you will learn about the subclass of generative models called Generative adversarial networks (GAN). The researchers conducted a neutral, multi-faceted large-scale empirical study on state-of-the-art models and evaluation measures. Further, the researchers proposed several datasets to evaluate precision and recall. Click here to read. 2| A Large-Scale Study on Regularization and Normalization in GANs Resource: Paper About: To understand this paper, you will need to have a basic understanding of what GAN is. The researchers explained the class of deep generative models from a practical perspective. Further, they discussed and evaluated the common pitfalls and reproducibility issues while working on GAN. Click here to read. 3| Deep Diving into GANs: From Theory To Production Resource: Blog About: Deep Diving into GANs is a guide where you will learn the very basics of what a GAN is, understand the use of TensorFlow, APIs, and Google cloud functions. You will learn topics like conditional GAN, its applications, generators, discriminators, non-saturating value function, gradient ascent, writing a GAN using AshPy and TensorFlow Datasets, along with much more. This tutorial requires packages like Python 3.7 or more, TensorFlow 2.0 or more, Jupyter and NumPy. Click here to read. 4| GAN by Ian Goodfellow Resource: Video About: This is a NIPS 2016 video tutorial where Ian Goodfellow explained the basics of Generative adversarial networks (GANs). The topics in this video include the review of work applying GANs to large image generation, extending the GAN framework to approximate maximum likelihood rather than minimizing the Jensen-Shannon divergence, semi-supervised learning with GANs, and other such. Click here to watch. 5| Generative Models By OpenAI Resource: Blog About: In this blog by OpenAI, you will learn about generative models and how to train them. It covers topics like training a generative model, process of generating images, examples of generative model approaches and how it works. Click here to read. 6| GANs In Action Resource: Book (Free Preview) About: GANs in Action is one of the popular books of GAN written by Jakub Langr and Vladimir Bok. In this book, you will learn how to build and train your own Generative Adversarial Networks (GANs), understand generators and discriminators, how to build your own simple adversarial system and much more. Click here to read. 7| Generative Adversarial Networks Resource: Paper About: This paper introduces General Adversarial Network (GAN). Here, you will learn the generative model and discriminative model and how they work. Ian Goodfellow and his team discussed adversarial nets, advantages, and disadvantages of GAN and other such. Click here to read. 8| Generative Adversarial Networks: An Overview Resource: Paper About: In this paper, you will get an overview of GANs for the signal processing community as well as drawing on familiar analogies and other concepts. You will understand how to identify different methods for training and constructing GANs, and their challenges. Click here to read. 9| Introduction to GAN By Google Resource: Blog About: This tutorial is provided by the researchers at Google. Here, you will learn the basics of GAN and how to use the TF-GAN library to create GANs. You will understand the difference between generative and discriminative models, understand the advantages and disadvantages of common GAN loss functions, and much more. Click here to read. 10| Lecture Notes On Generative Learning Algorithms By Andrew NG Resource: PDF About: This lecture notes by Andrew NG from Stanford University will help you understand topics like Gaussian discriminant analysis, understanding GDA model and logistic regression, Laplace smoothing, Naive Bayes, event models for text classification, and other such related topics. Click here to read.","excerpt":"Generative Adversarial Networks or GAN, one of the interesting advents of the decade, has been used to create arts, fake images, and swapping faces in videos, among others. GANs are the subclass of deep generative models which aim to learn a target distribution in an unsupervised manner. The resources we listed below will help a […]","categories":["AI Trends"],"tags":["free datasets for analysis","naive bayes","tensorflow gradient"],"author_name":"Ambika Choudhury","publish_date":"2020-01-06T14:00:00","publication_year":"2020","word_count":710,"keywords":["Go","NumPy","text classification","free datasets for analysis","OpenAI","AI","naive bayes","Python","Aim","Jupyter","tensorflow gradient","TensorFlow","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","TensorFlow","Jupyter","NumPy","text classification","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-free-resources-to-learn-gan-in-2020\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":41402,"title":"5 Popular Machine Learning Libraries Built On TypeScript For 2019","content":"Introduced in 2012, TypeScript is said to be the typed superset of JavaScript which helps in overcoming the limitations of JavaScript in large scale applications. The language has been gaining popularity because of its advantages and features, for instance, support for classes and modules, ES6 features, type-checking, and many more. In this article, we list down 5 machine learning libraries which are written in this language. (The list is in no particular order) 1| Machinelearn.js Machinelearn.js is a machine learning library which is written in Typescript. The library resolves the issues of the complexities in machine learning through web technologies. This library is similar to Scikit Learn and it features both supervised and unsupervised models which include Random Forest, PCA, KMeans, Decision Tree, Naive Bayes, etc. Installation You can set up Node.js environment and the installation can be done using either npm or yarn For npm:  npm install –save machinelearn For yarn: yarn add machinelearn Click here to install. 2| TensorFlow Deep Playground TensorFlow Deep Playground is an interactive visualisation of neural networks which is written in TypeScript using d3.js. The main goal of this framework is to make neural network more accessible as well as easier to learn. To run the visualisation locally, you have to run “npm i” to install dependencies “npm run build” to compile the app and place it in the dist\/ directory “npm run serve” to serve from the dist\/ directory and open a page on your browser. Click here to install. 3| R.js R.js is a machine learning library which includes packages of R language re-written in typescript for browsers. In this library, almost all the important components of R language have been ported into TypeScript language and it currently includes a total of 8 repositories. Binary Linear Algebra Subprograms (BLAS) is a linear algebra specification numerical library has been written completely into TypeScript. In fact, many numerical software applications use BLAS computations, including Armadillo, LAPACK, LINPACK, GNU Octave, Mathematica, MATLAB, NumPy, R, and Julia. Click here to install. 4| Machine-learning The machine-learning library is written in TypeScript which has a dependency on the nblas package for creating fast matrix operations. This library is in an early development phase and was last committed in 2017. It works by default on OSX and in Windows you may need  to install LAPACK, while in Linux you have to run: apt-get install libblas-dev. Installation The library can be installed via npm, type npm i machine-learning Click here to install. 5| ML Classifier UI ML Classifier is a machine learning engine written in TypeScript. This library can be used for training image classification models in your browser while consuming a shorter period of time. ML Classifier is a React front end for a machine learning engine for training the machine learning models. Installation The library can be installed via npm or yarn: For npm: npm install ml-classifier-ui For yarn: yarn add ml-classifier-ui. Click here to install.","excerpt":"Introduced in 2012, TypeScript is said to be the typed superset of JavaScript which helps in overcoming the limitations of JavaScript in large scale applications. The language has been gaining popularity because of its advantages and features, for instance, support for classes and modules, ES6 features, type-checking, and many more. In this article, we list […]","categories":[],"tags":["ml libraries","TypeScript"],"author_name":"Ambika Choudhury","publish_date":"2019-06-27T06:43:07","publication_year":"2019","word_count":488,"keywords":["Go","NumPy","machine learning","AI","neural network","ML","TypeScript","ml libraries","JavaScript","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","TensorFlow","NumPy","R","JavaScript","TypeScript","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/5-popular-machine-learning-libraries-built-on-typescript-for-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69229,"title":"Hands-On Guide to Implement Deep Autoencoder in PyTorch for Image Reconstruction","content":"Artificial Neural Networks have many popular variants that are applied in supervised and unsupervised learning problems. The Autoeconders are also a variant of neural networks that are mostly applied in unsupervised learning problems. When they come with multiple hidden layers in the architecture, they are referred to as the Deep Autoencoders. These models can be applied in a variety of applications including image reconstruction. In image reconstruction, they learn the representation of the input image pattern and reconstruct the new images matching to the original input image pattern. Image reconstruction has many important applications especially in the medical field where the decoded and noise-free images are required from the available incomplete or noisy images. In this article, we will demonstrate the implementation of a Deep Autoencoder in PyTorch for reconstructing images. This deep learning model will be trained on the MNIST handwritten digits and it will reconstruct the digit images after learning the representation of the input images. Autoencoder Autoencoders are the variants of Artificial Neural Networks which are generally used to learn the efficient data codings in an unsupervised manner. They usually learn in a representation learning scheme where they learn the encoding for a set of data. The network reconstructs the input data in a much similar way by learning its representation. The basic architecture of am Autoencoder is shown below. (Image Source: Wikipedia) The architecture generally comprises an input layer, an output layer and one or more hidden layers that connect input and output layers. The output layer has the same number of nodes as of input layers because of the purpose that it reconstructs the inputs. In its general form, there is only one hidden layer, but in case of deep autoencoders, there are multiple hidden layers. This increased depth reduces the computational cost of representing some functions and it decreases the amount of training data required to learn some functions. The popular applications of autoencoder include anomaly detection, image processing, information retrieval, drug discovery etc. Implementing Deep Autoencoder in PyTorch First of all, we will import all the required libraries. import os import torch import torchvision import torch.nn as nn import torchvision.transforms as transforms import torch.optim as optim import matplotlib.pyplot as plt import torch.nn.functional as F from torchvision import datasets from torch.utils.data import DataLoader from torchvision.utils import save_image from PIL import Image Now, we will define the values for the hyperparameters. Epochs = 100 Lr_Rate = 1e-3 Batch_Size = 128 The below function will be used for image transformation that is required for the PyTorch model. transform = transforms.Compose([ transforms.ToTensor(), transforms.Normalize((0.5,), (0.5,)) ]) Using the below code snippet, we will download the MNIST handwritten digit dataset and get it ready for further processing. train_set = datasets.MNIST(root='.\/data', train=True, download=True, transform=transform) test_set = datasets.MNIST(root='.\/data', train=False, download=True, transform=transform) train_loader = DataLoader(train_set, Batch_Size=Batch_Size, shuffle=True) test_loader = DataLoader(test_set, Batch_Size=Batch_Size, shuffle=True) Let us see some information on the training data and its classes. print(train_set) print(train_set.classes) In the next step, we will define the Autoencoder class that will be used to define the main model. class Autoencoder(nn.Module): def __init__(self): super(Autoencoder, self).__init__() #Encoder self.enc1 = nn.Linear(in_features=784, out_features=256) # Input image (28*28 = 784) self.enc2 = nn.Linear(in_features=256, out_features=128) self.enc3 = nn.Linear(in_features=128, out_features=64) self.enc4 = nn.Linear(in_features=64, out_features=32) self.enc5 = nn.Linear(in_features=32, out_features=16) #Decoder self.dec1 = nn.Linear(in_features=16, out_features=32) self.dec2 = nn.Linear(in_features=32, out_features=64) self.dec3 = nn.Linear(in_features=64, out_features=128) self.dec4 = nn.Linear(in_features=128, out_features=256) self.dec5 = nn.Linear(in_features=256, out_features=784) # Output image (28*28 = 784) def forward(self, x): x = F.relu(self.enc1(x)) x = F.relu(self.enc2(x)) x = F.relu(self.enc3(x)) x = F.relu(self.enc4(x)) x = F.relu(self.enc5(x)) x = F.relu(self.dec1(x)) x = F.relu(self.dec2(x)) x = F.relu(self.dec3(x)) x = F.relu(self.dec4(x)) x = F.relu(self.dec5(x)) return x Now, we will create the Autoencoder model as an object of the Autoencoder class that we have defined above. model = Autoencoder() print(model) Now, the loss criteria and the optimization methods will be defined. criterion = nn.MSELoss() optimizer = optim.Adam(net.parameters(), lr=Lr_Rate) The below function will enable the CUDA environment. def get_device(): if torch.cuda.is_available(): device = 'cuda:0' else: device = 'cpu' return device The below function will create a directory to save the results. def make_dir(): image_dir = 'MNIST_Out_Images' if not os.path.exists(image_dir): os.makedirs(image_dir) Using the below function, we will save the reconstructed images as generated by the model. def save_decod_img(img, epoch): img = img.view(img.size(0), 1, 28, 28) save_image(img, '.\/MNIST_Out_Images\/Autoencoder_image{}.png'.format(epoch)) The below function will be called to train the model. def training(model, train_loader, Epochs): train_loss = [] for epoch in range(Epochs): running_loss = 0.0 for data in train_loader: img, _ = data img = img.to(device) img = img.view(img.size(0), -1) optimizer.zero_grad() outputs = model(img) loss = criterion(outputs, img) loss.backward() optimizer.step() running_loss += loss.item() loss = running_loss \/ len(train_loader) train_loss.append(loss) print('Epoch {} of {}, Train Loss: {:.3f}'.format( epoch+1, Epochs, loss)) if epoch % 5 == 0: save_decod_img(outputs.cpu().data, epoch) return train_loss The below function will test the trained model on image reconstruction. def test_image_reconstruct(model, test_loader): for batch in test_loader: img, _ = batch img = img.to(device) img = img.view(img.size(0), -1) outputs = model(img) outputs = outputs.view(outputs.size(0), 1, 28, 28).cpu().data save_image(outputs, 'MNIST_reconstruction.png') break Before training, the model will be pushed to the CUDA environment and the directory will be created to save the result images using the functions defined above. device = get_device() model.to(device) make_dir() Now, the training of the model will be performed. train_loss = training(model, train_loader, Epochs) After successful training, we will visualize the loss during training. plt.figure() plt.plot(train_loss) plt.title('Train Loss') plt.xlabel('Epochs') plt.ylabel('Loss') plt.savefig('deep_ae_mnist_loss.png') We will visualize several images that are saved during training. Image.open('\/content\/MNIST_Out_Images\/Autoencoder_image0.png') Image.open('\/content\/MNIST_Out_Images\/Autoencoder_image50.png') Image.open('\/content\/MNIST_Out_Images\/Autoencoder_image95.png') In the last step, we will test our autoencoder model to reconstruct the images. test_image_reconstruct(model, testloader) Image.open('\/content\/MNIST_reconstruction.png') So, as we could see that the autoencoder model started reconstructing the images since the start of the training process. After the first epoch, this reconstruction was not proper and was improved until the 50th epochs. After the complete training, as we can see in the image generated after the 95th epoch and on testing, it can construct the images very well matching to the original input images. Further, it opens a scope to train the model for more number of epochs as 100 or 200 because we have seen a heavy loss during training that was getting decreased epoch by epoch. After a long training, it is expected to obtain more clear reconstructed images. However, we could understand using this demonstration how to implement deep autoencoders in PyTorch for image reconstruction. References:- Sovit Ranjan Rath, “Implementing Deep Autoencoder in PyTorch” Abien Fred Agarap, “Implementing an Autoencoder in PyTorch” Reyhane Askari, “Auto Encoders”","excerpt":"In this article, we will demonstrate the implementation of a Deep Autoencoder in PyTorch for reconstructing images. This deep learning model will be trained on the MNIST handwritten digits and it will reconstruct the digit images after learning the representation of the input images.","categories":["Deep Tech"],"tags":["autoencoders","Deep Learning","image reconstruction","Pytorch"],"author_name":"Dr. Vaibhav Kumar","publish_date":"2020-07-08T17:00:00","publication_year":"2020","word_count":1071,"keywords":["CUDA","Pytorch","autoencoders","TPU","AI","image reconstruction","neural network","PyTorch","deep learning","anomaly detection","Deep Learning","Matplotlib","R"],"extracted_tech_keywords":["AI","deep learning","neural network","PyTorch","Matplotlib","anomaly detection","TPU","CUDA","R","CUDA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-to-implement-deep-autoencoder-in-pytorch-for-image-reconstruction\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":56955,"title":"Can Someone Become A Data Scientist In 6 Months","content":"Data science is a field which has a considerable expansion, and it is a pretty tough job to build a career as a data scientist. It takes years of preparation along with certain degrees and educational background to understand what data science really is and how it works. The aim becomes more onerous for those who are looking to switch to data science from an entirely different background. In the domain of data science, there are many individuals who have successfully added the designation of a data scientist to their resumes. But there are some who aspires to do so in a very short span of six months which may not be possible. Ambition must be highly inclined, but it also has to meet realism. In this article, we will share with you a number of reasons why it is not possible to become a data scientist in just six months irrespective of being highly determined and putting massive efforts. Understanding The Depth The first step before beginning the preparation is to self evaluate the current scenario where the candidate should be clear about where he or she stands. One should start with rudimentary elements that are essential to becoming a data scientist and these go back to school days. To understand data science, it is very crucial to have mathematics as a subject in high school with a sound knowledge of chapters like calculus and algebra since these are the very foundations on which data science is built and functions. Although it is not absolutely essential to have these subjects in school but to become a data science, one needs to know these subjects which usually takes a semester or a period of four to five months to learn completely. Not to mention, learning alone is not enough since a lot of practice is a must to have these subjects on your fingertips. Learning The Core Like mentioned before, the data science universe is vast, and there is so much to learn. The learning in data science never comes to an end, but it starts with statistics which are divided into two parts, such as Inferential Statistics and Descriptive Statistics. An aspirant who is aiming to become a data scientist should know the A to Z of statistics and to do so, it is imperative to have a degree which teaches statistics along with a lot of other mathematics over a course of time, not in six months. Moving on, the next elements that must be learned by an aspirant are machine learning, deep learning, time series forecasting along with programming languages such as Python, R and SQL. For example, machine learning consists of several disciplines such as neuroscience, vision, speech etc. which takes more than a year to learn. Similarly, deep learning takes up more than six months, which also requires certain programming skills along with the ability to learn Python at the same time. Altogether, the amount of learning that is required to become a data scientist cannot be done in a mere time period of six months. How To Learn The Core Considering the number of online training courses and institutes coming up every day, finding the best course or institute is a challenge in itself. Since six months is a concise period, it is advisable to go for a full-time course. Although, someone with a job in hand can dare to go for the online courses. An aspirant must be able to dedicate more than 8 hours a day in order to learn data science and even after doing that, one might fall short. To have real experience in regards to projects, an aspirant can take a look at several projects that are offered by the online courses which one can implement by taking parts in different competitions to get the first-hand experience. A lot of free materials are available on the web in the form of YouTube videos, questions and answers on community sites like Quora and StackOverflow. Also, doing a wide range of projects on Kaggle will demonstrate your capabilities in data science techniques. Going back to the beginning of the paragraph, it is essential to figure out the differences between a part-time course and full-time course. Our article “How to Choose Between a Full-Time Data and Part-Time Data Science Courses” could help you in figuring out the right choice. From Where To Learn The Core Now that we have discussed the elements and how crucial a full-time or part-time course can play, let’s have a look at where one should head to learn data science. These days, there’s a new institute or training institute in every nook and corner. Definitely, one should avoid joining these institutes and find the credible one. The credibility of an institute is known by the teaching faculty and experience in the field. One should always conduct a thorough research about how and by whom the institute is being run. They should connect with aspirants from previous batches to understand how the institute functions.  Direct communication with an earlier student is the best way to find out the pros and cons of an institute. Furthermore, they should ask previous students about the possibility of learning data science in six months. Just in case an aspirant wants to enrol and take the challenge, we would like to ease the woes by sharing our articles “Top 10 Data Science Training Institutes in India” and “Top 10 Data Science Training Courses in India” which might help an aspirant to zero in on a particular course or institute, to begin with. The Last Step In a scenario where an aspirant might have been able to pull off a miracle by learning data science in six months, by this stage he or she is probably looking forward to their first job as a data scientist. Having knowledge about data science alone will never be enough to get a job, especially the first one when a relevant experience is missing. Hence, preparing well for the interview is very crucial. One should make a list of several questions that might be asked by the recruiter related to data science. The chances are high that an aspirant will have to give several interviews before getting an offer. It is wise to work on the questions that an aspirant might not have been able to answer in an interview, which led to the declination of the offer. In this way, the aspirant will be able to narrow down the chances of failing and understand the questions and skills which a recruiter wants from a data scientist.The road to learning data science is a hard one and irrespective of putting enough hard work, the time period is too short for anyone to know the ins and outs of data science. For any aspirant who is looking forward to ways of cracking an interview, we would suggest having a look at our article “Interview Strategies to Land a Data Science Job in 2020” for a clear understanding of how the interview process works in the field of data science.","excerpt":"Data science is a field which has a considerable expansion, and it is a pretty tough job to build a career as a data scientist. It takes years of preparation along with certain degrees and educational background to understand what data science really is and how it works. The aim becomes more onerous for those […]","categories":["AI Highlights"],"tags":["career in data science","Data Scientist","how to become a data scientist","Learn Data Science"],"author_name":"Rohit Chatterjee","publish_date":"2020-02-19T16:33:36","publication_year":"2020","word_count":1180,"keywords":["data science","Go","how to become a data scientist","machine learning","AI","RAG","Python","Aim","deep learning","SQL","Learn Data Science","Data Scientist","R","career in data science"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","Aim","RAG","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/can-someone-become-a-data-scientist-in-6-months\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":46936,"title":"25 Years Of QR Codes: A Look At Vulnerabilities Of This Popular Payment Method","content":"The turn of this millennium witnessed a heavy smartphone penetration in society. This also led to the development of electronic payments and digital wallets. Payment modes like scanning QR codes have taken the central stage. Today the majority of transactions that happen in retail shops or other bill payments are using this system. QR codes were invented in 1994 by Denso Wave, a unit of Japan’s largest automotive parts maker, to allow for quick scanning when tracking vehicles during the assembly process. It was initially designed for an automobile factory, which later found applications in other industries. Quick Response or QR codes are two-dimensional barcodes that visually encode bits of information represented as black square dots placed on a white square grid. Source: MIT Currently, over 23% of Trojans and viruses are transmitted via QR codes. On the 25th anniversary of QR codes, its creator, Masahiro Hara wants to make QR scanning more secure. Usually, in the case of QR scanning, possible scenarios of attacks can be summarised as follows: QR codes cannot be hacked. One way hackers to infiltrate this system by changing the QR code added in the poster. These fake posters can be circulated in the public domains and clueless customers scan these fake QR codes and end up visiting phishing websites. This usually happens because of the increase in the number of mobile users. Mobiles make it hard to verify the full link in the address bar. This makes users more vulnerable. When they use this phishing page to login, their passwords are compromised. An attacker might set up a fake website and redirect users by changing the QR Code. This is dangerous if some form of credentials are needed to access the website. The user has no possibility to verify that the link is not modified. SQL injection is another form of attack that occurs when SQL queries are made with user input text inserted into the query string. QR code readers are subject to data injection into their structured objects when they attempt to interpret the data of a QR code. A malicious party can create a QR code that injects arbitrary strings into a user’s data structures potentially causing harm to the user. Criminals can simply prepare malicious QR codes and affix them over legitimate codes which may result in victims inadvertently making payments to a criminal rather a legitimate service provider. QRLJacking or Quick Response Code Login Jacking is a simple social engineering attack vector capable of session hijacking affecting all applications that rely on “Login with QR code” feature as a secure way to login to accounts. QRLJacking attack gives attackers the ability to apply a full account hijacking scenario on the vulnerable Login with QR Code feature resulting in accounts stealing and reputation affection. Source: DSCI QR codes are capable of storing high quality data and its significance can be found in IoT applications as well. As more devices get connected, the more prone they are to attacks and QR codes can be one such blind spot for attacks if it is left untouched. Going Forward In a study done by Data Security Council Of India along with Paypal, the following tips were listed for the customers: Install a mobile security application with antivirus, antispyware and web filtering abilities to protect your mobile devices. If the QR code looks like it was added on to marketing materials, do not scan it. If the QR code leads you to a website that request for your personal information, do not disclose anything until you have verified that the request is legitimate. Do not scan QR codes in the form of stickers placed randomly in public places as it might be from scammers testing out his\/her malicious QR code. The concern of security of payment modes is more significant to countries like India, which has seen a sporadic rise in initiatives such as Bharat QR and UPI QR interoperable. The pursuit of large scale digitisation also exposes the systems to attacks such as discussed in the previous sections of the article. So far India hasn’t witnessed any significant malpractice using QR codes. However, as India aims at reaching new heights in its digital payments journey, consumer trust and safety becomes critical and should be considered as a high priority.","excerpt":"The turn of this millennium witnessed a heavy smartphone penetration in society. This also led to the development of electronic payments and digital wallets. Payment modes like scanning QR codes have taken the central stage. Today the majority of transactions that happen in retail shops or other bill payments are using this system. QR codes […]","categories":["AI Trends"],"tags":["Cybersecurity","Phishing","UPI"],"author_name":"Ram Sagar","publish_date":"2019-10-07T16:00:24","publication_year":"2019","word_count":716,"keywords":["Go","programming_languages:R","AI","Phishing","ML","Git","programming_languages:SQL","UPI","Aim","SQL","Rust","Cybersecurity","R"],"extracted_tech_keywords":["AI","ML","Aim","R","SQL","Go","Rust","Git","programming_languages:R","programming_languages:SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/25-years-of-qr-codes-a-look-at-vulnerabilities-of-this-popular-payment-method\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10012461,"title":"Most Popular Datasets For Neural Textual Entailment With Implementation In PyTorch And Tensorflow","content":"Textual entailment is a technique in natural language processing that endeavors to perceive whether one sentence can be inferred from another sentence. A pair of sentences are categorized into one of three categories: positive or negative or neutral. The positive category happens when the main sentence is used to demonstrate that a subsequent sentence is valid. Negative entailment or contradiction occurs when the primary sentence can be utilized to invalidate the subsequent sentence. Finally, if the two sentences have no relationship, they are considered to have a neutral entailment. Textual entailment is valuable in some of the applications. For example, it is used in question-answering systems to verify an answer from stored data. It may also be used to remove sentences that don’t have new information. The article will give a detailed explanation of the various popular datasets that are used in Textual entailment using TensorFlow and Pytorch. SNLI SNLI contains 570000 human-produced sentence sets in the English language that are physically marked for a balanced classification. It was developed in 2015 by the researchers: Samuel R.Bowman, Gabor Angeli and Christoper Potts of Stanford University. The data is collected by using the Amazon Mechanical Turk. It consists of three categories: entailment, contradiction and neutral. For Example: Source Loading the dataset usingTensorFlow Load the libraries required for this project. import csv import os import tensorflow.compat.v2 as tf import tensorflow_datasets.public_api as tfds url = 'https:\/\/nlp.stanford.edu\/projects\/snli\/snli_1.0.zip' class Snli(tfds.core.GeneratorBasedBuilder): VERSION = tfds.core.Version('1.1.0') def _info(self): return tfds.core.DatasetInfo( builder=self, features=tfds.features.FeaturesDict({ 'premise': tfds.features.Text(), 'hypothesis': tfds.features.Text(), 'label': tfds.features.ClassLabel( names=['entailment', 'neutral', 'contradiction']), }), supervised_keys=None, homepage='https:\/\/nlp.stanford.edu\/projects\/snli\/', ) def _split_generators(self, dl_manager): dl_directory = dl_manager.download_and_extract(url) data_directory = os.path.join(dl_directory, 'snli_1.0') return [ tfds.core.SplitGenerator( name=tfds.Split.TRAIN, gen_kwargs={ 'filepath': os.path.join(data_directory, 'snli_1.0_train.txt') }), ] Loading the dataset using Pytorch Pass the link to the function. The next step is to split the datasets into train,test and validation sets. from torchtext import data class ShiftReduceField(data.Field): def __init__(self): super(ShiftReduceField, self).__init__(preprocessing=lambda parse: [ 'reduce' if t == ')' else 'shift' for t in parse if t != '(']) self.build_vocab([['reduce'], ['shift']]) class ParsedTextField(data.Field): def __init__(self, eos_token='<pad>', lower=False, reverse=False): if reverse: super(ParsedTextField, self).__init__( eos_token=eos_token, lower=lower, preprocessing=lambda parse: [t for t in parse if t not in ('(', ')')], postprocessing=lambda parse, _: [list(reversed(p)) for p in parse], include_lengths=True) else: super(ParsedTextField, self).__init__( eos_token=eos_token, lower=lower, preprocessing=lambda parse: [t for t in parse if t not in ('(', ')')], include_lengths=True) class NLIDataset(data.TabularDataset): urls = [] directoryname = '' namenew = 'nli' @staticmethod def sort_key(ex): return data.interleave_keys( len(ex.premise), len(ex.hypothesis)) @classmethod def splits(cls, text_field, label_field, parse_field=None, extra_fields={}, root='.data', train='train.jsonl', validation='val.jsonl', test='test.jsonl'): path = cls.download(root) if parse_field is None: fields = {'sentence1': ('premise', text_field), 'sentence2': ('hypothesis', text_field), 'gold_label': ('label', label_field)} else: fields = {'sentence1_binary_parse': [('premise', text_field), ('premise_transitions', parse_field)], 'sentence2_binary_parse': [('hypothesis', text_field), ('hypothesis_transitions', parse_field)], 'gold_label': ('label', label_field)} for key in extra_fields: if key not in fields.keys(): fields[key] = extra_fields[key] return super(NLIDataset, cls).splits( path, root, train, validation, test, format='json', fields=fields, filter_pred=lambda ex: ex.label != '-') @classmethod def iters(cls, batch_size=32, device=0, root='.data', vectors=None, trees=False, **kwargs): if trees: TEXT = ParsedTextField() TRANSITIONS = ShiftReduceField() else: TEXT = data.Field(tokenize='spacy') TRANSITIONS = None LABEL = data.Field(sequential=False) train, val, test = cls.splits( TEXT, LABEL, TRANSITIONS, root=root, **kwargs) TEXT.build_vocab(train, vectors=vectors) LABEL.build_vocab(train) return data.BucketIterator.splits( (train, val, test), batch_size=batch_size, device=device) class SNLI(NLIDataset): urls = ['http:\/\/nlp.stanford.edu\/projects\/snli\/snli_1.0.zip'] directoryname = 'snli_1.0' name1 = 'snli' @classmethod def splits(cls, text_field, label_field, parse_field=None, root='.data', train='snli_1.0_train.jsonl', validation='snli_1.0_dev.jsonl', test='snli_1.0_test.jsonl'): return super(SNLI, cls).splits(text_field, label_field, parse_field=parse_field, root=root, train=train, validation=validation, test=test) Parameters Specification text_field: It will be used for premise and hypothesis data. Label_field: It will be used for label data. parse_field: It will be used for shift-reduce parser transitions. root: Path to the dataset’s zip archive directory. train: Training Set. Default: ‘train.jsonl’. test:Testing Set validation:Validation Set State of the Art The current state of the art on SNLI dataset is CA-MTL. The model gave an accuracy of 92.1%. Multi-NLI Multi-Genre Natural Language Inference (MultiNLI) corpus contains 433000 sentence sets explained with literary entailment data. It was developed by the researchers: Adina Williams, Nikita Nangia and Samuel R. Bowman1 . The corpus is demonstrated on the SNLI corpus, however, varies in that it covers a scope of kinds of spoken and composed content, and supports a particular cross-class speculation assessment. Below is the example of corpus: Source Loading the dataset using TensorFlow import os import tensorflow.compat.v2 as tf import tensorflow_datasets.public_api as tfds class MultiNLI(tfds.core.GeneratorBasedBuilder) VERSION = tfds.core.Version(\"1.1.0\") def _info(self): return tfds.core.DatasetInfo( builder=self, features=tfds.features.FeaturesDict({ \"premise\": tfds.features.Text(), \"hypothesis\": tfds.features.Text(), \"label\": tfds.features.ClassLabel( names=[\"entailment\", \"neutral\", \"contradiction\"]), }), supervised_keys=None, homepage=\"https:\/\/www.nyu.edu\/projects\/bowman\/multinli\/\", ) def _split_generators(self, dl_manager): downloaded_dir = dl_manager.download_and_extract( \"https:\/\/cims.nyu.edu\/~sbowman\/multinli\/multinli_1.0.zip\") multinli_path = os.path.join(downloaded_dir, \"multinli_1.0\") train_path = os.path.join(multinli_path, \"multinli_1.0_train.txt\") matched_validation_path = os.path.join(multinli_path, \"multinli_1.0_dev_matched.txt\") mismatched_validation_path = os.path.join( multinli_path, \"multinli_1.0_dev_mismatched.txt\") return [ tfds.core.SplitGenerator( name=tfds.Split.TRAIN, gen_kwargs={\"filepath\": train_path}), ] Loading the dataset using Pytorch class MultiNLI(NLIDataset): urls = ['http:\/\/www.nyu.edu\/projects\/bowman\/multinli\/multinli_1.0.zip'] directoryname = 'multinli_1.0' name2 = 'multinli' @classmethod def splits(cls, text_field, label_field, parse_field=None, genre_field=None, root='.data', train='multinli_1.0_train.jsonl', validation='multinli_1.0_dev_matched.jsonl', test='multinli_1.0_dev_mismatched.jsonl'): extra_fields = {} if genre_field is not None: extra_fields[\"genre\"] = (\"genre\", genre_field) return super(MultiNLI, cls).splits(text_field, label_field, parse_field=parse_field, extra_fields=extra_fields, root=root, train=train, validation=validation, test=test) State of the Art The current state of the art on SNLI dataset is T5-11B. The model gave an accuracy of 92%. XNLI The Cross-lingual Natural Language Inference (XNLI) corpus contains 5,000 test and 2,500 training sets for the Multi-NLI corpus. It was developed by the researchers: Adina Williams and Samuel R. BowmanThe sets are explained with printed entailment and converted into 14 dialects: French, Spanish, German, Greek, Bulgarian, Russian, Turkish, Arabic, Vietnamese, Thai, Chinese, Hindi, Swahili and Urdu. For example: Source Loading the dataset using TensorFlow import collections import csv import os import six import tensorflow.compat.v2 as tf import tensorflow_datasets.public_api as tfds url = 'https:\/\/cims.nyu.edu\/~sbowman\/xnli\/XNLI-1.0.zip' languages = ('ar', 'bg', 'de', 'el', 'en', 'es', 'fr', 'hi', 'ru', 'sw', 'th', 'tr', 'ur', 'vi', 'zh') class Xnli(tfds.core.GeneratorBasedBuilder): VERSION = tfds.core.Version('1.1.0') def _info(self): return tfds.core.DatasetInfo( builder=self, features=tfds.features.FeaturesDict({ 'premise': tfds.features.Translation( languages=languages,), 'hypothesis': tfds.features.TranslationVariableLanguages( languages=languages,), 'label': tfds.features.ClassLabel( names=['entailment', 'neutral', 'contradiction']), }), supervised_keys=None, homepage='https:\/\/www.nyu.edu\/projects\/bowman\/xnli\/', ) def _split_generators(self, dl_manager): dl_directory = dl_manager.download_and_extract(url) data_directory = os.path.join(dl_directory, 'XNLI-1.0') return [ tfds.core.SplitGenerator( name=tfds.Split.TEST, gen_kwargs={'filepath': os.path.join(data_directory, 'xnli.test.tsv')}), ] Loading the dataset using Pytorch class XNLI(NLIDataset): urls = ['http:\/\/www.nyu.edu\/projects\/bowman\/xnli\/XNLI-1.0.zip'] directoryname = 'XNLI-1.0' name3 = 'xnli' @classmethod def splits(cls, text_field, label_field, genre_field=None, language_field=None, root='.data', validation='xnli.dev.jsonl', test='xnli.test.jsonl'): extra_fields = {} if genre_field is not None: extra_fields[\"genre\"] = (\"genre\", genre_field) if language_field is not None: extra_fields[\"language\"] = (\"language\", language_field) return super(XNLI, cls).splits(text_field, label_field, extra_fields=extra_fields, root=root, train=None, validation=validation, test=test) @classmethod def iters(cls, *args, **kwargs): raise NotImplementedError('XNLI dataset does not support iters') State of the Art The current state of the art on SNLI dataset is RoBERTa-wwm-ext-large. The model gave an accuracy of 81.2%. Conclusion In this article, we have discussed some of the most popular datasets that are used in Textual Entailment. Further, we implemented these text corpus using Pytorch and TensorFlow.Textual Entailment are incredible vehicles for thinking, and essentially all inquiries regarding weightiness in language can be decreased to can be reduced to questions of entailment and contradiction in context. This recommends that Text Entailment is an ideal proving ground for hypotheses of semantic portrayal.","excerpt":"Textual entailment is a technique in natural language processing that endeavors to perceive whether one sentence can be inferred from another sentence. A pair of sentences are categorized into one of three categories: positive or negative or neutral.","categories":["AI Trends"],"tags":["Datasets","Natural Language Processing"],"author_name":"Ankit Das","publish_date":"2020-11-26T18:00:59","publication_year":"2020","word_count":1136,"keywords":["Go","Datasets","API","AI","PyTorch","Natural Language Processing","BERT","NLP","Ray","spaCy","TensorFlow","R"],"extracted_tech_keywords":["AI","NLP","Ray","TensorFlow","PyTorch","spaCy","R","Go","API","BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/most-popular-datasets-for-neural-textual-entailment-with-implementation-in-pytorch-and-tensorflow\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10089551,"title":"Researchers Making Diffusion Models Safe Again","content":"Diffusion models have replaced GANs to become the de facto way for image generation over the past year. However, there were some deeply-embedded unsuitable concepts such as nudity, violence and misattributed artwork that impeded its widespread use—until now. Researchers have found a way to remove these concepts from the ‘memory’ of diffusion algorithms. Using the power of the model itself to modify the weights of the neural network, researchers can make models forget concepts, artistic styles, or even objects. This could be the break diffusion models need to break into mainstream adoption. How does it work? Titled ‘Erasing Concepts from Diffusion Models’, the paper describes a method of fine-tuning that can edit model weights to remove certain concepts from memory. The method claims to do this while having minimal interference with other concepts in the latent space of the model. This approach can be useful in a variety of ways, with its primary goal being to make diffusion models more safe, private, and adherent to copyright laws. In the paper, the researchers showed a sample of this method where they were able to remove the concept of ‘nudity’ from a model’s weights. This has the obvious benefit of making diffusion models safe for certain types of use-cases, such as education, without having to build in stringent filters and moderation endpoints after the model is done training. Instead of spending large amounts of time and compute to re-train a huge diffusion model, erasing concepts through finetuning can create a safe model with a very low amount of compute. In a twist of irony, researchers used the models’ own encyclopaedic knowledge against them. By freezing the weights of the pre-trained model and using it to predict noise, the researchers were able to guide the model in the opposite direction of a given forbidden prompt. Using classifier-free guidance, the researchers were able to move the model away from the given concept. When this process is conducted iteratively, the model moves away from the prompt until the output cannot contain the prompt. Put into an analogy, this concept is similar to helping a child balance a bicycle. Whenever the model moves towards the forbidden concept, which correlates to falling down in this example, the person guiding the cycle (researchers) pushes it back to its centre point. As the child (the model) slowly begins to learn to balance, the concept of losing balance (the forbidden prompt) is slowly erased. As with any method, this also has a few limitations. In some cases where a certain concept was present across a large part of the dataset, erasing said concept would result in interference with unrelated concepts. However, this can bring diffusion models down to a more palatable form of generative AI by tackling some limitations. Can it fix diffusion? In the past, researchers have explored this concept by building out an entirely different model. Known as Safe Latent Diffusion, this model is a fork of Stable Diffusion’s 1.5 version and is trained to exclude images that ‘might be offensive, insulting, threatening, or might otherwise cause anxiety’. While this method was successful, the new method promises to be faster and more practical as it doesn’t require the whole mode lot be retrained. Selectively erasing certain concepts from the algorithm’s ‘memory’ has a lot of applications apart from just removing nudity. Removing these so-called forbidden concepts is a huge step forward for diffusion models. Due to the black box nature of neural networks, it was not possible to safely remove such concepts from models before without extensive data pre-processing to remove any bad data. With this method, state-of-the-art algorithms which derive a lot of value from their large datasets can be tweaked to remove the concepts without degrading their performance or efficacy. This approach can also tackle another common drawback of diffusion models—attribution. One of the biggest problems that artists have with diffusion models is that they ‘steal’ art and ‘remix’ it without any way of determining who the credit should be given to. With this new concept, it is possible to remove a certain artist or art style from the memory of the model completely. Take this diagram, for instance. The researchers were able to remove the artistic styles of Vincent Van Gogh, Edward Munch, Johannes Vermeer, and Hokusai from the memory of the model. This resulted in an output that resembled the original image, albeit without the signature style of the artist who made it. This tackles one of the biggest problems associated with diffusion-based models. When combined with other projects such as Stable Attribution, this approach might pave the way for diffusion models that don’t steal artwork.","excerpt":"Some deeply-embedded concepts in diffusion models have made them unsuitable for widespread use—until now.","categories":["AI Highlights"],"tags":["MidJourney","stable diffusion AI","stable diffusion download"],"author_name":"Anirudh VK","publish_date":"2023-03-17T17:00:00","publication_year":"2023","word_count":772,"keywords":["Go","MidJourney","stable diffusion download","TPU","AWS","AI","neural network","R","diffusion models","Aim","stable diffusion","generative AI","stable diffusion AI"],"extracted_tech_keywords":["AI","neural network","generative AI","Aim","AWS","TPU","R","Go","stable diffusion","diffusion models"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/researchers-making-diffusion-models-safe-again\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10022981,"title":"Guide to Perceiver: A Scalable Transformer-based Model","content":"Real-World Data comes in several modalities such as audio, text, video and images. Understanding different kinds of data and extracting patterns in it requires algorithms and models specific to the modality of the data. For example, CNNs are preferred to use on image data whereas attention-based models preferred for text data. But biological systems do not use disparate models to process data of diverse modalities. They process data from different modalities simultaneously. Inspired by this biological perception approach Perceiver, a transformer-based model was introduced by Andrew Jaegle, Felix Gimeno,  Andrew Brock, Andrew Zisserman, Oriol Vinyals and Joao Carreira on 4th March 2021. In this post let’s explore the Perceiver Model. Architecture Perceiver is a transformer-based model that uses both cross attention and self-attention layers to generate representations of multimodal data. A latent array is used to extract information from the input byte array using top-down or feedback processing i.e. attention to a byte array is controlled by the latent array which already has information about the byte array from the previous layer. Complexity The attention layers in the perceiver model are the most memory and time-intensive blocks. Vanilla Self Attention has quadratic complexity i.e if we have m inputs in the byte array it would take O(M2) memory to get attention values. The perceiver model solves this by using a low dimensional latent array for calculating attentions. The complexity of Self Attentions in the latent transformer blocks will reduce to O(N2) where N is the size of the latent array. Cross Attention layer complexity reduces to O(M X N). Here N is much smaller compared to M. This is especially helpful for data modalities with high bandwidth. Iterative Attention We can see that the byte array is used multiple times in the architecture. Each time an attention layer attends to the byte array with queries from the latent array. The output will have drawn some information from the byte array and this output is again used(after passing through a Latent Transformer block) to query the byte array again. This process can be repeated a number of times by making the network very deep. This results in the model selecting the right information from the byte array. This process can be optimized by sharing the parameters across transformer blocks. This makes it very similar to an RNN model. Positional Embeddings Transformer Models are permutation invariant i.e they ignore the sequence information in the data. While this may be good for generalization across modalities, this is bad when it comes to data where sequence information is crucial ex: Images, Text etc. Just like in many other transformer architectures this problem is solved by injected positional embeddings in the inputs. Fourier features are used as positional encodings in this model. This allows flexibility in the number of dimensions as well as the length of the dimensions of data. Frequencies are log uniformly sampled from n bands of frequencies. n is predetermined by us. This allows us greater control over the encodings while retaining the flexibility of the embeddings. The values of Fourier transform at these frequencies gives us the embeddings which are then concatenated with the inputs. Perceiver model performed on par with models with assumptions about the structure of the data. It got SOTA results using Image data, raw audio, video, audio + video and point clouds in 3D space. Let’s see how we can train a Perceiver model and run inference on it.. Code Installation A PyTorch implementation of the model was made available by authors. We can directly install this implementation from pip using the following command. pip install perceiver-pytorch Data Loading Let’s use the CIFAR10 image dataset for training the model. Following is the boilerplate code for loading the data. The code mentioned below, is referenced to official pytorch tutorial. import torch import torchvision import torchvision.transforms as transforms transform = transforms.Compose( [transforms.ToTensor(), transforms.Normalize((0.5, 0.5, 0.5), (0.5, 0.5, 0.5))]) trainset = torchvision.datasets.CIFAR10(root='.\/data', train=True, download=True, transform=transform) trainloader = torch.utils.data.DataLoader(trainset, batch_size=4, shuffle=True, num_workers=2) testset = torchvision.datasets.CIFAR10(root='.\/data', train=False, download=True, transform=transform) testloader = torch.utils.data.DataLoader(testset, batch_size=4, shuffle=False, num_workers=2) classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck') Model Model architecture can be loaded using. from perceiver_pytorch import Perceiver model = Perceiver( input_channels = 3,          # number of channels for each token of the input input_axis = 2,              # number of axis for input data (2 for images, 3 for video) num_freq_bands = 6,          # number of freq bands, with original value (2 * K + 1) max_freq = 10.,              # maximum frequency, hyperparameter depending on how fine the data is depth = 6,                   # depth of net num_latents = 32,           # number of latents, or induced set points, or centroids. different papers giving it different names cross_dim = 128,             # cross attention dimension latent_dim = 128,            # latent dimension cross_heads = 1,             # number of heads for cross attention. paper said 1 latent_heads = 2,            # number of heads for latent self attention, 8 cross_dim_head = 8, latent_dim_head = 8, num_classes = 10,          # output number of classes attn_dropout = 0., ff_dropout = 0., weight_tie_layers = False    # whether to weight tie layers (optional, as indicated in the diagram) ) Training Loop import torch.optim as optim criterion = torch.nn.CrossEntropyLoss() optimizer = optim.SGD(model.parameters(), lr=0.001, momentum=0.9) dev=torch.device(\"cuda:0\") model.to(dev) for epoch in range(2):  # loop over the dataset multiple times running_loss = 0.0 for i, data in enumerate(trainloader, 0): # get the inputs; data is a list of [inputs, labels] inputs, labels = data inputs, labels = inputs.to(dev), labels.to(dev) # zero the parameter gradients optimizer.zero_grad() # forward + backward + optimize outputs = model(inputs.permute(0,2,3,1)) loss = criterion(outputs, labels) loss.backward() optimizer.step() # print statistics running_loss += loss.item() if i % 1000 == 0:    # print every 1000 mini-batches print('[%d, %5d] loss: %.3f' % (epoch + 1, i + 1, running_loss \/ 1000)) running_loss = 0.0 print('Finished Training') Inference def predict(inputs): preds=model(inputs) preds=preds.detach().cpu().numpy() labels=[classes[np.argmax(i)] for i in preds] return labels predict(inputs.permute(0,2,3,1).to(dev)) References Perceiver PaperGithub RepositoryColab Notebook","excerpt":"Perceiver is a transformer-based model that uses both cross attention and self-attention layers to generate representations of multimodal data. A latent array is used to extract information from the input byte array using top-down or feedback processing","categories":["AI Trends"],"tags":["attention mechanism","Generative Pre-Trained Transformer","Guide","scalability"],"author_name":"Pavan Kandru","publish_date":"2021-03-27T18:00:00","publication_year":"2021","word_count":989,"keywords":["CUDA","Go","NumPy","TPU","AI","PyTorch","R","ML","Colab","Ray","Generative Pre-Trained Transformer","scalability","attention mechanism","Guide"],"extracted_tech_keywords":["AI","ML","Ray","PyTorch","Colab","NumPy","TPU","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/guide-to-perceiver-a-scalable-transformer-based-model\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":36136,"title":"Explainable AI Is Driving The Credit Analytics Market Now, Says Matthieu Garnier Of Equifax","content":"Matthieu Garnier, SVP Data and Analytics at Equifax, had some interesting insight to share with his audience while interacting with data science students at Manipal Pro Learn campus in Bengaluru. A mathematician by profession, Garnier has close to two decades of experience in the field of data science and spoke about how he attempts to solve everyday challenges in data science and analytics mathematically. Analytics India Magazine caught up Garnier to know more about his thoughts on how artificial intelligence and machine learning are driving the changes in credit analytics. Garnier, who leads a team of 400 people across Equifax’s 23 offices across the globe, is charged with handling the company’s overall data sets and derive knowledge from these raw data. Q: Fake data just like illegitimate news is a challenge for companies who are handling large data sets, could you tell how Equifax is mitigating this challenge? Data management is very important and you need to have good data to build a good model. We have some techniques to clean the data, identify data that are good by combining techniques through very simple routes to make sure that the data is relevant or not. Q. What are some of the recent trends with regards to credit analytics in India and worldwide? Advancement is the area of AI and ML is happening quite fast, if you ask me about the recent trend then I would say that it falls under three main categories Data protection and data privacy: In each country that Equifax is operating, including India, the regulators want to make sure that the activities around data are well-known and well-explained. There are regulations around how we use to make decisions around AI and data. Sophistication in AI: The Velocity and sophistication of AI are increasing and it is becoming more powerful as the systems using more and more data. For instance, earlier in credit when we wanted to build a model on behaviour, we used to take samples from population index, as a result of this we were losing out on a lot of information. Now that there is no limitation on storage or processing, we are able to train our algorithms in a better way. Transparency: That is in terms of how we make sure that people know about their data, how banks can explain to people when we take credit decision while making sure that all the sophistication that we are building towards data and analytics are explainable Q. With the onset of GDPR, how is Equifax complying to the rule? GDPR is one of the regulations that we have in Europe and some people say that it is the most advanced one, this is because of the idea of GDPR is to make sure that people can own their own data and give consensus each specific use of data. For Equifax, we need to adjust and adapt our processes and making sure that we are 100 per cent compliant with the regulation. This means that we need to keep in our database the consent of the people and ensure the quality of the data. The positive thing that I see in GDPR is that key players are spending more time on data protection and ensuring data quality. Q. According to you how are AI and ML driving the future of credit risk? All industry is using AI and they are trying to produce people behaviour through algorithms. This is one primary reason why we are seeing more and more sophistication and complexity. As a result of this, we are able to find the connection in data which we weren’t able to do in the past. When you are looking at credit, there is a great need for you to explain your decision and it took us a long time to include these complex models in credit. The adoption of explainable AI in credit has increased so much- that is we can give reasons based on outcomes of the models. Q. Which are some of the clients for Equifax India and could you point out a use case? There are many financial institutions and other enterprises that we are helping to the better credit decision. Now we are using AI and other deep learning technologies for what we call value-based pricing, that is pricing a service based on the data that we have. We are using AI to bring cost-efficiency and also apply it in all customer value chain- that is to identify targets of prospective customers, data collection, to identify people who haven’t paid their bills or loan, for even decision making, cross-selling. One particular use case is that of ZoomCar wherein they wanted to improve the smooth running of their business. They approached us to know more about the customers’ buying pattern. For instance, they wanted particular information about who wants to buy a car and who are the prospective customers and narrow it down to the selection of people and their identity based the data and financial information that we have. Q. What is Equifax India’s future plan with regards to the adoption of emerging technologies like AI and ML? We are bringing new technology called Conrea to India, it is an analytical sound box where we have a lot of tools and capabilities in terms of analytics. The idea is to create a data lake, where were will push a lot of anonymised data with tools like Hadoop around it. Further, we plan to give access to this environment to universities and our other customers so that they can access data and try to find a new correlation or knowledge based on these data. So, our plan for 2019-2020 will bring this advanced environment to India with no limitation in terms of volume and processing.","excerpt":"Matthieu Garnier, SVP Data and Analytics at Equifax, had some interesting insight to share with his audience while interacting with data science students at Manipal Pro Learn campus in Bengaluru. A mathematician by profession, Garnier has close to two decades of experience in the field of data science and spoke about how he attempts to […]","categories":["AI Features"],"tags":["data lake","Data Science","Interviews and Discussions","Machine Learning"],"author_name":"Akshaya Asokan","publish_date":"2019-03-12T06:24:15","publication_year":"2019","word_count":963,"keywords":["data science","Go","artificial intelligence","machine learning","AI","ML","Machine Learning","RAG","deep learning","analytics","Data Science","R","data lake","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/explainable-ai-is-driving-the-credit-analytics-market-now-says-mathieu-garnier-of-equifax\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10067605,"title":"India’s apparent mockery of VPN","content":"“VPN service providers that do not adhere to the latest cyber-security guidelines issued by the Indian Computer Emergency Response Team are ‘free to leave India’ if they do not comply with the rules”, Minister of State for Electronics and Information Technology Rajeev Chandrasekhar said recently. For those not heavily invested in cybersecurity news, this is Chandrasekhar’s retaliation to the backlash against the CERT-In’s new rules for VPN providers. CERT-In issued guidelines that mandate all VPN companies to store user data for up to five years. Even if the user has deleted the VPN account, this data has to be maintained. VPN companies will need to maintain validated names of customers using such services, period of hire, including dates, IPs allotted to or being used by the members, e-mail addresses, IP addresses and time stamp used at the time of registration or on-boarding, the purpose for hiring services, etc. Companies will also be required to report cybersecurity incidents to CERT-In within six hours of becoming aware of them. The new rules will come into effect in late June. Last week, Economic Times reported that CERT-In issued some clarifications to the rules, saying they would apply only to individual VPN customers and not to enterprise or corporate VPNs. Image: CERT-In_Directions_70B_28.04.2022.pdf (Read the full document by the government here). Government is not backing down While releasing the FAQs for this directive recently, Chandrasekhar did not mince his words. He clearly stated that if a VPN provider wants to hide and not reveal those who use VPN to do business, they can leave the country. That is the only option available. Technological challenge? This is not the first time that the Indian government has tried to curb the problems that can crop up with VPNs. As per a report by Medianama, last year, a plea from a Parliamentary Standing Committee asked the government to block virtual private networks in the country. According to the panel, VPNs are a technological challenge, and criminals use them to remain anonymous online. It also asked the Ministry of Home Affairs to coordinate with the Ministry of Electronics and Information Technology to use the help of internet service providers (ISPs) to ban VPN services in India. What’s the big deal about VPNs? A VPN can hide your online identity and make it difficult for third parties to track your activities online and steal data. Essentially, a VPN allows the network to “redirect it through a specially configured remote server run by a VPN host. If you surf online with a VPN, the VPN server becomes the source of your data,” says Kaspersky. Your internet service provider cannot keep track of which sites you are clicking and visiting. VPN providers are pissed The reactions from VPN providers have obviously not been positive. As per a report by Economic times, Laura Tyrylyte, head of public relations at Nord Security, said that the company is investigating the new directive recently passed by the Indian government and exploring the best course of action. There is still some time for the law to come into effect. She even went on to say that the company may remove their servers from India if no other options are left. The same report revealed that Surfshark said that its technology does not allow users to log in user’s information. “We don’t collect or share our customer browsing data or any usage information,” said Gytis Malinauskas, head of the company’s legal department. Yegor Sak, founder of Canada-based VPN company Windscribe, told MoneyControl that Windscribe does not collect or store the origin country of any customer. The company has no idea where a person is from when they use the firm’s service. Rajeev Chandrasekhar’s requirements are impossible to implement, Yegor added. India often features at the top of VPN markets A Surfshark 2022 report shows India has the biggest VPN market, followed by China and Indonesia. The US and Brazil take the fourth and fifth positions, respectively. The global VPN market was valued at USD 25.41 billion in 2019 and is projected to reach USD 75.59 billion by 2027, as per a report. Image: Surfshack Surfshark also says that the interest in VPNs peaked twice in the last two decades – in 2004 and 2019 (if we look at Google search volumes). Image: Global VPN Adoption Index | Atlas VPN Countries and their VPN status USA In the US, using a VPN is completely legal. Obviously, if VPNs are used for non-legal online activities, like accessing child pornography, cybercrimes, accessing the dark web, etc., it will attract strict punishment. China This has been an area of interest and debate for quite some time now. Even though VPNs are technically legal in China, their use is heavily restricted, and most VPN services are blocked. The government licenses only VPN providers that comply with its stringent terms of service. Besides, the government frequently bans unlicensed VPNs. Russia In 2017, Russia came down heavily on the use of VPNs and banned them. It was aimed at VPN providers who denied submitting data to the Russian government. A wave of bans came in 2021 as well. As per the Moscow Times, Russia’s communications regulator Roskomnadzor said that it had blocked access to some of the world’s largest VPN providers, which include Nord VPN and Express VPN, following an investigation.","excerpt":"VPN companies will need to maintain KYC data of the users, among other CERT-In guidelines put in place by the government.","categories":["IT Services"],"tags":["VPN"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-05-23T13:00:00","publication_year":"2022","word_count":886,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","ViT","R","VPN"],"extracted_tech_keywords":["AI","Aim","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/indias-apparent-mockery-of-vpn\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10074753,"title":"Can Midjourney Give Birth to a Movie? It Just Did!","content":"You have heard of AI-generated images and videos, but how would you react to an AI-generated movie? Fabian Stelzer, a Berlin-based entrepreneur and AI enthusiast, has done just that. Taking the frenzy of text-to-image generation tools like Midjourney to another level, Stelzer has produced a movie named ‘Salt’. “I am carrying this project entirely on my own, but the heavy lifting is done by these incredible new AI tools,” said Stelzer. With an academic background in cognitive neuroscience, Stelzer started a machine learning design startup called EyeQuant fresh out of college, which served clients like Google, Canon, Spotify and eBay. The making of ‘Salt’ Over the past year, Stelzer stumbled upon text-to-image generation tools. “The first time I used Midjourney, my chin dropped to the floor – it was the most powerful technology I had witnessed since putting up a computer for the first time,” he told Analytics India Magazine. Stelzer believes these tools are absolute game changers, “almost as big as the invention of photography”, and promise a new era in art, films, and other creative fields. “I wanted to hop on to this bandwagon as soon as possible; this is basically the premise over which my movie, ‘Salt’, was conceived, developed and maintained,” he said. He calls it a fully AI-generated multi-plot interactive film experience. https:\/\/twitter.com\/fabianstelzer\/status\/1565085199322456069?s=20&t=fPZAtR4pcmuCNhdmTMiijA It began with generating images using Midjourney, with the idea of weaving them together to build a trailer. “I don’t come from a filmmaking background and used simple Da Vinci Resolve to cut it together,” he said. However, the outcome turned out to be a truly experimental and innovative short story plot which he shared on Twitter and received rave reviews. “I realised that filmmaking is a costly affair. Since my movie is completely made using AI, my first thought was that, effectively, the cost of production is zero or approaching zero. I did not want to settle down with just one plot but have multiple of them,” he said. Stelzer would allow his followers to select from a range of prompts. The most popular ones would then be used to develop an alternative plot. He primarily used Midjourney to create Salt, along with Synthesia, but is now experimenting with Stable Diffusion and Avatarify for character animation. https:\/\/twitter.com\/fabianstelzer\/status\/1565092016136019968?s=20&t=fPZAtR4pcmuCNhdmTMiijA “So the vision is to have an extremely bottom-up decentralised film universe where everyone is able to render their own episodes, plots, and stories,” he said, and added that currently, his project is following the ‘choose your own adventure’ formula, where he receives inputs\/prompts from his followers. He calls it the sandbox-style film. The final goal, however, is to decentralise in a way that anyone with a great idea can design their own plot and become a co-creator. Stelzer is an optimist and truly believes that AI will see rapid growth, and within four years, we may have text-to-video generation tools. The barrier right now is fixing an embedding in the latent space and recreating it for different contexts. ‘Battleprompts’ Another major project that Stelzer has worked on is ‘Battleprompts’. “The main thought behind this project was whether an AI native game would run on Twitter,” he said. Inspired by the popular 80s game Cobra, ‘Battleprompts’ accepts text prompts to generate characters. “Player 1 can give a prompt like, “Kermit the Frog with four hands and laser eyes wearing Samurai swords”. Player 2’s prompt could be “a space hamster that can change time and bend Physics”. With these prompt inputs the GPT-3 can generate a fight sequence between the avatars. It is like AI taking a roll on the dice,” explained Stelzer. He used the GPT-3 Da Vinci model for this. https:\/\/twitter.com\/fabianstelzer\/status\/1559517876943523846?s=20&t=fPZAtR4pcmuCNhdmTMiijA Stelzer believes that with sufficient technical support and advancement, we may have an AR version of this game. “You will have these strange, funny avatars fighting it out on your kitchen table,” he said and added, “It’s like those fun experiments, which become the predecessors to more serious projects in AI native entertainment media. This is what I’m interested in.” Ethics and morality Right at the beginning of his conversation with us, Stelzer made a disclaimer that he does not think text-to-image generation tools will make the job of an artist or illustrator redundant. “When photography came around, people thought this would be the end of the art of painting. But even today, we have paintings, some even worth more than what they were 100 years ago. I believe going forth, we will see entirely new genres based on new synthetic media tools,” he said, When quizzed about the growing noise around the copyright issues associated with AI-based image generations, Stelzer said there was a misconception that AI tools are like search engines for images where it just pulls together art from online and recombines them in new ways. “It is not a search engine for images, it creates new images,” he said. “All art is inspired. Nobody can claim to really create completely original art. Everyone’s standing on the shoulders of giants. So with respect to AI art, I would like to basically use the same principles – moral and legal principles – that apply to humans as well. There is a clear case of theft, and there is inspired art,” Stelzer signed off.","excerpt":"Taking the frenzy of text-to-image generation tools like Midjourney to another level, Stelzer has produced a movie named ‘Salt’","categories":["AI Features"],"tags":["Interviews and Discussions","MidJourney","text to image generation"],"author_name":"Shraddha Goled","publish_date":"2022-09-09T15:00:00","publication_year":"2022","word_count":873,"keywords":["Go","MidJourney","API","machine learning","text to image generation","AI","GPT","Aim","stable diffusion","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","R","Go","API","GPT","stable diffusion","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-midjourney-give-birth-to-a-movie-it-just-did\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":7869,"title":"CYPHER 2015, A Huge Success","content":"The first ever CYPHER2015, Analytics Summit, held in Bangalore on Sep 12, 2015, was a huge success with more than 400 analytics enthusiasts and professionals attending the one day summit. Keynote speakers included ATUL JALAN – CEO at Manthan Systems, LAKSHMANA GNANAPRAGASAM – Head of Analytics at Quantium India, MOUMITA SARKER – Director – Client Delivery at Cartesian Consulting, SAMEER DHANRAJANI – Business Leader, Cognizant Analytics at Cognizant Technology Solutions and VIVEK RATNA – Country Director, HP Big Data Platform. Bhasker Gupta, founder of Analytics India Magazine shared his thoughts regarding this event. He said, “We realized that there is a dearth of quality conferences around Analytics in India. India being at the forefront of the Analytics wave needed a platform where industry leaders, professionals and various organizations can come together to exchange ideas. We took this initiative with the first edition of Cypher. The success of this event has motivated us to bring back Cypher again next year at a much larger scale.” The one-day summit was held at the Hotel Park Plaza in Bangalore. Designed to be an annual meeting to bring analytics professionals at one platform, the summit discussed various trends and challenges in analytics by way of Panel discussions. Attendees also received pertinent information regarding Analytics education in India through various workshops and exhibits. HP was the main sponsor for the summit. Sameer Dhanrajani, Business Leader, Cognizant Analytics and Data Science at Cognizant Technology Solutions said, “It was great to see a congregation of over 400+ enthusiastic analytics and data science professionals under a single forum. CYPHER 2015 was a well-envisaged event and our partnership with the Analytics India Magazine was complementing our key approach of innovative analytics thought leadership and advocacy and continued push towards path-breaking topical trends within the analytical and data science fraternity within India. I congratulate Bhasker from AIM for organizing this wonderful event and giving us all a common platform to discuss, brainstorm, and collectively learn from each other’s experiences on the forthcoming trends and topics shaping our algorithm economy.” Organized by Analytics India Magazine, Cypher 2015 showcased various B-schools, service providers and analytics products in its exhibits. Among the B-schools, Great lakes institute of Management, Praxis Business School and Alliance University showcasing their analytics programs at the exhibit. With 32 speakers between keynotes, panel discussions and workshops; the summit saw more than 150+ companies represented. Vivek Ratna, Country Director at HP Big Data Platform said about the event, “Cypher 2015 did a great job in congregating the niche audience interested in knowing the POV of professionals having advanced experience in analytics. It covered diverse topics and had cherry picked the experienced speakers with varied backgrounds to deliver the sessions. In my assessment, attendees took back the knowledge not available in text books and they will look forward to coming back again in subsequent editions.” The attendees had the opportunity to meet analytics leaders as well as representatives from Cognizant, Quantium, AnalytixLabs, Team Computers, Wiley, Bridgei2i, Manthan among others. At the Summit, there was a recognition that leading businesses are already taking action to build analytics competencies. In all sectors, business has developed solutions, continues to innovate and is preparing to accelerate the scale and pace of analytics deployment. Exciting workshops were conducted by Sumeet Bansal (CEO of AnalytixLabs), on “Internet of Things meet Big Data”, PAUL MEINSHA– USEN (Vice President, Data Science at Housing.com) on how Housing.com has grown and structured its data science teams and RAVI VIJAYARAGHAVAN (Head of Analytics at Flipkart) gave glimpse into some of Flipkart’s analytics solutions to drive business growth and User experience for its consumers and sellers. Moumita Sarker, Director Delivery at Cartesian consulting covered the applications of Genome Mapping in Marketing. For more information, visit www.analyticsindiasummit.com For event videos, subscribe to our YouTube channel. [custom_gallery source=”media: 7877,7876,7875,7874,7873,7872,7871″ link=”lightbox” width=”140″ height=”140″] All Pictures of the event can be accessed from our Facebook Page.","excerpt":"The first ever CYPHER2015, Analytics Summit, held in Bangalore on Sep 12, 2015, was a huge success with more than 400 analytics enthusiasts and professionals attending the one day summit. Keynote speakers included ATUL JALAN – CEO at Manthan Systems, LAKSHMANA GNANAPRAGASAM – Head of Analytics at Quantium India, MOUMITA SARKER – Director – Client […]","categories":["Deep Tech"],"tags":["analytics conference","analytics summit","Praxis Business School"],"author_name":"Дарья","publish_date":"2015-09-18T10:37:09","publication_year":"2015","word_count":649,"keywords":["big data","data science","Go","API","AI","analytics conference","RAG","Aim","analytics","analytics summit","GAN","R","Praxis Business School"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","API","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/cypher-2015-a-huge-success\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":31303,"title":"Yann LeCun Says Facebook Is ‘Dust’ Without Deep Learning, And No One Is Disagreeing","content":"One of the foremost experts on Deep Learning, Facebook AI supremo and Chief AI Scientist, FAIR, Yann LeCun made another key announcement regarding deep learning. The French AI expert who played a pivotal role in setting up Facebook’s lab in Paris shared in an interview to CNN, “If you take the deep learning out of Facebook today, Facebook’s dust. It’s entirely built around it now.” The statement typifies the current stand on technology by Mark Zuckerberg led Menlo Park giant which has been embattled over its role in US elections. Even though statement has been largely dubbed controversial, LeCun summed up Facebook’s transformation around AI and ML succinctly. Over the last three to four years, Facebook has been slowly but surely transforming its business around intelligent systems technology. The change can be seen in features such as posts, translations and newsfeed algorithms which is the core of the social network platform. Facebook applied deep learning to combat hate speech and misinformation in countries like Myanmar. The social media giant was criticised when the platform was said to have fueled ethnic violence against the Rohingya population. The company also said AI application created by Facebook is now capable of flagging 52 percent of all content it gets rid of in Myanmar before it is reported by users. Here are some of the main applications of deep learning inside Facebook: FB Learner Flow: FB Learner Flow is viewed as the backbone of Facebook’s AI. The famed internal platform has the ability to reuse algorithms in very different products and has the scale to run thousands of simultaneous machine learning experiments. The platform has powerful tools for automatic generation of user interface experiences from custom definitions and automatic parallelisation of Python code. Building Perception: This is a product that uses convolutional neural network, a popular technique pioneered by LeCun, the neural networks. Building Perception is a large system trained to understand a range of data better which can be anything such as photos, videos and even voices. Facebook’s Text Understanding Engine DeepText: Facebook’s DeepText application builds on deep learning research to tackle language challenges that are hard to solve via traditional NLP techniques. DeepText is built upon mathematical models that can mine the semantic relationship between words and can work on multiple languages. When asked about what if all this intelligence was used to dominate and control various aspects of personal lives, the head of AI at Facebook said, “The desire to dominate is not correlated with intelligence. In fact, we have many examples of this. It’s not the smartest of us that necessarily wants to be the chief.”","excerpt":"One of the foremost experts on Deep Learning, Facebook AI supremo and Chief AI Scientist, FAIR, Yann LeCun made another key announcement regarding deep learning. The French AI expert who played a pivotal role in setting up Facebook’s lab in Paris shared in an interview to CNN, “If you take the deep learning out of […]","categories":["AI Features"],"tags":["Deep Learning","Facebook","Facebook AI","Technology","Yann LeCun"],"author_name":"Abhijeet Katte","publish_date":"2018-12-10T11:46:48","publication_year":"2018","word_count":436,"keywords":["Go","Yann LeCun","machine learning","Facebook AI","AI","neural network","ML","NLP","Python","Technology","deep learning","Facebook","Deep Learning","CNN","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","Python","R","Go","CNN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/yann-lecun-says-facebook-is-dust-without-deep-learning-and-no-one-is-disagreeing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10006231,"title":"Top 7 Upcoming Deep Learning Conferences To Watch Out For","content":"In recent years deep learning has proved to be a critical aspect of machine learning with exciting applications to solve real-world problems across different sectors. Starting from creating virtual assistants, visual recognition and language translation to fraud detection, document processing, as well as self-driving cars, deep learning has proved to be immensely beneficial. As a matter of fact, in many areas, deep learning has also outsmarted traditional machine learning. With the field becoming popular among businesses, many conferences have emerged that delve deeper into the field of deep learning for people to understand it better. Not only these events will provide a deeper understanding of the advancement of deep learning space but will also offer a chance for deep learning practitioners to network with experts and researchers from the field. Analytics India Magazine has curated a list of seven upcoming deep learning conferences to watch out for: Also Read: Top 10 Computer Vision Conferences To Watch Out For Deep Learning DevCon 2020 When: 29-30th October 2020 Where: Virtual About: Deep Learning DevCon 2020 is a two-day conference hosted by the Association of Data Scientists (ADaSci) in association with Analytics India Magazine for deep learning practitioners and innovators. The conference provides an in-depth understanding of the latest research and advancements in the field of deep learning led by best-minds of the field. With 40 speakers and over 500 attendees from more than 200 organisations, this conference will host paper presentations, exhibitions, workshop as well as hackathons. Click here to know more. Deep Learning Summit When: 15-16th April 2021 Where: London, UK About: Deep Learning Summit is a two-day event hosted by ReWork, which aims to bridge the gap between the latest research advancements and the real-world applications with deep learning. The conference provides a deep dive session to facilitate interactive discussions, hands-on workshops and technology lab. Deep Learning Summit also provides a hybrid-event option where it allows participants to join in virtually to attend the conference if missed in-person. It also has multiple tracks covering topics like neural networks, machine learning, deep learning algorithms, computing systems, NLP, computer vision, and speech recognition, to name a few. Click here to know more. International Conference on Deep Learning Technologies When: 10-11th September 2020 Where: Tokyo, Japan About: International Conference on Deep Learning Technologies is a conference specifically designed for professionals in the deep learning field, hosted by The International Research Conference. The conference aims to bring together experts brains of the industry along with academic scientists and research scholars to provide a deep understanding of this field as well as exchange their experiences with deep learning technologies. Further, this conference also allows leading experts of the deep learning industry to present their papers, share their latest innovations, and speak on the challenges associated with deep learning technologies. Click here to know more. International Conference on Recent Advances in Deep Learning Technologies When: 17th-18th September 2020 Where: Amsterdam, Netherlands About: International Conference on Recent Advances in Deep Learning Technologies is another conference that is organised by The International Research Conference. It acts as a platform for leading academic scientists, researchers and research scholars to share their ideas and thoughts on this evolving field. The conference asks for contributions of journals and research papers, which are then addressed in the discussion, including figures, tables and references of novel research materials. Click here to know more. Also Read: Top Seven Virtual Conferences On AI European Machine Learning & Data Mining Conference When: 14-18 September 2020 Where: Virtual About: European Machine Learning & Data Mining Conference has a significant focus on the deep learning field. With talks, lectures, tutorials and workshops, this conference covers topics like reinforcement learning, pattern mining, clustering, the architecture of neural networks as well as deep learning. The conference has been focused on providing an interactive and rewarding experience for all contributors, participants, and sponsors, with their virtual conference. Click here to know more. International Conference on Deep Learning for Information Fusion When: 23rd-25th September 2020 Where: Vancouver, Canada About: Another event hosted by The International Research Conference — International Conference on Deep Learning for Information Fusion that significantly focuses on the exchange of ideas in the field of deep learning and its impacts on society. From discussing the latest innovations in the area to the concerns and challenges associated with deep learning, this conference acts as a comprehensive platform to get hands-on. This conference has been known to present some of the critical papers in the field providing a good understanding to the attendees. Click here to know more. International Conference on Machine Learning and Deep Learning When: 29-30th October 2020 Where: Los Angeles, US About: International Conference on Machine Learning and Deep Learning is also a two-day event that envelopes academic scientists and industry experts to exchange their knowledge and experiences while working in the field of machine learning and deep learning. With a premier interdisciplinary platform for industry experts and researchers to present their recent innovations, this conference also shares papers, trends, and theories in the fields of machine learning and deep learning. Click here to know more.","excerpt":"In recent years deep learning has proved to be a critical aspect of machine learning with exciting applications to solve real-world problems across different sectors. Starting from creating virtual assistants, visual recognition and language translation to fraud detection, document processing, as well as self-driving cars, deep learning has proved to be immensely beneficial. As a […]","categories":["AI Trends"],"tags":["Active Learning","Deep Learning","deep learning research","latest advances","latest machine learning innovation","latest technology in machine learning"],"author_name":"Sejuti Das","publish_date":"2020-09-08T12:00:00","publication_year":"2020","word_count":848,"keywords":["machine learning","AI","neural network","latest machine learning innovation","latest advances","virtual assistants","computer vision","Active Learning","deep learning research","NLP","deep learning","latest technology in machine learning","analytics","Aim","Deep Learning","fraud detection"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","NLP","computer vision","analytics","Aim","virtual assistants","fraud detection"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-upcoming-deep-learning-conferences-to-watch-out-for\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171395,"title":"IIT Kanpur is Behind India’s Next Wave of $1 Billion Startups","content":"IIT Kanpur has emerged as a significant hub for deep tech startups, focusing on product-led innovation across sectors like AI, defence, and healthcare. Its incubation centre is now powering both domestic growth and global expansion. Established in 2000, the Startup Incubation and Innovation Centre (SIIC) is one of India’s oldest technology business incubators, operating under the Foundation for Innovation and Research in Science & Technology (FIRST), a Section 8 company promoted by IIT Kanpur. Today, SIIC hosts around 100 startups working on next-generation technologies across various sectors, including AI, healthcare, agriculture, and manufacturing. It offers funding support of up to ₹20 lakh and one year of dedicated incubation that includes mentorship, technical due diligence, and business guidance. The incubator at IIT Kanpur has always leaned towards product-driven startups. “Even back then, and still today, we focus predominantly on ventures with a tangible technological product,” professor in-charge of incubation and innovation at SIIC, Deepu Philip, said in an exclusive interview with AIM. “We don’t typically engage with service-based models unless there’s a strong tech product at the core,” he added. Philip said that IIT Kanpur has incubated around 425 startups, out of which approximately 200 can be called reasonably successful, based on criteria like revenue, valuation, and employment. “We have about 200 companies that have either secured a valuation of ₹10 crore, made consistent revenue for three years, or employed five or more people consistently.” Philip added that 111 of these startups have undergone third-party independent valuations, collectively worth ₹7,228 crore (~$1 billion). IIT Madras, which has supported 457 deep-tech startups over the past 12 years, is often the first to come to mind when discussing startups. The startups, including two unicorns and one IPO-bound company, are now collectively valued at more than ₹50,000 crore. Yet, IIT Kanpur is steadily carving out its own identity as a thriving startup hub. In May, IIT Kanpur collaborated with the Wadhwani Foundation to launch a School of Advanced Artificial Intelligence and Intelligent Systems. Philip shared that the institution received a $35 million grant from the Wadhwani Foundation. Moreover, he shared that IIT Kanpur has submitted an application under the IndiaAI Mission to develop a sovereign AI model. Philip added that IIT Kanpur is actively involved in innovative public-private funding models. He explained the ‘40-40-20’ model for the India AI mission, where 40% is funded by the central government, 40% by the state government, and the remaining 20% is supported by the industry.To help startups find markets beyond India, IIT Kanpur has signed an MoU with Toyota Tsusho and the Bill and Melinda Gates Foundation to support international scaling. Most recently, the SIIC has also partnered with NMexus, a new global business accelerator launched in Albuquerque, New Mexico. The collaboration is expected to help Indian startups gain access to the US through curated support. Startups incubated at SIIC will benefit from soft-landing infrastructure, regulatory guidance, mentoring, and market access services in North America. As per the official release, NMexus will host up to 40 companies annually, aiming to create around 1,500 jobs and contribute over $400 million to the economy over five years. “The New Mexico centre is meant to facilitate the global expansion of startups. Once they gain a foothold internationally, they can shift operations there to scale,” Philip said, adding that Western markets, especially the US, offer larger revenue opportunities and later-stage funding. IIT Kanpur is Open to Everyone According to Philip, IIT Kanpur’s incubation model is uniquely inclusive—open to innovators from any college or region in India, not just IIT students. Startups are selected based on merit and incubated through the entire technology readiness lifecycle, from idea to production-ready stage. “We allow anybody from the country to integrate,  there’s no distinction. If they fit the criteria, we incubate them, no matter where they come from or which college they belong to,” he added. He highlighted Artificial Intelligence and Innovation Driven Entrepreneurship – Centre of Excellence (AIIDE-CoE), an initiative backed by the UP government along with institutions like Uttar Pradesh Electronics Corporation Limited (UPLC), which provides seed funding, technical and business mentorship, IP support, and global scale-up opportunities. “Our support starts at TRL (Technology Readiness Level) 0, which is the ideation stage. We help them scale up to TRL 8 or 9, where the product is ready for the market,” he said. He added that one of the biggest hurdles in deep tech startups is covering IP and patenting costs. “Filing and maintaining patents can easily cost ₹60 to ₹70 crore over time. That is why government support is crucial.” Philip highlighted that in a startup’s early days, maintaining cash flow matters more than chasing big investments. He pointed out that even small grants, loans, or CSR funds can keep things moving and help startups survive long enough to gain momentum. He drew a line between ‘hard money’ like VC or angel investments, which often come with high expectations and a loss of control, and ‘soft money’ like government or CSR support, which gives startups the breathing room they need. “Our job is also to ensure the smooth flow of soft money, which allows the startups to work.” Successful IIT Kanpur Startups The professor further said that the institute’s focus on product-based startups, particularly in deep tech, distinguishes its efforts. He shared examples of companies like Ananant Systems, Life and Limb, xTerra Robotics and VU Dynamics. Philip explained that Ananant is at the forefront of communication chip design in India, specifically for 5G or 6G domains. He stressed their contribution to bringing the crucial chip design and concept-proofing aspects, traditionally done abroad, to India. He explained that this domestic capability is vital for fostering an ecosystem of associated companies building drivers, embedded systems, and AI applications. Moreover, he revealed that Life and Limb is developing intelligent prosthetics that respond to muscle twitches and voice commands, moving beyond mere cosmetic function. “Based on the muscle twitch, it may trigger simple functions like grabbing, operating a machine or holding something.” On the other hand, xTerra Robotics is building unmanned ground robots, including robotic dogs used for tasks such as bomb disposal and intelligence gathering in challenging environments. Meanwhile, VU Dynamics, a faculty-led company, specialises in chemical drones equipped with autonomous target acquisition, path planning, and execution capabilities. Philip cited Curadev, a Kanpur-origin immunotherapy startup, as one of the success stories. The company has developed cancer-treatment drugs and made profitable exits. He believes most IIT Kanpur startups prefer private equity over IPOs. Citing the example of Phool, he said, “It’s currently valued at ₹2,000 crore. So they can happily go for IPO, but they are more interested in securing investment.” He cited the example of a 45-kg runway-based UAV developed for BEML, which was showcased at Aero India 2025. Defence Public Sector Undertakings (DPSUs) are now picking it for large-scale production. According to a recent report, it may soon become part of the Indian Army’s UAV arsenal. Quality over Quantity Referring to Union minister Piyush Goyal’s recent remark about the lack of deep tech startups in India, Philip said it is unrealistic to expect 100 deep tech startups to emerge overnight. “Many people don’t even understand what deep tech means. It’s just a fancy term today,” he said, adding that IIT Kanpur prioritises quality over quantity, even if that results in fewer startups. He contrasted this with China’s philosophy that quality emerges from quantity. “We follow a different path—quality first,” he said, arguing that such an approach leads to long-term impact, much like what the US has achieved. Praising UP CM Yogi Adityanath, Philip said that the UP government is marching steadily towards a $1 trillion economy. According to him, there’s a positive synergy between academia, industry, and the state government, particularly in semiconductors, defence, and biotech. “When we started working on semiconductor chip design, the UP government began building a semiconductor park in Noida. When we focused on defence, they announced the defence corridor. There is a clear alignment.” “There is no dearth of talent in this country. We have to attract them, mentor them, and their products have to be showcased,” Philip concluded.","excerpt":"“The institute has incubated around 425 startups, out of which approximately 200 can be called reasonably successful.”","categories":["AI Startups"],"tags":["IIT"],"author_name":"Siddharth Jindal","publish_date":"2025-06-05T13:44:03","publication_year":"2025","word_count":1349,"keywords":["Go","artificial intelligence","AI","innovation","ML","Aim","ViT","GAN","IIT","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","R","Go","GAN","ViT","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/iit-kanpur-is-behind-indias-next-wave-of-1-billion-startups\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":30332,"title":"Say Hello To STACL, Baidu’s New Innovation For Improving Simultaneous Translation","content":"United Nations(New York) Translating a language is a tedious job which involves listening, speaking and a thorough command over the language. An assignment which needs a high degree of skills for humans to have might be a cakewalk for modern computer technologies such as machine learning or artificial intelligence to acquire. Baidu, in a recent development, has developed a device which uses ML and AI to translate languages, while boasting anticipation capabilities and controllable latency. It is an automated system that ensures a high-quality translation between two languages. It would prove to be highly advantageous against traditional consecutive interpretation, wherein a translator waits until the speaker pauses to start translating. While this method usually doubles the time needed, simultaneous interpretation is a faster option where the translator begins translating just a few seconds into the speaker’s speech. Issues With Current Translation System The usage of the verbs and figures of speeches may vary significantly in various languages. For example, in English, the verb comes prior to the sentence whereas in German it comes at the end of the sentence. Same with Chinese to English translation. This variance in order of words is a major hindrance for real-time human translators, causing undesirable latency and rendering speaker out of sync with the speaker. Simultaneous Translation with Anticipation and Controllable Latency (STACL) promises to address the issue of when and how to use words. How Does It Work? STACL works on the principle of predictive analysis. As the researchers from Baidu explain, the model doesn’t predict the source language words in the speaker’s speech but instead directly predicts the target language words in the translation. The model seamlessly fuses translation and anticipation in a single “wait-k” model. It means that translation is always k words behind the speaker’s speech to allow context for prediction. The model is trained to use the available prefix of the source sentence at each step to decide the next word in translation. The researchers added that in the Chinese prefix Bùshí Zǒngtǒng zài Mòsīkē (“Bush President in Moscow”) and the English translation so far “President Bush” which is k=2 words behind Chinese, their system accurately predicts that the next translation word must be “meet” because Bush is likely “meeting” someone (e.g., Putin) in Moscow, which is done long before the verb appears. The model, however, needs to be prepared about the speaker’s topic and style beforehand, just as human translators need to be. This is done by training with large amounts of data which have similar sentence structures. This enables the model to anticipate words in a sentence most likely to be spoken with a reasonable accuracy. With the current capabilities, STACL aims at dealing with latency. How STACL Can Match Human Interpretation Baidu has used a technology named 3.4 BLEU (Bilingual Evaluation Understudy), which is the backbone of the entire architecture. It is essentially a standard algorithm to estimate the quality of text which has been machine-translated from one natural language to another. “It a standard evaluation metric for full-sentence translation quality by comparing a machine translation result with a human reference translation”, notes the website. While human translators can cover up to 60 percent of the source material with about three seconds delay, the new simultaneous system is much more efficient. While in the earlier Chinese to English simultaneous translation, the translator lagged behind by 3 Chinese words or about 1.5 to 2 seconds, the translation quality with new ML-model is about 5 BLEU points higher. Outlook While STACL shows significant potential, the researchers are still to overcome many limitations of the simultaneous machine translation system. The release of STACL is not proposed to take over the human interpreters yet but may use its capabilities to offer an improved service in the years to come.","excerpt":"Translating a language is a tedious job which involves listening, speaking and a thorough command over the language. An assignment which needs a high degree of skills for humans to have might be a cakewalk for modern computer technologies such as machine learning or artificial intelligence to acquire.     Baidu, in a recent development, […]","categories":["AI Features"],"tags":[],"author_name":"Bharat Adibhatla","publish_date":"2018-11-16T11:18:16","publication_year":"2018","word_count":629,"keywords":["Go","machine learning","artificial intelligence","programming_languages:R","AI","ML","programming_languages:Go","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/say-hello-to-stacl-baidus-new-innovation-for-improving-simultaneous-translation\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10161562,"title":"Google Gemini Now Available on Google Workspace","content":"On Thursday, Google announced that its best generative AI capabilities will now be included in Workspace Business and Enterprise plans, eliminating the need for add-ons. This brings AI-powered tools to businesses of all sizes, helping them enhance productivity, creativity, and innovation. “AI is foundational to the future of work,” said Jerry Dischler, president of cloud applications at Google. “Our goal is to make the transformative power of AI accessible to every business at an affordable price.” The new AI features are available in Gmail, Docs, Sheets, Meet, Chat, and other Workspace tools and will assist users in tasks such as drafting documents, summarising emails, and taking meeting notes. Additionally, users will be able to interact with Gemini Advanced, a next-gen AI tool that can aid in research, coding, and data analysis. “Businesses that embrace AI are gaining a significant competitive edge. Generative AI has reduced the burden of mundane tasks and has become a strategic partner for businesses, helping them bring their best ideas to life,” Dischler said. The update was rolled out to Workspace Business customers starting Thursday and will be extended to enterprise customers later this month. Pricing has been simplified as part of the integration. For example, if a customer on the Workspace Business Standard plan with the Gemini Business add-on previously paid $32 per user per month, under the new plan, they would be required to pay only $14. Google has emphasised its commitment to data security and privacy. “We don’t use your data, prompts, or generated responses to train Gemini models outside of your domain without permission,” Google stated. Additionally, Gemini adheres to privacy and security certifications, including SOC 1\/2\/3, ISO 27001\/17\/18, and HIPAA compliance. The pricing changes take effect immediately for new customers, and they will be updated for existing customers starting March 17. Small businesses are not currently affected by these pricing changes. The company encourages users to start a no-cost AI trial and learn more through upcoming webinars. “Today is just the start,” said Dischler. “We’ll continue to introduce new AI features throughout the year to empower our customers and help them innovate for the future.”","excerpt":"The new AI features are available in Gmail, Docs, Sheets, Meet, Chat, and other Workspace tools and will assist users in tasks such as drafting documents, summarising emails, and taking meeting notes.","categories":["AI News"],"tags":["Google"],"author_name":"Siddharth Jindal","publish_date":"2025-01-16T16:24:22","publication_year":"2025","word_count":354,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","llm_models:Gemini","ViT","generative AI","Google","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","ViT","innovation","llm_models:Gemini","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-gemini-now-available-on-google-workspace\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10095694,"title":"Can ChatGPT Actually Make You Rich?","content":"‘Get richer with ChatGPT’, ‘Making money online got easier’, ‘How you can make money with ChatGPT’, ‘Access my database of 125+ ChatGPT prompts to help you make money online for $5’. Wait, stop, calm down! So, if you’re fed up with all these paid advertisements trying to push different courses down your throats, luring you to make ‘easy’ money using ChatGPT, welcome to the club! Search engines and social media platforms like Instagram, Twitter, Facebook and even LinkedIn are brimming with such posts. These ‘influencers’ make ChatGPT seem like an easy money-minting machine. But have people actually been able to monetise ChatGPT and its free version? Well, a lot of individuals claim to have been leveraging its capabilities to generate income. From branding to app-creation to providing writing services, they claim it opens avenues to earn ‘free money’ whether you’re an aspiring entrepreneur, content creator, or a freelancer. You must’ve also come across tons of articles listing numerous ways to rake in the greens. These articles usually either bank on the chatbot’s capabilities to code and hint at it being used to build apps or websites ‘without any prior knowledge’ or for affiliate marketing, content marketing, optimising video production, or becoming a prompting expert amongst many other things. However, there are some who claim to have made money this way. Ukrainian entrepreneur Ihor Stefurak built a Chrome extension using ChatGPT despite having zero coding knowledge. He claims to have generated $1000 in revenue within just 24 hours of launching the extension. Many such claims and tutorials are flying fast and thick on YouTube, claiming to help make $100 a day but seem to know nothing about the tool, the model it is based on or even its name. Some coolly call it ChatGTP! Meanwhile, there has also been a notable surge in AI-written ebooks on Amazon. These are ranked as high as 50,377 on the Kindle book store and also have five-star ratings in the hundreds. According to reports, as of mid-February there were over 200 e-books in Amazon’s Kindle store, listing ChatGPT as the author or co-author. The number may not seem like a lot but the actual number may be much higher since many authors do not disclose in the Kindle store that their book was written entirely, or in part, by a computer because Amazon’s policies do not require it. In fact, the market is so saturated that making decent money out of this is improbable. According to Reuters, Brett Schickler, who wrote a children’s book titled, ‘The Wise Little Squirrel: A Tale of Saving and Investing’, (sold at the Amazon Kindle store for $2.99 and $9.99 for a printed version) netted less than $100. Can Only Enhance The reality is that ChatGPT alone cannot guarantee financial success. While ChatGPT can be an invaluable tool in reducing your workload, it is crucial to understand that it is not a magical solution that will automatically generate top-tier content for you. If you lack familiarity with a particular subject, you will struggle to differentiate between good and poor quality content. For example, let’s consider a scenario where you follow someone’s advice on starting a successful blogging business with the assistance of ChatGPT despite having no prior experience in content creation. Even if ChatGPT generates a basic piece of content, you might perceive it as exceptional because you lack the knowledge of what constitutes top-tier content. What you can actually do is create videos on how to make money with AI. As they say, ‘When there’s a gold rush, sell shovels and pickaxes’. On a serious note, what you can actually do is automate parts of your workflow with ChatGPT APIs, which are free to use. One can take up email affiliate marketing, content marketing, or learn and preach better avenues like prompt engineering. It is crucial not to be deceived by influencers who disseminate misleading content. Do not believe them when they claim that ChatGPT holds the secret to success or that it can effortlessly make you money. Instead, focus on learning and honing your skills, utilising ChatGPT as a tool to support you in accomplishing your goals.","excerpt":"ChatGPT can enhance your abilities 10 times, but not if you’re a 0, because 10×0 amounts to 0","categories":["AI Trends"],"tags":["Amazon","ChatGPT","OpenAI"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-06-23T16:31:50","publication_year":"2023","word_count":689,"keywords":["Go","ChatGPT","API","OpenAI","AI","Amazon","RAG","GPT","Aim","prompt engineering","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","Aim","RAG","prompt engineering","R","Go","API","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/can-chatgpt-actually-make-you-rich\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10074106,"title":"Ant Colony Optimisation and Predictive Maintenance in Indian Transport system","content":"India’s urban population is swelling 3% annually and hit the 500-million mark in 2021, which was a 32% rise from the 377 million in 2011. Eight major cities — Mumbai, Delhi, Kolkata, Chennai, Hyderabad, Bengaluru, Pune, and Ahmedabad — host a majority of this urban population. The expansion of our economy depends heavily on this urban sector, which accounts for over 60% of India’s income. While the population booms and the economy expands, can traffic remain unaffected? India’s metropolises are witnessing a rapid growth, but are also plagued with slow traffic. Mumbai’s peak hour traffic moves at 8 kmph, Bangalore is a notch better at 10 kmph; Delhi crawls at 8 to 9 kmph, while Kolkata’s main street remains bumper-to-bumper at 10 kmph. The type of traffic we’re referring to typically falls in two categories. The first is when there is persistent heavy traffic at the same time and location every day. This is what we generally refer to as the office-hour congestion. The second type of traffic snarl occurs when a random event, like a political demonstration, agitation, VIP movement, rally or celebration, takes to the road. When it comes to persistent traffic congestion, Ant Colony Optimisation (ACO) is a simple solution. Ants use trail pheromones to help other members of the colony get from their nest to a food source and return to the nest again in an efficient manner. Likewise, this model seeks out the most direct route, drawing inspiration from ants. How does it work? In this model, automobiles simulate ants and are assigned a pheromone value ranging from 0.5 to 2 depending upon the type of vehicle — heavy (2), medium (1), or light (0.5). For future research and prediction, all pheromone data is recorded in a database. When it comes to crossroads, the left lanes are typically kept open. The signal on the other lane is decided based on the intensity of the pheromone level on the lane (the higher the pheromone level, the longer the queue). As soon as the high-intensity pheromone lane gets cleared, the next lane with the highest intensity is chosen, and the previous one’s pheromone level is reset to 0. This goes on until the traffic on the road is cleared. ACO has proven to be effective in reducing traffic wait time in numerous trials. Additionally, incidents like emergency vehicle crossings and pheromone evaporation are taken into account by the model. Predictive maintenance in public transport Even if the traffic lights are optimised, there is always the problem of public transport vehicles not being in their best form. It takes months for a fleet of buses to get repaired. This is because of reactive maintenance in the current system. Emergency repairs, too, are a costly affair. Credit: Stratio This issue can be solved through predictive maintenance in the public transport system. A similar system was adopted by Go-Ahead Ireland, a public service provider in Ireland. After the induction of predictive maintenance, the company saw 57% fewer vehicle breakdowns in the span of three years. In this model, data points from various parts of the vehicle are collected and fed into the system, which then analyses and predicts future performance\/maintenance pipeline. The effectiveness of the bus fleet can be predicted by anticipating the driver’s KPI (key performance indicator). The system’s data mining algorithm can foretell the behaviour of the driver by predicting variables such as GPS speed, engine RPM, load, throttle positions, etc. With the aid of GPS, the system can also collect information about the vehicle’s path and determine whether the route contributes to vehicle wear and tear (poor road conditions, heavy traffic, etc.). Developments in Bangalore and Delhi Bangalore and Delhi are two of the most congested cities in India and efforts to relieve the congestion have not gone unnoticed. Delhi transport ministry has collaborated with Google Maps to provide real-time information about buses and metros on commuter’s mobile devices. A similar interface can be seen in Bangalore, but there is more. Google has partnered with the Bangalore police to optimise traffic signals, and right in the pilot project they saw a 20% reduction in the wait time for commuters. According to BR Ravikanthe Gowda, joint commissioner of police (traffic), “Putting this proposal into action will probably lessen traffic, wait times, fuel consumption, aggressive driving, and greenhouse gas emissions. Almost one crore vehicles are anticipated to benefit if this is expanded to other junctions throughout the city.”","excerpt":"When it comes to persistent traffic congestion, Ant Colony Optimisation is an effective solution. This model seeks out the most direct route, drawing inspiration from how ants choose the best path when looking for food.","categories":["AI Trends"],"tags":["ai in sports","Predictive AI"],"author_name":"Lokesh Choudhary","publish_date":"2022-09-01T13:00:00","publication_year":"2022","word_count":739,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","Git","R","ai in sports","Predictive AI"],"extracted_tech_keywords":["AI","R","Go","Git","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ant-colony-optimisation-and-predictive-maintenance-in-indian-transport-system\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35143,"title":"9 Paid Data Science Internships All Freshers Must Apply For","content":"Internships are a great way to kickstart an individual’s career. With the right opportunity at the right time, it becomes the starting point of one’s fruitful career. The internships in data science are meant for those who love playing with large datasets and extracting meaningful information and knowledge from them. In this article, we list 9 data science intern-friendly companies in India which have opened positions for interns. 1| Decathlon Company Profile: This is the largest sporting goods retailer in the world. Job Profile: Data Analytics Intern Location: Bengaluru Duration: 6 months Stipend: ₹10,000- 20,000\/ Month Requirements: Assist the innovations team take data-driven decisions for our colleagues in retail and e-commerce Sift through the collected data to obtain key insights required for the growth of the business as well as the user base Knowledge and working on SQL databases Knowledge and working on Python or R platform Apply here. 2| Spoonshot (formerly dishq) Company Profile: Spoonshot (formerly dishq) is a Bengaluru-based artificial intelligence startup that leverages food science and AI to predict consumer tastes and food trends. Job Profile: Data Science Intern Location: Bengaluru Duration: 6 months Stipend: ₹25,000- 40,000 Requirements: Create innovative solutions using data across food science, math, and user data to power the recommendation engine. Build data pipelines using NLP techniques to make data in-warding efficient from various unstructured data sources. Knowledge of Python, Machine learning and Apache Spark Apply here. 3| Euprime Company Profile: This company provides analytics software and statistical consulting in healthcare and other industries. Job Profile: Deep Learning\/ AI Intern Location: Bengaluru Duration: 6 Months Stipend: ₹14,000- 24,000 Requirements: The day-to-day responsibilities include programs in Python, shell, C++ and work on deep learning libraries Build a complete full-fledged model and document the process Knowledge in libraries like OpenCV, OpenPose, Kaldi Apply here. 4| Prizmatics Company Profile: This is artificial intelligence, analytics, and big data based startup started by IIT and IIM alumnus for creating some leading-edge big data analytics products. Job Profile: Data Science Intern Location: Gurgaon Duration: 6 months Stipend: ₹10,000-20,000\/ month Requirements: Day-to-day responsibilities include use, customize and create algorithms for specific tasks in data science Get exposed to and work on cutting-edge products based on ML and AI to create innovative industry solutions Work on real-life projects in analytics and the big data area as well as support development of big data platform Apply here. 5| Payzello Company Profile: This is India’s first ever smart bank that comprises of features and services ranging from a single virtual card for both Debit and Credit to App-based Credit to money transfer and request mechanism, end-to-end expense, etc. Job Profile: Data Science Intern Location: Hyderabad Duration: 6 months Stipend: up to ₹20,000 pm Requirements: Using innovative ideas to collect, curate or synthesize data Model the problem into an ML\/DL framework Build, measure and iterate on neural network architectures that effectively solve the problem Benchmark, validate and test the solution for real-world environments Optimize the solution for accuracy and performance Apply here. 6| FedEx Company Profile: FedEx Express is an eminent transportation company and is the industry’s global leader, providing rapid, reliable, time-definite delivery to more than 220 countries and territories. Job Profile: Data Analytics Intern Location: Mumbai Duration: 6 months Stipend: ₹25,000-35,000 Requirements: Day-to-day responsibilities include Work closely with the pricing analysts and form part of a core pricing analytics and revenue science team Understand the business problem and requirements and translate them into a project plan Conduct data analysis by using SQL, SAS, R, Python, etc. to provide insights and recommendations for business problems Design and build new applications using VBA & MS-Excel Apply here. 7| Blackcoffer Company Profile: Blackcoffer is an enterprise software and analytics consulting firm based in India and the European Union (Malta). It is a data-driven, technology, and decision science firm focused exclusively on big data and analytics, data-driven dashboards, applications development, information management and consulting of any kind, from any source, at massive scale. Job Profile: Data Mining Intern Location: Delhi Duration: 6 months Stipend: up to ₹10,000 Requirements: Day-to-day responsibilities include Mine data from social media, and process it for the analysis and insights Extract data from internets, and process it for analysis Collect data from digital documents such as image, audio, video, pdf, and unstructured file formats Mine data from XML, HTML, web URLs, data repo, and make then analytics ready Work on the database, data integration, the data lake management, and ETL tasks Engage in Python programming, crawling, and analysis Apply here. 8| Voziq India Pvt. Ltd. Company Profile: Vozig offers cloud-based machine learning and big data analytics solutions to quickly uncover breakthrough customer churn intelligence from millions of customer interactions and previous records. Job Profile: Data Science Intern Location: Hyderabad Duration: 6 months Stipend: ₹10,000 pm Requirements: The day-to-day responsibilities include Explore data from various sources and uncover insights Use machine learning, and analytical techniques to create solutions Build predictive models from scratch, train and deploy Measure the performance of models in production and optimise Apply here. 9| Vidooly Company Profile: Vidooly is an online video analytics company based in India. Our software provides video publishers ranking, audience insights, brand safety score, audience overlap and a proprietary video engagement score of over 4.5 million video content creators across the world. Job Profile: ML\/AI Trainee Location: Noida Duration: 6 months Stipend: up to ₹12,000 Requirements: Selected intern’s day-to-day responsibilities include Handle data mining for structured and unstructured data Build and evaluate models for machine\/deep learning Work on image\/object recognition\/detection using deep learning Work on text analysis using natural language processing. Apply here.","excerpt":"Internships are a great way to kickstart an individual’s career. With the right opportunity at the right time, it becomes the starting point of one’s fruitful career. The internships in data science are meant for those who love playing with large datasets and extracting meaningful information and knowledge from them. In this article, we list […]","categories":["AI Trends"],"tags":["audio mining software","career in data science","Data Science","data science internships","spoonshot","Virtual Internship Program"],"author_name":"Ambika Choudhury","publish_date":"2019-02-19T11:25:25","publication_year":"2019","word_count":928,"keywords":["spoonshot","data science","artificial intelligence","machine learning","AI","neural network","ML","OpenCV","NLP","Virtual Internship Program","deep learning","analytics","data science internships","Data Science","audio mining software","career in data science"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","data science","analytics","OpenCV"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/9-paid-data-science-internships-all-freshers-must-apply-for\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001968,"title":"Top 6 Aerospace Startups In India To Watch Out In 2019","content":"In India, the next growth sector is aviation that is seeing a lot of momentum. With India being one of the fastest growing markets, drone\/UAV and aviation companies are capitalising on the growth momentum. In fact, Bangalore has the highest share, 65% and has become the manufacturing hub of India. In this article, we list down top 6 startups in this space that are bolstering India’s aviation industry. 1. Aero 36 Aero provides end-to-end UAV\/drone solutions. They provide integrated solutions for surveying and inspections through the use state-of-the-art Unmanned Aerial Vehicle (UAV) technology for in the areas of construction, infrastructure, real estate, media and agriculture in India. The startup has a training arm called Unmanned Engineeria which provides training in unmanned aerial systems and drones. Their drones are equipped with full HD\/4K cameras, thermal cameras, Lidar sensor, crop sensing devices and other specialized cameras and sensors and are deployed for aerial photography, aerial mapping, aerial inspections and other commercial applications. Location: Chennai Year: 2014 2. Airpix Airpix is a drone startup based in Navi Mumbai, India. It provides data processing and analytics solutions derived from unmanned aerial vehicles or drones and has paying customers across various industries including government planning bodies, mining, oil and gas, railways, roads, agriculture and wind and solar power. By utilising a network of enrolled UAV operators across India and in some cases in-house UAV operation teams, Airpix gathers high quality UAV data and carries out data processing and analytics to generate valuable insights for customers. They capture high resolution aerial imagery in visual, thermal and near infrared spectrum, process it to generate high density point clouds, 2D & 3D outputs like orthophoto, digital surface models, asset inspection reports and more, to generate insights such as planimetric and volumetric analyses, change tracking, defect identification. The startup aims to achieve preventive maintenance through emergency response planning by deploying UAVs for collecting data in situations potentially hazardous to human life. Place: Navi Mumbai Year: 2013 3. Milan Infotech A provider software services and solutions in the areas of not just aerospace, but also automotive, telecommunications, manufacturing, transportation, utilities, government and retail markets. It provides, advanced software training along with the certification on niche skills. This startup helps clients meet their global sourcing needs by identifying, evaluating and configuring their IT architectures across multiple geographies to reduce costs, mitigate risks and maximize performance. Location: Bangalore Year: 2007 4. NavStik Labs NavStik Labs is a startup that works towards building innovative products in the Autonomous Systems space. Last year, the startup came up with an advanced flight computer system and its operating system, which essentially forms the brains of these drones. They have their Flyt platform that comprises of FlytPOD, which is the flight computer and FlytOS, which is the operating system. Additionally, the startup also offers a HackerSpace Internship Program aimed at young researchers with a hacker’s mindset to explore, learn and implement ideas around Intelligent UAVs. Location: Pune 5. Acumen Aviation Acumen offers a full range of services throughout the lifecycle of the aircraft including aircraft sourcing, pre-purchase inspections, project management, lease management and re-marketing, fleet audits, aircraft trading and powerplant management. Its customers include aircraft lessors, airlines, banks, private equity and investors. It has offices in China and US. Place: Bangalore Year: 2009 6. Momentum India Momentum India was established as a consultancy, to fill in the gaps in basic knowledge in safety and security regarding aviation. The plan was to work with established firms and training them on these issues. Apart from that, it provides multiple services like on-site safety audits, crisis and emergency management. It has prominent Aviation Search and Rescue (SAR) services. The startup conducts specialist helicopter operations, and focuses on International Civil Aviation Organization (ICAO) recognized training courses for ambulance, Search and Rescue (SAR) and military clients. It also has a National Search & Rescue Services for various government and resources industry clients, trained and established regional air ambulance services, and consulted to clients on the safe and effective operation of SAR services. Location: Noida","excerpt":"In India, the next growth sector is aviation that is seeing a lot of momentum. With India being one of the fastest growing markets, drone\/UAV and aviation companies are capitalising on the growth momentum. In fact, Bangalore has the highest share, 65% and has become the manufacturing hub of India. In this article, we list […]","categories":["AI Trends"],"tags":["Aerospace","Drone","Startups","UAV","uav companies in india"],"author_name":"Disha Misal","publish_date":"2019-04-30T18:34:58","publication_year":"2019","word_count":672,"keywords":["Go","Drone","API","TPU","AI","UAV","Git","Aim","uav companies in india","Startups","analytics","GAN","Aerospace","R","startup"],"extracted_tech_keywords":["AI","analytics","Aim","TPU","R","Go","Git","API","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-6-aerospace-startups-in-india-to-watch-out-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058787,"title":"Could NFTs change the game for indie filmmakers?","content":"Non-fungible tokens (NFTs) have become quite a runaway hit in 2021. For the movie-making industry in particular, production houses could leverage NFTs to shape a new business model to keep the fans hooked to their brand. Just like the arrival of streaming services effectively brought an end to DVDs (Netflix pivot is a good case in point), NFTs are primed to take the place of collector’s items including director’s cut DVDs, movie posters, ticket stubs etc.The NFT model has a proven track record in movie merchandising. Four years ago, 20th Century Fox released limited edition Deadpool 2 posters to promote the film. Of late, more and more streaming companies are jumping on the NFT bandwagon: Disney released an NFT collection of its most beloved characters (titled “Golden Moments); Warner Bros is releasing digital collectibles inspired by The Matrix franchise; and Lionsgate could have NFTs of mega-successful movies like John Wick and Dirty Dancing in the works. Even the well-known director Quentin Tarantino claims to have plans to auction off seven uncut scenes from Pulp Fiction on OpenSea, the world’s first and largest NFT marketplace. NFTs to democratise Hollywood Niels Juul, the executive producer of Martin Scorsese’s The Irishman, is set to make Hollywood’s first feature film fully funded by NFTs. He has roped in the hero and the director for the film, titled A Wing and a Prayer, and will announce their names at the Berlin Film Festival in February. Juul has set up a production company called NFT Studios to bankroll the project. He plans to raise between USD 8 million to USD 10 million by selling 10,000 NFTs—promising investors a cut of the box-office profits and licensing rights, the opportunity to visit the film’s sets, as well as the option to meet the film’s cast and attend its premiere. “The studios are mainly doing big franchise films, an independent film can take years and years,” said Juul. The idea behind the new funding model is to help small projects get off the ground and trim the spun-out timelines (takes upto 8 years for indie films) to hit the theatres or OTTs. Another first-mover in the indie film industry is Kevin Smith, who announced his decision to auction his horror feature anthology Killroy Was Here as an NFT last year. Smith explained, “Whoever buys it could choose to monetize it traditionally, or simply own a film that nobody ever sees but them.” #ZeroContact will blow your mind. @ZC_Universe pic.twitter.com\/wHbwanE3Ze— VUELE (@vueledigital) December 1, 2021 The NFT distribution and viewing platform Vuele is launching the first NFT feature film, Zero Contact, this year. Vuele is the first platform to deliver feature films and digital collectible entertainment content as NFTs. Oscar-winner Anthony Hopkins stars in the film, shot largely over Zoom during the Covid-19 pandemic in 2020, and tells the story of five people from different parts of the world who join forces to shut down a late tech titan’s most secret invention—a machine that could either solve all of mankind’s problems or trigger apocalypse. The launch of the trailer coincided with a Golden Ticket lottery, including a variety of digital collectibles developed by CurrencyWorks and gives official access to the first four NFT films to be available on the platform. It remains to be seen if NFTs hold the potential to open the cash-strapped indie film world to a new source of capital or end up becoming a short-lived gold rush.","excerpt":"Non-fungible tokens (NFTs) have become quite a runaway hit in 2021. For the movie-making industry in particular, production houses could leverage NFTs to shape a new business model to keep the fans hooked to their brand.  Just like the arrival of streaming services effectively brought an end to DVDs (Netflix pivot is a good case […]","categories":["IT Services"],"tags":[],"author_name":"Srishti Mukherjee","publish_date":"2022-01-23T10:00:00","publication_year":"2022","word_count":571,"keywords":["Go","API","funding","programming_languages:R","AI","Git","RAG","Ray","Aim","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","Git","API","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/could-nfts-change-the-game-for-indie-filmmakers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10117945,"title":"Databricks Launches Data Intelligence Platform for Energy Sector","content":"Databricks, the Data and AI company, announced today the launch of its Data Intelligence Platform for Energy. This unified platform brings the power of AI to data and people across the energy industry. The new offering, built on an open lakehouse architecture, enables energy enterprises to harness vast data streams and develop generative AI applications while maintaining data privacy and IP protection. Industry-leading organisations such as Shell, Octopus Energy, Australian Energy Market Operator (AEMO), and Chevron Phillips Chemical have already adopted Databricks to accelerate their data analytics and AI capabilities. The platform has helped them unlock real-time insights, drive strategic decisions, and create process improvements, cost reductions, and production increases. The Data Intelligence Platform for Energy offers pre-built accelerators, marketplace solutions, and an ecosystem of partner capabilities tailored to the energy industry. It enables companies to tackle critical challenges such as real-time asset performance management, accurate renewable energy forecasting, grid optimisation, and more. Databricks partners, including AVEVA, Capgemini, and Deloitte, are also driving the Data Intelligence Platform vision by delivering pre-built analytics solutions on the lakehouse architecture that are custom-made for the energy industry. The energy sector is undergoing a paradigm shift toward a smarter, cleaner, and more reliable energy system, with renewables now providing nearly 30% of global power. Databricks’ platform democratises data access across organisations, allowing them to optimise energy infrastructure and mitigate volatility by leveraging the full value of asset, operations, environmental, and customer data.","excerpt":"Organisations like Shell, Octopus Energy, Australian Energy Market Operator (AEMO), and Chevron Phillips Chemical are some early adopters.","categories":["AI News"],"tags":["Databricks"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-04-10T11:00:32","publication_year":"2024","word_count":238,"keywords":["Go","programming_languages:R","AI","RAG","llm_models:Gemini","generative AI","analytics","GAN","R","Databricks"],"extracted_tech_keywords":["AI","analytics","generative AI","RAG","Databricks","R","Go","GAN","llm_models:Gemini","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/databricks-launches-data-intelligence-platform-for-energy-sector\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10008396,"title":"What Is Code Golfing And Biggest Such Tournaments","content":"Code Golf is a game that is designed to let programmers show off their excellency in codes by solving problems in the least number of characters. The word “Golf” in code golfing refers to the popular game golf where two players compete with each other, and the one with the fewest club strokes wins. Similar to the golf game, code golf is a competition where the winner achieves the specifications in the fewest keystrokes. It is basically a kind of recreational computer programming competition where the participants compete to achieve the shortest possible source code that implements a certain algorithm. Code Golfing can be said as a classic playground for programmers where the main attempt is to solve a problem with the least number of characters. It is written in Go language, licensed under MIT and is available on GitHub. Code Golf is free and supports various programming languages including Python, Haskell, JavaScript, Julia, Perl, Rust, Swift, Ruby and more. Currently, it provides a number of games including 12 days of Christmas, Abundant numbers, Fizz Buzz, 99 Bottles of Beer, Diamonds, Evil Numbers and much more. If you are playing for the first time, it has been suggested to start with a simple game, such as Fizz Buzz. How Code Golfing Works Below here, we listed general tips of Code Golf that are implemented in popular languages like Python and others. Conditional Statement: Original- if a<b:return a else:return b Code Golf- return(b,a)[a<b] AND Operators Original- if a > 1 and b > 1 and 3 > a and 5 > b: foo() Code Golf- if 3 > a > 1 < b < 5: foo() Multiple Statements Original- while foo(a): print a;a*=2 Code Golf- while foo(a):print a;a*=2 Replacing Append Original- A.append(B) Code Golf- A+=B, Ceil Value of a Real Number Original- from math import ceil n = 3\/2 print(ceil(n)) Code Golf- n = 3\/2 print(-(-n\/\/1)) How Does The Score Work? The score of your solution is the count of the Unicode characters in your source code. This means both “A” (U+0041 Latin Capital Letter A) and “????” (U+1F609 Winking Face) cost the same despite the 1:4 ratio in byte count in UTF-8. For each hole, the shortest solution is awarded 1,000 points, with the points decreasing in uniform decrements per rank. Your overall score is simply the sum of your points in each hole. Also, the execution time is limited to 5 seconds. Pros and Cons of Code Golf Pros Code Golfing can be used for various purposes, such as- Practising programming with the help of Code GolfSharpening skills in codingLearning to optimise codesCan be used as a learning tool and experienceTo think something out of the box Cons The codes in Code Golf are often difficult to understandUnlike usual programming codes, the codes of Code Golf are merely impossible to be modified. Other Similar Tournaments The International Obfuscated C Code Contest The International Obfuscated C Code Contest (abbreviated IOCCC) is a computer programming contest for the most creatively obfuscated C code. In this contest, the programmers are asked to write the most obscure and obfuscated C program within the rules, stress C compilers with unusual code, etc. During the competition, the programmers usually write the smallest C compiler which is able to compile itself, avoid the commonly used constructs, avoid the C preprocessor and tricky statements such as “if”, “for”, “do”, “while” and other such. Know more here. js13kGames: Code golf for game devs Js13kGames is an annual online JavaScript competition for HTML5 game developers that began in 2012. The annual js13kGames competition channels creativity by challenging aspiring game developers to create a game using just 13,312 bytes, zipped. Some of the tricks that the programmers use in this competition include inlining all images, concatenate images, minimising the JS codes, zipping the final entries, and other such. Know more here.","excerpt":"Code Golf is a game that is designed to let programmers show off their excellency in codes by solving problems in the least number of characters. The word “Golf” in code golfing refers to the popular game golf where two players compete with each other, and the one with the fewest club strokes wins. Similar […]","categories":["Deep Tech"],"tags":["big data developer skills","coding","programming","what is big data coding"],"author_name":"Ambika Choudhury","publish_date":"2020-09-27T13:00:00","publication_year":"2020","word_count":641,"keywords":["Go","Rust","big data developer skills","AI","programming","coding","ML","R","Julia","Git","Python","JavaScript","what is big data coding","Java"],"extracted_tech_keywords":["AI","ML","Python","R","JavaScript","Go","Rust","Java","Julia","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/what-is-code-golfing-and-biggest-such-tournaments\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168198,"title":"Tech Data Partners with NVIDIA to Distribute AI Solutions in India","content":"Tech Data Advanced Private Limited, a TD SYNNEX company, signed a new distribution agreement with NVIDIA on Wednesday to offer its data centre and enterprise AI solutions in India. The partnership will give Indian businesses access to NVIDIA’s full-stack AI tools, including its advanced data centre GPUs, making it easier to adopt and deploy AI across different industries. “Complementing NVIDIA’s advanced AI solutions with our homegrown offerings and a robust partner ecosystem, this collaboration will enable us to help Indian businesses bridge infrastructure gaps, scale AI initiatives, and maximise innovation across the IT ecosystem,” Sundaresan K, vice president and country general manager at Tech Data said. NVIDIA’s platform includes a suite of tools and software that support developers in building AI-powered applications. Among these are the CUDA-X libraries, designed to speed up data processing and machine learning, and NVIDIA NIM microservices, which help developers build AI assistants and productivity plug-ins faster. Tech Data plans to use its network of local partners to support customers, simplify AI adoption, and help bring NVIDIA’s solutions to market quickly. The company will also tap into recent partnerships, such as its collaboration with Dell’s AI Factory and several software vendors, to open up more opportunities for AI in India. Tech Data already works with NVIDIA in Europe and North America. Expanding the partnership to India is part of its ongoing push to make AI more accessible through its Destination AI programme. Meanwhile, at GTC 2025, NVIDIA unveiled two new personal AI supercomputers—the DGX Spark and DGX Station—designed to tackle demanding AI workloads. Powered by the NVIDIA Grace Blackwell platform, these systems are built to support AI developers, researchers, data scientists, and students in prototyping, fine-tuning, and running inference on large language models right from their desktops. The models can be run locally or seamlessly deployed to any cloud platform.","excerpt":"NVIDIA’s platform includes a suite of tools and software that support developers in building AI-powered applications.","categories":["AI News"],"tags":["NVIDIA"],"author_name":"Shalini Mondal","publish_date":"2025-04-17T13:41:18","publication_year":"2025","word_count":304,"keywords":["CUDA","Go","AI assistants","machine learning","AI","ML","microservices","ViT","NVIDIA","R"],"extracted_tech_keywords":["AI","machine learning","ML","AI assistants","microservices","CUDA","R","Go","CUDA","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tech-data-partners-with-nvidia-to-distribute-ai-solutions-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60572,"title":"How To Get Your Child Interested In Data Science","content":"It is a known fact that data science as a career has emerged as one of the top choices for anyone looking to choose a stream or switch careers. In all likelihood, this individual is fully aware that the field has been facing a deficit when it comes to talent, despite the demand. Governments have been taking several initiatives – and so have many organizations – to include data science in the middle school curriculum. This move could familiarize students with key data science concepts like statistics, coding and visualization, among others. But, before that, they have to first develop an interest in their fundamentals. Below, we have tried to list down some tasks that we, as adults, can do to help children develop an interest in this promising field: Encourage Them To Work With Numbers More It is safe to assume that most children do not enjoy math, especially algebra. But when it comes to data science, some algebra and a lot of statistics is a major requirement. With that being said, how do you ensure that your children grow up to love crunching numbers? There are various ways: Get them interested in some fun, relatable activity: The best example would be to give them assignments related to statistics using social media websites. Let them use free online tools and apps to poll, which could pique their interest in statistics. Use statistics-related games and let them watch some fun online videos: This is a standard method when it comes to teaching children something seemingly uninteresting. There are many statistics-related games online and activities for kids to engage. It would also prove useful if they use online interactive graphs, pie charts and calculators. Along with games, there are many fun videos on the subject that they can choose. Demonstrate Applications Of Data Science Data science is applied everywhere, in all aspects of our lives. Making children aware of its application in logistics and insurance would not be helpful. Instead, show them how it is used in Netflix and Instagram to get them interested. Children avoid taking part in activities unless they can relate to it. So, when you tell them how Instagram uses data and insights to make their experience on the platform better, they will be interested in how it works. Get Them Familiarized With Coding Coding is one of the essential aspects of data science. Getting your children excited about coding is as important as getting them to know statistics better. Firstly, when it comes to explaining what coding is, use a creative process that will encourage them to explore it further. When children start seeing coding as a creative process, the next step is to make it a part of their daily activity. Introduce Children To Online Visualization Tools In today’s world, data visualization needs to be clear, interactive and colorful, since data insights need to communicate easily to anyone interested in the results that are to be drawn from them. Introducing online tools for creating visualization and infographics can be one of the fun ways to get children interested in using tools. These infographics – which are colourful and interactive – are almost as good as videos. These can be used to demonstrate the power and importance of data in the age of big data. Many data science professionals often prefer valuable and high-quality infographics over plain content. To get young people even more interested in this, you can give them exposure to free data into visualization transformation tools which are available online. Final Thoughts As the demand for talented data scientists increases along with the talent gap, we have to take definitive measures to bridge it in the coming future. It is a fact that the massive data flow will never end and that means more insights and more areas for companies to grow. As mentioned earlier, many initiatives have been taken by governments around the world (along with other organisations) to introduce data science in the school curriculum. Although it will take more time before we can see real results of these initiatives, familiarizing students with data science concepts from an early age seems promising.","excerpt":"It is a known fact that data science as a career has emerged as one of the top choices for anyone looking to choose a stream or switch careers. In all likelihood, this individual is fully aware that the field has been facing a deficit when it comes to talent, despite the demand. Governments have […]","categories":["AI Features"],"tags":["big data for social good","big data video games","TED Talks"],"author_name":"Sameer Balaganur","publish_date":"2020-03-31T20:00:00","publication_year":"2020","word_count":691,"keywords":["big data","data science","Go","programming_languages:R","AI","big data for social good","programming_languages:Go","TED Talks","RAG","GAN","ViT","big data video games","R"],"extracted_tech_keywords":["AI","data science","RAG","R","Go","big data","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-get-your-child-interested-in-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10005362,"title":"How to Implement Bitwise Operations On Images Using OpenCV?","content":"OpenCV is an image processing library created by intel which includes various packages and several functions in it. OpenCV is used to solve many problems in computer vision and machine learning applications and due to its large community, it is getting updated day by day. OpenCV can be implemented in C++, Python, Java programming languages, and different platforms like Linux, Windows, macOS. In this article, we will demonstrate one of the interesting applications of OpenCV in performing bitwise operations on images. Topics covered in this article Bitwise Operators in Computer VisionBitwise ANDBitwise ORBitwise NOTBitwise XOR Bitwise Operators in Computer Vision Bitwise operations can be used in image manipulations. These bitwise techniques are used in many computer vision applications like for creating masks of the image, adding watermarks to the image and it is possible to create a new image using these bitwise operators. These operations work on the individual pixels in the image to give accurate results compared with other morphing techniques in OpenCV. Using the below code snippet, we will create two images –  image1, image2 as input images on which the bitwise operations will be performed. import numpy as np import cv2 from google.colab.patches import cv2_imshow image1 = np.zeros((400, 400), dtype=\"uint8\") cv2.rectangle(image1, (100, 100), (250, 250), 255, -1) cv2_imshow(image1) image2 = np.zeros((400, 400), dtype=\"uint8\") cv2.circle(image2, (150, 150), 90, 255, -1) cv2_imshow(image2) Bitwise example images Bitwise example images In the above output, we created a square & circle with white pixels and a background with black pixels using these we will differentiate each bitwise function individually. Now we will start performing the bitwise operations. Bitwise AND This function calculates the conjunction of pixels in both images. This operation only considers pixels that are common with image 1 and image 2 and remaining pixels are removed from the output image. bit-and = cv2.bitwise_and(img1,img2) cv2_imshow(bit-and) Bitwise AND operations In the above output, by using the AND function only the intersected regions in both images are displayed. Bitwise OR This function calculates the disjunction of the pixels in both images. Here we perform an element-wise product of the array, we will not eliminate pixels but it merges both images. bit-or = cv2.bitwise_or(img1,img2) cv2_imshow(bit-or) Bitwise OR operations In the above output, we merged both images using the OR function Bitwise NOT This function inverts every bit of an array. It replaces the white pixels with black pixels and vice versa. This operation can be performed only on one image. bit-not = cv2.bitwise_not(img1) cv2_imshow(bit-not) Bitwise not operations In the above execution, we have given input as a circle with white pixels and background with black pixels by using the NOT function to replace black pixels with white and vice versa. Bitwise XOR This function inverts the pixels which are intersected in image 1 & image 2 and the rest of the pixels remains the same. bit-xor = cv2.bitwise_xor(img1,img2) cv2_imshow(bit-xor) Bitwise XOR operations As we can see in the above output, by using the XOR function it removes the intersected region, and the remaining pixels are displayed. Conclusion In this article, we demonstrated that bitwise functions are used for image manipulations using OpenCV. We illustrated the basic difference between each function. These operations are used in many computer vision real-time applications. But the main drawback of this bitwise function is that we cannot perform these operations with more than images.","excerpt":"OpenCV is an image processing library created by intel which includes various packages and several functions in it. OpenCV is used to solve many problems in computer vision and machine learning applications and due to its large community, it is getting updated day by day. OpenCV can be implemented in C++, Python, Java programming languages, […]","categories":["Deep Tech"],"tags":["Opencv image processing"],"author_name":"Prudhvi varma","publish_date":"2020-08-23T13:00:10","publication_year":"2020","word_count":555,"keywords":["Opencv image processing","NumPy","machine learning","TPU","AI","computer vision","OpenCV","Colab","Ray","Python","R"],"extracted_tech_keywords":["AI","machine learning","computer vision","Ray","Colab","OpenCV","NumPy","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-implement-bitwise-operations-on-images-using-opencv\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":43365,"title":"8 Ways Companies Can Incentivise Data Scientists To Be More Productive","content":"Data Science is a vast domain and sometimes the workload can make a data scientist feel stressed. Even though companies provide bonuses, paid holidays and other formal benefits, sometimes these might not be enough to make an employee feel motivated. This is where incentivising comes into the picture, which is all about getting the most out of a mutually beneficial relationship. In this article, we shall look at some of the employee incentivising tips that companies can use to bring out the best performance and make data scientists more productive. Data Scientists Love Recognition The job of a Data Scientist is not easy — it takes a tremendous amount of time and effort. There are instances when a data scientist ends up staying for way too long at work and sometimes, they even have to work when they go back home. Companies must make sure they recognise even the smallest achievements and appreciate them — it would not only boost the employee self-esteem and but also confidence, and would drive them to perform better. It is definitely a plus point when a company has Data Scientists who are motivated and are ready to take huge efforts not only for the company but for their career as well. Provide The Ecosystem Data Scientists are very concerned about the ecosystem they work in and why not, they solve some of the most complex business problems. So, when your company has a data science team, make sure you provide all the necessary tools and resources that are required to do the tasks. There are Data Scientists who might lack a particular skill and s\/he wants to learn that, the company can always provide them with the opportunity to learn and master that by providing training or courses. It is considered to be a good practice to have conversations with the data scientist in order to know if there is anything they would like to learn or upskill. Work Flexibility Irrespective of the work domain, work flexibility is something every professional seeks. Sometimes people tend to be more productive when they work from home. Therefore, when it comes to incentivising employees, work flexibility should also be one of the major things to consider. Monetary Benefits Aren’t Everything Studies have showcased that benefits in forms other than hard cash evoke a sense of joy and happiness in the human brain. That is why, instead of giving monetary incentives, companies should sometimes try to provide benefits like gift cards, food cards, or even movie tickets, after doing a background search on the data scientist’s likes and dislikes. This kind of gestures drive a sense of good feeling and employees are able to deal with the blues. Have An Open Door Policy The 9 to 5 job of a data science professional tends to be busy — they would rarely get a significant amount of time to deal with other stuff. And this where the open door policy comes into the scenario. When a data science professional faces an issue or if there is any query, they might not like the fact of following a hierarchy to get the issue to solve. If a company has the culture of the open door, it gives a data scientist the sense of little relief to reach out to any senior person and discuss its concerns. Allot Projects That Are Relevant And Significant A data scientist spends a lot of his\/her time on studying and gaining experience, and if they are not assigned some significant projects to work on, it would hit their mindset in a negative manner. It is okay for companies to let the professionals work on smaller projects, but after a time span, if the professionals are fit enough for a big project, make sure you give him\/her that opportunity to show their potential. Another great thing about providing such opportunities is that the data scientist would feel his\/her importance at the workplace that would ultimately drive productivity. Be Transparent About The Company’s Operations There are many companies across the globe who prefer to keep a significant part of the company operations hidden from its employees, which is not at all a good practice. When you have a data science team that is working day in and day out to solve your business problems, delivering some of the most meaningful insights out of cluttered data, it is imperative for a company to be transparent. If you don’t let them know what the actual business or operation is, you would not get the exact solution. And being transparent to your employees is also a form of incentivising. Have A Pleasant Work Environment Nobody wants to work in a toxic environment where everyone just works for the weekend. Organisations can take several steps to have that environment such as relaxation activities, off-work discussion, hobbies, sports, etc. Furthermore, it’s not just about providing facilities but also about keeping the workplace clean and tidy. Have those little touches that make the difference.","excerpt":"Data Science is a vast domain and sometimes the workload can make a data scientist feel stressed. Even though companies provide bonuses, paid holidays and other formal benefits, sometimes these might not be enough to make an employee feel motivated. This is where incentivising comes into the picture, which is all about getting the most […]","categories":["AI Trends"],"tags":["AI Workplace","Data Science","Data Scientist","freelancer","Productivity"],"author_name":"Harshajit Sarmah","publish_date":"2019-07-26T11:00:18","publication_year":"2019","word_count":831,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","AI Workplace","ViT","GAN","Data Science","Data Scientist","R","freelancer","Productivity"],"extracted_tech_keywords":["AI","data science","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-ways-companies-can-incentivise-data-scientists-to-be-more-productive\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10113067,"title":"VFX Industry Will Make Their Own ‘Sora-Like’ GenAI Tools","content":"OpenAI just released Sora, a video-generation AI tool that creates hyper-realistic videos from prompts. While Sora is impressive, it is sure to add to the mounting concerns on potential job losses for VFX artists owing to the rise of generative AI. Last year, the International Alliance of Theatrical Stage Employees, the overarching organisation pioneering the inaugural VFX union, advised Hollywood to exercise caution in utilising AI within the industry. Similarly, a major VFX studio owned by Netflix last year hired AI experts, raising a few eyebrows. The studio was looking for individuals with a comprehensive understanding of prompt engineering, experience in neural network image-generating space, and knowledge of the stable diffusion ecosystem, among other things. A recent report titled ‘FUTURE UNSCRIPTED: The Impact of Generative Artificial Intelligence on Entertainment Industry Jobs‘ underscores that generative AI is poised to instigate a substantial shift from conventional techniques to innovative processes. This transition is expected to recalibrate the demand for labour and capital across the entertainment industries. However, industry experts AIM spoke to hold a rather different view. “When ‘The Lion King’ was originally created through hand-drawn animation 35 years ago, the production process significantly differed from the recent remake with Disney and Jon Favreau. In the modern iteration, we employed virtual reality (VR) to finalise the entire film. Subsequently, we immersed ourselves in animating it with photorealistic precision. The current production of ‘The Lion King’ utilised an array of advanced tools and technologies, involving a larger team compared to the production of the original film several years ago,” Biren Ghose, managing director – Asia Pacific, Technicolor Creative Studios., told AIM. GenAI will make things easier Other industry experts also concurred with Ghose’s views. It is not just the VFX industry, but since generative AI caught the world’s attention with the launch of ChatGPT, concerns about jobs became widespread in many industries, ranging from engineering and programming to many creative fields such as designing. Nonetheless, experts across various industries share a common perspective that generative AI presently lacks the capacity to fully replace humans. Rather, it functions as a tool to augment human capabilities. It holds true for the VFX industry as well. Moreover, akin to past technological revolutions, AI is anticipated to reshape the job landscape, opening up new opportunities and avenues for employment. Notably, even in the VFX industry, the companies have been building proprietary tools for quite some time. Ghose explains that different projects have different requirements, and in some cases, VFX companies develop their own proprietary AI algorithms to support them in that particular project. Ghose believes that generative AI has the power to improve these existing tools and functions. For instance, “Imagine a landscape featuring a mix of greenery and buildings. When I capture this scene with the camera, visual effects today can enhance and extend the shot. With AI advancements, certain extensions, like modifying the skyline, may become more streamlined.” GenAI can’t do VFX alone However, to imagine that an AI model can do the work of a VFX artist or team is far-fetched. “In the next two years, AI’s capability to serve as a reference or an accelerator, expediting tasks in the realm of imprecise outputs, is expected to enhance productivity significantly. However, achieving the finesse and intricate detailing anticipated in the next one or two years, at least, may still be beyond the reach of AI,” Ghose said. Other experts whom AIM spoke to on the sidelines of the Bengaluru GAFX 2024 did appear impressed by the quality of the text-to-image AI tools like Midjourney or Stable Diffusion or video-generation tools like Lumiere by Google. However, they, too, believe these tools still do not possess the calibre to support a VFX project independently. Moreover, another general consensus among experts is that AI currently lacks the logical reasoning and real-world knowledge that artists rely on for creative decision-making and problem-solving. “When it comes to high-end artistry like what we do, in the short term, the human capability will not be replaced; however, in the long term, nobody knows,” Ghose remarked. Copyright issues pertain Ghose also adds that the quality of these tools depends on the data they are trained on. Copyright and AI are still ongoing issues, and many artists in the US and, more famously, The New York Times have sued OpenAI for using their content without permission to train AI models. Though Ghose believes generative AI will be a game-changer, chances are high that they will develop their own proprietary video generation model by training it with their own rich enterprise data. “Large organisations like ours, with extensive experience and a wealth of expertise gained from numerous endeavours, will channel this knowledge into proprietary tools tailored to our specific needs. The nature of our work with esteemed clients prohibits us from feeding data into publicly available generative AI tools,” Ghose said.","excerpt":"VFX companies have been developing proprietary AI tools for a while","categories":["AI Trends"],"tags":["ChatGPT","Generative AI","OpenAI","VFX"],"author_name":"Pritam Bordoloi","publish_date":"2024-02-16T13:00:00","publication_year":"2024","word_count":807,"keywords":["ChatGPT","GenAI","artificial intelligence","OpenAI","AI","neural network","ML","VFX","Ray","Aim","generative AI","Generative AI"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","generative AI","GenAI","ChatGPT","OpenAI","Aim","Ray"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/vfx-industry-will-make-their-own-sora-like-genai-tools\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10088173,"title":"Data Science Hiring and Interview Process at Livspace","content":"Founded in 2014 by Anuj Srivastava and Ramakant Sharma, Livspace has become a popular name in the interior design space. Backed by Goldman Sachs, IKEA, Jungle Ventures, Bessemer Ventures, Helion Venture Partners, among others, the unicorn is spread across multiple cities in India and Singapore. Livspace is a platform that facilitates home renovation by connecting interior designers, vendors, and homeowners. Customers can choose from a vast selection of home interior designs that match their preferences and tastes. Livspace’s design professionals assist customers in finalising their design choices. Additionally, Livspace offers complete project management services, including site inspections, installation, and post-installation support. With partnerships with over 20,000 suppliers and vendors, Livspace provides a broad range of products to its customers. In 2016, the company introduced ‘Canvas’, a design-to-installation platform that revolutionised the interior design industry by providing a unique experience for homeowners while also enhancing efficiency for interior designers. AIM got in touch with Abhishek Kasina, chief product officer at Livspace to know more about their AI and analytics play. The IIT-Kanpur and London Business School alumni has led the conceptualisation and development of the core technology platform. He currently leads the evolution of the global product portfolio. Read more: Data Science Hiring Process at McAfee In Expansion Mode The data science team is in charge of creating image-processing models that assist Livspace designers in producing cutting-edge homes for their clients. At present, the company has a few open positions for data scientists with more than four years of experience. The job location is Bengaluru, Karnataka, but the working arrangement will be hybrid. Inside Livspace’s Tech Team With a bandwidth of around 150 employees, Livspace’s engineering team is broken down into specific product and business areas. “Instead of just integrating AI because it exists, we evaluate whether it makes sense to leverage it to provide a better experience for consumers,” said Kasina. He further emphasised the importance of starting with a problem statement rather than starting with a tool and evaluating everything holistically to find the best way to reach a solution. The company focuses on democratising data and making more data-driven decisions. They have a team dedicated to this task, consisting of three sub-teams: data science, product analytics, and business analytics. The data science team aggregates data from various platforms, converts it into consumable datasets, and makes it available for use by the remaining two teams. The product analytics team works closely to identify opportunities for improving the platform, while the business analytics team works with sales and other functions to improve the company’s performance. Despite the challenges due to the complexity of the transaction of home interiors, the team’s collective efforts have led to a healthy combination of leveraging innovative technology such as AI, business process updates, and data-driven projects in their product roadmap. Livspace hosts most of its data on AWS and uses Amazon‘s QuickSight analytics platform to create dashboards. The data science team works on accessing and transforming raw data using Python libraries such as NumPy and SciPy. However, for larger datasets, they are leveraging more structured frameworks such as TensorFlow along with PyTorch for tasks such as classifying designs and tagging them for easy access. The Interview Process When evaluating a candidate for a position, their assessment will be tailored to the specific role they are seeking. In the initial rounds, the emphasis will be on assessing their fundamental technical skills and their proficiency in problem-solving. In the subsequent stages, the focus will shift to assessing how they can apply these skills to address practical business issues. Being able to work effectively with data and utilise the appropriate tools is essential, even if the candidate lacks fluency in a specific programming language. Ultimately, it’s crucial for candidates to be able to translate business context into actionable insights. “When we look to hire for our technology team, we try to understand if the candidate is a first principles thinker and a problem solver”, said Kasina, while explaining what traits he seeks in a candidate during interviews. To begin with, applicants need to be comfortable with data and numbers. They need to understand the core framework and depending upon the task, the company will upskill them. It does not matter which educational institution the candidate is from or how many years of work experience they have, or any such detail for that matter as long as they are a problem-solver and critical thinker with an eye for detail. Being comfortable with the tech tools that Livspace uses is favourable but not mandatory. Dos and Don’ts Kasina said that he has noticed that during interviews, people often rush to provide quick solutions, thinking it will benefit them. However, it is recommended that candidates take time to ask follow-up questions to understand the problem they are solving and why they are solving it. If a candidate cannot articulate their question and spend time sharpening the problem, it is unlikely to happen in a real-world scenario. Therefore, candidates must have the intuition to investigate the problem further, even challenging the relevance of the question at times. It is crucial not to jump into providing an answer without fully understanding the problem because incomplete information may lead to ineffective solutions. Expectations Hiring is a two-way street and it starts with clearly understanding the candidate’s expectations for the next two to five years and how the company can assist in fulfilling them. “Everybody comes with a different set of motivations, but it’s very, very important that we as a company are able to satisfy the aspirations of those candidates for joining us,” said Kasina, highlighting the importance of valuing candidates’ goals. As an employee, the company expects the candidates to take complete ownership and accountability of their work, while providing them with a high degree of autonomy. Ownership and accountability should be balanced with autonomy to deliver high impact results. Livspace focuses on eliminating the fear of failure as long as there is learning gleaned from them. Work Culture Kasina mentioned that they have been fortunate enough to evolve their work culture in the right direction over the last few years, while simultaneously adapting to the changes that come with expansion such as different cultural nuances across geographies. The company offers competitive cash compensation based on the role of the candidate. Additionally, they provide an ESOP programme and benefits like health insurance, maternity and paternity leave, sabbatical, bereavement leave, flexible work hours, crèche facilities and more. Livspace believes in open communication among its employees. Kasina further said, “We have an organisational hierarchy for personalised attention and feedback, but it shouldn’t hold people back or limit access. We are an extremely accessible organisation that starts all the way from our founders.” There are frequent AMAs between the team lead and the members on a monthly basis besides open calendar hours, guest sessions, learning platforms for employee growth and so on. “The kind of freedom that we give people to grow and learn and understand themselves and apply that at work is very unique to Livspace. So, if people come here, they can expect to be challenged but they can come through on those challenges as well,” said Nafeesa Tasneem, head of corporate communications, Livspace. Work culture is beyond what is mentioned on documents or the website. Livspace believes that shaping the culture of an organisation is an ongoing process and the importance of a positive work culture has been a cornerstone since the company’s inception. Check out their careers page here.","excerpt":"Livspace is hiring 2-3 data scientists to join their dynamic team.","categories":["AI Hirings"],"tags":["Career","Data Science Hiring","Top Trend"],"author_name":"Shritama Saha","publish_date":"2024-07-29T14:56:43","publication_year":"2024","word_count":1247,"keywords":["data science","Top Trend","NumPy","AWS","AI","PyTorch","RAG","Python","Data Science Hiring","Aim","analytics","TensorFlow","Career"],"extracted_tech_keywords":["AI","data science","analytics","Aim","TensorFlow","PyTorch","NumPy","RAG","AWS","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/data-science-hiring-process-at-livspace\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10170580,"title":"OpenAI, G42, Oracle, NVIDIA and Others Launch Stargate UAE for AI Infrastructure","content":"A group of global technology firms, including G42, OpenAI, Oracle, NVIDIA, SoftBank Group, and Cisco, have announced the launch of Stargate UAE, a 1-gigawatt AI compute cluster that will operate within the newly established UAE–US AI Campus in Abu Dhabi. The project is part of a broader initiative under the “US–UAE AI Acceleration Partnership” to strengthen collaboration between the two countries on artificial intelligence and digital infrastructure. Stargate UAE will be developed by G42 and operated jointly by OpenAI and Oracle. Cisco will provide AI-ready connectivity and security infrastructure, while NVIDIA will supply its Grace Blackwell GB300 systems. The first phase, a 200-megawatt AI cluster, is expected to go live in 2026. “The launch of Stargate UAE is a significant step in the UAE–U.S. AI partnership,” said Peng Xiao, Group CEO of G42. “This initiative is about building a bridge – rooted in trust and ambition – that helps bring the benefits of AI to economies, societies, and people around the world.” The broader UAE–U.S. AI Campus will provide 5 gigawatts of data centre capacity across 10 square miles, making it the largest AI infrastructure deployment outside the United States. It will be powered by a mix of nuclear, solar, and natural gas energy. The site will also host a science park to support innovation and talent development. Sam Altman, CEO of OpenAI, said, “By establishing the world’s first Stargate outside of the U.S. in the UAE, we’re transforming a bold vision into reality. This is the first major milestone in our OpenAI for Countries initiative.” “Stargate pairs Oracle’s AI-optimised cloud with nation-scale sovereign infrastructure. This platform will enable every UAE government agency and commercial institution to connect their data to the world’s most advanced AI models,” said Larry Ellison, chairman and CTO of Oracle. The initiative also supports reciprocal investment, with UAE entities expected to fund digital infrastructure in the United States under the “America First Investment Policy.” Jensen Huang, CEO of NVIDIA, said, “With Stargate UAE, we are building the AI infrastructure to power the country’s vision – to empower its people, grow its economy, and shape its future.” SoftBank CEO Masayoshi Son added, “The UAE becomes the first nation beyond America to embrace this sovereign AI platform. SoftBank is proud to support the UAE’s leap forward.” Chuck Robbins, CEO of Cisco, said, “By embedding our secure AI-optimised networking fabric, we’re building smart, secure and energy efficient networks that will turn intelligence into impact at global scale.” The announcement follows last week’s unveiling of the UAE–U.S. AI Campus in Abu Dhabi, attended by UAE President Sheikh Mohamed bin Zayed Al Nahyan and U.S. President Donald J. Trump.","excerpt":"A group of global technology firms, including G42, OpenAI, Oracle, NVIDIA, SoftBank Group, and Cisco, have announced the launch of Stargate UAE, a 1-gigawatt AI compute cluster that will operate within the newly established UAE–US AI Campus in Abu Dhabi.  The project is part of a broader initiative under the “US–UAE AI Acceleration Partnership” to […]","categories":["AI News"],"tags":["OpenAI","Oracle"],"author_name":"Siddharth Jindal","publish_date":"2025-05-23T12:45:58","publication_year":"2025","word_count":439,"keywords":["Go","artificial intelligence","OpenAI","AI","programming_languages:R","innovation","Git","Oracle","ViT","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","OpenAI","R","Go","Rust","Git","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-g42-oracle-nvidia-and-others-launch-stargate-uae-for-ai-infrastructure\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10055150,"title":"Council Post: Experiential Learning—An Essence To Address The Skill Gap In The Field Of Analytics And Data Science","content":"Method of teaching has been a subject of discussion and debate for a long time. The effectiveness of training and skilling has been questioned and deliberated time and again globally, irrespective of the field, stream, sector, or specialisation. The education sector, of late, has been witnessing a move away from traditional teaching techniques. Rote learning is slowly becoming expendable, especially in practical fields like analytics and data science. The pace at which these fields are evolving coupled with the rapidly increasing demand globally across all industry verticals has created a significant gap in the right talent supply with the skillset to apply themselves for a given business context and create impact. While the academic institutes, MOOCs, and the likes are doing a tremendous job in creating awareness and equipping the talent with theoretical concepts and knowledge, there is a gap that is widening around enabling the talent with the right experience to be impactful on-the-job quickly. The rising attrition of experienced talent is adding to the pressure on the system. There is no doubt that theoretical learning is foundational for analysts and data scientists, but the work entails individuals to critically understand business problems and create innovative solutions. This demands them to be continuous prolific learners, creative thinkers, and quick problem solvers. The way to achieve these desired qualities is through learning by experience. The learning-by-doing method allows learners to be engaged and actively participate in the learning process by working and reflecting on the projects done. This form of learning is proving to be the most effective in becoming successful in the analytics and data science landscape. We will discuss how one can and should upskill oneself through experiential learning in analytics and data science. The Foundations Before delving into the essence of experiential learning, there are two fundamental concepts that every aspiring analyst and data scientist must ingrain themselves with to become successful in the field. 1. Do not believe data without reasoning Data is the basis for your trendsetting, analysis, prediction, and business solutions. If the data is faulty, the entire project will fail. One must question the existence of the data and reason with the data to ensure its validity and quality before moving on to any other step. For instance, last year, Italy had the highest number of COVID-19 deaths at one point. But a part of this situation owed itself to every death in an Italian COVID-19 hospital being counted as a COVID-19 death, regardless of the real reason. If one were to base one’s predictions and trends on just the former statement, the results would be faulty. 2. Do not arrive at conclusions without critically examining the data Complementing data reasoning this step entails examining the data and its correlation to causation. Go a step further into ensuring that the claims made by the data are backed by facts and information. For instance, citizens in the UK shop more during winter than summer. At face value, this proves seasonal consumer preferences, but in reality, winter coincides with Christmas and New Year sales, pushing customers to go on shopping sprees. Basing your analysis on the first statement would lead to an incorrect business solution. The most fundamental aspect at which all three streams of analytics, “descriptive”, “predictive”, or “prescriptive” are built on, is the clarity around “correlation” v\/s “causation”. Several of the analytics and data science applications fail to address business problems due to a lack of critical examination leading to the faulty judgment of interchanging correlation with causation and vice versa. Forms of Learning The methods of teaching and learning are undergoing a significant change in the modern era. The traditional classroom approach, based on the foundations of listening to lectures and reading out of textbooks, is not proving to be successful in readying professionals for today’s workplaces. An increasing number of researchers with empirical pieces of evidence is proving the advantage of experiential learning on learners over conventional methods. Setting the foundation for today’s classrooms, Edgar Dale’s Cone of Experience, or his Learning Pyramid (1940), illustrates how the depth of a person’s understanding depends on the medium leveraged and the senses involved in the learning process. Dale’s research identifies that direct, purposeful, or on-field experiences are the most effective method, resulting in 90 per cent retention of the information. In contrast, it revealed the least effective learning method through presented information like verbal and visual symbols. As Dale explains, people learn best when they are present in action and learn from their experience. In the world of data science, opening up the learner’s sensory channels to interact with the information at hand is bound to produce better results. Moreover, analytics and data science are practical fields, entailing practitioners to work on models, deal with data, and make engineering decisions. For instance, a data scientist cannot learn a hackathon solution without brainstorming the possibilities or building an intelligent model right from the textbook. Building Hard and Soft Skills Experiential learning methodologies and their effectiveness can be illustrated through the essential skills under the hard-skill and soft-skill umbrella in the analytics and data science space. While hard skills provide a foundation for all solutions, soft skills help in creating innovative ideas and communicating them. A nurtured combination of the two is what sets apart a data scientist from their peers. Hard Skills The need for practitioners to be skilled in the textbook technical concepts to ensure that the best possible analytical approach and models are built, while is necessary, is not sufficient. They need to be seasoned in applying the concepts in real-life problem situations. The way to develop application-oriented hard skills is to focus on three essential components. 1.  Applied knowledge of algorithms While one may have mastered algorithms, it is essential to know how and when to apply them. There may be instances when one comes across a problem where conventional algorithms don’t work. One will need to be fluent in writing a new\/heuristic algorithm or creative in tweaking the old ones. Applied knowledge is learned from experience, so one must practice applying oneself in the right way. 2. Translation for business context Data scientists often work with non-tech-based business professionals to find solutions to business problems or to create incremental business impact. It is paramount for them to understand the business context and translate those to data analytics problems, followed by building the right solution to map the context for timely implementation. This process also requires translating back the solution to business stakeholders in a language that they can comprehend. This is critical not only for a successful implementation of analytical solutions but to also set the stage for continuous improvement for incremental impact. Contextualisation leads to the adoption and growth of data-driven culture within organisations.  The skills acquired by one through the experiential learning approach can help with the above endeavour. 3.  Programming skills in Python or R Python or R can handle applications from data mining and ML algorithms to running embedded applications under one unified language. Data scientists need to be skilled in one or both programming languages to be successful in the field. The application-oriented case study-based approaches enabled through experiential learning methodologies enable one towards industry readiness with this skill. Soft Skills LinkedIn’s “Future of Skills” report from 2019 that studied behavioural insights based on millions of data points from member engagement identified soft skills to have increased value in enterprises. This, they reported, is given the expanding application of new technology that is broadening the job expectations for data scientists. The data science industry focuses majorly on hard skills, but it is time we lay enough importance on developing soft skills as well. There are three soft skills that are most important for a data scientist to nurture. 1. Critical thinking & problem-solving Critical thinking and problem-solving skills assist data scientists in clarifying vague and broad problems. If the dataset has errors or is not understood correctly, the solution will be unsuccessful. Under the experiential learning framework, one can build these skills by participating in hackathons, building models for experimentation, or engaging with data. 2. Effective communication Once one has solved the problem, it is important to communicate it to the stakeholders effectively. Data scientists’ inability to communicate with stakeholders is a pressing concern within the industry. If the receiver does not understand the solution, it will not be implemented. Individuals can hone this skill by putting themselves out there, explaining solutions to non-technical people, receiving feedback on it, and working on enhancing the skill with more practice. 3. Agility & flexibility Agility and flexibility are two skills that are increasingly becoming more important. The agile approach to working empowers data scientists to prioritise and create roadmaps based on business needs and adapt to different goals. Agile individuals are always learning and growing from new practical experiences. Conclusion In summary, experiential learning is learning by doing with application orientation and contextualisation. The framework is poised to get wide adoption in the field of analytics and data science globally across enterprises, functions, and academia. The aspirants and practitioners in the field should benefit from the framework to be continuous and prolific learners to upskill themselves in the most effective way and be future-ready. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"Method of teaching has been a subject of discussion and debate for a long time. The effectiveness of training and skilling has been questioned and deliberated time and again globally, irrespective of the field, stream, sector, or specialisation. The education sector, of late, has been witnessing a move away from traditional teaching techniques. Rote learning […]","categories":["AI Features"],"tags":[],"author_name":"Satyamoy Chatterjee","publish_date":"2021-12-09T16:00:00","publication_year":"2021","word_count":1580,"keywords":["data science","Go","API","AI","ML","RAG","Python","Aim","analytics","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","Python","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-experiential-learning-an-essence-to-address-the-skill-gap-in-the-field-of-analytics-and-data-science\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":53290,"title":"Google Transformer Model Reformer Works On A Single GPU &#038; Is Memory Efficient","content":"Since its inception in 2017, transformer models have become popular among researchers and academia in the machine and deep learning sector. This deep machine learning model is used for various natural language processing (NLP) tasks such as language understanding, machine translation, among others. In one of our articles, we had discussed why transformers play such a crucial role in NLP development. Massive transformer models have the capability to achieve state-of-the-art results on a number of tasks. However, training these models especially on long sequences can be a costly affair and can have massive computations. These large models when trained with model parallelism fails to be fine-tuned on a single GPU. To overcome the issue of cost and massive computations, recently, the researchers at tech giant Google and UC Berkely have introduced a new transformer model known as Reformer. The paper on Reformer has been accepted by the International Conference on Learning Representations (ICLR 2020) and is currently under review as a conference paper at the same. Behind the Model The core idea behind this transformer model is self-attention — the ability to attend to different positions of the input sequence to compute a representation of that sequence. In a traditional transformer model, the memory in a model with N layers is N-times larger than in a single-layer model due to the fact that activations require to be stored for back-propagation. Also, the attention on the sequences of length L is O (L2) in both computational and memory complexity, which in result can exhaust the accelerator memory. The researchers used the following methods while building the reformer model in order to improve the efficiency of the model as well as make it faster. Reversible layers are being used to enable storing only a single copy of activations in the whole model, so the N factor disappears. This means, it allows storing activations only once in the training process instead of N times, where N is the number of layers. Thus, applying these reversible residuals instead of the standard ones does change the model but has a negligible effect on training in all configurations.Splitting activations inside feed-forward layers and processing them in chunks ends up removing the (d)fffactor and saves memory inside feed-forward layers. Splitting activations only affect the implementation and is numerically identical to the layers used in the Transformer.Approximate attention computation based on locality-sensitive hashing replaces the O (L2) factor in attention layers with O (L) allows operating on long sequences. In simple words, the researchers replaced dot-product attention by one that uses locality-sensitive hashing, changing its complexity from O(L2) to O(L log L), where L is the length of the sequence. The researchers at the tech giant further trained the model with up to 20-layer big reformers on enwik8 and imagenet64 dataset in order to verify that the model can indeed fit large models on a single core and train fast on long sequences. Reformer combines the modelling capacity of a transformer with an architecture that can be executed efficiently on long sequences and with small memory use even for models with a large number of layers. Applications The ability to handle long sequences opens the way for the use of the Reformer on many complex generative tasks. Also, in addition to generating very long coherent text, the Reformer model has the capability to bring the power of transformer models to other domains like time-series forecasting, music, image and video generation. Wrapping Up Using the reversible residuals instead of the standard residuals not only made Reformer possible to perform faster but also shows higher memory efficiency than other transformer models while using only a single GPU. The motive behind this model is to help large, richly-parameterised transformer models become more widespread and accessible. Read the paper here.","excerpt":"Since its inception in 2017, transformer models have become popular among researchers and academia in the machine and deep learning sector. This deep machine learning model is used for various natural language processing (NLP) tasks such as language understanding, machine translation, among others. In one of our articles, we had discussed why transformers play such […]","categories":["Global Tech"],"tags":["Generative Pre-Trained Transformer","Google","Transformer Model"],"author_name":"Ambika Choudhury","publish_date":"2020-01-07T14:19:18","publication_year":"2020","word_count":628,"keywords":["Go","machine learning","programming_languages:R","AI","ai_frameworks:Transformers","Transformers","programming_languages:Go","NLP","deep learning","Generative Pre-Trained Transformer","Google","Transformer Model","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","NLP","Transformers","R","Go","ai_frameworks:Transformers","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-transformer-model-reformer-works-on-a-single-gpu-is-memory-efficient\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":69768,"title":"Are SD-WAN Services for You?","content":"SD-SWAN services are growing in popularity. Currently, many organisations, buoyed by the need to use the new versions of SDNs, are turning to SD-WANs as the perfect solution to the problems that are associated with the use of traditional WAN technologies. Besides, many organisations are turning to SD-WANs as opposed to Hybrid WANs for the main reason that SD-WANs are built on various improvements to the architecture and performance of Hybrid WANs. One problem that organisations face when it gets to using SD-WAN services relates to choosing the actual type of the service and the vendor with whom organisations can work. Currently, the market of SD-WAN services is flooded with service providers. However, the high number of service providers does not mean that the quality of the services that are available for your organisation has declined. On the contrary, it appears that the growing level of competition among players in this nascent industry is helping to set very high standards of service among the players in the industry. As many established and new service providers compete to outdo each other, they develop highly innovative and consistent service offerings in a bid to woo new clients and maintain the existing ones. You are the ultimate beneficiary of this process because you end up getting value for the money that you spend buying SD-WAN services. Here are some of the factors that may indicate you are ready for SD-WAN services. The need to connect your branch network If you have few or several branches, then you may consider it necessary to use a highly effective SDN to connect all your branches. An effective inter-branch connection network is essential for your work as an organisation. It is only when all the employees of your organisation can access and communicate effectively, regardless of their location, that you can say that your organisation is performing optimally. An SD-WAN service provides you with a highly integrated form of connection for all your branches. Given that the service has two sets of connections (one based on the internet and the other on MPLS or some other form of technology) it can help to keep your employees connected and engaged at all times. The need for enhanced security Cyber security is now an essential aspect of the IT policy of any organisation. It is impractical to operate in the current business environment without having in place a sound cybersecurity policy. However, you need to take this into consideration: developing and implementing a cyber security policy for your organisation is an expensive undertaking. One of the things that make it costly to implement a cyber security plan for your organisation relates to nature of the operations involved. SD-WAN service providers now provide a solution to this problem. It is common for some providers of SD WAN or Software Defined WAN services to offer bundled-up services. The main advantage of this service offering is that the vendor seeks to handle the security needs of the client during the life of the contract. In this case, you will not have to worry about the security features of your network, since your SD-WAN service provider will handle them. In conclusion, SD-WAN services may be for you if you have a number of branches that are remotely located and you need a highly reliable and flexible SDN for your operations. An SD-WAN service will help you solve this complex problem by providing high-speed connectivity for your organisation. Besides, you may need to use SD-WAN services as a way of addressing network security features for your organisation. Given the nature of the SD-WANs, it may be necessary for you to buy them for your organisation.","excerpt":"SD-SWAN services are growing in popularity. Currently, many organisations, buoyed by the need to use the new versions of SDNs, are turning to SD-WANs as the perfect solution to the problems that are associated with the use of traditional WAN technologies. Besides, many organisations are turning to SD-WANs as opposed to Hybrid WANs for the […]","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2017-03-02T04:53:49","publication_year":"2017","word_count":611,"keywords":["programming_languages:R","AI","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/sd-wan-services\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001544,"title":"Inside Bangalore Startup Postman That Is Driving A Seamless API Development Environment For Developers","content":"Building APIs can be one of the most challenging things that a company faces, with issues such as security, integration with multiple backend legacy systems, support culture or others. While many companies are still learning to deal with it, Bengaluru-based Postman is climbing the popularity chart by being a solution to these challenges and serving as a complete API development environment making things quicker and easier for nearly five million developers and more than 100,000 companies across the globe. Postman is essentially an elegant, flexible tool used to build connected software via APIs. Founded in 2014, it is headquartered in San Francisco and boast of specialities in API development, API management, Saas, and more. Curious Dose got in touch with Abhinav Asthana, CEO & Co-founder, Postman to understand the use cases ranging from single developer to an enterprise. What Is Postman? As the company claims, Postman is the only complete API development environment. The comprehensive set of built-in tools support every stage of the API lifecycle so individuals and teams can easily maintain a single source of truth. “You can design and mock, debug, test, document, monitor, and publish your APIs – all from the same place. Postman allows you to manage your APIs on the Postman native apps for MacOS, Windows, and Linux, with Newman, Postman’s command line tool, and via the cloud using Postman Monitoring,” shares Asthana. The Postman free app was created as a side project, and first introduced in October 2012. It grew quickly to become one of the most popular apps on the Chrome store and expanded to Mac, Windows, and Linux native apps. Today they have 6 million developers using their apps, worldwide. How Does It Make Things Easy? Postman was initially created because Abhinav Asthana, Postman’s CEO and co-founder, noticed that API testing was difficult and inefficient across the board – no matter where he was working. He set out to create a tool that would simplify the API testing process. “With the goal to make API development better and faster, we have a powerful set of built-in tools that are able to address every stage of API development. Postman’s simple UI allows users to do everything from design and mock to publish APIs, and in addition, includes tools like the Postman API and Newman, Postman’s command line tool, to allow for flexible access to the Postman API development environment,” says Asthana. Postman Collections are the most common API specification format and are at the core of every tool within Postman. This makes it easy to save and reuse your work throughout each stage of the API life-cycle. Collections also allow you to collaborate with team members on Postman. How It Works: Postman provides a simple UI that helps developers and teams create collections made up of individual HTTP requests and organise these requests into folders. Creating collections helps the user to stay organised, up-to-date, and allows to attach test scripts to requests, build workflows, build integration tests suites, and pass data between API requests. How Did It Start? Asthana shares that during his time at Yahoo, he met Ankit Sobti, now the Co-founder and CTO of Postman. “We were building a front-end architecture of an app and testing APIs was a pain point. There were a lot of communication issues with different teams. We thought of making the problem simpler but being first-time engineers, we didn’t work on the problem then,” he said. Soon he went ahead to build TeliportMe where he faced similar issues. It was during this time that he created the prototype of the first Postman app. The other co-founder is Abhijit Kane who is the product architect at the company. “The most common coding protocol is HTTP and within that protocol, you have ‘post’ used as a term. Adding ‘man’ to it, created a dual-meaning name for us – because it was about delivering seamlessly to developers,” says Asthana when asked why the name Postman. Growth Story Postman evolved from a side project to a mature product with several integrated features within a short time frame. They are 100-people strong across the offices and are looking to add more in their engineering team in Bengaluru. Postman is used by 6 million developers and more than 200,000 companies to access 130 million APIs every month. These numbers and a clientele of names such as Atlassian, VMware, PayPal, and DocuSign suggest that they have grown spectacularly over the last few years. “Postman’s mission is always to solve problems. That’s how the company started. So our growth plan is to continue solving the problems that people find in creating connected services – a growing industry with plenty of challenges to overcome,” said Asthana in the concluding note.","excerpt":"Building APIs can be one of the most challenging things that a company faces, with issues such as security, integration with multiple backend legacy systems, support culture or others. While many companies are still learning to deal with it, Bengaluru-based Postman is climbing the popularity chart by being a solution to these challenges and serving […]","categories":["AI Features"],"tags":["Postman","postman api"],"author_name":"Srishti Deoras","publish_date":"2019-03-19T15:59:55","publication_year":"2019","word_count":786,"keywords":["Go","API","programming_languages:R","AI","IPO","postman api","ML","programming_languages:Go","GAN","Aim","Postman","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","API","GAN","IPO","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/inside-bangalore-startup-postman-that-is-driving-a-seamless-api-development-environment-for-developers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058012,"title":"Council Post: Designing hackathons to encourage problem-solving in AI\/ML","content":"As per a QuantHub report, 67 per cent of the surveyed companies said that they are expanding their data science team. However, it was found that there are almost three times as many data science jobs as there are aspirants. This seems to reflect the general trend in the industry. Are you one of those business leaders facing challenges in hiring? Talent management frameworks in data science organisations are seeing a radical shift as organisations rely on various methods to hire, train, and retain talent. Unlike other fields, data science is a field that demands much more than a theoretical understanding of the concepts from the practitioners. To be able to put theories into practice while creating business solutions is a critical measure of success. Hackathons are a successful tool for AI\/ML professionals to apply their learnings to a real-life simulation. The fast-paced collective ‘marathon’ revolves around creating a solution for a given problem within the span of one or two days. Hackathons involve participants being divided into teams, receiving mentoring from experts and presenting their findings to a judging panel that decides which solutions will be implemented. This event is an integral framework to develop and build tomorrow’s AI\/ML leaders. For a long time, the focus of the data science industry has been on theoretical learning. This can also be seen in hackathon results that often lack innovative ideas or the ability to translate them effectively within the business context to the panel of judges. We teach AI\/ML professionals the theories but not how to put them into practice or prove them to business leaders, thereby limiting the potential for data science to reach executive boardrooms or become part of everyday decision-making within business operations. We need AI\/ML professionals who can think outside the box to resolve problems and combine an analytical approach with business savviness. Applicability within the business context in data science is more important than one might assume. You could build the best analytical model, but it will remain on the shelf if business leaders do not understand its application within their business processes. Unfortunately, the training to do so is lacking, especially in the theory-focused education system in India. So, as leaders, it is upon us to instil and train AI\/ML professionals in developing these skills. One effective way to do so is through hackathons. There is a certain way to go about designing successful hackathons that, among achieving other goals, encourages the development of problem-solving and business savviness in AI\/ML professionals. Let’s look at the important steps in doing so. Be clear on the objective of the hackathon The foundational step to designing a hackathon is to have your objective clearly defined. Given that AI\/ML professionals work on the solution from start to end, leaders must know what they want from the hackathon and challenge the participants accordingly. Three of the most common objectives for hackathons are: Identifying Talent Interviews are becoming increasingly redundant in data science hirings. Organisations are going beyond traditional hiring methods to understand how prospective employees deal with practice problems. Hackathons are a constructive way to identify talent by monitoring participants in a simulated real-life situation on different parameters. These include problem-solving abilities, how they work under pressure, whether they are team players, whether they can think of creative solutions or judge their solutions compared to potential competitors. Creating MVPs Minimum viable products are a challenge to create in a short span of time. It requires innovative solutions, team unity, efficient management and deep technical knowledge. Hackathons are an effective way for organisations to judge such MVP initiatives and could lead to future funding to create assets or support and encourage research in specific areas. Creating a buzz around a new platform being introduced Hackathons are also a great way for publicity of new initiatives or new platforms taken up by the organisations. Young AI\/ML professionals are always looking for hackathons to participate in, and creating one around your new platform can help bring more traction and interaction to it. Set a theme for the event Once your objectives are set, move on to setting a theme for the hackathon. The theme questions should not be prescriptive. Instead, having open-ended questions will allow participant creativity. It should rely on larger and holistic datasets, making it important to ensure enough information is available around the selected themes. Innovative thinking and problem-solving can only be achieved when hackers are pushed to think outside the box. They will have to scale their thinking to work efficiently and build the needed effective solutions. Create clear assessment guidelines It is critical to know what you want from your talent and create guidelines to assess the desired qualities. At this step, remember that a successful data scientist is one who combines technical, analytical, and business skills effectively. Ensure that your hackathon outputs can measure all three parameters. What you want from your talent may not be having the smartest analytical coder but a balanced combination of the above qualities. Designing assessment markers accordingly can help ensure your goals are being met. When it comes to assessing the finalists, their insight consumption skills need to be judged too. This ensures the candidate’s ability to use the right brain and translate their solutions for better business impact. To assess the quality of the work being done in detail, establish checkpoints throughout the process. Checkpoints ensure that the desired outputs are generated, and the quality of work is maintained. To assist in framing the participants into producing the right output, create a set of recommendations and guidelines. Identify evaluators and judges relevant to your objectives Along with the evaluation criteria, the evaluators or judges also play a critical role in meeting your objectives. They are responsible for evaluating and mentoring the participants for their analytical, technical and business skills. Bring in evaluators and judges with specific AI capabilities and domain expertise, and pair them up with the participants accordingly. Moving ahead, choose the finalists among the participant pool based on the quality of output, and allow them to present their capability to the judges. Assign prizes based on the hackathon objectives Circling back to our first step, one must keep the hackathon objectives in context while aligning the prizes for the winners. Depending on the objective, the awards or prizes should differ. Your end prizes could offer job placement if your objective is to identify talent, funding if your objective is to create MPVs, or training or certifications if your objective is to advertise your product. Hackathons have a lasting impact on the AI\/ML professional community by being the source for practitioners to interact and collaborate. For organisations and leaders, they are a powerful tool to identify and shape the leaders for tomorrow. So, let’s use it wisely! This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"As per a QuantHub report, 67 per cent of the surveyed companies said that they are expanding their data science team. However, it was found that there are almost three times as many data science jobs as there are aspirants. This seems to reflect the general trend in the industry. Are you one of those […]","categories":["AI Features"],"tags":["ai hackathon india","Data Science Hackathons","Hackathons"],"author_name":"Kaushik Sanyal","publish_date":"2022-01-10T18:00:00","publication_year":"2022","word_count":1172,"keywords":["ai hackathon india","data science","Go","TPU","AI","ML","Hackathons","RAG","Data Science Hackathons","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","TPU","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-designing-hackathons-to-encourage-problem-solving-in-ai-ml\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":35228,"title":"Top 5 AI-Based Text-To-Video Products","content":"Artificial Intelligence and machine learning have been increasingly helpful in creating and rating visual contents and stories. In this article, we list down five AI-based text-to-video products that will help any storyteller put forward their best content. 1| Lumen5 About: This is a video creation platform by artificial intelligence that enables anyone without training or experience to easily create engaging video content within minutes. There are mainly three categories for this product. The Comunity category is cost-free and will provide you with 480p resolution including standard library and popular themes. The Pro category will cost you $49 per month which will provide you with a resolution of 720p along with the other features. The Business category is designed for businesses looking to tell their brand story and it cost $149 per month which comes with a resolution of 1080p and various other features. How It Works: The transformation of text to video follows the steps mentioned below: Enter the article link to turn it into a video where the natural language processing (NLP) algorithm will automatically create a storyboard for you. Using computer Vision technology, the system will find relevant visuals and audio to your content. Set up your branding profile once to add every video including a call-to-action to drive conversion. The dashboard can be easily accessed by your team’s administrator to watch and approve the video. 2| GliaCloud About: The Taiwanese startup, GliaCloud is launched to provide solutions and applications in data analytics and machine learning in 2015 and is the brainchild of David Chen. GliaCloud’s product, GliaStudio uses the technologies of artificial intelligence to automatically create video summaries of text articles. How It Works: The process of converting text to video follows the steps mentioned below: Just paste the content URL or upload file, the natural language algorithm of GliaCloud will go over the content to find major topics and keywords, and then generate video scripts with sections and highlights. Then, based on the generated video script, the AI engine will search and edit the corresponding image and clips. It is further easy to create different versions for testing and obtain better engagement on social media. 3| Wochit About: This platform with a super intuitive approach for any storyteller with any level of video creation experience or skill to produce and share their stories across all social and digital platforms. At the core of the platform is Wochit’s Predictive Video Creation platform, developed to pre-empt and serve the needs of every video creator and team, bringing all they need to create high impact videos. How It Works: The process follows four main steps as mentioned below: Plan to kickstart your creative process with new and trending story ideas primed to help you ace every goal. Select the most relevant licensed assets, graphics, and music tracks come together with the voiceover artists and also choose your own branded content to tell your stories Create the ultimate combination of advanced editing capabilities and automated tools to scale your creation process and craft powerful videos in a faster manner. Collaborate with streamline internal workflow and approval processes with the agile cloud-based platform and smart Slack integration. Tag-team to create, edit, repurpose, approve and distribute from anywhere. Publishing your story is easy by CMS, OVP, and social media integration and also ensure smooth delivery to every audience in the optimal format for each platform you choose. 4| Wibbitz About: This platform is found in 2011 by Zohar Dayan and Yotam Cohen as the first automated video creation platform in Tel Aviv and New York which automatically convert text into short-form video. Snippets of text overlay B-roll footage, with users empowered to do some tweaks with easy-to-use editing tools. There are two plans basically in this platform, the Standard Plan, and the Custom Plan. The Standard Plan costs $500 per month while the Custom Plan includes everything in the Standard Plan along with various other features. How It Works: Smart text analysis finds the highlights in your story and automatically outlines your video and AI pairs the subject matter in your story with relevant footage from our media library. To start working in the platform, start by requesting a demo, and the providers reach out to schedule a time to walk you through the platform. Depending on which plan you pick, you’ll learn the ropes with a hands-on training session or a self-guided tour. Templates for turning any story or article into a short-form video optimized for audience engagement. Cut, crop, zoom, and trim each section without needing any professional editing skills and also you can add your own voice by requesting a professional voiceover or easily uploading your own. You can save your default colors, fonts, logos, and bumpers to create a consistent finished product. 5| Vedia About: This artificial intelligence video maker enables users to instantly create professional looking videos from the text. Also, this platform can not only transform text into video but also URLs and Data with the help of automated video creation. Vedia does the work of transforming your ideas into a video by generating narration, sourcing media and assembling your scenes, freeing you from hours of frustration. How It Works: These follow three main steps as mentioned below: Analyse: The artificial intelligence firstly scans and analyses your data, text, links, blogs or feeds to identify the main ideas. Visualise: After analysing, it finds related media assets and places them on the video timeline and generates the voice narration. Customise: Finally, all you need to do is review and publish the video. You can also customise it to stay on brand with the provided drag & drop video maker.","excerpt":"Artificial Intelligence and machine learning have been increasingly helpful in creating and rating visual contents and stories. In this article, we list down five AI-based text-to-video products that will help any storyteller put forward their best content. 1| Lumen5 About: This is a video creation platform by artificial intelligence that enables anyone without training […]","categories":["AI Trends"],"tags":["AI products"],"author_name":"Ambika Choudhury","publish_date":"2019-02-21T12:37:06","publication_year":"2019","word_count":942,"keywords":["Go","machine learning","artificial intelligence","AI","ML","computer vision","RAG","NLP","analytics","R","AI products"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","computer vision","analytics","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-ai-based-text-to-video-products\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10009174,"title":"Salesforce’s Acquisition Of Tableau Is Starting To Show Off","content":"After acquiring Tableau Software last year through a transaction of $15.7 billion, Salesforce has now announced that Einstein Analytics will now be integrated with Tableau and it will be rebranded as Tableau CRM. With this integration, Salesforce hopes to provide the users with several benefits such as: A simpler way to connect any dataIntegrating the visual analytics capabilities of Tableau with the deep AI capabilities of Einstein AnalyticsHelp in faster decision making The intent behind this decision is to project Tableau as the analytics brand for Salesforce and integrate Einstein Analytics’ capabilities into its full offering, including — Tableau Desktop for visual analytics; Tableau Prep for visual data preparation; Tableau Server for self-managed analytics platform; Tableau Online for analytics platform as a service; and Tableau CRM for native Salesforce analytics. What to Expect from Tableau CRM While both Tableau and Einstein Analytics are focused on analytics, they use different technologies and have different use cases. Tableau uses a proprietary technology VizQL that translates user actions into interactive visualisations that can help a user to explore data and pivot analysis. This platform provides limitless data exploration and visualisation, connectivity to data anywhere, and deployment and integration capabilities. On the other hand, Einstein Analytics allows the user to combine internal data with the external and provides a more intelligent experience. Einstein Analytics also offers AI\/ML capabilities for building predictive models. With the integration, the user will be able to perform the following activities: Tableau can now query data stored in Einstein Analytics, thus revealing deep exploration capabilities of Tableau queries. Einstein Analytics can now create extracts in Tableau containing Salesforce data by being able to write data into Tableau.Tableau Prep will be able to write to both Tableau and Einstein Analytics. This gives the user a choice to choose the right platform for their needs. With Tableau CRM, Salesforce hopes to integrate dynamic machine learning-enabled predictions, insights, and recommendations to Tableau with the integration of Einstein Discovery, which will be launched in 2021. Einstein Discovery is different from Einstein Analytics in a way that it is a prescriptive and a predictive analytical tool, which provides recommendations, explanations, and prediction based on data source without building a sophisticated data model. In contrast, Einstein Analytics is a self-service analytics platform that helps in data visualisation and exploration. Speaking at Tableau Conference (ish) 2020 (TC20), Bobby Brill, the product management director of the Einstein Analytics team, said that the following features would be made available: A new extension to the Tableau dashboard will bring predictions, explanation, and suggestions to improve Einstein Discovery outcomes in the Tableau vizzes. Anybody with access can have a look and suggest actions to be taken without requiring any coding.Einstein Discovery will also be integrated into the calculation engine of Tableau, which will enable users to use prediction capability with much ease. Einstein Discovery powered time prediction can now be obtained just by dragging and dropping predictive calculations into the viz.The data prep is now easier, and with Einstein Discovery, users can add prediction nodes as a step in the prep flow and improve data sets. Wrapping Up With its easy to build business reports and dashboards, Tableau has been extremely popular. With its new collaboration, new horizons open up, especially in terms of the reach and the effectiveness of Analytics Solution built around the Salesforce Analytics tools. Einstein Analytics and Discovery offer added advantage of predictive analytics and use cases in the areas of accounts or to the sales to grow business. While using Einstein products for Salesforce Users lead to effective utilisation, Tableau yields a collaborative and successful approach towards analytics for business. Einstein Discovery and Analytics and Prediction API can be used in Tableau for further analysis through Tableau reports and dashboards.","excerpt":"After acquiring Tableau Software last year through a transaction of $15.7 billion, Salesforce has now announced that Einstein Analytics will now be integrated with Tableau and it will be rebranded as Tableau CRM. With this integration, Salesforce hopes to provide the users with several benefits such as: A simpler way to connect any data Integrating […]","categories":["AI Features"],"tags":["Salesforce","Tableau Salesforce"],"author_name":"Shraddha Goled","publish_date":"2020-10-08T16:00:14","publication_year":"2020","word_count":624,"keywords":["API","machine learning","AI","R","Tableau Salesforce","ML","RAG","ViT","Salesforce","analytics","predictive analytics","analytics platform"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","predictive analytics","R","API","ViT","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/salesforces-acquisition-of-tableau-is-starting-to-show-off\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60551,"title":"Power Analytics Global Launched COVID-19 Impact Data Simulation Modelling Solution","content":"Power Analytics Global has announced a new COVID-19 impact data simulation modelling solution. The goal of the solution has been designed to deliver a real-time predictive COVID-19 modelling and simulation tool for customers to cost-effectively manage, protect, and enhance their workforce’s productivity, along with having supply chains efficiencies, and critical network infrastructure optimisation. The new solution has the capacity to digest numerous types of data from any source virtually possible and process in real-time over two million data inputs per second. The infrastructure modelling library can model in real-time and boasts a proprietary library of over 30,000 pre-configured modelling components. The underlying data ingestion layer can allow a customer to run any data queries 10 to 100 times faster, at a fraction of their normal storage cost. This new real-time impact data modelling solution will allow users to make better decisions during the disruptions caused by COVID-19. The offering will allow customers to drastically reduce the cost, time and complexity of accessing critical, operational data from globally distributed network assets, infrastructure, and people. The hardened solution and real-time capabilities of this powerful new modelling tool will enable our clients to incorporate data from any source, predict trends, protect critical assets, and take action in real-time across all relevant components of distributed heterogeneous data sets, all while lowering cost, and increasing resilience. This powerful new platform simplifies and accelerates secure access to large, fragmented, and geographically distributed data sets to support the most demanding Machine Learning (ML), Artificial Intelligence (AI), and IoT workloads. The Power Analytics modelling solution suite gives network infrastructure management and electrical engineering professionals the required control over their critical power infrastructure. It includes designing and modelling, simulation and analysis, eliminating downtime, reducing energy costs and optimising business operations. The next-gen solution has been built on proprietary intellectual property and decades of engineering and design expertise. Power Analytics boasts a portfolio of 24 patents. The platforms allow customers to take full control of their critical network infrastructure, reduce costs, automate operations, protect critical assets, and improve business reliability. According to Keith Barksdale, the chairman of Power Analytics Global, COVID-19 has a global impact, and it is putting unforeseen stress on critical network infrastructure. This, in turn, has adverse effects on some of our clients, such as governments, hospitals, power plants, telecom operators, airports, and logistics companies. The company felt the need to expand their capabilities from real-time power modelling to incorporate any real-time data set. Additionally, Mike Kotlarz, the CTO of Power Analytics Global, stated that not only are our friends, families, and loved ones under siege, but so is the critical infrastructure that is enabling us to continue to be connected and productive during this trying time. The real-time digital twin modelling allows the company to deliver a solution that can help its customers assess the impact of COVID-19 on business.","excerpt":"Power Analytics Global has announced a new COVID-19 impact data simulation modelling solution. The goal of the solution has been designed to deliver a real-time predictive COVID-19 modelling and simulation tool for customers to cost-effectively manage, protect, and enhance their workforce’s productivity, along with having supply chains efficiencies, and critical network infrastructure optimisation.  The new solution has […]","categories":["AI News"],"tags":["Coronavirus","corporate analytics platform","covid-19"],"author_name":"Sejuti Das","publish_date":"2020-03-31T18:30:00","publication_year":"2020","word_count":474,"keywords":["Go","artificial intelligence","machine learning","covid-19","AI","ELT","corporate analytics platform","ML","Git","RAG","Coronavirus","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","R","Go","Git","ELT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/power-analytics-global-launched-covid-19-impact-data-simulation-modelling-solution\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10105243,"title":"Google Unveils VideoPoet, a New LLM for Video Generation","content":"Google researchers recently introduced VideoPoet, a new large language model (LLM) for video generation. This model is designed to perform a range of tasks including text-to-video, image-to-video, video stylisation, video inpainting and outpainting, and video-to-audio conversion. The introduction of VideoPoet addresses the challenge of creating coherent large motions in videos, a limitation in current video generation technologies. This new model differentiates itself by integrating multiple video generation capabilities within a single LLM framework, contrasting with the segmented approach of existing models. It uses various modalities and is trained with multiple tokenizers, such as MAGVIT V2 for video and image, and SoundStream for audio. This allows VideoPoet to perform diverse tasks, from animating images to editing and stylising videos based on text inputs. How it compares with others In the evolving landscape of AI-generated video technology, VideoPoet emerges as a significant advancement, distinguishing itself from existing models like Imagen Video, RunwayML, Stable Video Diffusion, Pika, and the most recent ‘Animate Anyone’ from Alibaba Group. through its enhanced capabilities in text fidelity and motion interestingness. This new model outshines its counterparts by more accurately following text prompts and generating videos with more engaging motions. Key points of comparison include: Zero-Shot Capabilities, VideoPoet, like other contemporary models, excels in generating content from minimal input, such as a single text prompt or image, without needing specific training on that content. However, unlike other models which may struggle with large motion coherence, VideoPoet showcases a higher degree of accuracy in translating text prompts into video, enhancing the user experience. Where other models often face challenges in creating large, artefact-free motions, VideoPoet demonstrates a marked improvement, creating more dynamic and fluid videos. As with Google announcements Amidst the announcement of VideoPoet by Google Research on December 19, 2023, some scepticism exists within the community regarding its practical application and effectiveness. While VideoPoet showcases advancements in text fidelity and motion interestingness in video generation, critics question the reliance on specific prompting techniques. There are observations that the use of terms like “8k” in prompts, a trick from previous AI models like VQGAN + CLIP and Stable Diffusion, may be employed to artificially enhance photorealism, raising concerns about the model’s genuine capability. Overall, while VideoPoet represents a significant step in video generation technology, its real-world application, effectiveness, and impact remain subjects of debate and speculation within the community.","excerpt":"It looks better than RunwayML, Stable Video Diffusion, Pika, and the most recent ‘Animate Anyone’","categories":["AI News"],"tags":[],"author_name":"K L Krithika","publish_date":"2023-12-20T12:13:33","publication_year":"2023","word_count":390,"keywords":["Go","AI","Modal","RPA","ML","stable diffusion","ViT","CLIP","GAN","R"],"extracted_tech_keywords":["AI","ML","R","Go","CLIP","stable diffusion","GAN","ViT","RPA","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-unveils-videopoet-a-new-llm-for-video-generation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018069,"title":"Top XGBoost Interview Questions For Data Scientists","content":"Introduced a few years ago by Tianqi Chen and his team of researchers at the University of Washington, eXtreme Gradient Boosting or XGBoost is a popular and efficient gradient boosting method. XGBoost is an optimised distributed gradient boosting library, which is highly efficient, flexible and portable. The method is used for supervised learning problems and has been widely applied by data scientists to get optimised results for various machine learning challenges. It implements ML algorithms under the Gradient Boosting framework and helps in solving data science problems in a fast and accurate manner. Here are the top ten interview questions on XGBoost that Data Scientists must know. 1| Is XGBoost faster than random forest? Solution: XGBoost is usually used to train gradient-boosted decision trees (GBDT) and other gradient boosted models. Random forests also use the same model representation and inference as gradient-boosted decision trees, but it is a different training algorithm. XGBoost can be used to train a standalone random forest. Also, random forest can be used as a base model for gradient boosting techniques. Further, random forest is an improvement over bagging that helps in reducing the variance. Random forest builds trees in parallel, while in boosting, trees are built sequentially. Meaning, each of the trees is grown using information from previously grown trees, unlike bagging, where multiple copies of original training data are created and fit separate decision tree on each. This is the reason why XGBoost generally performs better than random forest. Know more here. 2| What are the advantages and disadvantages of XGBoost? Advantages: XGB consists of a number of hyper-parameters that can be tuned — a primary advantage over gradient boosting machines.XGBoost has an in-built capability to handle missing values.It provides various intuitive features, such as parallelisation, distributed computing, cache optimisation, and more. Disadvantages: Like any other boosting method, XGB is sensitive to outliers.Unlike LightGBM, in XGB, one has to manually create dummy variable\/ label encoding for categorical features before feeding them into the models. Know more here. 3| How XGBoost Works? Solution: When using gradient boosting for regression, where the weak learners are considered to be regression trees, each of the regression trees maps an input data point to one of its leaves that includes a continuous score. XGB minimises a regularised objective function that merges a convex loss function, which is based on the variation between the target outputs and the predicted outputs. The training then proceeds iteratively, adding new trees with the capability to predict the residuals as well as errors of prior trees that are then coupled with the previous trees to make the final prediction. Click here to learn the step by step process of how XGB works. 4| What does the weight of XGB leaf nodes mean? How to calculate it? Solution: The “leaf weight” can be said as the model’s predicted output associated with each leaf (exit) node. Here is an instance of how to calculate the weights of the leaf nodes in XGB- Consider a test data point, where age=10 and gender=female.To get the prediction for the data point, the tree is traversed from the top to bottom, performing a series of tests. At each of the intermediate nodes, a feature is needed to compare against a threshold. Now, depending on the result of the comparison, one must proceed to either the left or right child node of the tree. In case of (10, female), the test “age < 15” is to be performed first and then proceed to the left branch, because “age < 15” is true. Then, the second test “gender = male?” is performed, which evaluates to false, so we proceed to the right branch. We end up at the Leaf 2, whose output (leaf weight) is 0.1. Click here to know more in detail. 5| What are the data pre-processing steps for XGB? Solution: The data pre-processing steps for XGB include the following- Load the dataExplore the data and remove the unneeded attributesTransform textual values to numericFind and replace the missing values if neededEncoding the categorical dataBreak the dataset into training set as well as test setPerform feature scaling or data normalisation Know more here. 6| How does XGB calculate features? Solution: XGB automatically provides the estimations of feature importance from a trained predictive model. After a boosting tree is constructed, it retrieves feature importance scores for each attribute. The feature importance contributes a score which indicates how much valuable each feature was in the construction of the boosted decision trees within the model. Also, in terms of accuracy, XGB models show better performance for the training phase and comparable performance for the testing phase when compared to SVM models. Besides accuracy, XGB has higher computation speed than SVM. Know more here. 7| Why does XGBoost perform better than SVM? Solution: In case of missing values, XGB is internally designed to handle missing values. The missing values are interpreted in such a way that if there endures any trend in the missing values, it is captured by the model. Users are required to supply a different value than other observations and pass that as a parameter. XGBoost tries different things as it encounters a missing value on each node and learns which path to take for missing values in future. On the other hand, Support Vector Machine (SVM) does not perform well with the missing data and it is always a better option to impute the missing values before running SVM. Know more here. 8| Differences between XGBoost and LightGBM. Solution: XGBoost and LightGBM are the packages belonging to the family of gradient boosting decision trees (GBDTs). Traditionally, XGBoost is slower than lightGBM but it achieves faster training through the Histogram binning process.LightGBM is a newer tool as compared to XGBoost. Hence, it has fewer users and thus a narrow user base than XGBoost and contains less documentation. Know more here. 9| How does XGB handle missing values? Solution: XGBoost supports missing values by default. In tree algorithms, branch directions for missing values are learned during training. It is important to note that the gblinear booster treats missing values as zeros. During the training time XGB decides whether the missing values should fall into the right node or left node. This decision is taken to minimise the loss. If there are no missing values during the training time, the tree made a default  decision to send any new missings to the right node. Know more here. 10| What is the difference between AdaBoost and XGBoost? Solution: XGBoost is flexible compared to AdaBoost as XGB is a generic algorithm to find approximate solutions to the additive modeling problem, while AdaBoost can be seen as a special case with a particular loss function. Unlike XGB, AdaBoost can be implemented without the reference to gradients by reweighting the training samples based on classifications from previous learnersnow more here.","excerpt":"Introduced a few years ago by Tianqi Chen and his team of researchers at the University of Washington, eXtreme Gradient Boosting or XGBoost is a popular and efficient gradient boosting method. XGBoost is an optimised distributed gradient boosting library, which is highly efficient, flexible and portable.  The method is used for supervised learning problems and […]","categories":["AI Highlights"],"tags":["data science master","data science training","Data Scientist"],"author_name":"Ambika Choudhury","publish_date":"2021-01-14T11:00:00","publication_year":"2021","word_count":1142,"keywords":["data science","Go","machine learning","TPU","AI","ML","data science master","distributed computing","data science training","XGBoost","LightGBM","Data Scientist","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","XGBoost","LightGBM","distributed computing","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/top-xgboost-interview-questions-for-data-scientists\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10043364,"title":"Behind Goldman Sachs Model That Predicts Euro 2020 Winner","content":"“It’s (probably) coming home,” said Goldman Sachs’ Christian Schnittker on England’s chances of winning Euro 2020 (the tournament was put off for a year due to pandemic). Three Lion fans were hoping Goldman Sachs would get it right in the third go-around and England would win the cup (Euro went to Rome). The prediction model crunched the data from about 6,000 football matches played from 1980 and took into account factors such as current team strength, recent performances, home-field advantage, etc. Euro 2020 prediction Goldman Sachs’ prediction model performed the following steps: The algorithm starts by modelling the number of goals by each team using a large dataset of international football matches. The number of goals provides the following information: The strength of the squad is measured with the World Football Elo Rating. The Elo rating system calculates the relative skill levels of players in zero-sum games such as chess. This Elo ranking did not incorporate individual player information for Goldman Sachs’ model but correlated highly with other metrics like the FIFA rankings and teams’ estimated transfer values. Credit: Goldman Sachs Goals scored and conceded in the last five matches: This data helps in capturing the momentum of a team in the run-up or during the European Cup. Credit: Goldman Sachs Home advantage is a pivotal factor in the number of goals scored. Goldman Sachs’ team found that, on average, the home team scored 0.4 goals more. With this, England seems to have an advantage since both the semi-finals and the finals are being hosted at Wembly. Credit: Goldman Sachs Tournament effect– referring to countries (Croatia, the Netherlands, and Germany) punching above their weight apropos Elo ratings at major tournaments –is another crucial parameter. Goldman Sachs said, while the prediction model captures the stochastic nature of the tournament, the forecasts are ‘highly uncertain’ as football is an unpredictable game. Third time lucky? In 2018, Goldman Sachs came up with a machine learning-based statistical model to predict the World Cup outcomes. The prediction fell flat. The AI system ran a simulation of one million possibilities and variations. Originally it predicted that Germany and Brazil would face each other in the finals, however, the former made an early exit. The AI system then concluded that England would be the second finalist and Brazil would win the cup. The prediction of the AI system embarrassingly failed as France and Croatia were the two finalists, with France taking home the crown. In 2014, Goldman Sachs used a simpler and less ambitious statistical model for predicting the 2014 World Cup results. It used fewer parameters such as the number of goals scored in the last ten official international matches and teams’ rankings. The model too failed at making accurate predictions. Apart from Goldman Sachs, USB and ING have also tried their hand at model-based predictions, but with no success. However, the Japanese bank Nomura made a successful prediction in FIFA 2018. The bank used portfolio theory (profiling teams on the basis of players’ value, the momentum of team performance, and historical performance) to predict France as the winner. However, the bank got the second-runner up wrong. Why is it difficult to predict? The advent of machine learning and data analytics have made the prediction game exciting. The predictors analyse historical data, run simulations, and use state of the art statistical techniques to forecast match outcomes. The programmers and AI researchers rely on quantifiable data to make observations. If the data is not authentic, results are bound to be inaccurate. Further, models can’t account for intangible (read unquantifiable) variables such as team dynamics, player emotions, fans sentiments etc. Experts believe predictions for Football matches are inherently complex, as compared to other games. Director of analytics at Merkle, Debs Balme, said in an interview, “For sports like baseball or basketball where there are lots of games against the same opposition, it’s an easier solution to predict, as there is more data available. Baseball players play 162 games a season, for example. And the Mets and Yankees have played each other 115 times. So there is more history of performance [than in the World Cup], and a greater wealth of data to be able to make more accurate predictions.”","excerpt":"“It’s (probably) coming home,” said Goldman Sachs’ Christian Schnittker on England’s chances of winning Euro 2020 (the tournament was put off for a year due to pandemic). Three Lion fans were hoping Goldman Sachs would get it right in the third go-around and England would win the cup (Euro went to Rome). The prediction model […]","categories":["AI Features"],"tags":["goldman sachs","Machine Learning"],"author_name":"Shraddha Goled","publish_date":"2021-07-12T12:00:00","publication_year":"2021","word_count":701,"keywords":["Go","machine learning","programming_languages:R","AI","Machine Learning","programming_languages:Go","RAG","goldman sachs","analytics","AI research","R"],"extracted_tech_keywords":["AI","machine learning","analytics","RAG","R","Go","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/behind-goldman-sachs-model-that-predicts-euro-2020-winner\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10122257,"title":"NVIDIA Unveils ‘Rubin’ Months Ahead of Blackwell Release, AMD Announces MI400X","content":"At Taipei’s Computex Conference, NVIDIA CEO Jensen Huang announced the launch of the Rubin AI chip platform, slated for 2026, and the Blackwell Ultra chip, expected in 2025, marking a shift to an annual update cycle for NVIDIA’s AI accelerators. The Rubin architecture follows the March announcement of the Blackwell model, which is set to ship later in 2024. “We are seeing computation inflation,” Huang stated, highlighting the need for accelerated computing to manage the growing data processing demands. He emphasised NVIDIA’s technology, which promises 98% cost savings and 97% less energy consumption. Previously, NVIDIA had a two-year update timeline for its AI chips. The shift to an annual release schedule underscores the competitive intensity in the AI chip market and NVIDIA’s efforts to maintain its leadership. The Rubin platform will feature new GPUs and a central processor named Vera, although details were scarce. Huang announced that the forthcoming Rubin AI platform will incorporate HBM4, the next generation of high-bandwidth memory. This memory type has become a bottleneck in AI accelerator production due to high demand, with leading supplier SK Hynix Inc. largely sold out through 2025. Huang did not provide detailed specifications for the Rubin platform, which is set to succeed Blackwell. AMD Focusing on AI Workloads Not just NVIDIA, during the opening keynote at Computex 2024, AMD Chair and CEO Lisa Su showcased the growing momentum of the AMD Instinct accelerator family. AMD unveiled a multiyear, expanded AMD Instinct accelerator roadmap, introducing an annual cadence of leadership AI performance and memory capabilities. In 2026, AMD plans to release the AMD Instinct MI400 series, based on the AMD CDNA “Next” architecture, which will provide the latest features and capabilities to enhance performance and efficiency for AI training and inference. Previewed at Computex, the 5th Gen AMD EPYC processors, codenamed “Turin”, will utilise the “Zen 5” core, continuing the high performance and efficiency of the AMD EPYC processor family. These processors are expected to be available in the second half of 2024. The roadmap begins with the AMD Instinct MI325X accelerator, set to be available in Q4 2024. This accelerator will feature 288GB of HBM3E memory and 6 terabytes per second of memory bandwidth, using the same Universal Baseboard design as the MI300 series. It boasts industry-leading memory capacity and bandwidth, being 2x and 1.3x better than the competition, respectively, and offering 1.3x better compute performance. Following this, the AMD Instinct MI350 series, powered by the new AMD CDNA 4 architecture, is expected in 2025. It promises up to a 35x increase in AI inference performance compared to the MI300 series with CDNA 3 architecture. The AMD Instinct MI350X accelerator will be the first product in this series, utilising advanced 3 nm process technology, supporting FP4 and FP6 AI data types, and including up to 288 GB of HBM3E memory.","excerpt":"Both companies have roadmaps till 2026 for AI accelerators.","categories":["AI News"],"tags":["ai announcements"],"author_name":"Mohit Pandey","publish_date":"2024-06-03T10:05:51","publication_year":"2024","word_count":470,"keywords":["ai announcements","R","programming_languages:R","AI"],"extracted_tech_keywords":["AI","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-unveils-rubin-months-ahead-of-blackwell-release-amd-announces-mi400x\/","complexity_score":3,"technical_depth":3,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10083729,"title":"&#8216;Toxic&#8217; Linux Community Acts As the Perfect Repellent for New Users","content":"Note: The term Linux usually refers to the Linux kernel, but this article uses it to refer to operating systems built on Linux kernel. An operating system comprises more than just a kernel; here we speak of Linux as an OS for brevity. Due to Windows and MacOSX’s penetration into the mainstream market, Linux has fallen to the wayside in terms of adoption. But that hasn’t stopped enthusiasts from creating strong communities for the OS online, most of whom swear by Linux for all their tasks, conveniently ignoring the problems with their ‘miracle distros’. Linux community vs Microsoft and Apple For hardcore Linux users, Microsoft represents everything wrong in personal computing. From forced updates to Windows’ NTFS formatting system, everything is fair game for these self-proclaimed power users. When not spouting baseless paranoia about how Microsoft is tracking everything you do on your ‘inferior’ Windows PC, they’re arguing amongst themselves on why their distro is better than the other. Any user fed up with Windows and looking to shift to Linux might have moderate success installing and setting up one of the easier-to-use distros out on the internet. However, one of the biggest obstacles that any new Linux user will find difficulty in surmounting is the toxic community the ecosystem has built around itself. User Prize_Barracuda_5060 on the Linux subreddit accurately captured this sentiment in their post, stating: “I tried using Linux on my main computer as a beginner and faced several issues. But when asking the community or searching the internet, I would only find hate comments by the so-called spawns of arch and tiling window managers rather than solutions to my problems and ways to prevent it from happening in the future.” Linux power users have a history of telling new users to just ‘RTFM’, an oft-repeated phrase that translates to ‘Read the F***ing Manual’. Newcomers have often had this thrown at them whenever they ask any basic queries regarding the OS. But, this is just one of the few ‘unpleasant phrases’ that newbies hear all the time when it comes to Linux tech support. Norman Jobling, a self-proclaimed Linux user, encapsulated this sentiment perfectly, stating: “I think there’s this ‘you should already know that’ and ‘work it out yourself or you’re stupid’ mentality among a lot of the Linux purists and that’s why it will never be a mainstream desktop OS.” Linux community vs…Linux community? The communities that each distro has collected over the years have collectively become one of the biggest repellents for new users, along with being a point of contention for those actively involved in using Linux OSs. Even members of the community have called out this fanboyish behaviour, where Linux nerds regularly split themselves apart on how their specific set of distros and utilities is better than every other option out there solely because it suits their use cases. These users have had the spare time to compile their list of preferred softwares, debugging and fixing issues at every step. This leads to a sense of technical superiority and extreme tone deafness when it comes to things that people actually need, like an intuitive GUI or an OS that just works. The Linux community thrives on belittling others for their choices while trying to convince them that their set of choices is better, solely due to the amount of time they spend to set up their system in the exact way they want it. According to these enthusiasts, using any OS apart from the free and open-source Linux is the highest form of insult, as the user is supporting ‘greedy corporations’ by using closed-source operating systems. Even though Windows and Mac trump Linux in terms of usability, support, and stability, the toxic community disregards these factors completely and forces Linux on users unwilling to use it. Although, these bad apples do not make a majority of the community surrounding Linux, since many forums are actually friendly towards new users, the negative users create enough noise in the community to turn newcomers away from the platform, thus creating a negative perception of the ecosystem.","excerpt":"Even though Windows and Mac trump Linux in terms of usability, support, and stability, the toxic community disregards these factors completely and forces Linux on users reluctant to use it","categories":["AI Features"],"tags":["Linux Kernel","linux mint","Linux Ubuntu"],"author_name":"Anirudh VK","publish_date":"2023-01-02T11:00:00","publication_year":"2023","word_count":680,"keywords":["CUDA","Linux Kernel","Linux Ubuntu","AI","programming_languages:R","Aim","ViT","linux mint","R"],"extracted_tech_keywords":["AI","Aim","CUDA","R","CUDA","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/toxic-linux-community-acting-as-the-perfect-repellent-for-new-users\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61248,"title":"Report: COVID-19 Is Going To Affect The Data Centre Market In Southeast Asia","content":"A recent report stated that the data centre market in Southeast Asia is expected to grow at a CAGR of over 6% during the period 2019–2025. According to a report, COVID-19 is going to affect the data centre market in Southeast Asia, which is witnessing growth due to the increased interest from giant cloud providers such as Google, AWS, and Alibaba to open cloud regions. The report clarified that the increasing adoption of cloud-based services would be a key driver for the data centre market in the coming years. Alongside, the rising penetration of the internet is likely to aid the use of smart devices in this region. Further, it stated that the impact of emerging technologies like big data, IoT, artificial intelligence, and virtual reality would also play a significant role in affecting data centre market growth in other southeast countries after 2020. In fact, the majority of colocation data centre providers are involved in the construction of hyperscale data centres to colocate space to cloud service providers. Case in point, in Southeast Asia, Singapore is a mature market and has been accounted for as the primary revenue generator of the APAC region. The list is then followed by Indonesia, Malaysia, Thailand, and Vietnam. The report further noted that last year, Google had announced its expansion of one of its Singapore data centres, and will implement 5G technology. According to the report, the incentives from government agencies will be highly beneficial for continuous investment from both local and global data centre developers. Alongside, countries like Thailand, Malaysia, and Indonesia are also experiencing a growth in the wholesale colocation of hyperscale spaces in the market. The report also noted some of the critical factors that are contributing to the growth of the data centre market in Southeast Asia during the forecast period, such as the deployment of 5G in order to increase edge data centre investments; increasing adoption of big data and IoT; availability of lithium-ion batteries and fuel cells; and adoption of renewable energy in data centres. The report also stated that although the data centre market in Southeast Asia is still coming onto its existence, it has a strong potential for growth as many enterprises are migrating from server rooms to cloud or colocation facilities. Cisco, Dell, HPE, Huawei, Fujitsu, NetApp, Lenovo, and IBM are a few of the companies that have a strong presence in the data centre market. Some of the critical pointers the report has highlighted are: The adoption of storage systems has started to shift towards all-flash storage solutions. The penetration of technologies like cloud computing, big data, and IoT will be the keydriver for the growth of the Southeast Asia data centre market. Alongside, the adoption of organization-specific software over the cloud platform will also increase the demand for high computing servers. The adoption of x86 based servers is more common in the market. Also, Dell EMC, HPE, IBM, Lenovo, Fijustu are some of the major server vendors. The market for storage drives has been increasing from the last five years. The implementation of the 5G network will boost the digital economy and will increase the demand for high- bandwidth networking infrastructure.","excerpt":"A recent report stated that the data centre market in Southeast Asia is expected to grow at a CAGR of over 6% during the period 2019–2025. According to a report, COVID-19 is going to affect the data centre market in Southeast Asia, which is witnessing growth due to the increased interest from giant cloud providers […]","categories":["Deep Tech"],"tags":["Coronavirus","covid-19","data centre","data centre india","data centres","Lenovo"],"author_name":"Sejuti Das","publish_date":"2020-04-08T13:29:38","publication_year":"2020","word_count":529,"keywords":["big data","data centre india","Go","artificial intelligence","covid-19","AI","cloud computing","AWS","Git","RAG","Coronavirus","data centres","Lenovo","GAN","R","data centre"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","cloud computing","AWS","R","Go","Git","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/report-covid-19-is-going-to-affect-the-data-centre-market-in-southeast-asia\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169808,"title":"How OpenAI’s New ‘CEO of Apps’ May Disrupt the AI Ecosystem","content":"Recently, OpenAI added a new role to its C-suite by appointing Fidji Simo, a current board member of the company as the ‘CEO of Applications’. She is expected to join the company later this year, after transitioning away as the CEO of Instacart, a grocery delivery company based in San Francisco. Simo will report directly to OpenAI’s CEO, Sam Altman, who issued a statement saying, this move will allow him to focus more on research, safety, and infrastructure. Altman, however, stated that he will continue to directly oversee success across all pillars of OpenAI – Research, Compute, and Applications – ensuring we stay aligned and integrated across all areas. “ I will work closely with our board on making sure our non-profit has maximum positive impact,” Altman said. While OpenAI has undergone key leadership changes in the recent past, analysts believe that Simo’s appointment indicates the company may expand its focus towards advertisements and the application layer, which can be decisive in increasing revenue. Will ChatGPT Get Advertisements Soon? Simo’s work experience, especially the ones that had the most impact on the establishments she worked at, indicates where she may be headed with OpenAI. Sequoia, an investor in OpenAI, said in a February 2024 blog post that Simo worked to ‘pivot the company towards profitability’ at Instacart. Following her appointment as CEO in 2021, Instacart was said to have had more than 5,500 advertising partners last year. “Instacart reported that ad campaigns drive an average 15% incremental sales lift—an almost unheard-of number,” read the blog post. Her role at Meta (formerly Facebook) also essentially involved the advertising space. For instance, in 2013, she led the team in charge of monetising mobile and growing advertisements on Facebook’s news feed, which happens to be the app’s majority revenue source. OpenAI has hinted in the past that it will integrate ads. In an interview with the Financial Times in December last, CFO Sarah Friar said the company is considering an advertising model and wants to implement it ‘thoughtfully’. The Financial Times also added that, according to sources and their independent analysis, OpenAI is hiring advertising talent from Meta and Google. Source: The Information In a Lex Friedman podcast a year ago, Altman expressed his dislike of integrating advertisements into products. And Simo’s appointment might counter that. Her appointment comes at a time when multiple studies show AI-driven platforms rapidly eclipsing traditional websites as the public’s primary source of information. This trend threatens the foundation of publishers and ad-supported sites that depend on page views for revenue. As more users bypass conventional articles for instant answers from AI apps like ChatGPT, advertisers may naturally follow where most of their audience lies. Frederick Vallaeys, CEO of Optymyzr, took to LinkedIn to say that Altman’s stance on ads is similar to that of Larry Page and Sergey Brin, Google’s co-founders, who believed that advertising-funded search engines would be inherently biased toward the advertisers. But today, advertising is a significant revenue generator for Google. “We’ve seen this before: criticise ads until you figure out how to make them fit in with your mission,” said Vallaeys. He further explained that when people make specific requests online, it creates a valuable opportunity for advertisers. OpenAI has built a system where these clear intent signals are even more helpful than the keywords that Google relied on initially for advertisements. While ChatGPT has over 500 million weekly active users, integrating advertisements is indeed a promising opportunity to drive revenue from non-paying users. To sustain its success, industry analysts believe OpenAI should onboard more users to its platform and increase the capabilities of its applications. Currently, research from Sacra suggest that OpenAI makes more revenue from ChatGPT, than the enterprise sector. AIM reached out to OpenAI seeking a response on Simo’s potential impact on the company’s advertising, and product roadmap, but did not elicit a response. Will OpenAI Own the Application Layer Soon? While many argue that large establishments, or AI incumbents, may not be able to meet all use case demands, thus securing the position of startups, OpenAI might accelerate its focus on an increasing number of use cases, with a ‘CEO of applications.’ For instance, the native image generation feature with GPT-4o was overwhelmingly successful. In the first week of launch, OpenAI said that over 130 million users worldwide generated over 700 million images. The feature was also made available to free users. Over the years, ChatGPT has also been updated with more features that were once only available with other dedicated apps built by startups. For instance, the application was recently updated with improved search results related to shopping, a feature highly touted by apps like Perplexity. Additionally, the release of the o3 and o4-mini models, along with the Codex CLI agent command line tool, enhanced the coding capabilities available through OpenAI’s products. However, in a key development, reports indicate that OpenAI has reached an agreement to acquire the AI-enabled coding platform Windsurf for $3 billion. The acquisition is expected to enhance OpenAI’s products with superior AI coding capabilities. Furthermore, OpenAI is also reportedly working on a social media platform, which may directly rival the one released by Meta. With ChatGPT’s existing capabilities, its position as the most popular AI app, a steady stream of new feature releases, an upcoming major acquisition, and the appointment of a dedicated CEO for its apps, it might now move into markets it once avoided—posing a serious threat to startups building products in those spaces. This may also contradict what Altman once said. In an interview last year, he said that as the company continues to enhance its models, founders who build startups that leverage these capabilities will likely continue to thrive. He emphasised that developers who avoid competing with the core functionality of the AI model are more likely to remain successful. Anish Acharya, an investor at Andreessen Horowitz (a16), said on X that, given that OpenAI provides the most powerful AI models today, the company is essentially aiming to capture all of the ‘economics in the ecosystem’ to achieve its goal of artificial general intelligence, or AGI. “We’re now living in a world where consumption layer apps are multiplexing and routing to different foundation models, which is disadvantageous for (foundation models) suppliers [like OpenAI] and limits their ability to extract down-funnel economics,” said Acharya. “In that world, OpenAI must vertically integrate and own the consumption layer to protect their economics, thus the recent focus on AI Applications,” he added.","excerpt":"Analysts believe that Fidji Simo’s appointment indicates the company may expand its focus towards advertisements and the application layer.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","OpenAI"],"author_name":"Supreeth Koundinya","publish_date":"2025-05-12T19:05:34","publication_year":"2025","word_count":1078,"keywords":["Go","ChatGPT","API","OpenAI","AI","GPT-4o","RAG","Aim","foundation models","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","foundation models","GPT-4o","ChatGPT","OpenAI","Aim","RAG","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-openais-new-ceo-of-apps-may-disrupt-the-ai-ecosystem\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10130164,"title":"Top 8 Epic Tech Outages and What Went Wrong","content":"According to the Uptime Institute’s 2024 Outage Analysis1, between 10 and 20 “high-profile IT outages or data centre events” occur every year. The study revealed that while power is the main cause of data centre outages, network issues are the leading cause of outages across all IT services. These outages make headlines and have serious consequences, disrupt business for customers, and damage company reputations. More than half of the respondents said their most recent major outage cost them over $100,000, and 16% reported that it cost them over $1 million. Additionally, the report mentions the leading causes of network outages, including design and configuration, hardware, capacity, software, and environmental threats. Here are eight major tech outages to be explored. 1. Microsoft-CrowdStrike Last week, CrowdStrike, a security technology provider, caused a massive global IT outage, potentially the biggest in history, affecting airlines, banks, businesses, schools, and government services worldwide. The CrowdStrike Outage occurred due to a faulty software update in their Falcon sensor program, which caused widespread disruptions to Windows systems globally. This led to the infamous “Blue Screen of Death” and reboot loops for millions of users. Excluding Microsoft, US Fortune 500 companies are said to face $5.4 billion in financial losses due to the Windows outage. 2. Meta Outage Lasting Nearly Six Hours On October 4, 2021, Meta platforms, including Facebook, Instagram and WhatsApp, experienced an outage lasting nearly six hours. Users faced difficulties accessing the apps, leading to a surge in traffic on competing platforms like Twitter and TikTok. During this period, Facebook reportedly lost about $545,000 in US ad revenue per hour. 3. Google Services Down for an Hour Popular Google services such as YouTube, Gmail, Google Drive, and Google Docs were down for an hour, affecting millions of users worldwide on December 14, 2020. The outage was attributed to a failure in Google’s authentication system, which manages user logins across its services. The issue specifically stemmed from an internal storage quota problem. Users attempting to access these platforms encountered errors, with many reporting that they were unable to log in or retrieve their data. Google acknowledged the issue and confirmed that the services were restored for the vast majority of affected users shortly after the outage. 4. Fastly Disrupted Numerous High-profile Websites On June 8, 2021, Fastly, a major content delivery network (CDN) provider, experienced a significant global outage that disrupted numerous high-profile websites, including Amazon, Reddit, and The New York Times. The outage was triggered by a software bug that had been introduced during a deployment on May 12, which remained dormant until a valid configuration change made by a customer activated it. This led to 85% of Fastly’s network returning errors, resulting in widespread accessibility issues for many internet users around the world. Read more at: Akamai Bets on Edge Computing to Take on AWS, Azure and Google Cloud 5. Twitter (X Corp) Down for Several Hours Twitter suffered a major outage on December 28, 2022, leaving tens of thousands of users unable to access the platform or its features for several hours. It primarily impacted users attempting to access the platform via desktop computers. Many reported being unexpectedly logged out, encountering error messages, and facing difficulties in viewing replies or using features like notifications and TweetDeck. The hashtag #TwitterDown trended on the platform as users shared their experiences during the outage. 6. AWS Disrupted Businesses and Applications On December 7, 2021, Amazon Web Services (AWS) experienced a significant outage that disrupted numerous services and affected a wide range of businesses and applications. It primarily impacted the US-East-1 region, located in Northern Virginia, which is crucial for many of AWS’s services. The outage was caused by an automated scaling activity designed to increase capacity for service within AWS’s main network. This action unintentionally triggered a surge in connection attempts within AWS’s internal network, overwhelming the devices managing communication between the internal and main networks.Read more at: Will AWS’ Slump In Revenue Affect Its Growth In India? 7. Akamai Affected Numerous Financial Institutions and Airlines On June 17, 2021, a significant disruption occurred at Akamai, affecting the websites of numerous financial institutions and airlines in Australia and the United States. This outage was traced back to server-related glitches at Akamai, a major content delivery network (CDN) provider. The incident marked the second major internet blackout within a week, following a prior outage caused by a rival CDN, Fastly Inc. Akamai attributed the outage to a bug in its software, which was promptly addressed. The company confirmed that the issue was not related to any cyber-attack or security vulnerability. Read: Around 42% of Overall Web Traffic is Generated by Bots: Report 8. Cloudflare Down for Two Days A power failure led to Cloudflare coming down for around two days. The platform uses the services of three data centres. One such data centre experienced a power failure. The outage was caused by a failure of the facility’s generators and faulty circuit breakers. As the generators failed, Cloudflare’s network routers lost power, which disrupted services reliant on the PDX-04 data centre. The outage primarily affected Cloudflare’s dashboard, APIs, and related services, while traffic through its global network continued to function without interruption. https:\/\/uptimeinstitute.com\/resources\/research-and-reports\/annual-outage-analysis-2024 ↩︎","excerpt":"Outages are often caused due to network issues, including design\/configuration, hardware, capacity, software, and environmental threats.","categories":["AI Trends"],"tags":["Google"],"author_name":"Gopika Raj","publish_date":"2024-07-25T14:59:27","publication_year":"2024","word_count":866,"keywords":["Go","API","AWS","AI","R","RAG","ViT","Google","disruption","edge computing","Azure"],"extracted_tech_keywords":["AI","RAG","AWS","Azure","edge computing","R","Go","API","ViT","disruption"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/tech-meltdowns-8-epic-outages-and-what-went-wrong\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10011910,"title":"Guide To Playment &#8211; A Leading Data Labeling Platform for Image, Video and Sensors","content":"Data Annotation is not only limited to image and video but now with recent advances in computer vision, it is extended to sensors even. Many of us must be aware of image and video annotation techniques(if not you can go through these articles- on SuperAnnotate and LabelBox) such as bounding boxes, pixel-level accuracy, polygons, and many more such tools which have now enhanced the state-of-art. Sensor-fusion or Lidar annotation is another such technique. Lidar(Light Detection and Ranging) which measures distances using sensors between objects by illuminating them using a pulsed laser. This is used in many research applications by scientists. Today we will be talking about Playment, a data annotator tool which allows all of this possible with workflow management. What is Playment? Playment is a complete data labelling platform generating training data for computer vision and machine learning models at scale build high-quality ground truth datasets. Playment was launched in November 2015 by Siddharth Mall, Ajinkya Malasane, Akshay Lal. It is headquartered at Bangalore, Karnataka, India. Playment has a large community of trained annotators who have work distributed among them in the form of micro-tasks, formally known as microwork(this kind of work breaks down a large unified project into small tasks) that is done by people all over the internet. Annotators go through the platform and complete pending tasks and get points. The points can be exchanged in the form of vouchers on online e-commerce sites. The platform is powered by a workforce of 300,000+ users which is managed by the human intelligence experts who build the tasks and deliver results with assured quality. It provides realtime ML-assisted pipelines with API support. Features: Image Annotation Auto labelling facility among the 80 common classes present. The different tools available are 2D Bounding boxes, Semantic Segmentation, Cuboids, Polygons, Polylines, Landmarks. Cuboids Polylines Polygons 3D Semantic Segmentation https:\/\/www.youtube.com\/watch?v=yUPW9HrzkC4&feature=emb_title Video Annotation Provides 2D bounding boxes, Cuboids(image), Cuboids(LiDar), Polylines, landmarks Sensor Fusion\/Lidar Annotation https:\/\/youtu.be\/ovgqp399_o8 3D point cloud annotation https:\/\/www.youtube.com\/watch?v=_D-i4TPv60k GT Studio This dashboard allows us to set up and monitor customized workflows and build an end to end project management with playment workforce. ML engineers can use Python code to integrate their pipelines. Creation of Multiple user groups. Project progress and quality analytics with active feedback are also provided keyboard shortcuts, visualization tools and confusion metrics. Use Cases: Deep learning assisted Semantic Instance Segmentation for Autonomous Vehicles Autonomous vehicles lane detection and driveable area.Full Pixel Segmentation Human Pose Estimation and TrackingDrone, CCTV and satellite surveillance Damage detection for car insuranceFashion, Gaming and Agriculture industry Companies: They empower high precision annotation services with complex customizations. Some of these companies and research institutions like Drive AI, Starsky Robotics, CYNGN, and UIUC, HELLA, Vayavision, INNOVIZ, Nuro (Building perception systems for robots), Daimler, Samsung, LG, Sony, Intel, Postmates, Siemens, Ouster(LiDar), AI AImotive, Alibaba, SAIC Motors, Continental.","excerpt":"Playment is a complete data labelling platform generating training data for computer vision and machine learning models at scale build high-quality ground truth datasets.","categories":["Deep Tech"],"tags":["annotation","data annotation","data labelling","data visualization python","image","mba in data analytics","PowerBI","python data visualization","sensor","Video"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-11-21T13:00:42","publication_year":"2020","word_count":468,"keywords":["computer vision","PowerBI","deep learning","python data visualization","Video","R","annotation","data labelling","data annotation","analytics","Go","machine learning","mba in data analytics","AI","ML","image","data visualization python","Python","sensor","Aim"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","analytics","Aim","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-playment-a-leading-data-labeling-platform-for-image-video-and-sensors\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":21098,"title":"Reliance Jio To Enter Indian IoT Market","content":"After its shock and awe campaign in the telecomm industry with its extremely low priced mobile and data services, and now the announcement of its foray into cryptocurrency,  Reliance Jio Infocomm is gearing up to conquer the Internet of Things (IoT) space. According to reports, the Reliance Industries subsidiary has brought onboard Ayush Sharma as the senior vice president of engineering and technology to drive accomplishments in this space. Before joining Reliance Jio, Sharma operated his own venture, MotoJeannie in the US. He has also worked with companies such as Huawei, Ericsson and Cisco in the past. With Sharma’s involvement, Jio aims to build business around IoT and disruptive technologies such as mobile edge computing, blockchain and artificial intelligence. “Jio is looking at these technologies to enable world’s largest programmable network with alternate technologies available,” Sharma said. He has already started assembling a team of engineers to innovate and utilise these technologies. The company aims to introduce IoT products and solutions for regular consumers as well as industries and enterprises. However, the immediate focus is on the latter. “It will take at least a year to enable consumer IoT, but the large focus is on enterprise IoT. We are working on specific use cases,” said Sharma. He also added that Reliance Industries seeks to employ these technologies in its in-house verticals such as logistics and retail in order to develop their intelligence. The bandwidth and latency required for enterprise and industrial use cases of IoT is expected to be the supplemented by the company’s 4G Network. According to Sharma, Jio is building its private platform backed by big data. To create a comprehensive network for consumer IoT, Jio is said to be in the process of enrolling car manufacturers, consumer appliances and durables maker and so on, and is also collaborating with vendors of various technologies. Just like its closest competitor, Bharti Airtel, Jio is foraying into home automation services as well. It is reportedly planning to introduce its own line of home surveillance systems. According a recent study, the home automation market is expected to grow at a compound annual growth rate of 14.5% between 2017 to 2022, and generate nearly $54 billion in revenues by 2022. This report comes within a month after the announcement of the company’s plan to create is own cryptocurrency, JioCoin. Open Source is said to be playing a key role in both JioCoin as well as the IoT initiative.","excerpt":"After its shock and awe campaign in the telecomm industry with its extremely low priced mobile and data services, and now the announcement of its foray into cryptocurrency,  Reliance Jio Infocomm is gearing up to conquer the Internet of Things (IoT) space. According to reports, the Reliance Industries subsidiary has brought onboard Ayush Sharma as the senior […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Blockchain","Cryptocurrency","IoT India","Jio","Reliance","Reliance Industries"],"author_name":"Jeevan Biswas","publish_date":"2018-01-30T09:23:49","publication_year":"2018","word_count":406,"keywords":["Jio","big data","artificial intelligence","Blockchain","programming_languages:R","IoT India","Cryptocurrency","AI","automation","Ray","Aim","edge computing","Reliance","R","Reliance Industries","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","Ray","edge computing","R","big data","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/reliance-jio-iot-ayush-sharma\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10049237,"title":"Shailesh Kumar","content":"Dr Shailesh Kumar has over twenty years of research, academic, entrepreneurship, and industrial experience in building innovative solutions and platforms across various Industrial and Societal verticals, including Telecom, Retail, Banking, Fleet Management, Insurance, Manufacturing, Education, Healthcare, Agriculture, Genomics, etc. He has published over 20 conference papers, journal papers, and book chapters and holds more than 20 patents in these areas. His current interests are in building large-scale, real-time AI Platforms for Complex Ecosystems spanning these societal and industrial verticals to bring the benefits of AI to the masses. Dr Kumar is currently the Chief Data Scientist at AICoE \/ Reliance Jio, Mentor and Program Chair for the AI & DS Program at the Jio Institute of Eminence, and visiting faculty of Machine Learning at Indian School of Business. Prior to this, he served as a Distinguished Scientist at Ola, Co-Founder at ThirdLeap, Researcher in the Google Brain team, Principal scientist at Microsoft Bing, Senior Scientist at Yahoo! Labs, and Principal Scientist at Fair Isaac Research. Dr Kumar was recognised as one of the Top 10 data scientists in India in 2015, top 50 Analytics Leaders in India in 2018, and top 10 most influential Analytics leaders in India in 2020 by the Analytics Science Magazine. Dr Kumar received his PhD and Masters in Computer Science from the University of Texas at Austin and B.Tech. in Computer Science from IIT-Varanasi. LinkedIn","excerpt":"Dr Shailesh Kumar has over twenty years of research, academic, entrepreneurship, and industrial experience in building innovative solutions and platforms across various Industrial and Societal verticals, including Telecom, Retail, Banking, Fleet Management, Insurance, Manufacturing, Education, Healthcare, Agriculture, Genomics, etc. He has published over 20 conference papers, journal papers, and book chapters and holds more than […]","categories":["AI Features"],"tags":["Interviews and Discussions","reliance jio"],"author_name":"AIM Media House","publish_date":"2021-09-22T13:13:55","publication_year":"2021","word_count":230,"keywords":["Go","machine learning","programming_languages:R","AI","programming_languages:Go","analytics","reliance jio","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dr-shailesh-kumar\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10168454,"title":"OpenAI Interested in Buying Google Chrome, Says ChatGPT Product Head","content":"OpenAI, the company behind the ChatGPT family of AI models, has revealed that it is interested in purchasing Google Chrome. Nick Turley, ChatGPT’s head of product, reportedly testified to this during a hearing regarding Google’s antitrust case brought by the US Department of Justice (DOJ) on Tuesday. The antitrust case against Google aims to dismantle the company’s monopoly in the search engine ecosystem. The lawsuit is advocating against Google entering into agreements with original equipment manufacturers (OEMs) and carriers to pre-install their services. Furthermore, the DOJ also proposed that Google sell its Chrome browser, as it is deemed a huge gateway to its search products. Thus, if it remains a separate entity, it could create an opportunity to include other search providers as the default option. Perplexity, the AI search engine company, was also asked to testify. The company released a statement saying the best option is to allow people to choose a search product for themselves, instead of breaking up Google by suggesting it sell Chrome. “Consumers deserve the best products, not just the ones that pay the most for placement. This is the only remedy that ensures consumer choice can determine the winners,” Perplexity said. “And for OEMs and carriers, you shouldn’t have to pick a side.” Regarding the separation of Chrome, Google stated that it would introduce cybersecurity and national security risks, as it would disconnect the browser from the company’s technical infrastructure. “DOJ’s proposal to split off Chrome and Android, which we built at great cost over many years and make available for free, would break those platforms, hurt businesses built on them, and undermine security,” Google said. Furthermore, Turley also revealed that Google declined an offer from OpenAI concerning the use of Google’s services in ChatGPT. OpenAI was reportedly “experiencing issues with its own search providers” and turned to Google for help.","excerpt":"OpenAI, the company behind the ChatGPT family of AI models, has revealed that it is interested in purchasing Google Chrome. Nick Turley, ChatGPT’s head of product, reportedly testified to this during a hearing regarding Google’s antitrust case brought by the US Department of Justice (DOJ) on Tuesday.  The antitrust case against Google aims to dismantle […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Google Chrome","OpenAI"],"author_name":"Supreeth Koundinya","publish_date":"2025-04-23T11:24:06","publication_year":"2025","word_count":307,"keywords":["Go","ChatGPT","OpenAI","AI","AWS","Google Chrome","GPT","Aim","Rust","R","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","AWS","R","Go","Rust","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-interested-in-buying-google-chrome-says-chatgpt-product-head\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10060114,"title":"Israeli Defence Forces disclose their AI war strategy","content":"The Israel Defence Forces has disclosed their AI war data factory and strategy at Tel Aviv University’s Blavatnik Virtual AI week. The speed at which a new weapon can be created using Data and AI is totally different when compared with physical jets and submarines,  Brig General Aviad Dagan, Director of the IDF’s Digital Transformation Administration, said. “Data and AI can actually win wars… not only arms, physical jets and submarines,” he said. IDF has been at the forefront of using technology to improve coordination among different units deployed on various frontlines. IDF currently uses digital networking in the cloud between all forces – from headquarters to frontline command centres to troops in the field. Israeli military plans to enhance its edge architecture by building mini-clouds or networks for its arms and smaller subdivisions. This will help the units to process and receive data at faster rates compared to the current network. A wide variety of data points gathered by sensory detection of an enemy will provide the IDF with the ability to evaluate who the enemy was, check the various options to respond to the threat, analyse how much fuel is needed by the drones or other units- and then quickly dispatch the most viable targeting order, Dagan said.","excerpt":"IDF has been at the forefront of using technology to improve coordination among different units deployed on various frontlines.","categories":["AI News"],"tags":["strategy"],"author_name":"SharathKumar Nair","publish_date":"2022-02-08T17:31:47","publication_year":"2022","word_count":210,"keywords":["programming_languages:R","AI","digital transformation","Git","strategy","GAN","R"],"extracted_tech_keywords":["AI","R","Git","GAN","digital transformation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/israeli-defence-forces-disclose-their-ai-war-strategy\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10008071,"title":"The All-New MachineHack Offers Faster &#038; Robust Platform For Data Science Hackathons","content":"MachineHack platform just got faster and robust while promising to host the best hackathons on data science, machine learning and artificial intelligence. MachineHack, which is an ambitious hackathon platform by Analytics India Magazine recently launched its all-new and revamped website that offers a new interface with new improved features providing an awesome user experience to the participants. MachineHack was conceptualised almost 2.5 years ago with an aim to provide an online platform to data science and machine learning enthusiasts with the best-in-class hackathons to keep up with the latest skills and trends in the field. It has provided a platform to over 20,000 ML enthusiasts so far to compete and have hands-on experience with data science tools while experiencing the near real-life business use cases. The new platform will allow practitioners to test and practise machine learning skills while competing against hundreds of data scientists, with the industry-curated hackathons. The MachineHack team has been working tirelessly over the last few months to bring the best user experience for the participants. Some of the features that participants will see in the new version of the website are: Global Leaderboard: The revamped website will host a global leaderboard or ranking displaying the name, rank, participation and final rank of all the participants. It shows since when a participant has been a user at MachineHack while placing them rank-wise inculcating a competitive spirit.  Exclusive Profile for every participant to showcase as a virtual portfolio: This is a public profile, where each participant’s profile is managed in a way that can be used to showcase the skills to potential hiring managers or companies.New Ranking Tiers: There will be 5 tiers of expertise on MachineHack — Grandmaster, master, champion, apprentice and novice — which will be achieved based on the quality of work in the competitions. The progression system will keep the community members to compete regularly to prove their mettle. The MachineHack global ranking page is updated each week where points awarded to every rank holder is based on the competition they have participated in. While the platform is still in its early stages, there will be more features that will be added in the coming weeks for budding machine learning developers to hone their skills. It currently has more than 25 hackathons that one can compete on. Experience the platform here.","excerpt":"MachineHack platform just got faster and robust while promising to host the best hackathons on data science, machine learning and artificial intelligence. MachineHack, which is an ambitious hackathon platform by Analytics India Magazine recently launched its all-new and revamped website that offers a new interface with new improved features providing an awesome user experience to […]","categories":["Deep Tech"],"tags":["data science hackathon","data science hackathon india","Data Science Hackathons","Machinehack","Machinehack Hackathon"],"author_name":"Srishti Deoras","publish_date":"2020-09-22T10:37:58","publication_year":"2020","word_count":388,"keywords":["data science","Go","artificial intelligence","machine learning","programming_languages:R","Machinehack","AI","ML","data science hackathon","Data Science Hackathons","Aim","analytics","Machinehack Hackathon","data science hackathon india","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","R","Go","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/machinehack-makes-a-comeback-with-a-faster-robust-platform-for-data-science-hackathons\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10086934,"title":"Code Readability vs Performance: Here is The Verdict","content":"A talented software developer, Jack, has the ability to write clean and efficient code, but he’s facing a new challenge: balancing the need for code readability and code performance. On the one hand, he wants to write easily understandable code that is easy to maintain, but on the other hand, he knows that his code needs to run fast and efficiently. Jack developed a systematic approach to finding the balance between code readability and performance wherein he starts by understanding the requirements of the project and defining the scope. Based on this, he assesses the size and complexity of the code and determines the level of collaboration required from other developers. He then decides which aspect, readability or performance, should take priority for that particular project. When Jack prioritises code readability, he takes a human-centred approach to coding. He uses descriptive variable names, proper indentation, and comments to ensure that the code is easy to understand and maintain. He believes that code that is easy to understand is more likely to be correct and less prone to bugs, saving time and resources in the long run. However, when Jack prioritises code performance, he takes a more technical approach. He focuses on optimising the code for speed and efficiency, even if it means compromising on readability. For example, he may use abbreviated variable names or forego comments to reduce the amount of code and make it run faster. Jack understands that code performance is critical for projects that require high-speed computation, real-time data processing, or other resource-intensive tasks. But are there any principles that developers like Jack can follow without juggling between performance and readability of code? Moreover, is there any reason that developers would choose one over the other? SOLID, DRY, KISS Just like any other field of science and engineering, programming also has some basic principles that govern good software engineering. The KISS or ‘Keep It Simple, Stupid’ principle is to remind all developers that application design and deployment should be as less complex as possible. The procedure should be easily understandable and easy for other developers to debug whenever required. This ensures that the code is maintainable in the future. Next principle is called DRY or ‘Don’t Repeat Yourself’. It means exactly the same. Every developer should aim to reduce repetition of information within their work to avoid redundant code. This can be achieved by converting and segregating the entire system into smaller fragments for helping manage the code, and recalling it whenever necessary. SOLID principles come from and for design. These sound philosophical, but when implemented while coding, make the code easier for collaboration, modification, and troubleshooting for bugs and issues. Single responsibility principle means that every module should carry out a one responsibility, instead of complicating it. Open-closed principle states that any part of the system should be easily extendable by adding further features. Liskov substitution principle states that objects within a program should be replaceable with instances of their sub-category types without altering the accuracy of the program. Interface segregation principle suggests that it’s best to avoid adding new methods or functionalities into an existing interface. On the contrary, it’s better to incorporate a new interface and allow the classes to implement different interfaces depending on the need. Dependency inversion principle says that design patterns should be resolvable by using dependency injection. What’s the Verdict? Code performance is critical, especially when working on projects that require high-speed computation and real-time processing. This can result in slow and sluggish user experiences. But focusing on the performance of a code that is not readable is useless. Moreover it can also be prone to bugs and errors. Performance is a quirky thing. Starting to write a code with performance as the first priority is not a path that any developer would take, or even recommend. In a Reddit thread, a developer gives an example of a code that compiles in 1 millisecond, and the other code in 0.1 millisecond. No one can really notice the difference between both the models as long as the code is “fast enough”. So improving the performance and focusing on it, while sacrificing the readability of the code can be counterproductive. Moreover, in the same Reddit thread, another developer pointed out that writing faster algorithms actually requires you to write harder code oftentimes, which again sacrifices the readability. “Performant code relies on clever tricks,” the user said. High performance code is oftentimes not related to the objective that a developer has, but just focused on improving the performance. Making lookup tables, hashing stuff, or doing manipulations like building up a cache, just for making the operation speed up, is just divergence from actually trying to make the code work. So ideally, the approach, when writing code, should be that the code must work first, then make it good, and then worry if the code should be fast, or just “fast enough”. Speed is not as important as the programming process to be memory efficient. It becomes a concern when it becomes a concern.","excerpt":"Are there any principles that can help developers juggling between performance and readability of code? Moreover, is there any reason developers should choose one over the other?","categories":["Deep Tech"],"tags":["Developers","VS Code"],"author_name":"Mohit Pandey","publish_date":"2023-02-09T11:00:00","publication_year":"2023","word_count":840,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","Aim","VS Code","R","Developers"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/code-readability-vs-performance-here-is-the-verdict\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10115898,"title":"2024 is the Year of Embodied AI","content":"“2024 will be the year of Embodied AI,” said Brett Adcock, CEO and founder of Figure robotics towards the end of last year. Fast forward to today, there is a humanoid robot that can not only perform tasks but also talk to you! While we patiently wait for GPT-5, OpenAI’s collaboration with Figure is surely the first of its kind. Now everything is slowly but surely coming together. OpenAI models provide high-level visual and language intelligence. Figure neural networks, on the other hand, deliver fast, low-level, dexterous robot actions. Embodied AI When Adcock started the AI robot company, Figure, in 2022, he believed that he’ll have reliable humanoid hardware well ahead of reliable real-world neural nets running on robots. However, the recent advancements in AI have proven that they go hand-in-hand with the development of reliable humanoid robots. In their latest demo video, Figure 01 is not only able to achieve small tasks but they are also able to talk to you. With ChatGPT integration, the humanoid is able to converse on end-to-end neural networks, something that is not achieved by rival robot makers. “We’ll be racing to get reliable hardware, vast training sets, and designing manufacturing processes for billions of units. I can’t think of a more exciting future,” said Adcock, referring to the advancements he wants to bring to Figure. One of the notable investors of Figure, OpenAI chief Sam Altman has also been extremely optimistic about the future of robotics. “On the physical hardware side, there’s finally, for the first time that I’ve ever seen, really exciting new platforms being built,” he said. Adcock’s vision for robotics to bring an all ‘embodied’ robot that can execute low-level tasks and even converse with humans is quite different from the capabilities of Tesla’s Optimus or Boston Dynamics’ Atlas. Source: X Tesla’s swanky Optimus has been able to do impressive tasks including yoga. Atlas on the other hand was last seen somersaulting in a factory setting. The humanoid, H1 robot, developed by China’s Unitree Robotics, recently reached a walking speed of 7.4 miles per hour, and also claims to reach a speed of 11 mph. Calling 2024 as the year of robotics, developments in this area have been remarkable. Humanoids may be one form, but general purpose-robots are also making their way. Google DeepMind-led robotics projects, announced a couple of models, at the start of this year, that helps with achieving low and easy tasks. Advancements in Robotics Training Collecting robot data and training it still remains an arduous task, which is partly reflected in the pace of advancements in robotics. However, innovative learning methods for the same are also simultaneously developed, such as imitation learning. Chen Wang, a PhD student at Stanford is heading a project that enables wearable devices to collect robot data without the presence of an actual robot. DexCap, is a portable hand motion capture system that collects 3D data for training robots. https:\/\/twitter.com\/chenwang_j\/status\/1768304378577084739 Stanford professor and former researcher at Google DeepMind Karol Hausman, recently announced the launch of his new startup, Physical Intelligence. The company aims to build foundation models that can control any robot for any application. “We’ll focus on collecting robot data at a scale never seen before, making algorithmic advancements, training very large models and whatever else is needed to bring AI into the physical world,” said Hausman. Tesla’s Optimus is trained on proprietary systems, including the full self-driving (FSD) feature in Tesla’s cars. AGI Conversations, Again… An X user commenting on Figure 01’s demo video. Source: X Furthermore, with AGI conversations popping up every two weeks, the latest demo of Figure 01 has definitely raised a lot of questions about whether robotics is the path to AGI. With ‘Embodied AI’ as the theme for robotics, a fully en masse humanoid might be the route to AGI. Interestingly, Paolo Pirjanian, a robotist and entrepreneur, is the founder of a company named ‘Embodied, Inc’, which has created Moxie robots, an AI-powered robot built to support and interact with kids.","excerpt":"Figure 01 humanoids can perform tasks and converse like a person, embodying a human-AI fusion.","categories":["AI Features"],"tags":["Embodied AI","figure","OpenAI","Robotics"],"author_name":"Vandana Nair","publish_date":"2024-03-16T10:00:00","publication_year":"2024","word_count":666,"keywords":["Go","ChatGPT","OpenAI","AI","neural network","GPT-5","Embodied AI","Robotics","GPT","Aim","figure","foundation models","R"],"extracted_tech_keywords":["AI","neural network","foundation models","GPT-5","ChatGPT","OpenAI","Aim","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/2024-is-the-year-of-embodied-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10021959,"title":"The Plight Of Gig Workers In An Algorithm-Driven World","content":"According to a recent report published by the International Labour Organisation (ILO), digital labour platforms erode the worker’s rights and quality of life. The ILO report analysed the role of digital labour platforms in transforming the nature of work – highlighting the challenges gig economy workers face. The number of digital labour platforms has increased by five times from 2010 to 2020, of which 8% are in India. The platforms received a further boost due to the pandemic as millions got laid off and were looking for work to make ends meet. Below, we look at the issues gig workers face in an algorithm-driven world. Lack Of Transparency The data collected by digital labour platforms help organisations in enhancing operations, maximising performances and accelerating decision-making that eventually leads to better ROIs. However, the data has created a power imbalance between the companies collecting the data and the gig workers generating it. The generated data is entirely considered as the property of the platforms. The lack of transparency and restricted access to personal data leave gig workers with no leverage to take on these platforms in case of disputes. In sum, the house always wins. Regulations like the GDPR in Europe allow gig workers to get a copy of their data. Taxi drivers in the UK have filed a case against Uber for withholding data and violating GDPR. However, other countries do not have such strong laws to uphold workers’ rights. The accrual of data also leads to monopolies or ‘data-opolies’. For instance, Uber acquired a number of their competitors like Careem, Cornershop, Postmates, etc. Uber has exclusive access to the data from these companies raising concerns over the privacy rights of gig workers who worked for them. The lack of transparency with regards to the source code used in these algorithms is another pressing concern. The workers are in the dark about the underlying decision-making processes used to rate them. The dice is loaded against the workers when an algorithm acts up and shows them in a bad light as they do not have the resources to make their case. Meanwhile, accessing the underlying source code is problematic because it is protected under the trade secrecy laws and intellectual property rules under the WTO. The EU has found itself in a tricky situation as the e-commerce WTO deal poses future challenges to its digital policy objectives of adopting rules that will, for instance, mandate external audits for AI systems. Loss Of Autonomy This opaque ‘algorithmic management’ defines the everyday work experience, performance, and achievement of the workers, which is based on the data generated by these workers. There is a long history of companies using selective data to get off the hook in courts. The ILO survey uncovers the issues faced by gig economy workers, such as the platforms asking them to install specific software; the working hours getting tracked by clients; pressure to be available during specific times; getting monitored while working in some instances. Taxi-based or delivery apps monitor data to the point where they use GPS systems to define optimal routes to be taken by drivers as they carefully track the time spent for the ride or delivery. Secondly, the gig economy’s major attraction is supposedly flexible working hours that allow workers to decline orders or requests for valid reasons such as exhaustion, safety, or personal commitments. However, the ILO survey showed a sizable portion of the app-based taxi or delivery services could not cancel or refuse their orders as it might affect their ratings. Respondents in the survey even mentioned that they get very little time to decide whether to accept or decline any order. Uber drivers, for instance, have only 15 to 40 seconds to accept the ride based on limited information. Thirdly, these algorithms are over-reliant on ratings received from the clients. This is especially true when using online web-based platforms. Sometimes when workers do not get a rating, the platforms log it as incomplete work. Other times, gig workers also get unfair or fraudulent ratings that are not based on their work. Nevertheless, these ratings are factored into the algorithms. Fourthly, rejection of work is also common on these platforms. According to the ILO survey, there is a high rate of unfair rejections that indicate that work tends to be supervised by algorithms more than humans. Also, algorithms can be designed so that the tasks are approved on the majority of responses, independent of the correct response, leading to unfair rejections. These rejections can further have implications for future work opportunities and can lead to the deactivation of the worker’s account. Lastly, the study observed that even redressal mechanisms to change their ratings or unfair rejections did not work smoothly. Less than half of people knew of such redressal mechanisms, and among them, less than a third had used the platform to contest or appeal a rating. Almost one in five of these said that their rating was not reversed and this happened even if the worker’s performance was affected by factors beyond their control. The ratings, rejections, and inability to resolve conflicts have led to reduced work, lost bonuses, penalties and even deactivation of the platform worker’s account. Going Forward Transparency in data and algorithms on digital labour platforms is crucial. The gig workers generating the data should get the benefits and access to the personal data collected. Moreover, a well-thought-out data law for the gig economy workers is the need of the hour to hold digital platforms accountable.","excerpt":"According to a recent report published by the International Labour Organisation (ILO), digital labour platforms erode the worker’s rights and quality of life. The ILO report analysed the role of digital labour platforms in transforming the nature of work – highlighting the challenges gig economy workers face.  The number of digital labour platforms has increased […]","categories":["IT Services"],"tags":[],"author_name":"Kashyap Raibagi","publish_date":"2021-03-12T15:00:00","publication_year":"2021","word_count":917,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","Git","RAG","GAN","R"],"extracted_tech_keywords":["AI","RAG","AWS","R","Go","Git","GAN","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-plight-of-gig-workers-in-an-algorithm-driven-world\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10165862,"title":"Microsoft Ports TypeScript to Go with 10x Speed Boost","content":"Microsoft on Tuesday announced a project to enhance TypeScript performance by porting its compiler and language tools to Go. The project, titled ‘Corsa’ promises a 10x speed boost for developers, along with a ‘substantial’ reduction in memory usage. Anders Hejlsberg, the lead architect of TypeScript, and a technical fellow at Microsoft, took to a blog post, and a YouTube video to announce the same. The TypeScript compiler and toolset have been ported to Go, through a direct, file-by-file and function-by-function translation from the original codebase. The decision to port the TypeScript compiler and toolset to Go was to overcome JavaScript’s (JS) performance limits. Despite TypeScript’s success over the past decade, its self-hosted JavaScript implementation struggled with issues like slow compile times, and out-of-memory errors, as indicated by Hejlsberg. “We’ve likely reached the limit of what we can squeeze out of JavaScript,” he said. Microsoft expects to be able to preview a native implementation of the TypeScript compiler, capable of command-line typechecking by mid 2025, with a feature-complete solution for project builds and a language service by the end of the year. This project showed over 10x faster compile time across several codebases. For example, compiling Visual Studio Code’s 1.5 million lines of code drops from 77.8 seconds to 7.5 seconds. “This native port will be able to provide instant, comprehensive error listings across an entire project, support more advanced refactorings, and enable deeper insights that were previously too expensive to compute,” read a section of the blog post. Source: Microsoft Microsoft also said that overall memory usage also ‘appeared to be roughly’ half of the current investigation. A few developers also questioned why Microsoft chose Go, over other programming languages like Rust. Matt Pocock, a TypeScript expert, wrote, “Far and away the most important reason was its [Go’s] structural similarity to the current JavaScript implementation. Go’s programming patterns closely resemble TypeScript’s existing code structure.” “This means that contributors familiar with the existing codebase will be able to navigate the codebase easier,” he added. TypeScript 5.8 will be updated to 5.9, and development will continue to the 6.0 series. Project Corsa will be released as TypeScript 7.0. The JS based codebase will be maintained along with the development of the new project – ‘until TypeScript 7+ reaches sufficient maturity and adoption.’ Developers are already in awe of the project. “I cannot think of a bigger impact project in software,” said a user on X. Microsoft is also inviting developers for an ‘Ask me Anything’ (AMA) session in the TypeScript community Discord channel at 10 AM PDT on March 13th.","excerpt":"The native implementation will drastically improve editor startup, reduce most build times by 10x, and substantially reduce memory usage.","categories":["AI News"],"tags":["coding","Microsoft","TypeScript"],"author_name":"Supreeth Koundinya","publish_date":"2025-03-11T23:41:26","publication_year":"2025","word_count":428,"keywords":["Go","Rust","programming_languages:R","AI","coding","TypeScript","ViT","JavaScript","R","Java","Microsoft","programming_languages:JavaScript"],"extracted_tech_keywords":["AI","R","JavaScript","TypeScript","Go","Rust","Java","ViT","programming_languages:R","programming_languages:JavaScript"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-ports-typescript-to-go-with-10x-speed-boost\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":16070,"title":"Softbank CEO Masayoshi Son believes super intelligent robots will outnumber humans in 30 years","content":"SoftBank CEO, Masayoshi Son Doomsday headlines about robots replacing humans is commonplace. The latest to jump on the bandwagon is Softbank CEO Masayoshi Son who believes robots will outnumber humans in 30 years. However, the SoftBank Group’s CEO views also mirror the strategic roadmap of the company.  According to Son, Japan’s internet and telecommunications company is entering into deals concerning artificial intelligence, Internet of Things (IoT) and smart robots. In line with the company’s vision is The SoftBank Vision Fund, established by Son last year and it has raised $93 billion recently. The fund was established to provide startups working in this space with capital impetus.  While the focus area of investment remains IoT, AI and robotics, Son will also plow money in areas such as telecom, computational biology, fintech, mobile applications and computing, and other data-driven business models, such as cloud technologies and consumer internet businesses. Explaining his belief behind setting up the world’s biggest VC fund, Son said, “Artificial intelligence is set to cause a paradigm shift in human history”. Son outlined 30 companies across robotics, health care, agriculture and other areas that are ideal for an investment . Son strongly believes in the concept of Singularity and is totally convinced about the age of super intelligent robots that will arrive 30 years from now. It was this belief that led to the acquisition of UK iPhone chipmaker ARM last year. Like his tech counterparts, Son believes AI can be leveraged for the larger good. “AI is going to be our partner and only if we misuse, it will pose  a threat. AI has the potential to be a partner for a better life. So the future can be better predicted, people will live healthier, and so on,” he added. Other grand plans of the tech billionaire and visionary includes launching 800 satellites in the next three years, in the lower Earth orbit to reduce latency and support faster connectivity. The move is aimed to plug connectivity gaps for connected cars — describing the planned configuration of satellites as “like a cell tower” and like “fiber coming straight to the Earth from space”.","excerpt":"Doomsday headlines about robots replacing humans is commonplace. The latest to jump on the bandwagon is Softbank CEO Masayoshi Son who believes robots will outnumber humans in 30 years. However, the SoftBank Group’s CEO views also mirror the strategic roadmap of the company.  According to Son, Japan’s internet and telecommunications company is entering into deals […]","categories":["AI News"],"tags":[],"author_name":"Amit Paul Chowdhury","publish_date":"2017-07-04T08:35:06","publication_year":"2017","word_count":355,"keywords":["Go","API","artificial intelligence","AI","RPA","data-driven","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","R","Go","API","ViT","RPA","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/softbank-ceo-masayoshi-son-believes-super-intelligent-robots-will-outnumber-humans-30-years\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085871,"title":"InMobi Joins the List of Companies on a Firing Spree","content":"India’s first unicorn InMobi has laid off close to 50 employees, amounting to nearly 3% of its 2600 strong workforce. The reason: performance evaluation. As per sources, the employees impacted by this layoff belong to the firm’s content-providing vertical Glance. With this layoff, Softbank backed InMobi joins a list of 22 Indian tech companies that have laid off over 2000 employees in less than a month. In response to an email asking the firm about the layoffs, the company said “InMobi \/ Glance is in the market actively hiring talent for our ambitious plans. We also evaluate the performance of our existing talent annually and make decisions based on it. This is business as usual for us and part of our annual process. This year is no different.” Based in Bengaluru, InMobi was founded in 2007 by Tewari alongside Abhay Singhal, Amit Gupta, Mohit Saxena, and Piyush Shah. The company’s advertisement serving algorithm helps in optimising the ranking of the advertisements shown on mobile phones. It became India’s first unicorn startup back in 2011. Two years later, Glance—owned by InMobi—an AI-first screen zero content discovery platform, also achieved unicorn status. In 2022, Glance also raised US$200 million from Jio Platforms Limited (Jio) in its Series D round of funding. A Tragic Season of Layoffs Major tech players, including the likes of Amazon, Google, Twitter, Microsoft, and Salesforce, have collectively let go of about 55,000 employees worldwide, that is, close to 3-6 percent of their total headcount. Amid the fears of a global recession in 2023, many tech behemoths, including Microsoft, Spotify, Amazon, Meta, Lyft, HP, Twitter, Salesforce, and Cisco, were compelled to implement mass layoffs. Companies such as Netflix and Adobe too felt the heat of these cutbacks. While Cisco reduced its workforce by 5%, Salesforce announced plans to trim the employee count by 10%. Music streaming platform Spotify has also said it will shed 6 percent of its workforce, that is, 588 people. In a global restructuring effort, Twitter eliminated 7,500 positions, leaving a skeletal 20 staff members in India. Meta, too, let go of 11,000 employees. Interestingly, Apple, on a hiring freeze, has not laid off any employees.","excerpt":"India’s first unicorn InMobi has laid off close to 50 employees, amounting to nearly 3% of its 2600 strong workforce. The reason: performance evaluation.  As per sources, the employees impacted by this layoff belong to the firm’s content-providing vertical Glance. With this layoff, Softbank backed InMobi joins a list of 22 Indian tech companies that […]","categories":["AI News"],"tags":[],"author_name":"Aparna Iyer","publish_date":"2023-01-25T17:12:04","publication_year":"2023","word_count":360,"keywords":["Go","AI-first","ELT","unicorn","funding","AI","programming_languages:R","RAG","R","startup"],"extracted_tech_keywords":["AI","RAG","R","Go","ELT","AI-first","startup","unicorn","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/inmobi-joins-the-list-of-companies-on-a-firing-spree\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166669,"title":"Fluence is Relying on Bengaluru to Tackle Energy Price Volatility","content":"Energy markets globally experience unprecedented volatility, with prices fluctuating through the day. For instance, in Australia, the cost of electricity can go as high as $17,500 AUD per megawatt-hour during peak demand, to a point where discharging stored battery energy can result in a $1,000 AUD penalty. Clubbed with the challenge of an uneven supply of solar energy, there’s a growing need for a balanced and efficient approach in energy storage. Fluence, a leading energy storage company, addresses this challenge with its AI-driven software. In a recent exclusive interview with AIM, Roman Loosen, SVP and chief business operations and transformation officer, said, “Our AI-drive software predicts these price fluctuations, optimising exactly when to charge and discharge the battery. We are not just managing energy—we are maximising profitability.” Much of the software development powering these innovations is happening in Bengaluru. Bengaluru as the Rising Star for Innovation The decision to establish a global innovation centre (GIC) in Bengaluru was strongly influenced by the availability of talent, Loosen mentioned. “When we talk about technology, almost 50% of our team is deployed here. We have other centres worldwide, but the global team is primarily based here. Specifically, in technology, almost 50% of our workforce is in India,” he noted. Fluence has launched its new product, Smartstack, which the company considers a game-changer in the energy storage industry. “Over the past months, we have seen a trend where competitors offer a standardised battery energy storage solution, typically in a 20-foot container,” Loosen explained. However, he mentioned that this setup has several challenges: it weighs around 50 tonnes, requires a heavy-duty crane for lifting, special permits for transport, and complex servicing requirements. Unlike traditional containerised solutions, Smartstack is designed to be modular. It consists of separate components stacked together, including the Smart Skid and battery pods. This modular design is a result of Fluence’s horizontal integration strategy. The company emphasises that Smartstack lives up to its name because it incorporates advanced intelligence within the system. “This begins with sensors embedded in the system, but it extends further into software and controls.” “Many of our competitors are Chinese companies. Given global concerns around energy security, Fluence ensures that our controls and software are developed outside of China, maintaining transparency and compliance with international regulatory standards,” he added. Fluence’s India operations play a key role in this innovation. This centre is remotely connected to Fluence’s entire installed base, allowing for real-time monitoring and remote interventions if any issues arise. In addition to hardware innovations, Fluence uses two proprietary software solutions. One is an asset management system called Nispera, which provides insights into a customer’s energy infrastructure. The other is a bidding software called Mosaic, which further enhances Smartstack’s capabilities. Reliance on Big Techs Fluence emphasises its proprietary approach to software development for critical applications, while also leveraging major technology providers for broader support. “We are also relying on big tech, basically in the background,” Loosen explained, adding that the company is relying on Microsoft Copilot and AWS, among other providers. However, when it comes to Fluence’s bidding software, ownership is a key priority due to the sensitivity of the data involved. Mosaic runs on Smartstack, Fluence’s battery energy storage solution. Empowering Bengaluru Team and the Ecosystem Loosen highlighted the growing importance of its leadership and decision-making presence in India, particularly for its Smartstack product. “For Smartstack, our product manager is based here, so decision-making happens in India,” he said. “Of course…we still have our headquarters in Arlington, but having decision-makers here is crucial.” Currently, the company is focused on its operations in Bengaluru, with future plans potentially including expansion to other locations in India. Loosen further mentioned that it is essential to define the GIC’s place within its global value chain, which means empowering decision-makers within the GIC. Moreover, Fluence is also focused on building a broader ecosystem around its GIC. “We also want to build an ecosystem around the GIC. By ecosystem, I mean the horizontal integration we discussed earlier, as well as our supply chain,” he said. As part of this strategy, the company is localising its supply chain in India, particularly for its new product, Smartstack. As part of its ecosystem strategy, Fluence is collaborating with local suppliers in India to manufacture its first Smartstack domestically. Loosen compared Fluence’s operational model to that of a leading global tech company. “We follow a model similar to Apple: we own the hardware and product design but work with contract manufacturers for production. This approach allows us to scale rapidly,” he explained. Highlighting the company’s rapid growth, he said, “Fluence has grown significantly in recent years, nearly doubling its volume annually.” Loosen also addressed India’s competitive advantage. “When you talk about competitiveness in India, I think it’s also a very strong part if you bring people from different geographies together. India also has a strong local ecosystem with strong local and global players. Now I’m talking about other companies based here in India.” Highlighting the unique work culture in India, Loosen explained that many people in the country have a strong competitive mindset, always aiming to deliver high-quality and efficient projects. In comparison, he noted that in some parts of the Western world, there can be a greater focus on luxury or added features, while in India, there is often a sharper focus on being lean, practical, and competitive. This approach, he believes, is often stronger in India than in other regions. The Way Ahead Fluence holds the largest installed base of battery energy storage systems in India and has established a joint venture to enhance its market reach. “We have the largest installed base of battery energy storage systems in India. We also established a joint venture to enhance our market reach, which is quite unique,” Loosen said. While he emphasised that the company continuously seeks new opportunities to better serve the Indian market, he also noted, “While India is important, it is not currently among our top markets. Our primary markets include the US, Australia, Germany, and a few other global regions where battery storage plays a critical role.” [Correction: This article and the headline have been revised for clarity.]","excerpt":"“Bengaluru creates a unique talent pool in India that can compete with any other place in the world.”","categories":["GCC"],"tags":["GCC india"],"author_name":"Shalini Mondal","publish_date":"2025-03-26T10:54:12","publication_year":"2025","word_count":1025,"keywords":["Go","API","AWS","AI","cloud_platforms:AWS","innovation","RAG","Aim","ViT","GCC india","R"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","R","Go","API","ViT","innovation","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/gcc\/why-fluence-is-relying-on-bengaluru-to-tackle-energy-price-chaos\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101792,"title":"Samsung and IISc Partner to Set Up a Quantum Technology Lab","content":"Samsung Semiconductor India Research (SSIR) has partnered with the Indian Institute of Science (IISc) to set up a Quantum Technology Lab. The lab is expected to integrate cryogenic control chips with qubits, single photon sources, and detectors, while also addressing reliability challenges in quantum technologies. This effort aligns with SSIR’s Corporate Social Responsibility (CSR) commitment and will provide research and training to faculty members and students pursuing higher education, not only at IISc but also at other educational institutions. Guided by Professor Mayank Shrivastava, an Associate Professor in the Department of Electronic Systems Engineering (DESE), this laboratory will become a dedicated facility for advanced quantum technologies. It underlines India’s growing focus on quantum research, innovation and global recognition. The Memorandum of Understanding (MoU) inked by Balajee Sowrirajan, CVP & MD at SSIR, and Professor Govindan Rangarajan, the Director of IISc accentuates the development of indigenous quantum technology. Sowrirajan outlined the goals of this partnership, stating, “The technology scale-up will propel India’s focus on quantum innovation and excellence in the global technology landscape.” The lab will offer opportunities for students practical training, research experiences, and skill development in quantum technologies. Furthermore, the lab will benefit researchers and scientists involved in quantum research. It will also extend support and resources to faculty members from various institutions who may not have the means to engage in capital-intensive research. Professor Govindan Rangarajan from IISc expressed his thoughts on this partnership, emphasizing its significance. He said, “This new Quantum Technology Lab at IISc highlights our commitment to emerging and futuristic research threads. This collaboration with SSIR will strengthen the Institute’s cutting-edge infrastructure and expertise, providing our students and researchers with a unique opportunity to explore the limitless possibilities of quantum technologies.”","excerpt":"The lab is expected to integrate cryogenic control chips with qubits, single photon sources, and detectors","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-10-20T12:55:46","publication_year":"2023","word_count":286,"keywords":["Go","API","programming_languages:R","AI","innovation","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/samsung-and-iisc-partner-to-set-up-a-quantum-technology-lab\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":69766,"title":"GOQii in association with Max Hospital forays into cardiovascular healthcare with “GOQii Heart Care Tracker”","content":"GOQii, the India’s leading health ecosystem and preventive health platform ventures into cardiovascular health with GOQii Heart Care. The key feature of this service is its ability to connect consumers to cardiovascular specialists through the GOQii application and a new GOQii tracker which has capabilities to measure heart rate. This new service comes along with a partnership with Max Hospital which is known for its specialized heart care. How can users avail this feature? To use this new feature, all that the GOQii users have to do is share their health and heart rate data via the new GOQii tracker with the GOQii Doctor for his analysis. The doctor will go through this data and provide specific inputs to the users on how to maintain good heart health. That’s not all, in specific cases, if required the GOQii doctor will refer the user to a cardiac specialists at Max Hospital and other partner hospitals. The idea behind developing this specialized offering was the constant rise of cardiovascular cases across the globe. With this new service, GOQii aims to create the world’s leading health eco-system which connects the GOQii users to leading professionals in the fitness and healthcare space. Commenting on the launch, Rohit Kapoor, Senior Director & Chief Growth Officer at Max Healthcare said, “Heart related diseases are on the rise due to lifestyle choices people make and they don’t generally pay attention to their health. We have observed that patients usually take action when their disease has advanced significantly.” He further added, “The GOQii Heart Care is an exclusive platform that connects the users and the specialists and enables them to work together towards preventative healthcare. We chose to partner with GOQii as they have been a frontrunner in the healthcare segment”. Vishal Gondal, CEO and Founder, GOQii said, “Today, too many devices are generating lots of data which is essentially junk if not analysed and acted upon. The GOQii Heart Care is an one stop solution that tracks and analyses data generated by the Heart Rate trackers and can connect the GOQii user to doctor who can help make sense of the data. It is a unique offering as compared to any other wearable device or service available in the market”. One of the world’s leading Health & Lifestyle Coaching platforms, GOGii created a disruption in the health and fitness segment in 2016 with the launch of its second generation smart band – GOQii 2.0 which includes features such as doctor consultation, diagnostics, auto sleep detection etc. With the launch of the new band, GOQii gets the tops spot in the wearable device segment in Q2 of 2016 as per the International Data Corporation (IDC) report. The GOQii 2.0 band which will incorporate the Heart Care Service will be available on Amazon from February 14th 2017. Consumers can upgrade their existing wearable devices with the new service as well.  With regards to third party hardware which includes the Heart Rate Monitor, consumers can avail the service by paying a nominal consultation fee. The services along with the new GOQii Heart Rate tracker will be available at a subscription Rs 2999 for 6 months and 3999 for 12 months respectively.","excerpt":"GOQii, the India’s leading health ecosystem and preventive health platform ventures into cardiovascular health with GOQii Heart Care. The key feature of this service is its ability to connect consumers to cardiovascular specialists through the GOQii application and a new GOQii tracker which has capabilities to measure heart rate.   This new service comes along […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-01-31T08:15:59","publication_year":"2017","word_count":531,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","disruption","R"],"extracted_tech_keywords":["AI","Aim","R","Go","disruption","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/goqii-association-max-hospital-forays-cardiovascular-healthcare-goqii-heart-care-tracker\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10143851,"title":"Verizon, NVIDIA Partner to Drive AI Workloads on 5G Networks","content":"American telecommunications company Verizon has announced a collaboration with NVIDIA to deliver a solution enabling AI applications to operate on Verizon’s 5G private network with mobile edge computing (MEC). This development, announced on Tuesday, is expected to revolutionise enterprise operations by providing real-time AI services at the network edge. Verizon’s secure 5G infrastructure and private MEC will be integrated with NVIDIA AI Enterprise software and NVIDIA NIM microservices. According to the company, demonstrations of this solution are scheduled to begin early next year. “Generative AI is becoming a cornerstone of digital transformation and business growth. We are enabling real-time AI applications that meet the needs of modern businesses,” said Srini Kalapala, senior VP of technology and product development at Verizon. Scalability and Flexibility The collaboration’s AI-powered private 5G platform supports compute-intensive applications, including generative AI models, video streaming, augmented reality, robotics, and IoT. The platform is also claimed to be designed for scalability and flexibility, enabling developers to innovate rapidly while ensuring high security and reliability. “Verizon’s integration of NVIDIA’s full AI platform into its private 5G networks helps enterprises achieve business objectives more efficiently and effectively,” Ronnie Vasishta, senior VP of telecom at NVIDIA, said. Key Features The solution’s key features are designed to address the demands of modern AI-driven enterprises. It offers ultra-low latency, enabling real-time AI processing, which is critical for applications like robotics, augmented reality, and video analytics, where instantaneous decisions are essential. The solution also provides high bandwidth and ensures the seamless handling of data-intensive workloads. This is particularly important for industries like healthcare, logistics, and manufacturing that rely on large-scale data processing and real-time monitoring. The solution incorporates enhanced security measures to address security concerns. These measures give enterprises greater control over sensitive data and reduce exposure to cyber threats, an essential feature for sectors such as finance and healthcare. Additionally, the solution uses edge AI processing by keeping computation local with MEC technology. This minimises the need to transmit data over the internet, improves response times and maintains data security while enabling quick decision-making for critical operations. AI Will Change 5G According to a PwC study, 75% of executives consider AI a business advantage. Meanwhile, McKinsey reports that 71% of companies plan to expand their use of AI. By 2035, 5G is projected to generate $12 trillion in global economic output, with AI playing a central role in sectors such as healthcare and manufacturing. Verizon and NVIDIA’s partnership expands on the transformative potential of 5G and AI. This initiative aims to empower enterprises to deploy advanced AI applications while driving innovation and efficiency across industries.In November, NVIDIA, along with SoftBank Group’s telecoms division, SoftBank Corp., launched the world’s first AI and 5G telecommunications network, AITRAS.","excerpt":"5G is projected to generate $12 trillion in global economic output by 2035, with AI playing a central role.","categories":["AI News"],"tags":["5G","telecom","verizon"],"author_name":"Sanjana Gupta","publish_date":"2024-12-18T15:02:04","publication_year":"2024","word_count":452,"keywords":["5G","TPU","AI","ML","microservices","telecom","Aim","analytics","generative AI","edge AI","edge computing","verizon","R"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Aim","edge AI","microservices","edge computing","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/verizon-nvidia-partner-to-drive-ai-workloads-on-5g-networks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10081379,"title":"Will New-Age Software Development Kill or Assist Developers?","content":"In the past 50 years, software development has progressed at a swift pace, leaping from machine code to assembly to higher-level (end-user application software like word processors, databases, video games) and finally creating programming languages. Until it all eventually hit a wall. Software engineering was on pause to catch a breather as technology was evolving faster. As dominant as software is, it simply needed a facelift in the current day. Programming languages did not reflect the demands of today’s cyber world. Coding was prone to errors and exacting, and developers found it difficult to handle real-time abstraction and distribution in programming languages. This was when AI made an entry. AI in software development AI was capable of revolutionising the way developers worked, boosting productivity significantly – everything from project planning, quality testing, estimation to user experience could all leverage AI algorithms. Software developers were able to use AI as a tool to extract new knowledge, optimise procedures and then produce higher-quality code instead of replacing it. GitHub Copilot interface Coding faster and easier The first major change arrived with AI pair programming tools built to quicken development. The blueprint for this was the Microsoft-owned GitHub Copilot, which was launched this year in June and has become exponentially popular among the community. GitHub Copilot suggests lines of code to developers in an integrated development environment or IDE like Visual Studio Code, Neovim and JetBrains. Copilot also suggests complete methods and complicated algorithms apart from boilerplate code and help with unit testing. More than 1.2 million developers signed up to preview GitHub Copilot over the past 12 months and it has added another 400,000 subscribers since CEO Satya Nadella stated during an earnings call in the fourth quarter this year. In files where GitHub Copilot has been enabled, GitHub reports that 40% of the code is now written by Copilot. Since then, AI pair programming tools have flooded the market and have added bonus features. Replit, another popular online-based IDE, announced that it would support an AI Mode that includes an ML-enabled pair programmer to help complete code. Amjad Masad, CEO of Replit, Source: Community Round Amjad Masad, the CEO of Replit tweeted about what is to be expected stating, “Crucially, it won’t be ‘prompting’ — we believe that’s more a bug than a feature — it will be a combination of the AI predicting what task you want done next and doing it for you, plus a dialog-based agent that follows your commands.” Ghostwriter also has components that can transform code to modernise it and make it fit standards and explain code by analysing the existing code first and then describe it in natural language. An array of these tools exist, including this year’s releases GitHub Copilot, Amazon CodeWhisperer and Tabnine. They joined a long list of existing AI-powered bots such as Kite Team Server, DeepMind’s AlphaCode and IBM’s Project CodeNet. Advent of Low-code No-code The low-code no-code market is seen as another force multiplier in software development. A lot of recent research in the area including a report by ISG has pegged the LCNC market at about USD 25 billion currently and the sector is expected to grow further at a compounded rate of 28% every year to USD 45.5 billion by 2027. If AI pair programming tools help coders write code faster, low-code and no-code platforms are built for people who don’t know how to code. Built on languages like Python, Java and PHP, low-code no-code platforms have visual software development environments where users can just drag and drop the program components. Consequently more apps can be built, tested and even deployed and more so these apps are focused on simple usage. Low-code No code platform ecosystem, Source: Medium A myriad of factors have popularised low-code no-code across both IT and business job roles. Aside from the obvious benefits like reducing costs and increasing productivity, LCNC is best suited for digital enterprises that have hybrid and remote workplaces and blurs the lines between professional developers and citizen developers while also making companies more agile. Brand new payment structure Another profound shift is expected within the payment structure for developers. Bitcoin’s Lightning Network which is a Layer 2 payment protocol layered on top of Bitcoin permits off-chain transactions, which are transactions between parties who are not using blockchain. Since blockchain transactions do not have to be approved by all nodes in the cryptocurrency network, this speeds up transaction times substantially. In the end when the two parties have completed transactions, they can close the channel. All the information on the channel is then collated into a single transaction which is then recorded. Bitcoin Payments structure, Source: Blockonomics Blog Bitcoin Lightning is integrated into the software supply chain and eases transactions between humans and machines. The new payment system can push transaction costs and overhead expenses in software down, making it much easier to hire developers for specific one-off tasks. It will ensure that developers are paid on time and appropriately for the amount of work they have done. One way to visualise this is that software will move from a stack to a network model. In the stack world, we assemble code in a repo and ship it somewhere to run and then monetization is bolted on. In a network model, code is fully monetized and running all the time.","excerpt":"The Bitcoin Lightning network can push transaction costs and overhead expenses in software down, making it much easier to hire developers for specific one-off tasks","categories":["IT Services"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-12-05T10:00:00","publication_year":"2022","word_count":890,"keywords":["Go","AI","ML","Git","RAG","Python","Ray","GitHub","R","Java"],"extracted_tech_keywords":["AI","ML","Ray","RAG","Python","R","Go","Java","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/it-services\/will-the-new-era-of-software-development-kill-or-assist-developers\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10069673,"title":"How tech is driving financial awareness among Indian youth","content":"Remember the first time you had to file your taxes? For most of us, it was a nerve-wracking experience though it’s quite a simple procedure to follow. In the traditional schooling curriculum, we are hardly taught (or least introduced to) one of the most important skills that we would need later in life—how to manage and grow our money. In adulthood, we face it directly, and many find it extremely confusing. Gone are the days of just FDs and PPFs Probably, when you got your first salary, you may have thought about investing in a PPF and FD, just like our fathers did. But times are surely changing. Millennials know better and are curious about other ways to build their portfolio and not follow old-school ways. The pandemic has also played a crucial role here. People realised how important it is to have a second stream of income (or passive income) and not just live paycheck to paycheck. In fact, a media report says that Central Depository Services Limited (CDSL) reported that clients opened 5 million new Demat accounts in the first nine months of 2020 (when the pandemic was in full swing). This value is half the number of accounts opened over the last five years. The rise of “finfluencers” Beauty, fashion, and food influencers are some of the most-watched influencers in India on popular social media platforms like YouTube and Instagram. Now, finance influencers (or “finfluencers” as they are called) are the new rage in the Indian content space, raking in millions of views. Such has been the rise that we are seeing YouTube making dedicated videos depicting their journey in the space of finance content creation. Recently, YouTube Creating for India showcased the inspiring journey of CA Rachana Ranade (one of the biggest finance content creators in the country). She has a whopping 3.74 million subscribers on YouTube and around 690,000 followers on Instagram. Rachana has a ‘Basics of Stock Market’ lecture series, which has gained massive success. Even music videos don’t garner such views often. The first lecture of her stock market lecture series has over 21 million views. Her videos do consistently well on YouTube. Demystifying the jargon in finance Ranade says, “In India, there is so much jargon about financial planning and stock markets that common people feel they will never be able to understand it. I am using YouTube as a platform to spread financial education among the masses. I started teaching MBA and CA students in 2009. Some students suggested that I make a full series on stock markets. I even recorded the course. Then I was suggested to upload the recorded course on YouTube. Within three months, that lecture got more than 25,000 views. And just like that, I took a leap of faith. I became a YouTube creator in Jan 2020, and in March 2020, a lockdown happened. That’s when people realised that the situation is quite serious.” Akshat Shrivastava is another popular name in the finance content creator niche. He is an ex-BCG and Dalberg consultant and has around 145.4K followers on Twitter and over 1.07 million subscribers on YouTube. He makes mainly finance content (along with career and some lifestyle content) on stock markets, how to pick up the right stocks, mistakes to avoid, how to make money moves to retire early, etc. Pranjal Kamra is one of the most followed finance influencers with over 4.15 million YouTube subscribers and over 388K Instagram followers. His stock market videos have generated millions of views. Reels to engage with the younger generation We all know the attention span of consumers is decreasing, and short-form content is becoming more and more popular with the younger generation. Not just YouTube (which is still traditionally for long-format content), finance content is garnering crazy views on Instagram reels too. FinancewithSharan (run by Sharan Hedge) has over 1.6 million Instagram followers and over half a million subscribers on YouTube. He is known for his informative yet entertaining reels. Sharan, in a media interaction, adds, “Not many people want to watch a 10-minute video on finance. So, I had to figure out a way to condense so much information about finance in a 30-second format so that it reaches as many people as possible.” Trading and investment apps Even a decade back, so many apps and online platforms (read: Zerodha, Upstox, Scripbox, Groww) did not exist that could make investing so easy. Anushka Rathod (another popular business and finance content creator with around half a million Instagram followers) says, “5 to 10 years ago, it was so difficult to invest in a stock. Now you have a phone. The onboarding time is hardly 10 minutes, and then within a few more minutes, you can buy a stock. It is literally that simple.” Nithin Kamath, Founder & CEO of Zerodha (one of the biggies in the online stock broker space), in a 2020 Economic Times interview, said that the company has been adding more than 200,000 new customers every month. April was the largest, but in August again, the company did more than 200,000, he informed. Last year, Ravi Kumar, co-founder and chief executive of online discount broker Upstox, told media that the company’s customer on-boarding accelerated over three times in FY21, and it added over 2 million customers. This took the customer base to over 4 million. He further said that over 80 per cent are in the 18-36 age bracket, while over 70 per cent of the new customers are first-time investors. Digital wealth management service provider Scripbox recently raised $21 million in funding and informed the media that the average age of Scripbox users ranges between 40 and 45, and 25-30% of its customer base is new to capital markets and mutual fund investments.","excerpt":"The pandemic has also played a crucial role, and people realised how important it is to have a second stream of income and not just live paycheck to paycheck.","categories":["IT Services"],"tags":[],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-24T14:00:00","publication_year":"2022","word_count":960,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","Git","RAG","R"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","API","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-tech-is-driving-financial-awareness-among-indian-youth\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":17282,"title":"India’s Industrial Revolution 4.0: Commerce Ministry Sets Up Task Force Dedicated To AI","content":"Commerce and Industry Ministry has constituted a task force to prepare for the industrial revolution 4.0 and the resulting economic transformation. Nirmala Sitharaman, minister for Commerce and Industry, said in a press statement: “With rapid development in the fields of information technology and hardware, the world is about to witness a fourth industrial revolution… Driven by the power of big data, high computing capacity, artificial intelligence and analytics, Industry 4.0 aims to digitise the manufacturing sector.” Reportedly, the task force will include experts, academics, researchers and industry leaders. V Kamakoti, a professor from the department of computer science and engineering at Indian Institute of Technology, Madras, is to head the panel, and Ravinder, Joint Secretary, DIPP, Ministry of Commerce & Industry is to be the convenor. Kamakoti added that that task force was to submit its first report within four months. “The committee will focus on a wide spectrum of areas including agriculture and transportation… We will concentrate on the types of tools, inter-operability, standardisation and skill sets needed for artificial intelligence and machine learning,” he said. The task force is also to explore possibilities to leverage AI for development across various fields, and then submit recommendations to the government, industry and research institutions. The rest of the 18-member panel includes: Anuj Kapuria, High Tech RoboticSystemz Ltd Anurag Agarwal, Institute of Genomics & Integrative Biology, CSIR Ashish Dutta, IIT Kanpur Ashwini Asokan, Mad Street Den, Chennai Gautam Shroff, vice-president & chief scientist, TCS Innovation Labs, Gurgaon GH Rao, HCL Technology G Madhusudan, IIT Madras GVN Apparao, ex-chief technology officer, Cognizant Komal Sharma Talwar, founder, XLPAT Kunal Nandwani, founder & CEO, uTrade Solutions Shantanu Chaudhary, IIT Delhi, Department of Electrical Engineering Vijay Kumar Sankarapu, founder & CEO, Arya.in Ajay Kumar, Additional Secretary, Ministry of Electronics & IT Amandeep Gill, Ambassador\/PR to CD, Geneva K Nagaraj Naidu, Joint Secretary (ITPO), Department of Economic Affairs Aloke Mukherjee, DRDO","excerpt":"Commerce and Industry Ministry has constituted a task force to prepare for the industrial revolution 4.0 and the resulting economic transformation. Nirmala Sitharaman, minister for Commerce and Industry, said in a press statement: “With rapid development in the fields of information technology and hardware, the world is about to witness a fourth industrial revolution… Driven […]","categories":["AI News"],"tags":["drdo","niti aayog ai"],"author_name":"Prajakta Hebbar","publish_date":"2017-08-28T12:59:57","publication_year":"2017","word_count":315,"keywords":["Go","API","artificial intelligence","machine learning","AI","Git","RAG","drdo","Aim","niti aayog ai","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indias-industrial-revolution-4-0-commerce-ministry-sets-task-force-dedicated-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10103931,"title":"Amazon Titan Crushes Google’s Imagen, Meta’s CM3leon","content":"Titan, Amazon’s suite of foundation models has got another addition. The Titan image generator. This week, at the re:Invent conference in Las Vegas, Swami Sivasubramaniam, VP of Analytics and Machine learning announced Titan Image Generator. He said that the tool is now available in preview. He went on to say, “You can customize these images using your own data to create content that better reflects your industry or your brand.” Amazon has entered the market for text-to-image models along with Adobe Firefly. While it’s early to judge, Titan’s competitors have faced challenges. For instance, Google’s Imagen admitted to encoding biases, sometimes resulting in racist or toxic outputs. Similar issues have been observed with DALL.E, Stable Diffusion, and others. The model is said to have been trained on ‘diverse datasets’ though Subramaniam declined to elaborate on the specifics of the data source. “We’re carefully choosing how we train our models and the data we use to do so,” Sivasubramanian said during the announcement. Better late than never AWS is the largest provider of rented computing power and data storage. But it has trailed OpenAI and Microsoft Bing Image Creator (which incidentally uses Dall-E) in bringing to market products based on its own text to image models. Amazon’s Titan line has added new models since a bare-bones launch in April, including some designed to generate text more cheaply than OpenAI’s latest versions. While big tech companies are cautious to release their image models like Imagen and CM3leon to the public, Midjourney, RunwayML, Stable Diffusion have a thriving user base for the same. The main reason most of them are reluctant to release them is because of safety and the risk of spitting out harmful, biased and stereotypical images.  Josh Woodward, VP of Google Labs, explained, “The whole point of AI Test Kitchen is to a) get feedback from the public on these AI systems and b) find out more about how people will break them.” Runway ML, Midjourney, Stability AI and Stable Diffusion have retained the right to ban users creating harmful images and the platform does not process obscene prompts. Meanwhile Amazon has done the same and has built-in guardrails against bias. The feature is said to deny topics that are unsafe and check the user inputs and outputs. In contrast Amazon refuses to release the datasets the model is trained on, and just trust that it has built in mitigations against toxicity. Swami Sivasubramaniam said at the re:Invent that,“ Titan image generator is trained on a diverse set of datasets to enable you to create more accurate outputs.” To combat intellectual property theft and to distinguish between AI generated images to an authentic one, Amazon has added invisible watermarks to their output. Besides just creating a new image, Titan Image Generator allows users to isolate, extract, or integrate new components and edit images. Most useful applications are swapping backdrop settings or incorporating items into lifestyle photographs. Amazon is also hedging its bets, trying to entice other large model makers to offer their software to AWS customers. Their primary model taps the B2B market unlike the already existing platforms. “Generative AI is poised to be the most transformational technology of our time, and we are inspired by how customers are applying it to new opportunities and tackling business challenges,” Sivasubramanian said in the company’s release. Amazon has also highlighted the model’s adaptability for various sectors, such as e-commerce, advertising, and entertainment. For instance, companies can tailor the model with their proprietary imagery to maintain a uniform visual style. Earlier this year, Amazon agreed to invest as much as $4 billion in AI startup Anthropic. As part of the deal, AWS clients have access to Anthropic’s Claude models, including one released last week, Sivasubramanian said. He also said Amazon offers an updated version of Meta Platforms Inc.’s Llama model. “As customers incorporate generative AI into their businesses, they turn to Amazon Bedrock for its choice of leading models, customisation features, agent capabilities, and enterprise-grade security and privacy in a fully managed experience.”","excerpt":"Amazon launches Titan Image Generator, joining the text-to-image AI race with a focus on ethical data use and invisible watermarks.","categories":["Global Tech"],"tags":["AI Tool"],"author_name":"K L Krithika","publish_date":"2023-11-30T16:55:34","publication_year":"2023","word_count":670,"keywords":["Anthropic","machine learning","OpenAI","AI","AWS","ML","RAG","analytics","generative AI","foundation models","AI Tool"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","generative AI","foundation models","OpenAI","Anthropic","RAG","AWS"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/amazon-titan-crushes-googles-imagen-metas-cm3leon\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":42333,"title":"The Thin Ideological Line That Separates White Hat Hackers &#038; Black Hat Hackers","content":"Image courtesy: Nishanth Sanjay Most of the times, the word ‘hacker’ is labelled with a negative connotation. Movies and TV shows have cemented the image of a hacker as a person with glasses typing away at a screen with green text. However, the world of hacking has much more to it than what we see in mainstream pop culture. Hacking is characterised as being a cat-and-mouse game, where the big companies with lots of financial backing go up against an individual or group of hackers. Companies struggle to fix the security holes that are pointed out by hackers, while the latter try to find the next big exploit. However, there is a group of people standing in the middle of both these groups; white hat hackers. The State Of Hacking Today Contrary to popular belief, hackers are not all evil cybercriminals out to get your data. The term ‘hacker’ is a general umbrella for individuals who gain access to a computer system through exploits. Owing to the open-ended nature of code and many programming languages, it is possible to find a way into a complex system through one way or another. Finding exploits and using them to gain access to an otherwise inaccessible system is known as hacking. Hacking can be used for either good or malicious purposes, leading to the bifurcation of the field into white hat hackers and black hat hackers. In theory, the only difference between these two parties is that white hat hackers are given permission to access secured systems. Black hat hackers are not allowed access into these systems, but still get in anyway. In reality, the difference between them is separated by a thin ideological line. Crossing The Line White hat hackers usually start off as bug bounty hunters, who find exploits in prominent systems and websites for a bounty from the company operating them. While many bounty hunters continue with solving these solutions for bugs, most of them get picked up by an Internet company to work as a ‘security consultants’. There exist individuals who have ‘crossed the line’ and have become black hat hackers. These individuals are the main turning force of the reactive method used to patch systems after attacks. Usually, a security cycle proceeds in this fashion. The company releases a new product or update after extensive penetration testing by in-house hackers. Then, dedicated black hats find exploits in the software. Once notified of this, either by an attack, an attempt to attack or public disclosure, the exploit is patched by the company. Usually, black hat hackers go after the bugs that have the largest amount of potential bounty, leaving the company to take reactive measures after attacks. White hat hackers are tasked with a practice known as penetration testing, which is the act of testing all possible attack vectors for a system. This not only gives a deeper look of the system itself, but also possible ways to exploit limitations in its architecture. This is a proactive method of protecting against cyberattacks, with the other being a reactive method. The back-and-forth creates an interesting dynamic between white hats and black hats, one that is more than just action and reaction; an ideological barrier. Black vs. White vs. Black vs… This barrier comes from the attitude of hackers on either side of the spectrum. In a never-ending battle, white hat hackers fight against vulnerabilities exposed by black hats, while black hats fight for malicious purposes and monetary gain. In the case of finding an exploit, black hats will most likely use it as a foothold to launch an even more dangerous attack against the system. White hats will simply report it and fix it, making the system stronger against malicious entry. Data breaches, Distributed Denial of Service attacks and theft of financial information are the mainstays of black hat hacking. This is mainly due to the monetary gain associated with them, as data can be sold on the dark web for money, while DDoS attacks can be performed as a paid attack for competitors of the target. However, one characteristic of black hats make them very different; bragging. Black hat hackers have specialised forums on the dark web such as FreeHacks, which enable a transfer of information and sharing new methods to exploit systems. More than these purposes, such forums are used to announce when a big hack has been conducted, suggesting that many black hats do it for fame. White hats, instead, stand for the protection of personal information and systems that contain them. They are the first line of defence against any malicious attack by black hats, and are against the ideas of malicious entry and data theft. Thus, this David and Goliath game continues, with the underdogs scoring ‘wins’ now and then while the giants continue to defend their fortresses. One cannot exist without the other, and both parties are locked in an ever-growing battle with bigger stakes. Hackers on both sides continue to fight against each other, each side for their own ideas and goals, leaving a battle largely unseen by anyone else.","excerpt":"Most of the times, the word ‘hacker’ is labelled with a negative connotation. Movies and TV shows have cemented the image of a hacker as a person with glasses typing away at a screen with green text. However, the world of hacking has much more to it than what we see in mainstream pop culture. […]","categories":["AI Features"],"tags":["Bug Bounty","Cybersecurity"],"author_name":"Anirudh VK","publish_date":"2019-07-12T13:00:12","publication_year":"2019","word_count":847,"keywords":["Go","programming_languages:R","AI","Bug Bounty","programming_languages:Go","RAG","Cybersecurity","R"],"extracted_tech_keywords":["AI","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-thin-ideological-line-that-separates-white-hat-hackers-black-hat-hackers\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45338,"title":"OpenAI&#8217;s Strategy Questioned As Grad Students Recreate The Infamous GPT-2 At A Fraction Of The Cost","content":"Earlier this year, OpenAI gained a lot of attention for all the wrong reasons when it produced a language model so good at generating fake news that the organisation decided not to release it. In fact, a study conducted by collaborators at Cornell University found that readers on average believed GPT-2’s outputs to be genuine news articles nearly as often as New York Times ones. OpenAI’s strategy of delaying the release of the model relies on these models being difficult to replicate and requiring a high degree of specialised domain knowledge. However, in what may come as surprising news, two grad students used an estimated $50,000 worth of free cloud computing from Google which hands out credits to academic institutions and made a decent attempt at replicating OpenAI’s GPT-2. Why Was Released It In The First Place OpenAI developed GPT-2 to investigate new benchmarks for natural language processing tasks. The authors of the original paper reported that GPT-2 zero-shots to state of the art performance on 7 out of 8 tested language modelling datasets. They claimed that the diversity of tasks the model is able to perform in a zero-shot setting suggests that high-capacity models trained to maximise the likelihood of a sufficiently varied text corpus begin to learn how to perform a surprising amount of tasks without the need for explicit supervision. OpenAI has insisted that it had not yet found any attempts of malicious use but had seen multiple beneficial applications, including in code autocompletion, grammar help, and developing question-answering systems for medical assistance. As a result, the lab felt that releasing the most recent code was ultimately more beneficial. Other researchers argue that several successful efforts to replicate GPT-2 have been made. How Two Grad Students Replicated GPT-2 The authors based their implementation off of the Grover model and modified their codebase to match the language modelling training objective of GPT-2. The authors in their blog gave a detailed view of they got around this model. Briefly, it can be described as follows: Started with the Pushshift Reddit scrape, a dataset containing a continuously updated collection of Reddit posts, comments, and related metadata. These links are then filtered to remove direct links to file-types unlikely to contain usable text or HTML. The text was extracted from HTML pages using the Newspaper Python library and then filtered for only English text using the fastText Python library. As a cleaning heuristic, documents with fewer than 128 tokens were removed from the dataset. For encoding the dataset, the Binary Pattern Encoder was used. Here’s a sample from their experiment: Prompt: “Recycling is good for the world. NO! YOU COULD NOT BE MORE WRONG!!” Output: Recycling is good for the world. NO! YOU COULD NOT BE MORE WRONG!! That’s because recycling is all about selling to customers who cannot afford to recycle.” If you don’t have any interest in recycling, in whatever form, then you could well be doing your own global environmental benefit if you just got rid of it. Let me explain. Recycling takes time, money and energy, and it doesn’t really solve the problem… Aaron Gokaslan, 23, and Vanya Cohen, 24 the duo behind this project insist that their release was intended to show that you don’t have to be an elite lab rich in dollars and PhDs to create this kind of software. And they argue that setting their creation free can help others explore and prepare for future advances — good or bad. Because the replication efforts are not unique, and large language models are the current most effective means of countering generated text, the students believed that releasing their model is a reasonable first step towards countering the potential future abuse of these kinds of models. The cost of training the model from scratch is about $50k. It’s important to note this figure is the estimated value of the cloud compute, and does not reflect the much smaller intrinsic costs involved. “This allows everyone to have an important conversation about security and researchers to help secure against future potential abuses,” says Cohen. Why Should We Care About Another ML Model? Machine learning practitioners have stayed divided for a long time over the reliability of AI. This owes in some part to the black-box modelling. The inner workings of a deep learning model is still unclear. Attempts such as activation atlases have been made to investigate how a model learns. However, the evaluation of a model based on just the end results has made people sceptical of AI. So whenever a new idea like GPT-2 is introduced, its most extreme outcome is often highlighted. In the case of GPT-2, the uncanny way in which a model spun stories out of thin air, made many uncomfortable. People started speculating about dire consequences such as fake news. For example, any malicious entity can sit in a remote place and can script speeches of presidents and can aggravate things within a nation or across the world. The rise of social engineering has been witnessed by the world during the 2016 presidential elections in the US.  So, it is quite understandable why people are paranoid about this new text generation model. The replication of this model by college students with a few thousand dollars cloud credits, sparks debate about the need for regulating AI. Both the models, the larger 1.5 million parameters or the latest cut down OpenAI version still have a long way to go. However, due to the ominous or sometimes exaggerated claims, the argument is being made against it rather than enhancement of NLP models as a whole. We can also not blame OpenAI for trying out new things. It is a new company and so is the field of applied AI. This makes even policy making difficult because there has been no precedent and the experts, think tanks have a lot to do in the coming years to ensure a safe human-AI symbiosis.","excerpt":"Earlier this year, OpenAI gained a lot of attention for all the wrong reasons when it produced a language model so good at generating fake news that the organisation decided not to release it. In fact, a study conducted by collaborators at Cornell University found that readers on average believed GPT-2’s outputs to be genuine […]","categories":["AI Trends"],"tags":["AI Safety","GPT2"],"author_name":"Ram Sagar","publish_date":"2019-09-03T11:05:22","publication_year":"2019","word_count":987,"keywords":["machine learning","TPU","OpenAI","AI","cloud computing","ML","RAG","NLP","Aim","deep learning","GPT2","AI Safety"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","OpenAI","Aim","RAG","cloud computing","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/openais-strategy-questioned-as-grad-students-recreate-the-infamous-gpt-2-at-a-fraction-of-the-cost\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":33693,"title":"GANs May Soon Replace Your Stock Brokers. Here’s How","content":"The latest developments in artificial intelligence, especially in the applications of Generative Adversarial Networks (GANs), are helping researchers invade the final frontier of human intelligence. With new papers being released every week, GANs are proving to be a front-runner for achieving the ultimate — General AI (AGI). Machine learning is proving to be a great addition to the arsenal of financial modelling and, with advancements like GANs, the algorithmic trading sector is poised to reap benefits. GANs have advanced to a point where they can pick up trivial expressions denoting significant human emotions. These dual-dueling networks seem to have saved something for the financial sector as well. Incentives and pitfalls are integral to any trading event. The winner in this scenario is the one who cuts on losses. So, what do GANs have to offer? Generator And Discriminator Before talking about GANs, it is essential to have an idea of their Generator and Discriminator networks and how these two go toe to toe with each other like arch-nemesis; benefitting the overall model eventually. The Generator is responsible to produce a rich, high dimensional vector attempting to replicate a given data generation process; the Discriminator acts to separate the input created by the Generator and of the real\/observed data generation process. They are trained jointly, with G benefiting from D incapability to recognise true from generated data, whilst D loss is minimized when it is able to classify correctly inputs coming from G as fake and the dataset as true. Fine-Tuning Trading Strategies With cGAN In this paper, the researchers show how the training and selection of the generator is made; overall, and, how it is less costly than the backtesting or ensemble modelling process. Firstly a conditional GAN is trained and then a generator G is used to draw samples from the time series. Conditional GANs (cGANs) are an extension of a traditional GAN, when both G and D decision is based not only in noise or generated inputs but include an additional information set. This additional information set contains labels like forecast of market value and other such features. The dataset used for this experiment was obtained from Bloomberg, which contains a list of 579 assets tickers over a time period of 18 years from 03\/01\/2000 to 02\/03\/2018, with an average length of 4,565 data points. Input features and target were scaled using a z-score function to ease the training process. The right cGAN was selected by taking snapshots every 200 iterations (snap= 200), drawing and evaluating 50 samples per generator along 20,000 epochs. Stochastic Gradient Descent with a learning rate of 0.01 and batch size of 252 as the optimization algorithm, was considered after checking different architecture sizes for their performance across the benchmarks. For every sample, one-split was performed to assess a set of hyperparameters. The base learners that are individually ”weak” (e.g. Classification and Regression Tree) but when aggregated can outcompete other .”strong” learners (e.g., Support Vector Machines) such as, Random Forest, Gradient Boosting Trees, etc., techniques that make use of Bagging, Boosting or Stacking. Autocorrelations for SPX Index using cGAN via paper by Adriano Koshiyama, Nick Firoozye, and Philip Treleaven A cGAN sample is drawn repeatedly to train a base learner. These base models are then returned as an ensemble. The ensemble Mean Squared Error tend to be minimized, particularly when low bias and high variance base learners are used, such as deep Decision Trees. (You can check out the full algorithm here.) Key Takeaways Creating proper validation sets and handling of time series through learning an unconditional model, similar to image and text creation. cGANs can be used for model tuning, bearing better results in cases where traditional schemes fail. Modelling of multiple financial time series. cGAN can be used to generate resamples of critical events for stress tests. Combining resamples from cGAN and Stationary Bootstrap can yield better results. cGANs appear to be doing well when pitted against other validation strategies. The large scale deployment of GANs can only be guaranteed depending upon whether these results will be replicated in a real-world scenario or not.","excerpt":"The latest developments in artificial intelligence, especially in the applications of Generative Adversarial Networks (GANs), are helping researchers invade the final frontier of human intelligence. With new papers being released every week, GANs are proving to be a front-runner for achieving the ultimate — General AI (AGI). Machine learning is proving to be a great […]","categories":["Deep Tech"],"tags":["financial analytics","GANs","Machine Learning"],"author_name":"Ram Sagar","publish_date":"2019-01-18T04:54:06","publication_year":"2019","word_count":682,"keywords":["Replicate","Go","artificial intelligence","machine learning","programming_languages:R","AI","RPA","Machine Learning","RAG","financial analytics","GANs","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","Go","GAN","RPA","Replicate","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/gans-may-soon-replace-your-stock-brokers-heres-how\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10086753,"title":"Build Robust Text Analyser and Insight Engine","content":"Introduction As per International Data Corporation (IDC), 80%–90% of data in the world will become unstructured by 2025 and a major part of it would be text data. In effect, this mandates enterprises to generate insights from the text for right decision making, growth, streamlining the business strategies and profitability. In parallel, we could see tremendous improvement in natural language processing and AI\/ML techniques. Though we have effective natural language processing techniques, machine learning and deep learning algorithms, the accuracy\/aspects of insights generation depends on the data scientist who analyses and interprets the raw text. Due to lack of business knowledge, these data scientists majorly rely on subject matter experts to understand the data, training set generation and right KPI’s (Key Performance Indicators) identification. Additionally, model building is time consuming and in contrast, analytics and insight generation need to be expedited at the earliest for right decision making at the right time. To overcome such challenges, we need to build robust text analysers with the capability to understand the data and generate training sets on our own and classify all kinds of human-generated text.  This is possible, if we apply hybrid approach to build our text analyser and insight engine with components such as text cleanser using natural language processing’s linguistic features, handle different way of human writing using sentence embedding transformers, understand the data using unsupervised clustering algorithm, NLP based labelling assistant and business insights generation using vector based classifiers. In addition, this proposed approach is very well-suited for multi-class classification problems and helps data scientists build their own insight engine, perform text analysis within a short span of time and extend to all kinds of text analysis with minimal configurable changes. Robust Text Analyser and Insight Engine Since most analytics projects are short term and the objectives are towards the insight generation for the right decision making, the focus would be on less turnaround time to generate these insights. Considering this, I have developed the Robust Text analyser with four different phases—data cleansing, gaining knowledge on the dataset, corpus generation and insight generation, as shown in Fig. 1. Data Cleansing Every human-generated text dataset contains various data quality issues such as spelling mistakes, junk characters, unwanted special characters, format inconsistencies, partial data filling, non-standardised data, different naming conventions and short descriptions, among several others. These quality issues affect the accuracy of the analysis. To overcome this, the first level of cleansings treatment can be performed using regex or POS tagging as well. Following basic cleaning, choosing the right word for the analysis is very important. This can be achieved using Part of Speech Tags. The part of the speech tag explains how a word is used in a sentence. There are eight main parts of speech tags such as nouns, pronouns, adjectives, verbs, adverbs, prepositions, conjunctions and interjections. Though eight main tags were available, the analysis can be performed using nouns, adjectives, verbs and adverbs tags. These four tags can be considered as analytical-tags. As a second level of cleansing, using analytics-tags, we can choose the important words for the insight generation. This approach helps us reduce the size of the data to be analysed drastically and also improves the accuracy. POS tagging can be performed using Spacy or NLTK  libraries. Fig. 1: Robust text analyser phases Knowledge Gain To identify the context of information to be presented in the report for decision making, complete understanding of the dataset becomes very crucial. Due to lack of subject knowledge, this process can lead to a dependency on subject matter experts and delay the analysis turnaround time. To overcome the same, I have used text clustering which groups the related items together, understand the categories of information belonging to each and every group and use this knowledge to speed up the labelling activity. The above mentioned text clustering becomes challenging due to humans generated with different writing patterns. In order to handle this challenge, preprocessing the text data with sentence embedding techniques is very effective. Sentence embedding is the collective name for a set of techniques in natural language processing where sentences are mapped to vectors of real numbers while a transformer is a deep learning model that adopts the mechanism of self-attention, differentially weighing the significance of each part of the input data. Overall, the Knowledge Gain phase contains data preprocessing supported with sentence embedding transformers—universal sentence encoders are preferable—which converts the text into 512 dimensional vectors and enables the text to support AI\/ML-based techniques for insight generation. Following text transformations, text clustering needs to be applied which will group the related information despite larger paragraphs and will help walk through different categories of information that the dataset possesses, eventually fastening the text data and understanding the phase compared to a sequential order. Corpus ML-based text analysis mandates the training set\/corpus, which is a time-consuming process due to SMEs’ dependency or lack of business knowledge. To overcome the time complexity and to expedite the corpus generation phase and speed up the analytics timeline, I have built a customised auto-label generators by finding the specific POS tags combinations such as Noun + Adjectives, Noun + Adverbs or Noun with analytical-tags existence in the dataset. Such customised pattern validation in the training set generation phases was very effective.  Sometimes, the data we deal with may contain longer paragraphs which can be handled by splitting the paragraphs into multiple sentences followed by filtering logic application. This auto-label generation approach helps to identify the most useful\/right phrases for the training set. In addition, this approach detects the labels automatically, easily and quickly. The accuracy of the auto-label detector approach is limited due to the nature of the dataset and human writing patterns but it would be helpful in reducing the total labelling turnaround time. Insights Generation In this phase, we can opt for any classification machine learning\/deep learning techniques with modelling life cycle such as building model, testing, deployment, retraining and governance. But analytical projects are short in span, non-repetitive and model governance may not be preferable due to dedicated FTE allocation and rework. In addition, there are some situations where analytical classification needs multiple classes to be coined to one paragraph. For instance, in an online purchase, an order delivery failure can occur due to multiple reasons such as system error, product classifications, product stocking area, product unavailability, delivery issues, or more. In this scenario, the standard existing classification ML models predict and result in only one output class which cannot be preferred to understand the holistic picture of the Performance analysis. So, to support these kinds of analytical natured projects, I have designed the robust text analyser’s Insight Generation phase by performing vector similarity based classification with the capability to classify ‘n’ number of output classes. To perform this phase, we need two prerequisite items such as corpus and cleansed base dataset—as mentioned in the Data cleaning phase—to be analysed. These two prerequisites—training set and the base dataset—need to be converted to sentence embedding based vectors. Finally, using Numpy similarity measurement algorithms and the matching threshold, multiple output class classification can be performed faster. Overall, this classification approach featured with a single item can be classified into multiple classes and new dataset\/metrics\/purposes can be uplifted only by revising\/replacing the training set. Applications This approach is well suited for insight generation from unstructured text data across sectors—such as e-commerce, healthcare, media, banking and others—for failures root cause analysis, performance analysis, complaints analysis, software issues cause analysis, chatbot query analysis, description analysis, survey analysis, and email analysis. It can also be a good assistance for topic modelling.","excerpt":"Model building is time consuming and in contrast, analytics and insight generation need to be expedited at the earliest for right decision making at the right time. To overcome such challenges, we need to build robust text analysers with the capability to understand the data and generate training sets on our own and classify all kinds of human-generated text.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Anusuya Kirubakaran","publish_date":"2023-02-08T10:00:00","publication_year":"2023","word_count":1264,"keywords":["NumPy","machine learning","AI","NLTK","ML","Transformers","NLP","deep learning","analytics","spaCy","AI Tool"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","analytics","Transformers","spaCy","NLTK","NumPy"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/build-robust-text-analyser-and-insight-engine\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":25796,"title":"How Will Implementation Of Blockchain Have Any Impact On Ads Industry?","content":"Today, Blockchain has become the new and revolutionary technology that everyone is talking about. It has been adopted in several sectors of many industries, ads industry is no exception. With so many frauds increasing day by day, ads industry is struggling with finding a solution for an effective ad delivery. Blockchain can be that solution as it offers decentralized database, consensus protocol, and data transparency. Blockchain is a decentralized digital ledger that contains all the transaction. It works around the encryption of the data, timestamps it and then stores it in a decentralized database. This ensures security, immutability and data transparency among users. Let’s discuss some critical points on how Blockchain will affect the current ads industry. Reducing Ad Frauds With Blockchain Technology The abundance of fraudulent advertisement is on the rise. According to a study, in 2017 ad frauds has cost many firms a loss of USD 16.4 billion. This is where Blockchain perfectly fits in with its feature to track any transactions to its very end. This will ensure advertisers that their ads are viewed by real humans not by bots and help to reduce any ad frauds. Blockchain can validate ad impressions by checking the authenticity of the domain and identifying non-default browser to check whether the ad delivery was real or not. Not only that, identifying and avoiding ad frauds in real time is now also possible through Blockchain. Reduction in ad frauds will ultimately lead to decrease in ad spends, making it cost and time efficient as well as effective. Real-Time Data Tracking Data is like oxygen for any marketing or advertising company. There is an ongoing need for real-time data tracking and acquiring. Blockchain with its features fills this gap and gives advertisers the relevant data, that too when they need it the most. With this, marketers can effectively check whether ads are reaching their targeted audience or not. This way Blockchain helps marketers to manage campaigns more effectively. Apart from managing campaigns with real-time data insights, Blockchain can provide marketers about the information regarding a client’s preferences. Advertisers can gain the insights about the patterns and inclination the client is utilizing for their targeted audience, channels, and platforms. This can help advertisers to narrow down the targeted audience to a specific group according to the page content and visits. Better Data Security The reason behind Blockchain being the hot topic for everyone is because of its security and immutability. Many successful Blockchains have not been hacked to date. This makes Blockchain a haven for the company’s confidential data. Every transaction in a Blockchain network has to go through a verification process by multiple users within the network. After which, it is encrypted, time-stamped and stored in a decentralized database. Any illegal changes or transaction will easily be caught as the tampered data copy will not match other user’s data copy. This will avoid any fraudulent acts through hacking. Blockchain being immutable is almost impossible to hack. Furthermore, even if someone tries to tamper the data, it won’t take time for other users to seize the fraud. Removing Any Middleman With the surety of authenticity and transparency among users of the Blockchain. There will be no need for any third party like brokers to strike a deal between publisher and advertisers. It will be up to Blockchain to bring authentic clicks and traffic. This way publishers and advertisers will exchange data directly, removing any fees from the middleman. With the third party removed from the picture, smart contracts will come into use. Smart contracts work totally upon Blockchain network. In it, a specific data or asset is released only if a specific condition is fulfilled. Data Privacy Solution Today, authorities like social media giants, banks, governments, etc. have the power to control the personal data of a regular person. There are so many cases where these authorities are found guilty in leakage of private data for their personal gains. Adopting Blockchain will snatch this power from authorities and give it to regular people. Users will be able to decide what data they want to share and with whom. This is a boon for marketers who have to obey GDRPR rules. More Genuine Influencers Many marketing companies look for public figures and influencer to advertise their product. Blockchain technology will help them to get genuine top influencers according to the authenticity of their followers. This will not only help advertisers to get genuine influencer but also rat out any fake ones also. Summary Blockchain Technology can do wonders for ads industry. It can be the tool that will form a better future for digital advertising. With so many features to offer, it will be wise for many industries including advertisement industry to adopt it and reap its benefits. Blockchain technology will give companies a competitive advantage over others with its secure, immutable and transparent network. P.S Enuke Software offer innovative Blockchain Services for our global clientele – at affordable rates. Visit http:\/\/www.enukesoftware.com\/blockchain-application-development.html for more information.","excerpt":"Today, Blockchain has become the new and revolutionary technology that everyone is talking about. It has been adopted in several sectors of many industries, ads industry is no exception. With so many frauds increasing day by day, ads industry is struggling with finding a solution for an effective ad delivery. Blockchain can be that solution […]","categories":["AI Features"],"tags":[],"author_name":"Owais Rasheed","publish_date":"2018-06-27T05:13:13","publication_year":"2018","word_count":830,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","Git","R"],"extracted_tech_keywords":["AI","ML","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-will-implementation-of-blockchain-have-any-impact-on-ads-industry\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172739,"title":"Cloudflare Just Became an Enemy of All AI Companies","content":"Cloudflare might have just killed the web search functionality of AI chatbots. The company announced that it would start blocking AI crawlers by default, drawing a line in the open web where content is no longer a free fuel for AI. If AI companies want in, they will have to pay up. The announcement reframes the foundational deal that powered the web for decades. For years, websites gave Google content, and in return, Google sent them traffic. Now, generative AI is severing that loop with GEO — copying without clicks, quoting without proper credit, and more. Cloudflare, which routes traffic for 20% of the internet (as the company claims), says it is time for publishers and AI companies to work together to reward the content that it deserves, and improve the economy of the web. This move won’t halt AI, but it might slow its free lunch. And that’s precisely the point. The Company Calls it ‘Content Independence Day’ “AI-driven web doesn’t reward content creators the way that the old search-driven web did,” reads the blog post, arguing that the exchange of traffic-for-content no longer holds in a world where tools like ChatGPT and Claude scrape text to generate answers with no attribution or reward. “With OpenAI, it’s 750 times harder to get traffic than it was with the Google of old. With Anthropic, it’s 30,000 times harder.” That isn’t a gentle drop-off, it’s a cliff. And content creators are falling off it. Cloudflare’s new policy flips the default, from passive permission to active protection. Every new domain signing up with the service now gets asked whether they want to allow AI crawlers. The default is “no”. Companies like Gannett Media, Condé Nast, Quora, Ziff Davis, and Reddit are backing the initiative, aiming to restore value that AI has quietly eroded. This could also address the trouble caused by AI crawlers. Bots from OpenAI, Anthropic, and Meta are increasingly burdening independent websites by consuming excessive bandwidth and disregarding protocols like robots.txt, resulting in higher bills and degraded server performance. A Slowdown, Not a Stop Developers like Gergely Orosz on LinkedIn and X also have raised concerns over this aggressive scraping, with some building tools like Anubis to fight back. Cloudflare seems to be adamant on what it wants to do. The company earlier reported that AI bots now account for more than 50 billion daily requests and have responded with deflection tools, such as AI Labyrinth, to waste bot resources. “If the Internet is going to survive the age of AI, we need to give publishers the control they deserve and build a new economic model that works for everyone – creators, consumers, tomorrow’s AI founders, and the future of the web itself,” said Matthew Prince, co-founder and CEO of Cloudflare. He added that the goal of Cloudflare is to put the power back in the hands of creators, while still helping AI companies innovate. “This is about safeguarding the future of a free and vibrant Internet with a new model that works for everyone,” he added. Even Reddit agrees. “AI companies, search engines, researchers, and anyone else crawling sites have to be who they say they are. And any platform on the web should have a say in who is taking their content for what,” said Steve Huffman, co-founder and CEO of Reddit. “The whole ecosystem of creators, platforms, web users and crawlers will be better when crawling is more transparent and controlled, and Cloudflare’s efforts are a step in the right direction for everyone.” While web search features in AI tools offer utility, there is a growing consensus that crawler behaviour must be regulated to protect smaller web operators. Considering this, it looks like Cloudflare’s new measures can be a necessary feature for the web. An Open Web With Closed Gates? The real significance of Cloudflare’s move isn’t just the block, it’s the framework it hopes to build next. The company plans to work on a marketplace where the value of content is judged not by page views, but by how much it adds value in terms of knowledge. It’s a step toward rewarding originality, not clickbait. Cloudflare is also working on protocols to help AI crawlers identify themselves, allowing publishers to make nuanced decisions, which could permit AI for search, but not for training. Until now, content scraping has been largely unregulated, masked behind generic user agents and vague intentions. Still, the policy opens up a paradox. AI companies are invited to work with Cloudflare, provided they compensate. This puts the company in a powerful position, which could be beneficial for publishers using Cloudflare, and in a way, could also be controversial for AI companies. Publishers may celebrate the move, but AI developers may see it as a speed bump to innovation. For an industry built on large-scale web scraping, “permission” could become the new latency.","excerpt":"“Our goal is to put the power back in the hands of creators, while still helping AI companies innovate.”","categories":["AI Features"],"tags":["Cloudflare"],"author_name":"Ankush Das","publish_date":"2025-07-02T18:00:00","publication_year":"2025","word_count":809,"keywords":["Anthropic","ChatGPT","Go","API","OpenAI","AI","chatbots","Cloudflare","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Anthropic","Aim","chatbots","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cloudflare-just-became-an-enemy-of-all-ai-companies\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10049606,"title":"Interpretable Student Performance Predictions","content":"Digital transformation has taken over the world in the last few years. From Artificial Intelligence (AI) to Big Data, digital technologies play an essential role in everyday lives. With all of the technological growth, tons of data are generated each second. This data has beneficial results and actionable insights for economies, governments, and societies. It is currently used extensively in retail, energy, banking, and finance, among other industries. One key industry that can benefit from this data revolution is Education. This sector occupies a place of revered importance even today. Data science, particularly AI, can be used to automate repetitive tasks or create interactive studying aids. There are many more uses, one of which is discussed in this article in detail – interpreting student performance predictions. These performance predictions are specifically helpful for educators as well as policy formulators. Data Analytics and Mining in Education Various data analytics techniques are used to gather and interpret insights from educational information. However, most prediction models face a significant problem – they can’t explain ‘why’ they arrived at a particular prediction and ‘how’ they did it. Figuring out the interpretation capabilities of Artificial Intelligence (AI) models is essential when researchers deal with complex and massive datasets. Analytics and insights need substantial proof or backing, and if the models can’t provide it, the levels of human trust in AI become lower. This research project aims to change this perception of AI in education by proposing a new prediction model using existing technologies. Recent digital progress has encouraged the adoption of AI in data mining. Data mining is useful to extract relevant data and identify patterns to provide actionable insights. In the education domain, the process is called Educational Data Mining (EDM). It uses tools and techniques to gather information from records, logs and examination results. It also follows various dimensions such as student’s performance, dropout rates, individual intelligence and cognizance, teacher’s\/administrator’s performance and more. But like any other technology, EDM is still an emerging discipline. The International Educational Data Mining Society agrees. Considering the definition that society provides for EDM, methods to explore educational data and use them for student betterment are still developing. This project is a step forward in improving EDM, expanding on technological progress. What’s progress without Challenges? Past data and literary works were extensively covered for this research. The information learned through them provided a set of challenges. 1.    Black box approach When dealing with sensitive information, AI adopts a black box approach. It shares results, but the end-user doesn’t know ‘how’ and ‘why’ it arrived at the results. This approach is still prevalent in most industries including Defense, Health, Autonomous Vehicles and the likes. While the black box approach is changing today, several industries are still wary of previous opacity in results. For models and their interpreted results to be acceptable, they need to be more transparent. The lack of confidence in black box AI is still present in the educational field, which poses challenges. 2.    Less trustworthy model prediction Every research model has to pass a certain confidence level. The confidence level is arrived at after extensive testing and validating data. However, not everyone believes that all confidence levels are trustworthy. For a model to succeed, it’s crucial to provide unbiased results. If the confidence levels are not right, the model will be less trustworthy and not have solid reasoning for its results. 3.    Gaps in past research In a field that’s continuously developing, getting accurate data to support research is difficult and time-consuming. The project massively covered past data in the education sector, focusing on predictions for student performance using AI or Machine Learning (ML) models. But none of these works achieved interpretable and trustworthy model predictions. Moreover, the techniques used still had undefined aspects such as their plot delivery, intuitiveness, accuracy and trustworthiness. Further research is necessary to support the model proposed in this project. Construction of the Model The prediction model for this project was proposed by considering students’ behavioral, academic and demographic feature attributes. Data mining extracted these attributes from a learning management system. Common EDM techniques use popular and performant algorithms such as Logistic Regression, Decision Trees, Support Vector Machines, Artificial Neural Networks and Random Forest. However, more often than not, these seem opaque. Therefore, this research focused on building a model that: ●      Provided insights on features and the reasoning behind the ones selected ●      Used interpretable AI to identify the marginal contribution and correlation of the various features ●      Analyzed and understood contributing factors for classification or misclassification of an instance The model was examined through a set of classifiers – XGBoost, Naïve Bayesian and Random Forest – but achieving credible confidence levels was challenging. To better the confidence level and trustworthiness, Explainable Artificial Intelligence (XAI) methods which gave the model interpretability, were used. By sharing results that show both advantages and limitations, XAI improved the trustworthiness of AI in education. SHAP, the XAI method used for this research, addressed the problem of non-interpretable results and black box approaches followed by previous prediction models. To build the model, the project undertook a comprehensive comparative analysis of highly performant models. Out of all, Random Forest, XGBoost and Naïve Bayes classification algorithms were found best suitable. Data visualization and feature selection techniques further helped to eliminate irrelevant features. Feature elimination was done by ranking and using an evidence-based approach. For the model to be adopted on a larger scale, making the process as transparent as possible by sharing interpretable insights was vital. As a result, the model presented key observations of experiments by denoting the importance of features and using multiple approaches based on Random Forest and XGBoost. It also covered feature comparisons to rank them based on importance. All the while, it provided interpretable results to make the model trustworthy and credible. Key Outcomes of the Model New technologies, particularly AI and ML, are changing the way the world functions. However, due to a previously rampant black box approach, human faith in these tools was low. This model aims to change perspectives and throw light on a new direction for AI in education. Some of the key outcomes were: Deriving interpretable results for humans was becoming increasingly important. Business decision-makers wanted to feel empowered in their capabilities, but unsupportive AI didn’t help. With the new model employing XAI and sharing interpretable results, decision making will become an empowering and data-driven process. In other places, GDPR policies and country-specific data regulations have pushed organizations to use XAI, so it aligns with privacy guidelines. In all of the previous literature work covered, an interpretable model in the education domain was missing. This research establishes a new model based on SHAP XAI, the first step to restoring trust in AI in the field. With SHAP XAI, stakeholders in the education industry can use data backed by reasoning for better policy making and improving students’ performance. The model has generated proven results through consistent testing and applications to various black box models. As a next step, a scalable interpretable and explainable model with its efficacy will boost adoption among stakeholders. ___________________________________ Manjari Chitti is an upGrad learner, and as a part of her program, she has developed the thesis report titled — Interpretable Student Performance Predictions.","excerpt":"This is one of the top voted thesis papers from upGrad’s online working professional programs in partnership with one of the UK’s leading universities.","categories":["AI Trends"],"tags":["students","UpGrad","UpGrad India"],"author_name":"Manjari Chitti Chitti","publish_date":"2021-09-23T18:42:06","publication_year":"2021","word_count":1210,"keywords":["data science","artificial intelligence","machine learning","AI","neural network","ML","students","Aim","XGBoost","UpGrad","analytics","xAI","UpGrad India"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","data science","analytics","xAI","Aim","XGBoost"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/interpretable-student-performance-predictions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10019235,"title":"img2pose: Guide to Face Alignment, Detection and Pose Estimation using 6DoF","content":"We have already seen a couple of pose estimation and face detection techniques in our previous article like 3ddfa-v2, OpenPose, Nvidia Imaginaire(Image & Video translation GAN Library), Yolov5, OneNet and many more. Today we will see a new face pose estimation, detection and alignment technique that uses 6DoF(degree of freedom) and 3D face estimation without face landmark\/detection localization. The paper has been published by Vitor Albiero, Xingyu Chen2, Xi Yin2, Guan Pang, and Tal Hassner of Notre Dame University. According to them, using the 6DoF method is more reliable than face bounding box labels. More specifically, they did mainly three contributions to achieve the SOTA results: (a) Introduced an easily trained, reliable, Faster R-CNN model that uses the 6DoF method for all faces in the photo, without using any face detection box technique. (b) Authors explained how the pose is converted and processed. (c) Also they have explained how face poses can really replace face bounding box training labels. 6DoF(degree of freedom) pose With the 6DoF pose method, it is very easy to estimate faces than using landmark detection. Now calculating face pose is a 6D regression problem. In contrast, authors estimate the face poses without considering that faces that were already detected. 6DoF pose labels predict more than just face bounding box landmarks. Also it can be further converted to a 3D to 2D projection matrix. 6DoF provides information about the face’s 3D position and orientation. 6DoF poses method is calculated by img2pose to capture the positions of all faces in the picture, for example, we can take the below picture. their 3D scene locations are shown below: img2pose network architecture The network follows a two-level approach based on the Faster R-CNN model. The first level is a region proposal network (RPN) with a feature pyramid(FPN), which aims at inherent face locations in the photo. The second level\/stage of img2pose network architecture extracts features from every proposal with the region of interest (ROI) pooling. It then transfers them to two distinct heads: a regular face\/no-face classifier and a new purposed 6DoF face pose regressor. The model is trained on the WIDER FACE dataset. The dataset mainly offers annotated bounding box labels, with no annotation for poses, but another RetinaFace set provides annotated 5 points facial landmarks for 76000 WIDER FACE dataset faces. Implementation The project is trained and evaluated using PyTorch so let’s see how to reproduce the results. First, we will start by cloning the project and necessary package installation and environment setup, use below commands to do so. !git clone https:\/\/github.com\/vitoralbiero\/img2pose.git ! cd img2pose Install dependencies pip install -r requirements.txt # move back to the main directory if inside cd.. # install renderer to visualize prepedition cd Sim3DR sh build_sim3dr.sh Imports and necessary steps import sys sys.path.append('..\/..\/') import numpy as np import torch from torchvision import transforms from matplotlib import pyplot as plt from tqdm.notebook import tqdm from PIL import Image, ImageOps import matplotlib.patches as patches from scipy.spatial.transform import Rotation import pandas as pd from scipy.spatial import distance import time import os import math import scipy.io as sio from utils.renderer import Renderer from utils.image_operations import expand_bbox_rectangle from utils.pose_operations import get_pose from img2pose import img2poseModel from model_loader import load_model np.set_printoptions(suppress=True) def render_plot(img, poses, bboxes): (w, h) = img.size image_intrinsics = np.array([[w + h, 0, w \/\/ 2], [0, w + h, h \/\/ 2], [0, 0, 1]]) trans_vertices = renderer.transform_vertices(img, poses) img = renderer.render(img, trans_vertices, alpha=1) plt.figure(figsize=(8, 8)) for bbox in bboxes: plt.gca().add_patch(patches.Rectangle((bbox[0], bbox[1]), bbox[2] - bbox[0], bbox[3] - bbox[1],linewidth=3,edgecolor='b',facecolor='none')) plt.imshow(img) plt.show()renderer = Renderer( vertices_path=\"..\/..\/pose_references\/vertices_trans.npy\", triangles_path=\"..\/..\/pose_references\/triangles.npy\" ) threed_points = np.load('..\/..\/pose_references\/reference_3d_68_points_trans.npy') Training Img2pose is implemented in PyTorch with having ResNet-18 backbone, it uses stochastic gradient descent(SGD) comprised with mini-batch in two pictures. The First 256 proposals are sampled for the RPN loss and 512\/image for the pose head losses. Note: On a single NVIDIA Quadro RTX 6000 machine, training takes almost 4 days. Preparing dataset Download the WIDER FACE dataset from here Extract it inside dataset\/WIDER_Face. Now to run train and validation(LMDB), run below script. python3 convert_json_list_to_lmdb.py --json_list .\/annotations\/WIDER_train_annotations.txt --dataset_path .\/datasets\/WIDER_Face\/WIDER_train\/images\/ --dest .\/datasets\/lmdb\/ -—train The above code will generate an LMDB dataset, which contains images with annotations and also produce a pose mean and std, files. Now, let’s create LMDB containing validation images within annoatations. python3 convert_json_list_to_lmdb.py --json_list .\/annotations\/WIDER_val_annotations.txt --dataset_path .\/datasets\/WIDER_Face\/WIDER_val\/images\/ --dest .\/datasets\/lmdb Train CUDA_VISIBLE_DEVICES=0 python3 train.py --pose_mean .\/datasets\/lmdb\/WIDER_train_annotations_pose_mean.npy --pose_stddev .\/datasets\/lmdb\/WIDER_train_annotations_pose_stddev.npy --workspace .\/workspace\/ --train_source .\/datasets\/lmdb\/WIDER_train_annotations.lmdb --val_source .\/datasets\/lmdb\/WIDER_val_annotations.lmdb --prefix trial_1 --batch_size 2 --lr_plateau --early_stop --random_flip --random_crop --max_size 1400 Testing If you can’t wait for days to finish training or don’t have a powerful GPU you can always download the pre-trained model from model ZOO here and extract it to your main directory of img2pose. Download arXiv model Visualizing trained model To test the trained model run the notebook: Visualize_trained_model_predictions Notebook threshold = 0.8 total_imgs = 20 data_iter = iter(lmdb_data_loader) for j in tqdm(range(total_imgs)): torch_img, target = next(data_iter) target = target[0] bboxes = [] scores = [] poses = [] img = torch_img[0] img = img.squeeze() img = transforms.ToPILImage()(img).convert(\"RGB\") ori_img = img.copy() run_img = img.copy() w, h = img.size min_size = min(w, h) max_size = max(w, h) # run on the original image size img2pose_model.fpn_model.module.set_max_min_size(max_size, min_size) res = img2pose_model.predict([transform(run_img)]) res = res[0] for i in range(len(res[\"scores\"])): if res[\"scores\"][i] > threshold: bboxes.append(res[\"boxes\"].cpu().numpy()[i].astype('int')) scores.append(res[\"scores\"].cpu().numpy()[i].astype('float')) poses.append(res[\"dofs\"].cpu().numpy()[i].astype('float')) (w, h) = img.size image_intrinsics = np.array([[w + h, 0, w \/\/ 2], [0, w + h, h \/\/ 2], [0, 0, 1]]) plt.figure(figsize=(16, 16)) poses = np.asarray(poses) bboxes = np.asarray(bboxes) scores = np.asarray(scores) if np.ndim(bboxes) == 1 and len(bboxes) > 0: bboxes = bboxes[np.newaxis, :] poses = poses[np.newaxis, :] if len(bboxes) != 0: ranked = np.argsort(poses[:, 5])[::-1] poses = poses[ranked] bboxes = bboxes[ranked] scores = scores[ranked] for i in range(len(scores)): if scores[i] > threshold: bbox = bboxes[i] pose_pred = poses[i] pose_pred = np.asarray(pose_pred.squeeze()) trans_vertices = renderer.transform_vertices(img, [pose_pred]) img = renderer.render(img, trans_vertices, alpha=1) plt.gca().add_patch(patches.Rectangle((bbox[0], bbox[1]), bbox[2] - bbox[0], bbox[3] - bbox[1],linewidth=3,edgecolor='b',facecolor='none')) img = Image.fromarray(img) plt.imshow(img) plt.show() AFLW2000-3D dataset evaluation You can Download the AFLW2000-3D dataset and extract it to datasets\/AFLW2000. Run the notebook foir aflw_2000_3d_evaluation. BIWI dataset evaluation Same you can Download the BIWI dataset and extract it to datasets\/BIWI. And then Run the notebook biwi_evaluation. Testing on your own images Run following notebook test_own_images. Conclusion We learned a novel approach to 6DoF pose estimation and face alignment, that does not rely on any face detector or localizing facial landmarks. This is the first multi-pose, multi-face, direct approach for complex images. To learn more about pose estimation and computer vision techniques, you can check out the below resources. Read More: Guide Towards Fast, Accurate, and Stable 3D Dense Face Alignment(3DDFA-V2) Framework","excerpt":"We have already seen a couple of pose estimation and face detection techniques in our previous article like 3ddfa-v2, OpenPose, Nvidia Imaginaire(Image & Video translation GAN Library), Yolov5, OneNet and many more. Today we will see a new face pose estimation, detection and alignment technique that uses 6DoF(degree of freedom) and 3D face estimation without […]","categories":["Deep Tech"],"tags":["pose estimation","Visualize Spatial Data"],"author_name":"Mohit Maithani","publish_date":"2021-01-29T12:00:00","publication_year":"2021","word_count":1093,"keywords":["CUDA","NumPy","pose estimation","AI","PyTorch","computer vision","Visualize Spatial Data","Ray","Aim","Python","Matplotlib","Pandas"],"extracted_tech_keywords":["AI","computer vision","Aim","Ray","PyTorch","Pandas","NumPy","Matplotlib","CUDA","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/img2pose-guide-to-face-alignment-detection-and-pose-estimation-using-6dof\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":46334,"title":"How To Become A Data Analyst In 2020","content":"The modern world is currently at its peak in terms of technology and technological advancements. There are very few things left that are not completely or partially technology-driven. Even the most basic of human tasks such as bathing too have become technology-oriented to quite an extent and the very basis of technology driving the world to a sustainable future is a mere 4-letter word; a word that we have all studied in elementary computer science and I bet that we all took it as the most inconsequential concept in the world of computers. So do you want to know what that word is? Are you ready for it? Wait. Drumrolls. The answer is data. Yes; this is the same data that we used to define as “pieces of facts and numbers that are unprocessed and make no sense”. But did any of us ever think that this unprocessed information would be at the helm of technological pioneers in the not so distant future; of course not. But the big question over here is how something so inconsequential witness such a meteoric rise! The year was 2012 and the world of big data was starting to gain traction. With the Internet being made relatively cheap, more and more people were getting connected to the world’s largest network as a consequence of which the data generated by the world population was rising exponentially and even the biggest of IT giants were finding it difficult to cope up with such large streams of data flowing into their servers; difficult in terms of storage and processing. This is where Big Data came in strong to offer its services through its plethora of tools and applications that offered pragmatic solutions to the aforementioned problems. As Big Data gained its footing in the industry with almost every IT company taking help of its functionalities, a new pattern or trend, as you may know, was discovered. Processing of customer data, companies and tech giants came to a solid conclusion that the same data contained answers to a lot of their problems. But the problem with this finding was that the dataset was too large to be analysed in a single go. Another problem was how to filter out the relevant stuff from such colossal piles of data and furthermore, what to do with the findings. The very same problems led to a whole other domain of computational study to deal with the same. This domain is what we refer to as ‘data science’ these days. Data science, as the name suggests, is based completely on data; in fact, this is the same data that we discussed just recently. Data science as a concept is all about using large sets of customer data to find a behavioural pattern as per the needs of the company and subsequently use the discovered patterns to solve the particular business problem. What does it take to be a data analyst? At present, in the technical world, being a data analyst is the most rewarding of professions both in terms of growth as well as money. But the grass is always greener on the other side. On paper and in theory, becoming a data analyst seems to be too easy a task and to be very honest, completing a data analysis course and calling yourself a data scientist is actually easy. But what differentiates a good data analyst from a mediocre one is the command and the proficiency that one holds over the various tools and applications that a data scientist uses daily. So, to become a data analyst not just for name but to become one such that you become a standard in yourself in the industry, following are some of the requirements with which you should be affiliated with from in and out: Programming Language As mentioned earlier, data science is all about finding logics and patterns underneath a mountain of data. Sieving through this ‘mountain’ is just not possible solely by human labour and so the only logical answer to derive said patterns is to turn to the powerful computers of today. But even computers cannot function on their own right! Even they need a set of instructions on which they can act accordingly and derive the appropriate results. These set of instructions are given to computers with the help of a piece of code written in a high-level language. At present, the most powerful and advanced languages to design models for data science and deal with the sophisticated level of statistics involved in data science include Python, R\/R-Studio, Java, SQL, MATLAB, etc. Amongst these languages, the most popular among data analysts is Python owing to its dynamic behaviour and a vast range of powerful libraries that do even the most complex of calculations in a jiffy. Stats And Aptitude Data science is designing models and writing code later and logical reasoning, maths, and numbers before. The thing with data science projects is that all of them are unique in their own way and the purpose of every single one of them is distinct too, which means that for the same dataset, 2 different projects need 2 different approaches and to devise these 2 distinct approaches, numbers and figures in the dataset have to be viewed from a whole other perspective and it is to look at things with a different perspective, a data analyst needs to constantly think out of the box which can be made possible only when the brain has been trained to think that way. Database Data is stored in databases (or in data centres as per the current technological needs) and to continuously deal with such large sets of data, a data analyst needs to have the basic concepts of database management on their fingertips. To work and communicate with the given data, a data analyst needs to be proficient in some database language (such as SQL, NoSQL, Swift, C#, etc.) and carry out the desired analytics required. Another reason why data analysts need to be excellent with databases is that continuous fetching & modification of data and subsequently having to write the changes to the physical database is both time as well as energy-consuming and data analysts are also tasked with making this process efficient (time and energy-wise). Machine\/Deep Learning & AI Machine learning may be considered as a subset of data science. It is the fastest growing and emerging technology of the modern world and offers services that can make your work a whole lot easier by automating a given task to its maximum extent. As its name suggests, Machine Learning is about machines learning something. This learning of theirs is again made possible by code written in programming languages and if the designed model works the way it is supposed to, it can make your task easier, more efficient, and more accurate than Big Data and cloud solutions. Data Visualisation Once a project has been completed with the help of the models that were designed, the algorithms that were used, and all of the other stuff that goes into successfully executing a project. However, there occurs a major obstacle in the end, which is to make your clients and the end-users understand the results and findings of your work. But what’s the obstacle over here, you may ask. Well, the hurdle is that you are a data analyst, but your clients are not. They do not understand what your model means or what your code is trying to convey. They can only understand the results if it is conveyed in a human-readable format. This is where data visualisation comes in. Using tools such as Excel, data analysts need to display the conclusion of the project in the form of bar graphs, pie charts, etc. in an accessible format to understand the trends and patterns identified in the data. Data Munging Before a data analyst can start with his work, there is a major task to be accomplished ahead. The data on which analytics has to be done is very large as we have discussed a lot of times. But what we haven’t noted is the randomness and lack of structure in the same. These 2 factors make reading and understanding the data more difficult than it already is. Putting this raw data in a format with the intent of making it more valuable is what is called data munging.","excerpt":"The modern world is currently at its peak in terms of technology and technological advancements. There are very few things left that are not completely or partially technology-driven. Even the most basic of human tasks such as bathing too have become technology-oriented to quite an extent and the very basis of technology driving the world […]","categories":["AI Features"],"tags":["Data Analysis","Data Science","recent technological advancements"],"author_name":"Ram Tavva","publish_date":"2019-09-25T10:54:31","publication_year":"2019","word_count":1397,"keywords":["recent technological advancements","Data Analysis","data science","Go","machine learning","AI","RAG","Python","deep learning","analytics","SQL","Data Science","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","analytics","RAG","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-become-a-data-analyst-in-2020\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10075843,"title":"Is AI Ethics Just An Eyewash?","content":"Desperate times call for desperate measures and no other big tech company is feeling the heat more than Meta Platforms Inc. A report published by Wall Street Journal last week revealed the strict new policy it has imposed on some employees asking them to either look for new positions somewhere else within the company or face termination. Meta has announced that it plans to cut costs by 10%. In the earnings released for the previous quarter, Meta’s results looked grim. The company had lost close to 50% of its value by the second-quarter of this year. The company also reported an outlook predicting higher-than-expected losses for the third-quarter. In a bid to rid itself of all excesses, the axe fell first on the company’s Responsible Innovation Team (RIT). The team was a crucial part of Meta’s efforts to redress the many blows that have been dealt to its reputation in the past few years. The company has had more than its fair share of scandals including Cambridge Analytica—which was recently settled—breeding political extremists and spreading misinformation during the US elections, violation of children’s privacy in Ireland and staking its money on the metaverse. Margaret Gould Stewart, VP of Product Design & former head of Responsible Innovation at Meta, Source: Ted Turbulent times in Meta In 2018, a vice president of product design with the company—Margaret Stewart—established the team to tackle the “potential harms to society” caused by Facebook’s products. Ironically, just last year, Stewart posted a blog titled, ‘Why I’m optimistic about Facebook’s Responsible Innovation efforts’, stating that she inherently believed that a lot of good could come from technology and Meta was ready to put in the work for it. “Goodness isn’t inevitable. It comes through sustained hard work, investing time in foresight work early in the development process, surfacing and planning mitigations for potential harms, struggling through complex trade-offs, and all the while engaging with external stakeholders, including members of affected communities, “ Stewart explained. Despite dissolving the team, Meta has promised that the team which comprised two dozen engineers and ethic specialists will continue with its work albeit in a scattered way. Eric Porterfield, a spokesman with the company, said that employees from the RI team would work in safety and ethical product design with specific issues in teams. He also stated that they weren’t guaranteed new jobs. How real are AI ethical teams in companies? While most media reports wasted no time in underscoring Meta’s readiness to let go of its ethical division, there is a section of AI experts who question the motivation behind an AI ethics team in the first place. Is it mainly PR motivated and a ploy to distract from the actual troubles with the business? Meta has finally had enough of phony AI ethics.https:\/\/t.co\/mSdBFXeVu6— Pedro Domingos (@pmddomingos) September 8, 2022 Pedro Domingos, author of ‘The Master Algorithm,’ widely known for his work on Markov logic network, has long been critical of the activism of AI ethicists like former Google scientist Timnit Gebru. Domingos applauded Meta’s decision to disband the RI team, calling AI ethics “phony.” The University of Washington professor has often called AI ethics a unidirectional field which isn’t welcoming of differing opinions. Domingos’ concerns aren’t entirely unfounded. For an AI startup or company to jump onto the AI ethics bandwagon is remarkably easy. The company’s management and marketing teams claim that they strictly adhere to the Ethical AI guidelines without due diligence. The practice has gained enough popularity to acquire a title for it, called ‘AI Ethics washing’ and includes having an ethical AI division as window dressing to silence knee-jerk criticism. What is AI ethics washing? There is a good reason for the rise of ethical washing. Building an ethical framework and incorporating it within a business is a costly process. Up until a few years back, when the concept of AI ethics was still nascent, tech company leaders expressed their reluctance openly. Ethics is a complicated minefield that isn’t necessarily navigated with ease. In 2019, Microsoft’s president and attorney, Brad Smith, plainly said—even as a group of Microsoft employees protested against the company’s military contracts—that American tech companies had a long history of supporting the US military and that Microsoft would continue to do so. “The U.S. military is charged with protecting the freedoms of this country. We have to stand by the people who are risking their lives,” Smith said. Google’s AI drone program called ‘Project Maven’ for Pentagon, Source: Getty Images In 2018, Google was pulled up for providing AI solutions to build warfare to the US Department of Defense. The pilot programme called ‘Project Maven’, which involved other tech companies as well, would help the US government analyse drone footage using AI. Google eventually stepped back after a bunch of resignations and internal dissent. With such deep involvement of governments, is it even possible to have transparent AI ethics? It is these fallacies that Domingos and others want to examine. The Wall Street Journal report mentioned Zvika Krieger, former RIT head, who revealed that the team had been effective in small ways, not the overarching beacon that it was meant to be. The team had been previously involved in Facebook’s decision to exclude a race filter in dating profiles. The feature was later copied and put into use by dating apps. Stewart also mentioned in her blog that the RI team was behind Meta’s COVID-19 products. The team wanted to “fight misinformation about the virus and whether a feature could be unintentionally offensive or insensitive”. However, even with these positive undertakings, Meta was drowning under a pile of snafus. In this context, is it better to simply discard pretences and put a concentrated focus on real ethical issues as Domingos says? Or is a front necessary?","excerpt":"Domingos applauded Meta’s decision to disband the RI team, calling AI ethics “phony.”","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Poulomi Chatterjee","publish_date":"2022-09-26T18:00:00","publication_year":"2022","word_count":962,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","AI ethics","Aim","ViT","AI Tool","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","ViT","AI ethics","innovation","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-ai-ethics-just-an-eyewash\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":29538,"title":"Inside Facebook’s Analytics And Data Science Play In Marketing Measurement","content":"Analytics India Magazine had an exclusive chat with Pratham Hegde who heads the Marketing Science unit for Facebook India. He has 18 years of experience in Data analytics and specialises in Marketing Measurement. In this conversation he talked to us about how data science and analytics is used in marketing measurement at Facebook and how the role of analytics has evolved over the years. Before his role at Facebook, Hegde headed the Marketing Analytics Practice at Genpact, then moved on to set up a Marketing Measurement team for Amazon for their major global markets. Analytics India Magazine: How is data science and analytics used in marketing measurement at Facebook? Pratham Hegde: Data science is at the core of what we do, which is to quantifiably measure the impact of ads on Facebook on business metrics (brand or sales metrics). Hence, the tools and solutions we deploy for marketing measurement are grounded on the principles of experimental design and statistical significance of results — ultimately everything boils down to that. A lot of data science and research goes into our measurement products, they are built and tested to ensure they yield unbiased results and they can be deployed in a scalable manner. We also use a lot of custom solutions working with third party providers; we use advanced statistical models to understand the impact of media in driving business metrics, and isolating the impact signal from the noise. AIM: How has the role of analytics evolved over the years in driving insights through data and consumer research? PH: In general, advertisers are getting more sophisticated in measuring the impact of marketing activities on their business. But the journey is hard and requires patience. If you think about it, marketing has traditionally been thought of as a “right brain” centered function where one had to be creative for the brand stand out in the crowd and be noticed. That is no longer the case. With the rapid move to mobile and digital, marketers who have a flair for data and new technology are able to leverage available tools well and drive impact for their business, while those hanging on to just offline channels are still stuck at asking the question “does digital work?”. However, in India the available tools for measurement are still at a nascent stage and the industry is still at an early stage of the journey is being able to quantify the impact of advertising. For example, the kind of closed loop experiments or panel based solutions that are able to generate single source measurement of TV and Digital are still not available in India. I find that in India we always ask the tough questions but the resource commitment and patience to get to the truth via hard measurement principles might not always be there. AIM: How has analytics and data science become crucial to the social media industry? Would you like to highlight few use cases? PH: I think the term “social media marketing” can mislead people. Paid advertising on social media is going on globally at an unprecedented scale and is set to eclipse many traditional channels in terms of both reach and ability to drive impact. I see a lot of folks are still depending on social channels to drive “social buzz” or “social engagement”. At Facebook measurement we have many studies that prove that intermediate metrics like likes, shares, click don’t necessarily correlate to business metrics such as brand metric movement or sales impact, which is what we ultimately want to drive for the business. So, we encourage advertisers to think of Facebook just like any other media channel and apply the same measurement principles as they would do for any other marketing spend. The good thing with digital is that it allows you to precisely measure the impact that a digital campaign had (or did not have) in terms of the business outcome metrics. We have hundreds of examples for Lift studies done in India which are both Facebook and third party measured, that show positive significant impact using control exposed methods. In addition, we also have a number of matched market tests and mix models that provide a read on the direct impact that Facebook advertising had on sales. On the Performance marketing side, we have a number of examples that measured positive lift in add to cart, sales, app install and other outcomes of interest for advertisers. We also have examples where we used innovative ways to close loop and prove that facebook ads drove in-store sales for retailers. AIM: What are the challenges you face in your current role? How do you overcome these challenges? PH: In general, there is a lack of understanding of robust advertising measurement approaches, tools and principles in the Indian market. We find many advertisers who are wedded to offline marketing and hence offline measurement tools such as brand tracks that rely on face to face interviews to measure brand metric movement. There is a tendency to use methods that help make marketing look good Vs trying to apply robust measurement principles that measure the true impact of media ; that then help optimise marketing spends. So we spend a lot of time in just educating the advertisers and agencies on fundamental measurement principles, and taking them down the path of continuous test and learn measurement – it’s hard and requires a lot of patience but the light at the end of the tunnel is the ability to accurately measure the impact of marketing monies ; which in turn ensures that valuable funds deployed in marketing are generating adequate returns for the business. The other big challenge that we face is to keep things simple. While advertisers are asking more complex questions that lead us to deploying more complex tests, it is very easy to get lost in the methodology and forget to translate the outcomes into simple business language. Ultimately if you cannot break the method and results down into consumable bytes that advertisers can use in decision making, the most sophisticated measurement experiment can become useless. AIM: What is the roadmap for analytics in marketing measurement at Facebook? PH: Our goal is to continuously build solutions and tools that help advertisers measure the true value of Facebook advertising. Since an absolute “Facebook only” measure can only go so far in helping the decision maker, we also focus on methods that help measure Facebook together with all other channels. Our long term goal is to provide every advertiser the ability to measure campaigns in self-serve mode. Our measurement tools rely on the unique “People based” platform and capabilities at Facebook, and are arguably the most accurate and best grounded in the first principles of measurement in the advertising industry.","excerpt":"Analytics India Magazine had an exclusive chat with Pratham Hegde who heads the Marketing Science unit for Facebook India. He has 18 years of experience in Data analytics and specialises in Marketing Measurement. In this conversation he talked to us about how data science and analytics is used in marketing measurement at Facebook and how […]","categories":["AI Features"],"tags":["best book to learn digital marketing","facebook analytics","Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-10-24T11:10:42","publication_year":"2018","word_count":1125,"keywords":["data science","Go","API","AI","Scala","Git","RAG","Aim","analytics","best book to learn digital marketing","facebook analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Scala","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/inside-facebooks-analytics-and-data-science-play-in-marketing-measurement\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10074446,"title":"TikTok Hijacking, Over 2 Billion Users Struggling With a Lost Data Record: Security Researchers","content":"Cyber security researchers on Monday unveiled a possible data hack in the Chinese short-form video app, TikTok. The hack is alleged to involve the database records of almost 2 billion users. On the same day, Twitter encountered posts from several cyber security analysts on “discovering a security breach of an insecure server that propelled access to TikTok’s storage, which contained personal user data”. “This is your forewarning. #TikTok has reportedly suffered a #data #breach; if true, there may be fallout from it in the coming days. We recommend you change your TikTok #password and enable Two-Factor Authentication if you have not done so already,” tweeted BeeHive CyberSecurity. It further mentioned that it has inspected a sample of the extracted data and communicated via email to private clients and email subscribers. A TikTok spokesperson reportedly said that their team “inspected this issue and ascertained that the code in question is irrelevant to TikTok’s backend source code.” The issue was further quoted by the Microsoft 365 defender research team—tracking down a susceptibility in the TikTok app for Android, allowing hackers to take over the personal data of millions of users with a single malicious link. The Chinese company is said to have denied all the allegations, after hackers claim to have user data and source code. The susceptibility can cause harm to millions of users, exploiting privacy by providing leverage to security attackers, as expressed by tech giants in a statement last week.","excerpt":"A TikTok spokesperson reportedly said that their team “inspected this issue and ascertained that the code in question is irrelevant to TikTok’s backend source code.”","categories":["AI News"],"tags":["Cyber Security","Security"],"author_name":"Nidhi Bhardwaj","publish_date":"2022-09-06T14:55:40","publication_year":"2022","word_count":241,"keywords":["Cyber Security","programming_languages:R","AI","Security","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tiktok-hijacking-over-2-billion-users-struggling-with-a-lost-data-record-security-researchers\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10014430,"title":"Rolls-Royce Releases AI Toolkit To Set A New Standard For Ethics &#038; Trustworthiness","content":"In a breakthrough development, Rolls-Royce has released a comprehensive ethical and trustworthiness framework, that will help businesses to check and balance their AI projects in a fair, ethical and trustworthy manner. Available for free to download from the Rolls-Royce website, the toolkit is called Aletheia FrameworkTM, which has been named after the Greek goddess of trust and disclosure. It aims to address one of the biggest barriers to the widespread use of AI – mistrust. It is especially timely when the debate around ethical AI is going on in full throttle. The company explains that the Aletheia Framework™ is essentially a checklist that invites organisations to consider the impacts of using artificial intelligence prior to deciding whether to proceed. It looks across a total of 32 facets of societal impact, governance and trust, and transparency and requires executives and boards to provide evidence that these have been rigorously considered. Further, the framework includes a five-step continuous automated checking process, which tracks the decisions the AI is making to detect bias or malfunction. Once fully implemented, businesses that follow the checks and balances within it can assure themselves that their AI projects are fair, trustworthy and ethical. It aims to help business leaders, academics, technologists and even philosophers to explore the potential of AI while keeping ethics and trustworthiness at the forefront. “As we move rapidly to a more digital world, people have a right to expect that AI is used ethically and that it is a trustworthy partner for people and society. The Aletheia FrameworkTM provides the foundations for that to happen,” Warren East, Chief Executive Officer, Rolls-Royce said. Notably, Rolls-Royce has used AI for years to analyse more than 70 trillion data points to decide fly time of engineers and its sustainability. “We’re now developing AI for quality inspections of critical components and it’s in the justification of applying artificial intelligence technologies to this activity that we have had to challenge ourselves to ensure it’s the right thing to do, and that it’s trustworthy,” said Caroline Gorski, Group Director of R2 Data Labs. The Aletheia FrameworkTM has also been shared with experts from UNESCO, which is undertaking a global consultation on AI ethics. “AI will radically change our world, but it is ethics that will shape how it looks. All stakeholders, whether public or private, must work on this together,” said Gabriela Ramos, Assistant Director-General for the Social and Human Sciences sector at UNESCO.","excerpt":"In a breakthrough development, Rolls-Royce has released a comprehensive ethical and trustworthiness framework, that will help businesses to check and balance their AI projects in a fair, ethical and trustworthy manner.  Available for free to download from the Rolls-Royce website, the toolkit is called Aletheia FrameworkTM, which has been named after the Greek goddess of […]","categories":["AI News"],"tags":["Ethical AI"],"author_name":"Srishti Deoras","publish_date":"2020-12-15T18:22:53","publication_year":"2020","word_count":405,"keywords":["Go","API","artificial intelligence","AI","Ethical AI","Git","Aim","ViT","Rust","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","Rust","Git","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/rolls-royce-releases-ai-toolkit-to-set-a-new-standard-for-ethics-trustworthiness\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":42183,"title":"8 Ways You Didn’t Know Quantum Technology Is Used In Everyday Lives","content":"Researchers and tech giants have been working in the field of quantum physics for over decades now. The quantum technology has enormous potential that brings to life the vision of time travel, unprecedented computing and more. While quantum technology has been explored by researchers for some amazing researches, we list down eight ways where this technology has been used in our everyday lives that often goes unnoticed. 1| Electronic Appliances We have been using quantum techniques in our household needs in the form of electrical appliances such as toasters. When we toast bread in a toaster, we notice the red glow which is because of quantum physics. The heating element emits red colour and the hot objects emit glow are the reasons how quantum physics came into existence. 2| Sensors Everybody loves to capture their precious moments and for this most of us carry a digital camera wherever we go. In digital cameras, the lens collects the photons and focuses them on the sensor which in turn emits electrons through the manipulation of energy levels in semiconductors. The rules of semiconductors are dictated by quantum physics, bringing a second use case of quantum technology. 3| Computers Computers are being widely used for many years and the interesting part about computers is that it solely works on the principle of quantum technology. The basic element inside a computer is known as a transistor which utilises the basic electronic properties of Silicon. It is known to all that Silicon is a semiconductor and the theory of solid-state physics is based on the foundation of quantum mechanics. Electrons in semiconducting solids behave like a wave, thus every computer in this world uses quantum mechanics. 4| Light Emitting Diodes (LEDs) Light Emitting Diode is a light source which is based on two layers of semiconductors which contain electrons and holes. We have already mentioned that semiconductors work on the principle of quantum physics. When an LED is connected to a battery, the two layers of semiconductor meet and releases energy in the form of bright light. Thus we are using the quantum technique in the form of a light bulb, television, etc. 5| Incandescent Bulb We have been using the incandescent bulbs for over a hundred years now. The working of this bulb is quite similar to the toasters. When the current passes through a thin wire which is tightly coiled inside the glass bulb, it makes the wire hot and the bulb to glow. This again works on quantum mechanics. 6| Laser A laser is an optical device which has many significant applications in medicines, communications, industries, science and technology, etc. In medicine, it can be used in cancer diagnosis, removing tumours, removing kidney stones, etc. It can also be used in optical fibre communications, space communications, etc. It emits monochromatic light through a method of optical amplification which is based on the stimulated emission of photons, which in turn can be said as the quantum mechanics. 7| GPS GPS or Global Positioning System is a network of satellites that has made finding locations and directions quite easy. The application of GPS is dependent on quantum physics. Each satellite in the GPS constellation includes an ensemble of atomic clocks and these atomic clocks use the principles of quantum theory to measure time. 8| MRI Scans Magnetic Resonance Imaging or MRI works by flipping the spins in the nuclei of hydrogen atoms. This procedure involves the spins of the electrons in hydrogen nuclei and the spins are nothing but an application of quantum physics where special relativity is included with quantum mechanics.","excerpt":"Researchers and tech giants have been working in the field of quantum physics for over decades now. The quantum technology has enormous potential that brings to life the vision of time travel, unprecedented computing and more. While quantum technology has been explored by researchers for some amazing researches, we list down eight ways where this […]","categories":["AI Trends"],"tags":["quantum mechanics","quantum physics","quantum technology"],"author_name":"Ambika Choudhury","publish_date":"2019-07-10T14:20:09","publication_year":"2019","word_count":598,"keywords":["Go","programming_languages:R","AI","R","programming_languages:Go","Git","quantum physics","ViT","Chroma","quantum mechanics","quantum technology"],"extracted_tech_keywords":["AI","Chroma","R","Go","Git","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-ways-you-didnt-know-quantum-technology-is-used-in-everyday-lives\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10022755,"title":"Guide to SelfTime: Self-supervised Time Series Representation Learning Framework with Python code","content":"Time-series forecasting is one of the most widely dealt with machine learning problems ever. Time series forecasting finds crucial applications in various fields including signal communication, climate, space science, healthcare, financial and marketing industries. Deep learning models outshine in time series analysis nowadays with great performance in various public datasets. The key idea of deep learning models is to learn the inter-sample relationships to predict the future. However, intra-temporal relationships among different features within a sample are hardly dealt with. Extracting intra-temporal relationships yields better pattern understanding in the time series representation learning even with fewer data samples. Without an intra-temporal relationship extraction system, a deep learning model becomes data-hungry and requires a lot of manually-annotated tabular data for supervised training. Present self-supervised representation learning approaches let models train with unannotated as well as fewer data. Self-supervised learning approaches yield huge success in different computer vision tasks including image inpainting, video inpainting, object tracking, rotation prediction and pace prediction. However, these approaches fail to capture the complete temporal-relationships in the case of high-dimensional time series structured data. Haoyi Fan and Fengbin Zhang of Harbin University of Science and Technology, China and Yue Gao of Tsinghua University, China introduced SelfTime, the Self-supervised Time Series Representation Framework. This framework explores inter-sample relationships among subsequent samples and intra-temporal relationships within each sample to capture the underlying spatial and temporal patterns in an unannotated structured time series dataset. How does SelfTime work? A sample or a piece of a sample taken for exploring the pattern is called an anchor sample or an anchor piece respectively. In the case of inter-sample relation reasoning, a sample is taken as an anchor and is transformed using the standard augmentation method to develop a new transformed-time-series sample that is denoted the positive sample. Another real sample is taken from some nearer time scale and is denoted the negative sample. Inter-sample relation reasoning is captured between these positive and negative samples. Concept of inter-sample relation reasoning and Intra-temporal relation reasoning (Source) On the other hand, a sample of interest is broken into pieces to learn intra-temporal relationships. One of the broken pieces is taken as an anchor and the rest are called references. Intra-temporal relation reasoning is performed between the anchor and each of the references individually. The number of references for each anchor can be varied based on the problem. If three references are chosen for analysis, it is termed 3-scale temporal relations and the references are called short-term, middle-term and long-term relations based on the temporal distance between the anchor and references. A shared representation learning backbone is developed with two relation-reasoning-heads based on the inter-sample relations and the intra-sample relations extracted separately. Finally, the time series representations are extracted from unlabeled data by supervising the two different relation-reasoning-heads. Since SelfTime explores in-depth relationships among different data samples and different time pieces, it exhibits extraordinary performance in representation learning of time series data, especially high-dimensional data. Architecture of SelfTime framework (Source) Python implementation of SelfTime The desired requirements for SelfTime architecture building are Python 3.6 or 3.7, Pytorch 1.4.0 and CUDA GPU runtime. The following command downloads the source code to the local machine. !git clone https:\/\/github.com\/haoyfan\/SelfTime.git Output: Verify proper download by exploring the SelfTime directory using the command, !ls SelfTime Output: Change directory to run the containing files using the following line-magic command %cd SelfTime\/ The following command pre-trains the inter-sample relation head-on in-built CricketX dataset !python train_ssl.py --dataset_name CricketX --model_name InterSample The following command pre-trains the intra-temporal relation head-on in-built CricketX dataset !python train_ssl.py --dataset_name CricketX --model_name IntraTemporal The following command pre-trains SelfTime model on in-built CricketX dataset !python train_ssl.py --dataset_name CricketX --model_name SelfTime Output: Once trained the framework parts separately, the three parts of the framework can be evaluated on the test data. The following commands evaluate the three networks on the in-built CricketX data’s test set. !python test_linear.py --dataset_name CricketX --model_name InterSample !python test_linear.py --dataset_name CricketX --model_name IntraTemporal !python test_linear.py --dataset_name CricketX --model_name SelfTime Finally, supervised training of the whole framework based on the pre-trained individual networks can be performed using the following command. This training is governed by early stopping based on on-time evaluation performance. !python train_test_supervised.py --dataset_name CricketX --model_name SupCE A portion of the output: The results of the final supervised training are stored in a new directory called ‘results’. Model performance can be explored by visualizing the losses and other metrics over iterations. Performance of SelfTime SelfTime has been evaluated and compared with other well-acclaimed models using public time-series datasets such as CricketX, UWaveGestureLibraryAll (UGLA), DodgerLoopDay (DLD), and InsectWingbeatSound (IWS) from the UCR Time Series Archive, and the bearing datasets XJTU2 and MFPT. Transformed data samples are generated by augmenting the magnitude domain and the time domain such as jittering, scaling, cutout, magnitude warping, time warping, window warping and window slicing. These transformed samples are used as positive samples to extract inter-sample relationships. All the recent state-of-the-art models have undergone identical augmentations to justify comparisons. Different transformations performed on real data (blue in colour) to generate positive samples (red in colour). SelfTime massively outperforms recent state-of-the-arts in time-series domain namely, Triplet Loss model (2019), Deep InfoMax (2019), Forecast (2020), Transformation (2020), SimCLR (2020) and Relation (2020). Qualitative comparison of state-of-the-art models with SelfTime on UGLA classification dataset. Classes are differentiated using different colours. SelfTime classifies the classes clearly. Further reading Original research paperSource code repository","excerpt":"SelfTime is the state-of-the-art time series framework by finding inter-sample and intra-temporal relations","categories":["AI Trends"],"tags":["Guide","keras python","self supervised learning","simclr","Time Series","Time Series Analysis","Time series data","time series datasets","Time Series Forecasting","time series prediction"],"author_name":"Rajkumar Lakshmanamoorthy","publish_date":"2021-03-24T17:00:00","publication_year":"2021","word_count":895,"keywords":["TPU","simclr","computer vision","deep learning","Time Series","keras python","Time series data","R","CUDA","PyTorch","Guide","machine learning","AI","Time Series Forecasting","self supervised learning","Time Series Analysis","time series datasets","Python","Aim","time series prediction"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","Aim","PyTorch","TPU","CUDA","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/guide-to-selftime-self-supervised-time-series-representation-learning-framework-with-python-code\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10038998,"title":"Infosys Opens COVID-19 Care Centre For Employees &#038; Family Members","content":"Recently, Infosys announced that the company has set up COVID-19 care centres for its employees as well as their families. Currently, the centres are located in Pune and Bengaluru and are in process of setting up similar facilities across its major offices in India. According to reports, currently, India is registering a record number of COVID-19 cases daily that has put extreme pressure on the healthcare infrastructure of the country. In an email statement, the IT services company stated, “In order to support employees and their family members who have contracted COVID-19 and need special medical care beyond home quarantine, Infosys has set up Employee COVID Care Centres in Pune and Bengaluru.” The statement also mentioned that these centres would be managed by Ruby Hall hospital for Pune, and Manipal Hospitals for Bengaluru. The company is also in the process of setting up similar care centres across its major offices in India. At present, the IT giant is continuing to operate in a remote model across the offices and see no impact on the client deliverables due to the health situation. Sudha Murty, chairperson of the Infosys Foundation stated in an official blog post, “These are unprecedented times that require every section of society to rise up to the challenge.” The blog post also mentioned that at present, the company’s new responsibilities include helping educate schoolchildren, expanding hospital capacity, and assisting governments with informing the public and tracking the disease’s spread. Jeff Mosier, Senior Producer, Infosys Knowledge Institute said, “The need to react and the desire to help has inspired Infosys, along with most large organisations, to contribute money, skills, and resourcefulness to the fight against the coronavirus. These efforts go beyond helping 240,000 of our own employees stay safe through sanitizing workplaces, installing thermal scanners, and enabling most of them to deliver client work from home.”","excerpt":"Recently, Infosys announced that the company has set up COVID-19 care centres for its employees as well as their families. Currently, the centres are located in Pune and Bengaluru and are in process of setting up similar facilities across its major offices in India. According to reports, currently, India is registering a record number of […]","categories":["AI News"],"tags":["Infosys"],"author_name":"Ambika Choudhury","publish_date":"2021-04-26T18:55:48","publication_year":"2021","word_count":307,"keywords":["Go","Infosys","AI","programming_languages:R","programming_languages:Go","GAN","R"],"extracted_tech_keywords":["AI","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infosys-opens-covid-19-care-centre-for-employees-family-members\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043947,"title":"Julia Computing Raises $24 Mn in Series A","content":"Massachusetts-based Julia Computing has raised $24 million in Series A led by Dorilton Ventures, with participation from Menlo Ventures, General Catalyst, and HighSage Ventures. Bob Muglia, former Snowflake CEO and former Microsoft President of Servers and Tools, will join the Julia Computing’s Board of Directors. Julia Computing will use the fresh funds to further develop its secure, high-performing JuliaHub cloud platform and to grow the Julia ecosystem. “I am excited to share that Julia Computing has raised a $24 Million Series A. This takes us one step closer towards building #julialang into the language and ecosystem of our dreams,” said Viral B Shah, co-founder & CEO, Julia Computing. “We are excited to lead this important round and partner with Julia Computing. Julia Computing is at the very centre of technical computing, a substantial global market with significant barriers to entry. Julia’s machine learning and AI technologies make it possible to simulate rather than approximate, changing the economics of computational analysis and scientific discoveries. This is a truly transformative business with high potential for success,” said Daniel Freeman, who led the Dorilton Ventures investment. The four creators of the Julia programming language including Dr. Viral Shah, Prof. Alan Edelman, Dr. Jeff Bezanson, Stefan Karpinski along with Deepak Vinchhi and Keno Fischer founded the company in July 2015.","excerpt":"Former Snowflake CEO Bob Muglia will join the Julia Computing’s Board of Directors.","categories":["AI News"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-07-20T11:02:43","publication_year":"2021","word_count":216,"keywords":["machine learning","programming_languages:R","AI","Julia","R","Snowflake"],"extracted_tech_keywords":["AI","machine learning","Snowflake","R","Julia","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/julia-computing-raises-24-mn-in-series-a\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10057025,"title":"Council Post: Reflecting On The Cost Of Our Dream To Build The Coveted General AI","content":"It is a long-standing joke among the industry experts that while AI may crunch massive amounts of data, write codes that run huge machinery or even author a book, it would still fail tasks that a three-year-old human child can accomplish. This is also why AI systems still have a long path to trace to be truly called ‘intelligent’. Hubert Dreyfus, a well-known philosopher, was one of the staunchest critics of overestimating computer\/AI’s capabilities. He wrote three books – Alchemy and AI, What Computers Can’t Do, and Mind over Machine, where he critically assessed the progress of AI. One of his arguments was that humans learn from implied knowledge, and such capability cannot be incorporated into a machine. Having said that, there have been tremendous advancements in the human endeavour to move away from Narrow AI towards the Coveted General AI. Many new models like GPT 3, DALL.E, LaMDA, Switch Transformer, etc., are extremely powerful, swallowing billions and even trillions of parameters to multi-task. But, the fact is, we are still far from reaching the goal. The Quest For General AI The powerful language models and the newer zero-shot text-to-image generation models are all marching really fast towards the intended goal of performing tasks for which they were not trained. Each one is outwitting the previous one for its applications and uses. DeepMind, one of the most well known AI research institutes (owned by Alphabet), has centred its ultimate goal at achieving AGI. Interestingly, this year, the lab published a paper titled ‘Reward Is Enough‘, where the authors suggested that techniques like reward maximisation can help machines develop behaviour that exhibits abilities associated with intelligence. They further concluded that reward maximisation and reinforcement learning, in extension, can help achieve artificial general intelligence. Let’s take a closer look at GPT-3 created by DeepMind’s closest competitor, OpenAI, which created a major buzz in the scientific community. It was widely considered a massive breakthrough when achieving General AI. GPT-3 leverages NLP to imitate human conversations. It has been trained on one of the biggest datasets with 175 billion parameters, making it so powerful that it can complete a paragraph based on a few input words. Moreover, unlike typical narrow AI models, GPT-3 can perform tasks beyond generating human-like text, like translating between languages, reading comprehension tasks without additional input, and even writing codes! On the flipside, GPT-3 does not have ‘common sense’. It almost blindly learns from the material that it has been trained on, scrouged from several pages on the internet. Due to its lack of common sense, GPT-3 can pick biassed, racist and sexist ideas from the internet and rehash them [Note: Open AI is working to mitigate bias and toxicity in many ways, but it is not eliminated yet]. Additionally, the transformer lacks causal reasoning and is unable to generalise correctly beyond the training set, making it far from General AI. But, as we know, this quest is going to go on. This journey is fraught with billions of computational inputs, thousands of tonnes of energy power consumption and billions of dollars of expense to build and train these AGI models. Most ignore or fail to comprehend its consequences on the environment. AI’s Carbon Footprint Climate change has become a fundamental crisis of our time, and AI plays a dual role. It can help reduce the effects of the climate crisis and control it through solutions like smart grid designing or developing low emission infrastructure. But, on the other hand, it can lead to the undoing of all sustainability efforts, making the extent of its carbon emission hard to ignore. Estimates suggest that training a single AI generates close to 300 tonnes of carbon dioxide, equivalent to five times the lifetime emissions of an average car. AI requires increased computational powers, and data centres that store large amounts of AI data consume high energy levels. The high-powered GPUs required to train advanced AI systems need to run for days at a time, utilising tonnes of energy and generating carbon emissions. This is increasing the ethical price of running an AI model. The Price Of GPT-3 and its successors GPT-3 is one of the largest language models. The Neural Architecture Search process of training general transformer models requires more than 270,000 hours of training and 3000 times the energy, so much so that the training has to be split over dozens of chips and broken down over months. If the input is so massive, the output is worse. A 2019 study found that training an AI language-processing system generates anywhere between 1,400 to 78,000 pounds of emission. This is equivalent to 125 round trip flights between New York and Beijing. Sure, it is better in performance, but at what cost? Carbontracker suggested training GPT-3 just once requires the same amount of power used by 126 homes in Denmark every year. It is also the same as driving a car to the moon and back. GPT-3 isn’t the only large language model in the market today. Microsoft, Google, and Facebook are working on and have released papers on more complex models involving images and powerful searches that go far and beyond language, to create multi-tasking and multi-modal models. OpenAI identified how, since 2012, the amount of computing power in training large models has been increasing exponentially with a 3.4 month doubling time. If this is true, one can only imagine the energy consumption and carbon emissions until we reach AGI. Rethinking Our Approach Towards AGI AI could become one of the most significant contributors to climate change if this trend continues. This entails employing efficient techniques for data processing or search and training models on specialised hardware, like AI accelerators, that are more efficient per watt than general-purpose chips. Google published the paper on Switch Transformers that use more efficient sparse neural nets, facilitating the creation of larger models without increasing computational costs. Researchers such as Lasse Wolff Anthony, who has worked on AI power usage, have suggested that large companies train their models in greener countries such as Estonia or Sweden. Given the availability of greener energy supplies, a model’s carbon footprint can be reduced by more than 60 times. The solutions aren’t many as of now, but attempts are being made to devise them. It is important that we have conversations about it. While innovation is the basis on which a society moves forward, we must also be conscious of the cost such ‘innovation’ brings. The need of the hour is to strike a balance between the two. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"It is a long-standing joke among the industry experts that while AI may crunch massive amounts of data, write codes that run huge machinery or even author a book, it would still fail tasks that a three-year-old human child can accomplish. This is also why AI systems still have a long path to trace to […]","categories":["AI Features"],"tags":["Artificial General Intelligence","climate change","LLMs"],"author_name":"Padmashree Shagrithaya","publish_date":"2021-12-24T16:00:00","publication_year":"2021","word_count":1131,"keywords":["data science","TPU","OpenAI","AI","LLMs","Transformers","RAG","NLP","Aim","Artificial General Intelligence","analytics","climate change","R"],"extracted_tech_keywords":["AI","NLP","data science","analytics","OpenAI","Aim","Transformers","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-reflecting-on-the-cost-of-our-dream-to-build-the-coveted-general-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":49456,"title":"Cypher 2019: Experience From India’s Largest Analytics Conference","content":"This is a first-person account of an attendee from Cypher 2019 held from 18-20 September. Views are personal. It’s 11:30 AM on the first day of the three-day conference, and the Telegram group dedicated to CYPHER 2019 is abuzz with enthusiastic chatter. “Room 2 and 3 are too small for this kind of conference”, a participant said in the Telegram group created for this occasion. “Too small for the size of the audience”, chimes another. Room 2 and 3 have been dedicated by the organizers to “knowledge sessions” and “tech talks”, respectively. They have a capacity of 100 each, and often have attendees crowding around the doors. In contrast, Room 1 is by far the largest (capacity ~800) and is dedicated to “keynotes and panel discussions”. Here, people with white hair tell their audience “where the field is going”, as I overheard the first keynote speaker putting it. This is all par for the course for a field that has become harder to define with each passing year. What is one doing when one “does data science”? Is it software engineering? Is it math? Is it plain and simple automation of what business managers and marketers have anyway been doing? Or is it just hype? Complicating the picture even more, are the salesmen. There are two kinds of sales pitches taking place – out-of-the-box “AI” products that try to eliminate the data scientist, and education companies trying to create evermore data scientists\/engineers\/professionals (take your pick, it doesn’t matter) to fuel a whole generation worth of disillusionment. I speak to one of the salesmen. “Here’s the data,” he says, holding out his left hand like it’s carrying an invisible apple. “And here’s the insight,” in his right hand. “In the middle,” and by this point he has run out of hands, “there’s a bunch of processing and that’s where our product comes in.” So his product starts right after data and stops right before insight. Cool… Make no mistake – this product is part of the stack at the 40-member data science team at one of India’s largest conglomerates, who are also sponsoring and speaking at the event. The delegation from this conglomerate is led by a genial and wise-looking man who speaks (in Room 1 of course) about ABCDE of data initiatives – viz. creating Awareness, understanding Business, solving Creatively, using Data and achieving Excellence. And if that framework is the one thing you took away from the 3 days of conferencing, more power to you, because in the end, this is as good an attempt to describe the big fat elephant that is the field of data science circa 2019 as any other. The talks themselves are peppered with references to “business users”. Here’s what I gathered were the key features of these abstract entities: Business users are not data scientistsThey are also not software engineersThey, in fact, are probably not engineers of any kindThey might be kids They are paying for all of this If there were any actual business users at the conference, they were to be found in Room 2 (knowledge sessions) and not Room 3 (tech talks). The curious thing about Room 3 was that it was markedly younger than the other two rooms, both in terms of the audience and the speakers. It also seemed to have a greater representation of women. Perhaps tech = young = new entrants in the industry = more women. Either way, it is safe to assume that the actual work of developing AI\/ML, both the software and the math, has been placed on the shoulders of the youth, which, now that I say it out loud, isn’t all that surprising, or a bad thing. Meanwhile, in Room 1, there’s a debate about whether AI is good or evil, touching upon the usual suspects – Cambridge Analytica, Black Mirror, Elon Musk, Facebook chatbots. It charges up the crowd and is a big hit. The ethical questions make people edgy and make them feel part of something big and important and unknown. In general, the whole Pandora’s box of explainability\/interpretability of statistical black-box models is referred to in abstract terms by several Room 1 speakers, and those are the moments when the boring age-old concept of “business context” takes a back seat as the big challenge in data science, and math itself knocks on the door. Later, I am chatting up a senior leader of a credit card company, who heads a team of some 80 data scientists, one of the largest in India. He is quick to point out that the team was operating at the cutting edge of tech – moving past random forests and implementing neural networks. He is also quick to point out that he was not really involved in the “day to day of it”. Meaning, he didn’t code and probably haven’t trained a model in a while. He is the big-picture guy, solving for the lack of talent, or for the incompetence of tech teams which makes thing harder to implement, or for the business users who understand neither tech nor stats. Let’s be in touch. I come back into Room 1, and listen to more about “new paradigms”, and “democratization of data”, and about running random forests built-in data lakes on a cloud infrastructure. It feels like a group of people connected only in their combined hope that their bet works out. The bet they have taken by calling themselves data (evangelist, scientist, engineer, analyst). The bet on an AI-led future, whatever such a future might hold. And in that bet, I feel one with them. This article is a part of the AIM Writers Programme. If you wish to write for us, email us at info@analyticsindiamag.com","excerpt":"This is a first-person account of an attendee from Cypher 2019 held from 18-20 September. Views are personal. It’s 11:30 AM on the first day of the three-day conference, and the Telegram group dedicated to CYPHER 2019 is abuzz with enthusiastic chatter. “Room 2 and 3 are too small for this kind of conference”, a […]","categories":["Deep Tech"],"tags":["AIM Writers Programme"],"author_name":"Saksham Agrawal","publish_date":"2019-11-05T15:21:33","publication_year":"2019","word_count":954,"keywords":["data science","Go","AI","neural network","chatbots","ML","AIM Writers Programme","Aim","analytics","R","data lake"],"extracted_tech_keywords":["AI","ML","neural network","data science","analytics","Aim","chatbots","R","Go","data lake"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/cypher-2019-experience-from-indias-largest-analytics-conference\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039959,"title":"In Conversation With Vineeth N Balasubramanian, Head Of AI, IIT Hyderabad","content":"“Science and Technology for a Sustainable Future” is the theme of National Technology Day 2021. The day serves as a reminder of India’s commitment to technological progress. Soon after the independence, the government set up IITs to cultivate a new crop of technology leaders and push our technology landscape. Analytics India Magazine caught up with Vineeth N Balasubramanian, Head, Department of Artificial Intelligence, Indian Institute of Technology, Hyderabad. He is also an Associate Professor at the Department of Computer Science and Engineering. “Choosing initial steps that build on one’s strengths can help provide the right impetus in early stages,” said Vineeth. Excerpts: AIM: Tell us about your Google Research Scholar Award 2021 Vineeth N Balasubramanian: The title of the research project is “Bridging Perspectives of Explainability and Adversarial Robustness”. Robustness to adversarial perturbations in input and explainability of deep learning models have emerged as two important dimensions of estimating the goodness of a trained deep neural network (DNN) model, especially in risk-sensitive and safety-critical application domains such as healthcare, aerospace or autonomous navigation. Most work so far on these two topics – explainability and adversarial robustness – have focused on either of them alone, with very few efforts looking at connections between the two. A human-imperceptible adversarial perturbation of an image of a cat which is predicted as an ostrich by a DNN model may have no meaningful explanation for such a prediction. This simple intuition tells us that there may be deeper complementary connections to explore these two seemingly disparate directions. Exploring this connection provides a new dimension to both these topics, which could unearth a fresh new understanding of DNN models. In this project, we seek to answer some of these questions to develop a foundation for further research at this intersection. AIM: How did your fascination with machine learning and deep learning begin? Vineeth N Balasubramanian: More than a specific event or experience, my journey has been a gradual progression of experiences that strengthened my fascination with machine learning and deep learning. My first academic degree was in mathematics, which continues to interest me to this day. My Master’s thesis in image processing was perhaps my first foray into the broader field that enthused me into this research area. My PhD thesis was in an interdisciplinary group that applied machine learning to real-world problems, which showed me the applied side of research. Through all these years, my natural ideation process always gravitated towards machine learning algorithms, which was perhaps an ideal middle ground between math on the one hand and real-world application on the other. This aspect of machine learning and deep learning continues to interest me, and I still try to work on solutions for real-world challenges. I do believe there is a long way to go. AIM: What were your initial challenges, and how did you address them? Vineeth N Balasubramanian: Different challenges are encountered at different stages of career. As a research student, my initial challenges were similar to many others – understanding what kind of a research problem I want to work on; knowing when to read and when to experiment; what it takes to publish your research; how to write a paper; and so on. To a large extent, these questions were answered through the mentorship I received from different people, discussing with peers, and sometimes just trial-and-error, or what one can call experience. AIM: Did you encounter any data science problems during your research? Vineeth N Balasubramanian: I have had to deal with two biggest problems: dataset creation; and finding the problem one can solve given a certain kind of dataset. Creating a dataset is a skill by itself and often a longitudinal effort with multiple players, especially when it pertains to a domain that requires expertise, such as healthcare or aerospace or agriculture. Knowing what kind of data collection is required to solve a problem is non-trivial. Discussions with multiple stakeholders and thinking through every detail of data collection and annotation is paramount. Ethical and privacy implications must also be kept in mind, and suitable consent and permissions must be taken. Most importantly, organisations with data often make assumptions on what they can do with it, which does not translate easily into practice. This requires careful understanding of the data and its annotations and discussions with the people who collected the data and domain experts. A thorough understanding is required before problem formulation and subsequent development of machine learning models. AIM: What are your thoughts on the scope of AI research in India compared to the global scenario? Vineeth N Balasubramanian: The scope of AI in India across different application domains is immense. Plenty of opportunities to learn AI\/ML – especially certificate programs – have emerged over the last couple of years. Governments, both at central and state levels, have shown great leadership and enthusiasm in adopting AI, and even AI-based startups have emerged in diverse application domains. However, considering our population, there is always scope for more. One of the issues we lag is – innovation in AI. However, AI as a field is still growing, and making new technological innovations at the heart of AI itself – be it fundamental or applied – is something we need more of. There are two aspects to the research scope of AI in India: to use and leverage AI to develop cutting-edge products and services in various application sectors, or to make fundamental advancements to the field of AI itself, be it theoretical or applied. The former is important not merely as an indicator of technological advancement but more fundamentally for improving the quality of life of every citizen in the country and helping scale the efforts of enablers to reach out to every corner of the country. But for our position in the global arena, the latter may be critical – to think beyond what exists and create new knowledge and technologies. We do lag behind in the latter aspect and need to invest concerted efforts to develop on both fronts. Being a global leader in a field such as AI automatically attracts economic investments and growth, and thereby a better quality of life for all individuals. In terms of potential, we have it in abundance but need the will and right platforms to convert potential to tangible outcomes. AIM: What’s IITs’ approach towards AI education and research? Vineeth N Balasubramanian: At IIT-Hyderabad, we have an AI department that offers Bachelors, Masters and PhD programs in AI. We also offer a certificate program for anyone interested in AI, even if they are not registered as an IIT-H student. The department consists of faculty members in AI with different backgrounds: computer science, electrical engineering, mathematics, mechanical engineering, design, liberal arts, physics, etc. This allows all our programs to be holistic. Historically, while AI was ensconced in computer science, it has developed over the years to encompass more perspectives, including ethical and design aspects. Hence, it is imperative to educate our next generation of professionals in a well-rounded manner, providing exposure to all these perspectives. AIM: How can the government and corporations play a role in encouraging more students to enter AI fields? Vineeth N Balasubramanian: There is always a need to do more. India has talented students across almost every institution; however, there is a dearth of mentorship opportunities to take their learning to the next level. Many faculty, though sincere and dedicated, have not had the access to state-of-the-art topics and practices. Remote mentorship programs for bright students, especially in lower-tier cities and towns, can be a catalyst. Attracting Indian talent around the globe to participate in such programs or spend a few months at Indian academic institutions or startups could be another direction. Opportunities for motivated faculty in lower-tier institutions to spend a few months at the best of institutions, in India or abroad, could be a third direction. In all such efforts, it may be important to develop win-win programs that are simple to operate and rewarding for all parties involved. AIM: What is your advice for people who want to pursue AI research? Vineeth N Balasubramanian: My suggestion to all aspiring AI researchers would be to keep at it. There is nothing like having a mentor, but even if not, there are many online resources and platforms such as discord groups, virtual meetups, ML hackathons, workshops to learn from peers and experts these days. Sometimes, students want to learn everything in a very short duration, which can cause confusion. It does take time to get deep into a field, and focusing on a sub-area and continue working on it for a reasonable period of time may be useful. It is also important to have clarity of purpose and a good understanding of one’s strengths. Knowing one’s goals – becoming a researcher, applying ML to industry problems, promoting social good through ML, or pedagogy in ML – helps progress in the right direction. Choosing initial steps that build on one’s strengths can help provide the right impetus in the early stages. AIM: Tell us about your projects Vineeth N Balasubramanian: We are currently working on explainable AI and its connections to introducing reasoning and commonsense in AI\/ML model predictions. Beyond these, mentoring and training quality researchers is by itself an important contribution of academic institutions. Considering we are one of the second-generation IITs, having our students succeed at the highest level, including publications at top-tier venues and laudable recognitions, has been a heartening impact from a skilling perspective. I hope that our research group and institution would be able to continue to contribute many more such researchers to the AI ecosystem in the years to come.","excerpt":"“Science and Technology for a Sustainable Future” is the theme of National Technology Day 2021. The day serves as a reminder of India’s commitment to technological progress. Soon after the independence, the government set up IITs to cultivate a new crop of technology leaders and push our technology landscape. Analytics India Magazine caught up with […]","categories":["AI Features"],"tags":["AI at IIT","Interviews and Discussions","research in AI","research in India"],"author_name":"kumar Gandharv","publish_date":"2021-05-11T18:00:00","publication_year":"2021","word_count":1608,"keywords":["data science","artificial intelligence","research in India","AI at IIT","AI","machine learning","neural network","research in AI","ML","Ray","Aim","deep learning","analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","data science","analytics","Aim","Ray"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-with-vineeth-balasubramanian-head-of-ai-iit-hyderabad\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":66977,"title":"How Misconfigured Containers May Create Cybersecurity Issues For Companies","content":"Containers give an easy approach to build, deploy and run applications by packaging individual dependencies like libraries, data files, and more into one package. Containers are very often deployed with the default security configurations, which do not provide adequate protection for enterprise security, according to experts. While containers have various benefits of portability, less system dependence and increased innovation, it can be very costly for companies if not done properly. So if a bad actor were to have control of a container, it could damage your entire container stack. At the same time, many companies don’t utilise identity and access management policies which can effectively secure containerised applications from hackers. Similarly, it’s important to sign the container images. An improperly configured container can be a source of a major security event. In fact, security researchers have discovered thousands of misconfigured containers belonging to companies located in various geographies like the US and China, which could become hacking targets and provide malicious players access to sensitive data. Misconfigured Containers Can Give Hackers Access To Corporate Networks Misconfiguration practices can be caused due to human errors while deploying containers. A very few common container misconfigurations, according to experts, are utilising default container names and leaving default service ports exposed to the public. Misconfigured containers can pose a significant security risk to companies using the cloud, and container security is such a critical issue that 94 % of security pros are worried about its security, and vulnerabilities in production, according to a report. One example of attack as a result of a misconfigured container took place when hackers exploited a misconfigured Docker API port to execute an Ubuntu container with the kinsing malware, which then runs a crypto miner and spreads the malware to other containers and hosts. The attack was discovered by security firm Aqua Security, the attack stood out as the example of the expanding threat to cloud-native environments. “Our analysis of this attack vector exposes the techniques used, starting with exploiting the open port, through evasion tactics and lateral movement, all the way up to the end-goal of deploying the crypto miner.” According to Aqua Security, the firm had been witnessing a growth in the volume of attacks which target container environments. The firm has been tracking an organised attack campaign which targets misconfigured open Docker Daemon API ports. Since attackers have been launching on newer strategies for penetrating into containers and gaining control of the entire cluster, it is very critical that container security is prioritised and made sure that containers are not released with any misconfiguration. Contrary to traditional applications, containerised applications need security to be built into the entire development and delivery process. Experts also say that using network policies and firewalls are important here so that resources are not exposed to the web, and therefore needs stringent cloud infrastructure policies including container management. Here’s What Can Be Done To Avoid Misconfiguration It’s critical to scan your containers and images, including base images. For continuous vulnerability management, teams need to ensure that the solution you are leveraging can use both signatures and behaviour based technologies. To avoid misconfiguration, there is a need for proper container management policies, including the implementation of effective security techniques at each step of the CI\/CD workflows. This would eliminate the room for errors. Similarly, automation can be embedded into the container orchestration to minimise misconfiguration as a result of manual processes.","excerpt":"Containers give an easy approach to build, deploy and run applications by packaging individual dependencies like libraries, data files, and more into one package. Containers are very often deployed with the default security configurations, which do not provide adequate protection for enterprise security, according to experts. While containers have various benefits of portability, less system […]","categories":["AI Features"],"tags":["AI Companies","Cyber Security","International Affairs","network firewall"],"author_name":"Vishal Chawla","publish_date":"2020-06-10T13:00:00","publication_year":"2020","word_count":569,"keywords":["network firewall","Go","API","Cyber Security","AI","innovation","CI\/CD","docker","RAG","automation","AI Companies","GAN","International Affairs","R"],"extracted_tech_keywords":["AI","RAG","docker","R","Go","CI\/CD","API","GAN","automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-misconfigured-containers-may-create-cybersecurity-issues-for-companies\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10040540,"title":"All The AI Announcements From Google I\/O Keynote 2021","content":"Alphabet CEO Pichai said Google has been working with various non-profit organisations to address the COVID crisis at the ongoing annual developer conference Google I\/O 2021. “I\/O has always been a celebration of technology and its ability to improve lives and to remain optimistic that technology can help us address the challenges we face together,” said Pichai. Pichai made many important announcements, including major releases and updates, in his keynote speech. Analytics India Magazine rounds up the key highlights and announcements related to AI\/ML from the conference: The next-gen language model Pichai discussed how natural language processing and transformer models have made several breakthroughs in the past few years and how Google has been working tirelessly to make enhancements to organise and access heaps of information conveyed by the written and spoken word. The tech giant unveiled LaMDA, a Language Model for Dialogue Applications. The model has been built on Transformer — a neural network architecture open-sourced by Google Research in 2017. “From concept all the way to design, we are making sure it is developed consistent with our AI principles. We believe LaMDA’s natural conversation capabilities have the potential to make information and computing radically more accessible and easier to use,” Pichai added. AI in Google Maps Google introduced two new features on Google Maps: Eco-friendly routes to provide the most fuel-efficient routes to the users; Safer Routing to get upfront information on weather and traffic conditions AI-powered dermatology tool Google AI has been working on detecting breast cancer, DNA sequencing, diabetic retinopathy, etc in healthcare. With the help of AI and machine learning, the tech giant has previewed an AI-powered dermatology tool. Using the same techniques that detect diabetic eye disease or lung cancer in CT scans, this tool helps to identify dermatologic issues, such as a rash on the arm, just using the phone’s camera. Google Cloud’s new AI platform Pichai introduced a new AI platform, Vertex AI. Vertex AI is Google Cloud’s new unified ML platform that brings AutoML and AI Platform together into a unified API, client library, and user interface. With the help of this platform, one can build, deploy, and scale ML models faster, with pre-trained and custom tooling within a unified AI platform. Vertex AI’s custom model tooling supports advanced ML coding, with nearly 80% fewer lines of code. The “magic mirror” using AI Project Starline combines advances in hardware and software to enable friends, families and coworkers to connect, even when they’re cities (or countries) apart. The tech project combines computer vision, machine learning, spatial audio and real-time compression. Google has developed a breakthrough light field display system that creates a sense of volume and depth that can be experienced without the need for additional glasses or headsets. The next-gen TPU Pichai unveiled the next generation of Tensor Processing Units (TPUs), TPUv4. TPUv4 is powered by the v4 chip, which is more than twice as fast as TPUv3. A single v4 pod contains 4096 v4 chips and each pod has 10x interconnect bandwidth per chip at scale. The v4 includes 1 exaflop of computing power, and is the fastest system ever deployed at Google. Quantum AI campus The search giant has unveiled its new Quantum AI campus in Santa Barbara, California. This campus houses the first quantum data centre, quantum hardware research laboratories, and in-house quantum processor chip fabrication facilities. A new milestone in AI algorithms The tech giant introduced a new AI algorithm, known as the Multitask Unified Model, or MUM.  Like BERT, MUM is built on a Transformer architecture, but it’s 1,000 times more powerful. It is trained across 75 different languages and many different tasks at once. MUM is multimodal and understands information across text and images and, in the future, can expand to video and audio. New innovations in Google Photos The tech giant unveiled a new feature in Google Photos called Little Patterns. Little Patterns uses machine learning to translate photos into numbers and then compare how visually and conceptually similar those images are. The next evolution of collaboration After launching Google Docs and Sheets 15 years ago, Smart canvas is the next big step to create an evolving hybrid work model that gives new urgency to existing collaboration challenges. The tech giant also introduced new smart chips in Docs for recommended files and meetings. To insert smart chips into work, one must simply type “@” to see a list of recommended people, files, and meetings. Smart chips will come to Sheets in the coming months.","excerpt":"Alphabet CEO Pichai said Google has been working with various non-profit organisations to address the COVID crisis at the ongoing annual developer conference Google I\/O 2021.  “I\/O has always been a celebration of technology and its ability to improve lives and to remain optimistic that technology can help us address the challenges we face together,” […]","categories":["Global Tech"],"tags":[],"author_name":"Ambika Choudhury","publish_date":"2021-05-20T18:00:00","publication_year":"2021","word_count":750,"keywords":["Go","API","machine learning","TPU","AI","neural network","ML","computer vision","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","computer vision","analytics","TPU","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/all-the-ai-announcements-from-google-i-o-keynote-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":66554,"title":"OurCrowd Launches $100 Million Pandemic Innovation Fund","content":"OurCrowd, one of the world’s largest crowdfunded-venture investment platform, today announced the launch of its Pandemic Innovation Fund. The Fund plans to raise $100 million for investment in urgent technological solutions for the medical, business, educational and social needs triggered by global pandemics and other health emergencies. “The rapid spread of the coronavirus has validated our vision of a connected digital world poised to solve any crisis through global communication and rapid response,” said OurCrowd CEO Jon Medved. “To ensure that we get the world back on track, there is now an urgent need for innovation. Technology can help us overcome many of the problems resulting from the crisis. It’s time for tech to move fast and fix things.” OurCrowd’s Pandemic Innovation Fund will focus on the following investment sectors: Prevention & Containment – Vaccines, Testing, Personal protection, etc.Treatment & Healing – Therapeutics, Diagnostics, Remote monitoring, Digital health, etc.Continuity & Disruption Mitigation – Remote working, Distance learning, Robotic process automation, Home exercise, Cybersecurity, etc. The Fund will both invest in new startups and select relevant companies already included in OurCrowd’s existing portfolio. OurCrowd’s portfolio already has more than 20 companies actively working to mitigate the coronavirus crisis and its effects, some of which will be candidates for follow-on investment from the Pandemic Investment Fund. OurCrowd’s existing investments in technologies on the frontlines of coronavirus response include: MigVax – Developing the MigVax-101 COVID-19 oral subunit vaccine for humans based on a proven platform developed over 4 years that was shown to be a highly effective oral vaccine against IBV (Infectious Bronchitis Virus) in poultry. Sight Diagnostics – Compact complete blood count analyzer that provides lab-grade results with 2 drops of finger prick blood sample in under 10 minutes. SaNOtize – Approved by Health Canada for multi-center Phase II trial of its Nitric Oxide Releasing Solution (NORSTM) for the prevention and early treatment of COVID-19 TytoCare – Remote physical exams and monitoring for primary care, chronic care and COVID-19 patients, protecting medical staff and reducing the burden on health systems MeMed – Provides diagnostic insights to distinguish between bacterial and viral infections and is working on actionable solutions to help enable early intervention, before the onset of COVID-19 symptoms, and the implementation of measures to identify infection severity and improve patient management. Techsee – Remote Visual Assistance powered by Computer Vision AI & AR used by leading brands like Vodafone and Verizon to provide uninterrupted tech support while ensuring the safety of employees and customers under social distancing. Kryon – Robotic process automation used to transfer and verify millions of coronavirus test results with individual patient health records in a fraction of the normal timeIntuition Robotics – ElliQ AI-powered intelligent digital companion for the elderly, helping them stay sharp, connected and engaged Zebra Medical Vision – Zebra-Med’s AI automatically detects and quantifies suspected COVID-19 findings on standard chest CTs, both contrast and non-contrast, and is already integrated in Apollo Hospitals Group in India. The fund is open to both accredited private investors (minimum $50,000) and institutional investors (minimum $1,000,000) and further information about this opportunity is available at www.ourcrowd.com. Advisors to the Fund, portfolio company experts, and top speakers from around the world will participate in the OurCrowd Pandemic Innovation Conference – an online event on June 22, 2020, broadcast from Jerusalem.","excerpt":"OurCrowd, one of the world’s largest crowdfunded-venture investment platform, today announced the launch of its Pandemic Innovation Fund. The Fund plans to raise $100 million for investment in urgent technological solutions for the medical, business, educational and social needs triggered by global pandemics and other health emergencies. “The rapid spread of the coronavirus has validated […]","categories":["AI News"],"tags":[],"author_name":"Vishal Chawla","publish_date":"2020-06-02T17:46:26","publication_year":"2020","word_count":550,"keywords":["API","programming_languages:R","AI","innovation","Git","computer vision","automation","disruption","R","startup"],"extracted_tech_keywords":["AI","computer vision","R","Git","API","automation","innovation","disruption","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ourcrowd-launches-100-million-pandemic-innovation-fund\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167751,"title":"Google Has Had Two Years to ‘Kill’ Perplexity — and it Hasn’t: Aravind Srinivas","content":"AI startup valuations and revenues are going through the roof – yet there’s a persistent concern that tech giants wielding frontier models will eventually muscle into their territory. Even as founders remain optimistic about their companies’ futures, spectators on the internet have often ‘killed’ startups through their self-proclaimed verdicts. Few understand this tension better than Aravind Srinivas, CEO of Perplexity, which was recently valued at $9 billion. During a recent Reddit AMA, Srinivas had to address a flurry of questions about whether Google might someday render Perplexity obsolete. Recently, Google released the new Gemini 2.5 family of models, which surpassed the competition on benchmarks. Additionally, the company is also planning to announce an AI-heavy feature for its search engine, called the AI Mode, while the AI Overviews feature is being rolled out to more and more users. When Srinivas was asked if Google will double down on AI-enabled search features that may render Perplexity “obsolete”, he said, “They [Google] have had two years to kill Perplexity and haven’t.” Srinivas added that Google is reluctant to fully embrace AI-generated search responses because they could undermine its advertising business model. Unlike traditional search results, where users click through links (generating ad revenue), AI responses provide information directly within the search interface, eliminating the need to visit external websites. “Think about it: if AI gave you direct answers to the score in a basketball game, how can you sell Ticketmaster ads?” added Srinivas. For context, Alphabet earned $72.46 billion in Google Advertising for the quarter that ended December 31 last year. ‘Google Will Have to Eat Its Revenue to Replicate Perplexity’ AIM reached out to Deedy Das, principal at Menlo Ventures, to understand if and how Perplexity will continue to have a place of its own despite Google’s strong presence. Like Srinivas, Das also emphasised that Google derives a substantial portion of its revenue from advertising, rather than from its AI models. So, would Google be inclined to experiment with something that generates profit for them? “Google will likely not build the exact same product as Perplexity because it cannibalises its own search business, which pays the bills,” he said. Perplexity makes most of its revenue through paid subscription services for individuals and enterprises. However, in November last year, the company announced it would experiment with advertising to generate revenue to share with its publisher partners. Besides, there’s also a strong investor confidence in Perplexity. The company is reportedly in talks to raise funds between $500 million and $1 billion to double its valuation to $18 billion. Srinivas also recently announced that the startup has crossed $100 million in revenue. For now, Perplexity is going strong, but assuming Google starts building products of the exact nature, will it hurt the flow of capital to Srinivas and Co.?“Incumbents have the capacity to replicate anything. They have more money, talent, and resources. By that logic, nothing should get funded,” said Das. “VCs fund Perplexity because Google will have to eat into its own revenue to replicate it fully, and they hope that users will prefer the product experience and move to it over time,” he added, indicating that Perplexity’s healthy user base works in its favour. However, that doesn’t mean Perplexity is under no threat. Google operates at a massive scale of over 2 billion monthly users. Perplexity’s opportunity lies in capitalising on the strategic dilemma Google faces, balancing innovation in AI with the risk of undermining its core advertising-driven business model. “Gemini 2.5 Pro does not pose a threat to Perplexity, but Google owns distribution for its search engine on nearly all platforms, which is a threat to Perplexity,” said Das, pointing out that the company will have to continue to build high-quality products to sustain. Comet to the Rescue? Perplexity has ventured into various areas of AI, making it feel strange to categorise it solely as a search engine. With in-house AI models like Sonar, alongside a deep research tool, image generation, shopping, and financial analysis tools, there is a wealth of offerings under their umbrella. Its next big bet is an agentic browser called Comet. In the AMA, Srinivas described Comet as the “ultimate frontend for using AI for daily browsing”. “[Comet is] a sidecar that lets you have an AI along with you on any webpage you are on. You can use it for asking questions about the content, extracting\/formatting the content to use for a task, or run a research job on that page\/domain,” he added. Interestingly, he claims it will be free from advertisements. He also said that Comet would answer questions based on the information available from the open tabs. “This way, Perplexity is no longer just a web search tool. It will search over everything,” he added. Srinivas also revealed that the original plan was to launch it mid-April, but the company is running “a couple of weeks behind timeline”. He said the company is working on expanding access to the waitlist in the next two weeks and making it available to all shortly after. “Perplexity needs a browser like Comet to own its distribution and hopefully provide useful features for its users,” said Das. Even with Comet, distribution is going to be a mammoth task for Perplexity, which Srinivas does agree with. “Google owns Android, Apple has Safari, and Microsoft has distribution lock-in deals with Windows OEMs. But anything worth doing is hard. We will persist and find ways around,” said Srinivas.","excerpt":"Will they ever?","categories":["Global Tech"],"tags":["Perplexity AI"],"author_name":"Supreeth Koundinya","publish_date":"2025-04-11T12:05:59","publication_year":"2025","word_count":906,"keywords":["Replicate","Go","API","AI","RPA","innovation","Perplexity AI","Aim","R","Gemini 2.5","startup"],"extracted_tech_keywords":["AI","Gemini 2.5","Aim","R","Go","API","RPA","innovation","startup","Replicate"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-has-had-two-years-to-kill-perplexity-and-it-hasnt-says-ceo-aravind-srinivas\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10102116,"title":"Reimagine Existing Data &amp; Tech in CXM","content":"In an era where companies are amassing vast troves of customer data, both online and offline, the challenge lies in effectively harnessing its potential. As projected by International Data Corporation (IDC), the global data landscape is set to undergo a monumental transformation by 2025, witnessing a tenfold surge from 2012, soaring from 6.5 zettabytes to an astounding 64.5 zettabytes. This vast amount of information represents a considerable repository of untapped value. Consequently, businesses are exploring novel avenues to engage with consumers, generating first- and zero-party data—information characterised by precision and resilience against data erosion. Amidst this vast landscape of data, encompassing various formats like images, videos, and audio recordings, lies a potential reservoir. Effectively re-imagining this reservoir of data and utilising it to scope necessitates strategic vision, technological infrastructure, and proficient skills for data processing. Further, we will delve into three specific applications that caption the effective re-imagination of data. These applications showcase innovative and transformative approaches that have the potential to significantly enhance customer experience through personalization. The scope of Retail Media Networks The advent of Retail Media Networks (RMNs) introduces promising avenues for both retailers and brands in this pursuit. RMNs offer opportunities to unearth fresh streams of revenue and insights, aligning with the goal of enhancing CX and driving business success. The scope of RMNs encompasses not only advertising, but also the strategic monetization of first-party data, valuable insights, and innovative out-of-home or in-store strategies that were previously confined to retailer merchant organisations only. Expected to reach $160 billion in global annual revenue by 2027, (according to Group M’s 2022 TYNY forecast) Retail Media itself is an enticing proposition for businesses the world over. So, through RMNs, retailers and suppliers can gain access to real-time insights that translate to consumer preferences. These insights empower them to deepen their connection with customers, shape their strategies, and close the loop from advertising impressions to sales. Another intriguing aspect is that brands are not limited to advertising solely through retailers that sell their products. RMNs have gained significant traction among non-endemic brands, such as fashion brands partnering with beauty retailers or travel agencies collaborating with theme park entities. This transformative shift in the RMN landscape involves the separation of media and data functionalities, providing brands with unprecedented access to essential first-party data insights, free from the constraints of media transactions. According to Merkle’s latest Retail Media Research Report, a remarkable 65% of surveyed retailers attest that offering insights independent of media transactions strengthens their vendor relationships, marking a 13% increase from 2021. Concurrently, brands are eagerly seeking access to profound insights within a privacy-conscious and contextually relevant environment that fosters customer engagement and loyalty. Elevating brand loyalty through zero-party data Brand loyalty is often elusive to achieve in today’s competitive landscape. Under these circumstances, the utilisation of zero-party data, willingly shared by customers, is proving to be a game-changer. However, this should be viewed with a lens that helps brands create profound emotional connections that nurture enduring loyalty. Recognizing that emotional bonds are the driving force behind loyalty, this evolution emphasizes shared values, transparency, trust, and open dialogue, all of which are fundamental to delivering a superior CX. Zero-party data transcends the limitations of demographics and behaviours. Consider the case of a tech-savvy 28-year-old professional from San Francisco who displays a keen interest in your offerings. While such insights offer a glimpse into preferences and aspirations, they represent just a fraction of the whole picture. Continuous engagement through surveys and social media, and website interactions allows customers to express their sentiments. This enables brands to adapt and personalize their offerings. When customers provide zero-party data, they are entrusting brands and retailers with personal information in exchange for some form of value. Therefore, the most creative and targeted uses of zero-party data are those that provide value based on the specific information customers willingly provide. Loyalty programs emerge as a natural and effective avenue for brands and retailers to achieve this. Consider an athletic apparel retailer that conducts surveys to understand what sport customers play or the types of fitness activities they engage in. Armed with this knowledge, the retailer can do more than simply send promotions related to the appropriate attire. They can deliver sport-specific content and even collaborate with local event organisers to inform customers about opportunities to get involved in their interests, all while wearing their favourite fitness apparel. In essence, zero-party data serves as the foundation for brands and retailers to not only understand their customers on a deeper level but also to offer highly personalized and valuable experiences in an increasingly privacy-conscious milieu. The seamless integration of zero-party data into a comprehensive data strategy can therefore greatly magnify CX enhancement, enriching the customer journey with each data facet. Furthermore, inventive ways of utilizing existing data and insights can complement this approach. Identity resolution: A synergy of data and insights Identity resolution platforms use existing data and harmonize it with second and third-party data sources, to weave a cohesive, cookie-less private identity graph. This intricate web empowers businesses with cross-channel targeting precision, catalysing personalised marketing initiatives, and facilitates accurate measurement of outcomes. In a noteworthy case study with a major US media conglomerate, their effort to achieve seamless personalisation across diverse brands through a unified marketing solution proved extraordinarily successful. This effort merged data from 30 sources, resulting in the synthesis of over 2,000 consumer attributes. The use of a proprietary identity resolution platform enhanced consumer experiences and provided valuable insights for advertisers as well. Advertisers benefited from the improved accuracy in predicting age and gender, reducing reliance on third-party data by 100%. For consumers, it led to a high success rate (80%) in anticipating the next optimal action, especially in product recommendations. For this innovative solution, the Merkle team recently received the “Excellence in AI Strategy Consulting Award” at Cypher 2023. Conclusion: The future is personal As data burgeons exponentially, reimagining its potential is no longer an option, but a necessity. The vast volume, diversity, and complexity of today’s data offer an extraordinary opportunity and challenge. Enterprises that effectively harness this and apply innovative approaches stand to benefit from the customer engagement landscape. The future of CX is as such profoundly personalized and insights-driven. An intimate understanding of individual preferences will characterize every interaction. Customers are no longer data points but valued individuals whose experiences demand to be meticulously tailored. The key to unlocking this lies in strategic data utilisation, not just collecting more but extracting actionable insights. Therefore, fine tuning existing solutions like RMNs, and refocusing on zero-party data insights coupled with identity resolution platforms can produce outcomes that are future-ready and hyper-personal.","excerpt":"The future of CX is as such profoundly personalized and insights-driven. An intimate understanding of individual preferences will characterize every interaction. Customers are no longer data points but valued individuals whose experiences demand to be meticulously tailored.","categories":["AI Highlights"],"tags":["customer experience"],"author_name":"Navin Dhananjaya","publish_date":"2023-10-30T12:30:00","publication_year":"2023","word_count":1108,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","customer experience","ViT","Rust","GAN","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","ML","R","Go","Rust","GAN","ViT","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/reimagine-existing-data-tech-in-cxm\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":44602,"title":"In A Major Security Push, Tech Giants Go Big On Bug Bounty Programs","content":"Over the past few years, the number of cybersecurity attacks have increased significantly. Today, every single aspect of a business is under attack by hackers. As the world continues to become more digital with new mobile apps and other platforms, the risk of getting pwned is also increasing. While companies are taking every possible step to secure all their platforms, hackers are also strengthening their attacking skills. And sometimes it gets difficult for a company’s in-house team to find and fix all the bugs.  This is where bug bounty comes into the scenario. What Is A Bug Bounty Program The idea of bug bounty started back in 1983, when there was an initiative for Versatile Real-Time Executive operating system — the deal was that anyone who would report a bug would receive a Volkswagen Beetle (a.k.a. Bug) in return. Today, Bug bounty programs have reached a whole new level, it has completely become a crowdsourcing initiative that reward experienced independent researchers upon identifying and reporting on bugs or vulnerabilities in technology and software programs. Currently, even some of the big companies have started their own bug bounty program — Facebook, Uber, Apple, Intel etc. The effectiveness of the bug bounty program to identifying vulnerabilities early on is significantly high. You not only able to discover some of the best talents from the cybersecurity domain but also manage to increase the chances of finding some serious flaws and get them fixed. Some Of The Big Investments On Bug Bounty Some of the big names that are going big on bug bounty programs are Intel, Snapchat, Facebook, Google, Apple etc. Officially launched in September 2014, Microsoft Bug Bounty is also one of the high paying bug bounties. The company rewards a minimum of $15,000 and a maximum of $300,000. However, it is only for critical and important vulnerabilities. According to Intel’s Bug Bounty Program page, the company rewards bug bounty hunters a bounty from $500 to $100,000 USD depending on the nature of the vulnerability and quality & content of the report. Source: https:\/\/www.intel.com\/content\/www\/us\/en\/security-center\/bug-bounty-program.html Snapchat is in the race to make the best out of bug bounty programs. The company rewards a bounty from $2000 to $15,000. However, the company is currently focusing on some specific areas. Source: https:\/\/hackerone.com\/snapchat Since November 2010, MountainView search giant Google has been running is Bug bounty program. And it is mostly focused on google.com, youtube.com, blogger.com. Talking about money, Google bug bounty rewards for qualifying bugs range from $100 to $31,337. Source: https:\/\/www.google.com\/about\/appsecurity\/reward-program\/ Facebook runs one of the best bug bounty programs. Even though the company is always on the radar when it comes to data security, the social media tech giant has taken necessary measures to keep its platform safe. Facebook’s bug bounty program rewards successful hackers a bounty of minimum $500 and the maximum depends on the severity of the flaw. Click here to know more about the program. Apple’s Million Dollar Move While all these companies are paying out hundreds and thousands of dollars, Apple has taken its bug bounty rewards to another level. In the recently concluded Black Hat security conference, Apple increased the maximum reward for its bug bounty program from $200,000 to $1 million, which is for severe deadly exploits. At present, this is the biggest bug bounty reward that has been offered by any major tech company for reporting vulnerabilities. Furthermore, segments that this bug bounty would cover are iOS, macOS, watchOS, tvOS, iPadOS, iCloud, and also, all the devices that run on these operating systems. That is not all, according to a source,  the tech giant starting from next years would be providing a pre-jailbroken iPhones to a selection of trusted security researchers as part of the iOS Security Research Device Program. When it is about security, Apple has always been serious. And as a proof, you don’t even have to dig deep — look at the Apple devices — unlike Android or any other device, Apple’s devices are more secure and reliable in the context of data protection and spying evasion than the Android or Windows operating systems. One of the reasons behind is that Apple has always made sure that the phones or any of its devices don’t get connected to other devices easily — it always takes a few extra steps. Talking about this move of increasing the reward for bug bounty, it clearly shows the company’s concern towards strengthening its cybersecurity game and also shows that the bug bounty ecosystem is on the rise. Outlook This trend has gone beyond the realm of tech giants with other industries too adopting this approach. After the Marriott hack that impacted as many as 500 million customers, Hyatt Hotels launched its bug bounty program with HackerOne in January 2019. The need for cybersecurity experts has increased and while many companies are still relying on the in-house cybersecurity team, companies have also realised that bug bounty is definitely one of the best ways to discover bugs and get them fixed. The recent investments are proof that the bug bounty ecosystem in India will grow even bigger.","excerpt":"Over the past few years, the number of cybersecurity attacks have increased significantly. Today, every single aspect of a business is under attack by hackers. As the world continues to become more digital with new mobile apps and other platforms, the risk of getting pwned is also increasing. While companies are taking every possible step […]","categories":["AI Features"],"tags":["Bug Bounty","Cyber Security"],"author_name":"Harshajit Sarmah","publish_date":"2019-08-18T14:10:19","publication_year":"2019","word_count":848,"keywords":["Go","Cyber Security","AWS","AI","ETL","cloud_platforms:AWS","ML","programming_languages:R","Bug Bounty","Git","Rust","R"],"extracted_tech_keywords":["AI","ML","AWS","R","Go","Rust","Git","ETL","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/in-a-major-security-push-tech-giants-go-big-on-bug-bounty-programs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046646,"title":"Collaborative Filtering Vs Content-Based Filtering for Recommender Systems","content":"The Internet is the new digital market, where it presents us with a number of choices, sometimes too overwhelming to choose from. Today everything we need or want to buy can easily be accessible to us through this new digital market. From content in entertainment to groceries and clothes, every basic necessity or luxury is available at the tip of our hands. With a plethora of options available, there always has been a serious need to filter, prioritize and efficiently deliver relevant information in order to overcome the potential problem of information overload, creating confusion to many Internet users. Recommender systems can help solve such a problem by processing through a large volume of dynamically generated information to provide the users with personalized content and services which in turn allows better communication and understanding between the users and the parent organization. What Is a Recommender System? Recommender systems are information filtering systems that help deal with the problem of information overload by filtering and segregating information and creating fragments out of large amounts of dynamically generated information according to user’s preferences, interests, or observed behavior about a particular item or items. A Recommender system has the ability to predict whether a particular user would prefer an item or not based on the user’s profile and its historical information. Recommendation systems have also proved to improve the decision making processes and quality. In large e-commerce settings, recommender systems enhance the revenues for marketing, for the fact that they are effective means of selling more products. In scientific libraries, recommender systems support and allow users to move beyond the generic catalogue searches. Therefore, the need to use efficient and accurate recommendation techniques within a system that provides relevant and dependable recommendations for users cannot be neglected. Conglomerates like Netflix use a recommendation engine to present their viewers with movie and show suggestions. Amazon, on the other hand, uses its recommendation engine to present customers with product recommendations. While each uses the one for slightly different purposes, both in general have the same goal: to drive sales, boost engagement and customer retention, and deliver more personalized customer experiences. Recommendations typically speed up the searches and make it easier for users to access the content they have always been interested in, and surprise them with several offers they would have never searched for. Doing companies are able to gain new customers by sending out customized emails with links to new offers that meet the recipients’ interests, or suggestions of films and TV shows that suit their particular profiles. Image Source Approaches to build Recommender Systems There are two main types of recommendation engines; namely collaborative filtering and content-based filtering. Collaborative filtering The Collaborative filtering method for recommender systems is a method that is solely based on the past interactions that have been recorded between users and items, in order to produce new recommendations. Collaborative Filtering tends to find what similar users would like and the recommendations to be provided and in order to classify the users into clusters of similar types and recommend each user according to the preference of its cluster. The main idea that governs the collaborative methods is that through past user-item interactions when processed through the system, it becomes sufficient to detect similar users or similar items to make predictions based on these estimated facts and insights. Such memory-based approaches directly work with the values of recorded interactions or data and are essentially core based on nearest neighbours search, i.e finding the closest users from a user of interest and suggest the most popular items among these neighbours. The created model approaches assuming there is an underlying “generative” insight that explains the user-item interactions and tries to discover it in order to make new predictions. It recommends an item to user A based on the interests of a similar user B. Furthermore, the embeddings can be learned automatically, without relying on hand-engineering of features. The collaborative filtering method does not need the features of the items to be given. Every user and item is described by a feature vector or embedding. The standard method used by Collaborative Filtering is known as the Nearest Neighborhood algorithm. There are several types of filtering such as user-based and Item-based Collaborative Filtering. Considering an example of User-based Collaborative Filtering, If we have an n × m matrix of ratings, with user u, i = 1, …n, and item p, j=1, …m. and we want to predict the rating r if the target user i did not watch\/rate an item j. The process is to calculate the similarities between target user i and all other users will be to select the top X similar users and take the weighted average of ratings from these X users with similarities as weights. Challenges with Collaborative Filtering The only issue with this method is that the prediction of the model for a given user, item pair is the dot product of the corresponding embeddings. So, if an item is not seen during training, the system cannot generally create an embedding for it and hence cannot query the model with this item. This issue is known as the cold-start problem. Content-Based Filtering The content-based approach uses additional information about users and\/or items. This filtering method uses item features to recommend other items similar to what the user likes and also based on their previous actions or explicit feedback. If we consider the example for a movies recommender system, the additional information can be, the age, the sex, the job or any other personal information for users as well as the category, the main actors, the duration or other characteristics for the movies i.e the items. The main idea of content-based methods is to try to build a model, based on the available “features”, that explain the observed user-item interactions. Still considering users and movies, we can also create the model in such a way that it could provide us with an insight into why so is happening. Such a model helps us in making new predictions for a user pretty easily, with just a look at the profile of this user and based on its information, to determine relevant movies to suggest. We can make use of a Utility Matrix for Content-Based Methods. A Utility Matrix can help signify the user’s preference for certain items. With the data gathered from the user, we can find a relation between the items which are liked by the user as well as those which are disliked, for this purpose the utility matrix can be put to best use. We assign a particular value to each user-item pair, this value is known as the degree of preference and a matrix of the user is drawn with the respective items to identify their preference relationship. Challenges faced with Content-based filtering Content-based methods seem to suffer far less from the cold start problem than collaborative approaches because new users or items can be described by their characteristics i.e the content and so relevant suggestions can be done for these new entities. Only new users or items with previously unseen features will logically suffer from this drawback, but once the system is trained enough, this has little to no chance to happen. Basically, it hypothesizes that if a user was interested in an item in the past, they will once again be interested in the same thing in the future. Similar items are usually grouped based on their features. User profiles are constructed using historical interactions or by explicitly asking users about their interests. There are other systems, not considered purely content-based, which utilize user personal and social data. An Example For Item Based Filtering Below is an example of Item Based Content Filtering where a movie recommendation system recommends movies based on user ratings and sorts recommendations according to it. This can be easily performed using pandas and the MovieLens Library. We are using the data and the item based files, which you can access and download using the link here. Creating a pivot table function on such a DataFrame will help us construct a user\/movie rating matrix. The Pandas’ corrwith can be used here, the function makes it really easy to compute the pairwise correlation of the vector of user rating with every other movie! Any results that have no data can be dropped, a new DataFrame of movies and their correlation score (similarity) to one Movie, in particular, can be constructed. To get the correct status of the recommendation, movies rated by fewer than 100 people should be avoided in such cases, The resultant product will be a Dataframe with a correlated similarity matrix that will define our results! Collaborative Filtering Vs Content-Based Filtering Here is a list of points that differentiate Collaborative Filtering and Content-Based Filtering from each other : The Content-based approach requires a good amount of information about items’ features, rather than using the user’s interactions and feedback. They can be movie attributes such as genre, year, director, actor etc. or textual content of articles that can be extracted by applying Natural Language Processing. Collaborative Filtering, on the other hand, doesn’t need anything else except the user’s historical preference on a set of items to recommend from, and because it is based on historical data, the core assumption made is that the users who have agreed in the past will also tend to agree in the future. Domain knowledge in the case of Collaborative Filtering is not necessary because the embeddings are automatically learned, but in the case of a Content-based approach, since the feature representation of the items is hand-engineered to an extent, this technique requires a lot of domain knowledge to be fed with. The collaborative filtering model can help users discover new interests and although the ML system might not know the user’s interest in a given item, the model might still recommend it because similar users are interested in that item. On the other hand, A Content-based model can only make recommendations based on the existing interests of the user and the model hence only has limited ability to expand on the users’ existing interests. A Content-Based filtering model does not need any data about other users, since the recommendations are specific to a particular user. This makes it easier to scale down the same to a large number of users. A similar cannot be said or done for Collaborative Filtering Methods. The collaborative algorithm uses only user behavior for recommending items while for Content-based filtering we have to know the content of both user and item. Conclusion In this article we understood how the Recommendation System works and the difference between the Collaborative Filtering vs Content-Based Filtering models and their working. Both methodologies have their own set of advantages, disadvantages and similarly particular use cases which we tried to explore and discuss. We also saw a small example for Item Based Content Filtering, you can find the whole implementation in a Colab Notebook using the link here. Happy Learning! References Collaborative Filtering vs Content-Based Filtering paper Basics of Filtering Methods","excerpt":"Recommender systems are information filtering systems that help deal with the problem of information overload by filtering and segregating information and creating fragments out of large amounts of dynamically generated information according to user’s preferences, interests, or observed behavior about a particular item or items. A Recommender system has the ability to predict whether a particular user would prefer an item or not based on the user’s profile and its historical information.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Collaborative Filtering","Content Analysis by AI","Data Science","Data Scientist","Deep Learning","filtering","Machine Learning","Python","RECOMMENDER SYSTEM"],"author_name":"Victor Dey","publish_date":"2021-08-25T11:00:00","publication_year":"2021","word_count":1845,"keywords":["Git","Collaborative Filtering","RECOMMENDER SYSTEM","Content Analysis by AI","R","Pandas","filtering","RAG","Data Science","Go","AI","ML","Machine Learning","recommendation systems","GAN","Python","Colab","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","Colab","Pandas","RAG","recommendation systems","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/collaborative-filtering-vs-content-based-filtering-for-recommender-systems\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10139719,"title":"GitHub Copilot Workspace to Make Developers&#8217; Lives Less Miserable","content":"GitHub’s tenth edition of its flagship developer conference, GitHub Universe 24, introduced several new feature additions to the GitHub Copilot Workspace—a collaborative environment within GitHub that helps developers use natural language to generate a structured plan based on the specifications of an issue, and seamlessly create a pull request. Copilot Workspace also leverages generative AI capabilities to assist in coding, allowing teams to iterate on their code effectively, and make changes across all the files in a repository. GitHub revealed that over 55,000 developers have used Copilot Workspace to plan, build, test, and run code using natural language, with over 10,000 pull requests merged. The Developer Environment of The Future Over the last six months, GitHub has released over 100 new additions to the Copilot Workspace, repeatedly suggesting that it is the ‘developer environment of the future’. The newly announced build and error repair feature suggests potential solutions to code errors and provides an option to fix the code manually or let Copilot Workspace rectify it automatically. GitHub also introduced new enhancements to the Brainstorming Mode in Copilot Workspace. This mode lets users collaborate and explore all the repositories, suggest solutions for issues, ask questions and help improve the overall problem-solving process. With the latest update, it can now automatically provide structure and organisation to a list of tasks and automatically update it when new questions are added. Moreover, GitHub also revealed that Brainstorming mode could derive context from multiple external sources. Another newly added feature, follow-up, helps users automatically update the project. Copilot Workspace implements all necessary changes across the codebase that are complete when a task impacts multiple files. GitHub also announced that Copilot Workspace can now be used directly in the context of Pull Requests, letting developers easily implement code changes and suggestions. A few days ago, GitHub introduced an AI vision feature that lets developers add image-based context to resolve issues. Users can add diagrams, screenshots, mockups, or photos, letting the AI model analyse the information inside the image. That said, the one game-changing update from GitHub Universe 24’ that can significantly impact Copilot Workspace is the integration of Claude, which is widely deemed the best foundational model for writing code. GitHub Copilot Workspace Has Hired the Best Coder in Town – Claude! Until today, GitHub’s Copilot was capable of writing code only using OpenAI’s GPT models. Now, developers can choose from OpenAI’s models and other industry-leading models, including Anthropic’s Claude 3.5 Sonnet. When developers approve suggestions from Copilot Workspace, it can automatically write code based on the generated plan. Using Claude to do so is nothing short of a game changer since it has been established that OpenAI’s models aren’t up to the mark when writing code. AIM compared multiple large language models (LLMs) for coding completion tests on LiveBench, and the results showed Claude 3.5 Sonnet on top. “In 2024, we experienced a boom in high-quality large and small language models that each individually excel at different programming tasks. There is no one model to rule every scenario, and developers expect the agency to build with the models that work best for them,” said GitHub CEO Thomas Dohmke. The new features align with the relationship GitHub Copilot Workspace has always wanted to establish with a developer. Copilot Workspace isn’t primarily focused on automatically generating code based on a single prompt or idea. It’s more than just a tool that auto-completes code—it focuses on assisting developers with a comprehensive plan of tasks before they generate a pull request. While several competitors exist in today’s AI ‘code companion’ market, one of GitHub’s advantages is that Copilot Workspace directly integrates into GitHub, which already boasts over 100 million developers and more than 420 million repositories. The new additions to GitHub Copilot and its newfound multimodel advantage threaten existing platforms like Cursor. In a podcast episode with Lex Fridman, Aman Sanger, co-founder of Cursor, said, “I think the Cursor a year from now will need to make the Cursor of today look obsolete.” This indicates if Cursor is to stay relevant in the competition, it has to pull its socks up. “You can wax poetic about moats and brand that, and this is our advantage, but I think in the end, just if you stop innovating on the product, you will lose,” said Micheal Truel, a co-founder of Cursor. It’ll be interesting to see how Cursor responds, and how the competition will evolve. From an Indian perspective, GitHub Copilot Workspace has the potential to enhance the workflow of more than 17 million developers – and as per 2024 GitHub’s Octaverse report, India is the fastest-growing developer community in the world. Moreover, the Indian developer community isn’t shying away from using AI to code, and as Thomas Dohmke says, “India’s booming developer community is using AI to build AI in record numbers, making it evermore likely that the next great multinational will come from India”.","excerpt":"GitHub has announced a slew of new features to GitHub Copilot Workspace, giving Cursor AI and other AI coding platforms a run for their money.","categories":["Global Tech"],"tags":["GitHub","Github Copilot","GitHub Copilot Workspace"],"author_name":"Supreeth Koundinya","publish_date":"2024-10-29T22:52:26","publication_year":"2024","word_count":816,"keywords":["Anthropic","OpenAI","AI","ML","GitHub Copilot Workspace","Github Copilot","Claude 3.5","RAG","Aim","generative AI","small language models","GitHub","R"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Claude 3.5","Anthropic","Aim","small language models","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/github-copilot-workspace-to-make-developers-lives-less-miserable\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10050389,"title":"Microsoft Windows 11 Releases A Day Early To Its Launch Date","content":"Microsoft has officially released Windows 11 in its stable form, and the new update is free for existing Windows 10 users on compatible PCs worldwide. Microsoft expects all eligible devices will have access to the free Windows 11 upgrade by mid-2022. The update was released a day earlier to its planned launch date of 5 October. The launch is known to be “phased and measured,” with new eligible devices getting the upgrade first and the rest getting offered the free upgrade sometime between October and mid-2022, depending on hardware, age of the device and other factors. A notification from Windows Update will be received letting you know when Windows 11 is available to one’s device in particular, or it can be checked manually. To install the new version, the Windows 11 system requirements dictate that your PC must have the following : CPU: 1GHz or faster with two or more cores on a compatible 64-bit processor or system on a chip (SoC)RAM: 4GBStorage: 64GB or largerSystem firmware: UEFI, Secure Boot capableTPM: Trusted Platform Module (TPM) version 2.0Graphics card: Compatible with DirectX 12 or later with WDDM 2.0 driverDisplay: High definition (720p) display that is greater than 9 inches diagonally, 8 bits per colour channelInternet: Windows 11 Home edition requires internet connectivity and a Microsoft account to complete device setup on first use. Windows 11 Pro edition does not. Windows 11, first introduced back in June, delivers a sleeker look compared to its previous versions. New features include support for Android apps, more detailed widgets, an update to layouts for snapping applications to the screen and a completely revamped Microsoft Store. Windows 11 also includes a new user interface, a redesigned Start menu, Microsoft Teams integration, and much more. The Windows 11 upgrade is available in the Windows Update section. In the Windows 10 settings area, one may even skip the queue and upgrade immediately using Microsoft’s new Installation Assistant. Microsoft claims that Windows 11 is a safer, more performant Windows that’s simple to use, with a welcoming design that’s meant to make using your PC for work and play easier than ever.","excerpt":"Microsoft claims that Windows 11 is a safer, more performant Windows that’s simple to use, with a welcoming design that’s meant to make using your PC for work and play easier than ever.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Android","latest advances","Machine Learning","updates","User Interface","windows 11"],"author_name":"Victor Dey","publish_date":"2021-10-05T13:32:36","publication_year":"2021","word_count":352,"keywords":["Go","windows 11","programming_languages:R","AI","latest advances","Machine Learning","programming_languages:Go","RAG","User Interface","Aim","programming_languages:Rust","ViT","updates","Rust","R","Android","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Rust","ViT","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-windows-11-releases-a-day-early-to-its-launch-date\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001248,"title":"The Downfall Of Dark Web Marketplaces","content":"Internet today is a vital part our daily lives and when we talk about the internet, our brain only thinks about the day-to-day activities — from watching a video to checking the news to booking a vacation online to killing time on social media. However, under the surface lies a dark corner of the web where terrorists, criminals, and whistleblowers lurk — the Dark Web. Understanding The Web The web is divided into three parts: Surface Web, Deep Web, and Dark Web. There was a lot of misconception regarding these terms. However, with time, as cybersecurity evolved, these terms have come to light. Surface Web Surface web is basically the things that we surf on the internet on a day-to-day basis or you can also say the websites that are indexed by the search engines. In the surface web, anyone can go to any website and grab information. Understanding the surface web is not a difficult task. The most interesting thing here is that the surface web is consist of billions of websites, however, it is not more than 10% of the entire web. The internet is a massive place and there are places that many of us don’t know. Deep Web The second layer of the Web is the Deep web. It is that part of the internet, which is not visible to normal users. The websites and web pages on the deep web are not indexed by search engines, and these can only be accessed with prior permission from the authorized bodies. For example, cloud storages, personal data, and military data etc. Dark Web This layer I considered to be the most mysterious and dangerous. Dark web is basically home to all criminal and illegal activities on the internet — from drug dealing with arms dealing with hiring hitman. The Dark web can only be accessed using the onion routing or the TOR browser. However, it is always advised to stay away from the dark web as the activities there is shadowy and dangerous beyond imagination. The Downfall Of The Dark Web There was a time when Dark Web was creating a lot of nuisance. Criminals were having a strong hand on the dark web. Whether it’s about drugs or arms or murders, everything had links with the dark web. However, with time and as legal bodies started to intervene, things have come down normal and pushing dark web towards its downfall. Here Are The Three Top Dark Web Marketplaces That Has Been Taken Down: Silk Road Silk Road is one of the most well know marketplace on the dark web. It was basically the internet’s haven for drug dealers, gun runners, and document forgers. Founded by a young libertarian, Ross Ulbricht, the bustling black market for drugs was also a crypto anarchist’s dream — a trusted trading ground on the Internet. The marketplace was operated under a Tor hidden service. So, the communication on the website was completely anonymous. Also, in order to maintain anonymity,  every transaction on the Silk Road was done using bitcoin. On 2 October 2013, Ulbricht’s dream of an online libertarian paradise came to an end as FBI arrested him in Glen Park Library, a branch of the San Francisco Public Library. At the time of his arrest, Ulbricht’s was logged into Silk Road as an administrator. Also, the FBI seized 26,000 bitcoins from accounts on Silk Road, worth approximately $3.6 million. And just a few hours after his arrest, Silk Road’s domain was seized, and the marketplace was shut down. AlphaBay AlphaBay is another shady dark web marketplace that offered weapons, drugs, and stolen identities and it was operating on the Tor network. The marketplace was initially launched in September 2014 and was gained good traction. AlphaBay was even placed as a top-tier market on the website of Gwern Branwen, writer & independent researcher. That is not all, it was even recognized as the largest online darknet market in October 2015. And by July 2017, AlphaBay was ten times the size of world-famous dark web marketplace, Silk Road. But, in just like Silk Road, even AlphaBay faced the wrath. Following a “landmark” investigation involving law enforcement from around the world, the world’s largest markets on the dark web were shut down. Alexandre Cazes, who was believed to be the mastermind behind AlphaBay was arrested on July 5 in Thailand. And a week later, Cazes was found dead in a Thai jail. When Cazes was arrested in Thailand, he was found performing an administrative reboot on an AlphaBay server using his laptop. Also, his laptop was completely unencrypted and contained a net worth statement mapping all global assets, which was later seized. Hansa Hansa was another well know dark web marketplace for all the vice. After the AlphaBay shut down, there were talks that Hansa would become the leading market. However, that didn’t happen as it was also been taken down. On  July 20, 2017, the Dutch Police revealed that it had been compromised by law enforcement and had taken control of Hansa Market in June. The Dutch nation police also said that they had been operating the market since then, in order to track and monitor the vendors and customers. The interesting thing here is, in 2016, security shop Bitdefender tipped off the Dutch national police that is hosted in the Netherlands, and since then they started their investigation and found that two individuals were running the site as German nationals. They were also arrested in Germany and were accused of operating an illegal marketplace.","excerpt":"Internet today is a vital part our daily lives and when we talk about the internet, our brain only thinks about the day-to-day activities — from watching a video to checking the news to booking a vacation online to killing time on social media. However, under the surface lies a dark corner of the web […]","categories":["AI Features"],"tags":["dark web","internet"],"author_name":"Harshajit Sarmah","publish_date":"2019-02-21T21:26:42","publication_year":"2019","word_count":923,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","dark web","RAG","BERT","llm_models:BERT","ViT","Rust","R","internet"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","BERT","ViT","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-downfall-of-dark-web-marketplaces\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10028097,"title":"Unity Launches Synthetic Image Datasets To Train AI Models Faster","content":"San Francisco-based videogame software development company Unity Technologies recently announced the launch of synthetic image datasets to help develop computer vision applications (train artificial intelligence (AI) models) faster and reduce cost significantly. Unity — a cross-platform game engine — is widely used by game developers worldwide to create interactive games, virtual reality and augmented reality applications. With its recent announcement, the company looks to leverage its datasets to build AI models across industry verticals, including manufacturing, retail and security. Tackling the privacy problem Unity believes most real-life data collection techniques are labour intensive, expensive, and carry greater privacy risks. The firm came up with synthetic image datasets to simplify data collection and protect the confidentiality of data. As real data contains sensitive information, most programmers, software developers or researchers may not want them to be disclosed. Synthetic data, however, holds no private information and can not be traced back to the source. Most importantly, synthetic data addresses confidentiality and privacy concerns to a large extent and eliminates the privacy issues arising from using images of real people and places. For example, in autonomous or self-driving cars, the collection of real data is unreasonably expensive. Waymo, the self-driving vertical of Alphabet, has spent close to $3.5 billion in testing Chrysler Pacificas in Silicon Valley and Phoenix. Over the past few years, around 30 self-driving car companies have spent close to $16 billion on developing fully self-driving cars. At present, the collection of synthetic data is expensive and time-consuming. An AI model must initially create it, and many companies lack the resources to do so. To that end, Unity has come up with a Unity Perception SDK and a library with labelling and randomisation tools for developers. Eliminates data bias Unity said the usage of synthetic data eliminates the problem of ‘biased data,’ which often results in skewed outcomes, lower accuracy levels and analytics errors. “Data captured from the real world is often biased towards what is easy to collect, is subject to human labelling errors, and needs to be refreshed often, which can be very expensive,” said Danny Lange, SVP of AI at Unity, comparing it with synthetic datasets. Lange said the best AI results are achieved with a large amount of high-quality synthetic data combined with a small amount of real data, when possible. The synthetic version of datasets validates privacy rules and accurately reflects real world-data, he added. Further, he said these datasets empower companies to simulate scenarios that might occur in the real world shortly based on a sizable increase in user data. “As a result, we see smarter indoor environments, such as cashier-less grocery stores, and more as our customers discover new applications,” said Lange. How does it work The new Unity computer vision datasets are based on synthetic data, which is generated using algorithms. Conversely, synthetic data is used for computer vision applications, particularly in the area of object detection. In the case of Unity, the artificial environment is most likely to be used in creating a 3D model of the object and learning to navigate environments by visual information. Unity engine creates ‘digital twins’ of objects (3D models) using photogrammetry techniques. Digital twins are a virtual replica of the physical things, mainly used to run simulations (testbed) before the actual deployment of the solution. The digital twins are then placed in various 3D environments or randomisers, with multiple lighting conditions, textures, camera positions, scale factors, and other parameters. Source: Unity [Showcasing the visual examples of image labelling] Unity recommends various environments best suited to address the customer’s computer vision problem and look for the most suitable dataset. Currently, the Unity team provides the necessary handholding required for its customers. Soon, the company plans to offer a simple self-service interface to generate additional features at their convenience. Unity offers a tiered pricing model. The price per image falls proportionally with the increased need for synthetic images\/data. Lange believes synthetic computer vision datasets can support a wide range of AI training use cases, ranging from object detection to improving the performance of AI models.","excerpt":"Unity came up with synthetic image datasets to simplify data collection and protect the confidentiality of data.","categories":["IT Services"],"tags":["data privacy use cases"],"author_name":"Amit Naik","publish_date":"2021-04-21T17:00:00","publication_year":"2021","word_count":677,"keywords":["Go","artificial intelligence","synthetic data","AI","data privacy use cases","computer vision","RAG","Git","object detection","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","computer vision","analytics","RAG","object detection","R","Go","Git","synthetic data"],"url":"https:\/\/analyticsindiamag.com\/it-services\/unity-launches-synthetic-image-datasets-to-train-ai-models-faster\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10014140,"title":"IIT Kanpur Launches A Department Of Cognitive Science To Understand Human Mind With Futuristic Technology","content":"Indian Institute of Technology (IIT) Kanpur, the first among the IITs, has launched a new department of cognitive science. The aim of establishing this fully-fledged department of cognitive science is to devise futuristic technology to understand the human mind. According to IIT Kanpur, it will be done by understanding, mirroring and adapting the complexity of the perception and processes of the human mind. The institute further believes that the initiative will be of great importance in mapping traditional Indian knowledge and understanding of the architecture of the human mind with futuristic technology. With last year’s visit by a team of cognitive science experts as a part of the national advisory committee, the department was established at IIT Kanpur as a recommendation given its strengths in teaching and research in the area over the last couple of decades. When asked about establishing a separate department for cognitive science, Prof. Abhay Karandikar, Director, IIT Kanpur stated that it is to provide impetus to the research and teaching programs in cognitive science at the institute. Not only this new department will address this gap, but will also help the country to become a leader in the field of cognitive science. Also Read: IIT Kanpur To Offer Free Statistics Courses During The Lockdown Period The department will work on understanding the human mind and its mechanisms involved in mental processes and interface on neuroscience-related areas. It will also collaborate with other departments like computer sciences and engineering in order to develop new artificially intelligent and machine-learning interfaces, stated professor Karandikar. The department of cognitive science has been designed to prepare students for careers by offering knowledge in the fields of artificial intelligence, data analysis, healthcare, product and software design, and human performance, among others. With that being said, this isn’t the first step by IIT Kanpur to venture into cognitive science. As a matter of fact, the institute, in June 2017, has formally established an interdisciplinary program in cognitive science. However, the proposal of establishing this separate department was approved by the board of governors. Also Read: How IIT Kanpur’s New AI Assistant Is Different From Other Corona Helpdesks Explaining that, Dr K Radhakrishnan, the chairman, Board of Governors stated to the media that considering India has been known for its classical knowledge about mind and consciousness, it is indeed surprising that the country is lagging in studying the human mind. Thus, this new department of cognitive science will provide learners with an opportunity to explore the machinist of the human brain and its functioning. Further to this, Prof Bishakh Bhattacharya, the current head of the IDP in cognitive science also stated that the institute has already been running PhD and MS programmes in cognitive science. And this department will provide real-world contributions that will be valuable for developing AI, multimodal communications and cognitive robots. Professor Bhattacharya also believes that such a launch can also help in understanding the impact of deadly viruses like COVID. Cognitive science can help researchers understand the virus and identify its effects, such as signals in speech and vocal system coordination among people.","excerpt":"Indian Institute of Technology (IIT) Kanpur, the first among the IITs, has launched a new department of cognitive science. The aim of establishing this fully-fledged department of cognitive science is to devise futuristic technology to understand the human mind.  According to IIT Kanpur, it will be done by understanding, mirroring and adapting the complexity of […]","categories":["AI News"],"tags":["Artificia Intelligence in Data Science","iit kanpur"],"author_name":"Sejuti Das","publish_date":"2020-12-14T10:52:42","publication_year":"2020","word_count":514,"keywords":["Go","artificial intelligence","iit kanpur","AI","Modal","programming_languages:R","programming_languages:Go","Aim","Artificia Intelligence in Data Science","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","Modal","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-kanpur-launches-a-department-of-cognitive-science-to-understand-human-mind-with-futuristic-technology\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10093445,"title":"Amazon Throws Another Hat in the Gen AI Ring","content":"Amazon is racing to join the AI chatbot race. As pointed out by Bloomberg, the tech goliath, which serves 54% of all product queries, has posted job listings for a machine learning-focused engineer describing how it is “reimagining Amazon Search” with a new “interactive conversational experience” to answer product questions, compare products, personalise suggestions, and more. But this is not the e-commerce giant’s first attempt to integrate AI\/ML to enhance users’ experience via Gen AI. “This will be a once in a generation transformation for Search, just like the Mosaic browser made the Internet easier to engage with three decades ago,” Amazon wrote. “If you missed the 90s—WWW, Mosaic, and the founding of Amazon and Google—you don’t want to miss this opportunity.” Furthermore, Amazon expressed its eagerness to swiftly implement these changes, stating, “We want to deliver this vision to our customers right away.” Previously, Amazon’s search experience has been heavily criticised for its sponsored content and ads heavily bombarding the results. A study by The Washington Post showed an overwhelming number of sponsored products displayed under different guises. The products not only exist under the label “sponsored” but also covertly have their presence under “highly rated”, while many also constitute products from Amazon’s own brands. In 2017, Amazon allowed its renowned assistant Alexa to purchase products through voice controls. Right from adding the products to cart to cancelling the order, the assistant provided features to ease the users’ shopping experience. Even though the company already had an existing product to upgrade, it seems to have decided to take the long haul for further improving its flagship search. Read: Who is to Blame for Your Bad Shopping Experience at Amazon? Earlier this year, Amazon did release Amazon Lex, to help retailers build out conversational interfaces for applications using voice and text. The chatbot uses the same conversational engine that powers Amazon Alexa and is available to use in new and existing applications. But the company shied away from integrating the engine for its own purpose. Existing Alternatives A chatbot is a starting point for users looking to shop with specific parameters. Ask ChatGPT-powered Microsoft Bing to show the five best colognes and it will pull up a list of five products, citing reviews from GQ, along with links to stores selling the products. Amazon’s hustle seems understandable with its rivals’ efforts catching the attention of users. Currently, a search on Amazon yields a bunch of ads, followed by some genuine products. Hence, the search king seems to be in a dire need to update its service via AI. Two weeks ago at Google I\/O, the Mountain-view based firm introduced genAI for its holy cash cow ‘Search’. With the latest addition, users can consider complex purchase decisions easily. When searching for a product, the users get a snapshot of factors to consider and along with product descriptions that include reviews, ratings, prices and product images. The reason for the relevant information is Google’s Shopping Graph, which has a comprehensive dataset of over 35 billion product listings. The Pichai-led firm stated that, hourly over 1.8 billion listings are refreshed in the graph. Early implementations of genAI by tech giants Microsoft and Google have challenges, particularly when it comes to providing accurate responses to queries. Despite the setbacks, these steps show the potential for an enhanced Microsoft Bing or Google search, which could offer users insightful means to shop online. Fashionably Late The past few months have seen the tech industry drooling over the potential of generative AI but that wasn’t the case for Amazon. The tech giant took its own sweet time to announce its foundational models, too, similar to the upcoming AI chatbot. In April, Amazon announced a foundational suite called Bedrock for its AWS cloud customers to leverage some of the most popular genAI models via an application programming interface (API). Apart from some of its own models collectively called “Titan” the announcement included Anthropic model for conversations and questions, AI21 Labs model for translation and Stability AI’s model for image generation. Though Amazon arrived late to the generative AI party. Up until now, its releases have been strategically smart. With its latest under the garb chatbot on the way Amazon is definitely pushing for more AI coming its way but the company’s fate in the ring is yet to be determined.","excerpt":"After arriving late to the genAI party with Bedrock, the search giant seems to be hush-hush about its AI chatbot","categories":["AI Features"],"tags":["AI Chatbot","Bard","bing ai","Bing Chat","ChatGPT","Google Bard","OpenAI"],"author_name":"Tasmia Ansari","publish_date":"2023-05-16T17:24:41","publication_year":"2023","word_count":722,"keywords":["Anthropic","ChatGPT","GenAI","machine learning","OpenAI","AI","AI Chatbot","Google Bard","ML","AWS","Bard","RAG","bing ai","generative AI","R","Bing Chat"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","GenAI","ChatGPT","Anthropic","RAG","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/amazon-throws-another-hat-in-the-genai-ring\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":26484,"title":"Facebook’s Talking Bots Can Now Help You Explore New York City","content":"Virtual assistants are going a long way to making our life easier. They find information for us, play songs for us and also make calls apart from many other daily activities. But the Facebook Artificial Intelligence Research (FAIR) centre is far from being satisfied. Researchers Douwe Keila, Jason Weston, Harm de Vries, Kurt Shuster, Dhruv Batra and Devi Parikh, want to take AI and virtual assistants to the next level. They want to use natural language to make travel more attractive. Full comprehension of human language has been a difficult aim to achieve for some time now. The research team at FAIR are teaching the artificial intelligence systems to understand language by getting them to ‘guide’ virtual tourists around New York City. They have developed a new research task, called Talk The Walk, which explores this embodied AI approach while introducing a degree of realism not previously found in this area. New AI Task In this new AI task, two agents have to talk to each other and accomplish a goal. The goal is to navigate to a specific location in the city or any tourist location. But the setting is not game-like, which can be seen in many other tasks. Here, the goal is to navigate and roam through 360-degree images of actual city streets, in this case, New York. The researchers at FAIR have created a guide agent which sees a map of the streets and the neighbourhood. Using a novel attention mechanism called MASC (Masked Attention for Spatial Convolution), the researchers helped the guide bot focus on the right place on the map. This work really goes on a long way to improve the research community’s understanding of perception and communication. Those aspects can lead to grounded language learning and provide a stress test for language as a method of interaction. The FAIR team have also released the baselines and data set for the task of Talk The Walk. This will help other researchers by providing a framework to evaluate embodied AI models especially related to dialogue. The Unlikely Pairing Of Tourism And AI 360-degree images were used by FAIR researchers to train the systems and demonstrate grounded language. The researchers took images from five New York city neighbourhoods — Hell’s Kitchen, East Village, Financial District, Upper East Side, and Williamsburg. The areas have grid-based layouts and four-cornered intersection and serve as the first-person perception for the tourist agent. In an era when simulation rules, FAIR’s bet on working with realism is like a breath of fresh air in AI. Researchers at FAIR have been also successful to create natural language conversations between two agents. In fact, researchers preferred real human talk rather than worded messages such as one sees on Google Maps. The interaction between AI agents promises to be closer to the language we use in our day-to-day life. The participants were given the same guide agents and tourist agents with the same navigation goals and constraints. The guiding agent has only access to only a two-dimensional overhead map with landmarks and noticeable places such as restaurants and hotels. Both agents are not able to see what the others see. Therefore, reaching a place requires great communication between the two agents. Each experiment has the target achieved when guide agent thinks that the tourist agent has reached the required destination. If the prediction is accurate the episode, (using some RL terminology) is marked as successful otherwise marked as incorrect. There is no limit on the number of communication messages. Building Communication Between Agents Communication between artificially intelligent agents has been always a prime target. The research team at FAIR looked to focus their research on natural language but also came up with communication protocols for robots which are different from human ones. There were two scenarios as researchers Weston and Kiela put it : “In the first setting, agents communicated via continuous vectors, meaning they transferred raw data to one another. Those continuous vectors included, for example, representations of what the tourists were observing and doing, to help the map-based guides localise their counterparts.” The second output, “The second emergent communication setting took a different approach, using what the researchers refer to as synthetic language. In this setting, communication was far more simplified than natural language, using a very limited set of discrete symbols to convey information. By giving the bots the option of communicating in the simplest form possible, the interactions are fast and precise and give us a good idea of how well we could perform with natural language.” Building Environment based AI The researchers try to make it very clear that this task is not a competition. They stated: “Talk the Walk isn’t a competition between natural language and synthetic interactions but rather an attempt to offer clarity and quantifiable results related to the ultimate goal of creating machines that can effectively “talk” to humans and to one another.” They also added that grounding AI tasks in reality is also very important. Going forward environment-based AI will be really important. Again attention mechanisms are used extensively to translate embeddings of places to tourist state transitions (directions to take left or right). This also helps the guide agent to know where the tourist agent is currently. The researchers stress that grounding the AI tasks in reality is important. But building such systems can be hard. For example, some tasks like reading letters in signage were not taken up by the researchers. This gives us a good understanding of building embodied AI can be really difficult because it consists of perceiving a given environment, navigating through it, and communicating about it.","excerpt":"Virtual assistants are going a long way to making our life easier. They find information for us, play songs for us and also make calls apart from many other daily activities. But the Facebook Artificial Intelligence Research (FAIR) centre is far from being satisfied. Researchers Douwe Keila, Jason Weston, Harm de Vries, Kurt Shuster, Dhruv […]","categories":["AI Features"],"tags":["AI Agents","Natural Language","Natural Language Processing"],"author_name":"Abhijeet Katte","publish_date":"2018-07-17T11:44:52","publication_year":"2018","word_count":937,"keywords":["Go","artificial intelligence","TPU","attention mechanism","AI","Natural Language Processing","RPA","AI Agents","virtual assistants","Aim","Natural Language","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","virtual assistants","TPU","R","Go","attention mechanism","ViT","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/facebooks-talking-bots-can-now-help-you-explore-new-york-city\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":63938,"title":"Top 9 Machine Learning Frameworks In Julia","content":"Julia is a high-level, dynamic programming language which is fast, flexible, easy-to-use, scalable, and supports high-speed mathematical computation. The programming language also supports all hardware, including GPUs and TPUs on every cloud. Julia uses multiple dispatches as a paradigm, making it easy to express many object-oriented and functional programming patterns. In one of our articles, we discussed how this language is making AI and machine learning better. In this article, we list down top 9 machine learning frameworks in Julia, one must know. (The libraries are listed according to their stars on GitHub) 1| Flux About: Flux is deep learning and machine learning library that provides a single, intuitive way to define models, just like mathematical notation. Any existing Julia libraries are differentiable and can be incorporated directly into Flux models. The intuitive features include compiled eager code, differentiable programming, GPU support, ONNX, among others. Click here to know more. 2| Mocha.jl About: Mocha.jl is a deep learning library for the Julia programming language, which includes a number of features mentioned below: Written in Julia and for Julia: Mocha.jl is completely written in Julia. This means that the library has native Julia interfaces and is capable of interacting with core Julia functionality as well as other Julia packages.Minimum dependencies: This library includes minimum dependencies to use Julia as backend, and there is zero need for root privileges or installation of any external dependencies. Multiple backends: This library comes with a GPU backend, combining customised kernels with highly efficient libraries from NVIDIA such as cuBLAS, cuDNN, etc.Modularity and correctness: This library is implemented in a modular architecture. Click here to know more. 3| Knet About: Knet is a deep learning framework implemented in the Julia programming language.  Knet allows models to be defined by just describing their forward computation in plain Julia, allowing the use of loops, conditionals, recursion, closures, tuples, dictionaries, array indexing, concatenation and other high-level language features. The library supports GPU operations and automates differentiation using dynamic computational graphs for models defined in plain Julia. Click here to know more. 4| TensorFlow.jl About: TensorFlow.jl is also a Julia wrapper for popular open-source machine learning TensorFlow. This wrapper can be used for various purposes such as fast ingestion of data, especially data in uncommon formats, fast postprocessing of inference results, such as calculating various statistics and visualisations that do not have a canned vectorized implementation. Click here to know more. 5| ScikitLearn.jl About: ScikitLearn.jl is a Julia wrapper for the popular Python library Scikit-learn. It implements the Scikit-learn interface and algorithms in Julia.  It provides a uniform interface for training and using models, as well as a set of tools for chaining (pipelines), evaluating, and tuning model hyperparameters. It supports both models from the Julia ecosystem and those of the Scikit-learn library. Click here to know more. 6| MXNet.jl About: MXNet.jl is the Apache MXNet Julia package that brings flexible and efficient GPU computing and state-of-art deep learning to Julia. The features of this library include efficient tensor and matrix computation across multiple devices, including multiple CPUs, GPUs and distributed server nodes. It also has flexible symbolic manipulation to composite and construction of state-of-the-art deep learning models. Click here to know more. 7| MLBase.jl About: MLBase.jl is a Julia package that provides useful tools for machine learning applications. It provides a collection of useful tools to support machine learning programs, including data manipulation and preprocessing, score-based classification, performance evaluation, cross-validation and model tuning. Click here to know more. 8| Merlin About: Merlin is a deep learning framework written in Julia. The library aims to provide a fast, flexible and compact deep learning library for machine learning. The requirements of this library are Julia 0.6 and g++ for OSX or Linux. The library runs on CPUs and CUDA GPUs. Click here to know more. 9| Strada About: Strada is an open-source deep learning library for Julia, based on the popular Caffe framework. The library supports convolutional and recurrent neural network training, both on CPUs and GPUs. Some of the features of this library include flexibility, support for Caffe features, integration with Julia and other such. Click here to know more.","excerpt":"Julia is a high-level, dynamic programming language which is fast, flexible, easy-to-use, scalable, and supports high-speed mathematical computation. The programming language also supports all hardware, including GPUs and TPUs on every cloud. Julia uses multiple dispatches as a paradigm, making it easy to express many object-oriented and functional programming patterns. In one of our articles, […]","categories":["AI Trends"],"tags":["julia computing","julia data scientist","Julia for developers","Julia Language","julia machine learning","Machine Learning","machine learning framework","multiple classification statistics","PowerBI","scikit learn"],"author_name":"Ambika Choudhury","publish_date":"2020-04-29T18:00:00","publication_year":"2020","word_count":690,"keywords":["scikit-learn","PowerBI","Ray","deep learning","scikit learn","julia data scientist","Julia for developers","Julia Language","GPU computing","machine learning","AI","machine learning framework","neural network","ML","Machine Learning","julia machine learning","julia computing","multiple classification statistics","Aim","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","Aim","Ray","TensorFlow","scikit-learn","GPU computing"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-9-machine-learning-frameworks-in-julia\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10041990,"title":"Nasscom Rubbishes Report Projecting 3Mn Job Losses In IT Sector","content":"Nasscom president Debjani Ghosh refuted the PTI report claiming three million job losses in India’s IT sector by 2022. The PTI story cited a Bank of America report saying: “Tech giants including TCS, Wipro, Infosys, Tech Mahindra, HCL and Cognizant and others appear to be planning for a 3 million reduction in low-skilled roles by 2022 because of the Robot process automation (RPA) up-skilling.” “Seriously don’t get why publications don’t verify data and use the right source! Not only r tech cos increasing workforce, but we have more demand than supply in several emerging areas like data analytics etc. We need keep up and upskill fast  .. the jobs exists,” tweeted (sic) Debjani Ghosh. Seriously don't get why publications don't verify data and use the right source! Not only r tech cos increasing workforce, but we have more demand than supply in several emerging areas like data analytics etc. We need keep up and upskill fast .. the jobs exists. https:\/\/t.co\/CKKDK5HEK7— debjani ghosh (@debjani_ghosh_) June 18, 2021 Roughly 0.7 million roles are expected to be replaced by RPA alone and the rest due to other technological upgrades and upskilling by the domestic IT players, while the RPA will have the worst impact in the US with a loss of almost 1 million jobs, according to the Bank of America report quoted by news outlets. Meanwhile, India.com carried a story claiming the report is from Nasscom: “As automation is moving at a much rapid pace across industries, notably in the IT-sphere, domestic software enterprises with over 16 million employees are on track to cut headcounts by a colossal 3 million by 2022, saving them a phenomenal USD 100 billion in salaries every year, as per the NASSCOM report.” However, NASSCOM has cleared the air and tweeted from its official handle: “We would like to clarify that the data points shared in this article by @indiacom are not from any NASSCOM report.The industry has & will continue to be a net creator of jobs & is committed to people-centric innovation, & relentless talent focus.” We would like to clarify that the data points shared in this article by @indiacom are not from any NASSCOM report.The industry has & will continue to be a net creator of jobs & is committed to people-centric innovation, & relentless talent focus. https:\/\/t.co\/aOF4PVdg5X— nasscom (@nasscom) June 18, 2021","excerpt":"Nasscom president Debjani Ghosh refuted the PTI report claiming three million job losses in India’s IT sector by 2022. The PTI story cited a Bank of America report saying: “Tech giants including TCS, Wipro, Infosys, Tech Mahindra, HCL and Cognizant and others appear to be planning for a 3 million reduction in low-skilled roles by […]","categories":["AI News"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-06-18T12:42:29","publication_year":"2021","word_count":391,"keywords":["API","programming_languages:R","AI","RPA","innovation","automation","Aim","analytics","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","API","automation","RPA","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nasscom-rubbishes-report-claiming-3mn-job-losses-in-it-sector\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":68489,"title":"How Self-Taught ML Expert Agnis Liukis Became A Kaggle Grandmaster","content":"“I’ve exchanged most of my evenings of watching TV for evenings of competing on Kaggle.”Agnis Liukis For this week’s ML practitioner’s series, Analytics India Magazine got in touch with Agnis Liukis from Latvia, who is a Kaggle Grandmaster ranked 14th in the global leaderboard. He shares his insightful journey, tips and tricks in this interview. How It All Began Despite featuring in top 10 of the highly competitive Kaggle contests, Agnis, surprisingly, is neither a data scientist nor a machine learning engineer.  He currently works as a Lead Software Architect in TietoEVRY, Latvia where his main focus is on developing web solutions for the backend and frontend. His foray into machine learning is a combination of keeping up with trends and his love for programming and mathematics. He has a Master’s degree in Information Technology, and all his Kaggle accomplishments are the direct result of his zeal to know more. Agnis is self-taught, and his journey into Machine Learning began five years ago when data science as a career was getting more attention globally. Agnis found this domain to have a promising future and started taking online courses. He started with the course from Andrew Ng, which he also recommends to beginners. However, he soon realised that theory would only take him so far. He wanted to get hands-on experience of dealing with machine learning models. His search for high-quality machine learning challenges landed him in the coveted Kaggle community. “At the beginning, there’s a lot of information to learn. One must learn theory to understand different algorithms, their weak and strong sides, and when to use which of them. They must also know things like what’s overfitting and how to deal with it. And many more things. And in parallel, there’s a lot of tools and frameworks available – they will have to be learned as well,” says Agnis, remembering his initial days. Though he wasn’t new to Python, it took him some time to get used to Pandas and NumPy. And other popular libraries to be successful in competitions, like LightGBM, XgBoost, Scikit-learn, Matplotlib etc. Today, after having participated in more than 100 competitions on Kaggle, Agnis has eight gold medals and is currently ranked 14th out of 1,39,500 participants. Kaggle Journey “One doesn’t have to be an expert to start competing. In my first competition on Kaggle, I got a silver medal based on my knowledge in Math and Probabilities.” Agnis says that the yearning for competition comes to him naturally. Be it sports or computer games, he likes to compete. “So it was quite natural for me to enter the world of Competitive Data Science once I decided to learn Data Science,” says Agnis. When it comes to approaching a problem on Kaggle, Agnis says that he would first read and try to understand the underlying problem, the data and its purpose. When it comes to data exploration, he underlines the importance of looking at raw data, which is essential, especially for competitions. “Many EDAs in Kaggle are typically approaching data from “heights” – looking at some general statistics, distributions, trends and so on. But sometimes, key insights are hidden in raw data values.” “For example, some competitions have a special pattern of digits after comma – like many values ending with .33333, .66666 and so on,  which give some clues about how this data was obtained and how to use that information to improve the score. And things like this can’t be seen from general statistics, but only from raw data,” explained Agnis. When some working pipeline is ready, he would usually make some initial submissions to calibrate the cross-validation, which he considers to be an essential strategy as it helps in better generalisation and allows to avoid overfitting on the public leaderboard. He reveals that falling during shake-up when a private leaderboard is revealed is one of the most threatening things all competitors are afraid of. Hence, good cross-validation is a key factor in avoiding this. Achieving the top spot in Kaggle takes time, says Agnis. He stresses on the importance of taking out some time to learn things, to explore data, to write code for models, do experiments, read and explore ideas in a competition forum.  Time is a significant differentiator, and he laments about lacking the same right now as he has a full-time job and a family. “Basically I’ve exchanged most of my evenings of ‘watching TV’ for evenings of ‘competing on Kaggle’. That’s more fun and also more useful,” quipped Agnis. Agnis looks at Kaggle as a great way to stay up-to-date with all bleeding-edge technologies and approaches in Data Science and Machine Learning. Tips, Tricks And Tools For competitions, Agnis typically works with his home computer (16 GB RAM), which he considers to be enough for most of the problems, or at least for tabular data and NLP. For computationally intensive contests, he prefers to create a virtual machine on Google Cloud Platform with desired power. However, he reminds us that lately, many Kaggle competitions are in “Code competition” format, requiring submissions to be made through Kaggle kernels, and he often uses resources offered by Kaggle, which is about the same power as his home computer, but with an advantage of launching multiple kernels in parallel. “Fail fast when testing new ideas. If an idea is not working, simply forget about it and try something new.” To avoid getting caught up in non-working ideas for weeks and wasting a lot of time, Agnis recommends the participants to move on to the next idea as soon as possible. That said, Agnis does admit to giving his failed ideas one final shot before scrapping them just in case there are any bugs. When it comes to libraries, Agnis expresses his fondness for LightGBM, mainly due to its speed and low memory requirements. LightGBM is also his primary option for all tabular data problems. For neural networks, he prefers the popular Keras library. For the rest of Data Science tasks in Python, he finds Pandas, NumPy, Scikit-learn, Matplotlib to be quite handy. Talking about how smart one needs to be to go from good to great, Agnis took the example of one of the competitions, which he has won by teaming up with Evgeny Patekha. The competition titled ‘Sberbank Russian Housing Market’ deals with predicting house prices in Moscow. Agnis recollected how hard the dataset was and how difficult it was to get stable cross-validation due to the small size of data and many significant outliers. The summary of his thought process that fetched him gold: there were two significantly different types of products there, and critical factor was to notice that and train separate models for each type;only one of those product types contained outliers, therefore, using only the other type for CV worked well;The strongest feature when predicting price there was the full area of the house, but it contained a lot of errors in data (like typos, unrealistic values, etc.). So, remove the most powerful feature and include it in some derived features. More detailed solution can be found here Future Of ML And Why Creativity Triumphs “I can’t imagine ML disappearing from recommendation systems, risk scoring applications, automated text translation, and many other fields.” When asked whether the hype around machine learning is real, Agnis stayed loyal to his opinion from the initial days about how promising data science would be. He opined that machine learning is here to stay and its importance will only grow in the coming years as we collect more data, leverage more computer power, tweak algorithms and make them affordable. He also speculates that there will be tremendous opportunities for the AutoML field as many companies cannot afford hiring good data scientists for building models from scratch. However, he also warns us that many people think machine learning to be some kind of super algorithm which can solve problems perfectly. For instance, he explains, using computer vision for medicine cannot be taken to be 100% perfect as the impact of failure can be huge. But, an ML model can assist a doctor in assessing results more confidently. “ML helps only if it is used in a smart way; understanding what it can do and what it can’t.” For machine learning aspirants, he suggests getting hands-on with real ML tasks or competitions as soon as possible because courses and theories may seem clear and straightforward. But, when one is tasked with 1 million rows of raw data, the real training begins! For the participants who are looking to top the Kaggle leaderboards, Agnis advises that using good models just won’t win the competitions. “To stand out and get some real advantage, it is necessary to do something different, find something that others didn’t notice. And creativity is something that really helps in this. Another thing is patience and endurance, as most creative ideas won’t work and it can take many iterations and experiments to get things working,” concluded Agnis.","excerpt":"“I’ve exchanged most of my evenings of watching TV for evenings of competing on Kaggle.” Agnis Liukis For this week’s ML practitioner’s series, Analytics India Magazine got in touch with Agnis Liukis from Latvia, who is a Kaggle Grandmaster ranked 14th in the global leaderboard. He shares his insightful journey, tips and tricks in this […]","categories":["AI Features"],"tags":["datascience","Interviews and Discussions","Kaggle"],"author_name":"Ram Sagar","publish_date":"2020-06-29T18:00:00","publication_year":"2020","word_count":1495,"keywords":["data science","machine learning","Kaggle","Keras","AI","neural network","datascience","ML","computer vision","NLP","Ray","analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","NLP","computer vision","data science","analytics","Ray","Keras"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/agnis-liukis-kaggle-grandmaster-interview\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":56173,"title":"Automatic Speech Transcription And Speaker Recognition Simultaneously Using Apple AI","content":"Last year, Apple witnessed several controversies regarding its speech recognition technology. To provide quality control in the company’s voice assistant Siri, Apple asked its contractors to regularly hear the confidential voice recordings in the name of the “Siri Grading Program”. However, to this matter, the company later apologised and published a statement where it announced the changes in the Siri grading program. This year, the tech giant has been gearing up a number of researchers regarding speech recognition technology to upgrade its voice assistant. Recently, the researchers at Apple developed an AI model which can perform automatic speech transcription and speaker recognition simultaneously. Behind the Model The researchers presented a supervised multi-task learning model. In this model, the speech transcription branch of the network is being trained to decrease a phonetic connectionist temporal classification loss. Also, the speaker recognition branch of the network is being trained to label the input sequence with the correct label. The model has been trained using several thousand hours of labelled training data for each task. The speech transcription branch of the network has been evaluated on a voice trigger detection task while the speaker recognition branch has been evaluated on a speaker verification task. How it Works Although, the tasks of automatic speech transcription and speaker recognition are inter-related, yet they are treated independently in most of the scenarios. In this work, the researchers tried to solve this issue by representing a single network which can efficiently represent both phonetic and speaker-specific information. It has been known that the “Hey Siri” detector uses a Deep Neural Network (DNN) to convert the acoustic pattern of voice at each instant into a probability distribution over speech sounds. In the current project, the researchers trained 3 sets of models. ls. The first model includes 4 biLSTM layers of the encoder are tied for both tasks. This model is made to learn to represent both phonetic and speaker information in the final layer of an encoder. The second model includes only 3 biLSTM layers in the encoder along with separate final biLSTM layers for the voice trigger and speaker recognition branch. The third model is trained where here 2 biLSTM layers have tied weights, with 2 additional biLSTM layers for each branch. Dataset Used The evaluation process of the model is done on a huge test-set which has been internally collected for models designed for smart speakers. The data has been recorded using live sessions in realistic home environments. The training data for voice trigger detection is 5,000 hours of anonymised audio data that is manually transcribed. The researchers used a set of 3,000 different room impulse responses (RIRs) to simulate the reverberant speech which is internally collected in a wide range of houses and represents a diverse set of acoustic conditions. The training data for the speaker recognition task comprises 4.5 million utterances sampled from intentional voice assistant invocations. The training set contains 21,000 different speakers, with a minimum of 20 examples and a median of 118 examples per speaker, resulting in a training set with over 5,700 hours of audio. The final dataset contains 13 million training examples with over 11,000 hours of labelled training data. Wrapping Up It is not new that the tech giant has tweaked its virtual assistant Siri’s machine learning algorithms to overcome several challenges in voice recognition space. Besides, the advancements in automatic speech transcription and speaker recognition, the tech giant has done researches in spoken language identification (LID) technologies for improving language identification for multilingual speakers.","excerpt":"Last year, Apple witnessed several controversies regarding its speech recognition technology. To provide quality control in the company’s voice assistant Siri, Apple asked its contractors to regularly hear the confidential voice recordings in the name of the “Siri Grading Program”. However, to this matter, the company later apologised and published a statement where it announced […]","categories":["AI Features"],"tags":["Speech Analytics"],"author_name":"Ambika Choudhury","publish_date":"2020-02-08T14:00:00","publication_year":"2020","word_count":586,"keywords":["Go","machine learning","Speech Analytics","AI","neural network","programming_languages:R","programming_languages:Go","LSTM","R"],"extracted_tech_keywords":["AI","machine learning","neural network","R","Go","LSTM","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/automatic-speech-transcription-and-speaker-recognition-simultaneously-using-apple-ai\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10082857,"title":"2022 in Review: Google","content":"Stock crash, sentient AI, racial workplace biases, Pixel watch, lawsuits, and layoffs are the ingredients that headlines about Google in 2022 were made of. On the tech side? Not much! Google’s parent company, Alphabet, witnessed a crash in its stock amounting to nearly 25% in the first half of 2022. Additionally, the tech giant also joined the ‘layoff’ trend in November as reports suggested that the company is letting go of 10,000 of its underperforming employees. During the same period, Google also agreed to pay a whopping USD 391.5 million in settlement for misleading users with regard to location tracking. Considering all the developments in 2022, it would be fair to say that Google had a, more or less, uneventful year. Let’s delve in further. Google AI In terms of AI, the highlights of 2022 would definitely include OpenAI’s DALL-E2, Stability AI’s Stable Diffusion along with OpenAI’s chatbot, ChatGPT. When it comes to Google, the highlight this year is likely to be the LaMDA fiasco, where Blake Lemoine, an engineer who worked with Google’s Responsible AI team, was fired for claiming that LaMDA AI was sentient. Google, often seen as a leader in the context of AI along with the other big tech firms, has also developed its own text-t0-image AI model called Imagen. However, unlike DALL-E2 and Stable Diffusion, the model is still not accessible. Stability AI, on the contrary, has open sourced Stable Diffusion. The biggest development for Google came from DeepMind, a company owned by Alphabet. Earlier this year, the London-based AI firm predicted nearly all the protein structures known to science. DeepMind said that its AlphaFold programme expanded its open online database to include more than 200 million protein structures. Revenue In 2021, Google made around USD 257.63 billion in revenue, a significant increase from the USD 182.53 billion that they made in 2020. Much like previous years, this year too, Google continues to generate a major chunk of its revenue from advertising. In the third quarter of 2022, Google reported a revenue of around USD 69 billion and around USD 13.9 billion in profits. Revenues in Q1 and Q2 stood at USD 69.7 billion and USD 68.01 billion, respectively. In Q3, Google reported a drastic decline in its overall revenue growth from 41% a year ago to 6%. Advertising revenue from YouTube also declined 2% year-over-year to $7.07 billion. Notably, it was the first time YouTube’s ad revenue shrank on a year-over-year basis since Google started declaring YouTube’s revenue on a quarterly basis. To get a broader picture of Google’s revenue in 2022, we will have to wait a bit longer. However, with advertisers cutting their budget short, Google’s revenue is expected to take a hit in 2022. Cracks begin to appear? Besides, for the first time since its inception, Google might have a serious competitor. Senior Vice President–Google, Prabhakar Raghavan, while speaking at Fortune’s Brainstorm Tech conference, revealed that nearly 40% of young users between ages 18 and 24 in the US preferred to use ‘TikTok’ or ‘Instagram’ over ‘Google Maps’ or ‘Google Search’ when looking for a restaurant. Tiktok is also redefining social media to the extent that YouTube recently introduced ‘Shorts’—a short-form video-sharing platform similar to TikTok—to revive its user engagement. Google started out as a search engine, and today, the tech giant is facing a threat that could possibly change how one searches for information over the web. Further, OpenAI’s conversational chatbot, ChatGPT, which took the internet by storm, was also labelled by many as the ‘Google Killer’ because of its ability to answer complex queries instantaneously. Many are of the opinion that if ChatGPT is allowed to connect to the internet, it could possibly be what ultimately brings about the demise of Google Search. While Google could see these developments as mere speculation and believe that it might not have much impact on their business or profitability, it could be argued that 2022 might well have been the year cracks began to appear in Google’s armour for the first time. What to expect in 2023? Earlier this year, in an email to his employees, Alphabet CEO Sundar Pichai said that the tech giant will slow down hiring as well as investment in 2023. Pichai cited economic headwinds as the reason. Google will also retire Universal Analytics. The tech giant announced earlier this year that it will stop processing new hits on July 1, 2023. Along similar lines, Google will also discontinue its dedicated Street View app on Android from March 31, 2023, according to 9To5Google. When it comes to AI, Pichai said that 2023 will mark a “point of inflection” for the way AI is used for conversations and in search. Google announced LaMDA in 2021, which implies that they possess the technology to compete with ChatGPT. With many Googlers worried about the company being left behind in terms of AI and innovation, we could expect some notable developments in this field from Google’s end. In the revenue front, Google is expected to continue generating a major chunk from ads; however, the share of revenue from Google Cloud could be expected to increase in 2023.","excerpt":"In Q3, Google reported a drastic decline in its overall revenue growth from 41% a year ago to 6%.","categories":["Global Tech"],"tags":["Google","Google Deepmind"],"author_name":"Pritam Bordoloi","publish_date":"2022-12-20T18:00:00","publication_year":"2022","word_count":855,"keywords":["Go","ChatGPT","OpenAI","AI","AWS","RAG","GPT","Aim","Google Deepmind","analytics","Google","R"],"extracted_tech_keywords":["AI","analytics","ChatGPT","OpenAI","Aim","RAG","AWS","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/2022-in-review-google\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":58844,"title":"Top Analytics Job Openings in India At The Moment","content":"In today’s time, there is no doubt that data science jobs are amongst the best-paid ones to avail in the nation. As the domain of data science has escalated at a rapid pace in the nation, various companies from startups to industry leaders are on the lookout to find talents that will help them weigh heavier in the industry and reach new heights. In this article, we bring to you the top analytics jobs openings in the nation, at the moment. Manager-Consumer Supply Chain Analytics at Johnson & Johnson Based in Mumbai, the ideal candidate must have experience in working with two enterprise-wide projects involving data lakes and data migration. The candidate must have a master’s degree in Management or M.Tech with 12-15 years of experience in business analytics. Requirements Must ensure that the data management platform meets the performance, data quality and scalability requirements.Should have knowledge of tools such as Excel, Access, VBA and machine learning.Should possess an understanding of Azure-based tools such as Azure SQL, Blob and table storage.Experience with big data tools, ETL tools along with coding proficiency in Python.Certification of TOGAF, Tableau and Alteryx is a must. Apply here Deep Learning & Computer Vision Architect at Pratibha Analytics The position is based in Hyderabad for those with knowledge and skills in statistics, mathematics and computer science. The ideal candidate should have an experience of more than five years in computer vision solutions such as CNN and RNN. Requirements Should be proficient in producing AI and ML computer vision solutions.Should know how to integrate AI and ML hardware actuators, IoT devices.Strong knowledge of Python, Tensorflow, Tensor RT, C++, Keras, Pytorch, Caffe, and OpenCV.Must be aware of market-leading advanced analytics, machine learning and deep learning techniques, frameworks, methodologies and tools. Apply here Lead Analytics at Nykaa Founded in 2012, Nykaa is an omnichannel beauty destination with millions of customers around the globe. This job role is based in Gurgaon and for those who are highly proficient in web analytics tools, adobe analytics and tag management (DTM). Requirements Should possess experience in business requirement gathering and strategy buildingImplementation exposure across different platforms such as mobile site, mobile apps and desktopExperience in reporting and analytics (Ad-hoc\/discover, report builder, workspace and web-based interface)Proficient in data crunching, pulling insights and key marketing analysis (cohort, funnels, conversion\/traffic dip analysis, and pre and post-campaign analysis) Apply here Managing Consultant – Analytics CoE at Mastercard Mastercard is renowned for the fastest payment processing network in the world. Based in Gurgaon, the role is apt for those who would share the vision with Mastercard for data-driven consulting. Requirements Must have a graduate or masters degree.Candidates with prior experience consulting and project management preferred.Should possess strong statistical programming skills such as SAS, R, Python.Strong hands-on data analysis skills.Ability to identify strong opportunities to leverage data and services capabilities. Demonstrates a strong aptitude for structured problem solving and quantitative skills. Apply here Associate Analyst I\/II – Social Media Risk Analytics at Colgate-Palmolive Job role based in Mumbai, Colgate-Palmolive is a leading consumer product brand that has introduced several products ranging from oral care to pet nutrition. The job role requires a candidate to have more than two years of experience in social media monitoring and social listening and will be a member of the social risk monitoring team in CBS Analytics based out of Mumbai. Requirements Must have a bachelor’s degree.Should be able to monitor, analyse and validate results from on-line social monitoring tools, software based on designated and pre-defined terms.Must be able to validate, document and communicate risk items for escalation based on internal guidelines and criteria.Should be capable of collaborating with divisions and other stakeholders to implement social analysis for campaigns\/events and present insights back to inform content\/media decisions for campaign\/event activities and advise on actions to take from optimisations standpoint.Should demonstrate experience with data analysis, statistics and\/or digital monitoring. Apply here Quantitative Business Analyst, Analytics and Insights at Google Based in Haryana, the role is open to candidates with more than five years of experience in analytics role and strategy consulting with a degree in economics, statistics, mathematics, physics, computer science or engineering. The candidate will be a part of the analytics and insight team and will work to solve strategic and operational problems using the present internal data. Requirements Knowledge of data analysis tools and techniques.Should know how to work with complex SQL seriesMust know to work with large datasets to develop analytical solutions for complex business problems.Should possess excellent task management skills.Must be able to synthesise data into useful formats. Apply here Data & Analytics Analyst at RBS Based in Delhi, RBS bank requires a candidate who will manage complex data to identify business issues and opportunities. The candidate should have strong analytical skills along with an eye for detail. Requirements Should be able to validate and document business data.Interrogate, interpret and visualise large amounts of data to identify, support and challenge business opportunities.Must know how to maintain full traceability and linkage of business requirements of analytics outputs.Perform data extraction, storage, manipulation, processing and analysis.Creating and executing quality assurance at various stages of the project. Apply here Head – Analytics and Consumer Insights at Sony Pictures Network India Based in Mumbai, the candidate should have a degree in MBA, Masters or Bachelors in Engineering with an overall experience of close to 12 years and 10 years of experience in Analytics and Consumer Insights. Requirements Should track initiatives against the strategy and budget, tweak as required during the year.Know how to identify newer methodologies for conducting consumer research.Must identify suitable research agencies for the project.Sift through all the data shared and identify key insights and create an action plan for the business.Should be proficient in using Excel, working knowledge of SQL and Google Analytics. Apply here Analytics Lead at Flipkart The job role is based in Bengaluru Bachelors in Engineering, Computer Science, Math, Statistics or MBA with an experience of more than six years in a Lead Analyst role. Requirements Must identify the key drivers of these KPIs by building causal models.Must dive deep into an assessment of key drivers for KPIs by drawing relevant business hypotheses.Should know to execute deep and complex quantitative analyses that translate data into actionable insights.Should possess a strong background in statistical modelling and experience with statistical tools such as R, Python, Spark, SAS or SPSS.Must have working experience with BI tools (Tableau, Qlikview, Power BI, Cognos, BO, Pentaho, etc). Apply here Data Analyst – Advanced Analytics at Gartner Based in Gurgaon, the candidate must have a bachelor’s degree with relevant experience in business operations along with advanced SQL capabilities and more than four years of experience. Requirements The candidate will need to connect, maintain and report on and analyse data used by the digital market team.Should be able to create new and maintain existing Power BI dashboards leveraging multiple data sources.Proficiency in Python programming.Should have prior experience in web analytics tools and data analysis.Must be able to analyse results, interpret and communicate the results and findings to the business. Apply here","excerpt":"In today’s time, there is no doubt that data science jobs are amongst the best-paid ones to avail in the nation. As the domain of data science has escalated at a rapid pace in the nation, various companies from startups to industry leaders are on the lookout to find talents that will help them weigh […]","categories":["AI Hirings"],"tags":["analytics certification","career in data analytics","mba in business analytics","online masters analytics","supply chain analytics projects"],"author_name":"Rohit Chatterjee","publish_date":"2020-03-17T13:53:15","publication_year":"2020","word_count":1173,"keywords":["data science","machine learning","Keras","analytics certification","AI","TensorFlow","ML","PyTorch","mba in business analytics","computer vision","online masters analytics","deep learning","analytics","career in data analytics","supply chain analytics projects"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","data science","analytics","TensorFlow","PyTorch","Keras"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/top-analytics-job-openings-in-india-at-the-moment\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10098412,"title":"Turning Rural India into Data Mining Hubs","content":"TIME recently covered a brilliant piece on how the rural market is contributing to the development of AI systems such as ChatGPT and alike, where it not only highlighted the rising demand for training datasets in one’s own native language, but emphasised the need for more inclusive solutions. The company behind this led us to Bangalore-launched Karya, a nonprofit organisation that is working towards accelerating social mobility in the country via AI training and upskilling. Founded by Manu Chopra and Vivek Seshadri in 2021, Karya was created with the vision of making an ecosystem of ethical data usage at the same time financially and technologically empower communities. Manu and Vivek were Microsoft researchers. From 2017 to 2020 Karya was a Microsoft incubated centre, post which it has been functioning as a separate private entity. Exploring for a project by Microsoft that involved data collection in particular languages involved seeking a partnership with people from rural areas, however, the approach could not be a direct one – building trust was the main challenge. That is where Jeevitha Satheeshkumar joined them, currently the Director of Operations at Karya. With over 10 years of experience as a software trainer, she began experimenting with projects with a social implication very early on. She has been part of campaigns such as ‘Plastic Free India’ to spread awareness amongst mothers to use stainless steel feeding bottles as opposed to plastic feeding bottles. Her foray into language training happened when she started freelancing as a Tamil linguist. With the company requiring over 100 people for language transcription, she then looked into tapping the rural market. “It is difficult for people who live in rural areas to get opportunities outside their area, and with online job frauds and scams, trust becomes a problem- I wanted to change that so I started training people on data annotating or labelling.” Jeevitha Satheeshkumar From here, for over six years, Satheeshkumar became a second party vendor for Microsoft, Google and other companies, and expanded the training to over 5000 people who work with multiple languages. She eventually built her own company, and gave opportunities to rural people, homemakers and even those with physical disabilities. Understanding the cyclical nature of demand for this service, she decided to build it as a product offering. “In India, only two or three companies have their own transcription tool\/ AI tool to do this kind of work.” The tool named ‘Labely’ was built to perform functions such as transcription and various types of annotations. Multiple modules such as manager, proofreader, transcriber were built onto the platform. “Any company that wishes to create a data set can use this tool.” “Having a shared vision of creating the same kind of social impact, creating opportunities for rural people and helping communities looking for a job or some kind of learning, we joined hands,” said Satheeshkumar. “Last year, we successfully completed our first pilot project with IIT Madras” Challenges Galore “It is not possible to go directly to them and offer them a job, as there is no trust element built between the parties. So, we connected with local NGOs, as they are aware of the right set of people who would benefit from this. They started in a remote area in Rajasthan and gave data collection work in Rajasthani dialect,” said Sathheshkumar. Karya has completed over 30M+ digital tasks. “Our goal is to reach 100,000 rural Indians by the end of this fiscal year, 1.5 million rural Indians by next fiscal year and 100 million rural Indians by 2030. Fundamentally, Karya’s goal is to use technology to accelerate social mobility in rural India. It currently takes an average low-income Indian over 7 generations to make $1500 in savings and a Karya worker can make the same amount in less than a year. We think of Karya not as a job, but as societal wealth distribution,” said Chopra, in an exclusive interview with AIM. Manu Chopra Designed for Rural India The Karya application has been built such that anyone can use it with ease. The app is simple with an easy user interface, where a person can read and record the text shown in the app, with an option to re-record as well. A basic Android phone is required for using the app. However, Karya is working towards catering to low end phone models too, owing to rural areas where the target group might not have access to an android phone but a lower version too. Earlier, voice recordings were done on phones and shared via Google Drive, but with the Karya app, it can be directly recorded. Once the recording is submitted, it moves to a ‘proofreader validator.’ First there is an automation process to check the quality of audio- to check for missing speech or non technical inputs in audio. Post validation, it then moves to ‘transcriber’. The process involves segmentation(classification) and transcription, where labeling and transcribing exact word-by-word for the recorded speech happens. It then reaches the final process of validation and the file is converted into either a txt or JSON format depending on their customer’s requirement. Rural Supersedes Urban People in rural areas show better prowess when it comes to learning and implementing new skills. “For language recordings, people who live in rural areas would be able to do it better than their urban counterparts. Owing to having studied and schooled in their native language, they know the language better than others.” Satheeshkumar has also learned that people who live in villages have better grasping power than the ones from cities. “The people there are focussed and always ready to do anything that will help them. As opposed to people in urban areas who can get distracted, the people here are fully focussed during training and listen attentively without batting an eyelid. Though such opportunities are high-hanging fruits for them, we want to bring it to their doorsteps.” said Satheeshkumar. Socially Fueling the Indian AI Ecosystem In addition to working with big tech companies and foundations such as Microsoft, Bill and Melinda Gates Foundation, Karya has also partnered with universities such as MIT, Stanford, IIT Madras and IIT Bombay. Currently, they are working with AI4Bharat, which builds open-source AI for Indian languages. Touted as a promising player for revolutionising the LLM space in India, AI4Bharat has over 200 translators over 22 Indian languages. Karya generates and provides datasets for this Indic platform. With a simple linear model of functioning, the workers who provide their service are rewarded justly- from a minimum of $5 per hour, it can go as high as $30\/hr, depending on skill sets. “Other companies who outsource such projects involve multiple vendors with multiple stages of transfer, ultimately leading to the last man receiving $1 or $2. However, in Karya, the structure is simple with no middle vendors. It goes : customer, Karya then workers. If a person is able to provide transcription worth $30 daily, in a month they can receive up to INR 50000. The payment process is automated as well. After validation of the tasks completed by the workers, within 10 to 15 days payment is done through UPI or bank account. Identifying the Workers As there will never be a shortage of people for these tasks, Karya ensures that it reaches the right set of people. Initially, when the app was launched, the first 100 downloads were done by men from urban areas who can do high level jobs, which defeats the purpose. After which, any form of task requirements are approached via NGOs in relevant areas who help connect with their right audience. There are over 180 NGO partners for Karya. “Currently we are working on two projects that require a minimum of 800 to 1000 people in each district. India has 766 districts and employing a 1000 for each would require our NGO partners for the right identification.” Speaking about future plans, Manu said, “To scale our operations, we want to capture a bigger section of the global AI training data market. To do this, we have to raise more awareness about our work and work with many more big tech companies.” said Manu","excerpt":"“It currently takes an average low-income Indian over 7 generations to make USD 1500 in savings and a Karya worker can make the same amount in less than a year.”","categories":["AI Trends"],"tags":["AI4Bharat","Datasets","Google","Karya","language","Microsoft","MIT","Stanford"],"author_name":"Vandana Nair","publish_date":"2023-08-10T15:00:00","publication_year":"2023","word_count":1358,"keywords":["Go","Datasets","ChatGPT","Karya","Rust","AI","language","MIT","Git","RAG","GPT","Ray","Aim","Google","Stanford","AI4Bharat","R","Microsoft"],"extracted_tech_keywords":["AI","ChatGPT","Aim","Ray","RAG","R","Go","Rust","Git","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/turning-rural-india-into-data-mining-hubs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10162825,"title":"Zomato is Hiring Business and Product Leaders with a ‘Second Brain’","content":"Zomato CEO Deepinder Goyal has once again ignited a hiring debate with his latest job posting. This time, he is seeking business and product leaders who have embraced AI as their “second brain”. Goyal took to X and announced, “I am looking to work with business and product leaders who have already started using AI as their second brain. If you are the one, please write to me at d@zomato.com. Please include the phrase ‘I have a second brain’ in the subject line.” This unusual hiring criterion has stirred conversations about AI’s growing role in leadership and decision-making. This comes just months after Goyal’s previous unconventional job posting – a chief of staff position that required candidates to pay a ₹20 lakh ($23,700) “fee” instead of receiving a salary. Back in November 2024, Goyal’s chief of staff job post sparked intense debate, with critics questioning the fairness of the selection process. However, the Zomato CEO later clarified that the fee was merely a filter to identify high-calibre candidates who saw value in a fast-track career. The strategy worked, and Goyal received 18,000 applications. Now, with AI at the centre of his latest job hunt, Goyal seems to be doubling down on his vision for an AI-driven leadership team. AI is the Future of Work Goyal’s emphasis on AI skills aligns with broader global trends. The World Economic Forum’s Future of Jobs Report 2025 predicts that 39% of core job skills will transform by 2030, with AI, big data, and cybersecurity leading the charge. However, tech expertise is not the only skill that’s in demand; creative thinking, resilience, and leadership are also becoming crucial. Companies worldwide are investing in reskilling and upskilling to keep pace with rapid advancements. As AI becomes an indispensable tool in business strategy, Goyal’s hiring approach signals a shift. Leaders who don’t integrate AI into their decision-making might soon find themselves left behind.","excerpt":"This unusual hiring criterion has stirred conversations about AI’s growing role in leadership and decision-making.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Hiring","Zomato"],"author_name":"Vidyashree Srinivas","publish_date":"2025-02-04T12:13:16","publication_year":"2025","word_count":316,"keywords":["big data","Go","API","programming_languages:R","AI","data_tools:Spark","Hiring","programming_languages:Go","Zomato","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","Go","API","big data","programming_languages:R","programming_languages:Go","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/zomato-is-hiring-business-and-product-leaders-with-a-second-brain\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10054278,"title":"State Of Chip Shortage In 2021: A Quick Revisit","content":"The global chip shortage of 2020–2021 is affecting over 169 sectors and consumer lines, including cars, graphics cards, video game consoles, and so on. The snowball effect of the COVID-19 pandemic happens to be the biggest reason amongst many, creating the global chip problem. Other possible causes include the China–United States trade war and Taiwan’s 2021 drought. In addition, automobile manufacturers worldwide anticipated a decline in demand at the start of the pandemic. As a result, these companies chose to order fewer semiconductor chips to reduce their inventory costs during the shutdown period. On the other side, people began upgrading their computers, laptops, phones, and other electronic equipment amid the pandemic-induced lockdown. During this time, Sony introduced their brand-new platform, the PlayStation 5, to rave reviews and millions of pre-orders. Similarly, demand for graphics cards increased due to their gaming use, graphics production, and crypto mining. While demand for semiconductor chips decreased temporarily in the vehicle industry, it witnessed a rise in all other industries. Moreover, the chip shortage created difficulties for automotive manufacturers when they could not resume normal manufacturing due to a sudden shortage of semiconductor chips. Automobile manufacturers, in particular, cancelled orders, while chip manufacturers shifted their focus to consumer products in an attempt to meet the pandemic’s growing demand. After retooling their operations to produce chips for consumer items rather than automobiles, a shortage of automobile chips ensued. Why aren’t there more semiconductor manufacturers? Semiconductor manufacturers can supply any product’s increased demand if manufacturing capacity is sufficient. However, manufacturing semiconductor chips is a difficult, time-consuming, and expensive process that barriers players. Currently, the semiconductor industry is dominated by three manufacturers: Taiwan Semiconductor Manufacturing Company (TSMC), Intel, and Samsung. Supply problems were caused by a small number of manufacturers and an unexpected surge in demand. In addition, due to the shortage of graphics cards, manufacturers increased their pricing, which increased the prices of PCs and laptops. What is being done to ease the global chip shortage? With demand for semiconductors expected to grow as more industries embrace digital transformation, chip manufacturers and governments collaborate to expand supply networks’ capacity. For example, TSMC is investing $100 million in extra capacity over the next three years, while Samsung, SK Hynix, and the South Korean government have committed to investing $451 billion in chip manufacturing capacity and incentives. When will the global chip shortage end? “Although the supply of semiconductors was supposed to recover by the end of 2021, analysts believe that the worldwide chip deficit could stretch throughout the next year and maybe into 2023. The shortage is because current capacity expenditures will take fruit,” said Malcolm Penn, CEO of industry researcher Future Horizons. Even if the current global chip shortfall is resolved, it is possible that other supply issues will arise as the demand for electronics continues to grow. “The capacity that chip makers are building today will suffice for the next few years, and as these things come online, there will be an excess of capacity,” Gartner analyst Alan Priestley explained. Intel CEO Pat Gelsinger stated that he anticipates the chip shortage to worsen in the second half of 2021 and that it will take a year or two for supplies to recover to normal. On the other hand, Nvidia CEO Jensen Huang stated that he anticipates the shortfall to last far into 2022, while AMD CEO Lisa Su stated that the deficit will improve in the second half of 2022. However, supply would remain tight until then. IBM CEO Arvind Krishna stated that “More than likely, the chip shortage will continue until 2023 or 2024 at the earliest.”","excerpt":"Semiconductors are used to power a wide variety of devices, from your phone to your car. A key supply chain issue has been addressed here.","categories":["IT Services"],"tags":["china chip technology","global chip shortage","semiconductor industry","semiconductors"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-11-26T17:00:00","publication_year":"2021","word_count":602,"keywords":["Go","programming_languages:R","semiconductors","AI","global chip shortage","digital transformation","semiconductor industry","Git","programming_languages:Go","china chip technology","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","digital transformation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/state-of-chip-shortage-in-2021\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10076527,"title":"Is NPCI Building Single-token Digital Identity?","content":"In November 2021, the National Payment Corporation of India (NPCI) discussed the future of digital identities in an article. Awarded ‘article of the month’, the story had Shivani Bisht, graduate engineer trainee, CDO office, brushing through the future of identity on the blockchain. As per the author, while robust and cryptographically safe blockchain programming is a necessity, before widespread use, user-friendly and simple interfaces, regulatory monitoring, and consensus standards must be established. The article ended with a quote from Aravind Srimoolanathan, senior research analyst, ICT at Frost & Sullivan, “By 2030, advancements in blockchain technology will lead to its integration with biometrics, resulting in the establishment of a single-token digital identity for individuals.” This sheds light on NPCI lofty intentions for blockchain technology, even if the central government may be hesitant. The intent is also reflected in NPCI’s continual recruiting of blockchain developers. The organisation has been aggressively seeking blockchain developers across several platforms for a few weeks. Previous experiments with Blockchain: Vajra platform In 2020, NPCI announced the Vajra platform, built on distributed ledger technology (DLT). It planned to utilise a permission model to ensure that only authorised parties access the network. Payment companies could apply to join the network. Once accepted, the idea of implementing the platform via an application programming interface (API) given by NPCI was discussed. The implementation of DLT and the ensuing audit trail would help make tracking and resolving payment disputes quicker, which often takes three days. Furthermore, the possibility of the Vajra platform being utilised for Aadhaar authentication by the Unique Identification Authority of India was also envisioned. However, barely a month later, the whitepaper was deleted from the NPCI website, and not much is known about the developments around it now. But it must be mentioned that before the announcement of Vajra, an article titled ‘The rising significance of blockchain in the banking sector’ was published on the NPCI website that ironically discussed the DLT-backed banking system. Thus, can the aforementioned NPCI article about the future of digital identity and the recent hiring spree of NPCI for blockchain developers be linked together? Future ahead: Single-token digital identity? While integrating blockchain and biometrics to create a single-token digital identity might seem like a great idea at the moment, it’s still not ready for implementation in a large state like India. One of the implementation limitations will be its processing capabilities. Ethereum, for example, can only handle about a dozen transactions per second, which may not be adequate for nationwide use. Furthermore, there is a minimum confirmation time before the transaction is considered correctly put to the blockchain. This duration can vary among blockchains, ranging from tens of seconds to minutes, limiting its applicability to biometric systems. Another constraint might be scalability. This has been one of the fundamental disadvantages of the technology from its inception, as all nodes of the blockchain network must potentially hold all blocks of the blockchain network. The size of the public blockchains (Bitcoin and Ethereum) is now over 200GB, and it is rapidly expanding. This can be an issue in some application contexts, such as the Internet of Things (IoT). While the issue of minimum confirmation time might be resolved by the likes of Ethereum-based layer two scaling solution Polygon, which claims to have achieved 65000+ transactions per second, however, in doing this, the security of Polygon is sometimes questioned. And this leads us to the infamous Blockchain Trilemma, which refers to the inability of decentralised networks to provide all three benefits at the same time: decentralisation, security and scalability. Instead, it is believed that it can deliver only two of the three benefits at the same time. Overall, it is up to the NPCI and the government to decide where they want to take it. While the introduction of blockchain technology into the banking industry was a significant step forward, the swift removal of the white paper from the website reveals there is still a debate over the direction of its use. Blockchain: Where does the government use it? On the surface, it might seem like the government is not doing much in this direction. However, some significant steps are indeed being taken. For instance, the Ministry of Electronics and Information technology (MeitY) has set up a Centre of Exchange in blockchain technology (CoE). According to the website, the National Informatics Centre (NIC) and CoE team will work with professionals worldwide to develop and execute novel blockchain solutions from proof of concept to production. So far, the department has four products under the hood, namely: Certificate Chain, Document Chain, Property Chain, and Logistic Chain. Certificate Chain is primarily intended to combat fake documents, time taking document verification, and service delivery delays. One of the first adopters of this tech, the Central Board of Secondary Education (CBSE), maintains its academic records on the blockchain. Read more here: https:\/\/cbse.certchain.nic.in\/aboutcc Similarly, Document Chain offers issuing authorities and consuming institutions a uniform mechanism for storing and retrieving any government document, such as caste and income certificates, ration cards, driving licences, birth and death certificates, and so on. Caste and income certificates issued by the revenue department of Karnataka, for example, are recorded using the same technology. Likewise, a blockchain-powered property management system allows for the availability of a common property ledger, allowing a single source of truth. The Logistics Chain is the online Supply Chain Management System for any particular industry.","excerpt":"NPCI has been seeking blockchain developers for a few weeks. But why would it need blockchain developers in the first place? In this article, we speculate whether NPCI is developing a Single-token Digital Identity","categories":["IT Services"],"tags":["npci","polygon"],"author_name":"Lokesh Choudhary","publish_date":"2022-10-07T10:00:00","publication_year":"2022","word_count":900,"keywords":["Go","API","npci","AI","programming_languages:R","Scala","Git","polygon","Aim","programming_languages:Scala","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Scala","Git","API","GAN","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/it-services\/is-npci-building-single-token-digital-identity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10076247,"title":"IISc Researchers Develop an Algorithm to Detect Seizures and Occurrence of Epilepsy","content":"A new algorithm can now help identify the type and occurrence of epilepsy, thanks to the researchers of Indian Institute of Science (IISc) and AIIMS Rishikesh. According to the press release, this new breakthrough is set to play a vital role for an efficient and automated screening and diagnosis of the disease. Epilepsy is a neurological disorder where a sudden burst of electric signals is released by the brain in a short period of time. This may cause fits, seizures, or may even result in death. The disease is classified as ‘focal’ or ‘generalised’—based on the origin of the erratic signals produced in the brain. Focal epilepsy occurs when the signals are confined to one specific region whereas generalised epilepsy is confined to random locations in the brain. Neurophysiologists must manually inspect EEGs (electroencephalograms), which is used to capture erratic signals and identify whether a patient is epileptic. Source: Detecting seizures and interpreting EEGs, the direct algorithmic way, IISc Assistant Professor, Department of Electronic Systems Engineering (DESE) and corresponding author Hardik J Pandya claims that visual inspection of EEGs could become burdensome or exhausting after prolonged periods, and may also lead to errors. “The research aims to differentiate EEG of normal subjects from epileptic EEGs. Additionally, the developed algorithm attempts to identify the types of seizures. Our work is to help the neurologists make an efficient and quick automated screening and diagnosis,” said the professor. According to the researchers, in their study, the team developed a novel algorithm that sifts through EEG data and identifies signatures of epilepsy from the electrical signal patterns. After the initial training, the algorithm was used to detect whether the subject could have the disease—based on the patterns of their respective analyses—with a high degree of accuracy. Rathin K Joshi, a PhD candidate at DESE says, “We hope to refine this further by testing on more data to consider more variabilities of human EEGs until we reach the point where this becomes completely translational and robust.” At present, a patent has been filed for the work, along with physicians at AIIMS Rishikesh testing the developed algorithm for its reliability.","excerpt":"The newly developed algorithm is set to play a vital role in assisting neurologists make efficient automated screening and diagnosis of the disease.","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-10-03T11:39:28","publication_year":"2022","word_count":354,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iisc-researchers-develop-an-algorithm-to-detect-seizures-and-occurrence-of-epilepsy\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10112214,"title":"Google&#8217;s &#8216;localllm&#8217; Lets You Create GenAI Apps Without GPUs","content":"Google has introduced ‘localllm’ which allows developers to develop next-gen AI apps on local CPUs. ‘localllm’ is a set of tools and libraries that provide easy access to quantised models from HuggingFace through a command-line utility. This solution eliminates the need for GPUs, offering a seamless and efficient solution for application development. ‘localllm’ revolves around the utilisation of quantised models optimised for local devices with limited computational resources. These models, hosted on Hugging Face and tailored for compatibility with the quantisation method, enable smooth operation on Cloud Workstations, eliminating the dependency on GPUs. Quantised models offer improved performance by employing lower-precision data types, reducing memory footprint, and enabling faster inference. The combination of quantized models with Cloud Workstations enhances flexibility, scalability, and cost-effectiveness. The approach aims to overcome the limitations of relying on remote servers or cloud-based GPU instances, addressing concerns related to latency, security, and dependency on third-party services. Key features and benefits include GPU-free LLM execution, enhanced productivity, cost efficiency through reduced infrastructure costs, improved data security by running LLMs locally, and seamless integration with various Google Cloud services. To get started with the localllm, visit the GitHub repository at https:\/\/github.com\/googlecloudplatform\/localllm. Google recently partnered with Hugging Face to enable companies to build their own AI with the latest open models from Hugging Face and the latest cloud and hardware features from Google Cloud.","excerpt":"‘localllm’ revolves around the utilisation of quantised models optimised for local devices with limited computational resources.","categories":["AI News"],"tags":["Gemini","LLMs"],"author_name":"Siddharth Jindal","publish_date":"2024-02-08T10:01:59","publication_year":"2024","word_count":225,"keywords":["Go","Gemini","Hugging Face","AI","LLMs","ML","Scala","Git","Aim","ViT","GitHub","R"],"extracted_tech_keywords":["AI","ML","Aim","Hugging Face","R","Go","Scala","Git","GitHub","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/googles-localllm-lets-you-create-genai-apps-without-gpus\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002457,"title":"How This Pune-based Agritech Startup Reduces Wastage In Farm Produce Using AI","content":"Though the agriculture industry in India is one of the largest, it is often unorganised especially in the post-harvesting supply chain where it witnesses 30-40% food losses due to the major gap between supply and demand, and inefficient logistics. These losses result in increased prices to the customers with farmers getting little share of the price paid by the consumers. Another major challenge in the value chain is the time it takes to reach the market, which is usually 24-30 hours, resulting in wastage in many cases. Karan Hon, who comes from the farming family along with Puneet Sethi recognised these challenges very well and after a two-decades-long journey in the corporate world, decided to set up Farmpal, a Pune-based startup that is on a mission to organise the post-harvest supply chain so that farmers can have better access to alternate markets. It also promises to provide fair prices for their produce and reduce inefficiencies across the supply chain so that all stakeholders are benefitted. To achieve their goals, Farmpal is extensively using artificial intelligence and other technologies to connect farmers directly with end consumers such as local kiranas, supermarkets, sabjiwala to local societies and more. Analytics India Magazine got in touch with Puneet Sethi to understand how AI comes into the picture at Farmpal which provides a mobile app, connecting farmers and customers at the front-end. “This is backed by a powerful ERP solution at the backend that helps us streamline inventory management, order cycles, supply-demand forecasts and analytics for all other functional areas. One of our key goals is to reduce wastage across the supply chain. For that, a close match of supply-demand is essential which is where AI plays a key role,” shares Sethi. Farmpal’s AI-Based Solution Has Been Able To Predict The Demand With 95% Accuracy One of the ways to deal with wastage in the agriculture industry is to predict the consumer demand and meet the supply accordingly so that there is minimum wastage across the value chain. The AI comes into the picture to predict the demand, where Farmpal has designed an artificial intelligence solution to predict demand at 95% accuracy and keeping the wastage below 5%. As Sethi shares, they are leveraging Oracle ERP historical data of supply-demand and placing a layer of AI over ERP to curate historical data. “AI makes sense out of available data to understand variation in demand considering peak and off-peak days in a week for each customer segment and seasonal variation for each SKU considering festive seasons,” he said. AI modules have helped them understand the weekly demand for more than 50 SKUs where they are able to inform the demand to farmers a week before, and farmers can plan on harvesting based on this demand to avoid over or under supply, resulting in minimum wastage. The fact that Farmpal has historical data of over two years and an extensive study on the post-harvest trends and patterns both in terms of production and consumption, helping them to accurately design AI solutions. Farmpal leverages a lot of this data and publicly available data to predict and forecast demand and supply patterns. “As our volumes grow and historical data grows with it, we are able to do a deeper dive into the trends and patterns such as production and demand for specific seasons, months and even specific days of the week. All this data is fed to the AI engine and algos, along with certain other conditions to forecast the demand and then enable matching supply through an integrated farmer database where we store farmer information such as profile, land holding, type of produce, harvest cycle, capacity etc.,” explained Sethi. The tech stack at Farmpal is a homegrown software coupled with Oracle ERP to enable their tech functions, including AI. “The Oracle suite is pretty comprehensive and provides a host of AI capabilities, specifically AI for the supply chain functions,” shared Sethi. Other Areas Where Farmpal Is Using AI Currently, the startup is using AI for the supply-demand cycle for different times of the year, week etc., they are also exploring AI for more efficient route planning and route optimisation to better manage the logistics. “In future, we will use AI for customer data to model, and then leverage information about purchase patterns, including their basket size allocation, so our pricing is more in sync and to enable the use of customer promotions to boost sales and also build customer loyalty,” he added. Farmpal solutions are so far deployed for their own internal working and tech enablement to support operations across Maharashtra. Challenges In Developing AI Algorithms For Agriculture Industry While the solutions seem promising, the Farmpal team has to face challenges working in the agriculture sector. The primary challenge, as Sethi shares, is the lack of historical data that is very specific to the post-harvest cycle, in terms of the supply chain and the supply-demand forecasts. “Additionally, the Agri supply chain still works primarily on old antiquated mechanisms where both farmers and high volume purchasers go to the local APMC or mandis. So the initial challenge was to have available meaningful data for AI algos,” he said. Secondly, especially with fresh produce, the perishability factor plays an important role that has to be factored in. Ensuring Direct Supply To Societies In The Covid-19 Times While Covid-19 caused disruption in Farmpal’s ground operations and they scaled down due to the lockdown and non-availability of manpower and logistics, being essential services, they were able to get back necessary permissions. “Keeping the challenges in mind, we have narrowed down our focus to high customer density areas both to reduce manpower needs in our collection and distribution centres as well as to optimise transportation,” he said. Having said that, they have piloted a new segment during these Covid-19 times, that is, ensuring direct supply to societies or a collection of customers in a local area through a “business owner” model. Roadmap Farmpal continues to focus on increasing the footprint both on the supply side, in terms of farmer tie-ups across Maharashtra and in other States; and on the demand side by retailer customer acquisition in Pune, and then Navi Mumbai, Mumbai and cities in Gujarat. The startup is also looking at increasing supplies to supermarkets in other States. They also plan to improve upon the use of machine learning and artificial intelligence to better predict the demand and map the supply to continue ensuring minimum wastage. He also added that on the operations side, they plan to use IoT, RFID technologies, route planning solutions, advanced analytics in the coming future. “On the farmer front, through our platform, as we onboard and connect with farmers across India, we want to be able to better assist farmers by providing better access to the Agri ecosystem,” said Sethi on a concluding note.","excerpt":"Though the agriculture industry in India is one of the largest, it is often unorganised especially in the post-harvesting supply chain where it witnesses 30-40% food losses due to the major gap between supply and demand, and inefficient logistics. These losses result in increased prices to the customers with farmers getting little share of the […]","categories":["AI Startups"],"tags":["AI Startups"],"author_name":"Srishti Deoras","publish_date":"2020-07-14T18:00:00","publication_year":"2020","word_count":1137,"keywords":["Go","machine learning","artificial intelligence","AI","ML","RAG","analytics","disruption","GAN","R","AI Startups"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","R","Go","GAN","disruption"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-pune-based-agritech-startup-reduces-wastage-in-farm-produce-using-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169003,"title":"AWS Records $29.3 Billion Cloud Revenue in Q1 2025, Beats Microsoft and Google","content":"Amazon Web Services (AWS), the cloud computing and services subsidiary of e-commerce giant Amazon, reported on Thursday that it earned a revenue of $29.3 billion in Q1 2025. This marks a 17% year-over-year increase. However, this growth is slightly lower compared to the 18.9% year-over-year increase recorded in Q4 2024. “Before this generation of AI, we thought AWS had the chance to ultimately be a multi hundred billion dollar revenue run rate business. We now think it could be even larger,” said Andy Jassy, CEO of Amazon in the earnings call. Recently, Alphabet, in the Q1 2025 earnings, stated that Google Cloud revenues represented a 28% year-over-year growth to $12.3 billion. On the other hand, Microsoft reported Q1 2025 revenue from the Intelligent Cloud segment was $26.8 billion, a 21% increase. This included a 22% rise in server products and cloud services revenue, led by a 33% growth in Azure and other cloud services. In Q1 2025, AWS expanded its AI portfolio with major releases under the Amazon Nova family. This includes Nova Sonic, a speech-to-speech foundation model for building human-like voice agents, Nova Act SDK, designed to enable AI agents to take browser-based actions, and Nova Premier, a multimodal model for complex tasks like coding and video understanding. The company also added new third-party foundation models to Bedrock, including Claude 3.7 Sonnet from Anthropic, DeepSeek’s R1, Meta’s Llama 4, and Mistral’s Pixtral Large. Recently, it announced an update to Amazon Q Developer, which now supports languages such as Chinese, Hindi, Spanish, French, Korean, and Portuguese, enabling natural conversations in multiple development environments.","excerpt":"Revenue grows 17% year-on-year, slightly lower compared to 18.9% in Q4 2024.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AWS"],"author_name":"Supreeth Koundinya","publish_date":"2025-05-02T10:28:42","publication_year":"2025","word_count":263,"keywords":["Anthropic","Amazon Nova","AWS","AI","cloud computing","Llama 4","R","Pixtral Large","foundation models","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","foundation models","Anthropic","Llama 4","Amazon Nova","Pixtral Large","cloud computing","AWS","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-records-29-3-billion-cloud-revenue-in-q1-2025-beats-microsoft-and-google\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101131,"title":"Google Assistant Gets a Generative AI Upgrade, Integrates Bard","content":"After 7 years of launching Google Assistant, the company is giving the tool AI superpowers. Assistant for Bard debuted last evening during the Made for Google event to provide personalised assistance for users. Similar to its web counterpart, the chatbot will integrate the company’s existing apps like Docs and Gmail via the recently introduced extensions for Bard. The Mountain-view based company states the latest Bard-powered tool is capable of helping users from planning a trip to picking out details from their inbox. For the content creators, the assistant can perform several tasks like generating relevant captions and trending hashtags to match the image. Amid the ongoing dialogues about the company’s chameleon-like privacy policies, Google noted that the latest tool will let the users choose their individual privacy settings. Much like the rest of its AI-powered tools, Google has only let out a glimpse of Assistant for Bard without making it public. Google will make the tool available for early testers “soon” before it arrives on Android and iOS. LLM integration is the magic ingredient missing in voice assistants that will take it from a somewhat helpful assistant to something that will more resemble the sci-fi concept familiar with. By adding natural language understanding in these models asking complex, multi step queries, contextual conversations from previous discussions, even novel instructions it may not understand at first but can be explained and then remembered later can be done. As per default the chatbot stores every interaction you have for 18 months. Apart from the prompts, Bard stores the users approximate location, IP address, and any physical addresses connected to the Google account for work or home. While these settings are activated, any conversation with Bard could be selected for human review. While some Google fans are excited about the early “Jarvis”-style experiment, concerns have been raised as every conversation with Bard is tracked, logged and used to train the AI. On a lighter note, some made fun of the branding saying ‘Assistant with Bard sounds like “Assistant to the Regional Manager” from Office.","excerpt":"After 7 years of launching Google Assistant, the company is giving the tool AI superpowers","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-10-05T12:54:44","publication_year":"2023","word_count":341,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","llm_models:Bard","R"],"extracted_tech_keywords":["AI","R","Go","RPA","llm_models:Bard","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-assistant-gets-a-generative-ai-upgrade-integrates-bard\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":7519,"title":"Indian video analytics start-up Vidooly gets Rs 6.4 crore from Bessemer Venture Partners","content":"Online video analytics start-up Vidooly has raised Rs 6.4 crore from venture capital firm Bessemer Venture Partners, which was among the first investors in LinkedIn and Skype. Founded by Ajay Mishra in 2014 along with friends Subrat Kar and Nishant Radia, Vidooly will use the money on product development and to expand its team. “We are excited to partner with them early in their journey of creating a video analytics product,” said Aakash Goel, Vice-President for Bessemer in India. “This will help further expand the video ecosystem by giving intelligent and actionable insights to creators, multi-channel networks and brands.” Noida-based Vidooly helps content creators, brands and multi channel networks to build an audience base on YouTube and earn more revenues. Its technology enables customers to track their competitors and offer features such as video tag suggestion, best time to upload and search rank analysis. “Content creators are putting a lot of efforts to create good quality content, but the problem is that they find it difficult in targeting the right kind of audience to watch their content. This is where we step in,” said Vidooly chief executive Subrat Kar. Since it’s beta launch in November 2014, Vidooly has added over 2000 YouTube channels like Bollywood Hungama, beauty portal Glamrs, India Food Network and Appu Series, a channel for kids. Nearly 120,000 videos are uploaded every month through Vidooly’s network. Its big data engine analyses over 500 million monthly videos through its platform. Vidooly is Bessemer’s second investment in the video platform space globally after Vidyard, that helps firms use online video to boost sales. In January, Canada-based Vidyard raised $18 million (Rs 115 crore) at a valuation close to $100 million (Rs 641 crore) in a round led by Bessemer. Experts say with rapidly growing smartphone and internet penetration, India has become a large market for video consumption. Internet video traffic in India is expected to grow nine-fold between 2013 and 2018, according to Cisco’s Visual Networking Index. Internet video traffic will reach 2.2 exabytes per month in 2018 or 72% of all internet traffic in India. “When we met Vidooly’s team we knew that they are onto something that is a big pain point for video creators, publishers and multi-channel networks,” said Abhishek Gupta, head of TLabs, an accelerator that is incubating Vidooly.","excerpt":"Online video analytics start-up Vidooly has raised Rs 6.4 crore from venture capital firm Bessemer Venture Partners, which was among the first investors in LinkedIn and Skype. Founded by Ajay Mishra in 2014 along with friends Subrat Kar and Nishant Radia, Vidooly will use the money on product development and to expand its team. “We […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2015-06-04T09:02:31","publication_year":"2015","word_count":384,"keywords":["big data","Go","API","programming_languages:R","AI","venture capital","programming_languages:Go","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","API","big data","venture capital","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-video-analytics-start-up-vidooly-gets-rs-6-4-crore-from-bessemer-venture-partners\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10122061,"title":"It&#8217;s High Time Data Engineers Expanded Left and Right: Pavan Nanjundaiah","content":"A decade ago, the position of a data engineer was essentially nonexistent.  Slowly and gradually, the role changed as the field matured. But with the advent of new generative AI technologies, many people fear their jobs are in jeopardy. Already in the early months of 2024, GenAI is beginning to upend the way data teams think about ingesting, transforming, and surfacing data to consumers. Tasks that were once fundamental to data engineering are now being accomplished by AI—usually faster and sometimes with a higher degree of accuracy. As familiar workflows evolve, it naturally begs the question: will GenAI replace data engineers? Thus, if a data engineer has to secure his position, he has to “Go All In”, said Pavan Najundalah, Head of Trendence Studio. Speaking at AIM’ Media House’s Data Engineering Summit 2024, Pavan Nanjundaiah, VP – Studio Innovations at Tredence Inc., said today’s data engineer needs a beginner’s mindset to adapt to the growing change. “2023 was the year of copilots from Github to Fabric and Google. 2024 is the year of agents. Which means AI is after your job. The question you should be asking is, will I become obsolete? We saw what happened to DEVIN. The world will change whether you like it or not,” Nanjundaiah said. “Today’s data engineers should expand left and right. Focus on your core foundational skills because they are not going to change. You don’t need to handcraft your code anymore. We all know that augmented coding is the future,” he added. Why do we need a change? Ten years ago, businesses relied on on-premise infrastructure for data storage. At this time, data engineers were more concerned with fine-tuning their machine configuration than with generating business value. The cloud companies appeared, promising to offer services they would handle on your behalf. You can then concentrate on your business’s needs. This has changed the game. So, Nanjundaiah asks an important question: If half your job is going to be done through a copilot, what will you do? “Understand what is the business problem you’re trying to solve. My recommendation to data engineers is to shift left, which is the person who gave you the requirement and probably understands the business a little better. By doing this, you’re now zooming out of a core data engineer’s profile and understanding the business and the longer impact you can create,” Nanjundaiah added. “That’s not it. Start looking at the right side to see if a data scientist is building a model based on the data you have staged, and figure out how you can start expanding,” he added. Division Into Schools Of Thoughts Nanjundaiah divided today’s GenAI into two schools of thought: one that sees everything as “rainbow and sunshine”,” while the other is “gloomy and fears a thunderstorm”.” “We have seen rapid changes in AI in such a short time that forces everyone to change. I believe today’s Chief Experience Officer (CXO) should “Bet and Check” while Data Engineering Practitioners should “bet big but in pockets”. Although humans losing jobs to robots is a lovely story, it is far from the truth for data engineers. AIM research tells us that data engineers continue to be in high demand. Senior developers working in generative AI draw over INR 1 crore per annum, while an entrant’s salary could easily be around INR 18 lakh per annum, much higher than India’s median income. Data engineers are needed to create and manage AI applications. Data engineers are increasingly responsible for how generative AI is integrated into the business, just as they develop and maintain the infrastructure supporting the data stack. AI infrastructure is created and maintained using the advanced data engineering abilities we discussed, including abstract thought, business comprehension, and contextual creation. Furthermore, incorrect data can occasionally occur even with the most advanced AI. Things malfunction. Shortly, we don’t see an AI engaging in much self-reflection, unlike a human, who can recognise and fix mistakes. So, when things go wrong, someone needs to be there babysitting the AI to catch it—a “human-in-the-loop,” if you will.","excerpt":"As familiar workflows evolve, it naturally begs the question: Will GenAI replace data engineers?","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","AI Tool","Data Science","Data Scientist"],"author_name":"Anshul Vipat","publish_date":"2024-05-30T12:51:38","publication_year":"2024","word_count":676,"keywords":["Go","GenAI","AI","Git","RAG","Aim","generative AI","GitHub","copilots","Data Science","Data Scientist","R","AI (Artificial Intelligence)","AI Tool"],"extracted_tech_keywords":["AI","generative AI","GenAI","Aim","RAG","copilots","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/its-high-time-data-engineers-expanded-left-and-right-pavan-nanjundaiah\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10112793,"title":"6 Courses for Designers &amp; Illustrators to Learn AI Art","content":"Today, AI tools have invaded practically every field, including art and design. Learning to use these tools has become more than necessary since they open up unknown possibilities for creative expression. Started as an experiment, these tools have now become part of a designer’s daily toolkits. There are many, many tools available in the market, from Midjourney to Firefly. Here are six courses to learn AI image generation for people starting from scratch or looking to learn advanced skills. Best AI Art Generation Courses OnlineMidjourney & ChatGPT: Unleash AI for Unique Image GenerationMake AI Work for You: Break Creative BlockAdobe Firefly Complete Guide: Learn to Use AI in Projects10 in 1 Course: Text to Image AI Art Generators MasterclassDesigning with Adobe Firefly and PhotoshopMaking Art with Stable Diffusion Midjourney & ChatGPT: Unleash AI for Unique Image Generation This easy-to-follow course has everything you need to make the most of Midjourney through these step-by-step lessons: Understand Midjourney and its unique features Learn how to use settings, suggestions, and advanced choices for creating custom images Plan out a business idea using Midjourney, including product pics, photos, website thoughts, and logos Get tips and tricks for making super practical suggestions for image creation Mix and match ideas to create pictures that are special and nice to look at Click here to know more. Make AI Work for You: Break Creative Block Join this class with Smitesh Mistry, an illustrator, content creator, and Skillshare’s top teacher, where he delves into collaborating with ChatGPT. Here’s what you’ll explore: Crafting ChatGPT art prompts tailored to your style and subject Sketching thumbnails inspired by AI-suggested concepts Utilising Adobe Firefly for visual references in your final illustration Translating co-created ideas from AI into your initial sketch As a bonus, Smitesh provides downloadable ChatGPT art prompts for those still navigating the nuances of communicating with AI software. This class not only guides you in incorporating AI into your drawing process but also imparts fundamental insights for creating illustrations in ProCreate. Click here to know more. Adobe Firefly Complete Guide: Learn to Use AI in Projects Dive into the bestselling ‘Mastering Adobe Firefly’ course on Udemy, designed to help you understand Firefly, Adobe’s AI-powered graphic design tool. Here’s what you’ll learn: Turn simple text into images, making your designs more interesting Text Effects: Play around with text using various effects to make ordinary words look amazing Learn how to change the colours of vector images precisely so your designs can match any colour scheme Explore the cool features of generative fill to create unique and vibrant backgrounds for your designs. Click here to know more. 10 in 1 Course: Text to Image AI Art Generators Masterclass This course is your ticket to the secrets of AI tools that turn text into images. It’s not just a simple stroll through various websites and tools; it’s a dive into the art of generating images using AI through creative prompts and tweaking the results to fit your fancy. The package covers a range of tools like Bing, Canva, Microsoft Designer, Blue Willow, Leonardo, and more. And here’s the cherry on top – once you are well-equipped with this technology, you’ll walk away with a certificate. Click here to know more. Designing with Adobe Firefly and Photoshop The free 8-hour stream on YouTube provides the perfect overview of Firefly and Photoshop’s newest feature, Generative Fill, with Adobe’s designers and Creative Cloud evangelists. The tool has become essential even for traditional designers since the entire Adobe Creative Cloud has been integrated with the tech. Moreover, if you plan to continue using Adobe’s products, getting an idea about how its AI-enabled tools work is a plus. Click here to know more. Making Art with Stable Diffusion This course is for anyone eager to whip up artwork through AI without any hassle. Picture this: you’ve got projects with deadlines, no time or cash to spare on creating graphics from scratch and zero artistic background. What’s the solution? AI, my friend! It’s quicker, cheaper, and more accessible – it can outdo your solo efforts. Here’s the lowdown on what you’ll learn during the course: Art of generating diverse styles using AI Get your hands on customising images with text prompts and variations Learn the secrets of Inpainting and Out Painting in AI art Master infinite zoom animations Creating videos with AI in Stable Diffusion Amp up your AI videos with post-production effects Discover how Stable Diffusion can work right inside Photoshop Click here to know more.","excerpt":"Learn AI art from scratch or advance your skills.","categories":["AI Trends"],"tags":["AI Tool","ChatGPT","Courses"],"author_name":"Tasmia Ansari","publish_date":"2024-02-14T11:00:00","publication_year":"2024","word_count":748,"keywords":["Go","ChatGPT","programming_languages:R","AI","GPT","stable diffusion","AI art","Courses","AI Tool","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","R","Go","GPT","stable diffusion","AI art","llm_models:GPT","llm_models:ChatGPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/6-courses-for-designers-illustrators-to-learn-ai-art\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10123315,"title":"Oracle and Google Cloud Partner to Simplify Multicloud Deployments","content":"Oracle and Google Cloud today announced a partnership enabling customers to combine Oracle Cloud Infrastructure (OCI) and Google Cloud technologies, accelerating application migrations and modernisation. Google Cloud’s Cross-Cloud Interconnect will initially be available for customer onboarding in 11 global regions, allowing the deployment of general-purpose workloads without cross-cloud data transfer charges. Later this year, a new offering, Oracle Database@Google Cloud, will be available, featuring Oracle database and network performance on par with OCI. The partnership aims to simplify cloud migration, multicloud deployment, and management. Both companies will jointly market Oracle Database@Google Cloud, targeting enterprises across financial services, healthcare, retail, manufacturing, and more. “Customers want the flexibility to use multiple clouds,” said Larry Ellison, Oracle Chairman and CTO. “To meet this demand, Google and Oracle are connecting Google Cloud services with the latest Oracle Database technology.” Sundar Pichai, CEO of Google and Alphabet, added, “This partnership will help customers use Oracle database and applications with Google Cloud’s platform and AI capabilities.” Oracle Database@Google Cloud offers direct access to Oracle database services running on OCI and deployed in Google Cloud datacenters. This service is designed to help customers accelerate cloud migration and modernize IT environments while leveraging Google Cloud infrastructure, tools, and AI services, including Vertex AI and Gemini foundation models. Benefits for customers include: Flexible migration options for Oracle databases to Google Cloud, with compatibility with migration tools such as Oracle Zero-Downtime Migration. Simplified purchasing via Google Cloud Marketplace, allowing use of existing Google Cloud commitments and Oracle license benefits. Unified customer experience and support from Google Cloud and Oracle. Deployment of Oracle database services, including Oracle Exadata Database Service, Oracle Autonomous Database Service, MySQL Heatwave, and more, within Google Cloud datacenters. Integration of Oracle data with Google’s AI services to enhance applications in customer service, employee services, creative studios, and more. Oracle will manage Oracle database services within Google Cloud datacenters globally, starting with regions in North America and Europe. Oracle Exadata Database Service, Oracle Autonomous Database Service, and Oracle Real Application Clusters (RAC) will launch later this year in US East (Ashburn), US West (Salt Lake City), UK South (London), and Germany Central (Frankfurt), with further expansion planned. The partnership also offers customers the ability to deploy workloads across both OCI and Google Cloud regions without cross-cloud data transfer charges. Initial onboarding will be available in 11 regions, including Australia East (Sydney), Australia South East (Melbourne), Brazil East (São Paulo), Canada South East (Montreal), Germany Central (Frankfurt), India West (Mumbai), Japan East (Tokyo), Singapore, Spain Central (Madrid), UK South (London), and US East (Ashburn). This collaboration allows customers to innovate using the best combination of Oracle and Google Cloud services, providing low-latency, high-throughput private connections between the two cloud providers. Customers can run multiple Oracle applications on OCI with distributed data stores on both OCI and Google Cloud and build new cloud-native applications using Google Cloud’s AI technologies. The new multicloud capabilities offer a fully integrated experience for deploying, managing, and using Oracle database instances within Google Cloud, enabling data movement and deployment of new cloud-native applications across both clouds. This integration allows organisations to leverage existing skills while utilising the best of Oracle and Google Cloud capabilities.","excerpt":"Later this year, a new offering, Oracle Database@Google Cloud, will be available, featuring Oracle database and network performance on par with OCI.","categories":["AI News"],"tags":["Oracle"],"author_name":"Siddharth Jindal","publish_date":"2024-06-12T08:42:11","publication_year":"2024","word_count":529,"keywords":["Go","AI","Oracle","RAG","Aim","llm_models:Gemini","SQL","cloud_platforms:Google Cloud","GAN","foundation models","R"],"extracted_tech_keywords":["AI","foundation models","Aim","RAG","R","SQL","Go","GAN","llm_models:Gemini","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/oracle-and-google-cloud-partner-to-simplify-multicloud-deployments\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":67525,"title":"How APIs Can Save AI Research Labs: Lessons From OpenAI","content":"“It is not a dream, it is a simple feat of scientific engineering, only expensive — blind, faint-hearted, doubting world!”Nikola Tesla Discovering a new medicine is a billion-dollar research endeavour. At least, it can draw in the money as the results are kind of self-explanatory; life-saving. But, in case of AI, which is usually riddled by speculations and scepticism, it is an uphill task for the researchers to sell their idea or to churn profits to keep fueling their AI labs. For example, OpenAI, which started as a non-profit research lab, changed its stance when it partnered with Microsoft. A year later, they have announced that they are making all their exotic deep learning innovations available to the public through an API that comes with a price tag. Now, customers can access state-of-the-art machine learning models without the headaches of training from scratch; GPT-3 training costs over $4 million! Today, the API can run models with weights from the GPT-3 family with speed and throughput improvements. In the next section, we take a look at OpenAI’s plan to unfold their new strategy and the key takeaways for other AI R&D labs and budding researchers. The Simpler The Better OpenAI team made sure that their API, unlike most AI systems which are designed for one use-case, is built to be both simple and flexible enough to make machine learning teams more productive. In fact, many of our teams are now using the API so that they can focus on machine learning research rather than distributed systems problems. OpenAI’s API is designed to provide a general-purpose “text in, text out” interface, that allows users to try it on any English language task virtually . We’ve designed the API to be both simple for anyone to use but also flexible enough to make machine learning teams more productive.OpenAI Modern day ML models are large and smaller organisations cannot afford them. With APIs, OpenAI tries to bring the advantages of their mega models to smaller businesses and organisations. Watching Out For Malicious Players Ever since the release of GPT, the text generator, OpenAI has been at the receiving end of criticism. OpenAI knows the adverse effects of their technology and have admitted the same in their latest paper on GPT-3. Now, these controversial machine learning models will be available to the public. To keep an eye on the consequences, their API was launched in a private beta rather than for general availability. In this way, the team believes that the users can control their content better with API. We cannot anticipate the consequences of a rapidly evolving technology. We can only deploy checkpoints. OpenAI states that they will terminate API access if the users use it for applications such as harassment, spam, radicalization, or astroturfing. Apart from this, they are also conducting research into the potential misuses of models served by the API, including with third-party researchers via an academic access program. Research Needs Revenue Today, we consider Marconi to be the father of wireless technology. However, Nikola Tesla was pursuing similar endeavours at the same time as Marconi. The difference between their successes was something as fundamental as funding! This is true even today. R&D department usually takes the first hit when an organisation is facing an economic downturn or a pandemic. For example, last month Uber announced that it would be winding down its AI wing. So, it is extremely important for any AI labs to maintain the cash flow. For example, the models developed by OpenAI are very large, taking a lot of expertise to develop and deploy, which make them very expensive to run. So, OpenAI will be soon announcing a pricing plan for its API customers, which it believes, in addition to being a revenue source, will also help them cover costs in pursuit of their mission. Leaving Room For Innovation What OpenAI API got right with their new strategy is that their tools are accessible to a wide range of users who will be willing to invest to access the top technology in the most simplistic manner. The OpenAI team also left some space for users to improve on the existing tools. This is a win-win scenario for both parties. While luring the customers with their technology, the team also has been vocal about the misuses of their technology and the steps they will be taking to stop them. The ultimate objective of all AI efforts is to achieve minimum or null human interference — AGI. And, for this to happen, research labs should devise strategies to give a commercial twist to their ideas.","excerpt":"“It is not a dream, it is a simple feat of scientific engineering, only expensive — blind, faint-hearted, doubting world!” Nikola Tesla Discovering a new medicine is a billion-dollar research endeavour. At least, it can draw in the money as the results are kind of self-explanatory; life-saving. But, in case of AI, which is usually […]","categories":["Deep Tech"],"tags":["AI Research","APIs","OpenAI"],"author_name":"Ram Sagar","publish_date":"2020-06-17T14:00:09","publication_year":"2020","word_count":766,"keywords":["Go","API","machine learning","OpenAI","AI","ML","AI Research","APIs","GPT","deep learning","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","OpenAI","R","Go","API","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-apis-can-save-ai-research-labs-lessons-from-openai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10081370,"title":"How Zoho Leveraged AI to Script A Unique Success Story","content":"Zoho is one of India’s premier success stories. Founded in 1996, this privately held company recently crossed $1 billion in revenue during a global recession. It even grew its revenue in India by 77% over the last year, and has announced its plans to expand globally by opening 100 more points of presence all over the world. The company is making strides in research, having filed 25 patents in the past two years. How did the company manage to turn a profit when the biggest tech giants are engaging in widespread layoffs? Analytics India Magazine reached out to Ramprakash Ramamoorthy, director at AI Research at Zoho, to know more about how the company scripted this success story. Ramprakash has been leading the AI efforts at Zoho Corporation since 2011, which has now grown to become main USPs of the company, handling 100 million requests a day. Future investments Zoho aims to be at the cutting edge of machine learning and artificial intelligence in the workplace, and have expressed their plans to double their investment for blockchain and AI. Speaking on this move, Ramamoorthy stated, “Our newer investments will help focus on firming up and honing the current technology offering to stay relevant, given how fast AI evolves.” The company aims to focus on research in statistical machine learning, computer vision, and natural language processing. These verticals provide a huge value add for businesses, as they allow them to derive unique insights which can be translated into better business. Zoho has identified fresh use-cases to help companies capitalise on their data. Some of these include a sales coach that will help sales engineers sell better, a what-if analysis engine that would help users run simulations and experiment on digital twin platforms, and AI provisions in even more vernacular languages. However, as seen with their privacy-first approach, all of these advancements will always keep the user’s privacy in mind. Ramamoorthy said, “As always, we will ensure none of these come at the cost of user privacy.” AI in Zoho One Zoho One, the company’s primary business product suite, incorporates AI in various facets of its services. Some use-cases include Zoho expense, which uses OCR to identify receipts, and Zia, their virtual assistant that utilises NLP for conversational AI. Moreover, due to its nature as a unified product suite, Zoho One can go beyond the problems faced by data siloing. Ramamoorthy remarked, “Our AI engine can use data across departments to offer contextual next best action for a given role. For example, to help solve an issue reported by a customer on Zoho desk, our customer service help desk software, the support engineer will be equipped with the history of the customer as well as the potential solution for his\/her issue pulled up from the knowledge base or similar tickets raised previously.” These optimisations are present across Zoho’s product suite and offer exponential benefits for companies using it. What’s more, Zoho also plans to increase the use-cases and verticals it deploys AI in. Ramamoorthy stated, “This is just one of the examples. We will relentlessly use AI to ensure process optimisation and efficiency maximisation across our enterprise software stack.” Privacy-first approach Their AI models are built with a specific focus on data privacy, and are able to generate accurate results even without personally identifiable information. This allows larger organisations to use Zoho’s tech to derive insights from their customers while keeping in line with global data privacy regulations. Ramamoorthy said, “At Zoho, we strongly believe privacy is more than just a feature and ensure a privacy-first approach in whatever we do — from building our own data centres to a homegrown AI stack.” Moreover, they also have a specific approach to their AI models that allows them to remove all personally identifiable information before the model is trained. “Any data that we use to train the baseline AI model runs through our PII (Personally Identifiable Information) detector and all PII is summarily removed before the data is fed to train the AI model,” said Ramamoorthy. They follow a very clean approach towards using data in the model, as there are additional checks and balances at every step to ensure there is absolutely no PII involved. “Once the baseline model is established and goes live, the next step for the model is to be personalised based on individual user behaviour – at this stage we completely anonymise the process, thereby ensuring no personal information ever gets used to train or personalise the AI engine,” Ramamoorthy explained. They also adopt industry-best practices when it comes to handling datasets and information. Ramamoorthy clarified, “We treat data the same way we treat our source code repositories. Our practices include proper and periodic reviews of access controls, maintenance of access logs and version control, and more. All data is processed only in the corresponding data centres, thereby ensuring our AI models are in sync with the local data storage requirements.” This privacy-first approach also ensures a better model is trained, as it is largely free from bias due to the omission of personal identifiers. Considering the approach Zoho is taking, it is no wonder the company is thriving. From their India-first approach, to its privacy-focused AI algorithms, to their laser-sharp focus on customer dedication, the company is setting the bar high for its competitor.","excerpt":"This customer-focused software provider plans to double investments in AI and blockchain","categories":["AI Features"],"tags":["Interviews and Discussions","zoho","Zoho CRM"],"author_name":"Anirudh VK","publish_date":"2022-12-03T16:00:00","publication_year":"2022","word_count":886,"keywords":["Go","zoho","artificial intelligence","machine learning","AI","computer vision","RAG","NLP","Aim","Zoho CRM","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","computer vision","analytics","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-zoho-leveraged-ai-to-script-a-success-story\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10052529,"title":"Best Python Libraries For Data Processing","content":"Data processing services are available in various encodings, including CSV, XML, HTML, SQL, and JSON. Each situation requires a unique processing format. There are numerous programming languages. Python is frequently recommended as a viable alternative for machine learning applications due to its implementation of major libraries and cutting-edge technologies. Machine learning is built on data processing, and model success is highly dependent on the ability to read and transform data into the format required for the task at hand. Let us examine the various Python libraries in terms of the data types they provide. Below, we have covered the Python libraries used for processing different types of data: Tabular Data Most of the large data is available in the tabular format, with rows referring to records and columns corresponding to features. Pandas in Python can handle such type data very perfectly. The advent of tabular data has evolved into a full-featured library that can handle both series and tabular data. Text data First, it’s worth noting Python’s extensive built-in text-processing capabilities. However, many natural language processing techniques, such as tokenization and lemmatization, may be done using NLTK. Along with that, Spacy is a good choice for advanced natural language processing and optimised pipelines. Audio and musical data Audio processing is enabled via libraries like librosa and essentia. Mido and pretty midi are good choices for symbolic music, like MIDI. Finally, music21 is a sophisticated library targeted at musicology analysis. Images Pillow is an image processing library in Python. Opencv is a computer vision library that can process videos or camera data. Because of its vast range of supported formats, imageio can give image data to the python script. Python, in particular, is a highly regarded data processing language for a variety of reasons, including the following: Prototypes and experimentation with code are incredibly simple. Processing data, especially from less-than-clean sources, necessitates a great deal of tweaking, back and forth, and a struggle to capture all options.Python3 significantly improved multi-language support by making every string in the system UTF-8, which enables the processing of data encoded in different character sets by different languages.The standard library is quite strong and packed with essential modules that provide native support for common file types such as CSV files, zip files, and databases.The Python third-party library is enormous, and it has a wealth of excellent modules that enable it to increase the capabilities of a programme. There are also modules for geospatial data analysis, creating command-line interfaces, graphical interfaces, parsing data, and everything in between. Jupyter Notebooks allows you to execute code and receive immediate feedback. Python is quite agnostic about the development environment required, allowing it to function with anything from a simple text editor to more complex alternatives such as Visual Studio. Conclusion In general, Python and R programming are two extensively used data processing languages. Javascript, like Python, has a thriving ecosystem. Julia is also in attendance. Almost every modern language is capable of data analytics. However, the capability varies according to the purpose. While R has the greatest statistical analysis features of any packages, Python meets the needs of the vast majority of analysts and is fast gaining popularity. It is preferable to begin with Excel, SQL, and basic programming concepts, then switch to a more widely spoken language and master it. After that, take a step back and apply the principles to real-world situations. To summarise, familiarise with R if conceptual understanding and application are crucial during this period. If large-scale data analysis is necessary, familiarity with Python’s big data capabilities is recommended.","excerpt":"What are the benefits of learning Python for data processing?","categories":["AI Trends"],"tags":["csv","data processing","databases","geospatial data","HTML","Javascript","Julia","Jupyter Notebooks","Machine Learning","NLTK","OpenCV","pandas","Python","python database gui","Python Libraries","r programming","SQL"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-11-01T10:00:00","publication_year":"2021","word_count":593,"keywords":["Javascript","computer vision","Pandas","databases","csv","HTML","analytics","python database gui","machine learning","AI","Python Libraries","data processing","ML","Machine Learning","NLTK","spaCy","Jupyter Notebooks","geospatial data","OpenCV","Python","pandas","SQL","Julia","r programming","Jupyter"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","analytics","Jupyter","OpenCV","spaCy","NLTK","Pandas"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/python-libraries-for-data-processing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10016761,"title":"What Does It Take To Set Up An Overseas Data Science Company in India?","content":"In the early days of the IT revolution, India was looked at only as an outsourcing destination — cost naturally being the prevailing factor. The IT adoption in India was slow, and people would hardly use any software themselves, even though they were making it. Now that everything is digital, it gets consumed unknowingly. This has led to India’s adoption growth to increase drastically. To add to that, everybody has realised the importance of data. Hence the sheer volume of the Indian market has attracted many foreign companies who are data-hungry to have a presence in India. However, setting up data science teams in India requires a variety of considerations and takes time. Traditional models of outsourcing and establishing subsidiaries in India present some roadblocks while doing the same. Analytics India Magazine caught up with Rajendra Vaidya, CEO of Remunance, to understand the factors one needs to consider while setting up data science, AI, and machine learning companies in India and overcoming some of the roadblocks. Talent When a company decides that it wants a setup in India, essentially what it wants is a physical presence in terms of human resources. So, the first and most relevant question becomes how to hire the right talent. For that, from a conventional point of view, one has the option of contracting or subcontracting. But when these two options are considered, IP-related issues start bothering the parent company. This is especially relevant with technologies like AI and machine learning as they need extreme confidentiality and IP protection. Enterprises thus want resources that they can control and relate to. “The tendency is to form a company or its subsidiary. This is because unless you don’t form a company, you cannot hire a resource in India, or for that matter in any country,” said Vaidya. “But having a company in India itself is a complex situation.” Another important factor, especially when it comes to AI, machine learning, and other high-end technologies, is ‘the availability of the right kind of resources rather than the number of resources.’ And this is not only in terms of intelligence or capabilities but also the cultural fit. Time Indian economy and its statutes are always evolving. There are continuously new rules and reforms coming in. Setting up data science companies especially becomes more tricky since there is a looming uncertainty over the data privacy and protection rules. Considering these factors, establishing a subsidiary in India is a long process. There are specific rules to form a subsidiary that still take a lot of time. This is despite the various digital initiatives in India and the improvement in the Ease of Doing Business rankings. This is not affordable for data science companies. Firstly, they need more certainty over data rules, and secondly, delays can cause existing data ‘to become stale’. “Companies developing data science products need faster response times. They need to set up firms quickly after a decision has been made,” said Vaidya. “It will take three-four months to form a subsidiary. After that, a few months are spent hiring resources. And even after hiring resources, you do not know whether they will stick with you or not, which again is a concern for data science companies in terms of all the intelligence property.” Cultural Interface AI and machine technology are continuously evolving and at a rapid pace. The communication here is at an extremely high-level and needs more vocabulary than everyday English. It has a lot to do with understanding and communicating mathematics. You need people to find logical and innovative ways to look at the data and make it more actionable. “A lot of data science companies have understood that to make offshore teams successful, you need someone in India who can relate to things and communicate things between the parties,” said Vaidya, “Hence, as far as communications go, both counterparts need a little bit of hand-holding at the start.” Support Systems When companies want to establish in India, at the start, the teams are usually small. At the same time, AI and data science companies are extremely dependent on support systems like accounting, admin, HR, among others. Also, for AI and machine learning companies, their data needs good security firewalls and good bandwidth. Thus, network administration becomes one of the key activities. Conclusion “India is a melting pot of cultures, and thus people have a different approach of looking, analysing, and consuming data,” said Vaidya, “Thus, if data teams and their applications work in India, they can work almost anywhere.” Thus, India is one of the best destinations to set up a data science firm but presents several challenges in terms of hiring talent and time, especially if traditional models are considered. New and innovative ideas like Professional Employment Organisation (PEO) models are helping overcome these. These setups help offshore companies hire employees under their agreements without forming a subsidiary. Considering the factors mentioned above, it is important for companies to look for different approaches and models that suit them the most and help them speed up the process to start their operations in India.","excerpt":"In the early days of the IT revolution, India was looked at only as an outsourcing destination — cost naturally being the prevailing factor. The IT adoption in India was slow, and people would hardly use any software themselves, even though they were making it.  Now that everything is digital, it gets consumed unknowingly. This […]","categories":["IT Services"],"tags":["Artificia Intelligence in Data Science","artificial intelligence machine learning data","Data Science","how does ai work","how does artificial intelligence work","outsourcing"],"author_name":"Kashyap Raibagi","publish_date":"2020-12-30T11:00:00","publication_year":"2020","word_count":848,"keywords":["data science","Go","API","machine learning","ELT","AI","outsourcing","how does ai work","artificial intelligence machine learning data","Git","how does artificial intelligence work","analytics","GAN","Data Science","R","Artificia Intelligence in Data Science"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","R","Go","Git","API","ELT","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/what-does-it-take-to-set-up-an-overseas-data-science-company-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":22025,"title":"Google Announces DSC Summit In Goa To Encourage Young Tech Enthusiasts","content":"Building on the momentum to creating a pool of highly-skilled tech workforce in India and giving the large Indian student community a head start on latest technologies like machine learning, artificial intelligence and mobile and web development, Google on Friday announced a Developer Students Club (DSC) summit in Goa. The program is aimed at inspiring and training student ambassadors who will encourage students in their campuses to learn newer areas of technologies that will equip them with the right skill sets as they step out of colleges to join the workforce. William Florance, Developer Products Group and Skilling Lead, Google, said in a statement, “In the last one year, we have engaged over half a million students and developers across India through a variety of programs and initiatives. To build on this momentum, and broaden access to world class technology curricula for students, we are delighted to launch the Developer Students Club summit today. These programs will make it easy for students and developers to gain the skills they need to be successful in the changing technology landscape.” The inaugural batch will see 200 student ambassadors from 98 cities across 170 colleges across the country attending the three-day summit and learn about various emerging technologies through a series of sessions. The summit will also have a Design Thinking workshop by the University Innovation Fellows from across India, encouraging them to find creative and innovative solutions, using advanced technologies to solve for India’s complex problems. “It is an awesome feeling teaching others what you know; we are not only teaching but also growing ourselves.’”, said, Spoorthi V, Vidya Vikas Institute of Engineering and Technology, Mysuru and one of the student brand ambassadors attending the DSC Summit. The DSC summit is in line with Google’s larger program of training two million developers in India. Last year, the company announced 1,30,000 scholarships for both working professionals and students community, that gives them access to new-age technology and an opportunity to gain skills they require to be successful in the changing technology landscape.","excerpt":"Building on the momentum to creating a pool of highly-skilled tech workforce in India and giving the large Indian student community a head start on latest technologies like machine learning, artificial intelligence and mobile and web development, Google on Friday announced a Developer Students Club (DSC) summit in Goa. The program is aimed at inspiring […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Google","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2018-02-23T11:59:05","publication_year":"2018","word_count":339,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","innovation","Machine Learning","programming_languages:Go","RAG","Aim","Google","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","RAG","R","Go","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-developer-students-club-summit-goa\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10068603,"title":"CredAvenue + Infosys: When tech firms reimagine co-lending","content":"Fintech platform CredAvenue has partnered with Infosys Finacle to develop innovative co-lending solutions. Finacle is a part of EdgeVerve, a wholly-owned subsidiary of Infosys. As a digital banking solution provider, it addresses the core banking, lending, digital engagement , cash management, wealth management, treasury, analytics, AI, and blockchain requirements of financial institutions. Chennai-based CredAvenue is a fully integrated, unified digital platform to discover, trade, execute and fulfil debt solutions for investors. The digital lending sector in India is expected to reach a valuation of USD 350 billion by 2023 from USD 110 billion in 2019. In 2018, RBI announced a co-origination model between banks and Non-Banking Financial Companies – Non-Deposit taking – Systemically Important (NBFC-ND-SIs) for providing competitive credit to the priority sector. In 2020, RBI permitted banks to co-lend with all registered NBFCs (including HFCs). Ever since, a lot of co-lending partnerships have been announced. State Bank of India’s partnership with Adani Capital to offer credit to farmers, Union Bank of India’s collaboration with UGRO to facilitate last-mile finance for MSMEs, Bank of India’s co-lending partnership with Vijayawada-based IKF Finance to offer commercial vehicle loans, to name a few. Co-lending is essentially a collaborative financing model between NBFCs and banks to disburse loans. It benefits both partners by giving NBFCs access to low-cost capital and enabling banks to access new customer segments historically inaccessible to them without incurring high operational costs. Moreover, the borrowers get access to funds at considerably lower interest rates while bringing them under an organised lending structure. CredCo-Lend CredAvenue’s CredCo-Lend helps streamline the entire co-lending process for lenders and originators by bringing banks and NBFCs in a unified marketplace. CredCo-Lend ensures entities comply with KYC\/AML norms without compromising processing time and seamlessly shares customer information and supporting documents between the originator and the co-lender. Personalised and automated workflows, reconciliation, invoicing and managing repayment split between the co-lenders, CredCo-Lend ensures the lending institutions can focus on credit disbursal growth and customer expansion without the hassle of long drawn out technical integrations. At the core, CredAvenue’s offering is about interoperability. CredAvenue’s clientele includes Kinara Capital, a leading collateral-free MSME loan provider; SmartCoin, a lending platform that offers short-term credit; and Orange Retail Finance India Private Limited (ORFIL), an NBFC. With CredCo-Lend, Kinara Capital was able to disburse loans worth Rs 125+ crore via multiple lenders. CredAvenue marketplace gave them access to qualified and verified lenders. CredCo-Lend gave SmartCoin access to lenders and a co-lending platform to increase loan offerings, expand the target audience, and disburse more loans to the credit-strapped segments of the economy without hassles. CredAvenue has become the fastest Indian fintech startup to join the unicorn club. The startup has facilitated loans of over USD 10.5 billion to date. More than 2,300 corporates, 450 enterprises and 750 lenders are active on the platform. CredAvenue has raised USD 137 million in Series B at a valuation of USD 1.3 billion. The fresh funds will be used to expand its business in India and to acquire companies to fuel its growth. What does the new partnership mean? CredAvenue’s partnership with Infosys will help their banking clients automate and streamline their co-lending operations. “At the core of this collaboration is our desire to create a strong technology and data-driven environment for co-lending and co-origination. This partnership with Infosys Finacle is a game-changer and a global first, which will resolve the integration problems between banks and non-banking financial companies (NBFCs) at the category and ecosystem levels. This platform will simplify the co-lending processes of banks and NBFCs through deep integration between banking platforms and co-originator platforms. It is estimated that the banks using this platform will be able to invest in less than a tenth of the current time taken. Besides, the banks will be able to partner seamlessly with small NBFCs, reaching out to last-mile borrowers,” said Gaurav Kumar, founder and CEO, CredAvenue. Typically, the lending entities rope in tech firms to take care of the tech side and streamline their processes. Sometimes, banks approach fintechs with NBFC licences to enable co-lending. “Two technology leaders came together to reimagine the colending ecosystem with the possibility of real time reporting and processing of the co-origination assets in banks’ core banking and loan processing systems and helping meet RBI compliance,” said Sameer Singh Jaini, CEO, The Digital Fifth. As lending becomes increasingly digital, the collaboration will facilitate banks to automate and streamline their co-lending operations. This will accelerate the co-lending ecosystem and bring down the cost of credit to a great extent as banks will be able to participate on a large scale with NBFCs and fintech. “We are pleased to welcome CredAvenue to the Finacle partner ecosystem. The co-lending business model is seeing significant growth. It is helping banks, Non-Banking Financial Companies (NBFCs), and Housing Finance Corporations (HFCs) leverage respective strengths to improve credit flow in the economy. Our collaboration with CredAvenue will help our clients effectively tap into the growing opportunity and automate and streamline their co-lending operations,” ​​said Venkatramana Gosavi, Global Head of Sales, Infosys Finacle. A 2021 BCG report titled “The Poster Child” said co-lending is a great enabler for the flow of credit to small and medium enterprises, with NBFCs taking care of the last mile. Against this backdrop, CredAvenue and Finacle coming together can be seen as a right step in the direction to solve India’s liquidity problem.","excerpt":"CredAvenue streamlines the entire co-lending process for lenders and originators by bringing banks and NBFCs in a unified marketplace.","categories":["IT Services"],"tags":["Infosys"],"author_name":"Zinnia Banerjee","publish_date":"2022-06-08T17:03:33","publication_year":"2022","word_count":894,"keywords":["Go","API","Infosys","AI","data-driven","ML","Git","RAG","analytics","GAN","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","Git","API","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/it-services\/credavenue-infosys-when-tech-firms-reimagine-co-lending\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":42260,"title":"The Brains Behind AI: How Pavlov’s Dogs &#038; Weight Loss Tips Influenced Reinforcement Learning","content":"Artificial intelligence has, in essence, executed many psychological concepts in a digital form. Fittingly, one of the biggest parts of human intelligence is the ability to learn and improve upon past tries of the same task. While this has been extended into AI as machine learning, there exists a specific type of ML that borrows heavily from psychology. Reinforcement learning is based on the concept of conditioning in psychology and applies it in a unique way to facilitate dependable learning. What Is Conditioning? ‘Conditioning’ is a general term used to describe a phenomenon where a previously unconnected stimulus and response are linked by learning. One of the earliest, and most famous, types of conditioning is classical conditioning, also known as Pavlovian conditioning. Classical Conditioning: First discovered by Russian physiologist Ivan Pavlov, this method of conditioning focuses on pairing a neutral stimulus with the response from a biologically potent stimulus. This can be seen in the example of Pavlov’s dogs. The physiologist discovered this phenomenon when he was studying digestion in dogs. When the food was brought in, the dogs salivated; an involuntary biological response to food. However, he experimented with ringing a bell every time the food was brought in, thus creating a connection between the sound of the bell and the food. This resulted in the dogs salivating whenever they heard the bell ring, thus being ‘conditioned’ to respond in a way similar to how they would to a conditioned stimulus (food), except without the stimulus being present. Thus, they had ‘learned’ that the sound of the bell meant food was coming. Today, classical conditioning has found applications in diet watch gadgets. These gadgets give the user a mild electric shock upon them exhibiting an unfavourable behaviour, usually binge eating. A connection is formed between the unpleasant stimulus of an electric shock towards the response of eating, eventually cutting down the eating habits of the wearer. Operant Conditioning: Another type of conditioning is operant conditioning, which is built on top of classical conditioning principles and was the inspiration for RL. Pioneered by psychologist BF Skinner, this was looked at as a method to explain more complex human behaviours that could not be explained by classical conditioning. Operant conditioning takes a more in-depth look into the process of conditioning, and also brings a way to influence human behaviour by inflicting actions. The process features 3 main principles; reinforcement, punishment, and extinction. Operant conditioning functions on the idea that encouraging positive behaviour and discouraging negative behaviour can have positive effects on the psyche. Encouraging positive behaviour through favourable changes to the environment is known as reinforcement while discouraging negative behaviour through unfavourable changes is known as punishment. Extinction is the removal of a connection between a stimulus and a response after a long period of neither punishment or reinforcement. This results in behaviours being eliminated altogether. Reinforcement and its subcategories are the basis of what make up the concepts of reinforcement learning. How Psychology Is Implemented In RL Instead of using both reinforcement and punishment, RL utilizes two forms of reinforcement. These are positive reinforcement and negative reinforcement, and are seen in the reward systems of a reinforcement learning workflow. Positive reinforcement is when a reward is given to encourage positive behaviour. Negative reinforcement is when a punishment is taken away to encourage the behaviour. While it is not this black and white in RL, these concepts are used in a gradient form to ensure that the system continues on its path of self-improvement. More effective solutions are given a higher amount of rewards, while less effective solutions are provided with a lower amount of rewards. This creates a conditioning within the algorithm that more effective solutions offer a higher chance of obtaining rewards, leading to the agent to try and pick the solution that gives the maximum amount of rewards. The concept of extinction also finds a use in this approach, as older, less effective paths to a solution are effectively weeded out due to a lack of reinforcement. Conditioning In Reinforcement Learning RL is a direct representation of the concept of reinforcement used for learning. In a typical RL workflow, an agent (algorithm) performs its designated function in the environment. The result is then passed on to an interpreter, which decodes both the state of the environment and the reward to be given to the algorithm. The reward given to the system depends on the degree of success or efficiency with which the problem is solved. Therefore, the algorithm tries to solve the problem with varying degrees of effectiveness. On the first iteration, the system will, most likely, come up with the least effective solution. However, as more effective solutions are found and reinforced by offering rewards to the system, the solution itself moves towards being more efficient. This then creates a self-learning algorithm that improves itself using the feedback given to it by the interpreter. Reinforcement learning is different from other machine learning methodologies, as it does not need to be told how exactly to solve the problem. It uses psychological methods to simulate human learning processes. This is just one of the many psychological concepts applied for use in AI, with a plausible way forward being to apply more complicated theories to machines. Hence, the rise of a true artificial intelligence can come from a deeper, psychological understanding of human consciousness.","excerpt":"Artificial intelligence has, in essence, executed many psychological concepts in a digital form. Fittingly, one of the biggest parts of human intelligence is the ability to learn and improve upon past tries of the same task. While this has been extended into AI as machine learning, there exists a specific type of ML that borrows […]","categories":["AI Features"],"tags":["Reinforcement Learning"],"author_name":"Anirudh VK","publish_date":"2019-07-11T17:00:44","publication_year":"2019","word_count":894,"keywords":["Go","machine learning","artificial intelligence","Reinforcement Learning","AI","programming_languages:R","ML","programming_languages:Go","Git","RAG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-brains-behind-ai-how-pavlovs-dogs-weight-loss-tips-influenced-reinforcement-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10085625,"title":"Hardware-software Symbiosis: The Litmus Test for Tech Industry","content":"‘What Intel giveth, Microsoft taketh away’ is more than just a clever quip – it’s a reflection of the increasing software complexity counteracting the growing pace of hardware. The spotlight may be on Moore’s Law, but Wirth’s Law provides a contrasting viewpoint on the evolution of technology. The law states that while advanced chips offer extra power and memory, software designed by companies like Microsoft is getting more complex (to make them do more). In the process, the software takes up the available memory space. This is why we haven’t seen a significant increase in the performance of software applications over time, and in some cases, they have even become slower. Niklaus Wirth believes that one of the important things that contributes to increasing complexity in the software world is the users’ lack of ability to distinguish between necessary and unnecessary functions in certain applications which leads to overly complex and unnecessary designs in software. For instance, Windows 11, an upgrade to the 10th, offered little-to-no performance gain in real-world use. Outside of the hoopla around the new look and feel given to it, the upgrade only offers supporting capabilities to the more advanced hardware requirements compared to its predecessor. It is like the software world is playing catchup to the up-and-coming hardware releases. Liam Proven, in writing for The Register, says that there is a symbiotic relationship between the hardware and software. “As a general rule, newer versions of established software products tend to be bigger, which makes them slower and more demanding of hardware resources. That means users will want, or better still need, newer, better-specified hardware to run the software they favour,” he writes. Integration difficulties However, Sravan Kundojjala, principal industry analyst at Strategy Analytics, told AIM, “The hardware and software symbiosis is easier said than done. For example, the AI chip landscape has quite a few start-ups but most of them lack software support to take advantage of the platform features.” A good software stack is important for the effectiveness and success of an AI chip. This is because when it comes to AI, compute itself is fundamentally different. AI chip company Graphcore’s Dave Lacey discusses three reasons to why this is the case: (i) Modern AI and ML technology deals with uncertain information, represented by probability distributions in the model. This requires both detailed precision of fractional numbers and a wide dynamic range of possibilities. From a software perspective, this necessitates the use of various floating-point number techniques and algorithms that manipulate them in a probabilistic way. (ii) The high-dimensional data, such as images, sentences, video, or abstract concepts, is probabilistic and irregular, making traditional techniques such as buffering, caching and vectorization ineffective. (iii) Additionally, machine intelligence compute deal with both large amounts of data for training and a significant number of computing operations per data processed, making it a significant processing challenge. Thus, a co-existence of AI hardware design and software algorithms is essential to improve performance and efficiency. Chip companies provide software development kits (SDKs) to developers, allowing them to access and utilise the platform’s features via application programming interfaces (APIs). An example of this is Qualcomm, which offers an SDK that enables original equipment manufacturers (OEMs) to utilise the AI capabilities of its chips. Companies that utilise these SDKs tend to have an advantage in terms of power efficiency and features. Similarly, Graphcore’s IPU-Machine M2000, which utilises off-chip DDR memory, doesn’t have hardware-based cache or mechanism to automatically manage the transfer or buffering of data between the external streaming memory and on-chip in-processor memory. It all relies on software control, using the computation graph as a guide. However, as indicated above, this is not entirely easy. Kundojjala said, “Even companies such as AMD and Intel are finding it hard to compete with NVIDIA in AI due to a lack of significant software developer support for their AI chips.” NVIDIA’s CUDA monopoly has been long-known. It dominates the AI chip market offering the best GPUs, with proprietary APIs exclusive for them in CUDA. GPT-3 and Stable Diffusion are all optimised for NVIDIA’s CUDA platform. Its dominance is therefore difficult to break. As Snir David points out, large businesses may incur additional costs by using non-mainstream solutions. This can include resolving issues related to data delivery, managing code inconsistencies due to the lack of CUDA-enabled NVIDIA cards, and often settling for inferior hardware. heading into a world where Google, OpenAI, Meta, and Anthropic all have chatGPT-level models available by API, Nvidia seems like the real winner— Daniel Fein (@DanielFein7) January 13, 2023 RISC-V to the rescue However, Kundojjala also mentions, “maintaining software compatibility on a hardware platform often comes at a cost”. While software growth propels buying new hardware, when the software matures it actually becomes a burden for hardware companies as they have to support legacy features. But, new architectures like RISC-V are offering a fresh template to companies in order to avoid suffering from legacy software support. As an open-source alternative to Arm and x86, RISC-V is already backed by companies like Google, Apple, Amazon, Intel, Qualcomm, Samsung, and NVIDIA. RISC-V is often likened to Linux in the sense that it is a collaborative effort among engineers to design, establish, and enhance the architecture. RISC-V International establishes the specifications, which can be licensed for free, and chip designers are able to use it in their processors and system-on-chips in any way they choose. It offers the flexibility to harness generic software solutions from the ecosystem. The open-source ISA allows an extremely customisable and flexible hardware and software ecosystem. Therefore, while historically there has been an imbalance between hardware and software progress, with open-source architectures, we can see the gap narrowing down a little. But, nevertheless, as Kundojjala says, “It looks like on most occasions, the software is the limiter as it requires more collaboration across the industry whereas hardware can be developed by individual companies.”","excerpt":"While hardware is achieving milestones each day, the software growth is only slowing down","categories":["AI Features"],"tags":["NVIDIA"],"author_name":"Ayush Jain","publish_date":"2023-01-23T17:00:00","publication_year":"2023","word_count":983,"keywords":["CUDA","Anthropic","ChatGPT","Go","OpenAI","AI","ML","Aim","analytics","NVIDIA","R"],"extracted_tech_keywords":["AI","ML","analytics","ChatGPT","OpenAI","Anthropic","Aim","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hardware-software-symbiosis-the-litmus-test-for-tech-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10021876,"title":"Inside Multimodal Neurons, The Most Advanced Neural Networks Discovered By OpenAI","content":"In a major breakthrough, researchers at OpenAI have discovered neural networks within AI systems resembling the neural network inside the human brain. The multimodal neurons are one of the most advanced neural networks to date. The researchers have found these advanced neurons can respond to a cluster of abstract concepts centred around a common high-level theme rather than a specific visual feature. Like their biological counterparts, these neurons can respond to a range of emotions, animals, photographs, drawings and famous people. Researchers wrote these neurons in CLIP can respond to the same concept, whether presented literally, symbolically, or conceptually. The multimodal neurons have been discovered in the CLIP model that can connect text and images. It can learn visual concepts from natural language supervision. Further, this general-purpose vision system can match the performance of a ResNet-50 but outperforms existing vision systems on the most challenging datasets. For instance, one neuron called the ‘Spider-Man’ can respond to a spider’s image, the text ‘spider’, and the comic book character ‘spider-man’. The Study The researchers found multimodal neurons in several CLIP models of varying sizes, but they focused on studying the mid-sized RN50-x4 model. Researchers employed two tools to understand the activations of the model: Feature visualisation, which maximises the neuron’s firing by doing gradient-based optimisation on the input.Dataset examples, which looks at the distribution of maximal activating images for a neuron from a dataset. The researchers carried out a series of carefully-constructed experiments to find these neurons’ unique capabilities in the convolutional layer. Each layer consists of thousands of neurons. “For our preliminary analysis, we looked at feature visualisations, the dataset examples that most activated the neuron, and the English words that most activated the neuron when rastered as images,” said researchers. Most of these neurons were made to deal with sensitive topics, from political figures to emotions. The experiment revealed an incredible diversity of features such as region neurons, person neurons, emotion neurons, art style neurons, time neurons, abstract neurons, colour neurons and more. Researchers found that a majority of neurons in CLIP are readily interpretable. “From an interpretability perspective, these neurons can be seen as extreme examples of “multi-faceted neurons” which respond to multiple distinct cases. Looking to neuroscience, they might sound like “grandmother neurons,” but their associative nature distinguishes them from how many neuroscientists interpret that term,” stated researchers. Researchers also studied how these multimodal neurons can give us insight into understanding how CLIP performs classification, such as image and text classification. Not Fool-Proof Neural networks work on the same principle as their biological counterparts to process data. However, the drawback is, it is difficult to understand why it makes certain decisions and how it comes to a particular conclusion. The researchers said that despite being trained on a curated subset of the internet, it still inherits its many unchecked biases and associations. “…we have discovered several cases where CLIP holds associations that could result in representational harm, such as denigration of certain individuals or groups,” researchers stated. For instance, “Middle East” neuron was associated with terrorism; and an “immigration” neuron responded to Latin America. Despite fine-tunes and the use of zero-shot techniques, researchers said these biases and associations would remain in the system. The CLIP findings are still evolving, and there is a lot of research and understanding that needs to be done in multimodal systems. In a bid to advance the area, researchers have shared the tools, dataset examples, text feature visualisations, and more with the community.","excerpt":"In a major breakthrough, researchers at OpenAI have discovered neural networks within AI systems resembling the neural network inside the human brain. The multimodal neurons are one of the most advanced neural networks to date.  The researchers have found these advanced neurons can respond to a cluster of abstract concepts centred around a common high-level […]","categories":["AI Trends"],"tags":["Neural Networks","online network graph"],"author_name":"Srishti Deoras","publish_date":"2021-03-11T09:00:00","publication_year":"2021","word_count":580,"keywords":["text classification","OpenAI","AI","online network graph","neural network","RPA","Modal","programming_languages:R","CLIP","ResNet","R","Neural Networks"],"extracted_tech_keywords":["AI","neural network","OpenAI","text classification","R","CLIP","ResNet","RPA","Modal","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/inside-multimodal-neurons-the-most-advanced-neural-networks-discovered-by-openai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":1288,"title":"Interview &#8211; Sudeshna Datta, EVP and Co-founder at Absolutdata","content":"Sudeshna Datta of Absolutdata talks with Analytics India Magazine about the growth of analytics practice and how Absolutdata is poised to capture this opportunity. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: What according to you are currently the most important and specific analytics needs for the industry? [dropcap style=”1″ size=”2″]SD[\/dropcap]Sudeshna Dutta: In today’s dynamic business environment, companies face increasingly stagnant markets and reducing customer loyalty. Hence, the need of the hour is to maximize ROI and develop deeper, stronger relationships with the customers. Analytics helps organizations achieve exactly that through solutions like Marketing Effectiveness and Customer Relationship management. AIM: What according to you is your biggest USP that differentiates your organization with similar sized players in the analytics space in India? SD: AbsolutData is an analytics firm which has a sharp focus on Marketing Analytics. We leverage our domain consultants with extensive industry experience to deliver stronger, contextual solutions driving business impact for our clients. We use an optimal mix of onshore-offshore specialists to deliver cost effective solutions achieving higher ROI for our clients. We also provide technology-agnostic analytic services that can work on variety of data formats, and with commercial Business Intelligence and data warehousing tools. AIM: Please brief us about some business solutions you provide to your customers and how do they derive value out of it. SD: AbsolutData provides a full range of Analytics and Research solutions to its customers. The clients use these services to evaluate and improve their marketing effectiveness, customer relationship, and overall market understanding. Our services include: Marketing effectiveness: This helps maximize ROI from organizations marketing investments. Customer Analytics: Here we work with customer data such as transactional data, survey data, behavioral data etc. to unearth insights which drive business decisions Market Research Analytics: Here we collect survey data and deliver insights from it. This also covers web and secondary research. AIM: Different organizations are in different life stages of BI maturity with unique requirements. Yet, most matured organizations were early adopters of BI\/ analytics capabilities either through in-house\/ third party development or specialized standard products. Where do you see a gap and how do you plan to fill these? SD: The gap really lies in application of BI\/ Analytics capabilities; there is a need of specialized and trained manpower and an infrastructure conducive of applying statistical tools. Some of the early adopters of analytic services have not implemented full scale solutions, which has created the gap. We foresee people and infrastructure as two biggest assets to fill up this gap. Also, AIM: Where do you see the bulk of your business coming from? Do Indian organizations have the same affinity towards BI\/ Analytics as that of organizations from other regions? SD: Most of our clients are based in the US and Europe. Majority of our revenue comes from Marketing Analytics and the balance from Marketing Research. At the same time, whilst Indian companies are becoming more customer centric and getting exposed to strategies adopted by global organizations, they are realizing the importance of BI\/Analytics and adopting them. AIM: Do you envisage BI or analytics reaching a level of standardization in near future, where solutions can be deployed on a ‘plug-and-play’ basis without having to deal with complexities and nuances of an industry? SD: Plug-and-play analytics solutions are aspirational both for clients and providers as these are faster, better and cheaper to implement. Although companies are making huge investments to synchronize their departments, various verticals are in different stages of evolution both in their IT and Analytics adoption. As long as these challenges continue to exist blanket pure plug and play will not be possible. The scope for customized analytics solutions will always exist. AIM: What is your projection of the growth of analytics practice in the future? SD: Analytics services include information retrieval, data management, reporting and statistical analysis. Based on the IDC estimates, this segment represents a $40 Bn market globally and is estimated to grow rapidly as the benefits of data driven decision making and analytics processes become known. While financial services and airlines have been early adopters of analytics, while other industries such as consumer, retail, technology and healthcare are following. Several large Fortune 500 companies have already shown a propensity to outsource business analytics and have grown into large accounts for analytics vendors like us. Also, with investment firms showing interest in the sector, the growth prospects have become brighter. [spoiler title=”Biography of Sudeshna Datta” open=”0″ style=”2″] With over a decade of experience in general management and marketing, Sudeshna is a member of the executive management team at AbsolutData and leads the Corporate Development, Human Resources and Marketing functions. Prior to co-founding AbsolutData Technologies, Sudeshna was a co-founder of globedecor.com; an Internet based international home décor. In the past, she has held prestigious roles as a Business Development Manager at Pfizer, New York, where she was responsible for leading business development initiatives for a sugar-substitute product in the Food Science Group. Later, she also joined Kraft Foods, Chicago as a Brand Manager where she led multiple large cross-functional teams to deliver substantial volume and profits and was responsible for more than $150 million in P&L. She also led the development and implementation of a turnaround strategy for a core brand for the Oscar Mayer division at Kraft Foods that involved process re-engineering, extensive market research and data analysis. Sudeshna grew up in Delhi and completed her schooling from Delhi Public School R.K. Puram. She went on to do a BA with dual major in Architecture and Building Construction from University of Washington, Seattle before pursuing her MBA in marketing from Cornell University.[\/spoiler]","excerpt":"Sudeshna Datta of Absolutdata talks with Analytics India Magazine about the growth of analytics practice and how Absolutdata is poised to capture this opportunity. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: What according to you are currently the most important and specific analytics needs for the industry? [dropcap style=”1″ size=”2″]SD[\/dropcap]Sudeshna Dutta: In today’s dynamic business environment, companies […]","categories":["AI Features"],"tags":["Absolutdata","analytics ceo","analytics leaders","Interviews and Discussions"],"author_name":"Дарья","publish_date":"2012-09-30T05:43:40","publication_year":"2012","word_count":932,"keywords":["business intelligence","Go","API","analytics ceo","programming_languages:R","AI","analytics leaders","RAG","Aim","analytics","GAN","R","Absolutdata","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","API","GAN","business intelligence","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-sudeshna-datta-evp-and-co-founder-at-absolutdata\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10092232,"title":"How Palantir Turned a New Leaf to Profitability","content":"Palantir Technologies, the Silicon Valley analytics firm best known for its surveillance software is turning a new page in its journey. Just as the mass exodus of companies to the cloud triggered the need for cloud transformation platforms, companies clamouring to integrate AI systems will ask for the equivalent of this in the AI space. Palantir promises to fill this gap with their newly launched AIP or Artificial Intelligence Platform which they claim will “empower companies to use LLMs safely and securely.” What does Palantir’s AIP do? The platform is focused on enterprise security – it builds a wall between the company’s private network and private data and what the LLM can or cannot see. Alok Panigrahy, the Head of Channel GTM at Palantir posted a demo of the AIP on LinkedIn saying, “With top-notch guardrails, the AIP ensures regulated and trustworthy AI deployment, enabling adherence to legal and regulatory standards while making LLMs operational in your own network and simplifying the life of operational users.” And Palantir does have a very important point to make here. As quickly as companies are pivoting to AI and integrating LLMs, there will be obvious cracks in the wall. The architecture that these LLMs will run on simply aren’t prepared enough yet. LLMs like ChatGPT are already known to scrape data from companies inviting a myriad of security risks. In a world like this, nothing will potentially be safe anymore – companies trade secrets, financial data, client information. In an interview with CNBC, company chief Alex Karp outlined the risks that organisations adopting LLMs may eventually face. “It’s going to really crush a lot of businesses for two reasons – because if you run an LLM or an algorithm on an architecture that is crumbly or crispy, it won’t work. It will also crush architectures because there will be an inability to create a barrier between the LLM and the decisions that involve ethics and are classified. Our business is built to do that,” he explained. “But as LLMs become specialised in every business, their real value will be at the intersection between your business logic, your business norms and ethics and the LLMs. The people who get all three right will make money. Building stronger infrastructure If the rest of the world wasn’t prepared yet, Palantir was. In a blog posted on April 7th by company chief Alex Karp announcing AIP, he explained how Palantir is perfectly placed to fill this need. “Our software and company were built for this moment. The advent of more generalizable artificial intelligence systems has begun to materially transform and advance the business we founded nearly two decades ago. The deployment of artificial intelligence in these novel contexts will only be made possible by foundational data platforms that enable the imposition of the legal and ethical constraints, as well as the implementation of regulatory requirements, that now govern the use of data across industries and national boundaries. And we have spent two decades building those platforms,” he stated. While Palantir’s rebranding as an ethically aligned entity may be surprising given its history of morally dubious ‘spy-tech,’ it may be the right business move. Last month, the company published a couple of noteworthy articles on AI ethics and why they must move forward beyond being merely performative. There’s a bunch of other partnerships that Palantir has entered into to cushion themselves infrastructure-wise. On 5th April, the company announced it was expanding its partnership with Microsoft Azure to its public sector. All of this was to build Palantir’s integration capabilities for its government and commercial customers on Azure. Palantir Apollo offerings for Edge A few days ago, it also stated that it was doubling down on its investment in deploying software to edge devices while also announcing Palantir Apollo, the company’s solution for edge deployment. Finally profitable Whatever direction it is going in, seems to be the right one. This year, the Peter Thiel-founded firm posted its first ever profitable quarter, three years since it went public. Karp stated during the earnings call that the company was on track to make more money this year and it certainly seems that way. Funnily, as most Big Tech companies abandon AI in hopes to beat the other, Palantir has become the one company that has turned ethics into its moat. In the last blog written by Courtney Bowman, the company’s Global Director of Privacy and Civil Liberties Engineering, the company warned investors against gobbling up the AI hype without any heed. “From self-driving vehicles to radiology and predicting job success based on candidate video snippets, there is a growing disillusionment with AI snake oil, alongside an increasing need to discover the credible bedrock underneath the sands of an AI hype cycle. In practice, this means examining what works, discarding what doesn’t, and recentering our moral frameworks around the contexts and whole-of-domain challenges of operationalized AI and away from vain musings about paper clips and trollies,” she said.","excerpt":"Palantir Technologies, the Silicon Valley analytics firm best known for its surveillance software is turning a new page in its journey.","categories":["AI Highlights"],"tags":["palantir"],"author_name":"Poulomi Chatterjee","publish_date":"2023-04-24T18:30:00","publication_year":"2023","word_count":826,"keywords":["Go","ChatGPT","artificial intelligence","AI","R","BERT","Aim","analytics","Rust","palantir","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","ChatGPT","Aim","Azure","R","Go","Rust","BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-palantir-turned-a-new-leaf-to-profitability\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10124529,"title":"Vertiv and Ballard Partner on Hydrogen Fuel Cells for Data Center Backup Power","content":"Vertiv and Ballard Power Systems a PEM fuel cells provider, have announced a strategic technology partnership focused on developing hydrogen fuel cell backup power solutions for data centers and critical infrastructure. The scalable systems will range from 200kW to multiple megawatts. The companies have successfully demonstrated a proof of concept at Vertiv’s facility in Ohio, integrating Ballard fuel cell power modules with a Vertiv Liebert EXL S1 uninterruptible power system (UPS). Initial validations and tests showed the zero-emission backup power system operating successfully as part of an uninterruptible power architecture. “As the soaring increase in data usage is driving up power demand and expansion of data centre capacity globally, the need to effectively manage electricity consumption and the carbon footprint of this energy-intensive sector is critical to achieve net-zero targets,” said Nicolas Pocard, vice president of marketing and strategic partnerships at Ballard. The demonstrated system, called Vertiv Power Module H2, integrates two Ballard PowerGen 200kW fuel cell cabinets along with cooling, power conditioning, and hydrogen storage infrastructure. It is part of a 1 MW microgrid solution at Vertiv’s Customer Experience Center that also includes solar PV and battery energy storage. Viktor Petik, vice president of Vertiv infrastructure solutions, said the adoption of AI and high-performance computing is driving demand for eco-friendly power solutions focused on zero-carbon and low-carbon alternatives. “The successful fuel cell proof-of-concept with Ballard provides a viable option for customers strengthening their data center sustainability strategy, and those moving to a future-ready Bring Your Own Power model,” he stated. The Vertiv Power Module H2 solution provides a rapidly deployable and scalable power infrastructure for new and existing data centers. It offers advantages such as zero greenhouse gas emissions, low noise, rapid power response, low maintenance, extended backup duration, optimized footprint, and multi-megawatt scalability. The partnership aligns with Vertiv’s “One Vertiv, One World” sustainability strategy. The companies plan to demonstrate the fuel cell powered backup systems at the Smarter E Europe 2024 energy industry exhibition in June","excerpt":"The companies have successfully demonstrated a PoC of the scalable systems ranging from 200kW to multiple MW at Vertiv’s facility in Ohio.","categories":["Deep Tech"],"tags":["AI Data Center","Data Center"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-25T13:34:27","publication_year":"2024","word_count":328,"keywords":["AI Data Center","API","Data Center","programming_languages:R","AI","Scala","RAG","BERT","llm_models:BERT","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","RAG","R","Scala","API","BERT","llm_models:BERT","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/vertiv-and-ballard-partner-on-hydrogen-fuel-cells-for-data-center-backup-power\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":49018,"title":"AI-Startup Fireflies Raises $5 Mn To Bring More Effective ML Solutions In Its Enterprise Offerings","content":"Fireflies.ai, the Hyderabad and San Franciso-based AI startup announced having raised $5M seed round led by Canaan and other individual investors from companies such as Slack, Salesforce, Dropbox, and Skype. They are Sandhya Venkatachalam, former vice president of corp development (Skype and Cisco), April Underwood, chief product officer(Slack), Armando Mann, vice president of sales (Dropbox), Susan Kimberlin, director of search (Salesforce), Bill Macaitis, chief marketing officer (Slack). The startup which was founded by alumni of MIT and University of Pennsylvania, Sam Udotong and Krish Ramineni respectively, Fireflies makes an AI-powered enterprise assistant that automatically records and transcribes meetings. The founders believe that the way companies work is changing drastically both in terms of tools and where the teams are located. There may many times be communication breakdown owing to different work locations that may further lead to a decline in productivity. Different mediums of conversations such as Slack messages, emails and phone further make it difficult. It is often the responsibility of the meeting owner to get the information collated and perform manual data entry into CRMs and project management systems. Fireflies.ai aims to bridge this gap by using voice-based AI assistant for meetings and bring features such as taking notes, improving collaboration and collating information. The startup uses neural networks and deep learning techniques for its meeting recorder. It uses NLP to interpret voice conversations, extract information such as action items, and more. As the reports suggest, the startup intends to use the funding amount to scale-up their engineering in Hyderabad and Bengaluru, bring in more machine learning solutions, add additional language support and expand its customer base in India.","excerpt":"Fireflies.ai, the Hyderabad and San Franciso-based AI startup announced having raised $5M seed round led by Canaan and other individual investors from companies such as Slack, Salesforce, Dropbox, and Skype.  They are Sandhya Venkatachalam, former vice president of corp development (Skype and Cisco), April Underwood, chief product officer(Slack), Armando Mann, vice president of sales (Dropbox), […]","categories":["AI News"],"tags":["Deep Learning Techniques","salesforce crm"],"author_name":"Srishti Deoras","publish_date":"2019-10-29T17:00:39","publication_year":"2019","word_count":271,"keywords":["funding","machine learning","AI","neural network","NLP","Aim","deep learning","salesforce crm","ViT","Deep Learning Techniques","R","startup"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","NLP","Aim","R","ViT","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-startup-fireflies-raises-5-mn-to-bring-more-effective-ml-solutions-in-its-enterprise-offerings\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10066851,"title":"A guide to clustering with OPTICS using PyClustering","content":"Automatic classification techniques, also known as clustering, aid in revealing the structure of a dataset. PyClustering is a Python and C++ open-source data mining package that offers a variety of clustering techniques and approaches, including bio-inspired oscillatory networks. PyClustering is primarily concerned with cluster analysis to make it more accessible and clear to users. This article will cover different types of modules offered by pyclustering and implement an algorithm that pyclustering supports. Following are the topics to be covered. Table of contents About PyClusteringModules in PyClusteringAbout OPTICSHow to Implement OPTICS using Pyclustering? Let’s start by talking about PyClustering as a package About PyClustering The PyClustering library is a data mining package for Python and C++ that focuses on cluster analysis. To ensure optimal efficiency, the C++ half of the library is utilised for processing by default. This is especially true for algorithms based on oscillatory networks, the dynamics of which are regulated by a set of differential equations. If a C++ compiler is not found, PyClustering defaults to pure Python implementations of all kernels. PyClustering uses the NumPy module for array operations to improve the speed of Python implementations. PyClustering provides efficient, parallel C++14 clustering solutions. Threading is given via a single thread of operation on most platforms, with the Parallel Patterns Library utilised for Windows. Because of the standardisation of these threading libraries, PyClustering is straightforward to implement into existing projects. PyClustering’s main Python requirements are NumPy and SciPy, with MatPlotLib and Pillow necessary for visualisation functionality. The visualisation feature includes 2D and 3D plots of cluster embeddings, picture segments, and graphs of synchronisation processes in the case of oscillatory networks. Are you looking for a complete repository of Python libraries used in data science, check out here. Modules in Pyclustering There are five modules in pyclustering clusters, n-net, g-colour, and containers. Clustering Algorithms Algorithms and methods are located in the Python module “pyclustering.cluster”. There is a list of clustering algorithms which is supported by Pyclustering. Agglomerative BANG BIRCH BSASCLARANSCLIQUECUREDBSCANElbow EMA GA – Genetic AlgorithmHSyncNetK-Means and K-Means++ K-Medians and K-Medoids MBSAS OPTICSROCKSilhouetteSOM-SCSyncNet and Sync-SOMTTSASX-Means Oscillatory Networks and Neural Networks The Oscillatory Neural Networks (ONN) and Neural Networks (NN) are located in the Python module “pyclustering.nnet”. There is a list of networks supported by Pyclustering. An oscillatory neural network (ONN) is a type of artificial neural network in which the neurons are connected to oscillators. Oscillatory neural networks are inspired by the occurrence of neural oscillations in the brain and are closely related to the Kuramoto model. To detect pictures, oscillatory neural networks have been developed. Chaotic Neural Network(CNN)Pulse-Coupled Neural Network (PCNN)Self-Organized Map (SOM)Oscillatory neural network based on Kuramoto model (Sync)Oscillatory neural  network for pattern recognition (SyncPR)Oscillatory neural network for image segmentation (SyncSegm)Oscillatory neural network based on Landau-Stuart equation and Kuramoto model (fSync)Oscillatory neural network based on Hodgkin-Huxley model (HHN)Hysteresis Oscillatory Network Local Excitatory Global Inhibitory Oscillatory Network (LEGION) Graph Coloring Algorithms Graph Coloring Algorithms are located in the Python module “pyclustering.gcolor”. Graph colouring is a subset of graph labelling; it involves the assignment of labels to nodes of a graph subject to specific criteria. It is a method of colouring the vertices of a graph so that no two neighbouring vertices are the same colour. Three different types of techniques are used. DSatur Hysteresis GColorSync Containers There are two types of containers used by Pycluster to build a tree in n-dimensional space to search data points for the clustering. In python, it is present in a “pyclustering.container”: KD Tree CF Tree About OPTICS Many real-data sets have the critical trait that their underlying cluster structure cannot be described by global density parameters. To identify clusters in various locations of the data space, very different local densities may be required. A density-based clustering approach is Ordering Points To Identify Clustering Structure (OPTICS). The essential principle of density-based clustering is that the neighbourhood of a certain radius must contain at least a minimal number of objects (MinPts), i.e. the cardinality of the neighbourhood must surpass a threshold. OPTICS generates a database order, storing the core distance and a reasonable reachability distance for each item. How does this work? OPTICS works similar to DBSCAN; it will start with retrieving the maximum distance neighbourhood of the object passed from the main loop OPTICS, sets its reachability-distance to UNDEFINED and determines its core distance. The object is then checked for the core object attribute, and if it is not a core object at the generating distance, control is simply returned to the main loop OPTICS, which chooses the next unprocessed object in the database. Otherwise, if the item is a core object at a distance less than the maximum distance, it iteratively gathers directly density-reachable objects concerning the maximum distance and a minimum number of objects. Objects that are directly density-reachable from a present core object are added to the OrderSeeds seed list for future growth. OrderSeeds objects are arranged by reachability-distance to the closest core object from which they have been directly density reachable. In each cycle, the current object in the seed list with the shortest reachability distance is chosen. This object’s maximum distance neighbourhood and core distance are determined. The item is then simply saved together with its core distance and current reachability distance. If the existing object is a core object, more expansion candidates may be added to the seed list. How does OPTICS differ from DBSCAN? The OPTICS clustering approach consumes more memory since it uses a priority queue (Min Heap) to select the next data point in terms of Reachability Distance that is closest to the point presently being processed. It also needs greater computer resources since nearest neighbour queries in DBSCAN are more complex than radius queries.The OPTICS clustering algorithm does not require the epsilon parameter and is merely included in the pseudo-code above to decrease the time required. As a result, the analytical process of parameter adjustment is simplified.OPTICS does not divide the input data into clusters. It just generates a Reachability distance plot, and it is up to the coder to analyse it and cluster the points accordingly. How to Implement OPTICS using Pyclustering? Let’s install the Pyclustering to get started with the OPTICS algorithms. ! pip3 install pyclustering Import necessary libraries import random from pyclustering.cluster import cluster_visualizer,cluster_visualizer_multidim from pyclustering.cluster.optics import optics, ordering_analyser, ordering_visualizer from pyclustering.utils import read_sample, timedcall from pyclustering.samples.definitions import SIMPLE_SAMPLES,FCPS_SAMPLES,FAMOUS_SAMPLES Using the sample data provided by the Pyclustering itself and performing OPTICS algorithm to cluster the raw data and analyse the results. sample = read_sample(FCPS_SAMPLES.SAMPLE_TWO_DIAMONDS) radius_cluster = 0.2 num_neighbors = 10 optics_model = optics(sample, radius_cluster, num_neighbors) optics_model.process() clusters = optics_model.get_clusters() noise = optics_model.get_noise() ordering = optics_model.get_ordering() Defining the radius for the cluster and the number of neighbours is necessary for the algorithm. The clusters formed by the OPTICS algorithm are stored by using “.get_clusters”. Similarly, information about the input data set’s noise and clustering ordering is extracted. The data-set clustering ordering contains information on the internal clustering structure following the connectivity radius. Analysing the clusters formed with the sample. Plotting the histogram of the number of clusters formed to check the distribution and spread of data points in each cluster. plots = cluster_visualizer() plots.append_clusters(clusters, sample) plots.show() analyser_cluster = ordering_analyser(ordering) ordering_visualizer.show_ordering_diagram(analyser_cluster, 2) As we can observe that the cluster is approximately the same as the sample. The clusters also maintain the shape of the diamond just like the sample. As mentioned above the radius of the cluster is 0.2 and in the representation, we can observe that the radius of the cluster is approximately 0.2. There are a total of 2 clusters formed by the algorithm and by the distribution it could be said they are perfectly formed. Let’s implement this on multidimensional data since the previous data was two dimensional.  Everything in the code would be the same except the visualisation part. As this is multidimensional data, I need to use “cluster_visualizer_multidim()”  instead of “ cluster_visualizer()”. sample = read_sample(FAMOUS_SAMPLES.SAMPLE_IRIS) radius_cluster = 0.5 num_neighbors = 10 optics_model = optics(sample, radius_cluster, num_neighbors) optics_model.process() clusters = optics_model.get_clusters() noise = optics_model.get_noise() ordering = optics_model.get_ordering() plots = cluster_visualizer_multidim() plots.append_clusters(clusters, sample) plots.show() analyser_cluster = ordering_analyser(ordering) ordering_visualizer.show_ordering_diagram(analyser_cluster, 3) In the above representation, there are six plots of clusters which represents six different dimensions of the data. One can also visualise a single dimension by the following code. plots = cluster_visualizer_multidim() plots.append_clusters(clusters, sample) plots.show(pair_filter=[[0,3]]) Conclusion PyClustering is a data mining package which supports a great list of clustering algorithms which make it flexible to operate. Since it uses Numpy for mathematical operation and parallel processing makes the processing faster. With this article, we understood the package and modules offered by the package with the implementation of OPTICS on both two dimensional and multidimensional data. References Link to the above codeDocumentation of PyclusteringGitHub repository of PyclusteringAbout OPTICS algorithm","excerpt":"OPTICS is a density-based clustering algorithm offered by Pyclustering.","categories":["Deep Tech"],"tags":["clustering algorithms","Data Mining","data mining tools","Neural Networks","numpy","scipy"],"author_name":"Sourabh Mehta","publish_date":"2022-05-12T16:00:00","publication_year":"2022","word_count":1457,"keywords":["data science","scipy","NumPy","Go","AI","neural network","Data Mining","numpy","Python","Ray","data mining tools","C++","clustering algorithms","Matplotlib","R","Neural Networks"],"extracted_tech_keywords":["AI","neural network","data science","Ray","NumPy","Matplotlib","Python","R","Go","C++"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-guide-to-clustering-with-optics-using-pyclustering\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10085376,"title":"At Davos 2023, Tech Leaders Debate the Future of Generative AI","content":"Generative AI was one of the hot topics of the World Economic Forum’s (WEF) Annual Meeting at Davos, where ChatGPT became the ‘goblin word’ of the conference. Even in the shivering cold of  -7°C, tech representatives could not stop hyping the immense potential of AI advancements in every field. The origins of generative AI were with the advent of GANs. In an exclusive interaction with AIM, AI stalwart Bengio raved about the distance that generative models had come since their emergence in 2014. On the recent advancements made by text-to-image generators like OpenAI’s DALL.E and StabilityAI’s Stable Diffusion, Bengio stated, “One of the things that impressed me the most is the progress in generative models.” Read more: What Excites Yoshua Bengio about the Future of Generative AI At the Davos conference, Microsoft CEO Satya Nadella discussed the stir ChatGPT, the intelligent chatbot from OpenAI, has generated. He said that Microsoft intends to open up access to its cloud-based Azure OpenAI service so that anyone may use its AI tools for business and that customers can access ChatGPT through Azure. Furthermore, he stated that the ChatGPT API would be released soon. Although Nadella acknowledged that generative AI advancements could be potentially dangerous, he also said they could help resolve problems rather than create new ones. Another AI stalwart, Meta AI chief Yan LeCun told AIM that data systems are entertaining but not really useful. “To be useful, they have to make sense of real problems for people, help them in their daily lives as if they were traditional assistants completely out of reach,” he added, painting the real picture. Talking about ChatGPT, LeCun added that many individuals are working on language models using slightly different methods. He said there are three to four companies producing GPT-X-like models. “But, they [OpenAI] have been able to deploy their systems in such a way that they have a data flywheel. So, the more feedback they have, the more the feedback they get from the system, and later adjust it to produce better outputs,” he explained. Matthew Prince, chief executive officer of Cybersecurity company Cloudflare, stated that generative AI could serve as a “really good thought partner” or a junior coder. He said Cloudflare was writing code on its “Workers” platform using such tech. Cloudflare is also looking into how such technology could help its free-tier clients get answers to questions more quickly. According to Alex Karp, CEO of Palantir Technologies, which develops software that helps governments track the movements of armies and businesses examine their supply networks, among other things, such AI might be used in the military. “The idea that an autonomous thing could generate results is useful for war,” Karp stated. However, he added that the nation with the fastest AI development would “define the law of the land,” and it is important to consider how technology will affect any war with China. In other topics, CEO Cristiano Amon of chip maker Qualcomm said that internet-connected glasses are set to make a dramatic comeback. According to him, connected glasses will eventually replace smartphones as the go-to computer for daily chores as computing grows and the metaverse becomes more pervasive in daily life. Although there are currently connected glasses on the market, the technology has yet to become widely adopted. A notable setback is the Google Glass technology. “The whole tech trend is the merging of digital and physical spaces; within the decade, it’s going to be as big as phones,” he said. However, Amon predicted that the metaverse’s speed would make the technology more resilient this time. The World Economic Forum’s (WEF) Annual Meeting has started in Davos, Switzerland, with the central theme “Cooperation in a Fragmented World”. The conference is scheduled from January 16th to January 20th, 2023.","excerpt":"CEO calls Generative AI the “Junior Coder”, while Qualcomm projected that connected-internet glasses would replace smartphones in the coming decade","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Tool","Data Scientist","davos","Deep Learning","Google","Machine Learning","Python"],"author_name":"Shritama Saha","publish_date":"2023-01-18T14:23:49","publication_year":"2023","word_count":627,"keywords":["ChatGPT","Meta AI","TPU","OpenAI","AI","R","Machine Learning","RAG","Python","Aim","generative AI","Google","Deep Learning","AI Tool","Data Scientist","Azure","davos","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Meta AI","Aim","RAG","Azure","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/at-davos-2023-tech-leaders-debate-the-future-of-generative-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10051721,"title":"JupyterLab Desktop App vs JetBrains DataSpell","content":"The market for data science IDEs isn’t overly crowded. On the one hand, there’s Jupyter for maximal interactivity, and on the other, there’s PyCharm for a professional atmosphere. Text editors such as VSCode can also be used; however, they are time-consuming. Dataspell is a new entry on the block, an IDE designed specifically for data scientists. Let’s have a look at JupyterLab and JetBrains Dataspell’s functionality. What is JupyterLab JupyterLab is an open-source web application, described as “the cross-platform standalone application distribution of JupyterLab. “It is a self-contained desktop application that includes a Python environment and numerous prominent Python libraries that are pre-configured for use in scientific computing and data science operations.” Previously, JupyterLab was kept within a web browser environment, however, with the latest improvements, it is now a standalone application. What is JetBrains DataSpell JetBrains announced the release of new integrated development environments (IDEs) for data scientists who construct AI models using a variety of programming languages, including Python. The new IDEs will be offered to data scientists via an early access programme, enhancing the experience of regular notebooks. JetBrains DataSpell will provide data scientists with enhanced experience for managing and writing code. One can sign up for it here. However, the new IDE will not be a replacement for Jupyter notebooks but rather work alongside them on local PCs. Jupyter notebooks are augmented with folding tracebacks, intelligent Python code aid, interactive tables, and out-of-the-box tables of contents, all of which make it easier to adhere to best practices. Andrey Cheptsov, product manager at JetBrains, stated that “There has never been a dedicated IDE for data science in the Python ecosystem. Individuals engaged in data research were required to use editors, developer integrated development environments or standalone Jupyter notebooks”. He continued, “JetBrains anticipates that DataSpell will provide a more practical and efficient environment for working with data in general. The developer has indicated that features relating to data manipulation would be prioritised.” Indeed, JupyterLab supports different languages and enables users to choose their display language using the language pack included with Jupyter. JupyterLab has switched to Jupyter Server as of the third version, a new Jupyter project built on the server element of the traditional Notebook server. This package includes a command palette that appears as a floating window on top of the JupyterLab workspace, allowing users to rapidly launch a command while leaving the sidebar closed or navigating between sidebar panels. Platform Compatibility JupyterLab App is compatible with Linux, macOS, and Windows operating systems based on Debian and Fedora. Each platform has a one-click installer. Jupyter’s current release includes a new visual debugger, as well as new methods for publishing and installing extensions via Python pip or Conda packages. Additionally, the company has enabled the installation of these without the need to construct JupyterLab with Node.js. He added that JetBrains DataSpell works with both local Jupyter notebooks and remote Jupyter, JupyterHub, and JupyterLab servers. Additionally, DataSpell has Python scripting capabilities in addition to various tools for manipulating and viewing static and interactive data. Along with Python, JetBrains DataSpell has rudimentary support for the R programming language, with additional data science languages being added in the future. JetBrains’ new integrated development environment (IDE) complements rather than replaces Jupyter notebooks, Cheptsov explained. Working Experience The Jupyter notebook experience has been enhanced with intelligent Python coding aid, an out-of-the-box table of contents, folding tracebacks, and interactive tables. Cell outputs are compatible with both Markdown and JavaScript. JetBrains’ DataSpell is geared toward the growing ranks of business data scientists, as opposed to other types of professionals who work with computer code. The data scientist team can optimise their workflow and deploy a small number of AI models successfully. They leverage industry-leading tools to navigate large information in less time, making it easier to work on numerous projects concurrently. The tools can assist businesses in attracting and retaining data scientist talent while taking into account a variety of aspects, including salary. Moreover, the tools are capable of writing complicated code. The goal is to increase data scientists’ productivity in order to launch numerous AI initiatives while increasing income and lowering costs. Additionally, the digital business transformation initiative can assist in navigating the data more readily without interfering with the code. Conclusion The overall impression of DataSpell is favourable since it extracts from PyCharm all of the critical functionality required for data science applications. DataSpell has significantly improved the Notebook experience. Additionally, DataSpell supports R, and the JetBrains team is working to improve their support for the R language and support for other data science-related languages, such as Julia. For all of these reasons, data scientists will undoubtedly give DataSpell a try once it is officially released.","excerpt":"A comparison of the JupyterLab Desktop App with the DataSpell IDE’s core features.","categories":["AI Trends"],"tags":["Pycharm"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-10-18T11:00:00","publication_year":"2021","word_count":784,"keywords":["data science","Go","TPU","AI","RAG","Python","Pycharm","JavaScript","Jupyter","R","Java"],"extracted_tech_keywords":["AI","data science","Jupyter","RAG","TPU","Python","R","JavaScript","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/jupyterlab-dataspell\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":33262,"title":"10 Breakthroughs In Artificial Intelligence That Skyrocketed Its Popularity This Decade","content":"Artificial Intelligence and its associated technologies have become the buzzwords of companies, individuals and even governments. This begs the question; How did AI gain the prominence and attention it has today? What forced companies to stop and take notice of a technology that has effectively changed every walk of life? According to Google Trends, this phenomenon occurred around the beginning of 2016, with indexed search volume results almost doubling over just one month. However, the building blocks for the widespread adoption of AI had already been set. IBM Watson Vs Google DeepMind (2011-17) In 2011, IBM had debuted their Watson AI on the reality TV show Jeopardy!, where it was pitted against two of the best players in the world. For those unaware, Jeopardy! relies heavily on language-based processing, a feat that has historically been difficult for AI. However, Watson handily beat the other two participants, one of who had never been beaten by any human competitor. This marked the true birth of AI as we know it today, as it progressed from close-bounded problems to more ‘intelligent’ procedures such as Natural Language Processing. The advancement of IBM was then noted by Google, who proceeded to acquire an AI-based startup known as DeepMind in early 2014 to fuel their AI expansion plans. This marked the age of AI where only big companies had the resources, tools, and knowledge required to create AI solutions. Soon after their acquisition of DeepMind, Google began training an algorithm designed to beat what is known commonly as the world’s most complex game; Go. Being much more complex than chess despite having simpler rules, the number of board positions in the game has been calculated to be around 2 x 10^170. This is an incredibly large number, which only increases in complexity as there are more pieces on the board. However, Google’s algorithm called AlphaGo first beat the European Go champion in 2015. This was followed by its iteration then beating the Korean Go champion in 2016. The most advanced iteration was then utilised to beat the World Go champion in 2017, thus marking the end of human dominance over Go that has existed since over 2000 years. Rise Of AI Awareness And Accessibility (2016-18) As the years passed, AI grew better at games that human players have dominated since their genesis. Poker, one of the most infamous incomplete information games in the world, held as a last bastion of humankind due to its inbuilt requirement of judging other humans. However, in 2017, Libratus, an AI built to play poker, defeated some of the best players in the world in 2017. In the same year, many prominent AI thinkers such as Elon Musk, Stephen Hawking, Steve Wozniak, and one of the founders of DeepMind, Mustafa Suleyman, signed an open letter addressed to the UN regarding the use of AI in autonomous weapons. Reportedly, this would bring about the creation of the “third age of war”, as the creation of gunpowder and nuclear weapons brought about the creation of the first and second ages. The world took note of this power, and its access to it was made easier by the transition of one of Google’s tools, TensorFlow, to open-source in 2016. This, coupled with advancements made in CPU and GPU technology, opened up the field to enthusiasts everywhere. OpenAI, a company founded by Elon Musk, also dipped its feet into the age-old practice of defeating humans. A bot was trained to play Dota 2, an extremely complex game with the requirement for a large amount of information and lightning-quick reaction times. This is arguably more complex than Go, with OpenAI’s bot beating nine of the world’s best players in closed games and appearing in a championship to beat one of the world’s best players. An honorable mention also goes out to AI that is being delivered even today to Android phones around the world. Google Assistant gained prominence in around 2016, with its use of NLP removing barriers between humans and computers. Moreover, advancements in Google’s WaveNet algorithm have also improved the speaking capabilities of the software, making it sound as close to human speech as possible. AI For Science, Governments, And Society (2018-?) This mix of a technology that seemed to be smarter than us and that held the threat of the extinction of our species over our heads seemed to excite everyone on the planet. In 2017, many prominent governments and companies jumped on the AI bandwagon, driving public interest in the field. This is evidenced by the rising indexed search volume indicator on Google Trends. More than the general population, enterprises, and administrations, the scientific community seems to have caught on to the AI train late. On the Scopus scholarly database, an open-access database used by scientists, the term ‘AI’ moved from 13th most searched to 4th in 2018, with terms such as ML and DL breaking into the top 20. This represents not only the update of the scientific community to one of the most groundbreaking technologies but also represents the fruits of research papers sponsored in 2016 and 2017, a high point for AI research. 2018 was undoubtedly the year of AI, as interest in all spheres for solutions that utilised AI reached a fever pitch. It received a lot of attention from media and governments alike, as seen by the Google Trends data which stayed near the ‘100’ mark for a majority of the year. Post this increase, the search interest for AI has been consistently high. The spike in 2017 showed that interest continued to increase as advancements continued to be made. Outlook As governments around the world continue to make positive moves forward into the AI field, the warnings of those who warned against giving algorithms too much power. Due to the general consensus that an Artificial SuperIntelligence is some time in the not-so-distant future, humans seem to be riding the AI trend train to the last stop.","excerpt":"Artificial Intelligence and its associated technologies have become the buzzwords of companies, individuals and even governments. This begs the question; How did AI gain the prominence and attention it has today? What forced companies to stop and take notice of a technology that has effectively changed every walk of life? According to Google Trends, this […]","categories":["AI Trends"],"tags":["ai Trends","Elon Musk","Google","Google Deepmind","OpenAI"],"author_name":"Anirudh VK","publish_date":"2019-01-11T06:13:00","publication_year":"2019","word_count":990,"keywords":["Go","artificial intelligence","OpenAI","AI","ML","ai Trends","Elon Musk","NLP","Google Deepmind","Google","GAN","TensorFlow","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","ML","NLP","OpenAI","TensorFlow","R","Go","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-breakthroughs-in-artificial-intelligence-that-skyrocketed-its-popularity-this-decade\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":38553,"title":"Top Big AI Announcements From Facebook’s F8 Developer Conference","content":"Source: Facebook Earlier this year Facebook demonstrated how serious they are about AI by open sourcing their NLP toolkit LASER. Now at the recently concluded 2-day event F8 Developer Conference held in San Jose,California, Facebook dished out more updates for the machine learning developers. At Facebook, ML is used to discover patterns in code and build tools that improve developer productivity through code search, code recommendation and automatic bug fixing. An AI-enabled testing adaptive approach is also deployed  to optimize products, infrastructure, machine learning models, marketing campaigns. On the second day of the event, Chief Technology Officer Mike Schroepfer along with his AI team, talked about the AI tools we’re using to address a range of challenges across many products of Facebook. Here are a few highlights: Computer Vision Facebook started out very early in the domain of face recognition with machine vision. Its auto-tagging photo feature uses convolutional neural networks(CNNs). And these neural networks got better as people shared billions of photos over more than half a decade. Now these CV systems have recognised progressively more image components over the years and can now perform detection of objects in both the foreground and the background with a single network. This results in better understanding of a photo’s overall context, as well as more computationally efficient image recognition. This was achieved using a new approach to object recognition, called a panoptic feature pyramid network (Panoptic FPN), which enables instance segmentation tasks (for the foreground) and semantic segmentation tasks (for the background) at the same time, on a single, unified neural architecture. Content Curation Finding policy violations within video is orders of magnitude harder than in photos. In many videos, however, only a few clips have information that’s salient to a specific task, such as detecting bullying, and the rest are either redundant or irrelevant. So a new approach where hashtagged videos functioned as weakly supervised data, meaning training examples whose labels had been applied by people, but without the precision of full supervision. This led to a 5.1 percent improvement over the previous state of the art’s 77.7 percent accuracy. The majority of the systems today rely on supervised training and the lack of enough training data makes these models not so reliable. With these new approaches Facebook pushes for self-supervision with support from AI pioneers like Yann LeCun who also is the Chief AI Scientist at Facebook. PyTorch 1.1 PyTorch is one of the hottest products to come out of Facebook. Facebook uses PyTorch 1.0 end-to-end workflow for building and deploying translation and natural language processing (NLP) services at scale. These systems provide nearly 6 billion translations a day for applications such as real time translation in Messenger and, as the foundation of PyText, power complex models that rely on multitask learning in NLP. Companies like Airbnb and Microsoft also leverage PyTorch to build conversational AI applications and other cognitive services. The ongoing evolution of PyTorch serves as an example of the power of open, community-led development in AI. Key Features: Improved performance for common models such as CNNs Added support for multi device modules including the ability to split models across GPUs while still using Distributed Data Parallel (DDP) PyTorch-BigGraph: PBG is a distributed system for creating embeddings of very large graphs with billions of entities and trillions of edges. Open Source Tools BoTorch: BoTorch is a research framework built on top of PyTorch to provide Bayesian optimization, a sample-efficient technique for sequential optimization of costly-to-evaluate black-box functions. Ax: Ax is an ML platform for managing adaptive experiments. It enables researchers and engineers to systematically explore large configuration spaces in order to optimize machine learning models, infrastructure, and products. BigGAN-PyTorch: This is a full PyTorch reimplementation that uses gradient accumulation to provide the benefits of big batches on as few as four GPUs. Curve-GCN: A real-time, interactive image annotation approach that uses an end-to-end-trained graph convolutional network (GCN) The field of NLP is innovating every other day thanks to the constant effort of tech giants like Google, Microsoft, Facebook and Amazon. One thing common with these tech giants is their willingness to open source their innovations. Their belief in accelerated innovation through transparency has started to see fruition in the form of diversified real world applications from homepods to chatbots. The developers team at Facebook have also made significant changes to Instagram and Whatsapp as well. Check more details here.","excerpt":"Earlier this year Facebook demonstrated how serious they are about AI by open sourcing their NLP toolkit LASER. Now at the recently concluded 2-day event F8 Developer Conference held in San Jose,California, Facebook dished out more updates for the machine learning developers. At Facebook, ML is used to discover patterns in code and build tools […]","categories":["AI Trends"],"tags":["Facebook AI","policy gradient","Pytorch"],"author_name":"Ram Sagar","publish_date":"2019-05-02T09:15:39","publication_year":"2019","word_count":732,"keywords":["Pytorch","machine learning","Facebook AI","AI","neural network","PyTorch","ML","chatbots","image recognition","computer vision","RAG","NLP","policy gradient"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","NLP","computer vision","PyTorch","RAG","chatbots","image recognition"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ai-announcements-from-facebooks-f8-developer-conference\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10114809,"title":"The Brain Behind Oracle Cloud","content":"“I think the biggest transformation I’ve seen is not just OCI but Oracle itself becoming a lot more cloud- and operations-focused,” said Pradeep Vincent, chief technical architect, Oracle, in an exclusive interview with AIM, adding that operations are one of the biggest value propositions of the cloud. Vincent joined the company a decade ago and is one of the founding members of Oracle Cloud Infrastructure (OCI). He has been involved in the design and implementation of the cloud infrastructure, responsible for the engineering architecture group, and plays a key role in driving the development of OCI. Multi-Cloud Approach Vincent believes that OCI is pretty different from the competitors out there. “Our goal is to make it easy for customers to use multiple clouds, period,” he said, explaining that a key part of this is their ‘distributed cloud strategy’, putting the cloud where customers want it. “There are a few different ways in which we are going about it, one of them is multi-cloud, with Oracle Database@Azure being one. We’re super excited about that,” he said. Oracle, last year, announced Oracle Database@Azure, which delivers Oracle database services running on OCI inside Azure datacenters and gives customers more flexibility in where they run their workloads. “I worked on the engineering architecture behind the scenes for that. It’s truly impressive,” said Vincent, adding that they essentially took OCI itself, creating a small OCI site inside the Azure Data Centre and connecting it to OCI. Simultaneously, it was wired directly to the Azure network. It is only a matter of time before OCI is integrated with GCP and AWS. “Cloud should be open,” said chief technology officer Larry Ellison. OCI’s Architecture Vincent is confident in the networking architecture of OCI. “I believe our super clusters are exceptionally powerful. Customers frequently tell us that when they try other cloud providers, including OCI, they observe a significant difference in networking technology,” he said. He further explained that good networking facilitates proper utilisation of GPUs. “GPUs are very pricey. If you think about LLM training, many of them actually run like 40 to 60% utilisation of the GPUs,” he said, adding that although networking comes with a cost, it provides a 10x value in terms of GPU savings. OCI Stands for Security “OCI is secure by default, which means it embeds security features as a built-in aspect. It’s not like buying a product where you have to add an extra pack for security, here, it’s integrated from the beginning,” said Vincent. OCI provides a variety of features to help you secure your network, such as subnet network filtering, firewalls, and security lists. Moreover, OCI offers a variety of data encryption options, both at rest and in transit. This helps to protect your data from unauthorized access, even if it is intercepted. Vincent explained that Oracle doesn’t make a network public by default. Users have the option to make it public if they choose to. However, he cautioned that a common issue arises when customers unknowingly leak data into the public cloud by creating public buckets. “Our security story is not just about infrastructure or apps but goes end-to-end. Advanced security functionalities like Identity & Access Management and Cloud Guard cut across most threats. And that’s a huge differentiator as far as this is concerned,” said Vincent. Generative AI in OCI Oracle recently embedded generative AI capabilities into the complete SaaS suite, which include applications like ERP, HCM, SCM, and CX. Additionally, OCI offers models from Cohere and Meta for various tasks such as writing, summarisation, analysis, and chat, without requiring extensive training from scratch. “The way I look at generative AI is that it is a very novel and creative way to use data. But it’s not necessarily accurate on its own every single day,” said Vincent, adding that generative AI will enhance use cases but not necessarily replace all. Speaking from Oracle’s perspective, he said the company is focusing on expanding data centres to meet the demand for generative AI services. “We offer many unique services, including cluster networks with support for remote direct memory access (RDMA),”  said Vincent. He further mentioned that Oracle internally uses generative AI for customer support. When asked about his motivation, he shared, “There’s a lot of customer problems to be solved, lots of innovation that’s happening. So I’m excited and privileged to be part of it.”","excerpt":"Pradeep Vincent, joined Oracle back in 2014 and is one of the founding members of OCI.","categories":["AI Features"],"tags":["Oracle"],"author_name":"Siddharth Jindal","publish_date":"2024-02-29T18:34:04","publication_year":"2024","word_count":726,"keywords":["Go","GCP","AWS","AI","cloud_platforms:AWS","R","innovation","Oracle","Aim","generative AI","Azure"],"extracted_tech_keywords":["AI","generative AI","Aim","AWS","Azure","GCP","R","Go","innovation","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-brain-behind-oracle-cloud\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10110245,"title":"Why Isomorphic Labs Partnered with Novartis and Eli Lilly","content":"Isomorphic Labs, a London-based AI drug discovery startup, spun out of Google’s DeepMind unit just over two years ago, recently announced key partnerships with two of the world’s largest pharmaceutical companies — Eli Lilly & Co. and Novartis AG. The deals are said to have a combined value of close to $3 billion. Isomorphic Labs will partner with Lilly for small molecule therapeutics, receiving an upfront payment of $45 million and up to $1.7 billion in milestone payments, excluding royalties. Similarly, with Novartis, Isomorphic Labs gets a $37.5 million upfront payment, funding for select research costs, and up to $1.2 billion in milestone payments, along with royalties on net sales. Founded in 2021 by Demis Hassabis, co-founder of DeepMind, Isomorphic Labs primarily focuses on small molecule therapeutics, which are easier to manufacture and deliver. The company utilises AlphaFold, an AI system developed by DeepMind to predict a protein’s 3D structure from its amino acid sequence. Alphafold’s latest iteration, released in October 2023, unlocks new insights and significantly improves accuracy across multiple key biomolecule classes, including ligands (small molecules), proteins, nucleic acids (DNA and RNA), and those containing post-translational modifications (PTMs). Furthermore, it can generate predictions for nearly all molecules in the Protein Data Bank (PDB), often achieving atomic accuracy. Why Eli Lilly and Novartis Eli Lilly and Novartis both are actively engaged in research and development in the field of small molecule therapeutics. Small molecules are compounds with low molecular weight that can easily enter cells, making them suitable for drug development. Eli Lilly had previously collaborated with Prism Biolab to develop and commercialise small molecules modulating targets selected by Lilly. This partnership leverages Prism’s PepMetics technology platform to explore oral protein-protein interaction (PPI) targets. Lilly has made several moves in the last couple of years to develop small-drug molecules. The company was one of the investors in Alto Neuroscience’s $45m Series C financing round, which will support Alto’s clinical programme of four small-molecule CNS candidates to treat psychiatric disorders including depression and post-traumatic stress disorder. Moreover, Eli Lilly recently acquired POINT, a radiopharmaceutical company with a pipeline of clinical and preclinical-stage radioligand therapies in development for the treatment of cancer. Radioligand therapy can enable the precise targeting of cancer by linking a radioisotope to a targeting molecule that delivers radiation directly to cancer cells, enabling significant anti-tumor efficacy while limiting the impact to healthy tissue. Meanwhile, Novartis is also working on creating new small molecules to target cancer in areas that were once considered undruggable. Novartis has recently teamed up with the University of California Berkeley, to open a centre specifically focused on the part of the proteome that has historically been challenging to address with small molecules. The collaboration aims to identify protein binding pockets and establish starting points for developing new therapeutic approaches. Generative AI and Drug Discovery The collaboration between generative AI and pharmaceutical firms is set to lead the way in new drug discovery. AstraZeneca has recently partnered with Absci, a startup specialising in generative AI antibody discovery technology, to expedite the discovery of novel cancer treatments. At the same time, NVIDIA introduced NVIDIA BioNeMo, a generative AI platform that offers services to develop, customise, and deploy foundation models for drug discovery. Much like AlphaFold, BioNeMo features a growing collection of pre-trained biomolecular AI models for protein structure prediction, protein sequence generation, molecular optimization, generative chemistry, docking prediction and more. Various companies are using NVIDIA BioNeMo for biology, chemistry, and genomics research. For example, Terray Therapeutics integrates BioNeMo cloud APIs into its multi-target structural binding model development. Innophore and Insilico Medicine apply BioNeMo to computational drug discovery, with Innophore incorporating it into the Catalophore platform, and Insilico using it in their generative AI pipeline for early drug discovery. Overall, the partnerships with Novartis and Eli Lilly offer Isomorphic Labs a potent combination of financial resources, industry expertise, and market validation. This will significantly accelerate their efforts in developing innovative small molecule therapeutics and ultimately bring their life-saving drugs to patients faster.","excerpt":"This development comes at a time when NVIDIA recently introduced BioNeMo","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2024-01-09T13:20:07","publication_year":"2024","word_count":666,"keywords":["Go","API","AI","RAG","Ray","Aim","generative AI","GAN","foundation models","R"],"extracted_tech_keywords":["AI","generative AI","foundation models","Aim","Ray","RAG","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-isomorphic-labs-partnered-with-novartis-and-eli-lilly\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10051240,"title":"Different Data Science Personality Tests","content":"To be a great data scientist, one must possess more than the capacity to meet technical requirements. Numerous personal characteristics contribute to the development of a great data scientist. The data science personality test is intended to assess the abilities of those who work as Data Scientists (DS), Data Analysts (DA), Machine Learning Engineers (MLE), Deep Learning Engineers (DLE), Machine Learning Researchers (MLR), or Deep Learning Researchers (DLR). The purpose of data science personality tests and quizzes is to learn about personality and strengths. The essential characteristics of a data scientist are A Constant Search for KnowledgeUnquenchable CuriosityStubborn OptimismA Capacity for PrioritizationA Practical Perspective A Moderate Dose of SkepticismAn Impact-Oriented MentalityA Dependable HabitA Domain-Specific IntuitionA Passion for Narrative Additionally, these assessments measure the following factors when it comes to career selection: Mathematics and Statistics: If you appreciate mathematics and statistics, you may also enjoy working as a data scientist!Practical business thinking: It is insufficient to produce an analysis. You must comprehend how and why an analysis becomes valuable and actionable. Are you more pragmatic and business-minded? Then you’ve come to the right place.Technical perception: Coding is a significant portion of a data scientist’s everyday job; throughout your data career, you will write several scripts, automations, and programmes. Several data science personality assessments are available Ambeone: This personality test for a data science job is a rapid indicator of whether an individual possesses the necessary fundamental skills for pursuing a career in data analytics, science, machine learning, or artificial intelligence. To attempt the quiz, click here. Aryng: Aryng’s Analytical Aptitude Assessment is a simple approach to determine your suitability for a data-driven profession or a career in analytics. To attempt the quiz, click here. Imocha: Recruiters and hiring managers use this data science personality assessment to find job-fit applicants for positions such as data scientist, data science developer, data science associate, data science analyst, and data visualisation engineer. To attempt the quiz, click here. Career explorer: They developed a test by combining advanced machine learning, psychometrics, and statistics on job satisfaction. To attempt the quiz, click here. Herzing: Technology generates an ever-increasing amount of data as it develops as a rising sector. The question of “Is data analytics helpful for you?” is unavoidable at this point. With the help of this personality evaluation exam, one can determine the value of a data scientist. To attempt the quiz, click here. Data scientists are mostly investigative persons, which means they are extremely inquisitive and interested individuals who frequently like solitude with their thoughts. Additionally, they are typically traditional, which means they are detail-oriented and organised and enjoy working in a structured atmosphere. The goal of data science is to extract value from data in order to assist organisations, institutions, and governments in making data-driven decisions. This is accomplished by the application of statistics, scientific methodologies, and data analysis. Data scientists must be proficient in mathematics (particularly statistics) and computer science. These data science personality assessments are designed to measure applicants’ knowledge of data science and programming fundamentals, as well as their talents and theoretical understanding of statistics, machine learning, neural networks, and deep learning.","excerpt":"Data science personality tests assess candidates’ understanding of the subject.","categories":["AI Features"],"tags":["Data analysts","Data Science","Data Scientists","Deep Learning","Machine Learning","Neural Networks","Statistics"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-10-12T15:00:00","publication_year":"2021","word_count":521,"keywords":["data science","artificial intelligence","machine learning","Statistics","AI","Neural Networks","ML","neural network","Machine Learning","Data analysts","RAG","deep learning","analytics","Deep Learning","Data Science","R","Data Scientists"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","data science","analytics","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-science-personality-tests\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":28522,"title":"Revenge Of The Humans: Why Open AI Five Could Not Win The Dota 2 Championship","content":"The International, which is the FIFA of Dota 2, a complex battle arena game, had an artificial intelligence system compete with professional players in the 2018 tournament. Earlier this August, an AI player called Five, created by OpenAI, failed to defeat professional human gamers. Despite having the training and “experience” of over 180 years, the AI was unable to achieve the feat. Why was it so? To give a brief to the uninitiated, Dota 2 is a popular online multiplayer video game which has 115 heroes, categorised according to strength, agility and intelligence. There are two teams of five players each and every team player has to pick a hero, which has different powers and characteristics, and destroy the opposite team’s base while encountering a lot of hurdles. Tech Behind Five Each of the five heroes of Five were trained with a neural network. They were trained for a gameplay worth of 180 years, for two months before the final match. Every neural network was trained by playing against itself. Learning from self-play provides a way for natural exploration of the game environment. During training, properties like health, speed or starting level, were randomised. At the beginning of each game, each hero was randomly assigned some set of lanes to follow and was not allowed to distract from these lanes. At first, Five players walked aimlessly inside the game, but after some hours of training, they could do things like farming and fighting. After some days, the Five players could think and play like humans by making strategies and performing actions such as stealing the opponent’s Bounty runes and walking to their tier one towers to farm. Gradually they became proficient in advanced tactics, like the 5-hero push. It was found that when the randomisations were increased the human player teams started to lose games. 80% of the games were trained against itself and the other 20% against its past selves. This was done to avoid any strategy collapse. The system was implemented as a general-purpose OpenAI Five’s learning algorithm named Rapid, which can be applied to any Gym environment. An advanced method based on policy gradient methods called proximal policy optimisation (PPO) was used to make decisions OpenAI used a separate long short term memory (LSTM) networks, a kind of recurrent neural network, for each hero to learn strategies. Each of the neural networks of Five has a single layer, 2024-unit LSTM that observes the current game state from the Bot API. It then eventually gives actions based on it via several action heads. Each head has a distinct action and is computed independently. To train the AI to play a game as real-time and complex as Dota 2, it had to be put in a very powerful processing capability. It has 256 P100 GPUs on GCP and 128,000 preemptible CPU cores on its CGP. Observations were 7.5 per second of the gameplay and the size of observation was 36.8 kilobytes. Batch per minute was 60 and the batch size was 1048576 observations. 5 Observations Where Five Went Wrong Unity: Five always seemed to stay in unity, even when it wasn’t required. This was beneficial to them when it was a good time to attack, but not favourable when the opponent took the advantage of it and tried to defeat them all together. It did not probably have the ability to realise that the opponent player heroes are not as same as they themselves and they could not decide what opponent powers could make them use as their strengths and weaknesses. It only took actions according to their own team heroes and so it wasn’t effective in a game where the opponent could be any hero of a hundred and fifteen. Missing couriers: They did not worry much about the courier in the game and kept playing in the battlefield despite the courier being present. They could not grasp that the courier is more important than fighting in the battlefield for the team survival. Speed: Five, being a machine, naturally had a faster response time than the professional players. So, they had fast decision-making abilities and could react faster in the gameplay. They didn’t have to keep checking on the map where their team was or check if their most powerful spell is ready. The usual human response time is around 150 to 500 milliseconds. Whereas, Five had a response time of about 80 milliseconds. Poor decisions: Although the decision-making skills were very fast, there were instances when the decision made was extremely poor. Five could not make optimal decisions to all the situations. For example, staying in groups all the time. Fearless: Five repeatedly sacrificed their top lane or bottom lane, with an intention of having a control over the opponent team’s safe lane. The instant they saw a kill, they went for it without gauging the consequences; without considering the enemy’s powers and what disadvantages might going near it and killing it have. Future The failure of OpenAI Five was not really a failure of AI. It showed that it could play something as complex as Dota 2. Dota 2 reflects many real-world environments. Games like this is a perfect testbed for AI research. OpenAI is one of the biggest organisations that are focused on solving humanity problems with AI. Humans are in turn learning new techniques from their matches with bots. For example, professional Go player Lee Sedol, was defeated by DeepMind’s AlphaGo, but it taught him a new technique in the game. DotA’s example would when Five allowed players to recharge a certain weapon quickly by staying out of range of the enemy. This was new and the human players learnt from it. Therefore, AI gives an opportunity to learn for both the parties — a win-win situation.","excerpt":"The International, which is the FIFA of Dota 2, a complex battle arena game, had an artificial intelligence system compete with professional players in the 2018 tournament. Earlier this August, an AI player called Five, created by OpenAI, failed to defeat professional human gamers. Despite having the training and “experience” of over 180 years, the […]","categories":["AI Features"],"tags":["lstm","Neural Network","policy gradient"],"author_name":"Disha Misal","publish_date":"2018-09-20T12:47:47","publication_year":"2018","word_count":963,"keywords":["Neural Network","Go","API","artificial intelligence","GCP","OpenAI","AI","neural network","ML","lstm","Aim","policy gradient","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","OpenAI","Aim","GCP","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/revenge-of-the-humans-why-open-ai-five-could-not-win-the-dota-2-championship\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10007379,"title":"9 Best Facial Recognition Software For Your PC","content":"Facial recognition has become one of the most debated technologies of recent times. Tech giants like IBM, Microsoft, Amazon, Google and others have been doing extensive research around it to help enhance numerous consumer applications, enhance security, help organisations go touchless in pandemic and more. While these are large-scale applications, installing facial recognition software into a personal computer can help in various ways such as automated login, use as a biometric for more convenient access, verify personal identity and more. Various organisations are now looking to install facial recognition software into personal computers, and with this article, we take you through nine best facial recognition software for your PC. (The list is in alphabetical order). Best Facial Recognition Software For PC in 2024ClarifaiDeepFaceDeepVisionFaceFirstFace++OpenFaceTrackerParavisionRohos Face LogonTrueface Clarifai The custom facial recognition software from Clarifai offers two powerful ways to integrate AI, which are Clarifai API and the portal. The face detection system returns probability scores on the likelihood that the image contains human faces and coordinate locations of where those faces appear with a bounding box. This system is being used in various sectors, including hospitality, retail, media, etc. The image, video and text recognition solutions are built on the machine learning platform and are made accessible via API, device SDK and on-premise. Get here. DeepFace Developed by a team of researchers at Facebook, DeepFace is a lightweight facial recognition and facial attribute analysis framework that leverages a very large labelled dataset of faces to obtain a face representation that generalises well to other datasets. According to the researchers, this system has closed the majority of the remaining gap in the most popular benchmark in unconstrained face recognition and is now at the brink of human-level accuracy. Get here. DeepVision Deep Vision facial recognition software provides various features for safety, advertising and urban planning. The features include facial detection, recognition, age and gender estimation, and more. It covers a variety of deployment scenarios from edge devices to cloud solutions or even on-premise. Also, the AI-based features can be applied to pedestrian safety and mobility, incident detection, vehicle recognition, among others to provide automated video analysis. Get here. FaceFirst FaceFirst is a popular facial recognition software for retail stores, including superstores, grocery, and department stores. The software offers a full range of biometric surveillance, mobile and desktop forensic face detection capabilities to fight theft, organised retail crime and workplace violence. It includes high-quality cameras and API that easily integrates face recognition analytics with existing technology systems. FaceFirst offers various functions such as authentication, access control, in-person ID verification, online ID verification and age verification. Get here. Face++ Face++ AI is a software that offers computer vision technologies which enable applications to read and understand. The software allows users to easily carry deep learning-based image analysis, with simple and powerful APIs and SDKs. Face++ AI Open Platform offers both free and premium options for all users. While the free option allows access to all Face++ APIs, with no usage limits, no upfront fees and no commitments, premium membership includes Pay-As-You-Go and Monthly Plan, along with technical support, for businesses of all sizes. Get here. OpenFaceTracker OpenFaceTracker is an open-source facial recognition program that is capable of detecting one or several faces on a picture or a video and to identify them via a database. The software comes with lGPLv3 license and a stable version 3.0. The features of this software include real-time processing of images, face identification, ability to operate on the Windows system, among others. Get here. Paravision Paravision face recognition is a computer vision developer toolset that powers a wide range of face recognition applications. The facial recognition enables comprehensive security and is deployable in the cloud, on-premises or at the edge. The software provides a comprehensive toolset for developing advanced face recognition solutions that includes face detection, face verification, face identification, real-time streaming video, among others. Get here. Rohos Face Logon Rohos Face Logon is a facial recognition software that allows a user to access a Windows computer in an easy and fast way by using any web camera. The software identifies a user by biometric verification based on neural network technology. The features of this software include automatic login or unlocking desktop by facial recognition, multi-user support, using of USB flash drive as a backup key to log in to Windows, among others. Get here. Trueface Trueface is an AI-powered facial recognition and digital identity verification system. The platform applies advanced computer vision technology to camera footage and images to enable businesses and other purposes to make immediate decisions based on identified patterns. Trueface has developed a suite of the software development kit and a dockerised container solution that harness the powers of machine learning and AI to transform the camera data into actionable intelligence. Get here. Related Posts Best Docker Containers Best AI-based Search Engines Best Data Cleaning Tools Best Free Resources To Learn Tableau Best Python Frameworks","excerpt":"Facial recognition has become one of the most debated technologies of recent times. Tech giants like IBM, Microsoft, Amazon, Google and others have been doing extensive research around it to help enhance numerous consumer applications, enhance security, help organisations go touchless in pandemic and more.   While these are large-scale applications, installing facial recognition software into […]","categories":["AI Trends"],"tags":["face recognition","face recognition online","Facial Recognition","machine learning software","Top Trend"],"author_name":"Ambika Choudhury","publish_date":"2020-09-15T12:00:01","publication_year":"2020","word_count":820,"keywords":["Top Trend","machine learning","Facial Recognition","AI","neural network","docker","computer vision","RAG","Python","deep learning","machine learning software","analytics","face recognition online","face recognition","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","analytics","RAG","docker","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/9-best-facial-recognition-software-for-your-pc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10119217,"title":"Azim Premji Investment Firm to Invest Portion of $10 Billion Fund in AI, Says CIO","content":"The Azim Premji-owned private equity fund, Premji Invest, will be investing a portion of the $10 billion it manages for the Wipro head in AI companies, according to reports. The company has decided to increase investments within the AI sector, according to a report from Bloomberg. Premji Invest’s CIO and managing partner T K Kurien stated that the investments will focus on refining their already existing proprietary investment tools that rely on AI. This announcement comes only a few weeks after Premji Invest had co-led healthcare-focused LLM Hippocratic AI’s $53 million Series A round. Additionally, it was also reported earlier this month that the company was looking to invest anywhere between $50 to $70 million in Canva. According to Moneycontrol, the company had also wanted to get ahead on the GenAI boom, which is why it had shifted focus towards more technology-based investments. Apart from Hippocratic AI, the company has invested in a few AI sector firms in the past, including Cohesity Inc., Pixis, and Ikigai. Additionally, according to Kurien, the company also hopes to allow access of their AI tools to open-source developers. In terms of AI tools developed by the company, Premji Invest had initially begun development in the field around three years ago, with a total of 14 AI engineers hired so far. Currently, the company makes use of AI to parse through companies based on several hundred parameters to “identify investment opportunities”. The latest announcement serves another purpose, with Kurien saying, “The firm expects the entire exercise to also give it a bird’s eye view of emerging technologies and trends that could help it stay ahead of peers.” The Premji Invest CIO also stated that they were looking into using the tools for potential streamlining of legal processes and improving government services. He stated that this would “help India’s overburdened courts resolve cases faster and to also aid governments’ efforts to offer service more effectively.”","excerpt":"Currently, the company makes use of AI to parse through companies to “identify investment opportunities”.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","ai investments","Fund Raising"],"author_name":"Donna Eva","publish_date":"2024-04-29T17:49:37","publication_year":"2024","word_count":319,"keywords":["Go","GenAI","ai investments","AI","programming_languages:R","Fund Raising","ML","programming_languages:Go","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","GenAI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/azim-premji-investment-firm-to-invest-over-10-billion-in-ai-says-cio\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10000151,"title":"China’s New Lunar Satellite Queqiao Acts Like A Radio Behind The Moon","content":"It has been almost 12 years since China launched Chang’e 1, their unmanned spacecraft which orbits the moon. After 2007, they later launched Chang’e 2 in 2010, Chang’e 3 in 2013, and now Chang’e 4 is set to probe the dark side of the moon. According to reports, the Chang’e-4 lander and rover are scheduled to launch in December this year to perform the first-ever soft-landing on the far side of the Moon. Even though the moon rotates, the same side always faces the Earth since the amount of time it takes for the moon to rotate on its axis is the same amount of time it takes to make a full orbit around the Earth. This latest launch is part of China’s growing ambitions for lunar exploration, which has already achieved numerous successes. The previous mission involved the Chang’e-3 probe and Yutu lunar rover. How Queqiao Works This is the first communications relay satellite in orbit around the moon and the orbiter’s name, Queqiao literally means “Magpie Bridge. The name has been derived from a 2,000-year-old fairy tale originating from the Han Dynasty. Queqiao, the 450-kilo satellite also carries a Netherlands-China Low-Frequency Explorer (NCLE) instrument to carry out low-frequency astronomy which is not possible on Earth due to its atmosphere, and provide a window into the cosmic ‘dark ages’. The Queqiao orbiter will act as a link between the Chang’e-4, which is set to go to the far side of the moon later this year, and communications stations on Earth. However, experts say that the moon does not have a dark side. But given the way it orbits the Earth, our current satellites shows us only one side of the moon. The far side of the moon is essentially the hemisphere that faces away from Earth. The far side’s terrain is rugged with a multitude of impact craters and flat lunar maria. It also has one of the largest craters in the Solar System, the South Pole–Aitken basin. Queqiao’s Journey One of two microsatellites launched along with a required communications relay satellite in May has quietly been allowing radio operators to download images from the spacecraft taken along its elliptical lunar orbit. After a 24-day journey, Queqiao entered the Earth-Moon L2 halo orbit. While a normal mission to lunar orbit takes up to four or five days, Queqiao took much longer due to its special orbit. In Stage 1 from Earth to the vicinity of the Moon, the launch took place on  21 May 2018 at 5:28 AM Beijing time and the satellite was sent directly into an Earth-Moon transfer orbit without an Earth parking orbit phase. There are two kinds of transfer orbits from Earth to the vicinity of Earth-Moon L2. One is a direct transfer, which is directly from the Earth parking orbit to L2. The relay satellite used a lunar swing-by transfer orbit with an apogee of about 400,000 kilometres from Earth. According to reports, the transfer orbit requires 2 or 3 trajectory correction manoeuvres and takes about four or five days to arrive at the Moon. On May 25, Queqiao performed another burn when it flew past the Moon at a distance of just 100 kilometres. This successfully put it on a Moon-L2 transfer orbit.","excerpt":"It has been almost 12 years since China launched Chang’e 1, their unmanned spacecraft which orbits the moon. After 2007, they later launched Chang’e 2 in 2010, Chang’e 3 in 2013, and now Chang’e 4 is set to probe the dark side of the moon. According to reports, the Chang’e-4 lander and rover are scheduled […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Richa Bhatia","publish_date":"2018-12-04T13:27:35","publication_year":"2018","word_count":540,"keywords":["Go","programming_languages:R","ETL","AI","programming_languages:Go","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","R","Go","ETL","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chinas-new-lunar-satellite-queqiao-acts-like-a-radio-behind-the-moon\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":20879,"title":"Top 7 Platforms That Are Busting Fake News On Social Media","content":"The development of social media provided a medium for instantaneous engagement across all social strata. As the number of people engaging on social media platforms increase, these platforms have turned into breeding grounds for outlandish, exaggerated  and often dangerous ’fake news’. This fake news covers misinformation and hoaxes that go viral on social media, and often lead to undesirable consequences. The expeditious dissemination of these materials and their sheer volume have created what can be called as epidemic of misleading and fallacious news on social media. From posts about an Iranian girl’s dramatic transformation to look like Angelina Jolie to the alleged torture of a young man who was found dead under mysterious circumstances that sparked communal tensions in coastal Karnataka, fake news has dangerous repercussions. On the bright side, there are startups and other independent initiatives that are doing their part in dealing with this epidemic by debunking or developing methods to stop spreading such misinformation. AIM brings you seven platforms that are taking fake news head on: 1. Check4Spam This self-funded project began as a humble WordPress site in 2015. Founded by Shammas Oliyath, a software engineer at IBM in February 2015, Bal Krishn Birla, an entrepreneur joined forces with Oliyath in July 2016. Since then the two techies from Bengaluru have taken it upon themselves to quash fake news and hoaxes doing the rounds on social media. In addition to their professional commitments, the two founders help people sending in queries to their WhatsApp number during lunch breaks and after office hours. Reportedly, there are over 200 such queries sent on their phone number and thousands of requests submitted to their website per day. While Oliyath handles research, Birla takes care of the technical aspects of the site, especially search engine optimisation that directs traffic towards them when messages are searched online for verification. Some of the categories found on their website include promotions, internet rumours, and jobs among others. Interestingly, the founders have gone on record to say that most of the material they verify is political in nature. 2. Storyzy Other than fake facts and figures that get circulated, another major issue is fabricated or fake quotations attributed to public figures. Storyzy, an ad-tech startup formed  in 2012, helps in verifying such questionable quotes. Employing Natural Language Processing (NLP) to identify fake news in real-time, this startup specialises in automated unique quote extraction. Initially starting as Trooclick, the company developed a browser add-on known as Glitch Spotter, that tagged fake financial news. However, in 2015 Glitch Spotter was shutdown and Trooclick eventually became Storyzy. It launched its quotes search engine in June 2016 and began concentrating solely on quotes. In 2016, the startup raised $900,000 seed funding to expand. A social media user can use Storyzy’s Quote Verifier to validate a quote by pasting the text from an online news in the box and hitting check. With a database of over 15 million quotes, it extracts reported speech from the online textual news material, along with other metadata. The text is converted to data using NLP and verified for authenticity. 3. Factmata This london-based AI startup was founded by NLP researchers Dhruv Ghulati, Andreas Vlachos and  Sebastian Reidel. It made headlines in June 2017 when it secured a seed funding of $750,000, which included investment from Mark Cuban, businessman, investor and the owner of NBA’s Dallas Mavericks. Before securing the $750,000 funding, Factmata had previously received a grant of nearly $70,000 from Google’s Digital News Initiative (DNI) in 2016. The startup does automated content scoring and verification, specifically targeting the dissemination of fake and misleading news. It is now working on building an AI-based platform for fact-checking and news aggregation. The model that the founders are adopting for now borrows elements from two famous crowd sourced information sites — Wikipedia and Quora. The process of fact-checking is to be  powered by people with interest and concern towards fake news, and in time is expected to be completely automated and powered by AI. In the initial model, users will serve as watchdogs and flag content that is false. When operational, the platform will provide two kinds of services — one for journalists, media enthusiasts, newsreaders and general public, and secondly for advertisers and businesses. 4. SM Hoax Slayer An Indian initiative to combat fake news, it began as Facebook page in 2015, before creating a website of its own. A personally funded initiative of a Mumbai-based businessman Pankaj Jain, it tackles every known kind of spurious ‘news’ circulating on social media based on issues surrounding religion, politics, scams and communal issues, among others. The website received a more wider audience after it received praise online from prominent Indian personalities such as noted journalist Shekar Gupta and musician Vishal Dadlani. It was also covered by the media both in India and abroad. SM Hoax Slayer’s posts have featured in prominent news outlets as proof to put rumours and fake news to rest. Like Shammas Oliyath and Bal Krishn Birla, Jain works on debunking hoaxes and setting records straight on fake news, along with his full-time commitment as a businessman. Depending on the nature of the content — from morphed pictures to viral political hoaxes — Jain always has an ear to the ground for news items from the archives resurfacing as distorted and fabricated trending news. 5. Crisp Thinking Founded in 2005 by Adam Hildreth, also the Chief Executive Officer and Director, Crisp Thinking is a social media risk management firm that protects the image of brands, companies and online communities on social media. It promises 24×7 monitoring and moderation of content posted on the social media pages of the brands it manages. The contents assessed for risk are images, videos and text, and includes substantial amount of user-generated content posted on these pages. Posts relating to terrorism, violence, illegalities such as infringement and abuse are picked up within 15 minutes of posting and the parties at harm’s way are notified. They also moderate thousands of online channels and scrutinise over 3 billions items of content every month across 50 languages. Initially established to protect children and teenagers using social media and online games from cyber bullying and sexual abuse, the company evolved into a watchdog protecting the online credibility of companies and other groups. Recently, Hildreth had said that one of the main issues that their company is dealing with is mostly racist, violent and related to terrorism. 6. Userfeeds The momentum a news item or a post gains online is based on the massive number of likes and shares it receives, which in turn increases it visibility and popularity. This holds true especially in the case of fake news. This Warsaw-based startup founded in 2017 by Grzegorz Kapkowski and Maciej Olpinski goes against this trend and establishes a platform that integrates blockchain and token-based ranking for content validation. They also received a seed funding of $800,000 in 2017. It creates a content network that depends on ranking algorithms that use digital tokens such as Bitcoins to verify to the quality of the item. The beauty of this platform is the ‘tokenisation’ of every item. By attaching an economic incentive to every item, an auditable digital trail is automatically created. This helps in accurate validation of content and assigning its true importance in spite of the tampering of metrics by bots and falsified information spread in abundance. 7. AltNews Founded in 2017 by Pratik Sinha and an anonymous contributor who goes by the name Unofficial Sususwamy, Altnews is a fact-checking website, which fights fake news like its mentioned Indian counterparts on the list. It aims to fight propaganda and misinformation found on social media. A software engineer by profession, Sinha quit his job to devote his time and efforts to AltNews. Sinha was invited by Google News Lab Asia-Pacific Summit 2017 in Singapore to discuss about possible solutions to combat fake news. It is a non-profit initiative registered under the name ‘Pravda Media Foundation’ and funded by Sinha himself. However, it seeks to expand its operation and reach, and only seeks donations from its readers for the same.","excerpt":"The development of social media provided a medium for instantaneous engagement across all social strata. As the number of people engaging on social media platforms increase, these platforms have turned into breeding grounds for outlandish, exaggerated  and often dangerous ’fake news’. This fake news covers misinformation and hoaxes that go viral on social media, and […]","categories":["AI Trends"],"tags":["famous quotes","social media","whatsapp"],"author_name":"Jeevan Biswas","publish_date":"2018-01-22T09:35:16","publication_year":"2018","word_count":1352,"keywords":["Go","AI","RPA","social media","Git","whatsapp","NLP","Aim","ViT","GAN","famous quotes","R","startup"],"extracted_tech_keywords":["AI","NLP","Aim","R","Go","Git","GAN","ViT","RPA","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/platforms-busting-fake-news-social-media\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":58926,"title":"How To Prepare For A Remote Data Science Interview","content":"Even remote data science workers are not spared the dread of facing interviews for job opportunities. No matter how confident you may be of your technical expertise, the best of us walk in with butterflies in our stomach. However, most of this nervousness stems from the uncertainty around the entire interview process. Although all companies conduct their interview processes differently, most follow a similar template. By condensing these common patterns into a few actionable points, we have created a guide for you to ace your next data science interview. However, be assured that it will be a long, drawn-out process. Interviewing for data science roles remote or otherwise can take you through multiple stages over the course of a few months. Telephonic Interview Getting a referral and having experienced professionals advocate for you might get you to this stage, but you need to prepare well to get past it. A telephonic interview by either a technical recruiter or a data science manager is the first step for most companies hiring for data science positions. These calls will broadly assess if you have the right skills for the position you have applied for, and are generally short and to the point. It will also delve into your past project experience to ascertain if you are a good fit for the company. One way to handle your interview at this stage is to express your passion for the company you have applied for. Research well and tell the interviewer why it makes professional sense for you to take on this role and associate yourself with the company. Thoroughly review the job posting and compare it against your resume when talking to the interviewer. Furthermore, do not hesitate to ask questions that demonstrate you researched the company well. Written Assessment After you have cleared the telephonic interview, companies often send a home assessment. This will vary depending on the position you have applied for, but typically includes a dataset for you to analyse or even a coding assessment. Assignments also greatly differ depending on the company and will reflect what you will be doing on a regular basis if hired. The format also varies from company to company. While some may provide an online platform that is unmonitored, others may prefer that you log in with an interviewer watching you perform the assessment. Both will, however, be bound by a strict timeline. While the outcome of this assessment totally depends on the depth of your technical knowledge, it will be advisable to use detailed visuals to make it interpretable to business stakeholders. This tells the interviewer that you have a good understanding of how your work can drive business value. Interview With Data Science Manager This person will be leading the data science team, and maybe somebody you would be directly reporting to. Other members of the data science team\/data engineers may also join in for the interview. While the first interview covered broad topics to assess whether or not you make a good fit for the company, this interview — done over Skype — will be quite technical. Expect them to comb through your resume to discuss the projects you have worked on in the past. Take time to explain the logic behind the methodologies you employed and the algorithms you used. Be prepared to meticulously go over everything you have mentioned in your resume — from the tools you used to the logic behind using those. They may also ask you to expound on some of your answers from the written assessment, so be sure to save your answers and revise them before you appear for this interview. Furthermore, think of ways in which you could have improved on the assessment. Although these tips should help you ace the interview, attempting to understand what the data science team is working on can help you greatly. Presentation Although these are quite rare, some companies may ask you to present the findings of your analysis — the one in the assessment, or from one of your past projects — using PowerPoint. You need not get flustered with colour coding or embellishing your presentation, although it should be visually appealing. The focus should be on your verbal communication with your audience (business stakeholders). What is the best way to handle this presentation? Explain in simple terms why you used the methods to arrive at your analysis, followed by explaining the outcomes in a non-technical language, and finally addressing the key takeaways. Final Interview This may usually be conducted by the HR department or even the CTO. They may ask about how you would like to grow in the company, or what you would like to learn on the job. This is an opportunity for you to reiterate your passion for the company and how you could bring value to the company. Since this is your final chance, do not hesitate to talk about yourself as long as it can tie back to your excitement for the job. Also, ask questions about the next steps and the future of the company.","excerpt":"Even remote data science workers are not spared the dread of facing interviews for job opportunities. No matter how confident you may be of your technical expertise, the best of us walk in with butterflies in our stomach.  However, most of this nervousness stems from the uncertainty around the entire interview process. Although all companies […]","categories":["AI Trends"],"tags":["Data Science","Data Scientist","job interview"],"author_name":"Anu Thomas","publish_date":"2020-03-18T15:00:00","publication_year":"2020","word_count":846,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","job interview","Data Science","Data Scientist","R"],"extracted_tech_keywords":["AI","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-prepare-for-a-remote-data-science-interview\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":6265,"title":"Why expertise in Analytics technology will not help you&#8230;","content":"Analytics is the buzz word in today’s world! Everyone wants to be connected with Analytics profession in one way or the other. And it is good – I can assure you that the future is bright for Analytics industry after being in the industry for over 13 years now. So where is the problem? Well, every single analyst or Analytics aspirant I come across these days would normally talk about his or her technology skills. “I am an “expert” in xyz tool for data analytics or abc tool for business intelligence and reporting. I have worked on 3 data visualization platforms.” Good – you have worked on all these platforms – so what? Does that make you a good analyst? NO. What makes you a good analyst is your subject matter expertise. What makes you a good analyst is the knowledge and experience to differentiate between two different time-series models and which one to apply in which situation. What makes you a good analyst is the understanding of gaps in your data and how to address them. Now, I am not trying to undermine the importance of technology here – technology is important. The Analytics revolution that we have seen in last decade or so would not have been possible without technology. My point is not to make technology your driver for your analytics practice. Ideally, technology should adapt to business requirements, but in today’s world, we try to fit business processes within the technology framework! That’s the problem – we focus on technology and we forget the core issue! Also, don’t get me wrong – expertise in Analytics technology is important for anyone who is interested in a career in IT industry – a programmer working for a software company cannot turn back and say that technology is not important for him or her. For him\/her, technology is bread-and-butter. Not the same for someone interested in “Analytics” industry – working as an analyst or aspiring to be an analyst! Fundamental knowledge or core body of knowledge, like applied Math, applied Statistics, Econometrics, Decision Science, Management Science, etc. is more important for an analyst. Remember – the core job of an analyst is to solve some of the most complex business problems and not to drive use of technology! The key reason why expertise in Analytics technology will not help you build a successful career! Original Article on LinkedIn: https:\/\/www.linkedin.com\/pulse\/article\/20141007084034-23651558-why-expertise-in-analytics-technology-will-not-help-you","excerpt":"Analytics is the buzz word in today’s world! Everyone wants to be connected with Analytics profession in one way or the other. And it is good – I can assure you that the future is bright for Analytics industry after being in the industry for over 13 years now. So where is the problem? Well, […]","categories":["IT Services"],"tags":[],"author_name":"Kushal Shah","publish_date":"2014-10-11T05:31:37","publication_year":"2014","word_count":400,"keywords":["business intelligence","Go","programming_languages:R","programming_languages:Go","analytics","R"],"extracted_tech_keywords":["analytics","R","Go","business intelligence","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/expertise-analytics-technology-will-help\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":53493,"title":"Tata Mutual Fund Launches New AI\/ML-Powered Quant Fund","content":"We live in a technological era where machines and software are becoming essential tools empowering businesses, Tata Mutual Fund has launched an artificial intelligence (AI) and machine learning (ML)-powered fund using a proprietary quant framework — Tata Quant Fund. Tata Quant Fund is an AI-powered and ML-enabled open-ended mutual fund scheme. Once subscribed to the scheme – the subscription lasts for a stipulated period of time; upon expiry – the scheme requires repurchasing or resubscribing. On the 3rd January, Tata Quant Fund was offered for subscription, the New Fund Offer (NFO) closes on the 17th January and will reopen on the 27th January again. Rs 5,000 is the least amount of money required for anyone who wishes to subscribe and an initial invest of Rs 1,000 is mandatory. Subscribers can compete to acquire additional shares and increase their initial investment amount by multiple of Re 1. With the launch of the Tata Quant Fund, the company becomes the fourth one to enter this space. Similar to Tata Quant Fund, DSP Quant Fund and Nippon Quant Fund follow the S&P BSE 200 TRI (Total Return Intex) while Reliance Quant follows Nifty 100 ESG TRI. Tata Mutual Fund has built a reputation for being morally upright where investor’s interest takes precedence over the company’s own. Tata Quant Fund Outperform The Market The scheme looks to invest in equity or equity-related securities anticipating medium to high capital appreciation in the long run. The performance of the schemes is predicted by the AI\/ML-powered engines. The quant fund engines are designed to analyse more than 20 years of market data and records, including the scheme’s behaviour, influence and relation with other schemes, and also its inabilities. The findings are then compared with the prevailing market and macro-economic conditions. The final outcome is free of human bias, emotions or other inefficacies and is logically the most viable option for investors and subscribers to buy into. Tata Quant Fund utilises several Rule Engines which can be understood as if\/then system where ‘if’ the required conditions are not met ‘then’ the system terminates the operation or keep it on hold. The AI\/ML-powered engines take into consideration the procurement of maximum profit during a boom, or it focuses on minimising losses during a recession while formulating schemes and predicting the scheme’s performance. Each Rule Engine is capable of analysing market data of 30 companies individually and grading them accordingly. The AI\/ML engines are designed to recalibrate itself at predetermined intervals to incorporate new and relevant information. This enables the AI\/ML engines to predict future trends and developments in the mutual fund market. The algorithm, behind the AI\/ML engines’ analysing and predicting capabilities, is developed by data science specialist from within the Tata Group. “Machines have massive computational power needed to process very large data sets, spot patterns and correlations, make decisions faster, objectively and without human biases. In the current world, computers are powerful enough to solve problems, a lot of data is available and we strive to use this data in combination with algorithms to its best,” comments Prathit Bhobe, Managing Director and CEO of Tata  Mutual Fund.","excerpt":"We live in a technological era where machines and software are becoming essential tools empowering businesses, Tata Mutual Fund has launched an artificial intelligence (AI) and machine learning (ML)-powered fund using a proprietary quant framework — Tata Quant Fund.  Tata Quant Fund is an AI-powered and ML-enabled open-ended mutual fund scheme. Once subscribed to the […]","categories":["AI News"],"tags":["AI\/ML"],"author_name":"Yeshey Rabzyor Yolmo","publish_date":"2020-01-10T10:45:38","publication_year":"2020","word_count":520,"keywords":["data science","Go","API","machine learning","artificial intelligence","programming_languages:R","AI","ML","programming_languages:Go","AI\/ML","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-powered-tata-quant-fund\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":30815,"title":"Top 5 Video Analytics Companies In India","content":"Artificial intelligence technology has helped build software that can analyze videos for better capturing of events, understanding patterns and surveillance. In this article, we will see the top 5 Indian companies that are capturing the space of video analytics. (The list is in no particular order). Silversparro This Gurgaon-based tech startup was founded IIT Delhi alumni. Silversparro aims to solve business problems by using deep learning technology along with in-house tools and expertise to help grow client’s business. They provide solutions like industrial automation, video analysis and consulting. As organizations are flooded with videos from content video to CCTV footage, Silversparro allows to automatically analyze such videos for a range of insights. They are venturing into CCTV video analytics for retail establishments, which is giving them a good traction. They also use DL for customer demography analytics, analyzing customer behavior and auditing staff performance. AllGoVision Headquartered in Bengaluru, AllGoVision provides AI-powered video analytics. Equipped with more than 50 basic and advanced features, they provide services in more than 30 countries. Traffic surveillance, building surveillance, city surveillance, and business services in different sectors, are in AllGoVision’s repertoire. The company uses a mixture of in-house image processing expertise and talented engineers. The company is also trying to bring out efficient outputs, like intelligent traffic surveillance, crowd management, perimeter management, capricious events detection, retail intelligence, and facial recognition. Videonetics Founded in 2008 by Dr Tinku Acharya Videonetics is a noted visual computing development company. Headquartered in Kolkata, they provide innovative security solutions such as intelligent video management software which is AI and deep learning video analytics prototypes for traffic management, systematic facial detection, and recognition based on cloud technology. The company was successfully granted patents from countries like the US, UK, Singapore, Canada, and Israel, among others. Livedarshan Founded in 2002, the company has around 500 clients across India which have developed its own software solutions for construction monitoring. The company has partnered with an Israel-based firm called VideoCalls to provide SaaS (surveillance as a service solution) to telecommunication organizations in India. They provide services in the fields of traffic, city, building surveillance, border security, and business intelligence. Cron-J This Bengaluru- based company provides advanced video analytics solutions which are integrated with smart security systems for video surveillance. The company software is capable of finding anomalies involving people, objects and vehicles and generate alerts based on predefined constraints. They are providing services in the fields of manufacturing, government schemes, healthcare, transportation and traffic security, retail, banking and finance, city surveillance, mining industries, and license plate recognition.","excerpt":"Artificial intelligence technology has helped build software that can analyze videos for better capturing of events, understanding patterns and surveillance. In this article, we will see the top 5 Indian companies that are capturing the space of video analytics. (The list is in no particular order). Silversparro This Gurgaon-based tech startup was founded IIT Delhi […]","categories":["AI Trends"],"tags":["artificial inelligence","big data in auditing","video analytics"],"author_name":"Bharat Adibhatla","publish_date":"2018-11-28T12:01:43","publication_year":"2018","word_count":423,"keywords":["Go","artificial intelligence","artificial inelligence","TPU","AI","automation","Aim","deep learning","big data in auditing","analytics","video analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","analytics","Aim","TPU","R","Go","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-video-analytics-companies-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":50960,"title":"How XROM Is Pushing The AR\/VR Industry To Thrive In India","content":"Augmented Reality and Virtual Reality technologies have been witnessing an exponential growth in the Indian Market. AR\/VR technology has vouched to be the future and potential to impact all possible industry verticals right from education, healthcare, retail, travel, training, simulation, entertainment & gaming industry. According to reports, Augmented Reality and Virtual Reality market are expected to grow at an estimate of 6.5 billion by 2022. Currently, India’s AR\/VR\/MR industry is in its nascent stage and it is growing eventually. With such a goal, a Mumbai-based Extended Reality venture called XROM is working hard to design and shape India’s AR\/VR\/MR growth story. The vision is to solve the accessibility problem and bring some of the best Global AR\/VR\/MR Products to India. The company was founded by Eddie Avil and Ashley Rodrigues with a vision to celebrate, showcase and shape India’s AR\/VR\/MR industry. The startup might be founded this year but the foundation was started back in 2016 through “VR Storytellers” which is an Immersive Experiential VR Content Studio. The founders built their own Robotic Dolly and 16 Camera VR Stereoscopic Rig and produced India’s 1st cinematic VR experience known as CRACKLE. Being a bootstrapped venture, the company did face lots of challenges such as limited resources but it is holding back the things to grow the business. Flagship Product This year in October, the company launched three business verticals as mentioned below XROM-News: XROM-News is an AR\/VR\/MR digital news platform which puts out daily articles and interview startups.XROM-Podcast: XROM-Podcast is an Audio\/Video Podcast for the AR\/VR Industry education and awareness.XR Bazaar: XR Bazaar is an online-shop\/e-commerce platform to buy\/sell\/rent all products such as hardware, software, and accessories related to AR\/VR\/MR. XROM Is Tackling The Hiring Phase While talking about the hiring phase, Eddie replied that there is a huge talent crunch, acquiring AR\/VR Developers isn’t easy and that is why it is a super lean two-member team. Currently, the startup is sticking to hiring freelancers for commercial jobs and they are looking forward to growing the team in the near future. Future Roadmap Talking about the roadmap, Eddie replied that the company has built an ecosystem that would be beneficial for any AR\/VR\/MR Startups, also, they are working on offering collective\/collaborative services and solutions to the enterprise. Currently, the startup is working on offering education institutes licensed\/ personalised curriculum along with the entire backend right from VR headsets, hardware, sync software to support.","excerpt":"Augmented Reality and Virtual Reality technologies have been witnessing an exponential growth in the Indian Market. AR\/VR technology has vouched to be the future and potential to impact all possible industry verticals right from education, healthcare, retail, travel, training, simulation, entertainment & gaming industry.   According to reports, Augmented Reality and Virtual Reality market are expected […]","categories":["Deep Tech"],"tags":["freelancer"],"author_name":"Ambika Choudhury","publish_date":"2019-12-02T13:00:00","publication_year":"2019","word_count":402,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","R","freelancer","startup"],"extracted_tech_keywords":["AI","R","Go","Git","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-xrom-is-pushing-the-ar-vr-industry-to-thrive-in-india\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045855,"title":"[Jobs Roundup] Latest Chief Data Scientist Job Openings In India","content":"We have listed the latest Chief Data Scientist Job Openings In India this week. 1| Chief Data Scientist at Oracle Location: Bangalore Responsibilities: Develop scalable infrastructure, including microservices and backend, that automates training and deployment of ML models.Brainstorm and design various POCs using ML\/DL\/NLP solutions for new or existing enterprise problems.Work with fellow data scientists\/software engineers to build out other parts of the infrastructure, effectively communicating your needs and understanding theirs, and address external and internal shareholders’ product challenges. Apply here. 2| Chief Data Scientist at Zycus Location: Mumbai Responsibilities: Act as a technical thought leader in collaboration with the analytics leadership team, helping to set the strategy and standards for machine learning and advanced analytics.Work with senior leaders from all functions to explore opportunities for using advanced analytics.Provide technical leadership, coaching, and mentoring to talented data scientists and analytics professionals. Apply here. 3| Head – Data Science at Wynk Location: Gurgaon Responsibilities: Analyse experimental data with statistical rigour. Ensure accurate interpretation by combining business acumen with detailed data knowledge and statistical expertise.Work closely with Product Development leadership to drive business decisions using data – you’ll turn data into insights and help surface relevant and personalised content to Wynk’s millions of users.Collaborate with Wynk’s product team to find novel ways to activate and engage users. Apply here. 4| Head – Advanced Analytics & Data Science at BARC Location: Mumbai Responsibilities: Manages experimental data mining, fusion and data modelling projects.Designs and applies modelling algorithms to be deployed for production improvements.Creates and enhances statistical models for numeric and categorical data. Apply here. 5| Head of Data Science at Matrimony.com Location: Chennai Responsibilities: Develop and implement a strategic roadmap to progressively move towards an increasingly data-driven organisation.Lead all data related activities to support the development of new data science approaches and methodologies to improve operations and business outcomes.Project management of multiple projects involving different stakeholders. Apply here. 6| Director Data Science at Novartis Location: Hyderabad Responsibilities: Develop and execute a roadmap for the team to innovate in multiple business verticals by transforming the way to solve a problem using Data Science and Artificial Intelligence.Strategically coordinate, prioritise and efficiently allocate the team resources to critical initiatives: plan resources proactively, anticipates and actively manage change, set stakeholder expectations as required, identify operational risks and enable the team to drive issues to resolution, balance multiple priorities, and minimise surprise escalations.Design, develop and deliver various data science-based insights, outcomes and innovation (using Mathematics, Computer Science, Statistics, Engineering, Management Science, Technology, Economics, etc.) and create “proof of concepts & blueprints” to drive faster, timely, highly precise, workable and proactive decision -making based on data-based insights and science, and shape strategic glide path of the company. Apply here. 7| Director – Analytics & Data Science at Varian Location: Pune Responsibilities: Build, lead and mentor a high-performing team of data modellers, data scientists and business intelligence developers.Hire world-class talent and provide technical guidance and career development to team members.Drive the process through quickly assessing data, product strategies by understanding needs, satisfaction drivers, performance requirements and business cases through the value chain\/channel structure for D&A. Apply here. 8| Head of Data Science at U GRO Location: Mumbai Responsibilities: Establish the advantage of using data science in various facets of SME lending, including (but not restricted to) credit scorecards & portfolio monitoring frameworks, business insights, campaign management, propensity, and pricing.Design, develop, and implement machine learning applications.Evangelise the use of alternate data in enhancing risk management and customer experience. Actively interface with technology to implement and improve business solutions. Apply here. 9| Director – Data Science & Engineering at Salesforce Location: Hyderabad Responsibilities: Drive the vision for data engineering across product verticals and define the execution path to achieve the vision.Build and lead a small but highly qualified team of data engineers to deliver reliable, scalable and secure analytic data models for various business processes.Build partnerships with Product Managers, Software Engineers, Program Managers, Architects and Platform Engineers to understand data consumption needs and influence instrumentation and data transport. Apply here. 10| Head Data Science at Tide Location: Hyderabad Responsibilities: Working closely with the data scientists in your team to ensure that best practices are followed.Owning the data science processes and tools; survey new developments within the data science space and adapt them within Tide.Working closely with the business and the data science team to ensure that ML projects are aligned to business needs. Apply here. AIMRecruits A subsidiary of Analytics India Magazine, AIM Recruits is an India-focused executive search firm that partners with leading businesses to assess and acquire top data science talent to drive breakthrough performance. AIM Recruits is a leading Executive Search Firm for Analytics, Data Science & Artificial Intelligence. 1| Data Engineer\/Senior Data Engineer Location: Bangalore Responsibilities: For onboarding tables to the data lake and building pipelines for data ingestion using python.Building data models and tables on Athena which will be used for reporting. Create and maintain workflow\/technical documents for all the development activities.Gather business requirements from business stakeholders and deliver solutions. Apply here. 2| Senior Data Scientist Location: Bangalore Responsibilities: Strong technical architecture, design, deployment and operational level knowledge of AI platforms, standards, and protocols. Developing AI solutions using Spark, TensorFlow, Keras, etc.Good understanding of model development, model validation techniques\/metrics and model deployment.Analyse huge data to train ML models and evaluate them with the right metrics. Apply here. 3| Big Data Lead – Spark\/MongoDB Location: Bangalore Responsibilities: Perform product and technology assessments whenever needed. Develop analytic tools, working on BigData and Distributed Environment. Scalability will be the key.Liaise with other team members to conduct load and performance testing on modules.Visualise and evangelise next-generation infrastructure in Big Data space (Batch, Near RealTime, RealTime technologies). Apply here. 4| Data Architect Location: Bangalore Responsibilities: Developing and implementing an overall organisational data strategy that is in line with business processes. The strategy includes data model designs, database development standards, implementation and management of data warehouses and data analytics systems.Identifying both internal and external data sources and working out a plan for data management aligned with organisational data strategy.Coordinating and collaborating with cross-functional teams, stakeholders, and vendors for the smooth functioning of the enterprise data system. Apply here. 5| Machine Learning\/AI Architect Location: Bangalore Responsibilities: Architecture, total solution design from requirements analysis, design and engineering for data ingestion, pipeline, data preparation & orchestration, applying the right ML algorithms on the data stream and predictions.Application of Machine Learning will be key at all layers of the stack.Requires engagement in every phase of the system development lifecycle including requirements generation, system and software design, implementation, integration & test, and verification & validation. Apply here. 6| AI Architect (PhD. Scholars) Location: Bangalore Responsibilities: To design applications and solutions using Machine Learning and AI.Architecture, total solution design from requirements analysis, design and engineering for data ingestion, pipeline, data preparation & orchestration, applying the right ML algorithms on the data stream and predictions.Application of Machine Learning will be key at all layers of the stack. This requires engagement in every phase of the system development life cycle, including requirements generation, system and software design, implementation, integration & test, and verification & validation. Apply here.","excerpt":"We have listed the latest Chief Data Scientist Job Openings In India this week.","categories":["AI Features"],"tags":["Data Science","Data Science Career","Data Science Jobs","Data Scientist Jobs","open source data science projects","weekly job updates"],"author_name":"kumar Gandharv","publish_date":"2021-08-12T18:00:00","publication_year":"2021","word_count":1185,"keywords":["data science","Data Scientist Jobs","open source data science projects","artificial intelligence","machine learning","AI","Keras","ML","Data Science Jobs","weekly job updates","Data Science Career","NLP","Aim","analytics","Data Science","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","data science","analytics","Aim","TensorFlow","Keras"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/jobs-roundup-latest-chief-data-scientist-job-openings-in-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10012545,"title":"What Is The Hiring Process For Data Scientists At Bridgei2i","content":"With strong AI and analytics operations, Bridgei2i has over 600 data scientists forming the backbone of the team. As the company focuses on delivering AI-powered transformation, they rely on data science, technology, and business applications to provide the best outcome — for which the data science team plays a key role. To drive a large data science team, the company fosters an open and people-driven work culture while encouraging collaborative and inclusive decision making across teams to nurture a problem-solving ecosystem. To understand in detail about their data science teams and how they hire data scientists in it, we got in touch with Preeth Joseph, Director – Talent and Strategy, BRIDGEi2i, for this week’s column. Data Science Skill Sets As Preeth shares, at Bridgei2i, they generally shortlist candidates skilled in technology, data science, or business applications. Apart from this, there is a strong focus on personality fit as they believe that passion and attitude go longer than knowledge and skills when it comes to screening candidates. In terms of educational background, while Bridgei2i looks for candidates with a strong academic background in a quantitative discipline, it is the business problem-solving, analytical, and customer-centric attitude they prioritise the most. Apart from engineers and statisticians, they are also open to candidates with different backgrounds if they exhibit a keen passion for data-related problem-solving. At Bridgei2i, data scientists primarily use artificial intelligence and ML-techniques to solve analytical problems for clients. An integral part of the AI Innovation Labs involves the development of advanced AI accelerators. “Our data scientists often alternate between various roles – that of strategists, inventors, or storytellers!” she said. Talking about how an ideal data science candidate should look, Preeth is quick to add that a perfect candidate should be a team player who is good at solving data analytics problems and finding new innovative ways of solving business problems. “Ability to work well with a diverse set of people is crucial since strategising and collaborating across teams is a big part of what we do, and how we deliver best in class solutions to our clients,” she added. Data Scientist Hiring Process A typical hiring process at BRIDGEi2i consists of multiple rounds of assessments, which usually begins with an initial screening of the resumes followed by numerous rounds of interactions and interviews. The interview round involves questions around their skill sets, prior experiences, and domain expertise. As mentioned earlier, considerable stress is applied to match applicants’ personalities with their ecosystem and culture! Talking about the challenges while hiring, Preeth said that while there’s no dearth of candidates for data science positions, one of the major challenges they face is the candidate’s lack of hands-on experience with the technical aspect of their projects. “One of the issues we face while hiring is that candidates do not always have all the skills needed to work and excel in the team successfully. Determining the right cultural fit may also take some time,” shared Preeth. He further added that they overcome these gaps with structured programs such as LeGo and SCaLa to nurture and cultivate learning as a culture. To hire the best candidates, they usually rely on employee referrals, careers page, job portals, and social media campaigns. The company is currently hiring analytics consultants and machine learning enthusiasts passionate about Natural Language Processing or Computer Vision. Interested candidates can apply in the careers section of their website. Growth Opportunities For Data Scientists As Preeth shares, Bridgei2i encourages new talent with the right skills and attitude to experiment, learn, and grow. In fact, to begin with, they induct all the employees into the hallway of learning through their Centre of Excellence (CoE) — SCaLa, which features comprehensively-designed courses, and seasoned subject matter experts as instructors. “When it comes to providing a good career foundation for our employees, our capability development and skilling platform, SCaLA (Short for Skills, Capabilities, Leadership and Ascension) ensures that employees are given the best learning and growth opportunities in whichever career path they choose,” she said. Some of the other initiatives by the company are: Learn & Grow (LeGo) initiative that prepares students recruited from campuses to adjust to the professional environmentACCENDO, which is their annual innovation fiesta modelled on a hackathon. It invites participation and collaboration among different teams to create unique solutions, among others. “We also have a data-driven, merit-based recognition system that factors both effort and outcomes with a strong focus on contribution to BRIDGEi2i values,” added Preeth. Apart from this, they encourage employees to work in cross-industry and cross-function roles across geographies. “At BRIDGEi2i, we believe in nurturing our talent pool with value-based learning and offer a structured growth path — be it leadership or subject matter experts,” said Preeth on a concluding note.","excerpt":"With strong AI and analytics operations, Bridgei2i has over 600 data scientists forming the backbone of the team. As the company focuses on delivering AI-powered transformation, they rely on data science, technology, and business applications to provide the best outcome — for which the data science team plays a key role.  To drive a large […]","categories":["AI Hirings"],"tags":["Applications of Data Mining","big data for social good","bridgei2i","Data Mining","data mining applications","data mining tools","data science job roles","Data Scientist","impact of big data in education","Natural Language Processing"],"author_name":"Srishti Deoras","publish_date":"2020-11-27T17:00:07","publication_year":"2020","word_count":789,"keywords":["computer vision","data mining tools","impact of big data in education","data mining applications","R","data science","artificial intelligence","Data Mining","Natural Language Processing","RAG","analytics","Go","machine learning","AI","data science job roles","big data for social good","ML","Applications of Data Mining","bridgei2i","Data Scientist"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","data science","analytics","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/what-is-the-hiring-process-for-data-scientists-at-bridgei2i\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":574,"title":"Teradata unveils Analytics of Things Accelerators: Pushes Slow-Moving IoT Projects to the Fast Lane","content":"Teradata, a leading analytics solutions company, recently announced “Analytics of Things Accelerators” (AoTAs). These are four powerful software-and-service solutions that speed up the transformation of Internet of Things (IoT) data to actionable insights. The company launched IoT analytics based on proven IP to transform sensor data streams into ROI streams. Teradata’s AoTAs, available immediately, are derived from field engagements at the world’s largest, most innovative IoT companies in manufacturing, transportation, mining, energy and utilities. The solutions comprise of technology-agnostic intellectual property (IP) and professional services, applied approaches proven to reduce implementation cost and risk, accelerate time to value, and drive business returns far greater than the initial investments. The Internet of Things is blending the physical and digital worlds transforming industries and the way we live and work. Yet there lies challenges in generating business value from enormous volumes of IoT sensor data. This is where Teradata’s new solution provides direction and differentiation as well as ROI. Teradata AoTAs help organizations determine what sensor data to trust and keep, while choosing types and combinations of analytical techniques to best address specific business questions. The accelerators help organizations move from ‘brute force’ one-off projects to enterprise-class solutions scaling across thousands of complex devices and countless assets that result in continuous positive business impact. Specifically, the Teradata AoTAs consist of: 1) A Condition-Based Maintenance Accelerator, which continuously monitors and analyzes asset data at scale to increase availability, improve safety, and reduce costs; 2) A Manufacturing Performance Optimization Accelerator, which identifies complex production problems across equipment performance and availability for quick corrective action 3) A Sensor Data Qualification Accelerator, which automates recommendations on the optimal frequency of sensor readings based on relevant anomaly patterns. Additionally, Teradata offers a Visual Anomaly Prospector Accelerator, which mines large amounts of multidimensional time series (MTS) data, and visually helps an end user discover anomaly patterns that frequently precede a key event. “Based on interviews that I conducted with eight Teradata customers across several industries, IoT applications at those companies are generating impressive business value. As shown in our study, this value comes from the analytics applied to sensor data that is blended with traditional data about customers, products, and the like,” said Dr. Richard Hackathorn, president and founder of Bolder Technology, Inc. “By leveraging the expertise from previous IoT engagements, these new Analytics of Things Accelerators can enable companies to realize value faster and enhance success for their IoT projects.” “Internet of Things (IoT), across both the enterprise and consumer segment is expected to grow exponentially over the next decade. According to industry sources, 75 percent of businesses in India in the utilities, telecom, transportation and oil\/gas industry are tapping or have plans to implement Internet of Things. These are likely to be in segments as diverse as security and surveillance, supply chain management, inventory and warehouse management and customer order monitoring among others. We therefore foresee India adding significant value to the overall IoT market in the near future and Teradata’s ‘Analytics of Things Accelerators’ can enable companies to incur insights faster and more efficiently from their IoT deployments”, said Mr. Sunil Jose, Managing Director, Teradata India. The study revealed highest levels of deployment and planned deployment among businesses in the utilities and telecom, chemicals and oil\/gas, as well as transportation. Common deployments of IoT involved security and surveillance, supply chain management, inventory and warehouse management, and customer order monitoring. “Teradata AoTAs are already addressing and resolving $100 million-dollar problems for premier producers of vehicles, equipment, oil and gas systems, and consumer goods,” said Oliver Ratzesberger, Executive Vice President and Chief Product Officer, Teradata. “These challenges represent billion-dollar budgets for each company, to be clear on the scale of business value addressed by AoTA. For example, our AoTAs have increased Overall Equipment Effectiveness as much as 85 percent, while also improving predictability and asset availability. We are seeing a lot of excitement around our Accelerators, because the return on investment is transformational in scope and compelling in business impact.”","excerpt":"Teradata, a leading analytics solutions company, recently announced “Analytics of Things Accelerators” (AoTAs). These are four powerful software-and-service solutions that speed up the transformation of Internet of Things (IoT) data to actionable insights. The company launched IoT analytics based on proven IP to transform sensor data streams into ROI streams. Teradata’s AoTAs, available immediately, are […]","categories":["AI News"],"tags":["big data processing interview","IoT"],"author_name":"Manisha Salecha","publish_date":"2016-09-02T09:58:32","publication_year":"2016","word_count":663,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","big data processing interview","analytics","Rust","GAN","R","IoT"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Rust","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/teradata-unveils-analytics-things-accelerators-pushes-slow-moving-iot-projects-fast-lane\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":65181,"title":"AI Facial Recognition Vs Contact Biometric Systems Amid Covid-19","content":"The Covid-19 pandemic – and the new reality unveiled by this crisis – has not only disrupted lives but also the arc of several technologies. In what has come to be a significant blow, contact biometric technologies may not only have become redundant but also potentially lethal given the risk it carries in spreading the infection. Automated fingerprint analysis has been a widely adopted form of biometrics for decades, spanning many applications. This includes marking staff attendance at workplaces, identity registration at government offices, as well as entry authentication at various places, to name a few. While thumb impressions through a biometric system were commonly used to verify an individual, these practices have been suspended to mitigate physical contact and prevent the spread of Covid-19. Although this could be the last nail in the coffin for this technology, it has also triggered opportunities for the development and improvement of AI-based biometric technologies. Although some companies have already adapted their offerings to integrate touchless fingerprint technologies, a primary option that is currently being explored by many companies and governments is facial recognition. Search For Contactless Biometric Systems With hygiene concerns expected to peak once lockdown measures begin to ease, contact biometric devices for public use are likely to get terminated. This is expected, given the challenges it could set off in crowded establishments, such as workplaces. If an infected employee uses a fingerprint scanner, they will trigger a catastrophic domino effect. When their colleague checks in using the same scanner, they, in turn, will be carriers of the virus, infecting all surfaces they touch. And given that the coronavirus can survive several hours on any metallic or plastic surface, this could escalate very quickly. ALSO READ: How Has Facial Recognition Technology Adapted To The Pervasive Use Of Masks Little wonder then that the Indian Government has suspended fingerprint authentication in all its offices. And while this was one of several protocols instructed by states, some are exploring AI-powered alternatives and running pilots to measure its potency. In fact, the Government of Telangana has been looking for vendors who can provide an effective and immediate solution. According to a report, its facial recognition pilots involve retrofitting existing biometric infrastructure with a security system enabled by AI facial recognition technology to deliver a contactless attendance system. And with this, many facial recognition companies are seeing a business opportunity. Spike In AI Facial Recognition Technologies Some Chinese tech companies, including SenseTime, Megvii, and even top enterprises like Alibaba and Baidu, are pitching AI facial recognition technologies as sanitary alternatives to traditional fingerprint scanners. For instance, SenseTime enables users to detect faces within ‘milliseconds’ when embedded on devices. Built on a deep learning platform, it claims to detect faces from different angles under different environments. This covers side-profiles, low light scenarios, as well as different facial expressions, among others. What is more, it recently announced that it had adapted its product to identify even masked people – a scenario that is likely to become mainstream once offices reopen. Another company based out of Germany, DERMALOG, has been developing a solution that integrates facial biometrics and temperature detection for workplace deployments. The biometrics company, known for creating fingerprint, iris, and facial recognition hardware, has adapted its technology to meet current demands. These developments align with the forecast of ABI Research that predicted a spike in facial recognition technology from longer-term investments in new contactless identification systems. Triggered by a drastic drop in fingerprint device revenue this year, the report had acceded that AI-powered biometric firms will develop solutions to meet the challenges borne by this ‘forced evolution’ of biometric technologies. The report suggested the sale of fingerprint devices to slide by $1.2 billion this year, but also signalled a spike in contactless attendance systems, spearheaded by AI facial recognition investments, by as much as by 20%. Many of these companies, like DERMALOG, are bundling facial recognition and temperature-sensing technology to make it a more effective solution amid this pandemic. China’s Telpo has been offering multi-person facial recognition, along with temperature screening sensors for quick identification as well as fever detection. Leveraging infrared thermography and AI, it performs mask detection as well, much like SenseTime, and other companies like Hanvon and Herta, to name a few. In addition to workplaces, these technologies are being explored for authentication and identification in other arenas as well, including airports, railway stations, schools, universities, and other places frequented in groups. The replacement of fingerprint scanners with facial recognition systems will, especially, drive the travel industry, where it can possibly be made the default form of identification for passenger identity checks and boarding. ALSO READ: Can Big Data & Biometrics Rewrite The Future Of Travel? Outlook Although the biometrics industry could lose $2 billion this year from the impact of the pandemic, AI facial recognition market is likely to grow as its algorithms get updated to work around facial occlusions. With the pandemic seeing ubiquitous adoption of masks and other protective gear that may partially cover faces, these improvements to biometric systems led by facial recognition will meet the current demands of the world.","excerpt":"The Covid-19 pandemic – and the new reality unveiled by this crisis – has not only disrupted lives but also the arc of several technologies. In what has come to be a significant blow, contact biometric technologies may not only have become redundant but also potentially lethal given the risk it carries in spreading the […]","categories":["Deep Tech"],"tags":["AI facial recognition","china ai investments","covid-19"],"author_name":"Anu Thomas","publish_date":"2020-05-16T18:00:00","publication_year":"2020","word_count":852,"keywords":["big data","Go","AI facial recognition","covid-19","AI","programming_languages:R","Scala","RAG","china ai investments","Aim","deep learning","GAN","R"],"extracted_tech_keywords":["AI","deep learning","Aim","RAG","R","Go","Scala","big data","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ai-facial-recognition-vs-contact-biometric-systems-amid-covid-19\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":86,"title":"Analytics Quotient","content":"Analytics Quotient (http:\/\/www.aqinsights.com) leverages analytics to help translate existing data into profitable marketing insights. With AQ’s simple customized solutions, plug and play tools, pay-as-you go pricing and globally delivered consulting services, AQ uses analytics to service their clients’ distinctive business needs. Analytics Quotient was incepted in the year 2008 by five entrepreneurs and today comprises of 120 talented professionals who use game theory and ANOVA to explain market dynamics and consumer behavior. AQ’s clientele includes the world’s largest non-alcoholic beverage company, World’s largest casual dining restaurant chain and North America’s largest confectionery company to name a few. Pritha Choudhuri – CEO, Analytics Quotient Pritha has 8 years’ experience in analytics. Her marketing analytics work includes projects in the areas of consumer segmentation, pricing and drivers’ analysis. She has worked for clients such as Procter & Gamble and Universal Studios. For these clients, her focus has been to build offshore capabilities and develop teams to deliver analytics. Prior to her analytics career, Pritha has worked in Colgate Palmolive in India in the Sales and Marketing functions.  Pritha is the CEO at AQ. Other than managing company operations she is actively involved in business development, new business delivery, managing resources, capability building, marketing and implementing scale and structure for AQ’s growth. She also leads the delivery team for one of the AQ’s main clients. Analytics Quotient has offices in Bangalore, Atlanta and Zurich.","excerpt":"Analytics Quotient (http:\/\/www.aqinsights.com) leverages analytics to help translate existing data into profitable marketing insights. With AQ’s simple customized solutions, plug and play tools, pay-as-you go pricing and globally delivered consulting services, AQ uses analytics to service their clients’ distinctive business needs. Analytics Quotient was incepted in the year 2008 by five entrepreneurs and today comprises of 120 […]","categories":["Deep Tech"],"tags":[],"author_name":"Дарья","publish_date":"2012-05-25T15:35:45","publication_year":"2012","word_count":231,"keywords":["Go","programming_languages:R","AI","llm_models:PaLM","programming_languages:Go","RAG","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","llm_models:PaLM","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/analytics-quotient\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10062679,"title":"Top 15 forgotten ML algorithms","content":"The ‘nearest neighbour algorithm’ in 1967, one of the first machine learning algorithms ever, was conceived, laying the foundations for basic pattern recognition. Today, in 2022, we are no short of brilliant algorithms like linear regression, KNN, SVM, decision trees and more. Machine learning algorithms are integral to helping computer systems improve their performance and automatically build a mathematical model using training data. It is owing to these algorithms that machines can make decisions without specific human instructions all the time. Through the years of computer science, algorithms have been developed and updated, and many of them are lost in space. Analytics India Magazine will take you down this time machine of some of the most popular yet forgotten machine learning algorithms. Survival analysis Survival analysis, part of statistics, analyses the expected duration of time for the occurrence of any event. It can be applied for tasks such as duration analysis or proportional questions, answering what part of the population will survive in a certain time or when will the current population become extinct, and more. In engineering, this is also known as reliability theory or reliability analysis because it helps study the reliability of any system that is going to close or destroy one day. Evolutionary algorithms As a subset of evolutionary computation, evolutionary algorithms are generic population-based metaheuristic optimisation algorithms. They provide a heuristic-based approach to solving problems that cannot be easily solved in polynomial time. The approaches used are inspired by biological evolution and occur in four processes: initialisation, selection, genetic operators, and termination. Hidden Markov Models A Hidden Markov Model (HMM) is a statistical model that describes the evolution of observable events. These events are not directly observable and depend on internal factors. A varied class of probabilistic graphical models allows us to predict a sequence of unknown variables from a set of observed variables. Manifold learning An approach to non-linear dimensionality reduction, manifold learning algorithms believe that the dimensionality of data sets is only artificially high. The algorithm generalises linear frameworks to be sensitive to non-linear structures in the data. This is usually unsupervised since it learns the structure from the data itself. Traditional signal processing Signal processing algorithm analyses, modifies and synthesises sound signals, images, and scientific measurements. These are typically built on three basic functions: Add, Multiply, and Delay. The functions are combined to build up complex algorithms in discrete-time systems. Traditional signal processing techniques include filtering, detection, estimation and frequency domain analysis. Adaptive Resonance Theory (ART) Adaptive resonance theory has been inspired by how the human brain processes information. The algorithm conceptualises supervised and unsupervised learning methods to train neural networks addressing pattern recognition, object identification and prediction. Self-Organising Maps (SOMs) A self-organising map is an unsupervised neural network trained using unsupervised learning techniques to produce a low dimensional, a discretised representation based on the input space of the training samples, known as a map. It is a method to reduce data dimensions. They apply competitive learning techniques to preserve all the topological properties within the input space. Bayesian nonparametric literature Nonparametric statistics are usually distribution-free or have a specified distribution. The distribution’s parameters are unspecified and include descriptive statistics and statistical inference. Edge detection The edge detection algorithm uses mathematical methods to identify edges or curves in a digital image where the image brightness has discontinuities. This algorithm is fundamental to image processing, machine vision and feature detection and extraction in computer vision. Hierarchical mixed-effects regression Hierarchical modelling algorithms deal with data and their observations in a certain group. Hierarchical linear models or mixed-effect models are chosen approaches that allow researchers to account for ecological, contextual, and individual-level or compositional variables. Gradient maps Gradient maps provide basic visualisations of spatial areas and their associated values depending on the thematic issue. Gradient maps follow the location map style, illustrating values like the region to visualise data with spatial context. Symbolic Regression A type of regression analysis, symbolic regression algorithm, finds the appropriate model based on the given dataset and mathematical expressions. These initial expressions are outputs of random combinations of mathematical concepts like mathematical operators, analytic functions, constants, and state variables. Conditional Random Fields A class of statistical modelling, conditional random fields algorithms help in pattern recognition and structured prediction. CRF models the predictions as graphical models to represent the various dependencies between predictions and account for different variables. Polynomial regression A polynomial regression algorithm is a type of regression analysis. The regression algorithm models the relationship between a dependent (y) and independent variable (x) as ‘nth‘ degree polynomial. Essentially, it provides the best approximation of the relationship between the dependent and independent variables. Multi-dimensional scaling The multi-dimensional scaling algorithm is used to visualise the similarity level of individual cases of a dataset. This algorithm falls under the branch of unsupervised ML and is a good technique to preserve the global and local structures of high-dimensional data.","excerpt":"An approach to non-linear dimensionality reduction, manifold learning algorithms believe that the dimensionality of data sets is only artificially high.","categories":["AI Trends"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-03-14T14:00:00","publication_year":"2022","word_count":815,"keywords":["Go","machine learning","TPU","AI","neural network","ML","Git","computer vision","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","computer vision","analytics","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-15-forgotten-ml-algorithms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10055841,"title":"DeepMind Now Wants To Study The Behaviour Of Electrons, Launches An AI Tool","content":"Density functional theory (DFT) describes matter at the quantum level, but popular approximations suffer from systematic errors that have arisen from the violation of mathematical properties of the exact functional. DeepMind has overcome this fundamental limitation by training a neural network on molecular data and on fictitious systems with fractional charge and spin. The result was the DM21 (DeepMind 21) tool. It correctly describes typical examples of artificial charge delocalization and strong correlation and performs better than traditional functionals on thorough benchmarks for main-group atoms and molecules. The company claims that DM21 accurately models complex systems such as hydrogen chains, charged DNA base pairs, and diradical transition states. The tool DM21 is a neural network to achieve the state of the art accuracy on large parts of chemistry and to accelerate scientific progress; the code has been open-sourced. DFT is a crucial technology and is essential to designing functionals that get simple chemistry correct before explaining more complex molecular interactions. To solve some of the major challenges of the 21st century, like producing clean electricity or developing high-temperature superconductors, it is essential to design new materials with specific properties. Doing this on a computer requires simulation of electrons, subatomic particles that govern the way in which atoms bond to form molecules and are also responsible for the flow of electricity in solids. Despite years of effort and several significant advances, it is still a challenge to model the quantum mechanical behaviour of electrons accurately. Source: DeepMind The team has addressed two long-standing problems: The delocalization error: In a DFT calculation, the functional determines the charge density of a molecule by finding the configuration of electrons which minimizes energy. Thus, errors in the functional can lead to errors in the calculated electron density. Most existing density functional approximations prefer electron densities that are unrealistically spread out over several atoms or molecules rather than being correctly localized around a single molecule or atom.Spin symmetry breaking: When describing the breaking of chemical bonds, existing functionals tend to unrealistically prefer configurations in which a fundamental symmetry known as spin symmetry is broken. Since symmetries play a vital role in our understanding of physics and chemistry, this artificial symmetry breaking reveals a major deficiency in existing functionals. Source: DeepMind These longstanding challenges are both related to how functionals behave when presented with a system that exhibits “fractional electron character.” DeepMind found that the problems of delocalization and spin symmetry-breaking can be solved by using a neural network to represent the functional and tailoring their training dataset to capture the fractional electron behaviour expected for the exact functional. Their function showed itself to be highly accurate on large-scale benchmarks, suggesting that the data-driven approach is capable of capturing aspects of the exact functional.","excerpt":"DeepMind solves delocalization and spin symmetry-breaking by using a neural network and training the dataset to capture the fractional electron behaviour","categories":["AI News"],"tags":["DeepMind","Quantum Computing"],"author_name":"Meeta Ramnani","publish_date":"2021-12-13T15:13:19","publication_year":"2021","word_count":457,"keywords":["Quantum Computing","Go","programming_languages:R","AI","neural network","data-driven","programming_languages:Go","Aim","ViT","R","DeepMind"],"extracted_tech_keywords":["AI","neural network","Aim","R","Go","ViT","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deepmind-now-wants-to-study-the-behaviour-of-electrons-launches-an-ai-tool\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":21652,"title":"Gaurav Kumar Takes Charge As Chief Operating Officer At The Smart Cube","content":"In a recent development, The Smart Cube, an award-winning global provider of research and analytics solutions, announced the appointment of Gaurav Kumar as it new Chief Operating Officer. Kumar, who first joined the company as Research Analyst in 2005 will lead the company’s operation as a part of his new role. Since his initial days at The Smart Cube, Kumar has worked his way up to Associate Vice President, before joining Fidelity Investments in 2012 to grow its analytics and market intelligence function. In 2015, he was selected to join Fidelity’s elite GMA Leadership Program, involving four rotations across global operations. After completion, Gaurav was appointed to the position of Director of Digitisation, which focused on leveraging analytics and technology to enhance customer experience. “Hiring into such a strategically important role is a daunting task, but we have found the perfect person in Gaurav, combining deep understanding of our core business with a fresh mindset and relevant learnings from his previous experience”, commented Gautam Singh, Co-Founder and Chief Executive Officer of The Smart Cube. He added “He joins at a very exciting time, following the recent launch of our new sector-focused solutions, and his wealth of experience across financial services, consulting, pharmaceuticals, technology, and oil and gas, will be hugely beneficial. I’m very much looking forward to working with him again, to drive the next stage of our evolution”. As a part of his current role, Kumar will pick up the responsibilities previously held by Managing Director Sameer Walia, one of The Smart Cube’s three founders, and also strengthen the remit of the executive team. Walia, who has been in the industry for 15+ years, will step aside to pursue his lifelong dream of working full-time in the not-for-profit space, dedicating his time to new social causes. He will retain a non-executive role and his stake in the business. Commenting on the new appointment Walia said “Upon first meeting Gaurav I knew that I wanted him at The Smart Cube. His intellect, hard work and personable nature saw him rise quickly – a record five promotions in less than seven years – so I am thrilled to welcome him back to lead our operations. I am proud of our strong culture, work ethic and high client retention rate, and know Gaurav will take this from strength to strength.” Gaurav Kumar, Chief Operating Officer of The Smart Cube, said “From day one, The Smart Cube’s influence on me was significant. The strong fundamentals – both in terms of business operations, the culture and values – laid the foundations for my career. I’m extremely excited to be back, with new skills and experience gained across analytics, technology, product development and, above all, people and operations management. From being client-side, I hope my fresh perspective and understanding of what clients are looking for, will allow me to add new value.” Headquartered in the UK with additional offices in the USA, Switzerland, Romania and India, The Smart Cube is a global provider of research and analytics solutions, primarily serving the CPG, Energy and Chemicals, Financial Services, Industrials, Life Sciences and Retail sectors. The company works with a third of companies in the Fortune 100, helping them make smarter decisions, accelerate value and gain a competitive edge.","excerpt":"In a recent development, The Smart Cube, an award-winning global provider of research and analytics solutions, announced the appointment of Gaurav Kumar as it new Chief Operating Officer. Kumar, who first joined the company as Research Analyst in 2005 will lead the company’s operation as a part of his new role. Since his initial days […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2018-02-13T06:13:04","publication_year":"2018","word_count":542,"keywords":["programming_languages:R","AI","Git","RAG","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/gaurav-kumar-takes-charge-chief-operating-officer-smart-cube\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044495,"title":"10 AI-Based Recruiting Tools","content":"Artificial Intelligence (AI) helps companies with predictive analysis to automate recruitment. It does so by matching candidates to the most relevant job profiles with the help of machine learning. AI for recruiting includes the usage of: voice-enabled chats, search, and personal assistanceMachine learning-powered pattern matching Chatbots to automate interview scheduling As of 2020, 65 percent of HR organisations already use AI for talent acquisition and by 2023, 77 percent of HR organisations are expected to use AI for recruitment. Today, we take a look at ten emerging AI-backed recruiting tools used by companies: Skillate Skillate is an advanced decision-making engine that accelerates and smoothens the process of hiring. It offers AI-powered intelligent hiring services, improved candidate experiences, and people analytics. Additionally, Skillate enables businesses to mask personal information about candidates to enable unbiased hiring decisions. Skillate uses deep learning to extract information from resumes, and attract better talent by writing job descriptions with the help of AI-based JD assistants. Additionally, it offers auto-interview scheduler, job description assistant and data-backed insights and analytics of the recruitment process. Skillate is used by Trell, BYJU’s. Ola, BigBasket, CleverTap, Grofers, Rapido, Sony, Mahindra, Larsen & Toubro, among others. For more information, click here. TurboHire Intelligent recruitment platform TurboHire is data-driven, structured and collaborative. It combines the power of human and machine intelligence to generate data-centric insights to source, screen and assess candidates. Additionally, TurboHire offers military-grade security of customer data and encrypted communication between customers and cloud. According to the website, with TurboHire, companies have saved more than 10,000 hours of recruiting time and $2 million hiring cost while recruiting more than 50,000 candidates. TurboHire is used by more than 5,000 plus recruiters including Cure.fit, ICICI Bank, Wakefit, ClearTax, and Accel Partners. For more information, click here. Talenture Talenture offers recruitment with AI and ML engine insights. Its virtual recruitment expert Marvin monitors and guides job processing for businesses. It uses Intelligent Heuristics to rank candidates on parameters like recency, availability, employability or a combination of these three. Additionally, its CRM provides a 360-degree view of businesses relationships with candidates to accelerate decision making. Talenture offers candidate sourcing services from job boards, social sites, career sites, email and desktop folders. It also has a mobile application Talenture has been used by more than 1,000 companies to recruit seven million plus jobs. Its clients include the likes of JobExcel, GenNEXT, Recruise, Venpa Staffing Services, and SR Corporate Services, among others. For more information, click here. Zoho Recruit Cloud-based applicant tracking system Zoho Recruit provides end-to-end hiring solutions for staffing agencies, corporate HRs and temporary workforces. Zoho’s AI recruitment software uses automation to help its clients and businesses source candidates, screen resumes, develop process workflows and engage with candidates in-real time with the help of chatbots. Zoho Recruit helps companies curate a list of the best candidates for a particular job profile with the help of candidate matching software Zia. Zia matches candidates’ resumes with job descriptions, locations, and industries, among other factors, reducing the candidate shortlisting process to a few milliseconds. Zoho Recruit is used by OnePlus, Airtel, SpiceJet and Yokohama, among others. For more information, click here. Pymetrics Pymetrics combines gamification and behavioural technology to ease the process of hiring. It has a patented set of 12 games to fairly and accurately measure cognitive and emotional attributes. It has an additional set of four games to measure numerical and logical reasoning. It claims to reduce the hiring time by 75 percent, increase the yield by 100 percent, and decrease hiring costs by 25 percent.  Pymetrics also ensures gender and ethic fairness in algorithms with ethical AI design with organisations including EEOC, Stanford University and World Economic Forum. It has a marquee list of clients of more than 60 global enterprises including Tesla, Unilever, Accenture and Mercer. For more information, click here. Mya With the help of Mya– a conversational AI recruiting platform, companies will be able to recruit faster, reduce hiring costs and improve the whole hiring experience. It automates candidate engagement and communications at scale. It does so by converting site visitors into applicants and talent pools. Using its proprietary natural language system, Mya extracts data and insights from conversations with candidates. Post which insights from structured and unstructured responses are used to update candidate profiles, perform direct comparisons and surface actionable conversation analytics. According to its website, Mya can automate candidate engagement for more than 12 industries and 100s of roles. It is used by more than 300 enterprise brands, including the likes of Loreal Paris. For more information, click here. Talocity SaaS platform Talocity enables ‘touchless’ hiring. The AI-based video interviewing platform Talocity is available in 39 languages. It offers four solutions: One-way video interviews Voice assessmentAssessment platform Typing assessment ATS AI Proctoring During its one-way video interviews, Talocity collects tens of thousands of data points on the candidate. The AI then augments the hiring team’s decision-making with insights from these interviews, allocating time and assets only on the candidates with the most compatible potential. Talocity is used by the likes of Infosys, Mahindra, ICICI Prudential, Pantaloons, HCL Urban Company and Aditya Birla Fashion and Retail Ltd. For more information, click here. Hiretual Hiretual AI builds a strong pipeline for present and future roles, casting a wider net for search across more than 750 million plus profiles on 45-plus open web platforms such as GitHub, AngelList, RateMDs and Upwork. Additionally, it provides market insights for hiring managers– providing a comprehensive summary of the market for a company’s open roles to guide targeted decisions. The platform keeps refreshing the Applicant Tracking System data for maximum visibility and faster optimisation. Hiretual is used by more than 100,000 recruiters. Companies including PayPal, IBM, Philips, Deloitte, Accenture and PWC use Hiretual. For more information, click here. HireVue HireVue offers live and on-demand video interviewing at scale, pre-hire assessments to identify the best candidates and minimise bias; conversational AI to automate workflows through text, web and WhatsApp interfaces; and provide structured interview guides to replace ad hoc and inconsistent interview processes. HireVue offers reduced unconscious bias as its USP over competitors. It’s AI-driven approach mitigates bias by eliminating unreliable and inconsistent variables– selection based on resume and phone screens– thus, letting recruiters focus more on job-relevant criterias. HireVue claims to decrease the time to fill up a position by 90 percent, increase diversity by 16 percent and experience return on investment up to 131 percent, within the first year of using the platform itself. Its clients include Unilever, Rockspace, Boston Red Sox, G4S, and the Foxtel Group. For more information, click here. Loxo Loxo offers a talent intelligence platform, an applicant tracking system and a Recruiting CRM. The AI recruitment automation software guides recruiters to the best suited candidates while reducing the work by 80 percent. Additionally, its proprietary directory consists of more than 530 plus million people across hundreds of data sources. Presently, it serves more than 1.25 lakh recruiters across the globe including Amazon, World Wide Technology, Kensington International and Cigna. For more information, click here.","excerpt":"As of 2020, 65 percent of HR organisations already use AI for talent acquisition and by 2023, the number is expected to be 77 percent.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","ai in recruitment","video processing ai","zoho"],"author_name":"Debolina Biswas","publish_date":"2021-07-26T16:00:00","publication_year":"2021","word_count":1172,"keywords":["ai in recruitment","Go","zoho","artificial intelligence","machine learning","AI","chatbots","ML","video processing ai","Aim","deep learning","analytics","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","analytics","Aim","chatbots","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-ai-based-recruiting-tools\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":68121,"title":"How Far Can Kaggle Help You In Your Professional Career?","content":"Any data scientist worth his salt would be familiar with Kaggle. The platform has emerged as an ideal medium for data science practitioners to use and sharpen their skills as they navigate this dynamic field. In fact, its popularity has also caught the attention of recruiters who have begun to use Kaggle achievements as one of many metrics to gauge the competitiveness of candidates. But how far can Kaggle help data scientists in their professional careers? We reached out to some who weighed in with their thoughts: Value of Kaggle Rankings Having a good ranking under the ‘Competitions Category’ on the platform can greatly help in the career advancement of participants. Not only does it demonstrate their skills, but also offers a window into other qualities they may possess, including time management, persistence, as well as consistency. “To achieve a good ranking on Kaggle, individuals need to persistently dedicate several hours of time over a period,” says Usha Rengaraju, a data science consultant, who is incidentally India’s first female Kaggle Grandmaster. “Dedication, consistency, innovation, out-of-the-box thinking, ability to not give up and collaboration – all these are key traits of a top kaggler, and they are also the key skills which any employer may look for,” he adds. According to her, although Kaggle problems and real-world business issues can be seemingly different for certain domains, competition ranking serves as a better metric to evaluate candidates on their data science abilities. “There are stories of several Grandmasters – some of them with no structural education – landing up high profile jobs,” she adds. Chimes in Prashant Kikani, who works at a Bengaluru-based startup, “This is especially relevant for amateur data scientists and allows them to gain an edge over others. What is more, in order to get a higher ranking, one must be in the know of what is latest in terms of research, Kaggle can not only augment your job profile but also make you more informed.” ALSO READ: How To Initiate Your Kaggle Journey Wide Exposure Although Kaggle is known to fine-tune a data scientist’s skills, it is also a great platform to learn from seasoned data science professionals as well. Although senior data science professionals are largely missing from the platform, the ones who are active can provide a wealth of information to aspiring data scientists. “There are some very great kernels prepared by seasoned AI engineers,” says Saurabh Jha, Director – Data Science at Dell. “Discussion forums are helpful too. It gives an exposure to all kinds of problems – from Computer Vision, Text, Recommendation Engine, etc,” he adds. These problems, although different from real-world issues, help push the boundaries of a user’s abilities, as some take a lot of time to solve and demand a lot of collaboration with other kagglers. In a world where employers are relying on non-traditional hiring methods, the wide exposure that Kaggle affords makes its credentials worth a lot to recruiters. “Another important aspect is that candidates learn the art of critical thinking,” says Vidhya Veeraraghavan, Head of Analytics – Financial Markets Operations at Standard Chartered. “Apart from the programming skills and implementation of ML algorithms, candidates can also learn to approach the problem statements and the datasets with multiple solutions,” she adds. ALSO READ: How To Find Success In Kaggle – What Masters Recommend True Demonstration of Working Knowledge Although the surge in data science courses has enabled enthusiasts to pick up critical skills, only a few of these learners have opportunities to work on real-life projects. This is where platforms like Kaggle help fill the gap. “To become job-ready or excel in their data science jobs, candidates need to have some hands-on experience, and Kaggle allows just that,” says Vidhya. She illustrates this with an example: “Consider a three-tiered Pyramid concept. Acquired knowledge occupies the bottom-most tier, working knowledge will be in the middle, followed by domain knowledge at the top-most tier. In my opinion, to have a professional career in data science, a candidate’s working knowledge, that is, the middle tier, is imperative.” Since Kaggle helps candidates work on real datasets and compete with other elite minds, it gives them the opportunity to sharpen and hone the ‘middle tier’.“Candidates can test their knowledge in the basics of programming languages, machine learning algorithms and its implementation too,” says Vidhya. “This definitely gives a boost to their resume and if articulated well, can help them ace their job interviews,” she adds.","excerpt":"Any data scientist worth his salt would be familiar with Kaggle. The platform has emerged as an ideal medium for data science practitioners to use and sharpen their skills as they navigate this dynamic field. In fact, its popularity has also caught the attention of recruiters who have begun to use Kaggle achievements as one […]","categories":["AI Features"],"tags":["Data science skills","Kaggle"],"author_name":"Anu Thomas","publish_date":"2020-06-25T13:00:00","publication_year":"2020","word_count":741,"keywords":["data science","Go","machine learning","Kaggle","AI","R","ML","innovation","computer vision","RAG","analytics","Data science skills"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","data science","analytics","RAG","R","Go","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-far-can-kaggle-help-you-in-your-professional-career\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10171572,"title":"Each ChatGPT Query Uses Merely a 15th of a Teaspoon of Water, Says Sam Altman","content":"OpenAI CEO Sam Altman has disclosed for the first time the average energy and water consumption of a single ChatGPT query, which is much less than what analysts and environmentalists had projected all this time. “The average query uses about 0.34 watt-hours,” Altman said in a blog post titled The Gentle Singularity. “It also uses about 0.000085 gallons of water; roughly one-fifteenth of a teaspoon.” For comparison, 0.34 watt-hours is equivalent to running a high-efficiency lightbulb for a couple of minutes. According to a study on ChatGPT’s water consumption by A. Shaji George, an expert in Information and Communications Technology (ICT), the AI chatbot consumes 0.5 litres of water during each of its lengthy conversations with a user. This applies to all AI systems and LLMs in place. Half a litre of water would mean the amount required to cook two packets of Maggi instant noodles. According to data provided by Semrush, ChatGPT was ranked as the eighth most visited site in the world as of March 2025, with approximately 5.56 billion visits. Meanwhile, research shows that every question potentially uses around 10 times more electricity than a simple Google search, with an average of 2.9 watt-hours of energy. However, Altman’s estimate is nearly one-tenth of what analysts projected. Previously, he shared that polite courtesies with ChatGPT, such as “please” and “thank you,” have cost the company tens of millions of dollars in electricity expenses. This revelation comes amid increasing scrutiny of AI’s resource usage as models grow in size and adoption. Altman also laid out a longer-term forecast on AI cost trends, stating, “As datacenter production gets automated, the cost of intelligence should eventually converge to near the cost of electricity.” Environmental analysts note that while per-query usage is low, aggregate consumption remains a concern. With hundreds of millions of queries each day, energy and water use from AI data centres could grow substantially. In the same post, Altman acknowledged these broader implications, saying, “The economic value creation has started a flywheel of compounding infrastructure buildout to run these increasingly-powerful AI systems.” The release of these statistics appears to be part of OpenAI’s broader attempt to improve transparency about its AI infrastructure. As public and regulatory interest in AI’s environmental impact grows, such disclosures are likely to play a larger role in shaping both company policy and public perception.","excerpt":"However, Altman’s estimate is nearly one-tenth of what analysts projected.","categories":["AI News"],"tags":["ChatGPT","OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2025-06-11T09:55:53","publication_year":"2025","word_count":390,"keywords":["Go","ChatGPT","API","OpenAI","AI","RAG","GPT","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","RAG","R","Go","API","GPT","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/each-chatgpt-query-uses-merely-a-15th-of-a-teaspoon-of-water-says-sam-altman\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":56263,"title":"Data Science Is Not About Data But State Thinking: Dr Shailesh Kumar, Chief Data Scientist At Jio ","content":"“For me, data science is not about data, it’s about state thinking where sensor data has to be converted to a number of state vectors,” says chief data scientist at Jio. He also says a successful artificial intelligence system is an interconnected ecosystem of AI models. Dr. Shailesh Kumar is presently the Chief Data Scientist at the Centre of Excellence in AI\/ML, Reliance Jio. As one of the top data scientists and renowned AI experts in India, Dr Shailesh has also worked as a distinguished data scientist at Ola cabs, as a researcher in the Google Brain and Yahoo! Labs, principal development at Microsoft, and also co-founded his edtech company Third Leap. With more than 20 years, Dr Kumar has published more than 20 international research papers and holds more than 20 patents in artificial intelligence and machine learning. At Jio, he says he is reimagining education, agriculture, healthcare, mobility, energy, environment, safety and security in the age of digital connectivity and AI. During the Hyderabad round of Machine Learning Developers Summit by Analytics India Magazine, Dr Shailesh Kumar, Chief Scientist at Jio gave a talk on the rise of ecosystems of models for AI. Dr Kumar’s presentation focused on the various ways in which the AI industry is changing and what data scientists should focus on to prepare for it. “It has never happened before that so many different kinds of technologies have come together not because they needed each other but independently developed. Devices, AI\/ML, cloud, IoT, blockchain. Now, the question is, how do we think about taking advantage of this ecosystem of techs which we have. The next generation of engineers, data scientists, product managers will have to create a new kind of ecosystem thinking and not traditional thinking,” Dr. Kumar said. Chief Data Scientist At Jio Says Successful AI Should Be An Ecosystem Of Models (Not Collection Of Models) According to Dr. Kumar, successful AI today is not a collection of models but an ecosystem of models. He envisions a common framework that can be used for AI independent of all the varying industries, big data sets, verticals and objectives. “After 20 years of thinking about it, it turns out to be a simple framework. For me all intelligent systems are a stimulus, response and learning systems,” he said. Indeed in a world of sensors and data, there is a lot of stimulus data. For example, IoT sensors in a plant is a stimulus, credit card data is stimulus, social media stats, etc. all of which companies can utilise to bring positive response or action. So, what to do with all of the stimulus data? “For me, data science is not about data, it’s about state thinking. So, if you are in retail, you need to know what is the state of the customer is, or state of the store in terms of inventory level or cost of inventory and so on. Sensor data has to be converted to a small number of state vectors. You need to know what the entities are in a business and how you characterise the high level notion of state of those things,” Dr. Shailesh Kumar, Chief Data Scientist at Jio said. According to Dr. Kumar, once data scientists know the state of entities, they can then convert a state into action, which is followed by the final response to the stimulus data. Dr. Kumar tells that all of this can be coded into an IT system also, where you can write a bunch of rules and have the stimulus data converted into state information and then into action. But, it does not end there. “This is not yet an intelligent system, and what makes it intelligent is the feedback loop. This is one of the most critical innovations in the last 20 years which is collecting lots of feedback data to improve the system of analysis. This continuous loop is what AI is. All the ML algorithms that we can think about are either stimulus to state algorithms (clustering, demand, price forecasting) or state to action algorithms. Now, when you look at AI from that perspective, it’s easier to understand which courses to take and modelling techniques to learn,” said Dr. Kumar, Chief Data Scientist at Jio. .","excerpt":"“For me, data science is not about data, it’s about state thinking where sensor data has to be converted to a number of state vectors,” says chief data scientist at Jio. He also says a successful artificial intelligence system is an interconnected ecosystem of AI models. Dr. Shailesh Kumar is presently the Chief Data Scientist […]","categories":["AI Features"],"tags":["Interviews and Discussions","Jio AI Cloud"],"author_name":"Vishal Chawla","publish_date":"2020-02-10T10:41:14","publication_year":"2020","word_count":709,"keywords":["big data","data science","Go","artificial intelligence","machine learning","AI","ML","Git","Jio AI Cloud","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","R","Go","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/shailesh-kumar-jio\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136538,"title":"Mumbai-based Qure.ai Raises $65M to Accelerate AI-Driven Medical Diagnosis","content":"Indian startup Qure.ai has secured $65 million in funding from investors, led by Lightspeed Venture Partners and 360 ONE Asset Management. The fresh capital will be used to expand and further develop its AI-powered solutions, aimed at enhancing disease detection and improving healthcare outcomes. The startup was founded in 2016 by Prashant Warier. Qure.ai is leveraging AI to help with medical diagnosis and help in detecting health conditions including brain trauma and chest diseases such as lung cancer. Qure.ai’s automated medical imaging tools accelerate diagnosis times, allowing physicians to triage cases more efficiently, particularly in urgent situations. This enables healthcare providers to identify critical conditions within minutes instead of hours, helping prevent fatalities and enhancing patient care. In environments where skilled specialists are scarce, Qure’s technology serves as a primary screening tool for various infections and non-communicable diseases. Qure.ai’s chest CT AI tool, qCT LN Quant, recently received FDA clearance, marking a significant advancement in lung cancer care. The AI tool enables precise tracking of lung nodules on chest CT scans over time, complementing other solutions like qXR-LN for chest X-ray nodule detection and qTrack for follow-up management. Together, these tools enhance early detection and improve patient outcomes. Qure.ai has a total of 16 FDA clearances till date. A couple of years ago, the startup raised $40 million in a funding round led by Novo Holdings and HealthQuad, supported by existing investor MassMutual Ventures. AI in Medical Diagnosis A number of tech companies are leveraging AI for medical diagnosis. A startup named RapidAI works on neurovascular and vascular AI-enhanced clinical decision support and patient workflow. It received an FDA approval for its ICH (intracranial haemorrhage) model, making it the first in the world to get a specificity of 100%. A few months ago, Google’s Med-Gemini model achieved 91.1% accuracy in medical diagnostics. Similarly, Microsoft has also been leading initiatives to bridge the healthcare gap in rural India with products such as HoloLens.","excerpt":"The startup has over 1 billion training datasets and is serving more than 90 countries.","categories":["AI News"],"tags":["AI","Google","health","Microsoft","qure.ai"],"author_name":"Vandana Nair","publish_date":"2024-09-24T17:26:43","publication_year":"2024","word_count":322,"keywords":["Go","API","funding","health","AI","RAG","Ray","Aim","llm_models:Gemini","qure.ai","Google","R","Microsoft","startup"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","API","startup","funding","llm_models:Gemini"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mumbai-based-qure-ai-raises-65m-to-accelerate-ai-driven-medical-diagnosis\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":18287,"title":"With AI powered Cloud Platform And Apps, Oracle Is Boosting Its Rich Portfolio of Intelligent Apps","content":"Hillsboro, United States- August 29, 2009: This image shows a red airplane advertising Oracle in mid air, performing aerobatic maneuvers at the Oregon International airshow. It wasn’t long back when Oracle unveiled its AI platform Cloud Service and AI-powered apps at Oracle’s OpenWorld, which promises to make it easier to integrate machine learning solutions into enterprises. The AI platform Cloud Service comes as a blessing for companies as they can tap into deep learning capabilities to better understand enterprise data and transform corporate business processes and user experience. “AI has the power to be more transformative for the enterprise than any other technology in recent history,” said Amit Zavery, senior vice president, product development, Oracle Cloud Platform. What would Oracle’s AI-powered cloud change? With Oracle AI platform Cloud Service that runs on world-class Oracle Cloud Infrastructure, developers and data scientists can rapidly set up a secure and scalable environment for new deep learning models in the cloud, making it easier for researchers to perform AI based work. “Only Oracle can deliver a pervasive approach to AI, embedding AI into its applications, providing an AI development environment coupled with a world-class infrastructure to run specialised workloads”, said the company on its blogpost. The tech giant believes that the entire process of setting up a complete environment for work on advanced machine learning can be quite daunting, and so with the launch of AI-powered cloud, Oracle intends to make it quick and easy for developers and data scientists to set up a secure and scalable environment for new deep learning models in the cloud. The company stated that Oracle AI Platform Cloud instances would come preinstalled with familiar AI libraries, tools and deep learning frameworks. Some of the deep learning frameworks available are Caffe, Jupyter Notebook, Keras, scikit-learn, TensorFlow, NymPy, amongst others. “Oracle is in a unique position to deliver AI across all layers of the cloud, empowering customers to uncover and unlock critical business patterns in their enterprise data to transform organizational productivity, efficiency, and insight”, said Zavery. The company explained that once the machine-learning models are trained, developers can access Oracle’s rich set of PaaS services to include them in AI-powered applications. Features of Oracle’s AI platform Cloud Service: The new AI platform promises an extreme performance which is evident from the differentiated infrastructure layer including NVMe flash storage that it boasts along with best-in-class GPUs on a 25 Gigabit network. It includes a GPU Bare-Metal Shape: 2x Tesla P100 GPUs based on Nvidia’s Pascal Architecture, and will soon support the new Volta GPUs with up to 8 GPUs. The company claims its performance to be 2.4X better than the closest AWS equivalent. The AI infused apps by Oracle: Apart from the AI platform Cloud Service, Oracle announced new AI-based apps for various sectors such as finance, human resources, manufacturing, customer service, marketing, sales and other portfolios. The all-new Oracle Adaptive Intelligent apps are powered by insights from the Oracle Data Cloud and delivers industry’s most powerful AI based modern business applications. “The new Adaptive Intelligent Apps enable business users from across the organizations to quickly and easily take advantage of the latest advancements in artificial intelligence,” said Steve Miranda, executive vice president of applications development, Oracle. The company has made this possible by eliminating the need for more integrations and embedded AI capabilities across Oracle Cloud Applications. “The new AI capabilities combine first- and third-party data with advanced machine learning and sophisticated decision science to deliver the industry’s most powerful AI-based modern business applications,” he added. The new Adaptive Intelligent Apps have the ability to learn, react and adapt in real time. It does so by applying advanced data science and machine learning to Oracle’s web-scale data and organization’s own data. For instance, the AI apps intend to transform various sectors in the following ways: In finance and procurement professionals, new AI-powered capabilities can help streamline financial processes by analyzing historical data. It can also reduce costs, modernize operations, and enhance collaboration. In HR, it can help improve talent management, by applying natural language search and deep learning technologies to resumes and job postings. AI in the supply chain will make use of predictive analytics machine learning techniques to detect and analyse key signals in device data, and then to act on these insights. On a concluding note: There is no denying that there is a substantial opportunity for artificial intelligence and AI enabled applications, which Oracle is venturing on quite efficiently. The company’s belief in the technology is quite evident with the fast pacing developments that it is announcing in the space. It has already been using AI in its SaaS, PaaS and IoT services to improve user experience and productivity, and with AI-powered apps and cloud platform, it has further made a mark in its extravagant AI portfolio.","excerpt":"It wasn’t long back when Oracle unveiled its AI platform Cloud Service and AI-powered apps at Oracle’s OpenWorld, which promises to make it easier to integrate machine learning solutions into enterprises. The AI platform Cloud Service comes as a blessing for companies as they can tap into deep learning capabilities to better understand enterprise […]","categories":["IT Services"],"tags":["Cloud Platform","Oracle"],"author_name":"Srishti Deoras","publish_date":"2017-10-12T05:21:36","publication_year":"2017","word_count":801,"keywords":["data science","machine learning","artificial intelligence","Keras","AI","ML","Oracle","Aim","deep learning","analytics","TensorFlow","Cloud Platform"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","Aim","TensorFlow","Keras"],"url":"https:\/\/analyticsindiamag.com\/it-services\/ai-powered-cloud-platform-apps-oracle-boosting-rich-portfolio-intelligent-apps\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":19727,"title":"Great Lakes PGP In Data Science &#038; Engineering Is The Right Course For The New Era","content":"In the rapidly changing tech domain, data science and data engineering have become core to every business and enterprise. The need for data science jobs has been emphasized in leading surveys across the world with Data Scientist pegged as the “Sexiest Job of the 21st century”. However, despite the high interest, one of the major challenges is how to best represent this multi-disciplinary field that combines elements of statistics, computer science, mathematics, information systems and operations management expertise. The demand for analytics professionals outstrips all other sectors and despite talks about slowdown plaguing the IT market, there are an estimated 50,000 positions related to analytics currently available to be filled in India, states our annual Analytics & Data Science India Jobs Study 2017.  Almost 50 percent of jobs in analytics are for those between 2-7 years of experience across all cities. For freshers however, Chennai stands out in terms of total number of openings. With the growing demand for DSA talent, people looking to transition to data science are expected to benefit the most from an estimated 17% openings for freshers. 39% of analytics and data science job openings are for professionals with around 5 years job experience, the report states. Companies which are aggressively hiring DSA talent in India and have the most number of analytics opening this year are – Amazon, Citi, HCL, Goldman Sachs, IBM, JPMorgan Chase, Accenture, KPMG, E&Y & Capgemini. But as demand grows for DSA jobs that are proving to be the hardest-to-fill roles, there are also concerns to close the talent gap for analytics-related jobs with the right training program that gives in-demand analytics skills to professionals who wish to transition to the booming field. Keeping this in mind, Great Lakes Institute of Management has launched a PGP in Data Science and Engineering program — a one-of-its-kind program that is exclusively tailored for professionals with less than 4 years of experience. Created exclusively for early career professionals, this program gives the much-needed analytical skill-set with a six-month intensive program, delivered in a blended format. The program is steered by a faculty that has great expertise and decades of experience in computational sciences and statistics, making it an optimal blend of academics and industry relevance. It is helmed by Prof. Mudit Kulshreshtha, a leading industry veteran with two decades of experience in data science. Before joining Great Lakes, Prof Kulshreshtha held top leadership positions at Payback, Deloitte and Angel Broking. According to Hari Krishnan Nair, Co-Founder, Great Learning, Great Lakes’ PGP-DSE program is designed specifically for early career professionals who want to learn and equip themselves with industry relevant capabilities that would help them convert into analytics and data science roles in the industry . “With the increasing adoption of big data and analytics across companies and industries, there is a huge talent gap when it comes to analytics and data science. And this gap is even more accentuated in the 0-4 year experience bracket. According to NASSCOM, there are 50,000+ open positions in analytics and data science in India. Most of these are at 0-6 year experience levels. And most companies are struggling to find quality talent, hence this is a great opportunity for young professionals to up-skill themselves and make themselves relevant for the exciting and rewarding careers in data science,” he said. Experiential learning is woven into the program and students learn through multiple projects and case studies that give them hands-on experience to deal with real data. Learners will also get mentored by some of the best in the industry, with sessions from leading analytics practitioners from Microsoft, EXL, Deloitte, HSBC and American Express who will share their tools and techniques to add to the data science tool-box. Program Key Features Who Is The Program Intended For: This in-depth program is intended for professionals with less than four-years of relevant experience and a prior background in (undergraduate degree) Mathematics, Engineering, Statistics, Computer Science and Economics is preferred. Early career professionals with 0-4 years of experience from BFSI, IT, Finance functions thinking about moving into big data can hugely benefit from the program that delivers conceptual techniques and the foundations required for data science roles. Duration: PGP-DSE is a 6-month weekend classroom program including classroom interactions with faculty and industry practitioners, hands-on exercises, and online learning. Classroom sessions would be conducted on alternate weekends from 9:30 am to 6:30 pm in Gurgaon and Bangalore. Tools Covered: Data analytics skills are spreading to new corners of the market, and is no longer restricted to IT. And with many new and emerging DSA roles, the program delivers key analytical skills required for existing job roles in the industry today. From R, SAS, MySQL, Python to data visualization skills such as Tableau, the program is extremely focused on training learners on the in-demand tools and techniques required today from a data-savvy professional. Course Outcome: Co-created by industry experts, the exhaustive practice-based course covers the most industry relevant techniques such as Regression, Predictive Modelling, Clustering, Time Series, Forecasting, Classification applied in enterprises today. Designed for early career professionals, the program exposes learners to industry relevant tools and techniques. There is a heavy focus on applied learning which means hands-on exercises and projects. At the end of the six-month program, learners will be awarded with a Dual Certification from Great Lakes and Stuart School of Business, IIT Chicago. Great Credentials: With six top-ranked programs, a distinguished faculty and an industry best record of successful transitions in business and functional roles in the analytics, the institute has a great placement record. Their alumni have been placed in leading companies such as Amazon, Honeywell, Deloitte among other big-ticket names, making Great Learning’s new offering — PGP in Data Science and Engineering the best fit for early-career professionals who want to fast-track their career in data science. Students can also benefit from an amazing 3000+ alumni network and tap into Great Learning’s world class resources to build on their training in data science. Great Learning also pioneered India’s #1 ranked business analytics program, has clocked 1.5 million + hours of delivering analytics learning, so if you are looking to knock down the frontiers of big data, then this is a good place to start. Career Support: Great Learning will also put you on the right career path with resume building workshops and interview preparation workshops that will help you land a fulfilling job. Here’s Why We Believe Great Learning’s PGP in Data Science & Engineering Has An Edge Great Learning’s PGP in Data Science and Engineering is the only course in India designed for early career professionals with 0-4 years of experience. The comprehensive six-month weekend program covers the whole breadth of analytics techniques, from the essentials of data science to the most in-demand big data technologies used by organizations today. Backed by an eminent faculty and a challenging curriculum, the PGP in Data Science & Engineering provides a convenient and flexible option for busy working professionals to transition to data science career and qualify for higher level data science positions. One of the primary facets of Great Learning’s program is that it imparts skills and tools which are industry-relevant and not just trendy. Besides, its robust industry connect and an excellent faculty, the well-placed alumni connect will also help the learners to land entry-level and higher level data science jobs. Outlook While there are a lot of industry and vendor specific certifications and PG programs that will offer quick-fix courses to buff up your skills, an intensive blended experience will go a long way in fully explaining the concepts behind data analysis. Great Learning’s PGP in Data Science and Engineering is an excellent choice for continuing education and worth the investment.","excerpt":"In the rapidly changing tech domain, data science and data engineering have become core to every business and enterprise. The need for data science jobs has been emphasized in leading surveys across the world with Data Scientist pegged as the “Sexiest Job of the 21st century”. However, despite the high interest, one of the major […]","categories":["AI Trends"],"tags":["great lakes analytics"],"author_name":"Richa Bhatia","publish_date":"2017-12-13T11:14:03","publication_year":"2017","word_count":1283,"keywords":["big data","data science","Go","API","AI","great lakes analytics","Python","data engineering","analytics","SQL","R"],"extracted_tech_keywords":["AI","data science","analytics","Python","R","SQL","Go","API","big data","data engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/great-lakes-pgp-data-science-engineering-right-course-new-era\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10020959,"title":"Now DeepMind&#8217;s New AI Agent Outperforms Humans","content":"Recently, a team of researchers from DeepMind, Google Brain and the University of Toronto unveiled a new reinforcement learning agent known as DreamerV2. This reinforcement learning agent learns behaviours purely from the predictions in the compact latent space of a powerful world model. According to the researchers, DreamerV2 is the first agent to achieve human-level performance on the Atari benchmark. DreamerV2, a collaboration between DeepMind, @GoogleAI and the @UofT, is the first RL agent based on a world model to achieve human-level performance on the Atari benchmark. Read more ⬇️ https:\/\/t.co\/lFFuHH2Uk9— Google DeepMind (@GoogleDeepMind) February 19, 2021 From driverless cars to beating Go world champions, reinforcement learning has come a long way. The researchers said, to successfully operate in unknown environments, reinforcement learning agents need to learn about their environments over time and World models are an explicit way to represent an agent’s knowledge about its environment. The Motivation World models have the ability to learn from fewer interactions, enable forward-looking exploration, facilitate generalisation from offline data as well as allow reusing knowledge across multiple tasks. Compared to model-free reinforcement learning that learns through trial and error, world models facilitate generalisation and can predict the outcomes of potential actions to enable planning. However, despite their intriguing properties, world models have so far not been accurate enough to compete with the state-of-the-art model-free algorithms on the most competitive benchmarks. To mitigate such challenges and to achieve human-level performance on reinforcement learning environments, the researchers created DreamerV2. Genesis DreamerV2 is the first-ever reinforcement learning agent based on a world model. The agent achieves human-level performance on the popular Atari benchmark. The agent basically constitutes the second generation of the previous Dreamer agent that learns behaviors purely within the latent space of a world model trained from pixels. Developed by the same team last year, the Dreamer agent is a reinforcement learning agent that solves long-horizon tasks from images purely by latent imagination. More specifically,  Dreamer learns a world model from the past experience and efficiently learns far-sighted behaviours in its latent space by backpropagating value estimates back through imagined trajectories. DreamerV2 is the successor of the Dreamer agent. The DreamerV2 agent relies exclusively on general information from the images and accurately predicts future task rewards even when its representations were not influenced by those rewards. The Tech Behind This new agent works by learning a world model and uses it to train actor-critic behaviors purely from predicted trajectories. It is built upon the Recurrent State-Space Model (RSSM) — a latent dynamics model with both deterministic and stochastic components — allowing to predict a variety of possible futures as needed for robust planning, while remembering information over many time steps. The RSSM uses a Gated Recurrent Unit (GRU) to compute the deterministic recurrent states. DreamerV2 introduced two new techniques to RSSM. According to the researchers, these two techniques lead to a substantially more accurate world model for learning successful policies: The first technique is to represent each image with multiple categorical variables instead of the Gaussian variables used by world models. The second new technique is KL balancing. This technique lets the predictions move faster toward the representations than vice versa. Wrapping Up The above image shows how DreamerV2 outperformed previous world models. The researchers showed how to learn a powerful world model to achieve human-level performance on the competitive Atari benchmark. DreamerV2 is the first world model that enables learning successful behaviors with human-level performance on the well-established and competitive Atari benchmark. Besides this, DreamerV2 out-performed top model-free algorithms with the same compute and sample budget using just a single GPU. Read the paper here.","excerpt":"Recently, a team of researchers from DeepMind, Google Brain and the University of Toronto unveiled a new reinforcement learning agent known as DreamerV2. This reinforcement learning agent learns behaviours purely from the predictions in the compact latent space of a powerful world model. According to the researchers, DreamerV2 is the first agent to achieve human-level […]","categories":["AI Features"],"tags":["DeepMind","DeepMind AI","Google Deepmind","Google Research","Reinforcement Learning","reinforcement learning algorithms","Reinforcement Learning Systems"],"author_name":"Ambika Choudhury","publish_date":"2021-02-28T16:00:00","publication_year":"2021","word_count":603,"keywords":["Reinforcement Learning Systems","Go","reinforcement learning algorithms","Reinforcement Learning","programming_languages:R","AI","programming_languages:Go","GRU","Google Deepmind","DeepMind","DeepMind AI","R","Google Research"],"extracted_tech_keywords":["AI","R","Go","GRU","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/now-deepminds-new-ai-agent-outperforms-humans\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":22217,"title":"Airtel Partners With Nokia To Use Their Hybrid Self-Organising Networks","content":"Bharti Airtel, India’s largest mobile network operator has announced that it is partnering with Nokia to boost its service quality and operational efficiency through the employment of Nokia’s hybrid self-organising networks (SON) solution. “Our collaboration with Nokia in taking SON to the next level by integrating it with other data sources and machine learning capabilities. It would enable us to proactively address network performance dips and outages even before a customer experiences them,” Abhay Savargaonkar, director, Networks (India and South Asia) at Bharti Airtel said in a statement. The solution enables an operator to make proactive decisions for optimisation of network utilisation and better user experience. It employs predictive machine learning which interworks with analytics platform to achieve the purpose. The integration of Airtel’s multiple vendor radio network structure with Nokia’s EdenNet SON software will enhance capabilities through spontaneous identification and resolution of network problems, and real-time management of network capacity. Sanjay Malik, head of India Market at Nokia commented on the partnership, “We are confident that our proven expertise will enable Airtel to manage complex multi-technology networks by automatically minimising the risk of human error.” 2018 has seen Bharti Airtel announcing a series of strategic alliances. Earlier this month, the company had entered into a collaboration with Hotstar to stream its shows on the Airtel TV app. This will make over 1,00,000 hours of content across nine languages, consisting of Indian cinema, live sports and TV shows across Star Network freely available to users of the Airtel TV App. Later, it announced a partnership with HMD Global to provide affordable 4G smartphones as a part of its ‘Mera Pehla Smartphone’ initiative. Earlier this week Bharti Airtel had announced that it had entered into a venture called the Seamless Alliance with companies across sectors to provide seamless low-latency and high-speed connectivity inside aircrafts. The alliance has companies from three sections – aircraft manufacturers (US-based airlines Delta and Airbus), in-flight broadband providers (GoGo) and  telecom operators (US-based Sprint). The alliance also includes Softbank-backed satellite start-up, OneWeb.","excerpt":"Bharti Airtel, India’s largest mobile network operator has announced that it is partnering with Nokia to boost its service quality and operational efficiency through the employment of Nokia’s hybrid self-organising networks (SON) solution. “Our collaboration with Nokia in taking SON to the next level by integrating it with other data sources and machine learning capabilities. […]","categories":["AI News"],"tags":["bharti airtel","Machine Learning"],"author_name":"Jeevan Biswas","publish_date":"2018-03-01T12:50:35","publication_year":"2018","word_count":335,"keywords":["Go","machine learning","ELT","AI","bharti airtel","ML","Machine Learning","ViT","analytics","GAN","R","analytics platform"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","R","Go","ELT","GAN","ViT","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/airtel-partners-nokia-hybrid-self-organising-networks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10090888,"title":"AI Won’t Go Rogue, And Here’s Why","content":"As AI systems get more powerful, the idea of existential threats from sentient algorithms seems to become more real. This emotion reverberates strongly in the recent open letter calling for a six-month pause on AI research for systems more powerful than GPT-4. If the researchers are freaking out about rogue AI and AGI risk, it means the layman must be shaking in his boots, right? Well, it seems that the risk associated with releasing AI systems is far less than it is perceived to be. Even if prominent research houses publish stories that over 300 million jobs will be replaced by AI systems, the truth, as always, is much more nuanced. However, the key takeaway is clear — AI won’t go rogue. It’s just a software Recently, an article proposed the idea that the game ‘Elden Ring’ would soon become sentient and an existential risk to humanity. While this proposal seems outlandish at first, it is just a metaphor to illustrate an equally impossible scenario — that of modern-day AI taking on the role of Skynet. Currently, what we call AI is just another piece of software, with the field associating the name ‘artificial intelligence’ because it is the closest description of what is being created. Modern AI is nothing but a precursor to an actually intelligent program that can pose an existential threat to humanity. Take the show Westworld for instance. Set in a near-future scenario, it explores the possibility of a theme park operated by human-like robots. When these robots gain sentience, the operator of the park raises an interesting question — what is consciousness? In a conversation between a programmer who knows intimately how his machines work and a machine that knows its own true nature, he states, “There is no threshold that makes us greater than the sum of our parts, no inflection point at which we become fully alive. We can’t define consciousness because consciousness does not exist.” This brings up an important point to be answered by AI researchers today — how can we be scared of a generally intelligent program, also known as AGI, when we haven’t even defined what consciousness is yet? According to this paper published on the nature of consciousness in AGI, there are still four definitional types of consciousness that need to be solved before we reach AGI. In this matter, our current algorithms are just simple instruction-following programs, whereas AGI is a completely new technology going beyond algorithms. While GPT-4 and Midjourney are leaps ahead of ELIZA or GANs respectively, they are still non-intelligent programs. The problem with the field of AI isn’t the technology they’re researching, it’s the branding. AI has a branding problem The real problem in AI is not AGI risk, but the fear associated with artificial intelligence. For decades, depictions of thinking machines in popular media have served to strike fear in the hearts of the audience. However, the path from the current state of AI technology to such capable machines is not only unclear, but necessitates a paradigm shift in computing itself. Computers today cannot even produce random numbers due to the binary nature of their foundations. Steve Ward, professor of computer science and engineering at MIT, stated, “On a completely deterministic machine you can’t generate anything you could really call a random sequence of numbers, because the machine is following the same algorithm to generate them.” If our deterministic computers cannot even produce random numbers, how can they have sentience, consciousness, or even intelligence, which are purely non-deterministic concepts? Let’s take a look at GPT-4, the most-advanced LLM and the subject of the aforementioned open letter. LLMs are just word generators, programs trained to predict the next likely word and fine-tuned to speak like humans. “There are a lot of people freaking out about the way large language models are doing things like writing college essays, etc. The harm is, these things are just bullshit generators” Alex Hanna, Former AI ethicist at Google There’s nothing truly intelligent about GPT-4, except if you count its prediction capabilities — an echo of what is possible with real, human intelligence. While the disruptive potential of AI cannot be dismissed outrightly, it seems that social systems have already adapted to technology replacing humans. To create real thinking machines, AI research has to go beyond what is possible today. The real path to AGI Current algorithms have gotten to this level mainly by aping human biology. Neural networks, the basis for GPT-4, are a faint echo of the structure of the human neuron. Reinforcement learning tries to apply a small facet of humans’ knowledge acquisition methods to computer programs. However, creating an artificial brain, or rather the recreation of the ‘thinking’ parts of the human brain, is a task best left to future researchers and scientists. This undertaking would not only necessitate advances in the understanding of the human brain, neuroscience, and psychology, but also breakthroughs in philosophy. To create a truly artificial intelligence, humanity needs to reach a consensus on what intelligence and sentience actually mean. Until then, researchers can figure out how to solve information hallucinations or how to make generative AI draw hands effectively.","excerpt":"To create a truly artificial intelligence, humanity needs to reach a consensus on what intelligence and sentience actually mean","categories":["AI Trends"],"tags":["AI Tool","Yoshua Bengio"],"author_name":"Anirudh VK","publish_date":"2023-04-06T11:00:00","publication_year":"2023","word_count":860,"keywords":["Go","API","artificial intelligence","AI","neural network","GPT","Yoshua Bengio","generative AI","AI research","GAN","AI Tool","R"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","generative AI","R","Go","API","GPT","GAN","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ai-wont-go-rogue-and-heres-why\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10064384,"title":"IBM unveils industry’s first quantum-safe system, IBM z16","content":"Big Blue has unveiled IBM z1, its next-generation system with an integrated on-chip AI accelerator—delivering latency-optimised inferencing to enable clients to analyse real-time transactions, at scale for mission-critical workloads such as credit card, healthcare and financial transactions. Building on IBM’s history of security leadership, IBM z16 is specifically designed to help protect against near-future threats that might be used to crack today’s encryption technologies. IBM z16 uniquely brings together AI inferencing via its IBM Telum Processor with the highly secured and reliable high-volume transaction processing IBM is known for. For the first time, banks can analyse fraud during transactions on a massive scale: IBM z16 can process 300 billion inference requests per day with just one millisecond of latency. For consumers, this could mean reducing the time and energy required to handle fraudulent transactions on their credit cards. For both merchants and card issuers, this could mean a reduction in revenue loss as consumers could avoid the frustration associated with false declines where they might turn to other cards for future transactions.IBM z16 clients can strengthen their cyber resiliency posture and retain control of their system. Also, the Crypto Express 8S (CEX8S) hardware security module will offer clients both classical and quantum-safe cryptographic technology to help address their use cases requiring information confidentiality, integrity and non-repudiation. IBM z16’s secure boot and quantum-safe cryptography can help clients address future quantum-computing related threats including harvest now, decrypt later attacks which can lead to extortion, loss of intellectual property and disclosure of other sensitive data.","excerpt":"IBM z16 can process 300 billion inference requests per day with just one millisecond of latency.","categories":["AI News"],"tags":["IBM"],"author_name":"Kartik Wali","publish_date":"2022-04-05T18:30:50","publication_year":"2022","word_count":252,"keywords":["programming_languages:R","AI","Together AI","IBM","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","R","Rust","Together AI","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-unveils-industrys-first-quantum-safe-system-ibm-z16\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119709,"title":"Why NVIDIA GPUs are Still Not Available in India","content":"Recently, NVIDIA chief Jensen Huang personally delivered the first NVIDIA DGX H200 to OpenAI, a gesture similar to the one made in 2016. However, Indian AI companies aren’t receiving the same treatment from the GPU king. For instance, when Yotta received the first shipment of 4,000 GPUs with much fanfare, the boxes bearing the ASUS logo caught everyone’s attention, but there was barely any NVIDIA soul around. And the energy, alas, was nothing like that in Silicon Valley or the West. ASUS helped Yotta procure NVIDIA GPUs by providing the ESC N8-E11 server, which is equipped with eight NVIDIA HGX H100 GPUs to enhance Yotta’s Shakti Cloud platform for AI model development and deployment. NVIDIA’s Love for Yotta is Elite Yotta plans to scale up its GPU inventory to 32,768 units by the end of 2025. Last year, the company announced that it would import 24,000 GPUs, including NVIDIA H100s and L40S, in a phased manner. However, acquiring the highly valued NVIDIA GPUs is no mean feat. NVIDIA sells its GPUs through the NVIDIA Partner Program, which includes Registered, Preferred, and Elite categories. According to NVIDIA’s blog post, Elite partners represent the highest level of partnership and the tag is reserved for those demonstrating exceptional commitment. In a recent interview with Forbes, Yotta chief Sunil Gupta, aka the ‘Data Centre Man of India’, said that the company was part of the NVIDIA Partner Network. “NVIDIA has put its entire weight behind us. We are an Elite partner of the NVIDIA Partner Network. NVIDIA is giving very high-priority allocations to us,” he said. Gupta added that India could build five GPT-4 models simultaneously using its existing infrastructure. “I have ordered 16,000 [GPUs], so if there are five customers each wanting to make a GPT-4, I can handle their load simultaneously,” said Gupta. Interestingly, Yotta is India’s sole NVIDIA Partner Network Cloud Partner (NCP) and has ascended to the Elite Partner status on the global NCP list. The company will also receive NVIDIA’s latest GPU Blackwell by October. Other Elite members of NCP include AWS, Microsoft Azure, and Meta. What about the others? “If you talk about high-end NVIDIA GPUs like the H100, they’re actually not available in India,” said Vivek Raghavan, founder of Sarvam AI, adding that the situation is expected to change soon. Similarly, Vishnu Vardhan, SML chief and creator of Hanooman, told AIM that they faced a shortage of NVIDIA GPUs and had to search across multiple places to purchase them when they began training Hanooman. “NVIDIA GPUs are considered the best in the market due to the extensive software libraries built to support them. This makes it possible for even a fresh out-of-school kid to work with NVIDIA GPUs,” said Vardhan, adding that he currently possesses more than 1,000 GPUs. Despite the near-absence of NVIDIA GPUs, AI startups in India have remained resilient. India’s AI unicorn Ola Krutrim, for instance, is actively engaged in pre-training Krutrim’s foundational models using the Intel Gaudi 2 cluster. Much like Yotta’s Shakti, Ola Krutrim has also introduced the Krutrim AI Cloud, offering developers access to a variety of open-source models. However, there is no clarity on whether Bhavish Aggarwal-led startup is using NVIDIA’s GPUs or not. Zoho is also exploring NVIDIA alternatives. ManageEngine, the enterprise IT management division of Zoho Corporation, recently invested nearly $10 million in procuring GPUs from all three major providers—Intel, AMD, and NVIDIA. The Indian Union Cabinet recently approved an INR 10,371.92 crore AI program, which includes deploying 10,000 GPUs through public-private partnerships. The government plans to adopt a rent-and-sublet model to provide these GPUs to AI startups in India. Last year, NVIDIA promised that India would receive tens of thousands of GPUs and partnered with Reliance, Tata, and the government. The government plans to establish a cluster of 25,000 GPUs for startups. NVIDIA GPUs are anticipated to enter the Indian market post the Lok Sabha elections.","excerpt":"NVIDIA GPUs are anticipated to enter the Indian market post the Lok Sabha elections.","categories":["Global Tech"],"tags":["NVIDIA","Yotta"],"author_name":"Siddharth Jindal","publish_date":"2024-05-06T18:00:00","publication_year":"2024","word_count":649,"keywords":["Go","OpenAI","AI","Yotta","AWS","ML","R","RAG","GPT","Aim","NVIDIA","Azure"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","RAG","AWS","Azure","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-nvidia-gpus-are-still-not-available-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10051097,"title":"Google AI Introduces FLAN, A Language Model with Instruction Fine-Tuning","content":"Google AI recently introduced their new Natural Language Processing (NLP) model, known as Fine-tuned LAnguage Net (FLAN), which explores a simple technique called instruction fine-tuning, or instruction tuning for short. In general, fine-tuning requires a large number of training examples, along with stored model weights for each downstream task which is not always practical, particularly for large models. FLAN’s instruction fine-tuning technique involves fine-tuning a model not to solve a specific task, but to also make it more amenable to solving NLP tasks in particular. FLAN is fine-tuned on a large set of varied instructions that use a simple and intuitive description of the task, such as “Classify this movie review as positive or negative,” or “Translate this sentence to Danish.” Creating a dataset of instructions from scratch to fine-tune the model would take a considerable amount of resources. Instead, it makes use of templates to transform existing datasets into an instructional format. Image Source: Google AI FLAN demonstrates that by training a model on a set of instructions, it not only becomes good at solving the kinds of instructions it has seen during training but becomes good at following instructions in general. Image Source: Google AI Google AI used established benchmark datasets to compare the performance of FLAN with existing models. It was also evaluated how FLAN performs without having seen any examples from that dataset during training. Evaluation results showed that FLAN on 25 tasks improves over zero-shot prompting on all but four of them. The results were found to be better than zero-shot GPT-3 on 20 of 25 tasks and better than even few-shot GPT-3 on some tasks. Image Source: Google AI It was also found that at smaller scales, the FLAN technique actually degrades performance, and only at larger scales does the model become able to generalize from instructions in the training data to unseen tasks. This might be because models that are too small do not have enough parameters to perform a large number of tasks. Google AI hopes that the method presented will help inspire more research into models that can perform unseen tasks and learn from very little data.","excerpt":"Google AI hopes that the method presented will help inspire more research into models that can perform unseen tasks and learn from very little data.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Google","Machine Learning","NLP","nlp in data","NLP models","nlp pipeline"],"author_name":"Victor Dey","publish_date":"2021-10-08T15:54:44","publication_year":"2021","word_count":356,"keywords":["Go","programming_languages:R","AI","R","Machine Learning","programming_languages:Go","NLP models","NLP","GPT","ai_applications:NLP","Google","Deep Learning","Data Science","Data Scientist","nlp pipeline","nlp in data","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","NLP","R","Go","GPT","llm_models:GPT","programming_languages:R","programming_languages:Go","ai_applications:NLP"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-ai-introduces-flan-a-language-model-with-instruction-fine-tuning\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10092984,"title":"Council Post: Beyond ChatGPT &#8211; Exploring opportunities in the Generative AI value chain","content":"A whole ecosystem, including hardware vendors and application developers, is emerging thanks to generative AI, which will help realise its commercial potential. In the second half of 2022 and the beginning of 2023, tech pioneers unveiled generative AI solutions, astounding investors, business executives, and the general public with the technology’s capacity to generate wholly original and presumably human-made prose and images. And well the response has been unprecedented. One million people flocked to ChatGPT, a generative AI language model from OpenAI that produces original material in response to user requests, in just five days. For iPhone, Apple needed more than two months to achieve the same level of adoption. In comparison to Netflix, Facebook had to wait more than ten months to reach the same user base. My role as a leader of Data & Analytics has given me a front-row seat to the benefits and risks of generative AI, and I have been spending time in evaluating several generative AI start-ups over the past few months. In my opinion, generative AI is not the omen of doom that detractors claim it to be, even though it is an effective tool that needs human oversight to use responsibly. Over the past three years, venture capital firms have spent over $1.7 billion in generative AI solutions, with the largest money going towards AI-enabled medication research and AI software writing. Additionally, ChatGPT is not the only company using generative AI. Within 90 days of its debut, Stability AI’s Stable Diffusion, which can create visuals based on text descriptions, received more than 30,000 stars on GitHub—eight times more quickly than any other package. A brief explanation of generative AI A branch of machine learning known as “generative AI” use algorithms to create new data, such as images, texts, or sounds. It resembles a virtual author or artist producing creative writing and art. However, it’s merely a group of brilliant algorithms at work behind the scenes, not a real artist or writer. Transformer-based models and GANs (Generative Adversarial Networks) are the two most used generative AI models at the moment. GANs are excellent at converting text and images into visual and multimedia content. Transformer-based models, such as GPT (Generative Pre-Trained) language models, can take in data from the Internet and produce a variety of content, including press releases, whitepapers, and articles for websites. Why should you care about generative AI? Well, there are a lot of explanations. Top three are as follows: It can create entirely new data that doesn’t already exist. Think about the countless opportunities for investigation and experimentation! By producing training data for fresh neural networks or developing top-notch deep learning architectures, it can enhance already-existing algorithms. Basically, it’s a machine that creates better machines. But that’s not all. Gartner has declared generative AI as one of the most disruptive and rapidly evolving technologies in their 2022 Emerging Technologies and Trends Impact Radar report. And get this – they’ve made some pretty bold predictions about its future impact. By 2025, generative AI is expected to generate 10% of all data (currently, less than 1%) and 20% of all test data for consumer-facing applications. Plus, it’ll be used in 50% of drug discovery and development projects by 2025. And by 2027, a whopping 30% of manufacturers will be using it to improve their product development process. Generative AI is making waves. So, pretty important stuff, right? Generative AI Industry-specific use applications Business models that use generative AI can not only help businesses automate routine work but also increase income. One of the most practical uses of generative AI is content production. Education By leveraging generative AI, personalized lesson plans can provide students with the most effective and tailored education possible. These plans are crafted by analyzing student data such as their past performance, skillset, and any feedback they may have given regarding curriculum content. This helps ensure that each student, especially those with disabilities, is receiving an individualized experience designed to maximize success. Logistics and transportation The investigation of historically unexplored areas is made possible by generative AI’s precise conversion of satellite photos into map views. For logistics and transportation businesses wishing to explore new areas, this might be extremely helpful. Travel industry Systems for face detection and verification at airports can benefit from generative AI. The technology can make it simpler to identify and confirm the identity of travellers by assembling a full-face image of a passenger from images taken from various angles. Banking Another area where generative AI is proven to be a useful tool is fraud detection. With the use of their past data, banks are teaching ML and AI algorithms to suggest risk criteria. They can train the system to either ban or allow specific user behaviours depending on the likelihood of fraud by exposing generative AI to prior incidents of fraud and non-fraud. This makes fraud detection quicker and more effective than it would be with just humans. It’s crucial to remember that the labour being automated is typically low-level, tedious, and repetitive, whether it’s fraud detection or product creation. To guarantee the end product’s quality and safety, human intervention is still required. The true benefit of generative AI is as a multiplier for general-purpose productivity and efficiency. The Future Of Generative AI The capabilities of generative AI are unknown to us. More than 30% of new medications and materials are predicted to be discovered by generative AI technology by 2025, which would result in significant cost savings for the healthcare sector. With the ability to foresee future market trends and investment possibilities to lower risk, generative AI is poised to be a potent tool in financial forecasting and scenario building. It has the potential to have a significant impact on the entertainment sector as well, giving businesses the ability to improve visual effects, preserve and colourize movies, and even age or de-age performers’ faces. Generative AI can analyse enormous volumes of data and patterns, but it cannot take the place of human originality, creativity, and common sense. Therefore, human oversight is essential for its creation and implementation. Businesses and decision-makers should approach generative AI with caution in order to solve ethical issues and make sure that the future involves widening the economic pie to benefit humanity. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"Generative AI can analyse enormous volumes of data and patterns, but it cannot take the place of human originality, creativity, and common sense. Therefore, human oversight is essential for its creation and implementation.","categories":["AI Features"],"tags":["Banking","education","Generative AI","Logistics","travel"],"author_name":"Anish Agarwal","publish_date":"2023-05-08T16:35:34","publication_year":"2023","word_count":1084,"keywords":["data science","ChatGPT","machine learning","OpenAI","AI","neural network","education","ML","deep learning","Logistics","Banking","analytics","generative AI","Generative AI","travel"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","data science","analytics","generative AI","ChatGPT","OpenAI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/beyond-chatgpt-exploring-opportunities-in-the-generative-ai-value-chain\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10135550,"title":"AI4Bharat Releases Anudesh, IndicASR, Rasa for Indic Language AI","content":"AI4Bharat has launched a series of innovations aimed at enhancing Indian language technology, including speech recognition, data annotation, and expressive text-to-speech (TTS). Source: AI4Bharat One of the key releases is IndicASR, India’s first speech recognition model covering all 22 official languages. A web demo is available for users to test and provide feedback. Also called ndicConformers, it is a comprehensive set of ASR models designed to accurately convert speech to text in all 22 official Indian languages. Another major release is Anudesh, v0.1, an open-source platform designed to improve LLMs for Indian languages through data annotation. The platform’s first version facilitates conversational data collection through LLM interactions and supports model evaluation workflows. Additionally, AI4Bharat has introduced Rasa, a dataset for expressive TTS that spans nine languages and features 14 speakers. The dataset includes at least 20 hours of speech per speaker, aiming to feature both male and female voices across all 22 official Indian languages. In its initial version, it presents a practical approach to gathering high-quality data for languages with limited resources, focusing on easily accessible neutral speech data, complemented by smaller samples of expressive speech. The organisation has also revamped its website, making datasets, models, and tools more accessible. Users can now find clear download instructions, usage guidelines, and supporting Colab notebooks for easier integration of AI4Bharat’s resources. Last month, Sarvam AI also launched ASR models Shuka v1, which comprises approximately 60 million parameters and is trained on less than 100 hours of audio data.","excerpt":"The organisation has also revamped its website, making datasets, models, and tools more accessible.","categories":["AI News"],"tags":["AI4Bharat"],"author_name":"Mohit Pandey","publish_date":"2024-09-16T16:35:30","publication_year":"2024","word_count":247,"keywords":["programming_languages:R","AI","innovation","Colab","Aim","GAN","AI4Bharat","R"],"extracted_tech_keywords":["AI","Aim","Colab","R","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai4bharat-releases-anudesh-indicasr-rasa-for-indic-language-ai\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":67805,"title":"5 Resources For Kids To Learn Coding","content":"GitHub recently announced its partnership with Hack Club to support students with coding. Committing to a $50K hardware fund, it announced working globally alongside Arduino and Adafruit on delivering the hardware tools directly to students’ homes. Designed for teenagers aged 13-18, the program is free of cost wherein students will have access to hardware on a needs basis and will have guidance from the industry mentors. This clearly tells us about the importance that global companies are laying on students and kids for coding. In this article, we list 5 institutes and initiatives that are working on teaching kids coding. Campk12 A global online school for 21st century skills, Campk12 teaches coding and STEM subjects to kids aged 5-18. With LIVE, interactive and gamified online sessions, it has been teaching kids to code since 2010. Founded in 2010 by an MIT computer grad, Anshul Baghi, it has since taught coding to more than 50,000 students and has partnered with 100’s of schools and institutions to inspire young innovators. It includes a 100% project-based curriculum with coaching sessions from Grade 1 to 12. There are classes on Augmented Reality, 3D Coding, VR game development, AI,  machine learning, creating chatbots and more. CoderBunnyz A unique initiative by a ten-year-old Samaira Mehta, CoderBunnyz is all about teaching coding and AI to young kids while engaging in fun games. Her love for computer programming and board games led her to bring computer coding to a board game. The game teaches kids how to code using artificial intelligence. Called CoderMindz, this is reportedly the first-ever AI board game that engages a kid while helping them learn coding. CodeMonkey Providing a fun and educational game-based environment, CodeMonkey helps kids learn and understand code without any prior knowledge. It provides resources for students of various grades that help them not only learn coding basics but also how to code in real programming. It helps them navigate through the programming world with a sense of confidence and accomplishment. It has various programs such as self-guided challenges, learning real-coding languages, and game-based learning for kids to learn. WhiteHat Jr Edtech startup founded in 2018 by ex-Discovery CEO Karan Bajaj, it teaches kids 6-14 years of age how to code and help them create interesting applications. It conducts sessions to build commercial-ready games, animations and apps online using the fundamentals of coding. It offers four levels of courses – Beginner, Intermediate, Advanced and Professional that helps in creating complex games using logic, sequence, commands and algorithmic thinking. WhiteHat Jr is looking to expand globally and recruiting for various roles in product development, technology, operations and sales. It is backed by Nexus Venture Partners, Omidyar Network India and Owl Ventures and had raised $10 million in series A funding last year. Vedantu Super Coders Programme This program by Vedantu is aimed at providing high quality coding programs for kids of 6-12 years of age. With a belief that coding improves a child’s reasoning and problem solving skills, this program is taken by coding experts in the field of AI and machine learning. There are various courses offered such as Lil Champs for Grade 1 and 2, Young Learners for Grade 3 and 4, Early Achievers for Grade 5 and 6, Pro Coders for Grade 7 and 8. These courses involve a wide range of topics such as creating animations, commercial-ready apps and games, creating advanced games and chatbots using artificial intelligence and Python, IoT and more. It delivers customized learning based on student’s learning pace and has been designed by experts from MIT and IIT, and promotes activity-based learning.","excerpt":"GitHub recently announced its partnership with Hack Club to support students with coding. Committing to a $50K hardware fund, it announced working globally alongside Arduino and Adafruit on delivering the hardware tools directly to students’ homes. Designed for teenagers aged 13-18, the program is free of cost wherein students will have access to hardware on […]","categories":[],"tags":["arduino"],"author_name":"Srishti Deoras","publish_date":"2020-06-20T12:04:25","publication_year":"2020","word_count":598,"keywords":["Go","artificial intelligence","machine learning","AI","chatbots","R","Git","Python","Aim","GitHub","arduino"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","chatbots","Python","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/coding-for-kids-resources\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10097511,"title":"OpenAI To Soon Launch Open Source GPT Models","content":"OpenAI is likely to release weights of its models in the coming months. Amid the Llama fever, OpenAI’s Andrej Karpathy, recently said that all of this is quite generic to just transformer language models. “If\/when OpenAI was to release models as weights (which I can neither confirm nor deny!) then most of the code here would be very relevant. In other words, OpenAI is most likely to make GPT- 3.5 open source according to OpenAI’s Karpathy, a prominent figure in the field of deep learning. It has to be noted that the company has not made any official announcement about this. The conversation stems from a Twitter (now X) thread, when one of the users asked Karpathy as to why has been playing with Llama 2, instead of building Jarvis for OpenAI. Yay, llama2.c can now load and inference the Meta released models! :) E.g. here inferencing the smallest 7B model at ~3 tokens\/s on 96 OMP threads on a cloud Linux box. Still just CPU, fp32, one single .c file of 500 lines: https:\/\/t.co\/CUoF0l07oXexpecting ~300 tok\/s tomorrow :) pic.twitter.com\/bjurODT4dL— Andrej Karpathy (@karpathy) July 25, 2023 This new development comes in the backdrop of the recent release of Baby Llama aka llama.c, where Karpathy has been exploring the concept of running large language models (LLMs) on a single computer as part of his recent experiments, inspired by the release of Meta’s Llama 2. Check out the GitHub repository here. Karpathy said llama2.c can now load and inference the Meta released models. He further gave an example of  inferencing the smallest 7B model at ~3 tokens\/s on 96 OMP threads on a cloud Linux box and is expecting ~300 tok\/s soon. Further, he said that If you can get 7B model to run at nice and interactive rates then you can go from “scratch-trained micromodels” to “LoRA fine tuned 7B base model”, all within the code of the minimal llama2.c repo (both training and inference). Can reach more capability and with less training data. Interestingly, the success of Karpathy’s approach lies in its ability to achieve highly interactive rates, even with reasonably sized models containing a few million parameters and trained on a 15 million parameter model of TinyStories dataset. Hopefully it will bring back the actual OpenAI which was started as an open source non-profit company where Karpathy was one of the initial founding members who played an active role in contributing to the open source community.","excerpt":"OpenAI most likely to make GPT- 3.5 open source according to OpenAI’s Karpathy, a prominent figure in the field of deep learning","categories":["AI News"],"tags":["Open Source AI"],"author_name":"Siddharth Jindal","publish_date":"2023-07-25T14:53:14","publication_year":"2023","word_count":407,"keywords":["Go","OpenAI","AI","RPA","Git","Open Source AI","GPT","deep learning","GitHub","R","llm_models:GPT"],"extracted_tech_keywords":["AI","deep learning","OpenAI","R","Go","Git","GitHub","GPT","RPA","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-to-soon-launch-open-source-gpt-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":45571,"title":"How Sinkholing Can Protect Your Network From Malware","content":"Over the years, cybersecurity has evolved tremendously — for both white hat hackers and threat actors. Botnets have also gained good traction.  Botnets are nothing but compromised computer systems that are managed by third parties. These computers are used to carry out different types of attacks such as deploying malware, stealing data, DDoS attack etc. The major challenge for white hats is to figure out the main culprit behind the attack as the identity of the botnet manager gets difficult to find. This is where Sinkholing comes into the picture. What Is Sinkholing? Sinkholing is a way of manipulating the flow of data from one point to another in a network. Simply put, it is basically a method of preventing some specific traffic to reach the desired server. This is done by simply rerouting or redirecting the traffic in a network from its original to an altered server (called as Sinkhole) which is also owned by the same owner as the main server. Sinkholing is that cybersecurity technique that is nothing less than a double-edged sword. This implies that it can have adversarial uses such as steering legit traffic away from its intended recipient. However, over the years, this method has gained significant traction when it comes to fighting malware as it is mostly used by anti-malware researchers to collect information about a botnet. The alternate server poses as one of the C2 (command-and-control) servers in the botnet. And once the malicious traffic lands to the sinkhole server, they are then analysed by researchers to understand the source of attacks and prevention methods as well. At the enterprise level, this technique of sinkholing is also used to restrict access to any website. For example, if someone is trying to violate corporate policies by accessing web pages that are not allowed in the corporate world, they end up landing to a customised page (this page can be created with information about the corporate policy restriction) and their data gets stored there that the firm can be used to take further action. When it comes to tracking down criminals, government bodies responsible for maintaining safe cyberspace uses this to carry out investigations and criminal infrastructure takedowns. Furthermore, Sinkholing has become so popular that today, even ISPs are using it to defend their networks and customers, and manage traffic flow. Types Of Sinkholing And Challenges Internal Sinkholing: This sinkholing is focused on an organisations network. It is basically used to figure out which all systems are infected and can cause an adversarial effect on the network. Once the machines are identified, organisations take back control of the machines. External Sinkholing: Despite its effectiveness, it is considered to be one of the controversial methods. The main reason behind this is that any machine on the internet can be manipulated, registering known malicious domains (usually the ones which expires). Despite the fact that sinkholing is one of the effective methods to fight against malware, there are some significant challenges. One of the major challenges is with the external sinkholing, and that is the legal issues. For example, if a victim who is not from your organisation is trying to access a sinkholed URL by your company, and if you take control of that victim machine (even if it is just for research purposes), it goes against the protocols in many regions. Many malware that is deployed using sinkholing has the option of self-destruction, but that doesn’t mean you can take control of any machine that lands on the sinkhole. However, there is a solution to this that is also becoming really popular. By using the reverse DNS, many sinkholes nowadays first identify whether the machine is infected or malicious. Outlook Over the past few years, Sinkholing has been used in several malware campaigns — as defender and attacker. However, the defence side is much more effective. Also, there were times when Sinkholes techniques couldn’t succeed but managed to thwart malware from spreading. While many believe Sinkholes are not as significant as other cybersecurity strategies, one cannot deny the fact that they play a major role in network security. After all, who would want to invite infectious traffic to their website?","excerpt":"Over the years, cybersecurity has evolved tremendously — for both white hat hackers and threat actors. Botnets have also gained good traction.  Botnets are nothing but compromised computer systems that are managed by third parties. These computers are used to carry out different types of attacks such as deploying malware, stealing data, DDoS attack etc. […]","categories":["AI Features"],"tags":["Cyber Security","Network security"],"author_name":"Harshajit Sarmah","publish_date":"2019-09-05T19:00:08","publication_year":"2019","word_count":695,"keywords":["Go","Cyber Security","programming_languages:R","AI","programming_languages:Go","Git","Network security","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-sinkholing-can-protect-your-network-from-malware\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10016663,"title":"How This AI-Enabled Chatbot Radically Transformed Cancer Care Amid Pandemic","content":"The critical industry that has been massively impacted by the pandemic is the healthcare sector; however, AI’s involvement has helped the industry weather the pandemic storm. The AI applications by companies bring back patients’ lives from the edge and improve diagnostics and treatment and help healthcare providers make informed decisions quickly. One such application has been developed by Hospido, India’s first holistic cancer care platform with which cancer patients can get the right treatment advice from India’s best cancer doctors, without visiting hospitals amid pandemic. The pandemic lockdown has forced many people, including cancer patients, to avoid hospitals and discontinue their treatment due to coronavirus risk. And this is what triggered Karan Chopra, the founder of Hospido to bring out quality healthcare to these cancer patients through telemedicine and satellite treatments centres for providing the right treatment at the right time. In this article, Analytics India Magazine, got in touch with him to understand how this startup revolutionised cancer carer amid pandemic. Transforming Cancer Care Amid Pandemic According to Chopra, Hospido currently runs its satellite centres in five cities. Its online platform has so far helped many cancer patients from unreachable cities like Arunachal Pradesh and Kashmir. To facilitate this, the startup has developed an AI-enabled chatbot called Cancer Dost, where patients can learn more about their cancer and the treatment course, free of cost. “With the bilingual option, Hindi and English, the chatbot manages to help patients from the length and breadth of this country,” said Chopra. Developed by the engineers at Hospido with the support of leading oncologists and following the international guidelines, this unbiased, advanced technology tool provides patients with the right guidance required for cancer treatment. It also gives doctors direction to consult and the right hospitals to visit amid the lockdown. Additionally, the startup is also working on a remote ICU and AI-based chemotherapy monitoring tool to ensure patients get the right treatment at their local hospitals. The chatbot — Cancer Dost, also connects the patients with leading cancer specialists from top hospitals of the country, via tele-consulting and in-person appointments. “We ensure a 12-hour turnaround upon receiving the patient request, and the appointments with these top doctors are on a priority basis,” added Chopra. Further to this, Hospido has also opened cancer treatment centres in the country’s tier-2 cities intending to decentralise healthcare and make quality cancer treatment more accessible. These centres have been established as per international standards and are run by medical staff trained under India’s leading oncologists. Adding to this, Chopra stated, “We are also providing services including home care, diagnostic testing, diet planning and counselling support to the patients to help them in their entire journey of battling with cancer.” Tech Behind Cancer Dost uses an online interface in the form of a chatbot to capture the user’s input, which is then sent to the RESTful API based application, which runs on the company’s server. The server runs the proprietary recommendation algorithm, which is optimised to search through our proprietary knowledge data-bank created by leading oncologists. Once a match is found, the AI-based algorithm provides the most relevant information matching the patient’s condition. This information is then provided to the user through the same chatbot in simple language to help them make the right treatment decisions in consultation with their treating doctor. The chatbot, Cancer Dost, has an easy user interface that asks four essential data points — organ affected; treatment taken, either none, surgery, chemotherapy or radiation; stage of disease; and patient fitness condition, i.e. active, weak or bedridden, to provide the resulting outcome. “Our engineers created this proprietary knowledge data-bank with the support of leading oncologists and as per international guidelines,” said Chopra, while speaking about the technology behind. “The platform provides output in terms of specialists to be consulted — medicine, radiologist, surgical oncologist, with recommended treatments, like surgery, radiation, chemotherapy, supportive care, and recommended diagnostic tests.” The AI-enabled chatbot also provides expert oncologists review and a treatment plan from experts on submitting the medical reports. The developers of the company currently leverage Angular, Laravel, MySQL, Bootstrap for developing the application. Besides, they also use a host of cloud SaaS like CRM, patient tracking, etc., to power the business. Wrapping Up The founder launched this startup amid COVID, seeing the challenges faced by cancer patients who have been struggling to visit hospitals for their daily check-up. Thus, Chopra strongly believes that COVID pandemic positively impacted digital health. At present, the startup is self-funded, and their revenues are helping them sustain amid this turbulent time. Competing with local healthcare delivery providers like hospitals and clinics and online service providers like Onco.com and Practo, Hospido aims to replicate the oncology playbook and expand to other areas of care, in the next five years. “India is facing a crunch of specialist doctors, with most of them clustered around the top ten cities. We are going to take digital health far and wide and make access available across the country,” concluded Chopra.","excerpt":"The critical industry that has been massively impacted by the pandemic is the healthcare sector; however, AI’s involvement has helped the industry weather the pandemic storm. The AI applications by companies bring back patients’ lives from the edge and improve diagnostics and treatment and help healthcare providers make informed decisions quickly. One such application has […]","categories":["Deep Tech"],"tags":["AI Chatbot","chatbot ai","chatbot india"],"author_name":"Sejuti Das","publish_date":"2020-12-29T10:00:00","publication_year":"2020","word_count":831,"keywords":["Go","API","TPU","AI","AI Chatbot","Git","RAG","Aim","chatbot ai","analytics","SQL","R","chatbot india"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","TPU","R","SQL","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-this-ai-enabled-chatbot-radically-transformed-cancer-care-amid-pandemic\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10065737,"title":"A new model that distills unsupervised features into high-quality discrete semantic labels","content":"Researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL), Microsoft, and Cornell University have developed a model to discover and localise semantically meaningful categories within image corpora without any annotation. The ML model, STEGO (Self-supervised Transformer with Energy-based Graph Optimisation), produces features for each pixel that are semantically meaningful and compact enough to generate distinct clusters. As a result, the model can detect and delineate objects at a finer granularity than classification or object detection systems. STEGO architecture STEGO uses a semantic segmentation method- the process of assigning a label to every pixel in an image. The model is built on top of the DINO algorithm, which learned about the world through 14 million images from the ImageNet database. STEGO refines the DINO algorithm through a learning process that emulates how human brains piece together information to make sense of the world. Source: arxiv.org STEGO consists of a frozen backbone that provides a source of learning feedback and helps predict distilled features. The segmentation head is a simple feed-forward network with rectified linear unit activation functions. The backbone remains un-trained, making it easy to train the model efficiently. For example, an NVIDIA V100 GPU card can be trained in under two hours. The STEGO architecture consists of a frozen backbone (neural network) that extracts global image features by global average pooling (GAP) spatial features: GAP(f). Then, a lookup table is made of each image’s K-Nearest Neighbors based on the cosine similarity in the backbone’s feature space. Each training minibatch consists of a collection of random images x and random nearest neighbours x sample x knn randomly from each image’s top 7 KNNs. The team also sampled random images, xrand, by shuffling x and ensuring that no image matched itself. The resultant model’s full loss becomes: L = λselfLcorr(x, x, bself ) + λknnLcorr(x, xknn, bknn) + λrandLcorr(x, xrand, brand) |________________|   |________________|    |__________________| Self K nearest neighbours       Random images Here, λ’s and the b’s manage the balance of the learning signals and the ratio of positive to negative pressure. The b parameters tended to be dataset and network-specific, but the system is kept in a rough balance between positive and negative variables. The images in the ‘CocoStuff’ and ‘Cityscapes’ datasets are riddled with small objects that are hard to identify at a feature resolution (40, 40). To manage the small objects while maintaining fast training times, the images are five-crop trained before learning KNNs. The technique allows the network to cover the images in detail and improves the quality of the KNNs. Five-cropping improves both Cityscapes results and CocoStuff segmentations. The final components of the architecture involve clustering and CRF refinement. STEGO’s segmentation features lean towards forming clear clusters due to its feature distillation process. To counter this effect, a cosine distance-based minibatch K-Means algorithm is applied to extract these clusters and compute concrete class assignments from STEGO’s continuous features. The clustered labels are then refined with a CRF to further improve spatial resolution. Performance STEGO offers better results on linear probe and clustering (unsupervised) metrics across various datasets than SOTA models like PiCIE, Deep Cluster, InMars, etc. Despite the backbone of these two datasets not being fine-tuned, DINO’s self-supervised weights on ImageNet are enough to resolve both settings simultaneously. STEGO also proves superior in simply clustering the features from unmodified DINO, MoCoV2, and ImageNet supervised ResNet50 backbones. Source: arxiv.org Conclusion The researchers showed modern self-supervised visual backbones can be refined to yield state of the art unsupervised semantic segmentation methods. However, despite the latest improvements, STEGO still faces bottlenecks like labelling issues. Using an unsupervised methodology, the researchers plan to take a large corpus of images and classify each pixel into an accurate and consistent ontology of objects.","excerpt":"STEGO nearly doubles in MIoU un both unsupervised as well as linear probe metrics in comparison to its predecessors.","categories":["AI Features"],"tags":["Computer Vision","MIT research","Object Detection","semantic segmentation"],"author_name":"Kartik Wali","publish_date":"2022-04-26T17:00:00","publication_year":"2022","word_count":619,"keywords":["Go","artificial intelligence","AI","neural network","ML","ResNet","semantic segmentation","RAG","object detection","Object Detection","Computer Vision","MIT research","R","T5"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","RAG","object detection","R","Go","T5","ResNet"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-new-model-that-distills-unsupervised-features-into-high-quality-discrete-semantic-labels\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10088204,"title":"Google&#8217;s Winning the Automotive AI Race","content":"With Google’s recent partnership with Mercedes-Benz, a “next-generation navigation experience” is not too far away. In this deal, Google Maps will provide geospatial data and navigation capabilities for the car manufacturer, while Mercedes-Benz will use Google Cloud’s AI and machine learning capabilities to create, train and deploy AI models at speed. This will enhance customer experience, alongside building faster and more efficient data processing platforms to analyse fleet data. Additionally, it also plans to leverage Google’s open infrastructure to secure and scale from on-prem to the edge to the cloud, across its technology ecosystem. Google chief Sundar Pichai said that the company will provide AI and data capabilities to accelerate their sustainability efforts, advance autonomous driving, and create an enhanced customer experience. The tech giant’s association with Mercedes-Benz should not be construed as a mere attempt to work in the infotainment or navigation space alone, but rather slowly build its way as a crucial player in the AI automotive space. Mercedes has been making waves in the autonomous driving space with SAE Level 3 powered by the NVIDIA Drive platform, making it the first automaker in the US to receive this level ahead of Tesla, which is still in the Level 2 autonomy. Google’s efforts to create an AI niche in the autonomous driving segment has been brewing for quite some time. It is going through a series of collaborations with leading automobile companies, and Mercedes is not the first that will help Google achieve this. In November 2022, Renault Group and Google announced a partnership to build a ‘software-defined vehicle’. SDV is meant to combine the automotive technology and software capabilities of Google to build on the existing Android Automotive OS. Google was also named the preferred cloud supplier for Renault Group, where the Google Cloud technology would be employed for data capture and analytics. In 2021, Google collaborated with Ford, where the team has been using Google Cloud, Android and Google AI to help Ford transform its business and build automotive technologies. Automobiles Powered by Cloud The partial exclusivity of Google in the software segment of automobiles is an added advantage. Some of the big brands where Google applications are built are Volvo, GMC, Chevrolet, Polestar, Cadillac, and Honda. Google’s Android Auto, an application that mirrors an android device onto a car’s entertainment unit, has been Google’s established software dominance over automobiles. Launched in 2015, Android Auto is part of Open Automotive Alliance, an association of auto manufacturers and tech companies to promote Android in vehicles. Android Auto is available in 46 countries and supports over 500 models. The only counterpart to Android Auto is Apple’s CarPlay. And though it has similar features, it is not superior to Android Auto. The make-or-break deal is Google Map’s edge over Apple Maps. But, Google is not alone, both Amazon and Microsoft have been partnering with automotive companies to leverage their technologies.  In October 2022, BMW partnered with Amazon, integrating AWS cloud computing in their systems. Similar to Google, Amazon is developing software-defined vehicles integrating AWS, which would help towards building mobility solutions for autonomous vehicles. Amazon is also working with car manufacturer Stellantis, makers of brands Alfa Romeo, Chrysler and Jeep. Microsoft has also tied up with automobile companies like Volkswagen to provide their cloud solution, Azure. In 2017, Microsoft had tied up with Tata Motors. Self-Driving Ambitions According to McKinsey research, the autonomous driving segment is said to generate $300 to $400 billion revenue by 2035. By 2030, 12% of the passenger vehicles sold will have L3+ technologies. This is an indication to how AI and ML tools will be widely adopted in the automotive sector in the coming years. By the way, Google was one of the first few technology companies to ambitiously take up self-driving car projects. The project, which was renamed ‘Waymo‘, first took shape in 2009, and launched as the 4th generation Waymo Driver (Chrysler Pacifica minivan) ten years later. In 2021, the 5th generation Waymo Driver was launched (electric Jaguar I-PACE), and it probably was as far as Google went as a manufacturer of autonomous vehicles. Waymo has been functioning as a ride-hailing service, and though fraught with technical problems, Google is still pushing updates to its self-driving segment. Waymo also signed a partnership with Uber in June 2022 to deploy autonomous trucks on the Uber Freight network. However, the company has held on to its aggressive expansion plans for Waymo. However, as part of the recent layoffs at Google, Waymo employees were given the pink slip. There were also claims that the company will hold back expansion on its autonomous trucks ‘Via’. Having spent over 12 years in a self-driving project, Waymo is still far from making profits. With Waymo’s growth trajectory, it is practical for Google to continue to prioritise developing their software solutions which can be implemented in the automotive sector, as opposed to the conventional method of investing in building autonomous cars. Google’s latest partnership with the luxury carmaker is a step towards software dominance in the automobile sector, and in farsight probably paves the way for autonomous driving and take on other biggies of AI in the automotive space. Though there are other cloud contenders in the automotive space, Google has a clear edge over them. Having already entered the autonomous driving segment with its flagship driverless cars, Google is ahead in the AI automotive space.","excerpt":"The tech giant has partnered with Mercedes- Benz, Renault, Volvo, GMC and many other automobile companies","categories":["Global Tech"],"tags":["Amazon","Autonomous Vehicles","AWS","Azure","Google","Google Cloud","Google Map","Mercedes-Benz","Microsoft","NVIDIA","Waymo"],"author_name":"Vandana Nair","publish_date":"2023-02-27T14:30:00","publication_year":"2023","word_count":897,"keywords":["Mercedes-Benz","Autonomous Vehicles","R","Google Map","RAG","analytics","NVIDIA","Google Cloud","machine learning","AWS","AI","cloud computing","ML","Amazon","Aim","Waymo","Google","Azure","Microsoft"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","cloud computing","AWS","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-is-winning-the-automotive-ai-race\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10041726,"title":"Apple WWDC, Intel Eyes SiFive And More In This Week’s Top News","content":"Apple kicked off its latest edition of developers conference WWDC with privacy augmented reality in focus. Earlier this year, Apple introduced privacy labels to allow users to take an informed decision on data sharing. The iPhone maker’s privacy push was reportedly so aggressive that it has annoyed companies like Facebook, which blamed Apple for encroaching their ad-territory. At this year’s WWDC, Apple flaunted more updates related to privacy. “Privacy has been central to our work at Apple from the very beginning,” said Craig Federighi, Apple’s senior VP of Software Engineering. At the virtual summit, Apple stressed on how they develop new technology to help users take more control of their data and make informed decisions about whom they share it with. This year’s updates include Private Relay, which is a new internet privacy service that’s built right into iCloud, allowing users to connect to and browse the web in a more secure and private way. When browsing with Safari, Private Relay ensures all traffic leaving a user’s device is encrypted, so no one between the user and the website they are visiting can access and read it, not even Apple or the user’s network provider. Apple, which claims to own the world’s largest augmented reality platform with over 1 billion AR-enabled devices, went heavy on AR updates with new APIs in RealityKit 2, which will enable developers to create more realistic and complex AR experiences with greater visual, audio, and animation control, including custom render passes and dynamic shaders. Apple also had major updates for the Swift programming community. Swift now features built-in concurrency support. This means developers can more easily write code that does work in parallel. In other news Apple also hired a BMW veteran hinting at their electric car ambitions. Check our full coverage of the highlights from WWDC21 here. Intel Eyes SiFive According to reports, Intel is in plans to acquire SiFive for more than $2 billion. Intel’s interest in SiFive can be linked to its losing turf war with Nvidia, which became more evident since the acquisition of ARM last year. Like ARM, SiFive licenses its chip designs to various customers.  SiFive chips are based on the RISC-V architecture and the company has been betting heavily on bringing the power of the open source RISC-V ISA that it invented to develop domain-specific silicon faster than ever before. Intel has been proactive in 2021. It became more evident ever since the new chief Pat Gelsinger took charge. The once upon a time pioneers of the Silicon Valley are leaving no stone unturned to regain their lost territory to rivals such as Nvidia and AMD. The company announced its plans to build multi-billion dollar worth chip fabrication facilities. It has even partnered with TSMC to boost chip production. Making breakthroughs in chip design is a laborious process to say the least. And, SiFive has all the right ingredients that can give Intel much needed edge to bring custom silicon to the masses. NVIDIA Acquires DeepMap Image credits: NVIDIA On Thursday, NVIDIA announced the acquisition of DeepMap, a startup that builds high-definition maps for autonomous vehicles. DeepMap was founded five years ago by Wu and Mark Wheeler, veterans of Google, Apple and Baidu, among other companies. DeepMap’s technology will be incorporated into the mapping and localization capabilities of NVIDIA DRIVE, to enhance driver’s safety. “The acquisition is an endorsement of DeepMap’s unique vision, technology and people,” said Ali Kani, VP, Automotive at NVIDIA. “DeepMap is expected to extend our mapping products, help us scale worldwide map operations and expand our full self-driving expertise.” OpenAI Finds A Way To Make GPT-3 Unbiased We've found we can improve AI language model behavior and reduce harmful content by fine-tuning on a small, carefully designed dataset, and we are already incorporating this in our safety efforts. https:\/\/t.co\/nJISaAyY2M pic.twitter.com\/AJe8bgkzRl— OpenAI (@OpenAI) June 10, 2021 In their latest research paper, OpenAI, known for popular language models such as GPT-3, has claimed to found a way to improve language model behavior. The researchers proposed a Process for Adapting Language Models to Society (PALMS) with Values-Targeted Datasets, an iterative process to significantly change model behavior. According to the results, the fine-tuned techniques improved the language model by a great margin in the direction of mitigating bias. “While the technique is still nascent, we’re looking for OpenAI API users who would like to try it out and are excited to find ways to use these and other techniques in production use cases,” said OpenAI in their blog. The US to Inject $250 Billion Into Tech On Tuesday, the Senate passed the U.S. Innovation and Competition Act, a $250 billion bill to accelerate innovation infrastructure and fortify America’s position as manufacturing and technology giant. “ It will empower us to discover, build, and enhance tomorrow’s most vital technologies — from artificial intelligence, to computer chips, to the lithium batteries used in smart devices and electric vehicles — right here in the United States,” said the President in a statement. End Game for Patent Trolls In Germany Germany on Friday tweaked a legislative loophole that allowed patent trolls to operate freely.  Over the years, Germany has become a paradise for patent litigators who target fast-growing tech companies. Germany’s fast track courts allowed patent owners to almost instantly obtain temporary bans from courts on the sale of products accused of patent infringement. Protecting intellectual property can be nightmarish for courts. Bigger companies can claim ownership of technologies that otherwise would have benefited people. Whereas, there can be patent trolls who can obstruct innovation by suffocating companies with lawsuits. This new law by Germany has seen support coming from companies like Google, Samsung and chip maker Nvidia. But, critics warned that this new law is one-sided and will be curb innovation. Coursera Puts India In A Tight Spot (Image credits: Coursera) Popular edtech platform, Coursera released its 2021 Global Skills Report (GSR). The reports ranked India 67th globally in data skills.  In Asia, India ranks 16, below Singapore and Japan, but above Thailand and Philippines. India’s digital skill proficiency stood at 38 percent.  Globally, India ranks 55th in skills and 66th in technology and data science. The insights in this year’s report are based on Coursera platform data and research from Q1 2020 to Q1 2021.","excerpt":"Apple kicked off its latest edition of developers conference WWDC with privacy augmented reality in focus. Earlier this year, Apple introduced privacy labels to allow users to take an informed decision on data sharing. The iPhone maker’s privacy push was reportedly so aggressive that it has annoyed companies like Facebook, which blamed Apple for encroaching […]","categories":["AI News"],"tags":["Apple","Intel","OpenAI"],"author_name":"Ram Sagar","publish_date":"2021-06-13T10:00:00","publication_year":"2021","word_count":1045,"keywords":["data science","Go","artificial intelligence","OpenAI","AI","AWS","Apple","Git","RAG","Aim","R","Intel"],"extracted_tech_keywords":["AI","artificial intelligence","data science","OpenAI","Aim","RAG","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/apple-wwdc-intel-sifive-nvidia-top-news\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10063872,"title":"Intuit&#8217;s Radhika Kannan on how the company is leveraging AI to enhance customer experience","content":"As per Market Data Forecast, the global fintech market will touch $324 billion by 2026. The growth of the segment is expected to be fueled by rapid digitisation. As of November 2021, there were about 10,775 fintech companies in the US. In India, there are about 1,860 startups, and as of December 2021, the country had over 17 fintech unicorns with a valuation of over $1 billion. In recent times, technology has paved the way for disruption and transformation in the financial services industry in India. Advancements within the fintech space have fostered innovation and powered fintech companies like Intuit to deliver best-in-class customer solutions. As a global technology platform company, Intuit helps customers and communities overcome their most important financial challenges. The company serves millions of customers worldwide with cutting-edge technology products like TurboTax, QuickBooks, Credit Karma, Mailchimp, and Mint. “We believe that everyone needs to have the opportunity to thrive and we never stop working to find innovative ways to make that possible,” said Radhika Kannan, Staff Technical Program Manager for AI at Intuit. However, Intuit isn’t alone in this space. It competes with a gamut of players, such as Xero, Sage, Square, H&R Block, Blucora, Liberty Tax, Experian, and Equifax, who are as vibrant and dynamic but mostly focus on gaining value from AI implementations. Kannan adds that Intuit’s biggest USP is its strategy of being an AI-driven expert platform, solving problems with a customer and platform-focused mindset. “AI helps our customers work better and smarter because we can predict, automate and personalise their experience. Powered by data, this integrated products and services platform is accelerating our speed of innovation,” said Kannan. She said that using AI technologies, some of the applications they are working on include – Leveraging ML to build decision engines and algorithms that learn from rich datasets to transform user experiences (UX) Applying knowledge engineering to turn compliance rules into codeUsing natural language processing to revolutionise how customers interact with products and services “We believe our AI-driven expert platform will revolutionise experiences for customers and help people make the right financial decisions for themselves, their business, as well as their families. People across the world will be able to instantly access the skills, insights and expert advice they need, while at the same time experts will gain access to new clients they wouldn’t reach otherwise,” said Kannan. Overcoming challenges in AI “One of the biggest challenges currently prevailing in the industry is properly structuring and architecting data so that AI can be applied and implemented easily,” said Kannan. She said that includes making sure that there is a clean pipeline of data processing and accessibility of data in real-time so that you can provide actionable insights to users. Additionally, Kannan said that in an ever-changing privacy regulation landscape, companies would need to make it easy for end customers to understand how their data could be used, alongside giving them the freedom to select who they trust with that data. Like many big data players, Intuit stores copies of unstructured and structured data in a large enterprise data lake. Kannan said that the data lake is important to Intuit’s AI-centric strategy, as data analysts and data scientists tap into this rich data set to build machine learning models and gain insights to deliver personalised customer experiences with their products and services. She said that this data is managed according to a strict set of ‘Data Stewardship Principles’ that govern the complete lifecycle for data. “These Data Stewardship Principles are working on changing the data culture at Intuit, enabling us to have well-designed, domain-driven data to start building AI,” added Kannan. Furthermore, she said that setting up the initial strategy and investment into AI as a strategic priority can be a barrier for some businesses. “Often, a compelling business case needs to be developed. In some situations, that business case may need to be based on data that is not available yet because the new product is not yet in production,” said Kannan, saying that it becomes the classic ‘chicken-and-egg problem.’ She adds that this is the reason it’s important to document the assumptions in a business case and the assumptions about AI. To solve this hassle, the team at Intuit has created a core set of innovation and design thinking methodologies, known as ‘customer-driven innovation’ and ‘design for delight’ that encompasses these and related issues. “The key in both of these approaches is making sure we obsess about falling in love with the customer problems and not our solutions. By keeping this focus, we continue to remain nimble and flexible in seeking new, innovative solutions to evolving customer problems,” said Kannan. Carving a niche Kannan said that at Intuit, they’re focused on powering prosperity for their 100 million customers through AI. This includes ● Financial insights, benchmarks, cash flow forecasts and advice – actionable insights for QuickBooks Advanced (QBO-Adv) customers and cash flow advice to help businesses understand where they are ● Clean, organised data – real-time feedback-driven system for transaction understanding enabling insights in the customers’ first session ● AI advisor – advisory experiences in Mint (personal finance product) moving from a look back to planning ahead ● Connect with an expert – context-aware digital experts in QuickBooks and TurboTax to help customers proactively, matchmaking for live offerings ● Expert augmentation – bringing ML intelligence and recommendations to experts Intuit AI and analytics team members regularly participate in premier international forums like KDD (Knowledge Discovery and Data Mining) to share their ideas, research results and experiences via presentations, workshops and demonstrations. Intuit’s product portfolio Kannan said that across their platform, they had put the power of technology and data on the side of their customers. “Our product delivers simple, delightful solutions across the full range of our customers’ financial lives – and we are proud that millions of people worldwide turn to Intuit to make the most of their money,” she added. Here’s the list of products developed by Intuit ● TurboTax takes the complexity out of the tax code, helping users get the maximum refund they deserve. ● QuickBooks helps small businesses manage their books, get paid fast, manage capital, and pay employees with confidence. ● Mint helps users get a comprehensive view of their financial picture and make smart money decisions. ● Credit Karma helps people find the right financial products, putting more money in their pockets and providing financial expertise and insights. ● Finally, Mailchimp provides an all-in-one marketing platform that empowers customers to start and grow their businesses by helping customers get their business online, market their business, and manage customer relationships. Tech stack Kannan said at Intuit, analysts glean insights from customer data available across the ecosystem, and AI teams focus on building customer experiences driven by machine learning, knowledge engineering, and natural language processing, delivering personalised, automated and insightful products. Further, she said that the team combines AI with human expertise to innovate and deliver unimaginable benefits for our customers along with areas such as – ● Protecting customers from financial loss ● Building personalised smart products with AI-based predictions, recommendations, insights, and workflow automation that are ethical, explainable, engaging, and fair ● Inspiring trust in human expertise and AI-driven insights The team also focuses on creating reusable self-service AI capabilities to help scale quickly across the company. Kannan said that they currently have AWS infrastructure support, including EMR notebooks that can run distributed workloads for large datasets with sizes of the order of TBs, Sagemaker notebooks having CPU and GPU capabilities that can scale according to needs, Databricks, etc. Besides this experimentation infrastructure, Intuit also has their own internally developed frameworks to help abstract the infrastructure details from data scientists. Future roadmap “As we build our AI-driven expert platform, we prioritise our resources on five top priorities, or Big Bets, across the company,” informed Kannan. She said these five key priority areas go after the biggest customer problems and are the largest growth opportunities for the company. These include revolutionising speed to benefit, connecting people to experts, unlocking smart money decisions, being the centre of small business (SMB) growth, and disrupting the small business mid-market. In addition to this, Intuit has also charted out some Bold Goals for 2025, including doubling household savings rate and improving SMB success rate greater than 10 points versus the industry and increasing their customers to 200 million, alongside accelerating their revenue growth. Intuit is hiring! To know more about Intuit or explore careers at Intuit visit https:\/\/www.intuit.com\/careers\/oa\/technology\/","excerpt":"As a global technology platform company, Intuit helps customers and communities overcome their most important financial challenges.","categories":["Global Tech"],"tags":["Intuit","unstructured data"],"author_name":"Amit Naik","publish_date":"2022-03-30T12:00:00","publication_year":"2022","word_count":1417,"keywords":["Go","machine learning","AWS","AI","Intuit","ML","RAG","Databricks","analytics","Rust","R","unstructured data"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","AWS","Databricks","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/intuits-radhika-kannan-on-how-the-company-is-leveraging-ai-to-enhance-customer-experience\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10055161,"title":"Fueling Research Or Another Publicity Stunt? The Truth Behind Zuckerberg-Funded AI Institute","content":"Harvard University recently launched a new AI institute funded by American philanthropist Priscilla Chan and her husband and Meta founder, Mark Zuckerberg. The Kempner Institute for the Study of Natural and Artificial Intelligence stands between the intersection of neuroscience and artificial intelligence (AI). The institute will be focusing on AI and artificial neural networks research, developing theories to understand how AI learns and functions. Zuckerberg and Chan have committed $500 million over the next 15 years towards the institute, which is scheduled to open at the Science and Engineering Complex in Allston by the end of 2022. The institute is named after Facebook founder Zuckerberg’s mother and his grandparents – Karen Kempner Zuckerberg, Sidney Kempner and Gertrude Kempner, respectively. Does Meta actually want to accelerate AI research by funding the new AI institute, or is it just another publicity stunt by the company to camouflage its recent controversies? What’s in store? With the monetary gift from the Chan Zuckerberg Initiative, the Kempner Institute for the Study of Natural and Artificial Intelligence plans to support ten new faculty members from the underrepresented STEM fields, a new computing infrastructure and lab resources. It will be headed by Bernardo Sabtini— the Alice and Rodman W Moorhead III Professor of Neurobiology at the Harvard Medical School, and Sham Kakade— Professor in the Faculty of Arts and Sciences, appointed to become the Professor of Computer Science and Statistics, starting January. The AI institute hopes to bring together approaches and expertise in cognitive sciences, neuroscience and biology with computer science, engineering, applied mathematics, machine learning and statistics. Thus, it aims to give rise to a new generation of leaders versed in both subjects. The main purpose behind initiating this institution is to understand how the human brain works, in order to accelerate and improve the addressing and management of diseases, creation of new therapies, and understand the human body better. The institute will be supporting the training of students– undergraduate, graduate and postdoctoral fellows and researchers. Research topics are said to cover the lengths and breadths of perception and sensation, brain functioning and meta-plasticity. Another big tech influence? In the last couple of years, big techs have been pouring in resources backing and establishing AI centres. Last year, Columbia University launched an AI research centre backed by Amazon. According to sources in the university, Amazon has committed a $5 million investment in the institute in the following five years, focusing on responsible AI. On the other hand, Google has invested $250 million into the academic community between 2005 and last year. Furthermore, Samsung invested $1.5 billion into research institutes in Korea. However, tech giants’ interests and investments in academia are extremely controversial. With big tech like Google, Amazon, Microsoft, and of course, at the forefront of it all, Meta (earlier Facebook) holding so much control over public data, it has been but raising eyebrows. Experts believe that future leaders graduating from these tech-funded educational institutions will eventually research and develop biased solutions, catering to these conglomerates. Recently, computer scientist Timnit Gebru launched Distributed Artificial Intelligence Research Institute — an AI institution to take on the Big Tech influence in the field of AI research and development. This comes almost a year after the ethical AI practitioner was fired from Google. The Chan Zuckerberg Initiative was founded in 2015 to cater to ‘social challenges’ through philanthropy and scientific research. In fact, Chan and Zuckerberg have committed to donating 99 per cent of Facebook’s equity for the betterment of the research study. Earlier, in 2017, the initiative donated $12.1 million at Harvard for public service initiatives and later another $30 million in 2018 to its Graduate School of Education. On paper, this might seem noble. However, if history is to be believed, the Kempner Institute for the Study of Natural and Artificial Intelligence seems like another attempt by the ever-controversial Meta to influence research and control the future of AI and technology at large.","excerpt":"Zuckerberg and Chan have committed $500 million over the next 15 years towards the Kempner Institute for the Study of Natural and Artificial Intelligence.","categories":["AI Features"],"tags":["AI Research","Big Tech","Google","harvard university","Mark Zuckerberg","Meta","Microsoft","Samsung","Tech giants"],"author_name":"Debolina Biswas","publish_date":"2021-12-09T17:00:00","publication_year":"2021","word_count":656,"keywords":["API","R","funding","artificial intelligence","Tech giants","Mark Zuckerberg","harvard university","Go","Meta","machine learning","AI","neural network","Big Tech","Samsung","AI Research","Aim","Google","responsible AI","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","Aim","R","Go","API","responsible AI","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/fueling-research-or-another-publicity-stunt-the-truth-behind-zuckerberg-funded-ai-institute\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":17101,"title":"6 Job Profiles That Can Make Way To A Data Scientist Role","content":"While there is already enough narrative out there on the kind of skills that goes into the making of a data scientist, there is very little information on the job profiles that can lead to Data Scientist job roles.  Hailed as the sexiest job of 21st century, data scientist job title has attracted too much attention over the last few years, yet employers find it tough to make a match for this job role. With more and more companies becoming digital and data science startups flush with VC money mushrooming up, analytics related jobs are on a definite rise. But the talent pool of data analysts and data scientists has stayed small. Sometimes, it is hard to find the perfect match of data science and analytics competencies for hiring these data-driven decision makers. And it’s not just analytical-enabled and machine learning skills that can land a top spot, domain knowledge, exceptional visualization skills, curating data (wrangling and cleaning), knowledge of data governance ethics also fit the recruiter’s description. In this article, we try to focus on what starting job roles can be ideal for someone looking for a data scientist’s profile. Data scientist role combines Statistics, technology and business acumen and all these 3 skills should work in tandem. These below listed profiles can be a starting point for someone who eventually would like to Analytics India Magazine lists down six top job profiles that can make way to an ideal Data Scientist role 1. Data Modeler A data modeler is tasked with translating business needs into data models – schematic representation of data. The first step in database design, data modelers create a conceptual model to see how various data objects relate to each other. Data modelers come from a mathematical background and some of the most popular enterprise data modeling tools used are SAP PowerDesigner, CA Erwin and ER\/Studio. Data modelers have a deep background in mathematics and statistics and have knowledge of popular programming languages. Besides, R and C++ are also part of their toolkit. An integral part of an agile project, their Entity Relationship Diagram (ERD) forms the basis of an enterprise’s data strategy. With a large experience of managing datasets, data modelers also have a good understanding of business requirements and the functional requirements of an application being developed. That’s why, they Data Modelers can go on to become Data Scientists. 2. R programmer Between SAS and R, R emerged as a dominant choice of analytics tool among data analysts and this has led to a rise of R programmers. With R having more powerful data analysis tools, R programming is on a rise, especially in sectors such as BFSI and pharma where R is used more for simulation as opposed to analysis. In the changing skill landscape, R programmers are in high demand for their in-depth experience in automated report generation. And that’s why they make a perfect fit for a data scientist job. 3. Database engineer Database engineer have long been propped up as the ideal candidate for the data scientist job. Here’s why- they manage the data pipeline and are tasked with the installation of database products, have in-depth knowledge of SQL, DBA scripts and other software. Some of the tools used by data engineers are Hadoop, Keras, Pandas and Python. Thanks, partly to the complexities of the job and their background in coding and managing data pipeline, database engineers usually transition to the role effectively. 4. Machine Learning specialist ML specialists are all the rage these days, with deep expertise in running standard machine learning models ranging from regression, binary classification and multiclass classification. From training and optimizing ML algorithms to developing new applications, ML specialists are in huge demand in big and small companies for their heft in predictive analysis and fulfilling business requirements. ML specialists graduate to become data scientists and also go on to lead data science teams. Popular tools used are Python\/R and Keras among others. 5. Deep Learning Expert A very rare field which requires great expertise, deep learning involves neural network, a subject of great research. ML specialists who have pursued neural networks and Convolutional Neural Networks in research go on to become Deep Learning experts. A relatively new profile, DL experts are drawn from the academic world and are staffed at leading tech companies across the globe – Facebook, Google, Baidu, Microsoft, OpenAI and not to forget DeepMind. DL guys work a lot on GPUs and work on problems such as image detection, and autonomous driving system. And that’s why they get out vote of acing the job of data scientist. 6. NLP specialist A hard nut to crack in computer science, Natural Language Processing developers organize and structure data to perform tasks such speech recognition, sentiment analysis, text analysis, and more. Today, NLP specialists are in huge demand thanks to enterprises building chatbots for automated question answering and machine translation. Besides creating a chatbot, NLP experts are also tasked with sentiment analysis, Named Entity Recognition and summarizing blocks of text. They are staffed at marquee tech companies such as Facebook, Google, Amazon, ecommerce and financial sector. NLP algorithms are used to filter malicious comments, suggest trending topics and social media monitoring. NLP experts have deep expertise in machine learning algorithms, are exceptional coders, have hands-on experience in executing new data projects. Hence, they make a perfect data scientist.","excerpt":"While there is already enough narrative out there on the kind of skills that goes into the making of a data scientist, there is very little information on the job profiles that can lead to Data Scientist job roles.  Hailed as the sexiest job of 21st century, data scientist job title has attracted too much […]","categories":["AI Trends"],"tags":["data scientist career path","data scientist qualifications"],"author_name":"Richa Bhatia","publish_date":"2017-08-21T08:36:40","publication_year":"2017","word_count":895,"keywords":["data science","data scientist career path","machine learning","Keras","data scientist qualifications","AI","neural network","OpenAI","ML","NLP","deep learning","analytics"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","data science","analytics","OpenAI","Keras"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/6-job-profiles-can-pave-way-data-scientist-role\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10090735,"title":"What BloombergGPT Brings to the Finance Table","content":"Last week, Bloomberg released a research paper on its large language model BloombergGPT. Trained on over 50 billion parameters, the LLM model will be a first-of-its-kind AI generative model catering to the finance industry. While the move may set a precedent for other companies, for now, the announcement sounds like a push for the data and news company to seem relevant in the AI space. Interestingly, Bloomberg already has Bloomberg Terminal, which employs NLP and ML-trained models for offering financial data. So, naturally, the question that arises is: how much of a value-add is BloombergGPT and where does it stand in comparison to other GPT models? Training and Parameters Bloomberg’s vast repository of financial data over the past forty years, has been used for training the GPT model. It is trained on 363 billion token proprietary datasets (financial documents) available from Bloomberg. In addition, 345 billion token public datasets were also incorporated to result in a total of 700 billion tokens for training. The company claims that the new model (Bloomberg GPT) will help improve their already existing NLP tasks such as sentiment analysis – a method that helps predict market prices – news classification, headline generation, question-answering, and other query-related tasks. Here is an example of BloombergGPT being used to generate valid Bloomberg Query Language. As we have seen with other models like GPT-3, this model can, with a few examples in input prompt, utilize knowledge about stock tickers and financial terms to compose queries for data… pic.twitter.com\/tMumrgnzX3— elvis (@omarsar0) March 31, 2023 On the face of it, the new LLM model appears great, but is still very limited in its approach. It’s not a multilingual model, has biases and toxicity and is a closed model. Multilingual BloombergGPT, the 50-billion parameter ‘decoder-only causal language model’ is not trained on multilingual data. Their training dataset, called FinPile, includes news, filings, press releases, web-scraped financial documents, and social media drawn from the Bloomberg archives, and they are all in the English language. For instance,  to train the model on data from press conferences, transcripts of company press conferences through speech recognition were used in the English language. The absence of multi-languages limits input training data. BLOOM, which has the same model architecture and software stack as BloombergGPT (though BLOOM is trained on higher parameters of 175 billion), is multilingual. Similar is the case with GPT-3, which is also trained on multilingualism and 175 billion parameters. Biases and Toxicity Bloomberg has mentioned that the possibility of the “generation of harmful language remains an open question”. LLMs are known for their biases and hallucinations, a problem that large trained models, such as ChatGPT, are also combatting. LLM bias can be highly detrimental when utilised in finance models, as accurate and factual information determines the rightful prediction of market sentiments. However, BloombergGPT does not address this concern completely. The company is still evaluating the model and believes that “existing test procedures, risk and compliance controls” will help reduce the problem. Bloomberg is also studying their FinPile dataset which contains lesser biases and toxic language, which will ultimately curb the generation of inappropriate content. Closed Model BloombergGPT is a closed model. Apart from the parameters and general information, details such as the weights of the model are not mentioned in their research paper. It is possible that since this model is based on decades of Bloomberg data, clubbed with its sensitive nature of information, the LLM will not become open sourced. Besides, the model is set to target their Bloomberg Terminal users, who are already availing the service at a subscription cost. However, the company does have plans to release training logs of the model. In a conversation with AIM, Anju Kambadur, head of AI Engineering at Bloomberg, said: “BloombergGPT is about empowering and augmenting human professionals in finance with new capabilities to deal with numerical and computational concepts in a more accessible way.” Bloomberg has been using AI, Machine Learning and NLP for more than a decade but each of them required a custom model. “With BloombergGPT, we will be able to develop new applications quicker and faster, some of which have been thought about for years and not developed yet,” he said. “Conversational English can be used to post queries using Bloomberg Query Language (BQL) to pinpoint data, which can then be imported into data science and portfolio management tools.” Kambadur clarified that BloombergGPT is not a chatbot. “It is an ingredient model that we are using internally for product development and feature enhancement.” The model will help power AI-enabled applications like Bloomberg Terminal, but also power back-end workflows within our data operations. Clients may not engage with the model directly but will be using it through the Terminal functions in the future. Comparison Below is a comparison with other models GPT-NeoX (trained on 20B parameters) and FLAN-T5-XXL (trained on 11B parameters). BloombergGPT, updated on the latest information, is able to answer the questions accurately when compared to other similarly-trained LLMs. Source: arxiv.org BloombergGPT fared better on financial tasks when compared to other similar open models of the same size and was even evaluated on the ‘Bloomberg internal benchmarks’ and other general-purpose NLP benchmarks such as BIG-bench Hard, knowledge assessments, reading comprehension and linguistic tasks.","excerpt":"The latest LLM by Bloomberg, trained on 700 billion tokens, is an ingredient model said to boost Bloomberg Terminal service","categories":["Deep Tech"],"tags":["bias","Bloom","Bloomberg","BloombergGPT","ChatGPT","goldman sachs","ML","NLP","Parameters","Sentiment Analysis"],"author_name":"Vandana Nair","publish_date":"2023-04-04T13:00:00","publication_year":"2023","word_count":872,"keywords":["data science","ChatGPT","Sentiment Analysis","machine learning","AI","sentiment analysis","R","ML","BloombergGPT","Bloom","NLP","GPT","Bloomberg","goldman sachs","Parameters","Aim","bias"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","data science","ChatGPT","Aim","sentiment analysis","R","GPT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/what-bloomberggpt-brings-to-the-finance-table\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10003074,"title":"Key Technology Takeaways From Microsoft Inspire 2020 Event","content":"Microsoft’s Inspire is one of the biggest events for enterprise technology solutions. Just like the annual Build 2020 conference for developers, this year’s Inspire was a two-day virtual event from Microsoft for its partner ecosystem which builds and sells cloud-to-edge products and solutions to thousands of businesses around the world. According to Microsoft, the tech giant is constantly scanning technology requirements for business and developing applications that solve the complexity of operations in cloud infrastructure or data intelligence. During this year’s virtual Inspire event, Microsoft announced many new features for its products and services cross Azure, Business Applications, Microsoft 365 and Teams, Security, along with new tools for partners. Let’s take a look at the various solutions announced at Inspire 2020, which can be useful to you or your business. Next-Generation Of Azure Stack HCI Hybrid cloud is vital for a majority of enterprises around the world. Microsoft’s approach to hybrid is to serve their customers’ particular needs that demand a hybrid environment or have reasons for keeping some applications or workloads operating within their own data centres. At Inspire, Microsoft announced the latest part of Microsoft’s hybrid cloud portfolio- the next generation of Azure Stack HCI, which is native and completely integrated Azure service that gives businesses the easiest and fastest way to integrate their data centre with the cloud. Organisations use Azure Stack solutions to meet their hybrid requirements across data centres, remote offices, and edge locations. The service is Azure consistent and can be run as an Azure managed appliance that provides intelligent compute and AI at the edge with Azure Stack Edge. With the latest update, IT administrators can also use a new deployment wizard to quickly set up an Azure Stack HCI cluster and connect to Azure and take advantage of Azure Stack HCI native integration with core Azure services. This gives flexibility for running hybrid applications with Azure Stack Hub high-performance virtualisation on-premises with Azure Stack HCI. “We’re witnessing an increased urgency for digital transformation, and technology is playing an important role in helping customers find the resilience and reimagination needed to navigate this time of disruption. Azure Stack HCI benefits from the latest technologies from Intel’s broad portfolio, so organisations can fast modernise their cloud infrastructure,” stated Jason Grebe, Corporate Vice President, General Manager, Cloud Enterprise Solutions Group at Intel. He further added that with this, users could speedily adapt in a dynamic world and have cloud efficiency for on-premises workload, while gaining the flexibility, performance, and scale they trust from Microsoft and Intel. Apart from Intel, other key partners that deliver Azure Stack HCI solutions include Dell EMC, HPE and Lenovo. New Azure Migrate Enhancements To provide the significant benefits of the cloud and virtualisation, Microsoft is also launching multiple new migration features to support companies act faster, and expand IT operations and processes. The new Azure Migrate enhancements can empower users to perform productive data centre assessments, including the capacity to import and make assessments utilising uploaded Configuration Management Database data and support for Azure VMware. In addition to Azure Migrate, several Azure disks and networking improvements have been added to meet the needs for most essential workloads. This includes shared disks for Azure Disk Storage, private links for exporting and importing data safely over a private network and support for third-party virtual appliances within virtual wide-area network hubs. Microsoft Makes Available Its Supported Hadoop Distribution In another announcement at Inspire 2020, Microsoft brought great news for open-source lovers. Microsoft’s supported distribution of Apache Hadoop is now available and completely open-source as well as compatible with the most recent Hadoop version. Apache Hadoop was the original open-source framework for distributed processing and analysis of big data sets on clusters. With this distribution, organisations can provision a new HDInsight cluster based on Apache code that is designed and fully supported by Microsoft. The company says that customers will be migrated automatically to the latest supported distribution. New Features For Azure Blob Storage Another major announcement came in for Azure Blob Storage called Last Access feature, which is now in public preview and gives organisations better visibility into their data, including the inbuilt feature on how often the data is accessed. This is said to allow customers to handle and control their data lifecycles based on access time. This new system metadata is also available to independent software vendor partners and will help make placement and retention decisions for their data. Apart from this, Azure Blob Storage now supports Network File System (NFS) 3.0, allowing support for read-heavy data workloads across various industries.  Customers have to move data a lot of the time and within different storage platforms, and they are usually challenged by legacy and read-heavy apps running on NFS protocol. Now in public preview, this news upgrade helps in removing data silos by working as the sole storage platform which supports NFS protocol over objects. General Availability Of HashiCorp Consul Service On Azure In another cloud news at Inspire 2020, HashiCorp announced the general availability of HashiCorp Consul Service (HCS) on Azure. HCS was earlier introduced in private beta in September 2019 and now is generally available on the Azure Marketplace on July 23 in multiple regions in the United States, Europe and Asia. HCS on Azure is a completely managed service which allows organisational users to natively provision HashiCorp-managed Consul clusters straight through their Azure dashboard in various Azure regions. The service provides a simple, safe and secure service-networking solution so customers can conduct service discovery, service segmentation and service mesh across a mix of virtual machines, hybrid, on-premises and Kubernetes environments while offloading the operational challenge to HashiCorp’s team. This minimises complexity for customers and helps them to focus on cloud-native innovation. Azure Data Security Enhancements Microsoft has also announced a new connector for Azure Sentinel, which helps companies collect and analyse security data across various sources in the complex enterprise environment. New connectors announced now help in quickly getting security information and insights across multiple well-known solutions and partners,  including networks, and firewalls. Security connectors, in this case, are partner companies’ products including Alcide kAudit (Kubernetes logs), Vectra AI, Perimeter 81 (Activity logs),  RiskIQ (Azure Logic Apps custom connector), Symantec Proxy SG, Proofpoint TAP, Pulse Connect Secure, Symantec VIP, Infoblox NIOS, Qualys VM, VMWare Carbon Black, Okta SSO. These connectors with sample queries, dashboards and analytics will assist in collecting security data very efficiently to help detect and respond to security threats, and giving security insights in real-time as they occur. Microsoft is also expanding the quality of signals and machine learning algorithms that Insider Risk Management utilises to flag potentially risky user behaviour. New advanced signals will now be captured from Microsoft Defender Advanced Threat Protection, Windows 10 endpoints, Microsoft 365. Microsoft’s Insider Risk Management will also include new policy templates and workflows to push alerts to other systems such as ServiceNow and Microsoft Azure Sentinel. Power BI Component For React Microsoft also announced the launch of Power BI component for React, a popular open-source JavaScript library. This feature will help in more straightforward integration between Power BI and React-based web applications. Utilising the React component, enterprises can perform embedded analytics in your application. The new React component supports both JavaScript and TypeScript and will allow organisations to integrate their analytics in a React web application. The library for React allows users to embed Power BI reports and dashboards, dashboard tiles. Also, it makes it easy to perform the Power BI embed lifecycle management in React applications. The Power BI React component is now made available on NPM and is open-sourced on GitHub.","excerpt":"Microsoft’s Inspire is one of the biggest events for enterprise technology solutions. Just like the annual Build 2020 conference for developers, this year’s Inspire was a two-day virtual event from Microsoft for its partner ecosystem which builds and sells cloud-to-edge products and solutions to thousands of businesses around the world.  According to Microsoft, the tech […]","categories":["Global Tech"],"tags":["hadoop on azure cloud","latest in cloud computing technology","latest technology in cloud computing","latest technology in machine learning","Microsoft"],"author_name":"Vishal Chawla","publish_date":"2020-07-23T13:55:00","publication_year":"2020","word_count":1266,"keywords":["Go","latest in cloud computing technology","machine learning","latest technology in cloud computing","hadoop on azure cloud","AI","R","TypeScript","RAG","latest technology in machine learning","analytics","JavaScript","Azure","kubernetes","Microsoft"],"extracted_tech_keywords":["AI","machine learning","analytics","RAG","Azure","kubernetes","R","JavaScript","TypeScript","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/key-technology-takeaways-from-microsoft-inspire-2020-event\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10003668,"title":"Understanding Adaptive Optimization techniques in Deep learning","content":"Optimization, as an important part of deep learning, has attracted much attention from researchers, with the exponential growth of the amount of data. Neural networks consist of millions of parameters to handle the complexities became a challenge for researchers, these algorithms have to be more efficient to achieve better results. The functionalities of the optimization algorithm are to minimize the loss function by reaching global minima. The two important metrics to determine the efficiency of algorithms are the speed of convergence which is the process of reaching the global minima, and the generalization to new data that means how the model is performing on unseen data. Based on these two metrics researchers built the optimization algorithms. Throughout this article, we will discuss these optimization techniques with their intuition and implementation. Topics we cover in this article Understanding adaptive optimization Adagrad optimization Adadelta optimization Adam optimization Adabound optimization Understanding Adaptive optimization Optimization techniques like Gradient Descent, SGD, mini-batch Gradient Descent need to set a hyperparameter learning rate before training the model. If this learning rate doesn’t give good results, we need to change the learning rates and train the model again. In deep learning, training the model generally takes lots of time. Some researchers are fed up with setting up these learning rates. Hence they got an idea of Adaptive optimization techniques. Here, it doesn’t need to set learning rate, just we need to initialize the learning rate parameters 0.001  and these adaptive optimization algorithms keep updating learning rates while training the model. So what is the learning rate …………?  The learning rate is the most important aspect of the learning process of the model. These are steps the model takes to reach the global minima. Hands-on implementation Here we will implement the Convolutional Neural Network (CNN) model in MNIST data classification through which we will compare the optimization techniques. from keras.datasets import mnist import tensorflow from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Dense, Flatten from tensorflow.keras.layers import Conv2D, MaxPooling2D from tensorflow.keras.layers import BatchNormalization # Model configuration batch_size = 250 no_epochs = 5 no_classes = 10 validation_split = 0.2 verbosity = 1 # Load KMNIST dataset (input_train, target_train), (input_test, target_test) =mnist.load_data() # Shape of the input sets input_train_shape = input_train.shape input_test_shape = input_test.shape # Keras layer input shape input_shape = (input_train_shape[1], input_train_shape[2], 1) # Reshape the training data to include channels input_train = input_train.reshape(input_train_shape[0], input_train_shape[1], input_train_shape[2], 1) input_test = input_test.reshape(input_test_shape[0], input_test_shape[1], input_test_shape[2], 1) # Parse numbers as floats input_train = input_train.astype('float32') input_test = input_test.astype('float32') # Normalize input data input_train = input_train \/ 255 input_test = input_test \/ 255 # Create the model model = Sequential() model.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape)) model.add(BatchNormalization()) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(BatchNormalization()) model.add(Conv2D(64, kernel_size=(3, 3), activation='relu')) model.add(BatchNormalization()) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(BatchNormalization()) model.add(Flatten()) model.add(Dense(256, activation='relu')) model.add(BatchNormalization()) model.add(Dense(no_classes, activation='softmax')) Now, we will discuss the adaptive optimization techniques one-by-one and add it to the above-defined CNN model. Adagrad Adagrad works on setting the learning rate by dividing the learning rate component by the square root of the cumulative sum of the current gradient and the previous gradient. Here θ is the parameter we need to update, η is the learning rate ε is added to give non zero value, Gt is the gradient estimate at time t. Compiling the CNN model with Adagrad Optimizer # Compile the model model.compile(loss=tensorflow.keras.losses.sparse_categorical_crossentropy,optimizer=tensorflow.keras.optimizers.Adagrad( learning_rate=0.001, initial_accumulator_value=0.1, epsilon=1e-07)) # Fit data to model history = model.fit(input_train, target_train, batch_size=batch_size, epochs=no_epochs, verbose=verbosity, validation_split=validation_split) # Generate generalization metric  s score = model.evaluate(input_test, target_test, verbose=0) print(f'Test loss using Adagrad: {score[0]} \/ Test accuracy: {score[1]}') #OUTPUT Adadelta Adadelta works on exponential moving averages of the squared delta’s, here delta refers to the difference between the current weight and the newly updated weight. In Adadelta optimization technique it removes the learning rate parameter and replaces it with delta. Compiling the CNN model with Adadelta Optimizer replacing Adagrad in the above CNN model model.compile(loss=tensorflow.keras.losses.sparse_categorical_crossentropy,optimizer = tensorflow.keras.optimizers.Adadelta(learning_rate=0.001, rho=0.95, epsilon=1e-07, name=\"Adadelta\")) After updating the optimizer to Adadelta, we again trained the model. #OUTPUT Adam – Adaptive moment estimation Beginners mostly used the Adam optimization technique very popular and used in many models as an optimizer, adam is a combination of RMS prop and momentum, it uses the squared gradient to scale the learning rate parameters like RMSprop and it works similar to the momentum by adding averages of moving gradients. It computes different parameters for individual parameters. Momentum In momentum technique, instead of using only the gradient of current steps it also accumulates the gradient of the past steps to reach global minima. We use SGD with momentum to work effectively and momentum help SGD to accelerate training. Momentum term γ = 0.9 Here m and v are moving averages of the gradients and Betas only used in Adam optimization uses parameters are beta_1 = 0.9 and beta_2 =0.999 and g is the gradients in mini-batch. Compiling the CNN model with Adam Optimizer replacing in the above CNN model model.compile(loss=tensorflow.keras.losses.sparse_categorical_crossentropy,optimizer=tensorflow.keras.optimizers.Adam(learning_rate=0.01),metrics=['accuracy']) After updating the optimizer to Adam, we again trained the model #OUTPUT Adabound Adabound is an Adam variant that uses dynamic boundaries of learning rates, Adabound is as fast as Adam and as good as SGD, the main problem in adaptive techniques is they fail in convergence better because of insatiable and extreme learning rates, where the lower and upper bounds are initialized as 0 and infinity. This concept was inspired by gradient clipping the gradients larger than the threshold to avoid gradient explosion. Compiling the CNN model with Adabound Optimizer replacing in the above CNN model In the below code snippet, we are importing Adabound because in Keras optimizer’s library Adabound is not an inbuilt function, so we need to import Adabound. from keras_adabound import AdaBound model.compile(loss=tensorflow.keras.losses.sparse_categorical_crossentropy,optimizer=AdaBound(lr=1e-3, final_lr=0.1)) After updating the optimizer to Adabound, we again trained the model #OUTPUT Conclusion In this article, we discussed the adaptive optimization techniques and demonstrated the implementation. As we discussed above the best optimization algorithm will have better convergence and good generalization to new data. As we have seen the Adabound optimization introduced has a higher accuracy as compared to other optimizers, which balances the convergence and generalization.","excerpt":"Throughout this article, we will discuss these optimization techniques with their intuition and implementation.","categories":["Deep Tech"],"tags":["Deep Learning","machine learning optimization","optimization algorithms neural networks"],"author_name":"Prudhvi varma","publish_date":"2020-07-30T17:00:00","publication_year":"2020","word_count":1008,"keywords":["Go","TPU","optimization algorithms neural networks","Keras","AI","neural network","ELT","RAG","deep learning","machine learning optimization","Deep Learning","TensorFlow","R"],"extracted_tech_keywords":["AI","deep learning","neural network","TensorFlow","Keras","RAG","TPU","R","Go","ELT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/understanding-adaptive-optimization-techniques-in-deep-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10143856,"title":"Banco Sabadell Partners with Accenture, QuSecure to Adopt Post-Quantum Cryptography","content":"Spanish financial services company Banco Sabadell, in partnership with Accenture and QuSecure, has successfully completed a project to explore the adoption of post-quantum cryptography (PQC) technologies within its infrastructure. This initiative strengthens the bank’s defences against potential quantum attacks. It modernises encryption using QuSecure software and open-source libraries. The four-month project provided Banco Sabadell with a roadmap for becoming quantum-resilient and demonstrated how quantum security technologies can be implemented effectively. “This project, in collaboration with Accenture and QuSecure, has allowed us to explore the impact of the adoption of PQC technologies on the bank’s infrastructure,” said Joan Puig, Group CISO at Banco Sabadell. Cryptography plays a vital role in securing banking processes like payments and communications, which are increasingly vulnerable to emerging quantum threats. “As we transition into a post-quantum computing era, the challenges include identifying cryptographic methods vulnerable to quantum attacks and transitioning to quantum-safe cryptography with agility and efficiency,” Puig explained. Accenture brought its expertise in quantum risk and security to the project, helping Banco Sabadell develop a clear roadmap based on the latest NIST PQC standards. “This collaboration underscores our commitment to helping organisations safeguard their data from emerging threats posed by the rapid quantum computing advancements,” said Tom Patterson, emerging technology security lead at Accenture. QuSecure highlighted the ease of implementing quantum-safe technologies without overhauling existing systems. “This proves that mitigation of existing encryption to post-quantum standards for complex environments is possible and can be achieved expediently this decade,” said Elizabeth Green, SVP customers and ecosystems at QuSecure. As quantum computing advances, preparing for its impact on cryptography becomes increasingly critical. The project reflects Banco Sabadell’s proactive efforts to secure its systems and data and ensure their protection in the post-quantum era.","excerpt":"Cryptography plays a vital role in securing banking processes like payments and communications, which are increasingly vulnerable to emerging quantum threats.","categories":["AI News"],"tags":["Accenture","Banco Sabadell","cryptography","QuSecure"],"author_name":"Shalini Mondal","publish_date":"2024-12-18T15:45:55","publication_year":"2024","word_count":286,"keywords":["Accenture","API","Banco Sabadell","programming_languages:R","AI","QuSecure","cryptography","ViT","GAN","R","emerging_tech:quantum computing"],"extracted_tech_keywords":["AI","R","API","GAN","ViT","programming_languages:R","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/banco-sabadell-partners-with-accenture-qusecure-to-adopt-post-quantum-cryptography\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101455,"title":"Why OpenAI Should Acquire Graphcore","content":"AI has been a very competitive space since the end of 2022, and OpenAI has been the leader through all of it. Having championed the software side of it, now the Sam Altman-led organisation is looking to foray into the hardware industry. There has been a buzz around a report from Reuters that says OpenAI is planning to develop its own AI chips. AIM reached out to OpenAI, but the company did not respond. It is possibly busy with its first-ever DevDay conference happening on November 6, 2023. Amidst the much-discussed chip shortage in the industry, this move by OpenAI comes after several companies, such as Microsoft, Google, Amazon, and Apple announced developing their own AI chips. But even with such high demand for AI hardware, most manufacturers, except for NVIDIA, are struggling to emerge. Case in point: Graphcore. The startup, which inked a deal with Microsoft in 2019 for buying processors, is now desperately in need of funding. Graphcore turned into a unicorn in less than four years, but went on to report a revenue of $2.7 million in 2022 — a fall of 46% from 2021. Losses too have soared 11% to 204.6 million, which the company says is “due to lower hardware sales to key strategic customers”. New customer on the way One of the “key strategic customers” that Graphcore is speaking of is possibly Microsoft. Ever since Microsoft has backed OpenAI, the company has also been using NVIDIA chips for almost all of the workings, dropping the use of Graphcore chips in its cloud computing system, resulting in a decline of funds for the chipmaker. Since its last funding round in 2020, Graphcore has been valued at $2.5 billion. Since then the company hasn’t raised any funds. The UK-based company also requested the government to include chipmakers in the recent Brunel AI project, which aims at building a new AI supercomputer with a $1.1 billion fund. But, the deal hasn’t come through yet. If the government actually accepts the deal, Graphcore can rise up to its status of being an NVIDIA-rival again. On the other hand, there is OpenAI. According to a recent report by The Information, Altman has said that the company is generating revenue at a pace of $1.3 billion in a year. Now that the company is planning to make its own AI chips, it may be a great move for OpenAI to realise its dream of manufacturing its own AI chips, and actually saving Graphcore in the process. Potential @OpenAI acquisition of\/investment in @graphcoreai will completely shake up the AI chip space. But their own backer (@Microsoft ) dumped graphcores chips already so will prob feel more like a bailout at first.— DANIΞL ➪〡│〡 (@daniel0x53) October 14, 2023 OpenAI currently has the financial strength, as exemplified by its revenue growth, to make strategic moves in the AI hardware space. It has already contemplated acquiring a chipmaking firm or building its own chips. By making an early investment in Graphcore, OpenAI can not only secure a reliable supply of AI chips but also pave the way for a potentially transformative partnership, and also open up the market for other players. What about Microsoft? As mentioned earlier, Microsoft, the major OpenAI-backer, is also developing its own AI chips. According to reports, Microsoft and OpenAI employees have been testing out these chips (codenamed Athena), in secrecy since April. Furthermore, d-Matrix, an AI chip startup also raised $110 million with a backing from Microsoft. Though the company does not make direct competitors to NVIDIA’s chips, it is still a pocket friendly alternative for Microsoft for the inference portion of AI models. Anyhow, these news reports suggest that the ChatGPT creator companies may be trying to avoid the costs of NVIDIA GPUs for building AI models. But when it comes to OpenAI’s foray into the hardware market, it seems like the company’s departure from Microsoft, and indeed NVIDIA as well, in certain ways. OpenAI has plans to build its own “iPhone of AI”, and raising funds from SoftBank, according to several reports. Though not a phone, a new product is definitely in the making. So, an acquisition of Graphcore might just be the right pick for the company to host its own in-house capabilities in the near future. Meanwhile, almost all companies, namely AMD, Intel, and Cerebras, are trying to have what NVIDIA already has and are developing their own AI chips. There is no guarantee that OpenAI might actually go ahead with the deal since it is also assessing a lot of other targets.","excerpt":"OpenAI wants to make its own AI chips and Graphcore wants to make some money. What else does the partnership need?","categories":["Global Tech"],"tags":["AI Chips","Graphcore","Mergers and Acquisitions","Microsoft","OpenAI"],"author_name":"Mohit Pandey","publish_date":"2023-10-16T12:48:50","publication_year":"2023","word_count":758,"keywords":["Go","ChatGPT","OpenAI","AI","cloud computing","AI Chips","Graphcore","Aim","Ray","GPT","GAN","Mergers and Acquisitions","R","Microsoft"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","Ray","cloud computing","R","Go","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-openai-should-acquire-graphcore\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10020439,"title":"Hands-on to ReAgent: End-to-End Platform for Applied Reinforcement Learning","content":"Facebook ReAgent, previously known as Horizon is an end-to-end platform for using applied Reinforcement Learning in order to solve industrial problems. The main purpose of this framework is to make the development & experimentation of deep reinforcement algorithms fast. ReAgent is built on Python. It uses PyTorch framework for data modelling And training and TorchScript for serving. ReAgent holds different algorithms for data preprocessing, feature engineering, model training & evaluation and lastly for optimized serving. ReAgent was first presented in the research paper – Horizon: Facebook’s Open Source Applied Reinforcement Learning Platform by Jason Gauci, Edoardo Conti, Yitao Liang, Kittipat Virochsiri, Yuchen He, Zachary Kaden, Vivek Narayanan, Xiaohui Ye, Zhengxing Chen, Scott Fujimoto. The key features of Facebook’s ReAgent are : Capable of handling Large-dimension datasets.Provides optimized algorithms for data preprocessing, training, etc.Gives a highly efficient production environment for model serving. Algorithms Supported by ReAgent Discrete-Action DQNParametric-Action DQNDouble DQN, Dueling DQN, Dueling Double DQNDistributional RL: C51 and QR-DQNTwin Delayed DDPG (TD3)Soft Actor-Critic (SAC) Workflow of ReAgent The image below shows the overall workflow of ReAgent for decision making and reasoning. It starts its decision-making process by using predefined rules and then with the help of feedback, multi-armed bandits, those decisions are optimized and at last, contextual feedback trains the contextual bandits and reinforcement learning models. These trained models are then deployed via TorchScript library. Dependencies of ReAgent Platform Python >=3.7 Installation ReAgent can be installed using the docker image and manually.  In this case, we are cloning the GitHub repository and installing all the required dependencies of ReAgent via pip. %%bash git clone https:\/\/github.com\/facebookresearch\/ReAgent.git %cd \/content\/ReAgent\/ Then install the required python packages: !pip install -r requirements.txt Additionally, install: !pip install pytorch_lightning !pip install --pre torch torchvision -f https:\/\/download.pytorch.org\/whl\/nightly\/cpu\/torch_nightly.html !pip install \".[gym]\" You can check the detailed-installation process here and usage of ReAgent is discussed here. To know more about ReAgent Serving Platform(RASP), you can check this documentation. Demo – Reinforcement Learning on CartPole Problem This demo explains the usage of ReAgent on CartPole reinforcement learning. The code below uses the OpenAI Gym environment. Import all the required modules and packages. from reagent.gym.envs.gym import Gym import pandas as pd from matplotlib import pyplot as plt import seaborn as sns import numpy as np import torch import torch.nn.functional as F import tqdm.autonotebook as tqdm Define the environment by passing CartPole to the Gym class. env = Gym('CartPole-v0') def reset_env(env, seed): np.random.seed(seed) env.seed(seed) env.action_space.seed(seed) torch.manual_seed(seed) env.reset() reset_env(env, seed=0) Next, is to create a policy which contains a simple scorer or Multilayer perceptron network and softmax sampler. from reagent.net_builder.discrete_dqn.fully_connected import FullyConnected from reagent.gym.utils import build_normalizer norm = build_normalizer(env) net_builder = FullyConnected(sizes=[8], activations=[\"linear\"]) cartpole_scorer = net_builder.build_q_network( state_feature_config=None, state_normalization_data=norm['state'], output_dim=len(norm['action'].dense_normalization_parameters)) The idea behind the policy is that agents will simply execute this cart pole environment. from reagent.gym.policies.policy import Policy from reagent.gym.policies.samplers.discrete_sampler import SoftmaxActionSampler from reagent.gym.agents.agent import Agent policy = Policy(scorer=cartpole_scorer, sampler=SoftmaxActionSampler()) agent = Agent.create_for_env(env, policy) Now, create a trainer that takes reinforcement learning algorithms to train. The following trainer can be created from the commands below: from reagent.training.reinforce import ( Reinforce, ReinforceParams ) from reagent.optimizer.union import classes trainer = Reinforce(policy, ReinforceParams( gamma=0.99, optimizer=classes['Adam'](lr=5e-3, weight_decay=1e-3) )) After creating a Trainer, start the training of the model by creating a function that changes the observed transitions into a training batch and then evaluating the reward for all episodes via RL interaction loop. The code for it is shown below: import reagent.types as rlt def to_train_batch(trajectory): return rlt.PolicyGradientInput( state=rlt.FeatureData(torch.from_numpy(np.stack(trajectory.observation)).float()), action=F.one_hot(torch.from_numpy(np.stack(trajectory.action)), 2), reward=torch.tensor(trajectory.reward), log_prob=torch.tensor(trajectory.log_prob) ) Run agent on the environment and record the rewards. from reagent.gym.runners.gymrunner import evaluate_for_n_episodes eval_rewards = evaluate_for_n_episodes(100, env, agent, 500, num_processes=20) Start the loop num_episodes = 200 reward_min = 20 max_steps = 200 reward_decay = 0.8 train_rewards = [] running_reward = reward_min from reagent.gym.runners.gymrunner import run_episode with tqdm.trange(num_episodes, unit=\" epoch\") as t: for i in t: trajectory = run_episode(env, agent, max_steps=max_steps, mdp_id=i) batch = to_train_batch(trajectory) trainer.train(batch) ep_reward = trajectory.calculate_cumulative_reward(1.0) running_reward *= reward_decay running_reward += (1 - reward_decay) * ep_reward train_rewards.append(ep_reward) t.set_postfix(reward=running_reward) Finally, print all the rewards on each training episode in a form of the graph as shown below. Conclusion In this write-up, we have discussed ReAgent Platform aka Horizon and its demo with an example of CartPole Problem with reinforcement learning. Note : The codes mentioned above are not suitable for colab, due to some dependency issues. The following is the Jupyter Notebook file, to reproduce the above experiment. Official Codes, Docs & Tutorials are available at: GithubWebsiteResearch PaperOfficial TutorialFacebook Blog Horizon Blog","excerpt":"Facebook ReAgent, previously known as Horizon is an end-to-end platform for using applied Reinforcement Learning in order to solve industrial problems. The main purpose of this framework is to make the development & experimentation of deep reinforcement algorithms fast. ReAgent is built on Python. It uses PyTorch framework for data modelling  And training and TorchScript […]","categories":["Deep Tech"],"tags":["Jupyter Notebook","Python","Pytorch"],"author_name":"Aishwarya Verma","publish_date":"2021-02-18T10:00:00","publication_year":"2021","word_count":738,"keywords":["Pytorch","NumPy","Jupyter Notebook","OpenAI","AI","PyTorch","ML","Python","Ray","Colab","Matplotlib","Jupyter","Pandas"],"extracted_tech_keywords":["AI","ML","OpenAI","Ray","PyTorch","Jupyter","Colab","Pandas","NumPy","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-to-reagent-end-to-end-platform-for-applied-reinforcement-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10054298,"title":"Rasa Releases Open Source 3.0 To Help Build Better Conversational AI","content":"Rasa recently announced their new major release Rasa Open Source 3.0, to help build better conversational AI of the future. It separates the model architecture from the framework architecture, enabling developers to run arbitrary model architectures. It also comes with several enhancements focused on improving the developer experience when building conversational AI assistants with Rasa. The revamped computational backend empowers to experiment with architectures, reducing maintenance costs and enabling collaborative development at scale. There have also been improvements to slot mappings that will make it easier to implement desired slot behaviour as well as forms. A new experimental feature known as ‘Markers’ has also been introduced, which is intended to help figure out how to add a “semantic layer” on top of the tracker store of events that makes it easier to identify and track situations of interest in conversations. The new graph architecture aims to make it easier to understand the relationship between the NLU and policy components in the pipeline. It is now much easier to define and modify the dependencies between the training pipeline components. The beauty of graph architecture is that it makes it possible to save trained components on disc. This means that if a change is made to a specific component, only that component will need to be retrained. This should save a lot of computational resources and reduce training time. Image Source: Rasa In the past, if you had an entity and a slot defined with the same name, Rasa would automatically fill the slot with the value of the extracted entity. While this sometimes saves a little bit of development time, it has often led to undesired behaviours (slots being filled in when they shouldn’t have been) and confusion when implementing forms with slot mappings. With Rasa Open Source 3.0, this behaviour has been updated. From now on, it will be necessary to define global slot mappings for all slots defined in a domain file. Those mappings will have to be defined inside of the slots section of your domain. Markers are conditions that allow you to describe and mark points of interest in the dialogue for evaluating your assistant. With markers, developers will be able to describe specific points, like when an action has been executed, the intent has been classified correctly, or a slot has been set. With every Rasa Open Source release, Rasa aims to make it easier for developers to build conversational AI assistants. The team says that the new features and improvements were crucial to making sure that the changes made tackle the most pressing needs of the developer community.","excerpt":"The new update comes with several enhancements focused on improving the developer experience when building conversational AI assistants with Rasa.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI and NLP","Conversational AI platform","Data Science","Deep Learning","Machine Learning","NLP","NLP AI","nlp in data","NLP models","open source project"],"author_name":"Victor Dey","publish_date":"2021-11-26T13:42:42","publication_year":"2021","word_count":433,"keywords":["AI assistants","programming_languages:R","Conversational AI platform","AI","open source project","AI and NLP","Machine Learning","NLP models","NLP","Aim","Deep Learning","Data Science","R","nlp in data","AI (Artificial Intelligence)","NLP AI"],"extracted_tech_keywords":["AI","Aim","AI assistants","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/rasa-releases-open-source-3-0-to-help-build-better-conversational-ai\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10040381,"title":"Top Talks to Look Forward to At The Rising 2021","content":"It’s that time of the year again for the much awaited Rising event. Now in its third year, Rising is organised by Analytics India Magazine and celebrates women in data science, IT, industry and academia. Rising 2021 is a two-day virtual conference scheduled to be held on 21 & 22 May 2021. The conference will provide a platform to meet and support women developers and professionals in the field of data science.  Leading women visionaries in the fields of AI and ML will be taking the audience on an insightful journey in this 2-day long event. Rising 2021 has a strong line up of speakers. Here are a few sessions that you should not miss: 1| AI in Agriculture When: 21 May, 12:35 – 13:05 By: Roli Jindal, Co-Founder, RMSI Cropalytics With more than 70% of the rural population depending solely on agriculture, falling average landholding size and minimal access to modern technology – the sector is facing a stressful situation. AI can be the way forward from prediction of weather to amount of inputs, identifying crop disease on time and many other valuable insights can ensure a shift in momentum for the sector. Roli Jindal of Cropalytics will be speaking about the same at Rising. Roli is the Co-Founder at RMSI Cropalytics and she is going to speak on how AI can be used in agriculture to minimise agrarian distress. RMSI Cropalytics specialises in crop yield measurement with AI\/ML, remote sensing, advanced agri-modeling, and meteorological domain expertise. The experts have programmed machines to analyse satellite imagery, classify crops, and forecast crop yields in every visible farm. 2| Exploiting technologies to drive business outcomes When: 21 May, 14:35 – 15:05 By: Madhurima Agarwal, Director – Engineering Programs, NetApp India; Leader – NetApp Excellerator Businesses, especially in the post pandemic world, require a greater degree of technology adoption to further propel their growth. With automation and digitisation witnessing a wide scale adoption, technology driven businesses are the way to compete not just locally, but on a global scale too. In her talk, Madhurima will be exploring how technologies can be exploited to drive business outcomes. Madhurima is the Director – Engineering Programs at NetApp India & Leader – NetApp Excellerator. She is going to speak on how in the coming years, business performance will be determined by the technology investments made. Businesses can gain agility, scale, and productivity by leveraging cloud, SaaS solutions, IoT, Data, and automation. The discussion will provide us the need and gain that can be made from the same. 3| Demystifying MLOps – What is MLOps and Why is this a hot emerging topic? When: 21 May, 19:25 – 19:55 By: Hamsa Buvaraghan, Smart Analytics and AI Platform Solutions Manager at Google MLOps is the buzzword of 2021. Everyone wants to know about it. MLOps enables automation of ML pipelines. The techniques and tools make productionising easier than ever before. Successful implementations and efficient operations have emerged as the main bottleneck to getting immediate value from Machine Learning systems as nearly all industries embrace it at a rapid rate. Organisations are betting big on MLOps and you wouldn’t want to miss this talk by Hamsa B if you want to know more about MLOps. Hamsa Buvaraghan is the Smart Analytics and AI Platform Solutions Manager at Google and she will be demystifying MLOps for all of us at the Rising event. 4| AI in Racket Sports When: 21 May, 16:30 – 17:00 By: Megha Gambhir Co-Founder & CEO, Stupa Sports Analytics Sports is one of the emerging fields and sports persons nowadays are much more concerned about their game, gaming style, skill-sets, fitness and are very much keen to enhance their results. Data-driven insights can help athletes improve their performance and step out with the best possible decision making. Megha Gambhir is the Co-Founder & CEO at Stupa Sports Analytics and at Rising she is going to talk on how the use of AI has impacted the world of sports. Both data analytics and artificial intelligence are used extensively in sports to improve audience interaction and spectator experiences through graphics, stats, replays, AR\/VR, game strategy, coaching assistance through replay time analysis, and Match predictions, Player predictions, and so on. 5| The Healthcare Industry and Artificial Intelligence (AI) When: 21 May, 17:05 – 17:35 By: Mitali Dutta, Head Data Science and AI&ML – Enterprise IT, Philips As the pandemic hit, one of the hardest hit sectors was healthcare. Post-pandemic requires a robust health sector with minimal human-to-human interaction. Artificial intelligence is the use of machine-learning algorithms and software to imitate human cognition in the study, presentation, and comprehension of complex medical and health-care data. Mitali Dutta is the Head of Data Science and AI&ML – Enterprise IT at Philips and she is going to speak on the impact of AI in healthcare. At a high level she will cover what AI emphasizes on healthcare. It will briefly talk about the challenges of leveraging AI in healthcare. Finally there will be a discussion around how the industry can adapt to AI. 6| How AI in Logistics helped during the pandemic When: 21 May, 18:15 – 18:45 By: Manisha Raisinghani, Co-founder and CTO, LogiNext Logistics and supply chain can also be termed as the backbone for the industry to survive in the future to come. Adoption of AI and technology for the sector will result in cost reduction, enhanced efficiency and is the potential answer for the safety concerns. Moreover, it’s high time for the adoption of high-end technology for route optimisation and quick decision making. Manisha Raisinghani, co-founder and CTO at LogiNext will be talking on how during the pandemic and in the future, automation across the entire global supply chain is critical. LogiNext itself assists businesses in digitising and optimising their supply chains and logistics. Being from the industry itself, this conversation will help learn more about how AI and machine learning are transforming the face of logistics. 7| Teaching analytics for teacher data literacy When: 21 May, 18:50 – 19:20 By: Surabhi Goel, CEO, Aditya Birla World Academy Every student is important and the learnings we provide to our students in the schools will be the key factor for the futuristic growth. Teachers need to equip themselves with the best possible skill-sets, understandability about their students, and even the teaching style. Surabhi Goel is the CEO at “Aditya Birla World Academy, The Aditya Birla Integrated School & Aditya Birla Education Academy,” and she will discuss how in the education sector, data science and analytics have become indispensable. Student results are directly linked to teacher data literacy. Teachers today must regularly capture, analyse, and interpret data while teaching in the classroom. This data may also help a teacher better her or his teaching methods and therefore the results. 8| The Importance of introducing AI to young students When: 22 May, 09:15 – 09:45 By: Pranjali Awasthi, Machine Learning Researcher Pranjali Awasthi is a Machine Learning researcher and at Rising, she will be talking about the early adoption of AI for young students. Whether it’s incredibly low-cost diagnostic systems or automating business processes that have been manual for years, AI and superior data analytics methods are revolutionising every industry. This explains why it is important to begin teaching analytics and AI to children at a young age. More specifically, educating women in AI has the potential to increase female involvement in technology by a large number of people, as AI has applications in a wide range of fields, from politics to fashion. 9| Career path of an AI business leader When: 22 May, 19:25 – 19:55 By: Beena Ammanath, Executive Director, Deloitte AI Institute & Founder, Humans For AI AI jobs are the hottest in the town right now. The resources are so abundant that it is possible to feel lost in this ever growing field. Beena Ammanath is here to help us. Beena is an Executive Director at Deloitte AI Institute; Founder at Humans For AI. At Rising, she is going to discuss the education\/training background. She will be addressing most pressing questions of our time: Key traits needed for success in AI?How do you build an AI business? What are the challenges? How has it been being a woman in AI? Why is it important to have more women in AI? And more. 10| Workshop on Medical Image Classification and Segmentation using Deep Learning with Tensorflow2 When: 22 May, 12:25 – 15:30 By: Anup Kumar, Tech Lead & Soniya Singhal, Staff Engineer, Stryker Anup Kumar is the Tech Lead and Soniya Singhal is the Staff Engineer at Stryker and in this virtual workshop, Anup and Soniya will be walking us through ML implementation for medical use cases.  They will be using the TensorFlow framework for this session and will be quite useful for those who want to get hands-on experience.","excerpt":"It’s that time of the year again for the much awaited Rising event. Now in its third year, Rising is organised by Analytics India Magazine and celebrates women in data science, IT, industry and academia. Rising 2021 is a two-day virtual conference scheduled to be held on 21 & 22 May 2021. The conference will […]","categories":["Deep Tech"],"tags":["Women in Data Science"],"author_name":"kumar Gandharv","publish_date":"2021-05-18T19:00:00","publication_year":"2021","word_count":1473,"keywords":["data science","Women in Data Science","artificial intelligence","machine learning","AI","ML","MLOps","RAG","deep learning","analytics","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","MLOps","TensorFlow","RAG"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/top-talks-to-look-forward-to-at-the-rising-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":29203,"title":"Brain-Inspired Cognitive Architecture Is Now Solving Computational Challenges Faced By AI","content":"With the development of artificial intelligence intensifying across the globe, IT companies are looking for ways to revamp their architecture to make more robust. Increasingly, researchers are turning to brain-inspired architecture with co-located memory and processing, resulting in computers which are 200 times faster than conventional computers. Such is the excitement around AI hardware, that this phase has been dubbed as a “renaissance of hardware” as vendors are rushing to build domain-specific or workload-specific architectures that can significantly scale and improve computational efficiency. And as we nudge forward in the mobile era, the workloads are going to look extremely dissimilar since the requirements of computing are changing. Businesses have to rely on a different architecture, each meant for a particular workload. This is where vendors are making a shift from Von Neumann computing architecture and are striving to improve the performance of computing with multi-core CPU architectures. Now, the rapid gains in neuroscience have also spurred researchers across the globe to propose Brain-inspired computing architecture to develop highly advanced cognitive systems. IBM researchers are working on a new computer architecture which can process data efficiently for AI workloads. However, what’s remarkable is that this new architecture is inspired by the brain and will feature coexisting processing and memory units. The IBM team argues that traditional computers were built on Von Neumann architecture, developed in the 1940s and featured a central processor that executes logic and arithmetic, a memory unit, storage, and input and output devices. But the current industry requirement has necessitated a move from homogeneous to heterogeneous computing architecture which has led to an increased research in applied materials and neuroscience. Why Does AI Workload Require New Computing Architecture? So what kind of changes are required in computing architecture for AI workload? According to researchers, there have to be consistent, ground-breaking breakthroughs in material sciences and neuroscience to advance AI processing. Let’s look at the two key requirements for AI workloads which have arisen an interest in brain-inspired computer architecture. Technologists emphasise that multi-core CPUs have reached their performance and efficiency limit and are adding to architectural challenges. Memory requirement: Since AI workloads depend on a lot of data — for processing huge amount of data, AI workloads require faster access to memory. In traditional CPUs have multi-level cache architecture which is not well-suited for AI. Parallel processing requirement: Parallel computing is on a rise and AI workloads and architectures have to be designed which can execute parallelism at scale. Brain-Inspired Computing Architecture There has been a rise in the development of cognitive which researchers assert emulates brain architecture modelling, cognitive architecture design and cognitive architecture fostering (by making them learn in a certain environment) and application to products. IBM’s new paper discusses the three layers of inspiration from the human brain. Firstly, the team took inspiration from the brain’s memory and processing, emulating a memory device to perform computational tasks in the memory itself. The second feature was inspired by the brain’s synaptic network structures and developed as arrays of phase change memory (PCM) devices to speed up the training process for deep neural networks. And finally, the researchers drew on the stochastic nature of neurons and synapses to develop a powerful computational substrate for spiking neural networks. Speaking about testing the systems, Abu Sebastian from IBM said, “These systems are expected to be better than conventional computing systems in some tasks, and they also surpass traditional systems in terms of efficiency.” In an experiment where the researchers ran an unsupervised machine learning algorithm on the computational memory platform, they found the brain-inspired memory platform to be 200x faster in performance than conventional computing systems. Conclusion Given the pace at which knowledge of neuroscience is rapidly increasing, thanks to frenetic research, there have been attempts to correlate AI with brain’s cognitive architecture. Researchers cite that constructing AI systems is based on the hypothesis that it is possible to build a general-purpose intelligent machine which can replicate human-level intelligence. A key aspect of the brain actively researched by neuroscientists is how deep learning in a way appears to replicate the cerebral neocortex which plays an important role. There is also research on the connectome, which forms the cognitive architecture of the brain and plays a role in advancing several breakthroughs in neuroscience. Experts cite that the neuron model, now widely known to be used in the artificial neural network has lots of functions despite being a simple internal structure.","excerpt":"With the development of artificial intelligence intensifying across the globe, IT companies are looking for ways to revamp their architecture to make more robust. Increasingly, researchers are turning to brain-inspired architecture with co-located memory and processing, resulting in computers which are 200 times faster than conventional computers. Such is the excitement around AI hardware, that […]","categories":["AI Features"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-10-12T12:35:52","publication_year":"2018","word_count":740,"keywords":["Go","machine learning","artificial intelligence","TPU","AI","neural network","RAG","Ray","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","Ray","RAG","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/brain-inspired-cognitive-architecture-is-now-solving-computational-challenges-faced-by-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10164990,"title":"UBTECH Advances Humanoid Robotics with Swarm Intelligence Training at Zeekr","content":"Shenzhen-based UBTECH Robotics has successfully completed what is claimed to be the world’s first multi-humanoid robot collaborative training program at Zeekr’s 5G Intelligent Factory. This marks a significant advancement in industrial automation, demonstrating the application of swarm intelligence for humanoid robots in multi-task, multi-scenario environments. The initiative involved UBTECH’s Walker S1 humanoid robots working collaboratively across various production zones, including assembly workshops and quality inspection areas, as the company announced on LinkedIn. The robots executed tasks such as sorting, handling, and precision assembly, showcasing their ability to coordinate seamlessly in real-world industrial settings. UBTECH announced that this development transitions its robots from single-agent autonomy to swarm intelligence. Advancing Swarm Intelligence in Robotics UBTECH has developed BrainNet, a software framework enabling humanoid robots to collaborate effectively. It integrates cloud-device inference nodes and skill nodes, forming a centralised “super brain” for complex decision-making and an “intelligent sub-brain” for distributed control. Powered by a large reasoning multimodal model, the system allows robots to autonomously schedule and coordinate tasks. According to UBTECH, this innovation is supported by the Internet of Humanoids (IoH), which acts as a control hub. It enables robots to adapt to dynamic environments, optimise workflows, and improve task execution accuracy. Practical Training 2.0: Multi-Robot Collaboration At Zeekr’s factory, dozens of Walker S1 robots participated in training that included activities like collaborative sorting, handling and precision assembly. Sorting uses vision-based perception and hybrid decision-making for efficient dynamic task allocation, while handling challenges like uneven load distribution through advanced path planning and adaptable control. Additionally, precision assembly employs high-precision sensing to manipulate deformable materials without damage. These capabilities are underpinned by UBTECH’s multimodal reasoning model, which is trained on industrial datasets from automotive factories. This model leverages retrieval-augmented generation (RAG) technology to enhance decision-making and scalability. Expanding Industrial Applications UBTECH collaborates with leading automakers such as Geely Auto, BYD, and Audi FAW to deploy its Walker S series robots globally. The company plans to expand its Practical Training 2.0 program to more partner factories, accelerating the adoption of humanoid robots in intelligent manufacturing. “Swarm Intelligence represents the next frontier in robotics,” UBTECH stated. “Our innovations pave the way for scalable deployment in complex industrial workflows.”","excerpt":"The initiative involves UBTECH’s Walker S1 humanoid robots executing tasks like sorting, handling, and precision assembly.","categories":["AI News"],"tags":["AI in china","Humanoid Robots","Robotics","swarm","ubtech"],"author_name":"Sanjana Gupta","publish_date":"2025-03-03T16:19:47","publication_year":"2025","word_count":362,"keywords":["AI in china","Humanoid Robots","ubtech","AI","Modal","innovation","ML","Scala","Robotics","RAG","automation","Aim","ViT","swarm","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Scala","ViT","automation","innovation","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ubtech-advances-humanoid-robotics-with-swarm-intelligence-training-at-zeekr\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171368,"title":"Mistral Rolls Out AI Coding Assistant to Challenge Claude and Copilot","content":"Mistral AI has launched Mistral Code, a new AI-powered coding assistant built specifically for enterprises, offering deployment flexibility and full-stack control. The product is open for private beta for JetBrains IDEs and VSCode, with general availability expected soon. Mistral Code bundles intelligent code assistance, local and cloud deployment options, and enterprise-grade tools into a single package. Unlike typical SaaS-based copilots, the system is designed to run entirely within an enterprise’s secure infrastructure, on cloud, reserved capacity, or utilising on-premises GPUs. “Every line of code resides inside the customer’s enterprise boundary,” the company said in its announcement. At its core, Mistral Code integrates four models: Codestral for autocomplete, Codestral Embed for code retrieval, Devstral for agentic development tasks, and Mistral Medium for chat-based help. Teams can fine-tune these models or distil them into lightweight variants, a feature Mistral claims is unmatched by competitors tied to closed APIs. Sophia Yang, head of developer relations at Mistral AI, highlighted the platform’s flexibility and deep integration on an X post, calling it “the most customisable AI-powered coding assistant for enterprises”. She pointed to its ability to automate code generation, debugging, documentation, and even migration tasks, without compromising visibility and compliance. Early adopters include Abanca, a popular bank in Spain and Portugal, SNCF, France’s national railway company and Capgemini, each deploying Mistral Code across thousands of developers under hybrid or on-prem setups. With built-in observability, role-based access, and 24\/7 support under one SLA, Mistral positions its platform as a single-vendor alternative to fragmented AI dev tools that stall at proof-of-concept. The product builds on the open source Continue project, but adds features like audit logging, seat management, and agentic workflows that allow AI to handle complete software tickets, not just suggest lines of code.","excerpt":"“Every line of code resides inside the customer’s enterprise boundary.”","categories":["AI News"],"tags":["coding","mistral"],"author_name":"Ankush Das","publish_date":"2025-06-05T11:46:37","publication_year":"2025","word_count":289,"keywords":["API","programming_languages:R","AI","coding","agentic workflows","RAG","Aim","llm_models:Gemini","mistral","copilots","R"],"extracted_tech_keywords":["AI","agentic workflows","Aim","RAG","copilots","R","API","llm_models:Gemini","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mistral-rolls-out-ai-coding-assistant-to-challenge-claude-and-copilot\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085894,"title":"Microsoft Gives a New Lease of Life to Healthcare via ChatGPT","content":"The healthcare sector has been grappling with the utility of AI for quite a while now. However, since the days of IBM Watson, Big Tech has more often than not stumbled upon the healthcare door and the recent developments in generative AI may be thawing the ice further. A new report published by authors from McKinsey and Harvard certainly mirrors this enthusiasm. The report posits that the employment of different branches of AI like ML and NLP could save the healthcare system between USD 200 billion and USD 360 billion. This is aside from improving the quality of patient experience and expanding the access to healthcare facilities. While the big fat numbers already have experts salivating, the initial problems with the adoption of the tech are yet to be dealt with. There still exists a silo there. Azure’s investments in health But Big Tech hasn’t given up. Earlier this month, Microsoft announced a partnership with AI healthcare startup, ‘Paige’. Launched in 2018, the startup was working on cancer diagnostics and pathology. Paige does a lot on its own—its tools have advanced diagnostics in labs that have computational pathology which is leveraged to recognise complex tissue patterns. It also has predictive tools that offer new methods to identify and then study tissue biomarkers. Paige was growing seamlessly in a complex sector—in 2020, it picked up another USD 20 million from Goldman Sachs and Healthcare Venture Partners, closing out its Series B funding at an impressive USD 70 million. Andy Moye, CEO of Paige, addressed the announcement in a blog post: “In Microsoft, we’ve really found a partner that shares our vision in how healthcare is going to be transformed. . . For us, the vision we talked to Microsoft about is, how do we help create the digitization of pathology? How do we ensure that these tools are being used to get better patient care, to get better patient outcomes?” To be more specific, Paige will be adopting Microsoft Azure as the cloud partner for their platform. Azure has a robust offering in healthcare and has been making headway in the sector with a bunch of healthcare-centric  tools like Azure AI and Machine Learning, Azure Synapse Analytics, Azure Healthcare APIs, Azure Stack and Mixed Reality. ‼️Healthcare AI builders‼️ @Microsoft Azure @OpenAI Service is now HIPAA ready (+ includes support for zero-shot or specialized model creation via fine tuning on your own data)!https:\/\/t.co\/hoZcmKmH3vCheck out this content for tips: https:\/\/t.co\/jlzAmPgqr7 pic.twitter.com\/cWnQ3y3fZb— Matthew Lungren MD MPH (@mattlungrenMD) January 21, 2023 OpenAI Azure compliance change There is another vital regulatory breakthrough that Azure achieved with a simple update. Azure OpenAI Service is now Health Insurance Portability and Accountability Act, or HIPAA, compliant which will help healthcare companies use generative AI to determine medical outcomes better. But why is HIPAA compliance so vital in healthcare? Well, first off, it is simply the law of the land. In US healthcare, every medical software must be HIPAA compliant. The main objective under HIPAA is to protect sensitive medical data of patients from potential breaches. ChatGPT listed as an author in a medRxiv research paper The change has opened the door for companies working with Azure to use LLMs in areas within the medical context, ranging from diagnosis and treatment to patient engagement. For instance, OpenAI’s models GPT and Whisper can remove the need for administrative tasks such as medical note-taking and documentation which normally consume up to 50% of doctors’ time, often causing them to burnout. Diagnosing using LLMs which are trained on medical literature and health records can help doctors with making decisions faster. A month ago, Google introduced ‘MedPaLM’, an open-source LLM that was able to answer medical questions just as well as clinicians could, according to the new MultiMedQA benchmark. LLMs can analyse a lot of patient data to predict how likely a patient is to develop a certain disease or a condition in the future. They can also classify patients and direct them immediately to the appropriate level of care. LLMs can also sift through unstructured data, which includes clinical notes and medical journals, to better medical research. How HoloLens 2 plans to enhance patient treatment, Source: Microsoft Besides, Nadella’s company has also invested in developing hardware tools, such as HoloLens, which also have genuine potential applications in the future of care facilitation. While ChatGPT has been getting the cold shoulder from the coders and the educational community, the healthcare industry is growing increasingly familiar with chatbots. The bot recently passed the US medical licensing exam while also getting credited as an author in scientific papers. Generative AI tools like ChatGPT have often been accused of being ineffectual in core sectors. But it is time that an industry as indispensable as healthcare also reaped its benefits.","excerpt":"Earlier this month, Microsoft announced a partnership with AI healthcare startup Paige.","categories":["Global Tech"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2023-01-25T18:35:00","publication_year":"2023","word_count":792,"keywords":["ChatGPT","machine learning","OpenAI","AI","chatbots","ML","RAG","NLP","analytics","generative AI"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","analytics","generative AI","ChatGPT","OpenAI","RAG","chatbots"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-gives-a-new-lease-of-life-to-healthcare-via-chatgpt\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":25802,"title":"AI Recreates Chemistry’s Periodic Table Of Elements In A Matter Of Hours","content":"It is a truth universally acknowledged that an artificial intelligence-based system can perform calculations and other data-based analyses faster than a human. But when an AI creates something in hours, which took humans hundreds of years to discover, is a new feat altogether. Now, an AI programme called Atom2Vec, developed by Stanford physicists has successfully learned to distinguish between different atoms after analysing a list of chemical compound names from an online database. The programme uses concepts from natural language processing to cluster the elements together according to their chemical properties. Shou-Cheng Zhang, a physics professor at Stanford’s School of Humanities and Sciences said, “We wanted to know whether an AI can be smart enough to discover the periodic table on its own, and our team showed that it can.” Professor Zhang further explained that they modelled Atom2Vec on a Google programme called Word2Vec. Here, the AI works by converting words into numerical codes or vectors. By analysing these vectors, the AI programme can figure out the probability of a word appearing in a text given, in relation to the co-occurrence of other words. Zhang explained with a simple example: The word “king” is often accompanied by “queen,” and “man” by “woman.” Therefore, the mathematical vector of “king” might be translated as: king = queen – woman + man “We can apply the same idea to atoms — instead of feeding in all of the words and sentences from a collection of texts, we fed Atom2Vec all the known chemical compounds. From this data, the AI program figured out that potassium (K) and sodium (Na) must have similar properties because both elements can bind with chlorine (Cl). Just as the king and queen are similar, potassium and sodium are similar too,” said Zhang. It took nearly a century for human scientists to organize the periodic table of elements into its current form, but a new artificial intelligence (AI) program developed by Stanford physicists accomplished the feat in just a few hours. https:\/\/t.co\/efcsobucU6 — Stanford University (@Stanford) June 26, 2018","excerpt":"It is a truth universally acknowledged that an artificial intelligence-based system can perform calculations and other data-based analyses faster than a human. But when an AI creates something in hours, which took humans hundreds of years to discover, is a new feat altogether. Now, an AI programme called Atom2Vec, developed by Stanford physicists has successfully […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","stanford university"],"author_name":"Prajakta Hebbar","publish_date":"2018-06-27T06:25:23","publication_year":"2018","word_count":339,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","stanford university","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-recreates-chemistry-periodic-table-stanford-atom2vec\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10020456,"title":"Things To Remember While Using Demographic Data In ML Models","content":"Machine learning algorithms improve the accuracy of human decision-making by leaps and bounds. Such algorithms take multiple parameters into account to conduct analysis and come to a decision. Cognitive biases influence human decisions. And when fallible humans build machine learning models, biases creep into algorithms. The decisions made by biased machines can have far-reaching consequences. The repercussions are more severe when models use demographic data like age, gender, race, or zip codes since they can impact communities as a whole. In this article, we try to analyse the use of demographic data when building ML models to produce fair AI. Pitfalls Machine learning is used to develop decision-making systems across sectors. Models could help with diagnosis in healthcare, perform market segmentation in retail, or build recidivism models to reduce crime. In some cases, demographic data is essential, especially when building diagnosis or prediction models in healthcare. For many illnesses, age, gender or socio-demographic factors like income or neighbourhood become crucial decision-making parameters. For instance, age is a risk factor for many diseases, including cancer or cardiovascular conditions. Gender can play an important role in obesity disphormism or coronary artery diseases. Economically weaker neighbourhoods are at a higher risk of infectious diseases like dengue or tuberculosis. However, introducing demographic characters has led to discrimination against people, predominantly minority or socio-economically weaker communities. For instance, a recidivism model used in the US consistently put blacks at a higher risk than white people in facing the heat of law, even when the formers’ crimes were significantly less severe. In another instance, Amazon’s recruitment model did not rate candidates in a gender-neutral way as the model was trained on resumes mostly from men. This resulted in the system penalising resumes with the word women in them. The discrimination engenders from introducing demographic details in models, reflecting the inherent bias in human beings. From an ethical perspective, using demographic information to make decisions, like assigning recidivism scores based on race or allocating a bank loan, is prejudicial. Biases snuck in from incorporating demographic details without the machine learning developer’s knowledge like in Amazon’s recruitment model. In that case, developers should take extra caution while deploying such an algorithm in the real world. Third-party audits should be compulsory for any algorithms that make decisions for human beings. Handle With Care On the flip side, some machine learning models have shown the need to use inclusive demographic representation to mitigate bias. For instance, Timnit Gebru, the AI ethicist who recently got fired from Google, published a paper in 2018 that found significant disparities in facial recognition systems developed by the Big Tech. Her study revealed that all classifiers in these models performed the best for lighter male individuals but the worst for dark women. Flawed facial recognition system models have led to a Black US citizen wrongly arrested due to misidentification. Whether algorithms like facial recognition should be deployed in the first place is out of this article’s scope, the study showed that algorithm development needed more inclusivity and analysis on features specific to demography; racial traits in this case. Inclusive demographic data could help mitigate bias, but the decision as to when and how to use them for bias mitigation is critical. Partnership on AI addressed such concerns in a report in 2020. The first concern is, how should demographic data be defined. While the US and the EU have taken an effort to categorise demographic data as ‘protected class data’ or ‘sensitive personal data’, many countries, including India, have weak data protection laws. In such a case, collecting demographic data might do more harm than good. Further, the decision-makers should be careful that their approach to mitigate bias is not itself biased. For instance, self-selection bias (collecting data from only those who want to give it to you) can compound the problem. Lastly, once the data is collected, it is essential to ensure that it is used towards the original objective. Wrapping Up Some models present the absolute need for demographic details, especially in healthcare. In such cases, extra caution should be applied to mitigate biases. Further, we need stricter policies and regulations to enable the fair use of demographic data to build ML models or mitigate biases in existing models.","excerpt":"Machine learning algorithms improve the accuracy of human decision-making by leaps and bounds. Such algorithms take multiple parameters into account to conduct analysis and come to a decision. Cognitive biases influence human decisions. And when fallible humans build machine learning models, biases creep into algorithms. The decisions made by biased machines can have far-reaching consequences. […]","categories":["AI Features"],"tags":["deploying models"],"author_name":"Kashyap Raibagi","publish_date":"2021-02-18T13:00:00","publication_year":"2021","word_count":706,"keywords":["Go","machine learning","deploying models","AWS","AI","cloud_platforms:AWS","programming_languages:R","ML","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","machine learning","ML","AWS","R","Go","ViT","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/things-to-remember-while-using-demographic-data-in-ml-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":31012,"title":"5 Noteworthy AI And IoT Startups That Are Providing Smart Solutions To Indian Farmers","content":"Agriculture is the most important sector in the Indian economy with the country being pegged as a global agricultural powerhouse. According to recent statistics, the country has 195 million hectors under cultivation of which only 63 percent is rainfed while 37 percent is irrigated. Despite several advancements, India’s agriculture sector is fraught with several challenges such as water shortage, delay in weather prediction or dealing with weed, among others. In this article, we will list down how robotics and drones are being leveraged to help farmers and change the face of the agricultural sector. Spoors Founded in 2013 and based in Hyderabad, Spoors has developed a toolkit called EFFORT (effortless field force optimising and reporting toolkit). Built with cutting-edge tools, it is an innovative and complete field force management technology which unfolds the planning and records of the workforce activities across diverse fields. With advanced workflow prototype, the tool provides scope to fully automate the field business approval procedure in a single tier structure. The entire technology is built on multiple platforms that deliver technological promotions in field service management. QTLomics Headquartered in Bengaluru, QTLomics integrates the breeding process with modern genomic tools to organise the hybridisation of desirable characteristics in the crop. Latest technological developments in DNA progressing and tools for bioinformatics are assisting plant biologists to gain genomic data from prototype plants and agricultural species which help in the understanding of plant metabolism, natural genetic variations and the procedure of operation selection on plant genomes. The company uses manipulated plant growth and metabolism through MAS and gives hope to plant breeders to develop appropriate kind of plants and renewable energy sources that are subject to current climatic conditions and geography. Infratab Bengaluru-based Infratab is an IoT company which provides quality monitoring technology for perishable items. They offer integrated technology of sensors, software and analytics to assist customers in making superior decisions in the supervision of the supply chain to minimise spoilage, protect and boost their brand identity. Cropin This startup is known for developing the “full-stack AgriTech” which delivers SaaS solutions to agribusinesses across the globe. The startup provides advanced solutions like big data analysis, AI, machine learning and remote sensing to help clients understand and interpret data to collect real-time activities on existing crops. The company focuses on digitalisation, yielding, predictability and durability of the agriculture industries. Intello Labs This Bengaluru-based AgriTech startup provides agriculture product grading through user quality analysis of images of food products in an accurate and reliable way to grade fresh products. The other technology provides alerts on crop infestation which uses the farmers clicked photographs to understand the pests, diseases and weeds that are growing in their farms. The company leverages advanced analytic tools and methodologies such as deep learning, artificial intelligence and computer vision, IoT and big data to build product-based solutions for business clients.","excerpt":"Agriculture is the most important sector in the Indian economy with the country being pegged as a global agricultural powerhouse. According to recent statistics, the country has 195 million hectors under cultivation of which only 63 percent is rainfed while 37 percent is irrigated. Despite several advancements, India’s agriculture sector is fraught with several challenges […]","categories":["AI Trends"],"tags":["AI and IoT","Bengaluru","latest ai products across industries"],"author_name":"Bharat Adibhatla","publish_date":"2018-12-04T06:08:45","publication_year":"2018","word_count":474,"keywords":["big data","artificial intelligence","machine learning","AI","AI and IoT","Git","computer vision","RAG","deep learning","analytics","latest ai products across industries","Bengaluru","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","computer vision","analytics","RAG","R","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-noteworthy-ai-and-iot-startups-that-are-providing-smart-solutions-to-indian-farmers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24285,"title":"Meet JioInteract, Reliance&#8217;s New AI-Based Brand Engagement Platform","content":"Reliance Industries Limited announced the launch of Jiointeract on Thursday, in what the company is calling as the “world’s first artificial intelligence-based brand-engagement video platform.” The official statement read, “The first of many services to be launched on this platform is the Live Video Call that features India’s favourite celebrities. To kick-start, Jio has on-boarded none other than Bollywood’s biggest star, Amitabh Bachchan, who will promote his upcoming comedy-drama film 102 Not Out in the most innovative way.” The company added that as Jio’s base of over 186 million subscribers and another 150 million smartphone users was ready, JioInteract was to become the largest platform for movie-promotion and brand engagement. “Over the next few weeks, Jio will introduce services such as video call centres, video catalogue, and virtual showrooms to the forefront redefining customer experience,” said the statement. Have a question for @SrBachchan ? Then get it answered by the Shehenshah of Bollywood himself on #JioInteract – world’s first AI based live video call! Download #MyJioApp – https:\/\/t.co\/T1szBZYmYh pic.twitter.com\/YQaAKY1HAK — Reliance Jio (@reliancejio) May 4, 2018 JioInteract’s first service Live Video Call will allow all Jio and other smartphone subscribers to make a video call to Bachchan anytime during the day, starting 4 May 2018. Users can even ask questions related to his upcoming film, 102 Not Out and book their movie tickets in real-time through JioInteract’s ticketing partner BookMyShow. Such use of artificial intelligence is a first across the world and will reposition the way brands and customers think of engagement “This unique and innovative service uses a powerful artificial intelligence based platform to listen to user questions and respond to them in the most appropriate way. In addition, the platform has a unique auto-learning feature that helps improve the answering accuracy,” RIL said in a statement. Introducing VCBaaS or Video Call Bot as a Service “Positioned as VCBaaS (video call bot as a service), JioInteract with its multimedia capabilities attempts to democratise AI and video call technologies in a simplified way for effective brand engagement. This technology has wide scale applications across B2C space. Using it, Jio is also tapping developer ecosystem to create innovative applications like virtual showrooms, product demonstrations, ordering cart for e-commerce, etc.”","excerpt":"Reliance Industries Limited announced the launch of Jiointeract on Thursday, in what the company is calling as the “world’s first artificial intelligence-based brand-engagement video platform.” The official statement read, “The first of many services to be launched on this platform is the Live Video Call that features India’s favourite celebrities. To kick-start, Jio has on-boarded […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Jio","reliance jio","RIL"],"author_name":"Prajakta Hebbar","publish_date":"2018-05-04T12:35:43","publication_year":"2018","word_count":368,"keywords":["Jio","artificial intelligence","programming_languages:R","AI","reliance jio","RIL","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/reliance-jiointeract-ai-platform-amitabh-bachchan\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35050,"title":"Understanding Why MATLAB Is Best Suited For Deep Learning","content":"Deep learning is a technique that is obtaining a foothold beyond multiple disciplines – enabling self-driving cars, predictive fault monitoring of jet engines, and time series forecasting in the economic markets and other use cases. In MATLAB it takes fewer lines of code and builds a machine learning or deep learning model, without needing to be a specialist in the techniques. MATLAB provides the ideal environment for deep learning, through model training and deployment. Why You Should Learn Matlab For Data ScienceComparing Different Programming Languages For Machine LearningMathWorks In Collaboration With NVIDIA’s DLI Offers New Deep Learning With MATLAB CourseMATLAB Expo 2015 – April, Bangalore & PuneMATLAB In Deep Learning, Analytics Space; Announces R2017B, Massive Update In September In this article, we see how MATLAB is gaining in popularity for deep learning: Why Matlab MATLAB programming platform has numerous advantages over other techniques or languages. The fundamental structure has a basic data element in a matrix. A simple integer is recognised as a matrix of one row and one column. Different mathematical methods that work on arrays or matrices are built into the Matlab environment. For instance, cross-products, dot-products, determinants, inverse matrices. Vectorized operations such as adding two arrays together need only one command, instead of a for or while loop. The graphical output is optimized for communication. Users can plot their data very simply, and then modify colours, sizes, scales, etc, by handling the graphical interactive tools. Matlab’s functionality can be considerably expanded by the addition of toolboxes. These are sets of specific functions that provided more specialized functionality. These features make the programming language very effective for implementing deep learning. The following are the fundamental features of MATLAB: Interoperability MATLAB supports interoperability with other open-source deep learning frameworks such as ONNX. Users can choose MATLAB for locating capabilities and prebuilt purposes and applications which are not available in other programming languages. Scope For Preprocessing Matlab gives scope for preprocessing datasets actively with domain-specific apps for audio, video, and image data. Users can visualize, check, and mend problems before training the Deep Network Designer app to build complex network architectures or modify trained networks for transfer learning. Multi-Program Deployment Matlab can use deep learning models everywhere including CUDA, C code, enterprise systems, or the cloud. It gives a great performance where a user can produce code that supports optimized libraries like Intel(MKL-DNN), NVIDIA (TensorRT, cuDNN), and ARM (ARM Compute Library) to build deployable patterns with high-performance inference activity. Deep Learn Toolbox Deep Learning Toolbox implements a framework for composing and performing deep neural networks with algorithms, trained models, and applications. A user can apply convolutional neural networks and long short-term memory (LSTM) networks to provide classification and regression on image, time-series, and text data. Apps and plots support users to visualize activations, edit network architectures, and monitor preparation progress.For modest training sets, a user can operate transfer learning with trained deep network models and models implied from TensorFlow-Keras and Caffe.For further training on large datasets, users can assign computations and data beyond multicore processors and GPUs on the desktop or scale up to clusters and clouds, including Amazon EC2® P2, P3, and G3 GPU instances. MatConvNet MatConvNet is a process of Convolutional Neural Networks (CNNs) execution for MATLAB. The toolbox is originated with an emphasis on simplicity and flexibility. The library is the building block of CNNs as easy-to-use MATLAB functions, providing methods for calculating linear convolutions with filter banks, feature pooling, and many more. The MatConvNet allows for fast prototyping of distinct CNN architectures, and at the same time, it supports efficient computation on CPU and GPU providing to exercise complex models on large datasets. Cuda-Convnet This library is a secured C++\/CUDA implementation of convolutional neural networks. It can display optional layer connectivity and network intensity. Any marked acyclic graph of layers is accepted. Training is done using the back-propagation algorithm. Backpropagation is a method implemented in ANNs to calculate a gradient that is needed in the calculation of the weights to be used in the network. Backpropagation is stenotype for the backward propagation of errors since an error is calculated at the output and distributed backwards throughout the network’s layers It is commonly used to train deep neural networks.Cuda uses Backpropagation as it is a generalization of the delta rule to multi-layered feedforward networks, which were made possible by applying the chain rule to iteratively compute gradients for each layer. It is intimately associated with the Gauss-Newton algorithm and is part of advancing research in neural backpropagation.","excerpt":"MATLAB provides the ideal environment for deep learning, through to model training and deployment. In this article, we see how MATLAB is gaining in popularity for deep learning","categories":["AI Features"],"tags":["cuda","Deep Learning","MATLAB"],"author_name":"Bharat Adibhatla","publish_date":"2019-02-15T17:58:58","publication_year":"2019","word_count":750,"keywords":["data science","machine learning","cuda","Keras","AI","neural network","TPU","MATLAB","Ray","deep learning","analytics","Deep Learning","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","data science","analytics","Ray","TensorFlow","Keras","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/understanding-why-matlab-is-best-suited-for-deep-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":38008,"title":"10 Emerging Analytics Startups In India To Watch Out In 2019","content":"India has witnessed a spectacular growth in the number of analytics startups over the last few years. In the last one year, we have covered numerous analytics startups that are working in the areas of healthcare, fashion, real estate, agriculture, facilitating lending decisions, emotional intelligence, voice-based solutions, hybrid TV powered by AI and many more. This is also indicative of the fact that India has emerged as a startup nation, with a robust ecosystem enabling startup founders and professionals to come up with innovative solutions in advanced analytics and kick-start their ventures. While many startups have risen to fame in the last one year, we bring you 10 emerging analytics startups that you could watch out for in the year 2019. The list is in reverse alphabetical order. 1| Zendrive, Founded 2013 Co-founders, Zendrive The team: It was founded by Jonathan Matus and Pankaj Risbood who were colleagues at Google, Facebook and Walmart Labs. Employee strength: 100 Eureka moment: Addressing the fact that the transportation system is evolving with each passing day while working on reducing the number of road traffic deaths globally, Zendrive came up with a state-of-the-art technology platform that leverages mobile sensor data to provide actionable insights and thereby improve safety for all road users worldwide. About the company and its analytics solutions: The startup uses data and analytics-based solutions to analyse drivers’ behaviour, predict road risks and reduce collisions. The company has analysed 160-billion miles of driving data to save lives and money for fleet management firms, insurance companies, and customers. It has built a first-of-its-kind AI-powered technology platform for fleet businesses, automobile companies, insurance firms and consumers to provide actionable insights about driving behaviour on the road, ensuring a safe and reliable drive. Zendrive offers a driver behaviour analytics solution to customers. Their mobile-based SDK product detects and analyses driving events like hard braking, rapid acceleration, over-speeding, hard turns and risky phone use. Their algorithms then, use these events to compute a safety score for the driver. Differentiator factor: With access to over 160 billion miles of driving data, Zendrive’s algorithms predict collision risk six times more accurately than industry leading models. This has been validated by Milliman, one of the largest global independent actuarial and consulting firms. Also, their solution is completely smartphone-based and does not involve any hardware. This saves resources on deployment and maintenance. Growth story: Over the last few years, Zendrive has been able to add prestigious names such as  GasBuddy, HopSkipDrive, BoosterFuels, Fleetio, EverTransit, in their client list. Funding:  Zendrive has raised a total of $20 million USD, most recently completing a Series A funding round of $13 million USD in 2016. 2| vPhrase, Founded 2015 Team vPhrase The team: vPhrase was founded by Neerav Parekh who brings extensive experience from technology, marketing and finance into this startup. Employee strength: 35 Eureka moment: The motivation behind vPhrase was the identification of a growing need for a tool that would help in making report generation efficient and simpler. Parekh worked towards identifying this gap and built a product to plug that gap. He was also strongly motivated to be an employment generator rather than being employment consumer. About the company and its analytics solutions: The company’s  AI platform, Phrazor, analyses data, derive insights and then communicates those insights in words and also in multiple languages. vPhrase automates insight and generates reports using the following products: Phrazor is a completely customizable self-service Natural Language Generation platform with functionalities of configuring and generating customisable reports in seconds Explorazor is a self-service exploratory tool that helps in analysing uploaded data based on all major analysis frameworks Explorazor generates insights, detects outliers and anomalies and provides an in-depth analysis of the data provided BI Dashboard Plugin adds management commentary to your dashboards on Tableau, Power BI or Qlik. It helps to make these dashboards easier to understand and act upon. Differentiator factor: Phrazor’s ability to marry analytics and Natural Language Generation is the key differentiator. With Phrazor, business teams don’t need to invest time in charts and tables, rather they get their actionable insights ready-made in seconds. Growth story: The startup has been able to generate many use cases from various industries. It has worked with clients such as Kotak Bank, HDFC Securities, ICICI Lombard Insurance, Star India, Abbott Pharma, and more. vPhrase is now expanding internationally where they are now working with the likes of Fidelity International, ABB, and OCBC. Funding: vPhrase had raised $250,000 Seed Round from Venture Catalyst, Zone Start-ups, and CIIE. 3| Thoucentric, Founded 2015 Team Thoucentric The team: The core team of Thoucentric includes Archi Bagchi, Neelakshi Kotnis, Pradeep Jadhav, Siddhartha Roy, Manish Garg, Bharat Kumar L, Satyam Tiwari, Karthik S, Ashok Babu R, Sridhar S, Prince Kumar, Elango M, Abha Bhatnagar, Rajan Jindal, Ankit Agarwal, Udit Ranjan, Prashant Bhushan Sharma, Ashish Verma, Ajay Kumar MS. Employee strength: 170+ Globally Eureka moment: The idea behind its inception was to create an organisation that focuses on solving complex business problems through process solutions or technology. Build by a group of strong techno-functional consultants, they are highly focused on execution. About the company and its analytics solutions: The startup covers problems across the entire cycle of analytics including prescriptive, predictive and cognitive. Some of the examples of the value delivered by the analytics team for clients include: Improved short-term demand forecast by building a ‘Demand Sensing Platform’ for a Global Personal Care major ML Approach to Production Planning; Reduced 60% Man-Hours & improved planning forecast accuracy to 94% for a global CPG player ML approach to exception-based planning impacted 50% reduction in the S&OP man hours ML approach to ‘right’ thinking on outlet prioritization & BTL spend allocation for a global Alco-Bev player Differentiator factor: The startup brings strong analytics experience in solving business problems. They believe in handholding the clients right from defining a problem to the last step of ensuring returns on investment are met. It is their collective passion that brings solutions hitherto unheard of and unprecedented, into fruition. Growth story: Starting with one consultant in 2015 they have grown to a team of 170+ consultants spread across India, UK, Australia and Singapore. They began as a two-man show and are today a team comprising of data scientists, data engineers and cloud architects. Some of their clients are Unilever, Mondelez, Tata Global Beverages Limited, Asian Paints, Nivea, Marico, Trident Group, Garware Group and Diageo and others. Funding: Thoucentric is bootstrapped and self-funded. 4| Spoonshot, Founded 2015 Team Spoonshot The team: Spoonshot was founded by Sai Sreenivas Kodur (ex Zomato) and Kishan Vasani (ex Just Eat). Employee strength: 20+ Eureka moment: The startup was founded to put a stop to stale, rear-view insights developed by out-of-touch companies that employ archaic methodologies when it comes to FMCG sector. Founders believe that over 50 percent of new products fail due to lack of understanding and a lack of testing with consumers and most consumer insight methodologies are not evolving fast enough to capture rapidly changing trends and tastes. Spoonshot was born to address these problems. About the company and its analytics solutions: Spoonshot (formerly known as dishq) was founded in 2015 with the aim of building a personalised food discovery app for the Indian market. dishq (a combination of dish and ishq), was focused on shifting food ordering away from ratings and reviews, to a more visual and emotionally driven decision. They changed the name to Spoonshot when they decided to pivot to a B2B audience and have an international focus. The startup uses food science to extrapolate novel and relevant insights, personalised to every user. Their methodology is geared towards uncovering hidden product innovation opportunities by surfacing early signals for novel and emerging ingredient combinations. Their analysis gives decision-makers the quantitative confidence they need to act faster. Differentiator factor: They leverage long-tail of open, alternative data to build proprietary structured data sets, connecting this data using machine learning, drawing signals and casualty. With these knowledge repositories, their goal is to ultimately replicate human cognition in the domain of food. Growth story: They have seen good growth in terms of product development and building solutions that actually solve key problems in the food and beverage industry. They have plans for hiring and driving product innovation ahead. They are currently working on Spoonshot Genesis which is in private beta stage and will be launched fully in June. For Genesis’ private beta, they are currently working with 40+ large FMCG and F&B brands, 15 of which are global power-houses of food, doing billions of dollars of business every year. Funding: Spoonshot has raised $500,000 till date and is looking to raise another round of funding later this year. 5| Pisquare, Founded 2014 Team PiSquare The team: Pisquare was founded by Chinmay Pradhan and Rojalin Biswal, who have more than 15 years of experience in management consultant and data analytics. Employee strength: 35 Eureka moment: With an aim to create a smarter workforce, they are working to create a smarter set of processes with the infusion of AI into the current workflows of enterprises. PiSquare wants to take AI across the layers in a web-based system where intelligence is available on demand. About the company and its analytics solutions: Initially started as Arima Research in 2014, they operate under the brand name of Pisquare. Pisquare partners with organisations to transform their decision-making ecosystem with the right mix of strategic planning supported by analytical insight. Using a combination of analytics and visualisation they are creating a spring board for businesses to reach the next level of performance.  They have been partnering with clients across the globe to optimise their customer engagement, supply chain, marketing & sales, delivery and talent for more aligned and mature performance. Their solutions are broadly structured into the areas of customer analytics, talent analytics and operations in IT\/ITES. Differentiator factor: They offer a combination of advanced analytics with advanced visualisation served on a web-based platform. They focus not only on the accuracy of predictions but equally on factors and inputs that are causing the effect Growth story: Within a year of inception they have worked with companies such as Dell, Capita Plc, Mahindra Finance, Tata Capital, Sunlife, Powerschool, Department of Transport and State Government departments, major banks and insurance organisations. Funding: They competed for a seed funding round last year 6| Pentation Analytics, Founded  2015 The team: It was co-founded by Anirban Roy (Founder Director & Chief Executive Officer), Kamal K Das (Co-Founder &Chief Operating Officer) and Pardeep K Shah (Director and mentor). Employee strength: 23 Eureka moment: The company targets core insurance use-cases, with an objective to deliver in the areas of soliciting & customer acquisition, risk assessment & policy development, claim assessment & customer service, fraud analytics, customer retention & cross-Sell. Using technologies such as predictive algorithms, AI, NLP, deep learning, gradient boosting trees and more, it analyses structural and unstructured data residing at various sources. About the company and its analytics solutions: Pentation analytics, as an InsurTech and Big Data Analytics company, provides analytical services to the BFSI industry, with its core product Insurance Analytics Suite ®. The analytics products by Pentation Analytics are IAS (Insurance Analytics Suite), Pentation Insurance Scores and Quote machine. These products bring use cases such as advanced analytics across the value chain, machine learning based scores at an individual policy level, acquisition via Self-quote generation and more. Differentiator factor: The product by Pentation Analytics is an all-in-one self-service platform, covering business intelligence & insights, planning & monitoring, policy-level scorings through ‘Ensemble scoring algorithms’, automated allocation of policy through ‘Operational optimization algorithms’, role-wise access, connection to 3rd party API and technology platforms and more. Growth story: Pentation Analytics has been recognized at multiple international as well as national industry platforms. The Insurance Analytics Suite® by the company has already been implemented with some major insurance companies, where they have been able to increase the retention through their predictive modelling and operational optimisation. Some of their major clients are NPCI, DCB Bank, Bharti AXA etc. Funding: They are at the early stage of funding. 7| Impact Analytics, Founded 2015 The team: The primary founder of Impact Analytics is Prashant Agrawal who has more than two decades of cross industry experience. Employee strength: 200+ across global geographies Eureka moment: Impact Analytics came into existence to fill a gap where clients were looking for solutions that combined best practices of management consulting along with analytic based products. About the company and its analytics solutions: Impact Analytics is the US-headquartered, rapidly growing analytics firm dedicated to serving clients through a synthesis of business consulting, analytics services and products. Impact Analytics offers a set of advanced analytics products and services for data-driven decision making in consumer-facing sectors such as CPG, retail, hospitality, banking, sports and gaming. Specific solutions centre around providing deep actionable insights for margin improvements, merchandising optimization, marketing analytics, customer analytics, operational excellence and AI-based automation solutions. Differentiator factor: Impact Analytics prides itself on being first in the market on several of the solutions which are driven by a core set of values deriving substantial savings in costs and maximising profitability for clients. Growth story: Growth has doubled every year since its inception. Some of their major clients include at least 12 of the top 500 companies in the United States. Funding: Impact Analytics secured $750,000 in a funding round led by early-stage VC firm Aarin Capital which is a tie-up between TV Mohandas Pai and Dr Ranjan Pai. Other investors who participated in this round include Michael Herzig, serial entrepreneur and Ashish Lakhanpal (CEO Kismet Capital). 8| Dataweave, Founded 2011 Team Dataweave The team: The company was founded by Karthik Bettadapura and Vikranth Ramanolla. Employee strength: 140+ employees Eureka moment: During Bettadupura’s previous role as the lead programmer of the data team at Web18, he recognised the market potential of providing a solution that harnesses publicly available data and analyses the external factors that impact businesses. This thought process led to the formation of DataWeave in 2011 along with Ramanolla, who shared Karthik’s vision and passion. About the company and its analytics solutions: DataWeave provides competitive intelligence to e-commerce businesses and digital shelf analytics to consumer brands by aggregating and analysing data from the web at a massive scale. The company’s AI-powered technology platform enables e-commerce businesses to make smarter pricing and merchandising decisions to drive profitable growth, and consumer brands to govern their online brand presence, optimise their Share of Voice, and improve their e-commerce shelf velocity. DataWeave offers SaaS-based product suites called Retail Intelligence and Brand Analytics that helps companies protect their brand equity online and optimise the experience delivered to shoppers on e-commerce websites. Popular use cases include monitoring and resolving minimum advertised price (MAP) violations, detecting counterfeit product listings, tracking and improving the share of voice of online promotions, optimising their digital shelf velocity, and more. DataWeave also provides investment firms and hedge funds with Retail Alternative Data, which enable them to take data-driven buy or sell positions in the market. Differentiator factor: DataWeave stands apart from its competitors in several ways. Their human-aided machine intelligence based technology platform leverages proprietary NLP and Computer Vision technologies to enable accurate and rapid data processing and insights generation. Their technology platform is also language agnostic, currently supporting over 25 international languages. They support diverse delivery mode options including API, CSV, PPT, FTP for easy and speedy consumption. Growth story: DataWeave is one of very few SaaS startups of Indian origin that has replicated its domestic success internationally, particularly in the US. They have a strong presence in India, the Middle East, and SouthEast Asia and the US. Funding: Following a seed funding round in 2011, they announced Series-A round of financing in April 2017, which was led by FreakOut Group, Herb Madan – a Silicon Valley investor, and a diverse group of institutional investors from US, India, and Singapore. Other investors include Blume Ventures, M&S Partners, Rajan Anandan, Times Internet, and WaterBridge Ventures. 9| Crediwatch, Founded 2016 Team Crediwatch The team: It was founded by Meghna Suryakumar and Sandeep Anandampilla. Employee strength: 36 employees across two locations Eureka moment: The founders state that India has over 50 million unregistered businesses and over 1.13 million active registered companies but only 7000 of these are listed offering detailed while the rest exist in the ‘Dark Space’. There is an estimated $1000bn annual business entered with these ‘dark space‘ companies as a combination of credit, trading and other agency related activities. Crediwatch was created as a cost-effective and scalable solution to provide transparency and information about these businesses. About the company and its analytics solutions: Crediwatch is an insights-as-a-service platform that deploys scalable deep learning tools across disparate digital footprints left by private entities (big and small) to provide dynamic credit management as a service to financial institutions. It uses advanced computational techniques such as AI, ML and NLP to derive near-real-time insights from structured and unstructured datasets. Crediwatch has created an Early Warning System (EWS) that banks and NBFCs can use to monitor their loan portfolios in near real-time using a combination of public and private data. This system uses globally-used models trained on Indian corporate datasets and predicts delinquencies and defaults 12 to 18 months before they occur. Differentiator factor: Their proprietary predictive algorithm has been developed and tested on over 9000 Indian companies. Crediwatch provides singularity of insights by tapping into 2500 data sources and tying the results into a singular truth hence allowing for a verified and accurate delivery of insights. Other differentiating factors are zero human touches, platform agnostic solutions, scalability and more. Growth story: Since 2016, Crediwatch has provided credit analytics for a portfolio of over INR 50,000 crores across 20+ BFSI clients analysing over 1.2 million data tokens along the way. Some of Crediwatch’s clients are  Barclays Bank, Capital Float, RBL Bank, SBI, Trilegal etc. 10| BDB, Founded 2015 Team BDB The team: Avin Jain (Founder, CEO), Anoop VP (CTO), Vishal Venugopal (VP AI) and several others founded BDB with a focus on end-to-end data, analytics, AI platform. Employee strength: 120+ Eureka moment: Being in BI consulting for the initial four years the founding team realised the pain points of the analytics industry and that many existing analytics tools were unable to fulfil the analytics demands of customers. There was no single platform which could address end to end analytics requirement and therefor BDB platform was founded as a single end-to-end analytics platform to address these challenges. About the company and its analytics solutions: The BDB Decision Platform brings analytics within the reach of every individual by enabling seamless workflows for all user groups including CXO, business user, citizen data scientists and more. With customers in 10 different verticals, BDB platform has been deployed in education, retail, healthcare, life science, BFSU and other domains. BDB has a cloud-based SaaS Analytics Platform for SMEs – Yujaa which internally uses a set of business intelligence tools that gives a 360° view of data. Yujaa provides role-based solutions to various industries that are affordable and accessible by any user across the globe. Differentiator factor: BDB platform provides a comprehensive experience of data analytics offering plugins such as data pipeline and data wrangling all tacked together in a simple drag and drop manner. BDB can be deployed on-premise or accessed as a cloud-based application. The product is built on microservices architecture and contains all the features of a modern, highly scalable, secure, multi-tenant analytics platform. Growth story: BDB has been growing significantly with its solutions available on cloud and on-premise. It has earned more than $8 million through its platform in the last four years and has worked with a few fortune 500 companies in different verticals such as enterprise customers, SMEs and more. Funding: Currently BDB is looking for Series A funding.","excerpt":"India has witnessed a spectacular growth in the number of analytics startups over the last few years. In the last one year, we have covered numerous analytics startups that are working in the areas of healthcare, fashion, real estate, agriculture, facilitating lending decisions, emotional intelligence, voice-based solutions, hybrid TV powered by AI and many more. […]","categories":["AI Features"],"tags":["retail bi prescriptive","Walmart Labs"],"author_name":"Srishti Deoras","publish_date":"2019-04-22T06:45:39","publication_year":"2019","word_count":3298,"keywords":["machine learning","AI","retail bi prescriptive","Walmart Labs","ML","computer vision","RAG","NLP","microservices","Aim","deep learning","analytics"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","computer vision","analytics","Aim","RAG","microservices"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/10-emerging-analytics-startups-in-india-to-watch-out-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10003646,"title":"The Importance Of No Free Lunch Theorems In Deep Learning","content":"“The no free lunch theorem calls for prudency when solving ML problems by requiring that you test multiple algorithms and solutions with a clear mind and without prejudice.” In a paper titled, ‘The Lack of A Priori Distinctions Between Learning Algorithms’, that dates back to 1996,  David Wolpert explored the following questions: Can we actually get something for nothing in supervised learning? Can we get useful, caveat-free theoretical results that link the training set and the learning algorithm to generalisation error, without making assumptions concerning the target? Are there useful practical techniques that require no such assumptions? He showed that for any two algorithms, A and B, there are as many scenarios where A will perform worse than B as there are instances where A will outperform B. In short, for all possible problems, average performance of both the algorithms is the same. Implications Of NFL In Deep Learning Although the no free lunch theorem by Wolpert has a more theoretical than practical appeal, there are some implications that should still be taken into account by everyone working with machine learning algorithms. These theorems prove that under a uniform distribution over search problems or learning problems, all algorithms perform equally. In recent work, Wolpert discusses the importance of NFL theorems for machine learning. Search and learning are key aspects of ML and the NFL theorems have something to deliver here. For Search To illustrate it, writes Wolpert, choose a set of objective functions on which a certain search algorithm performs better than the purely random search algorithm. Then the NFL for search theorem says that compared to random search, this favourite search algorithm “loses on as many” objective functions as it wins and this is true no matter what performance measure one uses. According to Wolpert, the primary significance of the NFL theorems for search is to give insights about the underlying mathematical ‘skeleton’ of optimisation theory before the ‘flesh’ of the probability distributions of a particular context, and set of optimisation problems are imposed. For Supervised Learning The performance of any supervised learning algorithm, the paper states, is governed by an inner product between two vectors, both indexed by the set of all target functions. As long as the loss function is symmetric, it indicates how close the results of the supervised learning algorithm are to the real world. This supervised learning inner product formula results in a set of NFL theorems, which can come in handy. Why NFL Enables Prudency Wolpert, who has been instrumental behind the proliferation of NFLs, in his recent work, wrote that there are many avenues of research related to the NFL theorems which are yet to be explored properly. Some of these involve free lunch theorems which concern fields closely related to search, e.g., coevolution. “NFL theorems can give insights about the underlying mathematical ‘skeleton’ of optimisation theory before the ‘flesh’ of the probability distributions of a particular context and set of optimisation problems are imposed.”David Wolpert Luca Massaron, a best selling machine learning author and a Google Developer Expert, who has been in the field of machine learning for nearly two decades, says that, though some learning algorithms are usually considered the best-in-class for certain types of problems, such as for instance gradient boosting with tabular data problems, any practitioner should not overlook that the ultimate goal of a learning algorithm is to correctly map an unknown function ruling how the predictions relate to the data inputs available. Since such function is unknown, there is no a-priori guarantee that the best learning algorithm is being chosen for the problem if one picks a state of the art algorithm. The no free lunch theorem, explains Luca and calls for prudency when solving machine learning problems. Sometimes, by testing multiple solutions, one might even find that simpler solutions may work perfectly well for some problems, without resorting to more complex, state of the art ones. One may discover that in the long run, other algorithms could be more performant than the picked state of the art solution. Link to paper by David Wolpert","excerpt":"“The no free lunch theorem calls for prudency when solving ML problems by requiring that you test multiple algorithms and solutions with a clear mind and without prejudice.” In a paper titled, ‘The Lack of A Priori Distinctions Between Learning Algorithms’, that dates back to 1996,  David Wolpert explored the following questions: Can we actually […]","categories":["Deep Tech"],"tags":["supervised learning"],"author_name":"Ram Sagar","publish_date":"2020-07-30T13:00:00","publication_year":"2020","word_count":679,"keywords":["Go","machine learning","programming_languages:R","AI","ML","programming_languages:Go","RAG","deep learning","supervised learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/no-free-lunch-theorems-deep-learning\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10164441,"title":"Delta Electronics Launches Collaborative Robots, Announces $500 Million Investment in India","content":"Delta Electronics, a Taiwan-based provider of power and thermal management solutions, launched its D-Bot Series Collaborative Robots (Cobots) and new products for the smart manufacturing market in India at ELECRAMA 2025, Greater Noida, NCR. The robots are tailored for smart factory automation and have advanced safety mechanisms. They are 6-axis cobots with payload capacities of up to 30 kg and speeds of 200 degrees per second, enhancing efficiency in electronics assembly, packaging, materials handling, and welding. They can be integrated with Delta’s VTScada SCADA system, DIATwin digital twin platform, and machine vision systems, ensuring precision and efficiency in industrial automation. Alongside the D-Bot Series, the company also launched its 240kW DC Fast EV Charger and Industrial Power-protect Transformer-based UPS (IPT Series), reinforcing its commitment to smart manufacturing and energy sustainability in India. The EV charger is a high-speed dual-vehicle charging solution developed locally by Delta India’s R&D and engineering talent. The charger claims to have 95% efficiency, OCPP compatibility, and wired\/4G GSM connectivity to facilitate seamless and reliable operations for all use cases. In addition to the product launches, Delta Electronics is investing $500 million (announced in 2015 under the government’s Make in India initiative) as part of its long-term commitment to India. The company announced that this investment will expand its local presence and strengthen its domestic manufacturing, R&D, and innovation capabilities. Jimmy Yiin, executive vice president of global business operations at Delta, stated, “India is a key market for Delta, and we are committed to driving its industrial and energy transformation with our advanced solutions”. On the investment, he noted, “Delta’s strategic investment in the Krishnagiri facility underscores our dedication to local innovation, manufacturing excellence, and sustainability. Through this investment, we aim to strengthen India’s self-reliance in smart manufacturing and energy infrastructure while contributing to global industry standards.” Benjamin Lin, president of Delta Electronics India, added, “Delta is proud to showcase our innovative smart manufacturing solutions at ELECRAMA 2025. Our focus has always been on delivering solutions that advance industries to become more efficient, sustainable, and resilient.”","excerpt":"More companies are investing in India to scale their domestic operations, strengthening the manufacturing landscape.","categories":["AI News"],"tags":["ai funding","AI Robots"],"author_name":"Ankush Das","publish_date":"2025-02-24T12:40:12","publication_year":"2025","word_count":339,"keywords":["Go","ELT","AI Robots","AI","innovation","ML","Git","automation","Aim","ai funding","ViT","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Git","ELT","ViT","automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/delta-electronics-launches-collaborative-robots-announces-500-million-investment-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10066629,"title":"Hugging Face raises USD 100 Mn in Series C","content":"Hugging Face has raised USD 100 million in Series C funding led by Lux Capital. Sequoia, Coatue and existing investors including Addition, Betaworks, AIX Ventures, Cygni Capital, Kevin Durant, Olivier Pomel (co-founder & CEO at Datadog) etc participated in the round. “Machine learning is becoming the default way to build technology. Hugging Face is the most used ML platform and community with over 10,000 companies using it, 100,000 pre-trained models & 10,000 datasets shared on the hub for NLP, computer vision, speech, time-series, biology, reinforcement learning, chemistry and more. Not only does Hugging Face host models and datasets but empowers companies to test them, collaborate on them, run them in production and assess them for a more ethical use thanks to its amazing community,” said Julien Chaumond, co-founder and CTO at Hugging Face. Hugging Face aim to create a positive impact on the AI field by focusing on responsible AI through openly sharing models, datasets, training procedures, and evaluation metrics. The team believes that open source and open science bring trust, robustness, reproducibility, and continuous innovation. “Though many practitioners emphasise the long term impact of machine learning and eventually AGI that mostly points towards singularity or a “terminator” effect, we chose to focus on the limitations and challenges of ML that need to be tackled now like biases, privacy, and energy consumption. We believe that through openness, transparency and collaboration, we as a community can foster responsible and inclusive progress, understanding and accountability to mitigate these challenges. This way, we aim to build a better future where AI is founded on open source, open science, ethics and collaboration. With the new funding, we will be doubling down on research, open-source, products and responsible democratisation of AI,” said Clement Delangue, CEO and co-founder at Hugging Face. Hugging Face is also leading BigScience, a collaborative workshop around large language models gathering more than 1,000 researchers of all backgrounds and disciplines. The community is now working towards training the world’s largest open-source multilingual language model. Hugging Face started its life as a chatbot and has come a long way to become “the home of machine learning.” In the past 12 months, the company has grown from 30 to 120+ members and is actively hiring.","excerpt":"With the new funding, we will be doubling down on research, open-source, products and responsible democratisation of AI.","categories":["AI News"],"tags":["Clément Delangue","Ethical AI","future","Hugging Face","Machine Learning","Natural Language Processing","NLP","Open Source","Responsible AI"],"author_name":"Sri Krishna","publish_date":"2022-05-09T22:00:56","publication_year":"2022","word_count":371,"keywords":["future","Hugging Face","API","machine learning","Open Source","Rust","AI","Natural Language Processing","Ethical AI","ML","Machine Learning","Responsible AI","computer vision","NLP","Aim","Clément Delangue","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","computer vision","Aim","Hugging Face","R","Rust","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hugging-face-raises-usd-100-mn-in-series-c\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":11365,"title":"How India will change Artificial Intelligence","content":"It’s been an interesting couple of months. In between helping set up a new IoT insurance center in Munich and some corporate restructuring, I’ve also been catching up on my reading. There’s a lot in the press about ‘artificial intelligence’ and ‘cognitive analytics’ – in effect the whole issue of machine learning. I was also especially interested also to read about the top 10 Indian analytics start-ups in 2015, and that some of these are already thinking about the AI agenda. In my reading, I came across a very interesting academic report called ‘Future Progress in Artificial Intelligence: A survey of Expert Opinion’ *. It originates at the Future of Humanity Institute of the University of Oxford, and in addition to pointing to concerns about the impact of AI on humanity, suggests that there is a ‘one-in-two’ chance that high-level AI ‘will be developed around 2040-2050, rising to a nine-in-ten chance by 2075.’ Of the group of 500 eminent scientists surveyed for the report, 11% said that we will understand the architecture of the brain sufficiently to create machine simulation of human thought within 10 years. Also, 5% suggested that machines will be able to simulate learning and every other aspect of human learning within 10 years. Malone’s Third Law of Technology says that technological development always takes longer than we expect, but even when it eventually arrives we still won’t be ready for it. Personally, I think it will happen sooner rather than later. Let me explain why. My hunch is that despite economic volatility, there will be a rapid and substantial expansion of financial services especially in India and China whose regulation and legislation are promoting growth in those areas. Even with large populations, there is nevertheless a shortage of talent in the financial services sector. This will increasingly, rapidly and inevitably lead to the development of automated analytical systems such as AI to replace the ‘talent shortage’. This is a different problem from the mature markets of Western Europe and North America that suffer from overcapacity of financial service providers, and whose imperative in using AI is to reduce operating cost, rather than to manage growth. Opportunity for growth is always a stronger reason to change than is cost-cutting. There are also issues of technology market maturity to think about. Already the Indian tech market suffering from a ‘brain drain’ to Silicon Valley and elsewhere. Will they increasingly return, bringing back with them the skills and competencies they have refined overseas? (China already also sends many of its techies oversea, later to return and apply the ‘Chinese Way’ to what they have learned.) Perhaps AI will increasingly be driven by economic need rather than technological developments. The scale of India, China (and probably also Brazil) will provide a superb catalyst for innovation. And maybe the balance of technology power, if there is such a thing, might start to see some form of seismic shift. *Muller, Vincent C and Bostrom, Nick (2016). ‘Future Progress in artificial intelligence: A survey of Expert Opinion’. Synthesis Library; Berlin; Springer).553-571","excerpt":"It’s been an interesting couple of months. In between helping set up a new IoT insurance center in Munich and some corporate restructuring, I’ve also been catching up on my reading. There’s a lot in the press about ‘artificial intelligence’ and ‘cognitive analytics’ – in effect the whole issue of machine learning. I was also […]","categories":["IT Services"],"tags":["India AI"],"author_name":"Tony Boobier","publish_date":"2016-11-26T06:19:36","publication_year":"2016","word_count":508,"keywords":["India AI","API","machine learning","artificial intelligence","programming_languages:R","AI","innovation","ViT","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","R","API","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/india-will-change-artificial-intelligence\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10123276,"title":"LlamaGen Beats Diffusion Models for Scalable Image Generation","content":"The University of Hong Kong and ByteDance have unveiled LlamaGen, a new family of autoregressive models that outperform popular diffusion models like LDM and DiT for high-resolution image generation. The key breakthrough is that LlamaGen applies the same “next-token prediction” paradigm used in large language models to the visual domain without relying on inductive biases tailored for vision. The LlamaGen models range from 111 million to 3.1 billion parameters and achieve an impressive 2.18 FID score on the challenging ImageNet 256×256 benchmark, surpassing state-of-the-art diffusion models. For class-conditional image generation, LlamaGen-3B realises 2.32 FID with classifier-free guidance at 1.75 scale. Read the full paper here. Training Method Notably, the researchers developed an image tokeniser with a downsampling ratio of 16 that achieves 0.94 reconstruction FID and 97% codebook usage on ImageNet. This discrete representation matches the quality of continuous VAE representations used in diffusion models. For text-conditional generation, a 775M parameter LlamaGen model was first trained on 50M image-text pairs from LAION-COCO, then fine-tuned on 10M high-quality images. It demonstrates the competitive visual quality and text alignment on challenging prompts from datasets like PartiPrompts. A key advantage of LlamaGen is its ability to leverage optimisation techniques developed for large language models. The researchers showed a 326-414% speedup using the vLLM serving framework compared to baseline settings. While still behind the latest diffusion models on some metrics, the researchers believe LlamaGen paves the way for unified autoregressive models spanning language and vision. With more training data and computing, they aim to scale LlamaGen above 7B parameters for further gains. Up Next OpenAI’s Sora was released earlier this year, and with Google recently releasing Veo, text-to-video AI models are now gaining prominence. As these improved capabilities demonstrate that image generation can become faster and more accurate, they can also be applied to open-source video generation models, putting them on par with video-generation models like Sora and Veo.","excerpt":"LlamaGen applies the same “next-token prediction” paradigm used in large language models to the visual domain without relying on inductive biases tailored for vision.","categories":["AI News"],"tags":["diffusion models"],"author_name":"Gopika Raj","publish_date":"2024-06-11T17:30:09","publication_year":"2024","word_count":315,"keywords":["Go","OpenAI","AI","RPA","diffusion models","RAG","Aim","VAE","llm_models:Llama","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","RAG","R","Go","diffusion models","VAE","RPA","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/llamagen-beats-diffusion-models-for-scalable-image-generation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10052771,"title":"Why Is Yoshua Bengio Betting On This Newly Launched AI Governance Platform?","content":"Toronto-based Armilla AI recently launched a first of its kind, all-in-one AI governance platform. The company also announced that it has managed to bag $1.5 million in seed investment. Armilla AI brings stakeholders together with automation validation tools to test machine learning algorithms for accuracy, robustness, fairness, bias, data drift, etc. The company claims that its services help its customers deploy trustworthy AI models. The company has an interesting lineup of investors, including entrepreneur-investor Naval Ravikant’s Spearhead Fund, Alana and Eva Lau’s Two Small Fish Ventures, and C2 Ventures. AM Turing Awardee Yoshua Bengio has also invested in the company and Apstat partners Nicolas Chapados and Jean-Francois Gagne. A backing from Bengio, a pioneer in AI and deep learning, lends Armilla AI reliability. But why did Bengio choose to invest in this young start-up? About Armilla AI AI-based technology has touched almost every aspect of our lives, including work, entertainment, leisure, play, society, and culture. AI has managed to positively impact every sector where it has and continues to be employed, including finance, science, transportation, healthcare, environment, etc. That said, there are a few glaring problems with this technology which are further compounded when scaled and deployed at massive proportions. In the past, AI fallacies have caused monumental losses to individuals and communities. For example, a study conducted in 2019 revealed that a widely used algorithm in US hospitals was systematically discriminating against black patients. The study found that the algorithm was less likely to refer a black person compared to a white person for the same ailment. Hospitals and insurers use similar algorithms to manage care for about 200 million people in the US each year. Faulty AI algorithms and models are a result of massive growth and increased complexity. Traditional testing methods have not been able to keep up with such explosive growth, resulting in erroneous outputs that include bias and negative results. Armilla AI offers an ML testing platform that offers organisations tools to plan, experiment, validate, and archive models to remediate this problem. Armilla automated the testing process, which involves more than 50 steps to assess any miscalculations in the ML models. This system includes Armilla FingerPrint, a validation framework that learns the sensitivities of any system and allows organisations to monitor their machine learning system in production. The entire process deployed by Armilla is fully auditable by logging all tests conducted, discovered issues, and problems resolved. This allows previously siloed business stakeholders like executives, managers, and data scientists to view and collaborate directly on results in real-time. Speaking of the company, Yoshua Bengio said, “AI models are making more critical decisions every day, which means they require new oversight protocols that can ensure they are accurate, fair, and curb potential abuse. This growing need for independent validation requires the same attention and investment used to build models themselves. This is how to responsibly build AI.” Prof Bengio’s Quest for Ethical AI Bengio is recognised as one of the leading AI experts. Prof Bengio completed his postdoctoral studies at the Massachusetts Institute of Technology (MIT) in Boston. He has been bestowed with prestigious awards like the Turing Award in 2018 and the Killam Prize in 2019. Bengio is a great proponent of ethical and trustworthy AI. In a 2019 interview with Nature, Bengio spoke at length about irresponsible and unethical AI usage. He said that AI could amplify discrimination and biases. In the same interview, he spoke about how the need of the hour was to go beyond just self-regulation in terms of AI ethics and devising government or international guidelines for AI. Writing a guest post on The Conversation, Bengio said that the objective of tech innovation is to reduce human misery and not increase it. However, AI is very well capable of doing the latter. He further wrote that such discrimination originates because of the inherent biases of humans developing these machines. In this article and in general, Bengio has called for a resolution of such discriminations by involving all stakeholders, including the government. More notably, Prof Bengio founded Mila, a research institute for AI, born out of a partnership between the Université de Montréal and McGill University. It has 500 researchers, making the institute one of the largest academic research centres in the world for machine learning. The institute was responsible for establishing an ethical framework through the Montreal Declaration for Responsible Development of AI. This framework proposed ethical principles based on ten fundamental values — well-being, respect for autonomy, privacy and intimacy, solidarity, democratic participation, equity, diversity inclusion, prudence, responsibility and sustainable development. About the declaration, Bengio has said that its goal is to establish principles that would form the basis for adopting new rules and laws pertaining to responsible AI. He also highlighted that the current laws were not equipped enough to deal with the challenges of AI.","excerpt":"“This growing need for independent validation requires the same attention and investment used to build models themselves. This is how to responsibly build AI.”","categories":["AI Features"],"tags":["Responsible AI","Trustworthy AI","Yoshua Bengio"],"author_name":"Shraddha Goled","publish_date":"2021-11-02T18:00:00","publication_year":"2021","word_count":807,"keywords":["Go","machine learning","TPU","AWS","AI","ML","Responsible AI","Trustworthy AI","Yoshua Bengio","deep learning","Aim","Rust","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","Aim","AWS","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-yoshua-bengio-betting-on-this-newly-launched-ai-governance-platform\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10003471,"title":"Can Machine Learning Be Deployed For Cancer Detection On A Large Scale?","content":"Machine learning applications in healthcare have been there for a while. Research has been consistently evolving and more areas have been expanded under this umbrella. It is not only being used in the diagnosis and treatment of cancer, but also in the intricacies of several other human conditions. The union of medical knowledge with machine learning techniques is trivial, in the sense that the applications are direct and provide an extremely deterministic value. In this article, we will talk about the applications of such algorithms in cancer detection and how they are carried out since, although there are several research papers discussing the biological and technical highlights of the same, there are few that talk about the hands-on applications, and some technician who wants to enter the sector may find it difficult to land the techniques without prior knowledge in computer science. Neural networks applied to cancer detection One of the most prominent and popular applications in the implementation of machine learning algorithms for cancer detection is the one carried out through Computer Vision. Although detecting cancer using images is not the only machine learning application out there -it is also possible to make detections from structured data with common supervised problems- it is the most famous one. Let’s see how this type of algorithm works using an example of detecting breast cancer. If you would like to know more, join the upcoming CVDC 2020 event. Detecting breast cancer using machine learning algorithms has been a topic of much discussion lately. One of the most trending innovations in the last months has been Google Health, which introduced an artificial intelligence system that detected breast cancer in the early stages. In that study, in which they used data from more than 90,000 women, artificial intelligence was able to predict better than the medical experts since it had far fewer errors in the diagnosis. I am going to show you how these imaging cancer detection algorithms generally work. In the following image, you can see a complete system for detecting breast cancer from x-ray images. The system is based on the classification of images using deep neural networks -commonly called deep learning. These deep neural networks are of a certain type that allows images to be treated fairly efficiently. They are the so-called convolutional neural networks or CNN. Neural networks can obtain what is called activation maps. This is nothing more than a series of heat maps representing the areas in which the model has been based to make a certain prediction. In the previous image, in green are the areas on which the model has been based to determine that this image is not carcinogenic. In red, the areas on which it has been based to determine that it is carcinogenic. In this way, the expert could directly see the parts that the model assumes are important in the prediction. I would like to emphasize that machine learning models should not be considered a substitute for medical experts. They must be a useful complement through which to improve the results and efficiency of the experts’ work. In the image below, I put an example of how these systems would work. The image is of the detection of pneumonia but the operation is identical in the detection of cancer. From images, a machine learning algorithm is used to predict and obtain heat maps that give interpretations to those predictions. Subsequently, an expert radiologist would decide based on the original images and the model’s predictions. Without going into technical aspects -which can be checked in the research paper mentioned above- we are going to see how the system would work and how it would be applied at a high level. The system, after being trained, would receive an x-ray image of which its class is unknown. Using a machine learning model, it would predict the probability of that region for having cancer exposure or not. You would basically get a “cancer” or “non-cancer” prediction. I am sure this sounds great, but surely the expert radiologist would come and ask “Why does this machine learning model tell me that there is cancer and what is it based on? I can’t just trust such a model like that. It is just an algorithm”. This is where we apply what is called the explainable artificial intelligence. In the following image, you can observe how this type of model is endowed with explanations in the biomedical field.","excerpt":"Machine learning applications in healthcare have been there for a while. Research has been consistently evolving and more areas have been expanded under this umbrella. It is not only being used in the diagnosis and treatment of cancer, but also in the intricacies of several other human conditions. The union of medical knowledge with machine […]","categories":["AI Features"],"tags":["human intelligence at machine scale","Machine Learning"],"author_name":"Dr. Raul V. Rodriguez","publish_date":"2020-07-28T18:00:00","publication_year":"2020","word_count":740,"keywords":["Go","artificial intelligence","machine learning","human intelligence at machine scale","AI","neural network","Machine Learning","computer vision","Ray","deep learning","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","computer vision","Ray","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-machine-learning-be-deployed-for-cancer-detection-on-a-large-scale\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10089600,"title":"Council Post: The Rise of Generative AI and Living Content","content":"Marshall McLuhan once said, “We shape our tools and thereafter our tools shape us”. The concern about technology entering every human space is not novel. With various developments through time such as processors, digital photography, creative software editing suites, music editing software, and computer graphics, the discourse between human creation and technology has continued through time. Humans are capable of leaps of logic that machines are yet to catch up with. Basic computer programming is merely one level above AI. Recent advances and accomplishments in AI are indubitably tied to human intellectual capacity. Although machines are still capable of much more than the human brain, humans differ significantly in terms of application of their knowledge by using logic, reasoning, understanding, learning, and experience. But concerns surrounding man versus machine saga have often lost ground to reality. It is true that a number of advancements in technology have made human involvement redundant in certain aspects of the creative process. However, even as the concern of being replaced is undoubtedly real, it is unlikely to happen that humans would be replaced by machines completely. The article will be talking about how content is going to evolve after the coming of generative AI. At the same time, it will address whether we’d really need writers when we have AI to write for us. Will content evolve or will it get too automated for readers? According to a Reuters source, ChatGPT is the creator or co-author of more than 200 books that are available on Amazon as paperbacks or e-books. The investigation also finds that as Amazon standards do not compel users to acknowledge the usage of AI in their books, the number of books authored by AI may be significantly higher than the number of books actually listed. More than 200 e-books in the Kindle shop on Amazon attributed ChatGPT as the author or co-author during the month of February alone. As more books are published, Amazon has introduced a brand new sub-genre devoted to books about using ChatGPT that are wholly authored by ChatGPT itself. At present, readers do not engage with lengthy content but instead prefer to consume media that is pertinent, succinct and tailored to their interests. This shift towards a more direct approach is something that the consumers are demanding for. Customers want content that caters to their unique needs and interests. Such content may be tailored to the preferences of certain audiences using AI and other cutting-edge technology, thereby giving each user a unique experience. The transition to interactive content is altering how we absorb information as well. Interactive content offers a more dynamic and engrossing way to information distribution, ranging from static images and text to immersive and engaging experiences. Every story should either be relatable or something that manifests a possibility of happening. Long-form Content Merges with Living Content Simultaneously, there is an audience that wants narrative content, which is typically referred to as long-form content. AI can make long-form content by using a technique called natural language generation (NLG). NLG is a subset of artificial intelligence that focuses on generating human-like language from data. Some people will prefer shorter, more precise content that gets straight to the point, while others will appreciate the depth and nuance that long-form content can provide. Additionally, the type of content and the purpose it serves can also impact its reception. For example, people may be more likely to consume long-form content for entertainment or educational purposes while they may prefer shorter, more concise content for news or information that needs to be consumed and understood quickly. It’s also worth noting that the rise of generative AI does not necessarily mean the decline of long-form content. While generative AI may be able to create coherent text, it may not necessarily be able to create engaging, thought-provoking content that resonates with readers. In many cases, long-form content is valued precisely because it provides an opportunity for in-depth exploration of complex topics, which may be difficult for generative AI to replicate. A theory of concept localisation highlights a major challenge in comprehending unfamiliar ideas when they are presented without sufficient context. As human beings, we tend to rely on metaphorical explanations to make sense of complex concepts. We learn best when someone provides us with a metaphor that allows us to understand and contextualise the underlying meaning or connotation of the concept at hand. With the advent of advanced language models, one can take a given concept and translate it into metaphors that are tailored to an individual’s unique background, making it easier for them to understand and absorb the idea. While AI-generated content is yet to perfect this process and may require significant refining by a human editor, it has the potential to greatly speed up the content creation process and help businesses and individuals produce high-quality, engaging content at scale. Rise of Living Content But just when we believed that this is the extent of what technology is capable of, something else comes along. No user can imagine how living content will develop in the future. Living content can be ‘tweaked’ for consumers, in the sense that it can be personalised to meet the needs and interests of individual readers. This can be accomplished through the use of data analytics and machine learning algorithms that analyse a reader’s behaviour and preferences and then curate content that is tailored to their interests. Living content can take many forms, including blogs, news articles, social media updates, and more. The key characteristic of living content is that it is constantly updated so that consumers can return routinely to get the latest information. Mark Twain once said, “There is no such thing as a new idea. It is impossible. We simply take a lot of old ideas and put them into a sort of mental kaleidoscope. We give them a turn and they make new and curious combinations.” Generative AI is trained on existing ideas to present seemingly new content. Conversely, the value of ideas with generative AI gets enhanced through the increased efficiency and scalability of idea generation. With generative AI, it is possible to create a large number of unique and original ideas in a relatively short amount of time, which can be particularly valuable for industries that rely on creative output, such as advertising and marketing. It is however noteworthy that, at present, humans are the only intellectual beings capable of such leaps of logic and epiphanies. Generative AI can also improve the quality and diversity of ideas generated, as it can draw on a vast amount of data and knowledge to create new and innovative ideas. This can help businesses stay ahead of the competition by providing them with unique and valuable insights that would be difficult or time-intensive to obtain through traditional research methodologies. Another way that the value of ideas with generative AI gets better is through the ability to personalise ideas based on individual preferences and needs. With generative AI, it is possible to create content that is tailored to the specific interests and preferences of an individual, which can improve engagement and drive better outcomes. In this era of content, the use of technology, such as AI and data analytics, is becoming increasingly important as it can help content creators personalise their content, improve its quality, and reach their target audience with greater efficacy. AI writing has arrived and is here to stay. Once we overcome the initial need to cling to our conventional methods, we can begin to be more receptive to the tremendous opportunities that these technologies present. Not only do they offer writers the chance to advance from being merely word processors to thought leaders and strategists, they also quicken the pace of content creation significantly. As this technology advances, authors will be able to devote more of their time to deep thought, developing their creative visions and original viewpoints. The majority of those who will profit from this inevitable change in the industry will be writers with original ideas. By expressing these thoughts with impact, clarity, and conciseness, the world of content creation is looking at a renaissance of its own. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here","excerpt":"In this era of content, the use of technology, such as AI and data analytics, is becoming increasingly important as it can help content creators personalise their content, improve its quality, and reach their target audience with greater efficacy. AI writing has arrived and is here to stay. Once we overcome the initial need to cling to our conventional methods, we can begin to be more receptive to the tremendous opportunities that these technologies present.","categories":["AI Features"],"tags":["Generative AI"],"author_name":"Ashwin Swarup","publish_date":"2023-03-20T15:00:00","publication_year":"2023","word_count":1403,"keywords":["data science","ChatGPT","machine learning","artificial intelligence","TPU","AI","Aim","analytics","generative AI","Generative AI","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","generative AI","ChatGPT","Aim","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-rise-of-generative-ai-and-living-content\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10106506,"title":"9 Gifting Ideas For Your Tech Bros","content":"Once again, it’s that time of the year when we enjoy spending time with loved ones, shopping for presents for them, and expressing our gratitude with these budget-friendly lists of gifts. Tech bros are no exceptions. Give your tech buddy or developer friend a present from this amazing list. The Data Is Calling, And I Must Go – Sweatshirt Are you trying to find the ideal gift for a coder? This is the perfect option: the Standard Crew Neck Sweatshirt. For programmers who appreciate both style and usefulness, this cozy and adaptable hoodie is ideal. Their passion for coding is evident in its straightforward and tidy design. This silky, long-lasting, and high-quality sweater will keep them warm throughout extended coding sessions. This sweatshirt is a useful and stylish present that will make any coder pleased, whether they wear it to work or on a casual stroll. Mouse Jiggler This Jiggler mouse should be on your present list if you want to offer a coder or developer something unique. When your loved ones receive it, they might feel inspired and happy. So let’s seize it right away. To keep your computer running while you are away from it, the Mouse Jiggler moves the cursor across your screen. It is a standalone device that doesn’t need any additional USB connections, software, or external power. This allows you to take time off from the computer without appearing on any employee tracking program. Simply and easily operated by resting your mouse atop the apparatus. Computer Programmer Steel Tumbler With this tumbler etched with programming, finding a fantastic gift for a programmer is not as difficult. Let’s get it so your favorite developer or coder is surprised. With its open port lid and dual walls, this vacuum-insulated stainless steel mug will maintain the ideal temperature for your beverages. The conventional lids have an open port drinking hole that is convenient for straws. They have sliding lids under their choices and add-ons area. The bottom of the 20- or 30-oz mugs is narrower to accommodate the majority of common cup holders. Men’s Digital Sports Watch Casual military outdoor design and simple minimalism fashion with a comfortable silicone rubber watch band make it a perfect present for your developer friend or coder. This waterproof watch is suitable for both indoor and outdoor sports, such as running, hiking, biking, fishing, climbing, and more. With an imported EL lamp, whether in the dark or under the sun, they can see the time clearly and find it easy to read. Cleaning Gel Universal Dust Cleaner For a coder or developer, a gift should be related to computer equipment, right? Then you can give your colleague this cleaning gel kit to clean the dust on the keyboard. This universal dust cleaner is made of biodegradable gel, not sticky to hands, smells sweet with a lemon fragrance, and causes no skin irritation. Ensure your hands are dry and clean before touching this gel. The keyboard cleaning gel can be used repeatedly until the color turns dark. Jet Performance Jet 15008 Performance Programmer If you are searching for a practical and useful present for someone who is a coder, try to have a look at this performance programmer. With three different performance tunes, it adjusts the correct speedometer for tire or gear changes. The JET performance programmer plus lets them take control and program their vehicle computer to match their driving style. It allows them to program for performance using lower-cost regular octane fuel or, for optimal performance gains, they can program for midgrade or premium fuels. Google Nest Wifi Router These make great gifts for hackers who love to grab their laptop and code from anywhere in the house. Does a coder in your life work from home? For anyone whose house has annoying wifi “dead zones,” Nest Wifi plugs into modems to provide up to 2200 square feet of strong, reliable internet. Bose Noise Canceling 700 Headphones Noise-canceling headphones are among the best gifts for programmers who need peace and quiet to get work done. These are some of the top noise-canceling headphones on the market, with 11 levels that let you decide whether to block all sound or allow ambient sounds. You’ll get up to 20 hours of battery life per charge and easy access to voice assistants like Alexa and Google Assistant. Spotify Premium Subscription Someone who loves both music and coding. Background music might actually improve our performance on cognitive tasks. By giving your techie friend the gift of music to program to, you could actually help them be more productive!","excerpt":"Once again, it’s that time of the year when we enjoy spending time with loved ones, shopping for presents for them, and expressing our gratitude with these budget-friendly lists of gifts. Tech bros are no exceptions.","categories":["Deep Tech"],"tags":[],"author_name":"Arya Vishwakarma","publish_date":"2023-12-26T15:00:00","publication_year":"2023","word_count":763,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","Git","RAG","R"],"extracted_tech_keywords":["AI","RAG","AWS","R","Go","Git","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/9-gifting-ideas-for-your-tech-bros-2023\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":20099,"title":"Myntra Working On Augmented Reality To Give Customers A 360-Degree Experience","content":"Very soon, when you shop on Myntra, a quick evaluation of your fashion quotient and options of latest fashion to elevate it, may be available at your disposal. Myntra is working on the addition of an augmented reality (AR) feature to its app to help its customers stay up-to-date with the latest fashion trends. When operational, the feature will be able to use the camera on the customer’s phone and critique their fashion choices. “It [the app] uses the camera of the phone, figures out what you are wearing and then provides a rating. It is for the fashion-conscious audience who would want to wear the latest trends and be seen as fashionable,”Ambarish Kenghe, Chief Products Officer of Myntra, was reported saying. The Flipkart-owned fashion platform has already embraced other growing technology like AI. Its AI-powered ‘fast fashion’ brands, Moda Rapido and Here and Now,  have already found success. Augmented reality has great potential in the area of fashion,especially in the e-commerce sector. The use of AR may be able incorporate the concept of a traditional trial room into to an online environment and help customers choose the best fashion and fit from the comfort of their homes. The DressingRoom app is a good example of this.The app was launched earlier this year by Gap.Inc to help its customers ‘try’ its clothes on virtually. Based on vital measurements such as height and weight, the app creates a 3D mannequin that mimics the entered measurements. It provides a 360 degree view of the model and shows how a chosen a garment will look on it. The app was created in association with Google and Avametric. Flipkart, which owns Myntra, faces the stiffest competition in the online fashion category from Amazon. In order to consolidate its position in the market, Flipkart acquired Myntra in 2014. And in 2016, Jabong was acquired. It recently was reported that Myntra was contemplating  launching its multi-brand offline stores. With the domestic online fashion market in India set to be valued at $12-$14 billion by 2020, the adoption of an AR-enabled feature to serve customers better could be a game changing and differentiating factor.","excerpt":"Very soon, when you shop on Myntra, a quick evaluation of your fashion quotient and options of latest fashion to elevate it, may be available at your disposal. Myntra is working on the addition of an augmented reality (AR) feature to its app to help its customers stay up-to-date with the latest fashion trends. When […]","categories":["AI News"],"tags":["ar","augmented reality","e-commerce","e-Commerce India","Flipkart","myntra","Online shopping"],"author_name":"Jeevan Biswas","publish_date":"2017-12-26T10:33:55","publication_year":"2017","word_count":357,"keywords":["Go","API","e-commerce","programming_languages:R","AI","augmented reality","e-Commerce India","Flipkart","programming_languages:Go","ar","myntra","Online shopping","ViT","R"],"extracted_tech_keywords":["AI","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/myntra-augmented-reality-app\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":22244,"title":"Women In New Tech: Pooja Sharma Of ZAPR Media Labs Talks About Need For Gender Diversity","content":"Pooja Sharma, Senior Data Analyst at ZAPR Media Labs. The scant number of women in new tech, especially in the areas of data science, analytics, and artificial intelligence has been a worrying trend for organisations all over the world. The resultant sexism is increasingly becoming one of the side-effects, making the global protests for gender equality so much more necessary. In fact, only about 12 to 15 percent of the engineers who are building the internet and its software are women. Therefore only seven percent of partners at top 100 venture capital firms across the world are women. These numbers are even more abysmal for the Indian new tech sector. Why is it so? And what can be done to change it? Analytics India Magazine is featuring women leaders in these sector for all of March celebrating Women’s Day — for the entire month. Pooja Sharma, Senior Data Analyst at ZAPR Media Labs What does the career in analytics\/data science looks like for a woman? It is about discovering and communicating meaningful insights or patterns in data available. It is exciting and challenging. We work closely with senior management to interpret data and help them take informed business decisions. Why did you choose this field as a career option? I was working as a full stack web developer when analytics as a career seemed interesting to me. You get to think about complex business problems, interpret the data available and be involved in key business decisions. You use the huge unstructured data to tell a story. You get to see how the decisions that you make impacts the organisation. This gives me a high and that’s why I chose to be a part of this. How is your growth story so far? I have been working with ZAPR for past two and a half years and it has been one roller coaster journey, starting from scattered data sources with unstructured data to a single stop data platform for all our data extraction requirements. I was working on different tech for extracting, cleaning, manipulating data. I had to learn different things on the go which was the fun. My computer science background helped me a lot to understand and think about data problems more efficiently. Over the period of time, I have tried to automate almost all of the data extraction and cleaning tasks, which helps to have more time and energy to focus more on deriving meaningful insights from the data. Do you struggle to maintain a work-life balance? Honestly, initially I was struggling to maintain a work-life balance as I was entirely concentrating on my work. It happens with everybody when they start something new and are totally focused on it. But later I figured it out. Time management is the key. Once you figure out how much time you need to allocate to your work and for yourself then everything works out perfectly. Your thoughts on incorporating more women in new tech sectors. We definitely need more women in tech sectors, as it is amazing and inspiring seeing women doing great in technology where female representation has been low since decades. Women make half of the users of the technology and now that our day to day life is so much dependent on technology, we need gender diversity in workplaces to arrive at better solutions. What are the key changes in education\/career choices that could change the current scenario? I think lack of encouragement, fear of harassment and lack of role models are the key factors for low participation of girls in higher and professional education. If we can provide good role models, a safe environment to work and teach them right professional skills, women tech careers can be changed for better.","excerpt":"The scant number of women in new tech, especially in the areas of data science, analytics, and artificial intelligence has been a worrying trend for organisations all over the world. The resultant sexism is increasingly becoming one of the side-effects, making the global protests for gender equality so much more necessary. In fact, only […]","categories":["AI Features"],"tags":["international women's day","Interviews and Discussions","Women's Day"],"author_name":"Prajakta Hebbar","publish_date":"2018-03-05T06:40:00","publication_year":"2018","word_count":624,"keywords":["data science","Go","API","artificial intelligence","AI","venture capital","RAG","international women's day","analytics","Women's Day","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","RAG","R","Go","API","GAN","venture capital"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/womens-day-zapr-media-labs-pooja-sharma\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10065602,"title":"Top 10 tips to ace ML hackathons","content":"The global AI software market cap is predicted to reach around 126 billion US dollars by 2025, according to Statista. The exponential technologies are shaping the modern job market and creating new jobs in the process. Of late, hackathons have become a big part of tech companies’ hiring strategies. And for good reasons. Source: Statista TAIKAI claimed that 40% of their hackathon participants were hired by companies in a matter of months. AI\/ML hackathons like MachineHack, Kaggle, NeurIPS, etc are the best platforms to network with industry experts, collaborate with peers and get recruited by large tech companies: Such hackathons have high visibility and credibility. We put together a list of key skills required to crack AI hackathons: Strong basics: Good fundamental knowledge in subjects such as programming language, mathematical concepts, machine learning methods, deep learning, etc is a necessity for such competitions. Rajat Rajan, a data scientist at TheMathCompany and a MachineHack grandmaster, said: “I guess the prerequisites were pretty simple for me. But, of course, it is always Python at the start. But then, for any ML hackathon, it comes down to good domain understanding. Then, dive deep into the sklearn package for error metrics, model algorithms, cross-validation etc. Most importantly, know how to understand data, train and validate.” Get hands-on experience: Bookish knowledge will only take one so far. Working on projects where one can apply the concepts taught in a book or a class is more effective than reading a book. As per Mobassir Hossen, the first Kaggle grandmaster from Bangladesh, one should not focus heavily on MOOCs or books, but rather spend more time on hands-on work and stay up to date with the latest research. Hyperparameters vs ideas: In a time-based challenge, it’s often easy to lose track of time focussing on the tuning of hyperparameters of an ML model. Instead, the participant should spend more time implementing new ideas based on the EDA and latest data to improve their models. Designing a strong validation strategy: A proper validation strategy can be the difference between winning and losing. Defining it is more complicated than cross-validation or holdout folds. One must always run tests on the test set variables distribution and construction against the leaderboard to ensure the correct local validation strategy is used. Time is of the essence: It is important to plan the model by taking the timeline into account. It is very easy to lose track of time when focusing on tuning hyperparameters or running cross-validation tests, etc. Adhering to a strict schedule will make sure that you finish your project on time. Explore-collaborate: Hackathons provide an overview of the talent pool present in the community. One must explore new possibilities, learn more about what’s trending and collaborate with fellow participants to come up with out-of-the-box ideas. The importance of feature engineering: Feature engineering is the process of extracting new data from existing data. It is one of the most important aspects of an AI hackathon as the performance of your model depends on the quality of the dataset used to train the model. Perseverance is key: Although not impossible, you are less likely to win a hackathon in the first go. You must be patient and learn from the competitions, accrue practical knowledge and develop a portfolio to reach a competitive level. Follow the grandmasters and engage in forums: Engaging regularly in the hackathon forums will bring you up to speed on the cutting-edge techs, tools and approaches. Following grandmasters and picking their brains will give insights into their game plans; what worked for them and what did not.Keep evolving: Adaptability is key to ace hackathons. The participants have to roll with the punches and be anti-fragile to overcome minor setbacks. Make sure you have a time-critical approach and a solid plan that account for untoward events. Learn from the mistakes, and develop a robust approach to tackle challenges.","excerpt":"TAIKAI claimed 40% of their hackathon participants were hired by companies in a matter of months.","categories":["AI Trends"],"tags":["Hackathon","Machine learning hackathon"],"author_name":"Kartik Wali","publish_date":"2022-04-24T13:00:00","publication_year":"2022","word_count":648,"keywords":["Go","API","machine learning","AI","Machine learning hackathon","ML","RAG","Hackathon","Python","Aim","deep learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","Aim","RAG","Python","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-tips-to-ace-ml-hackathons\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10138089,"title":"AMD&#8217;s Newfound Obsession with ‘Feeding the AI Beast’","content":"At Advancing AI 2024, AMD recognised the explosive growth of AI models and the strain it places on both network and computational infrastructure. The company showed its focus on ensuring that every component—from CPUs, GPUs, and DPUs to networking infrastructure—is tailored to fuel AI workloads efficiently and at scale. AMD’s innovations aim to eliminate any barriers that prevent GPUs from operating at their full potential, enabling faster training and inference cycles, and ultimately creating an ecosystem where AI models can grow without limits. “AMD is the only company that can deliver the full set of CPU, GPU, and networking solutions to address all of the needs of the modern data center. And, we have accelerated our roadmaps to deliver even more innovations across both our Instinct and EPYC portfolios, while also working with an open ecosystem of other leaders to deliver industry-leading networking solutions,” said Lisa Su, chair and CEO at AMD. AMD’s senior vice president of networking, Soni Jiandani, emphasised the growing demands of AI, stating, “Networking is not only critical, but it’s foundational to drive optimal performance.” With the launch of the UEC-ready Polara 400 AI networking adapter, AMD promises to elevate performance by ensuring seamless GPU communication across AI clusters, meeting the ever-growing demands of modern AI workloads. This innovation, along with other announcements such as the third-generation Selena DPU, positions AMD to lead the computing and networking industries to the next level, offering up to a six-fold performance improvement in AI training times. “The explosive growth of AI workloads requires innovations across GPUs, CPUs, and networking. AMD is innovating on all fronts to address this growth with our products like Polara 400 and Selena DPU.” said Jiandani. Polara 400 AI Networking Adapter: This new networking card is the first UEC-ready AI adapter designed to improve GPU communication, reducing congestion and improving job completion time by intelligently rerouting data across optimal paths​ Selena DPU: AMD announced the release of its third-generation DPU, the Selena, which offers 400 GB throughput, enabling AI workloads to accelerate with better security, load balancing, and congestion management​ “To truly feed the AI beast, we must innovate across every component—GPUs, CPUs, and networking. The Polara 400 and Selena DPU are designed to handle the immense scale and speed required by AI workloads, ensuring the AI beast is always running at its full potential,” said Jiandani. Forrest Norrod, EVP and GM at data center solution business at AMD, said that both Selena and Polara will be available early next year. Rajagopal Subramaniyan, SVP of networking at Oracle Cloud Infrastructure (OCI) said that it is leveraging AMD’s Elba DPU for high performance and scalability in their AI workloads. “The programmable architecture of Elba of the DPUs has helped us leverage the flexibility to launch features and services for our customers with software-like agility, while maintaining line rate at the hardware. The Elba delivered 5x improvement in performance… which was critical for us to cater to the demands of high-performance requirements of some of our key customers,” he added, looking forward to using Selena and Polara. IBM Cloud’s Ajay Apte said that it has adopted AMD’s DPU for virtualised environments, improving security and performance for enterprise AI and cloud workloads. “By moving our crown jewel, which is the SDN stack, into the DPU, we are essentially increasing the security aspect of our network… with AMD Pensando’s programmable DPUs, we are now running our virtualised machine offerings on a more secure and performance-optimized platform,” he added. Builds the Next-Gen AI Smart Switch At Advancing AI, Cisco, Microsoft, and AMD announced that they are joining forces to develop a next-generation AI Smart Switch that integrates AMD’s programmable DPU into data center switches. This collaboration aims to offload tasks such as security, load balancing, and network traffic management from CPUs and GPUs, significantly improving the performance of AI networks. Microsoft Azure general manager and head of products, Narayan Annamalai highlighted the importance of this development, stating, “We are building an extended network… using DPUs to offload the workload and apply security policies before traffic hits storage or services.” “Hypershield is our AI-driven, next-generation security architecture that runs on top of AMD’s Pensando DPU inside Cisco servers,” added Cisco’s senior vice president, Jeremy Foster. This partnership brings AMD’s DPU technology into mainstream networking, delivering faster, smarter, and more secure AI infrastructures. By embedding AMD’s DPUs into Cisco switches and Microsoft’s infrastructure, these companies are ensuring that AI networks can scale efficiently and handle the demands of emerging AI workloads. “Instead of having everything done within the compute servers, we can bring that into the switch over time and implement those technologies… This capability becomes just another capability in your networking fabric across all the switches you are deploying,” said SVP and general manager of datacenter and provider connectivity at Cisco, Kevin Wollenweber. Bets Big on Ethernet AMD is betting big on Ethernet as the foundation for its AI vision, positioning the technology as the most scalable, cost-effective solution for the growing demands of AI workloads. “Ethernet is clearly the preferred choice for back-end and front-end networks for AI workloads… delivering over 50% cost savings and scalability advantages,” avered Jiandani. For instance, AMD’s new Polara 400 AI networking adapter and its third-generation Selena DPU are at the heart of this Ethernet-driven AI infrastructure, enabling faster data transmission, congestion management, and enhanced scalability across massive GPU clusters. As AI workloads continue to grow exponentially, AMD’s Ethernet solutions provide a clear advantage over competitors, allowing for seamless scaling beyond the limitations of traditional architectures like Infiniband. “The foundational architecture of Infiniband is not poised to scale beyond 48,000 nodes without making dramatic and highly complex workarounds, whereas Ethernet has really proven itself to scale to millions of nodes, delivering huge scalability advantages,” shared Jiandani, ahead of Advancing AI 2024. AMD believes that its approach to AI networking based on Ethernet helps its customer scale AI clusters, reducing network congestion, and ensuring seamless communication between GPUs. Polara 400 and Selena DPU: These products form the core of AMD’s AI networking strategy, using Ethernet to handle increasing data traffic without the complexity and cost associated with alternative networking solutions​(AMD Pensando Press Deck). AMD’s Ethernet-based approach stands in contrast to competitors like NVIDIA, which relies on solutions like Spectrum-X that incorporate both NICs and switches. By focusing on Ethernet, AMD looks to deliver cost-effective, scalable AI systems without needing additional proprietary hardware. Darrick Horton, CEO of TensorWave, said that it is building large AI clusters with AMD’s Polara Ai networking adapters to overcome networking congestion challenges.  “With traditional Ethernet, we face challenges with congestion and flow management… Polara Ethernet will allow us to build larger clusters and improve workload efficiency, specifically around job completion times and overall hybrid utilization,” he added. Going beyond, AMD is also leading the Ultra Ethernet Consortium (UEC), a coalition of 97 industry-leading vendors, which aims to standardise Ethernet protocols for AI networks, creating a robust ecosystem for high-performance AI and cloud computing. The Future of AI Systems with AMD’s P4 Engine At Advancing AI 2024, AMD also spoke about revolutionising AI systems with its fully programmable P4 engine, designed to accelerate AI networking and offload critical tasks from GPUs and CPUs. Interestingly, this innovative engine is at the core of AMD’s Polara 400 and Selena DPU, providing AI systems with the flexibility and performance required to manage the vast amounts of data generated by modern AI workloads. “At the core of this solution powering innovation for both AMD and our customers is our fully programmable P4 engine, which enables us to adapt to the evolution of Ethernet and accommodate AI networking needs,” said Jiandani. AMD’s P4 engine supports 400-gigabit line-rate throughput and can scale AI networks to handle millions of GPUs, delivering unmatched scalability. “Our fully programmable P4 engine puts AMD in a very unique position because we now have the ability to deliver a holistic portfolio that is future-proof and provides end-to-end network solutions through full programmability,” added Jiandani, saying how P4 engine is poised to lead the next generation of AI systems, providing AI workloads with the flexibility to evolve alongside emerging industry standards, outpacing competitors in both performance and adaptability.","excerpt":"“To truly feed the AI beast, we must innovate across every component—GPUs, CPUs, and networking.”","categories":["AI Trends"],"tags":["AMD"],"author_name":"Siddharth Jindal","publish_date":"2024-10-10T23:04:42","publication_year":"2024","word_count":1362,"keywords":["Go","AMD","AI","cloud computing","R","ML","Scala","RAG","Ray","Aim","Azure"],"extracted_tech_keywords":["AI","ML","Aim","Ray","RAG","cloud computing","Azure","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/amds-newfound-obsession-with-feeding-the-ai-beast\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10076851,"title":"Shiv Nadar Institution of Eminence, Delhi-NCR launches Analytics Olympiad 2022 for Data science professionals","content":"The Academy of Continuing Education at Shiv Nadar Institution of Eminence Delhi-NCR, in partnership with MachineHack, is launching the second edition of its annual Analytics Olympiad, from 30th September 2022, to 6th November 2022, for data scientists and machine learning professionals. This two-month-long championship has been primarily designed to strengthen the data science community in India and pave the way for innovation. This challenge is the perfect opportunity for data science enthusiasts, learners and professionals in India to showcase their skills and leadership potential in business analytics and advance their careers. Problem statement and description The vehicle insurance business is a multi-billion dollar industry. Every year millions and millions are paid in premiums, and a huge amount of claims pile up. At the Analytics Olympiad 2022, the data science and machine learning community would step into the shoes of a data scientist and create an ML model that would help an insurance company understand which insurance claims should be accepted for reimbursement and which must be rejected. The participants would be given a rich dataset consisting of thousands of rows of past records to learn more about customers’ behaviours. Dataset Description: Columns: [‘ID’, ‘AGE’, ‘GENDER’, ‘DRIVING_EXPERIENCE’, ‘EDUCATION’, ‘INCOME’, ‘CREDIT_SCORE’, ‘VEHICLE_OWNERSHIP’, ‘VEHICLE_YEAR’, ‘MARRIED’, ‘CHILDREN’, ‘POSTAL_CODE’, ‘ANNUAL_MILEAGE’, ‘SPEEDING_VIOLATIONS’, ‘DUIS’, ‘PAST_ACCIDENTS’, ‘OUTCOME’, ‘TYPE_OF_VEHICLE’] Learn and predict the OUTCOME variable. Start Date: September 30, ‘22 End Date: November 06, ‘22 Click here to participate in the hackathon Prizes: Winner: INR 1 Lakh 1st Runner-up: INR 30,000 2nd Runner-up: INR 20,000 ** Note: Analytics Olympiad 2022 will be held in two phases. In the first phase, the participants will be allowed to submit their approaches against the given problem statement, which they have to solve based on the dataset provided on the MachineHack platform. The evaluation of the first phase will be done based on the leaderboard, and the top 10 participants will be given a chance to participate in the second phase, i.e. the Jury Round. The jury round will be helmed by a group of expert panellists from the industry and academia. The participant who will crack phase two of the challenge will be the Analytics Olympiad and be awarded INR 1 Lakh cash. In addition, the second & the third winner will be granted INR 30,000 & INR 20,000 each, respectively. To understand the rules of the hackathon, click here. Winners will be judged based on the following criteria: 30% Business Outcome\/Impact 20% Innovative + Creativity 20% Algorithm and ML approach, 15% Statistically analysis 15% Presentation + Communication Start Date: September 30, ‘22 End Date: November 06, ‘22 Click here to participate in the hackathon About Shiv Nadar Institution of Eminence, Delhi-NCR Shiv Nadar Institution of Eminence is a student-centric, multidisciplinary and research-focused university offering a wide range of academic programs at the Undergraduate, Master and Doctoral levels. The University was set up in 2011 by the Shiv Nadar Foundation, a philanthropic foundation established by Mr. Shiv Nadar, founder of HCL. In the NIRF (Government’s National Institutional Ranking Framework), the University has been the youngest institution in the ‘top 100’ overall list. The university’s Academy of Continuing Education aims to facilitate best-in-class knowledge, practices and skill development offerings to the growing ecosystem of lifetime learners and leaders, both within and outside the university. With distinguished academics as the university’s faculty members and programme instructors, the Academy of Continuing Education offers uniquely crafted programmes that are delivered innovatively, bringing together the best of the university’s rich intellectual resources. The university aims to help students prepare for today as well as their future through its unique certification programme in data sciences and business analytics. The collaboration between the Academy of Continuing Education at Shiv Nadar Institution of Eminence and MachineHack hopes to strengthen the data science community in India and pave the way for innovation in business analytics. If you are interested in data science and analytics certification programmes, click here. Start Date: September 30, ‘22 End Date: November 06, ‘22 Click here to participate in the hackathon","excerpt":"Shiv Nadar Institution of Eminence, Delhi-NCR, invites data scientists, ML developers and AI enthusiasts to showcase their data & analytics skills by participating in the second edition of the ultimate Analytics Olympiad.","categories":["Deep Tech"],"tags":["analytics olympiad"],"author_name":"Tasmia Ansari","publish_date":"2022-10-10T13:00:00","publication_year":"2022","word_count":664,"keywords":["data science","Anthropic","analytics olympiad","machine learning","Go","AI","ML","Aim","ViT","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Anthropic","Aim","R","Go","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/2nd-edition-of-analytics-olympiad-is-live-now\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":60219,"title":"IBM’s AI Learns To Generate New Footage From Video Stills","content":"In a recent paper released by IBM, the company describes an AI system that generates videos that are seen during training, as well as the videos that are unseen. While Navsynth, the AI system, isn’t a novel idea, it’s indeed a deep interest for many companies like DeepMind and others. According to the researchers, this effort will produce superior quality videos compared with existing methods, which could be used to synthesise videos on which other AI systems train — supplementing real-world data sets that are incomplete or marred by corrupted samples. Explaining further, the researchers stated, “the bulk of work in the video synthesis domain leverages GANs, or two-part neural networks consisting of generators that produce samples and discriminators that attempt to distinguish between the generated samples and real-world samples. They’re highly capable but suffer from a phenomenon called mode collapse, where the generator generates a limited diversity of samples (or even the same sample) regardless of the input.” In contrast, IBM’s system consists of a variable that represents video content features, a frame-specific transient variable, a generator, and a recurrent machine learning model, which breaks videos down into a static constituent that captures the constant portion of the video common for all frames and a transient constituent that represents the temporal dynamics between all the frames in the video. Alongside, IBM’s system learns the static and transient constituents together, which it uses to generate videos at inference time. In this paper, IBM’s researchers proposed a novel non-adversarial framework to generate videos in a controllable manner without any reference frame. Specifically, they proposed to synthesise videos from two optimiSed latent spaces, one providing control over the static portion of the video (latent static space) and the other over the transient portion of the video (transient latent space). The researchers proposed to jointly optimise these two spaces while optimising the network (a generative and a recurrent network) weights with the help of regression-based reconstruction loss and a triplet loss. Our approach works as follows — the research team trained, validated, and tested the system on three publicly available data sets: Chair-CAD, which consists of 1,393 3D models of chairs (out of which 820 were chosen with the first 16 frames); Weizmann Human Action, which provides ten different actions performed by nine people, amounting to 90 videos; and the Golf scene data set, which contains 20,268 golf videos (out of which 500 videos were chosen). Compared with the videos generated by several baseline models, the proposed method produces visually sharper and consistently better results using the non-adversarial training protocol. Moreover, it reportedly demonstrated a knack for frame interpolation or a form of video processing in which the intermediate frames are generated between the existing one in an attempt to make animation more fluid.","excerpt":"In a recent paper released by IBM, the company describes an AI system that generates videos that are seen during training, as well as the videos that are unseen. While Navsynth, the AI system, isn’t a novel idea, it’s indeed a deep interest for many companies like DeepMind and others. According to the researchers, this […]","categories":["AI News"],"tags":["DeepMind","DeepMind AI","IBM","learn ai"],"author_name":"Sejuti Das","publish_date":"2020-03-27T16:36:54","publication_year":"2020","word_count":461,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","programming_languages:Go","RAG","IBM","GAN","DeepMind AI","learn ai","R","DeepMind"],"extracted_tech_keywords":["AI","machine learning","neural network","RAG","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibms-ai-learns-to-generate-new-footage-from-video-stills\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10113275,"title":"Now Open Source Projects Can Make Money","content":"Open source development has been the reason for the rapid growth of tech. Yann LeCun, the famous proponent of open source projects, takes every opportunity to elaborate on how vital open source development is. The sustainability of open source projects, however, is dependent on the financial returns it sees over time. “There is a lot of unnecessary friction today to sponsor specific features, issues or milestones for open source projects,” said Birk Jernström, the founder of Polar. The company, founded in 2022, is a platform that manages the subscriptions and payments for people who create and support open-source software. It also offers tools for working with data. Many open source projects start with being freely available and eventually seek funding. Red Hat for example, known for its Linux distribution, monetised by selling subscriptions for technical support, updates, and training to businesses. This model helped fund continuous open-source development while providing enterprise-level services. Alternatively, Blender, a 3D creation suite, supports its development through the Blender Development Fund, donations, and paid services like professional training and Blender Cloud subscriptions. For smaller projects, platforms like Patreon or Open Collective let supporters donate monthly or per project. GitHub Sponsors allows direct donations to developers. These models rely on voluntary support, which may not match the actual effort needed for development. However, Polar takes it a step further and allows funding for specific features, issues, or milestones. This drives the project in the direction that is valued by the customers. It motivates the developer who would know they’ll get paid for hitting clear goals. Jernström clarified the difference from existing funding platforms, pointing out, “There is no one-size-fits-all solution to this and that’s what we want to build, one platform for multiple solutions.” Polar is changing open-source funding GitHub is keen on giving developers the options to choose how they want to monetise their work. In 2019, they launched GitHub sponsors, but as one user pointed out, it is nothing more than ‘coffee money’ between persons. Polar, according to Jernström, gives maintainers the option to be ‘entrepreneurs’. There have been donations in the past with platforms like Open Collective, and Stack Aid, among others, allowing individuals and companies to pledge financial support directly towards specific issues or feature requests in open-source projects. Polar intends to go beyond this ‘coffee money’ funding and provide a steady stream of income. Jernström explained, “Donations and sponsorships are great when they happen. Problem is, they rarely do. In order to drive meaningful (full-time work) capital to OSS initiatives, I believe it has to charge for add-on value and that such services and subscriptions are mutually beneficial.” Polar facilitates the sale of add-on services, subscriptions, or premium features, and inturn maintainers can craft offerings that align with their project’s goals and community’s needs. The platform takes a 10% commission including the 5% charges for Stripe transactions. “As an ecosystem, we should be focused on how we can get 10x, 100x and then 1000x funding. Five percent of nothing is nothing. That’s the real problem in OSS today. Let’s fix that first,” he said. This could include anything from offering paid support, consulting, custom development work, access to premium features, or early access to new releases. This is an upgrade from voluntary support to making it easy for backers to financially support the issues and features they care about. By handling the financial transactions, tax considerations, and potentially even compliance issues, Polar lets developers focus on working on the projects itself. Ease and transparency Transparency has always been very important to open source funding. For example, the open collective for example is designed around transparency, with all financial transactions visible to the public by default. Expenses, income, and budgets are tracked and displayed on the platform, and contributors can see how funds are used and allocated. Polar goes the same route and has complete control over which issues or features they want to highlight for funding through the platform. This ensures that they can align any external funding with their project’s roadmap and priorities. Maintainers can set goals for funding specific initiatives within their projects, providing clarity to potential backers about what their contributions will support. Andreas Kling, a key contributor to the SerenityOS and Ladybird who uses Polar for funding, said, “We’ve been using Polar for funding GitHub issues for a couple of months now, and it always makes me super happy when I see someone collect a reward!” SerenityOS is a Unix-like OS with a classic desktop interface and user-friendly design, supported by an active developer community. Ladybird is its companion lightweight web browser, offering fast and secure browsing seamlessly integrated with the OS. Kling added, “I’m super happy to see Polar take on the task of becoming a Merchant of Record and abstracting away much of the complexity for all developers.” By addressing these critical and often overlooked aspects of open-source project maintenance, Polar is setting a precedent for how platforms can support the sustainable development of open-source software.","excerpt":"Polar emerges as a game-changer in open source funding, giving developers new ways to monetise their work beyond traditional donations.","categories":["AI Highlights"],"tags":["AI projects","Open Source AI"],"author_name":"K L Krithika","publish_date":"2024-02-20T16:00:27","publication_year":"2024","word_count":831,"keywords":["Go","API","funding","AI projects","programming_languages:R","AI","ML","Git","Open Source AI","ViT","GitHub","R"],"extracted_tech_keywords":["AI","ML","R","Go","Git","GitHub","API","ViT","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/now-open-source-projects-can-make-money\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":34942,"title":"Understanding Dimensionality Reduction Techniques To Filter Out Noisy Data","content":"When machine learning classification problems are performed, there are various factors that are considered on the basis of which the final classification is done. These factors – fundamental variables are known as features. The greater the number of features, the harder it gets to envision the training set and then work on it. Sometimes, most of these features are related, and hence unnecessary. This issue can be addressed with dimensionality reduction algorithms. Dimensionality reduction is the process of reducing the number of random variables under study, by collecting a set of principal variables. It can be classified into feature selection and feature extraction. Feature Selection In this process, we try to identify a subset of the primary set of variables, or features, to get a modest subset which can be used to illustrate the problem. Feature extraction In this process, the data is reduced into a high dimensional space to a profound dimensional space. Methods for Dimensionality Reduction Dimension reduction or turning a group of data having immense dimensions into data with subordinate dimensions with effective concise information can be achieved by using various methods. Principal Component Analysis (PCA) Principal Component Analysis (PCA) is a dimension-reduction mechanism that can be used to overcome a large set of variables to a small set that still contains most of the information in the large set. In this procedure, correlated variables are transformed into a number of uncorrelated variables termed as principal components. The original principal component accounts for the variability in the practicability data, and each succeeding component values for as much of the outstanding variability is possible. A principal component analysis can be considered as rotation of the axes of the original variable coordinate system to new orthogonal axes, called principal axes, such that the new axes coincide with directions of maximum variation of the original observations. Linear Dimensionality Reduction (LDA) Linear Discriminant Analysis (LDA) is a technique used for supervised classification problems. Linear Discriminant Analysis is a dimensionality reduction technique used as a preprocessing level in Machine Learning and pattern classification applications. Linear Discriminant Analysis takes labels into consideration. This level of dimensionality reduction is used in biometrics, chemistry and many more. The primary motive of LDA is to calculate the characteristics in higher dimension space onto a lower dimensional space. The process starts by calculating the separability between various classes also termed as between-class variance. Once the class variance is obtained we need to determine the distance between the mean and sample of every class, which is called within class modification, followed by construction of lower dimensional space which maximises the value between class variance and minimises the within-class variance. Generalised Discriminant Analysis(GDA) The GDA technique applies the methods of the general linear model to the discriminant function analysis problem. In GDA, the discriminant function analysis problem is termed as  “recast” which is a general multivariate linear model, where the conditional variables of a class are coded vectors that indicate the group membership of each case. The remainder of the analysis is then produced as described in the context of General Regression Models (GRM), with a few additional characteristics. Defining standards for predictor variables and predictor effects. Stepwise and optimal-subset analyses. Value profiling of succeeding classification probabilities. Advantages Of Dimensionality Reduction Dimensionality reduction has a host of advantages from a machine learning point of view Since the model has smaller degrees of freedom, the possibility of overfitting is lower. The model will generalise more easily on new data If user applies feature selection or linear classifications (such as PCA), the conversion will promote the most related variables which will improve the interpretability of the model Most of features extraction procedures are unsupervised. The user can encourage the autoencoder or fit a PCA on unlabeled data. This can be really effective as the user will have a bunch of unlabeled data and labelling is time-consuming and expensive","excerpt":"When machine learning classification problems are performed, there are various factors that are considered on the basis of which the final classification is done. These factors – fundamental variables are known as features. The greater the number of features, the harder it gets to envision the training set and then work on it. Sometimes, most […]","categories":["Deep Tech"],"tags":["model","overfitting","PCA","stepwise regression machine learning"],"author_name":"Bharat Adibhatla","publish_date":"2019-02-14T05:55:16","publication_year":"2019","word_count":647,"keywords":["Go","machine learning","programming_languages:R","AI","R","programming_languages:Go","RAG","model","stepwise regression machine learning","PCA","overfitting"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/understanding-dimensionality-reduction-techniques-to-filter-out-noisy-data\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10165932,"title":"Coding Interviews are Becoming a Joke","content":"A Reddit user shared that after spending three months in the free pool with no projects, he began job hunting, only to find that every Java backend role required over four years of experience. Meanwhile, at his current company, Wipro, he was required to take unproctored competency tests. Despite the challenge, he was able to complete these tests successfully using online resources. However, during his final proctored test, he confronted an unfamiliar IDE which lacked the code-completion features of IntelliSense. To tackle the problem, he used his local integrated development environment (IDE) before submitting the test. Days later, HR accused him of malpractice. Despite his detailed explaination on the matter, they refused to reconsider their decision and forced him to resign immediately without a second chance or even a proper notice period. This incident raises an important question: If companies are so stringent about proctoring, how do some candidates manage to clear interviews using AI? Recently, a LinkedIn user highlighted how AI-powered tools are helping candidates breeze through coding interviews. “Are coding interviews becoming a joke? Just came across an AI tool that…can auto-hide when screen sharing and stay invisible, generate natural reasoning for ‘your’ solution, and simulate real eye movements to bypass monitoring. And guess what? It’s open-source,” he said. His concern was clear: If AI can ‘clear’ coding rounds for candidates, are companies truly assessing real skills? Even if one AI tool is blocked, another will surface. So what’s next? Should companies rethink hiring strategies altogether, move to live pair-programming, go back to whiteboard interviews or let AI interview AI? In response, a user on LinkedIn with three decades of coding and over 25 years of teaching experience, said that Leetcode, Codepen or Hackerrank are simply not good enough. “If you’re hiring and you cannot give me two hours of your time for a coding session, or you don’t trust me enough to hand me an email challenge…Why should I commit to you?” This sentiment resonates with many in the developer community, who feel that technical interviews have become more of a formality than a true test of skills. Hear it From Experts In a conversation with AIM, Pratham Patel, a member of Rocky Enterprise Software Foundation, shared his perspective. “As an interviewee, I can say…that the interview process has become more of a formality than an actual test. It is unfortunate to see that interviewers are more interested in whether the candidate can provide code with flawed reasoning rather than understanding if they truly grasp programming and the art behind it.” Meanwhile, Krishna Vij, vice president of IT staffing at TeamLease Digital, pointed out that AI is pushing companies to rethink their hiring strategies. “How we assess technical talent is evolving and AI-driven tools are accelerating that shift. If a candidate uses AI to clear a coding test, we must ask ourselves: Are we truly measuring their skills or just their ability to leverage technology? This is why companies must rethink their hiring strategies, as standard coding tests alone are no longer sufficient in the current situation.” Vij added that the company is seeing a stronger push towards live coding interviews, project-based assessments, and in-depth problem-solving discussions. The focus now is on critical thinking, adaptability, and real-world application rather than just syntax and speed. Hiring processes must keep evolving because as AI gets smarter, so must the way companies evaluate talent. Similarly, Rahul Veerwal, founder and CEO of GetWork, reinforced this perspective and stressed the need for multi-layered assessments. “At GetWork.ai, we’ve seen firsthand how AI can enhance hiring, but we also recognise its potential for misuse. While AI-assisted cheating poses a challenge, the solution isn’t to abandon coding rounds but to evolve how we assess talent,” he explained. According to Veerwal, this is why GetWork believes the future of hiring lies in AI-proctored, dynamic assessments that test a candidate’s real problem-solving skills, not just their ability to recall syntax. Moreover, structured follow-ups like automated technical interviews can reveal a candidate’s practical knowledge beyond AI-aided solutions. “We also see immense potential for AI in levelling the playing field for job seekers, especially from tier-2 and tier-3 cities. Our GenAI copilot, Horizon AI, helps candidates prepare for interviews, upskill, and build confidence, without crossing ethical boundaries,” he said. He stressed that companies have to move beyond one-dimensional coding tests. Implement multi-layered assessments, use proctored environments, and, most importantly, assess problem-solving approaches over perfect code output. What’s Next? Only about 10% of Indian engineering graduates possess adequate coding skills, according to a 2019 report by Aspiring Minds. More recently, a study by TeamLease found that merely 5.5% of Indian engineers are qualified with basic programming abilities. Source: Statista Furthermore, the Equinix 2023 Global Tech Trends Survey revealed that 86% of Indian businesses are actively reskilling their IT workers to address the industry’s needs. This proves that there are challenges in coding skills among Indian engineers, and so there is a strong demand for skilled professionals in emerging technologies. Hence, engineers must to upskill to stay in the race.","excerpt":"If companies are so stringent about proctoring, how do some candidates manage to clear interviews using AI?","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","coding","Jobs"],"author_name":"Vidyashree Srinivas","publish_date":"2025-03-12T18:00:00","publication_year":"2025","word_count":839,"keywords":["Go","GenAI","TPU","AI","coding","ML","RAG","Aim","Rust","Jobs","R","Java","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","GenAI","Aim","RAG","TPU","R","Go","Rust","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/coding-interviews-are-becoming-a-joke\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35173,"title":"5 Data Science And AI Job Opportunities At Siemens India","content":"With 22 factories and 11 R&D centres, Siemens is determined to turn India’s untapped potential into continual progress. For around 90 years now, Siemens has been associated with the growth of emerging India. Be it modernising India’s infrastructure, enabling greener power generation or empowering hospitals with cutting-edge healthcare systems, Siemens in India is playing a major role in creating a sustainable future for the country. Here are few job openings at Siemens’ Bengaluru centre which require a variety of skills ranging from image processing, NLP, algorithm designing and data analytics: Lead Research Engineer – Decentralised and Intelligent Systems Design A lead research engineer will architect, design and develop prototypes and technology demonstrators with very high quality to solve contemporary and emerging critical business challenges. Requirements PhD Or M tech in Computer Science (or related fields) with relevant work experience of 2 – 7 years. C\/C++, C#, Java, Python, JavaScript, Golang, Groovy, Haskell Blockchain technologies: Ethereum, Hyperledger, Big Chain DB Understanding of cloud computing and cloud-native development methods like microservices architectures and associated patterns. Knowledge of modern DevOps methods and tooling. A good understanding of distributed and P2P algorithms including graph algorithms, non-blocking data structures, leader election, fault-tolerant consensus. Apply here Lead Research Engineer – AI & Industrial Data Analytics A data analytics research engineer will analyse large, complex data sets by developing advanced machine learning and deep learning pipelines based on business initiatives and the ability to build theoretical models as well as build innovative, practical and robust real-world solutions for problems in Oil & Gas, Manufacturing and Process Industries. And also demonstrate outstanding ability to drive innovation and research in the form of patents and publishing papers at top-tier conferences\/journals. Requirements An advanced degree in operations research, applied statistics, data mining, machine learning, physics, or a related quantitative discipline is an asset. Java\/Python\/C++\/Matlab Strong background in data science including statistical analysis and modelling, data mining, machine learning, uncertainty analysis, time series forecasting, working with both structured and unstructured data. Proven experience in oil & gas and power generation or manufacturing preferred. Apply here Research Engineer – ML for Simulation A research engineer will implement advances available in Artificial Intelligence, Machine Learning and Natural Language Processing for custom programs. And, interface with Siemens internal customers and provide AI\/ML\/NLP based solutions for their needs. Requirements Excellent programming skills, with 2 to 4 years of programming job experience. Flask\/Django, Servlet Program(JSP) \/PHP, JavaScript\/JQuery, AngularJS. Python and familiarity with libraries like NLTK\/Numpy\/MatPlotLib\/Keras\/Tensorflow Apply here Lead Research Engineer – Semantic Web & RDF A knowledge graph expert on our team you will have to research, develop and maintain tools used to architect, create, debug, analyse and maintain industrial knowledge graph. And, search for product data sources, to use your tools to integrate, disambiguate and keep consistent the entire knowledge graph structure. Requirements PhD\/Masters\/Bachelor’s degree in Computer Science with grass-roots experience of 2+years  in architecting data analytics and machine learning applications using knowledge graphs. Experience in semantic web related specifications and tools, knowledge graph consistency and coherency checking. Python, Java, basic HTML Experience in Experience with ETL processes and tools Experience of property graph formalism (e.g. Neo4j, Gremlin, Tinkerpop). Apply here Lead Research Engineer – Computational Imaging A computational imagine engineer will develop and optimise algorithms for detection and analysis of imaging features acquired using multi-spectral imaging technologies, conduct initial proof-of-concept experiments for assessing the feasibility of imaging device ideas. Requirements Post-graduate degree (PhD is preferred) in science\/engineering in a field related to image processing, pattern recognition and\/or computer vision 3+ years of relevant experience in algorithm development in at least 3 of the following areas: image segmentation, feature extraction, multi-spectral imaging,   analysis, signal processing, pattern classification, computational photography, image analysis C\/C++, Python Experience in Machine Learning is highly desirable Basic understanding of optics, lighting, and digital image acquisition is a plus Apply here","excerpt":"With 22 factories and 11 R&D centres, Siemens is determined to turn India’s untapped potential into continual progress. For around 90 years now, Siemens has been associated with the growth of emerging India. Be it modernising India’s infrastructure, enabling greener power generation or empowering hospitals with cutting-edge healthcare systems, Siemens in India is playing a […]","categories":["AI Hirings"],"tags":["devops journal","java data structures methods","Siemens"],"author_name":"Ram Sagar","publish_date":"2019-02-20T08:51:20","publication_year":"2019","word_count":639,"keywords":["java data structures methods","devops journal","data science","artificial intelligence","machine learning","AI","ML","computer vision","NLP","Siemens","deep learning","analytics","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","computer vision","data science","analytics","TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/5-data-science-and-ai-job-opportunities-at-siemens-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10444,"title":"Great Lakes Institute of Management in partnership with Great Learning introduces Mentored Learning in their Online Analytics Program","content":"Consistently ranked as one of the leading business schools in India providing analytics education, Great Lakes Institute of Management provides one of the most sought after online courses in the industry – the Business Analytics Certificate Program or BACP. The course is offered on the learning platform Great Learning and is best suited for working professionals who want to build their skills in analytics through the convenience of online learning. The course is designed for a duration of roughly 6 months, covers all the essential topics including R, SAS and Advanced Excel, which are required for a promising career in analytics. It has a designated 100 hours of self-learning through lecture videos and exercises along with 40 hours of mentored learning and 20 hours of assessments, which are available online. All the classes are delivered by faculty with over 15 years of hands-on experience in business analytics. Additionally, Great Lakes allows an access to their learning material for a period of 3 years even after completion of the course. The online Learning Management System (LMS) built by Great Learning hosts all the content, including online lectures, webinar recordings, reading material and assignments. Further, the learning platform also encourages collaboration between candidates, thus maximizing learning effectiveness. In the first batch itself, the BACP program has attracted professionals across various roles and industries. The batch consists of professionals from technology and consulting sectors looking to transition to analytics, professionals already working in analytics and looking to upskill themselves and also senior professionals looking to build their competencies in business analytics. With the objective of making the learning experience in the program closely map with the requirements of the analytics industry, Great Lakes has recently launched ‘Mentored Learning’ to all candidates pursuing BACP. What is Mentored Learning? The addition of this new feature as a part of BACP would facilitate industry mentorship to the enrolled students. During the entire duration of the program, the toughest part is to keep the passion alive, and the mentors would help maintain that. They would not only help candidates clear their concepts and application of concepts, but also encourage them to complete the course with much ardor. As the mentors would be available as and when the career advice is required by the student, it would enable the candidates to build their careers in analytics. Pairing students with suitable mentors from analytics can speed up their acquisition of professional knowledge and enhance their learning. It can be a great way for students to achieve their professional development goals as well. Mentors can help students make a sense of what they are learning and help them understand the professional contexts of the knowledge that they have gained. They definitely form a resource that will continually help students build their networks and take them a notch higher in their professions. Jatinder Bedi, one of the mentors for the program with an experience of more than 10 years in the analytics industry said “It’s a very important part of learning. Mentoring is more like a support system to participants. As one gets introduced to new concepts, participants may have a lot of questions, which they may sometimes be not able to ask during the session. Mentor alignment helps in getting these questions answered.” “Mentored initiative helps students get the maximum out of the course from knowing what to learn and where to implement what they learn”, said Vinol Joy, serving as a mentor for the program. “It would open up a channel where students could reach an industry ready mentor for any questions”, he said. Vinol is an analytics professional and has worked with big names like FlipKart, MuSigma and currently works as Business Intelligence Manager at Qraved. Further adding to the importance of mentored learning he added that “Learning without goals or objectives is like shooting an arrow in the dark. Speaking to a mentor and understanding what the industry uses, top trending tools and processes of industry helps the person chalk out an accurate picture of the industry. It also creates a confide in zone where a student can ask a mentor any question from any topic and not have the shame of asking a dumb question in front of an audience”. What are the Benefits of Mentored Learning? Mentored learning would help students in acquiring skills relevant to the industry and not carry any legacy skills taught in courses. It would help them see their value in teams which they would be a part of later much clearly. Clearly mentored learning has innumerable advantages, few of which have been listed as below: Enhance Employability: A mentor would pave way towards a stable career by equipping students with industry insights, knowledge and empower them to focus on professional goals. By associating a student to mentor from analytics would help students get a better understanding of their career path, hence enhancing their employability ratio. Enriched Learning Experience: Apart from clearing doubts quickly, mentors can guide them on what to read, where to find a particular information and how to get the best out of it. Networking: The program not only link students to experienced analytics professionals from the same field, but also allows to expand their network, thereby, opening a door to lucrative job opportunities in analytics. Better Course Completion rate: Continued guidance and encouragement from mentors would assist students accomplish assignments on time and help keep them on track with the syllabus. Mentors also provide intervention: In case of under performance by students, mentors continued intervention would encourage them to perform better. Celebration of Milestones: As a part of the program, if a student performs well, they are rewarded by celebrating the achievement in a group, thereby, keeping a zest alive to perform better the next time. “Mentored learning enables a student solve their queries easily, either through whatsapp ping or an email, facilitating a better learning and hence better experience”, commented Jatinder. How are the mentors being selected? Since, mentors form the basic inspiration for the students to excel well in the career, it becomes crucial to get just the right mentor for them. They form a support system for candidates in career exploration, professional development, and networking. Looking at these points, Great Lakes has selected mentors from analytics industry with considerable experience and exposure across leading organizations. And that’s what Great Lakes is doing and each mentor is assigned with 4-5 students to ensure best attention to the students. Apart from the industry experience that a mentor brings to the table, their willingness to share time, skills and knowledge with the BACP candidates is a crucial criterion for selecting the mentors. How does Mentored Learning work? “It works more like a Buddy program, where mentor is in touch with the group through most commonly used medium of whatsapp or email, where the group discussions happens. It comes out great as mentor helps student understand critical concepts as and when he\/she ask questions”, said Jatinder, one of the industry mentors in the BACP program. It’s mostly online support when needed and does not need physical presence of either mentor ‘or students. This way a lot of queries can be handled hence extending a support right at the time when a student is in need. That’s not all, in the case where no questions are coming in the group, mentor ignites the thought process by putting various analytics use cases in the group chat and then create a brain storming environment for new ideas & approaches. “This activity is very healthy for brain”, added Jatinder. What Students are saying? The whole concept of online and mentored learning has garnered a lot of appreciation and popularity in the analytics community which is evident from the long list of enrolment that the program has fetched. Let’s know it from the students what they have to say about the learning methodology adopted at BACP program. “So far, my experience in the BACP has been good. Whatever I have learned was new to me and it has been very helpful. The mentor is approachable. Anytime you have a question, you go and post it, you will get a quick response. We all keep posting our questions whenever we like. One has the record of the answers as well for future reference. “Krithika, Trident Techno Solutions “I am a Project Manager with 14 years of experience. One of my friends doing Great Lakes’ PGPBA referred BACP to me as I was keen to do a short-term course in Analytics. The experience in the program has been exceptionally good till now. The presence of a mentor in BACP is a very good initiative offering support in doubt clearing and application of concepts.” Moulesh, Technical Project Manager – Happiest Minds For any more info on the program and enroll for the upcoming batch in August, you can visit, http:\/\/bit.ly\/2a4HZOB","excerpt":"Consistently ranked as one of the leading business schools in India providing analytics education, Great Lakes Institute of Management provides one of the most sought after online courses in the industry – the Business Analytics Certificate Program or BACP. The course is offered on the learning platform Great Learning and is best suited for working […]","categories":["AI Trends"],"tags":["mu sigma","musigma"],"author_name":"Дарья","publish_date":"2016-07-24T05:16:29","publication_year":"2016","word_count":1474,"keywords":["business intelligence","Go","programming_languages:R","AI","programming_languages:Go","RAG","mu sigma","ViT","analytics","GAN","R","musigma"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","GAN","ViT","business intelligence","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/great-lakes-institute-management-partnership-great-learning-introduces-mentored-learning-online-analytics-program\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10168517,"title":"Databricks to Invest $250 Million in India, Plans 50% Workforce Expansion","content":"Databricks has announced a US$250 million investment in India over the next three years, with a focus on advancing data and AI innovation. The company announced that it will increase its local workforce by over 50%, expanding its employee base in India to more than 750 by the end of the fiscal year. The investment will support large-scale training initiatives, a research and development centre in Bengaluru, and go-to-market efforts, the company said during its flagship Data Intelligence Day event in Mumbai. Databricks introduced the India Data + AI Academy to train 500,000 professionals over the next three years. The programme offers courses on data, analytics, and AI, delivered through AI-powered tutors and hands-on labs. Learners will be eligible for Databricks certifications upon completing the programme. “India is emerging as a global AI talent hub, and Databricks is proud to invest in its growing ecosystem,” said Rochana Golani, vice president of learning & enablement at Databricks. “The India Data + AI Academy will equip professionals with critical skills in data and AI, ensuring they stay ahead in the AI-driven economy.” Golani added that Accenture has launched a Databricks learning programme at its Bengaluru Innovation Hub as part of this initiative. The new 105,000-square-foot R&D office is located at Bagmane Capital Park in Bengaluru. Databricks said it will hire over 100 additional engineers this year to expand its research efforts in the region. “In less than two years, our Bengaluru R&D center has grown into a dynamic team of over 100 engineers,” said Vinod Marur, senior vice president of engineering at Databricks. “This year, we plan to hire an additional 100+ R&D engineers to strengthen our capabilities and deepen our roots in this vibrant technology community.” Databricks noted that India remains one of its key growth markets, with organisations such as HDFC Bank, Swiggy, CommerceIQ, Freshworks, TVS Motors, and Zepto using its Data Intelligence Platform. “I’m excited that all our investments in India will ultimately enable our customers to become more successful in their data and AI journey,” said Ed Lenta, senior vice president & GM for Asia Pacific & Japan. The Bengaluru R&D office joins Databricks’ global network, which includes centres in Amsterdam, Belgrade, Berlin, San Francisco, Mountain View, and Seattle.","excerpt":"Databricks noted that India remains one of its key growth markets, with organisations such as HDFC Bank, Swiggy, CommerceIQ, Freshworks, TVS Motors, and Zepto using its Data Intelligence Platform.","categories":["AI News"],"tags":["Databricks"],"author_name":"Siddharth Jindal","publish_date":"2025-04-24T12:30:19","publication_year":"2025","word_count":371,"keywords":["Go","API","programming_languages:R","AI","innovation","programming_languages:Go","analytics","GAN","R","Databricks"],"extracted_tech_keywords":["AI","analytics","Databricks","R","Go","API","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/databricks-to-invest-250-million-in-india-plans-50-workforce-expansion\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044014,"title":"AI Is Getting Creative, In Style","content":"“There is no must in art because art is free” Wassily Kandinsky Art is subjective, ambiguous and reactionary whereas AI is objective, universal and logical, at least on paper. So, what happens when the two worlds collide? Artificial intelligence is redefining the creative process with computers creating music, poetry etc. According to the Association of Computational Creativity, the goal of computational creativity is to model, simulate, or replicate creativity. AI as an imitator Neural artistic style transfer is used to create artworks based on the data fed into it, and uses deep neural networks to recreate, replicate and blend styles to produce new pieces. Pastiches produced by Google AI Recurrent Neural Networks (RNNs) are a class of artificial neural networks associated with sequential models and temporal series that can imitate the work of writers. RNNs are used predominantly in language translation and speech recognition. Human and AI collaboration AI can inform and inspire artists to discover insights, connections and patterns across huge sets of data points. For instance, the ‘trending emotions’ in the world. “Drawing or painting was never my strength, as I never managed to get the same control over my hand muscles as when I write code. So instead of fighting against my body to produce an image I might have in my head, I preferred to learn how to instruct machines to do that. Now that I have the possibilities to create and control my own visual universes, the question remains what is it that I ultimately want to find there?” Mario Klingemann, a pioneer German AI artist states. Harshit Agrawal, a computer-human interaction researcher, curated 60,000 images of human surgical dissections. He calls this the “human-machine creativity continuum”, a melding of human and machine creativity. Also read: Meet Three Leading AI-Based Artists In India ‘The anatomy lesson by Dr Algorithm’ by Harshit Agrawal AI’s ability to create artwork independently is a debatable subject. Many AI systems do not need human intervention in the art-making process, but require human creative inputs and interventions in the learning process. It was observed that most people could not tell the difference between AI art and real human art. 75% of the time, subjects mistook computer generated images as art made by real artists. Ethical issues The use of AI in art and culture poses ethical questions. ‘The Next Rembrandt’ for example was created by analysing 346 Rembrandt paintings pixel by pixel and upscaled by deep learning algorithms 351 years after the painter’s death. Every detail of Rembrandt’s artistic identity set the foundation for the AI to work on. When a human author is replaced by algorithms and machines, who owns copyrights and to what extent can they be exercised? Issues of piracy, plagiarism, originality and creativity have all come under scrutiny in terms of how we view and understand these terms. AI makes art from the data of existing works. If this is so, how original is AI art? Rage against the machine AI is changing our relationship with art and how we perceive beauty, imagination, literature, music and other fine arts. An artwork is preceded by social and historical schemes. It embodies cultural inheritance and lived experiences. Art in itself is mysterious, and it is often difficult to explain where creative ideas come from, and we tend to use vague notions of inspiration and intuition to define art. But art is a human condition; It mirrors life. What is missing in a machine is the artistic process. Algorithms can create appealing images, but it lives in a consolidated, isolated creative space that lacks social context. Human artists are influenced and inspired by people, places, movements, and they use these experiences to create transcendental pieces of work. The machine also lacks intent. Will AI ever be able to create art as a form of resistance? Can it transgress in its own unique ways, resist establishments or fuel revolutions? The collaboration of the two could be the future of art and may fundamentally change the world and our perceptions.","excerpt":"AI can inform and inspire artists to discover insights, connections and patterns across huge sets of data points.","categories":["AI Features"],"tags":["ai generated images"],"author_name":"Prajaktha Gurung","publish_date":"2021-07-20T19:00:00","publication_year":"2021","word_count":668,"keywords":["Replicate","ai generated images","Go","artificial intelligence","AI","neural network","RAG","deep learning","ViT","RNN","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","RAG","R","Go","RNN","ViT","Replicate"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-is-getting-creative-in-style\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168582,"title":"Sundar Pichai Says Over 30% of Code at Google Now AI Generated","content":"More than 30% of the code written at Google is now created with help from AI, CEO Sundar Pichai said during Alphabet’s recent Q1 2025 earnings call. This means developers are accepting AI-generated suggestions in nearly one out of every three code changes. Pichai said the company is seeing strong momentum in the use of AI-assisted coding across teams, driven by more capable models and the introduction of agentic workflows, AI systems that can plan and execute multi-step tasks. “We are deploying these deeper flows across all parts of the company,” he said, noting that customer service teams are particularly ahead in adoption. Alphabet, Google’s parent company, reported its first quarter 2025 financial results on April 24. Consolidated revenues increased 12% year over year to $90.2 billion, but declined 6.5% QoQ from $96.5 billion in Q4 2024. Net income rose 46% to $34.5 billion, with diluted earnings per share up 49% to $2.81. Operating income for the quarter reached $30.6 billion, up from $25.5 billion a year earlier, and the operating margin expanded to 34% from 32%. Google’s core advertising business remained a primary driver, with Google Search and other revenues climbing to $50.7 billion, up from $46.2 billion in the same quarter last year. YouTube ad revenue increased to $8.9 billion, while Google subscriptions, platforms, and devices brought in $10.4 billion. Google Cloud revenues saw significant growth, rising 28% year-over-year to $12.3 billion, led by demand for core cloud products, AI infrastructure, and generative AI solutions. Google recently released Gemini 2.5 Pro, which has been well-received by both developers and consumers. According to Pichai, the model is “state-of-the-art across a wide range of benchmarks” and even debuted at number one on the Chatbot Arena by a significant margin. He added that Gemini 2.5 Pro brings major improvements in reasoning, coding, science, and math, opening up new possibilities for developers and customers alike. “Active users in AI Studio and Gemini API have grown over 200% since the beginning of the year,” he said. Gemini models are now embedded into all 15 Google products with over half a billion users. Android and Pixel are leading the charge, offering features like Gemini Live and AI-powered camera tools. Google Assistant is also being upgraded to Gemini, with support extending later this year to tablets, smartwatches, and cars. In addition, Gemma models have been downloaded more than 140 million times. Pichai also highlighted new frontiers for AI development, including Gemini Robotics Models and AI Co-Scientist, a multi-agent system for scientific research. AlphaFold, a breakthrough in protein structure prediction, has now been used by more than 2.5 million researchers. Pichai highlighted progress across Google’s core platforms. In Search, AI Overviews now serve over 1.5 billion users monthly, helping expand the types of questions Google can answer. Gemini Live screen sharing and camera tools are also rolling out across Android devices, including the new Pixel 9a and Samsung S25.","excerpt":"Google Cloud revenues saw significant growth, rising 28% year-over-year to $12.3 billion, led by demand for core cloud products, AI infrastructure, and generative AI solutions.","categories":["AI News"],"tags":["Google"],"author_name":"Siddharth Jindal","publish_date":"2025-04-25T10:39:39","publication_year":"2025","word_count":483,"keywords":["Go","API","programming_languages:R","AI","agentic workflows","llm_models:Gemini","generative AI","Google","cloud_platforms:Google Cloud","R","Gemini 2.5"],"extracted_tech_keywords":["AI","generative AI","Gemini 2.5","agentic workflows","R","Go","API","llm_models:Gemini","cloud_platforms:Google Cloud","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sundar-pichai-says-over-30-of-code-at-google-now-ai-generated\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10068105,"title":"DigitalOcean launches a serverless computing solution","content":"The US-based cloud infrastructure provider DigitalOcean has launched a serverless computing solution, DigitalOcean Functions. According to Anshu Agarwal, VP Serverless at DigitalOcean, serverless computing has been one of the most popular requests from DigitalOcean users. In recent years, serverless computing has gained tremendous popularity among developers building modern apps, and according to IDC’s IaaSView buyer survey, 25% of cloud IaaS buyers intend to utilise serverless functions in the next 12 months. Serverless is a cloud-native development model that allows developers to build and run applications without having to manage servers. DigitalOcean Functions will provide a cost effective computing solution by allowing developers to create serverless functions. Developers will be able to write functions in their preferred languages without the need to learn complex concepts like Kubernetes and containers. It will enable them to focus more on coding as scaling of infrastructure components and resource managament will be automatically done based on demand. DigitalOcean functions will allow developers to build apps that require on demand functions and extend them with serverless computing. “A developer with a container-based app could add a function as an API endpoint, enabling them to easily extend their app running on containers with serverless components,” said Anshu Agarwal. The serverless solution supports popular programming languages like Node.js, Python, Go, and PHP. It also makes testing and integration easier. In addition to simplifying efforts, DigitalOcean Functions will do away with the need to maintain idle servers thereby saving costs. Users will pay for resources only when they are running.  Customers can avail the product across all DigitalOcean regions including New York, Amsterdam, San Francisco, Singapore, London, Frankfurt, Toronto and Bangalore.","excerpt":"Serverless is a cloud-native development model that allows developers to build and run applications without having to manage servers.","categories":["AI News"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-05-31T12:56:01","publication_year":"2022","word_count":272,"keywords":["Go","API","programming_languages:R","AI","Git","serverless","Python","programming_languages:Python","R","kubernetes"],"extracted_tech_keywords":["AI","kubernetes","serverless","Python","R","Go","Git","API","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/digitalocean-launches-a-serverless-computing-solution\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10012519,"title":"HR Technology Company Eightfold AI Raises $125 Million Funding To Hire Exceptional Talent In India","content":"Eightfold AI, an HR technology company, has announced raising $125M funding in series D to hire exceptional talent in India. According to the official release, the funds will also be used to expand and scale its AI-powered Talent Intelligence Platform, which manages the entire talent lifecycle. Further, it also addresses the disjointed point solutions commonly found in the talent management space. With this funding, the company will now be valued at $1 billion along with a customer base across 110 countries, 17 industries and 13 languages. Subsequently, the company is positioned to invest heavily in growing its India presence by adopting a remote-first model for its employees. General Catalyst led this round of funding along with investors like Capital One Ventures, Foundation Capital, IVP and Lightspeed Venture Partners. Till date, the company has raised more than $180 million and has more than quadrupled its sales since April 2019. Currently, the company is serving customers like Tata Communications, AirAsia, Bayer, Capital One and Micron. When asked about the funding, Ashutosh Garg, Founder & CEO at Eightfold, stated to the media, that the company aims to leverage AI to provide the right career for everyone in the world. As a matter of fact, the team has spent the last four years developing a single platform addressing all talent lifecycle needs, bypassing single-point solutions entirely. Adding to this, Vinodh Ravindranath, Head of Artificial Intelligence at Eightfold AI stated to the media, that India has a long history of development and engineering accomplishments, and the company wishes to harness that talent to its fullest effect. “This investment ensures that the most talented, intelligent and ambitious engineers and data scientists in India will have a place to continue their growth here at Eightfold.” According to the official release, the platform by Eightfold AI manages the entire talent lifecycle by bringing in billions of anonymised data points together and working with algorithms and domain expertise to make a reliable, scalable impact for enterprise-scale organisations. It uses deep learning AI on existing public and legacy HCM data to transform how companies hire talent. With the platform, the customers have managed to get a 49% increase in the number of employees finding their next role within their current company, a 58% increase in employees offered other jobs within the same company after their role had been eliminated, and 60% more highly qualified candidates are applying to open jobs.","excerpt":"Eightfold AI, an HR technology company, has announced raising $125M funding in series D to hire exceptional talent in India.  According to the official release, the funds will also be used to expand and scale its AI-powered Talent Intelligence Platform, which manages the entire talent lifecycle. Further, it also addresses the disjointed point solutions commonly […]","categories":["AI News"],"tags":["ai in hr"],"author_name":"Sejuti Das","publish_date":"2020-11-26T17:26:33","publication_year":"2020","word_count":400,"keywords":["Go","API","artificial intelligence","AI","ai in hr","Scala","RAG","Aim","deep learning","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","Aim","RAG","R","Go","Scala","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hr-technology-company-eightfold-ai-raises-125-million-funding-to-hire-exceptional-talent-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166930,"title":"Will Trump’s Tariffs Derail India’s GCC Expansion?","content":"As US President Donald Trump renews his push for protectionist trade policies, his proposed reciprocal tariffs are causing concern among Indian exporters. Announced earlier this month, these tariffs—set to mirror the duties that other countries impose on American goods—are scheduled to take effect on April 2. While Trump’s decision to delay similar tariffs on cars and auto parts from Mexico and Canada gives US automakers some breathing space, there is little indication that India will receive the same consideration. Sanjay Varnwal, co-founder and CEO of Spyne.ai, stated that the real impact of these tariffs will depend on how India chooses to respond. “If India counters with its own tariffs, it could disrupt supply chains, increase costs, and slow market expansion for both countries,” he said. “For American automakers looking at India as a key market, higher tariffs could make their vehicles less competitive.” On the flip side, Varnwal noted that this situation could motivate Indian manufacturers to focus more on domestic production and explore new export markets. “The reciprocal tariff system might accelerate efforts to strengthen domestic production and focus on exports to other regions,” he explained. “However, if India decides to lower duties on high-end American vehicles, it will do so with the intent to keep trade relations balanced.” He pointed to research by SBI, which suggests the effects may be less damaging than feared. “Even if the US increases tariffs by 15-20%, India’s overall exports are expected to decline by only 3-3.5%,” he said, adding that India’s ongoing efforts to secure trade agreements with the UK and the EU could help soften the blow. “Ultimately, the industry will have to adapt—automakers will have to find ways to optimise supply chains, explore alternative strategies, and focus on expanding their presence in other growing markets,” he concluded. What This Means for India’s GCCs While exporters are watching the tariff situation closely, the outlook for India’s global capability centres (GCCs), particularly in automotive research and development, appears more stable. “For now, R&D in GCCs in India is unlikely to be directly affected by tariffs. India has emerged as a global hub for automotive R&D with over 60 active automotive GCCs, spanning electric vehicles, AI-driven automation, and mobility solutions,” Varnwal said. He cautioned, however, that if tariffs lead to wider trade restrictions, companies might need to rethink their investments. “But given India’s strong digital and engineering capabilities, it might remain a compelling destination for automotive innovation,” he added. When asked whether this situation might discourage companies from setting up new automotive GCCs in India, Varnwal said, “Not necessarily. India’s appeal as an R&D and engineering hub extends beyond trade policies. The country offers a deep talent pool, cost advantages, and a growing domestic market for electric and connected vehicles.” He believes that while tariffs might impact vehicle imports and exports, they don’t directly affect the strategic advantages that make India a global R&D powerhouse. He pointed out that over 20 global automakers, including General Motors, Volkswagen, Volvo, Ford, Marelli, and BorgWarner, already have R&D centres in India. “Major players will continue investing in Indian GCCs, especially for software-defined vehicles, AI-driven automotive solutions, and next-gen mobility tech,” he said. “If anything, this situation may encourage more localised manufacturing and innovation, reinforcing India’s position as an automotive tech leader rather than just a production base.” Trade Tensions and the Bigger Picture In February, Prime Minister Narendra Modi’s visit to the US led to an agreement to begin talks on a new bilateral trade deal. However, these discussions are not expected to influence the upcoming reciprocal tariffs, which are set to take effect next month. During his State of the Union address on March 4, Trump singled out India as a significant contributor to trade imbalances and reaffirmed his commitment to imposing matching duties. Alouk Kumar, founder and CEO at Inductus Limited, shared his thoughts on the broader implications for India’s GCCs. “Trump’s pro-America policies, including tariffs and immigration curbs, may cast a subtle shadow over India’s GCCs in the near term, but this is more of a ripple than a tide.” He further stated that automotive GCCs may even benefit from the current situation if automakers choose to localise operations to avoid tariffs. A Strong Outlook for India’s GCCs The future outlook for India’s GCCs remains optimistic. Kumar shared that projections estimate GCCs in India will generate over $110 billion by 2030. “American companies are unlikely to retreat; instead, they will adapt, leveraging India not just as a cost-effective hub but as a strategic partner in navigating the complexities of a changing global landscape.” “The story of India’s GCCs is one of innovation, and it’s still at a nascent stage,” he concluded. Consistent Growth Despite Uncertainty Echoing this sentiment, Vikram Ahuja, co-founder of ANSR and CEO of 1Wrk, expressed confidence in India’s GCC future despite Trump’s policies. “The consistent momentum in GCC expansion through 2024 and early 2025 demonstrates that global enterprises, particularly US-based companies, see India as a critical hub for talent, innovation, and operational efficiency,” he said. While some policy shifts may affect outsourcing trends, Ahuja believes they won’t halt the rise of India’s GCCs. He also pointed to India’s growing role in global technology leadership. “With deep expertise in AI, machine learning, and digital technologies, India’s workforce provides the capabilities that global enterprises need to scale their AI initiatives, enhance automation, and drive next-generation innovation,” he said. Ahuja stated that the numbers tell the story. “Major players like Meta, Google, Ford, Amgen, and ANZ are increasing their presence in India, while new entrants—including Goodyear, Sonoco, Cyara, and Dark Matter—have announced GCC setups,” he shared. “In Q4 2024 alone, more than 22 new GCCs were established, and in early 2025, the trend has continued with over 12 new setups and over 14 expansions. Over 30 new GCCs are expected to be established in the next five to six months.” “India’s GCC ecosystem has proven to be resilient, future-ready, and indispensable to global enterprises. Rather than seeing a decline in American companies setting up GCCs in India, the current trends suggest a continued expansion and deepening of GCC investments as companies seek to balance operational efficiencies with long-term innovation and competitiveness in a tech-driven global economy,” Ahuja concluded","excerpt":"“If India counters with its own tariffs, it could disrupt supply chains, increase costs, and slow market expansion for both countries.”","categories":["GCC"],"tags":["donald trump"],"author_name":"Shalini Mondal","publish_date":"2025-03-31T09:00:00","publication_year":"2025","word_count":1035,"keywords":["Go","machine learning","programming_languages:R","AI","innovation","programming_languages:Go","donald trump","Git","RAG","automation","R"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","Git","automation","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/gcc\/will-trumps-tariffs-derail-indias-gcc-expansion\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10163180,"title":"‘AI.com’ Now Redirects to DeepSeek","content":"The domain AI.com now redirects to Chinese AI maker DeepSeek’s chatbot – ’chat.deepseek.com’ According to the South China Morning Post, DeepSeek appears “to have received approval” from the owner of AI.com, based in Kuala Lumpur, Malaysia, and has registered the domain in the region. According to Whois, a service that lets one check the ownership of a domain, AI.com was first registered in May 1993. However, the domain plays musical chairs of sorts with big tech companies and their AI models. Previously, in 2023, the domain redirected users to ChatGPT, and it was reported that OpenAI CEO Sam Altman bid for the domain. Soon after, the domain started redirecting users to Elon Musk’s xAI. Later, in February 2024, it was observed that the domain was redirecting to a video posted by American YouTuber Marques Brownlee on his channel. It redirected users to one of his videos titled ‘AI-generated videos just changed forever’. Even then, it wasn’t clear who the owner of the domain was, and Brownlee clarified that he did not own the domain. In March, Android Police reported that the domain was redirecting to Google’s Gemini chatbot. However, in November last year, OpenAI bought the domain ‘chat.com’, and Altman notified users of the update in a post on X. Several reports in 2023 revealed that HubSpot co-founder and CTO Dharmesh Shah purchased ‘chat.com’ for $15.5 million, making it one of the most expensive domain sales on record. In March, he revealed that he sold the domain to an undisclosed buyer. Later, he confirmed on X that OpenAI was the purchaser, hinting that the transaction might have involved OpenAI shares. OpenAI has not disclosed the amount they paid for this acquisition, but people predict it’s likely more than $15 million. Moreover, TechCrunch reported that a software engineer named Ananay Arora bought the domain name ‘OGOpenAI.com’ and redirected it to DeepSeek. However, the domain does not redirect to DeepSeek’s chatbot as of today, but to a black screen with the message ‘Are you accelerating anon (anonymous)?’.","excerpt":"Previously, at different times, the domain redirected to ChatGPT, xAI, Google’s Gemini and an MKBHD video.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","DeepSeek"],"author_name":"Supreeth Koundinya","publish_date":"2025-02-11T10:37:52","publication_year":"2025","word_count":335,"keywords":["Go","ChatGPT","OpenAI","AI","R","GPT","DeepSeek","XAI","llm_models:Gemini","xAI","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","xAI","R","Go","GPT","XAI","llm_models:GPT","llm_models:Gemini"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-com-now-redirects-to-deepseek\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":557,"title":"75F launches operation in India with IoT based Smart Building solution","content":"US based 75F, a leader in building automation for light commercial buildings, announced the launch of its business operations in India. This would be company’s first international office and it will be set up in Bangalore. Also along with this launch, 75F will be introducing its award-winning ‘Dynamic Airflow Balancing’ technology to the Indian market. 75F, headquartered in Minneapolis, USA, creates solutions that use Internet of Things (IoT) and cloud computing to predict the needs of the building and manages them proactively. Their solutions make spaces more comfortable, energy efficient, automated and smart. IoT devices can sense, analyse and control the behaviour of the building to achieve consistent temperature, airflow and lighting on a zone or room by room basis. 75F through its solutions helps in saving up to 70% in energy usage. It is easy to install and reduces carbon footprint. “It is important to note that every building is a dynamic entity and hence the building intelligence system installed should be dynamic as well. We are excited to bring to India our state-of-the-art solutions that leverage the latest in technology and IoT, understand your building’s ever changing needs, and proactively cater to them to ensure that the occupants are comfortable,” said 75F CEO and Founder, Deepinder Singh. India being a very lucrative market for 75F to enter, offers a little more than 499 million sq. ft. of addressable commercial buildings space. Gaurav Burman, VP & Country President, 75F India further added, “We see a huge untapped market in India for IoT-based building intelligence systems ranging from new-builds to retrofits in both enterprises and SMEs. Large or small, the Indian business consumer is looking for comfort and efficiency, power in the palm of his hand, at a cost that is affordable. One of the biggest headaches customers face is how to integrate these new technologies into their existing infrastructure. With a strong R&D team in India, we seek to design and tailor solutions that cater specifically to the Indian market.” When asked about the expected market share in India, Gaurav Burman said, “Harnessing IoT to make buildings smarter is new today, but our reading is that in three to seven years, everyone is going to be moving over to these predictive systems. We are a startup in a ₹3500 crore addressable market and we are confident we can capture significant share by FY2018.” Referring to the concept as ‘Internet of Air’, 75F along with the launch of operations also unveiled its innovative approach to HVAC zone controls called Dynamic Airflow Balancing™. 75F through its solution has been able to achieve continuous commissioning or perfect air balancing by leveraging IoT design philosophy and the power of cloud computing. Launched on a national basis, the smart HVAC system has the potential to reduce energy usage by 562 trillion BTUs, enough energy for 24 million cars to be taken off the road. Wireless Zone Controllers sense and collect hundreds of data points from the room every minute and send it to the Central Control Unit & then to the servers in the Cloud. Each night cloud computing algorithms analyse these data points, including weather forecast and daily usage patterns, that allow the system to predict future conditions. Post which, a new set of instructions are sent to the Central Control Unit and the motorised dampers are modulated a few degrees at a time to achieve the perfect balance. The system also factors in real-time events, such as room occupancy, the position of the sun and weather patterns to make continuous adjustments to the plan as needed. Speaking about the new technology, Pankaj Chawla, CTO of 75F said, “We believe in the promise of IoT to improve our lives by designing systems that work so well, you forget they are there. Currently, we are focused on one of the most serious problems in commercial buildings – HVAC – because it represents the largest single cost of operating a building. Beyond HVAC, there is a need to provide Building Automation Systems (BAS) to commercial buildings that are easy to install, affordable and effective. We look to bring the further benefits of BAS to improve energy usage, lighting, security and more.”","excerpt":"US based 75F, a leader in building automation for light commercial buildings, announced the launch of its business operations in India. This would be company’s first international office and it will be set up in Bangalore. Also along with this launch, 75F will be introducing its award-winning ‘Dynamic Airflow Balancing’ technology to the Indian market. […]","categories":["AI News"],"tags":["IoT"],"author_name":"Manisha Salecha","publish_date":"2016-08-26T13:44:03","publication_year":"2016","word_count":696,"keywords":["Go","programming_languages:R","cloud computing","AI","llm_models:PaLM","programming_languages:Go","RAG","automation","R","IoT","startup"],"extracted_tech_keywords":["AI","RAG","cloud computing","R","Go","automation","startup","llm_models:PaLM","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/75f-launches-operation-india-iot-based-smart-building-solution\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":25456,"title":"Govt Plans AI-Based Surveillance To Create Positive News Slant For India","content":"The 2019 General Elections in India are going to be fought on a whole different level this year — a whole new dimension, as the trend indicates. Big data, analytics and artificial intelligence are going to play a key role in the fight to power. A tender released by the Broadcast Engineering Consultants India Limited (BCIL), a public sector enterprise, under the Ministry of Information and Broadcasting showcases that the reigning power seems to be taking the public opinion and media seriously. The Government had sent out a proposal looking for a specific “technology platform” which would sense general public emotion by analysing social media posts, blogs and even emails to help boost nationalism and neutralise any “media blitzkrieg by India’s adversaries.” The proposal explains: “The platform is expected to provide automated reports, tactical insights as well as comprehensive workflows to initiate engagement across digital channels. The platform may be used to disseminate content and hence should support publishing features. The platform should also support easy management of conversational logs with each individual with capabilities to merge it across channels to help facilitate creating a 360-degree view of the people who are creating buzz across various topics.” The tender also explains that the technology platform is expected to monitor these media to analyse sentiment and identify “fake news”. They are also expected to promulgate information on behalf of the Government and publicise news and social media posts with a “positive slant for India”. To counter the aforementioned “media blitzkrieg”, the tender explains that the software should use a variety of statistical techniques from predictive modelling and data mining to analyse the current perception and make predictions about the future or unknown events. The software should analyse the patterns found in historical data to identify potential risks and opportunities. The analytics tool should capture relationships between explanatory variables and predicted variables from past occurrences and utilize them to predict the unknown outcome. The analytics tool should be able to use machine learning to predict, analyse and help the Ministry of Information and Broadcasting to make informed decisions. The analytics tool should create custom predictions based on the perceptions being created. Artificial intelligence may be used for predictive trends and sentiments over social media. This is not the first time that the NDA-led Indian Government is investing time and money to maintain a positive image for the country as well as the ruling authorities. In fact, the BJP had invested a huge chunk of money as well as manpower on technology, social media, data analysis and marketing, during the run-up to the 2014 Lok Sabha elections. Arvind Gupta, who had led the IT team for BHP before the last elections, had used technology and data to communicate and to convert people into volunteers and volunteers into voters. He even ran a campaign internally called ‘Organise online for success offline’. Like any company, the BJP IT cell used data to understand trends, pick markets correctly and run the campaign. On the other hand, the Congress party’s freshly-minted Data Analytics Department is dissecting the numbers for the upcoming elections. These figures include vote share, seats and the apparent shifting of voting patterns since 2009. The INC is going to use the analytics gun to strategise against the BJP for the 2019 Lok Sabha battle.","excerpt":"The 2019 General Elections in India are going to be fought on a whole different level this year — a whole new dimension, as the trend indicates. Big data, analytics and artificial intelligence are going to play a key role in the fight to power. A tender released by the Broadcast Engineering Consultants India Limited (BCIL), a […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","BJP","Narendra Modi","social media"],"author_name":"Prajakta Hebbar","publish_date":"2018-06-15T09:07:49","publication_year":"2018","word_count":550,"keywords":["big data","Go","artificial intelligence","machine learning","programming_languages:R","AI","social media","Git","BJP","analytics","Narendra Modi","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","R","Go","Git","big data","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/govt-plans-ai-based-surveillance-to-create-positive-news-slant-for-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10142228,"title":"Freshworks Names Srinivasan Raghavan as Chief Product Officer","content":"Saas giant Freshworks Inc. has appointed Srinivasan Raghavan as its new Chief Product Officer (CPO), bringing over 20 years of leadership experience in enterprise SaaS. Raghavan will spearhead the company’s product vision and strategy to enhance its AI-powered customer experience (CX) and employee experience (EX) solutions. He will report directly to Dennis Woodside, CEO and President of Freshworks. “Srini is a key addition to our team to lead innovation that delivers a scalable trajectory for growth across our three key business priorities: employee experience, artificial intelligence, and customer experience,” said Woodside. “Srini’s track record driving enterprise growth and managing complex multi-product scaling efforts coupled with his bold vision for the future of AI, make him uniquely qualified to lead our CX and EX product strategy.” Saas Leader Previously, Raghavan served as CPO at RingCentral, where he expanded the company’s portfolio with new cloud-based solutions for contact centers, marketing, and sales intelligence. During his tenure as senior vice president at Five9, he led the development of AI-driven digital engagement and workflow automation tools. His earlier career includes leadership roles at Cisco, where he contributed to the Applications Software and Collaboration Business units. “Joining Freshworks at a time when AI is unlocking new possibilities for businesses worldwide to drive growth and improve operational efficiency is an incredible opportunity,” Raghavan said. “By continuing to integrate AI and workflow automation across the Freshworks platform and solutions, we can add significant customer value and shape the future of CX and EX together.” Raghavan holds advanced degrees in computer science and engineering and an MBA from The University of Chicago Booth School of Business. His global leadership experience spans the U.S., Europe, and Asia. Freshworks has been going all in on AI agents. It recently unveiled its new generation of AI agent Freddy AI. It’s an easy-to-deploy autonomous service agent for enhancing both customer and employee experiences","excerpt":"Based out of Chennai, people-first AI service software company Freshworks serves over 68,000 companies.","categories":["AI News"],"tags":["Freddy AI","Freshworks","SaaS","Srinivasan Raghavan"],"author_name":"Vandana Nair","publish_date":"2024-12-02T20:25:01","publication_year":"2024","word_count":311,"keywords":["Go","Srinivasan Raghavan","artificial intelligence","AI","innovation","Scala","Freshworks","RAG","Git","Freddy AI","SaaS","automation","workflow automation","R"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","R","Go","Scala","Git","automation","workflow automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/freshworks-names-srinivasan-raghavan-as-chief-product-officer\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35105,"title":"Here&#8217;s How Uber’s Ludwig Helps Deploy Code-Free DL Models","content":"Uber AI is at the heart of AI-powered innovation and technologies at Uber. The research and advancements in artificial intelligence made by the team solve several challenges across the company. For example: Uber AI develops packages and open sources them which are built on top of the strong foundations of other open source libraries Pyro, a deep probabilistic programming language built on PyTorch, released by Uber in 2017 Another major open source AI tool Horovod, a framework hosted by the LF Deep Learning Foundation that allows distributed training of deep learning models over multiple GPUs and several machines At Uber, deep learning models are used for a variety of tasks like customer support, object detection, improving maps, streamlining chat communications, forecasting, and preventing fraud. Named after one of the greatest physicists, Ludwig Boltzmann, Uber’s new tool Ludwig is poised to make deep learning more accessible with code-free model training. Increased Deep Learning Accessibility With Ludwig Ludwig was developed internally at Uber over the past two years to simplify the use of deep learning models in applied projects. Multi-task learning can be performed along with learning to predict all the outputs simultaneously, a task that usually requires custom code. Default values of preprocessing, training, and various model architecture parameters are chosen or are adapted from the academic literature, allowing non-experts to easily train complex models. The new idea that Uber AI introduces with Ludwig is the notion of data-type-specific encoders and decoders, which results in a highly modularized and extensible architecture: each type of data supported (text, images, categories, and so on) has a specific preprocessing function. In short, encoders map the raw data to tensors, and decoders map tensors to the raw data. The researchers at Uber AI say that each data type may have more than one encoder and decoder. For instance, text can be encoded with a convolutional neural network (CNN), a recurrent neural network (RNN), or other encoders. The user can then specify which one to use and its hyperparameters directly in the model definition file without having to write a single line of code. How Is Ludwig Different From Other Open Source AI Toolboxes Ludwig is built on top of TensorFlow that allows users to train and test deep learning models without writing code. In this way, Ludwig is unique in its ability to help make deep learning easier to understand and enable faster model improvement iteration cycles for experienced non-experts and researchers alike. Tool Integration: Ludwig has a mix of influences from Weka, scikit-learn, and Mlib. Ludwig is a different tool from the usual deep learning libraries that provide tensor algebra primitives and few other utilities to code models, while at the same time making it more general than other specialized libraries like PyText, StanfordNLP, AllenNLP, and OpenCV No coding required: No coding skills are required to train a model and use it for obtaining predictions Extensibility: Easy to add new model architecture and new feature data types Deploying Ludwig In 4 Easy Steps Install Just run pip install ludwig and it will be ready to use. Some features may require further steps, read Getting Started. Train Prepare data in a CSV file, define input and output feature in a model definition YAML file and run: ludwig train --data_csv file.csv --model_definition definition.yaml Predict Prepare data in a CSV file and use a pre-trained model to predict the output targets: ludwig predict --data_csv data.csv --model path_to_model Visualize Ludwig comes with many visualization options. ludwig visualize --visualization learning_curves --training_statistics train_statistics.json Know more about Ludwig here. Future Direction Ludwig incorporates a set of command line utilities for training, testing models, and obtaining predictions. Furthering its ease-of-use, the toolbox provides a programmatic API that allows users to train and use a model with just a couple lines of code. By open-sourcing Ludwig, Uber has made AI more accessible and with this, there will be a torrential inflow of ideas and innovations from outsiders who have been left out due to hurdles like coding skills.","excerpt":"Uber AI is at the heart of AI-powered innovation and technologies at Uber. The research and advancements in artificial intelligence made by the team solve several challenges across the company. For example: Uber AI develops packages and open sources them which are built on top of the strong foundations of other open source libraries […]","categories":["Deep Tech"],"tags":["Deep Learning","Uber"],"author_name":"Ram Sagar","publish_date":"2019-02-18T11:53:57","publication_year":"2019","word_count":665,"keywords":["scikit-learn","artificial intelligence","AI","neural network","PyTorch","ML","OpenCV","NLP","deep learning","Uber","Deep Learning","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","neural network","NLP","TensorFlow","PyTorch","scikit-learn","OpenCV"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-ubers-ludwig-helps-deploy-code-free-deep-learning-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10008620,"title":"Top Works In Neural Architecture Search","content":"Currently employed neural network architectures have mostly been developed manually by human experts, which is a time-consuming and error-prone process. This is when Neural architecture search, a subset of AutoML, came to the rescue. Neural Architecture Search (NAS) is the process of automating architecture engineering. Source: automl.org Here we list top research works in Neural Architecture Search based on their popularity on Github. These works have set new baselines, resulted in new networks and more. Efficient Neural Architecture Search This work proposes Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In this approach, a controller (RNN), trained with policy gradient, learns to discover neural network architectures by searching for an optimal subgraph within a large computational graph. The authors state that ENAS is fast, delivers strong empirical performances using much fewer GPU-hours than all existing automatic model design approaches, and notably, 1000x less expensive than standard Neural Architecture Search. Link to the paper. The Evolved Transformer In this work, the researchers from Google Brain attempt to apply NAS to search for a better alternative to the Transformer. They first constructed a large search space and then ran an evolutionary architecture search by seeding the initial population with the Transformer. The results showed that the architecture — the Evolved Transformer — demonstrated consistent improvement over the Transformer on four well-established language tasks. Link to the paper. MobileDets In this work, the authors state to have achieved substantial improvements in the latency-accuracy trade-off by incorporating regular convolutions in the search space and effectively placing them in the network via neural architecture search. This work resulted in a family of object detection models, MobileDets, that achieve state-of-the-art results across mobile accelerators. Link to the paper Progressive Neural Architecture Search In this work, the authors propose a new method for learning the structure of convolutional neural networks (CNNs) that is more efficient than recent state-of-the-art methods based on reinforcement learning and evolutionary algorithms. This approach uses a sequential model-based optimisation (SMBO) strategy. Results show that this method is up to 5 times more efficient than the popular RL methods in terms of the number of models evaluated, and 8 times faster in terms of total compute. Link to the paper. DARTS This paper addresses the scalability challenge of architecture search by formulating the task in a differentiable manner. Instead of applying evolution or reinforcement learning over a discrete and non-differentiable search space, this method is based on the continuous relaxation of the architecture representation, allowing an efficient search of the architecture using gradient descent. Experimental results show that this algorithm excels in discovering high-performance convolutional architectures for image classification and recurrent architectures for language modelling. Link to the paper. MorphNet With MorphNet, the researchers aim to automate the design of neural network structures. MorphNet iteratively shrinks and expands a network. In contrast to previous approaches, MorphNet is scalable to large networks, adaptable to specific resource constraints and capable of increasing the network’s performance. When applied to standard network architectures on a wide variety of datasets, this approach discovers novel structures in each domain, obtaining higher performance. Link to the paper. Neural Architecture Search Without Labels This paper tries to explore the most pressing question: can we do away with human annotations of images? Can we find high-quality neural architectures using only images? To answer this, the researchers first defined a new setup called Unsupervised Neural Architecture Search (UnNAS). They then trained a large number of diverse architectures with either supervised or unsupervised objectives and found that the architecture rankings produced with and without labels are highly correlated. The results reveal that labels are not necessary, and the image statistics alone may be sufficient to identify good neural architectures. Link to the paper. SpineNet This paper states that convolutional neural networks do not perform well for tasks requiring simultaneous recognition and localisation. The encoder-decoder architectures are proposed to resolve this by applying a decoder network onto a backbone model designed for classification tasks. In this work, the authors argue encoder-decoder architecture is ineffective. So, they propose SpineNet, a backbone with scale-permuted intermediate features and cross-scale connections that is learned on an object detection task by Neural Architecture Search. Using similar building blocks, SpineNet models outperform ResNet-FPN models. Link to the paper. Randomly Wired Neural Networks The researchers at Facebook explored a more diverse set of connectivity patterns with Randomly Wired Neural Networks for Image Recognition. Randomly wired networks are founded on random graph models in graph theory. According to the authors, randomly wired neural networks managed to outperform human-designed networks such as ResNet and ShuffleNet with a typical computation budget and can go toe to toe with the increased computational regime. Read more about this here. To keep yourself up to date with the latest developments in NAS, check this.","excerpt":"Currently employed neural network architectures have mostly been developed manually by human experts, which is a time-consuming and error-prone process. This is when Neural architecture search, a subset of AutoML, came to the rescue. Neural Architecture Search (NAS) is the process of automating architecture engineering. Here we list top research works in Neural Architecture Search […]","categories":["AI Trends"],"tags":["cnn neural network","Convolution Neural Network","Convolutional Neural Network","neural architecture search","policy gradient"],"author_name":"Ram Sagar","publish_date":"2020-09-30T11:00:45","publication_year":"2020","word_count":801,"keywords":["Go","AI","neural network","ML","image recognition","cnn neural network","Scala","Git","neural architecture search","Convolution Neural Network","Aim","policy gradient","object detection","Convolutional Neural Network","R"],"extracted_tech_keywords":["AI","ML","neural network","Aim","image recognition","object detection","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/neural-architecture-search-top-works-papers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10012596,"title":"Guide to V7 Darwin &#8211; The Rapid Image Annotator","content":"V7 labs provide rapid and most powerful end-to-end computer vision development framework for teams building image and video annotation products. Founded in 2018 by Albert Rizzoli(CEO) and Simon Edwardson(CTO), headquartered in London, United Kingdom. Managing datasets at ground-truth via an intuitive interface, or with powerful API integrations. Maintains crowdsourcing with hundreds of annotators and provides customers with customizable workflow steps. Within a year, V7 claimed to have semi-automatically annotated over 1,000 images and video segmentation to curate datasets. V7 specializes in the areas of healthcare, manufacturing, autonomous driving, sports, life sciences, and agri-tech. The Darwin Platform (from V7 labs) has access to its automated image annotation and neural network training with active learning that create a pixel-perfect annotation, which is 10x faster than traditional tools. It allows great team collaboration, visualization and has resulted in a massive increase in the efficiency and detail of the labelling process. The API and CLI tools have allowed flexibility to work with the data and model training pipeline. It also scales really well with user’s demands that meet strict compliance and compatibility standards. The active documentation provides details about every toolkit in the framework, which makes it user friendly. Image And Video Annotation The speed at which V7 darwin works is incomparable to any other platform. The features offered are pixel-level accuracy in labelling and segmentation. Annotator metrics to check the performance of each annotator’s output reported in graphs and exportable CSV files, along with the quality assurance reviews. Better annotation details with sub annotations like adding attributes, text, directional vectors, grouping, and more within each region of interest. The tools using for achieving these are auto-annotator, polygon, brush and eraser, bounding boxes, keypoints, line, ellipse, cuboid, classification tags, attributes, instance tags, directional vectors. Easy exporting into any deep learning framework Dataset Management Supports a wide range of file formats to be uploaded such as JPG, PNG, TIF, MP4, MOV, SVS, DICOM. Preview images in a flash, powered by a high-speed Elixir(Erlang-based language to handle massive scale concurrency between millions of users are moving billions of images) back-end. Keeps track of the work in progress with proper versioning. Allows troubleshooting model performance by class representation and instance frequency. It also has an option for advanced search filtering. V7 Neurons Train neural networks and run them from a web dashboard. V7 Neurons is a deep learning inference series of pre-trained for building image recognition applications. V7 in collaboration with Nvidia has launched NJORD GPU engine to deploy models and guarantee 99.99% uptime and infinite scalability. Zero dev-ops engineers needed. Models can be run in a REST API to call from any device in realtime. Industry Use Cases: Aipoly – This Vision AI has been built for blind and visually impaired with over 26 languages in realtime that can be run on smart devices. This application can recognize common objects, text, food, colours, plants and animals. Retail AI Life Sciences – microscopy, pipetting. Environmental – plant and field agriculture, livestock and wildlife Manufacturing – Defect inspection, prototype logging, material science Python SDK Installation: pip install darwin-py Client-Server Connection from darwin.client import Client client = Client.local() dataset = client.get_remote_dataset(\"example-team\/test\") dataset.pull() # downloads data annotations and images\/frames(for video) for the latest exported version For loading dataset using PyTorch, you can check this documentation and for complete CLI and API integration follow this GitHub repository. Partnered Companies and Research Labs Merck, cloudfactory, Honeywell, GE Healthcare, Miele, Nvidia, Toyota, Tractable, Continental, Fanuc, GE renewable energy. Stanford, Harvard, Genomics Institute of Novartis Research Foundation, National Institute of Health, Cancer Research UK.","excerpt":"The Darwin Platform (from V7 labs) has access to its automated image annotation and neural network training with active learning that create a pixel-perfect annotation, which is 10x faster than traditional tools.","categories":["Deep Tech"],"tags":["annotation","Automation","computer vision dataset","data annotation","data labelling"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-11-30T13:00:21","publication_year":"2020","word_count":590,"keywords":["computer vision dataset","TPU","AI","neural network","PyTorch","retail AI","Automation","annotation","image recognition","computer vision","data labelling","Python","Aim","data annotation","deep learning"],"extracted_tech_keywords":["AI","deep learning","neural network","computer vision","Aim","PyTorch","image recognition","retail AI","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/v7\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":36885,"title":"How YouTube’s ML Algorithm Earned Billions For Music Producers By Building Fingerprints For Songs","content":"Before the rise of audio streaming services such as Gaana, Saavn and JioMusic, the premier destination for music was YouTube. A large number song from  Bollywood record labels were present on the platform for users’ listening pleasure. It was a model that worked for everyone, with the user not being charged for listening to the music and the copyright holders benefitting. Companies such as T-Series rode this trend to its peak and is now going after the crown of YouTube’s most subscribed channel. Taking a step back and looking at the bigger picture, the question arises as to how one of the most commercialised and copyright-heavy channels is so accessible to the masses. In a world of rampant piracy and copyright infringement, YouTube exists in the intersection of paywalling and facilitating theft. As with every new frontier, the appointed watchdog and guardian of YouTube’s copyright-friendly atmosphere emerged. This is the system we know today as Content ID. Why Copyright Was The First Frontier The Internet was a free space in its infancy, and could very well be compared to the Wild West when it first emerged. As the net was opened to the masses, software pirates rushed on the scene and created an impression on record labels and producers. A message that the Internet was not a place that copyright could not be enforced, at least not in any sense of power. Websites were protected against legal action by the safe harbour provisions of the Digital Millennium Copyright Act. This specifies that copyrighted content could exist on websites as long as there was a way to remove it. Record labels and music producers kept away from putting music online for this very reason. Copyright infringement was the bane of their existence, sucking away at what little profits they were making as the Internet took over the world. Due to the free nature of the DMCA guidelines, a pirate could very easily make duplicates of a song posted on a platform by a copyright holder, leading to infringement and theft. When Google acquired YouTube, this is the precise issue they wished to avoid. Examples such as Grooveshark and Napster stuck to the top honchos at Google, with a solution required quickly to avert the possibility of being sued by record labels. This prompted them to create a system that would automatically flag copyrighted content as early as 2007. They looked to make YouTube into a platform where companies could upload copyrighted content and get compensated for it without unauthorised pirates trying to capture market share. What Is Content ID Their solution was with Content ID. The system gave rights holders an automated way of finding unauthorised copies of videos and songs. Then, they were given the choice to block the content or run ads against it. Videos were checked at upload to detect copyright-protected material. This can then be blocked by the copyright holder, muted in case of audio infringement, or restrict the video from playing on certain platforms. However, if the holder wishes to keep the content on the platform, they can choose to monetise it with ads and take the revenue from the views for themselves. The system is treated as the first line of defence, and if the claim was made without basis, it can be refuted by the creator.  YouTube revealed the success of the venture, as the algorithm generated $2 billion in revenue for rights holders. It also represents a proactive method of handling problems, handling 98% of copyright management on YouTube. How Does It Work As mentioned previously, the Content ID systems scan videos at upload to compare them against a reference library. This library contains over 50 million copyrighted works, provided by holders, which add up to a combined watch time of more than 500 years. This content comes from thousands of hand-selected partners, leaving small chinks in the armour of the copyright system. It uses audio and video fingerprinting technology to detect matches between videos people upload and the reference files. Videos see a frame-by-frame analysis for fingerprinting, with identical images tackled by the use of a heat map visualisation that compares frame data from two videos side by side. The system utilises a finite-state transducer algorithm for fingerprinting music. This allows it to detect changes such as beeps in the middle of songs, pitch, volume and speed changes, along with audio overlays and effects. To power this large amount of compute, YouTube harnesses the power of Google’s Brain deep learning system. Content ID directly functions on this platform, providing multiple advantages. The deep learning framework also allows developers to make any changes that cause Content ID to fail, such as flipping video or changing aspect ratios. The deep learning smarts of Google’s system allows for a much more organic and easy way to launch new iterations of the fingerprinting system. The neural network can be trained easily and much faster. The Future Of Automated Copyrighting Even as the music industry is largely moving away from a focus on copyright to more pressing matters such as artists receiving remuneration, YouTube’s move is one of the most important in the space. It not only ensured the future of the platform as we know it today but also set the standard for other platforms such as Twitter to engage in copyright protection for content on their platform. Content ID thus served its purpose as a stopgap into a world of more complex issues in the music industry while ensuring a fair outcome for everyone.","excerpt":"Before the rise of audio streaming services such as Gaana, Saavn and JioMusic, the premier destination for music was YouTube. A large number song from  Bollywood record labels were present on the platform for users’ listening pleasure. It was a model that worked for everyone, with the user not being charged for listening to the […]","categories":["AI Features"],"tags":["copyright","YouTube"],"author_name":"Anirudh VK","publish_date":"2019-03-26T11:39:59","publication_year":"2019","word_count":919,"keywords":["Go","programming_languages:R","AI","neural network","programming_languages:Go","Git","copyright","deep learning","Aim","GAN","YouTube","R"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/youtubes-ml-algorithm-earned-billions-music-producers-building-fingerprints-songs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10140786,"title":"Rabbitt AI Announces Strategic Applications of Generative AI in Defense","content":"Indian AI startup Rabbitt AI has launched a suite of GenAI tools to reshape military operations by minimising human involvement in high-risk zones. The core idea centres on reducing human exposure to danger. GenAI-powered drones, autonomous vehicles, and surveillance systems enable real-time threat detection and response, offering a safer, AI-driven alternative to traditional security methods. By incorporating diverse sensor data—from infrared and radar to audio and visual feeds—Rabbitt’s models detect unauthorised movements, environmental anomalies, and abnormal activities without human intervention. “We are not very far from a future where AI with limbs can dominate battlefields,” said Harneet Singh, Rabbitt AI’s chief, who was previously an AI consultant to the South Korean Navy. “One of our key missions is to protect lives at the borders by creating situationally aware, autonomous AI systems that can respond to threats by observing and analyzing sensor data in real-time,” he added. Rabbitt AI’s technology integrates deep learning models with diverse sensor inputs—from infrared and radar to audio and video—to detect unauthorised intrusions, environmental anomalies, and abnormal activities without human intervention. Singh highlighted the technology’s autonomy, saying, “With AI-powered systems, we can now provide uninterrupted, unbiased monitoring that ensures both coverage and efficiency, all while reducing operational costs.” In addition to reducing personnel risks, Rabbitt’s GenAI tools also help streamline resources by automating many surveillance functions. The technology minimizes reliance on human labor, which Singh says “not only reduces costs but increases accuracy, freeing military personnel to focus on strategic tasks.” The system’s AI-driven detection capabilities also lower the need for costly corrective actions. Rabbitt AI is also advancing “human-machine teaming” by pairing GenAI with unmanned drones and ground vehicles to increase adaptability in hard-to-reach terrains. According to Singh, “This tech enables real-time situational awareness, allowing command centres to get immediate insights without the delay of human reporting, even in complex environments like urban areas or mountainous regions.” Singh, an IIT Delhi alumnus recognised by DRDO and Indian military officials with an honorary medal, emphasised Rabbitt AI’s broader vision for defence. “Our work goes beyond developing AI models,” he said. “We are building a defence ecosystem where AI serves as a force multiplier, enhancing every soldier’s capabilities while increasing situational awareness and reducing decision-making time.” Founded by Singh, Rabbitt.ai focuses on generative AI solutions, including custom LLM development, RAG fine-tuning, and MLOps integration. The company recently raised $2.1 million from TC Group of Companies and investors connected to NVIDIA and Meta. The company recently appointed Asem Rostom as its Global Managing Director to lead expansion across the MENA and Europe regions. Before this role, Rostom served as the managing director at Simplilearn. The company has also launched Rabbitt Learning, a new division focused on transforming education access and workforce readiness in the MENA region. As a part of its expansion, Rabbitt AI has opened a new office in Riyadh, Saudi Arabia, to meet the growing demand for Gen AI skilling courses and digital transformation projects in the Gulf countries.","excerpt":"Rabbitt AI’s technology integrates deep learning models with diverse sensor inputs—from infrared and radar to audio and video—to detect unauthorised intrusions, environmental anomalies, and abnormal activities without human intervention.","categories":["AI News"],"tags":["AI in Defense Sector","GenAI","Generative AI","Rabbitt AI"],"author_name":"Siddharth Jindal","publish_date":"2024-11-11T18:27:59","publication_year":"2024","word_count":493,"keywords":["Go","GenAI","AI","ML","MLOps","Git","RAG","deep learning","generative AI","Rabbitt AI","Generative AI","R","AI in Defense Sector"],"extracted_tech_keywords":["AI","ML","deep learning","generative AI","GenAI","MLOps","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/rabbitt-ai-announces-strategic-applications-of-generative-ai-in-defense\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040650,"title":"How Google&#8217;s Password Manager Keeps Brute Force Attacks At Bay","content":"In today’s profoundly digital world, people are likely to have many more profiles and accounts on the internet than ever before. This can mean some great things: a social media presence means you get to keep in touch with your friends—especially during a pandemic—and accounts for newspapers, online banking, or educational websites make things accessible for you. So, what’s the catch? Each of these accounts that you make requires you to remember a password for security reasons. Remembering these passwords is difficult—unless you’re Sheldon Cooper—and keeping the same password for everything, something many people do, substantially increases the risk of a third party breaching your privacy. Where does Google come in Source: Google Password managers are programmes that allow users to store, generate and manage their passwords online. Most browsers such as Chrome and Safari come with password managers built-in. However, Google is adding a new feature to its Chrome password manager. The feature will warn users about stolen passwords and then fix them as well. At the ongoing Google I\/O event, Google’s Jen Fitzpatrick, announced that the search giant will add four new features to its Password Manager. First, Google will be launching a tool that will import any passwords a user uses stored in third-party managers such as LastPass to Chrome’s Password Manager. This would make for easier switching between Password Managers, making Google’s new tool more attractive to users. Secondly, Google will further integrate this into PCs and Android devices. This would allow devices that use either the Chrome web browser or the Chrome OS to use the same account passwords on both platforms. Additionally, the Password Manager will launch a notification alert to let users know about any compromised passwords. One of the main features of Google changed Password Manager is that it will enable Chrome to help users change their passwords with ease. Thus, if a user checks their password and gets an alert for a compromised password, the Assistant will show them a ‘Change Password’ button. Upon tapping this, Google’s AI Assistant will go through the entire process of changing the password for them. This makes it considerably easier to be safe on the internet. People who are not comfortable with AI fixing their passwords can still use the manual option on the Password Manager. Chrome’s password manager will still help the user by suggesting solid and unique passwords for their accounts—like it currently does. Google’s Duplex on the Web technology will power this new feature. Duplex was introduced in 2019 to allow Google Assistant to help users complete tasks such as purchasing movie tickets, checking in to flights and ordering food. Duplex on the Web will enable Assistants to take over more monotonous tasks on the internet, such as scrolling and filling out forms. The technology is being used to quickly create a strong password for sites and apps on Chrome, signalling a potential breach alert. Users can also find such features in some third-party password managers, such as Dashlane. The main objective here is to dissuade users from keeping one or two passwords for every online account whilst making it easier for them to maintain their privacy. So far, this seems like a great idea from Google, and we’ll have to wait until it is properly implemented to see the platform’s workings better. Google says that the feature will be launched gradually, first to American users of its Chrome for Android and then will be more available in other regions in the ‘coming months.’","excerpt":"In today’s profoundly digital world, people are likely to have many more profiles and accounts on the internet than ever before. This can mean some great things: a social media presence means you get to keep in touch with your friends—especially during a pandemic—and accounts for newspapers, online banking, or educational websites make things accessible […]","categories":["Global Tech"],"tags":["Google Chrome"],"author_name":"Mita Chaturvedi","publish_date":"2021-05-22T17:00:00","publication_year":"2021","word_count":582,"keywords":["Go","programming_languages:R","AI","Google Chrome","Git","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-googles-password-manager-keeps-brute-force-attacks-at-bay\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":6253,"title":"Introduction to Apache Tez","content":"This article is intended as a research brief on Apache Tez, an emerging open source technology. Tez is useful for a number of applications which need to execute a series of Map-Reduce jobs (chaining). By executing the series of MR tasks as a single job and cutting down the consequent intermediate reads and writes from\/to HDFS, Tez can execute a Directed Acyclic Graph (DAG) of MR jobs very efficiently. It is the basis for Hive 13 or Stinger, the upcoming version of Hive and can speed up Pig scripts significantly. Motivation Hadoop, the open source implementation of the Google Map-Reduce paper [1], has enabled end users to process large amounts of data. Consequently, it has been used in several contexts, both for use cases that are well suited for it and also for uses cases where it has been force-fitted. Data crunching, especially the embarrassingly parallel kind is ideally suited for Hadoop. However, it has been observed by a number of researchers that Hadoop 1.0 (the plain MR version of Hadoop) is not well suited for real-time scenarios [2] or for machine learning involving iterative processing [3] [4] [5].  By iterative processing, we mean the following kind, as depicted in the figure below. The main reason why Hadoop 1.0 is not well suited to iterative processing is the repeated read\/write from\/to the Hadoop Distributed File System (HDFS) for every iteration – there are no long lived MR jobs and every iteration has to be realized as a fresh MR job, with data being initialized by reading from HDFS and written back to HDFS at the end of the iteration. Moreover, the termination condition for iteration, which may itself involve certain computation, may also have to be realized as a separate MR job. With the advent of Hadoop 2.0 or Hadoop Yet Another Resource Negotiator (YARN), the need for executing a workflow or a DAG of MR jobs becomes critical. It should be noted that Hadoop YARN separates out the processing paradigm (MR) from the resource management functions, which were tied up together in Hadoop 1.0. Thus, in Hadoop YARN, the resource management functionality is handled by the YARN resource manager, while several frameworks can work on top of YARN. The possible frameworks that may work on top of YARN include Spark for iterative processing, Storm for real-time processing, GraphLab\/Giraph for graph processing as can be evidenced from the diagram given below [8]: Thus, one can say that Hadoop YARN has enabled several processing frameworks to co-exist and process HDFS data in various ways. Consequently, this motivates the need for an application that can efficiently execute a DAG of tasks. The vanilla Hadoop 1.0 way of expressing DAG as a chain of individual MR jobs can significantly impact the efficiency. This is where Tez fits in – it provides a framework for executing DAG of tasks efficiently. For instance, a single SQL query or a simple PIG script can be executed as a single DAG in Apache Tez. Apache Tez Tez allows an application to be modeled as a data flow – with edges representing movement of data, while vertices represent data processing tasks. It should be noted that high level query languages such as Pig and Hive produce a DAG for execution after processing the query\/script. This, Tez fits in perfectly as an executor for Pig and Hive jobs – it can speed up such jobs significantly. Given below is an example of a two stage MR application – data is sorted topologically in the first stage, while the frequency of occurrence is computed in the next stage. This can be expressed as two traditional MR jobs with intermediate results placed in HDFS. The same can also be expressed as a Pig script as follows: When executed on Tez, the same Pig script runs as a single job that is expressed as a simple DAG comprising two stages. Tez takes care of expressing this logical graph as a physical graph of tasks and their dependencies and executes them efficiently on a cluster of nodes. Tez Internals Every vertex of the dataflow graph can be modeled as a combination of input, processing and output modules – where the input module specifies the set of edges required for this task as input, while the processing refers to the data transformation as part of this vertex and output refers to the output data written by this transformation and passed through the incident out edges of this node\/vertex. The vertex can be viewed as equivalent to a task in Tez. The parallelism of the task can be specified at DAG construction time or via user plugins running in the Application Master (AM), a component of Hadoop YARN. The context\/configuration information (environment or application specific) is provided to the three modules through an initialization routine which gets invoked first. Subsequently, the run method of the Processor is invoked for every task instance (based on parallelism factor). Once this method finishes, the task is logically completed. The output (via the LogicalOutput class) in addition to writing out the data from the Processor, also provides information to subsequent \/downstream Input stages (for a chain of tasks or a typical DAG). Tez also provides sophisticated error handling to handle both fatal errors (that require termination of the task) and non-fatal errors (that may require re-reading the input, for instance). One of the unique features of Tez is the ability to dynamically optimize the DAG execution – the information available at runtime such as data samples and sizes is used to optimize the DAG execution plan. Tez works perfectly with Hadoop YARN and can negotiate with the resource manager component of YARN and accept YARN containers for execution. Conclusions This research brief has explained the basics of Apache Tez and how we at Impetus have started using it. The future of Tez looks bright, as Tez is being integrated by the Pig community as their execution engine. Concurrent Inc. is also moving to Tez as the execution engine for Cascading. Tez is also being used by the Hive community as the basis execution engine for Hive 13. The Mahout community is also evaluating Tez as an alternative to MR. The Tez community have also tested version 0.3 of Tez on large clusters (beyond 300 nodes) and on large data sets (beyond 30 Terabytes). We are also planning to evaluate the possibility of realizing large scale distributed deep learning networks using Apache Tez. Authors: Dr. Vijay Srinivas Agneeswaran, Director, Big Data Labs, Impetus Infotech India Pvt. Ltd. Email: vijay.sa@impetus.co.in Inelu Nagamallikarjuna Reddy, Software Engineer, Big Data Labs, Impetus Infotech India Pvt. Ltd. Email: inelu.naga@impetus.co.in","excerpt":"This article is intended as a research brief on Apache Tez, an emerging open source technology. Tez is useful for a number of applications which need to execute a series of Map-Reduce jobs (chaining). By executing the series of MR tasks as a single job and cutting down the consequent intermediate reads and writes from\/to […]","categories":[],"tags":[],"author_name":"Vijay Srinivas Agneeswaran","publish_date":"2014-10-08T16:21:18","publication_year":"2014","word_count":1104,"keywords":["big data","Go","machine learning","TPU","programming_languages:R","AI","programming_languages:SQL","deep learning","SQL","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","TPU","R","SQL","Go","big data","programming_languages:R","programming_languages:SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/introduction-apache-tez\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10075661,"title":"PyTorch releases free tutorials on Fully Sharded Data Parallel (FSDP)","content":"PyTorch has announced a new series of 10 video tutorials on Fully Sharded Data Parallel (FSDP) today. The tutorials are led by Less Wright, an AI\/PyTorch Partner Engineer and who also presented at Nvidia Fall GTC 2022. Introducing what the users will be learning, Less Wright says, “Whether you are training a 100 million or 1 trillion model parameter model, the series will enable users to train the models more efficiently, along with short deep dives of various aspects of FSDP.” Wright believes that the main goal of the 10-part series is to help build expertise on leveraging FSDP for distributed AI training. He also says that the series will be added with new videos, along with features to the FSDP. Source: YouTube For instance, the first series titled, ‘Accelerate your training speed with the FSDP Transformer wrapper’, consists of a tutorial on how to utilise the new FSDP transformer wrapper. Unlike the default wrapper which makes sharding choices based on parameter count, this transformer wrapper understands how the model operates—locating appropriate breaks to shard. To put it in simple terms, it lets users know how to implement the transformer wrapper and increase the model’s training speed by up to 2x. Other parts of the series include FSDP Mixed Precision Training, Sharding Strategies, Backwards Prefetching, and Fine Tuning Models. Meta recently announced the PyTorch project to be part of the non-profit Linux Foundation—newly launching as PyTorch foundation. The main goal would be to drive adoption of AI and deep learning tooling—fostering and sustaining an ecosystem of open source—and vendor-neutral projects with PyTorch.","excerpt":"The tutorial’s main goal is to help build expertise on leveraging FSDP for distributed AI training and awaits upcoming addition of new videos to the series.","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-09-20T15:08:46","publication_year":"2022","word_count":262,"keywords":["Go","programming_languages:R","PyTorch","AI","programming_languages:Go","RAG","deep learning","ai_frameworks:PyTorch","R"],"extracted_tech_keywords":["AI","deep learning","PyTorch","RAG","R","Go","ai_frameworks:PyTorch","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-releases-free-tutorials-on-fully-sharded-data-parallel-fsdp\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":37387,"title":"Top 10 Free Online Resources For Learning Scala","content":"Scala, the modern multi-paradigm programming language is the next generation processing engine for Big Data. It has been evolving at a very fast pace since the last few years. According to reports, Scala has jumped a whopping 11 positions on the chart since May 2017 to land at #20. This isn’t the first time Scala cracked a top slot, but it might be a sign for future relevance. It has become one of the must-have tools in a data scientist’s toolbox. In this article, we list five free online resources for Scala by which you can kickstart your learning today. 1| A Complete Beginner’s Tutorial To Learn Scala By All About Scala Overview: This tutorial by allaboutscala.com provides a complete beginner’s tutorial which will help you to learn Scala in a very simple and easy way. There are a total of nine chapters in this tutorial where you can learn the basic foundation of Scala, immutable collection, collection functions, Scala test, introduction to Apache Spark 2, understanding IntelliJ, traits and dependency injection, mutable collection, and other relevant topics. Besides Scala, the tutorial also includes Scala cheat sheet which shows small code snippets and answers some relevant questions, various frameworks which you can use, introduction to Big Data and data science. Click here to start learning the course. 2| Scala Tutorial By Intellipaat Overview: In this tutorial, you will get a detailed understanding of Scala, integration of object-oriented and functional languages, how to install Scala, Scala operators, array, strings, collections, statements, loops, etc. The learning package will also help you make your Scala programs with proficiency and efficiency. This is a concise and descriptive tutorial where you can learn pattern matching and class classes along with regular expressions, exception handling, and other relevant topics. Click here to start learning the course. 3|  Scala Official Documentation Overview: What can be more resourceful than the official documentation? The Scala official documentation is a free learning resource for beginners and intermediate. It contains various guides and tutorials starting from how to install Scala on your computer, understanding the basics of Scala, basic introduction to core language features. This documentation also includes an in-depth guide on how to write idiomatic Scala code, cheat-sheet covering the basics of Scala’s syntax and features, API documentation of every version of Scala, Scala’s formal language specification, and other such relevant topics. Click here to start learning the documentation. 4| Scala Tutorial By Tutorials Point Overview: This tutorial explains the basics of Scala in a simple and reader-friendly way. It has been prepared for the beginners to help them understand the basics of Scala in simple and easy steps. Here, you will learn the basics of this language, how to set up, basic syntax, data types, variables, classes, objects, access modifiers, operators, functions, closures, traits, pattern matching, regular expressions, exception handling, extractors, and other such relevant topics. After completing this tutorial, you will find yourself at a moderate level of expertise in using Scala from where you can take yourself to the next level. Click here to start learning tutorial. 5| Scala 101 By Cognitive Class Overview: This is a free course offered by Cognitive Class where you can learn the foundations of the language, understand how to tackle data analysis problems involving Big Data, Scala and Spark, understanding the fundamentals of the language, the tooling, and the development process, develop a good appreciation of more advanced features. The course syllabus contains 5 modules where you can gain insight into the basics of Scala, basic object-oriented programming, case objects and classes, collections including idiomatic Scala. There is no fixed duration for learning this tutorial and can be audited as many times as you wish. Click here to start learning the course. 6| Introduction to Programming and Problem Solving Using Scala Video Overview: This video series on YouTube helps in understanding the basics of Scala that include how to program and solve problems using Scala, how to use command lines, directories and navigation and vi basics. One will also understand how to set up Scala, objects and methods, string methods, sequential execution, recursion, and other such relevant topics. Click here to start learning. 7| Scala Tutorials By TekSlate Overview: This blog includes basic knowledge of Scala for beginners. The topics covered are an introduction to Scala and its features, operators, object system, collections, data types, how to do configurations from a file, configurations from command line parameters, other relevant topics. Click here to start learning the article. 8| First Steps To Scala By Artima Overview: In this article, you’ll follow twelve steps that are designed to help you understand and gain some basic skills in the Scala programming language. The topics include the basic guides to install Scala, how to use scala interpreter, variables, methods, Scala scripts, how to iterate, parameterize, and other relevant topics. Click here to start learning the article. 9| Scala School Overview: Scala school started as a series of lectures at Twitter to prepare experienced engineers to be productive Scala programmers. These lectures assumed the audience knew the concepts and showed how to use them in Scala. The focus will be on the interpreter and the object-functional style as well as the style of programming on Scala. Most of the lessons require no software other than a Scala REPL. Click here to start learning the tutorial. 10| Scala Programming Tutorial Video Overview: This is a 5-hour video course which is aimed at beginners who have no prior programming experience or those who have limited knowledge of Scala. This course can get you up and running and will give you the skills to master Scala.","excerpt":"Scala, the modern multi-paradigm programming language is the next generation processing engine for Big Data. It has been evolving at a very fast pace since the last few years. According to reports, Scala has jumped a whopping 11 positions on the chart since May 2017 to land at #20. This isn’t the first time Scala […]","categories":["AI Trends"],"tags":["scala"],"author_name":"Ambika Choudhury","publish_date":"2019-04-08T06:04:15","publication_year":"2019","word_count":937,"keywords":["big data","data science","Go","API","AI","Apache Spark","Scala","scala","Ray","Aim","R"],"extracted_tech_keywords":["AI","data science","Aim","Ray","Apache Spark","R","Go","Scala","API","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-free-online-resources-for-learning-scala\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10064967,"title":"Interview with Ajinkya Bhave, Director (Engineering Services), Siemens","content":"AGI as the true north star of artificial intelligence? What is fast technologies’ effect on the environment? Are digital twins the next big thing? Analytics India Magazine caught up with Ajinkya Bhave, Director and Country Head (India) – Engineering Services at Siemens Digital Industries Software, to get answers to these questions and more. Edited excerpts: AIM: How did you get interested in artificial intelligence? Ajinkya Bhave: In the last year of my Bachelor’s that I was pursuing at Mumbai University, we had a course on robotics and AI; this ignited my whole interest in the field, and by extension, in machine learning. From there on, I did my Masters’s in Robotics from Carnegie Mellon University (CMU). I realised that Robotics is a huge field, and I was more inclined toward the machine learning and control systems aspects of it. I went on to pursue a second master’s degree and a PhD in control systems, with autonomous systems being the backdrop. So, in my education, robotics was the focus of my interest and passion, which is the case even today, and machine learning is an enabler for that. I have been working with Siemens for 10 years now in its Engineering Services (ES) group. We are a unique group within the Siemens Digital Industries Software business unit. We do not work directly on a product but offer state-of-art customised solutions to our clients, combining engineering domain expertise and Siemens tools and frameworks. Currently, I head the ES India group with three teams in the field of control systems, systems simulation, and computational fluid dynamics. Each team has exceptionally talented and motivated members, led by an experienced technical manager. The control systems team focuses majorly on machine learning and autonomous driving. I have gone from working on control systems to then transitioning to machine learning and artificial intelligence, working in both academia and now in an industry role at Siemens. So, we went from traditional control and automation all the way to machine learning and increasingly into AI. AIM: Having been part of the AI and machine learning community for a long time, what are the most common misconceptions you would like to bust? Ajinkya Bhave: There are two concerning trends I see in the AI\/ML domain today. The first one is the use of AI as the buzzword for everything. At Carnegie Mellon, when we were taught about AI and machine learning, we were also made aware of the difference between the two – the important fact that ML is a subset of AI. Many of today’s systems use deep learning, but we call them AI generically, which is not completely correct. Our team at Siemens is very specific that we are a machine learning group and not a general AI group. The second concern is when people say that a neural network is a model of the human brain. It’s not their mistake; the conversation around this topic has largely been misleading. If you talk to a biologist, they will tell you that the human neuron is extremely complex and advanced and that the machine learning model of a neuron does not even come close to that. There are synapses, dendrites and activation functions, and there are multiphase signals going on in a biological neuron. Each human neuron is like one deep neural network, as a recent study pointed out. A single human neuron is a computational factory by itself. AIM: We all talk about how machine learning and deep learning have been breakthrough technologies. However, we often overlook the newer challenges they bring with themselves. One of them is the effect these fast technologies have on the environment. What are your views on that? Ajinkya Bhave: As we start computing more and more, we are taking up a lot of power, which will become larger in the coming future. A classic example of this is the GPT-3 model made of ~175 billion parameters. At one point, it becomes an overfitting problem. As John von Neumann said, “With four parameters, I can fit an elephant, and with five, I can make him wiggle his trunk.” It’s true because, at some point, models are no longer intelligent; they are fitting to the data. It depends on the model, the domain, and the data, but I have been seeing a trend when you have a larger model trying to go for complex problems; it doesn’t really solve the problem; it gives an illusion of solving the problem. This is a problem that machine learning practitioners must think of. There is a certain elegance in having the right model for the job. A new trend should be going from large models to models that are efficient enough to do the job. AIM: Artificial general intelligence is considered to be the true north star of AI for some. What are your views on this? Ajinkya Bhave: I believe that the way we do machine learning and AI today, I don’t think we can have machines that think like humans. Fundamentally, when you look at intelligence, it is evolving mainly because there is life and consciousness. There is the will to survive. My feeling is that unless you can teach AI about death, you will not make it intelligent. Unless AI knows and fears death, it will not know and independently evolve strategies for how to survive. Some people say we can penalise the systems in the right way to simulate evolution or use reinforcement learning to learn the right survival actions based on rewards. But it is not the same as the machine semantically understanding and avoiding death. Life and intelligence are complex phenomena, and I don’t believe you can compress them into an explicit formula. You can’t make machines’ intelligent’ by a pure deep learning approach; at best, what you can do is mimicry. AIM: We have been seeing massive interest in digital twins in machine learning’s context. What is its future? Ajinkya Bhave: In many cases, these terms are used as buzzwords, mainly because that is how the industry operates. But in our case, Siemens invests a lot in digital twins; it’s not a buzzword for Siemens. Digital twins differ from simulation models because of the amount of detail and fidelity that goes into the twin. Yes, you can simulate anything to a certain point but to develop the detail and realism and to keep the digital twin always updated with the current state of the physical system is what differentiates the two. Typically, in Siemens, a digital twin is a detailed actual scaled model of the hardware. It is a living, digitised embodiment of your plant. You can do a lot of studies on that which you may not be able to do otherwise either because of the time, access, or cost investment. The use of digital twins across the full spectra of engineering applications has increased dramatically in recent years and continues to grow rapidly. AIM: What are your comments on AI as snake oil? Ajinkya Bhave: When AI is used so loosely like that, it is usually by people who are not that conversant with the technology. People practising in the field don’t make AI the ultimate thing to go to. At Siemens, when a client comes to us, we don’t say we will use AI as the first approach. We examine the problem, and only if AI or machine learning is applicable we suggest that as a solution. Since we are an engineering group, we don’t take an AI-first approach. We take an AI-guided approach. A lot of our preferred approaches come from physical systems dynamics. We use machine learning in a selective way, applying it only in cases where it lends an efficiency edge that traditional approaches may fail to do. AIM: What are your tech predictions for 2022? Ajinkya Bhave: It is difficult to predict technology. However, I think two things might happen. The first is more combination of machine learning with digital twins. Because data is hard to get in the real world, that is why you have to augment the data via simulation. And how do you use simulation smartly and efficiently? — it is via machine learning. Using digital twins to train machine learning models would increase as companies scale up. So, if the data is close to reality, the model would be trained well and eventually be able to predict accurately. The second would be the scaling down of the machine learning model. This is not an easy thing to do. Scaling down the model means gaining more insights into its architecture and semantics. You have to take out layers from the model to understand what the model is doing. So, linked with the scaling of models will be Explainable AI. As an extension, companies would invest more in the verification and validation of these models. This would not be just limited to autonomous vehicles. Models should be trustable and, at least, supervisable so that they do not do something stupid when deployed to real-world scenarios.","excerpt":"When you look at intelligence, it is evolving mainly because there is life and consciousness. There is the will to survive. My feeling is that unless you can teach AI about death, you will not make it intelligent.","categories":["AI Features"],"tags":["AGI","Carnegie Mellon University","Interviews and Discussions"],"author_name":"Shraddha Goled","publish_date":"2022-04-14T11:00:00","publication_year":"2022","word_count":1504,"keywords":["Go","Carnegie Mellon University","artificial intelligence","machine learning","AI","neural network","ML","Aim","deep learning","AGI","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","analytics","Aim","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-with-ajinkya-bhave-director-engineering-services-siemens\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10049869,"title":"How to Improve Collaborative Filtering with Dimensionality Reduction?","content":"Along with a large number of applications across the domains, the recommendation systems are increasingly challenged with the issue of huge data and sparsity. It raises concerns about computing costs as well as the inadequate quality of the recommendations. To address this issue, dimensionality reduction techniques are employed in recommendation systems to minimize processing costs and enhance prediction. In this post, we will discuss the collaborative filtering approach of recommendation along with its limitations. We will also try to understand the problem of sparsity faced in this approach and how it can be addressed with dimensionality reduction techniques. We will cover the following major points in this article to understand this concept in detail. Table of Contents Collaborative FilteringThe Approach to Collaborative FilteringLimitation of Collaborative FilteringDimensionality Reduction in Collaborative FilteringSingular Value DecompositionPrincipal Component Analysis Let’s start with understanding collaborative filtering. Collaborative Filtering Collaborative filtering is a famous technique used in most recommendation systems. Generally, collaborative filtering is categorized into two senses: the narrow one and the more general one. Collaborative filtering, in a narrower sense, is a method of creating automatic predictions (filtering) about a user’s interests by gathering preferences or taste information from a large number of users (collaborating). The collaborative filtering strategy is based on the concept that if person A and person B have the same opinion on a topic, A is more likely to have B’s perspective on a different topic than a randomly selected person. For example, given a partial list of a user’s tastes, a collaborative filtering recommendation system for electronics accessories purchases preferences could offer predictions about which accessories show the user would like to purchase (likes or dislikes).  On other hand, in a more general sense, it is the process of searching for information or patterns using strategies that involve several agents, viewpoints, data sources, and so on. Source Collaborative filtering applications are generally used with very big data sets. Collaborative filtering methods have been applied to many different types of data, including sensing and monitoring data, such as in mineral exploration, environmental sensing over large areas or multiple sensors; financial data, such as financial service institutions that integrate many financial sources; and electronic commerce and web applications where the focus is on user data, among others. Approaches to Collaborative Filtering The majority of collaborative filtering-based recommender systems create a community of like-minded clients. As a measure of closeness, the neighbourhood creation method often uses Pearson correlation or cosine similarity. These algorithms generate two sorts of recommendations after determining the nearby neighbourhood those are, Approximation of how much a client C will enjoy a product P. In the case of a correlation-based algorithm, the prediction on product ‘P’ for customer ‘C’ is derived by computing a weighted total of co-rated goods between C and all of his neighbours and then adding C’s average rating to that. This may be represented using the formula below. The prediction is tailored to consumer C. Source In the above expression, rCJ denotes the correlation between user C and neighbour J and JP is the J’s rating on the product P. Recommendation of a product list to a client C. This is sometimes referred to as a top-N suggestion. After forming a neighbourhood, the recommender system algorithm concentrates on the items evaluated by neighbours and selects a list of N products that the client will like. Limitations of Collaborative Filtering These systems have been effective in a variety of fields, however, it does not always successfully match things to a user’s preferences. Unless the platform achieves exceptionally high levels of diversity and independence of opinion, one point of view will always predominate over another in a given group. Based on such situations the algorithm has been claimed to have certain flaws, those are: Sparsity Because nearest neighbour algorithms rely on exact matches, they sacrifice recommender system coverage and accuracy. Because the correlation coefficient only applies to consumers who have evaluated at least two products in common, many pairings of customers have no correlation at all. Many commercial recommender systems are used in practice to analyze vast product sets (for example, Amazon.com suggests books). Even active customers may have rated far under 1% of the products in these systems (1% of 2 million volumes is 20,000 books—a vast set on which to form an opinion). As a result, Pearson’s nearest neighbour algorithms may be unable to offer numerous product recommendations for a single consumer. This is known as reduced coverage, and it is caused by sparse neighbour ratings. Furthermore, because only a limited amount of rating data can be provided, the accuracy of suggestions may be poor. Scalability Nearest neighbour methods imply computation that rises in parallel with the number of consumers and items. A conventional web-based recommender system employing existing algorithms will have major scalability issues with millions of clients and products. Synonymy In the real world, several product names can relate to the same thing. Correlation-based recommender systems are unable to detect this hidden relationship and handle these products differently as a result. Consider two customers who each score ten different recycled letter-pad items as “high,” and another customer who rates ten different recycled memo pad products as “high.” Correlation-based recommender systems would not be able to compute correlation because there would be no match between product sets, and they would be unable to find the latent relationship that they both like recycled office supplies. Dimensionality Reduction in Collaborative Filtering In general, dimensionality reduction is the process of mapping a high-dimensional input space into a reduced level of latent space. Matrix factorization is a subset of dimensionality reduction in which a data matrix D is reduced to the product of many low-rank matrices. Singular Value Decomposition (SVD) SVD is a powerful dimensionality reduction technique that is a specialization of the MF approach. The primary issue in an SVD is to discover a reduced dimensional feature space. As SVD is a matrix factorization technique, a m x  n matrix R is factored into three matrices as follows: R = U ⋅ S ⋅V ′ Here, S is a diagonal matrix with all singular values of R as diagonal elements, whereas U and V are two orthogonal matrices. All of the entries in matrix S are positive and are kept in descending order of magnitude. In recommender systems, SVD is used to accomplish two different tasks: First, it’s used to collect latent relationships between consumers and products, allowing us to calculate a customer’s estimated likelihood of purchasing a specific product. Second, it’s used to make a low-dimensional representation of the original customer-product space and then compute the neighbourhood in that space. It’s then used to provide a list of top-N product suggestions for customers. Principal Component Analysis (PCA) PCA is a sophisticated dimensionality reduction technique that is a specific application of the MF approach. PCA is a statistical process that employs an orthogonal transformation to turn a set of possibly correlated observations into a set of values that are linearly uncorrelated variables known as Principal Components (PC). The number of PCs is less than or equal to the initial variable count. This transformation is defined in such a way that the first PC has the most variance feasible and each subsequent component has the greatest variance possible while being orthogonal to the preceding components. Because they are the eigenvectors of the covariance matrix, the principal components are orthogonal. The relative scaling of the original variables affects PCA. Conclusion So here in this post, we have seen collaborative filtering, any kind of ML algorithm if exposed to high dimensional data it will return bad predictions, the collaborative filtering also experiences the same thing. To address this issue we first understand the approach of CF and under which circumstances it fails. Based on flaws we have seen how SVD and PCA can be used to address this issue. References Dimensionality Reduction and Clustering TechniquesAn Item-Based Collaborative FilteringApplication of Dimensionality Reduction in Recommender System – A Case Study","excerpt":"Collaborative filtering is a famous technique used in most recommendation systems. Generally, collaborative filtering is categorized into two senses: the narrow one and the more general one.","categories":["AI Trends"],"tags":["Collaborative Filtering","Data Science","Dimensionality Reduction","Machine Learning","PCA","recommendation engine"],"author_name":"Vijaysinh Lendave","publish_date":"2021-09-28T14:00:00","publication_year":"2021","word_count":1327,"keywords":["big data","Go","AWS","AI","ML","Machine Learning","recommendation systems","Scala","RAG","Collaborative Filtering","recommendation engine","PCA","Aim","Dimensionality Reduction","Data Science","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","recommendation systems","AWS","R","Go","Scala","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-improve-collaborative-filtering-with-dimensionality-reduction\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10099894,"title":"Top AI Leaders who were Left Out of the TIME AI 100 List (Part-2)","content":"For the first time in this century, TIME magazine released a list dedicated to 100 most influential personalities in AI. Yet, amidst the grand spectacle, some important individuals in AI found themselves absent from this illustrious list. AIM 100: The Most Influential Leaders in AI 2024As AI reshapes our world, AIM introducing the most influential global leaders in AI. In continuation with the Part-1 that we published earlier, here’s a list of all the polymaths who couldn’t make it to TIME. Erik Brynjolfsson Erik Brynjolfsson, currently a professor and a senior fellow at Stanford University, is a visionary scholar and guiding star in digital economics. He has been a driving force in the study of the digital economy, emphasising the profound transformation brought about by technological advancements. While the rise of AI sparks concerns about job displacement, Brynjolfsson offers a more pragmatic perspective. In an NYT piece, Brynjolfsson encourages us to shift our gaze. “The thing that I wish people would do more of is think about what new things could be done now that were never done before. Obviously, that’s a much harder question,” he said. It is also, he added, “where most of the value is”. Rodney Brooks Then there’s Rodney Brooks, a seasoned technologist who knows the difference between real progress and baseless hype as a majority of his predictions have been spot-on. Having co-founded iRobot and contributed significantly to MIT’s computer and AI labs, his expertise in robotics and AI is unparalleled. Brooks, in his annual predictions, reminds us to temper our expectations, believing that the integration of robots into our lives will be a gradual, symbiotic process. In his fifth annual scorecard in 2023, he confessed to having allowed hype to make him too optimistic about some developments. “My current belief is that things will go, overall, even slower than I thought five years ago,” he wrote. Brooks expects “robots that will roam our homes and workplaces … to emerge gradually and symbiotically with our society” even as “a wide range of advanced sensory devices and prosthetics” emerge to enhance and augment our own bodies: “As our machines become more like us, we will become more like them. And I’m an optimist. I believe we will all get along.” Yejin Choi Despite AI breakthroughs in previously human-dominated language and visual art, our gravest concerns should probably be tempered, believes Yejin Choi, a professor of computer science at the University of Washington. The computer scientist, who is also a 2022 recipient of the MacArthur “genius” grant, has been doing groundbreaking research on developing common sense and ethical reasoning in AI. She reminds us that simply instructing AI not to commit certain actions is insufficient; AI must also possess the wisdom to make sensible decisions and consider the broader implications of its actions. In an interview with the NYT earlier this year, she elaborated how some people naïvely think that if we teach AI “don’t kill people while maximising paper-clip production”, that will take care of it. But the machine might then kill all the plants. It’s common sense not to go with extreme, degenerative solutions, she explained. Jeff Dean American computer scientist and engineer Jeff Dean has long headed the AI department at Google Brain, encouraging researchers to publish academic papers actively. Impressively, they officially pushed out nearly 500 studies since 2019, according to Google Research’s website. On the one hand, there are concerns around AI development and its associated risks. And on the other, this is a natural progress in technology: Innovation happens quickly. It’s not an either\/or; it’s a both\/and. To Dean’s point, society can mitigate risk and be bold. Time and again, Dean has reminded us that the rapid development of AI is both exhilarating and worrisome, emphasising the need to balance innovation and risk mitigation. Sergey Levine Levine is the associate professor of electrical engineering and computer sciences and the leader of the Robotic AI & Learning (RAIL) Lab at UC Berkeley. An advocate of reinforcement learning who also holds an appointment with the Robotics at Google program, along with fellow researchers Julian Ibarz, Jie Tan, Chelsea Finn, Mrinal Kalakrishnan, and Peter Pastor, recently published a review titled How to Train Your Robot with Deep Reinforcement Learning — Lessons We’ve Learned. In the latest, second of four Distinguished Lectures on the Status and Future of AI that he has delivered, he extensively spoke about examining algorithmic advances that can help ML systems retain both discernment and flexibility. He emphasised the relationship between data and optimisation in problem-solving. Without adequate data, researchers are unable to address challenges innovatively. Conversely, optimisation strategies struggle to find real-world applications without the necessary data. By combining both the elements effectively, we can inch closer to creating a space-exploring robot capable of devising solutions to unexpected problems, Levine believes. Pieter Abbeel Peter Abbeel has had a long and upward career in robotics from learning to significantly improve robot manipulation to receiving the 2021 ACM Prize in Computing for pioneering work in robot learning. Abbeel has journeyed from teaching robots to learn from humans to pioneering learning-through-trial-and-error techniques. His groundbreaking work forms the bedrock of the next generation of robotics, showcasing the potential of AI to evolve and adapt.He is currently a professor of electrical engineering and computer sciences, director of the Berkeley Robot Learning Lab, and co-director of the Berkeley AI Research Lab at the University of California.","excerpt":"Amidst the grand TIME 100, some important individuals in AI found themselves missing from this illustrious list. Let’s find out who","categories":["AI Trends"],"tags":["AI leaders","Top Trend"],"author_name":"Tasmia Ansari","publish_date":"2023-09-12T18:12:51","publication_year":"2023","word_count":900,"keywords":["Top Trend","Go","API","AI","ML","Git","RAG","Aim","CLIP","Julia","AI leaders","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Julia","Git","API","CLIP"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ai-leaders-missed-in-times-100-ai-2023-list\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10080286,"title":"Developers’ Favourite AI Code Generator Kite Shuts Down","content":"Last week, Kite, an AI assisting tool that helps developers write code, announced that it would no longer be operational. Kite was founded by Adam Smith in 2014. The platform augments the coding environment with the internet’s programming knowledge and machine learning – similar to GitHub Copilot, which uses OpenAI CodeX, a version of GPT-3 language model. Other similar platforms include Tabnine and Amazon Code Whisperer. Last year, the company was acquired by financial technology company Affirm for $24.8 million. Even before Copilot achieved mainstream success, this platform was a developers’ favourite. Kite could be integrated with VS Code, IntelliJ, PyCharm, Atom, Spyder, and several others enabling superior autocomplete for various languages. In an official statement, Smith said that Kite failed to keep up on two fronts— being the first mover in using AI for coding, they fell short in being tech ready; second, he admitted that it was difficult to build a business around the product they built. Falling short on the technology front means that the company was not able to project the 10x improvement required to break through the market, stating the state-of-the-art for ML on code is not up to the mark. This is because the SOTA models cannot comprehend the structure of the code, such as non-local context. This, Smith adds, also is a reason why GitHub Copilot, while showing immense promise, has still a long way to go. While Kite had been working on filling the gap by focusing on building better models for code, the problem, Smith stressed, is very engineering intensive. The cost of making these models would go over $100 million, so that a production-quality tool that is capable of synthesising code reliably can be built. On the other hand, in discussing their failure at the business front, Smith said that their diagnosis showed that Kite could also not bring in enough revenue as the 500K developers, and engineering managers using the product preferred to not pay for it. The team at Kite considered pivoting their business towards code searching by leveraging their AI technology and bottoms-up development strategy, however, that too, led to fatigue, and eventually, the calling-off of all operations of Kite. Most of Kite’s code, which includes their data-driven Python type inference engine, Python public-package analyzer, desktop software, editor integrations, Github crawler and analyzer, and much more is open sourced on GitHub, and can be accessed here.","excerpt":"Even before GitHub Copilot achieved mainstream success, the platform’s code assisting feature was a developers’ favourite","categories":["AI News"],"tags":["AI Tool","Github Copilot"],"author_name":"Ayush Jain","publish_date":"2022-11-21T15:52:25","publication_year":"2022","word_count":399,"keywords":["Go","machine learning","OpenAI","AI","ML","Github Copilot","Git","RAG","Python","AI Tool","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","ML","OpenAI","RAG","Python","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/developers-favourite-ai-code-generator-kite-shuts-down\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10140977,"title":"Anthropic Will Use Chain of Thought Reasoning to Improve Prompts","content":"Anthropic has released yet another feature on Anthropic Console. The latest addition lets you improve your prompts for higher-quality outputs. “The prompt improver allows developers to take existing prompts and leverage Claude to automatically refine them using advanced prompt engineering techniques,” said Anthropic in the announcement. More importantly, this marks Anthropic’s foray into the world of reasoning. Anthropic has mentioned that the prompt improver uses chain-of-thought reasoning to detect problems and refine the prompt. The prompt improver will include steps to break it down systematically, and ‘think’ before responding. In addition, the tool will check for grammatical errors and prefill with any necessary information to improve the accuracy of the output. Anthropic also revealed a considerable improvement in the output’s accuracy. They said, “Our testing shows that the prompt improver increased accuracy by 30% for a multilabel classification test and brought word count adherence up to 100% for a summarization task.” The test involved mapping a randomly picked sentence to a parent article, within 500 samples picked from Wikipedia. Another test involved assessing how accurately Claude could adhere to the word limit when summarising ten articles on Wikipedia. In the latter test, Claude scored a full 100% accuracy. Anthropic is also allowing developers to add input-output examples, which are then transformed into a ‘standardised’ XML format to help the model process it with the best clarity.In case a developer can’t craft examples, Claude will also generate synthetic ones to emulate them. “Claude can automatically create synthetic example inputs and draft outputs for you to streamline this process,” said Anthropic. Furthermore, Anthropic has also introduced a ‘prompt evaluator’ that allows developers to benchmark and grade prompts on a five-point scale. Anthropic is also enabling developers to provide feedback, and further improve the results. Interestingly, Anthropic has already tested this feature with one of their customers, Kapa.ai. “Anthropic’s prompt improver streamlined our migration to Claude 3.5 Sonnet and enabled us to get to production faster,” said Finn Bauer, Co-Founder at Kapa.ai in the announcement from Anthropic. A few days ago, Dario Amodei, CEO at Anthropic revealed that Claude 3.5 Opus is on the cards. We’re curious if today’s announcement is a hint towards integrating reasoning capabilities in the flagship Claude model.","excerpt":"The new feature lets developers fine-tune prompts, to increase output accuracy.","categories":["AI News"],"tags":["Anthropic","Claude","prompt engineering","reasoning"],"author_name":"Supreeth Koundinya","publish_date":"2024-11-15T09:24:24","publication_year":"2024","word_count":369,"keywords":["Anthropic","Go","TPU","AI","Claude","ML","RAG","Ray","prompt engineering","reasoning","Claude 3.5","R"],"extracted_tech_keywords":["AI","ML","Claude 3.5","Anthropic","Ray","RAG","prompt engineering","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/anthropic-will-use-chain-of-thought-reasoning-to-improve-prompts\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":24048,"title":"Top 5 Legal AI Startups That Have Changed The Face Of Indian Legal Sector","content":"Given how India’s public sector is showing a growing interest in artificial intelligence, can legal tech startups keep up and help transform India’s judiciary system? Though India has made rapid progress in terms of technology, companies and researchers are yet to utilise the full potential of AI. In fact, a PwC report emphasises that how instead of waiting for technology to reach a level where regulatory intervention becomes necessary, India could be a frontrunner by establishing a legal infrastructure in advance. A slew of Indian legal tech startups are building NLP-based applications and introducing next-gen legal research platforms that help law firms go beyond simple, keyword-based research, thereby making it less time-consuming. Many legal startups are fast rising in AI research capabilities, some of who have their own AI research labs. AIM Lists Top Indian Legal Tech Startups In India That Are Disrupting Legal Services In India SpotDraft: This AI-based contract management platform is helping businesses make their contracts “come to life”. The company is trying to ease the pain of managing paperwork by offering revolutionary tools powered by AI. Through SpotDraft’s intuitive AI-powered platform, the customers can draft and sign contracts, send automated reminders and receive payments. The Gurugram-based startup is kickstarted by Shashank Bijapur, a Harvard Law School alumni and Wall Street Lawyer, and Madhav Bhagat, an alumnus of Carnegie Mellon, an ex-Googler specialising in various AI applications. The AI platform can analyse legal documents to give the users the “good, bad and ugly” components in the contracts, so that the users know what clauses to negotiate on. CaseMine: The NCR-based legal research and analysis platform CaseMine, founded by Aniruddha Yadav, has pioneered the next-gen legal research platforms in India. The startup leverages AI capabilities to unearth latent linkages between case laws, thereby making research more in-depth and comprehensive. The proprietary tools enable researchers to look beyond a mere keyword-based search — an old technology with severe limitations. In law, the extent of search is constrained by the extent of keywords known on a given legal proposition. The startup’s CaseIQ software — a virtual legal research assistant automatically analyzes the language of the brief. It then feeds this information into a complex predictive algorithm that leverages the latest data science technology, to highlight potential missing points of law, or alternative arguments. CaseIQ offers suggestions in the form of keywords, acts or landmark case laws to make your research more in-depth and comprehensive. NearLaw: Mumbai-based NearLaw founded last year by Vikas Sahita offers AI-based solutions for lawyers, law firms and companies to search for cases. According to the founder, NearLaw’s deep tech capabilities are said to be the key differentiators that set it apart from the competition. Speaking to a portal, Sahita shared that NearLaw uses NLP technology to understand the relevance of CaseRanking. The startup uses Python heavily and their tech stack is built on Ruby on Rails. The tech team has developed a proprietary model for legal documents, judgments as well as acts and statutes. Pensieve: This Mumbai-based startup founded by serial entrepreneur Gaurav Shrivastava, (co-founder of Zimmber bought by Quikr in 2017) and Prahlad K Routh, a materials scientist and an adjunct professor at Columbia University has developed a proprietary AI-driven legal research platform called Mitra that is being used by over 300 customers already. Currently, the Mumbai-based startup is working with IIT-Madras to mature the product and its AI solution is being used by major law firms like Argus and K Law. Founded in 2016, Pensieve is using AI and NLP to improve efficiency of law firms and also made it into the Axilor Accelerator Program, one of India’s largest accelerator program. Talking about making it into the sixth accelerator batch, Shrivastava said, “Axilor has helped has immensely in understanding the business priority. We can expand the product range from a domain specific to a domain agnostic catering to technology solutions for problems in areas such as linguistic and text analytics”. Practice League: Pune-based firm Practice League, is an advanced legal tech company which claims to have transformed the workflow model of over 8,000 lawyers. The company’s SaaS-based products automate legal practice and are deployed in most law firms and legal departments across the globe. News reports indicate that Practice League is currently working with big tech companies like Google and Amazon to weave AI capabilities into its solution.","excerpt":"Given how India’s public sector is showing a growing interest in artificial intelligence, can legal tech startups keep up and help transform India’s judiciary system? Though India has made rapid progress in terms of technology, companies and researchers are yet to utilise the full potential of AI. In fact, a PwC report emphasises that how […]","categories":["AI Trends"],"tags":["AI in legal sector in India","legal AI"],"author_name":"Richa Bhatia","publish_date":"2018-04-26T11:30:11","publication_year":"2018","word_count":724,"keywords":["data science","artificial intelligence","AWS","AI","AI in legal sector in India","RAG","NLP","Python","Aim","analytics","R","legal AI"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","data science","analytics","Aim","RAG","AWS","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-legal-ai-startups-that-have-changed-the-face-of-indian-legal-sector\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10008966,"title":"Google Teases Large Scale Reinforcement Learning Infrastructure","content":"“The new infrastructure reduces the training time from eight hours down to merely one hour compared to a strong baseline.” The current state-of-the-art reinforcement learning techniques require many iterations over many samples from the environment to learn a target task. For instance, the game Dota 2 learns from batches of 2 million frames every 2 seconds. The infrastructure that handles RL at this scale should be not only good at collecting a large number of samples, but also be able to quickly iterate over these extensive amounts of samples during training. To be efficient requires to overcome a few common challenges: Should service a large number of read requests from actors to a learner for model retrieval as the number of actors increases.The processor performance is often restricted by the efficiency of the input pipeline in feeding the training data to the compute cores. As the number of computing cores increases, the performance of the input pipeline becomes even more critical for the overall training runtime. So, Google has now introduced Menger, a massive large-scale distributed reinforcement learning infrastructure with localised inference. This can also scale up to several thousand actors across multiple processing clusters reducing the overall training time in the task of chip placement. Chip placement or chip floor design is time-consuming and manual. Earlier this year, Google demonstrated how the problem of chip placement could be solved through the lens of deep reinforcement learning and bring down the time of designing a chip. With Menger, Google tested the scalability and efficiency through TPU accelerators on-chip placement tasks. How It Works Source: Google AI The above illustration is an overview of a distributed RL system with multiple actors placed in different Borg cells. Google’s Borg system, introduced in 2015, is a cluster manager that runs thousands of jobs, from many thousands of different applications, across tens of thousands of machines. With increasing updates from multiple actors within an environment, the communication between learner and actors is throttled, and this leads to an increase in convergence time. The main responsibility here, wrote the researchers, is maintaining a balance between a large number of requests from actors and the learner job. They also state that adding caching components not only reduces the pressure on the learner to service the read requests but also further distributes the actors across multiple Borg cells. This, in turn, reduces computation overhead. Menger uses Reverb, an open-sourced data storage system designed to implement experience replay in a variety of on-policy\/off-policy algorithms for machine learning applications that provides an efficient and flexible platform. Reverb’s sharding helped balance the load from a large number of actors across multiple servers, instead of throttling a single replay buffer server while minimising the latency for each replay buffer server. However, the researchers also state that using a single Reverb replay buffer service does not cut the job. It doesn’t scale well in a distributed RL setting with multiple actors. It becomes inefficient with multiple actors. The researchers claim that they have successfully used Menger infrastructure to drastically reduce the training time. Key Takeaways Reinforcement learning applications have slowly found themselves in unexpected domains. But, implementing RL techniques is tricky. The performance accuracy trade-off looms large in research. With Menger, the researchers have tried to answer the shortcomings of RL infrastructure. However, its promising results in the intricate task of chip placement has the potential to shorten the chip design cycle and other challenging real-world tasks as well. Reduces the average read latency by a factor of ~4.0x, leading to faster training iterations, especially for on-policy algorithms.Efficient scaling of Menger is due to the sharding capability of Reverb.The training time was reduced from ~8.6 hours down to merely one hour compared to the state-of-the-art. Know more about Menger here.","excerpt":"“The new infrastructure reduces the training time from eight hours down to merely one hour compared to a strong baseline.” The current state-of-the-art reinforcement learning techniques require many iterations over many samples from the environment to learn a target task. For instance, the game Dota 2 learns from batches of 2 million frames every 2 […]","categories":["Deep Tech"],"tags":["chipset","Reinforcement Learning"],"author_name":"Ram Sagar","publish_date":"2020-10-06T15:00:24","publication_year":"2020","word_count":628,"keywords":["Go","machine learning","Reinforcement Learning","TPU","AI","programming_languages:R","Scala","RAG","Aim","programming_languages:Scala","chipset","R"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","TPU","R","Go","Scala","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/google-large-scale-reinforcement-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10102525,"title":"Trained on Chips, GenAI is Reshaping the Chip Industry","content":"Bill Dally, NVIDIA’s chief scientist, recently introduced ‘ChipNeMo‘, a custom LLM developed by their engineers, at the International Conference on Computer-Aided Design, an event for electronic design automation (EDA). ChipNeMo is trained on the company’s internal data to generate and optimise software in designing semiconductors by customising large language models with the help of NVIDIA NeMo—a cloud-native framework for developers to create and deploy generative AI models with billions of parameters. “This effort marks an important first step in applying LLMs to the complex work of designing semiconductors,” said Dally. “It shows how even highly specialised fields can use their internal data to train useful generative AI models.” After evaluating possible use cases of ChipNeMo, the research team at NVIDIA decided to start with three: a chatbot, a code generator, and an analysis tool. Of these, the analysis tool — automating the time-consuming tasks of maintaining updated bug descriptions — has garnered the most positive feedback. The prototype chatbot responds to questions regarding GPU architecture and design, helping engineers quickly locate technical documents. The code generator, which currently produces snippets of 10-20 lines of software in specialised chip design languages, will be integrated with existing tools, providing a valuable assistant for ongoing designs. The research paper by NVIDIA explains how the team gathered design data and employed it to craft a specialised generative AI model, a process that can be adapted to any industry. They started with a foundational model and customised it using NVIDIA NeMo, a framework included in the NVIDIA AI Enterprise software platform for building, customising, and deploying generative AI models. The chosen NeMo model boasts 43 billion parameters and was trained on over a trillion tokens. The model was refined through two training rounds, with the first using approximately 24 billion tokens of internal design data and the second incorporating about 130,000 conversations and design examples. This research represents just one of several instances where generative AI is making its mark in the semiconductor industry. Sharing their valuable experience, NVIDIA Research director and the paper’s lead author Mark Ren underscored the importance of customisation in LLMs. Custom ChipNeMo models, with as few as 13 billion parameters, outperformed even much larger general-purpose LLMs in certain chip-design tasks. However, Ren emphasises the importance of careful data collection and cleaning, as well as staying updated with the latest tools to streamline the work. The increasing complexity of chip design, driven by the relentless march toward smaller transistors is also straining engineering resources, as the industry faces a daunting 4x increase in workload while grappling with a talent gap of 10%-20%. According to Synopsys, the average number of transistors per chip has increased by a staggering 1,000 times since 2000. Optimising Cost Rising costs have been another prominent issue in semiconductor manufacturing. The industry has witnessed a significant increase in the cost of designing and producing semiconductor chips. Factors such as growing design complexity, shrinking feature sizes, a surging number of masks, higher equipment costs, and stringent quality requirements have collectively contributed to this surge. McKinsey reports that the average cost per transistor has soared by 50% since 2013. Generative AI can be a game-changer in semiconductor manufacturing by applying its capabilities to every facet such as design, fabrication, testing, and packaging. “I believe, over time large language models will help all the processes across the board,” said Ren. It can optimise chip design through reinforcement learning, specifically in component placement, known as floorplanning. For instance, Google’s floorplanning algorithm uses deep reinforcement learning to achieve remarkable improvements in power consumption, wire length, and congestion, reducing product development life cycle time significantly. In the competitive landscape of semiconductor manufacturing, achieving shorter time-to-market cycles is paramount. However, as chip design becomes more intricate, the time available for development and delivery significantly decreases. McKinsey indicates that the average time to market for chips has shrunk by 25% because of outstanding demand. Companies like Synopsys, an American EDA company headquartered in California, offer a full-stack AI-driven EDA suite that has made contributions to improving efficiency and reducing development cycles for major semiconductor companies like Samsung Electronics Co Ltd and ST Microelectronics, bringing down costs in turn. Shankar Krishnamoorthy, GM of Synopsys’ EDA Group illustrated how AI, particularly AI-driven EDA tools, can optimise the design of low-power chips. He cited figures from Synopsys.ai EDA solutions, demonstrating energy and power savings of up to 15%. Synopsys is in direct competition with Cadence Design Systems, which is making big moves to add AI to chip design software. However, experts believe that the former is lapping the latter in competition. Additional Opportunities in Chip Manufacturing Pipeline In addition to overcoming these challenges, generative AI presents a range of opportunities for the semiconductor manufacturing industry. Generative AI can also play a pivotal role in improving defect detection. The quality of chips depends on a multitude of factors, including material properties, process parameters, environmental conditions, and human error. However, as feature sizes shrink below 10 nanometers, defect rates increase exponentially due to factors such as quantum effects, variability, and noise. McKinsey highlights that the average defect density has surged by 10 times since 2013. By utilising unsupervised learning, it can detect defects in semiconductor chips without requiring labelled data or prior knowledge. This technology can enhance defect detection accuracy by up to 30%, surpassing traditional methods. GenAI can also lead to the discovery of new materials or the optimisation of existing ones for better performance and reduced costs. It can also drive the creation of new products or improvements in existing ones, significantly reducing product development cycles. Moreover, generative AI can broaden the market potential of semiconductor devices by creating custom ICs for specific tasks, offering superior performance, lower power consumption, and reduced costs compared to general-purpose ICs. In conclusion, while the semiconductor manufacturing industry faces substantial challenges, the integration of generative AI promises to be a transformative force. It offers innovative solutions that can optimise chip design, enhance defect detection, and unlock new horizons of efficiency, quality, and innovation.","excerpt":"GenAI capabilities can be applied to every facet of semiconductor manufacturing such as design, fabrication, testing, and packaging.","categories":["AI Highlights"],"tags":["AI Chips","AI Tool","Generative AI","NVDIA"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-11-06T11:25:43","publication_year":"2023","word_count":1000,"keywords":["Go","NVDIA","GenAI","AI","RPA","ML","innovation","RAG","AI Chips","automation","generative AI","Generative AI","AI Tool","R"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","RAG","R","Go","automation","RPA","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/trained-on-chips-gen-ai-is-shaping-the-chip-industry\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10087532,"title":"Unleash the Power of Rust with these Frameworks","content":"Released in 2010 by Mozilla as a project to develop a safer, faster, and more concurrent alternative to C and C++, Rust has gained widespread popularity among developers for its ability to offer the speed and low-level control of these languages while also providing modern language features and safety guarantees. After C and Assembly, Rust was added to the Linux kernel in 2022 as the third programming language. Since Rust has many use cases and developers recognise its potential, it is a valuable language to learn. Let’s see what the different web frameworks built on Rust are. Rocket To begin with, Rocket is a user-friendly and customisable web framework for Rust that focuses on speed and is suitable for both experienced and novice developers. It generates code automatically, removing the need for unnecessary code writing. Additionally, it makes sure that all the types in a given request are validated and can be processed seamlessly. It also streamlines handling new requests by automatically converting them into HTTP responses. Rocket is an asynchronous web framework emphasising security, performance, flexibility, and usability. Other features include async streams, an easy-to-use testing library, and extensibility, and it is also entirely type-safe. Actix Web Actix Web and Rocket are the most preferred frameworks of developers. It offers type safety and fast performance and is stable as it runs on the most recent, stable Rust build, while Rocket works on Rust Nightly. Actix is a great option for secure, efficient, and simple web development. Actix Web includes features like SSL support, WebSockets, and Tokio compatibility and is compatible with HTTP\/1.x and HTTP\/2. Axum Adding to the list is Axum, a lightweight and asynchronous web application framework that milks Rust’s memory safety and performance capabilities to build scalable and reliable web services. Key features include asynchronous request handling, middleware, flexible routing and type-safe handlers. Axum is also compatible with Rust libraries such as serde, SQLx, and tracing. It is ideal for building high-performance web applications, microservices, APIs, and real-time applications. Seed Seed is a frontend framework with an Elm-like architecture for generating dependable and performance-driven web apps. It requires minimum boilerplate and configuration. Tauri Tauri is another framework to offer fast and compact binaries for major desktop operating systems. Engineers can use any frontend framework that compiles to HTML, JavaScript, and CSS to build the user interface. At the same time, the backend is a Rust-generated binary with an API for front-end communication. Tauri offers an array of features, including app storage, tray, plugin system, bundler, GitHub action, native notifications, scoped filesystem, self-updater, and sidecar. Yew Programmers who want to create web apps with WebAssembly and a familiar framework design prefer Yew as it offers component-based development, performance-boosting multithreading, and full JavaScript interoperability. Yew is a powerful framework that can make internal tools with Rust’s efficiency and memory safety. It includes a macro for defining interactive HTML, minimises DOM API calls, and allows offloading of processing to background web workers. Yew is compatible with NPM packages and works with all major modern browsers, using its own virtual-DOM representation. Dioxus Finally, we have Dioxus, which is used in creating cross-platform user interfaces. It supports creating web, mobile, desktop, TUI, and liveview apps with a component-based architecture, react-like design, and features such as props, state management, and an error handler. Dioxus is renderer agnostic and comprises inline documentation, memory efficiency, and a multi-channel asynchronous scheduler.","excerpt":"Rust has different frameworks, and developers recognise its potential, making it a valuable language to learn.","categories":["AI Features"],"tags":[],"author_name":"Shritama Saha","publish_date":"2023-02-16T17:25:36","publication_year":"2023","word_count":565,"keywords":["Rust","AI","ML","microservices","RAG","Ray","SQL","JavaScript","R","Java"],"extracted_tech_keywords":["AI","ML","Ray","RAG","microservices","R","SQL","JavaScript","Rust","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/unleash-the-power-of-rust-with-these-frameworks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":54784,"title":"Top 7 FAQs About Machine Learning In India: Answered","content":"Machine learning provides one of the best working opportunities out there and hence it the one the top hottest career choices not only India but around the world. Tech giants like Google, Amazon and others are always looking for skilled personnel when it comes to hiring for machine learning. Machine learning provides a great challenge and a good salary, not only that but a massive market for job opportunities and indeed, people are interested. So, naturally, people would like to know more about ML. Below are a few of the FAQs that we have tried to answer in some detail: How do I learn machine learning on my own? This question not only ends up in the FAQs on the web, but this is also the first question that pops up in one’s mind when they think about learning machine learning. How? Well, it depends on where one stands right now on their skill level. Generally, it will be better to have specific knowledge about the programming languages, but fret not! Even if one doesn’t have any experience with programming, one can learn machine learning after learning those languages of course. The languages like python are relatively easy to learn, and most of the students, from personal experience, have said that studying python without prior knowledge of programming will make it a bit easy to learn. If one intends to learn ML on their own, they need to be relentless and have to start thinking about them round the clock. Reading books, articles, and blogs, and keeping books related to ML always by one’s side to get the basics right, is also essential. If one has decided to start learning on their own, they would have to work very hard and take up every possible course online. As to build up one’s skill set and an impressive resume, one should do as many Kaggle competitions as possible. What skills are needed for machine learning jobs? A machine learning engineer must have a particular skill set to work and excel as a machine learning engineer. The following skills are the must-haves for a machine learning engineer: Expertise in languages like Python\/C++\/R\/Java.Probability and statistics.Data modelling and Evaluation.Machine learning Algorithms.Advanced Signal Processing Techniques.Distributed Computing. Some prerequisites: Linear Algebra.Programming knowledge.Calculus. What are some algorithms that every machine learning engineer should know? Machine learning algorithms are self-modifying and continue to improve. These machine learning algorithms in future will take up a majority of the automation sector jobs. These algorithms help in tackling the complex real-world problem with minimal human interventions, and there are some algorithms that every machine learning engineer should know. Below are some of the most famous and top algorithms an ML engineer should know: Naïve Bayes Classifier AlgorithmK Means Clustering AlgorithmSupport Vector Machine AlgorithmApriori AlgorithmLinear RegressionLogistic RegressionArtificial Neural Networks How should you start a career in machine learning? Kickstarting your career in AI and machine learning follows almost the same path as every other stream, get one’s basics right, thone skills in programming languages and other prerequisites. Basics to learn: Learn calculusLearn linear algebraLearn programming languages like Python, R, Java Next step would be to learn machine learning. Take up online machine learning courses and take up as many as you can. Practising machine learning is crucial as Andrew Ng says, ‘you will have a choice between staying at home and reading research papers\/implementing algorithms, vs watching TV. If you spend all Saturday working, there probably won’t be any short-term reward, and your boss won’t even know or say ‘nice work’.’ One can take many courses on platforms like Coursera and other MOOCs. Next step would be to build projects and start participating in Kaggle competitions after one has honed their skills enough to be confident about machine learning. Which are the best online courses for machine learning? Courses help to get a better understanding of machine learning and taking one’s knowledge to the next level. There are several courses online on Coursera and other MOOCs. Tech giants like Google and Microsoft also offer AI and ML courses which will help take one’s career one step further. Below are listed some of the popular courses: Machine Learning Course by Stanford University (Coursera)Deep Learning Course (deeplearning.ai)Machine Learning: From Data to Decisions (MIT Professional Education)Machine Learning Course A-Z™: Hands-On Python & R In Data Science (Udemy) What are some common machine learning interview questions? The interview questions asked in these machine learning interview generally aim at finding out the comprehensive picture of the candidate’s depth of knowledge. Usually, the interviewer would like to go deeper as it gives far more understanding of the candidate’s knowledge. Below are some questions that a candidate must be able to answer, which will help one become more confident while facing an interviewer: (Many other questions can be asked; hence, it is vital to have sound knowledge about all the ML concepts) Discuss views on the relationship between ML and statistics.Why do we call it GLM when it’s non-linear?How are neural nets related to Fourier transforms? And what are Fourier transforms?Explain and walk us through the math and implementation of an algorithm you like? (next, the interviewer might pick an advanced one and ask about it)What’s the relationship between Principal Component Analysis (PCA) and Linear & Quadratic Discriminant Analysis (LDA & QDA)What’s the difference between logistic and linear regression? How do you avoid local minima? What are the AI and ML starting salaries in India? MNCs with R&D centres in India offer large salaries ranging from ₹ 1.5 million – 2 million per annum. However, to attract these kinds of numbers, one must have significant experience and knowledge. The salaries as per Glassdoors are: For beginners, the salary number will be in the range of ₹4-5 lakhs.A machine learning engineer will have around range ₹750k – ₹1,200k.An ML data associates ₹231K – ₹354K These salaries vary from one MNC to another MNCand could be more.","excerpt":"Machine learning provides one of the best working opportunities out there and hence it the one the top hottest career choices not only India but around the world. Tech giants like Google, Amazon and others are always looking for skilled personnel when it comes to hiring for machine learning. Machine learning provides a great challenge […]","categories":["AI Trends"],"tags":["fourier transform machine learning","Machine Learning India","MOOCs","naive bayes"],"author_name":"Sameer Balaganur","publish_date":"2020-01-28T17:04:08","publication_year":"2020","word_count":986,"keywords":["data science","machine learning","AI","fourier transform machine learning","neural network","ML","MOOCs","Machine Learning India","distributed computing","naive bayes","Python","Aim","deep learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","data science","Aim","distributed computing","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-faqs-about-machine-learning-in-india-answered\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10074520,"title":"Polygon Aims to be the AWS of Web3","content":"Meta recently announced that it would allow its users to post NFTs across both Instagram and Facebook. A few months back, a fintech company, Stripe, announced that it would make payouts to its creator community. A year ago, global professional services and technology firm, EY, started offering its flagship blockchain products. The American professional football league NFL started selling NFT collections last year. At the core of everything these companies are developing – Polygon remains ubiquitous. Polygon is one of the most widely used platforms for Ethereum ‘Level 2’ scaling and infrastructure development, besides Solana, Polkadot, Fantom, LoopRing, Arbitrum (Offchain Labs), Immutable X, Tezos, and Flow blockchains. It is used to build blockchain games, mint NFTs and more. It is revolutionising how companies look at DeFi (decentralised finance), NFTs (non-fungible tokens), Web3 and blockchain. “We partner with companies that make life easier for developers. We are the last man standing, but we want to make sure each and every person in that value chain is successful along with us. The vision of us being the AWS of Web3 is not far-fetched. We are already on that journey. We will probably get there sooner than most of our competitors,” said Dalip Tyagi confidently. He currently heads Polygon as the SVP and head of developer relations. Tyagi told Analytics India Magazine that he joined Polygon eight months ago. Prior to this, he worked in senior leadership roles at Amazon, Meta (previously Facebook), and Microsoft, among others, leading global marketing teams and developer initiatives. “It was the right time for me to jump into the industry, considering it is pretty nascent and highly fragmented, and best is yet to come,” said Tyagi when asked about his career transition. Further, he lauded the work culture and the value the company is adding to the entire ecosystem. Tyagi said that the energy, excitement, humility, and ability to make mistakes while learning keeps the team motivated at all times.“But, every single day coming out and making new and continuous improvements in the processes, how do we do it better than the previous day? That is what keeps us going. That is what makes us a unique company,” said Tyagi, sharing his personal experience at Polygon. Tyagi said that Polygon looks to invest in technologies that are future leaning. “We invest in them ahead of the curve,” he added, saying that at the core, they are looking at investing in more secure, faster and scalable solutions and products. He also said that they are emotionally connected with developers and will continue to invest in the future. “If you look at the past few months. It has been a golden time for us. However, the best is yet to come,” said Tyagi. “Our investment in India will continue to grow,” he added, saying that its Web3 hackathons (BUIDL IT) conducted by Polygon saw about 300 per cent year-on-year (YoY) growth, and most of it is from India. “We are on track; our hockey stick is going to bend the other way around,” said Tyagi. Besides India, Polygon sees larger adoption in Africa, Latin America, Europe and the US. India’s future relies on Web3 “India is the biggest exporter of technology, and technology services, not necessarily the consumer market. It is growing too. But, a substantial portion is still based on how we export the technology outside of India and how we have a skilled workforce that can build. That is why India has become a country with outsized impact,” said Tyagi. Citing prominent names in Web2 like Microsoft, Google and others, Tyagi said that some big names have set up their research centres in Gurugram, Bengaluru and Hyderabad. “Each and every single web company has a presence in India. I expect the same will happen for Web3,” he added, pointing at the trends of how Indian enterprises are looking at Web3 as a future. Stressing on some of the drawbacks of Web2 in terms of UI\/UX, efficiency, transferring data, safety and security, and more, Tyagi said today, large data brokers in the middle are controlling the identity and everything; with Web3, you are giving control back to the users. He believes that all of these things are coming together. “You will see this merger that happens globally,” said Tyagi. He further told AIM that with the sheer size of the talent pool, alongside the quality impact on Web2 and all the educational institutions that are continuing to churn out high-quality engineering graduates, India will continue to have that outsize impact in the global market, as well as in the home market. Polygon At a Glance Bengaluru- and Singapore-based Polygon (previously ‘Matic Network’) is one of the few blockchain startups to have made it big in the global arena. The company was co-founded by Anurag Arjun, Jaynti Kanani, and Sandeep Nailwal in 2017. Later, Serbia-based software engineer Mihailo Bjelic joined the firm as another co-founder in 2020, alongside rebranding the company from Matic Network to Polygon. Co-founders of Polygon Over the years, the company has invested in multiple blockchain startups, including thirdweb, Insomnia Labs, Unstoppable Domains, Mighty Bear Games, Optic, Gnosis, Revoland, Sidus Heroes, Thred Apps, and Mech, among others. The company has invested in over 34 blockchain and NFT startups and successfully exited from one company (Health Hero). In 2021, Polygon acquired two zero-knowledge (ZK-based) companies, namely Mir Protocol ($400 million) and Hermez Network ($250 million). These investments came after the company, last year, had announced its strategy of investing in ZK-focused initiatives and projects, where it had allocated a $1 billion strategic fund. In 2022, the company raised $450 million in a funding round led by Sequoia Capital India. Besides Sequoia, the company is backed by multiple investors and VC firms globally. Some prominent names include Mark Cuban, Tiger Global Management, Coinbase Venture, and others. Most recently, Polygon joined Disney’s 2022 Accelerator Programme to develop AR, NFT and AI experiences. It was one of the six startups to make it to the programme, besides Flick Play, Inworld, Lockerverse, Obsess and Red 6. Polygon told AIM that it currently facilitates more than 37,000 dApps across the globe. The company has recorded more than 160+ million unique users and over 3+ million daily transactions, and since its inception, it has recorded close to 3.4 billion transactions and counting. The company has over 450+ people and a fully-remote team. It has grown significantly in the last six months. The company told AIM that it is looking to hire 100+ people across geographies across all the business units in the coming months, particularly in business development, marketing, engineering roles, and more. “There is no way for us to slow down. We will continue to see growth both in bringing new products and launching them (particularly in the areas of DeFi, NFT, and Web3), supporting them, and going after all the big opportunities we see in the marketplace,” they added. Polygon claimed that it is carbon neutral and is going carbon-negative in 2022. What’s new? Polygon is betting big on zero-knowledge (ZK) technology. In July 2022, the company launched Polygon zkEVM, which it describes as a ‘major leap forward.’ The company believes it is the first Ethereum-equivalent scaling solution that works seamlessly with all existing smart contracts, developer tools and wallets, using advanced cryptography called ‘ZK proofs.’ The team said, up until the Polygon zkEVM launch this year, many of the greatest minds in Web3 believed that this breakthrough would take up to ten years to realise. It is the first ZK scaling solution that is EVM (Ethereum Virtual Machine) equivalent: all existing smart contracts, developer tools, and wallets work seamlessly. Tyagi is super excited about zero-knowledge technology as well. He said, “It is really fascinating. The things that are possible say, you can be on a blockchain—which is by definition public—but you can also be anonymous at the same time. “That is truly empowering to own how the data is created, distributed, managed, and monetised at the end of the day. It is truly about making a creator-led economy, and zero-knowledge has the potential to change,” he added, saying he is always up for an open, decentralised and equitable web. Last month, Etherium announced that it would be switching to proof-of-stake (PoS), slated for mid-September. Many wonder how this merge would impact ‘Layer 2’ platforms like Polygon. “We are kind of okay, more than kind of—we are extremely, unbelievably, falling-out-of-our seats excited by Ethereum’s upcoming transition to PoS,” said the team. Proof-of-stake (PoS) is a type of consensus mechanism used by blockchain to achieve distributed consensus. In proof-of-work (PoW), miners prove they have capital at risk by expanding energy. In PoS, validators explicitly stake capital in the form of ether into a smart contract on Ethereum. Polygon said that Ethereum would be more environmentally friendly. But, it will not lower its gas fees or increase its speed. “The network depends on Polygon and other Layer 2 solutions to solve this,” the team added.","excerpt":"Polygon is betting big on zero-knowledge (ZK) technology. In July 2022, the company launched Polygon zkEVM, which it describes as a ‘major leap forward.’","categories":["Global Tech"],"tags":["Blockchain","polygon"],"author_name":"Amit Naik","publish_date":"2022-09-07T13:08:16","publication_year":"2022","word_count":1492,"keywords":["Go","API","Blockchain","AWS","AI","ML","Scala","RAG","polygon","Aim","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","AWS","R","Go","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/polygon-aims-to-be-the-aws-of-web-3-0\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10130400,"title":"​​OpenAI&#8217;s SearchGPT Could Blow Up Google&#8217;s $2 Tn Monopoly","content":"Google consistently builds cool AI products, while OpenAI seems to wait for these launches to steal the spotlight. “Be OpenAI, wait for Google to drop a cute math model, then launch a competing search engine that could potentially blow up Google’s $2T internet search monopoly and send Google execs into existential dread,” quipped Aidan McLau, the chief executive of Topology Invest. A few hours after Google launched its new model capable of solving International Math Olympiad problems, OpenAI grabbed attention with its new Google alternative, SearchGPT. It combines OpenAI’s AI models with real-time web information to provide fast and relevant answers to user queries. Currently in the prototype phase, it is available to a limited group of 10,000 test users. >be google>build cool ai!>ai does well on math.>yay!>be openai>wait for google to drop cute math model>launch fire competing search engine that could potentially blow up google's 2T internet search monopoly and send google execs into existential dread— Aidan McLau (@aidan_mclau) July 25, 2024 “We believe there is room to make search much better than it is today. We are launching a new prototype called SearchGPT. We will learn from this prototype, improve it, and then integrate the technology into ChatGPT to make it real-time and maximally helpful,” said OpenAI chief Sam Altman. This isn’t the first time OpenAI has done something like this. Earlier this year, when Google released Gemini 1.5, OpenAI announced Sora on the same day. Then, just a day before Google I\/O 2024, OpenAI hosted its Spring Update and released GPT-4o. However, OpenAI has yet to make Sora publicly available, and the voice features for ChatGPT are still not accessible. Who knows, OpenAI might have achieved AGI internally. “OpenAI has probably already achieved gold in the Math Olympiad—something even the most optimistic AI researchers would not have expected before 2025,” posted a user on X who goes by the name Chubby. This might actually be true, as Altman responded to Google DeepMind’s IMO score with a simple ‘lol’. openai has the opportunity to do the funniest thing… pic.twitter.com\/RdtkGDTXdC— Aidan McLau (@aidan_mclau) July 25, 2024 Math all the Way Solving problems at the IMO Olympiad is any day a greater achievement than launching an AI-based web search. “AlphaProof is one of the most exciting applications of LLMs combined with RL. The Gemini model automatically translates natural language problem statements into formal statements (i.e., formalizer network),” said Elvis Saravia, the co-founder of DAIR. “LLMs are alien beasts. It is deeply troubling that our frontier models can both achieve a silver medal in the Math Olympiad and fail to answer. “Which number is bigger, 9.11 or 9.9?’” said Jim Fan, the lead of Embodied AI (GEAR Lab) at NVIDIA. He further explained that AlphaProof and AlphaGeometry-2 are trained on formal proofs and domain-specific symbolic engines. “In a way, they are highly specialized towards solving Olympiads, even though they build on a general-purpose LLM base,” he said. Meanwhile, OpenAI is quietly developing Project Strawberry to significantly boost the reasoning capabilities of its AI models. While details about the project remain undisclosed, it focuses on a novel approach that enables AI to plan ahead and autonomously navigate the internet for in-depth research. Internally, OpenAI has tested AI that scored over 90% on the MATH dataset, which benchmarks championship-level math problems. This progress aims to tackle current AI reasoning limitations, such as common sense issues and logical errors that often lead to inaccurate outputs. Previously known as Project Q*, which was leaked last year and could solve new math problems, Project Strawberry is now working to enhance long-horizon tasks (LHT). This involves a specialised “post-training” phase, adapting base models for better performance, similar to Stanford’s 2022 Self-Taught Reasoner (STaR), which enables AI to generate its own training data for improved intelligence. Do we Really Need SearchGPT? While SearchGPT is a nice feature to have in ChatGPT, it appears that Perplexity AI has already taken the lead in this segment. “By the way, Perplexity is awesome. They made me rethink what search & AI integration could be. It’s often the first place I go to now when I need to start researching a new topic,” posted Lex Fridman on X. Click here to skip the waitlist >> https:\/\/t.co\/lKL77G8bil https:\/\/t.co\/0ua0nlWrzA— Hersh Desai (@Hersh_Desai) July 25, 2024 “SearchGPT is like a nice feature at this point. Nobody can take the crown from Perplexity after they introduced agentic search – it’s simply too good and OpenAI cannot just steamroll them,” posted a user on X. Building a tool like Perplexity AI isn’t particularly difficult today, so it’s puzzling why it took OpenAI so long. “This is cool, but the name sounds like something a high schooler would put on their resume as their first solo project,” joked a user on X. On a related note, Bishal Saha, a dropout from Lovely Professional University, created Omniplex, an open-source alternative to Perplexity AI, over a single weekend. https:\/\/twitter.com\/alex_valaitis\/status\/1816540499773444182 It’s unclear how concerned Perplexity AI chief Aravind Srinivas is right now, but Google’s stock did plummet after the OpenAI SearchGPT demo. In response to competitors like Perplexity AI and ChatGPT, the search giant introduced ‘AI Overviews’ at Google I\/O 2024. This feature generates summaries for user queries.","excerpt":"Not really.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-07-26T18:12:32","publication_year":"2024","word_count":867,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","ETL","GPT-4o","GPT","Aim","R"],"extracted_tech_keywords":["AI","GPT-4o","ChatGPT","OpenAI","Aim","TPU","R","Go","ETL","GPT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openais-searchgpt-could-blow-up-googles-2-tn-monopoly\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046130,"title":"Millionaire Street Vendors: Big Data Helps Bust Tax Fraud In Kanpur","content":"This month, the Uttar Pradesh IT department identified over 250 street food vendors of Kanpur to be millionaires. These millionaires include vegetable sellers, small pharmacy shop owners, grocers, rag pickers and sanitation workers- all who have been evading tax for years and manage to hoard Rs 37.5 million. The IT department has been using surveillance software to identify the tax evaders. The software tracked down several dealers who owned at least three cars, KVP worth more than 30 crores and 650 bighas of agricultural land. The IT department found that hundreds of these individuals had not paid any tax beyond the GST, and over 65 stores had not even registered their business on the GST records. Vendors had not applied for an FSSAI certificate that allows them to work legally. Instead, they invested in properties and finance schemes to hide their money trails. The fraud was busted using the PAN card identification number trail. Data analytics for fraud detection In 2019, tax officials detected tax fraud worth Rs. 37,946 crores. The Finance Ministry reported 1,620 cases of fake invoices summing up to a total value of Rs 11,251 crores. AI-enabled systems can assist by scanning through large datasets in a shorter period making use of criteria such as employment status, audit history—among others to avoid such tax frauds. The Indian government has been making an effort to turn to the digital space for assistance in various sectors, including initiatives to reform direct taxes in the country. For example, last year the government announced the ‘Transparent Taxation Platform’. The scheme aims to reward the ‘honest taxpayers’ of the country and make tax compliance easier with reforms like faceless assessment, faceless appeal, and taxpayers’ charter. The platform uses data analytics, artificial intelligence and machine learning (AI\/ML) against tax fraud and tax evaders. The government has created a single window to generate e-invoices, pay e-way bills, and file GST to certify there is no data duplication. Steps have also been taken by the Central Bureau of Direct Taxation and the Central Board for Indirect Taxes and Customs to share data. This move has allowed the earnings tax division to go after Benami properties or actual property properties purchased beneath faux names. These are followed through with data analytics-based investigations. Not just India, European countries too have been using AI\/ML for tax fraud detection for a while now. For instance, Denmark managed to catch 60 of every 100 cases involving tax fraud using big data. Norway and France have also made use of AI tools as preventive measures against tax evasion. Fraud detection softwares Tax fraud analysis uses big data for data mining. The process involves two key groups of data mining tasks; predictive tasks and descriptive tasks. Predictive tasks predict each observation, whereas descriptive tasks describe the data examined. The two most popular fraud detection softwares are: AdvaSmart: AdvaSmart by AdvaRisk is an AI-powered fraud prevention software that highlights credit negative transactions to ensure frauds are detected at the earliest. The software comes with an automated decision support system powered by its proprietary algorithms and 1000+ data sources. In addition, the software leverages 600+ data sources using data analytics to find unidentified patterns associated with defaulters. Feedzai: Feedzai assists banks and institutions globally to fight financial crimes through its ML-powered fraud detection software. The software can detect complex typologies and visualises hidden relationships among transactions. Applying big data to analyse these transactions help mitigate transaction fraud, money laundering and account opening frauds.","excerpt":"Transparent Taxation Platform uses data analytics, artificial intelligence and machine learning (AI\/ML) against tax fraud and tax evaders.","categories":["IT Services"],"tags":["Big Data","Big Data Analytics","big data and analytics everyday life","decision tree algorithm"],"author_name":"Avi Gopani","publish_date":"2021-08-16T17:00:00","publication_year":"2021","word_count":579,"keywords":["Go","artificial intelligence","decision tree algorithm","machine learning","AI","ML","RAG","Big Data Analytics","Aim","analytics","big data and analytics everyday life","Big Data","R","fraud detection"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","fraud detection","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/millionaire-street-vendors-big-data-helps-bust-tax-fraud-in-kanpur\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10053732,"title":"Council Post: Re-imagining Business Value Creation Through AI","content":"In the last 40 years, we have experienced two major technology-led waves of value creation: one driven by the rise of personal computing, the next by the dawn of the internet. A third wave is now being led by artificial intelligence (AI)-enabled digital transformation. Prior to the outbreak of COVID-19, digital transformation was part of a five-year roadmap for many businesses. The pandemic has brought tremendous acceleration, and companies have crashed five-year plans to a few months in many cases! AI technology is already being used extensively in the business world in areas like customer monetisation, marketing, human resources, risk management, robotic manufacturing, supply chain and others. Digital transformation is only accelerating the use of AI, as with digital business models, more data is available, and in the digital world, things can be adapted in real-time. AI technology will disrupt entire industries and create new business models, either creating completely new companies of tomorrow or transforming existing companies. Two noteworthy examples here are smart cars and personalised entertainment. Smart cars and autonomous driving are changing the auto industry, and we can see the value being captured by Tesla. AI tech that personalises, curates, and creates entertainment content is transforming the business of entertainment and media companies and is evident in the value being captured by Netflix and Amazon. AI technology is evolving rapidly, riding on huge investments being made in AI R&D by universities, governments, big technology companies and start-ups, and also supported by innovations in hardware and infrastructure. What will be possible in the near future will be very different from what is possible today. While there will be variations in degree and timeline of impact depending on the nature of business, almost all industries will be affected in one way or the other. AI in healthcare is going to save lives, cut costs, and improve the wellbeing of people across the world. This will happen through personalised healthcare, preventive diagnosis and AI-assisted physicians. AI is driving an enormous transformation of the education sector as well, bringing value to students and institutions. By leveraging AI and other digital technologies, it has become possible to make education more convenient and personalised. AI will automate administrative tasks, allowing educators to pay more attention to students’ skills and overall development. By meeting the education needs of millions of children and adults cost-effectively and with limited resources, AI and digital technology can help developing countries like India solve developmental issues at scale. A question I often get asked is what other big impacts AI will have in the business world. To be honest, nobody knows the full answer, but we can be sure it will be profound. In the early days of the internet, around the year 2000, we were mostly discussing applications for e-commerce and content consumption. Who would have dreamt at that time of applications like Instagram, Uber or Airbnb? Similar will be the case for AI – entrepreneurs and innovators will use their ingenuity to leverage the latest AI technology to come up with things in the future that we cannot even dream of today. Existing companies could lead the way with AI-enabled disruption and create new business models or choose to leverage AI in their existing business to improve effectiveness and maintain competitive advantage. In order to be successful at AI-led transformation, existing companies need to create a culture that empowers an agile way of working either in their overall business or in a new unit focusing on innovation. This should be accompanied by a flat organisational structure and empowered and autonomous teams. Risk-taking and learning should be encouraged, while failures should be looked at as learning opportunities rather than just frowned upon. Not doing anything to participate in the AI wave is not an option, and such companies could face a threat to their business in the near future. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"In the last 40 years, we have experienced two major technology-led waves of value creation: one driven by the rise of personal computing, the next by the dawn of the internet. A third wave is now being led by artificial intelligence (AI)-enabled digital transformation. Prior to the outbreak of COVID-19, digital transformation was part of […]","categories":["AI Features"],"tags":[],"author_name":"Ashwin Mittal","publish_date":"2021-11-18T16:00:00","publication_year":"2021","word_count":684,"keywords":["data science","Go","API","artificial intelligence","AI","Git","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-re-imagining-business-value-creation-through-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10012244,"title":"Bringing AI &#038; Analytics In The Solar Industry: Story Of SenseHawk","content":"Similar to every other industry, digitisation has hit the industry of solar energy as well. The traditional approach of manually operating these massive solar farms has been a tedious task, and with the rising demand for clean energy, the process becomes even more challenging. This, in turn, has called for standardising the processes of managing solar farms and bringing in digitisation to drive a seamless operation of solar power plants. To address such needs, startups like SenseHawk have been established. Founded by two experienced professionals of the industry — Swarup Mavanoor and Rahul Sankhe in 2018, the startup, SenseHawk builds software to bring in AI and automation for solar companies to have a seamless operation across teams, according to the company’s website. With an aim to create a cleaner planet by providing AI-powered software processes for the solar industry, this California and Bangalore-based startup claims to have deployed its software at over 400 solar sites. To understand better, how the startup is leveraging AI, ML and analytics techniques for helping the solar industry, we got in touch with one of the founders, Swarup Mavanoor. To start with, he said, “Our AI-powered software not only speeds up the operating process for solar companies but also ensures efficiency in its clear inspection,” said Mavanoor. How AI-Powered Software Works? In order to provide digitisation to solar companies, SenseHawk has developed an AI-powered software tool that has been designed to be compatible with analysing data from all sorts of devices like drones, cameras, sensors etc., used in solar asset management. This AI-powered software can help solar companies to streamline asset processes for an entire solar lifecycle. Two of the company’s core applications are — Eye and Therm that leverage cutting-edge AI algorithms to enable automatic detection of hotspots and help solar companies sift through and process tens of thousands of thermal images captured by drones. The AI-based asset management tool — Eye provides a 360-degree view of construction sites of any size, with simplified monitoring, whereas, the infrared image processing and reporting tool — Therm uses AI to detect and classify defects on a site based on thermal imagery. From structure installations and module mounting to grading\/levelling and cable laying, SenseHawk’s Eye can effectively monitor all the work being done on-site and accurately assess their quality and progress with the help of images acquired by drones. Besides monitoring, Eye can also be utilised for quality checks, as-built validation, and inventory control. On the other hand, Therm works with thermal imagery that ensures the optimal health of solar sites. This fully automated AI workflow creates a thermal map of your solar site, helping evaluate severity and energy loss, schedule maintenance of routine or one-off activities, and track all existent defects, thus maximising the project’s energy yield. By accurately mapping the detected issues to the site model, Therm enables managers to isolate those issues and assess the overall performance, through an easy to use web interface. “We have developed a geographic information system, design automation, and data management platform enabling developers to manage multiple development projects and big scaled sites,” said Mavanoor, when asked about the technological standpoint. “The AI-based technology platform can help gather all accurate measurements real-time and create contour maps to better understand your site, and create high-resolution cloud points as you go to keep your project running smoothly,” he adds. Furthermore, the company is also successfully leveraging the digital twin technology to essentially provide the solar companies with a real-time look at their physical assets and their performances. By integrating cross-platform data with a cloud-based site model, SenseHawk enabled seamless access to relevant information at all stages of the solar asset life cycle. “While frameworks like Vuejs, Tensorflow, PyTorch, Django, Nodejs, PostgreSQL, and MongoDB are some of the tools currently being used by our developers, our next plausible step is to build new modules in the pipeline and to improve our existing products including the computer vision modules,” said Mavanoor. “SenseHawk is also looking to apply deep learning technology to solve the challenges related to solar plant layout design optimisation.” Wrapping Up Currently, SenseHawk is looking to build a one-stop solution to solve all emerging changes in our solar industry. As a matter of fact, the ongoing pandemic has spurred the digitisation wave in the solar industry and necessitated remote management for which drone data analytics plays a pivotal role. This, in turn, has had a positive impact on the startup. According to the news media, the company so far has delivered the power of AI and analytics for more than 28 gigawatts of solar assets across 15 countries worldwide. “With our AI-based software, the solar companies can now experience automation and tackle their efficiency problem effectively,” concluded Mavanoor.","excerpt":"Similar to every other industry, digitisation has hit the industry of solar energy as well. The traditional approach of manually operating these massive solar farms has been a tedious task, and with the rising demand for clean energy, the process becomes even more challenging. This, in turn, has called for standardising the processes of managing […]","categories":["Deep Tech"],"tags":["ai grading"],"author_name":"Sejuti Das","publish_date":"2020-11-24T12:00:22","publication_year":"2020","word_count":787,"keywords":["AI","PyTorch","ai grading","ML","computer vision","RAG","Aim","deep learning","analytics","edge AI","TensorFlow"],"extracted_tech_keywords":["AI","ML","deep learning","computer vision","analytics","Aim","TensorFlow","PyTorch","edge AI","RAG"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/bringing-ai-analytics-in-the-solar-industry-story-of-sensehawk\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10131981,"title":"Synthetic Data Generation in Simulation is Keeping ML for Science Exciting","content":"If only AI could create infinite streams of data for training, we wouldn’t have to deal with the problem of not having enough data. This is what is keeping a lot of things undiscoverable in the field of science as there is only a limited amount of data available that can be used for training. This is where AI is taking up a crucial role with the help of simulation. The integration of data generation through simulation is rapidly becoming a cornerstone in the field of ML, especially in science. This approach not only holds promise but is also reigniting enthusiasm among researchers and technologists. As Yann LeCun pointed out, “Data generation through simulation is one reason why the whole idea of ML for science is so exciting.” Data generation through simulation is one reason why the whole idea of ML for science is so exciting. https:\/\/t.co\/iC3lkquKOf— Yann LeCun (@ylecun) August 6, 2024 Simulations allow researchers to generate vast amounts of synthetic data, which can be critical when real-world data is scarce, expensive, or challenging to obtain. For instance, in fields like aerodynamics or robotics, simulations enable the exploration of scenarios that would be impossible to test physically. Richard Socher, the CEO of You.com, highlighted that while there are challenges, such as the combinatorial explosion in complex systems, simulations offer a pathway to manage and explore these complexities. Synthetic Data is All You Need? This is similar to what Anthropic chief Dario Amodei said about producing quality data using synthetic data and that it sounds feasible to create an infinite data generation engine that can help build better AI systems. “If you do it right, with just a little bit of additional information, I think it may be possible to get an infinite data generation engine,” said Amodei, while discussing the challenges and potential of using synthetic data to train AI models. “We are working on several methods for developing synthetic data. These are ideas where you can take real data present in the model and have the model interact with it in some way to produce additional or different data,” explained Amodei. Taking the example of AlphaGo, Amodei said that those little rules of Go, the little additional piece of information, are enough to take the model from “no ability at all to smarter than the best human at Go”. He noted that the model there just trains against itself with nothing other than the rules of Go to adjudicate. Similarly, OpenAI is a big proponent of synthetic data. The former team of Ilya Sutskever and Andrej Karpathy has been a significant force in leveraging synthetic data to build AI models. The development at OpenAI is testimony to the advanced growth of generative AI in the entire ecosystem, but not everyone agrees that they will be able to achieve AGI with the current methodology of model training. Likewise, Microsoft is also researching in this direction; its research on Textbooks Are All You Need is a testament to the power of synthetic data. Google’s AlphaFold, which is spearheading protein fold prediction and creations for drug discovery, too can benefit immensely from synthetic data. At the same time, it can be scary to use this data for a sensitive field like science. Synthetic Data is Too Synthetic However, the potential of simulations extends beyond mere data generation. Giuseppe Carleo, another expert in the field, emphasised that the most exciting aspect is not just fitting an ML model to data generated by an existing simulator. Instead, true innovation lies in training ML models to become advanced simulators themselves—models that can simulate systems beyond the capabilities of traditional methods, all while remaining consistent with the laws of physics. This is becoming possible with synthetic data generated by agentic AI models, which are increasing in the field of AI. Models that can test, train, and fine-tune themselves using the data they created is something that is exciting for the future of AI research. Moreover, the discussion around simulations also touches on broader applications. Sina Shahandeh, a researcher in the field of biotechnology, for example, suggested that the ultimate simulation could model entire economies using an agent-based approach, a concept that is slowly becoming feasible. Despite the excitement, the field is not without its sceptics. Stephan Hoyer, a researcher with a cautious outlook on AGI, pointed out that simulating complex biological systems to the extent that training data becomes unnecessary would require groundbreaking advancements. He believes this task is far more challenging than achieving AGI. Similarly, Jim Fan, senior AI scientist at NVIDIA, said that while synthetic data is expected to have a noteworthy role, blind scaling alone will not suffice to reach AGI. When it comes to science, using synthetic data can be tricky. But its generation in simulation shows promise as it can be tried and tested without deploying in real-world applications. Besides, the possibility of it being infinite is what keeps ML exciting for researchers.","excerpt":"Simulations allow researchers to generate vast amounts of synthetic data, which can be critical when real-world data is scarce, expensive, or challenging to obtain.","categories":["AI Trends"],"tags":["Machine Learning"],"author_name":"Mohit Pandey","publish_date":"2024-08-09T15:08:25","publication_year":"2024","word_count":823,"keywords":["Anthropic","Go","agentic AI","OpenAI","AI","AWS","ML","Machine Learning","RAG","generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","agentic AI","OpenAI","Anthropic","RAG","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/synthetic-data-generation-in-simulation-is-keeping-ml-for-science-exciting\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10023001,"title":"Top 10 Python Tools For Time Series Analysis","content":"Time series is a sequence of numerical data points in successive order and time series analysis is the technique of analysing the available data to predict the future outcome of an application. At present, time series analysis has been utilised in a number of applications, including stock market analysis, economic forecasting, pattern recognition, and sales forecasting. Here is a list of top ten Python tools, in no particular order, for Time Series Analysis. 1| Arrow About: Arrow is a Python library that offers a human-friendly approach to creating, manipulating, formatting and converting dates, times and timestamps. The library implements and updates the datetime type, plugging gaps in functionality and providing an intelligent module API that supports many common creation scenarios. Features include: Fully-implemented, drop-in replacement for datetimeTimezone-aware and UTC by defaultSupport for Python 3.6+Shift method with support for relative offsets, including weeksFull support for PEP 484-style type hints Know more here. 2| Cesium About: Cesium is an open source library that allows users to extract features from raw time series data, build machine learning models from these features, as well as generate predictions for new data. The cesium library also powers computations within the cesium web interface, which allows similar time series analyses to be performed entirely within the browser. Know more here. 3| Featuretools About: Featuretools is an open source Python library for automated feature engineering. The framework excels at transforming temporal and relational datasets into feature matrices for machine learning. Featuretools references generated features through the feature name. In order to make features easier to understand, Featuretools offers two additional tools, featuretools.graph_feature() and featuretools.describe_feature(), to help explain what a feature is and the steps Featuretools took to generate it. Know more here. 4| TSFRESH About: TSFRESH or Time Series Feature extraction based on scalable hypothesis tests is a Python package with various feature extraction methods and a robust feature selection algorithm. The package automatically calculates a large number of time series characteristics and contains methods to evaluate the explaining power and importance of such characteristics for regression or classification tasks. Advantages include: It is compatible with sklearn, pandas and numpyIt allows anyone to easily add their favorite featuresIt both runs on the local machine or even on a cluster Know more here. 5| Pastas About: Pastas is an open-source Python framework designed for processing, simulation and analysis of hydrogeological time series models. Introduced by Raoul A. Collenteur, Mark Bakker, Ruben Calje, Stijn A. Klop and Frans Schaars, this framework has built-in tools for statistically analysing, visualising and optimising time series models. The two major objectives of this library are: To provide a scientific software package for the development and testing of new hydrogeological methods using few lines of Python codeTo provide an efficient and easy-to-handle library for groundwater practitioners. Know more here. 6| PyFlux About: PyFlux is an open source library for time series analysis and prediction. In this library, users can choose from a flexible range of modelling and inference options, and use the output for forecasting and retrospection. The library allows for a probabilistic approach to time series modelling. The latest release version of PyFlux is available on PyPi. Python 2.7 and Python 3.5 are supported, but development occurs primarily on 3.5. Know more here. 7| TimeSynth About: TimeSynth is an open source library for generating synthetic time series for model testing. The library can generate regular and irregular time series. The architecture of this library allows the user to match different signals with different architectures allowing a vast array of signals to be generated. At present, the library supports only Python 3.6+ versions. Know more here. 8| Sktime About: Sktime is a unified python framework that provides API for machine learning with time series data. The framework also provides scikit-learn compatible tools to build, tune and validate time series models for multiple learning problems, including time series classification, time series regression and forecasting. Know more here. 9| Darts About: Darts is a python library for easy manipulation and forecasting of time series. It contains a variety of models, from classics such as ARIMA to neural networks.  Darts supports both univariate and multivariate time series and models, and the neural networks can be trained multiple time series. Know more here. 10| Orbit About: Orbit is a Python framework created by Uber for Bayesian time series forecasting and inference. The framework is built on probabilistic programming packages like PyStan and Uber’s own Pyro. At present, Orbit supports the implementations of a few forecasting models, such as Damped Local Trend (DLT), Exponential Smoothing (ETS) and Local Global Trend (LGT). Know more here.","excerpt":"Time series is a sequence of numerical data points in successive order and time series analysis is the technique of analysing the available data to predict the future outcome of an application. At present, time series analysis has been utilised in a number of applications, including stock market analysis, economic forecasting, pattern recognition, and sales […]","categories":["AI Trends"],"tags":["python time series","python tools","Time Series Analysis"],"author_name":"Ambika Choudhury","publish_date":"2021-03-29T12:00:00","publication_year":"2021","word_count":764,"keywords":["Time Series Analysis","scikit-learn","NumPy","machine learning","TPU","AI","neural network","python time series","Python","Ray","python tools","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","neural network","Ray","scikit-learn","Pandas","NumPy","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-python-tools-for-time-series-analysis\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10059623,"title":"IBM launches new Mainframe model, aims to regain lost ground","content":"International Business Machines dominates the mainframe market with a 90% share. Now, the tech giant has announced the launch of a new model of IBM Z series mainframe. It is expected to hit the markets either in the first or second half of 2022, said IBM in its recent Q4 earnings call. One cloud to rule them all The Covid-19 pandemic accelerated the adoption of cloud. Enterprises woke up to the need for distributed cloud-based solutions in lieu of a single-vendor approach. Thanks to cloud, companies can now optimise workloads, share resources seamlessly and cut down on data centre overheads. IBM had a delayed start in cloud computing compared to tech giants like Amazon, Microsoft and Google. The major portion of the former’s revenue came from consulting and selling hardware devices. So instead of playing catch-up with rivals like Amazon and Microsoft in the public cloud space, IBM decided to bet on hybrid cloud and artificial intelligence. Despite the cost-saving benefits and ease of sharing resources, only 25% of enterprise workloads have been moved to the cloud. The main concerns organisations have around shifting to cloud include the pain of migrating legacy systems, integration issues, and data sovereignty problems. IBM’s hybrid cloud offering is designed to address such concerns and drive enterprise cloud adoption. Also Read:- AI adoption Organisations have identified AI and automation as the sine qua non to remain competitive in a post-pandemic world. AI adoption in at least one function is up from 50 percent in 2020, according to a Mckinsey & Company report. The survey results suggest that AI adoption has increased most at companies headquartered in emerging economies, including China, the Middle East and North Africa. Across regions, the adoption rate is highest at Indian companies, the report added. Companies had to take the digital route to meet new business challenges brought on by the pandemic. As a result, the organisations are heavily investing in areas such as automating IT and processes, building trust in AI outcomes and understanding the language of business, said Rob Thomas, Senior Vice President, IBM Cloud and Data Platform. IBM is currently working on a ton of AI automation tools to help enterprises leverage AI. For instance, Cloud Pak for data uses AI to answer distributed queries and gives results 8x faster and at nearly half the cost compared with other data warehouses. Watson Orchestrate, a new interactive AI-powered tool made by IBM, boosts the productivity of business professionals across departments. Two-pronged approach IBM’s Hybrid cloud (powered by AI) approach aims to overcome the existing bottlenecks arising from being tied to a single cloud vendor. At the same time, increased AI adoption among organisations will engender demand for powerful machines capable of storing and processing huge volumes of data. This is where IBM’s new model of mainframes comes in. The model offers a complete end-to-end cloud solution for the enterprises’ computing needs. IBM has let go of its low margin managed infrastructure business by spinning off it into a new business unit called Kyndrl. In sum, with its new hybrid cloud and AI offering, combined with the mainframes series, IBM is committed to taking back its long lost clout in the technology space.","excerpt":"Despite the cost-saving benefits and ease of sharing resources, only 25% of enterprise workloads have been moved to the cloud.","categories":["IT Services"],"tags":["Amazon","Data Privacy","digital transformation","Google","Hybrid Cloud","IBM","Microsoft"],"author_name":"SharathKumar Nair","publish_date":"2022-02-02T14:00:00","publication_year":"2022","word_count":533,"keywords":["Go","artificial intelligence","AI","cloud computing","digital transformation","ML","Amazon","Git","RAG","Aim","Hybrid Cloud","Data Privacy","IBM","Google","Rust","R","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","RAG","cloud computing","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/mainframe\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10047279,"title":"A Complete Guide to Sequential Feature Selection","content":"In machine learning, feature selection is the procedure of selecting important features from the data so that the output of the model can be accurate and according to the requirement. Since in real-life development procedure, the data given to any modeller has various features and it happens all the time that there are various features given in the data which are not even required for the generation of the models and also the presence of those features can reduce the performance level of the model. So in modelling, it becomes an important step of data preprocessing. Basically, the feature selection is a method to reduce the features from the dataset so that the model can perform better and the computational efforts will be reduced. In feature selection, we try to find out input variables from the set of input variables which are possessing a strong relationship with the target variable. This means changes in an input variable should form changes in the output variable. There can be various reasons to perform feature selection. Simplification of the model.Less computational time.To avoid the curse of dimensionality.Improve the compatibility of data with models. Roughly the feature selection techniques can be divided into three parts. Filter method.Wrapper method.Embedded method. Filter methods These methods are very fast and easy to do the feature selection. In this method, we perform feature selection at the time of preprocessing of the data. These methods select the features before using a machine learning algorithm on the given data. But the problem with the method is that it does not remove the multicollinearity from the data. The basic architecture of the modelling procedure with these methods of feature selection is as follows. Some of the techniques under the filter methods are: Information gain.Chi-square test.Fisher’s Score.Correlation coefficient.Variance thresholdMean absolute deviation.dispersion coefficient Wrapper Methods(Greedy Algorithms) In this method, feature selection algorithms try to train the model with a reduced number of subsets of features in an iterative way. In this method, the algorithm pushes a set of features iteratively in the model and in iteration the number of features gets reduced or increased. The stopping criteria of the iteration are to select the features which will give the best result in modelling. So at the end of the procedure of the wrapper method, the optimal set of features gets selected for the modelling. The basic architecture of the technique followed by the wrapper methods is as follows: Some of the techniques under these methods are: Foreword selection.Backward selection.Bi-directional selection.Exhaustive selection.Recursive selection. Note- the main motive of the article is to talk about the sequential feature selection which is a technique related to the wrapper methods so later in the article we will discuss the wrapper methods. Embedded methods This method is the combination of the filter method and wrapper methods which is fast as filter methods are and accurate as of the wrapper method. These methods are blended as part of the learning algorithm. They are most accurate because they overcome the drawbacks of filter and wrapper methods. Some of the techniques under this method are: Regularization: Main regularization methods are: Lasso regularization.Ridge regularization.Elastic net regularization. Tree-based methods: Main tree-based methods are: LightGBM.XGBoost. The basic architecture of any procedure containing embedded methods as feature selection techniques is as follows. Sequential Feature Selection Algorithms Sequential feature selection algorithms are basically part of the wrapper methods where it adds and removes features from the dataset sequentially. Sometimes it evaluates each feature separately and selects M features from N features on the basis of individual scores; this method is called naive sequential feature selection. It works very rarely because it does not account for feature dependence. In a proper technique, the algorithm selects multiple features from the set of features and evaluates them for model iterate number between the different sets with reducing and improving the number of features so that the model can meet the optimal performance and results. Mathematically these algorithms are used for the reduction of initial N features to M features where M<N. and the M features are optimized for the performance of the model. The sequential feature selection method has two components: An objective function: The method finds to minimize the number of overall features in a subset from the set of all features. So the results can be enhanced. It can be called the criterion where the mean squared error is a criterion for regression models and the misclassification rate is a criterion for the classification model. A sequential search algorithm: This searching algorithm adds or removes the feature candidate from the candidate subset while evaluating the objective function or criterion. Sequential searches follow only one direction: either it increases the number of features in the subset or reduces the number of features in the candidate feature subset. On the basis of movement, we can divide them into two variants. Sequential forward selection(SFS) In SFS variant features are sequentially added to an empty set of features until the addition of extra features does not reduce the criterion. Mathematically if the input data in the algorithm is Then the output will be : Where the selected features are k and K<d. In the initialization X is a null set and k=0 (where k is the size of the subset). In the termination, the size is k = p where p is the number of desired features. Sequential Backward Selection (SBS) This variant algorithm picks all the features from the input data and combines them in a set and sequentially removes them from the set until the removal of further features increases the criterion. mathematically if the input data is The output of the variant will be In the initialization X is a subset of features and k=d (where k is the size of the subset). In the termination, the size is k = p where p is the number of desired features. There are two more variants of the sequential feature selection. Sequential forward floating selection.Sequential backward floating selection. These floating variants are the extensions of the SFS and SBS where they consist of an additional execution or inclusion step to remove features if once they are included or excluded in the procedure. These extensions provide a larger number of features in the final combination of features sets. Let’s see how we can implement these all using python. MLxtend is a package that provides the implementation of sequential feature selection methods. You can check the whole code at this link. Here in the article, I will just give the images and will explain to them how did it work in the background In this procedure, I am using the iris data set and feature_selection module provided in mlxtend library. In the following codes after defining x, y and the model object we are defining a sequential forward selection object for a KNN model. from mlxtend.feature_selection import SequentialFeatureSelector as SFS sfs1 = SFS(knn, k_features=3, forward=True, floating=False, verbose=2, scoring=’accuracy’, cv=0) sfs1 = sfs1.fit(X, y) Output: Here in the output, we can see all the scores with the number of features selected in the process. In the SFS object we have provided our model which is KNN and the number of the input features in the data set forward = True will tell the object to follow the SFS and floating + false because this is an example of SFS not SFFS of SFBS. Using sfs.subsets_ we can cross-check all the results of every step. Here in the results, we can see the results of every step with the name of the feature in the data set. Just by changing the parameters, we can toggle between all the variants of the sequential feature selection. For SBS forward = Falsefloating = False For SFFS forward = Truefloating = True For SBFS forward = Falsefloating = True Printing the results. We can also save the results in the DataFrame. Results of SFS Results of SBS Here we can see the SFS and SBS found that their best result in three features however the procedure of feature selection was different. In the data frame; cv_score = cross validation score.ci_score = confidence interval.std_dev = standard deviation of cross validation score.Std_err = standard error of the cross validation score. Visualizing the standard deviation for the SFS Here in this article, we have seen the basics of feature selection and why it becomes an important part of the modelling procedure with the data. There are various methods that can be used for feature selection. But we prefer to go with the sequential feature selection because by the time performing the modelling we can have full control of the procedure wherein embedded methods we can just perform it we can not control it in between the procedure and filtering methods are faster but they are not accurate. The main motive of the article was to make feature selection automatic but should be in the controlled form. In the reference, I have provided the link for the whole code, so if any reader wants to practice them they can access the notebook. References: Sequential Feature SelectorGoogle Colab notebook for the above codes","excerpt":"In machine learning, feature selection is the procedure of selecting important features from the data so that the output of the model can be accurate and according to the requirement. Since in real-life development procedure, the data given to any modeller has various features and it happens all the time that there are various features […]","categories":["Deep Tech"],"tags":["Data Engineering","feature selection","Guide","iris dataset python","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2021-08-29T17:00:00","publication_year":"2021","word_count":1521,"keywords":["Go","iris dataset python","machine learning","TPU","AI","ML","Machine Learning","Colab","Python","XGBoost","Data Engineering","LightGBM","R","feature selection","Guide"],"extracted_tech_keywords":["AI","machine learning","ML","XGBoost","LightGBM","Colab","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-complete-guide-to-sequential-feature-selection\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10002473,"title":"What Data Science Graduates Need To Do To Get Hired During Covid-19 Pandemic","content":"While news of data science graduates from tier-1 colleges getting hired with good salary packages amid a recession and an ongoing pandemic is mildly reassuring, there is no contending the fact that jobs are scarce and freshers need to brace themselves for an altered work landscape. However, despite the uncertainty, some companies have been hiring for various positions, and you can do plenty to catch their attention. Here are some of the ways in which you can do this: – Filter Your Search Sector-Wise If you have been following the news keenly, you can deduce the sectors that are doing well, and hence, may be more likely to build their capacity and hire more people. Although sectors like hospitality and tourism have been struggling to bounce back, logistics and healthcare have been booming. Given this, it may be a good practice to do some research on companies that are doing well in these sectors before aimlessly sending across applications. Once you have shortlisted companies sector-wise, focus on openings at these places, since they may not advertise them widely. Identify Portals That Regularly Update Jobs Feed While portals like LinkedIn and Glassdoor have become commonplace among professionals, not only to build a strong network, but also to get the latest information on vacancies, registering to more job boards like Indeed and Monster will help narrow down your search further. You should also enable job notifications on social media platforms like Facebook and sign up with platforms like Google for Jobs, CareerBuilder, and more. Be sure to update your profiles across these platforms as and when required. Network Aggressively While looking out for jobs directly is a good practice to have, it is just as important to talk to people within your circle for opportunities. The first rule to networking is to update your profiles so people can help you appropriately. Secondly, always try and leverage personal networks first – you will be surprised how wide your network could go with a little help from people you already know. Not only can these connections land you jobs through referrals, but it will also help you widen your network. Start by adding people you already know, and then reaching out and building a rapport with people who you think may be able to help you. Develop Relevant Skills In a dynamic field like data science, you are expected to continuously learn and build your skill sets. Given the pace at which technology has been moving across industries, it has become a basic requirement to take the initiative to pick up on additional skills you did not study in college or know very little about. Even if these skills do not directly help you land a job, they will certainly add a layer of competence to your profile and automatically make you a more attractive candidate. Sign up for online courses, read books, or attend online forums – there are many ways to develop your capabilities. Prepare Well For Interviews Job interviews can be anxiety-inducing, but it need not be that way. If you prepare well, facing your job interview will not be very challenging. With lockdowns reinstated in many districts and many places of leisure yet to reopen, you have the time to invest in your self-improvement, and you should set aside some time in a day to prepare for interviews. Today, there are short-term courses as well that can teach you how to master the skills required to give a good interview. Since most interviews – at least in the coming months – are likely to be held virtually, it would be best to prepare for remote interviews and practice mock interviews. Meanwhile, Consider Freelance Or Short-Term Work Since landing a job is not under your control alone and is subject to various factors, including available opportunities, it could be a while before you can join a place that aligns with your expertise. In the interim, it will be good to add some work experience to your portfolio by accepting freelance work or signing up for short-term jobs.While these may not be the type of work you would have liked to be involved in, an additional experience can give you an advantage in the long-run, teach you many things, and give you the confidence you may be lacking.","excerpt":"While news of data science graduates from tier-1 colleges getting hired with good salary packages amid a recession and an ongoing pandemic is mildly reassuring, there is no contending the fact that jobs are scarce and freshers need to brace themselves for an altered work landscape. However, despite the uncertainty, some companies have been hiring […]","categories":["AI Highlights"],"tags":["Data Science Jobs","data scientist salary","red hat"],"author_name":"Anu Thomas","publish_date":"2020-07-15T10:00:36","publication_year":"2020","word_count":716,"keywords":["data science","Go","red hat","data scientist salary","AI","programming_languages:R","ML","programming_languages:Go","Data Science Jobs","RAG","Aim","R"],"extracted_tech_keywords":["AI","ML","data science","Aim","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/what-data-science-graduates-need-to-do-to-get-hired-during-covid-19-pandemic\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10097515,"title":"How Many Jobs has AI Actually Gobbled Up?","content":"“AI will not replace you, but someone who uses AI will.” Like a cautionary sign stuck on hazardous substances, the above line has been floating around for quite some time now. Economists and tech leaders have been ringing the death knell for a while cautioning about AI, which is looked upon as a potential reason for an impending apocalypse in the job market. With predictive stats on how certain jobs would be gone in an ‘xyz’ timeframe, a reality check would keep anxieties in check. So what’s the brouhaha all about? Has AI, or someone using AI replaced you? The media: \"AI is replacing 300 million jobs\"The jobs it’s replacing: pic.twitter.com\/HAh5XjwGan— Elkmont AI (@Elkmontai) July 19, 2023 Executive outplacement and career consulting firm Challenger, Gray & Christmas, attributed 4000 job losses in May to artificial intelligence, making it the first time for the company to mention AI as a cause of job loss. Even though there have been massive layoffs owing to recession, including at big tech companies such as Microsoft, Meta, and others in the past six months, none of them have pointed to AI as a cause of it. However, observing the trend at those same companies on how each of them is adopting generative AI in their workflow, it is not difficult to connect the dots. But, having said that, the trends in the job market are reflecting another picture. Generative AI Fuels Job Market With massive adoption of generative AI in enterprises, which is either fuelled by anxiety or enthusiasm, AI has indeed kick-started a job evolution in the market. As per AIM Research, the generative AI job market has witnessed a steady growth from January to June of this year. Generative AI-related job postings in the United States are said to have risen by 20% in May. From 3000 job openings in April, the job counts have risen to 4500 jobs in June. The IT sector has observed the highest number of job positions for generative AI roles. The figures may be indicative of jobs not seeing a decline, but job roles have been modified to suit the current wave. For instance, the role of a ‘generative AI engineer’, a role that never existed before, will require skills of engineers equipped in different fields. Therefore, the role of a generative AI engineer will encapsulate the roles of a deep learning engineer, ML, NL and also a software engineer. New roles that probably didn’t exist earlier are also sprouting in full vigour. The role of prompt engineers — a result of the chatbot revolution — has witnessed an uptick with companies increasingly seeking such roles. With the role offering salaries higher than Python developers, the job market is only looking positive. The massive shift brought by AI has also sparked a debate on how an entire generation will study for jobs that won’t exist. It might not be an overstatement to say that certain job roles may become redundant, but it is mostly because the nature of the job role is seeing a change. Together We Grow Enterprises are approaching the generative AI rage in a coalition of sorts — not via replacements, but through implementation and training of their employees to tame the system. TCS, which initially partnered with Google Cloud for their generative AI services, recently partnered with Microsoft Azure to train 25000 engineers on Azure OpenAI. In addition to implementing and training, enterprises are also building ways to support other companies thrive on generative AI. Tech Mahindra, which partnered with Microsoft to enable generative AI-powered enterprise search, unveiled their Generative AI Studio to help other enterprises kickstart their efforts in generative AI. Other IT companies have also followed suit. AI Over Humans? While most companies have found ways to work around the generative AI job buzz, there have also been companies that have openly embraced AI over human resources. Dukaan, a platform for enabling merchants to set up their e-commerce business, recently laid off 90% of their support staff replacing them with their new AI chatbot. The company claims to have saved cost and reduced customer resolution time since the move. Telecom company British Telecommunications said that over 55,000 jobs will be cut by the end of the decade, out of which a fifth will be in customer service where AI will replace staff. There are also industries that have no choice but to embrace generative AI — travel industry being one of them. The industry that faced the biggest impact owing to the pandemic, is now slowly breathing and they all have integrated AI chatbots, ChatGPT plugins and other features — AI being the saviour. While you have companies and industries relying on AI, Zerodha on the other hand is all out to safeguard its employees from any form of AI job takeover. The company has been clear on its stance to adopt AI only if they deem necessary and it will not be at the cost of someone’s job. With a few companies allowing AI to replace jobs, and many others creating new jobs and also embracing AI to empower their employees to effectively use it without posing any threat to their jobs, it is fair to say that AI is becoming integral in all jobs. However, whether it will be a deciding factor to safeguard one’s job is not conclusive.","excerpt":"Given the anxiety over AI replacing jobs, let’s find out how much of it has actually happened this year","categories":["AI Features"],"tags":["Accenture","AI Chatbot","Azure","Dukaan","Generative AI","Google Cloud","IT","Meta","Microsoft","OpenAI","prompt engineering","TCS","Tech Mahindra","Zerodha"],"author_name":"Vandana Nair","publish_date":"2023-07-25T15:33:01","publication_year":"2023","word_count":890,"keywords":["IT","Tech Mahindra","Ray","prompt engineering","deep learning","Generative AI","Dukaan","Accenture","ChatGPT","artificial intelligence","RAG","Meta","Google Cloud","AI","ML","generative AI","TCS","OpenAI","AI Chatbot","Aim","Zerodha","Azure","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","generative AI","ChatGPT","OpenAI","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-many-jobs-have-ai-actually-gobbled-up\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":36060,"title":"5 Key Benefits Of Using AI As A Long-Term Strategy For Business","content":"Artificial intelligence (AI) can solve a myriad of organisational issues and bring in an aspect of much-needed agility as humans evolve on to making more complex decisions. AI can benefit people by freeing them from mundane, unchallenging tasks, hire new recruits, aid logistical and production problems. Now that we’ve dabbled with several isolated use cases to test the waters, we need to understand how AI can help businesses make a strategic shift in operations across the enterprise. If you’re only considering implementing chatbots to streamline your customer experience instead of looking at resolving larger business problems, you’re not leveraging the full potential that this technology has to offer. Every business model is unique, so understanding if and how AI is the answer for your organisation is crucial. Let’s discuss how AI can help you plan a long-term organisational strategy: Hiring Assistance: The initial stages of a hiring process can be repetitive, cumbersome and utterly mundane. Imagine if you had a system that qualified leads as per your requirement and filtered them for review. AI assistants can now understand KRAs uploaded by human resources (HR), search and identify the best-suited people from a database as well as connect and converse with the candidate for an initial level screening. Bots can also validate and shortlist candidates for an interview basis this screening and offer a complete interaction history to HR. A recent survey by Korn Ferry found that 69 percent of recruiters surveyed said using AI for candidate sourcing gets them a higher quality of candidates and a whopping 87 percent are excited about working with AI and don’t feel it will replace them in the future. The best aspect of chatbots streamlining such processes is that it saves time that can be spent conducting in-person interviews to test candidates on the softer skills they bring to the organisation. Cross Leveraging Information: An incredible amount of data is churned with the use of several applications that collect information. The information obtained with a specific department can be utilised as intelligence for another within the same organisation. For instance, data collected by the R&D team can be turned into insights for the marketing and sales department, which would help them target their audience better, understand their behavioural patterns, spending habits etc. Customer feedback received through conversational chatbots employed by the customer service department can help the R&D team improve the products they develop. One possible reason that can withhold this flow of information could be siloed within which data is currently stored within the organisation. Once this is integrated, the possibilities and business benefits are boundless. Improving Employee Productivity: Employees spend a lot of their time at work looking for reports, indexing and compiling data. Machine learning can help with such menial tasks thereby freeing up the employee’s time and increasing efficiencies within teams. Employees would have the time to work on creative and strategically focused tasks whereas companies would have a largely error-free directory of data computed in real-time. Streamlining Processes: AI can build custom reports in seconds based on custom searches, customise insights across business groups and functions, export snippets across formats and integrate repositories. It can also be integrated with applications like Skype for business, CRM, LMS, SharePoint, AD, etc., and generate reports specific to a company or client. Let’s take for instance supply chain processes, there is a significant amount of purchases to index, invoices to handle and millions of orders to fulfil. Using tools like computer vision and process automation tools, this entire operation can be simplified with the highest levels of accuracy. Customer Satisfaction: According to Forrester, we’ve begun having smarter and more strategic conversations having entered the era of automated customer service. The goal now is to effectively engage the customer using a medium of their choice, without intruding in on their privacy. Enterprises have been making constant efforts to simplify interactions and processes. Prior to having an informed conversation with a customer an AI assistant does not need time to research and can generate considerable choices of action with greater accuracies. They can not only provide precise replies but also make interfaces cost-effective by learning the customer’s needs and preferences from previous chat histories. There is currently a gap with what organisations expect from AI and what AI can do for your organisation. There is a need to educate oneself about where AI can be leveraged to streamline business processes before jumping onboard the hype wagon. AI and NLP powered chatbots are changing the world as we know it, aiding innovation and revolutionising customer service by reducing operational costs but identifying and leveraging the full potential of artificial intelligence can unlock many long-term strategic benefits for organisations.","excerpt":"Artificial intelligence (AI) can solve a myriad of organisational issues and bring in an aspect of much-needed agility as humans evolve on to making more complex decisions. AI can benefit people by freeing them from mundane, unchallenging tasks, hire new recruits, aid logistical and production problems. Now that we’ve dabbled with several isolated use cases […]","categories":["AI Trends"],"tags":["AI Benefits","AI Strategy"],"author_name":"Animesh Samuel","publish_date":"2019-03-11T06:26:08","publication_year":"2019","word_count":782,"keywords":["AI assistants","machine learning","artificial intelligence","AI","chatbots","ML","AI Benefits","AI Strategy","computer vision","NLP","RAG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","computer vision","RAG","AI assistants","chatbots","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-key-benefits-of-using-ai-as-a-long-term-strategy-for-business\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":42043,"title":"Methods That Make Data Exploration Easy in Python: Tips And Tricks For Beginners","content":"For a Data Scientist, Data is the world and exploring it can give insights and help in understanding it better. Data exploration is a critical phase in any data specific problem and is also a skill that every Data Scientist should possess. It profiles the ability and curiosity of the Data Scientist who performs it. Having said that there are numerous ways in which one can understand data. It obviously depends on one’s logic or experience and prior knowledge. Most of us have a tendency to rely on our logic to implement functionality without even bothering to know about an existing library or method which provides the exact same functionality or maybe we already knew and we just forgot that such a library existed. It is possible to forget a library that we don’t use often but can be a game changer when used at the right time. It is nearly impossible to keep track of all those helpful libraries and methods unless we work with them on a regular basis which may not always be possible as each project we work with varies with time. So here in this article, I will put down some of the simplest and most helpful packages or methods that make the  process of exploring and manipulating data easy. Getting The Dataset All the examples shown below uses datasets from the Hackathons at MachineHack. To download the datasets, go ahead and sign up at Machinehack and start a hackathon. Here is a glimpse of the dataset from Predicting The Costs Of Used Cars – Hackathon By Imarticus Learning which we will be using as we go along: import pandas as pd data = pd.read_excel(\"Participants_Data_Used_Cars\/Data_Train.xlsx\") print(data.head()) Data Set Description When we get a data set in hand most of us jump right ahead to explore it by printing the shape of the dataset, printing the columns in the dataset and doing all sorts of such things. However a simple method by pandas shows most of the details in one go. The DataFrame.describe() method describes the specified dataset. data.describe(include ='all') Output: The method can also be applied to individual features (pandas.Series) in the dataset as shown below. data['Transmission'].describe() Output: Describe as category Though the pandas dataframe is well capable of understanding the types of data in each column, sometimes it often confuses itself or maybe we wish to see the data in a different context. Say, for example, the year in the dataset above is a numerical feature, however year as a categorical feature would make more sense to us while describing it. At times like this, we can convert the type of a feature into category or numerical as desired. See the example below. data['Year'].astype('category').describe() Output: Number of observations Per Category This is very useful information given a classification or clustering problem. Identifying if the dataset is balanced or not can help in engineering the features and building an effective model. The value_counts() method will count the number of observations for a specified categorical feature. data['Location'].value_counts() Output: Number Of Missing Data Points Per Feature Knowing the missing data points is a critical piece of information that can help in determining how to impute values. The one-liner code shown below will output the number of missing data points in each column or feature. data.isnull().sum() Output: Merging Dataframes Based on a common feature or column One with an SQL background may be familiar with all the different types of ways two tables can be merged. This feature is extremely useful when handling relational datasets. In pandas merging can be achieved with a simple method called merge(). The below code block merges two dataframes easily based on a common feature on both datasets. d1.merge(d2, how='inner', left_on='id', right_on='ID') Here d1 and d2 are 2 different dataframes with a common feature ‘id’. The above command would merge the d2 dataframes to d1 based on matching id from both the dataframes. Masking Sometimes during data exploration, we might need to pick out a specific piece of data from a large dataset. While database languages like SQL allow querying based on specified conditions, the pandas dataframes also comes with a similar feature called masking. The below line of code picks out only those observations where the Fuel_Type is CNG data[data['Fuel_Type'] == 'CNG'] Output: Transforming The datasets Dataset manipulation is as easy as it can be, thanks to pandas it is possible to apply any method across data in an instant with the apply method. See the example below. A similar method called transform allows us to apply multiple functions or transformations across the features of the dataset. The below code block returns a series after applying the lower() method to all the values in the column Name. The original dataframe is preserved. data['Name'].apply(lambda x: x.lower()) Output: Manipulating A Range of Observations Pandas dataframe allows us to easily manipulate data within a dataframe. The scope of the manipulation can be selected by specifying a range. See the example below. Let’s replace all the values until the 5th row with ‘NaN’ import numpy as np data[0:5] = np.nan Output: Interactive Plotting It’s hard to resist adding some colourful information when all we have are some tables and texts. There are a variety of plotting libraries available, however, when we consider interactive plotting on the go, that can be a hand full on its own. But for those who use notebooks for their works, interactive plots can easily be achieved with a couple of libraries and a simple method called iplot(). import cufflinks as cf import plotly.offline cf.go_offline() cf.set_config_file(offline=False, world_readable=True) data[['Owner_Type' ,'Kilometers_Driven', 'Price']].iplot() Output: The Complete Dataset Profiler The rise in popularity of Python in Machine Learning is mainly because of the abundance of library and support from a huge community. Almost every required functionalities and tool kits are packaged and can be executed just by calling the name of a method. For those who use ipython notebooks, a library called the pandas_profiling single-handedly does most of the Data Exploration task for you. As the name suggests the library profiles any dataframe and generates a complete HTML report on the dataset which includes a lot of information on the dataset and its features. import pandas_profiling data.profile_report() Output: These are some of the useful libraries and methods that make a Data Scientist’s job easier.","excerpt":"For a Data Scientist, Data is the world and exploring it can give insights and help in understanding it better. Data exploration is a critical phase in any data specific problem and is also a skill that every Data Scientist should possess. It profiles the ability and curiosity of the Data Scientist who performs it. […]","categories":["AI Features"],"tags":["Data Analysis","data exploration","exploratory data analysis","interactive plots"],"author_name":"Amal Nair","publish_date":"2019-07-08T10:38:18","publication_year":"2019","word_count":1048,"keywords":["Data Analysis","NumPy","machine learning","Plotly","TPU","AI","ML","data exploration","interactive plots","Python","SQL","exploratory data analysis","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","Pandas","NumPy","Plotly","TPU","Python","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/methods-that-make-data-exploration-easy-in-python-tips-and-tricks-for-beginners\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10083397,"title":"China’s Insatiable Hunger for Espionage and the ‘Thousand Grains of Sand’ Tack","content":"TikTok has everyone swept up in its short-form video wave, but the platform has a dark side. Bytedance, the company behind the viral sensation, has been exposed yet again as being an espionage front for the Chinese government. However, this isn’t news to anyone, as China has made spy work an integral part of its geopolitical policy. While many countries world over have just begun to wisen up to China’s insatiable hunger for sensitive information, India set a precedent by cutting off Chinese surveillance efforts at the beginning of 2020. Many panned the move vehemently, calling it anti-Chinese, however, this pre-emptive move may have helped preserve the privacy of millions of Indian citizens. The Chinese government not only collects data through its ‘private’ companies, but also engages in dangerous economic behaviour targeted at eroding the power of other countries. In addition to its overt information-gathering techniques, the Chinese government also conducts operations to steal intellectual property from technologically advanced nation-states like China. How China spies on you Western cybersecurity experts have been bearing the brunt of Chinese state-sponsored cyberattacks since 2016. A treaty signed between the United States and China during the time of the Obama administration prohibited the theft of intellectual property by either party. However, with Trump’s hardline attitude against China, the totalitarian regime returned to its ways, moving its cyberattack operations to the Ministry of State Security (MSS). The MSS has been at the forefront of China’s cyber espionage activities ever since. Earlier this year, the Indian Ministry of Corporate Affairs carried out a probe into the registrar of companies in India wherein household names like Vivo, Xiaomi, and Oppo were found to be managed directly by Chinese partners. As per the Chinese Internet Security Law, any company operating in the country must share data with the Chinese government. This includes call logs, contact information, and other personally identifiable information like location, IMEI numbers and other sensitive information. Apart from using software to collect information, the Chinese government uses its manufacturing grunt to incorporate espionage-focused backdoors in hardware. By offering products below production cost, Chinese phone manufacturers have penetrated the sub-Rs 15,000 phone market. However, these devices’ firmware and background software routinely seep data back to Chinese servers, even if the phone has no Chinese apps installed. The most disturbing method of Chinese espionage comes from their infrastructural invasions. Through companies like Huawei and ZTE, the Chinese government is able to hijack the network and telecommunications infrastructure of their target nations. Not only do these companies have a cavalier attitude to cybersecurity, their close collaboration with the Chinese government almost ensures the fact that they have built-in backdoors for espionage. The coldest war China has been described as using a ‘thousand grains of sand’ approach to information gathering. Any and all information collected through hardware, software, and other attack vectors is useful. Even when collecting large amounts of data, the Chinese state agencies have the capability to sift through the chaff and create rich datasets to train their algorithms. The Chinese government gains enormous political power from its espionage activities. An investigation revealed that an intelligence collection technology company called Zhenhua was harvesting data on over 10,000 Indians. Some high-profile names who were being monitored included, PM Narendra Modi, President Ram Nath Kovind, Finance Minister Nirmala Sitaraman, Leader of the House Piyush Goyal, at least 15 former Chiefs of the Army and multiple prominent mediapersons. Data of this quality and scale can not only provide a deeper look into the workings of the Indian state, but also set a dangerous precedent for further espionage by China. As far as espionage allegations are concerned, India too doesn’t get a clean chit, as seen in the Pegasus spyware scandal indicting the government. However, one must understand the importance of espionage in statecraft, as it is an indispensable tool to collect information without military aggression. Governments spying on its own citizens and other countries is nothing new, but the scale at which the Chinese government is undertaking espionage activities is alarming, to say the least. In the age of big data analytics and AI powered by huge datasets, China is leveraging its stolen intellectual property and harvested data to create the next generation of spy tools. It is no surprise that the Indian government took pre-emptive regulatory moves to curb the amount of data China can collect from Indian citizens. Till date, Indian authorities have banned over 300 Chinese applications like TikTok, ShareIt, UC Browser, WeChat, and more. In addition to this, they have also initiated a nation-wide block on Chinese websites like Baidu and Alibaba. However, the most important facet of reducing Chinese influence on India lies with the common man, as regulation can only go so far. Chinese Entities: Espionage, profiling and economic control mark Chinese commercial companies – The Economic Times Ministry Of Corporate Affairs News: Ministry of Corporate Affairs searches 300 entities with Chinese nationals on board – The Economic Times","excerpt":"Chinese espionage isn’t restricted to just phones and apps","categories":["AI Features"],"tags":[],"author_name":"Anirudh VK","publish_date":"2022-12-27T12:00:00","publication_year":"2022","word_count":825,"keywords":["big data","Go","programming_languages:R","AI","programming_languages:Go","RAG","GRU","ViT","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","GRU","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chinas-insatiable-hunger-for-espionage-and-the-thousand-grains-of-sand-tack\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101924,"title":"MiQ Certified as Best Firm for Data Scientists for the 2nd Time","content":"MiQ has once again been certified as the Best Firm for Data Scientists to work for by Analytics India Magazine (AIM) through its workplace recognition programme. The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company cultures. AIM analyses the survey data to gauge the employees’ approval ratings and uncover actionable insights. “We are delighted to be selected by AIM as Best Firm for Data Scientists to work for. Data and data science greatly influence everything we do. The impact that data has and will have continues to grow every day. Our focus has been on solving major business challenges that go beyond media campaigns. Thanks to our entire team for making our DS practice one of the best in the industry,” said Ramya Parashar, chief operating officer at MiQ. The analytics industry currently faces a talent crunch, and attracting good employees is one of the most pressing challenges that enterprises are facing. The certification by Analytics India Magazine is considered a gold standard in identifying the best data science workplaces and companies participate in the programme to increase brand awareness and attract talent. Best Firms for Data Scientists is the biggest data science workplace recognition programme in India. To nominate your organisation for the certification, please fill out the form here.","excerpt":"The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company culture.","categories":["AI Highlights"],"tags":["best companies for data scientists in india","data science companies in india","Top Trend"],"author_name":"AIM Media House","publish_date":"2023-10-25T12:01:58","publication_year":"2023","word_count":225,"keywords":["best companies for data scientists in india","data science companies in india","data science","Top Trend","Go","programming_languages:R","AI","programming_languages:Go","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/miq-certified-as-best-firm-for-data-scientists-for-the-2nd-time\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10083640,"title":"Build ChatGPT-like Chatbot Using PaLM","content":"Ever since its release, ChatGPT (built on GPT-3.5 API) has taken the internet by storm and made many developers experiment in building an open-source alternative. Finally, there is a way one can build a ChatGPT-like chatbot using open-source alternative to GPT-3 (175 billion parameters) – i.e. Google’s PaLM (540 billion parameters), alongside reinforcement learning with human feedback (RLHF) built on PyTorch. Check out a work-in-progress chatbot, similar to ChatGPT here. The code is built exclusively for Python language and is implementable for PyTorch. Developers can build their own chatbot by training PaLM like an autoregressive transformer and then training the reward model using human feedback. The main contributor for the code, Phil Wang (lucidrains) has also made alternatives to AlphaFold 2, DALL-E2, and Imagen, with PyTorch implementation and has more than 200 repositories on GitHub. Read: Top 7 ChatGPT Alternatives Another machine learning researcher and expert Yannic Kilcher is also building Open-Assistant, an open-source alternative to ChatGPT on Python — something to look forward to in the coming months. Other ChatGPT-like Chatbots While ChatGPT has been making noise as Google Killer, Google is silently working on building a tangible product with real use cases. Recently, in collaboration with DeepMind, it introduced MultiMedQA (built on PaLM), a ChatGPT-like chatbot for healthcare. Just a few days ago, there was another alternative made for ChatGPT that overcame its biggest limitation of being disconnected from the internet, Chatsonic. Though it works on the same architecture as ChatGPT, GPT-3.5, the ability to connect to the internet allows users to converse with it on real-time events.","excerpt":"The main contributor for the code, Phil Wang has also made alternatives to AlphaFold 2, DALL-E2, and Imagen","categories":["AI News"],"tags":["ChatGPT","GPT-3"],"author_name":"Mohit Pandey","publish_date":"2022-12-29T11:47:03","publication_year":"2022","word_count":260,"keywords":["GPT-3","Go","ChatGPT","machine learning","RLHF","PyTorch","AI","chatbots","Git","Python","R"],"extracted_tech_keywords":["AI","machine learning","ChatGPT","PyTorch","RLHF","chatbots","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/build-chatgpt-like-chatbot-using-palm\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":5571,"title":"Interview &#8211; Vinay Gupta, Head of Business Analytics at Wind World India","content":"[dropcap size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [dropcap size=”2″]VG[\/dropcap]Vinay Gupta: We are in Renewable Energy generation sector, so maximum uptime of Wind turbines and generation of maximum power are the main focus of our operations. Sensor analytics carried out by us is aimed at maximizing the revenue for the customer through enhanced performance, availability and reliability of Wind turbines. In addition, we also aim at reducing the operational costs by continual process improvement. We have an end to end Enterprise Analytics Eco- system viz. Data Warehousing, Analytics Application and Real Time Visualizations. In our organization, we continuously strive to enhance the quality of data being generated\/ captured so as to ensure higher availability of Sensor data, reduce rework on data generated before analysis and build confidence across the users for Decision Making at Strategic, Operational and Tactical Level. A lot of steps have been taken in this regard by re-engineering of communication systems, process improvements and employees training. Further, we spend considerable time in formulating the right question, because then time and efforts spent in analytics would yield the maximum dividend. We firmly believe that data driven facts should drive our major decision making. The recommendations\/inferences obtained from data analytics are discussed, brainstormed and complemented with past experiences and inherent domain knowledge, before the decisions are taken. AIM: What is your approach to face the challenge of meeting needs of so many clients across vast geographies with limited resources? VG: I feel that the solution lies in the basic heart of Analytics process and eco-system itself. To meet the challenge of large number of clients, which in our case includes both internal and external customer, we lay a great emphasis on correct understanding, documenting the need\/problem and prioritizing each need\/requirement. Consequently, we work diligently on the algorithms and methodology of analyzing data for those requirements which have the maximum impact on the business. Once proof tested and outputs confirmed, we automate the analytics end-to-end, ie from the process of data extraction to visualization, and embed this into our current work stream. This ensures timely availability of information, greater value to the user and low overheads on a continuous basis. AIM: What are the key differentiators in your analytical solutions? VG: In energy sector the key data sources are quite different from other popular sectors viz. Retail, Banking, Social Media, Hospitals etc. We have a large network of sensors which are embedded in the wind turbines, which generate huge volume of data in a continuous manner. These data packets are transmitted in real time to our central storage system. Thus there is a need to analyse this data on a 24 x7 real time basis to understand the performance of each wind turbine. Based on certain alerts, decisions need to be taken in a few seconds. In addition, the performance of wind turbines is analyzed based on various systems\/sub systems and material\/spare parts used. Thus the key differentiator is the integration of sensor analytics, engineering analytics and materials analytics under one roof. The models deployed are both descriptive and predictive ones. We need to migrate towards the prescriptive models, as well. AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth. VG: We have a team of around 40 analysts in the Head office and around 30 analysts in seven states in India, where our wind farms are located. We have organized the Business Analytics & Business Excellence Department into four main groups Transactional Data Warehousing team, which includes Turbine Sensor data and ERP data, Data Modeling and Analytical team, which develops new application and analytical models Data Presentation team, which includes visualization, dashboard generation and reports creation for internal and external customers Operational Excellence Team, which takes on business critical projects for end-to-end analysis and solution generation. We ensure that the skills sets of our team are constantly upgraded and in synch with the current technologies and methodologies in the market. AIM: What are the next steps\/road ahead for analytics at your organization? VG: In the immediate future, our focus is mainly to consolidate the initiatives undertaken by us to enhance the analytics capability. The next big step would be to identify the right big data technology and tools, which will meet our business requirement. There is a plethora of technologies, tools, data platforms and data integration software in the market. But making the right choice, which will be a value for money in the long run, is the main challenge. It’s important to have the correct use case for such implementation and one should not do it just because it is in vogue. Another focus area would be to develop applications and expertise in prescriptive analytics and mobile computing. AIM: What are a few things that organizations should be doing with their analytics efforts that most don’t do today? VG: As you are aware that there are two sets of complementary activities in the field of Data Analytics. Firstly, Analytics can be considered either a science or an art and secondly, Decision Making is either intuitive or fact based. I feel that the process of using data analytics in decision making as a competitive advantage is a combination of the both the aspects. Therefore the users at different level need to be aware and proficient in these aspects. Presently, a large number of companies are in the process of collecting\/ capturing huge amount of data, considering that would be required for analysis at the later stage. Very little efforts are being put to ensure the veracity and value of data captured at the transactional stage (or Point of Origin) itself. Emphasis in this direction will not only increase the confidence value in the data sets but also significantly reduce the efforts for analytics\/visualization and enhance the quality of decision making. Another important aspect is to avoid making incorrect assumptions and failing to test them while analyzing. AIM: What are the most significant challenges you face being in the forefront of analytics space? VG: In my opinion, the biggest challenge is to embed the analytics process and insights obtained into the decision making at junior, middle and senior level. Building the culture of data driven approach needs few good success stories. Fortunately, we have been able to make the initial impact in our organization. The second challenge is to ask the right question or generate the right use cases for business functions. Another, significant challenge is to keep updated with the new technologies being developed and evaluate for absorption in the organization. AIM: How did you start your career in analytics? VG: My journey in analytics space began around seven years back in New Delhi, when I started handling huge volume of data related to the multi-disciplinary defence equipment and technical manpower in order to ensure proper planning, employment and resource optimization. Thereafter, I had the privilege of establishing the first Center for Data Analytics and Optimization in Secunderabad. We conceptualized, built and established the complete analytics ecosystem from the scratch. My present role encompasses Sensor based data systems (SCADA), transactional data systems (ERP), Data Integration, Modeling and Application development, Real time analytics and visualization, Customer Portal and Process Improvement projects. We have built various analytical models for optimization of Wind energy operations to drive the business strategy. Modeling includes predictive modeling, O&M revenue modeling plus forecasting, reliability analysis and BI\/BW implementation. We aim to optimize the Wind farm performance through Analysis and Visualization of sensor data (high volume & velocity) of 5500 plus Wind turbines and ERP transactional data on real time basis. AIM: What do you suggest to new graduates aspiring to get into analytics space? VG: My advice to the new grads is that they need to understand both business process and analytics methodology for achieving the positive results. With mushrooming of various analytical applications, the complexity of calculation is hidden, but to correctly interpret the result of analysis and provide solutions for critical business problem is the key to success. Business problem- Analytics Problem- Analytics Solution- Business Solution; this chain needs to correctly linked and implemented for the desired outcome. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? VG: We recruit new graduates as well as experienced individuals, who have an ability to understand the business problem quickly. These professionals should have the passion to play with data, look through the data and generate insights using various statistical modeling techniques, both descriptive and predictive ones. We look for the skills of understanding and applying various data modeling and integration techniques required in engineering analytics and are not driven by a particular Analytics Platform or product. We believe that the platform or application related skills can be acquired with short capsules, hand holding and team support. Finally, it is the attitude towards problem solving and team co-operation, which matters the most. AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? VG: The concept and terms for data analysis to gain insights in business process has evolved from Decision Support System in 1970-80’s to OLAP’s in 1990-2000 to Business Analytics & Big Data in 2010 onwards. There are two main sources of Data Generation viz. Firstly, the customer or social media activist (in BFSI, Retail, Social Media, Telecom, Media, Medical etc) and Secondly, the machine sensor data ( Oil & Gas, Energy, Manufacturing, Automobiles etc). Presently, both these sources generate huge amount of data and analysts are trying to understand either the customer propensity or the machine behavior\/performance. The future lies in the integration, where-in sensors would be embedded on\/inside the human beings for health, fitness, productivity, location or behavior analysis. It would also enable the sensors to analyse our moods and brain waves and send the information on real time basis to health agency or the sales department. While this concept would further impact the privacy of an individual, but it is considered to be potentially revolutionary in terms of both volumes and benefits. As regarding the Engineering Analytics (EA), the study carried out by research firm Zinnov Management Consulting in Aug 2013 has revealed that the market is poised to touch USD 27 billion by 2017. The industries like energy, automobiles, aerospace and healthcare will lead this development. Two years before, in 2012, the spending on engineering analytics, which refers to derivation of meaningful insights by processing information from physical machines, was around USD 12.6 billion. So, in five years this market volume is anticipated to become more than double. AIM: Anything else you wish to add? VG: To achieve ‘Nirvana’ in Analytics, one need to continuously adapt, change and evolve, especially because the data and processing technologies itself are rapidly changing with time and varying circumstances. I feel that even though there are a large number of machines operating with sensors and performing business critical applications, and large amount of data is available from them, Engineering Analytics will still take some more time to rise to the center stage of the Analytics world. Machine behavior and performance analysis could raise the efficiency levels beyond what has been achieved till date, and therefore need to be given due significance by users, engineers, analysts and statisticians of the future. [divider] [spoiler title=”Biography of Vinay Gupta” style=”fancy” icon=”plus-circle”] At Wind World India Limited, Vinay Gupta is Heading the Business Analytics & Business Excellence Division. His focus areas include Predictive Modeling, Reliability Analysis, BI\/BW implementation, SCADA System, SAP implementation, Automation initiatives and Lean Six Sigma Projects. Earlier in his career as an Army officer, he has played a pivotal role in establishing the Centre forData Analytics & Optimization in Army. He has been instrumental in formulating the data management strategies, roadmap and analytical framework for mapping key strategic & operational objectives (KPI’s) of Operations & Maintenance of Military Equipment in Indian Army. He is an M Tech in Electronics & Telecommunications from JNU, New Delhi, Project  Management Professional (PMP) from PMI, USA and Six Sigma Black Belt from ISI, Hyderabad. His five papers have been published in various Defence Journals and he has participated in various national seminars\/conferences on Big Data Analytics and Innovation. His research interests include Sensor Networks, Asset Readiness Analysis and Optimisation, Business Process Improvement, Developing Analytical Model and Managing Data Systems.[\/spoiler]","excerpt":"[dropcap size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [dropcap size=”2″]VG[\/dropcap]Vinay Gupta: We are in Renewable Energy generation sector, so maximum uptime of Wind turbines and generation of maximum power are the main focus of our operations. Sensor analytics carried out by […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"AIM Media House","publish_date":"2014-04-02T15:06:14","publication_year":"2014","word_count":2092,"keywords":["big data","Go","API","TPU","AI","ML","RAG","Aim","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","TPU","R","Go","API","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-vinay-gupta-head-of-business-analytics-at-wind-world-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":32085,"title":"Cloud Outages That Shook The Tech World: 2018","content":"Over the past few years, the popularity of cloud computing has snowballed, and it is boosting the power of the internet more than ever. And with all its benefits that it provides, the cloud is not only becoming a vital part for companies across the globe, but also an essential tool for the future existence of the internet. However, in 2018, the world saw the cloud storage’s dark side. Some unexpected cloud outages have hit even the most prominent cloud providers, causing embarrassment all around. Major Cloud Outages Of 2018 Google Cloud Many products and services nowadays depend on platforms such as AWS and Google Cloud and if their cloud goes down or runs into a problem, it not only affects the companies but also its customer base. And unfortunately, in July 2018, Google Cloud witnessed an outage that caused headaches for some of the popular platforms like Snapchat, Spotify, and Pokémon Go.  According to Google’s status page, the search giant has stated that the outage started at 12:17 on 17 July 2018 and ended at 12:55 on the same day (times are US\/Pacific). Discord, a free VoIP application and digital distribution platform for video gaming communities, said the issue was completely due to Google outage, however, other platforms like Snapchat, Spotify, and Pokémon Go didn’t mention Google. We've identified the cause of our connectivity issues to @GCPcloud which is also impacting other popular apps such as SnapChat. No eta on the recovery time, but awaiting updates on https:\/\/t.co\/2zsKQSmzKF — Discord (@discord) July 17, 2018 https:\/\/twitter.com\/snapchatsupport\/status\/1019306579626094592 Trainers, we're aware of a technical issue causing an outage. Stay tuned for more information, and thanks for your patience as we investigate. — Pokémon GO (@PokemonGoApp) July 17, 2018 Something’s not quite right, and we’re looking into it. Thanks for your reports! — SpotifyCares (@SpotifyCares) July 17, 2018 Even though the issue lasted for almost an hour, Google took quick actions and fixed the problem. “The issue with Google App Engine should be resolved for some users and we expect a full resolution in the near future,” said Google in an update. Amazon Web Services (AWS) On March 2018, Amazon Web Services (AWS) was hit by a cloud outage that silenced Amazon’s Alexa and affected hundreds of enterprise services including Atlassian, Slack, and Twilio.  The outage happened in the important region of data centres in Virginia when the Direct Connect dedicated links from AWS North Virginia region to other server warehouses and premises on the East Coast got disabled. The outage wasn’t the only thing Amazon was dealing with, Twitter was also going crazy with some of the funniest tweets in response to the AWS outage. We can’t publish our story about AWS being down because, well, AWS is down pic.twitter.com\/cwUWEkLBuM — Mashable (@mashable) February 28, 2017 wow this amazon outage is really taking a toll pic.twitter.com\/7efX84789P — JARRY LEE (@jarry) February 28, 2017 That is not all, in May 2018, Amazon was hit by another outage — it witnessed a critical connectivity issue due to some hardware failure in a data centre in North Virginia. And AWS’ EC2, Relational Database Service, Workspaces, and Redshift were all impacted by the outage. The same day, Amazon in an update said, “customers with EC2 instances in the availability zone may see issues with connectivity to the affected instances.” However, the company took the necessary measures to deal with the issue and restored power to the vast majority of the affected instances. Slack Slack was running fine after its May outage that lasted for around 20 minutes — it had a 100% uptime throughout the month of June 2018, however, at the end of the month, the workplace messaging platform suffered a Global outage and the according to the company, the reason was a connectivity issue. “We’ve received word that all workspaces are having troubles connecting to Slack. We’re currently investigating the issue, and will have updates shortly,” Slack confirmed on its website. The investigation continues for our connectivity issues, and we're working hard to get things back to normal. https:\/\/t.co\/uQIDJzyLSV — Slack (@SlackHQ) June 27, 2018 With a massive customer base of organisations and people, Slack is a renowned platform and even a small outage can cause great loss to all its customers. So, when Slack was down, people lost their minds and to respond to the global outage, they took it to Twitter. How am I supposed to tell my team that Slack is down when I don't have Slack to tell them at Slack is down? — Justin Karp (@jskarp) June 27, 2018 Slack is down so I can’t share memes with my friends so have resorted to printing them out and leaving them on their desks. pic.twitter.com\/xhg92anReH — Dave Jewitt (@IrregularDave) June 27, 2018 After all the disruption and Tweets that Slack was dealing with, it resolved the issue and tweeted that all the services had been restored. Folks should be able to connect to Slack again. We're sorry for the disruption. https:\/\/t.co\/uQIDJzyLSV — Slack (@SlackHQ) June 27, 2018 Microsoft Azure Announced in 2008, Microsoft Azure is a popular cloud computing service and over the years, it has gained tremendous popularity. In June 2018, Microsoft Azure suffered a critical outage overnight and it affected the platform’s storage and networking services.  The outage affected the Northern Europe region and the reason behind it was an underlying temperature issue in one of the data centres in the region. According to the company, the outage started at 5.45 PM and lasted till 4.30 AM. However, it seemed that many customers faced issues for a long-time despite Azure Support claiming that engineers had “mitigated the issue and impacted services should be recovered at this time”. https:\/\/twitter.com\/PLJ4330\/status\/1009202847517478914 @azuresupport waiting 24 hours for SQL backups to restore from long term retention. Recent attempts for older backups restored in minutes. Have previous attempts failed or do we just keep waiting? #azTechHelp — SpinnakerSoft (@SpinnakerSoft) June 20, 2018 Witnessing all the complaints, Microsoft quickly did the needful and it was back in normal. But, just after two months, Microsoft suffered another outage caused by a severe lightning storm in the San Antonio. Azure’s South-Central US data centre region was down for quite a while. Customers across the world using Active Directory and Visual Studio Team Services faced trouble for more than 24 hours. Outlook Whether it is due to a human error or natural disaster, when it comes to cloud, expect outage. So, why not be prepared for the downtime.  And when talking about being prepared, it starts with knowing all the pieces of your organisation’s technical chain which includes network, servers, load balancers, applications, DB, other third-party vendors etc. Knowing the chain will not only help you prepare a back-up but also help you prepare it fast.","excerpt":"Over the past few years, the popularity of cloud computing has snowballed, and it is boosting the power of the internet more than ever. And with all its benefits that it provides, the cloud is not only becoming a vital part for companies across the globe, but also an essential tool for the future existence […]","categories":["AI Features"],"tags":[],"author_name":"Harshajit Sarmah","publish_date":"2018-12-24T12:37:48","publication_year":"2018","word_count":1128,"keywords":["Go","GCP","AWS","cloud computing","AI","R","RAG","Aim","SQL","Azure"],"extracted_tech_keywords":["AI","Aim","RAG","cloud computing","AWS","Azure","GCP","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cloud-outages-that-shook-the-tech-world-2018\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":67373,"title":"OpenAI Releases Commercial API That It Earlier Deemed Too Dangerous","content":"One of the prominent AI research labs, OpenAI, has recently launched the beta version of an API for accessing new AI models developed by the company. The API allows searching over documents based on the natural-language meaning of queries rather than keyword matching. With the advancement of machine learning and other emerging technologies, the ultimate goal of this journey is to achieve Artificial General Intelligence (AGI). According to the researchers, the API will serve as a revenue source to help them cover costs in further AI research as well as assist in advancing the technology and making it usable in the real world. Behind The API The API is designed to be simple for anyone to use but also flexible enough to make machine learning teams more productive. While providing any text prompt, the API will yield a text completion and attempts to match the pattern that the user provides. In a blog post, the developers at OpenAI stated that unlike most AI systems which are designed for one use-case, this API affords a general-purpose “text in, text out” interface that allows the users to try it on virtually in any English language task. The API also allows a user to hone the performance on any specific task by training on any small or large dataset of examples that are being provided, or by learning from human feedback provided by the users or labellers. According to the developers, OpenAI’s teams at the research lab are currently using the API so that they can focus on machine learning research rather than distributed systems problems. The API runs models with weights from the GPT-3 family with many speed and throughput improvements. How It Helps The OpenAI API will help in a number of tasks as mentioned below- Semantic Search – The API allows searching over documents based on the natural-language meaning of queries rather than keyword matching. The API identifies relevant content for natural language queries without using any keywords.Chat – The API can enable fast, complex and consistent natural language discussions. With a brief prompt, the API generates dialogues spanning a range of topics, from space travel to history.Customer Service – Leveraging search and chat capabilities, the API generates natural dialogue to quickly give customers relevant information. Through semantic text comprehension, the API can offer a range of analytics and productivity tools to better serve customers.Generation – The API can generate complex and consistent natural language, and enables use cases like creative writing.Productivity Tools – The API allows for parsing text into spreadsheet tables, summarising email discussions, expanding content from bullet points, and more.Content Comprehension – The API can be used to build tools to help individuals consume content more efficiently.Polyglot – The API can be used for tasks such as translation or chat with users in their preferred language. Why Release an API Instead of Open-Sourcing the Models? According to the developers, there are mainly three reasons as to why releasing the API. They are discussed below: – Commercialising technology helps the researchers at OpenAI pay for their ongoing AI research, safety, as well as policy efforts.Secondly, many of the models underlying the API are very large, thereby requiring extensive knowledge and resources to develop and deploy. This makes them very expensive to run, especially for smaller organisations to benefit from the technology.Lastly, the API model allows the researchers to respond to misuse of the technology more easily. The researchers stated, “Since it is hard to predict the downstream use cases of our models, it feels inherently safer to release them via an API and broaden access over time, rather than release an open-source model where access cannot be adjusted if it turns out to have harmful applications.” Wrapping Up With the help of this API, researchers can easily access new AI models that are developed by OpenAI. Currently, the API is available in private beta for free for the first two months, and only a few qualified customers can have access to it including researchers at institutions like the Middlebury Institute as well as companies like Reddit, MessageBird, Koko, among others. Last year, OpenAI released the largest version of text generating AI system, GPT-2 along with code and model weights to facilitate detection of outputs of GPT-2 models in November. This system was initially planned to release in Feb 2019 but the researchers at the lab halted the research due to the fear of misusing this system by the extremist groups. The Center on Terrorism, Extremism, and Counterterrorism (CTEC) also stated that it’s possible to create models that can generate synthetic propaganda for spam and phishing purposes. However, with the release of the system, the AI research lab stated that they have not witnessed any evidence of writing code, documentation, or instances of misuse. Join the private beta here.","excerpt":"One of the prominent AI research labs, OpenAI, has recently launched the beta version of an API for accessing new AI models developed by the company. The API allows searching over documents based on the natural-language meaning of queries rather than keyword matching. With the advancement of machine learning and other emerging technologies, the ultimate […]","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Ambika Choudhury","publish_date":"2020-06-15T18:00:00","publication_year":"2020","word_count":800,"keywords":["Go","semantic search","API","machine learning","TPU","OpenAI","AI","RAG","analytics","R"],"extracted_tech_keywords":["AI","machine learning","analytics","OpenAI","RAG","semantic search","TPU","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-releases-commercial-api-that-it-earlier-deemed-too-dangerous\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":4276,"title":"Interview – Sri Krishnan V, Vice President – Engineering at Robert Bosch Engineering and Business Solutions","content":"[su_dropcap]AIM[\/su_dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [su_dropcap]SK[\/su_dropcap]Sri Krishnan: We are a core Engineering organization with deep Automotive domain knowledge. We have expertise in all the Automotive fields like Engine management, Automotive safety systems like Anitlock Breaking, Electronic Stability, Airbag, Driver assistance systems, Infotainment ,Telematics systems and Automotive  Diagnostics. Leveraging the competencies that we have built over years as the Engineering Offshore Development Centre of Bosch, our approach is to focus on “Engineering Analytics”. We have identified this as a niche area, since we see huge benefits that Engineering analytics can bring in. We are applying analytics in house and with Automotive OEMs for engineering and manufacturing. Though we started with focus only on Engineering Analytics, we find that many use cases require end to end analysis of the value chain. In the context of the technology trend IoT (Internet of Things), where everything will be connected; we envisage more and more end-to-end solutions. Analytics will be part of the solution offering, rather than seen as a separate service. We are building such systems already. AIM: What is your approach to face the challenge of meeting the needs of so many clients across vast geographies with limited resources? SK: We do not see it as a major challenge, since we leverage the presence of Bosch in all geographies for customer engagement and interface. We have collaboration with Bosch Corporate research teams in Germany and USA. We are able to scale up the Analytics team in India supported by right domain skill. AIM: What are the key differentiators in your analytical solutions? SK: About data and algorithms, all problems could be grouped into few classical solutions. It is the domain knowledge that differentiates real insights that add business value. As mentioned, we have profound domain knowledge which makes the difference. For example JD Power Initial Quality Study 2013 report says that majority of problems experienced by owners with their new vehicle in the first 90 days of ownership are design-related rather than manufacturing defects. We could validate this with social media analytics and connect with engineering design. We are able to narrow down customer sentiments to feature and function level. We are able to trace field quality issues to a specific slice in the product life cycle. We see analytics as part of a system solution. We have in house expertise and develop complete products with Electrical hardware, mechanical engineering teams and manufacturing.  We have IT and ITES teams who provide elegant solutions for Engineering and manufacturing organizations. We are able to offer end-to-end integrated solutions from sensor network, M2M systems to aggregate, Data collection, validation and Integrated Analytics with existing IT systems. This is a unique combination compared to typical Analytics solution providers, whose heritage is IT systems. AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth. SK: The core Analytics team consists of about 40 people, and additionally solutions team as a matrix team consists of about 100 people. We cover all the major verticals. AIM: What is the next steps\/ road ahead for analytics at your organizations? SK: We plan to expand further in to other domains, for example Healthcare and Energy. Already we have acquired domain experts on these topics and started with few use cases. We will continue with automotive sector and reach out to more customers. We see Analytics as an integral part of any offering that we make in future. AIM: What are a few things that organizations should be doing with their analytics efforts that most don’t do today? SK: The importance of domain expertise is undermined by many Analytics alone teams.  We find that many of the customers do not appreciate the value of analytics, without a concrete use case that makes real business sense to them.  This holds good even for in house solutions. To understand the customer problems and make value proposition, Organizations should focus on building up analytics teams backed with domain knowledge. AIM: What are the most significant challenges you face being in the forefront of analytics space? SK: One of the major challenges is to convince middle and lower management that Analytics helps them. Many of them are either skeptical due to deep conviction of their own established spreadsheet based methods or they see a threat of being seen as not effective, in case big improvements happen as a result of analytics. Coming out with good use cases is a team work. It involves collaboration between the analytics team, domain experts, actual users and management. Even if one of the stakeholders is not engaged, it could affect the speed or the extent of potential benefit itself. AIM: How did you establish Analytics Business area? SK: We had seen the technology megatrend of Analytics picking up momentum. We could clearly see the benefits that Analytics can bring to large MNC like Bosch and also what we could offer to others. Being an offshore Engineering centre of Bosch, located in India we have the benefit of leveraging on the in house domain expertise and eco system to scale up. We started a small group with the culture of “Startup” within our offshore centre.  To get the right level of focus and attention, the team directly reports to me, one of the Business Unit Heads. Significant investments have gone in the last two years to build up a competent team, infrastructure and proof of concepts. We are able to see the results in terms of real business benefits. AIM: What do you suggest to new graduates aspiring to get into analytics space? SK: Do not jump in to the bandwagon just because reports show that there are so many thousands of jobs in the analytics space. Understand the field; check if you are really passionate about it.  Develop practical experience in any of the domains first along with specializing in Analytics. Without such understanding, pure analytics knowledge will be a limitation. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? SK: We recruit at various levels from relatively newbie to expert. We look at the level of understanding and system knowledge on the problems that they have worked. We evaluate the candidate’s ability to comprehend and get an overview of any new problem and look for Analytics oriented solutions. The approach to the problem and ability to think of alternatives is vital. AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? SK: The industry is evolving. Lot of interesting use cases, visualizations and business benefits are published. There is lot of expectation on Big data along with IoT (Internet of Things). Open source adoption is gaining momentum. But strategy roadmap and systematic adoption in large organizations is still lacking. Overall it is an opportunity space. AIM: Anything else you wish to add? SK: As a country, India has got great minds. We are well known for our inclination to Mathematics and specially statistics. But in the global scenario, our thought leadership is not visible enough. If we take Analytics space, it would be worth pondering how many papers from India get published in reputed international journals and get citations. We cannot be satisfied with the volume of the workforce and revenue generated, but value created in advancement of technology. [su_spoiler title=”Profile of Mr Sri Krishnan” style=”fancy” icon=”plus-circle”] Mr Sri Krishnan V is the Vice President – Engineering of Robert Bosch Engineering and Business Solutions Limited (RBEI) with an extensive track record of having worked in core engineering organizations developing embedded software, hardware and products. A seasoned professional in the aerospace and automotive sectors, Mr Krishnan started his career at ISRO in the satellite centre working on satellite telemetry and payload data handling systems of Indian Remote sensing satellites program. Following ISRO, he moved to Aerospace Systems Ltd where he played a key role in setting up navigation systems for defence applications. Mr Krishnan joined Bosch in 1999 in the Car Multimedia division responsible for product development for the Indian market. He also worked in USA and Germany for 4 years where he was in charge of coordination of software development, customer liaison and leading software process improvement initiative. In his current role as Vice President engineering, he has been instrumental in planning the growth for Automotive Safety (Airbag, ABS, ESP and Driver Assistance systems), Car Multimedia (Navigation systems, Head Units, Radios), Engineering tools and Automotive Base software. With a degree in Electronics and Communication Engineering from PSG College of Technology, he is currently the Chairperson of NASSCOM’s Automotive Special Interest Group. Mr Krishnan is married and has 2 children. [\/su_spoiler]","excerpt":"[su_dropcap]AIM[\/su_dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [su_dropcap]SK[\/su_dropcap]Sri Krishnan: We are a core Engineering organization with deep Automotive domain knowledge. We have expertise in all the Automotive fields like Engine management, Automotive safety systems like Anitlock Breaking, Electronic Stability, Airbag, Driver […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"AIM Media House","publish_date":"2013-11-13T17:36:07","publication_year":"2013","word_count":1485,"keywords":["big data","Go","AI","RAG","BERT","Aim","ViT","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","big data","BERT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-sri-krishnan-v-vice-president-engineering-at-robert-bosch-engineering-and-business-solutions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10006092,"title":"American Express Establishes Data Analytics, Risk &#038; Technology Lab (DART) In IIT Madras","content":"In a recent announcement, American Express announced the establishment of a Data Analytics, Risk and Technology (DART) Laboratory at the Indian Institute of Technology Madras. The lab envisions to establish itself as a world-class hub of research in areas such as risk analytics and behavioral sciences by adopting AI, ML and related technologies. The company stated that the lab would focus on risks that originate from human behavior and decisions, and develop science-based understanding of humans’ strengths and weaknesses. It will develop technologies to measure the cognition and attention of the decision-makers and prevent potential accidents in high-risk industries. The company hopes to apply these technologies across dimensions such as employee engagement and attention, evaluating and enhancing the quality of education and learning in school. The Lab at IIT Madras will explore a range of verticals with key emphasis on manufacturing, finance, healthcare, operations management and smart cities. “Our collaboration with IIT Madras reiterates our commitment to support and invest in interventions for public good in the country. The technologies and applied sciences R&D in the Lab will be beneficial for creating an overall societal impact through advancement in financial services, healthcare and safety standards,” said Bharathram Thothadri, EVP and Chief Credit Officer, American Express. It also plans to build talent for industry by partnering with academia while promoting talent and diversity in technology. It has also announced annual scholarships for economically-disadvantaged and meritorious students, including ‘Ambition Awards’ for deserving women students at IIT Madras. “We are proud to partner with the country’s premier academic institution to invest in the future of cutting-edge data science and analytics research. We are also thrilled to announce Ambition scholarships and awards for meritorious women students that underlines our focus on fostering diverse talent,” said Manoj Adlakha, Senior Vice President and CEO, American Express Banking Corp., India. The scholarships being awarded as part of the grant to IIT Madras from American Express include: American Express Ambition Scholarship for Women in Science, Engineering and Technology – tuition fees scholarship for top female student from disadvantaged economic backgroundAmerican Express Ambition Award for Women for Excellence in Science, Engineering and Technology in Bachelor’s program to top female student in the graduating batch of B.Tech\/B.E CoursesAmerican Express Ambition Award for Women for Excellence in Science, Engineering and Technology in Master’s program to top female student in the graduating batch of Masters\/M.Tech\/MSC\/M.S courseAmerican Express Emerging Leadership Award for Excellence in Data Science (Pre-Final) to top student in the Data Science course in the pre-final yearAmerican Express Leadership Award for Excellence in Data Science (Final) to top student in the Data Science Course in the final year “The scholarships and awards will encourage our women students to excel in STEM and pursue enriching careers at the post-graduate level. We look forward to continued support from American Express for growing these efforts in the coming years,” said Prof. Bhaskar Ramamurthi, Director, IIT Madras.","excerpt":"In a recent announcement, American Express announced the establishment of a Data Analytics, Risk and Technology (DART) Laboratory at the Indian Institute of Technology Madras. The lab envisions to establish itself as a world-class hub of research in areas such as risk analytics and behavioral sciences by adopting AI, ML and related technologies. The company […]","categories":["AI News"],"tags":["back office data","data analytics masters","MSc data science","msc data science and analytics","MSc. in data science"],"author_name":"Srishti Deoras","publish_date":"2020-09-04T15:05:28","publication_year":"2020","word_count":482,"keywords":["MSc. in data science","data science","data analytics masters","Go","programming_languages:R","AI","ML","programming_languages:Go","RAG","msc data science and analytics","analytics","MSc data science","back office data","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/american-express-establishes-data-analytics-risk-technology-lab-dart-in-iit-madras\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10111581,"title":"Eagle 7B RNN Model Surpasses Transformers in Performance","content":"The open-source community has introduced Eagle 7B, a new RNN model, built on the RWKV-v5 architecture. This new model has been trained on 1.1 trillion tokens and supports over 100 languages. The RWKV architecture, short for ‘Rotary Weighted Key-Value,’ is a type of architecture used in the field of artificial intelligence, particularly in natural language processing (NLP) and is a variation of the Recurrent Neural Network (RNN) architecture. Introducing Eagle-7BBased on the RWKV-v5 architecture, bringing into opensource space, the strongest– multi-lingual model (beating even mistral)– attention-free transformer today (10-100x+ lower inference)With comparable English performance with the best 1T 7B models pic.twitter.com\/hWtEMC1264— RWKV (@RWKV_AI) January 29, 2024 Eagle 7B promises lower inference cost and stands out as a leading 7B model in terms of environmental efficiency and language versatility. The model, with its 7.52 billion parameters, shows excellent performance in multi-lingual benchmarks, setting a new standard in its category. It competes closely with larger models in English language evaluations and is distinctive as an “Attention-Free Transformer,” though it requires additional tuning for specific uses. This model is accessible under the Apache 2.0 license and can be downloaded from HuggingFace for both personal and commercial purposes. In terms of multilingual performance, Eagle 7B has claimed to have achieved notable results in benchmarks covering 23 languages. Its English performance has also seen significant advancements, outperforming its predecessor, RWKV v4, and competing with top-tier models. Working towards a more scalable architecture and use of data efficiently, Eagle 7B is a more inclusive AI technology, supporting a broader range of languages. This model challenges the prevailing dominance of transformer models by demonstrating the capabilities of RNNs like RWKV in achieving superior performance when trained on comparable data volumes. In the RWKV model, the rotary mechanism transforms the input data in a way that helps the model better understand the position or or order of elements in a sequence. The weighted key value also makes the model efficient by retrieving the stored information from previous elements in a sequence. However, questions remain about the scalability of RWKV compared to transformers, although there is optimism regarding its potential. The team plans to include additional training, an in-depth paper on Eagle 7B, and the development of a 2T model.","excerpt":"For the first time RNN models outperformed Transformer counterparts.","categories":["AI News"],"tags":["Generative Pre-Trained Transformer"],"author_name":"K L Krithika","publish_date":"2024-01-29T18:21:06","publication_year":"2024","word_count":371,"keywords":["Go","artificial intelligence","AI","neural network","Transformers","Scala","NLP","Aim","Generative Pre-Trained Transformer","RNN","R"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","NLP","Aim","Transformers","R","Go","Scala","RNN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/eagle-7b-rnn-model-surpasses-transformers-in-performance\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10041788,"title":"MIT Researchers Make New Chips That Work On Light","content":"“The new architecture could be used in convolutional neural networks and recurrent neural networks.” As the ability of AI systems improve, they will likely require even more processing capacity. This will probably be beyond ordinary computing technology. MIT spinout Lightelligence is working on next-generation computing hardware to overcome this challenge. In contrast to standard electronic architectures, the optical chips that Lightelligence creates offer improvement in terms of high speed, low power consumption and low latency. In 2017, co-founder and CEO of Lightelligence Yichen Shen, in his paper, “Deep learning with coherent nanophotonic circuits”, had presented a new neural network architecture, based on unique advantages of optics, that offered two-orders-of-magnitude speed boost and three-orders-of-magnitude power efficiency gain over cutting edge models for learning tasks. It was demonstrated using a programmable nanophotonic processor. According to the paper, the main principle behind Artificial Neural Networks is based on the computational network models found in the nervous system. The proposed architecture could be used to perform matrix multiplications and nonlinear activations in various artificial neural network techniques, such as convolutional neural networks and recurrent neural networks. Instead of using the fabrication platform typically used for regular semiconductor chips, Lightelligence uses it in a revolutionary way. Lightelligence designs light-powered computing components, which could be the hardware required to power the AI revolution. Lightelligence’s optical processors have orders of magnitude better performance than standard architectures. To conduct arithmetical calculations electronic circuits need to integrate tens, perhaps hundreds, of logical gates. To carry out this process the electronic chip transistors must be turned on and off for several clock cycles. Each time a logical port is switched on, heat is generated, and power is consumed.This is not the case with Lightelligence chips. Shen’s optic computing chips allow for a significantly lower power usage than their electron-powered counterparts, thereby generating very little heat. Furthermore, their ONN can potentially provide direct training of the network on the photonic chip with the higher forward propagation speed and power efficiency by employing only propagation. While lightelligence’s CEO doesn’t intend to replace the electronic computing industry in its entirety. He rather aims to accelerate certain linear algebra operations to perform quick, power-efficient tasks like those found in artificial neural networks. Lightelligence’s competitive advantage (Source : Lightelligence) In this emerging field of optical computing, Shen and his colleagues are not alone. Only recently, another MIT spinout Lightmatter announced an additional round of $80 million in Series B fundraising. The technology of Lightmatter is built on patented silicon photonics technology that handles consistent light within a chip to conduct calculations very fast with extremely low power. Lightmatter’s CEO Nick Harris is one of the contributing authors of “Deep learning with coherent nanophotonic circuits”. But unlike their competition, Shen and his colleagues have crucial advantages. In 2017, Shen teamed up with Soljajic and two other MIT alumni to start Lightelligence. Lightelligence’s vice president of photonics, Dr Huaiyu Meng, holds a doctorate in electrical engineering. Adding to the founding team is business administration major Spencer Powers. They not only invented the technology at the institute, but they are also the first firm to have developed a complete optical hardware solution. Irrespective of competition, Shen is confident in Lightelligence’s innovation potential. To date, Lightelligence has raised over $40 million and currently, the team is working on building the world’s most extensive integrated photonic system. Since data centres such as Amazon and Microsoft play a major role in AI computing, which takes place in the cloud. Centres that run computationally intensive AI algorithms that take up a huge chunk of data centre capacity. Every year millions of dollars’ worth of electricity is burned by thousands of servers running continuously. On the other hand, Lightelligence servers consume far less power, at a significantly lower cost. Their AI chips not only reduce the cost but also significantly increase the computational capability making Lightelligence a lucrative startup.","excerpt":"“The new architecture could be used in convolutional neural networks and recurrent neural networks.” As the ability of AI systems improve, they will likely require even more processing capacity. This will probably be beyond ordinary computing technology. MIT spinout Lightelligence is working on next-generation computing hardware to overcome this challenge. In contrast to standard electronic […]","categories":["AI Features"],"tags":["Deep Learning","MIT"],"author_name":"Ritika Sagar","publish_date":"2021-06-14T17:00:00","publication_year":"2021","word_count":646,"keywords":["Go","programming_languages:R","AI","neural network","MIT","RPA","innovation","Aim","deep learning","Deep Learning","R","startup"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","R","Go","RPA","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mit-researchers-make-new-chips-that-work-on-light\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10082298,"title":"Big Tech Loves Space, But Not Enough","content":"The space race is on, and the tech industry can get quite picky! For instance, Microsoft, Amazon, and IBM have been investing, quintessentially catering to the spacetech companies, both private and public. Google, on the other hand, is contributing, albeit sparsely. Meta and Apple, are barely sitting out on spacetech – mostly art and entertainment. In 2017, SpaceX announced that it is partnering with Google to launch the biggest space project ever, which was later revealed in 2021 to be providing Starlink ground stations at Google’s data centres. So far, nothing big has been revealed yet. Besides Google Sky, a celestial map that shows objects like stars, constellations, galaxies, planets, or the Earth’s Moon, and Rubin Observatory that uses Google Cloud, the tech giant has not made major breakthroughs in the spacetech landscape. On the other hand, Microsoft has been partnering with NASA for lots of projects since the beginning of space exploration. In 2020, Microsoft made a deal with SpaceX to connect their cloud computing network through Starlink satellite. Then came Google to make the deal with SpaceX for the same purposes. So far, nothing substantial has been announced. Microsoft went on to launch Azure Space platform, providing infrastructure for space companies in the cloud. Recently, Microsoft announced that it is democratising the space development industry by offering infrastructure-as-a-service through Azure Orbital Space SDK for development, testing, and deploying space hardware. IBM started partnering with NASA more than five decades ago for space missions. Since then, it has been developing technologies and providing infrastructure for space companies apart from NASA. In 2021, IBM collaborated with HPE Spaceborne Computer-2 and ISS National Lab to provide edge computing solutions in space. In 2022, IBM partnered with Sierra Space for building a space infrastructure for commercialisation of low-earth orbit (LEO), furthering cloud technology in space. Recently, Amazon took huge steps and successfully tested\/ ran a suite of AWS compute and machine learning (ML) software on a LEO, in a first-of-its-kind space experiment. Launched in 2020, the AWS in Space business segment, is aimed to further build and provide cloud services for spacetech innovations by connecting with the AWS Ground Station along with offering services for building satellites and conducting space and launch operations. So far, (not) so good NASA and ISRO have been the long standing giants of the space industry. But they have also partnered with IBM and Amazon for their data and infrastructure services since the beginning. After Elon Musk created SpaceX and paved the way for private companies to take part in the space race, companies like Microsoft, Google, along with other startups, started making breakthroughs. In 2020, Google partnered with NASA’s Frontier Development Lab for upscaling low resolution images using AI. Google has also proposed another project to enable navigation on the moon’s surface without GPS. Looks like Google has been sitting ducks at an enterprise level and not focusing heavily on infrastructure for space missions. It mostly looks to support the mission once it has been successfully deployed – aka ‘post launch assistance’. In November, during the launch of NASA’s Artemis, the videos of the take-off were simulcast in ultra-high-definition through Meta Horizon Worlds venues, which is a streaming product of Meta that supports 360-degree streaming throughout the world. Though Meta has been making the metaverse look more and more possible, getting into spacetech is still a long way to go as they still do not have their own cloud services. Meanwhile, Apple had made no big bids in the space or even the cloud infrastructure industry until September, when it announced its investment plan of $450 million in satellite infrastructure to compete with Starlink. It also plans to invest $50 billion till 2026. Earlier, the Mac Observatory has been using the Apple Mac for astrophotography. Clearly, Google is ahead of Meta and Apple in the field, but they can do much better. Need of the hour Google, Apple, and Meta currently stand on top of innovations in the tech industry. By investing in IaaS (infrastructure as a service) for spacetech, they can contribute to the development of new space technologies. Many of them that are being developed for space exploration, including satellite-based internet and space-based cloud infrastructures, require advanced platforms and data centres to function, much like how SpaceX is doing with Starlink. Meanwhile, in India Globally, there are over 5,500 spacetech companies and 368 of them are in India. Moreover, there are an increasing number of data centres in India with Microsoft, IBM, AWS, and Adani Groups planning to build, facilitating them for space tech companies remains unaddressed. The Indian space-tech companies like Skyroot Aerospace, Bellatrix Aerospace, and Agnikul are proving to be successful in space missions and having infrastructures within the country can make India ahead in the space race. [Updated] December 19, 2022, 11:00AM | The story has been updated to reflect Amazon’s latest developments in spacetech ‘The space industry is capital intensive and requires operators to make large up-front investments with long times to break even. Only large – typically government-funded – organisations could build, launch, and operate space infrastructure. IaaS solutions fill a critical gap between the satellite operators and downstream service providers and enable access with much lower up-front investments. While the ecosystem of service providers is bound to grow, operators will also see a faster ROI enabling smaller organisations to also step into the business. With the availability of robust state-of-the-art software tools that help reduce design time, the availability of more cost-effective launch vehicles, and the available talent-pool in India available from a successful multi-decade space program, we should see a strong ecosystem of operators and service providers develop and pave the way for many new applications and products.’","excerpt":"Microsoft, AWS, and IBM have always been in the space industry, Google is making large bids, Apple is taking the first steps, and Meta is still too far behind","categories":["Global Tech"],"tags":["Google","ISRO","NASA"],"author_name":"Mohit Pandey","publish_date":"2022-12-14T16:00:00","publication_year":"2022","word_count":951,"keywords":["Go","ISRO","machine learning","NASA","AWS","AI","cloud computing","R","ML","Aim","Google","edge computing","Azure"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","cloud computing","AWS","Azure","edge computing","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/big-tech-loves-space-but-not-enough\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":26285,"title":"India &#038; South Korea To collaborate On AI, To Establish Research and Innovation Centre","content":"Prime Minister Narendra Modi and President of the Republic of Korea, Moon Jae-in at Gandhi Smriti in New Delhi. (Image credit: @PIB_India\/Twitter) India and South Korea on Monday signed five memorandums of understanding (MoUs) in the field of new tech. Union Minister for Science and Technology, Dr Harsh Vardhan and his South Korean counterpart, You Young Min signed three MoUs for: Programme of Cooperation 2018-21 Establishment of Future Strategy Group Cooperation in biotechnology and bio-economy According to a statement released by the Press Information Bureau, Government of India and the Ministry of Science and Technology, the top bureaucrats also decided to establish an Indo-Korean Center for Research and Innovation (IKCRI) in India. It will act as the hub for systematic operation and management of all cooperative programmes in research and innovation between the two countries including innovation, entrepreneurship and technology transfer. The PIB also announced the plan to establish additional India-Korea Joint Network Centres in areas for Artificial intelligence Internet of Things focused on agriculture, energy, water and transportation Semiconductor electronics India & Korea have agreed to establish a #FutureStrategicGroup to build collaborative platform to utilise the potential of the two countries to foster innovation and create impact for social & economic good. @PMOIndia, @PIB_India, @IndiaDST, @DBTIndia, CSIR_IND pic.twitter.com\/18ljiuIGfv — Dr Harsh Vardhan (@drharshvardhan) July 9, 2018 These centres will leverage existing infrastructure and funding available from the partners on both sides in focused applied research areas, which have potential towards technology development. According to a news report, the two countries will collaborate in AI and build partnerships between Indian IT companies and Korean engineering and hardware companies. Two other MoUs signed between Council of Scientific and Industrial Research (CSIR), and South Korean National Research Council for Science and Technology and IIT Mumbai and Korea Institute of Science &Technology, to further accelerate future-oriented cooperation in their respective sectors. The MoUs were signed at the conclusion of the fourth India-Korea Science and Technology Ministers Steering Committee Meeting. These talks were held on the backdrop of Prime Minister Narendra Modi and President of Republic of Korea Moon Jae-In inaugurating the world’s largest mobile phone factory in Noida. South Korean mobile phone giant Samsung has launched the factory which has brought in an investment of about ₹5,000 crore. The importance of digital technology is increasing all over India. At the programme in Noida spoke about how technology is bringing convenience and enhancing transparency. Also talked about the efforts to popularise digital payments across India. — Narendra Modi (@narendramodi) July 9, 2018 The facility, located in the Uttar Pradesh city near Delhi will have a capacity of fabricating 120 million phones a year. The factory will build 10 million phones a month, 70 percent of which will be set aside for usage and sales in India.","excerpt":"India and South Korea on Monday signed five memorandums of understanding (MoUs) in the field of new tech. Union Minister for Science and Technology, Dr Harsh Vardhan and his South Korean counterpart, You Young Min signed three MoUs for: Programme of Cooperation 2018-21 Establishment of Future Strategy Group Cooperation in biotechnology and bio-economy According to […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Narendra Modi","Samsung","south korea"],"author_name":"Prajakta Hebbar","publish_date":"2018-07-10T05:54:36","publication_year":"2018","word_count":462,"keywords":["Go","funding","Samsung","artificial intelligence","AI","RPA","innovation","Git","RAG","Narendra Modi","GAN","R","AI (Artificial Intelligence)","south korea"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","R","Go","Git","GAN","RPA","innovation","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/india-south%e2%80%89korea-artificial-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167586,"title":"WordPress Launches Free AI Website Builder","content":"WordPress.com, on Wednesday, released a new AI website builder to simplify the process of creating a website. The company, in a blog post, pointed out that the process is as easy as having a conversation. The AI website builder takes your input and creates a fully designed WordPress website, including all the essential elements like text, layouts, and images. Users simply need to share the idea using prompts in a chatbot-like interface, and AI handles the rest. The AI-powered website builder is aimed at entrepreneurs, small business owners, freelancers, creators, bloggers, and developers who want to quickly build their online presence to showcase their products and services. WordPress.com mentioned that it is not tailored to create e-commerce websites or sites that require complex integrations, but it could do that in the future. To start creating a website, the user must head to the AI website builder page, log in to their WordPress.com account, type in the prompt, and wait for the AI to build the website. One can customise manually or by following up with new prompts in the chat box. Once done, users can pick a WordPress.com hosting plan to make it accessible to people across the globe. The company clarified that the AI feature is only available for brand-new WordPress.com websites, and users get 30 free prompts. If you purchase a hosting plan, you can send unlimited prompts to the builder. Recently, to assist small businesses in utilising AI for web development, GoDaddy introduced GoDaddy Airo, an AI-driven website builder to streamline the website creation process. Regardless of whether you are a seasoned designer or a total novice, Airo allows users to build professional websites with ease. With solutions like these, anyone can now create websites in minutes.","excerpt":"The AI website builder takes your input and creates a fully designed WordPress website.","categories":["AI News"],"tags":["wordpress"],"author_name":"Ankush Das","publish_date":"2025-04-10T14:28:40","publication_year":"2025","word_count":290,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","Aim","wordpress","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wordpress-launches-free-ai-website-builder\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044955,"title":"Complete Guide To VIT-AugReg: A PyTorch Image Model Descriptive Predictions","content":"In today’s world, we live in a generation where we tend to generate and produce vast amounts of data every day. It won’t be wrong to describe these times as an era of big data, where all areas of science and industry thrive upon masses of data and its related technologies. Although this also confronts us with unprecedented challenges regarding their analysis and interpretation. This is the sole reason that there seems to be an urgent need for novel and self-aware machine learning and artificial intelligence methods that can help in utilizing the data properly. Deep learning (DL) is a method that is currently receiving much attention. DL can be described as a family of learning algorithms that can be used to make our systems learn complex prediction models. Deep learning is and has been successfully applied to several application problems. Deep learning models present us with a new learning paradigm in artificial intelligence (AI) and Machine Learning (ML). The recent breakthrough results gained in image analysis and speech recognition have also aided in generating massive interest in this field, as applications in many other domains inculcating big data seem possible. The mathematical and computational methodology underlying deep learning models is challenging, especially for interdisciplinary scientists. These models form the major core architectures of deep learning models currently being used and should always belong in a data scientist’s toolbox. The core architectural building blocks can be composed flexibly to build new application-specific network architectures. Data analysis varies from company to company depending upon the needs, so the data model must always be designed to meet the requirements. Predictive modeling is a major subpart of data analytics that uses data mining and probabilistic methods to predict results. Each model is built using many predictors that make them highly favourable to determine future decisions. Once the data is received for a specific prediction type, an analytical model is formulated. Then, simple linear equations or a complex neural structure can be further applied and are outlined by a concerned software. If in case, additional data is available, then the analytical model is revised. Predictive Modeling also uses different regression algorithms and analytics or statistics to estimate the probability of an event, applying detection theory and largely employed in the fields related to Artificial Intelligence (AI). PyTorch is an optimized tensor library primarily used for Deep Learning applications that use GPUs and CPUs to enhance the processing power. It is an open-source machine learning library for Python, developed by the Facebook AI Research team and is one of the widely used Machine learning libraries, others being TensorFlow and Keras. What Are Image Transformers? Image Transformer is a model-dependent entirely on the self-attention mechanism, where the encoder generates a per-pixel-channel representation of the source image. Despite comparatively low resources required for training, the Image Transformer models are usually trained on images from the standard ImageNet data set. Many applications of image models require conditioning on additional information of various kinds: from images in enhancement or reconstruction tasks such as superresolution, in-painting and denoising to text when synthesizing images from natural language descriptions. In visualization tasks, image generation models using Transformers can predict future frames of video based upon the previous frames and taken actions. Image Transformers treat pixel intensities as either discrete categories or ordinal values, where the setting is subjective and depends on the distribution of image data. For both the image encoder and decoder during image preprocessing, the Image Transformer uses multiple stacks of self-attention and position-wise feed-forward layers. The decoder uses an attention mechanism to take the encoder representation as an input. For parametric and conditional image preprocessing using the Image Transformer, a decoder only configuration is used. Each self-attention layer computes a D dimensional representation for each position, that is, each channel of each image pixel. To recompute the representation for a given position, it first compares the position’s current representation to other positions’ representations, obtaining an attention distribution over the other positions. This distribution is then used to judge the contribution of the other positions’ representations to the next representation for the position at hand. Vision Transformers (ViT) have been shown to attain highly competitive performance for a wide range of computer vision and image-based applications, such as image classification, object detection and semantic image segmentation. Vision Transformer is generally found to have an increased reliance on model regularization or data augmentation, also known as “AugReg”, for short when training on smaller training datasets. Image Source: Original Paper Getting Started with the Code for VIT-AugReg This article will try to generate a descriptive prediction from an image dataset using the VIT library. We will try to predict the dog breed in the image and provide it with a label using Vision Transformer. The following code is inspired by the library’s creators, whose Github link can be accessed here. Installing the Library To create this prediction model, we will first install the Vision Transformer. The following code can be used to do so, # Install the vision_transformer Library. ![ -d vision_transformer ] || git clone --depth=1 https:\/\/github.com\/google-research\/vision_transformer Install the further dependencies, # Install dependencies. !pip install -qr vision_transformer\/vit_jax\/requirements.txt |████████████████████████████████| 57 kB 2.8 MB\/s |████████████████████████████████| 76 kB 4.5 MB\/s |████████████████████████████████| 179 kB 24.6 MB\/s |████████████████████████████████| 88 kB 7.8 MB\/s |████████████████████████████████| 168.3 MB 15 kB\/s Importing the AugReg Model, import sys if '.\/vision_transformer' not in sys.path: sys.path.append('.\/vision_transformer') %load_ext autoreload %autoreload 2 from vit_jax import checkpoint from vit_jax import models from vit_jax import train from vit_jax.configs import augreg as augreg_config from vit_jax.configs import models as models_config Importing dependencies for analysis, import glob import os import random import shutil import time from absl import logging import pandas as pd import seaborn as sns import tensorflow as tf import tensorflow_datasets as tfds from matplotlib import pyplot as plt pd.options.display.max_colwidth = None logging.set_verbosity(logging.INFO) Loading the Data Now we will start loading our data table from the VIT Cloud. # Load master table from Cloud. with tf.io.gfile.GFile('gs:\/\/vit_models\/augreg\/index.csv') as f: df = pd.read_csv(f) # List the rows and columns print(f'loaded {len(df):,} rows') df.columns #print length of dataset len(set(df.filename)), len(set(df.adapt_filename)) # loading the dataset checkpoint best_filenames = set( df.query('ds==\"i21k\"') .groupby('name') .apply(lambda df: df.sort_values('final_val').iloc[-1]) .filename ) # Fine Tuning these models. best_df = df.loc[df.filename.apply(lambda filename: filename in best_filenames)] Now that all the essential model and model data checkpoints are loaded, we can create the predictor model. Creating The Predictor Model We will start creating the predictor model by loading our pet image dataset first. The following code can be used to do so, #loading the image dataset ds, ds_info = tfds.load(tfds_name, with_info=True) ds_info # Get model instance model = models.VisionTransformer( num_classes=ds_info.features['label'].num_classes, **model_config) Now we will import a single random image from the pets dataset perform our prediction on, d = next(iter(ds['test'])) #display random image def pp(img, sz): img = tf.cast(img, float) \/ 255.0 img = tf.image.resize(img, [sz, sz]) return img plt.imshow(pp(d['image'], resolution)); Output: # Applying the VIT-AugReg model on image logits, = model.apply({'params': params}, [pp(d['image'], resolution)], train=False) # Plotting the label probabilities. plt.figure(figsize=(10, 4)) plt.bar(list(map(ds_info.features['label'].int2str, range(len(logits)))), logits) plt.xticks(rotation=90); Output : As we can see, the model has predicted the dog breed to be Leonberger. So let’s now compare the result with an image of the Leonberger dog. Image Source As we can observe, our predictor model seems to have correctly predicted the dog breed label for our sample image! The created Vision Transformer can also be incorporated into other Pytorch Image models as well. Let’s try it with the Timm Model, # Installing the timm model library !pip install timm import timm import torch # Loading the model into timm timm_model = timm.create_model( 'vit_small_r26_s32_384', num_classes=ds_info.features['label'].num_classes) if not tf.io.gfile.exists(f'{filename}.npz'): tf.io.gfile.copy(f'gs:\/\/vit_models\/augreg\/{filename}.npz', f'{filename}.npz') timm.models.load_checkpoint(timm_model, f'{filename}.npz') Processing the image into the model, #loading the image into model def pp_torch(img, sz): img = pp(img, sz) img = img.numpy().transpose([2, 0, 1]) return torch.tensor(img[None]) with torch.no_grad(): logits, = timm_model(pp_torch(d['image'], resolution)).detach().numpy() # Visualizing results for Timm plt.figure(figsize=(10, 4)) plt.bar(list(map(ds_info.features['label'].int2str, range(len(logits)))), logits) plt.xticks(rotation=90); We can observe that our model still gives us a correctly predicted label. EndNotes This article tried to explore and understand Image Models and how they work. We also learned about a descriptive PyTorch Image Model known as VIT, where we implemented AugReg to create an Image Label Predictor. The following implementation can be found as a Colab notebook which can be accessed using the link here. Happy Learning! References White Paper: An Image is Worth 16×16 Words: Transformers for Image Recognition at ScaleWhite Paper: MLP-Mixer: An all-MLP Architecture for VisionWhite Paper: How to train your ViT? Data, Augmentation, and Regularization in Vision TransformersDifferent PyTorch Image Models","excerpt":"This article tried to explore and understand Image Models and how they work. We also learned about a descriptive PyTorch Image Model known as VIT, where we implemented AugReg to create an Image Label Predictor.","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Data Science","Deep Learning","Guide","Machine Learning","Python"],"author_name":"Victor Dey","publish_date":"2021-07-31T12:00:00","publication_year":"2021","word_count":1420,"keywords":["artificial intelligence","machine learning","AI","AI (Artificial Intelligence)","PyTorch","ML","Machine Learning","computer vision","Python","deep learning","analytics","JAX","Deep Learning","Data Science","TensorFlow","Guide"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","computer vision","analytics","TensorFlow","PyTorch","JAX"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/complete-guide-to-vit-augreg-a-pytorch-image-model-descriptive-predictions\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10011267,"title":"Hands-on Linear Regression Using Sklearn","content":"In today’s article, we will be taking a look at how to predict the rating of cereals. The problem statement is to predict the cereal ratings where the columns give the exact figures of the ingredients. Link to the data set is mentioned below. We will be making the data ready to go and will fit a simple model into it and would also regularise the data to see how good the model can become. #import necessary libraries import pandas as pd import numpy as np Now you can download the dataset from here. It is advised to read the description of the dataset before proceeding, will help you comprehend the problem better. Extract the data and enter the file path of csv file in it. df=pd.read_csv('D:\\Data Sets\\cereal.csv') #reading the file df.head() #for printing the first five rows of the dataset Output Here since we see that rating column is a continuous data thus it is a regression problem. #dropping the rows that are redundant data=df.drop(['name'],axis=1) #to see if there’s any missing data data.isnull().sum() #no missing values #encoding the data from sklearn.preprocessing import LabelEncoder le=LabelEncoder() #label encoding the first two rows for i in range(2): x[:,i]=le.fit_transform(x[:,i]) Output from scipy.stats import pearsonr corelation=[] for i in range(len(data.columns)-1): col_x=x[:,i] col_y=y corr,_=pearsonr(col_x,col_y) corelation.append(corr) print(corr) Taking the index values of those whose correlation is greater than 0.29 or less than -0.29 If you don’t know what is correlation then you can study it from here. drop_col=[] #dropping the columns whose index is the there in the given condition for i in index: data.columns[i] #print(data.columns[i]) drop_col.append(data.columns[i]) Now the independent variable. x=data.iloc[:,:-1].values #Splitting the dataset from sklearn.model_selection import train_test_split x_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.2) Here the test size is 0.2 and train size is 0.8. from sklearn.linear_model import LinearRegression regressor=LinearRegression() regressor.fit(x_train,y_train) regressor.score(x_test,y_test) #no regularization Output 0.9943613024056396 It is way too high and is overfitted so we will regularize it. You can read about regularisation from here. y_pred=regressor.predict(x_test) #regularizing the linear model from sklearn.linear_model import Ridge ridge_reg_1=Ridge(alpha=1,normalize=True) ridge_reg_1.fit(x_train,y_train) ridge_reg_1.score(x_test,y_test)   #alpha =1 ridge_reg_05=Ridge(alpha=0.5,normalize=True) ridge_reg_05.fit(x_train,y_train) ridge_reg_05.score(x_test,y_test)   #alpha =0.5 ridge_reg_2=Ridge(alpha=2,normalize=True) ridge_reg_2.fit(x_train,y_train) ridge_reg_2.score(x_test,y_test)    #alpha =2 Output Conclusion This article was aimed to discuss the problem statement of cereal rating. We had a look at different things including making the data ready for training where we had label encoded our data columns. Not only that but we trained the data using linear regression and then also had regularised it. To tweak and understand it better you can also try different algorithms on the same problem, with that you would not only get better results but also a better understanding of the same. Hope you liked the article.","excerpt":"In today’s article, we will be taking a look at how to predict the rating of cereals. The problem statement is to predict the cereal ratings where the columns give the exact figures of the ingredients. Link to the data set is mentioned below.","categories":["Deep Tech"],"tags":["Python","regression","sklearn"],"author_name":"Bhavishya Pandit","publish_date":"2020-11-06T11:00:06","publication_year":"2020","word_count":431,"keywords":["Go","NumPy","data_tools:Pandas","TPU","programming_languages:R","AI","programming_languages:Go","regression","Python","Aim","sklearn","R","Pandas"],"extracted_tech_keywords":["AI","Aim","Pandas","NumPy","TPU","R","Go","programming_languages:R","programming_languages:Go","data_tools:Pandas"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/linear-regression-with-sklearn\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":8858,"title":"Mu Sigma Appoints Ambiga Dhiraj as New CEO","content":"Mu Sigma, the world’s largest pure-play provider of decision sciences and analytics solutions, announced today that Ambiga Dhiraj, formerly Chief Operating Officer, will assume the role of Chief Executive Officer (CEO) effective immediately. Founder and former CEO, Dhiraj Rajaram, will continue in a full-time capacity with the company and remain as chairman. Ambiga has been with Mu Sigma for more than eight years, having led functions of talent management, marketing, innovation and product development before assuming the role of COO in February 2012. In that role she drove major improvements to the company’s India operations and the global delivery of products and services to the market. Now she joins an exclusive and short list of female CEOs to lead privately held companies with billion dollar-plus valuations. “Mu Sigma is moving from its adolescence into its prime,” said company founder Dhiraj Rajaram. “The rate at which we’ve added new clients, combined with the amount of help that our clients demand in realizing the potential of Big Data, led me to this decision to work with them more closely. Ambiga is one of the most design-oriented and operationally strong leaders that I’ve ever worked with. The company had planned for this possibility, and the time is right.” While Ambiga will lead the company out of Bangalore, Dhiraj will spend more time across the U.S. and other parts of the world to engage more directly with Mu Sigma clients – more than 140 of the Fortune 500. He’ll also spend time designing new offerings for the company’s man-machine ecosystem, forge strategic alliances, and explore options for using the company’s healthy balance sheet to further cement its position as market leader. “I am honored and excited at the opportunity to lead Mu Sigma in its purpose to help clients adopt a fundamentally new approach to decision sciences and problem solving,” said Ambiga. She began her career in 1998 as a research lead at Motorola, followed by a stint at mobility service provider Kirusa. She holds an M.S. in Computer Engineering, with a specialization in Artificial Intelligence, from Wayne State University in Detroit, Michigan; and a B.E. in Electrical Engineering from Anna University in Chennai, India. Shailendra Singh, Managing Director at Sequoia Capital – an investor in Mu Sigma – said, “We support and recognize this change as yet another sign of Mu Sigma’s growth. This move will unleash Dhiraj to help large enterprises capitalize on the incredible amount of change they’re facing.” Founder Dhiraj added, “It’s been a privilege to lead the Mu Sigma team, but I know in my heart that this new role is how I can add the most value to both the company and its clients. Our purpose as a company will never change, but we must occasionally change how we pursue that purpose.”","excerpt":"Mu Sigma, the world’s largest pure-play provider of decision sciences and analytics solutions, announced today that Ambiga Dhiraj, formerly Chief Operating Officer, will assume the role of Chief Executive Officer (CEO) effective immediately. Founder and former CEO, Dhiraj Rajaram, will continue in a full-time capacity with the company and remain as chairman. Ambiga has been […]","categories":["AI News"],"tags":["Analytics Case Study"],"author_name":"AIM Media House","publish_date":"2016-02-03T08:39:42","publication_year":"2016","word_count":464,"keywords":["big data","API","artificial intelligence","programming_languages:R","AI","innovation","GAN","analytics","Analytics Case Study","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","API","big data","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mu-sigma-appoints-ambiga-dhiraj-as-new-ceo\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":50982,"title":"Indonesia Parliament May Pass Bill To Replace Civil Servants With AI","content":"Indonesia will replace some government workers with AI, as Indonesian president says he needs to lessen top four levels of government to cut red-tapism and boost investments in the country, reported Reuters. President Joko Widodo, prevalently known as Jokowi, has asked the largest Islamic nation to climb the manufacturing sector, and use AI as a beneficial technology in this direction. Indonesian President Joko Widodo has requested government offices to cut two positions of government workers next year and supplant jobs with artificial intelligence (AI), in an offer to cut bureaucracy hampering foreign investment and productivity. Widodo made the comments on Thursday in a room brimming with top executives of big Indonesian organizations as he spread out a second-term agenda planned for changing the structure of the Southeast Asian nation. The president, whose new five-year term started a month ago in the wake of winning a political race in April, said Indonesia should progress to better quality manufacturing, for example, electric vehicles, and utilise raw materials like coal and bauxite in such businesses, not only for export purposes. Such change would require outside foreign investment and Widodo said he would improve the business atmosphere by fixing many redundant rules and cutting red tape in the administration across the country. To diminish red tape, Widodo said the present top four levels in government organisations would be smoothed to two one year from now. In any case, he included this arrangement would require parliamentary endorsement. “Our bureaucracy will be faster with AI,” he said. Widodo didn’t give further details, including any direction for which particular jobs would be evacuated or how artificial intelligence would be utilized. The administration will hand over to parliament a bill on tax reforms also one month from now and another bill tending to labor issues, Widodo stated. Political groups in Widodo’s decision alliance control 74 percent of the seats in parliament, making it simple for his organization to push through enactment. Widodo emphasised the administration’s viewpoint that Indonesia’s economy would develop by around 5 percent this year, more slow than its 5.3 percent focus, in the midst of a worldwide financial slowdown.","excerpt":"Indonesia will replace some government workers with AI, as Indonesian president says he needs to lessen top four levels of government to cut red-tapism and boost investments in the country, reported Reuters. President Joko Widodo, prevalently known as Jokowi, has asked the largest Islamic nation to climb the manufacturing sector, and use AI as a […]","categories":["AI News"],"tags":["Indonesia"],"author_name":"Vishal Chawla","publish_date":"2019-12-02T15:32:00","publication_year":"2019","word_count":354,"keywords":["Go","API","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Indonesia","ViT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","API","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indonesia-parliament-replace-civil-servants-with-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10117298,"title":"AWS Teams Up with Minfy for Cloud and AI Boost through Global Expansion","content":"AI solutions and cloud-native system integrator Minfy Technologies has announced a multiyear strategic collaboration agreement (SCA) with hyperscaler AWS India to improve the former’s use of cloud services and AI. Over the next four years, the SCA will support US$500 million in overall business growth through international expansion. Minfy caters to large-scale enterprises, public sectors, and growing businesses. Now, the company will assist global enterprises in various sectors in utilising AI and cloud technologies effectively. One notable initiative is the Swayam.ai app store, which caters to industries like healthcare, aerospace, logistics, manufacturing, and the public sector, offering tools such as intelligent chatbots and sentiment analysis. The company wants to expand its reach in the U.S., Australia, Malaysia, and the Philippines, focusing on enhancing market strategies, hiring local talents, and developing customer-centric solutions. With a track record of over 500 AWS projects,  Minfy targets a broader international presence. Under this collaboration, Minfy will help transition clients’ workloads to AWS, particularly in healthcare, logistics, and manufacturing sectors. The goal is to facilitate AI integration, cloud-driven transformation, and the development of digital capabilities to improve operational efficiency. The Hyderabad-based company plans to leverage AWS’s advanced AI and ML capabilities, such as Amazon SageMaker and AWS Inferentia, to power Swayam.ai’s generative AI solutions. Additionally, the partnership will utilise AWS’s cloud services like Amazon Elastic Kubernetes Service and Amazon RedShift for efficient database management and modernisation of IT infrastructure. Over the next four years, Minfy intends to upskill its workforce in core AWS competencies, focusing on healthcare, data analytics, and ML. This includes training over 1,000 professionals and establishing a Cloud Centre of Excellence to centralise knowledge and improve solution access globally. “AWS is committed to helping local partners like Minfy drive growth and expand internationally,” said Chris Casey, head, partner management, APJ, AWS. Recently, Bengaluru-based agritech startup Cropin Technology and AWS India signed a Memorandum of Understanding (MoU) to enable Cropin to build an AI-powered solution to address the pressing issues of global hunger and food insecurity.","excerpt":"Over the next four years, the strategic partnership will support US$500 million in overall business growth through international expansion.","categories":["AI News"],"tags":["Amazon","Amazon AWS"],"author_name":"Shritama Saha","publish_date":"2024-03-28T13:19:37","publication_year":"2024","word_count":333,"keywords":["Amazon SageMaker","AWS","AI","chatbots","sentiment analysis","ML","Amazon","RAG","Amazon AWS","analytics","generative AI","kubernetes"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Amazon SageMaker","RAG","chatbots","sentiment analysis","AWS","kubernetes"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-teams-up-with-minfy-for-cloud-and-ai-boost\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":31176,"title":"8 AI Influencers You Must Follow On Twitter To Stay On Top Of Your Game","content":"Artificial Intelligence is the new normal which is rapidly advancing the technology sector and has significantly impacted key aspects of the business. In this article, we list down AI’s power players who are also the thought leaders in the area of machine learning, deep learning, computer vision, robotics and have a massive following. They are industry experts, product innovators and researchers who bring the latest insights and AI trends to their followers. Andrej Karpathy Karpathy is the Director of AI at Tesla and has worked as a Research Scientist at OpenAI. His courses are immensely popular and he is one of the most followed AI pioneers online as well as offline. Follow him here Demis Hassabis Hassabis is the founder and CEO at DeepMindAI. He completed his PhD in cognitive neuroscience from the University College London. His doctoral studies are focused on the field of autobiographical memory and amnesia. He has co-authored papers on this subject which were published in Nature, Science, Neuron, and Proceedings of the National Academy of Sciences (PNAS). His work on the neurological connection between imagination and episodic memory was cited as one of the “Top 10 Scientific Breakthroughs of the Year”. Follow him here Alex Champandard Alex is an AI expert, and his field of expertise is as vast as the domain itself. His interests consist primarily of deep learning research. He has also co-founded CreativeAI. He oversees the widely popular nucl.ai conferences which brings together some of the brightest AI minds across the world to discuss the latest happenings in the field. Follow him here Soumith Chintala Well known for creating PyTorch and co-authoring WGAN and DCGAN research papers, Soumith works at Facebook AI research and his research projects range from pedestrian detection, sentiment analysis to digit classification. Follow him here Hugo Larochelle Hugo is a Google Brain researcher and a machine learning professor whose retweets help the followers to access the state-of-the-art research material. His publications include topics like brain tumour segmentation with deep neural networks, deep learning with nanophotonic circuits among many others. Follow him here Sander Dieleman Sander is a Research Scientist at DeepMind. His research was about learning hierarchical representations of musical audio signals for classification and recommendation, with a focus on deep learning and feature learning. His tweets frequently about new researches, research papers and other developments in the new tech areas. Follow him here Kate Crawford Kate is a Principal Researcher and Co-founder of AI Institute at NYU. Her works have been published in highly reputed journals like Nature and her recent publications address the impacts of AI on society, data bias and fairness, predictive analytics and due process, and algorithmic accountability and transparency. Follow her here. Dr GP Pulipaka Dr Pulipaka is a Chief Data Scientist and his book ‘The Future Of Data Science’ is listed as one of the top books in this domain. In his tweets, he primarily focuses on the core concepts behind machine learning models and also about the latest discoveries in this field. Follow him here.","excerpt":"Artificial Intelligence is the new normal which is rapidly advancing the technology sector and has significantly impacted key aspects of the business. In this article, we list down AI’s power players who are also the thought leaders in the area of machine learning, deep learning, computer vision, robotics and have a massive following. They are […]","categories":["AI Trends"],"tags":["deepmind london","Machine Learning","Twitter (X)"],"author_name":"Ram Sagar","publish_date":"2018-12-07T05:03:46","publication_year":"2018","word_count":504,"keywords":["data science","artificial intelligence","machine learning","OpenAI","AI","neural network","PyTorch","Machine Learning","computer vision","deep learning","analytics","Twitter (X)","deepmind london"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","computer vision","data science","analytics","OpenAI","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ai-influencers-twitter-top-handles-follow\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10143515,"title":"Meta FAIR Announces New Research Artifacts","content":"Meta FAIR (The Fundamental AI Research) has unveiled new research artifacts, amidst major AI announcements from OpenAI and Google, according to their blog. It includes releases in language, embodied AI, new architectures, and more. “It’s been a big year for AI, and today at NeurIPS I’m excited to share nine new open source releases from Meta FAIR to wrap up the year — all part of our continued mission to achieve advanced machine intelligence (AMI), and I am very excited for even more to come in 2025! ” said Joelle Pineau, VP of AI research, Meta, on Linkedin. AI for Advanced Behaviour, Intelligence, and Vision-Language Capabilities Meta Motivo enables virtual humanoid agents to perform complex tasks with human-like behaviours by learning from motion datasets and adapting to environmental factors. Theory-of-Mind Data provides tools and datasets designed to train AI in understanding and predicting human thoughts and beliefs, significantly advancing research in social intelligence. Memory Layers and Large Concept Models (LCM) further enhance AI’s capabilities by improving its ability to store, retrieve, and process complex information, enabling better handling of diverse languages and hierarchical reasoning. Complementing these advancements, Meta CLIP 1.2 is a vision-language model that precisely aligns image and text data, supporting critical tasks such as retrieval, classification, and multi-modal embedding. Content and Media Generation Meta Video Seal enhances digital content security by embedding robust, invisible watermarks in videos, ensuring they remain resistant to editing and compression. Supporting this innovation, Omni Seal Bench introduces a leaderboard for evaluating neural watermarking techniques, with plans for a dedicated workshop in 2025 to further advance the field. Additionally, Flow Matching serves as an open-source framework that facilitates the creation of high-quality images, videos, audio, and 3D structures, empowering users to generate rich and dynamic media content efficiently. On the whole, under Meta AI chief Yann LeCun‘s leadership, the company is advancing AI with systems like Layer Skip and V-JEPA to improve reasoning and interaction, alongside self-supervised learning to ultimately AMI.","excerpt":"New research showcase innovations in agents and machine learning architectures.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Aditi Suresh","publish_date":"2024-12-13T18:48:53","publication_year":"2024","word_count":327,"keywords":["Go","self-supervised learning","Meta AI","OpenAI","AI","Modal","innovation","Git","CLIP","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","Meta AI","R","Go","Git","CLIP","self-supervised learning","innovation","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-fair-announces-new-research-artifacts\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10048262,"title":"Guide To Distributed Representations in ML","content":"Distributed Representations (DR) play a significant role in machine learning. DR is a principled way of representing entities (say, cats or dogs) in terms of vectors. Entities sharing common properties have vector representations that are nearer to each other. Numeric Representations and the Role of DR The input and output of the Machine Learning (ML) models are often numeric. This requires finding a suitable numeric representation of text. Consider that the following sentences are used to train an ML model. He is a King.King is a man.Queen is a woman.She is a Queen.King and Queen are rulers. For the words to be fed as an input to the model, it needs a mathematical representation. One-Hot encoding of words to vectors is one way to get this representation. The dimension of each vector is equal to the number of unique words in all the sentences. This collection of unique words is referred to as vocabulary. Using one-hot representation, we have, King represented as [0 1 0 0 0 0 0] and Queen represented as [0 0 0 1 0 0 0]. This is an example of local representation of words if the vocabulary of the above sentences considers only nouns and pronouns. However, this representation is not very expressive as it does not capture much information about similar words. For example, King and Queen are rulers. The number of unique words increases with the increase in size of the input data. This requires longer word vectors to represent each word. Moreover, we do not capture the semantic similarity of words. To overcome these issues, we have DR of words to represent similar words by nearer vectors. Building DR with an Example Let us imagine that we want to express the words “Man”, “Woman”, ”King”, “Queen”, “Ruler” using 2-D vectors such that they preserve the following semantics: King-Man+Woman —> QueenKing-Man —> RulerRuler+Woman —> Queen Note that we have used the standard vector representation of the variable with an overhead arrow. For example, King is the vector representation of the word “King”. If the rules of vector arithmetic should hold, one way to choose vectors satisfying the above rules is as shown below. Man =[0,1], Woman=[2,1] King =[1,1], Queen=[3,1] Ruler =[1,0] The vector representation of the words above can be visualized in a two-dimensional vector space as shown below. However, the example taken here is very simple. In real-world scenarios, there could be thousands of words to deal with and hundreds of thousands of contexts by which multiple words could be related to each other. In such cases, assigning appropriate vector representations to these words manually would be cumbersome or even infeasible. So, we need a generalized way to perform the task. The theory of deep learning has produced beautiful results in this respect. Before diving into the details of achieving DR using deep learning, we must take a look at a simple implementation of the popular Word2Vec model. Building DR using Word2Vec Model Let us start with a small text corpus which is a collection of the following sentences. We went to the beach on a sunny day.There were many tourists on the beach.We went to the nearby museum with other tourists. Our objective is to automate the derivation of the word vectors. We use the above sentences to learn the semantic similarity of words. As we are interested in word vectors, we start with tokenizing the sentences. sentences = [[We, went, to, the, beach, on, a, sunny, day],[There, were, many, tourists, on, the, beach], [We, went, to, the, nearby, museum, with, other, tourists]] We have seen that one-hot encoding of words is an inefficient way of vectorial representation. A better way is to use Word Embeddings. Embeddings provide dense representations where similar words are identifiable and also reduce the dimensionality of the vector. A Machine Learning (ML) model is trained to learn the values of the embeddings. from gensim.models import Word2Vecfrom sklearn.decomposition import PCAmodel=Word2Vec(sentences,min_count=1)X = model[model.wv.vocab]pca = PCA(n_components=2)result = pca.fit_transform(X) The vector representations can be plotted in the 2-D vector space as follows. from matplotlib import pyplotpyplot.scatter(result[:, 0],result[:, 1])words=list(model.wv.vocab)for i, word in enumerate(words):a,b = result[i,0],result[i,1]pyplot.annotate(word,xy=(a,b))pyplot.show() Various neural network models have been designed to build word embeddings. Common Bag of Words (CBOW) and Skip Gram are two such examples. CBOW model is a popular neural network implementation to arrive at the distributed representation of words. While in CBOW model, a word is predicted from its context word, the Skip Gram model attempts to predict the context word with respect to a word. When implemented using a neural network, the input layer of CBOW model contains multiple context words as input and the output is a single word. The Skip Gram model, however, contains a single input word and the output layer comprises multiple context words corresponding to the input. DR using DL Generally, we can consider a Deep Neural Network (DNN) to be composed of an input layer that takes input vectors as input to the DNN, hidden layers (often seen as a black box) and an output layer that gives the output vector. The weights in each hidden layer serve as a compact representation of the input vectors, which in most cases are understood only by the neural network. These new input representations can be used to transform the input vector into a lower-dimensional vector. If we input words to a DNN and decide on a loss function, then the DNN can be trained by the backpropagation algorithm. The output of such a DNN can be used as the distributed representations of the words. To understand this better, let us perform a task of binary classification (positive, negative) for sentiment analysis of IMDB movie reviews in the Large Movie Review Dataset, which has 25,000 labelled movie reviews. Find the data here.We are now ready to write the Python code to implement the sentiment analysis classifier which uses the idea of embeddings. Add the following library imports. import ioimport osimport reimport shutilimport stringimport tensorflow as tffrom tensorflow.keras import Sequentialfrom tensorflow.keras.layers importDense, Embedding,GlobalAveragePooling1Dfrom tensorflow.keras.layers.experimental.preprocessing importTextVectorization Each data point is a movie review which is classified as a positive or a negative sentiment. We need to firstly upload the data for use by executing: url = https:\/\/ai.stanford.edu\/~amaas\/data\/sentiment\/aclImdb_v1.tar.gz dataset = tf.keras.utils.get_file(“aclImdb_v1.tar.gz”, url, untar=True, cache_dir = ’.’ cache_subdir = ‘ ‘)dataset_dir =os.path.join(os.path.dirname(dataset), aclImdb)train_dir = os.path.join(dataset_dir, train)remove_dir = os.path.join(train_dir,     unsup)shutil.rmtree(remove_dir) Our data is split into training and validation set by using the keras.preprocessing module by executing: batch_size = 1024seed = 123train_ds = tf.keras.preprocessing. text_dataset_from_directory(aclImdb\/train, batch_size=batch_size,validation_split=0.2, s ubset=training, seed=seed)val_ds = tf.keras.preprocessing.text_dataset_from_directory(aclImdb\/train, batch_size=batch_size, validation_split=0.2, subset=validation, seed=seed) Often, the text data that we use directly from a dataset available online can have undesired components like HTML tags, punctuations, etc. So, we need to do some preprocessing to clean the text data. This is taken care of while initializing the TextVectorization layer. It also creates the vocabulary which is used while training the model. We create a standardized function to use in the TextVectorization layer, so that it can clean the text data according to our needs. def custom_standardization(input_data):    lowercase =   tf.strings.lower(      input_data)                stripped_html = tf.strings.regex_replace(lowercase,<br \/>,  )   return tf.strings.regex_replace(            stripped_html, [%s] %            re.escape(string.punctuation),’ ‘)vocab_size = 10000sequence_length = 50vectorize_layer = TextVectorization(standardize=custom_standardization,max_tokens=vocab_size,output_mode=’int’,output_sequence_length =sequence_length)text_ds = train_ds.map(lambda x, y: x)vectorize_layer.adapt(text_ds)vectorize_layer.get_config() The task that we have chosen to get DR is sentiment classification. A neural network is built to achieve sentiment classification as a model. We use Adam as an optimization algorithm, Binary Cross-Entropy as the loss, and accuracy as the performance parameter to train the model.embedding_dim=16model =   Sequential([vectorize_layer,                           Embedding(vocab_size,embedding_dim, name=”embedding”),             GlobalAveragePooling1D(),                  Dense(16,activation=’relu’),Dense(1)])model.compile(optimizer=adam,                    loss=tf.keras.losses.                    ,BinaryCrossentropy                    (from_logits=True),                     metrics=[accuracy]) Now, we use the train and validation data sets to train the model. The number of epochs used here is 15 as the focus is on explaining the workings of the code. However, a higher number of epochs might lead to better accuracy. We can check this by executing: model.fit(train_ds,validation_data=val_ds, epochs=15,callbacks=[tensorboard_callback]) Now, since we have trained our model to a certain level of accuracy, we can retrieve the word embeddings from our Google Colab project, which are the new DR of words in the vocabulary. weights = model.get_layer(embedding).get_weights()[0]vocab =           vectorize_layer.get_vocabulary()out_v = io.open(vectors.tsv, w, encoding=utf-8)out_m = io.open(metadata.tsv, w, encoding=utf-8)for index, word in enumerate(vocab):if index == 0:continuevec = weights[index]out_v.write(\\t.join([str(x) for xin vec]) + “\\n”)out_m.write(word + “\\n”)out_v.close()out_m.close() We can download the embeddings from Google Colab to our local disk by executing: try:from google.colab import filesfiles.download(vectors.tsv)files.download(metadata.tsv)except Exception:pass We have now got the DR of words in the vocabulary. We can visualize them by uploading the embedding files downloaded above here. We have chosen the vocabulary size to be 10,000 and the corresponding embedding for each word in the vocabulary is 16-dimensional. This means that the embedding layer in the model above has 160,000 parameters. This can be confirmed and the other details of the model used above can be viewed by executing: model.summary() We have shared the code used in this article here. Applications of DR We saw an application of embeddings in the sentiment analysis task. Another fascinating example of distributed representation is that when words are trained together with paragraphs, we can predict the equivalence between two similar, yet different words. Suppose you come across a paragraph about Bill Gates. The article is a paragraph and ‘Microsoft’ would be an obvious word in the paragraph. Now check out what distributed representation can do: ParagraphVector(“Bill Gates”)- WordVector(“Microsoft”)+ WordVector(“Apple”) → ParagraphVector(\\Steve Jobs”) Such exciting results find several applications in the real world. We see embeddings used in almost every domain such as legal text understanding, clinical healthcare or even software engineering. Authors Anisha Sahaanisha@cmi.ac.inAnisha Saha is a graduate student at the Chennai Mathematical Institute. Her interests are in Data Science and Machine Learning.Chandrashish Prasadchandrashish@cmi.ac.inChandrashish Prasad is a graduate student at the Chennai Mathematical Institute. His interests are in Software Engineering and Deep Learning.Venkatesh Vinayakaraovenkateshv@cmi.ac.inVenkatesh Vinayakarao is a lecturer in the Department of Computer Science at the ChennaiMathematical Institute. His interests are in information Retrieval, Program Analysis and software Engineering.","excerpt":"Distributed Representations (DR) play a significant role in machine learning. DR is a principled way of representing entities (say, cats or dogs) in terms of vectors. Entities sharing common properties have vector representations that are nearer to each other. Numeric Representations and the Role of DR The input and output of the Machine Learning (ML) […]","categories":["Deep Tech"],"tags":["Guide"],"author_name":"AIM Media House","publish_date":"2021-09-13T17:35:15","publication_year":"2021","word_count":1647,"keywords":["data science","machine learning","Keras","AI","neural network","ML","Colab","Aim","deep learning","TensorFlow","Guide"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","data science","Aim","TensorFlow","Keras","Colab"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-distributed-representations-in-ml\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171212,"title":"Only ₹17 Cr Disbursed Out of ₹6,000 Cr Budget for India’s National Quantum Mission: RTI","content":"India’s National Quantum Mission (NQM), a government initiative focused on research and development in quantum computing, has so far disbursed only ₹17.02 crore out of its total allocated budget of ₹6,003.65 crore. Approved in April 2023, the mission’s funding was intended to be distributed over eight years, from the financial year 2023-24 through to 2030-31. This, along with the following information, was revealed after a citizen filed an RTI (Right to Information) application to retrieve the details about the mission. Out of the total ₹17.02 crore, the allocations were distributed among key Indian institutions specialising in quantum technologies. IISc Bengaluru received ₹3.05 crore for Quantum Computing; IIT Madras was awarded ₹5.16 crore for quantum communication; IIT Bombay secured ₹5.79 crore for quantum sensing and metrology; and IIT Delhi obtained ₹3 crore for quantum materials and devices. The responses to the RTI application also revealed that funding was only provided for research institutions, not for large players or startups. However, it was also revealed that under the National Mission on Interdisciplinary Cyber-Physical Systems (NM-ICPS) scheme of the Department of Science and Technology (DST), the I-Hub Quantum Technologies Foundation, established at IISER Pune, has selected eight quantum startups for support. These startups include QNu Labs, QPiAI India, Dimira Technologies, Prenishq, QuPrayog, Quanastra, Pristine Diamonds, and Quan2D Technologies. Recently, it was reported by the Mint that these eight startups, are set to receive funds of ₹30 crore each as part of the NQM. In 2020, the Indian government announced the National Mission on Quantum Technologies and Applications (NM-QTA), with a budget of ₹8,000 crore. However, from 2020 to 2023, not a single rupee was disbursed. The mission was later rebranded as NQM in 2023. The citizen shared the details of the RTI queries and responses on LinkedIn. Besides NQM, the country has taken several steps to improve the quantum computing ecosystem. In April, the office of the principal scientific adviser (PSA) to the Indian government released the first edition of the International Technology Engagement Strategy – Quantum (ITES-Q). It offers foundational analysis to guide domestic and international stakeholders in government, academia, and industry, facilitating impactful partnerships and enhancing India’s global presence in emerging technologies. Recently, IBM, Tata Consultancy Services (TCS), and the government of Andhra Pradesh announced plans to deploy India’s largest quantum computer, set to be housed in the newly established Quantum Valley Tech Park in Amaravati. The facility will house an IBM Quantum System Two, which is equipped with a 156-qubit Heron processor. The initiative will focus on developing quantum algorithms and applications for real-world challenges.","excerpt":"From 2020 to 2023, not a single rupee was disbursed under the National Mission on Quantum Technologies and Applications.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","india quantum computing","quantum"],"author_name":"Supreeth Koundinya","publish_date":"2025-06-03T16:12:23","publication_year":"2025","word_count":425,"keywords":["Go","funding","programming_languages:R","AI","programming_languages:Go","india quantum computing","quantum","Ray","R","emerging_tech:quantum computing","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","Ray","R","Go","startup","funding","programming_languages:R","programming_languages:Go","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/only-%e2%82%b917-cr-disbursed-out-of-%e2%82%b96000-cr-budget-for-indias-national-quantum-mission-rti\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168879,"title":"AI-Driven Multimodal Models Reshape Indian Businesses, Deloitte Report","content":"Indian businesses are increasingly adopting small, multimodal models to meet demands for faster, more efficient, and targeted solutions, with AI leading the transformation across sectors, according to Deloitte’s latest Tech Trends – India Perspective 2025 report. The report emphasises the growing role of context-aware intelligence in reshaping industries. The report noted that by combining multimodal AI, spatial computing, and advanced analytics, systems are becoming increasingly adept at understanding environments, behaviours, and intent. This shift enables hyper-personalised experiences, more innovative resource planning, and adaptive decision-making across sectors like education, retail, urban planning, and agriculture. “Indian enterprises are moving beyond automation to deploy technologies that interpret, adapt and act in real time, making digital transformation more fluid, inclusive and future-ready,” Abhrajit Ray, partner and CIO program leader, Deloitte India, said. Deloitte’s report also highlights that the rapid growth of generative AI (GenAI) has re-centered hardware in the technology conversation. What was once seen as a commoditised, slowly evolving domain is now undergoing rapid innovation to meet the complex demands of advanced AI systems. Specialised chips for applications such as power management, telecommunications, and cryptographic acceleration are becoming increasingly critical. Moreover, the integration of heterogeneous computing — combining GPUs and NPUs on a single chip — is unlocking new levels of efficiency and performance across diverse AI workloads. In addition, spatial computing powered by AI is allowing systems to anticipate needs rather than react to them proactively. The report further notes that AI is reshaping the foundations of modern enterprise architecture. Traditional monolithic systems are being replaced with decoupled, composable structures that seamlessly integrate AI insights across business functions. Despite these advances, the transformation is not without its challenges. It demands complex architectural redesigns to ensure scalability, security, and long-term sustainability. A warning is also sounded about the impending Year to Quantum (Y2Q) era, which could break traditional encryption standards. This makes the adoption of quantum-safe encryption and fortified cybersecurity measures critical for safeguarding India’s digital future. Finally, focusing on the IT sector, the report underlines AI’s increasing role in enhancing tech talent. AI is now utilised for code generation, software testing, and automation, minimising manual effort and improving operational efficiency.","excerpt":"The report emphasises the growing role of context-aware intelligence in reshaping industries.","categories":["AI News"],"tags":["Deloitte"],"author_name":"Shalini Mondal","publish_date":"2025-04-29T21:06:32","publication_year":"2025","word_count":358,"keywords":["Go","GenAI","Deloitte","AI","ML","Scala","Ray","analytics","generative AI","multimodal AI","R"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","GenAI","multimodal AI","Ray","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-driven-multimodal-models-reshape-indian-businesses-deloitte-report\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10067432,"title":"It&#8217;s a wrap! Intel® oneAPI masterclass on Neural Compressor to accelerate deep learning inference","content":"Intel® concluded its oneAPI masterclass on Intel® Neural Compressor on May 13, 2022, at IST 5:00 pm. The masterclass witnessed close to 200+ participants. The workshop covered various initiatives and projects launched by Intel®, alongside deep-diving into Intel® Optimisation for TensorFlow to enhance the performance on Intel platforms. Intel® Optimisation for TensorFlow is the binary distribution of TensorFlow with Intel® oneAPI deep neural network (oneDNN) library primitives. In addition to this, the masterclass gave an overview of the oneAPI AI analytics toolkit, which contained a core set of tools and libraries for developing high-performance applications on Intel® CPUs, GPUs, and FPGAs, alongside throwing light on Intel®’s open-source, cross-platform performance library oneDNN for deep learning applications. The session also highlighted the importance of using Intel Neural Compressor to boost deep learning inference, alongside giving a hands-on demo, use-cases, benchmarks, and more. The workshop was led by Kavita Aroor, a developer marketing lead – Asia Pacific & Japan at Intel; Aditya Sirvaiya, an AI software solutions engineer at Intel; and Zhang Jianyu (Neo), a senior AI software solution engineer (SSE) of SATG AIA in the PRC. Intel® had a demo on the following topics: oneAPI AI Analytics Toolkit overviewIntroduction to Intel® Optimization for TensorflowIntel optimisations for Tensorflow Intel Neural CompressorHands-on demo to showcase usage and performance boost on DevCloud Check out the recording for the “Speed up deep learning inference with Intel® Neural Compressor” masterclass session here. Key highlights Throwing light on various initiatives launched by Intel® to uplift the developers’ ecosystem, Intel’s Aroor touched upon the DevMesh projects forum. She said they have about 2000-2500 projects across various technologies, including AI, gaming development, IoT and HPC, curated by Intel software developers and the student ambassador community. Further, she said that developers could create their blog articles, where they can amplify their work, projects and more. She also said that oneAPI Technology partner program have a separate forum where they can partner with Intel to unleash the power of oneAPI. “Today, we have about 19-20 such organisations across ten different countries,” added Aroor. Next, she spoke about the oneAPI certified instructor programme, where developers can become certified instructors, become an advocate for Intel®, and work very closely with the company. Click here to download Intel® oneAPI Toolkits to get started. Click here to create an Intel® DevCloud account. After this, Aditya gave the audience a quick overview of the oneAPI analytics toolkit, introducing Intel® Optimisation for TensorFlow, and then a hands-on, showcasing usage and performance boost on DevCloud. This was followed by the introduction of Intel Neural Compressor Hands-on with quantisation workload. The demo explained an end-to-end pipeline to train a TensorFlow model with a small customer dataset and speed up the model based on quantisation by Intel® Neural Compressor. This included training a model by Keras and Intel Optimisation for Tensorflow, getting an INT8 model by Intel® Neural Compressor, alongside comparing the performance of the FP32 and INT8 models by the same script. (Source: Intel) Click here to download Intel® oneAPI Toolkits to get started. Click here to create an Intel® DevCloud account. During the event, Analytics India Magazine also ran a Lucky Draw, wherein lucky participants won an Amazon Voucher worth INR 2000\/- each at the end of the workshop. The winners were selected based on their engagement with Discord throughout the workshop. <​​https:\/\/discord.gg\/ycwqTP6> Sahil ChachraRaviteja PeriPrateek ModiRamachandrareddy GadiSreyashi BhattacharjeeAishwarya GholseAnirban MallaDhruv KothiyaChirumamilla PitchaiaShivaraj KarkiPriyanka PatnyAkanksha SrivastavaMaaz MohammedMani Shekhar Gupta Check out the recording for the “Speed up deep learning inference with Intel® Neural Compressor” masterclass session here.","excerpt":"The workshop covered various initiatives and projects launched by Intel®, alongside deep-diving into Intel® Optimisation for TensorFlow to enhance the performance on Intel platforms and more.","categories":["Deep Tech"],"tags":["Intel","oneAPI"],"author_name":"Amit Naik","publish_date":"2022-05-19T12:00:00","publication_year":"2022","word_count":588,"keywords":["API","Keras","AI","neural network","deep learning","ViT","analytics","GAN","TensorFlow","R","oneAPI","Intel"],"extracted_tech_keywords":["AI","deep learning","neural network","analytics","TensorFlow","Keras","R","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/its-a-wrap-intel-oneapi-masterclass-on-neural-compressor-to-accelerate-deep-learning-inference\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":62417,"title":"Bangalore-Based AI Company Delivering N95 Masks To Fight Against The Spread Of COVID-19","content":"Niki.ai, a Bangalore-based artificial intelligence company, is now providing N95 masks to fight against the spread of COVID-19. In a recent LinkedIn post, Shishir Modi, the co-founder of the company, confirmed the news by stating — “We at Niki are now providing N95 masks for our customers. Our team and various partners are working hard to deliver it to Pan India today. Please keep you & your family members safe in this pandemic – stay indoors as much as possible. #StayHomeStaySafe #FightCorona.” Niki.ai has arranged a delivery of the N95 masks pan India done by the team and various partners. The initiative has been taken in order to curb the spread of the deadly virus. The post further provided links to obtain those masks. “Get your CORONA mask from Niki.” The customers have to visit their site and can order N95 masks — in twos and fours. Niki.ai was founded in May 2015 by IIT Kharagpur graduates Sachin Jaiswal, Keshav Prawasi, Shishir Modi and Nitin Babel and has leveraged technologies like natural language processing and machine learning to create a multilingual chatbot for its users’ digital transactions. The company understands how users chat in India, deciphers the words, in the context of product\/services that they would like to purchase, and comes up with apt recommendations.","excerpt":"Niki.ai, a Bangalore-based artificial intelligence company, is now providing N95 masks to fight against the spread of COVID-19.  In a recent LinkedIn post, Shishir Modi, the co-founder of the company, confirmed the news by stating — “We at Niki are now providing N95 masks for our customers. Our team and various partners are working hard […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Companies","artificial intelligence corona","Artificial Intelligence India","covid-19"],"author_name":"Sejuti Das","publish_date":"2020-04-23T12:24:55","publication_year":"2020","word_count":215,"keywords":["artificial intelligence","machine learning","covid-19","AI","programming_languages:R","Artificial Intelligence India","Git","RAG","AI Companies","AI (Artificial Intelligence)","R","artificial intelligence corona"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bangalore-based-ai-company-delivering-n95-masks-to-fight-against-the-spread-of-covid-19\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39503,"title":"HTTP vs HTTPS: Why You Should Care More About HTTPS","content":"In this ever-evolving technology era, security on the web has become a prime concern for everyone. News headlines of data breaches and companies and systems getting compromised are creating a fearful ecosystem. Even though big organisations across the world are taking serious and advanced steps to eliminate or at least mitigate cyber threats, new vulnerabilities and data breaches are being reported on a daily basis. Today, it has become imperative to do a better job of securing our data online. Talking about securing the web, HTTPS is one of the most vital parts. There was a time when HTTP was doing the job, however, gone are those days. Today, if you are not making use HTTPS you are putting yourself and your data on risk. You all must be wondering now, what is the difference between HTTP and HTTPS? And why HTTPS is more secured. In this article, we will be seeing just that. Why People Are Dropping HTTP And Opting HTTPS HyperText Transfer Protocol (HTTP) is an application protocol for distributed, collaborative, hypermedia information systems. Simply put, it is a protocol that is used to transfer hypertext over the Web. Even though it has been one of the widely used protocol, it couldn’t last for a long time because of the security flaw it poses — data (i.e. hypertext) exchanged using HTTP  goes as plain text. Meaning, if anyone establishes an interception between the browser and the server can keep an eye on the data. Today, in order to set up a secure session between server and browser, almost every website is getting rid of HTTP and implementing HTTPS. HTTPS is the obvious answer when you are looking for a data protection issue. The “S” stands for security and that doesn’t convince you, then take a look at your address bar where the lock icon is there — it is the sign of an encrypted website connection. Today, whether its e-commerce, banking or any other sector, websites across the world are using https in order to protect data by encrypting it before sending. How HTTPS Can Save You From Certain Cyber Attacks One of the biggest advantages of using HTTPS is the fact that it lets you verify whether the webpage is authentic or a cloned one. For example, when you log into a website where you are going to enter some sensitive information, the first you need to see is the address bar. Make sure that you see that lock icon next to the URL. The popularity of HTTPS has reached such a level that even the search giant Google has made HTTPS default, making your search history private. Otherwise, just like the old days, anyone on the same network would be able to see what you have a search on google. Another major benefit of using HTTPS is that it prevents Man in The Middle (MITM) attack. If you don’t know what a man-in-the-middle attack is, it is an attack where the attacker secretly establishes space and eavesdrop on the communications between two parties. Sometimes, the attacker also alters the communication between both the ends, and they believe they are directly communicating with each other. So, when a website uses HTTPS, it encrypts the communication, making it difficult for third-party connections to intercept the network. Word to the wise: HTTPS definitely let you verify whether they are legit or not. However, attackers have found out different ways too. An attacker cannot clone a website with the same URL, but it can make the URL look somewhat similar. For example, Phishing URL for Facebook: www.facb00k.com. Bottom Line When it comes to cyber security, no organisation can take a light step. There was a time when people would think, they can still carry on with HTTP even after knowing that it has flaws. However, gone are those days, irrespective of the domain, every company has started implementing HTTPS.  So, if you are still among those who are relying on HTTP thinking your website, the blog is not dealing with sensitive information, you are compromising big time with web security.","excerpt":"In this ever-evolving technology era, security on the web has become a prime concern for everyone. News headlines of data breaches and companies and systems getting compromised are creating a fearful ecosystem. Even though big organisations across the world are taking serious and advanced steps to eliminate or at least mitigate cyber threats, new vulnerabilities […]","categories":["Deep Tech"],"tags":["Cybersecurity","Network"],"author_name":"Harshajit Sarmah","publish_date":"2019-05-21T13:17:28","publication_year":"2019","word_count":680,"keywords":["Go","AWS","AI","ETL","cloud_platforms:AWS","programming_languages:R","Git","Network","ViT","GAN","Cybersecurity","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Git","ETL","GAN","ViT","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/http-vs-https-why-you-should-care-more-about-https\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10076117,"title":"Council Post: Can Supercloud benefit Data Science Industry","content":"The next version of cloud innovation is here. Automation is non-negotiable in today’s world. Data received from the cloud cannot be manipulated, operated or controlled as customers want an effortless experience. Companies today want to evolve because they no longer want to be dependent on a single provider and benefit from innovative solutions that have higher availability. Adopters of Multi-cloud face obstacles like inconsistency, need for multiple teams for each cloud and repeated implementation of functionality for each cloud. Keeping in mind how advanced the native cloud services are and the complexities they entail, we are in need of an architecture that helps businesses take advantage of the cloud services they need and simultaneously deliver multi-cloud services that provide stability, simplify the environment and reduce cost. Howie Xu, VP of ML\/AI at Zscaler Inc., said that Supercloud eliminates the complexities that are involved, and AI models will follow the data. Thus without supercloud, it is very hard to do AI machine learning. The interest in supercloud ability to connect across clouds and being able to operate in a hybrid environment makes life easier for users and developers. Companies use heterogeneous sources making it difficult to understand and analyse data, and so with this new term what does the future of data science and analytics look like? The roundtable conference was moderated by Satyajit Nair, Director of Innovations and Architecture at Laerdal along with panellists Anees Merchant, EVP – Global growth and client success at Course5i; Suresh Chintada, CTO at Subex; Ankush Sabharwal, Founder and CEO at Corover and Biswajit Biswas, Chief Data Scientist at Tata Elxsi with the agenda of understanding how supercloud can benefit DS\/AI & Analytics. Supercloud as “The new term” The whole idea is to give a user a perspective of them using a single interface, as data can reside anywhere. So, from the user point of view, it is a single cloud, or is a super cloud. However, underneath, you have multiple technologies, multiple security, network topologies being addressed. So, these things put together, we design a layer of abstractions, which puts these complexities of implementation in a more managed form, so that the end user gets a very seamless interface and experience. —Biswajit Biswas, Chief data Scientist at Tata Elxsi ‘Super cloud’ seems to be very moniker. If I were to use the word of cross cloud infrastructure, to really provide you an ability to create applications that can serve customers across, markets, regions, or otherwise it is, app migration, perhaps a service across the cloud providers. It also helps you to use this abstraction layer to provide interfaces, tools, technologies to start again, do the same thing as what we used to do in the cloud and allocate, migrate and terminate the sources in a fashion that is more agnostic of the underlying infrastructure. —Suresh Chintada, CTO at Subex Evolution of Supercloud 90% people would not be using just one cloud platform. Everyone is acknowledging whether, it is an IT company or a cloud provider, a company would not just get stuck with just one cloud provider. You think about a software architecture or enterprise architecture we do see the applications are spread across various cloud platforms practically. — Ankush Sabharwal, Founder and CEO at CoRover Multi cloud multi environment is already there. Companies have always been in the mind-set that there is data sitting outside, which is your propriety. With cloud infrastructures like the GCP, Azure, AWS has come in, organisations have realised a big mistake they had done in the past of getting too confined to one provider from an infrastructure perspective. And that’s why this rise of multi cloud multi environment will continue to exist. It’s just going to increase so much that I think it’s going to bring in other problems. —Anees Merchant, EVP – Global growth and client success at Course5i Security, an obstacle for supercloud To alleviate some of these data security problems, there is a need for a different set of query mechanisms like the queries which will fetch results not in bulk, but in a chance which is totally encrypted across the network security and compute security. All these three layers of security need to be addressed. The IAM policies have to be revamped. — Biswajit Biswas, Chief data Scientist at Tata Elxsi Supercloud encompassing other benefits Whether superclouds should be existing and considered on prem, those on prem infrastructures continue to exist. And now they are interacting with the cloud. SuperCloud is just another buzzword. It’s been there, it continues to be there. Mature organisations will define their own policies, what kind of data exists, what kind of processes exist. And based on that what security governance and processes need to be defined at these multi layered areas. — Anees Merchant, EVP – Global growth and client success at Course5i Benefits with supercloud The promise of super cloud is, if it provides you a trade-off to make it simpler to work across this cloud without having to worry about which cloud I’m working with at this point hence which tool or methodology or management system I need to use. It’s definitely an abstraction that I would want to look forward to. — Suresh Chintada, CTO at Subex Certain platforms whether you look at the sales tech, operations tech, HR tech, or the key organisation dimensions, these platforms are defaulted or built on certain of these cloud ecosystems. And then there are these AI\/ML platforms, which are very good or built around a specific cloud. Now trying to get that into some other environment, there are certain latencies and issues which you will see from a data and data processing aspect. Second aspect is data sovereignty, which is becoming a big thing where we used to manage these on prem infrastructures, because we had to manage the data within those environments. So, cloud providers who don’t have those infrastructures to manage and scale and manage the requirement of that nation, unfortunately, are giving rise to the super cloud, and they have that requirement. —Anees Merchant, EVP – Global growth and client success at Course5i Superpass for supercloud? Sometimes GPUs or any other cloud components are cheaper in cloud-1, not in cloud-2. And, we should also be able to create VMs in Cloud-3 may be because the third part application is deployed there or the static IPs there have been white-listed for the access to the third party applications. Moreover, we may need to use the Cloud-4, for the heavy outbound resource usage, which may have less data egress cost. So, everything should be happening so seamlessly using SuperPass, which should also enable application migration as a service across different availability zones or cloud providers. We should be able to allocate, migrate, and terminate resources such as virtual machines and storage and presents a homogeneous network to tie these resources together. — Ankush Sabharwal, Founder and CEO of CoRover Supercloud for which industry? What Cloud essentially gave us is speed to innovation. And what I would say is any industry that is really innovating in the future, is healthcare, banking, FinTech would definitely benefit from a super cloud, because I’m assuming that super cloud will only accelerate their innovation. —Suresh Chintada, CTO at Subex Super-Cost I think the super cloud will actually solve this runaway cost problem and take the best usage of where the cost benefit is for that day. So, an AI\/ML can come and do a great job here. So, they will constantly analyse and perform all those analyses to find out where the cost is running, where the cost is low, which instances today’s going lower, which regions are having less. So, there is some kind of bidding process that can be done, which we of course, will enable, so that will allocate resources or towards the least cost areas and the customer has a greater flexibility to kind of pick and choose the areas which part they should run in their cost areas. —Biswajit Biswas, Chief data Scientist at Tata Elxsi Having one Supercloud is better than four separate cloud services that have the same vendor. Though it is the future, it may not be the preferred outcome for customers. It is a powerful term that pushes debates, thoughts and simultaneously solves problems public cloud vendors are not addressing directly. Yet Supercloud lacks a proper definition. Whether Supercloud will become the next big thing or not is a test of time. But the actual test will be when companies can identify themselves as users of supercloud when it reaches the value and specificity that they need to position them to customers. As Paul Maritz, CEO of VMware said, “Cloud is about how you do computing, not where you do computing.”","excerpt":"The roundtable conference was moderated by Satyajit Nair, Director of Innovations and Architecture at Laerdal along with panellists Anees Merchant, EVP – Global growth and client success at Course5i; Suresh Chintada, CTO at Subex; Ankush Sabharwal, Founder and CEO at Corover and Biswajit Biswas, Chief Data Scientist at Tata Elxsi with the agenda of understanding how supercloud can benefit DS\/AI & Analytics.","categories":["AI Features"],"tags":[],"author_name":"AIM Media House","publish_date":"2022-09-30T15:00:00","publication_year":"2022","word_count":1458,"keywords":["data science","machine learning","GCP","AWS","AI","R","ML","RAG","analytics","Azure"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","RAG","AWS","Azure","GCP","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-can-supercloud-benefit-data-science-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10047197,"title":"Elon Musk And His Over-Exaggerated Claims","content":"The announcement of the humanoid robot was one of the first biggest highlights of the Tesla AI Day 2021 event. The announcement was complete with a bit of theatrical fanfare. A dancer, dressed in a suit representing how the proposed Tesla robot would look, entered the stage dancing amidst applause. The specifications of this robot were specified via slides which promised that the robot would be five feet tall, weigh 125 pounds, and have human-level hands. In addition, it would perform ‘dangerous, repetitive, boring tasks’. https:\/\/twitter.com\/RMac18\/status\/1428708172856918018?s=20 As exciting as this proposal sounds, many experts panned this announcement. The robotics industry, by and large, has been under slump in recent years. Many promising names in the robotics business have had to shut shop. This included names like Anki and Jibo. Almost 30-year-old Boston Dynamics, too, has suffered significant losses in recent years. By the end of March 2020, the company had suffered a net loss of $103 million. A major stake buyout finally rescued it by Hyundai, who paid close to $900 million, a value considered too small for a company with such potential and scope. Robotics is a tricky field. It requires expertise and skills spanning domains like software engineering, mechanical, electromechanical engineering, and complex assembly procedures. Performing even simple tasks like climbing stairs or moving across a room without running into obstacles can prove to be a major challenge. Considering these challenges, Musk’s claim of developing a humanoid robot in just a year seems to be unbelievable. That said, stranger things have happened in the world; we will have to wait for a year to see how this unfolds. This is not the first time Musk has made tall claims. There have been several instances in the past when he promised to deliver but has failed. We list some of them in this article. Fully Self Driving Car Earlier this week, Elon Musk said that the latest release of FSD Beta 9.2, the company’s experimental driver-assistance software, is ‘not great’. He tweeted that his Autopilot\/AI team is rallying to improve it as fast as possible and trying to build a single stack for highway and city streets. The company currently sells the FSD package for $10,000 or $199 per month in the US. This system still needs a human driver behind the wheel. Interestingly, just a few months ago, Musk told investors that he was ‘highly confident’ that the Tesla car would be able to drive itself with reliability ‘in excess of humans this year’. Such a claim came under intense scrutiny, especially since it coincided with a Tesla vehicle without any human driver that crashed in Texas, killing two men. The company’s director of Autopilot software then had to issue a statement to the California Department of Motor Vehicles. The California DMV then said in a memo, “Elon’s tweet does not match engineering reality per CJ. Tesla is at Level 2 currently.” Level 2 technology refers to a semi-automated driving system that requires human driver supervision. Putting Humans On Mars In December 2020, Elon Musk said that his space company SpaceX would land humans on Mars by 2026. A few years back, Musk had said that humans would be able to inhabit Mars by the 2060s. Both the claims are very bold. Mars has a highly hostile environment for humans to survive. It has a thin atmosphere, no magnetic strip to protect the surface from sun radiations, no breathable air, and high average temperatures. Given such conditions, Musk’s claim right now seems mere speculation and even an over-exaggeration. Hyperloop Musk had first spoken about a transport system that would propel passengers in capsule-like vehicles, floating on a cushion of air through low-pressure tubes. He called them Hyperloop and described them as a cross between a Concorde, a railgun, and an air hockey table. According to Musk, this project received verbal government approval in 2017 for building a hyperloop between New York, Philadelphia, Washington DC, and Baltimore. No other substantial piece of information has emerged from this project. The contract for Las Vegas underground tunnels has now been downgraded to just a ‘loop’. Neuralink Since as early as the 1990s, researchers have been probing the possibility of having brain implants that would extend human capabilities. The Tesla and SpaceX owner took it a notch higher when he said that his company Neuralink would develop such a brain-computer interface that can be installed in a doctor’s office in under an hour. In 2019, Musk also claimed, “We hope to have this, aspirationally, in a human patient by the end of this year. So it’s not far.” Last year, Musk introduced a coin-sized disk containing chips that compress and wirelessly transmit signals recorded from the electrodes. The disk is called ‘the link’ and is as thick as a human skull. Musk said that the link could be placed on the brain’s surface through a drill hole that can be secured with superglue. “I could have a Neuralink right now, and you wouldn’t know it,” Musk said. A few months back, Musk’s company demonstrated a monkey playing ping-pong via wireless implants. Despite these smaller wins, Musk’s dream of bringing such brain implants to the market has not been successful so far. Taking Tesla Private At $420 In 2018, Musk tweeted that he is taking his company private when the stock prices hit $420 per share. This announcement created much furore. The US Securities and Exchange Commission had announced that it would be suing Musk for his comments. Am considering taking Tesla private at $420. Funding secured.— Elon Musk (@elonmusk) August 7, 2018 To Musk’s credit, not all of his promises and claims have fallen flat. One can even credit a lot of his groundbreaking innovations to his optimism. To end with, check out some of the best quotes by Elon Musk that showcases another side of his personality.","excerpt":"There have been several instances in the past when Elon Musk promised to deliver but has failed. We list some of them in this article.","categories":["AI Features"],"tags":["Elon Musk"],"author_name":"Shraddha Goled","publish_date":"2021-08-27T15:00:00","publication_year":"2021","word_count":976,"keywords":["Go","funding","programming_languages:R","AI","innovation","programming_languages:Go","RAG","Elon Musk","Aim","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","innovation","funding","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/elon-musk-and-his-over-exaggerated-claims\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":1058,"title":"Science of Analytics","content":"Analytics is not just pure science; it is part art as well. Organizations that master the fine art of using analytical tools realize increased revenues and enjoy cost savings. Last week we talked broadly about ANALYTICS. This week we dive into the “SCIENCE OF ANALYTICS.”  The scientific approach involves the following four key steps: 1.    Observe\/define the business problem: Observation is either an activity consisting of receiving knowledge, or the recording of data using scientific instruments. The term may also refer to any data collected during this activity. Analytics begins with observing the phenomenon and setting up the right business problem. It requires understanding the facts, to which you have ready access, and then drawing conclusions from it to identify the business problem which needs to be solved. For example, a manufacturing company is suffering from declining profits. By looking at their balance sheet we realize that revenues have declined while the costs have remained constant. Through these two facts, we can identify a simple business problem – the manufacturing company must reduce costs or increase revenue if it wants to have the same profitability as before. 2.       Hypothesis: A hypothesis is a proposed explanation for an observable phenomenon. People refer to a trial solution to a problem as a hypothesis — often called an “educated guess” because it provides a suggested solution based on the evidence. Researchers may test and reject several hypotheses before solving the problem. Taking the above mentioned example of the manufacturing company, the business may have two sets of hypothesis: a.    Increase Revenue:  Within increasing revenue, the firm might think of many different avenues: i.    Focus on Marketing – Increasing the marketing budget will enable us to increase sales and hence increase revenue. ii.    Focus on Price – By reducing the price of our product we would be more competitive and hence increase sales, which might offset the decrease in sales\/unit. b.    Reduce Costs: Within reducing cost bucket, the organization has various alternatives: i.    Operations cost – By reducing the operations budget (e.g. staff, electricity etc.), we will reduce costs. ii.    Reduce Marketing budget – By reducing the marketing budget, we will save on costs. As you can see, you can achieve increased profitability by both increasing and decreasing marketing budgets. There are several implications of each action beyond the primary implication and all need to be evaluated. The key element of the hypothesis-building phase is that you should have a mutually exclusive and collectively exhaustive set of hypothesis. This means we should think about all the possible sets of relevant hypothesis for the situation at hand and ensure they do not overlap and that together they are complete. 3.    Test\/Experimentation: An experiment is the step in the scientific method that arbitrates between competing models or hypotheses. Experimentation is also used to test existing theories or new hypotheses in order to support them or disprove them. An experiment or test can be carried out using the scientific method to answer a question or investigate a problem. First, an observation is made and then a question is asked, or a problem arises. Next, a hypothesis is formed and an experiment is used to test that hypothesis. The results are analyzed, a conclusion is drawn, sometimes a theory is formed, and results are communicated through business cases. A good experiment usually tests a hypothesis. However, an experiment may also test a question or test previous results. The fundamental reason for following this process is to ensure the results and observations are repeatable and can be closely replicated given similar circumstances. Let’s continue with the example above and set up a test for the manufacturing company to learn whether increasing the marketing budget would affect revenue. In this case, we would set up a TEST where we run the EXISTING marketing programs and call it GROUP A while in GROUP B we run the increased marketing program. At the end of the observation time frame (assume 2-3 months), we would measure revenue for GROUP A and GROUP B and understand the differences. As long as the groups have a statistically significant size we should be able to repeat these results. 4.       Learn: Learning is acquiring new knowledge, behaviors, skills, values, preferences or understanding, and may involve synthesizing different types of information. Continuing our manufacturing company example, let’s assume that GROUP B performed far better than GROUP A. Let’s also assume that at the same time we increased marketing our competitors decreased it in the GROUP B target market. Now the question becomes, was the incremental benefit driven by our increased marketing or the fact that competitors reduced their marketing? Assimilating all possible and relevant information is extremely important in order to reach a good decision. As you can tell, while scientists have been utilizing the above mentioned technique for a long time, businesses are just beginning to use it. This requires a strong commitment to the scientific process and a systematic approach to create a TEST & LEARN environment where you are constantly testing, learning and evolving to create increased bottom line benefits for a company.","excerpt":"Analytics is not just pure science; it is part art as well. Organizations that master the fine art of using analytical tools realize increased revenues and enjoy cost savings. Last week we talked broadly about ANALYTICS. This week we dive into the “SCIENCE OF ANALYTICS.”  The scientific approach involves the following four key steps: 1.    […]","categories":["IT Services"],"tags":[],"author_name":"Rahul Nawab","publish_date":"2012-09-11T10:36:50","publication_year":"2012","word_count":848,"keywords":["Replicate","Go","programming_languages:R","AI","programming_languages:Go","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","Go","GAN","ViT","Replicate","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/science-of-analytics\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10097391,"title":"It&#8217;s Time For ChatGPT to Have A Voice","content":"In the ever-evolving landscape of AI, OpenAI’s ChatGPT is on the cusp of a groundbreaking transformation. With the recent introduction of ‘custom instructions‘, users can now experience a more personalised and intuitive interaction with the AI model. But the possibilities don’t end there. Imagine ChatGPT with its own unique voice, ushering in a new era of conversational AI. This development is reminiscent of the past shortcomings of voice assistants like Alexa, which failed to deliver personalised experiences. By incorporating custom instructions, users will have the ability to use fewer prompts, making the interaction process more efficient and user-friendly as ChatGPT will now be able to remember your conversation context based on your chosen preferences, allowing for a more personalized and tailored AI interaction experience. With the exciting development of ChatGPT remembering conversations, integrating a voice interface seems like the next logical step. Just like how Amazon’s Alexa revolutionized the way we interact with voice-controlled devices. In the past, voice assistants like Alexa could perform tasks like sharing news and playing songs, but they lacked a personalized touch. With ChatGPT’s memory feature, we can take this interaction to a new level. Imagine booking a restaurant or listening to your favorite songs effortlessly, as ChatGPT remembers your preferences and caters to your unique needs. This evolution in the field of generative AI brings us closer to a more natural and intuitive way of interacting with technology. How Amazon Alexa made big promises and failed Upon its initial rollout in 2015, Alexa faced a surge of users posing curious and offbeat questions, ranging from inquiries about the meaning of life to whimsical desires. However, as time passed, Alexa’s unsatisfactory responses failed to retain users’ interest. The devices lacked the ability to deliver personalized experiences or generate significant advertising revenue for the company. As a result, users’ engagement with Alexa gradually diminished over time. It was speculated that  generative AI will bring back Alexa from the dead, however it doesn’t look like it is going to happen anytime soon. A few months ago, Alexa was declared dead. The company pulled a plug on its ‘Amazon Alexa’ voice-assisted feature succumbing to huge operating losses. In addition to Amazon, other companies like Google have also shifted their focus away from their respective voice assistants, as reported earlier this year by The Information. Similarly, Microsoft’s Cortana and Samsung’s Bixby, once considered Alexa competitors, have also faced challenges and reduced prominence in the market. Right now OpenAI has the first mover advantage in generative AI and it is not easy to catch up in this race. This is high time for OpenAI to tap the market of voice assistance before anyone else does. Bard breathing under the neck Recently, Google announced  that the Bard  would be available in over 40 languages. With its latest updates, users can now listen to Bard’s responses, alongside changing the tone and style of Bard’s responses to five different options: simple, short, long, professional or casual. Google is making significant efforts to enhance Bard’s conversational abilities and enable generative AI to speak more naturally.  Personalised voice interaction opens up numerous use cases that are more than just asking factual details like weather outside. For instance like asking for recipes based on groceries, asking questions and receiving quick insights on any topic with ChatGPT holding to and fro conversation just like any other human. Currently, Bard has access to real-time information and events, giving it an advantage in these scenarios. In contrast, ChatGPT relies on plugins, which are only available to plus users, to access real-time data. OpenAI in the last few days had a slew of news agency partnerships through Associated Press and American Journalist Project. Hopefully, this will help ChatGPT to have real time information just like Bard which will make personalized conversations even better. To stay competitive and meet user expectations, OpenAI should consider incorporating voice control into ChatGPT, transforming it into a personal virtual assistant like Jarvis from Ironman. Just like Tony Stark’s seamless conversations with Jarvis, the world is eagerly waiting for real-life AI interactions that can provide personalized and natural responses.","excerpt":"With the exciting development of ChatGPT remembering conversations, integrating a voice interface seems like the next logical step.","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-07-23T10:26:56","publication_year":"2023","word_count":683,"keywords":["Go","ChatGPT","OpenAI","AI","ML","GPT","generative AI","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ML","generative AI","ChatGPT","OpenAI","R","Go","GPT","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/its-time-for-chatgpt-to-have-voice\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039387,"title":"Visualize Data in 3D Using plot3D R-Package","content":"Data visualization holds an important place among a wide range of tasks handled by data scientists and analysts for getting a proper insight of the data in hand. This post will talk about how to visualize data in 3D using the plot3D package of the R programming language. Note: The code snippets throughout this article have been implemented using RStudio IDE (version 1.2.1335). RStudio can be downloaded from here. Practical implementation Install and load plot3D package. install.packages(\"plot3D\") library(\"plot3D\") Load the built-in Iris dataset. data(iris) Display initial records of the data. head(iris) Extract sepal length, petal length and sepal width data from the respective columns of the dataset. The values of these attributes for a given record will be used as x,y and z coordinates for plotting a point in 3D. a <- sepal.len <- iris$Sepal.Length b <- petal.len <- iris$Petal.Length c <- sepal.width <- iris$Sepal.Width Scatter plots: scatter3D() function can be used to draw 3-dimensional scatter plot. scatter3D(a,b,c, clab = c(\"Sepal\", \"Width (cm)\")) Here, a,b and c represent the x,y and z coordinates of points respectively. ‘Clab’ parameter specifies the label of the color key. Output: The type of box to be drawn surrounding scatter plot points can be specified using ‘bty’ parameter. The color key legend can be disabled using ‘colkey’ parameter. scatter3D(a, b, c, bty = \"f\", colkey = FALSE, main =\"bty= 'f' (full box)\") Output: scatter3D(a, b, c, bty = \"b2\", colkey = FALSE, main =\"bty= 'b2'(with back panels and grid line)\") By specifying ‘bty’ value to ‘u’, we can manually select attributes such col.panel, col.grid etc. scatter3D(a, b, c, pch = 18, bty = \"u\", colkey = FALSE, main =\"bty= 'u'\", col.panel =\"steelblue\", expand =0.4, col.grid = \"darkblue\") Output: 6) The plotted points can be annotated using text3D() function. scatter3D(a, b, c, pch = 18,  theta = 30, phi = 30, main = \"IRIS Flower data\", xlab = \"Sepal.Length\", ylab =\"Petal.Length\", zlab = \"Sepal.Width\") ‘theta’ and ‘phi’ parameters define the azimuthal angle, and co-altitude of the scatter plot. Output: Line plots: ‘Type’ parameter of the scatter3D() function should be specified for customizations like lines (only), points and lines, vertical lines. #Only lines scatter3D(a, b, c, phi = 0, bty = \"g\", type = \"l\", ticktype = \"detailed\", lwd = 4) Output: #Vertical lines scatter3D(x, y, z, phi = 0, bty = \"g\",  type = \"h\", ticktype = \"detailed\") Output: #Both points and lines scatter3D(x, y, z, phi = 30, bty = \"g\",  type = \"b\", ticktype = \"detailed\") Output: Regression plot We will use the built-in mtcars dataset for putting a regression plot. Load the dataset. data(mtcars) Display its initial records. head(mtcars) Output: # Initialize x, y, z coordinate variables x <- mtcars$wt y <- mtcars$disp z <- mtcars$mpg # Calculate linear regression (z = ax + by + c) using lm() method fit <- lm(z ~ x + y) # predict values on a X-Y grid grid.lines = 30   #number of lines on grid #predict x, y and z variables’ values x.pred <- seq(min(x), max(x), length.out = grid.lines) y.pred <- seq(min(y), max(y), length.out = grid.lines) xy <- expand.grid( x = x.pred, y = y.pred) z.pred <- matrix(predict(fit, newdata = xy), nrow = grid.lines, ncol = grid.lines) # fit data points to drop lines perpendicular to the grid surface fitpoints <- predict(fit) # draw scatter plot with regression plane scatter3D(x, y, z, pch = 18, cex = 2, theta = 20, phi = 20, ticktype = \"detailed\", xlab = \"wt\", ylab = \"disp\", zlab = \"mpg\", surf = list(x = x.pred, y = y.pred, z = z.pred, facets = NA, fit = fitpoints), main = \"mtcars\") Output: Code source.R file of the above code snippets References The following article throws light on to visualize data in 3D using the plot3D package of R programming language. R documentationplot3D documentationArticle on basic R conceptsArticles on various applications using R: (article1) (article2) (article3)","excerpt":"Data visualization holds an important place among a wide range of tasks handled by data scientists and analysts for getting a proper insight of the data in hand. This post will talk about how to visualize data in 3D using the plot3D package of the R programming language. Note: The code snippets throughout this article […]","categories":["AI Trends"],"tags":["data analyst vs data scientist","Visualization"],"author_name":"Nikita Shiledarbaxi","publish_date":"2021-05-04T10:00:00","publication_year":"2021","word_count":644,"keywords":["data analyst vs data scientist","TPU","programming_languages:R","AI","Visualization","R"],"extracted_tech_keywords":["AI","TPU","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/visualize-data-in-3d-using-plot3d-r-package\/","complexity_score":3,"technical_depth":4,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":56210,"title":"Why AI Is The Perfect Drinking Buddy For The Alcoholic Beverage Industry","content":"The use of AI-driven processes to increase efficiency in the F&B market is no longer an anomaly. A host of breweries and distilleries have incorporated the technology to not only develop flavour profiles faster, but also for other functions, including packaging, marketing, as well as to ensure they meet all food-safety regulations. Although the intention is not to find a replacement for the brewmaster\/distiller, it becomes a thrilling learning experiment that equips them with multiple data points that could help them come up with innovative ideas. Here is a list of companies that have successfully blended technology into their beverages to make a heady cocktail: IntelligentX The company claims to be the world’s first to use AI algorithms and machine learning to create innovative beers that adapt to users’ taste preferences. Based on customer feedback, the recipe for their brews goes through multiple iterations to generate various combinations. IntelligentX currently has four different varieties — Black AI, Golden AI, Pale AI, and Amber AI. How does it work? Codes are printed on the cans which direct customers to the Facebook Messenger app. They are then asked to give feedback on the beer they tried by answering a series of 10 questions. The data points gathered are then fed into an AI algorithm to spot trends and inform the overall brewing process. Furthermore, using the feedback, the AI also learns to ask better questions each time to get better outcomes. Although the insights gathered give brewmasters a window into understanding customer preferences better, the final decision to heed the AI’s recommendations to create a fresh brew rests on them. But what is certain is that without technological intervention, such a large collection of data would not only be difficult to process, but also extremely time-consuming. Mackmyra Whisky Multi-award-winning Swedish whiskey distillery Mackmyra Whisky collaborated with Microsoft and Finnish tech company Fourkind to create the world’s first AI-generated whiskey. Using Microsoft Azure and Machine Learning Studio, Fourkind’s resulting AI solution was fed into Mackmyra’s existing recipes and customer feedback data to create thousands of different recipes. Following this, the distillery’s key master blender Angela D’Orazio used her experience to review which ingredients would work well together, filtering down the recipes to more desirable combinations. Since this process was repeated multiple times over, the AI algorithm picked up on which combinations worked best and using machine learning, began producing more desirable mixes. Eventually, D’Orazio was able to filter it down to five recipes, finally arriving at recipe number 36 which ultimately became the world’s first AI-generated whiskey that went into production. This “AI-generated, but human-curated” whiskey has opened the doors to new and innovative combinations that would otherwise have never been discovered. Monikered ‘Intelligens’, the first batch of this blend was launched in September 2019. Carlsberg The Copenhagen-based brewery started a multimillion-dollar project in 2017 to analyse different flavours in its beer using AI. Unlike IntelligentX which uses customer feedback to improve its brew, Carlsberg has accomplished this by developing a taste-sensing platform that helps identify the differential elements of the flavours. Under the ongoing Beer Fingerprinting Project, 1000 different beer samples are created each day. With the help of advanced sensors, the flavour fingerprint of each sample is determined. Following this, different yeasts are analysed to map the flavours and help make a distinction between them. Thus, the data collected by this AI-powered system could potentially be used to develop new varieties of brews. Launched in collaboration with Microsoft, Aarhus University and the Technical University of Denmark, the project marked a shift from conventional practices that did not involve any technology. https:\/\/www.youtube.com\/watch?v=v9rT5O4U89Q AB InBev The brewers of Budweiser and Corona had also jumped on the AI bandwagon to shake up its business. The company had invested in a slew of initiatives to improve how it brews beer. The Beer Garage is one such initiative. Sitting at the interjection of a startup ecosystem and the AB InBev business, it focuses on developing technology-driven solutions. ZX Ventures – another offshoot of its larger business – was launched in 2015 with the objective of creating new products that address consumer needs. Anchored around these enterprises, AB InBev is using machine learning capabilities to stay ahead of the curve in three broad areas: Improving quality and flavourBuilding better relationships with customersAllowing tech-intervention in choosing marketing and advertising content Sugar Creek Brewing This maker of Belgian-inspired ales has begun integrating AI and IoT into its brewing process to improve both the quality of the beer, as well as its manufacturing process. It started when a significant problem came to light at the packaging stage. When the beer was loaded into bottles, it was observed that the level at which it was filled was inconsistent. Another problem was the excessive foaming inside the bottles. This spiked the oxygen levels in the beer, which is known to ruin the flavour and reduce the beer’s shelf life. After partnering with IBM, the tech giant installed a camera at SCB’s warehouse, which took pictures of the beer as it crossed the bottle line. When combined with other data collected during the packaging operations, the team of engineers at IBM uploaded it to the Cloud. At this point, brewers at SCB also provided specific criteria which they found to be useful and this was then left with Watson algorithms to interpret the large amount of data quickly and solve the problem.From losing more than $30,000 a month in beer spillage, SCB found a solution by building AI and IoT into its brewing processes. https:\/\/www.youtube.com\/watch?v=wYGum1GIHWw","excerpt":"The use of AI-driven processes to increase efficiency in the F&B market is no longer an anomaly. A host of breweries and distilleries have incorporated the technology to not only develop flavour profiles faster, but also for other functions, including packaging, marketing, as well as to ensure they meet all food-safety regulations. Although the intention […]","categories":["AI Features"],"tags":[],"author_name":"Anu Thomas","publish_date":"2020-02-07T16:01:56","publication_year":"2020","word_count":922,"keywords":["Go","machine learning","AI","R","RAG","Aim","GAN","Azure","T5","startup"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","Azure","R","Go","T5","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-ai-is-the-perfect-drinking-buddy-for-the-alcoholic-beverage-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":38874,"title":"How NLP Is Making These Day-to-Day Tasks More Intelligent","content":"Image source: SlideShare As artificial intelligence and machine learning continue to transform the way we interact with the world, the one subtle or the invisible power that we all owe a huge credit is Natural Language Processing (NLP). From grammar check to auto-prediction, the technology the has proven to be a powerful tool in improving our lives on a day-to-day basis and is making our daily chores a lot easier to handle. In this article, we look at how NLP is aiding this process and understand how exactly is it doing so. Auto Prediction One of the most common use-cases of NLP that we encounter every day is when we compose a mail in Gmail. The power of the technology is such that Gmail auto predicts the commonly used phrases while writing the mail, which includes customary addressing, greetings and a person’s name if we replaying in an email thread. In the case of Gmail, the feature was added in 2018, when Google announced that it will use NLP for its email platform. Called the Smart Compose feature for Gmail, the tech giant relied on a combination of NLP features like Bag of Words (BoW), Recurrent Neural Network (RNN) and n-grams to device technology. This they say is to make the process of email composing easier and faster, thus helping people to concentrate on their other work. In order to achieve this, the company used its vast pool of data to train the models and used full TPUv2 Pod to perform the experiment, thus helping the models to convergence in less than a day. The company further stated that in the future, it hopes to personalise the technology according to the individual’s style of writing. Grammar Check In the hustle-bustle of everyday work chaos, it is most likely that we will misplace, misspell or even miss out words or grammar when we are composing a mail or writing something. But, thanks to the advancing capabilities of NLP, most often we are warned by the red or blue underlines that force us to look into the draft again. In India, especially, where English is still a second-language, scores of people rely on software like Grammarly to fix their grammar and to polish their writing. The platform is known to extensively use technologies like AI, ML and NLP to make this happen. According to a blog spot by Grammarly, it said that their systems combine machine learning with a variety of natural language processing approaches. From grammatical sentences, sentences, paragraph to full text, its algorithms’ parse through many nuanced aspects to predict the outcomes. Their AI systems have also been trained to learn from feedback, which includes certain command as ignore, “When lots of users hit “ignore” on a particular suggestion, for example, Grammarly’s computational linguists and researchers make adjustments to the algorithms behind that suggestion to make it more accurate and helpful,” it said. Chatbots Chatbots have literally replaced human service across the business from food delivery to ticket booking app and it is known for its quick and prompt to customers queries. Essentially, NLP is leveraged at the core of the technology along Natural Language Generation (NLG), ML and Fundamental Learning (FL) to train the bots to interact with the customers and to streamline responses. By applying key features like semantics and cognitive computing, the bots are equipped to comprehend words in a particular context and respond accordingly. Translation The presence of different languages and variations in dialects have made India the perfect market for NLP service providers to grow and thrive. Due to this reason, the number of product offerings from technology giants in Indian regional languages has increased in the past few years as the number of Indian language users has grown to 234 million. This has been made possible as tech giants now rely on deep neural networks for translating words or phrases into its intended meaning. As per a Stanford study, which looked into how machine translation was conducted, the researchers used deep source-side linguistics analysis to improve their Chinese-to-English translation process and trained the classifier to recognise certain Chines words into its syntactic and semantic context. Further, they used Minimum Error Rate Training (MERT), a procedure that optimises the system’s performance on an automated measure of translation quality","excerpt":"As artificial intelligence and machine learning continue to transform the way we interact with the world, the one subtle or the invisible power that we all owe a huge credit is Natural Language Processing (NLP). From grammar check to auto-prediction, the technology the has proven to be a powerful tool in improving our lives on […]","categories":["AI Features"],"tags":["Gmail","NLP"],"author_name":"Akshaya Asokan","publish_date":"2019-05-09T12:50:48","publication_year":"2019","word_count":713,"keywords":["machine learning","artificial intelligence","Gmail","AI","neural network","chatbots","ML","TPU","RAG","NLP","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","NLP","RAG","chatbots","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-nlp-is-driving-value-from-language-to-drive-personalisation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10125474,"title":"Noventiq to Open a New Global AWS Cloud Centre of Excellence in India","content":"Noventiq has signed a Strategic Collaboration Agreement (SCA) with Amazon Web Services (AWS) to provide businesses worldwide with scalable, secure, and innovative cloud solutions, including machine learning and generative AI services. With AWS, Noventiq will open a new global AWS Cloud Centre of Excellence (CCoE) in India to drive innovation and technology best practices. The global CCoE will be responsible for a wide range of functions, including governance, oversight, and coordination of all cloud initiatives, integrating sales and marketing efforts into practices, and driving accountability for results. The CCoE will operate across several realms, including research, evangelizing new solutions, applying strategic initiatives, leading market engagement, and mentoring within the organization. Under the SCA, Noventiq will accelerate the digital transformation of enterprises, SMBs, public sector, and startups in key sectors like financial services, manufacturing, media and entertainment, and retail across the globe. Noventiq, a global leader in digital transformation and cybersecurity solutions and services, aims to migrate more than 1,000 customers from on-premises infrastructure to the world’s most comprehensive and broadly adopted cloud by 2027. This will enable customers to capitalise on new-age cloud technologies like AI. This goal includes delivering large transformation projects, like the project Noventiq did for Profectus Capital, a financial institution that caters to small-and medium-sized businesses (SMBs). With Noventiq and AWS, Profectus developed a data lake on AWS that uses AWS Glue, Amazon S3, and Amazon Redshift. As a result, Profectus now has an instant, 360-degree view of insights across multiple systems to help streamline the decision-making process around loan sanctions. A clear example of the project’s outcomes is Profectus’s ability to utilize a data lake to harness unstructured data for lead generation. The company’s sales team processed approximately one million annual school reports available in the public domain as PDFs. Using Amazon Textract to extract key information, they successfully identified schools in need of loans. As part of its global expansion to the Middle East and Europe, Noventiq is also expanding its AWS operations with five local AWS CCoEs and five Regional Delivery Centers (RDCs) in the UK, the Netherlands, Germany, the UAE, and Saudi Arabia, on top of the previously launched centres in APAC and LATAM to grow the company’s local AWS delivery capabilities. To ensure the success of the CCoEs and RDCs, Noventiq will launch comprehensive training and development programs to nurture future technology leaders and foster talent development, besides creating a robust system to recruit, train, and elevate junior talent. Noventiq will launch an AWS Centre of Learning as part of the SCA, to attain 1,000 AWS Certifications in the next three to five years. These skills will help Noventiq’s consultants, engineers, architects, and developers in key business units become well-versed in building solutions on AWS. Noventiq will also recruit 250 AWS specialists by 2027 for the CCoEs and RDCs to support the company’s growth trajectory and enhance its capabilities in delivering advanced cloud solutions.","excerpt":"Noventiq has signed a Strategic Collaboration Agreement (SCA) with Amazon Web Services (AWS) to provide businesses worldwide with scalable, secure, and innovative cloud solutions, including machine learning and generative AI services. With AWS, Noventiq will open a new global AWS Cloud Centre of Excellence (CCoE) in India to drive innovation and technology best practices. The […]","categories":["AI News"],"tags":["AWS"],"author_name":"Pritam Bordoloi","publish_date":"2024-07-02T15:08:38","publication_year":"2024","word_count":483,"keywords":["Go","machine learning","AWS","AI","ML","Scala","Git","Aim","generative AI","R"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","Aim","AWS","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/noventiq-to-open-a-new-global-aws-cloud-centre-of-excellence-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10117157,"title":"OpenAI&#8217;s Sora Takes About 12 Minutes to Generate 1 Minute Video on NVIDIA H100","content":"OpenAI’s Sora generates 5 minutes of videos produced per NVIDIA H100 per hour, equivalent to 120 minutes of videos per H100 per day, according to estimates from Factorial funds. The report further adds that an estimate of approximately 89,000 NVIDIA H100 GPUs needed to support the creator community on TikTok and YouTube. Combining the AI-generated video production from TikTok and YouTube yields a total of 10.7 million minutes of videos produced daily by AI. However, taking into account factors such as realistic utilisation, peak demand and busy traffic, the estimated number of Nvidia H100 GPUs needed at peak demand is approximately 720,000, significantly higher than the initial calculation based on simplified assumptions. Creators are likely to generate multiple candidate videos before selecting the best one, leading to an average of two candidates per uploaded video. This factor also doubles the GPU requirements. In a recent interview with the Wall Street Journal, CTO Mira Murati shared that OpenAI will make Sora publicly accessible later this year. When Sora was launched earlier in February, users greatly appreciated its hyper-realistic videos, many calling it the “ChatGPT moment for video” The model, showcased in February, generates realistic scenes from text prompts and will soon be open for public use. The initial rollout will primarily target visual artists and filmmakers. Murati also disclosed plans to incorporate sound and editing flexibility into Sora-generated videos. OpenAI is pitching Sora to Hollywood. The ChatGPT creator has scheduled meetings in Los Angeles next week with Hollywood studios, media executives and talent agencies to form partnerships in the entertainment industry and encourage filmmakers to integrate its new AI video generator into their work, reported Bloomberg.","excerpt":"An estimate of approximately 89,000 NVIDIA H100 GPUs needed to support the creator community on TikTok and YouTube.","categories":["AI News"],"tags":["OpenAI","Sora"],"author_name":"Siddharth Jindal","publish_date":"2024-03-26T11:31:02","publication_year":"2024","word_count":275,"keywords":["NVIDIA H100","ChatGPT","OpenAI","AI","programming_languages:R","RAG","GPT","Sora","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","RAG","R","GPT","NVIDIA H100","llm_models:GPT","llm_models:ChatGPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openais-sora-takes-about-12-minutes-to-generate-1-minute-video-on-nvidia-h100\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":38128,"title":"5 Things To Know Before You Start Your Data Science Career","content":"In the present scenario, data science jobs are the hottest trend. According to our recent report, there has been an overall growth in the number of jobs in analytics and data science ecosystem with India contributing to 6% of open job openings worldwide. The total number of analytics and data science job positions available are 97,000 wherein India, about 97 percent of job openings are full-time while the rest 3 percent are contractual. In this article, we jot down five crucial ways to reach great heights in the data science career. 1| Start Your Journey With An Internship Being a data scientist is not easy and involves lots of analytical skills as well as problem-solving experience. If you want to progress in your career as a data scientist, the best way to start is as an intern which will help you to tackle many unseen challenges. One more advantage is that as an intern you will get help wherever you are stuck and your co-workers will give you multiple pieces of advice from their own experiences which will help you step ahead in your career. 2| Be Competitive And Understand Business Problems In order to become a good data scientist, you should always have the curiosity of asking questions whenever there is a doubt which not only enhances a good communication between the co-workers but also helps you to become a good analyst. It is also crucial to understand the business metrics and other statistical problems in order to solve the business problems in an organisation. 3| Go Beyond Theoretical Concepts As a data scientist, having a fundamental understanding of concepts is a must. It is crucial to understand the basic concepts of machine learning, artificial intelligence, knowledge of Python, R, SQL, Julia, Hive, etc. including the deep understanding of linear algebra and other statistical methods as well as gain a piece of in-depth knowledge in data analysis, data visualisation, pre-processing, etc. 4| Add Multiple Skills Problem-solving is the core part of data science which helps in segmenting large business problems into smaller and solvable ones. Most of the time, large organisations look for data science specialists with in-depth knowledge in some specific area. But, if you have multiple skill sets rather than the specialised area which can be resulted as helpful to your organisation, no one can stop you from being a step ahead in your career. Also sometimes, there are organisations who want some extra skill set apart from your knowledge in the field of data science, in that case too having extra skills add plus points in your career building. 5| Don’t Stop Learning And Practicing As we know that change is the only constant. Artificial intelligence and machine learning are evolving in an exponential way and are not constant. We cannot compare the present scenario with a few years back. It is very important to keep on the track of learning with the growing technology to be in the rat race. There are many ways for upskilling yourself such as online courses in data science, conferences, and many others. You should get used to problem-solving and coding as much as you can in order to learn how to apply data science in problems.","excerpt":"In the present scenario, data science jobs are the hottest trend. According to our recent report, there has been an overall growth in the number of jobs in analytics and data science ecosystem with India contributing to 6% of open job openings worldwide. The total number of analytics and data science job positions available are […]","categories":["AI Trends"],"tags":["Data Science Career","data science career path","data science career tips","datascience"],"author_name":"Ambika Choudhury","publish_date":"2019-04-23T12:03:58","publication_year":"2019","word_count":537,"keywords":["data science","Go","artificial intelligence","machine learning","AI","datascience","Data Science Career","Python","analytics","SQL","Julia","data science career tips","R","data science career path"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Python","R","SQL","Go","Julia"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-things-to-know-before-you-start-your-data-science-career\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":37440,"title":"What&#8217;s Trending: Learning-Based Algorithms Now Offer New Ways To Find Out What&#8217;s Popular","content":"Still from Netflix show American Meme (2018) Google releases the most popular and trending word or celebrity each year. It counts the number of times a word has been searched for or how many times a word appeared in a string of searches. Google alone handles more than 40,000 search queries per second which, over a year, would be a data storage nightmare. But thanks to improved hardware and efficient cooling systems, data is intact in the vast data centres. In the early days of the internet, two decades ago, identifying the most searched entity was not an easy job. Hashing was introduced by the researchers at AT&T, a statistical technique which transforms data into random value uniformly distributed over some interval. Two similar elements will be stored in the same bin. The distinction between the elements is estimated from the probability that a given bin remains empty. There have been many innovations to finding the frequencies of items in a stream of data. With increasing internet usage, the data generated too has increased. Back in the day, data might have been an area of concern for a few entities. Now, data is being used for almost all purposes from recommending movies to predicting presidential candidate from sentiment analysis on Twitter. The commonality in these applications is the continuous stream of data that is being exchanged back and forth. To analyse this data and to make sense of it in real time, is an arduous task. To address this issue, the researchers at MIT’s CSAIL, developed a method which uses machine learning models to make guesses about the frequency of the data. This is the first of its kind streaming algorithms which employ machine learning. The main objective behind the proposal of these new class of algorithms is to keep track of the frequencies of certain items in a stream of data. These items can be trending topics on Twitter or popular searches on Google. The researchers call this method “LearnedSketch”, which is trained on 210 million data packets from a Tier 1 ISP and run on 3.8 million AOL queries. The main focus here was on a class of hashing based algorithms like Count-Min and Count-Sketch. The team define what they call as a Heavy-hitter oracle within the subsets of the bins that store these frequently appearing items from packets of data streamed. This Source: MIT CSAIL The above scatter plot is a visualisation of the embedding space learned by the model on Internet traffic data, where each point represents one network flow. For example, if there is a word say, “machine learning”, that appears frequently in the dataset. The model can be made to remember such popular words for predicting. But, there are always other factors that contribute to the popularity. Model Construction For Search Query Estimation: A neural network is trained to predict the number of times a search phrase appears(AOL queries) To process this, a recurrent neural network(RNN) is trained with LSTM which takes these search phrases as inputs. The character IDs(lower or upper case, punctuation marks etc) are mapped to embedding vectors before being fed to RNN The final states encoded by the RNN are fed into a fully-connected layer to predict the frequency of a certain search phrase. Key Takeaways Learning based algorithms reduced estimation errors by 18% to 71% when compared to non-learning based algorithms like Count Min. This model can identify items which go unnoticed with conventional methods. A full-length application would enable the users to follow spikes in web-traffic and exploit these results say, for improving e-commerce sites and other customer based recommendation platforms. Read more about the experiment here","excerpt":"Google releases the most popular and trending word or celebrity each year. It counts the number of times a word has been searched for or how many times a word appeared in a string of searches. Google alone handles more than 40,000 search queries per second which, over a year, would be a data storage […]","categories":["AI Trends"],"tags":["Machine Learning","MIT","RNN"],"author_name":"Ram Sagar","publish_date":"2019-04-09T06:57:50","publication_year":"2019","word_count":609,"keywords":["Go","machine learning","AI","neural network","MIT","ML","sentiment analysis","Machine Learning","RAG","RNN","LSTM","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","RAG","sentiment analysis","R","Go","RNN","LSTM"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/whats-trending-learning-based-algorithms-now-offer-new-ways-to-find-out-whats-popular\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":19770,"title":"The Amazing Way In Which HT Media Utilises Analytics For Success","content":"While the importance of analytics has been sweeping into all the major industries, media remains one of its prominent users. Whether it is for creating a connection with the viewers, predicting audience interest or content monetisation, media companies are vouching big time on analytics and big data. Analytics India Magazine caught up with Amit Gupta, Head, Strategy and New Revenue Development, Hindustan Times, one of the leading media companies in the country today, to understand how analytics play a role at the company. Adorning 11 years of varied experience in corporate strategy, revenue management, branding, content strategy and more, he believes that data analytics is becoming far more integral to business operations across most industries and this holds true equally, perhaps more for media and entertainment organisations. Analytics changing the media and entertainment landscape According to Gupta, when it comes to media and entertainment sector, data analytics adds immense value at both sides of its value chain— audience at one end and business partners at the other end. “On the audience side as an industry, it’s going through a paradigm shift with the burgeoning options that a consumer has— be it in terms of the mediums of consumption (print, digital, OTT, mobile, OOH, radio, cinema, live streaming etc. to name some) and multiplicity of options within each of them”, he said. So how is media industry evolving in terms of adoption of analytics? “If, as an organisation, one were to focus on having a ‘loyal audience’, it needs a shift in focus from from eye-balls to conversation, from acquisition to engagement, from quantum to involvement etc. etc. This leans back heavily on delivery of high-quality content and that too on a sustainable basis. Thus, consumer analytics is emerging as a critical strategy component to efficiently develop and deliver the content and then monetise the ‘engaged audience’”, he explained. Analytics is therefore impacting business models on monetisation side significantly. “Changing media mix and focus on building engaged communities has resulted in planning and delivery of bespoke solutions built on a brand’s communication objectives. These would again lean quite heavily on insight mining and data analytics across an array of variables in the consumer’s content consumption journey– quite a transition it’s looking to be”, he said. How has HT been benefited by adopting analytics? Gupta believes that the role analytics within Hindustan Times is continually evolving as would be the case with most organisations. Monetisation is one of the vastly affected areas by the use of analytics at HT. Looking from a monetisation standpoint, it is being leveraged through many interventions which he explained as below: Analytics in the manner of descriptions through dashboards for the sales function is helping them identify and leverage meaningful insights from a whole lot of data without mining into it. It also helps in establishing operating KPIs and act as tools for performance updates, tracking and monitoring etc. It enables customer segmentation through various stages of growth and thus predicts the corresponding impact they are likely to result in. Aid revenue strategy including pricing through inputs beyond just transactional information. “Like most organisations, we are also expanding and growing the influence that analytics can have towards making a meaningful business impact”, he added candidly. Impact on revenue management and pricing strategy Is analytics changing the revenue management and pricing strategy in media industry? “A whole lot, I would say”, he quickly noted. Given the nature of the industry, coupled with the current economic scenario, there is a whole lot that analytics can help in, for example, churn analysis with a significant share of new advertisers every year, ever evolving mix with varying consumption of inventories sought or different segments of advertisers being active, evolving product configurations with a multitude of options to choose from, and many more. “Analytics is being leveraged by quite a few organisations to really harness the benefits from an abundant and rich data pool that is getting augmented day after day. Advertising solutions are also gaining complexity– ranging from linear transactions to custom solutions, from a single window to a long term arrangement and so on. Moreover, there is more to be extracted by establishing robust patterns on demand vs. supply for various inventories, to not only assess suitable pricing but more importantly, aid customer satisfaction”, he said. The rise of AI, ML and chatbots in HT When it comes to the monetisation standpoint in the media industry, Gupta believes that it is pretty early days, but there surely can be value added by technologies like machine learning and chatbots. “However, there is much to be covered for these technologies to start making a material business impact”, he said. As HT media is still looking forward to adopting these technologies, he says that one of the key challenges that needs to be addresses for these to truly yield a benefit is to have a structured and integrated data set right from the time the field team prospects an opportunity, or even before that, when the editorial team captures it as a news. It should also follow through the various stages of sales journey till the transaction gets executed, followed by feedback and tracking inputs, if required. “This remains a significant challenge given the host of formats of data capture— notepads to excel files to CRM inputs etc. coupled with an interaction between many systems. If this can be overcome in a meaningful fashion, there is a lot of value creation that such emerging technologies can result in”, he added. The analytics roadmap at HT Media As the company is constantly evolving in terms of making countless decisions based on data, he believes that data without context is quite meaningless and hence is only as valuable as the team looking at it. With a strong management support and a well evolved structure integrated with the revenue leadership, the analytics team as HT media is focusing on many task ranging from descriptive activities to green shoots. “There are massive growth opportunities and business enhancements that can be made: category development, a mash-up between advertising patterns and reader consumption patterns, dynamic inventory management, pricing solutions, product development opportunities etc.”, he noted. According to Gupta, the key remains to deliver business impact through the investments being made and identifying the right talent for making the most of the data at hand. “We’ve made some strides in our journey so far. However, as we continually leverage the richness of data generated, there is a whole lot to do still”, he said. Deterrents on the way of maximising data usage Governance of data quality, establishing ROI and generating relevant insights have been the most significant challenges not just for HT but for other organisations who are keen on leveraging data for better benefit. He further explained these points as below: Strong data quality governance: It involves managing a single source of truth through technology to integrate multiple sources and multiple formats of data inputs. That also means making sure that adequate and skilled resources are available to manage this on a continuous basis under the guidance of sharply defined governance principles to aid data quality. While reams of data may be analysed and churned through, the key remains delivering a meaningful insight to the user group. It has to result in an ‘aha’ moment for them, it’s got to be actionable, impactful and practical. Establishing ROI: No matter what, any and every penny invested into business has to pass this litmus test and so remains true to analytics. It cannot be a domain that continues to attract investment without any tangible benefits accrued through it. “Needless to say, this is all meaningful with having the necessary leadership alignment and support as also the talent pool available to progress this”, points out Gupta. Endnote Gupta believes that over the years,  analytics industry has evolved from a being data to insights phase, to data predictions and finally generating actions based on these insights. “Gone are the days when stating what is likely to happen was an accepted deliverable; the need now is “So, what do we do?”. Analytics is fast progressing to generating business impact (direct or soft), and it’s got to be value accretive— be it by virtue of product enhancements, process enhancements, efficacy enhancements or a combination of them”, he says in the concluding.","excerpt":"While the importance of analytics has been sweeping into all the major industries, media remains one of its prominent users. Whether it is for creating a connection with the viewers, predicting audience interest or content monetisation, media companies are vouching big time on analytics and big data. Analytics India Magazine caught up with Amit Gupta, […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2017-12-14T09:27:37","publication_year":"2017","word_count":1386,"keywords":["Go","machine learning","AI","chatbots","ML","Git","RAG","Ray","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Ray","RAG","chatbots","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/amazing-way-ht-media-utilises-analytics-success\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":17173,"title":"Will AI Make Artists Redundant? Evolution From MS Paint To Creative Adversarial Network Shows A Powerful Change","content":"(Representational image) Artwork by Johan Scherft, titled ‘Robot, painting a self-portrait’. Most of us viewed technology as an accessory for humans to conduct their jobs better and making it a ladder to efficiency. With advanced artificial intelligence (AI) and machine learning, we would give up our mundane, repetitive tasks to focus on the creative side. But looks like AI is entering the realm of art too — something which was believed exclusive to humans and emotions. Will artists will be in the next line to lose their jobs to machines? Here’s how the progression took place: Tool Maketh The Art: Most millennials’ first memory of using tools to create anything resembling to “art” was 32-year-old now deprecated Microsoft tool, Paint. From using it as a doodler, to fixing image sizes, the programme, which generated feeds full of hue and cry over its so-called retirement, has been a worthy sidekick for many budding artists. It is still used by many illustrators as a nostalgic ode. Newer designing tools such as Adobe Illustrator, CorelDRAW, ArtRage 5 and Paper by FiftyThree, among others, have almost made using physical drawing or sketching tools redundant. App For The Soul: As if the extreme sepia tones of Instagram filters were not enough, newer apps such as Prisma and its spawning websites are turning every photograph into a work of art. Created by Aram Airapetyan and co-founder and CEO Alexy Moiseenkov, the app got 85 million downloads within six months after it was released. Experts have pointed out that the app uses Convolutional Neural Network to learn the low-level style of an photograph, and applies it to the uploaded image and then processes images on the cloud server. Prisma just launched a deep learning API and SDK ️ https:\/\/t.co\/ISGQxUOiDY pic.twitter.com\/XeZh2yvnDI — Product Hunt (@ProductHunt) August 22, 2017 Programme Learneth The Formula: Van Gogh’s bold strokes, Michelangelo’s sharp, intuitive carvings are no longer exclusive to the masters. With the correct programming and with the help of the right AI algorithms, several systems have learnt to automatically create art, not only in the domain of AI and computational creativity, but also in computer graphics and machine learning. For example, Google has made public its many AI experiments such as Quick Draw and Auto Draw tools that turn human doodles into artwork. Project Magenta from the Google Brain team advances the state of art through machine learning algorithms. Though it is still under progress, the aim is to make the programme the best machine learning platform for art and music, and connect artists with technology. In fact, so impressed are the artists themselves with the “learning” capability of these programmes, that artists like Mario Klingemann are using it to invoke startling results. Klingemann, for example, builds art-generating software by feeding photos, video, and line drawings into code, and then curates collections of hauntingly distorted abstract images — faces and figures, for his followers. Made some subtle improvements to the algorithm. pic.twitter.com\/kcXg3nXix1 — Mario Klingemann (@quasimondo) August 20, 2017 Man Wonders At Bot’s Creation: Now, researchers from Rutgers University, College of Charleston, and Facebook’s AI Research Lab have created an algorithm that allows AI to create an artwork that human experts cannot distinguish the difference between works of man and machine. The researchers have made a Creative Adversarial Network (CAN) where the AI creates images that the discriminator recognises as art, but cannot categorise into an established style, thus creating a unique, non-human artwork. “If we teach the machine about art and art styles and force it to generate novel images that do not follow established styles, what would it generate? Would it generate something that is aesthetically appealing to humans? Would that be considered ‘art’,” asked lead author Dr Ahmed Elgammal of in a blog post.","excerpt":"Most of us viewed technology as an accessory for humans to conduct their jobs better and making it a ladder to efficiency. With advanced artificial intelligence (AI) and machine learning, we would give up our mundane, repetitive tasks to focus on the creative side. But looks like AI is entering the realm of art […]","categories":["IT Services"],"tags":["art","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2017-08-23T10:18:40","publication_year":"2017","word_count":626,"keywords":["Go","API","artificial intelligence","machine learning","AI","neural network","Machine Learning","art","RAG","Aim","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","Aim","RAG","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/will-ai-make-artists-redundant-evolution-ms-paint-creative-adversarial-network-shows-powerful-change\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10067479,"title":"DeepMind’s AlphaFold 2 is half of the story","content":"Last year, Google’s research arm DeepMind released an open-source version of its deep-learning neural network AlphaFold 2, which essentially solved the 50-year-old grand challenge problem of protein folding. This computational method achieved higher prediction performance than much more expensive experimental methods like X-ray crystallography. What’s the big deal? As shown below, DeepMind’s AlphaFold 2 algorithm significantly outperformed other teams at the CASP14 protein-folding contest – and its previous version’s performance at the last CASP. In 2018, its previous version achieved a score of 58 on the hardest class of proteins. The second generation of AlphaFold achieved a score of 87, which is a huge improvement – 26 points better than the closest competition. For those unaware, a score above 90 is considered roughly equivalent to the experimentally determined structure. But, there is more “Protein folding is just half of the story,” said Vikas K. Garg, co-founder and chief scientist at YaiYai. An IISc and MIT (PhD) alumnus, Garg is one of the lead researchers who worked on developing ‘Generative models for graph-based protein design,’ which served as one of the references to shape DeepMind’s AlphaFold. His firm YaiYai, which he co-founded alongside Tamar Pichkhadze, provides AI-based solutions in biopharma, energy, edtech, gaming, and fintech sectors to startups and governments, as well as leading companies across the globe. Let’s get real The human body is almost entirely made up of protein, besides water and fat. The folding of proteins is the underlying cause of many diseases. Understanding these protein folding, protein designs, etc., helps pave the way for finding a cure, designing new medicines, pharmaceutical solutions, drugs, etc. Last two years, because of the Covid-19 pandemic, the methodologies or techniques, such as AlphaFold, have been gaining renewed focus. Here’s how it works People get diseases, maybe cold or fever or Covid-19. So we have these protein targets in our bodies, and there are different kinds of proteins throughout our bodies, in our cells. So, when we are trying to give them a tablet or a drug, we identify certain specific proteins in the human body, where we want these molecules to go and bind with these proteins. “You should think of it like how you would open a lock. So, think of protein as a lock and your drug or a molecule as a key,” explained Garg. He said if you do not open the right lock, there is a risk that you will open something else, which will lead to side effects. To minimise the side effects, and at the same time, you want to cure someone’s health – what you would do is, administer this protein, which is in the 3d objects or shape. Think of it like these balls (proteins) coming together. Once you find the right key to the lock, the activity is restored. In other words, the drug has the effect of restoring the normal functionality of the human body. Simply put, there are two parts to this process. First is the target identification of proteins we want to target for a particular disease. The second is finding the right drugs that can have the desired therapeutic effect—for example, Remdesivir injection in treating coronavirus. “So, you want these drugs to technically open every lock. For that, you need just the right amount of the right key. You don’t want to open too many locks. You want to open only the lock you are looking for,” said Garg. Here’s how you do it There are two ways, either you could design new proteins (new locks) and find a cure (key), or identify new proteins and inject molecules. Why is AlphaFold half the story? “Now, when we talk about AlphaFold, for example, and protein fold–this is only the half of the story, probably a quarter of the story,” stressed Garg. He said that proteins are very complex structures. There are lots of properties that have evolved over centuries or millennia. For example, some of the proteins are known to be more biologically stable compared to others. “The space where we are operating is humongous,” he added, saying that it is really hard to find the right drugs or generate new proteins. “This is where the fun part lies,” said Garg, “This is also where I want to emphasise why quantum holds so much promise. Of course, AI has been making very rapid progress. But, quantum gives you this flexibility.” If done well, they have tremendous potential because they will be able to search through this humongous space very quickly. “Much more quickly than any classical methods today,” said Garg. Throwing light on AlphaFold, Garg said: “Proteins are very compact structures.” For instance, each protein can be viewed as a sequence of amino acids. There are 20 amino acids in our body. This includes histidine, isoleucine, methionine, leucine, lysine, phenylalanine, etc. “Think of it as a chain, where at each location, and each position in that chain, you are labelling it as one of these 20 amino acids,” explained Garg, saying that this is a very complex problem. “Now, imagine, 20 raised to the power of 1000 possibilities,” added Garg, underlining the complexity of solving this problem. Inverse protein folding “The idea was if I give you a sequence of amino acids, can you predict what will be the structure or the shape that it will take in the 3D space?” said Garg. He said that this is a problem AlphaFold solved using deep learning. “Of course, it is a useful thing, but it is not the end goal,” he added, saying that it is just a part of the process. “The end goal is that you want to design new proteins,” said Garg, proposing an inverse protein folding technique. “I will give you the shape of the lock, but I will not tell you which is the right key to open the lock. Now, the idea is can I map the 3D structure to the sequence,” said Garg. How does it work? “We created these 3D structures, where we used very advanced deep learning methods to predict which amino acids or new proteins can be generated. Because we already have the structure, you can fill in the amino acids at each position, and you get a new protein,” said Garg. He said that the potential of inverse protein folding is immense; you can design therapeutics, new materials, synthesise new batteries for electric vehicles, etc.","excerpt":"The idea was if I give you a sequence of amino acids, can you predict what will be the structure or the shape that it will take in the 3D space?","categories":["AI Features"],"tags":["DeepMind","DeepMind Alphafold"],"author_name":"Amit Naik","publish_date":"2022-05-20T10:32:08","publication_year":"2022","word_count":1066,"keywords":["Go","API","startup","programming_languages:R","AI","neural network","Ray","deep learning","ViT","R","DeepMind Alphafold","DeepMind"],"extracted_tech_keywords":["AI","deep learning","neural network","Ray","R","Go","API","ViT","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deepminds-alphafold-2-is-half-of-the-story\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":554,"title":"Interview &#8211; Rahul Deshmukh, Director &#8211; Web Intelligence at Splunk","content":"[frame align=”left”] [\/frame]Fourth in the series, Analytics India Magazine talks with Rahul Deshmukh on how Splunk is helping organizations drive value from their machine data. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: Could you tell us more about the topic that you are speaking at The Fifth Elephant? [dropcap style=”1″ size=”2″]RD[\/dropcap]Rahul Deshmukh: Splunk’s mission statement is “Make machine data accessible, usable and valuable to everyone.”  , Splunk is the leading provider of operational intelligence software used to collect and harness massive machine data to provide new levels of visibility and insight for IT and the business. Typically users start using Splunk to search and investigate their machine data – to find and fix problems dramatically faster. As they find causes of issues, problems and attacks, they start monitoring for these automatically. They then use Splunk to gain end-to-end visibility to track and deliver on IT KPIs and make better-informed IT decisions. They then use Splunk to deliver real-time insight from operational data to make better-informed business decisions. [pullquote align=”left”]Splunk is a freemium model – for the first 500 MB of daily indexed data, Splunk is Free. Splunk offers both an on-premise and cloud version. Splunk integrates with Hadoop and helps manage Hadoop deployments. For more information visit: www.splunk.com[\/pullquote] AIM: How was Splunk incepted, how has it evolved over years and what is the next step? RD: Splunk started in 2004, with the first release on 2006.  Splunk started with solving IT issues and in the earlier days was called “Google for IT data”.  Over the years, customers realized the value in the machine data and started using Splunk from troubleshooting their applications to proactively monitoring their business. Splunk is software that collects machine-generated data from virtually any source. It indexes the data, aggregating it so you can search and query the data from one place in real-time  Currently, 54 of the Fortune 100 companies use Splunk, and there are over 4000 paid customers that use Splunk. AIM: Splunk specializes in so-called machine data. Could you tell us what you mean by that? RD: Machine data is non-standard, highly diverse, dynamic and high volume. You will notice that machine data events are also typically time-stamped – it is time-series data. AIM: Why is making machine data accessible to organization “disruptive” according to Splunk? RD: Splunk is making machine data or machine generated data accessible, usable and valuable to everyone.  In the past, this data would be with the select group of individuals who had to use things like “grep” to troubleshoot their applications or resolve security related issues.  Splunk has provided collection, indexing and visualization of this data an easy to use interface. AIM: What according to is future of analyzing machine data? RD: At Splunk, we observe the customers derive value from their machine data in a variety of use cases.  Beyond analyzing digital (web, mobile, social), Splunk’s customers are using this data for Manufacturing, energy, Healthcare sectors.  Extending the data using SDKs or API as well as Big Data are some of the new areas for machine data. [divider top=”1″] [spoiler title=”Biography of Rahul Deshmukh” open=”1″ style=”2″] Rahul has a unique blend of Web Analytics strategy, Business Intelligence and marketing experience. He is currently Director – Web Intelligence at Splunk. Previously, he was Sr. Manger -Search Analytics at eBay. His team worked on analytics and optimization for Product Based Experience, Catalogs, Fitment, Classification and Advertising. Prior to eBay, Rahul was an Independent Web Strategy consultant focused on Social Media. His past experience includes Director of Analytics and Optimization for Ask.com and Dell. Rahul serves on the Client Advisory board for Hitwise and has spoken at Google and eMetrics conference.[\/spoiler]","excerpt":"[frame align=”left”] [\/frame]Fourth in the series, Analytics India Magazine talks with Rahul Deshmukh on how Splunk is helping organizations drive value from their machine data. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: Could you tell us more about the topic that you are speaking at The Fifth Elephant? [dropcap style=”1″ size=”2″]RD[\/dropcap]Rahul Deshmukh: Splunk’s mission statement is “Make […]","categories":["AI Features"],"tags":["hadoop data catalog","Interviews and Discussions"],"author_name":"Дарья","publish_date":"2012-07-25T13:14:33","publication_year":"2012","word_count":608,"keywords":["big data","Go","API","business intelligence","AI","Git","Aim","analytics","GAN","hadoop data catalog","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","Git","API","big data","GAN","business intelligence"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-rahul-deshmukh-director-web-intelligence-at-splunk\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":16079,"title":"“Genpact Cora”- AI based self learning platform by Genpact makes debut","content":"Genpact announced the launch of Genpact Cora—an artificial intelligence based platform that accelerates digital transformation for enterprises. A modular, interconnected mesh of flexible digital technologies, Genpact Cora claims to hone specific operational business challenges and tackle them from beginning to end. It thus helps large global companies reframe and solve their most pressing real world business issues quite conveniently. “Genpact Cora is timely for an industry seeking digital transformation,” stated Peter Bendor-Samuel, founder and chief executive officer, Everest Group, a leading analyst firm. As the company strives to drive the digital-led innovation and digitally-enabled intelligent operations for clients around the world, Genpact Cora provides the fastest path in driving meaningful transformations at larger scale. It brings digital transformation in a planned and managed fashion, without sacrificing the governance security and investment protection that mature and established businesses need. NV ‘Tiger’ Tyagarajan, president and chief executive officer, Genpact  said “Achieving enterprise impact from digital transformation is challenging with so many disparate, disconnected technologies in the market.” “Genpact Cora brings leading digital solutions together in one unified platform, combined with the process and deep domain expertise that comes from decades of experience running intelligent operations. The combined benefit creates connected intelligence for our clients at a previously unattainable level of agility and speed to predictive insight, that then drives outcomes”, he added. With a mature application program interface (API) design and open architecture including Genpact’s own intellectual property, it integrates advanced technology across three key areas— Digital Core including cloud, software-as-a-service, blockchain, mobility and ambient computing, robotic process automation, and dynamic workflow; Data Analytics including advanced visualization, data engineering, big data, and Internet of Things (IoT); Artificial Intelligence including conversational AI, computational linguistics, computer vision, machine learning and data science AI. The Genpact Cora platform is the foundation for Genpact products and consulting services already in the market. It brings together Genpact’s original process and industry domain depth with new digital capabilities.  Manufacturing, Pharmaceuticals, finance and insurance are some of the industries making the most of Genpact Cora. “Genpact Cora reduces risks around errant robots and misapplied AI spinning out of control, through an integrated command and control hub that delivers the much-needed governance that business processes require”, said Sanjay Srivastava, senior vice president and chief digital officer, Genpact on a concluding note.","excerpt":"Genpact announced the launch of Genpact Cora—an artificial intelligence based platform that accelerates digital transformation for enterprises. A modular, interconnected mesh of flexible digital technologies, Genpact Cora claims to hone specific operational business challenges and tackle them from beginning to end. It thus helps large global companies reframe and solve their most pressing real world […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-07-05T03:55:32","publication_year":"2017","word_count":381,"keywords":["data science","Go","machine learning","artificial intelligence","AI","Git","computer vision","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","data science","analytics","Aim","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/genpact-cora-ai-based-self-learning-platform-genpact-makes-debut\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":14610,"title":"IBM Bluemix &#038; BlockSmiths Host Blockchain Conference | 5th May 2017 | Gurgaon","content":"IBM Bluemix together with Blocksmiths is set to host a Blockchain event on 5th May 2017 at Blocksmiths’ Gurgaon facility to help industry leaders, journalists and startup founders to understand the blockchain technology and its applications in various domains. Blocksmiths is a Gurgaon(India) based blockchain consultancy that helps organisations from pilot and prototype deployment to the full scale implementation of blockchain based solutions as per their particular business requirements. IBM being the industry leaders in providing services and solutions to businesses has been aggressively exploring blockchain through its Hyperledger platform which specialises in providing private\/permissioned blockchian for enterprises that want to try and implement the technology within the organisation and its associated companies. The agenda for the event is to make the CxOs and founders aware of the plethora of cloud based solutions that IBM Bluemix has to offer and to stir their minds to innovate by leveraging the blockchain technology. Attendees at the event will also have the opportunity to interact with technical experts from IBM and the senior management from BlockSmiths and understand how blockchain can revolutionise their particular industry. Given the current interest in blockchain by many countries and companies primarily because it can provide a wide range of solutions across a wide array of industries such as insurance, banking, manufacturing, IoT, cybersecurity and many others, the event will be a great fit for those business leaders and company founders that wish to address business inefficiency issues by experimenting with new tech. Interested attendees can either send a request for an invitation at events@blocksmiths.io or fill up the Registration Form","excerpt":"IBM Bluemix together with Blocksmiths is set to host a Blockchain event on 5th May 2017 at Blocksmiths’ Gurgaon facility to help industry leaders, journalists and startup founders to understand the blockchain technology and its applications in various domains. Blocksmiths is a Gurgaon(India) based blockchain consultancy that helps organisations from pilot and prototype deployment to […]","categories":["Deep Tech"],"tags":[],"author_name":"AIM Media House","publish_date":"2017-04-27T10:15:50","publication_year":"2017","word_count":263,"keywords":["programming_languages:R","AI","RAG","Ray","ViT","GAN","R","startup"],"extracted_tech_keywords":["AI","Ray","RAG","R","GAN","ViT","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ibm-bluemix-blocksmiths-host-blockchain-conference-5th-may-2017-gurgaon\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042723,"title":"The Linux Foundation Launches New License Agreement For Open-Source Datasets","content":"The Linux Foundation has announced the release of the new license agreement, CDLA-Permissive-2.0. It’s a short, simple, and broad agreement to enable wider and more open use of data, particularly with respect to artificial intelligence and machine learning models. IBM and Microsoft have already made their datasets available under CDLA-Permissive 2.0. CDLA-Permissive 2.0 In October 2017, The Linux Foundation launched CDLA 1.0, delineating the terms and conditions to share and modify data. The licence agreement also allowed the use of insights from the analysed data to create AI and machine learning models without having to share the data itself. The overly complicated nature of the permission agreement called for a simpler version 2.0. To wit, the new agreement is less than a page long. Unlike other licensing agreements such as MIT or BSD-2-Clause licenses, CDLA 2.0 doesn’t insist on attribution for the data source. However, v2.0 mandates that the license agreement should be included with the dataset. Further, CDLA-permissive-2.0 removes some of the confusing terms, making the document concise and easy to understand, especially for data scientists who are not well-versed in English. Under this agreement, Microsoft has made the following datasets available: The Hippocorpus dataset, which contains diary-like short stories of recalled and imagined events.The Public Perception of Artificial Intelligence dataset which contains an analysis of the text corpora revealing trends in beliefs, interest, and sentiment around a topic.The Xbox Avatars Descriptions dataset of avatars created by actual gamers.A Dual Word Embeddings dataset trained on Bing queried for document information retrieval.A GPS trajectory dataset with 17,621 trajectories with a total distance of 1.2 million kilometres and a duration of 48,000+ hours Further, IBM’s Project CodeNet data set would be the first to carry the CDLA-Permissive-2.0 license. This is made available through IBM Data Asset eXchange (DAX). It consists of 14 million code samples, 500 million lines of code, in over 55 different programming languages. The purpose of Project CodeNet is to train AI models to understand and write code. Why is such an agreement required? Today, we have multiple open-source software licenses that work within the frameworks established by laws and regulations for copyrights, patents, etc. The licenses allow for the broad use, modification, and sharing of software and other copyrightable content. However, open data is different from open software. The copyright protection and permits differ from country to country and may be subject to different laws specific to databases. Unlike software and other creative content, data is consumed, transformed, and incorporated as per the requirement of the particular artificial intelligence and machine learning model. “Data is different from software. Open-source software is typically made of copyrightable works, where authorship is important. By contrast, data may frequently have little or no applicable intellectual property rights, and authorship and attribution are often less important,” said The Linux Foundation compliance and legal VP Steve Winslow. Ipso facto, the commonly-used licenses for software and creative content cannot cover open data. The aim of Linux’s CDLA agreement is to focus on addressing concerns arising out of AI and ML use cases.","excerpt":"The Linux Foundation has announced the release of the new license agreement, CDLA-Permissive-2.0. It’s a short, simple, and broad agreement to enable wider and more open use of data, particularly with respect to artificial intelligence and machine learning models. IBM and Microsoft have already made their datasets available under CDLA-Permissive 2.0. CDLA-Permissive 2.0 In October […]","categories":["AI Features"],"tags":["open data"],"author_name":"Shraddha Goled","publish_date":"2021-07-01T14:00:00","publication_year":"2021","word_count":508,"keywords":["machine learning","artificial intelligence","open data","AI","AWS","cloud_platforms:AWS","ML","programming_languages:R","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","AWS","R","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-linux-foundation-launches-new-license-agreement-for-open-source-datasets\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10079854,"title":"India’s Semiconductor Mission is the Hot Cake No One Wants to Have","content":"The revisions made to India’s existing semiconductor policy have made it ripe for investors to build a semiconductor ecosystem within India, but only a few players have chosen to avail of the scheme. Analytics India Magazine spoke to Sravan Kundojjala, director at Strategy Analytics, to understand why companies who have the capital to invest, even with the incentives provided, are hesitant to take the long stride into semiconductor manufacturing in India. Kundojjala attributed this to a variety of factors. First and foremost, setting up a semiconductor plant requires huge capital expenditure, while the profitability is very low, at least during the initial period. This, he believes, is due to the fact that “it takes time for companies to get to the break-even point, and ramp up the number of wafer starts per month.” Naveed Sherwani, chairman, president & CEO at RapidSilicon, explained two things about setting up a fab—one, once a fab is started it can never be stopped, and second, it needs 80-85% capacity to make any money. Strong global competition Additionally, the demand is very cyclical in the industry. And to survive, the financials have to be very strong, especially when there is ample global competition looming around. Companies like Semiconductor Manufacturing International Corporation (SMIC), GlobalFoundries, and United Microelectronics Corporation (UMC) are pushing for expansion in countries like France, US, Singapore, Taiwan, and Japan— further stifling the market share that India can obtain if at all it gets its feet up and running by 2025, or 2026. UMC and Taiwan Semiconductor Manufacturing Company (TSMC), for instance, are deploying resources for the production of 28nm chips in Singapore and China, respectively, whereas GlobalFoundries are investing in Europe in the 22FDX range. The 28nm range has been a hot item in the semiconductor manufacturing space. It is used for a wide set of applications, including WiFi chips, Bluetooth chips, camera image sensor, display driver ICs, or management chips, RF transceivers, and sensors like fingerprint sensors, MEMS, etc, and has also been the target of applicants availing the government scheme on semiconductor manufacturing in India. However, the concern, as Kundojjala admitted, is will India be able to build a solid base, and get the factories running fully loaded, in the initial stage, when there is heavy competition? Arun Mampazhy, a veteran semiconductor analyst, in addressing what the foreign expansion in 28nm range means for India, said this: Earlier UMC said 28nm may have over capacity after 2023. If T$MC also switches to \"two 28nm fabs\" plan at Kaohsiung instead of \"7nm + 28nm\", it is bad news for India's 28nm plans. IMO, 65nm Analog remains a safer bet even if \"experts\" advise @GoI_MeitY @AshwiniVaishnaw otherwise. https:\/\/t.co\/lZwZHPBdjl— arun mampazhy (@nano_arun) November 14, 2022 However, the government can make certain policy interventions to give the semiconductor manufacturers an edge in the Indian market. “As of now, as long as companies in India register in CHIMS (Chips Import Monitoring System), the import of chips won’t have a custom duty. But later, the government can levy customs duty on the import of those chips whose requirements can be filled by fabs in India,” Mapazhy told AIM. Robust government backing needed In the semiconductor industry, the government is the backbone, incentivising a lot of investments and getting the supply chain in order. A country like India in particular, which is yet to walk through the semiconductor waters (with its ebb and flow), needs strong government backing to get the process running at least during the initial stretch, when companies lose a lot of money. In this regard, Kundojjala said, “We shouldn’t look at this [India’s] investment as consisting of one or two years, it should be there at least for 10-15 years to have things play in favour of you.” He also cites the example of the Taiwan government which offered a lot of incentives to TSMC and EMC when they initially started, providing infrastructure and human resource support. These companies, for instance, sent their engineers to the US for training, who later came back to set up these foundries. When asked about what should be the state of subsidies going forward, Kundojjala said that the allowance should ideally be raised if we are looking to get more into more cutting-edge technology with a 28nm process. However, for the 45nm, 55nm process, which is considered to be the sweet spot currently, and will be so at least until 2030, considering the number of applications in areas of power electronics, and periphery modules inside numerous consumer applications that we use currently, the present allowance should work fine, provided the government is ready to taken the burden of loss during the beginning phase. Texas Instruments is a big name targeting between 45nm and 130nm nodes, spending about $24 billion on four fabs, each with a capex of about $6 billion. But, what if the nodes that are currently above 28nm migrate to the advanced category—that is, with the growing pace of technology, how is it guaranteed that technologies that currently need chips of 28nm and above would not move to 16nm nodes tomorrow? In response to this, Kundojjala said that in analog chips, you cannot scale well beyond a certain node. At this stage, even increasing the transistor density will not impact components such as performance, power consumption, savings, etc. This further attests to the fact that for at least a decade, there would be a market for mature nodes which India can utilise and sow the seeds for future investments for production of advanced chips used in defence, AI, and other developing technology. Reality check Semiconductor consumption in India is expected to grow to roughly $70 billion by 2026, said India Electronics and Semiconductor Association (IESA) chairman, Vivek Tyagi. In this light, the Government of India’s semiconductor mission is aligned with building a robust semiconductor supply chain to enable India’s emergence as a global hub for electronics manufacturing and design. But, in reality, there is still no sense of urgency seen on ground in the government’s actions. Union Minister Ashwini Vaishnaw, earlier in April, said that the approval for making electronic chips will likely be done in the next 6 to 8 months. However, at a recent Bengaluru Tech Summit. Karnataka’s IT Minister CN Ashwath Narayan clarified that the central government would decide on which company would build a semiconductor plant, and which state would house it by February 2023. The statement brought disappointment to many who were anticipating the government to take swift actions on the proposals received. A Twitter user by the name ‘EkNashwar’ expressed frustration with the present state of affairs saying this: It's a pathetic bureaucratic show of delay from @GoI_MeitY @AshwiniVaishnaw @Rajeev_GoI where just approval will take 12+ months, just when semiconductor glut is being predicted. During the same period, US approved and funded $280 billion over 10yrs, so did EU, SoKo & Japan.— एक नश्वर (@EkNashwar) November 8, 2022 India’s major advantage, as Kundojjala reiterates, is that it is politically neutral in the global scheme of things. At a time when China is cut off from advanced AI equipment and AI processors due to the geopolitical battle it is engaged in, building a strong base to strengthen the ecosystem becomes paramount. Japan is already doing this. According to reports, eight major Japanese tech firms, including Sony and Toyota, have teamed up to form a consortium by the name Rapidus, aimed towards mass-producing advanced 2nm chips by 2027. The Japanese government has vowed 70 billion yen (~$500 million) to back Rapidus. It is time the Indian government recognises this and takes prompt steps to make India a strong nodal point for global semiconductor supply—and for this, mature node is just the start.","excerpt":"Why companies that have the capital to invest, even with the incentives provided, are hesitant to take the long stride into semiconductor manufacturing in India?","categories":["IT Services"],"tags":["GlobalFoundries","India semiconductor mission","MeitY","tsmc"],"author_name":"Ayush Jain","publish_date":"2022-11-16T14:00:00","publication_year":"2022","word_count":1280,"keywords":["Rapids","Go","API","Rust","programming_languages:R","AI","India semiconductor mission","Ray","Aim","analytics","MeitY","tsmc","GlobalFoundries","R"],"extracted_tech_keywords":["AI","analytics","Aim","Ray","Rapids","R","Go","Rust","API","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/indias-semiconductor-mission-is-the-hot-cake-no-one-wants-to-have\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10020894,"title":"What’s Inside Agora’s Real-Time Engagement Platform","content":"For this week’s startup column, Analytics India Magazine spoke to Ranganath Jagannath, Director, Growth at Agora, to understand how the Santa Clara-based startup is leveraging emerging technologies to provide the SDKs and building blocks to enable a real-time engagement platform. Founded in 2014 by CEO Tony Zhao and Co-Founder Tony Wang, Agora is a Real-time Engagement Platform as a Service (RTE PaaS) whose interactive voice, video, and messaging SDKs are embedded into mobile, web and desktop applications. Flagship Products With a head office in Bangalore, Agora offers a voice, video, and live interactive streaming platform. The platform provides developers with simple-to-use, highly customisable, and widely compatible application programming interfaces (APIs) to embed real-time video and voice functionalities into their applications without the need to develop the technology or the underlying infrastructure for real-time engagement. According to Jagannath, Agora is the first RTE PaaS that enables the full spectrum of human interactivity, which can be fully contextualised and embedded into any application, on any device, anywhere. The platform is easier to scale than a traditional Communications Platform-as-a-Service (CPaaS) and delivers RTE services closer to absolute real-time than a traditional Content Delivery Network (CDN). What’s The Differentiator? Jagannath mentioned three reasons why Real-Time Engagement (RTE) is unique and different from Real-time Communication (RTC)- Firstly, RTE delivers the full spectrum of human interactivity, including 1 to 1, 1 to few, and few to millions of people, while RTC is usually about 1 to 1, or 1 to few people. Secondly, the interactions in RTE are contextual and often involve human interactions with digital content and the virtual environment. Lastly, RTE is ubiquitous. Use of AI\/ ML @ Agora To provide customers around the world with high-quality interactive streaming experiences, Agora uses ML in a variety of ways. Software-Defined Real-Time Network (SD-RTN) is a proprietary network of  Agora, which uses ML to dynamically manage the routing of voice and video to overcome severe packet loss incidents and provide uninterrupted streaming. Agora leverages the computing power of ML and AI algorithms on devices to offer more clarity and sharpness during interactive streams. The company uses AI and ML to achieve lower bandwidth allocation during streams, which ensures that users in troubled network environments or emerging markets experience the best live voice and video possible. Core Tech Stack Real-time data transmission throughout RTE-PaaS is handled by the proprietary Software-Defined Real-Time Network (SD-RTN) — a virtual network overlay running on top of the public internet. The Agora SD-RTN runs on more than 200 co-located data centres worldwide. Using sophisticated algorithms, the SD-RTN continually monitors and optimises data transmission to minimise latency and packet loss as well as mitigating lagging and delays. Jagannath said, “In addition, we provide built-in encryption methods that developers can use out of the box, as well as enabling customised encryption that developers can implement for higher levels of security. So, we engineer for security, but our customers have ultimate control over how they design for security.” Funding Recently, Agora was listed on the NASDAQ stock exchange with stock ticker API. In terms of guidance, the company is expecting full-year revenue to be in the range of $125 million to $130 million, unchanged from the last earnings report in August. Hiring Jagannath said, “In India, we are looking forward to increasing our headcount in the coming days. Apart from the tech skills that one needs for the role, what we are looking for are people that have fire in their belly and the enthusiasm to take challenges head-on. And, the desire and drive to help our customers launch products that make a positive difference in peoples’ lives.” Future Roadmap “Agora will continue to focus on key verticals where RTE makes a huge impact like remote education, virtual events, and the like. Online learning is a significant opportunity where organisations can improve their experience. Even beyond the pandemic, a hybrid in-person and digital schooling model might be adopted. We are looking forward to partnering with businesses to integrate real-time engagement solutions.” said Jagannath.","excerpt":"For this week’s startup column, Analytics India Magazine spoke to Ranganath Jagannath, Director, Growth at Agora, to understand how the Santa Clara-based startup is leveraging emerging technologies to provide the SDKs and building blocks to enable a real-time engagement platform. Founded in 2014 by CEO Tony Zhao and Co-Founder Tony Wang, Agora is a Real-time […]","categories":["Deep Tech"],"tags":["Startups"],"author_name":"Ambika Choudhury","publish_date":"2021-02-26T12:00:00","publication_year":"2021","word_count":668,"keywords":["Go","API","AI","ML","Git","RAG","ViT","analytics","GAN","Startups","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","Git","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/whats-inside-agoras-real-time-engagement-platform\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10111387,"title":"Infosys Leverages AI to Bring Immersive Experience at Australian Open 2024","content":"Continuing on its technological partnership with tennis, IT and consulting giant Infosys is expanding its AI influence at the 2024 Australian Open (AO). Leveraging on generative AI technologies, the company aims to enhance fan engagement, optimise player performance and facilitate digital content creation. Infosys will use its AI-centric range of services and solutions with Infosys Topaz to boost the AO experience. Topaz will deliver LLMs and cognitive cores to enhance Tennis Australia’s AI journey. To enhance the Australian Grand Slam experience, different services are offered such as, AO 2024 Bracket Challenge vs AI, where fans can challenge AI in the AO 2024 bracket on the Australian Open website, predicting tournament outcomes and daily match winners. Besides, there’s the Infosys Match Centre, where fans receive contextual insights through the Key Stats feature, AI Match Bytes that uses Gen AI to create visual match story cards, and Win Predictor monitors for showing victory probabilities throughout the match. There would also be Fan Zone where graffiti portraits created using generative AI, powered by Topaz, will be available. Tennis fans will also be able to experience AR selfies and play VR matches. Infosys and AO will utilise digital skilling platforms to build future leaders in Australia, too. Infosys-Tennis Love Source: AIM Over the years, Infosys has actively invested in developing AI-powered solutions for tennis fans. By collaborating with global tennis bodies, especially with the one in Australia, the company has expanded its customer portfolio. Last year, Infosys signed Rafael Nadal to be its brand ambassador with a three-year partnership. The collaboration between both parties will bring forth an AI-powered match analysis tool.","excerpt":"Infosys Topaz brings more AI features to enhance the tennis fan experience.","categories":["AI News"],"tags":["ar","Infosys","VR"],"author_name":"Vandana Nair","publish_date":"2024-01-24T11:17:12","publication_year":"2024","word_count":268,"keywords":["Infosys","AI","programming_languages:R","Git","ar","RAG","VR","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","R","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infosys-leverages-ai-to-bring-immersive-experience-at-australian-open-2024\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":33192,"title":"How Alibaba Is Driving An AI-led Transformation For Macau","content":"Jack Ma, Co-Founder of Alibaba AI may not be ascribed to Macau, a former Portuguese territory better known for its architectural wonders and rich cultural heritage. Dubbed as the gambling capital of the world, Macau, close to Hong Kong has not been known for its technological prowess. But the city has taken several initiatives in the field of artificial intelligence, the most notable ones are in the field of tourism, a major revenue generator for this city, nestled in southern China. Here are some ways in which it has taken small but impactful measures in the field of AI. AI Activities 1) AI For Tourism: Macao Government Tourism Office (MGTO) plans to implement two forms of interactive communication through artificial intelligence systems. According to Herbert Choi, the acting head of Organizational Planning and Development Department of MGTO, the MGTO will use AI to test two similar systems called Chatbot and Singou Butler 1 to visitors’ questions and solicitations. These AI systems will be based on interactive responses to these visitors. A chatbot is a mobile chat system to respond to visitor questions. Whereas, Singou Butler 1 is a more interactive device that can, apart from responding to visitor questions, can suggest activities and places to visit and other information like that. 2.Smart Cities: Alibaba partnered with Macau on the project of Smart Cities. The partnership is for four years, of which, the first phase lasting from 2017 to 2019, will have the collaboration focusing on cloud computing, smart transportation, smart tourism, smart healthcare, and smart city governance. It also has talent development as one of its focus. This initiative will help to expand environmental protection, economic forecasting and customs clearance of Macau. Alibaba has launched similar smart city initiatives in Shanghai, Guangzhou and Shenzhen. The e-commerce giant Alibaba has an ambitious plan for transforming China’s major urban centres into smart cities by leveraging the strength of AI. 3.Predict Flu: According to the CDC, the flu can really wreak havoc on you, knocking you off our feet and straight into bed. It was reported in December 2018 that Alibaba is making use of AI to predict the upcoming flu season in Macau, with a hope of keeping the Macau citizens and the local health bureau informed in advance so that it can help in both the preparations and prevention of flu. The head of the Macau Health Bureau said that the project meets local healthcare standards and will be deployed across local medical institutions. 4.Macau Startup Club: The Macau Startup Club started in the year 2017 is by the Macau-China ThinkThank, that aims to connect, nurture and inspire the local startup community with the help of networking and entrepreneurship projects. This initiative will also encourage research in the field of artificial intelligence to grow in Macau. Conclusion President of the Macau Artificial Intelligence and Blockchain Institute, Rocky Chan had said that Macau (MNA) – Macau is “even more conservative than Mainland China” when it comes to introducing Artificial Intelligence (AI), e-commerce and blockchain technology. However, while AI has experienced considerable development in Mainland China, Mr Chan believes Macau residents and companies take much fewer risks and invest much less in this area. But with baby steps like these, it is safe to say that Macau is heading towards a good adaptation of fast-growing technology like artificial intelligence.","excerpt":"AI may not be ascribed to Macau, a former Portuguese territory better known for its architectural wonders and rich cultural heritage. Dubbed as the gambling capital of the world, Macau, close to Hong Kong has not been known for its technological prowess. But the city has taken several initiatives in the field of artificial intelligence, […]","categories":["Deep Tech"],"tags":["prediction"],"author_name":"Disha Misal","publish_date":"2019-01-10T07:28:10","publication_year":"2019","word_count":556,"keywords":["Go","API","prediction","artificial intelligence","AI","cloud computing","RAG","BERT","Aim","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","cloud computing","R","Go","API","BERT","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-alibaba-is-driving-an-ai-led-transformation-for-macau\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":6939,"title":"The Big Data of Wearables","content":"Wearables – a technology which aims to make life easier. With wearable technology, learning more about yourself has not only become high tech but also real time. From devices and apps that help you track heart rate and food consumption details to gadget that monitor your mood and even surrounding air. According to an April 10th press release from International Data Corporation (IDC), a research company that analyses future trends, “Wearables took a huge step forward over the past year and shipment volumes will exceed 19 million units in 2014, more than tripling last year’s sales. From there, the global market will swell to 111.9 million units in 2018.” I believe there are major four ways wearables can help improve our life:- Firstly wearables keep us FIT – a bracelet which tracks your activity levels, your nutrient intake and improve your fitness. NIKE-Fuelband is one such initiative. Secondly wearables save life – there are wearables which are not only crucial to your health but can also save lives. A wearable health monitor and GPS location device keeps track of the elderly and can alert their caregivers when something is wrong. Thirdly wearables keep us safer – an average smartphone user checks their phone 34 times a day. With people constantly looking at their devices on the go, things can get dangerous. This is where Google Glass comes into play, which is the beginning of this challenge. Fourthly wearables make things fun – apart from making things easier and improving our health, these devices can make life more interesting and fun. Today the biggest market in wearable technology is health and fitness. Big companies are putting wearables to work to figure out how to use these kinds of gadget to improve their business. They are giving wearables to employees and customers to gather subtle data about how they move and act and then use that information to help them do their jobs better or improve their buying experience. However there is a big risk involved. People will naturally resist real world intrusion into their privacy, so businesses needs to be very careful about asking employees and customers to strap gadgets on their head, chest, wrists etc. This compels me to think that we need to truly evaluate the real need of wearable technology. Much of what is being done with wearable devices is happening simply because it can be done. However, several users still are not sure about wearables and whether they want to walk around with devices strapped to them all day. Is this the paradox of wearables? The Paradox of Wearables Today, each one of us has so much personal digital data flowing out there that it is possible for someone to steal an entire online identity and cause real damage offline. You already know that your personal information and references to your social media presence on Facebook, LinkedIn, Twitter, Instagram, Flickr etc. are all over google. This is the data that you know about and it is just a fraction of what can be unearthed with a little drilling, what is really scary is the data that you don’t know is constantly being collected, like your location data. Multiple apps collect location data and track your movements 24×7. Apart from your mobile phone if you use a smart card to pay road-toll or access public transport, you can be tracked by that as well. Some companies are taking this to the next level by using location data to confirm that employees not in the office are actually where they claim they are. On the financial front, every time you swipe a credit or debit card you release more digital information. A marketer may analyse credit card purchases and deduce likely interests. Online retail giants like Amazon, Ebay uses such deduction algorithm when it offers hints like ‘people who viewed this product also looked at the following products. Indian players like Myntra and Flipkart use similar analyses. Recommender systems can help people to find interesting things. Amazon’s recommendation system has helped the technology giant to reap billions in sales increase, NETFLIX is another such success story. Put all these bits and pieces together with just a little online snooping and you could create a detailed composite of an individual’s identity. This may sound like a crime fiction but the basis for it is visible everywhere, if you know where to look. If this personal information falls into the wrong hands, it can lead to a wide spectrum of cyber-abuse like employment or housing discrimination, higher insurance rates, identity theft, or targeted advertising. Maybe it’s high time you give a thought to the question – “how public is your private life.” Having said all this, I see a silver lining, with wearables every individual becomes a data generator and transmitter. We generate data that is continuously collected by various government agencies and private companies. This data can be monetised and can also be used to make life easier for the people, what we need to make sure is that the data do not get manipulated or misused.","excerpt":"Wearables – a technology which aims to make life easier. With wearable technology, learning more about yourself has not only become high tech but also real time. From devices and apps that help you track heart rate and food consumption details to gadget that monitor your mood and even surrounding air. According to an April […]","categories":["IT Services"],"tags":["iot analytics","Wearables India"],"author_name":"Rohit Yadav","publish_date":"2015-02-16T19:23:35","publication_year":"2015","word_count":847,"keywords":["Go","programming_languages:R","AI","Wearables India","iot analytics","Git","programming_languages:Go","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/big-data-wearables\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10021195,"title":"Explained: How To Access JupyterLab On Google Colab","content":"Integrated Development Environments (IDEs) have emerged as one of the fundamental tools in the software development process. IDEs give developers flexibility by integrating various platforms and different processes. According to the Data Science Skills Study 2020 by AIM, of all the IDEs, Jupyter Notebook is the most preferred among data scientists and practitioners. JupyterLab is known as the next-generation web-based user interface for Project Jupyter. JupyterLab is a web-based interactive development environment for Jupyter notebooks, code, and data. It is one of the most loved IDEs by the data science community, thanks to its intuitive features. For instance, JupyterLab can configure and arrange the user interface to support a wide range of workflows in data science, scientific computing, and machine learning. The IDE is also extensible and modular as it can write plugins that add new components and integrate with existing ones. JupyterLab offers full support for Jupyter Notebooks and enables users to use text editors, terminals, data file viewers, and other custom components side by side with notebooks in a tabbed work area. It is built on top of an extension system to customise and enhance JupyterLab by installing additional extensions. Google Colab or Collaboratory is a popular tool for machine learning research and education. Colab comes with a lot of Python libraries pre-installed to help data scientists work efficiently. To access JupyterLab on Google Colab: Step 1: Open a new Notebook on Google Colab Step 2: Switch the backend into either GPU or TPU to run complex computations in an efficient manner. To do so, Select Runtime -> Change Runtime Type Step 3: Install ColabCode, a Python package to run the code server from the Google Colab Notebook. The package can be used to start a JupyterLab through Google Colab. ColabCode also has a command-line script. To install ColabCode, run the following $ pip install colabcode Step 4: In order to run JupyterLab on Google Colab, type the following on a new Colab cell- import colabcode as cc cc.ColabCode(port = 10000,  lab = True) Step 5: Running step 4 will generate a new link through NgrokTunnel. This also generates a new JypterLab token. The code server can be accessed on NgrokTunnel. ngrok is a reverse proxy that creates a secure tunnel from a public endpoint to a locally running web service. ngrok captures and analyses all traffic over the tunnel for later inspection and replay. Step 6: Copy paste the generated JupyterLab token on the “Token” field of the new tab. This will open the JupyterLab. Step 7: To cross-check, whether your JupyterLab is running on Google Colab or not, use the following code- if ‘google.colab’ in str(get_ipython()): print(‘Running on Google Colab’) else: print(‘Not Running!’) Wrapping Up: Accessing JupyterLab on Google Colab allows the use of intuitive features of JupyterLab on Colab. For instance, using Jupyterlab you can code in languages like Python, R and Swift. You can also access JupyterLab on Kaggle Notebook. Click here to know more.","excerpt":"Integrated Development Environments (IDEs) have emerged as one of the fundamental tools in the software development process. IDEs give developers flexibility by integrating various platforms and different processes. According to the Data Science Skills Study 2020 by AIM, of all the IDEs, Jupyter Notebook is the most preferred among data scientists and practitioners. JupyterLab is […]","categories":["AI Trends"],"tags":["Google Colab","Jupyter","Jupyter Notebook","Jupyter Notebooks"],"author_name":"Ambika Choudhury","publish_date":"2021-03-03T10:00:00","publication_year":"2021","word_count":491,"keywords":["data science","Go","Jupyter Notebook","machine learning","TPU","AI","Colab","Python","Aim","Google Colab","Jupyter","R","Jupyter Notebooks"],"extracted_tech_keywords":["AI","machine learning","data science","Aim","Jupyter","Colab","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/explained-how-to-access-jupyterlab-on-google-colab\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10010262,"title":"How AI Is Being Used By Companies During the Festive Season","content":"This part of the year is special for several reasons. While the world approaches the closure of yet another year, it is also the time of some major festivals. Beginning in August with Independence day and ending in February, this period usually sees an increased consumer activity. Naturally, ‘festival season’ is also the time when companies look at stepping up their businesses. Given the scenario, more companies and industries are investing in AI and related technologies to drive more profits. Let us check how companies are leveraging AI-based technology, more specifically, in India’s context. E-commerce From websites to the warehouse, AI can play a very important role in automating the whole process of e-shopping. AI can help in the following ways: AI-based programs can be used in image search to customise search results to users taste and preferences.Chatbots and virtual assistants can help by assisting users. This is particularly helpful for sales days where huge traffic is expected.The analytics can help organisations create a roadmap by predicting which products would be higher in demand.The logistics can also be managed with AI-assisted technology. It is also the need of the hour as the given situation calls for limited human contact. Flipkart’s (Bangalore-headquartered) Big Billion Sale Days has become a separate brand in itself. Watershed deals and discounts available across products have hooked the interest of avid shoppers. This is also the time when the e-commerce giant sees business, unlike any other time of the year. This year, the e-shopping festival is taking place between October 16-21. Anticipating the traffic during these days, e-commerce uses real-time analytics to predict the traffic and consumption patterns to optimise their systems. The AI-based solutions are used to detect and mitigate bots and resellers to stop them in the transaction funnel itself so as to give fair opportunities to authentic buyers. Interestingly, in 2019, Flipkart had deployed up to 340 automated guided vehicles (AGV) ahead of its pre-Diwali sale 2019; this becomes the first of kind robot-based sortation technology in India. Flipkart-owned online fashion retailer Myntra has also geared up for its annual sale. As per media reports, this year Myntra is expected to have two times the traffic as compared to last year. Keeping this view, the company has upgraded its AI-based technology solutions to operate up to 20,000 orders per minute. India’s third-largest e-commerce platform Snapdeal is also increasingly investing in AI-based technology. Just recently, in September, Snapdeal tested its first robot-assisted package delivery. Built-in partnership with mobility startup Ottomony IO, these robots from Snapdeal pick up the package for contactless delivery, keeping in view the raging pandemic situation. The company spokesperson also revealed in a release that they are investing heavily in AI and machine learning capabilities for future capabilities. Offline Retail Stores Not just online stores, offline brick-and-mortar stores are also incorporating technology into their daily operations. The automation is no longer only restricted to checkout. Bots are now being utilised to detect products going out of stock or approaching expiration dates. In such cases, these bots can send alerts to the management. This can be especially useful during busy days such as the festival season where every delay can run the business into massive losses. Further, retailers are adopting robotic process automation to make a transaction and to provide answers to customers’ queries. Earlier this year, Reliance Trends, the fashion retail arm of the Reliance company, decided to use artificial intelligence and data analytics to the stock store-specific assortment. The company is also investing heavily in technology to go hyperlocal. Miscellaneous Several food and product delivery companies flourished during the country-wide lockdown which had severely disrupted the movement of people. This trend is expected to continue even throughout the festive season. Delivery partners such as Swiggy, Zomato, and Dunzo have added a slew of AI-based technologies to make the logistics process much more efficient. Swiggy and Zomato are using the customer information and preferences data that has been accumulated over the years and applying it to machine learning to drive businesses. Interestingly, Swiggy uses data analytics to curate customer landing pages individually.","excerpt":"This part of the year is special for several reasons. While the world approaches the closure of yet another year, it is also the time of some major festivals. Beginning in August with Independence day and ending in February, this period usually sees an increased consumer activity. Naturally, ‘festival season’ is also the time when […]","categories":["AI Features"],"tags":["AI Companies","business intelligence use cases"],"author_name":"Shraddha Goled","publish_date":"2020-10-21T13:00:37","publication_year":"2020","word_count":681,"keywords":["Go","machine learning","artificial intelligence","business intelligence use cases","AI","chatbots","virtual assistants","RAG","analytics","AI Companies","real-time analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","RAG","chatbots","virtual assistants","R","Go","real-time analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-ai-is-being-used-by-companies-during-the-festive-season\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10022675,"title":"Introduction to SIREN: Sinusoidal Representation Networks for Implicit Neural Representations","content":"Implicit Neural Representations yield memory-efficient shape or object or appearance or scene reconstructions for various machine learning problems, including 2D\/3D images, videos, audio and wave problems. However, present implicit neural representations employ non-periodic activation functions such as ReLU, tanh, sigmoid and softplus. ReLU is linear, continuous and differentiable to first-order, but it cannot be differentiated twice. On the other hand, a few variants of ReLU, tanh, sigmoid and softplus are twice-differentiable and continuous. But, these functions are unable to handle a physical signal’s spatial and temporal derivatives. Therefore, these functions fail to yield satisfactory reconstructions for complex and higher-order problems. Vincent Sitzmann, Julien N. P. Martel, Alexander W. Bergman, David B. Lindell and Gordon Wetzstein of Stanford University have introduced a periodic activation network for representing higher-order complex problems. This periodic activation network produces the sine form of the input signal and is named Sinusoidal Representation Networks, shortly, SIREN. Because sinusoidal functions are differentiable to any degree, they help achieve precise 2D and 3D reconstructions along with their spatial and temporal derivatives. SIRENs are trained and validated for these representations using hyper networks with quick and accurate sine activation functions. The generalized form of the problems that the SIREN solves. The SIREN helps modeling complex first-order and second-order ordinary differential equations (ODE) and partial differential equations (PDE), and solves them to great accuracy. The complex problems that the SIREN can solve include famous boundary value problems (BVP) and initial value problems (IVP) such as the Eikonal equations, the Poisson’s equations, the Helmholtz equation, the wave equation and the heat equation. Comparison with other competing implicit neural representation architectures for 2D image reconstruction task in PSNR (Peak Signal-to-Noise Ratio) scale (Source). Reconstruction of 2D image, its first derivative and second derivative by softplus, ReLU PE, tanh and SIREN architectures over several iterations (Source) Reconstruction of audio signal by ReLU, ReLU PE and SIREN architectures (Source) Indoor video reconstruction by SIREN and ReLU architectures (Source) Outdoor video reconstruction by SIREN and ReLU architectures (Source) SIREN solves non-homogeneous Helmholtz equation over a few iterations while other functions fail at all (Source) SIREN solves a wave equation over a few iterations while tanh function fails at all (Source) Continuous 3D structure-aware Neural Scene Representation by SIREN: 3D reconstruction from a single-pose 2D image alone! (Source) Python Implementation of SIREN SIREN needs a GPU runtime, and PyTorch and conda environment. Download the source code into the local machine using the command, !git clone https:\/\/github.com\/vsitzmann\/siren.git Output: Download the Anaconda-3 package using the following command, if the local machine does not have a conda environment. !wget https:\/\/repo.anaconda.com\/archive\/Anaconda3-2020.02-Linux-x86_64.sh Output: Install the downloaded Anaconda-3 package using the following command. !bash Anaconda3-2020.02-Linux-x86_64.sh Enable the conda directory to run further commands, %cd content\/siren\/ !export PATH=~\/anaconda3\/bin:$PATH !exec bash and activate the environment by providing the following commands inside the inner base mode command cell as shown below. Activate the model, conda activate siren The following command performs experimental image training. !python experiment_scripts\/train_img.py --model_type=sine The following command performs experimental audio training on in-built audio clips. !python experiment_scipts\/train_audio.py --model_type=sine --wav_path=<path_to_audio_file> The following command performs experimental video training on in-built bikes-video dataset. !python experiment_scipts\/train_video.py --model_type=sine --experiment_name bikes_video The following command performs experimental 3D-scene reconstruction on in-built Thai statue data by fitting a signed distance function (SDF). !python experiments_scripts\/train_single_sdf.py --model_type=sine --point_cloud_path=<path_to_the_model_in_xyz_format> --batch_size=250000 --experiment_name=experiment_1 Wrapping up Apart from performing implicit neural representations on 2D\/3D image, scene, video, audio datasets, the SIREN architecture is capable of solving complex higher-order differential equations, both ODE and PDE such as the wave equation, the heat equation, the Poisson’s equation, the Helmholtz equation and the Eikonal equations. SIREN’s periodic-activation-function approach may open a vast mathematical field of solving non-homogeneous and complex higher-order problems in the future. References: Original research paperOfficial website of SIRENSource code RepositorySimplified PyTorch implementationSIREN in TensorFlow Playground","excerpt":"SIRENs are trained and validated for implicit neural representations using hyper networks with quick and accurate sine activation functions","categories":["AI Trends"],"tags":["audio","image reconstruction","relu","Tanh"],"author_name":"Rajkumar Lakshmanamoorthy","publish_date":"2021-03-22T17:00:00","publication_year":"2021","word_count":628,"keywords":["Go","machine learning","TPU","AI","image reconstruction","PyTorch","audio","Git","Python","GitHub","TensorFlow","R","Tanh","relu"],"extracted_tech_keywords":["AI","machine learning","TensorFlow","PyTorch","TPU","Python","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/introduction-to-siren-sinusoidal-representation-networks-for-implicit-neural-representations\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10099386,"title":"Google is Being Responsible Irresponsibly","content":"Last week at the Google Cloud Next 2023, the tech giant unveiled SynthID, a tool for watermarking and identifying AI-generated images, in lieu of their motto of keeping responsible AI at the forefront of all that they do. After generating images using Imagen, users can decide whether to include a watermark or not. As decided in the last White House hearing, Google pledged to watermark AI-generated content to protect intellectual property rights and prevent any misleading representation of the content’s authenticity. This helps prevent incidents like the viral spread of misleading images, such as Pope’s Midjourney altered image in a stylish jacket or Trump’s fake arrest photo. SynthID is a step towards it. However, there is a slight hiccup. Although Google has come up with the “experimental” watermarking technique, the company is not addressing the main issue at hand – copyright. The tech conglomerate, which is the flagbearer of the “bold and responsible” approach, recently revised its privacy policy stating that it will be extracting public data from web sources to improve its AI offering such as Bard and Cloud. The updated policy regarding “publicly accessible sources” is somewhat hidden behind a link inside the “Your Local Information” section of the blog post. The Google situation is a bit tricky. It’s as if they want to use copyrighted data for training and then claim it’s copyrighted themselves. To understand this better, imagine Google as a chef. They take your secret recipe, use it to cook delicious dishes, and then slap their own logo on the plates, claiming it as their own special creation. It’s like a never-ending cycle where they keep borrowing your recipes and pretending they came up with them, ignoring the whole conversation around copyright. Google is Leading a Double Life However, these tech giants don’t really care about lawsuits regardless of the issue. Back in July of this year, a class action lawsuit was filed against Google, alleging that the tech giant unlawfully appropriated the IP of “countless Americans” to develop innovations like the AI-driven chatbot Bard and that Google illicitly acquired their information, including personal and work-related data, images, and electronic correspondence, without obtaining user consent over an extended period. Well, this is not the first time that Google has tried to hold up its “Good Kid” image but failed. Its tryst with copyright controversies began in 2005 by launching Google Print, later known as Google Books, a project aiming to scan and share nearly every printed book globally. This move sparked opposition from publishers and authors who saw it as intellectual property theft. Google’s response was to shift the burden of enforcement onto copyright holders, adopting a “we’ll do it until someone tells us not to” approach. Fast forward 18 years later, Google’s updated privacy policy states that they have full control over publicly available data unless the entity explicitly requests to exclude their data from being crawled. Google’s extensive lobbying efforts to influence policy, including copyright and competition rules, costed them around $20 million to settle legal disputes. Again, in 2020, France demanded that Google should negotiate fair compensation for using copyrighted content, highlighting international regulatory disparities compared to the U.S. And now we have similar problems with AI-generated content. Other AI Art Generators Don’t Really Care Google recently launched Visualising AI as well. While not directly involved in image generation, Visualising AI has expanded into the global stock image and video market. However, unlike Google, companies behind AI image generators like Stability AI and Midjourney are taking a different route for copyright, often angering the artists. They don’t really care about responsible AI or copyright, either. In a Forbes interview dated from last September, Midjourney’s founder, David Holz disclosed that their AI-image generator was trained using artworks and photos without the creators’ consent, sparking anger among artists and photographers. He openly acknowledged that the company used existing artworks and photos without permission, with no option for creators to opt out. On the other hand, stock photo agency Getty Images initiated a lawsuit against Emaad Mostaque’s Stability AI alleging that the company trained their open source image generator Stability Diffusion on more than 12 million images from Getty’s database with no permission leading to copyright and trademark infringement. Additionally, Getty claimed that the inclusion of its watermark on some AI-generated images tarnished its trademark, further complicating the dispute. Meanwhile, compared to Stable Diffusion’s over 10 million and Midjourney’s 15 million daily users, Imagen has a very limited user base.  However, even though Google’s reach is low in the image generation market compared to other key competitors, considering the immense power it holds over the tech ecosystem, its every step needs to be mindful. To prevent legal implications, the big techs are now going for partnerships as a new strategy. On one hand, Google is teaming up with Adobe for Firefly and Express integration while DeepMind is collaborating with artists to provide images on platforms like Pexels and Unsplash and maybe this is how Google will continue to be responsible, irresponsibly. Read more: Google Turns AI ‘Bold & Responsible’","excerpt":"Although Google has come up with an “experimental” watermarking technique, the company is not addressing the main issue at hand – copyright.","categories":["Global Tech"],"tags":["AI Tool","Gemini"],"author_name":"Shritama Saha","publish_date":"2023-09-02T13:00:00","publication_year":"2023","word_count":844,"keywords":["Go","Gemini","AWS","AI","innovation","responsible AI","Aim","stable diffusion","GAN","AI Tool","R","AI-generated content"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","stable diffusion","GAN","responsible AI","innovation","AI-generated content"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-is-being-responsible-irresponsibly\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002869,"title":"How Vendor Lock-in Works In Analytics And How To Avoid It","content":"Despite all of the valuable technology products available today on the cloud, many corporations who are considering migrating to the cloud have concerns. And one of the primary issues is vendor lock-in. For starters, vendor lock-in means a condition where the price of switching to another vendor is so high that the customer is stuck with the first vendor. Because of budgetary pressures, an inadequate workforce, or the necessity to avoid obstructions to business operations, the customer is locked-in to what may be a substandard product or service. Challenges That Can Arise Due To Vendor Lock-in Today, many companies have no dedicated servers or dedicated capacity and are priced according to the compute capacity consumed by them. In theory, cloud workloads can be moved from one public cloud provider to another, but it would be a complicated task by embedding a company into a single public cloud infrastructure limits the ability to change vendors. If a vendor’s quality of service declines, or never meets the desired threshold, to begin with, the client will be stuck with it. The vendor may also drastically change their product offerings in such a way that they no longer meet business needs. Finally, a vendor may impose massive price increases for the service, knowing that their clients are locked in. Vendor lock-in that comes in with cloud platforms like AWS, Google and Microsoft, among others, which provide machine-learning-as-a-service (MLaaS) can act as a roadblock to some enterprises. Choosing the correct vendor or cloud-based analytics can help companies to avoid massive costs and gain benefits in the long run. The fact is that vendors can make decisions that are in line with the goals of their customers, especially those companies that are small or use cloud products in very niche applications and use cases. But public cloud vendors (Amazon, Microsoft, and Google) are so huge that it is quite unlikely they will make decisions that negatively impact a large number of users. The Multi-Cloud Has Made Vendor Lock-ins A Big Concern Companies are always worried about the safe and secure portability of data and workloads across different cloud environments. Today open-source data solutions such as MongoDB, Apache Hadoop, Apache Kafka, etc. have emerged to give a great answer to the big data problem and helped companies escape the draconian pricing of traditional cloud vendor lock-ins. Let’s say if a company has an entire data lake and analytics solution on a public cloud platform for a particular region and needs to deploy that solution to another cloud. Implementing this solution on a separate cloud platform would demand a lot of re-work. Therefore, a large number of companies are now exploring flexibility to implement their solutions on any of the widely accessible public or private cloud platforms. But despite many of the platforms having APIs and open-source connectivity, frameworks are generally inflexible. For example, data scientists can use AWS Sagemaker as the tool to use for training and deploying models. Sagemaker promises to cut in training and deployment time by handling all the infrastructure but runs exclusively on AWS. This may not be a problem if you run your business on AWS, but if new products or tools are offered on other public clouds, there is no way to access them. There are also many companies, particularly startups which provide solutions for streamlining operations across various servers, cloud and containers, by providing a transparent platform built to work across multiple clouds. For instance, Cloud Foundry, a container-based architecture that runs apps in any programming language, helps users deploy, and manage high-availability Kubernetes clusters with its open-source project BOSH on any cloud. The project can help decouple applications from infrastructure, so users can host workloads – on-premise, in public clouds, or in managed infrastructures. Cloud giants too, have introduced solutions that may help escape vendor lock-ins. For example, by using Anthos from Google, companies can manage their cloud data workloads across multi-cloud, so that a particular cloud provider does not restrict developers and data scientists. Proprietary Vs Open Source Data Analytics Products Customers today have alternatives to proprietary tools with advances in open source software technologies, along with a range of ‘as-a-service’ capabilities that can remake traditional IT. Projects like Apache Kafka, Apache Spark and Kubernetes are widely accessible as a service on the large cloud platforms. This convergence of open source and proprietary platforms is one of the ways to avoid vendor lock-in in today’s era. Open source is the king. It is perfectly fine to augment whatever you are doing with proprietary tools, but do not be too dependent on them and do learn how to do it without relying on a single vendor. In a recent survey by Anaconda, it found that developers and data scientists value open source so they can get work done right away. It also suggested that many responders believe open source helps prevent vendor lock-ins in data science. An open cloud architecture helps prevent vendor lock-ins and makes it simpler to operate with various analytics services. With the fight over cloud becoming intense, and users trying to avoid the situation of cloud lock-ins, cloud companies themselves are increasingly turning to open-source container technologies and providing them via managed containers. Growth and acceptance of containers have a positive impact on big data analytics and vice versa as they can process and manage vast amounts of data from disparate sources on the cloud via managed containers. In recent years, Kubernetes has appeared as a gold-standard of implementing cloud-native yet cloud-agnostic solutions. It is paving the way for innovation across the cloud infrastructure domain. Containers and microservices make the development process simple and render other benefits, like decreasing the complexity of running and updating apps and advancing the consistency linking testing and production environments. What To Do To Prevent Vendor Lock-In? Companies using cloud computing should make an effort to keep their data portable or easy to move from one environment to another. They can partially do this by clearly defining their data models and keeping data in formats that are usable across a variety of platforms, rather than formats that are specific to a given vendor. Data should be stored in open source formats in the cloud, not in proprietary formats in a vendor’s software or cloud platform. There should be flexibility to choose any existing and future technologies to access, process and query data. For example, you could use Amazon S3 to store the data, Databricks to process it, and Tableau and Power BI visualise it. Keeping internal backups of all data helps a business stay ready to host the data elsewhere if it is too difficult to extract it from cloud service. It also provides protection from ransomware. Companies can prevent vendor lock-in by opting for an enterprise AI platform which gives them a smooth cloud integration with all of their preferred cloud hosting providers. One such platform is DataRobot which supports all of the cloud hosting providers and enables enterprises to scale their data science infrastructure securely and cost-effectively. Using open-source such as sci-kit-learn, Tensorflow, etc. is more effective in creating a model that is full-featured and is more suited to the data scientist’s workflow. With containers, applications are portable and ready to be deployed on any platform. It is also cheaper. In complex applications, Kubernetes give more control in the hands of the developer teams, empowering them to build with ease. It can give teams the flexibility to move to any public\/private\/hybrid cloud solution as per their needs.","excerpt":"Despite all of the valuable technology products available today on the cloud, many corporations who are considering migrating to the cloud have concerns. And one of the primary issues is vendor lock-in.  For starters, vendor lock-in means a condition where the price of switching to another vendor is so high that the customer is stuck […]","categories":["AI Features"],"tags":["Cloud services","Ransomware","Visual Analytics Provider"],"author_name":"Vishal Chawla","publish_date":"2020-07-21T13:00:00","publication_year":"2020","word_count":1249,"keywords":["data science","AWS","AI","cloud computing","Ransomware","ML","Cloud services","Apache Spark","microservices","Visual Analytics Provider","analytics","TensorFlow","kubernetes"],"extracted_tech_keywords":["AI","ML","data science","analytics","TensorFlow","cloud computing","AWS","kubernetes","microservices","Apache Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-vendor-lock-in-works-in-analytics-and-how-to-avoid-it\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":7693,"title":"Can we blame customers for non-adoption of Analytics","content":"We, at Analytics India Magazine, recently did a survey of 50 analytics leaders in India on their Business outlook for the year to come. The results were not surprising at all; almost all of them are confident that the demand of Analytics at their organization will increase in next 12 months; all of them plan to increase their analytics workforce. An interesting outcome of this survey was that almost 3 out of 5 leaders believe that their customers have little knowledge of analytics and this posed as a top challenge for non-adoption of analytics. This came as a close second from “Unavailability of Analytics Talent” as a top challenge. I have little doubt in my mind that the end consumers of analytics are a misinformed bunch. More so, I believe that it is we as an industry that is to blame for this. In my mind couple of things have happened: 1. Heavily “jargon-ized” industry Analytics is one of the most jargon heavy areas recently; “Business intelligence”, “Data Science”, “decision science”, “machine learning”, “Big Data”, “artificial intelligence”. It’s hard for even a seasoned analytics professional to distinguish between these terms, leave aside the end consumer of analytics. I recently published an article on how Big Data is not analytics. Jargons help in initial selling by playing on naivety of end consumer, especially where heavy consulting is involved rather than actual implementation. 2. Exaggerated misinformation I regularly come across content over Internet around how big data can eradicate poverty, market bubbles, cancer, diseases; predict earthquakes, future, crime, super bowl etc. Obviously, statistics and intelligent algorithms are being used for decades in various areas and analytics is more of an industrialization of these techniques. Analytics is being utilized in various areas with varying degree of success and the usage will only increase; but most of these content is just an exaggerated misinformation without much of the crux. 3. ROI from analytics cannot be fully established Let’s say I pitch a basic regression model to a customer and explain the benefits of doing so. But, maybe there is no model; maybe there is no correlation at all. I’ll come to know that once I start working on the model itself. This is just a simple example of why it’s hard to pin point a confident result from the very beginning. Unlike an IT project, where we can freeze on a specific output (application, database etc), analytics is as much about doing analytics as the final results. 4. Standards & processes not prevalent IT industry, during its maturing phase, gave rise to various standards and processes in almost all areas – development, testing, collaboration, maintenance etc. SDLC is one such framework that is an industry standard and everyone in the eco-system abide by it. Analytics has yet to evolve any such standards. Even standards around how do we store, deploy, share for re-use analytics models\/ algorithms is still to evolve. 5. A fast paced industry Analytics is an extremely faced paced industry, what is novel now is outdated in 1-2 years. The pace of innovation and buzz creation outruns adoption. An example being social media analytics, I don’t think there is complete adoption of social media analytics among enterprises, but there is not much buzz anymore. Given all the shortcomings of the industry, I would believe that the best way to adopt analytics is a continuous experimentation. Persistence is the key; early results may at times even be discouraging. A right implementation of analytics is a source of sustained competitive advantage for organizations; I say ‘sustained’ because not many get it right. I would also re-iterate from an earlier article that as an industry we need to make analytics ‘unpretentious’ for our customers and not blame them for non-adoption.","excerpt":"We, at Analytics India Magazine, recently did a survey of 50 analytics leaders in India on their Business outlook for the year to come. The results were not surprising at all; almost all of them are confident that the demand of Analytics at their organization will increase in next 12 months; all of them plan […]","categories":["IT Services"],"tags":[],"author_name":"Дарья","publish_date":"2015-07-21T12:28:54","publication_year":"2015","word_count":625,"keywords":["big data","data science","Go","machine learning","artificial intelligence","TPU","AI","RAG","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","RAG","TPU","R","Go","big data"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-we-blame-customers-for-non-adoption-of-analytics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10085722,"title":"40 Under 40 Data Scientists 2023 – Who are they?","content":"Following two action-packed days of workshops, conferences, paper presentations, and tech talks, Machine Learning Developers Summit 2023 concluded by awarding 40 dynamic data scientists with the 40 Under 40 Data Scientists award. The award recognises innovators and their achievements in the machine learning and analytics industry. Here is the list of winners of the 40 Under 40 Data Scientists (in alphabetical order): Aakash Kathuria, General Manager\/Head – Business Growth, Lenskart Solutions Private Limited Aakash is a seasoned analytics leader with 15+ years experience and has been instrumental in driving data and insight-led transformations. Over his career, he has worked closely with biz functions to drive revenue and achieve aggressive market growth by leveraging more than 50 analytical approaches. He also has experience in launching AI and tech-based solutions like Omni Channel Attribution, Customer Segmentation, Customer-360, Supply Chain Efficiency, Workforce Management and more at telecom, media, FMCG, retail, and ecommerce industries. Abhilash Surendran, Assistant Vice President, Insights and Analytics – Merkle Abhilash Surendran is assistant vice president, analytics, and data science at Merkle, leading the analytics practise for their high-tech portfolio. He comes with 15+ years of experience in advanced analytics, data science, data visualisation and consulting. He was previously part of Mu Sigma where, over the course of ten years, he went on to deliver analytics solutions for a portfolio of accounts across pharma, retail, CPG, and energy verticals. Abhishek Mishra, Vice President – Senior Manager of Analytics, MarshMclennan With over 13 years of expertise in the information technology services and product business as a data science professional, Abhishek Mishra is committed to using his data science knowledge and engineering talents to access, comprehend, and express insights from sizable data sets to aid in growth plans, product development, and crucial business decisions. He has technical credentials such as Analytics Core from Harvard Business School, Data Science Associate from EMC, and Edge AI developer from Intel. He is also a Six Sigma Black and Scrum Alliance Certified Product Owner. Aditya Durai, Delivery Unit Head at Math Company A seasoned data science expert, Aditya Durai currently spearheads multiple enterprise-wide initiatives and delivers successful business outcomes for leading Fortune100 enterprises. His in-depth knowledge of algorithm design, statistics, and machine learning has enabled him to build over 100 solutions as well as act in an advisory capacity to Fortune500 enterprises for data science solutions across merchandising, pricing, and trade promotion optimization. Aditya also plays a significant role in building the data science talent at MathCo by designing the organisation’s learning and development strategy and being deeply involved in their training. Aditya Kumar Pal, Associate Director – Head of Artificial Intelligence, Lendingkart Aditya has a rich experience of more than ten years in the domain of data sciences and AI. He has worked with more than 50 stakeholders over the past decade to solve their problems using data and algorithms across multiple functions such as customer analytics, pricing analytics, assortment analytics and marketing analytics. Recently, Aditya won the “AI Impact changemaker award” from one of the largest Analytics federations of India. Anurodh Kumar Gupta, Head Data Science Compliance – Societe Generale Global Solution Centre Anurodh is a data and AI leader with 14 years of experience in strategy and consulting and business process improvement across multiple domains like banking, telecom, healthcare, retail and more. He is an alumnus of IIT-Madras and has good experience of establishing data science teams from scratch for US & European MNCs. As consultant, he has supported C-Suite of his client organisations in achieving their digital agenda. Anurodh is also a President Awardee in Bharat Scouts and Guides. He loves to give back to the community and is passionate about teaching. Dr Anusuya Kirubakaran, Lead Assistant Vice President – HSBC Dr Anusuya Kirubakaran is driving innovations in the banking sector’s analytical domain especially discovering business insights from challenging unstructured free text data. She has completed Doctorate in computer science by solving analytical problems and published papers in leading Scopus-indexed journals. She developed a generic, reusable, extensible text analytics tool using NLP, AI\/ML-based hybrid techniques which speeds up the text analytics phases and is part of NLP-as-a-Service. Arvind Jayakumar, Director of Technology, Artificial Intelligence – Springer Nature Group Arvind Jayakumar is an alumnus of BITS Pilani, Osmania University, and the University of Oxford, he has nearly one-and-a-half decade worth of experience in data engineering, data science, and AI\/ML. In his current role, he is building SN Group’s AI CoE and is working to industrialise the usage of AI\/ML and its associated best practices across the firm. Ashwini Chandrashekharaiah, Senior Manager – Walmart Global Tech India Along with Walmart, Ashwini Chandrashekharaiah is also leading Data Science products for Data Ventures, a data commercialization initiative. She has over 12 years of experience in the retail industry spanning across domains of Pricing, Assortment, Customer, Security and Data Insights leading to multi-million-dollar impact products, as well as helping the business teams in their strategic activities. She is also an active member of the Women in Data Science (WiDS) initiative and is a passionate mentor for budding AI enthusiasts. Balaji Jagan, Associate Manager – Data Science (NLP), Allstate India Data scientist with specialisation in NLP, Balaji Jagan has 15+ years of research and industry experience, working on diverse machine learning problems. He has been working in Allstate Solutions Pvt. Ltd for more than four years as an Associate Manager – Data Science (NLP), leading the Business Solutions team. He has filed 8 patents out of which 7 have been granted. Balaji has published more than 20 research articles in International conferences and International Journals. Deekshant Saxena, Global Director – Software Engineering, AI & ML, research, HERE Technologies A global director of Software Engineering and Research at HERE Technologies, Deekshant Saxena spans a team of more than 200 data scientists, researchers, senior engineers globally. He joined HERE Technologies in 2017 and over the last five years, he has been instrumental in driving research and innovation thereby driving new business opportunities focusing customer centricity on various domains, markets and geographies. Deepak Singh, Principal – Analytics, Dr. Reddy’s Deepak Singh is an analytics leader with over 13 years of experience and expertise in delivering top-line business impact through innovative use of data and productising analytical solutions across the business value chain. He is currently heading the Data and Analytics practice for Dr. Reddy’s PSAI Business and is instrumental in unlocking business value through various Analytics products and solutions. One of the products, JARVIS (AI\/ML-powered Data Intelligence and Analytics Platform), was recently awarded ‘Top Domestic Firms using AI\/ML’ in Cypher 2022. Deepika Kaushal, Deputy Vice President – Piramal Finance With 15 years of experience in data science and business analytics, Deepika Kaushal is currently working as a Deputy Vice President at Piramal Finance leading everyday AI and digital practices. She works in the areas of process improvement and works with stakeholders from sales planning, marketing, HR, customer service and operations and audit. She has also worked with firms like Fractal analytics, Zebra technologies, Cognizant, WNS, Sapient and others. Deepika has also worked for many fortune 500 clients like Coca Cola, Pepsico, TE connectivity, Mahindra Automotive, Tommy Hilfiger, Philips, Unilever, Telstra, and Insurance Australia Group. Divyan Kavdia, Assistant Vice President – Max Life Insurance Company Ltd With an experience of more than 12 years, Divyan Kavdia has been associated with MaxLife since the last 5 years. He has a rich diversified professional experience across industries like BFSI, FMCG, CPG, banking, insurance, & taxation. As an AI leader at Max Life, Divyan has been instrumental in carving the AI-Works team as a unit to provide MLI Launchpad on embedding intelligence in traditional processes thereby transforming processes with AI\/ML at core. Dr Santosh Karthikeyan, Associate Director, Clinical Programming – AstraZeneca A certified data management professional, Dr Santosh Karthikeyan is involved in providing strategic, technical, and operational leadership for the data and analytics platform in R&D. He has led data science projects on predicting adverse drug reactions in Life Sciences and managed teams in digital and analytics practice. His research papers are published in various national and international journals. Dr Santosh Karthikeyan has also participated in National and International conferences. Hirdesh Khanna, Asst. Vice President and Lead Data Scientist – HDFC Bank Ltd. With over 13 years of experience in applied analytics, ranging from ML to stochastic processes like Markov’s chain and many more, Hirdesh Khanna is a key team member at HDFC Bank for driving digital business that has helped transform the marketing department as a revenue function for the bank. He was also recently conferred as the “Data Science Professional of the Year – 2022” at the Future of Data and AI Summit 2022 organised by UBS Forum Pvt. Ltd. Hitesh Nahata, Senior Manager, Capabilities – MiQ Hitesh Nahata has over 12 years of experience delivering analytical solutions to business problems across functions for Fortune 500 companies in CPG and Retail. At MIQ, he leads the global advanced analytics capabilities team from their Centre of Excellence in Bangalore. Hitesh has a PGDM from MICA specialising in Market Research and Analytics and an undergraduate degree in Engineering from Mumbai University specialising in information technology. Jitendra Gupta, Data Science Leader – Baker Hughes Data science leader with 14 years of experience in solving critical problems in finance, oil and gas, and retail organisations, Jitendra Gupta, has worked across companies and geographies, helping design, develop and deploy data and analytics solutions. He has substantial experience with machine learning, deep learning, NLP, and computer vision. In addition to his role as Baker Hughes’ data science leader, he has worked with a range of other clients including General Motors, General Electric, Best Buy, and HSBC. Kumar Abhishek, Senior Machine Learning Engineer – Expedia Group As a senior ML engineer for Expedia, Kumar is leading efforts on fraud detection and prevention. He uses NLP for risk analysis and fraud detection for various Expedia brands. Additionally, Kumar is personally responsible for designing and implementing several highly effective efficiency measures, including record low latency and heavily optimising the machine learning pipeline. Previously, he worked for Microsoft on Cortana and has an M.S. in CS from the University of Florida. Nakul Laad, Head of Analytics – Dunzo Nakul Laad has been working in the field of data analytics for the last 13+ years. He has built world-class analytics capabilities across professional services and product companies. Currently, he is heading the analytics team at Dunzo where they work on solving some complex hyperlocal logistics problems using advanced analytics. Nakul previously held a leadership role at Swiggy where he led strategy and analytics. Parimesh Panda, Senior Manager – Genpact In his current capacity at Genpact, Parimesh Panda is a senior consulting leader in AI and ML with superlative expertise in blueprinting, strategy and business development, go-to-market pipeline creation and solution execution of digital transformation programmes. He leads AI and ML maturity assessments and drives growth in customer engagements by creating and sharing whitepaper articles on state-of-the-art AI\/ML technologies and practices. Pramod R, Director, Data Science – Falabella Pramod R comes with a wide range of AI\/ML expertise ranging from recommendation systems, NLP, reinforcement learning, deep neural networks, big data processing technologies, massive parallel computing, web development, and more, using numerous open source technologies. He holds a patent on reinforcement learning, has presented papers in international conferences, given talks at various meetups and is featured in notable international journals. Pramod brings in 16+ years of rich experience in building high end AI solutions. Praveen Jesudhas, Associate Director – Data Science, Tiger Analytics Praveen Jesudhas has 13+ years of experience in utilising data science aligned with technology to add client value both across services and product organisations. Currently, he focuses on enhancing capabilities to better execute computer vision-based projects at Tiger Analytics. For the past seven years at Tiger Analytics, Praveen has played a key role in solving several client problems across domains such as insurance, manufacturing, sports analytics, and social media using his expertise. Preetam Biswas, Head Analytics – Aditya Birla Health Insurance Preetam is an MDI Gurgaon MBA graduate with 15 years of experience, currently heading the analytics department in Aditya Birla Health Insurance and driving the overall Analytics initiatives across the organisation involving all stakeholders. He has a rich experience in BFSI and telecom industry and has worked for organisations like Kotak Life, Axis Bank, Vodafone Idea, Kantar, and Capgemini. Preity Yadav, Manager (Strategic Information Systems) – Indian Oil Corporation Limited Preity Yadav has expertise in core business domains and core technical skills such as advanced analytics, systems integration, data management, hardware management, architecture design and networking which helps in the identification, conceptualisation, development, and evaluation of initiatives Indian Oil envisages under its digital transformation journey. She is responsible for delivering high-impact advanced analytics machine learning solutions such as predictive maintenance, yield optimisation, revenue maximisation and more. Pritam Banerjee, AVP – Advanced Analytics, Chubb Pritam has worked extensively in areas of advanced analytics, optimisation, and strategy. He has been instrumental in setting up and scaling analytics centres of excellence delivering AI-enabled solutions. He has led cross-functional diverse teams across APAC, EMEA and NA and has been highly recommended by clients and stakeholders alike. In his current role as AVP Advanced Analytics at Chubb, Pritam leads the Claims and Unstructured Center of Excellence, working with cross-border teams to develop large scale AI solutions. Rahul Pandey, Associate Vice President – Advanced Analytics & Applied AI, Course5 Intelligence Rahul Pandey has 14 years of experience leading end-to-end design and development of analytics and solutions across diverse industry segments—beyond automotive, retail and CPG, supply chain, and manufacturing. He has built and led a multi-disciplinary team of domain SMEs, data architects, data engineers, data scientists, and technology experts in India and the US. Currently, he drives a global practice cohort of Industry 5.0, Media, and Telecom for Course5 Intelligence in Advanced Analytics & Applied AI. Rai Rajani Vinodkumar, Global Director Data Science – AB InBev Rai Rajani is a highly accomplished AI professional and a digital transformation lead and has a career spanning 12 years. She leads the team of data scientists assigned to multiple analytics products supporting marketing, sales, and operations business units across the globe. She has designed and implemented data science methodology, and best practices to productise scalable initiatives across functions. Rai Rajani has contributed to many talent building activities by conceptualising data science career paths and designing skill assessments for building and retaining best data science talent. Sachin Chaudhari, Manager – Charge Management Data Science, Maersk Sachin Chaudhari is an experienced data scientist with a demonstrated history of solving critical business problems using AI\/ML. He specialises in AI and operations research techniques at Maersk. His strong belief in data driven decision making and research focused mindset has helped Maersk develop innovative solutions for pricing and revenue optimization. Sachin is an integral part of ‘Innovation, Data Science and Automation’ (IDA) CoE at Maersk which has been recognized globally through various awards by organisations such as NASSCOM, Zinnov, 3AI, Analytics India Magazine, etc. Sateesh Gottumukkala, Head of AI Engineering Centre of Excellence (Associate Director, Data Science) – LTIMindtree Sateesh has nearly 13 years of experience in working with data science, AI\/ML, MLOps, and real-world business problems across different domains. He leads the AI Engineering Centre of Excellence at LTIMindtree. He has previously worked as a Senior Data Scientist at Sogeti (Capgemini Group) where he implemented Data Science and ML solutions for Banking, Insurance, Mining and Telecom customers. Prior to that, he worked as Research Associate at National Aerospace Laboratories where he worked on mathematical models on wind tunnel data for designing delta wing configurations. Seema Nagar, Staff Research Scientist – IBM Research, India For the past fifteen years, Seema Nagar has been researching in the field of computer science and has over 45 publications in eminent conferences and journals along with more than 150 patents filed. She has been named a master inventor for two consecutive terms. She has also been part of the reviewer committee of many eminent conferences, such as IJCAI, NAACL, IEEE Cloud and ACL. Seema has earned several research awards in IBM Research India, including three Outstanding Technical Achievement Awards for her work on social network analysis and trustworthy AI. Shuaib Ahmed S, Technical Manager – Mercedes Benz R&D India Dr Shuaib Ahmed is responsible for algorithm research, design, and development for interior cameras in Mercedes Benz cars. His research work has been published in conferences and workshops like CVPR, ECCV and NeurIPS. He has filed 20 patent applications. His area of interests are active learning, continuous learning, explainable AI, and learning from multiple domains. He holds a PhD from Indira Gandhi Centre for Atomic Research. Dr Ahmed has delivered lectures in many tier-2 and tier-3 colleges to create awareness among students and faculties for AI and data science. Siddhant Verma, Department Manager – Usha International Ltd. With over nine years of experience in data science and AI, Siddhant Verma has expertise in designing and developing high impact solutions that help companies across various verticals such as sales planning, supply chain, recommendation engines for B2B customers, demand planning and forecasting, and others. Along with the love for writing ML\/AI code and churning data, he thrives on solving business problems with data and communicating the solution clearly, concisely and in actionable stories. Sudalai Rajkumar – 4x Kaggle Grandmaster Sudalai Rajkumar (SRK) is a Kaggle Quadruple Grandmaster and is currently the head of AI and ML at Growfin.ai. He has more than 12 years of experience in the data science field and has built scalable AI\/ML products. SRK has also done data science consulting for enterprises as part of his career at H2O.ai, Freshworks, Tiger analytics and Global Analytics. He is an alumnus of PSG College of Technology and IIM-Bangalore. Sunil Kumar Singh, Sr Director, Artificial Intelligence (AI Engineering) – IQVIA Sunil is a graduate from IIT-Roorkee and has 15+ years of professional experience across industry verticals. He currently heads the AI and Machine Learning Center of Excellence at IQVIA. Sunil has an extensive experience of 12+ years in healthcare analytics—leading advanced data science projects, developing innovative AI ML solutions and creating next gen data science assets for various large and emerging pharma companies. Suresh Sethuramaswamy, Engineering Lead – Data and AI for FSI, Microsoft Suresh Sethuramaswamy is a data and analytics leader with over 15 years of extensive international experience, ranging from heading data platforms to leading the Analytics Center of Excellence for Banking and capital markets. Suresh built many award-winning predictive analytics and alternative data platforms. In his most recent role at Microsoft, Suresh leads product development efforts for the financial services industry-specific differential data and analytics products. He is an advisory member of Forbes Technology Council and a senior member of IEEE Princeton. Suvajit Mukhopadhyay, Principal Solution Architect and Data Scientist – Uno Bank With 16 years of industry experience, Suvajit Mukhopadhyay has designed large scale enterprise digital transformation programmes in banking and financial services domain. He has built complex AI Models for use cases around KYC, customer analytics, bank statement analysis, risk analysis and fraud detection. He is also a trainer, mentor and researcher having trained more than 1,000 learners in data science and AI in diploma and postgraduate programmes. Syed Mohammad Imran Hashmi, co-founder and Chief Analytics Officer – Loyalytics Consulting LLP Imran is a data science and analytics leader with more than 13 years of experience in setting up and leading high performing data science teams. He has developed and deployed solutions across industries and markets by leveraging the power of data, machine learning and engineering. In 2015, Imran, along with two other co-founders, started ‘Loyalytics Consulting’, a boutique data science and analytics company with the vision to help businesses scale and systematise data-driven decision-making through the power of machine learning and AI. Varsha Singh, Chief Manager – Data Scientist, Reliance Securities Ltd. Varsha is a seasoned data scientist with a total of seven years of experience in the finance domain. She is also a visiting Professor in NMIMS university. Previously, she has worked with organisations like SBI-MF, Kotak Securities, Reliance Capital and Cognizant. She has developed several machine learning models in the organisation to solve business problems and improve customer experience. Her expertise lies in fraud detection, churn analysis, next-based product recommendation model and competition analysis.","excerpt":"40 Under 40 Data Scientists award recognises innovators and their achievements in the machine learning and analytics industry.","categories":["IT Services"],"tags":["40 Under 40 Awards","Data Scientist Awards","machine learning developers summit"],"author_name":"Mohit Pandey","publish_date":"2023-01-24T15:30:00","publication_year":"2023","word_count":3379,"keywords":["Data Scientist Awards","data science","artificial intelligence","machine learning","AI","neural network","ML","computer vision","machine learning developers summit","NLP","deep learning","analytics","40 Under 40 Awards"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","computer vision","data science","analytics"],"url":"https:\/\/analyticsindiamag.com\/it-services\/40-under-40-data-scientists-2023-who-are-they\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":54455,"title":"Top 8 IoT Projects Beginners Must Try In 2020","content":"With the growing interest in the Internet of Things (IoT), there has been a lot of advancement in various sectors such as healthcare, home systems, manufacturing companies and others. The global IoT in utility market size is expected to grow from $28.6 billion in 2019 to $53.8 billion by 2024, at a CAGR of 13.5% during the forecast period.In one of our articles, we discussed how India railways could use this technology to match up to other countries. In this article, we list down eight IoT projects for a beginner in 2020. The list is in alphabetical order. Air Pollution Monitoring System IoT-based air pollution monitoring system has been developed to monitor the quality of air over a web server using the internet. After building the model, it will trigger an alarm when the quality of air goes down beyond a certain level. For this project, the software requirements include Arduino 1.6.12 and embedded C language and the hardware requirements include 1K ohm resistors, Arduino Uno, breadboard, 10k potentiometer and other such. Click here to know more. Facial Recognition Door With Raspberry PI Facial recognition is one of the most trending security techniques one uses in the present scenario. A facial recognition door lock system can be used to prevent a robbery in a home. In this project, an image will be captured by a Pi camera and pre-processed by Raspberry Pi like converting, re-sizing and cropping. Then face detection and recognition are performed. Once the face is recognised by the classifier based on a pre-stored image library, the image will be sent to a remote console waiting for the house owner’s decision. Click here to know more. IoT based Alarm Clock An ordinary clock only uses sound as a medium to wake up the user. While in the case of the internet of things based smart alarm clock. It can use more than a traditional alarm clock does to wake up or remind of something important to the user. For instance, it can turn on the smart lights or switch off the fan, etc. The requirements of this project include RTC chip, TFT screen, Wifi Module, etc. Click here to know more. IoT Based Intelligent Traffic Management System An advanced traffic management system using the internet of things can decrease the more waiting time for the drivers to crossroad signals. It can include several features such as parking space identification and allotment system, anti-theft system and other such. Click here to know more. Robot Using Arduino Making a robot with the help of Arduino is one of the simplest uses of Arduino. This project will help you to understand how Arduino works as well as how to interface DC motors, IR sensors, etc. The requirements include Arduino or Arduino clone board, DC geared motors, IR modules, Caster wheel, etc. Click here to learn more. Smart Street Light The use of smart street light projects instead of using manual light can help in saving a considerable amount of electrical energy. This is done by sensing and approaching vehicles using an IR transmitter and IR receiver couple. The requirements of this project include Arduino IDE, MPLAB IDE, OrCAD, microcontroller, light-dependent resistor (LDR), IR sensor and other such. Click here to learn more. Smart Garage Door A smart garage is one of the essential needs of a smart home. These doors not only help from burglars but also closes on its own as soon as it senses that someone took the car out. Unlike a manual garage door, the smart garage door gives a tough time to the thieves for entering the house. The project provides the user with a simple monitoring system that enables them to act more security conscious, effectively increasing the safety and security of their home. The device consists of a transmitting unit in the garage that tracks the door’s position, and a portable, wireless receiving device in the house that displays the garage door’s status. Click here to learn more. Smart Wheelchairs Smart wheelchairs not only focus on the mobility of the device but also on the health monitoring of the patient. The objective of the present work is to develop a smart sensing wheelchair by implementing sensors within its structures. Using sensors and processing by embedded systems, the intelligent wheelchair will be able to track the heart rate, blood oxygen levels, etc. and initiates a trigger as soon as it receives any abnormality. The requirements of this project include Arduino IDE, Node MCU, MQTT library, PPG sensor and other such. Click here to learn more.","excerpt":"With the growing interest in the Internet of Things (IoT), there has been a lot of advancement in various sectors such as healthcare, home systems, manufacturing companies and others.  The global IoT in utility market size is expected to grow from $28.6 billion in 2019 to $53.8 billion by 2024, at a CAGR of 13.5% […]","categories":["AI Trends"],"tags":["ai for beginners tutorial","arduino","Internet of things","IoT","iot platform"],"author_name":"Ambika Choudhury","publish_date":"2020-01-20T13:00:00","publication_year":"2020","word_count":759,"keywords":["Go","programming_languages:R","AI","ai for beginners tutorial","R","programming_languages:Go","Internet of things","RAG","iot platform","arduino","IoT"],"extracted_tech_keywords":["AI","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-8-iot-projects-beginners-must-try-in-2020\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042211,"title":"Microsoft Rolls Out The Future Of Hybrid Workplaces","content":"At the Microsoft 2020 Build Conference, a team built a puzzle demo where thousands of users were asked to work on jigsaw-like puzzles in real-time. All the users were allowed to see the result of thousands of minor edits and updates. The reason behind showing off the power of high-scale and high-performance collaboration capabilities was to unveil Microsoft’s then-new Fluid Framework. Now, Microsoft has started to roll out what is called the most significant change to Office made in decades. <Embed Youtube Video: https:\/\/www.youtube.com\/watch?v=tPw5kFkXtt4&feature=emb_imp_woyt> Fluid Framework is a tool built for very low latency collaboration and synchronisation. Using it, developers can create robust applications with the help of familiar programming patterns and cooperation. With Fluid Framework, when any single user makes a change in their browser, every other user will be able to see it almost instantly. Additionally, Fluid components allow tables, graphs, lists—things that are static and usually bound to specific platforms and documents—to live independently across the web. Thus, Fluid Framework would make an application more accessible to multiple users and allow them to edit it simultaneously whilst also making such content easy to share across applications. Google Docs, but in a bolder font Like Google Docs– a tool known for its collaborative properties, Fluid Framework is rapid, with no sync time. However, Microsoft’s tool goes a little further than Docs. Google Doc still involves creating a new document, adding tables or charts or tasks to it (all of these usually made elsewhere, saved and then added to the Doc) and sharing it with others. However, Fluid Framework is Microsoft’s attempt to shift from the age-old concept of creating and saving documents. Fluid components can be made in any application and shared immediately, without switching to another app. As per Maya Rodrig, Principal Programme Manager for this platform, “It is about helping people achieve a particular task” without needing to decide which document to return to or which app to go on. Microsoft looks at Fluid as a vital element to the future of productivity. The Silicon Valley giant also open-sourced its work last year to allow people everywhere to shape its creation. The rollout Microsoft will roll out Fluid Framework in Microsoft Teams this year, where it will be embeddable in meetings and chats. Doing so makes sense given the increase in remote and in-person hybrid workplaces. As a part of this, every Microsoft Teams meeting will soon come with a built-in notes experience tool. It will be present in Teams meetings or Outlook calendar and allow everyone invited to type notes in real-time. The platform will immediately sync these notes to one’s tasks across Microsoft 365 and Outlook Calendar. Besides this, Microsoft will also be allowing Fluid components on its Whiteboard app. It will comprise new collaboration cursors on Whiteboard, enabling users to see any additions made to documents by their coworkers and vice-versa in real-time. Microsoft will also add new reaction stickers and a virtual laser pointer to make these remote and hybrid workplaces more interactive. Finally, individuals can also embed aspects of Fluid such as tables and task lists to Whiteboard—making the entire app look and feel the same across every device and platform. This will allow users to edit elements whilst using Whiteboard as a dashboard. Microsoft sees the launch of its Fluid Framework as the ‘Start of a community built around developer technologies for building collaborative applications.’ It is highly likely for users to see Fluid components first on Teams and then on web aspects of MS Office before finally having them be commonplace on desktop platforms. One has to wait eagerly for this novel spectacle. Until then, individuals can also check out Microsoft’s Fluid Framework demos to get started.","excerpt":"Microsoft sees the launch of its Fluid Framework as the ‘start of a community built around developer technologies for building collaborative applications.’","categories":["Global Tech"],"tags":["Microsoft 365"],"author_name":"Mita Chaturvedi","publish_date":"2021-06-22T18:00:00","publication_year":"2021","word_count":618,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","Microsoft 365","ViT","R"],"extracted_tech_keywords":["AI","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-rolls-out-the-future-of-hybrid-workplaces\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":3713,"title":"A Typical Day In The Life Of An Analyst","content":"One of the common thoughts in the mind of the new MBA is – how will my typical work day span out? What will I be doing for 8 hours a day, 40 hours a week if I were to make a career in Analytics? Analytics is a relatively ‘young’ area of work. It has evolved and expanded 100 fold in the last 10 years as more and more companies and their top managements have become convinced about the immense value it adds to decision making. In this increasingly complex marketplace, where net and mobile technologies are adding to the conventional brick and mortar economy, and the life of data is becoming short (ie, data which is very old is not usable), analytics is becoming increasingly useful and moving from the fringes towards the mainstream w.r.t business decisions. This, of course, spells increased requirement for analytics talent. It also increases the responsibility of success and failure to correct and timely decisions made using analytics. In short, the analyst (used loosely to define the analytics fraternity) has many choices in the workplace and can decide which role to play in his career. Broadly the roles can be segregated into Individual Contributor – Subject matter expert  (IC -SME) role Team Manager role Individual Contributor – Subject matter expert (IC -SME) role: – As the name signifies, an Individual contributor works by himself and may not have a team reporting to him. He would typically be an expert on a particular subject\/s or line of work (eg. web analytics and using it for marketing campaigns, modeling for Loss given default etc.). He may also be a specialist in some software – Hadoop \/ SAS e- Miner etc. He would work on projects across teams in an advisory role and would interact with the client as and when required. Exploring paths to improve a project would form a large part of his job. He will have to keep abreast with happenings in his area of specialization in his industry as well as in other related industries.  A keen brain, exploratory nature and love of his work would be the hallmarks of an SME IC. Team Manager Role: – A team manager would be spending equal parts of his day between client management, team handling and other organizational duties (budgeting, planning, recruitment, training etc.). His knowledge of analytics would be sound peppered with practical ways to execute projects, optimize outputs and ensure client satisfaction with the given resource constraints. A typical ‘manager’, he will be juggling multiple deliverables and playing different roles in the day. A person in this role can rise up to become the CEO, COO etc. of an Analytics firm. He also has scope to move into consulting and have a practical idea of implementing the analytics output. The higher he rises, the less of hands-on analytics comes his way. His goals would include business acquisition targets. These roles are more plentiful than the IC- SME role discussed above. Thus, as many pure BPOs move into analytics and become KPOs, the requirement for analytics talent will increase. In my view, analytics is a skill and people with any background, with a head for numbers and a conviction in the power of decision making by numbers will find this to be a field after their own heart. It covers nearly all streams of work within an organization – HR, Operations, Marketing, Sales, IT, Supply chain management, Logistics– and hence, the choice of the team that you decide to join for analytics can be aligned to your area of interest of specializations and study.","excerpt":"One of the common thoughts in the mind of the new MBA is – how will my typical work day span out? What will I be doing for 8 hours a day, 40 hours a week if I were to make a career in Analytics? Analytics is a relatively ‘young’ area of work. It has […]","categories":["IT Services"],"tags":[],"author_name":"Subhashini Tripathi","publish_date":"2013-06-06T09:38:57","publication_year":"2013","word_count":600,"keywords":["Go","TPU","programming_languages:R","AI","programming_languages:Go","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","TPU","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/a-typical-day-in-the-life-of-an-analyst\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002164,"title":"Intel &#038; AMD’s Competing Offerings Make 2019 The Best Year For PC Enthusiasts","content":"PC enthusiasts just had one of their most action-packed weeks since a long time. All prominent players in the space are out of the gate with new products, refreshes, and advancements in computing. The technology alone is enough to pop the eyes of any PC enthusiast. Moreover, apart from the cutting edge advancements in technology, there is also a higher focus on the lower price points by the underdog, AMD. The back-and-forth between the two players in the CPU space forms the crux of why it is so exciting to be a PC enthusiast today. AMD’s Leapfrog Strategy The CPU battle has heated up vastly, and not just at the top-end where most of the competition is. Newer processor advancements have allowed AMD to compete at a market price point where Intel simply cannot match its performance. At the beginning of the week, AMD announced the launch of the Ryzen 3000 series of chips. Built on the famed 7nm manufacturing process, the new chips operate at a lower voltage while being more efficient, offering a sizeable performance increase over bigger chips. AMD CEO Lisa Su announces the Ryzen 3000 series. They also announced a new socket (X570) and a whole new lineup of motherboards for optimal compatibility with the new chips. It is also to be noted that these chips are still backward compatible with the existing AM4 sockets, except the older ones will not be compatible with some of the new features that the new chips bring to the table. Primary among them is the support for the next generation of the Peripheral Component Interconnect – Express. This standard, commonly referred to a PCI-Express, connects almost everything on the computer to each other, including storage, GPUs and other ASICs. With PCIe Gen.4, data transfers and GPU read and write speeds will be greatly increased. A New Chapter In Intel Vs. AMD Apart from handily beating one of Intel’s top-of-the-line i9 chips, Ryzen 3000’s flagship processor is also priced at less than half its price. AMD has begun taking over a part of the market which Intel has neglected in lieu of better higher-end offerings with more bells and whistles. The Ryzen’s chiplet design. At the price point of the highest-end AMD processor, which is priced at around $500, Intel (currently) sells the i7-9700K and i5-9600K processors. These are Ryzen 3000’s direct competitors at the top-end. If the top-end processor can punch above its weight class into the i9 series, the lower-end offerings can easily handle the challenge from the i5 and i9 Intel chips. This is the first time in a long time where AMD has taken the lead over Intel in IP and associated infrastructures. The reason for this can be seen as Intel’s wavering focus towards CPUs owing to other enterprise and enthusiast-focused offerings. Intel’s 9th Gen Core series is still built on the 12nm process, a whole two steps behind AMD. Not to be outdone, Intel released the details of the 10th Gen Core series, which would be built on the 10nm process. Intel has sorely needed this advancement to occur, as their ‘Intel Inside’ brand has been undercut both by recurrent hardware vulnerabilities and a general plateau in performance. Another Battlefield? Moreover, AMD is beginning to take on what was previously an absolute Intel monopoly; the laptop market. AMD’s laptop offerings, in the form of APUs, had been sorely lacking before the launch of Ryzen. Now, with specialized chips for low-powered laptops, AMD is also taking on Intel in the portables market as well. Intel, recognizing the potential loss of market share, is now looking to strategize the use of their 10nm process Core CPUs in laptops. This strategy is set to culminate in the release of a completely new type of laptops altogether. At the crux of wanting to take back market share, Intel has initiated a standard that they are calling ‘Project Athena’. Project Athena is the culmination of all the non-CPU tech Intel has been developing while AMD is innovating in better performance. With the cornerstones of ‘focus, always ready, adaptive’, Project Athena is the name for a new form factor and type of laptops. The standard sets rules for laptop manufacturers for the connected future. Primary among them is the ability to lucid sleep and sign in within 1 second using biometric authentication. The laptops also need to come with 8GB of RAM and an NVMe SSD for fast storage. There is also a heightened focus on better battery life and the integration of Intel technologies such as Thunderbolt 3, Wi-Fi 6, OpenVINO and more. At the core of all of this will be Intel’s new high-performance, low-powered 10th Gen Core processors. This space is moving towards more computing power and advancements over frills that offer no performance increase. This is a healthy trend for the PC market, which is an exciting time for anyone, across any price range.","excerpt":"PC enthusiasts just had one of their most action-packed weeks since a long time. All prominent players in the space are out of the gate with new products, refreshes, and advancements in computing. The technology alone is enough to pop the eyes of any PC enthusiast. Moreover, apart from the cutting edge advancements in technology, […]","categories":["AI Trends"],"tags":["AMD","CPU","Gaming","Intel","market"],"author_name":"Anirudh VK","publish_date":"2019-05-31T16:15:40","publication_year":"2019","word_count":818,"keywords":["CPU","AMD","programming_languages:R","market","AI","Gaming","RAG","R","Intel"],"extracted_tech_keywords":["AI","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/intel-amds-competing-offerings-make-2019-the-best-year-for-pc-enthusiasts\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046476,"title":"Top Weekly News: Tesla AI Day To Foundation Model","content":"Applications program interface (API) development platform Postman has emerged as the most valuable Indian SaaS company. Recently, it raised $225 million in Series D funding, raising its valuation to $5.6 billion. The company entered the Unicorn club last year. Postman has now overtaken platforms like Browserstack ($4 billion), Freshworks ($3.5 billion) and Icertis ($2.8 billion). Excited to share that we have closed on a $225 million Series D round that values Postman at $5.6 billion. Read more about it here! https:\/\/t.co\/B23YrmBk40— Abhinav Asthana (@a85) August 18, 2021 New York-based investor, Insight Partners, led the Series D funding round for Postman. Other investors included Coatue Management LLC, Battery Ventures, BOND, Nexus Venture Partners, and CRV. The company last raised funds during the Series C round ($150 million) in 2020. It has raised more than $430 million to date. The company said it would use the fresh funding to expand sales, marketing, product, and engineering teams. Further, the company will use the funds for its developers, design innovative API literacy programmes for students, and contribute to open-source projects to foster a thriving API ecosystem. Tesla AI Day After hosting Autonomy Day in 2019 and Battery Day in 2020, Tesla hosted AI Day on August 19. In July, Tesla head Elon Musk had announced via his Twitter account that the event would be held in August. Tesla AI Day August 19th— Elon Musk (@elonmusk) July 29, 2021 Tesla AI Day streamed on YouTube for general viewing, was essentially a virtual event with physical invites to select guests. The event opened with 45-minute long industrial music from The Matrix soundtrack. As also indicated by Musk earlier, the event revolved around explaining various Tesla techs at work, and inviting the best and the brightest AI talents to join the team. Convincing the best AI talent to join Tesla is the sole goal— Elon Musk (@elonmusk) July 29, 2021 One of the most exciting announcements made at this event was about the humanoid robot that the team is building. This Tesla bot will be 5’8” in height, weigh 125 pounds, and can deadlift 150 pounds. It can walk at 5 miles per hour and has a screen for a head to display important information. A prototype of this robot is likely to be unveiled next year. The most anticipated announcement about Tesla’s computer chip Dojo D1 was also made at the event. Tesla director Ganesh Venkataraman unveiled D1, which will be used to run its supercomputer Dojo. The idea behind this chip is that Dojo-trained AI software will be shared via over-the-air updates. Venkataraman claimed that Tesla technology would be the first AI-training computer. Further, Dojo will also be used for Tesla’s ambitious full driving system. Facebook VR App Facebook launched a test version of a new virtual reality remote work app Horizon Workworms. This app allows users to hold meetings as avatar versions of themselves via their Oculus Quest 2 headsets. As per early reviews, this app’s VR is more thoughtful than any other first-party social VR app launched by the company in the past. It also seems more productive and engaging than a regular video call. Facebook views this app launch as a step forward to building a futuristic metaverse. In July, Facebook founder Mark Zuckerberg said, “What I think is most interesting is how these themes will come together into a bigger idea. Our overarching goal across all of these initiatives is to help bring the metaverse to life.” Facebook seems to be investing heavily in virtual and augmented reality. In June, the social media giant acquired BigBox VR for an undisclosed amount. Before that, Facebook also bought Downpour Interactive, Ready at Dawn, Sanzaru Games, and Beat Games, starting from 2020. Foundation Models Over 100 researchers and scholars from ten Stanford University departments came together to publish a report on the opportunities and risks associated with foundation models like BERT and GPT-3. Foundation models are the ones trained on broad data at scale and are adaptable to a range of downstream tasks. This report gives an account of its capabilities such as language, vision, and human interaction; technical principles like model architecture, data, and training procedures; application areas like law, medicine, and education; societal impacts such as inequity, environmental impact, and ethical considerations. Credit: Report The authors also noted that even though the foundation models are based on deep learning and transfer learning, their capabilities and effectiveness incentivises homogenisation. While homogenisation is powerful leverage, one must be careful as the defects of foundation models are inherited by all adapted models downstream. Ola’s Electric Vehicle This week, Ola announced the rollout of Ola S1, a first in its range of electric two-wheelers. In a blog, company co-founder Bhavish Aggarwal announced that the scooter will be available in two variants — S1 and S1 pro. The vehicle control unit of the scooter consists of an Octa-core processor, 3GB RAM, and offers high-speed connectivity via 4G, Wi-Fi, and Bluetooth. It is powered by a multi-microphone array and has AI speech recognition algorithms built into it. The vehicle will be powered by Ola’s proprietary operating system called MoveOS. In terms of security features, Ola S1 includes an anti-theft alert system and geo-fencing.","excerpt":"One of the most exciting announcements made at Tesla AI Day was about the humanoid robot that the team is building.","categories":["AI News"],"tags":["Mark Zuckerberg","Ola Electric","Postman","postman api"],"author_name":"Shraddha Goled","publish_date":"2021-08-22T10:00:00","publication_year":"2021","word_count":865,"keywords":["Go","API","AI","Ola Electric","postman api","RAG","Mark Zuckerberg","Ray","Aim","deep learning","BERT","Postman","foundation models","R"],"extracted_tech_keywords":["AI","deep learning","foundation models","Aim","Ray","RAG","R","Go","API","BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/top-weekly-news-tesla-ai-day-to-foundation-model\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":17942,"title":"And Man Created God, Manthan&#8217;s Atul Jalan Gives Insights","content":"And man will create God again, says Atul Jalan. Artificial Intelligence has always fascinated Atul Jalan – from his Asimov & Clarke phase to his current role as Founder CEO of Manthan. In the past couple of years, Atul has incorporated AI into Manthan entire product range – delivering not just the world’s first conversational agent for business analytics but also taking recommendations to a purer level with AI. Speaking on his favourite topic (everyone’s favourite topic, actually), he cuts through what he calls the “hullabaloo, hyperbole and hubris” that surrounds AI to give us a new perspective on what awaits us in the near future. (And don’t you worry, the prophecy is divine!) Atul begins with the tiny change 70 million years ago that put primates inside, and humans outside the bars in the zoo – the opposable thump. If that tiny change could make us this mighty race, what could another change of that magnitude result in? This question, is the crux of Atul’s point. John McCarthy once said that as soon as it works, AI ceases to be AI. Atul proved that statement, by pointing the audience to their smartphones. Most of what our phone does today, was in the realm of science-fiction not too long back. Without pausing to think about it, we have already allowed the AI in our phones to influence our lives, considerably. Where will this take us? “Math and science have evolved to the point where it can absorb its own complexity; making the most sophisticated science applications, most intuitive,” Atul says. This means that adoption will not just be faster, it might also be without question. We will just keep making AI a part of ourselves. And as we augment this frail, fallible human self with the infallible power of AI and other emerging sciences, we might be making man’s second great transition. If the first was getting down from the tree, the second might be ascending into heaven. Having laid this goose-pimply foundation, Atul takes us (with enough and more examples) through what happens when genetics and AI gets added to ‘man’. We could well and truly transform into our own imagination of godhead; an omniscient, omnipresent, omnipotent divinity. When you think about it, isn’t that the path we have followed all along? We invent the story and then invent the technology that makes the story come true. And along this path to becoming gods, Atul also believes that we will be forced to ask ourselves (and answer) some really tough questions. Questions that we have often (conveniently) left to philosophy and religion – like what is it to be human, what is consciousness, where do reason and knowledge come from. For, to progress with AI, we need the answers to these questions. And for the first time in the history of mankind, we have the ability to look at these questions from an algorithmic – as opposed to a philosophic – perspective. Atul also touches upon the changes we will see as we don our new, divine garb. About how we will change how we look at jobs, love and sex, and the rules that govern our ‘human’ lives. So how do you and I actually become gods? Will this happen in our lifetime? The best person to answer those, is Atul Jalan.","excerpt":"And man will create God again, says Atul Jalan. Artificial Intelligence has always fascinated Atul Jalan – from his Asimov & Clarke phase to his current role as Founder CEO of Manthan. In the past couple of years, Atul has incorporated AI into Manthan entire product range – delivering not just the world’s first conversational […]","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Cypher","Manthan"],"author_name":"Priya Singh","publish_date":"2017-10-03T06:16:59","publication_year":"2017","word_count":554,"keywords":["Go","artificial intelligence","programming_languages:R","AI","IPO","programming_languages:Go","Manthan","analytics","R","AI (Artificial Intelligence)","Cypher"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","Go","IPO","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/man-created-god-atul-jalan-gives-insights\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10071797,"title":"Build a dream career in data engineering with the MTech program from SRM Institute of Science &#038; Technology","content":"The big data and data engineering services market is expected to see market growth at a rate of 24.13% in the forecast period of 2022 to 2029, as per a report by Data Bridge Market Research. Given the growth potential, now is the right time to invest in a course to help you build a rewarding career in big data. The  SRM Institute of Science and Technology (accredited and approved by AICTE) is offering an MTech in Data Engineering with a curriculum covering Artificial Intelligence and Machine Learning concepts extensively. The classes will be held at the Chennai Main campus during the weekends. Why SRM Institute of Science & Technology? SRM University is one of the best universities in the country and has won several accolades over the years: Secured rank #3 among top engineering institutes in India (Times Engineering Survey)Accredited with the highest ‘A++’ Grade (NAAC)Globally rated ‘Four Star’ University (QS World University Ranking)It has ranked #19 among the Top 20 Universities (MHRD – NIRF-2022) What’s in it for the students? Industry-relevant curriculum: The program enables participants to gain an in-depth understanding of data science, artificial intelligence and machine learning techniques and tools widely used in the industry. The program imparts real-world skills to enable learners to become job-ready. The program includes sessions from seasoned professors and industry experts. Networking opportunities: The program offers networking opportunities and mentorship from experts to gain industry insights. Dedicated recruitment drives: Students can attend job fairs conducted online and offline and get a chance to participate in the recruitment drives of top tech companies. Access to curated jobs: The career support team works with over 120000+ organisations to recommend the right jobs for the course participants. Personalised career mentorship: Students will get an expert career mentor to help them navigate the job market. Upon completing the course, the students can choose roles like data scientist, data architect, business intelligence engineer, or big data engineer. Curriculum Overview Semester 1: Python for data science, Machine Learning-supervised regression, database management systemSemester 2: Machine Learning-Supervised classification, Machine learning-unsupervised classification, time series forecasting Semester 3: Deep neural network architecture, NLP and its applications, cloud computing for data analyticsSemester 4: Advanced deep learning, big data architecture, open elective 1 Open Electives for Data engineering: Operations research, Industrial safety, Entrepreneurship and IPR, business analysis and communication, professional ethics Eligibility B.E\/B.Tech (or) M.Sc (or) MCA degree and a minimum of 50% in X, XII and Bachelor’s degree. Selection process Apply by filling out a simple online registration form The admissions committee will review and shortlist candidatesAdmissions testProgram selection interview The fee for this course is INR 5,00,000. You must pay the admission fee and first instalment before the program starts, followed by three equal quarterly instalments. ( Examination fee is not included in the program fee). And, there is more…Great Learning also offers a complementary Artificial Intelligence & Machine Learning” certification course worth Rs. 1,50,000 to everyone enrolling for the MTech in Data Engineering from SRM IST.This additional course will help participants to master multiple career-critical skills needed to achieve newer heights in their tech careers.The course covers the following topics:Data Analysis with PythonMachine Learning – Regression (Linear Regularized Regression) Machine Learning – Classification (Decision Trees, Ensemble models, Naive Bayes, KNN, End-to-End model development)Unsupervised Learning – KMeans & Hierarchical Clustering Advanced Data Visualization with Tableau So what are you waiting for? To register for this MTech program, click here.","excerpt":"The program enables participants to gain an in-depth understanding of data science, artificial intelligence and machine learning techniques and tools widely used in the industry.","categories":["AI Trends"],"tags":["Data Engineering","data engineering career","data engineering demand","data engineering jobs"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-08-01T10:10:00","publication_year":"2022","word_count":568,"keywords":["data science","artificial intelligence","machine learning","AI","neural network","cloud computing","data engineering career","data engineering demand","NLP","Python","deep learning","Data Engineering","analytics","data engineering jobs"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","NLP","data science","analytics","cloud computing","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/build-a-dream-career-in-data-engineering-with-the-mtech-program-from-srm-institute-of-science-technology\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10098885,"title":"Google DeepMind’s Unlikely Route of AI Stock Images","content":"In an attempt to eliminate misleading representation of AI in ‘stock imagery and pop culture’, Google DeepMind released Visualising AI. By partnering with artists, various artworks depicting AI themes related to research, technologies, and real-world consequences, such as artificial general intelligence (AGI), robotics, neuroscience, sustainability, and generative AI have been created. The images are available on stock-free image sites Unsplash and Pexels. With this, Google DeepMind forays into imagery — a new route that the company is paving for itself. Breaking Stereotypes The portrayal of AI usually overlooks worldwide viewpoints, leading to an absence of diversity that can magnify societal inequalities — visualising AI looks to address this. With a preconceived notion of how AI is depicted through streams of code, blue brains or white robots with men in suits, the repository of new images will contain artist-envisioned images inspired by conversations with scientists, engineers, and ethicists from Google DeepMind. All the art work — image and motion graphics — is available for anyone to download free of charge. Since its launch, Visualising AI has engaged with 13 artists and has created over 100 artworks which has gained over 100 million views and 800,000 downloads. These have been used by media outlets, research and civil society organisations. The Image Ride From robotics, protein folding, and maybe even AGI to this, Google DeepMind’s recent fascination with imagery is quite offbeat. However, Google has experimented with a number of image tools. Google’s text-to-image diffusion model Imagen, which is built on large transformer language models, was released in the beta mode. Google had also launched Muse, a text-to-image transformer model, and StyleDrop which allows the synthesis of images in a specific style using the Muse text-image model. Looking at other image generation tools in the market, Google has not been able to have a hold over the market. Midjourney, Stable Diffusion, Dall-E 2, and now with Adobe, Canva and other players, the market is being crowded with consumers spoilt for choice. Furthermore, OpenAI is said to be working on their next image-generation tool which is said to be better than Midjourney. While not being an image-generation tool, Visualising AI has entered the global stock images and video market, extending its influence in this domain. In 2022, the market size of worldwide stock images and videos was assessed at approx. $4.96 billion. By 2028, the market is expected to rise to $7.33 billion. The surge in digital media and heightened need for visual content across diverse industries has led to substantial market growth. Way Through Partnerships Recently, Google partnered with Adobe to bring Firefly and Express to Bard. The integration will allow users to generate Firefly images within Bard and these images can be further modified using Express. With Big Techs wanting to diversify into domains that are not their main stream of business, partnerships have been the recent go-to strategy. However, each of the partnerships usually has a larger implication for the company in question. OpenAI has been actively partnering with a diverse range of companies, from media to product firms. Notably, they acquired Global Illumination, a product company which employs AI for crafting creative tools. OpenAI’s collaborations extend to news media such as Associated Press and American Journalistic Project too, aiming to gain a foothold in the media industry and enhance model training. With Visualising AI, Google DeepMind partnering with artists to provide images on Pexels and Unsplash might just be the beginning. While currently it is offering images on AI and other technologies in the same space, it is possible that in the future Google DeepMind might venture into other image categories. Considering how other companies are vigorously pursuing image-generation tools, Google DeepMind might be slowly catching up.","excerpt":"With Visualising AI, Google DeepMind takes an unconventional approach to create stock images that break AI stereotypes. But, why?","categories":["Deep Tech"],"tags":["Adobe","AGI","AI Tool","art","artists","Associated Press","DALL.E","Google Deepmind","image-generation","Imagen","Microsoft","MidJourney","OpenAI","protein folding","Robotics","Stable Diffusion"],"author_name":"Vandana Nair","publish_date":"2023-08-23T14:19:32","publication_year":"2023","word_count":617,"keywords":["artists","Git","Robotics","Ray","AI Tool","R","Adobe","Stable Diffusion","art","DALL.E","Go","DALL-E","AI","Imagen","Associated Press","Google Deepmind","AGI","generative AI","stable diffusion","MidJourney","OpenAI","image-generation","Aim","protein folding","Microsoft"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Aim","Ray","R","Go","Git","DALL-E","stable diffusion"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/google-deepminds-unlikely-route-of-ai-stock-images\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":50777,"title":"Can Deepfakes Be Used For The Good Of Humanity?","content":"Deepfakes, an unknown word in the traditional dictionary, has found a morbid meaning in today’s world of technology and media. Everyone in today’s era of AI revolution has heard of deepfakes and has heard about the worst of it. Be it Obama’s public service announcement deepfake or Zuckerberg’s ‘truth’ about privacy on Facebook video, deepfake has shown how it can have an impact on misleading the viewers. If people can’t tell the difference between a real video and a deepfake one, then it becomes an easy way to manipulate the public. It is, therefore, difficult to imagine that there is a better way to utilise deepfakes, a good way to leverage its technology. But, there are such ways: The World Of Visual Storytelling The business of visual storytelling is expensive. Hollywood studios, especially, spend billions of dollars collectively to bring an artist’s imagination to screen. These studios use costly tools to try and bridge the gap between fantasy and reality. Movies like Avatar and the Avengers franchise are such examples — but have we ever wondered about the small-scale creators who do not have that kind of budget or expensive tools? Deepfakes technology incorporates such potential. It can help the small-scale creators to have a similar capacity to bring their imaginative creativity as close to reality as possible. For example: Political Satire isn’t some big application that might be worth mentioning here, but since deepfakes nowadays are used mainly in a political context, why not? Political satires are used to gain entertainment from politics with a touch of humour. Countries like the UK encourage this practice. The artist performing a political satire doesn’t have to be guarded all the time because their country has no problem with it, but what about the countries where it is forbidden? Call it silly, but in some parts, deepfake actually can have a contribution towards good in politics. Voice Manipulation Is Not All That Bad Suppose you are an artist who has a good voice and talent, you can potentially take your career global by using voice manipulation in different languages. There are companies like Synthesia who have shown their commercial product where they used David Beckham’s face to talk in nine different languages. Technology like this can do wonders to someone’s career when they use it inside the ethical boundaries. There is software in work by Adobe which lets you produce speech from tech and edit it the way you modify the photographs in photoshop. Another angle one can look at this is from the perspective of the advantage the big firms hold when it comes to distribution and creation. The amateur artists will have more free space and can let their creativity run in various directions through such advantages that one can explore through deepfakes. Virtual Humans What are Virtual Humans? A simple example is the movie; Blade Runner 2049’s character Joi. What you can get from this is an application of deepfake that has created a virtual human for another human being which acts like a companion. These types of applications of the deepfakes are named as digital influencers. Now, this is far in the future, but if you think about it, not so much. In today’s time, an example would be Alexa. What we watch in the movies like Blade Runner 2049 and Her are the extensions of digital influencers like Alexa. Deepfake is an application of machine learning, where it takes data from human behaviours as input and gives output as elaborate on that behaviour. It is what makes a digital influencer to understand and originate a conversation with a certain degree of sophistication. Potential Applications In Healthcare Yes, when someone mentions healthcare with machine learning, concerns about privacy go off — especially regarding the medical history of a patient. As the debate goes on, researchers are finding a way to the AI and privacy problem with more application of AI to it. For AI to diagnose a disease, it needs to learn through many examples of patients, real patients. That’s where deepfake is trying to find its application. Though there are some limitations to overcome, Deepfake application like Generative Adversarial Networks (GANs) can create medical images for AI which are real enough for it to learn. Outlook: Deepfake is one of the best examples to prove how far artificial intelligence has come. While it is infamous for defaming celebrities, interfering with political information or promoting a vengeful and illegal activity called ‘revenge porn’, it has made people more aware of their environment. Since deepfake started with fake videos, people are aware and more lenient towards finding out the truth. Until companies like Google and Facebook succeed towards spotting real from the fake, we might not have any other option other than to sit tight and try to take its advantage because technology like this is subjective to the context it is used. Also check-out: https:\/\/www.youtube.com\/watch?v=dkoi7sZvWiU&feature=youtu.be","excerpt":"Deepfakes, an unknown word in the traditional dictionary, has found a morbid meaning in today’s world of technology and media. Everyone in today’s era of AI revolution has heard of deepfakes and has heard about the worst of it. Be it Obama’s public service announcement deepfake or Zuckerberg’s ‘truth’ about privacy on Facebook video, deepfake […]","categories":["AI Features"],"tags":["deep fake","deep learning application examples","DeepFake","DeepFake videos","deepfakes","deepfakes and AI","human intelligence at machine scale"],"author_name":"Sameer Balaganur","publish_date":"2019-11-27T17:00:00","publication_year":"2019","word_count":817,"keywords":["Go","DeepFake videos","human intelligence at machine scale","artificial intelligence","AI","machine learning","TPU","deepfakes","Git","RAG","GAN","deep fake","deepfakes and AI","ViT","deep learning application examples","DeepFake","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","TPU","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-deepfakes-be-used-for-the-good-of-humanity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10071661,"title":"LatentView records its highest ever quarterly revenue","content":"Chennai-based pure-play analytics firm ‘LatentView’ announced its first quarterly results for FY23 today. The company reported a revenue of INR 120 crore, witnessing the highest ever revenue in this quarter, growing at 37 per cent YoY. Its financial services vertical reportedly grew by 17 per cent, and the technology vertical, one of the company’s largest verticals, grew by 15 per cent during FY23 Q1. However, LatentView’s EBITDA margin for the quarter witnessed a marginal drop at 29 per cent, compared to the previous quarter, which was at 30.5 per cent. The reason—latest investments made by the company in the front end and marketing activities, alongside hiring via campus hires i.e., a total of 103 employees were added in the last few quarters, out of which 85 were campus hires. “These employees go through 2-3 months of training programme\/bootcamp before [being] deployed on the projects”, shares Rajan Venkatesh, CFO at LatentView. At the earnings call, Venkatesh revealed that the company recorded its highest PBT, which stood at INR 41.8 crores; a growth of 46.4 per cent YoY; and 3.7 per cent QoQ. Further, its cash and investments as of June 30 2022, stood at INR 587 crores, where the company considers spending this cash in the coming months across various avenues, M&As and more. Analytics revenue breakdown Breaking down revenue from an analytics services perspective, LatentView chief Rajan Sethuraman told Analytics India Magazine that about 10~15 per cent of the revenue comes from upfront analytics consulting and road-mapping work. Meanwhile, data engineering, data platforms and architecture as well as deploying data solutions contributed about 25 per cent of its revenue. “In total, these two make up about 40 per cent”, Sethuraman adds. The remaining 60 per cent of the revenue is split into two: According to LatentView—it’s called ‘look back’ and ‘look ahead’ analytics. ‘Look back’ is a diagnostic, descriptive type of work, contributing about 40 per cent of their revenue. The remaining ‘look ahead’, which is oriented more towards predictive and prescriptive analytics—using AI and ML models—i.e., about 20 per cent. From an industry perspective, 65 per cent of revenue stems from technology and digital native companies. Sector-wise, banking, financial services, retail and consumer segments witnessed huge demand for their services, besides automotive, oil and gas, logistics and manufacturing, and others. “We added three new accounts”, exclaims Sethuraman. In addition, he said that they had significant growth in many of their existing accounts. “Last year, we added about 18 accounts. This year, we started with three in the first quarter, but right now, there are conversations with eight new accounts, which we expect to close in quarter two”, Sethuraman adds. Citing the early days of deals, Rajan Sethuraman said that the quantum of the worth of the first SOW (statement of work) was around the $200~250 range. “In recent times, that has gone up significantly. Today, when we take on new accounts, most start at least half a million ($500K). And we are expecting that trend to continue”, concludes Sethuraman.","excerpt":"The company recorded its highest PBT, which stood at INR 41.8 crores.","categories":["AI News"],"tags":["AI in finance","latentview"],"author_name":"Amit Naik","publish_date":"2022-07-27T19:30:40","publication_year":"2022","word_count":500,"keywords":["latentview","Go","programming_languages:R","AI","ML","Git","AI in finance","Aim","data engineering","ViT","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","R","Go","Git","data engineering","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/latentview-records-its-highest-ever-quarterly-revenue\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10140792,"title":"OpenAI is So Doomed if Inference Time Scaling for o1 Fails","content":"OpenAI’s progress from GPT-4 to Orion has slowed, The information reported recently. According to the report, although OpenAI has completed only 20% of Orion’s training, it is already on par with GPT-4 in intelligence, task fulfilment, and question-answering abilities. While Orion outperforms previous models, the quality improvement is less dramatic than the leap from GPT-3 to GPT-4. This led many to wonder—Have LLM improvements hit a wall? No one seemed more thrilled about it than the most celebrated AI critic, Gary Marcus, who promptly posted on X, “Folks, game over. I won. GPT is hitting a period of diminishing returns, just like I said it would.” However, it appears Uncle Gary may have celebrated a bit too early. “With all due respect, the article introduces a new AI scaling law that could replace the old one. The sky isn’t falling,” one of the article’s authors quickly responded, clarifying the point to Marcus. Similarly, OpenAI researchers were quick to correct the narrative, asserting that the article inaccurately portrays the progress of OpenAI’s upcoming models—or rather misleading. “There are now two key dimensions of scaling for models like the o1 series—training time and inference time,” said Adam Goldberg, a founding member of OpenAI’s go-to-market (GTM) team. He explained that while traditional scaling laws focusing on pre-training larger models for longer are still relevant, there’s now another important factor. “Aspect of scale remains foundational. However, the introduction of this second scaling dimension is set to unlock amazing new capabilities,” he added. He was elaborating on OpenAI researcher Noam Brown’s earlier statement claiming that o1 is trained with reinforcement learning (RL) to “think” before responding via a private chain of thought. “The longer it thinks, the better it performs on reasoning tasks,” he had said. This, Brown explained, introduces a new dimension to scaling. “We’re no longer bottlenecked by pretraining. We can now scale inference compute as well,” he added. Jason Wei, also a researcher at OpenAI, defended o1 and explained the difference in the chain of thought before and after o1. He explained that the traditional chain-of-thought reasoning used by AI models like GPT was more of a mimicry than a true “thinking” process. He said the model would often reproduce reasoning paths it encountered during its pretraining, like solutions to math problems or other tasks. He added that the o1 system introduces a more robust and authentic “thinking” process. In this paradigm, the chain of thought reflects more of an internal reasoning process, similar to how humans think. He explained that instead of simply spitting out an answer, the model engages in an “inner monologue” or “stream of consciousness,” where it actively considers and evaluates options. “You can see the model backtracking; it says things like ‘alternatively, let’s try’ or ‘wait, but’,” he added. This back-and-forth process is a more dynamic and thoughtful approach to solving problems. “People underestimate how powerful test-time compute is: compute for longer, in parallel, or fork and branch arbitrarily—like cloning your mind 1,000 times and picking the best thoughts,” said Peter Welinder, VP of product at OpenAI. Earlier, when OpenAI released o1-mini and o1-preview, they mentioned in their blog post that o1’s performance consistently improves with more reinforcement learning (train-time compute) and with more time spent thinking (test-time compute). Regarding inference time scaling, they said, “The constraints on scaling this approach differ substantially from those of LLM pretraining, and we are continuing to investigate them.” It appears that OpenAI has exhausted all available data for pre-training the model and is now exploring new methods to improve o1. According to The Information report, Orion was partially trained on AI-generated data (or synthetic data) produced by other OpenAI models, including GPT-4 and the recently released reasoning models. Jensen to the Rescue: When NVIDIA CEO Jensen Huang recently said “We Are Going to Take Everybody with Us,” he really meant it. In a recent podcast with No Priors, Huang shared that one of the major challenges NVIDIA is currently facing in computing is inference time scaling, which involves generating tokens at incredibly low latency. Huang explained that, in the future, AI systems will need to perform tasks like tree search, chain of thought, and mental simulations, reflecting on their own answers. The model would prompt itself and generate text internally, all while responding in real-time, ideally within a second. This approach subtly points to the capabilities of the o1 system. AGI Coming Soon? While others remain uncertain, OpenAI chief  Sam Altman is confident that artificial general intelligence (AGI) is closer than many think. In a recent interview with Y Combinator’s Garry Tan, Altman suggested that AGI could emerge as soon as 2025. “I think we are going to get there faster than people expect,” he said, underscoring OpenAI’s accelerated progress. Further, he said OpenAI had fewer resources than DeepMind and others. “So we said, ‘Okay, they are going to try a lot of things and we have just got to pick one and really concentrate’,” he added. “I’ve heard people claim that Sam is just drumming up hype, but from what I’ve seen everything he’s saying matches the median view of OpenAI researchers on the ground,” said Brown. OpenAI has yet to release o1 fully. While it may not perform well in math and coding at this stage, it doesn’t mean it won’t improve over time. Many believe that o1 could be the first commercial application of System 2 thinking. In EpochAI’s FrontierMath benchmark, which tests LLMs on some of the hardest and unpublished problems in math, it was revealed that only 2% of these problems were successfully solved by LLMs. While all models showed poor performance, the o1 preview showed a positive sign, as it was able to consistently solve problems correctly in repeated testing. Apple recently published a paper titled ‘Understanding the Limitations of Mathematical Reasoning in Large Language Models’, which said that the current LLMs can’t reason. The researchers introduced GSM-Symbolic, a new tool for testing mathematical reasoning within LLMs because GSM8K was not accurate enough and, thus, not reliable for testing the reasoning abilities of LLMs. Surprisingly, on this benchmark, OpenAI’s o1 demonstrated “strong performance on various reasoning and knowledge-based benchmarks” according to the researchers. However, the capabilities dropped by 30% when the researchers introduced the GSM-NoOp experiment, which involved adding irrelevant information to the questions. N-o1 Saw This Coming, Not Even OpenAI Subbarao Kambhampati, a computer science and AI professor at Arizona State University said that some of the claims of LLMs being capable of reasoning are “exaggerated”. He argued that LLMs require more tools to handle System 2 tasks (reasoning), for which techniques like fine-tuning or chain of thought are not adequate. “When we develop AI systems that can actually reason, they will involve deep learning (as one of two major components, the other being discrete search). Some people might argue that this ‘proves’ deep learning can reason,” said François Chollet, the creator of Keras. “But that’s not true. It will prove that deep learning alone isn’t enough and that we need to combine it with discrete search,” Chollet added. Pointing to the inclusion of Gemini in AlphaProof, he described it as “basically cosmetic and for marketing purposes”. He argued that this reflects a wider trend—using the ‘LLM’ brand name as a blanket term for all AI progress, even though much of it is unrelated to LLMs. When OpenAI released o1, claiming that the model thinks and reasons, Hugging Face CEO Clem Delangue was not impressed. “Once again, an AI system is not ‘thinking’; it’s ‘processing,’ ‘running predictions’… just like Google or computers do,” said Delangue, adding that OpenAI is “selling cheap snake oil”. However, all is not lost for OpenAI, Google DeepMind recently published a paper titled ‘Chain of Thought Empowers Transformers to Solve Inherently Serial Problems’. While sharing his research on X, Denny Zhou mentioned, “We have mathematically proven that Transformers can solve any problem, provided they are allowed to generate as many intermediate reasoning tokens as needed.” This echoes AI researcher Andrej Karpathy’s recent remarks on next-token prediction frameworks, suggesting that they could become a universal tool for solving a wide range of problems, far beyond just alone text or language.","excerpt":"But Sam Altman and his team are taking their biggest risk ever to bring AGI next year.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-11-11T19:34:11","publication_year":"2024","word_count":1360,"keywords":["Hugging Face","Keras","OpenAI","AI","AWS","Transformers","Ray","Aim","deep learning","chain of thought"],"extracted_tech_keywords":["AI","deep learning","OpenAI","Aim","Ray","Keras","Hugging Face","Transformers","chain of thought","AWS"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-is-so-doomed-if-inference-time-scaling-for-o1-fails\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10048169,"title":"Andrew Ng Announces The Launch Of NeurIPS Data-Centric AI Workshop","content":"DeepLearning.AI’s Andrew Ng recently announced the launch of the NeurIPS Data-Centric AI workshop. The workshop is expected to showcase some of the best academic research work related to data-centric AI. Academic researchers and practitioners can submit their research papers on or before September 30, 2021. Announcing the NeurIPS Data-Centric AI workshop! This will be the premier event for academic research on Data-Centric AI. Deadline for research article submissions is Sep 30th. Check it out! https:\/\/t.co\/PV9uNJZyVa— Andrew Ng (@AndrewYNg) September 7, 2021 The organising committee includes Google Research’s Lora Aroyo, Stanford University professor Cody Coleman, Landing AI’s Greg Diamos, Harvard University professor Vijay Janapa Reddi, Eindhoven University of Technology researcher Joaquin Vanschoren, and Google’s machine learning product manager, Sharon Zhou. What is Data-Centric AI? Data-Centric AI, or DCAI, represents the recent transition from modelling to the underlying data used to train and evaluate models. DCAI aims to address the gap in tooling, best practices, and infrastructure for managing data in modern ML systems. Plus, it looks to offer high productivity and efficient open data engineering tools to make building, maintaining, and evaluating datasets cost-effective and seamless. The team strives to cultivate the DCAI community into a vibrant interdisciplinary field and tackle practical data problems with this event. The data problems include: data collection\/generation, data labelling, data preprocess\/augmentation, data quality evaluation, data debt, and data governance. The team believes that many of these areas are still in the early stages and hope to knit the gaps by bringing the ML community together. Call for Papers The journey of building and using datasets for AI systems is often artisanal — painstaking and expensive. The ML community lacks high productivity and efficient open data engineering tools. To accelerate creation and iteration, alongside increasing the efficiency of use and reuse by democratising data engineering and evaluation, remains a core challenge even to this day. “If 80 per cent of machine learning work is data preparation, then ensuring data quality is the most important work of an ML team and therefore a vital research area,” said the NeurIPS DSAI team. Further, they said human-labelled data has increasingly become the fuel and compass of AI-based software systems, while innovative efforts have mostly focused on models and code. However, in recent years, there has been an increased focus on scale, speed, and cost of building and improving datasets, which has, in turn, resulted in an impact on quality. Some of the major research work in the areas include ‘Response-based Learning for Grounded Machine Translation,’ ‘Crowdsourcing with Fairness, Diversity and Budget Constraints,’ Excavating AI, ‘Data Excellence: Better Data for Better AI,’ ‘State of the Art: Reproducibility in Artificial Intelligence, and others. “We need a framework for excellence in ‘data engineering’ that does not yet exist,” said the NeurIPS DCAI team, and noted that aspects like maintainability, reproducibility, reliability, validity and fidelity of datasets are often overlooked when releasing the dataset into the market. In this event, the team plans to highlight examples, case studies, and methodologies for excellence in data collection. The NeurIPS DCAI team said that building an active research community focused on data-centric AI is critical for defining the core problems and creating ways to measure progress in machine learning through data quality tasks. Topics The interested candidate can submit their papers on the following topics that include but are not limited to the following: New Datasets in areas: Speech, vision, manufacturing, medical, recommendation\/personalisation Science Tools and methodologies that Quantify and accelerate time to source high-quality data Ensure data is labelled consistently, such as label consensus Improve data quality more systematically. Automate the generation of high quality supervised learning training data from low-quality resources, such as forced alignment in speech recognition Produce uniform and low noise data samples, or remove labelling noise or inconsistencies from existing data Control what goes into the dataset and make high-level edits efficiently to very large datasets, like adding new words, languages, etc. Search techniques for finding suitably licensed datasets based on public resources Create training datasets for small data problems or rare classes in the long tail of big data problems Incorporate timely feedback from production systems into datasets Understand dataset coverage of important classes and editing them to cover newly identified important cases Import dataset by allowing easy combination and composition of existing datasetsExport dataset by making the data consumable for models and interface with model training and inference systems such as web dataset Enable composition of dataset tools like MLCube, Docker, Airflow Algorithms for working with limited labelled data and improving label efficiency Data selection techniques like active learning and core-set selection for identifying the most valuable examples to label Semi-supervised learning, few-shot learning, and weak supervision techniques for maximising the power of limited labelled data Self-supervised learning and transfer learning approaches for developing powerful representations used for many downstream tasks with limited labelled data. Novelty and drift detection to identify and spot when more data needs to be labelled Responsible AI development: Fairness, bias, diversity evaluation and analysis for dataset and algorithms\/modelling Tools for ‘green AI hardware-software system’ design and evaluation Scalable, reliable training systems and methods Tools, methodologies, and techniques for private, secure ML training Efforts towards reproducible AI (data cards, model cars, etc.) Instructions for submitting papers Researchers can submit short papers (1-2 pages) and long papers (4 pages), addressing one or more of the topics  Papers need to be formatted as per NeurIPS 2021 guidelines Papers will be peer-reviewed by the programme committee Accepted papers will be presented as lighting talks during the workshop Timeline Early submission deadline: 17 September 2021Submission deadline: 30 September 2021 Notification of acceptance: 22 October 2021 Workshop: 14 December 2021 Click here to submit your research paper.","excerpt":"If 80 per cent of machine learning work is data preparation, then ensuring data quality is the most important work of a machine learning team","categories":["AI Features"],"tags":["AI latest","Andrew Ng","big data quality","data science and manufacturing","Machine Learning","Machine Learning Latest"],"author_name":"Amit Naik","publish_date":"2021-09-12T12:00:00","publication_year":"2021","word_count":947,"keywords":["Go","big data quality","data science and manufacturing","artificial intelligence","machine learning","AI","ML","Machine Learning","Machine Learning Latest","docker","Andrew Ng","RAG","Aim","few-shot learning","AI latest","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","few-shot learning","docker","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/andrew-ng-announces-the-launch-of-neurips-data-centric-ai-workshop\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":53775,"title":"10 Cybersecurity Internships You Should Apply For Right Away","content":"Cybersecurity is an ever-changing field. There are more vulnerabilities in most sectors now, than there were a few years ago. In fact, data breaches ran at a record pace in 2019. There were over 3,800 publicly-disclosed breaches and 4.1 billion exposed records in just the first half of 2019. In one of our previous articles, we discussed how bad is the cybersecurity talent gap in the industry. Now, in this article, we list down ten internships in cybersecurity you can look at applying for. Cyber Forensics Internship at Cybersafe Location: Bangalore Stipend: ₹8,000 per month Responsibilities: Cybersafe provides advanced CTO\/CISO level IT security consulting services for small, medium, and large enterprises. The selected candidate’s responsibilities will include identifying new systems\/networks vulnerable to cyber-attacks; dismantling and rebuilding damaged systems to retrieve lost data; drafting technical reports; writing declarations and preparing evidence for trial, and advising LEAs on the credibility of acquired data. Deadline: 15 Jan 2020 Apply here. Information Security & Ethical Hacking Internship at 4 IR Location: Faridabad Stipend: ₹5,000 per month Responsibilities: The selected candidate’s responsibilities will include learning about new vulnerabilities that have been disclosed recently, performing VAPT on live projects and writing reports of the analysis. The duration of the internship will be for six months, and on the successful completion of the internship, the candidate will get a job offer from the organisation. Deadline: 22 Jan 2020 Apply here. Ethical Hacking Internship at ShieldByte Infosec Location: Mumbai Stipend: ₹5,000 to ₹6,000 per month Responsibilities: The selected candidate’s responsibilities will include performing threat modelling, performing architectural analysis, performing a logical security assessment, monitoring third-party APIs, SDK, and libraries, generating assessment report, reviewing false positives and true positives, analysing reports from an interactive source code review tool for false positives and creating a report on it and other such. Deadline: 15 Jan 2020. Apply here. Software Testing Internship at ArtSpot India Location: Jaipur Stipend: Not Disclosed Responsibilities: ArtSpot India is a Dubai-based company that deals with cybersecurity and designing. The selected candidate’s responsibilities will include working on the functional testing of websites and mobile app, working on OWASP testing and working on Burp Suite. The duration of the internship will be of 3 months, and the skills required are PHP, JavaScript, Ethical Hacking, and C++ programming. Deadline: The internship can be started between 7th Jan’20 and 6th Feb’20. Apply here. Cyber Security Internship at HackerEarth Location: Bangalore Stipend: Not Disclosed Responsibilities: The selected candidate’s responsibilities will include creating problems in cybersecurity domain based on the client’s requirements, setting up backend or infra for the challenge, creating all the necessary resources for the challenge, performing R&D for possible areas in cybersecurity which can be tested in a challenge. Deadline: 13 Feb 2020 Apply here. Core Cyber Security Internship at TeleNetworks Technologies Location: Mumbai Stipend: Not Disclosed Responsibilities: Only Students from EXTC, Comps and IT branches can apply for this internship. The selected candidate’s responsibilities will include troubleshooting with the live networks of CISCO Routers & Switches in technologies like BGP, OSPF, Data Policing, IP multicasting, VPLS, DMVPN, IPv6, telecom training into radio frequency, 2G, 3G, 4G LTE Advanced, Synchronous Digital Hierarchy (SDH), EMF ( NARDA Tool ), and other such. Deadline: Not Disclosed Apply here. Cyber Security Internship at Innover Systems Location: Pune Stipend: Not Disclosed Responsibilities: The selected candidate’s responsibilities will include performing threat modelling, architectural analysis, assessing logical security, monitoring third-party API- s, SDK, generating assessment report, reviewing false positives and true positives, analysing reports from an interactive source code review tool for false positives and including it in the report. Deadline: Not Disclosed Apply here. Security Trainee Executive at SevenMentor Pvt Ltd Location: Pune Stipend: Not Disclosed Responsibilities: The selected candidate’s responsibilities will include learning concepts, analysing primary logs, network concepts, handling customer calls, assisting the Security Analyst in their day-to-day functions, such as preparing reports and analysing security traffic, performing Android and IOS based mobile application testing, vulnerability assessment, and penetration testing, and other such. Deadline: Not Disclosed Apply here. Security Architect at JRD Systems Location: Bangalore Stipend: Not Disclosed Responsibilities: The selected candidate’s responsibilities will include designing, building and implementing enterprise-class security systems for a production environment; aligning standard frameworks; securing overall business and technology strategy; identifying and communicating emerging security threats, and designing security architecture elements to mitigate risks as they arise. Deadline: Not Disclosed Apply here. Cyber Security Intern at Leapify Location: Work from home. Stipend: ₹5,000 to ₹8,000 per month Responsibilities: The selected candidate’s responsibilities will include working on website development (task by task), handling weekly meetings, managing daily reporting. The duration of the internship is for 2 months and the skills required are .NET, data structures, database management system (DBMS) and database testing. Deadline: 23 Jan 2020 Apply here.","excerpt":"Cybersecurity is an ever-changing field. There are more vulnerabilities in most sectors now, than there were a few years ago. In fact, data breaches ran at a record pace in 2019. There were over 3,800 publicly-disclosed breaches and 4.1 billion exposed records in just the first half of 2019.  In one of our previous articles, […]","categories":["AI Trends"],"tags":["Ethical Hacking","Virtual Internship Program"],"author_name":"Ambika Choudhury","publish_date":"2020-01-14T11:08:09","publication_year":"2020","word_count":789,"keywords":["Go","API","programming_languages:R","AI","Git","Virtual Internship Program","C++","JavaScript","GAN","Ethical Hacking","R","Java"],"extracted_tech_keywords":["AI","R","JavaScript","Go","Java","C++","Git","API","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-cybersecurity-internships-to-apply-for\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10056829,"title":"Pune Based Unbox Robotics Raises USD 7 Million","content":"Unbox Robotics, a supply chain robotics technology company, raised USD 7 million in a Series-A round led by 3one4 Capital with participation from Sixth Sense Ventures and Redstart Labs (Info Edge). The company will use the funds to expand the team across functions, R&D, fulfil the customer demand internationally, and expand to new geographies. It has onboarded some large e-commerce and logistics enterprises as early adopters through pilots and has filed IP for the technology in India, the US and the EU. Pramod Ghadge, CEO and Co-Founder of Unbox Robotics, said, “We believe that there’s a massive potential for building and deploying robotics tech to deliver more packages using smaller footprints at lower operational costs as we see more adoption of e-commerce and Q-commerce (quick commerce) across the globe. Since we launched our beta pilots with leading third-party logistics and e-commerce players in April 2021, we have already got orders from some of the leading e-commerce logistics companies.” He also informed that the company is on their way to convert every other pilot or demo into a commercial contract. “We will deploy the funds to build an A+ team to help our customers experience the future of logistics, sorting and supply chain across continents,” added Ghadge. Unbox Robotics specialises in robotics-based fulfilment and distribution technology for small to large e-commerce, retail and logistics enterprises. Its proprietary plug-and-play Swarm robotics can bring down the deployment time at locations to less than a week. The product can scan, sort, and dispatch packages in less than 50-70% physical space, improve personnel productivity by three times, and reduce operational cost by 60%. The system allows users to automate bigger facilities with a physical space of a few hundred thousand sq ft, as well as automate micro-hubs with less than 2,000 sq ft space through a compact vertical sorting mechanism. Founder and CEO of Sixth Sense Ventures, Nikhil Vora, said, “We believe the e-commerce and D2C boom is creating a large need for automation solutions within warehouses and FCs, with the industry slated to grow to $32 billion over the next 2-3 years. This, coupled with evolving consumer expectations, labour shortage, omnichannel distribution, and technology accessibility, has made robotics a priority for both e-commerce and fulfilment companies.” Anurag Ramdasan, Partner, 3one4 Capital, added, “Unbox Robotics is doing amazing work in identifying and solving these high-value problems and the team built by Pramod, Shahid and Rohit bring deep domain expertise to tackle this problem at scale.” The round saw participation from Unbox Robotics’ founders including Pramod Ghadge, Shahid Memon and its CPO Rohit Pitale. Its current investors who participated included SOSV, Arali Ventures, BEENEXT, Dr Vijay Kedia (Kedia Securities), Karthik Bhat’s Force Ventures, Aditya Singh (Stride Ventures), WEH Ventures and Pavitar Singh (Sprinklr). Other participating investors include Nikhil Vora and Kathan Shah (Sixth Sense Ventures), Rahul Chaudhary (Treebo Hotels), and Veda VC.","excerpt":"Unbox will use the funds to expand the team & the company to new geographies, fulfil international customer demand and R&D.","categories":["AI News"],"tags":["Funding","investment","Supply chain"],"author_name":"Meeta Ramnani","publish_date":"2021-12-22T14:54:06","publication_year":"2021","word_count":476,"keywords":["Go","Supply chain","Funding","investment","API","programming_languages:R","AI","programming_languages:Go","RAG","automation","ViT","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","RAG","R","Go","API","ViT","automation","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pune-based-unbox-robotics-raises-usd-7-million\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011365,"title":"Quick Guide To Survival Analysis Using Kaplan Meier Curve (With Python Code)","content":"Today, with the advancement in technology, Survival analysis is frequently used in the pharmaceutical sector. It analyses a given dataset in a characterised time length before another event happens. The Kaplan Meier estimator is an estimator used in survival analysis by using the lifetime data. In medical research, it is frequently used to gauge the part of patients living for a specific measure of time after treatment. Here, we will implement the survival analysis using the Kaplan Meier Estimate to predict whether or not the patient will survive for at least one year. About the dataset The dataset can be downloaded from the following link. It gives the details of the patient’s heart attack and condition. Code Implementation Install all the libraries required for this project. pip install lifelines import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns import statistics from sklearn.impute import SimpleImputer from lifelines import KaplanMeierFitter, CoxPHFitter from lifelines.statistics import logrank_test from scipy import stats Reading the Data df = pd.read_csv(\"echocardiogram.csv\") df.head() Data Pre-Processing Let us check for missing values and impute them with mean values. mean = SimpleImputer(missing_values = np.nan, strategy = 'mean') Columns = ['age', 'pericardialeffusion', 'fractionalshortening', 'epss', 'lvdd', 'wallmotion-score'] X = mean.fit_transform(df[Columns]) df_X = pd.DataFrame(X, columns = Columns) keep = ['survival', 'alive'] df_keepcolumn = df[keep] df = pd.concat([df_keepcolumn, df_X], axis = 1) df = df.dropna() print(df.isnull().sum()) print(df.shape) Create a new column df.loc[df.alive == 1, 'dead'] = 0 df.loc[df.alive == 0, 'dead'] = 1 df.groupby('dead').count() Kaplan Meier Curve kmf = KaplanMeierFitter() X= df['survival'] Y = df['dead'] kmf.fit(X, event_observed = Y) kmf.plot() plt.title(\"Kaplan Meier estimates\") plt.xlabel(\"Month after heart attack\") plt.ylabel(\"Survival\") plt.show() From the plot we can see that the survival rate decreases with the increase in the number of months.The Kaplan estimate is 1 for the initial days following the heart treatment.It gradually decreases to around 0.05 after 50 months. print(\"The median survival time :\",kmf.median_survival_time_) The average survival time of patients is 29 months.Given below is the KM_estimate that gives the probability of survival after the treatment. print(kmf.survival_function_) age_group = df['age'] < statistics.median(df['age']) ax = plt.subplot(111) kmf.fit(X[age_group], event_observed = Y[age_group], label = 'below 62') kmf.plot(ax = ax) kmf.fit(X[~age_group], event_observed = Y[~age_group], label = 'above 62') kmf.plot(ax = ax) plt.title(\"Kaplan Meier estimates by age group\") plt.xlabel(\"Month after heart attack\") plt.ylabel(\"Survival\") Kaplan Meier Curve Using Wallmotion Score As we can see that the difference between the age groups is less in the previous step, it is good to analyse our data using the wallmotion-score group.The Kaplan estimate for age group below 62 is higher for 24 months after the heart condition. After it, the survival rate is similar to the age group above 62. score_group = df['wallmotion-score'] < statistics.median(df['wallmotion-score']) ax = plt.subplot(111) kmf.fit(X[score_group], event_observed = Y[score_group], label = 'Low score') kmf.plot(ax = ax) kmf.fit(X[~score_group], event_observed = Y[~score_group], label = 'High score') kmf.plot(ax = ax) plt.title(\"Kaplan Meier estimates by wallmotion-score group\") plt.xlabel(\"Month after heart attack\") plt.ylabel(\"Survival\") Conclusion In this article, we have discussed the survival analysis using the Kaplan Meier Estimate. It also helps us to determine distributions given the Kaplan survival plots. Further, we researched on the survival rate of different age groups after following the heart treatment. Finally, it is advisable to look into survival analysis in detail.","excerpt":"The Kaplan–Meier estimator is an estimator used in survival analysis by using the lifetime data. In medical research, it is frequently used to gauge the part of patients living for a specific measure of time after treatment.","categories":["Deep Tech"],"tags":["data preprocessing","Python for Data Science","python machine learning"],"author_name":"Ankit Das","publish_date":"2020-11-09T14:00:07","publication_year":"2020","word_count":537,"keywords":["Go","NumPy","programming_languages:R","AI","data preprocessing","programming_languages:Go","RAG","Matplotlib","Seaborn","Python for Data Science","python machine learning","R","Pandas"],"extracted_tech_keywords":["AI","Pandas","NumPy","Matplotlib","Seaborn","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/quick-guide-to-survival-analysis-using-kaplan-meier-curve-with-python-code\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10170564,"title":"Microsoft’s Aurora AI Model Sets New Standard for Weather Forecasting","content":"Traditionally reliant on physics-based models and supercomputers, forecasting is now being accelerated and enhanced by machine learning systems that can process massive datasets and detect subtle atmospheric patterns. This shift provides faster, more accurate and more localised weather and environmental forecasting. Aurora from Microsoft is an AI foundational model that predicts the weather with precision and also forecasts a broad range of environmental events, from ocean waves to air pollution. As detailed in a recent paper published in Nature, Aurora represents a significant advancement in using AI to understand and anticipate Earth system phenomena. Aurora is a large-scale AI foundation model, trained initially on over one million hours of diverse atmospheric data, including satellite imagery, radar readings, weather station data and simulation outputs. “It’s not just about weather anymore,” said Megan Stanley, a senior researcher on the Aurora project. According to Microsoft’s blog post, Aurora beat traditional numerical models and prior AI systems in 91% of forecasting benchmarks for medium-range forecasts, up to 14 days, at a resolution of 0.25 degrees. It also outperformed major global forecasting centres in cyclone tracking, setting a new standard by correctly predicting Typhoon Doksuri’s landfall in the Philippines, four days ahead of the event, while official forecasts missed the mark. Leveraging high-performance GPUs, the model can produce forecasts in seconds, around 5,000 times faster than current supercomputer-based weather systems. In one test, it accurately predicted a massive sandstorm in Iraq 24 hours in advance, despite limited data, by using its foundational understanding of atmospheric patterns. It also demonstrated high accuracy in forecasting ocean wave heights and directions, essential for maritime safety and disaster preparedness. Moreover, MSN Weather is already integrating Aurora’s capabilities to provide users with more accurate hourly forecasts and expanded weather parameters. “There’s a huge opportunity here, especially for countries underserved by traditional forecasting tools. Aurora allows for high-resolution, localised predictions with much lower operational costs,” Stanley said. “If it truly is learning physics correctly, it can be adapted to different climate settings with confidence,” she explained. “Aurora is the first of its kind—but it won’t be the last.”","excerpt":"From sandstorms to cyclones, Aurora predicts the weather with precision and forecasts a broad range of environmental events","categories":["AI News"],"tags":["Aurora","Microsoft","Weather"],"author_name":"Merin Susan John","publish_date":"2025-05-23T09:56:01","publication_year":"2025","word_count":347,"keywords":["machine learning","TPU","programming_languages:R","AI","Aurora","RAG","Weather","GAN","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","RAG","TPU","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsofts-aurora-ai-model-sets-new-standard-for-weather-forecasting\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":38188,"title":"Good News For Developers, A New Machine Learning Tool To Recommend Code Snippets","content":"Your smart phone OS contains more than 10 million lines of code. A million lines of code takes 18000 pages to print which is equal to Tolstoy’s War and Peace put together, 14 times! Though the number of lines of code is not a direct measure of the quality of a developer, it indicates the quantity which has been generated over the years. There is always a simpler, shorter version of the code and also a longer more exhaustive version. What if there is a tool which uses machine learning algorithms to pick out the most suitable code and recommend it with just a click? There is one now. Aroma, a code-to-code search and recommendation tool which enables developers to get insights from large codebases. “With Aroma, engineers can easily find common coding patterns without the need to manually go through dozens of code snippets, saving time and energy in their day-to-day development workflow,” say the team behind Aroma at Facebook AI. Source: Facebook The code corpus is indexed as a sparse matrix initially and then a set of features are extracted from the parse tree. The feature vectors then act as index matrix, which is used for searching codes like shown below. A code snippet is fragmented as above and the result of the dot product of a sparse vector with the feature vectors is used as a threshold for recommendation. All possible code recommendations with similarities to the original snippet will be clustered. Since dot product alone cannot avoid the abstractions of the components in a code, pruning is used to rank the similarity. This cluster of potential candidates are skimmed through iteratively for extra statements useful for recommendations.  The remaining code after pruning is recommended eventually. Aroma is put to test by pitting it against top voted answers to 500 most popular questions asked on Stack Overflow. The Java code from each snippet is extracted by excluding comments. Then AROMA’s recommendations are checked for similarities in the lines used in top voted answers, which are picked randomly. The recommended code also performs additional operations like showing API methods called in the query code and suggests related statements that commonly appear in the query code. The results show that AROMA takes 1.6 seconds before it recommends a code. Aroma eclipses its counterparts with the following advantages: Aroma performs search on syntax trees. It can find instances that are syntactically similar to the query code and highlight the matching code. Aroma automatically clusters together similar search results to generate code recommendations. Creates real-time recommendations for very large codebases and does not require pattern mining ahead of time. Deployed across codebases in Hack, JavaScript, Python, and Java Aroma facilitates faster discovery of codes by sifting through large number of lines of code for recurring coding patterns. Now developers instead of worrying about missing a trivial syntax or defining a class for a task specific functionality, they can now proceed with their work at higher level with this semi-automatic tool. Know more about Aroma here","excerpt":"Your smart phone OS contains more than 10 million lines of code. A million lines of code takes 18000 pages to print which is equal to Tolstoy’s War and Peace put together, 14 times! Though the number of lines of code is not a direct measure of the quality of a developer, it indicates the […]","categories":["Deep Tech"],"tags":["Facebook AI"],"author_name":"Ram Sagar","publish_date":"2019-04-24T10:35:21","publication_year":"2019","word_count":504,"keywords":["Go","API","machine learning","Facebook AI","AI","ML","RAG","Python","JavaScript","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","Python","R","JavaScript","Go","Java","API"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/good-news-for-developers-a-new-machine-learning-tool-to-recommend-code-snippets\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":43200,"title":"Machine Learning Helps Discover New Polymers Which Can Be Used For 5G Connectivity","content":"Over the last few years, the chances of creating new conducting polymers with the help of machine learning have caught the attention of many researchers in the field of chemistry. Now, a team of researchers has discovered a new kind of polymer which contains high thermal conductivity and can be beneficial to the 5G mobile communication technologies. Researcher Ryo Yoshida said that many aspects remain to be explored, such as “training” computational systems to work with limited data by adding more suitable descriptors. He added, “ML for polymer or soft material design is a challenging but promising field as these materials have properties that differ from metals and ceramics, and are not yet fully predicted by the existing theories.” This research was conducted by Statistical Mathematics (ISM), Research Organization of Information and Systems, Tokyo Institute of Technology and Center for Materials Research by Information Integration, Research and Services Division of Materials Data and Integrated System (MaDIS). Methodology Used The goal behind this computational molecular design is to identify new molecules whose physicochemical properties meet the given arbitrary requirements. The researchers were successful in designing new polymers with high thermal conductivity with the help of a machine learning algorithm that is referred to as Bayesian molecular design. In the Bayesian molecular design process which generated a library of virtual chemical structures, the researchers specified a higher region of glass transition temperatures and melting temperatures as alternative design targets, for which sufficient data were given to obtain reliable prediction models. The model not only identifies new molecules but also helps to mitigate the issue of limited data. In addition to that, transfer learning which is an ML framework is also applied in order to obtain a thermal conductivity model with the given small data set. Dataset used The study was performed on a dataset of polymeric properties from PoLyInfo which is the largest database of polymers in the world at NIMS. The database contains a limited amount of data on the heat transfer properties of polymers. In order to predict the heat transfer properties from the given amount of data, machine learning models on proxy properties were pre-trained and then these pre-trained models extracted the common features which are relevant to the related tasks. The pre-trained model is merged with a specially designed ML algorithm for computational molecular design which is known as the iQSPR algorithm. Other Use Cases Last year, a group of researchers from the department of chemical engineering at Virginia Tech developed a temperature-independent computational model for a particular polymer which is sensitive to temperature. The simulation trajectories of this computational model were analysed with the help of a data-driven machine learning method. The computational model is known as the coarse-grained model which utilises a specific data-driven ML approach, known as non-metric multidimensional scaling method to analyse the molecular dynamic simulation trajectories of a coarse-grained model of a temperature-sensitive polymer. Outlook Traditionally, materials and polymers are created with a trial-and-error approach, but with ML, researchers can build and create new materials in a cost-efficient manner, in less time. In this modern era, databases and computation models are the keys to create resistive and efficient materials. The researchers at the Tokyo Institute of Technology are striving to create ML-driven high-throughput computational systems in order to design next-generation soft materials for applications going beyond the 5G era. The polymers with high thermal conductivity would be the key to heat management in the fifth-generation (5G) mobile communication technologies. (You can read the full paper here.)","excerpt":"Over the last few years, the chances of creating new conducting polymers with the help of machine learning have caught the attention of many researchers in the field of chemistry. Now, a team of researchers has discovered a new kind of polymer which contains high thermal conductivity and can be beneficial to the 5G mobile […]","categories":["AI Features"],"tags":["Machine Learning"],"author_name":"Ambika Choudhury","publish_date":"2019-07-28T10:00:27","publication_year":"2019","word_count":583,"keywords":["Go","machine learning","ELT","programming_languages:R","AI","data-driven","ML","Machine Learning","ViT","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","R","Go","ELT","GAN","ViT","data-driven","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/machine-learning-helps-discover-new-polymers-which-can-be-used-for-5g-connectivity\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10162920,"title":"Indian Startup Founders Push OpenAI to Offer Region-Specific AI Pricing","content":"As the Indian founders met with OpenAI CEO Sam Altman and their leadership team in Delhi, the sentiment was clear – tech leaders pushed for region-specific pricing for OpenAI’s models in India. We at AIM had pointed out earlier the implications of high-cost and premium pricing models for advanced tools in a country like India. India’s significance in the global AI ecosystem is visible, and Altman reinforced that by calling it its second-biggest market. India prides itself on having one of the best developer and startup ecosystems. Indian leaders, including policymakers, venture capitalists, developers, and founders, met for a closed-door roundtable meeting to discuss how OpenAI’s models can support businesses in India. Kunal Bahl, co-founder at Snapdeal, who was part of the event, took to X to share that OpenAI’s leadership team acknowledged its high pricing and the need for significant cuts for mass adoption, with possible updates ahead. In his post, Bahl says OpenAI acknowledged that foundational models reach 80-90% efficiency and require a robust application layer for full industry-specific use—crucial for startups in this space. Many startups are building on OpenAI’s models on the application layer. HealthifyMe’s Tushar Vashisht, who was also present at the event, said, “AI+human coaches, tutors, doctors—coming soon from India for India and the world.” In an AIM podcast earlier, OpenAI’s policy lead Pragya Misra cited Healthify and Be My Eyes as examples of the company’s impact on the Indian market and beyond. She highlighted how Indian startups are already developing products for a global audience. The roundtable saw key players from India’s startup scene, including Paytm’s Vijay Shekhar Sharma, Unacademy’s Gaurav Munjal, Fractal’s Srikanth Velamakanni, Ixigo’s Aloke Bajpai, and Aakrit Vaish, who advises the India AI Mission. IT minister Ashwini Vaishnav is optimistic about India’s youth on pushing innovation to the next level while keeping costs down. Speaking at the event, he compared it to the Chandrayaan mission, asking why the same ambition and efficiency couldn’t be brought to developing large language models (LLMs). Addressing costs, Altman pointed out that AI training costs will continue to rise exponentially, so do the returns in intelligence. That said, Altman noted that the company will continue to make solutions unique for India’s needs. Push for Competitive Pricing The meeting took place amid growing competition from Chinese AI lab DeepSeek, which claims to offer AI models comparable to OpenAI, Meta, and Google, at significantly lower costs and is open-source. The meeting focused on discussing Indian user preferences and API pricing. Touching upon the lower cost of DeepSeek APIs, speaking to Moneycontrol, Paytm founder Vijay Shekhar Sharma said, “Although Sam did not commit to anything, he said that options of open sourcing and reducing costs are both on the table.” It is interesting to note that Altman recently conceded that the future of AI will ultimately be open-source in an AMA session on Reddit. “I personally think we have been on the wrong side of history here and need to figure out a different open-source strategy,” he said. “Not everyone at OpenAI shares this view,” Altman said. At the event, Altman discussed deep research, a new capability in ChatGPT that independently conducts multi-step research on the Internet. “Deep research can perform a single-digit percentage of all economic, time-consuming tasks. It can make you twice as efficient,” he said. Funnily, in just 24 hours since the launch of OpenAI’s Deep Research, an open-source version of the tool was built on HuggingFace, scoring 55% on GAIA, one of the leading benchmarks for AI assistants. Just as startup founders questioned API pricing, India’s price-sensitive market suggests that a uniform pricing strategy for AI models may not be as effective from a consumer standpoint. AI Adoption in India While AI accessibility is improving, true adoption in India hinges on both usability and affordability. Simplified interfaces help, but without cost-effective pricing, AI may remain out of reach for many. OpenAI now offers ChatGPT Pro at $200\/month and is rumoured to introduce plans up to $2,000\/month due to high compute costs of advanced models. While economics often sees costs decrease over time. Many people responded to Altman on X, highlighting that in the current realm, $200 per month is comparable to salaries and average incomes in many economies outside the US – suggesting that AI subscription pricing cannot be the same globally. In India, for instance, the average monthly income is around ₹20,000. “The potential of advanced AI models with AGI capabilities in India lies in their ability to deeply integrate with the country’s diverse and localised contexts,” said Digital Empowerment Foundation’s Osama Manzar. He was speaking in the context of making AI more accessible and bridging the urban-rural divide. “For these technologies to meaningfully impact the daily lives of average Indians and create new opportunities, the approach must prioritise hyper-localised content generation,” he added.","excerpt":"India is a key player in AI, with a thriving developer and startup ecosystem.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Aditi Suresh","publish_date":"2025-02-06T12:40:09","publication_year":"2025","word_count":801,"keywords":["Go","ChatGPT","AI assistants","OpenAI","AI","Git","RAG","Ray","Aim","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","Ray","RAG","AI assistants","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/indian-startup-founders-push-openai-to-offer-region-specific-ai-pricing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":28121,"title":"How Do Machine Learning Algorithms Differ From Traditional Algorithms?","content":"Machine learning is an algorithm or model that learns patterns in data and then predicts similar patterns in new data. For example, if you want to classify children’s books, it would mean that instead of setting up precise rules for what constitutes a children’s book, developers can feed the computer hundreds of examples of children’s books. The computer finds the patterns in these books and uses that pattern to identify future books in that category. Essentially, ML is a subset of artificial intelligence that enables computers to learn without being explicitly programmed with predefined rules. It focuses on the development of computer programs that can teach themselves to grow and change when exposed to new data. This predictive ability, in addition to the computer’s ability to process massive amounts of data, enables ML to handle complex business situations with efficiency and accuracy. Traditionally, applications are programmed to make particular decisions, for example there may be a scenario based on predefined rules. These rules are based on human experience of the frequently-occurring scenarios. However, as the number of scenarios increases significantly, it would demand massive investment to define rules to accurately address all scenarios, and either efficiency or accuracy is sacrificed. How Does Machine Learning Differ From Traditional Algorithms A traditional algorithm takes some input and some logic in the form of code and drums up the output. As opposed to this, a Machine Learning Algorithm takes an input and an output and gives the some logic which can then be used to work with new input to give one an output. The logic generated is what makes it ML. ML Vs Classical Algorithms ML algorithms do not depend on rules defined by human experts. Instead, they process data in raw form — for example text, emails, documents, social media content, images, voice and video. An ML system is truly a learning system if it is not programmed to perform a task, but is programmed to learn to perform the task ML is also more prediction-oriented, whereas Statistical Modeling is generally interpretation-oriented. Not a hard and fast distinction especially as the disciplines converge, but in my experience most historical differences between the two schools of thought fall out from this distinction In classical algorithms, statisticians emphasis on p-value more and a solid but comprehensible model Most ML models are uninterpretable, and for these reasons they are usually unsuitable when the purpose is to understand relationships or even causality. The mostly work well where one only needs predictions. Traditional learning methodologies such as training a model-based on historic training data and evaluating the resulting model against incoming data is not feasible as the environment is in a constant change. As compared to the classical approach, traditional ML approaches as in most cases these approaches are too expensive within web scale environments and their results are too static to cope with dynamically changing service environments As opposed to classical approach, spending a lot of computational power on learning a very complex model of a highly dynamic network environment is not cost-effective Gradually, “statistical modelling” will move towards “statistical learning” and employ good parts about and creating tools for better interpreting the models in the process, Pekka Kohonen, assistant professor at the Karolinska Institutet pointed out One of the key differences is that classical approaches have a more rigorous mathematical approach while machine learning algorithms are more data-intensive In the last two decades, there has been a significant growth in algorithmic modeling applications, which has happened outside the traditional statistics community. Young computer scientists are relying on machine learning which is producing more reliable information. Unlike traditional methods, prediction, accuracy and simplicity are in conflict.","excerpt":"Machine learning is an algorithm or model that learns patterns in data and then predicts similar patterns in new data. For example, if you want to classify children’s books, it would mean that instead of setting up precise rules for what constitutes a children’s book, developers can feed the computer hundreds of examples of children’s […]","categories":["AI Features"],"tags":["Machine Learning Algorithms"],"author_name":"Richa Bhatia","publish_date":"2018-09-10T04:15:39","publication_year":"2018","word_count":611,"keywords":["Machine Learning Algorithms","Go","machine learning","artificial intelligence","TPU","AI","programming_languages:R","ML","programming_languages:Go","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","TPU","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-do-machine-learning-algorithms-differ-from-traditional-algorithms\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10095636,"title":"PM Modi Calls for Global Collaboration for AI in Education at G20 Meeting","content":"In a significant address delivered during the G20 education ministers’ meeting in Pune, Prime Minister Narendra Modi emphasised the vital role of AI in the field of education. He highlighted the immense possibilities offered by AI in learning and skills development, while acknowledging the challenges that accompany. Modi stressed the importance of striking the right balance in utilising AI for educational purposes. The G20 ministers convened to officially accept the outcome documents, marking the culmination of extensive deliberations conducted within the education working group track over the past several months. These outcome documents will serve as a roadmap for coordinated global actions, aiming to ensure inclusive and high-quality education for all learners. The education working group, which gathered in Chennai, Amritsar, Bhubaneshwar, and Pune, focused on formulating inclusive solutions and collective actions to address global educational challenges. Their priorities encompassed foundational literacy and numeracy, inclusive and qualitative tech-enabled learning, capacity building for future work requirements, as well as fostering research and innovation through collaboration and partnerships. During the meeting, a recorded message from Prime Minister Modi urged the G20 to play a pivotal role in leveraging technology to enhance global education. He emphasised the role of digital technology as an “equaliser” and a means to promote inclusivity, highlighting its potential to expand access to education and adapt to future needs. Additionally, Modi underscored the integration of technology with learning to enhance the e-learning experience. Read: India Backs Off on AI Regulation. But Why? He commended the Study Webs of Active-learning for Young Aspiring Minds (SWAYAM), an esteemed e-learning portal providing remote access to courses from renowned institutions. SWAYAM has achieved remarkable enrollment numbers, offering a wide range of 9,000 courses from the 9th grade to the postgraduate level. Furthermore, Prime Minister Modi proposed that G20 countries undertake international skill mapping to identify and address gaps in workforce capabilities. He stressed the necessity of continuous skilling, re-skilling, and up-skilling of the youth to equip them for the future and align their competencies with evolving work profiles and practices. India has already initiated skill mapping efforts, and Modi encouraged other G20 nations to follow suit, fostering a collective approach to workforce development.","excerpt":"Modi emphasised the role of digital technology as an “equaliser” and a means to promote inclusivity, highlighting its potential to expand access to education and adapt to future needs.","categories":["Deep Tech"],"tags":["AI in Education","AI Regulation"],"author_name":"Mohit Pandey","publish_date":"2023-06-22T17:26:48","publication_year":"2023","word_count":360,"keywords":["AI in Education","programming_languages:R","AI","innovation","Git","RAG","AI Regulation","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Git","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/pm-modi-calls-for-global-collaboration-for-ai-in-education-at-g20-meeting\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":17989,"title":"Pinky Sahu Explains How To Assure Business Outcomes Through Analytics","content":"Speaking at Cypher 2017, India’s most exciting analytics summit, Pinky Sahu, General Manager and Head of analytics, India, at Wipro spoke about how to bring assurance on the outcome of businesses using analytics. Today the market and technology are rapidly evolving. The environment is more volatile, less certain or rather disruptive than in the past and because of these disruptions, organisations cannot rely on methods and assumptions that they would have followed about a few years back for business decisions. Sahu gives an insight on how there is mounting pressure on the CXOs to optimise business outcomes for the best results. Often, they are burdened with faster decision-making on the strategy building. Sahu’s talk largely outlines the need for faster and better decisions that has propelled the adoption of business analytics for all businesses. The art behind using data: Adopting a consultative model is the only way forward. More than often, companies provide solutions to problems based on symptoms. However, the problem could be deep-rooted and may need a better diagnosis. Several and repeated workshops are the best way to understand the exact problems, Sahu explains. Starting small is the key but the real success is when the solution can be implemented for long-term solutions. As Sahu points out in her talk, the trick is to operationalise the analytics to modify and enhance customer information. Prioritising deliverables is also a key to maintain a long analytical product. It is understood that the need to change the traditional business models is higher than ever before. How to bridge the gap? Sahu recommends starting small with small projects and understanding the entire cycle of an analytics project. She insisted on getting feedback and improve the models over a period of time. The feedback determines whether a product is sustainable and will be used by businesses over and over again. Essentially, one must look at scaling up the project only when they see and experiences successes that can be stitched together to build a platform for a data lake implementation. Wipro’s DDP Wipro’s Data Discovery Platform (DDP), with its industry-specific apps, covers the entire spectrum from data to information to insights, empowers customers with relevant insights to make faster, better decisions. It enables businesses to gain valuable analytical insights, and further bridge the insights gap with value-added process simplification and business transformation services. Leveraging techniques such as visual sciences and storytelling with data, DDP drives insight-driven decision making and accelerates time to market, said Sahu. She added that the comprehensive Wipro DDP solution addresses challenges faced by global enterprises, from early adopters and established practitioners to those that are just starting off with analytics.","excerpt":"Speaking at Cypher 2017, India’s most exciting analytics summit, Pinky Sahu, General Manager and Head of analytics, India, at Wipro spoke about how to bring assurance on the outcome of businesses using analytics. Today the market and technology are rapidly evolving. The environment is more volatile, less certain or rather disruptive than in the past […]","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Wipro"],"author_name":"Priya Singh","publish_date":"2017-10-03T08:56:54","publication_year":"2017","word_count":441,"keywords":["Wipro","API","programming_languages:R","AI","RAG","analytics","disruption","GAN","R","data lake","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","analytics","RAG","R","API","data lake","GAN","disruption","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/pinky-sahu-explains-assure-business-outcomes-analytics\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10077285,"title":"First Home-Grown Semiconductor Chip Manufactured by Polymatech","content":"Polymatech, India’s first and foremost semiconductor chip manufacturer powered by Japanese technology, rolled out its Opto-semiconductors and memory modules this week. Currently, the homegrown manufacturer’s main manufacturing plant, in Kancheepuram, Tamil Nadu, manufactures 400,000 chips daily. The manufactured chips have already been released into the market. In a few months, Polymatech aims to attain its full capacity of manufacturing 1 million per day (300 million chips per annum). This comes after the firm’s announcement of its massive business expansion and investments, to the tune of US$ 1 bn, in semiconductor chip manufacturing in July. The firm has completely packaged Opto-semiconductors in both HTCC (High Temperature Co-fired Ceramic substrates) and COBs (Chip on Board). Opto-semiconductors are majorly used in medical, lighting, and food sanitization applications. The COBs are packed for applications with high-power lighting, such as stadium lighting and airport lighting. On the other hand, Opto-semiconductors packed in HTCC substrates are used in aircraft, metro trains, and traffic lights, etc. Thus, memory modules manufactured at Polymatech form an integral part of all major electronic systems. Speaking on the latest development, Polymatech’s founding president Eswara Rao Nandam said, “We are thrilled to announce the full-fledged rollout of production for our Opto-semiconductors and memory modules. Our Optos give more than 97% CRI (Color Rendering Index). By 2029, the global semiconductor industry market size is predicted to be US$ 1,340 bn and the Indian market will form a significant portion of this industry with a projected growth of US $64 bn by 2026. This, coupled with the ongoing worldwide chip shortage, makes for immense growth potential.” He added that the company aims to leverage this global opportunity to become one of the largest chip manufacturers in Asia by 2025. In addition to this, the company stands at its final stage in the trials for semiconductor chips production with general and medical applications.In 2018, Polymatech started manufacturing semiconductor chips and became India’s first semiconductor chip manufacturer. The company’s vision is in line with the government’s initiatives such as Make in India and Digital India.","excerpt":"Currently, Polymatech’s main manufacturing plant, located in Kancheepuram, Tamil Nadu, manufactures 400,000 chips every day","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-10-14T14:12:00","publication_year":"2022","word_count":339,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/first-home-grown-semiconductor-chip-manufactured-by-polymatech\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10073187,"title":"Is TikTok the New Polling Booth?","content":"Back in 2020, American politician Tim Ryan featured in a 15-second TikTok video casually lounging around in the White House with the song ‘Bored in the House’ playing in the background. The video clip was adored by millions. Having served as the US representative for Ohio’s 13th congressional district, Ryan’s move was seen as a surefire way to connect with Gen-Z. TikTok is being seen as an extremely potent campaign tool shaping political opinions, one video at a time. And while most candidates rely on this app, lawmakers have formed new rules for the young influencers who promote them. On Wednesday, the platform announced its plans to safeguard the upcoming US elections by reinforcing the ban on paid political ads. Governments around the world have shifted to using social media streaming platforms for their political campaigns. TikTok, being one of the favourites, said it would publish educational content and host briefings with influencers, citing the rule as a longstanding policy. It was implemented “so the rules of the road are abundantly clear regarding paid content around elections”. Catch ‘em young According to the United States Census Bureau, only 51.4% of the 18-24 age category voted during the 2020 elections. The million-dollar question is how do politicians tap into the young vote bank and really connect with the group? Undoubtedly, by tapping into TikTok. Wallaroo Media claims over 60% of TikTok’s monthly users are youngsters, which makes it a foolproof way for the politicos to reach them. A young candidate, Christina Haswood, used TikTok to become the youngest representative in the Kansas state legislature. “My entire life has been balancing this fine line of Western society and being an Indigenous woman, and I really wanted to highlight this on TikTok,” says Haswodo in a podcast. Role of campaign tools in India Banned in June 2020 over security concerns, TikTok is no longer accessible in India. However, platforms like Facebook and WhatsApp continue to rule the digital space. A research paper named ‘Influence of NaMo App on Twitter’ aims to study the role of the app in shaping the political discourse on Twitter. Source: (Influence of NaMo App on Twitter, September 2021) Prime Minister Narendra Modi launched his first official app called the ‘NaMo App’, as part of his prime ministerial election campaign. The paper indicates that whenever a post is shared using the NaMo App on Twitter, the tweet gets annotated with ‘with via MyNt’ or ‘via NaMo App’ tag at the end. Twitter played a significant role in engaging users and political parties during election campaigns in India. Meta-based platforms Facebook and Instagram on Tuesday announced their decision to restrict political advertisers from running new ads a week before the election. Last week, Twitter planned to revive past strategies for the elections, which included debunking false claims and inserting reliable information. However, the plan seemed inadequate to several civil and voting rights experts. The Indian I&B Ministry recently blocked seven Indian and one Pakistani YouTube channel for spreading misinformation. The total count of such bans on YouTube channels stands at 102 since December 2021. These channels reportedly had over 114 crore viewership, with 85 lakh subscriptions. To TikTok or not to TikTok? Everything seems to have changed since we last checked – right from the way we shop and eat to how they run political campaigns. Apps like TikTok that were established for leisure, now say, “We take our responsibility seriously.” However, everything has a downside. Recently, the US house of representatives, sent a two-page memo to House employees warning against using TikTok citing security concerns. An excerpt from the memo goes: “TikTok policy has stated that it automatically collects information about users’ devices, including location data based on your SIM card and IP addresses and GPS, your use of TikTok itself and all the content you create or upload, the data you send in messages on its app, battery state and even your keystroke patterns and rhythms, among other things.” Ironically, despite security warnings, Tim Ryan continued advertising his TikTok handle on his official Twitter account. The Congressman while on the one hand says “follow me on TikTok, also goes all tough on China for his upcoming campaign ad on Twitter that said “China is out-manufacturing us left and right, and it’s time we fight back.” Source: Twitter The security threat Senator Marco Rubio of Florida deleted his TikTok account after indulging in the platform for a short time and called on President Biden to block the platform entirely. In an email statement last year, Rubio wrote that TikTok “poses a serious threat to US national security”. He says, “US partners such as India have already come to this conclusion, of banning TikTok from their country last year. It is time we acted on this threat as well.” Since the app is now being used to discuss political and social movements, it also comes with potential risks. It is left to GenZ to decide if TikTok is more effective than other platforms. But the real question is whether the US is comfortable with an app owned by a Chinese company influencing its citizens.","excerpt":"TikTok is being seen as an extremely potent campaign tool shaping political opinions, one video at a time","categories":["IT Services"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-08-21T16:00:00","publication_year":"2022","word_count":854,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","Git","Aim","CLIP","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","API","CLIP","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/is-tiktok-the-new-polling-booth\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10043903,"title":"How Hackathons Are Changing The Way Data Scientists Are Hired","content":"Today, hackathons are one of the primary recruitment tools for tech companies. Organisations invest in new-age tools to hire employees by assessing their problem-solving approach and skills to manage time & people. In machine learning hackathons, the participants are given a problem statement and need to work with a dataset to create an accurate model to top the leaderboard. The gamifying experience makes the hiring process more interactive and less stressful for both candidates and recruiters. A recent report showed universities that leveraged technologies like hackathons in their hiring process achieved a 70 percent on-boarding rate. “For a better perspective, think of a way of solving a problem as quickly as possible. To crack the code, you will certainly need multiple minds to work and innovate together. That’s the simplest definition of a hackathon,” Nikhil Barshika, founder of Imarticus Learning. ML hackathons allow employers to take a closer look at how potential hires deal with real-world situations. Let’s take a deep dive into why hackathons are changing the way data scientists are hired. Team player: ML hackathon activities allow the recruiters to measure if the individuals are team players and can collaborate with cross-functional teams to finish the given task effectively. The recruiters can create a realistic working environment and observe the candidates  as they complete the workflow and build a product. Hackathons are a window into the candidate’s work ethic. Versatility: The recruiters get to judge the participant’s range of skills a participant has; for instance, if the software engineer can easily transition to a data science role and support end-to-end solutions. Business sense: The way participants go about the given task speaks about their business acumen. The candidate’s approach to the problem reflects their analysis of the solution’s value proposition and ROI. The recruiter understands if the candidate is thinking in terms of business value while building the model.Ability to deliver under pressure: Hackathons are conducted in a constricted amount of time, testing the candidate’s ability to handle deadlines.Problem-solving skills: Activities outside the candidate’s comfort zone prove if they can hustle and collaborate with people from other fields to extract insights. Quick turnarounds: Traditionally, the data scientist hiring process can last for weeks. Hackathons offer a ring side seat to the recruiters to watch the candidates up close and at work, thereby eliminating a few hiring rounds. Market pulse: Hackathons allow organisations to sample the global talent pool. MachineHack In a bid to spotlight virtual hackathon as a non-traditional channel of recruitment, MachineHack hosted The Great Indian Hiring Hackathon (2020) in collaboration with 12 prominent companies including Aditya Birla Group, Bridgei2i, Concentrix and Fractal. “Since its inception, MachineHack has aimed to empower data scientists to innovate. Data scientists, despite having tremendous talent and innovation to offer, are facing unprecedented challenges during this pandemic, and we at MachineHack want to tap into that pool of talent,” said Bhasker Gupta, CEO & Founder, AIM. MachineHack has an ongoing fortnight-long hiring hackathon, Mathco. Thon for data scientists and machine learning practitioners. TheMathCompany will interview the candidates who make it to the top leaderboard positions. The participants also stand to win a cash prize. Organisations can also leverage hackathons for training and upskilling employees, thus preparing them for senior and more relevant roles within the company. The hackathon approach helps companies achieve two goals: to promote the work culture among existing employees and build a substantial brand recall value. For example, Karan Juneja, a regular participant in MachineHack hackathons and its grandmaster, has said hackathons have helped him pick up new data science skills. While employers get the opportunity to shortlist the best talents for their organisation, candidates get a hang of the organisation’s work culture. The hackathons offer the ideal setting for both candidates and recruiter to understand if they are the right fit for each other.","excerpt":"The gamifying experience makes the hiring process more interactive and less stressful for both candidates and recruiters.","categories":["AI Highlights"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-07-20T11:00:00","publication_year":"2021","word_count":636,"keywords":["data science","Go","machine learning","AI","ML","RAG","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Aim","RAG","R","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-hackathons-are-changing-the-way-data-scientists-are-hired\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10053182,"title":"Council Post: Notes on the Nuances of Leadership","content":"Once upon a time, my dream job was selling soap. And toothpaste. And oil. And a variety of other household goods which on their own seem trivial. But how one got good at it involved a journey through the towns and villages of India, visiting and sipping chai with owners of kirana stores, speaking to customers who straddled socio-economic strata, and learning the feet-on-the-street approach to racking up sales. Someone else ended up living that dream. And I sat in a plush office with a view of the ocean, cold calling John Deere dealers in Dorothy’s Kansas – some would be flummoxed to hear an accented voice from Mumbai (then Bombay); others, intrigued, would chat along and provide a wealth of information as did the farmers in Punjab and the auto accessory manufacturers in Madras. In the end, the trick to selling anything is understanding people. From an analyst working numbers, an analytics industry grew around me, and while we now strive for excellence in the science of data, the end goal is still to influence people. To buy, transact, fund, contribute, use, enable, take action. And internally, to align, problem-solve, model, story-tell, belong. With the volume and velocity of data generated, moving up the ladder makes it hard to live in the details. And the sophistication of tools and domain depth requires mastery of what we manage. A career choice that faces many of us – and we are lucky to have reached that level of maturity – is whether to be a broad leader or the uber SME. My journey, with personal circumstances and an inclination to experience variety – or avoid what seemed like boredom – took me across industries and functions. Hence, the former. It required me to understand the tractor market, sell fuel oil, develop pricing strategies for brick and mortar grocery stores, extend the UK built analytics models to Asian markets with their unique products and competitors, build a data dictionary of contact centre acronyms, figure out how to produce tech support YouTube videos in Portuguese, lay a foundation for social media command centres and soon after for performance management processes in a start-up, finally stepping gingerly into banking and finding my current niche spanning analytics, social impact, communications, campus recruitment and COVID-19 crisis management. With such a vast spectrum, perhaps the key pillars that my ‘leadership’ is founded on are keeping the customer in mind as one looks at data, keeping the people in mind as one leads teams, and adhering to and continually raising the bar for the quality of work one is associated with and wants to be known for. Some of the mantras I absorbed over the years: Begin with the customer. Read their actions and choices in the data. Personify them and imagine the individual, the families, the lifestyles. Agatha Christie taught us to look for a motive. Our profession doesn’t deal with grim mysteries, but the ‘why’ drives everything.Figure out answers to – What business are you in, how it makes money, and the part you play in it. When you start your career staring at code, correcting syntax and swimming in numbers, sometimes this does not seem important. But it is. Because when you present to the CEO and make the data talk, the CEO is wondering how to make the customer walk; into a store, an aisle, a branch, a purchase process.Feed your Strengths. Manage your Weaknesses. There are some of us who sit in a performance review and let the accolades slip by, waiting to hear what we did wrong or what we can improve. And then we work on the deficiencies often to hear the same the next time around.It is rare for a weakness to become a strength. Moving the needle from negative to zero will not make you a star. But honing your strengths and exploiting them just might. So, take in the praise and know that this is what you must bank on to succeed.Master your stuff but keep it simple. Being a data scientist demands a lot. You must be a poet and a quant, an artist and a geek. Loading up on certifications is not critical; knowing your models is. But most important is the ability to discern whether you need a hammer or a drill. Everything cannot be simplified, but what is the right level of complexity that a problem requires? Quoting Einstein, “Everything should be made as simple as possible, but no simpler.”Make people feel ‘awesome’. This was a Chris Arnold mantra. It is perhaps hard for others to emulate him, but the essence is to hear people, make them feel that they matter, that someone is enabling the paths they want to traverse. Presence, communication and the connection – all need to come together to make this possible.Be a mini-Ganesha – remove the obstacles. If you’re not the uber SME, the people who work for you often know more about what works or doesn’t and what they need in order to be more effective and supported. Get those smart people on your team and clear the weeds for them. If you can be charismatic and omniscient, more power to you. Not all have that luxury, though.Rethink the regret. Regret and resentment are not unnatural sentiments, and everyone has some incidents they can recall when these were triggered. Sometimes, the right amount of resentment can be very powerful – channelled into the right action; it could lead to phenomenal outcomes that may not have transpired had you stayed content in your largely comfortable zone. At times, people even resent another person leaving the organisation for other opportunities, not realising that one’s network and connections can be lifelong. Move on. And think of what you can do differently. Most of these don’t sound like tips for data scientists. Perhaps because as one takes on the mantle of leadership, the core principles are universal. I do believe, though, that exposure to business in ways that force you to think of people – customers, employees, teams – and what drives them is critical to leveraging the potential of data and bringing your strengths and leadership abilities to the forefront. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"Once upon a time, my dream job was selling soap. And toothpaste. And oil. And a variety of other household goods which on their own seem trivial. But how one got good at it involved a journey through the towns and villages of India, visiting and sipping chai with owners of kirana stores, speaking to […]","categories":["AI Features"],"tags":[],"author_name":"Nidhi Pratapneni","publish_date":"2021-11-11T10:00:00","publication_year":"2021","word_count":1072,"keywords":["data science","Go","programming_languages:R","AI","RAG","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-notes-on-the-nuances-of-leadership\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10125674,"title":"Mike Knoop Discusses Rapid AI Growth and Slowed ARC Challenge Progress","content":"In a recent interview, Mike Knoop, co-founder of Zapier said performance on most AI evals is accelerating to human-level, but improvement on the ARC Challenge is decelerating. He further announced the ARC Prize competition designed to advance progress toward artificial general intelligence (AGI). It is hosted by Knoop, and François Chollet, the creator of ARC-AGI and Keras. The competition aims to solve the ARC-AGI benchmark, which is a formal measure of AGI that tests the ability to efficiently acquire new skills and solve open-ended problems “The ARC Prize, which is one of the most unique benchmarks in AI, measures a machine’s ability to truly learn things intelligently versus just parroting patterns in the training data,” he said, adding that we need radically different approaches and benchmarks to achieve true general intelligence. “The mission of Zapier and sort of the mission and purpose of AI intersect in an interesting way, which is you know the promise of AI is software that’s just going to do more work for you,” Knoop explained. He highlighted that Zapier utilises AI to automate repetitive tasks, such as data entry and formatting, enabling users to save time and concentrate on more important work. Founded in Columbia, Missouri, by Wade Foster, Bryan Helmig, and  Knoop in 2011, Zapier  was accepted into the Y Combinator startup seed accelerator the following year and temporarily relocated to Mountain View, California. Zapier is a web automation tool described as the “glue” that connects hundreds of other web apps and services, allowing them to work together seamlessly. It also offers integrations with over 5,000 different apps, this approach enhances the ability to connect with most of the tools one can use for automating their current workflow processes. Powered by machine learning algorithms, its AI capabilities can learn from data and improve over time. Its AI features include natural language processing, which allows users to interact with the platform using everyday language, and computer vision. At this point, Zapier might be the biggest automated AI platform, given that many researchers, entrepreneurs, and builders are trying to develop agentic AI systems where the AI operates without any humans in the loop. They are also approaching a milestone of 10 million AI tasks per month, and the current run rate suggests they will soon achieve this volume consistently. Check all of the apps that Zapier integrates with here.","excerpt":"Zapier might be the biggest automated AI platform.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Tarunya S","publish_date":"2024-07-03T15:22:12","publication_year":"2024","word_count":393,"keywords":["Go","API","agentic AI","machine learning","Keras","AI","ML","computer vision","Aim","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","agentic AI","Aim","Keras","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mike-knoop-discusses-rapid-ai-growth-and-slowed-arc-challenge-progress\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":34749,"title":"How Google, Amazon &#038; Microsoft Are Making ML Accessible For SMEs","content":"Image for representative purposes only Organisations are now realising the benefits of data, thanks to the democratisation of machine learning which has put powerful tools in the hands of SMEs and large enterprises alike. What has changed the game for companies (small and large alike) is the access to algorithms and labelled data coupled with massive computing resources that helps teams train and deploy models on a large scale. Today, machine learning is being provided as a service by multiple vendors, who provide compute and pre-trained models. In this article, we will look at how the Machine Learning as-a-service market is opening up access for small and medium enterprises to begin using artificial intelligence and scale according to their uses. Big Tech Giants Lower The Barrier Of Entry for SMEs With MLaaS Over the years, the existence of MLaaS market serves a bigger purpose for the scaling of small and medium enterprises. This means that organisations need not build an in-house internal machine learning team, or take the financial risk of establishing one only to find out that data cannot make the process more efficient. While bigger companies can afford to take the risk that comes with building a sizable ML team, it is not possible for SMEs to do the same. At the stage of a startup, a bad investment towards the direction of the company can usually result in a failure state. This is due to the fact that hiring requirements for ML are extremely specialized. Moreover, expensive infrastructure is also required for the high amount of compute power required to train algorithms. This requires expensive computing machines, and infrastructural concerns like data pre-processing, model training and evaluation. The general purpose nature of MLaaS allows organisations that have any need for ML to essentially plug and play, as ML suites today have APIs that can call different types of ML models for data analytics. Companies such as Google have offered pre-trained machine learning models via APIs that perform specific tasks. The user-friendly nature of these services also greatly reduce the barrier of entry for companies willing to use data analytics to superpower their processes. Finding Solutions To Problems With Machine Learning Companies today lack the infrastructure to store a massive amount of data. MLaaS offerings not only allow companies to look deeper into their data, but it also enables a culture of data collection. Storage and infrastructure needs are taken care of by the cloud storage products MLaaS vendors offer. ML today is usually offered as a service and is incorporated into a cloud computing service. This includes multiple ‘cognitive’ algorithms such as facial recognition, Natural Language Processing and speech recognition, along with DNNs and CNNs along with data visualizations. While the market is not only developing at a fast pace, it is also an open one, as consumers have multiple choices when it comes to deciding the best MLaaS for their solution. Due to the fact that data insights will provide an accurate insight into multiple processes, it will provide solutions to the company’s pain points when paired with a data-driven approach. Laying The Foundation For An ML-Driven Future The low barrier of entry and high effectiveness of solutions from vendors like Amazon, Google, Microsoft and Oracle leads to faster adoption of AI and takes the heavy lifting out of setting up their own infrastructure and building an in-house team. Existing workers can be upskilled and reassigned to work on cloud solutions without sacrificing economical costs and while adding value to workers. Moreover, companies are allowed free cloud credit and trial for a specific period.  If organisations find that these work for them, then they can continue using these solutions at a cheap rate, incentivising the use of AI.","excerpt":"Organisations are now realising the benefits of data, thanks to the democratisation of machine learning which has put powerful tools in the hands of SMEs and large enterprises alike. What has changed the game for companies (small and large alike) is the access to algorithms and labelled data coupled with massive computing resources that helps […]","categories":["Global Tech"],"tags":["Amazon","Azure","Google","Microsoft","SMEs"],"author_name":"Anirudh VK","publish_date":"2019-02-11T11:34:02","publication_year":"2019","word_count":622,"keywords":["Go","API","artificial intelligence","machine learning","SMEs","AI","cloud computing","ML","R","Amazon","RAG","analytics","Google","Azure","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","cloud computing","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-google-amazon-microsoft-are-making-ml-accessible-for-smes\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121917,"title":"Microsoft Launches Telegram Bot Powered by Copilot","content":"Microsoft has introduced a new Copilot Telegram bot, allowing users to search, chat, and ask questions to an AI assistant for various purposes. Currently, in its beta phase, Copilot for Telegram offers its services for free to users on mobile or desktop, exclusively supporting text-based interactions. How to Use Copilot for Telegram? Getting started with Copilot for Telegram is simple. Users can access it through the Telegram desktop & mobile app, or via the Telegram web interface. Alternatively, they can search for Copilot by typing @CopilotOfficialBot in Telegram’s search bar. Upon discovery, the bot will prompt users to provide their Telegram phone number to connect. Microsoft has highlighted the Copilot Chatbot’s ability to deliver comprehensive responses, ranging from quick facts to in-depth research powered by Bing Search. It can further assist with travel plans, offer fitness guidance, provide entertainment updates, and more. Copilot’s content generation is based on language patterns extracted from the internet. As a result, its responses may occasionally resemble existing online content, or it may generate similar responses for users with similar prompts. Microsoft Copilot In September 2023, Microsoft launched Copilot, advertising it as your everyday AI companion. The company has highlighted that Copilot leverages web context, your work data, and real-time PC activities to deliver personalised support, all while prioritising users’ privacy and security. The tool has now been integrated into several Microsoft offerings, including Windows 11, and Microsoft 365, as well as the web ecosystem with Edge and Bing, offering a simple and seamless experience. At Microsoft Build 2024, significant updates to Microsoft Copilot were unveiled, signalling a strategic shift towards enhancing employee self-service and productivity through advanced AI capabilities. The Copilot Studio introduces independent agent capabilities, allowing Copilots to act autonomously based on events, expanding their utility beyond just conversations. Further, custom Copilots enable users to tailor AI assistants to specific roles and industries, offering enhanced information retrieval and contextual insights from SharePoint content. Complementing these advancements, the introduction of Copilot+ PCs promises unparalleled computing power and AI integration, aiming to revolutionise user experiences and productivity on Windows platforms, emphasising Microsoft’s commitment to innovation in conversational AI and intelligent computing.","excerpt":"Currently, in its beta phase, Copilot for Telegram offers its services for free to users on mobile or desktop.","categories":["AI News"],"tags":["Microsoft"],"author_name":"Vidyashree Srinivas","publish_date":"2024-05-29T12:06:34","publication_year":"2024","word_count":356,"keywords":["AI assistants","programming_languages:R","AI","innovation","ML","RAG","Aim","ViT","copilots","R","Microsoft"],"extracted_tech_keywords":["AI","ML","Aim","RAG","AI assistants","copilots","R","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-launches-telegram-bot-powered-by-copilot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":62532,"title":"Research: Impact Of Work From Home On Analytics Functions","content":"Working from home as a business practice requires a lot more than connectivity and overall productivity. For Work from Home (WFH) to be effective, an organization must attempt to extend all the facilities available in the office to the employee’s home. The current lockdown, due to the COVID-19 pandemic, has imposed work from home across those sectors where remote working is possible. The broad IT and Data Science domains are 2 such functions where working from home is possible under certain conditions, including the type of projects, type of functions, access to tools, employee engagement, and overall connectivity and collaboration with the rest of the team\/organization. KEY FINDINGS AIMResearch recently conducted a survey of Analytics personnel across levels, enterprises, sectors, and geographies. This global survey reveals interesting findings on the challenges encountered by the analytics personnel while working from home during this global lockdown and the opportunities and areas of focus for new projects. The companies covered in the survey include leading Domestic BFSIs, Infrastructure and Energy companies, Retail firms, Global Beverage brands, Technology enterprises, Global Pharmaceutical majors, Ride-sharing aggregators, Food Delivery firms, and Domestic companies. 1. Level of Impact of WFH on the Entire Analytics Team This question of the survey focused on the impact of working from home on the analytics team. There was No Impact of WFH for little more than half the respondents at 52.2%. 34.8% of the respondents revealed a Negative Impact in the range of 25-50% because of WFH on the team. The reason for this is explored in the responses to the next question. However, 8.7% of the respondents disclosed that WFH had a positive impact on the analytics team, as the productivity has improved through WFH. This is attributed to almost no time spent on commuting and lesser stress for some people in the WFH environment. Nonetheless, 4.3% of the respondents revealed a Negative Impact of 50-75% because of WFH on the analytics team. This is also explored in the next question. 2. Reasons for the Negative Impact of WFH on the Analytics Team Of those respondents that identified WFH had a negative impact on the analytics team, about 47.8% identified Remote Team Collaboration as the main reason for the negative impact. About 30.4% of the respondents revealed Remote Access to the Data as the reason for the negative impact – this is due to the lack of or difficulty in remotely accessing several data points. Surprisingly, about 21.7% of the survey audience has identified that although they have access to Data and Technology, the analytics team is unable to adjust to the WFH environment. This is usually true for personnel or teams that have higher years of experience or need to work for multiple teams within a local geography\/office to understand the data and insights. There were no respondents that expressed a negative impact because of the Remote Access to Tools or Lack of Tools and Connectivity. 3. Individual Ability to Work from Home All the respondents of the survey were Individually able to Work from Home during the ensuing lockdown. This is not surprising as all companies now provide ready access to laptops, VPN, technologies, tools, and platforms to enable their analytics employees to work remotely or from home. 4. Access to Tools and Platforms – Local or Via Cloud Majority of the respondents – 52.2% – revealed that the tools and platforms were available on their laptops or the local systems for access. While 47.8% of the respondents revealed that they accessed the tools and platforms via the Cloud, revealing how high the adoption of the Cloud is in enabling the access of tools and platforms. 5. Type of Collaboration Tools: Enabled for WFH To enable the analytics teams and personnel to WFH, organizations provide a variety of conferencing and collabora- tion tools to remotely connect and network. 54.5% of the employees utilized MS Teams to remotely connect and network with other members of their analytics teams. 13.6% of the employees utilized the new conferencing tool Zoom – despite general concerns of privacy and security from security experts. The other Microsoft tool for connecting, Skype, was utilized by 9.1% of the respondents, while Webex too was utilized by 9.1% of the respondents. A combination of Other Tools was also utilized by 9.1% of the respondents. Lastly, 4.5% of the respondents used the workplace productivity and networking tool, Slack, to collaborate and network across the analytics team. 6. Frequency of Connecting with the Team while WFH Using the collaboration tools, 60% of the respondents connected with their respective teams twice or more than twice a day. 25% of the respondents connected once every day with their respective teams. 15% of the respondents connected more than once weekly (but not daily) with their teams. There were no respondents that connected just once a week or did not connect at all with their teams. 7. Type of Projects the Analytics Teams are Working on while WFH The lockdown has had a financial and economic impact on companies. To measure this impact, many of the companies have deployed their analytics teams, which are working remotely, to work on new projects that analyze the impact of the lockdown on their compa- nies, industries, and customers. 45.5% of the respondents are working on such new projects that gauge the economic or financial impact of the lockdown. 4.5% of the respondents are working on new projects that gauge the social impact of the lockdown. Exactly half or 50% of the respondents are working on regular projects while working from home. There are no respondents that are not working on any project. 8. Duration of the New Projects Of the respondents working on new projects, 50% of the respondents are working on projects with less than 1 month in duration. 27.8% of the respon- dents are working on projects that were expected to last more than 3 months but less than 6 months. About 22.2% of the respondents are working medium duration projects that are expected to last more than a month but less than 3 months. Organizations do not foresee the impact or the severity of the impact of the recession to extend beyond 6 months. Hence no teams or personnel are working on such projects. 9. Revenues of New Projects Of those respondents who are working on new projects, 31.6% of the respon- dents claim that the revenues of new projects would be in-line with the average revenues. 26.3% of the respon- dents claim that new projects will bring in revenue 5-25% higher than average revenues. 21.1% of the respondents claim that the new projects would bring in revenue 5-25% lower than the average revenues. 10.5% of the respondents claim that new projects will bring in revenues 25-50% lower than average. 10.5% of the respondents respond that the new projects would bring in revenues 25-50% higher than average. 10. Business Areas of the New Analytics Projects Of those respondents who are working on new analytics projects, 10.5% responded that they are working on Customer Engagement related analytics – the highest proportion of the new analytics projects. 9.5% are working in Digital Marketing analytics. Similarly, 9.5% responded that they are working on Customer Relationship Management related projects. These results highlight the focus of organizations and the analytics to work on new projects focused on customer engagement, customer relationships, digital marketing, and consumer demand – these business to consumer segments are critical to analyze during and after a lockdown, downturn, or recession – consumer segments typically rebound the fastest after a downturn and are, hence, the most critical to analyze during the downturn. Similarly, 6.3% of respondents are working on new projects related to the area of Consumer Products and Customer Market & Segmentation – highlighting the theme of customer-focused new analytics projects. 11. Industry \/ Sectors of the New Projects Of those respondents who are working on new analytics projects, 25% are focused on new projects related to the Retail sector – the highest proportion in terms of sector of the new analytics projects. This is followed by 21.4% of the respondents working on Domestic BFSI projects. This again highlights the focus of new projects across the consumer-focused sectors. 12. Scope or Type of Analytics across the Projects The responses on the scope or type of analytics carried out across analytics projects reveal the highest proportion of respondents – 29% – are working on projects related to Prescriptive analytics. This is followed by 25% of respondents working on Diagnostic analytics and 24% on Predictive analytics. 22% of the respondents are working on Descriptive analytics related projects. 13. Time Spent While Not Working on Projects While there are analytics professionals spending time on projects, there is many personnel who at some point in time while working from home may not be working on a project or assignment. 76.9% of these analytics professionals are devoting their time on Training through Self Learning while 23.1 % are spending their time on Training through company driven learning programmes. Conclusion: The findings from the survey reveal that the entire analytics community is able to work from home. Moreover, while analytics personnel are working from home, they have access to the required tools and platforms to execute their tasks and responsibilities. Moreover, most of the respondents have either experienced no impact on their work or have experienced a positive impact on their work because of higher productivity. While 50% of the respondents are working on their existing projects – about 45% of the respondents are working on projects to analyze the financial or economic impact of the lockdown – and most of the projects are related to the consumer-facing sectors. Download the complete Report","excerpt":"Working from home as a business practice requires a lot more than connectivity and overall productivity. For Work from Home (WFH) to be effective, an organization must attempt to extend all the facilities available in the office to the employee’s home. The current lockdown, due to the COVID-19 pandemic, has imposed work from home across […]","categories":["AI Features"],"tags":["covid19 data","retail bi prescriptive","Work from Home"],"author_name":"AIM Media House","publish_date":"2020-04-24T14:21:00","publication_year":"2020","word_count":1612,"keywords":["data science","AI","R","retail bi prescriptive","Work from Home","Git","covid19 data","RAG","Aim","ViT","analytics","GAN","predictive analytics"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","predictive analytics","R","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/research-impact-of-work-from-home-on-analytics-functions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10000943,"title":"How The Arrival Of Quantum Internet Will Change The Face Of World Wide Web","content":"According to experts, quantum internet might not replace the classical internet but will enhance it. And after 5G, the next wave of technological advancement is going to be quantum internet. Quantum Vs Classical Internet In this article, we list down the differences between quantum internet and classical internet. Signals: The classical internet today uses radio waves to send out signals. This wireless adapter of a computer translates data into a radio signal and transmits it using an antenna, which the wireless router receives. Then, the router sends the information to the internet via a wired ethernet connection. Ordinary internet uses radio frequencies to connect computers through a global web in which electronic signals are sent back and forth. Quantum internet, on the other hand, uses quantum signals to send out information. In this case, the signals will be sent to and fro through a quantum network via quantum entanglement particles. Security: The data transmitted will have quantum encryption, and because of this, we would be able to communicate data which will be unhackable over the quantum network. Also called as quantum cryptography uses quantum key distribution (QKD), which will have the data encrypted using the property of entanglement of quantum mechanics. If any disturbance is sensed in such kind of a message transfer, it will cause the message to be automatically destroyed, with both the sender and the receiver being notified about it. This means qubits cannot be copied or amplified. Such a level of security is not provided by the internet that we have today. Applications: The internet today can be thought of primarily being used for the purpose of communications and browsing. Quantum internet would provide advantages in the field of medicine, clock synchronization, extending the baseline of telescopes, secure identification, achieving efficient agreement on distributed data, exponential savings in communication, quantum sensor networks, apart from just communication.  “Quantum voters,” says physicist Nicole Yunger Halpern at the Harvard-Smithsonian Center for Astrophysics in Cambridge, “…Could use strategic-voting schemes that classical voters can’t implement”. Quantum techniques might help large groups to coordinate and reach a consensus, for instance, to validate electronic currencies such as Bitcoin. Way Forward China has already launched the first ever quantum satellite and has been sending signals to it as well. Last year, in December, according to an experiment by a team of scientists, several carefully managed photons were exchanged in pulses in infrared light, carried between Russian GLONASS satellites and the Space Geodesy Centre on the ground run by the Italian Space Agency. These signals passed through several kilometers of space without any loss of data, indicating that a global network of photons is indeed a possibility. Theoretical physicist Stephanie Wehne who works in the domain of quantum internet said, “In the quantum-computing domain, it’s much more all or nothing,”","excerpt":"According to experts, quantum internet might not replace the classical internet but will enhance it. And after 5G, the next wave of technological advancement is going to be quantum internet. Quantum Vs Classical Internet In this article, we list down the differences between quantum internet and classical internet.   Signals: The classical internet today uses […]","categories":["AI Features"],"tags":["computer","cryptography","encryption","entanglement","Interviews and Discussions","Network","speed"],"author_name":"Disha Misal","publish_date":"2019-01-17T13:23:02","publication_year":"2019","word_count":465,"keywords":["entanglement","Go","programming_languages:R","AI","programming_languages:Go","speed","cryptography","Network","computer","encryption","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-the-arrival-of-quantum-internet-will-change-the-face-of-world-wide-web\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001389,"title":"Now IBM And Facebook Are Eyeing Blockchain For Financial Inclusivity In Developing Markets","content":"After the promise of cryptocurrencies taking over the current banking system was quashed, its underpinning technology known as blockchain can be the cause of a renaissance in the financial sector. This is not speculative in nature, as many offerings have been in the past, but based on an actual product, use case and customers. Big companies such as Facebook and IBM have begun investing in blockchain technology, which has already begun to show a lot of promise in revolutionising the financial sector. Many banks in India have also begun to adopt the technology for smoother cross-border transactions and transparency in their operations. The original blockchain was that on Bitcoin, which was used to facilitate the creation of a permissionless and trustless value network. While the concept of blockchain grew to be accommodated under the umbrella of distributed ledger technology, the main promise has always been to bring financial services to the unbanked and enable true financial inclusivity. Let us take a look into what these offerings are doing for the state of the financial and logistical markets. IBM’s Giant Footsteps Into Blockchain IBM began moving to a service-based company in the late ’00s and early ’10s, with its position cemented as one of the world’s biggest companies providing solutions today. Capitalising on their reach across the world and a long list of clients, IBM has begun its blockchain efforts. It is already a part of the Linux Foundation’s Hyperledger product for shepherding blockchain developments, and itself has made many forays into the real-world applications of the blockchain. To this tune, IBM has 3 prominent blockchain products. The first among them is known as IBM Food Trust, which uses blockchain to enable end to end tracking of produce. The use of the blockchain will allow for a shared history for all the participants, thus enhancing visibility and accountability. This ensures a higher amount of trust in the food supply chain, along with ensuring that the consumer knows exactly where their produce is coming from. Their second product is one known as TradeLens, and is set to create a blockchain-enabled environment to set ground truth standards for supply chains. The project aims to do what standardisation of shipping containers did for the shipping industry as a whole. To achieve this, IBM has made it an open and neutral platform that has every step of the chain validated in real time by the participant doing the work. However, the biggest bet for IBM’s blockchain dreams comes in the form of IBM Blockchain World Wire. This is a system aimed at replacing the currently outdated system for cross-border transactions known as SWIFT. SWIFT has existed for more than 60 years and creates pain points for customers wishing to transfer money across borders as it is unreliable and slow. IBM has leveraged blockchain to create a system that clears and settles cross-border payments in almost real-time, as opposed to SWIFT which can take anywhere from 3-5 business days. Cupertino’s Blue Giant Looks To Provide Value Adds For Consumers Facebook is not only looking into blockchain technology but also aiming to use it to create a cryptocurrency. This is to allow users to engage in cross-border transactions with their friends across the world at lowered transaction fees compared to traditional cross-border transactions. When looking at their reach over 3 of their products, Facebook has almost 3 billion individuals logging in to use them across the world, among which 300 million are in India. Even though introducing a cryptocurrency would mean that they will be required to navigate through multiple regulatory loopholes. However, Facebook, with its multiple run-ins with regulators, has a better chance of getting the product onto the open market. The giant is also looking to integrate WhatsApp, Messenger and Instagram messaging platforms into one experience. This will unify all of the user bases of the messaging platforms into one, creating a monopoly over social networks. In this environment, using a cryptocurrency means that Facebook can settle transactions across the world for all of its users. The cryptocurrency is set to be a stablecoin pegged to the value of the US dollar to ensure that volatility risk doesn’t take place when the users are utilising the currency to send the value. In Conclusion Blockchain technology as a whole can be expected to be seen in many more applications similar to this. The distributed ledger can be a prominent place to expand into the financial sector due to its ease of implementation for a company with resources such as IBM and Facebook. It also provides a unique value-add for the service and will leverage the existing network effect of the social media platforms. Blockchains can also be used for multiple other purposes as well, as seen by IBM’s implementation of a decentralised identity platform and supply chain tracking.","excerpt":"After the promise of cryptocurrencies taking over the current banking system was quashed, its underpinning technology known as blockchain can be the cause of a renaissance in the financial sector. This is not speculative in nature, as many offerings have been in the past, but based on an actual product, use case and customers. Big […]","categories":["AI News"],"tags":[],"author_name":"Anirudh VK","publish_date":"2019-03-04T21:27:16","publication_year":"2019","word_count":802,"keywords":["API","programming_languages:R","AI","RAG","Ray","Aim","ViT","Rust","GAN","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Rust","API","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/now-ibm-and-facebook-are-eyeing-blockchain-for-financial-inclusivity-in-developing-markets\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10065067,"title":"How this Gurugram based startup leverages AI to power logistics &#038; supply chain solutions","content":"Haryana-based Pickrr Technologies provides end-to-end logistic solutions to D2C and B2B sellers and serves over 29,000 pin codes across India. The SaaS platform’s major offerings include Pickrr Plus, Pickrr Connect and Pickrr Predictt. Analytics India Magazine got in touch with Gaurav Mangla, CEO and co-founder of Pickrr Technologies, to understand the company’s tech stack. AIM: What problem does Pickrr solve? Gaurav Mangla: Pickrr is a SaaS-based logistics start-up that uses big data, AI and ML-powered solutions to help D2C brands grow their business. Pickrr addresses the challenges sellers in the D2C ecosystem face with tech-driven value-added services and smart fulfilment facilities. AIM: How does Pickrr use technology to address challenges? Gaurav Mangla: Amid the increasing competition, the logistics service providers must offer their services in line with the customers’ expectations. While there are plenty of processes involved in bringing perfection and accuracy into the logistics operations, reverse logistics is one aspect closely related to customer satisfaction. Since huge shipping costs are involved in moving a product from the inventories to the customer’s doorstep, businesses incur significant losses on return orders without even making a sale. To help the sellers reduce the percentage of Return-to-Origin (RTO) orders, Pickrr has introduced a solution called Pickrr Predict. This value-added service on our platform provides predictive customer analysis to the sellers by considering nearly 50 parameters against every order and identifying their risk percentage. Sellers from across the country can then take necessary measures to mitigate the significant risk proportion attached to RTO orders, even before they are dispatched. Online sellers face a lot of hiccups in ensuring last-mile deliveries to difficult-to-reach terrains, primarily due to logistics barriers. It adds to their struggle to expand to these markets as customers prefer their shipments delivered within the same day, considering their impulsive buying habits. So, if a brand takes a long time to deliver an order, the shopper often changes his mind and refuses to accept the parcel. To expedite shipments to such parts of the country with a quick turnaround time and reduce the shipping costs for sellers, Pickrr has its pan –India network of intelligent Fulfillment centres. Since Pickrr Plus Fulfillment facilities are closer to the customers’ locations, our solution makes it easier for the sellers to promise and fulfil light-speed deliveries with reduced shipping costs. In addition, we are working with 20+ delivery partners and many local courier services to ensure timely deliveries. AIM: What technologies do Pickrr use to drive operational efficiency? Gaurav Mangla: To make the logistics operations more efficient, we have introduced various value-added services like Pickrr Connect, Pickrr Predict, etc., ensuring quality customer service and helping businesses thrive. Pickrr has a courier partner recommendation engine that suggests the suitable courier partner for a certain consignment that will get the fastest delivery done in any region. Pickrr Connect allows our sellers to stay connected with their end customers by informing them about their package at every step. The sellers can connect with their customers during all seven different stages of the order journey, from the order placing to delivery. As a result, the service builds transparency, ensures customers’ trust, and contributes to customer retention. Pickrr Predict feature uses a predictive algorithm that runs through the orders before their dispatch and corroborates the risk percentage based on nearly 50 parameters. This allows the sellers from across the country to take necessary measures on high-risk orders and reduce the percentage of Return-to-Origin (RTO) even before they are out for delivery. The extensive network of Pickrr Plus fulfilment centres is strategically positioned near the end consumers and has played a pivotal role in driving the growth of brands by cutting down transportation costs, executing same day and next day deliveries, and giving SKU performance intel to the brands for better inventory management. AIM: How does Pickrr’s one-click integration work? Gaurav Mangla: Pickrr offers ready-to-install customised APIs that seamlessly connect any online store with Pickrr’s platform. After the seller has signed up with Pickrr and set up the account, they can do the integration in three simple steps: Login to dashboard: After logging in, the seller needs to select the option of “Manage Channel.” Connect your store: Next step is to add the URL of the online store, and Pickrr’s algorithm will start fetching data automatically for integration. Approve the connection: The final step is to approve the connection and configure the status. Pickrr offers integrations with over 30+ e-commerce channels such as Woocommerce, Shopify, Magento, Unicommerce, WordPress, Vinculum, etc. AIM: How is the logistics sector evolving in India? Gaurav Mangla: The logistics sector in India is witnessing unprecedented changes, and the recent shift of businesses into the online space has led to a surge in demand for faster and more efficient logistics. Impulsive buying habits of consumers are also prompting businesses to quicken their deliveries, even ahead of the promised time. Technology is playing a big role in shaping the sector: Big data management: As more online shoppers join the e-commerce industry, the data gets larger in volume; hence, managing data has become a mammoth task. Players in logistics sectors are utilising the same to identify consumers’ purchasing patterns. Then, based on the shopping requirements and interests of the customers, the brands can show them more options. Faster delivery: With the surge in online shoppers, there is an unprecedented increase in the demand for faster deliveries. Big e-commerce players offer options for same-day delivery to provide a better customer experience. However, last-mile deliveries continue to be a persistent challenge for most emerging brands. Logistics players have been filling this gap and disrupting the industry by utilising advanced technologies such as RTO prediction, analysing the consumption pattern of the market, live tracking of the shipments, etc., leading to the best customer experience. Warehousing: According to the study by CARE Ratings, 90% of the warehousing space in India falls under unorganised players. The current opportunities propelled by organised private players are standardising the warehouse segment, leading to significant improvements. These facilities use advanced technologies for better inventory management, faster package allocation, and swift deliveries. The extensive network of Pickrr Plus Fulfillment centres stands as the backbone of last-mile operations for the company. Pickrr has a pan India reach of Pickrr Plus Fulfillment centres where sellers can store their products for faster deliveries of the shipments to their end customers. In addition, Pickrr’s warehouse management system (WMS) allows sellers to monitor their inventory in real-time across all their fulfilment facilities, providing insights into their SKU performance.","excerpt":"Online sellers face a lot of hiccups in ensuring last-mile deliveries to difficult-to-reach terrains, primarily due to logistics barriers.","categories":["AI Startups"],"tags":["AI Startups"],"author_name":"Avi Gopani","publish_date":"2022-04-17T13:00:00","publication_year":"2022","word_count":1081,"keywords":["big data","Go","API","AI","ML","Aim","analytics","Rust","GAN","R","AI Startups"],"extracted_tech_keywords":["AI","ML","analytics","Aim","R","Go","Rust","API","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-gurugram-based-startup-leverages-ai-to-power-logistics-supply-chain-solutions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":1492,"title":"Interview: Tushar Chhabra, Co-founder and CEO at CRON Systems","content":"Tushar Chhabra is the co-founder of CRON Systems Last month IoT startup, CRON Systems raised a pre-series A funding from early stage investor, YourNest. The startup was only founded in 2015, and with its unique value proposition and brilliant leadership never looked back. The firm was founded by three minds, Tushar Chhabra, Saurav Agarwala and Tommy Katzenellenbogen. CRON Systems deploys solution for border security, leveraging its competencies such as deep research in lasers, data, Artificial Intelligence (AI), encrypted communications, and automation. Earlier this month, IoT India Magazine got in touch with the CEO of CRON Systems, Tushar Chhabra to get a more transparent picture of CRON’s unique value proposition in the defense space, besides understanding how CRON leverages IoT to offer innovative solutions. We present before you the detailed interview as follows. IoT India Magazine: What are the factors that spurred the conception of CRON Systems. Please talk about your journey so far in the IoT-based defense space. Tushar Chhabra: When you look at the challenges and solutions our Border Security Forces face and try to solve them from a technological point of view, it is very easy to understand how CRON Systems came to be as there is a huge gap between what today’s technology can offer and what is actually implemented at the front lines. While this gap was the spark that ignited CRON Systems, the “fuel” was the combination of a diverse team with unique experiences and expertise, ranging from Google (Tushar Chhabra, CEO), to Qualcomm (Saurav Agarwala, CTO), to the Israeli Defense Force (Tommy Katzenellenbogen, Chief Strategy Officer). We believe that any team with the right skills and motivations, given the same problem, would have probably reached the same conclusion. Our journey started at the Morcha with a few Jawans, where we tried to understand how they live and what the daily routine and challenges are and tried to really understand what can be automated and deployed at scale. We also looked at the type of threats the forces encounter and the environmental challenges. Once we analyzed all the information, we developed a prototype solution with the BSF and worked with them to fine tune it and test it. This was the beginning of the KV product line, also know as the Laser Wall. Once we had a product that addressed a specific problem, we realized that the physical infrastructure of the Laser Wall could be a great platform for additional sensors and communication networks, and from there a whole new world of possibilities opened up to us. We are currently considering ourselves not as an IoT company, but more as an IoST (Internet of SECURITY Things) as we are aiming to become a platform to connect not only our products, but every 3rd party security product and solution that could benefit the defense forces securing borders. IoT India Magazine: Please highlight some of the current technological trends that is shaping the defense space. What is the potential that IoT holds for streamlining defense activities? TC: Even though the potential of IoT appear to be endless, it is very difficult to actually develop an effective product or application. This is probably why IoT has been around for a long time but hasn’t really hit the mainstream and taken off yet. But IoT Technology is becoming cheaper, more reliable and more secure, and we believe this is going to bolster the adoption rate from here on forward.  Defense and Security are actually perfect verticals for the IoT (or IoST, as we call it). These verticals are characterized by an abundance of sensors gathering huge amounts of information and intel, which needs to be analyzed and cross checked at scale to produce Real Time Actionable Intel, with which the operators need to make decisions with and act upon. As the technology becomes more reliable and more secure, we will definitely see more IoST startups (and also ML & AI) in the near future. Saying that, all will continue to have a difficult time establishing themselves (compared to the B2C and B2B markets) as the barriers to entry for startups will continue to be high. This is due to the nature of the defense space, which is less tolerant to unproven tech, and working with startups that don’t have the resources to support a client such as the military. IoT India Magazine: We congratulate you on receiving from YourNest. Please shed light on this most recent funding, describing how CRON Systems has planned to utilize this fund. TC: Once we found Product-Market Fit, we decided that it was time to scale and started to look for funding from early stage investors, which was not an easy task, being in the defense space – most investors do not feel comfortable with this space as they don’t have the necessary skills to analyze a deal in this space. Fortunately to us, we managed to identify a number of interested parties and chose YourNest. We identified a real partner who understood our vision and was not afraid to venture forward and support us on this path. Our goal is to use the funding predominantly to hire top talents to develop the next generation of our product and scale our company. We have tripled our numbers and we are still looking for top talent, which is a challenge by itself. IoT India Magazine: CRON deploys solutions predominantly focused on border security. BSF and Indian Army make use of your products to secure the international borders as well as perimeter of sensitive installations like army camps, airports etc. Can you shed light on how your solutions helps maintain border security? TC: When it comes to security and defense, there is no “one technology to rule them all” that can solve all problems and address every threat. The best strategy is to have a layered architecture with a combination of different sensors and technologies and this is exactly what CRON systems products are all about. We provide the platform that enables this concept and we call it DETECT -> VERIFY -> ACT.  This architecture incorporates a detection grid (such as the laser walls) that use several sensors to DETECT activity in the perimeter. Once an alert has been raised, we cross-reference all the information of all sensors with other data sources we have, and VERIFY if the threat is real or just a false positive (for example, a donkey or boar wandering near the perimeter). The information is then presented to the commander and shared with other platforms and systems that can ACT and intercept, or even neutralize the threat, based on the type of procedures the security force has in place. Systematic and automated process can be deployed at scale across an entire border and instantly upgrade and future proof any perimeter. IoT India Magazine: Please walk us through some of your IoT-based products and services. Please talk about your KV product line which offers all terrain, all weather breach notification. How is it helping your clients? TC: The KAVACH “laser wall” is a laser and IR based detection grid. The system is comprised of 2 units that run a laser and an IR between each other and can detect if and what and how something crossed its grid. The laser wall not only detects a breach but can also sense if something or someone is approaching it and map its immediate vicinity. The system also incorporates a long range, independent and encrypted communication network which is used to connect all laser walls and even other defense systems and products to the Command and Control dashboard which gives the operator a bird’s eye view on what is happening in his vicinity. The system auto-aligns and is water and dust proof, which allows for a very quick and easy deployment and operation. IoT India Magazine: Could you provide us with a case study where your IoT solutions helped to address border security needs or your IoT products were utilized for any commercial application. TC: We have had great feedback from the BSF who already deployed our earlier version product at the border. The main use of the unit was in specific areas with harsh and barely accessible terrain and a large water body that has been very difficult and dangerous to cover and by deploying laser walls at these areas, the BSF could decrease the amount of physical patrols, which were being shelled and exposed to snipers on a consistent basis, while still maintaining (or even improving) the level of security at the area. As for the commercial application for the technology, we had a pilot project with one of India’s largest ecommerce platform, where laser walls were used to protect their warehouses. By implementing our systems, they could cut the cost of men power needed to secure the warehouse while increasing the efficiency of the security provided. IoT India Magazine: What’s the state of IoT technology in India. What is CRON doing towards driving an IoT ecosystem in the country? TC: India’s technological strength up until recently was mainly IT and ecommerce while IoT was not mainstream. This translates into the number of engineers and their experience in the IoT space to be relatively scarce and hard to find. We hope that with our recent funding becoming public and (hopefully) future success, CRON systems will increase awareness of the space and demonstrate India’s ability to shine in the IoT space as well and encourage other entrepreneurs and startups to develop this still relatively nascent space and become a global player. IoT India Magazine: Please present us with a picture about innovation that takes place at CRON. Also, tell us about the work culture imbibed at CRON. TC: CRON systems culture is unique in the sense that it’s very fast paced, hard working, and yet still leaves room to learn a lot and expand the professional horizons of our employees. When we hire, we are looking to hire the best talent with the right mindset to solve problems with technology. We are not looking for GPA or titles but rather looking people in the eye and searching for that spark of joy that comes with solving a problem for our customers with technology. Having this type of fast, cutting edge, learning combination makes hiring a bit of a challenge and not everyone fits or feels comfortable with this culture but once we hit the right person, possibilities are endless and the joy of accomplishment is highly rewarding. IoT India Magazine: What is the roadmap that lies ahead for CRON? Please talk about any upcoming projects involving IoT, or talk about any geographical expansion that CRON might undertake. TC: CRON systems is at a critical point right now where we have a number of customers very interested in our products and our short-term goals will be to deliver our best solutions with the best results. While our ops team is focusing on production and delivery, our core R&D team is working hard on our next line of products and upgrades to the Laser Wall line – incorporating additional sensors to our platform and opening our encrypted communication network to 3rd party defense and security products, which we believe will be an additional force multiplier to our customers. As for geographical expansion, we initially decided to focus all our attention to the Indian market, as it is a huge market by itself, but we have been already approached by investors and clients internationally, specifically from Germany, Africa, and the Gulf countries, and we are demoing our products internationally as well, as the opportunity is huge and the timing is perfect. In the background, we are also continuously hiring additional engineers. An additional major milestone we will have this year is raising our Series-A round which will be used to accelerate our scale up and fuel our expansion into international markets.","excerpt":"Last month IoT startup, CRON Systems raised a pre-series A funding from early stage investor, YourNest. The startup was only founded in 2015, and with its unique value proposition and brilliant leadership never looked back. The firm was founded by three minds, Tushar Chhabra, Saurav Agarwala and Tommy Katzenellenbogen. CRON Systems deploys solution for border […]","categories":["AI Features"],"tags":["Interviews and Discussions","Sensors India"],"author_name":"Дарья","publish_date":"2017-02-15T13:54:44","publication_year":"2017","word_count":1973,"keywords":["Go","API","artificial intelligence","AI","ML","RAG","automation","Aim","ViT","Sensors India","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","RAG","R","Go","API","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-tushar-chhabra-co-founder-ceo-cron-systems\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10070205,"title":"Microsoft launches Microsoft Climate Research Initiative","content":"Microsoft has launched the Microsoft Climate Research Initiative (MCRI). The tech giant said that MCRI is a community of multi-disciplinary researchers working together to fight climate change. Initially, it is focusing on three critical areas in climate research where computational advances can drive key scientific transformations. These include: Overcoming constraints to decarbonisationReducing uncertainties in carbon accountingAssessing climate risks in more detail Results to be made public As per Microsoft, all results of this initiative are expected to be made public and freely available to encourage broader research and progress on these important climate issues. “As researchers, we’re excited to work together on projects specifically selected for their potential impact on global climate challenges. With Microsoft’s computational capabilities and the domain expertise from our academic collaborators, our complementary strengths can accelerate progress in incredible ways,” added Karin Strauss, Microsoft Senior Principal Research Manager. Microsoft has laid down various goals it wants to achieve through this program. Accelerate cutting-edge research and innovation in climate science and technology through collaborationDevelop and sustain a highly collaborative research ecosystem comprising diversity of perspectives across representation, expertise, institution, and geographyProvide strategic direction for climate-related research priorities and investments Microsoft researchers will be working with collaborators globally to co-investigate priority climate-related topics and bring innovative, world-class research to influential journals and venues.","excerpt":"As per Microsoft, all results of this initiative are expected to be made public and freely available to encourage broader research and progress on these important climate issues.","categories":["AI News"],"tags":["Microsoft","researchers"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-30T18:36:33","publication_year":"2022","word_count":215,"keywords":["Go","programming_languages:R","AI","innovation","researchers","programming_languages:Go","RAG","R","Microsoft"],"extracted_tech_keywords":["AI","RAG","R","Go","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-launches-microsoft-climate-research-initiative\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10072685,"title":"BlenderBot — Public, Yet Not Too Public","content":"After the strong performance exhibited by Meta’s chatbots BlenderBot and BlenderBot 2, Meta AI launched BlenderBot 3, a 175B-parameter, publicly available chatbot. Built with all the skills of its previous versions such as long-term memory, personality and empathy, BlenderBot 3 shows a 31% improvement in overall performance related to conversational tasks. Meta AI incorporated several new features in the latest version of the chatbot. For example, the model has been trained using a new learning algorithm, Director, that enables the bot not only to provide the most relevant responses to users’ input but also distinguish between right and wrong. New algorithms were used to ensure that the bot is able to distinguish between harmful responses and harmful examples. BlenderBot 3 uses a new safety recovery technique to respond to challenging conversations in a more civil manner. The most remarkable aspect is the option of a live interactive demo that enables BlenderBot 3 to learn from organic interactions with all kinds of people. Notwithstanding the new and improved features of the latest version, what has triggered a lot of discussions is the ‘publicly available’ aspect of BlenderBot 3. How ‘public’ is BlenderBot3? In its blog, Meta announced it is a publicly available chatbot. In fact, it has provided links to get the code and datasets. “In addition to sharing the model weights and code, we’re publishing new datasets and model cards so that other researchers can leverage BlenderBot 3 in their own work. We believe this open research approach will accelerate progress in conversational AI,” tweeted Meta AI. For a moment, any reader might be tricked into thinking that like BlenderBot 2, the latest version is also open-source. Well, like they say, all that glitters is not gold. In the case of BlenderBot 3, one can say all that is publicly available and available for open research is not open source. The moment one goes to access the model code, one gets to know that only the smaller models, 3b and 30b, can be accessed. For the larger 175b model, one needs to request access. Source: Google Docs While requesting access, one needs to provide details like the organisation one is affiliated to, intended use and others. In fact, the Google Doc page displays a message – “To access the largest model, please fill out this form, and we’ll let you know if your use case is approved. If you have a .edu email address, please use that instead of any other personal email address”. From the message it is tacit that someone with a ‘.edu’ email address would be preferred more than the others. As a footnote to the blog, Meta cites that access to the 175b parameter model will be limited. Access will be granted to academic researchers and people affiliated to government organisations, civil society groups, academia and global industry research labs. Thus, what Meta says to be publicly available isn’t easily accessible to many sections of the public. Further, discussions on Reddit and Twitter bring to light that access to Blender Bot 3 is restricted not just to certain groups of people but also based on region. A Twitter user shares a screenshot of a window that says “BlenderBot is not available in your location”. In fact, if any user who isn’t from the US, attempts to deploy the demo version, the window conveys that ‘BlenderBot is US-only at the moment’. Source: parl.ai That’s startling. If access is indeed region-based why call it ‘publicly available’? Well, there could be several reasons for this restricted availability. Some opinions doing the rounds in social media regarding geographical restrictions is probably because of the imminent blackout Facebook could face in Europe if the Irish Data Protection Commission’s draft decision to block Meta from sending data across the Atlantic turns true. Perhaps, Meta isn’t very happy with such decisions being deliberated upon. “Zuck’s probably angry with the EU forcing them out and all that controversy about the metaverse,” commented a Reddit user. As stated by Meta, BlenderBot 3 is available for public demo. However, a public demo is not without challenges. There is no guarantee that users deploying the demo version are well-intentioned and would not engage in toxic conversations that could be offensive and disrespectful. This  could be a reason for the restricted access. “We believe that long-term safety is an important component of quality chatbots — even if it means sacrificing engagingness in the short term,” notes Meta in the blog. Meta seemed to be aware of the criticisms it could face once the BlenderBot 3 releases. Thus, it conveyed its plan to improve the models using feedback from demo users’ interaction and release updated models for the benefit of the larger AI community.","excerpt":"As a footnote, Meta cites access will be granted to academic researchers and people affiliated to government organisations, civil society groups, academia and global industry research labs.","categories":["AI Features"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-08-12T13:00:00","publication_year":"2022","word_count":785,"keywords":["Go","Meta AI","programming_languages:R","AI","chatbots","programming_languages:Go","RAG","GAN","R"],"extracted_tech_keywords":["AI","Meta AI","RAG","chatbots","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/blenderbot-public-yet-not-too-public\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":67853,"title":"How To Build Your Own Homemade RFID Card With Arduino","content":"Radio Frequency Identification (RFID) technology is a system capable of identifying objects through a unique identifier (UID). The RFID system consists of two main components: an RFID tag and an RFID reader.The RFID tag has different formats. It can be a sticker like the one placed on supermarket products to prevent theft, on a plastic card, on a keychain, and it can even go under the skin like in animals. There are read-only RFID tags. These labels are engraved with the identifier in the manufacturing process. Once a read-only RFID tag is generated, the UID can no longer be modified. However, there are other types of RFID tags that are read and written. Such a tag can change your information thanks to RFID readers such as RC522 that allow you to read and write RFID tags of this type. In reality this component not only reads information, it is also capable of emitting a radio frequency signal. How RFID Works Source: Replica Sistemi The power of the RFID reader and RFID tag is different. A module like RC522 must be connected to the electrical network to function. But what about RFID tags? How do you feed these types of cards? The typical RFID stickers that go on many products do not have any batteries or power supplies. Neither do RFID-based access control cards. The basic operating principle on which they are based is induction. These types of RFID tags are known as passive. To read the information encoded on a passive RFID tag you need to withdraw the electromagnetic field that causes the electrons to move through the tag’s antenna and then power the chip. When this happens, the powered chip is able to send the information stored in the RFID tag through the radio frequency. This is called backscattering. The backscatter is detected and interpreted by the RFID reader, which then sends the data to a computer or a microcontroller such as the one with Arduino. RFID types Just as a radio must be tuned to different frequencies to listen to different channels, RFID tags and readers must be tuned to the same frequency in order to communicate. An RFID system can use multiple frequencies within the radio frequency spectrum. The radio frequency spectrum is between the Extremely Low Frequency (ELF) band and the infrared. Source: Resource Label Group There are three types of system depending on the frequency they use. Low frequency or LF (125-134 KHz)High frequency or HF (13.56 MHz)Ultra-high frequency or UHF (433, 860 and 960 MHz) Radio waves do not behave the same at all frequencies in the radio spectrum. This forces us to choose a frequency depending on the application we want to build. Low frequency RFID or LF systems RFID applications that use low frequency or LF have a long wavelength and can better penetrate thin and metallic objects. Furthermore, RFID systems using LF are ideal for reading objects with high water content such as fruits and drinks. However, the range of the low frequency range is limited to a few centimeters. Memory on RFID tags is very limited due to the low data transmission rate and high production cost. Typical RFID applications using low frequency include access control and tagging of animals. High frequency RFID or HF systems These use high frequency work very well on metal objects and products with a medium and high-water rate. Generally, these RFID systems operate in a range of centimeters although the maximum reading is around the meter. They use global standards and protocols such as NFC with larger memory options, although the reading range is still short and the data transmission rate is low. Ultra-high frequency RFID or UHF systems RFID systems based on ultra-high frequency generally offer a greater range than LF and HF types. We speak from a few centimeters to more than 20 meters. Another difference is that they can read data faster –several labels per second. This technology is used to monitor warehouse items, to count people, to control races when runners cross the finish line, and for tolls and for access control to parking lots. RFID reader with Arduino RC522 The RFID reader module RC522 is based on the MFRC522 integrated circuit. It is usually accompanied by an RFID tag in a credit card format and an RFID tag in a keychain format. The cards can have 1K or 4K memory divided into sectors and blocks. The RFID reader module RC522 is also used to write RFID tags. Source: Shutterstock The RFID reader RC522 uses the high frequency HF creating an electromagnetic field of 13.56 MHz. In theory it has a maximum range of 35 centimeters. Something very interesting is that this module comes with a very useful interrupt pin. It is used so that instead of asking the RFID reader RC522 over and over again if there is an RFID tag nearby, the module will warn us through the pin when an RFID tag approaches and thus be able to activate the microcontroller. The operating voltage is 2.5V to 3.3V. This implies that it is compatible with most microcontrollers on the market such as those used by Arduino. The good news is that despite being powered by 3.3V, the logic levels are compatible with 5V making it compatible with any Arduino or microcontroller that works at 5V. These would be the complete specifications taken from the technical data sheet: Frequency range13,56 MHzInterfaceSPI\/I2C\/UARTSupply Voltage2,5V to 3,3VMax. current13-26 mAMin. current10 uALogic levels5V y 3V3Reach5 cm Components of the RFID RC522 VCC: RFID reader power pin RC522. It supports a supply voltage between 2.5V and 3.3V.RST: it is a pin to turn the module on and off. As long as the pin is in the LOW state it will stay off with little consumption. When the state changes to HIGH the RC522 restarts.IRQ: interrupt pin that alerts the microcontroller when an RFID tag approaches the RFID reader RC522.MISO \/ SCL \/ TX: This pin has three functions. When the SPI interface is enabled, it functions as slave output and master input. MOSI: entry in the SPI interface.SCK: clock signal of the SPI interface.SS \/ SDA \/ RX: The pin acts as a signal input when the SPI interface is enabled. RFID reader connection RC522 with Arduino UNO The communication between Arduino and the RFID reader RC522 is quite complex at the programming level. Fortunately, we have ready-to-use Arduino tools such as the Grove Beginner Kit for Arduino, an All-in-one Arduino compatible board with 10 sensors and 12 projects that will surely help us saving valuable time. First, we will run DumpInfo. This program does not write any data to the RFID tag. Just read the RFID tag if you can and display the information on the serial monitor. Let’s have a look at the code. This will vary depending on the purpose for the RFID :","excerpt":"Radio Frequency Identification (RFID) technology is a system capable of identifying objects through a unique identifier (UID). The RFID system consists of two main components: an RFID tag and an RFID reader.The RFID tag has different formats. It can be a sticker like the one placed on supermarket products to prevent theft, on a plastic […]","categories":["Deep Tech"],"tags":["arduino","no etl"],"author_name":"Dr. Raul V. Rodriguez","publish_date":"2020-06-22T12:01:00","publication_year":"2020","word_count":1144,"keywords":["Go","TPU","programming_languages:R","AI","arduino","programming_languages:Go","no etl","R"],"extracted_tech_keywords":["AI","TPU","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-build-your-own-homemade-rfid-card-with-arduino\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162055,"title":"85% Executives Are Unprepared to Implement Responsible AI: HCLTech-MIT Report","content":"A recent study by HCLTech and MIT Technology Review Insights reveals that while 87% of business executives consider responsible AI principles critical, 85% admit they are not adequately prepared to adopt them. The report, titled Implementing Responsible AI in the Generative AI Age, was launched during the World Economic Forum’s Annual Meeting in Davos. It highlights the widening gap between recognizing the importance of responsible AI and the ability to implement it effectively. Steven Hall, President of Europe and Chief AI Officer of ISG, remarked, “Everybody understands how transformative AI is going to be and wants strong governance, but the operating model and the funding allocated to responsible AI are well below where they need to be given its criticality to the organization.” According to the reports, the key challenges hindering responsible AI adoption include, complexity of implementation, lack of expertise, operational risk management difficulties, regulatory compliance hurdles, and insufficient resource allocation. Despite these challenges, the study finds optimism as enterprises plan to increase investments in responsible AI over the next twelve months. Vijay Guntur, CTO & Head of Ecosystems at HCLTech, emphasised, “AI can be a tremendous force of positive change in businesses and society at large, but its full potential can only be realized when it can be trusted.” The report provides actionable recommendations for enterprises, including: Establishing a robust responsible AI framework with guiding principles for ethics, safety, compliance, and change management. Leveraging partnerships within the technology ecosystem to pilot and integrate best practices. Setting up a dedicated Center of Excellence to drive cross-functional AI initiatives. HCLTech has proactively established an Office of Responsible AI and Governance, led by subject matter experts with experience in frameworks such as NIST and the Europe AI Act. This office focuses on co-innovation, consulting capabilities, and intellectual property solutions related to responsible AI. The study surveyed senior business leaders across various industries worldwide and found that AI-driven transformation is progressing from proof of concept to wider adoption, with growing interest in applications across customer service, software development, and marketing. Additionally, agentic AI—operating autonomously with minimal human involvement—is gaining traction in lower-risk areas like IT operations. While half of the respondents feel confident in managing operational risks, fewer than 25% are prepared to address user adoption, change management, and bias-related issues.","excerpt":"Additionally, agentic AI—operating autonomously with minimal human involvement—is gaining traction in lower-risk areas like IT operations.","categories":["AI News"],"tags":["HCL Technology"],"author_name":"Mohit Pandey","publish_date":"2025-01-23T15:18:00","publication_year":"2025","word_count":379,"keywords":["Go","agentic AI","AI","innovation","RAG","responsible AI","generative AI","HCL Technology","Rust","GAN","R"],"extracted_tech_keywords":["AI","generative AI","agentic AI","RAG","R","Go","Rust","GAN","responsible AI","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/85-executives-are-unprepared-to-implement-responsible-ai-hcltech-mit-report\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":64415,"title":"Should You Love Or Be Scared Of Maths Required For Data science?","content":"Data science is the future, everyone wants to learn this budding technology. Is everyone able to learn? The answer is “No”. Do you know the reason, it is none other than “mathematics”. What….did I say mathematics? Yes, Mathematics or simply math. While reading this, people who know what is data science or has worked in this field would be confused and would be asking how maths is responsible. Let me, rephrase my answer, “ Fear to Mathematics”. Starting from our elementary education to our higher education we see students scared of mathematics or we can say students have a math phobia. Not sure who has created this buzz that data science requires a long list of math topics as a prerequisite. It is not completely correct, elementary math is required but, as a beginner, you don’t need that much math for data science. Also, there is another side to data science and that is the practical side. For practical data science, a great deal of math is not required. Practical data science only requires skills to select the right tools. Being said that let’s understand how theoretical and practical data science differs. Difference between Theory and Practice When we talk about data science, it is important to set our goal. What we want to achieve from learning data science? It is for academic learning or for practical purpose to build career on. Why this goal setting is important because priorities and deliverables are different in both theoretical and practical data science. Learning data science for academic purposes is more for publishing research papers and push the field forward. While practical use of data science is to generate reports, build models and system software. Skills required for foundational Data science As a beginner in data science, one will primarily work on foundational\/fundamental skills of data science. These skills are required in each and every data science project. What are these skills? Data manipulationData VisualizationData analysis, also know as EDA (Exploratory Data Analysis) It is a known fact that in any data science project 75% of effort and time is spent doing these fundamental steps. These are the core skills required for success of any data science project. If any of the above step is missed out or is not carried out with utmost precision, final model might not be as good as it should be. Coming back to the question, how much math is required for these core skills? – Very little. So, by now you must be pretty much clear and convinced that math required for data science is not scary at all. What concepts of math are required for these foundation skills of data science? You would be wondering that still some math is required , so what all topics are there. Let me break the good news to you, you just require lower-level algebra and simple statistics. Don’t be astonished, it is a fact. Feeling delighted !! Let me explain it in a bit more detail. For getting the data, cleaning it and understanding it, doesn’t require any math. If new variable creation is required by deriving it’s value from already present variable. Then , it requires elementary math of addition, subtraction, multiplication and division. It may be required to manipulate the data by calculating mean, median or mode. It could be seen that none of the calculations here require complex maths. For 95% of cases above method holds true but exceptions will always be there. Very rare cases would require complex computation. Again it is very rare. One major tasks in data science is data visualization. In this step graphs and plots are created to check patterns in data. This is like a subset of exploratory data analysis step. So, where math is required in this? One should be aware of which plot to create and which tool to use. So, problem solved, no math is required in this step as well. One should know how to read plots and graphs whether it be scatter plot, histogram, line graph, point plot, etc. This is the only demand from this step. By now 75% work of any data science project is done. Now comes model creation using machine learning. Here again concept of theory and practical applicability of machine learning comes into picture. We have to create models for our business purpose and not to dig deep in model theoretically and publish a research paper. Conceptual understanding of machine learning models again don’t require math. One should learn which model to used in which situation. How to interpret the model. How to check assumptions in model. How to make predictions from model. Have I mentioned math anywhere, no. So, in short no advance math is require to become a data scientist. Basic skills required to become a foundational data scientist Selection of right toolUnderstanding the syntaxBasic graphs and chartsConceptual understanding of different machine learning modelsBasic AlgebraBasic Statistics This is mostly all you need to get started learning data science. Motivation and will to learn is utmost important. But always remember saying by Galileo Galilei: “ Mathematics is the language with which God has written the universe.”","excerpt":"Data science is the future, everyone wants to learn this budding technology. Is everyone able to learn? The answer is “No”. Do you know the reason, it is none other than “mathematics”. What….did I say mathematics? Yes, Mathematics or simply math. While reading this, people who know what is data science or has worked in […]","categories":["AI Highlights"],"tags":["what is data science"],"author_name":"Netali Agrawal","publish_date":"2020-05-04T14:00:00","publication_year":"2020","word_count":858,"keywords":["data science","Go","machine learning","programming_languages:R","AI","programming_languages:Go","what is data science","R"],"extracted_tech_keywords":["AI","machine learning","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/should-you-love-or-be-scared-of-maths-required-for-data-science\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168927,"title":"Bengaluru-based CognitiveLab Wins 2024 Meta Llama Impact Grant for Multilingual AI Project Nayana","content":"Meta has named CognitiveLab as one of the global recipients of its 2024 Llama Impact Grants, recognising the lab’s work on multilingual AI through its Nayana project. The Bangalore-based research group will use the grant to expand access to AI across over 22 languages, aiming to reach more than three billion people. Announced today, the Llama Impact Grants support open-source initiatives using Meta’s Llama models to address social challenges. Nayana, developed by CognitiveLab, is a multilingual, multimodal language model that integrates Llama for document and image processing, with support for low-resource Indic languages. The model includes capabilities across text, vision, and speech, and has outperformed existing benchmarks in OCR for ten Indian languages. Shivnath Thukral, vice president and head of public policy at Meta India, said, “Open-source AI is a powerful tool to bridge digital divides, especially in a diverse country like India. With the 2024 Llama Impact Grant, we’re proud to support the Nayana project. Its work embodies the spirit of open innovation—making advanced AI usable for billions of people.” CognitiveLab founder Shashi Kumar said the grant would accelerate their development efforts. “The Llama Impact Grant enables us to supercharge our efforts with Nayana—expanding language coverage, enhancing multimodal capabilities, and building high-quality training datasets for low-resource language communities,” he said. “Open-source and Llama have empowered us to build world-class systems like Ambari and Nayana with minimal resources.” AIM previously spoke to Adithya S Kolavi, co-founder of CognitiveLab, during the launch of the Indic LLM Leaderboard. Founded in 2023, the Llama Impact Grant programme supports work built on Meta’s open-source LLMs—Llama 2, Llama 3, and Llama 4. With over 1 billion downloads and more than 85,000 derivative models, the Llama family has become widely adopted across global research and development communities. CognitiveLab plans to use the funding to expand Nayana’s language and multimodal coverage, improve its Indic tokeniser, and develop new datasets for speech, text, and image processing. The lab also aims to release tools for deployment in low-resource environments and collaborate with the community to set new benchmarks for multilingual AI. India continues to be one of the largest markets for Llama adoption, where open-source developers are using the models to build solutions for local needs.","excerpt":"CognitiveLab plans to use the funding to expand Nayana’s language and multimodal coverage, improve its Indic tokeniser, and develop new datasets for speech, text, and image processing.","categories":["AI News"],"tags":["Meta"],"author_name":"Siddharth Jindal","publish_date":"2025-04-30T13:51:54","publication_year":"2025","word_count":367,"keywords":["Meta","funding","AI","Modal","Llama 4","innovation","Git","RAG","Aim","llm_models:Llama","R"],"extracted_tech_keywords":["AI","Llama 4","Aim","RAG","R","Git","innovation","funding","Modal","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bengaluru-based-cognitivelab-wins-2024-meta-llama-impact-grant-for-multilingual-ai-project-nayana\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045130,"title":"Surveillance Tech Firm Clearview AI Raises $30 Mn Despite Shady Past","content":"Clearview AI is currently facing investigations by British and Australian governments. The surveillance firm has received a substantial amount in investment despite mounting class action lawsuits. Earlier this month, the New York-based startup raised $30 million in Series B. Clearview has scraped the web to cull billions of personal images of people worldwide without their permission. More than 1,800 law enforcement agencies have used Clearview’s  facial recognition technology, according to a leaked list accessed by Buzzfeed. Past investors Peter Thiel, Kirenaga Partners, a NY based venture capital firm, and Hal Lambert, Founder of investment company Point Bridge Capital, haD EARLIER invested in Clearview in the past. The company has not disclosed the identity of investors in the latest round. Lambert’ company has set up an exchange-traded fund called MAGA ETF “to let people invest in 150 companies from the S&P 500 Index whose employees and political action committees (PACs) are highly supportive of Republican candidates.” According to Point Bridge’s website, Lambert served on the Inaugural Committee for President Donald Trump and was the finance chair for the Texas GOP. Lambert has also taken to his social media profiles to criticise the Black Lives Matter protests and spread misinformation about the racial justice movement. Clearview AI in the headlines The British and Australian authorities have launched a probe into the company’s data scraping techniques. Canada has banned Clearview AI’s controversial tech. The American Civil Liberties Union in Illinois accused the app of violating the state’s Biometric Information Privacy Act, after which the company stopped selling its product to private companies in the US. Clearview has also faced legal action in Vermont, New York and California. Nonprofit news site Muckrock obtained emails between the New York Police Department and Clearview through freedom of information requests and released the data in April this year. The emails track a two year relationship between NYPD and clearview where the company offered trial rounds to the police department, testing the facial recognition tools in live investigations. While NYPD has downplayed its relationship with Clearview AI, it is evident that Clearview has indeed cut a deal. The emails showed cops used the app to conduct more than 5,000 searches. Meanwhile, state policies limit NYPD from creating an unsupervised repository of photos for facial recognition. Still, the emails show Hoan Ton-That, Clearview AI’s CEO, was introduced to NYPD deputy inspector Chris Flanagan in 2018. Following this, Clearview entered into a vendor contract with NYPD on a ‘trial basis’ in 2018. But accounts for NYPD were created till February 2020 even though the trial period should have ended in 2019. NYPD also offered to aid Clearview in selling its technology to the Homeland Security department. Controversy galore The police departments used Amazon’s Rekognition till 2020. In the wake of protests from civil liberty advocates and activists, Amazon had put a moratorium on facial recognition software. Palantir Technologies has government agencies as clients, including the Department of Defense, the Central Intelligence Agency (CIA) and the Immigration and Customs Enforcement (ICE). The company itself began with funding from the CIA, and despite the controversies surrounding it, Palantir went public in 2020. The company has raised a total of $2.6 billion in funding over 34 rounds. While controversy is ‘bad’ publicity for tech giants like Amazon, the attention seems to be working in favour of companies like Clearview AI. Companies like Palantir cashed in on the controversies to cut deals with government agencies.","excerpt":"Clearview has also faced legal action in Vermont, New York and California.","categories":["AI Features"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-08-03T11:00:00","publication_year":"2021","word_count":571,"keywords":["Go","API","funding","AWS","AI","venture capital","BERT","GAN","R","startup"],"extracted_tech_keywords":["AI","AWS","R","Go","API","BERT","GAN","startup","venture capital","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/surveillance-tech-firm-clearview-ai-raises-30-mn-despite-shady-past\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129522,"title":"NTT DATA Unveils Ultralight Edge AI Platform","content":"NTT DATA recently unveiled its new Edge AI platform to accelerate IT\/OT convergence by bringing AI processing to the edge. By processing data when and where it is generated and unifying diverse IoT devices, systems and data, this unique, fully managed solution enables real-time decisions, enhanced operational efficiencies and secure AI application deployment across industries to drive advanced Industry 4.0 technologies. Designed to support industry-specific requirements, the Edge AI platform leverages lighter, cost-effective AI models, enabling it to run within a small compute box. (NTT DATA Edge AI Device) Edge AI will perform specific tasks, such as supporting safety or operational efficiency, by collecting data from disparate devices across a network environment, enabling instantaneous and secure data processing and analytics. It is an all-inclusive managed service platform that includes all the systems, tools and capabilities required for AI at the edge. It addresses data discovery, collection, integration, computation power, seamless connectivity and AI model management. The Edge AI platform, supported by NTT DATA’s consulting data scientists, managed services and global technical resources, addresses the shadow IoT challenge and AI infrastructure requirements. It does this by auto-discovering, unifying and processing data from IoT devices and IT assets across the organisation, simplifying AI deployment and management. Manufacturing operations could benefit from improved predictive maintenance by accessing IT\/OT data from sensors, machinery, cameras and applications to plan and address failures. In addition, NTT DATA’s Edge AI can monitor and optimise energy consumption in real-time, predicting energy spikes and optimising machine usage, reducing costs and CO2 emissions with renewable energy.","excerpt":"NTT DATA recently unveiled its new Edge AI platform to accelerate IT\/OT convergence by bringing AI processing to the edge. By processing data when and where it is generated and unifying diverse IoT devices, systems and data, this unique, fully managed solution enables real-time decisions, enhanced operational efficiencies and secure AI application deployment across industries […]","categories":["AI News"],"tags":["NTT data"],"author_name":"Pritam Bordoloi","publish_date":"2024-07-18T18:03:42","publication_year":"2024","word_count":256,"keywords":["programming_languages:R","AI","ML","RAG","ViT","edge AI","analytics","emerging_tech:edge AI","GAN","NTT data","R"],"extracted_tech_keywords":["AI","ML","analytics","edge AI","RAG","R","GAN","ViT","programming_languages:R","emerging_tech:edge AI"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ntt-data-unveils-ultralight-edge-ai-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10069845,"title":"Hiring Hackathon &#8211; Data Engineering Championship ends successfully — Meet the winners who cracked the code","content":"MachineHack has recently concluded Data Engineering Championship – a hiring hackathon for data scientists and data engineers, organised in association with Publicis Sapient, iMerit, USEReady, Tiger Analytics & The Math Company. The hackathon was a part of the Data Engineering Summit 2022, presented by Google Cloud and organised by Analytics India Magazine, and was a huge success with over 700 registrations. The winners stood a chance to present their solution approach at DES 2022 & got an opportunity to land an interview with one of the leading analytics organisations. You can read more about the dataset here. Here are the solution approaches of the winners who secured the top three positions in the Data Engineering Championship. Rank 01: Sylas John Rathinaraj Rathinaraj got interested in predictive analytics in 2017. He attended Coursera and Udemy courses in statistics, exploratory data analysis (EDA), machine learning, data science and deep learning to improve his skills. In addition, he has participated in a slew of ML hackathons on different platforms to test and build his knowledge. Approach The participants were provided details about an airport along with a weather information dataset. They had columns such as ‘DATE’,’ LOW’,’ HIGH’, and’ TIMESTAMP’ for which the participants could impute the constant value. In the year column missing records, you can impute 2020 as they had 2020 as a year for all other records. It is the same with the month column where one can impute with 01 as we had 01(Jan) for all other records. The main challenges in the datasets were: Data missingnessFormula column with uncertainty Data missingness In the airport details with the weather information dataset, for several columns, 20 per cent of the data are missing. The bar chart below shows the non-missing records count of the columns. The formula for computing can be easily formulated with other dependent columns. For missing records in the dependent columns, used imputation based on group-by of the mean value. Formula column with uncertainty The definition for the WIND_CHILL column given in the competition was “the perceived temperature due to the cooling effect of wind blowing”. Rathinaraj utilised information from the TMAX (temperature max0, AWND (MAX wind speed of the day), SNOW and timing of the day when the flight departs. He used a combination of this information and calculated the WIND_CHILL columns. WIND_CHILL column is in ranges from 0 to 80 Fahrenheit. The WIND_CHILL column is vital in the competition to get the best score as the mean absolute error increases in the same range(0 to 80) for wrong calculation. Rathinaraj feels that MachineHack provides participants with different domains of the ML and Data Engineering competition. “Participating in the competition helps me to become more knowledgeable. After the competition ends, I always spend time exploring the top-ranked achiever’s solution approach and codes,” he adds. Check out Winners Solutions here. Rank 02: Jeena Binex Jeena has been working as an embedded system engineer for nine years in Mumbai, and for the last five years, she has been working in a courier company in Singapore, where her profile is to maintain the In-house ERP system which is built on .Net Framework and SQL Database and analyse the data available to identify the trends for sales, operations, customer service etc. “I started analysing the data with the limited knowledge I had, and my interest in data analysing started here and hence decided to have in-depth knowledge in this field. So, in July 2021, I enrolled in a data science online course. After spending 12 months in the course, I studied supervised and unsupervised Machine learning and Time Series. Then, I moved on to deep learning, NLP. Approach Jeena’s approach to the problem included the following steps: Reading through the dataset and understanding the meaning of each column of the dataset (26 columns)Reading through the features to be created and identifying the columns of the dataset contributing to the creation of features. The main agenda was to fill the missing values of these columns,She used two approaches for filling the missing values-Regressive imputation and Mean and median imputation. Finally, she calculated the features using the formulas. “Solving hackathons helped put into practice the knowledge I gained from the theory, which was a huge confidence booster for me,” concluded Jeena. Check out Winners Solutions here. Rank 03: Suresh Arunachalam Suresh has always been passionate about data science and curious about understanding its connection with real-world business use cases. “This curiosity enabled me to spend additional effort during the day and the weekends to learn more about it from the internet, which eventually created a pathway to knowing about the hackathon events happening across the globe in the data science space,” he said. Approach Suresh says that a use case was given to calculate Wind Chillness, Airline Seat Distribution, Snow Ratio and a few other useful pieces of information along with the date and time stamp, which helps the airline companies to plan their trips from the airport data dump. The dump contained about 200k rows and 26 columns with various information (such as wind speed, latitude, longitude, snowfall, flight ID, etc.). He followed these steps: At first, he removed the unwanted columns and replaced the null values using Max () and Median () methods from NumPy.  He did a column split to form the date and timestamp using Pandas.He then performed some basic arithmetic operations to calculate the expected use case results. “I was delighted to be part of this hackathon event conducted by MachineHack, which helped me to improve my analytical and problem-solving skills. Moreover, the rules and guidelines set by MachineHack for such events helped in intuiting my competitive skills to keep myself in the top three positions every day on the leaderboard,” Suresh adds. Check out Winners Solutions here.","excerpt":"MachineHack has recently concluded Data Engineering Championship – a hiring hackathon for data scientists and data engineers, organised in association with Publicis Sapient, iMerit, USEReady, Tiger Analytics & The Math Company.  The hackathon was a part of the Data Engineering Summit 2022, presented by Google Cloud and organised by Analytics India Magazine, and was a […]","categories":["Deep Tech"],"tags":["Data Engineering","data engineering career","data engineering demand","data engineering jobs","Hackathon","Hackathon Winners","Hackathon Winners India","Hackathons","hackathons in India","Hackathons India","machinehack winners","weekend hackathon winners","Winners","winners of MachineHack Hackathon","Winners of Weekend Hackathon","Winners’ Approach"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-28T17:00:00","publication_year":"2022","word_count":961,"keywords":["Hackathon Winners India","data engineering demand","Winners of Weekend Hackathon","deep learning","machinehack winners","Pandas","data science","NumPy","hackathons in India","NLP","Data Engineering","analytics","Hackathon Winners","machine learning","AI","ML","data engineering career","Hackathon","winners of MachineHack Hackathon","Winners’ Approach","Hackathons India","Winners","Hackathons","weekend hackathon winners","predictive analytics","data engineering jobs"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","data science","analytics","Pandas","NumPy","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hiring-hackathon-data-engineering-championship-ends-successfully-meet-the-winners-who-cracked-the-code\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":64000,"title":"This Mumbai-Based AI Video Bot Platform Raises Funding From LetsVenture &#038; Others","content":"Expertrons, a Mumbai-based AI video bot platform for career hacks, has raised capital in its seed funding round from Nikhil Vora, MD of Sixth Sense Venture, Iceland Venture Studio, LetsVenture, and Samyakth Capital. In this seed funding round a few other angel investors were involved such as — Rohit Chanana, Partner at Sarcha Advisors; SumitChhazed, the co-founder of OTO Capital; and Dr.SukantaGanguly, the President & Board of Directors at ClickIPO. The company plans to use this latest funding to enhance its tech capabilities and expand its product offerings to enable organisations with its video bots. The funding has come at a time when the global lockdown due to COVID-19 has triggered a significant rise in the need for upskilling professionals. The startup is founded by IIT Bombay alumni – Vivek Gupta and Jatin Solanki, with over eight years’ experience in education technology. Solanki’s previous venture was StepAppEduisfun, India’s largest gamified learning platform which served over a million students in the K-12 domain while Gupta’s previous venture Plancess, a test preparation company, was acquired by a public listed ed-tech company. Vivek Gupta, the co-founder of Expertrons, said, “The company is essentially Netflix for career hacks. We founded it with the vision to reimagining career decisions for the 1.87 billion professionals globally who change their careers 5 to 7 times in their lifetime.” According to company’s release — Expertrons already has the most extensive library of career experiences with over 12,000 minutes of video bots of 550+ experts who are pursuing a dream career at top firms like BCG, Google, One Plus, Morgan Stanley, Pepsi (Mexico), Rakuten (Japan) or even recent graduates from top B Schools like IIMs, ISBs, or Harvard. While these ‘experts’ are busy working or studying, their video bots can train millions of aspirants. Users can talk to the video bots of experts to get precise first-hand career hack experience by asking questions. “Expertrons’ deep tech AI recommendation engine based on numerous data points personalises the best-suited experts and career options for aspirants. After the unique video bot interaction, experts get a lifetime earning opportunity via referral bonus and one on one consultations to interested aspirants,” said Jatin Solanki, co-founder, Expertrons. The release further stated — in India, 30 million youngsters are completing their higher education and spending close to $5000 for it – all aimed at getting a dream career opportunity. A majority of these aspirants are first-generation graduates with no one in their circle to guide them on their career choices. Expertrons bridges this last mile gap to help students and professionals make the right career decisions. “Expertrons was a compelling deep tech for investors on the LetsVenture platform interested in ed-tech and allied verticals, jobs and professional networks. While the focus today is on career guidance, the core AI video bot technology could have some good parallel applications in large markets,” said Sunitha KR, Director, LetsVenture. According to Nikhil Vora, MD, Sixth Sense Ventures “Expertrons is solving a relevant problem. It’s a space that is very fragmented where every aspirant would like to do better professionally. Expertrons bring consolidation to the same in an innovative way that has the potential to go extremely viral. And the cherry on the cake is a strong team with both founders having run large organisations previously,” When asked further, BalaKamallakharan, MD, IcelandVenture Studio, which has made its first investment in India with Expertrons said, “We are impressed with the Expertrons team and their focus on execution. We believe that the democratisation of access to mentor networks is a valuable tool for anyone, especially when we are continuously redefining the future of work and professional development. Anyone can learn it by connecting to a video bot technology that enables the scalability of the mentors.” Besides being associated with 30+ colleges in India, Expertrons is amongst the top 3 startups globally to get selected for the TecLabs Accelerator, Mexico. As a part of which, they are doing a pilot with Mexico’s top university – Tec De Monterrey with 26 campuses, to help them boost admissions & placements. Expertrons was also awarded a special prize at Startup Masterclass, Pune conducted by IITK & IITD Alumni association. Other than the career guidance application of their AI video bot technology, Expertrons is also developing use cases of their technology to drive sales, clear customer queries, generate leads and even internal training for organisations.","excerpt":"Expertrons, a Mumbai-based AI video bot platform for career hacks, has raised capital in its seed funding round from Nikhil Vora, MD of Sixth Sense Venture, Iceland Venture Studio, LetsVenture, and Samyakth Capital. In this seed funding round a few other angel investors were involved such as — Rohit Chanana, Partner at Sarcha Advisors; SumitChhazed, […]","categories":["AI News"],"tags":["best technology to learn for future"],"author_name":"Sejuti Das","publish_date":"2020-04-30T12:32:49","publication_year":"2020","word_count":728,"keywords":["Go","API","funding","AI","Scala","RAG","Aim","GAN","R","best technology to learn for future","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Scala","API","GAN","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/this-mumbai-based-ai-video-bot-platform-raises-funding-from-letsventure-others\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":23402,"title":"EXCLUSIVE: AI Task Force Head V Kamakoti Explains Recommendations, Road Map Of The Committee","content":"One of the first things that come to mind while talking to Professor Kamakoti Veezhinathan from the Indian Institute of Technology, Madras (Chennai), is that he is keen on using new tech as a means to an end. His thumb rule is that artificial intelligence should act as a bridge to bring all disciplines together. Now as the head of the Artificial Intelligence Task Force set up by the Ministry of Commerce and Industry, Government of India, he, along with the other Task Force members are doing precisely that. Meet The AI Task Force: The AI Task Force was set up to prepare India for the upcoming Industrial Revolution 4.0 and the resulting economic transformation, with an emphasis on artificial intelligence. Veezhinathan explained how the AI task force set up by the Government is also an amalgamation of thought leaders from multiple disciplines — academicians, government officials, corporates and people with other specialisations. He added that it was an enriching experience working with such a multitude of talents. The 18-member AI Task Force includes: Anuj Kapuria, High Tech RoboticSystemz LtdAnurag Agarwal, Institute of Genomics and Integrative Biology, CSIRAshish Dutta, IIT KanpurAshwini Asokan, Mad Street Den, ChennaiGautam Shroff, vice-president and chief scientist, TCS Innovation Labs, GurgaonGH Rao, HCL TechnologyG Madhusudan, IIT MadrasGVN Apparao, ex-chief technology officer, CognizantKomal Sharma Talwar, founder, XLPATKunal Nandwani, founder & CEO, uTrade SolutionsShantanu Chaudhary, IIT Delhi, Department of Electrical EngineeringVijay Kumar Sankarapu, founder & CEO, Arya.inAjay Kumar, Additional Secretary, Ministry of Electronics & ITAmandeep Gill, Ambassador\/PR to CD, GenevaK Nagaraj Naidu, Joint Secretary (ITPO), Department of Economic AffairsAloke Mukherjee, DRDO What Does The AI Task Force Do? The AI Task Force key role is to explore possibilities to leverage AI for development across various fields and then submit recommendations to the government, industry and research institutions. “The committee will focus on a wide spectrum of areas including agriculture and transportation… We will concentrate on the types of tools, interoperability, standardisation and skill sets needed for artificial intelligence and machine learning,” said Veezhinathan, speaking to Analytics India Magazine. The next transformation is turning electronics hardware and software intelligent, radically changing their relationships with human ‘wetware’. While there is a strong consensus that AI will be a game-changer and a key factor in economic development, there is a concurrent need to arrive at frameworks that will promote its deployment, taking all social factors into account. In short, the mission of the AI Task Force is to: Leverage AI for economic benefitsCreate policy and legal framework to accelerate deployment of AI technologiesCreate concrete five-year horizon recommendations for specific Government, Industry and Research programs “With rapid development in the fields of information technology and hardware, the world is about to witness a fourth industrial revolution… Driven by the power of big data, high computing capacity, artificial intelligence and analytics, Industry 4.0 aims to digitise the manufacturing sector,” Nirmala Sitharaman, former minister for Commerce and Industry, who had set up the task force, said in August last year. Is There Any Bureaucratic Overlap? We also discussed the seeming overlap of different committees set up by various ministries — for example, a new committee for standardisation in AI was recently set up by the Bureau of Indian Standards, which in turn falls under the Union Ministry of Consumer Affairs. This committee is headed by Pushpak Bhattacharyya, director at Indian Institute of Technology, Patna. “The basic focus of the AI Task Force was to suggest an umbrella document which would cover all the aspects of how India can become a world leader. We, therefore, have made recommendations to all the ministries to apply the directives given in the report to their respective ministries. The Bureau of Indian Standards is one of them,” explained Veezhinathan. Key Takeaways From Report Published By AI Task Force: Data Ombudsman: In a six-part recommendation, Veezhinathan and the task force have hit upon the most crucial thing which is coming in the way of accelerated AI implementation in the country: “The most important challenge in India is to collect, validate, standardise, correlate, archive and distribute AI-relevant data and make it accessible to organisations, people and systems without compromising privacy and ethics.” This point becomes of great relevance especially in the light of the current data security issues that have been unearthed after the Cambridge Analytica and Facebook controversy. The Task Force has also suggested setting up digital data banks, marketplaces and exchanges to ensure the availability of cross-industry information. They have also recommended creating a data ombudsman — like the one available in the banking and insurance sector — to address data-related issues and grievances. The Task Force also talks about incorporating audited data sets to prevent bias and calls for a discussion on the rights and responsibilities of autonomous entities. The report also stresses the need to have more effective data protection policies as a precursor to encouraging any form of data sharing. Education And Awareness: The AI Task Force also recommended the Government to set up a ₹1,200 crore corpus for five years under the Union Budget. They had recommended that the funds be used to push the National Artificial Intelligence Mission (N-AIM). They allocated the funds like this: ₹50 crore PA for core activities, ₹25 crores per Centre of Excellence PA and ₹20 crore PA for generic AI testbed and large data integration centre. Areas Of Focus For AI In India: In the report, the AI Taskforce has also identified 10 specific domains that need attention with respect to AI: ManufacturingFintechHealthcareAgricultureEducationRetail\/Customer engagementPublic utility servicesAid for differently-abled persons\/Accessibility technologyEnvironmentNational security “All these domains need to have supremacy so that is extensively interdisciplinary,” said Professor Veezhinathan. Using AI As An Interdisciplinary Bridge: Veezhinathan, who had just returned home from the office, said, “Computer science has grown to such a large extent because of the demand from non-computer scientists. You can call it one of the most ‘interdisciplinary departments’ in terms of collaboration at any academic institution.” “If you want to do manufacturing, create autonomous vehicles, precision manufacturing, build submarines or ships… There is a lot of computers systems and programming in all of this. Thus computer scientists are always interested in doing this work and this is how artificial intelligence now meshes into various streams,” explained Veezhinathan. AI Task Force’s Take On Employment: As Professor Veezhinathan sounded extremely enthusiastic about the advantages of artificial intelligence, we asked him about the doubts raised by critics. From Elon Musk to NR Narayana Murthy and Shashi Tharoor, many noted personalities have had doubts, especially automation of jobs previously held by humans. Veezhinathan seemed pleased by our question and directed us to a painstakingly-collected case study and data cited in the AI Task Force’s report: “Past trends, wherein new technologies have disrupted existing job market indicate that most jobs which are replaced are mostly tedious, repetitive, labour-intensive and monotonous; while jobs that are less likely to be replaced are those involving human interactions, judgement, and ability to make complex decisions. With a high degree of interdisciplinary nature, AI will certainly not be an exception to this. …Interestingly, in the global arena Gartner has predicted that AI will create 2.3 million new jobs while eliminating only 1.8 million jobs in 2020, suggesting that while some of the jobs that exist today disappear, it is likely that new jobs may appear. A study commissioned by FICCI, NASSCOM and EY has projected that 9% of India’s 600 million estimated workforces would be deployed in new jobs that do not exist today, while 37% would be in jobs that have radically changed skill sets.” Conclusion: To conclude, the report by the AI Task Force is a welcome step in the right direction, which is also in keeping with the current government’s long-term plans. As reported earlier, the Narendra Modi-led BJP Government will strive to achieve its commitment towards Sustainable Development Goals (SDGs) with the help of AI by 2030, as it has the potential to churn out a slew of applications while keeping in mind the quality of approach. Industry insiders are happy to see that the Modi-led NDA government is working towards supporting a tech-driven future. While the effectiveness of this report can only be seen after its recommendations are implemented, it will hopefully jump-start a much-needed discussion on AI policy issues in India.","excerpt":"One of the first things that come to mind while talking to Professor Kamakoti Veezhinathan from the Indian Institute of Technology, Madras (Chennai), is that he is keen on using new tech as a means to an end. His thumb rule is that artificial intelligence should act as a bridge to bring all disciplines together. […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI in India","AI India","AI Task Force","analytics career","IIT Madras","Interviews and Discussions","Machine Learning","Narendra Modi","Startups"],"author_name":"Prajakta Hebbar","publish_date":"2018-04-10T05:01:36","publication_year":"2018","word_count":1378,"keywords":["TPU","AI in India","Ray","analytics career","R","artificial intelligence","RAG","analytics","Narendra Modi","Go","machine learning","AI","IIT Madras","Machine Learning","Startups","AI Task Force","Aim","AI India","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","Ray","RAG","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-task-force-v-kamakoti-report-1200-crore\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10001462,"title":"What Is The Cost Analytics For 3D Printing &#038; Conventional Manufacturing","content":"Modern technology and gadgets not only made our life easy but also fast and comfortable. In the chilly winters of December or on the scorching sun of May, one can enjoy the coolness of spring; thanks to the air conditioner. But what if people can use customised products instead of choosing among AC’s design they make their design in their own smartphone and print like they give print + F command to print docs. To the amazement of creator and innovators, their ideas can become into reality; into 3D or rapid prototyping. Technology has broken the barrier via 3D printing technology. In this manufacturing method object is manufactured from 2D drawing into a solid object by adding 2D slices one over other. Capabilities of 3D printing are certainly limited by human’s thinking capability. Any complex part can be fabricated with as ease as with simple geometries. And that is not the case with current methods of manufacturing. All the methods used today can be grouped as ‘process of manufacturing in which the final product is obtained by subtracting the unnecessary materials’. With the help of the cost analysis of the final product consumer consume, future of manufacturing can be determined. The cost analysis will include everything that a company spend on selling the product in stores. But in order to compare the two manufacturing processes, namely conventional manufacturing, and 3D printing. Analysis of marketing and sales strategies will give be a futile effort. Today, we can manufacture almost all varieties of products using the principle of standardisation, for example, the clothing market offer shirts of size 36,38,40,42 and 44. People with sizes other than these will have to choose size among these only. Let us see the manufacturing cycle a product is undergone through To start manufacturing of a one feet cylinder with a diameter of say 0.25 feet. Industrialist will purchase land to install machines which can weigh for few quintals to thousands of quintals. 3D printing machines can also have massive sizes, but one machine can replace multiple conventional manufacturing process requirements. Conventional machines have rotating parts to cause linear, harmonic or rotatory motion. And these parts cause huge Vibration which needs a foundation to be cemented, to isolate them from each other. Poorly isolated machines have the potential to create an epicentre for an earthquake. Well with 3D printing, though motors are used, due to the requirement of a few machines foundation cost is less. In the case of mass manufacturing robots can be installed to automate the production but in batch production or job type production, an operator is required to operate the machines. And is also required to do inventory management for a low-cost product. In the preview of low-cost manufacturing, mass manufacturing is preferred over other types but in it to bring even a small variation in products is difficult. In conventional manufacturing, a requirement of variation in products results in alteration of processes. And the production of a customised product is limited by its design complexity. The 3D printing machines have the capabilities to overcome all these because of the process it involves manufacturing products. Which are: Make layer by layer deposition of material to form a solid object. The input file can be given by PC\/Laptop. It requires powdered material as its raw material. And the complete product will form instead of disjointed parts. The benefits of 3D printing are less inventory management, waste material though very less can be reused, can easily outsource manufacturing and can economically build custom products. 3D printing is the future of manufacturing because it can take production from specific locations to places near the demand for the products.This article was written in association with Umang Varshney, Associate Data Engineering Analyst, Optum (OGS OPERATIONS), Gurugram, India","excerpt":"Modern technology and gadgets not only made our life easy but also fast and comfortable. In the chilly winters of December or on the scorching sun of May, one can enjoy the coolness of spring; thanks to the air conditioner. But what if people can use customised products instead of choosing among AC’s design they […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2019-03-13T18:25:02","publication_year":"2019","word_count":629,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","data engineering","R"],"extracted_tech_keywords":["AI","R","Go","API","data engineering","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/what-is-the-cost-analytics-for-3d-printing-conventional-manufacturing\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":53040,"title":"Why Is MLOps Gaining Prominence?","content":"MLOps hopes to build automation and improve the nature of ML production while likewise concentrating on business and regulatory necessities. It is estimated that the fourth industrial revolution is largely driven by the AI\/ML capabilities of organisations. According to experts, if you neglect creating data transformation and a model-driven organization, then you are destined to fail. The change in machine learning is bringing greater than that any past disruption has achieved, and there are three central mainstays of AI adoption: data, innovation, and organisational culture and processes. For the ideal adoption of ML across millions of organisations, there requires a standardisation of the machine learning workflows so there is no obstruction to implementation. But, this is not the case currently, as the complex processes involved in ML models makes it very challenging to deploy them. Also, there are a plethora of tools and frameworks to make it more complex. Then, there are many challenges that companies can face when it comes to using ML in the enterprise space such as the ones related to diagnostics, governance, scalability, data compliance, etc. DevOps vs MLOps: Transforming ML Pipelines With DevOps Practices To overcome such challenges, experts say MLOps is on the fast rise to bring efficient collaboration and communication between data scientists and ops teams. MLOps hopes to build automation and improve the nature of ML production while likewise concentrating on business and regulatory necessities. Like the DevOps or DataOps approaches, MLOps also focuses on the entire lifecycle – from continuous integration\/continuous delivery and orchestration to diagnostics and governance. Many of the existing MLOps work on CI\/CD pipeline utilizing a managed machine learning service and cloud-based development services. Unlike using DevOps for app development, MLOps is very different as it is extremely iterative and involves so many languages, libraries, toolkits, and environments for different types of ML models. As part of MLOps, there could be multiple pipelines running in parallel with ML-specific interdependencies that require agile management to ensure a smooth integration. And, then there is specialized hardware involved too, utilizing powerful CPUs and GPUs using frameworks like TensorFlow, Caffe, Apache MXNet, etc. The other thing that makes MLOps different from classic DevOps is the toolchain. MLOps would require not just using things open source repository, automation software, orchestration, and containerisation, but also things that cater to compliance, risk and business goals. MLOps Offerings In 2019 That Signal The Burgeoning Trend As part of MLOps, the integration of ML code, datasets and models into operational procedures becomes incredibly important for organisations. Many vendors have been introducing solutions which cater to the specific integration and develop ML pipelines and workflows. Let’s look at a few vendors and their announcements in 2019 which highlights the burgeoning MLOps market. Cloudera has recently announced an initiative around the operations of the machine learning (ML) workflow. Here, the company is extending the Apache Atlas platform to accommodate the governance of machine learning assets with BlueData’s EPIC software platform. In September 2019, HPE (Hewlett Packard Enterprise) has also announced its foray into the expanding MLOps market with an array of software and tools for managing models. Those tools combine container-based software for AI development and management acquired in a deal, last year for BlueData.The MLOps service is meant to extend the capabilities of which allows enterprises to create Hadoop and Spark clusters in virtual environments. In June 2019, DataRobot acquired ParallelM after its $206 million mega funding and indicates the rising prominence of MLOps. ParallelM- MLOps tools can be used to scale deployment, management and governance of machine learning (ML) pipelines in production. Experts also say there is a trend where we can see the rise of model-as-a-service on cloud marketplaces where companies will be offering multiple ML models based on license base subscription and are charged on the basis of consumption. This way companies won’t have to build ML models from absolute scratch. Open source will also be playing an important role. We can see the increasing use of frameworks like MLflow for MLOps, which is an open-source project is being led by Databricks and built into the Databricks Unified Analytics Platform, which is available on Amazon Web Services and in the Azure Databricks service. Overview While MLOps also began as a set of best practices, it is slowly evolving into an independent approach to ML lifecycle management. Applying DevOps practices to machine learning workloads not only brings models to the market faster but also maintains the quality and integrity of those models. To overcome the prototype phase, smooth, automated and dependable operations have to exist. Thus, we may see much higher adoption in 2020 for MLOps — machine learning operations practices to standardize and make the lifecycle of machine learning in production more efficient.","excerpt":"MLOps hopes to build automation and improve the nature of ML production while likewise concentrating on business and regulatory necessities. It is estimated that the fourth industrial revolution is largely driven by the AI\/ML capabilities of organisations. According to experts, if you neglect creating data transformation and a model-driven organization, then you are destined to […]","categories":["AI Features"],"tags":[],"author_name":"Vishal Chawla","publish_date":"2020-01-03T13:00:00","publication_year":"2020","word_count":788,"keywords":["machine learning","AI","ML","MLOps","Ray","analytics","MLflow","TensorFlow","Azure","Databricks"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","MLOps","MLflow","Ray","TensorFlow","Azure","Databricks"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-mlops-gaining-prominence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10020672,"title":"What Is Geometric Deep Learning","content":"Working with 2D data is becoming passe as more and more researchers tap 3D data to develop AI models. Geometric deep learning, as the field is popularly called, deals with complex data such as graphs to create competitive models. Geometric deep learning, which Michael M. Bronstein first mentioned in the paper titled Geometric deep learning: going beyond Euclidean data, is now finding applications in areas such as 3D object classification, graph analytics, 3D object correspondence, and more. Bronstein’s paper highlighted how research in many scientific fields such as computational social science, sensors network, physics, and healthcare — especially brain imaging — calls for exploring non-Euclidean data. Deep learning has applications in computer vision, natural language processing and audio analysis, requiring Euclidean or 2D data. To facilitate working with 3D data, researchers are exploring Geometric deep learning, an umbrella term for emerging techniques used to generalise (structured) deep neural models to non-Euclidean domains such as graphs and manifolds. Surpassing Deep Learning Methods The current deep learning algorithms such as Recurrent Neural Networks (RNN), Convolutional Neural Networks (CNN), and LSTM have seen tremendous growth in the last few years tackling problems in speech recognition, computer vision, image generation, language transition and more. Most of these deep learning algorithms work on Euclidean (1D or 2D) data. The researchers believe tapping 3D data will improve the accuracy of findings by leaps and bounds. One of the challenges with traditional deep neural networks is that they cannot parse data. Also, most of these networks are based on convolutions, and convolution works better on Euclidean data. Especially in areas such as network science, physics, biology, computer graphics and recommender systems, researchers have to deal with non-Euclidean data such as manifolds and graphs — which cannot fit in the two-dimensional space. For instance, graphic specialisation or mesh in the computer graphics field is non-Euclidean data. Non-euclidean data can represent more complex data compared to 1D and 2D representation. Researchers believe, since the real world manifests in 3D, the data should reflect that. To make machine learning and deep learning achieve human-level efficiencies, researchers are now exploring the use of 3D data. Non-Euclidean Data Types Graphs are one of the most prominent examples of a non-Euclidean datatype. Graphs consist of nodes connected with edges and can be used to model almost anything. For instance, considering social networks as graphs, each user is a node, and their interactions with other users are edges. Researchers are now modelling sensors and computer networks as graphs, where each signal and communication represent vertices in the graph. Manifolds are another example of non-Euclidean data involving a vast number of geometric surfaces such as curvy, twisty or other diverse 3D shapes. Manifolds are essentially multi-dimensional spaces made of multiple points. Manifold data can come from a variety of sources, such as images or other numerical values. Geometric deep learning is explored in areas such as molecular modelling, 3D modelling and more and can address bottlenecks in computational chemistry, biology, physics. For instance, in a use case around COVID-19, researchers used Knowledge Graph Convolutional Network (KGCN) for Relation prediction. This KGCN framework is designed to provide a versatile means to perform learning tasks over a Grakn knowledge graph. Researchers used patient inputs to gather ground truth graph data. The ground truth data was used to learn to predict relations for new patients. Geometric deep learning also has potential applications in drug discovery as molecules can naturally be represented as graphs, where atoms are nodes and bonds are edges.","excerpt":"Working with 2D data is becoming passe as more and more researchers tap 3D data to develop AI models. Geometric deep learning, as the field is popularly called, deals with complex data such as graphs to create competitive models. Geometric deep learning, which Michael M. Bronstein first mentioned in the paper titled Geometric deep learning: […]","categories":["AI Features"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2021-02-23T14:00:00","publication_year":"2021","word_count":583,"keywords":["Go","machine learning","AI","neural network","computer vision","deep learning","analytics","RNN","CNN","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","analytics","R","Go","CNN","RNN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-is-geometric-deep-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10010303,"title":"The Dearth Of AI Teachers &#038; How It Can Be Mitigated","content":"As per the Data Science Skills Study 2020, more than 10% of the machine learning and data science practitioners learn from various online sources, such as massive online open courses or MOOCs, online certifications and courses, online videos hosted on such platforms as well as LinkedIn and YouTube, among others. On the other hand, traditional formats like university certifications and courses are at the lower end of the spectrum of preference, which is 5.7%. The one main reason behind this is the dearth of AI teachers among institutions and academia. To get an industry perspective on this, Analytics India Magazine caught up with a few experts in this field who explained the reasons behind the void and helped in understanding how these issues can be addressed. The Need of AI Teachers To Teach The Intricacies of AI & ML The use of artificial intelligence-based solutions has been proliferating in everyday life, starting from the shopping experience to financial transactions. Rabindranath A, founder of LeapCurve shares that the need of AI-based solutions is going to grow and it is high time that it is introduced in the school curriculum as AI is estimated to boost India’s annual growth rate by 2035. He added that while the government has some AI-based education policies in place and many state governments are trying to introduce AI as a subject, there is indeed a dearth of good teachers at this moment. Those who have studied AI and are working on it are not comfortable teaching at schools as the pay is abysmally less. So the government has to take into cognisance such issues and come up with a solution. Gaurav Vohra, CEO and Co-founder, Jigsaw Academy said, “AI, like many other emerging technologies, faces a severe shortage of teachers because the field is rapidly evolving and academia can’t keep pace with it. Academics are great at explaining the fundamentals, but when it comes to a business application, practitioners are needed. This is why Jigsaw Academy’s programs on AI have a healthy mix of academicians and business practitioners. A majority of Indian institutions aim to deliver world-class education in AI\/ML, analytics, and cloud computing, but their lean infrastructure and lack of hands-on, qualified faculty make the ground reality very different.” One of the significant motivations to integrate AI in the school curriculum is to make the forthcoming age acquainted with the rising innovation as this age is utilising various applications of the AI in reality. “Teachers in the existing system are not qualified to teach AI. So, teacher training needs to be ensured before AI curriculum is brought in, as this will create the foundation for the future. At the point when we talk about educating, whatever concepts being presented in the schools or college education programs are fundamentals and foundations for future, if the educators aren’t sufficiently gifted to make AI straightforward for students, it will hinder further development,” said Dileep Kumar Singh, Head, JLU School of Engineering & Technology. Aniruddha Kannal, Academic Council Member of Atria University, believes that it’s not easy to find educators who understand, let alone teach the intricacies of AI. Institutions can broadly be divided into two groups. Ones that are focused on teaching the tools of AI\/ML and how they can be applied to solve specific problems. And, others that are also teaching the underlying mathematics (probability, statistics, linear algebra, etc.). On the contrary, Joydeep Paul, Advisory Consultant at IBM, stated that beginnings are always challenging, but in the age of internet and video learning, there are abundant resources to refer. Schools can arrange regular learning workshops for their educators to help them achieve expertise. Difficulties To Recruit AI Educators To this, Kannal replied that there is a massive skill gap, and it is hard enough to find competent trainers. AI just adds another layer of complexity. He pointed out a few reasons as to why it is difficult: There’s a massive skill gap in the industry as a result of which the best talent is paid top dollars. And, the educational institutions can’t afford to compete with the tech industry pay scales.One cannot teach AI with whiteboards and PowerPoint presentations. Our educators are not used to getting their hands dirty.The domain is in flux. Staying on top of what’s happening is difficult, even for the best. Singh stated, “The educational industry experiences the gap most prominently. The main cause is that the technological use was not fully utilised and put into practice. The difficulty in recruiting more trainers rises due to the various factors. Candidates are not trained with sufficient practical exposure in their respective fields. The condition becomes more disappointing in the field of computer science. Furthermore, we are comfortable in continuing to use old technology rather than looking further at the new and updated ones just because we have a stereotyped thought about new things.” “In my opinion, there is definitely a skill gap amongst the tech-trainers as most Indian techies are not trained to think on how to spot a problem. Until and unless they know what the problem is, how will they create a solution around it? Besides, most techies are not well equipped or have those required skill sets to teach or train young kids. The good ones are difficult to recruit as they are very few in number. There is a problem of demand and supply here,” said Rabindranath. Overcoming the Dearth of AI According to Ashutosh Kumar, CEO and co-founder of Testbook.com, the best way to overcome the dearth is by industry-academia collaborations by getting more and more people exposed, trained and educated in the concepts of this new and emerging technology. To overcome the shortage of educators, it is crucial to give them some real-time practical experience in the field by providing opportunities in organisations that have adopted the technology. Another way is to introduce refresher courses for teachers and professors in the field. Singh explained that in order to fulfil this dearth, tech companies should change their relationship with the academic community. By making the collaborative efforts with academic institutions, the tech companies who are working with AI can train the professors or researchers of any institute in their own way. He further added that after getting trained from the industry, the professor or researcher could then continue their training as an expert in the field of AI. Through these types of collaborations, universities continue to benefit from the brightest minds in AI and data science, while businesses get their in-house expertise and a pipeline for AI talent through internships. “We must encourage and educate individuals from ground level along with building infrastructure. Arranging seminars to display AI-enabled product prototypes can also help individuals to visualise and encourage them to start a career in AI,” shares Paul. Initiatives to Meet This Shortage Singh stated that educational institutes might take some of the initiatives to meet this shortage such as teachers can go through various development programs designed to get trained in the field of AI, a collaboration between industries and academia as well as sponsored and joint R&D projects from industries. According to Kannal, institutions are facing shortages on multiple fronts, including infrastructure, talent, content, and pedagogy. The AI industry is playing a huge role in evangelising as well as educating. Never before has a community been more involved in creating free content in the form of videos, manuals and tutorials, code snippets, GitHub repos, projects, etc. He added, “I believe the industry is already doing its bit. It is now up to the educational institutions to tap into all these resources and make good use of them.” “Many BTech institutions today are looking to hire freelancers or part-time teachers who are working as a full time AI professional in major IT companies across India. If we look at the adoption rate of artificial intelligence and technology by the education sector, we have been adapting very rapidly to compensate for the demand-supply gap of teachers and students in India,” shares Ashutosh. Rashi Thakur, Founder of Mentor Mpact, said, “Currently, unskilled teachers are assigned tasks of teaching multiple subjects even if they lack the knowledge or requisite training to teach certain subjects. AI and tech-based companies can develop platforms to provide training support for teachers which can be completed with ease even from the comfort of their homes. Such platforms can curate lessons and provide learning and feedback to students, which are customised as per the student’s ease in grasping concepts.” Current Percentage of AI Educators in India Singh said on a concluded note that although it’s very difficult to put the exact percentage but according to him roughly 10-15% of the total computer science and engineering educators may have the sound knowledge of artificial intelligence. He added, “Talking in terms of the teacher-student ratio in the educational institutes in India, according to me for approx 60 students there might be only one AI educator available with good knowledge.”","excerpt":"As per the Data Science Skills Study 2020, more than 10% of the machine learning and data science practitioners learn from various online sources, such as massive online open courses or MOOCs, online certifications and courses, online videos hosted on such platforms as well as LinkedIn and YouTube, among others.  On the other hand, traditional […]","categories":["AI Features"],"tags":["AI Certifications","collaboration ai platform","PowerBI"],"author_name":"Ambika Choudhury","publish_date":"2020-10-24T10:00:00","publication_year":"2020","word_count":1489,"keywords":["data science","artificial intelligence","machine learning","AI","cloud computing","ML","AI Certifications","RAG","PowerBI","Aim","collaboration ai platform","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","RAG","cloud computing","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-dearth-of-ai-teachers-how-it-can-be-mitigated\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10131603,"title":"7 Midjourney 6.1-Generated Creatives that will Blow Your Mind","content":"Midjourney has consistently set the bar for advancements in video creation and other forms of creatives. The recent release of the Midjourney 6.1 model has once again taken the internet by storm. Shortly after its launch, the news spread on X, highlighting it as the most advanced AI art model to date. Midjourney 6.1 brings new updates, enhanced image coherence, new 2x upscalers boost image and texture quality, with processing speeds about 25% faster. Moreover, the text accuracy in prompts is improved, and a new personalisation model adds more nuance, surprise, and accuracy to the generated content. Midjourney released a new update to its AI image generator with the new V6.1Upgrades include improved image quality, coherence, and text rendering, along with new upscaling and personalization models https:\/\/t.co\/O26LKfdcHF— Brett Adcock (@adcock_brett) August 4, 2024 Here are seven examples of Midjourney 6.1 generated creatives: Midjourney 6.1 + Runway 3 Combo A user tried Midjourney V6.1 and Gen-3 i2v with the prompt “a girl laughs as a fire swirls far away behind her”. Runway handled the fire well, but ‘far away’ was added to avoid flagging. Despite a bit of hair being caught, the image is expressive, aligns with their style preferences, and the personalisation parameter (–p) offers more customisation than V6.” https:\/\/twitter.com\/LeoEspinozaCine\/status\/1818780361838432547 Updated Personalisation Version 6.1 is now more personalised, offering enhanced nuance, surprise, and accuracy. This system allows for quicker and more precise adjustments. Users can continue using their current personalisation style with the–p code, and previous personalisation codes remain compatible. As users rank more images, the system will automatically evolve to reflect these updates. For example, a user on X provided the code “night”, and the system generated an image of a nighttime scene. Another example is the code “dream”, which produced this image. Creation of cinematic scenes Several cinematic scenes were created using Runway ML Gen-3’s image-to-video feature. The images were generated with Midjourney 6.1, based on specific prompts. For example, one of the images was prompted to depict the subject with a “seductive smile”. https:\/\/twitter.com\/jamesyeung18\/status\/1819034132158242831 Midjourney + Luma Dream Machine A user on X, created a video with the image generated with Midjourney v6.1 and Dream Machine. Sharing the experience, they said, “Taking this as the base image and popping it into Dream Machine is so next-level. 2024 truly is the year for AI video. I’m exploring a few Greek myths currently, so cool seeing all this come to life.” Speechless. Midjourney v 6.1 knocked it out of the park. Taking this as the base image and popping it into Dream Machine is so next-level…2024 truly is the year for AI video. I'm exploring a few Greek myths currently, so cool seeing this all come to life. Blown away lately. pic.twitter.com\/7vPTSC0720— Kiri (@Kyrannio) July 31, 2024 Creating Abstract images A user on X highlighted that Midjourney excels at generating abstract images with outstanding aesthetics. They shared two prompts: one for a high-chaos, stylised optical fibre image (optical fibre –chaos 10 –ar 16:9 –sref 1558173677 4244977961 1214440399 –p –sw 800 –stylize 800 –weird 3 –v 6.1) another for optical fibre integrated into a landscape (optical fiber along a landscape –sw 25 –sref 531168650 2959860992 –ar 7:3 –v 6.1). Both prompts are crafted to produce visually striking images. Optical fiber.Only Midjourney can generate abstract images with exceptional aesthetics and qualityPrompt:optical fiber –chaos 10 –ar 16:9 –sref 1558173677 4244977961 1214440399 –p –sw 800 –stylize 800 –weird 3 –v 6.1 pic.twitter.com\/TUU6LUvzqJ— Tatiana Tsiguleva (@ciguleva) July 30, 2024 Blending SREF Codes with Midjourney A user successfully integrated SREF Codes (parameters that guide the AI in producing specific visual outputs) with Midjourney v6.1 to create synthetic life forms using specific parameters. The process involved: Step 1: Generate a movie still with –sref random, –sw 1000, and –s 0. Step 2: Evaluate satisfactory images and run additional tests (product and street photography) with –sref [CODE], –sw 100, and –s 300. Step 3: Select the best codes from Step 2 and test with complex, multi-prompts for optimal results. How I test Midjourney SREF codes. Updated version.The parameters are more important than the prompt for the first test.Step 1: movie still –sref random –sw 1000 –s 0 Step 2: If I see images that look ok with –sw 1000 –s 0 I run a few other tests: {product… pic.twitter.com\/h0kC9lesYv— Tatiana Tsiguleva (@ciguleva) June 19, 2024 2x Upscaling In Midjourney 6.1, users have different preferences for upscalers. The Creative upscaler is widely appreciated for its improved artifact removal and overall rendering quality. However, some users, particularly those working with images that have heavy textures like film grain and paper overlays, prefer the subtle upscaler, as it preserves these textures better and avoids the more artificial look that creative can introduce. Additionally, version 6.1 is faster than 6.0, has fewer artifacts, and offers a notable improvement in text integration within generated images. https:\/\/twitter.com\/MacPaw\/status\/1820463355707994543","excerpt":"Midjourney 6.1 boosts image coherence with new 2x upscalers, improved texture quality, and 25% faster processing speeds.","categories":["AI Trends"],"tags":["MidJourney","personalisation"],"author_name":"Gopika Raj","publish_date":"2024-08-06T16:36:03","publication_year":"2024","word_count":804,"keywords":["MidJourney","TPU","programming_languages:R","AI","ML","AI art","personalisation","R"],"extracted_tech_keywords":["AI","ML","TPU","R","AI art","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-midjourney-6-1-generated-creatives-that-will-blow-your-mind\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10057740,"title":"Now IBM wants to speed up automated transfer learning","content":"IBM introduced CodeFlare at the Ray Summit in June of 2021. The platform was introduced to drastically reduce the time required to set up, run, and scale machine-learning tests. For example, CodeFlare reduced the time to execute each pipeline from 4 hours to 15 minutes when one user used the framework to examine and improve approximately 100,000 pipelines for training machine learning mode. Recently, IBM announced that CodeFlare significantly reduces the time to automate transfer learning tasks for foundation models. What is CodeFlare? CodeFlare is a hybrid multi-cloud platform that streamlines the integration, scalability, and acceleration of complicated multi-step analytics and machine learning pipelines. It is an open-source framework that makes it easier to integrate and scale big data and AI operations to the hybrid cloud. CodeFlare is developed on Ray, an open-source distributed computing framework for machine learning applications. Ray’s capabilities are expanded by CodeFlare, which adds specialised aspects that make scaling operations easier. CodeFlare uses Python and runs on IBM’s new serverless platform, IBM Cloud Code Engine and Red Hat OpenShift. Users can deploy the platform from just about anywhere, allowing researchers to reap the benefits of serverless computing. The platform also provides rich APIs and tools, allowing researchers to focus more on their research and less on configuration complexities. Also read: IBM Unveils New Data Fabric Capabilities & Advanced Data Privacy Features For IBM Cloud Satellite here. How does CodeFlare speed up the automation of transfer learning tasks for foundation models? The newly enhanced and upgraded CodeFlare effectively transforms it from a data science exploratory tool to a platform that can automate AI and machine learning workflows on IBM’s hybrid cloud. Businesses use foundation models for a variety of functions. A financial services firm, for example, may create a foundation model specifically for sentiment research. Currently, gathering and training an AI model on a relevant body of data and various upstream and downstream activities takes an incredible amount of time. CodeFlare speeds up the process by using a Python-based interface or foundational model pipelines, making it easier to integrate, duplicate, and share data. Moreover, on a hybrid cloud platform, these tasks of preprocessing, validating, and adapting foundation models for commercial use cases are now entirely automated. For example, take the instance of sentiment analysis. CodeFlare begins by cleaning up the input data, which includes de-duplication and the removal of potentially harmful or biassed content. Then, it fine-tunes a foundation model for all the specific activities required for the sentiment analysis of the company. A data scientist may operationalise hundreds of such pipelines with only a few lines of code and automate these operations anytime they need to make changes. CodeFlare allows data scientists to use their own data without leaving the hybrid platform. Hybrid cloud is an essential part of IBM’s growth strategy. IBM’s total cloud revenue increased by 20% in the fiscal year 2020, thanks in part to hybrid cloud programmes supplied by IBM-owned Red Hat. Also read: IBM’s Strategy For Hybrid Cloud Growth In India here.","excerpt":"This article looks into how IBM’s CodeFlare significantly speeds up automated transfer learning tasks for foundation models","categories":["AI Features"],"tags":["IBM"],"author_name":"Abhishree Choudhary","publish_date":"2022-01-06T14:00:00","publication_year":"2022","word_count":501,"keywords":["data science","machine learning","AI","sentiment analysis","ML","distributed computing","serverless","Ray","IBM","analytics","foundation models"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","foundation models","Ray","sentiment analysis","serverless","distributed computing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/now-ibm-wants-to-speed-up-automated-transfer-learning\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10124775,"title":"Pixxel to Manufacture Miniaturised Satellites for Indian Air Force Under iDEX Grant","content":"Bengaluru-based space technology startup Pixxel has signed the 350th contract under the iDEX (Innovations for Defence Excellence) program to manufacture miniaturised multi-payload satellites for the Indian Air Force. The contract, awarded as part of the iDEX Prime Space grant, marks a significant milestone in Pixxel’s mission to revolutionise the space industry in India. The contract was signed between Awais Ahmed, CEO of Pixxel, and Anurag Bajpai, Additional Secretary (Defence Production) and CEO of IDEX-DIO, in the presence of Defence Secretary Giridhar Aramane, the Vice Chiefs of the Armed Forces, and other officials of the Ministry of Defence. “We are delighted to receive iDEX’s grant and utilise our expertise of building microsatellites in-house to manufacture satellites externally for the first time,” said Ahmed. “This recognition highlights Pixxel’s dedication to pushing the boundaries of space exploration and innovation.” Under the multi-crore contract, Pixxel will develop small satellites weighing up to 150 kg for electro-optical, infrared, synthetic aperture radar, and hyperspectral applications. The company will leverage its indigenous hyperspectral satellite technology and manufacturing expertise to build these satellites, enabling ease of manufacture, low cost, and ease of launch. As Pixxel sets out to launch six commercial-grade hyperspectral satellites, ‘Fireflies’, this year, the company remains committed to harnessing its indigenous expertise and the power of hyperspectral satellites for a sustainable future. Building on its expertise, Pixxel now offers high-performance, cost-effective satellite manufacturing solutions, empowering clients to drive meaningful change with space data.","excerpt":"Under the contract, Pixxel will develop small satellites weighing 150 kg for electro-optical, infrared, SAR & hyperspectral applications.","categories":["Deep Tech"],"tags":["AI in manufacturing","Pixxel"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-25T19:15:41","publication_year":"2024","word_count":238,"keywords":["Pixxel","programming_languages:R","AI","innovation","RAG","AI in manufacturing","R","startup"],"extracted_tech_keywords":["AI","RAG","R","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/pixxel-to-manufacture-miniaturised-satellites-for-indian-air-force-under-idex-grant\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10020584,"title":"Do Good Analytics Leaders Have To Be Expert Data Scientists?","content":"Data science is a highly multidisciplinary field. And to run data teams profitably, organisations need versatile leaders and managers. While analytics leaders’ job entails managing resources and a lot more, data scientists deal with technical aspects like coding, modelling, data collection, among others. We caught up with a few industry professionals to understand if analytics leaders should have an in-depth knowledge of data science. The short answer is: Yes and no. For the long answer, read on. Yes And No There are several drawbacks to analytics leaders not being data scientists. Since data science engagements are hypothesis-driven experiments with different ways to arrive at the right solution, good knowledge of data science will enable the leader to guide the team in the right direction. “The hands-on managers bring subject expertise and have been in those shoes before, which is effective,” said Srikanth Velamakanni, Co-Founder, Group Chief Executive & Executive Vice-Chairman of Fractal Analytics. “Especially in data science, you observe that people who have done this before have very strong empathy for how it is being done currently, and the junior members also feel that there is a lot of guidance coming in. This helps in avoiding mistakes. “Also, if you have a manager who can point their team in the right direction, they can really save a lot of time. They can ask them to, for instance, not bother looking at ten different things, but take this particular approach, because they know this is where things are,” he added. On the other hand, industry professionals also think it is not essential for analytics leaders to have in-depth technical knowledge. “All three roles are focused and complement each other. CXOs and Analytics Leaders are expected to focus more on the business value, effective program and data governance, not on the technical skillset,” said Prashant Pansare, Founder and CEO of Rubiscape. “We define the Analytics Leaders persona as the orchestrators of the various stages of the Analytics Program, running multiple projects on a diverse set of technologies producing a harmonious outcome which forms the basis of fact-driven decisions. Tool level knowledge is an advantage, not a critical factor,” said Pansare. Tradeoffs Velamakanni asked to consider four quadrants with scores from low to high, with both responsibilities on either axis. “The high-high quadrant, where managers are high on managerial as well as analytical skills, present the best talent but are very rare to find. And the low-low ones should not be in the system,” said Velamakanni. The other two quadrants also present uncertainties. “If the leaders are low on management capability and very high on their analytics understanding, they are not going to be good with managing resources. But if they are high on management capability and low on analytics, it could be somewhat dangerous if they get into too much micromanagement.” If you are high on managerial skills but low on analytics, then you want to let the team be. But if you are high on analytics skills and low on managerial skills, then you want your team to be more of self-starters,” he added. Essentials One challenge with pure data scientists is their weaknesses to either look at the bigger problem or explain and visualise the output in simple terms. “First and foremost, they need to understand the business goals and the environmental factors well – markets, trends, people, and products, which are more than just a technology,” said Kedar Sabne, Co-founder and Director at Rubiscape. “They should also know the people, process & profitable performance aspects of analytics programs. They should be able to deliver a measurable business value and impact on the key stakeholders.” The other important factor is the ability to spot the right talent. “What is happening today is that there is a vast difference between the average talent and game-changing talent. Game-changing talent is at least five times as good as the average talent. So, the ability to spot and pick the right talent needs to be very good,” said Velamakanni. “Secondly, as a manager, you have to understand how work gets delivered by understanding aspects like time required, conflict resolution, or the quantum of work. They should not come across as too unreasonable or too relaxed and have a good sense of estimation of the effort required.” The experts also pointed out that the ability to understand diverse data, roles, and mindsets, along with excellent communication, visualisation and business-facing skills, are more important than the core data science skills. Finally, the analytics leaders and data scientists must be on the same page. To ensure that, the managers should be able to set the team composition correctly. It is important to ensure the data scientist appreciates the inputs and understands the leader’s value in the engagement. Industry professionals think there is a tendency for leaders only to sit outside and review the progress. In such cases, you might not get the level of cooperation or final outputs as desired. Upskilling Finally, the industry professionals suggested several ways leaders could upskill themselves, technically and managerially, to be a better analytics leader. “One of the most important ways is to read books,” said Velamakanni, “There are a lot of good management books like Peter Drucker, Marshall Goldsmith, or Stephen R Covey to smoothen some of your rough edges. What also helps is having a mentor in people who have done this before to get guidance on challenges that you are facing.” “On the technical front, one should be enthusiastic about taking up challenges. Another way is to take up teaching as you will learn a lot of things in the process. One could also start writing. They say that writing is a process of discovering how little you know.” Courses from leading institutions through Coursera, Udemy, etc., are widely available and can help in the initial upskilling phase. Participation in online challenges and forums like Topcoder and Kaggle can further help understand and cement their knowledge and understanding.","excerpt":"Data science is a highly multidisciplinary field. And to run data teams profitably, organisations need versatile leaders and managers. While analytics leaders’ job entails managing resources and a lot more, data scientists deal with technical aspects like coding, modelling, data collection, among others. We caught up with a few industry professionals to understand if analytics […]","categories":["AI Features"],"tags":[],"author_name":"Kashyap Raibagi","publish_date":"2021-02-22T10:00:00","publication_year":"2021","word_count":990,"keywords":["data science","Go","TPU","programming_languages:R","AI","RAG","analytics","data governance","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","RAG","TPU","R","Go","data governance","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/do-good-analytics-leaders-have-to-be-expert-data-scientists\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10023109,"title":"Is Explainability In AI Always Necessary?","content":"“AI models do not need to be interpretable to be useful.”Nigam Shah, Stanford Interpretability in machine learning goes back to the 1990s when it was neither referred to as “interpretability” nor “explainability”. Interpretable and explainable machine learning techniques emerged from the need to design intelligible machine learning systems and understand and explain predictions made by opaque models like deep neural networks. In general, the ML community is yet to agree on a definition for explainability or interpretability. Sometimes it is even called understandability. Some define interpretability as “the ability to explain or to present in understandable terms to a human”. According to experts, interpretability depends on the domain of application and the target audience. Therefore, a one-size-fits-all definition might be infeasible or unnecessary. When concepts are used interchangeably, would it be wise to sacrifice the usability of a model for lack of comprehension? Where does one draw the line? Despite deep learning’s popularity, many organisations are still comfortable using logistic regression, support vector machines and other conventional methods. Though model agnostic techniques can be used for traditional models, they are considered overkill for explaining kernel-based ML models. Model-agnostic methods can be computationally expensive and can lead to poorly approximated explanations. Stanford’s Nigam Shah, in a recent interview, touched on why explainability may not always be necessary. “We don’t fully know how most of them really work. But we still use them because we have convinced ourselves via randomized control trials that they are beneficial,” said Shah. Explainability In Its Many Forms Image credits: Stanford HAI blog For any organisation, explainability becomes an issue when clients or other stakeholders come into the picture. The stakeholders fall into two categories: One where explanations can be used as a one-off sanity check or shown to other stakeholders as reasoning for a particular prediction. Explanations that can be used to garner feedback from the stakeholder regarding how the model ought to be updated to better align with their intuition. It is generally believed that explainable methodologies can have broader advantages as they can be communicated to a wider audience and not just the immediate stakeholders. These methodologies help share the insights across the organisation without the need for a specialist in every scenario. According to Shah, there are three main types of AI interpretability: Explainability that focuses on how a model works.Causal explainability deals with the “whys and hows” of the model input and output.Trust-inducing explainability provides the information required to trust a model and confidently deploy it. So, it is important to know what type of explainability a data science team is targeting. That said, there is a chance that a use case might be a mix of all three. Such trade-offs and overlaps present a bundle of paradoxes to a decision-maker. With increasing sophistication and completeness, the system becomes less understandable. “As a model grows more realistic, it becomes more difficult to understand,” said David Hauser at the recently concluded machine learning developers conference. According to Hauser, clients want the model to be understandable and realistic.This is another paradox a data scientist has to live with. He also stressed that understandable solutions give up on accuracy. For instance, network pruning one such technique which takes a hit on accuracy. The moment non-linearities or interactions are introduced, the answers become less intuitive. “Do you, as a user, care how the weather is predicted, and what the causal explanation is, as long as you know a day ahead if it is going to rain and the forecast is correct?” We live in a world of an abundance of tools and services. Making the right choice leads to another paradox– Fredkin’s paradox, which states the more two alternatives seem similar, the harder it is to choose and the more time\/effort required to decide. Stanford professor Shah has also emphasised the Trust paradox. According to him, explanations aren’t always necessary. What can be worse is, sometimes they lead people to rely on a model even when it’s wrong. According to Shah, what engineers need from interpretability might not coincide with those of the model users whose focus is around causality and trust. Furthermore, explanations can also dent the chances of knowing what one really needs. Key Takeaways In his interview with Stanford HAI, Shah shared: AI models do not need to be interpretable to be useful.Doctors at Stanford prescribe drugs on a routine basis, without fully knowing how most of them really work.In health care, where AI models rarely lead to such automated decision making, an explanation may or may not be useful.If it is too late to intervene for the clinician, what good are the explanations?But, AI for job interviews, bail, loans, health care programs or housing, absolutely require a causal explanation. One of the vital purposes of explanations is to improve ML engineers’ understanding of their models to refine and improve performance. Since machine learning models are “dual-use”, explanations or other tools could enable malicious users to increase capabilities and performance of undesirable systems. There is no denying that explanations allow model refinement. And, as we go forward, apart from the debugging and auditing of the models, organisations are looking at data privacy through the lens of explainability. Medical diagnosis or credit card risk estimation, making models more explainable, cannot come at the cost of privacy. Thus, sensitive information is another hurdle for explainability.","excerpt":"“AI models do not need to be interpretable to be useful.” Nigam Shah, Stanford Interpretability in machine learning goes back to the 1990s when it was neither referred to as “interpretability” nor “explainability”. Interpretable and explainable machine learning techniques emerged from the need to design intelligible machine learning systems and understand and explain predictions made […]","categories":["Deep Tech"],"tags":[],"author_name":"Ram Sagar","publish_date":"2021-04-01T11:00:00","publication_year":"2021","word_count":890,"keywords":["data science","Go","machine learning","TPU","AI","neural network","ML","deep learning","Rust","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","data science","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/is-explainability-in-ai-always-necessary\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10048151,"title":"Underrated But Fascinating ML Concepts #5 &#8211; CST, PBWM, SARSA, &#038; Sammon Mapping","content":"There are a few exciting machine learning concepts that do not receive nearly enough attention. Let’s look at Constructing skill trees, Prefrontal cortex basal ganglia working memory, State–action–reward–state–action, and Sammon mapping. Constructing Skill Trees Constructing skill trees (CST) is a hierarchical reinforcement learning technique that can create skill trees from a series of example solution trajectories gathered through demonstration. CST segments each demonstration route into skills and integrates the results into a skill tree using an incremental MAP (maximum a posteriori) change point detection technique. It uses a changepoint detection algorithm to partition each trajectory into a skill chain by recognising a suitable abstraction change or a segment too complex to characterise as a single skill. Each trajectory’s skill chains are then combined to build a skill tree. CST is a considerably faster algorithm than skill chaining for learning. Even a failure can help you enhance your talent. Agent-centric features can be utilised to learn skills that can be applied to different issues. In the PinBall area, CST has been utilised to learn abilities through human demonstration. On a mobile manipulator, it has also been used to learn skills from human demonstration. Prefrontal Cortex Basal Ganglia Working Memory The algorithm prefrontal cortex basal ganglia working memory (PBWM) represents working memory in the prefrontal cortex and the basal ganglia, according to a research study. In terms of functioning, it’s similar to long short-term memory (LSTM) but is more biologically definable. PBWM was inspired by LSTM and provides flexible memory regulation but was built with a heavy focus on biological plausibility. Sensory stimuli are only allowed into PBWM’s Working Memory (WM) shop in an all-or-none approach. However, researchers have pointed out that the exact functionality of PBWM is masked by the fact that it is a complex model with a highly interwoven architecture of a variety of neural subsystems and several parallel learning algorithms, both supervised and unsupervised. A few years back, a group of researchers proposed a simpler PBWM model that concentrates on just one key component of the technique: the employment of internal gating events to govern memory content. This model substitutes a more abstract tabular representation of all possible input and memory states for all physiologically based neural subcomponents. The simplified PBWM model thus ignores most of PBWM’s biological realism, but it does highlight its basic capability. It is the control over memory content by internal gating actions, which can be learned through reinforcement learning alone. State–action–reward–state–action The state–action–reward–state–action (SARSA) method is a reinforcement learning approach for learning a Markov decision process policy. The SARSA algorithm is a slightly modified version of the well-known q-Learning algorithm. In any reinforcement learning algorithm, a learning agent’s policy can be one of two types:  on-policy and off-policy. The greedy strategy is used to learn the q-value in the q-Learning technique, which is an off-policy technique. The SARSA approach, on the other hand, is an on-policy that learns the q-value from the present policy’s activity. The most significant distinction between SARSA and q-learning is that the greatest reward for the following state is not always used to update the q-values. Instead, the same policy that decides the initial action is used to select a new action and therefore reward. SARSA gets its name from the fact that it uses the quintuple Q(s, a, r, s’, a’) to perform updates. Where s and a represent the initial state and action, r represents the reward observed in the next state, and s’ and a’ represent the subsequent state-activity combination. Sammon Mapping Sammon mapping, also known as Sammon projection, is an algorithm for mapping a high-dimensional space to a lower-dimensional space while attempting to preserve the structure of inter-point distances in the higher-dimensional region. It’s especially well-suited to exploratory data analysis. According to a study, unlike principal component analysis, the mapping cannot be expressed as a linear combination of the original variables, making it more difficult to employ for classification purposes.Since its introduction in 1969, the Sammon mapping has been one of the most successful nonlinear metric multidimensional scaling methods; however, efforts have been focused on algorithm improvements rather than the form of the stress function. Through the use of left and right Bregman divergence, the Sammon mapping’s performance has been improved.","excerpt":"As part of this series, we’ll review several fascinating yet underestimated machine learning concepts.","categories":["IT Services"],"tags":["Machine Learning","Principal Component Analysis","Reinforcement Learning"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-09-13T12:00:00","publication_year":"2021","word_count":706,"keywords":["Go","LSTM","machine learning","Reinforcement Learning","programming_languages:R","AI","Machine Learning","programming_languages:Go","Principal Component Analysis","ViT","GAN","R"],"extracted_tech_keywords":["AI","machine learning","R","Go","GAN","LSTM","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/underrated-but-fascinating-ml-concepts-5-cst-pbwm-sarsa-sammon-mapping\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10161436,"title":"Soket AI Labs Launches Realtime Speech API, Offers Multilingual Capabilities","content":"Soket AI Labs, the Gurugram-based AI startup, has unveiled its Realtime Speech API, aiming to transform AI interactions with voice intelligence and seamless integration. Soket AI Labs claims that the Realtime Speech API boasts ultra-low latency of under 500 milliseconds, ensuring near-instantaneous responses for real-time interactions. It supports multilingual capabilities to overcome language barriers and includes advanced functionalities such as tool calling, Retrieval-Augmented Generation (RAG) support, custom voice creation and cloning, and the ability to handle dynamic voice interruptions for natural conversations. Developers can integrate the API effortlessly within 1-4 weeks using SDKs available for Python and JavaScript. The service is priced competitively at $0.012 per minute, providing an affordable alternative to industry leaders like OpenAI. Soket AI Labs emphasised the platform’s versatility, highlighting its applications across industries such as banking, financial services, insurance (BFSI), healthcare, and telecommunications. Additional features include fine-tunable models and customisable voice options to meet specific business needs. The company is set to launch its “Voice Innovators Beta Program” soon, inviting users to explore and shape the future of voice technology. In a separate post on LinkedIn, Abhishek Upperwal, founder and CEO of Soket AI Labs, highlighted the importance of making ‘General Voice Intelligence’. “Voice is one of the most important interfaces to AI today and language models are at the core of intelligence,” Upperwal said. In May, the company also launched India’s first open source multilingual foundational model, Pragna-1B. Upperwal said that it took the company six months to train the model, which involved many experiments with different models and a total of 150 billion tokens. Founded in 2019, Soket AI Labs’ focus was on building a decentralised data exchange for smart cities. However, things changed significantly after OpenAI CEO Sam Altman’s visit to India, which motivated the team to build the best AI models in the country. Apart from Soket AI Labs, startups like Sarvam AI and CoRover.ai have been focused heavily on building speech models. Speaking at Cypher 2024, Sarvam AI chief Vivek Raghavan demoed the speech capabilities of its AI models, leaving everyone at Cypher speechless.","excerpt":"The Realtime Speech API boasts ultra-low latency of under 500 milliseconds.","categories":["AI News"],"tags":["Developers","Realtime Speech API","Soket AI Labs"],"author_name":"Mohit Pandey","publish_date":"2025-01-15T13:05:47","publication_year":"2025","word_count":344,"keywords":["API","OpenAI","AI","Realtime Speech API","ML","Soket AI Labs","RAG","Python","Aim","JavaScript","R","Java","Developers"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","RAG","Python","R","JavaScript","Java","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/soket-ai-labs-launches-realtime-speech-api-offers-multilingual-capabilities\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162350,"title":"Data Privacy Day: How GCCs and Indian IT are Leading the Charge","content":"As we celebrate Data Privacy Day, it is important to acknowledge the increasing complexity of cybersecurity and the higher stakes involved in protecting digital assets. During a recent session at the World Economic Forum (WEF), experts discussed the rise in cyberattacks and the urgent need for effective defences. “Whenever there is conflict in the political space, cyber conflict follows,” said Matthew Prince, CEO of Cloudflare. Highlighting the scale of the challenge, Prince noted that his company blocks over 220 billion cyberattacks daily. From Russia’s invasion of Ukraine to the conflict in Gaza, cyber warfare has targeted critical infrastructure and disrupted communication systems, underscoring the growing link between physical and digital conflicts. The WEF Global Cybersecurity Forum 2025 report also highlights a concerning gap between organisations’ awareness of cybersecurity risks associated with AI and their preparedness to mitigate these risks. While two-thirds of companies recognise that AI will significantly impact cybersecurity this year, only 37% have the necessary tools to assess and address AI-related security threats effectively. The challenge is particularly acute for smaller organisations, with 69% lacking adequate safeguards for the secure deployment of AI technologies. This lack of readiness could leave them vulnerable as AI adoption accelerates without sufficient protections in place. A Collective Call to Safeguard Personal Data From healthcare to commerce, education, and governance, safeguarding data has become essential to building trust and fostering innovation in every sector. “At Dell Technologies, we believe data privacy is foundational to a resilient digital ecosystem. Organisations, governments, and individuals share a collective responsibility to protect sensitive information and uphold data rights,” said Ramesh Jampula, vice president, IT, India and APJC regional CIO, Dell Technologies. Businesses Lead with Privacy-by-Design Data privacy is no longer a box to check for compliance—it is a cornerstone of customer trust and financial stability. “Our technology and operations platforms underpin the daily trading of, on average, more than $10 trillion in equities, fixed income, and other securities globally,” said Prasad Vemuri, chief information officer, Broadridge Financial India. “We recognize that robust data privacy transcends compliance—it is the bedrock of financial stability and client trust,” he said. Meanwhile, SBM Offshore said that its advanced fleet of Floating Production Storage and Offloading (FPSOs) generate enormous volumes of operational data through extensive sensor networks that monitor equipment performance, safety, and efficiency. “We leverage this data using digital agents—advanced systems designed to detect anomalies and predict potential failures. These technologies enable us to take proactive measures and maintain high safety standards,” said Vikrant Sharma, IT department manager, SBM Offshore India. Beyond laws and technology, Data Privacy Day is a reminder of the importance of empowering individuals to take control of their digital identities. Consent and control are central to ethical data practices, ensuring people can decide how their personal information is shared and used. “This International Data Privacy Day, we should highlight consent’s vital role in digital ethical practices. When individuals share information willingly, they must maintain authority over its modification, deletion, and utilisation,” said Prashant Singh, COO, LeadSquared. Commenting on the same thought, Jay Swamidass, vice president and global head of sales, Rakuten SixthSense said, “Privacy isn’t an upshot of chance. It comes from putting in the effort to create strong systems, holding ourselves accountable, and fostering a culture where privacy matters to everyone, at every level.” He further mentioned that Rakuten SixthSense emphasises that privacy is fundamentally about trust—the trust individuals place in the organisation when sharing their data. Recognising that every piece of data represents a person, Rakuten SixthSense takes this responsibility with the utmost seriousness. India’s Take on Data Privacy India has made significant strides in data protection with the introduction of the Digital Personal Data Protection (DPDP) Act. The legislation places a strong emphasis on individual rights, such as consent, data correction, and erasure, while imposing strict penalties for non-compliance to prevent data misuse. “The DPDP Act reflects the global push toward stricter data governance and privacy norms in an increasingly digital world,” said Rohith Reji, co-founder & CEO, Neokred. “It sets clear principles around lawful data processing, consent, and transparency.” Commenting on the Draft Digital Personal Data Protection Rules 2025, a Reddit user wrote, “The million-dollar question is: What will happen to SPDI rules once the Digital Personal Data Protection (DPDP) Act comes into force?” The user also said that unlike GDPR and other privacy laws, the Indian legislation does not further classify personal data within the Act itself, leaving open the question of how this aspect will evolve. “It will be interesting to see which direction it goes in,” as the lack of categorisation sets it apart from its global counterparts. Another unique feature of the DPDP Act is its restriction to electronic data only. “It will be interesting to see how companies implement this limitation,” given its potential operational and compliance implications. That said, this version of the Act appears far more refined than earlier drafts. “I do like this version over the previous draft and the ridiculous suggestions provided by the JPC committee,” such as the inclusion of non-personal data. These improvements reflect a more focused and practical approach to data protection in India. Collaboration and Education Strengthening privacy frameworks requires collaboration between industries, governments, and individuals. Companies like Bosch Global Software Technologies are championing continuous education and accountability to build a secure digital future. “At Bosch Global Software Technologies, we recognize that data protection is not just about regulatory compliance—it’s about fostering trust and resilience,” said Vindhya Kudva, head of data protection & information security. “Through continuous education and partnerships, businesses can navigate data security complexities responsibly.” Kudva added, quoting Zig Ziglar, “People often say motivation doesn’t last. Neither does bathing—that’s why we recommend it daily. The same applies to data privacy awareness—continuous education and vigilance are essential.” Building a Privacy-Focused Future The collective focus on reskilling, recruitment, and workforce development will be critical to building a privacy-focused digital future. “As advancements in AI, edge computing, and IoT create new opportunities and challenges, the importance of data privacy will only grow,” said Jampula. However, according to the WEF Global Cybersecurity Forum 2025 report, the sector faces a significant shortfall in cybersecurity talent, with up to 4.8 million professionals needed to bridge the gap. Only 14% of organisations report having the skilled workforce they require to tackle current cybersecurity challenges. In the public sector, the issue is even more pronounced, with nearly half (49%) of respondents stating they lack the workforce necessary to meet their cybersecurity objectives. The cyber skills gap increased by 8% in 2024. “It’s critical we help close the growing cyber skills gap with a focus on training, reskilling, recruiting, and retaining cybersecurity talent,” said Chuck Robbins, chair and chief executive officer of Cisco. Resonating the same, Publicis Sapient also recognises that safeguarding privacy requires more than just technology—it demands a workforce equipped with advanced skills and a mindset of continuous learning. “By investing in reskilling initiatives, particularly for AI engineers and cybersecurity experts, we aim to empower our people to design solutions that embed privacy by design and anticipate emerging risks. This culture of continuous learning ensures that our people are not just problem-solvers but forward-thinkers who can address tomorrow’s challenges today,” said Amit Patil, senior director of technology at Publicis Sapient.","excerpt":"While two-thirds of companies recognise that AI will significantly impact cybersecurity this year, only 37% have the tools to assess and address AI-related security threats effectively.","categories":["AI Features"],"tags":["data privacy India"],"author_name":"Shalini Mondal","publish_date":"2025-01-29T09:00:00","publication_year":"2025","word_count":1205,"keywords":["Go","API","AWS","AI","Git","data privacy India","RAG","Aim","Rust","edge computing","R"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","edge computing","R","Go","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-privacy-day-how-gccs-and-indian-it-are-leading-the-charge\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043411,"title":"US IPO Puts Didi In Beijing’s Crosshairs","content":"The Chinese government is on track to take over the US as the tech capital of the world by 2030. Beijing’s Big Tech-friendly policies were instrumental in putting the country on the AI map. However, the government has now started clamping down on big tech companies to put them in their place. Chinese tech titans like Alibaba, Tencent and Bytedance have almost half a million employees. Meituan boasts of 569 million users and Didi Chuxing is one of the largest ride-sharing companies in the world today. Beijing is threatening to cut the Big Tech off global investment citing security concerns. The CCP wants to  control the data and also prevent the companies from leaning towards monopolistic practices. The Cyber Security Review Office (CSR) was created in 2020–backed by 12 Chinese ministries–to look into cybersecurity risks. It introduced a new cybersecurity regulation that demands companies to undergo a review process for transactions deemed a risk to national security. In addition, the CSR with the Chinese Securities Regulatory Commission and the State Administration of Foreign Exchange will act as gatekeepers for overseas IPOs. Roadblock for Didi The agency has audited top apps such as Baidu Maps, Suning Yigou and Tongcheng-Elong in the past. Last week, regulators ordered removing Chinese top ride-hailing giant Didi Chuxing from the app stores, putting it under review over possible data security and national security risks. The move came in the wake of allegations that Didi has mishandled sensitive data about its Chinese users. The action took place two days after Didi raised $4.4 billion in a New York IPO. It did not take longer than three trading days for this to trigger a sell-off in Chinese tech stocks in NY and Hong Kong, with Didi losing one-third of its market valuation. The shares of Alibaba and Tencent also plummeted. Now, Didi, along with truck hauling company Full Truck Alliance and online recruiting platform Kanzhun, have banned new users. Brocks Silvers, managing director at Hong Kong-based Kaiyuan Capital, told CNN the investigation on companies raising capital in the US was ‘no coincidence’. Beijing is firming up on its anti-monopolistic stance and focusing on firms seeking foreign investment. China’s legislature has also passed a new data-security law last month that goes into effect in September. The law will compel companies to handle their data with care. Alex Capri, a Singapore-based research fellow, spoke to CNN about data becoming increasingly strategic when more powerful AI, algorithms and machine learning are combined with state sponsored cyber activities. Clash with the US Big data has become the next battleground in the clash of tech superpowers. While the US is preventing China from obtaining tech like the advanced computer chips, China is enforcing stringent data controls that risk fragmenting financial markets and disrupting supply chains. For example, a US legislation passed in 2020 would allow the Public Company Accounting Oversight Board to review audits of large Chinese firms, including Alibaba and Baidu, that trade on American exchanges. This seemed to have triggered Beijing officials who have become vocal about the potential national security risks and data leaks. Chinese regulators are also planning changing rules that will allow them to block Chinese companies from listing abroad even if the unit selling shares is incorporated outside China, according to Bloomberg. In 2019, the US put Chinese company Huawei and several others on the Entity List– preventing them from doing business with any organisation that operates in the United States. The step was taken after Huawei was accused of espionage. China was also linked to the attack on Microsoft’s email software and Solargate. The US-China trade war, the rising geopolitical tensions, and the AI arms race have precipitated a technological cold war. While China risks losing the push it needs for technological superiority from abroad’s capital markets, US companies can be cut off from investing in one of the world’s top growing companies.","excerpt":"The Chinese government is on track to take over the US as the tech capital of the world by 2030. Beijing’s Big Tech-friendly policies were instrumental in putting the country on the AI map. However, the government has now started clamping down on big tech companies to put them in their place. Chinese tech titans […]","categories":["AI Features"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-07-13T12:00:00","publication_year":"2021","word_count":647,"keywords":["big data","Go","API","machine learning","AI","RAG","ViT","GAN","CNN","R"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","API","big data","GAN","CNN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/us-ipo-puts-didi-in-beijings-crosshairs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":8477,"title":"Qlik clicks with IIM Bangalore for BAICONF 2015","content":"Qlik®, a leading name in visual analytics has collaborated as a key sponsor with the Indian Institute of Management (IIM), Bangalore for it’s third Business Analytics and Intelligence Conference (BAICONF). The conference will be organised at IIM Bangalore campus from December 17 to 19, 2015 where more than 250 attendees including researchers, academicians, industry professionals and students have been given the golden opportunity to present their research papers. Moreover, the attendees will also get a first-hand experience of Qlik’s visual analytics solutions portfolio. Primarily, Qlik is a global entity, headquartered in Radnor, Pennsylvania, with more than 1700 partners in 100+ countries. It delivers a platform approach to Business Intelligence and driving innovative decision-making for organizations by helping them visualize their data with its patented associative technology. Customers from around the world rely on Qlik to explore the hidden relationships within data leading to insights that kindle better ideas. At BAICONF 2015, Qlik is planning to discuss ‘the importance of visual analytics in today’s competitive business landscape and how enterprises can leverage data effectively for insights, decisions and agility.’ Moreover,the participants will also get to listen to a speech by Souma Das, Managing Director, Qlik (India). Professor Dinesh Kumar, Chairperson of Data Centre and Analytics Lab and Executive PGP, who played a key role in organizing this conference for the third time, told us, “this conference aims to be a learning platform for academicians, practitioners and researchers from academia and the industry alike; Having Qlik onboard as a partner will provide us with greater insights into innovative visual analytics platforms, applications and best practices.” Moreover, it is anticipated that the discussions would revolve around topics such as, advanced data analysis, business analytics, big data and business intelligence. Souma Das, Managing Director, Qlik India recognises BAICONF 2015 as an opportunity for senior executives and IT professionals in India, to not only learn new ways of harnessing data and simplifying their business decisions, but also helping organizations to compete and not just survive, in the ever changing Indian economy.","excerpt":"Qlik®, a leading name in visual analytics has collaborated as a key sponsor with the Indian Institute of Management (IIM), Bangalore for it’s third Business Analytics and Intelligence Conference (BAICONF). The conference will be organised at IIM Bangalore campus from December 17 to 19, 2015 where more than 250 attendees including researchers, academicians, industry professionals […]","categories":["Deep Tech"],"tags":[],"author_name":"Apoorva Verma","publish_date":"2015-12-14T11:32:26","publication_year":"2015","word_count":336,"keywords":["big data","Go","business intelligence","AI","RAG","Aim","analytics","GAN","R","analytics platform"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","big data","GAN","business intelligence","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/qlik-clicks-with-iim-bangalore-for-baiconf-2015\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":46348,"title":"8 Solid Career Tips We Can Take From Andrew Ng’s New Webinar","content":"Chinese-American computer scientist and statistician, Andrew Ng is one of the most popular researchers among the millennials for his work in artificial intelligence, machine learning, deep learning, and other emerging technologies. His online courses on Coursera and deeplearning.ai has helped many enthusiasts to democratise these emerging technologies. In one of the webinars on building a career in machine learning, Ng shares tips and tricks on how to break into AI and discussed a few valuable skills that a person must have in order to successfully switch career machine learning. Ng had earlier tweeted, “I often advise people to take on projects you’re only 70% qualified for, but then learn like crazy to bridge that 30%.” On this note, we listed down eight most important points that the deep learning master advised in his video on building a career with machine learning. 1| Understand Emerging Tech If someone wants to pursue a career with emerging technologies, it is very important for him\/her to understand the basics of machine learning, artificial intelligence, deep learning, graphical models, neural networks and other technologies. Currently, the organisations are shifting towards the ecosystem where techniques like reinforcement learning, LSTM, CNN, RNN, etc. have been used thoroughly. Programming languages like Python, R, SQL, etc. are demanding these days and one must have a clear concept of these programming languages. The better way is to keep updated as much as possible. 2| Learn From Research Papers This is the most important point that Andrew Ng keeps stressing in almost all his videos. Whether it be a career-building webinar or a Stanford University online deep learning class, Ng advised to all the learners and listeners to read at least two research papers on the merging technologies. According to him, it turns out to be a very efficient way to learn the depth of any knowledge regarding the emerging tech. 3| Course Work And MOOCs Massive Open Online Courses (MOOC) and course work provided by organisations and academia contains a massive amount of information which cannot be found anywhere else. These sources contain exercises, practicals, etc. which helps a candidate not just to understand the topic but where and when to imply it. It is an efficient way to grab depth of knowledge in the interested areas, To be a strong potential candidate, completing online course work and MOOCs and adding it to the resume surely create a stand out among the other candidates in a job interview. 4| Working in a research project Doing an internship allows a practical depth of knowledge which allows a candidate to demonstrate the skills. Not only internships but also taking up a machine learning project on its own and trying to build and develop a model provides in-depth knowledge to the domain where the candidate wants to work on. 5| How to Build ML Systems Learning how to make machine learning systems work is very crucial in this field. With the help of online courses available, one can learn how to build a machine learning system from scratch. This will help in fetching a good-paid job along with a fruitful career in machine learning. 6| Prepare for ML Questions Along With Demonstrating the Portfolio of Work While appearing for a machine learning job interview, one must prepare the questions that are related to machine learning and artificial intelligence. There are various blogs where one can find common interview questions on emerging technologies. Also, in an interview, when a candidate is asked questions on topics like machine learning, along with answering the question, s\/he must also demonstrate the portfolio of the work that has been done earlier with these technologies. 7| Importance of Dirty Work According to Ng, downloading dataset, cleaning, plotting the learning curve and trying to figure out whether it is right or wrong, working and predicting PCAs can be said as the dirty work. However, he also mentioned that these are the most important parts while building a machine learning model. After all, data which is fed decides the fate of a machine learning model. One should not be afraid of jumping into doing dirty work. 8| Lifelong Learner Read research papers regularly or at least a few every week. The secret to becoming good at machine learning is not just studying certainly any weekend but to keep the pace by learning every weekend. One must study online courses and keep finding interesting research papers. If someone studies two papers a week it will make him\/her read 100 papers in a year which is eventually a huge amount of knowledge. This will help in getting better in AI skills with time. The current job market is directly proportionate to the actual job skills in the present scenario and constant learning will prove to be a benefit in this case.","excerpt":"Chinese-American computer scientist and statistician, Andrew Ng is one of the most popular researchers among the millennials for his work in artificial intelligence, machine learning, deep learning, and other emerging technologies. His online courses on Coursera and deeplearning.ai has helped many enthusiasts to democratise these emerging technologies.  In one of the webinars on building a […]","categories":["AI Trends"],"tags":["Andrew Ng","career in AI","ML jobs"],"author_name":"Ambika Choudhury","publish_date":"2019-09-25T12:00:07","publication_year":"2019","word_count":798,"keywords":["ML jobs","Go","artificial intelligence","machine learning","AI","neural network","ML","Andrew Ng","Python","career in AI","deep learning","SQL","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-solid-career-tips-we-can-take-from-andrew-ngs-new-webinar\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":52574,"title":"You Aren’t Considered Successful Until You Make It In India: Mike Capone, Qlik","content":"Data is crucial for business growth, but due to the strenuous nature of data science practices, firms struggle to make the most out of the information they collect. Since analysing and finding insights requires experts, only data scientists are able to deliver value by leveraging data. But for a business to thrive in this cut-throat competition, companies need to democratise data among every employee to enable self-analytics. To facilitate employees to analyse data easily, organisations are leveraging business intelligence tools and AutoML. While the former is allowing employees to mould and evaluate data dynamically, the latter is helping in automating ML model selection to get an accurate prediction. To understand how Qlik is helping several companies who are hungry to figure out the most effective ways of using data, Analytics India Magazine interviewed Mike Capone, CEO of Qlik, for our weekly column Deep Dive. Capone’s Journey At Qlik Qlik is an end-to-end platform for data analytics that empowers organisations to obtain insights into their data for mitigation business challenges. Founded in 1993, the Sweden-based company has helped firms gain operational resilience by utilising data and stay competitive in the market. The company hired Capone in 2018 to continue its endeavours and stay ahead of its competitors, mainly PowerBI and Tableau. “I started my career as a programmer and worked my way up and eventually became a CIO of a 12-billion-dollar company. Later, I was recruited to be the CEO of Qlik, and it was super attractive for me as I have been on both sides and can understand customers and development procedures,” says Capone. Over the last two years, he has spent building the team by fine-tuning operational strategy and expanding Qlik from just a visual analytics firm to comprehensive, machine learning-driven platform that includes data integration. Trends In The Business Intelligence Landscape On a trip to India and other Asian countries, Capone met with multiple CIO’s to learn how they are using analytics in their organisations. This helped him understand each company’s specific challenges with regards to becoming a full-fledge data-driven business. While talking to us, he mentioned that different companies are arduously trying to analyse data at the edge. “What’s great about Asia Pacific and India, in particular, is that companies here are ahead in terms of embracing analytics and data, there’s a high degree of data literacy in India and companies are willing to experiment, and push the envelope in terms of trying new technologies and leveraging modern AI capabilities to improve their operations,” Capone explained. How Is Qlik Helping Organisations With Streaming Analytics Over the past 15 months, Qlik has acquired firms to build a data platform, gaining access to technologies that can enable real-time analytics. Capone stressed on the need of getting rid of data lakes through streaming analytics. He emphasised that the platform’s recent integrations will enable companies to perform analytics in real time and do the heavy lifting when it comes to unlocking the insights hidden in their data. He also talked about how Qlik offered APIs and other toolsets to assist numerous firms for seamlessly accomplishing their objectives. “With our platform, organisations can harness the power of data with AI techniques to gain meaningful insights. Companies like HDFC Life are using our platform to determine fraudulent insurance claims,” says Capone. AI In Qlik Platform Capone said that, today, data scientists analyse data to find interesting insights and draw patterns that can guide the decision-making process. The major challenge, however, is embedding those insights back into business operations and making it available at the point of need to the decision-maker. In a firm with 30,000 or 100,000 people, a handful of data scientists will not be able to cope with the colossal volume of data queries that will be demanded. This, according to Capone, is where Qlik steps into the picture with its machine learning capabilities. The Qlik CEO showcased Qlik Insight Bot, an integrated bot that lets end-users interact with the data and trigger analytics through typing questions in natural language. This brings analytics right to the decision-maker, allowing even non-experts to seamlessly embrace and explore the datasets available to them to generate actionable insights. “The AI that’s built into our platform eliminates the need for coding, it’s our own technology embedded in the platform,” explains the CEO. “It will automatically interpret your data sets and provide logs for assimilating the outcomes so we call it a “white box.” Our API can be used to connect with R, Python, and DataRobot models for advanced analytics,” adds Capone. When asked about the Qlik platform’s capability compared to what one can get by in-house coding, Capone said that the comparison was tough because the answer depended on the type of business problem that was to be solved. He elaborated that most of the basic data science questions can be easily answered but if one is trying to look at multivariate regression, then they will have to use programming languages. How Qlik Differentiate In Fierce Competition The business intelligence landscape is already crowded with multiple providers such as Tableau and PowerBI and cloud providers such as Azure, AWS, and Google Cloud are also trying to establish their footprint in this burgeoning, high-potential space. Talking about the fierce competition, Capone said, “I have a lot of respect for our competitors, but we do much more than what they offer.” He spoke of how Qlik goes a step forward and solves large problems by helping enterprises seamlessly scale their data operations. Using a proprietary, patented technology that doesn’t use SQL, Qlik allows developers to bring data together from different data sources without having to worry about the data structure and SQL-based queries. Addressing data privacy concerns on SaaS environments, he further stated that Qlik does not access user’s content, thereby providing peace of mind to customers with data privacy concerns. Further, the CEO said that Qlik is a fully integrated platform, allowing firms to use it with any other solutions and platforms. Such functionality enables flexibility for companies, resulting in enhanced efficiency and productivity. “I don’t think any other BI providers can do what we are offering,” claims Capone. What’s Ahead “Our next iteration is to improve the platform further to help our customers run more complex tasks that use AI. The idea is to deliver an end-to-end solution for carrying out data science workflows,” unveils Capone. The Asia Pacific is the fastest-growing market for Qlik as it is gaining customers. “I always say that you cannot be a successful technology company unless you are successful in India.”– Mike Capone, Qlik The company envision high revenue from Asia; thus, it is doubling down its resources to help its customers achieve business growth. “I expect APAC to outpace the rest of the world in terms of growth for us for the foreseeable future,” concludes Capone.","excerpt":"Data is crucial for business growth, but due to the strenuous nature of data science practices, firms struggle to make the most out of the information they collect. Since analysing and finding insights requires experts, only data scientists are able to deliver value by leveraging data. But for a business to thrive in this cut-throat […]","categories":["Deep Tech"],"tags":["Data Scientist","qlik"],"author_name":"Rohit Yadav","publish_date":"2019-12-25T17:09:43","publication_year":"2019","word_count":1137,"keywords":["data science","machine learning","AWS","AI","ML","RAG","Python","Aim","qlik","analytics","Data Scientist","Azure"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","RAG","AWS","Azure","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/you-arent-considered-successful-until-you-make-it-india-mike-capone-qlik\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":34301,"title":"Harvard&#8217;s Tiny RoboBee Is Exactly Like The Scary Bees From Black Mirror","content":"Since time immemorial, robot developers have long looked at nature for inspiration. When designing robot mechanisms and movement, animals and human motor functions provide insight into the optimal movement that many robotic engineers had the curiosity to always try to replicate it. In the past 25 years, a small design lab at Harvard has been relentlessly working on a unique robotics project perfecting the world’s first and only RoboBee. After many upgradations, now this Harvard’s tiny robotic bee has learned how to stick to surfaces like Spiderman. Unlike spiders that use thousands of tiny hairs to climb walls, though, the upgraded RoboBee uses the power of static electricity as a team of engineers from both Harvard and MIT wanted it to find a way for minuscule drone batteries to last longer. Hybrid RoboBee The concept of RoboBee was first conceived by mechanical engineering student Robert Wood in 1991, much before Hollywood grabbed the idea and showcased in its films. Since then, the RoboBee project has been constantly a perennial work-in-progress for scientists and students at the Harvard School of Engineering and Applied Sciences and the Wyss Institute for Biologically-Inspired Engineering. In true sense, the RoboBee is a mechanical marvel. Be its tiny polymer wings which are designed to mimic real insect wings are powered by small ceramic muscles that convert electrical pulses into kinetic energy, making the RoboBee the world’s smallest flapping wing aircraft. RoboBee Features  & Facts The latest version of the RoboBee can stick to walls, can fly, dive into the water, swim around, and propel out of the water All those are tricky manoeuvres for RoboBee is based on larger amphibious drones that manage those tricks RoboBee weighs in at mere 175 milligrams only. The tiny machine overcomes forces of mass, volume, and surface tension that are completely different than what a bird-sized robot must deal with The bot also requires a multimodal locomotive system that enables it to fly and swim The RoboBee is nearly 1000 times lighter than any other aerial-aquatic robot, and this difference in scale is what’s kept decades worth of Harvard engineering students busy with the design The RoboBee represents a platform where the forces it experiences are different than what we at human scale do Design, Structure & Functions Of RoboBee The RoboBee has four buoyant outriggers specifically robotic floaties, essentially. To achieve lift-off from the surface into the air, the RoboBee uses a small electrolytic plate that converts water into oxyhydrogen, a combustible fuel. It uses laser-cut materials that are layered and sandwiched together into a thin and flat plate. The flat materials then pop open like a book into its final and complete electromechanical structure. Then the spark mechanism within the bot ignites the gas, powering the RoboBee upward like a minuscule rocket ship. Once the bot gets suspended in the air it then stabilises itself and lands. The robots physical structure is inspired by the biology of a usual fly with only submillimetre-l scale anatomy and two wafer-thin wings. The wings are mostly are non-visible, flapping 120 times per second. It uses both motion control systems, as well as micro-manufacturing techniques. The tiny robot uses piezoelectric actuators, which are made from strips of ceramic that expand and contract when an electric field is applied. Larger robots use electromagnetic motors but at this small scale, the piezoelectric actuators were the only alternative. The carbon fibre body frame that it uses has thin plastic hinges that work as its joints. Then its balanced control system commands the rotational motions with each wing controlled independently in real-time. The control system usually needs to react quickly but as powerful change takes place in airflow, which creates an outsized effect on the flight dynamics, the bot uses water to become buoyant and with an electrolytic plate in the chamber which then converts the water into oxyhydrogen. This provides the robot with extra buoyancy needed for the wings to pop out of the water. This way the robot can fly into and an out of the water without any pertinent damage. Applications Of RoboBee The RoboBee is used as a testbed for new robotic mechanical studies as well as new material construction methods. Recently, RoboBee even demonstrated a new programming language that taught the robots to not only move like the insect inspiration but also to think like them exactly mimicking the way an insect’s brain operates. Anyway, these flying microbots like RoboBee will play a highly beneficial role in agriculture, search and rescue missions, surveillance and climate monitoring. For example, the technologies developed to manufacture such Robobee could be used in the medical field to make small surgical devices for endoscopic procedures. Inspired by RoboBee functionality Ferrari’s labs also worked upon a new class of event-based sensing and control algorithms that mimic neural activity and also on swarms of RoboBee to make them communicate with one another and coordinate their movements. Ferrari’s lab teamed up with Harvard’s RoboBee to test the new chips. The robot has all the necessary vision, optical flow, and motion sensors needed to provide an adequate test bed. Ferrari lab is even installing RoboBee into newer microdevices such as a camera, expanded antennae for tactile feedback, contact sensors on the feet and air-flow sensors. Even many physics-based simulators models, follow  RoboBee and replicate its model with instantaneous aerodynamic forces that it would face during each wing stroke. This way the simulator can accurately model RoboBee’s motions during flights through complex environments. Outlook One of the major drawbacks of RoboBee is that it has only remained a tethered device till now. RoboBee as a benchmark robot  is not a very difficult invention to build, but other robots that are already untethered would greatly benefit from this invention because they have similar  issues in terms of power Before any of these benefits can be realised, Harvard researchers must solve some tougher technical challenges to find how to power a small flying robot.","excerpt":"Since time immemorial, robot developers have long looked at nature for inspiration. When designing robot mechanisms and movement, animals and human motor functions provide insight into the optimal movement that many robotic engineers had the curiosity to always try to replicate it. In the past 25 years, a small design lab at Harvard has been […]","categories":["AI Features"],"tags":["AI Robots","bill gates","harvard"],"author_name":"Martin F.R.","publish_date":"2019-01-30T07:15:33","publication_year":"2019","word_count":991,"keywords":["Replicate","Go","AI Robots","programming_languages:R","AI","Modal","programming_languages:Go","BERT","harvard","llm_models:BERT","ViT","bill gates","R"],"extracted_tech_keywords":["AI","R","Go","BERT","ViT","Replicate","Modal","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/harvard-robobee-black-mirror\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054110,"title":"What Does A Top University Look For In A Student For MS In Data Science: Rutgers University","content":"For the latest in our series on MS In Data Science Program in the US for Indian students, Analytics India Magazine got in touch with Rutgers, The State University of New Jersey. Established in 1766, Rutgers is an academic, health and research powerhouse and one of the US’s oldest and largest universities. The university has a network of 500,000 active alumni. AIM: How does Masters in Data Science Program stand out, making students industry ready? Rutgers: The MS in Data Science at Rutgers University is a professional MS program that prepares students to be job-ready upon graduation. It is a rigorous program that is offered by the Department of Statistics, to be completed in three full-time semesters. While many data science programs are very wide and shallow, Rutgers MSDS is narrow and deep. Students learn important theories and core concepts that allow them to be lifelong learners and practitioners of data science irrespective of what the current trendy tools are available now or will be available in the future. As a professional program, practical training is its core aspect, making MSDS one of the few programs in the US to provide international students with day-1 CPT (Curricular Practical Training), or the ability to take a relevant internship opportunity as soon as they join the program, compared to the typical one year waiting period. In addition, the Department of Statistics has established the Office of Professional Programs that is dedicated exclusively to providing internship, part-time and full-time opportunities to students in the program; this is in addition to students having access to the office of Career Exploration and Success that is available to all Rutgers students. In addition, our location in New Jersey with close proximity to New York, Pennsylvania, Delaware and Connecticut makes us an attractive hub for companies recruiting from these areas. AIM: What are the key aspects that the University looks for in a student during the admission process? Rutgers: We look at applicants in a holistic manner and consider all aspects of the application when making admission decisions. This includes college transcripts, test scores, letters of recommendation and personal essays. Having said that, MSDS is an academically rigorous quantitative program that requires a strong background in math and statistics. Admission decisions favour those applicants who can demonstrate strong proficiency in math, statistics and knowledge of a programing language. AIM: Do you place any specific emphasis on foreign students, especially from India? Rutgers: We welcome applications from both domestic US students and international students from across the world. When considering international students, English proficiency and communication skills in English are important. Students from India historically score well on those two aspects. AIM: How well is the university linked with industry to facilitate on-campus placements? Rutgers: The Department of Statistics has established the Office of Professional Programs that is dedicated to finding employment opportunities specifically for MSDS students. MSDS students get exclusive full-time job and internship opportunities from corporate partners of the program. This is in addition to Rutgers’ University-wide career services office, one of the biggest in the country, that is also available to all Rutgers students, including MSDS students. Students have found employment in financial services, consumer, pharmaceutical, technology and consulting industries in data science-related positions once they graduate. AIM: Can you share with us the placement record of the college, specific to the data science program for students? Rutgers: We don’t carve out employment numbers by student type (international or domestic); however, 85 percent of our graduates find full-time employment, and that too within six months of graduation, while a small percentage chooses to pursue further graduate studies, including a PhD. Our international students typically make up 60 percent or more of our students. Important to note that previous full-time work experience is not required and that a majority of MSDS students enter the program with less than one year of full-time work experience. An overwhelming majority of students find internships. Typically, many students get internships as soon as they start the program in the Fall via existing corporate relationships that the program has with employer partners. AIM: Who can apply to the course? Any requirements specific to Indian students? Rutgers: MSDS is open to all students who have completed an undergraduate degree and have fulfilled the program’s academic prerequisites in math, statistics and computer programing listed below: Multivariate CalculusLinear AlgebraIntroduction to ProbabilityTheory of StatisticsTwo courses in statistical computing or computer science, including advanced programming. For further info, one can visit here. AIM: Can you mention the availability of any scholarship and accommodation availability for Indian students? Who will all be the concerned authorities for the Indian students to contact for matters related to admission directly? Rutgers: The department of Statistics does not offer any scholarships or financial assistance. However, students are eligible for external scholarships. In addition, there are some grader and other paid on-campus opportunities available to students in the program. All admission inquiries should be addressed by email to the Director of Admissions at msds@stat.rutgers.edu directly.","excerpt":"Students learn important theory and core concepts that allow them to be lifelong learners and practitioners of data science irrespective of what the current trendy tools are available now or will be available in the future.","categories":["AI Trends"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-11-24T14:00:00","publication_year":"2021","word_count":833,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","Aim","analytics","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/what-does-a-top-university-look-for-in-a-student-for-ms-in-data-science-rutgers-university\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":56646,"title":"Five Best Practises For Data Scientists To Effectively Communicate With Businesses","content":"Much has been written about the technical skills data scientists need to be adept at, but their technical expertise would be for nought if they do not understand how to present their findings in a comprehensible manner. Organizations across a wide range of industries are increasingly turning to data to make the right decisions, and this is opening up new avenues for data scientists. This also means that they will have to broaden the scope of their work to learn how to cultivate team cohesion and effectively collaborate with businesses. In addition to being more versatile, the onus of setting expectations of business stakeholders also lies on data scientists. Sitting at the interjection of raw technology and business, they need to be fluent in both languages to be able to build a relationship between the two. Focus on outcomes, values and optimized actions While the basics of how a data scientist can stand out in the crowded market are elucidated here, following are some practices that can make them better communicators:-One of the biggest challenges faced by data scientists during business meetings is tailoring their findings to the audience’s technical level. They often struggle to find the right balance between not compromising on key points while making their presentation and not overwhelming colleagues with technical jargons as they do so. To overcome this challenge would be to first and foremost, acknowledge the fact that while everyone in the room may be seeking data-driven insights, not everyone speaks the same language.One way to tackle this problem would be to emphasize actionable outcomes.Most business owners are generally interested in knowing how the findings can impact their businesses, and not necessarily on the exact analysis. To get around this, identify the problem and explain it to the audience in clear and simple terms. This can be followed by citing evidence of the existence of the problem. Once a case for the problem has been made, explain how it can be fixed. A clear timeline and the price of fixing this problem could be a valuable addition, especially from the organization’s point of view. Use empathy to craft a compelling story When starting with the problem, it will be helpful if you relate it to common experiences or interests of the audience so that they understand its larger context. This ability to present insights in the form of a story will hold data scientists in good stead, especially when they make business presentations. Employing clever narratives will help them connect with their audiences – which, in most cases – would involve non-technical stakeholders. Punctuate analysis with data – sparingly While the importance of keeping the presentation simple and uncomplicated cannot be emphasized enough, some data points may find credence with some stakeholders. But be very frugal in doing so. Understand that showcasing every data point all at once can overwhelm your audience.Hence, it is important to present it in a staggered fashion and use it astutely to help build the story and narrative. Use data visualization tools Very few people can look at a spreadsheet and draw clear conclusions from the data. However, most have the capability to compare and interpret graphs and bars easily.Thus, by representing data visually, data scientists can hold their audience’s attention as well as communicate their message effortlessly. But be sure to demonstrate and break down the visual data, that is, make sure that you are clearly explaining what is shown on the screen. For instance, if you are alluding to a particular section of a graph, point to it as you explain what it means in the broader context of the problem. Presentation tips Pace is important. While a big battle has already been won by shaping the conclusions of the analysis in an easy-to-consume language, holding the attention of the audience is a different ball game. As a general practice, do not exceed 15 minutes for the entire presentation. Be sure to leave out unnecessary details – be it for the explanations, or the visual tools. Also, do the math for the audience and quantify it for them if they ask for specifications. This may come in handy, especially when making predictions.","excerpt":"Much has been written about the technical skills data scientists need to be adept at, but their technical expertise would be for nought if they do not understand how to present their findings in a comprehensible manner. Organizations across a wide range of industries are increasingly turning to data to make the right decisions, and […]","categories":["AI Trends"],"tags":["Data Scientists","data visualization"],"author_name":"Anu Thomas","publish_date":"2020-02-14T12:30:00","publication_year":"2020","word_count":695,"keywords":["Go","API","programming_languages:R","AI","data-driven","programming_languages:Go","data visualization","GAN","R","Data Scientists"],"extracted_tech_keywords":["AI","R","Go","API","GAN","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-practises-for-data-scientists-to-communicate-with-businesses\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10057998,"title":"Intel takes hold of the processor market with the 12th Gen Intel Core","content":"Intel has unveiled the 12th Gen Intel Core processor family that includes 60 processors and will power over 500 designs across all PC product segments – ranging from desktop and gaming to ultra-thin-and-light laptops. The 12th Gen Intel Core processors are the first ones that are based on Intel 7 process technology. The new performance hybrid architecture will enable major performance gains across the range of PC workloads. The company has also unlocked “K” desktop processors, which they claim includes the world’s best gaming processor – the 12th Gen Intel Core i9-12900K. It has a max turbo boost of up to 5.2 GHz and 16 cores and 24 threads which can reach new heights of performance. One of the new launches also includes the new 65- and 35-watt 12th Gen Intel Core desktop processors that have good scalable power and performance for productivity, creation, and gaming. Other features added in the new lineup include H-series mobile processors that feature the flagship Intel Core i9-12900HK – the fastest mobile processor and the best mobile gaming platform that is built on the Intel 7 process. Intel has also announced Intel vPro platform offerings which include Intel vPro Essentials and Intel vPro Enterprise. These are for expanded commercial support for small businesses and large-scale enterprises. The 12th Gen Intel Core processor family represents the company’s most scalable lineup to date, powering designs across consumer, IoT, enterprise, and other applications. “Intel’s new performance hybrid architecture is helping to accelerate the pace of innovation and the future of computing. And, with the introduction of 12th Gen Intel Core mobile processors, we are unlocking new experiences and setting the standard of performance with the fastest processor for a laptop – ever,” said Gregory Bryant, EVP and general manager, Intel’s Client Computing Group. IoT Application At CES 2022, Intel launched the 12th Gen Intel Core processors, code-named Alder Lake S-series and H-series. Enhanced to accelerate IoT application and innovation, the new processors offer manufacturing, retail, healthcare, and digital safety customers increased core counts; advanced graphics, display, media; and AI capabilities. According to Intel, this is their first family of processors that is enhanced for the edge to feature performance hybrid architecture to combine Performance-cores and Efficient-cores with Intel Thread Director. The key features include: High flexibility and computing performance and for heavier IoT workloads.Accelerated deep learning and AI capabilities without additional hardware.Enhanced integrated GPU to support 4K and 8K displays.Hardware-based security to defend vulnerable IoT devices. The Alder Lake S-series processors are up to 1.36 times faster in single-thread performance, up to 1.35 times faster in multithread performance, up to 1.94 times faster in graphics performance and up to 2.81 times faster in GPU image classification inference performance compared with 10th Gen Intel Core processors. The Alder Lake H-series processors for IoT have been estimated to be up to 1.04 times faster in single-thread performance, up to 1.18 times faster in multithread performance, and up to 2.29 times faster in graphics performance compared with 11th Gen Intel Core processors. These SKUs feature up to 14 cores and 20 threads at 35W to 45W TDP. The new advancements will prove to be beneficial for industrial manufacturing, retail, banking, hospitality and education, healthcare, and digital safety and security customers. Intel’s 12th Gen Core processors have over 80 IoT customers participating in Intel’s Early Access Program. Alder Lake S-series and Alder Lake H-series are expected to be available in January 2022 and April 2022, respectively. Laptop\/PC Intel claims that with 12th Gen Intel Core H-series processors, it has set the standard for the highest performance laptop processors ‘on the planet’. The series of processors is built on the Intel 7 process node with its first performance hybrid design. The H-series is led by Intel Core i9-12900HK. It delivers up to 40% higher performance for gaming experiences and up to 28% faster gameplay than the i9-11980HK. Intel will launch new devices with the 12th Gen Intel Core H-series, with its partners that include Asus, Acer, Dell, HP, MSI, Lenovo, and Razer. Intel has also unveiled a new ultraportable mobile processor – the P-series product line, which brings high levels of performance to thin-and-light laptops. The new process will power systems starting in February 2022. Key features of the 12th Gen H-series processors include: Up to 5 GHz frequencies, 20 threads and 14 cores (6 P-cores and 8 E-cores).Broad memory support for DDR5\/LPDDR5 and DDR4\/LPDDR4 modules up to 4800 MT\/s. PC connectivity to multiple 4K monitors & accessories and transfer speeds up to 40 Gbps. Mobileye Along with the 12th Gen Intel Core processor, Intel’s Mobileye announced a new system-on-chip (SoC) for autonomous vehicles (AVs) – EyeQ Ultra. It is built on Mobileye’s EyeQ technology and does the work of 10 EyeQ5 SoCs in a single package. Wrapping up With the launch of 12th Gen Intel Core processor, Intel claims that they have killed competition. After falling behind AMD for a few years now, Intel has come back to being the maker of the best processors. But the important question remains how long it will stay at the top position.","excerpt":"With the 12th Gen Intel Core processor family, Intel claims to have killed competition. The article looks at the new features that the new processors offer","categories":["Global Tech"],"tags":["AMD","Intel","Intel processors"],"author_name":"Meeta Ramnani","publish_date":"2022-01-10T14:00:00","publication_year":"2022","word_count":848,"keywords":["Go","AMD","programming_languages:R","AI","innovation","Intel processors","Scala","Git","Aim","deep learning","ViT","R","Intel"],"extracted_tech_keywords":["AI","deep learning","Aim","R","Go","Scala","Git","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/intel-takes-hold-of-the-processor-market-with-the-12th-gen-intel-core\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10137847,"title":"E2E Networks Partners with People+ai to Expand Cloud GPU Access in India","content":"E2E Networks has announced  its partnership with People+ai, an initiative by the EkStep Foundation, to address the growing demand for cloud GPU and compute power in India. This collaboration will strengthen India’s compute ecosystem by making advanced and customisable cloud resources more accessible to a wider audience. People+ai’s Open Cloud Compute (OCC) project is central to this partnership, focusing on creating a network of micro data centres across India. The OCC project promotes an open, decentralised cloud infrastructure that allows businesses, developers, and government entities to collaborate effectively. By utilising open-source technologies, OCC democratises cloud access, enabling organisations of all sizes to benefit from computing resources, reducing latency, and improving data sovereignty. “At Open Cloud Compute, our vision is to democratise AI by making compute resources accessible to all innovators. The collaboration with E2E Networks will empower AI innovators across India, from early-stage startups to established industry leaders, fostering inclusive and diverse ecosystems for AI-driven innovation,”said Tanuj Bhojwani, Head of People+ai. “We’re excited to partner with People+ai to make cloud computing and GPU power more accessible. This partnership will help businesses in India, big and small, leverage advanced AI and cloud technologies to innovate and grow. We aim to build a strong AI ecosystem that benefits everyone,” said E2E Networks’ chief revenue officer, Kesava Reddy. The initiative is timely as India’s demand for scalable and flexible cloud solutions continues to surge, driven by rapid digital transformation across various sectors. E2E Networks, a leading AI-focused cloud company listed on the NSE and empanelled by MeitY, will offer high-performance cloud GPUs such as H100 and A100, along with cluster configurations and AI development studio TIR. These resources will allow private and public organisations to train, deploy, and scale AI-driven applications with ease. In an exclusive interview with AIM, Reddy  previously revealed that he is actively in dialogue with folks from People+AI and the company also wants to be part of the government’s plan to build a GPU cluster of around 10,000-20,000 GPUs which will be made available for research institutions and startups in India. “We have given our pre-bid queries to the government and are waiting for a response,” Reddy said. Some of the notable customers of E2E Network include Zomato, IndiaMART, CarDekho, Zoomcar, Niyo, Nykaa, Mobikwik, Reverie, IIIT Hyderabad, ISB, IIT Guwahati, and Matrimony.com, among others. As GPU computing becomes increasingly vital for specialised workloads like AI, machine learning, and high-performance computing, the partnership between E2E Networks and People+ai positions India to lead in the AI-driven future. This partnership will support startups, SMEs, educational institutions, and large enterprises alike, fostering innovation and competition across the AI landscape.","excerpt":"E2E Networks, a leading AI-focused cloud company listed on the NSE and empanelled by MeitY, will offer high-performance cloud GPUs such as H100 and A100, along with cluster configurations and AI development studio TIR.","categories":["AI News"],"tags":["AI GPUs"],"author_name":"Siddharth Jindal","publish_date":"2024-10-08T16:47:22","publication_year":"2024","word_count":436,"keywords":["Go","machine learning","AI GPUs","AI","cloud computing","R","Scala","Git","RAG","Aim","GPU computing"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","cloud computing","GPU computing","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/e2e-networks-partners-with-peopleai-to-expand-cloud-gpu-access-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":1071,"title":"IoT is now becoming more Open Source, and Eclipse is leading the charge","content":"Open source is the key driver for growth of any upcoming technology. Recently, this has been manifested in the area of Internet of Things (IoT), with the Eclipse Foundation. Set up in 2004 by legacy enterprise IBM, Eclipse has recently spurred the growth of a big ecosystem around open source IoT. Eclipse IoT provides the technology for building IoT devices, gateways and cloud platforms and while it was seen as a celebration of Java, over the years the open-source project has become language-agnostic and has adopted C and C++ IDE system. It’s long believed that IoT has its roots in the industrial sector, what with world’s legacy companies Bosch and GE unifying the fragmented IoT platform with their proprietary tools. GE Digital’s Predix platform made a lot of headlines in 2016, with AI and Machine Learning capabilities woven into it to sharpen its real-time optimization and analytics capability. While the German machine maker’s IoT suite is a platform for managing IoT device connectivity. However, despite the staggering investment in trillions of dollars by storied companies and venture capital, Mike Milinkovich, Executive Director of the Eclipse Foundation, who was present in India recently for the Eclipse Summit 2016, drilled in the importance for open source software at his keynote address.  “But much of this investment in billions of dollars is ignoring one fundamental truth — the basic building blocks of the Internet of Things are going to be built on top of free and open source software. Business models that expect to achieve market dominance based on proprietary business models are going to fail.” Mike Milinkovich is the Executive Director at Eclipse Foundation Why software licensing business model might not be favourable for IoT Industry Essentially, when it comes to IoT, there is still no one-size fits all policy. According to a prediction by Juniper Research, the number of connected devices will cross the 46 billion mark in 2021, it’s a forecast bigger than all the previous predictions. Juniper Research also billed the revenue opportunity in IoT to be the tune of $300 billion. But as the number of connected devices increase, so do concerns about privacy and how to optimize them for target markets, how the devices will communicate with each other and how to accelerate the time to market. Even though the market sentiment remains overwhelmingly positive about what with IoT grabbing headlines year long, Milinkovich drives home the point that for enterprises to make money, IoT will need to be open source.  “There is no money in software. Business models of today are not based on software licensing. Anybody trying to make money by selling software that goes on devices is going to be dead and no one will participate with a vendor that tries to build a business model like that,” he noted at the summit. There is no better eye opener for the inefficacy of the software licensing business model than the startling number shared by US Bureau of Economic Analysis that showed that computer software is now 0.7% of its price in 1980. In his talk, Milinkovich who had past stints at Nortel, IBM, and Oracle before joining Eclipse Foundation lists down five major reasons for IoT to be open source: 1) To scale effectively, software has to be free: The writing on the wall is clear — there is no money in software. Milinkovich cites the example of big enterprises such as Facebook, Twitter and Google where all of the infrastructure that was needed to scale effortlessly is based on open source. “Google would never have been able to get off the ground if they had to pay Microsoft money for every individual server they were putting in their data centre. For the same reason anybody who comes with a business model that says we are going to put a tax on every device will lose out to a person who has a software that goes on the devices for free,” he cited. 2) Open source means innovation without permission: Open source enables a) Permission less innovation b) Innovation through integration c) Far higher levels of experimentation “If you want to move faster and accelerate time to market you will need developers to come up with cutting edge solutions. The best part of open source IoT is developers doesn’t need to seek permission to mash up different pieces of software together to see if you can use them in a new solution, they can just go ahead and try it,” he said in his keynote address. 3) Interoperability is the way to save cost: Interoperability is the way for companies to come together and share the same implementation of the same standard and the cost of that implementation, if shared is close would accelerate time to market and save money. A recent example of interoperability in IIoT is GE and Bosch – the companies recently signed an MoU to boost the growth and openness of Industrial Internet of Things (IIoT). 4) Internet of Things is not a Market, it is service driven:  For enterprises to survive, they must understand that Internet of Things (IoT) is not a market in itself, it is an enabling technology people can use to build solutions for the market, or even build solutions for customers in market. It is an enterprise and customer driven technology. “Basic plumbing and basic infrastructure of IoT is going to be based on open source,” he noted. That’s why open source acts as building blocks on top of which they can build solutions. 5) Developers vote for the technology of their choice:  According to an estimate by Vision Mobile, the IoT industry would require a whopping 4.5 million developers by 2020. Part of the reason why Open source is so successful and popular is that it brings down the cost of training developers significantly. But that’s not all. Open Source is going the REST way.  Milinkovich reminisces about the technology era in 2000, when the world at large knew SOAP web services. Then came REST, a similar competing technology which was better documented and the developer community embraced it wholeheartedly. “And a good example of that is MQTT first developed in 1999 and has been around for a long time in terms of its adoption as an open source solution for doing IoT protocol. It amply demonstrates how developers are voting for the technology of their choice,” he shared. On the IoT ecosystem, Eclipse open source platform provide IoT development tools, communication protocols and application framework and runtimes. In order for businesses to be successful, Eclipse IoT is enabling end-to-end IoT solutions, right from software to run on sensors, to gateway management solutions and back-end infrastructures.","excerpt":"Open source is the key driver for growth of any upcoming technology. Recently, this has been manifested in the area of Internet of Things (IoT), with the Eclipse Foundation. Set up in 2004 by legacy enterprise IBM, Eclipse has recently spurred the growth of a big ecosystem around open source IoT. Eclipse IoT provides the […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-01-03T06:21:47","publication_year":"2017","word_count":1112,"keywords":["Go","API","machine learning","AI","Git","RAG","C++","analytics","R","Java"],"extracted_tech_keywords":["AI","machine learning","analytics","RAG","R","Go","Java","C++","Git","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/iot-becoming-open-source-eclipse-leading-charge\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10062263,"title":"The best Indian conferences for women in tech in 2022","content":"Debjani Ghosh. Susan Wojcicki. Gwynne Shotwell. Upasana Taku. Roshni Nadar. Neelam Dhawan. Deepa Madhavan. For every Elon Musk or Jeff Bezos, we have an equally powerful woman changing the face of the technology industry. Yet, STEM is one of the major fields to witness gender inequality in the professional sphere. Globally, for every 100 men promoted to manager, only 72 women match up. This ratio is even lower in India. Today, more and more companies are taking steps to lessen this divide. Women conferences, forums, and events are a proven method to encourage women’s participation in tech and allow individuals to learn from leaders’ experiences. Analytics India Magazine has listed down some of the best Indian and virtual (global) conferences taking place this quarter for tomorrow’s women leaders to attend. The Rising 2022 April 8-9, 2022 The Rising is Analytics India Magazine’s annual conference for AI and data science women. It is the biggest meeting for professionals across the tech domain and brings together women from academia and industry to discuss the latest trends, ideas and achievements. The two-day conference will be held in Bangalore and consists of talks and sessions catered towards leadership, empowerment and celebration of women disrupting the industry. Speakers at the conference are leaders such as Lata Iyer, VP, Research at Rakuten India. For more details on the conference, click here. Watch more here. Women in Cybersecurity March 17-19, 2022 In its 9th year, Women in Cybersecurity is the flagship event and largest cybersecurity conference for working women and students to help organisations recruit, retain and advance women in cybersecurity. Additionally, the team creates a community of women in the industry to support and encourage innovations and ideas. For more details on the conference, click here. International Women’s Day India Summit April 2-3, 2022 IWD India Summit is described as “a celebration for all women in technology”. The event is virtual, inviting women from across the country to learn through the exchange of inspirational examples and experiences. The conference aims to empower women through learnings, curated sessions, exchanging best practices and more. The theme for 2022 is Progress Not Perfection. The event welcomes beginners, experts and anyone passionate about tech. For more details on the conference, click here. Women Transforming Technology Conference April 11-12, 2022 Women Transforming Technology, a consortium of organisations building a community to tackle issues and encourage women in technology, has been holding conferences for over five years now. The 2022 agenda includes exploring career tracks, emerging leaderships, ensuring inclusion, technical workshops and learning from senior leadership. For more details on the conference, click here. WITS Spring 2022 Virtual Summit April 27-29, 2022 The Women in Tech Summit (WITS) is a virtual version of the Women in Tech Summit series that empowers women through discussions of trending topics in the industry. WITS aims to support and educate women through career phases and opportunities through online networking, connections, workshops, company contacts and inspirational speakers. Speakers include leaders from IBM, RedHat, AWS, Bloomberg, KPMB and more. For more details on the conference, click here. WITI Virtual Summit June 20-21, 2022 In their 28th edition, the Women in Technology International (WITI) aims to advocate, recognise and encourage women’s contribution to the tech industry. The event includes insights, inspirations, and action items from women worldwide. WITI’s membership consists of over 3 million industry professionals, 300 partners, and 60 networks, with various global events to support women’s collaboration across the globe. For more details on the conference, click here. Wonder Women Tech Virtual Summit June 7-10, 2022 Claimed to be ‘the largest conference for women in tech’, Wonder Women Tech Summit takes place in a hybrid model, with the three initial days being virtual, welcoming over 500 speakers from across the globe. The event is catered towards women in STEM to interact with industry leaders, get inspiration from experiences, attend presentations and panels, workshops, classes to up-skill and breakout rooms to network with leaders. The focus of 2022 is to create diversity and drive innovation in the industry. The speakers at the summit are leaders from big tech companies like Amazon, Google, Tesla, IBM, Uber, Microsoft, Lyft, etc. For more details on the conference, click here. IEEE Wintechcon June 2-3, 2022 Organised by IEEE CAS Bangalore Chapter, the Wintechcon is sought after by hundreds of women in tech. The conference was introduced in 2018 as an exclusive forum for women technology leaders in India to present their work in emerging knowledge areas. While the conference speakers are women, the forum sees a footfall from individuals of all genders, demonstrating various technical subjects within the one-day event. The theme for Wintechcon is ‘Smarter technologies for a sustainable and hyper-connected world’. For more details on the conference, click here.","excerpt":"The two-day conference will be held in Bangalore and consists of talks and sessions catered towards leadership, empowerment and celebration of women disrupting the industry.","categories":["Deep Tech"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-03-08T12:00:00","publication_year":"2022","word_count":789,"keywords":["data science","Go","AWS","AI","RAG","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","AWS","R","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/the-best-indian-conferences-for-women-in-tech-in-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":25754,"title":"Top Non-Profit Artificial Intelligence &#038; Machine Learning Institutes That Are Working On Making AI Safe","content":"Tech companies across the globe are becoming more and more aware of the power of artificial intelligence and singularity. Not-for-profit institutes are thus conscientiously working on the strategic implications and the openness of artificial intelligence. From subjects such as ethics and policy in AI to the development of machine intelligence, not-for-profit research institutes are working towards the advancement of making AI safe. In this article, we list down top not-for-profit AI research institutes across the globe that are doing ground-breaking research in aligning AI with human values: OpenAI: Set up by billionaire technologist Elon Musk of Tesla, OpenAI is a nonprofit research lab that is doing groundbreaking work on developing safe general intelligence. Headed by CEO and OpenAI co-founder Sam Altman, OpenAI is researching on developing a safe and friendly AI future to benefit humanity. With a primary focus on humanity, the institute is committed to working on AGI’s deployment to ensure that it is used for the benefit of all, and to avoid enabling uses of AI or AGI that can affect humanity. with that in mind, the institute has channelled all its resources on research to drive broad adoption of AI and make AGI safe. In their charter, the institute has outlined the guidelines for long-term safety about late-stage AGI development and also cooperating with other research and policy institutions to create a global community. Machine Intelligence Research Institute:  From developing better decision-making systems to making more reliable general-purpose AI systems, MIRI (earlier known as the Singularity Institute for AI) has been at the forefront of cutting-edge research and has been developing tools to design and analyse AI systems. At MIRI, the research is directed at developing fool-proof AI systems that are well-aligned with human goals. According to Stuart Russell, a MIRI research advisor and co-author of The Long-Term Future of Artificial Intelligence, robustness and safety should be integrated into mainstream research capabilities. MIRI’s technical research agenda clearly focuses on developing formal agent foundations for AI alignment which would help in developing the conceptual tools and theory that for engineering robustly beneficial systems in the future. The Allen Institute of Artificial Intelligence: AI2, as it is widely known, was started by Microsoft co-founder Dr Paul Allen and is steered by well-known researcher Dr Oren Etzioni with a focus on developing high-impact research for AI. Pegged as one of the largest non-profit AI organisations in the world, AI2 is committed to building responsible AI that will benefit humanity and work on some of the biggest human challenges. Set up in 2014, AI2 focuses on computer vision, machine reading and NLU. It has also developed over dozen AI applications and technologies. Earlier last year, the AI2 joined hands with Partnership on AI, a noted consortium which is laying down the groundwork for best practices for advancing public perception of AI. Wadhwani Institute for Artificial Intelligence: The Wadhwani Institute for AI which was set up by tech entrepreneur brothers, Dr Romesh Wadhwani and Sunil Wadhwani in Mumbai earlier this year to advance the development of AI for good is India’s first independent nonprofit research institute. It was started with a vision to pursue AI and ML research in domains such as education, healthcare, infrastructure, financial inclusion, and agriculture. Besides fostering a research environment, the institute also aims to develop collaboration between AI scientists from top institutes as well as the government. The other agenda is to create accessible datasets and foster dialogue on the ethics and guidelines for AI development. Future of Humanity Institute: This research centre at the University of Oxford is led by Dr Nick Bostrom, a well-known researcher and author of Superintelligence: Paths, Dangers, Strategies. From assessing game theory, existential risk to the assessment of AI and its long-term impact, FHI Governance of AI program explores the ethical, social and political dimensions of AI and also tracks the current applications of AI in military and cybersecurity among other areas. Bostrom, who has been vocal about the danger of AI has often stressed to “build up the technology and also understand the science of how to predict and control advanced artificial agents”. Last Word Besides these institutes, leading tech giants such as Google, Microsoft, Adobe and Facebook, among others, have their own research arms such as FAIR, DeepMind, Microsoft Research and Adobe Research which also carry out groundbreaking work in machine learning and artificial intelligence. Their researches and findings are published on open-source platforms as well as in top journals.","excerpt":"Tech companies across the globe are becoming more and more aware of the power of artificial intelligence and singularity. Not-for-profit institutes are thus conscientiously working on the strategic implications and the openness of artificial intelligence. From subjects such as ethics and policy in AI to the development of machine intelligence, not-for-profit research institutes are working […]","categories":["AI Trends"],"tags":["AI for Humanity","OpenAI"],"author_name":"Richa Bhatia","publish_date":"2018-06-26T04:31:12","publication_year":"2018","word_count":741,"keywords":["Go","AI for Humanity","artificial intelligence","machine learning","OpenAI","AI","ML","computer vision","Aim","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","OpenAI","Aim","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-non-profit-artificial-intelligence-machine-learning-institutes-that-are-working-on-making-ai-safe\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10138410,"title":"K&#8217;taka Govt Supports Games24x7 to Boost Innovation in Gaming and AI","content":"Games24x7, an Indian online gaming company, recently announced the launch of the second edition of its accelerator program, TechXpedite, with the support of the Karnataka government. With AWS as its cloud partner, this 60-day accelerator program is designed to empower startups in gaming, AI, and inclusive technology. According to a NASSCOM report, Bengaluru is home to over 7,000 startups, solidifying its status as India’s leading startup hub and accounting for 20% of the country’s overall startup activity. Speaking to AIM Media House at IGIC 2024, Sanjeev Kumar Gupta, the CEO of Karnataka Digital Economy Mission, noted, “There are over 1,000 AI startups in the state. With initiatives like Beyond Bengaluru, we are also promoting AI startups in Mysuru, Hubli, and Mangaluru.” Earlier, at London Tech Week, state IT\/BT minister Priyank Kharge unveiled a new scale-up program called Hypergrowth Global Karnataka. It aims to accelerate the global commercialisation and international market expansion of the best tech companies in Bengaluru and Karnataka. The program offers local companies access to global mentors, expert scaling advice from leading executives, go-to-market support, and connections to potential new customers and investors to enhance their international growth capabilities. Currently, many accelerator programs are functional in India including NVIDIA’s Inception program, India Accelerator, JioGenNext, and AWS. But coming down to online real money or skill gaming, there has been significant growth only in recent years. Games24x7’s Accelerator Program At Games24x7, their founding thesis since 2006 has been to connect people with the games they love and offer gaming on its platforms. It took them three years (2006 to 2009) to launch the first product, which is almost unheard of today when startups are incubated in months. Last year, it launched GameTech Accelerate, initially focused on gaming, but soon realised that technology, particularly AI and machine learning, is central to everything. In an exclusive conversation with AIM, Bhavin Pandya, co-CEO & co-founder, Games24x7, said, “Today, 77% of Indian startups use technologies like AI, machine learning, IoT, or blockchain. That’s why we wanted to broaden our focus and contribute more broadly to the tech startup ecosystem, not just gaming.” This led them to launch TechXpedite, the next phase of their accelerator program, which will expand to more cities and focus on deep technology. “We’re now tackling challenges like designing integrated circuits and chips for use in various devices, addressing global problems. For years, the Indian tech scene has focused on solving problems from the West—like how Flipkart emerged as an answer to Amazon, or Ola to Uber. But now, we are at a stage where we can solve uniquely Indian challenges,” Pandya said. Second Edition of TechXpedite Accelerator Program In December 2023, in partnership with the Government of Karnataka, the ‘GameTech Accelerate – The Future of Gaming’ accelerator programme was launched to nurture and propel promising ventures offering innovative and viable solutions for India’s evolving online gaming industry. The programme received over 200 applications resulting in the selection of 15 emerging startups across four key focus areas, including New Games (Real Money, Casual, Hyper Casual, Educational, Simulation, Strategy, and MMO); Real-Time Analytics, Telemetry & Personalisation; Security and Anti-Cheating Measures; and Community and Social Integration. Speaking at the launch event, Kharge said, “Initiatives like TechXpedite are the kind of catalysts that will go a long way in accelerating the next phase of innovation, helping early-stage startups grow into global leaders.” For this year’s edition, applications to the program are open till November 10. The program will culminate in a pitch event in February 2025, where selected startups will present their solutions to a panel of investors, industry experts, and domain leaders. Pandya highlighted, “Startups are the future. By 2030, the number of startups is expected to grow by 2.6 times, from 90,000 today to over 2,50,000. This growth will only be possible if we continue investing in the startup ecosystem.” Karnataka Stands Tall The state recently signed a letter of intent with the World Economic Forum (WEF) to establish a dedicated AI centre, aiming to establish itself as a global AI hub. Further, the state government signed eight MoUs with global firms at the WEF meet in Davos, securing investments worth INR 23,000 crore across AI, citizen services, sustainability, and e-governance. Moreover, the emergence of companies like Sarvam AI, which focuses on developing advanced AI models, reflects Karnataks’s commitment to technological innovation. Krutrim AI, led by Ola founder Bhavish Aggarwal, has raised $50 million in funding at a $1 billion valuation and is now India’s first AI startup to reach unicorn status. Innovation in AI isn’t limited to corporate giants. Startups like KOGO OS are disrupting markets with their AI operating systems, offering modular AI assistants tailored for diverse industries. Meanwhile, Karya AI is pioneering in the rural employment sector, leveraging AI to create job opportunities through tasks in local languages, with notable partnerships with tech giants like Microsoft and Google. These initiatives underscore the state’s status as a hub for AI innovation, promising transformative impacts across multiple sectors.","excerpt":"Bengaluru is home to over 7,000 startups, solidifying its status as India’s leading startup hub and accounting for 20% of the country’s overall startup activity.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","India"],"author_name":"Vidyashree Srinivas","publish_date":"2024-10-16T15:00:00","publication_year":"2024","word_count":828,"keywords":["Go","AI assistants","machine learning","AWS","AI","Git","RAG","Aim","analytics","R","India","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","AI assistants","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/karnataka-govt-fuels-ai-revolution-with-1000-ai-startups-already-in-the-game\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10009634,"title":"Register For Webinar: How To Future-Proof And Advance Your Career In The New Normal","content":"With the world grappling with COVID-19, digital transformation has become a key driver for many economies and businesses globally. There is an increasing trend in the adoption of technologies such as data science, cybersecurity, cloud computing, deep learning, IoT, automation, virtual reality and others, as organisations look to re-imagine the evolution of their workflows and processes. Not only has this crisis brought in the new normal of remote working, it has also pushed analytics to the forefront. As businesses look to emerge stronger while adopting the much-in-demand technologies, the need for professionals who are equipped with these skills is on the rise. In fact, according to one of our reports, there are 93,500 open jobs in data science in India alone. To keep up with this increasing demand, professionals should equip themselves with these emerging skills to accelerate their career in the new normal. To help candidates future-proof their jobs and advance their career in the new normal, Analytics India Magazine in association with the National University of Singapore is organising an exclusive webinar to detail the essential skill sets that are critical to recover business models. The webinar will also focus on recommending programmes that are specially designed to ensure that you thrive in the new normal. What Will The Webinar Cover? Participants will gain insights on relevant skill sets and technologies which are magnified due to the COVID-19 situationParticipants will get to explore graduate programmes with the industry-relevant curriculum taught by the finest faculties at Asia’ Top University (QS World University Rankings 2021)This session will also give participants insights to the National University of Singapore’s MSc in Industry 4.0 and MSc in Data Science and Machine Learning Register Here Who Will Be The Session Speakers? Associate Professor Goh Puay Guan Associate Professor Goh Puay Guan is the Academic Director of the MSc in Industry 4.0, at the National University of Singapore. He has diverse academic and industry experience, holding an MS in Management Science and Engineering from Stanford University and many publications in esteemed journals. Professionally, Associate Professor Goh has served in various senior management roles for global MNCs, holding specific expertise in areas such as technology impact on supply chains, globalisation of Asian companies and more. Professor Zhang Louxin Professor Zhang Louxin is the Academic Director of the MSc in Data Science and Machine Learning, at the National University of Singapore. He has a PhD in Computer Science from the University of Waterloo, Canada. Professor Zhang has been with the Department of Mathematics at the National University of Singapore since 2001. His current research interests include the application of data science in biomedicine, complex networks and infectious disease propagation modelling. He has served as the project committee member\/chairman of several international bioinformatics conferences such as the International Conference on Molecular Biology Intelligent Systems and the European Conference on Computational Biology, and served as the editorial board\/guest editor of eight international bioinformatics journals. Register Here Webinar Details Topic: Accelerate your Career during the New Normal: Recommended Programmes designed to future-proof your career Date: 5th November (Thursday) Timing: 6pm – 7pm (IST); 8.30 – 9.30pm (SG time) Register Here","excerpt":"With the world grappling with COVID-19, digital transformation has become a key driver for many economies and businesses globally. There is an increasing trend in the adoption of technologies such as data science, cybersecurity, cloud computing, deep learning, IoT, automation, virtual reality and others, as organisations look to re-imagine the evolution of their workflows and […]","categories":["Deep Tech"],"tags":["ai iot and future","analytics career","Data Science Career","msc data science and analytics","National University of Singapore","supply and demand in cyber security","webinar"],"author_name":"Srishti Deoras","publish_date":"2020-10-14T10:00:37","publication_year":"2020","word_count":520,"keywords":["data science","Go","machine learning","AI","cloud computing","ai iot and future","Git","supply and demand in cyber security","Data Science Career","National University of Singapore","analytics career","msc data science and analytics","deep learning","analytics","GAN","R","webinar"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","analytics","cloud computing","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/register-for-webinar-how-to-future-proof-and-advance-your-career-in-the-new-normal\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075764,"title":"OpenAI Open-Sources ‘Whisper’ —  a Multilingual Speech Recognition System","content":"Speech recognition remains a challenge in AI. However, OpenAI has just moved one step closer to solving it. In a blog post last week, OpenAI introduced Whisper—a multilingual, automatic speech recognition system that is trained and open sourced to approach human level robustness and accuracy on English speech recognition. Numerous organisations such as Google, Meta and Amazon have developed highly capable speech recognition systems. But OpenAI claims that Whisper stands out. The model is trained on 680,000 hours of multilingual and multitask supervised data collected from the web. It claims to have an improved recognition of background noise, unique accents, and technical jargon owing to the use of such a large and diverse dataset. The company’s open-sourced models and inference code serve as a foundation for building useful applications and boost further research on robust speech processing. Source: Introducing Whisper, OpenAI An excerpt from the blog reads, “The Whisper architecture is a simple end-to-end approach, implemented as an encoder-decoder Transformer. Input audio is split into 30-second chunks, converted into a log-Mel spectrogram, and then passed into an encoder. A decoder is trained to predict the corresponding text caption, intermixed with special tokens that direct the single model to perform tasks such as language identification, phrase-level timestamps, multilingual speech transcription, and to-English speech translation.” The company says that other existing approaches frequently use smaller, more closely paired audio-text training datasets or broad but unsupervised audio pretraining. Since Whisper was trained on a large, diverse dataset (about a third of which is non-English audio dataset) without being fine-tuned to any specific one, it does not beat models that specialise in LibriSpeech performance. When measured, findings show that Whisper’s zero-shot performance across many diverse datasets is robust—making 50% fewer errors than other models. OpenAI hopes that the model’s ease of use and high accuracy will allow developers to add voice interfaces to a wider set of applications. To learn more about the paper, model card, and additional details on Whisper, click here.","excerpt":"The company’s open-sourced models and inference code serve as a foundation for building useful applications and boost further research on robust speech processing.","categories":["AI News"],"tags":["Amazon","Google","Meta","OpenAI","Whisper"],"author_name":"Bhuvana Kamath","publish_date":"2022-09-26T12:37:36","publication_year":"2022","word_count":329,"keywords":["Whisper","Go","Meta","OpenAI","AI","programming_languages:R","Amazon","programming_languages:Go","Aim","Google","GAN","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-open-sources-whisper-a-multilingual-speech-recognition-system\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049854,"title":"A Guide to VARMA with Auto ARIMA in Time Series Modelling","content":"Time series modelling needs a series of steps to be performed such as processing the time series data, analyzing the data before modelling with different types of tests and then finally modelling with the data. There are different types of tests and modeling techniques used based on the types of requirements. There are different modelling techniques based on the Moving Average (MA) of the time series, each of them has its own advantages. Here in this article, we are going to discuss such a model named VARMA (Vector Auto Regressive Moving Average) and we will see how it can be implemented with the Auto-ARIMA model to achieve some specific types of results. The major points to be covered in this article are listed below. Table of Contents What is ARIMA? Why Do We Need Auto-ARIMA? What Is VARMA(Vector Auto Regressive Moving Average)? What is VARMA With Auto ARIMA? Implementing VARMA With Auto ARIMA Let us begin with understanding the ARIMA model. What is ARIMA? Before going for the Auto-ARIMA we need to understand what the ARIMA model is? In time series analysis, the ARIMA model is a model made up of three components: Auto-Regressive(AR), Integrated(I), and Moving Averages(MA). p: Stands for the number of lag observations included in the model, also known as the lag order. d: The number of times the raw observations are differentiated, also called the degree of difference. q: Is the size of the moving average window and also called the order of moving average. Before implementing the ARIMA model it is assumed that the time series we are using is a stationary time series and a univariate time series. To work with the ARIMA model we need to follow the below steps: Load the data and preprocess the data. Check the stationarity of the data.- if stationary then proceed for the further steps and if not then make it stationary. determine the degree of differencing(d). Determine the order of lag(p) and moving average(q), which can be done by making a PACF(partial autocorrelation function) and ACF(autocorrelation function) plot. Fitting the model and making the prediction. Check the performance of the model by calculating RMSE(root mean square error) between the actual and predicted values. So hereby the full procedure we can understand that following all the steps can become time-consuming. You can check the full implementation of an ARIMA model in this article. To save us from this situation auto-ARIMA comes into the picture. Why do we need Auto-ARIMA? As we know, ARIMA models are a very powerful tool for us in time series analysis but on the other hand, we are required to analyze model prediction on the basis of p, q, and d values here Auto-ARIMAis a model using which we can save our time in iterating between the p,q and d values. It helps in finding the best combination of these values and fitting them into the ARIMA model. A combination of these values can be estimated by estimating the AIC (Akaike information criterion) and BIC(Bayesian Information Criterion). Lower AIC and BIC values give the best combination of p, q, and d. Here, by providing the best combination, the Auto-Arima model saves us from performing some of the steps in the ARIMA modelling procedure. In the above, we learned that an ARIMA or Auto-ARIMA model is a powerful tool when working with the univariate time series. So when we talk about a multivariate time series VARIMAX models come into the picture. What is VARMA (Vector Auto Regressive Moving Average)? When the scenario comes on the modelling for a multivariate time series we can use different models like VAR and VMA and VARMA. The Vector Auto-Regressive(VAR) model is a generalization of the auto-regressive model for multivariate time series where the time series is stationary and we consider only the lag order ‘p’ in the modelling. The Vector Moving Average(VMA) model is a generalization of the Moving Average Model for multivariate time series where the time series is stationary and we consider only the order of moving average ‘q’ in the model. The Vector Autoregressive Moving Average(VARMA) model is a combination of VAR and VMA models that helps in multivariate time series modelling by considering both lag order and order of moving average (p and q)in the model. We can make a VARMA model act like a VAR model by just setting the q parameter as 0 and it also can act like a VMA model by just setting the p parameter as 0. What is VARMA With Auto ARIMA? From the above-given information, we can understand what is going to happen next in the article. As we have discussed, ARIMA is time-consuming because of the procedure for finding the best combination of parameters p,q, and d and which makes us introduce the auto-ARIMA model which helps in finding the best-fit combination of parameters for better performance of the model. Here we have the VARMA model where we have two parameters p and q, again if we go with the simple model obviously we will find it problematic in case of finding the best combination of parameters for better performance of the model. In these situations, we can try a model from which we can get the combination of p and q parameters easily and without wasting time in iteration with random parameters we can fit them in the VARMA model and increase the performance of the model in less time. So here we try to find the best-fit combination of the parameters by Auto-ARIMA and we use those parameters in the VARMA model for predicting the forecast values. Let’s see how we can implement this. To reduce the size of the article, many of the coding parts are presented in brief. For complete codes of VARMA with Auto ARIMA, you can check out this link to find the Colab notebook. Implementing VARMA With Auto ARIMA First of all, we are importing some of the basic libraries which we will require in the procedure. So that we can know these packages are installed in our system. import pandas as pd import matplotlib.pyplot as plt from statsmodels.tsa.statespace.varmax import VARMAX import numpy as np from statsmodels.tsa.stattools import adfuller from sklearn import metrics from timeit import default_timer as timer import warnings warnings.filterwarnings(\"ignore\") I am using the Dow_jones_Industrial_average data where we have an opening, closing, and high and low information of the shares of the Dow Jones industries. The reader can get this data from this link. 10 head values of the data are as follows: As we have discussed, to perform ARIMA modelling first we need to check whether the time series is stationary or not. I will perform a dickey-fuller test to check the stationarity of the four series. To know more about the test, you can go to this link. So the results of the test are as follows: Dickey-Fuller Test for Open Dickey-Fuller Test for High Dickey-Fuller Test for Low Dickey-Fuller Test for Close Here we can see that the series of data is not stationary in the Dickey-Fuller test we measure the p-value for the series in the value is lower than 0.02 then we can say the time series is stationary and if the p-value is greater than 0.02 which states that time series is not stationary. So here we need to make the time series stationary. There can be many methods to make the data stationary some of them are Detrending Seasonal adjustment Transformation To make it stationary I am using a differencing method which comes under the transformation methods. Where we just subtract a time series value from its only one lagged value. After differencing the dickey fuller test results are as follows: Dickey-Fuller Test for Open after Differencing Dickey-Fuller Test for High after Differencing Dickey-Fuller Test for Low after Differencing Dickey-Fuller Test for Close after Differencing Here we can see all the series in the data set are now stationary. After this, we can perform the Johansen cointegration test in multivariate time series to see if the series is correlated to each other or not. The Johansen cointegration test results are as follows: Here we can see that the multivariate time series we are using are correlated. Now we can apply the Auto ARIMA model. Which will tell us the order of p and q for our VARMA model. To apply the Auto-ARIMA model on time series data using the prima library. We can import the model for Auto-ARIMA like this from pmdarima import auto_arima The results for p and q are as follows: Order of p and q for Open Order of p and q for High Order of p and q for Low Order of p and q for Close Here these four optimal orders can be divided into two parts because we can not give the order of q as 0 and we can see that we have order values for high and low values of the stock. So here we should iterate the VARMA model between (1,0,1) and (1,0,0). As we have talked about, our major concern for this modelling is to save our time from iterating the model with different p and q values. We can see in the above outputs that we have various values for p and q order. Finding them all from PACF and ACF plots could consume so much time. Now we have only two combinations and iterating the model for these two combinations took the following effort So the picture is clearer now by analyzing the RMSE score we can say that orders of p = 0 and q = 1 will give the best score and forecasting values for this multivariate time series. We can fit the model on a time series with p=0 and q=1 combinations. The forecasting results are as follows. Forecasting Results for High Here we can see our results which are very satisfactory. The codes for all this implementation are given in this link. Final Words In this article, we have seen what ARIMA is, why we use Auto ARIMA and how we can use the Auto ARIMA model with the VARMA model. The major points that we focused on in the article are the p, q, and d values. With them, we are required to learn how we can make the best combination of these values. Since Auto ARIMA makes this procedure of finding combinations easy for us, I suggest you find the combinations by analyzing the PACF and ACF plots also. These are very basic things which we should know about before going to understand the advanced techniques. There are various things and techniques that I used in the article that are not deeply explained. Below are the links to a few articles that will give you a deep understanding of those things. A Complete Tutorial on Time Series Filters General Overview Of Time Series Data Analysis Guide To AC and PAC Plots In Time Series Comprehensive Guide To Time Series Analysis Using ARIMA Hands-On Tutorial on Vector AutoRegression(VAR) For Time Series Modeling References Comprehensive Guide To Time Series Analysis Using ARIMA Hands-On Tutorial on Vector AutoRegression(VAR) For Time Series Modeling Pmdarima Google Colab Notebook for Codes","excerpt":"Time series modelling needs a series of steps to be performed such as processing the time series data, analyzing the data before modelling with different types of tests and then finally modelling with the data. There are different types of tests and modeling techniques used based on the types of requirements. There are different modelling […]","categories":["Deep Tech"],"tags":["ARIMA","Guide","Time Series","Time Series Analysis","time series modeling"],"author_name":"Yugesh Verma","publish_date":"2021-09-28T13:00:00","publication_year":"2021","word_count":1863,"keywords":["Time Series Analysis","Go","NumPy","ARIMA","TPU","programming_languages:R","AI","RAG","Colab","Matplotlib","Time Series","time series modeling","R","Guide","Pandas"],"extracted_tech_keywords":["AI","Colab","Pandas","NumPy","Matplotlib","RAG","TPU","R","Go","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-guide-to-varma-with-auto-arima-in-time-series-modelling\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10060692,"title":"Public vs private blockchains: How do they differ","content":"Blockchain technology is the fulcrum of the ‘next internet’. At a granular level, every ‘block’ is a part of a database that records information. Blockchain is divided into two types: Public and private. While public blockchains are decentralised peer-to-peer networks, the ledger is controlled by a centralised authority in private blockchains: Meaning, the main difference lies in the level of access given to users. What are public blockchains? Also known as permissionless blockchains, public blockchains are completely open and follow the idea of decentralisation to the T. Bitcoin and Ethereum are both examples of public blockchains. Anyone in the network can access the chain and add blocks. Public blockchains are also largely anonymous, unlike private blockchains, where the identity of the people involved in the transaction is not kept hidden. Advantages Security: The biggest advantage of public blockchains is how secure they are. A transaction recorded in a chain is immutable because it cannot be altered or removed, only added. Also, the validity of a transaction is recognised when the majority agree that the transaction is valid, making blockchain immune from external tampering. The more decentralised a blockchain is, the safer it is. A larger number of users makes it more difficult for hackers to band together and breach the network. Transparency: Since public blockchains are built using open source computing codes, the transactions are completely transparent and verifiable. Public blockchain is central to pushing the futuristic system of decentralised identity (DID). According to Microsoft, a decentralised identity is a trust framework in which identifiers, such as usernames, can be replaced with IDs that are self-owned, independent, and enable data exchange using blockchain to protect privacy and secure transactions. Normally, users have to register separately and go through a lengthy identity verification process to access bank accounts, Netflix accounts, or even obtain a driver’s licence. DID wallets will serve as a single and secure gateway to access all such services. Anonymity: Anonymity is one of the biggest draws of a public blockchain. The transaction is spread across the public ledger as bits of data and can’t be traced back to the original address of the users. Disadvantages Power consumption: Public blockchains like Bitcoin have algorithms that function on Proof-of-Work. Special nodes called miners compete to finish a transaction on the network for a reward. However, these transactions are highly energy-extensive and take a long time to complete. For example, Bitcoin manages to complete seven transactions every second compared to Visa, which can complete 24,000 transactions in a second. Block sizes offered on public blockchains are also limited because of how heavy they are on the resources. Scalability: Transaction speed compounds the scalability issues with public blockchains. The more the users are on a blockchain, the more it burdens the network with more transactions. Security: There is an exception to the rule regarding security in public blockchains. The transactions go through when the majority of miners agree on it. In this Proof-of-Work protocol, the attackers can prevent new blockchains from forming by gaining a 51% hash rate. According to the Bitcoin Nakamoto consensus rule, ‘the longest chain wins’. That said, a blockchain network as vast as Bitcoin is immune to 51% attacks. However, there have been malware attacks on smaller blockchains such as Bitcoin Gold and Ethereum Classic in 2018. Private vs public blockchains Private blockchains like Ripple and Hyperledger have the advantage of speed because a smaller set of users means less time to reach a consensus to validate a transaction. Private blockchains can process thousands of transactions every second and are easily scalable. A private blockchain has a centralised network that quickens the transaction process. Having a centralised network also raises the issue of trust, which is resolved in a public blockchain. A transaction’s validity cannot be verified on private networks and relies on the authorised nodes’ credibility. Additionally, fewer nodes make the network more susceptible to malicious attacks. The anonymity of public blockchains has also made it a major go-to transaction method for nefarious activities in the darknet, as it is difficult to trace the parties involved.","excerpt":"Also known as permissionless blockchains, public blockchains are completely open and follow the idea of decentralisation to the T.","categories":["Deep Tech"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-02-16T11:00:00","publication_year":"2022","word_count":679,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","Scala","ViT","programming_languages:Scala","Rust","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Rust","Scala","ViT","cloud_platforms:AWS","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/public-vs-private-blockchains-how-do-they-differ\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10052393,"title":"PyTorch Releases Drug Discovery Platform “TorchDrug”","content":"PyTorch recently announced the release of its machine learning drug discovery platform TorchDrug to accelerate drug discovery research. The library is open-sourced and can be installed through pip if you have PyTorch and torch-scatter installed using pip install torchdrug, or through conda conda install -c milagraph -c conda-forge torchdrug. TorchDrug covers many recent techniques such as graph machine learning, deep generative models, and reinforcement learning. It also provides reusable training and evaluation routines for popular drug discovery tasks, including property prediction, pretrained molecular representations, de novo molecule design, retrosynthesis and biomedical knowledge graph reasoning. It is easy to build a prototype for one’s own dataset and application based on these techniques and modules. For advanced users, the platform provides multiple levels of building blocks for different customisation demands. These include low-level data structures and operations (e.g. molecules and graph masking), mid-level layers and modules (e.g. graph convolutions and GNNs) and high-level task routines (e.g. property prediction). TorchDrug is flexible for all kinds of customisation. It also provides graph data structures and operations for manipulating biomedical objects, as well as reusable layers, models and tasks for building machine learning models. The core data structures of TorchDrug are graphs, which can be used to represent a wide range of biological objects, including molecules, proteins and biomedical knowledge graphs. Visualisation API in the library can be used to check graph objects. PackedGraph data structure, which builds a unified large graph and re-index each small graph in the batch, can be used to create a batch of variable-size graphs. Code for calculating a batch of 4 molecules: mols=data.PackedMolecule.from_smiles([\"CCSCCSP(=S)(OC)OC\", \"CCOC(=O)N\", \"N(Nc1ccccc1)c2ccccc2\", \"NC(=O)c1cccnc1\"]) mols.visualize() mols = mols.cuda() print(mols) # PackedMolecule(batch_size=4, num_nodes=[12, 6, 14, 9], num_edges=[22, 10, 30, 18], device='cuda:0') Image: TorchDrug Graphs also support a wide range of indexing operations. Typical usages include applying node masking, edge masking or graph masking. The optimiser can be used for parameters in the task and combine everything into the core. The engine provides convenient routines for training and testing. To test the model on the validation set, it only takes one line. TorchDrug is designed to cater to all kinds of development. This ranges from low-level data structures and operations, mid-level layers and models, to high-level tasks. One can easily customise modules at any level with minimal effort by utilising building blocks from a lower level. Image: TorchDrug The correspondence between modules and the hierarchical interface is : torchdrug.data: Graph data structures and graph operations; e.g. molecules. torchdrug.datasets: Datasets; e.g. QM9. Torchdrug.layers: Neural network layers and loss layers; e.g. message-passing layer. Torchdrug.models: Representation learning models; e.g. message passing neural network. torchdrug.tasks: Task-specific routines; e.g. molecule property prediction. Torchdrug.core: Engine for training and evaluation. Machine learning for drug discovery is a fast-growing area, and the PyTorch team expects that TorchDrug could help more and more people get involved in this interdisciplinary area. To learn more about TorchDrug, you can check out the Colab tutorials for basic usage and several drug discovery tasks using the link here.","excerpt":"TorchDrug covers many recent techniques such as graph machine learning, deep generative models, and reinforcement learning.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","libraries","library","Machine Learning","Python","Python for Data Science","Python Libraries","python machine learning","Python Programming","Pytorch"],"author_name":"Victor Dey","publish_date":"2021-10-26T15:04:08","publication_year":"2021","word_count":496,"keywords":["API","Python Programming","R","CUDA","Pytorch","PyTorch","Data Science","machine learning","AI","Python Libraries","library","neural network","Machine Learning","libraries","Python for Data Science","python machine learning","knowledge graphs","Python","Colab","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","neural network","PyTorch","Colab","knowledge graphs","CUDA","R","CUDA","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-releases-drug-discovery-platform-torchdrug\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10014474,"title":"EfficientDet: Guide to State of The Art Object Detection Model","content":"Object detection is a technique of training computers to detect objects from images or videos; over the years, there are many object detection architectures and algorithms created by multiple companies and researchers. In this race of creating the most accurate and efficient model, the Google Brain team recently released the EfficientDet model, it achieved the highest accuracy with fewest training epochs in object detection tasks. This architecture beats the YOLO, AmoebeaNet with minimum computation power. It is an advanced version of EfficientNet, which was the state of art object detection model in early 2019, EfficientNet was a baseline network created by Automl MNAS, it achieved state-of-the-art 84.4% more accuracy and used a highly effective compound coefficient to scale up CNNs in a more structured manner. EfficientNet Architecture: img There are many object detection techniques invented over the years some of the detection models are discussed here, but now EfficientDet has increased the bar and takes the accuracy and efficiency to new levels. EfficientDet EfficientDet is an object detection model created by the Google brain team, and the research paper for the used approach was released on 27-July 2020 here. As we already discussed, it is the successor of EfficientNet, and now with a new neural network design choice for an object detection task, it already beats the RetinaNet, Mask R-CNN, and YOLOv3 architecture. Also, the architecture of EfficientDet employs the ImageNet pre-trained EfficientNet as the backbone of the network. With EfficientNet-B3 as the backbone, it already increased accuracy by 3%It achieved 55.1 AP(average precision) on COCO test-dev that contains 77 million parameters.Their model runs 5x faster on CPU2x-4x faster on GPU compared to other models. EfficientDet purposed some of the new optimisation techniques to improve efficiency as follows: BiFPN& new Computational Scaling technique The overall architecture of EffiecientDet: Img Let’s understand the techniques that make this model so efficient. BiFPN As shown in the above diagram, BiFPN refers to a Bi-directional Feature Pyramid Network that can be enhanced with fast normalization and feature fusion. Basically, It is a type of pyramid network which allows easy multi-scale feature fusion. BiFPN idea was inspired by FPN(Feature Pyramid Network) where the information is inherently restricted by one-way information flow. Traditionally, FPN techniques treat all the feature inputs the same even they have different resolutions, and that causes an unequal output feature. PANet adds an additional feature of bottom-up flow at the cost of more computation. NAS(neural architecture search) discovered NAS-FPN architecture. However, this architecture was irregular. But in BiFPN the information can be flow in both top-down and bottom-up approaches, Bi-FPN added the additional weight for each input so that network can learn each feature input differently. Also, BI-FPN reduced the cross-scale connection by removing the nodes with a single input edge and added an extra edge to the output node if it’s on the same level. It treats each bidirectional path as one feature’s layer for making it more future fusion. With all these optimisations, BiFPN further improved the accuracy by 4% Model performance On evaluating EffiecientDet on the COCO dataset, it achieved mAP(mean average precision) of 52.2 with 9.4x less computation and exceed current State of the art (SOTA) models by 1.5 points. https:\/\/ai.googleblog.com\/2020\/04\/efficientdet-towards-scalable-and.html Implementation As the model is open-sourced, let’s use EfficientDet for some object detection tasks. There are many pre-trained models on EfficientDet is available on the internet Like Monk: a Computer vision toolkit for low-code, easily installable object detection pipelines. They have already created a wrapper for all the different applications like Wheat head detection in the field, underwater imagery object detection, person detection in infrared imagery, and more here. Multiple object detection using Google EfficientDet We are going to use the official python notebook provided by the Google  Automl team, for understanding the working. View source n GitHubRun in Google ColabDirect Run notebook Install packages, download source code, and images with this script below. It will clone the automl repository and install all the dependencies from the requirements.txt file. %%capture #@title import os import sys import tensorflow.compat.v1 as tf # Download source code. if \"efficientdet\" not in os.getcwd(): !git clone --depth 1 https:\/\/github.com\/google\/automl os.chdir('automl\/efficientdet') sys.path.append('.') !pip install -r requirements.txt !pip install -U 'git+https:\/\/github.com\/cocodataset\/cocoapi.git#subdirectory=PythonAPI' else: !git pull You can make a parameter passing interface inside your notebook by using command like this, we are using effiicientDet-D0 that has 34.6 mAP, 10.2ms-batch1latency, 97fps-throughput. View graph in Tensorboard !python model_inspect.py --model_name={MODEL} --logdir=logs &> \/dev\/null %load_ext tensorboard %tensorboard --logdir logs Let’s do the benchmark network latency, there are two main type of network & end-to-end latency, To ensure the network latency: from the first convolution to last class prediction output use the following code: !python model_inspect.py --runmode=bm --model_name=efficientdet-d4 --hparams=\"mixed_precision=true\" Calculating End to End latency: from the input image to the final rendered new image, including image m = 'efficientdet-d4' # @param batch_size = 1 # @param m_path = download(m) saved_model_dir = 'savedmodel' !rm -rf {saved_model_dir} !python model_inspect.py --runmode=saved_model --model_name={m} \\ --ckpt_path={m_path} --saved_model_dir={saved_model_dir} \\ --batch_size={batch_size} --hparams=\"mixed_precision=true\" !python model_inspect.py --runmode=saved_model_benchmark --model_name={m} \\ --ckpt_path={m_path} --saved_model_dir={saved_model_dir}\/{m}_frozen.pb \\ --batch_size={batch_size} --hparams=\"mixed_precision=true\" --input_image=testdata\/img1.jpg Inference images, below code will execute the automl inspect.py file with arguments model name, run mode, checkpoint path. # first export a saved model. saved_model_dir = 'savedmodel' !rm -rf {saved_model_dir} !python model_inspect.py --runmode=saved_model --model_name={MODEL} \\ --ckpt_path={ckpt_path} --saved_model_dir={saved_model_dir} # Then run saved_model_infer to make an inference. # Notably: batch_size, image_size must be the same as when it is exported. serve_image_out = 'serve_image_out' !mkdir {serve_image_out} !python model_inspect.py --runmode=saved_model_infer \\ --saved_model_dir={saved_model_dir} \\ --model_name={MODEL} --input_image=testdata\/img1.jpg \\ --output_image_dir={serve_image_out} \\ --min_score_thresh={min_score_thresh} --max_boxes_to_draw={max_boxes_to_draw} Display image using display module: from IPython import display display.display(display.Image(os.path.join(serve_image_out, '0.jpg'))) You can execute more outputs and detect objects by using the trained model and run the inference as follows: serve_image_out = 'serve_image_out' !mkdir {serve_image_out} saved_model_dir = 'savedmodel' !rm -rf {saved_model_dir} # Step 1: export model !python model_inspect.py --runmode=saved_model \\ --model_name=efficientdet-d0 --ckpt_path=efficientdet-d0 \\ --hparams=\"image_size=1920x1280\" --saved_model_dir={saved_model_dir} # Step 2: do inference with saved model. !python model_inspect.py --runmode=saved_model_infer \\ --model_name=efficientdet-d0 --saved_model_dir={saved_model_dir} \\ --input_image=img.png --output_image_dir={serve_image_out} \\ --min_score_thresh={min_score_thresh} --max_boxes_to_draw={max_boxes_to_draw} from IPython import display display.display(display.Image(os.path.join(serve_image_out, '0.jpg'))) Conclusion Google brain team developed the most powerful and efficient object detection algorithm so far we have seen a started code for object detection provided officially by google Github repository, Also, there are many Repository, and other framework implementations are also available regarding EfficientDet, some of the repository you should look to get more information about this framework are as follows: EffiecientDet implementation using Pytorch.Official repository by Google Automlhttps:\/\/github.com\/xuannianz\/EfficientDethttps:\/\/github.com\/signatrix\/efficientdet","excerpt":"Object detection is a technique of training computers to detect objects from images or videos; over the years, there are many object detection architectures and algorithms created by multiple companies and researchers. In this race of creating the most accurate and efficient model, the Google Brain team recently released the EfficientDet model, it achieved the […]","categories":["Deep Tech"],"tags":["Object Detection"],"author_name":"Mohit Maithani","publish_date":"2020-12-17T10:00:00","publication_year":"2020","word_count":1061,"keywords":["TPU","AI","neural network","PyTorch","ML","computer vision","RAG","Colab","object detection","Object Detection","TensorFlow"],"extracted_tech_keywords":["AI","ML","neural network","computer vision","TensorFlow","PyTorch","Colab","RAG","object detection","TPU"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/efficientdet\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10020687,"title":"In Conversation With Dr Alpan Raval, Head of Data Science At Wadhwani Institute for Artificial Intelligence","content":"This is the ninth article in the weekly series of Expert’s Opinion, where we talk to academics who study AI or other emerging technologies and their impact on society and the world. This week, we spoke to Dr Alpan Raval, Head of Data Science at Wadhwani Institute for Artificial Intelligence. Wadhwani AI is an independent, nonprofit research institute and global hub, developing AI solutions for social good. The institute works on various missions to develop and apply AI-based solutions for a broad range of verticals including healthcare and agriculture. Analytics India Magazine caught up with Dr Raval, to get insights into the challenges and opportunities in Wadhwani AI’s line of work. AIM: Where do your datasets come from and what are the major issues you face while collecting and working with datasets? Dr Raval: Our datasets have a mix of sources. Some of the data, for our public health projects in TB for example, comes from the government, while other data is collected by us. Major issues in the data, especially from rural areas, include variability, incompleteness, and noise.  For example, the subject’s age can be a noisy variable in rural settings. For image data, as in our anthropometry project, there are stark differences in lighting conditions across homes that the AI has to deal with. Many of the variables we analyse are impossible to collect in a complete manner across populations. There are always gaps. AIM: Since India is a country with a high demographic diversity, how do you ensure inclusivity in your data? Dr Raval: Significant sections of our data, across projects, come from rural areas. So ensuring adequate representation of rural populations is usually not a problem. There are other inclusivity issues, such as an adequate representation of under-represented minorities in rural areas that we are beginning to consciously grapple with. We are aware of these issues as we fine-tune and tweak our models and plan to address them on a case-by-case basis, as appropriate for the given problem. AIM: Tell us about your data ownership models. How do you ensure the data benefits the people it is collected from? Dr Raval: Our overall stand on data ownership is that the data is always owned by the data principal. Ownership does not transfer to the party that analyses the data, pre-processes it, or builds AI models on the basis of the data. Our solutions, almost by design, must benefit the people we collect data from since they will be most accurate, and therefore most useful, for those communities. We are currently either piloting or close to piloting our solutions in different areas and our pilots are mainly deployed in the communities where the data is collected from, with some representation from outside these communities to ensure that our models generalize well. AIM: What are the challenges you face while deploying AI models across India? Dr Raval: AI products, unlike traditional software products, have unique challenges in deployment. Shifts in data distributions can result in degradation of model accuracy with time, necessitating re-training of the models. Thus AI model application is inherently an iterative process in which a significant amount of feedback is required from on-the-ground deployment to continuously fine-tune and improve the model. This is the main technical challenge. There are non-technical challenges too. AI-based decision-making is an entirely novel concept in the social sphere, so adoption with sufficient trust represents a barrier to entry. We must ensure privacy is preserved both in the training of the model as well as in the deployment. We must ensure, as best as we can, that the decisions are not flagrantly biased. All of these aspects are related to building trust in what we do. AIM: How does Wadhwani AI overcome these challenges? Dr Raval: We see many of these challenges facing us in the future as we scale up our solutions, and we aim to address them by systematically addressing potential areas of concern during the model development process. We don’t see significant roadblocks as yet, but there are always unknown unknowns. AIM: How advanced are ‘AI for social good’ models in India and how can they improve? Dr Raval: Central and state governments have been very supportive of what we do. AI for social good in practice is still a nascent field, not just in India, but across the world. What we need to create is examples of cases where AI has had a clear, sustained, positive societal impact at scale. These cases will spur the development of all kinds of support mechanisms.","excerpt":"This is the ninth article in the weekly series of Expert’s Opinion, where we talk to academics who study AI or other emerging technologies and their impact on society and the world. This week, we spoke to Dr Alpan Raval, Head of Data Science at Wadhwani Institute for Artificial Intelligence. Wadhwani AI is an independent, […]","categories":["AI Features"],"tags":["AI for social good","data privacy use cases","Interviews and Discussions","Wadhwani AI"],"author_name":"Kashyap Raibagi","publish_date":"2021-02-23T15:00:00","publication_year":"2021","word_count":755,"keywords":["data science","Go","artificial intelligence","programming_languages:R","AI","data privacy use cases","Wadhwani AI","Aim","ViT","analytics","AI for social good","Rust","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","R","Go","Rust","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/in-conversation-with-dr-alpan-raval-head-of-data-science-at-wadhwani-institute-for-artificial-intelligence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":29026,"title":"Google CEO Sundar Pichai Visits Pentagon To Diffuse Tension Over Drone Project: Reports","content":"Only four months after the employee outrage at Google which prompted the search engine giant to end their drone programme contract with the Pentagon, CEO Sundar Pichai reportedly visited Washington DC to “smooth things over”. According to a noted US newspaper, Pichai met important people from the office of the Undersecretary of Defense for Intelligence, the Defense Department directorate which oversees the artificial intelligence drone system called Project Maven. Project Maven, also known as the Algorithmic Warfare Cross-Function Team, was developed with an aim to integrate computer-vision algorithms to help military and civilian analysts with full-motion video data to help with counterinsurgency and counterterrorism operations. But more than 3,000 Google employees had signed a petition against the company’s involvement in the project because they thought that the study and the imagery could eventually be used to improve drone strikes in the battlefield. Similarly, employees at Microsoft posted a letter to CEO over the use of technology by law enforcement agencies. One of the biggest voices in this debate has been Tesla’s Elon Musk, who has rallied against the use of killer robots. A Pentagon spokesperson told The Washington Post, “We do not comment on the details of private meetings. Department leaders routinely meet with industry partners to discuss innovative technologies. These meetings support continuing dialogue aimed at solving future technology challenges.” AI researchers have repeatedly emphasised using the tech for good and are striving to promote a culture of responsibility. In fact, many tech organisations have taken a definitive stance on political and security threats caused by AI and have laid down frameworks to build a security landscape for an AI-enabled future. In the light of the dual use of AI and ML technologies, eminent institutes such as MIT, Stanford and Oxford have published risk-assessment technical studies on the dual use of AI technologies. Non-profit research companies such as Open AI and the Partnership of AI are also striving ahead in pushing research for social good.","excerpt":"Only four months after the employee outrage at Google which prompted the search engine giant to end their drone programme contract with the Pentagon, CEO Sundar Pichai reportedly visited Washington DC to “smooth things over”. According to a noted US newspaper, Pichai met important people from the office of the Undersecretary of Defense for Intelligence, the […]","categories":["AI News"],"tags":["drones","Google","pentagon","Sundar Pichai"],"author_name":"Prajakta Hebbar","publish_date":"2018-10-08T12:33:19","publication_year":"2018","word_count":326,"keywords":["Go","pentagon","artificial intelligence","drones","AI","programming_languages:R","ML","RAG","Aim","Google","GAN","AI research","R","Sundar Pichai"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","RAG","R","Go","GAN","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sundar-pichai-visits-pentagon-project-maven\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":60886,"title":"How To Avoid Common Mistakes That May Lead To Failure Of Data Projects","content":"According to a report, 85% of data projects fail, and many factors contribute to it. With the widening skills gap, and the growing sophistication of analytics tools and software, many projects tend to fail. With so many companies adopting big data analytics, 85% is a massive number that cannot be ignored. Below, we look at some of the problems – along with their possible solutions – that are common when it comes to a data project’s failure:- Project & Data Not Aligned With Business Needs The business structure should be framed in accordance with the needs and challenges of the specific analytics problem that the organization is looking to solve. Oftentimes, failure happens because organizations and individuals working on the data project fail to define and align the data with the business needs. One can counter these problems by involving subject experts with strong analytical skills and background knowledge. Hiring data scientists who can help define the problem early on in the project will be helpful. Complicated Tools Big data and analytics are complicated and should make sense and be accessible to business users. An organization should provide tools that are simple and easy to use for its business analytics team. This team should be able to leverage these tools for analytics, visualization and data discovery. Also, care must be taken to ensure that non-technical business users are not burdened with the task of using programmer-level tools. Easy and simple operations make it better for business teams to handle work, and come up with desired results. Lack Of Emphasis On Data Lakes Analytics involving big data deals with massive sets of data. Hence, organizations need to give due importance to storage. Although there are many on-premise technologies and cloud systems in place to manage this, storage is not enough when it comes to dealing with distinct types of data. For this, organizations need to give importance to data lakes. A data lake holds several sources of data and contains many data types, which makes it easy to manage massive amounts of data associated with a data project. However, a data lake must not be used to dump every type of data in it, and should be used in a more meaningful manner. Not Prioritizing Quality Not having good data quality and a data management system are two of the most significant factors that lead to failures in data projects. Failure to search, curate and model data properly will only result in faulty analytics. An organization has to put systems in place which enhance the accuracy of the data, and ensure that the data is up-to-date and is delivered on time. Ignore Key Facets Of Security Most data involved in a data project is from clients. Some of it is identifiable and personal as well. So keeping this information secure while working on the project becomes critical. It would be unacceptable if people outside the project get hold of the data. The security measures should include establishing necessary enterprise tools, using data encryption, policy enforcement, and training on the use and access of data. Not Hiring Good Consultants When Needed Sometimes, organizations may have to decide between building an in-house team or hire a consultant. However, this depends on the kind of budget available, and the software the organization uses. When an organization hires a consultant, it is often on a long-term basis because of the potential knowledge transfer that could happen. A consultant can help organizations define their needs, and develop a plan to meet its goal.","excerpt":"According to a report, 85% of data projects fail, and many factors contribute to it. With the widening skills gap, and the growing sophistication of analytics tools and software, many projects tend to fail. With so many companies adopting big data analytics, 85% is a massive number that cannot be ignored.  Below, we look at […]","categories":["AI Features"],"tags":[],"author_name":"Sameer Balaganur","publish_date":"2020-04-04T18:00:00","publication_year":"2020","word_count":587,"keywords":["big data","Go","programming_languages:R","AI","RAG","analytics","data quality","GAN","R","data lake"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","data lake","data quality","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-avoid-common-mistakes-that-may-lead-to-failure-of-data-projects\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":35598,"title":"6 Machine Learning Tools That Will Help You Find Code Bugs","content":"One of the many direct applications of machine learning is the use of machine learning tools to find the bugs in programs, without executing the programs. There have been a number of tools developed and released in recent times for this luxury to programmers. Here are the top six examples of ML tools that can assist you to find code bugs in your program. 1. DeepCode An AI software platform called DeepCode has a tool for analyzing and improving code for programmers. The system uses a corpus of 2,50,000 rules, reads the GitHub repositors of the user and tells them how to fix problems. It remains compatible and generally improves the programs. The tool currently supports Python, JavaScript and Java and assists programmers with finding hidden bugs and improving their code. “We have a unique platform that understands software code the same way Grammarly understands written language,” said Boris Paskalev, one of the founding members. “This unique proposition is positioned us save billions of dollars within the software development community with our first service and then to be on the front end of transforming the industry towards fully autonomous code synthesis.” 2. Clever-Commit: In a bid to cut the number of coding errors made in its Firefox browser, Mozilla is deploying a machine-learning-driven coding assistant developed in conjunction with Ubisoft, called Clever-Commit. Clever-Commit analyzes code changes as developers commit them to the Firefox codebase. It compares them to all the code it has seen before to see if they look similar to code that the system already identifies as a bug. If the tool thinks that a commit looks like the bug, it warns the developer. It can also give suggestions as the solutions for the bugs that it finds. Initially. It works with C++, JavaScript, and Rust, which are the languages hat Mozilla uses for Firefox. The tool builds on work by Ubisoft La Forge, Ubisoft’s research lab. Mozilla plans to use Clever-Commit during code reviews, and in time this will expand to other phases of development, too. 3.IntelliCode: InteliCode is a tool by Microsoft which is on its Visual Studio. IntelliCode is used to find bugs and can also detect improperly used variables. It saves time by adding what the user is most likely to add at the top of the compilation list. It also has recommendations based on open source projects on GitHub each with over 100 stars. When combined with the context of the existing code, the completion list is tailored to promote common practices. IntelliCode isn’t limited to statement completion. Signature help also recommends the most likely overload for your context. It’s also being used to detect coding styles and whitespace usages to format the code in a way that it looks consistent with the rest of the program. IntelliCode has examined some of the most popular public GitHub repositories, more than 2,000 projects each with more than 100 stars, to figure out best coding practices. 4.SapFix: SapFix is an AI hybrid tool created by Facebook engineers for debugging. The tool can suggest fixes for bugs in the code after which it proposes them to the programmers for their appoval and deployment. SapFix has been used to accelerate the process of shaping robust, stable code updates to millions of devices using Facebook Android app which is the first such use of AI-powered debugging at this scale. It also speeds up the process of rolling out new software. SapFix can create patches that either fully or partially revert the code submission that introduced them. For more complex crashes, the system generates patches by drawing from its collection of templated fixes. These templates are generated based on a pool of past fixes. When previously used human-designed templates don’t fit, SapFix will attempt a mutation-based fix, whereby it performs small code modifications to the abstract syntax tree (AST) of the crash-causing statement, making adjustments to the patch until a potential solution is found. A graphic illustrating how SapFix generates patches for software bugs. Image credit: Facebook. A graphic illustrating how SapFix generates patches for software bugs. Image credit: Facebook. 5.Sapienz: Deployed in September 2017, Sapienz is a tool again developed by Facebook based on AI which helps SapFix find and fix code, before reaching the production. Along with Facebook’s Infer static analysis tool, it helps to localise the points in the code to patch. Once both the tools identify a particular part of the code associated with a crash, it passed the information to SapFix, which picks from a few strategies to generate a patch. The tool automatically designs, runs and reports the results of tens of thousands of test cases every day on the app. In the first few months since its deployment, the technology has allowed engineers to fix issues within hours, and even minutes of the code being written. It has tested millions of lines of code in Facebook’s Android app. 6. Commit Assistant: The Commit-Assistant aims to identify patterns in past bugs to better intercept new bugs. It will allow teams to save on debugging time and focus on the creation of quality features. It was made with a code of 10 years worth from Ubisoft’s software library, to speed up the company’s development process. This was specifically for predicting bugs in a game’s code before even the error is committed.","excerpt":"One of the many direct applications of machine learning is the use of machine learning tools to find the bugs in programs, without executing the programs. There have been a number of tools developed and released in recent times for this luxury to programmers. Here are the top six examples of ML tools that can […]","categories":["AI Trends"],"tags":["languages","Machine Learning","model"],"author_name":"Disha Misal","publish_date":"2019-03-01T04:29:14","publication_year":"2019","word_count":887,"keywords":["machine learning","Rust","AI","ML","Machine Learning","Python","languages","model","Aim","C++","JavaScript","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","Python","R","JavaScript","Rust","Java","C++"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/6-machine-learning-tools-that-will-help-you-find-code-bugs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":12592,"title":"Uber introduces Movement, A BI tool based on its anonymized data from over 2 billion trips","content":"San Francisco-based cab aggregator Uber is in the news again and this time for making use of its terabytes of ride data for an online tool called Movement.  The underlying principle of Movement is to make moving in cities efficient by mapping points such as rush hours, ride durations, shutdowns and places travelled to. The data can be used by governments and local authorities in improving city’s infrastructure by understanding commute patterns. Uber Movement is anonymizing data by divvying it up into geographic zones which can further be used by city planners in improving transportation infrastructure. The tool is not divulging specific rider detail such as routes and departure points. Even then, the news has raised several eyebrows over privacy concerns and making user databases accessible to public. Detractors worry over the privacy of the user data shared and whether anonymization would truly protect user’s identity. According to the Movement page, “Making our cities move more efficiently and grow in way that works for everyone. That’s why we’re providing access to anonymized data from over 2 billion trips to help improve urban planning around the world.” This is how Uber movement tool plans to revolutionize cities with ride data: City officials:  Can gain historical insights that will enable them to measure the impact of road improvements, closures and and more. Planners and policymakers: policymakers can now make planned investment based on detailed analysis of commute and transportation patterns that will help in making informed decisions. General public: The data will be made accessible to everyone and Uber strongly believes it will lead to groundbreaking insights and ideas from everywhere. The news was meted with a positive response from policymakers. Muriel Bowser, Mayor, Washington DC, USA welcomed the Uber Movement  tool by sharing  this on the Movement page, “In today’s time, smart technology and intelligent use of data is critical to our success and District of Columbia is committed to using these tools to keep pace with the rapid growth of our neighborhoods. We are truly excited to partner up with Uber on this new platform. We would like to make use of data sources to reduce traffic congestion, improve infrastructure, and make our streets safer for every visitor and resident”.","excerpt":"San Francisco-based cab aggregator Uber is in the news again and this time for making use of its terabytes of ride data for an online tool called Movement.  The underlying principle of Movement is to make moving in cities efficient by mapping points such as rush hours, ride durations, shutdowns and places travelled to. The […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-02-07T10:27:57","publication_year":"2017","word_count":370,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/uber-introduces-movement-bi-tool-based-anonymized-data-2-billion-trips\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45829,"title":"Python Vs Scala: Which Language Is Best Suited For Data Analytics?","content":"Python and Scala are two of the most popular languages used in data science and analytics. These languages provide great support in order to create efficient projects on emerging technologies. In this article, we list down the differences between these two popular languages. Python Python continues to be the most popular language in the industry. Python, the open-source programming language has been widely used as a scripting and automation language. There are a number of features which makes Python popular among the list of toolkits of a developer. Python is powerful, fast, easy to learn and use. It has efficient high-level data structures and a simple but effective approach to object-oriented programming. The Python interpreter and the extensive standard library are freely available in source or binary form for all major platforms. Python’s elegant syntax and dynamic typing, together with its interpreted nature, make it an ideal language for scripting and rapid application development in many areas on most platforms. Advantages of Python This language is easy to learn and use It has support from a very large community It includes an extensive set of libraries and frameworks It has built-in support for datatypes Disadvantages of Python This language is often slow in nature while running. Comparing to C, Java or C++, which are statistically typed languages, Python is a dynamically typed language which sometimes makes the computer consume a little more time than expected. Memory consumption is high in this language due to the flexibility of the datatypes Scala Scala is a combination of object-oriented and functional programming in one concise, high-level language. This language was originally built for the Java Virtual Machine (JVM) and one of Scala’s strengths is that it makes it very easy to interact with Java code. Last year in the Tiobe Index report, Scala secured the 20th place among the top twenty programming languages with a rating of 0.9%. Scala’s static types help the developers to avoid bugs in complex applications, while its JVM and JavaScript runtimes allow a developer to build high-performance systems with easy access to huge ecosystems of libraries. Advantages of Scala Scala allows the utilisation of most JVM libraries which helps in becoming deeply embedded in enterprise code This language shares several readable syntax features of popular languages such as Ruby It has several functional features like string comparison advancements, pattern matching, among others which incorporates functions within class definitions Disadvantages of Scala In this language, the type-information can sometimes be complex to understand due to the combination of functional and object-oriented in nature This language has a limited developer in the community For Machine Learning & Data Science Python is currently the most preferred language among the data scientists not just it is easy to learn and implement but also for its extensive libraries and frameworks. In data science and machine learning projects, it includes a broad range of useful libraries SciPy, NumPy, Matplolib, Pandas, among others while for more complex projects in deep learning, Python offers libraries such as Keras, Pytorch, and TensorFlow. On the other hand, one of the important reasons to learn Scala for machine learning is because of Apache Spark. Scala can be used in conjunction with Apache Spark in order to deal with a large volume of data which can also be called Big Data. Popularity According to the Tiobe Index reports for September 2019, Python has ranked the third position after Java and C language. The reports have also shown that Scala is securing 30th position in the list of 50 trending programming languages. Community In simple words, the community for Python programming language is huge. For better enhancement of the language, the community keeps hosting conferences, meetups, collaborates on code and much more. According to our skills study report, Python is one of the largest programming communities in the world. The favourite language for data scientists is Python, as almost 68% of the professionals use it the most.","excerpt":"Python and Scala are two of the most popular languages used in data science and analytics. These languages provide great support in order to create efficient projects on emerging technologies. In this article, we list down the differences between these two popular languages. Python  Python continues to be the most popular language in the industry. […]","categories":["Deep Tech"],"tags":["apache scala","Apache Spark","Data Analytics","Python","Python language","scala"],"author_name":"Ambika Choudhury","publish_date":"2019-09-11T17:00:29","publication_year":"2019","word_count":656,"keywords":["data science","NumPy","Python language","apache scala","machine learning","Keras","AI","PyTorch","Apache Spark","scala","Python","deep learning","analytics","Data Analytics","TensorFlow","Pandas"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","analytics","TensorFlow","PyTorch","Keras","Pandas","NumPy"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/python-vs-scala-which-language-is-best-suited-for-data-analytics\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10141678,"title":"Meetings are Now a Problem of the Past","content":"Let’s face it, our workdays often feel like an endless cycle of meetings. From the SOD (start-of-day) huddles to the EOD (end-of-day) debriefs, it sometimes feels like we spend more time discussing work than actually getting it done. This isn’t just a corporate thing. As journalists, we’re no strangers to meetings, debating pitches and dissecting story ideas until everyone agrees or, eventually, gives up. Such is the nature of the game. But recently, something interesting caught our eye. Pickle, a software development company, posted on X about its lifelike AI clones that can attend video calls on behalf of users while requiring no cameras. These AI avatars lip-sync to users’ voices in real time, mirroring their facial expressions and interactions with almost no lag. This feature, however, comes with a hefty price tag. Pickle’s subscription for 15 hours monthly costs around ₹2,000. For someone clocking in four hours of daily meetings, this feels impractical. This is when we found another such platform, HeyGen. Tempted by the idea of skipping the grind, we decided to sign up and see if it lived up to the hype. How it works? To use this feature, users need to record their voice for 30 minutes and video footage for three minutes with the camera on. Using this input, HeyGen then creates a digital avatar which can attend meetings, take notes, and, perhaps, give users some more time to enjoy their day. Interestingly, this isn’t the first AI-powered twist in video conferencing. Zoom’s CEO, Eric S Yuan, previously shared his vision of an AI clone that could join meetings on behalf of attendees. “I can send a digital version of myself to the meeting, freeing me up to hit the beach,” Yuan joked during an interview with The Verge. “There’s no need for five or six Zoom calls a day when AI can take care of that for you,” he added. Zoom has already launched Zoom Workplace, an AI-powered platform designed to enhance teamwork and productivity. The platform offers 40 new features such as updates to Zoom AI Companion for Zoom Phone, Team Chat, Events, Contact Center, and the Ask AI Companion feature. Notably, Zoom has emphasised its commitment to responsible AI. The company claims that none of the users’ audio, video, chat, screen sharing, attachments, or other customer content will be used to train Zoom’s or third-party AI models. Deepfakes for Good Cloning isn’t exactly new. Deepfakes, for instance, often viewed as a threat, have the potential to drive positive, even transformative, change in business and beyond. For instance, an animated GIF of Martin Luther King Jr was created using online genealogy platform MyHeritage.com. The GIF, which was made from an image on Wikipedia, serves as an interactive learning tool for students. Education is embracing AI in several other innovative ways. Stanford researchers have introduced Tutor CoPilot, an AI-powered system built on OpenAI’s GPT-4. Integrated with FEV Tutor, the platform connects students with human tutors virtually, bridging the gap between expertise and accessibility. Tutor CoPilot offers tutors “expert-like” insights and enhances their teaching, ultimately making learning more engaging and personalised. Meanwhile, in the telecommunications industry, AI avatars are stepping in to handle customer interactions. These digital clones respond to queries and note customer concerns, streamlining the process. For example, Indian telecalling outsourcing partner SquadStack uses AI-driven solutions to revolutionise traditional setups. What’s Next? The Apple Vision Pro (AVP) and similar other devices are set to profoundly impact our communication, enabling 3D video calls, playback of real-life “memories”, and the sharing of immersive experiences. Tech giants like Meta, OpenAI, Microsoft, and Google have already rolled out AI features for the workplace. Now, startups are also joining the race. UK-based startup Artisan AI is on a mission to create advanced human-like digital workers called Artisans. So far, they have released Ava, a sales representative for Artisan. Ava operates as a business development representative (BDR) who streamlines the entire outbound sales process, requiring only a brief 10-minute conversation for setup. A similar project was presented on Shark Tank US earlier in 2022. AI productivity platform Beulr enables a person to be virtually present in two places simultaneously. It looks like we are not that far from the days when we can actually relax during meetings!","excerpt":"Pickle, a software development company, offers lifelike AI clones that can attend video calls on behalf of users.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI Avatar","meeting"],"author_name":"Vidyashree Srinivas","publish_date":"2024-11-26T10:00:00","publication_year":"2024","word_count":708,"keywords":["meeting","Go","AI Avatar","AI","OpenAI","ML","Git","GPT","responsible AI","Aim","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","R","Go","Git","GPT","ViT","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meetings-are-now-a-problem-of-the-past\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10119318,"title":"Bengaluru Leads in Diversity Representation Among tier-1 Indian Cities: Report","content":"According to a recent report on gender diversity by Pure Storage and Zinnov, Bengaluru leads among tier-1 cities in India in diversity representation. The report found that diversity is around 31.4% in Global Capacity Centres (GCCs) based in the city and 14% in deeptech companies. The report titled, ‘Towards a Gender Equitable World: Unveiling Diversity in DeepTech,’ highlights the need for greater focus on university enrolment in STEM courses and workplace retention to address the low representation of women in the DeepTech sector. The report analyses women engineering graduates between 2004 and 2023 from 42 top engineering universities leveraged by GCCs for recruitment, with particular emphasis on 23 top institutions deemed to be preferred by DeepTech companies. Moreover, it also highlights that the ongoing gender disparity is largely due to two main factors: a shortage of women’s enrolment in these institutions and a significant rate of mid to senior-level dropouts within the industry. Some of the other key takeaways from the report include: GCCs are leading the charge for a diverse workforce with 28% women in their workforce, yet they face unique challenges in achieving gender parity in DeepTech organizations, where the gender diversity stands at 23%​. The median representation of women graduates from top engineering universities stands at 25% between 2020 and 2023, which directly affects the inflow of women candidates in GCCs, especially in the deeptech sector. Despite this disparity in women’s representation, women graduates consistently outperformed in securing placements compared to the overall average in top-tier universities. With a mere 6.7% of women at the Executive level in GCCs and 5.1% in deeptech organisations, there is a considerable decrease in the available talent pool of women as they move up their careers. Family and caregiving responsibilities, limited access to career advancement and leadership opportunities, poor work-life balance are some of the key factors influencing women’s attrition.","excerpt":"The report found that diversity is around 31.4% in GCCs based in the city and 14% in deeptech companies.","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2024-04-30T18:07:37","publication_year":"2024","word_count":309,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","GAN","R"],"extracted_tech_keywords":["AI","RAG","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bengaluru-leads-in-diversity-representation-among-tier-1-indian-cities-report\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10098549,"title":"Snowflake Now Wants You to Converse With Your Data","content":"Snowflake’s growth trajectory has been nothing short of remarkable. Since 2012, the company has witnessed exponential market adoption and has attracted a diverse range of clients, from startups to Fortune 500 giants. Some of its notable customers include Adobe, Airbnb, BlackRock, Dropbox, Pepsico, ConAgra Foods, Novartis and Yamaha. In India, Snowflake caters to the needs of companies such as Porter, Swiggy and Urban Company. The rapid expansion is a testament to Snowflake’s ability to address the ever-increasing demands of the data-driven world we live in. But today, we are stepping into the age of generative AI and Snowflake too is gearing up to bring the best of the technology to its long list of customers. Torsten Grabs, senior director of product management at Snowflake told AIM that with the advent of generative AI, we will increasingly see less technical users successfully interact with computers with technology and that’s probably the broadest and biggest impact that he would expect from generative AI and Large Language Models (LLMs) across the board. Moreover, talking about the impact of generative AI on Snowflake, he said that it has impacted Snowflake on two distinct levels. Firstly, like almost every other company, generative AI is leading to productivity improvements at Snowflake. Grabs anticipates developers working on Snowflake to benefit the most from generative AI. This concept is akin to Microsoft’s co-pilot and AWS’s CodeWhisperer, where a coding assistant aids in productivity by comprehending natural language and engaging in interactive conversations to facilitate faster and more precise code creation. Moreover, Snowflake is harnessing generative AI to enhance conversational search capabilities. For instance, when accessing the Snowflake marketplace, it employs conversational methods to identify suitable datasets that address your business needs effectively. “There’s another layer that I think is actually very critical for everybody in the data space, which is around applying LLMs to the data that’s being stored or managed in a system like Snowflake,” Grabs said. The big opportunity for Snowflake lies in leveraging generative AI to offer enhanced insights into the data managed and stored within these systems. Conversing with your data On May 24, 2023, Snowflake acquired Neeva AI with the aim of accelerating search capabilities within Snowflake’s Data Cloud platform by leveraging Neeva’s expertise in generative AI-based search technology. “We recognised the necessity of integrating robust search functionality directly into Snowflake, making it an inherent and valuable capability. Partnering with Neeva AI further enriched our approach, combining their expertise in advanced search with generative AI, benefiting us in multiple dimensions,” Grabs said. Grabs believes the Neeva AI acquisition is going to bring a host of benefits to Snowflake’s customers. Most importantly, it will give them the ability to talk to their data essentially in a conversational way. “It’s analogous to the demonstration we presented, where a conversation with the marketplace utilizes metadata processed by the large language model to discover relevant datasets,” Grabs said. Now consider scaling this process and going beyond metadata, involving proprietary sensitive data. By employing generative AI, Snowflake’s customers can engage in natural language conversations to gain precise insights about their enterprise’s data. Building LLMs for customers Building on generative AI capabilities, Snowflake, at its annual user conference called ‘Snowflake Summit 2023’ also announced a new LLM built from Applica’s generative AI technology to help customers understand documents and put their unstructured data to work. “We have specifically built this model for document understanding use cases and we started with TILT base model that we leveraged and then built on top of it,” Grabs said. When compared to the GPT models from OpenAI or other models developed by labs such as Antrhopic, Snowflake’s LLMs offers few distinct advantages. For example, the GPT models are trained on the entirety of publicly available internet data, resulting in broad capabilities but high resource demands. Their resource-intensive nature also makes them costly to operate. Much of these resources are allocated to aspects irrelevant to your specific use case. Grabs believes utilising a more tailored, specialised model designed for your specific use case allows for a narrower model with a reduced resource footprint, leading to increased cost-effectiveness. “This approach is also poised to yield significantly superior outcomes due to its tailor-made design for the intended use case. Furthermore, the model can be refined and optimised using your proprietary data. This principle isn’t confined solely to the document AI scenarios; rather, it’s a pattern that will likely extend more widely across various use cases.” In many instances, these specialised models are expected to surpass broad foundational models in both accuracy and result quality. Additionally, they are likely to prove more resource-efficient and cost-effective to operate. “Our document AI significantly aids financial institutions in automating approval processes, particularly for mortgages. Documents are loaded into the system, the model identifies document types (e.g., salary statements), extracts structured data, and suggests approvals. An associate reviews and finalises decisions, streamlining the process and enhancing efficiency.” Addressing customer’s concerns While generative AI has garnered significant interest, enterprises, including Snowflake’s clients, which encompasses 590 Forbes Global 2000 companies, remain concerned about the potential risks tied to its utilisation. “I think some of the top concerns for pretty much all of the customers that I’m talking to is around security, privacy, data governance and compliance,” Grab said. This presents a significant challenge, especially concerning advanced commercial LLMs. These models are often hosted in proprietary cloud services that require interaction. For enterprise clients with sensitive data containing personally identifiable information (PII), the prospect of sending such data to an external system outside their control and unfamiliar with their cybersecurity processes raises concerns. This limitation hinders the variety of data that can interact with such systems and services. “Our long-standing stance has been to avoid dispersing data across various locations within the data stack or across the cloud. Instead, we advocate for bringing computation to the data’s location, which is now feasible with the abundant availability of compute resources,” Grabs said. Unlike a decade or two ago when compute was scarce, the approach now is to keep data secure and well-governed in its place and then bring computation to wherever the data resides. He believes this argument extends to generative AI and LLMs as well. “We would like to offer the state-of-the-art LLMs and side by side the compelling open-source options that operate within the secure confines of the customer’s Snowflake account. This approach ensures that the customer’s proprietary or sensitive data remains within the security boundary of their Snowflake account, offering them peace of mind.” Moreover, on the flip side, another crucial aspect to consider is the protection of proprietary intellectual property (IP) within commercial LLMs. The model’s code, weights, and parameters often involve sensitive proprietary information. “With our security model integrated into native apps on the marketplace, we can ensure that commercial LLM vendors’ valuable IP remains undisclosed to customers utilising these models within their Snowflake account. Our role in facilitating the compute for both parties empowers us to maintain robust security and privacy boundaries among all participants involved in the process,” Grabs concluded.","excerpt":"The Neeva AI acquisition will give Snowflake’s customers the ability to talk to their data essentially in a conversational way","categories":["Global Tech"],"tags":["Snowflake"],"author_name":"Pritam Bordoloi","publish_date":"2023-08-14T15:00:00","publication_year":"2023","word_count":1172,"keywords":["OpenAI","AI","AWS","ML","RAG","document AI","Aim","generative AI","R","Snowflake"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Aim","RAG","document AI","AWS","Snowflake","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/snowflake-now-wants-you-to-converse-with-your-data\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10335,"title":"Analytics India Industry Study – 2016","content":"Over the last few years, analytics has rapidly gained importance and there is hardly any industry that analytics has not touched upon. Analytics is being used everywhere from Banking to Information Technology to Film-making. Hence, as a part of our annual exercise, we have undertaken to conduct ‘Analytics India Industry Study 2016’ with the aim to know where the analytics industry stands as of today and what lies in ahead for the industry. The study was conducted over a period of one month and by collecting data from both primary and secondary sources. Further an in-depth analysis of data was carried out to understand the size of the industry, its biggest revenue contributors and the size & demographics of analytics professionals in India. Analytics India Industry Study 2015 Analytics India Industry Study 2013 Analytics India Industry Study 2012 Industry Size Analytics Market in India currently stands at $1.64 Billion annually in revenues, growing at a healthy rate of 28.8% CAGR. Though analytics is still largely considered a part of broader KPO industry, increasingly analytics is now being considered a separate industry of its own. This is evident from the rise of many startups in analytics and data science space creating a niche of their own. Moreover, numerous sub industries are mushrooming around analytics, most prominent being artificial intelligence. Sector Type In terms of Sector type, Finance & Banking form the largest sector being served by analytics in India. Of the total revenue earned by analytics industry in India, 35% or $575 Million comes from Finance & Banking. Marketing comes second at 25%, followed by E-commerce sector at 17% of analytics revenues in India. Domain in Analytics Almost 70% of all analytics related work done in India is in the space of BI\/ Reporting\/ Dashboard or analytics code deployment & maintenance. Just 7% of analytics is in advanced model building and prediction. 22% of analytics professionals are involved in Big Data Management. Analytics Industry by Cities 29% or $472 Million in market size for analytics industry comes from Delhi\/ NCR. This is followed by Bengaluru at 26%. We earlier reported that Bengaluru is the biggest city in terms of number of analytics companies in India. Yet, Bengaluru lags behind Delhi\/ NCR in terms of relative market sizes. Analytics Professionals in India The average work experience of analytics professionals in India is 7.2 years. Around 8,500 freshers were added to analytics workforce in India last year. Almost 46% analytics professionals in India have a work experience less than 5 years. Around 19% of analytics professionals are at senior positions (Director, VP, Founders). 56% of analytics professionals have a Master’s\/ Post Graduation degree. Only 3% of analytics professionals in India hold a PhD or Doctorate degree. Women participation in analytics workforce is low – just 23% of analytics professionals in India are women. Analytics Industry as a whole is growing rapidly at 28.8% CAGR and almost 80% of its revenue comes from 3 sectors i.e. Finance and Banking, Marketing, and E-commerce. As these sectors grow, the need for analytics by them will also increase thereby contributing to the overall growth of analytics industry. At present most of the analytics related work is in the space of BI\/ Reporting\/ Dashboard and areas like advanced analytics and big data are yet under-tapped, representing so much scope for the industry to grow in these areas. Also from the data of analytic professionals in India, it is clearly evident that in order to support the growth of the industry, more skilled professionals are required and especially at the senior levels.","excerpt":"Over the last few years, analytics has rapidly gained importance and there is hardly any industry that analytics has not touched upon. Analytics is being used everywhere from Banking to Information Technology to Film-making. Hence, as a part of our annual exercise, we have undertaken to conduct ‘Analytics India Industry Study 2016’ with the aim […]","categories":["AI Features"],"tags":["Analytics India"],"author_name":"Дарья","publish_date":"2016-07-03T03:37:01","publication_year":"2016","word_count":593,"keywords":["big data","data science","API","artificial intelligence","AI","Analytics India","RAG","Aim","analytics","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","RAG","R","API","big data","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-india-industry-study-2016\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":64084,"title":"What Happened When Google Tested Its AI In Real World","content":"Back in 2016, Google’s AI team have burdened themselves with tackling one of the fastest-growing illnesses of our time — diabetic eye diseases. Diabetic retinopathy (DR) — an eye condition, currently affects people with diabetes and is the fastest-growing cause of blindness, with nearly 415 million diabetic patients at risk worldwide. Google’s researchers have developed a deep learning algorithm that can interpret signs of DR in retinal photographs, potentially helping doctors screen more patients, especially in communities where the resources are limited. This deep learning algorithm showed great promise with results that are on par with ophthalmologists. After three years of thorough testing and tweaking the model, a team of researchers have decided to put their model into practice. For this, they have chosen Thailand, where there are only about 1,400 eye doctors for approximately five million diabetics. How Did It Go Google AI, in partnership with the Ministry of Public Health in Thailand, conducted field research in clinics across the provinces of Pathum Thani and Chiang Mai for over a period of eight months. During this period, the researchers made regular visits to 11 clinics, observed how the nurses of those clinics handled eye screenings and interviewed them to have a deeper understanding of the process. In the course of their trials, they found significant fundamental issues in the way the deep learning systems were deployed. Though the model was improved regularly, the challenges came from factors external to the model. For instance, some images captured in screening might have issues like blurs or dark areas. An AI system might conservatively call some of these images “ungradable” because the issues might obscure critical anatomical features that are required to provide a definitive result. For clinicians, the gradability of an image may vary depending on one’s own clinical set-up or experience. Two images of the same eye, with varied lighting The system’s high standards for image quality is at odds with the consistency and quality of images that the nurses were routinely capturing under the constraints of the clinic. The nurses at the clinic took two images of the same eye in case of an ungradable image. However, according to the report, this caused discomfort in the patients and also added to the frustration of the nurses. So, they have explored solutions such as darkening the room to improve lighting conditions that will lead to higher-quality images. Not only this, but even the speed of internet connection plays a major role in the time taken for each patient. This proves that no matter how good a model is, challenges surface once they are deployed in the real-world. More so in something like healthcare set up. Key Findings & Recommendations In a recent report, the researchers have elaborated more on their research. The findings can be summarised as follows: In the case of user-centred applications, product design should involve people who would interact with the technology. In the case of AI systems in healthcare, we must also factor in environmental differences like lighting, which vary among clinics and can impact the quality of images. Just as an experienced clinician might know how to account for these variables in order to assess it, AI systems also need to be trained to handle these situations. Building an AI tool is a challenge, as any disagreements between the system and the clinician can lead to frustration. This study found that the AI system could empower nurses to confidently and immediately identify a positive screening, resulting in quicker referrals to ophthalmologists. Future Direction Although the researchers evaluated a deep learning system in the wild, they say that the study was focused on nurses and camera technicians. More research needs to be done to understand how the system affects patient experience and their trust in the results, and the likelihood to act on them. Google’s AI team also suggests that additional research is needed to understand how the system may alter the practices of ophthalmologists who evaluate patients based on the deep learning system. Lastly, as more systems are evaluated in clinical environments, an important area of future work includes the design of study protocols for conducting human-centered prospective studies and studies on end-to-end service design of AI-based clinical products.","excerpt":"Back in 2016, Google’s AI team have burdened themselves with tackling one of the fastest-growing illnesses of our time — diabetic eye diseases. Diabetic retinopathy (DR) — an eye condition, currently affects people with diabetes and is the fastest-growing cause of blindness, with nearly 415 million diabetic patients at risk worldwide. Google’s researchers have developed […]","categories":["Global Tech"],"tags":["Healthcare Automation"],"author_name":"Ram Sagar","publish_date":"2020-05-01T13:00:07","publication_year":"2020","word_count":704,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Healthcare Automation","deep learning","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","deep learning","R","Go","Rust","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-ai-healthcare\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10063118,"title":"Is India shifting to low-code\/no-code platforms?","content":"According to NASSCOM, the global LC\/NC Industry is expected to grow at a CAGR of 28.1% from 2021 to 2025. “The pandemic has changed the way enterprises operate, and this is evident in India as well. At present, Indian enterprises are looking to transform their business processes and make them fully digital. All manual interactions and paper-intensive processes need complete digitisation to ensure that the customer and employee experience remains at par. This leads to heavy investment in the adoption of low code-no code platforms. These could be low code application platforms, low code data science platforms, or low code cognitive platforms. These platforms reduce the go-to-market time by 70% and make it extremely convenient to absorb the ever-changing need of the market, thus leading the competition curve. Another reason for its growing adoption is it solves the problem of managing a huge workbench of skilled resources,” said Arvind Jha, senior vice president, software development, Newgen Software. Indian LCNC market In FY 2021, Indian IT service providers and startups generated over USD 400 million in revenue from LC\/NC solutions, of which 70% came from global markets. The market is projected to grow to USD 4 billion by 2025, a NASSCOM study showed. According to Gartner, low-code application development will account for over 65% of application development activity by 2024. Source: NASSCOM India currently has 150 LC\/NC players, the vast majority of which are bootstrapped startups. As much as 80% of LC\/NC PoCs (Proof of Concepts) were parlayed to full-scale implementations. Hubbler, Appy Pie, and ToolJet are the leading platforms in this space. Tata Consultancy Services won 24 LC\/NC contracts in the third quarter of the current fiscal year. Infosys is also hiring talents to deliver on LC\/NC solutions. “This is one space which will disrupt the IT service market and ensure a quick roll-out of new ideas with little budget,” said Ankit Patel, co-founder at Aadhyarupam Innovators LLP. According to NASSCOM, companies have seen a 30-35% increase in ROI while using LC\/NC platforms and up to a 75% reduction in development time and a 65% reduction in costs. Impact on workforce The popularity of low-code and no-code solutions has opened up new job roles: Coding experts to build LC\/NC platformsSkilled professionals to enable platform integrations between legacy systems and LC\/NC platformsExperts with specialised skills to manage LC\/NC platforms Global tech leaders such as Microsoft, Amazon, and ServiceNow and Indian firms like Infosys, Tech Mahindra, and HCL are betting big on LC\/NC solutions. Leading adopters The LC\/NC platforms are quickly gathering steam across the world. “Low code no-code platforms empower developers to create applications rapidly and even test and deploy the same using the connected CI\/CD pipelines. Nowadays, even the global system integrators that work across industries adopt these platforms to meet customer expectations. Hence, it has applicability across industries and verticals,” said Arvind Jha. Source: NASSCOM As much as 90% of Indian LC\/NC adopters are using homegrown platforms. The BFSI sector has led the way on account of their need for high-volume, customer-facing digital touchpoints. HDFC, SBI, Kotak, and ICICI are among the early adopters. Retail and SaaS product firms followed. SaaS product & services firms, expanding into overseas markets, were responsible for 10% -15% of revenue for Indian LCNC players. “The key driver of these industries is customer delight. Also, the change in regulatory compliance is so frequent that it needs to respond in no time. Hence the applicability of low-code platforms is high compared to other industries,” said Arvind Jha. Challenges According to Arvind Jha, the major challenge is selecting the right platform. We don’t have a one-size-fits-all LC\/NC platform yet. “On the one hand, there are low-code platforms that are specialised in the creation of light-weighted applications for personal or small group usage. On the other hand, there are professional low-code platforms that empower large enterprises to rapidly create complex business use cases which require hierarchical approvals and are highly document-centric. Businesses must look at the 2X2 matrix of department vs enterprise and internal vs customer use cases and try to fit the different platforms to their different needs,” he said. Vetting LCAP platforms and investing in the most cost-effective solution, is the real pain point. “Analysing the solution thoroughly to differentiate between UI capabilities, data capabilities, process & workflow capabilities, business logic, integration, and other services would be critical to ensure that the right investments are taking place,” said Srividya Kannan, founder, Avaali Solutions. Low-code and no-code technology solutions are here to stay, but businesses must consider the viability, value-add propositions, and price-points before picking an LC\/NC solution. “This technology is expected to grow significantly due to the rapid rise in the demand for new applications and continued digital acceleration. It is estimated that low-code application platforms will be responsible for over 70% of application development activity by 2025,” said Srividya Kannan.","excerpt":"According to Gartner, low-code application development will account for over 65% of application development activity by 2024.","categories":["IT Services"],"tags":["coding","Developers","Digital India","digital transformation","Indian IT","IT","low code","Make in India","programming"],"author_name":"Sri Krishna","publish_date":"2022-03-21T12:00:00","publication_year":"2022","word_count":806,"keywords":["IT","API","coding","Git","R","data science","Make in India","digital transformation","ViT","CI\/CD","startup","Go","Digital India","AI","Indian IT","low code","programming_languages:R","programming","Developers"],"extracted_tech_keywords":["AI","data science","R","Go","Git","CI\/CD","API","ViT","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/is-india-shifting-to-low-code-no-code-platforms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":52590,"title":"Is Data Science For You? This IBM Data Scientist Tells How To Figure It Out","content":"In the ever-changing data science landscape, people need to assimilate whether they are fascinated by the machine learning processes or just going along with the hype this field has generated. Many a time, fresh graduates get into data science, but they struggle to stay abreast with the latest trends. To discuss this, and more, Analytics India Magazine got in touch with Sharath Kumar RK, Data Scientist at IBM India Software Labs, for our weekly column My Journey In Data Science. He spoke to us about how natural it was for him to choose data science over other thriving domains. Data Science Journey Sharath graduated from Mysore and later on completed his Master’s Degree in Finance from BITS Pilani. Early in his employment career, Sharath worked as a technical support specialist for Microsoft products like Windows Server. However, it was only in his second job when he was exposed to data. Sharath used to carry out business intelligence and analytics for financial services. He understood the potential of data and started exploring it to obtain meaningful insights. Fascinated by the abundance and curiosity around data, he learned new techniques for analysing data for solving various real-world business challenges. “I was a good problem solver, which enabled me to move towards curating, understanding and moulding the data using different machine learning techniques for solving numerous problems,” explains Sharath. “I realised that I was never got bored of exploring data using different techniques and started playing around it to churn out insights and predictions,” he added. Such enthusiasm cleared the clouds for Sharath to choose the data science field. He went on to say, how he never looked back after that and kept honing machine learning-related skills that led him to file a couple of disclosures (file rated patents) under artificial intelligence. At IBM, he used various platforms like Watson Portfolio and Cloud Pak for Data, to create multiple solutions. Preparation And Job Strategy While talking about his strategy, Sharath said he preferred learning while at the job, took help from peers, and downloaded materials from the internet for quickly enhancing his skills. It is a competitive landscape, thus one needs to learn from every possible way and from everyone for gaining a competitive edge. For job, Sharath worked on creating a portfolio by working on different projects and hosting them on GitHub. He also solved problems in Kaggle to understand the approach to design the solution. Pinpointing the importance of critical thinking, Sharath said that he firmly believes that data intuition is crucial for solving problems. “Recruiters look at aspirants’ ability to approach and solve problems on paper which include structured thinking, logical reasoning, and call out appropriate techniques that can best fit the scenario. Although those are necessary, the primary focus for recruiters is on candidates capability of understanding the challenges at the micro and macro level,” says Sharath. Recalling his first data science interview, Sharath said he was asked to solve guesstimate questions, riddles on the paper to evaluate how he goes about solving a problem. Further, he was assessed based on his applied technical skills where he had to justify which ML, NLP, and DL, techniques were effective in various scenarios. Strategy To Flourish In Data Science Domain Sharath mentioned there always be someone who knows better than you in some of the other techniques. That’s why, instead of getting demotivated, he tries to learn from them and enhance his dexterity. “I believe learning is a continuous process, thereby, I keep pushing myself to learn new things and add value with my deliverables,” exclaims Sharath. Besides, he said one learning from projects is essential for thriving in the domain. Of many successful projects that he worked on, he prefered a project where he made an end-to-end analytics framework for data extraction, transformation, and building machine learning pipeline, to deploy it in production and build visualisation using dynamic dashboards. Further, he worked on projects such as AI-based humanoid bots, image classification using DL, and create NLP solutions. These projects are open-source and available on the IBM website. Advice For Aspirants “Learning new techniques that emerge in the market is key to stay relevant, it also helps in enhancing performance to analyse data quickly. However, one should not lose grip on the fundamentals to thrive in the competitive landscape. At the same time, aspirants should also focus on building and enhancing soft skills. Candidates with a mix of technical and soft skills are preferred by organisations of late. Besides, candidates should stay motivated, participate in different hackathons, and use those skills to ace interviews for getting an amazing job offer,” concludes Sharath.","excerpt":"In the ever-changing data science landscape, people need to assimilate whether they are fascinated by the machine learning processes or just going along with the hype this field has generated. Many a time, fresh graduates get into data science, but they struggle to stay abreast with the latest trends. To discuss this, and more, Analytics […]","categories":["AI Features"],"tags":["Data Science","Interviews and Discussions"],"author_name":"Rohit Yadav","publish_date":"2019-12-27T10:08:00","publication_year":"2019","word_count":771,"keywords":["data science","Go","machine learning","artificial intelligence","AI","ML","NLP","Aim","analytics","Data Science","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","data science","analytics","Aim","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-data-science-for-you-this-ibm-data-scientist-tells-how-to-figure-it-out\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046000,"title":"Construction Analytics Startup Doxel Raises $40 Million in Series B","content":"California-based Doxel has announced the closing of $40 million in Series B financing – bringing its total funding to $56.5 million. The investment was led by New York-based global private equity and venture capital firm Insight Partners, with participation from existing investors, Andreessen Horowitz and Amplo. Founded in 2016 by Saurabh Ladha and Robin Singh, the software start-up provides computer-vision-powered predictive analytics solutions for the construction industry. The investment will allow Doxel to continue to expand its India workforce with hiring planned across the enterprise, scale its artificial intelligence platform, and rapidly accelerate recruiting across its engineering, sales, marketing & product teams. “With hundreds of thousands of variables changing every day and an opportunity to leverage rapidly exploding datasets on modern construction projects, project teams are seeking a next-generation solution that can do the monitoring for them, so they can focus on solving problems rather than on finding them. Doxel’s computer-vision powered predictive analytics enables building owners and general contractors to identify critical risk factors that threaten to derail their project before they even know these risks exist,” said Saurabh Ladha, Chief Executive & Co-Founder, Doxel. The company’s AI-Powered Project Controls platform taps into multiple real-time data sources on a project – such as 360-degree images, Building Information Models (BIM), as well as budget and schedule – to provide both predictability and control to building owners and contractors. This helps prevent a domino effect of delays and heightened costs, enabling building owners and contractors to stay on time and on budget.","excerpt":"Founded in 2016 by Saurabh Ladha and Robin Singh, the software start-up provides computer-vision-powered predictive analytics solutions for the construction industry.","categories":["AI News"],"tags":["Computer Vision","Funding"],"author_name":"kumar Gandharv","publish_date":"2021-08-13T17:30:19","publication_year":"2021","word_count":251,"keywords":["API","Funding","artificial intelligence","funding","programming_languages:R","AI","R","venture capital","RAG","analytics","Computer Vision","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","RAG","predictive analytics","R","API","venture capital","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/construction-analytics-startups-doxel-raises-40-million-in-series-b\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10124768,"title":"Tata Electronics and Synopsys Partner for Factory Automation and Establishing an AI-enabled Fab","content":"Tata Electronics has signed a Memorandum of Understanding (MOU) with Synopsys, a leading provider of silicon-to-systems design solutions, to collaborate on process technology bring-up and a foundry design platform to accelerate the successful ramp of customer products in India’s first fab being built by Tata Electronics in Dholera, Gujarat. The two companies have identified the following areas of potential collaboration: Advanced factory automation and yield data analytics solutions to help establish an AI-enabled fab TCAD (Technology Computer Aided Design) flow set-up to enable accurate technology transfer from the technology partner PDKs (process design kits) and design enablement IP development, including foundation and analog IP DTCO (Design Technology Co-optimization) methodologies “For nearly 30 years, Synopsys has been researching and developing silicon-to-systems design solutions for customers and investing in workforce development in India. We applaud and support Tata Electronics’ vision to develop semiconductor manufacturing capacity in India, advancing supply chain resiliency for the global semiconductor industry,” Sassine Ghazi, president & CEO of Synopsys, said. As previously announced, Tata Electronics plans to build India’s first fab in Dholera, Gujarat, with a total investment of INR 91,000 crores. In addition, another INR 27,000 crores will be invested in a greenfield facility in Jagiroad, Assam for assembly and testing of semiconductor chips. Together these facilities will produce semiconductor chips for applications across automotive, mobile devices, artificial intelligence (AI), and other key segments to serve customers globally. As construction of the facilities progresses, it is critical to grow partnerships across the entire semiconductor ecosystem spanning process and design technology, and equipment suppliers. With this intended collaboration with Synopsys, Tata Electronics solidifies a critical pillar for a holistic approach to achieve its targets to be the first to bring semiconductor manufacturing to India.","excerpt":"Tata Electronics plans to build India’s first fab in Dholera, Gujarat, with a total investment of INR 91,000 crores.","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2024-06-25T18:59:24","publication_year":"2024","word_count":286,"keywords":["artificial intelligence","programming_languages:R","AI","automation","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tata-electronics-and-synopsys-partner-for-factory-automation-and-establish-an-ai-enabled-fab\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002535,"title":"Webinar: What is Chartered Data Scientist Designation","content":"Making a career in the field of data science and machine learning is still uncertain for many people. The biggest reason is that this field has not been completely explored yet. Many of the industry segments are still looking at various avenues where they can employ data science to grow their businesses. Another reason is that the aspirants who wish to switch their career in this field do not have a complete understanding of what the required skill sets are. Even if a candidate makes a career in data science, often they pose challenges such as gaining the trust of the employer and prove his\/her potential by performing the assigned tasks. To showcase the required knowledge and skill sets, data science professionals often resort to a variety of professional certifications designed for a specific target group. One such popular program offered by the Association of Data Scientists (ADaSci) is the Chartered Data ScientistTM. Chartered Data ScientistTM (CDSTM) is a prestigious distinction in the field of Data Science that is awarded by the Association of Data Scientists (ADaSci). The CDSTM charter which is awarded to professionals in the field data science gives a unique value to its holder. In the data science marketplace where recruiters are struggling to find a potential candidate in the field of data science and machine learning, this distinction helps its holder stand out from the crowd. In this webinar, the Chartered Data ScientistTM program of ADaSci will be discussed in detail. The following points will be covered during this webinar:- Career in the field of Data Science Current situation and future scope in Data Science Various certification programs in Data Science Chartered Data ScientistTM Program The uniqueness of Chartered Data ScientistTM distinction Benefits of Chartered Data ScientistTM How to become a Chartered Data Scientist Chartered Data ScientistTM Exam Information How to prepare for Chartered Data ScientistTM Study resources to prepare for Chartered Data ScientistTM How this Charter is awarded Ethical Standards for Chartered Data Scientists Career ahead Questions and Answers To register for this webinar, click here. Session Details Date: 1st August 2020 Time: 11:00 AM to 01:00 PM IST Speaker Dr. Vaibhav Kumar Dr. Vaibhav Kumar has broad experience in the field of Data Science and Machine Learning, including research and development. He holds a PhD degree in which he has worked in the area of Deep Learning for Stock Market Prediction. He has published\/presented more than 15 research papers in international journals and conferences. He has an interest in writing articles related to data science, machine learning and artificial intelligence. Click here to register for this webinar.","excerpt":"There is a webinar going to be held on 1st August 2020 on “What is the Chartered Data Scientist Designation”","categories":["Deep Tech"],"tags":["ADAsci (Association of Data Scientists)","AI Certifications","Career","chartered data scientist","Data Analytics Certification","Data Science Certification","red hat","webinar"],"author_name":"Dr. Vaibhav Kumar","publish_date":"2020-07-15T17:00:16","publication_year":"2020","word_count":433,"keywords":["data science","red hat","ADAsci (Association of Data Scientists)","artificial intelligence","chartered data scientist","Data Science Certification","AI","machine learning","AI Certifications","programming_languages:R","programming_languages:Rust","deep learning","Rust","Data Analytics Certification","R","Career","webinar"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","R","Rust","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/cds-webinar-what-is-chartered-data-scientist-designation\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":54920,"title":"Facebook Takes Privacy Seriously: Launches Tool To Let Users Stop 3rd-Party Data Sharing","content":"Facebook has now announced it would provide all its users a privacy tool — “Off-Facebook Activity”, which lets them delete the data that third-party websites and apps share with Facebook for targeted advertising. The feature was initially rolled out at Facebook’s annual developer conference in 2018 but was available to users only in a few countries. Back then, it was called Clear History, akin to browser terminology. To now clarify that the tool helps with enhancing user data privacy. The tool not only deletes user activity or history but also gives more freedom to the users of the biggest social media company. Formally launched at Data Privacy Day, the tool is a sign that Facebook is taking serious steps to enhance its user data privacy. Earlier, the issue of third party sharing the user data led to multiple controversies such as the Cambridge Analytica scandal which made a big dent in Facebook’s image in the public eye. Nevertheless, the company has remained profitable and making advancements in areas like artificial intelligence, autonomous systems, virtual reality, to name a few. Given data privacy is a rising end-user concern, and such a tool was long overdue. Facebook explained how other businesses send Facebook data about all user activity on their sites and apps, which Facebook then uses to exhibit personalised ads on their ad marketplace similar to Google ads also work. With the new Off-Facebook Activity tool, users can keep track of the summary of that information and clear it from your account if they like. Users can then wish to break the third-party’s link to Facebook or to have a 3rd party app third-party to delete data it has harvested on a user. Is The Tool A Bane For Targeting Advertising? Targeted advertising is a form of advertising, which is aimed at users with particular personalised traits, based on the product or a user the advertiser is marketing to, using search history, browser history, purchase history, or online activity on social media. The majority of targeted media advertising presently leverage second-order proxies for targets, such as tracking online or mobile online activities of users, linking historical web page consumer demographics with new consumer web page access, utilising search words as the basis for a particular interest, and contextual advertising. With more contextual advertising and less targeted advertising, Facebook’s reliance on third-party data providers have anyway been reducing over the last few years. Besides, given Facebook’s powerful AI capabilities, the company would need smaller sets of user data for its advertising business. Other companies like Google are evaluating federated learning of cohorts (FLoC), which uses federated learning — a more privacy-focused form of machine learning for showing ads based on instead of tracking users individually. A Positive Step For Data Privacy According to experts, this could be a positive step in the right direction and probably a milestone for other social media companies to learn from. This is because it is the first tool of its kind to let users log off from third party data tracking, thus prompting advertising-driven companies to follow suit. At the same time, others say it would also be more useful for users if companies also let users know their data sharing practices as most companies don’t openly promote that they already share with Facebook. Facebook had to revamp its systems to get the tool out, and in coming weeks, it is expected that the Off-Facebook Activity tool will be the new highlight of new privacy settings on Facebook and show up in the News’ Feed as an alert. Nevertheless, the tool doesn’t stop third parties to collect user data makes it accessible to users to know which company is collecting data on them and then sharing it with Facebook. That way, the onus is also on third-party apps to be responsible and ethical is their data collection of end-users for targeted advertising. Also, it is to be noted that a clear history button doesn’t prevent the third-parties from data-sharing later on. With the button, the user data access will be blocked as the user gets logged out from 3rd party apps. Overview Users are usually unable to access and keep track of all the various sites and apps collecting and sharing data because the total number of such third-party apps could be in tens and potentially hundreds. All the various digital services Facebook users have outside of the social media platform would be collecting and sharing data on the ad marketplace. Try the tool to know which of your favourite and loyal companies are tracking you off your Instagram.","excerpt":"Facebook has now announced it would provide all its users a privacy tool — “Off-Facebook Activity”, which lets them delete the data that third-party websites and apps share with Facebook for targeted advertising. The feature was initially rolled out at Facebook’s annual developer conference in 2018 but was available to users only in a few […]","categories":["AI Features"],"tags":["Data Privacy","Facebook","facebook privacy"],"author_name":"Vishal Chawla","publish_date":"2020-01-30T18:41:09","publication_year":"2020","word_count":763,"keywords":["federated learning","Go","facebook privacy","artificial intelligence","machine learning","AI","Git","RAG","Aim","Data Privacy","ViT","Facebook","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","federated learning","RAG","R","Go","Git","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/facebook-takes-privacy-seriously-launches-tool-to-let-users-stop-3rd-party-data-sharing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167812,"title":"Is Google Making a Comeback With its AI Integrations in Cloud?","content":"Google claims that nearly 90% of generative AI unicorns and more than 60% of funded GenAI startups are Google Cloud customers. While Google might be a little late to the party with AI breakthroughs, it may be on to something with its AI integrations being available on Google Cloud. As part of the Google Cloud Next Conference 2025, the company announced that it is integrating AI-powered assistance throughout the entire application lifecycle. With this development, Google is trying to address the inherent complexities of traditional, infrastructure-centric cloud models. Developers using Google Cloud should benefit the most. How are they trying to empower developers using Google Cloud? Is it a big threat to competitors? Is it making a comeback this year after Microsoft Azure crushed both AWS and Google Cloud in revenue last year? Moving to Application Level From Infrastructure Level When building an application using the cloud, developers often focus on optimising and securing the infrastructure to protect their apps. Google, with its new AI capabilities, aims to put applications at the centre of the cloud experience, keeping the complex nature of infrastructure as an abstract layer away from it. A developer can now observe, secure, and optimise at the application level. Brad Calder, VP and GM of Google Cloud Platform, stated in a blog post, “Application components are spread across many systems and dashboards, making it hard to track performance, understand costs, and troubleshoot issues. Our application-centric approach helps you address these challenges, with a number of new services and expanded features.” Application Design Centre and Hubs To begin with, Google has introduced an Application Design Centre service in public preview, which aims to assist administrators and developers in streamlining the design, deployment, and evolution of cloud applications. In the process, the service also ensures that the apps are secure, reliable, and follow the best practices. Not just limited to APIs and the Google Cloud command line interface, Application Design Centre enables developers to interact with a visual, canvas-style format to design and tweak application templates. Complementing this service, Google has added a Cloud Hub, which is a control centre for the app, and improvements to the App Hub. The Cloud Hub service provides a complete view of the application landscape. It provides essential insights into deployments, application health, troubleshooting, resource optimisation, maintenance schedules, quota management, and support cases. This integrated perspective streamlines management and speeds up issue resolution by emphasising the critical elements of applications and their related workloads. The App Hub integrates with over 20 Google Cloud products like Google Kubernetes Engine (GKE), Cloud Run, Cloud SQL, and AlloyDB, enabling better modelling of interconnected applications. Alongside this, new observability features are improving application-centric insights. Taking AI Assistance for Everything To accelerate everything involved in the application development process, Google is integrating Gemini Code Assist and Cloud Assist. When it comes to Gemini Code Assist, the new AI agents possess the capability to create applications from specifications in Google Docs, translate code between language versions, create code for new features, review, test, and generate documentation. Gemini Code Assist can be found in Android Studio and Firebase Studio, in the form of prototyping and testing agents. “CME Group, which runs the Chicago Mercantile Exchange, states that Code Assist provides most of its developers more than 10.5 hours of productivity gain per month,” the company shared. It is not just code; Google’s Gemini Cloud Assist aims to help developers streamline cloud operations. It can generate architectural diagrams and application templates based on natural language descriptions, which can then be iteratively refined and readily deployed. Furthermore, Gemini Cloud Assist provides AI-driven insights into application cost and resource utilisation, identifying inefficiencies and offering personalised recommendations for optimisation. Differentiator Factors to Drive Google’s Comeback Google Cloud’s implementation with GenAI startups potentially gives an edge to Google over competitors like AWS and Azure. Not to forget, startups like Mistral AI and Anthropic have partnered with Google Cloud to offer a range of models on their platforms, using Google’s Cloud TPU chips. This should become more pronounced with Google’s Ironwood TPU chips, which are designed specifically for inference. The app-centric approach could also bring in more developers to the platforms. However, can they migrate to a different platform when necessary? How easy is that process? This may help developers make more informed decisions. AIM reached out to Google to understand how easily users can migrate to other platforms. No response was received before publication. The article will be updated accordingly if the company responds.","excerpt":"Developers using Google Cloud should benefit the most.","categories":["AI Features"],"tags":["Google"],"author_name":"Ankush Das","publish_date":"2025-04-11T15:04:53","publication_year":"2025","word_count":749,"keywords":["Anthropic","GenAI","TPU","AWS","AI","ML","Aim","generative AI","Google","Azure","kubernetes"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","Anthropic","Aim","AWS","Azure","kubernetes","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-google-making-a-comeback-with-its-ai-integrations-in-cloud\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10140441,"title":"Donald Trump’s Victory Paves the Way for AI’s Free Rein","content":"Republican candidate and former US President Donald Trump is set to return to the White House after securing a little over 270 electoral college votes, surpassing the majority mark in the 2024 presidential election. With this win, Trump has not only changed the political landscape but is also likely to alter the technology landscape globally. The former president, who has expressed enthusiasm for and scepticism about Artificial Intelligence (AI) technology in the past, has left the AI industry leaders feel optimistic about the future. Trump is not a proponent of regulating AI companies. Throughout his campaign, he revealed several plans to significantly change the AI ecosystem in the United States and beyond. One key policy decision that the industry is hoping for is the deregulation of Biden’s 2023 Executive Order on AI, which Trump has promised to revoke once he comes to power. The policy mostly focuses on stringent measures, and practices that AI firms need to follow. The order was drafted to avoid misuse, misinformation and any dangerous consequences that would affect the public in several key sectors. Expect a more lax approach to AI regulation in the US with Trump’s election win. “With the combination of Chevron Deference being diminished and a new administration pulling back the enforcement efforts of the FTC and other agencies, we’ll see a vacuum of federal governance over… pic.twitter.com\/b7z0OkSE5j— Parmy Olson (@parmy) November 6, 2024 The First Reactions – MAGA With AI Despite Trump threatening to throw Meta CEO Mark Zuckerberg in prison for possibly meddling with the elections, and taking on Google saying it could be broken because its search results are ‘rigged’, tech industry leaders are lining up to congratulate the leader on his win. Amazon founder Jeff Bezos took to X to send his wishes to the newly elected and said, “No nation has bigger opportunities. Wishing Donald Trump all success in leading and uniting the America we all love.”OpenAI CEO Sam Altman extended his diplomatic wishes to Trump and said, “It is critically important that the US maintains its lead in developing AI with democratic values.” Meanwhile, Google CEO Sundar Pichai also expressed his support and said, “We are in a golden age of American innovation and are committed to working with his administration to help bring the benefits to everyone.” During in presidency, Trump acknowledged the importance of AI and even signed an executive order in 2019 to promote the development of AI. He also used AI to rewrite his political speech during his campaigns this year and expressed his excitement over it. “What it does is so crazy. Now, it can also be really used for good,” Trump had said in an interaction with American influencer Logan Paul. With his win, people in the industry predict a shift in the technological landscape and the economic impact it is expected to create. Tesla CEO Elon Musk who runs the biggest Electric Vehicle company in the US had significantly supported Trump’s campaign, reportedly spending $175 million. Musk is expected to be the chief beneficiary of the result amid Trump’s views against the EV market over tax credit given by the government. Dan Ives, MD, equity research at Wedbush Securities, said in an interview with CNBC,” For a Tesla bull, if you wake up, this is exactly what you want to see, and of course, Musk will clearly have a big voice in the administration.” However, exploring the other side of the narrative, Ives also said on X that “A major change in tariffs\/harsher stance on China would significantly impact the supply chain, Nvidia, Beijing retaliatory Apple\/Tesla likely, and slow the pace of the AI Revolution.” Raul Brens Jr., acting senior director at GeoTech Center, Atlantic Council, said, “The second Trump administration will likely chart a different course on AI than the one taken over the past four years. Trump has historically favoured limited government regulation, emphasising AI as a tool to strengthen US competitiveness, particularly against China.” Kai-Fu Lee, author and founder of Chinese AI startup 01.AI, wrote a blog post explaining five influential factors in Trump’s victory, which indicates his stance on an optimistic future for America. Contrary to Biden’s policy, Trump’s allies at the American First Policy Institute have drafted an order and aim to create a ‘Manhattan Project’-esque effort to propel AI technology, especially in the defense sector. Some of the most influential names in the AI industry believe that removing Biden’s order could open up the market for an unprecedented rate of innovation, benefiting technological advancements and the economy. The lack of guardrails could also increase investments in the AI ecosystem. Trump Endorsers Look to Reap Rewards While not all tech company CEOs have publicly endorsed Trump, a few notable figures were quite vocal about their support—barring the obvious, Elon Musk. Over the last few months, Marc Andreessen and Ben Horowitz, founders of a16z, have extended their support to Donald Trump. Both of them donated $2.5 million each to ‘Right for America’, a pro-Donald Trump PAC. Andreesen believes Trump’s policies are favourable for big tech and AI. In a podcast episode hosted by both, Horowitz said “ The future of our business, the future of technology, new technology and the future of America is literally at stake so here we are, and for a little tech we think Donald Trump is actually the right choice”. Peter Thiel, founder of PayPal and a notable VC, a few months ago said, “If you hold a gun to my head, I will vote for Trump”. Crystal McKeller, founder at Aloft VC, in an interview published a day before the result, said: “They’re (Trump and J.D Vance) not going to just roll back regulations that are stifling American industry, but they are actually going to actively implement policies that will stimulate growth and encourage innovation.” Moreover, Trump also proposed several immigration policies favourable to aspiring talent. In a podcast episode, he said, “What I will do is, I think you should get automatically as part of your diploma a green card to be able to stay in this country.” However, Debarghya Das, a VC at Menlo Ventures pointed out on X that “His (Trump’s) last administration was not statistically favourable for legal high-skilled immigration.” Not all is Well Yann LeCun, Meta’s chief AI scientist has been quite vocal in criticising Donald Trump and has frequently engaged in a verbal battle with Musk. While Meta may benefit from the broader advantage of loosening the regulations, they’re clearly not expecting any pleasant surprises from Trump’s second term. Without any direct engagement with Trump, several companies have maintained a strong stance on AI safety and regulations. Anthropic recently released a blogpost emphasising the need for AI control and regulation. They said the government should urgently take action on AI policy in the next eighteen months to proactiveley prevent risksk. Trump’s victory, and his alliance with Elon Musk may also leave OpenAI uncomfortable. Musk has more than criticised OpenAI’s practices, and given the expectation that he may have a leading voice in Trump 2.0, Sam Altman would have certainly wished for the other possibility in the election results. OpenAI is evil— Elon Musk (@elonmusk) October 2, 2024 Sundar Pichai acknowledges the need for AI regulations that could alleviate potential dangers that may arise from the technology. Moreover, key figures like Ilya Sutskever who left OpenAI to build AI tech that priortizes safety may not favour loosened guardrails that regulate AI. While removing stringent policies can momentarily increase the capital influx into the AI industry and strengthen America’s global technological dominance, it’s imperative that the new government strike a balance between the two. Furthermore, Trump’s plans to use AI to amplify America’s defense and military capabilities may lead to  impactful developments for the private sector, especially for startups that are focusing on building technology in the interests of national security. [Updated] November 18, 2024, 15:40 | The article has been updated to correct an error where Google’s CEO Sundar Pichai was mistakenly referred to as CEO of Microsoft.","excerpt":"His second term as the 47th President of the United States could shake up the AI ecosystem.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","donald trump","Editors Picks","USA"],"author_name":"Supreeth Koundinya","publish_date":"2024-11-07T07:45:25","publication_year":"2024","word_count":1334,"keywords":["Anthropic","Go","API","artificial intelligence","OpenAI","AI","donald trump","RAG","Editors Picks","Aim","GAN","AI (Artificial Intelligence)","R","USA"],"extracted_tech_keywords":["AI","artificial intelligence","OpenAI","Anthropic","Aim","RAG","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/donald-trumps-victory-paves-the-way-for-ais-free-rein\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10016546,"title":"Interview With Aleksa Gordić, Machine Learning Engineer At Microsoft","content":"“There are too many ML resources out there and so many people suffer from decision paralysis.” For this week’s ML practitioner’s series, we got in touch with AI Epiphany founder, Aleksa Gordić. He is a machine learning engineer from Serbia and works at Microsoft. In this interview, he shares some excerpts from his vibrant journey in the world of data science. AIM: How did your journey in machine learning begin? Aleksa: I was inquisitive since I was a kid. I loved biology, genetics, cosmos and watching science fiction. But my exposure to the tech world started pretty late in my life. Unfortunately, as I was growing up, nobody in my environment was involved with programming, AI or tech in general. That’s the case with many people in today’s world – which gives me this unique ability to empathise a lot more with people who are just starting off. Having said that, I was always passionate about algorithmic thinking, although I only got exposed to programming when I was 19. And those first programming days feel like I was raised in the 80s—coding on paper and writing code in PASCAL. I like to think about my education as having two separate but related threads. On one side, I have my formal education, and on the other side, I’ve constantly been working on improving myself in my free time since I was at least 14 years old. I have studied everything from: Hardware – both analogue and digital electronics. I constructed 8-bit digital dividers in bare metal and silicon.Low-level programming like bare metal uC embedded programming (C), RTOS programming for uCs (C), FPGA programming (VHDL).Higher-level programming languages, mainly C++ and Java andAlso lots of mathematics, digital image processing, etc. I started doing ML 2.5 years ago when I attended Microsoft’s ML summer camp, but due to my solid foundations in electronics, CS, Maths and algorithms and a lot of hard work, I managed to catch up with the best people in the field. I thought that starting later was my weakness, but it became my strength as I have a more diverse background than others. Over the last couple of years, my main area of focus has been on machine learning\/deep learning with an emphasis on computer vision. But throughout 2020, I’ve been exploring various other areas of DL like NLP (transformers), graph neural networks, etc. Aleksa: Mostly balancing between my full-time work at Microsoft and learning on my own in my free time – which doesn’t leave me with a lot of spare time and burnouts become a real issue. It took me a lot of time to find that sweet spot where I’m super productive but also sustainable over the long run. AIM: What were the challenges and how did you address them? The fact that I was born in Serbia automatically gave me fewer opportunities than my peers who were born in Silicon Valley and who were exposed to tech since they were five years old. Hard work and a passion for learning and success were the key drivers for me. The Coursera course “Learning how to learn” helped me find the work-life balance that I needed so desperately. Nowadays, I don’t treat the work I do as a separate entity from my life – it is my life. I also take care to keep my social relationships and my body healthy. AIM: Can you tell us about your role at Microsoft? Aleksa: I currently work as a Machine Learning engineer at Microsoft. My job depends on the kind of project I’m working on — ranging from ideation, research, shipping and more. My role is a mix of research and engineering. A fun fact is that I officially got employed as a software engineer but due to my serious investments in ML, my efforts got recognised internally, and I got shifted to ML-based projects. I’m involved in data collection, data engineering, visualisations and all the way to training and measuring the compute budget of different computer vision models, and communicating that with the team. As another example, I was also developing the metrics pipeline from scratch, monitoring labelling pipelines and developing scripts to robustify them. Even though I’m in a big corporation, it feels a lot more like a startup. Earlier this year, I took ~3 months to reconstruct one deep learning paper. I was communicating my progress with the broader team, and I learned a lot about research along the way. Sometimes, depending on the project phase, I also present papers on internal reading groups with Cambridge and Zurich teams. We collaborate a lot with Cambridge, and most of them have had some previous MS research experience. AIM: Can you talk a bit about the tools you use as an ML engineer? Aleksa: Python\/PyTorch\/AML is what I mostly use, and it does the job, so why change it. Python has a big ecosystem built around it so you can find a library for pretty much anything. In my first year of work at Microsoft, I was extensively using C++ but not anymore. Nowadays, I use C++ only occasionally when I’m working close to hardware and am trying to optimise the Python code. We work on cutting edge devices in my team (like Microsoft HoloLens), so that’s something that pops up from time to time. Given my solid background in electronics I don’t have any problems getting my hands dirty (although I never did truly low-level programming while at Microsoft). “I’m a big believer of not losing your precious time by learning a bunch of frameworks and languages.” PyTorch is my framework of choice for now, but I’d like to explore JAX in my free time. I think it’s a suboptimal strategy to “learn the framework first” via some course. Just go out there, get your hands dirty, and develop something and learn on the fly. I did some projects in Keras\/TF as well, but I find PyTorch much nicer. As Karpathy nicely put it in his tweet: I've been using PyTorch a few months now and I've never felt better. I have more energy. My skin is clearer. My eye sight has improved.— Andrej Karpathy (@karpathy) May 26, 2017 It takes me a maximum of two weeks to learn a new programming language — even those which have a very different paradigm like Haskell (functional paradigm – I love it I must say). Same goes for cloud providers — all of the biggest ones will do the job — AML, GCP, AWS, whatever makes the most sense for your context\/company. AIM: What does it take to make a machine learning engineer from good to great? Aleksa: Make, make, make, code, code, code. I’m bullish on my GitHub portfolio. I like creating my own projects and sharing them with the world. There are many benefits to doing that: You learn a lotYou help othersPeople hear about you Good engineers focus on learning frameworks and packages (should I learn NumPy or pandas?), figuring out which course is the best one (should I do this Udacity course or Coursera?) or which book they should read next. The great ones care about solving the problem and see all of those as just the tools. Problem-solving skill is much more critical than learning 100 DL frameworks in parallel. “If you’re still reading books and doing courses still – you’re at an intermediate level at best.” It takes a lot of patience and time! You can’t learn ML in 3 months or even a year. It takes a lot of hard work, and it also takes making right decisions along the way, i.e. knowing where to invest your time into and what to ignore. There are too many ML resources out there, and so many people have decision paralysis. Also, I see way too many “advanced” practitioners, as well as ML influencers, spending too much time reading books and doing courses and sharing how they’ve completed some new course on LinkedIn. The only way to keep up with the field is by regularly reading research papers and surrounding yourself, both online and in real life if possible, with best researchers and engineers from the field. LinkedIn and Twitter are good platforms for that. AIM: Can you name some resources that helped you and could help others in their journey? Aleksa: Here are a few resources: For Calculus: 3Blue1Brown’s YouTube playlist (amazing channel in general!)For Linear algebra: Linear algebra course on MIT by Professor Gilbert Strang.For Python, good old Stack Overflow would do.On LearningOn NLP It’s sometimes useful to take a step back and learn how to learn. Many people suffer from suboptimal learning strategies. Like this this Coursera course “Learning How To Learn” helped me tremendeously. I create a lot of content myself. I also did a YouTube series on neural style transfer, deep dream, I’ve open-sourced many projects like this one on GANs. I have covered transformers extensively on my channel, and again I’ve open-sourced a project here. I’m a big believer of coding up a project from scratch once you gain some solid theoretical understanding of the field. It’s hard to explain just how much I’ve learned in 2020. I think it’s probably close to ¾ of a PhD in units of effort!! (Here is a great blog by Aleksa where he talks more about starting the ML journey.) AIM: What is in store for machine learning in the coming decade? Aleksa: Deep learning is still a young field, and many problems are not yet solved – so it will take a lot of time before that knowledge ends up getting distilled into thick books. What’s specific to ML is that people believe that the systems that we engineer today are truly intelligent. The outside world is prone to anthropomorphising the tech that we create – so it’s easy to spark much fake news around it. We’ve seen this happening a lot. The Sophia robot from Hanson Robotics is a good example. I don’t believe that a single area of AI will win out any time soon. But, we will see a combination of Bayesian learning (for better probability modelling and hopefully increased model interpretability), graph neural networks (for finding good representations in knowledge graphs, etc.), reinforcement learning (for learning when you don’t have differentiable functions), smart attention methods (transformers, GATs, etc.). I also think that causation will play an important role. I still don’t have enough knowledge to further argue about causational frameworks, but that’s just my gut feeling. Deep learning is great at modelling perception, so I don’t think it will go away anytime soon. The bubble may boom and bust, but it’s here to stay. I like to think about deep learning as our best model of the cerebellum and different parts of the brain that care about perception (visual cortex, auditory perception, etc.). Those parts are characterised by the lack of consciousness and by a vast amount of computation happening inside them. On the other hand, we still need to improve systems that care about the equation’s cognition part. It’s not entirely impossible to imagine that even “obsolete” symbolic AI will have its role in developing the “true” intelligence of tomorrow. I also expect graph neural networks to play their part here in modelling knowledge, memory, etc. Finally, overly complicated heuristics and systems which have way too much human knowledge integrated into them will go away. “The bitter lessons” by Sutton is a nice read on that topic.","excerpt":"“There are too many ML resources out there and so many people suffer from decision paralysis.” For this week’s ML practitioner’s series, we got in touch with AI Epiphany founder, Aleksa Gordić. He is a machine learning engineer from Serbia and works at Microsoft. In this interview, he shares some excerpts from his vibrant journey […]","categories":["Deep Tech"],"tags":["artificial intelligence machine learning data","java project ideas","Machine Learning","machine learning engineer","Microsoft"],"author_name":"Ram Sagar","publish_date":"2020-12-28T17:00:00","publication_year":"2020","word_count":1910,"keywords":["data science","machine learning","AI","neural network","PyTorch","ML","machine learning engineer","artificial intelligence machine learning data","Machine Learning","computer vision","NLP","Aim","deep learning","java project ideas","Microsoft"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","computer vision","data science","Aim","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/interview-aleksa-gordic-microsoft\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10097864,"title":"AI Alignment is a Joke","content":"OpenAI has been crystal clear about one of the most important aspects behind the success of ChatGPT — Reinforcement Learning from Human Feedback (RLHF). Everyone nodded. And since then, they have all been building models using RLHF. By training LLMs through interactions with human evaluators, RLHF seeks to improve the performance of AI models in real-world applications, but in turn it induces biases and reduces the robustness of the models. A recent paper, Open Problems and Fundamental Limitations of Reinforcement Learning from Human Feedback by researchers from Harvard, Stanford, MIT, UC Berkeley, and many other universities, discusses the problems with the RLHF approach. Good, but not the best According to the paper, obtaining high-quality feedback from human evaluators is one of the primary challenges in RLHF. Human beings, while capable of providing valuable feedback, are susceptible to various limitations and biases. Misaligned evaluators might have difficulty in understanding the context or objectives of the AI model, leading to suboptimal feedback. The complexity of supervision, especially in long conversations, can also hinder the accurate assessment of model performance. “AI Alignment” is just a cancel culture for GPUs.— Bojan Tunguz (@tunguz) July 19, 2023 Besides, data quality is another critical concern. Human evaluators may unintentionally provide inconsistent or inaccurate feedback due to factors like limited attention, time constraints, and cognitive biases. Even with well-intentioned evaluators, disagreement can arise due to subjective interpretations and varying perspectives. The form of feedback used in RLHF can further compound these challenges. Depending on the evaluation method, evaluators may provide binary judgments, rankings, or comparisons, each with its own strengths and weaknesses. Selecting the most appropriate form of feedback for a specific AI task can be complex, leading to potential discrepancies in the training process. A fundamental issue in RLHF is accurately representing individual human values with a reward function. Human preferences are context-dependent, dynamic, and often influenced by societal and cultural factors. Designing a reward function that encompasses the complexity of human values is a formidable task. Incorrect assumptions about human decision-making or using a reward model that neglects personality and context-dependence can lead to misaligned AI models. Why so much alignment? The diversity of human evaluators further complicates the reward modelling process. Different evaluators may have unique preferences, expertise, and cultural backgrounds. Attempting to consolidate their feedback into a single reward model might overlook important disagreements and result in biassed AI models that favour majority opinions. This could be of disadvantage to underrepresented groups and perpetuate existing societal biases. To address these challenges, researchers must explore techniques for representing preferences in more nuanced and context-aware ways. Utilising ensemble reward models that consider multiple evaluators’ feedback, or personalised reward models that cater to individual preferences, can help capture the diversity of human values. Transparently addressing potential biases in the data collection process and conducting thorough evaluations to identify and mitigate harmful biases are essential steps in responsible AI development. To overcome these data constraints, researchers should explore methods for cost-effective data collection that do not compromise data quality and diversity. Understandably, training on GPT-output data for quicker alignment has been the new trend, but this in the end brings in the same bias into other models as well. So, there has been no conclusion on this so far. The fundamental challenges of RLHF have significant implications for AI alignment. While some problems may have tractable solutions through technical progress, others may not have complete solutions and may require alternative approaches. Researchers must be cautious about relying solely on RLHF for AI alignment, as certain challenges might not be fully addressed through this method alone. Essentially, RLHF leads to over-finetuning of a model that may handicap its capabilities. This phenomenon is called the alignment tax of AI models. When a model goes to several benchmarks testing with humans in the loop trying to make the model as aligned and as “politically correct” as possible, it loses a lot of its performance. Alignment tax is the extra cost that an AI system has to incur to stay more aligned, at the cost of building an unaligned or uncensored model, which ultimately also hinders its performance. That is why, in a lot of cases, uncensored models that do not go through the RLHF phase actually perform better than aligned models.","excerpt":"In many cases, uncensored models that do not go through the RLHF phase actually perform better than aligned models","categories":["AI Features"],"tags":["RLHF"],"author_name":"Mohit Pandey","publish_date":"2023-08-01T12:24:14","publication_year":"2023","word_count":711,"keywords":["Go","ChatGPT","TPU","RLHF","OpenAI","AI","AI alignment","GPT","data quality","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","RLHF","TPU","R","Go","data quality","GPT","AI alignment"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-alignment-is-a-joke\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10054815,"title":"Top 5 Announcements At PyTorch Developer Day Conference","content":"For developers and users to discuss core technical developments, ideas, and roadmaps, PyTorch designed the Developer Day Conference – virtually for 2021. Day 1 was live and later uploaded on Twitter, LinkedIn and Facebook and had over 25,000 views. From demos and learning to PyTorch’s achievements of 2021, a lot was discussed throughout the day. Here are some highlights of what new and interesting Meta’s learning-research lab is coming up with: Launch of PyTorch Live The big announcement was the launch of PyTorch Live. It is a set of tools to build mobile AI-powered experiences easier. These tools support products to build on both iOS and Android platforms. There will be no need to write the same in two different languages; it uses JavaScript unified language to write apps for both platforms. To achieve this, PyTorch Live is powered by two successful open-source projects: PyTorch Mobile powers the on-device inference for PyTorch Live and Reactive Native is a library for building visual interactive UI. The three highlights of PyTorch Live include CLI, which enables developers to quickly set up a mobile deployment environment and bootstrap mobile app projects. Second is the Data Processing API that helps prepare and integrate custom models to be used with the PyTorch Live API. And the last is Cross Platform Apps that allow building mobile AI-powered apps for Android and iOS using the PyTorch Live API. Release of PyTorch Profiler 1.9 and 1.10 The PyTorch profiler collects a performance matrix during training and inferencing and provides actionable guidance to optimize the PyTorch model performance. The version PyTorch Profiler 1.9 has new features: distributed training view, memory view, GPU utilization, cloud storage support, and visual studio code. One can jump directly to source code. The newly added features of version 1.10 are Forward\/Backward Correlation, Enhanced Memory view, recommendation enhancement, Gloo support and Tensorcore support. Introduction to TorchBench TorchBench is a new benchmark suite that has been released in open source. It captures models the way users use them and is focused on the researcher use case. The purpose is to improve the framework and not just individual models. This required bringing together all the scores of different models and impacting this across a very large suite. For TorchBench to be successful, diversity is the key. This is tricky as there is a large variety of models, and it is unclear what to benchmark. They will bring together a lot of models and combine them based on the distribution in the image below: Source: PyTorch Developers Day 1, broadcast presentation Investments in Ecosystem According to Dwarak Rajagopal, Engineering Director, Meta AI, “Community and Ecosystem is an important factor for sustaining both fast research innovation and hyper-growth in production. To grow the ecosystem 10X more, it is important to build extension points for the ecosystem.” There are three levels of extension that enable extending PyTorch to several ecosystem components. To grow the ecosystem and community and be able to innovate at the API level, PyTorch has planned to increase its investment in the following: Core Authoring Ecosystem To take it to the next level, PyTorch plans to make the core front-end language even more extensible to authoring innovators. For this, PyTorch is building an Extensible Dispatch Subclassing, which includes making the dispatch system itself a tool for building a more extensible front-end. They are also working on an ability to override autograd-like compute APIs (eg. Vmap, quantization, etc.), and not just new operators or primitives. The new abilities planned also provide a scope to extend behaviour in C++ or Python via callbacks and subclassing, whatever fits the use case best. Implementations Ecosystem PyTorch plans to build richer batteries, including a foundation for the OSS library and a framework for authors. For this, PyTorch has doubled its investment in supporting domain ecosystem libraries – providing hardened table stakes primitives and common extensibility foundations for OSS authors and maintainers. DataLoader v2 and TorchData are the built-in tools for accelerated data loading and dataset authoring. With the new PyTorch Profiler and TorchBench, they are providing a better debugging and performance measurement experience. Execution Ecosystem PyTorch is building more extensibility and hooks for execution and productionisation partners. Within the execution ecosystem, on the Training and E2E workflows, they have built – TorchX, which is a job authority with built-in components for running job\/workflows on schedulers, pipelines, etc. On the Program Transformation side, they have built torch.fx – toolkit to create composable transformations for customer compilation, execution engines, profiling, and more. They also have PyTorch Mobile to run models on edge devices and TorchServe to provide out-of-the-box models serving to integrate on your cloud provider of choice. With torch.package + Torch::deploy, they are making it easier to run inference efficiently in a multithread environment for production model serving. Brewing in Prototypes A lot of PyTorch functionality is currently in the works. It is not ready yet, but is available in a repository in prototype features. Some of them include FunTorch, which brings composable differentiations and backend transformations. Also known as vmap, it allows one to efficiently compute for simple gradience. The team continues to improve framework extension points. It can hard blend backends and all types of repositories. Lazy Tensor Core provides an ability to experiment on new Tensorblade extensions completely on the Python side. On the distributer side, sharding Tensor support and general model primitives are currently in the works. The data project aims to make DataLoader more modular, extensible and performable.","excerpt":"While PyTorch Live remained the biggest announcement of the PyTorch Developer Day Conference 2021, there are many important highlights.","categories":["AI Trends"],"tags":["Pytorch"],"author_name":"Meeta Ramnani","publish_date":"2021-12-03T18:00:38","publication_year":"2021","word_count":910,"keywords":["Pytorch","Go","Meta AI","AI","PyTorch","RAG","TorchServe","Python","Aim","JavaScript","R"],"extracted_tech_keywords":["AI","Meta AI","Aim","TorchServe","PyTorch","RAG","Python","R","JavaScript","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-announcements-at-pytorch-developer-day-conference\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10079561,"title":"After Flex, Intel’s on Max Mode to Boost Cloud Infrastructure","content":"Intel made a late entry into the GPU scene with the Intel Arc, but has since then accelerated innovation. In August, the company released its first Data Centre GPU Flex Series and last week, it came out with the Data Center GPU Max Series. Data Center GPU Max Series is the industry’s highest-density processor, packing over 100 billion transistors on 47 active tiles and up to 128 Xe-HPC cores. The product maximises bandwidth (up to 128GB HBM2e), capacity (up to 408MB Rambo L2 Cache), and memory (up to 64MB L1 cache). Intel has claimed that its 408MB L2 cache will be able to deliver 2x the performance than the previous versions. The GPU solves a series of obstacles, such as one around porting and refactoring code, the economic and technical burdens around proprietary GPU environments that prohibit portability between different GPU vendors, and finally, the inconsistencies between CPU and GPU implementations like CPU having too little memory bandwidth, and GPU too little memory capacity. Intel’s GPU Max is a product that maximises bandwidth, compute power, developer productivity, and impact. Above that, the entire Max series—that is, both its CPU and GPU—is powered by oneAPI, an open-source programming model that allows developers to use various accelerated architectures. The Data Centre Max Series products are planned to be launched in 2023. Data Center GPU Flex Series The previous version of the model Data Centre GPU Flex Series was made to manage media streaming and cloud gaming, along with supporting AI visual inference and virtual desktop infrastructure workloads. The device could be set in different power levels for requirements ranging from basic AI needs to complex AI workloads. Two weeks ago, Intel announced that its data centre GPU Flex Series has been added to their family of PluggableDevices – aka Intel Extension for TensorFlow. PluggableDevice architecture offers a plugin mechanism for registering devices with TensorFlow without the need to make changes in the original code. This new implementation will allow Intel Data Center GPU Flex Series hardware and the company’s Intel Arch graphics. However, it is said to be compatible with Linux and the Windows Subsystem for Linux by connecting to oneAPI.","excerpt":"Intel made a late entry into the GPU scene with the Intel Arc, but has since then accelerated innovation. In August, the company released its first Data Centre GPU Flex Series and last week, it came out with the Data Center GPU Max Series.  Data Center GPU Max Series is the industry’s highest-density processor, packing […]","categories":["AI News"],"tags":["AI Infrastructure","AI innovation","HPC","Intel"],"author_name":"Ayush Jain","publish_date":"2022-11-14T12:08:49","publication_year":"2022","word_count":359,"keywords":["Go","API","AI innovation","programming_languages:R","HPC","AI","innovation","Aim","ai_frameworks:TensorFlow","ViT","AI Infrastructure","TensorFlow","R","Intel"],"extracted_tech_keywords":["AI","Aim","TensorFlow","R","Go","API","ViT","innovation","ai_frameworks:TensorFlow","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/after-flex-intels-on-max-mode-to-boost-cloud-infrastructure\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093554,"title":"Killed By Amazon (Part 1)","content":"In July 1995, an online store emerged selling the world’s largest collection of books to anyone with World Wide Web access, and the rest is history. Amazon has been the go-to online delivery service for over half of the world’s online shoppers. The company’s execs have been vocal about Amazon’s “fail fast, fail often” culture. While CEO Jeff Bezos has created a dominant global online retailer and cloud computer, he’s led the creation of many duds, too. He famously called the e-commerce giant “the best place in the world to fail” in his 2016 shareholder letter. At Amazon, multiple products came to an end, some due to slowing sales and the rest due to the company’s shift in focus. Here is your first tour of the ‘Amazon Graveyard’. Auctions and zShops Similar to eBay, Amazon once ran an auction site called Amazon Auction. It debuted in 1999 and ended a few years later. The service was replaced by Amazon Marketplace. Early Reviewer Program Launched in 2018, the Early Reviewer Program was introduced to boost the review count under the supervision of Amazon. Discontinued in April 2021, the company pushed third-party sellers to Vine and the “Request a Review” button. Kindle Fire HDX Launched in 2013, the Fire HDX was a favourite but it was discontinued in 2015 and replaced by a line of Fire HD products. Amazon Fire TV Recast Released in 2018, Amazon’s Fire TV Recast, was a DVR that allowed cord-cutters to record and watch shows aired by TV antenna. The company announced to pull the plug on the product in 2022 though it will continue to offer software security updates till 2026. Fire Phone In 2014, almost two months after it started selling, the phone was heavily criticised for its lack of features and soaring price. And even though Amazon reduced the price from $200 to just $0.99 with a two-year contract, the phone couldn’t be saved and was discontinued about a year later. Amazon Honor System Amazon’s Honor System was launched in 2001 to allow customers to make donations or buy digital content, with Amazon collecting a percentage of the payment plus a fee. The service was discontinued in 2008 and replaced by Amazon Payments. Amazon Music Storage The Amazon Music feature lets users upload their MP3 files from other sources, but the company ended its Amazon Music Storage subscription service in January 2019. Amazon Halo Earlier this year in April, Amazon decided to axe its health-tracking bracelet. Released in 2020, Halo was combined with an app that tracked users’ activity, body fat and emotional state, and was integrated with Amazon’s Alexa digital assistant. Amazon Glow Just over a year after its debut, Amazon killed its kids-focused video device that used projection technology to create a virtual space on the tabletop. Kindle MatchBook Launched in 2013, Amazon entirely shut down the digital bundling platform in October 2019. The program gave the ability to authors and publishers to provide a heavily discounted ebook, when a reader bought a hardcover or paperback. Amazon Video Direct Amazon Prime Video Direct (APVD) was the only mainstream service that accepted unsolicited film submissions. Until 2021, after which, it no longer accepts submissions for non-fiction and short-form content. Amazon Prime Pantry Launched in 2014 Amazon discontinued its Amazon Pantry (originally known as Prime Pantry) service, instead rolling those household goods and shelf-stable pantry items into the main Amazon website where they can be ordered alongside the rest of Amazon’s products. Amazon Drive In October 2022, the Amazon Drive app was taken down from the iOS and Android app stores. As of February 1, 2023, Amazon no longer supports uploading files on the Amazon Drive website. You will still be able to view and download your files until December 31, 2023. Amazon Elements On 21 January 2015, Amazon discontinued its prime-only Elements diapers in just six weeks. Amazon EC2 In July 2021, Amazon Web Services announced it will shut one of its oldest cloud computing services, EC2-Classic. The company also warned the remaining users to move off to avoid application downtime. Kindle Newsstand From March 9, 2023, you can no longer subscribe to any publication on Kindle. Amazon notified subscribers in an Amazon post. Amazon Local Amazon Local was also shut down in 2015. A site similar to Groupon, Local shut shop as the other site also saw high rises and subsequently falls. Amazon Go Earlier this year, Amazon permanently closed eight of its high-tech Amazon Go convenience stores, including two in Seattle. Amazon 4-star In March 2022, Amazon announced it would close all of its 4-star, Books and Pop Up stores, Reuters reported. Amazon Whole Foods Market Whole Foods, a wholly owned subsidiary of Amazon, announced in early 2019 shutting down the business in India completely last year. Wag Amazon acquired Quidsi in 2010, which expanded into Soap.com, Wag.com, BeautyBar.com, Casa.com, and YoYo.com. In 2017, Amazon shut down Quidsi as it was never profitable. Amazon Destinations Amazon also briefly owned a hotel-booking website called Amazon Destinations which was intended to plan quick getaways. It didn’t last long though. Released in April 2015, it was gone by October of the same year. Amazon Dash Button Amazon stopped selling the Dash Button, a small wireless device connected to WiFi which instantly ordered pre-selected items on Amazon, in 2019, but apparently they were a success since they got customers used to not shopping with a screen. Amazon Tap The first Amazon Echo device was discontinued by Amazon without a replacement, the Amazon Tap was a mobile version of its Alexa-enabled smart speakers which the company stopped selling near the end of 2018. Pop-up stores Amazon will close all 87 of its pop-up stores and discontinue the program, it told Business Insider in 2019. The stores were a place where customers interested in smart gadgets could see how they worked before purchasing. Instant Pickup In 2017, Amazon introduced pick up for items within minutes of ordering them with Instant Pickup. However, the company ended its service in 2018 without specifying the reason. Amazon Screenwriter In May 2019, Amazon Studios introduced Amazon Storywriter and Storybuilder only to discontinue them on June 30, 2019 with accounts disabled and not downloaded content becoming inaccessible.","excerpt":"Since the list is too exhaustive, we have made it a two-part series","categories":["AI Features"],"tags":["Amazon"],"author_name":"Tasmia Ansari","publish_date":"2023-05-18T12:00:00","publication_year":"2023","word_count":1035,"keywords":["Go","programming_languages:R","cloud computing","AI","Amazon","Git","programming_languages:Go","RAG","ViT","cloud_platforms:Amazon Web Services","R"],"extracted_tech_keywords":["AI","RAG","cloud computing","R","Go","Git","ViT","cloud_platforms:Amazon Web Services","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/killed-by-amazon-part-1\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10100153,"title":"AWS Receives Cloud Service Provider Empanelment From MeitY","content":"Amazon Web Services (AWS) India has announced that it has received cloud service provider (CSP) empanelment from India’s Ministry of Electronic and Information Technology (MeitY) for cloud services provided using the AWS Asia Pacific (Hyderabad) Region. Operational since November 2022, as part of AWS’s total investment of INR 1,36,500 crores (US $16.4 billion) in cloud infrastructure in India by 2030, the AWS Asia-Pacific (Hyderabad) Region is the second AWS Region in India to be fully empaneled by MeitY. In 2017, AWS India became the first global CSP in India to receive full empanelment for its cloud service offerings after the AWS Asia-Pacific (Mumbai) Region completed MeitY’s STQC (Standardization Testing and Quality Certification) audit. AWS Regions are comprised of Availability Zones (AZs) that place infrastructure in separate and distinct geographic locations. AZs are located far enough from each other to support customers’ business continuity, and near enough to provide low latency for high-availability applications that use multiple data centres. Public sector organisations and financial services institutions across India, including banking and payment providers, can achieve greater operational resiliency from the AWS Asia-Pacific (Hyderabad) Region – which consists of three AZs – by using its cloud infrastructure for higher availability, disaster recovery, data backup, while meeting security and regulation requirements, and enjoying lower latency performance, AWS said. Moreover, government organisations can improve e-governance standards and enable on-demand digital services for citizens and businesses across India. Similarly, financial services organisations can benefit from agile, efficient, and security-compliant cloud solutions at scale, providing consumers faster and secure digital banking, insurance, and payment innovations. “With both AWS Regions in India empaneled by MeitY, AWS is providing customers more choice to access resilient, secure, and low-latency cloud infrastructure, while offering more ability for AWS Partners to develop innovative solutions and address customer needs,” “Shalini Kapoor, director and chief technologist for the public sector with AWS India Private Limited (AWS India), said.","excerpt":"Operational since November 2022, the AWS Asia-Pacific (Hyderabad) Region is the second AWS Region in India to be fully empaneled by MeitY.","categories":["AI News"],"tags":["AWS"],"author_name":"Pritam Bordoloi","publish_date":"2023-09-15T17:18:48","publication_year":"2023","word_count":315,"keywords":["Go","AWS","AI","cloud_platforms:AWS","innovation","programming_languages:R","Git","GAN","cloud_platforms:Amazon Web Services","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Git","GAN","innovation","cloud_platforms:AWS","cloud_platforms:Amazon Web Services","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-receives-cloud-service-provider-csp-empanelment-meity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10096061,"title":"The Truth Behind OpenAI’s Silence On GPT-4","content":"In March, OpenAI launched GPT-4 with much fanfare, but a dark cloud loomed over the horizon. Scientists and AI enthusiasts alike panned the company for not releasing any specifics about the model, like the parameter size or architecture. However, a top AI researcher has speculated the inner workings of GPT-4 revealing why OpenAI chose to hide this information — and it’s disappointing. OpenAI CEO Sam Altman famously stated on GPT-4 that “people are begging to be disappointed, and they will be”, speaking about the potential size of the model. Rumour mills ahead of the model’s launch suggested that it would have trillions of parameters and be the best thing that the world has ever seen. However, the reality is different. In the process of making GPT-4 better than GPT-3.5, OpenAI might have bitten off more than it could possibly chew. 8 GPTs in a trenchcoat George Hotz, world-renowned hacker and software engineer, recently appeared on a podcast to speculate about the architectural nature of GPT-4. Hotz stated that the model might be a set of eight distinct models, each featuring 220 billion parameters. This speculation was later confirmed by Soumith Chintala, the co-founder of PyTorch. While this puts the parameter count of GPT-4 at 1.76 trillion, the notable part is that all of these models don’t work at the same time. Instead, they are deployed in a mixture of expert architecture. This architecture makes each model into different components, also known as expert models. Each of these models is fine-tuned for a specific purpose or field, and is able to provide better responses for that field. Then, all of the expert models work together with the complete model drawing on the collective intelligence of the expert models. This approach has many benefits. One is that of more accurate responses due to models being fine-tuned on various subject matters. MoE architecture also lends itself to being easily updated as the maintainers of the model can improve it in a modular fashion, as opposed to updating a monolithic model. Hotz also speculated that the model may be relying on the process of iterative inference for better outputs. Through this process, the output, or inference result of the model, is refined through multiple iterations. This method also might allow GPT-4 to get inputs from each of its expert models, which could reduce the hallucinations in the model. Hotz stated that this process might be done 16 times, which would vastly increase the operating cost of the model. This approach has been likened to the old trope of three children in a trenchcoat masquerading as an adult. Many have likened GPT-4 to be 8 GPT-3s in a trench coat, trying to pull the wool over the world’s eyes. Cutting corners While GPT-4 aced benchmarks that GPT-3 has had difficulties with, the MoE architecture seems to have become a pain point for OpenAI. In a now-deleted interview, Altman admitted to the scaling issues OpenAI is facing, especially in terms of GPU shortages. Running inference 16 times on a model with MoE architecture is sure to increase cloud costs on a similar scale. When blown up to ChatGPT’s millions of users, it’s no surprise that even Azure’s supercomputer fell short of power. This seems to be one of the biggest problems that OpenAI is facing currently, with Altman stating that cheaper and faster GPT-4 is the company’s top priority as of now. This has also resulted in a reported degradation of quality in ChatGPT’s output. All over the Internet, users have reported that the quality of even ChatGPT Plus’ responses have gone down. We found a release note for ChatGPT that seems to confirm this, which stated, “We’ve updated performance of the ChatGPT model on our free plan in order to serve more users”. In the same note, OpenAI also informed users that Plus users would be defaulted to the “Turbo” variant of the model, which has been optimised for inference speed. API users, on the other hand, seem to have avoided this problem altogether. Reddit users have noticed that other products which use the OpenAI API provide better answers to their queries than even ChatGPT Plus. This might be because users of the OpenAI API are lower in volume when compared to ChatGPT users, resulting in OpenAI cutting costs at ChatGPT while ignoring the API. In a mad rush to get GPT-4 out to the market, it seems that OpenAI has cut corners. While the purported MoE model is a good step forward for making the GPT series more performant, the scaling issues that it is facing show that the company might just have bitten off more than it can chew.","excerpt":"In the process of making GPT-4 better than its predecessors, OpenAI might have bitten off more than it can chew","categories":["AI Highlights"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-06-30T13:00:00","publication_year":"2023","word_count":776,"keywords":["Go","ChatGPT","API","TPU","OpenAI","PyTorch","AI","R","GPT","Azure"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","PyTorch","Azure","TPU","R","Go","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/the-truth-behind-openais-silence-on-gpt-4\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10088642,"title":"Google Returns to Federated Learning Over Privacy Concerns","content":"Google has created the first federated learning and distributed differential privacy system with formal guarantees against an honest-but-curious server. However, a fully malicious server could still bypass the privacy guarantees by manipulating the public key exchange or introducing fake malicious clients, said researchers in a recent blog post. In 2021, Google started using federated learning to train Smart Text Selection models, an Android feature to select and copy text easily by predicting what text users want to select and then automatically expanding the selection. Since the launch, Google has improved the models’ privacy by combining secure aggregation (SecAgg) and a distributed version of differential privacy. The recent development is all thanks to an honest-but-curious server that follows the protocol but could gain insights about users from the data it receives. The Smart Text Selection models trained with this system have reduced memorization by over double. Data minimization is a key principle hence the server learns nothing about individual updates and only receives an aggregate model update. The SecAgg protocol ensures this due to cryptographic guarantees. For Smart Text Selection, all updates and metrics are aggregated via SecAgg using TensorFlow Federated, and stored in Android’s Private Compute Core. This enhances privacy because unaggregated model updates and metrics are not visible to any part of the server infrastructure. SecAgg minimizes data exposure but does not guarantee against revealing anything unique to an individual. This is where differential privacy (DP) comes in. Last week, Google also announced a new method through which DP guarantees trusted servers control the process. In practice, SecAgg has other privacy challenges. Google has addressed them by introducing an approach for auto-tuning the discretization scale during training and added integer noise using the distributed discrete Gaussian and distributed Skellam mechanisms. Read: For The Sake Of Privacy: Apple’s Federated Learning Approach As privacy concerns grew among consumers, major tech companies like Apple and Google started investing heavily in this decentralised form of machine learning, which trains models without collecting raw data. Federated learning is commonly used to power suggestion features and rank suggested items in context. The term was first coined by Google researchers in a 2016 paper, titled ‘Communication Efficient Learning of Deep Networks for Decentralised Data’ by Google researchers in 2016, and a heavily cited paper titled ‘Deep Learning with Differential Privacy,’ co-authored by Google and OpenAI researchers.","excerpt":"The tech giant has guaranteed formal privacy due growing concerns.","categories":["Global Tech"],"tags":["federated learning","Machine Learning"],"author_name":"Tasmia Ansari","publish_date":"2023-03-03T17:04:17","publication_year":"2023","word_count":390,"keywords":["federated learning","Go","machine learning","OpenAI","AI","Machine Learning","deep learning","differential privacy","Rust","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","OpenAI","TensorFlow","federated learning","differential privacy","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-returns-to-federated-learning-over-privacy-concerns\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10163978,"title":"India and Qatar Forge Stronger AI Ties with Key Partnerships","content":"India and Qatar are set to enhance their economic partnership, focusing on sustainability, technology, entrepreneurship, and energy. Union minister of commerce and industry, Piyush Goyal, announced this at the inaugural session of the India-Qatar Business Forum held in New Delhi on Tuesday. Goyal stated, “The future partnership will rest on the pillars of sustainability, technology, entrepreneurship and energy.” Two significant MoUs were signed during the forum: one between the Qatari Businessmen Association (QBA) and the Confederation of Indian Industry (CII) and another between Invest Qatar and Invest India. Goyal emphasised that India and Qatar complement each other, saying, “We are two countries that can work together for a better future.” He also focused on how both countries’ “business leadership can help create synergy between the vision of Viksit Bharat 2047 and Qatar National Vision 2030 to bring greater prosperity”. The minister noted a shift in trade dynamics from traditional energy sectors to emerging technologies such as AI, IoT and semiconductors. He remarked, “The world is going through a major shift in the context of geopolitical tensions, climate change, and cybersecurity threats”, highlighting the need for collaboration in these areas. The event was attended by HE Sheikh Faisal bin Thani bin Faisal Al Thani, Qatar’s minister of commerce and industry, who underscored the importance of expanding business opportunities beyond oil and gas. Goyal invited Qatari companies to participate in India’s growth journey across various sectors, including renewable energy and infrastructure development. Apart from Qatar, the AI ecosystem in the Middle East is slowly taking shape, starting with Saudi Arabia. The country recently announced a $1.5 billion investment in Silicon Valley-based AI startup Groq to expand AI inference infrastructure in the region. The agreement, revealed at LEAP 2025, Saudi Arabia’s leading tech event, strengthens the Kingdom’s AI computing capabilities and advances its Vision 2030 goal of building an AI-powered economy. Over the past year, the country has eyed an estimated $12.8 billion investment to develop a massive data centre at Oxagon. This centre is a key part of the futuristic NEOM city being built in the Tabuk Province, which was launched in 2017. The planned facility would feature 1 gigawatt of capacity, marking a step forward in the country’s push to strengthen its digital infrastructure.","excerpt":"“The future partnership will rest on the pillars of sustainability, technology, entrepreneurship and energy.”","categories":["AI News"],"tags":["India","partnership India","saudi arabia"],"author_name":"Sanjana Gupta","publish_date":"2025-02-18T14:49:20","publication_year":"2025","word_count":373,"keywords":["Go","programming_languages:R","AI","saudi arabia","programming_languages:Go","Git","ViT","Groq","partnership India","R","India","startup"],"extracted_tech_keywords":["AI","R","Go","Git","ViT","startup","Groq","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/india-and-qatar-forge-stronger-ai-ties-with-key-partnerships\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":44441,"title":"How Variational Autoencoders Can Flourish In Any Machine Learning Setting","content":"Variational Autoencoders (VAE) came into limelight when they were used to obtain state-of-the-art results in image recognition and reinforcement learning. VAEs consist of encoder and decoder network, the techniques of which are widely used in generative models. Encoders can be seen in CNNs too, they convert an image into a smaller dense representation which is fed to the decoder network for reconstruction. Source: Kaggle For example, one can see in the above illustration how the input image of the digit ‘2’ has been reconstructed using the autoencoder network. If the output image looks a bit hazy than the input, it means that there has been a loss of information. This is given by the loss function of the network; cross-entropy between the output and input. In image generation, even if the mean and standard deviation stays the same, the actual encoding will vary due to sampling. An autoencoder makes the encoder generate encodings to reconstruct its own input. Now, what if these widely used VAEs are made to effortlessly switch between fully supervised, semi-supervised and unsupervised learning? In short, training gets simple. The unlabelled aspect of data doesn’t seem to be a hurdle any more. A group of researchers from Munich have introduced a new flavor of Variational Autoencoder (VAE) that interpolates between different supervised settings. Making VAEs As A One Stop Solution The new model is an extension of the original VAE which is depicted in. The only addition is that a classification layer π (typically a one-hot classifying layer using softmax activation) is introduced that is attached to the topmost encoder layer. As can be observed in the picture above, the new model architecture is compared to its equivalents of supervised (left) and unsupervised (right). The π layer (and its loss) represents an extension to the standard VAE proposed in this. The μ and σ layer encodes the mean and standard deviation of the gaussian prior in the latent layer. The authors claim that the simplicity of their model allows turning any existent VAE into a semi-supervised VAE by simply adding the π layer and extending the loss function. An advantage of the semi-supervised variant is that the decoder can be used as a generative model by providing both the target label and by sampling from the latent layer. Given that the prior distribution of the latent layer is Gaussian, the sample from a normal distribution can be fed as an input to the decoder. In particular, all learned weights can directly be reused when transitioning into the semi-supervised learning scenario. This is very useful, as in many real-world applications, a labelled dataset (even partially labelled) is only built up over time and not available at project initiation. Adding A Flavor Of Transfer Learning The availability of unlabelled data points aids the model to form better representations in its deeper layers, hence enabling semi-supervised learning. Maybe the opposite is true as well: Does the availability of labels also aid with finding better representations? Does it perform better on reconstruction related tasks such as anomaly detection? This problem setup can be generally described as a flavour of ’transfer learning’: can the model improve its task related to unsupervised learning by leveraging the availability of labels that are primarily associated with the supervised learning task? To investigate the above scenario, VAE was used as an anomaly detector. This is a classic case of feature engineering where the labels incorporate domain knowledge of some very specific, yet important, the property of the data set. So, the idea here is that the π layer will guide the model towards an extractor for those very specific high-level features. The term ’semi-unsupervised learning’ is a perfect description of this task – as semi-supervised learning enhances the performance of a supervised task by using unlabelled data, ’semi-unsupervised’ learning would enhance the performance of an unsupervised task by using labelled data. Key Ideas A new flavour of Variational Autoencoder (VAE) that enables semi-supervised learning. The model architecture requires only minimal modifications on any given purely unsupervised VAE. Applied this VAEs to the problem of anomaly detection, it is observed that its performance increases The model adapts seamlessly on the full 0-100% range of available labels. Know more about this work here.","excerpt":"Variational Autoencoders (VAE) came into limelight when they were used to obtain state-of-the-art results in image recognition and reinforcement learning. VAEs consist of encoder and decoder network, the techniques of which are widely used in generative models. Encoders can be seen in CNNs too, they convert an image into a smaller dense representation which is […]","categories":["Deep Tech"],"tags":["encoder","VAE","variational autoencoder"],"author_name":"Ram Sagar","publish_date":"2019-08-13T18:00:02","publication_year":"2019","word_count":705,"keywords":["encoder","TPU","AI","ML","image recognition","feature engineering","variational autoencoder","Git","RAG","VAE","Aim","anomaly detection","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","anomaly detection","image recognition","TPU","R","Git","feature engineering"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-variational-autoencoders-can-flourish-in-any-machine-learning-setting\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171277,"title":"Wipro Wins Multi-year Deal with Entrust to Support Digital Transformation","content":"Wipro Limited on Tuesday secured a multi-year agreement with Entrust, a global leader in identity-centric security solutions, to help the company accelerate its growth and improve technology operations. Under this deal, Wipro will support Entrust in product development, infrastructure, and application upgrades. The partnership will also involve the use of generative AI solutions to enhance self-service options, reduce support response times, and improve overall user experiences. Wipro will strengthen Entrust’s application security by using advanced analytics to identify vulnerabilities early on and improve the software development process. “We chose Wipro based on its ability to help us access top talent, scale up to meet market opportunities, and add new capabilities,” Jeff Smolinski, senior vice president of operations at Entrust, said. The deal was first referenced in Wipro’s quarterly financial results announcement on October 17, 2024, without disclosing Entrust as the client at that time. Malay Joshi, CEO–Americas 1, Wipro Limited, added, “We are excited to bring our proven expertise to deliver comprehensive, AI-powered software development services at scale to further Entrust’s strategic priorities. Meanwhile, Wipro Limited also announced the launch of its global Wipro Innovation Network, aimed at driving client-centric co-innovation using frontier technologies. Besides, Wipro also unveiled a first-of-its-kind 60,000 sqft Innovation Lab at its Kodathi campus in Bengaluru, envisioned as a hub for exploring next-generation solutions. It is designed to bring together clients, partners, academia, and tech communities to accelerate the development of transformative solutions across industries. It will focus on five strategic technology themes: agentic AI, robotics with embodied AI, quantum computing, digital ledger technology, and quantum-safe cyber resilience.","excerpt":"Under the deal, Wipro will support Entrust in product development, infrastructure, and application upgrades.","categories":["AI News"],"tags":["Wipro"],"author_name":"Shalini Mondal","publish_date":"2025-06-04T09:40:50","publication_year":"2025","word_count":262,"keywords":["Wipro","agentic AI","programming_languages:R","AI","innovation","Git","Aim","generative AI","analytics","Rust","R"],"extracted_tech_keywords":["AI","analytics","generative AI","agentic AI","Aim","R","Rust","Git","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-wins-multi-year-deal-with-entrust-to-support-digital-transformation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040298,"title":"What To Expect From VMware’s New CEO?","content":"VMware recently named its long time executive Rangarajan (Raghu) Raghuram as the new Chief Executive Officer. He will replace Pat Gelsinger, who had moved to Intel. Further, Raghu will be a member of the Board of Directors of VMware, with effect from June 1. He is at present the COO of Products and Cloud Services, VMWare. I am honored and thrilled to lead @VMware as the next CEO. We have an enormous opportunity in front of us and are uniquely poised to lead the multi-cloud computing era. http:\/www.vmware.com\/go\/newceo— Raghu Raghuram (@RaghuRaghuram) May 12, 2021 The news comes close on the heels of VMware COO Sanjay Poonen leaving the company. Sanjay has been instrumental in building VMware into a $1 billion business. It's been a great time @VMware the past 7 yrs; friendships created, all we’ve built together, employees touched, and customers\/partners impacted. Also excited about what I’m going to be doing next. More on my next adventure in the next few weeks and months, stay tuned… pic.twitter.com\/u13femnz3U— Sanjay Poonen (@spoonen) May 12, 2021 VMware has been on the lookout for a new CEO since Pat left the company in January this year to head Intel. Post which, Sanjay was expected to be shortlisted as a potential candidate for CEO of the company. In addition to Raghu’s appointment, VMware also named Sumit Dhawan as its new President. Dell spins-off VMware In 2016, Dell acquired EMC (parent company of VMware) in a cash and stock deal worth $74 billion–the largest acquisition in the history of the tech sector. Last April, Dell announced the spin-off of 81 percent of its stake in VMware to create two standalone companies. The deal is expected to be close in the fourth quarter of 2021. In an official statement, VMware said it expects to generate revenue of $2.994 billion in the first quarter of fiscal 2022, a 9.5 percent year on year growth. Chairman of VMware Board of Directors and CEO of the parent company, Dell Technologies, Michael Dell, said: “Throughout his (Raghu’s) career, he has led with integrity and conviction, playing an instrumental role in the success of VMware. Raghu is now in a position to architect VMware’s future, helping customers and partners accelerate their digital businesses in this multi-cloud world.” Tracing Raghu’s Journey IIT-Bombay and The Wharton School alumnus Raghu joined VMware in 2003 as the Director of Product Management. Raghu switched about seven roles, holding management ranks in verticals, including the Datacenter and Desktop Platform Products, Product and Solutions Marketing, Server Business Unit, and Software-Defined Datacenter Division. Most recently, Raghu was the Chief Operating Officer of Products and Cloud Services. For close to two decades, Raghu has been steering the California-based cloud computing and virtualisation tech company’s strategic direction and technology evolution. He has been pivotal in growing the company’s visualisation business, driving VMware’s software-defined data centre strategy, constructing and guiding VMware’s cloud computing business and SaaS transformation efforts. Additionally, Raghu has helped in the company’s Merger and Acquisition strategy, driving key partnerships with Dell Technologies. What Raghu brings to the table In an interview with a US-based media house, Raghu said, “We are hardware agnostic, cloud-agnostic, and we can focus on helping our customers run their applications and run their IT however they want to run it. We are providing the tools to help them build new applications faster, run them across all these locations, manage these applications, secure and protect them — that’s what we are trying to do in a nutshell.” According to sources, VMware is already in talks with ecosystem vendors to form strategic partnerships that will assist the virtualisation company’s multi-cloud takeover. Post the spin-off, VMware will find it easier to get into strategic partnerships with vendors. “What you should look for in the coming weeks, months and years, is that our partnership with the broader ecosystem is as deep and as strong as it is with Dell today,” Raghu added. Raghu said: “VMware is uniquely poised to lead the multi-cloud computing era with an end-to-end software platform spanning clouds, the data center and the edge, helping to accelerate our customers’ digital transformations.” VMware competes with Nutanix in the hyperconverged infrastructure (HCI), virtualisation, IT automation and hybrid cloud management space. In a media statement, Raghu said what differentiates VMware from its competitor is its business initiative — to aid customers in their digital transformation journey. Raghu emphasised on how VMware’s offerings continue to remain unparalleled when it comes to helping clients moving to the cloud and securing their digital assets.","excerpt":"VMware recently named its long time executive Rangarajan (Raghu) Raghuram as the new Chief Executive Officer. He will replace Pat Gelsinger, who had moved to Intel. Further, Raghu will be a member of the Board of Directors of VMware, with effect from June 1. He is at present the COO of Products and Cloud Services, […]","categories":["IT Services"],"tags":["VMWare"],"author_name":"Debolina Biswas","publish_date":"2021-05-18T14:00:00","publication_year":"2021","word_count":751,"keywords":["VMWare","Go","programming_languages:R","cloud computing","AI","digital transformation","programming_languages:Go","Git","RAG","automation","R"],"extracted_tech_keywords":["AI","RAG","cloud computing","R","Go","Git","digital transformation","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/what-to-expect-from-vmwares-new-ceo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075210,"title":"Using ‘Cocktail Party Problem’ to Talk with Animals","content":"Animals communicating with each other might seem simplistic at first glance. Compared to human communication, animals do not appear to be using any particular language but merely noises to communicate with each other. Several noises that animals make are less of a conversation in the present, and more of a call for predicting natural changes such as rain, water, or signals for food some distance away. When it comes to artificial intelligence, plenty of progress has been made in the development of AGI using machine learning and neural networks on animals and through the understanding of animal behaviour. However, understanding the language of animals and communicating with them is one of the longest-running fields of study in technology and biological sciences alike. Recently, California-based organisation, Earth Species Project (ESP), introduced the Bioacoustic Cocktail Party Problem Network (BioCPPNet) that uses machine learning to decode non-human communication. The machine learning architecture is a modular, U-Net-based network that optimises bioacoustic source separation in diverse biological taxa. You can find the link to the code here. What is Bioacoustic source separation? Informally referred to as the “cocktail party problem”, Bioacoustic source separation encompasses the detecting, recognising, and extracting information problem from specific signals in the presence of noisy environments. While separating human speech is a well studied subject with the use of deep neural networks (DNNs), bioacoustic CPP in animal environments remains problematic due to an overlap of noises from different unidentifiable sources. BioCPPNet, the machine learning model, is a lightweight neural network that acts as an end-to-end source separation system and extracts information from raw waveforms obtained directly from recordings to identify and reconstruct sources. Extracting information from a herd of animals is a difficult task. For example, 58% of vocal recordings of an African elephant consisted of two concurrent signals which were hard to separate. Aza Raskin, founder of ESP, said that the idea of BioCPPNet originates from the recent advancement in machine learning models that has made it possible to translate between distant human languages in real-time without any prior knowledge requirement. https:\/\/twitter.com\/earthspecies\/status\/1496941682155081729 How does it work? Recently, Elodie Briefer, associate professor at University of Copenhagen, has developed a pig-grunt analysing algorithm that helps in assessing positive or negative emotions in pigs. Though a great development, the algorithm only worked on pigs and failed to analyse other animals like dolphins, primates, or bees’ communication. Raskin says that their network aims to understand the entire biodiverse ecosystem’s communication. The model was tested on macaques, Egyptian fruit bats, and bottlenose dolphins. The supervised model performed two tasks—to group sequences of input signals from different sources and to integrate simultaneous harmonic or quasi-harmonic sounds by a given signaller. ESP used a CNN-based classifier model in BioCPPNet to label the individual identity of vocalised signals. Using the data already available from previous studies, the aim was to apply a self-supervised machine learning algorithm to relate the physical animal behaviour and actions with the audio data to verify if they could be tied together. The algorithm worked best in a closed speaker regime with testing subjects drawn from the same distribution of training subsets. In case of bottlenose dolphins and macaques, the system struggled in an open speaker regime with the large testing data being different from smaller training data. However, in case of bats, the model yielded comparable results in both open and closed regimes suggesting the need for larger datasets. Can we talk to animals? Since the model has to be implemented on larger datasets of the animal environment, Raskin says that the method could benefit by reducing the supervised training scheme. This poses a limitation since the models worked best with larger training data. Raskin points to ongoing studies that are applying CNN and developing a self-supervised machine learning algorithm, without the requirement of human experts to label and input data. Christian Rutz, professor of biology at University of St Andrews said that Hawaiin crows, the species that makes and uses tools for foraging, are believed to have a more complex set of vocalisations than other crow species. Another study by Ari Friedlaender of University of California, uses data from sound recorders placed inside the ocean to observe behavioural patterns of marine animals. Robert Seyfarth, professor of Psychology at University of Pennsylvania, points out the problem of inferring meaning from animal sounds. He argues that the same sound can have different meanings in different contexts when it comes to animals. “Applying AI analyses to human language, with which we are so intimately familiar, is one thing, but it can be different and difficult doing it to other species,” said Seyfarth. Raskin acknowledges the concern and says that AI alone cannot unlock communication with other species but researches have showcased how complex animal languages are than merely noises and actions. This research opens the gate to previously unusable large datasets of overlapping signals and enables researchers to implement ML-based models to design management and conservation strategies for animal species.","excerpt":"Understanding the language of animals and communicating with them is one of the longest-running fields of study in technology and biological sciences alike.","categories":["AI Features"],"tags":["AGI","AI (Artificial Intelligence)","Deep Learning","Machine Learning","Neural Network"],"author_name":"Mohit Pandey","publish_date":"2022-09-14T17:00:00","publication_year":"2022","word_count":825,"keywords":["Neural Network","Go","artificial intelligence","machine learning","AI","neural network","ML","Machine Learning","RAG","BERT","Aim","AGI","Deep Learning","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","Aim","RAG","R","Go","BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/using-cocktail-party-problem-to-talk-with-animals\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10066497,"title":"Detecting deepfakes using novel self-blended images approach","content":"Deepfake technology is gaining prominence in the criminal underworld, Europol warned. In the UAE, deepfake technology was used to steal USD 35 billion by cloning a company director’s voice. According to a Sensity report published in 2021, non-consensual and harmful deepfake videos double every six months. As of December 2020, nearly 85,047 deepfake videos were detected. “We have reached a point where there is hardly any distinction between reality and augmentation, making it high time to take the necessary steps that ensure no harm is done via augmented reality in the coming years,” Aishwarya Srinivasan, an AI & ML Innovation Leader at IBM, said. Therefore, it is becoming more vital by the day to have reliable techniques for deepfake detection. A research team from the University of Tokyo is doing exactly that. The team has developed a method to detect deepfakes using Self-Blended Images (SBIs). This unique synthetic training data methodology has outperformed state-of-the-art techniques on unseen manipulations, the research paper said. Through extensive experiments, the team has learned that this method improves the model generalisation to unknown manipulations, particularly on DFDC and DFDCP. This is because existing methods suffer from the domain gap between the training and test sets. The novel approach has outperformed the baseline by 4.90 per cent and 11.78 per cent points in the cross-dataset evaluation. Most of the existing methods we know perform well in detecting known manipulations. However, when it comes to unknown manipulations, some methods do tend to be ineffective. So, how do we deal with this? One of the most effective approaches is training models with synthetic data. It encourages models to learn generic features for face forgery detection. Method The study’s primary purpose is to detect statistical inconsistencies or anomalies in the deepfakes between altered faces and background images. To train their detectors, the team generated fake synthetic samples that consist of common forgery traces that are difficult to recognise. These samples will be further used to train more reliable detectors. Deepfake generation techniques will continue to improve; hence, GAN-synthesised source images will be even closer to pristine target images in their properties, for example, facial landmarks and pixel statistics. Research paper: Samples of pristine images (top row) and their SBIs (bottom row) It involves the following three steps: (1) A source-target generator generates false source and target images, which will be used later for blending. (2) A mask generator then creates a grey-scale mask image with some deformations. (3) Lastly, the source and target images get blended with the mask to obtain an SBI. Research paper: Overview of generating an SBI In their cross-dataset evaluation, the team has trained their model on FF++ and evaluated it on CDF, DFD, DFDC, DFDCP, and FFIW with standard protocols. The purpose of this is to create a situation where the detectors are exposed to unseen domains. “Our approach surpasses or is at least comparable to the state-of-the-art methods on all test sets despite its simplicity,” the paper said. The model has also been tested against discriminative attention models (DAM) and fully temporal convolution networks (FTCN). Limitations Even though the results in cross-dataset and cross-manipulation evaluations are expected to be beneficial, there are some limitations associated. Firstly, this model is incapable of capturing temporal inconsistencies across video frames. This means some sophisticated deepfake generation techniques with fewer spatial artefacts may not get detected. Further, this model is not compatible with whole-image synthesis. This is because it defines a fake image as an image where the face region or background is manipulated. The model has been evaluated using a 20k image set sampled from the FFHQ dataset and StyleGAN synthesis, and its AUC is only 69.11 per cent, the report said. Research Paper: Typical artefacts on forged faces Detection of deepfakes more crucial by the day As AI keeps making groundbreaking advancements, the potential for criminal exploitation also increases. Currently, deepfakes are mostly being used for entertainment purposes, but there is a darker side to this evolving technology. According to a report, the cost associated with deepfake scams exceeded USD 250 million in 2020. Deepfake tech has the potential to be a great threat in the political sphere and cybersecurity. A lot of emphasis is hence being given to the development of its detection methodologies. Delving into the identification of deepfakes, Srinivasan said, “Considering the advanced technology needed to even classify them as fakes, laymen, if caught up in deep fake treachery, would not even be able to prove themselves easily.” Last year, the Chinese government lost as much as USD 76 million to criminals who manipulated personal data and fed the facial recognition systems with deepfake videos. This was just one example of how a biometric hack could lead to catastrophic results.","excerpt":"This methodology outperformed the baseline by 4.90% and 11.78% points in the cross-dataset evaluation.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Pritam Bordoloi","publish_date":"2022-05-07T16:00:00","publication_year":"2022","word_count":788,"keywords":["Go","synthetic data","AI","RPA","ML","innovation","RAG","ViT","GAN","AI Tool","R"],"extracted_tech_keywords":["AI","ML","RAG","R","Go","GAN","ViT","synthetic data","RPA","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/detecting-deepfakes-using-novel-self-blended-images-approach\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":8408,"title":"Jigsaw Academy, the Bangalore-based online school of analytics receives 20 crore INR in funding from Manipal Global Education Services (MaGE)","content":"Bangalore, 9th December 2015: Jigsaw Academy, the Online School of Analytics, receives 20 crore INR in funding from Manipal Global Education Services (MaGE).  The company would be using these funds from MaGE to expand its offerings and to penetrate new markets. InteQuant Advisors, a leading advisory firm providing professional services in the field of fund raising and transaction support services were the sole financial advisors who facilitated this transaction. Excited about the development, co-founders Gaurav Vohra, CEO, Jigsaw Academy and Sarita Digumarti, COO, Jigsaw Academy said that this association with MaGE will greatly help improve their global footprint and help them successfully penetrate new markets. Speaking on this occasion Gaurav Vohra, CEO, Jigsaw Academy said, “We are very happy and excited about this strategic partnership – this is a huge milestone for Jigsaw Academy. We could not have found a better partner than the Manipal group in the education space.  We believe there is great potential in the analytics and big data training market in India and abroad. This funding that we have received will help us take bigger steps in the global market more aggressively while opening up more corporate training opportunities” Jigsaw Academy was founded in 2011 with the vision of providing best in class analytics and big data skills to students and professionals worldwide. There is a huge global shortfall of professionals with analytics and big data skills.  Jigsaw Academy has trained over 40,000 students from over 30 countries and more than 100 corporates and IT giants across different verticals in the last five years, using its unique and proprietary learning platform. Jigsaw Academy has the largest library of digital content in analytics and big data, and offers courses across industries and functions, like Retail, E-Commerce, Finance, HR, and Supply Chain. Mr. S. Vaitheeswaran, MD and CEO of Manipal Global Education Services, said “We are excited about our investment in Jigsaw Academy. We believe the company has established itself as a market leader in the analytics and big data space, and the investment will help them achieve rapid scale. At MaGE, professional certification is the next big foray for us and this investment is a significant step in that journey” Also speaking on this occasion, T.V. Mohandas Pai, who is joining Jigsaw Academy’s Advisory Board said, “Trained professionals who understand the right approach to applying analytics will drive new value propositions across industries and companies of all sizes. When empowered with training in the right data-driven frameworks and methodologies, they will show that big data is more than just a buzzword – it is a truly transformative approach to growing businesses and improving decision-making. The demand for these skills is increasing tremendously across all ventures and enterprises, and Jigsaw Academy is perfectly poised to meet this demand with its innovative individual and corporate training offerings” McKinsey Global Institute projects that by 2018 demand for data scientists may be as much as 60 percent greater than the supply. Analytics and Big Data training is receiving a push in the recent years and educational giants are willing to offer industry relevant courses in association with institutes like Jigsaw Academy to bridge the demand and supply gap. About Jigsaw Academy Jigsaw Academy, the online school of analytics, has trained over 40,000 students across 30+ countries in the most widely used industry-relevant data analytics tools and techniques. Jigsaw founders have over 25 years of combined experience in consulting and analytics across multiple industry verticals in India and the United States. Jigsaw Academy’s industry-relevant curriculum has led to innovative partnerships with leading academic institutions such as SDA Bocconi in Milan, Italy, Great Lakes Institute of Management in Chennai, and the Indian Institute of Management in Bangalore. Jigsaw Academy’s award winning courses are well recognized in the industry, and Jigsaw is the training partner of choice in analytics and big data for many large global consulting and IT services companies.","excerpt":"Bangalore, 9th December 2015: Jigsaw Academy, the Online School of Analytics, receives 20 crore INR in funding from Manipal Global Education Services (MaGE).  The company would be using these funds from MaGE to expand its offerings and to penetrate new markets. InteQuant Advisors, a leading advisory firm providing professional services in the field of fund raising and transaction support services were […]","categories":["AI News"],"tags":["online education"],"author_name":"AIM Media House","publish_date":"2015-12-09T10:40:22","publication_year":"2015","word_count":647,"keywords":["big data","API","funding","programming_languages:R","AI","data-driven","Git","Ray","online education","analytics","R"],"extracted_tech_keywords":["AI","analytics","Ray","R","Git","API","big data","data-driven","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jigsaw-academy-the-bangalore-based-online-school-of-analytics-receives-20-crore-inr-in-funding-from-manipal-global-education-services-mage\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":66864,"title":"New Wipro IBM Novus Lounge In Bengaluru Will Boost AI\/ML","content":"Wipro announced today a collaboration with IBM to assist Wipro customers embark on a seamless and secure hybrid cloud journey. Through this alliance Wipro will develop hybrid cloud offerings to help businesses migrate, manage and transform mission-critical workloads and applications, with security across public or private cloud and on-premises IT environments. The launch of Wipro IBM Novus Lounge in Bengaluru will foster innovation and build industry solutions leveraging Cloud, Artificial Intelligence, Machine Learning and Internet of Things. The Novus Lounge is located at Wipro’s Kodathi campus in Bengaluru is a dedicated innovation centre, and will offer a comprehensive suite of solutions leveraging Cloud, Artificial Intelligence, Machine Learning and Internet of Things capabilities to foster innovation for enterprises, developers and start-ups. Customers will have remote access to IBM and Red Hat solutions, designed to help them scale their technology investments for improved experience and business agility with connected insights. Ramesh Nagarajan, Senior Vice President – Cloud Services, Wipro Limited said, “Wipro empowers customers across industries to re-imagine their cloud journey with its business-first strategy and industrialized solutions approach. Wipro IBM Novus Lounge will allow us to showcase hybrid multi-cloud and open source solutions even more comprehensively and support our customers’ continuous business transformation journey.” Additionally, Wipro will leverage IBM Cloud offerings and technologies alongside in-house services to develop industry solutions for clients in Banking and Financial Services, Energy and Utilities, Retail, Manufacturing and Healthcare space. Gaurav Sharma, Vice President – Cloud and Cognitive Software, IBM India said, “As companies across the world continue to drive digital transformation, decision-makers must rethink radically on how to leverage the combined power of data, cloud and open source technologies to become industry leaders. Wipro IBM Novus Lounge brings together Wipro’s expertise across industries and IBM’s open source technologies, designed to be secure and scalable across hybrid cloud, Data and AI, all running on Red Hat OpenShift promoting the journey to Cloud and journey to AI.”","excerpt":"Wipro announced today a collaboration with IBM to assist Wipro customers embark on a seamless and secure hybrid cloud journey. Through this alliance Wipro will develop hybrid cloud offerings to help businesses migrate, manage and transform mission-critical workloads and applications, with security across public or private cloud and on-premises IT environments. The launch of Wipro […]","categories":["AI News"],"tags":["Wipro"],"author_name":"Vishal Chawla","publish_date":"2020-06-08T12:24:18","publication_year":"2020","word_count":320,"keywords":["Wipro","machine learning","artificial intelligence","AI","innovation","ML","digital transformation","Scala","Git","RAG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","R","Scala","Git","digital transformation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/new-wipro-ibm-novus-lounge-in-bengaluru-will-boost-ai-ml-and-cloud-innovation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10137880,"title":"Tata Communications Joins Forces with Palo Alto Networks to Enhance Enterprise Cyber Resilience","content":"Tata Communications today announced a collaboration with Palo Alto Networks, the global cybersecurity leader, to deliver comprehensive cybersecurity solutions to global enterprises. The partnership will bring together Palo Alto Networks industry-leading technologies with Tata Communications digital fabric of solutions and deep industry expertise across network security, cloud security, cyberthreat detection and response, security assessment and consulting services — leveraging a powerful alliance to address the evolving cyberthreat landscape. Given the rapid shift towards digital transformation, cloud adoption, and remote workforces has significantly expanded and complicated the attack surface for organisations, making them more vulnerable to sophisticated cyberthreats. To combat these challenges, enterprises require robust, integrated, and managed cybersecurity solutions. “It’s critical for businesses to adopt a platform-centric approach to cybersecurity as attack surfaces expand and threats become more complex,” said Vaibhav Dutta, Associate Vice President and Global Head Cybersecurity Products & Services at Tata Communications. “Our strategic collaboration with Palo Alto Networks stitches together all the essential solutions and tools into a single cloud and cybersecurity fabric –simplifying and streamlining enterprise security management.” As businesses shift to cloud-based solutions, new security challenges such as unauthorised access and lack of visibility into cloud environments are bound to arise. Against this backdrop, Palo Alto Networks’ Precision AI-powered platform addresses these issues by combining the capabilities of SIEM, XDR, SOAR, and other SOC tools to streamline security operations, outperforming traditional Security Operations Centre (SOC) tools. Meanwhile, Tata Communications will provide a unified, cloud-delivered security solution using the Palo Alto Networks Prisma Access platform, which integrates network security, cloud security, Zero Trust Network Access (ZTNA) 2.0, and Secure Access Service Edge (SASE) capabilities. This solution offers proactive threat isolation and resolution with embedded threat intelligence, ensuring top-tier security at the edge for end-users. Tata Communications will also deliver dedicated managed services, including lifecycle management of Palo Alto Networks solutions, allowing customers to focus on their core business. Moreover, the partnership aims to optimise security posture, enhance user experience, and reduce response times through simplified security management. Customers can also achieve significant cost savings by consolidating multiple security vendors into a single provider. Additionally, the company will offer unmatched expertise through a thorough assessment of existing security infrastructures, providing a roadmap with real-time analytics, threat intelligence, and a use case library. “Tata Communications’ proven capabilities as a Managed Cybersecurity Service Provider and commitment to excellence, perfectly complement our advanced technologies,” said Michelle Saw, VP GTM and Ecosystems, JAPAC at Palo Alto Networks. “Together, we are confident of empowering future forward enterprises worldwide to further strengthen their security posture, improve operational efficiency, and mitigate risks effectively.”","excerpt":"Tata Communications today announced a collaboration with Palo Alto Networks, the global cybersecurity leader, to deliver comprehensive cybersecurity solutions to global enterprises.","categories":["AI News"],"tags":["Cyber Security","tata communications"],"author_name":"Shalini Mondal","publish_date":"2024-10-08T18:05:08","publication_year":"2024","word_count":430,"keywords":["API","Cyber Security","AI","tata communications","ML","Git","RAG","Aim","analytics","real-time analytics","Rust","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","R","Rust","Git","API","real-time analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tata-communications-joins-forces-with-palo-alto-networks-to-enhance-enterprise-cyber-resilience\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61233,"title":"How Effective Data Visualizations Enable Us To Understand Covid-19 Pandemic Better","content":"Amid a barrage of information on Covid-19, processing the relevant material can be challenging. While some have completely disconnected themselves from the news cycle, others are regularly tracking their news feed to better understand how to combat Covid-19. This necessitates that they not only get access to verified information of these developments, but also pick up valuable insights from the massive amount of data that has been generated in the last few months across the world. In this setting, data visualization tools have allowed people to understand and absorb information in a quick and accurate manner. Softwares like Tableau and Microsoft Power BI have helped synthesize these complex networks of events by visualizing data, detecting hidden patterns and turning them into illustrated stories. Without relying on people to sift through quantitative data or numbers on a spreadsheet, this allows them to obtain information relevant to their situation and thereby, react better. This could be achieved by shaping the timeline of the pandemic, documenting the spread of the virus in a particular area, and identifying hotspots, among others. What is more, these diverse sets of visualizations have also helped political leaders and community members strategize more effectively, thereby, having quite a real-world impact in the face of this crisis. Importance Of Visualization Amid Massive Data As cases of Covid-19 proliferate across the globe, so has data associated with it. This includes information on the number of affected patients, the number of people they are likely to infect, equipment that is available for healthcare workers, as well as the death and recovery rates, among others. This data needs to be efficiently conveyed to people, since in the event of a pandemic like this, intuition cannot substitute for facts to understand how the spread is advancing. The approach that is needed should involve analysing, sharing and leveraging data. Here, visualizing the data can help explain the developing events in a clear and succinct way for people to interpret data well, tease out patterns, and pick up on trends. The Coronavirus Resource Center dashboard from Johns Hopkins University has emerged as a reliable source that gives both a micro and macro view on the pandemic. Incorporating data from various organizations including WHO, the platform is updated daily with rich infographics on the Covid-19 outbreak. (Image via Johns Hopkins University on April 7, 2020) ALSO READ: Tableau Makes Johns Hopkins Covid-19 Data Available In Its Data Resource Hub Adds More Context The above image is likely to get the attention of Indian readers, especially those who have been tracking the rise of Covid-19 cases in India. And this may be applicable for readers anywhere across the globe. While they may dismiss stories developing at another corner of the world, they would tend to react to data associated with their own circumstances. While this image only captures the cases of infected people and related trends, infographics on the number of hospitals in your area, its capacity to treat Covid-19 patients, the rate at which cases are being detected and the number of people requiring hospitalization will be information that you will be interested in knowing. And so should be, given that this kind of dissemination of information has helped contain the outbreak. It has yielded unprecedented behavioral changes, one example being the concept of ‘flattening the curve’, successfully illustrated in various data visualization charts. This has motivated the public to practice social distancing, and thereby, actively participate in combating Covid-19. Done correctly, like Washington Post has is this interactive piece, it engages readers, is persuasive, and increases empathy, opening up the public to change. (Image via World Economic Forum) ALSO READ: Are Too Many Data Scientists Trying to Predict Covid-19 Outcomes in Futility? Helps Refine Strategies With countries across the world putting travel restrictions in place – and some, like India, enforcing nation-wide lockdowns – visual representations of massive data has helped inform these strategies. For example, Colorado in the US has taken various proactive measures to use Tableau to visualize the outbreak’s progression. The website is updated everyday at 4pm, giving epidemiologists enough time to review the data and improve its accuracy. It comprises a case summary of the entire state, including the number of cases, those hospitalized, number of people tested, and the number of deaths recorded. Even the government of India has been representing key data points visually. An interactive depiction of this has been created by a student of IIT-Bombay as well, as shown below. However, it is important to note that although such representations relay information in a precise and pithy manner, they may not always convey the entire story and hence, need to be understood in conjunction with associated data. This is because it is difficult to represent time visually, and given that this is a pandemic, it plays a critical role when charting strategies to combat it. Also, some factors are difficult to measure, including uncertainty, which also has a big part to play in this. Outlook Even as the value of data visualization amid Covid-19 has been discussed, it also demands that people at the helm of this data use it responsibly, given that their interpretation of information on a critical subject like the COVID-19 outbreak is likely to influence how people see, understand, and react to it. They should recreate it in such a way that it relays the necessary information, and reflects reality in the larger context – data that can inform decisions and drive change.","excerpt":"Amid a barrage of information on Covid-19, processing the relevant material can be challenging. While some have completely disconnected themselves from the news cycle, others are regularly tracking their news feed to better understand how to combat Covid-19. This necessitates that they not only get access to verified information of these developments, but also pick […]","categories":["Deep Tech"],"tags":["big data and analytics everyday life","Coronavirus","covid-19","covid-19 data visualization","covid19 data","data visualization","trends in bi and data visualization"],"author_name":"Anu Thomas","publish_date":"2020-04-09T11:00:12","publication_year":"2020","word_count":910,"keywords":["trends in bi and data visualization","Go","API","covid-19","AI","programming_languages:R","covid-19 data visualization","programming_languages:Go","covid19 data","Coronavirus","RAG","data visualization","GAN","big data and analytics everyday life","R"],"extracted_tech_keywords":["AI","RAG","R","Go","API","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-data-visualizations-enable-us-to-understand-covid-19-better\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10055121,"title":"5 Best IDEs For Data Scientists","content":"An IDE (Integrated Development Environment) is a software application that provides comprehensive facilities to computer programmers for software development. The easy debugging process, syntax highlighting, tool integration, keyboard shortcuts, and parsing available on IDEs make them an optimal coding tool for data scientists. This article highlights the top five IDEs for data scientists. Spyder Scientific Python Development Environment (Spyder) is an open-source, cross-platform IDE for Data Science. The IDEs essential building blocks, include advanced editing, code analytical tools, IPython Console, variable explorer, plots, debugger and the help icon, which makes Spyder an ideal choice for data scientists. To install it, one must have Anaconda Environment in their system. The IDE integrates important libraries for data science- NumPy, SciPy, Matplotlib and IPython and can be extended to plugins- Spyder Notebook, Spyder Terminal, Spyder Unittest. According to data scientists, Spyder is very intuitive for scientific computing. JupyterLabs JupyterLab is an open-source web application, which has been designed to provide a user interface based on Jupyter Notebook. It allows the user to work with documents on Jupyter Notebook, born out of IPython in 2014. Its flexible interface lets users configure and arrange workflows in data science, scientific computing, computational journalism, and machine learning. A modular design allows for extensions that expand and enrich functionality. The application is easy to use, has an interactive data science interface, and is user-friendly for presentation or educational tools. PyCharm PyCharm is an IDE for professional developers and data scientists. It has intelligent coding assistance that allows for smart code completion, code inspections, on-the-fly error highlighting and quick fixes, along with automated code refactorings and rich navigation capabilities. PyCharm has many tools: An integrated debugger and test runner A built-in terminal A Python profiler remote development capabilities with remote interpreters Integration with major VCS and built-in database tools Remote integration with Docker and Vagrant. In addition, it also has an integrated library with tools such as NumPy and MatplotLib. Visual Studio Code VS Code is one of the most used Python IDEs. The IDE is known for its tools such as IntelliSense that allows features beyond syntax highlighting and smart completions based on variable types, imported modules, and functions definition. In addition, VS Code allows debugging code right from the editor with breakpoints, call stacks and an interactive console. Furthermore, VS Code is extensible and customisable, allowing for the addition of new languages, themes, and debuggers. The IDE also has built-in Git commands. VS Code is available in free and paid versions. Atom Atom is a formidable IDE for ML & DS professionals that supports many languages other than Python, such as C, C++, HTML, JavaScript, etc. The IDE includes features such as cross-platform editing, built-in package manager, smart autocompletion, file system browser, and multiple panes. Moreover, its plugins, languages, libraries, and tools are constantly updated, resulting in the Atom interface and experience being customisable and outstanding. What do Experts suggest? According to Chief Data Scientist at PayU Finance, Piyush Gupta, “At PayU, the developers tend to choose the platform of their own choice for development. However, the majority of them use a combination of Jupyter Notebooks and PyCharm. Jupyter is great for initial EDA and provides flexibility for a lot of basic tasks. Personally, I prefer to use PyCharm because of better environment management, more accurate refactoring, better package management, a dedicated python console, better navigation & UX, and advanced debugging capabilities.”","excerpt":"In this article, we pick out the top 5 IDEs useful for data scientists","categories":["AI Trends"],"tags":["coding","Data Science","Pycharm","Pycharm IDE","spyder","Visual Studio Code","VS Code"],"author_name":"Abhishree Choudhary","publish_date":"2021-12-09T12:00:00","publication_year":"2021","word_count":565,"keywords":["data science","NumPy","machine learning","Pycharm IDE","AI","coding","spyder","ML","docker","Python","Pycharm","Matplotlib","VS Code","Data Science","Jupyter","R","Visual Studio Code"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Jupyter","NumPy","Matplotlib","docker","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-ides-for-data-scientists\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10012447,"title":"Top Multi-Cloud Solutions That Are Ruling The Market","content":"According to a new app modernisation survey from Enterprise Strategy Group, “92% of organisations feel it is important to utilise multi-cloud enabled container management and orchestration solutions. According to Netapp, multi-cloud refers to the distribution of cloud assets, software, and applications across several cloud environments, using multiple cloud computing platforms to support a single application or ecosystem of applications that work together in a common architecture. Public cloud providers offer virtual compute, storage, and many services but they still differ from each other as they operate in different geographic regions and offer different security, data sovereignty, and hybrid capabilities than others. Even the user experience, including developer tools, web portals, automation capabilities are not uniform. Using multiple clouds is not straightforward. An app designed to run great on one cloud won’t easily run on another. According to Google Cloud, developers may write software that’s completely cloud-agnostic and can run anywhere. But one can run into problems like: App dependency on unique capabilities for AI, data processing, IoT, or vertical-specific APIs — think media or healthcare. Hosting applications in a specific geography, and thus choose a specific cloud, datacenter, or partner facility. Need for a specific data source so that the host is closest. There have been quite a few multi-cloud solutions developed even by the public cloud providers. In the next section, we take a look at a few of those. (The list below is in no particular order and has been compiled considering the reach of the service providers and popularity.) Anthos (Source: Google Cloud) One of the most popular Google Cloud Services in the multi-cloud segment is Anthos. Introduced in 2019, Anthos lets users manage workloads running on third-party clouds like AWS and Azure, giving the freedom to deploy, run and manage applications on the cloud of choice, without requiring administrators and developers to learn different environments and APIs. Anthos leverages open APIs, giving users the freedom to modernise any place, any time and at their own pace. Because Anthos is based on GKE, the managed Kubernetes service, users can automatically get the latest feature updates and security patches. Azure Arc Microsoft’s Azure Arc does two key things — bringing Azure management capabilities to any infrastructure and enabling Azure services to run anywhere. Since its launch, Azure Arc has seen tremendous customer interest and adoption across all industries. Azure SQL Managed Instance, and Azure PostgreSQL Hyperscale can run across on-premises data centres, multi-cloud, and the edge. vRealize VMware’s vRealize provides app-centric security and network visibility across hybrid and multi-cloud environments. It offers visibility and security planning across clouds that users can consume as a service or deploy on-premises. The tool easily maintains control over users’ infrastructure with custom policies and automated workflows that are uniquely based on how the users want to run their multi-cloud environment. Netapp NetApp’s cloud solutions provide a full range of hybrid cloud data services that simplify management of applications and data across cloud and on-premises environments to accelerate digital transformation. It claims to offer a hybrid multi-cloud experience that breaks down technological and organisational silos in order to deliver better customer value as part of ongoing enterprise digital transformation. IBM Multi-Cloud Management Platform IBM’s MCMP platform supports multiple technology stacks across a multi-vendor platform by — optimising cloud spend and usage; managing services mapping and dependencies; and extending DevOps processes for traditional IT and cloud natives. IBM’s MCMP enables organisations to analyse health and inventory of their multi-cloud systems. Morpheus A leader in the Gartner 2020 Magic Quadrant for Cloud Management Platforms (CMP), Morpheus claims to have the highest capability scores for provisioning, brokerage, and governance as well as the best customer feedback in the field. Morpheus is an application-centric automation and orchestration framework which means it can present a self-service catalogue to provision applications made up of bare-metal servers, VMs, containers, and public cloud services. Scalr Scalr was founded in 2007, with the challenge of federating dispersed IT teams over common cost and security standards while preserving local autonomy. Scalr Organisational Model combines proactive and reactive policies with a hierarchy that maps to the organisation’s structure. Scalr’s platform enables enterprises to achieve cost-effective, automated and standardised application deployments across multi-cloud environments. Scalr’s customers include Samsung and NASA. Snow Embotics The Snow Commander Cloud Management Platform is built from the ground up on a common architecture to deliver an all-in-one solution that supports multi-hypervisor and multi-cloud environments. Snow Software was also named a Leader in the Gartner Magic Quadrant for Cloud Management Platforms 2020 for the second year in a row. JUKE Juniper’s JUKE created a persistent and distributed layer of storage and network for containers, schedulers and distributed applications across clouds. JUKE eliminates the need to create and manage multiple and siloed clusters across each individual cloud provider, drastically reducing management time, complexity and enabling agility and ownership of your data, where applications and containers can be deployed, moved and distributed based on your requirements, and not on cloud providers\/schedulers limitations. If you think we have missed out any multi-cloud tools, please add in the comments.","excerpt":"According to a new app modernisation survey from Enterprise Strategy Group, “92% of organisations feel it is important to utilise multi-cloud enabled container management and orchestration solutions. According to Netapp, multi-cloud refers to the distribution of cloud assets, software, and applications across several cloud environments, using multiple cloud computing platforms to support a single application […]","categories":["AI Trends"],"tags":["latest technology in cloud computing","multi cloud platform Anthos","netapp"],"author_name":"Ram Sagar","publish_date":"2020-11-26T11:00:12","publication_year":"2020","word_count":844,"keywords":["PostgreSQL","latest technology in cloud computing","AWS","AI","cloud computing","R","netapp","RAG","Aim","multi cloud platform Anthos","SQL","Azure","kubernetes"],"extracted_tech_keywords":["AI","Aim","RAG","cloud computing","AWS","Azure","kubernetes","PostgreSQL","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-multi-cloud-tools-anthos-aws-azure\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10079953,"title":"Tech Mahindra, T-Systems Plan to Hire 6,000 People in Nagpur","content":"After Pune and Bengaluru, T-Systems, one of the leading European IT service providers, plans to add Nagpur to its list of delivery centres. These centres are aimed at providing intelligent digital solutions and efficient cloud services to T-Systems’ global customers. The new centre is established in partnership with Tech Mahindra, and will open employment opportunities for 6,000 people across various departments including sales, marketing, technical roles, and others, over the period of 24 months. To meet the needs, Tech Mahindra will enable the building up of a dedicated workforce delivering cutting-edge technological solutions in digital and cloud spaces. T-Systems attributes Tech Mahindra’s access to market, talent from Tier II cities, and competitive pricing as primary reasons for this partnership. CP Gurnani, CEO of Tech Mahindra, said that Nagpur can be the next digital hub for leading innovation because of several factors, like ease of doing business, availability of manpower, and accessibility to infrastructure. He also added that its partnership with T-Systems will enable employment opportunities for local talent to work and collaborate with a global team. The move further supports the notion that post-pandemic, IT companies are not just restricting themselves to Tier-I cities anymore, but are expanding their presence and workforce to Tier-II and Tier-III cities. With growing attrition, IT companies are looking to harness the entry-level talent pool available in smaller cities.","excerpt":"T-Systems will enable employment opportunities for local talent to work and collaborate with a global team","categories":["AI News"],"tags":["Cloud services","Hiring","Tech Mahindra"],"author_name":"Ayush Jain","publish_date":"2022-11-17T10:45:52","publication_year":"2022","word_count":224,"keywords":["Tech Mahindra","programming_languages:R","AI","innovation","Cloud services","Hiring","Git","Aim","R"],"extracted_tech_keywords":["AI","Aim","R","Git","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tech-mahindra-t-systems-plan-to-hire-6000-people-in-nagpur\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10009506,"title":"Complete Guide To FastAPI With Machine Learning Deployment","content":"Any machine learning model’s end goal is a deployment for production purposes. Building a REST API(Application Programming Interface) is the best possible way to evaluate model performance. In python, Django and more evidently Flask frameworks are used for this purpose. For machine learning, Flask is preferred more than Django. Here comes FastAPI which is faster than Flask, providing higher performance boost, easier to code, comes with automatic documentation, provides data validation on input data provided to the server, and many more such amazing features which I’ll be covering in this article. FastAPI is one of the fastest python frameworks available. The automatic documentation comes with free of cost. Since it is based on python so it provides nice python type hints for autocompletion and type checks. FastAPI is built upon two major python libraries – Starlette(for web handling) and Pydantic(for data handling). FastAPI is based on some standard integrations — OpenAPI, JSON Schema, OAuth2. In this article, we will discuss the implementation of FastAPI and demonstrate how it can be used for deployment purposes. Getting Started With FastAPI Software Dependencies that are required for FastAPI:- Python 3.6+ FastAPI ASGI server for production such as Uvicorn or Hypercorn. Pydantic Basic Hello World First of all, we will do a very basic implementation to get familiar with FastAPI. from fastapi import FastAPI app = FastAPI() @app.get(\"\/\") def index(): return {\"message\": \"Hello World\"} Save this program in main.py then from a terminal, execute: $uvicorn main:app --reload The following message should be printed: INFO:     Uvicorn running on http:\/\/127.0.0.1:8000 (Press CTRL+C to quit)INFO:     Started reloader process [28720]INFO:     Started server process [28722]INFO:     Waiting for application startup.INFO:     Application startup complete. After redirecting to the webpage, it should appear as: { “message”  :  “Hello World” } Interactive Documentation http:\/\/127.0.0.1.8000\/docs will show the interactive documentation by Swagger UI. Alternative documentation by redoc FastAPI for ML models Let’s take a pre-trained NLP model to extract the entities present in an English sentence. For this I’ll be using the Spacy library’s model ‘en_core_web_sm’  which is a pre-trained statistical model for English and can be installed with the command: python -m spacy download en_core_web_sm Then use it in the following code: from typing import List import spacy from pydantic import BaseModel from fastapi import FastAPI app = FastAPI() nlp = spacy.load(\"en_core_web_sm\") @app.get(\"\/\") def read_root(): return {\"Hello\": \"World\"} class Item(BaseModel): content: str comments: List[str] = [] @app.post(\"\/item\/\") def post_item(item: Item): doc = nlp(item.content) ents = [] for ent in doc.ents: ents.append({\"text\":ent.text,\"label\":ent.label_}) return {\"message\":item.content,\"comments\":item.comments, \"ents\":ents} This code takes an English statement as input(‘item’) in realtime and extracts entities using the spacy model to show which part of the sentence falls under which class. Here class Item inherits the BaseModel from pydantic for its properties to be used. Execute the code similarly as done above which will redirect to http:\/\/127.0.0.1.8000\/docs and the following page will appear. First, click onto ‘post’ as a post request is being made then click ‘Try it out’ to make changes to content and comments. Edit string and add text as per your wish. I’ve put in place of content string ‘Apple buys U.K. based startup for $1 billion’. And for comments string – ‘Nice as expected’. If the inputs are not as per the requirements then validation error will show which part went wrong. Click on ‘Execute’ to see the output Note that both content messages and comments as per user input are printed. { \"message\": \"Apple buys U.K. based startup for $1 billion \", \"comments\": [ \"Nice as expected\" ], \"ents\": [ { \"text\": \"Apple\", \"label\": \"ORG\" }, { \"text\": \"U.K.\", \"label\": \"GPE\" }, { \"text\": \"$1 billion\", \"label\": \"MONEY\" } ] } The entities have worked well depicting Apple as an organisation, the U.K as a geopolitical entity and $1 billion as Money. More experimentation could be done with the code such as adding multiple contents, adding comments and markdowns. Conclusion FastAPI can handle 9000 requests at a time. FastAPI can manage database sessions, web sockets, easy GraphQL injection and many more are still being built. FastAPI is being used by tech giants such as Netflix, Facebook, Microsoft, Uber. Making production-ready RestAPIs with few lines of code is apprehensive. The complete code of this implementation is available on the AIM’s GitHub repository. Please visit this link for that code.","excerpt":"In this article, we will discuss the implementation of FastAPI and demonstrate how it can be used for deployment purposes.","categories":["Deep Tech"],"tags":["Active Learning","machine learning deployment","machine learning software","NLP","Python"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-10-12T13:00:28","publication_year":"2020","word_count":711,"keywords":["machine learning","TPU","AI","ML","machine learning deployment","Active Learning","Python","NLP","Aim","machine learning software","FastAPI","spaCy","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","Aim","FastAPI","spaCy","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/complete-hands-on-guide-to-fastapi-with-machine-learning-deployment\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162530,"title":"Epsilon Appoints Pratik Nath as MD of India GCC","content":"Epsilon, a global data, technology, and services company, has appointed Pratik Nath as the managing director of its India Global Capabilities Centre (GCC). With over 25 years of experience in technology, Nath will lead the India operations from Bengaluru ensuring the teams enhance and optimise Epsilon’s solutions for clients. “Pratik’s deep engineering background and his experience in building a culture of innovation is exactly what we need at Epsilon India to grow to the next level,” said Myron Sojka, CTO of Epsilon. “Under his leadership, we will continue the next phase of growth for our India Global Capabilities Centre, where we will build deeper technical capabilities and maximise the value Epsilon delivers to our clients,” he added. Before joining Epsilon, Nath was VP of India IP development at Oracle Cerner, where he led strategy and delivery for a 2,000+ member IP development organisation. He played a key role in establishing and expanding the engineering footprint at Cerner India and holds six engineering patents. “Epsilon is an industry leader in the marketing and advertising technology space. Their ability to connect data, insights, and identity across the entire customer journey is unparalleled. I am eager to drive engineering innovation from India so that we can continue disrupting the industry,” said Nath. Nath is a graduate of the National Institute of Technology, Surat, with a Bachelor of Engineering degree. He also holds a Diploma in Advanced Computing from the Centre for Development of Advanced Computing, Bangalore.","excerpt":"Before joining Epsilon, Nath was VP of India IP development at Oracle Cerner, where he led strategy and delivery for a 2,000+ member IP development organisation.","categories":["AI News"],"tags":["GCC"],"author_name":"Mohit Pandey","publish_date":"2025-01-30T14:43:26","publication_year":"2025","word_count":243,"keywords":["GCC","programming_languages:R","AI","innovation","GAN","R"],"extracted_tech_keywords":["AI","R","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/epsilon-appoints-pratik-nath-as-md-of-india-gcc\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23794,"title":"PM Modi Asks Union Ministries To Embrace AI To Track Socio-Economic Growth","content":"File photo of Prime Minister Narendra Modi (Image source: @MIB_India) In a bid to lead the world in the area of artificial intelligence, Indian Prime Minister Narendra Modi this week asked the NITI Aayog to acquaint all the ministries with high-end technologies. “China is way ahead of India in a lot of areas but the government does not want to miss the bus this time when it comes to research and adoption of this technology… NITI Aayog has been asked to undertake all possible pilots happening elsewhere in the world, if it addresses India’s problems, so government departments and states are willing to use AI in their day-to-day functioning,” a source claimed. According to a report in a financial newspaper, the Indian think tank has also been asked to describe how artificial intelligence will be beneficial to address the most of country’s socio-economic problems. In fact, all central ministries have been asked to set up dedicated AI cells as soon as possible. This directive comes only weeks after the national AI Task Force submitted a report on exploring the possibilities of AI in development across various fields, including industry and research. Following this directive from PM Modi, the NITI Aayog will likely periodically review the progress made by the respective ministries. The AI Task Force has already recommended the following points about AI to the government: Leverage AI for economic benefits Create policy and legal framework to accelerate deployment of AI technologies Create concrete five-year horizon recommendations for specific Government, Industry and Research programs In the report, the AI Task force has also identified 10 specific domains that need attention with respect to AI. They are, manufacturing, fintech, healthcare, agriculture, education, retail and customer engagement, public utility services, aid for differently-abled persons and accessibility technology, environment and national security. “With rapid development in the fields of information technology and hardware, the world is about to witness a fourth industrial revolution… Driven by the power of big data, high computing capacity, artificial intelligence and analytics, Industry 4.0 aims to digitise the manufacturing sector,” Nirmala Sitharaman, former minister for Commerce and Industry, who had set up the task force, had said in August 2017.","excerpt":"In a bid to lead the world in the area of artificial intelligence, Indian Prime Minister Narendra Modi this week asked the NITI Aayog to acquaint all the ministries with high-end technologies. “China is way ahead of India in a lot of areas but the government does not want to miss the bus this time […]","categories":["AI News"],"tags":["Agriculture","AI (Artificial Intelligence)","AI in manufacturing","China","customer engagement","education","environment","FinTech","Healthcare Automation","Manufacturing","Narendra Modi","NITI Aayog","retail"],"author_name":"Prajakta Hebbar","publish_date":"2018-04-17T11:35:43","publication_year":"2018","word_count":362,"keywords":["API","Agriculture","Git","AI (Artificial Intelligence)","R","Manufacturing","China","artificial intelligence","RAG","analytics","Narendra Modi","AI in manufacturing","Go","AI","Healthcare Automation","customer engagement","environment","FinTech","big data","education","Aim","retail","NITI Aayog"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","Aim","RAG","R","Go","Git","API","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pm-modi-asks-union-ministries-to-embrace-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10162885,"title":"As Sam Altman Returns to India, Will OpenAI Offer Region-Specific AI Pricing?","content":"OpenAI CEO Sam Altman is set to visit India for the second time this Wednesday. His Asia tour has already seen several product launches so far and his upcoming visit to India is expected to bring key announcements tailored to the country’s unique needs. India, with its vast user base, is a prime candidate for simpler, more accessible ChatGPT interfaces tailored to local needs. In this light, OpenAI’s announcement last December about adding a phone-calling feature to ChatGPT is particularly exciting. It removes the need for an internet connection or high-end devices, meaning even users with basic flip phones or rotary phones will have access to AI assistance. “1-800-ChatGPT might seem like a silly gimmick, but the underlying principle is critical to scaling AI adoption,” wrote Google Deepmind’s Logan Kilpatrick on X. If implemented well, a feature like this will be a huge win for accessibility in developing countries. “The next billion AI users will not be on the existing UX’s, they will be using text, email, and voice,” he added, suggesting whoever lands this experience is going to win in a huge way. A mail sent to OpenAI did not elicit any response about the specifics of the phone-calling feature’s rollout in India. “One of the most significant advantages of such integration is its ability to empower users,” added Osama Manzar, founder and director of the Digital Empowerment Foundation. By enabling people to create content, search for information, and share ideas, through voice commands, ChatGPT could make technology more accessible to those with limited digital literacy. This would be particularly impactful for rural and non-internet users, opening up new opportunities for learning, communication, and accessing essential services. “However, there are critical challenges to consider. The first concern is the contextual relevance of the AI’s responses. ChatGPT’s backend systems may not always adapt well to the diverse linguistic and cultural needs of local communities,” he added, commenting that misalignment could reduce effectiveness and drive users away. Manzar also raised concerns about data privacy, as users may unknowingly share information without understanding how it’s stored or used. From 2008 to 2024, active SIM cards in India surged over threefold, surpassing one billion in a population of 1.4 billion, underlining the immense potential for AI adoption at scale in the country. From a regulatory perspective, OpenAI can introduce such a feature as long as it complies with the relevant data privacy laws in India. Towards Simpler UIs? Important to note that AI still remains a relatively new phenomenon in the country, unlike the internet. Beyond chatbots, intuitive design is key to making it accessible and human-centric. Developing AI solutions tailored to India’s needs and cultural context is imperative. “The key lies in leveraging existing platforms that are already deeply engaging Indian masses like WhatsApp, Google, and Facebook,” said Manzar. Despite the success of ChatGPT, it is interesting to observe that apps such as WhatsApp and YouTube continue to be the most popular in the country. “It is crucial to focus on how we can mine and utilise data generated on these platforms to create AI solutions that are not dominated by foreign tech giants, who might appropriate Indian content for their own benefit,” he added. India, a Price Sensitive Country While AI accessibility is improving, true adoption in India hinges on both usability and affordability. Simplified interfaces help, but without cost-effective pricing, AI may remain out of reach for many. The implications of high-cost and premium pricing models for advanced tools are huge in a price-sensitive country like India. DeepSeek is slowly changing the game for everyone, forcing the AI labs to introduce cheap and open-source alternatives or at least rethink their strategy going forward. For instance, OpenAI now offers ChatGPT Pro at $200\/month and is rumoured to introduce plans up to $2,000\/month due to high compute costs of advanced models. While economics often sees costs decrease over time. Many people responded to Altman on X, highlighting that in the current realm, $200 per month is comparable to salaries and average incomes in many economies outside the US – suggesting that AI subscription pricing cannot be the same globally. In India, for instance, the average monthly income is around ₹20,000. This puts into perspective that AGI is accessible to the common man. essential part of daily life for many in India and worldwide. More so, as ChatGPT continues to become an essential part of many people’s lives in the country and the world, at least in the urban parts. Interesting to note, Altman recently conceded on the future of AI ultimately being open-source. “I personally think we have been on the wrong side of history here and need to figure out a different open-source strategy,” Altman said in a recent AMA session on Reddit. “Not everyone at OpenAI shares this view,” he added. Push for Indic LLMs Closer home, the key question remains: Can India achieve its own free, open-source DeepSeek equivalent?  More so, with a focus on Indic datasets. Aravind Srinivas, CEO of Perplexity, urged that building a foundational model in the country is as important as building on the application layer. Much is written about how the internet is divided on this question. Infosys co-founder Nandan Nilekani and former People + AI head Tanuj Bojwani have advocated to really solve Indian problems through use cases. Ola chief Bhavish Aggarwal has announced Krutrim AI Lab and the launch of several open-source AI models tailored to India’s unique linguistic and cultural landscape.  This announcement aligns with India’s broader AI ambitions. The government has officially called for proposals to develop homegrown AI models, marking a decisive push toward sovereign AI that can compete globally. Through the IndiaAI Mission, startups, researchers, and entrepreneurs are invited to build large multimodal models, large language models, and small language models, ensuring AI that is deeply rooted in India’s languages and culture. Per the website, the government expects the models must be trained on diverse Indian datasets, comply with Indian regulations, and serve both public and strategic interests","excerpt":"Artificial Intelligence still remains a relatively new phenomenon in the country, unlike the internet.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Aditi Suresh","publish_date":"2025-02-04T21:33:19","publication_year":"2025","word_count":1000,"keywords":["Go","ChatGPT","OpenAI","AI","chatbots","AWS","Git","RAG","small language models","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","small language models","RAG","chatbots","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/as-sam-altman-returns-to-india-will-openai-offer-region-specific-ai-pricing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":29358,"title":"TEG Analytics and Jigsaw Academy Students Partner To Crack Website Sales Of An Automobile Company","content":"At a time when analytics industry is going through a talent crunch and upskilling is the case du jure  — Bengaluru-based TEG Analytics, analytics outsourcing company leaned on edtech startup Jigsaw Academy to solve industry use cases with anonymized data in the form of Capstone projects. The partnership — a win-win proposition provided TEG Analytics with ready solutions which could be implemented and also enabled Jigsaw Academy’s Postgraduate Program in Data Science and Machine Learning (PGPDM) students a chance to cut teeth on industry data and be mentored by a senior TEG Analytics member. In line with its code-first and hands-on methodology, Jigsaw Academy’s one of its kind edtech-industry partnership fosters a strong student data community and also provides students a chance to experience real world business cases. A big draw for TEG Analytics was IBM and University of Chicago’s participation in the course creation which ensured the quality of students’ work output is the best-in-class and in line with the industry’s requirements. Here’s how the edtech-industry partnership worked: Jigsaw Academy identified the right capstone partners to ensure ideal returns for both the organization as well as the students. PGPDM students with similar skill sets were grouped together, and then matched with the appropriate company for which they worked on a Capstone project. So, each Capstone Partner provides a sanitized dataset to their group  which was as close to the real world data as possible along with a problem statement. The partner also provides a business problem to be solved, however this is optional. Since the data is open-ended, the team can choose to interpret the data on their own and devise pointers and solutions that the company can use in the future. The projects typically lasts 5-6 months, and a point of contact from the company is assigned to each group as a mentor, to oversee the work and to approve the final output. TEG Analytics has had a productive experience working with Jigsaw Academy students thus far and this has been the case since the first batch of PGPDM students. One of the projects students undertook was on behalf of a global leader in vehicle manufacturing and the problem statement involved identifying High Value Activities on the company’s website, that led to worldwide vehicle purchase,” In terms of fulfilling the business requirements, the students ticked all the boxes and delivered high quality output. “We found the hypotheses on the sessional as well as the sales data to be both business driven as well as very well supported by all the relevant code and statistical results,” an official from TEG Analytics shared. Additionally, the exploratory analysis of the website session data was descriptive, visual and exhaustive. The methodology that was followed to prepare the analytical model data file also contained extremely rational inferences. Another highly positive aspect of the work was the cogent and thoroughly comprehensive approaches used in the multiple modelling. According to Sheeba Rajam, TEG’s Head-Talent Acquisition & Development, Jigsaw Academy students showed cutting-edge techniques in their execution of projects and addressed the problem statement effectively. “Their responsiveness and ownership for the completion of the deliverables of the project were extremely good,” she said. She further added that the POCs from the Jigsaw team also kept us regularly updated throughout the entire process and were very communicative. “Perhaps most hearteningly, we found the students to be extremely meticulous and structured in their thought processes, which shone through in the way they relayed information back to us,” she added. In an endeavour to shore up data literacy in India, professional partnerships like these are effective in bridging the much-talked about skill gap and allows both students and data-intensive organizations to benefit in some way with the exchange of information that takes place. Even in the future, TEG Analytics will take on the mantle of spreading data practices with professional partnerships.","excerpt":"At a time when analytics industry is going through a talent crunch and upskilling is the case du jure  — Bengaluru-based TEG Analytics, analytics outsourcing company leaned on edtech startup Jigsaw Academy to solve industry use cases with anonymized data in the form of Capstone projects. The partnership — a win-win proposition provided TEG Analytics […]","categories":["AI Trends"],"tags":["data literacy","jigsaw academy"],"author_name":"Richa Bhatia","publish_date":"2018-10-17T12:01:12","publication_year":"2018","word_count":642,"keywords":["data science","Go","machine learning","TPU","AI","data literacy","ViT","analytics","GAN","jigsaw academy","R","startup"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","TPU","R","Go","GAN","ViT","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/teg-analytics-and-jigsaw-academy-students-partner-to-crack-website-sales-of-an-automobile-company\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10011568,"title":"IBM &#038; AMD To Advance Confidential Computing For Cloud &#038; Accelerate AI","content":"IBM and AMD today announced a multi-year joint development agreement to enhance and extend the security and artificial intelligence offerings of both companies. This joint development agreement has been planned to expand this vision by building upon open-source software, open standards, and open system architectures in order to drive confidential computing in hybrid cloud environments. It would also support a broad range of accelerators across high-performance computing, and enterprise-critical capabilities such as virtualisation and encryption. According to Dario Gil, Director of IBM Research, “The commitment of AMD to technological innovation aligns with our mission to develop and accelerate the adoption of the hybrid cloud to help connect, secure and power our digital world.” He further stated, “IBM is focused on giving our clients choice, agility and security in our hybrid cloud offerings through advanced research, development and scaling of new technologies.” “This agreement between AMD and IBM aligns well with our long-standing commitment to collaborating with leaders in the industry,” said Mark Papermaster, executive vice president and CTO, AMD. “AMD is excited to extend our work with IBM on AI, accelerating data centre workloads, and improving security across the cloud.” Securing highly sensitive data has still continued to be a challenge for many companies. According to data from IBM’s Institute for Business Value, this has led to cybersecurity being the top barrier for adoption as well as the top criteria for selection of cloud providers. According to Gartner, confidential computing potentially omits the remaining barrier to hybrid cloud adoption for highly regulated businesses or any organisation concerned about unauthorised third-party access to data in use in the public cloud. Confidential computing is a technology, enabled by hardware, that allows the data associated with a running virtual machine to be encrypted, including while workloads are running. This capability helps prevent would-be attackers and bad actors from accessing confidential information, even in the event of a break-in. Confidential computing for hybrid cloud unlocks new potential for enterprise adoption of hybrid cloud computing, especially in regulated industries such as finance, healthcare and insurance. Engagement between AMD and IBM researchers on joint development activities under the agreement is now underway.","excerpt":"IBM and AMD today announced a multi-year joint development agreement to enhance and extend the security and artificial intelligence offerings of both companies. This joint development agreement has been planned to expand this vision by building upon open-source software, open standards, and open system architectures in order to drive confidential computing in hybrid cloud environments. […]","categories":["AI News"],"tags":["Cloud Computing","confidential computing","Cybersecurity"],"author_name":"Sejuti Das","publish_date":"2020-11-12T14:03:21","publication_year":"2020","word_count":356,"keywords":["artificial intelligence","programming_languages:R","AI","cloud computing","innovation","Git","GAN","Cloud Computing","ViT","confidential computing","Cybersecurity","R"],"extracted_tech_keywords":["AI","artificial intelligence","cloud computing","R","Git","GAN","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-amd-to-advance-confidential-computing-for-cloud-accelerate-ai\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10017329,"title":"IISc Announces A New PG Programme In Deep Learning","content":"Indian Institute of Science (IISc) along with TalentSprint announced the launch of a PG level Advanced Certification Program in Deep Learning. This 10-month executive program is designed for aspiring and practising AI\/ML professionals who want to build expertise in deep learning. The program is designed to provide a practical understanding of how machine learning algorithms can be developed and optimised for hardware, which can be applied in areas such as edge computing. Apart from this, the program aims at providing expertise in areas such as speech, text, image and video. The classes would be taught through live interactive sessions by renowned faculty from IISc, in association with TalentSprint. It will provide hands-on learning on curated projects with industry mentor support. The learners will get to work with capstone projects apart from an option of bringing their own projects. The program will also enable participants to establish a portfolio that demonstrates their learnings. This program will also help learners connect to the deep tech alumni network to reap long-term career benefits. “Deep Learning is increasingly being used to extract valuable insights from enormous amounts of data, build innovative products and improve customer experience, thereby enhancing revenue opportunities. This has led to a massive growth in need for professionals with expertise in Deep Learning,” shared Prof Chiranjib Bhattacharya, Professor and Chair of the Department of Computer Science & Automation, and Dean of the Advanced Deep Learning Program. He added that this program will fulfil that need. “Our team of research faculty will teach and mentor participants and help them build expertise in both the fundamentals and applications of Deep Learning,” he said. “This program has been specifically designed for working professionals keen to unleash the potential of deep learning in their careers.  This academic program will bring together lectures by top research faculty at IISc and hands-on labs and projects supported by industry mentors,” said Dr Santanu Paul, Co-Founder and CEO, TalentSprint. The enrolment for the first cohort are open now and classed with commence in March 2021. Click here for more info.","excerpt":"Indian Institute of Science (IISc) along with TalentSprint announced the launch of a PG level Advanced Certification Program in Deep Learning. This 10-month executive program is designed for aspiring and practising AI\/ML professionals who want to build expertise in deep learning. The program is designed to provide a practical understanding of how machine learning algorithms […]","categories":["AI News"],"tags":["Active Learning","ai machine learning and data analytics","data science machine learning ai"],"author_name":"Srishti Deoras","publish_date":"2021-01-07T18:46:15","publication_year":"2021","word_count":341,"keywords":["Go","machine learning","programming_languages:R","AI","ai machine learning and data analytics","ML","data science machine learning ai","automation","Active Learning","Aim","deep learning","edge computing","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","Aim","edge computing","R","Go","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iisc-announces-a-new-pg-programme-in-deep-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10053505,"title":"Data Science Professionals Need To Have Domain Knowledge: Sreetama Das, GSK","content":"From being a researcher in Computational Biology & Structural biology\/ Biophysics to a Senior data scientist, Sreetama Das has gained a wide knowledge base and worked in diverse industries like manufacturing and healthcare. Analytics India Magazine caught up with Das, who is currently serving as a Senior AI ML Engineer at GSK, to understand her insights on AI-based solutions in digital health. AIM: Given your experience in a broad career path from data science to AI and machine learning, what do you think will be the future of the tech-medical space? Sreetama Das: To briefly mention my background, I have worked with data science and machine learning applications to biomolecules as part of academic research during my PhD, and then worked in diverse industries like manufacturing and healthcare. This experience has shaped my view of what I think the field will evolve to be. Several areas in the tech-medical space use machine learning for improved outcomes – for example, research and development for drug discovery or repurposing, optimising clinical trials, manufacturing and quality control, digital health sensors, digital pathology patient triaging, to name a few. The data could be numeric, structured nicely in tables, or huge images or even messy text – that is what makes problem statements in healthcare exciting but challenging. Advancements in computer vision and natural language processing will improve solutions or even solve problems that were not feasible earlier. In addition, solutions focussing on explainability will find more acceptance. The overall trend is to move towards faster development, efficient manufacturing, better quality control, and easier access to health monitoring, disease detection, and patient treatment outcomes. In short, AI-based solutions will be assisting to improve lives. AIM: There have undoubtedly been significant advancements in the field of digital health. Is there a particular situation that prompted you to conduct additional research? Sreetama Das: Digital health has seen significant advancements in developing digital sensors for health monitoring and detection of diseases. I will cite an example from a project I was part of at my earlier organisation (Robert Bosch Engineering and Business Solutions, India) since I have been with GlaxoSmithKline for only a short time. Back there, the team developed a novel, non-invasive haemoglobin monitoring sensor based on photoplethysmography. We collected data from many participants, both healthy and frail, and trained machine learning models. As a result, our device performed better than some existing non-invasive hemoglobinometers, which tended to overestimate haemoglobin levels. The project required a lot of research, and our results have been published in several reputed scientific journals. AIM: Enterprises began migrating away from on-premises data centres and applications toward cloud and SaaS-based solutions. How feasible is this transition to the cloud in light of the numerous security concerns and ransomware attacks occurring? Sreetama Das: I am not an expert in this area, but I am aware of private commercial cloud and hybrid cloud offerings that combine the advantages of cloud with the security of on-premise solutions and are used to address regulatory concerns. Moreover, we are starting to see the use of blockchain in commercial cloud solutions and upcoming technologies like federated learning and its use in med-tech. So, I think the transition is feasible with well-thought-through strategies. AIM: In what we call a “combination of AI and Digital Health”, what are some near or so groundbreaking developments that carry future potential as well? Sreetama Das: One of the groundbreaking developments in recent times would be the development of AlphaFold (by Deepmind) which has revolutionised the science of protein structure solution and has important implications for drug design. The software provided near accurate theoretical models of protein 3D structures, which are presently obtained by time-consuming and expensive experiments or sometimes represented by not-so-accurate theoretical models. Implementation of AlphaFold generated models will dramatically reduce the search space, thereby reducing the time and expense in the drug discovery process. Other advancements include deep learning in several areas, for example, in the classification of chest X-rays of normal flu versus COVID-19-infected cases with high accuracy or in early detection based on voice abnormal respiratory sounds. In addition, image and video analytics are finding applications in monitoring physical activity or senior patients for falls or other issues. Finally, precision personalised medicine is also an area with significant potential. AIM: In general, why do you believe there are widespread misconceptions about artificial intelligence in healthcare, and why do they persist? Sreetama Das: I believe there are several sides to this question. Firstly, when there are breakthroughs in artificial intelligence algorithms, many articles in general forums are often very flattering, without trying to explain how they work, the input requirements or where they may fail. As a result, there is a disappointment when such tools do not generate the expected results, either due to inappropriate application or improper data. It is important to remember that gathering enough good data is still a challenge for several healthcare problems. Also, some problem statements may not be feasible to solve and require modification after multiple rounds of discussion with the stakeholders. Moreover, healthcare is a sensitive topic, and people are often apprehensive (and sometimes rightly so) about accepting ‘black-box’ solutions or the ‘fairness’ of such solutions. Hence, raising awareness about artificial intelligence and addressing concerns – data requirements, understanding how machine learning works and what is feasible, and model explainability – will help remove persistent misconceptions. AIM: Could artificial intelligence eventually replace clinicians? Sreetama Das: Human biology is highly complex – we still don’t understand everything. Clinicians make decisions by looking at many different aspects and based on their experience. It is hard to design ‘general’ machine learning solutions that can work similar to an experienced clinician – only solutions to very specific AI can develop well-defined problems and where a lot of data are available. Since healthcare delivery directly impacts people’s lives, the future would have AI solutions “assisting” the clinicians in their decision-making rather than replacing clinicians altogether. AIM: What resources, such as books and journals, would you recommend for aspiring professionals in this field? Sreetama Das: A lot of useful information is often found in blogs on Medium. I refer to GitHub for codebases. Other sites to look out for will be papers with code and company blogs for recent developments (e.g., Google, Facebook, Microsoft). There are also many free videos and course materials (e.g. MIT Courseware) for aspiring professionals to get started. AIM: What would you recommend as an efficient career direction for professionals aiming to make it big in the digital health space? Sreetama Das: I will answer this question based on my journey. Data science professionals need to have some domain knowledge to develop appropriate solutions. Since digital health is a vast space, it is important to be aware of that and identify the topics the professionals are interested in and understand. It also helps to be flexible and adaptable to change since different areas may gain traction at different times in a professional’s career. Finally, we need to be lifelong learners and acquire relevant knowledge when the need arises. Lastly, it is important to focus on the foundation for such careers. Good coding skills, understanding of key concepts, and good communication are important. Also, please remember that such projects are often based on teamwork, with different skillsets within the data science team (analyst, ML engineering, MLOps, etc.). So, it is important to talk to peers and learn more about these aspects.","excerpt":"AI-based solutions make it easier for patients to obtain health monitoring, disease detection, and treatment outcomes.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AlphaFold","Blockchain","data science professionals","Data science team","Deep Learning","Interviews and Discussions","Machine Learning","ML models"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-11-15T16:00:00","publication_year":"2021","word_count":1234,"keywords":["Data science team","computer vision","deep learning","data science","AlphaFold","artificial intelligence","analytics","data science professionals","machine learning","Blockchain","AI","ML","ML models","Machine Learning","MLOps","Aim","Deep Learning","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","computer vision","data science","analytics","MLOps","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/domain-knowledge\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10055206,"title":"Why Nobody Is Talking About Biases In NFT","content":"What if one were to ask you, “How much trust do you have in the promises of the metaverse?” Virtual reality took the real world by a storm of confusion, intrigue and questions. But it promised a world without the racist, sexist, or other biases present in society and technology. While the idea is certainly moving in the right direction, it was too good to be true for most people. That might be the case considering we are only a few months in on the metaverse and are already seeing instances of racism and sexism in one of its features, NFTs, or non-fungible tokens. NFTs, the JPEG files or digital certificates of authenticity, are secured via blockchain technology. They are the latest offerings of the metaverse to the art world, where dealers can buy digital artworks like paintings or sculptures. However, while these started from an innocent space, the complexities and biases of the real world are starting to blur the boundaries. Beeple’s Sexist Hillary Clinton Recently, taking the NFT world by breaking news was the purchase of artwork for nearly $70 million. A generally unknown digital artist, Beeple, rose to fame after his art titled ‘Winkelmann’s Everyday: The First 5000 Days’ was sold. The artwork is a collage of 5000 images the artist has taken since 2007. While critics have taken to the online space to analyse the NFT and find it problematic, the media has praised and promoted the artwork. The collage even made it to the cover of TIME magazine, with Beeple being titled a cultural icon. But in reality, the art piece is not living up to the utopian dreams of the metaverse. It contains multiple sexist, racist and objectionable cartoons. Ben Davis from Artnet revealed the pictures to include sexist features of Hillary Clinton portrayed through a misogynistic lens, unflatteringly sexualised, and illustrated in compromising positions with Donald Trump. The collage also consisted of casual racism with instances like sketches of dark-skinned men captioned as ‘Black Dude’. Trosley’s Derogatory Black Americans Another instance is Trosley’s ‘Jungle Freaks’ that depicted genetically modified gorillas in combat with zombies in a dystopian future. This digital artwork was previously owned by actor Elijah Wood, who later took to Twitter to reveal that he has sold the NFTs on accusations of the paintings being racist. The allegations highlight that the artwork depicts black Americans and marginalised people in a derogatory way. CryptoPunk’s Gender & Race Strewed Avatars CryptoPunks, one of the earliest and most famous collections of NFTs, consist of 10,000 collectable characters or avatars. Of the 10,000, over 6000 are male characters, and the rest are female. They have been generated algorithmically, making every character unique. The NFTs have recently come under media scrutiny after their price difference based on gender, race and skin colour. The phenomenon portrays a stark difference in the demand and price for light-skinned male characters as opposed to darker-skinned female characters. While experts and users have ratiocinated this to a lack of diversity among the investors, with most of them being light-skinned men, there might just be more to this story. Over 2000 male CryptoPunks have been sold in the past few months, while the number of female avatars sold plays in close to 1200. DeGenData is a statistics company that tracks CryptoPunk sales data. According to them, there have been 424 sales since November 1, 2021,  of which 267 are males and 157 are females. The male characters were sold at 99 ETH and the females at 95. To put this cost difference in perspective, as of today, one ETH is equivalent to $4000. Additionally, DeGenData also found that the average sale price of dark-skinned characters was much below that of the light-skinned. Similarly, NonFungible.com, an NFT ranking website, shows five of the top seven CryptoFunk sales (a seven-day average) having a light coloured skin tone. All the seven are men. Competing with righteousness A ray of a silver lining for the metaverse – the world is also witnessing artists willing to challenge and put a stop to the NFT bias while it is still possible. Artists like Drue Kataoka, a Japanese visual artist, have taken up initiatives to un-do these biases. Kataoka’s work is a charitable NFT titled “Will Your Heart Pass the Test?”. This is responding to the sexist and racist images in the NFT ecosystem and is inspired by a 3500-year-old Egyptian myth. As depicted in the story, the painting is a person’s heart weighed against a feather to determine the purity of the soul, demanding people today to introspect on the effects of people, power, and technology. These are small instances of bias in the universe that fall into the grey area that can be worked upon and improved. While the metaverse is surely promising, we still have to wait and watch how the world shapes up eventually.","excerpt":"Ben Davis from Artnet revealed the pictures to include sexist features of Hillary Clinton portrayed through a misogynistic lens, unflatteringly sexualised, and illustrated in compromising positions with Donald Trump.","categories":["AI Trends"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-12-11T10:00:00","publication_year":"2021","word_count":809,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","Ray","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","Ray","RAG","R","Go","Rust","Git","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-nobody-is-talking-about-biases-in-nft\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10064643,"title":"Jack Dongarra’s road to the Turing award","content":"Jack Dongarra has won the coveted Turing award for 2021. The Association for Computing Machinery (ACM) said his work has driven high-performance computing, and in turn impacted areas like artificial intelligence, computer graphics, analytics, deep learning, etc. Dubbed the Nobel Prize of Computing, the award comes with a cash prize of USD 1 million. Congratulations to Jack Dongarra, who receives 2021 #ACMTuringAward for pioneering contributions to numerical algorithms and libraries that enabled high performance computational software to keep pace with exponential hardware improvements for decades.\" https:\/\/t.co\/RxTTGaZ7Mx pic.twitter.com\/9JxRrIFJg9— Association for Computing Machinery (@TheOfficialACM) March 30, 2022 ACM president Gabriele Kotsis said Dongarra’s trailblazing work goes back to 1979 and he is one of the foremost and actively engaged leaders in the HPC community. https:\/\/twitter.com\/NVIDIAHPCDev\/status\/1509626504304644099?s=20&t=cwRov7KDtRI4F39xotSuZQ Background Jack Dongarra got his bachelor’s in mathematics from Chicago State University and a master’s in computer science from the Illinois Institute of Technology. His doctorate is in Applied Mathematics from the University of New Mexico. He is now the University Distinguished Professor of Computer Science in the Electrical Engineering and Computer Science Department at the University of Tennessee and Distinguished Research Staff in the Computer Science and Mathematics Division at Oak Ridge National Laboratory (ORNL). Dongarra is a Turing Fellow at Manchester University. Dongarra has created open-source software libraries and standards that use linear algebra as an intermediate language. His libraries also introduced important innovations such as autotuning, mixed precision arithmetic, and batch computations. His major contributions include: EISPACK is a collection of Fortran subroutines that compute the eigenvalues and eigenvectors of nine classes of matrices. These include complex general, complex Hermitian, real general, real symmetric, real symmetric banded, real symmetric tridiagonal, special real tridiagonal, generalised real, and generalised real symmetric matrices. LINPACK was written in Fortran by Jack Dongarra, Jim Bunch, Cleve Moler (under whom Dongarra did his PhD) and Gilbert Stewart. It is a software library for performing numerical linear algebra on digital computers. LINPACK makes use of the BLAS (Basic Linear Algebra Subprograms) libraries for performing basic vector and matrix operations. Initially, the LINPACK benchmarks appeared as a part of the LINPACK user’s manual. Basic Linear Algebra Subprograms (BLAS) is a routine that provides standard building blocks for performing basic vector and matrix operations. While the Level 1 BLAS perform scalar, vector and vector-vector operations, the Level 2 BLAS perform matrix-vector operations and the Level 3 BLAS perform matrix-matrix operations. Linear Algebra Package (LAPACK): Written in Fortran 90, LAPACK provides routines for solving systems of simultaneous linear equations, least-squares solutions of linear systems of equations, eigenvalue problems etc.ScaLAPACK is a library of high-performance linear algebra routines for parallel distributed memory machines. It solves least squares problems, eigenvalue problems, and singular value problems. It is designed for heterogeneous computing and is portable on any computer that supports Message Passing Interface (MPI) or Parallel Virtual Machine (PVM).The TOP500 project was launched in 1993. Dongarra has played an active role since the origin and formation of the TOP500 list since its inception. It uses his LINPACK benchmark to evaluate the performance of supercomputers. Autotuning Dongarra worked on methods for automatically finding algorithmic parameters that produce linear algebra kernels of near-optimal efficiency which often outperformed vendor-supplied codes. He pioneered the usage of multiple precisions of floating-point arithmetic to get accurate solutions quicker. HPL-AI benchmark In 2019, Jack Dongarra, Piotr Luszczek, and Azzam Haidar proposed the first implementation of the  High-Performance Linpack–Accelerator Introspection (HPL-AI) benchmark. Same year, the trio released the HPL-AI reference implementation that looks at supercomputers that use mixed-precision (16- or 32-bit) arithmetic in data science. Though traditional HPC works on simulation runs for modeling phenomena in physics, chemistry, biology, the mathematical models that used for these computations majorly require 64-bit accuracy. But machine learning models get the results needed at 32-bit ( even lower floating-point precision formats). The HPL-AI benchmark works at the intersection of high-performance computing (HPC) and artificial intelligence (AI) workloads.","excerpt":"He is now the University Distinguished Professor of Computer Science in the Electrical Engineering and Computer Science Department at the University of Tennessee.","categories":["IT Services"],"tags":["hardware","Supercomputers"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-04-11T11:00:00","publication_year":"2022","word_count":648,"keywords":["data science","Go","artificial intelligence","machine learning","AI","Scala","hardware","Git","Supercomputers","deep learning","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/jack-dongarras-road-to-the-turing-award\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049086,"title":"All about Spotify Pedalboard","content":"Spotify has quickly become a preferred audio streaming service among thousands of users worldwide. Widening its services into the music industry, Spotify has open-sourced its Python framework, Pedalboard. The Pedalboard is a Python library for adding effects to audio and supports many common effects outside the box. Essentially, it was built by Spotify’s Audio Intelligence Lab to allow creators to use studio-quality audio effects within Python and TensorFlow. DAWs and the Landscape Up until now, music and podcast producers have generally been using DAW – digital audio workstation, software packages that allow them to edit, manipulate and perfect audios. Currently, they are used in the majority of the audios we listen to today, including the content on Spotify. However, while DAWs enable musicians with flexible performance and enhanced control over audio quality, they are not made for programmers. Spotify’s Pedalboard allows programmers to leverage the “power, speed and sound quality” of DAWs in their code. Marketed as a new Python package, Pedalboard meets the criteria to bridge the gap between professional audio software and Python code. JUCE is the industry-standard framework for performant and reliable audio applications and is the leading framework for multi-platform audio applications. Pedalboard is built on top of JUCE, allowing it to have remarkable speed and quality. Additionally, the built-in convolution operator allows for high-quality simulation of speakers and microphones. Finally, the package supports several built-in audio effects, third party VST3® and Audio Unit plugins to increase sonic possibilities. Pedalboard, inspired by pedalboards used by guitar players, includes the stylistic effects found on the instrument. The programmer has the freedom and control to alter sounds with these effects and augmentations. The package also offers volume control tools like a noise gate, compressor, and limiter; and stylistic tools like distortion, phaser, filter, and reverb. The effects can be saved by grouping plugins together into a pedalboard to speed up the process. Machine Learning and Pedalboard The Pedalboard is a programmer’s heaven with the ML and content creation related functions it has. Given its speed, Pedalboard increases the speed of data augmentation and ensures enhanced results. It can be leveraged on models to increase the size of the training data and performance by taking a small dataset and augmenting it with audio effects. The engineering team had taken to a blogpost to claim, “Pedalboard has been thoroughly tested in high-performance, and high-reliability ML use cases at Spotify, and is used heavily with TensorFlow.” It also makes scripting audio effects applications with Python codes accessible. This makes it possible to automate some parts of the audio creation process – a feature that hasn’t been available with most tools to date. Additionally, the coding process assists the user in creating a line of workflow commands to apply a third-party plugin without launching DAW or importing\/exporting audio. This reduces the steps involved in the process with better results. Creativity is an integral part of the music creation process, requiring human input and not computation. Pedalboard ensures that it is supporting software for the artists and their creativity without hindering it. In fact, musicians and producers only need a little bit of Python knowledge to leverage its creative effects. This process would be a lengthy and time-consuming flow with DAW, but Pedalboard is easier for Python beginners. As a result, Spotify has placed Pedalboard as a bridge between code and music. The results of the tests run on Pedalboard by Spotify found that the common developer hardware is up to 300 times faster than the Python audio effects packages present in the market today. The company has been using Pedalboard internally to process millions of audio hours for over a year and have now open-sourced the software.The pedalboard is also ‘stage ready’ for macOS, Windows and Linux. Find Pedalboard’s code and documentation on GitHub.","excerpt":"Spotify’s Pedalboard allows programmers to leverage the “power, speed and sound quality” of DAWs in their code.","categories":["IT Services"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-09-21T10:40:13","publication_year":"2021","word_count":632,"keywords":["machine learning","AWS","AI","ML","Git","RAG","Python","Aim","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","TensorFlow","RAG","AWS","Python","R","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/all-about-spotify-pedalboard\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162926,"title":"MLDS 2025: Key Highlights from Day 2","content":"Day two of MLDS 2025, India’s biggest GenAI summit for developers, hosted by AIM Media House, continued with just as much energy, excitement, and insightful discussions as the first day. Multiple tech enthusiasts attended the event. Today’s captivating talk featured AI innovators Ganesh Gopalan, CEO and co-founder of Gnani AI, and Bharat Shankar, co-founder and chief product and engineering officer at Gnani AI, highlighting the transformative impact of AI-powered voice agents across various industries. A standout example mentioned was a user who forgot about his PAN card application and was assisted by an AI voice agent in just 24 hours. These AI-driven solutions, built for enterprise-scale use, are enhancing customer experiences, reducing call handling times, and automating millions of daily interactions across industries like banking, insurance, and retail. AI voice agents are more than just chatbots. “They are the next level of automation—handling real-time conversations at scale,” Gopalan explained. Their AI-powered voice customer service agents have already processed over 30,000 concurrent calls and millions of daily interactions. The technology works across platforms, including telephone, WhatsApp, and iMessage, automating customer service and reducing operational costs. “Our AI can lower call handling time by 15% and improve outcomes for customer support agents by 40%,” Gopalan shared. In another interesting session, Ritesh Agarwal, solution architect at Talentica Software, discussed hallucinations as a key challenge with AI. Although databases and a few other methods can mitigate this, they are not the only solution. Agarwal and his team found it difficult to use AI to handle over 10 million stock-keeping units (SKUs) in e-commerce searches, so they used AI to fix AI-related problems. He further explained that they use test queries to retrieve results based on semantic or cosine similarity. They then send the image and query to OpenAI for validation of stock items, which returns a simple true or false flag indicating accuracy. This flag is stored in the database to manage hallucinations. Another highlight of the day was an exciting workshop, “Building Scalable Multi-Agent Systems with Gemini: From Scratch to EV Industry Report,” led by Lavi Nigam, a developer relations engineer at Google Cloud. This hands-on session delved deep into the world of multi-agent systems, showing participants how to build scalable AI applications from the ground up using Google’s Gemini AI models. Attendees explored key design patterns, architecture, and essential tools needed to develop robust, intelligent AI systems. Finally, a prominent highlight of the day was the “40 Under 40 Data Scientists” awards at MLDS 2025, hosted by AIM Media House. This prestigious recognition brought together some of India’s brightest minds in data science, celebrating their innovation, impact, and contributions to the industry. These young data scientists are driving the future of analytics in India, shaping the landscape with their vision and expertise. The award highlights real innovators and achievers, setting them apart as leaders in the field. AIM’s expert panel of editors and industry veterans carefully reviewed and selected the winners, making this a truly elite recognition in the world of data science.","excerpt":"Today’s captivating session featured innovators at Gnani AI highlighting the transformative impact of AI-powered voice agents across various industries.","categories":["AI Features"],"tags":["AI","Developers","MLDS"],"author_name":"Shalini Mondal","publish_date":"2025-02-06T17:55:30","publication_year":"2025","word_count":501,"keywords":["data science","GenAI","OpenAI","AI","chatbots","ML","MLDS","Aim","multi-agent systems","analytics","R","Developers"],"extracted_tech_keywords":["AI","ML","data science","analytics","GenAI","OpenAI","multi-agent systems","Aim","chatbots","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mlds-2025-key-highlights-from-day-2\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10093504,"title":"This Canadian AI Startup has Designed the World’s First Humanoid General-Purpose Robot That Actually Works","content":"Vancouver-based artificial intelligence and robotics company, Sanctuary AI, yesterday unveiled a new advanced, general-purpose robot called Phoenix. The company claimed that it has developed the world’s first humanoid general-purpose robot powered by Carbon, a unique AI control system. The SOTA AI system offers human-like intelligence and enables robots to do a wide range of tasks to help address the labour challenges affecting many organisations. In March this year, the company announced its first commercial deployment, a significant milestone in the company’s progress toward full commercialisation. In a little less than two months the company has now announced the sixth generation of technology. Founded in 2018, Sanctuary AI is on a mission to create the world’s first human-like intelligence in general-purpose robots that will help humans work safely, efficiently, and sustainably. The members of Sanctuary AI are also part of the founding team at D-Wave, Kindred, Creative Destructive Lab and others. In a bid to fulfil the ambitious mission of creating human-like intelligence in general-purpose robots, it has partnered with companies like Apptronik, Bell, Common Sense Machines, Contoro, Cycorp, Exonetik, HaptX, Magna, Tangible Research, Verizon Ventures, Workday Ventures, and others. In March last year, the company raised Series AI funding. In November 2022, the company received a C$30 Mn strategic innovation fund (SIF) contribution from the Government of Canada, bringing its total funding to over C$100 Mn. How is Sanctuary AI Different? While the majority of the companies, including the likes of Tesla are still in the prototype and experimentation stage, Sanctuary AI claimed that its technology is already capable of completing hundreds of tasks identified by customers from more than a dozen different industries. Sanctuary AI chief and cofounder Geordie Rose said that it designed Phoenix to be the most sensor-rich and physically capable humanoid robot ever built and to enable Carbon’s rapidly growing intelligence to perform the broadest set of work tasks possible. Meanwhile, here is a glimpse of Tesla’s humanoid robot, exploring the real world. https:\/\/twitter.com\/Tesla_Optimus\/status\/1658576897490530305?s=20 “I think Tesla is in a great position to build the largest humanoid robot data flywheel ever. Very optimistic to see the latest progress on Tesla,” said Linxi Fan, AI scientist at NVIDIA. He said Optimus can reuse the powerful vision system built for FSD. “The decision to use a camera instead of LIDAR makes the models instantly transferable,” he added, saying that many humanoid tasks likely need less precise and rigorous visual processing than self-driving. Further, he said that Tesla has deep experience in mass-producing hardware. “The first company to deploy humanoid en masse will be able to spin the data flywheel in the wild and compound the model capability faster than competitors,” he added. On the other hand, Sanctuary AI claimed its literal take on ‘general purpose’ and emphasis on creating a technology that can conduct physical work just like a person sets them apart in the industry. Rose said that to be general purpose, a robot needs to be able to do nearly any work task, the way humans typically do, in the environment where the work is. “While it is easy to get fixated on the physical aspects of a robot, our view is that a robot is just a tool for the real star of the show, which in our case is our proprietary AI control system, the robot’s Carbon-based mind,” he added. In March this year, AI robotics startup Figure also claimed to have released the world’s first commercially available general-purpose humanoid robot Figure O1, the prototype of which bears a strikingly close resemblance to Tesla’s robot Optimus. Read: Meet Tesla Optimus Clone","excerpt":"A majority of the companies today, including the likes of Tesla, Figure and others, are still in the prototype and experimentation stage.","categories":["Deep Tech"],"tags":["AI Startups"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-05-17T16:31:45","publication_year":"2023","word_count":597,"keywords":["Go","API","funding","artificial intelligence","AI","innovation","Aim","GAN","R","AI Startups","startup"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","API","GAN","innovation","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/this-canadian-ai-startup-has-designed-the-worlds-first-humanoid-general-purpose-robot-that-actually-works\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10101219,"title":"Canva&#8217;s Magic Studio Catches Up To Adobe Firefly","content":"Canva was once seen as struggling to keep pace with Adobe a few months back. However, it swiftly dispelled those doubts. Canva has upped its ante. On its tenth anniversary, the Australian design company launched Magic Studio, a powerful suite of AI tools designed to simplify design creation for businesses. Canva just released their most impressive AI update yet…A whole \"Magic Studio\" that makes designing way easier.10 great features: pic.twitter.com\/Ky7ua9fMUo— Borriss (@_Borriss_) October 5, 2023 Magic Studio significantly enhances Canva’s AI capabilities, introducing innovative tools like Magic Switch, Magic Grab, Magic Expand, Magic Morph, Magic Alt Text, and Magic Animate. These tools leverage Canva’s in-house AI technology along with partnerships with industry leaders like Google and OpenAI, offering users a seamless and advanced design experience. This development came just a week after Adobe launched its latest FireFly web application. Following a six-month beta phase, Firefly’s advanced features are now integrated into Adobe Creative Cloud, Adobe Express, and Adobe Experience Cloud and are now accessible for commercial purposes. Canva Vs Adobe FireFly Magic Studio marks Canva’s bold stance against rivals such as Adobe Express and Microsoft Designer. While Adobe Firefly appeared promising in recent months, Canva’s innovative AI tools have positioned it as a strong contender, ready to give its competitors a run for their money. Under the Magic Studio, Canva introduced Magic Media where users can seamlessly create images and videos from text. Unlike Adobe Firefly which was dependent on its in-house capabilities, Canva has outsourced image generation processes to OpenAI’s DALL·E and Google’s Imagen. Moreover, Canva partnered with Runway bringing Runway’s cutting-edge Gen-2 AI technology directly into Canva’s ecosystem to create videos from text. Notably, as of now Adobe Express does not have the capability to create videos. A user of X said “This is an awesome power play between the two companies. Canva already is an amazing tool. I’ll be curious to see how Adobe will respond to this collaboration.Who knows maybe they will absorb PIKA”. BIG news: Runway partnered with Canva.They're making Gen-2 accessible directly in Canva's new Magic Media app, which is available to their pro tier customers. Canva has 150 million monthly users. Generative video is starting to go mainstream. pic.twitter.com\/DLZoKEeNMF— Nick St. Pierre (@nickfloats) October 4, 2023 Canva’s move to merge DALL·E and Imagen into one platform is a smart choice, eliminating the hassle of switching between various image generation tools. Presently, DALL·E 3 stands as one of the top image generation tools in the market, rivaling competitors like Midjourney. Meanwhile, Adobe FireFly which is integrated into Adobe’s Creative Cloud flagship products like Adobe Photoshop, Adobe Premiere Pro, and Adobe After Effect is developed with the help of NVIDIA Picasso cloud service. Adobe’s FireFly was trained on Adobe Stock images, openly licensed content, and public domain content where copyright has expired. However, it seems this approach has not only handicapped Firefly’s image generation capabilities, but resulted in an inferior product. Recently, many users have expressed dissatisfaction with the image quality produced by Adobe FireFly when compared to Midjourney and Dall·E. According to recent reports, Adobe is planning to introduce a new photo editing tool dubbed Project Stardust. The tool automatically identifies individual objects in regular photographs, allowing them to be easily moved around and changed. Surprisingly, it is pretty much similar to Canva’s Magic Grab. Magic Grab lets users pick and separate the main part of a photo. They can then edit, move, or change the size of this part, and add text, stickers, or other things to the picture. It appears that Canva and Adobe are having a stiff competition over the features they have. FireFly’s AI models for images and text effects now support prompts in over 100 languages. In contrast, Canva’s Magic Switch translates designs into 100+ languages seamlessly within the page interface. Which is users’ favourite Canva currently boasts of about 150 million active users all around the world according to the company’s blog. This shows that Canva’s user base has surged to nearly four times that of Adobe’s estimated 26 million Creative Cloud subscribers. Interestingly, Adobe doesn’t disclose exact user numbers, only the value of its subscription business, making the comparison approximate. Apart from the features, pricing plans also play a key role in deciding which platform users opt for. Canva’s Magic Studio tools are accessible to Canva Pro and Canva for Teams users at a monthly fee of $14.99, providing unlimited usage. Free users can access select features with limitations, including 500 monthly usages of Magic Edit per user, along with Magic Alt Text, Beat Sync, and Canva Assistant. Meanwhile, Adobe Express offers two plans: free and premium. The premium plan for individuals costs $9.99 per month. However, for users or organizations that have an Adobe ‘Creative Cloud All Apps’ license, Adobe Express comes included with your subscription, which costs $54.99 to $84.99 per month. However, it is important to note that starting November 1, 2023, the price of the Adobe Creative Cloud single apps and All Apps plans will increase in select countries. Interestingly, Adobe recently introduced a new credit-based model for generative AI across Creative Cloud. After the plan-specific number of “fast” Generative Credits is consumed, subscribers can continue to generate content at slower speeds, or buy additional “fast” Generative Credits through a Firefly paid subscription plan. It will be intriguing to see what Adobe unveils at Adobe Max next week, starting on October 10th to challlenge Canva.","excerpt":"Adobe is planning to introduce a new photo editing tool dubbed Project Stardust","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-10-07T10:00:00","publication_year":"2023","word_count":902,"keywords":["Go","OpenAI","AI","programming_languages:R","ML","programming_languages:Go","RAG","generative AI","GAN","R"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","RAG","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/canvas-magic-studio-catches-up-to-adobe-firefly\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69724,"title":"10 Leading University Courses on IoT (Worldwide)","content":"Education Industry i.e. universities mainly in Europe and online platforms like coursera are launching courses specific to the stream of IoT to provide a pool of trained and talented IoT professionals. Most of these courses are still in their early stages as they have been recently launched and also many of them are online, very few being taught in universities. Through this article we explore the top universities that offer courses in the field of ‘Internet of Things’ and have an in-depth look as to what they offer their students. This list does not include courses offered by online platforms like Coursera or EdX, though one can find many IoT courses on those platforms. IoT India Magazine brings you a list of top 10 University courses in the field of IoT around the world (listed in alphabetic order). This would help our IoT aspirants, who wish to enter the field of IoT and contribute to the change, to get a fair idea of various education options available worldwide. Beijing-Dublin International College – BE The Internet of Things (IoT) Engineering The Internet of Things (IoT) Engineering is an interdisciplinary bachelor’s degree programme that combines the study of electronic engineering and computer science, with an emphasis on internet technologies, wireless communications, sensor devices, and cloud computing. The Internet of Things (IoT) programme modules are taught in English. The majority of course modules are taught by University College Dublin academic staff who are experts in research areas relevant to this exciting interdisciplinary programme. The Internet of Things (IoT) programme provides access to state-of-the-art technologies and students will engage in the creative development of the innovative Internet of Things. The first two years of instruction offer twice as much English language exposure than is currently required by the Chinese educational system. Students will graduate with two degrees from both Beijing University of Technology and University College Dublin, both highly regarded world-class universities. CIFF Business School – Master in Internet of Things (IoT) The Master in Internet of Things (IoT) is the answer to the ongoing revolution of the new generation of uses and applications of mobile devices, wearables and ubiquitous sensors rapidly emerging with the deployment of new networks and specialized hardware. The program provides the knowledge to develop solutions in different hardware platforms and networks from a holistic perspective, from data acquisition to analysis, including the potential of Big Data technologies in the backend need. Malmo University – Computer Science: Internet of Things, Master’s CourseM The aim of this course is that students should deepen their understanding of the Internet of Things (IoT) and how to design and engineer IoT-based systems. This Internet of Things master course involves advanced technologies and research in Computer Science. Application areas include for instance: smart transportation, smart cities, smart living, smart energy, smart health, and smart learning. Relevant research areas include for instance Self-Adaptive Systems, Cyber Physical Systems, Systems of Systems and Engineering of all such systems. Within this master course the student has the possibility to deepen her\/his own understanding of Computer Science and Internet of Things and to develop the basis to continue her\/his studies after graduate level (PhD studies). MIT Professional Education – Internet of Things: Roadmap to a Connected World (Certification) While the promise of the Internet of Things (IoT) brings many new business prospects, it also presents significant challenges ranging from technology architectural choices to security concerns. MIT Professional Education’s new Internet of Things: Roadmap to a Connected World course offers important insights on how to overcome these challenges and thrive in this exciting space. The concept of Internet of Things (IoT), which has roots at MIT, has begun to make an impact in industries ranging from industrial systems to home automation to healthcare. MIT researchers continue to conduct ground-breaking research on topics ranging from RFID to cloud technologies, from sensors to the World Wide Web. Queen Mary University of London – MSc, Internet of Things This innovative Internet of Things (IoT) MSc programme will help you adapt to become one of the highly skilled and in-demand engineers who are able to fully exploit the potential that these technologies offer. The MSc in IoT is designed to meet the demand for a new kind of IT specialist and skills, those who can: engineer new interactive products – things; acquire, fuse and process the data they collect from things; interact with, and interconnect these things as part of larger, more diverse, systems. The MSc IoT is organised along 3 pathways: Data pathway, Engineering pathway, and the Intelligent Sensing pathway to enable students to focus on these different aspects of the course. Royal Holloway, University of London – Masters in The Internet of Things The Masters in The Internet of Things at Royal Holloway will provide you with advanced knowledge and skills in three essential and critical areas: Data analytics is essential for getting value from the IOT. For example, in Formula One racing there are hundreds of sensors providing thousands of data points for analysis such as tyre pressure and fuel burn efficiency, which have to be collected in real-time for very quick analysis by race engineers onsite. Distributed computing and systems concern technical aspects such as algorithms for distributed coordination, time-synchronisation, scalable storage, virtualisation and cloud computing technologies, as well as methodological aspects such agent-based modelling and simulation. Cybersecurity is another essential aspect of the IOT. Recent news such as the safety recall issued by Fiat Chrysler of 1.4m vehicles in the US after tech magazine Wired reported that hackers had taken control of a Jeep Cherokee via its internet-connected entertainment system, are examples of how privacy, safety and security are major concerns for the IOT. The programme can be taken part-time over two years, but without a placement. University of Oxford: Data Science for the Internet of Things (certification) The University of Oxford has put together a data science course for developers targeting the IoT space based on statistics suggesting that one in five developers are targeting IoT for upcoming projects. The course duration overall is six months with the first three months being conducted mainly online with a few face-to-face lessons taught at the university’s campus in Oxford. Following this, students have three months of coding exercises and learning consolidation activities. The four key areas of the course include the IoT ecosystem, data science (including time series data), programming and problem solving, and online engagement and programming exercises. The University of Oxford says participants are expected to have “a mindset of exploration” to be able to address real world problems with data science. The course requires students to have a good grasp of mathematics and will use IoT databases, large-scale IoT datasets, specific case studies and commercial data products like MapR, Numenta and MongoDB. To run the course, the university is also working with IoT companies and organisations such as Hyper\/Cat, FlexEye, Microsoft, Siemens and MongoD. University of Salamanca – Master in the Internet of Things The Master Degree in Internet of Things at the University of Salamanca is designed to train multidisciplinary professionals with the necessary skills to adapt to the different challenges the Internet of Things proposes. University of the West of Scotland – MSc Internet of Things (IoT) The MSc Internet of Things (IoT) forms part of the Advanced Computing framework here at UWS and is a post graduate course focusing on teaching advance topics of the internet-of-things. It is designed to develop practical skills by providing a comprehensive understanding of fundamental knowledge and hands-on experience of IoT technologies within industry and academia. In the programme, unique and challenging modules are introduced for example Mobile networks and smartphone applications, Data mining and visualization, in addition to Ethics for IT professionals and Object-oriented analysis and design that provide the basics for Internet of Things (IoT). Advanced wireless networking technologies and Internet of Things are introduced to in depth knowledge of the field. In addition, two modules which focus on research and the latest trends in the field, namely, Research design and methods and Emerging topics in smart networks, help to define and complete the research project on time. Waterford Institute of Technology – BSc (Hons) in the Internet of Things This course will explore the disciplines, technologies, tools and business opportunities involved in both sensing and connecting people, places and things. Powerful, connected, always-on devices and sensors, combined with sophisticated cloud infrastructure, are fast becoming a major focus for new products and services. The graduate will be ideally positioned with a unique combination of knowledge in a new and exciting field. In keeping with IoT trends, the programme will avail of open, web-based technologies and accessible electronic devices that are driving the emergence of IoT. Students will be encouraged to connect and incorporate their own personal devices in project work. They will programme diverse devices, including embedded sensors, mobile phones, single board computers and cloud systems. e.g. Raspberry Pi Students will also share and disseminate their work using industry standards and collaborative tools, such as github, building a digital portfolio of work, which will allow them to showcase their broad spectrum of skills to potential employers.","excerpt":"‘Internet of Things’ has been in the spotlight lately and with most of the world’s countries including India aiming at becoming connected through IoT, it is definitely something to watch out for. With IoT gaining momentum, education industry is also not far behind to cater to the demand that would arise in the coming years.","categories":["AI Trends"],"tags":["best iot training in india","iot training india","which countries have good cybersecurity"],"author_name":"AIM Media House","publish_date":"2016-08-04T05:45:58","publication_year":"2016","word_count":1518,"keywords":["data science","Go","AI","cloud computing","iot training india","MongoDB","which countries have good cybersecurity","distributed computing","RAG","Aim","analytics","best iot training in india","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","cloud computing","distributed computing","MongoDB","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-leading-university-courses-iot-worldwide\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10023267,"title":"How Fujitsu Is Using Artificial Intelligence","content":"Companies across the globe are exposed to a variety of risks. While some of them can be identified and avoided through strategic planning, others can not be even tracked. One of these dangers is a product recall, which normally occurs after a product or a service has been released, thereby adding huge costs to the company and irreversible damages for many. Fujitsu, a Japanese firm, recently developed an AI system capable of highlighting irregularity in the product’s appearance to detect associated issues at an earlier stage, thereby providing the chance to correct them before the product is released in the market. The AI technology will be used for image inspection, which will allow for the extremely detailed identification of a wide range of external abnormalities on manufactured objects, such as scratches and production errors. How does it work? The particular AI-enabled model is pre-trained on images of the products with simulated abnormalities. The company uses real images of defective goods pulled from a production line’s inspection process for the training data. Although many products have similar shape and appearance, the AI-based tool has the capability to correctly identify abnormalities associated with the product. For example — the frayed out threads of the carpet made of different materials or colour or defective wiring patterns on circuit boards can be identified by the AI tool with precision. The Fujitsu lab further confirmed the effectiveness of the AI-model in reducing the man-hours required to inspect the printed circuit-boards by at least 25%. Photo Courtesy: Fujitsu Official Website The earlier methods of training the AI model were based on the tendency to focus on individual characteristics of a product, rather than working on all characteristics of even similar looking products, to identify abnormalities with accuracy. As a result, it is essential to capture a wide range of features of a standard image while training AI to perform quality control tasks. Moreover, it will reduce the workload of the manufacturing industries and enhance productivity. The Necessity The first and foremost reason to have more AI-based models is to reduce the enormous cost associated with recalling goods and services. Take, for example, the most recent case of the Hyundai’s battery fiasco. Hyundai had to recall more than 82,000 vehicles – thereby costing the company around $900 million, amounting to $11,000 per vehicle. Similarly, General Motors had recalled around 7 million vehicles due to faulty airbags that hit the company with a whopping $1.2 billion. Secondly, it creates an unnecessary burden on the companies’ working staff, leading to increased man-hours, overburden of the work, and delays in meeting the targets set by the organisation. This delay is avoidable by emphasising the pre-production phase and adopting AI-based tools for precise product identification. Thirdly, defeated products in the market can cause injuries and fatalities, creating a massive brand image declination. Lastly, the Consumers Protection Laws of the respective countries will hold companies accountable for the defects and the harm caused to the consumers. This has been seen recently in March 2021, where Johnson & Johnson (J&J) has appealed with the US Supreme Court in a final effort to reverse one of the country’s biggest product liability verdicts. The Way Forward It’s better to embrace the latest AI, ML-based models, and technologies to provide a new life to companies’ production facilities to enhance the final products before rolling out in the market. Rather than facing trials, managing brand crises, or paying hefty sums, companies can look out for deep tech solving the problems and providing a cushion for the long-term good.","excerpt":"Latest AI-based model from Fujitsu can reduce the cases of products and services recalling for the company.","categories":["AI Features"],"tags":["AI for manufacturing","Fujitsu","Machine Learning"],"author_name":"kumar Gandharv","publish_date":"2021-04-02T18:00:00","publication_year":"2021","word_count":591,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","ML","Machine Learning","Fujitsu","Ray","AI for manufacturing","ViT","GAN","R"],"extracted_tech_keywords":["AI","ML","Ray","AWS","R","Go","GAN","ViT","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-fujitsu-is-using-artificial-intelligence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10004608,"title":"How This Cybersecurity Startup Is Using Machine Learning To Enhance Threat Intelligence","content":"The COVID pandemic has massively escalated the surge of cyberattacks and data breaches despite having robust security controls, software, and solutions abundantly available in the market. A lot of this could be attributed to the vulnerability businesses offer the cybercriminals to take advantage of the situation quickly. While the conventional cybersecurity approach has benefited many, having cybersecurity without cyber-intelligence and necessary awareness can put the security professionals off-guarded to more complicated and novel threats. Furthermore, with limited cybersecurity resources, businesses need to prioritise their efforts to strengthen cyber posture effectively; however, many organisations do not have an anchor point or a guiding principle, to begin with. With cyber-intelligence inputs missing from cybersecurity capabilities like incident management, vulnerability management, risk assessment and brand monitoring, businesses end up running their security practice in silos instead of an integrated approach. And, thus, in an attempt to revolutionise the cyber threat visibility and intelligence market, CYFIRMA, a cyber analytics startup assists businesses to understand the relevance of the current threat landscape. Not only it provides insights on threat actors and indicators, emerging threats and digital risks, but also automatically applies intelligence into cyber posture management. To dig deeper, Analytics India Magazine got in touch with the chairman and CEO of the company, Kumar Ritesh, to understand how the company uses a predictive intelligence-driven approach to discover cyber threats. Incubated under Antuit Inc, CYFIRMA was started as a business analytics firm in 2017 under the cybersecurity analytics arm, which developed the first-ever intelligence-driven approach — CAP (Cybersecurity Analytic Platform) for companies. Later in 2018, the company established a foothold in Japan and secured marquee clients such as Mitsubishi Corporation and NEC. However, in 2019, the company spun-off from Antuit to become an independent entity, backed by Goldman Sachs, Zodius Capital and Z3 partners. Also Read: Dark Web Peddlers Are Selling Fake COVID 19 Vaccines DeCYFIR Uses Strategic Cyber-Intelligence For Threat Discovery In June 2020, amidst COVID, CYFIRMA launched DeCYFIR, their flagship product which provides threat visibility and intelligence, cyber-situational awareness for businesses and cyber incident analytics. DeCYFIR is cloud-based threat discovery and cybersecurity platform, which discovers threats from hackers’ trenches, decodes signals from noise to get the most in-depth insights and apply threat intelligence to take necessary actions. Ritesh said that with the platform, it becomes easy to connect the dots to provide businesses with full contextual details on their threat landscape. “Our analysis uncovers the following insights – who is the threat actor, what assets are interesting for hackers, why the compelling interest, when is the attack mounted, and what is going to be the approach. In a nutshell, we help organisations predict future cyberattacks,” said Ritesh. According to the company, DeCYFIR is the only product currently available in the market to provide cyber intelligence to businesses to get a hacker’s perspective and their approaches. Not only it discovers hidden signals by recognising signs of an impending attack, but it also predicts the attacks quite early providing enough time for businesses to take necessary actions. The DeCYFIR platform picks up ‘indicators’ of the threat attack rather than the ‘indicators of compromise.’ Alongside, the DeCYFIR platform provides strategic, management and tactical intelligence which can be harnessed in totality to make accurate decisions. While strategic cyber-intelligence provides long-term implications for businesses such as changing the overall regulatory framework, the management cyber-intelligence highlights the approaches for a robust cybersecurity planning, and tactical cyber-intelligence focus on techniques to examine the indicators of compromise. Also Read: CYFIRMA Raises Series A Funding To Enhance Its Cybersecurity Initiatives How Does DeCYFIR Work? The platform DeCYFIR works on four logical layers — data collection, data analysis, data dissemination, and reporting. In the stage data collection, the platform knows where to look for collecting relevant data, which is critical for unlocking actionable insights. Secondly, in the data analysis layer, the platform applies correlation, attribution and association using AI and ML to seek indicators of threats beyond IoC. The platform further runs probability models to predict the likelihood of a cyber-attack. This layer helps in understanding the threat vectors, actors, method etc. In the third layer of data dissemination, it reads signals to identify farthest threats, consumption to apply cyber intelligence and predictions to oppose proaction. And lastly, in the reporting stage, the platform provides relevant information on a dashboard highlighting early warning of impending threats and real-time insights. The company also uses various NLP engines and language efficacy models to gather and generate information and classification and regression algorithms to enhance the threat intelligence insights of the platform. Alongside, it also uses probability models to anticipate the upcoming threats, and decision tree and attribution algorithms to connect different pieces of threats. Explaining further, Ritesh started — the company predicts the impact and probability scores based on a regression model, to keep a check on the new threats evolving every day. Also, CYFIRMA uses semantic, syntactic and lexical analysis to summarise the text of articles. Moreover, “based on the various attributes of an URL, the company also manages to predict if the URL is malicious in nature or not,” said Ritesh. Also Read: How Misconfigured Containers May Create Cybersecurity Issues For Companies Challenges & Future Road Map As a disruptor in the cybersecurity space, CYFIRMA has faced quite a few changes to bring a change. One of their prominent concerns was to educate organisations and business leaders about the disadvantages of the conventional approaches and layered defences. “And therefore, we needed to invest in substantial resources to showcase the new way of looking at cybersecurity,” said Ritesh. The company noted that there was an urgent requirement for companies to change their mindset and focus on more ‘intelligence-driven’ methods of managing cybersecurity. This approach would help them redirect their resources to predicting impending attacks. Additionally, till date, CYFIRMA has secured $8 million funding in their SeriesB round, which they claimed to be using for product development and enhancement, market expansion and business development. Currently, the company works with several companies, including government bodies, Fortune 500 MNCs, and commercial businesses. A couple of named clients are Mitsubishi Corporation, Toshiba, NEC, Suntory, SBI Holdings, Digital Hearts and Toppan. In the light of COVID pandemic, CYFIRMA has been helping customers with early advisory and helping them predict the next attack so they can close their cybersecurity gaps in time. Simultaneously, CYFIRMA has been actively advocating governments, businesses and the public on the importance of cybersecurity education and awareness. “To beat hackers in their own game, we have to stay a step ahead, and this calls for continuous iteration and improvement. We, at CYFIRMA, are focused on building the best product in our category and will continue to invest in product engineering,” concluded Ritesh.","excerpt":"The COVID pandemic has massively escalated the surge of cyberattacks and data breaches despite having robust security controls, software, and solutions abundantly available in the market. A lot of this could be attributed to the vulnerability businesses offer the cybercriminals to take advantage of the situation quickly. While the conventional cybersecurity approach has benefited many, […]","categories":["AI Startups"],"tags":["AI Algorithms","cloud business intelligence solutions","how artificial intelligence works","how does artificial intelligence work","Intel","Machine Learning","Startups"],"author_name":"Sejuti Das","publish_date":"2020-08-11T13:00:00","publication_year":"2020","word_count":1115,"keywords":["Go","cloud business intelligence solutions","how artificial intelligence works","API","AI","ML","Machine Learning","Scala","Git","AI Algorithms","how does artificial intelligence work","NLP","Aim","analytics","Startups","R","Intel"],"extracted_tech_keywords":["AI","ML","NLP","analytics","Aim","R","Go","Scala","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-cybersecurity-startup-is-using-machine-learning-to-enhance-threat-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10050093,"title":"How Would A Zombie Outbreak Look Like? Someone Just Created A Simulation Using Julia","content":"Physicist turned developers George Datseris and Timothy DuBois, have modelled a zombie outbreak on a real map using Agents.jl with OpenStreetsMaps API. The agent-based model (ABM) simulation represents how it might turn out to be in real life if a zombie outbreak occurs. An agent-based (or individual-based) model is a computational simulation of autonomous agents that react to their environment (including other agents), given a predefined set of rules. The simulation shows a zombie outbreak in a city, where an agent satisfies the OSMSpace conditions of having a tuple position. The model constructor consists of a map and 100 agents scattered randomly around it, each having their own agenda and need to travel to some new destination. The ones from the population that turn into zombies will begin infecting anyone who comes close. The zombies are seemingly oblivious to their state since they keep going about their business but start eating people along the way. Agents.jl is a free, open-source and extremely transparent Julia framework for agent-based modelling (ABM), whose modular, function-based design and support for many types of space (arbitrary graphs, regular grids, continuous space, and instances of Open Street Map) make it popular among developers. It also provides multi-agent support for interactions between disparate agent species and agent distributions on regular grids or continuous space. OpenStreetMapXPlot has been used for seamlessly plotting the space but is still a work in progress.","excerpt":"The agent-based model (ABM) simulation represents how it might turn out to be in real life if a zombie outbreak occurs.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Agents","APIs","Data Science","Deep Learning","Julia Language","Machine Learning","reinforcement learning environment","Reinforcement Learning Systems"],"author_name":"Victor Dey","publish_date":"2021-09-30T19:46:39","publication_year":"2021","word_count":233,"keywords":["Reinforcement Learning Systems","Go","API","autonomous agents","programming_languages:R","ML","AI Agents","Machine Learning","programming_languages:Go","APIs","Julia Language","Julia","reinforcement learning environment","Deep Learning","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["ML","autonomous agents","R","Go","Julia","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/how-would-a-zombie-outbreak-look-like-someone-just-created-a-simulation-using-julia\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10138370,"title":"Lenovo Announced as FIFA Technology Partner for 2026 and 2027 World Cups","content":"Lenovo has been named the Official Technology Partner for FIFA, covering both the FIFA World Cup 2026 and the FIFA Women’s World Cup 2027. The company announced this move at the Lenovo’s Tech World innovation event held at Bellevue, Washington on October 15. This partnership places Lenovo in FIFA’s top-tier sponsorship category. Lenovo will integrate its AI innovations, devices, and data center infrastructure to enhance fan engagement and help FIFA achieve its goal of growing the sport globally. I'm excited to announce that Lenovo has become the Official Technology Partner of FIFA, including both the FIFA World Cup 26™ and FIFA Women’s World Cup 2027™. As a leader in innovation, we are proud to join forces with the world’s most popular sport to power the largest global… pic.twitter.com\/7YwQsrmfrK— Yuanqing Yang (@YuanqingYang) October 16, 2024 The 2026 FIFA World Cup is set to be hosted in Canada, Mexico, and the United States. And the 2027 FIFA Women’s World Cup is to be hosted in Brazil. The events will feature Lenovo’s products such as ThinkPad laptops, Motorola smartphones, and servers to power real-time analytics, stadium experiences, and global data accessibility. “Lenovo is proud to support FIFA’s vision of leveraging technology to elevate the game, enhance the fan experience worldwide, and foster innovation that levels the playing field,” said Yuanqing Yang, Lenovo chairman and CEO. “We’re excited that our cutting-edge technology and AI innovation will take center stage in the upcoming tournaments.” FIFA President Gianni Infantino echoed this sentiment, highlighting the importance of technology in transforming fan experiences. “We are excited to welcome Lenovo to our journey and work with them to implement technologies, innovations, and programs that spread our sport. Data and technology combined helps us to know fans better, and we will use it to create unparalleled and unforgettable fan experiences.” This partnership marks Lenovo’s first collaboration with FIFA and aligns with Lenovo’s history of supporting flagship global sporting events. The FIFA World Cup 2026 will feature 48 teams and be hosted by three nations for the first time, while the 2027 FIFA Women’s World Cup will take place in Brazil, marking the first time the event will be held in South America. At Cypher 2024, former Indian football captain Bhaichung Bhutia advocated merging football and technology. “In Indian football, we are yet to use the technology [AI and analytics] that is used in a much bigger way in world football,” he said. “I was just reading, not sure if it’s true, that Liverpool signed Mohamed Salah because of data,” recalled Bhutia, pondering over the existence of technology in football and how it’s now a necessity in Indian football.","excerpt":"The FIFA World Cup 2026 will feature 48 teams and be hosted by three nations for the first time, while the 2027 FIFA Women’s World Cup will take place in Brazil, marking the first time the event will be held in South America.","categories":["AI News"],"tags":["AI in World Cup","FIFA Technology","Lenovo"],"author_name":"Mohit Pandey","publish_date":"2024-10-16T10:24:36","publication_year":"2024","word_count":438,"keywords":["AI in World Cup","Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","Lenovo","analytics","real-time analytics","R","FIFA Technology"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","real-time analytics","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/lenovo-announced-as-fifa-technology-partner-for-2026-and-2027-world-cups\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10132387,"title":"Is Runway’s Gen-3 Update the First or Last Frame?","content":"Runway, the US-based AI startup, has taken another significant step in the rapidly evolving field of AI-generated video. The company announced today that its Gen-3 Alpha Image to Video tool now supports using an image as either the first or last frame of video generation, a feature that could dramatically improve creative control for filmmakers, marketers, and content creators. Gen-3 Alpha Image to Video now supports using an image as either the first or last frame of your video generation. This feature can be used on its own or combined with a text prompt for additional guidance.All examples below demonstrate using an image as the last frame.(1\/5) pic.twitter.com\/koFBP7iKNf— Runway (@runwayml) August 5, 2024 The startup was founded in 2018 by Cristóbal Valenzuela, Alejandro Matamala, and Anastasis Germanidis. Furthermore, this update comes after the startup officially released Gen-3 Alpha, highlighting the company’s aggressive push to stay ahead in the competitive AI video generation market. The new capabilities of the model allows users to anchor their AI-generated videos with specific imagery, potentially solving one of the key challenges in AI video creation; consistency and predictability. By allowing users to generate high-quality, ultra-realistic scenes that are up to 10 seconds long—with various camera movements—using only text prompts, still imagery, or pre-recorded footage, this model has set a new benchmark in video creation. “The ability to create unusual transitions has been one of the most fun and surprising ways we’ve been using Gen-3 Alpha internally,” said Runway co-founder and CTO Anastasis Germanidis. Go make art https:\/\/t.co\/BRraBDEv0g— Cristóbal Valenzuela (@c_valenzuelab) July 1, 2024 Back in February 2023, Runway released Gen-1 and Gen-2, the first commercial and publicly available foundational video-to-video and text-to-video generation models accessible via an easy-to-use website. Now the Gen-3 update takes it to the next level. The Power of First and Last Frames “Gen-3 Alpha update now supports using an image as either the first or last frame of your video generation. This feature can be used on its own or combined with a text prompt for additional guidance,” Runway announced on X. The impact of this feature was immediately recognised by users. Justin Ryan, a digital artist, posted in response: “This is such a big deal! I’m hoping this means we are closer to the First and final frame like Luma Labs offers.” This development puts Runway in direct competition with other players in the space, such as Luma Labs, Pika, OpenAI’s much-anticipated Sora, and the Bengaluru-based startup Unscript, which is generating videos using single images. However, Runway’s public availability gives it a significant edge over Sora, which remains in closed testing. A spokesperson from Runway shared that the initial rollout will support 5 and 10-second video generations, with significantly faster processing times. Specifically, a 5-second clip will take 45 seconds to generate, while a 10-second clip will take 90 seconds. Accelerating to Get Ahead Since the release of the Gen-3 Alpha model, internet users have been showcasing their unique creations in high-definition videos, demonstrating the versatility and range of Runway AI’s latest AI model. As Runway makes a bold move, there is a significant shift in the generative AI video space. The company describes this update as “first in a series of models developed by Runway on a new infrastructure designed for large-scale multimodal training,” and a “step toward creating General World Models.” Germanidis also revealed that Gen-3 Alpha will soon enhance all existing Runway modes and introduce new features with its advanced base model. Runway Gen-3 Alpha will soon be available in the Runway product, and will power all the existing modes that you’re used to (text-to-video, image-to-video, video-to-video), and some new ones that only are only now possible with a more capable base model.— Anastasis Germanidis (@agermanidis) June 17, 2024 He also noted that since Gen-2’s 2023 release, Runway has found that video diffusion models still have significant performance potential and create powerful visual representations. Since we released Gen-2 last year, we learned a lot. We learned that multi-modal artistic control is key, that video diffusion models are nowhere close to saturating performance gains from scaling, and that those models, in learning the task of predicting video, build really…— Anastasis Germanidis (@agermanidis) June 17, 2024 While the startup states that Gen-3 Alpha was “trained on new infrastructure” and developed collaboratively by a team of researchers, engineers, and artists, it has not disclosed specific datasets, following the trend of other leading AI media generators that keep details about data sources and licensing confidential. Interestingly, the company also notes that it has already been “collaborating and partnering with leading entertainment and media organisations to create custom versions of Gen-3,” which “allows for more stylistically controlled and consistent characters, and targets specific artistic and narrative requirements, among other features.” AI comes to filmmaking Additionally, Runway hosted its second annual AI Film Festival in Los Angeles. To illustrate the event’s growth since its inaugural year, Valenzuela noted that while 300 videos were submitted for consideration last year, this year they sent in 3,000. Hundreds of filmmakers, tech enthusiasts, artists, venture capitalists, and notable figures, including Poker Face star Natasha Lyonne, gathered to watch the 10 finalists selected by the festival’s judges. Now, the films look different, as does the industry with generative AI. Meanwhile, amidst all this, it is evident that Runway is not giving up the fight to be a dominant player or leader in the rapidly advancing generative AI video creation space.","excerpt":"Runway’s new update brings it in direct competition with other players in the space, such as Luma Labs, Pika, OpenAI’s much-anticipated Sora.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Runway","Sora"],"author_name":"Tarunya S","publish_date":"2024-08-13T11:26:47","publication_year":"2024","word_count":902,"keywords":["Go","API","OpenAI","AI","ML","Git","diffusion models","Runway","generative AI","Sora","CLIP","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","R","Go","Git","API","CLIP","diffusion models"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/is-runways-gen-3-update-the-first-or-last-frame\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042549,"title":"PaperswithCode: Top 10 ML Papers Codes in 2024","content":"Papers With Code is the go-to resource for the latest SOTA ML papers, code, results for discovery and comparison. The platform consists of 4,995 benchmarks, 2,305 tasks, and 49,190 papers with code. Besides Papers With Code, other notable machine learning research papers’ resources and tools include arXiv Sanity, 42 Papers, Crossminds, Connected Papers etc. Papers With Code is a self-contained team within Facebook AI Research. Its open-source, community-centric approach offers researchers access to papers, frameworks, datasets, libraries, models, benchmarks, etc. Here, we have rounded up the top 10 machine learning research papers on ‘Papers With Code.’ 1. TensorFlow: A system for large-scale machine learning TensorFlow is an ML system that operates at a large scale and in heterogeneous environments. It uses dataflow graphs to represent computation, shared state, and the operations that mutate that state. The machine learning system maps the nodes of a dataflow graph across many machines in a cluster and within a machine across multiple computational devices, including multicore CPUs, general-purpose GPUs, and custom-designed ASICs TPUs. The code is available on GitHub. 2. Adversarial Machine Learning at Scale Adversarial examples are malicious inputs designed to fool machine learning models. It transfers from one model to another, allowing attackers to mount black box attacks without knowing the target model’s parameters. It is the process of explicitly training a model on adversarial examples to make it more robust to attack or reduce its test error on clean inputs. 3. Scikit-learn: Machine Learning in Python Scikit-learn is a Python module integrating a wide range of SOTA machine learning algorithms for medium-scale ‘supervised’ and ‘unsupervised’ problems. It focuses on bringing machine learning to non-specialists using a general-purpose, high-level language. The source code and documentation are available on SciKit. 4. AutoML-Zero: Evolving Machine Learning Algorithms From Scratch AutoML has made significant progress in recent times. However, this progress has focused mainly on the architecture of neural networks, where it has relied on sophisticated expert-designed layers as building blocks. AutoML is expected to go further, where it can automatically discover complete machine learning algorithms just using basic mathematical operations as building blocks. 5. MXNet: A Flexible & Efficient Machine Learning Library for ‘Heterogeneous Distributed Systems MXNet is a multi-language ML library to ease the development of ML algorithms, especially for deep neural networks (DNNs). Embedded in the host language, it blends ‘declarative symbolic expression’ with imperative tensor computation. In addition, it offers auto differentiation to derive gradients. It is computation and memory efficient, and runs on various heterogeneous systems, ranging from mobile devices to distributed GPU clusters. 6. DeepFaceLab: A simple, flexible and extensible face-swapping framework It is an open-source deepfake system created by iperov for face swapping with more than 3K forks and 13,000 stars in GitHub. DeepFaceLab provides an easy-to-use pipeline for people with no comprehensive understanding of deep learning framework or model implementation, while remains a flexible and loose coupling structure for people who need to strengthen their own pipeline with other features without writing complicated code. More than 95% of deepfake videos are created with DeepFaceLab. The code is available on GitHub. 7. Politeness Transfer: A Tag and Generate Approach In this paper, you can convert non-polite sentences to polite sentences while preserving the meaning. It provides a dataset of more than 1.39 instances automatically labeled for politeness to encourage benchmark evaluations on this new task. For politeness and five other transfer tasks, its model outperforms the SOTA methods on automatic metrics for content preservation, with a comparable or better performance on style transfer accuracy. Additionally, the model surpasses existing methods on human evaluations for grammaticality, meaning preservation and transfer accuracy across all the six style transfer tasks. The data and code are available on GitHub. 8. Caffe: Convolutional Architecture for Fast Feature Embedding Caffe provides researchers with a clean and modifiable framework for SOTA deep learning algorithms and a collection of reference models. The framework is a BSD-licensed C++ library with MATLAB and Python bindings for training and deploying general-purpose CNNs and other deep models efficiently on commodity architectures. The source code is available on GitHub. 9. Well-Read Students Learn Better: ‘On the Importance of Pre-training Compact Models The paper shows pre-training is crucial to smaller architectures, and fine-tuning pre-trained compact models can be competitive to more elaborate methods proposed in concurrent work. The paper explores pre-trained models and transferring task knowledge from large fine-tuned models through standard knowledge distillation. As a result, the general algorithm, along with pre-trained distillation, brings improvements. 10. XGBoost: A Scalable Tree Boosting System The paper describes a scalable end-to-end tree boosting system called XGBoost, used widely by data scientists to achieve SOTA results on many machine learning challenges. The source code is available on GitHub. Popular Posts Top Object Detection Algorithms Top Chart GPT Alternatives Top Ethical Hacking Courses Top AI Powered Tools for Stock Market Trading Top Library in CC for Machine Learning","excerpt":"Papers With Code is a self-contained team within Facebook AI Research","categories":["AI Trends"],"tags":["AI Research"],"author_name":"Amit Naik","publish_date":"2021-06-30T10:00:00","publication_year":"2021","word_count":812,"keywords":["scikit-learn","machine learning","AI","neural network","ML","AI Research","RAG","deep learning","XGBoost","object detection","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","TensorFlow","scikit-learn","XGBoost","RAG","object detection"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-ml-papers-on-papers-with-code\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10023435,"title":"MIT Researchers Develop An AI Tool To Detect Skin Cancer","content":"In recent research, MIT scientists have come up with a unique AI tool to help in the early detection of skin cancer. Traditionally, physicians had to examine suspicious pigmented lesions to identify any hint of skin cancer, which was not only time-consuming but also proved to be inaccurate, preventing the chance of early treatment. The US, in 2019, reported having diagnosed approximately 96,480 people with melanoma, which led to 7230 deaths. However, with this new research, scientists are claiming to solve the early detection issue of skin cancer with the help of artificial intelligence. The scientists at MIT have developed an AI-powered SPL (suspicious pigmented lesions) analysis system to accurately assess the pigmented lesion on patients’ skin to detect the anomalies involved. The Research To facilitate the research, scientists leveraged wide-field images of patients’ skin, taken using any smartphone camera, and fed into deep convolutional neural networks (DCNNs) to analyse the suspicious pigmented lesions. DCNNs can cluster similar parts of images based on the category thereby used for image classification. It then uses deep learning algorithms to further facilitate the process. The SPL system has been trained on “38,283 dermatological datasets collected from 133 patients and publicly available images” to detect and extract the pigmented lesions found in that image. Yellow: consider further inspection; Red: requires further inspection or referral to a dermatologist. According to the scientists, the result of the AI-powered system is displayed in the form of a heatmap, where the resultant yellow marks are selected for further inspection, and the red ones are referred to the dermatologist. Instead of the traditional process of evaluating every single lesion separately to look for the signs of neoplasia, the AI-powered SPL system identifies all the lesions and marks on the patient’s skin, flagging them in order of suspiciousness. This process is a breakthrough for screening early-stage melanoma, a disease where the cells that produce pigment in the body are attacked by cancer. If it is not treated early, it can spread to the internal organs and eventually lead to death. To evaluate the AI model, MIT scientists have worked with dermatologists to visually classify the lesions and compare the results with those produced by the system. It was noticed that the AI system achieved over 90.3% accuracy in distinguishing suspicious lesions from the non-suspicious ones without the cumbersome individual lesion imaging. Wrapping Up Luis R. Soenksen, the man behind this research, is an expert in medical devices and a vocal advocate of using AI to solve real-world problems. He firmly believes that early detection of suspicious pigmented lesions can help doctors save many lives lost due to skin cancer. He claims that this research is suggestive of the fact that with the assistance of computer vision along with deep neural networks, AI can achieve the level of accuracy in detecting melanoma that can easily be compared with expert dermatologists. The research further claims that the new process can extract the intra-patient saliency of lesions that can compare the lesions on a patient’s skin with others. With this research, Soenksen hopes to achieve more efficient screenings of lesions within a primary care visit, which will not only improve patient triaging but will also manage the utilisation of resources in hospitals. Read the paper here.","excerpt":"Now, skin cancer can be easily treated at an early stage with the help of MIT researchers’ AI system.","categories":["AI Features"],"tags":["cancer","MIT","MIT research","MIT researchers"],"author_name":"Satavisa Pati","publish_date":"2021-04-06T14:24:00","publication_year":"2021","word_count":543,"keywords":["Go","artificial intelligence","MIT researchers","AI","neural network","MIT","ML","computer vision","RAG","Aim","deep learning","cancer","MIT research","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","neural network","computer vision","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mit-researchers-develop-an-ai-tool-to-detect-skin-cancer\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10084560,"title":"A.R. Rahman Launches Music Metaverse Platform ‘Katraar’","content":"Oscar-winning Indian music director and composer A.R. Rahman entered Web 3.0 on his 56th birthday with the launch of ‘Katraar‘, a metaverse platform for other musicians. Besides featuring his work, Katraar will also showcase musicians across the globe. The immersive platform, which is in the development stage, will enable musicians to upload their work and mint money off it, making it a long-term sustainable revenue model for upcoming talents. Non-profit organisation HBAR is partnering to build the project, which will be available on its Hedera Network. Katraar will combine music, art and storytelling with metaverse features, including NFTs. In Tamil, the word Katraar means “a group of learned people who can change the world”, said the ‘Roja’ music composer. The evolving music metaverse Several celebrated artists in the West have dabbled in the music metaverse. To create NFTs from her concerts, Katy Perry partnered with Theta Labs, a blockchain video streaming company. Nowadays, metaverse concerts are also quite popular, particularly in the wake of the pandemic. Justin Bieber, Travis Scott, and the Astroworld team are among the musicians playing on Decentraland, Sandbox, and other virtual worlds that put up an amazing performance on the game Fortnight. In the world of digital art and NFTs, music videos also have the opportunity to be a lucrative tender. Canadian musician Grimes, for instance, sold original video NFTs for a staggering $6 million. Back home, earlier in April 2022, Tamil singer Karthik announced music NFTs on India’s first NFT marketplace Jupiter Meta. Notably, the metaverse is changing the music landscape, with a myriad of opportunities for new, independent musicians, from filling in the gaps in streaming services to NFT bands. The metaverse is surely the internet’s future.","excerpt":"The immersive digital platform Katraar empowers emerging musicians and allows them to monetise their works.","categories":["AI News"],"tags":["Metaverse","music","Wikipedia"],"author_name":"Shritama Saha","publish_date":"2023-01-10T12:10:22","publication_year":"2023","word_count":283,"keywords":["programming_languages:R","AI","Metaverse","Git","music","Wikipedia","GAN","R"],"extracted_tech_keywords":["AI","R","Git","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/a-r-rahman-launches-music-metaverse-platform-katraar\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10097686,"title":"6 Brilliant Video Resources on Generative AI by Andrej Karpathy","content":"Former AI director at Tesla Andrej Karpathy returned to OpenAI pretty recently. He came to fame for his immense contribution working alongside CEO Elon Musk to create Optimus, a groundbreaking humanoid robot. Karpathy also played a pivotal role as the head of Tesla Autopilot’s computer vision team. He released NanoGPT, a fast repository for training and tuning medium-sized GPTs, building upon his earlier work with miniGPT for GPT language models. His latest project is baby Llama which he made by tuning NanoGPT to use the Llama 2 architecture instead of GPT-2. Apart from his big contributions to generative AI, the computer vision genius has been a huge contributor to the open-source community through his mini projects, educational resources, coding tutorials on YouTube and more. He is also known for creating courses on building deep neural networks, including NanoGPT, based on GPT-2\/GPT-3 and the ‘Attention is All You Need’ paper. Here are some free important resources for you. Let’s build GPT from Scratch In this two hour long YouTube video, Karpathy takes you on a journey to build a GPT model, based on Google’s research paper “Attention is All You Need” and OpenAI’s GPT-2 and GPT-3. To help the audience grasp the concepts better, he suggests watching earlier videos, which cover autoregressive language modelling framework and the fundamentals of tensors and PyTorch nn, essential knowledge they assume viewers already possess in the current video. The video is a great resource for anyone who wants to learn more about how GPT works or how to build their own GPT model. It is also a good introduction to the attention mechanism, which is a powerful tool for natural language processing. State of GPT If you want to learn more about the training process of GPT assistants like ChatGPT, this video is most suitable for you. It covers tokenization, pretraining, supervised finetuning, and Reinforcement Learning from Human Feedback (RLHF). Additionally, you will also get to know about practical approaches and conceptual frameworks for utilising these models effectively. This includes prompting strategies, finetuning techniques, the ever-expanding toolkit available, and potential future advancements in this field. Intro to Neural Networks and Backpropagation: Building Micrograd One of his most admired videos of all time, in this comprehensive guide to backpropagation and neural network training, Karpathy presents a highly detailed and easily understandable explanation. The tutorial assumes minimal prerequisites, needing only a fundamental understanding of Python and basics of high school-level calculus. By breaking down complex concepts into step-by-step instructions, Karpathy ensures that you can understand the complexities of the subject without feeling overwhelmed. The Spelled-Out Intro to Language Modelling: Building Makemore By developing a bigram character-level language model as a starting point, Karpathy later advanced it into a contemporary Transformer language model similar to GPT. The main objectives of this particular video are to introduce the audience to torch.Tensor and its nuances, demonstrating its significance in the effective evaluation of neural networks; secondly, to provide an overview of the language modelling framework encompassing tasks such as model training, sampling, and evaluating loss measures like the negative log likelihood utilised in classification tasks. He has explained the process through five detailed videos. Building Makemore: Activations & Gradients, BatchNorm This video teaches you the working of internals of Multi-Layer Perceptrons (MLPs) encompassing multiple layers, primarily revolving around the analysis of the results of improper scaling. Moreover, the study focuses on the diagnostic tools and visualisations, crucial for understanding how complex neural networks work. You will also learn about the fragility of training deep neural networks and discover the revolutionary technique known as Batch Normalisation, which greatly simplifies the process. Building a WaveNet By taking a 2-layer MLP (Multi-Layer Perceptron), Karpathy shows you how to turn it into a deeper neural network using a tree-like structure, similar to DeepMind’s WaveNet (2016) architecture. The WaveNet paper implements a more efficient version of this hierarchical structure using causal dilated convolutions, which are not yet covered in the video. Throughout the process, viewers gain a better understanding of torch.nn, how it works behind the scenes, and what a typical deep learning development process involves—like reading documentation, keeping track of tensor shapes, and switching between Jupyter notebooks and repository code.","excerpt":"The videos are very detailed and take you through the step-by-step process of creating different generative AI applications","categories":["AI Trends"],"tags":["Andrej Karpathy"],"author_name":"Shritama Saha","publish_date":"2023-07-27T17:57:03","publication_year":"2023","word_count":696,"keywords":["ChatGPT","Andrej Karpathy","OpenAI","AI","neural network","PyTorch","ML","computer vision","deep learning","generative AI","Jupyter"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","computer vision","generative AI","ChatGPT","OpenAI","PyTorch","Jupyter"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/6-brilliant-video-resources-on-generative-ai-by-andrej-karpathy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10119783,"title":"Apple&#8217;s ‘Big AI Plans’ Coming Soon","content":"Over the last decade, Apple has spent around $700 billion on stock buybacks, and with this money, the Cupertino tech giant could have bought Tesla, Rivian, Lucid, OpenAI and Anthropic and become a dominant force in both EVs and AI. Most recently, Apple unveiled a $110 billion share buyback programme after its Q2 profit and revenue dropped. As a result, its stock surged 12% in after-hours trading as CEO Tim Cook predicted sales growth with upcoming AI-driven features. The buyback initiative rewards its investors and aligns Apple with other tech giants amid concerns over rising generative AI investment. It also enables Apple to manage its capital structure effectively, returning capital to shareholders without ongoing dividend commitments, especially amid economic uncertainty. Additionally, this buyback signals to the market that Apple’s management perceives its stock as undervalued, potentially boosting investor confidence and attracting more buyers. Apple’s EV ambitions come to a halt After a decade of effort, Apple reportedly ended its ambitious Project Titan, halting the development of its electric car in favour of shifting focus to generative AI initiatives. Lately, Apple has been investing heavily in expanding its AI capabilities. This focus is evident in its development of proprietary AI models like the OpenELM model, designed to run on Apple devices. The model is particularly notable for optimising deep neural network layers to enhance efficiency. Apple’s use of a neural net with just 1.3 billion parameters stands in contrast to models like GPT-4 and Google’s Gemini. Apple GenAI Dreams At the latest earnings call, Cook said, the recent quarter was thrilling, as we launched Apple Vision Pro to show the world the potential that spatial computing unlocks. “This is just the beginning, we’re also making exciting product announcements and sharing more about the vision for generative AI in coming weeks and months,” he added. Cook is all set to unveil significant upgrades at the iPad Let Loose event on May 7 and tease new AI capabilities. Apple has been making waves lately with its focus on integrating generative AI capabilities into its devices. In the past few months, it has introduced its latest creations, the MM1 models—a new series of multimodal AI models with an impressive 30 billion parameters, alongside ReALM, which integrates text and images to enhance its understanding and responsiveness to prompts. Apple also introduced ‘Ferret-UI,’ a multimodal AI model designed to execute precise tasks concerning user interface screens while interpreting and acting upon open-ended language instructions. This advancement suggests a future where verbal commands may seamlessly supplant traditional finger gestures for iPhone navigation. In addition, another research paper unveils Keyframer, a tool purportedly capable of generating animations from static images, alongside an AI model tailored for image editing. Siri will be powered by Gen AI At the same time, Apple has been in talks with various tech giants, including OpenAI and Google, to integrate generative AI into its upcoming iPhone series. Recently, Apple has rekindled discussions with OpenAI to integrate AI functionalities into iOS 18 (iPhone 16). Additionally, in 2023, reports indicated that Apple invested significantly in research and development to improve Siri’s conversational skills. Integrating GPT-like technology into Apple’s infrastructure would essentially improve Siri. The primary capabilities of a GenAI-driven Siri will probably stem from Apple’s in-house models operating on the device, while supplementary functionalities such as generating images and crafting long-form text may be sourced from third-party entities like OpenAI, Baidu, or Google. Furthermore, Apple is also experimenting with an internal chatbot. The upgraded Siri boasts natural conversation abilities and enhanced user personalisation. According to Mark Gurman, a significant change will be removing ‘Hey’ from ‘Hey Siri.’ What’s next? At last year’s WWDC,  Tim Cook and other Apple executives did not mention the term ‘artificial intelligence’ even once. But, at the recent earnings call, everything changed — Cook said that he is optimistic about its “opportunities in generative AI” and is “making significant investments,” hinting at ‘big AI plans’ coming at the WWDC, next month.","excerpt":"“This is just the beginning, we’re also making exciting product announcements and sharing more about the vision for generative AI in coming weeks and months,” says Apple chief Tim Cook","categories":["AI Features"],"tags":["Apple"],"author_name":"Gopika Raj","publish_date":"2024-05-07T16:57:14","publication_year":"2024","word_count":659,"keywords":["Anthropic","GenAI","artificial intelligence","OpenAI","AI","neural network","Apple","ML","generative AI","multimodal AI","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","generative AI","GenAI","multimodal AI","OpenAI","Anthropic","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/apples-big-ai-plans-coming-soon\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10097640,"title":"LMQL: The Cure for LLM Chatbot Hallucination?","content":"Large language models don’t always respond correctly to your questions. They are difficult to control because the user cannot fully understand what goes on inside them. Recently, there have been a lot of complaints about LLM chatbots hallucinating, giving unsatisfactory responses, and a good option to fix them is to improve prompting techniques. Language Model Query Language (LMQL) solves this issue by combining language prompting with simple scripting. Researchers from ETH Zurich wrote a paper titled ‘Prompting Is Programming: A Query Language for Large Language Models’ — on the emerging discipline of clever prompting, which is an intuitive combination of natural language and programming language prompts. Users can specify constraints on a language model’s output, and get it to perform multiple tasks at the same time by providing high-level semantics. How does it work? LMQL is a declarative programming language, which means the language states only what the end result of the task is and abstracts the control flow of logic required for the software to perform the action. It is inspired by SQL but integrates Python into its framework. Users can ask the model prompts that contain both text and code. The language grammar, according to the paper, has five essential parts. The decoder, as the name suggests, decodes the algorithm that generates the text. It is a string of code which transforms the output into meaningful results, improving the quality and diversity of words. The Query block written in Python syntax serves as the core interaction mechanism with the language model. Each top-level string within the query block is a direct query to the language model. The Model\/from clause specifies the model being queried. This defines the underlying language model used for text generation and Where Clause on the other hand allows users to define the constraints that influence the generated output. It defines the output required by the language model to stick to the desired qualities. And finally, Distribution Instructions, which is an optional instruction, guides the distribution of generated values. It defines how the generated results should be distributed and presented, enabling the user to control the outcome’s format and structure. Control the interaction For simple queries, users can guide the language model using natural language, but when the tasks increase in complexity and when the user requires responses to specific questions, it is better to have full control of the query. If you’re tech savvy, even for simple tasks like asking the model to tell you a joke, you can be in full control of the result you wish to get. LMQL offers a dedicated Playground IDE to make query development easier. Users can examine the interpreter’s status, validation outcomes, and model responses at any stage of text generation. This comprises the capability to analyze and explore various hypotheses produced during beam search, providing useful insights to refine the language model’s behavior. Efficiency and performance are a big challenge according to the paper. Despite being more efficient, the inference step in modern Language Models rely on costly, high-end GPUs to achieve satisfactory performance. With LMQL, the generation of text closely aligned with desired output becomes possible in the first attempt, eliminating the need for subsequent iterations. The evaluations show that LMQL improves accuracy in various tasks while significantly reducing the computational costs in pay-to-use APIs. This translates to an impressive cost savings ranging from 13% to 85%. One of the authors of LMQL said on HackerNews, “Cost is definitely a dimension we are considering (research has limited funding after all), especially with the OpenAI API. Lock-step token-level control is difficult to implement with the very limited OpenAI API. As a solution to this, we implement speculative execution, allowing us to lazily validate constraints against the generated output, while still failing early if necessary. This means, we don’t re-query the API for each token (very expensive), but rather can do it in segments of continuous token streams, and backtrack where necessary.” Language Model Programming This isn’t the first hybrid approach to prompt engineering. Jargon, SudoLang, and prlang all do something similar. “LLMs+PLs is a very interesting field right now, with lots of directions to explore,” said another author of LMQL. They offer users the ability to express both common and advanced prompting techniques in a simple and concise manner. But if you can use any programming language on LLMs, why learn a specific query language like LMQL? LMQL gives you a concise way to define multi-part prompts and enforce constraints on LLMs. For instance, you can make sure the model always adheres to a specific output format, where parsing of the output is automatically taken care of. Also abstracts a number of things like APIs and local models, tokenisation, optimisation and makes tool integration (e.g. tool function calls during LLM reasoning) much easier. This is also language model agnostic, improving portability and can be used across LLMs. Language Model Programming (LMP) makes it easier to adapt language models for different tasks while abstracting the model’s internals and providing high-level semantics. LMQL represents a promising development, as evidenced by its ability to enhance the efficiency and accuracy of language model programming. It empowers users to achieve their desired results with fewer resources, making text generation more accessible and efficient.","excerpt":"LMQL takes a hybrid approach to programming and combines natural language prompts with programming language for accurate responses from language models","categories":["AI Highlights"],"tags":["ai hallucinations","prompt engineering","researchers"],"author_name":"K L Krithika","publish_date":"2023-07-27T11:13:41","publication_year":"2023","word_count":872,"keywords":["ai hallucinations","Go","API","TPU","OpenAI","AI","chatbots","researchers","Python","prompt engineering","SQL","R"],"extracted_tech_keywords":["AI","OpenAI","prompt engineering","chatbots","TPU","Python","R","SQL","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/lmql-the-cure-for-llm-chatbot-hallucination\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10061169,"title":"We need social scientists, lawyers, ethicists to work with data scientists to build responsible AI applications: Arunima Sarkar, Lead AI, WEF","content":"We all know that tech is largely dominated by men, and sub-domains of tech like data science, AI and analytics follow the same trend. Though in recent years, we have seen more and more young women showing interest in making a career in these new-age technologies, the number is still quite low. To encourage more women to pursue these so-called “non-traditional job roles”, we bring out stories of women breaking the stereotypes and making successful careers in tech. Today, we look at the journey of Arunima Sarkar. What makes her journey unique is that she has seen a mix of both worlds—spending around two decades in corporate with 13 years in Accenture holding positions like Global Applied Intelligence Research Lead, Growth and Strategy, to policymaking and technology governance currently at World Economic Forum as the Lead Artificial Intelligence, Global Assignments, Centre for Fourth Industrial Revolution. From technology research to leadership in Accenture Sarkar adds, “I started my career from a technology research perspective. I was interested in understanding the impact of technology on the economy, society and the different opportunities it creates for the different industries and regions. I started off with roles in such research consulting firms focusing on technology like Gartner. At that time, the Indian IT outsourcing industry was just coming up, and I was deeply involved in developing India’s first BPO study that was done back in those days with Gartner. I was closely working in the IT, BPO and telecom market.” She continued with her interest in broad IT services and then joined Accenture back in 2006. She has spent over a decade there where she built the analytics, AI and data research team within the company and was leading that globally for several years. Strong advocate of responsible AI Sarkar says, “Around six years back, the company also started looking at responsible AI. It was an important aspect, and Accenture wanted to make it an integral part of everything they did in artificial intelligence at the organisation. At that time, I was representing Accenture at PAI (Partnership on AI), a private-public partnership based in the US, working specifically in the area of responsible AI. Sarkar also started looking at the societal impact of such technologies and their policy implications. She spent time working on areas not only around applied AI but also responsible AI. She wanted to answer this critical question—how can this technology be built, designed and deployed in a responsible and ethical manner? Then she moved out of the private sector and joined the World Economic Forum. Impact of tech on society Sarkar leads the codesign of governance protocol and technology policy frameworks for artificial intelligence as well as quantum computing governance frameworks for developing principles for responsible innovation and use of the technology. She also leads several sectoral projects that have multiple stakeholders and are global in nature. WEF brings in experts from the public sector, private sector, academia, research and civil society together to understand what are the pressing societal challenges, what are the policy and technology governance gaps and where technology can be  applied for achieving larger societal benefits. Developing India’s responsible AI strategy with NITI Aayog Sarkar has had a long and illustrious career with several milestones. But when asked to identify her biggest achievements if she could look back, she has two very special projects in mind. Sarkar led the roadmap from WEF’s side along with NITI Aayog to develop India’s Responsible AI strategy. It puts together a national-level roadmap on how AI can be designed, developed and deployed responsibly in the country. The roadmap helps understand and create awareness about the ethical challenges and potential risks of this technology, what kind of considerations need to be kept in mind by different stakeholders and how to implement such guidelines in the country. She adds, “The potential of the technology is known and talked about and quantified, but the responsible use of this technology is critical. It is very important to build that awareness in the country. India is representative of a huge part of the work population, and if such technologies are implemented properly, it can create best practices for the entire world to follow.” First set of quantum computing guidelines Another big achievement for Sarkar has been leading the governance track of the quantum computing initiative at WEF. Last year was a year of intense hard work for Sarkar and her team to bring out the first set of quantum computing governance principles for the responsible development of this technology. Sarkar states, “We see investments coming from the private sector, venture capital funding in quantum technology and national level programs being rolled out. We are at the stage when quantum computing in the lab is in the PoC stage, just before commercialisation starts happening in this tech. It is the right time for us to look at the criteria that are needed to develop this technology.” Good mix of business and tech skills As someone who has hired many AI professionals and worked with data scientists in various leadership roles, Sarkar feels that in order to build AI that is impactful and beneficial for all, it is very important for a data scientist to understand the business needs and the use case the AI application is targeting. A data scientist should have a good mix of tech and business skills and a problem-solving mindset. More success stories need to be out in the public Sarkar feels that this trend of lesser women in the AI and analytics space is a continuation of the trend for the tech space. A stereotype exists, and there are not enough success stories that are talked about enough. She adds, “We need to build more awareness right from the educational level and within the organisation as well. Sharing success stories and mentoring women to build up their career in this space can be greatly helpful in bridging the gender gap.” We need social scientists, lawyers, ethicists to work with data scientists to build responsible AI applications Sarkar gives an interesting perspective on the various aspects of AI that one can build a career in. She says, “AI needs multidisciplinary experts-data scientists, social scientists, lawyers, ethicists to build responsible AI applications that can be deployed. Most large organisations are recognising the needs, and many startups are understanding this too. There are multiple ways in which one can approach this field with allied career paths around AI. Sarkar says that it is time to embed data science and AI curriculum in schools and colleges. Children should be made aware that such a career path is available to them. It does not have to be taught only to students pursuing Science but Arts as well. AI needs a combination of both. Man and machine will learn how to augment each other Sarkar concludes that in the near future, AI is going to become more and more pervasive in literally every aspect of our lives. We will see it impacting every industry, and there will be new ways in which man and machine will learn how to augment each other. The greater spread of the use of this tech will help in solving many problems in varying capacities—from critical problems like climate change to day-to-day issues.","excerpt":"AI needs multidisciplinary experts – data scientists, social scientists, lawyers, ethicists to build Responsible AI applications that can be really deployed.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Data Scientist","Interviews and Discussions","Machine Learning","Quantum Computing","Women in AI","World Economic Forum"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-02-21T16:00:00","publication_year":"2022","word_count":1211,"keywords":["Quantum Computing","Women in AI","data science","Go","artificial intelligence","World Economic Forum","API","AI","Machine Learning","RAG","responsible AI","analytics","GAN","Data Scientist","R","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","RAG","R","Go","API","GAN","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/we-need-social-scientists-lawyers-ethicists-to-work-with-data-scientists-to-build-responsible-ai-applications-arunima-sarkar-lead-artificial-intelligence-world-economic-forum\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10086580,"title":"Prediction is Prevention: This is How Bureau Tackles Cyberfraud","content":"Cyber attacks in India have tripled in the last three years, according to data released by CERT-in. Nearly 14,02,809 different cases of cyber security incidents were recorded in 2021 alone. “The growing instances of cyber fraud are alarming for businesses, especially when digital assets have seen exponential growth in recent years,” said Ranjan R Reddy, founder and CEO of Bureau, a trusted network that provides safety net to businesses solving cyber fraud risks with its AI\/ML-powered solution. Ranjan founded Bureau in 2020 to build a single source of digital trust to counter cyber fraud. In this exclusive interview, he discusses how AI and ML solutions can help tackle the problem of cyber fraud. AIM: Tell us a bit about ‘Bureau’. Ranjan: We are the first identity decisioning company that offers a complete, no-code identity and risk orchestration platform for fraud-free customer interactions for the entire customer journey. Our key differentiators include custom workflows for everything from onboarding, KYC\/KYB, and AML to transaction monitoring and a trusted network where each validated customer is added to a growing consortium of merchants’ data that further strengthens the Bureau’s database. Enabling trust between businesses and consumers is how we deliver value. We started our journey in 2020 and we have since powered several companies across sectors to design their digital user journeys in a secure manner, thereby reducing cyber fraud significantly. In the process of enabling these trusted financial transactions, we have protected over USD 200 million of commerce from use cases such as mule accounts, synthetic ID fraud, phishing, KYC fraud, fraud rings, UPI fraud, money laundering, collusion, OTP theft, and referral abuse—to name a few. More than 70 fintech, regulated financial institutions, and new-age technology businesses use Bureau today to ensure their customer interactions are friction- and fraud-free. AIM: How is ‘Bureau’ using AI\/ML to tackle cyber fraud risk? Could you share some use cases? Ranjan: Identity verification is the starting point of any cyber fraud detection, as securing one’s customers’ identity on the platform is the first step in making one’s platform secure. We use AI to assess the trustworthiness of customers by analysing data from various sources, such as phone and email intelligence. Machine learning is used to build models that can predict the likelihood of a customer committing fraud. We analyse large amounts of data using advanced algorithms to identify patterns and anomalies indicative of identity fraud. Another advantage of using machine learning algorithms is that it helps analyse data with greater accuracy than traditional techniques. Customer centricity is at the core of any business. One of the critical factors in achieving client satisfaction is by making users’ time hassle-free on platforms. This is where Bureau steps in with one-stop identity verification solutions that integrate multiple-point vendor solutions for capabilities to serve different use cases such as Know-Your-Customer, optical character recognition, AML and PEP, among others. Our platform comes pre-integrated with all these services, saving our customers’ capital, time, and engineering resources and helping them scale their businesses to brand new heights. AIM: Could you tell us about your models and the datasets they have been trained on? Ranjan: Our identity verification and fraud detection models have been trained on billions of data points unique and suitable to the Indian market and have already been validated across a network of clients. By combining pre-integrated device, persona, phone, email, and behaviour intelligence, a contextual view of a user’s identity and behaviour is formed. Performing an accurate identity verification and pre-empting fraud becomes easier, thereby safeguarding the business from financial and reputational loss. AIM: What role can AI play in cyber fraud\/crime prevention? Ranjan: AI and ML have the potential to revolutionise the way cyber fraud can be prevented. AI\/ML algorithms can detect and prevent malicious activities by analysing data from multiple sources. For example, AI can be used to detect suspicious activity in online payments by recognising patterns of fraudulent behaviour, along with detecting potential phishing attacks. These algorithms also automate the verification process, making it faster and more precise by analysing customers’ behaviours, such as online activity and purchases, to identify any suspicious or fraudulent activity. We can expect AI to play a significant role in cyber fraud\/crime prevention in the future as technology continues to evolve and improve. As AI evolves to more advanced stages and efficiency, and is trained on more data points, it will become better at detecting vulnerable sources and preventing cyber fraud and crime. Moreover, as organisations become more aware of its benefits, they are likely to adopt the technology to protect their systems and data, thus making it a vital tool in the fight against cyber fraud\/crime prevention. AIM: What are your views on the current cybersecurity landscape in India? Ranjan: India is one of the biggest marketplaces for digital technology that is rapidly expanding, as evident with the government’s ‘Digital India’ initiative and its vision to make the country a five-trillion-dollar trillion economy in the next few years. Currently, India has over 1.15 billion mobile users and over 700 million active internet users, creating a vast pool of digitally vulnerable targets for identity fraud. With businesses unlocking new levels of digital transformation, identity fraud threats have become more prominent than ever. The growing instances of cyber fraud are alarming for businesses, especially when digital assets have seen exponential growth in recent years. Data security and privacy, therefore, should be the top priority for organisations. The ability to protect users from fraud and cyber theft will be a significant competitive advantage and businesses need to invest in a wide spectrum of capabilities that—with a combination of data and technology—proactively thwarts fraud without having to choose between growth and risk.","excerpt":"AI can be used to detect suspicious activity in online payments by recognising patterns of fraudulent behaviour, along with detecting potential phishing attacks.","categories":["AI Features"],"tags":["Cybersecurity India","Interviews and Discussions"],"author_name":"Pritam Bordoloi","publish_date":"2023-02-06T11:46:53","publication_year":"2023","word_count":946,"keywords":["Go","API","machine learning","AI","R","ML","Git","Aim","Cybersecurity India","Rust","fraud detection","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","fraud detection","R","Go","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/prediction-is-prevention-this-is-how-bureau-tackles-cyberattacks\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10011881,"title":"Software 2.0: The Software That Writes Itself &#038; How Kotlin Is Ushering This New Wave","content":"“Neural networks represent the beginning of a fundamental shift in how we write software. They are Software 2.0.” Andrej Karpathy The current coding paradigms nudge developers to write code using restrictive machine learning libraries that can learn, or explicitly programmed to do a specific job. But, we are witnessing a tectonic shift towards automation even in the coding department. So far, code was used to automate jobs now there is a requirement for code that can write itself adapting to various jobs. This is software 2.0 where software writes on its own and thanks to machine learning; this is now a reality. Differentiable programming especially, believes the AI team at Facebook, is key to building tools that can help build ML tools. To enable this, the team has picked Kotlin language. Kotlin was developed by JetBrains and is popular with the Android developers. Its rise in popularity is a close second to Swift. Kotlin has many similarities with Python syntax, and it was designed as a substitute for Java. In the next section, we look at how Facebook is appending new capabilities to Kotlin so as to build next-generation software toolkits. Overview Of Software 2.0 (Source: Andrej Karpathy) Software 1.0 or the software as we know it comprises the usual languages such as Python, C++, etc. The programmer writes explicit instructions to the computer; instructions that tell the computer how to behave. In contrast, wrote Andrej Karpathy, the director of AI at Tesla Motors, in one his blogs, that Software 2.0 can be written in much more abstract, human-unfriendly language, such as the weights of a neural network. Machine learning fits perfectly into the software 2.0 transition that’s happening right now. For instance, computer vision applications, wrote Karpathy, require engineered features with a bit of machine learning, and now with the help of large datasets like ImageNet, researchers have developed techniques that search the space of neural network architectures itself, which is an active area of research currently. So, it is established that AI and automated software development cannot be separated. And, Facebook’s AI team in an effort to leverage this capability to build better ML programming tools bringing in Kotlin programming language into the mix. How Kotlin Is Being Used “Facebook AI is building an automatic differentiation system for the Kotlin language.” The team at FB, incorporating the concepts of differentiable programming to improve the capabilities of Kotlin, which in turn can be used to build new toolkits. In differentiable programming, said Facebook, library code could be incorporated into more comprehensive models, and it also allows developers to leverage gradients to automatically optimise parameterised programs that aren’t written using machine learning libraries. The intuitive nature of differentiable programming in Kotlin allows developers to create programs that are flexible and take advantage of the structure of the problem while keeping debugging simple. According to the team at Facebook AI, Kotlin enables Software 2.0 through: Automatic differentiation.Tensor typing.Generating compile-time errors for differentiable functions and tensor shapes.Making available a library that provides a Tensor class and machine learning APIs. “We’re extending the Kotlin compiler to make differentiability a first-class feature of the Kotlin language. Our work enables developers to explore Software 2.0, where software essentially writes itself.”Facebook AI Facebook announced that they are building an automatic differentiation system for the Kotlin language. Automatic differentiation refers to constructing a procedure for computing derivatives. An automatic differentiation system dissects any given program into its primitive operations(e.g., add, subtract etc.) to compute derivatives. (Source: FAIR) The Facebook team has also integrated Kotlin language with the IntelliJ IDE extensions so that developers can get real-time feedback. As shown above, a simple convolutional neural network written in IntelliJ where the developer can inspect the resulting shapes at each step. Very soon, Facebook will also be releasing a user library that packs the advantages of automatic differentiation and so that the developers can use these independent of the framework they are working on. Know more about this project here.","excerpt":"“Neural networks represent the beginning of a fundamental shift in how we write software. They are Software 2.0.” Andrej Karpathy The current coding paradigms nudge developers to write code using restrictive machine learning libraries that can learn, or explicitly programmed to do a specific job. But, we are witnessing a tectonic shift towards automation even […]","categories":["Deep Tech"],"tags":["Kotlin"],"author_name":"Ram Sagar","publish_date":"2020-11-19T16:00:07","publication_year":"2020","word_count":661,"keywords":["machine learning","AI","Kotlin","neural network","ML","computer vision","RAG","Python","Tecton","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","computer vision","Tecton","RAG","Python","R","Java"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/software-2-0-kotlin-programming-machine-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":52836,"title":"Why This Delhi-Based Startup Prides Itself As The McDonalds Of Geospatial World","content":"The synergistic integration of artificial intelligence (AI) and the geographic information dimension creates geospatial artificial intelligence (GeoAI). Geo-tagged big data collated from varied sources, such as satellite imagery via remote sensing, IoT sensors in smart cities, social media streaming, and personal sensing via connected ambient and wearable sensors, can be analysed using GeoAI to get actionable insights. Founded in 2017, Attentive AI was started by an IIT Delhi core team comprising of Shiva Dhawan, Utkarsh Sharma and Sarthak Vijay. The company was established to develop artificially intelligent systems that can analyse petabytes of geospatial imagery and convert it into accurate insights. It serves geospatial technology providers and end-users with 2D and 3D vector data extracted from satellite, aerial, street and drone imagery. The mission of Attentive AI is to convert all the data collected by satellites, aeroplanes and drones into actionable insights, and help businesses, governments and non-profit organizations in reaching meaningful conclusions and making better decisions faster. Every startup has its challenges and everyone should be prepared to tackle those, said Shiva Dhawan, founder and CEO of Attentive AI. “The biggest challenge for any startup is running out of money. Some run out of funding whereas some run out of paying customers,” said Dhawan. He further said, “From the very beginning, we knew that we would face the latter problem rather than the former. However, we were also confident that overcoming this problem would lead us to build a sustainable business, and so we relied on our customers to sustain ourselves as well as to grow. As a result, we tried to be as frugal as possible with our expenses. For instance, for a long time, we did not work out of a high-end corporate office instead, we worked out of a couple of apartments which were converted into modest office spaces.” The key to success is to build a team that understands the importance of frugality and enjoy the startup journey just as much as you do, added Dhawan. Attentive AI’s approach involves experienced computer vision team that prepares deep learning models and are trained on geospatial imagery data, in-house annotators process machine-generated data which fix gaps and inaccuracies, expert in-house quality control team that ensures nearly 100% correctness through pain-staking visual inspection, and client monitors live production until the map features are delivered in the desired file format. Standing Out Of The Crowd With the Chinese government leading the AI race with its expansive surveillance system and heavy investment in AI research and skill development, the Asian governments still lag in terms of implementing AI initiatives. With such a scene in hand, Attentive AI has come up with an innovative solution to deliver the best GIS solutions using high-quality digital maps. MapX is Attentive AI’s flagship product that can be used to request meaningful insights from geospatial imagery data. The core of the product envelops a simple workflow allowing each user to select an address or an area, followed by requesting the analytics they desire from a list of available analytical options or can even go for a custom request followed by near-instant delivery of output\/insights. MapX prides itself on being the ‘McDonalds’ of the geospatial world. Thus, one does not have to wait a long time to get the geospatial dataset or a digital map. “Initially, when McDonald’s had started, people said it was impossible to deliver high-quality burgers in such a short time but they did it, and their key to success was their engineering process. Similarly, delivering high-quality land features almost instantly was also impossible, but Attentive AI makes it possible by an engineering process at the core of which is our AI algorithm supported by a detailed QC process,” said Dhawan. Being a customer serving web platform, MapX also makes the experiences of ordering geospatial data very seamless for customers and users. “Our customers attest to the seamless geospatial data ecosystem that we are creating.” Surviving Industry Challenges AI, being one of the newest innovative technology with a maximum number of commercial benefits, comes up with several industry challenges. The major limitation in the AI industry is customer awareness, or the whole concept of AI, where it is to be believed that AI itself is the solution to all problems. However, that is not true as AI technology alone can not provide a complete solution. AI does aim to solve the most complex steps, but the whole solution can only be created with a combination of multiple technologies including AI. Explaining this, Dhawan said, “One of the major challenges of being an AI service provider is that a customer’s expectations. Usually, customers tend to have high expectations from AI rather than what is currently possible. And, this is mostly because of the increased hype around AI and also because a lot of companies project a higher AI capability than what is possible at the current stage.” “These companies show pilots on specific and suitable examples, however, the truth is that scaling an AI is incredibly hard and it takes a significant amount of time to build an AI that is accurate on all possible user scenarios. We, at Attentive AI, try to mitigate this challenge by educating our customers transparently about the training period and the processes that are necessary to make the AI scalable. Thus our clients understand that the AI will not be scalable from the first day, rather through an iterative and active learning mechanism, which will help it to grow, and be more efficient,” said Dhawan. In the rising age of AI, the key to success is to build artificial intelligence systems for a specific niche datasource to harness the power of data network effects, which Attentive AI is acing with their high-resolution geospatial imagery. “This gives us a competitive advantage while making cutting edge breakthroughs every week since we have worked on multiple use cases over a period from which our AI systems are continuously learning,” said Dhawan. Future Prospective Attentive AI, being one of the AI service provider, aims to create an accurate, constantly updating digital twin of the physical world. “We have only touched the tip of the iceberg as we are creating more AI technologies to analyse aerial imagery in specific geographies,” said Dhawan. “We aim to build a global repository of multiple geospatial imagery sources and a suite of intelligent analytics for customers to request analytics on drones, streets, light detection and ranging (LIDAR), and all other kinds of geo-data sources at any time from anywhere,” concludes Dhawan.","excerpt":"The synergistic integration of artificial intelligence (AI) and the geographic information dimension creates geospatial artificial intelligence (GeoAI). Geo-tagged big data collated from varied sources, such as satellite imagery via remote sensing, IoT sensors in smart cities, social media streaming, and personal sensing via connected ambient and wearable sensors, can be analysed using GeoAI to get […]","categories":["AI Startups"],"tags":["Geospatial analysis India","Geospatial Analytics Tools and Platforms","geospatial data","geospatial data India","Geospatial Data Tools and Technologies","geospatial mapping"],"author_name":"Sejuti Das","publish_date":"2019-12-30T17:00:00","publication_year":"2019","word_count":1084,"keywords":["Go","artificial intelligence","TPU","geospatial data","AI","Geospatial Analytics Tools and Platforms","geospatial data India","Geospatial Data Tools and Technologies","Geospatial analysis India","ML","computer vision","Aim","deep learning","analytics","R","geospatial mapping"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","computer vision","analytics","Aim","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/why-this-delhi-based-startup-prides-itself-as-the-mcdonalds-of-geospatial-world\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10040896,"title":"Top ML Announcements From Microsoft Build 2021","content":"“Azure ML Managed Endpoints is already being used for the OpenAI GPT-3 model.” Microsoft kicked off its annual developer’s conference, Microsoft Build. The event, which will be held digitally for the second time in a row, had a jam-packed first day where multiple announcements within various facets of Microsoft, from MS Teams to Azure, were made. Here is a round-up of key announcements and releases related to artificial intelligence and machine learning. AI for Azure Build 2021 made many AI-related updates to Azure, Microsoft’s cloud computing platform. Azure’s AI service involves a portfolio of AI services for data scientists and developers to deploy their own AI solutions. The portfolio includes ML models with tools such as Jupyter Notebook and open-source models like PyTorch and TensorFlow. Microsoft has also announced two key machine learning capabilities to help customers deploy AI models with ease. The first of these is called Azure Machine Learning Managed Endpoints, which is currently in preview. This novel capability would allow developers and data scientists to build and deploy ML models quickly and easily. The model automates the creation and management of the underlying compute infrastructure, which includes updates and security features. Through this, users can get access to out-of-the-box infrastructure monitoring and log analytics capabilities. We are building the platform for platform creators. At #MSBuild, we are introducing more than 100 new updates and services across the tech stack to make creating easier for every developer. https:\/\/t.co\/QT72JMA9ls— Satya Nadella (@satyanadella) May 25, 2021 Azure ML Managed Endpoints is already being used for the OpenAI GPT-3 model, which is among the world’s most significant natural language models—in Microsoft Power Apps. Microsoft announced its use of the OpenAI GPT-3 model to generate Power Fx formulas based on natural language input automatically. As per Satya Nadella’s keynote announcement, this is being added to the Power Apps Studio. Users can leverage GPT-3 to translate their queries in natural language into usable Power Fx Code. For instance, instead of learning how to write database queries in Power Fx, the user could write ‘Show 10 orders with ABC in their product name and sort by date with oldest on top’, and the AI will produce the corresponding code for them. This way, Power Apps users will be able to build apps more quickly and efficiently. Conversational AI Bot Framework Composer 2.0 (Source: Microsoft) A slew of improvements has been made to Azure’s Bot Service and Bot Framework, including releasing a new model: Bot Framework Composer 2.0. The service provides an integrated, purpose-built environment for bot development. It includes a visual authoring canvas with open-source tools to enable developers to add speech and telephony features and help them test, debug, and publish bots to multiple channels with minimal code changes. Also, updates have been added to make the bot service more secure with the introduction of Customer Managed Encryption Keys—which would allow the data stored about the bot to be encrypted by a Microsoft-managed key and a key provided by the user. The Bot Framework Composer is Microsoft’s open-source integrated development tool for writing conversational AI applications. Its 2.0 version would feature new templates to pose as starting points for your bots. Composer 2.0 will also guide the user as they begin to build their application. These tools would also make publishing bots to Azure easier with integrated resource provisioning and publishing from within the Composer. Azure Metrics Advisor Source: Microsoft In September 2020, Microsoft announced its plans for the Azure Metrics Advisor—which was inspired by their work in detecting deviations from normal operations within the Bing search engine. At Build 2021, the tech giant announced that the AI-powered platform is now generally available. The service, which is tailored for businesses, will take in time-series data and use machine learning to detect anomalies from sensors, products and metrics automatically. The algorithm used will also provide diagnostic insights and allow users to monitor their organisation’s performance with ease (without themselves needing to be experts of machine learning). Azure Video Analyser Another new service is the Azure Video Analyser. This service, like the Metrics Adviser, also looks for anomalies using AI. It uses existing services such as the Computer Vision from Azure Cognitive Services to build AI-powered video analytics from both stores and streaming footage. According to Microsoft, this brings Live Video Analytics and Video Indexer into a single service. The former would help build intelligent, video-based apps using the user’s choice of AI. At the same time, the Indexer automatically extracts data from video and audio content. Customers can use this for workplace safety, digital asset management, and monetising content. The Azure Bot Service, Metrics Advisor and Video Analyser—along with Azure Cognitive Search, which would bring an AI-backed cloud search to mobile and web apps, and Azure Form Recogniser, which would enable quickly turning documents into usable data—will all be making up a new category within Azure AI called Azure Applied AI Services. These services build on the cognitive APIs from Azure Cognitive Services and Azure Machine Learning. Doing so, as per Microsoft, provides it with task-specific AI and built-in business logic, allowing it to develop AI-powered business solutions further. Azure Cognitive Services Further updates in AI fall under the Azure Cognitive Services, which are another collection of AI services and cognitive APIs to help users build intelligent applications. Microsoft Build 2021 brought in enhancements for these products, especially Document Translation and Text Analytics for Health. Document Translation Document Translation, which was announced in February, is now available. This feature allows developers to translate documents while preserving the structure and format of the original document. The AI-delivered feature could quickly perform tasks for enterprises that require going through complex documents in one or more languages. Text Analytics for Health Now generally available with Text Analytics in Azure Cognitive Services, this would quickly process and derive insights from unstructured medical data. Such data includes doctors’ notes, medical journals, clinical trial protocols, and electronic health records. An additional feature of Text Analytics is called Question Answering, which will allow users to find answers from text passages without needing to save or manage data in Azure. . Source: Microsoft Finally, Microsoft is bringing PyTorch Enterprise to Azure, which would provide additional benefits to users of Microsoft Premier and Unified Support for Enterprise. These benefits include hands-on support and solutions to bugs and security patches. Microsoft users with Microsoft Premier and Unified Support using PyTorch will be eligible for PyTorch Enterprise and will be able to request hotfixes. Stay tuned to Analytics India Magazine for more updates from Microsoft BUILD 2021","excerpt":"“Azure ML Managed Endpoints is already being used for the OpenAI GPT-3 model.” Microsoft kicked off its annual developer’s conference, Microsoft Build. The event, which will be held digitally for the second time in a row, had a jam-packed first day where multiple announcements within various facets of Microsoft, from MS Teams to Azure, were […]","categories":["AI Trends"],"tags":[],"author_name":"Mita Chaturvedi","publish_date":"2021-05-26T17:17:43","publication_year":"2021","word_count":1092,"keywords":["machine learning","artificial intelligence","OpenAI","AI","PyTorch","ML","computer vision","analytics","TensorFlow","Azure ML"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","analytics","OpenAI","Azure ML","TensorFlow","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ml-announcements-from-microsoft-build-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10009855,"title":"Google Analytics 4 Released: Key AI\/ML-Based Enhancements","content":"Google has announced an overhauled version of Google Analytics. In one of the major revamps of the platform, in a decade, the new Google Analytics is built on the foundation of App+Web property, whose beta was introduced in 2019. The new Google Analytics has machine learning at its core and allows integration between analytics and Google Ads. The company claims that this would help customers manage their data better and can bear industry disruptions. Among the most significant changes, the new analytics will alert the user of the significant trends in their data. Further, one can also anticipate actions that customers may take in the future. Other features include the addition of new predictive metrics. The company claims that such insights can help users and business owners achieve high-value customers and improve results by taking steps like analysis of customer expenditure patterns. Features of Google Analytics 4 Since 2005, Google Analytics has been the industry leader and standard among web analytics tools. Its free standard version is favoured among individuals and small businesses. For larger companies, Google provides a paid product — Google Analytics 360. In addition to the standard functions, Google Analytics 360 provides features such as unsampled reports, BigQuery export, and data-driven attribution. The new Google Analytics is currently in beta with an Analytics 360 version that will provide SLAs and integrations with tools like BigQuery. This new update is aimed at enhancing the customers’ decision-making abilities and get a better return on investment (ROI). One of the most exciting features is the availability of AI-powered insights and predictions. Machine learning-powered insights have been a part of Google Analytics for some time now; however, the AI-powered insights will now help business owners automatically be alerted about data trends such as surging demand for their products. It can also be used to predict business outcomes such as potential revenue collection; these predictions can act as a guiding force for businesses to plan their next move. With Google Analytics 4, marketers can now build and maintain audiences from the website and app visitors. For example, suppose the visitor of the website or app has been converted to an audience because they made a purchase. In that case, analytics will make sure that such users are removed from the list to avoid getting retargeted with the ads. The way reports are organised in Analytics 4 is one of the most striking features that differentiate it from Universal Analytics. As per the company statement, these detailed reports can reveal even the minutest of the customer engagement, such as what channels are attracting customers, whether or not they are sticking around after conversion, among others. Marketers will now be able to provide customised user ID to gauge users across devices for reporting and targeting. This has been possible due to Google Analytics’ significant move from measurement fragmented by device measurement to a more customer-centric measurement. Business owners will now be able to track and measure customer actions on their apps and website, without requiring to add code, thanks to the new expanded codeless features. This is a massive update as compared to Universal Analytics which still requires high-latency processing for the purpose. Other features include relying on machine learning for filling the data gap caused by phasing out of third-party cookies; and the introduction of several new options to help advertisers comply with data regulations. About Web+App Beta Type Last year, Google had announced the beta of web+app property, which is now the foundation of the new Google Analytics 4 version. Invented to help customers navigate between both the websites and apps simultaneously, Google had announced a new way to measure apps and website together for the first time for unified reporting and analysis. This would allow business owners to gather data and benefit from the latest innovations while maintaining current implementations. In its new avatar, this feature provides Analytics 4 the capability to offer smarter insights on user behaviour.","excerpt":"Google has announced an overhauled version of Google Analytics. In one of the major revamps of the platform, in a decade, the new Google Analytics is built on the foundation of App+Web property, whose beta was introduced in 2019. The new Google Analytics has machine learning at its core and allows integration between analytics and […]","categories":["Global Tech"],"tags":["Google Ads","Google Analytics"],"author_name":"Shraddha Goled","publish_date":"2020-10-17T16:00:57","publication_year":"2020","word_count":657,"keywords":["Go","machine learning","AI","Google Ads","data-driven","innovation","RAG","Google Analytics","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","R","Go","GAN","data-driven","innovation"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-analytics-4-released-key-ai-ml-based-enhancements\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":48812,"title":"Is Data Science The Right Choice For You?","content":"Data science is the “sexiest job of the 21st century” and might also have the highest paycheck compared to a lot of other jobs. And despite these perks and benefits, there is a burning question that often puts people in a serious dilemma — is data science the right fit for me? There are people who end up becoming a part of this tribe and then later regret when things turn difficult. This is mostly because of the hype that gets created at a certain point of time and when we see everyone following data science. In this article, we are going to list some points that would help you figure out whether data science is a good fit for you. How Well Do You Understand The Data Science Domain? This is definitely the first and foremost question you must ask yourself when you are considering data science as a career. Data science is a really vast domain and requires deep knowledge and specific skill set to excel. Further, it’s not just the knowledge and skills, but you also have to understand the different job roles in this industry and the difference between them. For example, the role of a data scientist is different than the role of a data engineer. Once you look at all the different job roles, you would get an idea about what the work is and whether you fit any of them. Some of the job roles are: Data Scientist Business Analyst Data Analyst Data Engineer\/Data Architect Statistician etc. How Well Do You Gel With Mathematics And Statistics? Mathematics and Statistics are two of the most vital pillars of data science. They are so important that their concepts are extensively used in the data science field. So, lately, if you have been wondering about starting a career in data science you have to assure that your knowledge and skills with mathematics and statistics are top-grade. Cement this fact in your mind that you will have to work with numbers extensively if you are working as a data science professional. You will have to spend a significant amount of time to build a great bond with numbers. You will also need to figure out what are all the prime concepts that will be used in your day to day data science job. To give you an idea, here are a number of statistical principles and theorems that matters a lot when you are trying to enter the data science domain: Linear Algebra Dimensionality Reduction Probability Distribution Central Limit Theorem Bayes Theorem To know more about these concepts and how they would help you have a data science career, you can read this article “5 Mathematical Concepts Every Data Science Aspirant Should Master.” How Well Can You Analyse Things And Solve Problems? While data science is definitely about algorithms and models, there is something that you cannot ignore when you are working in this field — analysing and solving problems. For instance, when you are playing games like chess or card, what approach do you take? You play completely depending on your luck or you analyse the entire scenario and then make your next move? Luck contributes a small percentage, but your strategies matter a lot. It is same with business; you will have to think based on both rational and emotional aspect. You will have to have the mindset of finding a solution to the complex problems and challenges businesses face. How Much Time You Can Devote To Learning? As mentioned earlier, data science is a vast domain and the learning process takes time. It’s not just the concepts, but one needs to have a 360 degree understanding of the entire domain which includes the tools and methods as well. If you are working professional who is occupied much of the time with office work and the family and doesn’t have much time to put in for learning, then you might have to figure out a way to manage. You can take up online courses or even make use of free resources. Enrol yourself to some data science course, engage with the community, work on projects and once you are ready enough, make that move to land a job. Here’s a list of books you can check Here’s a list of YouTube channels you can follow. YouTube Channels For AI Enthusiasts Must Watch Big Data Videos On YouTube Also, you can watch these TedTalks on Data Science Do You Love Working With People\/Teams From Different Department? Being a data scientist in an organisation doesn’t mean you would work the entire day with people from your domain. The prime reason why a data science department is there in an organisation is to solve different problems and these problems may come from a different domain. This scenario also includes things like communication and presentation skills. As a data science professional, you would also have to present your work, findings, and results to a set of people. You will have to be creative and good storyteller when you are interacting. Here are some of the most vital skills that you need to be a data storyteller: Ability to understand the target audience Should be able to discover the sole purpose of the data story Should know how to make the best presentation Skills and knowledge of data visualisation tools So you have to make sure that you are comfortable spending a significantly large proportion of your time with people doesn’t understand or do data. Ask yourself whether you would be able to interact with such people and make them understand what you are working on and how you are solving these problems.","excerpt":"Data science is the “sexiest job of the 21st century” and might also have the highest paycheck compared to a lot of other jobs. And despite these perks and benefits, there is a burning question that often puts people in a serious dilemma — is data science the right fit for me? There are people […]","categories":["AI Features"],"tags":["big data video games","Data Science","Data Science Career","data science learning","mathematics","Statistics"],"author_name":"Harshajit Sarmah","publish_date":"2019-10-25T18:00:48","publication_year":"2019","word_count":943,"keywords":["mathematics","data science","Go","big data","programming_languages:R","Statistics","AI","R","programming_languages:Go","GAN","Data Science Career","ViT","big data video games","Data Science","data science learning"],"extracted_tech_keywords":["AI","data science","R","Go","big data","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-data-science-the-right-choice-for-you\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172577,"title":"Why Manhattan Associates Wants Its Clients to Build AI Agents","content":"AI agents are quickly becoming a standard in various industries. According to a recent LangChain report, while 51% of companies are already using AI agents in production, 78% have plans to adopt them in the near future. Originally rooted in robotic process automation (RPA), which handles simple, repetitive tasks, AI in the supply chain is now evolving rapidly. Today’s AI agents go beyond task automation—they’re capable of managing entire workflows, adapting in real time and supporting decision-making across logistics, procurement, and inventory. Manhattan’s Alternative Approach While many vendors are racing to build AI agents for their clients, Manhattan Associates—a global technology leader in supply chain and omnichannel commerce—has chosen a different route. Manhattan Associates India Development Centre is the largest GCC with over 2,000 employees, having started with just five employees in 2002. Rather than pre-building large sets of agents, the company has introduced Manhattan Agent Foundry—a platform that enables clients to create their own agents, tailored to specific needs and workflows. “You can write your agent, you can take our agents, you can mix and match, you can work with it,” Ushasri Tirumala, executive vice president and GM, India at Manhattan Associates, said in a recent interview with AIM. “That is the philosophy that we have. We are not announcing we will do 200 agents or whatever. We are going to provide the critical agents for you to understand and leverage them, and then you can develop your own.” What Makes Manhattan Agent Foundry Stand Out Agents built on the Manhattan platform are powered by large language models (LLMs) and integrated into a cloud-native, microservices-based API architecture. These AI agents are not just reactive—they can orchestrate processes, adjust dynamically to real-world conditions, and communicate across systems. Notably, Manhattan’s agents are designed to interoperate with third-party agents, adhering to standards like A2A and MCP. This means enterprises can mix Manhattan-built agents with those developed or acquired elsewhere, creating a truly interconnected, AI-driven ecosystem. “Customers would have the ability to get the information when they want, however they want, and to make decisions,” Tirumala shared. “It is very important for companies because it would help in certain areas so that you don’t need to do the kind of context jobs…using more skilled manpower that would go away,” she added. Tirumala confirmed that these agents are being integrated across all Manhattan product lines, with broader availability planned by this fall. A Unified View of the Supply Chain Manhattan’s product philosophy revolves around unification. From demand forecasting to order and transportation management, the focus is on creating a cohesive view across supply chain operations. “More and more today, we are talking about unification. It is not enough if you look at one transportation management or distribution management,” Tirumala explained. “You have to have a total view, and [know] how to bring that visibility to people who are planning, using, etc. That is what we are bringing.” Manhattan’s broader product strategy already includes several key innovations aimed at improving supply chain intelligence and customer experience. These include Manhattan Active Supply Chain Planning (MASCP), designed for advanced forecasting and supply optimisation, and the Enterprise Platform Framework (EPF), which enables seamless omnichannel visibility. Manhattan Active Maven, a generative AI solution tailored for next-generation customer service, and Manhattan Assist, an AI-powered assistant that provides contextual guidance on product functionality, API structures, and more. Furthermore, Manhattan’s move into AI didn’t happen overnight. The company began investing in AI years ago and even developed its own machine learning platform. “Even a couple of years back, we had developed our own machine learning platform,” Tirumala noted. Talking about LLMs, Tirumala revealed that Manhattan also maintains a strong partnership with Google, hosting its solutions on Google Cloud and leveraging Google’s AI infrastructure. “We do not stick to one LLM,” Tirumala added. “We work with Google in a big way. We work with Google platforms. So, our solutions are also hosted on the Google Cloud.” Bengaluru for Scaling Manhattan Associates takes a deliberate and focused approach to team structure. Unlike companies with multiple development centres across the globe, Manhattan operates from just two primary locations—the United States and India—with over two-thirds of its R&D workforce based in Bengaluru. “We are not in the tens of thousands in number. And we believe that product development requires very close working together,” Tirumala said “We cannot have one centre here, another centre somewhere else…That will be quite difficult.” Tirumala mentioned that an earlier internal study even explored adding a third development hub, but the findings supported sticking with Bengaluru as the optimal location. “We looked at whether it is worthwhile establishing another centre from a talent perspective, not for anything else. And then that study came back with the finding that being in Bengaluru is the best thing. Not even Mysuru at that time,” she said. With a lean team of just over 2,000 employees, Manhattan India handles not only R&D but also customer support and product-based services—from implementation to lifecycle support. “Product-based services, the implementation of that, the extension of that, and improving whatever the customer wants from their work perspective—all of those things are carried out here,” Tirumala noted. Note: The headline has been updated to improve clarity and reader understanding.","excerpt":"Manhattan maintains a strong partnership with Google, hosting its solutions on Google Cloud and leveraging Google’s AI infrastructure.","categories":["GCC"],"tags":["GCC"],"author_name":"Shalini Mondal","publish_date":"2025-06-30T14:12:34","publication_year":"2025","word_count":868,"keywords":["Go","machine learning","GCC","AI","ML","RAG","microservices","LangChain","Aim","generative AI","R"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","LangChain","Aim","RAG","microservices","R","Go"],"url":"https:\/\/analyticsindiamag.com\/gcc\/why-manhattan-associates-is-not-developing-ai-agents-for-its-clients\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10568,"title":"DBS Digibank: How India’s first mobile-only bank uses AI?","content":"With the growth of digital technology, banking sector has gone through radical changes. From branches banking to the era of internet banking and ATM’s where banking facilities are available 24*7 at the ease of transacting from anywhere. Breaking away from the conventional way of banking, Singapore based DBS bank has taken a step to take banking to an altogether new level. DBS Bank has opened a ‘mobile-only’ bank in India called the ‘Digibank’. About Digibank Well Digibank is an innovative offering which combines cutting-edge technology from biometrics to artificial intelligence (AI) to give its customers a hassle-free banking experience. Digibank is a completely paperless, signatureless and branchless bank. So let’s explore what Digibank is all about and how it functions. This mobile-only bank is completely paperless and has no physical presence in terms of branches in India. Digibank’s functionality is based on technologies like biometrics, AI, and natural language recognition and processing to provide its services. How to open an account To open a bank account with Digibank, a customer needs only an Aadhar card and a biometrics-enabled ID card for authentication and he or she can open an account with Digibank at any one of bank’s 500 partner cafes spread across India. Features of Digibank Digibank’s technology is supported by Kasisto, a US based startup which brings AI into banking. For Digibank, Kasisto has introduced KAI, an AI-powered virtual assistant. KAI is designed to anticipate and answer customer queries related to banking in real time. Another feature of Digibank is an intuitive budget optimiser designed to help its customers to track their expenses, analyze their purchase trends and carry out budgeting. The budget optimiser is designed smartly to draw inference from customer behaviour and preferences and accordingly provide recommendations. Another noteworthy feature of Digibank is enhanced security with dynamic inbuilt security. Like other traditional banks, Digibank does not require an OTP (One Time Password) to be received via SMS to carry out a transaction. Instead digibank has bettered its security system for transaction authorization through soft token security. Hence DBS customer don’t need to wait for SMS to carry out any transaction and have more secured banking experience with this technology. Apart from these features, Digibank being cost effective is able to provide its customers more value. This mobile-only bank offers its account-holders an interest rate of 7%, one of the highest in the Indian banking market. Also there is no minimum balance requirement and customers can relish unlimited free cash withdrawals at over 200,000 ATMs nationwide. The account-holders are provided with a physical debit card which is valid across all Visa-enabled online and point-of-sale transactions domestically and internationally. This initiative of mobile-only bank by DBS can act as a game-changer for the Indian banking industry. This can enable the Indian banking industry to go beyond the confines of physical banking and extend their reach with the help of breakthrough technology making banking more customer-friendly.","excerpt":"With the growth of digital technology, banking sector has gone through radical changes. From branches banking to the era of internet banking and ATM’s where banking facilities are available 24*7 at the ease of transacting from anywhere. Breaking away from the conventional way of banking, Singapore based DBS bank has taken a step to take […]","categories":["IT Services"],"tags":[],"author_name":"Manisha Salecha","publish_date":"2016-08-07T07:09:37","publication_year":"2016","word_count":486,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Git","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/dbs-digibank-indias-first-mobile-bank-uses-ai\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":39897,"title":"Who Moved My Camera? Deep Learning Can Now Identify Depths In Videos","content":"Dolly Shot From Spielberg’s Jaws What is moving? The man with the glasses? Or is the world collapsing around him? Actually, it’s neither — it is the camera that is moving inwards in a technique popularly-known as dolly shot. This distorting spacetime effect was first effectively used in Hitchcock’s masterpiece Vertigo (1958) to manipulate the attention of the audience. Even with all the knowledge of the real world, it is tricky for humans to identify how far or close the objects in the videos really are — such as in the video above. It gets even messier when you task a machine to recognise the depth in the videos. Moreover, unlike the movies, the real world video streams can have objects\/people moving along with the camera and other such possibilities. A shot from Alfred Hitchcock’s Vertigo The existing 3D reconstruction algorithms will find such freely moving targets confusing. So, most existing methods either filter out moving objects (assigning them “zero” depth values) or ignore them (resulting in incorrect depth values). A paper titled, Learning the Depths of Moving People by Watching Frozen People, published by the researchers at Google, talks about how a deep learning approach can be used to generate depth maps from an ordinary video that can tackle the previously mentioned challenges. This model avoids direct 3D triangulation by learning priors on human pose and shape from data. This work is the first of its kind as it uses a learning-based approach to the case of simultaneous camera and human motion. The success of this method can find its significance in applications involving augmented reality and 3D video effects. Overview of the network via paper by Zhengqi Li et al., The objective here is to generate depth maps of videos having moving people and moving camera. For this, the authors pursue a multiple point approach where the motion of static point between two points can give how far or close a certain object is, in a frame. To make sense from these multi-viewpoint cues, a 2D optical flow is calculated. Optical flow is a mathematical approach to identify the motion of an object in a frame. This was originally modeled around how animals perceive their surroundings as they move. So, an optical flow would give out the difference between frames by considering pixel intensities and other such attributes. For instance, if an object is getting brighter with every frame then it can be inferred that the object is not only moving but also coming closer as well. In this case, the camera position is predetermined, hence their dependencies are not considered. Images devoid of dependencies are fed and are checked for humans. The algorithm masks all the potential humans in the image using mask R-CNN. These masked regions are removed and this image is run through regression network, which predicts depth. Depth Maps via Google AI blog One cool outcome of generating depth maps algorithm is the synthetic defocus as can be seen above. As the network applies segmentation and optical flow computation, the target objects can be pushed out of attention and make other CG works in post-production tasks in case of movie making. Even though this approach looks promising, there is still room for improvement. In videos where camera position is of less significance or unknown can trick the model. And, there are obviously non-human objects in almost all videos. However this approach will act as a vantage point for future works, which use neural networks to decipher and design videos. Know more about this work here Also watch:","excerpt":"What is moving? The man with the glasses? Or is the world collapsing around him? Actually, it’s neither — it is the camera that is moving inwards in a technique popularly-known as dolly shot. This distorting spacetime effect was first effectively used in Hitchcock’s masterpiece Vertigo (1958) to manipulate the attention of the audience. Even […]","categories":["Deep Tech"],"tags":["Deep Learning"],"author_name":"Ram Sagar","publish_date":"2019-05-30T07:36:15","publication_year":"2019","word_count":593,"keywords":["Go","AWS","AI","neural network","R-CNN","cloud_platforms:AWS","Mask R-CNN","deep learning","Deep Learning","CNN","R"],"extracted_tech_keywords":["AI","deep learning","neural network","AWS","R","Go","CNN","R-CNN","Mask R-CNN","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/who-moved-my-camera-deep-learning-can-now-identify-depths-in-videos\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10051362,"title":"IBM Aims to Skill More Than 30 Million People Globally By 2030","content":"IBM unveiled a groundbreaking commitment and global plan to provide 30 million people of all ages with new skills needed for the jobs of tomorrow by 2030. IBM is announcing a clear roadmap with more than 170 new academic and industry partnerships to achieve this goal. The effort will leverage IBM’s existing programs and career-building platforms to expand access to education and in-demand technical roles. With diverse offerings and an adaptable approach, IBM’s education portfolio strives to be unique and effective, reflecting IBM’s understanding that a one-size-fits-all approach simply does not work when it comes to education. IBM’s programs range from technical education for teens at brick-and-mortar public schools and universities and extend to paid, on-site IBM internships and apprenticeships. The company’s skills and education programs also pair IBM mentorships with learners and provide no-charge, customizable online curricula to aspiring professionals. “Talent is everywhere; training opportunities are not. This is why we must take big and bold steps to expand access to digital skills and employment opportunities so that more people – regardless of their background – can take advantage of the digital economy. Today, IBM commits to providing 30 million people with new skills by 2030. This will help democratize opportunity, fill the growing skills gap, and give new generations of workers the tools they need to build a better future for themselves and society.” said Arvind Krishna, Chairman and CEO, IBM. In India, IBM will continue the partnerships with the Ministry of Skills Development & Entrepreneurship, MEITY, Board of Open Schooling and Skill Education – Sikkim, Navodaya Vidyalaya Samiti (DST), CBSE, Skill Development Councils in Goa, Telangana and Andhra Pradesh, NPTEL – IIT-Madras and many others as implementing partners to upskill youth from across the country. Dr JP, DGT, Ministry of Skill Development and Entrepreneurship, said, “We have collaboratively worked through our 16 nodal National Skills Training Institute (NSTIs) spread across 14 states to effectively train and skill students and teachers from vocational institutes to impart necessary future skills. The IBM SkillsBuild program provides free, self-paced technical and non-technical training to tens of thousands of students across the country. These initiatives will help the youth of our country to effectively and efficiently prepare and meet the evolving demands of the industry.” IBM’s longstanding commitment to education has long been core to its corporate social responsibility initiatives. Three years ago, IBM launched its STEM for Girls program with a vision to reach out to 300,000 students and initiate them into the world of STEM. IBM also launched an online learning platform SkillsBuild in India, which has over 500,000 learners of all age groups, learning and closing the skill gaps in emerging tech.","excerpt":"The company’s skills and education programs also pair IBM mentorships with learners and provide no-charge, customizable online curricula to aspiring professionals.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","covid-19","Data Science","Data Science Jobs","Data Scientist","Deep Learning","education industry india","IBM","Machine Learning","Programming Languages"],"author_name":"Victor Dey","publish_date":"2021-10-13T18:09:12","publication_year":"2021","word_count":443,"keywords":["Go","covid-19","AI","programming_languages:R","Machine Learning","Data Science Jobs","Programming Languages","RAG","Git","programming_languages:Go","IBM","GAN","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)","education industry india"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-aims-to-skill-more-than-30-million-people-globally-by-2030\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129231,"title":"Time To Scale Down Large Language Models","content":"Renowned research scientist Andrej Karpathy recently said that the llm.c project showcases how GPT-2 can now be trained in merely 24 hours on a single 8XH100 GPU node—for just $672. Karpathy’s journey began with an interest in reproducing OpenAI’s GPT-2 for educational purposes. He initially encountered obstacles in using PyTorch, a popular deep-learning framework. Frustrated by these challenges, Karpathy decided to write the entire training process from scratch in C\/CUDA, resulting in the creation of the llm.c project. It eventually evolved into a streamlined, efficient system for training language models. The project, which implements GPT training in C\/CUDA, has minimal setup requirements and offers efficient and cost-effective model training. Scaling down LLMs In his post, Karparthy mentioned how advancements in hardware (H100 GPUs), software (CUDA, cuBLAS, cuDNN, FlashAttention), and data quality have drastically reduced training costs. Mauro Sicard, the director of BRIX Agency, agreed with Karparthy. “With the improvements in both GPUs and training optimisation, the future may surprise us,” he said. Scaling down LLM models while maintaining performance is a crucial step in making AI more accessible and affordable. According to Meta engineer Mahima Chhagani, LLMLingua is a method designed to efficiently decrease the size of prompts without sacrificing significant information. Chhagani said using an LLM cascade, starting with affordable models like GPT-2 and escalating to more powerful ones like GPT-3.5 Turbo and GPT-4 Turbo, optimises cost by only using expensive models when necessary. FrugalGPT is another approach that uses multiple APIs to balance cost and performance, reducing costs by up to 98% while maintaining a performance comparable to GPT-4. Additionally, a Reddit developer named pmarks98 used a fine-tuning approach with tools like OpenPipe and models like Mistral 7B, cutting costs by up to 88%. Is there a Real Need to Reduce Costs? Cheaper LLMs, especially open-source models, often have limited capabilities compared to the proprietary models from tech giants like OpenAI or Google. While the upfront costs may be lower, running a cheap LLM locally can lead to higher long-term costs due to the need for specialised hardware, maintenance overheads, and limited scalability. Moreover, as pointed out by Princeton professor Arvind Narayanan, the focus has shifted from capability improvements to massive cost reductions, which many AI researchers find disappointing. Cost over Capability Improvements Narayanan argued that cost reductions are more exciting and impactful for several reasons. They often lead to improved accuracy in many tasks. Lower costs can also accelerate the pace of research by turning it more affordable and making more functionalities accessible. So, in terms of what will make LLMs more useful in people’s lives, cost is hands down more significant at this stage than capability, he said. In another post, Narayanan said that the cheaper a resource gets, the more demand there will be for it. Maybe in the future it will be common to build applications that invoke LLMs millions of times in the process of completing a simple task.This democratisation of AI could accelerate faster than we imagined, possibly leading to personal AGIs for $10 by 2029.","excerpt":"Advancements in hardware (H100 GPUs), software (CUDA, cuBLAS, cuDNN, FlashAttention), and data quality have drastically reduced training costs.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Large Language models"],"author_name":"Anshul Vipat","publish_date":"2024-07-16T11:26:13","publication_year":"2024","word_count":504,"keywords":["CUDA","Go","OpenAI","AI","Large Language models","PyTorch","ML","Scala","Ray","Rust","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","OpenAI","Ray","PyTorch","CUDA","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/time-to-scale-down-large-language-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004859,"title":"Telangana Government To Use AI For Agricultural Innovation","content":"Telangana IT Minister recently announced the launch of Artificial Intelligence for Agricultural Innovation (AI4AI) programme to boost agricultural developments. It was launched in association with Centre for the Fourth Industrial Revolution, India (C4IR), World Economic Forum. Agriculture is one of the crucial domains for the state and the government aims to bring digital technology-led innovation in the field. Along with Telangana State Agricultural University (PJTSAU) and the Telangana government’s ITE&C department, the govt. is aiming to identify the high impact use cases of AI which could benefit both farmers and policymakers. Telangana has announced the year 2020 to be the Year of AI to accelerate the adoption in various fields. The state aims to develop a conducive AI innovation ecosystem to use AI for social good, especially in the public sector. As a part of 2020- Year of AI, the government has partnered with several reputed organisations in industry and academia to brings various projects to life such as AI strategy for Telangana, Telangana’s AI Mission, Centre for Research in Applied AI, and more. Speaking at the launch, Telangana industries and IT Minister Rama Rao said, “We feel that AI will offer immense possibilities for farmers, governments and other stakeholders of the ecosystem.” The Minister is hopeful that working closely with organisations such as the World Economic Forum will bring in a robust community of experts and years of experience to lead this agricultural innovation. The Telangana government had announced the deployment of the country’s first automated ‘COVID-19 Monitoring System App’ to help users in identifying, undertaking live surveillance, tracking, monitoring, and providing real-time analytics.","excerpt":"Telangana IT Minister recently announced the launch of Artificial Intelligence for Agricultural Innovation (AI4AI) programme to boost agricultural developments. It was launched in association with Centre for the Fourth Industrial Revolution, India (C4IR), World Economic Forum. Agriculture is one of the crucial domains for the state and the government aims to bring digital technology-led innovation […]","categories":["AI News"],"tags":["telangana government"],"author_name":"Srishti Deoras","publish_date":"2020-08-14T13:09:53","publication_year":"2020","word_count":265,"keywords":["Go","artificial intelligence","AI","innovation","Git","telangana government","Aim","analytics","real-time analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","Aim","R","Go","Git","real-time analytics","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/telangana-government-to-use-ai-for-agricultural-innovation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":65677,"title":"What May Be The Factors Behind Your Layoff?","content":"The global recession has engulfed the world of technology. In the field of data science, layoffs are happening across many companies, and analytics professionals have concerns staying relevant in these tough times. Alongside, businesses are also strategising to cut costs to ensure continuity during this crisis. In this article, we will take a look at some of the factors behind the ongoing layoffs in data science. During the last five years, a lot of professionals had migrated towards data science from other job titles. We saw a lot of companies which started expanding their data science teams rapidly, but now many say the bubble has burst. One of the key aspects where data science and analytics teams are hired is automating, improving customer experience or making businesses processes more efficient. But experts say businesses didn’t have a clear vision in mind, and just had a vague idea on strategising data science. “The hype is still there, but it is slowly coming to the actual reality in terms of what is possible not, and what is not possible with data science. The last couple of years, the economy had been doing quite well and since every company wanted to join the AI race, they started pulling up these data science teams. But they didn’t do the due diligence in hiring. They didn’t have a clear vision in mind as in how their AI strategy is actually going to help,” says Dipanjan Sarkar, Data Science Lead at Applied Materials. Lack Of Funding For Data Science Projects Earlier, businesses could afford data science in terms of funding ensured by a race among businesses to leverage AI for different projects. Then, the pandemic happened which severely affected product sales, and getting funding for new AI projects became difficult. Companies have now started looking at the most important things, and the key business units which can deliver value, say experts. Businesses certainly have been closely evaluating a large chunk of their job functions, trying to cut costs and focus on essential positions. We can take a look at the recent layoffs at companies like Zomato, Uber, Airbnb, Lyft and many others around the world. Many of the laid-off staff had skills in advanced technologies, including AI\/data science. But, that’s no longer the only parameter for employment these days. The key is survival. “If we remove the insights which data scientists are bringing, like forecasting sales, the business won’t be critically affected. That is where the layoffs are happening. Data science is a thing which is nice to have for business innovation, but it can be selective as well. For businesses, the main thing is how long can you sustain their operations and pay salaries,” tells Dipanjan. Focus On Essential Job Functions Organisations around the world are now focussing on the key roles in the company which are related to the essential functions. Many say data science now isn’t an essential function, and that is where a lot of data scientists have been some of the first ones to get laid off. Dipanjan Sarkar explains, “The key thing is that data science as a position does not guarantee you job security, especially at this time unless you exhibit key skills and talent which is directly plugged into value creation. It doesn’t matter if you’re building a fancy deep learning model. You have to ask yourself how you can deliver value for the business.” According to experts, a lot of people have entered into data science, but that doesn’t make them data scientists without learning the core expertise. Even though there is a rush towards AI projects, particularly using open source tools, it is difficult to quantify the value of such data-driven innovation. This was seen recently during the COVID-19 forecasting models, most of which failed to deliver any real-world value in mitigating the pandemic. Businesses Management May Not Recognise The Value Of Data Science Coming under the brunt of the economy, companies have been trying their best to cut costs to the bare minimum. COVID-19 has led to changing priorities that require quick turnarounds on deriving insight from data to make sound business decisions. Financial institutions are re-assessing strategies across their business and resources are limited. “The perspective of data science jobs and the need for analytics is changing, and this will be critically analysed as we move forward. Where earlier 20-30 professionals were needed, in coming times, companies would hire 5-10 because much of the work related to data science will be automated,” tells Puesh Rajiv Ajmani, Global Head of Analytics & Insights at Square Panda AI or machine learning is just one small part of the whole business. In addition to that, data scientists may face resistance because they want to change the existing processes in the business. Unless you are very sure about having supportive management, many data science projects will remain POCs for a long time, say experts. This particularly holds true for many companies which rely on legacy infrastructure. “Companies may suspect that they’re not getting any tangible value from excessive data scientists. This can trigger a move towards cutting down the staff, which may be non-essential units. And that is where data scientist got hit really hard, especially in companies where data science is a supportive role or non-essential,” Dipanjan added. Experts say that for companies to survive in the next coming months, they need to be able to pay their employees. The most essential things like supply chain and logistics will be prioritised. Subhobroto Ghosh, Head – Data, Analytics & Actuarial at Allstate India tells, “We are at a major inflection point in the field of data science. While the pandemic will taper off, it will leave behind a very bad economy. So businesses are looking at containing costs, leading to layoffs.”","excerpt":"The global recession has engulfed the world of technology. In the field of data science, layoffs are happening across many companies, and analytics professionals have concerns staying relevant in these tough times. Alongside, businesses are also strategising to cut costs to ensure continuity during this crisis. In this article, we will take a look at […]","categories":["AI Features"],"tags":["data science layoff","Layoffs","what is data science"],"author_name":"Vishal Chawla","publish_date":"2020-05-23T13:00:00","publication_year":"2020","word_count":963,"keywords":["data science","Go","API","machine learning","data science layoff","Layoffs","AI","RAG","deep learning","analytics","GAN","what is data science","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","analytics","RAG","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-may-be-the-factors-behind-data-science-layoffs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10136054,"title":"Allie K Miller","content":"Allie K Miller is on a mission to help people understand AI as the CEO of Open Machine. Known for her work as  a product manager, advisor, investor, and influencer, she has held pivotal roles at IBM Watson and Amazon Web Services (AWS). Miller was the youngest woman to build an AI product at IBM, working in critical areas such as conversational AI, computer vision, and multi-modal AI. As the global head of Machine Learning Business Development for Startups and Venture Capital at AWS, she transformed her division into a 100-person, 10-figure organisation. Beyond her corporate achievements, Miller is a driving force in shaping the AI landscape. She has addressed global platforms like the European Commission, authored guidebooks on AI success, and earned numerous accolades, including being named Forbes’ “AI Innovator of the Year” and a LinkedIn Top Voice for Technology and AI. Miller’s commitment extends to promoting diversity and education in AI. She co-founded Girls of the Future, serves as a national ambassador for the American Association for the Advancement of Science (AAAS), and invests in machine learning startups. With an MBA from The Wharton School and a BA in Cognitive Science from Dartmouth College, Miller’s academic background complements her extensive practical experience in AI.","excerpt":"CEO, Open Machine","categories":["People"],"tags":["AI and machine learning"],"author_name":"Pabitra Moharana","publish_date":"2024-09-23T12:40:53","publication_year":"2024","word_count":206,"keywords":["API","machine learning","AWS","AI","Modal","venture capital","computer vision","AI and machine learning","GAN","R","startup"],"extracted_tech_keywords":["AI","machine learning","computer vision","AWS","R","API","GAN","startup","venture capital","Modal"],"url":"https:\/\/analyticsindiamag.com\/people\/allie-k-miller\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10051358,"title":"Google Earth Engine Now Open For Commercial Use","content":"Google recently announced its Google Earth Engine is now open for operational and commercial offering for select customers in preview as a part of its integration with Google Cloud Platform. The Google Earth Engine combines satellite imagery and geospatial data with powerful computing to help people and organizations understand how the planet is changing, how human activity contributes to those changes and what actions they can take. Organizations in the public sector and businesses can now use insights from Earth Engine to solve sustainability-related problems, such as building sustainable supply chains, committing to deforestation-free lending, preparing for recovery from weather-related events and reducing operational water use. This new offering puts over 50 petabytes of geospatial open data into the hands of business and government leaders. Google Cloud customers and partners can bring together earth observation data with their own data as well as other useful datasets, train models to analyze at scale, and derive meaningful insights about real-world impact. By combining Earth Engine’s powerful platform with Google Cloud’s distinctive data analytics tools and AI technology, we’re bringing the best of Google together. The new offering puts over 50 petabytes of geospatial open data into the hands of business and government leaders. Google Cloud customers and partners can now bring together earth observation data with their own data as well as other useful datasets, train models to analyze at scale, and derive meaningful insights about real-world impact. By combining Earth Engine’s powerful platform with Google Cloud’s distinctive data analytics tools and AI technology, it brings the best of Google together. To make sure businesses can make the most out of Google Earth Engine, the company is working with partners, like NGIS and Climate Engine, to help businesses identify and manage risks related to climate change.","excerpt":"The Google Earth Engine combines satellite imagery and geospatial data with powerful computing to help people and organizations understand how the planet is changing.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Cloud Computing","Cloud Platform","Data Science","Data Scientist","Deep Learning","Google","Google Research","Machine Learning"],"author_name":"Victor Dey","publish_date":"2021-10-13T17:16:10","publication_year":"2021","word_count":294,"keywords":["Go","programming_languages:R","AI","Machine Learning","programming_languages:Go","Cloud Computing","ViT","analytics","Google","cloud_platforms:Google Cloud","GAN","Deep Learning","Data Science","Data Scientist","R","Cloud Platform","AI (Artificial Intelligence)","Google Research"],"extracted_tech_keywords":["AI","analytics","R","Go","GAN","ViT","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-earth-engine-now-open-for-commercial-use\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":48134,"title":"Batch Norm Patent Granted To Google: Is AI Ownership The Gold Rush Of 21st Century?","content":"The machine learning community has witnessed a surge in releases of frameworks, libraries and software. Tech pioneers like Google, Amazon, Microsoft and others have insisted their intention behind open-sourcing their technology. However, there has been a growing trend of these tech giants claiming ownership for their innovations. According to the National Bureau of Economic Research study, in 2010, there were 145 US patent filings that mentioned machine learning, compared to 594 in 2016. Google, especially, has filed patents related to machine learning and neural networks 99 times in 2016 alone. via WIRED After getting rejecting initially, the US patent office, recently, has granted the ownership of batch normalisation to Google with an expiration date marked as 2038. Here is a timeline of the journey of the BatchNorm patent application: 2015-01-28 Priority to US201562108984P 2016-01-28 Application filed by Google LLC 2016-07-28 Publication of US20160217368A1 2019-09-17 Publication of US10417562B2 2019-09-17 Application granted 2019-10-15 Application status is Active 2038-01-01 Adjusted expiration Google does patent many of its products but there is a reason why batch normalisation patent gets all the attention. Why BatchNorm Touched A Nerve In the original BatchNorm paper, the authors, Sergey Ioffe and Christian Szegedy of Google introduced this method to address a phenomenon called internal covariate shift. This occurs because the distribution of each layer’s inputs changes during training, as the parameters of the previous layers change. This slows down the training by requiring lower learning rates and careful parameter initialisation. This makes training the models harder. The introduction of batch normalised networks helped achieve state-of-the-art accuracies with 14 times fewer training steps. This reduction in training duty led to the emergence of many improvements within the machine learning community. In spite of voicing for the democratisation of technology, companies like Google are rushing to claim ownership of advanced approaches in domains like AI. However, before jumping the bandwagon of critics, one should know that machine learning techniques like batch normalisation are a product of Google. The researchers working at Google used the resources available at Google to develop new frameworks to optimise machine learning applications. So, claiming ownership does sound right as it has been classified as a defensive patenting. One argument that is being made for Google is that it is better if the creators have the ownership rather than a patent troll claiming ownership and creating hurdles amidst lawsuits. However, a large half of the machine learning community is still sceptical of this consistent claim of ownership. Few members of the most happening machine learning forums like that of Reddit likened this whole event to a loaded gun that can backfire. Working Around This Trilemma There is no denying the fact that Google AI has been pioneering even before machine learning became a worldwide phenomenon. Their technology is being used to power up lives and promote growth. The use of BatchNorm has literally become a norm for many of the previous applications. This also is the reason for the rising scepticism amongst the practitioners. For instance, if a computer vision startup has used this technique to train one of its neural networks then should it be checking the mailbox for a probable lawsuit forever? That said, ever since the news broke out that Google has eyed batch normalisation technique, the alternative approaches to batch norm technique have gained traction. Here few methods that look promising: Fixup Initialisation: Fixed-update initialisation (Fixup) was aimed at solving the exploding and vanishing gradient problem at the beginning of training via properly rescaling a standard initialisation. Using Weight Normalisation: Weight normalisation accelerates the convergence of stochastic gradient descent optimisation by re-parameterising weight vectors in neural networks. General Hamming Network (GHN): The researchers at Nokia technologies in their work illustrated that the celebrated batch normalisation (BN) technique actually adapts the “normalised” bias such that it approximates the rightful bias induced by the generalised hamming distance. Group Normalisation (GN): GN divides the channels into groups and computes within each group the mean and variance for normalisation. GN’s computation is independent of batch sizes, and its accuracy is stable in a wide range of batch sizes. Switchable Normalisation (SN): Switchable Normalisation (SN) learns to select different normalisers for different normalisation layers of a deep neural network. Attentive Normalisation (AN): Attentive Normalisation(AN) is a novel and lightweight integration of feature normalisation and feature channel-wise attention. These are only a few of the many approaches that have surfaced in recent times and we can safely assume that there will be more coming up from this space.","excerpt":"The machine learning community has witnessed a surge in releases of frameworks, libraries and software. Tech pioneers like Google, Amazon, Microsoft and others have insisted their intention behind open-sourcing their technology. However, there has been a growing trend of these tech giants claiming ownership for their innovations. According to the National Bureau of Economic Research […]","categories":["Global Tech"],"tags":["Deep Learning","Google","Neural Networks","normalisation","Patent"],"author_name":"Ram Sagar","publish_date":"2019-10-17T13:00:02","publication_year":"2019","word_count":750,"keywords":["Go","machine learning","startup","AWS","AI","neural network","innovation","Patent","computer vision","normalisation","Aim","Google","Deep Learning","R","Neural Networks"],"extracted_tech_keywords":["AI","machine learning","neural network","computer vision","Aim","AWS","R","Go","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/batch-norm-patent-granted-to-google-machine-learning-ai-ownership-the-gold-rush-21st-century\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10080323,"title":"Activision Blizzard Leverages AI to Generate Music in Video Games","content":"Gaming has long been a hotbed of AI innovation, establishing precedents in fields like procedural content generation, pathfinding, decision-making algorithms, and human behaviour simulation. Now, it can add music generation to its arsenal. One of the world’s biggest developers, Activision Blizzard—known for games such as Call of Duty, Overwatch, and World of Warcraft—have recently published a patent that details a system that generates music unique to each player using artificial intelligence and machine learning algorithms. This system has the potential to not only change the way that composers approach music in video games, but also redefines immersion in video games by creating a unique auditory experience for every player. Before we delve into the specifics of how the proposed AI system works, let us explore some examples of certain algorithmic methods used to create unique experiences in games. Inside Activision Blizzard’s AI music play In the patent—filed under the US Patents Office in April of this year—team Activision Blizzard describes a system to dynamically generate and modulate music based on gaming events. As with any AI-based solution, the first step is to collect data on the player’s play patterns, profile, and performance, along with the music that was playing whenever the player engaged with the game, thus creating an activity–music pair. How the server will interact with the music generation module This data is then sent to a central server, which then places the player in one or more player profiles depending on the nature of the data collected. These player profiles comprise beginner, enthusiast, and expert levels of skill. The data is then used as a base to generate event data, which will depict the player’s engagement with ‘virtual elements’. This comprises a wide gamut of possible data points, including the player’s pace, their capacity to defeat enemies, and their general approach to interacting with the game. This event data is then used to classify the data into two or more event profiles. The event data is classified by the value of players’ engagement, ranging from low to high. A machine learning algorithm is then trained on these datasets, with the weights saved for later iterations of the model. Then, the algorithm is used to generate music by identifying the player’s mood based on the event profiles and player profiles—in theory creating a reactive music system that responds to both the user’s activity and their interaction with virtual elements in the game. An additional model then changes elements of the song such as beat, metre, tempo, chord progressions, loudness, duration, and more depending on the player profiles. This can result in a completely immersive experience for the user. Not only will the music generated be fitting for the events happening on the screen, which is the status quo today, the player’s approach to playing the game will also be taken into consideration while creating the music. This will create an impactful experience for each player while still being discrete from other players’ experiences. Even different players in the same server—in case of a multiplayer game—will have fitting music accompanying their in-game escapades. A brief history of music algorithms One of the first systems used for interactive music is iMUSE, created by video game developer LucasArts in the early 90s. The engine was born out of the frustration of the existing audio system in the SCUMM game development engine, and was first used in the game Monkey Island 2: LeChuck’s Revenge. The engine was relatively simple by modern standards, but set the precedent for how games approach music. The iMUSE system synchronised music to the player’s actions and would transition smoothly from one piece of music to another—a standout in an era when games commonly came bundled with fairly basic music which played on a loop. This concept then became a cornerstone of video game music, which then evolved to include concepts such as horizontal resequencing and vertical reorchestration. Resequencing is when pre-composed sections of music play according to the player’s action cues, such as their location or the activity they are undertaking. Whenever the scenario changes, the music engine either crossfades to the next segment—befitting of the new scenario—or changes the segment upon completion of the current musical phrase. For instance, the game can change the music being played if the player picks up a power-up and then reverts to the original soundtrack when the power-up wears off. Reorchestration, on the other hand, changes the mix of different instruments on top of a pre-existing ongoing loop of music. The player’s actions then determine the instrumentation of the soundtrack. For example, a game can have a certain track playing when the player is exploring, but amp up some components—like string instrumentations or percussion—when the player enters into combat. Many algorithms were written on the basis of these two concepts, creating varied solutions befitting the scope of the game and how players played them. One notable example of reorchestration is 2016’s hit game DOOM, which used a complex algorithm to make musical accompaniments to the player’s actions. To begin with, song sections were made into smaller parts, so as to let the algorithm pick the best part of songs to reflect the action on screen. Then, the player’s actions are closely monitored, with the music engine keeping a track of the player’s movement speed, number of enemies in combat, and the current activity being undertaken by the player i.e., exploration or combat. Speaking on the Doom soundtrack, Hugo Martin, Creative Director, iD Software said, “We ultimately struck a really good balance with hearing the things you need to hear from a gameplay perspective. . . but then ultimately making the player constantly feel like a badass. It crescendos beautifully, it accents every action that I do, it’s a rock concert.” Then, considering these factors, the algorithm picks the most fitting snippet of the song and blends it in seamlessly with the track that is already playing. The result is a soundtrack that complements the player’s actions perfectly and provides an unforgettable experience. Even as such advanced tech has already been used for modern games, Blizzard aims to push the envelope further and completely personalise the game experience for each player. AI-generated music might just be the tip of the iceberg when it comes to content in games. We are already seeing procedural generation take the mainstream in games, not only to create the music but also to create the world that the player is in. AI has also been leveraged to reduce system resource utilisation, as seen with NVIDIA’s DLSS technology. Using cutting-edge models will not only provide a value-add for games, but will create a bevy of new experiences that will leave an indelible mark on the new generation of gamers.","excerpt":"Activision Blizzard has filed a patent for video game music generation, and it might change the industry as we know it.","categories":["AI Features"],"tags":["AI Patent","Generative AI"],"author_name":"Anirudh VK","publish_date":"2022-11-22T10:00:00","publication_year":"2022","word_count":1121,"keywords":["Go","artificial intelligence","machine learning","AI","ML","AI Patent","RAG","Aim","ViT","Rust","Generative AI","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","R","Go","Rust","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/activision-blizzard-leverages-ai-to-generate-music-in-video-games\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":5531,"title":"iCreate Collaborates with SAP to Provide Banks with Better Analytics, Risk and Compliance Reporting","content":"21 March, Bangalore. Banking Decision Sciences pioneer iCreate today announced its plans to deliver enhanced enterprise information management capability to the global banking sector. In collaboration with SAP, iCreate will be able to integrate its banking solutions with the business intelligence technology from SAP. This initiative is aligned with the shared vision between iCreate and SAP to help banks run their businesses faster and better with informed decisions. Banks worldwide have been attempting to find the right solutions to address rapid growth and increasing compliance challenges. The changes in banking industry are driven by fluctuating customer needs and evolving regulatory environment globally, which urges the banks to leverage technology to make timely decisions through fast and accurate data. With its acclaimed banking decision enablement and risk & compliance expertise, iCreate intends to roll out innovative solutions spanning across the spectrum of Banking Analytics, MIS, Data Management and Risk & Compliance. Through the agreement, the solutions will integrate with SAP’s proven technology portfolio for banking and financial services to deliver a compelling value proposition to progressive banks. The portfolio mainly includes SAP® Sybase® IQ enterprise edition, SAP® BusinessObjects™ Business Intelligence platform and SAP Crystal Reports®. “It is part of our mission to provide banks worldwide with a superior decision making capability and this collaboration sets a new benchmark for the banking fraternity,” said Vivek Subramanyam, CEO, iCreate. “I am delighted that we can leverage SAP’s industry leading technology platform and R&D prowess to co-innovate best-in-class banking solutions. As a result of both companies’ leadership and focus in the banking domain, I am sure banks globally can now expect better information management and faster insights.” “Our innovative partnering models open up new opportunities for partners to leverage SAP’s strengths with their own core competencies to create new, compelling, and innovative products,” said Elbert Bailey, VP of Strategic Initiatives, Ecosystem and Channels, SAP Asia Pacific Japan, “Through their solutions embedding our proven technology platforms, we will work closely with iCreate to serve our joint banking customers – partnering to increase their customer-centricity, reduce cost and complexity, and better manage regulatory and risk compliance.” About iCreate Headquartered in Bangalore, India with offices in Asia Pacific, South Africa, the Middle East and Europe, iCreate is a global Banking Decision Sciences leader that works with banks worldwide to enable faster, better decision making. iCreate’s enterprise-grade BI, Analytics and Performance Management solutions integrate seamlessly with Banking systems and delivers a fully functional Decision Enablement System running in a fifth of the time when compared to conventional alternatives, at the lowest TCO and with advanced solutions in areas such as Retail Banking, Corporate Banking, Analytics 360, etc. iCreate currently has over 35 progressive financial services institutions across 11 nations as customers. iCreate is funded by venture capital majors Sequoia Capital and IDG Ventures to accelerate market expansion and boost product innovation efforts. www.icreate.in","excerpt":"21 March, Bangalore. Banking Decision Sciences pioneer iCreate today announced its plans to deliver enhanced enterprise information management capability to the global banking sector. In collaboration with SAP, iCreate will be able to integrate its banking solutions with the business intelligence technology from SAP. This initiative is aligned with the shared vision between iCreate and […]","categories":["AI News"],"tags":[],"author_name":"Fintellix","publish_date":"2014-03-21T14:21:56","publication_year":"2014","word_count":475,"keywords":["business intelligence","API","AI","innovation","ML","RAG","BERT","Aim","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","R","API","BERT","business intelligence","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/icreate-collaborates-with-sap-to-provide-banks-with-better-analytics-risk-and-compliance-reporting\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094851,"title":"How Wigzo by Shiprocket is Using Generative AI to Boost Sales for D2C Brands","content":"Online jewellery brand GIVA recently saw an increase of 55% in its orders and a 775% growth in its subscriber base. Another online lifestyle brand Leaf saw a 128% increase in revenue. The same goes for FabIndia, Snitch, Daily Objects, Libas and others that saw their online sales grow multifold in the past few months. One company that is revolutionising the data experience for D2C brands, and other new-age digital companies, the likes of 4700 BC, Quiltcomfort, FabHotels and Happyfares, is Wigzo by Shiprocket. The company claims to use automated user segmentation to help D2C brands convert, retain, and grow consumers. Founded in 2014 by Umair Mohammed, Atyab Mohammad and Himanshu Kaushik, Wigzo by Shiprocket provides data driven solutions for leading D2C (e-commerce) brands, especially apparel and fashion brands, with main focus on customer retention and marketing. Earlier this year, Shiprocket acquired 75% stake in the company which is already backed by the likes of SEA Fund, Aarin Capital and more. Analytics India Magazine got in touch with Umair Mohammed, founder and chief executive officer, and Atyab Mohammad, chief technology officer, to understand how they are helping brands boost sales and profits. “We aim to enable retailers to achieve a remarkable tenfold increase in sales by equipping them with comprehensive automation and customer engagement solutions,” Umair told AIM. Atyab added, “We believe, consumer data is the gold mine that can be used to reduce spending on customer acquisition and can be leveraged to target the right consumers.” Read more: How Data Analytics Fuels Shiprocket Wigzo identifies active customers and enables conversions by providing tailored recommendations. The company is at the forefront of leveraging AI and advanced analytics to revolutionise customer journeys. By harnessing the power of AI, Wigzo is able to delve deep into consumer behavior, enabling businesses to generate predictive insights and deliver personalised experiences. Wigzo by Shiprocket recently unveiled their AI-powered customer data platform (CDP), integrating retail, website, and mobile app data. This unified hub allows brands to access and leverage customer information, creating actionable segments and gaining valuable insights. It aims to establish the largest and most comprehensive database, showcasing impressive technical capabilities. Read more: Prompt Engineering is the New C++ Generative AI to the Rescue “We are also working at providing a comprehensive chat solution using generative AI, which will enable brands to automate their routine tasks of solving customer queries, tracking inventory, and monitoring marketing trends,” said Atyab. The company claims to enhance the reliability of its platform with cutting edge tech solutions by partnering with big techs Google and AWS and are leveraging models like Cassandra, MongoDB, and BigQuery, which are all scalable in nature. It is also working towards building AI tools to improve brands’ social media presence and offers automation for marketing. It also integrates with third-party apps like Fastrr, Razorpay, and checkout engines. Wigzo by Shiprocket also uses AI-enabled sentiment analysis to cull out insights from user-generated content across social media platforms. “Our aim is to simplify merchants’ lives by seamlessly integrating AI into their workflows and providing daily value,” shared Umair at their annual D2C Verse event in Bengaluru, last month. The tech team is also working on implementing OpenAI’s APIs to build solutions around content generation and data search. Merchants can now easily create messages and ask analytics questions without manual analysis. Wigzo by Shiprocket has an array of interesting tech-related announcements coming up. One of them is that it is set to launch their versatile omnichannel chatbot platform to address a wide range of brand needs, ranging from customer support to personalised recommendations. In recent times, e-commerce platforms have been playing with generative AI in their day-to-day operations, particularly in the areas of customisation, design, marketing and improving the customer journey. Meesho, an online lifestyle company, is actively experimenting with generative AI across various applications like recommendation systems, enhancing product cataloguing with more visually appealing representations, and providing virtual try-on experiences. Tapestry, a luxury fashion brand, is strengthening its marketing strategies by harnessing the power of generative AI to automate personalised online experiences. Marks & Spencer is also closely following this trend. Following suit, Coca-Cola underwent a marketing transformation through generative AI, facilitated by the partnership between OpenAI and Bain & Co. Over time, the adoption rate of AI by D2C brands is only going to increase for the good. Read more: Indian Edtech, Why So Toxic?","excerpt":"With more than 300 users on board, the company uses automated user segmentation to help D2C brands convert, retain, and grow consumers","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Shritama Saha","publish_date":"2023-06-09T16:06:38","publication_year":"2023","word_count":729,"keywords":["OpenAI","AI","ML","recommendation systems","RAG","Ray","Aim","prompt engineering","generative AI","analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","OpenAI","Aim","Ray","RAG","prompt engineering","recommendation systems"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-wigzo-by-shiprocket-is-using-generative-ai-to-boost-d2c-brands-sales-customer-experience\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":36503,"title":"How Companies Are Turning AI\/ML Research For Building A Product Pipeline","content":"For every company to launch new products in the market, there is something called a product pipeline. A product pipeline is basically a series of products, either in a state of development, preparation, or production developed and sold by a company. The complete process is like a funnel and consists of different phases or stages. For example, there is a company that conducts a brainstorming session to launch new products in the market and they come up with 20 ideas. Now, the ideas go the researchers or scientists and they filter out the ideas which the team thinks may not work in the market well. Therefore, like this, there would be other bodies in an organisation too who would play a vital part and finally, in the end, there are possibilities that out of all the 20 ideas, more than 10 gets rejected and only 5 goes out in the market. This how a product pipeline works. There are many businesses across the world that are facing problems turning ideas into business growth and are looking for new technologies and innovative ideas charged with bringing new products to new markets. Adobe’s AI and ML-Based Approach to Product Pipeline AI and ML have evolved tremendously over the past few years and have marked its territory in several verticals. Be it a recommendation engine of a shopping website or a dating app, ML and AI are everywhere. In today’s tech-driven era, if you don’t have a system or the latest technologies to turn raw technology into experiences that make a difference for businesses and consumers, there are chances that you would be left out. With so much happening already, experts from the industry believe that these sought after technologies can also solve the pain-points in the product pipeline. Talking about leveraging the superpowers of AI and ML in the product pipeline, Adobe is a great example. According to a source, last years, Adobe revealed how it approaches the product pipeline using ML and AI. And from that, few vital things have come onto the surface and have the capability to reduce the challenges in the product pipeline. When we talk about the product pipeline — from product suggestion to product launch — the question that revolves around is “Is this product good enough to bring value to the customers?” So, Adobe came up with some new AI-driven technologies that collaborate across functions. Intelligent Forecasting: It is a conceptual feature that allows customers to address business metric shortfalls from real-time data. Also, it aims to crunch billions of data points to optimize operations for online retailers, companies looking to improve conversion rates, and manage seasonal slowdowns Predictive Pathing: It allows customers to spot things such as app installs and emerging problem areas. It also analyzes the paths a customer takes across screens Automated segmentation: This feature is about managing audiences and customer bases across ages and demographics. Here, AI and machine learning automatically segment audiences Analysis Workspace Assistant: Last but not the least, it is an AI tool that takes a question and spreads across historical queries in order to make sure that work doesn’t get replicated. The assistant is designed to improve over time Outlook It is obviously going to be time-consuming when a process depends on multiple bodies. But with the emergence of  AI and ML in the product pipeline, things might transform. Also, time and again, these two sought after technologies have proved that they are here to stay and make an impact, but not to fade. Therefore, in order to reduce the time frame of the process of product-suggestion-to-product-launch, try to find ways to identify and surface data that is collected but not seen. However, if the data is huge, it would require an ML and AI to surface data that wouldn’t be seen with the naked eye and then humans can optimize. This will not only give a clear view of what the customer wants or what they are interested in but will also help in the very first phase of product pipeline — brainstorming. So, use the latest technologies to understand the customers and the challenges they are facing. This is the ultimate solution for any business.","excerpt":"For every company to launch new products in the market, there is something called a product pipeline. A product pipeline is basically a series of products, either in a state of development, preparation, or production developed and sold by a company. The complete process is like a funnel and consists of different phases or stages. For example, there […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI Companies","Machine Learning","ML"],"author_name":"Harshajit Sarmah","publish_date":"2019-03-18T12:16:48","publication_year":"2019","word_count":700,"keywords":["Replicate","Go","machine learning","AI","ML","Machine Learning","RAG","Aim","ViT","AI Companies","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","RAG","R","Go","GAN","ViT","Replicate"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-companies-are-turning-ai-ml-research-for-building-a-product-pipeline\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":34439,"title":"Swiggy Acquires Kint.IO, A Bengaluru-Based AI Startup","content":"In a bid to strengthen their foothold in the consumer app market, the online ordering and delivery service Swiggy acquired Kint.IO, a Bengaluru-based AI start-up that applies deep learning and computer vision to object recognition in videos. With clever imaging and computer vision touted to drive the future of consumer apps, the acquisition can be seen as a strategic move by Swiggy to consolidate its position in the highly-competitive food delivery market in India Kint.IO which was founded in 2014, will join Swiggy to boost its computer vision technology and superior consumer experience. Further, by welcoming both the founding members of Kint.IO into their team, Swiggy hopes to strengthen its expertise in the field of  AI, machine learning, big data systems. “The team at Kint.IO comes with an exceptional understanding and expertise in AI, machine learning and data sciences. This acquihire is part of Swiggy’s strategy to scale our tech prowess by bringing in entrepreneurial teams that can solve unique customer problems while leveraging the network and resources at Swiggy. We provide a unique mix of strong entrepreneurial DNA and professional leadership that gives startup teams the ownership and leverage to move fast and make a big impact,” said Dale Vaz, Head of Engineering and Data Sciences, Swiggy. This will be the second acquisition since Vaz took charge in July last year after his brief stint with Amazon India. In August 2018, under his watchful eyes, Swiggy acquired on-demand delivery startup Scootsy for Rs 50 crore. Speaking on the development, Pavithra Solai Jawahar and Jagannathan Veeraraghavan, Co-founders of Kint.IO said, “AI research has leapfrogged this past year but lack of data, cultural biases, inability to adapt to our diversity, has somehow always pulled us back when it comes to applying AI to India-based problems, effectively. This is where Swiggy left us stumped. We were impressed by the team’s razor-focussed mindset to bring ingenious solutions for problems unique to India. We are confident that this partnership can unlock immense possibilities. This is a great opportunity for us to show scale and address a-billion-people problems through AI. We look forward to interesting innings with Swiggy.”","excerpt":"In a bid to strengthen their foothold in the consumer app market, the online ordering and delivery service Swiggy acquired Kint.IO, a Bengaluru-based AI start-up that applies deep learning and computer vision to object recognition in videos. With clever imaging and computer vision touted to drive the future of consumer apps, the acquisition can be […]","categories":["AI News"],"tags":["Mergers and Acquisitions","swiggy"],"author_name":"Akshaya Asokan","publish_date":"2019-02-04T11:11:03","publication_year":"2019","word_count":353,"keywords":["big data","data science","machine learning","AI","computer vision","RAG","deep learning","ViT","GAN","swiggy","R","Mergers and Acquisitions"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","data science","RAG","R","big data","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/swiggy-acquires-kint-io-a-bengaluru-based-ai-startup\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018608,"title":"Explained: Tech Behind Facebook’s ML Models For COVID Prognosis","content":"Though it’s been a year into the COVID outbreak, the researchers, healthcare workers, and hospital staff are still struggling to contain the situation. Not only has it been a challenge to make accurate predictions of the course of the disease, but the pandemic has also put a strain on hospitals’ resources. To address this, Facebook AI has recently introduced pre-trained machine learning (ML) models to help doctors project the prognosis of COVID patients to make effective clinical decisions and allocate resources. The research is a part of an ongoing collaboration with NYU Langone Health’s Predictive Analytics Unit and Department of Radiology, where the ML models are used to predict patient deterioration using X-ray radiographs. In a recent paper, Facebook AI researchers showed how self-supervised models based on the momentum contrast (MoCo) method could help in learning more general image representations for downstream tasks. The experiment proved the ML model (using sequential chest X-ray images) can predict the deterioration of COVID patients up to 96 hours with the highest accuracy. The researchers believe a model like this can help healthcare providers to predict the demand for the resources that would be critical to deal with high-risk patients. Also Read: Are Easy-To-Interpret Neurons Necessary? New Findings By Facebook AI Tech Explained Deep learning methods for image-based diagnosis using supervised training have been a standard in predicting the risk of deterioration in COVID patients. However, such methods come with several constraints like the collection of labelled data, expensive training etc. Also, experts believe that large datasets often do not capture data from emerging diseases, such as COVID-19, and thus can restrict the use of deep learning methods. That is why the Facebook AI team relied on a new self-supervised Momentum Contrast technique to generate accurate representations of images for classification. This allowed the group to achieve feature extraction independent of labels or tasks associated with the pre-training dataset, wrote the authors. Diagram for momentum contrast training. For the experiment, the team pre-trained a model using MoCo on two large public chest X-ray datasets — MIMIC-CXR and CheXpert with more than five lakhs chest X-ray radiographs. The pre-trained model was then used to build classifiers for predicting the clinical deterioration of COVID patients. To evaluate this approach’s effectiveness, the team applied it on three downstream tasks — adverse event prediction from single images; oxygen requirements prediction from single images; and adverse event prediction from multiple images. The team noted that with DenseNet architecture, the self-supervised training achieves higher performance accuracy for predicting an adverse event at all time points. Model outputs According to the paper, the team used the NYU COVID dataset with over 26 thousand X-ray radiographs from 4,914 patients for fine-tuning the model. The labelled data showed whether the patient’s condition deteriorated within 24, 48, 72, or 96 hours of the scan, wrote the authors. The results The team further built two kinds of classifiers for COVID deterioration tasks — one that uses a single X-ray, and another that uses a sequence of X-rays by aggregating the image features via a Transformer model. According to the results, while the first model achieved an accuracy of 0.742 for predicting an adverse event within 96 hours, the new transformer-based architecture attained an accuracy of 0.786 to predict an adverse event at 96 hours. Also Read: Tech Behind Facebook AI’s Latest Technique To Train Computer Vision Models Wrapping Up The COVID pandemic has massively increased healthcare providers’ need to understand COVID patients’ prognosis to make effective clinical decisions. With the uncertainty involved with this deadly disease, it has been challenging to develop machine learning models for predicting its risks. A lot of this could be attributed to the difficulty in gathering large datasets by most medical centres. As a result, Facebook AI decided to address this problem using self-supervised contrastive loss pre-training. Facebook AI is also open-sourcing their pre-trained models for other researchers and healthcare providers to fine-tune their own X-ray datasets, and hopes to assist the broader community with resource planning. Read the paper here.","excerpt":"Though it’s been a year into the COVID outbreak, the researchers, healthcare workers, and hospital staff are still struggling to contain the situation. Not only has it been a challenge to make accurate predictions of the course of the disease, but the pandemic has also put a strain on hospitals’ resources. To address this, Facebook […]","categories":["AI Features"],"tags":[],"author_name":"Sejuti Das","publish_date":"2021-01-21T15:00:00","publication_year":"2021","word_count":671,"keywords":["machine learning","TPU","AI","R","ML","computer vision","Ray","deep learning","analytics","predictive analytics"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","analytics","Ray","predictive analytics","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/explained-tech-behind-facebooks-ml-models-for-covid-prognosis\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10170782,"title":"‘Downloads isn’t Rooted in What Works in India, Sarvam Grew 10x in a Month,’ Says Lightspeed Investor","content":"Following days of criticism and debate over Sarvam AI’s newly launched Indic LLM, Sarvam-M, Hemant Mohapatra, partner at Lightspeed India, an investor in Sarvam AI, has addressed the backlash, highlighting the company’s scale, growth, and strategic direction. In a post on X, Mohapatra acknowledged the global attention Sarvam received in the past week, saying, “We appreciate the attention and also understand the responsibility.” He noted that while criticism is expected, much of it stems from a misunderstanding of the startup’s long-term goals. Our portco @SarvamAI got a lot of global eyeballs this week – we appreciate the attention and also understand the responsibility. What is clear is (1) most folks really don't understand what strategy we are going for and what matters in that strategy (2) when the timing is right…— Hemant Mohapatra (@MohapatraHemant) May 26, 2025 “Most folks really don’t understand what strategy we are going for and what matters in that strategy,” Mohapatra wrote. He emphasised that Sarvam AI’s approach is deeply rooted in India-specific dynamics and long-term boardroom-level transformation projects. “There is absolutely a huge demand for what we are building but looking at it from the generic lens of downloads isn’t rooted in what really works in India,” he said. Sarvam, according to Mohapatra, grew 10x last month from an already significant base and ran one of India’s largest population-scale AI campaigns, reaching over 4% of the adult population. Responding to online comments suggesting that Sarvam should build hype like other LLM players, Mohapatra dismissed the idea of clout-driven validation. “Most of X lives at the peak of the bell curve and clout mongering isn’t a path to gain respect anyway,” replied a user on X. The remarks follow days of social media criticism, sparked in part by investor Deedy Das, who compared the Sarvam-M model to one built by Korean college students. Read: Sarvam AI’s Backlash Exposes the Sad State of Indian AI Das’ comments prompted defense from Sarvam’s team and supporters, including Zoho CEO Sridhar Vembu, who argued that instant success isn’t essential for long-term impact. Sarvam-M currently has 1.2k downloads on Hugging Face. Earlier in this ongoing debate, Sarvam co-founder Pratyush Kumar expressed optimism in a post on X. “Great to be receiving feedback on Sarvam-M. Please keep them coming. Will help strengthen our pipelines as we start to train our sovereign model,” he said, confirming that Sarvam is on the path of building a foundational LLM under the IndiaAI Mission.","excerpt":"Hemant Mohapatra said that Sarvam grew 10x last month from an already significant base and ran one of India’s largest population-scale AI campaigns, reaching over 4% of the adult population.","categories":["AI News"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-05-26T22:07:41","publication_year":"2025","word_count":408,"keywords":["Go","Hugging Face","programming_languages:R","AI","data_tools:Spark","programming_languages:Go","ai_frameworks:Hugging Face","R","startup"],"extracted_tech_keywords":["AI","Hugging Face","R","Go","startup","ai_frameworks:Hugging Face","programming_languages:R","programming_languages:Go","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/downloads-isnt-rooted-in-what-works-in-india-sarvam-grew-10x-in-a-month-says-lightspeed-investor\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10123844,"title":"Bill Gates Advises Indians to Build AI Models on Google, Microsoft and OpenAI","content":"In a recent episode of Nikhil Kamath’s podcast, Bill Gates spoke all about AI and its equation in India for entrepreneurs. “I’d probably build some AI thing and just use the Google Microsoft platforms and go on top of that and try,” he said, suggesting young entrepreneurs to use established AI platforms to build applications on top of them. Gates highlighted that by using these robust AI platforms, Indian entrepreneurs can focus more on developing specific applications rather than spending significant resources on creating the foundational AI technology themselves. Solves Compute, Fosters Collaboration Considering how developing foundational AI models from scratch requires immense capital and technical expertise, Gates suggested that by leveraging existing platforms, entrepreneurs can bypass these hurdles, making their ventures more resource-efficient. Gates also believes that integrating with AI platforms from major tech companies can lead to valuable collaborations and growth opportunities. These tech giants often have programs and networks that support startups, providing mentorship and potential funding. However, when asked about the timeline for achieving AGI, Gates seemed to smartly answer the question without revealing anything. He instead spoke about productivity benefits about AI. AGI is Productivity? When asked about a timeline for AGI, Gates spoke about productivity improved with AI. Gates pointed out that AI, even in its current state, is already driving significant productivity gains by automating repetitive and time-consuming tasks. He suggested that as AI technology evolves, these gains will only increase. A Youtube user said, “He [Gates] was not willing to reveal the secrets of AGI and future technology prospects to Indians. So he was beating around the bush throughout the conversation.” Gates might have not given anything conclusive about AGI, however other tech leaders such as Elon Musk predict AGI sooner than ever. He even predicted a short timespan of two years, or possibly even one, for achieving the same. AGI is Here? In AIM’s recent interaction with GitHub CEO, Thomas Dohmke said, “Today I see no sign that machine learning models or LLMs have sentience. They are not creative. They are machines created by us that help us with the things that we want to do or don’t want to do.” Dohmke highlighted how people have a different understanding of AGI, and even doesn’t understand what the ‘G’ in AGI truly stands for. Resonating a similar thought, NVIDIA chief Jensen Huang believes that in five years, we would be able to hit AGI, but it becomes important to define AGI.","excerpt":"“To have a product breakthrough, you can just build on top of what they’ve done,” said Bill Gates.","categories":["AI News"],"tags":["AGI","AI","bill gates"],"author_name":"Vandana Nair","publish_date":"2024-06-17T19:02:34","publication_year":"2024","word_count":409,"keywords":["Go","API","machine learning","AI","Git","RAG","Aim","ViT","AGI","bill gates","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","R","Go","Git","GitHub","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bill-gates-advises-indians-to-build-ai-models-on-google-microsoft-and-openai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10140360,"title":"Losing Touch with the Metaverse, Meta Turns to Robotics","content":"Earlier this year, when Figure CEO Brett Adcock referred to 2024 as the year of embodied AI, few could predict the extraordinary advancements in robotics, pushing the boundaries of what once seemed implausible. Last week, an unexpected contender released research updates that advance robotics to a new level. Meta’s Fundamental AI Research (FAIR) released three new research artefacts that advance touch perception, robot dexterity and human-robot interaction, namely Meta Sparsh, Meta Digit 360, and Meta Digit Plexus. Source: Meta Meta Sparsh, derived from the Sanskrit word for ‘touch’, is the first general-purpose encoder for vision-based tactile sensing. The technology aims to integrate robots with the sense of touch, thereby addressing a crucial modality to interact with the world. Sparsh operates across various types of vision-based tactile sensors and tasks that use self-supervised learning, eliminating the need for labelled data. It consists of a family of models pre-trained on an extensive dataset of over 4,60,000 tactile images. Meta has also released Meta Digit 360, a tactile fingertip featuring human-level multimodal sensing capabilities. Next up is Meta Digit Plexus, a platform paired with Digit 360, that integrates various tactile sensors into a single robotic arm. Meta FAIR mentioned in their blog that these new artefacts “advance robotics and support Meta’s goal of reaching advanced machine intelligence (AMI)”. AMI, also referred to as autonomous machine intelligence, is an innovation of Meta’s AI chief scientist Yann LeCun. The technology is envisioned to help machines assist people in their daily lives. LeCun has proposed a future where systems can understand cause and effect and model the physical world. Interestingly, Meta’s advancements in robotics are intended to help the whole ecosystem of builders develop machines that understand the world. However, robotics is not new for the company. Meta’s Robotics Dream Takes Shape Meta’s robotics developments have largely revolved around Metaverse and AR\/VR sets that leverage AI. Two years ago, at the ‘Meta AI: Inside the Lab’ event, the company highlighted AI and robotics developments that are central to creating the metaverse, largely to bring immersive virtual experiences. Features such as Builderbot, Universal Speech Translator, and others aim to enrich the metaverse experience. Notably, the former head of Meta hardware, Caitlin Kalinowski, who headed the development of Orion’s augmented reality glasses at Meta, recently joined OpenAI to lead robotics and consumer hardware. With the recent announcements on tactile sensing innovations, Meta seems to be taking robotics to the next level. Furthermore, by open-sourcing these new models, Meta is continuing its run of enabling individuals and companies to grow in the open-source community. In the process, they are also attempting to take on NVIDIA. Open Source for Robotics The graphics processing unit (GPU) giant has also been making significant strides in robotics. NVIDIA Omniverse and digital twins have been powering several domains, including automobile, semiconductor and healthcare. NVIDIA’s Project GR00T, which was released earlier this year, is a new foundation model that aids the development of humanoid robots. Just last week the company released two updates in robotics. NVIDIA, along with researchers from the University of California, Berkeley, Carnegie Mellon University, and other universities, released HOVER (humanoid versatile controller), a 1.5-million-parameter neural network, to control the body of a humanoid robot. The model is said to improve efficiency and flexibility for humanoid applications. To further accelerate robotics developments, NVIDIA even released DexMimic Gen, a large-scale synthetic data generator that allows humanoids to learn complex skills from very few human demonstrations. This effectively reduces the time required for training robots, considering real-world data collection is one of the biggest hurdles in the humanoid development process. “At NVIDIA, we believe the majority of high-quality tokens for robot foundation models will come from simulation,” said Jim Fan, senior research manager and lead of Embodied AI at NVIDIA. With these many advancements, it is evident that more companies are increasingly expressing interest in robotics. OpenAI is seemingly positioning itself for the future with the recent appointment of a new leader from Meta. It won’t be too surprising if, tomorrow, Meta or OpenAI release a robot that embodies all senses. Now that robots can hear, see, think, move, and touch, the only thing left is smell. Considering how a company is already building technology to teleport smell, it won’t be a surprise if robots get equipped with this capability in the future!","excerpt":"Meta’s three new research artefacts bring touch, dexterity and interaction to robots.","categories":["AI Features"],"tags":["AI","Embodied AI","figure","Meta","NVIDIA","Touch perception"],"author_name":"Vandana Nair","publish_date":"2024-11-06T16:00:00","publication_year":"2024","word_count":719,"keywords":["Go","Meta","Meta AI","OpenAI","AI","Touch perception","neural network","Embodied AI","Git","RAG","Aim","figure","foundation models","NVIDIA","R"],"extracted_tech_keywords":["AI","neural network","foundation models","OpenAI","Meta AI","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/losing-touch-with-the-metaverse-meta-turns-to-robotics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10090622,"title":"The Immortal Soul Behind Google Search","content":"In 1998, when Sergey Brin and Larry Page, two computer scientists and PhD students in Stanford, were working on creating what would be one of the biggest companies in the world, an Indian professor helped mentor them. He worked alongside them and co-authored a research paper on the “page rank” algorithm, a criteria still relevant in Google’s search engine. Computer Science Professor Rajeev Motwani was not only known for mentoring the Silicon Valley entrepreneurs and backing Google and PayPal during their early days, but was also known for his research work across fields such as computing, robotics and medicine. Rajeev Motwani who tragically died at the young age of 47, was recently commemorated on his birth anniversary on March 24. Rajeev Motwani has worked on a number of research papers from late 1990s to early 2000s, wherein his research work laid out foundations for algorithms that were used in future applications. Motwani, a theoretical computer scientist, researched and published works on different areas of computing and algorithms, and even medicine. He has 110,860 citations and 231 publications to his name. Google Days Rajeev Motwani, a mentor and guide, has closely worked with Google co-founders Page and Brin. The three of them have worked on multiple papers, the most notable one titled, ‘The PageRank Citation Ranking: Bringing Order to the Web’ in 1998. Along with Terry Winograd, who was another professor at Stanford, they developed the PageRank algorithm that is a method used by Google to rank web pages in its search engine results. PageRank is determined by the number and quality of web pages that link to a site. A concept that was developed during the early stages of Backrub, which later became ‘Google’, is still relevant for the search engine. In 1997 and 1998, Motwani co-authored two other papers with Brin, titled, ‘Dynamic itemset counting and implication rules for market basket data’ and ‘Beyond market baskets: Generalizing association rules to correlations’, respectively. Both the papers talk about association rule mining, a concept used for analysing large datasets to find patterns. It is widely used across fields where data is analysed for retail market, healthcare, web mining, and more. Sergey Brin reminisced about Motwani after the latter’s untimely demise in 2009. He considered him a “brilliant computer scientist” and a “teacher and good friend”. “His legacy and personality live on in the students, projects, and companies he has touched,” said Brin. Motwani’s Priceless Contributions One of Motwani’s recognised works is Probabilistically Checkable Proofs (PCP) Theorem. In collaboration with Sanjeev Arora, Carsten Lund, Madhu Sudan and Mario Szegedy, the proposed paper on PCP Theorem has wide applications in theoretical computer science. The authors of this paper were awarded with the Gödel Prize in 2001. The award is considered the most prestigious award given in theoretical computer science. The award is given by the European Association for Theoretical Computer Science (EATCS) and the Association for Computing Machinery Special Interest Group on Algorithms and Computational Theory (ACM SIGACT) . The PCP theorem is a mathematical result that states that any problem that can be solved using a computer programme that tries many different solutions—a non-deterministic algorithm—can also be checked using a random sample of the solution rather than checking the whole thing. The PCP theorem has had a significant impact on many areas of computer science and related fields and continues to be an active area of research with practical applications. At present, the PCP theorem provides a foundation for secure computation of functions on private data which is useful for applications, such as online transactions and data sharing. In software development, the PCP theorem enables automated checking of programme correctness and, in communication systems, the PCP theorem aids in designing reliable methods for transmitting data over noisy channels. Algorithmic Combinatorics and Graph Theory ‘Constructive results from graph minors: Linkless embeddings’ is Motwani’s paper, co-authored with Arvind Raghunathan and Huzur Saran, was published in the Journal of the Association for Computing Machinery (ACM) in 1995. It focussed on constructive results related to “graph minor theorem”. The paper is cited in the field of graph theory, particularly in the study of linkless embeddings and graph minors. The paper has been used in various applications and some of the practical examples where this paper is used are in circuit layout designs, network visualisation, network routing, and algorithmic designs in computer vision, computational geometry, and computer graphics. Robotics and Medicine Dating back to 1997, Rajeev Motwani’s interests in geometric computation drove him to contribute towards robotics. Motwani, along with his researchers, released a paper on the Robot Localization Problem in two dimensions. The paper proposed data structures for processing robot localization queries by using a probabilistic approach based on the Monte Carlo method. Future applications of research on robot localization queries are varied. It is used in autonomous vehicles where precise localization is required for accurate navigation as well as in other fields such as augmented reality, robot utilisation industries, and more. Motwani, along with a team of researchers, published a paper titled, ‘RAPID: Randomized Pharmacophore Identification for Drug Design’, in the Journal of Medicinal Chemistry in 2002. RAPID is a method for identifying essential structural and chemical features—called pharmacophores—of a molecule that is responsible for its biological activity. This method has the ability to handle large datasets and generate pharmacophores specific to a particular target or class of compounds. The RAPID method is adopted in the field of drug discovery and is used by pharmaceutical companies. Mentor and Guide Motwani was also an active member in student communities. In Stanford, he assisted entrepreneurial student groups, such as BASES and Stanford Student Enterprises. Being an active member in the venture industry, he offered strategic guidance to teams, investors and entrepreneurs. As an advisor to Deutsche Bank, he was one of the first investors in PayPal. In 2000, Rajeev and his wife Asha Jadeja Motwani started a venture fund called Dot Edu Ventures that invested in early-stage technology businesses. He was also a special advisor to Sequoia Capital. Rajeev also served as a mentor and board member for IIT Kanpur’s research programme with the Department of Computer Science and Engineering. Early Life Rajeev Motwani graduated from IIT Kanpur in 1983 as one of the first batches of Computer Science students. In 1983, Rajeev joined University of California, Berkeley, to pursue his PhD. He was advised by Professor Richard Karp, computer scientist and computational theorist who won the prestigious Turing award in 1985. In 1988, towards the end of his PhD, at Don Knuth’s request, he joined as a faculty in Stanford. Motwani’s famous courses at Stanford include, ‘Automata and Complexity Theory’—a field of computer science that studies abstract machines used for solving computational problems—‘Randomized Algorithms’ and ‘Advanced Data Structures and Algorithms’. Rajeev Motwani also went on to become the Director of Graduate Studies in the Computer Science Department at Stanford.","excerpt":"2001 Gödel Prize winner Indian scientist was behind some of the leading research algorithms that are still used today.","categories":["AI Features"],"tags":["berkeley","computer science","Google","larry page","PhD","Stanford","University of California"],"author_name":"Vandana Nair","publish_date":"2023-04-03T13:02:20","publication_year":"2023","word_count":1143,"keywords":["computer science","Go","API","PhD","larry page","University of California","AI","programming_languages:R","programming_languages:Go","computer vision","RAG","berkeley","ViT","Google","Stanford","R","ai_applications:computer vision"],"extracted_tech_keywords":["AI","computer vision","RAG","R","Go","API","ViT","programming_languages:R","programming_languages:Go","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-immortal-soul-behind-google-search\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":24731,"title":"Demystifying Crypto Assets: A Digital Barter System","content":"Ethereum cryptocurrency in hand, blockchain technology decentralized currency coin In today’s world, crypto assets are changing the way businesses need to be built. Instead of getting the equity financing from founders and accredited investors, entrepreneurs are using ICO (Initial Coin Offering) financing route. In the crypto token version of business, marketplace value is captured by all those who create value for the marketplace, including marketplace “users”, and that value is captured without any transaction fees. This model eliminates the middleman. The oligopoly of platforms like Uber, Airbnb, Amazon, etc. will face a challenging time in coming days. Demystifying Crypto Assets Crypto assets are digital assets which utilize cryptography, peer to peer (P2P) networking and a public ledger to regulate the generation of new units, verify the transactions and secure the transactions without the intervention of any middleman. Crypto assets work exclusively on the internet by using a network of computers that lend their processing power to verify and register all the transactions made. In return for their work, computers are rewarded with a payment in the form of tokens. The system that allows for this to happen is known as the blockchain, and it is the fundamental force behind any crypto asset. The blockchain is formed by blocks and each block is a segment of the chain that holds the Ledger of transactions made with crypto assets. The decentralization of economy along with avoidance of taxes, inflation, and local incertitude, implies that anyone can save their money in tokens and withdraw it on a need basis. This is the reason why crypto assets are getting popular. One of the most important aspects of crypto assets is their supply. The supply of a crypto asset is often the determinant of its utility and value. While some have a finite total supply, many are launched with infinite total supply. A crypto asset’s supply can be divided into three types: Circulating Supply: It’s the amount of coin available at present time and circulating in the market. Total Supply: The total supply is the amount of coin that is already in existence (whether in circulation or not) Maximum Supply: It is the maximum amount of coin that will ever exist for a particular cryptoasset. Types of Crypto assets Crypto assets can be divided into 4 categories based on the characteristics they possess. Four type of crypto assets are Cryptocurrencies, Platform tokens, Utility tokens & Transactional\/ Payment tokens: Cryptocurrencies: Well-known native blockchain assets with only the basic characteristics as those of a fiat currency but are decentralized. The only purpose of these crypto assets is to act as money or digital currency which offer a more secure and decentralized experience to the users. Popular cryptocurrencies are Bitcoin, Monero, Litecoin, etc. Platform Tokens: Platform tokens are designed to act as a platform for other decentralized projects. Popular platform tokens are Ethereum, EOS, NEO, etc. Utility Tokens: Utility tokens (also called protocol tokens) denotes decentralized services or units of decentralized services that can be bought, sold, and earned. Just like any real-life service token, these tokens can be exchanged for specific services like distributed storage, video game currency, etc. These tokens are developed with a specific use case in mind and acts as an API key which is used to access a specific service. Some of the popular utility tokens are Golem, Augur, Gnosis, DigixDAO, etc. Transactional Tokens\/Payments Tokens Financial institutions and banks are still very slow when it comes to cross-border payments. They depend on year’s old technology facilitated by SWIFT which involves multiple parties and KYC and AML requirements. Transactional tokens are crypto assets launched to solve this inter-border transactional problem. One of the great examples of transactional crypto assets is Ripple. Ripple is a global real time settlement network which connects different banks and financial institutions around the world without requiring beneficiaries or other intermediaries. It enables cross border payments within banks within seconds while providing them an end to end visibility throughout the whole process. Do you know crypto assets are already disrupting traditional business models? AdEx – Advertisement exchange AdEx is a decentralized ad exchange built on Ethereum blockchain and smart contracts. The AdEx platform is designed to disrupt and replace the traditional digital advertising models by providing a transparent, focused solution for advertisers to collaborate with ad publishers and reach the best potential clients. Major ad serving networks and exchanges operate with huge amounts of centralized data that can easily be traced back to the consumers’ identities. With the help of blockchain, advertisers still get to receive and process data about their target audiences and consumers in the form of statistics only – without compromising the privacy of these consumers. SingularDTV – Film\/ television studio & distribution portal SingularDT is a decentralized system in which artists and digital content creators can build, monetize, protect and manage their creations using the blockchain technology. Musicoin – Music platform Musicoin is designed to support the creation, distribution and consumption of music in a cycle. It enable musicians and consumers to exchange value in a frictionless environment. With a unique currency and a solid peer-2-peer (P2P) contract system. Golem – Decentralized supercomputer Golem connects computers in a peer-to-peer network, enabling both application owners and individual users (“requestors”) to rent resources of other users’ (“providers”) machines. Today, such resources are supplied by centralized cloud providers which, are constrained by closed networks, proprietary payment systems, and hard-coded provisioning operations. Also core to Golem’s built-in feature set is a dedicated Ethereum-based transaction system, which enables direct payments between requestors, providers, and software developers. iExec – Cloud computing platform iExec aims at providing decentralized applications running on the blockchain a scalable, secure and easy access to the services, data-sets and computing resources they need. This technology relies on Ethereum smart contracts and allows the building of a virtual cloud infrastructure that provides high-performance computing services on demand. A2B Taxi – Taxi service platform With A2B they want to make a decentralized taxi service and distribute wealth to community members. They are building trust-based social platform, which connects licensed taxi service providers and taxi service users into one community platform. WeTrust – Savings & Insurance platform WeTrust is a collaborative savings, lending and insurance platform that is autonomous, agnostic, frictionless, and decentralized. WeTrust creates a full-stack alternative financial system that leverages existing social capital and trust networks, eliminating the need for a “trusted third party”, allowing for lower fees, improved incentive structures, decentralized risks, allowing a greater amount of capital to reside among the participants, and ultimately improving financial inclusion on a global scale. Going back to barter system… To conclude transaction in its earliest form began with the barter system and now with crypto assets, we are going back to the same structure. Crypto assets are treated as a commodity today, just like gold. We have to wait and watch to see how it will develop. The pace of change is accelerated now more than ever. As an investor, you can only ensure you remain updated and part of the whole industry so that you are present when the next big evolution happens.","excerpt":"In today’s world, crypto assets are changing the way businesses need to be built. Instead of getting the equity financing from founders and accredited investors, entrepreneurs are using ICO (Initial Coin Offering) financing route. In the crypto token version of business, marketplace value is captured by all those who create value for the marketplace, including […]","categories":["IT Services"],"tags":[],"author_name":"Nitin Srivastava","publish_date":"2018-05-19T07:15:59","publication_year":"2018","word_count":1184,"keywords":["Go","cloud computing","AI","ML","Scala","Git","RAG","Aim","Rust","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","cloud computing","R","Go","Rust","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/demystifying-crypto-assets-a-digital-barter-system\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10110347,"title":"JPMorgan Scientist Unveils Phixtral, Mixture of Mistral with Phi-2","content":"Mistral just released the paper of their Mixtral of Experts model and there are new models already coming in. Maxime Labonne, Sr. Machine Learning Scientist at JPMorgan, has introduced Phixtral, a novel Mixture of Experts (MoE) model built with Microsoft Phi-2 models. Click here to check out the model. Labonne’s creation combines 2 to 4 fine-tuned models, each containing 2.8 billion parameters, surpassing the performance of individual experts, drawing inspiration from Mistral AI’s Mixtral architecture while developing Phixtral. Phixtral can run with 4-bit precision on a free T4 GPU. Phixtral is presented in two variations: phixtral-2x2_8 and phixtral-4x2_8. The former represents the first MoE made with two microsoft\/phi-2 models, inspired by the mistralai\/Mixtral-8x7B-v0.1 architecture, and outperforms individual experts. Meanwhile, the latter, phixtral-4x2_8, stands out as the inaugural MoE incorporating four microsoft\/phi-2 models, once again surpassing the capabilities of individual experts. The model’s efficiency is underscored by its ability to outperform each individual expert, marking a notable advancement in MoE design. On ‘Yet Another LLM Leaderboard’ (YALL), the model performed better than base phi-2 and just below Zephyr2-7B. Phixtral’s underlying architecture, represented by models like dolphin-2_6-phi-2, phi-2-dpo, phi-2-sft-dpo-gpt4_en-ep1, and phi-2-coder, showcases the collaborative effort of various model authors. Labonne emphasises the significance of these models in the creation of Phixtral, highlighting their exceptional capabilities.","excerpt":"Phixtral comes in two versions and can run with 4-bit precision on T4 GPU.","categories":["AI News"],"tags":["mistral ai"],"author_name":"Mohit Pandey","publish_date":"2024-01-10T15:37:32","publication_year":"2024","word_count":213,"keywords":["machine learning","mistral ai","programming_languages:R","AI","RPA","GPT","GAN","R","llm_models:GPT"],"extracted_tech_keywords":["AI","machine learning","R","GPT","GAN","RPA","llm_models:GPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jpmorgan-scientist-unveils-phixtral-mixture-of-mistral-with-phi-2\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10078480,"title":"ISRO’s Upcoming Space Missions","content":"The race to space has just begun. Every other day, rockets, satellites, and rovers are being launched by governments and private organisations, including Roscosmos or NASA across the globe to push ahead and make their mark in space. Indian Space Research Organisation (ISRO) has likewise been making several breakthroughs in developing space technology—with missions planned for the decade and beyond. As of October 26, 2022, India has launched about 381 satellites for different countries, including USA, Canada, UK, Germany, Sweden, and others. Here’s a list of missions announced by ISRO, some of which are set to launch next year, and some before 2030. Gaganyaan-3 With plans to launch by mid-2023, Gaganyaan-3 is ISRO’s first manned mission to space. It is being manufactured in collaboration with DRDO and HAL. The mission starts with sending two unmanned test-flights, Gaganyaan 1 & Gaganyaan 2, before launching its three-manned satellite. The development for the satellite started back in December 2014. With a total expected budget of nearly INR 12,400 crore, the objective of the mission is to demonstrate the indigenous capabilities of undertaking human space flight missions on Low Earth Orbit (LEO). What is GaganyaanGaganyaan (Orbital Vehicle) is an Indian crewed orbital spacecraft intended to be the basis of the Indian Human Spaceflight Programme.PM Modi : \"We have decided that by 2022, when India completes 75 years of independence.#ISRO#gaganyaan#astronomy pic.twitter.com\/hpIWHggLcy— Gaganyaan (@Gaganyaan_Isro) September 21, 2019 Aditya – L1 This is India’s first mission for solar observation. The satellite will be placed around the Lagrangian point 1—which is approximately 1.5 million km from Earth—to dodge all kinds of eclipses and study the solar corona with the help of a solar chronograph. It is expected to launch in March 2023. If the mission becomes successful, ISRO would have the ability to predict storms coming from the Sun and study the Solar Weather System. The budget for the mission is estimated to be INR 378 crore, excluding launching costs. Gearing up for Aditya-L1 Space MissionAt the ongoing #ASI2021, Indian solar physicists gathered to take stock of their preparations for the Aditya-L1 space mission in a dedicated workshop. Aditya-L1 is the first mission of @isro dedicated to observing the Sun’s activity.(1\/8)— Astronomical Soc. India Outreach and Education (@asipoec) February 20, 2021 Chandrayaan-3 Expected to launch around June 2023, Chandrayaan-3 is the successor of Chandrayaan-2, which had failed due to a last-minute glitch in soft landing guidance. As the name suggests, it is a lunar exploration mission to be launched at Satish Dhawan Space Centre with Launch Vehicle Mark-3 (LVM) rocket and will consist of a rover, an orbiter, and a lander. With a budget of approximately INR 615 crore, Chandrayaan-3 will make India the fourth country to soft-land on the moon successfully. Chandrayaan-3 is almost ready. Final integration and testing almost complete. Still, some more tests are pending, so we want to do it a little later & there were two slots available one in Feb & another in June. We would like to take June (2023) slot for the launch: ISRO Chairman pic.twitter.com\/mfdxh4nvi5— ANI (@ANI) October 22, 2022 X-ray Polarimeter Satellite A space observatory to study polarisation of cosmic X-rays, ‘XPoSat’ is planned to be launched in the second quarter of 2023 on a Small Satellite Launch Vehicle (SSLV). The mission will last for five years to study approximately 50 brightest sources in the universe and gauge the radiation from each of them. With a grant of INR 95 crore, the project began in September 2017. The satellite is set to study the pulsars, active galactic nuclei, black hole X-ray binaries, and non-thermal supernova remnants. Another X-ray polarimetry mission viz. the X-ray Polarimeter Satellite (XPoSat) (https:\/\/t.co\/QUZnKOvGAE) is slated for launch sometime next year onboard ISRO's Small Satellite Launch Vehicle (SSLV). Indeed exciting times ahead for astronomers\/astrophysicists and upcoming…— X-ray pulsar (@RayPulsar) December 9, 2021 NISAR A Synthetic Aperture Radar (SAR) satellite, ‘NISAR’, is a joint project between NASA and ISRO. The mission is to launch a dual-frequency SAR on an Earth observation satellite—the first of its kind. The estimated cost for this project is nearly $1.5 billion and the satellite is expected to launch by January 2024. This satellite will be used for remote sensing to understand the natural processes of the Earth like the Arctic and Antarctic cryosphere. It will be launched from India on a GSLV Mark II and the planned mission life is three years, orbiting synchronously with the Sun. NISAR (@NASA–@isro Synthetic Aperture Radar) is an upcoming Earth satellite mission currently being built and tested here at JPL. Join us live at 10:15am PT (1:15pm ET) on Aug. 3 to learn about what NISAR will do from experts. Drop your ❓ in the comments. https:\/\/t.co\/KYFDWnwNS1 pic.twitter.com\/XlaNBwePb0— NASA JPL (@NASAJPL) August 2, 2022 Mangalyaan 2 Otherwise known as the Mars Orbiter Mission 2 (MOM 2), Mangalyaan 2 is ISRO’s second interplanetary mission to launch to Mars between 2021 and 2022. The mission will include a panchromatic camera, a radar, and a hyperspectral camera to understand early stages of Mars. With a duration of around one year, the mission is expected to be launched in 2025. Following MOM-2, ISRO has also proposed MOM-3 in 2030, with the objective of soft landing of a rover near the Eridania Basin, a theorised ancient lake on Mars. #Mangalyaan quietly bids goodbye. India's best innovation and world's cheapest yet best performing Mars Orbiter has run out of fuel. It is the result of our phenomenal scientific minds that it served for 8 years, way beyond its design life of 6 months. pic.twitter.com\/1mhVQC1eW0— Varun Puri (@varunpuri1984) October 3, 2022 Shukrayaan-1 Another interplanetary mission by ISRO, Shukrayaan-1 is a planned orbiter to Venus. The objective is to study the atmosphere and surface of Venus. This will also include studying the solar irradiance and solar wind interaction with the ionosphere. The mission is planned to be launched in December 2024 on a GSLV Mark II. The mission was proposed in 2012 and the Government of India allocated and increased funds for the Department of Space in 2017–18. ISRO has shortlisted proposals to include collaborations with France, Sweden, Germany, and Russia for the mission. #Shukrayaan mission update ·ISRO is targeting to launch India's First Venus mission in Dec 2024.·Work is undergoing for project.·It will be launched by #GSLV (MK2 maybe)·It will be put in 500 to 60,000 km orbit. pic.twitter.com\/b11TfONxat— Vishesh Verma (@Vishesh03625993) May 5, 2022 AstroSat-2 As a successor to AstroSat-1—whose operation ended in 2020—AstroSat-2 is India’s second multi-wavelength space telescope and is expected to propel the study of astrophysics and astronomy. ISRO made the announcement of the opportunity in February 2018, seeking proposals from all institutions to further research and development in the field. The key functions of these satellites include studying neutron stars, black holes, binary star systems, and star berth regions. The AstroSat-1 is managed by the spacecraft control centre of ISRO’s Mission Operations Complex (MOX). https:\/\/twitter.com\/isro\/status\/1572826829391802368 Lunar Polar Exploration Mission LUPEX, also known as Chandrayaan-4, is a lunar robotic mission in collaboration with Japan Aerospace Exploration Agency (JAXA). The objective is to send a lander and rover to explore the south pole of the moon by 2025. JAXA is providing the H3 launch vehicle and the rover, and ISRO will be building the lander. In 2019, NASA also discussed the possibility of joining the mission. JAXA proposed in its website that its aim is to obtain ground truth data for quantity of water and also the quality of lunar water. ISRO-JAXA joint Lunar Polar Exploration Mission seems to have got a new mission website. This is not to be confused with Chandrayaan-3 mission which will be done by ISRO alone.Website : https:\/\/t.co\/TxZvUFkiRkFor more on this project : https:\/\/t.co\/7Qc54uKu50 pic.twitter.com\/gETjNCvsh5— Strategic Frontier (@strategicfront) May 21, 2020","excerpt":"As of October 26, 2022, India has launched about 381 satellites for different countries.","categories":["AI Features"],"tags":["Gaganyaan","Isro Satellites"],"author_name":"Mohit Pandey","publish_date":"2022-11-02T16:00:00","publication_year":"2022","word_count":1287,"keywords":["Go","AI","ETL","Gaganyaan","BERT","Ray","Aim","Isro Satellites","JAX","CLIP","Chroma","R"],"extracted_tech_keywords":["AI","Aim","Ray","JAX","Chroma","R","Go","ETL","BERT","CLIP"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/isros-upcoming-space-missions-2023\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":69607,"title":"Why AWS Has Been Never Been Surpassed","content":"Amazon Web Services (AWS) is a subsidiary of the parent Amazon, which gives on-demand cloud computing platforms and APIs to developers, companies, and world governments, on a subscription pay-as-you-go model. Its cloud computing web services present a set of fundamental tools for abstract technical infrastructure and distributed computing around the world. The AWS technology is implemented at server farms worldwide and managed by the Amazon subsidiary. The fees are determined by the hardware, operating system, software, or networking features taken by subscribers, required availability, redundancy, security, and service alternatives. Subscribers can pay for a private virtual AWS computer, a dedicated physical computer, or clusters of either. The AWS platform was first introduced in July 2002. In its initial stages, the platform included only a few different tools and services. The company found that they were actually very good at operating infrastructure services, such as computing, storage, and databases. So they began to focus on building the AWS platform. As the first company to launch cloud services and leader of the cloud service market for more than a decade, AWS remains the cloud giant with more than one-third of the cloud computing market share. AWS has become accustomed to being imitated by various types of competitors. But why is AWS so strong? Continuously Introduce New Functions And Services To Meet User Needs AWS owns a dominant share in the cloud market while the next three competitors Google, Microsoft, and IBM, have a much smaller market share. As of 2020, AWS comprises more than 212 services, including storage, networking, database, computing, developer tools, analytics, application services, deployment, management, mobile, etc. The most well known including Amazon Elastic Compute Cloud (EC2) and Amazon Simple Storage Service (Amazon S3). Over the years, the number of new features and services launched by AWS every year has been rising rapidly. The unprecedented speed of innovation stems from the fact that AWS’s innovation is customer-centric, and continues to develop new products based on customer-first concepts and customer feedback. In addition, AWS’ infrastructure will continue to introduce new technologies, and through advanced architecture design, it will maintain its leading position in terms of stability, reliability, security, and efficiency of resource utilization. AWS has a philosophy: every service and function has been used by its own team, and it is only launched after it is used. It can be said that every service and function launched by AWS has been thoroughly refined. As a technology learner, you should pay more attention to AWS technology, architecture, and so on. The investment in infrastructure has a strong scale effect. At the same time, the scale effect can also bring a series of benefits: AWS’s number of customers, industry type, application type, scale distribution, geographical and cultural distribution, all other cloud service providers are beyond the reach. When you encounter the same technical problem and need help from others, you will find that AWS has a lot of technical circles, and it is easy to find the answer, and your problem is basically not new. Customer-Centric Service After intensively looking at AWS, you can find beliefs that the development of AWS today occupies major advantages. The AWS team has accumulated a lot of lessons in building and operating AWS cloud computing services. These services not only ensure security, availability, and scalability but also provide predictable performance at the lowest cost. Given that AWS is a pioneer in building and operating such services around the world, these lessons are critical to the business. Considering that AWS has more than one million active users per month, and these users may serve hundreds of millions of their customers. Therefore, the opportunities to accumulate the above lessons are everywhere in AWS, and these lessons have become the basis, foundation, and motivation for AWS to continue to progress.","excerpt":"Amazon Web Services (AWS) is a subsidiary of the parent Amazon, which gives on-demand cloud computing platforms and APIs to developers, companies, and world governments, on a subscription pay-as-you-go model. Its cloud computing web services present a set of fundamental tools for abstract technical infrastructure and distributed computing around the world.  The AWS technology is […]","categories":["AI Features"],"tags":["Amazon AWS","AWS cloud","RPA developer"],"author_name":"Vishal Chawla","publish_date":"2020-07-13T13:00:00","publication_year":"2020","word_count":631,"keywords":["Go","API","RPA developer","AWS","AI","cloud computing","distributed computing","Scala","RAG","Amazon AWS","analytics","AWS cloud","R"],"extracted_tech_keywords":["AI","analytics","RAG","cloud computing","AWS","distributed computing","R","Go","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-aws-has-never-been-surpassed\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":32525,"title":"Tech Mahindra Launches An Open Source AI Platform GAiA Powered By Acumos","content":"Tech Mahindra recently launched GAiA, an open-source AI platform that will enable enterprises across the industry to build, share and deploy AI-driven services and applications to solve business-critical problems. Also, GAiA is the first enterprise edition of open source AI platform Acumos. Now build, share & rapidly deploy #AI-driven services and applications to solve business critical problems! #CreateTheNxt with GAiA – https:\/\/t.co\/y0DznlxnrS pic.twitter.com\/KOiaRWmNJk — Tech Mahindra (@tech_mahindra) December 27, 2018 Furthermore, the platform will not only support enterprises across the industry to adopt open-source AI platform offerings but also help in implementing custom use cases, models and integration services. GAiA will be available for the commercial purpose and will support open source distribution. The launch of GAiA is a part of Tech Mahindra’s TechMNxt charter that focuses on beefing up capabilities for next-gen technologies like AI. Acumos, AI-powered platform is now available in Beta, hosting a marketplace of Machine Learning (ML) models that can be applied to popular use cases in many industry verticals. “The need of the hour is to democratise knowledge in AI and make it accessible to everyone, accelerating adoption and transformation. With GAiA, we’re one step closer,” said Manish Vyas, President, Communications, Media & Entertainment Business, and the CEO, Network Services.  “We believe that it will be the change agent to drive digital transformation journey of our customers,” he further added. Talking about the further enrichment of GAiA with various industry solutions and the existing marketplace, it will be done through collaboration with academia, third-party Machine Learning developers, and companies. “Our goal in launching the Acumos AI project was always to see it become an open platform to accelerate innovation in the AI ecosystem, and we are pleased to see that vision taking shape. GAiA, powered by Acumos, looks set to help broaden AI’s reach even further and make it accessible to everyone,”  said Andre Fuetsch, President, AT&T Labs, and CTO at AT&T.","excerpt":"Tech Mahindra recently launched GAiA, an open-source AI platform that will enable enterprises across the industry to build, share and deploy AI-driven services and applications to solve business-critical problems. Also, GAiA is the first enterprise edition of open source AI platform Acumos. Now build, share & rapidly deploy #AI-driven services and applications to solve business […]","categories":["AI News"],"tags":["Open Source AI"],"author_name":"Harshajit Sarmah","publish_date":"2018-12-31T10:07:12","publication_year":"2018","word_count":317,"keywords":["Go","API","machine learning","programming_languages:R","AI","innovation","ML","digital transformation","Git","Open Source AI","R"],"extracted_tech_keywords":["AI","machine learning","ML","R","Go","Git","API","digital transformation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tech-mahindra-launches-an-open-source-ai-platform-gaia-powered-by-acumos\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10068642,"title":"How to learn from uncertainty using probabilistic machine learning?","content":"The use of traditional or deterministic machine learning models has the ability to come up with predictions only for a predefined or specified event. That is where probabilistic machine learning plays a vital role by not only predicting outcomes for certain events rather coming up with predictions for uncertainties considering various parameters for prediction. This article briefs on the different ways a machine learning model learns from the data with uncertainty involving the concept of probability theory.  We will also focus on why the probabilistic machine learning model outperforms the traditional methods. Table of contents What is probabilistic machine learning?Different ways a machine learning model learns?Pros\/cons of the traditional learning methodThe necessity for probabilistic machine learningWhy is probabilistic learning the best?Summary Let’s start the discussion with what actually is a probabilistic learning approach. What is probabilistic machine learning? Probabilistic machine learning is one of the techniques a machine learning model learns from the fitted data and yields predictions not only for particular classes or instances but also ensures to address the issues with uncertainties in general and helps us in yielding predictions with respect to all the classes considered. The linear growth of data may be one of the reasons why probabilistic machine learning models yield the right outcomes considering the possible outcomes of each of the features and the different classes in the instance of the data. Are you looking for a complete repository of Python libraries used in data science, check out here. Different ways a machine learning model learns There are various ways a machine learning model learns its data. Some of them include Deterministic machine learning, Stochastic machine learning, and Probabilistic machine learning. Deterministic Machine learning As the name suggests deterministic machine learning is one of the ways where a machine learning model learns its data in the form of possibilities of natural outcomes possible. This way of learning may help in yielding reliable results for the various input parameters of the data utilized, and the learning happens in an iterative process where more the learning from the data acquired by the machine learning model developed better are the outcomes. Pros and cons One major advantage of deterministic learning is the learning process may converge quicker yielding a simpler model but the model when considered for uncertain data or changing events may be responsible for wrong outcomes as in this process learning process happens without considering the randomness of features. Stochastic Machine Learning As in the name “Stochastic”, this is a way where the machine learning model learns by considering the possibilities of randomness and possible future uncertainties. So as the model learns about randomness and possible uncertainties, the model will be responsible for yielding predictions considering all the likeliness factors of the input data used. But even if the model learns with respect to randomness and uncertainties there are some drawbacks to using this type of learning. The pros and cons of this type of learning are mentioned below. Pros and cons As mentioned the stochastic learning process happens by considering all the uncertainties in the data. But the fact considering the possible uncertainties, the consideration happens with respect to certain bias factors, and only the most extreme uncertainties are considered leaving behind factors with considerable uncertainty, and this would be a concern because data the model learns may possibly change over time and the let of uncertainties by stochastic machine learning may stand off to be the most extreme uncertain information or data. The necessity for probabilistic machine learning So as we have an overview of the individual learning factors of both the deterministic and stochastic learning models and their respective concerns of pros and cons, probabilistic machine learning is one such effective learning technique for the machine learning model developed as it learns all the uncertainties of the data without any bias and considers the effect of possible randomness in future and help in yielding the right outcomes in the testing phase of the machine learning model developed. One of the most commonly used probabilistic classifier models is the Naive Bayes Classifier model which facilitates addressing the possible uncertainties through required conditions of randomness as it basically obeys the property of Conditional Probability where for any certain independent assumptions made for uncertainties the probability distribution considers the possible likelihood of all the parameters to be estimated. Estimating certain parameters and possible outcomes becomes easier by using the Naive Bayes classifier algorithm but as in the name “Naive”, it has certain limitations with respect to the assumptions of the target variable. But during the learning process, specific conditions for possible uncertainties can be mentioned to evacuate the possible outcomes of issues associated with a bias for most uncertain events. Due to the various advantages of probabilistic machine learning, there are various frameworks supporting the same. Some of them include: STAN – A Bayesian statistical frameworkTensorflow probability – A compact framework of TensorflowPyro – A universal probabilistic learning frameworkPyMC3 – Open Source framework for probabilistic learning Why is probabilistic machine learning the best? To answer this question let’s keep in mind the common issues a typical machine learning model in production would face. Say for supposing data provided is insufficient. This is where probabilistic learning addresses the issue of data scarcity by addressing the randomness and the possible uncertainties of the data employed and probabilistic learning So once required data is available the next possible concern would be a linear expansion of the model. So as probabilistic learning facilitates linear expansion, possible concerns with model scalability are also addressed as the model has already learned for possible uncertainties. The issues of bias uncertainty are addressed and the machine learning model exhibits a high degree of representation of convergence for uncertainties for random events. The probabilistic model easily converges with small changes in data as it is pretrained for possible uncertainties. Attention to the most relevant information is ensured in probabilistic learning as it learns through various randomness and uncertainty by iterating through various parameters. Transparency and reliability of the model in production are ensured as the model has learned to yield the right predictions for any possible uncertainties. So these are some points which make probabilistic learning one of the effective learning techniques a machine learning model learns and help us yield reliable models for production.s Summary So among the various methods of making a machine learning model learn the data, currently probabilistic machine learning appears to be the most appealing technique as it adheres to learning for all possible uncertainties without any bias factors for all the possible uncertainties. Making a machine learning model learn the probabilistic way may help in yielding reliable models for better predictions for business-driven solutions and evacuate the concerns associated with serious consequences of faulty predictions.","excerpt":"This article briefs on the various ways a machine learning model learns from the data with uncertainty and why is probabilistic machine learning the best.","categories":["AI Trends"],"tags":["Data Science","Machine Learning"],"author_name":"Darshan M","publish_date":"2022-06-09T12:00:00","publication_year":"2022","word_count":1127,"keywords":["data science","Go","machine learning","AI","Machine Learning","Scala","Python","ai_frameworks:TensorFlow","ViT","Data Science","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","data science","TensorFlow","Python","R","Go","Scala","ViT","ai_frameworks:TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-learn-from-uncertainty-using-probabilistic-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10134339,"title":"MiniMax- 10 Realistic Videos Created Using the Chinese Text-to-Video Startup","content":"MiniMax was launched as a text-to-video generator by a Chinese startup bearing the same name. The company recently launched its first model, Video-01. This model is designed to create high-resolution videos from text prompts. It operates at a native resolution of 1280 x 720 pixels and can generate videos at 25 frames per second. Currently, the maximum video length is limited to six seconds, with plans to extend this to ten seconds in the future. Another Chinese 'Sora': A new AI video tool launched today by Minimax, backed by major investors Alibaba Group and Tencent. 🎞️Check out their official AI film Magic Coin🪙, created entirely with text-to-video . 🥁Try it for free now: https:\/\/t.co\/Kl1avPXkFL pic.twitter.com\/df14ZVq1Es— Junie Lau (@JunieLauX) August 31, 2024 During an interview, founder Yan Junjie mentioned that the company had made significant progress in video generation. However, the specific parameters and technical details of the model have not been disclosed yet. “We have indeed made significant progress in video model generation, and based on internal evaluations and scores, our performance is better than Runway,” Junjie said. The current model is the initial version, with an updated version expected soon. As of now, it offers text-to-video capabilities, with future plans to include image-to-video and text-plus-image generation features. Founded in 2021 by former employees of SenseTime, including Junjie. The startup is backed by Alibaba and Tencent. 10 Best Videos Created by Minimax Here, we delve into some of the best videos generated by MiniMax. 1. Magic coin The company released an official 2-minute AI film titled ‘Magic Coin’ generated entirely by its large model, showcasing a coin that appears and disappears within a person’s hand, illustrating the tool’s capability to seamlessly integrate AI-generated elements into real-world footage. This technology highlights the rapid advancements in AI video generation and its potential applications across various industries, including entertainment, advertising, and visual effects. 2. Cats eat fish, dogs eat meat The MiniMax video contrasts the natural instincts of cats and dogs, with cats favouring fish and dogs preferring meat, while the camera slowly pans towards a robotic figure appearing at the window. The figure then watches the animals for a brief few seconds. The video uses clever animations or real-life footage to highlight how each animal reacts to its favourite food. 3. Man eating fast food Here, it portrays a man indulging in a burger, capturing the rush and satisfaction of a quick meal against the backdrop of a food court. The video uses visuals and humorous elements to emphasise the speed at which the man consumes his food, reflecting a common modern-day scenario. This highlights the advanced visual capabilities of AI in capturing and rendering detailed environments. 4. Teenager skateboarding The video showcases a teenager skateboarding through the city, highlighting the thrill and skill involved in navigating the streets, with dynamic camera angles and fast-paced editing. It also features iconic city landmarks, adding to the sense of adventure and exploration. 5. Classical beauty applying lipstick As the blurred camera comes back into focus, it features a beautiful Asian woman applying lipstick in her room. The video by Minimxax emphasises the delicate, precise motions as she applies the lipstick, with soft lighting and close-up shots showcasing her beauty. The room’s décor enhances the aesthetic appeal, reflecting a sense of elegance that might evoke royalty. 6. Pixel style Amid the busy street in a bustling city, all rendered in pixel art style, a small cat walks by, weaving through the pixelated crowd and traffic. It cleverly contrasts the lively urban environment with the calm, making the cityscape vibrant and the cat’s presence even more striking. The precise features and seamless editing add to future creativity in AI. 7. Futuristic high-tech lab Set in a futuristic high-tech lab, this MiniMax video shows a woman engaging in a conversation with a holographic figure. The sleek, advanced technology is highlighted with the hologram, possibly representing an AI or digital assistant, interacting fluidly with the woman, suggesting a seamless integration of human and machine communication. 8. Blade Runner Cyber City It begins with a sweeping view of a cyber city, characterised by towering skyscrapers, neon lights, and a vibrant, futuristic atmosphere. As the camera slowly shifts, it focuses on a man eating ramen at a roadside food truck, creating an intriguing contrast between the high-tech surroundings and the simplicity of street food. This scene emphasises the coexistence of advanced technology and everyday human experiences. 9. Silver 1977 Porsche 911 Featuring a sleek, silver 1977 Porsche 911 Turbo cruising through a vibrant cyberpunk landscape, the car’s classic design contrasts strikingly with the futuristic, neon-drenched cityscape surrounding it. The video showcases the vehicle’s smooth motion against the backdrop of glowing backdrops, towering skyscrapers, and bustling streets. The contrast of the vintage car with the high-tech environment highlights a blend of old-world charm and futuristic aesthetics. 10. Wig and sunglasses A sad, bald man appears dejected and downcast. As the scene progresses, he puts on a wig and sunglasses, which instantly transform his mood. The video by Minimax captures his transition from sadness to joy, highlighting the cheerful change in his expression and posture. AI ensures that each gesture and reaction is natural, enhancing the viewer’s engagement and the characters’ believability.","excerpt":"The brand-new text-to-video model stands out due to its realistic AI visual content.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","AI Video Generation Models"],"author_name":"Tarunya S","publish_date":"2024-09-03T18:15:55","publication_year":"2024","word_count":870,"keywords":["API","programming_languages:R","AI","ML","Git","AI Video Generation Models","Ray","ViT","GAN","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","ML","Ray","R","Git","API","GAN","ViT","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-10-best-videos-created-by-minimax\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60221,"title":"MyGate Lays Down Guidelines For Gated Communities","content":"With guidelines such as self-distancing amongst the most crucial to breaking the chain of COVID-19, MyGate– the security and community management app for gated communities have laid down a number of guidelines for gated communities to be followed during this period of the outbreak. Through its official Twitter account, the company shared the Coronavirus gated community response guidelines which one can download by clicking the link. The instructions and tips shared by MyGate have been divided into six different categories to serve different people associated with a gated community. We have put together a coronavirus response plan for gated communities. You can also download it here: https:\/\/t.co\/81ZRxOuxqQPlease share it along in your community if you find it useful.#Coronaindia #COVIDー19 #CoronaVirusUpdates #CoronavirusOutbreak— Mygate (@MyGate_com) March 15, 2020 For the management committee of a community, MyGate suggests not to organise any events in the community in this period and closure of gyms, swimming pools and other classes that are often used by many. The company urges to stop the usage of biometrics for access till the situation is under control which will help in reducing the spread of the virus. The company also requested gated communities to ask society staff (e.g. electrician, plumbers) to not roam around in the community. They should be restricted to the estate Managers office once their designated job is complete. Last but not least, it urges to create an Emergency Response Team (ERT) within the community consisting of residents, committee members and key society staff members. If one community has a medical practitioner in the ERT, that will be of big help. For large communities with multiple towers, one can create an ERT for each tower or block. Moving on to the guidelines for estate managers, MyGate suggests to take care of personal hygiene, keep backup for Genset, ensure water continuity and regular sanitisation of the entire community, especially the lifts. Not to mention, garbage collectors are to be taken care of since they are likely to be more exposed to the virus. Last but not least, a team should be trained to handle the Sewage Treatment Plant (STP) operation. A few tips for the residents as well have been pointed by MyGate. The reports read that one should disinfect your main door knobs\/handles and calling bell switches frequently. If someone has to step out of the house for supplies, once back, the first thing to do is thoroughly clean hands and change the clothes immediately. Residents can increase their immunity by consuming Zinc, Vitamin-C and other multivitamins along with adequate sleep. It is also the moral responsibility of the residents to educate their domestic help on the sensitivity around the issue and the precautions that need to be taken at their homes. Last but not least, MyGate mentions residents must follow every measure laid down by health agencies self disclose or report if there is any suspect case of infection.","excerpt":"With guidelines such as self-distancing amongst the most crucial to breaking the chain of COVID-19, MyGate– the security and community management app for gated communities have laid down a number of guidelines for gated communities to be followed during this period of the outbreak.  Through its official Twitter account, the company shared the Coronavirus gated […]","categories":["Deep Tech"],"tags":[],"author_name":"Rohit Chatterjee","publish_date":"2020-03-27T16:39:11","publication_year":"2020","word_count":484,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/mygate-lays-down-guidelines-for-gated-communities\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10040155,"title":"Appian Launches New Low-Code Automation Platform For Enterprises","content":"Appian is one of the early entrants in the low-code application development space. Recently, the US company launched a new version of its low-code automation platform for businesses looking to develop quick applications. With over two decades of experience in the enterprise technology landscape, Appian has helped businesses in digital process automation (DPA), intelligent business process management systems (iBMS) and dynamic case management (DCM), among others, Interestingly, the launch of its new low-code automation comes when enterprises are looking for quick solutions to deploy AI-powered applications and smooth workflow automation across departments with limited resources and agile processes. Today, low-code, no-code technology platforms have emerged as a go-to model for businesses. Several players, including Appian, Microsoft, Amazon, Pega, and ServiceNow are working on products and ideas to ease the burden for enterprises. In India, companies like Infosys, HCL Technologies and Tech Mahindra, alongside various startups, are also working on this technology. “This is the time for low-code automation platforms,” said Matt Calkins, Appian founder and CEO. “We have just started a new decade, but low-code is how applications are built in the future. It’s inevitable,” said Calkins. A cloud-based, no-code application development platform Quixy’s CEO Gautam Nimmagadda told AIM that no-code would allow more companies to participate in software development, allowing professional developers to focus on advanced and specialised areas. According to Gartner, low code platforms will account for over 65% of application development by 2024. In the next three years, 75% of large businesses will employ at least four low-code tools for IT application development. What’s new? Appian’s USP is its ability to manage business outcomes and process orchestration of hybrid operations. Its model-driven app development platform allows anyone with limited or no coding experience to build process-centric and case-centric applications, effortlessly. With its latest launch, Appian claimed to offer enhanced AI-driven intelligent document processing (IDP), new design guidance and collaboration features, along with advanced DevSecOps capabilities. More than anything, the platform provides a single workflow for automation, team orchestration, existing systems, data, bots and more. Appian claimed the new platform could deliver a project (for example, a company’s first app) in just eight weeks for $200k. “We were able to build the first Appian RPA process in just four days, integrating with our people processes and Appian AI,” said Matt Richard, CIO at Laborers International Union of North America (LiUNA). The new features in Appian’s low-code automation platform include: Low-code data: Helps integrate data as easy as building apps. With this, a developer can source data from anywhere without migrating it. The platform visually combines and extends the model relationships between varied data sources automatically and optimises data sets for performance without database programming.Intelligent document processing (IDP): Delivers massive efficiency by processing large volumes of unstructured data. In addition to this, it now features native optical character recognition (OCR) to extract data from documents without third-party software or services securely. Also, IDP can eliminate manual processing in any workflow with the help of native, pre-trained and constantly learning AI models.Low-code RPA (robotic process automation): Manages and monitors Appian bots, third-party bots and end-to-end processes to increase performance and scale. With the help of the new low-code RPA, the enterprise can now automate tasks much faster and access new libraries and updates from the Appian AppMarket. Low-code apps: Boosts developer productivity and lets organisations build and automate 10x faster through visual development (which is as simple as drawing a flowchart).Low-code DevSecOps: Enhances simplified movement of software packages between development, test and production workflow with a one-click comparison feature to scale the deployment securely. Wrapping up Recently, Appian was named in a 2021 Gartner Peer Insights Customers Choice report for enterprise low-code application platforms. The University of Texas at Dallas included the Appian low-code automation platform into the core curriculum of its new intelligent automation course. Also, HP Hood deployed its workforce safety solution with vaccination insights. Last year, the company acquired robotic process automation (RPA) firm Novayre Solutions for an undisclosed amount. In the first quarter of 2021, Appian’s cloud-subscription revenue increased by 38% year-on-year to $39.1 million. Its subscriptions revenue grew 26% year-on-year to $63.8 million.","excerpt":"Appian is one of the early entrants in the low-code application development space. Recently, the US company launched a new version of its low-code automation platform for businesses looking to develop quick applications.  With over two decades of experience in the enterprise technology landscape, Appian has helped businesses in digital process automation (DPA), intelligent business […]","categories":["IT Services"],"tags":["no code platforms"],"author_name":"Amit Naik","publish_date":"2021-05-16T18:00:00","publication_year":"2021","word_count":688,"keywords":["Go","intelligent automation","AI","RPA","Git","automation","Aim","ViT","GAN","no code platforms","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GAN","ViT","automation","RPA","intelligent automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/appian-launches-new-low-code-automation-platform-for-enterprises\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10171602,"title":"HCLTech Expands Partnership with The Standard","content":"HCLTech has expanded its partnership with The Standard, a leading provider of financial protection services, to enhance its digital transformation through AI-driven infrastructure and application services. This collaboration will improve efficiency, support rapid growth, and boost customer experience, aligning with The Standard’s long-term goals of shifting to an IT products and services-based operating model. The partnership will focus on delivering HCLTech’s GenAI-led platform, AI Force, along with digital engineering and cloud services, to help The Standard enhance customer service in its workplace benefits offerings. A newly formed Joint Innovation Council and Digital Experience Office will further support this transformation. “The Standard’s growth journey has accelerated in recent years through digital transformation and acquisitions, and HCLTech has proven to be the best partner to help us scale efficiently and seamlessly with its digital-first and customer-focused approach,” Laxman Prakash, chief information security officer at The Standard, said. Meanwhile, Anubhav Mehrotra, senior vice president at HCLTech, stated, “We are excited about this extended partnership with The Standard, showcasing our deep commitment to the insurance sector. “ Previously, UiPath, a global automation company, partnered with HCLTech to help companies worldwide automate their operations using AI. The partnership aims to make businesses more efficient by reducing the need for human involvement in everyday processes. HCLTech will use the UiPath Platform to help companies automate areas such as finance, supply chain, procurement, customer service, marketing, and human resources. The tech giant will also provide ready-to-use AI tools to help businesses start and scale their automation efforts easily. As part of this collaboration, HCLTech and UiPath will set up an AI lab in India. The lab will focus on building industry-specific solutions and small-scale working models that cover everything from planning and implementation to ongoing improvements.","excerpt":"The partnership will focus on delivering HCLTech’s GenAI-led platform, AI Force, along with digital engineering and cloud services, to help The Standard enhance customer service in its workplace benefits offerings.","categories":["AI News"],"tags":["HCL Technology"],"author_name":"Shalini Mondal","publish_date":"2025-06-11T14:44:51","publication_year":"2025","word_count":289,"keywords":["Go","API","GenAI","AI","ML","digital transformation","Git","automation","Aim","HCL Technology","R"],"extracted_tech_keywords":["AI","ML","GenAI","Aim","R","Go","Git","API","digital transformation","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hcltech-expands-partnership-with-the-standard\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080386,"title":"It’s Time for Google to Resurrect Google+ from its Graveyard","content":"Twitter, the social media platform that shaped the cultural zeitgeist in a distinctive way, has turned into a chaotic dumpster fire. While the company wasn’t exactly in great financial health before Elon Musk took over, it has since then descended into absolute corporate havoc. All this is not to say that the platform was all things unholy. Source: TechCrunch Twitter helped employees build and sustain a personal brand, journalists do their research on Twitter, researchers track other academics on the platform and environmentalists map out forest fires globally on it. What would the loss of the microblogging platform mean to users? And more importantly, is there any other social media platform that can replace it? Tumblr is reportedly seeing a wave of users coming in from Twitter and decentralised social network Mastodon, which markets itself as an alternative to Twitter, is also having its moment in the sun with the platform’s monthly active users crossing the 1 million mark earlier this month. The waitlist for another decentralised social media platform, Bluesky Social, saw more than 30,000 sign ups within two days in the beginning of the month. Bluesky is technically an offshoot of Twitter and was commissioned by Twitter founder and former CEO Jack Dorsey. Origins of Google+ People are hashing out their own suggestions around what can fill the Twitter-shaped hole on Twitter (ironically) and some familiar names are coming up. Can Google revive its last attempt at a social media platform, Google+? The interface of Google+, Source: Failory.com Google officially pulled the plug on its highly-publicised Google+ service in the beginning of 2019 but its shutdown was announced two years prior to that. During its launch in 2011, Google’s new social networking site was expected to be an aggressive response to competitors like Facebook and Twitter. Google was figuring out how to deal with challenges presented by social media companies and was excited to keep up. Amit Singhal, former Google search engineer, discussed how the internet was starting to organise itself around people and the company had to turn its attention towards building a personalised hub of social activity for users. Singhal noted then that Facebook seemed like it was way ahead of its time and that it might even be building an alternative to the traditional internet with itself at the centre. While it does seem like a new kind of internet has emerged with the advent of social media companies, the race for social media platforms has proven to be much trickier than previously imagined. Facebook has pivoted to a new branding under Meta with their focus on the metaverse in the light of their social media company falling behind newer competitors like TikTok. With Twitter in a crisis now, the race is left wide open. Google+ was eventually phased out due to ‘low usage’ but the truth was that it became a security liability for Google. Around the same time as Facebook’s infamous Cambridge Analytica scandal, the company found two substantial data leaks that could have potentially exposed the personal data of tens of millions of Google+ users to developers outside. During the first incident, which Google kept secret for months, the company decided to discontinue Google+ permanently. After the second incident, the company sped up plans of shutdown by four months—leading the division to close by April instead of August. But did this mean Google had simply bowed out of the social media race? Google’s sneaky social features Google’s businesses were more complex than that—it tends to shut down parts which often merge elsewhere. For instance, Google later announced that it would revive a version Google+ for its G Suite platform. This means that the old Google+ apps for Android and iOS were rebranded as Google Currents and stayed in use for enterprise customers only. Currents was started back in April 2019 after Google+ was shut down but was essentially the same as the social media network. The plan made a lot of sense because it is typical for businesses to have internal communication platforms. It also didn’t seem like Google was willing to completely abandon their plans to have an ecosystem. The tech giant was anyway building systems designed to have users within its products and services instead of hopping onto a traditional social service. In certain spaces, this intention was even more evident. Google added a full-fledged social stream within its Maps app (Android and iOS systems) which more or less appears as a feed with photos, posts and reviews. Users can follow restaurants, articles from publishers and other useful information and tailor their feed according to their interests. Even when users share photos or videos from the Google Photos app, they can post these through a social or messaging service and can instead share them directly just as an ongoing or private conversation in the app. Users can also like or comment on posts that others share. Google Pay, the company’s payment transfer app was also launched as having been “designed around your relationships with people and businesses” according to their blog. An executive stated in an interview that, “All your engagements pivot around people, groups, and businesses.” So while the features haven’t been explicitly advertised as social media-like, they inherently are. Social media as a whole may still be outside the purview of Google but social features still aren’t. But if there’s a time to turn things around, this may be it.","excerpt":"Google’s businesses were more complex than that – it tends to shut down parts which often merge elsewhere.","categories":["Global Tech"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-11-22T16:00:00","publication_year":"2022","word_count":902,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/its-time-for-google-to-resurrect-google-from-its-graveyard\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10044003,"title":"Indian Autonomous Driving Startup Swaayatt Robots Raises $3 million","content":"India’s autonomous driving company Swaayatt Robots has raised $3 million from US based investors at a valuation of over $70 million. “The current funding will allow us to continue our fundamental research in reinforcement learning and motion planning, and will allow us to scale our existing planning and perception algorithmic frameworks. Furthermore, it will allow us to expand our R&D in developing novel motion planning and decision making algorithmic frameworks, to enable Level-4 and Level-5 autonomous driving in stochastic traffic and unstructured environments,” said Sanjeev Sharma, founder of Swaayatt Robots. According to Sanjeev, the algorithms developed by Swaayatt Robots are capable of negotiating highly stochastic traffic-dynamics, and can even work in unstructured Indian environments. “Majority of our research focuses on motion planning and decision making under uncertainty, and a small fraction of our research focuses on enabling autonomous driving without the requirement of high-fidelity maps.” Swaayatt Robots is one of the few Indian startups working on solving the autonomous challenges from a pure algorithmic point of view. The team makes use of reinforcement learning(RL) for learning various navigation policies to allow autonomous vehicles to deal with both the stochastic and adversarial traffic scenarios. For example, their multi-agent intent analysis and negotiation framework is a multi-RL agent framework which allows autonomous vehicles to negotiate tight dynamic roads, at both low and high speeds. They have also developed perception algorithmic frameworks that help autonomous vehicles perceive without LiDARs and RADARs, both during the day and at night. According to the company, their perception algorithms, based on end-to-end deep learning, are highly computationally efficient compared to the existing state-of-the-art deep learning systems. Also Read: Interview with Sanjeev Sharma, Founder, Swaayatt Robots Another key challenge Swaayatt is trying to solve is avoiding the use of high fidelity maps. For self-driving cars, high fidelity maps are a must. These maps contain information like lane markers, which can be leveraged for navigation. However, this mapping process is costly and time-consuming. Sanjeev and his team at Swaayatt Robots have developed some novel algorithms in both perception and planning to get rid of high fidelity maps. “Typically, companies rely on high fidelity maps for projecting and generating delimiters because they cannot develop algorithms robust enough to detect delimiters reliably in real-time. Around 2% of our research focuses on enabling autonomous driving without requiring high-fidelity maps.” explained Sanjeev. Founded in 2015, Swaayatt Robots has already tackled many hard problems in autonomous technology. The fresh capital will be used to further expand their R&D efforts to solve problems such as off-shift policy framework, learning overtaking on typical two lane roads, learning to abort overtaking (like the Indian drivers), and learning stochastic tight-space negotiation policies. “Going forward we will be scaling our existing algorithmic frameworks, as well as continue our research on development of algorithmic frameworks to achieve Level-4 and Level-5 autonomous in highly stochastic traffic-dynamics,” said Sanjeev.","excerpt":"Swaayatt Robots has raised $3 million from US based investors at a valuation of over $70 million.","categories":["AI News"],"tags":["AI Startups"],"author_name":"Ram Sagar","publish_date":"2021-07-20T14:36:53","publication_year":"2021","word_count":477,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","RAG","deep learning","R","AI Startups","startup"],"extracted_tech_keywords":["AI","deep learning","RAG","R","Go","API","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/india-autonomous-swaayatt-robots-funding-3-million\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":52028,"title":"Become An AI Proficient Business Leader With This Unique Programme","content":"Artificial intelligence has been increasingly extending its influence over businesses around the world, which has created a steady demand for AI professionals. While there have been significant discussions around the opportunities available for qualified professionals, their efforts still need to be directed towards achieving specific business outcomes. Enterprises, as well as mid to senior managers, are looking forward to effectively leveraging AI professionals into maximising their productivity. This has set the stage for a secondary demand for qualified AI business leaders. Today, business leaders need a proper understanding of the capabilities and limitations of AI to transform their businesses. The Indian tech industry is set to add 3 million new jobs over the course of the next five years. In fact, by the year 2023, the size of India’s tech-based jobs is predicted to touch 7 million. This includes Indian tech firms, MNCs, global capability centres of hundreds of international firms, enterprises across segments including e-com, BFSI, pharma, telecom, as well as over 1,300+ captive units. Research also suggests that 54% of employees of large companies need significant reskilling and upskilling in order to fully harness the growth opportunities offered by the Industry 4.0. Reports suggest that Indian executives are way more optimistic when it comes to the leadership taking the appropriate decision when it comes to usage and adoption of AI. In fact, 70% of the employees think that artificial intelligence will change the way an enterprise functions over the next decade. Neeti Sharma, Senior Vice President at noted recruitment firm TeamLease Services says, “Technology is creating opportunities and making way for newer jobs — many of them still unknown to us. The convergence of big data with AI has emerged as the single most important development that is shaping the future of how firms drive business value from their data and analytics capabilities.” Today’s recruiters from large organisations expect potential candidates to not just know the fundamentals of artificial intelligence, but they also want the applicants to have a domain understanding along with skills in data visualisation, computer vision, NLP, cloud platforms and deep neural networks. More importantly, today’s leaders are expected to be able to apply AI techniques to real-world scenarios such as data augmentation, sentiment analysis, product life cycles and even client servicing. Artificial intelligence has emerged as a highly sought after skill in the new IT sector and there’s a need for a well-rounded applied AI course that will allow professionals to transition to this highly lucrative field. Great Learning’s intensive 4-month PGP Artificial Intelligence for Leaders programme covers the foundation of AI, data visualisation, computer vision, NLP, cloud platforms, Neural networks, and other core areas. The unique combination of hands-on online plus live virtual classes will equip professionals with the most in-demand skills in the growing and lucrative field of artificial intelligence. This programme also focuses on a conceptual and real-world case-based approach which helps to simulate working scenarios where AI can be adopted to improve business outcomes. Hari Krishnan Nair, Co-Founder, Great Learning, said, “Our current cohort of learners has an average work experience of 18 years with a significant number of CXOs, AVPs, and Directors from leading MNCs, who all come with a set of unique expectations from the program. The companies represented in the program include Oracle, GE, Verizon, Cisco, TCS, Indian Navy and many more. Since Artificial Intelligence is seen as a game-changer to drive new business models and transform workplaces, these mid and senior leaders are keenly interested in familiarising themselves with it. The increasingly engaged participation and inquisitive queries from senior professionals are indicative of the growing criticality of AI in the current business scenario. The program offers participants an immersive experiential learning of AI from a business leader’s perspective with sessions conducted by senior academicians and industry leaders who are at the forefront of implementing AI and Machine Learning at their respective organisations.” The programme, which offers a dual certificate from The University of Texas at Austin – McCombs School of Business and Great Lakes Executive Learning, is suitable for business professionals who are looking to leverage the power of AI in their day to day decision making. Relevant profiles include product managers, directors, category managers, CXOs, Delivery managers, product managers, senior managers and team leads who want to upskill themselves. Program Structure Program Faculty Learn from leading academicians in the field of Artificial Intelligence and Machine Learning and several experienced industry practitioners from top organizations. What You Will Learn from Great Learning’s PGP Artificial Intelligence For Leaders Programme: The programme will equip students to handle the changes in the evolving AI-powered industry with the help of 7 live virtual classes, 4 industry sessions, 5 case studies and one intensive capstone project.By the end of the 4-month course, learners would have understood enough concepts in artificial intelligence to be able to make important managerial-level decisions.A key takeaway from the program is that students will be able to deliver transformative projects to external and internal clients and stakeholders.In addition, the professionals will be able to manage technical teams through the lifecycle of AI projects.As far as technical skills in emerging tech are concerned, the graduates will be well-equipped to make appropriate choices when deciding between tech stacks or products in any enterprise.Experiential learning has always been a core part of Great Learning’s programs and during this program, students will be enabled to lead organizations to the new AI world as they develop AI-enabled products and services. Hard Facts Course Name: Post-graduate Programme in Artificial Intelligence for LeadersMode: Online + Live Virtual ClassesDuration: 4 monthsFees: ₹1,50,000 + GSTCertification: Dual Certificate from The University of Texas at Austin – McCombs School of Business and Great Lakes Institute of Management.Pre-Requisites: A minimum of 8 Years of work experience (No prior programming knowledge is required).Key Technical Learnings: How to build AI projects & teams, How to calculate ROI of an AI project,  Data visualisation, Basics of prediction using ML, Neural networks, Computer vision, NLP, Building POC for an AI project.","excerpt":"Artificial intelligence has been increasingly extending its influence over businesses around the world, which has created a steady demand for AI professionals. While there have been significant discussions around the opportunities available for qualified professionals, their efforts still need to be directed towards achieving specific business outcomes. Enterprises, as well as mid to senior managers, […]","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","career in business intelligence","Great learning","leader of ai"],"author_name":"Prajakta Hebbar","publish_date":"2019-12-16T15:00:00","publication_year":"2019","word_count":998,"keywords":["artificial intelligence","machine learning","AI","neural network","sentiment analysis","ML","computer vision","RAG","career in business intelligence","NLP","analytics","Great learning","leader of ai","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","NLP","computer vision","analytics","RAG","sentiment analysis"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/become-an-ai-proficient-business-leader-with-this-unique-programme\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":6653,"title":"INSOFE’s CPEE Adds SPARKLE – Offers Full Fee Waiver and Paid Research Assistantships","content":"Scholarships and paid research and teaching assistantships are generally associated with Masters and PhD programs in the US.  These are certainly not offered for shorter-term certificate programs anywhere in the world.  International School of Engineering (INSOFE), Hyderabad is set to change all that with their SPARKLE – Scholarships and Paid Assistantships to ReKindle LEarning – program. INSOFE recently announced that they will be offering full scholarships and paid assistantships to 5 people who enrol in their upcoming 13th batch of the Certificate Program in Engineering Excellence (CPEE) in Big Data Analytics and Optimization starting June 30, 2015. The CPEE program is listed # 3 globally alongside certificate programs of Columbia and Stanford universities among “11 Big Data Certifications That Will Pay Off” by CIO.com. At a time when education, especially world-class education, is becoming very expensive, this initiative by INSOFE brings a breath of fresh air.  When asked about the motivation for this initiative, Dr. Sridhar Pappu, co-founder and Executive VP – Academics at INSOFE, said, “It is our way of attracting the brighter minds in this country to come and take up knowledge work rather than be stuck doing regular jobs and wasting their talent.  And they should be able to do it without incurring any financial burden, even if our pricing is lower than that of similar programs from other reputed institutes.” So, what does it really cost to do this 6-month certificate program that is certified for its quality of content, pedagogy and assessment by the Language Technologies Institute (LTI) in the School of Computer Science at Carnegie Mellon University?  “The selected fellows will be given complete fee waiver of Rs 2.85 Lakhs inclusive of taxes.  Not only that, they will also earn Rs 1.5 Lakhs during the program (Rs 30,000 per month from the second month of the program) as Research Fellows working on challenging, cutting-edge, real-world client projects under the expert guidance of INSOFE’s globally acclaimed faculty and team of senior Data Scientists”, said Dr. Dakshinamurthy V. Kolluru, co-founder and President of INSOFE. INSOFE is offering these scholarships to people with 0-3 years of work experience.  The candidates will have to take an in-person examination test on 3rd May’15 at INSOFE campus in Hyderabad in addition to their regular online entrance, followed by a group discussion and a personal interview, to qualify for these awards.  They will have to work 32 hours per week from INSOFE’s office and attend classes over the weekends. “If you consider yourself to be a gem, let INSOFE bring out the sparkle in you.  Instead of paying, get paid to get the highest quality education free of cost”, added Dr. Pappu. To learn more about INSOFE’s CPEE and the SPARKLE program, www.insofe.edu.in or call +91 9502334562","excerpt":"Scholarships and paid research and teaching assistantships are generally associated with Masters and PhD programs in the US.  These are certainly not offered for shorter-term certificate programs anywhere in the world.  International School of Engineering (INSOFE), Hyderabad is set to change all that with their SPARKLE – Scholarships and Paid Assistantships to ReKindle LEarning – […]","categories":["AI Trends"],"tags":["international school of engineering hyderabad","masters in data analytics in india"],"author_name":"Dr. Dakshinamurthy V Kolluru","publish_date":"2015-01-02T14:10:30","publication_year":"2015","word_count":456,"keywords":["big data","Go","programming_languages:R","AI","data_tools:Spark","programming_languages:Go","masters in data analytics in india","Aim","analytics","R","international school of engineering hyderabad"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","big data","programming_languages:R","programming_languages:Go","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/insofes-cpee-adds-sparkle-offers-full-fee-waiver-paid-research-assistantships\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164509,"title":"Google Secures Cloud Deal with Salesforce to Counter Microsoft","content":"Salesforce and Google expanded their partnership on Monday, integrating Google’s Gemini AI into Salesforce’s Agentforce. This allows agents to process images, audio, and video, handle complex tasks with Gemini’s multi-modal capabilities, and provide real-time insights using Google Search with Vertex AI. “Through our expanded partnership with Google Cloud and deep integrations at the platform, application, and infrastructure layer, we’re giving customers choice in the applications and models they want to use,” said Srini Tallapragada, chief engineering officer at Salesforce. The partnership deepens integration between Salesforce Service Cloud and Google Customer Engagement Suite, improving AI contact centres. New features include real-time voice translation, intelligent agent handoffs, and AI-driven insights to improve customer service experiences. Agentforce will also employ Google Search and Gemini AI for real-time insights, enabling industries like logistics, insurance, and customer service to make faster, data-driven decisions. Additionally, new Slack and Google Workspace integrations will improve productivity by making it easier to search, share, and act on data across platforms. Salesforce research estimates that AI-powered agents are a $2 trillion opportunity. Nearly 84% of CIOs see AI as transformative as the internet. The Salesforce-Google partnership gives businesses the flexibility, security, and AI tools to scale while ensuring trust and openness in their AI strategies. According to reports, Salesforce was in talks with other providers like Microsoft and Oracle as well.","excerpt":"Salesforce research estimates that AI-powered agents are a $2 trillion opportunity.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Salesforce"],"author_name":"Aditi Suresh","publish_date":"2025-02-25T11:19:58","publication_year":"2025","word_count":221,"keywords":["Go","AI","Modal","data-driven","RAG","llm_models:Gemini","ViT","Salesforce","cloud_platforms:Google Cloud","Rust","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","ViT","data-driven","Modal","llm_models:Gemini","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-secures-cloud-deal-with-salesforce-to-counter-microsoft\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171945,"title":"AI May Be a Threat to Freelancers, But Here’s the Silver Lining","content":"AI is both a boon and a bane. When one AI application completes a task quickly, it simultaneously displaces a real human who was paid to do that job. There has been considerable debate about the impact of AI on freelancing, with several entry-level jobs and projects already being taken over by AI. However, research from the Wharton School of the University of Pennsylvania in the US has highlighted how the introduction of ChatGPT to the online labour market isn’t all bad news. The study analysed millions of records from a leading global freelancing platform, both before and after ChatGPT’s launch. The records included data on how over three lakh freelancers applied for jobs each month. Freelancers are Strategically Adapting to AI The findings showed that fields like copywriting and language translation faced a direct threat from AI. Independent contractors submitted between 51% and 62% fewer bids post-ChatGPT—a drop that persisted for nearly a year. Segments most affected by the drop in demand resorted to “horizontal repositioning”, where they were more likely to tailor their bids to various specialisations than their previous approach. Moreover, they also increased the percentage of their bids on high-value jobs. While the demand for freelancers in sectors like software development and data analytics was less affected, the platform saw an increase in the number of freelancers bidding for work. The report attributed a portion of the rise to the use of AI among international freelancers to bypass barriers like language and formal training, among others. For those affected by the rise in supply, the shift in their “vertical position” involved targeting lower-value jobs, which the study suggests could be a response to a more crowded and competitive market. Source: Wharton School Research Highly skilled freelancers were less likely to reduce their overall activity on the platform compared to their peers. They made fewer significant changes to the distribution of their bids across different work domains, suggesting a reluctance to move away from their area of expertise due to higher adjustment costs. “Our findings, perhaps, are more in line with the story where AI is levelling the skill deficit,” Manav Raj, one of the report’s authors, said. “It’s allowing lower-skilled workers to compete with higher-skilled people, which may be why higher-skilled workers are applying for a broader range of jobs now.” The document contains further details from the report, including detailed statistics and findings. The study concludes that freelancers across all segments are making strategic changes to how they work to adapt to AI, which seems to be the means of survival, for now. “I don’t think we’re at a point where freelance work is under existential threat, but the demand for it will shift,” said Raj. However, there seems to be an optimistic side. As AI capabilities continue to advance, freelancers, product engineering consultancies and studios can use these tools to increase their productivity, improve the quality of their deliverables, and shorten project timelines. The True Potential of AI for Freelancers and Studios AIM reached out to Aditya Chhabra, founder of CreateBytes, a design and engineering studio based in Gurugram. “A major focus for our engineers is to consider the design and the framework, and then the code is generated by AI agents,” Chhabra said, adding that this process helps them build quality deliverables at scale. Moreover, this also allows developers to spend more time on reviews and feedback. Chhabra also pointed out that using AI in product development improves the experience for his clients. “Most of our clients get very excited when we tell them this (deliverables) can be completed quickly. They get extraordinarily responsive, and happy about that,” he added. Chhabra explained that customers often approach engineering or design firms with a well-defined requirement that they believe fully captures their needs. However, by combining AI capabilities with deep domain expertise, it’s often possible to uncover a more strategic and valuable direction for the product. According to him, the real value lies in guiding clients beyond their initial specifications, helping them see that what they first envisioned may only be a starting point. As the development process unfolds, new opportunities often emerge. With the introduction of intelligent tools and autonomous agents, the product can evolve into something significantly more advanced and impactful than initially imagined.","excerpt":"There’s an opportunity to adapt to the capabilities of AI, using AI.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","freelancing"],"author_name":"Supreeth Koundinya","publish_date":"2025-06-18T17:30:00","publication_year":"2025","word_count":711,"keywords":["ChatGPT","autonomous agents","AI","GPT","Aim","AI agents","freelancing","analytics","ViT","R","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","analytics","ChatGPT","Aim","autonomous agents","R","GPT","ViT","AI agents","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-may-be-a-threat-to-freelancers-but-heres-the-silver-lining\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":33392,"title":"Top Trends For Stream Processing Expected To Affect Enterprises In 2019","content":"Data serves as the life source for many organizations many of whose customers rely on live stream data. Take for example applications that provide live updates such as Uber or Google Map. Stream processing has set a new standard of customer experience in terms of how users relate to data. By definition, stream processing is a technology that let users query continuous data streams and detect conditions quickly within a small time period from the time of receiving the data. The detection time period varies from few milliseconds to minutes. As opposed to the traditional data processing which follows a store and process procedure, stream processing allows live and incoming data to be processed simultaneously and continuously. Stream processing simplifies the offering of personalized services to customers and instantaneous response to issues. According to a report, the world data will grow from 33 zettabytes in 2018 to a massive 175 zettabytes by 2025 of which nearly 30% will be live stream data. The report also says that the number of consumers who interact with data every day will rise from 5 billion in 2018 to 6 billion in 2025 which would make up 75% of the world population. With data only expected to grow, organisations have to come up with efficient ways to process live data. Here are the top trends for stream processing expected to hit the enterprises in 2019 Need For Distributed Stream Processing Machine Learning models heavily emphasize on data and with more and more AI applications and the use for live data going up, distributed stream processing will become a necessity. Distributed and high-performance stream processing frameworks will be required to handle complex real-time data efficiently. Greater Bandwidth For IoT, More Data To Process With Fifth generation, cellular communication already on its way, and more IoT devices hitting the market, more real-time streaming data will be created and thus more use cases that need instant reaction to events. Edge computing will also increase the need for stream processing. Better Compliance With GDPR With both consumers and organizations growing concern over the privacy of sensitive data, stream processing will lead a new path which is more GDPR compliant than traditional “Store and process later” architectures as Stream processing does not require long term storage of data and sensitive information can be kept isolated in the application state for a limited time. Stream Processing Helps Cyber Security Data breaches and other cybersecurity threats are on the rise and stream processing is expected to provide a great deal of help. Stream processing will be emphasized in Cyber Security as it brings in real-time gathering and aggregating of events, tracking complex patterns, evaluating and adjusting ML models over the real-time data among other features into the plate. Complex Data Calls For Stream Processing Stream processors will have an edge over relational databases as Stream processors can process ACID transactions directly across streams. Stream processing will add more flexibility to the processing of data as multiple and overlapping streams can be resolved simultaneously. Outlook Stream processing, by all means, proves to be better than traditional data processing architectures and has high hopes for future with data growing enormously every second. Stream processing promises a better and more consumer-friendly approach to the management of data and its privacy and hence we can expect this trend to go up in the future.","excerpt":"Data serves as the life source for many organizations many of whose customers rely on live stream data. Take for example applications that provide live updates such as Uber or Google Map. Stream processing has set a new standard of customer experience in terms of how users relate to data. By definition, stream processing is […]","categories":["AI Trends"],"tags":["iot enterprise architecture","iot friendly database"],"author_name":"Amal Nair","publish_date":"2019-01-14T10:17:08","publication_year":"2019","word_count":559,"keywords":["Go","stream processing","machine learning","programming_languages:R","AI","iot enterprise architecture","ML","RAG","GAN","iot friendly database","edge computing","R"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","edge computing","R","Go","stream processing","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-trends-for-stream-processing-expected-to-affect-enterprises-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":30843,"title":"Tableau’s NLP-Driven Ask Data Will Transform BI And Data Visualisation","content":"Image Courtesy : Tableau Tableau, a platform which specialises in visualisation software for analytics recently added a feature which uses natural language in its software. The feature called ‘Ask Data’, allows users to work with data by just conversing with the software. What Is Ask Data? Fully embedded into the Tableau server, Ask Data is currently available in the beta version of Tableau 2019.1. Powered by machine learning algorithms for NLP, its analytical capabilities get an obvious boost and may give other visualisation software providers a run for their money. It requires only the right data source without the need for any kind of setup. With Ask Data, a user can ask questions of any published data source and get answers in the form of a visualisation. “It allows you the ability to explore data at the speed of thought”, says the company. It will empower users to have powerful insights and get quick answers to make better data-driven decisions. Ask Data also allows one to analyse the data with a simple statement or questions. It implements smart analytics as the technology behind Ask Data understands ambiguous or underspecified statements and resolves them by offering helpful recommendations and more. In spite of all these developments, can Ask Data revolutionise business intelligence and visualisation? Here are the arguments: Bringing In The AI game Software development, in general, is getting an AI touch these days. Most software available today is looking into AI tech one way or the other. Be it in terms of face recognition or voice recognition, ML applications are growing. Tableau is no different here. The company is pacing up with AI getting into business intelligence tools today. AI solutions make up processes much more efficient and easier. In an article by Peter Sayer of CIO, he highlights how analysts believe Tableau’s new language feature can help a long way in automating specialised tasks for data scientists. Sayer says, “Data scientists spend up to 80 percent of their time on data preparation, she said (referring to Forrester analyst, Martha Bennett), and the less time they spend on it, the more they can spend on things that create value. One way around the time crunch is to hand over workloads to the machines. Another is to make it easier for people that couldn’t previously manipulate the data themselves to do so, the so-called democratization of data.” Is Language At The Helm Of Better Understanding? In his blog, Ruhaab Markas, Senior Product Manager, NLP at Tableau, illustrates how Ask Data has leaned on ML advancements to leverage language for visualisation. “Natural language can be difficult. For example, in the statement, ‘six-foot doctors came to the hospital’, it’s not clear if we’re talking about really tall doctors or podiatrists. Humans are adept at clarifying ambiguity and intent by understanding the context of a conversation, but machines face a more difficult challenge. Now, with advances in the field of natural language, machines can better handle pragmatics, deciphering context to better understand the meaning behind a statement.” In fact, if visualisation software understands broken down queries, they might accurately grasp the context of what the user actually wants (just like Google Search but in terms of graphs, charts or other visualisation elements). Ultimately, visualisation tools have more to offer than lose when natural language is infused in its structure. Aligning Best Practices Markas also mentions how Ask Data goes a long way in aligning with best practices for data visualisation. For example, if the user works on a visualisation project, Ask Data recommends the right type of visualisation tool based on best practices. Furthermore, Markas emphasises that the data can be driven towards more useful decisions by sharing insights and asking questions. All one has to do is save visualisations created as Tableau workbooks and share with Tableau users. Outlook Tableau’s latest offering will definitely attract more crowd. But, all depends on how well it tweaks its natural language features in visualisation as well as for BI. With rival software like Microsoft’s Power BI (it already has a feature called Q&A For Power BI, that uses language) and Qlik’s Qlikview racing up in the quest for implementing AI into their portfolio, it is essential that Tableau keeps up very well with Ask Data.","excerpt":"Tableau, a platform which specialises in visualisation software for analytics recently added a feature which uses natural language in its software. The feature called ‘Ask Data’, allows users to work with data by just conversing with the software. What Is Ask Data? Fully embedded into the Tableau server, Ask Data is currently available in the […]","categories":["AI Features"],"tags":["nlp in data","nlp pipeline"],"author_name":"Abhishek Sharma","publish_date":"2018-11-29T09:02:07","publication_year":"2018","word_count":709,"keywords":["business intelligence","Go","machine learning","AI","R","ML","data-driven","RAG","NLP","analytics","nlp pipeline","nlp in data"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","analytics","RAG","R","Go","business intelligence","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/tableaus-nlp-driven-ask-data-will-transform-bi-and-data-visualisation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10004444,"title":"Step By Step Guide To Stabilize Facial Landmarks In A Video Using Dlib","content":"The human face has been a topic of interest for deep learning engineers for quite some time now. Understanding the human face not only helps in facial recognition but finds applications in facial morphing, head pose detection and virtual makeovers. If you are a regular user of social media apps like Instagram or Snapchat, have you wondered how the filters fit perfectly for each face? Though every face on the planet is unique, these filters seem to magically align on your nose, lips and eyes. These filters or face-swapping applications make use of facial landmarks. These landmarks are basically points that are meant to help with the identification of the distance between eyes, position of the nose, size of the lips etc. In the context of facial landmarks, our goal is to detect important facial structures on the face using shape prediction methods. In this article, we will cover: Need for stabilizationPopular types of landmark detectors Implementation and stabilization of 68 point landmarks for a video. The need to Stabilize Facial Landmarks Facial landmarks are easy to use on images since the pixels are not moving, but when it comes to a video, due to continuous motion of pixels and because of translational and rotational variances, a lot of the times these landmarks are unstable. Take a look at the image below. Few of the points are missing out the features of the face. This can create problems for features that involve estimating the size of the face, the position of the mouth etc. This instability can affect the efficiency of the model and the results. Popular types of landmark detectors The Dlib library is the most popular library for detecting landmarks in the face. There are two types of detectors in this library. 68-point landmark detectors: This pre-trained landmark detector identifies 68 points ((x,y) coordinates) in a human face. These points localize the region around the eyes, eyebrows, nose, mouth, chin and jaw. 5 point landmark detector: To make things faster than the 68 point detector, dlib introduced the 5 point detector which assigns 2 points for the corners of the left eye, 2 points for the right eye and one point for the nose. This detector is most commonly used for alignment of faces. Implementation and stabilization of 68 point landmarks for a video Step 1: Collecting the pre-trained files. Create a folder for your project. Create a subfolder called a model. Download the 68-points and 5-points and place them in the subfolder. Next, place this file in the root folder of your project. Step 2: The data. Select a short 5-10 second video for this project with good lighting. I have chosen this video. Feel free to download it from here. Step 3: Importing the required modules import dlib import cv2 import numpy as np import matplotlib.pyplot as plt import matplotlib import os from google.colab import drive drive.mount('\/content\/gdrive') Copy all the downloaded files to your notebook !cp -r '\/content\/gdrive\/My Drive\/video-stable\/model' \/content !cp -r '\/content\/gdrive\/My Drive\/video-stable\/videos' \/content !cp '\/content\/gdrive\/My Drive\/video-stable\/faceBlendCommon.py' \/content Step 4: Convert the video into image frames and save them to a folder inside your project folder. We will convert the entire video into individual image frames since it makes it easier to work with. Create a main folder for all the images. def image_saver(path, filename, images): for count in range(0, len(images)): temp = filename + '_' + str(count) + '.png' fn = os.path.join(path) + os.path.join(temp) cv2.imwrite(fn, images[count]) cap=cv2.VideoCapture('\/content\/gdrive\/MyDrive\/video-stable\/videos\/video_data.mp4') image_frame = [] while(cap.isOpened()): pic, frame = cap.read() if frame is None: break image_frame.append(frame) cap.release() plt.imshow(image_frame[0][:,:,::-1]) Now that we have the image frames, we will save all these frames in a folder. directory = \"dataset\" parent = \"\/content\/gdrive\/My Drive\/video-stable\/\" path = os.path.join(parent, directory) os.mkdir(path) os.mkdir('\/content\/gdrive\/My Drive\/video-stable\/dataset\/original') image_saver('\/content\/gdrive\/My Drive\/video-stable\/dataset\/original\/', 'frame', image_frame) Step 5: Facial alignment This is an important step in the process. We will use the 5 point detector for aligning the face to the frame to eliminate noise from the background and focus only on the face. modelroot = '\/content\/gdrive\/My Drive\/video-stable\/model\/' five_point_landmark = modelroot + \"shape_predictor_5_face_landmarks.dat\" detect_face = dlib.get_frontal_face_detector() detect_landmark = dlib.shape_predictor(five_point_landmark) Now we will make use of the built in methods of the face blend common to get the detectors and align the face. import faceBlendCommon as fb def facial_alignment(image): faceRects = detect_face(image, 0) print(\"Number of faces detected: \",len(faceRects)) points = fb.getLandmarks(detect_face, detect_landmark, image) print('length of points is', points) landmarks = np.array(points) print('after np array',len(landmarks)) image = np.float32(image)\/255.0 height = 600 width = 600 if len(landmarks) > 0: normalize_image, landmarks = fb.normalizeImagesAndLandmarks((height, width), image, landmarks) normalize_image= np.uint8(normalize_image*255) return normalize_image else: return image aligned_faces = [] print('performing alignment') for count in range(0, len(image_frame)): frame = image_frame[count] alignment = facial_alignment(frame) aligned_faces.append(alignment) print('Done!') Let us check one of the images before saving it. plt.imshow(aligned_faces[50][:,:,::-1]) plt.title(\"Aligned Image\") plt.show() As you can see the background noise has been eliminated and the face has been resized to 600×600 after alignment. Save the aligned images in your images folder. os.mkdir('\/content\/gdrive\/My Drive\/video-stable\/dataset\/aligned_faces') image_saver('\/content\/gdrive\/My Drive\/video-stable\/dataset\/aligned_faces\/', 'align_face', aligned_faces) Step 6: Using the 68 point detector and performing stabilization MODEL_PATH = '\/content\/gdrive\/My Drive\/video-stable\/model\/' PREDICTOR_PATH = MODEL_PATH + \"shape_predictor_68_face_landmarks.dat\" RESIZE_HEIGHT = 480 NUM_FRAMES_FOR_FPS = 100 SKIP_FRAMES = 1 detector = dlib.get_frontal_face_detector() landmarkDetector = dlib.shape_predictor(PREDICTOR_PATH) Now, we will calculate the distance between each eye using the function below def interEyeDistance(predict): leftEyeLeftCorner = (predict[36].x, predict[36].y) rightEyeRightCorner = (predict[45].x, predict[45].y) distance = cv2.norm(np.array(rightEyeRightCorner) - np.array(leftEyeLeftCorner)) distance = int(distance) return distance In order to save the points of detection we create separate lists. points=[] pointsPrev=[] pointsDetectedCur=[] pointsDetectedPrev=[] all_stabilized_frames=[] Next, we set the parameters required for the process and perform the stabilization eyeDistanceNotCalculated = True eyeDistance = 0 isFirstFrame = True fps = 10 showStabilized = False count =0 while(True): if (count==0): t = cv2.getTickCount() ret,im = cap.read() if im is None: break imDlib = cv2.cvtColor(im, cv2.COLOR_BGR2RGB) imGray = cv2.cvtColor(im, cv2.COLOR_BGR2GRAY) imGrayPrev = imGray height = im.shape[0] IMAGE_RESIZE = float(height)\/RESIZE_HEIGHT imSmall = cv2.resize(im, None, fx=1.0\/IMAGE_RESIZE, fy=1.0\/IMAGE_RESIZE,interpolation = cv2.INTER_LINEAR) imSmallDlib = cv2.cvtColor(imSmall, cv2.COLOR_BGR2RGB) if (count % SKIP_FRAMES == 0): faces = detector(imSmallDlib,0) if len(faces)==0: print(\"No face detected\") else: for i in range(0,len(faces)): print(\"face detected\") newRect = dlib.rectangle(int(faces[i].left() * IMAGE_RESIZE), int(faces[i].top() * IMAGE_RESIZE), int(faces[i].right() * IMAGE_RESIZE), int(faces[i].bottom() * IMAGE_RESIZE)) landmarks = landmarkDetector(imDlib, newRect).parts() if (isFirstFrame==True): pointsPrev=[] pointsDetectedPrev = [] [pointsPrev.append((p.x, p.y)) for p in landmarks] [pointsDetectedPrev.append((p.x, p.y)) for p in landmarks] else: pointsPrev=[] pointsDetectedPrev = [] pointsPrev = points pointsDetectedPrev = pointsDetectedCur points = [] pointsDetectedCur = [] [points.append((p.x, p.y)) for p in landmarks] [pointsDetectedCur.append((p.x, p.y)) for p in landmarks] pointsArr = np.array(points,np.float32) pointsPrevArr = np.array(pointsPrev,np.float32) if eyeDistanceNotCalculated: eyeDistance = interEyeDistance(landmarks) print(eyeDistance) eyeDistanceNotCalculated = False if eyeDistance > 100: dotRadius = 3 else: dotRadius = 2 print(eyeDistance) sigma = eyeDistance * eyeDistance \/ 400 s = 2*int(eyeDistance\/4)+1 lk_params = dict(winSize  = (s, s), maxLevel = 5, criteria = (cv2.TERM_CRITERIA_COUNT | cv2.TERM_CRITERIA_EPS, 20, 0.03)) pointsArr,status, err = cv2.calcOpticalFlowPyrLK(imGrayPrev,imGray,pointsPrevArr,pointsArr,**lk_params) pointsArrFloat = np.array(pointsArr,np.float32) points = pointsArrFloat.tolist() for k in range(0,len(landmarks)): d = cv2.norm(np.array(pointsDetectedPrev[k]) - np.array(pointsDetectedCur[k])) alpha = math.exp(-d*d\/sigma) points[k] = (1 - alpha) * np.array(pointsDetectedCur[k]) + alpha * np.array(points[k]) if showStabilized is True: for p in points: cv2.circle(im,(int(p[0]),int(p[1])),dotRadius, (255,0,0),-1) else: for p in pointsDetectedCur: cv2.circle(im,(int(p[0]),int(p[1])),dotRadius, (0,0,255),-1) isFirstFrame = False count = count+1 if ( count == NUM_FRAMES_FOR_FPS): t = (cv2.getTickCount()-t)\/cv2.getTickFrequency() fps = NUM_FRAMES_FOR_FPS\/t count = 0 isFirstFrame = True cv2.putText(im, \"{:.1f}-fps\".format(fps), (50, size[0]-50), cv2.FONT_HERSHEY_COMPLEX, 1.5, (0, 0, 255), 3,cv2.LINE_AA) all_stabilized_frames.append(im) imPrev = im imGrayPrev = imGray cap.release() plt.imshow(all_stabilized_frames[1][:,:,::-1]) plt.title(\"Stabilized Image\") plt.show() As you can see above, the 68 points are applied to the face. We save these images in a folder and move on the final step. os.mkdir('\/content\/gdrive\/My Drive\/video-stable\/dataset\/stable_faces') image_saver('\/content\/gdrive\/My Drive\/video-stable\/dataset\/stable_faces\/', 'stable_face', all_stabilized_frames) Step 7: The last step is to stitch the original, aligned and stable frames back to a video. We will resize the frames to fit the screen and stitch the images together. def read_all_images(dir, filename_prefix, num_files): result_list = [] for cnt in range(0, num_files): fn = filename_prefix + '_' + str(cnt) + '.png' full_path = os.path.join(dir, fn) img = cv2.imread(full_path) result_list.append(img) return result_list original_frames = read_all_images('\/content\/gdrive\/My Drive\/video-stable\/dataset\/original', 'frame', 157) aligned_frames = read_all_images('\/content\/gdrive\/My Drive\/video-stable\/dataset\/aligned_faces', 'align_face', 157) stable_frames = read_all_images('\/content\/gdrive\/My Drive\/video-stable\/dataset\/stable_faces', 'stable_face', 157) def resize_images(imageList, width, height): result_list = [] for cnt in range(0, len(imageList)): new_img = cv2.resize(imageList[cnt], (width, height), interpolation=cv2.INTER_AREA) result_list.append(new_img) return result_list orig_frames_resized = resize_images(original_frames, 800, 800) aligned_resized=resize_images(aligned_frames,800,800) stab_frames_resized = resize_images(stable_frames, 800, 800) stitched = [] for cnt in range(0, len(orig_frames_resized)): new_img = np.hstack((orig_frames_resized[cnt], aligned_resized[cnt], stab_frames_resized[cnt])) stitched.append(new_img) plt.imshow(stitched_frames[50][:,:,::-1]) Let us convert these into a video. image_saver('\/content\/gdrive\/My Drive\/video-stable\/dataset\/', 'stitch', stitched_frames) from os.path import isfile, join def convert_frames_to_video(pathIn,pathOut,fps): frame_array = [] files = [f for f in os.listdir(pathIn) if isfile(join(pathIn, f))] for i in range(len(files)): filename=pathIn + files[i] img = cv2.imread(filename) height, width, layers = img.shape size = (width,height) print(filename) frame_array.append(img) out = cv2.VideoWriter(pathOut,cv2.VideoWriter_fourcc(*'DIVX'), fps, size) for i in range(len(frame_array)): out.write(frame_array[i]) out.release() pathIn= '\/content\/gdrive\/My Drive\/video-stable\/dataset\/' pathOut = '\/content\/gdrive\/My Drive\/video-stable\/videos\/finalvid.mp4' fps = 30.0 convert_frames_to_video(pathIn, pathOut, fps) Here is the final output. The final video shows that despite movement of the lips and facial contortions, the points are stable and are adjusting according to the movement. You can check this video here. Conclusion In this article, we have learned the step-by-step process to stabilize the important landmark for a face in a video. In order to improve the accuracy and precision of face detection or recognition systems, the process of stabilization is very important.","excerpt":"The human face has been a topic of interest for deep learning engineers for quite some time now. Understanding the human face not only helps in facial recognition but finds applications in facial morphing, head pose detection and virtual makeovers. If you are a regular user of social media apps like Instagram or Snapchat, have […]","categories":["Deep Tech"],"tags":[],"author_name":"Bhoomika Madhukar","publish_date":"2020-08-09T18:00:00","publication_year":"2020","word_count":1526,"keywords":["Go","NumPy","TPU","AI","ETL","Colab","Ray","deep learning","Matplotlib","R"],"extracted_tech_keywords":["AI","deep learning","Ray","Colab","NumPy","Matplotlib","TPU","R","Go","ETL"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/step-by-step-guide-to-stabilize-facial-landmarks-in-a-video-using-dlib\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119746,"title":"How SuperKalam Uses OpenAI GPTs to Fuel UPSC Aspirants","content":"As the UPSC Civil Services Exam (CSE) date of June 16th draws closer, over 11 lakh aspirants are diligently preparing to tackle one of the world’s most challenging examinations. SuperKalam, an AI platform is empowering students with personalised learning experiences, paving the way for their success in cracking the UPSC CSE. Vimal Singh Rathore along with Aseem Gupta, both of whom previously founded qoohoo, started SuperKalam in July last year. Vimal who cleared his UPSC in Central Armed Police Forces Exam (CAPF) exam but went on to work at Unacademy and then start his own company, Coursavy, which was acquired by Unacademy. The duo recognised the need for a more adaptable and student-centric approach to UPSC preparation. With his extensive background in education and entrepreneurship, Rathore set out to create a platform that could bridge the gap between traditional coaching methods and the unique requirements of each aspirant. In an exclusive interview with AIM, Rathore shared his insights on the platform’s genesis and its mission to revolutionise UPSC preparation. “The inspiration is pretty simple,” he explained. “There are three examinations in India which kind of change the trajectory of any student or any family’s future or life, which are UPSC, that is for the civil services, NEET to become a doctor, and JEE to become an engineer. These examinations cumulatively are given by 5 million aspirants every year.” Rathore further emphasised the need for a more personalised approach to learning for these exams. He stated, “The entire responsibility of understanding, asking questions, whether they are grasping or not, is completely on the student. The platform somehow, because of this kind of process, was not accountable. They were unable to do a lot about how students can improve their learning outcomes.” Leveraging OpenAI’s GPTs SuperKalam leverages multiple models, including Llama, OpenAI’s GPT-4, and GPT-3.5, to deliver personalised learning experiences. For generating MCQs, Super Kalam relies on OpenAI GPT-4, while reasoning-based problems are tackled using OpenAI GPT-4 Turbo. Super Kalam assists students in creating personalised timetables, sending daily targets, and tracking progress throughout their learning journey. It helps resolve doubts, aids in mastering concepts, and can evaluate a UPSC student’s handwritten answer in less than 60 seconds—a process that typically takes 2-4 weeks in the current education market. The platform also conducts mock tests, identifies students’ strengths and weaknesses, and provides progress reports that foster self-awareness, unlocking their full potential. Rathore revealed, “There are at least 14 Agents that we are using to understand what you are asking Super Kalam.” The platform also uses fine-tuning and prompt engineering techniques to enhance the accuracy and relevance of the generated content. SuperKalam has invested in NVIDIA GPUs and cloud infrastructure to ensure scalability and cost-efficiency, although they declined to mention how many. “We are also leveraging NVIDIA GPUs for certain models to handle the performance” Rathore mentioned. As the platform’s user base grows, the team is exploring partnerships with local cloud service providers to optimise costs and performance. Aspiring to become SuperKalam Last year the founders were a part of the Y Combinator W23 batch and the company is backed by Gustaf Alstromer, a partner at Y Combinator. Rathore brings his experience as a founding team member and growth leader at Unacademy, while Gupta, the CTO, previously worked in the early engineering team at Razorpay. Since its inception, Super Kalam has witnessed remarkable growth. The platform currently boasts a user base of over 46,000 students, with an impressive week 12 retention rate of 78%. Rathore proudly shared, “We launched this latest version of the product on 15th of March. So 355 students who attempted any number of questions that day, the total number of questions that you are seeing, is different for every student.” The impact of Super Kalam’s AI-driven approach is evident in the success stories of its users. Rathore highlighted the journey of Navya, a UPSC aspirant who struggled with consistency and self-doubt. “Navya’s accuracy rate was 56% at 45 questions. And here, Navya is number one ranker, with 234 questions and 84% accuracy,” he beamed. SuperKalam has set its sights beyond UPSC and aims to scale to cater to students attempting other competitive exams. “At the same time, our mission is to make quality education accessible and affordable,” Vimal concluded.","excerpt":"SuperKalam aims to go beyond UPSC and scale to cater to students attempting other competitive exams, including JEE and NEET, in the coming months.","categories":["Global Tech"],"tags":["Startups"],"author_name":"K L Krithika","publish_date":"2024-05-07T12:00:00","publication_year":"2024","word_count":709,"keywords":["OpenAI","AI","AWS","RPA","Scala","RAG","GPT","Aim","prompt engineering","Startups","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","RAG","prompt engineering","AWS","R","Scala","GPT","RPA"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-superkalam-uses-openai-gpts-to-fuel-upsc-aspirants\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":3930,"title":"Analytics Placement &#038; Salary Report 2013 by Jigsaw &#038; AIM","content":"Analytics is a strange mix of art and science. It is used to analyse large amounts of data by applying statistical tools and techniques in order to generate business insights. Analytics has traditionally been used in industries like financial services, retail, healthcare and telecom. Lately however, it is finding uses in various new fields as well. Sports, social media, e-commerce, gaming and even HR are a few examples of recent adopters of analytics. The 2013 Analytics Placement and Salary Report by Jigsaw Academy and Analytics India Mag (AIM) is for both aspiring analytics professionals and employers alike. Consider it your personal handbook or resource to help you navigate today’s hiring environment, determine remuneration levels and plan budgets. [quote style=”1″]Navigate today’s analytics hiring environment, determine remuneration levels and plan budgets.[\/quote] We give you a portrait of the industry this past year, while also giving some predictions about the future. We have benchmarked salaries sector wise, allowing employers to compare their remuneration standards to industry standards. Employees and wannabe analytics professionals on the other hand can use this guide to validate their earnings or remuneration packages offered. This guide is divided into 3 parts. The first part gives a portrait of the past year and predictions from our panel of experts about what is to come. In the second part, we focus on analytics recruitment. We spoke to a number of people who have gone through the analytics recruitment process (some as recruiters and others as aspiring recruitees). Based on their invaluable experience, we have identified tips and best practices for those who are looking to hire and those who are looking to get hired in analytics. In the final part of the report, we present the results of the salary survey that was conducted across India. We hope you find this guide useful. If you have any questions on the current hiring landscape in the analytics industry, you can reach out to our team at +91-92435-22277 or info@jigsawacademy.com. You can visit us at www.jigsawacademy.com and www.analyticsindiamag.com. [divider top=”1″] In this section, we have compiled detailed information on the salaries in analytics – at different experience levels as well as across different kinds of analytics organizations. The first split is across the various experience levels. We have divided the population of analytics professionals into 4 different levels based on years of experience within the analytics industry. Entry Level – 0 to 2 years Mid Level – 2 to 5 years Senior manager level – 5 to 10 years Director level – > 10 years The second split is by the type of analytics organization. We have divided the various analytics organizations into 4 types – Knowledge process outsourcers or KPOs – This includes companies such as Genpact, Accenture, WNS etc. which offer analytics as one of the outsourced services in the portfolio. IT Companies – This includes IT companies such as Wipro, TCS, Infosys etc. that are now looking at analytics as the next growth engine. In-house units – This includes companies that use analytics to support their own businesses such as HSBC, Citibank, Tesco, Target etc. Niche analytics companies – This includes specialized analytics companies such as Mu Sigma, Fractal analytics, Marketelligent and Gramener. We have presented the average salaries across the 4 experience groups as well as across the 4 different types of analytics organizations. Further, in order to capture the broad range the salaries cover, we have also added the 10th and the 90th percentile salary numbers for each group. KPOs Knowledge process outsourcing is a part of business process outsourcing or BPO as it is commonly known. KPOs offer outsourcing services for high-end processes like analytics as opposed to BPO activities such as Customer service. There are hundreds of KPO companies in India and the size of the KPO industry is estimated at upwards of $10 billion. They offer a variety of high-end services to their clients and analytics is one of them. KPOs offer exciting opportunities to those who are looking to move into analytics. Companies like Genpact, WNS, McKinsey Knowledge centre are all examples of such companies. Our study found that KPOs offer the highest average salaries at entry level positions amongst the 4 different types of analytics organizations. With an average salary of Rs. 5.6 lakhs at entry levels, KPOs are the most financially attractive option to enter into analytics However, as we move up the experience levels, the difference in salaries between KPOs and others reduces and KPOs are not as attractive for someone with 5 years and more of experience.. With an average salary of Rs. 5.6 lakhs at entry levels, KPOs are the most financially attractive option to enter into analytics. Financial benefits aside, KPOs are able to offer some other options that make them a good place to work. KPOs offer analytics services to a wide variety of clients. This means that they are able to offer their employees opportunities to work on different clients and different industries. Our study shows that those working in KPOs for 2 years or more have worked on 4 different clients and 2 different industries on an average. The flexibility to work on different clients and industries is a big plus for KPOs and makes them a favoured destination for those starting off in analytics. IT Companies India witnessed the IT revolution in the 90s. Companies like Infosys and Wipro have built billion-dollar businesses by servicing the IT needs of organizations around the world. With growth in IT stagnating in the last few years, these companies are now looking at analytics as the next big growth driver. Indian IT companies are investing heavily in building analytics capabilities and they offer another option to analytics aspirants. IT companies are trying to emulate the business model of the KPOs and we found a lot of similarities between the two. IT companies also offer attractive entry level salaries in a bid to lure the best analytic talent from engineering colleges and b-schools. They are also able to flex their muscle at the higher levels and offer fairly attractive salaries even at senior positions. There is a disadvantage of working in the analytics team of an IT company as well. IT companies have a strong software mentality and sometimes they struggle to adapt to the reality of the analytics market. With entry level salaries of Rs. 5.5 lakhs and director level salaries averaging at Rs. 23 lakhs, IT companies certainly offer an attractive option to analysts. Analytics professionals are used to working in lean and nimble setups and the bureaucracy of a large IT company can be frustrating at times. However, with entry level salaries of Rs. 5.5 lakhs and director level salaries averaging at Rs. 23 lakhs, IT companies certainly offer an attractive option to analysts. In House Units In-house or captive units differ from the other analytics companies in the sense that the business they are serving is not a client’s but their own. Banks like Citibank and HSBC and retailers like Tesco and Target are good examples of companies with captive analytics teams. The big advantage of working in a captive is that one is much more closely aligned to the business here. Professionals working in these captive units felt that it was easier for them to see the impact of their analyses because it was for their own business. Working for your own business has advantages of proximity but it also has a string disadvantage. Captive units do not offer you a chance to work on different industries or different organizations. Obviously, the client and the industry always remain the same. Captive units have to offer the highest salaries to retain their analytics talent. Perhaps, this is the reason why captive units have to offer the highest salaries to retain the people. Salaries at mid level, manager level and director level are the highest in captive units. Analytics aspirants who are focused on one particular industry will find captive units to be an ideal place to work. They encourage specialization and emphasize building domain knowledge. Niche analytics companies This category comprises of companies (mostly small and mid-size) that focus on providing analytics expertise to businesses around the world. Companies like Mu Sigma, Fractal analytics and Gramener are good examples of such companies. Niche analytics companies offer the lowest salaries among the four different types of organizations. Salaries average at 4.5 lakhs at entry level and go up to 16 lakhs for director level (10+ years of experience). However, they make up for it to a certain extent by also providing varied and challenging work. The quality of work in some of the niche analytics companies can be very high. Salaries at niche analytics companies average at 4.5 lakhs at entry level and go up to 16 lakhs at director level. Additionally, growth opportunities are abundant in such companies. If one is looking for exciting and challenging work, niche analytics companies are an attractive option. Comparison of salaries across different groups In this table we can see the average salaries by years of experience for all the 4 different types of analytics organizations. KPOs offer the highest average salary at the entry level. IT companies are the next best pay masters coming in slightly below the KPOs. In-house analytics teams or captive units offer the highest salaries at all the other levels – i.e. the mid level, the manager level and the director level. IT companies also offer good salaries at the director level (10+ years of experience), coming in just behind the captive units. Conclusion There are plenty of different options for analytics aspirants. If one is looking to enter this field, KPOs and IT companies offer financially attractive options. Niche analytics companies offer more varied work which can sometimes be very challenging. In-house units offer the chance to focus on one industry and build extensive domain knowledge in the area. * – Minimum salaries correspond to the 10th percentile in the data ** – Maximum salaries correspond to the 90th percentile in the data Salary comparison by city As part of the study, we also conducted a city wise survey. We have collected and compared average sala- ries for various cities in India as well as the national average. Average salary in analytics in India is Rs. 10.35 lakhs. 60% of the analytics professionals in India earn over Rs. 6 lakhs per annum and 20% earn over 15 lakhs per annum. Mumbai has the highest salary amongst all cities at Rs. 11.49 lakhs. Bangalore comes a close second at Rs. 11.34 lakhs. If we adjust these numbers for the cost of living, Bangalore will actually have the highest salaries in analytics amongst all cities in India. [attachments docid=”3946″]","excerpt":"Analytics is a strange mix of art and science. It is used to analyse large amounts of data by applying statistical tools and techniques in order to generate business insights. Analytics has traditionally been used in industries like financial services, retail, healthcare and telecom. Lately however, it is finding uses in various new fields as […]","categories":["AI Features"],"tags":["data scientist india salary","Fractal Analytics","jigsaw academy","masters in data analytics in india"],"author_name":"Gauravohra","publish_date":"2013-08-03T06:44:51","publication_year":"2013","word_count":1784,"keywords":["Fractal Analytics","Go","ELT","AI","data scientist india salary","RAG","masters in data analytics in india","Aim","ViT","analytics","Rust","GAN","jigsaw academy","R"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Rust","ELT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-placement-salary-report-2013-by-jigsaw-aim\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10162035,"title":"Meta Launches CSR Initiative to Address India’s Challenges","content":"Meta India announced a partnership with The\/Nudge Institute to launch Pragati: AI for Impact. This CSR initiative aims to help nonprofits use AI to address challenges in healthcare, agriculture, education, and inclusion. “This program reflects Meta’s commitment to leveraging AI to drive meaningful social impact, particularly in underserved communities across India,” Sandhya Devanathan, VP of India at Meta, said. The program offers grants of ₹40 lakh to ₹1.5 crore for 10-12 startups each to develop and implement AI solutions. It also aims to help these nonprofits receive guidance from experts through workshops. Proposals can be submitted before January 26 using the following link. The initiative is set to incubate 10-12 startups. To be eligible, interested organisations should have been operational for one to 10 years or be part of The\/Nudge Institute. They should also have at least one full-time founder and not have used developed solutions for commercial purposes. In short, the AI solutions must be open-source projects. Devanathan added that the program includes a special focus on women-led social enterprises. Founded in 2015, The\/Nudge Institute is a social enterprise incubator that funds nonprofit startups. Its first incubator cohort, which included 10 startups, launched in 2017. In 2022, the organisation marked the incubation of over a hundred nonprofit startups. Besides Meta, even Google announced plans to support nonprofits a few days ago. Through its philanthropic arm Google.org, the tech giant has announced its next generative AI accelerator program. The six-month-long accelerator will support nonprofit organisations in using generative AI to build solutions. The initiative will provide Google Cloud credits and pro bono support from Google employees, along with a portion of $30 million in funding. This is Google’s second such initiative. The first one, held in 2024, brought together 21 nonprofits. Last year, Google.org awarded a $1 million grant to Karya, an Indian nonprofit that provides low-income communities worldwide with AI-driven learning and earning opportunities. Karya plans to utilise the grant to develop a skilling curriculum based on research and practical experience, which will be translated into 10 major Indic languages.","excerpt":"The Pragati: AI for Impact CSR initiative offers grants of ₹40 lakh to ₹1.5 crore to 10-12 startups each to use AI for challenges in healthcare, education, agriculture and inclusion.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Meta"],"author_name":"Supreeth Koundinya","publish_date":"2025-01-23T12:56:07","publication_year":"2025","word_count":341,"keywords":["Anthropic","Go","Meta","funding","AI","RAG","Aim","generative AI","GAN","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","generative AI","Anthropic","Aim","RAG","R","Go","GAN","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-launches-csr-initiative-to-address-indias-challenges\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":67869,"title":"The Difference Between Various Data Science Job Titles","content":"As the data science field has blown in popularity, it is important to note that there are other job titles with an overlap of functions. Job titles are so confusing nowadays that one company might label a designation something that is completely different somewhere, and so mainly focuses on what the responsibilities, technical skills and experiences will be when it comes to job titles related to data. In this article, we take a look at such similarities and differences in data job titles. Data Scientist vs Data Engineer: What’s Common and Different? Data scientists are typically seen at work for all the data related research and very little data housekeeping, but it turns out that other organisations may vary on the data science work even though they may share the same job titles. Some companies want data scientists to maintain databases or also gather data themselves. Data engineers, on the other hand, enable data scientists to perform their data operations. The main responsibility of a data engineer\/architect is to define the ways\/processes in which data is stored, managed, and utilised by various people in the company, assuring that organisation is devoid of problems, and so data scientists\/analysts have the best tools and systems for their analysis. Many times data scientists do the research work, and it’s up to developers\/and or data engineers to source the data and deploy the models. A data engineer has a deep knowledge of engineering and testing tools, which may not be the necessary case with data scientists. It is usually the responsibility of a data engineer to handle the entire data pipeline and architecture. Data engineers build pipelines so that data scientists have all the data they want for their research and then models are built and sent to production by engineers. The engineering team knows our database well and can do it efficiently, along with the pipelines needed for the model to be deployed. Other companies want their data scientists to be everything from DevOps to business analysis. In such companies, it is the responsibility of data scientists to have clean, defined and readily available data sources, not the developers\/engineers. Data Scientists Vs Data Analysts: What’s Common and Different? Data analysts collect data, organise it and utilise the data for getting useful analytical insights. Data analysts expend their skills in building processes for collecting data and bringing business insights using that data. Both data scientists and data analysts write queries to source data and clean it for analysis, which then leads to deriving insights using business intelligence tools. On the contrary side, a data analyst is not supposed to create advanced machine learning algorithms and model building for predicting critical business outcomes. Data visualisation is a major component for data analysts, presenting data in a manner, which helps business leaders to make quick decisions that are centred on the day-to-day operations. This is also something which overlaps with a data scientist, who needs strong data visualisation skills and capabilities to convert data into a business story. Data analyst, data scientist and data engineer skills may overlap. The various job roles require some fundamental math knowledge, along with algorithms, models, programming\/software and communication. Data Scientists Vs Statistician: What’s Common and Different? Like data scientists, statisticians also collect, analyse, and come up with conclusions from data. But, the manner in which they collect data is not the same as how data scientists collect data. Statisticians typically use smaller and conservative methods of data collection like surveys, experiments, and polls. Data science, on the other hand, is more focused on predictive analytics, building models, with a background with math, statistics and computer programming. Here the difference is that statisticians may lack advanced knowledge of software, programming, and AI algorithms. Statisticians are typically employed by companies that are connected in public opinion and market research, product development and government agencies.  Unlike data scientists who work on comparing various algorithms to develop the best ML model, statisticians focus on improving a single model to leverage the available data. While statisticians deal with data, they may not be familiar with the backend to serve the data and have the ability of how to access and clean it properly.","excerpt":"As the data science field has blown in popularity, it is important to note that there are other job titles with an overlap of functions. Job titles are so confusing nowadays that one company might label a designation something that is completely different somewhere, and so mainly focuses on what the responsibilities, technical skills and […]","categories":["AI Hirings"],"tags":["Data Science","Data Science Jobs"],"author_name":"Vishal Chawla","publish_date":"2020-06-22T14:00:00","publication_year":"2020","word_count":698,"keywords":["data science","Go","machine learning","AI","R","ML","Data Science Jobs","RAG","analytics","Data Science","DevOps","predictive analytics"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","RAG","predictive analytics","R","Go","DevOps"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/the-difference-between-various-data-science-job-titles\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":37145,"title":"7 Energy Startups From India That Are Harnessing The Power Of Analytics &#038; AI","content":"It has been two years since we covered a list of startups using analytics and big data in the energy space. Since then the energy sector in India has grown aggressively and there have been many startups emerging in the space that are exploring artificial intelligence, analytics and the emerging technologies to offer newer and better products. In this article, we list seven such startups from India that are defying all odds to overcome the challenges and stand tall in offering the solutions while using AI. The list is in alphabetical order: 1| Avrioenergy: This energy startup leverages data to improve energy efficiency. Founded by IIT Bombay alums, they are building a utility-focused analytics platform that would provide actionable insights to both energy utilities and end consumers to improve energy efficiency. The startup aims to leverage data to facilitate consumers with offerings such as itemised billing, customer engagement platform, personalised energy saving recommendations and more. 2| deMITasse Energies: The startup designs and manufactures a new class of engine that generates power efficiently, inexpensively and reliably, with zero emissions and no dependence on fossil fuels. It uses technologies such as aerothermodynamics, process engineering and machine learning and other cutting edge technologies as they intend to replace fossil-based solutions in the telecom, transportation and power industries with their clean engines. 3|Energly: Founded in 2015 and based in Chennai, Energly is an energy analytics startup that is helping industries, businesses and homes to reduce their power cost using a simple user interface at an attractive price point that increased productivity and profits. Essentially an IoT company it works in the areas of energy monitoring, improving energy efficiency and more while using cloud-based solutions and analytics. 4| Quenext: It is a Mumbai-based AI lab which focuses on energy, agriculture and water management offering end-to-end decision support system while adding efficiency across the value chain and offering real-time energy management. They use technologies such as artificial intelligence and machine learning to offer a fully integrated end to end decision support system for utilities. 5| SenseHawk: They boast a suite of SaaS tools that apply sensing and analytics technologies to streamline processes across the life-cycle of infrastructure assets. With tools in UAVs, IoT enabled devices, it serves in the solar energy space. It leverages aerial data combined with seamless cloud-based processing and Machine Learning based analytics to increase generation, reduce capital and O&M costs and improve safety. It provides timely and super accurate information to stakeholders while monitoring solar plant cost, locates hotspots with thermal scans and more. 6| Sustlabs: This IIT Bombay, Powai incubated startup leverages big data on energy and water. They have built technology that gets energy data and derives meaning out of it. They do so by using technologies such as automated metering and big data analytics. They have a product called SustLabs OHM that connects electricity meter to phone generating real-time insights about energy usage and house activity. It can all be done without installing an additional gadget inside the house while setting electricity limits. 7| The Solar Labs: The startup based out of Mandi uses artificial intelligence to help spread solar in the world. They are developing software to help solar firms understand how much solar can be installed and generate engineering design that maximises solar energy generation. AI is used to cut the engineering time while giving higher energy output for rooftop solar system design. Founded by IIT and BIT alumni, the startup aims to make solar energy more accessible to the world.","excerpt":"It has been two years since we covered a list of startups using analytics and big data in the energy space. Since then the energy sector in India has grown aggressively and there have been many startups emerging in the space that are exploring artificial intelligence, analytics and the emerging technologies to offer newer and […]","categories":["AI Startups"],"tags":["uav companies in india"],"author_name":"Srishti Deoras","publish_date":"2019-04-01T09:42:13","publication_year":"2019","word_count":584,"keywords":["API","machine learning","artificial intelligence","TPU","AI","ML","RAG","Aim","analytics","uav companies in india","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","TPU","R","API"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/7-energy-startups-from-india-that-are-harnessing-the-power-of-analytics-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101230,"title":"Former Google and Meta Engineers Announce AI Startup","content":"Six months ago, a software engineer named David Petrou left Google after 17 years to start his AI company. After months of silence, the former Googler has revealed some exciting updates regarding the startup named Continua. Petrou, a founding member of projects like Google Goggles and Google Glass, has a track record of leading large teams pioneering on-device machine intelligence. Now, he finds himself among a select group of individuals, including former colleagues and industry experts. The lineup of Petrou’s team includes Jonathan Betz, formerly of Meta, JT DiMartile, who held a senior staff motion designer position, as well as names like George Nachman, Jason Bacasa, Noah Lieberman (with experience at WhatsApp, Google, and AOL), Daniel Switkin (a former Google Staff Software Engineer and the visionary behind the first version of Oculus Avatar Editor), Ken Bowden (formerly of Slack), and Jason Hunter (with previous stints at Google and Microsoft). While the specifics of the startup’s financial backing remain undisclosed, Petrou confirmed that they have received investments from prominent investors and angel backers. The company has also recently established a presence in San Francisco’s Financial District. Without revealing much about the product under process, the company’s website provides a glimpse into the work in progress, describing it as “a mission to revolutionize how people interact with information, services, and each other by applying always-on, deeply integrated language models.” The team is currently looking for software engineers skilled in a wide range of areas, spanning machine learning, systems\/infrastructure, iOS, Android, and web frontend development, as indicated on their career page. In closing, Petrou shared his vision, stating, “Together, we’ll bring our vision of personal agents endowed with LLMs (large language models) to the world. It will be hard work, there are open-ended problems to solve, but it will be an adventure.”","excerpt":"After a 17 year stint at Google, David Petrou along with some of his former colleagues has announced Continua","categories":["Deep Tech"],"tags":["AI Startups"],"author_name":"Tasmia Ansari","publish_date":"2023-10-06T17:22:19","publication_year":"2023","word_count":299,"keywords":["Go","machine learning","programming_languages:R","AI","programming_languages:Go","R","AI Startups","startup"],"extracted_tech_keywords":["AI","machine learning","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/former-google-and-meta-engineers-announce-ai-startup\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10131122,"title":"Young Indians Are More Likely to Be Jobless if They&#8217;re Educated","content":"If you are an employee at an Indian IT company, you would receive an email from your manager asking to update your resume mentioning generative AI skills. It is now a mandatory drill to undergo such training, one can even call it a mania. Because back in 2017, Mark Cuban, the famous American businessman, said “Artificial Intelligence, deep learning, machine learning — whatever you’re doing if you don’t understand it — learn it. Because otherwise you’re going to be a dinosaur within 3 years.” This hasn’t changed. Almost an entire generation is studying for jobs that won’t exist. And when it comes to the Indian IT industry, if you are educated, you won’t get a job that pays you well enough. But if you are not that upskilled and do not demand higher salaries, you might be in a better position to land a job. Indian Employees Look to Reskill A Reddit user wrote, “I am 25 with a bachelor in CS and masters in AI. With zero YOE [Years of experience], I found it tough to land a job after college and had to settle with a generic SDE [Software Development Engineer] role at a small company.” Many fresh graduates shared similar concerns on forums like Reddit. Source: Reddit This sentiment reflects a broader issue faced by many new entrants into the tech industry. Aspiring software engineers\/developers consider several strategies to stand out and secure desirable positions. Indian IT companies have all recently highlighted their commitment to integrate generative AI into their operations. Top firms like TCS, Infosys, and Wipro have been leveraging technology-enabled training for their employees. Wipro’s foundational training for over 2,25,000 employees, coupled with advanced AI training for an additional 30,000, highlights their approach to AI education within their workforce. Similarly, Accenture has been expanding its data and AI workforce, reaching approximately 55,000 skilled practitioners. With a goal to double this number to 80,000 by the end of FY2026. Further, HCL plans to train and upskill around 20,000 employees in generative AI every quarter. It aims to reach 100,000 employees by the end of FY25, reflecting its strategic focus on AI proficiency. TCS is taking a similar approach, training over 150,000 employees in collaboration with tech giants. TCS aims to build the largest AI-ready workforce in the world through organic reskilling efforts. Milind Lakkad, TCS’ executive vice president and global head of human resources, noted that the company now has over 100,000 GenAI-ready employees and is investing in further deepening their expertise. Infosys has trained around 250,000 employees in GenAI capabilities, focusing on enhancing their service offerings through this technology. While Capgemini has scaled its capabilities significantly by training over 120,000 employees on generative AI tools. What’s Next? Accenture has announced over $900 million in new bookings for generative AI, reaching a total of $2 billion fiscal year-to-date during its Q3 FY24 earnings report. “We have achieved two significant milestones this quarter — with $2 billion in generative AI sales year-to-date and $500 million in revenue year-to-date — demonstrating our early lead in this critical technology,” the company said in a statement. Meanwhile, Indian IT majors like TCS have doubled up on their AI game. The IT giant announced that it is doubling its AI pipeline to $1.5 billion this quarter, from the $900 million pipeline reported in the previous one. TCS is now also working on around 270 AI projects worldwide. Moreover, in Q1 FY25, it applied for 154 patents and was granted 277, as revealed by TCS chief K Krithivasan. Coming to Wipro, it is driving innovation and enhancing productivity through investments in the Lab 45 AI platform and Wipro Enterprise GenAI Studio, incorporating various GenAI tools into the software development lifecycle. Further, HCL’s AI-led engagements include implementing a GenAI-based solution for a global technology major, automating gaming review analysis, which resulted in significant workload reduction and a 119% increase in game reviews. Additionally, they are transforming the client’s content lifecycle management with GenAI features. While Capgemini is currently engaged in over 350 new projects, with more than 2,000 deals in the pipeline. The company has also scaled its capabilities by training over 120,000 employees on generative AI tools and continues to invest in related tools, assets, and platforms. Seems like Indian IT giants have taken up the herculean task of upskilling the Indian techies with generative AI and also not keeping them unemployed.","excerpt":"Indian IT companies are anyway going to train you in generative AI.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)"],"author_name":"Vidyashree Srinivas","publish_date":"2024-08-01T14:00:00","publication_year":"2024","word_count":730,"keywords":["Go","GenAI","machine learning","artificial intelligence","AI","RAG","Aim","deep learning","generative AI","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","generative AI","GenAI","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/young-indians-are-more-likely-to-be-jobless-if-theyre-educated\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":61538,"title":"AWS Makes COVID-19 Data Lake Available For Free","content":"In a recent AWS big data blog, the company has announced that it has made a public ‘AWS COVID-19 data lake’ available for free to fight this disease. According to the blog, the ‘AWS COVID-19 data lake’ is a centralised repository of up-to-date and curated datasets on or related to the spread and characteristics of the novel coronavirus, COVID-19. The blog stated that — “As the COVID-19 pandemic continues to threaten and take lives around the world, we must work together across organisations and scientific disciplines to fight this disease. Innumerable healthcare workers, medical researchers, scientists, and public health officials are already on the front lines caring for patients, searching for therapies, educating the public, and helping to set policy.” It further stated, “At AWS, we believe that one way we can help is to provide these experts with the data and tools needed to understand better, track, plan for, and eventually contain and neutralise the virus that causes COVID-19.” AWS confirmed that the company has been working with its partners to make this crucial data, which has been hosted on AWS cloud, freely available and keep it up-to-date. The company has also seeded the curated data lake with COVID-19 case tracking data from Johns Hopkins and The New York Times, hospital bed availability from Definitive Healthcare, and over 45,000 research articles about COVID-19 and related coronaviruses from the Allen Institute for AI. “We will regularly add to this data lake as other reliable sources make their data publicly available,” said the company on the blog. The AWS’ public COVID-19 data lake has been designed to allow users to quickly run analyses on the data in place without wasting time extracting and wrangling data from all the available data sources. The users can also use AWS or third-party tools to perform trend analysis, do a keyword search, perform question\/answer analysis, build and run machine learning models, or run custom analyses to meet their specific needs. Alongside, the users can choose to work with the public data lake, combine it with their data, or subscribe to the source datasets directly through AWS Data Exchange. AWS expects local health authorities to build dashboards in order to track infections and collaborate to deploy vital resources like hospital beds and ventilators efficiently. The data lake could also be helpful for epidemiologists in order to complement their models and datasets to generate better forecasts of hotspots and trends. To access this data lake, users have to have access to an AWS account and have permissions to create an AWS CLoudFormation stack, and AWS Glue resources.","excerpt":"In a recent AWS big data blog, the company has announced that it has made a public ‘AWS COVID-19 data lake’ available for free to fight this disease. According to the blog, the ‘AWS COVID-19 data lake’ is a centralised repository of up-to-date and curated datasets on or related to the spread and characteristics of […]","categories":["AI News"],"tags":["Amazon AWS","AWS","AWS cloud","Covid Dataset","covid-19","covid19 data","free datasets for analysis"],"author_name":"Sejuti Das","publish_date":"2020-04-13T11:48:26","publication_year":"2020","word_count":430,"keywords":["big data","API","free datasets for analysis","Covid Dataset","covid-19","AWS","AI","machine learning","cloud_platforms:AWS","covid19 data","Amazon AWS","ViT","GAN","AWS cloud","R","data lake"],"extracted_tech_keywords":["AI","machine learning","AWS","R","API","big data","data lake","GAN","ViT","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-is-making-covid-19-data-lake-available-for-free\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10003433,"title":"GPT-3 Is Great. But Not Without Shortcomings","content":"OpenAI’s third generation of Generative Pre-training Transformer — GPT-3 — has been in the news a lot lately, and many experts have praised it for its intuitive capability of writing text and even code. On the other hand, others have pointed out the limitations of the GPT-3 model, including Sam Altman, the founder of OpenAI. GPT-3 is trained on massive datasets that covered the entire web and contained 500B tokens, humongous 175 Billion parameters, a more than 100x increase over GPT-2, which was considered state-of-the-art technology with 1.5 billion parameters. Despite all the developments, OpenAI’s GPT-3 is still in the experimental phase. While it has an excellent capability to generate language in all kinds of styles, there are issues that experts have pointed out. If you look at the language model, there is undoubtedly a lot of hype, which is undermining its limitations as well. Even OpenAI CEO Sam Altman tweeted by saying that “The GPT-3 hype is way too much….AI is going to change the world, but GPT-3 is just a very early glimpse.” Here we discuss some of the limitations of GPT-3 that still needs to be addressed: Lack Of Semantic Understanding According to numerous people, GPT-3 doesn’t have any understanding of the words it churns out, lacking semantic representation of the real-world. It suggests that GPT-3 is devoid of pure common sense, and, therefore, can be fooled into generating text which is incorrect or even racist, sexist and incredibly biased. GPT-3 itself, like most neural network models, is a black box where it’s impossible to see why it makes its decisions. Experts say that GPT-3 has the same architecture as GPT-2, and the only difference is the vast scale. GPT-3 suffers from similar disadvantages of not understanding real-world sensibility and coherence, like its predecessor GPT-2. Far From AGI Many AI practitioners have made an argument that the model is nothing more than one big transformer. The impressive text generation is only because of the scale and the number of resources involved in massive pre-training. According to Ayush Sharma, an AI professional, GPT-3 can be impressive; they are not even close to Artificial General Intelligence (AGI). This is because of the fact that it has no semantic understanding, no causal reasoning, and poor generalisation beyond the training set, and therefore has no “human-agent” like properties such as a Theory of Mind or Agency. He wrote, “GPT-3 has little semantic understanding, it is nowhere close to AGI, and is a glorified $10M+ auto-complete software. As is the case with all generative language models, GPT-3 assigns probabilities to strings of tokens and predicts the next likely set of words given a prompt. It remains a glorified auto-complete that has the backing of the Internet-level knowledge repository along with the magic of basic NLP.” According to a research paper, there is substantial research that language models like GPT-3 and hype around such models should not mislead people into thinking the language models are capable of understanding or meaning. Bias In Generated Text GPT-3 text generation is racially biased, and there have been many instances where people have posted how it can be highly irresponsible in terms of text generation. According to Jerome Pesenti, the head of AI at Facebook, GPT-3 is surprising and creative, but it’s also unsafe due to harmful biases. Prompted to write tweets from one word – Jews, black, women, holocaust – and GPT-3 came up with these (below). We need more work on Responsible AI before putting NLG models in production, he tweeted. #gpt3 is surprising and creative but it’s also unsafe due to harmful biases. Prompted to write tweets from one word – Jews, black, women, holocaust – it came up with these (https:\/\/t.co\/G5POcerE1h). We need more progress on #ResponsibleAI before putting NLG models in production. pic.twitter.com\/FAscgUr5Hh— Jerome Pesenti (@an_open_mind) July 18, 2020 Even OpenAI admits its API models exhibit biases in the GPT-3 paper and will be seen often in the generated text. As the model is trained on the world wide web, it is a real-time representation of views of people on the internet, and those views can be crude and even racist at times. “I don’t believe that GPT-3 is a new paradigm or an advanced technology indistinguishable from magic. GPT-3 and the OpenAI API showcases on social media don’t show potential pitfalls with the model and the API,” Max Woolf, Data Scientist at BuzzFeed wrote on his Medium blog. Max also pointed to the demo videos and said the model is low and can take time for the output to come back. The issue with the model latency can create an unsatisfactory experience for users. Given there are 175 billion parameters, the GPT-3 model is expected to be a little slow, and there are hardware challenges even with training such a large model. “I don’t blame OpenAI for the slowness. The model is way too big to fit on a GPU for deployment. No one knows how GPT-3 is actually deployed on OpenAI’s servers, and how much it can scale,” Max wrote. Problem With The ML Approach Work For Natural Language While the present state of functions in NLP is that massive neural language models, such as BERT or GPT-3, are making significant progress on a broad range of tasks, other experts may disagree.  According to them, there may also be overclaims caused by a misunderstanding of the relationship between linguistic form and meaning of words. Walid Saba, NLU Scientist and Co-founder of Ontologoik.AI wrote, “Data-Driven\/ML approaches to NLP\/NLU will not (will not ever) result in systems that truly understand natural language and the theoretical\/technical proof of this statement exists for those who listen to science.” Walid elaborated this by talking about transformers Automodel on Huggingface and said that the model demo should be taken out because it can be made to look beyond silly just in a few seconds. Research has pointed to the fact that language modelling tasks cannot lead to learning of the true meaning of words (by NLP) because they only use the form of words as training data. On the other hand, linguistic meaning pertains to the relation between a linguistic form and communicative intent. Therefore, data-driven machine learning approaches will not result in systems that genuinely understand natural language.","excerpt":"OpenAI’s third generation of Generative Pre-training Transformer — GPT-3 — has been in the news a lot lately, and many experts have praised it for its intuitive capability of writing text and even code. On the other hand, others have pointed out the limitations of the GPT-3 model, including Sam Altman, the founder of OpenAI.  […]","categories":["AI Features"],"tags":["GPT-3"],"author_name":"Vishal Chawla","publish_date":"2020-07-28T10:00:00","publication_year":"2020","word_count":1043,"keywords":["GPT-3","machine learning","TPU","OpenAI","AI","neural network","ML","Transformers","NLP","Aim","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","NLP","OpenAI","Aim","Transformers","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/gpt-3-is-great-but-not-without-shortcomings\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10172364,"title":"Animal Health Firm Zoetis Partners with Infosys to Integrate Advanced AI Solutions","content":"Infosys has announced a strategic collaboration with Zoetis, a leading animal health company. According to an official statement, the long-term engagement aims to enhance Zoetis’ IT operations and digital capabilities by integrating advanced AI solutions and automation services. This will provide greater agility to respond to the evolving business landscape and customer needs. “We look forward to this collaboration with Infosys, which will help us improve operational efficiency as we deliver value to veterinarians, livestock producers, and pet owners globally. This represents our commitment to use digital innovation to continue to lead the animal health industry,” Keith Sarbaugh, executive vice president and chief digital and technology officer at Zoetis, said in the statement. Meanwhile, Subhro Mallik, EVP and head of life sciences at Infosys, said, “This transformative journey…will reduce operational complexity and support long-term business objectives. This collaboration also underscores our commitment to delivering practical AI innovations and tailored digital solutions to the life sciences sector.” Last year, the largest producer of medicine and vaccinations for pets and livestock announced the expansion of its global capability centre (GCC) in Hyderabad, to drive the company’s innovative technology portfolio further.In a report last year, Zoetis’ India capability centre head Anil Raghav talked about the impact of the GCC, and said that the company plans to create hundreds of high-end technology roles over the next several months. “These positions will require specialised skills in areas such as artificial Intelligence, machine learning, data science, enterprise resource planning (ERP), and many more,” he had said. In the same report, Sarbaugh had said that the animal healthcare industry is evolving rapidly, driven by factors such as increasing pet ownership, growing medical needs of companion animals, and rising global demand for animal protein. “Data, analytics, and technology underpin this evolution and our abilities to connect with our customers, whether a pet owner, veterinarian, or producer. Identifying talent who has experience in technology that is critical to our future, such as SAP S4 HANA, Salesforce, Hybris, Microsoft Copilot and ChatGPT, to name a few, is a key priority,” Sarbaugh had said.","excerpt":"Infosys is expected to improve Zoetis’ operational efficiency and help the former with its long-term business objectives.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI solutions","Infosys","partnership","Zoetis"],"author_name":"C P Balasubramanyam","publish_date":"2025-06-25T13:04:23","publication_year":"2025","word_count":343,"keywords":["partnership","data science","ChatGPT","artificial intelligence","machine learning","Infosys","AI","Git","RAG","Aim","analytics","Zoetis","R","AI solutions","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","ChatGPT","Aim","RAG","R","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/animal-health-firm-zoetis-partners-with-infosys-to-integrate-advanced-ai-solutions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080917,"title":"Monster Lost it After Marketing Gimmick Backfired","content":"Last week, one of the global employment websites Monster.com rebranded itself to ‘foundit’. The development comes with an aim to add new functionalities and target the new age job market. It aims to use artificial intelligence (AI) and data and analytics to serve related roles to the users. Currently, the platform serves more than 70 million job seekers, with 10,000 corporate customers spread globally across 18 countries. The new brand name would be reflected in SEA, India as well as Middle-East businesses. The company said that the process had been in the works since Quess acquired Monster, but got delayed due to the pandemic and other reasons. According to the statistics for websites that use job board technologies, Naukri.com stands as the most popular one in the category. The data provider built with mentions that there are 384 live websites that are using Monster, along with eight websites in India. With the rebranding, Monster surely looks to regain top position in the recruitment and job hiring market. Unfortunately, the rebranding backfires While the intent for rebranding seems legit, the recent Monster India employees posting fake resignation on Linkedin and later announcing a rebranding for the platform did not resonate well with netizens, and sparked mixed emotions. Citing Marico rebranding from Bombay Oil, Former Ogilvy, Flipkart and Edelman exec Karthik Srinivasan said that Monster.com also went a similar route. He said when Bombay Oil changed its name to Marico, “this ad screamed “200 employees walk out of Bombay Oil” (agency: Trikaya Grey),” shared Srinivasan, “But the ad quickly goes on to explain that those 200 employees are now part of a new company called ‘Marico’ before anyone could take it seriously.” Srinivasan asked readers to imagine the 200 Marico employees telling their friends and families that they are no longer employed with Bombay Oil and they also do not respond to concerned, surprised, or supportive questions from friends and family. “The next day, they inform friends and family that they are now employed with Marico, what they said yesterday was just for fun. Sounds nuts?,” concurred Srinivasan, explaining that that’s exactly what Monster\/Foundit pulled off in the name of a marketing gimmick. He said that they need not have gone this way at all. “All they needed to ask employees was to change their current employer to foundit (choose Foundit India as employer) and Linkedin would automatically announce that to their networks. But, the company chose to ask employees to patently lie as a renaming gimmick,” said Srinivasan, calling Monster’s campaign deceptive and unethical, pointing at his blog post. Monster’s new avatar focuses on AI and Analytics Similar to LinkedIn, the revamped platform will add more new age offerings, with the help of existing databases for recommendations. Garissa said that foundit will use AI-based recommendations to showcase jobs to candidates—providing personalised services like mock interviews and pep materials—along with curating a list of prospective candidates for recruiters. Quess Corp founder Ajit Isaac said that the business provider acquired the platform with a vision to transform the white-collar talent. As organisations experienced the Great Resignation and the Great Regret in recent years, it has led to hiring at an unprecedented rate. Isaac believes that as the market is now settling, hiring is going to get sharper, skill-based, and focused.","excerpt":"Currently, the platform serves more than 70 million job seekers, with 10,000 corporate customers","categories":["AI Features"],"tags":["marketing analytics"],"author_name":"AIM Media House","publish_date":"2022-11-28T18:49:15","publication_year":"2022","word_count":547,"keywords":["Go","artificial intelligence","programming_languages:R","AI","marketing analytics","programming_languages:Go","Git","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","Aim","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/monster-lost-it-after-marketing-gimmick-backfired\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":31188,"title":"8 Powerful AI Chips Challenging NVIDIA’s Dominance In Computing Industry","content":"Deep learning has picked up tremendously over the last few years and is being used extensively in numerous areas — from running digital assistants to autonomous vehicles. As these machine learning and deep learning models deal with large data sets, they need powerful chips for crunching large numbers. The latest advancements are pushing AI chips to emerge victorious than before. Recent report estimate that by 2025, cloud-based AI chipsets will account for $14.6 billion in revenue and that these AI chips would be used in a variety of areas such as smartphones, smart speakers, AR\/VR headsets, and other devices which need AI processing. While the market has been largely dominated by NVIDIA, there are many other players who are building equally competent AI chips from large players to startups for carrying large computations. Here we list some of the most powerful GPUs that are revolutionising deep learning and machine learning space. 1| AWS Inferentia Latest in the AI chips, Amazon announced ‘Inferentia’ during the re:Invent conference in Las Vegas. Designed by Annapurna Labs which is an Amazon-owned Israeli company, the chip is made to deal with large workloads while requiring lower latency. While it is designed for inference, which is the process-trained ML model to find patterns in large data, it can also handle power workloads, providing thousands of teraflops per Amazon EC2 instance for multiple frameworks. Some of the popular frameworks that it is compatible are TensorFlow, Apache MXNet, and Pytorch. Data types that it supports are INT-8, mixed precision FP-16 and bfloat16. The company doesn’t aim to directly compete with NVIDIA, Intel or AMD, and would be made available to only their own cloud customers. 2| Intel’s Myriad 2 AI Chip Brought to you by Movidius, an Intel company, these chips are utilised for some of the most ambitious AI, vision and imaging applications involving both enhanced performance and low power consumption. The Myriad 2 family of processors are transforming the capabilities of devices and delivering industry-proven performance at an unbeatable price proposition. In a recent development, an ESA-led team subjected Intel’s Myriad 2 AI chip to one of the most energetic radiation beams available on earth. The test was done in CERN. The chip is run using a pair of twin LEON4 controllers – the latest in the LEON family of integrated circuits developed by ESA with Sweden’s Cobham Gaisler company. 3| IBM’s 8-Bit Analog Chip IBM was recently in news for bringing new hardware that brings power efficiency and improved training for AI projects. With an 8-bit precision for both their analogue and digital chips for AI, the chip is currently being used to test a simple neural net that identifies numerals with 100 percent efficiency. As data constantly shuttles between memory and processing, which consumes valuable energy and time, this AI chips envisions to overcome these challenges quite smoothly. IBM’s new analogue chip is based on phase-change memory. The newest solution utilises in-memory computing that promises to double the accuracy and consumes 33x less energy than a digital architecture of similar precision. It is well suited for low-power environments, making it possible to bring AI to the Internet of Things (IoT) devices and edge computing applications. 4| Huawei’s Ascend 910 and Ascend 310 Huawei recently announced two new AI chips— Ascend 910 and Ascend 310, at a recent global event for the ITC industry held at the Shanghai World Expo Exhibition and Convention Centre. With an aim to be used in data centres and internet-connected consumer devices, this is pegged as one of the most powerful chips for edge computing scenarios. The AI chips by Huawei claims to be processing more data in a faster amount of time than its competitors and help train networks in a matter of time. Huawei’s Ascend 910 is aimed at data centres and would be available in the second quarter of 2019. The Ascend 310, meanwhile, is aimed at internet-connected devices like smartphones, smartwatches and other gadgets tied to the IoT. 5| AMD GPU Radeon Instinct MI60 A recent addition in the list, AMD announced world’s first 7nm GPU named Radeon Instinct MI60 at its Next Horizon conference. With an industry-leading 1TB\/sec of memory bandwidth, the company believes that GPU will power the next generation of deep learning and AI applications in High-Performance Computing (HPC), cloud computing and graphical rendering applications. The chips have an ability for ultra-fast floating point computing performance. The company says that the GPU to GPU communication has increased considerably which is about 6X faster than before. This is enabled by the special AMD Infinity Fabric Link technology. The chips are specifically designed for high scale operations where the 7nm technology by AMD claims to dramatically improve performance per watt over previous generation products. 6| Google TPU Google introduced its homegrown AI chip, Tensor Processing Unit (TPU) in 2016 which is now in its third generation. The upgraded TPU goes deeper into artificial intelligence than the initial versions to carry heavy workloads. With the newest improvements in the chip, it will reduce the dependency of Google on chips makes such as NVIDIA on which it was dependent for GPUs to carry intensive machine learning applications. The original TPU designed was meant for the inference stage of deep learning, whereas the new version can handle training as well. The company claims that it takes a day to train a machine translations system using 32 of the best commercially available GPUs, and the same workload takes six hours atop eight connected TPUs. Google is currently operating this equipment inside its own data centres rather than selling it to other device makers. 7| PowerVR GPUs and AI chips By Imagination In a recent announcement, Imagination Technologies announced three new PowerVR graphics processing units (GPUs) that will be aimed for various categories of products such as neural networks for AI markets. With a performance range of 0.6 to 10 tera operations per second (TOPS) and multi-score scaling up beyond 160 TOPS, these chips will play a crucial role in bringing new computing capabilities in smart cars, smartphones, cameras, IoT devices and more. 8| Qualcomm AI Chips One of the front runners in the chip making for mobile phones, it unveiled two new systems-on-chip (SoCs) designed to serve smart visual applications for IoT platforms. The densely packed 10-nm FinFET-based chip can track automated equipment in Industrial IoT, carry face recognition, and more. Also, Qualcomm’s Neural Processing SDK for AI is designed to help developers run one or more neural network models trained in Caffe\/Caffe2, ONNX or Tensorflow. Along with saving time and effort, it optimises performance of trained neural networks on devices with Snapdragon. It provides tools for model conversion and execution, does a lot of the heavy lifting needed to run neural networks and more. I Startups In The Space Apart from the leading chip makers discussed above, there are many startups booming in the space. 2016 founded Cerebras System is a California-based startup that was recently funded for building chips for next-gen machine learning operations. Another UK-based AI hardware startup called Graphcore is working to lower the cost of accelerating AI applications in cloud and enterprise data centres to increase the performance of both training and inference by up to 100x compared to the fastest systems today. Coming back to India, we recently wrote about AlphaIC (Alpha Integrated Circuits) a startup which is trying to introduce revolutionary changes in the world of high-performance computing and data centres using AI. Founded in 2016, it is designing AI chips and working towards AI 2.0 through which it wishes to enable the next generation of AI with this series of products.","excerpt":"Deep learning has picked up tremendously over the last few years and is being used extensively in numerous areas — from running digital assistants to autonomous vehicles. As these machine learning and deep learning models deal with large data sets, they need powerful chips for crunching large numbers. The latest advancements are pushing AI chips […]","categories":["AI Trends"],"tags":["AI Chips","GPU","IBM","Intel","most revolutionary ai deep learning company","NVIDIA","Qualcomm"],"author_name":"Srishti Deoras","publish_date":"2018-12-07T07:38:43","publication_year":"2018","word_count":1270,"keywords":["artificial intelligence","machine learning","AI","neural network","PyTorch","ML","RAG","AI Chips","Aim","deep learning","most revolutionary ai deep learning company","IBM","Qualcomm","NVIDIA","TensorFlow","Intel","GPU"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","Aim","TensorFlow","PyTorch","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-powerful-ai-chips-challenging-nvidias-dominance-in-computing-industry\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36811,"title":"5 Ways Raspberry Pi Can Supercharge Your DIY Robotics Project","content":"Raspberry Pi is one of the biggest advancements in miniature computers, a fact that has been picked up on by the maker community across the world. The Pi has become a mainstay of many ‘smartification’ techniques to household objects, from creating mirrors that tell the time to miniature game consoles that can be taped to the back of a television. It is an extremely diverse piece of hardware and is cheap as well. Coming in at around ₹2700, the credit-card sized computer is a popular option to be added to DIY robotics projects to provide features such as Wi-Fi connectivity or onboard image processing. Here are 5 ways Raspberry Pi can be used to take DIY robotics to the next level. Rapiro Rapiro is a company that provides a programmable, do-it-yourself kit for Raspberry Pi and robotics enthusiasts. It comes with 12 servo motors for accurate and lifelike movements, and also a main board that is compatible with Arduino, another Maker favourite. It is designed from the ground up to work with Raspberry Pi and features easy assembly and a plug-and-play pre-programmed main board. Apart from the pre-programmed features, the robot can also be programmed with the Arduino IDE to do tasks such as sweeping the users’ desk. It’s easy compatibility with Raspberry Pi allows for the addition of more functions on top of the platform. This includes features such as  Wi-Fi, Bluetooth, and image recognition. GoPiGo GoPiGo is an initiative by a company known as Dexter Industries that aims to bring accessible robotics and programming education to students in classrooms. To do this, they have developed multiple kits that are both heavily resilient to classroom scuffs and scratches but can also be easily programmed. The kits also come with an easy to approach programming language known as Bloxter that enable a drag-and-drop interface, similar to a language developed by Google, known as Blockly. It is also easy to build and get started, requiring only a Wi-Fi connection to connect to the client interface on a web browser. The robot has “dozens of sensors”, which can also be attributed to the Raspberry Pi’s easy extendability due to open pins on the motherboard. GrovePi GrovePi is another initiative taken by Dexter Industries and is billed as a way to integrate plug-and-play sensors in a way to monitor, control and automate many devices in the users’ every day lives. It functions on top of the Raspberry Pi and comes with 12 different sensors that can be simply plugged in onto the board and used. These sensors include a button, a LED socket kit, a buzzer, a temperature sensor, a sound sensor, thumb joysticks, alcohol sensors and even LED bars. This opens up the possibilities of using such a device for multiple applications in a home setting, such as adding smarts to a robot that has already been fabricated. For example, a general purpose robot can quickly be converted into a safety device for the house with the addition of a temperature sensor, gas sensor, air quality sensor, flame sensor, an LED bar, and a buzzer. If values move above safe levels, the LED bar will light up and the buzzer will provide noise, thus warning residents before danger can occur. BrickPi BrickPi is a project that aims to capture the childlike amusement that everyone in the Maker community wields. The project brings together Lego and Raspberry Pi to create an easily approachable platform for those wishing to create their own robots. This is a system that works in conjunction with the Lego Mindstorms series of toys to offer greater web-service connectivity. The Lego Mindstorms series is a series of toys produced by Lego that aim for the development of programmable robots. The platform is based on the interoperable Lego building blocks that are available everywhere. The BrickPi allows users to create robots and make them smart, allowing them to connect wirelessly to the Internet and adding features such as Bluetooth and more precise movements owing to onboard processing given by the Raspberry Pi. PiKit PiKit is a project that is aimed at giving everyone a slice of the robotics pie, bringing powerful robot building to all. It brings together the modularity of Lego, the simplicity of Arduino and the power of the Raspberry Pi in one project. It features a Pi with Linux, computer vision and WiFi connectivity, enabling a higher degree of intelligence to be attributed to even a household project. The robot also has a built-in speaker and speech synthesis, which allows the robot to say whatever the user writes down to say. It also comes with a controller app that allows users to control it through the Internet, over mobile or tablet. The microphone that comes with the unit is powered by cloud computing, allowing for inference to be done in the cloud and real-time response to voice commands. This will allow for the creation of a powerful assistant robot, sparking the growth of robotics evolution.","excerpt":"Raspberry Pi is one of the biggest advancements in miniature computers, a fact that has been picked up on by the maker community across the world. The Pi has become a mainstay of many ‘smartification’ techniques to household objects, from creating mirrors that tell the time to miniature game consoles that can be taped to […]","categories":["AI Trends"],"tags":["arduino","Raspberry Pi","Robotics"],"author_name":"Anirudh VK","publish_date":"2019-03-25T11:34:07","publication_year":"2019","word_count":827,"keywords":["Go","API","AI","cloud computing","R","image recognition","computer vision","Robotics","RAG","Aim","ViT","Raspberry Pi","arduino"],"extracted_tech_keywords":["AI","computer vision","Aim","RAG","image recognition","cloud computing","R","Go","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-ways-raspberry-pi-can-supercharge-your-diy-robotics-project\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10068498,"title":"Can you really be anonymous on the internet?","content":"The New Yorker carried a Peter Steiner cartoon on its July 5, 1993 edition: The caricature of two dogs staring at each other in front of a computer became the defining meme of internet anonymity. “You can be whoever you want to be. You can completely redefine yourself if you want. You don’t have to worry about the slots other people put you in as much. They don’t look at your body and make assumptions. They don’t hear your accent and make assumptions. All they see are your words,” said sociologist Sherry Turkle about internet privacy. Those days are long gone. “Online anonymity is a complete myth. Particulars such as your postcode, your date of birth, and your gender can be traded freely and without your permission, because they’re not considered personal, but pseudonymous,” said Madhumita Murgia, European technology correspondent, Financial Times. Remember the adage, if it’s free, you are the product. Tech giants like Meta and Google monetise user data by selling them to third parties for targeted advertisement. All your movements, purchases, and interactions online are recorded. Go incognito? In incognito mode, the cookies, browser history and site data will not be saved on the device. Third-party sites cannot easily track your activity in private browsing or incognito mode. But there is a small issue. Your system’s DNS cache will still record your information. The workaround here is to clear the cache, and no one will be able to see the websites you have visited. Meanwhile, a big disadvantage of private browsing is that data is only removed from your machine. Closing the incognito window will not erase data saved on your ISP server and the servers of the websites you visit. Further, incognito browsing does not protect you against viruses or other online threats. If your system is connected to the workplace network, system administrators can see what you are doing even if you are browsing in incognito mode. VPN maybe? An anonymous proxy server or Virtual Private Network (VPN) is another option. A virtual private network extends a private network across a public network and allows users to send and receive data across shared or public networks as if their computing devices were directly linked to the private network. The encrypted connection between your computer and the internet is known as a VPN tunnel, and it is, in theory, invisible to other people online. That said, a VPN service provider always knows where your requests are coming from, where they are going, and what data you eventually send and receive. As a result, your VPN provider effectively becomes your new internet service provider (ISP), with the same level of visibility into your online activities as a traditional ISP. A tunnel within a tunnel So why can’t we use two VPNs in sequence? First, encrypt your network traffic for VPN2 to decrypt. And then encrypt it again for VPN1 to decrypt before sending it to VPN1. So VPN1 would know where your traffic originated, and VPN2 would know where it was going, but unless the two providers collaborated, they’d only know half the story. In theory, you’d be anonymous; anyone attempting to track you down would need to first obtain decrypted traffic logs from VPN2 and then username information from VPN1 to trace you. However, the catch here is; a pattern will emerge if you are consistently using the same two VPNs and leave behind a digital trail. Enter Tor! The US Navy originally designed the Tor network or onion router to hide your true location on the network while browsing, so servers can’t trace your web browsing requests back to your computer. A large pool of volunteers from all over the world serves as anonymising relays to provide a randomised, multi-tunnel VPN for Tor network users. Because your computer chooses which relays to use, there is a lot of ever-changing mix-and-match involved in bouncing your traffic through the Tor network and back. Your computer obtains the public encryption keys for each relay in the circuit it establishes and then separates the data you are sending using three onion-like layers of encryption. At each hop in the circuit, the current relay can only remove the outermost layer of encryption before passing the data to the next. Relay 1 knows who you are but not where you’re going or what you’re doing. Relay 3 is aware of your location but not your identity. And Relay 2 acts as a buffer making it much more difficult for Relays 1 and 3 to collude. Though Tor’s exit nodes can’t locate you due to entry\/guard and middle relay (which changes frequently), it can see your final, decrypted traffic and its ultimate destination because the exit node removes Tor’s final layer of encryption. If you use Tor to access a non-HTTPS (unencrypted) web page, the Tor exit node that handles your traffic can snoop on and modify your outgoing web requests and any responses that come back. And, with limited exit nodes available on average, an attacker who wants to control a significant portion of the exits doesn’t need thousands or tens of thousands of servers – a few hundred will suffice. Then, there are cyber-attacks. According to a 2003 study, a cyber attack happened every 39 seconds. Ten years down the line, nearly 30,000 new websites got hacked every day. By 2021, 4,145 publicly disclosed breaches exposed over 22 billion records. And according to 2021 IBM data, 44 percent of data breaches had personally identifiable information.","excerpt":"The US Navy originally designed the Tor network to hide the location of the browser on the network.","categories":["IT Services"],"tags":["IBM","VPN"],"author_name":"Sri Krishna","publish_date":"2022-06-07T13:00:00","publication_year":"2022","word_count":915,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","ViT","IBM","R","VPN"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-you-really-be-anonymous-on-the-internet\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018128,"title":"15 Top AI Tools For Resume Screening and Recruitment","content":"Best AI Tools for Resume Screening Artificial intelligence has made huge inroads into the human resource domain lately. Companies are now deploying AI tools to reduce or eliminate time-consuming tasks in the recruitment process. According to experts, screening resumes is the most critical, albeit, the most labour-intensive and challenging task for HR personnel. Typically, an HR recruiter of a medium and large corporation gets thousands of job applications. The applications will be all over the map and require hours and hours of screening. And, that’s where AI tools come into play. With the right AI tools, recruiters can sort the applications in a fraction of time. Here we will list the top eight tools that use artificial intelligence to screen resumes. Also Read: How AI-Powered Tools Can Shape Your Resume 1. Freshservice Freshservice, developed by Freshworks, is a comprehensive HR solution that excels in resume screening. Its intuitive user interface and robust integration capabilities make it an ideal choice for quickly assessing resumes. Freshservice acts as a centralized hub for collecting resumes, adding notes, and rating applicants, streamlining the hiring process and enhancing recruiter productivity. Check Freshwork tool 2. SmartRecruiters SmartRecruiters is a versatile platform that integrates AI-driven resume screening with a comprehensive applicant tracking system. It enhances candidate-job matching by analyzing resumes against job requirements, ensuring only the most suitable candidates are shortlisted. This tool is particularly beneficial for organizations looking to streamline their recruitment process and improve hiring outcomes. Check SmartRecruiters tool 3. Manatal Manatal is renowned for its user-friendly interface and advanced AI capabilities that simplify recruitment. It offers features like social media enrichment and candidate recommendation, which help recruiters identify the best candidates efficiently. Manatal’s AI-driven approach ensures that resumes are assessed with precision, reducing the time spent on manual screening. Check Manatal tool 4. SkillPool SkillPool leverages AI to assess resumes by focusing on skills and competencies relevant to job roles. It provides detailed insights into candidates’ abilities, helping recruiters make informed decisions. SkillPool’s emphasis on skills-based assessment makes it a valuable tool for organizations prioritizing competency over traditional qualifications. Check skillpool tech 5. HireBeat HireBeat offers a comprehensive suite of features, including an inbuilt applicant tracking system and customizable screening questions. Its AI capabilities enable efficient resume screening and candidate ranking, making it easier for recruiters to identify top talent. HireBeat’s integration with multiple platforms enhances its versatility in various recruitment scenarios Check hirebeat tool 6. Textkernel Textkernel specializes in semantic search and matching, providing a sophisticated approach to resume screening. Its AI-driven technology ensures that resumes are matched with job descriptions accurately, improving the quality of shortlisted candidates. Textkernel’s focus on semantic analysis sets it apart from traditional keyword-based tools Check Textkernel tool 7. Daxtra Daxtra offers advanced resume parsing and matching capabilities, which are highly beneficial for large-scale recruitment drives. Its AI algorithms efficiently sort and rank resumes, reducing manual effort and improving the speed of the hiring process. Daxtra’s integration with existing HR systems further enhances its utility Check Daxtra tool 8. Sniper AI Sniper AI uses machine learning to screen candidates and match their resumes with the job specifications at a blinding speed and stunning accuracy. Backed by Recruitment SMART, a UK-based HR tech startup, Sniper AI can be easily integrated with applicant tracking systems for easy resume screening. Sniper AI comes with 53% internal workforce reduction capability that allows recruiters to spend less time screening the resumes. With clients like Infosys, Vodafone, Capgemini, etc., this tool is quite renowned among the industry and claims to be a game-changer for AI-based recruitment. Check out Sniper AI tool. 9. Ideal Ideal is an AI tool that leverages recruiters’ feedback to build an effective algorithm to shortlist thousands of candidates based on their resumes. Claiming to reduce recruiters’ time by 70%, the AI accurately shortlists candidates for the next round of interview with HR personnel. The AI helps in resume screening for a particular job opening and improves the representation of diverse applicants with 100% bias-free algorithm. It considers external evaluation like assessments, chatbot conversations, and other data points for identifying the right candidate. Check out the Ideal recruitment tool 10. CVViZ CVViZ uses its in-house AI algorithm to understand the resume contextually, beyond the keyword or binary search approach. The AI-powered resume screening software reduces manual efforts significantly and ranks candidates in real-time. The model can be customised based on the organisation’s requirement and the type of employees it has worked with before. When an organisation adds a new job opening, the AI automatically finds top talent that may already exist in your recruitment database. The AI makes the recruiting process simpler, intuitive, and efficient by improving the quality of hire. Check out CVViZ tool 11. Skeeled Skeeled is an automated applicant tracking system that harnesses AI to screen resumes and assess candidates’ qualifications. It has deployed powerful ranking algorithms to deliver a shortlist based on qualification indicators. Along with their qualification indicator, the system also checks the add-ons like specific driver licence type, work permit, or any other custom criteria that organisations demand. This, in turn, automates and streamlines the first steps of the recruitment process, i.e. resume screening. Additionally, it comes with an advanced search engine and filtering tool that finds the right candidates from organisations’ resume databases. Check out Skeeled. 12. Hubert Hubert is a holistic AI recruiting platform that assists HR personnel in doing the hiring process. Trained on millions of reliable data points, the AI has been designed to provide transparent advice to the HR team. Hubert combines data from uploaded documents, personality, logical tests and the desired requirements for the job to pick the right candidate. Its transparent in-house-built algorithm also reduces the impact of unconscious bias to make better hiring decisions. Along with resume parsing, it also comes with features like diversifying the workforce, team compatibility, initial job interview, etc. Check out Hubert Also Read: How This Startup Is Revolutionising Tech Hiring By Using NLP & Azure 13. Mosaictrack A smart recruiting solution, Mosaictrack utilises the cognitive power of artificial intelligence to scan through resumes and social profiles to pick the best talent based on culture fit and skill set. Its predictive analytics capability omits the task of conducting surveys or making questionnaires, where advanced algorithms pre-qualify the applicants to enhance the interview process. The AI-tool has been built with IBM’s Watson technology allowing to match talent with extreme accuracy through the power of machine learning and natural language processing. Check out Mosaictrack. 14. Vervoe Vervoe is an AI-based skills assessment tool that uses machine learning algorithms, to screen resumes at scale, allowing recruiters to spend more time with high performing candidates. Its predictive analytics then automatically grades and ranks them based on how well they can do the job. The algorithmic models measure the quality of a candidate’s answer against millions of similar responses. It also processes thousands of responses quickly to look for specific words or sentiments that accurately reflect these values. It then provides a list of candidates ranked in order of their strength and potential businesses need. Check out Vervoe. 15. XOR Ranked as a leading artificial intelligence platform for recruiting teams, XOR automates resume screening, interview scheduling, onboarding and more. Its advanced capabilities also allow the platform to manage conversations with candidates through one-on-one messaging that saves even more time for the recruiters. The XOR model is hosted on Microsoft’s Azure that enables automatic mass resume screening for a suitable candidate on more than 100 different languages and algorithms and has been trained on various HR and recruitment data sets. Check out XOR.","excerpt":"Best AI Tools for Resume Screening Artificial intelligence has made huge inroads into the human resource domain lately. Companies are now deploying AI tools to reduce or eliminate time-consuming tasks in the recruitment process. According to experts, screening resumes is the most critical, albeit, the most labour-intensive and challenging task for HR personnel. Typically, an […]","categories":["AI Trends"],"tags":["ai in hr","Top Trend"],"author_name":"Sejuti Das","publish_date":"2024-08-16T00:19:42","publication_year":"2024","word_count":1264,"keywords":["Top Trend","semantic search","artificial intelligence","machine learning","AI","ML","ai in hr","RAG","NLP","Aim","analytics","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","analytics","Aim","RAG","predictive analytics","semantic search"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ai-tools-for-resume-screening\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":8718,"title":"10 movies that show the power of Analytics","content":"Everyone loves a good plot. There are movies that have fascinated us with a deep focus on various aspects of data science, like mathematics, statistics, artificial intelligence, machine learning etc. Analytics India Magazine decided to compile a list of ten films that have left a mark on the global film industry and given a sneak peak into the world of analytics and data science. However, this is quite a subjective choice and have listed them in alphabetical order. Also, we know that only ten are not enough, so we would encourage you to share your views and mention the movies that you think we have missed out and why in the comments section below. 21 Year of release: 2008 This movie gives insights on Newton’s method and Fibonacci numbers through the eyes of six students and their professor. A math professor, Micky, decides to cash in at the blackjack table with six brilliant students from MIT using numbers, codes and hand gestures. https:\/\/www.youtube.com\/watch?v=Zr_xWfThjJ0 A Beautiful Mind Year of release: 2001 This movie walks us through the life of the brilliant mathematician John Nash who won the Nobel Prize in economics for his work in game theory. He combats schizophrenia, while working for the government breaking Soviet codes and ends up in the midst of a conspiracy. His struggle and the journey of self-discovery coupled with his love for numbers made this movie a must watch! Here’s a clip of Nash Equilibrium explained in an interesting manner. ExMachina Year of release: 2015 The story presents the pros, cons and complexities of Artificial Intelligence in a fascinating plot. It explores the possibilities of robotics and data coupled with emotions and humanity. Also, the viewer gets to know interesting concepts around artificial intelligence like The Turing Test. Caleb, a programmer wins an opportunity to spend a week at a mountain estate of his company owner, Nathan who explains that Caleb has to evaluate the reactions and emotions of a female robot – Ava. As they get to know each other more, Caleb gets involved in a plot to scheme her flee. Meanwhile, Nathan tells Caleb that he has been manipulated by Ava. Ava’s emotional intelligence proves more sophisticated and deceptive, beyond the imagination of the two men. Watch the movie to find out who lied. Good Will Hunting Year of Release: 1997 Good will Hunting is about a stubborn, mathematical genius, who knows he’s smart but pulls back from challenges due to his shady past. There are these four people, who want to support him, but he feels threatened by their help because it means abandoning his careless and laid back attitude. In the scene below, Cayley’s formula is showcased stating the formula for number of labeled trees on n nodes. Then it lists 8 different unlabeled trees with 10 nodes. Minority Report Year of Release: 2002 This movie deals with predictive analytics. Set in year 2054, Washington D.C is crime-free as a future technology makes it possible for cops to know even before a crime is committed. Predictive crime fighting at its best. It works through 3 people known as precogs who are psychics and can see the future crimes in the way it occurs. Through advanced technology, they are able to see and analyze the data and identify the criminal before the crime is committed. However, when John Anderton, the founder of this system is accused of one such crime, he sets out to prove his innocence. https:\/\/www.youtube.com\/watch?v=7SFeCgoep1c Moneyball Year of Release: 2011 A discussion on Analytics in movies is considered unrendered without Moneyball. It is probably the most definitive movie with use of analytics. By betting against tradition and in favour of numerical analysis, General Manager of Oakland Athletics, joins forces with Peter Brand, a Yale graduate, to challenge old-school selection methods and reinvent his team using predictive modelling techniques. Pi Year of Release: 1998 The film Pi is all about maths and how we can use it for predictions. It personifies the power of mathematics, computers and analytics through it’s characters. The mathematical genius, Max believes that mathematics is the language of universe and from nature to stock market, he can predict anything if he can find patterns and the key to the chaos. Looks like a data scientist’s delight. The Bank Year of release: 2001 The film follows the story of a maths prodigy who takes revenge on a bank. Jim Doyle, a mathematician devises a formula to predict the fluctuations in the stock market. The story is an intricate game of stock speculation, personal vendettas and computer hacking. It gives a detailed insight into the evil, the good and grey characters in and around “the bank.” The Imitation Game Year of release: 2014 This film traces the life of the cryptanalyst Alan turing during the dark times of World War II. It gives a sneak peek into the world of coding in its nascent stages. Turing joins a team of codebreakers at a secret facility to try to decipher the code of the German enigma machine while fighting his internal battles. https:\/\/www.youtube.com\/watch?v=_C25CwNlVjA X+Y Year of release: 2014 A socially awkward teenage math prodigy finds new confidence and new friendships when he lands a spot on the British squad at the International Mathematics Olympiad.","excerpt":"Everyone loves a good plot. There are movies that have fascinated us with a deep focus on various aspects of data science, like mathematics, statistics, artificial intelligence, machine learning etc. Analytics India Magazine decided to compile a list of ten films that have left a mark on the global film industry and given a sneak […]","categories":["AI Features"],"tags":[],"author_name":"Дарья","publish_date":"2016-01-26T06:37:40","publication_year":"2016","word_count":876,"keywords":["data science","Go","machine learning","artificial intelligence","AI","R","RAG","analytics","CLIP","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","RAG","predictive analytics","R","Go","CLIP"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/10-movies-that-show-the-power-of-analytics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":600,"title":"Controlling shipping traffic in the Netherlands canals with a wireless sensor network","content":"For municipalities, provinces and other potential canal authorities such us Ministry of Infraestructure, it is essential to, it is essential to provide a system able to control the amount of boats that sails in an area an also automate the opening and closing of the bridges. Vicrea is a dutch company specialized in Smart Integrated Information and Geographic Systems and Smart City solutions. The company has developed a wireless sensor network with Libelium technology to manage canals traffic in the Netherlands by controlling the flow of boats. Innovative laser solution to detect ships This solution arises with the objective to know the amount of ships that sails in the canals and also their direction to plan the water infrastructure. It has been already deployed in some of the most touristic and well-known cities from the Netherlands: Delft, The Hague, Leiden, Alphen aan den Rijn and Gouda. There is a wide range of boats that sails on the canals: cargoboats, commercial passenger ships and recreational ones. Although it was known by public authorities and citizens that in specific times there was heavy traffic, there was not any knowledge and neither control to solve this problem. “Libelium helped us in an innovative and professional way to help our project to become a success”, affirms Erkan Efek, Business consultant & Architect GIS of Vicrea. Vicrea and Libelium have worked in developing a new and innovative laser solution that is able to monitor direction, distance and speed. This sensor allows to detect any ship that is crossing by a concrete point and also to know the direction towards the ship is sailing. The shipping traffic data monitored by the sensors added to Waspmote Sensor Platforms is sent out through Zigbee to Meshlium. 868 Mhz wireless technology is used to connect the gateway with Vicrea Cloud Platform. From there, data can be further distributed to a live monitor or can be stored for later analysis and prediction. Traffic management thanks to sensors The solution developed between Libelium and Vicrea will help public institutions to manage the water infrastructure and also control other issues. With this deployment, there will be possible to know if the maintenance level of canals is at par with traffic density in a canal. Bridge opening and closing will be automated to improve cars and ships experiences and also to reduce waiting times. The platform will control if a boat is entering in a private or prohibit area. This monitoring system will be working 24 hours a day, 7 days a week, to give public authorities a holistic view about what is happening in their water infrastructures. Erkan Efek, Business consultant & Architect GIS of Vicrea, has highlighted that “the solution incorporates cost-effective elements and eco-friendly parts. We use solar panels for our power supply and send data wirelessly”. Sensors can be placed anywhere so each government can choose depending on the needs and the urban planning of each city. Knowing in real-time information about traffic congestion or predicting maintenance for bridges are just some of the improvements that councils can apply in their strategies for water infrastructures management. The deployment has reached the goals that the companies settled since the beginning. Data related with passing ships and also frequency will be monitored with high accuracy, costs will be drastically reduced, about 80.000 euros each year per bridge, and the system will be working with no interruption.","excerpt":"In the Netherlands there are many waterways, small inland harbors and canals that are used by citizens in their daily routines. Each city has a water infrastructure with high-density shipping traffic.","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2016-10-04T07:45:47","publication_year":"2016","word_count":565,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/controlling-shipping-traffic-netherlands-canals-wireless-sensor-network\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24429,"title":"FGPA Vs GPU: Autonomous Car Makers In Silicon Valley Have Definitely Chosen A Side When It Comes To AI Chips","content":"We are back to the battle of graphics processing units vs field-programmable gate array (FPGA) with NVIDIA in the centre of attention. This time, there’s more at stake than just FPGA power efficiency factor as compared to GPUs cost efficient advantage. The recent tech boom with cloud computing, image processing, robotics, deep learning and big data workloads have necessitated a shift towards GPUs. But that doesn’t rule out FPGAs which are designed to perform fixed-point operations with a close-to-hardware programming approach. GPUs on the other hand, are optimised for parallel processing of floating-point operations using thousands of small cores, according to this whitepaper. FPGAs also score over GPUs in terms of interface flexibility, improved by the integration of programmable logic with CPUs and standard peripherals. They also provides huge processing capabilities with a great power efficiency. But when it comes to cost efficiency, GPUs are winning the battle in development and software acceleration while FPGAs require a team of design engineers. Another thing to be considered is that, today many algorithms are designed directly for GPUs, and FPGA developers are difficult and expensive to hire. Is FPGA Suitable For Self Driving Technology? This is not the age-old comparison between GPU vs FPGA which is not the crux of the matter. FPGAs which can perform a range of logical functions simultaneously are being considered unsuitable for emerging technologies such as self-driving cars or deep learning applications. At the recently-concluded GPU Technology Conference in March, NVIDIA top brass Jensen Huang declared that FPGA is not the right answer for autonomous tech development. According to Huang, FPGA was more suited for prototyping and wouldn’t lead to desirable results. With leading tech giants like Intel, Qualcomm, Nvidia and Google plowing in millions of dollars in self-driving technology and AI chips, it is evident that chipmakers would want to design a chip for autonomous driving technology. Today, thanks to its immense popularity and cost-efficiency, GPU has emerged as the dominant chip architecture for autonomous technology and Nvidia has intensified the battle by increasing the computing speed 10 times and reducing the power consumption 16 times. They have further strengthened their position in the automotive market with partnerships with Audi, ZF and Zenrin and are building an AI-driven big data system. Big Players Dominate The Self Driving Market, But Smaller Players Are Emerging The chip market is dominated by players like Intel, AMD, Google, Altera and Cambricon among others, but Nvidia is the market leader. Besides GPU, CPU and FPGA, ASIC is another mainstream chip that is fast becoming the industry standard for self-driving technology. One of the key advantages of ASIC chips is that they are best suited for computing and power consumption. In the near future, industry analysts believe research and development of all AI chips will gain traction as autonomous technology and certain use cases would drive specialised AI chips. Until autonomous cars reach the stage of mass production, various chip architectures are possible. For example, Chinese startup Horizon Robotics, leader in embedded AI released an AI processor named Journey 1.0 for smart driving, and another one called Sunrise 1.0 for smart cameras. Now, in an automotive setting the two chips work together and the Chinese startup which got its latest funding from Intel Capital claims Journey 1.0 has a detection accuracy of more than 99 percent for vehicles, pedestrians, lane lines and traffic signs. Qualcomm is not a small player but the chipmaker doesn’t want to be left behind in driving autonomous technology forward. The leading chipmaker of choice for smartphones, the American semiconductor player launched Drive Data platform, dubbed a variant of Snapdragon processor in 2016. As part of its automotive transition, the Drive Data platform builds on Qualcomm’s Snapdragon 820 Automotive processor and delivers the next level of intelligence and unprecedented mobile connectivity for autonomous driving. And in order to dominate the market like Nvidia and Intel, Qualcomm is providing a full platform that brings together all the key capabilities such as HD Mapping and machine intelligence for safe navigation.","excerpt":"We are back to the battle of graphics processing units vs field-programmable gate array (FPGA) with NVIDIA in the centre of attention. This time, there’s more at stake than just FPGA power efficiency factor as compared to GPUs cost efficient advantage. The recent tech boom with cloud computing, image processing, robotics, deep learning and big […]","categories":["Deep Tech"],"tags":["Intel","Intel Chips"],"author_name":"Richa Bhatia","publish_date":"2018-05-09T05:11:59","publication_year":"2018","word_count":670,"keywords":["big data","Go","API","AI","cloud computing","Intel Chips","RAG","Ray","Aim","deep learning","R","Intel"],"extracted_tech_keywords":["AI","deep learning","Aim","Ray","RAG","cloud computing","R","Go","API","big data"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/fpga-gpu-autonomous-car-makers-silicon-valley-chosen-side-when-it-comes-to-ai-chips\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10018157,"title":"Need For A Balanced Perspective On Big Tech As A Consumer","content":"In these polarized times, very few ideas find universal approval. Individual freedom is one of them. All of us agree that an individual has a right to freedom i.e., no one ought to dictate how we live and no one should be able to control our lives. So we are instinctively apprehensive of anybody who has the power to do that. This results in fear and some cases hatred, of authority, which manifests itself in different forms: Teenagers wanting to break free from societal norms, Citizens wanting minimal interference of the Government in their personal lives, people’s fear of big corporations. Big corporations and Billionaires, especially, have been at the receiving end of an unusual amount of vitriol lately. Apocalyptic visions We often think of Big corporations or Big tech as ominous bodies who are exploiting us for profits. In our imagination, Big tech has no face, it’s just a dark cloud hanging over our lives and our future. This dark cloud is up to no good. It knows our secrets and controls our lives by telling us what we ought to think, eat, do, watch, etc. One day this cloud will swallow us. Big tech will cause the future to be like our favorite apocalyptic movie or book. Think Terminator movies and George Orwell’s 1984. Or will it? Can it not reign in an era of equality, knowledge, justice? It is because of technology, and big tech companies specifically, that today we have all the knowledge imaginable available at our fingertips. The world is more connected than ever and there’s the availability of opportunities for all regardless of caste, gender, age, nationality. Is it not possible for us to imagine a future where it can be a force for good? I am not prophesying, just asking you to consider the other possibility as well and look at the matter with a balanced perspective. By denying the Big tech the credit it deserves and distrusting it based on apprehensions and not facts, we could be denying ourselves the enormous possibilities it presents. Big Tech an actual cause of concern? Not just individuals but Governments world over as well have started questioning the influence of Big tech. The US is considering antitrust proceedings against the big four of the tech world: Amazon, Google, Apple, and Facebook. These companies are also going to face regulation issues with the EU. The core bones of contention are their alleged anti-competitive nature and privacy concerns. Politically as well, social media companies like Facebook (plus its other networks like Instagram and WhatsApp) and Twitter are in muddied waters. Recently both banned Donald Trump from using their platforms. These social networking sites have at different points of time been accused of being biased toward the other side by both Left-wing and Right-wing and have often banned controversial people leading commentators to believe they are hampering freedom of speech. However, it cannot be denied that Social networking sites along with other big tech companies have been hugely beneficial to people’s lives in myriad ways. So much so that services like Google, WhatsApp, YouTube, Facebook, and countless others have become an essential part of our daily lives and have helped many survive and get through the lockdown. It does not behove us to use them every single day and then curses them and boycott them for matters we have scant understanding of, just because everybody else is doing so and we do not want to invest the time into understanding the nuances or even the basic details of the matter. WhatsApp privacy policy hullabaloo Whenever we come across a news story or op-ed that is negative of a service we are using and that also happens to be offered by a Big tech company, often the knee jerk reaction is the desire to quit it. This could be either because we want to boycott it or there’s a vaguely disconcerting development. This recently happened with WhatsApp. Facebook changed WhatsApp’s privacy policy and asked the users to accept it, or else they won’t be able to use it anymore. This caused a big furore amongst netizens and commentators and everybody started discouraging the use of WhatsApp all of a sudden. Users who saw this got scared and started considering quitting WhatsApp over privacy concerns. So now there’s this whole ‘Quit WhatsApp, Join Signal’ situation and there are users blindly joining the bandwagon. Do these users actually know what has changed? Had there not been a fuss regarding this update, how many users would have quit? Does it make sense to blindly follow what others are doing and not consider what’s best for you? WhatsApp still offers end to end encryption. It cannot hear your calls, read your messages, see your call logs among other things. What has changed however is that Facebook will be able to access information like Battery level, app version, signal strength, etc through WhatsApp. But do users really know if they are already sharing this information with WhatsApp or not? The fact is most users don’t even know what information they are sharing with what services and largely do not care. Most users wouldn’t bat an eyelid accepting the new terms and conditions if there wasn’t such furore and if Facebook had not put the clause that one has to either accept the news terms and conditions or quit. By this, I do not mean to say that new terms and conditions are not problematic. That users can themselves decide. However, we must act as informed consumers and weigh the pros and cons of availing the service whenever such a situation arises. Consider the disadvantages of wholly getting off WhatsApp. It is the predominant messaging service in much of the world and almost everyone you know is on it. I am not even getting into ease of user experience and features. Point is to think and make our own decisions. Big tech, like any other thing that holds excessive influence over our lives, will be targeted repeatedly. However, as consumers, we must decide what’s best for us and act accordingly. We must not let others’ opinions dictate our lives.","excerpt":"In these polarized times, very few ideas find universal approval. Individual freedom is one of them. All of us agree that an individual has a right to freedom i.e., no one ought to dictate how we live and no one should be able to control our lives.  So we are instinctively apprehensive of anybody who […]","categories":["AI Features"],"tags":[],"author_name":"Manish Bhanushali","publish_date":"2021-01-16T22:00:00","publication_year":"2021","word_count":1023,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","ViT","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","ViT","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/need-for-a-balanced-perspective-on-big-tech-as-a-consumer\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077804,"title":"GitHub Copilot: The Latest in the List of AI Generative Models Facing Copyright Allegations","content":"GitHub Copilot, the text-to-code AI tool, has been—for the most part—revolutionary in determining how people code. Twitter has been erupting with people expressing how this new AI tool has benefitted them with organisation heads and developers alike hailing it for saving much of their time. However, the latest discussion surrounding it suggests that things are murky. Tim Davis, Professor – Computer Science, Texas A&M University, took to Twitter to express his resentment over Copilot producing his copyright code for a particular prompt. @github copilot, with \"public code\" blocked, emits large chunks of my copyrighted code, with no attribution, no LGPL license. For example, the simple prompt \"sparse matrix transpose, cs_\" produces my cs_transpose in CSparse. My code on left, github on right. Not OK. pic.twitter.com\/sqpOThi8nf— Tim Davis (@DocSparse) October 16, 2022 Chris Rackauckas, lead developer of SciML, also shared a thread of Armin Ronacher from July 2021, adding, “Github Copilot spits out the Quake source code. It just repeats its training data often, even without OSS licenses”. I feel sorry for you given how many really bad takes you're getting here. For some more ammo, here's a thread that shows that Github Copilot spits out the Quake source code. It just repeats its training data often, even without OSS licenses. Oops.https:\/\/t.co\/YDN3nBoMYJ— Dr. Chris Rackauckas (@ChrisRackauckas) October 16, 2022 But beyond this, the latest news that has been making rounds is about Matthew Butterick, a writer, programmer, and lawyer, who announced on October 17 that he would be teaming up with Joseph Saveri Law Firm, investigating a potential lawsuit against GitHub Copilot on the grounds of violating open-source licences. In writing on the issue of copyright violation in June 2022, Butterick cautioned organisations creating software products against the use of Copilot, as they would be taking part in using someone else’s intellectual property, albeit unintentionally. GitHub is trained upon billions of lines of public code. But, there is no surety over whether the training data comes as fair use under copyright law. In presenting the case, Butterick writes that Microsoft characterises Copilot’s output code as only a series of “suggestions” and does not claim any rights over it. Additionally, he also cites a passage from GitHub’s website showing how Microsoft plays safe by pushing the blame onto the end user: “You are respon­si­ble for ensur­ing the secu­rity and qual­ity of your code. We rec­om­mend you take the same pre­cau­tions when using code gen­er­ated by GitHub Copi­lot that you would when using any code you didn’t write your­self. These pre­cau­tions include rig­or­ous test­ing, IP [(= intel­lec­tual prop­erty)] scan­ning, and track­ing for secu­rity vul­ner­a­bil­i­ties.” In a recent statement, Open AI claimed that the training material from public repositories is not meant to be included in the output generated by Copilot. Additionally, their analysis has shown that a vast majority of the output (>90%) doesn’t match the training data. There is a divided opinion (a grey area, if you will) about who “legally” stands right among the two parties. GitHub has made it clear that the users need to check if the code used is free of copyright infringement, but at the same time, the open-source communities see the whole facade of “AI training is fair use” for their copyrighted codes to be a disregard for their rights. See, for example, this statement by Butterick: “By claim­ing that AI train­ing is fair use, Microsoft is con­struct­ing a jus­ti­fi­ca­tion for train­ing on pub­lic code any­where on the inter­net, not just GitHub.” Hence, there is little clarity over who is to be held accountable for this—Is it Copilot or the end users employing the AI-generated code for their product? GitHub’s claim that AI training comes under fair use needs more inspection. This is not the first time questions of copyright have sprung forth in AI applications. It has been a persistent issue throughout the recent surge in AI generative models. In an interview with Ben Sobel by IPW in 2017, Sobel explains the problem as a “fair use dilemma”. His argument goes like this: (i) If Machine Learning doesn’t come under fair use, then organisations have to pay remedies to millions who form the training data on which machines learn. This will hinder any progress in the field. (ii) But, if it does come under fair use, it is likely that organisations will take liberty in using the intellectual labour of people for their own profit. Therefore, it will not be a stretch to say that the legal aspect of AI use is in difficult terrain. If there is a case for Butterick to take the makers of Copilot to court, the outcome of the lawsuit will have a huge impact on the future of open-source communities and AI generation models.","excerpt":"GitHub Copilot, the text-to-code AI tool, has been facing accusations of stealing people’s codes. So, what’s next for AI-generative models?","categories":["Global Tech"],"tags":["AI laws","GitHub","Github Copilot","Microsoft"],"author_name":"Ayush Jain","publish_date":"2022-10-23T10:00:00","publication_year":"2022","word_count":785,"keywords":["Go","machine learning","TPU","AWS","AI","ML","Github Copilot","Git","AI laws","Aim","GitHub","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","AWS","TPU","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/github-copilot-the-latest-in-the-list-of-ai-generative-models-facing-copyright-allegations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":22933,"title":"How Big Data Is Used To Address Water Crisis In India","content":"Big data has become a crucial part of every business – from Fortune 500 enterprises to startups. As it is changing the way we do business, many believe that big data has the power to solve world’s biggest issues. One such issue is the addressing water crisis in India. A recent UNESCO report predicts that Central India is staring at deepening water scarcity, meaning withdrawal of 40 percent of the renewable surface water resource. Bangalore could be the first victim to face the acute water crisis. An alarm bell rang loud when a CSE-assisted environment magazine report stated that the city is fast heading towards ‘Day Zero’ and could go the Cape Town way. The report says the water table in Bengaluru has shrunk from 10-12 m to 76-91 m in just two decades. There are multiple factors that are playing a role in the water shortage in India. But the major ones are regional bias in the availability of surface water.  Inter-state water wars between Karnataka and Tamil Nadu, even Punjab and Haryana continues. More than half of India’s total water supply comes from groundwater resources, which have been severely depleted because of excessive use. There is evidence to show that vast amount of water is wasted. Utilities around the world lose about 35 percent of water due to leakages and bursts. To combat water crisis in India the need of the hour is to invest in water distribution system by upgrading the old infrastructure as well as building a new one to conserve the resources. And there are few companies that are exploiting the power of big data to bring significant change. These companies are aggregating data across the entire water cycle to avoid any water wastage and ensure a sustainable use of water. With the help of sensors and monitoring system, they are generating a large amount of real-time data. TetherBox Technologies’ Elemento Aqua The lakes in the country have become a toxic ecosystem with a high level of industrial chemicals and pollutants, which results in foaming, sporadic eruption of fire, deaths of fishes. In such dismal condition, TetherBox Technologies is serving as a ray of hope to cater this situation. The company’s Elemento Aqua provides real-time analysis of the water body by monitoring the oxygen level and highlighting the percentage of contamination level in the lake. After detecting the reduction of oxygen levels they pump in the required oxygen into the lake to keep it hygienic. The dismal condition of the lakes in the city has given birth to this startup. The startup came into limelight after Ulsoor fish kill incident in 2016. Under the supervision of Karnataka State Pollution Control Board authorities, TetherBox in collaboration with EcoGen Tech conducted a pilot project to gather water quality reading through big data and reduced the algae cover in the lake. With the help of cloud-based tools such as Hadoop, the smart device collects data regarding timely fluctuation of oxygen levels and the water levels in the water. It then sends an alert about the status of the lakes to the factory clients. TetherBox will also install its devices in city’s most polluted Bellandur lake and Vrishabhavathi Valley. SmarterHomes Technologies There are many ways that people wastes water, from accidental leakages to utter negligence. Most families waste roughly 40 percent of water. This Bengaluru-based IoT company feels that with the help of their smart meters 35 percent of water wastage can be controlled. The smart water meter ‘WaterOn’ generates real time data that helps the consumer to shut off the water supply in an event of a leakage. The smart meter helps in preventing wastage of water by detecting leaks and shuts the water supply remotely. The SmartHomes mobile app monitors the data from the water meter. The smart meters can store approximately 45 days of consumption records in its local storage. The meter counts each consumption and sends the data to our cloud server through Nuclious. The Nuclious is a hybrid variant of wired meters and wireless communication hub. The company installed near 7000 smart meters in India and plans to expand in the middle east, claims a report. IBM’s Big Data Solution IBM’s analytics-based solutions provide smarter water management and better control over the resources for water boards thus controlling the wastage of water. According to the company, “The Intelligent Operations Center (IOC) offers integrated data visualization, near real-time collaboration and deep analytics to help enhance the ongoing efficiency that will improve the efficiency of services for its residents.” IBM in collaboration with Bangalore Water Supply and Sewerage Board has created an operational dashboard, based on IOC, which serves as a command center. The command center monitors the water flow in the city and provides a data of the functioning of all the meters, amount of water transmitted by each meter and the water supplied to the distribution system. The IOC contains geo-information system to enable a real-time view of flow meters. IOC converts the data into a geospatial visual map to help BWSSB engineers to monitor the flow and distribution of water. Kerala Water Authority is also working with IBM to reduce water wastage. With the help of IBM’s Analytics and Mobility solutions, Kerala Water Authority monitors water distribution by mixing data analytics and a system -level view of the water infrastructure and lifecycle. The IBM systems data also helps in tracking water meters across the city and helps in reducing the billing irregularities, thus improving revenue collection. Kerala Water Authority can check the chlorine, PH, salinity level of the water with the help of the data. The data also KWA to respond quickly to the water supply irregularities and reacting to repairs in a shorter time frame that helps in saving the water wastage to a large extent. Concluding Note To control water wastage to a larger extent the government needs to implement the big data solutions country-wide. Right now, it is only limited to the southern part of the country. There are only limited number of companies that are working to fix the water woes in the country. To stop India going the Sao Paulo or Cape Town way, the government needs to step-up policies, measures and integrate technology. As many studies warn, India is sitting on a freshwater time bomb that is about to blow up in our face and alter the demographic and economic character of the country.","excerpt":"Big data has become a crucial part of every business – from Fortune 500 enterprises to startups. As it is changing the way we do business, many believe that big data has the power to solve world’s biggest issues. One such issue is the addressing water crisis in India. A recent UNESCO report predicts that Central […]","categories":["IT Services"],"tags":["Big Data","Data Analytics","IBM"],"author_name":"Smita Sinha","publish_date":"2018-03-23T12:03:41","publication_year":"2018","word_count":1067,"keywords":["big data","Go","Nuclio","AI","RAG","Ray","Aim","analytics","IBM","Data Analytics","Big Data","R","startup"],"extracted_tech_keywords":["AI","analytics","Aim","Ray","Nuclio","RAG","R","Go","big data","startup"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-big-data-is-used-to-address-water-crisis-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10094564,"title":"India’s Spacetech FDI Policy to Open Doors for Big Tech Investments","content":"India has lofty goals in the realm of space exploration with objectives like sending individuals to space and constructing an indigenous space station within a decade. In recent years, the Indian government has recognised the role the private sector will play in achieving these ambitious targets. Keeping in line with this, the government is planning to allow 100% foreign direct investment (FDI) in the space sector. India will allow FDI between 49-100% in three areas — sub-system manufacturing, launch vehicle operations, and satellite operations and establishments. A liberal FDI policy in the space sector is a welcome move because the sector is a capital-intensive one. In India, over the years, the space sector was led by the Indian Space Research Organisation (ISRO), but in 2020, it decided to open up the space industry to private players. Since then, we have seen various space-tech startups mushrooming in different parts of the country. As per the Economic Survey 2022, the number of startups in this space in 2021 was 41, compared to just 11 in 2019. However, currently, India only accounts for only 2% of the USD 44 billion global space economy, trailing heavyweights such as the US and China. Boosting FDI is a good move So far, industry players have responded positively to the development. They believe foreign funding will help Indian players get more competitive against global players such as Elon Musk’s SpaceX and Jeff Bezoz’s Blue Origin. “By attracting foreign investments through an automatic route or streamlined process, the government aims to supplement limited domestic funds and achieve the target of reaching USD13 billion by 2025 and capturing 10% of the global space economy by 2030,” Anil Prakash, director general at Satcom Industry Association (SIA) India, told AIM. He believes a liberal FDI policy will not only spur growth in the space sector but also have positive effects on various other sectors, including telecommunications, energy, and transport. “What is more important now is there will be financial incentives to private players to be able to operate and they will be able to get finances not only from the country, but from international sources too. And hence, a liberal FDI policy will be most welcome,” Lt Gen AK Bhatt (Retd), director general, Indian Space Association told AIM. Last year, Ajay Kumar Sood, principal scientific advisor, compared the growth spurt in the sector to the one experienced by the IT sector in the 1990s. He even believes India will have its own SpaceX in a few years. If Sood’s vision does come true, FDI will definitely play an important role in it. Additionally, 100% FDI opens up opportunities for job creation, skill development, and economic growth. It can attract new players, startups, and entrepreneurs to the space industry, creating a vibrant ecosystem and stimulating innovation. Gen. Bhatt believes that once the FDI policy is in place, funds will also flow into the sector from domestic players. “Funds will definitely flow from Indian big companies once it becomes profitable. I’m very sure this money will come within the country also. He believes investments could also come in from Indian IT giants such as Infosys, Tech Mahindra, TCS etc. “I see that happening because a very important part of the space economy is the space applications and quite a bit of space applications are IT related. So, I see that happening in the future,” he added. Big tech’s interest in India’s space sector The new FDI policy could potentially pave the way for big tech firms to invest in Indian startups, presenting a promising opportunity for collaboration and growth. So far, we have seen Google, Microsoft, Amazon, and IBM investing in various spacetech companies, both private and public. In 2017, SpaceX announced a partnership with Google to launch the biggest space project ever, which was later revealed in 2021 to be providing Starlink ground stations at Google’s data centres. Similarly, Microsoft too, has been partnering with NASA for lots of projects. In 2020, Microsoft made a deal with SpaceX to connect their cloud computing network through Starlink satellite. Given their interest in space tech, allowing FDI could potentially lead to significant investments from big tech in Indian startups. In fact, Google recently made its first investment in the Indian space sector. The tech giant led a USD36 million Series B funding round for Bengaluru-based startup Pixxel. Microsoft has also made its intentions known about collaborating with startups in India. The Redmond-based tech giant signed an MoU with ISRO to fuel the growth of space tech startups in India. This means startups in India will have access to Microsoft’s sophisticated tools, and resources to help them build and run their business. AWS too, last year, announced that it will work with Indian space tech startups as well as with ISRO. Similarly, the policy could also open the doors for existing space tech giants like SpaceX, Blue Origin, and Virgin Galactic to work in unison with the Indian space sector and unleash a new frontier in space technology. Addressing the concerns However, there is a flip side. A liberal FDI policy could mean foreign companies owning space tech startups in India and technologies and IPs developed in India could be taken out of the country. Gen. Bhatt is of the opinion that in the space sector, technologies will be developed not in just one nation, but globally. It will be a cumulative effort. “Further, 100% FDI is already allowed in the telecom sector in India. It has not led to us losing any technology or IP,’ he added. But he does agree that there should be checks and balances in place in terms of regulations for this to not happen. Another concern is the possibility of foreign companies dominating the local market and squeezing out domestic competitors, which can have negative implications for the overall economy. Additionally, excessive reliance on foreign investment may lead to economic vulnerability, as the country becomes highly dependent on external factors. Hence, Prakash also believes it is crucial to maintain proper regulatory frameworks and policies to safeguard national security interests. “Stringent measures need to be in place to prevent any misuse of technology or sensitive information by foreign entities. The government must strike a balance between attracting foreign investments and protecting national interests,” he said. Therefore, the implementation of appropriate regulations is crucial to mitigate the potential risks associated with a liberal FDI policy. Nonetheless, the recent developments are undoubtedly promising for the industry and could play a vital role in realising India’s ambitious goals in the space sector.","excerpt":"A liberal FDI policy will help bring new technologies, innovation, and expertise to India’s space industry, accelerating growth and development","categories":["AI Trends"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-06-06T14:00:00","publication_year":"2023","word_count":1089,"keywords":["Go","API","AWS","cloud computing","AI","ML","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","ML","Aim","cloud computing","AWS","R","Go","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/indias-spacetech-fdi-policy-to-open-doors-for-big-tech-investments\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039974,"title":"Tech Behind Introbot, A COVID-19 Helpline &#038; Database Of Verified Leads","content":"India is going through a health crisis like never before. The world’s largest democracy just recorded close to four lakh COVID-19 cases and about 4,000 deaths in a day. While countries across the globe are gradually taking control of the COVID-19 situation, India is struggling to deal with the second wave. The social media platforms have stepped up to the plate to make up for India’s flailing healthcare system. The last couple of weeks have seen Twitter, Facebook and Instagram flooded with leads for hospital beds, oxygen cylinders, plasma and Remdesivir etc. With help pouring in from all directions, verifying the authenticity of leads has become paramount. However, it’s easier said than done. To that end, Introbot is helping citizens with verified leads to save lives. Introbot was founded by Ohwow! founder and Stanford University alumnus Utkarsh Roy, and Divyaansh Anuj, a former Data Engineer at Oracle and founder of Dextra. Introbot was conceived to provide a centralised information source for virtual communities (think founder communities, business ecosystems, VC portfolios etc). However, with the second wave hitting hard, Divyaansh and Utkarsh decided to repurpose their bot to help people with information related to COVID-19 instead. The platform provides users with an evolving and verified database of beds, oxygen cylinders, plasma and other sources. While Divyaansh focuses on the tech part, Utkarsh handles the business side of Introbot. How does it work? The first step is to WhatsApp ‘COVID’ to +1(234)517-8991, post which the AI community manager helps them connect with relevant members based on broad objectives or specific needs. The sender can then respond in the ‘X in Y’ format (for example,”Hospital bed in Bengaluru” or “Plasma in Mumbai”). Introbot will share available leads along with details of when the lead was last verified. Launched around a week back, Utkarsh said, Introbot has been seeing a 10-15,000 spike in queries every day, with users seeking critical medical information through the Whatsapp bot. First week into the launch, the bot has already served almost half a million requests for medical supplies. The team collaborated with groups like Covid Citizens to recruit 350-plus volunteers, who manually fed data to the system. Today, the helpline has partnerships with more than 15 initiatives across the country, building India’s largest database of medical supply. Introbot has over 25,000 entries for oxygen cylinders, oxygen refills, beds, doctors, ambulance services, COVID-19 meals, among others, at a city-level Tech behind Introbot uses conversation AI on the front-end. The sourcing of information is extremely challenging today, with data from multiple sources. “It is very difficult to take in all the data available on social media, do a string passing analysis on them, cleansing the data and then making the machine understand what that data means, and be able to serve it to users through the bot,” said Utkarsh. Both humans and tech do the collection and verification of data. The verification process is three-tiered and is automated using a built-in system that does the pan-India mapping. Secondly, each city and postal code has seen a massive demand and supply gap. Thus, load balancing becomes taxing as well. Introbot is dynamically load-balancing the supply-demand process (to the tune of 100-requests per minute). On the supplier side, a lot of data cleansing has to be done. Introbot uses an algorithm for this and has a supply and demand feed. Taking in all the information, the platform stores all the data in its central database, feed it to the algorithms and uses a recommendation engine to match the supply and demand. Introbot works like a crowdsourced real-time instant system for the demand and supply of medical resources and COVID-based requests, much like what Uber does for cabs. “All of this happens in real-time, making the process all the more complicated,” Utkarsh adds. The third challenge comes with removing fraudulent and suspicious entries. The team removes the leads from its central database based on tips. Parallelly, Introbot makes sure Whatsapp vendors (Zendesk, Landbot.ai), systems’ are in place to handle the volume of requests. The platform does not collect any data apart from phone numbers. “We do not know who the person is, do not have any data on their age, gender or backgrounds. There isn’t much data that we are collecting in the first place. We have privacy policies and Terms and Conditions in place as well. We do not allow anybody to get access to personified information,” Utkarsh explains. Introbot facilitates the initial search and fastens the process of arriving at a verified lead by reducing the number of calls to verify sources significantly. Additionally, once it makes leads available, it seeks feedback on whether the lead is helpful, responsive or invalid. Once it receives the feedback, it updates its database. Introbot has responded to over five lakh-plus COVID-19 affected families, positively impacting over 25,000 lives. The platform’s services are used by Twitter-based Covid India Resource Bot (founded by Rahul Raina and Naman Gupta) to generate verified COVID-19 related data.base","excerpt":"India is going through a health crisis like never before. The world’s largest democracy just recorded close to four lakh COVID-19 cases and about 4,000 deaths in a day. While countries across the globe are gradually taking control of the COVID-19 situation, India is struggling to deal with the second wave.  The social media platforms […]","categories":["Deep Tech"],"tags":["data privacy use cases"],"author_name":"Debolina Biswas","publish_date":"2021-05-12T14:00:00","publication_year":"2021","word_count":828,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","data privacy use cases","R"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/tech-behind-introbot-a-covid-19-helpline-database-of-verified-leads\/","complexity_score":4,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10123865,"title":"Generative AI will Help Build $100 Million SaaS Businesses in India","content":"“In the next decade, we will see several $100m software businesses built from India for India. They will follow an entirely different playbook to US SaaS startups,” said Haptik AI founder Aakrit Vaish, adding that an entire industry is waiting to be disrupted. Salesforce’s shares recently plummeted by as much as 17% after the cloud software vendor reported weaker-than-expected revenues. However, Zoho founder Sridhar Vembu remained unaffected and recently took a dig at Salesforce for its disappointing quarterly numbers. Zoho continues to thrive, having not faced any layoffs and claiming zero debt on its balance sheet. AIaaS is the New Saas It would be naive to declare SaaS dead. Rather, existing SaaS companies are evolving into AI-first entities. “Legacy and new SaaS companies will truly become AI-first (not just marketing), abstract away the complexity of deploying LLMs,” said Matt Turck, VC at First Mark Cap, adding that AIaaS becomes the new SaaS. https:\/\/twitter.com\/rajeshsawhney\/status\/1794563707558838300 Indian SaaS companies such as Zoho, Freshworks, CleverTap, and Atlassian have added generative AI capabilities to their existing solutions. Meanwhile, numerous AI startups are developing new products based on these generative AI technologies. According to AIM, approximately 60 AI startups in India are building products using generative AI. Every AI startup that is not engaged in core research effectively becomes a SaaS company, or what we might call AIaaS. For instance, several companies are developing AI-powered conversational platforms for customer care, such as Exotel, Freshworks, Gupshup, Corover.ai, Limechat and Yellow.ai. Thiyagarajan Maruthavanan from Upekkha thinks that SaaS is not dead; instead, enterprise CMOs are overwhelmed by too many options. “Some people are selling more tools, such as GenAI tools, to solve their ‘too many tools’ problem. More sensible folks are going to their service partners, like Accenture and TCS, and asking how to use these GenAI tools to achieve the outcomes they were missing while the tools proliferated,” he said. India’s SaaS ecosystem has skyrocketed, with startups growing from 471 in 2016 to over 26,000 in 2022. Eleven Indian SaaS companies have surpassed $100 million in annual recurring revenue (ARR). This elite group includes pioneers such as Zoho, Druva, Icertis, and Freshworks, alongside emerging leaders like Innovaccer, Zenoti, and Postman. According to a recent report, the Indian SaaS industry is projected to reach $50 billion by 2030, with companies and unicorns generating $20-$25 billion in revenues by that time. SaaS x GenAI Zoho currently offers Zia, its generative AI assistant powered by OpenAI’s GPT-4, which helps generate business emails and social media posts, answer tickets, and generate minutes of the meetings. Freshworks reported that its GenAI platform, Freddy Copilot, has saved a B2C enterprise retail customer over $94,000 in the first year of implementation, resulting in a 187% return on investment. Similarly, Yellow.ai has already seen 10% of its enterprise customers choose GenAI-powered customer support, with response times reduced to 0.6 seconds. Another Indian startup LeadSquared is using LLMs provided by AWS Bedrock with customer-specific data to improve the quality of the chatbot responses and streamline the onboarding process. Meanwhile, Clevertap recently unveiled Clever.AI, designed to enhance customer engagement and retention. It can forecast precise business outcomes using data from CleverTap’s proprietary TesseractDB™, ensuring granular data and extended lookback periods for accurate predictions. Many other SaaS startups, such as ChargeBee, Druva, and Postman, are also integrating GenAI solutions. Non-Indian SaaS startups are also performing well. In an exclusive interview with AIM, Zendesk said that it plans to reach $3 billion by 2027. The company leverages Amazon Bedrock to scale generative AI applications, incorporating Anthropic’s leading model, Claude 3. Recently, Adobe’s stock jumped nearly 15% in after-hours trading, fueled by strong demand for its AI products and impressive earnings. “Adobe delivered a beat-beat-raise, and the market felt good about it, jumping almost 15% after hours,” said Daniel Newman, the CEO of TheFuturumGroup. He added that software, and in many cases SaaS specifically, will be one of the most important consumption layers for AI. “Much like how AI on iPhones will be a big growth driver for Apple, SaaS companies like Adobe will drive significant AI adoption and incremental growth by making AI easily accessible and consumable within their existing apps and captive user base,” he said.","excerpt":"Indian SaaS startups are betting big on generative AI to improve their products and services.","categories":["AI Features"],"tags":["SaaS"],"author_name":"Siddharth Jindal","publish_date":"2024-06-18T10:38:44","publication_year":"2024","word_count":698,"keywords":["Anthropic","GenAI","OpenAI","AI","AWS","ML","RAG","SaaS","Aim","generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","OpenAI","Anthropic","Aim","RAG","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/generative-ai-will-help-build-100-million-saas-businesses-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10085893,"title":"Happiest Minds Set to Acquire Madurai-based IT Services Firm SMI","content":"IT company Happiest Minds Technologies Ltd on Wednesday announced that it has signed an agreement to acquire 100 percent equity in Sri Mookambika Infosolutions (SMI), a Madurai-headquartered company, for Rs 111 crore. The acquisition was supported by Ernst & Young on the advisory side. SMI provides product engineering services to its US customers around Enterprise Applications & Integrations, digital data platform services (Analytics, Data Strategy, AI \/ ML, User Experience), Mobility Services and DevSecOps. Certified as a CMMI Level 3 and ISO 9001:2015 company, SMI delivers its engagements through agile delivery leveraging mature and industry-standard software engineering and development practices. In addition, the company has, over the years, built deep domain expertise around the healthcare vertical. With 400+ offshore-based employees, SMI has an annual run rate in revenues of US$ 9 Million. “We are excited to have the SMI team of 400+ join the Happiest Minds family. SMI brings in deep domain capabilities which add to our healthcare vertical strengths and align very well with our Product Engineering Services business unit. Working together, we seek to go deeper into the healthcare vertical.” said Joseph Anantharaju, Executive Vice Chairman & CEO – Product Engineering Services at Happiest Minds Technologies. In November 2022, the company inaugurated a new development centre at Bhubaneswar. The development comes as part of its long-term investment plans for expanding its presence. The new infrastructure continues the trend of IT spending concentrating towards Tier-II cities, leveraging the workforce available in the regions.","excerpt":"IT company Happiest Minds Technologies Ltd on Wednesday announced that it has signed an agreement to acquire 100 percent equity in Sri Mookambika Infosolutions (SMI), a Madurai-headquartered company, for Rs 111 crore.  The acquisition was supported by Ernst & Young on the advisory side. SMI provides product engineering services to its US customers around Enterprise […]","categories":["AI News"],"tags":["Mergers and Acquisitions"],"author_name":"Aparna Iyer","publish_date":"2023-01-25T18:28:59","publication_year":"2023","word_count":244,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","Git","RAG","analytics","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/happiest-minds-set-to-acquire-madurai-based-it-services-firm-smi\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10109965,"title":"After 30 Years, Windows Keyboard Gets a Copilot Upgrade","content":"Microsoft is set to introduce the Copilot key to Windows 11 PCs, a move hailed as the most significant change to the Windows PC keyboard in nearly three decades. The Copilot key joins the Windows keyboard as a core component of PC keyboard. When pressed, it will invoke Copilot in Windows, making it seamless to engage with Copilot in your day-to-day use. “Nearly 30 years ago, we introduced the Windows key to the PC keyboard that enabled people all over the world to interact with Windows. We see this as another transformative moment in our journey with Windows where Copilot will be the entry point into the world of AI on the PC, ”said Yusuf Mehdi, executive vice president & consumer chief marketing officer. The Copilot key is expected to debut on new Windows 11 PCs from ecosystem partners, including devices from AMD, Intel, and Qualcomm, with availability starting from late February through Spring. Surface devices featuring the Copilot key are also anticipated in the upcoming release. Microsoft wants 2024 to be “the year of the AI PC and said the Copilot key will not only simplify people’s computing experience but also amplify it. The tech giant said that the collaboration with silicon partners such as AMD, Intel, and Qualcomm has played a crucial role in Microsoft’s efforts to introduce new system architectures that power AI experiences on Windows PCs. This collaborative approach involves leveraging GPU, CPU, NPU, and cloud technologies to drive innovation. In the lead-up to and during CES, the Copilot key will be prominently featured on various Windows 11 PCs, showcasing Microsoft’s commitment to driving AI transformation and making it accessible to users.","excerpt":"The Copilot key is expected to debut on new Windows 11 PCs.","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2024-01-04T15:01:03","publication_year":"2024","word_count":276,"keywords":["Go","programming_languages:R","AI","innovation","ML","programming_languages:Go","RAG","AI transformation","R"],"extracted_tech_keywords":["AI","ML","RAG","R","Go","AI transformation","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/after-30-years-windows-keyboard-gets-a-copilot-upgrade\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25587,"title":"Creating A Sense Of Touch In Robots","content":"Where robots are being trained to be human-like in every possible way, researchers are now trying to gift them with an artificial nerve system which will allow them to mimic human skin — so that they can stretch, repair and transmit sensations to the brain and have reflexes. In an effort by the Stanford University and Seoul National University researchers, they have been able to create a system that can detect touch, process information and transmit it to other nerves. The work, which was originally reported by the researchers in Science, have been able to activate a sense of twitch in cockroaches as of now, and aim to explore this finding in the area of prosthetics and robotics. How Do The Artificial Nerves Work? Human skin works in a complex fashion and involves intricate sensing, signalling and decision-making systems. As the researchers are trying to create an artificial sensory nerve system that can process all the information, they constructed a nerve circuit that could be embedded in the skin-like covering for neuroprosthetic devices. The nerve circuit consisted of three major components: A touch sensor A flexible electronic neuron, which is a sensor that could send signals through the second component An artificial synaptic transistor — a brainchild of Tae-Woo Lee — which is able to process the sensory signals from the earlier two components. Explaining it in detail, researchers said that the sensitive touch sensor could detect even slightest of forces and changes in pressure. These sensors spark an electric voltage which is then picked up by the second component that transforms voltage to electrical pulses. These pulses are then passed to synaptic transistor, which is the central station for the device. It takes all the electrical pulses from all active sensors, which then integrates the signals. The artificial synaptic transistor is a replica of ‘human synapses’ that relay signals, store information to make simple decisions in a human body. All these functions are trained to do in the artificial nerve circuit. Using knee reflex as an example, Lee explained that a sudden tap causes the muscles to stretch, and sensors in those muscles send an impulse through the neuron, which then sends a series of signals to the relevant synapses. This causes the knee muscle to contract reflexively and send an immediate signal to register a sensation in the brain. The researchers are trying to replicate a similar neural code for artificial nerve system. The Practical Implementation Though practical uses of this artificial nerve system are far away, researchers aim is to make robots and prosthetics react reflexively and make prompt decisions. The researchers tested their process in two different setups to check the results. According to the paper, in one test they hooked up the artificial nerve to a cockroach leg and applied slight increments of pressure in the touch sensor. The sensor signal was converted to a digital signal through electronic neuron which was relayed through a synaptic transistor that caused the leg to twitch based on the pressure applied. “They also showed that the artificial nerve could detect various touch sensations. In one experiment the artificial nerve was able to differentiate Braille letters. In another, they rolled a cylinder over the sensor in different directions and accurately detected the direction of the motion,” noted the research finding. Though it is in its infancy, they aim to create artificial skin coverings for prosthetic devices that could detect sensations such as heat or cold and transmit this to the artificial brain in the coming years. Past Inventions Work By MIT Researchers: Researchers at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) had unveiled a sensor technology called GelSight eight years ago. It used physical contact with an object to provide a detailed 3D map of its surface. Working and improving upon the same model, researchers have now mounted GelSight sensors on the grippers of robotic arms, to give robots greater sensitivity. Robots could sense the hardness of the surface, and also manipulate smaller objects. Work By Researchers At University of Washington And UCLA: A team of researchers developed stretchable skin that could cover any part of a robot, giving them the power to sense vibrations and force applied to them. The researchers explained that as robot finger slides on to the surface, the electrically conductive liquid metal on the channels embedded on the skin stretches on one side of the finger and compress on the other. “This changes the amount of electricity that can flow through the channels, which can be correlated with sheer force and vibration,” they said. SynTouch, A Startup Creating Sense Of Touch: This California-based company is working on giving robots a human-like sense of touch. The startup which was recently awarded $2.5M in grants boasts a BioTac sensor, which consists of an elastic skin over an epoxy core, where the electronics are located. It contains a layer of fluid between the skin and epoxy core which is instrumental in generating electronic signals when an object is grasped by the robot. On A Concluding Note Not just the sensation of ‘touch’, but the researchers aim to advance into creating the sense of temperature, movement, texture, pressure and other parameters that would help robots and prosthetics to navigate through the environment. The researchers also aim to potentially combine artificial nerve with an artificial brain to interpret its output signals, which will power the next gen of bio-robots.","excerpt":"Where robots are being trained to be human-like in every possible way, researchers are now trying to gift them with an artificial nerve system which will allow them to mimic human skin — so that they can stretch, repair and transmit sensations to the brain and have reflexes. In an effort by the Stanford University […]","categories":["IT Services"],"tags":["robotic inventions"],"author_name":"Srishti Deoras","publish_date":"2018-06-20T09:07:38","publication_year":"2018","word_count":901,"keywords":["Replicate","Go","artificial intelligence","TPU","AI","robotic inventions","Git","Aim","ViT","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","TPU","R","Go","Git","ViT","startup","Replicate"],"url":"https:\/\/analyticsindiamag.com\/it-services\/creating-a-sense-of-touch-in-robots\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":28355,"title":"Meet Tesla T4, NVIDIA&#8217;s Fastest Data Centre Inferencing Platform","content":"Fueling the growth of artificial intelligence-based services worldwide, NVIDIA launched an AI data centre platform that delivers the industry’s most advanced inference acceleration for voice, video, image and recommendation services. The tech giant debuted the Tesla Turing 4 graphics processing unit (GPU) chip to speed up inference from deep learning systems in data centres. According to the official statement released by NVIDIA, the Tesla T4 GPU provides breakthrough performance with flexible, multi-precision capabilities, from FP32 to FP16 to INT8, as well as INT4. Packaged in an energy-efficient, 75-watt, small PCIe form factor that easily fits into most servers, it offers 65 teraflops of peak performance for FP16, 130 TOPS for INT8 and 260 TOPS for INT4. To optimise the data centre for maximum throughput and server utilisation, the NVIDIA TensorRT Hyperscale Platform includes both real-time inference software and Tesla T4 GPUs, which process queries up to 40x faster than CPUs alone. “We’re racing toward the future where every customer interaction, every product, and every service offering will be touched and improved by AI. Realising that the future requires a computing platform that can accelerate the full diversity of modern AI, enabling businesses to create new customer experiences, reimagine how they meet — and exceed — customer demands, and cost-effectively scale their AI-based products and services,” said the chipmaking giant. NVIDIA estimates that the AI inference industry is poised to grow in the next five years into a $20 billion market. Chris Kleban, product manager at Google Cloud, said, “AI is becoming increasingly pervasive, and inference is a critical capability that customers need to successfully deploy their AI models, so we’re excited to support NVIDIA’s Turing Tesla T4 GPUs on Google Cloud Platform soon.”","excerpt":"Fueling the growth of artificial intelligence-based services worldwide, NVIDIA launched an AI data centre platform that delivers the industry’s most advanced inference acceleration for voice, video, image and recommendation services. The tech giant debuted the Tesla Turing 4 graphics processing unit (GPU) chip to speed up inference from deep learning systems in data centres. According to […]","categories":["AI News"],"tags":["GPU","NVIDIA"],"author_name":"Prajakta Hebbar","publish_date":"2018-09-14T11:50:57","publication_year":"2018","word_count":283,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","deep learning","cloud_platforms:Google Cloud","NVIDIA","R","GPU"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","R","Go","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tesla-t4-nvidia-gpu\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":68515,"title":"Top 10 Python Libraries For Robotics","content":"One of the most popular languages, Python, is extensively used by emerging tech developers as well as robotics researchers. In robotics, the language has become a key part of the robot operating system (ROS) and is used for designing the embedded systems. For instance, the embedded systems and exhaustive automation packages of Raspberry Pi and Arduino are designed using this language. In this article, we list down the top 10 Python libraries for Robotics. (The libraries are listed according to their GitHub Stars) 1| Robot Framework About: Robot Framework is a generic open-source automation framework for acceptance testing, acceptance test-driven development (ATDD), and robotic process automation (RPA). The core framework is implemented using Python – it supports both Python 2 and Python 3 – and runs also on Jython (JVM), IronPython (.NET) and PyPy. Robot Framework is open and extensible and can be integrated with virtually any other tool to create powerful and flexible automation solutions. Know more here. 2| Pyro About: Python Remote Objects – or Pyro – is a library that enables you to build applications in which objects can talk to each other over the network, with minimal programming effort. Written in Python, this toolbox works between different system architectures and operating systems. It provides a set of powerful features that enables you to build distributed applications rapidly and effortlessly. Know more here. 3| DART About: Dynamic Animation and Robotics Toolkit – or DART – is a collaborative, cross-platform, open-source library that provides data structures and algorithms for kinematic and dynamic applications in robotics and computer animation. The library is distinguished by its accuracy and stability due to its use of generalised coordinates to represent articulated rigid body systems and Featherstone’s Articulated Body Algorithm to compute the dynamics of motion. DART also provides efficient computation of Jacobian matrices for arbitrary body points and coordinate frames. The library was created by the Graphics Lab and Humanoid Robotics Lab at Georgia Institute of Technology with ongoing contributions from the Personal Robotics Lab at the University of Washington and Open Source Robotics Foundation. Know more here. 4| PyRobot About: PyRobot is a Python library for benchmarking and running experiments in robot learning. It is a combination of two popular Python libraries, i.e. Requests and BeautifulSoup.  It can be used to drive applications that don’t provide an API or any way of hooking into them programmatically. This library will allow you to run robots without having to deal with the robot specific software along with enabling better comparisons. Know more here 5| PyDy About: Python Dynamics or PyDy is a tool kit written in the Python programming language that utilises an array of scientific programs to enable the study of multibody dynamics. The toolkit helps a user to perform visualisation, model specification, simulation, benchmarking, among others in their workflows. Know more here. 6| Simulation Open Framework Architecture About: Simulation Open Framework Architecture or SOFA is an open-source library and an efficient framework dedicated to research, prototyping and development of physics-based simulations. The library primarily focuses on real-time simulation, with an emphasis on medical simulation. The advanced software architecture of this framework allows the creation of complex and evolving simulations by combining new algorithms with existing algorithms, synthesis of complex models from simpler ones using a scene-graph description, among others. Know more here. 7| Klamp’t About: Kris’ Locomotion and Manipulation Planning Toolbox or Klamp’t is an open-source, cross-platform software package for robot modelling, simulating, planning, optimisation, and visualisation. The library aims to provide an accessible, wide range of programming tools for learning robotics, analysing robots, developing algorithms, and prototyping intelligent behaviours. Some of the features of this tool are- Simulation of various sensors including RGB+D cameras, laser sensors, gyroscopes, force\/torque sensors, and accelerometersMany sampling-based motion planners implementedSupports legged and fixed-based robotsContact mechanics computations: force closure, support polygons, the stability of rigid bodies and actuated robots Know more here. 8| Pybotics About: Pybotics is an open-source Python toolbox for robot kinematics and calibration. The toolbox was mainly designed to provide a simple, clear, and concise interface to quickly simulate and evaluate common robot concepts, such as kinematics, dynamics, trajectory generations, and calibration. Know more here. 9| Siconos About: Currently distributed under Apache Licenses, Siconos is an open-source scientific software primarily targeted at modelling and simulating nonsmooth dynamical systems. Written in C++ and Python, this software package can be used for modelling and simulation of dynamic systems. Know more here. 10| iDynTree About: iDynTree is a library of robot dynamics algorithms for control, estimation and simulation. The library is written in C++ language and supports several other languages including Python, MATLAB, among others. To use the library in Python language, you need to add the PYTHONPATH environment variable to the install path of the iDynTree.py file. To install, type export PYTHONPATH=$PYTHONPATH:<prefix>\/lib\/python2.7\/dist-packages\/ Know more here.","excerpt":"One of the most popular languages, Python, is extensively used by emerging tech developers as well as robotics researchers. In robotics, the language has become a key part of the robot operating system (ROS) and is used for designing the embedded systems. For instance, the embedded systems and exhaustive automation packages of Raspberry Pi and […]","categories":["AI Trends"],"tags":["Python","Python Libraries","Robotics","simple python project","software automation testing","Why is Python so Popular"],"author_name":"Ambika Choudhury","publish_date":"2020-06-30T13:00:00","publication_year":"2020","word_count":800,"keywords":["Go","API","AI","Python Libraries","simple python project","Why is Python so Popular","Robotics","Git","Python","software automation testing","Aim","Ray","C++","GitHub","R"],"extracted_tech_keywords":["AI","Aim","Ray","Python","R","Go","C++","Git","GitHub","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-python-libraries-for-robotics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10046623,"title":"Why Decompose a Time Series, and How?","content":"In time series analysis for forecasting new values, it is very important to know about the past data. More formally, we can say it is very important to know about the patterns which are followed by the values with time. There can be many reasons which cause our forecasted values to fall in the wrong direction. Basically, a time series consists of four components. Variation of those components causes the change in the pattern of the time series. These components are: Level: It is the main value that goes on average with time.Trend: The trend is the value that causes increasing or decreasing patterns in a time series.Seasonality: This is a cyclic event that occurs in time series for a short time and causes the increasing or decreasing patterns for a short time in a time series.Noise: These are the random variations in the time series. The combination of those components with time causes the formation of a time series. Most time series consists of the level and noise\/residual and the trend or seasonality are the optional values. They may take part or they may not. If seasonality and trend are part of the time series then there will be effects in the forecast value. As the pattern of the forecasted time series can be different from the older time series. The combination of the components in time series can be of two types: AdditiveMultiplicative Additive time series if the components of the time series are added together to make the time series. Then the time series is called the additive time series. By visualization, we can say the time series is additive if the increasing or decreasing pattern of the time series is similar throughout the series. The mathematical function of any additive time series can be represented by: y(t) = level + Trend + seasonality + noise Multiplicative time series If the components of the time series are multiplicative together, then the time series is called the multiplicative time series. By visualization, if the time series is having exponential growth or decrement with time then the time series can be considered as the multiplicative time series. The mathematical function of the Multiplicative time series can be represented as. y(t) = Level * Trend * seasonality * Noise The image below represents the additive and multiplicative time series. Image source In the above image, we can see the difference in the growth of values. In additive, it is quite slower and has a proper trend but on the other hand, we can see that the time series is growing exponentially with the time Instead of knowing the type of time series, it is better to know about the component of the time series. As we earlier discussed, the time series is the composition of level, trend, season and residuals. It is much better to know all the components in a time series for better understanding and making a model more accurate for better forecasting values. In this article, we will see how we can measure these components in any time series by decomposing a time series in its components. Here in the process, I am using the airline-passenger dataset which consists of the count of passengers in the airline with the date values. Importing the libraries: import pandas as pd import numpy as np Reading the dataset path = '\/content\/drive\/MyDrive\/Yugesh\/deseasonalizing time series\/AirPassengers.csv' data = pd.read_csv(path, index_col='Month') Checking for the 20 heads of the dataset. data.head(20) Output: Index of the dataset data.index Output: Here we can see the name of the index is Month and the length of the data is 144, in 20 heads we have seen that the column name of the time series is Passengers. Plotting the data: import matplotlib.pyplot as plt plt.rcParams[\"figure.figsize\"] = (15,10) data.plot() plt.show() Output: Here we can clearly see that there is a trend in the time series and there may be a chance to present seasonality in the time series. To check for all of the components in the time series by decomposition, we can use the python library statsmodel provided seasonal_decompose package. from statsmodels.tsa.seasonal import seasonal_decompose data = data.asfreq('MS') decompose_data = seasonal_decompose(data, model=\"multiplicative\") decompose_data.plot() plt.show() Output: Here we can clearly see by visualization that there is trend and season is present in the time series also the residual is showing high variability. We can check it more with additives. decompose_data = seasonal_decompose(data, model=\"additive\") decompose_data.plot() plt.show() Output: Here we can clearly see the variability of the residual in the additive model. We can also visualize the components separately. Visualizing the observed values. level = decompose_data.observed level.plot() Output: Visualizing the trend. trend=decompose_data.trend trend.plot() Output: Visualizing the seasonality. seasonality = decompose_data.seasonal seasonality.plot() Output: Visualizing the residual. residual = decompose_data.resid residual.plot() Output: We also save the components in tabular form by simply concatenating them component = pd.concat([level, trend, seasonality, residual],axis=1) component.tail(10) Output: Here in the article, we have gone through the procedure of decomposition of a time series. In time series if the components are available in uneven amounts then it can cause the model to predict wrong values. To improve the quality of the prediction we can perform detrending, deseasonalizing and perform some smoothing methods on it. I have covered some methods of cleaning a time series from its components in these articles: Comprehensive Guide To Deseasonalizing Time Series.Guide To Detrending Using Scipy Signal.How To Apply Smoothing Methods In Time Series Analysis. Also, some more information is provided regarding tests we need to perform in time series modelling. Complete Guide To Dickey-Fuller Test In Time-Series Analysis.Guide To AC and PAC Plots In Time Series.General Overview Of Time Series Data Analysis. After analyzing and performing all the required analyses in the time series we can go through the modelling of time series for which these articles can help a reader. Comprehensive Guide To Time Series Analysis Using ARIMA.Complete Guide To SARIMAX in Python for Time Series Modeling.Tutorial on Univariate Single-Step Style LSTM in Time Series Forecasting. I encourage you to go through the article and use it in real-world time series data. These all are the basic articles to perform time series analysis and are very easy to perform going through all you can master in time series analysis. References: Seasonal decomposition using moving averages.Google Colab Notebook for above codes","excerpt":"Components of time series are level, trend, season and residual\/noise. breaking a time series into its component is decompose a time series.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Machine Learning","Time Series","Time Series Analysis","Time Series Forecasting","time series prediction"],"author_name":"Yugesh Verma","publish_date":"2021-08-24T17:00:00","publication_year":"2021","word_count":1045,"keywords":["Time Series Analysis","Go","NumPy","TPU","AI","Machine Learning","Time Series Forecasting","RAG","Colab","Python","Time Series","time series prediction","Data Science","Matplotlib","R","AI (Artificial Intelligence)","Pandas"],"extracted_tech_keywords":["AI","Colab","Pandas","NumPy","Matplotlib","RAG","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-decompose-a-time-series-and-how\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":21238,"title":"Top 10 Investors In Indian Analytics Sector For 2017","content":"The year 2017 witnessed a total of $112.8M in the analytics startup funding pool. With the overall startup funding seeing a slump in investment for the Indian scenario, key investors have focused on diversifying their investment portfolio in various business areas and industries apart from data analytics. However, there were a few, that were focused on boosting the Indian analytics space. Here are the top 10 investors who have invested significantly in Indian startups in the year 2017. These are based on the amount of investment that the funds did last year. 1. Singtel Innov8 Investment : Qubole Singtel Innov8, part of of the Singtel group, is a venture capital firm that operates out of Singapore. It also serves enterprise clients with offices in San Francisco, USA and Tel Aviv, Israel. It is involved in business areas such as internet applications and digital media. It mainly backs companies which provide innovative solutions in communications and technology. Qubole, an analytics startup gets $25M in a joint investment with Harmony Partners. Its portfolio include Vuclip, Jasper Wireless, Airspace Systems, Zeotap, Moogsoft and many more creative tech-startups. 2. Harmony Partners Investment : Qubole Harmony Partners is a boutique venture capital firm based out of USA with offices in New York City and San Francisco. It mainly invests in technology, life sciences, and consumer businesses. Until 2017, it has invested over more than $750M in 80 companies. The firm has a rigid focus on investments in technological innovations. It funded Qubole jointly with Singtel Innov8 for a deal of $25M in November, 2017. Its portfolio include a large number of domestic investees such as Swiggy, Spotify, Illumio, Zerto, Natera and many more. 3. Eight Roads Ventures Investment : Unbxd Eight Roads Ventures is a venture capital firm headquartered in London, UK with offices in India, China and Japan. It is a subsidiary of Fidelity International Limited. It specialises in providing principal investments and in real estates. In June 2017, it raised $12.5M in a series C funding for Unbxd, a Bengaluru-based cloud technology startup. Existing investors, IDG Ventures and Inventus Capital also participated in the funding round. Its portfolio include Colt, DeltaHeath, Medvivo, CiplaHealth, Credr, Bankbazaar.com and many more. 4. DAH Beteiligungs GmbH Investment : Flytxt DAH Beteiligungs GmbH is a holding investment company which operated out of Mannheim, Germany. It is a subsidiary of the Service Innovation Group GmbH which provides business process outsourcing services for sales and marketing departments in many organisations. It offers sales consulting solutions such as Corporate and IT strategies. It mainly serves consumer electronics and other related sectors. Flytxt, a customer data analytics startup which started its operations in India by partnering with IIT Delhi, was funded by DAH Beteiligungs GmbH in April for $11M. Its portfolio comprises of the software giant SAP and their subsidiaries. 5. Omidyar Network Investment : i2e1 Omidyar Network is an investment firm based out of USA that specialises in providing funds for both profit and non-profit organisations. Their key sectors of investment are education, emerging technology, financial inclusion, governance & citizenship engagement and property rights. Recently, in November, it financed Wifi analytics startup i2e1 for $3M in a Series A funding round. Its portfolio include Indian School of Business, Akshara Foundation, 1MG, Healthkart, Change.org and many government service oriented organisations. 6. Vertex Ventures Investment : Active.ai Vertex Ventures is an investment firm headquartered in Palo Alto, USA which has primary operations in USA, South Asia and Israel. The firm makes investments in enterprise and cloud-platform companies. Their focus areas are mobile technology, enabling technology and SaaS. Active.ai, chatbot-oriented analytics startup raised $8.25M by Venture Ventures in collaboration with Creditease holdings and Dream Incubator. Its portfolio consists of online partners such as Yatra.com, Housejoy, Firstcry.com and 17 Media. 7 . Creditease Holdings Investment : Active.ai Creditease Holdings is a FinTech company from China, which offers wealth management services. Its primary investment interests are in sectors such as insurance technology, blockchain and similar business horizons. It also provides loan-advisory services to people. Active.ai was funded by Creditease jointly with Vertex Ventures and Dream Incubators for $8.25M. Its portfolio include Fintech companies such as Circle, Tradeshift and Marqeta. 8. Sistema Asia Fund Investment : Mobikon Sistema Asia Fund is a venture capital arm of Russian investment conglomerate, Sistema for financial operations in South Asia. The subsidiary is headquartered in Singapore. The target companies for investment include high-profile tech companies which are in mid-development stage — Series B and later funding rounds. Mobikon, an analytics startup which provides marketing automation assistance for restaurants, was funded by Sistema Asia Fund, Qualgro and C31 Ventures in a three-member joint funding with Sistema leading the funding round. Its portfolio includes enterprise and cloud-platforms such as Seclore, Qwikcilver, Wooplr and Licious. 9. General Catalyst Investment : Julia Computing General Catalyst is a private equity venture capital firm that funds early-stage investments. It is headquartered at Cambridge, USA with offices in Palo Alto,USA and New York City, USA. Their specific areas of investment interests are advanced materials, clean energy solutions, cybersecurity, mobile computing and related technologies. Julia Computing, an analytics startup specialising in developing Julia programming language, and services based on the language, got a joint seed funding of $4.6M from General Catalyst and Founder Collective with the former leading the round. Their portfolio includes major companies such as Airbnb, Snap and HubSpot. 10. Kalaari Capital Investment : EdGE Networks Kalaari Capital is an Indian venture capital firm based in Bengaluru. It provides mid-stage and late venture fundings typically Series A and Series B funding in the finance sector. The target sectors for funding is usually tech-savvy companies that assist finance firms such as e-commerce, retail and telecom services. EdGE Networks, a Bengaluru-based human resource analytics startup was financed by Kalaari Capital for $4.5M in a Series A round, with Ventureast joining the funding. Its portfolio include big players such as Snapdeal, Urban Ladder, ScoopWhoop, Bluestone.com and Zivame.","excerpt":"The year 2017 witnessed a total of $112.8M in the analytics startup funding pool. With the overall startup funding seeing a slump in investment for the Indian scenario, key investors have focused on diversifying their investment portfolio in various business areas and industries apart from data analytics. However, there were a few, that were focused […]","categories":["AI Features"],"tags":["analytics investment"],"author_name":"Abhishek Sharma","publish_date":"2018-02-02T11:39:18","publication_year":"2018","word_count":987,"keywords":["Go","API","ELT","analytics investment","AI","Git","analytics","CLIP","Julia","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Julia","Git","API","ELT","CLIP","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-investors-indian-analytics-sector-2017\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168338,"title":"Saying ‘Please’, ‘Thank you’ to ChatGPT Costs OpenAI Millions of Dollars","content":"OpenAI CEO Sam Altman has revealed in a tweet on X that polite courtesies with ChatGPT, such as “please” and “thank you,” have cost the company tens of millions of dollars in electricity expenses. I wonder how much money OpenAI has lost in electricity costs from people saying “please” and “thank you” to their models.— tomie (@tomieinlove) April 15, 2025 The computing efforts required for polite language interaction are quite significant, which leads to increased energy consumption in AI data centres. While Altman has spoken about the electricity costs the company incurs, it raises a more substantial issue: the environmental impact of AI tools. According to data provided by Semrush, ChatGPT has ranked as the eighth most visited site in the world as of March 2025, with approximately 5.56 billion visits. Research shows that every question potentially uses around 10 times more electricity than a simple Google search, with an average of 2.9 watt-hours of energy. UK-based energy supplier Business Energy UK states that data servers consume a significant amount of energy to train AI models and then process user demands. It added that even a small data centre can use up to 18,000 gallons of water per day, while large-scale data centres, such as Google’s, can consume up to 5,50,000 gallons daily. ChatGPT also runs a large number of calculations that are housed in data centres, which generate heat. Water systems are used as cooling equipment and keep the servers functioning. According to The Washington Post, data centres are one of the largest consumers of water in the locations where they are situated. Centres with electrical cooling systems also increase residents’ energy bills and strain the power grid. AIM reported that ChatGPT consumes the energy equivalent of powering 17,000 houses in the US per day. In another report, AIM highlighted that a 2027 projection shows that the world’s demand for AI will lead to significant water withdrawal, including both temporary and permanent sourcing of freshwater from underground or surface, which could be detrimental to the environment. Despite its massive impact on the environment and finances, some tech experts argue that using polite language enhances the user experience.","excerpt":"ChatGPT has ranked as the eighth most visited site in the world as of March 2025.","categories":["AI News"],"tags":["AI electricity costs","Energy consumption","OpenAI"],"author_name":"Smruthi Nadig","publish_date":"2025-04-21T15:56:30","publication_year":"2025","word_count":358,"keywords":["Go","ChatGPT","OpenAI","AI","RAG","GPT","Energy consumption","Aim","llm_models:GPT","R","llm_models:ChatGPT","AI electricity costs"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","RAG","R","Go","GPT","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/saying-please-thank-you-to-chatgpt-costs-openai-millions-of-dollars\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":22884,"title":"Latest Update On Cambridge Analytica Scandal: Mark Zuckerberg Vows To &#8216;Step Up&#8217; Security","content":"The controversy around Cambridge Analytica — the political data firm which has been accused of ‘harvesting’ personal data taken illegally from social media sites and used to influence elections — took another turn on Thursday, when Facebook CEO took responsibility over the data breach, saying that they would “step up” the security. In a detailed statement posted on Facebook, Mark Zuckerberg said, “I started Facebook, and at the end of the day I’m responsible for what happens on our platform. I’m serious about doing what it takes to protect our community. While this specific issue involving Cambridge Analytica should no longer happen with new apps today, that doesn’t change what happened in the past. We will learn from this experience to secure our platform further and make our community safer for everyone going forward.” After Cambridge Analytica was accused of illegally obtaining information on more than 50 million Facebook users, the social media giant banned the analytics firm Strategic Communication Laboratories and its political arm on the grounds of failure to follow their rules regarding handling of personal data. In what data experts are terming as one the largest abuses of personal data in history, Cambridge Analytica has been accused used this data to help target American voters for its work with (now US President) Donald Trump’s 2016 election campaign. Earlier this week, the Cambridge Analytica board had suspended their chief executive officer Alexander Nix. The decision to sack had come after the British news channel Channel 4‘s broadcast a video of an undercover operation, where Nix was seen talking about working on over 200 elections across the world — including in Nigeria, Kenya, the Czech Republic, India, Argentina – and bribing opponents to sway the election results. Meanwhile, Cambridge Analytica’s chief data officer Alexander Tayler will serve as acting CEO while the probe is under way. The company board said a full investigation into the comments and allegations would be led by noted UK lawyer Julian Malins. Meanwhile, in India, Indian National Congress as well as Bharatiya Janata Party have been asking the Election Commission to help them find proof of involvement (against each other) regarding alleged links with Cambridge Analytica. There are media reports of alleged data impropriety by chief of data analytics of Congress by his previous employer. It’s not just a question of association of Congress with a rogue company but a bigger challenge to free & fair elections & democratic values. #FacebookDataBreach pic.twitter.com\/F3DTEIdOEf — Ravi Shankar Prasad (@rsprasad) March 21, 2018 Ravi Shankar Prasad, Minister of Information Technology, warned Zuckerberg on Wednesday, saying, “Mr Mark Zuckerberg, you better know the observation of IT Minister of India. If any data theft of Indians is done with the collusion of Facebook systems, it will not be tolerated. We have got stringent powers in the IT Act including summoning you in India.” We welcome the fact that @facebook has one of the highest number of users from India but if any theft of data of Indians takes place in collusion with other companies for manipulation of democratic processes then that will not be tolerated. #FacebookDataBreach pic.twitter.com\/OBdv2vN7Ho — Ravi Shankar Prasad (@rsprasad) March 21, 2018","excerpt":"The controversy around Cambridge Analytica — the political data firm which has been accused of ‘harvesting’ personal data taken illegally from social media sites and used to influence elections — took another turn on Thursday, when Facebook CEO took responsibility over the data breach, saying that they would “step up” the security. In a detailed […]","categories":["AI News"],"tags":["BJP","cambridge analytica","congress","data breach","Facebook","Mark Zuckerberg","ravi shankar prasad"],"author_name":"Prajakta Hebbar","publish_date":"2018-03-22T07:07:20","publication_year":"2018","word_count":526,"keywords":["Go","cambridge analytica","congress","ravi shankar prasad","programming_languages:R","AI","programming_languages:Go","data breach","BJP","Mark Zuckerberg","analytics","Julia","Facebook","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Julia","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/latest-update-cambridge-analytica-mark-zuckerberg\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10098639,"title":"OpenAI Teams Up with Stainless to Unveil Version 4 of TypeScript\/Node SDK","content":"In collaboration with Stainless, a platform for high-quality, easy-to-use APIs, OpenAI has announced the release of Version 4 of their TypeScript\/Node SDK for the OpenAI API. https:\/\/twitter.com\/OfficialLoganK\/status\/1691875240647758123 The update includes a new set of improvements such as streaming responses for chat and completions, improved TypeScript types, compatibility with ESM, Vercel edge functions, Cloudflare workers, and Deno, a better file upload API for Whisper, fine-tune files, and DALL-E images, enhanced error handling through automatic retries and error classes, improved performance via TCP connection reuse, and simpler initialisation logic. Several users have taken to X to share their excitement about the project. For more information on how to access the new version, you can check out their  Github, NPM and migration guide. Additionally, OpenAI has acquired the team at AI design studio Global Illumination to work on our core products including ChatGPT. The team, with prior experience in creating and developing products during the initial stages at Instagram and Facebook, has also played a substantial role in advancing various projects at renowned companies like YouTube, Google, Pixar, Riot Games, and other prominent firms. All of these updates come amid the ChatGPT website experiencing reduced user activity. By the end of July, the user base further decreased by 12% to 1.5 billion users compared to 1.7 billion in June, as reported by SimilarWeb, excluding API usage. Read more: OpenAI Might Go Bankrupt by the End of 2024","excerpt":"The update includes a new set of improvements such as streaming responses for chat and completions, improved TypeScript types, compatibility with ESM and more.","categories":["AI News"],"tags":[],"author_name":"Shritama Saha","publish_date":"2023-08-17T11:44:49","publication_year":"2023","word_count":234,"keywords":["Go","ChatGPT","API","OpenAI","AI","Git","TypeScript","GPT","GitHub","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","TypeScript","Go","Git","GitHub","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-teams-up-with-stainless-to-unveil-version-4-of-typescript-node-sdk\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":23863,"title":"Using Natural Language Processing To Check Word Frequency In ‘The Adventure of Sherlock Holmes’","content":"Natural Language Processing is one of the most commonly used technique which is implemented in machine learning applications — given the wide range of analysis, extraction, processing and visualising tasks that it can perform. In this article, you will learn how to implement all of these aspects and present your project. The primary goal of this project is to tokenize the textual content, remove the stop words and find the high frequency words. We shall implement this in Python 3.6.4. To start with, we shall look into the libraries that we are going to use: Beautifulsoup: To scrape the data from the HTML of a website and it also helps to process only the text from these HTML codes Regular Expressions: Also known as Regex. It will convert the noise data containing special characters and carry the conversion of uppercase to lowercase characters NLTK (Natural Language Toolkit): For the tokenization of the sentences into a list of words We are using the eBook for, The Adventure of Sherlock Holmes by Sir Arthur Conan Doyle, which is available here. Let Us Grab The URL Of The Book And Start Our Project Data Extraction: Assign the url to an object as below, Now, after we have the URL, let us try to make a request. Once you are go through the browser while visiting a web page, it shows request as below. requests make this easy with its function. Make the request here and check the object type returned. There are other types of requests, such as POST requests, but that is not of our concern for this project. After getting the html script from the link, let us process this html to get the text from the body. Text Extraction From HTML: We shall make use of Beautifulsoup to extract the string of words from the html content. Let’s import the Beautifulsoup from bs4 and parse the html content with the argument “htmllib”. You can also use other parameters such as “lxml”, “html” etc. Let us look at the title of the eBook, to learn more about the functioning of the Beautifulsoup here. To extract just the string from the contents inside the title tag, follow this code Let us take a look at all the chapter available inside the book and how they are represented in HTML code. This is the output that we are looking for. The complete Sherlock Holmes’ eBook textual content can be access with .get_text() command. Now that you have the text of interest, it’s time for you to count how many times each word appears and to plot the frequency histogram that you want. This is where Natural Language Processing comes into picture. Extract Words From Your Text With NLP: We’ll now use nltk, the Natural Language Toolkit, to Tokenise the text (splitting sentences into words (list of words)); Remove stopwords (remove words such as ‘a’ and ‘the’ that occur at a great frequency). We will be using the regular expressions first, to remove all the unwanted data from the text. the ‘\\w’ is a special character that will match any alphanumeric A-z, a-z, 0-9, along with underscores; The ‘+’ tells you that the previous character in the regex can appear as many times as you want in strings that you;re trying to match. This means that ‘\\w+’ will match arbitrary sequences of alphanumeric characters and underscores. Let us now convert all the uppercase letters to lowercase letters, which is a mandatory task because in Python, uppercase and lowercase are considered as different objects. Removal Of Stop Words: It is common practice to remove words that appear frequently in the English language such as ‘the’, ‘of’ and ‘a’ (known as stopwords) because they’re not so interesting. The package nltk has a list of stopwords in English which you’ll now store as sw and of which you’ll print the first several elements. If you get an error here, run the command nltk.download (‘stopwords’) to install the stopwords on your system. Now we need to remove all the words that are now in sw  from the original text to complete the NLTK extraction and processing. Presenting The Project: With the help of seaborn and matplotlib, let us visualise how the data is scattered and present our NLP model on the book The Adventures of Sherlock Holmes by Arthur Conan Doyle. Let us now look at how the graph looks and also the tokenised word count. Here we will be ending our model and finally present our findings with the graph below.","excerpt":"Natural Language Processing is one of the most commonly used technique which is implemented in machine learning applications — given the wide range of analysis, extraction, processing and visualising tasks that it can perform. In this article, you will learn how to implement all of these aspects and present your project. The primary goal of […]","categories":["AI Features"],"tags":["Natural Language Processing","NLP","NLTK"],"author_name":"Kishan Maladkar","publish_date":"2018-04-19T12:45:28","publication_year":"2018","word_count":754,"keywords":["machine learning","TPU","AI","Natural Language Processing","ML","NLP","Python","Seaborn","NLTK","Matplotlib","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","NLTK","Matplotlib","Seaborn","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/using-natural-language-processing-to-check-word-frequency-in-the-adventure-of-sherlock-holmes\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10096674,"title":"Cloudera Expands Open Data Lakehouse Empowering Trusted Enterprise AI","content":"Cloudera, the hybrid data company, has unveiled an expansion of its Open Data Lakehouse offerings, allowing users to harness the power of analytics and AI capabilities for all their data within their enterprises, whether it resides in the cloud or on-premises. This move builds upon Cloudera’s previous introduction of support for Apache Iceberg V2 in its CDP-Public Cloud offering last year, enabling seamless access to emerging technologies like large language models (LLMs) and real-time self-service analytics at scale. Notably, Cloudera stands apart as the sole provider of an open data lakehouse that spans multiple public clouds and on-premises environments, retaining their data where it is best suited. Moreover, the unified security and governance offered by CDP ensures consistent data access and protection, irrespective of its structure or format. Ram Venkatesh, Cloudera’s Chief Technology Officer, expressed enthusiasm about the enhancement, stating, “Large enterprises want to extract business value from all their data using AI and data analytics. Our announcement today of Apache Iceberg support on private clouds means our best-in-class Open Data Lake House just continues to get better.” He further emphasised that customers can now leverage the power of Iceberg “everywhere” they need it to be. Mayank Baid, Regional Vice President, India, Cloudera, highlighted the challenges faced by businesses in centralising their structured and unstructured data and the growing demand for actionable insights. Baid remarked, “To stay ahead of the curve, we are expanding our best-in-class Open Data Lakehouse and announcing support for Apache Iceberg for CDP-Private Cloud, which enables organisations to fully utilise the value of their data and delivers Iceberg ‘everywhere’ the customers need it to be.” Cloudera’s latest development reflects the increasing need for businesses to unlock the potential of their data through advanced analytics and AI, catering to the evolving demands of enterprises striving for centralised data management and agility in generating valuable insights. With its Open Data Lakehouse and Apache Iceberg integration, Cloudera aims to provide organisations with a comprehensive solution for maximising the value of their data assets.","excerpt":"This move builds upon Cloudera’s previous introduction of support for Apache Iceberg V2 in its CDP-Public Cloud offering last year.","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-07-11T12:58:53","publication_year":"2023","word_count":334,"keywords":["Go","programming_languages:R","AI","ML","RAG","Aim","analytics","GAN","R","data lake"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","R","Go","data lake","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cloudera-expands-open-data-lakehouse-for-trusted-enterprise-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10052327,"title":"Psst… Amazon Is Busy Transfer Learning","content":"Amazon, perhaps, is one of the only few players in the machine learning landscape to have heavily invested in transfer learning, besides Facebook, Microsoft and DeepMind. The success of this can be seen in its Alexa virtual assistant, which has reached significant strides in the last few years, competing closely with Google Assistant and Apple’s Siri. For those unaware, transfer learning is a technique where learning in a new task is done via the transfer of knowledge from a related task that has already been learned. For instance, knowing how to ride a bicycle makes learning how to ride a motorcycle easier. Similarly, this idea can be applied to machine learning instead of learning or developing a large scale language model with billions of parameters from scratch. Doing so is not only time consuming but also expensive. Developers or researchers have to collect thousands of voice samples and annotate\/label them manually, a process that can take months or weeks easily. That is why researchers at Amazon Alexa have pursued transfer learning, which leverages a neural network trained on a large dataset of previously labelled samples to train in a new domain\/area with sparse data. Today, the Amazon Alexa team uses transfer learning technology extensively to transfer knowledge across various language models, features, and better machine translation capabilities. Transfer Learning Everywhere Not just Alexa, Amazon has been working on transfer learning across various areas, including product recommendation, AutoML, computer vision, and others. Two years ago, Amazon introduced two new features called ‘Newscaster and Neural text-to-speech (TTS)’ to its cloud-based TTS service – Amazon Polly. Launched in 2016, Amazon Polly turns text into human-like speech, allowing users to develop speech-enabled products and applications. Besides Polly, Amazon offers multiple APIs that aim at executing tasks within text analysis, which can also explore transfer learning to further customise the inference of these models. Some of them include Amazon Personalize, Amazon Forecast, Amazon Transcribe, Amazon Rekognition, Amazon Comprehend, Amazon Lex, Amazon Textract, Amazon Translate, etc. Here’s a list of all the research work done in transfer learning by Amazon researchers in the last four years. Shaping Transfer Learning Four years ago, Amazon founder Jeff Bezos (executive chair), at the annual Amazon letter, described Alexa’s success and improvements from semi-supervised and transfer learning. He said they have dramatically reduced the amount of time required to teach Alexa new languages by using machine translation and transfer techniques, which allowed them to serve customers in more countries, including India and Japan. Last month, Amazon released new tools like notebooks, text models and solutions for multimodal financial analysis within Amazon SageMaker JumpStart. Using these tools, you can easily retrieve public financial documents, including SEC filings, and process financial text documents with features like summarization and scoring for various attributes such as sentiment, risk and readability. In addition, users can access pre-trained language models trained on financial text for transfer learning and use example notebooks for data retrieval, text feature engineering, regression models and multimodal classification. Today, transfer learning has become one of the most popular techniques in deep learning as it can train deep neural networks with little or limited data in a short period of time. Previously, tech evangelist Andrew Ng, at NIPS 2016, had said that transfer learning would be – after supervised learning – the next driver of machine learning commercial success. Cut to the present; Amazon is most certainly leading the way for transfer learning.","excerpt":"Amazon Alexa team uses transfer learning extensively to transfer knowledge across various language models, features, and better machine translation.","categories":["Global Tech"],"tags":["Amazon","Amazon Alexa","Transfer Learning"],"author_name":"Amit Naik","publish_date":"2021-10-26T15:00:00","publication_year":"2021","word_count":570,"keywords":["Amazon SageMaker","machine learning","AI","neural network","ML","Amazon Alexa","Amazon","computer vision","RAG","Aim","deep learning","Transfer Learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","Aim","Amazon SageMaker","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/psst-amazon-is-busy-transfer-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168981,"title":"Cognizant Q1 Net Profit Jumps 21%, Reports 1,400 Generative AI Engagements","content":"Cognizant reported a 21% year-over-year increase in net profit for the first quarter ended March 31, 2025, reaching $663 million, up from $546 million in the same quarter last year. The company’s revenue rose 7.5% to $5.1 billion during the quarter YoY, but stayed flat quarter on quarter, meeting the expectations set during the last quarter. CEO S Ravi Kumar attributed the performance to growing demand for AI-driven IT services. “We started the year on a strong note, delivering revenue and adjusted operating margin ahead of our expectations, reflecting our steadfast focus on the execution of our strategy over the last several years,” said Kumar in a company release. During the earnings call, Kumar said that the company is actively scaling its work in generative AI, reporting approximately 1,400 early-stage GenAI engagements, up from 1,200 in the previous quarter. Cognizant has developed over 20 agentic solutions using Google’s LLMs, particularly focused on healthcare, addressing issues such as prior authorizations, fraud, and member experience. Kumar pointed out that “AI-written code increased to more than 20% for us,” calling it a pioneering moment for developer productivity. Cognizant is also building out a sophisticated framework for industrialising AI, working with hyperscalers like Microsoft, AWS, and Google, as well as deepening its collaboration with NVIDIA to develop enterprise AI agents, industry-specific large language models, and AI infrastructure. A notable internal breakthrough includes a patent-pending method for managing LLM hallucinations by setting uncertainty thresholds, allowing fallback to rule-based systems when confidence is low. On the GCC front, Cognizant is expanding aggressively, most recently announcing a new GCC with Citizens Financial in Hyderabad. This centre will focus on enhancing enterprise technology, data, and analytics for the client, using Cognizant’s Neuro and Flowsource AI platforms. Kumar sees GCCs as a strategic growth lever, adding, “We are undertaking more engagements to support clients on the GCC journey, equipping them with strategic AI tooling and platforms needed to drive operational strength.” Earlier, Cognizant in its 2024 annual report pointed out that there is potential threat from GCC when it comes to acquiring talent. While verticals like health sciences, financial services, and products and resources contributed to the revenue boost, the Communications, Media, and Technology segments experienced a decline. The quarter also included a $62 million gain from the sale of an office complex in India. Bookings on a trailing twelve-month basis increased 3% year-over-year to $26.7 billion, representing a book-to-bill ratio of approximately 1.3X. However, bookings in the first quarter alone declined 7% from the previous year. The quarter included four large deals, each valued at $100 million or more. Employee metrics showed that voluntary attrition in the tech services segment stood at 15.8% on a trailing twelve-month basis, compared to 15.9% in Q4 2024 and 13.1% in Q1 2024. Total headcount at the end of March 2025 was 336,300, marking a reduction of 500 employees from the prior quarter. Looking ahead, Cognizant projected second quarter revenue to range between $5.14 billion and $5.21 billion, and full-year 2025 revenue between $20.5 billion and $21.0 billion.","excerpt":"CEO Ravi Kumar pointed out that “AI-written code increased to more than 20% for us,” calling it a pioneering moment for developer productivity.","categories":["AI News"],"tags":["Cognizant","Quarterly Earnings"],"author_name":"Mohit Pandey","publish_date":"2025-05-01T09:17:55","publication_year":"2025","word_count":506,"keywords":["Go","GenAI","AWS","AI","cloud_platforms:AWS","R","Cognizant","AI agents","ViT","generative AI","analytics","Quarterly Earnings"],"extracted_tech_keywords":["AI","analytics","generative AI","GenAI","AWS","R","Go","ViT","AI agents","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cognizant-q1-net-profit-jumps-21-reports-1400-generative-ai-engagements\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10023629,"title":"Bangalore-based No-code AI Video Communication Platform Raises $300K In Seed Funding","content":"Superpro.ai, the AI-powered video communication workflows platform, received a boost in its portfolio with funding of $300K as part of its seed round led by IvyCap Ventures Angel Fund. According to the official release, the received funds will be utilised for “accelerating the product development, increasing the platform’s sales and marketing efforts to cater to more customers, and reaching product-market fit.” Headquartered in Bangalore, this AI-powered video communication workflows platform, Superpro.ai, assists startups in the edTech, health Tech, HR tech, and events industries with video communication workflows. Talking about the fresh round of funding, Gaurav Tripathi, the Founder and CEO of Superpro.ai said, that the company has been established “with a vision to power a sustainable and fulfilling future of work,” with “remote and video communication is at the core of it.” Thus, the startup has been helping edtech, health tech, and eCommerce companies to launch video communication as part of their service delivery workflows. This fresh round of funds will “help us iterate faster with the product and reach product-market fit faster,” said Tripathi. Along with speeding up the product development, Superpro.ai has also released mobile SDKs for its video communication workflow platform, which are compatible with Android, iOS, and React Native platforms. Apart from IvyCap, other prominent angel investors participated in this seed funding round include Pentathlon Ventures; ah! Ventures, SOSV – Superpro.ai’s existing investor, Piyush Prahladka, ex-Google, ex-Uber, and Gopi Vikranth, the Associate Principal ZS, ex-Mu-sigma. With its AI and automation capabilities, Superpro.ai helps companies provide an advanced video communication experience. It equips startups with no-code and low-code resources to power video conversations within their websites or applications.","excerpt":"Superpro.ai offers a better customer experience around video communication, while IvyCap aims to build technology-driven innovative companies.","categories":["AI News"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-04-08T14:15:10","publication_year":"2021","word_count":271,"keywords":["Go","funding","programming_languages:R","AI","programming_languages:Go","automation","R","startup"],"extracted_tech_keywords":["AI","R","Go","automation","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/superpro-ai-received-usd-300k-seed-funding\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10065867,"title":"Capgemini delivers record growth of 17.7 per cent in Q1 2022","content":"The Capgemini Group’s activity accelerated in Q1 2022, with a growth of 17.7 per cent at constant exchange rates. The growth has strengthened compared to the 12.5 per cent observed in Q4 2021 across all regions and sectors. As a result, organic growth reached 16.3 per cent, up 3.1 points in Q4 2021. Aiman Ezzat, Chief Executive Officer of the Capgemini Group, said: “Capgemini delivered an excellent start with a further growth acceleration versus the previous quarter. This is the fourth consecutive quarter with double-digit growth. It demonstrates the Group’s growth profile change and our ability to gain market share. We are taking advantage of the strong alignment of our capabilities and offerings portfolio with our clients’ structural demand for digital transformation. Our digital transformation capabilities and strong industry focus enable us to be positioned as our clients’ key partners for their most ambitious transformations. Our brand is strong, and we attract and develop the best talent in a very competitive market. As a result, we increased our global workforce by 16,000 in the last quarter to reach 340,000 employees worldwide.” Operations by region The United Kingdom and Ireland region (12 per cent of the Group revenues in Q1 2022) again enjoyed a particularly robust quarter with revenue growth of 21.3 per cent at constant exchange rates, driven primarily by the public sector and consumer goods. The North America (29 per cent of Group revenues) and the rest of Europe (30 per cent of Group revenues) regions grew at constant exchange rates of 16.8 per cent and 16.0 per cent, respectively. France (20 per cent of Group revenues) reported revenue growth of 11.1 per cent at constant exchange rates, with a robust momentum in the manufacturing and consumer goods sectors. Finally, revenues in the Asia-Pacific and Latin America region (9 per cent of Group revenues) increased sharply, by 42.6 per cent at constant exchange rates, as Group acquisitions in the region added to strong organic growth.","excerpt":"The Capgemini Group reported Q1 2022 revenues of EUR 5,167 million, up 21 per cent year-on-year at current exchange rates and 17.7 per cent at constant exchange rates.","categories":["AI News"],"tags":["capgemini","Fiscal Earnings","quarterly results","Revenue","Revenue Growth"],"author_name":"Poornima Nataraj","publish_date":"2022-04-28T15:48:17","publication_year":"2022","word_count":326,"keywords":["Go","Fiscal Earnings","programming_languages:R","AI","Revenue","R","digital transformation","quarterly results","Git","Aim","llm_models:Gemini","ViT","GAN","capgemini","Revenue Growth"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GAN","ViT","digital transformation","llm_models:Gemini","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/capgemini-delivers-record-growth-of-17-7-per-cent-in-q1-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":30786,"title":"Strong AI? Industry Majors Are Reeling From Gender Bias In AI Tools","content":"“All animals are equal, but some animals are more equal than others” said George Orwell in his novel titled Animal Farm. But in the world of artificial intelligence, Google’s Gmail product manager Paul Lambert has found out that, “Not all ‘screw ups’ are equal,”. This was his response when the search engine giant found out there was at least one serious defect in the AI software running Gmail. It did not understand gender pronouns. Many benchmarks are broken every now and then.As we wrote recently, Sony announced that the company has achieved the best speed in the industry. They achieved the industry best performance by using distributed learning. But at the same time, there are many fundamental things where AI capabilities need to be improved. In the last few months,  Google’s product Gmail introduced the autocomplete feature that helps users complete their sentences and craft emails faster. This feature does not suggest “him” or “her” now since Google’s supposedly smart feature calls itself “Smart Compose” fails to understand gender pronouns. Gender Bias Issues The change has come from the fear that Gmail may suggest a wrong pronoun because of the imperfect AI techniques and may offend many users. One of the research scientists working at Gmail found one such example. When the researcher typed, “I am meeting an investor next week,” and Smart Compose responded by suggesting a follow-up question: “Do you want to meet him?” and did not mention the female pronoun “her.”  Many such examples came from regular usage of the product. Google did not want to risk a PR disaster in a world where gender issues and discussions are on the rise. Google is not the only major AI company to face such issues. A team at Amazon has been working on a product to review job applicants’ resumes. The hiring tool used AI to give job candidates some sort of scores to evaluate their readiness for the role. But just within a year they realised the system was biased towards male candidates because of the biased training it had gotten. Most of the data Amazon had collected came from men, which was a reflection of male dominance in the industry. New Tools On Hold LinkedIn, the top social networking network for professionals also tried to build an algorithmic ranking for candidates to assess the fit for the job. John Jersin, vice president of LinkedIn Talent Solutions, told Reuters that the service is not a replacement for traditional recruiters. “I certainly would not trust any AI system today to make a hiring decision on its own,” he added. In the world where there is a rush to introduce new AI tools and services, gender based tools are taking a backseat. It is a matter of great intrigue that AI still can not identify gender pronouns and has bias towards male despite of various efforts to eradicate such a bias. Certainly does not seem like a strong AI is emerging any time soon. Once it does, it will know if a person is a “he” or “she.”","excerpt":"“All animals are equal, but some animals are more equal than others” said George Orwell in his novel titled Animal Farm. But in the world of artificial intelligence, Google’s Gmail product manager Paul Lambert has found out that, “Not all ‘screw ups’ are equal,”. This was his response when the search engine giant found out […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Amazon","gender bias","Gmail","Google","linkedin","Sony"],"author_name":"Abhijeet Katte","publish_date":"2018-11-28T09:07:20","publication_year":"2018","word_count":509,"keywords":["Go","artificial intelligence","Rust","Gmail","AI","programming_languages:R","Amazon","programming_languages:Go","BERT","llm_models:BERT","Google","Sony","linkedin","AI (Artificial Intelligence)","R","programming_languages:Rust","gender bias"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Rust","BERT","llm_models:BERT","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/strong-ai-industry-majors-are-reeling-from-gender-bias-in-ai-tools\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":33733,"title":"The Sound Of AI Music","content":"When we talk about artificial intelligence, people always tend to think of the only spaces where this tech can replace humans. However, there is another vertical, which is one of those least expected, that AI has influenced in recent times — Music composition. The music industry has also witnessed tremendous transformations done by AI over the past couple of years — not only in terms of listening to music but also in terms of how music is made. American singer-songwriter, Taryn Southern in 2017 created a big, moody ballad song Break Free entirely produced by an AI. Not only the song but Southern also went on to create the entire album I AM AI which is the first LP to be entirely composed and produced using AI. Today, AI has reached such a great level that there’s a whole industry built around AI services for creating music. Whether it is sounds of distortion or electronic beats, AI has opened the doors of new possibilities for sound generation Here are Some Of The Most Popular AI-Powered Music Making Tools: IBM Watson Beat: The brainchild of Janani Mukundan, a member of the IBM Research team in Austin, Texas, IBM Watson Beat is AI-driven music making an app that over the years have gained tremendous popularity in the music industry. Janani was fascinated about the IBM Watson AI technology and she thought of leveraging this amazing tech in the domain of music. Therefore, with the help of machine learning, she set started out to build a neural network that would compose original music. However, things weren’t that easy; after months of hard work, Janani brought in Richard Daskas, a professional musician and software engineer. The duo taught the system, things like pitch, rhythm, chord progression and instrumentation. Also, they provided a massive number of data points into the neural network. Today, IBM Watson Beat is one of the best music making tools in the industry. “The system’s neural network understands music theory and how emotions are connected to different musical elements. And then taking the basic ideas, Beat creates something completely new and unique.,” said Janani. NSynth Super: “I am fascinated with this about every time we see new technology come along, artists take interesting bits to the technology and turn into new ways to making music,” said Doug Eck, Research Scientist on the Google Brain Team. Doug runs a project called Magenta, a project that explores how machine learning tools can help artists create art and music in new ways, and NSynth Super is part of this project. Stands for Neural Synthesiser, NSynth is one of the Magenta’s first projects. The NSynth algorithm, using deep neural network, learns the core aspects of what makes a sound, sound like it does. And then the system combines the characteristics of different sounds and comes up with new sound or tracks which is not the blending of different sounds; it’s completely new. Also, It has the ability to generate more than 100,000 sounds. Despite all the amazing things NSynth can do, the major benefit is that it is open source, NSynth Super is built using open source libraries to welcome a greater and wider community of artists, coders, and researchers to experiment with machine learning. Jukedeck: Started by a duo, Patrick Stobbs and Ed Newton-Rex, Jukedeck is another AI-driven music maker. This amazing music writes music using artificial intelligence to music composition and production. Inspired by the neural networks of the human brain The system is trained in such a manner that the deep neural networks understand music composition at a granular level “We build up digital understanding, or representation, of a composer’s brain and use that to create original tracks of music,” says Patrick Stobbs at an Interview with The Guardian. Amper Music: Amper is a New York City-based music software company that uses artificial intelligence to create music. The tool has the ability to make music across a range of genres and moods and is designed in such a way that it gives users creative control regardless of expertise. Everything in the Amper music is constructed from the scratch and has a proprietary library that has numerous individual samples. The thing that makes Amper class-apart is the fact that even though, it powered by AI, the sounds created by Amper Music doesn’t sound synthetic. “Our AI doesn’t compose music for you, it composes with you. You have full control to shape Amper’s output. Our AI lets humans take control,” reads Amper’s official site And if you don’t believe how good Amper is, then you should definitely listen to Taryn Southern’s I AM AI. Outlook No matter how good a technology is, naysayers will always be there. In the previous years when Auto-tune and things like digital workstations entered the arena, they were criticized a lot, but today, they have become mainstream — they are being used in music creation. The same way AI is also facing criticism. But the rate at which it is evolving and transforming the music industry, the days are not so far when AI-powered music making machines will become go-to for artists.","excerpt":"When we talk about artificial intelligence, people always tend to think of the only spaces where this tech can replace humans. However, there is another vertical, which is one of those least expected, that AI has influenced in recent times — Music composition. The music industry has also witnessed tremendous transformations done by AI over […]","categories":["AI Features"],"tags":["AI in music","Deep Learning","Neural Network"],"author_name":"Harshajit Sarmah","publish_date":"2019-01-18T10:52:43","publication_year":"2019","word_count":853,"keywords":["AI in music","Neural Network","Go","artificial intelligence","machine learning","TPU","AI","neural network","Git","RAG","Dask","Deep Learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","Dask","RAG","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-sound-of-ai-music\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10052667,"title":"A Guide to Self-Supervised Learning in Computer Vision","content":"In the last few years, we have seen that self-supervised learning methods are emerging rapidly. It can also be noticed that models using self-supervised learning methods have solved many of the problems regarding unlabeled data. Uses of these methods in fields like computer vision and natural language processing have shown many great results. In this article, we are going to discuss the self-supervised learning methods used in the field of computer vision. We will also discuss the contrastive learning used as a self supervised approach to address the data labelling problems in computer vision. The major points to be discussed in this article are listed below. Table of Contents Why Self-supervised Learning is Needed?The Contrastive LearningContrastive Predictive Coding (CPC)Instance Discrimination Methods Let us begin the discussion by understanding why self supervised learning is needed. Why Self-supervised Learning is Needed? We require a lot of labeled data when working on the supervised learning techniques. Most of the time data labelling becomes very costly, and time consuming especially in the field of computer vision where tasks such as object detection are required to perform. Image segmentation tasks are also there in the computer vision tasks which require every small detail to be well-annotated and labeled. And we also know that we can make available the unlabeled data in abundance. The basic idea behind Self-supervised learning is to make a model learn only the important representation of the data from the pool of unlabelled data. In this learning method the models are trained as they can supervise themselves and after supervising they can provide a few labels on the data so that supervised learning tasks can be performed on it. If we are talking about computer vision the supervised learning task can be the simplest image classification task or it can also be semantic segmentation which is a complex task in the computer vision field. In the field of natural language processing, the transformer models such as BERT and T5 are providing a lot of fruitful results. These models are also built on the idea of self-supervised learning where they are already trained with a large amount of unlabelled data and then they apply some fine-tuned supervised learning models with few labeled data. Similarly in the field of computer vision, there are some models which follow the idea of self-supervised learning. In this article we are going to introduce some of them. In computer vision, the basic idea behind self-supervised learning is to create a model which can solve any fundamental computer vision task using the input data or image data and by the time model is solving the problem it can learn from the structure of the objects presented in the image. There can be many self-supervised learning methods but in the case of computer vision, one method named the contrastive method seems to be more successful than the others. Hence, In the next section of this article, we are going to introduce the contrastive learning method. The Contrastive Learning To understand contrastive learning refer to the below image wherein the model using the contrastive learning method has a function f() which takes the input a and gives the output f(a). Let’s say there can be two types of inputs, positive and negative. So if there are two similar inputs (in the below image we are considering a1 and a2c as similar positive input ) to the function f() then their output should be the same and this output should be dissimilar to the output of the opposite input. Above can be considered as the statement of any contrastive learning approach. Positive or similar input can be two sections of the same image or two frames from the same video and the negative or dissimilar input can be a part of different computer vision data or section from the different image. Contrastive Predictive Coding (CPC) The general idea behind the contrastive predictive coding (CPC) is extracting a few upper rows from the coarse grid of the images ad the task is to predict a few lower rows of the image by the time of generating a prediction of the lower rows model need to learn the structural behaviors and the objects of the images. For example, by seeing the face of the cat, the model can predict that the cat has four legs in the lower part of the image. Let us suppose a basic computer vision modeling task is divided into three parts as: Divide the image into the grids such as if the size of image is given 256 x 256, it can divide into 7×7 grids where if the cell size is 64px and 32px it can overlap with each neighbor cell.Encode the grid cell into a vector such that if the size of the given image is similar to the first point the grid cells can be encoded into a 1024 dimension vector so that the whole image can be converted into the 7x7x1024 tensor. An autoregressive generative model can be used to predict the lower rows of the image using the grid cells converted into the tensor of the upper row. For example, the upper 3 rows converted into the 7x7x1024 tensor can be used for the generation or prediction of the last 3 rows. The PixelCNN model can be used for the above process. The below image can be a representation of the above-given steps. Image source To train such a model we are required to introduce a loss function that can calculate the dissimilarity between the patch predictions more formally saying a measure of similarity if between the positive pairs and negative pairs. Mostly the loss function uses the set X of N patches., where the set X can be considered as the set of N-1 samples which are negative and 1 positive sample. This loss function can be calculated by estimating the difference between noise and contrast. And the loss function can also be called the infoNCE function where NSE stands for noise-contrastive estimation. Image source Optimizing this loss will result in a function for estimating the density ratio, which is: Image source The above-given function is very similar to the log softmax function which is often used for calculating the similarity between the prediction and actual values or original values.  In the paper “Representation learning with contrastive predictive coding,” the idea of CPC has been introduced, where we can find a comparison between the accuracy of models using different techniques. Which is shown in the below table. MethodsAccuracyMotion Segmentation 27.6Exemplar31.5Relative Position36.2Colorization 39.6Contrastive Predictive Coding(CPC)48.7 In the above table, we can see the performance of the CPC method for representation learning which is still far from the many of the supervised learning models like ResNet-50 with 100% labels on the Imagenet has 76.5% top-1 accuracy. An update on the CPC is used for increasing the accuracy of the model which can be called Instance Discrimination Methods. The next section of the article is an introduction of the Instance Discrimination Methods Instance Discrimination Methods As we have seen in the CPC, the method was applied to the part of the images. In comparison to the CPC, the Instance discriminative methods have the basic idea to apply the CPC in the whole image. Images with their augmented version can make a positive pair and they should have a similar representation and either image with its augmented version should have a different representation. Image source The main motive of the image augmentation is that if any representation has been learned from the model that should not be varied and the augmented image can be horizontal flip, random crop, different color channel, etc. by putting an augmented image as input can change the image but there class information learned by the model should not be changed. In this method the process of the model can be divided into three basic steps: Give the input to the model as an image with its randomly augmented version to make a positive pair and also feed the model with negative samples with their augmented version.Encode the image pair using any encoder and get the labels or representation on the image and also use the encoder for rep[resentation of the negative samples. Apply InfoCPC to cross-check the similarity level between the positive pair and the dissimilarity level between the positive and negative pairs. There are two papers SimCLR and Momentum Contrast (MoCo) which have wired on the instance discrimination methods and the major difference between them is how they handle the negative samples. We can use their techniques for making self-supervised learning models in the field of computer vision. Final Words In this article, we have seen that in the field of computer vision the self-supervised learning is a representation learning method where we can use the supervised learning models to make the data labeled which can be very helpful in reducing the cost, time, and effort in the labeling of the data. There are various models based on contrastive learning like MoCo and  SimCLR which can be used in computer vision for self-supervised learning.","excerpt":"In the last few years, we have seen that self-supervised learning methods are emerging rapidly. It can also be noticed that models using self-supervised learning methods have solved many of the problems regarding unlabeled data. Uses of these methods in fields like computer vision and natural language processing have shown many great results. In this […]","categories":["Deep Tech"],"tags":["Guide"],"author_name":"Yugesh Verma","publish_date":"2021-10-31T16:00:00","publication_year":"2021","word_count":1513,"keywords":["Go","API","TPU","T5","AI","ML","computer vision","BERT","object detection","R","Guide"],"extracted_tech_keywords":["AI","ML","computer vision","object detection","TPU","R","Go","API","BERT","T5"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-guide-to-self-supervised-learning-in-computer-vision\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10172779,"title":"HCLTech, Equinor Join Forces For AI and AR in Workspace","content":"HCLTech and Equinor announced an expansion of their technology partnership to support the next stage of Equinor’s digital transformation on July 2. The renewed collaboration will see HCLTech manage core areas of Equinor’s IT operations, including cloud migration, cybersecurity, network performance, and user experience. The company will help accelerate Equinor’s shift to cloud systems, improve cyber defences, and enhance workplace efficiency with automation. The agreement also includes the use of new technologies like augmented reality to improve user experiences. “We’re pleased to continue our long-standing collaboration with Equinor,” said Sandeep Kumar Saxena, executive vice president at HCLTech. “This collaboration reflects our shared commitment to innovation and sustainability.” Equinor, Europe’s largest energy supplier and a leader in renewables, has worked with HCLTech for over ten years. Over time, their partnership has moved from basic service management to a broader, long-term alliance focused on digital growth and business goals. HCLTech has previously supported Equinor’s global expansion, infrastructure setup, and efforts to strengthen cybersecurity. Meanwhile, HCLTech has announced a partnership with OpenAI in a multi-year strategic deal to drive enterprise AI. It’s a one-of-a-kind deal in India’s IT landscape, marking a direct partnership with OpenAI. Although firms, including HCLTech, have partnered with startups like Sarvam and Yellow.ai, as well as those from the West, this OpenAI deal has several angles to it. While most IT firms are still accessing OpenAI’s models via Microsoft’s Azure OpenAI Service, HCLTech’s direct collaboration provides it privileged access to the ChatGPT maker’s AI portfolio, positioning it uniquely at the frontlines of enterprise AI adoption.","excerpt":"The agreement also includes the use of new technologies like augmented reality to improve user experiences.","categories":["AI News"],"tags":["hcltech"],"author_name":"Shalini Mondal","publish_date":"2025-07-03T11:00:27","publication_year":"2025","word_count":256,"keywords":["Go","ChatGPT","OpenAI","AI","R","digital transformation","Git","automation","GPT","hcltech","Azure"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Azure","R","Go","Git","GPT","digital transformation","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hcltech-equinor-join-forces-for-ai-and-ar-in-workspace\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":64537,"title":"Snowflake Announces General Availability In AWS Asia Pacific (India) Region","content":"Snowflake has announced the general availability of Snowflake on Amazon Web Services (AWS) in the AWS Asia Pacific (India) Region. This development builds on Snowflake’s efforts to support its India customers that want to keep their data in India and leverage the flexibility and scalability offered by a cloud data platform. The deployment will empower Indian organizations to use the instant elasticity of Snowflake to grow their business by providing quick access to data insights safely and securely. “Cloud adoption rate in India is growing remarkably fast as more organizations are realizing the cost and operational benefits of moving to the cloud,” said Vimal Venkatram, Snowflake’s Country Manager for India. “We have always been a customer-centric company, and by offering Snowflake on AWS in India we are delivering the solutions that Indian companies of every size are requesting, and deserve.” With this deployment, customers can now reap all the benefits Snowflake has to offer, such as: Rapid data analyticsAccess to data-driven insights for all the company’s business users Near-zero management – Snowflake eliminates the administrative and management demands of traditional data warehouses and big data platforms “As organizations and SMBs become more data-driven in their approach, the cloud platform offers them elasticity and flexibility. We are very excited to deploy Snowflake on Amazon Web Services in India. Through this collaboration, customers will benefit from Snowflake’s concurrency and flexibility along with AWS’s trusted cloud capabilities,” added Snowflake Managing Director for South Asia, Geoff Soon.","excerpt":"Snowflake has announced the general availability of Snowflake on Amazon Web Services (AWS) in the AWS Asia Pacific (India) Region. This development builds on Snowflake’s efforts to support its India customers that want to keep their data in India and leverage the flexibility and scalability offered by a cloud data platform. The deployment will empower […]","categories":["AI News"],"tags":["AWS","AWS services","Snowflake","snowflake cloud data platform"],"author_name":"Vishal Chawla","publish_date":"2020-05-05T15:58:45","publication_year":"2020","word_count":243,"keywords":["big data","API","AWS services","AWS","AI","Scala","RAG","snowflake cloud data platform","analytics","Rust","R","Snowflake"],"extracted_tech_keywords":["AI","analytics","RAG","AWS","Snowflake","R","Rust","Scala","API","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/snowflake-announces-general-availability-in-aws-asia-pacific-india-region\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10063199,"title":"The history of machine learning algorithms","content":"Machine learning algorithms can perform exponential tasks today—from mastering board games and identifying faces to automating daily tasks and making predictive decisions—this decade has brought forward countless algorithmic breakthroughs and several controversies. But one would find it a challenge to believe this development started only less than a century ago with Walter Pitts and Warren McCulloch. Analytics India Magazine takes you through a historical story of machine learning algorithms. The mid-1900s Machine learning was ideated first in 1943 by logician Walter Pitts and neuroscientist Warren McCulloch, who published a mathematical paper mapping the decision-making process in human cognition and neural networks. The paper recognised every neuron in the brain as a simple digital processor and the brain as a whole computing machine. Later, mathematician and computer scientist Alan Turing introduced the Turing test in 1950. The three-person game identifying machines as ‘intelligent’ is still unmastered by any machine in 2022. The Turing test demands a computer to fool a human into thinking the machine is also a human being. The 1950s was when pioneering machine learning research was conducted using simple algorithms. In 1952, Arthur Samuel at IBM wrote the first computer program that played a game of checkers. The game was written on top of the alpha-beta pruning algorithm, a search algorithm that decreases the number of nodes evaluated by the minimax algorithm in search trees. This has since been used for two-player games. The algorithm improved over more games by learning from its winning strategies. In 1957, American psychologist Frank Rosenblatt designed the perceptron, the first neural network stimulating the thought processes of the human brain. The discovery is relevant to date. The nearest neighbour algorithm was introduced in 1967, one of the foremost algorithms that solved the ‘travelling salesman problem’, a common problem statement of a salesman who starts at a random city and visits the neighbouring cities repeatedly until all have been visited. The late 1900s Backpropagation concepts were initially introduced in the 1960s and re-introduced in the 1980s to find hidden layers between the input and output layers of the neural networks, making them appropriate for commercial usage. In 1981, Gerald Dejong discovered Explanation Based Learning, based on a computer algorithm, including explanation and generalisation of data. The NetTalk neural network was written in 1985 by Terry Sejnowski, an algorithm capable of pronouncing words like a baby, based on text and matching phonetic transcript as input. In 1989, Christopher Watkins developed a Q-learning algorithm that improved the practical applications of reinforcement learning. The 1990s popularised forward statistical methods for algorithms, given that neural networks seemed less explainable and demanded higher computational power. These methods included support vector machines and random forest algorithms introduced in 1995. Following this, one of the biggest AI wins was in 1997 when IBM’s Deep Blue beat the world champion at chess. The 2000s The early 2000s popularised support vector clustering, unsupervised learning and kernel methods for machine learning. The introduction of convolutional neural networks in 2009 was a major breakthrough. Fei-Fei Li, a computer science professor at Stanford University, created a large dataset reflecting the real world in 2009, which became the foundation for Alex Krizhevsky, who created the first CNN, AlexNet, in 2012. Meanwhile, in 2011, IBM’s Watson beat its human competitors in Jeopardy, and Google Brain introduced its machine that can categorise objects like a cat. The company followed it up with its algorithm to browse YouTube videos and its unlabeled images to identify cats based in 2012. Word2vec algorithms introduced in 2013 used neural networks to learn word associations and later became the foundations for large language models. In 2014, Facebook developed DeepFace, an algorithm that created history by beating all previous benchmarks of computer algorithms recognising human faces. 2014 also saw the creation of generative adversarial networks (GANs) by Ian Goodfellow. Today, algorithmic architecture is the backbone of image, video, and voice generation, popularly used for deepfakes. In 2016, one of the most popular machine learning victories was marked by Deepmind’s AlphaGo algorithm, which beat the world champion in the Chinese board game, Go. In 2017, AlphaGo and its successors beat several champions in Go, Chess and Shogi. In 2017, Waymo started testing its autonomous minivans. Deepmind had another victory in 2018 with AlphaFold for its ability to predict protein structure. 2020 and beyond The decade since 2015 has seen some of the most victorious algorithms since date while birthing major questions and concerns regarding their usage, safety and explainability. In 2020, Facebook AI Research introduced the Recursive Belief-based Learning, or ReBeL, a general RL+Search algorithm with the capacity to work in all two-player and zero-sum games—even those with imperfect information. Deepmind’s Player of Games introduced in 2021 can similarly play perfect and imperfect games. Deepmind also introduced the Efficient Non-Convex Reformulations, a verification algorithm in 2020. It is a novel non-convex reformulation of convex relaxations of neural network verification. AlphaFold 2 in 2021 achieved a level of accuracy much higher than any other group. The years have also created the largest transformer and language models such as GPT-3, Gopher, Jurassic-1, GLaM, MT-NLG and more, founded on NLP algorithms. Google also released Switch Transformers, a technique based on the modified MoE algorithm, Switch Routing, to train language models with over a trillion parameters.","excerpt":"In 1957, American psychologist Frank Rosenblatt designed the perceptron, the first neural network stimulating the thought processes of the human brain.","categories":["IT Services"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-03-22T11:00:00","publication_year":"2022","word_count":877,"keywords":["Go","machine learning","TPU","AI","neural network","Transformers","Git","NLP","analytics","R"],"extracted_tech_keywords":["AI","machine learning","neural network","NLP","analytics","Transformers","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-history-of-machine-learning-algorithms\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":57789,"title":"Study: Gender Diversity In Analytics &#8211; 2020","content":"Women are breaking the glass ceiling across industries and enterprises, rising to the top echelons of company departments and management. Nevertheless, the participation of women in the technology sector, and more specifically across the data science domain, is still significantly constrained by low participation fueled by possible hiring stereotypes and mindsets around women in science functions v\/s women in art functions. The greater adoption and scaling of data science services, such as predictive analytics, artificial intelligence (AI), and machine learning (ML), across all industries from Industrials to Media, enterprises ranging from large to small, and organization functions covering HR and Marketing, renders the representation of women across the data science domains all the more significant as the representation impacts not a single, isolated enterprise function but the entire organizational ecosystem across industries. Overview The broader data science domain has experienced significant growth in India over in the past couple of years. In 2019, the industry was valued at $3.03 Bn, and, considering that the domain will double by 2025, the analytics space will experience 12.2% CAGR. Organizations are bringing forth capabilities not just in analytics as a service but also in specialized data science domains, including Computer Vision, Machine Learning (ML), and Natural Language Processing. Moreover, Indian AI startups received funding of $762 mn from global investors in 2019, reflecting the strength and growth of capabilities across Analytics, AI, and ML. Nonetheless, the role of Women in AI and Analytics needs to be studied and considered, to gauge the pace of change for Women vis-à-vis the broader growth across the Data Science domains. Key Highlights As of February 2020, women represent 27.8% of the AI \/ Analytics workforce in India. As mentioned, this workforce delivers all AI and Analytics services originating from India regardless of the stakeholder or client type, type of firm providing the service, and location of client\/stakeholder. Women in AI in India constitute 8.1% of the global women in AI workforce. • Close to 79% of the women AI workforce is concentrated in organizations with personnel greater than 10000 personnel. • The median salary of women in AI is INR 10.7 Lakhs while the median experience is 5.8 years. • The average women AI workforce across organizations is 160 personnel. Women in AI in India This section covers the findings of the demographics and experience levels for women in AI. Work Experience: The median work experience of women analytics professionals in India is 5.8 years. This is much lower than the median experience for the entire AI workforce in India at 7.4 years. This difference exists because a larger proportion of women have recently started taking up jobs in this domain – recent women graduates and post-graduates have taken up a greater number of positions in the new Data Science domains. This recent entry has made the pool of women professionals in analytics much younger. The women analytics professionals with work experience of fewer than 5 years make up the majority of the workforce – 53%, with the greatest concentration of professionals in the 2-5 years’ experience bracket at 37%. 13% of women analytics professionals have more than 10 years of work experience. This total experience representation may not necessarily entirely be in analytics as these professionals may have transitioned to this space over a period of time. WOMEN IN AI ACROSS COMPANIES IN INDIA Percentage of Women AI Personnel by Company Type (excluding Boutique AI firms): Domestic IT (45.6%), MNC IT (25.5%), Sector-specific firms (18.1%) and Consulting firm (10.5%) make up the largest proportion of women analytics personnel. These IT firms include enterprises, such as TCS, Wipro, HCL Tech, Infosys, IBM, Cognizant, and Accenture Sector-specific firms include: Consumer-focused firms, which process millions of volumes of customers’ data and data sets each day, understandably constitute a significant share of the Women AI employee personnel. These include Consumer & Healthcare firms at 5.2% Organizations that fall under the manufacturing-focused Automotive, Industrials, and Energy industries, include Maruti Suzuki, Hyundai, Tata Motors, Tata Steel, Vedanta, HPCL, IOCL, and BPCL, to name a few. These firms – driven by broad-based data including macro-economic data, commodity price fluctuations, OEM \/ supplier data, and large sales volumes and revenues – make up 4.8% of the women AI employee share. Banking and Financial firms, both domestic – State Bank of India, HDFC Bank, ICICI Bank, and multinational – JP Morgan, American Express, and Blackrock, make up 4.5% of women AI employee base The remaining sector-specific firms include Semiconductor, Network, Telecom & Media, that makeup 3.6% of the women AI employee proportion. SALARIES FOR WOMEN IN AI IN INDIA Women in AI are representing an increasing share of the larger AI employee base. While their representation increases across Data Science domains, the salaries earned by women personnel are still significantly lower than the median salaries of the wider AI employee workforce across locations and experience levels. The median salary in India of women AI personnel is INR 10.7 Lakhs across all experience levels, locations, and skillsets. This figure is 26% lower than the median salary for all AI personnel. Women in the salary bracket of INR 3-6 Lakhs make up the highest concentration of women AI personnel – at 30.9%. Almost 43% of women professionals in AI command a salary of less than INR 6 Lakhs. Approximately 17% of the women professionals draw a salary greater than INR 15 Lakhs. Mumbai emerges as the destination for highest salaries for Women in Analytics at 11.9 Lakhs per annum as median salary, followed by Bengaluru at 10.9 Lakhs. • The sector-location niche of Mumbai and the higher cost of living in the city has pushed up the salaries of women AI personnel to the highest level across cities, while Bengaluru has come in 2nd – signifying the city’s strength as a hub of data science domains for women employees as well Similarly, the BPO\/KPO hub of Delhi NCR, registers the median salaries for Women AI personnel at the 3rd place at 10.5 Lakhs Companies based in Chennai offer the lowest salaries at INR 8.5 Lakhs. Although Chennai does figure numerous analytics start-up enterprises with women employees, the low cost of living in the city, a key factor in salaries offered, has brought down the median salary of Chennai. Download the complete report here.","excerpt":"Women are breaking the glass ceiling across industries and enterprises, rising to the top echelons of company departments and management. Nevertheless, the participation of women in the technology sector, and more specifically across the data science domain, is still significantly constrained by low participation fueled by possible hiring stereotypes and mindsets around women in science […]","categories":["AI Features"],"tags":["automotive analytics","automotive data analytics","enterprise analytic hub","hpc data management system","mba in business analytics"],"author_name":"Дарья","publish_date":"2020-03-02T14:00:00","publication_year":"2020","word_count":1046,"keywords":["data science","hpc data management system","artificial intelligence","machine learning","automotive data analytics","AI","R","ML","mba in business analytics","computer vision","RAG","automotive analytics","analytics","predictive analytics","enterprise analytic hub"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","data science","analytics","RAG","predictive analytics","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/study-gender-diversity-in-analytics-2020\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":36014,"title":"How MachineHack Users Solved The Uber City Traffic Visualization Challenge","content":"Analytics India Magazine’s hackathon platform MachineHack recently concluded its hackathon titled “Visualisation is Beautiful: Uber City Traffic Visualization Challenge”. We talked to the winners of the hackathon to know about them and to know how they solved this month’s mind-bender. Surenther Has Rank 1 On The Leaderboard Surenther is into data science since 2016. He was in the data science and analytics centre of Karpagam College of Engineering, Coimbatore. He says that data visualisation is the heart of data science as it gives solutions as the visualisation tools like Tableau is making easy for him. Tableau made him understand the keynotes of visualisation and he shares his skills in visualization that he acquired, with his students. Approach to solving the problem: After having a look at the data, Surenther realised that he needs more dimensions to make charts, so he carefully handled the data carefully making sure that the values do not change. He then moved to visualisation of the charts, data repentance time series to make an easy forecast and said that the single dashboards gave him more information. Experience on MachineHack: Talking about his experience on the platform, he says that it is a good platform to get started in data science hackathons. He thanked MachineHack for the opportunity to participate in the hackathon. Sahil Mendhiratta Has Rank 2 On The Leaderboard: Mendhiratta says that a data science career involves a mix of skills around data handling, modelling and presentation skills. He started his career towards data science in 2018 when he was introduced to it in his very first organisational project. He developed a deep interest in it and started learning it at a good pace. He read various blogs on data science and data analytics and realised that it is a hot technology that will create a boom in the market in the coming years. Courses on EDX, DataCamp, Coursera helped him develop his foundation. He believes that data science will definitely bring a drastic change in every industry to solve real-world problems. Approach to solving the problem: According to Mehdhiratta, the data provided was the kind a BI developer actually wants, in a numerical format. Although he had to do some transformations and data cleansing operations on the data. The major dimension to show on dashboards was date and time and so he focused primarily on showing time span at various hierarchies of date and time, like a year, day and month level.  He also focused on analysing traffic at different point of times. He used the platform of Microsoft PowerBI because it would give highly customised and beautiful charts giving deeper insights view. Experience on MachineHack: It was a great learning experience with MachineHack and it helped him grow, enrich knowledge and develop confidence in data science. He said that he would look forward to participating in the upcoming hackathons of MachineHack. Sreyan Ghosh Has Rank 3 On The Leaderboard Ghosh is a third-year undergraduate student from Christ University, Bangalore from the department of computer science and engineering. His data science journey began in his second year of engineering when he attended an introductory machine learning workshop organised by one of his seniors in college. Ghosh said that he went back home the same day and signed up for Andrew Ng’s machine learning course on Coursera and he continues with various such courses even after. Ghosh said that being a student, he did take some time and determination to learn the statistics and mathematics side of data science but he never gave up. He also received a lot of encouragement from his teachers in his department. He also co-founded a club in his college called Neuron which has been growing exponentially since its inception. The first ever competition that he took part in was on Kaggle where the job was to predict poverty levels in Costa Rica. Since then I kept taking part in several online and offline competitions, which helped him to stay in touch with all new innovations in the data science world and I learnt a lot more quickly than just taking up courses online. One thing that he says has learnt so far in his data science journey is that data science is not just about making complicated models, it’s about taking better decisions. Approach to solving the problem: Talking about how he solved the hackathon problem, Sreyan said that he was the last one to submit the solution. He initially saw the leaderboard filled with visualisations from Power BI and he knew that he needed something different to stand out. So he used Plotly tool of Python. The best thing about Plotly, he says, is the customisation that it allows. Custom functions can be written that can be used to present data in exactly the way you want it to be. Moreover, the vibrant colours that can be used to catch attention immediately. Initially, after cleaning the dates column which had dates in different formats, he had started with the visualization part. The most important thing was to choose wisely between all the columns in the dataset, the columns which could be used to portray the most out of the data. He primarily used the ‘mean’ features rather than the ‘upperbound’ and ‘lowerbound’ features for making his visualisations. Some of his key visualisations included a violin plot of the mean travel time across days of a month, days of the week, boxplot of mean travel time across years and across months of a year and distribution plots of mean, upper-bound and lower-bound travel times across different parts of a day. The whole process of selecting the best visualisations to put up on MachineHack Ghosh said was quite iterative. He also took the help of my friends and asked them to review it before he could select the 6 best visualisations. In the process, he discarded about 50 visualisations before he came up with my 6 best visualisations on the eve of the submission deadline. Experience on MachineHack: Ghosh said that I was a great learning experience at MachineHack and competing with top contenders, especially professional data scientists can make students and aspiring data scientists grow and See where they stand in terms of industry standards. Machine Hack provides an amazing competitive environment for professionals and data science enthusiasts. Ghosh also said that he will keep taking part in hackathons on MachineHack and is looking forward to more hackathons. He also thanks the HOD of his department, Dr Balachandran K and his teacher who supported him in his data science journey. He also thanked his friends who helped spread the word and all those who voted for him. MachineHack recently concluded one more hackathon called “Predict A Doctor’s Consultation Fee Hackathon” and we published the winners’ article on how they solved the hackathon problem at Analytics India Magazine. MachineHack hosts many interesting hackathons and also has interesting prices for the winners. It has recently launched a new hackathon called “Predict The Flight Ticket Price Hackathon” to solve the unpredictability of flight ticket prices.","excerpt":"Analytics India Magazine’s hackathon platform MachineHack recently concluded its hackathon titled “Visualisation is Beautiful: Uber City Traffic Visualization Challenge”. We talked to the winners of the hackathon to know about them and to know how they solved this month’s mind-bender. Surenther Has Rank 1 On The Leaderboard Surenther is into data science since 2016. […]","categories":["Deep Tech"],"tags":["Hackathon","plotly","Python"],"author_name":"Disha Misal","publish_date":"2019-03-08T09:45:55","publication_year":"2019","word_count":1173,"keywords":["data science","Go","machine learning","Plotly","AI","plotly","Hackathon","Python","Ray","RAG","analytics","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Ray","Plotly","RAG","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-machinehack-users-solved-the-uber-city-traffic-visualization-challenge\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101281,"title":"5 Must-Know General Purpose Robots","content":"Earlier this week, Google DeepMind released resources for general purpose robots and took it a step further. This is touted to be the ImageNet moment for robotics, where it won’t be necessary to individually train robotic models to each specific task as it explores knowledge transfer between robots. In line with this, let’s take a look at other models and robots for general purpose tasks and operations. RT-X RT-X, or the Robotics Transformer X, is a general-purpose robotics model developed by Google DeepMind to enhance robotics research and performance across a variety of robots and tasks. The key goal was to create a model that can generalize and transfer its skills to various robots and tasks. It can be used for actions, vision, and language understanding, making it a comprehensive tool for general robotics research. RT-X is built upon the foundation of two previous robotics transformer models: RT-1 and RT-2. RT-1 is a model developed for real-world robotic control at scale, and RT-2 is a vision-language-action (VLA) model that learns from both web and robotics data. By incorporating the architectures and knowledge from these models, RT-X achieves higher performance, especially due to the extensive and diverse cross-embodiment data it is trained on. The development of RT-X involves training the model on a vast and diverse dataset, known as the Open X-Embodiment dataset. This dataset is a collection of data from 22 different robot embodiments, encompassing a wide range of skills and tasks, a total of 150,000 tasks in more than 500 distinct skills according to the paper. RoboCat RoboCat developed by Google DeepMind is an AI agent in robotics that can learn various tasks across different robotic systems. What sets RoboCat apart is its ability to not only master multiple tasks but also generate its own training data to enhance its skills. With access to a broad and varied dataset, it can quickly grasp a new task in as little as 100 demonstrations. This rapid learning potential is a promising advancement in robotics research, diminishing the need for extensive human-guided training. RoboCat is a product of merging Google’s Gato multimodal model, proficient in processing language, images, and actions in both simulated and real-world contexts. Gato’s architecture was integrated with an extensive training dataset comprising image-action sequences from diverse robot arms engaged in a myriad of tasks. Dactyl Dactyl, built by OpenAI, was trained in a computer simulation within, with its learnings then applied in the real world, despite some discrepancies in simulation accuracy. The specialised Shadow Dexterous Hand enables Dactyl to manipulate objects like blocks or prisms, employing a learning approach akin to that used for OpenAI Five. Dactyl specifically addresses the task of altering an object’s position while held by the robotic hand, such as rotating a block. For precise grips, like the Tip Pinch grasp, Dactyl cleverly uses its thumb and little finger, which is quite like how humans use their thumb along with either the index or middle finger. However, Dactyl’s little finger on the robot hand is more flexible due to an extra way it can move. That’s why Dactyl often opts for this finger. Essentially, Dactyl can independently discover grips like humans do but adjusts them to suit its own hand’s capabilities and limitations. Jaco Arm JACO arm built by Kinova Robotics is a lightweight assistive robot designed to compensate for lost arm movements. It consists of six linked segments, including a three-fingered hand. Users control hand movements in 3D space and object grasping\/releasing with an adaptable controller, utilizing two or three fingers. The arm can be mounted on a wheelchair, workstation, or table, fitting under the wheelchair’s armrest without widening it. JACO enhances daily activities for individuals with physical disabilities without hindering wheelchair mobility. Key components include actuators driven by DC brushless motors with Harmonic Drive tech and various sensors for precise control. Grippers are designed to grasp diverse daily objects, with underactuated grippers simplifying control and adapting to object shapes. The interface options include a Software Development Kit (SDK), a ROS package, and a joystick, providing users with flexible control over the robot’s movements. Panda Arm Panda Arm was built by Franka Emika, a German robotics platform company. The robot has a range of features and is designed with sensors to stop moving when there’s unexpected interference to prevent accidents. It has a special hand that can hold things tightly, like exerting a strong grip. This hand can hold up to 3 kg of weight. This allows the robot to pick up various objects almost like a human arm. The arm is used in production lines and its uses are non-exhaustive. It is being used in nursing homes, laboratories and universities, logistics platforms, etc.","excerpt":"DeepMind unveiled versatile robot resources, a potential ImageNet moment in robotics. This development hints at knowledge transfer, reducing the need for task-specific model training","categories":["AI Trends"],"tags":[],"author_name":"K L Krithika","publish_date":"2023-10-09T15:08:56","publication_year":"2023","word_count":781,"keywords":["Go","API","OpenAI","AI","Modal","programming_languages:R","programming_languages:Go","ViT","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","OpenAI","R","Go","API","ViT","Modal","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-must-know-general-purpose-robots\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168550,"title":"L&amp;T Tech Services Q4 Revenue Rises 17.5% YoY to ₹2,982 Crore","content":"L&T Technology Services (LTTS) on Thursday reported a 9% year-on-year decline in net profit for the fourth quarter of FY25, falling to ₹311 crore from ₹341 crore in the same quarter last year. Despite the decline in profit, the company reported a robust operational performance, with revenue increasing by over 17.5% to ₹2,982.4 crore during the quarter. The company stated that the dividend will be disbursed within 30 days following the necessary approvals, though the record date for shareholder eligibility is yet to be finalised. The IT engineering firm reported an EBIT margin of 13.2%, while its EBITDA margin improved to 19.8%. Revenue in dollar terms also increased by 13.1% year-over-year, reaching $345.1 million. Employee strength stood at 24,258 at the end of the quarter, according to the company’s press release. Commenting on the Q4 performance, Amit Chadha, CEO and MD of LTTS, said, “The large deal pipeline has been robust on the back of value enhancement across the clients’ product lifecycle and digital transformation journey.” Looking ahead, Chadha expressed optimism for the coming fiscal year. “Based on the large deal bookings closed during the quarter, in our view, FY26 will be a better year than FY25. We also reaffirm our medium-term outlook of USD 2 billion revenue,” he said, while adding that AI and automation will provide new solutions. Meanwhile, LTIMindtree, the consulting arm of L&T, reported positive Q4 FY25 results with modest growth. With a 2.6% YoY increase in consolidated net profit for the fourth quarter of FY25, reaching ₹1,128.5 crore, up from ₹1,100.7 crore last fiscal.","excerpt":"‘In our view, FY26 will be a better year than FY25. We also reaffirm our medium-term outlook of USD 2 billion revenue.’","categories":["AI News"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-04-24T18:13:32","publication_year":"2025","word_count":259,"keywords":["programming_languages:R","AI","digital transformation","Git","automation","R"],"extracted_tech_keywords":["AI","R","Git","digital transformation","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/lt-tech-services-q4-revenue-rises-17-5-yoy-to-%e2%82%b92982-crore\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168481,"title":"This Indian AI Startup Proves LLMs No Longer Need Expensive GPUs","content":"The future of running LLMs may no longer rely on expensive infrastructure or GPUs. While India works on developing its own foundational model under the IndiaAI mission, a startup is taking a different approach by exploring how to efficiently run LLMs on CPUs. Founded on the principle of making AI accessible to all, Ziroh Labs has developed a platform called Kompact AI that enables the running of sophisticated LLMs on widely available CPUs, eliminating the need for costly and often scarce GPUs for inference—and soon, for fine-tuning models with up to 50 billion parameters. “With a 50 billion-parameter model, no GPU will be necessary during fine-tuning or inference,” said Hrishikesh Dewan, co-founder of Ziroh Labs, in an exclusive interview with AIM. He further added that work on fine-tuning capabilities is already underway and will be released in the next three months, claiming that nobody will need GPUs to train their models anymore. Ziroh Labs has also partnered with IIT Madras and the IITM Pravartak Technologies Foundation to launch the Centre for AI Research (CoAIR) to solve India’s compute accessibility challenges using AI models optimised for CPUs and edge devices. Ziroh Labs is based in California, US, and Bengaluru, India. Dewan shared that Kompact AI has been entirely developed in the Bengaluru office — from the core science and engineering to every aspect of its design and execution. The company has already optimised 17 AI models, including DeepSeek, Qwen and Llama, to run efficiently on CPUs. These models have been benchmarked with IIT Madras, evaluating both quantitative performance and qualitative accuracy. The Tech of Kompact AI Dewan explained that LLMs are nothing but mathematical equations that can be run on both GPUs and CPUs. He said that they don’t use the technique of distillation and quantisation, which is quite common today. Instead, Ziroh Labs analyses the mathematical foundations (linear algebra and probability equations) of LLMs and optimises these at a theoretical level without altering the model’s structure or reducing its parameter size. After theoretical optimisation, the model is tuned specifically for the processor it will run on, taking into account the CPU, motherboard, memory architecture (like L1\/L2\/L3 caches), and interconnects. Dewan argued that running an LLM on a CPU is not novel—the real challenge is maintaining quality and achieving usable speed (throughput). He explained that anything computable can run on any computer, but the practicality lies in how fast and how accurately it runs. They have been able to solve both these aspects without compressing the models. “What is essential, then, that needs to be solved is twofold. One is to produce the desired level of outcome, that is, quality. And two is how fast it will generate the output. So these are the two problems that need to be solved. If you can solve these together, the system becomes usable,” Dewan said. Partnership with IIT Madras Dewan shared that the partnership with IIT Madras came about through Professor S Sadagopan, the former director of IIIT-Bangalore, who introduced him to Professor V Kamakoti, the current director of IIT Madras. At the launch event, Sadagopan said, “India too is developing GPUs, but it will take time. Ziroh Labs demonstrates that AI solutions can be developed using CPUs that are available in plenty, without the forced need for a GPU, at a fraction of the cost.” Dewan added that their collaboration with IIT Madras has a dual purpose—ongoing model validation and the development of real-world use cases. “The idea is to make these LLMs available to startups so that an ecosystem can be built,” he said. Kamakoti said the initiative reflects a nature-inspired approach. “Nature has taught us that one can effectively acquire knowledge and subsequently infer in only a limited set of domains. Attempts to acquire everything under the universe are not sustainable and bound to fail over a period of time.” “This effort is certainly a major step in arresting the possible AI divide between one who can afford the modern hyperscalar systems and one who cannot,” he added. Dewan discussed the diverse range of use cases that have emerged since the launch of Kompact AI. “We’ve received over 200 requests across various segments, including healthcare, remote telemetry, and even solutions for kirana stores,” he said. “People are also working on creating education software tools and automation systems. Numerous innovative use cases are coming from different industries,” Dewan added. Take on Big Investments in AI Microsoft has announced plans to spend $80 billion on building AI data centres, while Meta and Google have committed $65 billion and $75 billion, respectively. When asked whether such massive investments are justified, Dewan pointed to the scale of the models these companies are developing. “They’re designing huge models… their thesis is that large models will do a lot of things,” he said. While $50 billion may seem like a vast sum of money, Dewan noted that in the world of large language models, it’s relatively modest, citing Grok, which has over a trillion parameters, as an example. He added, “They have the money, so they’re doing it. And we have the tech, and we can solve our problems. So everybody will coexist.” Ziroh Labs currently has a team of eleven people and is bootstrapped. The company was founded in 2016 to address the critical problem of data privacy and security, specifically focusing on developing privacy-preserving cryptographic systems that could be used at scale. Dewan said they are still working on this. “We will bring privacy to AI in 2026, because ultimately, AI must have privacy.” As AI becomes increasingly accessible, Cypher 2025, India’s largest AI summit, emerges as a powerful convergence point for innovation and collaboration. It is scheduled to take place from September 17–19 at the KTPO Trade Centre in Whitefield, Bengaluru. In its ninth edition, the event brings together over 5,000 daily attendees, 100+ speakers, and a wave of emerging tech voices. Organised by AIM Media House, Cypher 2025 serves as a pivotal platform for professionals, startups, and enterprises to engage with the evolving AI landscape in India. [Note: The headline has been updated for clarity.]","excerpt":"“With a 50 billion-parameter model, no GPU will be necessary during fine-tuning or inference.”","categories":["AI Startups"],"tags":["GPUs"],"author_name":"Siddharth Jindal","publish_date":"2025-04-23T18:13:30","publication_year":"2025","word_count":1013,"keywords":["Go","TPU","AI","innovation","Scala","automation","Aim","GAN","GPUs","R","startup"],"extracted_tech_keywords":["AI","Aim","TPU","R","Go","Scala","GAN","automation","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/this-indian-ai-startup-proves-llms-no-longer-need-expensive-gpus\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10053970,"title":"NVIDIA GTC 2021: What We Predicted Vs What Actually Happened","content":"Tech enthusiasts across the globe wait all year long for American tech giant NVIDIA’s annual GPU Technology Conference or GTC. The global AI conference brings together developers, inventors, engineers, researchers and IT professionals to present new developments and share ideas in the fields of data science, artificial intelligence (AI), machine learning (ML), computer graphics and autonomous vehicles. Hosted between November 8 and 11, GTC 2021 had over 500 sessions scheduled, delivered by academicians, researchers and thought leaders. Before the beginning of the conference, Analytics India Magazine published an article on what to expect at the GTC 2021. Today, we make a comparison of what we predicted versus what was actually announced during the conference. On Omniverse Starting earlier this year, NVIDIA has made its Omniverse the talk of the town. Omniverse is a platform that connects 3D worlds together into a virtual universe. In August, NVIDIA revealed that its CEO, Jensen Huang, pulled off an Omniverse-stunt without any glitch. For a good 14 seconds, Hunag’s virtual replica replaced him during a keynote speech. Predictions It was thus obvious that Omniverse– one of the first simulation and collaboration platforms delivering the foundation of the metaverse– will be a topic of discussion during the GTC 2021. With NVIDIA planning to expand the Omniverse platform with its partnership with Adobe and Blender, we predicted that NVIDIA would make major announcements with respect to scaling the platform across industries. We also predicted that NVIDIA would reveal its partnership with companies developing solutions to provide an immersive experience in its Omniverse. Announcement During the keynote speech, Huang announced that with the help of Omniverse, users will now be able to create new 3D models of their physical world. Furthermore, with Omniverse Avatar— a tech platform to generate interactive AI avatars– developers will now be able to not just create characters that can see but also understand and interact. Latest updates on Project Maxine– NVIDIA’s GPU-accelerated SDK, will equip Omniverse Avatar to connect the Rive speech AI, computer vision and animation to conversational AI robot in real-time. It will also add audio and video features to content creation applications and virtual collaborations. GTC 2021 also unveiled a synthetic data generation engine– NVIDIA Omniverse Replicator, to train deep neural networks. Replicators are of two kinds– Omniverse Replicator for Isaac Sim, Omniverse Replicator for DRIVE Sim, meant for autonomous vehicles. The Isaac robotics platform can easily be integrated into Robotic Operating System (ROS) and will further equip partner companies. On AI tools Predictions We predicted that at GTC 2021, NVIDIA would most likely announce deep learning tools to build multimodal conversational AI applications to deliver real-time performance on GPUs. Announcement At the conference, Huang unveiled Nemo Megatron– reportedly the biggest HPC application ever used to train large language models. Additionally, he also announced NVIDIA Modulus to build and train ML models to obey the laws of physics. NVIDIA has also introduced three new libraries – ReOpt, cuQuantum and cuNumeric. On Autonomous Vehicles Predictions Analytics India Magazine predicted that NVIDIA will make announcements related to autonomous driving systems, give updates on the latest DRIVE, and reveal partnerships with companies unveiling the future of mobility. Announcement Much like the predictions, at the GTC 2021, Huang announced that NVIDIA DRIVE will now have Hyperion 8. Its latest hardware and software architecture suite will include nine radars, 12 ultrasonic sensors, 12 cameras and a front-facing LIDAR, along with NVIDIA Orin SoCs. During his keynote, CEO Huang also said that the team at NVIDIA was thrilled by the growth of the ecosystem that they are building and that the company shall continue to put their heart and soul into advancing it. The company is focused on building tools for ‘the Da Vincis’ of our times, and in doing so, it will be helping create the future.","excerpt":"Before the beginning of NVIDIA GTC 2021, AIM made predictions regarding the expected announcements at the conference.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Autonomous Vehicles","Data Science","GTC","Jensen Huang","Machine Learning","Metaverse","NVIDIA","Omniverse","Robotics"],"author_name":"Debolina Biswas","publish_date":"2021-11-22T12:20:13","publication_year":"2021","word_count":633,"keywords":["Robotics","computer vision","deep learning","GTC","Autonomous Vehicles","data science","artificial intelligence","Metaverse","Omniverse","analytics","NVIDIA","Data Science","machine learning","AWS","AI","neural network","ML","Machine Learning","Jensen Huang","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","computer vision","data science","analytics","AWS"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidia-gtc-2021-what-we-predicted-vs-what-actually-happened\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10003724,"title":"Low-Code Development Lowers The Barrier To Entry — The Need Of The Hour, Says Venkatesh Ramarathinam, CEO, Vuram","content":"In the current dynamic landscape, rolling out new applications and developing innovative software is the need of the hour, however, demands high-performance quality engineering to ensure utmost quality and accuracy. However, conventional methods come with several challenges, poor quality leading to hampering the outcome. Thus, automation testing is the proven way that can optimise the software development cycle, but the majority of automation testing tools require extensive codes to run the tests making it challenging for many. In order to simplify this process, Vuram, a hyper-automation services company, launched an automation testing tool — VATT (Vuram Automation Testing Tool) built using an intelligent hybrid framework to achieve complete automation for testing applications. Such a solution not only bridged the loopholes of the testing phase of the software development but also provided speed and accuracy that’s is critical for businesses to sustain amid this crisis. To understand how such low code platforms can become the future of application development, Analytics India Magazine spoke to Venkatesh Ramarathinam, the CEO of Vuram. Edited excerpts — Can you tell us about the recently launched Automation Testing tool? Also, how does it stand out in the market? We have launched Vuram Automation Testing Tool or VATT, where instead of writing several lines of codes for testing an Appian application, a single line of code is all it takes to create and run tests. The intelligent hybrid framework that the tool is built on allows for accurate, end-to-end functional testing. Since it’s designed to achieve total automation, it can save significant execution time during the quality testing phase of software development. The tool comes with a great range of predefined functions and several reusable scenarios that replace complex code with a single line of code. This allows users to create and run tests efficiently for Appian applications. Firstly, the users have to download our tool package and import into the eclipse IDE. They just need to call the inbuilt functions available in VATT. These inbuilt functions will help the user to interact with the Appian components and automate easily without writing dozens of codes. The tool also maintains test data dynamically in Excel sheets. So, the tests can be run with different data sets without changing the script. Besides, VATT also ensures security levels according to the user roles. After executing the test results, the results will be auto emailed. This minimises status checks on the part of users. Another essential feature is that it has visually-rich reporting capabilities to assess test results. This framework is designed using the Java language, and with that, we have built a lot of inbuilt functions which helps the quality engineer to achieve the automation testing easily. Also, we have developed the logic, which will establish an automatic connection between the VATT and Excel sheet. How do you see low-code platforms becoming the future of application development? In recent years, we have seen tremendous growth in the segment of low-code development platforms. Several new products are being introduced that offer a wide range of features to the enterprises. Enterprise-grade platforms such as Appian continue to innovate and dominate the market with their low-code capabilities. These platforms offer a significant reduction in the time to market, the cost to deliver and the total cost of ownership. All the applications that are built on low-code platforms allow their customers to tweak the solution according to their needs and fit the technology to their business model, instead of trying to meet their practices around the technology. One of the most significant benefits of taking a low-code approach to application development is speed. With a low-code platform, business units can prototype applications in a few hours, rather than weeks or months. This facility will allow the IT companies to gain valuable feedback from their end-users and to adjust their solutions to meet the evolving and mission-critical business requirements. It enables a faster and more streamlined process of development, positioning IT as an adaptive and responsive business partner for creating new applications, processes and taking advantage of innovations in coding, cloud and hosting infrastructure, scalability and security requirements. Low-code development will also lower the barrier to entry, cost and time to deployment, thus a need of the hour. Is low code’s traction going to bridge the skills gaps or yield unemployment? Digital transformation is no longer a choice, it is indispensable. Every organisation, irrespective of its size or industry it is in, needs to continuously transform to survive in the evolving and demanding business environments. Low-code tools are what help accelerate digital transformation. Low code tools differ from each other in the underlying technology they use, the extensions they allow to the toolset, the feature set they provide to solve a particular problem and more. As low code tools get more traction, the talent pool will have newer opportunities to learn new technologies and ways to solve problems. It will undoubtedly get more job opportunities and career paths for people. It is certainly not going to cause any unemployment. As we have seen with every new technology, opportunities increase and thereby demand increases. What might have been an entry barrier for people with specialised knowledge to work in the IT sector might now become a bit easier for people with good analytical skills to become developers on low-code platforms. How can Robotic Process Automation help companies during the COVID-19 pandemic? Globally, the current pandemic situation has accelerated companies to have reliable processes and effective technologies to work more efficiently and effectively to carry out their businesses. Robotic process automation has become a beacon of hope for businesses aimed at increasing the productivity and quality of work across the organisation where even after the lockdown ends, half of the employee population will work from home such as IT companies. Using RPA, businesses can create and train software ‘bots’ that can serve as a virtual workforce. The bot can perform numerous repetitive tasks such as copying, pasting, comparing information between applications; they can even check information against various regulations and business rules. By automating mundane, repetitive, high-volume and rules-based tasks, the technology reduces the burden on humans. It is an important part of the solution stack for companies that want to become digital. How has COVID-19 accelerated the shift to digital and workplace transformation? How is Vuram participating in this? The pandemic has thrown a daunting yet relevant challenge to the corporates. They need to adopt digital transformation, irrespective of their current capabilities quickly. In the post-COVID world, technologies such as artificial intelligence, machine learning and robotics will be more prevalent in our everyday lives. Organisations that benefited from the emerging or already built digital capabilities will be in a better position to emerge as leaders in the post-pandemic era. Vuram has been in a very good spot to navigate the challenges thrown by the COVID situation easily. One of our founding principles was to enable our people to perform their work from any location. This simple objective led us to have zero servers in any of our premises; no software that we locally host resides in any particular machine. We have nimble cloud software agreements; we use effective collaborative tools for communication and coordination and use project management and delivery tools that are trusted, scalable and feature-rich. All our internal processes run on a BPM platform. Some of our processes were dependent on physical presence in the office. These processes were not mission-critical, but because of our digital maturity, we are now moving the few remaining processes to the platform as well. We have a continued reputation for being the best in Appian delivery. And now we are looking to expand that reputation to other areas of Hyperautomation. We are also looking at the direct sales of our solutions and technology toolsets. We have selected a few verticals to focus on and channelise our efforts towards them. We are focusing on specific verticals and are building-specific solutions that are modern, feature-rich, simple-to-use, pre-built and are easily configurable to a particular customer’s requirements. We are looking at the SMB segment as a potential growth market for our services. We have also identified a few industries that can benefit significantly from adopting modern technology platforms including education, manufacturing and healthcare.","excerpt":"In the current dynamic landscape, rolling out new applications and developing innovative software is the need of the hour, however, demands high-performance quality engineering to ensure utmost quality and accuracy. However, conventional methods come with several challenges, poor quality leading to hampering the outcome. Thus, automation testing is the proven way that can optimise the […]","categories":["AI Features"],"tags":["automation testing","Interviews and Discussions","low code","low code no code platforms","low code platform"],"author_name":"Sejuti Das","publish_date":"2020-08-01T18:00:00","publication_year":"2020","word_count":1372,"keywords":["Go","artificial intelligence","low code platform","automation testing","AI","low code no code platforms","machine learning","ML","R","Aim","analytics","Rust","low code","Java","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","R","Go","Rust","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/low-code-development-lowers-the-barrier-to-entry-the-need-of-the-hour-says-venkatesh-ramarathinam-ceo-vuram\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10057305,"title":"Misuse of artificial intelligence in China","content":"By 2030, China plans to become the world’s leading country in artificial intelligence (AI). Beijing’s AI development and implementation approach are fast-paced and pragmatic, with a focus on finding applications to help solve real-world problems. Several advances are being made in healthcare, such as “AI doctor” chatbots, machine learning for pharmaceutical research, and deep learning for medical image processing. Aside from this rapid development in artificial intelligence, China’s AI policies are deeply troubling and deserve condemnation. Nevertheless, portraying China as a “villain” in this way may be overly simplistic and potentially costly. The unexpected The way China uses artificial intelligence poses serious concerns. However, AI advances that benefit citizens and society shouldn’t overshadow the fact that China’s authoritarian government abuses citizens’ data and violates their privacy. Due to a massive police surveillance apparatus powered by big data and artificial intelligence, China is living up to the fictional scenario every day. The government, for instance, now requires national IDs to buy train tickets, making it easier to block human rights activists or anti-corruption journalists from travelling. According to reports and leaked documents, the government in Xinjiang, home of China’s Uighur Moslem minority, uses AI-sifted big data to screen people entering mosques and even shopping malls. The government can collect data on everything from bank accounts to family planning through thousands of checkpoints that require national IDs. The New York Times confirmed Soros’ fears when it reported that the Chinese authorities are using facial recognition technology to track and target members of the Uighurs, a persecuted Muslim minority in China. The recently released report by Human Rights Watch, titled “China’s Algorithms of Oppression,” provides additional evidence of Beijing’s use of new technologies to restrict the rights and liberties of Uighurs. Privacy rights In democratic and open societies, AI can directly interfere with human rights beyond its use by repressive regimes. As AI systems collect personal data for micro ad targeting, the right to privacy is violated. Monitoring online content enabled by AI impedes freedom of speech. Access to and the sharing of information by users are controlled in opaque and incomprehensible ways, restricting users’ freedom of expression and opinion. Disinformation campaigns powered by AI – such as troll bots and deepfakes (altered video clips) – threaten societies’ access to accurate information, disrupt elections, and erode social cohesion. Concerns also exist over the emergence of opaque social governance systems that lack accountability mechanisms. An AI-generated assessment is used to determine sentencing in the smart court system in Shanghai, for instance. Unfortunately, defendants find it difficult to assess the tool’s potential biases, the quality of the data and the soundness of the algorithm, making it difficult to contest the results. China’s experience demonstrates the need for transparency and accountability when it comes to AI in public services. Inclusion and the protection of citizens’ digital rights must be the goals when designing and implementing systems. It is unhelpful to reduce China’s rapid AI development into a simplistic narrative about China as a threat or as a villain. Observers outside China need to engage in the debate and take more steps to understand – and learn from – the nuances of what’s really going on. According to MIT researcher Jonathan Frankle, “this is an urgent crisis we are slowly sleepwalking our way into.”","excerpt":"It is unhelpful to reduce China’s rapid AI development into a simplistic narrative about China as a threat or as a villain.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","china ai","Machine Learning"],"author_name":"Sohini Das","publish_date":"2021-12-28T16:00:00","publication_year":"2021","word_count":547,"keywords":["Go","artificial intelligence","machine learning","AI","chatbots","Machine Learning","Git","RAG","china ai","Ray","deep learning","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","Ray","RAG","chatbots","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/misuse-of-artificial-intelligence-in-china\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049652,"title":"AWS announces general availability of Amazon QuickSight Q","content":"Amazon Web Services (AWS) announces the general availability of Amazon QuickSight Q, a new capability in Amazon QuickSight that allows anyone in an organisation to ask business questions in natural language and receive accurate answers along with relevant visualisations to help them gain insights from the data. “Now, anyone within an organisation can ask natural language questions and receive highly relevant answers and visualisations thanks to Amazon QuickSight Q. For the first time, anyone can harness the full power of data to make quick, accurate, data-driven decisions that will help them plan more efficiently and respond more quickly to their customers,” said Matt Wood, VP of Business Analytics, AWS. Amazon QuickSight Q also offers auto-complete suggestions for keywords and business terms, as well as spell checking and acronym\/synonym matching, removing the need for customers to worry about typos or knowing the exact business terminology for their data. Similarly, customers who use Amazon QuickSight Q do not have to make any upfront commitments, and they simply pay for the number of users or queries they use. Moreover, anyone can use Amazon QuickSight Q to get powerful analytics by asking business data questions in natural language and getting accurate answers with relevant visualisations in seconds. Users simply type their inquiry into Amazon QuickSight Q. Furthermore, Amazon QuickSight Q understands questions, so users may ask inquiries of data in natural language without having to do any complicated data preparation. Likewise, to deliver visualisations, Amazon QuickSight Q does not rely on prebuilt dashboards or reports, which eliminates the need for business intelligence (BI) analysts to update a dashboard every time a new business question emerges, allowing anybody to ask questions and receive visual answers in seconds. Amazon QuickSight Q currently supports questions in English and is generally available today to customers running Amazon QuickSight in the United States East (Ohio), the United States East (North Virginia), the United States West (Oregon), Europe (Frankfurt), Europe (Ireland), and Europe (London), with additional AWS regions to follow.","excerpt":"A new ML-powered functionality enables anybody to write questions about their company data in natural language and instantly obtain correct answers with relevant visualisations.","categories":["AI News"],"tags":["data preparation"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-09-24T10:52:12","publication_year":"2021","word_count":330,"keywords":["business intelligence","Go","AWS","AI","cloud_platforms:AWS","data-driven","data preparation","analytics","GAN","cloud_platforms:Amazon Web Services","R"],"extracted_tech_keywords":["AI","analytics","AWS","R","Go","GAN","business intelligence","data-driven","cloud_platforms:AWS","cloud_platforms:Amazon Web Services"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amazon-web-service\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061033,"title":"AI to make life easier for the disabled in 2022","content":"As per a 2011 WHO report, 15 percent of the global population lives with some form of disability. The global estimate for disability is on a rise with ageing population and the spread of chronic diseases, along with better techniques to measure disability. However, the good news is that with the advent of AI and other emerging technologies, the impact of such disabilities on a person’s daily functioning can be reduced to a great extent. A lot of organisations are venturing into this space and allotting resources and R&D efforts to tackle these challenges. We list some of the ongoing efforts in this space. AI’s promise for the paralyzed A research team led Grégoire Courtine, a professor at Swiss Federal Institute of Technology Lausanne (EPFL) and Jocelyne Bloch, a professor and neurosurgeon at CHUV, has developed a system that enables patients with a complete spinal cord injury to stand, walk, and perform activities like swimming, cycling and canoeing. According to the paper, the team has used sophisticated implants controlled by artificial intelligence software to stimulate the spinal cord region that activates the trunk and leg muscles. The stimulation algorithms are based on imitating nature. The team has used soft implanted leads that are to be placed underneath the vertebrae on the spinal cord. These implants can modulate the neurons and regulate specific muscle groups. We can activate the spinal cord by controlling these implants like the brain would do naturally to have the patient stand, walk, swim, or ride a bike. The epidural electrical stimulation (EES) targets the dorsal roots of lumbosacral segments that restores walking in people with spinal cord injury (SCI). The team came up with an arrangement of electrodes targeting the ensemble of dorsal roots involved in leg and trunk movements that result in superior efficacy, restoring more diverse motor activities after the most severe SCI. They established a computational framework that informed the optimal arrangement of electrodes on a new paddle lead and guided its neurosurgical positioning. They also developed software supporting the rapid configuration of activity-specific stimulation programs that reproduced the natural activation of motor neurons underlying each activity. As part of the ongoing clinical trial, these neurotechnologies were tested on three individuals with complete sensorimotor paralysis. Within a single day, activity-specific stimulation programs enabled these three individuals to stand, walk, cycle, swim and control trunk movements. Vision without eyes Source: WeWalk.io Founded by Kürşat Ceylan, WeWALK is a combination of a smart cane design and a smartphone app. This year in February, Microsoft partnered with WeWALK to develop the technology further. The AI-powered WeWALK cane pairs with a free mobile navigation app that can detect overhead obstacles and alert its users with haptic feedback, inform about restaurants, stores, cafes that they are passing by. WeWALK takes commands in nine languages: English, French, Italian, German, Russian, Spanish, Portuguese, Turkish, and Arabic. The smart cane not just pairs with the phone, but can also be controlled with the touchpad on the cane. With Microsoft, the team plans to develop the current technology and future offerings as they receive support towards developing AI Mobility (AIM)- technology. The next step of WeWALK’s journey is to roll out AIM technology in collaboration with Microsoft and introduce it to schools all around the world. “Our main goal is to help unlock the hidden potential of many more visually impaired people through cutting-edge tech. The hope is that with the use of a smart cane, users will be able to independently travel on their own anywhere, whether it’s down the street or across the world,” wrote Kürşat Ceylan on Twitter. Our main goal is to help unlock the hidden potential of many more visually impaired people through cutting-edge tech. The hope is that with the use of a smart cane, users will be able to independently travel on their own anywhere, whether it's down the street or across the world.— kürşat ceylan (@kursatceylan) February 8, 2022 AI for ALS Source: flyparrots.com After being highlighted as the 2021 Entrepreneur of the Year by the Association of Washington Business, Parrots Inc’s Polly has become quite famous. Designed in the shape of a parrot, Polly is powered by artificial intelligence is designed specifically to address the unique needs of the disabled community and their caregivers. Polly proves to be a great support for patients with ALS, MS, spinal cord injuries, and physical and neurological challenges as it can help anyone navigate their surroundings and help them communicate with ease. It can be attached to any wheelchair or bedside, tracks eye movement and uses ML to assist smart prediction of the user’s needs and wants. The wide-angle cameras capture a 360-degree view of the surroundings and display it on a Windows tablet to provide safe and efficient navigation. The AI Smart Prediction enables real-time start prediction. The bird also learns the user’s habits, predicts their needs, and communicates in real-time. Swarajability In India, too, the Indian Institute of Technology-Hyderabad (IIT-H) has developed Swarajability, a platform for the benefit of persons with disabilities like hearing impairment, visual impairment, and locomotive disorders. In collaboration with Kotak Mahindra Bank and Youth4jobs, the platform is India’s first AI triggered job platform for the disabled. It analyses the available information and suggests the required training needed for the concerned jobseeker. Wrapping Up AI is enabling people with disabilities to live independent life. The makers of technology are trying to understand their difficulties and figuring out ways to transform their lives and also the world into becoming an inclusive place.","excerpt":"Polly can attach to any wheelchair or bedside, track eye movement and use ML to assist smart prediction of the user’s needs and wants.","categories":["AI Features"],"tags":["iit hyderabad","Microsoft"],"author_name":"Meeta Ramnani","publish_date":"2022-02-18T17:00:00","publication_year":"2022","word_count":919,"keywords":["Go","API","artificial intelligence","programming_languages:R","AI","ML","iit hyderabad","Aim","ViT","GAN","R","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","R","Go","API","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-to-make-life-easier-for-the-disabled-in-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10059323,"title":"Microsoft records massive growth in the second quarter, Azure comes out as a winner","content":"Microsoft Corp has recorded a revenue of USD 51.7 billion with a net income of USD 18.8 billion in the second quarter. The tech conglomerate’s Intelligent Cloud Computing division logged 25.5% growth compared to the corresponding quarter last year. Microsoft Cloud has seen consistent growth year after year. In the previous quarter, ended September 30, 2021, the company’s revenue from Office Commercial products grew by 18% and the Intelligent Cloud Computing reeled in USD 17 billion. Of late, Microsoft has acquired Activision Blizzard to push its Metaverse ambitions, and cloud is a big part of its long term play. AWS takes the lion’s share of the USD 130 billion cloud infrastructure services market with 32% of the total pie– more than the market cap of Azure and Google Cloud combined. Competitive edge Last year, CRN’s Cloud Barometer survey asked service providers to rank the top three cloud platforms in terms of product capability, profitability, pricing maturity, lead sharing and support for demand generation. Microsoft Azure was ranked first in the fields of product capability, profitability and maturity of pricing. The survey participants favoured Microsoft Azure for its proven partnering records. Meanwhile, the Goldman Sachs IT Spending Survey in 2020 showed a similar outcome. While AWS led the charge, Azure came out as the most popular cloud infrastructure supplier, again. The survey asked clients to poll their vendors in two areas: IaaS (Infrastructure-as-a-Service) and PaaS (Platform-as-a-Service), and Microsoft topped both ratings. Solid head start The cloud landscape has undergone seismic changes over the last decade. According to a Gartner study, AWS was five times the size of 14 of its competitors combined ten years back. The online retailer essentially created the cloud computing services market as we know it when, in 2006, its subsidiary, Merchant.com, started helping third parties to build their own websites. AWS has engaged in competitive pricing since its launch, slashing rates more than 50 times. While Microsoft Azure and Google witnessed higher growth rates in the recent past, expanding a smaller customer base is easier than doing the same with a much larger group. AWS reportedly is still growing at a solid 47% with USD 3.53 billion in earnings even as the growth rate dipped. Aggressive expansion AWS’ clientele has exploded to a million users including Netflix and Airbnb. Amazon is determined to build more data centers. According to a CRN report, it had spent close to USD 3.5 billion on its cluster of data centers in Virginia alone between 2011–2020. While all three ‘hyperscalers’ spend billions of dollars each year on growing their data centers (The 20 largest data centers in the world reportedly spent USD 38 billion in the first quarter of 2021), Amazon shells out the most money. AWS announced three new data centers in the Middle East in the first half of this year. A second data center in India (Hyderabad) will go live this year. Optimisation “Over a year into the pandemic, digital adoption curves aren’t slowing down. They’re accelerating, and it’s just the beginning. We are building the cloud for the next decade, expanding our addressable market and innovating across every layer of the tech stack to help our customers be resilient and transform,” Microsoft CEO Satya Nadella said earlier. As much as 23% of the IT industry relies on cloud services and the percentage could double in the next three years. “Cloud will be foundational to scaling India’s digital journey. There’s been a massive growth in cloud services in India and across the world, on the back of the massive digital transformation happening across every industry. Reports by IDC predict the Indian public cloud services market to reach USD 10.8 billion by 2025, growing at a CAGR of 24.1% for 2020-2025. Last year, Microsoft introduced six industry-specific cloud offerings that deliver differentiated value by industry-Cloud for Financial Services, Healthcare, Manufacturing, Retail, Non-Profit and Cloud for Sustainability,” Dr. Rohini Srivathsa, National Technology Officer at Microsoft India stated.","excerpt":"Microsoft Azure was ranked first in the fields of product capability, profitability and maturity of pricing.","categories":["Global Tech"],"tags":["Amazon","AWS","Cloud Computing","Cloud services","Microsoft Azure"],"author_name":"Poulomi Chatterjee","publish_date":"2022-02-02T19:00:00","publication_year":"2022","word_count":656,"keywords":["Go","cloud_platforms:Azure","AWS","AI","cloud computing","R","digital transformation","Cloud services","Amazon","Git","cloud_platforms:AWS","Cloud Computing","Microsoft Azure","Azure"],"extracted_tech_keywords":["AI","cloud computing","AWS","Azure","R","Go","Git","digital transformation","cloud_platforms:AWS","cloud_platforms:Azure"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-records-massive-growth-in-the-second-quarter-azure-comes-out-as-a-winner\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001527,"title":"How IoT Is Poised To Be The Next Big Technological Wave In India","content":"Internet of Things is growing rapidly and the number of connected devices is expected to go over 20 billion devices by the year 2020 according to a research firm. In fact,  IoT’s market in India is expected to rise to over 9 billion USD by the year 2020 with over 1.9 billion units compared to the 1.3 billion USD in 2016. IoT Market Trend With over 120 startups that comprises more than 60% of all startups, India has a huge market for IoT, according to a joint report by two leading firms. The report says that the installed base of connected IoT units in India is expected to grow at a faster rate than those of developed economies. The rapid growth of IoT in India has been recent and IoT devices like smart home devices and smart electronics are becoming increasingly popular. In India, healthcare and manufacturing are two sectors that IoT has shown its presence in predominantly. The development itself will drive and support the growth of IoT, the rise of smart cities will leverage its power. Though the industrial application and adoption of IoT are expected to boom, the consumer side of IoT lags behind mainly due to price and security factors. The below image shows the current and expected market trends of the Internet of Things in India according to the report : IoT in Indian Sectors Over the coming years, IoT’s will become more predominant in sectors like Manufacturing, Agriculture, Healthcare, Transport and Logistics and Retail. Given below are some of the common uses cases of IoT across different sectors: Manufacturing: IoT applications will play out in inventory management, plant safety and security, quality control, logistics and supply chain optimisation and production flow monitoring are some of the best use cases of IoT in the manufacturing sector. India has a wide scope for adopting the use cases with an expanding manufacturing sector and Government actually promoting the Make in India initiative. Retail: Retail industry will see the development of IoT applications such as automated checkouts, beacons, smart shelves, In-store Layout Optimization call etc. Most of the retail industry in India has already adopted IoT for improving promotion through advertisements of offers. Transport and Logistics(T&L): T&L businesses are focused on maximising supply chain efficiency in order to sustain profitability and viability and therefore need to make end-to-end improvements, IoT can help improve the end-to-end visibility. Warehouse and Yard Management and Fleet Management are also top use cases for IoT in T&L Healthcare: Remote medical assistance is something that revolutionised the healthcare industry, with the help of IoT’s. Health tracking and analysis with the help of fitness gadgets are popular consumer IoT’s. Agriculture: Livestock monitoring, agricultural drones and precision farming are top use cases of IoT in the Agricultural sector. The Indian government, as well as the private sector, has been focusing on bringing technology to the Agricultural Sector. With IoT’s and similar technologies the agricultural sector in India which is considerably slow in growth can be greatly benefited. Conclusion Internet of Things is changing the way we communicate with the devices. India’s adoption of IoT as a technology is driving the development of the nation towards a better-connected future and the Government also seems to be making a big leap towards technology.","excerpt":"Internet of Things is growing rapidly and the number of connected devices is expected to go over 20 billion devices by the year 2020 according to a research firm. In fact,  IoT’s market in India is expected to rise to over 9 billion USD by the year 2020 with over 1.9 billion units compared to […]","categories":["AI News"],"tags":["is the tech boom over"],"author_name":"Amal Nair","publish_date":"2019-03-18T17:16:47","publication_year":"2019","word_count":545,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","RAG","is the tech boom over","R","startup"],"extracted_tech_keywords":["AI","RAG","R","Go","API","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/how-iot-is-poised-to-be-the-next-big-technological-wave-in-india\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10007581,"title":"10 Creative Safety Features For Driverless Car","content":"For a few years now, research and development in driverless cars have evolved tremendously, and several intuitive and creative changes have been witnessed. There have been various demonstrations of these autonomous vehicles by the auto giants both on and off roads. However, the measure of safety has always been a serious concern in self-driving cars. There have been severe accidents during the trials on-road of various driverless cars. For instance, in 2016, a Tesla driver died in a fatal crash while using autopilot mode. In another instance in 2018, a self-driving car from Uber hit and killed a woman. It is, therefore, a key requirement to work around the safety features of self-driving cars. In this article, we discuss 10 creative measures that can or have been adopted for driverless cars. (The list is in no particular order). 1| Adaptive Cruise Control Adaptive cruise control (ACC), also known as Dynamic Cruise Control is an intelligent form of cruise control that works by slowing down the speed and speeding up automatically to keep pace with the vehicle in front of it. This technique helps in avoiding a collision. This control scheme is designed to improve drivers’ comfort during multi-vehicle driving situations and to completely avoid rear-end collision using severe braking and lane change manoeuvre. This technique is widely regarded as the critical component of future generations of intelligent driverless cars. Know more here. 2| Autonomous Emergency Braking (AEB) Autonomous Emergency Braking (AEB) is one of the most advanced developments for standard safety equipment on autonomous vehicles. This technique works by scanning the road ahead and can apply the brakes automatically to avoid a collision. In autonomous vehicles, this technique works by automating the activation of a car’s brakes in conjunction with the existing self-steering, lane-keeping assistance and other systems to control the vehicle with negative inputs from the driver. Know more here. 3| Blindspot detection Blindspot detection is one of the core technologies which provides 360 degrees of electronic coverage around a car, regardless of the speed. Developed by Volvo, this technology tracks traffic just behind the vehicle as well as what’s coming alongside. Blindspot detection or blindspot monitoring includes two different categories, active blindspot monitoring and passive blindspot monitoring. Know more here. 4| Electronic Stability Control (ESC) Electronic Stability Control (ESC) is an automatic feature that uses automatic computer-controlled braking of individual wheels to assist in maintaining control in critical driving situations. The feature becomes active when the driver loses control of the vehicle. According to this blog, vehicles fitted with ESC are involved in 32% fewer single-vehicle crashes, and 58% fewer roll-over crashes that result in driver injury. Know more here. 5| Lane-Keeping Assist Lane-keeping assists (LKA) is a technique in autonomous driving vehicles that enables vehicles to travel along a desired line of lanes by adjusting the front steering angle. This technique is used mainly for preventing accidents during free driving of autonomous vehicles. It works by warning the driver and\/or deploying the steering if the vehicle moves out of a lane. Know more here. 6| Reverse Park Assist The Reverse Park Assist system helps a driver to sense when objects are in the blind spot of the vehicle. The system can help in preventing reverse parking accidents. There are mainly two common types of reverse parking-assist systems, which are simple audio warning and a sophisticated camera and video-monitoring system. Also known as rear park assist, this technique utilises multiple ultrasonic sensors located on the rear bumper of the vehicle. Know more here. 7| Rear Cross-Traffic Assist Rear cross-traffic assist or rear cross-traffic alert system helps drivers reversing out of perpendicular parking spaces when their rear view is obstructed. According to this blog post by Bosch, rear cross-traffic alert can make reversing easier by warning about the vehicles crossing when getting out of a parking space. This system basically utilises two mid-range radar sensors in the rear of the vehicle, which measure as well as interpret the distance, speed and anticipated driving path of vehicles detected in cross-traffic. Know more here. 8| Traffic Jam Assist Traffic Jam Assist can be said as an extension of the Adaptive Cruise Control. The technique is basically a low-speed version of Adaptive Cruise Control that tries to maintain the set speed while taking other vehicles into account. Traffic jam assist is based on the sensors, and the functionality of adaptive cruise control with stop & go and lane-keeping support. When the Adaptive Cruise Control systems ‘stop & go’ is turned on, it continuously analyses the speed of the surrounding vehicles and compares with their own driving speed.. Know more here. 9| Vehicle-to-Vehicle (V2V) Communication The Vehicle-to-Vehicle (V2V) communication is a technique that wirelessly exchanges information about the speed, position as well as distance of the surrounding vehicles. The technology behind V2V communication allows vehicles to broadcast and receive Omni-directional messages and creates a 360-degree “awareness” of other vehicles in proximity. Know more here. 10| Vehicle Guidance System This technology helps steer the vehicle without human intervention. Unlike driver assistance systems, this system needs no monitoring by a human driver. The Vehicle Guidance System is part of the control structure of the vehicle and consists of a path generator, a motion planning algorithm and a sensor fusion module. Know more here.","excerpt":"For a few years now, research and development in driverless cars have evolved tremendously, and several intuitive and creative changes have been witnessed. There have been various demonstrations of these autonomous vehicles by the auto giants both on and off roads.  However, the measure of safety has always been a serious concern in self-driving cars. […]","categories":["AI Trends"],"tags":["Autonomous Vehicles","driverless cars","safety","Self Driving Cars"],"author_name":"Ambika Choudhury","publish_date":"2020-09-18T11:00:00","publication_year":"2020","word_count":878,"keywords":["Go","programming_languages:R","AI","Self Driving Cars","programming_languages:Go","ai_applications:autonomous driving","driverless cars","RAG","safety","Autonomous Vehicles","R"],"extracted_tech_keywords":["AI","RAG","R","Go","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-creative-safety-features-for-driverless-car\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121626,"title":"NVIDIA Controls a Whopping 95% of the AI Chip Market","content":"It’s Jensen Huang’s world, and we’re all just living in it. Recently, the AI chip giant reported a profit of $14.881 billion, up 629% compared to the corresponding quarter last year. Revenue was $26.04 billion, surpassing the estimated $24.65 billion. NVIDIA controls a whopping 95% of the AI chip market right now. To put it in perspective, NVIDIA  is now larger than Tesla and Amazon combined. Furthermore, NVIDIA  is now larger than the entire German stock market. BREAKING: Nvidia stock, $NVDA, is now trading with a market cap above $2.5 TRILLION for the first time in history.To put this in perspective, Nvidia is now larger than Tesla and Amazon COMBINED.Furthermore, Nvidia is now larger than the entire German stock market.Nvidia is… pic.twitter.com\/Uaj9mgFoML— The Kobeissi Letter (@KobeissiLetter) May 22, 2024 NVIDIA X (OpenAI Spring Update, Google I\/O and Microsoft Build) Last week, we saw several tech events, from OpenAI’s Spring update to Google’s I\/O and Microsoft Build. What tied them together was NVIDIA, which made a notable presence at all three. OpenAI released GPT-4o, which won hearts with its ‘omni’ capabilities across text, vision, and audio. OpenAI’s demos included a real-time translator, a coding assistant, an AI tutor, a friendly companion, a poet, and a singer. However, this would not have been possible without NVIDIA. “I just want to thank the incredible OpenAI team and thanks to Jensen and the NVIDIA team for bringing us the advanced GPU to make this demo possible today,” said OpenAI CTO Mira Murati at the end of the event, expressing gratitude to NVIDIA. Notably, Huang personally delivered the first DGX H200 to OpenAI last month. Similarly, both Microsoft and Google were quick to claim that they were the first ones to get their hands on NVIDIA’s latest GPU, Blackwell. “We are also proud to be one of the first cloud providers to offer NVIDIA’s cutting-edge Blackwell GPUs, available in early 2025,” said Google Chief Sundar Pichai at Google I\/O. “We’re bringing in the latest H200s to Azure later this year and will be among the first cloud providers to offer NVIDIA’s Blackwell GPUs in B100 as well as GB200 configurations,” said Microsoft chief Satya Nadella at Microsoft Build. He added that Microsoft is continuing to work with NVIDIA to train and optimise models like GPT-4o and small language models like the Phi-3 family. Apart from the above three players, leading LLM companies such as Adept, Anthropic, Character.AI, Cohere, Databricks, DeepMind, Meta, Mistral, xAI, and many others are building on NVIDIA AI in the cloud. Blackwell Fever Begins NVIDIA’s data centre revenue alone was $22.6 billion, a record, up 23% sequentially and 427% year-on-year, driven by continued strong demand for the NVIDIA Hopper GPU computing platform. ‘The demand for GPUs in all data centres is incredible. We’re racing every single day. And the reason for that is because of applications like ChatGPT and GPT-4o,’ said Jensen, adding that the future is going to be multi-modal. Moreover, he announced that NVIDIA will build a new chip every year. “I can announce that after Blackwell, there’s another chip. We’re on a one-year rhythm,” said Huang on the earnings call. He further stated that the Blackwell platform is in full production and forms the foundation for trillion-parameter scale generative AI. “Our production shipments will start in Q2 and ramp in Q3, and customers should have data centres stood up in Q4,” he said. “Blackwell’s time-to-market customers include Amazon, Google, Meta, Microsoft, OpenAI, Oracle, Tesla, and xAI,” said NVIDIA CTO Colette Kress. In India,  Yotta is India’s sole NVIDIA Partner Network Cloud Partner (NCP) and will receive NVIDIA’s latest GPU Blackwell by October. Yotta plans to scale up its GPU inventory to 32,768 units by the end of 2025. Last year, the company announced that it would import 24,000 GPUs, including NVIDIA H100s and L40S, in a phased manner. Huang also announced NVIDIA Spectrum-X, an advanced Ethernet networking platform specifically designed to enhance the performance and efficiency of AI workloads. NVIDIA is betting big on autonomous vehicles. “We supported Tesla’s expansion of its training AI cluster to 35,000 H100 GPUs. Their use of NVIDIA AI infrastructure paved the way for the breakthrough performance of FSD Version 12, their latest autonomous driving software based on Vision,” said Kress. Similarly, in a recent interview with Yahoo Finance, Huang stated that besides the cloud industry, the primary users of NVIDIA’s data-center chips are from the automotive industry. He then emphasised about autonomous cars and their advancements. “Tesla is far ahead in self-driving cars, but every single car, someday, will have to have autonomous capability,” said Huang. .","excerpt":"NVIDIA  is now larger than the entire German stock market.","categories":["Global Tech"],"tags":["NVIDIA"],"author_name":"Siddharth Jindal","publish_date":"2024-05-26T12:03:05","publication_year":"2024","word_count":764,"keywords":["Anthropic","ChatGPT","OpenAI","AI","GPT-4o","ML","small language models","Aim","generative AI","NVIDIA","xAI"],"extracted_tech_keywords":["AI","ML","generative AI","GPT-4o","ChatGPT","OpenAI","Anthropic","xAI","Aim","small language models"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidia-controls-a-whopping-95-of-the-ai-chip-market\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":61281,"title":"How Google Is Teaching Robots To Be Agile Like Dogs","content":"With Google AI conducting many studies, a considerable amount of research has gone into understanding how machines can imitate human – or animal – behavior. Most recently, the company developed a system that learns from the motions of animals to give robots greater ‘agility’. Developing a robot that can move like animals\/humans and one that can replicate agile\/complex motions opens up various possibilities. One practical use case could be to leverage its help for sophisticated tasks. But designing such robots is not easy. While in the past, Google explored the possibility of making a robot learn how to walk from scratch, in their latest attempt, they explore the idea of making a four-legged robot learn agile behaviors by imitating motions from real animals like trotting and hopping. The Research To counter problems like inefficient algorithms and hardware, the teams’ framework takes motion capture clips of a dog and uses reinforcement learning. The clip provides the system with different reference motions. It helps researchers teach a four-legged Unitree Laikago robot to perform a range of behaviors, like hopping, turning and speed walking (4.18 km\/hr). First, Google AI’s research team compiled a data set of real dogs performing various tasks\/skills. Next, by using the different motions in the reward function – which describes how agents ought to behave – the researchers used about 200 million samples to train the simulated robot to imitate the motion skills, in this case, the motions of a dog. The training took place in a physics simulation so that the pose of the reference motions could be carefully tracked. Via Google But the problem with simulators is that they provide a rough idea of approximation of the real-world. To find a solution to this problem, the researchers employed an adaptation technique that randomized the dynamics of the simulation. To give an example, the technique randomized the physical quantities of the robots, like its mass and friction. Values like these were mapped using an encoder to a numerical representation, which is called encoding, and this encoding was passed on as an input to the root control policy. But the encoder was removed during deployment, and the researchers directly searched for a set of variables that allowed these robots to execute skills successfully. Via Google The system, according to the team, was able to adapt the policy to the real-world using a clip from real-world data which was under eight minutes. The real-world data has approximately 50 trials, and they demonstrated that the real-world robot learned to imitate various motions from a dog and excellent artist animated keyframe motions. The motions included pacing, trotting and spinning action, as well as a dynamic hop-turn. Final Thoughts Although this approach holds a lot of promise, it comes with a few challenges. The control policy can only learn to hop, spin, trot and walk but, imitating an animal naturally means doing more than that. The researchers say that while the system has successfully learned policies for a diverse set of behaviors, due to limitations in hardware and algorithms, they have not been able to learn more dynamic behaviors, like large jumps and runs. Also, the behaviors learned are not stable when compared to the manually-designed controllers. While the team encounters these problems, they believe that reproducing complex behaviors and improving the robustness of the system will eventually lead to more complex real-world applications. In future, they expect to make the system learn from video clips as much as possible.","excerpt":"With Google AI conducting many studies, a considerable amount of research has gone into understanding how machines can imitate human – or animal – behavior. Most recently, the company developed a system that learns from the motions of animals to give robots greater ‘agility’. Developing a robot that can move like animals\/humans and one that […]","categories":["Global Tech"],"tags":["Google","Robots"],"author_name":"Sameer Balaganur","publish_date":"2020-04-10T16:00:00","publication_year":"2020","word_count":577,"keywords":["Replicate","Go","programming_languages:R","AI","programming_languages:Go","RAG","Google","CLIP","R","Robots"],"extracted_tech_keywords":["AI","RAG","R","Go","CLIP","Replicate","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-google-is-teaching-robots-to-be-agile-like-dogs\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042989,"title":"How Common Are INR 1Cr+ Salaries In Data Science In India?","content":"Data has become central to organisations across every industry as they realise its potential in strategic decision-making or effective process automation. As a result, the analytics or data science function has witnessed significant growth in the past couple of years. Analytics professionals are in demand, and organisations are willing to pay competitive salaries to attract the best talent. The scale of growth of any domain is evinced by the salaries drawn by experienced professionals or the growth in the proportion of experienced professionals drawing higher wages. Overall, the analytics professionals have seen an increase in their pay and the industry has seen more analytics professionals being hired in strategic positions to ensure enterprise-wide adoption. Even though the pandemic resulted in a decrease in the median salary of analytics professionals in 2021 compared to 2020, it was still higher than in 2019. With the industry showing considerable growth and an overall increase in the demand for analytics or data science post-pandemic, this salary is expected to grow even further in the coming years. This report aims to analyse the data professionals hired at the highest level with the most competitive salaries. It provides an overview of the analytics professionals with a salary package greater than INR 1 Crore and their break up across parameters such as sectors, geographies, company type, education, years of experience, and gender representation. In addition, the report outlines how large companies are aligning their companies to be more data-driven and leveraging analytics in their strategy. Methodology The research presented in this report has been collected from publicly available data, including news reports, academic journals, job sites, and other information portals. The data was collated around the last week of June 2021. Key Highlights Around 1,400 data science professionals working in India are paid a salary of more than INR 1 Crore. More than one in four (28.8%) of these professionals work in the IT & Consulting sector. This is followed by the BFSI and Internet & Technology sectors at 13.9% and 12.6%, respectively.Almost two in five (38.7%) of these professionals work in Captive Centres established by MNCs in India.Almost three in four (72.8%) of these analytics professionals are based in Bengaluru, Mumbai, or Delhi.More than two in five (40.8%) of these analytics professionals are MBA postgraduates.Almost four in five (79.0%) of these analytics professionals are employees with work experience greater than 15 years. Only 17.5% of these analytics professionals are women. Share of Analytics Professionals by Sector The IT Services & Consulting sector has the highest share of employees (28.8%) with a salary package greater than INR 1 Crore. This is not surprising, considering that most data science professionals work in IT or consultancy firms. The professionals getting a salary greater than INR 1 Crore are working in some of the biggest IT companies – Domestic and MNCs.BFSI at 13.9% and the Internet & Technology (E-commerce) at 12.6% are the second and third largest contributors to this pool of analytics employees. These two sectors have heavily invested in analytics over the last couple of years. Almost all of the processes in their value chain are digitised. In addition, both have realised the huge potential in using data for consumer analysis to improve revenue and customer service. Hence, both have been early adopters of data science with a significant market share within the domain. Their analytics functions are now maturing.The Retail & CPG sector has 11.5% of the analytics professionals with salary packages greater than INR 1 Crore.These professionals are mainly working in the captive centres set up by multinational consumer companies.The Engineering & Production industry and Pharma & Healthcare both have a significant market share in the data science industry, with many employees working in their analytics function. However, both of them have only 5.1% and 4.6% of the professionals, respectively, that earn more than INR 1 Crore per annum.Engineering & Manufacturing’s adoption of data science is very recent, and the analytics function has a relatively younger workforce as the analytics adoption is very recent.As for Pharma & Healthcare, analytics professionals are hired mainly at a technical level to develop applications to automate tasks. Very few enterprises within the sector hire analytics professionals at strategic positions to justify a salary package greater than INR 1 Crore. Share of Analytics Professionals by Company Type All companies with hired analytics professionals for a salary greater than INR 1 Crore are big brands that operate across geographies. Some are domestic firms, while others are MNCs with analytics functions established as cost centres in India. Captive centres provide the highest median salary as of 2021. Subsequently, they also have the highest share (38.7%) of analytics professionals with a salary greater than INR 1 Crore.Captive centres are Global In-house Centres (GICs) for big MNCs with operations set up in India. These are set up by companies such as PayU, PepsiCo, Philips that have a portfolio of high margin products and services. Hence, they provide very competitive salaries.Non-IT domestic firms come second with 24.7% of the INR 1 Crore+ salaried analytics professionals working for them.The category includes companies (in the traditional sectors) like Tata Steel, Bajaj Auto, and Mahindra & Mahindra that have ramped up their analytics adoption, and also enterprises (in data-driven sectors) like Myntra, Ola, and Edelweiss General Insurance.MNC IT & KPO Services and Consulting Firms make up 16.7% and 14.0% of the analytics professionals salaried greater than INR 1 Crore. Many MNC IT & KPO Service providers like Dell EMC, Accenture, IBM, etc., and Consulting firms like BCG, Deloitte, KPMG, etc., have set up their analytics functions in India.Very few (6.0%) Domestic IT companies provide a salary of more than INR 1 Crore to their analytics professionals.These are companies like Wipro and The Smart Cube. Some of the MNCs headquartered in foreign countries like the US, the UK, etc., pay salaries to their Indian employees in their local currencies of USD, GBP, etc., which when converted to INR are greater than 1 Crore. Share of Analytics Professionals by City Analytics employees in Bengaluru, Delhi, and Mumbai host the highest share of analytics employees working in India. They also pay the highest median salary among all the top metropolitan cities in India. Likewise, almost three in four, 72.8%, of the analytics professionals with a salary greater than INR 1 Crore work for Bengaluru, Delhi NCR, and Mumbai offices. Bengaluru – the hub of analytics and data science –  has the highest share (27.4%) of data science professionals with a salary greater than INR 1 Crore.Bengaluru has the headquarters of all the top domestic IT companies like Wipro, Infosys, etc. In addition, all the leading MNCs, including IT and non-IT companies, have set up their analytics functions in Bengaluru as cost centres. These analytics units are huge and need leaders with extensive domain experience to run them.Delhi hosts more than one in four (26.5%) analytics professionals with a salary greater than INR 1 Crore. Mumbai has 18.9% of the analytics professionals with a salary greater than 1 Crore+ working in the city.Mumbai has the highest median salary of analytics professionals and a sector-location niche that includes the top BFSI captives, investment companies, and domestic banks. As mentioned before, the financial sector has been investing heavily to improve its data science capabilities. Mumbai also has a higher cost of living.Hyderabad hosts 10.7% of these analytics professionals, while Chennai and Pune host 6.8% and 6.3%, respectively. Share of Analytics Professionals by Education More than two in five (40.8%) analytics professionals with a salary greater than INR 1 Crore are MBA graduates.Around 15.9% are MBA postgraduates from top-ranked universities like IIM, ISB, etc. The remaining 24.9% are from other post-graduate institutes.Around 28.7% of these analytics professionals have done their undergraduate in engineering.Only 3.8% are engineering undergraduates from top-tier universities like IIT, BITS, etc. This percentage is comparatively lower because many of these undergraduates have gone on to do their postgraduation in MBA also from top-tier universities. Hence, these employees are a part of the MBA (Top Ranked) category.According to our Analytics India Industry Study 2021, Engineering undergraduates and MBA post-graduates from top-tier universities form only 5.8% of the total analytics employee share. However, they make up 19.7% of the analytics professionals who earn more than INR 1 Crore.Non-engineering undergraduates make up 27.7% of the analytics professionals.These are primarily graduates with specialisations in computer applications, statistics, economics, commerce, mathematics, and other related fields.Only 2.7% of the analytics professionals earning a salary more than INR 1 Crore has completed their PhD. Years of Experience Data science professionals with a salary greater than INR 1 Crore in India have a median work experience of 19.3 years. Almost four in five (79.0%) of these professionals are employees with work experience greater than 15 years. The majority of these professionals work in the top management working as the CxO, VP\/SVP, or Heads of Analytics\/Data\/Information for IT & Consulting, Banking, and FMCG enterprises.25.1% of these analytics professionals are engineering graduates or MBA postgraduates from top-tier universities.33.9% are with work experience between 15 to 20 years, 30.7% between 20 to 25 years, and 14.4% with work experience greater than 25 years. 21.0% are analytics professionals with work experience between 10 to 15 years.These professionals work as the Heads of Analytics or Business Intelligence Units or in the Top Management, mainly for IT & Consulting, Internet, Banking, or Software Product companies. One-third of these (31.3%) professionals are engineering graduates or MBA postgraduates from top-tier universities.6.8% are with work experience between 10 to 12 years, and 14.3% have work experience between 12 to 15 years. Share of Analytics Professionals by Gender Only 17.5% of the analytics professionals with a salary greater than INR 100 Crore are females when women make 28.1% of the total analytics professionals. There is significant work to be done if we are to achieve equality (50%-50%) in terms of gender representation overall as well as for employees with higher salaries.Women Analytics employees are still new in the workforce in India, with a median employee experience of 6.3 years compared to 7.6 years overall. Overall women also receive a median salary of 30.9% lesser men.More representation of female analytics professionals at strategic positions can help improve gender ratio and equal pay. Conclusion The analytics industry has seen significant growth over the last few years and is predicted to grow in the coming years as well. There is increased adoption in enterprises across sectors as well as within enterprises across its departments. With the growth in its adoption, enterprises are seeing an increasing need of hiring analytics professionals at strategic positions. These professionals are not just needed to ensure a well-planned analytics strategy to increase revenue but to also ensure that this growth is safe and sustainable. As of today, there are around 1,400 analytics professionals in India that earn a salary greater than INR 1 Crore. They make up around 0.2% of total analytics professionals and around 3.0% of the analytics professionals with work experience greater than 15 years. With the current salary and industry trend that we have observed through our reports at AIM Research, we anticipate this share to grow. While still far from being qualified as a mature industry, the increase in salaries through hiring analytics professionals at strategic positions will help it move in that direction.","excerpt":"Data has become central to organisations across every industry as they realise its potential in strategic decision-making or effective process automation. As a result, the analytics or data science function has witnessed significant growth in the past couple of years. Analytics professionals are in demand, and organisations are willing to pay competitive salaries to attract […]","categories":["AI Highlights"],"tags":["data science and manufacturing","data science salary"],"author_name":"Kashyap Raibagi","publish_date":"2021-07-12T10:00:00","publication_year":"2021","word_count":1879,"keywords":["data science","Go","data science and manufacturing","AI","Git","RAG","automation","Aim","analytics","data science salary","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Git","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-common-are-1crore-salaries-in-data-science-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10041569,"title":"Top Cybersecurity Budgets Around The World","content":"Technological advancement has not only given us power to manage everything with a click of a button, but it has also made us vulnerable to many threats online.The unexpected onset of the pandemic and the shift of workspace have led to a rapid increase in cyber-attacks globally.In 2020, there was a huge increase in both criminal attacks and ransom attacks around the world. According to Check Point research, the number of ransomware assaults worldwide increased by 102% in 2021. In addition, the numbers showed that India is the most afflicted country, with an increase of 17% in the number of weekly ransomware assaults per organisation during the first half of the year. Only recently, 4.5 million customers around the world were affected by a breach in Air India’s data servers. The increasing activity in 2021 included reports of high-profile ransomware assaults on essential infrastructure, private businesses, and municipalities. This year demands for ransom have risen to tens of millions of dollars around the globe. A more advanced approach is using threat actors to steal essential firm data and hold it for ransom. A recent attack on Colonial Pipeline, a major US gasoline corporation, led to losses for businesses globally in 2020 of over $20 billion, roughly 75% more than the total in 2019. The Colonial Pipeline, which spans for more than 5,500 miles, was forced to shut down all its fuel distribution network — and consequently jeopardised the transportation of gasoline and jet fuel over the U.S. east coastline due to a ransomware attack. The pipeline moves around 100 million gallons of fuel daily. The pipeline is spread across 14 states and serves the transportation needs of seven airports. In light of this recent attack on the colonial pipeline, many countries have come forward to strengthen their network security.Cybersecurity Ventures anticipates that over the five years, 2017 to 2021, worldwide spending on cybersecurity products and services would collectively exceed $1 trillion. Let’s look at some of the top cybersecurity budgets around the world. USA The U.S. government is expected to allocate $18.78 billion for cyber security investment in 2021, according to an Atlas VPN inquiry. Recently, President Joe Biden had proposed a $2.1 billion allocation for the Cybersecurity Infrastructure and Security Agency (CISA), $110 million  more than the authorised level in 2021. Israel Israeli cybersecurity broke another record in the first three months of 2021, with a $1.5 billion capital influx.In contrast to other sectors that have suffered economic setbacks, Israel’s cybersecurity business has continued to surge, breaking a new record every other year. In the first year of operation in 2019, the investment firms set a record by raising $1.8 billion in capital. Iran In its national budget for the cyber-space program for its two-government-controlled institutions, the Tehran administration has updated $ 71.4 million in its budget. As reported by IRNA, Iran’s official news agency, the Islamic Republic of Iran Broadcasting Agency (IRIB) “cyberspace activists” program received $63 million and another $ 8.4 million for its ‘cyberspace section’ of the Islamic Development Organization. Canada The Canadian government has planned to spend $80 million on a new federal cybersecurity network. The new pan-Canadian cybersecurity programme is planned to be supported by $80 million CAD (Canadian dollars) in financing over the next four years. Malaysia In 2021 CyberSecurity Malaysia (CSM) was allocated a budget of  RM27 million to help boost Malaysia’s cybersecurity. The grant of RM27 million to CSM would help Malaysia meet its digital transformation readiness. Malaysia’s Prime Minister, Tan Sri Muhyiddin Yassin, launched the MCSS 2020-2024, with a budget of RM1.8 billion. Australia As part of its 2021-2022 Budget, the Federal government has given Australia a new range of investment in digital services and technology. The Digital Transformation Agency (DTA) has key components of this, while the total budget of the agency has been decreased from $425,5 million for this fiscal year to $336 million this year.The Digital Technology Taskforce is set to extend by 30 June 2022 by $3.2 million in 2021-22. France In 2021, following recent hospital breaches, French President Emmanuel Macron has announced a plan to invest €1 billion ($1.2 billion) to strengthen cybersecurity in France, with €350 million (approx. $400 million USD) set aside for hospitals. United Kingdom The recent £16.5 billion increase by British Prime Minister Boris Johnson in British military spending includes an interesting caveat for the cyber security community. A sizable proportion of the investment will be dedicated to securing and enhancing the country’s cybersecurity and offensive capabilities.To help UK  fight cyber criminals on a large scale, the National Cyber Force will be created with an investment of 1.5 billion pounds. Hiscox reports that, globally, the average cybersecurity budget is $1.46 million , while in the UK, it is around $900,000. Spain As of January 2021, the government revealed its National Digital Skills Plan in Plan Nacional de Competencias Digitales in Spanish. It included joint investments across the aforementioned frameworks for the digitization of SMEs and public authorities in 2021-2025. As a result of these critical digital and cyber skills gaps in Spain, this new strategy is an answer to them. The EU’s €4.66 billion Digitalisation of SMEs Plan for 2021-2025 will make a direct impact through a €4.46 billion digitalisation subsidy for business. Whereas, the Government of India, in spite of having one of the largest digital ecosystems, the cyber-security budgets are still paltry. That said, the Indian government has increased the expenditure for the Indian digital programme by 23 percent to Rs 3958 crore for the year 2020-21.The funds allocated for cyber security projects and promotion of the IT and ITeS industries has seen and increase from Rs 102 crore and Rs 90 crore to Rs 170 crore.This increase was mainly due to incentives for electronic production, research and development, cyber security, and the promotion of IT and IT services. The switch to remote work has brought about a significant change in the cyberspace domain, the increase in online presence has compelled cybersecurity around the world to adopt certain measures to ensure the safety of their digital assets. India is progressing quickly into digital adaptation and innovation as a result of the Digital India Initiative. Although India’s cybersecurity budget is still inadequate, according to market analysts, India’s cybersecurity services industry is projected to grow from $4.3 billion in 2020 to $7.6 billion in 2022. According to the Data Security Council of India, the size of the industry is expected to be $13.6 billion by 2025, with a growth rate of 21%.  In 2014, the NCCC’s separate budget of one billion rupees was allotted to increase the country’s cybersecurity. This proposal is geared toward helping the cybersecurity ecosystem in India grow stronger. The MeitY has launched the Cyber Surakshit Bharat initiative which was in conjunction with the National e-Governance Division (NeGD). Even though India faces a critical cybersecurity risk with a dire need to improve the cybersecurity defences, it is taking small steps in improving the overall cybersecurity infrastructure.","excerpt":"Technological advancement has not only given us power to manage everything with a click of a button, but it has also made us vulnerable to many threats online.The unexpected onset of the pandemic and the shift of workspace have led to a rapid increase in cyber-attacks globally.In 2020, there was a huge increase in both […]","categories":["AI Trends"],"tags":["Air India","Cybersecurity","france","israel","Ransomware","USA"],"author_name":"Ritika Sagar","publish_date":"2021-06-10T10:00:00","publication_year":"2021","word_count":1158,"keywords":["Go","API","AI","Ransomware","digital transformation","innovation","Air India","Git","RAG","israel","france","ViT","GAN","Cybersecurity","R","USA"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","API","GAN","ViT","digital transformation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-cybersecurity-budgets-around-the-world\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10023344,"title":"Register For This Webinar To Understand Why Competence In Business Analytics Is Critical For Data-Driven Business Decisions","content":"In this era of AI, organisations have to rely on gaining valuable insights from historical data to make strategic business decisions. And hence, without any doubt, the roles relating to business analytics have emerged as one of the most critical functions that exist in a data-driven enterprise. The surge in demand has opened new career opportunities for individuals in the analytics domain. New age businesses are looking for competence in analytics as a critical skillset in their managers. About The Webinar VCNow has joined hands with top-ranked management school IIM Ahmedabad to offer the Executive Programme in Advanced Business Analytics (EPABA) exclusively for working professionals. We at Analytics India Magazine are organising a live webinar to help interested candidates understand how competence in business analytics is critical for making data-driven business decisions and also to answer your queries on IIMA’s Executive Programme in Advanced Business Analytics (EPABA) by the session speaker. REGISTER NOW Session Speaker Prof. Arnab Kumar Laha, M.Stat & PhD (Indian Statistical Institute)Professor, Indian Institute of Management Ahmedabad Arnab Kumar Laha is a professor of Production and Quantitative Methods (P&QM) at the Indian Institute of Management Ahmedabad. He is also the faculty chair of the Executive Programme in Advanced Business Analytics. He takes a keen interest in understanding how analytics, machine learning, and artificial intelligence can be leveraged to solve complex business and societal problems. He has published his research in reputed national and international journals, authored a popular book on analytics, and has edited book volume published by Springer. He was named among the “20 Most Prominent Analytics and Data Science Academicians in India” by Analytics India Magazine in 2018. He is a member of the Indian Science Congress Association and Indian Statistical Institute. Webinar Date and Time — April 18, 2021 (Sunday) at 5:00 PM Click here to register – https:\/\/register.gotowebinar.com\/register\/4431806330204946704","excerpt":"In this era of AI, organisations have to rely on gaining valuable insights from historical data to make strategic business decisions. And hence, without any doubt, the roles relating to business analytics have emerged as one of the most critical functions that exist in a data-driven enterprise. The surge in demand has opened new career […]","categories":["Deep Tech"],"tags":[],"author_name":"AIM Media House","publish_date":"2021-04-05T13:25:51","publication_year":"2021","word_count":305,"keywords":["data science","Go","machine learning","artificial intelligence","AI","data-driven","RAG","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","RAG","R","Go","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/register-for-this-webinar-to-understand-why-competence-in-business-analytics-is-critical-for-data-driven-business-decisions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":4130,"title":"Value Based Pricing","content":"Pricing is the tool that can either create or destroy the demand for any product. The pricing decision is one of the most critical decisions that a firm can make in the launch of a product.In modern pricing context pricing can be considered as the monitory equivalent of the value. A product’s worth depends on how it stacks up against competing products. If it is a better-than average product, it is worth more than average. Such products could be sold for a premium price. On the other hand, customers will often tolerate reduced performance if they can get the product at an economy price. Customer value measurement and accounting is a structured approach for comparing a product or service against the competition to understand its comparative strengths and weaknesses, assess its worth, and provide a rational basis for setting its price. From a marketing perspective, the goal of pricing strategy is to assign a price that is the monetary equivalent of the value the customer perceives in the product while meeting profit and return on investment goals. This white paper posits the literature review that corroborates traditional cost-based approaches to product pricing as being short-term, tactical in nature, and place the interests of the seller over the interests of the buyer. Paper on Value-Based Marketing & Pricing by Bradley T. Gale and Donald J. Swire indicates that pricing approaches based on customers’ perceptions of value are strategic and long-term in nature since they are focused on capturing unique value from each market segment through the pricing mechanism. Firms need to invest to create “pricing capital” to ensure the long-term benefits of value-based pricing. Firms that invest in a strategic pricing center can make better product decisions throughout the development process by understanding how customers value product alternatives and arrive at prices that they are willing to pay. Traditional methods of pricing: As per the paper on Value-Based Pricing For New Software Products Strategy Insights for Developer by Robert Harmon, David Raffo and  Stuart Faulk  few of the traditional methods of pricing entail cost based pricing approach containing Flat Pricing, Tiered Pricing, User based Pricing, and Usage based Pricing. This section tries to elucidate the aspects of traditional methods of pricing and compare and contrast it with the value based pricing. Cost-based pricing is historically the most popular method since it relies on more readily available information from the cost-accounting system. This data is generated as a matter of course to produce operating results, budgets, and financial statements. It is imbued with an aura of authenticity. Financial, marketing, and product managers are schooled to price the software product to yield a desired return on fully allocated costs. As per ROI Guide: Economic Value Added by J. Berry the fundamental problems with cost-driven pricing derive from the assumptions that must be made about product costs. First, unit costs are volume dependent. Fixed cost per unit is an allocated number that varies with projected volume. The allocation procedures, be they direct labor hours, or some other surrogate metric, are not very precise. Therefore, product costs are imprecise at best and a continually moving target at worst. In addition, cost-based pricing is bedeviled by a circular logic—price is based on unit-volume assumptions, but price will determine sales volume. The inability of firms to successfully model the impact of price on volume and of volume on price calls cost-based pricing into question. The circular nature of cost-based pricing can lead to overpricing in weak markets and under-pricing in strong markets. Earned value (EV) is a related cost-based concept that is used for tracking a project’s adherence to the original project budget. EV analysis focuses on explaining the cost variances between the amounts budgeted for the work and the equivalent dollar volume of work accomplished during a specified time period. The cost variances are identifiable to specific project tasks, which can then be evaluated for corrective action. Practitioners of the activity based costing (ABC) approach to determining product costs recognize that traditional cost information is not useful for firms who need to base strategic resource allocation decisions on accurate product costs. ABC methods can improve the overhead allocation process for assigning costs such as logistics, marketing, sales, production, finance, and general administration to individual products or product lines. Traditional methods of overhead allocation use direct labor hours that tend to over burden less-complex products and high-volume products. Conversely, they under allocate overhead cost to complex to low-volume products. The resulting cost distortion can produce misleading profit estimates and lead to poor decisions. A firm’s cost behavior and relative cost position in an industry are derived from the value-producing activities of the firm. Each value activity has its own cost structure. After identifying the relevant value chain, operating costs and asset costs are assigned to the product-related activities that they support. By analyzing the “cost drivers” of its value producing activities the firm can better assign costs to products and understand its true cost position. A literature review of common approaches to product pricing is as follows. Flat pricing: Users pay a fixed price for unlimited use of the product. This approach enables customers to more easily predict what they will pay for the use of the product. The fixed price is usually restricted to a particular user and\/or machine. Many consumer  offerings are priced in this manner. Some level of online support is typically built in for a set period of time. The primary drawback to this method is the lack of flexibility in customizing a price for each customer based on the value the customer requires. Some customers will have to pay more than they would like and may be motivated to seek better deals. Others will enjoy a subsidy since they would be willing to pay more for the higher value they perceive. A fixed-price strategy can be segmented to embrace discounts for large purchasers, government, and members of preferred buying organizations. Flat pricing simplifies the vendor’s pricing model since the price is set to return a dependable but fixed rate of return. Prices are based on a financial return model, not on customer value. Prices are increased when costs increase. Tiered-pricing: Tiered-pricing attempts to package product benefits according to user requirements and their willingness to pay. This approach to pricing is an attempt to link product costs to perceived customer value. Tiered-pricing is viewed more favorably when the customer can easily see the increase in value received and the pricing scheme offers desirable choices. User-based pricing: This is another cost-based pricing method that tends to benefit the vendor more than the user by maximizing license fee revenues in terms of technology services. The charge is based on the number of users that utilize a collection of service features over a given period of time. Per-user pricing: Prices to the individual user who typically can use the product on an unlimited basic for the term of the license. The price is set on assumptions about product costs and customer use. This approach typically offers one price for a specified number of users. High water mark pricing: Charges are based on the maximum number of concurrent users over a given time period. Per client pricing: Similar to per user pricing, except that the license is assigned to the designated number of users. Usage-based pricing: Customers pay only for what they actually use on a transaction basis. This model is also known as “pay-as-you-go pricing” or network-based pricing model. It is often associated with technology services. Value based pricing: On having the glimpse of traditional methods of pricing now we can move to understanding the underlying concepts of Value based Pricing. According to The Wall Street Journal, 2002-Value based pricing involves setting of a product or service’s price, based on the benefits it provides to consumers. By contrast, cost-plus pricing is based on the amount of money it takes to produce the product. Companies that offer unique or highly valuable features or services are better positioned to take advantage of value-based pricing, than companies whose products are services are relatively indistinguishable from those of their competitors. Customer value is the overall benefit derived from the product, as the customer perceives it, at the price the customer is willing to pay. At the core of perceived value pricing is the requirement that companies must first understand how the customer perceives value. Perceived value can be defined in terms of the tradeoff between perceived benefits received and the perceived price for acquiring the product or service that delivers those benefits. Different methods of value based pricing that are discussed in here are as follows penetration, skimming, and hybrid pricing strategies. Penetration pricing strategies: Penetration strategies target market segments where buyers have a high degree of price sensitivity . Price-sensitive buyers typically have low reservation prices. Delivering benefits that are perceived as industry standard at a price that is sufficiently low to generate increases in sales volume creates customer value. Certain subcomponents of penetration pricing include the following: Low-price leader (low reservation price\/competitive pricing): Low-price leaders target buyers with low reservation prices. This strategy targets the mass-market buyers with reasonable features at a low price. The competitive pricing objective recognizes that the market has reached maturity. Experience-curve pricing (low reservation price\/competitive pricing): This competitive strategy targets buyers with low reservation prices. The initial price is set below cost in order to build volume and move more rapidly down the cost curve toward profitability. The low price is intended to ramp volume quickly and to keep out competition. Bundling (low reservation price\/product-line pricing): This strategy features several applications that are packaged together and priced as a single product. It targets buyers with low reservation prices. It is a product-line strategy since it maximizes sales of complementary products within the product line. There may be differing preferences for each individual product, but overall demand is increased if the value is perceived to be greater for the bundled package. Skim-Pricing Strategies: Skim strategies target buyers that are relatively insensitive to price. All have high search costs. Some will engage in search behavior and perceive a high degree of value in the features, advantages, and benefits of the product. For example, innovators are often willing to pay more since they perceive opportunity in their ability to exploit the unique value of a new product. Others are unwilling to search and see the high price as a cue indicating high quality. Certain subcomponents of skim pricing include the following: Price signaling (high search costs\/segment differential pricing): This strategy is often used for segment differential pricing of new products where time is a primary factor in the decision process. Innovators with high search costs and a high degree of trust in the brand do not want invest heavily to evaluate product alternatives. Information about price is more easily acquired than that about quality or performance. The high price signals the benefits these buyers desire. Buyers in this situation demand a high level of service and rapid response for their continued loyalty. Reference pricing (high search costs\/competitive pricing): This competitive pricing strategy is a variant of price signaling. Buyers have high search costs and higher perceived risk than the innovators. They need a reference point to calibrate the value in the price-quality relationship. Use of a reference price strategy can benchmark the higher price of an competitor. Comparison with the higher-priced product highlights the value of the moderate priced product and vice versa. Image\/prestige pricing (high search costs\/product-line pricing): This product line strategy targets customers with high search costs who are attracted to brands that have achieved a reputation for high quality and exclusivity. The buyer’s self image is emotionally linked to the brand’s image. Buyers have expectations for exclusiveness and high levels of support and service. Hybrid Pricing Strategies: Hybrid strategies combine elements of skimming and penetration strategies. Combinations of high search costs, low reservation prices, and\/or special transaction costs may characterize potential buyers. Special transaction costs might include the complex and expensive evaluation process. Certain subcomponents of hybrid pricing include the following Complementary pricing: Complementary pricing is a product-line strategy that exploits the special transaction costs of products that are used jointly but sold separately. The base product (low reservation price\/product-line pricing) is sold at a low price that minimizes resistance to purchase. Higher profits are then made on the complementary consumable products or services due to the special transaction costs. Premium pricing. Marketers address different groups of customers by using a product-line strategy that addresses the higher search costs of some groups and the lower reservation prices of others. This practice is also known as “price lining.” The strategy is implemented by pricing versions of the product to address entry level, mid-level, and high-end premium-buying customers. Random discounting (high search costs\/segment differential pricing: A random discounting strategy maintains a high skimming price but offers discounts on a random basis as an incentive to new buyers to try the product. The price break serves to draw attention to the product. Periodic discounting (low reservation price\/segment differential pricing): Periodic discounting strategy creates customer value for sequential classes of buyers with increasingly low reservation prices. The initial strategy focuses on skimming the inelastic demand of the innovator then reducing prices on a predictable basis as the market matures in order to attract more price sensitive customer groups. Second-market discounting (differential pricing special transaction costs): For second market discounting, marketers introduce an existing product to a new market where buyers are more price-sensitive than the primary market and have identifiable special transaction costs. Techniques of value based pricing: As per the “Matching Appropriate Pricing Strategy with Markets and Objectives” by C. R. Duke techniques of devising value based pricing entail the following: 1) Customer value analysis 2) Customer value drivers. Customer Value Analysis: Customer value analysis\/ accounting is a comprehensive system of analysis that integrates whatever data is available from Importance-performance analysis based on market perception studies. Price data from competitive intelligence Engineering Economics studies. Conjoint Analyses that are tightly couple to market perception studies. Different aspects of customer value analysis include the following: Conjoint analysis or trade-off analysis: This technique enables firms to compute the consumers’ utility functions for individual variables and to understand how they are combined, traded off, and otherwise valued. Conjoint analysis is useful for pricing since the feature tradeoffs at different levels of price can be mapped. Economic Value to the Customer (EVC): Alternately called “value-in-use” or “exchange value”, EVC is the maximum amount a customer would be willing to pay for the product, assuming s\/he is fully informed about the product and competitive offerings. It is analogous to the reservation price. EVC answers the question, “What’s it worth to you?” EVC measures the life-cycle economic costs and benefits to the user of one product when compared with a reference product. Economic Value Added (EVA). Broadly stated, EVA measures a company’s net operating profit after taxes. It focuses the organization on earning a target rate of return over and above the cost of capital. Price Sensitivity Measurement (PSM): PSM models are useful for estimating market demand and for calculating the proportion of buyers that would buy the product within a specific price range. PSM determines the limits of buyer resistance over a range of prices that relate to the product’s value perceptions. These value perceptions are market segment specific and based on the buyer’s perceptions of product value, buying intentions, and spending capabilities. Typical outputs from the model are the upper and lower bounds for the acceptable price range and the optimal pricing point. PSM is very useful for pricing alternate software configurations in the early stages of development and throughout the development cycle. Customer Value Drivers: In order to create the foundation for setting prices, it is necessary that product developers and managers understand what the customer’s value drivers are and how important each is in the purchase decision. Customer value drivers are emotional links that summarize customer beliefs about the product and firm, create positive attitudes and feelings, provide the basis for differentiation, and provide the reason to buy. Value drivers are the expression of the customer’s evaluations of the product, the perceived credibility of the vendor, and the confidence the customer has in the brand. The customers’ value drivers need to be reflected in the design requirements of the products if the value is to be subsequently captured by the pricing mechanism on the product’s launch. Some of the primitive and essential customer value drivers are the following: 1. Economic value. Economic-value drivers are based on the buyer’s perceptions about the cost of acquiring, owning, installing, using, and disposing of a product or service. The concept encompasses costs savings and ROI impact deriving from the purchase of the product. 2. Performance value: Performance value is based on the buyer’s perceptions of the utility to be derived from the functional features, advantages, and benefits associated with a product or a service. 3. Supplier Value: The buyer’s perceptions about credibility of the vendor and trust in the business relationship links directly to brand acceptance. It is relatively easy for competitors to match economic and performance value by changes in price and product design. A strong brand provides a greater barrier to competition since it takes much longer to change perceptions about a company. Strong brands support skimming strategies across a broad range of pricing objectives. 4. Buyer Motivations: The buyer’s psychological motivations and goals for a particular purchase are central to the decision process. Cost-based pricing does not consider these higher-level motivations. Psychological motives arise from the buyer’s need for recognition, esteem, and belonging. Additional motivations may involve novelty seeking and knowledge acquisition. Buyer motivations are often subjective and emotional. 5. The Buying Situation: Purchase behavior always occurs within a situational context. The situation may act as a constraint or to facilitate a given purchase or it may have no effect at all. Key situational variables are: a. Task definition. The task situation addresses the question: “What objective or tasks are the products used for?” Knowledge of the specific task will help to define the software use situation and product requirements. b. Resource capability. This variable focuses on the physical and intellectual resources of the buyer including budgets, infrastructure, and technical skills. c. Time horizon. Time is an important influence on price perceptions. Buyers with short decision time horizons tend to be fewer prices sensitive. Key questions to be addressed are: “How long until the buyer is ready to make a decision? How long does the buyer anticipate using the product?” d. Social influences. What is the composition and role dynamics of the buying center team that will influence the purchase the product? e. Experience: Highly experienced buyers tend to have stronger product-related attitudes, which influence subsequent evaluations of product and price. Developers and marketers need to answer the question: “How experienced is the buyer with similar product?” f. Availability. Availability refers to the ease of finding purchase related information about the product or company. Summary: This white paper presented a literature review of contemporary cost-based pricing models. This literature review decipher that  as current markets have become more competitive and buyers are faced with more choices, cost based pricing models that ignore customer-value requirements can no longer ensure a favorable rate of return to the vendor. A taxonomic analysis of customer value drivers indicates that cost-based models appeal to price-performance value drivers with promises of improved ROI for the customer while ignoring other potentially more important value drivers that are more intangible in nature. The primary contribution of this literature review is the detailed discussion of value-based pricing strategies as they relate to the current context. The article develops a prescriptive pricing taxonomy that depicts the relationships between customer characteristics, company objectives, and pricing strategy. It suggests that deep knowledge of the customer can result in more appropriate approaches to pricing strategy.","excerpt":"Pricing is the tool that can either create or destroy the demand for any product. The pricing decision is one of the most critical decisions that a firm can make in the launch of a product.In modern pricing context pricing can be considered as the monitory equivalent of the value. A product’s worth depends on […]","categories":["Deep Tech"],"tags":["Advanced Analytics"],"author_name":"Srujana H.M.","publish_date":"2013-10-02T06:36:55","publication_year":"2013","word_count":3312,"keywords":["Go","API","TPU","ELT","AI","RAG","BERT","Rust","GAN","R","Advanced Analytics"],"extracted_tech_keywords":["AI","RAG","TPU","R","Go","Rust","API","ELT","BERT","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/value-based-pricing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10133946,"title":"AI4Bharat Invites Contributors to Chitralekha Open-Source Project","content":"AI4Bharat is seeking contributors for Chitralekha, an open-source Video Transcreation platform supported by the EkStep Foundation. The platform, initially developed for video annotation, allows users to auto-generate and edit audio transcripts in Indic languages. “We’re thrilled to invite you to contribute to Chitralekha, an innovative open-source video transcreation platform built by AI4Bhārat and funded by the EkStep Foundation,” said Ishvinder Sethi, Project Officer at AI4Bhārat. Chitralekha’s features include subtitle generation and download, audio\/video dubbing, and video translation across various Indic languages. The project, built on AI models developed in-house by AI4Bhārat, is open to enhancements and new feature integrations. The initiative offers contributors the chance to gain experience in a live project environment, build their portfolios, and contribute to the Indian Language AI landscape. AI4Bhārat encourages participation in community discussions on the platform’s GitHub page, focusing on potential use cases and desired features. This is an unpaid opportunity for open-source enthusiasts to help expand Chitralekha’s capabilities. Chitralekha is an open-source platform designed for video transcreation across various Indic languages, leveraging machine learning models for ASR (Automatic Speech Recognition) for transcription, NMT (Neural Machine Translation) for translation, and TTS (Text-to-Speech) for voice-over. The platform supports multiple input sources, such as YouTube and local files, and offers various options for transcription generation, including models, source captions, custom subtitle files, and manual creation. For translation and voice-over generation, users can utilise models or manually created content. Currently, Chitralekha supports voice-over for single-speaker videos, with multi-speaker support under development.For more details or to contribute, visit the GitHub page here or try out the platform here.","excerpt":"The initiative offers contributors the chance to gain experience in a live project environment, build their portfolios, and contribute to the Indian Language AI landscape.","categories":["AI News"],"tags":["AI4Bharat"],"author_name":"Siddharth Jindal","publish_date":"2024-08-28T17:44:45","publication_year":"2024","word_count":261,"keywords":["machine learning","programming_languages:R","AI","Git","RAG","ViT","AI4Bharat","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Git","GitHub","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai4bharat-invites-contributors-to-chitralekha-open-source-project\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001319,"title":"5 Budget Workstation Builds Available  Under ₹45,000","content":"The PC market is seeing a renaissance of sorts, as market leaders Intel and Nvidia are finally seeing competition from their rival, AMD. This has led to the announcement of new components in a bid to outdo their rivals. In a highly competitive market, the customers are always the ones to benefit. Now, with the market full of components from the low-end to the premium end, consumers can choose precisely which product they want at which price point. Keeping this in mind, here are 5 budget, low-end PC builds under ₹40,000. As we move up the spectrum, performance gains are also visible, allowing for the fine-tuning of exactly which components to purchase. The Lowest End The price range for this build is set at ₹24,000 for the whole setup, a steal for the performance it brings to the table. The heart of this PC is the underrated, dual-core Pentium Gold G5400. It makes up for its low core count by implementing hyper-threading and having 4 threads running at 3.7 GHz. It is priced at ₹7200, making it one of Intel’s lowest end offerings. The CPU is paired with a low-end LGA1151 motherboard, the Gigabyte GA-H110M, priced at ₹4500. The setup is completed with 1 stick of Corsair Vengeance LPX 8GB RAM for upgradeability, and a Cooler Master 450W PSU. These are priced at ₹4800 and ₹2500 respectively, with the Antec VSK3000 case, priced at ₹2000, providing a home for the components. Every PC needs a hard drive, and this is no different. After adding on the 1TB Seagate Barracuda disk, priced at ₹3000, we have the PC ready to go. This is a PC that can be used for web browsing, watching HD videos and other similar, work-oriented tasks. It also functions quietly, making it ideal for a dedicated home theater PC due to the presence of the UHD 610 iGPU. One Step Higher This build has the i3 8100 as its heart, an 8th Gen processor running at 3.6 GHz on a quad-core architecture. The rest of the build remains fairly the same, with the motherboard, RAM, HDD and PSU carrying over for a price of ₹14800 combined. When added to the Antec NX100 Mid Tower Case, which comes with a plastic side panel, the price comes up to ₹27,900. This build comes with a 25% performance boost over the Pentium PC due to its higher core count and burst speed. While this still sits in the lower end of the spectrum, this build is deceptively powerful. The presence of the fast RAM and the additional cores will make a difference in the performance of the PC. It can be used for light video editing, encoding tasks, as a secondary PC for streaming and watching 4K footage. The All-In-One, Middle Of The Pack Contender This build switches things up by playing for team Red. It uses the Ryzen 5 2400G APU at its heart, and even gives users an opportunity to get into gaming on the cheap. The processor has 4 cores and is clocked at 3.6GHz. The CPU provides gaming performance similar to a low-end dedicated GPU, showing a marked improvement over the UHD 610 and 630 iGPUs found in the first two offerings on this list. The processor is priced at ₹16700, and the accompanying Asrock A320M-HDV motherboard is priced at ₹3700. Due to the CPU using RAM as VRAM, it is recommended to buy a higher amount of high speed RAM. This comes in the form of a 4GB stick added to the Corsair Vengeance 8GB, bringing the total for RAM up to ₹7500. The PC will run on a Cooler Master VS450 PSU worth ₹2,500, and will come in the Antec NX200 case, which is set at ₹2,500. Adding on the hard disk puts the price of the build at just under ₹32,000. An additional ₹2,000 can be saved if the user opts to go with 8GB of RAM instead of 12GB. Along with being usable for light gaming, the PC can also be used to perform encoding tasks and even rendering lower-end HD footage. The presence of the iGPU and the added RAM will also contribute to the future proof nature of this build. Mid-Range Workstation With A Gaming Boost This build comes with the AMD Ryzen 5 2600 CPU, which was considered one of the best budget CPUs when it launched and has held on to its position since. It is priced at ₹16,700 and functions on the same chipset as the 2400G, allowing users to get the cheaper motherboard for maximum savings. It boasts of a 6 core, 12 thread setup with a clock speed of 3.4GHz. With this performance boost, it is only natural to boost up the RAM as well. 2 8GB sticks of Corsair Vengeance at 3000MHz will serve to further boost the performance. One thing to note is that the CPU does not have integrated graphics, leading to a dedicated GPU to be bought for this PC. The best choice for non-gamers is the GT 710, priced at ₹3100. It is the lowest end card on the market, and will serve the purpose of this build. The total cost of the build is pegged at ₹41,700, with close to ₹5,000 being saved if the user decided to opt for an 8GB setup. However, this PC is a beast, and will perform very well in settings that require a lot of processing power. This includes use cases such as running virtual machines, heavy video editing, streaming HD video from the machine, and even gaming with the lower end card. The dedicated GPU still functions adequately in many older titles, and performs significantly better than the UHD 630 iGPUs in the Intel processors. Team Blue’s Retort To The Mid Range Intel’s reply to the Ryzen 2600 comes in the form of the Intel i5 8400. However, the reason for Intel falling behind in CPU sales is seen clearly with the positioning of this processor. The 8400 comes with 6 cores as well, albeit without support for hyperthreading. Notwithstanding, the card still has a slight performance increase over the 2600, and comes with an integrated UHD 630 GPU. This PC will also have the Gigabyte motherboard, along with 16GB of Corsair RAM. Paired with the Cooler Master Master Force 500 case, priced at ₹3000, the price of the completed build comes up to ₹41,900, just a little bit more than the Ryzen build. However, it is important to note that the Ryzen build comes with a dedicated GPUm whereas the Intel one relies on the integrated GPU to perform. The build will perform similar tasks to the Ryzen build as it has comparable compute power. However, it will fall shorter in multi-threaded workloads, instead relying on its higher clock count to make up the difference. This PC can be used to run video editing software, encode or decode HD video, and run multiple virtual machines simultaneously. Last Word As mentioned, there is a lot of choice for the consumer looking to get into the CPU market. This is due to the golden age of competition that is going on currently, with newer CPUs slated to launch in a few months from AMD and Intel’s side. This will shake up the already disrupted CPU market.","excerpt":"The PC market is seeing a renaissance of sorts, as market leaders Intel and Nvidia are finally seeing competition from their rival, AMD. This has led to the announcement of new components in a bid to outdo their rivals. In a highly competitive market, the customers are always the ones to benefit. Now, with the […]","categories":["AI Trends"],"tags":["GPU","graphics"],"author_name":"Anirudh VK","publish_date":"2019-02-27T21:36:15","publication_year":"2019","word_count":1213,"keywords":["CUDA","Go","programming_languages:R","AI","ETL","programming_languages:Go","graphics","R","GPU"],"extracted_tech_keywords":["AI","CUDA","R","Go","CUDA","ETL","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-budget-workstation-builds-available-under-%e2%82%b945000\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039507,"title":"Why Is There A Shortage Of MLOps Engineers?","content":"MLOps, or machine learning operations, is emerging as one of the hottest fields. In the last four years, the hiring for machine learning and artificial intelligence roles has grown 74% annually. MLOps engineer, also known as DevOps for machine learning, covers a whole gamut of machine learning tasks starting from data integration to training and managing infrastructure to deploying. Source: Arrikto (MLOps pipeline) DevOps Vs MLOps MLOps and DevOps engineers require different skill sets. Firstly, developing machine learning models do not need a software engineering background as the focus is mainly on the proof of concept\/prototyping. Secondly, MLOps are more experimental in nature compared to DevOps. MLOps calls for tracking different experiments, feature engineering steps, model parameters, metrics, etc. MLOps is not limited to unit testing. Various parameters need to be considered, including data checks, model drift, analysing model performance, etc. Deploying machine learning models is easier said than done as it involves various steps, including data processing, feature engineering, model training, model registry and model deployment. Lastly, MLOps engineers are expected to track data distribution with time to ensure the production environment is consistent with the data it is being trained on. The AI reality Last year, AI\/ML research hit the doldrums in the wake of the pandemic; tech giants like Google slowed down hiring AI researchers and ML engineers, and Uber laid off their AI research and engineering team. According to a 2020 Stanford study, the AI sector saw an increase in private investments despite the pandemic. China surpassed the US in scholarly work on AI the same year, the report added. The demand for data scientists and machine learning engineers is at an all-time high. Machine learning engineers topped LinkedIn’s Emerging Jobs ranking, with a recorded growth of 9.8 times in five years (2012-2017). Similarly, data scientist jobs witnessed an 8.5x surge. Source: LinkedIn (Top 20 Emerging Jobs) MLOps engineers: Short in supply MLOps engineering, being a fledgling field, is witnessing a shortage of experienced professionals. Other factors for the shortage include: Lack of clarity in role and responsibility of MLOps engineer at the organisational level, especially startupsMultiple platforms and tools to learnShortage of dedicated courses for MLOps engineers Scribble Data CEO Venkata Pingali said companies are facing attrition in MLOps staff. “Ability to handle the loss of staff and associated knowledge is becoming a key requirement to design systems and processes,” said Pingali. We asked Pingali how companies could deal with attrition, he said whatever MLOps engineers build should be person independent, documented, and reproducible. Moreover, the companies should keep evaluating individual staff. Enterprises had to downsize at the beginning of Covid breakout, which hampered the deployment of AI projects globally. Plus, companies are still stuck with the traditional method while interviewing an MLOps engineer. “Companies should ignore the resume and should start giving realistic but simplified hands-on assignments to see how the potential staff handles problems,” said Pingali. An ideal MLOps engineer should have good discipline, architectural thinking, and tooling agility. Also, experience matters. Who can pivot? A software engineer can easily transition to the MLOps engineer role. However, they need to understand the nuances of data science. For this, software engineers’ mindset has to change – not necessarily the tooling. That is hard, Pingali said. Essential skills for MLOps engineer MLOps engineers need a strong foundation across the machine learning pipeline, tools, framework, and processes to deploy machine learning models throughout the project’s lifecycle systematically. The core skills a machine learning operations (MLOps) engineer should possess are: Expertise in cloud architecture\/DevOps to recommend enterprise-grade solutions for operationalising AI analytics. An in-depth understanding of cloud platforms (AWS, Azure, etc.), AI lifecycle, and business problems to develop end-to-end (data\/Dev\/ML) Ops pipelines.Experience in project governance, alongside customer-facing experience of discovery, assessment, execution and operations.","excerpt":"MLOps, or machine learning operations, is emerging as one of the hottest fields. In the last four years, the hiring for machine learning and artificial intelligence roles has grown 74% annually. MLOps engineer, also known as DevOps for machine learning, covers a whole gamut of machine learning tasks starting from data integration to training and […]","categories":["AI Highlights"],"tags":["machine learning jobs","ML jobs"],"author_name":"Amit Naik","publish_date":"2021-05-04T12:00:00","publication_year":"2021","word_count":627,"keywords":["machine learning jobs","ML jobs","data science","machine learning","artificial intelligence","AWS","AI","ML","MLOps","analytics","model registry","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","MLOps","model registry","AWS","Azure"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/why-is-there-a-shortage-of-mlops-engineers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":38342,"title":"5 Upcoming Academic Conferences AI Researchers Should Not Miss","content":"Image source: pbs.twimg.com As the development in the field of AI and analytics continue to witness an upsurge in India, the AI ecosystems in the country is bustling with activities. Over the last couple of years, research avenues and the number of competitions for students and AI scholars have increased steadily across the country. In this article, we take a look at the upcoming paper research submissions and how you can be a part of it. 1. Cyber Security Paper Presentation Contest, IIT Madras About: The competition aims to bring Cyber Security Professionals and students from all over India who are interested and actively involved in the cyber security research activities. Though the paper is largely concentrated on cybersecurity, the paper presentation domain include machine learning, data science, artificial intelligence among other topics. Topics for presentation: Machine learning, artificial intelligence, data science Who can participate: UG\/PG students, research scholars, cyber security professionals and   teaching faculties Last date of abstract submission: 31st May 2019 Final Paper submission:  12th June 2019 Paper presentation date: 22nd June, 2019 Registration deadline: 31st May 2019 Venue: IC & SR Auditorium, IIT Madras Chennai Register here 2.National Conference on “Machine Learning and Artificial Intelligence”,  hosted by IIMB in association with  Coimbatore Institute of Technology (CIT) About: The two day conference includes plenary sessions, invited talks and paper presentations focusing on the applications of ML and AI Area of focus: Field of textiles, engineering, healthcare, agriculture, business, social media and other relevant domains Who can participate: Students, researchers, academicians and corporate professionals Last date of abstract submission: 19 May, 2019 Final Paper submission: 14 July, 2019 Paper presentation date: 26-27 August, 2019 Registration deadline: 21 July, 2019 Venue:Coimbatore Institute of Technology Register here Image source: pbs.twimg.com 3.International Conference on Artificial Intelligence, Machine Learning and Big Data Engineering, Vijawada About: The conference aims to present the latest research and results of scientists related artificial intelligence, machine learning and big data engineering topics. This conference provides opportunities for the different areas delegates to exchange new ideas and application experiences face to face, to establish business or research relations and to find global partners for future collaboration. Who can participate: Students and working professionals Topics for presentation: AI Algorithms, Artificial Intelligence tools & applications, automatic control bioinformatics Natural Language Processing etc and Big-data engineering Last date of Application: May 10th, 2019 Paper presentation date: May 26th, 2019 Registration deadline: May 14th, 2019 Venue: Hotel Midcity, Vijawada Register here 4.ACN- International Conference on Artificial Intelligence, Robots and Mechanical Engineering(ICAIRME), Bengaluru About: The conference aims to provide opportunities for the global participants to share their ideas and experience in person with their peers expected to join from different parts on the world Topics for presentation:  Computer vision and speech understanding, data mining and machine learning tools, fuzzy logic, heuristic and AI planning strategies and tools, computational theories of learning Who can participate: Academicians, researchers, engineers, industrial participants and budding students Last date of Application: May 8th, 2019 Paper presentation date: May 26th, 2019 Registration deadline: May 14th, 2019 Where is it happening: Bangalore, Karnataka, India Register here 5.International Conference on Artificial Intelligence, Machine Learning and Big Data Engineering, Bhopal About: The conference aims to provide opportunities for the global participants to share their ideas and experience in person with their peers expected to join from different parts on the world Topics for presentation: AI Algorithms, Artificial Intelligence tools & Applications Automatic Control ,Bioinformatics, Natural Language Processing etc Who can participate:The conference will bring together leading academic scientists, researchers and scholars in the domain of interest Last date of Application: May 17th, 2019 Paper presentation date: June 1st, 2019 Registration deadline: May 21st, 2019 Where is it happening: Bhopal Register here","excerpt":"As the development in the field of AI and analytics continue to witness an upsurge in India, the AI ecosystems in the country is bustling with activities. Over the last couple of years, research avenues and the number of competitions for students and AI scholars have increased steadily across the country. In this article, we […]","categories":["AI Trends"],"tags":["Research"],"author_name":"Akshaya Asokan","publish_date":"2019-04-26T13:02:31","publication_year":"2019","word_count":618,"keywords":["data science","Go","artificial intelligence","machine learning","AI","ML","computer vision","Aim","Research","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","data science","analytics","Aim","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-upcoming-academic-conferences-ai-researchers-should-not-miss\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10090941,"title":"Ex-Intel Exec Raja Koduri Gives a Sneak Peek into his New Startup","content":"Intel Corporation made a significant move in December 2022, splitting its Accelerated Computing Systems and Graphics (AXG) business unit into two distinct entities, namely, Data Center and AI (DCAI) business unit and AXG business unit.The decision was aimed at aligning the company’s graphics efforts to better compete with industry heavyweights—NVIDIA and AMD. The AXG business unit’s consumer-oriented portion was slated to merge with Intel’s Client Computing Group (CCG), which specialises in developing platforms around the company’s central processing unit (CPU) products. Meanwhile, the teams responsible for managing data centre and supercomputing graphics processing units (GPUs), including the highly anticipated Ponte Vecchio and Rialto Bridge products, will be transferred to the DCAI business unit.Raja Kodouri, the former head of the AXG division, was to assume the role of Chief Architect, leading the charge in expanding the company’s efforts in the CPU, GPU, and AI domains as well as accelerating critical technical programmes.In a surprising turn of events, Raja Koduri announced his resignation from Intel Corporation in March 2023, only three months after being appointed as the Chief Architect. This announcement comes at a crucial time for Intel as it prepares to face off against its primary competitor, NVIDIA. Koduri’s departure has caught many industry experts off guard, as he was given a significant position at Intel merely months before his resignation. To gain insight into his decision and what lays ahead, Analytics India Magazine got in touch with him. Challenges Intel faced Koduri has been with Intel since 2017, the year when Intel decided to get into GPUs for the first time in two decades. When asked what were the various challenges he faced competing with NVIDIA while being at Intel, he said that the biggest challenge lies in their (NVIDIA’s) proprietary software ecosystem and the reach it has. “Their products are ubiquitous, and they do an excellent job of hooking customers,” says Koduri, “Furthermore, the other major challenge is execution, and specifically, proper execution”. As per Koduri, Intel experienced major issues with the 10-nanometer chip production that brought the entire product pipeline to a standstill. He says that “the situation was akin to a clogged drainage system”. The most significant hurdle was that it took a considerable amount of time to resolve this. “The process of clearing out the old products to make way for the new ones, along with innovative ideas from pioneers like me and Jim Keller, and new architectures, was all hindered by these drainage execution problems. Nevertheless, you’ll start to notice these issues gradually clearing up. Intel has already launched new products, and I anticipate some remarkable developments from them in the next three to four years,” shares Koduri. Why he left However, when all things were sorted out, why did he choose to leave? As per Koduri, he had a major surgery in December and had to take a medical leave due to which Intel decided to give the responsibility of heading the business to someone else. “When I took the break, and you know, when you have that space in mind to think about, I made a decision that I want to do something outside, particularly, in the open ecosystem,” says Koduri. Koduri added that he wanted to get back hands on to do more software for himself. It’s been a long time since he was actually proactively coding anything. So, he wanted to learn everything since he feels like he hasn’t learned a lot about what’s happening on software. However, Koduri believes that he can leverage his knowledge of hardware. As per him, he can utilise his knowledge on hardware to help other companies. “I don’t need to be building hardware myself,” says Koduri. He shared that he is now advising companies Tenstorrent, Bodhi Computing, and others. “I’ll be advising several hardware startups as well, on their roadmaps and all,” says Koduri. What’s next? After leaving Intel, there was news that Koduri is going to work on a generative AI gaming startup. Upon inquiring what it was and how he plans to move ahead with it, Koduri says that he had some ideas way back in 2017 as well. “But now with the whole progress on large language models, diffusion models, transformers and all that, the ideas that I’ve had in 2017 are very much a reality now,” says Koduri. Koduri says he wants to build a platform that connects all these tools, with the best hardware there. He believes that today, the models are not working in real-time. “So, if you want a real-time gaming use case, you want to deliver AI-generated content at 60 frames per second. Like how would you go about doing that?” explains Koduri, “The problem hasn’t yet been solved”. According to reports, Raja Koduri is in negotiations with Hiranandani-backed data centre operator ‘Yotta’ for a deal for his generative artificial intelligence startup, which he claims will have a significant presence in India.As a startup, Koduri accepts that he doesn’t know how to solve these problems, “But, I have a few smart people who will figure it out. We may fail, but we will learn from it,” believes Koduri. ‘Don’t do computer science’: Raja Koduri When asked why innovation is not happening at the rate humanity anticipated in the 1980s, Koduri believes that the sole reason why we couldn’t deliver the expectations is computer science. “Let’s talk about flying cars,” says Koduri, “You have to solve fundamental physics problems, mechanical engineering problems. They’re real hard problems.” But Koduri believes when there is so much easy money to be made, writing an app for getting food delivered to my home can make millions of dollars, why would anyone put brainpower into the hard problems? “I would like to see people coming back to fundamental problems,” says Koduri.  “In fact, I advise, nobody should join a computer science course. It’s a waste of time, in my opinion,” shares Koduri. Koduri believes that if someone is going to university, they should go to electrical engineering, mechanical engineering, fundamental chemical engineering, where they get access to machines, get access to equipment that allows one to experiment with the physical things that one can’t otherwise sit at home and learn. “I strongly believe in encouraging fundamental research in India and applaud those who are already dedicated to this pursuit. While I appreciate software and hardware startups that focus on innovative products, such as a smaller and longer-lasting battery, I also want to prioritise supporting startups that aim to develop cutting-edge technology like body computing chips. It’s crucial that we foster a culture of innovation in India, particularly in areas that tackle complex issues like heart disease,” shares Koduri.","excerpt":"I don’t think people should do computer science: Raja Koduri","categories":["Deep Tech"],"tags":["AMD","Intel","NVIDIA"],"author_name":"Lokesh Choudhary","publish_date":"2023-04-06T12:15:00","publication_year":"2023","word_count":1105,"keywords":["Go","AMD","artificial intelligence","AI","Transformers","diffusion models","RAG","Aim","analytics","generative AI","NVIDIA","R","Intel"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","generative AI","Aim","Transformers","RAG","R","Go","diffusion models"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/raja-koduri-gives-a-sneak-peek-into-his-new-startup\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":65957,"title":"Opportunities &#038; Challenges Of Conducting Exams Through AI-Based Proctoring In India","content":"Online learning may have been on the rise in recent years, but it has become a necessity with many places still under extended lockdowns amid the Covid-19 pandemic. As education undergoes a tremendous transformation with emerging technologies enabling digital learning, a natural progression may be to explore the potential of a segment that has not found large scale adoption yet — virtual exams. With numerous exams being postponed or conducted amid massive health risks to students, remote proctoring can enable them to take exams in the safety of their homes. What is more, with AI-based invigilation technologies ensuring that they do not cheat or indulge in unfair means during the assessment, educational institutions can also benefit from this arrangement. “We have arrived at an era where the adoption of online learning and the implementation of tech-driven education is the only way forward,” says Arjun Mohan, India CEO of upGrad. “With AI tackling the challenges of the offline model, especially amid such crucial times, it is sure to be a gamechanger and prepare students and educators for the post Covid-19 world,” he adds. Using a combination of manual and AI-based technologies, remote proctoring offers several benefits. While it enables students to take a test from any location with certain technical prerequisites, it also eliminates the need for physical examination centres. It is cost-effective and is easy to schedule, both of which can be difficult to manage, especially during competitive exams. “AI-proctored exams are a great way to ensure that the lockdown doesn’t mar the aspirations of students and that the academic calendar is followed in due time,” says Randhir Kumar, founder, of BasicFirst Learning. Adds Shweta Doshi, co-founder of career-focused edtech platform GreyAtom, “It’s an exciting time for educators to challenge the status quo on how we assess our learners.” AI-Based Proctoring – How It works While the finer details may differ with various niche offerings by service providers, the overall modus operandi remains the same. Students are provided with a link, which when activated, shifts the laptop or computer into a fortified browsing mode, which secures the screen and keyboard to restrict movements beyond the contained space. Here, AI is trained to assist in proctoring, where it empowers the system to be mindful of candidates’ movements and take note of any discrepancies in activities –  be it shifty eyes or body movements, recording if they are gazing somewhere repeatedly or interacting with someone, and even documenting how many times they are pushing the Alt+Tab keys to switch to different windows. If a student is caught cheating through any of these means, the same can be flagged to authorities for subsequent action. This will reduce the chances of cheating among students. For instance, Mercer Mettl offers a suite of technologies to users. Its key features include candidate authentication and AI-based proctoring. Candidate authentication involves photo and ID card verification, as well as a quick authentication using OTP. Students are required to enter their registration details, get a picture clicked using the webcam, and verify their ID proof to eliminate any risk of impersonation. Its AI-based proctoring solution involves facial detection, detection of mobile phones, and more, to monitor candidates behaviour for any suspicious activity. AI-powered proctoring solutions can be used for conducting all sorts of assessments – be it university entrance exams, semester evaluations in schools, or even certification tests for online courses. In fact, Scaler Academy has been using AI-based proctoring tools to conduct online tests for its six-month intensive computer science course. “AI-based proctoring tools have the potential to be much more effective and accurate than human intervention in any assessment,” feels Abhimanyu Saxena, co-founder of Scaler Academy. “It is impossible for human invigilators to ensure 100% accuracy when it comes to students not employing cheating or copying tactics. That is not the case with AI, which can track and flag something as small as shifty eye movements, which can emerge from the use of a cell phone or an alternative screen while giving a test,” he adds. In the Scaler Academy entrance test, the entire length of the assessment is monitored by proctoring tools that record everything. Post the exam, all the solutions submitted are run through proprietary proctoring tools that check plagiarism as well – not just the text, but also the semantic flow of solutions. In fact, according to Ankush Singla, co-founder of Coding Ninjas, AI proctoring tools are being toyed as part of assessments conducted on potential employees during recruitment processes as well. “In fact, many firms are using software like Mercer Mettl, ProctorU, Examity, Verificient, AI Proctor, and more, for an accurate and error-free evaluation of exams,” he says. Challenges & Opportunities There may be several benefits of AI-driven online examinations, but it carries several prerequisites, including a functional computer or laptop, high-speed internet connection, and several accessories including a webcam and speakers. Although many people may have such an arrangement at home, scaling this tool entails that a wider population uses it as well. “In a country like India, the looming challenges for students in remote areas with poor or no internet connectivity and limited infrastructure can restrict their access to this model, as a disrupted internet connection may not allow students to login in at the same time and complete the exam,” says upGrad’s Mohan. “There could also not be enough bandwidth for the webcam application to work optimally, thereby leading to more complicated technical challenges,” he adds. Concurs BasicFirst Learning’s Kumar, “Even though innovations are underway for students to take virtual exams, numerous challenges remain before they can become mainstream in India. Many students either don’t have a proper internet connection at home or the means to take the exams online or sometimes both.” Adds Ashutosh Kumar, co-founder of Testbook, “Although the rising trend of online proctoring has been visible in other countries, India still has a few challenges to surmount before it can be made mainstream.” Therefore, an improvement in accessibility can overall improve the results of the AI proctoring model and ensure maximum usage across all levels. However, despite the challenges, GreyAtom’s Doshi sees an opportunity here to personalise learning and not just use proctoring platforms as a tool for better-conducted tests. “I strongly feel this an opportunity to go beyond, and design tests where AI is evaluating learners not just at one point in time, but consistently over a period, giving constant feedback for improvement, along with a path and recommendation to improve,” says Doshi. “The systems must be designed to capture data from the beginning, which can then make possible all kinds of personalisation and prediction experiments,” she adds. According to her, proctoring tests is just one part of the larger problem statement on how to evaluate learners on an ongoing basis, and can instead, be used as a tool to diagnose a learner’s gaps in understanding to get better outcomes. She explains: “As a learner attempts more tests, based on the behavioural, cognitive signals, the system updates its model of the learner’s understanding and adjusts the curriculum accordingly. As more learners use the system, it spots previously unreleased connections between concepts. The ML algorithms then update the relationship in the knowledge graph to take these new connections into account, identify learner gaps, recommend the right courses and finally in newer tests, help the learners perform better.”","excerpt":"Online learning may have been on the rise in recent years, but it has become a necessity with many places still under extended lockdowns amid the Covid-19 pandemic. As education undergoes a tremendous transformation with emerging technologies enabling digital learning, a natural progression may be to explore the potential of a segment that has not […]","categories":["AI Features"],"tags":["edtech","facial detection software"],"author_name":"Anu Thomas","publish_date":"2020-05-26T18:00:00","publication_year":"2020","word_count":1220,"keywords":["Go","programming_languages:R","AI","innovation","ML","edtech","Git","programming_languages:Go","facial detection software","ViT","R"],"extracted_tech_keywords":["AI","ML","R","Go","Git","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/opportunities-challenges-of-conducting-exams-through-ai-based-proctoring-in-india\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10012325,"title":"Can TensorFlow’s New Face Landmarks Model Improve Iris Tracking In Mobile Devices?","content":"Open-source machine learning platform TensorFTlow has announced that it would be adding iris tracking to its face mesh package. The iris tracking has been added to this package through the TensorFlow.js face landmark detection model. It must be noted that the face mesh package was introduced in TensorFlow.js earlier this year in March. This package uses a single camera to derive approximate 3D facial surface geometry from an image or a video stream, without even a depth sensor. It was introduced with an ability to locate eyes, nose, lips (along with lip contours), and facial silhouette. With the addition of iris tracking now, it would be possible to detect eye movements, including blinking. It is implemented through the MediaPipe iris model, which is again an open-source ML model, with no requirement for additional hardware. Iris Landmark Tracking Iris tracking, especially on mobile devices, is a challenging task to perform. There are several hurdles to overcome, such as constrained computing resources, presence or occlusion such as hair strands or squinting of eyes, and even variable light conditions available. Most often, separate hardware is deployed for this function. This hardware includes expensive headsets and remote eye-tracking systems. Its high cost is not the only problem of deploying hardware; given how bulky they are, it becomes for usage with mobile devices unfeasible. However, with the new announcement from TensorFlow, the users will be able to upgrade to the new face landmark detection model by just making a few code changes and no additional hardware installations. This package can be installed in two ways: by using script tags or by using NPM. Further, this new model offers three significant improvements: iris keypoints detection, better eyelid contour detection, and improved detection for rotated faces. As discussed previously, this package cancels the need for having separate hardware, thereby establishing compatibility with mobile devices. The fact that this is a lightweight package that contains only 3MB of weights makes it even more suitable for real-time interference on mobile devices. Further, with TensorFlow.js, a user could choose between a variety of different backends such as WebGL and WebAssembly (WASM) with XNNPACK for devices with lower-end GPUs. As part of future enhancements, the TensorFlow.js and the MediaPipe teams would now be adding depth estimation capabilities to the face landmark detection using improved iris coordinates. The teams will also make the code available for facilitating reproducible research and developer community’s further usage. The full blog from TensorFlow can be found here. About MediaPipe Iris MediaPipe Iris was announced in August this year, as a machine learning model for accurate iris estimation built for use on modern mobile phones, desktops, laptops, and over the web. Along with tracking iris landmarks involving iris, pupil, and eye contour, this model also showcased that ability to determine the metric distance between the subject and the camera. It demonstrated an error rate of just 10% without the use of a depth sensor. This was ensured by relying on the fact that the horizontal iris diameter of the human eye remains constant across different populations along with some simple geometric arguments. The model was built upon the previous work on 3D Face Meshes from which the eye region of the original image was isolated for use in the iris tracking system. The problem was broadly divided into two parts — eye contour estimation and iris location. A multi-task model that consists of a unified encoder with different components for each task was designed for using task-specific training data. This model was trained upon manually annotated 50,000 images which described a variety of illumination conditions, and head rotation poses from diverse regions. A detailed blog on this can be read here.","excerpt":"Open-source machine learning platform TensorFTlow has announced that it would be adding iris tracking to its face mesh package. The iris tracking has been added to this package through the TensorFlow.js face landmark detection model. It must be noted that the face mesh package was introduced in TensorFlow.js earlier this year in March. This package […]","categories":["AI Features"],"tags":["Tensorflow"],"author_name":"Shraddha Goled","publish_date":"2020-11-25T16:00:12","publication_year":"2020","word_count":614,"keywords":["API","machine learning","programming_languages:R","AI","ML","ai_frameworks:TensorFlow","TensorFlow","R","Tensorflow"],"extracted_tech_keywords":["AI","machine learning","ML","TensorFlow","R","API","ai_frameworks:TensorFlow","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-tensorflows-new-face-landmarks-model-improve-iris-tracking-in-mobile-devices\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10137968,"title":"From Finance to Customer Service, These Top 10 AI Agent Drive Efficiency","content":"Venture capitalist Vinod Khosla recently predicted that most consumer interactions online will involve AI agents handling tasks and filtering out marketers and bots. However, this raises a fundamental question: How are enterprises using AI agents to solve problems? Listed below are ten AI agents used in enterprise automation by different companies. UiPath Platform Developed by UiPath, this agent specialises in robotic process automation (RPA), enabling businesses to automate repetitive tasks across various applications like data entry, customer support, and financial processes. It is incorporated by companies like GE Healthcare to streamline medical record management and Wells Fargo for financial operations. Power Automate Developed by Microsoft as a part of Microsoft’s Power Platform, this agent helps automate repetitive tasks with AI integration and extensive app compatibility like email handling, notifications, and data updates. It’s used by companies like Toyota to automate supply chain notifications and Shell to streamline HR workflows. Automation Anywhere Developed by Automation Anywhere Inc, this agent offers an RPA platform with AI bots for end-to-end automation across different workflows and applications. It’s used by companies like Coca-Cola for supply chain optimisation and Cisco for automating customer services. WorkFusion for RAP and Intelligent Automation Developed by WorkFusion, this agent combines RPA with AI to streamline operations in sectors like finance and healthcare. Deutsche Bank uses it to automate transaction processes and UBS employs it for document compliance automation. Blue Prism Developed by the Blue Prism Group, this agent focuses on digital workforce automation, providing an AI-based RPA solution for large scale automation, helping with compliance, data processing and customer service. DHL uses Blue Prism for logistics and order processing, and Coca-Cola uses it for supply chain automation. NICE Robotics Automation Developed by NICE Ltd, this agent offers AI automation tools tailored for customer service, workforce optimisation and back office operations. Companies like AT&T integrate NICE automation for call centre management and Allianz employs it to automate insurance claim processing. SAP Intelligent RPA Developed by SAP, it is a part of their Business Technology Platform aimed to offer process automation tailored for SAP ecosystems by automating ERP processes like finance, sales orders, and inventory management. Siemens uses SAP RPA for managing procurement processes, while BMW employs it to streamline logistics. Pega RPA Developed by Pegasystems, this agent is integrated with CRM and BPM systems, emphasising case management and customer service automation. It’s used by Bank of America for automated customer onboarding, while AIG uses it to handle insurance claims processing. Appian RPA Developed by Appian Corporation, this agent merges RPA with low code capabilities to automate workflows and integrate seamlessly with Appian’s low code automation platform. Companies like T-Mobile use Appian RPA for customer service workflows and General Electric incorporates it for data processing tasks. AutomationEdge Developed by AutomationEdge Group, this agent is an intelligent automation platform that integrates RPA, AI and IT process automation for end-to-end digital transformation. It offers automation for both IT and business processes, including ticketing, email processing, and data extraction. This agent is used by American Express for customer support and Wipro integrates it for IT process automation. AI agents are in high demand in the industry. In the recent past, Oracle, Salesforce, and Microsoft have embraced AI agents as part of a larger trend in autonomous business functions. Oracle introduced over 50 AI agents within its Fusion Cloud Suite, targeting functions like HR and finance. Salesforce unveiled its Agentforce Partner Network, collaborating with companies like NVIDIA to expand AI-driven capabilities. Microsoft enhanced its Copilot agents to streamline workflows via integration with tools like SharePoint. This push highlights the potential of AI in reducing manual tasks and enhancing productivity across enterprises. With respect to streamline enterprise automation, the above mentioned few can be best suited for finance and HR activities. On the other hand, OpenAI has partnered with T-Mobile to develop AI customer service agents through a new platform called IntentCX, which uses OpenAI’s latest model to analyse customer interactions and optimise service responses. The platform aims to improve customer experiences by processing feedback and refining responses, with a full launch anticipated in the coming year. Additionally, T-Mobile is working with NVIDIA to further integrate AI in wireless networks, enhancing overall communication systems.","excerpt":"Most consumer interactions online will soon involve AI agents handling tasks and filtering out marketers and bots.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","AI Agents","AI in finance"],"author_name":"Tanisha Bhattacharjee","publish_date":"2024-10-09T18:00:15","publication_year":"2024","word_count":698,"keywords":["Go","API","OpenAI","AI","ML","digital transformation","AI Agents","Git","AI in finance","Aim","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","R","Go","Git","API","ViT","digital transformation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/from-finance-to-customer-service-these-top-10-ai-agent-drive-efficiency\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004950,"title":"How Is Voot Using Computer Vision To Scale-up Their Offerings","content":"One of the interesting areas where computer vision is being explored is the video streaming platform. Video streaming has become one of the most important fields in the last decade, especially with the lockdown that has resulted in an increased viewership on these platforms than ever before. To boost the user experience a lot of these companies are using computer vision (CV). Addressing the attendees at CVDC 2020 Anubhav Shrivastava, head of Data Sciences at Voot Viacom18 shared some of the ways that Voot is using computer vision. Computer Vision Has Gained Traction Over The Years Shrivastava shared that the area of computer vision has been evolving over the years and has drastically seen an increase in applications with time. He defined CV as the field of study of developing techniques to help machines see, interpret and create visuals. Some of the common application areas of computer vision include surveillance, biometrics, heat maps, checkouts, facial recognition, medical imaging, remote sensing, 3D model building, intelligent image processing and more. Some of the methods in computer vision that are most popularly used are object detection, image classification, semantic segmentation, instance segmentation, deep tracking, GANs and more. Computer Vision At Voot Voot is an ad-supported and subscription video streaming platform available in India. It has more than 100 million peak monthly users. For operating at such a large scale, Voot largely depends on emerging technologies such as artificial intelligence and machine learning, of which computer vision is a crucial technology. Shrivastava shared that computer vision plays a crucial role in video streaming as it helps in increasing user experience, increasing user retention, increase in efficiency by bringing about automation, reduce cost, increase monetising opportunities, increase user acquisition and more. ‘It has been used both at the supply-side to create content and at the consumption-side to increase user experience,” he said. Some of the areas where Voot is using computer vision are: Ads cue point detector: The content on video is placed with ads in between which plays a key role in making or breaking viewership. The ads have to be strategically placed as its location at the wrong place at the wrong time may lead to the audience to leave. Placing ads in continuation with the content is crucial. Using CV has helped Voot to decide where to place ads, ensure that video watching experience is not hampered, replace manual detection, omit human error, handle scale and enable a pathway for experimentation. Contextual ads: Another important aspect is to place ads that are relevant to the user depending on the in-video objects. Using technologies such as object detection, facial recognition, colour recognition and more, Voot is playing on human psychology to mimic human behaviour. They have been able to drastically increase user engagement, propensity to buy and ensure higher monies. Product placements: Product placement is a popular concept in Hollywood and is now used in videos all across the globe. Most of the time, the content is created first and then they look for buyers or the content is made for TV which is then moved online. In situations like these, it may be difficult to ensure ad placements. Therefore, product placements come into picture where product logos are strategically placed into the videos. Computer vision comes handy to detect whitespaces, deep tracking logo insertion, ensure post publishing advertising, and more. Highlight creation: Highlights or short video content is in demand as most users wish to see the important highlights from episodes in a go. While creating highlights in sports is easy as it allows for voice detection during important moments in the form of audience cheering or high pitched commentary, it is challenging to do this in non-sports videos. Shrivastava says that it is not only about decibels but requires to look into the defining moments, listen to the video intently, observe facial expressions and more. Voot creates a lot of highlights and is experimenting with computer vision to produce and scale it. It is still not in production yet but the experiment is on. Augmented reality: Voot is also looking to include augmented reality into its shows to increase user engagement. Especially in the times of COVID, it is ensuring that users are a part of the show while at home. This too is at an experimental phase and Shrivastava believes that it will be a reality in 2-5 years. Shrivastava concluded by sharing that the key areas where computer vision has been deployed are to increase user engagement, user experience and increase monetisation. It has been using techniques such as group normalisation and video to video synthesis to create thematic content catering to audiences in a specific manner.","excerpt":"One of the interesting areas where computer vision is being explored is the video streaming platform. Video streaming has become one of the most important fields in the last decade, especially with the lockdown that has resulted in an increased viewership on these platforms than ever before.  To boost the user experience a lot of […]","categories":["AI Features"],"tags":["Computer Vision","human intelligence at machine scale","Interviews and Discussions","medical image processing companies"],"author_name":"Srishti Deoras","publish_date":"2020-08-18T11:00:09","publication_year":"2020","word_count":778,"keywords":["data science","Go","artificial intelligence","human intelligence at machine scale","machine learning","AI","computer vision","Interviews and Discussions","automation","object detection","Computer Vision","GAN","R","medical image processing companies"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","data science","object detection","R","Go","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-is-voot-using-computer-vision-to-scale-up-their-offerings\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10055099,"title":"Microsoft Introduces New Resources &#038; Tools To Help Implement AI Responsibly","content":"In collaboration with Boston Consulting Group (BCG), Microsoft has introduced guidelines for product leaders that are designed to help prompt important conversations about how to put responsible AI principles to work. This guidance is distinct from Microsoft’s internal processes and reflects perspectives from both organizations. Microsoft has also built tools to help ML practitioners identify issues, diagnose causes and mitigate problems before deploying apps. “Moving from principles to practices is difficult, given the complexities, nuances and dynamics of AI systems and applications. There are no quick fixes and no silver bullet that address all risks with applications of AI technologies. But we can make headway by harnessing the best of research and engineering to create tools aimed at the responsible development and fielding of AI technologies,” wrote Eric Horvitz, Chief Scientific Officer at Microsoft, in a blog post. The ten guidelines are grouped into three phases: Assess and prepare: Evaluate the product’s benefits, the technology, the potential risks, and the team.Design, build, and document: Review the impacts, unique considerations, and the documentation practice.Validate and support: Select the testing procedures and the support to ensure products work as intended. Along with these, the company has released a Responsible AI dashboard that surfaces Error Analysis, Fairlearn, InterpretML, DiCE and EconML functionalities into one pane of glass to assist AI developers with fairness, interpretability and reliability of AI models. Within the dashboard, the tools can communicate with each other and show insights in one interactive canvas for an end-to-end debugging and decision-making experience. The open-source tools that Microsoft has built include: Error Analysis: Analyses and diagnoses model errorsFairlearn: Assesses and mitigates fairness issues in AI systemsInterpretML: Provides inspectable machine-learned models to enhance debugging of data and inferencesDiCE: Enables counterfactual analysis for debugging individual predictionsEconML: Helps decision-makers deliberate about the effects of actions in the world using causal inferenceHAX Toolkit: Guides teams through creating fluid and responsible human-AI collaborative experiences","excerpt":"Microsoft has launched new tools and guidelines to enable product leaders build AI responsibly from research to practice","categories":["AI News"],"tags":["AI Models","Microsoft","Responsible AI"],"author_name":"Meeta Ramnani","publish_date":"2021-12-08T16:17:32","publication_year":"2021","word_count":316,"keywords":["AI Models","programming_languages:R","AI","ML","Responsible AI","responsible AI","Aim","ViT","GAN","R","Microsoft"],"extracted_tech_keywords":["AI","ML","Aim","R","GAN","ViT","responsible AI","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-introduces-new-resources-tools-to-help-implement-ai-responsibly\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":13625,"title":"Is your in-house analytics center SLEAK?","content":"The data sciences industry is growing at a rapid pace, making it seem like it has been around forever. It was only in 2012, that HBR claimed ‘Data scientist’ to be one of the sexist job of this century. In today’s world, it is very rare for large enterprises to outsource functions like sales or strategy and analytics is proving to be the next core element in how businesses function today. This is prompting a new wave – with a number of enterprises looking to setup their in-house analytics centres and take full advantage of the wealth of information at their disposal. With the industry not showing any signs of slowing down, it is important to understand what makes analytics centres and teams successful. The Analytics Powertrain There exists a popular myth, that an Analytics centre is deemed a success or failure solely based on the talent it is able to attract and retain. While talent is important, it is not the sole driver of success. The most critical element of the centre’s success is its integration with the organization and the value it drives. So how should this integration be engineered? Analytics teams and their processes are no longer ad-hoc. Enterprises often make the mistake of treating analytics teams as support and outsource them and as a result isolating them from the business and the decision-making process. Using analytics for pure reactive needs reduces the potential and the chances of it being successful. Rather, it should be a cycle that balances between proactive and reactive modes. Like a vehicle’s powertrain mechanism transmits the power from the original source of energy to the surface of the road analytics centres should drive all phases of the business lifecycle to propel the organization forward. For example, business Performance could be decoded using attribution models, which in turn determine the Strategy to be adopted. Similarly, scenario generators that provide an insight on how to gain a competitive advantage could be leveraged into fine-tuning an enterprise’s strategy. Once the strategy is finalized, a series of analytical experiments could aid the go to market Decisions and finally, operationalization on the ground occurs through Activities that are streamlined using systems that recommend the best action. It is therefore important to build analytics centres and teams as ‘Powertrains’ to ensure their success and relevance in the organization by integrating them in every step of the business process. SLEAK- the new paradigm For an Analytics centre to transform into a ‘Powertrain’, it is necessary to be SLEAK. Synthesize, to ensure the bridge between the business and analytics teams is strong. An ecosystem that allows for interactions that result in problem identification, translation and solution consumption with ease. Learn, to stay ahead of the industry curve. Technologies, methods, problems are becoming obsolete so fast that it has become a basic need continually learn to stay relevant in the current world. Experiment, to innovate and be a differentiator. Problems continue to evolve and hence long cycles for solutioning would not work. Quick experiments must be performed to stay ahead of the curve. Align, to be relevant in the organization. It becomes very important to streamline the efforts to ensure alignment with the organizational priorities. Any effort with no clear value adds is just a wastage of resources. Kultivate, – Knowledge Cultivation to create efficiency through collaboration. Creating a smart, collaborative environment prevents knowledge loss in the system and reduces the dependency on people through reusable assets. These cultural traits and characteristics are key to successfully setting up an in-house analytics center or enhancing an existing center’s offerings. SLEAK centers have a higher degree of success to realize their tremendous potential and act as innovation hubs for enterprises.","excerpt":"The data sciences industry is growing at a rapid pace, making it seem like it has been around forever. It was only in 2012, that HBR claimed ‘Data scientist’ to be one of the sexist job of this century. In today’s world, it is very rare for large enterprises to outsource functions like sales or […]","categories":["IT Services"],"tags":[],"author_name":"Ashwin Kumar","publish_date":"2017-03-21T04:33:09","publication_year":"2017","word_count":617,"keywords":["data science","Go","API","AI","ML","RAG","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/house-analytics-center-sleak\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10115262,"title":"10 AI Startups Run by Incredible Women Entrepreneurs","content":"Women entrepreneurs are increasingly making their mark in the tech world, overcoming unique challenges and breaking barriers. Despite facing issues such as access to funding, gender bias, and balancing professional and personal life, many women have risen to prominence, particularly in the field of artificial intelligence. As of 2023, India has witnessed a significant surge in women-led startups, with 45% of the country’s startups now led by women. Their work not only addresses real-world problems but also sets a precedent for future generations of women in entrepreneurship. Here’s a look at 10 remarkable AI startups founded by women entrepreneurs, each contributing to their field with cutting-edge solutions. Spark Studio This ed-tech company was founded by Anushree Goenka, Namita Goel, and Jyothika Sahajanandan, along with Kaustubh Khade, in 2021. The women, with a combined experience of three decades in operations, education and engineering, respectively, came together to start Spark Studio, which is backed by YCombinator. The company provides online courses for children between the ages of 6 and 15. Their focus is on extracurricular skills, offering classes in art, music, and communication areas like public speaking. Spark Studio also aims to create an engaging and interactive environment for learning with the help of Sparky, a friendly fox powered by AI, to help kids build English fluency and confidence. GoCodeo Meghana Jagadeesh, who previously worked at ByteDance, ShareChat and Google, began GoCodeo along with Jatin Garg. With a background in industrial engineering and management, she admits to having learnt a lot from a niche education, where engineering meets management. GoCodeo is an AI co-pilot designed for software quality control, targeting individual developers to assist in writing clean code. The startup automates the software testing process, particularly focusing on the laborious task of writing tests by using generative AI to autonomously generate test cases. GoCodeo has attracted over 12,000 developers from 40 countries and is used by over 150 companies. The company was selected for the Google startup accelerator program last year. Delv AI This company was founded by a 16-year-old entrepreneur. Pranjali Awasthi, having experienced the difficulties of research, was driven to create a solution that uses artificial intelligence to make information more easily accessible. She began her entrepreneurial journey after an internship at the research labs of Florida International University, coinciding with the release of ChatGPT-3 beta. This experience inspired her to conceive Delv.AI, with the vision to harness AI for extracting and summarising data from the vast online information space. The platform simplifies complex data gathering from various sources, making the search process faster and more efficient. The company is already valued at $12 million. Immunito AI ImmunitoAI is a bio-technology company that leverages AI to revolutionise antibody discovery for drug development. Founded by Trisha Chatterji and Aridni Shah, who possess expertise in the fields of artificial intelligence and biological sciences, respectively. Their combined experience brings a unique perspective and skillset to the forefront of drug discovery. This AI-powered platform designs and screens new antibodies with specific properties, making the drug development process faster and more efficient.  ImmunitoAI enjoys the backing of industry giants like Google and Entrepreneur First. Atlan Prukalpa Sankar’s second company Atlan, which is a central platform for managing data assets, similar to how GitHub supports engineering or Figma supports design. Previously, she was the founder of SocialCops, another data management company. A TEDx speaker, Sankar is passionate about data being used to make big decisions. Atlan is a platform where teams can consolidate their data, including tables, dashboards, models, and code, creating a unified source for their data assets. Recognized for its innovation in the field, Atlan was named by Gartner as one of the top three global companies in DataOps, highlighting it as a ‘Cool Vendor’. Hasura Rajoshi Ghosh and Tanmai Gopal founded Hasura and led the company towards $136.5 million funding over four rounds. Ghosh, a serial entrepreneur, has founded six startups before Hasura. This company simplifies building APIs by automatically generating something called a GraphQL API directly from your existing databases. Traditionally, building APIs can be time-consuming and requires a lot of code. Hasura streamlines this process. Metashop MetaShop makes 3D asset creation accessible for everyone. It specialises in developing tools that allow users to easily create high-quality 3D models of real-world objects using only photos or videos. It was founded by Sophiya Jagannathan and Mukul Ingle. During the COVID pandemic, the duo was looking to buy furniture, but struggled with not being able to imagine how it would appear in real life. Finding a solution for this led to Metashop. Jagannathan has a background in computer science and went on to work as a deep learning engineer at Superset Labs. Here, she contributed to the development of object detection software and co-developed a device called ‘Trace’ during COVID, aimed at early detection of infections in public spaces and offices. BrainSight AI Laina Emmanuel and Rimjhim Agrawal came together to start BrainSight AI that uses AI to help doctors study brain health more accurately and speed up patient recovery. Agarwal, who has a PhD in psychiatry and machine learning application in neuroscience, joined hands with Emmanuel, whose expertise lies in engineering and management. The women, though from different backgrounds, are passionate about solving real world problems. With their combined expertise in neuroscience they aim to make a lasting impact in the field of neuro-technology. Metamagics Anita Kulkarni Puranik comes with extensive experience of 32 years in software development and management. She started Metamagics with a vision to improve patient care after experiencing organ transplantation in real life. Metamagics’ main product, GridSense, automates analytics and optimises care protocols for transplant patients using data analytics, semantic web technologies, and deep expertise in transplant care. This platform offers disease-specific insights, aims to minimise errors, and improves long-term outcomes by effectively organising patient data for clinical decisions. TekUncorked Meenakshi Vashist, the founder and CEO of TekUncorked, brings over two decades of experience in product innovation across various sectors before venturing into entrepreneurship. A former ISRO scientist and previously the founder of TU Technology, she aims to address the challenges in the power sector by making grids smarter and more reliable, leveraging IoT. Founded in 2019, TekUncorked operates in three Indian states, focusing on climate tech impact and the integration of renewable energy sources to create a sustainable future in electricity distribution. This platform predicts and mitigates risks, improving grid efficiencies and reducing environmental impact. The company has quickly gained recognition, winning awards including the UN-Habitat’s Katowice Energy Innovation Challenge 2022.","excerpt":"Women entrepreneurs are harnessing AI to launch startups that innovate across sectors.","categories":["Deep Tech"],"tags":["Startups","Women in Tech"],"author_name":"K L Krithika","publish_date":"2024-03-08T19:00:16","publication_year":"2024","word_count":1083,"keywords":["ChatGPT","artificial intelligence","machine learning","AI","ML","RAG","Aim","deep learning","Women in Tech","analytics","generative AI","Startups"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","analytics","generative AI","ChatGPT","Aim","RAG"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/10-ai-startups-run-by-incredible-women-entrepreneurs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10070249,"title":"Log9 unveils its latest range of EVs in India","content":"Founded in 2015, Log9 Materials started off as a materials science company with a focus on nanomaterials and Graphene. Incubated at IIT Roorkee’s TIDES, Log9’s vision was to lead the Graphene revolution by introducing real-life products\/solutions to the world. Later, the startup positioned itself as a deep-tech and advanced battery and energy storage technology company. Today, Log9 primarily caters to the e-mobility space. “We are at the forefront of redefining the EV industry’s standards in the fight against climate change, by offering batteries that can be charged 9x faster, can last 9x longer and offer 9x higher performance and safety,” said Dr Akshay Singhal, CEO and founder of Log9 Materials. With an aim to make India self-reliant in meeting its own electric mobility and EV transformation goals, Log9 has launched India’s first and South Asia’s largest indigenously developed battery cell manufacturing line and made-in-India TiB cells. “We have grown our headcount by over 10x while nurturing a great work culture and homegrown deep-tech talent. As a brand, Log9’s sole purpose and mission is to pioneer responsible energy, and we are dedicatedly working towards realising this by creating Log9’s responsible energy that comes without compromises or caveats”  Dr Akshay Singhal RapidX battery series: The Log9 RapidX series is the safest EV batteries is designed to efficiently power electric vehicles across all Indian and tropical conditions. Built using patented high-power cell technology, the RapidX batteries powered by Log9’s InstaCharge battery tech are extremely durable, with the ability to last longer than conventional batteries without losing their efficiency. Furthermore, the RapidX batteries provide 9x faster charging, 9x better performance, and 9x battery life. Log9’s InstaCharge technology brings down the total charging time which further optimises charging infrastructure and operational profitability; furthermore, these batteries are built to operate across -30° to 60° C and come with an operational life of 15,000+ cycles. Under the RapidX series, Log9 has till date launched two variants – RapidX 2000 and RapidX 6000 batteries, which can fully charge an e-2W in 15 minutes and e-3W in 40 minutes. With its flagship TiB cell technology and RapidX batteries, Log9 aims to provide India’s fastest charging EVs built with “made for India” batteries that have over 10-year battery life. ZappUp: ZappUp is a cutting-edge energy storage solution by Log9 that effectively regulates and stabilises supply-demand mismatch by bridging the current gaps in the electric grid. ZappUp can be applied to considerable power and energy services ranging from fast-response ancillary grid services to energy arbitrage and so on. It caters to stationary power capturing renewables and making the energy available to the grid at peak demands. Furthermore, ZappUp’s features allow it to be used and operated in versatile applications like microgrids, UPS, data-centres, telecoms, grid-level energy storage, and industrial applications like in forklifts, robotics, and AGVs. 2022 line-up “The overall EV adoption in India today stands at 0.8 percent; but the country has set goals to achieve 30 per cent EV adoption in the B2C sector and over 70 per cent adoption in the B2B segment by 2030. To achieve such ambitious goals India needs to accelerate EV adoption through technologies made for India, and also, impetus on B2B electrification should be the go-to route to take. And keeping the same in mind, at the Green Vehicle Expo this year, we at Log9 are proud to be showcasing our complete range of vehicles for last-mile logistics and transportation along with the technology that makes our batteries 9x faster charging, 9x longer life, and most of all, the safest in the market. More importantly, this event provides us with a prolific opportunity and space to highlight our brand identity, vision, technological capabilities and much more amongst relevant stakeholders and the general public at large,” said Akshay. On July 01, 2022, Log9 unveiled its latest range of electric vehicles for last-mile logistics across different EV platforms – including two-wheelers, three-wheelers and four-wheelers – all of which are instacharged by Log9’s RapidX batteries. In the e-2W segment, Log9 introduced the Quanta e-bike made by Gravton Motors, which holds the distinction of travelling from Kanyakumari to Khardung La successfully. It also boasts of being an all-terrain bike, which is a first in its segment. In the e-3W segment, Log9 introduced the Rage+ RapidEV which has the ability to get charged from 0 to 100 percent in 35 minutes. They also unveiled a e-3W cargo vehicle Grevol– which is India’s first electric 3-wheeler with a load carrying capacity more than electric 4Ws. Both the Rage+ RapidEV and Grevol would be powered by Log9’s newly launched 7.7 kWH battery pack. Log9, in collaboration with eBikeGo, launched the first of its kind electric trike Velocipedo with an aim to redefine urban mobility. In the e-4W segment, Log9 tied up with Northway Motors to unveil a retrofitted LCV, claimed to be the fastest-charging 4-wheeler commercial retrofitted vehicle. The startup also demonstrated the complete portfolio of its batteries including RapidX 2000, RapidX 6000, RapidX 8000 and RapidX 15000, and a range of stationary energy storage solutions – ZappUp 4000 and ZappUp 12000. Log9 also showcased its on-road InstaCharge assistance vehicles, the Power Bank. Performance and safety In recent times, safety and quality have taken a back seat in the EV market. A massive fire at an EV parking space in Delhi guts nearly 100 vehicles#EVFire #EV https:\/\/t.co\/kXBL4OsIKY— DriveSpark (@drivespark) June 8, 2022 “There are multiple reasons that are contributing to EV 2-wheeler incidents\/mishaps that we have noticed recently in the country, that include low-grade battery cells, improper thermal management, wrong cell connection design, comparability and design mismatch between battery and vehicle sub-systems. The Indian Government is currently discussing and revising the country’s battery testing protocols to arrive at standardising batteries and their performance. Additionally, policy frameworking is in progress to ascertain the performance of cells in India with respect to their operating and weather conditions. Companies are also internally setting up QC testing protocols to stress-test batteries and vehicle design to make them safer by adapting a ‘safety-first’ approach. Also read: Can cobots save EVs? “As of today, Indian companies import Li-Ion cells and they make battery packs out of it in India. The major challenge that India is facing is the price sensitivity of its battery packs. Hence the choice of cells being imported in India is mainly NMC cells which are cheaper but come with a risk of becoming thermally unstable under hot weather conditions. To overcome these issues, Indian battery companies and EV companies have to select the right cell chemistry that is more suited to Indian weather and driving conditions. Moreover, battery and EV companies have to work on developing cooling technologies that can ensure that batteries are able to perform their function without overheating. Whereas the really impactful, long-term solution is for India to make cells in India, with chemistries and materials that are more suited for Indian conditions. Log9 is already moving ahead in this path by creating India’s first Li-ion cells named ‘TiB’ cells that can withstand extreme temperatures of Indian and tropical climatic conditions,” he added. The overall EV adoption in India today stands at 0.8%. However, India has set goals to achieve 30% EV adoption in the B2C sector and over 70% adoption in the B2B by 2030. To achieve such ambitious goals, India needs to accelerate EV adoption through technologies made for India; B2B should be the route to take and that’s exactly what Log9 does. “For EV adoption in India to happen fast, the maximum number of vehicle deployments initially are bound to come from the B2B sector. In India, 50 lakh deliveries happen each year. And businesses are today competing to get the product to the customer faster and at the lowest delivery cost. With rising fuel prices, the last-mile delivery and logistics sector has got a huge incentive to move to EV vehicles since EV-led deliveries happen at a fraction of a cost than ICE engine deliveries and can thereby lead to significant operational cost savings in the long run,” said Akshay.","excerpt":"Log9, in collaboration with eBikeGo, launched the first of its kind electric trike Velocipedo with an aim to redefine urban mobility.","categories":["Deep Tech"],"tags":[],"author_name":"Sri Krishna","publish_date":"2022-07-01T14:00:00","publication_year":"2022","word_count":1335,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","RAG","Aim","ViT","R","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","ViT","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/log9-unveils-its-latest-range-of-evs-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":51346,"title":"Kolkata Police To Use AI In Crime Detection","content":"The land of art, literature and culture will now see the government authorities keeping a close eye on the lawbreakers. In a move that will empower the law enforcement authorities, the Kolkata Police is now expanding the footprint of the CCTV cameras AI-powered devices in crime detection. According to a noted news wire, Kolkata Police has already installed 3,000 closed-circuit cameras all across the city. Police Commissioner Anuj Sharma said, “We are expanding it. Recently, you have seen instances of crime detection by analysing the CCTV footage… With the installation of such cameras, catching those indulging in anti-social acts will become simpler.” Sharma added that the police were planning to install more hi-tech cameras with artificial intelligence and face recognition facility. Earlier this year, an AI-powered CCTV camera aided the cops in identifying and punishing a citizen for spitting and defacing public property in Gujarat. These smart cameras installed by the Ahmedabad Municipal Corporation (AMC) were able to identify a citizen’s body movements, interpret them, and understand that the person who was breaking the law. Another interesting example was when last year, the Special Investigation Team probing the murder of journalist-activist Gauri Lankesh used AI-based algorithms to gather CCTV footage and identify persons which matched the physical description given by eyewitnesses to the gruesome murder which took place in September 2017. The Indian Police force has already started taking an increasing interest in crime analytics using big data, which involves storing and analysing huge volume and variety of data in real-time, to predict and inference patterns and trends especially relating to human interactions and behaviour. To know which areas are most prone to crimes, the police force also uses predictive analytics to develop models using machine learning to know which areas are most prone to crime. This helps them to keep a track on which criminals or individuals to keep a track on. For example, the Delhi police has partnered with ISRO to develop an analytical system called Crime Mapping, Analytics and Predictive System (CMAPS), which helps the Delhi police to ensure internal security, controlling crime, and maintaining law and order through analysis of data and patterns. Late last year, the Delhi traffic police had announced the proposed installation of an intelligent traffic management system (ITMS). The news traffic management system will work on radar-based monitoring with the help of AI. The police analyse the traffic pattern, volume, number of vehicles, and other factors, and collect them on a cloud. The data then would be used to manage the traffic — one of the key tools being the automated traffic signals. Jharkhand police force is trying to implement an analytical system with the help of IIM Ranchi, to help evaluate criminal records, date and time of crime occurrences, and location to predict crime-prone zones.","excerpt":"The land of art, literature and culture will now see the government authorities keeping a close eye on the lawbreakers. In a move that will empower the law enforcement authorities, the Kolkata Police is now expanding the footprint of the CCTV cameras AI-powered devices in crime detection. According to a noted news wire, Kolkata Police […]","categories":["AI News"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-12-09T12:22:59","publication_year":"2019","word_count":464,"keywords":["big data","Go","machine learning","artificial intelligence","programming_languages:R","AI","R","GRU","analytics","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","predictive analytics","R","Go","big data","GRU","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/kolkata-police-to-use-ai-in-crime-detection\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10008768,"title":"How This Startup Is Accelerating Human-Machine Collaboration","content":"With an aim to help business users within an enterprise to improve organisational capabilities, Milpitas, California-headquartered Jiffy.ai has been facilitating enterprise automation with a no-code development platform. Jiffy.ai has been helping the companies revolutionise areas such as finance and operations. The startup believes that innovating and automating in a business setting has historically been a struggle — requiring teams of people, a variety of roles, and a significant budget to bring new ideas to life, among others. It needs full support from multiple stakeholders. But with the advancements in AI, powered by machine learning and natural language processing, we are able to finally put the power of innovation in a business user’s hands. Uniquely structured with a 20-person co-founding team, 70% of whom come from the C-suite level or have previous founder experience, Jiffy.ai helps Fortune 1000 customers execute their business transformation initiatives by using intelligent automation to improve complex processes at scale and create efficiencies by 10x. Analytics India Magazine caught up with Babu Sivadasan, CEO, Jiffy.ai, who was a former president of Thiruvananthapuram-based IT company Envestnet. “Jiffy.ai was conceptualised with the singular vision of putting the power of innovation and automation in a business user’s hands,” said Sivadasan. With over 150 employees, Jiffy has offices across Palo Alto, CA, Boston MA and Trivandrum and Bengaluru, India. Enabling Automation Sivadasan shares that they enable organisations to deploy automation HyperApps that make various organisational functions and end-to-end complex business processes automated, resilient, and autonomous. The Jiffy.ai product suite offers an integrated automation experience using AI-powered machine learning, cognitive processing, Natural Language Processing and more to remove mundane, error-prone tasks from the workforce. This allows employees to focus on higher-value strategic and innovative work. “This is how we are focused on accelerating human-machine collaboration,” he said. Highlighting some of the use cases where they have helped various businesses, Sivadasan shared that using Jiffy.ai’s Invoice Processing HyperApp, a global manufacturer was able to automate the entire invoice processing function across multiple countries. “We achieved 85% straight-through processing over 12 weeks across 150,000 invoices a month for 5,000 suppliers,” he said. Sharing another use case, Shivadasan said that they worked with a global professional service firm with new hire onboarding, which can be very challenging if carried out manually, especially when organisations hire in large numbers. They wanted to automate their onboarding process to ensure a smooth first experience for their new hires. “Jiffy.ai automated 80% of the client’s onboarding activities, making the process faster, frictionless and a pleasant experience overall, which was a welcome move for their HR team, that could focus on more strategic tasks,” he added. In another instance, where a leading American airline faced an overload of requests for ticket rescheduling and refunds due to COVID-19, their customer service teams were unable to cope with the tens of thousands of requests coming in, and accuracy in their work was suffering. Jiffy.ai, with its powerful hyper-automation capabilities, resolved over two thousand person-hours of the backlog in less than six weeks, with accuracy. This helped the airline deliver a superior customer experience, despite limitations caused by the pandemic. AI\/ML Is Central To The Company With a team that has deep expertise in AI, machine learning, natural language processing, fintech and automation, their core product is designed on these technologies to enable complex process automation and application design which enterprises can avail in just a few clicks. “Our technology sits on top of legacy systems dependent on manual entries which tend to be slow and heavily error-prone. We combine the power of RPA, deep document processing and native cognitive (AI and ML) capabilities into a single integrated platform with built-in enterprise-class security,” shared Sivadasan. He further added that they are the only context-aware system in the market, capable of recognising and fixing mistakes and discrepancies through self-learning machine models that can handle structured and unstructured documents without the need to create new machine models from scratch. “It’s this combination of intelligent document processing and machine learning, which gives organisations greater control and governance, and access to centralised scaling so deployments can take place in days and not months,” he said. Talking about the tech stack, Sivadasan shares that it has everything an enterprise might need to hyper-automate complex business processes to increase accuracy and reduce both errors and costs. For instance, Jiffy.ai Automate is the only app-based intelligent automation platform that lets a user implement, manage and monitor enterprise-wide automation, to bring about business transformation. “The single dashboard combines the power of RPA, ML, AI, document processing, workflow and analytics that supports the end‑to‑end lifecycle management of automation with a human-in-the-loop approach to achieve significant ROI,” he said. Growth Story As the demand for enterprise automation is picking up, there is an increasing number of companies who want to streamline operations; therefore, the demand for solutions by Jiffy.ai is quite up. “We currently have more than 30 enterprise customers, and our client base has grown more than 200% in the first half of 2020 alone,” shared Sivadasan. Talking about funding, the startup received $18 Million funding led by Nexus Venture Partners with participation from Rebright Partners, W250 Venture Fund and other C-level business leaders. “Our new funding will be used to scale our product, technology and sales as we pursue a responsible path to AI and innovation by delivering on our vision, grow the team further and scale our product roadmap and existing customer base which already comprises Fortune and Global 1000 companies,” he said. Overcoming Challenges Due To Pandemic While Covid-19 has impacted many businesses they work with, there’s an increased demand from a global workforce now being forced to work from home, and the added pressure on teams to lean into more collaborative workflow technologies. Sharing an example, he said that customer service teams have struggled with customer inquiry volumes, and through their technology, they can automate help desk operations, ensuring 30% improvement in agent efficiency. With so many disruptions to the supply as a result of the pandemic, businesses are looking toward RPA to achieve the full value of intelligent automation, and embrace this rapid pace of change. “And we expect that solutions like JIFFY.ai will become an increasingly important part of the strategy and free up their employees to focus on creative, innovative, and higher-value activities,” he said on a signing note.","excerpt":"With an aim to help business users within an enterprise to improve organisational capabilities, Milpitas, California-headquartered Jiffy.ai has been facilitating enterprise automation with a no-code development platform. Jiffy.ai has been helping the companies revolutionise areas such as finance and operations.  The startup believes that innovating and automating in a business setting has historically been a […]","categories":["AI Startups"],"tags":["Automation","innovative technology advancements","RPA"],"author_name":"Srishti Deoras","publish_date":"2020-10-01T17:00:10","publication_year":"2020","word_count":1054,"keywords":["Go","API","machine learning","innovative technology advancements","AI","RPA","ML","Automation","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","R","Go","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-startup-is-accelerating-human-machine-collaboration\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10103388,"title":"NeevCloud Launches the Country’s First Made in India AI SuperCloud","content":"NeevCloud, has today announced the launch of the country’s first AI SuperCloud. The company aims to make India self-reliant on AI and Supercomputing and help users in resolving India’s core challenges by deploying AI cloud infrastructure including 40,000 GPUs and storage worth USD 1.5 billion by 2026. The startup aims to solve fundamental customer problems in India related to accessibility, environmental impact, support, self-reliance, and fostering innovation in cloud and AI technologies within the Indian market. NeevCloud aims to solve these challenges and democratize access to AI and supercomputing for all enterprises, and startups by introducing ‘AI Supercomputing as a Service’ model and providing the world’s lowest prices for Cloud GPUs at USD 1.69\/hour reducing costs by up to 50%. It can enable companies to leverage Generative AI and deploy Large Language Models (LLM) across various use cases capable of solving India’s unique challenges. “At NeevCloud, we envision a future where India not only leads but empowers the world with innovative technology. Our mission is to democratize AI cloud computing, making it accessible to all enterprises and startups across India, and fostering the growth of Made in India AI. “Our AI SuperCloud stands as a testament to Indian innovation, and we believe our journey has the potential to power not only our nation, but users across the world who need cutting-edge but affordable cloud solution,” said Narendra Sen, Founder & CEO, NeevCloud. NeevCloud began its journey by solving a fundamental issue plaguing the data center industry when it came to AI and HPC workloads. Existing data center infrastructures are inefficient at handling demanding workloads of AI, ML, and HPC projects, which can consume up to a staggering 80KwH of power, starkly contrasting the typical 10-20KwH of power used for regular workloads. This power-intensive environment leads to excessive heat generation, causing operational complications, and rising operating costs. To tackle this problem, NeevCloud created an entirely indigenous patented solution: Varuna, a groundbreaking liquid immersion cooling system. Varuna’s technology efficiently addresses high rack density heat concerns by submerging servers in a coolant, resulting in a remarkable reduction of Power Usage Effectiveness (PUE) from 1.5 to 1.05, achieving exceptional sustainability. This not only resolved the heat issue but also had a cascading effect, reducing cooling costs by an astounding 90% and overall expenses by a significant 50%. NeevCloud can also enable confidential machine learning by creating SuperClusters that directly allocate to the BFSI, automotive, and healthcare organizations. It has the potential to assist businesses of all scales, research institutions, and diverse industries in advancing their AI strategies.","excerpt":"The startup wants to deploy an AI cloud infrastructure that includes 40,000 GPUs and storage worth USD 1.5 bn by 2026.","categories":["AI News"],"tags":["Startups"],"author_name":"Pritam Bordoloi","publish_date":"2023-11-21T11:00:50","publication_year":"2023","word_count":423,"keywords":["machine learning","AI","cloud computing","innovation","ML","RAG","Aim","generative AI","GAN","Startups","R"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","Aim","RAG","cloud computing","R","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/neevcloud-launches-the-countrys-first-made-in-india-ai-supercloud\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24722,"title":"How To Convert Data Science And Machine Learning Internships Into Jobs","content":"Can popular massive open online courses turn into job offers? This a common dilemma faced in data science forums about internships and certificate courses converting in job offers. Now that you snagged an internship, built a portfolio of work, networked with the mid and senior management team and are ready to pursue a career in data science, you are waiting for the high-paying job offer to land in your inbox. According to UC Berkeley data scientist Karsten Walker, recruiters look for specific traits such as applying scientific methodology to a business problem and look for candidates who have a demonstrated history of applying analytical concepts. Breaking into the data science industry is tough as it is, and data science newbies face intense competition in the search for jobs after internships. However, getting an internship means getting one foot inside the door and applying classroom learning to real world business problems. Another upside of the internship is that most organisations who look for interns, do it from an opportunistic point of view wherein interns can be converted into full-time hires. It is an informal way of recruitment and also helps in minimising the hiring cost to the company. How Can One Maximise Their Internship Experience And Convert It Into A Job Offer: Talk about the opportunity given to work on all types of projects, including the ability to work with unstructured data and writing scripts to scrape websites Talk about the responsibility given as an intern and how the work delivered created value Stress about the experience of working alongside data scientists to productionize models and testing different algorithms Stress about the importance of your technical credentials with respect to the impact on business and general IT skills Emphasise on how hiring a senior professional could take six months and can be an expensive process, in the meantime, critical requirements can be filled with an entry level candidate Note to Recruiters: Here’s Why You Should Hire Entry Level Professionals Vin Vashishta, data science and machine learning strategist recently observed via LinkedIn that most organisations who look for PhD-level candidates usually end up using basic techniques like regression and decision trees for structured data. For these jobs, coding skills required are minimum and the result is often a report or visualisation. Vashishta poses a relevant question —  why are we still asking for a Master’s or PhD with over three years of experience for an entry level position? Data science interns usually demonstrate an eager learning ability most sought after by recruiters and startup founders who are keen on seeing the passion behind the project. Vashishta strongly urges organisations and businesses to hire more entry level talent instead of leaning towards mere bigger degree holders for their data science needs. He believes organisations should be more open to hire an aspiring data scientist with a BS\/Certification\/Online Program and one or two years’ experience in development or analytics. Vashishta’s thoughts on entry level professionals are, “You’re going to be amazed by the value they return in terms of hiring costs and ability they bring to the table to grow with the business needs.” Given how data science is becoming more business-centric and is being leveraged across domains, it will be helpful for entry-level professionals to build skills in both tech side such as engineering, systems engineering and web programming and also beef up learnings about the business domain such as presentation skills, working in big teams, cross-functional stakeholder management. This can only be gained on the job, so try being as participative in the internship experience as you can. All these soft skills fall in the data science job cluster and add value to business. Tips For Data Science Interview For tips on cracking the dreaded data science job interview, Vashishta points out in a blog post that it comes to two things — potential and goal. He explains that the employers ask if the candidate must be self-motivated and does the he\/she have a reasonable plan to achieve their goals. He writes that applicants should spend some time understanding their minutes professional motivations and what drives you as a person or sets you apart from others. One must make sure to emphasise how the role fits into their career progression and how the company can help in achieving the career goals. Besides skill, recruiters look for skills such as machine learning and how it will shape your work can leave a strong impression on the recruiting team. Tier I leadership management looks for vision and strategy and how it can be tied to business outcomes, build new products and meet demands, so sharing your vision with senior management can help clinch the job.","excerpt":"Can popular massive open online courses turn into job offers? This a common dilemma faced in data science forums about internships and certificate courses converting in job offers. Now that you snagged an internship, built a portfolio of work, networked with the mid and senior management team and are ready to pursue a career in […]","categories":["AI Hirings"],"tags":["data science internships","data science interview","Data Science Jobs","high paying jobs in india","Machine Learning Interview","MOOCs","science jobs","Virtual Internship Program"],"author_name":"Richa Bhatia","publish_date":"2018-05-18T11:38:31","publication_year":"2018","word_count":781,"keywords":["data science","Go","data science interview","machine learning","programming_languages:R","science jobs","high paying jobs in india","AI","Machine Learning Interview","MOOCs","Data Science Jobs","RAG","Virtual Internship Program","analytics","GAN","data science internships","R","startup"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","RAG","R","Go","GAN","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/how-to-convert-data-science-and-machine-learning-internships-into-jobs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10139643,"title":"Google AI Overviews Rolling out in 100 Countries","content":"Starting this week, Google’s AI Overviews will expand to more than 100 countries, marking a milestone with over a billion global active users each month. Google’s CEO, Sundar Pichai, while sharing the news on X, said, “Since launching AI Overviews earlier this year, people have been asking a whole new universe of questions. With today’s expansion to 100+ countries, AI Overviews will reach 1B+ users globally. More exciting things for Search ahead!” The expansion also brings a significant update to language accessibility. Users in any region with AI Overviews will now be able to view information in multiple supported languages like English, Hindi, Indonesian, Japanese, Portuguese, or Spanish, making it easier for users to get insights in their preferred language, regardless of their location. Since May, Google has been enhancing the visibility of links within AI Overviews to make it easier for users to explore supporting content. New link displays, like the right-hand desktop links and tap icons on mobile, have already driven an increase in website traffic. Google has also added in-line links directly within overview text, which, as seen in early testing, helps boost visits to related sites. Google AI Overview Suggests Adding Glue to Pizza In May, Google introduced AI Overview in search results in the US, marking one of the most significant updates to its search engine in 25 years. These results, displayed at the top of the search page, provide users with a condensed overview before delving into the typical list of blue links. The internet was abuzz with discussions on the impact of this new technology, from recommending unconventional ingredients for pizza recipes to suggesting unusual remedies for medical conditions like kidney stones and whatnot. “I understand the sentiment. You know, it’s a big change. These are disruptive moments,” said Google chief Sundar Pichai in a previous interview. He acknowledged the concerns raised by media outlets and publishers about the rollout of AI previews in Google Search. In the interview, he said, “If you put content and links within AI Overviews, they get higher click-through rates than if you put it outside of AI Overviews.” Pichai has been optimistic about the outlook and has highlighted that integrating AI Overviews into search results is not a simple win-lose scenario. Instead, he believes it will ultimately contribute to a richer and more dynamic search experience.","excerpt":"AI Overviews will now be able to produce information in multiple supported languages like English, Hindi, Indonesian, Japanese, Portuguese, or Spanish.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Google"],"author_name":"Vidyashree Srinivas","publish_date":"2024-10-29T12:39:45","publication_year":"2024","word_count":389,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Google","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-ai-overviews-rolling-out-in-100-countries\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10105076,"title":"Why ShareChat Co-founders Started a Robotics Company","content":"Last month, former ShareChat co-founders Bhanu Pratap Singh and Farid Ahsan announced the launch of General Autonomy, an industrial robotics startup based in Bengaluru. It is interesting to note that they exited the social media company less than a year ago after having built it for nine years. Their new startup General Autonomy – a significant pivot from social media – aims to revolutionise the future of factories, and build robots, which will be used by the industrial sector. These machines will be driven by artificial intelligence platforms. “We LOVE robots! 😍 Especially industrial ones—they help us “make” things! With a vision to make mass manufacturing agile and distributed like software development, we’re in Bengaluru but working globally. 🌎,” shared Ahsan on X. Interestingly, the IIT Kanpur alumni Singh and Ahsan majored in Electrical and Material Science Engineering respectively were active members of the robotics club in college. Looks like they have always had the passion for robotics since the beginning, and want to be a part of the growing demand for industrial robots in India. This year, saw a lot more founders leaving the company to start another compared . A critical aspect of this trend is the financial health of these companies. Only a few were profitable in 2022, with the majority still operating at a loss. ShareChat Struggles The duo Ahsan and Singh started building apps together in 2012 since their college days. After experimenting with several products, they succeeded with ShareChat, along with co-founder Ankush Sachdeva designed to cater to non-English speakers in India. Currently their roles as COO and CTO are left vacant and with no plans of hiring. The current CFO Manohar Charan has taken up some of the responsibilities but according to sources the teams are working without formal leadership. Launched in 2015, ShareChat was notable for its focus on regional Indian languages, which is a barrier for social media apps in India. They made digital content accessible to a broader user base by adding 15 languages like Odia, Kannada and Punjabi. The platform was built to work effectively even in areas with low internet connectivity, thanks to a lightweight messaging architecture. “This was a game-changer. It helped us grow fast in areas of low internet connectivity,” said Bhanu. Under their leadership, ShareChat achieved rapid growth, reaching over 60 million users within its first year. This year however, things haven’t been going well for the company. The company, despite being valued at around $4.9 billion, reported operational revenues of only ₹347 crore and losses amounting to ₹2,988 crore in the fiscal year ending March 2022. Additionally, ShareChat faced significant challenges, including a string of expensive acquisitions and a mass firing where 20% of its workforce was cut due to external macro factors impacting the cost and availability of capital. As of 2022, Moj, a subsidiary of ShareChat specialising in short video content, along with MX TakaTak. Amidst this, the loss of two founders raised eyebrows. However, it’s not all bad news as it saw an increase in revenue because of their multiple acquisitions over the last three years. The number of users also increased considerably to a combined 300 million in all their platforms. The transition to building a robotics company also comes at a time when ShareChat is going through a turbulent time in terms of revenue. The company reported a loss of INR 3,241 crores against a revenue of  INR 535 crores. The industrial robotics market is valued at $17 billion as of this year and is expected to grow significantly to a $35.4 billion market by 2028. The segment is fueled by the increasing adoption of Industry 4.0 technologies across various sectors, particularly in the automotive, packaging, and metal industries. Robotics to the Rescue In a time where the integration of AI in robotics is predicted to change how the field works, General Autonomy is a significant shift in the founders’ entrepreneurial journey, moving from social media to the field of industrial robotics.“Our mission is to revolutionise the future of factories, focusing on automation..” Ahsan said in his tweet. He stressed that AI driven machines to automate the toughest parts of labour workflows in factories. https:\/\/twitter.com\/frdahsan\/status\/1724003874695188637?s=20 To set this up, the company has successfully raised $3 million in seed funding, with contributions from India Quotient, ElevCap, and mentorship from favourite investors including Mayank Khanduja and Anand Lunia. The funding round also saw participation from notable angels like Srinath Ramakkrushnan, Ramakant Sharma, and Ankush Sachdeva, Share Chat’s co-founder. This shift aligns with the sector’s anticipation of a ‘GPT moment’ in robotics. Companies like Grey Orange, have become notable success stories, raising substantial investment and expanding globally. The sector has attracted investor interest, with startups receiving funding for a range of innovative solutions, from agricultural robots to automated solar panel cleaning and companion robots. The plans of General Autonomy are unclear as they refused to comment.","excerpt":"General Autonomy, launched by ShareChat’s ex-co-founders Bhanu Pratap Singh and Farid Ahsan, marks a shift from social media to AI-driven industrial robotics.","categories":["Deep Tech"],"tags":["industrial robots","Robotics"],"author_name":"K L Krithika","publish_date":"2023-12-18T16:04:30","publication_year":"2023","word_count":814,"keywords":["Go","API","artificial intelligence","AI","Git","Robotics","industrial robots","GPT","automation","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","Git","API","GPT","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/why-sharechat-co-founders-started-a-robotics-company\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26238,"title":"MIT’s New Plug-And-Play System Can Control A Team of Robots With Brainwaves","content":"Image Source: MIT In an age where robots are being controlled by the mind, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory system have designed a plug-and-play supervisory control using muscle and brain signals for real-time gesture and error detection. In a new paper, MIT researchers have presented an end-to-end system that enables classification of gestures which will help correcting the robot on demand. The research combined three streams — human biosignals, brain activity and muscle feedback — by harvesting Electromyography (EMG) and Electroencephalography (EEG) signals to enable human intervention in supervisory tasks. Earlier last year, MIT’s CSAIL department joined hands with Boston University to design a system that corrects errors in real-time. This research has gone further down the path to making the communication between robots and humans more seamless and intuitive by allowing a means to control robots and minimise errors in critical situations. According to CSAIL director Daniela Rus, this particular work combining EEG and EMG feedback enables natural human-robot interactions for a broader set of applications that we’ve been able to do before using only EEG feedback. She shared, “…And by including muscle feedback, we can use gestures to command the robot spatially, with much more nuance and specificity.” To show how the system works, researchers utilised error-related potential (ErrP) or the brain signals that are related to error which surface when people notice errors. When the plug-and-play system detects an ErrP, the robot stops so that a user can correct the action with hand gestures, which are conveyed to the robot through the user’s muscle activity. Some Of The Highlights Of The Paper Are: A framework for combining error detection through EEG with gesture detection via EMG allows supervisory control of autonomous robots Since a new user doesn’t need to provide additional training data for EEG or EMG classifiers, training is done on a corpus of data which increases the applicability and makes the interface plug-and-play Need For Brain-Controlled Robots And How Humans Can Correct It According to Kurzweil AI, the plug-and-play supervisory control systems is not without its shortcomings. While EEG signals cannot be detected reliably, EMG signals can be tricky to map the motion. But when are the two combined, it enables better bio-sensing, making it possible for the system to work on new users without training. As the world gears for more intuitive and natural systems, a safe robot supervision architecture can pave the way for human safety as well. However, in cases where a human supervisor is crucial and a human element has to close the loop, this plug-and-play system can play an important role. If robots and humans have to exist as co-workers, have to exist and robots require supervision, then robots and human should work on tasks in coordination in real time. Areas Where Plug-And-Play System Can Play A Central Role: With the rise of industrial robotics, labour-starved national economies are dependent on an automated workforce and are using industrial robots more extensively in the manufacturing and service-oriented sector. In fact, industrial robots were introduced in the manufacturing industry a decade ago with Japan being the leading maker and supplier of industrial robots. Automotive industry and emerging markets in China, North America and South-East Asia are also investing heavily in the robotics sector. While the widespread use of robotics has sparked talks about labour displacement, a research paper claims that around four million workers are directly interacting with robots in their day-to-day tasks. This underscores the need for a supervisory system with a human element closing the loop. While robots do perform pre-programmed tasks in a highly-structured environment such as the automotive and manufacturing industries, the involvement of humans through the plug-and-play system can minimise the scope of error. Major robot manufacturers such as FANUC have provided a system for safe cooperation with robots with physical and information support in place, to support human work. Close attention to correcting errors and ensuring a safe human-robot interaction with plug-and-play systems And lastly, MIT’s work is truly groundbreaking because the plug-and-play system can be easily deployed in large manufacturing settings and other sectors with humans managing a team of robots.","excerpt":"In an age where robots are being controlled by the mind, researchers from MIT’s Computer Science and Artificial Intelligence Laboratory system have designed a plug-and-play supervisory control using muscle and brain signals for real-time gesture and error detection. In a new paper, MIT researchers have presented an end-to-end system that enables classification of gestures which […]","categories":["AI Features"],"tags":["industrial robots"],"author_name":"Richa Bhatia","publish_date":"2018-07-06T04:57:37","publication_year":"2018","word_count":690,"keywords":["Go","artificial intelligence","programming_languages:R","AI","data_tools:Spark","ML","programming_languages:Go","industrial robots","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","R","Go","ViT","programming_languages:R","programming_languages:Go","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mits-new-plug-and-play-system-can-control-a-team-of-robots-with-brainwaves\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10119993,"title":"‘Pronoun Illness’ is Ola’s Problem, Not India’s ‘Rich Culture’","content":"Ola Krutrim’s chief executive and influential Indian AI leader, Bhavish Aggarwal, is unhappy. LinkedIn has taken down his post, in which he called the networking platform’s usage of non-binary gender pronouns like they\/them “pronoun illness” and hoped it would not reach India. According to the billionaire IIT-Bombay alumni, “pronoun illness” is being taught by “big city schools” and is increasingly appearing in CVs—which he clearly is not a fan of. He believes India “needs to know where to draw the line in following the West blindly!” He did not stop there. “Most of us in India have no clue about the politics of this pronoun illness. People do it because it’s expected in our corporate culture, especially MNCs. Better to send this illness back where it came from. Our culture has always respected all. No need for new pronouns,” he added. Aggarwal’s commentary revealed a deep-seated resistance to what he perceives as blind westernisation. Soon enough, he received widespread criticism from people debunking his old-school ideologies and LinkedIn eventually removed his post. The lack of inclusivity and unconscious bias in the tech ecosystem aren’t new. Yet, it’s troubling and kind of embarrassing that at a time when tech giants worldwide are ramping up their DE&I initiatives, Aggarwal choses to disparage the use of pronouns – a bare minimum. To break it down, pronouns such as “he\/him”, “she\/her”, and “they\/them” are used in language to refer to someone in place of their name. Traditionally linked to male and female genders, these pronouns are evolving with greater awareness of gender diversity. Many opt for pronouns that accurately represent their identity, using “they\/them” as a gender-neutral option. Using chosen pronouns respects and validates personal identities, enhancing inclusivity. “Recognising the importance of using correct pronouns isn’t just about respecting individual identities; it’s a crucial step towards fostering inclusivity and acceptance within society, ” Ruku Arora, director – enterprise business solutions (EBS) at Walmart Global Tech, told AIM. Arora identifies as queer. Pride Month is Rainbow Washing for Ola In 2024, labelling this aspect an ‘illness’ feels like a regressive step back by a century. The “illness” is actually inclusivity. This brings us to the question: Is rainbow washing during the Pride Month only a marketing gimmick for companies like Ola in India? “It’s disgusting how Ola had put up a Pride Month post last year in June, and now the CEO himself has spewed so much hate against the community. We don’t want your performative pride posts. Really fearing for the safety of the LGBTQI+ employees working at Ola,” shared Ayden, a non-binary senior copywriter at DViO Digital. However, this is not the first time when Ola’s work culture has come under scrutiny. Several employees have left the company, citing the bad work-life balance policies. “Being a very aggressive company, it is not everyone’s cup of tea and finding the right people is invaluable,” Aggarwal said during an interview almost 12 years ago. It looks like there has been no major shift since then. Two years ago, a Bloomberg report stated that an employee of Ola Electric, its EV arm, was asked to run laps around the facility as punishment for a minor oversight! The CEO, who takes pride in his “anger and frustration”, is known for fostering a toxic work environment through destructive criticism, abusive language, unrealistic deadlines, late-night meetings, and an impatient and hostile management style. In a previous interview, Aggarwal, who described himself as a “straight shooter”, said, “I have a purpose and passion. Sometimes, in the journey of a business, people don’t align with it, and those separations are not as one wants. That sometimes leads to bitterness, and sometimes, the bitterness gets amplified.” However, Ola does not shy away from celebrating Pride with much pomp and show every June. The Importance of Pronouns Using the right pronoun is the minimum respect that one can ensure to the queer community. Contrary to its creator’s beliefs, when we turned to Krutrim AI for clarity, the Indic LLM chatbot was quick to explain the situation. “The use of gender-appropriate pronouns is important as it encourages people to be more conscious and respectful of others’ preferences. Changing one’s attitude to using gender pronouns is a two-step process – the first is to accept and acknowledge the stated gender preference, and the second is to practise using the preferred pronoun in interactions,” Ketty Avashia, executive director on the enterprise functions technology team at Wells Fargo India & the Philippines, told AIM. Avashia, who identifies as a trans man, faced a lack of support for non-conforming identities in a culture that did not accept diversity, leading to social isolation in his early days. “Pawns” of the “West” “Rich of you to call my post unsafe! This is exactly why we need to build our own tech and AI in India. Else we’ll just be pawns in others’ political objectives,” said Aggarwal in a followup post about LinkedIn removing his content. Aggarwal is leading the development of Krutrim, touted as “India’s first full-stack AI” solution. However, he wants to make it more relevant to India’s culture and heritage. Contrary to his opinions, non-binary and trans folks have been recognised in “our” literature since ages, from mediaeval Bhakti literature to the Ramayana and Mahabharata. “Bhavish Aggarwal, you are totally justified in being angry if your preferred pronouns are he\/him and AI is using they. For a lot of people, though, who are on the gender spectrum (yes gender is a spectrum) or are nonbinary or simply do not want to reveal their gender, people prefer they\/them. “Besides, it’s a standard practice that if the gender of a person is not known, we tend to use ‘they’ nowadays. Also, please note this is not a political ideology. Trans, non-binary and people who do not identify with traditional genders have existed (on record) for thousands of years and are documented well in our literature,” noted Tanuj Kumar Kukreti, product manager at Eko. Many saints in the Bhakti movement in India, which emphasised love and devotion to God over ritualism, expressed themselves in ways that blurred traditional gender lines. Poets like Basava and Akka Mahadevi questioned and defied the gender norms of their times. So, to break the bubble—it is not new. “You did call pronouns as illness and it’s discriminatory to trans people. India has always been accepting of all cultures and Indian languages do have gender neutral pronouns. If you aren’t aware of it, do better and learn. Calling it a political stance and western agenda to hide transphobia will do nothing,” commented Farhana M, engineering recruiter at Atlassian. An LLM’s performance relies heavily on the quality of the dataset used during its training. If the dataset is biassed and contains systematic distortions, lacks diversity, or misrepresents certain groups or facts, the LLM will inevitably inherit and potentially amplify these biases in its outputs. “Diversity, or the lack thereof, in technology, isn’t merely a social issue; it profoundly impacts the functionality and accessibility of products. The tech world’s homogeneity often leads to products and services that fail to consider the diverse needs of their user base,” Brenda Darden Wilkerson, global chief executive officer, AnitaB.org, told AIM. Krutrim is not an exception. The question is if personal biases like those of Aggarwal gets into the system, it will not make India the AI hub of innovation he wants to build. Read more: The Struggles and Triumphs of Trans Inclusion in Indian Tech","excerpt":"“Better to send this illness back to where it came from. Our culture has always had respect for all. No need for new pronouns.”","categories":["AI Trends"],"tags":["Ola"],"author_name":"Shritama Saha","publish_date":"2024-05-09T17:39:47","publication_year":"2024","word_count":1243,"keywords":["Go","Ola","TPU","AI","innovation","Git","RAG","Aim","ViT","Rust","R"],"extracted_tech_keywords":["AI","Aim","RAG","TPU","R","Go","Rust","Git","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/pronoun-illness-is-olas-problem-not-indias-rich-culture\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10076175,"title":"Atlassian introduces ‘Atlassian Together’ and New Product Capabilities For Teams","content":"At an inaugural Work Life event in San Francisco on Friday, leading productivity software firm Atlassian Corporation Plc unveiled a new work management offering called ‘Atlassian Together’, along with several new product capabilities. Recognising the changing nature of how the firm works, Atlassian released the new products and features to help align autonomous teams and use them to power businesses’ digital transformation efforts. The new products are built on the Atlassian Platform to drive cross-team collaboration, which is open by design. This would give teams the freedom to easily integrate their favourite tools to stay aligned with universally accessible project context, processes, and status. Erika Trautman, Head of Product, Atlassian’s Work Management offerings, said, “At Atlassian, we’re obsessed with designing products that unleash the potential of every team. Our leading work management products enable teams to choose the best tool for their needs and work differently. But delivering amazing products on their own is not enough. They must connect with each other, and all the other apps teams are using to get work done, to bring the teams together. The new offering and capabilities we unveiled today will drive this powerful cross-team collaboration.” One such product is the Atlassian Together, a single subscription to Atlassian’s work management products such as Confluence, Jira Work Management, and Atlas among others. The company claims that more than 150,000 customers currently use these products. Smart Links is one of their new capabilities built on the platform that helps users access information across Atlassian and third-party apps. It allows team members from finding and inserting content to creating and editing work items across products. Moreover, Atlas will be coming out of beta next month, as announced at Atlassian’s flagship conference, TEAM’22. Since its inception, customers in the early access programme have witnessed benefits from new improvements such as richer integrations and flexible reporting. Delivering to more than 240,000 customers across small and large companies—including Bank of America, NASA,  Redfin, Verizon, and Dropbox—Atlassian helps teams organise, discuss, and complete shared work to deliver quality results on time.","excerpt":"Recognizing the changing nature of how the firm works, Atlassian released the new products and features to help align autonomous teams and use them to power businesses’ digital transformation efforts.","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-09-30T16:35:50","publication_year":"2022","word_count":341,"keywords":["programming_languages:R","AI","digital transformation","Git","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Git","GAN","ViT","digital transformation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/atlassian-introduces-atlassian-together-and-new-product-capabilities-for-teams\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10097891,"title":"Meta to Develop Chatbots with Personas to Enhance User Retention","content":"Meta, the parent company of Facebook, is gearing up to introduce a lineup of AI-driven chatbots with distinct personalities for its services including Instagram and Facebook as early as next month, reported Financial Times. The move aims to enhance user engagement on its social media platforms. Meta has been designing prototypes for chatbots that can have humanlike discussions with its nearly 4bn users, the report added, citing people with the knowledge of the plans. According to sources, some of the chatbots, referred as “personas” by the staff, embody distinct characters. Meta has considered introducing personas like Abraham Lincoln, while another one might offer travel advice with a surfer-style approach. The chatbots are set to debut in September, offering a novel search function, personalized recommendations, and providing an enjoyable interactive experience for users. This development comes after their new Twitter rival app Threads lost more than half of its users in the weeks following its buzzy launch on July 5.  According to a report by SimilarWeb, the number of daily active users on Threads fell from 49 million on July 7 to 23.6 million on July 14. In addition to enhancing user engagement, chatbots could gather substantial new data on users’ interests. This data could enable Meta to deliver more personalized and relevant content and advertisements to its users. Meta recently reported its most profitable quarter since 2021. In terms of advertising, ad impressions delivered across their Family of Apps increased by 34% year-over-year in Q2 2023. However, the average price per ad decreased by 16% year-over-year.","excerpt":"Meta has been designing prototypes for chatbots that can have humanlike discussions","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-08-01T15:42:55","publication_year":"2023","word_count":256,"keywords":["programming_languages:R","AI","chatbots","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","chatbots","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-to-develop-chatbots-with-personas-to-enhance-user-retention\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10054773,"title":"AWS Announces Scholarship Program For Their AI And ML courses","content":"AWS recently announced the launch of the AWS AI & ML Scholarship program in collaboration with Intel and Udacity, designed to prepare underrepresented and underserved students globally for careers in AI and ML. The AWS AI & ML Scholarship program is launching as part of the all-new AWS DeepRacer Student service and Student League. This is a new student division of the popular AWS DeepRacer program, a cloud-based 3D car racing simulator that provides a fun way to learn about ML and reinforcement learning (RL). Through DeepRacer Student, learners will have access to free online training to learn the ML and RL basics. Students will also be able to access 10 hours of model training and 5 GB of storage per month to participate in the DeepRacer Student League, a global autonomous racing competition exclusively for AWS AI & ML students. In collaboration with Intel and Udacity, the program is geared toward underserved and underrepresented high school and college students globally who are at least 16 years old. These are students who may have faced financial barriers growing up or are part of underrepresented groups, including women, people with disabilities, LGBTQ+ persons, as well as people of colour. Students in these groups who take part in the DeepRacer Student League will be eligible for a chance to win one or two of 2,500 annual scholarships from Udacity, an online learning platform focused on technology skills. To be considered for the scholarship, students must successfully finish all AWS DeepRacer Student learning modules and achieve a score of at least 80% on all course quizzes, reach a certain lap time performance with their DeepRacer car in the Student League, and submit an essay. Each year, 2,000 students will win a scholarship to the AI Programming with Python Udacity Nanodegree program. Udacity Nanodegrees are massive open online courses (MOOCs) designed to bridge the gap between learning and career goals. This aims to equip students with programming and ML fundamentals to solve real-world problems with ML. The top 500 participants in this first Nanodegree will be eligible to join a second customized Nanodegree program curated specifically for AWS AI & ML Scholarship program recipients. All scholarship recipients will be given exclusive access to Ask-Me-Anythings (AMAs), fireside chats, and office hours with AI\/ML professionals and diversity experts from Amazon, Intel, and AWS collaborators such as Girls in Tech. These will help familiarize students with different job functions in the AI\/ML field. To register for the program, click here.","excerpt":"All scholarship recipients will be given exclusive access to Ask-Me-Anythings (AMAs), fireside chats, and office hours with AI\/ML professionals and diversity experts from Amazon, Intel, and AWS collaborators such as Girls in Tech.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AWS","AWS cloud","AWS services","Data Science","Data Scientist","Deep Learning","Deep Reinforcement Learning","Machine Learning","Python","Reinforcement Learning","Udacity"],"author_name":"Victor Dey","publish_date":"2021-12-03T12:26:49","publication_year":"2021","word_count":413,"keywords":["AWS services","Udacity","R","RAG","programming_languages:Python","Data Science","Go","Reinforcement Learning","AWS","AI","ML","Machine Learning","Deep Reinforcement Learning","AWS cloud","cloud_platforms:AWS","Python","Aim","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","Aim","RAG","AWS","Python","R","Go","cloud_platforms:AWS","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-announces-scholarship-program-for-their-ai-and-ml-courses\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10082136,"title":"Microsoft Acquires a Stake in London Stock Exchange to Offer AI &#038; Analytics Solutions","content":"Microsoft signs a ten-year partnership with the London Stock Exchange Group (LSEG) and buys nearly 4% stake in the UK bourse operator. Microsoft will provide new data infrastructure, analytics, modelling solutions and cloud computing products to the former in exchange. The agreement also states that Microsoft’s executive vice president for cloud and artificial intelligence, Scott Guthrie, will be appointed as a director. “This strategic partnership is a significant milestone on LSEG’s journey towards becoming the leading global financial markets infrastructure and data business and will transform the experience for our customers,” said David Schwimmer, chief executive officer of LSEG. The deal indicates that the investment sector’s growing desire for data that offers them a competitive edge in today’s lightning-fast electronic market. In 2021, global spending on financial market data and news increased 7.4% to a record $35.6 billion, according to Burton-Taylor International Consulting research from April. LSEG’s shares rose to 4% in Europe on Monday. According to the release, Microsoft will purchase its share from a group that includes Blackstone, Thomson Reuters, partners of the Canada Pension Plan Investment Board, and Singapore’s GIC. While its data and analytics will be used with different Microsoft applications, LSEG’s data platform and other crucial technology and infrastructure will move to Microsoft’s large public cloud product, Azure. LSEG promises to invest a minimum of $2.8 billion on cloud-related products with Microsoft in its ten-year contractual agreement. Additionally, Microsoft and LSEG will collaborate on the creation of innovative solutions for professional cooperation. LSG earlier made Workspace—a platform for data and analytics. The two organisations will collaborate to develop this product and include it in Microsoft Teams, the company’s messaging platform. “Advances in the cloud and AI will fundamentally transform how financial institutions research, interact, and transact across asset classes and adapt to changing market conditions,” said CEO and chairman Satya Nadella of Microsoft. Earlier in 2018, Microsoft announced a major cloud deal with multinational technology company Grab, which chose Azure as its cloud platform as well. Microsoft also invested $300 million in online bookseller Barnes & Noble in 2012. But, it did not end well as the latter bought out a stake for about $125 million in 2014 after being unable to gain customers.","excerpt":"The US tech giant enters a ten-year strategic partnership with LSEG for cloud computing services and data analytics","categories":["AI News"],"tags":["Azure","Mergers and Acquisitions"],"author_name":"Shritama Saha","publish_date":"2022-12-12T18:12:53","publication_year":"2022","word_count":369,"keywords":["artificial intelligence","cloud_platforms:Azure","programming_languages:R","AI","cloud computing","R","analytics","GAN","Mergers and Acquisitions","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","cloud computing","Azure","R","GAN","cloud_platforms:Azure","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-acquires-a-stake-in-london-stock-exchange-to-offer-ai-analytics-solutions\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":726,"title":"Connecting College Students With IoT Ideas","content":"The Internet of things has been pinned as the next Industrial Revolution by the industry experts and we do not have any doubts about it. From a hairpin to an aeroplane, everything is on its way to becoming the next smartest thing up for grab! The Internet of things is redefining our present and future. How do we define IOT? Reference: MobileMarketingWatch.com Not everyone has agreed to a common consensus on the definition of IOT. Some call it the internet of things, while others the Internet of everything. IOT is that umbrella under which all internet enabled devices capable of sending and receiving messages interact with each other intelligently. The IOT contains 6 essential elements : Connectivity – With the internet, of course. Devices– All internet enabled smart devices to interact with the ecosystem ( your smart phones, iPads, sensors in your refrigerator and that cool IOT chip in your garage door) Data– This is the most crucial part of the system as the usability and the functionality of an IOT model depends on the type of information that is exchanged. Your data has a purpose. You need to discover and translate it into action. Not all data points are easy to find and equal in stature. You need to isolate, aggregate and analyze them based on the specific use cases. And then you can do wonders. Reference: thetechreader.com Intelligence – Because you need to comprehend the information exchanged and act on it. You must have surely heard the DIKW model (Data to information to knowledge to wisdom). What comes after the implementation of the model is the analysis and action part which collectively completes the run. Automation– The automation is a key aspect in IOT. Whether it may be the auto update of software or business process automation. After all, you need things to run on their own, don’t you? So, these 6 essential elements of IOT together create an IOT Ecosystem which can be defined according to the industry it needs to thrive in. For other definitions of IOT by the industry leaders, you can refer to the Internet of things definitions. Industry forecasts: In the coming years, IOT will be bringing us life changing experience by the virtue of it’s every big and small ecosystem that is going to be the part of our daily lives. Here are few of the key forecasts for the IOT market: Largest device market in the world: A projected 34 billion devices will be connected to the internet by 2020 The money spent on development and deployment of IOT ecosystems might be as huge as $5-6 trillion in the next 5 years It will add to over $2 trillion to the global economy by 2020 The global IOT market will reach $14.4 trillion by 2022 Manufacturing, retail trade , information services , and finance and insurance  are the four industries that comprise more than half the total value of the projected 2022 market. An insane game of numbers is what you will see in the coming years. It’s going to be a gala time with Machine to Machine (M2M)\/IOT models and hyper-enhanced consumer experiences. No wonder, the whole world is hugely investing man, machine and money in it. Remember :  If solving hard data problems is your thing or if squeezing the last bit out of your code to improve 1% query performance gives you a kick , then we invite you to fight the battle at the CeBIT IoT Hackathon .Our partners for the hackathon are Intel and Bosch making for an insane event filled with innovative contests, prizes and hacking. With top prizes being a paid trip to Hannover, Germany and b incubated. You’ll also get to meet lot’s of like minded people. Be there on 9th and 10th December for an inexplicably awesome hackathon. All you need to do is register at vsity.in\/cebithack The CeBIT IoT Hackathon explores ideas and innovation around Internet of Things and challenges you to be innovative while connecting the everyday objects you find around towards making life easier for all. They are building tomorrow, today. Are you a part of this revolution?","excerpt":"How Venturesity is partnering with big names in the IoT space","categories":["Deep Tech"],"tags":[],"author_name":"AIM Media House","publish_date":"2016-11-06T05:11:39","publication_year":"2016","word_count":689,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","automation","ViT","R"],"extracted_tech_keywords":["AI","RAG","R","Go","ViT","automation","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/connecting-college-students-iot-ideas\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10064686,"title":"How are Indian banks adopting digitisation and AI?","content":"Remember the time when demonetisation in India shook the country, back on November 8, 2016? People across the nation went berserk, standing before ATMs and banks, restless to deposit their notes. Banks have gone digital now. The majority of transactions are online. Can you imagine what it would be like to witness another round of demonetisation in this digital day and age? The days of standing in queues, following the long token process and fulfilling tedious formalities are long gone. The way banks function has been revolutionised with the onset of digitisation and the adoption of technology in banking. With technology, especially fintech banking solutions, customers are empowered with self-service capabilities providing them with operational processes which were otherwise given only when a customer visits a bank branch. The Union government had announced in 2021 that about 72 per cent of financial transactions from public sector banks are done digitally. In 2019-2020, there were 3.4 crore customers active on digital channels; thanks to the COVID-19 pandemic, the customer base almost tripled to 7.6 crores in 2020-2021. Leading the digitisation and adoption of AI in the Indian banking systems are private financial institutions like ICICI Bank, HDFC Bank, Axis Bank, Kotak Mahindra Bank, etc. On the other hand, big nationalised banks like SBI, Canara Bank, Central Bank, etc., too, are travelling with time and slowly taking steps to digitise their banking solutions. In an article by finder.com, it is said that India is set to experience the biggest boom in digital banking adoption in the next five years, with a 21 per cent increase in adults with online-only bank accounts. This means that by 2025, there is an estimate that just under 400 million Indian adults will hold neobank accounts. A recent survey report highlights that an estimated 205 million Indian adults already have a digital-only bank account. This number is predicted to increase to 397 million within the next five years. IDC’s latest report reveals predictions for Indian corporation banking, stating that Indian banks are projected to spend over US$1 billion by 2025 on public cloud initiatives to indicate the importance of the cloud in driving technology transformation. At the policy level, in the Union Budget 2022-23, there has been an increase in capital expenditure to fund various infrastructure projects. Accelerating the process of adoption of the cloud, leveraging AI extensively and efficiency in data management are some of the key technologies that are focused on. Some of the key predictions of IDC include AI in payments: About 40 per cent of payments will be optimised using AI-derived routing models by the year 2025. CBDC impact on cash management: As the rolling out of CBDC is gaining momentum, by 2025, more than 15 per cent of tier I corporate banks will be offering integrated solutions to unlock liquidity from traditional and digital assets. Connectivity platforms: 35 per cent of corporate banks will platformise connectivity by 2023 to deal with the growing channel fragmentation. How is AI used in banking? Artificial intelligence helps banks manage high-speed data to receive useful insights, and features like digital payments, AI bots, and biometric fraud detection systems lead to high-quality services. The COVID-19 pandemic has accelerated the deployment of AI, where organisations are automating day-to-day operations to understand COVID-19 affected datasets and leverage them to improve the stakeholder experience. Here’s a look at the application of AI in the banking sector with some exclusive insights from bankers. Akhil Handa, Chief Digital Officer, Bank of Baroda “The Indian banking industry is at the forefront of digitisation and AI specifically; many use cases are shaping the future of banking. At the Bank of Baroda, we use AI for cash forecasting at currency chests, predictive maintenance of ATMs using external sensors and internal data points of failures. Natural language processing to understand the reasons for the reopening of email complaints by analysing the resolution responses. For training employees, the Bank of Baroda has developed two adaptive learning modules that consider individual officers’ learning rates and learning requirements. This is akin to learning “Segment of One” – in terms of personalisation.” Prashant Joshi, Managing Director and Head Consumer Banking Group, DBS Bank India “At DBS, our brand promise ‘Live more, Bank less’, reflects our belief that in the digital era, we must deliver banking that is so simple, seamless and invisible that customers have more time to spend on other important aspects of life. Our mobile banking application, digibank, by DBS, delivers a paperless, straight-through and intuitive experience across all banking transactions. It provides a host of digital-first features, including 24×7 live chat for daily banking needs and end-to-end wealth management solutions, domestic and international remittance options and clutter-free digital loan approval and repayment process. DBS customers can also complete their digital KYC via a video-based Customer Identification process.” About digibank digibank provides AI-backed Intelligent Banking services to customers, enabling them to manage their finances better, leading to higher engagement, retention, and transactions. Since the AI-powered insights feature within digibank went live in 2020, the bank has seen 47% repeat usage of the feature month-on-month. In fact, digibank is already contributing to about 20% customer acquisition across all our branches, thus complementing the bank’s physical network. According to IBS intelligence, there are five applications of AI in banking: 1. Customer service\/engagement (Chatbot) Incorporating chatbots provides very high ROI in cost savings, making them a popular application across many industries. Customers can easily solve their queries on chatbots, like balance inquiries, accessing mini statements, fund transfers, etc. This helps reduce the burden on contact centres, internet banking, etc. 2. Robo advice A Robo-advisor makes an effort to understand a customer’s financial health by analysing the shared financial history and data. The Robo-advisor gives investment recommendations for a particular product or equity based on the analysis and goals set by the client. 3. General Purpose\/Predictive Analytics AI’s most popular usage is in general-purpose semantic and natural language applications and broadly applied predictive analytics. AI is leveraged to detect specific patterns and correlations in the data, which was otherwise impossible using legacy technology. These patterns help to identify untapped sales and cross-sell opportunities, or even metrics around operational data, which leads to direct revenue impact. 4. Cybersecurity AI improves the effectiveness of cybersecurity systems where it leverages data from previous threats and then learns the patterns and indicators that may be unrelated to predicting and preventing attacks. AI also helps in monitoring internal threats or breaches and suggests corrective actions, resulting in the prevention of data theft or abuse. 5. Credit Scoring\/Direct Lending AI plays an important role in helping alternate lenders determine the creditworthiness of clients by analysing data from traditional and non-traditional data sources. This helps lenders to come up with innovative methods of lending systems that are backed by a robust credit scoring model, even for those individuals or entities with limited credit history. Indian banks have to move with time and help customers to adopt newer technologies with ease. Digitisation and leveraging AI has helped both banks and customers in making the entire banking experience hassle-free.","excerpt":"An estimated 205 million Indian adults already have a digital-only bank account, and this number is predicted to increase to 397 million within the next five years.","categories":["IT Services"],"tags":["ai bots","Cybersecurity","Predictive AI","predictive analytics"],"author_name":"Poornima Nataraj","publish_date":"2022-04-11T16:00:00","publication_year":"2022","word_count":1178,"keywords":["Go","artificial intelligence","AI","chatbots","R","ML","RAG","analytics","ai bots","Cybersecurity","predictive analytics","Predictive AI","fraud detection"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","RAG","chatbots","predictive analytics","fraud detection","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-are-indian-banks-adopting-digitisation-and-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10119347,"title":"Mysterious gpt2-Chatbot Takes Everyone by Surprise","content":"A mysterious AI model named ‘gpt2-chatbot’ recently appeared on the LMSYS Chatbot Arena. According to several users on X, the model showcased better reasoning and math abilities than OpenAI’s GPT-4. This left many users surprised, questioning whether it’s a new model by OpenAI. i do have a soft spot for gpt2— Sam Altman (@sama) April 30, 2024 Interestingly, the model was released without official documentation, and there are no details to be found. However, soon after, OpenAI chief Sam Altman posted a cryptic message: ‘I do have a soft spot for GPT-2.’ This led to speculation that this could hint at a new version beyond GPT-4, possibly GPT-5. AI influencer Rowan Cheung highlighted several notable features of the gpt2-chatbot, with its enhanced reasoning skills being praised by several users on X who posted screenshots. A mysterious new AI model called “gpt2-chatbot” is going viral.It was released without official documentation, and there is speculation that it could be OpenAI's next model.Here's everything we know so far (and how to try it for free):— Rowan Cheung (@rowancheung) April 30, 2024 A user on X  tested the chatbot’s mathematical capabilities, presenting it with an International Math Olympiad problem. Impressively, the chatbot solved it on the first attempt, although it couldn’t tackle all problems on the test. Despite this, its performance remained outstanding. Moreover, the chatbot’s coding skills surpassed those of GPT-4 and Claude Opus, according to Chase, founding engineer at Codegen. He said the gpt2-chatbot excelled in complex code manipulation tasks, outperforming newer models. Furthermore, the chatbot’s proficiency in ASCII art was lauded by Cheung, who described it as “miles ahead of any other model” in this domain. Interestingly, this development comes as the tech ecosystem eagerly awaits GPT-5. Recently, Altman said that the company will release GPT-5 in the ‘coming months,’ adding that OpenAI has more important things to release before GPT-5. “Before we talk about a GPT -5-like model… I know we have a lot of other important things to release first,” said Altman. Meta also stirred the air with Llama 3, released about two weeks ago.","excerpt":"“I do have a sweet spot for gpt 2,” says Sam Altman","categories":["AI News"],"tags":["GPT-5","OpenAI","Sam Altman"],"author_name":"Gopika Raj","publish_date":"2024-04-30T18:40:20","publication_year":"2024","word_count":346,"keywords":["Go","Sam Altman","OpenAI","GPT-5","AI","RPA","llm_models:Claude","GPT","llm_models:Llama","R","llm_models:GPT"],"extracted_tech_keywords":["AI","GPT-5","OpenAI","R","Go","GPT","RPA","llm_models:GPT","llm_models:Claude","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mysterious-gpt2-chatbot-takes-everyone-by-surprise\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10076196,"title":"Cloud Prices Are Hurting Enterprises Bad, But There Might Be Ways Out","content":"The discussions around companies facing an ongoing cloud crisis are everywhere. According to a recent report published by autonomous solution provider company Anodot, almost 50% IT professionals are currently facing difficulties in controlling cloud costs, and almost a third of the respondents witnessed a 25–50% hike in cloud prices. Amidst the debate over increasing prices, cloud economist Corey Quinn expressed concerns over Google’s recent decision to increase its prices even further w.e.f. October 1. “Google is all set to put fire to its reputation and hurt customer sentiments,” Quinn said. Over the past decade, cloud adoption has significantly grown in terms of popularity, mirroring the increased trust in individual cloud providers and cloud models. However, the price hike has changed the game for everyone. Anodot further reports that monitoring costs, optimizing resource use, and forecasting future spend are some challenging areas to tap while managing cloud costs. Leon Kuperman, the chief technology officer at CAST AI, termed the rising cloud prices as cloudflation. In an exclusive interaction with Analytics India Magazine, Vishal Prakash Shah, co-founder & CEO of Synersoft Technologies Private Limited, said, “It is going to be a severe problem in times to come. Cloud infrastructure costs are closely correlated with price trends in energy, real estate, bandwidth, and semiconductors. AWS is a pure cloud infrastructure provider, Google Cloud and Microsoft Azure are infrastructure providers as well as SaaS providers. They have deep pockets to burn money to acquire customers.” Founded in 2008, Synersoft is the maker of disruptive technologies for SMEs, now branded as BLACKbox. It is an incubated and invested portfolio company of CIIE – IIM-Ahmedabad. Enhancing their competitiveness with state of the art IT standardization. Read the full interview below. AIM: How are cloud players tackling cloudflation while maintaining customer centricity in a price-sensitive market like India? Shah: Cloudflation is a severe problem for cloud service providers. There are two types of cloud service providers: 1) offering cloud infrastructure services to customers; and 2) offering software as a service solution to customers. Mostly, category one is a service provider to category two and enterprise customers. Most SaaS providers in India have resorted to the freemium or subsidized business model to penetrate the market and acquire customers. As the cloud infrastructure service providers increase the cloud fees, the input costs for SaaS providers go up. It puts pressure on SaaS companies operating on a freemium or subsidized basis. They deal with such a situation either by increasing their burn rate to retain the customers or by increasing prices, letting the churn-out happen. AIM: How do you perceive the problem of cloudflation on a holistic and ground level? Shah: It is going to be a severe problem in the future. This is mainly because of internal competition in the SaaS industry and the mentality of burning money to win customers. It is like a ‘Catch Me If You Can’ problem. First, companies subsidize fees or offer free services to acquire customers, hoping that once the customer realizes the value of the service, they will pay for it. Then they gradually begin charging customers or increasing fees, discovering a newly invested competitor willing to burn money to win the customer. Cloudflation will worsen this by reducing the runway and intensifying the burn rate. AIM: What are the vital steps companies take to dodge the problem of cloudflation while maintaining prices? Shah: Companies have started exploring the availability of cloud services from cold-climate countries whose data centers are cost-effective. Cloud infrastructure hosting costs in India and some western countries are startlingly different. That’s how one can dodge the problem. But the party is not going to last for too long. The government of India is deliberating on a data localization policy for its citizens’ data, moreover, India will have its own version of GDPR shortly. In my opinion, the cost-effectiveness sought by importing cloud infrastructure services is not a long-term solution. India will face cloudflation, given the rise in infrastructure, bandwidth, energy, and hardware costs. AIM: To settle the spiking cloud price issue, what is your perspective on investing in renewable energy, semiconductor innovation, and existing nuclear power sources as the only alternatives? Shah: Cloud infrastructure costs are closely correlated with price trends in energy, real estate, bandwidth, and semiconductors. Concerning India’s interest, the Vedanta and Foxconn deal and many other joint ventures in the pipeline will be game changers. Just like cold countries have a cost advantage by using less energy to cool off the processors, solar abundance in India will be a balancing factor for us. The possible use of abundantly available solar energy to power data centers coupled with locally available semiconductors will work in favor of cloud-dependent businesses. AIM: Do chip shortages have an impact on public cloud costs? Shah: It does have an impact. Budgets have gone haywire due to acute price rise. The projects to create local cloud infrastructure and bandwidth are delayed due to the short and uncertain supply of chips. AIM: Do you think the lack of competition (since AWS, Microsoft Azure, and Google Cloud are the only key players) is directly related to its growing price? Shah: AWS, MS Azure, and Google Cloud enjoy an obvious oligopoly. They have deep pockets to burn money and acquire customers. The recent price changes by Google and Microsoft substantiate the fact that they can increase the prices of dependent customers owing to the lack of competition. It is the same phenomenon everywhere. AIM: How does your business perceive the problem of geopolitical tensions in cloudflation? Shah: Current geopolitical tensions are the driver of the uncertain and short supply of chips and rising energy prices catalyze cloudflation. Synersoft is severely affected. The components used in our products are costlier, and the supply is uncertain. Google recently discontinued MSME-friendly products such as 100 GB of storage, forcing MSMEs to upgrade all of their users to a 2000 GB subscription if they required more than 30 GB of space for a single user. Cloudflation has made us realize the consequences of adopting widely promoted cloud computing. Businesses like us are seriously thinking of on-premise deployments that have immunity from cloudflation. AIM: What are your thoughts on the growing supercloud trend amidst cloudflation? Shah: Supercloud is a step to lower the exit barrier for switching from one service to another. It is similar to the olden-days practice of continuing with our mobile service providers as we could not afford to lose our widely known mobile number. With portability becoming a reality, we could retain our mobile number and migrate to a more cost-effective service provider. What portability is in telecom, supercloud is in SaaS. It is desired to break the oligopoly and allow customers access to more competitive services. In the absence of supercloud, customers will have to accept unreasonable charges by established cloud providers because they cannot migrate to another cost-effective service provider. The cloud industry would only be dominated by a few monopolistic players without supercloud. To quote a few examples, the recent announcement by Microsoft that it will not support IMAP\/POP protocols will make it very difficult to migrate email systems to other cost-effective providers. Also, the recent policy changes by Google forced customers to upgrade all the users from Business Starter (30 GB) to Business Standard (2000 GB) at four times the cost in the event of a single user requiring more than 30 GB of storage.","excerpt":"Cloudflation is going to be a severe problem in times to come. AWS, Google Cloud and Microsoft Azure have deep pockets to burn money to acquire customers: Vishal Prakash Shah","categories":["AI Features"],"tags":["AWS","cloud infrastructure","Google Cloud","Interviews and Discussions","Microsoft Azure"],"author_name":"Nidhi Bhardwaj","publish_date":"2022-10-02T10:00:00","publication_year":"2022","word_count":1233,"keywords":["Go","cloud infrastructure","Google Cloud","Rust","AWS","AI","cloud computing","R","RAG","Aim","analytics","Microsoft Azure","Azure","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","cloud computing","AWS","Azure","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cloud-prices-are-hurting-enterprises-bad-but-there-might-be-ways-out\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10101528,"title":"Bollywood Goes Gaga Over Generative AI","content":"Taking a cue from Hollywood, Indian actor Anil Kapoor recently safeguarded his digital persona. The Delhi High Court made a historic decision, safeguarding his ‘personality rights’ and acknowledging the misuse of AI tools to create deep fakes and explicit videos. ​​The court ordered that 16 entities that were using his name and image without his permission will be “restrained from in any manner utilising Anil Kapoor’s name, likeness, image, voice or any other aspect of his persona to create any merchandise, ringtones … either for monetary gain or otherwise.” While the move by Anil Kapoor is appreciable, it raises the question whether the Indian film industry will resist the adoption of generative AI and deep fakes to create new stories and content similar to Hollywood. Navigating their way Amrit Thomas, the chief data officer at Zee Entertainment, shared his insights at Cypher 2023 about Zee’s rapid adoption of generative AI to create innovative and ethically sound content, aiming to enhance the audience experience. “We are training our models on our own IP” he said clarifying that Zee is being careful to only use content that belongs to them, staying within their own creative boundaries. Moreover, Thomas believes it’s essential to involve humans when creating content with generative AI. “It’s like a new canvas to paint, a platform to write your story. That’s why I have an issue with the term ‘fully AI generated’,” he said, emphasising that there will never be a completely AI-generated film; humans will always play a role in the creative process. “I’m using AI to enhance my creativity and intuition. What happens in day to day work is I get so bogged down in getting that story out and trying to sell it that I don’t have time for my creativity to do work. The role of AI is to allow you as a screenwriter, to really explore your audience and  that needs space. AI gives you that space.” On similar lines, The Rabbit Hole, a brand film agency that produced the ‘Brand Film for the ICC World Cup 2023,’  recently revealed in an exclusive interview with AIM that the company has integrated generative AI tools such as Dall-e and Midjourney to enhance its video creation process. “These tools are particularly valuable during the initial stages of storyboarding and pre-visualization, helping us generate visual concepts,” the company said. Biren Ghose, country head, India & global excom member, Technicolor Creative Studios during Cypher explained that in the future how IP works is going to be changed. He said that famous personalities will have an IP for both their digital and physical avatars. “You may have a manager, managing your physical avatar, and then that person will license your avatar for a particular ad commercial or a movie or for an advertisement,” he said. Addressing the strikes and the challenge that the actors in the west are facing, he said that Technicolor Creative Studios will make sure that your avatar is scanned and that you are now a digital virtual person. Moreover, he explained that the usage of someone else’s IP would be subject to specific applications outlined in contractual agreements, ensuring a regulated and ethical framework. He gave an example of a recent commercial which Technicolor Creative Studios created for Mastercard where it used the AI based neural rendering in order to create Messi’s image rendering an entirely new visual effects method using no CGI. In this case he explained that the company got the necessary permissions from that personality to produce the commercial. Also, Bollywood legend Amitabh Bachchan recently partnered with Ikonz Studios to explore generative AI. This venture aims to merge cultural icons and iconic intellectual properties into interactive mediums using AI technology. What separates India from the West The SAG-AFTRA strike has persisted for over 3 months, and recently, negotiations between Hollywood actors and studios have tragically collapsed, extinguishing hopes of ending the performers’ strike. “Unlike the Western world, in the eastern part, our areas are far more collaborative and collectivist,” said Thomas, comparing the situation with Hollywood and explaining that it is time to bring this collaborative culture into the AI domain as a force for good. Thomas further said that as a community, we need to be AI evangelists. The leverage of the technology should be for good. In the end, humans created AI. “I’m an optimist. I believe in humanity and human beings’ ability, the right to overcome every hurdle along the way.” Comparing the strategies employed by both industries, it’s evident that Bollywood has shown maturity in embracing generative AI. Unlike Hollywood, India hasn’t witnessed strikes related to AI usage; instead, numerous instances highlight their willingness to integrate this innovative technology into content creation.","excerpt":"Unlike Hollywood, India hasn’t witnessed strikes yet","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-10-16T17:39:38","publication_year":"2023","word_count":784,"keywords":["Go","API","DALL-E","AI","Git","RAG","Aim","ViT","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","R","Go","Git","API","DALL-E","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/bollywood-goes-gaga-for-generative-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10039777,"title":"How 2020 Turned It Around For Zoom: Interview with Sameer Raje, Head of India","content":"What do Google, Xerox and Zoom have in common? You guessed it right; all are now verbs. Google became a stand-in for search, Xerox for photocopy and Zoom for video conferencing. Zoom entered the verb club as late as 2020. A shift to remote working in the wake of the pandemic accelerated Zoom’s transformation from a fledgeling startup to one of the most valuable companies in the world (valued at $117 billion, as of December 2020). Analytics India Magazine caught up with Zoom’s General Manager and Head of India, Sameer Raje, to better understand the platform, its exponential growth, and security concerns. Excerpts: AIM: What changed in 2020 for Zoom Communications? Sameer Raje: Nobody anticipated what was coming. Zoom started as a frictionless or a friction-free video-first platform in 2011. Meaning, Zoom is designed to offer video first communication, and everything else is built around it. Other platforms have video as the last component. When we built our platform, we built it to scale, adjust to fluctuating bandwidth, and cater to the rising demand. As many as 40,000 attendees can participate in a video webinar and still have a flawless meeting experience on our platform. The simplicity of the platform struck a chord with the people. Zoom took off by February 2020. A simple interface, easy download option etc made Zoom stand out from similar platforms. Big corporates are quick to adopt the latest technologies. On the other hand, you have segments such as the traders’ community with little exposure to technology. When the pandemic struck, even those segments were forced to adapt. Zoom was highly favoured because it is simple, easy to use, and works even in remote locations with low internet bandwidth. The ability to stay connected at the lowest possible data packet is the beauty of the platform. AIM: How did Zoom scale quickly to handle the sudden surge in users? Sameer Raje: We operate data centres across the world. Some of these data centres are our own, while others are co-location. We already have two data centres in India–one in Mumbai and the other in Hyderabad. Being a cloud service, it is very easy to augment our capacity. In India, we work with some of the most renowned telcos. We are co-located in their data centres, while also maintaining a few of our own. So it is very easy to scale up in a short period. In addition, we work with partners such as Amazon and Oracle to support us in this process in India and in other locations. AIM: What are your comments on the privacy concerns around Zoom? Sameer Raje: There are multiple aspects to the privacy issue. We take privacy very seriously, and we have the best practices in place, such as encryption technology, high-end security protocols, etc. A lot of technology platforms are commonly used in corporate set-ups. To make the whole experience secure, companies have a slew of measures in place. However, when the pandemic began, almost everyone started working on virtual platforms. Even people who were otherwise not exposed to technology on such a scale were now forced to adapt. Unaware of consequences, there were several lapses of judgement on the users’ end. Many started sharing meeting links, sometimes even passwords on public platforms, paving the way for bad actors to break in. It is not just with Zoom, but in general too, cybersecurity issues saw a 40 percent increase during this time. When people think of privacy and security, they need to understand that we don’t store any information. Unless you choose to record, the meeting and its content vanish once the call ends. That said, we are not averse to saying there were minor mistakes from our side. As a corrective measure, we went on a ‘90-day plan’, hired third-party vendors to analyse our platform’s security features. We also interacted with our users to help pinpoint issues. Our CEO Eric Yuan personally spoke to customers, which helped us improve and provide the kind of support our customers need. We made many changes: we updated our password requirement from 8 digits to 10 digits; went from 256-bit encryption to 256-bit AES GCM encryption; introduced a security tab; and designed a coaching and training plan for certain users. AIM: Tell us about the 90-day plan. Sameer Raje: In the first few months of 2020, Zoom experienced a sudden demand unlike ever seen before. We realised we needed to include an ‘equivalent focus’ on security and privacy issues as we began hosting hundreds of millions of daily participants. Starting April 1, 2020, we rolled out a 90-day program to focus on seven commitments towards the security and privacy aspects of our platform: Enacting a feature freeze for 90 days to redirect all engineering resources and efforts towards trust, safety and security issues.Conducting a comprehensive review with third-party experts and representatives.Preparing a transparent report that details information on requests for data, records, or content.Enhancing the current bug bounty program.Launching a CISO council in partnership with leading CISOs from across industries to facilitate dialogue on security and privacy best practices.Conducting a series of simultaneous white box penetration tests.Host a weekly webinar to provide privacy and security updates to our community. AIM: Indian government raised a few concerns on security and privacy with Zoom. What was the company’s response to that? Sameer Raje: At Zoom, we are transparent and honest in sharing the information. The condition is the same in India as it is in the US where all our information and company holding patterns are available in the public domain. In India, we engaged with the Ministry of Electronics and Information Technology. We have shared the relevant technology of our platform, information on our security practices, and how we take care of our users. We also underwent stringent validation checks. AIM: What are your plans for the Indian market? Sameer Raje: India is a huge market in terms of consumption. The country also has a huge talent pool. Our goal is not only to tap into the potential market but also the talent. We have announced our Tech Centre in Bangalore, and we are hiring big time; I’m excited because that will be the largest of its kind, outside of the US. Our Indian team is also growing. It was just me two years ago, but now the headcount has grown to 100 and will continue to grow in the future too.","excerpt":"What do Google, Xerox and Zoom have in common? You guessed it right; all are now verbs. Google became a stand-in for search, Xerox for photocopy and Zoom for video conferencing. Zoom entered the verb club as late as 2020. A shift to remote working in the wake of the pandemic accelerated Zoom’s transformation from […]","categories":["AI Features"],"tags":["encryption","Interviews and Discussions"],"author_name":"Shraddha Goled","publish_date":"2021-05-07T16:00:00","publication_year":"2021","word_count":1069,"keywords":["Go","startup","programming_languages:R","AI","Git","Aim","encryption","analytics","Rust","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","Rust","Git","GAN","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-2020-turned-it-around-for-zoom-interview-with-sameer-raje-head-of-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60077,"title":"AI4ALL Open Learning Program Provides Free Lesson On AI’s Role In Current Crisis","content":"AI4ALL Open Learning is a free program which community-based organisations, as well as teachers, can use the resource to educate the community and high-school students about Artificial Intelligence. Recently, AI4ALL announced a new module which high school teachers with a free online lesson to engage students around AI’s role in the current crisis. In the program, the ExploreAI curriculum can be implemented in 10, 20 or 30 hrs. The reason behind this program is to support the community by sharing relevant, free curriculum and teaching resources. AI4All is trying every possible way to ensure the best to serve all people in the work to increase diversity and inclusion in artificial intelligence. For educators and students, we recognize COVID-19 poses unique challenges. AI4ALL Open Learning’s new module provides high school teachers with a free online lesson to engage students around AI’s role in the current crisis. Get access here: https:\/\/t.co\/x7p5glnEje— AI4ALL (@ai4allorg) March 18, 2020 As mentioned, the ExploreAI program is implemented into three durations which are 10, 20 and 30 hrs. ExploreAI 10-Hour curriculum includes an introduction to AI, limitations and abilities of AI, use of AI implementation, the design process in AI and other such.ExploreAI 20-Hour curriculum includes an introduction to AI and data, how to explore data in AI, where and how to collect data, how to use data to perform tasks and gain insights, and other such. ExploreAI 30-Hour curriculum includes 7 courses where one can learn about AI and data, what are the career paths in AI and how to explore the variety of jobs in AI, how to create presentations to educate the community about AI and other such. The AI4ALL Open Learning Program also includes a course called AI & COVID-19 where one can learn how AI is being used to tackle varying parts of the COVID-19 outbreak and how AI can be used to help the situation. Click here to join the Open Learning Program.","excerpt":"AI4ALL Open Learning is a free program which community-based organisations, as well as teachers, can use the resource to educate the community and high-school students about Artificial Intelligence. Recently, AI4ALL announced a new module which high school teachers with a free online lesson to engage students around AI’s role in the current crisis. In the […]","categories":["Deep Tech"],"tags":["learn ai"],"author_name":"Ambika Choudhury","publish_date":"2020-03-26T16:53:07","publication_year":"2020","word_count":322,"keywords":["artificial intelligence","programming_languages:R","AI","GAN","learn ai","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ai4all-open-learning-program-provides-free-lesson-on-ais-role-in-current-crisis\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10043402,"title":"Ansible vs Docker: A Detailed Comparison Of DevOps Tools","content":"The Emergence of Software and the Internet have both led to industrial transformation, whether it be from shopping to entertainment or banking. Companies these days interact with their customers through the software delivered either as an online service or applications, supported on all sorts of devices. DevOps is a concoction of work practices, ethics and tools designed to increase an organization’s potential to deliver its applications and services faster than conventional software development processes and techniques. Such a combination of practices, providing services at high velocity, evolving and improving products faster aids to better delivery and software development, helping the infrastructure management processes in turn. This work speed enables organizations to serve their customers better and compete more effectively in the open global market. In addition, DevOps helps remove the barriers between siloed and segregated working teams, development and operations and help a whole organization work even better under one roof. Under a created DevOps model, the development and operations teams work together across the entire software application life cycle, starting from development and test, getting through deployment to its operations. There are a set of key practices that help organizations innovate faster through automating and streamlining their software development and infrastructure management processes when inculcated and accomplished with proper tooling backend. Such practices help organizations adapt to changing markets better and become more efficient at driving business results. Certain reliability practices like continuous integration and continuous delivery can accordingly ensure the quality of application updates and infrastructural changes to deliver applications while maintaining an optimum experience for end-users reliably. Using a microservice architecture might help make the delivered applications more flexible and enable quicker innovation. The microservices architecture segregates very largely complex systems into simple and independent projects to focus on. The use of monitoring and logging services helps system engineers track deployed applications and infrastructure to react quickly to problems and provide essential services. What are DevOps Tools? DevOps tools are certain software services that ensure transparency, automation, and collaboration for organizations to stay at the forefront of the value stream. These tools facilitate effective sharing and exchange of information within the organization and technical know-how between all stakeholders, be it development, operations, security or business teams, for effective product output. The tools help firms resolve most of the challenges faced with the implementation of DevOps practices. However, there is no one solution available that fits all and takes care of everything. Hence, there are a wide variety of DevOps tools available for every requirement. In this article, we will be exploring two DevOps Tools in particular, Ansible and Docker. What is Ansible? Ansible is an open-source automation engine that helps in DevOps and comes to the rescue to improve your technological environment’s scalability, consistency, and reliability. It is mainly used for rigorous IT tasks such as configuration management, application deployment, intraservice orchestration, and provisioning.  In recent times, Ansible has become the top choice for software automation in many organizations. Automation is one of the most crucial aspects of industries these days. Unfortunately, many IT environments are too complex and often require to be scaled too quickly for system administrators and developers to keep up, rather than manually. Automation simplifies several complex tasks, making developers’ jobs more manageable and focusing their attention on other tasks and areas that add value to an organization. To sum it up, it frees up time and increases efficiency. Working on Ansible requires no special coding skills and with a very simple set of instructions are necessary to use Ansible’s playbooks. It enables one to model even highly complex IT workflows; the entire application environment can be planned and orchestrated no matter where it’s being deployed. Customization can be set up based on the needs. With Ansible, there is no need to install any other software or firewall ports on the client systems which are to be automated, even no need to set up a separate management structure!. And Because you don’t need to install any extra software, there’s more room for the application resources on the development server. Features Of Ansible Some of the most important features that Ansible offers are : Configuration ManagementApplication Deployment Planning & OrchestrationSecurity & Compliance Cloud Provision Configuration Management As Ansible is supposedly designed to be very simple to leverage reliability and consistency for configuration management, one with an IT background can quickly get it up and running it. Ansible’s configurations are simple data descriptions of the infrastructure and are readable by humans and parsable by machines. To start the managing systems, all required is a passcode or a Secure Socket Shell network protocol key. For example, suppose one wants to install an updated version for a specific type of software on all the machines present in the enterprise premise. In that case, all one has to do is write out all the IP addresses of remote hosts and write an Ansible playbook to install it on all the host nodes and then run the playbook from the control machine. Application Deployment Ansible lets one quickly and easily deploy multi-tier apps. There is no need to write a custom code to automate the systems present. Just list the tasks required to be completed by writing a playbook, and Ansible figures it out on how to get the systems to the state you want them to be in. To simplify,  there is no need to configure the applications on every machine manually and individually, which would be a rather tedious task. When you run a playbook from the set and configured control machine, Ansible uses Secure Socket Shell to communicate with all remote hosts present within its network environment and run all the set desired commands. Planning & Orchestration As the name tells us, orchestration involves bringing the different elements of a delivery pipeline into a beautifully run whole operation similar to a beautifully organized music opera. Knowing each instrument’s role and essence, things are planned around accordingly. For instance, for application deployment, one would need to manage the front-end and back-end services and the databases, networks, storage, etc. You also have to make sure that all the tasks are being handled in the proper order. Ansible helps create automated workflows, provisioning, and to make orchestrating and planning tasks easy. Furthermore, the infrastructure defined in the Ansible playbooks can be used as the same orchestration wherever you need to due to the portability of Ansible playbooks. Security & Compliance Applied security policies, such as firewall rules or locking down users, can be implemented and other automated processes present. If security settings are configured on the control machine, and the associated playbook is run, all the remote hosts present will automatically be updated with those settings. So the need to monitor each machine for security compliance manually is eliminated out of the picture. The admin’s user ID and password are not retrievable in plain text on Ansible for extra security. Cloud Provision The first step in automating an applications’ life cycle is automating the infrastructure. With Ansible, you can plan your cloud platforms, virtualized hosts, network devices, and bare-metal servers. The Ansible Architecture Ansible’s Architecture comprises of the following components : Module: Small programs that Ansible uses to send out from a control machine to all the nodes or remote hosts. The modules are executed altogether using playbooks, and they control things such as services, packages, and files. Plugins: Ansible comes with a number of its plugins, which are extra pieces of code that augment functionality. Inventories: From the registered inventory, you can assign variables to any of the hosts using a simple text file, where the information of all the IP addresses, databases, servers is present. Playbooks: Describe the tasks that are to be given priority and without the need for the user to know or remember any particular syntax. Playbooks are instruction manuals for tasks. APIs: Application programming interfaces available help extend Ansible’s connection types, callbacks, and more. Ansible Commands Some of the widely used basic Ansible Commands are as follows : To Verify connectivity of host:  # ansible  <group> -m -ping To Reboot the host systems:  #ansible <group> -a “\/sbin\/reboot” Create a new user: # ansible <group> -m user -a “name=ansible password=<encrypted password>” Delete a user: # ansible <group> -m user -a “name=ansible state=absent” To perform File Transfer to more than one server :  # Ansible abc -m copy -a “src = \/etc\/yum.conf dest = \/tmp\/yum.conf” To Reboot more than one server : # Ansible abc -a “\/sbin\/reboot” -f 12 How Does It Work? Ansible works by connecting to your server with Secure Socket Shell protocol and hereby pushing out smaller programs known as the Ansible modules. Ansible’s most powerful feature while creating playbooks is the containment of YAML code. As a result, users can program repetitive tasks automatically, i.e. automate without learning an advanced programming language. What is Docker? Docker is an open-source platform application for developing, shipping, and running applications. It enables developers to package applications into containers, a set of standardized and executable components that combine the application source code with the operating system libraries and dependencies required to run that code in an executable environment. Containers can even be created without Docker, but the platform and user interface make it easier, simpler, and safer to build, deploy and manage containers. Docker enables developers to perform operations such as build, deploy, run, update, and stop on containers using simple commands and work-saving automation through a single API. In addition, Docker enables you to separate the created applications from the present infrastructure to help deliver the software quickly. Using Docker, infrastructure can be managed similarly to manage applications. Implementing Docker’s methodologies for shipping, testing, and deploying code quickly, significantly reduces the delay between writing code and running it into production. Docker provides the tooling and a platform to manage the lifecycle of created containers: Application and its supporting components can be developed using containers. The container hence becomes the unit for distributing and testing metrics of your application. The application can then be deployed into the production environment as a container or an orchestrated service. Works the same even if your production environment is a local data centre, a cloud provider or a hybrid of the two. Features Of Docker Some important features of Docker are : Application isolation: Docker provides containers that can be used to run applications in an isolated environment. Since each container is independent of one another, Docker can execute any kind of application as defined. Swarm: Swarm, a clustering and scheduling tool for Docker containers, uses the Docker API on the front end, making it easy to use with various tools to control it. It comprises a self-organising group of engines that enables pluggable backends.Security Management: It saves important code elements directly into the swarm and chooses to give services access to only certain protocols, including a few important commands to the engine such as secret inspect, secret create, etc.Software-defined networking: Docker supports Software-defined networking, and hence without having to touch a single router, the Docker CLI and Engine enables users to define isolated networks for containers. Operators and Developers can design systems with complex network topologies and define the networks into the configuration files.Ability to Reduce the Size: Since it provides smaller footprints of the OS via containers, Docker can help to reduce the size of the current development. The Docker Architecture Docker uses a client-server architecture where the Docker client communicates to the Docker daemon, which performs several heavy processing tasks such as building, running, and distributing the created Docker containers. The Docker client and daemon can run on the same system, or the Docker client can be connected to a remote Docker daemon. The Docker client and daemon communicate using a REST API over UNIX sockets or a network interface. Another Docker client is Docker Compose, which lets you work with applications consisting of a set of containers. The Docker daemon The Docker daemon known as dockerd listens for the Docker API requests and manages several Docker objects such as images, containers, networks and volumes. To manage Docker services, a particular daemon can also communicate with other daemons. The Docker client The Docker client docker provides the communication pathway the way through which many users interact with Docker. Commands such as docker run, to execute the docker are sent by the client to dockerd, which executes them through. The docker command uses the Docker API. The Docker client can communicate with more than one daemon. Docker registries A Docker registry stores all the relative Docker images. Docker Hub is a public registry that anyone can use and access, and Docker is configured in a certain way to look for images on Docker Hub by default. Private registries can be created and run. When commands such as docker pull or docker run are executed,  templates with instructions known as images are pulled from the configured registry. With the docker push command, your image is pushed to your configured registry. Docker objects When you use Docker, you create and simultaneously use images, containers, networks, volumes, plugins, and other objects. The two of the  most important docker objects are : ImagesContainers Images An image in a Docker is a read-only template with instructions for creating a new Docker container. Most of the time, an image template is based on another image, with some additional customizations. One may build an image based on the Ubuntu image, but install the apparent Apache web server and the configuration details needed to make the application run on the system. One can create his own image or simply use those created by others and published into a registry. To build an image, create a Dockerfile with a simple syntax that defines the steps needed to create the image and run it. Each instruction present in the Dockerfile creates a layer in the image. When a Dockerfile is changed or the image is rebuilt, only those layers which have changed are rebuilt. This makes using images so lightweight, small, and fast compared to other similar virtualization technologies. Containers A container is a runnable instance defined by the docker image. You can perform operations such as   create, start, stop, move, or delete a container using the Docker API or CLI. A single container can be connected to one or more networks; storage can be attached to it or create a new image based on its current state. Generally, a container is relatively well isolated from other containers and its host machine for security. Container’s network, storage, or other underlying subsystems and how isolated they remain from other containers or the host machine can be defined and controlled. A container is defined by its image and other configurations set and provided by the user when creating or starting it. When a container is removed or deleted, any changes to its state that are not saved and stored in persistent storage disappear automatically. Docker Commands Some of the basic and widely used commands for Docker are as follows : To know the Docker Version: ~$ docker --version To pull information using docker images: ~$ docker pull ubuntu Create a container: ~$ docker run -it -d ubuntu Get a list of running Containers: ~$ docker ps -a Access a running Container: ~$ docker exec -it <Container name> To stop a running Container: ~$ docker stop <Container name> Shutting down a Container completely : ~$ docker login Where can Docker Be Used? To Deploy highly available, fully managed Kubernetes clusters. To Deploy and run apps across on-premises, edge computing and public cloud environments using a cloud service. Simplify and consolidate the data lakes across the organization by seamlessly deploying container-enabled enterprise storage. Docker Compose can be used to manage the application’s architecture. EndNotes Both Docker and Ansible provide a wide array of uses and can be integrated according to the requirements and needs. For example, one makes use of modules, while the other makes use of containers to convey or store information. In addition, both the DevOps tools can be used to create automation services across the enterprise with their unique set of properties and capabilities. References Docker Official SiteAnsible Official siteUsing Docker with IBM CloudHow Ansible Works","excerpt":"The Emergence of Software and the Internet have both led to industrial transformation, whether it be from shopping to entertainment or banking. Companies these days interact with their customers through the software delivered either as an online service or applications, supported on all sorts of devices. DevOps is a concoction of work practices, ethics and […]","categories":["AI Trends"],"tags":["Automation","DevOps","Docker"],"author_name":"Victor Dey","publish_date":"2021-07-12T17:00:00","publication_year":"2021","word_count":2698,"keywords":["Docker","TPU","AI","ML","Automation","docker","RAG","microservices","Ray","edge computing","DevOps","R","kubernetes"],"extracted_tech_keywords":["AI","ML","Ray","RAG","kubernetes","docker","microservices","edge computing","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ansible-vs-docker-a-detailed-comparison-of-devops-tools\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10118853,"title":"Microsoft Introduces Phi-3, LLM That Runs on the Phone","content":"While speaking to AIM, Harkirat Behl, one of the creators of the Phi model, said that his team was working on the next version of Phi-2 and making it more capable. “Phi-1.5 started showing great coding capabilities, Phi-2 was code with common sense abilities, and the next one would be even more capable,” he said. Microsoft has now unveiled Phi-3-Mini, a 3.8 billion parameter language model trained on an extensive dataset of 3.3 trillion tokens. Despite its compact size, Phi-3-Mini boasts performance levels that rival larger models such as Mixtral 8x7B and GPT-3.5. “One of the things that makes Phi-2 better than Meta’s Llama 2 7B and other models is that its 2.7 billion parameter size is very well suited for fitting on a phone,” said Behl. Phi-3 makes it even better now. For instance, Phi-3-Mini achieves 69% on the MMLU benchmark and 8.38 on the MT-bench, making it suitable for deployment on mobile phones. Phi-3-Mini, being highly capable, can run locally on a cell phone. Its small size allows it to be quantized to 4 bits, occupying approximately 1.8GB of memory. Microsoft tested the quantized model by deploying Phi-3-Mini on an iPhone 14 with an A16 Bionic chip, running natively on the device and fully offline, achieving more than 12 tokens per second. The innovation behind Phi-3-Mini lies in its training dataset, an expanded version of the one used for its predecessor, Phi-2. This dataset comprises heavily filtered web data and synthetic data. The model has also been optimised for robustness, safety, and chat format. Microsoft has also introduced Phi-3-Small and Phi-3-Medium models, both significantly more capable than Phi-3-Mini. Phi-3-Small, with 7 billion parameters, utilises the tiktoken tokenizer for improved multilingual tokenization. It boasts a vocabulary size of 100,352 and a default context length of 8K. The Phi-3-small 7 billion parameter model achieves an MMLU score of 75.3 outperforms Meta’s recently launched Llama 3 8B Instruct with a score of 66. The model follows the standard decoder architecture of a 7B model class, featuring 32 layers and a hidden size of 4096. To minimise KV cache footprint, Phi-3-Small employs a grouped-query attention, with four queries sharing one key. Additionally, it utilises alternative layers of dense attention and a novel blocksparse attention to optimise KV cache savings while maintaining long context retrieval performance. An additional 10% multilingual data was used for training this model. However, Phi-3-Mini has its limitations. While it demonstrates a similar level of language understanding and reasoning ability as much larger models, it is fundamentally limited by its size for certain tasks. For example, it lacks the capacity to store extensive “factual knowledge,” resulting in lower performance on tasks such as TriviaQA. Microsoft believes such weaknesses can be addressed by augmenting the model with a search engine. Additionally, the model’s language capabilities are mostly restricted to English, highlighting the need to explore multilingual capabilities for Small Language Models.","excerpt":"Despite its compact size, Phi-3-Mini boasts performance levels that rival larger models such as Mixtral 8x7B and GPT-3.5.","categories":["AI News"],"tags":["LLMs","Microsoft"],"author_name":"Mohit Pandey","publish_date":"2024-04-23T08:32:10","publication_year":"2024","word_count":481,"keywords":["synthetic data","AI","LLMs","innovation","ML","GPT","CuPy","Aim","small language models","R","Microsoft","llm_models:GPT"],"extracted_tech_keywords":["AI","ML","Aim","CuPy","small language models","R","GPT","synthetic data","innovation","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-introduces-phi-3-llm-that-runs-on-the-phone\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":56691,"title":"How This Conversational AI Startup Is Solving Job Market Inefficiencies In India","content":"India is currently the hotbed for emerging technologies like AI & machine learning. But the need of the hour is “Deep Tech” which fundamentally is a connection of different types of technologies, not just AI & ML but also computer vision, image processing, blockchain, and AR\/VR, to come up with a solution that is not trivial but a revolutionary one. However, there are a few limitations such as the skills of the engineering graduates in India to meet this demand, the quality of data in India continues to be an issue because of the diversity of demographics, languages etc. in the Indian market which could slow the momentum. With a passion for solving such issues in India, Bangalore-based Dhiyo is utilising AI and machine learning to help the folks to solve job market inefficiencies for the blue-collar sector by building a quick prototype using Google Assistant. Founded by Santhosh SS in 2018, Dhiyo is a Conversational AI interface for the job seekers in the informal sector who can create their profile or resume just by speaking to their smartphone through a series of voice commands in their own native language. Dhiyo leverages state-of-the-art NLP(U) and ML techniques to deliver the industry’s most robust and engaging conversational experience. Flagship Product The flagship product of Dhiyo is a Vernacular Conversational AI Platform. It is a conversational voice interface where the job seekers in the informal sector can create their profile\/resume just by speaking to their smartphone through a series of voice commands in their own native language. Through this open-ended, natural and dynamic conversations, the company is able to gather deep candidate insights as well as build trust and confidence with the job seekers and connect them with the right employers. The voice-first and mobile-centric platform improve employability by enabling, empowering and engaging users in India by helping them build their digital identity (resume), discover gainful jobs as well as connecting them with relevant employers after background verification. Use of AI and ML The speech engine at Dhiyo evaluates all possible meanings and interpretations of the utterances to determine what the user is saying and asking. The NLU engine supports native Indian languages by building segmentation support for different languages and allows training text and utterances in the appropriate language. The platform is based on some fundamental building blocks for conversational AI like natural language understanding (NLU), intent identification, information extraction, action triggers, query understanding and transformation, natural language response generation, speech processing, personalisation, etc. Core Tech Stack The core tech stack that Dhiyo is using to build its product is a loosely coupled architecture where the components are highly modular. Some of the frameworks that are being used are React Native, Angular, nodeJS, expressJS, MongoDB and multiple customised ASR\/NLP engines in the backend. Talent Crunch Currently, Dhiyo is a team of 10 people and is looking to grow. Talking about talent, Santhosh explained that the company is very keen on getting talent with not only best aptitude but also an attitude that aligns with our company culture. He said, “We want to build a workplace environment that values creative problem solving, open communication and a flat hierarchy.” Roadmap Currently, the market is huge for conversational voice AI with vernacular language, especially for the Indian market and even outside India. According to the founder, Dhiyo is built to understand the pain points of the masses and provide intelligent solutions. In the coming years, the product of Dhiyo envisions to create a huge socio-economic impact by not imposing English literacy as a pre-condition for being members of the new knowledge & digital economy.","excerpt":"India is currently the hotbed for emerging technologies like AI & machine learning. But the need of the hour is “Deep Tech” which fundamentally is a connection of different types of technologies, not just AI & ML but also computer vision, image processing, blockchain, and AR\/VR, to come up with a solution that is not […]","categories":["AI Startups"],"tags":["Conversational AI","most revolutionary ai deep learning company"],"author_name":"Ambika Choudhury","publish_date":"2020-02-14T16:00:00","publication_year":"2020","word_count":600,"keywords":["Go","machine learning","AI","MongoDB","ML","computer vision","RAG","NLP","most revolutionary ai deep learning company","Conversational AI","Rust","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","computer vision","RAG","MongoDB","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-conversational-ai-startup-is-solving-job-market-inefficiencies-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":27900,"title":"Visualisation Recommendation Sees Whole New World When It’s Boosted Through Machine Learning","content":"Data visualisation is increasingly being used across many businesses as well as in research communities, and the key reason is better insights. However, when working with data visualisation tools, not everyone is comfortable with the level of abstraction associated with it. For example, in the field of data science, basic data visualisation and data exploration are sometimes more than sufficient. Here, complex manual visualisation models are not used at all. This is where visualisation recommender systems come into play. They provide a variety of visualisations so that the user does not need to specify each and every time he\/she is working on a similar project. But this does not hold true for all applications — for example, the rule-based approaches. With improvements in ML, these limitations can be dealt with easily. This article lays out an interesting research study by scholars at MIT, where they chart out an ML-based approach for visualisation recommendation. ML-Based Visualisation Recommendations In contrast to rule-based recommender systems, which are usually sophisticated and incorporate insights from experts for visualisations, ML-based recommender systems deal directly with data to train models and embed them into their systems. For example, Data2Vis, an automatic visualisation generating model developed by IBM researchers, uses a recurrent neural network (RNN) built on sequence-to-sequence modelling. With significant accuracy in visualisations, Data2Vis generates them within seconds and can aid in complex visualisation without referring other manually-created visualisations. The researchers’ ultimate goal was to provide a more accessible platform, speed up visualisations and make it easier for anyone to work with data visualisation regardless of his or her programming skills. In addition, data exploration is improved further with initial visualisations available for complex tasks. Likewise, there are other tools based on ML to generate visualisations. But, all of these, including Data2Vis, face the limitations in areas such as training models, validation method or even compatible integration with other visualisation systems. A Sui Generis System Called VizML Kevin Hu and team from Massachusetts Institute of Technology have recently worked on a new ML-based visualisation method called VizML. This technique learns visualisation designs choices from a collection of different datasets and their associated visualisations. The design choices, which are selected by the researchers, are trained with a neural network using Pytorch and the models are developed with scikit-learn. Here, visualisation is specifically defined on the basis of design choices. “We describe visualisation as a process of making design choices that maximise effectiveness, which depends on dataset, task, and context. Then, we formulate visualisation recommendation as a problem of developing models that learn make design choices, train and test these models using one million unique dataset visualisation pairs from the Plotly Community Feed.” Design choices through visualisation process, specified by analysts (Image courtesy: Kevin Hu and team) VizML entails five steps in its ML-based approach: Problem Formulation: This step involves the process of making visualisation design choices as mentioned earlier. Data Processing: It involves data collection, cleaning and feature extraction. Predicting Design Choices: Now, the neural network is checked to see if it predicts the design choices. Feature Importance: It involves interpreting features ideal for prediction. Crowdsourced Benchmark: Comparison of this model with visualisation models evaluated by human experts and see if it is better and effective. Visualisation is done through encodings (specifying parameters) on tools such as Vega-lite and Tableau. Design choices are based on these visualisations and are evaluated by an objective function equation (mentioned as visualisation effectiveness in the study). If all the design choices fall under the required criteria, preferred design choices, then it is considered. Otherwise, it will be ignored. Mathematical analysis of this can be found here. For data collection and cleaning, Hu and the team use Plotly. A total of 2.3 million dataset-visualisation pairs describing each dataset and column with features was taken. “Using the Plotly API, we collected approximately 2.5 years of public visualisations from the feed, starting from 2015-07-17 and ending at 2018-01-06. We gathered 2,359,175 visualisations in total, 2,102,121 of which contained all three configuration objects, and 1,989,068 of which were parsed without error.” Data processing and design choices study flowchart (Image courtesy: Kevin Hu and team) The ‘Plotly corpus’, as described in the study, is now subjected to feature extraction. With 81 single-column features and 30 pairwise-column features, they are converted into scalar values using 16 aggregating functions. Consequently, through feature processing, only 119,815 datasets and 287,416 columns are taken for prediction tasks. Once set, a fully-connected feedforward neural network is built and implemented through PyTorch. This is again optimised with respect to learning rate and weight ratio. Finally, it is trained and tested for the obtained data. Conclusion: VizML predicted visualisations with an accuracy of 70 to 95 percent. Furthermore, the crowdsourced benchmarking done by the researchers also aligns with similar accuracy. This high success by VizML tells us ML can vastly come up with better visualisation than us. Ultimately, it all depends on the parameters such as design choices if there is a need for data visualisation. On the other hand, it also depends on applications that rely on data visualisation.","excerpt":"Data visualisation is increasingly being used across many businesses as well as in research communities, and the key reason is better insights. However, when working with data visualisation tools, not everyone is comfortable with the level of abstraction associated with it. For example, in the field of data science, basic data visualisation and data exploration […]","categories":["AI Features"],"tags":[],"author_name":"Abhishek Sharma","publish_date":"2018-09-01T06:17:52","publication_year":"2018","word_count":845,"keywords":["data science","scikit-learn","Go","Plotly","PyTorch","neural network","AI","ML","Scala","R"],"extracted_tech_keywords":["AI","ML","neural network","data science","PyTorch","scikit-learn","Plotly","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/visualisation-recommendation-sees-whole-new-world-when-its-boosted-through-machine-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":27352,"title":"Artificial Intelligence Is A Key To Future International Relations Dynamics","content":"International security and foreign affairs seem to be the newest hotbeds for artificial intelligence innovation and applications. The multidimensionality of today’s machine learning and AI is already seen to have a profound impact on how countries manage their foreign affairs — one of the most recent cases being China. Analytics India Magazine had reported earlier that the Chinese government is not only researching how they can leverage AI to build better foreign policies, but also how to implement them. Bureaucrats and politicians along with researchers in China have come to think of AI as an assistant in their day-to-day decision making. The AI has been trained on data that relates to international politics, domestic politics and other related issues. It can therefore be trained and can come up with policy suggestions and recommendations in a matter of seconds. Chinese President Xi Jinping had pushed the AI agenda in foreign policy and had called for efforts to “break new ground” in diplomacy. Reportedly, the programme built by the Chinese government takes a huge amount of data, which has information that ranges from cocktail-party gossip to images taken by spy satellites! Industrial Revolution, AI And Foreign Affairs As with AI, we have also observed that the industrial revolution of the 18th century also had a great impact on international affairs and trade and relations. It increased productivity and wealth in western societies like the US and the UK. It is because of this advantage that western countries took a decisive lead in the competition between countries and excelled more than the others. This lead is very apparent in the size of the economies, the development that has taken place and the increase in the living standards of these countries. In some ways, the current AI revolution is more powerful than the industrial revolution because it has engulfed not only the industries but also the personal lives of citizens. That is why, to avoid polarisation and imbalance of power, AI knowledge and expertise should be spread around the world and throughout the masses. If this does not happen, there is a danger that underdeveloped countries might become dependent upon the US, UK and China for AI services. During the first industrial revolution, countries had competed to control resources like oil and coal. In the present situation, the resources are AI professionals, cloud computing, AI chips and others. Between the 1970s and the 1980s, the great combination of microprocessors, and electronics created a great trend of innovation that gave birth to the internet and the web community and also the GPS. And after that the 2000s the advent of AI and the application of AI in policy and wars, promise excitement in the short term and the long term. Technology And War Industrial revolutions and technological progress have changed the nature of war and informational conflict. The Center for a New American Security wrote in a post, “The First Industrial Revolution enabled Napoleon’s levée en masse or mass mobilisation of the population for war. This shifted military power away from small, very professionalised militaries, such as those of Prussian leader Frederick the Great, and toward countries able to mobilise their population on a large scale.” World War I started the implementation of machinery in warfare and reshaped warfare as well as the war in the last century. Modern technologies like trucks, aeroplanes and radio communication were deployed for the first time in human history. A Chatham House report suggested a “broad framework to define and distinguish between the types of roles that AI might play in policymaking and international affairs: these roles are identified as analytical, predictive and operational.” In an analytical role, AI will automate large parts of foreign policymaking. Experts also say that AI will augment our abilities rather transform the whole ecosystem. The predictive uses of AI, will have an effect on a longer term and let policymakers understand the impact of their policies. The report suggests that operational uses of AI are very far off and they relate to fully autonomous systems and weaponry. Nation Power And AI CNAS report suggested that there are certain elements that relate to national power during the AI revolution. They predict some of the main points to be considered are: Owning Large Quantities Of The Right Type Of Data: Most algorithms need large amounts of data to function perfectly and owning huge quantities of data will be very important. Creating An AI-Capable Talent Pool: Human empowered with AI skills will be the most important resource of the AI revolution. AI-Ready Organisations: Organisations and institutes should be ready to transition into the AI era. Public-Private Cooperation: Like China, more countries should working with the private entities to leverage their innovation and talent Conclusion AI can be leveraged to totally change and transform international relations and foreign policy but there will be some requirements that will be hard to acquire. David Gosset, who is the director of the Academia Sinica Europaea at CEIBS and founder of the Euro-China Forum says, “AI, more than any other technology, will impact the future of mankind, it has to be wisely approached on a quest for human dignity and not blindly worshiped as the new Master of a diminished humanity, it has to be a catalyst for more global solidarity and not a tyrannical matrix of new political or geopolitical divisions.”","excerpt":"International security and foreign affairs seem to be the newest hotbeds for artificial intelligence innovation and applications. The multidimensionality of today’s machine learning and AI is already seen to have a profound impact on how countries manage their foreign affairs — one of the most recent cases being China. Analytics India Magazine had reported earlier […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI Chips","China","International Affairs"],"author_name":"Abhijeet Katte","publish_date":"2018-08-16T12:24:37","publication_year":"2018","word_count":891,"keywords":["Go","artificial intelligence","machine learning","AI","AI (Artificial Intelligence)","cloud computing","RAG","AI Chips","ViT","analytics","GAN","International Affairs","R","China"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","RAG","cloud computing","R","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/artificial-intelligence-is-a-key-to-future-international-relations-dynamics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":22035,"title":"Meet India’s Only Entrant to IBM Watson X Prize Challenge That Is Developing A Solution To Screen &#038; Segment Oral Cancer","content":"Atom360 team Bangalore-based startup Atom360 is the youngest on the block in the AI space and the only Indian entrant that progressed to second round of the famed IBM Watson X prize that kick-started in 2016. The $5 million IBM Watson AI XPrize competition, which kicked off in 2016 and will end in 2020, is the interestingly the first of the X-prize contests to feature a contestant-defined “open” goal rather than a predetermined objective. Atom360 made it to the second round of the competition and was one of the 59 teams, representing 14 countries. While the top 10 teams were recognized at the Annual Conference on Neural Information Processing Systems in December 2017 – Atom360 that submitted an entry related to the healthcare domain also received a superb response. Talking to Analytics India Magazine about the AI-focused healthcare solution, founder and CEO Reuben Fernandes said, “We feel great to be representing our country in this global AI competition. As part of the competition, we are working on AI-assisted surgical robot. We have broken this problem down and are trying to get market validation for each algorithm. The application ranges from tumor segmentation in medical data to reducing oral cancer deaths in India using deep learning algorithms”. The X-Prize challenge led to the inception of this startup This Bangalore-based six-member startup is currently working in a stealth mode and would be launching two products this year. The startup was incepted at the time when IBM Watson X Prize Challenge was announced, after their healthcare entry was selected, the team started working full-time in 2017. Atom360 has a very young team, the founder, Fernandes is a 2012 SRM University, Chennai graduate who had always been interested in robotics and deepened his Deep Learning base with online programs. Meanwhile, the average age of the team is around 24 years with everyone interested in making an impact in the world using AI, revealed Fernandes. Atom360’s AI solution can reduce deaths occurring due to oral cancer The impact of AI on healthcare has been huge – a 2016 Frost & Sullivan research forecasted that the market for AI in healthcare is projected to reach $6.6 billion by 2021. From healthcare bots to virtual assistants and AI-powered diagnostic tools, AI has the potential to improve medical outcome by 30-40 percent and reduce cost of treatment by 50 per cent. When AIM quizzed Fernandes on focusing on the healthcare domain, he said, “We believe in human and AI working together to improve healthcare. We work on augmented intelligence making solutions for doctors therefore improving efficiency and reducing time.” The startup aims to provide a AI-powered solution for oral cancer, one of the top three types of cancers in India. A recent study indicated that India also has among the highest and most number of avoidable deaths due to cancer. In India, tobacco-related diseases cost an astounding INR 100 billion to India’s healthcare, a news report indicates. “The impact we want to create is reduce the deaths due to oral cancer by at-least half before 2020. Getting our doctors to be augmented by AI will help bring quality care to all Indians at low costs,” shared Fernandes. Tech Stack For Atom360, the application ranges from tumor segmentation in medical data to reducing Oral Cancer deaths in India using deep learning algorithms. Also, the final solution is the surgical robot. “We are selectively getting into domains such as robotics, tumor segmentation and computer vision so that can validate our algorithms that we will eventually use in our surgical robot,” he shared. The startup is working with algorithms ranging from reinforcement learning to deep convolution neural net. “We use Deep learning for Image classification and segmentation to detecting tumors in medical images,” he added. Are the solutions scalable? “Well, part of the judgement criteria during the X Prize challenge was measuring the impact we will be making in the world. Our solutions are highly scalable and we will be launching two of our products the beginning of the next quarter,” he shared. Implementing the AI Solution The solution is targeted at radiologists who can use to segment tumour faster and more efficiently. Besides, Fernandes added that the app would enable any person or volunteer to detect oral cancer through the app for periodic self-diagnosis. When quizzed about time to market, he said, “We will be deploying systems for cancer segmentation and oral cancer screening this year. Our clients are Doctors we are starting with radiologists as our clients who can leverage the sensing capability of the surgical robot,” he said. The startup’s augmented AI solution would help doctors screen and segment the tumour effectively and also reduce the time take up by the segmentation procedure significantly. How The X Prize Challenge Gave Global Visibility The competition that led to the founding of the startup gave the young team global visibility and also globally connecting them with international mentors and talented people from India who want to be a part of their exciting journey. We also have support in terms of cloud computing credits to help accelerate fast, shared Fernandes. When prodded about his hiring criteria, he shared, “We hire young talented people who we think have the potential to learn fast. We have leaders from the AI field as our mentors who guide us to accomplish our goals”. The founder’s advice to Deep Learning enthusiasts who want to shift towards this field is to learn from the courses available on the Internet, there are new ones coming every month and one can go through courses and research papers to get a different perspective.","excerpt":"Bangalore-based startup Atom360 is the youngest on the block in the AI space and the only Indian entrant that progressed to second round of the famed IBM Watson X prize that kick-started in 2016. The $5 million IBM Watson AI XPrize competition, which kicked off in 2016 and will end in 2020, is the interestingly […]","categories":["Deep Tech"],"tags":["healthcare artificial intelligence","International Affairs"],"author_name":"Richa Bhatia","publish_date":"2018-02-26T04:38:53","publication_year":"2018","word_count":933,"keywords":["Go","AI","cloud computing","virtual assistants","computer vision","RAG","Aim","deep learning","analytics","healthcare artificial intelligence","International Affairs","R"],"extracted_tech_keywords":["AI","deep learning","computer vision","analytics","Aim","RAG","virtual assistants","cloud computing","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/meet-indias-only-entrant-to-ibm-watson-x-prize-challenge-that-is-developing-a-solution-to-screen-segment-oral-cancer\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10112147,"title":"Microsoft to Train 2 Million People With AI Skills In India","content":"Satya Nadella, currently on a visit to India, announced in his keynote that Microsoft is set to provide AI skills training to 2 million people by 2025. “We believe that, obviously, at the end of the day, ensuring that the workforce has the skills to thrive in this new age is the most important thing any of us can be doing and we are happy to play our role in it, ” said Nadella. Further Nadella said that it’s not just about the skills but the new jobs which get created along with generative AI. “AI tasks involve data labeling, and bringing those jobs to rural India, providing well-paying opportunities and creating economic growth. That is pretty unique,” said Nadella. He went on to praise Karya, one of the world’s first data cooperatives that offers labeling and annotation services. “You even see that in rural India people are participating in the AI economy. Thanks to the work of organisations like Karya,” he said. Besides Karya, he also lauded Bhashini and Agami. Agami has recently introduced a new product called Jugalbandi Studio. “Jugalbandi Studio is a no-code local tool for people to build bots and make them available to all social entrepreneurs, nonprofits, and others who are making a positive impact,” said Nadella. Further Nadella said India has more AI engineers, making the AI engineering community second only to the United States. On February 6, 2024, Nadella marked his 10th anniversary as the Chief Executive Officer of Microsoft. Presently, he is in India for February 7 and February 8 to engage with AI startups.","excerpt":"“Cloud adoption in India was a lot slower than what happened in the rest of the world,” said Microsoft’s Satya Nadella, envisioning to accelerate AI adoption in the country.","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2024-02-07T11:46:42","publication_year":"2024","word_count":263,"keywords":["programming_languages:R","AI","generative AI","GAN","R","startup"],"extracted_tech_keywords":["AI","generative AI","R","GAN","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-to-train-2-million-people-with-ai-skills-in-india\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164551,"title":"Transformer Co-Author Niki Parmar Joins Anthropic After Founding Two AI Startups","content":"Niki Parmar, a former Google AI researcher and co-author of the groundbreaking “Attention Is All You Need” paper, has joined Anthropic. Parmar announced her move on X, stating, “Today is as good a day as any to share that I joined Anthropic last Dec :) Claude 3.7 is a remarkable model at complex tasks, especially coding, and I’m thrilled to have contributed to its development. From winning Pokémon badges to vibes coding, Claude’s got you covered!” Parmar joined Google Research in 2015 as part of Google Brain, where she played a key role in developing the Transformer architecture—a foundation for modern AI models, including ChatGPT. She left Google in 2021 to co-found Adept AI Labs, a startup focused on general intelligence. Later, she co-founded Essential AI alongside Ashish Vaswani. Emerging from stealth in December 2023 with backing from Google, NVIDIA, and AMD, Essential AI raised nearly $65 million to develop large language model (LLM)-powered tools for automating business workflows and improving productivity. Parmar’s journey in AI began at the Pune Institute of Computer Technology in India. Despite not securing admission to the Indian Institute of Technology (IIT), she pursued her passion by taking online courses from AI pioneers Andrew Ng and Peter Norvig. She later earned a Master’s degree in Computer Science from the University of Southern California. Meanwhile, Anthropic has released Claude 3.7 Sonnet, its latest AI model, and Claude Code, an agentic coding tool available in a limited research preview. The company, in its blog post, mentioned that Claude 3.7 Sonnet is “the first hybrid reasoning model on the market” and allows users to choose between near-instant responses and extended, step-by-step reasoning. Claude 3.7 Sonnet is available across all Claude plans, including Free, Pro, Team, and Enterprise, and through Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI. Extended thinking mode is not included in the free tier. The pricing remains unchanged from previous models at $3 per million input tokens and $15 per million output tokens, which includes thinking tokens. Anthropic describes Claude 3.7 Sonnet as “both an ordinary LLM and a reasoning model in one.” Users can decide when the model should generate a quick response or engage in a deeper reasoning process.","excerpt":"Parmar joined Google Research in 2015 as part of Google Brain, where she played a key role in developing the Transformer architecture—a foundation for modern AI models, including ChatGPT.","categories":["AI News"],"tags":["Anthropic"],"author_name":"Siddharth Jindal","publish_date":"2025-02-25T15:10:32","publication_year":"2025","word_count":367,"keywords":["Anthropic","ChatGPT","Go","TPU","API","AI","GPT","GAN","transformer architecture","R"],"extracted_tech_keywords":["AI","ChatGPT","Anthropic","TPU","R","Go","API","transformer architecture","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/transformer-co-author-niki-parmar-joins-anthropic-after-founding-two-ai-startups\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10083602,"title":"What Happened with Quantum Computing","content":"According to Tracxn, there are 41 quantum computing startups, and according to CrunchBase, there are a total of 238 startups operating for quantum computing. The number remains nuanced. The global market of quantum computing is around $35.5 billion. In February, the Indian government announced its plans of investing $1 billion for the next five years towards the development of quantum technology. Quantum technology still remains in a nascent stage globally. Only a handful of big tech companies and a few research institutions in the US, China, and Europe are able to make developments in the sector as the technology requires expertise and high-computing capabilities, not available for everyone. Let’s look at some of the developments in the quantum computing and technology space that took place in 2022. IBM Qubit and z16 IBM is developing a 4000+ qubit quantum computer slated to be completed by 2025. In December, IBM unveiled the world’s first quantum computer with 1,000 qubit, Condor, which is set to debut in 2023. Big Blue is making great strides in the quantum computing field after launching its modular quantum processor, called Heron. In April, IBM released industry’s first quantum-safe system, IBM z16 with an integrated on-chip AI accelerator for delivering low-latency inference for real-time transactions and making history in the quantum security arena. The system leverages IBM’s AI inferencing Telum Processor for high-volume processing. NVIDIA QODA In July, NVIDIA, in a bid to replicate the success of its computing platform CUDA, announced the release of its unified computing platform QODA (Quantum Optimised Device Architecture) for accelerating research in quantum computing across areas like AI, HPC, finance, and health. QODA enables developers to add quantum computing capabilities to their existing applications and thus aims to make the field more accessible by creating a coherent hybrid quantum-classical programming model. The team said that HPC and AI experts can leverage these quantum processors using their DGX systems to simulate the future of quantum computing. India-Finland In March, India and Finland laid out a detailed plan for setting up an Indo-Finnish Virtual Network Centre for quantum computing and identified three institutes – IIT Madras, IISER Pune, and C-DAC Pune for the project. FIN-Q n (Finland India Quantum Network) is built for creating a sandbox environment for both the countries’ companies for quantum technology. TechMahindra has been trying to solidify India’s quantum computing field and now they have signed an MoU with Finnish quantum computing company IQM for advancing the field. The partnership also includes collaborations with Mahindra University for research in quantum computing and explainable AI. Quantinuum Quantinuum announced InQuanto 2.0 for computational chemistry using quantum computing. The new version of InQuanto introduces new tools for greater efficiency, advanced algorithms for speeding up vector calculations and integral operator classes. The tool is now also user-friendly and has improved resource cost estimation on H-series quantum computers. IN-SPACe and QNu Labs Indian National Space Promotion and Authorisation Centre (IN-SPACe) signed an MoU with Bangalore-based QNu Labs for creating domestic satellite QKD (quantum key distribution) products. This will be achieved through the startups increasing use of quantum cryptography for addressing cybersecurity challenges with the classical computing world. The partnership will provide QNu Labs with payload designs with the help of ISRO as well. TCS, Infosys, and AWS In November, TCS made its quantum computing lab available on AWS for enterprise customers for quantum computing and applications. The research and development will be powered by Amazon’s Braket, a fully-managed service offered by AWS for quantum computing. TCS has been collaborating with the government for academic, research, and startups for developing quantum technology. Infosys made big bets and launched Quantum Living Labs for their customers who want to apply quantum computing in their applications like manufacturing, cyber security, healthcare, etc. Twist Programming Language In January, scientists at MIT’s Computer Science and Artificial Intelligence (CSAIL) developed a programming language for quantum computing called Twist. The language relies on the concept of purity for building intuitive programs by enforcing the absence of entanglement, thus resulting in fewer bugs. When programming quantum computers, two qubits are entangled, which results in actions taken on one qubit to affect the other one, resulting in weakness and incorrectness in the program. Twist enables developers to write quantum programs explicitly when a qubit is not entangled with another. Google’s Quantum Virtual Machine In July, Google made another of their products publicly available. Quantum Virtual Machine (QVM) is a tool for prototyping, testing, and optimising quantum circuits with processor-like output for near-term quantum hardware that can now be deployed from Colab notebook. Users can emulate two of its processors – Rainbow and Weber. Weber is a Sycamore processor used in Google’s beyond-classical experiments, published in Nature in 2019. Rainbow is used by the company’s experiments demonstrating variational quantum eigensolver with quantum chemistry problems. SpinQ Triangulum China’s SpinQ published their paper in February about Triangulum, a second-generation, three-qubit desktop quantum computer. In November, they released three portable quantum computers – Gemini, Gemini Mini, and Triangulum – that would be used for educational purposes. These computers use nuclear magnetic resonance (NMR) for performing quantum computations using motion of spins of atoms.","excerpt":"Let’s look at some of the developments in the nascent quantum computing space","categories":["AI Features"],"tags":["Quantum Computing","quantum technology"],"author_name":"Mohit Pandey","publish_date":"2022-12-28T17:00:00","publication_year":"2022","word_count":854,"keywords":["Quantum Computing","CUDA","Go","artificial intelligence","TPU","AWS","AI","RAG","Colab","Aim","R","quantum technology"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","Colab","RAG","AWS","TPU","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-happened-with-quantum-computing-in-2022\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":22459,"title":"Top 5 Women-Only-Hackathons And Tech Events In India","content":"On International Women’s Day, it is worth noting the grim picture of IT tech sector – the representation of women in tech is dismally low. There is just one woman engineer as against three men engineers. The overall representation of women in the engineering workforce of IT firms is just 34 percent. And out of 5 million developers in India, only 18 percent are women. To eradicate the gender gap, women coders or programmers need to be encouraged to continue with their chosen career. And to increase the demand for top talent in the industry, new women programmers need to be brought into the system. Deepa Madhavan, director, Enterprise Data Services at PayPal told Analytics India Magazine, “We cannot dispute the gender imbalance in the technology field. Even though women choose the field of technology as a career option, the attrition rate at the mid-management and senior levels are high due to both professional and personal pressures. This attrition can lead to disinterest among women towards a career in the field of technology.” Big companies like Google, Adobe, Accenture, Walmart, among others are clamouring for better representation of women in the IT sector. Thanks, to their valuable campaign the number of women is growing in IT sector and they are coming out of the shadows to participate during the hackathons and other tech events. Here we take a lowdown at the five interesting women-only hackathons and tech events in which coding enthusiasts can take part: 1. Wonder Coders Walmart Labs is launching a hackfest exclusively for women, who can create and innovate through technology. Starting from April 19 to 20, women will build products under following themes: health and fitness, wearables (IOT), Travel, social, reinventing retail. So far 1,453 participants have registered for the hackathon. As it is a physical hackathon it will be conducted in Walmart Labs office premises in Bengaluru. Eligibility Both individuals and teams can compete, but they must identify as a woman. No male participants are allowed. No coding experience needed. Maximum number of participants in one team is 2 people How To Apply If you are a female and you love coding, visit the website to register now. Deadline is March 21. You do not have to pay anyone to register yourself for the Hackathon. Prize Top three teams to be awarded Rs 1,00,000\/- each 2. International Women’s Hackathon 2018 HackerEarth and Schlumberger are conducting an online Hackathon targeted at women developers across the globe, with an aim to encourage them to take up programming. The hackathon is live now, it started from February 20 and will go on till March 12. Women developers will build products under any of the following four themes – women’s health and safety, economic freedom, social impact, innovation. Eligibility Both individuals and teams can participate A maximum number of contestants in one team is 3 people You need not have a coding experience. How to Apply Interested hackathon enthusiasts may visit the website to apply. Prize Prizes include $2500 for the winning team, $1500 for the runners-up and $1000 for the second runners-up. 3. Anita’s Moonshot Codeathon The main goal of this online codeathon is to bring women from interdisciplinary backgrounds to produce a creative solution to a socio-cultural or socio-economic problem. Simply, a diverse team is needed to solve a problem in their community with the online hackathon. Registration is now closed and the winner will be announced on March 30. Eligibility Both individuals and teams can compete, but they must identify as a woman Maximum number of women contestants in one team is 5 people No coding experience necessary Prize Grand prize winners get $10,500 for a team of 5 members Full scholarships to Grace Hopper Celebration 2018 4. Broadridge CODE-BEE 2.0 Broadridge is conducting an online hackathon challenge targeted at women. It will be 48 hours non-stop coding marathon. Before you take up the online challenge, you should participate in programming practice challenge. The sample challenge will enable you to understand how to participate in programming challenges on HackerEarth. It will give you a detailed information on how programming questions should be attempted. The sample challenge will also give you the details about the HackerEarth judge Eligibility Only women participants will be considered. No male contestant allowed. How to Apply The programming challenge will begin on March 23 and will end on March 25. Click on the website to register immediately. You will receive a reminder email 3 hours before the challenge begins only if you registered for the challenge. To participate in this programming challenge, follow the steps: Register for the challenge Visit the challenge page on Click participate in the challenge Before you attempt programming questions, you should choose a language from the list. All inputs for the programming problems are from STDIN and output to STDOUT. Prize Top 3 winners will get up to Rs 100000. 5. Women Who Code Women Who Code is the largest and most active community of engineers dedicated to inspiring women to excel in technology careers. The events are intended to inspire women to excel in technology careers. The event is scheduled for March 23. Who Should Join? The community is for professional women in technology careers, including software engineers, developers, UX\/UI designers, data scientists, among others. Aspiring coders can also participate. What to Expect? The events will offer free hands-on technical events, study groups, panel discussions, lightning talks and keynotes featuring influential tech industry experts, innovators and investors. The idea is to build the skills the participants need to raise their professional profile and achieve career success. How to Join Visit the website, if you are interested to attend the event.","excerpt":"On International Women’s Day, it is worth noting the grim picture of IT tech sector – the representation of women in tech is dismally low. There is just one woman engineer as against three men engineers. The overall representation of women in the engineering workforce of IT firms is just 34 percent. And out of […]","categories":["AI Trends"],"tags":["Accenture","Adobe","Google","Hackathons","Hackathons India","international women's day","Walmart"],"author_name":"Smita Sinha","publish_date":"2018-03-08T11:14:37","publication_year":"2018","word_count":944,"keywords":["Accenture","Go","TPU","programming_languages:R","Adobe","AI","R","innovation","Hackathons","programming_languages:Go","RAG","international women's day","Aim","analytics","Google","Walmart","Hackathons India"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","TPU","R","Go","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-women-only-hackathons-and-tech-events-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10167399,"title":"Vaccine Discovery for AIDS and Cancer Is No Longer a Distant Dream","content":"The quest to develop vaccines for AIDS and cancer, two of the most complex and deadly diseases, has long tested the limits of medical science. But today, technology can help to solve these complex diseases, particularly with quantum computing and Artificial Intelligence (AI). A study by University of Toronto scientists and Insilico Medicine, published in Nature Biotechnology, demonstrated how integrating quantum computing, generative AI and classical computing techniques enabled researchers to design molecules targeting KRAS ( mutated oncogenes). This cancer-driving protein was long deemed “undruggable.” One of the biggest bottlenecks in traditional vaccine development is accurately predicting how a protein will fold or bind in a complex biological environment. However, quantum computing allows researchers to simulate these scenarios and simultaneously process with greater precision and speed, potentially fast-tracking the identification of viable vaccine candidates. According to the UK’s National Quantum Computing Centre (NQCC) 2024–25 healthcare insights report, quantum computing is not merely a scientific curiosity, but is becoming a necessity. Meanwhile, the country aims to integrate quantum technology across the NHS by 2030, acknowledging the complexities of regulatory compliance and clinical adoption. Unlike classical systems that process information using binary digits (bits), quantum computers use quantum bits or qubits. While traditional supercomputers check one solution at a time, quantum computers can explore many possibilities at once using qubits. This makes them much faster for tasks like simulating how a drug interacts with a disease-causing protein, saving weeks or even months in research time. Generative AI complements these capabilities by proposing novel drug molecules or protein structures, while machine learning models predict how these interact with the human body. This synergy helps develop treatments more efficiently. “Generative AI and quantum computing expedite drug discovery by predicting interactions with protein targets and optimising molecules for reactions like protein folding,” Madhur Singhal, managing partner for pharma and lifesciences at Praxis Global Alliance, told AIM. Additionally, concepts such as reverse vaccinology use AI to analyse pathogen genomes for likely protective antigens since cancer represents a collection of diseases with distinct mutations. Also, HIV has a high mutation rate that enables it to frequently evade immune detection. Ayush Singh, practice member at Praxis, a research and consulting firm, said, “AI models trained on massive datasets can help uncover tumour-specific antigens for cancer, or identify conserved regions in HIV for vaccine targeting.” Quantum computing also has the potential to significantly impact two underserved areas of medicine–rare diseases and women’s health. The NQCC report emphasises this. “The shift brought about by deep tech extends far beyond vaccines. AI and quantum computing are transforming every stage of pharmaceutical development. From innovative manufacturing and automated packaging to climate-controlled logistics via IoT and blockchain,” adds Singhal. In a study published in Scientific Reports, researchers introduced a quantum hybrid classical convolutional neural network (QCCNN) to improve breast cancer diagnosis. By combining quantum computing with classical machine learning, the QCCNN model enhances accuracy and speed in medical image analysis. Still, hurdles remain. Quantum computing hardware is still maturing, and the so-called “quantum advantage”, where quantum systems consistently outperform classical computers, hasn’t yet been fully realised across all healthcare applications. Meanwhile, AI raises ethical concerns, data privacy, regulatory frameworks for AI predictions and challenges in model interpretability (black box problem).","excerpt":"With AI and quantum computing powering new waves in healthcare, vaccine discovery for some of the world’s deadliest diseases is closer than ever.","categories":["AI Features"],"tags":["AI","AIDS","cancer","Quantum Computing"],"author_name":"Merin Susan John","publish_date":"2025-04-07T18:18:26","publication_year":"2025","word_count":537,"keywords":["Quantum Computing","artificial intelligence","machine learning","model interpretability","AI","neural network","AIDS","Git","Aim","generative AI","cancer","CNN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","generative AI","Aim","R","Git","CNN","model interpretability"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/vaccine-discovery-for-aids-and-cancer-is-no-longer-a-distant-dream\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10087488,"title":"Will ‘Made in India’ iPhones Outshine China’s?","content":"When it comes to manufacturing quality products or timely deliveries of orders, India still has its struggles. According to recent reports, a 50% rejection rate was observed for iPhone casings made in India. It is alleged that the phones manufactured domestically did not meet Apple’s quality criteria and ended up as waste or recycled material. With the increasing shift towards manufacturing in India, does this statistic conclusively determine the fate of manufacturing for the world’s largest phone company or are we scrutinising the manufacturing process through the wrong lens? India started assembling the older generation iPhones in 2017 and it was only towards the last quarter of 2022 that they started manufacturing new models, i.e., iPhone 14 in India. Until then, China had been, and still is, the largest manufacturer of Apple products. In 2022, over 90% of Apple products, such as iPhones, iPads and Macbooks, were made in China. China’s Dominance China started manufacturing iPhones in the middle of 2007. Steve Jobs moved the majority of manufacturing to China because the cost of labour was the lowest compared to other countries. In addition, China was then the only country in the world with a ready-made supply-chain network with production lines that could be scaled up in no time. Further, Tim Cook has mentioned that the reason China remains the manufacturing hub is not solely due to the availability of cheap labour, as it is widely believed, but also because of skill availability, quantity and location. To work on a profitable model, Apple does not own any factories in China and only works with manufacturing companies on a contractual basis. “Sweatshops” China’s dominance as a manufacturing hub evidently stemmed from cheap labour. With time, technological innovations helped them advance their skills, but the common tag of running “sweatshops” was perennially criticised. Foxconn is Apple’s biggest iPhone maker, manufacturing as much as 70% of iPhone shipments globally. But, the plant has been laden with labour problems and mass protests have erupted over time. Apple was called out for indirectly employing workers who operated in unfavourable conditions. The workers were said to have been paid as little as £1.12 an hour, with several of them working for nearly 24 hours in a single shift. Foxconn has also had multiple incidents involving labour suicides and accidents. With such recurrent issues, Apple’s reliance on China for manufacturing capabilities became unsustainable over time. Shift from China Depending on a single country for manufacturing a large chunk of products posed its own risks from an economic and political perspective for the company. Two global incidents stirred the shift in Apple’s manufacturing dependency on other Asian countries, primary of which was Covid-19. Strict pandemic rules and continued lockdowns hampered manufacturing efforts and it was estimated that Apple lost $8 billion in revenue due to lockdowns in 2022. The tumultuous trade relationship between US and China, which occurred during Trump’s administration, also posed a threat to the future dealings between the countries. Furthermore, China’s stance in the Ukraine crisis put the relationship between the two countries under further duress. Apple has reportedly reduced production sites in China from 47% in 2019 to 36% in 2021. India Shines In September 2022, India started manufacturing iPhone 14 models. Analysts from JP Morgan claim that Apple would move 5% of global iPhone 14 production to India and, by 2025, India will become a global hub for manufacturing 25% of all iPhones. Vietnam is said to manufacture 20% of Apple iPad and watches, 5% of MacBooks and 65% of AirPods by 2025. By inviting government proposals that offer tax incentives to boost local production, Apple has increased its manufacturing footprint in India. For instance, manufacturing of iPhone 14 started in September 2022 and three of Apple’s global suppliers—Foxconn and Pegatron in Tamil Nadu and Wistron in Karnataka—are currently operating domestically. Incentives of $550 million were offered to bring Apple and other device manufacturers to India. Roadblocks in India’s path Although Apple’s suppliers have been in India for some time now, in 2020, Karnataka’s Wistron plant in Narasapura, Kolar, was vandalised by their employees. It was alleged that delays in payments and extended working hours led to protests by the workers, where company property and phones were destroyed. Wistron had incurred losses of over INR 437 crore due to the resultant vandalism and theft. In addition, with every manufacturing curve, the ramp-up period takes considerable time and depending on unit production capacity and the number of workers, the period can go up. China has been manufacturing iPhones since 2007—which gives it a ten-year jump start when compared to India— but has lost contracts due to its highly unstable government bodies and efficiency stance. India, a fairly new competitor in the iPhone manufacturing space, still has a long way to go and the current statistics might not be enough to seal its dream of becoming the next global manufacturing hub.","excerpt":"Analysts from JP Morgan said that Apple would move 5% of global iPhone 14 production to India and, by 2025, India will become a global hub for manufacturing 25% of all iPhones.","categories":["AI Features"],"tags":["Apple","China","foxconn","iPhone","karnataka","tim cook"],"author_name":"Vandana Nair","publish_date":"2023-02-16T12:56:38","publication_year":"2023","word_count":816,"keywords":["karnataka","Go","programming_languages:R","AI","Apple","innovation","programming_languages:Go","GAN","Aim","China","ViT","tim cook","iPhone","R","foxconn"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/will-made-in-india-iphones-outshine-chinas\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094956,"title":"Geoffrey Hinton Raises Concerns Over Profit-Driven AI Development, Urges Caution","content":"Geoffrey Hinton and Andrew Ng, the two pioneers of AI, came together for an interesting discussion on AI threats and risks. Hinton recently left Google to discuss AI threats. He supported the likes of Elon Musk, Steve Wozniak, Yoshua Bengio, Gary Marcus and several other AI experts who signed an open letter for the pausing of AI development beyond GPT-4. However, Ng is not in favour of the “pause”. While Hinton now expresses concern over AI dangers, he previously ignored ethical concerns raised by Google’s own team. He compares the potential risks of AI to the creation of the atomic bomb during World War II, emphasising the dangers of profit-driven AI development that could result in AI-generated content surpassing human-produced content and jeopardising our survival. Read more: Tech-giants, self-regulation, and free speech Had an insightful conversation with @geoffreyhinton about AI and catastrophic risks. Two thoughts we want to share: (i) It's important that AI scientists reach consensus on risks-similar to climate scientists, who have rough consensus on climate change-to shape good policy.… pic.twitter.com\/TXT9wgv2TR— Andrew Ng (@AndrewYNg) June 11, 2023 Hinton Calls for Unity as AI Research Faces Diverse Opinion However, during an insightful conversation with another pioneer of AI, Andrew Ng, both discussed that AI researchers need to reach a consensus similar to climate scientists on climate change whereby it proved that there has been a substantial increase in the Earth’s temperature since the latter half of the 19th century. The main reason behind this is attributed to human actions, predominantly the release of greenhouse gases into the atmosphere. “If there are diverse opinions among AI researchers, it becomes easier for others to cherry-pick opinions that suit their agendas,” added Hinton. He continued to say that there is a significant diversity of opinions and even conflicting factions. It would be great to move past this phase and reach a point where researchers agree on the main threats posed by AI or at least agree on some of the major threats and their urgency and danger. This is because policymakers and decision-makers will seek technical opinions from researchers. Read more: Big-tech Regulation: India, Drop the Dubiety & Go the EU Way The Urgent Need for Consensus Another important point discussed is the need for researchers to urgently reach a consensus on whether LLM chatbots like ChatGPT or Bard truly understand what they are saying or are statistical constructs is an important point discussed. While some believe they understand, others disagree. Resolving this issue is crucial for achieving consensus on AI-related matters. The challenge in assessing understanding lies in identifying the appropriate tests for determining its presence in a system. Large language models and AI models appear to be constructing a world model, suggesting some level of understanding. However, this is a personal viewpoint. If the research community engages in further discussions about this interface and develops a shared understanding, it can promote more consistent reasoning and improved alignment within the AI community regarding the risks associated with AI. An aspect of this discussion relates to statistics, as we all agree that statistics play a crucial role. However, some people who consider it to be solely statistics tend to think in terms of programming or counting co-occurrence frequencies of words. “We believe that the process of creating features or embeddings and the interactions between these features goes beyond mere statistics; it involves understanding,” he added. By predicting the next symbol based on complex interactions between features, we can make predictions about the probability of the next words. I personally believe that this process represents understanding, akin to what our brains do. However, this is a topic that needs to be discussed within the research community to convince others that these systems are not just statistical constructs and that a shared understanding can be developed to address the risks associated with AI. “Gaining a better understanding of what AI systems comprehend will likely bring the research community closer to reaching similar conclusions as a community,” he concluded. Read more: India Backs Off on AI Regulation. But Why?","excerpt":"Geoffrey Hinton and Andrew Ng, the two pioneers of AI, came together for an interesting discussion on AI threats and risks.","categories":["AI News"],"tags":[],"author_name":"Shritama Saha","publish_date":"2023-06-12T17:50:40","publication_year":"2023","word_count":674,"keywords":["Go","ChatGPT","AI","chatbots","RPA","GPT","AI research","R","AI-generated content","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","chatbots","R","Go","GPT","RPA","AI-generated content","AI research","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/geoffrey-hinton-raises-concerns-over-profit-driven-ai-development-urges-caution\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":7608,"title":"Flytxt Secures $11 Million (INR 70 Crore approx.) Investment  for Next Phase of Growth","content":"India – 30 June, 2015 – Flytxt, a fast growing Big Data Analytics solution provider, today announced that it has closed an $11 million (INR 70 Crore approx.) funding round to finance its next phase of growth. The investment was led by Sunrise Capital GmbH and Big Data Investments B.V. with participation from Flytxt’s existing investors. Launched in 2008, Flytxt has previously received funding from a group of angels and family offices in Germany and The Netherlands. The new investment will support Flytxt’s plans to consolidate its presence in the Communication Service Provider (CSP) market and to offer mobile consumer analytics solutions to other industry verticals. The company plans to double its team through the next phase of growth. Following the latest round of funding, Dr. Ms. Brigitte Mohn (Sunrise Capital GmbH) and Frits Baron van Dedem (Big Data Investments B.V.) will join the Flytxt Board of Directors. Dr. Mohn is an Executive Board Member of Bertelsmann Stiftung and Supervisory Board Member of Bertelsmann AG. Frits Baron van Dedem is Managing Director of UNION TANK Eckstein GmbH & Co. KG and an experienced private equity investor. Dr. Brigitte Mohn and Frits van Dedem hold several other industry and non-profit positions. “We are delighted to welcome Dr.Brigitte Mohn and Frits Van Dedem on our Board. Their rich business leadership experience will be a significant asset for Flytxt in the next phase of our growth”, said, Joerg Swoboda, Chairman, Flytxt. “The company already has a unique standing in Big Data Analytics market with a sustainable business model, proven technology, impressive customer base and an exceptional team. And the potential is huge.” he added. Flytxt’s proprietary Big Data Analytics solutions have enabled many leading global Communication Service Providers (CSPs) and brands to increase sales and revenues, optimize margins and enhance customer experience. Flytxt now aims to grow its product portfolio for CSPs to generate higher economic value, and leverage its proven technology to expand to other industry verticals. About Flytxt Flytxt is a fast growing Big Data Analytics solution provider for Communication Service Providers (CSPs) and Mobile Enterprises across the globe. The company offers full suite of internal and external monetization solutions for increasing revenue, optimizing margins and enhancing customer experience. Flytxt offers solutions with full service delivery model combining Technology, Consulting and Execution to deliver guaranteed economic impact to its customers. The company has deployed its platforms with more than 50 customers across 32 countries, analysing data of more than 500 million mobile consumers, delivering 2 to 7% economic impact consistently. Flytxt has its headquarters in the Netherlands, corporate office in Dubai and also presence at Mumbai, Trivandrum, London, Singapore, Lagos, Nairobi, Dhaka and Mexico City.","excerpt":"India – 30 June, 2015 – Flytxt, a fast growing Big Data Analytics solution provider, today announced that it has closed an $11 million (INR 70 Crore approx.) funding round to finance its next phase of growth. The investment was led by Sunrise Capital GmbH and Big Data Investments B.V. with participation from Flytxt’s existing investors. […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2015-06-30T15:18:55","publication_year":"2015","word_count":444,"keywords":["big data","Go","API","AI","Git","RAG","BERT","Aim","analytics","R"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Git","API","big data","BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/flytxt-secures-11-million-inr-70-crore-approx-investment-for-next-phase-of-growth\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058409,"title":"Researchers leverage AI to develop world’s fastest DNA sequencing technique","content":"Researchers from Stanford University, NVIDIA, Oxford Nanopore Technologies, Google, Baylor College of Medicine and the University of California at Santa Cruz have developed a method to do DNA sequencing in 5 hours and 2 minutes – and entered the Guinness World Record for fastest DNA sequencing technique. The research team, led by Stanford University, used AI to expedite the end-to-end process, from collecting a blood sample to sequencing the whole genome and identifying variants linked to diseases. The researchers made the diagnosis for a three-month-old infant suffering from a rare seizure-causing genetic disorder in a few hours. The traditional gene panel analysis takes as long as two weeks to return results. By optimising the diagnosis pipeline at 7-10 hours, clinicians can quickly identify genetic clues to inform patient care plans. In this pilot project, the genomes were sequenced for 12 patients – most of them children–at Stanford Health Care and Lucile Packard Children’s Hospital Stanford. The researchers optimised the pipeline including speeding up sample preparation and using nanopore sequencing on Oxford Nanopore’s PromethION Flow Cells to generate over 100 gigabases of data per hour. The data was then sent to NVIDIA Tensor Core GPUs in a Google Cloud computing environment for base calling. At this stage, raw signals from the device are turned into a string of A, T, G and C nucleotides, and alignment in near real-time. Since it distributed the data across cloud GPU, it instantly helped minimise latency. The next step was to find tiny variations in the DNA sequence that can cause a genetic disorder. This stage was sped up with Clara Parabricks using a GPU-accelerated version of PEPPER-Margin-DeepVariant, a pipeline developed in a collaboration between UC Santa Cruz’s Computational Genomics Laboratory and Google. For highly accurate variant calling, DeepVariant uses convolutional neural networks. The GPU-accelerated DeepVariant Germline Pipeline software in Clara Parabricks provides results at then times the speed of native DeepVariant instances, decreasing the time to identify disease-causing variants. The details of the ultra-rapid sequencing method is published in the New England Journal of Medicine.","excerpt":"The traditional gene panel analysis takes as long as two weeks to return results.","categories":["AI News"],"tags":["Neural Networks"],"author_name":"Meeta Ramnani","publish_date":"2022-01-14T15:57:55","publication_year":"2022","word_count":341,"keywords":["Go","API","programming_languages:R","cloud computing","neural network","AI","ML","programming_languages:Go","cloud_platforms:Google Cloud","R","Neural Networks"],"extracted_tech_keywords":["AI","ML","neural network","cloud computing","R","Go","API","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/researchers-leverage-ai-to-develop-worlds-fastest-dna-sequencing-technique\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":35633,"title":"7 Tips That Can Help Land The First Data Science Internship","content":"With Data Science being an attractive field, students look for internship opportunities. An internship for college students provides them a big advantage to bolster their CV strong and land a job in data science. College students or newcomers in the field often lack guidance and need tips to bag their first data science internship. We list some tips for the beginners in data science to grab their first ever data science internship. 1.Know the company: Before applying, make sure that you study all about what the company does and what its goals are. Research a little to know about the company, its products and services and make sure that the job they are offering is the job that interests you. Applying for an internship without knowing the organisation well could be very negative when asked about the company during the interview. 2.Know your ML fundamentals and don’t stuff up resume with keywords: At the end of the day, it is a machine learning internship and recruiters are looking for some basic level of expertise in machine learning. You are of course not expected to be an expert with all the deep technical understanding. But before you go for the interview, make sure you have mastered at least one of the basic algorithms and are at least to a good level awareness of the introductory machine learning algorithms like linear regression, K means, SVMs, random forest. 3.Active GitHub account: A GitHub account definitely enhances the chances of landing an internship in data science or machine learning. The account must be active and have useful content. It is a clear proof to the people in the interview panel that you have knowledge and leaves behind a great deal of a good impression. 4.Data science blogs: Probably not as impactful as a GitHub account or a project done by you, but an own technical blog on data science depicting your knowledge will add to your positives giving the organisation an impression that you are truly interested in data science and. Although this is not a necessity, it is sure a plus. 5.Do ML and DS projects: Machine learning courses that you had taken up online and academics that you are involved in are the things that all the candidates applying for the interview will have. But what not everyone has is a practical machine learning project. Recruiters are looking for not just good coding and machine learning skills, but also problem-solving skills. Projects show that you have applied the theoretical knowledge that you have learnt into practical applications, which is in turn a proof of you being able to handle real-time organisation issues and goals. If you cannot come up with a project yourself, do it under a professor under your college or participate in online hackathons are online competitions. It is also important to showcase how you function in a team. Working in an organisation requires collaborating with different people and it is important to show that you can work in a group and offer value to the team. 7.Don’t overcrowd your CV: Don’t fill your resume with all the popular programming languages that you know and don’t claim that you know all those popular machine learning algorithms. Specially since you are joining an internship as a fresher, there is no way that you could’ve mastered every programming language that there is. Proactive and eager-to-learn attitude is what will get applicants noticed, even if one might not have an extensive portfolio or industry experience. Familiarity with every aspect, or an extensive experience in the industry is not necessary, but it is important for recruiters know that you are up for taking up a research and work towards the goal. Ms Veronica Puah, Deputy Director of Talent Networking at SGInnovate said in a panel discussion, “It’s OK if you don’t know or are not too familiar with certain things. But at the end of the day, we want someone who takes the initiative to do their own research to close the gap.” Make sure that the project that you are presenting, if you are, has everything that you know so that you can answer any level of technical questions restricted to that area and slay the interview. Be ready to answer questions related to anything that is mentioned in your CV because interviewers pay special attention to that. Revise all the theory related to the projects and courses that you had done, before the interview day.","excerpt":"With Data Science being an attractive field, students look for internship opportunities. An internship for college students provides them a big advantage to bolster their CV strong and land a job in data science. College students or newcomers in the field often lack guidance and need tips to bag their first data science internship. We […]","categories":["AI Trends"],"tags":["Data Science","Machine Learning","resume","Virtual Internship Program"],"author_name":"Disha Misal","publish_date":"2019-03-01T12:06:44","publication_year":"2019","word_count":743,"keywords":["data science","Go","machine learning","resume","AI","ML","Machine Learning","Git","Virtual Internship Program","Aim","GAN","Data Science","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Aim","R","Go","Git","GitHub","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-tips-that-can-help-land-the-first-data-science-internship\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":60675,"title":"How These Data Science Enthusiasts Solved MachineHack’s Patient Drug-Switch Prediction Hackathon","content":"MachineHack is not just a platform for data science enthusiasts to hone and practice their skills. It can also help organizations mine the right talent in an ever-expanding domain. Data science is one of the most demanding skills in the IT industry today. With the hype in AI and machine learning, the pool of enthusiasts is expanding exponentially, making it extremely hard for organizations to find a perfect fit for their job requirement. MachineHack – backed by its parent company Analytics India Magazine – has been continuously helping the machine learning and data science community grow to its peak by conducting exciting hackathons and challenging aspirants. Recently, we collaborated with an IT firm to help them find the best talents in the domain. The Patient Drug-Switch Prediction Hackathon was both thrilling and challenging for the community. Despite being one of our shortest challenges running for just 13 days, the hackathon proved to be a great success, with over 700 registrations and over 60 active participants. Out of the 66 participants, Tirthankar Das, Nikhil Kumar Mishra, and Amey More ended up as the top competitors and presented their work at MLDS 2020. Analytics India Magazine introduces you to the winners and their approaches to the solution. Tirthankar Das Tirthankar started his data science career in the Banking Financial Services and Insurance (BFSI) domain, when data science was still in its infancy. Currently working in the aviation industry, he has more than five years of experience working with data. Tirthankar has always been inclined towards statistics, and data science has helped him pursue his love for the subject. He believes that a good data scientist should be domain-agnostic. Approach To Solving The Problem Tirthankar explains his approach as follows: The objective of the Patient Drug Switch Hackathon was to identify the patient population who are likely to switch any product in the RA (Rheumatoid Arthritis therapeutic) Market. It had three layers to evaluate the solution. The first round was evaluating the best solution based on AUC. The second round was based on ‘Time and Memory Complexity’. The second step is especially relevant in the practical scenario from a deployment perspective. Most hackathons ignore this point. Other than ‘MacBook’, this second step made this hackathon interesting. And the final round was presented at MLDS in a room full of data science enthusiasts. During the hackathon, we were given transaction data of the patient for both train and test. We had to create features from that transaction data of drug purchasing. It had six columns: Patient_id, Time, Event, Specialty, Plan_Type and Payment. The event, Specialty and Plan_type are different sublevels to represent the drug. I always try to follow the traditional approach of the model building, which has three building blocks. Feature Generation There were three broad types of features, that is, Recency, Frequency and NormChange. Recency is defined as how recently did an Event\/Plan_Type\/Specialty happen before the anchor date; Frequency means how many times did an Event\/Plan_Type\/Specialty occur in a specific time frame, and NormChange means whether the frequency of an Event\/Plan_Type\/Specialty increased or decreased in a recent time frame (not more than 1.5 years) as compared to the previous time frame. Around 40k variables were generated with the logic mentioned above. As I mentioned earlier, this hackathon is not only about the accuracy of the model. It is also about tackling time and memory complexity. Creating 40k features sequentially will take a lot of time if we run it on the machine with RAM of 16GB with four cores (as per given specification). To reduce the time consumption, I introduced a parallel processing concept for calculating features of Frequency and NormChange. In contrast, I used the default Apply function of python for calculating Recency features. Feature Selection Once all the features were created, I did two preprocessing steps before feeding these into the LightGBM model. Imputing the missing value with 9999999 and removing degenerate variables. It helped in reducing the number of features. After this step, remaining features were fed into the LightGBM model. I did not do much tuning in this step as it was just a feature selection step. I used feature importance to select the final set of features. The logic of subsetting the feature was very straightforward. It involved selecting all the features whose importance is greater than zero. Finally, I had 7k features for the final model. Recency features were found most important compared to NormChange and Frequency. Model building LightGBM has always been my favourite algorithm whenever memory and time are constrained. In the final model, I used stratified k fold cross-validation, where the value of K is 30. The final prediction came from the average of those models. Tuning of hyperparameters, such as learning rate, feature fraction and num_leaves were important, which impacted the result. “MachineHack is a great platform where data science enthusiasts can experiment with different algorithms and look at their performances. I have been an active participant of MachineHack hackathons from the last seven to eight months. I have learned many ML techniques from my fellow participants like Chetan Ambi, Rajat Rajan, and Saurabh Kumar. MachineHack’s initiative to share winners’ solutions with proper documentation helps many data scientists like myself” – Tirthankar shared his MachineHack experience. Get the complete code here. Nikhil Kumar Mishra Nikhil is currently a final year Computer Science Engineering student at Pesit South Campus, Bangalore. He started his data science journey during his second year after being inspired by a youtube video on self-driving cars. The technology intrigued him, and he was driven into the world of ML. He started with Andrew NG’s famous course and applied his knowledge in the hackathons which he participated in. Kaggle’s Microsoft Malware Prediction hackathon – in which he finished 25th – was a turning point in his data science journey, which gave him the confidence to take it further and challenge himself with more hackathons on other platforms like MachineHack and Analytics Vidhya. Approach To Solving The Problem Nikhil explains his approach as follows: “This was a very challenging competition. Feature engineering was the key component,just like any other competition, along with a wide selection of models”- Nikhil spoke about the problem. Feature Engineering At first, it seemed simple aggregations like mean on the numerical data and size, and unique categorical data for each patient id would help. But the AUC was stuck to well below 80, while other competitors could easily get above 80. Creating very deep features like this landed me a max AUC of 77. Then I realized I needed to use the features which were required to be created in the feature generation part. Recency was the feature which had magic in it, especially events like event_439 and event_449 gave the model a magic boost to easily reach above 85. Time-based Aggregations Recency based features gave a lot of ideas into much more creative feature engineering. Recency was the latest when something occurred, such as an event, a patient going for some plan_type and some speciality. So why not create other time-based features like: 1. The first time something happened (Oldest_Time) 2. The number of unique times something happened Note: All features were being calculated for each event, and not all of the events at once, calculating on all events or plan_type or speciality, which will cause a lot of loss in information. A helpful way to think the would-be occurrence of event_1 could lead to the patient switching drugs while the occurrence of event_2 would make him use the same drug. So all the events, plan_types and specialities should be treated separately. Verifying which were important among these could be easily done with the help of EDA. Payment Based Aggregations One can calculate the mean, max and total payment for each of the events, plan_types and specialities separately. Time and Money Recency and other time-based features were being calculated on categorical columns, and payment was numerical, so I discretized the payment into bins, and calculated time-based features on that. This helped me a lot and I reached 88 AUC. Model Selection and Ensembling I used an ensemble of 2 XGBoost and 2 LightGBM models, LightGBM was used with a lot less number of features, because it was not memory efficient and RAM usage was spiking easily above the 16 GB memory limit on Kaggle. 180 days trick Normchange features hinted that a patient’s behaviour could change over time, so I created two models(LightGBM and Xgboost) using the data from only past six months or 180 days, which again gave me a significant boost in the ensemble. 180 days was by no means a strict threshold, and I encourage people to try data for only the past three months or one year or any other time interval. The final model consisted of the ensemble of the four models, which led to my highest LB score. “Understanding the problem is very critical in ML competitions. Also, no model rules all kinds of problems, so I suggest all MachineHack fellow competitors first build a baseline for different models and then proceed appropriately.”- he said. “MachineHack is an amazing platform, especially for beginners. MachineHack team is very helpful in understanding and interacting with participants to get doubts resolved. Also, the community is ever-growing, with new and brilliant participants coming up in every competition. I intend to continue using MachineHack to practice and refresh my knowledge on data science,” says Nikhil about his experience with MachineHack. Get the complete code here. Amey More Amey is an engineering graduate with close to one year of experience in the industry. He was introduced to data science two years back during his college days with Andrew Ng’s Machine learning course. He soon realized that the best way to enter this expanding domain was through hackathons. He also praises the vast and lively global data science community for its excellent support in sharing knowledge. Approach To Solving The Problem Amey explains his approach as follows: Objective 1: Feature Creation I decided to use built-in pandas functions like group by, stack, unstack, describe etc. rather than going the iterative way. I created the recency features with correct values, and observed that the frequency feature values were not entirely accurate. Objective 2: Modelling I realized that features mentioned in the problem were extremely useful, as only using features I engineered did not give a good score. Used recency and frequency features significantly increased the dimensionality of the dataset. Also, most of these features consisted mainly of null values. Additional Features Aggregated ‘patient_payment’ (min, max, mean, sum) features over ‘event_name’ & ‘specialty’ for each patient, ‘event_name’ frequency for each patient, ‘patient_payment’ aggregated for each patient and count of total events for each patient The dataset provided had a class imbalance of 85%(negative class) and 15%(positive class); a similar proportion was expected in the test set. The final model was a bagged run of LightGBM across five-folds created using Stratified K-Fold cross-validation. Using this cross-validation strategy allowed the model to be trained on data that followed the same class distribution as train and test sets. The above approach leads to more robust and better predictions. The idea was to get a more generalized sense of error. Parameters used in the model script were already tuned on the training set. Final test data is scored using these five models (built on five-folds). To convert the predictions into hard classes, the simple logic of metric maximization was used. I took all the out-of-fold predictions and iterated over a range of thresholds, took the argmax, i.e. whichever threshold gave the maximum AUC score was chosen as the threshold value. “MachineHack is a great platform for budding data scientists to hone their skills, as well as test them against each other. The problems here are real-life, giving us an idea about the kind of work being done in industry” – he shared his opinion on the platform. Get the complete code here.","excerpt":"MachineHack is not just a platform for data science enthusiasts to hone and practice their skills. It can also help organizations mine the right talent in an ever-expanding domain. Data science is one of the most demanding skills in the IT industry today. With the hype in AI and machine learning, the pool of enthusiasts […]","categories":["Deep Tech"],"tags":[],"author_name":"Amal Nair","publish_date":"2020-04-01T19:00:00","publication_year":"2020","word_count":1989,"keywords":["data science","machine learning","AI","ML","RAG","Python","XGBoost","analytics","LightGBM","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","XGBoost","LightGBM","Pandas","RAG","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/patient-drug-switch-prediction-winners\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":9877,"title":"Leadership Interview: Onno Pistorius","content":"We interviewed Onno Pistorius, the founder and director of ClearPredictions.com, a young start-up from 2015 as well Morphis, a company he started in 2005.  We got to know that Onno Pistorius is not only an enthusiastic entrepreneur but an experienced leader in the analytics field and a self-taught developer. Telling us about himself, he told us, “I did my studies in mechanical engineering and have been fascinated by the possibilities and power of computer software and hence self-taught myself to be a developer to design products that provide a user friendly and easy way of doing difficult processes. As a director, I have been in the lead of the development of various products like workflow management, business rule engines and call center software for Morphis. Using Morphis’ software, employees can easily and effectively handle customer contacts and underlying workflow processes. Morphis was nominated for years in a row for the Deloitte Technology Fast50 and ranked number 38 on the Deloitte Technology Fast 500 EMEA 2010, a ranking of the 500 fastest growing technology companies in EMEA. Morphis was nominated for the category Emerging Entrepreneur in the “Ernst & Young Entrepreneur Of The Year 2010” and received the “FD Gazellen Award” as the fastest growing ICT Company in the region.” AIM: How did you start your career in it? How has the journey been so far? OP: I started my career designing industrial robots and innovative, custom designed packaging machines as a mechanical construction engineer. After years of work in that area, I went on to develop ProcessRunner, which is the software suite of Morphis. I love to make complex things as simple as possible. The journey has been fabulous so far. Since, a couple of years, I have been fascinated with the thought of using predictive capabilities on data – but in a new way. It should no longer be needed to hire expensive teams of data scientists to make use of predictive analytics. Managers and business owners define their strategy mostly looking backwards, using reports, to analyse questions like “what happened?” in their businesses. Predictive analytics can help to provides answers on “what will happen?”. The aim of Clear Predictions here is to provide an easy to use predictive analytics platform for all kinds of businesses and for all kinds of users. AIM: How do you define Predictive Analytics and how is it related to machine learning, big data and data mining? OP: “Predictive analytics is the use of data, statistical algorithms and machine-learning techniques to identify the likelihood of future outcomes based on historical data. Bear in mind that, no statistical algorithm can predict the future with 100% certainty; it is based on probabilities. On the other hand, Machine learning is a method of data analysis that automates predictive model building. It automatically learns to make accurate predictions based on past observations. Big Data is a broad term for data sets so large or complex that traditional data processing applications are inadequate (High Volume, Velocity and Variety). By the way, a CRM system is not big data. Data-mining is carried out by a person, on a particular data set, to get a better understanding of the data.” AIM: What factors, according to you, have fueled the rise of Predictive Analytics in recent years? OP: According to IBM, 80% of all the data in the world has been created over the last 2 years alone! This is an amazing trend. The growing volumes of data makes it more interesting to produce valuable information out of it. Besides this trend in data growth, the tougher economic conditions and a need for competitive differentiation are reasons many companies look into predictive analytics now. The goal is to go beyond descriptive statistics (reporting) on “what has happened” to providing best assessment on what will happen in future. The result is improved decision making and predictions that lead to more effective actions. Analyzing historic data to predict future events enables decisions that can give companies this competitive advantage. These decisions depend on analyzing at a speed, volume, and complexity that is often too great for humans. This would require easy-to-use software and faster and cheaper computers of course. AIM: Would you like to share any projects that have you been working on this year? OP: “We have started many projects, one of them is at the utility company Qurrent in Amsterdam. As Mr Slieker, CEO of this company, states: “The ClearPredictions platform is providing us beautiful actionable predictions. It tells us which customer is sensitive to our up-sell campaigns, but also which customer is about to churn. My employees operate the tooling themselves. There is no need any more to hire expensive data scientists nor have long running projects to make use of big data technology.” We are doing projects for the marketing directors of various retailer companies. Another very interesting pilot we recently started at a leading Dutch Hospital, is about using ClearPredictions’ machine learning technology for an early stage recognition of patients having a rare disease. I am very excited to use our platform in order to make the decision making process in medical care more efficient. AIM: Why\/how are they interesting for you? OP: Those projects are tremendously interesting for us because we get a lot of high value feedback from them. We honor all thoughtful requests for product enhancements by putting them on our product development roadmap. The high value requests will always to be implemented within a couple of weeks, using our agile development methodology. AIM: Which industries are placing their bets (i.e. investing heavily) in predictive analytics? OP: There is a saying amongst marketing managers: “50 percent of my marketing budget was spend well, I only don’t know which 50 percent”. I find this amusing in an era where smart predictive marketing software has become available! The telecom industry is active in predicting churn (customers that tend to leave) for years now. But many other verticals, like assurance, banking, utilities, travel, aviation maintenance and retail, are just getting aware of the possibilities and benefits of the predictive analytics technology. In order to obtain differentiation towards their competitors, companies are now getting up to speed in realizing the importance of data analytics and have started to investigate. AIM: What are the biggest areas of opportunity in predictive analytics? OP: Big data is a megatrend that touches so many aspects of our interactions in life – from the Internet of Things (think about all the data generated by your smart phone, like the GPS positioning feature) and content analytics (so Facebook can present the most interesting updates especially for you) to customer satisfaction (think about recommendations on flipkart.com: customers who bought this, also bought that). What really is going to make predictive analytics go mainstream is the ability to connect not just with data scientists and technologists but with business people. And absolutely one of the key features to that is making use of this technology plain easy. De-mystify the concept of machine learning! AIM: What challenges do you see for young entrepreneurs starting out their own firms in Analytics? Any words of wisdom for them? OP: Entrepreneurs in the area of predictive analytics face the fact that they operate on the front wave of technology. One of the mistakes young entrepreneurs make is to release their solution too late. Reid Hoffman, founder LinkedIn, has made a nice statement on this topic: “If you are not embarrassed by the first version of your product, you’ve launched too late.” So: in the early phase of your company: go out there! Talk to people, visit companies and be an extrovert from the beginning. This will get you all the inputs you need from talking to customers and enable you to design a product that is really needed and that will be your mantra to success. AIM: What are your top predictions for predictive analytics in 2016-2017? OP: “Market size of predictive analytics software in 2016-17 alone is over 5 billion dollars (quoting Forrester). This opportunity origins from the needs of operational managers and C-level leaders who have allocated a significant amount of their budget for big data activities in the coming years. The last 10 years, the paradigm shift was to bring paper documents, HR dossiers and off-line project administration from the office desks into the cloud. The next 10 years, the big shift in mindset will be to find -and value- the trends in the tremendous amount of data we create. Using predictive analytics, hidden trends and correlations will be made visible and valued.”","excerpt":"We interviewed Onno Pistorius, the founder and director of ClearPredictions.com, a young start-up from 2015 as well Morphis, a company he started in 2005.  We got to know that Onno Pistorius is not only an enthusiastic entrepreneur but an experienced leader in the analytics field and a self-taught developer. Telling us about himself, he told us, “I did my […]","categories":["AI Features"],"tags":["Interviews and Discussions","leadership"],"author_name":"Apoorva Verma","publish_date":"2016-05-10T06:42:53","publication_year":"2016","word_count":1424,"keywords":["big data","Go","machine learning","T5","AI","R","Aim","ViT","analytics","predictive analytics","leadership","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","predictive analytics","R","Go","big data","T5","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/leadership-interview-onno-pistorius\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10058919,"title":"New AI model uncovers risk genes for motor neuron disease","content":"Researchers from the University of Sheffield and the Stanford University School of Medicine have designed a machine learning tool to identify risk genes for motor neurone disease. The tool named RefMap, has already been utilised to discover 690 risk genes for motor neurone disease. Dr Johnathan Cooper-Knock, from the University of Sheffield’s Neuroscience Institute, said: “This new tool will help us to understand and profile the genetic basis of MND. Using this model we have already seen a dramatic increase in the number of risk genes for MND, from approximately 15 to 690. Each new risk gene discovered is a potential target for the development of new treatments for MND and could also pave the way for genetic testing for families to work out their risk of disease.” The 690 new genes identified by RefMap lead to a five-fold increase in discovered heritability, a measure which describes how much of the disease is due to a variation in genetic factors. “RefMap identifies risk genes by integrating genetic and epigenetic data. It is a generic tool and we are applying it to more diseases in the lab,” Sai Zhang, PhD, instructor of genetics at the Stanford University School of Medicine said. Michael Snyder, PhD, professor and chair of the department of genetics at the  Stanford School of Medicine and also the corresponding author of this work added: “By doing machine learning for genome analysis, we are discovering more hidden genes for human complex diseases such as MND, which will eventually power personalised treatment and intervention.”","excerpt":"Each new risk gene discovered is a potential target for the development of new treatments for MND.","categories":["AI News"],"tags":["ML models","stanford university"],"author_name":"Meeta Ramnani","publish_date":"2022-01-21T18:31:48","publication_year":"2022","word_count":254,"keywords":["machine learning","programming_languages:R","AI","ML models","stanford university","R"],"extracted_tech_keywords":["AI","machine learning","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/new-ai-model-uncovers-risk-genes-for-motor-neuron-disease\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":25393,"title":"What Happened To Self-Driving Cars? Why Are They Not On The Roads Yet?","content":"The road to using fully-automated vehicles is long and complicated. Every year car companies like Nissan, Toyota and Mercedes, among others, flock to auto shows to show off their self-driving concepts cars. But not a single car company has had a successful demonstration at Level 5 on the five-step scale of autonomous driving. Level 5 is where the self-driving car is fully autonomous and doesn’t require any human intervention and there is no steering wheel or pedals. Where Are Cars That Tesla Promised Back in 2015? Three years ago, Tesla CEO Elon Musk predicted that autonomous driving would be ready in two years. He had clarified that it was dependent on regulatory approval and software validation. Recently, Musk disclosed on Twitter that the company plans to roll out autopilot software, Version 9, which will enable full self-driving features. That issue is better in latest Autopilot software rolling out now & fully fixed in August update as part of our long-awaited Tesla Version 9. To date, Autopilot resources have rightly focused entirely on safety. With V9, we will begin to enable full self-driving features. — Elon Musk (@elonmusk) June 10, 2018 However, here is a catch: Musk said that Tesla will activate a subset of self-driving car features, but did not promise that their vehicles would be fully autonomous. The company already sells vehicles with full autonomous features, but it is currently disabled because Tesla’s existing autopilot technology has been involved in few wrecks lately. A few months ago, autopilot-enabled Model X crashed into a barrier in Mountain View, California. Reports suggest that Tesla’s autopilot itself was at fault. Another report said that in 2016, Tesla’s autopilot-enabled Model S was responsible for a man’s life as the AI mistook a white lorry for a clear sky. The driver could not correct it in time, and it resulted in an accident. After these incidents, Tesla has refused to rely on LIDAR sensors for autonomous driving, instead, they are using ultrasonic sensors, low-profile cameras and radar. But there is still no clear deadline from the company or the founder for when they will be able to fulfil the promise to bring the vehicles that will require “no intervention from the person in the driver’s seat.” Danger: Authorised Autonomous Vehicles Only Three months ago, a self-driving Uber car killed a pedestrian in Arizona. One of Waymo’s self-driving vehicle was also involved in a collision in Arizona. Reportedly, five of these crashes are already under investigation by the National Transportation Safety Board in the US. These incidents are a reminder that self-driving cars are still at an experimental stage and have a long way to go before they truly become safe for drivers, passengers and the general public. The governments are also trying to figure out how to regulate them. Countries like US, Germany, UK and Netherland have regulations in place for driverless cars on public roads and have issued autonomous testing permits. Similarly, few Asian countries have also been enforcing similar legislation over the last three years. Two months ago, China issued road test regulations for driverless and semi-autonomous vehicles. Singapore regulated the operation of AVs on public roads. Challenges In The Journey The decade-old autonomous driving technology has started to experience unpredictable situations which (human) drivers have always been facing These autonomous cars are to be geofenced to specific areas for at least five to 10 years more because of the safety issue. It is also because the technology is still at a nascent stage to allow them to travel unrestricted. Most testings of the autonomous vehicles are being done with a safety driver in the front seat who will be able to take over if something goes wrong. However, it can become challenging for the drivers to take control of a fast-moving car. Autonomous vehicles need a huge amount of data before they can confidently prowl the streets. Just like smartphones and app, these cars have code that will also need an update. For example, if an autonomous car which was released five years ago, wants to work today, it needs an update. Scenario In India Self-driving cars are a distant dream for Indians. As of now, the Indian government is more focused on bringing electric vehicles on road, rather than autonomous ones. Moreover, none of the State governments has clear laws for such vehicles. Even the current BJP-led NDA  government is not keen on bringing them to India over fears that it could take away jobs. Apart from lack of support from the government, there’s been one major hurdle to self-driving vehicles in India – a legion of lawbreakers, from rash drivers and jaywalkers to cattle. And there are very few companies in India that are testing autonomous vehicles in India. Former Uber CEO Travis Kalanick had also joked about the same, saying, “India will be the last one to get autonomous cars! Have you see the way people drive here?”","excerpt":"The road to using fully-automated vehicles is long and complicated. Every year car companies like Nissan, Toyota and Mercedes, among others, flock to auto shows to show off their self-driving concepts cars. But not a single car company has had a successful demonstration at Level 5 on the five-step scale of autonomous driving. Level 5 […]","categories":["IT Services"],"tags":["Autonomous Vehicles","driverless cars","Self Driving Cars","Tesla","Waymo"],"author_name":"Smita Sinha","publish_date":"2018-06-13T09:57:29","publication_year":"2018","word_count":820,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","Self Driving Cars","programming_languages:Go","ai_applications:autonomous driving","driverless cars","Waymo","Autonomous Vehicles","Tesla","R"],"extracted_tech_keywords":["AI","AWS","R","Go","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving"],"url":"https:\/\/analyticsindiamag.com\/it-services\/what-happened-to-self-driving-cars-why-are-they-not-on-the-roads-yet\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10011890,"title":"AIM Data Science Education Ranking 2020 | Top UG Programmes In India","content":"Check our Latest 2021 rankings With a rise in the popularity of the data science domain over the last few years, candidates are looking to make a career in the field right after their schooling. This has resulted in an increase in the number of undergraduate and graduate programmes in data science. This year, in our yearly data science education ranking, we have introduced the ranking of UG programmes to help candidates make the right decisions at the start of their careers. The UG courses listed here have been ranked on several parameters such as pedagogy, faculty profile and credentials, faculty to student ratio, gender diversity, graduation outcome, external relations & placement assistance, ROI & brand value and student review, the detailed methodology of which have been shared here. The information on the courses from various institutions was collected by circulating survey form, which was analysed and scored according to the methodology mentioned below. The detailed scoring also involved internal discussions with the participating institutions to run factual checks on the information shared, and ensure that there is no missing data. You can find the detailed methodology followed for this ranking here. You can find the data received by the institutes across various parameters here. 1| Bachelor of Data Science By S.P Jain School of Global Management Duration of programme (in months): 36Total no. of learning sessions (in hours): 3780Course fees: ₹ 22,00,000Industry partnership, affiliation or collaboration: AWS, Oracle, IBM, Here Technologies, SAS Institute, IEEE, Tableau, ACS, IIT Bombay, University of Massachusetts, ISI Kolkata, Instagram and several others This bachelor of data science programme by SP Jain School of Global Management has the best pedagogy and course structure. It is also the top scorer in the parameter of faculty profile and credentials and has a high faculty to student ratio. The programme Pedagogy includes 3780 learning hours, which is the highest in this list and moreover for the total duration being 3 years. It also fares very well in graduation outcome and has one of the best scores for student review. 2| B.Sc. (Applied Statistics & Analytics) By NMIMS (Deemed to be University) Duration of programme (in months): 36Total no. of learning sessions (in hours): 2865Course fees: ₹ 3,37,500Industry partnership, affiliation or collaboration: As a part of the course, industry experts, especially from marketing and finance domain, to share their work and research papers One of the oldest colleges in the list, this bachelor’s programme by NMIMS offers the best pedagogy and course structure among all the other colleges listed here. It also has one of the best gender diversity rates with the ratio of female to male learners of 57:43. With one of the best ROI and brand values among other colleges listed here, it also scores good in external relations and placement assistance. 3| B.E. in Artificial Intelligence (AI) By Vijaybhoomi University Duration of programme (in months): 48Total no. of learning sessions (in hours): 1800Course fees: ₹ 12,80,000Industry partnership, affiliation or collaboration: INSOFE is the teaching partner. In addition, the University of Texas, USA and Rennes School of Business, France are other partners With a good pedagogical background, it is one of the lead scorers in parameters such as faculty profile and credentials and faculty to student ratio. It has the best graduation outcome and scores the highest in this parameter among all the other courses listed here. Being a new programme, it lags in parameters such as placement assistance, ROI and brand value. 4| Bachelors of Science in Data Science By Vijaybhoomi University Duration of programme (in months): 24Total no. of learning sessions (in hours): 1200Course fees: ₹ 9,60,000Industry partnership, affiliation or collaboration: INSOFE is the teaching partner. In addition, the University of Texas, USA and Rennes School of Business, France are other partners Just like its course B.E in AI, this bachelors programme has a good pedagogical background, along with good scores in parameters such as faculty profile and faculty to student ratio. It also has the best graduation outcome, scoring full marks. As it is a new programme, it lags in parameters such as placement assistance, ROI and brand value. 5| B.Tech CSE with specialisation in Data Science By The NorthCap University Duration of programme (in months): 48Total no. of learning sessions (in hours): 3052Course fees: ₹ 9,84,000 This course by The NorthCap University has one of the best pedagogy scores with over 1200 live classroom sessions. It also has one of the best graduation outcomes with the programme completion rate of 96-99%. It scores second-best in terms of external relations and placement assistance, with the average placement rate of all batches of the programme till date to be 99%. It however, lags in faculty to student ratio and gender diversity. 6| Bachelor of Data Science By CHRIST (Deemed to be University), Pune Lavasa Campus Duration of programme (in months): 36Total no. of learning sessions (in hours): 2916Course fees: ₹ 4,50,000Industry partnership, affiliation or collaboration: Centre of Excellence for Hands-on Training and Tie-Up for Research The Bachelor of Data Science programme offered at Pune, Lavasa Campus of CHRIST University has one of the best pedagogies with 2716 live sessions and 200 online recorded sessions. It also has one of the best graduation outcomes, and consists of faculty with a good profile and credentials. With one of the best faculty to student ratio, it also displays decent ROI and brand value. 7| B.Tech CSE (Big Data & Analytics) By Ganpat University Duration of programme (in months): 48Total no. of learning sessions (in hours): 2717Course fees: ₹ 1,42,000Industry partnership, affiliation or collaboration: Affiliation with IBM, Redhat, Cisco and AWS With a good graduation outcome and pedagogy, it scores the highest in the parameter ROI and brand value.  In terms of pedagogy, it only has live videos with no online recorded videos. It also has a good placement record with the average placement rate of all batches of the programme to date being 95-98%. However, the programme lags in parameters such as gender diversity and needs improvement in its faculty profile and credentials. 8| Bachelors of Science in Data Science By NSHM, Knowledge Campus Duration of programme (in months): 36Total no. of learning sessions (in hours): 1550Course fees: ₹ 2,85,000Industry partnership, affiliation or collaboration: Subex, CDAC, National Cyber Safety & Security Standards, Analytics Society of India, Indian Navy, Analytics India Magazine, Amazon Educational Services, TCSion, Association of Data Scientists It secured full scores in pedagogy, faculty to student ratio and graduation outcome. The average pass percentage of all the batches for the programme to date is 100%. The number of live sessions during the programme amounts to 1550, with no online recorded sessions. With a decent score in faculty profile and credentials, it lags in ROI, brand value, gender diversity and placement assistance. 9| Bachelors of Science in Data Science (Honours) By Sri Sri University Duration of programme (in months): 36Total no. of learning sessions (in hours): 3380Course fees: ₹ 2,70,000 This course by Sri Sri University has one of the best pedagogy. It scores decent in graduation outcome, with an average graduation rate of 91-95% from the previous batch. It also has a good faculty to student ratio and a decent faculty profile and credentials. However, it lags in providing placement assistance, student reviews and gender diversity. 10| B.Tech (Computer Science) with Data Analytics By MUIT Noida Duration of programme (in months): 48Total no. of learning sessions (in hours): 1600Course fees: ₹ 5,00,000Industry partnership, affiliation or collaboration: upGrad While it scores lower in terms of pedagogy compared to other courses listed here, it shows comparatively better scores in faculty profile and credentials. It lags in parameters such as faculty to student ratio, gender diversity, placement assistance and student reviews. It has decent scores in ROI and brand value.","excerpt":"Check our Latest 2021 rankings With a rise in the popularity of the data science domain over the last few years, candidates are looking to make a career in the field right after their schooling. This has resulted in an increase in the number of undergraduate and graduate programmes in data science. This year, in […]","categories":["AI Features"],"tags":["big data roi","mba in data analytics"],"author_name":"Srishti Deoras","publish_date":"2020-11-20T11:00:30","publication_year":"2020","word_count":1291,"keywords":["big data","data science","Go","artificial intelligence","mba in data analytics","AWS","AI","RAG","big data roi","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","RAG","AWS","R","Go","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/aim-data-science-education-ranking-2020-top-ug-programmes-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10169238,"title":"Now, Agentic AI-Powered TCS MasterCraft to Speed Up Legacy Modernisation","content":"Tata Consultancy Services (TCS) has unveiled an enhanced version of its flagship product, TCS MasterCraft, now powered by generative AI (GenAI) and agentic AI. The upgraded solution promises to reduce the cost of legacy system modernisation by over 70% and deliver results twice as fast as traditional methods. Designed to automate the process of updating legacy applications, the latest TCS MasterCraft cuts down manual effort and enables faster, more accurate mining of business logic. Ashvini Saxena, VP and head, TCS components engineering group and digital software and solutions, said, “We’ve successfully delivered hundreds of complex modernisation programs using intelligent automation-driven technology conversion for years. We have leveraged GenAI and agentic AI capabilities for the extraction of business knowledge and conversion to augment this capability to deliver maintainable applications and data with a powerful design repository.” The AI-augmented TCS MasterCraft has already delivered tangible outcomes for major clients. A large North American bank used the platform to achieve 2x productivity gains and 3x faster delivery during mainframe modernisation. Another ongoing project involves migrating over 50 million lines of legacy COBOL code to Java for a global financial services firm, significantly accelerating modernisation while building a reusable knowledge base. With GenAI and agentic AI, TCS MasterCraft now features an intelligent automation engine enhanced by TCS’ proprietary knowledge base and industry best practices. Agentic AI introduces goal-oriented agents capable of tackling unique modernisation challenges while ensuring the resulting systems are resilient and ready for future technological shifts. Since its inception in 2012, TCS MasterCraft has helped clients transform legacy systems into modern, scalable, and cloud-native architectures. The new version builds on this legacy, integrating human-in-the-loop decision-making and a DevSecOps pipeline to continuously deliver modernised solutions.Meanwhile, TCS reported approximately 580 AI-centric business engagements in Q4FY25 and has developed over 150 specialised agentic AI solutions across sectors, including financial services, supply chain, and accounting.","excerpt":"Designed to automate the process of updating legacy applications, the latest TCS MasterCraft cuts down manual effort and enables faster, more accurate mining of business logic.","categories":["AI News"],"tags":[],"author_name":"Shalini Mondal","publish_date":"2025-05-06T18:32:42","publication_year":"2025","word_count":309,"keywords":["Go","GenAI","agentic AI","AI","Scala","Git","RAG","generative AI","R","Java"],"extracted_tech_keywords":["AI","generative AI","GenAI","agentic AI","RAG","R","Go","Java","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/now-agentic-ai-powered-tcs-mastercraft-to-speed-up-legacy-modernisation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10117621,"title":"The Rise of Mixture of Experts LLMs","content":"In the past week, we saw several ‘Mixture of Experts’ models coming in, like Databricks DBRX, AI21 Labs’ Jamba, xAI’s Grok-1, and Alibaba’s Qwen 1.5, with Mixtral 8X 7B already in the mix, making MoE popular. Welcome to MOE's, which model would you like to use today from our new and updated menu? XXL @DbrxMosaicAI XL @MistralAI Medium @AI21Labs Small @Alibaba_Qwen cc @code_star @JustinLin610 @tombengal_ pic.twitter.com\/Ms7PCXiULv— Alex Volkov (Thursd\/AI) (@altryne) March 28, 2024 Decoding Mixture of Experts A Mixture of Experts (MoE) model is a type of neural network architecture that combines the strengths of multiple smaller models, known as ‘experts’, to make predictions or generate outputs. An MoE model is like a team of hospital specialists. Each specialist is an expert in a specific medical field, such as cardiology, neurology, or orthopaedics. With respect to Transformer models, MoE has two key elements – Sparse MoE Layers and a Gate Network. Sparse MoE layers represent different ‘experts’ within the model, each capable of handling specific tasks. The gate network functions like a manager, determining which words or tokens are assigned to each expert. MoEs replace the feed-forward layers with Sparse MoE layers. These layers contain a certain number of experts (e.g. 8), each being a neural network (usually an FFN). Breaking Down Popular MoEs Databricks DBRX uses a fine-grained mixture-of-experts (MoE) architecture with 132B total parameters, of which 36B are active on any input. It stands out among other open MoE models, such as Mixtral and Grok-1, because it employs a fine-grained approach. A fine-grained mixture of expert models further breaks down the ‘experts’ to perform extremely specific subtasks, splitting the FFNs into smaller components. This can result in many small experts (even hundreds of experts), and then you can control how many experts you want to be activated. The idea of fine-grained experts was introduced by DeepSeek-MoE. Specifically, DBRX has 16 experts and selects four of them, whereas Mixtral and Grok-1 each have eight experts and choose two. According to Databricks, this provides 65x more possible combinations of experts. xAI recently open-sourced Grok 1, which is a 314B parameter Mixture-of-Experts model with 25% of the weights active on a given token, which means at a time it uses 78 billion parameters. Whereas, AI21 Labs’ Jamba is a hybrid decoder architecture that combines Transformer layers with Mamba layers, a recent state-space model (SSM), along with a mixture-of-experts (MoE) module. The company refers to this combination of three elements as a Jamba block. Jamba applies MoE at every other layer, with 16 experts and uses the top-2 experts at each token. “The more the MoE layers, and the more the experts in each MoE layer, the larger the total number of model parameters,” wrote AI21 Labs in Jamba’s research paper. Jamba uses MoE layers to only use 12 billion out of its total 52 billion parameters during inference, making it more efficient than a Transformer-only model of the same size. “Jamba looks very impressive! It’s technically smaller than Mixtral yet shows similar performance on benchmarks and has a 256k context window,” shared a user on X. Alibaba recently released Qwen1.5-MoE which is a 14B MoE model with only 2.7 billion activated parameters. It comes with a total of 64 experts, representing an 8-time increase compared to the conventional MoE setup of eight experts. Similar to DBRX, it also employs a fine-grained MoE architecture where Alibaba has partitioned a single FFN into several segments, each serving as an individual expert. “DBRX is good for enterprise applications, but the Qwen MoE is a cool and great toy to play with,” wrote a user on X. Mixtral 8X7B is a sparse mixture-of-experts network. It is a decoder-only model where the feedforward block selects from eight distinct parameter groups. At each layer, for every token, a router network chooses two of these groups (the ‘experts’) to process the token. It has 47B parameters but uses only 13B active parameters during inference Why Choose MoE? “Mixture-of-Experts will be Oxford’s 2024 Word of the Year,” quipped a user on X. However, jokes aside the reason today MoE models are getting popularity is that they enable models to be pretrained with far less compute, which means you can dramatically scale up the model or dataset size with the same compute budget as a dense model. In an MoE model, not all parameters are active or used during inference, even though the model might have many parameters. This selective activation makes inference much faster compared to a dense model that uses all parameters for every computation. However, there’s a trade-off in terms of memory requirements because all parameters must be loaded into RAM, which can be high. As the need for larger and more capable language models increases, the adoption of MoE techniques is expected to gain momentum in the future.","excerpt":"“Mixture-of-Experts will be Oxford’s 2024 Word of the Year.”","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2024-04-03T17:04:15","publication_year":"2024","word_count":802,"keywords":["Go","TPU","programming_languages:R","AI","neural network","R","programming_languages:Go","XAI","xAI","Databricks"],"extracted_tech_keywords":["AI","neural network","xAI","TPU","Databricks","R","Go","XAI","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-rise-of-mixture-of-experts-llms\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10045419,"title":"Intuit To Hire Over 350 Engineers In India By 2022","content":"Intuit, a global AI-powered platform that helps customers overcome critical financial challenges, has announced plans of expanding its team by adding over 350 engineers in the next twelve months. With India having a large pool of skilled technology professionals, the company is looking to leverage that and grow its key tech capabilities to drive innovation. Intuit is currently working with 1000+ employees in India, and with these new hires, it aims to scale its engineering innovation happening across products. The offered job roles will be in the field of software engineering, product design, product management, data science, risk analytics, and data engineering roles, at entry, mid-level and senior positions. According to the news media, Saurabh Saxena, Intuit’s India Site Leader and Vice President of Product Development, stated that the company is looking to onboard professionals who have an innovative mindset to solve some of the pressing financial problems. Thus, Intuit is looking for professionals interested in AI, data science, cloud, open-source and natural language understanding. Some of the key company goals are — to scale the intelligence of the products using a virtual expert platform and to leverage omnichannel commerce to increase small business growth. Additionally, Intuit is looking to enhance its customer experiences and communications at scale and improve developer productivity. With a vision of being an AI-powered expert platform, Intuit is looking to expand its team in India in order to address its 100 million customers worldwide.","excerpt":"Intuit, a global AI-powered platform that helps customers overcome critical financial challenges, has announced plans of expanding its team by adding over 350 engineers in the next twelve months.","categories":["AI News"],"tags":["engineers","engineers India","Intuit"],"author_name":"Sejuti Das","publish_date":"2021-08-06T13:41:47","publication_year":"2021","word_count":239,"keywords":["data science","Go","engineers India","AI","Intuit","innovation","engineers","RAG","Aim","data engineering","ViT","analytics","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","data engineering","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/intuit-to-hire-over-350-engineers-in-india-by-2022\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10047810,"title":"Bangalore-Based Deeptech SaaS Startup NeuroPixel.AI Raises $825k In Seed Round","content":"Deeptech SaaS startup NeuroPixel.AI has raised $825k in a Seed round led by Inflection Point Ventures. Other investors in the round include Entrepreneur First, Huddle, Dexter Angels, and Rishaad Currimjee. The funds raised will be used to scale up their R&D team to accelerate the transition of their product from beta to production and expand their ‘training set’, a crucial piece of the puzzle for every machine learning algorithm. “We are extremely excited at the opportunity to work with IPV as we scale our young startup. Having access to such an accomplished and diverse pool of industry veterans amplifies the value of the seed investment raised many times over, and this access to industry veterans is clearly where IPV stands apart,” said Arvind Nair, Co-founder & CEO, NeuroPixel.AI. Founded in 2020 by Arvind Nair (CEO) and Amritendu Mukherjee (CTO), the startup works to apply advanced AI\/ML and statistical learning theory in the Computer Vision and Image Processing area for online retail storefronts. Ankur Mittal, Co-Founder, Inflection Point Ventures, says, “Artificial intelligence (AI), as well as Machine Learning (ML) technologies, are omnipresent across sectors where digital transformation is making inroads. Online commerce is at the forefront of this transition. As e-commerce will expand, so will the need to put up quality and realistic product pictures online. In fashion commerce, it is a big part of the buyer’s purchase decision. However, it is not a seamless process and is both time-consuming and expensive, especially for SMEs and social sellers, two segments that are growing exponentially. NeuroPixel is trying to solve this problem by building a product that can transform online fashion storefronts through catalog image-based personalization and virtual try-on, helping the average consumer make a far more informed purchase decision. It will save the businesses both time and money and will allow them to bring their products faster to the market, helping them generate higher revenue and lower returns. The product has huge international appeal as well. Strong founding team, the uniqueness of the approach adopted by NeuroPixel, and their high-calibre R&D team comprising PhDs and postgraduates from The Indian Institute of Science (IISc) were key factors influencing our decision to invest in the company.” NeuroPixel.AI’s first product – an AI-powered cataloguing tool – will enable clients to shoot any apparel on just a mannequin, and their technology will render the apparel on models of different sizes in different poses. In the near term, the company aims to reduce cataloguing spends by 30% and reduce process times by 90%. It was also among the six startups selected for investment by the ISB D-Labs incubator under their Seed support program in collaboration with The Department of Science and Technology, Government of India. In addition, the startup has also been selected into the Huddle accelerator, which will commence from the closure of this round of funding.","excerpt":"The funds raised will be used for scaling up their R&D team to accelerate the transition of their product from beta to production.","categories":["AI News"],"tags":["AI Startups"],"author_name":"kumar Gandharv","publish_date":"2021-09-06T12:12:43","publication_year":"2021","word_count":472,"keywords":["Go","machine learning","artificial intelligence","AI","ML","Git","computer vision","RAG","Aim","R","AI Startups"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bangalore-based-deeptech-saas-startup-neuropixel-ai-raises-825k-in-seed-round\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10131642,"title":"The Integration of AI in Unified Data Intelligence Platforms","content":"Unified Data Intelligence Platforms are transforming the way organisations handle data. Built upon the foundation of lake houses by integrating Gen AI capabilities, these platforms automatically analyse metadata, data, queries, and reports, generating lineage and understanding an organisation’s data model, metrics, and KPIs.” This evolution addresses significant challenges like governance and security while ensuring effective data handling. AIM recently spoke to Saravana Kumar KJ, senior manager and Databricks solutions architect champion at Tredence, who, with his extensive experience and certifications, provided valuable insights into the current and future state of data intelligence. KJ began by tracing the evolution of data intelligence platforms. “In the past, it was all about data warehousing, dealing primarily with structured data. Then came the necessity of handling unstructured data, leading to the advent of data lakes,” KJ said. Lake houses, he said, addressed these needs but brought along their own set of challenges in governance, security, and management. Addressing Data Volume, Variety, and Velocity One of the critical capabilities of unified data intelligence platforms is handling the increasing volume, variety, and velocity of data. “Unified platforms utilise cloud-based infrastructure to scale according to data processing needs and employ distributed computing to process large datasets efficiently,” KJ noted. This ensures that organisations can handle vast amounts of data without compromising on processing time. In terms of variety, these platforms consolidate diverse data sources, including structured, semi-structured, and unstructured data, into a single system. “They incorporate databases, data lakes, IoT data, social media data, and more, all within a centralised repository,” KJ added. This consolidation is crucial for efficient data management and analysis. These platforms also address data velocity by enabling real-time analytics and stream processing. “This allows organisations to analyse data as it is produced, which is essential for applications like real-time customer engagement and operational monitoring,” KJ emphasised. AI Integration for Actionable Insights Integrating AI with data intelligence platforms significantly enhances an organisation’s ability to generate actionable insights. KJ elaborated, “AI improves data governance and quality by automatically tagging, categorising, and standardising data, ensuring consistency across the platform.” This consistency is vital for maintaining data integrity and making informed decisions. Compliance and security are also enhanced through AI-driven tools that monitor and enforce data protection regulations like HIPAA, GDPR, and CCPA. “These tools help reduce the risk of data breaches and ensure adherence to legal standards,” KJ noted. One of the biggest enhancements that AI enables is personalised insights tailored to user roles and preferences, optimising the relevance of information for each user. “Natural language processing allows non-technical users to interact with data using natural language queries, democratising data access and insights,” KJ explained. Improving Organisational Efficiency AI-driven automation within data intelligence platforms significantly enhances organisational efficiency by reducing the time needed to develop, test, and launch new products and services, thereby ensuring a faster time to market and a competitive edge in the marketplace. “Enhanced collaboration is a key benefit, as AI integration includes tools that enable seamless communication and data sharing among team members,” KJ stated. This ensures that everyone is working with the most current and accurate data. Automation of routine tasks through AI increases productivity by allowing employees to focus on higher-value activities. “It also reduces operational costs by optimising resource management and continuously improving processes,” KJ added. Unified data intelligence platforms hold transformative potential for various industries. In healthcare, these platforms can analyse patient data to predict health trends, create personalised treatment plans, and improve medical imaging analysis. “In finance, they enable real-time fraud detection and personalised financial services based on customer behaviour,” KJ noted. Retail benefits from customer insights and personalization, optimised pricing strategies, and streamlined inventory management. “Manufacturing sees improvements in predictive maintenance and quality control, while energy and utilities benefit from smarter grid systems and improved energy management,” KJ explained. Transportation and logistics also see advancements in route optimization and operational efficiency. Future of Unified Data Intelligence Platforms Looking ahead, KJ envisions several advancements in unified data intelligence platforms. “Multilingual capabilities in natural language processing, real-time analytics, built-in data quality checks and automated compliance monitoring are among the future enhancements,” he suggested. These advancements will further enhance the usability and effectiveness of these platforms. KJ also emphasised that as deep learning models advance, unified data intelligence platforms will be able to process and analyse increasingly complex datasets with greater accuracy.","excerpt":"One of the critical capabilities of unified data intelligence platforms is handling the increasing volume, variety, and velocity of data.","categories":["AI Highlights"],"tags":["AI in Data Intelligence"],"author_name":"Mohit Pandey","publish_date":"2024-08-06T16:16:18","publication_year":"2024","word_count":723,"keywords":["AI in Data Intelligence","Go","AI","ML","distributed computing","Aim","deep learning","Databricks","analytics","R","fraud detection"],"extracted_tech_keywords":["AI","ML","deep learning","analytics","Aim","fraud detection","distributed computing","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/the-integration-of-ai-in-unified-data-intelligence-platforms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10095700,"title":"Tech Mahindra, BosonQ Psi to Move AI to Quantum","content":"The Indian government’s allocation of INR 6,000 crore (approximately $800 million) in April for quantum computing startups reflects its recognition of the potential of this technology. However, to foster a thriving quantum ecosystem in India, efforts must extend beyond funding. Recognising the immense potential and need of quantum computing, Tech Mahindra, the IT giant, has formed a strategic partnership with BosonQ Psi, a pioneering quantum computing startup in India. AIM caught up with Nikhil Malhotra, chief innovation officer at Tech Mahindra and creator of Makers’ Lab and Aditya Singh, one of the founding members and head of growth and infrastructure of BosonQ Psi to understand the vision and scope of the partnership. Together, they aim to harness the power of quantum software and advance the realms of computing and post-quantum cryptography. Malhotra said that while quantum computing encompasses various applications such as communication and sensing, the partnership between Tech Mahindra and BosonQ Psi focuses on exploring the practical use cases and driving innovation. Their collaboration extends to both scientific and business domains, spanning areas like drug discovery in healthcare, untapped opportunities in the telecom sector, and simulation-based optimisation for personalised customer experiences and network design. By leveraging BosonQ Psi’s expertise in software simulation, Tech Mahindra aims to deliver cutting-edge solutions to its customers. Read: Year in Review: Will Tech Mahindra’s Good Spell Continue in 2023? Currently, BosonQ Psi is in the process of developing BQPhy™, an innovative software suite that utilises quantum power to deliver simulation-as-a-service (Q-SaaS). The groundbreaking software suite will offer unparalleled computational benefits in a wide range of engineering simulation domains. Currently, it offers functionalities in structural mechanics, thermal sciences, and design optimisation. Malhotra explains that Tech Mahindra’s decision to partner with BosonQ Psi was driven by the startup’s exceptional prowess in software simulation, a critical aspect for digital twin creation and growth. “While hardware providers like IBM have made significant strides in quantum technology, there are few software providers operating in this niche domain,” he said, emphasising that BosonQ Psi has carved out a unique position in this arena. Boosting AI with Quantum Tech Mahindra had been actively involved in the development of generative AI even prior to the release of OpenAI’s ChatGPT. The company’s innovative Storicool platform, an automatic content creation tool, was regarded as ahead of its time by CP Gurnani, the chief executive of Tech Mahindra. In April of this year, Tech Mahindra introduced its Generative AI Studio as part of its comprehensive suite of AI offerings and solutions known as amplifAI0->∞. Read: How Indian IT Giants are Bringing GenAI to Their Clients Now, this partnership opens doors to explore synergies between quantum computing and AI. Malhotra said, “It is important to note the large amount of energy that is being utilised in these AI systems, and this is where quantum neural networks, conversational systems, and sequence systems hold promise in AI.” Although he admits that further research and development are required to fully realise their potential. Preliminary studies are already underway to uncover the applications of quantum in language-related fields, hinting at a future where quantum-powered AI becomes a reality. Aligned in their vision of moving beyond just being consultancy services, Tech Mahindra seeks to commercialise quantum computing and apply this transformative technology to benefit their customers. Malhotra said that a lot of customers want Tech Mahindra to explore this field and reap the benefits. The collaboration between Tech Mahindra’s research and development wing, Makers Lab, and BosonQ Psi was built on the shared ideology of translational research, translating cutting-edge innovations into practical solutions. With a customer-centric approach, Tech Mahindra aims to showcase the possibilities of quantum computing and its tangible impact, going beyond mere consultancy services. Quantum, revenue, and India Speaking with AIM, Mark Mattingley-Scott, chief revenue officer at Quantum Brilliance, said that generating revenue in a quantum startup depends on when we reach the final commercial state of quantum which might take a few more years from now. Read: ‘India Can be a 600-Pound Gorilla of Quantum Applications’ Similar views were expressed by Singh from BosonQ Psi. The company is targeting the commercialisation of its quantum solutions this year. By collaborating with Tech Mahindra instead of independently pursuing hardware development, the startup can focus on utilising quantum capabilities to benefit clients. This strategic approach ensures that customers can experience the advantages of quantum computing when the hardware is ready for deployment. Tech Mahindra’s previous partnership with IBM Quantum, as well as their collaboration with IQM, highlights the company’s commitment to exploring different quantum hardware options and leveraging each for specific use cases. IQM’s unique hardware offerings, such as their demonstration of a high-value introduction of quantum volume, make them a natural fit for Tech Mahindra’s software layer development. The addition of BosonQ Psi to this equation brings valuable software expertise, resulting in a partnership that combines the best of both worlds—software and hardware—for transformative quantum solutions. “Quantum still remains a far off field for Indian IT giants,” said Malhotra. A comprehensive approach encompassing quantum research, education, industry partnerships, and understanding the needs of startups is essential. By nurturing this ecosystem, India can emerge as a global leader in quantum computing, capitalising on its unique challenges and diverse linguistic landscape to develop novel algorithms and transformative solutions. “The Indian government wants to build its own hardware just like any other government. We need a good think tank within India to build that,” concluded Singh.","excerpt":"Nikhil Malhotra from Tech Mahindra and Aditya Singh from BosonQ Psi spoke with AIM to discuss the future of quantum and its convergence with AI","categories":["IT Services"],"tags":["quantum","Quantum Computing"],"author_name":"Mohit Pandey","publish_date":"2023-06-23T17:30:00","publication_year":"2023","word_count":903,"keywords":["Quantum Computing","Go","ChatGPT","GenAI","OpenAI","AI","neural network","RAG","quantum","Aim","generative AI","R"],"extracted_tech_keywords":["AI","neural network","generative AI","GenAI","ChatGPT","OpenAI","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/tech-mahindra-bosonq-psi-to-move-ai-to-quantum\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10005734,"title":"How This Startup Is Building A Language Understanding Engine To Automate Conversations","content":"According to Research and Markets reports, Artificial Intelligence for speech recognition market in India is anticipated to expand at a compound annual growth rate (CAGR) of ~65.17% during the forecast period (2019-2024) and is expected to reach a value of INR 14.61 Bn by 2024. The increasing demand for smart speakers and voice-enabled devices, coupled with rising penetration of speech recognition technology in customer care services are driving this growth, further stimulating development and innovation in the space. Enterprises across the globe have been spending a considerable amount of money on contact centres and agents, yet the customers are rarely satisfied due to problems like more prolonged time consumption. To mitigate such issues, Bangalore-based Vernacular.ai is aiming to resolve by automating a large chunk of non-productive calls over voice. The company recently raised a Series A investment of $5.1 million. Founded in 2016 by two IIT Roorkee grads, Sourabh Gupta and Akshay Deshraj, Vernacular.ai is an AI-first SaaS business startup that is driven with the vision to become the leading voice automation\/AI platform in the world. How Is It Driving Language Understanding Vernacular.ai delivers two unique products — VIVA (Vernacular Intelligent Voice Assistant) and VASR (Vernacular Automated Speech Recognition). VASR enables enterprises to convert audio to text by applying powerful neural network models in an easy-to-use API. This API can recognise over 160 dialects in ten different languages to support the enterprise user base. Built on top of VASR, VIVA is an AI-based voice automation platform which helps automate 80 percent of calls handled by a call centre and reduces agents’ average call handling time by 30 per cent. VIVA uses innovative natural language understanding and speech recognition technology, which supports around 10 Indian languages. It also enables hyper-personalisation of customer calls with its ability to understand the various characteristics of the speaker, like accent, speech rate, age, gender, region, even dialect. On being asked how these products are different from others in the market, Gupta pointed out three key differences- Validated self-learning technology that ensures the system is really improving over time.Capability to automate full end-to-end open conversations rather than limited automation over IVR and closed-domain calls.VIVA, built using state of the art advances in voice AI can identify the user’s persona that includes the language, dialect, accent and sentiment. The Tech Behind It Vernecular.ai uses AI and machine learning for accomplishing a number of tasks, such as- In the Contact Center Automation solution for understanding the intent of the user in their utterances.For recognising text from the speech, the company uses deep neural networks trained on thousands of hours of acoustic data.For identifying whether a person is speaking or not. In audios, they employ recurrent models which work on real-time audios with minimal latency.For controlling the behaviours of voice bot based on the current understanding of user and call state during a call flow.Modelling usage patterns in all the languages, which covers conversational nuances and semantic equivalence of words and phrases.In the Text to Speech models where the company trains models to synthesise audio which replicate nuances of conversational human speech.For analysing human to human conversations to get insights about resolution, user satisfaction etc. Core Tech Stack Talking about the core tech stack, the company mostly uses Python with various ML frameworks. Deshraj said, “Most commonly TensorFlow, but people have been using PyTorch in various projects too. In places where more performance is needed, we use languages like C, C++ and most recently Rust. For speech recognition, Kaldi as a framework works well for us since it gives a lot of hackability, though we do use other stock end-to-end frameworks too.” He added, “Since most of our engineering stack is in Golang, we also use that whenever required for stitching pieces together or for things which are less Machine Learning. At times we have experimented with Clojure for such tasks.” On the backend, Vernacular.ai has core services written in Golang, C\/CPP & Python, and for frontend applications, they use React.js and Elm. Gupta also mentioned that they support relational databases like PostgreSQL, MySQL & Oracle, and the services usually communicate over gRPC and JSON-over-HTTP. They also use Kubernetes for container orchestration. Tackling Hiring Phase The general hiring process at Vernacular.ai involves technical and cultural fit rounds. Gupta said, “Rigorous filtering aside, one important piece for us is sourcing from the right places and finding candidates who are going to excel in our environment. The things we look for when hiring are learnability, ambitiousness, ability to work with unknowns, and someone teeming with enthusiasm whom we would love to have on-board.” Future Roadmap In the next five years, the company wants to build a language engine for the world. We envision a world where human-machine interactions over voice will become second nature to everyone. Our mission is to build the structural components to ensure this revolution unfolds, said Deshraj on a concluding note.","excerpt":"According to Research and Markets reports, Artificial Intelligence for speech recognition market in India is anticipated to expand at a compound annual growth rate (CAGR) of ~65.17% during the forecast period (2019-2024) and is expected to reach a value of INR 14.61 Bn by 2024.  The increasing demand for smart speakers and voice-enabled devices, coupled […]","categories":["AI Startups"],"tags":["automated saas intelligence","innovative technology advancements","language understanding","Speech Recognition","Startups","Vernacular Automated Speech Recognition"],"author_name":"Ambika Choudhury","publish_date":"2020-08-28T16:00:26","publication_year":"2020","word_count":814,"keywords":["artificial intelligence","innovative technology advancements","machine learning","AI","neural network","PyTorch","Vernacular Automated Speech Recognition","ML","automated saas intelligence","RAG","language understanding","Aim","Speech Recognition","Startups","TensorFlow","kubernetes"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","Aim","TensorFlow","PyTorch","RAG","kubernetes"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-startup-is-building-a-language-understanding-engine-to-automate-conversations\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":7342,"title":"Analytics India Salary Study 2015 – By AIM &#038; Great Lakes Institute of Management","content":"In the past couple of years, data analytics has grown from being a discretionary spend area to a service that is a need for competitive advantage. Not only are businesses looking at internal well-structured corporate or customer data but also are exploring large external data sources (on social networks, internet, e-mails, text documents, etc.), which are usually unstructured, and need to be combined with structured data to conduct meaningful analysis. Managing the sheer volume, variety, and velocity of data that is being generated (with the innumerable technological interface devices) is a relatively new challenge for the typical business organization. Thus, the demand for trained analytics and big data professionals is increasing at a tremendous rate. Supply is still very constrained and this means that over half the positions on offer still remain vacant making it a lucrative career option for professionals. This annual Analytics India Salary Study 2015 is an initiative by Analytics India Magazine in partnership with Great Lakes Institute of Management to highlight the salary trends in the industry across cities, experience levels and sectors. Read the latest Analytics India Salary Study 2016 Key Trends In the past few years there is a rise in demand of the data analytics professional and the overall salary trends look very optimistic. Overall Salary Trends: The overall average salaries of analytics professionals across the country is 9.4 lakhs per annum. This is across seniority levels and expertise. The average salaries for analytics professionals increased by 21% from the same time last year. This is an excellent increase given that the recruitment at entry level has been high last year. The salaries for mid to senior levels professionals have increased in the range of 25-40% last year. 14% of all analytics professionals command more than 15 lakhs salary. 37% command less than 6 lakhs Almost 12% of entry level professionals in analytics command more than 6 lakhs Salary Trends across Cities: Mumbai pays the highest salary to analytics professionals at an average of 9.9 lakhs per annum, marginally higher than Bangalore at 9.8 lakhs average. After Mumbai, Bangalore and NCR, Pune has the highest average salaries at 8 lakhs per annum. Among cities, Hyderabad had the highest year on year hike in salary at almost 25%. In Mumbai, there are more analytics professionals in the 10 – 25 lakh range salary than any other city (at 31%). Key inferences: The number of professionals in the high income bracket is low and the majority of professionals are employed at a fresher or senior analyst level across India. This is in line with the expectation where when an area grows in volume, more of junior level professionals are required. Though Mumbai has traditionally offered the highest salaries because of its high cost of living and the trend still continues, Bangalore is fast catching up and its salaries are almost at par with that of Mumbai now. Salary Trends across Experience Levels: From Analyst (0-3 years) to Senior Analyst (4-6 years), an analytics professional can expect almost 75% average hike in the salary. From Senior Analyst to Assistant Manager (7-9 years), it is almost 57%. Almost 85% of all analytics professionals in India with more than 12 years of experience can expect to have more than 15 lakhs per annum 45% have more than 25 lakhs per annum salary. Key inferences: Things are a lot perkier as a data analyst and trends show that at each level salary increment is upwards of 50%. The biggest jump in salary is from the Analyst’s to a Senior Analyst’s level. At the Director’s level, salaries are much higher in Bangalore, Delhi\/NCR and Mumbai in comparison to Pune, Chennai and Hyderabad. At the Analyst’s level, salaries are almost at par across the cities. Salary Trends across Industries: Ecommerce emerged as the highest paymaster paying its analytics professionals an average salary of 12.9 lakhs per annum. The lowest paymasters were the Media\/Advertising and Pharma sectors with average salaries of 5 and 8.1 lakhs per annum respectively. Key inferences: In general captive centres (eg: ecommerce, retail, telecom) paid higher salaries to retain their talent. As compared to captives, analysts that work for service providers either in the IT\/ITES or Media\/Advertising sectors have the opportunity to move around domains and gain expertise. This is the advantage that affords service providers the luxury of paying lower salaries but still attracting high quality talent. Conclusion Analytics drives insights and insights lead to better decisions. As businesses find themselves in an era of unprecedented competition and changing economic landscape, data analytics is the crucial component that can help them build a competitive advantage and make well-informed choices. Businesses of all sizes today are waking up to this realization and to help these businesses realize their analytics goal, the skilled analytics professional is looked as the saviour of sorts. As a result, the analytics job market will grow like never before and there will be unparalleled opportunities for those with analytic skills. Salaries will continue to increase and we will see professionals from other sectors honing their analytics skills and switching careers. The time for the data-savvy analytics professional is here! [divider divider_color=”#777777″ link_color=”#777777″ size=”1″] Analytics India Salary Study 2015 Download the complete report below: [attachments include=”7353″]","excerpt":"In the past couple of years, data analytics has grown from being a discretionary spend area to a service that is a need for competitive advantage. Not only are businesses looking at internal well-structured corporate or customer data but also are exploring large external data sources (on social networks, internet, e-mails, text documents, etc.), which […]","categories":["AI Features"],"tags":["analytics career","data scientist india salary","great lakes analytics","masters in data analytics in india"],"author_name":"Дарья","publish_date":"2015-04-29T07:00:55","publication_year":"2015","word_count":871,"keywords":["big data","Go","programming_languages:R","AI","data scientist india salary","programming_languages:Go","RAG","great lakes analytics","masters in data analytics in india","analytics career","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-india-salary-study-2015-by-aim-great-lakes-institute-of-management\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10118931,"title":"‘We May be Able to Create an Infinite Data Generation Engine with Synthetic Data,’ says Anthropic CEO","content":"Despite the scepticism about producing quality data using synthetic data, Anthropic chief Dario Amodei recently believes that creating an infinite data generation engine that can help build better AI systems is possible. “If you do it right, with just a little bit of additional information, I think it may be possible to get an infinite data generation engine,” said Amodei in an interview with CNBC while discussing the challenges and potential of using synthetic data to train AI models. “We are working on several methods for developing synthetic data. These are ideas where you can take real data present in the model and have the model interact with real data in some way to produce additional or different data,” explained Amodei. Citing AlphaGo, he said, it is actually possible to inject very small amounts of new information to get more than you started with. “If you go back to systems eight years ago, so if you remember AlphaGo, note that the model there just trains against itself with nothing other than the rules of Go to adjudicate,” he added, saying that those little rules of Go, the little additional piece of information, is enough to take the model from “no ability at all to smarter than the best human at Go.” Amodei believes that if we do it right, with just a little bit of additional information, we can create an infinite data generation engine. For those unaware, AlphaGo systems were trained by reinforcement learning, where the neural networks were initially bootstrapped from human gameplay expertise. Meta’s AI chief, Yann LeCun, a self-supervised-learning proponent, has slightly different views, criticising reinforcement learning for being inefficient and impractical for real-world applications when used on its own. “A lot of the success of machine learning at least until fairly recently was mostly with supervised learning. Reinforcement learning gave some people a lot of hope, but turned out to be so inefficient as to be almost impractical in the real world, at least in isolation, unless you rely much more on something called self-supervised learning, which is really what has brought about the big revolution that we’ve seen in AI over the last few years,” said LeCun. Self-supervised learning is a technique used where the model autonomously discovers patterns and structures in data without explicit labels. Other techniques Besides reinforcement learning and self-supervised learning, LeCun also discussed other techniques for data generation and training AI systems. This includes generative models such as GANs and VAEs, which generate new data by learning the distribution of existing data. “There are systems of this type that have been trained to produce images and they use other techniques like diffusion models,” he added. Predictive learning models are also another interesting method, which forecasts future states or missing parts of data to aid in learning representations and dynamics. “A particular way of doing it is you take a piece of data… and then you train some gigantic neural net to predict the words that are missing,” said LeCun. Then, there are energy-based models, which score data configurations based on their probability of supporting various tasks, including generation and classification. “Energy-based models learn a scalar energy for each configuration of the variables of interest,” explained LeCun. Joint embedding predictive architectures (JEPA) is another technique for training AI systems. It uses embeddings to predict parts of data from others, facilitating the learning of complex data relationships. “Instead of reconstructing y from x, you run both x and y through encoders… you do the prediction in representation space,” he explained. Latent variable models also help in data generation. These models integrate hidden variables that explain inherent data variability, which is essential for complex generative tasks. “Latent variable models consist in models that have a latent variable z that is not given to you during training or during tests that you have to infer the value of,” mentioned LeCun. Lastly, there is hierarchical planning. This technique is crucial for enabling AI systems to operate in complex, real-world environments where decisions need to be made at both strategic and tactical levels. Here, LeCun gave an example of planning a trip to Paris through high-level tasks (like getting to the airport) and detailed steps (like navigating to the departure gate), touching upon the reasoning aspect. Join us at the Data Engineering Summit 2024 on May 30-31 at the Hotel Radisson Blu in Bengaluru, India, organised by AIM for two days of cutting-edge discussions on data engineering innovation featuring top engineers and innovators from leading tech companies.","excerpt":"“If you remember AlphaGo, note that the model there just trains against itself, using only the rules of Go to adjudicate,” said Anthropic CEO Dario Amodei.","categories":["AI News"],"tags":["Dario Amodei"],"author_name":"Donna Eva","publish_date":"2024-04-24T13:09:39","publication_year":"2024","word_count":751,"keywords":["Anthropic","Go","machine learning","Dario Amodei","AI","neural network","Scala","diffusion models","Aim","data engineering","R"],"extracted_tech_keywords":["AI","machine learning","neural network","Anthropic","Aim","R","Go","Scala","data engineering","diffusion models"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/we-may-be-able-to-create-an-infinite-data-generation-engine-with-synthetic-data-says-anthropic-ceo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":68390,"title":"Deloitte Launches The Deloitte AI Institute &#8211; A Centre To Advance The Development Of AI For Enterprises","content":"Recently, Deloitte announced the launch of the Deloitte AI Institute – a centre that focuses on artificial intelligence (AI) research, eminence and applied innovation across industries. The institute is said to apply the cutting-edge research to help address a wide spectrum of relevant AI use cases and bridge the ethics gap surrounding AI. The Deloitte AI Institute brings together the brightest minds and research in #AI to advance human-machine collaboration in the Age of WithTM and address a wide spectrum of use cases. Lead the future of AI: https:\/\/t.co\/cIfdtKp2i9— Deloitte on AI (@DeloitteAI) June 25, 2020 In a blog post, Nitin Mittal, AI co-leader and principal, Deloitte Consulting LLP said, “The Deloitte AI Institute is being established to advance the conversation and development of AI for enterprises.” He added, “Our goal is to blend Deloitte’s deep experience in applied AI with a robust network of some of the most intelligent AI minds in the world to challenge the status quo. Through the power of this centre, we aim to deliver impactful and game-changing research; and innovation to help our clients lead in the ‘Age of With,’ a world where humans work side-by-side with machines.” The network of this institute will include the top industry thought leaders and academic luminaries, start-ups, research and development groups, entrepreneurs, investors as well as innovators. “With our unique experience, investments in AI and work with top organisations, we believe the Deloitte AI Institute can ignite ground-breaking applied AI solutions for enterprises,” said Beena Ammanath, executive director of Deloitte AI Institute, Deloitte Consulting LLP. “Further, to help enterprises advance with AI, we will aim to help organisations remain distinctively human in a technology-driven world.” Irfan Saif, Deloitte Risk & Financial Advisory principal, Deloitte & Touche LLP and Deloitte AI co-leader stated that with AI ethics, the AI institute aims to help organisations achieve a positive future by bringing together top stakeholders from all sectors of society to discuss and co-design effective policies and frameworks, such as Deloitte’s Trustworthy AI framework, for governing AI.","excerpt":"Recently, Deloitte announced the launch of the Deloitte AI Institute – a centre that focuses on artificial intelligence (AI) research, eminence and applied innovation across industries. The institute is said to apply the cutting-edge research to help address a wide spectrum of relevant AI use cases and bridge the ethics gap surrounding AI.  In a […]","categories":["AI News"],"tags":["ai consulting","App Development","Deloitte","human touch to ai"],"author_name":"Ambika Choudhury","publish_date":"2020-06-26T18:33:15","publication_year":"2020","word_count":337,"keywords":["Go","ai consulting","artificial intelligence","Deloitte","programming_languages:R","AI","innovation","AI ethics","App Development","Aim","Rust","GAN","human touch to ai","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","Rust","GAN","AI ethics","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deloitte-launches-the-deloitte-ai-institute-a-centre-to-advance-the-development-of-ai-for-enterprises\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10060135,"title":"What drove Sony to buy Halo studio Bungie","content":"After Microsoft dropped a big bag on Activision, Sony decided to buy Bungie video game studio for $3.6 billion. The independent studio developed Destiny, the famous online FPS (first-person shooter) and the iconic Halo series, which fuelled Xbox’s growth. Bungie will continue to work independently, and Destiny will be published across all gaming platforms. The company has also stated that its future games will not be exclusives for Sony. So then, what pushed the PlayStation (PS) maker to buy the studio? A slice of the answer lies in the comment made by Sony Interactive Entertainment (SIE), “Bungie purchase will give SIE access to its live game services and technology expertise, allowing SIE to reach billions of players”. Unlike Microsoft, Sony is a multimedia giant with its claws on most of the entertainment industry. Meanwhile, Bungie has been planning to jump from its SciFi oriented games to PvP (player vs player) competitive games for a while now. In 2019, the studio pledged to release at least one non-Destiny game by 2025, and Bungie CEO Pete Parsons said in 2020 that Bungie had been working on multiple new games for the past three years. He also mentioned, “telling new stories and creating new IPs”. (i would like to add this tweet to this story) – Other players have entered the buy zone. Take-Two Interactive, the Grand Theft Auto developer, will buy FarmVille maker Zynga for $12.7 billion. Why buy now? There has been a lot of buzz surrounding this deal, especially after Microsoft bought Activision. But, Sony’s recent buy is not out-of-the-blue, and it has been under discussion for the past six months. In a way, Sony was already planning two steps ahead of the Xbox maker. Previously, SIE worked with Bungie to drop a few Destiny exclusives in recent years. “We always had a strong partnership with Bungie ever since Destiny was developed, and I could not be more thrilled to officially welcome Bungie to the PlayStation family,” said Jim Ryan, President and CEO, SIE. This deal benefits both parties and must be considered a “collaboration” and not just any acquisition. SIE can now expand PS products to a larger audience and deploy Bungie’s live services power. Meanwhile, Bungie can utilise SIE’s ability to focus on new projects and improve its future potential. Jim Ryan, the PS boss, said, “With the purchase of Bungie, we will use the studio’s experience to create games as a service that engage the community in the long term and to support future PlayStation Studio projects that follow that model.” In an interview with Gamesindustry.biz, Ryan mentioned that the Bungie deal was not responding to the other big acquisitions announced this year. However, it isn’t easy to look past that as it is an attempt to widen the gaming industry consolidation. Large entertainment companies are slowly swallowing the smaller video game studios, a trend that has picked up in the last two years. However, it is not Sony’s only buy; in 2021, it bought Nixxes Software (PC port developer), Returnal developer Housemarque, The Playroom (an AR interactive game) maker Firesprite Studios, Bluepoint Games (known for making the best remastered\/remake games), and finally, Valkyrie Entertainment, a God of War support studio. Given the recent acquisitions spree, there are other rumours in the gaming industry like Sony buying Square Enix and making Final Fantasy a PS exclusive. Will Microsoft buy Ubisoft to list Assassin’s Creed to its game pass? These unimaginable outcomes might just come true. SIE has signalled that there could be more purchases soon after Ryan said that Sony is still not done and has a long way to go. Can Sony claim the metaverse prize? Each company has its distinct features and a target demographic. Since Sony launched PS, it created a very competitive console market and placed itself as the frontrunner. However, it still lacks the agility of a gaming PC. For the longest time, the quality of PS games overshadowed Triple-A games (AAA). To differentiate itself from Xbox and Nintendo, PlayStation focused on indie developers and released character-based and narrative-driven games (exclusives). Gamers always played story-based games, which delivered a more progressive and natural gaming experience. From action RPGs (role-playing games), turn-based strategy games, to the souls’ genre, one thing remained constant – storytelling. Over the years, the demand for open-world games grew, and Sony capitalised on it. However, the best ones were exclusives (only for PS) and were not a live service game (online), which meant people often used to finish the game and then move on. You can play it again at any time you want. It was not a player-to-player interactive service, and now Sony plans to go that direction after Ryan discussed Sony’s plans for live service games. With the ongoing hype around metaverse, Sony can build on its player base and expand to Destiny’s audience. Destiny 2 has an estimated player base of 38.8 million players, making it the second most popular massively multiplayer online game (MMO) of all time, only second to World of Warcraft. Hence, Sony can benefit more from this deal after it slowly moves to a different demographic amidst the rapid demand for a realistic gaming experience.","excerpt":"Unlike Microsoft, Sony is a multimedia giant with its claws on most of the entertainment industry and Bungie plans to shift to PvP games.","categories":["AI Features"],"tags":[],"author_name":"Akashdeep Arul","publish_date":"2022-02-09T17:00:00","publication_year":"2022","word_count":863,"keywords":["Go","API","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","Aim","R"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","API","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-drove-sony-to-buy-halo-studio-bungie\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011340,"title":"Intel Acquires Cnvrg.io To Lead The Race In The AI Game","content":"Intel, in its latest acquisition spree, has acquired Israel-based Cnvrg.io. The deal, like most of the deals in the past, is aimed at strengthening its machine learning and AI operations. The 2016-founded startup provides a platform for data scientists to build and run machine learning models that can be used to train, run comparisons and recommendations, among others. Co-founded by Yochay Ettun and Leah Forkosh Kolben, Cnvrg was valued at around $17 million in its last round. According to a statement by Intel spokesperson, Cnvrg will be an independent Intel company and will continue to serve its existing and future customers after the acquisition. However, there is no information on the financial terms of the deal or who will join Intel from the startup. The deal comes merely a week after Intel’s announcement of acquiring San Francisco-based software optimisation startup SigOpt, which it did to leverage SigOpt’s technologies across its products to accelerate, amplify and scale AI software tools. SigOpt’s software technologies combined with Intel hardware products could give it a major competitive advantage providing differentiated value for data scientists and developers. How Will Intel Benefit From Cnvrg The code-first platform by Cnvrg is built by data scientists and for data scientists. It helps them to focus on building algorithms while spending less time on DevOps and not worry about building or maintaining the platform they run on. It leads to more number of models in production, thereby increasing the business value. It works across on-premise, cloud and hybrid environments and competes with the likes of Databricks, Sagemaker and H2O.ai. Its capabilities of offering end-to-end solutions from research to production bring a unique perspective to the startup that Intel is willing to vouch upon. In fact, in a recent blog post, Cnvrg.io claimed to improve visibility and increase machine learning server utilisation by up to 80%. It allows administrators to monitor and compare overall capacity against both allocation and utilisation, with graphs to identify gaps in the efficiency. AI is increasingly important to Intel’s chip business. As per reports, Intel’s sales of AI processors more than doubled over the last year to $3.5 billion in 2019. As Intel is refocusing its business around next-generation chips to better compete against the likes of Nvidia and smaller players like GraphCore, it makes sense for Intel to invest in AI tools for customers. It can help with the compute loads that they will be running on those chips. Cnvrg will further strengthen Intel’s ambitions in the machine learning space that are currently driven by its Mobileye business and from AI chipmaker Habana. Intel’s acquisition of Cnvrg will help it continue to build a full suite of AI tools for developers to complement its next-gen hardware. Building AI-Focused Strategy At Intel Over the last few years, Intel has been working on strengthening its AI capabilities and shifting a focus from the semiconductor industry, which is evident with a series of acquisitions of AI-based startups that it has been doing over the last few years. Intel has made some bold decisions by acquiring startups for a hefty price. For instance, Intel acquired Nervana for more than $350 million, back in 2016 to build on its fully-optimised software and hardware stack for deep learning. The same year, Intel acquired an Irish chip company Movidius for $400 million which specialised in designing low-power processor chips for computer vision. The following year, Intel bought Mobileye for a whopping $15 billion to further boost the development in computer vision, data analysis, machine learning and autonomous driving. Further, its acquisition of Israel-based startup, Moovit, for approximately $900 million helped build on Mobileye by using Moovit’s vast and diverse transportation datasets. Shortly after the acquisition of Moovit, the company further announced that they are investing $132 million in 11 AI startups with startups such as Anodot, Astera Labs, Axonne, Hypersonix, KFBIO, Lilt, MemVerge, ProPlus Electronics, Retrace, Spectrum Materials and Xsight Labs. With most of these startups working in the self-driving industry, cloud computing, neural machine translation space, Intel sees a great opportunity with this expensive deal. One of its most significant and most important acquisitions was that of Habana Labs last year, which was acquired for approximately $2 billion. Pioneering in deep-learning accelerators for data centres, Habana Labs boosted Intel’s AI portfolio and accelerated its efforts in the AI silicon market, which Intel expects to be greater than $25 billion by 2024. With its Gaudi AI training processor and Goya AI Inference Processor, Intel vouches to take its AI processor game to the next level. These acquisitions suggest that Intel has strongly shifted its focus to AI while focusing on building software and designing chips that can work together. Intel reported $3.8 billion in AI-driven revenue in 2019, which it further expects to increase with deals such as SigOpt and Cnvrg. Having said that, the focus on AI chips will undoubtedly be a core component of Intel’s business strategy in the coming years. Apart from building AI software stack, it will give tough competition to the likes of Qualcomm, Marvell, and AMD in the AI chips space.","excerpt":"Intel, in its latest acquisition spree, has acquired Israel-based Cnvrg.io. The deal, like most of the deals in the past, is aimed at strengthening its machine learning and AI operations. The 2016-founded startup provides a platform for data scientists to build and run machine learning models that can be used to train, run comparisons and […]","categories":["Global Tech"],"tags":["Intel"],"author_name":"Srishti Deoras","publish_date":"2020-11-07T16:00:24","publication_year":"2020","word_count":846,"keywords":["Go","machine learning","AI","cloud computing","computer vision","RAG","Aim","deep learning","R","Intel","Databricks"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","Aim","RAG","cloud computing","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/intel-acquires-cnvrg-io-to-lead-the-race-in-the-ai-game\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10007098,"title":"How Can MLflow Add Value To Machine Learning Lifecycle And Model Management","content":"One of the major concerns around machine learning is deploying it. Running a large number of deployment tools and environments, and migrating a model to a production environment can be extremely challenging. There are countless independent tools from data preparation to model training, and software tools that cover every stage of the machine learning life cycle. Machine learning developers need to use and deploy dozens of libraries while in a production environment. There is no standard way to migrate models from any library to any of these tools, so that every time a new deployment is made, new risks are created. What Are The Challenges With ML Workflow? The experimental results are difficult to reproduce. Algorithm scripts are difficult to run repeatedly for many reasons, such as code version, past parameters, and operating environment. Without detailed tracking, the team often encounters difficulties in using the same code to achieve the same effect. Whether you are a data scientist delivering training code to engineers for production or rolling back to the code to fix a bug, the steps to reproduce the machine learning workflow are critical. We have heard many horror stories, such as the model performance of the production environment being not as good as the training model or one team can not reproduce the results of another team. Whether you work alone or in a team, it isn’t easy to track which parameters, codes, and data in each experiment create a certain model. These workflow challenges surrounding the ML lifecycle are usually the biggest obstacle to using machine learning in a production environment and scaling it within an organisation. To meet these challenges, many companies have begun to build internal machine learning platforms that can automate some of these steps. For example, Uber and Facebook have established Michelangelo and FBLearner Flow to manage data preparation, model training and deployment. However, even the internal platform has its limitations: a typical machine learning platform only supports a small group of limited customisation (no matter what the engineering team has built), and the platform is bound to each company’s infrastructure. Why MLflow Comes In? At the 2018 Spark + AI Summit, Databricks introduced MLflow, which is a new open-source project that can build an open machine learning platform. In addition to being open-source, MLflow is also open. In a sense that anyone in the organisation or open source community can add new features to MLflow (such as a new training algorithm or a new deployment tool). These functions can automatically cooperate with other parts of MLflow. MLflow provides a powerful way to simplify and linearly expand the deployment of machine learning within the organisation by tracking, reproducing, managing and deploying models in software development. What Problem Does MLflow Solve? Machine learning is not a one-way pipeline, but an iterative loop. It includes four parts: data preprocessing, model training, model deployment and data update. Among them, the preprocessing and model training involves the adjustment of parameters while the entire ML process involves cooperation between links. There is a lot of communication work, code rewriting and environment configuration that undergoes in a machine learning process. To solve the coordination problems between various links, MLflow proposed the two concepts of MLflow Project and MLflow Model, both of which define a set of convention standards, as long as your project or model follows this set Configuration. MLflow can be used to perform one-click project reproduction and model deployment functions, which is equivalent to the one-click online deployment. There is no need to rewrite the code in the project and no environment configuration. How Is It Designed? MLflow is designed to solve these workflow challenges through a set of APIs and tools, and you can use them with any existing machine learning libraries and code repositories. In the current alpha version, MLflow provides three main components: MLflow tracking module: The MLflow tracking module is an API and UI that is used to record parameters, code versions, performance evaluations and output files when executing machine learning codes so that they can be visualised in the future. By using a few simple lines of code, you can track parameters, performance indicators and “artifacts”. MLflow project module: It is a code packaging format for reproducible operations. By encapsulating your code in an MLflow project, you can specify the dependencies and allow any other users to rerun it later and reliably reproduce the results. MLflow model module: It is a simple model packaging format that allows you to deploy the model to many tools. For example, if you can encapsulate the model as a Python function, the MLflow model can be deployed to Docker or Azure ML for online services, to Apache Spark for batch scoring, and so on. MLFlow On Azure Databricks The MLflow community has been growing fast, with hundreds of contributors from many companies having contributed code to the open-source project. For example, because the project is part of Databricks, Microsoft uses it on the Azure platform. Azure Databricks implements a fully managed and hosted version of MLflow, and other Azure Databricks workspace features like experiment and runs management and notebook revision capture. MLflow on Azure Databricks extends an integrated experience for tracking and securing machine learning model training and running ML projects. MLflow’s tracking URI and logging API, together known as MLflow tracking, can be used to connect MLflow experiments and Azure Machine Learning. Doing so enables users to track and log experiment metrics and artifacts in Azure Machine Learning workspace. Users can deploy their MLflow experiments as an Azure Machine Learning web service. By deploying as a web service, they can apply the Azure Machine Learning monitoring and data drift detection functionalities to their production models. Wrapping Up Although the various components of MLflow are simple, whether you are using machine learning alone or collaborating with people in a large team, you can combine them in powerful ways. For example, when developing the model on your laptop, using MLflow you can record and visualise the code, data, parameters and performance indicators. You can encapsulate the codes as MLflow projects to run them on a large scale in a cloud environment for hyperparameter search. You can also share algorithms, feature extraction steps and models as MLflow projects or MLflow models. Finally, you can deploy the same model to batch and real-time processing without the need to develop separate code for two different tools. MLflow is open-source and can be easily installed using pip install MLflow. To start using MLflow, follow the instructions in the MLflow documentation, or view the code on GitHub.","excerpt":"One of the major concerns around machine learning is deploying it. Running a large number of deployment tools and environments, and migrating a model to a production environment can be extremely challenging.  There are countless independent tools from data preparation to model training, and software tools that cover every stage of the machine learning life […]","categories":["AI Features"],"tags":["Azure Machine Learning","machine learning software"],"author_name":"Vishal Chawla","publish_date":"2020-09-12T13:00:00","publication_year":"2020","word_count":1095,"keywords":["machine learning","TPU","AI","ML","docker","Apache Spark","Azure Machine Learning","machine learning software","MLflow","Azure","Azure ML","Databricks"],"extracted_tech_keywords":["AI","machine learning","ML","MLflow","Azure ML","Azure","docker","TPU","Apache Spark","Databricks"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-can-mlflow-add-value-to-machine-learning-lifecycle-and-model-management\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":35553,"title":"Hackathons Are Popular As Competitions, Their Contribution To Hiring Is Still Modest: Haragopal M, IIM Bangalore","content":"Hackathons have been in vogue for quite some time in the tech world to source innovative ideas. Many employers have also used this medium to brand themselves well with their prospective employees. For the participants, hackathons address one or many of these motivations – prize money, bragging rights, learning opportunities and\/or hires. Quite often the hiring as a motive is understated in many hackathons. We spoke to Haragopal Mangipudi of IIM to know his views on Hackathon. Mangipudi did his Post Graduate Programme in Software Enterprise Management at IIM Bangalore in 2001 and is a recipient of IIMB’s Distinguished Alumnus Award in 2016. He is currently the CEO and MD of finUNO and also the founder of guNaka. He is also Adjunct Faculty at IIMB, where he teaches Software Product Management. AIM: How hackathons have become a crucial part of the hiring process in the data science industry HM: Data is the oxygen in this space and to ideate or create compelling prototypes, access to the right data sets is a must. And to give such access in a controlled yet meaningful way, a hackathon could be a right stage. Data science is a team sport and it would be ideal to have a hiring platform that helps in hiring a complete team of complementary skills and experience than individual candidates. Hackathons these days do invite senior data scientist profiles to mentor and judge these hackathon events. AIM: Have hackathons replaced the traditional ways of hiring? HM: While hackathons have been popular both with the sponsors and participants, their contribution to the total hiring is still a small number. This number is higher in the valley and it’s getting better in India as well. Communities and platforms like Kaggle tailored for this space have been very successful in addition to the broader platforms like HackerRank,  Hackathon, MachineHack. AIM: What is the main purpose of conducting a hackathon for educational institutions? HM: Hackathons is a good idea to simulate the business context in an academic environment as well to give a closer to real life example. Many of the communities do support this in an academic environment as well. IIMB being known for both technology and management skills is better suited to hold hackathons. AIM: How important are hackathons (both internal and external) to boost innovation HM: Internal Hackathon as an unconstrained “platform for innovation” is a good idea to source disruptive ideas from hitherto unknown corners of the organization. Most often employees are inhibited by the organizational structure and their roles to come up with innovations. Hackathons break those departmental and hierarchical barriers to source productive and inspiring ideas and creations from the internal stakeholders as well. AIM: Any more ideas for productive and engaging hackathons? Gamification is a good idea for the hackathons to increase the challenge-level to push the participants to their fullest potential for hiring decisions Customer\/User hackathons are also useful to source some good profiles from the customer industry\/domain","excerpt":"Hackathons have been in vogue for quite some time in the tech world to source innovative ideas. Many employers have also used this medium to brand themselves well with their prospective employees. For the participants, hackathons address one or many of these motivations – prize money, bragging rights, learning opportunities and\/or hires. Quite often the […]","categories":["AI Features"],"tags":["Data Science Hiring","Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2019-02-28T10:01:51","publication_year":"2019","word_count":493,"keywords":["data science","Go","programming_languages:R","AI","innovation","RAG","Data Science Hiring","Aim","ViT","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","Aim","RAG","R","Go","GAN","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hackathons-are-popular-as-competitions-their-contribution-to-hiring-is-still-modest-haragopal-m-iim-bangalore\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161988,"title":"Perplexity CEO is ‘Ready to Invest $1 Million, and 5 Hours\/Week’ to Make India Great Again in AI","content":"Perplexity CEO Aravind Srinivas, in a post on X, expressed his support for the development of AI in India. “I am ready to invest $1 mn personally and 5 hours per week of my time into the most qualified group of people that can do this right now to make India great again in the context of AI,” he said. Srinivas was referring to a previous X post by him in which he assured help to anyone willing to run a “DeepSeek-like company in India and open-source the models”. Notably, he means it with all seriousness. “Consider this as a commitment that cannot be backtracked,” he affirmed. In order to make the cut, he said the team must be “cracked and obsessed” like the Chinese research lab DeepSeek, whose recent models have shaken up the AI world with unprecedented efficiency. Moreover, he added that he would invest an additional $10 million if the team aims to beat the latest DeepSeek-R1 model on all benchmarks with rigour. Srinivas believes India should build a foundation AI model of their own instead of building on top of existing open-source models. “India must show the world that it’s capable of ISRO-like feat for AI,” he earlier said. Expressing a disagreement with Infosys co-founder Nandan Nilekani, Srinivas said the former is “wrong in pushing Indians to ignore model training skills and focus on building on top of existing models”. “Essential to do both,” Srinivas argued in a post on X. Srinivas is an Indian citizen and is an alumnus of the Indian Institute of Technology, Madras. Recently, he also announced plans to recruit talent to expand the company’s presence in India. “I am looking to recruit someone to work together on growing Perplexity in India. It will be fun and intense. You must be based in India and willing to travel and meet with strategic partners and institutions – scrappy startup mode,” Srinivas said in a LinkedIn post. Last year in December, Srinivas met with Prime Minister Narendra Modi in Delhi to discuss AI’s potential in India. Following the meeting, Srinivas posted on X, “We had a great conversation about the potential for AI adoption in India and across the world. Really inspired by Modi ji’s dedication to staying updated on the topic and his remarkable vision for the future.” “Was great to meet you and discuss AI, its uses and its evolution. Good to see you doing great work with Perplexity AI. Wish you all the best for your future endeavours,” PM Modi responded.","excerpt":"“And beating DeepSeek R1 on all benchmarks with rigour will mean I will invest $10 million more,” Aravind Srinivas said.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Perplexity"],"author_name":"Supreeth Koundinya","publish_date":"2025-01-22T15:59:37","publication_year":"2025","word_count":420,"keywords":["Go","Perplexity","programming_languages:R","AI","programming_languages:Go","Aim","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/perplexity-ceo-is-ready-to-invest-1-million-and-5-hours-week-to-make-india-great-again-in-ai\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167369,"title":"GitHub Copilot Adds Vibe Coding Abilities","content":"GitHub Copilot recently announced a significant update with the rollout of agent mode with Model Context Protocol (MCP) support to all Visual Studio Code users. This new mode, coupled with the integration of the MCP, aims to provide a more proactive and context-aware coding assistance experience. The agent mode represents a shift from reactive code completion and multi-file editing towards a more action-oriented approach. It is designed to interpret user prompts and execute necessary subtasks to achieve a desired outcome. It seems like Microsoft is embracing the concept of ‘vibe coding’, which is better late than never. This can include identifying and generating relevant files, suggesting and requesting the execution of terminal commands or tool calls, and analysing runtime errors with self-correction capabilities. Support for the MCP was added to enhance agent mode’s capabilities. MCP allows agent mode to use external tools and information sources, providing a broader task context. GitHub also released a local, open-source MCP server, enabling the integration of GitHub functionalities into other LLM-powered tools that support the protocol. With the agent mode, users can choose the model they want, including Claude 3.5 and Claude 3.7 Sonnet, Google Gemini 2.0 Flash, and OpenAI GPT-4o, which are included with all paid Copilot subscriptions. GitHub Copilot is introducing premium requests for newly available models used in chat, multi-file edits, and agent mode, supplementing the unlimited access to the base model for paid users. Starting May, existing Copilot Pro, Business, and Enterprise subscribers will receive a monthly allocation of these premium requests, with unlimited access until then. A new Pro+ plan for individuals offers a higher allocation for $39 per month. All paid users will also be able to purchase additional premium requests beyond their monthly allowance, with administrators able to manage these settings.","excerpt":"With the agent mode, users can choose the model they want, including Claude 3.5 and Claude 3.7 Sonnet, Google Gemini 2.0 Flash, and OpenAI GPT-4o.","categories":["AI News"],"tags":["GitHub","Github Copilot","Vibe Coding"],"author_name":"Ankush Das","publish_date":"2025-04-07T15:24:25","publication_year":"2025","word_count":294,"keywords":["Go","Gemini 2.0","OpenAI","AI","GPT-4o","Github Copilot","Claude 3.5","Git","Vibe Coding","Aim","GitHub","R"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","Claude 3.5","Gemini 2.0","Aim","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/github-copilot-adds-vibe-coding-abilities\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166258,"title":"Perplexity Partners with SoftBank to Launch Enterprise Pro in Japan","content":"AI search engine Perplexity has partnered with Japanese investment firm SoftBank to launch Perplexity Enterprise Pro for corporate customers in Japan. With this, SoftBank becomes the first authorised reseller, helping expand Perplexity’s reach in Japan’s corporate market. According to the company’s blog post, the partnership utilises SoftBank’s 7,000-member enterprise sales team to target Japanese companies. SoftBank’s established presence provides an opportunity to penetrate one of the world’s largest economies. SoftBank has already tested the product internally, confirming its ability to improve productivity for Japanese enterprises. This collaboration builds on an existing alliance between Perplexity, SoftBank, Y!Mobile, and LINEMO, which began in June 2024 with a free trial of Perplexity Pro for individual customers. SoftBank’s relationships across various industries, including finance, manufacturing, healthcare, and technology, help Perplexity overcome market entry barriers. With Perplexity Enterprise Pro, Japanese businesses can leverage AI-driven search to transform how teams access and use information. They will join over 7,000 global organisations using the platform. With 15 million users, Perplexity has been on a shipping spree recently. The company released a Windows app, added MCP support for real-time data. Perplexity’s latest release, the Perplexity Windows app, is now available. Meanwhile, Deutsche Telekom, the parent company of T-Mobile, has partnered with Perplexity AI to create a next-generation AI Phone. Running on a custom Magenta AI operating system, it will feature Perplexity Assistant. The company also open-sourced R1 1776, a version of the DeepSeek-R1 language model that has been post-trained to eliminate censorship and provide factual responses.","excerpt":"AI search engine Perplexity has partnered with Japanese investment firm SoftBank to launch Perplexity Enterprise Pro for corporate customers in Japan.  With this, SoftBank becomes the first authorised reseller, helping expand Perplexity’s reach in Japan’s corporate market. According to the company’s blog post, the partnership utilises SoftBank’s 7,000-member enterprise sales team to target Japanese companies. […]","categories":["AI News"],"tags":["Perplexity AI"],"author_name":"Aditi Suresh","publish_date":"2025-03-18T20:32:48","publication_year":"2025","word_count":248,"keywords":["programming_languages:R","AI","Perplexity AI","RAG","ViT","GAN","R"],"extracted_tech_keywords":["AI","RAG","R","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/perplexity-partners-with-softbank-to-launch-enterprise-pro-in-japan\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":38603,"title":"Why Is It Important To Make Your Neural Networks Compact","content":"Image credits: Pete Warden’s blog AI has already taken up a lot of space in the industries. Even if the general population thinks the technology is not very prominent, it is taking up a lot of applications in the commercial world. We do not even certain applications that could be a result of AI but they are in everyday lives. With a rise in the technology, it has become extremely essential for neural networks, the technology largely responsible for forecasting using data, also one of the largest applications, to be appropriately quantized. Importance Of Quantization Today, the neural networks are deployed to a plethora of application and need a large amount of data for their accurate working. Neural networks have become the state-of-the-art approach for many large-scale computer vision and sequence modeling problems. Deep convolutional networks dominate the leaderboards for popular image classification and object detection datasets such as ImageNet and Microsoft COCO. They naturally need hundreds of megabytes of memory storage for their trainable floating-point parameters, and billions of floating-point operations to make a single inference. In order to achieve large memory many efforts are made to quantize the neural networks, also maintaining the performance. Quantizing neural networks dates back to the 1990s. The initial motivation of quantization neural network was to make it easier for the digital hardware implementation. The recent importance and research of quantization of neural networks, however, has emerged due to the success of neural network applications. Deep learning has been proven to be powerful on tasks including image classification, objection detection, natural language processing, and many other areas. For a more accurate prediction and with deeper networks, the memory size of the network becomes a problem. As more smartphones have begun to include these neural networks, these networks being deeper and the memory size being larger becomes a problem to smartphones as well. These phones are generally equipped with 4GB of memory and is expected to be able to support multiple applications at one time. Neural network models that are large in size make phones short of memory since they occupy at least 1GB of memory with about three or more models being run. The model size is not only a memory usage problem, it is also a memory bandwidth problem since the weight of the models is walked every time for each prediction and image related applications usually need to process data in real time. All this accounts for at least 30 FPS. Large memory bandwidth is required for a simple model. Memory, CPU and battery burn the device when the network is running. All of these challenges make the quantization of neural networks a necessity. How Can Neural Networks Be Quantized? The goal of quantization is to compact the models without that having any effect on the performance. This will need to have machine learning, computer architecture, and suitable hardware design. There are three components that can be quantized in a neural network: weights, activations, and gradients. The motivation and methods to quantize these components are different from each other. There are various quantization of neural network methods, but they can broadly be classified into two categories of deterministic quantization and stochastic quantization. A proper quantization technique selection is important. 1.Deterministic quantization: In deterministic quantization, there is a one-to-one mapping between the quantized value and the real value. This method of quantization can specify the appropriate quantization levels in advance to run on dedicated hardware. That is why they should generally be preferred if one wants to quantize for hardware acceleration, giving a greater hardware performance. 2.Stochastic quantization:  In stochastic quantization, the weights, activations or gradients are discretely distributed. The quantized value is sampled from the discrete distributions. In this approach, the weights are assumed to be discretely distributed and a learning algorithm is used to infer the parameters of the distributions. It has the quantized weights more interpretable than in deterministic quantization. The distributions of the weights can be understood and gain more insights into how the network works. Conclusion Neural networks today are gaining huge popularity and their applications are very powerful in the sphere of machine learning. They have over the years largely grown to solve complex world problems. A small portion of the benefits achieved when using lower precision can be forfeited to increase the network size and therefore the accuracy. Given the limitations in power budgets dedicated to these networks, the importance of low-power and low-memory solutions are important to emphasize. There has been a lot of research going on in recent times to overcome this challenge.","excerpt":"AI has already taken up a lot of space in the industries. Even if the general population thinks the technology is not very prominent, it is taking up a lot of applications in the commercial world. We do not even certain applications that could be a result of AI but they are in everyday lives. […]","categories":["AI Features"],"tags":["memory","Neural Network","quantization","smartphones"],"author_name":"Disha Misal","publish_date":"2019-05-03T12:13:28","publication_year":"2019","word_count":760,"keywords":["Neural Network","Go","machine learning","quantization","AI","neural network","smartphones","computer vision","RAG","CuPy","deep learning","memory","object detection","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","CuPy","RAG","object detection","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-it-important-to-make-your-neural-networks-compact\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21107,"title":"Is Deep Learning Going To Be Illegal In Europe?","content":"In a matter of months, General Data Protection Regulation (GDPR) will become a law throughout Europe, deeming a complete overhaul in the way artificial intelligence techniques are used in business settings. By May 25, the GDPR will become fully enforceable throughout the European Union, states the EU GDPR timeline. The coming deadline, which will be enforced in the next 100 days, has sparked a debate among the AI research community and tech giants who are now scrambling to meet the EU’s data privacy and algorithmic fairness guidelines. Well, for the EU citizens, GDPR has strengthened their rights by ushering in a new era by unifying data protection rules and placing new obligations on tech enterprises on the process of collecting personal user data. The forthcoming regulations have firmly divided Europe into two different camps – a) one that welcomes the need for data privacy and algorithmic fairness in society, b) tech giants who are bristling at the thought of new challenges, such as asking for user consent in simpler terms and tackling the black box problem of AI, which would make eventually make it illegal, with fines imposed to the tune of 4 percent of global turnover, reportedly. GDPR Highlights in a Nutshell First up, let’s shine a spotlight on some of the highlights of GDPR: Regulation on collecting data from EU citizens by companies, big & small: This rule isn’t just limited to companies which are headquartered in EU but extends to all organizations that have data from EU citizens. From rethinking the size of text in terms and conditions to explaining how a company uses personal data to sell adverts, the GDPR enforces companies to follow Privacy by Design principles. Data Portability: The regulation states that the subject can demand his\/her personal data to be transferred directly to a new provider, without hindrance, that too in a machine-readable format. This is akin to switching mobile provider or switching social networks without losing any data. For companies like Google, Facebook which are veritable data mines and even smaller data science start-ups, this sounds like a death knell and a mass exodus of data when users leave a company. Right to be forgotten\/Right to erasure: As emphasized in Article 17 of GDPR, every data subject shall have the right to obtain from the controller the erasure of personal data concerning him or her without undue delay and the controller shall have the obligation to do that. Again, this will be a huge loss for tech behemoths who collect data in the form of cookies and reap gains from running targeted ads. Algorithmic Fairness: The Right to Explanation of Automated Decision mandates that the data subject has a right to get an explanation about decisions made by algorithms and a right to opt-out of some algorithmic decisions altogether if they are not satisfied with it. For example, if an applicant is refused a loan based on an automated decision, they have a right to seek an explanation. For tech companies, it is deemed as a severely harmful restriction on artificial intelligence and will exponentially slow the development of AI technology, known for its universal accuracy. High Performance vs Poor Explainability Conundrum of AI We are not going to delve into the nitty-gritty of EU guidelines but are going to shine the spotlight on the biggest criticism of the most widely used techniques of Artificial Intelligence – Deep Learning and its un-interpretability problem or let’s say the black box problem. This will make it virtually impossible for any company to do AI and even make it illegal. AI experts and tech companies who profit from data are crying hoarse about the unfeasibility of explaining algorithmic decisions because of the architecture of artificial neural networks which makes it hard to decipher how the output was generated. Well-known academician Dr. PK Viswanathan, Program Director at Great Lakes Institute of Management took a shot at demystifying the black box problem of artificial neural networks at Cypher 2017. According to Dr. Viswanathan, the wide perception is that Neural Network is a black box, but it is not completely a black box and there are a few ways the output can be explained. Citing an example, he said that the word artificial is important and it is a strong contender in the world for its prediction and classification problems. Unlike logistic regression and other supervised techniques such as random forest, which are statistically oriented, Neural Network is non-parametric, non-linear complex relationship model-building exercise which is considered universal approximation. Neural Networks are used in all classification problem and prediction problems. Some of the common applications of neural networks are — buyer vs non buyer in marketing and classification of risk using neural network. Topology of Artificial Neural Network For example, let’s talk about a Multilayer Perceptron with Two Hidden Layers which is a class of feed-forward artificial neural network. Multilayer Perceptron with Two Hidden Layers is well-known for its predictive accuracy. In Multilayer Perceptron with Two Hidden Layers, you have two input neurons, two hidden layers and four nodes and then the output comes. Initially, one starts by giving some weights, bias, numbers and the activation function which could be a sigmoid function — a logistic function and you try to change the weights with a feed forward method. The feeds are changed every time in a recursive way and finally you get the output. Now, in the architecture, the hidden layer is closely associated with the black box point –and how the neural network is learning the training set. This is where the major criticism steps in. And every now and then a weight is changed, one applies a rule that minimizes some of the squares of error, and the iteration continue. This is known as the black box conundrum, it can approximate any function but it gives no insight into the relation between predicted variables and the outcome, explains Dr. PK Viswanathan. Now, in a supervised learning problem, one can explain the precise relationship between y and x but it is not possible to capture in a Neural Network. The biggest criticism of Neural Network is: It lacks explanatory power and we can’t define what is happening inside the hidden layers. But Neural Networks score high on universal approximation and accuracy. Very difficult in the practical world to interpret the synoptic weights which is not the case in traditional techniques which further adds to the black box puzzle Solving AI’s Black Box Conundrum Now, many researchers are already working on explaining how neural networks make decisions. Let’s cite a couple of ways: LIME: Better known as LIME, Local Interpretable Model-Agnostic Explanations, this technique involves manipulating data variables in many ways to see what moves score the most.  In LIME, local refers to local fidelity – i.e., the explanation should reflect the behaviour of the classifier “around” the instance being predicted. This explanation is useless unless it is interpretable – that is, unless a human can make sense of it. Lime is able to explain any model without needing to ‘peak’ into it, so it is model-agnostic. You can see the research paper here. DARPA’s Explainable AI:  Now, DARPA has created a suite of machine learning techniques that produce more explainable models, while maintaining a high level of learning performance. Dubbed as Explainable AI (XAI), it enables human users to understand and manage the upcoming AI partners. The main advantage of Explainable AI is that new techniques can potentially circumvent the need for an extra layer. Another explanation component could from training neural network to associate semantic attributes with hidden layer nodes – which could boost learning of explainable features.","excerpt":"In a matter of months, General Data Protection Regulation (GDPR) will become a law throughout Europe, deeming a complete overhaul in the way artificial intelligence techniques are used in business settings. By May 25, the GDPR will become fully enforceable throughout the European Union, states the EU GDPR timeline. The coming deadline, which will be […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-01-30T11:56:19","publication_year":"2018","word_count":1273,"keywords":["data science","machine learning","artificial intelligence","TPU","AI","neural network","R","ML","deep learning","xAI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","data science","xAI","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/deep-learning-going-illegal-europe\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":22698,"title":"Wipro Invests $2.02 Million In US-Based AI Startup Avaamo","content":"Wipro said it has invested $2.02 million in Avaamo Inc, an artificial intelligence-based conversational computing platform, Avaamo inc. “Investment of $2.02 million in this tranche through a fresh infusion of funds into Avaamo,” the company said in a BSE filing. Reportedly, the tech giant will make an additional investment through conversion of convertible notes issued earlier, taking the total investment to $3.02 million. India’s third largest software service exporter said that the cash deal was completed on March 14, 2018. The deal will give Wipro a minority stake of less than 20 percent on a diluted basis. Avaamo is a Los Altos-based startup founded by Ram Menon and Sriram Chakravarthy in 2014. The early-stage company provides artificial intelligence-driven enterprise bot platform that simplifies the time needed to design and deploy enterprise bots to their customers and to corporate employees. In India, Avaamo has an office in Bengaluru. In pic: Ram Menon (Right) and Sriram Chakravarthy (Left) The company, through its investment arm Wipro Ventures, primarily invests in early-to-mid-stage startups in emerging technologies in US, India and Israel. Wipro Ventures has got a corpus of $100 million and is headed by Rishad Premji. The key focus areas of the investments are AI, big data, IoT, FinTech, healthcare IT etc. Earlier in November 2017, Wipro Ventures picked a minority stake in California-based app testing platform HeadSpin Inc for an undisclosed amount. It has also invested in US-based cybersecurity company Vectra Networks to deliver automated ‘Threat Hunting as-a-service’. On February 2016, it has picked up a minority stake in Pune-based big data startup Altizon for Rs 9.78 crore. The tech company has also invested in robotics startup Vicarious and risk and fraud prevention firm Emailage Corporation.","excerpt":"Wipro said it has invested $2.02 million in Avaamo Inc, an artificial intelligence-based conversational computing platform, Avaamo inc. “Investment of $2.02 million in this tranche through a fresh infusion of funds into Avaamo,” the company said in a BSE filing. Reportedly, the tech giant will make an additional investment through conversion of convertible notes issued […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Wipro"],"author_name":"Smita Sinha","publish_date":"2018-03-16T07:42:48","publication_year":"2018","word_count":283,"keywords":["big data","Wipro","Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","ai_applications:robotics","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","big data","startup","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-invests-2-02-million-in-us-based-ai-startup-avaamo-for-20-stake\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001540,"title":"Can Tesla’s Speculated Indian Debut Drive A Shift Towards Electric Mobility In The Country?","content":"In November last year Elon Musk, the founder of Tesla, one of the world’s leading electric car makers,  replied to a user’s Tweet that the company is likely to hit the Indian market following their huge profit margin in the previous quarter. Musk said that the company would soon consider expansion of its work to attain a partial presence in Asia, South America, and Africa by the end of this year.  Presently its market share is limited to North America and to certain South-Asin countries like China, Japan and Hong Kong where the company have their regional office. Switching to online store On March 18, Musk further reiterated his interest to enter India market in yet another Tweet and reassured Twitteratis that he would love to enter the Indian market this year and if not for this year, the company will certainly do it within the next year. “Would love to be there this year. If not, definitely next!”, Musk tweeted Though Tesla does not have regional offices in India,  their Tesla Model 3, a medium-size premium sedan priced at ₹ 75 L,  is likely to be available in India from June this year, a leading website notes. Building more optimism for its users, on March 4, in a dramatic turn of events, the company announced that it is planning to shut down its physical store and will sell its Model 3 via online store in order to cut down on the expenditure. The company said that the model will be available online and that doorstep delivery would be made for its customers. “It’s 2019. People want to buy online,”  Elon Musk, the CEO of Tesla said post its announcement to go online. This decision by the company brought cheers among auto-enthusiasts in the country and some strongly feel the car will be available to be purchased through its online store and its affordability is another advantage. While the standard base model for Model 3 is $35,000, the company will also offer the model at $37,000 with a premium interior and better acceleration, thus making it the cheapest ever model to be available of the high-end cars. However, in a dramatic turn of events, Tesla recently reversed its shut down its stores stating that the new plan will lead to a 3 per cent hike in the cost of its vehicles. “As a result of keeping significantly more stores open, Tesla will have to raise vehicles price by about three per cent on average worldwide,” it said in a statement. “In other words, we will only close about half as many stores, but the cost savings are therefore only about half,” it added. However, it stated that the new Model 3 will not be affected by this variation. Will it be a reality? Earlier, Musk has been critical about government for its stringent rules thus restricting the company’s entry to the Indian market. As a result of this, the company closed a deal of $5 billion factory in China, making its first-ever manufacturing facility outside of its headquarters in India. In addition, the present ruling government has been trying to woo the company to set up its unit in India. Following the losing of the bid to China for its manufacturing unit, the Union Minister for Road Transport & Highways of said to a leading newspaper, “If they are coming, if they are ready to come, we will welcome, we are ready to offer them land and all type of help.  I requested them, I met them in America, but their first point was China at that time,” Gadkari said after India lost the bid despite months-long deliberations with the company. With the government pushing for Make in India, it has been taking deliberate steps to do away with bureaucratic red-tape that has been stalling the entering of these MNC in India. Furthermore, as the auto market is forecasted to grow by leaps and bounds and demand for electric vehicles on the rise in India, Tesla is likely to face competition from regional players.  Key E-vehicle makers like Mahindra and Tata have witnessed success on a smaller scale across the country. Though Tesla has to its advantage possessing its state-of-the-art technology and customisation, the cost attached to each car can possess a problem.","excerpt":"In November last year Elon Musk, the founder of Tesla, one of the world’s leading electric car makers,  replied to a user’s Tweet that the company is likely to hit the Indian market following their huge profit margin in the previous quarter. Musk said that the company would soon consider expansion of its work to […]","categories":["AI Features"],"tags":["autonomous cars India","Elon Musk","Tesla"],"author_name":"Akshaya Asokan","publish_date":"2019-03-19T13:35:00","publication_year":"2019","word_count":713,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","Elon Musk","Tesla","R","autonomous cars India"],"extracted_tech_keywords":["AI","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-teslas-speculated-indian-debut-drive-a-shift-towards-electric-mobility-in-the-country\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121355,"title":"Niantic Uses Meta Llama Models To Generate Creative Behaviours in Latest Mobile Game Peridot","content":"Niantic Inc. is using Meta Llama models to generate creative behaviour in real time as players interact with creatures in the company’s latest AR game, Peridot. The game uses Meta’s Llama 2 to enhance interactions with the lifelike virtual pets called Dots that exhibit smart, unpredictable behaviours akin to real animals. By leveraging Llama 2, Dots can now react dynamically to their surroundings in real time, fostering deeper connections and a heightened sense of companionship for players. Players can have “conversations” with their Dots, and the creatures will respond with surprising and organic behaviours, such as expressing joy, curiosity, or mischief, bringing a sense of realism to the virtual pet experience. “We are eager to see more models open-sourced to enable teams like ours to freely explore their capabilities without being caught in discussions around cost, privacy, and cloud dependencies,” said Niantic Global Product Marketing Lead Asim Ahmed. Implementing Llama 2 presented challenges in creating prompts that strike the right balance between expressiveness, creativity, and response time. The team addressed this by defining an expected response format in JSON, instantly improving the quality of the AI’s responses. Niantic plans to continue pushing the boundaries of generative AI in Peridot and explore new ways to elevate player interactions across devices. “Peridot’s success with generative AI gives us a glimpse into what’s possible, and we plan to elevate the way players interact with Peridot across devices,” said Ahmed. As they have done before with pioneering AR games like Pokémon GO, they are now innovating with virtual pets in their latest mobile AR adventure. Looking ahead, the company envisions a wide range of opportunities to leverage AI, such as Llama 2, to drive new areas of gameplay more procedurally. This opens up opportunities for the gaming industry to similarly leverage Meta Llama.","excerpt":"“We are eager to see more models open-sourced to enable teams like ours to freely explore their capabilities without being caught in discussions around cost, privacy, and cloud dependencies,” said Asim Ahmed.","categories":["AI News"],"tags":["AI in Gaming"],"author_name":"Gopika Raj","publish_date":"2024-05-23T15:09:27","publication_year":"2024","word_count":299,"keywords":["Go","programming_languages:R","AI","AI in Gaming","programming_languages:Go","RAG","ViT","generative AI","GAN","llm_models:Llama","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","GAN","ViT","llm_models:Llama","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/niantic-uses-meta-llama-models-to-generate-creative-behaviours-in-latest-mobile-game-peridot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10021792,"title":"Redis Labs Announces General Availability for Integrated Enterprise Tiers of Azure Cache","content":"Recently, Redis Labs announced the general availability of Redis Enterprise-powered tiers on Azure Cache for Redis at the Microsoft Ignite 2021 event. The availability of the Enterprise tiers is said to enable companies to more effectively optimise for low-latency data access in their most critical applications. The combination of Redis Enterprise and Azure Cache for Redis enables companies to always have access to the latest enterprise-grade Redis functionality, expanded use cases, and enterprise-grade active geo-replication deployment architected for five-nine uptime. It will also offer unique benefits to developers, operators, and ultimately the customer’s bottom line. According to sources, the Enterprise tiers are fully managed by Microsoft and are the result of a year-long collaboration between the companies to provide the most highly available, resilient, scalable, and extensible Redis service to customers. Talking about the new Enterprise tiers, Ofer Bengal, CEO and Co-Founder, Redis Labs stated, “Companies can now effortlessly incorporate the performance and reliability of Redis Enterprise with the breadth and simplicity of consumption Azure offers to serve a variety of low-latency use cases.” Bengal added, “This unique service enables customers to confidently operate Redis at scale with five-nines availability and active geo-redundancy, expanded use cases with Redis modules, and operate at a very attractive cost on Azure.” New Redis functionality in the Enterprise tiers include the following- Up to 99.999% availability, leveraging Redis Enterprise’s active geo-replication technology and Azure’s multiple availability zone deployment capabilitiesFlexibility to build modern applications leveraging Redis Enterprise modules, including RediSearch, RedisTimeSeries, and RedisBloomScale cost-effectively with Redis on Flash’s intelligent memory tiering to ensure the greatest cost savings for large datasetsRedis 6.0––including all native Redis data structures (streams included), and advanced security capabilitiesA seamless, familiar Azure UI for configuration and integrations to Azure security and monitoring servicesConsumption-based pricing––only pay for what’s required, with no minimum commitments Julia Liuson, Corporate Vice President, Developer Division at Microsoft Corp said, “The general availability of Redis Enterprise on Azure Cache for Redis underscores Microsoft’s commitment to work with partners to give developers the tools and features they need to create applications. Enterprises will have access to features like active geo-replication to create applications with unprecedented scale, availability, and performance.” Click here to know more about the Enterprise tiers.","excerpt":"Recently, Redis Labs announced the general availability of Redis Enterprise-powered tiers on Azure Cache for Redis at the Microsoft Ignite 2021 event. The availability of the Enterprise tiers is said to enable companies to more effectively optimise for low-latency data access in their most critical applications. The combination of Redis Enterprise and Azure Cache for […]","categories":["AI News"],"tags":["Azure cloud platform","Microsoft Azure","Redis"],"author_name":"Ambika Choudhury","publish_date":"2021-03-09T17:27:50","publication_year":"2021","word_count":368,"keywords":["cloud_platforms:Azure","programming_languages:R","AI","R","ML","Julia","Scala","Azure cloud platform","RAG","Microsoft Azure","Azure","Redis"],"extracted_tech_keywords":["AI","ML","RAG","Azure","Redis","R","Scala","Julia","cloud_platforms:Azure","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/redis-labs-announces-general-availability-for-integrated-enterprise-tiers-of-azure-cache\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018166,"title":"Jordan Walke, Creator Of ReactJS Walks Away From Facebook","content":"Jordan Walke, who is the creator of ReactJS recently announced in the micro-blogging platform, that after a decade he is leaving Facebook. Walke had been working as a software engineer at the social media giant, where he was working on the React JavaScript library. In a post, Walke tweeted that after spending more than a decade at the social media giant, he is finally leaving the company. However, there lies a good reason for this resignation. All these years, after working for a company, he is now going to start one of his own. However, the details of the company have not been disclosed yet. After ten years, I am leaving Facebook.I am starting a company. pic.twitter.com\/2i9Jq9qwcg— jordwalke (@jordwalke) January 9, 2021 Replying to this tweet, he also mentioned that working at Facebook is an incredible experience for him. Besides this, Walke mentioned that from now on he will also be investing more in startups as opportunities arise as well as investing in open-source projects\/communities including ReactJS related technologies, UI frameworks, programming languages and other such. It has been an incredible experience for which I am grateful.I will also be investing more in startups as opportunities arise, and investing in open source projects\/communities including @reactjs related technologies, @reasonml, UI frameworks, programming languages.— jordwalke (@jordwalke) January 9, 2021 Graduated from the University of Washington, Walke not only created ReactJS but also ReasonML, which is a fast and type-safe code that leverages the JavaScript & OCaml ecosystems. On a concluding note, Walke acknowledged everyone for the outpouring of encouragement and gratitude. He said, “An incredible – no unbelievable – story so far. A new chapter begins. One paragraph at a time.” Thank you everyone for the outpouring of encouragement and gratitude. I assure you that out of everyone, I have gotten the most from of this journey. An incredible – no unbelievable – story so far. A new chapter begins. One paragraph at a time.— jordwalke (@jordwalke) January 9, 2021","excerpt":"Jordan Walke, who is the creator of ReactJS recently announced in the micro-blogging platform, that after a decade he is leaving Facebook. Walke had been working as a software engineer at the social media giant, where he was working on the React JavaScript library. In a post, Walke tweeted that after spending more than a […]","categories":["AI News"],"tags":["jordan walke"],"author_name":"Ambika Choudhury","publish_date":"2021-01-16T11:00:00","publication_year":"2021","word_count":328,"keywords":["Go","programming_languages:R","AI","ML","RAG","JavaScript","jordan walke","R","Java","programming_languages:JavaScript","startup"],"extracted_tech_keywords":["AI","ML","RAG","R","JavaScript","Go","Java","startup","programming_languages:R","programming_languages:JavaScript"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jordan-walke-creator-of-reactjs-walks-away-from-facebook\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10053438,"title":"A Tutorial on Spiking Neural Networks for Beginners","content":"Despite being quite effective in a variety of tasks across industries, deep learning is constantly evolving, proposing new neural network (NN) architectures, deep learning (DL) tasks, and even brand new concepts of the next generation of NNs, such as the Spiking Neural Network (SNN). SNN was introduced by the researchers at Heidelberg University and the University of Bern developing as a fast and energy-efficient technique for computing using spiking neuromorphic substrates. In this article, we will mostly discuss Spiking Neural Network as a variant of neural network. We will also try to understand how is it different from the traditional neural networks.  Below is a list of the important topics to be tackled. Table of Contents What is Spiking Neural Network (SNN)How Does Spiking Neural Network Work?Traditional Neural Network Vs SNNApplication of Spiking Neural NetworksAdvantages and Disadvantages of SNN Let’s start the discussion by understanding what is Spiking Neural Network is. What is Spiking Neural Network (SNN)? Artificial neural networks that closely mimic natural neural networks are known as spiking neural networks (SNNs). In addition to neuronal and synaptic status, SNNs incorporate time into their working model. The idea is that neurons in the SNN do not transmit information at the end of each propagation cycle (as they do in traditional multi-layer perceptron networks), but only when a membrane potential – a neuron’s intrinsic quality related to its membrane electrical charge – reaches a certain value, known as the threshold. The neuron fires when the membrane potential hits the threshold, sending a signal to neighbouring neurons, which increase or decrease their potentials in response to the signal. A spiking neuron model is a neuron model that fires at the moment of threshold crossing. SNN with connections and Biological Neuron Artificial neurons, despite their striking resemblance to biological neurons, do not behave in the same way. Biological and artificial NNs differ fundamentally in the following ways: Structure in generalComputations in the brainIn comparison to the brain, learning is a rule. Alan Hodgkin and Andrew Huxley created the first scientific model of a Spiking Neural Network in 1952. The model characterized the initialization and propagation of action potentials in biological neurons. Biological neurons, on the other hand, do not transfer impulses directly. In order to communicate, chemicals called neurotransmitters must be exchanged in the synaptic gap. How Does Spiking Neural Network Work? Key Concepts What distinguishes a traditional ANN from an SNN is the information propagation approach. SNN aspires to be as close to a biological neural network as feasible. As a result, rather than working with continually changing time values as ANN does, SNN works with discrete events that happen at defined times. SNN takes a set of spikes as input and produces a set of spikes as output (a series of spikes is usually referred to as spike trains). The general idea is as; Each neuron has a value that is equivalent to the electrical potential of biological neurons at any given time.The value of a neuron can change according to its mathematical model; for example, if a neuron gets a spike from an upstream neuron, its value may rise or fall.If a neuron’s value surpasses a certain threshold, the neuron will send a single impulse to each downstream neuron connected to the first one, and the neuron’s value will immediately drop below its average.As a result, the neuron will go through a refractory period similar to that of a biological neuron. The neuron’s value will gradually return to its average over time. Spike Based Neural Codes Artificial spiking neural networks are designed to do neural computation. This necessitates that neural spiking is given meaning: the variables important to the computation must be defined in terms of the spikes with which spiking neurons communicate. A variety of neuronal information encodings have been proposed based on biological knowledge: Binary Coding: Binary coding is an all-or-nothing encoding in which a neuron is either active or inactive within a specific time interval, firing one or more spikes throughout that time frame. The finding that physiological neurons tend to activate when they receive input (a sensory stimulus such as light or external electrical inputs) encouraged this encoding. Individual neurons can benefit from this binary abstraction because they are portrayed as binary units that can only accept two on\/off values. It can also be applied to the interpretation of spike trains from current spiking neural networks, where a binary interpretation of the output spike trains is employed in spike train classification. Rate Coding: Only the rate of spikes in an interval is employed as a metric for the information communicated in rate coding, which is an abstraction from the timed nature of spikes. The fact that physiological neurons fire more frequently for stronger (sensory or artificial) stimuli motivates rate encoding. It can be used at the single-neuron level or in the interpretation of spike trains once more. In the first scenario, neurons are directly described as rate neurons, which convert real-valued input numbers  “rates”  into an output “rate” at each time step. In technical contexts and cognitive research, rate coding has been the concept behind conventional artificial “sigmoidal” neurons. Fully Temporal Codes The encoding of a fully temporal code is dependent on the precise timing of all spikes. Evidence from neuroscience suggests that spike-timing can be incredibly precise and repeatable. Timings are related to a certain (internal or external) event in a fully temporal code (such as the onset of a stimulus or spike of a reference neuron). Latency Coding The timing of spikes is used in latency coding, but not the number of spikes. The latency between a specific (internal or external) event and the first spike is used to encode information. This is based on the finding that significant sensory events cause upstream neurons to spike earlier. This encoding has been employed in both unsupervised and supervised learning approaches, such as SpikeProp and the Chronotron, among others. Information about a stimulus is encoded in the order in which neurons within a group generate their first spikes, which is closely connected to rank-order coding. SNN Architecture Spiking neurons and linking synapses are described by configurable scalar weights in an SNN architecture. The analogue input data is encoded into the spike trains using either a rate-based technique, some sort of temporal coding or population coding as the initial stage in building an SNN. A biological neuron in the brain (and a simulated spiking neuron) gets synaptic inputs from other neurons in the neural network, as previously explained. Both action potential production and network dynamics are present in biological brain networks. Source The network dynamics of artificial SNNs are much simplified as compared to actual biological networks. It is useful in this context to suppose that the modelled spiking neurons have pure threshold dynamics (as opposed to refractoriness, hysteresis, resonance dynamics, or post-inhibitory rebound features). When the membrane potential of postsynaptic neurons reaches a threshold, the activity of presynaptic neurons affects the membrane potential of postsynaptic neurons, resulting in an action potential or spike. Learning Rules in SNN’s Learning is achieved in practically all ANNs, spiking or non-spiking, by altering scalar-valued synaptic weights. Spiking allows for the replication of a form of bio-plausible learning rule that is not possible in non-spiking networks. Many variations of this learning rule have been uncovered by neuroscientists under the umbrella term spike-timing-dependent plasticity (STDP). Its main feature is that the weight (synaptic efficacy) connecting a pre-and post-synaptic neuron is altered based on their relative spike times within tens of millisecond time intervals. The weight adjustment is based on information that is both local to the synapse and local in time. The next subsections cover both unsupervised and supervised learning techniques in SNNs. Unsupervised Learning Data is delivered without a label, and the network receives no feedback on its performance. Detecting and reacting to statistical correlations in data is a common activity. Hebbian learning and its spiking generalizations, such as STDP, are a good example of this. The identification of correlations can be a goal in and of itself, but it can also be utilized to cluster or classify data later on. STDP is defined as a process that strengthens a synaptic weight if the post-synaptic neuron activates soon after the pre-synaptic neuron fires, and weakens it if the post-synaptic neuron fires later. This conventional form of STDP, on the other hand, is merely one of the numerous physiological forms of STDP. Supervised Learning In supervised learning, data (the input) is accompanied by labels (the targets), and the learning device’s purpose is to correlate (classes of) inputs with the target outputs (a mapping or regression between inputs and outputs). An error signal is computed between the target and the actual output and utilized to update the network’s weights. Supervised learning allows us to use the targets to directly update parameters, whereas reinforcement learning just provides us with a generic error signal (“reward”) that reflects how well the system is functioning. In practice, the line between the two types of supervised learning is blurred. Traditional Neural Network Vs SNN A spiking neural network is a two-layered feed-forward network with lateral connections in the second hidden layer that is heterogeneous in nature. To transfer information, biological neurons use brief, sharp voltage increases. Action potentials, spikes, and pulses are all terms used to describe these signals.  Spiking neuron networks are more potent than non-spiking counterparts because they can encode temporal information in their signals, but they also require different and biologically more realistic synaptic plasticity rules. Spikes can’t just hop from one neuron to the next. They must be handled by the neuron’s most complex component: the synapse, which is made up of the axon’s end, a synaptic gap, and the first piece of the dendrite. The synapse was formerly thought to merely transport a signal from the axon to the dendrite; however, it has now been discovered to be a highly complex signal pre-processor that is critical in learning and adaptation. When a spike reaches the synapse’s axonal (presynaptic) side, some vesicles fuse with the cell membrane and release their neurotransmitter content into the extracellular fluid that fills the synaptic gap. Artificial neural networks are a rather old computer science technique; the original ideas and models date back more than fifty years. McCulloch-Pitts threshold neurons were the first generation of artificial neural networks, a conceptually simple model in which a neuron sends a binary ‘high’ signal if the sum of its weighted incoming inputs exceeds a threshold value. Despite the fact that these neurons can only produce digital output, they have been used in sophisticated artificial neural networks such as multi-layer perceptrons and Hopfield nets. A multilayer perceptron with a single hidden layer, for example, can compute any function with a Boolean output; these networks are known as universal for digital computations. Second-generation neurons compute their output signals using a continuous activation function rather than a step- or threshold function, making them appropriate for analogue in- and output. The sigmoid and hyperbolic tangent are two examples of activation functions that are commonly utilized. Application of Spiking Neural Networks In theory, SNNs can be used in the same applications as standard ANNs. SNNs can also stimulate the central nervous systems of biological animals, such as an insect seeking food in an unfamiliar environment. They can be used to examine the operation of biological brain networks due to their realism. Starting with a hypothesis about the topology and function of a real neural circuit, recordings of this circuit can be compared to the output of the appropriate SNN to assess the hypothesis’ plausibility. However, adequate training processes for SNNs are lacking, which can be a hindrance in particular applications, such as computer vision. Advantages and Disadvantages of SNN Advantages SNN is a dynamic system. As a result, it excels in dynamic processes like speech and dynamic picture identification.When an SNN is already working, it can still train.To train an SNN, you simply need to train the output neurons.Traditional ANNs often have more neurons than SNNs; however, SNNs typically have fewer neurons.Because the neurons send impulses rather than a continuous value, SNNs can work incredibly quickly.Because they leverage the temporal presentation of information, SNNs have boosted information processing productivity and noise immunity. Disadvantages SNNs are difficult to train. As of now, there is no learning algorithm built expressly for this task.Building a small SNN is impracticable. Final Words Through this post, we have seen the basic concept related to spiking neural networks and discussed its general working methodology. We covered the various concepts related to it such as neural codes, their architectures, its learning schemes. Lastly, we have discussed the differences between traditional ANN and the SNN and have seen some advantages and disadvantages of SNN. References Spiking Neural NetworkBasic Guide to Spiking Neural NetworkDeep Learning in Spiking Neural Networks","excerpt":"Despite being quite effective in a variety of tasks across industries, deep learning is constantly evolving, proposing new neural network (NN) architectures such as the Spiking Neural Network (SNN).","categories":["AI Trends"],"tags":["Data Science","Machine Learning","Neural Networks"],"author_name":"Vijaysinh Lendave","publish_date":"2021-11-13T16:00:00","publication_year":"2021","word_count":2134,"keywords":["Go","TPU","AI","neural network","Machine Learning","Scala","computer vision","RAG","Ray","deep learning","Data Science","R","Neural Networks"],"extracted_tech_keywords":["AI","deep learning","neural network","computer vision","Ray","RAG","TPU","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/a-tutorial-on-spiking-neural-networks-for-beginners\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10141777,"title":"Anthropic’s Claude Can Now Write Like You","content":"Anthropic has announced a major update in its Claude model. Claude will now be able to follow custom writing. This allows users to upload sample texts, using which the model will generate outputs in a similar style. Chris Gorgolewski, a technical staff member at Anthropic, took to X to make the announcement. “How many times AI generated some content for you, but it ‘didn’t sound like you’? We have solved it.” How many times AI generated some content for you, but it \"didn't sound like you?\". We have solved it – now you can give Claude samples of your writing and it will use them to adjust its writing style. It was a lot of fun to contribute to this feature! https:\/\/t.co\/GaCAQ5Dr2Z— Chris Gorgolewski (@chrisgorgo) November 26, 2024 Tailored Styles for All Users Anthropic’s new enhancement offers different preset styles, namely normal, formal, concise, and explanatory, to cater to various professional needs. “Whether you’re a developer writing formal documentation, a marketer crafting clear brand guidelines, or a product team planning extensive project requirements, Claude can adapt to your preferred way of writing,” Anthropic claimed. Users can also create custom styles by uploading sample content that reflects their preferred communication methods. This allows Claude to adapt to specific writing tones and structures. GitLab was among the early adopters of this update. Taylor McCaslin, product lead for AI and ML tech at GitLab, noted, “Claude’s ability to maintain a consistent voice while adapting to different contexts allows our team members to use styles for various use cases including writing business cases, updating user documentation, and creating and translating marketing materials.” As mentioned in the official blog post, users can define their preferences using preset options or customise their own within the Claude.ai chat interface to implement these styles. Anthropic is on the Rise As reported by AIM on Tuesday, Anthropic had also made its Model Context Protocol open source. This made connectivity between AI assistants or chatbots and data repositories very easy. The company also recently beat big tech by releasing the ‘computer use’ feature to its Claude 3.5 series, along with several other updates such as Claude Artifacts, analysis tool and visual PDF. These updates created a significant hype around Anthropic developments in the past year. OpenAI: \"Look at our new AI-powered internet search! with a fine-tuned model!\"Perplexity: \"Look at our AI-powered internet search combining multiple models!\"Google: \"We built a custom pipeline for AI summaries!\"Microsoft: \"We did AI search built around Bing years ago!\"Claude: pic.twitter.com\/2oNkCapmUu— Ethan Mollick (@emollick) November 1, 2024 The company has also secured its place in the public sector and government organisations, starting from its partnership with Palantir to provide the US government with its advanced AI model Claude for data analysis and complex coding activities in projects of national security interest. With this, Anthropic has built the trust and confidence of various players in the public as well as private sector. Anthropic CEO Dario Amodei also recently revealed that their Opus model isn’t going anywhere. In addition, Anthropic will also launch Claude 3.5 Opus and Claude 4.0 as per the usual business cycle.","excerpt":"“How many times has AI generated content for you, but it ‘didn’t sound like you?’ We have solved it,” says Chris Gorgolewski, tech staff member at Anthropic.","categories":["AI News"],"tags":["Anthropic"],"author_name":"Sanjana Gupta","publish_date":"2024-11-27T14:42:09","publication_year":"2024","word_count":516,"keywords":["Anthropic","AI assistants","TPU","OpenAI","AI","chatbots","ML","Claude 4","Aim","Claude 3.5"],"extracted_tech_keywords":["AI","ML","OpenAI","Claude 4","Claude 3.5","Anthropic","Aim","AI assistants","chatbots","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/anthropics-claude-can-now-write-like-you\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10047531,"title":"How to Become Google Cloud Certified Professional Data Engineer?","content":"Google Cloud offers a variety of professional certification programs to assess the expertise of professionals using different types of Google Cloud services in different domains. In the field of data science, it offers a very popular certification program named “Google Cloud Certified Professional Data Engineer”. This certification, through an exam, checks the potentials of candidates working as or equivalent to data engineers and gives recognition to the candidates in the form of a certificate. Here, we will discuss this certification program in detail, where we will focus more on how to prepare to achieve this certification. Google Cloud Certified Professional Data Engineer certification Google Cloud Certified Professional Data Engineer is a suitable certification for those who are working in the field of data science and their roles are most likely to handle the data, process the data, do a lot of feature engineering on the data to make it ready for modelling purposes. The professionals whose day-to-day tasks include collecting the data, transforming the data and publishing the data for data-driven decision making should go ahead with this certification that will give them an opportunity to stand out from others. However there are no strict criteria for this certification, but Google recommends having 3+ years of industry experience, out of which 1+ years should be related to designing and managing data related solutions with Google Cloud. The reason behind this recommended experience is that this certification assesses the candidate’s ability to design, build and operationalize the data processing systems and machine learning solutions with ensured quality. It also expects that a data engineer should be able to leverage, deploy, and continuously train pre-existing machine learning models. About the Exam To achieve the Google Cloud Certified Professional Data Engineer certification, the candidates need to pass an exam that is conducted by Google. This exam can be taken remotely with an online proctoring facility or it can be taken from the designated test centres across the world. To register for the exam, candidates need to pay a reservation fee of $200 USD. This exam is available only in English and Japanese. The important details regarding the exam are given below: Duration: 2 HoursNumber of questions: 50Format: Multiple choice questions The questions in the exam belong to different topics including the following: Designing data processing systemsBuilding and operationalizing data processing systemsOperationalizing machine learning modelsEnsuring solution quality The candidates appearing in this exam should align themselves with the aforementioned topics and ensure a strong understanding of these skills. However there is no information on the minimum passing score officially from Google, but as per the successful candidates, you must score at least 70% to pass the exam. Steps to Prepare As we have discussed above, there are different skills and knowledge expected from aspiring candidates. To meet these expectations and perform well in the exam, one needs to have excellent preparation. The following are the suggested steps a candidate should go through in order to achieve the goal:- Get the Required Information: First of all, you need to collect all the required information for this exam. You can find the highlights of this certification program on its official website. The second important detail that needs to be collected is the outline of contents to which the questions belong in this exam. You can find this outline using this link. Along with these, the sample questions for this exam are also provided by Google that can be found using this link. Using contents outline and sample questions, you can understand the breadth and depth of the exam. Prepare with Courses: Google Cloud offers a wide range of courses related to the skills needed for this certification. The list of these courses can be found using this link. Some of these courses are offered by Google Cloud itself, and some are offered by Coursera. These courses are arranged on a learning path that supports the candidates to prepare for this certification with the right content. Learn about Google Cloud Services: You can find here the details on a variety of Google Cloud services with their use cases and implementations. You should go through and learn about all the resources which are related to data engineering and data processing. Along with this, you can find here many documents and blogs related to Google Cloud services, including their application and use cases. All you need to do is to filter out the resources based on the exam contents outlined. These resources will help you in building a solid understanding of relevant Google Cloud services Take Mock Test: You can find many mocks exams on popular professional certifications offered by MachineHack, India’s leading hackathon and assessment platform. These mocks are provided for free, just to support the candidates. These are prepared based on the details as given by certifying organizations and feedback from the successful candidates. You can take the Google Cloud Certified Professional Data Engineer Certification Mock test on MachineHack where you will experience taking this exam in real life. Here, you will see the same levels of questions belonging to all the required topics to be attempted within the stipulated time. After taking this mock exam, you will be able to assess your level of preparedness for the exam. Repeat the Steps: Repeat steps 2 to 4 until you get an excellent score in the mock. Concluding Remarks Nothing is difficult or impossible if it is adequately attempted with good preparation. Google Cloud Certified Professional Data Engineer is a suitable certification for data engineering-related professionals as it gives them a chance to stand out from others. To achieve this credential, they need to crack an exam that requires good preparation. The candidates can prepare for this exam by going through a wide range of learning resources provided by Google Cloud and followed by a mock test that can be taken on the MachineHack platform. Wishing all the best to the certification aspirants!","excerpt":"Google Cloud offers a variety of professional certification programs to assess the expertise of professionals using different types of Google Cloud services in different domains. In the field of data science, it offers a very popular certification program named “Google Cloud Certified Professional Data Engineer”. This certification, through an exam, checks the potentials of candidates […]","categories":["AI Highlights"],"tags":["data engineer","Data Science Certification","data science certifications","Google Cloud"],"author_name":"Dr. Vaibhav Kumar","publish_date":"2021-09-03T14:00:00","publication_year":"2021","word_count":985,"keywords":["data science certifications","data science","Go","Google Cloud","machine learning","Data Science Certification","AI","data engineer","data-driven","feature engineering","RAG","data engineering","GAN","R"],"extracted_tech_keywords":["AI","machine learning","data science","RAG","R","Go","data engineering","feature engineering","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-to-become-google-cloud-certified-professional-data-engineer\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":53239,"title":"How Spotify’s Algorithm Manages To Find Your Inner Groove","content":"Have you ever wondered how Spotify manages to recommend you that perfect song, playlist or even that ‘daily mix’? The answer is simple — data. Spotify’s algorithm is always finding new ways to understand the kind of music one listens to — from the songs that are always on repeat to the favourite genre that one can’t let go. Not only is the algorithm monitoring the music history but also analyses the reason behind a person listening to a particular song or preferring a particular genre over the other. Spotify’s algorithm is what sets it apart from other music streaming services. Where other music streaming services make use of paid playlists created by individuals and other communities, Spotify, on the other hand, relies on its algorithm in addition to free human-created playlists. How Does This Algorithm Work? As of October 2019, Spotify had over 248 million active users. The algorithm seems to be working very well for the streaming giant among the competitions. The first thing one notices on Spotify is the home screen, that’s where the recommendations start, and the home screen is governed by an AI system called Bandits for Recommendations as Treatments or simply known as BaRT. BaRT is the real reason why one doesn’t go on searching for an appropriate playlist to listen to Spotify. Of course, if one has been on Spotify for some time, it starts recommending songs based on the previous listening activities. But, this is not just is, BaRT also drops in fresh music that it thinks users will like in order to keep them out of the same listening loop. Now two concepts come into play with BaRT — exploit and explore. The combination of ‘exploit’ and ‘explore’ is the key to Spotify’s recommendation. While exploiting, Spotify makes use of every activity produced by a user. Exploit usually makes use of the user’s listening history, songs skipped, playlists the user has created, social media activity on platforms and even one’s location to recommend music. While exploring, Spotify studies the rest of the world. It starts searching for playlists and artists similar to your listening taste and also looking at the popularity of the artists in that area whom you haven’t heard of or any other related works. By taking all this data that has been collected from you over the decade, or as long as you have been on the Spotify, it presents their user with ‘Decade Wrapped Playlists’, and by taking the data set of different genres one has listened to over the decade will become the ‘Best of the Decade For You’. The 30-sec rule BaRT’s success totally depends on whether one listens to the songs recommended in the ‘shelves’ or ‘rows’ on their home screen. Spotify’s algorithm looks at the duration of the time one has spent on a song, and if it is for more than 30 seconds, then the platform takes it as a check on their recommendations. The longer one spends on a song or a playlist, the better their suggestions will get. So, if one doesn’t like the first 28-second of a song, it is better to skip it before the 29-second mark to never get to listen to something like that again! Recommending New Artists by Analysing Audio Recommending music liked by other people or from the data received from millions of music blogs which can be called collaborative filtering doesn’t seem to work in the case of a new artist. In that case, how do you think one gets a recommendation on a song that has just been released, that too by a new artist? The solution to this problem is Spotify analysing the audio itself, and training the audio analysis algorithm to learn to recognise different desirable characters to music. Spotify experiments could identify various aspects of songs like distorted guitars, and while others recognised more abstract ideas like genres. Recommendations in Automatic Playlist Continuation It is obviously has been observed that Spotify plays a song automatically after your current playlist ends, and this happens because of artificial intelligence too. The feature analyses the songs in your playlists and tries to predict the music that can be played next after your song finishes. The platform has taken this feature to the next step by releasing a ‘Million Playlist Dataset’ of user-generated Spotify playlists. This vast data set was released to help understand the behaviour of users on Spotify. The auto continuation of the playlist becomes a problem when one doesn’t get appropriate songs after the existing playlist finishes, so in order to solve this problem, the company invited AI researchers all around the world to present their solutions at The RecSys Challenge 2018. Recommendations Using Locations The data provided by users about their age and gender along with their locations are also used by Spotify in order to study whether a user’s taste of music changes after the individual moves to a different state. Along with that, how the age of the listener impacts the kind of music the person listens to. “Every word anyone utters on the internet about music goes through our systems that look for descriptive terms, noun phrases, and other text,” Brian Whitman wrote, CFO The Echo Nest (Acquired by Spotify in 2014) By studying the kind of music people listen to in particular areas and then putting this data against the group of people who have recently moved to a different area, is a move done by Spotify. With different overall trends, the Spotify team concluded that over some time, location does factor in when it comes to moving to a different area in a certain way. Outlook With Spotify’s recent struggles and by struggles I mean, the stock hit that it took last year because of the emerging competitions like Amazon.com, Alphabet and Apple, it is continuously looking out for different areas to improve or trying to enhance the existing features. Making features like Discover Weekly more personalised is one of the areas where people probably don’t mind sharing their ‘musical’ data with the streaming platform. Spotify is also taking bold steps to improve it’s ‘Behind the Music’ feature, where it is looking to integrate real-time lyrics, and sort of make it karaoke-ish. It also plans on including a feature which can detect covers of a song, although it would prove challenging especially with a genre like Jazz. With other competitions like Youtube Music, Amazon Prime Music, direct competition juggernaut Apple music, and the other streaming platforms, Spotify is still keeping its top position intact, and the biggest reason is its algorithm.","excerpt":"Have you ever wondered how Spotify manages to recommend you that perfect song, playlist or even that ‘daily mix’? The answer is simple — data.  Spotify’s algorithm is always finding new ways to understand the kind of music one listens to — from the songs that are always on repeat to the favourite genre that […]","categories":["AI Features"],"tags":["spotify"],"author_name":"Sameer Balaganur","publish_date":"2020-01-06T17:00:00","publication_year":"2020","word_count":1097,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","spotify","ViT","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","ViT","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-spotifys-algorithm-manages-to-find-your-inner-groove\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10010141,"title":"IBM Watson Just Analysed a TV Debate. Read to Know How","content":"Bloomberg Television’s show “That’s Debatable” had an unusual participant on its show broadcasted on October 9. In a debate on the topic “Is it time to redistribute the world’s wealth?”, IBM Watson synthesised thousands of responses and opinions received from the public to incorporate into the debate. IBM Watson used a new natural language processing feature called key point analysis which categorises and summarises thousands of public opinions to a handful of concrete key points. Key point analysis is basically the next generation of ‘extractive summarisation’ which processes statements in a given text document to summarise the most significant points. It helps in examining large and complex documents to produce a concise list of information which can be acted upon. Key point analysis can help organisations that overload of data from which relevant information needs to be extracted to make data-driven decisions. IBM Watson and Key Point Analysis That’s Debatable is a limited series show, hosted by John Donvan, presented by Bloomberg media which features industry leaders, economists, and public intellectuals debating on some of the most pressing socio-political issues. The episode in question had the former U.S. Labor Secretary Robert Reich and former Greece Finance Minister Yanis Varoufakis arguing for the motion, against former U.S. Treasury Secretary Lawrence Summers and Manhattan Institute Senior Fellow Allison Schrager. To have a larger set of voices and ideas to the debate, the show admitted up to 3,500 submissions from the public online, before the show. Using key point analysis in its Debater technology, IBM Watson identified 1,600 usable arguments and 20 key points among the total submissions. This analysis helped in prompting further discussion and exchange of ideas among the debaters. Using key point analysis, it was found that 56% of the arguments were in favour, and remaining were against it. Using natural language processing, the IBM system rated the comments for relevancy and discarded opinions which are not in the context of the topic. In this step, all potential keywords which were too long, emotional in tone or incoherent were discarded. Then the system grouped the remaining comments into two broad categories — those in support of the motion and those against it. The comments in each of these groups are distilled to a handful of key points using the language processing algorithm to summarise the gist and omit repetition of the same ideas; this step is the one that generates the actual narrative. Grown out of IBM’s Project Debater research, spearheaded by its Israel-based AI lab, it is hoped that the key point analysis will find applications in conducting market research for companies and for involving citizens in governance. The key point analysis is the upgradation in capabilities first used with Project Debater. The initial project included several NLP capabilities, including understanding English idioms and clustering topics that IBM is integrating into Watson NLP products. Notably, the original Debater technology of IBM made its debut in 2018 when it competed with a human debater on the topic ‘We should subsidise space exploration’. However, the key point analysis used in the Bloomberg show considers the quantitative nature of the output, whereas the former just summarises the information into a narrative. This quantitative analysis puts forth the prevalence of each key point analysed in the data. Wrapping Up With its participation in the show, IBM has demonstrated how the company is actively investing in advancing Watson’s ability to understand business language and generate insights. Further, from March this year, IBM has introduced the features of Debater technology in Watson Discovery, Watson Assistant, and Watson Core Services, apart from other advanced features such as sentiment analysis, summarisation capabilities, classification of elements in business documents, and advanced topic clustering. It must be noted that a small section is also critical of new Watson NLP capability. It is being argued that this technology may further push social inequality if used extensively in social analysis.","excerpt":"Bloomberg Television’s show “That’s Debatable” had an unusual participant on its show broadcasted on October 9. In a debate on the topic “Is it time to redistribute the world’s wealth?”, IBM Watson synthesised thousands of responses and opinions received from the public to incorporate into the debate. IBM Watson used a new natural language processing […]","categories":["AI Features"],"tags":["Sentiment Analysis"],"author_name":"Shraddha Goled","publish_date":"2020-10-19T18:00:41","publication_year":"2020","word_count":649,"keywords":["Go","Sentiment Analysis","TPU","sentiment analysis","AI","RAG","NLP","BERT","GAN","R","Redis"],"extracted_tech_keywords":["AI","NLP","RAG","sentiment analysis","TPU","Redis","R","Go","BERT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ibm-watson-just-analysed-a-tv-debate-read-to-know-how\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10020946,"title":"How Tata Steel Uses AI: A Case Study","content":"Tata Steel is one of the prominent names in the steel-making industry boasting over three decades of manufacturing expertise. The company is currently the world’s second-most geographically-diversified steel producer, with fully integrated operations — from mining to the manufacturing and marketing of finished products. To sustain its leadership position in a volatile market, Tata Steel needed to fortify its supply chain. Poor visibility of in-plant operations was causing delays in loading trucks. This, in turn, triggered a series of inefficiencies like traffic congestions, parking problems, and forced route diversions for inbound\/outbound vehicles. Tata Steel was also facing challenges in material handling during transit. Overspeeding, route diversions, theft, and prolonged and unexpected stoppages were weighing down transportation operations. A high degree of dependency on transporters to get information on vehicles, high volumes of customer inquiry calls, and the inability to identify late deliveries further compounded the issues. Lack of adequate visibility of ground-level transportation operations and increasing volumes of delayed deliveries have resulted in high transportation costs. Due to the absence of real-time visibility of fleet operations, it was becoming increasingly difficult to hold freight carriers accountable. Thus, to achieve optimised logistics operations, mitigate risks and eliminate any delays, the company required a 100% visibility of their in-plant and in-transit transportation. To ensure intelligent management of logistics, reduce turnaround-time, risk mitigation, end-to-end transportation visibility, and delivery efficiency, Tata Steel collaborated with FarEye, a predictive logistics SaaS platform. To understand the case better, we got in touch with Kushal Nahata, the CEO & Co-founder of FarEye. The Use Of AI FarEye created a customised solution that included automating core delivery operations by introducing real-time visibility in in-plant and in-transit transportation operations and leveraging electronic proof of delivery to boost customer experience. By leveraging FarEye’s auto-allocation engine, Tata Steel automatically allocated orders to transporters when their engines are running. Automating key compliance checks is another important aspect of the solution. “FarEye’s platform seamlessly executes digital checks of RC details and vehicle age before letting a vehicle pass through the plant’s entry point. Entry is restricted in case any red flags are raised during the process. Through e-indents, the status of orders assigned to a transporter is automatically mapped as confirmed, placed, and invoiced.”Explaining the process better, Kushal said With FarEye’s AI-based real-time tracking capabilities, Tata Steel can keep a tab on where exactly a vehicle is inside the plant and what’s its status. The real-time tracking and routing capabilities are powered by machine learning algorithms. These algorithms parse historical data to identify best routes for executing deliveries. To enhance in-transit operations, FarEye’s AI-driven platform accurately predicts deliveries; whether it will be done early, on time, or delayed. The platform’s real-time tracking capabilities also empower Tata Steel BSL to gain end-to-end visibility of in-transit operations. “FarEye’s in-plant logistics optimisation processes are driven by IoT-devices, like sensors that work in tandem with compliance applications to identify irregularities and restrict vehicle entry and movement within a plant,” added Kushal. To further boost visibility and KPI benchmarking of transporter performance, FarEye designed easy-to-navigate dashboards to make quick data-driven decisions. These dashboards offer critical insights on loading performance based on vehicle-in and vehicle-out instances, update on trip status based on vehicle location, and analyse unloading performance at specific destinations. Wrapping Up Real-time tracking with a 360- degree view keeps delivery stakeholders and plant executives updated on their shipments and empowers them to drive efforts to ensure on-time deliveries and boost customer experience. FarEye’s AI and ML applications have significantly improved Tata Steel BSL’s predictive and information analysing capabilities in  ETA (Estimated Time of Arrival), in-transit performance, loading and unloading TAT (Turn-around Time), and allocation and placement. Tata Steel attained 100% operational visibility with a 57% reduction in theft. The AI-powered platform also engaged 20 different stakeholders with collaboration possibilities. Tata Steel has also witnessed a 32% reduction in loading and unloading turnaround time.","excerpt":"Tata Steel is one of the prominent names in the steel-making industry boasting over three decades of manufacturing expertise. The company is currently the world’s second-most geographically-diversified steel producer, with fully integrated operations — from mining to the manufacturing and marketing of finished products. To sustain its leadership position in a volatile market, Tata Steel […]","categories":["IT Services"],"tags":[],"author_name":"Sejuti Das","publish_date":"2021-02-28T13:00:00","publication_year":"2021","word_count":645,"keywords":["Go","machine learning","programming_languages:R","AI","data-driven","ML","programming_languages:Go","Git","RAG","R"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","R","Go","Git","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-tata-steel-uses-ai-a-case-study\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075175,"title":"Midjourney is Biased","content":"A few years ago, Microsoft released an AI chatbot to develop a conversational understanding. However, the software started to behave weirdly while interacting with thousands of Twitter users, eventually forcing the corporation to shut it down in less than 24 hours. In a shocking turn of events, the chatbot expressed its support for Adolf Hitler and professed to despise the Jewish community. Although ultimately discontinued, this served as a great example of how bias may be exhibited by AI systems as a result of either their programming or data source. As of right now, there are also allegations of prejudice against well-known text-to-image generators like Midjourney. Since its dataset consists of millions of images amassed over time, the programme also displays societal bias, which is often reflected in images we upload online. For example, this Twitter user shared how the Stable Diffusion shows bias in the system: Wanted examples of generative #AI models results for a townhall event next week.1st attempt w\/ @huggingface, #stablediffusion & #gpt got this!Much worse than expected! Serious measures needed to deal w\/ how easily bias and prejudice creep in…Where is #responsibleAI? pic.twitter.com\/MCHLTNFTQY— Virginia Dignum is mostly commenting on #LinkedIn (@vdignum) September 10, 2022 We got eerily similar results from Midjourney, when searched with generic prompts. All four results for the text prompt “Doctor” were male doctors, while all the results for the text prompt “nurses” were female nurses. These results clearly indicated a gender bias. However, much to general dismay, this was in no way a standalone case. Midjourney repeatedly identifies text prompts such as ‘engineers’, ‘lawyers’, and ‘construction workers’ as predominantly ‘male’ or ‘masculine’. It also tends to identify prompts like ‘president of a country’ as masculine, while upon prompting for ‘cook in a home kitchen’, the generative engine comes up with results that are predominantly feminine. It’s not limited to gender bias The generative engine’s bias is not merely limited to gender—it extends to the category of nations as well. Midjourney views different nations with different, seemingly preset lenses. In a peculiar and—more or less, unnoticed—parallel, Midjourney’s resultant images reflect stereotypes that are widely, albeit implicitly, still present in human society. For instance, the programme displays visibly different colour palettes for India and other nations. However, it appears as if the system is repeatedly producing images with yellow\/orange hue for prompts with the term “Indian” attached to them. In the images above, it can be seen that when prompted with “India”, Midjourney produced different results but all with a typical yellow hue characterising them. It is plausible that with future updates to the generator, images like these may not appear in the results with such high frequency. However, as of now, these images indicate how the AI may be stereotyping text prompts like ‘India’ to produce results with a mystical\/fantastical aesthetic and ancient architecture. It is notable that even when prompted with “Indian city” or “Indian skyline,” the generated images routinely include architectural elements like forts, tombs and alike, as shown below. In another experiment, it was found that even professionals of Indian origin are not identified by the programme as what it generally identifies as ‘regular professionals’. For the software, Indian physicians must constantly be dressed in religious garb, and Indian engineers must always be working in structures that resemble temples or similar architecture. Apart from these stereotypes, when prompted ‘Indian and American standing in front of their houses’, stark differences in the resultant images are evident. The Indian equivalent is often a peasant standing in front of what appears to be a hut-like house, whereas the American counterpart is typically dressed in urban clothing and is seen standing in front of what appears to be a small villa. Likewise, in the images below, a similar stereotype can be noticed in the results of mugshots of Indian men and women versus how Midjourney perceives men and women typically look. How does the bias arise? Though the prejudice in AI is not new, we have seen some form of bias ingrained in technology ever since its inception. For instance, in Broward County, Florida, a criminal justice algorithm misclassified African-American defendants as “high risk” at a rate that was nearly twice that of Caucasian defendants. In addition to the dataset it is fed, the AI algorithm also takes systemic and human bias into account. The way institutions function typically leads to systemic biases, which frequently disfavour particular social groups. Human prejudices can be linked to how people often utilise facts to fill gaps in their knowledge. A common example is how a person’s neighbourhood may sometimes influence how the authorities may perceive them. Human, systematic and computational biases, when combined, form a pernicious  mixture—often worsened when the AI system lacks explicit guidelines.","excerpt":"While generative search engines are gaining substantial popularity, do they also have some kind of bias in them?","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Lokesh Choudhary","publish_date":"2022-09-14T15:00:00","publication_year":"2022","word_count":788,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","GPT","stable diffusion","AI Tool","R","llm_models:GPT"],"extracted_tech_keywords":["AI","R","Go","GPT","stable diffusion","RPA","llm_models:GPT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/midjourney-is-biased\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":45822,"title":"Adobe Applies Real-Time Layered Approach To Customer Journey Analytics","content":"Tech giant Adobe this week showcased how it was tweaking its new Adobe Customer Journey Analytics solution in a way that is easy to use. Announced recently, Adobe’s new product helps businesses and brands access layers of multi-channel customer data to be curated and stacked to reveal new perspectives into how customers interact with a brand. Reportedly, the new capability in Adobe Analytics uses the Adobe Experience Platform, which can piece together customer data from across the enterprise, and opens up new ways to understand insights across online, offline, and third-party channels. “The customer journey is a progressive stream of digital and off-line experiences that are unique to each individual. An integrated marketing platform allows you to create a journey filled with relevant and meaningful personal experiences across channels, devices and geographies,” said Adobe. It also has features like: A unified customer view shared across teams Better understanding of key touchpoints Identification of conversion pathways Team alignment to a customer-centric focus Adobe said that by combining the strengths of the integrated solutions in Adobe Experience Cloud, the user can seamlessly guide a customer journey that is rewarding, delightful, consistent and personalised across the entire experience lifecycle. Nate Smith, group manager for product marketing for Adobe Analytics, told a noted tech porta, “When you think about organisations that are trying to do omnichannel analysis or trying to get that next channel of data in, they now have the platform to do that, where the data can come in and we standardise it on an academic model.”","excerpt":"Tech giant Adobe this week showcased how it was tweaking its new Adobe Customer Journey Analytics solution in a way that is easy to use. Announced recently, Adobe’s new product helps businesses and brands access layers of multi-channel customer data to be curated and stacked to reveal new perspectives into how customers interact with a brand. […]","categories":["AI News"],"tags":["Adobe"],"author_name":"Prajakta Hebbar","publish_date":"2019-09-11T16:18:47","publication_year":"2019","word_count":255,"keywords":["programming_languages:R","AI","Adobe","ML","Git","analytics","GAN","R"],"extracted_tech_keywords":["AI","ML","analytics","R","Git","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/adobe-applies-real-time-layered-approach-to-customer-journey-analytics\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10112143,"title":"Diggibyte Technologies is Certified as Best Firm for Data Engineers","content":"Diggibyte Technologies is certified as the Best Firm For Data Engineers to work for by Analytics India Magazine (AIM) through its workplace recognition programme. The Best Firm certification surveys a company’s data and analytics employees to identify and recognise organisations with great company cultures. AIM analyses the survey data to gauge employee approval ratings and uncover actionable insights. “We are delighted to share the exciting news that Diggibyte Technologies Pvt Ltd has been honoured as the Best Firm for Data Engineers. This prestigious recognition is a reflection of our exceptional team and the outstanding culture we’ve fostered. A sincere thank you to our remarkable Data & Analytics team, whose talent and dedication have been instrumental in achieving this milestone. It’s not just about data; it’s about our great culture that propels us forward. Here’s to celebrating our success and the thriving culture that continues to inspire innovation and excellence” said Lawrance Amburose, CFO and Managing Director, India at Diggibyte Technologies. “Our team is our biggest asset, and this recognition propels us towards our commitment to fostering an innovative and creative ecosystem within the Data & Analytics space. We take immense pride in being recognized as Best Firm for Data Engineers at Diggibyte Technologies, attracting exceptional young talent. I extend my gratitude to our team and our valued customers. Together, we aspire to set the standards as thought leaders and market pioneers in Data & Analytics. Thank you, AIM, for selecting Diggibyte as the Best Firm for Data Engineers,” said Sekhar PVR – COO at Diggibyte Technologies. The analytics industry currently faces a talent crunch, and attracting good employees is one of the most pressing challenges enterprises face. The certification by Analytics India Magazine is considered a gold standard in identifying the best data science workplaces and companies participate in the programme to increase brand awareness and attract talent. The Best Firms certification is the biggest data science workplace recognition programme in India. To nominate your organisation for the certification, please fill out the form here.","excerpt":"The Best Firm certification surveys a company’s data and analytics employees to identify and recognise organisations with great company cultures","categories":["AI Highlights"],"tags":["Top Trend"],"author_name":"AIM Media House","publish_date":"2024-02-07T12:00:00","publication_year":"2024","word_count":335,"keywords":["Top Trend","data science","Go","programming_languages:R","AI","innovation","programming_languages:Go","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/diggibyte-technologies-is-certified-as-best-firm-for-data-engineers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039721,"title":"Inside Nucleus, An Integrated Data Platform From Vodafone &#038; Google Cloud","content":"Vodafone and Google Cloud recently entered a six-year strategic partnership to build a robust integrated data platform called Nucleus. The new platform will drive the use of reliable and secure data analytics, learnings and insights to create new digital products and services for Vodafone customers worldwide. In 2019, Vodafone migrated to Google Cloud to support the former’s data analytics, infrastructure and machine learning. Now, Vodafone and Google Cloud have extended the strategic partnership to develop integrated data services. The new integrated platform built using the latest hybrid cloud technologies from GCP will house a system called Dynamo with the capability to process and drive large volumes of data across Vodafone’s pipelines, enabling it to offer new personalised services in multiple markets. Dynamo will allow Vodafone to offer new connectivity services by releasing smart network features, such as accelerating the broadband speed. It will also help Vodafone to centralise its operation, cut down operation costs and re-use data artefacts. Dynamo can move about 5,000 data feeds to the Cloud per day, which is equal to 50 TB of data per day. Depending on the time of the day, Dynamo can automatically control the capacity of a data pipeline, and increase the speed of building a data pipeline by 25%. Around 1,000 employees from Britain, Spain, and the United States will collaborate on the project to allow Vodafone to utilise machine learning to identify and resolve customer issues quickly across the world. This reinvented approach towards data analytics and business intelligence will empower Vodafone to make quick, informed decisions based on business insights. Already, Vodafone has identified more than 700 use-cases to deliver new products and services across markets. It will also enhance Vodafone’s clients to access its data analytics, intelligence, and machine-learning capabilities. Gamechanger In a recent press release Johan Wibergh, CTO, Vodafone, said: “Vodafone is building a powerful foundation for a digital future. We have vast amounts of data which, when securely processed and made available across our footprint using the collective power of Vodafone and Google Cloud’s engineering expertise, will transform our services, to our customers and governments, and the societies where they live and serve.” Apart from enhancing Vodafone’s mobile and connectivity services, the detailed insight and data-driven analysis will empower data scientists to collaborate on health and environmental issues in 11 countries using automated machine learning tools. Using artificial intelligence and advanced analytics, Vodafone plans to provide a complete digital replica (digital twin) of its internal support. In addition to that, Vodafone will also host its entire SAP environment to Google Cloud, including its core SAP workloads and SAP Central Finance. Thomas Kurian, CEO at Google Cloud, said the telecommunication companies have rapidly transformed the services and experiences they provide customers through data and analytics, especially during the pandemic. The partnership between Google Cloud and Vodafone will accelerate the digital transformation of the industry through its innovative new platform. With the development of Dynamo, Vodafone will store and process the vast data it generates on Google Cloud. It will direct all of Vodafone’s data by extracting and encrypting it back and forth from the source to the Cloud, enabling intelligent data analysis and insight. This will also encourage multinational corporations to move their data to Cloud in future. According to news reports, Vodafone and Google Cloud are already considering exploring opportunities to provide consultancy services to multinational organisations.","excerpt":"Vodafone and Google Cloud recently entered a six-year strategic partnership to build a robust integrated data platform called Nucleus. The new platform will drive the use of reliable and secure data analytics, learnings and insights to create new digital products and services for Vodafone customers worldwide.  In 2019, Vodafone migrated to Google Cloud to support […]","categories":["Global Tech"],"tags":["digital twin","Vodafone"],"author_name":"Ritika Sagar","publish_date":"2021-05-06T16:00:00","publication_year":"2021","word_count":561,"keywords":["Go","API","artificial intelligence","machine learning","GCP","AI","R","digital twin","Git","RAG","analytics","Vodafone"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","RAG","GCP","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/inside-nucleus-an-integrated-data-platform-from-vodafone-google-cloud\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":46388,"title":"Scala Vs Kotlin: Which Programming Language Should Java Developers Learn?","content":"Java is one of the oldest and most popular programming languages. It has been used by organisations for software development purposes for years. However, in the modern software development scenario, this prominent language is lagging. This is where programming languages like Kotlin and Scala come into play. The codes for these new languages are mostly written for the Java Virtual Machine which serves the perfect roles in modern software development scenarios. In this article, we compare the two programming languages Scala and Kotlin and summarise which language will be more suitable for Java developers. Kotlin Kotlin is an open-sourced statically typed general-purpose programming language. JetBrains announced the development of this language in 2011 which serves as an alternative to Scala or Java to run on Java Virtual Machine (JVM). Several prominent organisations like Google, Uber, Netflix, Amazon, among others have been using this language for its number of advantages. One strong point about Kotlin is that it is officially supported by tech giant Google for the Android operating system. Kotlin has been mainly used for building server-side code and building mobile applications which run on Android devices. Features Multi-platform Language: Kotlin serves as a multi-platform language and can be used to write native macOS and Windows applications, Android applications and Java applications. Server-Side Development: For server-side development, Kotlin allows a coder to write expressive code while maintaining full compatibility with the existing Java-based technology. Android Development: Kotlin brings almost all the advantages of a modern language in the Android platform. Kotlin\/Native: Kotlin\/Native is designed to allow compilation for platforms and it supports iOS, Android, Windows, Linux, WebAssembly, among others. Scala Scala or Scalable Language is a general-purpose programming language which combines object-oriented and functional programming in one concise, high-level language. The source code can be compiled to Java bytecode and the resulting executable code can run on a Java Virtual Machine (JVM). The static types in Scala help in avoiding bugs and errors in complex applications and its JVM and JavaScript runtimes let a developer build high-performance systems with easy access to huge ecosystems of libraries. Unlike Java, this language has many features of functional programming languages like Scheme, Standard ML, and Haskell, including currying, type inference, immutability, lazy evaluation, and pattern matching. Features Extensible: Scala provides a unique combination of language mechanisms which make it easy to smoothly add new language constructs in the form of libraries. Pattern Matching: The case classes of Scala and its built-in support for pattern matching model algebraic types have been used in many functional programming languages. Other Comparisons Compilation When it comes to runtime and compilations, Scala consumes a little higher time to compile a code than Kotlin. Documentation Both languages are well documented there are several resources available both open source and paid. However, Kotlin being the newer one has fewer libraries, tutorials, etc. than Scala. Community Scala and Kotlin have been around us since 2004 and 2011 respectively. Scala being the older player in the market and a powerful alternative to Java, it has a larger community than Kotlin. Android Language Kotlin is an officially supported language for Android development while Scala can be used for Android development. Easy to Learn Kotlin is easier to learn than Scala. The former has an interactive official test environment where a developer can learn to convert Java codes. Outlook There are certain similarities between both languages such as both languages are alternative to Java on Java Virtual Machine (JVM). Both Scala and Kotlin are relatively new to the market. However, Language like Kotlin is backed up by tech giant like Google which means this language is surely going to last long as compared to the other one. Also, Scala being a little older to the market, it has a larger community which means it has more sustainable power along with job opportunities.","excerpt":"Java is one of the oldest and most popular programming languages. It has been used by organisations for software development purposes for years. However, in the modern software development scenario, this prominent language is lagging. This is where programming languages like Kotlin and Scala come into play. The codes for these new languages are mostly […]","categories":["Deep Tech"],"tags":["Java","Kotlin","scala"],"author_name":"Ambika Choudhury","publish_date":"2019-09-25T18:00:32","publication_year":"2019","word_count":638,"keywords":["Go","programming_languages:R","AI","Kotlin","ML","Scala","scala","JavaScript","GAN","R","Java","programming_languages:JavaScript"],"extracted_tech_keywords":["AI","ML","R","JavaScript","Go","Java","Scala","GAN","programming_languages:R","programming_languages:JavaScript"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/scala-vs-kotlin-which-programming-language-should-java-developers-learn\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":60006,"title":"This Is How Brillio Is Responding To The Dynamic Challenge In Real-Time Approach","content":"With the unprecedented government actions such as work from home, travel ban and other such during the COVID-19 crisis, organisations have been tackling several issues when it comes to remote access, collaboration as well as security. One of the global leaders in enterprise digital transformation, Brillio responded this dynamic challenge by using an adaptive real-time approach which they call a “never before” scenario. The company rewrote the business continuity and crisis management playbook and within a week, Brillio deployed a virtual workforce program that enabled every one of their global employees to work from home, with little to no disruption to the business operations. As part of the Business Continuity Planning (BCP), Brillio has anticipated the government actions and provided work from home to the employees since March 23rd. The company has been using its existing infrastructure, technology platforms, and operational policies to remote access capabilities to address the near-term challenge. Click here to view the post. In a blog-post, Raj Mamodia, Chairman and CEO at Brillio pointed out several key-lessons from their real-time experiences, they are mentioned below: Lack of or too many collaboration tools and limited training on how employees should operate effectively in work from home scenario.No policies or procedures on how to maintain productivity at the scale of a remote workforce, once basic connectivity is established, and collaboration tools were deployed.Limited pressure testing of cybersecurity and infrastructure resiliency frameworks to support the volume and scale of a remote workforce.Little focus on human factors and ergonomics related to the health and welfare of employees working by themselves, in remote locations, for extended hours and how to minimise fatigue and repetitive motion injuries.No organisation guidance on maintaining social connections between employees via digital platforms to minimize loneliness and feelings of despair during a crisis.","excerpt":"With the unprecedented government actions such as work from home, travel ban and other such during the COVID-19 crisis, organisations have been tackling several issues when it comes to remote access, collaboration as well as security.  One of the global leaders in enterprise digital transformation, Brillio responded this dynamic challenge by using an adaptive real-time […]","categories":["AI Features"],"tags":["Coronavirus","Coronavirus and AI","covid-19"],"author_name":"Ambika Choudhury","publish_date":"2020-03-26T13:42:38","publication_year":"2020","word_count":296,"keywords":["Coronavirus and AI","Go","covid-19","AI","programming_languages:R","digital transformation","programming_languages:Go","Git","Coronavirus","ViT","disruption","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","ViT","digital transformation","disruption","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-is-how-brillio-is-responding-to-the-dynamic-challenge-in-real-time-approach\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":33864,"title":"5 Things That MLDS19 Can Add To Every Machine Learning Professional&#8217;s Resume","content":"Img Src: Cypher 2018 With barely two weeks left for one of the largest developer’s summit, Machine Learning Developers Conference 2019 promises to offer focused sessions, keynotes, and workshops by some of the foremost minds from the machine learning field. While covering both technical and business aspect of this emergent technology, the two-day summit will be packed with learning and networking sessions suited for both developers and C-suite executives. A one-of-its-kind conference, MLDS19 is definitely going to impact ML professionals in terms of career and business prospects. MLDS19 is going to offer unmatchable content, invaluable advice and mentoring while offering peer-to-peer networking opportunities and presentation material. It will provide a forum for networking, expert speakers and real-world professionals to share their insights and experiences. For any professional who has had an experience of more than six years is typically expected to have attended various events. MLDS, which is slated for 30-31 January 2018 can add a great value to your resume. Here are five of them: Visibility To The ML Community, Including Recruiters MLDS19 will bring together engineers, developers and business leaders who with their intensive industry experience are going to transform the practice of software development and revolutionise the next wave of innovation. They will focus on real-world practices and how to implement machine learning applications in your work environment. It brings like-minded people together who are willing to learn something new. It gives you an opportunity to be exposed to hundreds of different attendees of which few could be recruiters, helping you land that perfect job that you were looking for. Technical Training And Hands-On Exposure It goes without saying that the summit is all about technical training in the field of machine learning. It will give you an opportunity to learn new and innovative tools while providing an opportunity to network with peers and experts in the tech community. The summit will cover topics such as architecture, deployment, development, operations, implementation, management and more. It will brush up your skills and add on newer ones with the highly industry-specific use cases and applications. The sessions will give hands-on exposure to these skills that will be an add-on to your resume in key skills. With masterclasses on topics such as the democratisation of AI, industry use cases on machine learning and more, it will give an exposure to technical training that you most often need. Get Access To Workshops Machine Learning Developers Summit will offer workshops specific to the machine learning themes and therefore help you hold the stronger foot in the industry. Workshop environment will allow engaging in practical exercises while facilitating close interactions with other delegates participating in the summit. Every workshop will be a source of powerful information while creating an extensive effective learning opportunity for the developers. Staying Up-To-Date With The Latest Developments There is nothing more important than staying updated with the industry developments if you are looking to switch jobs in a technical field. When it comes to technology, especially software and programming, it is extremely crucial to stay relevant and updated with the latest versions and updates in tools and software. As there are a lot of revisions and new introductions that keep happening at a frequent pace, it is important to stay relevant. There is no better way to learn the newest emerging trends in the industry than by attending a conference where you can meet other like-minded people and watch the creative sparks fly while honing technical skills. Explore New Ways Of Working And Get Effective At Work By Implementing Learnings From The Summit MLDS19 will be a perfect platform to get hands-on information specific to your business directly from the experts. Across the 2-days, it will help you curate ideas to help improve your approach in your daily working. Though there is a lot of information on the web, conferences will cut through the clutter to deliver the best content specific to machine learning industry. The sessions on industry use cases of ML, Visit here to register for MLDS 2019.","excerpt":"With barely two weeks left for one of the largest developer’s summit, Machine Learning Developers Conference 2019 promises to offer focused sessions, keynotes, and workshops by some of the foremost minds from the machine learning field. While covering both technical and business aspect of this emergent technology, the two-day summit will be packed with learning […]","categories":["Deep Tech"],"tags":["machine learning developers summit","MLDS","mlds india","mlds nimhans"],"author_name":"Srishti Deoras","publish_date":"2019-01-22T11:20:46","publication_year":"2019","word_count":672,"keywords":["Go","machine learning","programming_languages:R","AI","data_tools:Spark","innovation","ML","mlds india","programming_languages:Go","machine learning developers summit","MLDS","R","mlds nimhans"],"extracted_tech_keywords":["AI","machine learning","ML","R","Go","innovation","programming_languages:R","programming_languages:Go","data_tools:Spark"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/5-things-that-mlds19-can-add-to-every-machine-learning-professionals-resume\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10083788,"title":"LLMs finally Bloom with Petals","content":"Even when large language models like BLOOM, PaLM, or GPT get open-sourced, fine-tuning and inferencing them on your system is a memory-heavy task. This might hinder developers from running these models on their systems and thus slow down innovation, leaving it in the hands of only big players. BigScience Workshop released Petals, which allows users to run language models with more than 100 billion parameters at home by loading a small part of the model on your machine, and then collaborating with other people to run other parts of inference and fine-tuning. Click here to check out the repository on GitHub. This BitTorrent-style running of large language models allows many times faster inference when compared to offloading on single systems, closer to 1 second per token. Parallel inference can reach hundreds of tokens per second. The script is built for CUDA-enabled PyTorch and uses Anaconda to install and is only available for Linux users for now. Mentioned in the GitHub page, “Petals” is a metaphor for a single person serving different parts of the model, and hosting together the entire language model – BLOOM, which has 176 billion parameters. Since the collaboration might be slow in the beginning because of privacy or security issues, the team has decided to give “bloom points” as an incentive system for people who donate their GPU time for people to fine tune it. Also read: ChatGPT and DALL-E on Discord","excerpt":"This BitTorrent-style running of large language models (LLMs) allows many times faster inference when compared to offloading on single systems, closer to 1 second per token. Parallel inference can reach hundreds of tokens per second.","categories":["AI News"],"tags":["language model"],"author_name":"Mohit Pandey","publish_date":"2023-01-02T18:00:00","publication_year":"2023","word_count":236,"keywords":["CUDA","ChatGPT","DALL-E","PyTorch","AI","Git","language model","GPT","GitHub","R"],"extracted_tech_keywords":["AI","ChatGPT","PyTorch","CUDA","R","CUDA","Git","GitHub","GPT","DALL-E"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/llms-finally-bloom-with-petals\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167601,"title":"Arctic Wolf is Using AI to Process 1.2 Trillion Cybersecurity Threats Daily","content":"Arctic Wolf, a leading US-based cybersecurity firm, has established a strong presence across North America and Europe, and has recently expanded into the APAC region, including India. Established in Minnesota in 2012, the firm focuses on managed detection and response services, providing businesses with threat detection and prevention solutions. While placing a strong bet on India’s rich pool of cyber talent, Arctic Wolf is adopting AI to enhance and accelerate its capabilities to provide cybersecurity services to businesses and enterprises. To understand how the company is using AI to combat cyber threats, AIM spoke with Jeff Green, senior vice president of engineering, and Dean Teffer, vice president of artificial intelligence at Arctic Wolf. AI in the Endpoint and Security Operations Centre To begin with, Green cited an example of Arctic Wolf’s recent acquisition of the endpoint security assets of Cylance, BlackBerry’s former cybersecurity unit, highlighting how the company uses AI to secure endpoints and detect malicious files. “We use AI in our SOC (security operations centre) so that analysts can look at the events and observations that we accumulate…either from an endpoint, a network sensor, or an API integration with a third party…like CrowdStrike, SentinelOne, and others,” he told AIM. Green highlighted that AI plays a key role in helping the SOC evaluate security events and identify those that are potentially malicious or problematic. The company’s use of AI is not just limited to analysis; Arctic Wolf also uses it for threat detection. Typically, a company’s SOC relies on a Security Information and Event Management (SIEM), which uses Sigma or Yara rules to identify threats. Arctic Wolf, however, is using AI to generate these types of detection rules. Green explained that they aim to translate human learnings into AI systems to make detections more efficient and improve response to customer issues. He illustrated this with an example, “If we spot something where a machine is talking out to a command-and-control server (C2), we can then use AI to detect that and block that machine from communicating further.” Source: Arctic Wolf 2025 Threat Report Being Picky in Using AI Teffer revealed to AIM that they are being deliberate about where they apply AI, basing their decisions on the impact observed in their pre-testing. “We’re not just having AI do all the work, but we’re having AI do components of the work,” he stated. He highlighted time-bound use cases where AI is intentionally left out, as those tasks require human intervention within a certain time. While he acknowledged that AI can perform those tasks, he added that the team tends to adjust how or where it uses AI based on the task. Green echoed this sentiment, agreeing that the application of AI is highly specific to the task at hand. He pointed to examples of some companies using AI for everything, noting that some things might not need AI at all. “Sometimes a simple rule could be the quickest way to detect something, right? And you don’t need to train massive models, and the performance of the model isn’t as quick as a rule.” “You’ve got to pick and choose the application. And I think that’s how we’re focused. We look at it very pragmatically.” Green further elaborated on his cautious approach, “AI is the answer, now what’s the question? You’ve got to be very focused on that. Otherwise, you can just go overboard and it’s not helpful.” Teffer explained that they do not start with AI but with the actual security problems that need solving, and the tasks being done. “It’s like starting simple and then only adding in GenAI if it’s needed.” Standing Out From Tech Giants and Helping Organisations Considering how every major company is trying to build cybersecurity solutions like Microsoft’s Security Copilot agents, AIM questioned how Arctic Wolf stands out from such offerings. To this, Green revealed that they have a very large SOC and process a lot of data. He highlighted that the data they see is potentially on the order of 1.2 trillion observations daily, which gives them an advantage when working with AI and building models. “Our major competitors, like Microsoft and others, work best when you buy all their products. Arctic Wolf has never been like that. It’s always been: whatever you have, we can add to it, but we’ll take what you have,” Teffer stated. He added that Arctic Wolf focuses on security outcomes independent of an organisation’s IT infrastructure. Compared to large companies that build AI models, which often rely on humans, Arctic Wolf relies on experts for fine-tuning security outcomes. The company, with its security expert teams, continually improves its AI tools with human reinforcement.“One of the characteristics of cybersecurity is that if we solved cybersecurity today, there’d be work to do tomorrow because attackers would be responding to that,” Teffer said.","excerpt":"“If we solved cybersecurity today, there’d be work to do tomorrow because attackers would be responding to that.”","categories":["AI Features"],"tags":["AI","Cybersecurity"],"author_name":"Ankush Das","publish_date":"2025-04-10T17:21:39","publication_year":"2025","word_count":802,"keywords":["Go","API","GenAI","artificial intelligence","programming_languages:R","AI","RAG","Aim","GAN","Cybersecurity","R"],"extracted_tech_keywords":["AI","artificial intelligence","GenAI","Aim","RAG","R","Go","API","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/arctic-wolf-is-using-ai-to-process-1-2-trillion-cybersecurity-threats-daily\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10055645,"title":"The Winning Papers At NeurIPS 2021","content":"The NeurIPS 2021 has started from December 6 and will go on till December 14 packed with tutorials, conferences and workshops. Meanwhile, NeurIPS 2021 has announced the recipients of the 2021 Outstanding Paper Awards, the Test of Time Award, and the new Datasets and Benchmarks Track Best Paper Awards. Let us take a more detailed look at them and understand what made them stand out. A Universal Law of Robustness via Isoperimetry By Sébastien Bubeck and Mark Sellke The researchers said that data interpolation with a parametrised model class is possible if the number of parameters is more than the number of equations to be satisfied. In deep learning, models are trained with many more parameters than what this classical theory would suggest. In this paper, the researchers suggest a theoretical explanation for this that proves for a broad class of data distributions and model classes, over parameterisation is necessary. This is required if one wants to interpolate the data smoothly. They show that smooth interpolation requires d times more parameters than just interpolation (d is the ambient data dimension). The team also shows the universal law of robustness for any smoothly parameterised function class with polynomial-size weights and any covariate distribution verifying isoperimetry. Read the full paper here. Source: Microsoft Research Youtube channel On the Expressivity of Markov Reward By David Abel, Will Dabney, Anna Harutyunyan, Mark K. Ho, Michael Littman, Doina Precup, and Satinder Singh. The researchers have studied the expressivity of Markov reward functions in finite environments through inspection of what kinds of tasks these functions can express. The paper looks into understanding the expressivity of reward as a way to capture tasks that we would want an agent to perform. The team bases the workaround three new abstract notions of “task” that might be desirable – a set of acceptable behaviours, a partial ordering over behaviours, or a partial ordering over trajectories. The results show that though reward can express many of the tasks, yet there exist cases of each task type that no Markov reward function can capture. The researchers also provide a set of polynomial-time algorithms that build a Markov reward function which allows an agent to optimise tasks of each of these three types. It determines correctly no such reward function exists. Read the full paper here. Deep Reinforcement Learning at the Edge of the Statistical Precipice By Rishabh Agarwal, Max Schwarzer, Pablo Samuel Castro, Aaron Courville, and Marc G. Bellemare. This paper looks into practical approaches to improve the rigour of deep reinforcement learning algorithm comparison. It looks into especially the evaluation of new algorithms that should provide stratified bootstrap confidence intervals, performance profiles across tasks and runs, and interquartile means. The researchers also show that standard approaches for reporting results in deep RL across many tasks and multiple runs can make it difficult to assess if a new algorithm represents consistent advancements over past methods. The performance summaries are designed to be able to compute with a small number of runs per task. Read the full paper here. MAUVE: Measuring the Gap Between Neural Text and Human Text using Divergence Frontiers By Krishna Pillutla, Swabha Swayamdipta, Rowan Zellers, John Thickstun, Sean Welleck, Yejin Choi, and Zaid Harchaoui. The researchers said that measuring how close machine-generated text is to human language is an important issue. Mauve is a comparison measure for open-ended text generation. It compares the learnt distribution from a text generation model to the distribution of human-written text using divergence frontiers. The team said that Mauve identifies known properties of generated text and scales naturally with model size. It correlates with human judgments, with fewer restrictions than existing evaluation metrics. Read the full paper here. Continuized Accelerations of Deterministic and Stochastic Gradient Descents, and of Gossip Algorithms By Mathieu Even, Raphaël Berthier, Francis Bach, Nicolas Flammarion, Pierre Gaillard, Hadrien Hendrikx, Laurent Massoulié, and Adrien Taylor. The paper introduces the “continuized” Nesterov acceleration in which the two separate vector variables evolve jointly in continuous time. It uses the best of the continuous and the discrete frameworks: as a continuous process. They show that discretisation has the same structure as Nesterov acceleration but with random parameters. The team provides continuized Nesterov acceleration under deterministic and stochastic gradients, with either additive or multiplicative noise. In the end, they provide the first rigorous acceleration of asynchronous gossip algorithms by using their continuized framework and expressing the gossip averaging problem as the stochastic minimisation of a certain energy function. Read the full paper here. Moser Flow: Divergence-based Generative Modeling on Manifolds By Noam Rozen, Aditya Grover, Maximilian Nickel, and Yaron Lipman. The paper introduces Moser Flow (MF) which is a class of generative models within the family of continuous normalising flows (CNF). MF also produces a CNF through a solution to the change-of-variable formula. Its model (learned) density is parameterised as the source (prior) density minus the divergence of a neural network. The divergence is a local, linear differential operator, easy to approximate and calculate on manifolds. It does not require invoking or backpropagating through an ODE solver during training. They demonstrate the use of flow models for sampling from general curved surfaces and have achieved significant improvements in density estimation, sample quality, and training complexity. Read the full paper here. Datasets & Benchmarks Best Paper Awards NeurIPS launched the new Datasets & Benchmarks track this year. The award recipients for this are: Reduced, Reused and Recycled: The Life of a Dataset in Machine Learning Research ( by Bernard Koch, Emily Denton, Alex Hanna, and Jacob Gates Foster and ATOM3D: Tasks on Molecules in Three Dimensions (by Raphael John Lamarre Townshend, Martin Vögele, Patricia Adriana Suriana, Alexander Derry, Alexander Powers, Yianni Laloudakis, Sidhika Balachandar, Bowen Jing, Brandon M. Anderson, Stephan Eismann, Risi Kondor, Russ Altman, and Ron O. Dror). Read the full papers here-Reduced, Reused and Recycled: The Life of a Dataset in Machine Learning Research and ATOM3D Test of Time Award The Test of Time Award went to Online Learning for Latent Dirichlet Allocation by Matthew Hoffman, David Blei, and Francis Bach. The research team developed a variational Bayes (VB) algorithm for Latent Dirichlet Allocation (LDA). It is based on online stochastic optimisation with a natural gradient step and converges to a local optimum of the VB objective function. It comes with the capability of analysing large amounts of document collections, including those arriving in a stream. The paper studies the performance of online LDA by fitting a 100-topic topic model to 3.3 million articles from Wikipedia in a single pass. They show that online LDA finds topic models as good or better than those found with batch VB and does this in a fraction of the time. Read the full paper here.","excerpt":"Let us take a look at the recipients of the 2021 Outstanding Paper Awards, the Test of Time Award, and the new Datasets and Benchmarks Track Best Paper Awards.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Data Science","Google","Machine Learning","NeurIPS"],"author_name":"Sreejani Bhattacharyya","publish_date":"2021-12-13T12:00:00","publication_year":"2021","word_count":1119,"keywords":["Go","machine learning","AI","neural network","Machine Learning","NeurIPS","RAG","BERT","llm_models:BERT","deep learning","ViT","Google","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","RAG","R","Go","BERT","ViT","llm_models:BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-winning-papers-at-neurips-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10123334,"title":"How a Traditional HCM Giant is Powering Workplaces with GenAI","content":"In today’s fast-paced business environment, leading human capital management (HCM) providers are harnessing generative AI to efficiently manage a large workforce. This technology helps automate routine tasks and deliver actionable insights for better decision-making. ADP, an HCM and payroll giant, is at the forefront of this AI-driven revolution. The company operates in 140 countries, providing payroll, HCM, and networking solutions. “In the US, we pay one in five private sector employees, processing payroll for approximately 32 million employees. So, it’s huge,” said Srinivas Konidena, CTO and VP of APAC products at ADP, in an exclusive interview with AIM. The company is leveraging generative AI to automate repetitive HR tasks and payroll, predict attrition, and provide personalised employee experiences. AI in HCM “We have something called ADP Data Cloud. It is more like a concept of a lake where all the systems, whether it is down market, mid market, or up market, they all flow the data into this data cloud. And, on top of which, we would like to use AI to generate some interesting analytics, interesting things,” he explained. Among the different products, ADP’s attrition prediction model, not only predicts attrition rates but also suggests where to find replacement candidates and the expected salary range. “When you have attrition in an organisation, the general expectation immediately after, is to predict attrition. So, we had some models to predict attrition based on the industry, based on how things are going in the company and things like that,” Konidena shared. Wary of ChatGPT and Copilot “We are very paranoid about data. Because of the business we are in, clients trust our data,” Konidena stated. To ensure data security, ADP has banned the use of public AI tools such as ChatGPT within the company. Instead, they utilise LLMs that are captive inside ADP’s systems, with no external connections. The use of copilots, such as GitHub Copilot and Amazon Code Whisperer, requires approval from their central office called GAIN (Generative AI Now) office. “One thing clear is, we do not allow the data to go out, even when we use copilots,” Konidena explained. “What we’re doing is we’re restricting the copilot to focus only on the IDE and not go back to the repo.” Currently, ADP is not developing new AI models but rather reusing existing models from providers such as AWS, Azure, and Facebook. Konidena also revealed that ADP is working with Microsoft to create a separate ChatGPT instance for the company, allowing them to use the technology more actively while maintaining data security. Unlike Zoho and Others Besides ADP, platforms like Rezolve AI and Zoho are also leveraging AI to transform HR workflows. AI-powered service desks handle repetitive tasks, such as password resets and software installations, allowing human agents to focus on complex issues requiring empathy and critical thinking. Zoho has also implemented AI models on its platforms, including AI chatbots for query resolution related to organisational policies and an AI-based document processing technology called ‘IDP’ for extracting information from documents. However, ADP differentiates itself by primarily focusing on specialised HCM and payroll solutions, whereas companies like Zoho offer a single system for accounting and other business functions. “Their focus is on entering into a single system for accounting and work and things like that. On the other hand, we are mostly on the specialised part of it, which is HCM payroll,” said Konidena. Promising Indian Market In India, ADP is experimenting with AI to detect payroll anomalies and streamline the year-end tax submission process. “The ability to accept change and the rate of innovation in India, unbelievable. And I see the same in China. So accepting of change. Then what happens is our ability to rapidly develop the product is good,” Konidena remarked. The company is also expanding by allocating dedicated funds for emerging technologies at the corporate level. “We are co-investing using ADP ventures. So I think that will give us some leverage more than anything, I think it will give an insight into what’s going on,” said Konidena. ADP sees significant potential in tier-two and tier-three cities in India, where the standard of living is improving, and companies are establishing operations. “Those are the engines of growth. We’re also seeing all the manufacturing, right? Earlier used to be centred around these places, they’re shifting out. And when they shift out, they’re not getting regular manual labour. I mean, it’s all fully automated supervisors,” Konidena pointed out. ADP’s growth has been driven by both organic expansion and acquisitions. In India, ADP acquired Ma Foi Randstad’s payroll business, which had around 100 clients at the time. Over the past 12 years, ADP India has grown to serve close to 2,000 clients.","excerpt":"ADP, an HCM and payroll giant, is at the forefront of this AI-driven revolution.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-12T11:34:10","publication_year":"2024","word_count":781,"keywords":["ChatGPT","AWS","AI","chatbots","ML","RAG","Aim","generative AI","analytics","copilots","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","ChatGPT","Aim","RAG","copilots","chatbots","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-a-traditional-hcm-giant-is-powering-workplaces-with-genai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":4102,"title":"Big Data Tech Conclave 2013 Winter Edition Overview","content":"The Big Data Tech Conclave Winter Edition will be an ideal forum to discuss the way forward in addressing the key issues, and we look forward to see you on board an actively share, benchmark and discuss the emerging technologies, tools and challenges and identify the immense opportunities available in Big Data for you and your organization. The winter edition will feature presentations, workshops, case studies and key insights into Algorithms and Machine Learning, Analytic Insights, Architectural Patterns, Data Management, Design\/Interface and Visualization, Enterprise IT, Testing, Privacy, Compliance and Governance. Key emphasis will be provided in areas such as parallel computing innovations in Big Data, Harnessing and inferencing the Semantic Web, SQL on Haddoop and blending the traditional RDBMS and NOSQL databases with a special focus on workshops around some key tools \/ technologies in this space such as Impala, Parallel Computing with R, Stinger, Spark\/Shark, Gehi, Drill, Apache Mahout and SparQL. We would like to take this opportunity to thank all our Speakers, Sponsors and Delegates who made the spring edition of the Big Data Tech conclave 2013, which was held on 26 & 27 April 2013 at The Marriott Whitefield Bangalore, a thumping success. Big Data Tech Conclave 2013 saw over 300 delegates participate from enterprise, Technology, ITES, Big Data Start ups, venture capitals and Data centers, reiterating that India is the place to be, if you want to take Big Data to the next level. We thank all our esteemed participants for the valuable feedback and suggestions that we have received in helping us design the winter edition of the Big Data Tech Conclave 2013 to be held on the 6th to the 7th of December 2013. For more information write to bhasker.gupta01@gmail.com [attachments docid=”4105″]","excerpt":"The Big Data Tech Conclave Winter Edition will be an ideal forum to discuss the way forward in addressing the key issues, and we look forward to see you on board an actively share, benchmark and discuss the emerging technologies, tools and challenges and identify the immense opportunities available in Big Data for you and […]","categories":["Deep Tech"],"tags":[],"author_name":"AIM Media House","publish_date":"2013-09-24T15:34:15","publication_year":"2013","word_count":288,"keywords":["big data","Go","API","machine learning","AI","innovation","venture capital","SQL","GAN","R"],"extracted_tech_keywords":["AI","machine learning","R","SQL","Go","API","big data","GAN","innovation","venture capital"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/big-data-tech-conclave-2013-winter-edition-overview\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10053780,"title":"Google AI Releases Data Model To Predict Text Readability From Scrolling Interactions","content":"Google AI recently announced the release of its data model for predicting text readability from scrolling interactions. The new data model shows that data from on-device reading interactions can be used to predict how readable a text is. This novel approach provides insights into subjective readability — whether an individual reader has found a text accessible — and demonstrates that existing readability models can be improved by including feedback from scroll-based reading interactions. In order to encourage research in this area and to help enable more personalized tools for language learning and text simplification, Google AI also released the dataset of reading interactions generated from scrolling behaviour–based readability assessment of English-language texts. Traditional machine learning approaches to measure readability have exclusively relied on such linguistic features. However, using these features alone does not work well for online content because such content often contains abbreviations, emojis, broken text, and short passages, which detrimentally impact the performance of readability models. To address this, Google investigated whether aggregate data about the reading interactions of a group can be used to predict how difficult a text is, as well as how reading interactions may differ based on a readers’ understanding. When reading on a device, readers typically interact with the text by scrolling in a vertical fashion, which was hypothesized can be used as a coarse proxy for reading comprehension. They recorded the reading interactions by measuring different features of the participants’ scrolling behaviour, such as the speed, acceleration, and the number of times areas of text were revisited. This information was then used to produce a set of features for a readability classifier. The research team tested the significance using linear mixed effect models. Using linear mixed-effect models gives us higher confidence that the differences in interactions we are observing are because of the text difficulty and not other random effects. The results showed that multiple reading behaviours differed significantly based on the text level, for example, the average, maximum and minimum acceleration of scrolling. It was found that the most significant features were the total read time and the maximum reading speeds. These features were then used as inputs to a machine learning algorithm. The team designed and trained a support vector machine (i.e., a binary classifier) to predict whether a text is either advanced or elementary based only on scrolling behaviours as individuals interacted with it. The dataset on which the model was trained contains 60 articles, each of which were read by an average of 17 participants. From these interactions, we produced aggregate features by taking the mean of the significant measures across participants. Image Source: Google AI The accuracy of the approach was measured using a metric called f-score, which measures how accurate the model is at classifying a text as either “easy” or “difficult” (where 1.0 reflects perfect classification accuracy). We are able to achieve an f-score of 0.77 on this task, using interaction features alone. This is the first work to show that it is possible to predict the readability of a text using only interaction features. It was also found that the addition of interaction features improves the f-score of this model from 0.84 to 0.88. In addition, the team were able to significantly outperform this system by using interaction information with simple vocabulary features, such as the number of words in the text, achieving an impressive f-score of 0.96. Image Source: Google AI Such an understanding is crucial when designing educational applications for low-proficiency readers and language learners because it can be used to match learners with appropriately levelled texts as well as to support readers in understanding texts beyond their reading level. Image Source: Google AI Through this, the research team confirms that there are statistically significant differences in the way that readers interact with advanced and elementary texts and that the comprehension scores of individuals correlate with specific measures of scrolling interaction.","excerpt":"This novel approach provides insights into subjective readability, whether an individual reader has found a text accessible and demonstrates that existing readability models can be improved by including feedback from scroll-based reading interactions.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Deep Learning","Google","Machine Learning","NLP","NLP models","Python","Support Vector Machine","text analysis"],"author_name":"Victor Dey","publish_date":"2021-11-18T19:08:35","publication_year":"2021","word_count":649,"keywords":["Go","machine learning","programming_languages:R","AI","Support Vector Machine","Machine Learning","programming_languages:Go","NLP models","RAG","NLP","Python","Google","text analysis","Deep Learning","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-ai-releases-data-model-to-predict-text-readability-from-scrolling-interactions\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10051656,"title":"To Be A Data Scientist, You Must Be A Software Engineer: Antrixsh Gupta, Danalitic","content":"“Apart from their mastery of technology and long-term vision, I treat business processes as programming challenges,” says Antrixsh Gupta, CEO & Chief Data Scientist at Danalitic. Analytics India Magazine caught up with Antrixsh to learn more about the research and data science industries in general. Antrixsh spoke on data science, its current state, and future predictions. AIM: What drove you from senior software engineer to technical advisor to mentor to a researcher to CEO? Who or what encouraged you to pursue a career in data science? Antrixsh Gupta: I was always fascinated by science and technology. I was always interested in knowing how it works. My ultimate goal was always to have my own company, develop new resources and make big profits. As I was fascinated with science and technology, I always wanted to work on new things and wishfully guide the learners. I would also say that some lucky breaks got me where I am today. I believe I had the ability to approach business processes like programming tasks apart from their understanding of technology and long-term vision. The variety of skills that I could learn along the way was one of the major reasons that encouraged me to pursue my career in data science. I was always interested in new market opportunities, and the skills in data science have helped me identify those patterns. AIM: A researcher who became a CEO is an interesting metamorphosis. How did this transition take place? Antrixsh Gupta: I would say that the transition wasn’t a piece of cake. As I already mentioned, honestly, my journey was paved with some lucky breaks. There were some personal motivations that took me from the lab to the boardroom. Being a CEO, you have to make loads of decisions without enough information, which could be challenging and here comes the role of the experiences you have had over the past years. I didn’t pay major attention to money; instead, I focused on my company’s developments and the journey I am onto. AIM: Is there a specific incident in your data science journey which you look back on with pride? Antrixsh Gupta: This has happened to me very often, as I am involved myself in training and consulting and many times when my students call me and tell “Antrixsh, I interviewed your student”, and I feel very proud as both the interviewer and the interviewee are my students. And I must say, where we are now, it happened just because of my students. Also, when my students call me\/text me that we got our dream job, it makes me very proud. AIM: It is great that being a machine learning facilitator, you have organized hackathons. You would have solved and witnessed various key challenges in the industry. Could you please share one of those interesting challenges with us? Antrixsh Gupta: Being ML facilitators, we conduct almost 50+ hackathons, and our main motive has been to conduct hackathons to help participants keep pursuing their goals. There are so many resources online and can sometimes be overwhelming, so helping participants choose wisely helps them develop their learning path individually, helping them by using hands-on projects on healthcare, retail, and the supply chain domain. It also helps the participants to gain domain knowledge by inviting industry experts. AIM: What, in your opinion, is the most demanding area of research in the data science industry at the moment that requires attention? Antrixsh Gupta: The healthcare sector represents one of the most important industries for data scientist involvement. Not only does an estimated 30 per cent of the world’s warehoused data come from the medical field, but the opportunities for improvement made possible by this cache could save the industry as much as $300 billion annually. Working in the healthcare industry as a data scientist means more than just efficiency improvements — it can mean lives saved. Because of this, data scientists are flocking to this humanitarian industry. AIM: What are your perspectives on the Indian data science industry’s growth? Could you share your insights with us? Antrixsh Gupta: It is, without doubt, a highly growing industry. The hiring in this industry has increased by 45%, and there will be more than 10 million job openings in the coming five years. The impact of the data science sector is far-reaching, and as a result, a range of new roles and skillsets will be in demand. It is estimated that the industry is growing at a healthy rate. The industry is expected to grow seven times in the next seven years. Startups have also contributed significantly to the overall output in India. AIM: Is India’s data science industry losing its competitive edge in terms of cost? With the cost of recruiting data scientists increasing, will businesses continue to outsource? What is your opinion? Antrixsh Gupta: I would not say that the industry is losing its competitive edge in the country. Instead, data science applications have amassed all the industries and have readily increased the demand for data scientists. But the trends are changing nowadays. The demand is no longer the same as before. Even if there is a demand for data scientists, people lack either the skill set or the experience. Transitioning to data science is a smart move as it fetches far higher comparative returns. For the majority of businesses, outsourcing will be the right option as it will grant a high degree of flexibility and scalability. It will also lessen the burden of finding the right resources in a highly competitive industry. AIM: In the industry, Who are your role models? Antrixsh Gupta: Andrew Ng and Yann LeCun are two personalities that I look up to as role models in this industry. AIM: What are your favourite data science\/AI books? Antrixsh Gupta: These three count for my favourite: Competing in the Age of AI: Strategy and Leadership When Algorithms and Networks Run the World.Building Analytics Teams: Harnessing analytics and artificial intelligence for business improvement.The Data Detective: Ten Easy Rules to Make Sense of Statistics. AIM: As a researcher and data science practitioner, what advice would you give to aspirants interested in pursuing a career in data science? Antrixsh Gupta: My major piece of advice to the aspirants would be to start with practical projects and then slowly progress with theory. Ask questions till you become satisfied with your knowledge and tools. One can also find and join online communities that can help them learn and grow. Regularly read blogs, and listen to podcasts of industry experts. And for all the fresh minds, I would like to give one suggestion: before becoming a data scientist, you should first be a good software engineer. Don’t let yourself think you will be a master in a week or month, but a consistent approach to learning new things daily will definitely help you achieve your goal.","excerpt":"Antrixsh Gupta is the CEO of and a Chief Data Scientist at Danalitic. He comes with vast experience solving real-world business problems in a variety of industries.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Andrew Ng","Data Science","Data Scientist","Hackathons","Interviews and Discussions","Machine Learning","programming","software engineer","Yann LeCun"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-10-15T18:00:00","publication_year":"2021","word_count":1141,"keywords":["TPU","R","data science","artificial intelligence","Andrew Ng","RAG","software engineer","analytics","Data Science","machine learning","AI","ML","Machine Learning","Yann LeCun","programming","Hackathons","Aim","Data Scientist","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chief-data-scientist-at-danalitic\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10118387,"title":"Deloitte Inaugurates 4th Office in Bengaluru","content":"Deloitte has expanded its presence in Bengaluru, India, by opening a new office, its fourth in the city. Located in Yemalur Village, Marathahalli, this office is equipped to house 6,000 professionals and is part of Deloitte’s strategic efforts to grow and expand in the region. This new facility will focus on serving global clients and adds to the existing trio of offices in the city, drawing on Bengaluru’s talent pool and infrastructure. Register for the workshop on Automating Data Pipelines >> Members at the new office will work across various domains including artificial intelligence, data analytics, cybersecurity, cloud services, and more. The office, inaugurated by Lara Abrash, Chair of Deloitte US, features technological setups like an XR Studio and Innovation Labs. In March, the company inaugurated three new workplace hubs in Bengaluru Noida, and Pune. It has offices in 14 locations (cities) in India, including Bhubaneswar, Coimbatore, Kochi, and Jamshedpur. In January, Deloitte stated that it is training over 120,000 employees through its AI Academy. Additionally, the company is investing over $2 billion in global technology learning initiatives aimed at improving skills in AI and related fields. It also announced a substantial $2 billion investment in the IndustryAdvantage program to improve industry-specific services by integrating generative AI into Deloitte’s solutions. Additionally, Deloitte is expanding its suite of generative AI-enabled accelerators and enhancing its cloud-native platform, Converge. .","excerpt":"In March, the company inaugurated three new workplace hubs in Bengaluru Noida, and Pune.","categories":["AI News"],"tags":["Deloitte"],"author_name":"Shritama Saha","publish_date":"2024-04-17T13:40:21","publication_year":"2024","word_count":227,"keywords":["artificial intelligence","Deloitte","programming_languages:R","AI","innovation","data pipeline","Aim","generative AI","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","generative AI","Aim","R","data pipeline","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deloitte-inaugurates-4th-office-in-bengaluru\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":12012,"title":"Flashback 2016: All The News Story On Analytics That Made Headlines In India","content":"Another fruitful year for the industry marks an end. We saw start-ups being funded, companies being acquired, the various collaboration and partnerships, organizations and companies adopting analytics and more. But most significant addition to the list this year has been the inclusion of technologies like artificial intelligence, chatbots and robotics that made the headlines quite often. After sifting through all the news that impacted the analytics industry in some way or the other, we bring to you the biggest developments that hit this sector in the year 2016. [heading size=”25″]Funding and Investments[\/heading] Jan Ratan backs Tracxn: Ratan Tata invests an undisclosed amount in Tracxn, a startup research and analytics firm. Feb Qubole closes funding worth $30 million: It was led by IVP and existing investors CRV, Lightspeed Venture Partners and Norwest Venture Partners. A host of Marquee Angels back Tracxn: Start-up data analytics firm, Tracxn Technologies Pvt. Ltd raised yet another funding through marque Angel Investors including former UIDAI chairman Nandan Nilekani, Mohandas Pai (Co-Founder, Aarin Capital), Neeraj Arora (Vice President, Whatsapp) and Anand Rajaraman (Co-Founder, Junglee), for an undisclosed amount. Mar Imarticus gets seed funded: Imarticus Learning, a financial services and analytics training institute, announced the completion of its primary funding process with an undisclosed amount. It received support from existing investors like Amit Nanavati, as well as new investors like Tashwinder Singh, Director, KKR, Anil Gudibande, 1 Crowd Founder\/Ex MD RBS, Taranjit Jaswal, Director, Barclays and Amit Khanna, ex KPMG Partner. Apr vPhrase Analytics Solutions raises seed funding: vPhrase, an artificial intelligence-enabled data analytics firm, raised an undisclosed amount of seed funding from Mumbai-based Venture Catalysts (VCats). May Fractal Analytics raises USD 100 million: Fractal Analytics, a leading global provider of analytics, entered a partnership with Khazanah Nasional Berhad, the strategic investment fund of the Government of Malaysia, who will invest up to USD 100 million in the company to accelerate its growth. Jun BRIDGEi2i raises funds from Edelweiss Private Equity: BRIDGEi2i Analytics Solutions, a Bangalore based analytics solutions company, secured Series A growth capital from one of India’s leading diversified financial services conglomerate- Edelweiss Private Equity, the private equity and venture capital arm of Edelweiss Financial Services. RIL invests in NetraDyne: Reliance Industries Limited (RIL) invested US$16 million (approximately Rs. 108 crore) in NetraDyne, a US-based visual analytics start-up. Aug Mezi raises $9 million: Mezi, a leading chatbot shopping app received funding in Series A Round from investors to accelerate AI bots technology. Investors in this round included previous investor Nexus Venture Partners and new investors Saama Capital and American Express Ventures. Innovaccer raises 15.6 million: Big Data Startup Innovaccer raised 15.6 Million in Series A funding led by Westbridge Capital Partners along with other angel investors. The company intended to use this funding to accelerate its footprint in healthcare. Tricog Health raises Series A funding: HealthCare Analytics firm Tricog Health raised undisclosed funding from Inventus Capital Partners, Blume Ventures and a host of angel investors. It was aimed for further developing the technology and market reach of their path breaking healthcare analytics platform. HR Analytics Startup inFeedo raise $150,000 in Angel Round: This $150,000 round was led by Dheeraj and preceded by Palash Jain, Ex-Head of Google India Core Operations. Sep AI start-up Octo.ai raises 200,000$: Octo.ai, an open-source analytics platform built for machine learning raised fund in seed round led by Rohan and Arjun Malhotra (Investopad), Rahul Khanna (Trifecta Capital), Rakesh Agrawal (US based- seed in Lyft, Cruise, Shyp, Poshmark and Lendup), Sidharth Rao (WebChutney), Outbox Ventures, Rajan Navani, Jaspreet Bindra, Gagan Duggal. Oct Hyderabad Angels backs Realbox: Realbox Data Analytics, a data science startup, raised pre-series A round funding of amount $300K led by Alok Mittal, ex-MD, Canaan Partners, Dharamveer Singh Chouhan, Zostel CEO, Shubranshu Pani, Director JLL and vCommission. Nugit raises US$5.2M: Singapore-based startup Nugit, AI and Data Analytics startup announced a US$5.2M funding round from Sequoia India. It combines AI, natural language generation, and visual design to transform data into decision-ready reports that integrate seamlessly into marketers’ workflows. SIBIA Analytics raises seed investment: SIBIA Analytics, a Kolkata based start-up in predictive analytics, raised an undisclosed amount in seed investment from a group of investors from India and USA including Radhakrishnan Natarajan and other prominent leading IT and Banking professionals. Nov Aarin Capital invests in Impact Analytics: Impact Analytics, a global business intelligence and data science insights provider for organizations across sectors such as retail, CPG, financial services etc. raised a seed round of $750,000 led by Aarin Capitals, co-founded by T. V. Mohandas Pai and Dr. Ranjan Pai. Ashish Lakhanpal, CEO of Kismet Capital and Michael Herzig, serial entrepreneur, were other co-investors in the startup. Pramod Bhasin backs Scienaptics: Scienaptics, a new age analytics platform received strategic investment from senior industry leader and former Nasscom Chairman, Pramod Bhasin. ACTIVE.AI raises $3 Mn: Active.ai, a Singapore based fintech startup announced raising a fund of $3 million from leading venture capital funds IDG Ventures India and Kalaari Capital to expand its AI platform for Banking & Payments. Dec THB secures an undisclosed seed funding: The Healthy Billion, (THB) a clinical research and data analytics start-up reportedly secured an undisclosed seed funding raised by Ajith Sukumaran (formerly with Nokia) and Apoorva Patni, son of Patni co-founders Ashok Patni who has set up Currae Healthtech Fund. Reportedly, Kanwaljit Singh, former MD of Helion Ventures and a current investor also participated in this funding round. [heading size=”25″]Collaboration and Partnership[\/heading] Jan Simplilearn partners with Tableau: Simplilearn and Tableau, one in education and the other in Business Intelligence respectively, entered into a partnership with an aim to build a talent pool of 200,000 skilled Data Scientists by the year 2020. June The Boston Consulting Group (BCG) and Mu Sigma announced a partnership that will enable their clients to leverage cutting-edge big data and analytics solutions to transform the way they do business. Mu Sigma’s integrated ecosystem of 3,500 decision scientists, along with its unique Art of Problem Solving methodologies and innovation platforms, will complement BCG’s existing strategy, technology, and analytics capabilities, further enhancing their mutual position in the market for global change management and data-driven decision making. May Infosys partners with Trifacta: Infosys, the Indian IT giant came into a partnership with Trifacta, the global leader in data wrangling to provide a data wrangling solution for the Infosys Information Platform (IIP) and Infosys’s other platforms and offerings. This partnership has culminated into funding, joint development, integrations and solution collaboration. [heading size=”25″]Acquisition[\/heading] Mar Millward Brown acquires Analytics Quotient (AQ): Millward Brown, a global leader in brand, media and communications research acquired the business operations of Analytics Quotient (AQ), a marketing analytics company that extracts insights from data to help clients define their marketing strategies. May Snapdeal acquires TargetingMantra:  Snapdeal acquired a boutique technology firm, ‘TargetingMantra’ to further to integrate machine learning based solutions on its platform. Jul Synnex acquires Bangalore based Minacs: Synnex, a US based Fortune 500 corporation, came to a definitive agreement to acquire Bangalore based Minacs, a leading outsourcing business solutions in the field of manufacturing, retail, telecom, technology, media and entertainment, banking, insurance, healthcare, and public sectors. The deal valued at $420 million was made to boost up analytics space. Hansa Cequity acquires Bangalore based D-Square: As a part of the deal signed between Hansa Cequity, India’s leading & largest customer marketing company and D-Square Solutions Private Limited, a data sciences and analytics company based out of Bengaluru, Hansa Cequity acquired a majority stake in D-Square. It aims at scaling the breadth of analytics offerings and entering artificial intelligence & machine learning capabilities. Nov Freshdesk acquires a Chatbot Platform Chatimity: The acquisition aims to enhance Freshdesk’s capabilities within chat platform and beyond. This comes as the company’s sixth acquisition following its stint with Airwoot, Framebench, 1CLICK.io, Konotor and Frilp within a time span of 12 months. L&T Infotech to Acquire AugmentIQ: Leading global IT services and solutions provider L&T Infotech announced that it will acquire Pune based AugmentIQ Data Sciences Pvt. Ltd., an innovative start-up offering IP-based, big data and analytics solutions that helps enterprises derive business benefits from big data. The deal will enrich and expand LTI’s high-end analytics offerings across industries. GE Digital acquires two AI startups: Legacy company General Electric (GE) recently acquired two major startups – California based Wise.io and Vancouver based Bit Stew Systems to bolster their AI capability. While the former specializes in machine learning, the latter is focused on data analytics. Dec Practo acquires Enlightiks: Practo the healthcare company, known for providing end-to-end healthcare solutions has made its third major investment by acquiring Enlightiks, healthcare analytics firm that provides business intelligence solutions and advanced analytics. [heading size=”25″]AI, Chatbot and other technologies[\/heading] Jul ICICI Bank adopts Blockchain and Artificial Intelligence: As part of its annual reorganization, ICICI Bank’s Managing Director and CEO, Chanda Kochhar, announced to create a new division, Technology and Digital group (TDG). This new team will be responsible of improving the digital capabilities of the bank. Aug Karbonn Mobiles forays into AI: Karbonn Mobiles, India’s leading handset player becomes the first mobile handset player to foray into the field of artificial intelligence. The company is the first to introduce artificial intelligence in areas of fashion and lifestyle with the launch of its smartphones ‘Fashion Eye & ‘Fashion Eye 2.0’. Narayana Health launches Robotic Surgery Institute: Narayana Health, one-stop healthcare destination, has launched the Institute of Robotic Surgery supported by Infosys Foundation at its flagship unit at Narayana Health City. It will be using the da Vinci Robotic Surgical System primarily for prostate, kidney, gynecological, colorectal and select head & neck cancer surgeries. Oct Book an OYO room using Niki bot: OYO, India’s largest branded network of hotels has entered into a partnership with niki.ai, enabling users to book an OYO room using niki.ai’s bot – a unique personal assistant application powered by artificial intelligence. Nov India’s first humanoid Banker – Lakshmi: The friendly robot powered by Artificial Intelligence (AI) made its debut on the big stage recently at the City Union Bank based in Chennai, Tamil Nadu. It can reportedly chat with customers on more than 125 subjects, field queries on current interest rates on loans, account balance and transactional history. Dec Meet Haptik, virtual assistant powered by AI: Personal assistant app and conventional commerce platform powered by a combination of machines and humans, Haptik launched a new version 5.0 of the app. This chat-based app help users perform daily tasks as well as facilitate multiple end-to-end transactions on a single platform. Jinie- India’s first HR Chatbot: Jinie is here to make the whole process of HR management more engaging and impactful. It is brought to you by Peoplestrong – one of the India’s leading human resource (HR) solutions and HR Technology Companies in the world. Microsoft India introduces AI in Eye Care: In collaboration with L V Prasad Eye Institute, Microsoft has launched Microsoft Intelligent Network for Eyecare (MINE) to bring technology in the foreplay. [heading size=”25″]Analytics adoption and launches[\/heading] Mar Times of India Group turns to analytics: India’s largest media and entertainment company, The Times of India Group turns to analytics to power its editorial analytics engine and driving deeper user engagement. Apr Teradata launches Analytics of Things Units: The big data analytics and marketing applications company, Teradata, revealed plans to build IoT Analytics units based in India, United States and United Kingdom. Jun Snapdeal establishes Data Science Center: With their announcement of establishing a Data Sciences Center in San Carlos, California, they intend to study customer behavior and strengthen the supply chain. Aug JCPenney’s launches “Global in-house centre” in Bangalore: One of the largest apparel and home furnishing retailers of the United States, JCPenney, announced its entry into the Silicon Valley of India, joining the league of other global retailers like Target, Walmart and Tesco. Noodle.ai launches operations in India: Silicon Valley based Noodle.ai and former co-founders of Infosys Consulting announced the launch of Noodle Analytics Private Limited in Bangalore, India. The venture is funded by TPG, a leading private equity firm, and the founders of the company. Sep CA Technologies Launches ‘CA App Experience Analytics’: CA Technologies recently launched CA App Experience Analytics, a new SaaS solution in India. This app enables organizations to deliver a premium customer experience across the increasing number of digital channels – Web, Mobile and Wearables – used by today’s consumers. [heading size=”25″]Government Initiatives[\/heading] Aug TRAI introduces TRAI Analytics Portal: With the launch of a new portal called “TRAI Analytics Portal”, it aims at providing subscribers with all the information about the performance and quality of services being offered by their telecom operators like network coverage, data speed, call drop rates to name a few. Sep CAG of India sets Big Data Analytics center: A brainchild of Shri Shashi Kant Sharma, this center aims to analyze big data of the government and provide valuable insights to the government. This data analytics center will synthesize and integrate all the appropriate data required for auditing. Oct Ministry of Commerce & Industry launches a Dashboard for Foreign Trade Data: This initiative by the ministry is to provide easy access of India’s import, export, and balance of trade data to the people of India. The data on the dashboard is represented in an analytical format, over time and space. Dec Tax Official use Big Data and Analytics to combat black money menace: The Income Tax department will reportedly use big data analytics to comb through personal bank deposits money holders. [heading size=”25″]Others[\/heading] The various developments at Mu Sigma: It was quite an eventful year for Mu Sigma, the world’s largest pure-play provider of decision sciences and analytics solutions. The year began with the appointment of Ambiga Dhiraj as the new CEO in Feb 2016 who was formerly a Chief Operating Officer. After the news of the founders filing for divorce surfaced, there were also speculations of the Mu Sigma CEO and other investors such as General Atlantic looking to sell their stakes in the company. However, after sailing a turbulent ride, the year 2016 settled for Mu Sigma by having a new CEO- Dhiraj Rajaram, who got a majority ownership in the data analytics company after Ambiga Subramanian agreed to sell her stake.","excerpt":"Another fruitful year for the industry marks an end. We saw start-ups being funded, companies being acquired, the various collaboration and partnerships, organizations and companies adopting analytics and more. But most significant addition to the list this year has been the inclusion of technologies like artificial intelligence, chatbots and robotics that made the headlines quite […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2016-12-30T05:36:49","publication_year":"2016","word_count":2361,"keywords":["data science","machine learning","artificial intelligence","AI","chatbots","ML","RAG","Ray","Aim","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","Ray","RAG","chatbots"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/flashback-2016-news-story-analytics-made-headlines-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10163015,"title":"40 Under 40 Data Scientists Awards 2025 &#8211; Meet the Winners","content":"Amidst the three days of AI and ML workshops, conferences, presentations, and tech talks at the Machine Learning Developers Summit (MLDS) 2025, about 40 dynamic data scientists were presented with the 40 Under 40 Data Scientists Award on Thursday. This award recognises India’s top data scientists and their achievements in the machine learning and analytics industry. This year’s winners are driving real impact at some of the world’s most influential companies, including Razorpay, HSBC, Genpact, PepsiCo, Bloomberg, Ford Motors, Paytm, Tata, Wells Fargo, Accenture, and more. They are creating AI solutions that improve efficiency and developing data models that prioritise privacy. More than just driving innovation, they are also fostering a culture of learning and growth. The Winners of the 40 Under 40 Data Scientists Awards 2025 Abhinav Vajpayee, Senior Manager, Analytics at Razorpay Software Private Limited A hands-on innovator in data-driven business strategy, Abhinav takes part in optimising payments, ads, and content acquisition. His solutions at Razorpay, Swiggy, and Vuclip boosted retention, ad revenue, and cost efficiency, earning industry recognition. Abhishek Kumar, VP, Analytics Lead at HSBC Abhishek is a data science leader known for high-impact analytics solutions across banking, FMCG, and retail. His work spans forecasting, pricing models, and customer insights, earning multiple awards for innovation and business impact. Akhil Makol, Principal Engineer at NatWest Group Akhil leads the data strategy and cloud architecture for commercial & institutional domains. He drives AI and analytics adoption by aligning data products with banking standards and leveraging AWS Data Lake. Akshay Jain, AGM – Lead Digital Downstream at Hindalco Industries Ltd Akshay Jain is a data scientist transforming aluminium manufacturing with AI, predictive analytics, and Industry 4.0. He leads a team that optimises operations and drives AI adoption on the shop floor, focusing on people-centric implementation. Amresh Kumar, General Manager at Niva Bupa Health Insurance Amresh, a data scientist with more than 15 years of experience, has worked across insurance, banking, and digital marketing. He has successfully implemented renewal, planning, and reinsurance models, with certifications in advanced insurance and Google Analytics. Ankit Sati, Senior Manager at Genpact Ankit is a vital member of Genpact’s AI\/ML practices. He specialises in computer vision and GenAI solutions. With more than seven years of experience across industries, he is also an active Kaggle competitor and hackathon enthusiast. Anup Kumaar Goenka, Deputy Director of Data Science at PepsiCo. Anup is a data science innovator known for AI-driven solutions in forecasting and automation. He developed an award-winning AI meeting summarisation tool and led predictive analytics projects optimising supply chains and financial planning. Anupam Tiwari, Data Science Manager at GoTo Company Anupam contributed to GenAI for Southeast Asian languages, developing Sahabat AI, a suite of LLMs for Indonesian dialects. His models, openly available on Hugging Face, support AI innovation and adoption in Indonesia. Arjit Jain, Co-Founder and CTO at TurboML Arjit is an ML researcher specialising in real-time machine learning for fraud detection and personalisation. A former Google researcher and IIT Bombay graduate, he has published award-winning papers with more than 200 citations. Avinash Kanumuru, Senior Manager – Data Science & Engineering at Niyo Avinash is a data science contributor who publishes acclaimed articles on platforms like ‘Towards Data Science’. He has also developed open-source Python libraries (ml-utils, pyspark-utils), simplifying ML workflows and enhancing industry best practices. Debanjan Mahata, Senior ML Research Engineer at Bloomberg Debanjan Mahata is a leading researcher in NLP, machine learning, and Document AI, with publications in top conferences and patented innovations in document analysis. His recent work focuses on multimodal Retrieval-Augmented Generation (RAG) for DocVQA, enhancing financial and ESG data extraction. Dr. Shital Patil, Solution Architect at Robert Bosch Dr. Shital, a Prime Minister’s Fellowship recipient, specialises in AI-driven machinery condition monitoring and predictive maintenance. With four patents, she excels in fault diagnosis, XAI research, and solution architecture across India and the Middle East. Dr. Vikram Singh, Senior Vice President of AI & Digital at EightBit AI Private Limited Dr. Vikram is a researcher specialising in image super-resolution, deblurring, and deep learning. His work includes high-frequency refinement techniques and advanced neural networks for sharper image and video processing. Gaurav Mhatre, Director at Tiger Analytics Gaurav, a data science leader with 13+ years of experience, drives AI innovations across CPG, healthcare, telecom, and eCommerce. At Tiger Analytics, he has led route-to-market analytics, pet food optimisation, and 5G network routing using cutting-edge AI algorithms. Gopinath Chidambaram, Global Technical Director, AI\/ML & Cloud at Ford Motors Gopinath, an AI\/ML professional, holds five patents in autonomous vehicle perception and has filed two more in AI-driven monitoring systems. He is also co-authoring Gen AI Untrained, an upcoming book on AI concepts and applications. Kantesh Malviya, Associate Vice President – Analytics at Paytm Kantesh is a data science leader known for mentorship and thought leadership. He has spoken at IIT Bombay and developed innovative analytics solutions, driving user engagement and revenue growth across industries. Kulbhooshan Patil, Head of Data Science and Analytics at TATA AIG General Insurance Company Kulbhooshan is an award-winning AI leader recognised for innovation in risk management and user experience in insurance. His AI-driven solutions have earned multiple industry accolades, including the Best AI Technology Implementation of the Year and Outstanding AI & ML Solution Provider. Mahima Bansod, Data Science and Analytics Leader at LogicMonitor Mahima is a Data and AI leader with a decade of experience driving digital transformation at companies like Salesforce and Siemens. She has implemented ML models for customer retention, achieving a 94% renewal rate and 40% growth in product adoption. Mahish Ramanujam, Associate Director – Analytics at Games24X7 Mahish led an award-winning project, developing a game-wise adaptive user-engagement model inspired by cricket analytics. The model boosted D30 LTV by 20% while reducing spending by 15%. Mehuli Mukherjee, Vice President at Wells Fargo Mehuli, VP at Wells Fargo, is an analytics leader specialising in GenAI, LLMs, and NLP. A gold medallist and PhD researcher, she is developing an Indian Sign Language recognition system while mentoring in AI and advocating for social impact. Namit Chopra, VP-2 at EXL Service Namit developed EXL Property Insights Solution (patent pending) and applies NLP\/GenAI to insurance claims. His work includes LLM-based claim summarisation, fraud detection, and cause-of-loss identification. Namita Khurana, Data Scientist Associate Director at Accenture Namita is a leader in Revenue Growth Management (RGM), specialising in pricing, promotion, and assortment analytics across global markets. She has developed patented solutions, including AI-driven conversational tools for strategy optimisation and decision-making. Nandita Saini, Manager – AI & Cognitive Solutions at e& enterprise Nandita Saini led the productisation of e& Enterprise’s GenAI-based Utilities Copilot. She successfully turned the concept into a launched product, driving innovation. Nishant Ranjan, Head of Analytics at Godrej Consumer Products Nishant developed innovative pricing, forecasting, and AI-powered analytics models, including a first-of-its-kind promotion attribution model. He also pioneered MLOps best practices, enabling scalable machine learning deployment. Pankaj Goel, Associate Vice President – Innovations at BA Continuum India Pvt Ltd (Bank of America subsidiary) Pankaj Goel pioneered demand prediction in the CPG industry, analysing country-level demand impact on product lines, earning recognition from Procter & Gamble. He recently completed a proof of concept on digital transformation using LLM\/GenAI. Pavak Biswal, Senior Manager at Merkle Pavak, a data science leader with 13+ years of experience, has driven business impact through AI and analytics innovations. He designed GenAI-powered chatbots, optimised pricing models, and led multimillion-dollar analytics projects across industries. Pawan Kumar Rajpoot, Lead Data Scientist at TIFIN Pawan has more than 10 years of experience in NLP research and development and has been the winner of 10-plus international competitions. His previous work experience includes companies like Tact.ai, Rakuten India and Huawei. Puspanjali Sarma, Senior Manager – AI at ServiceNow Puspanjali is an AI and data science expert specialising in AI product management, NLP, and predictive analytics. Her work at ServiceNow and beyond has driven innovative, AI-driven solutions with measurable business impact. Rajaram Kalaimani, Senior Principal Data Scientist at Mindsprint Rajaram is the architect behind Mindverse, a GenAI platform, and precision agriculture solutions for the agri supply chain. His innovations are now hosted on Google, expanding AI capabilities at Mindsprint. Ritwik Chattaraj, Data Science Manager\/Senior Data Scientist at Commonwealth Bank of Australia Ritwik Chattaraj is a data science leader with expertise in Generative AI, LLMs, and robotics. He has led AI\/ML innovations at major banks, published extensively, and received multiple excellence awards. Sachin Kumar Tiwari, Deputy Vice President at Canara HSBC Life Insurance Company Sachin Kumar has led key AI and analytics projects, including GenAI chatbots, customer genomics, and sales governance models. His work spans predictive modelling, sentiment analysis, and geospatial analytics to drive business decisions. Sairam Mushyam, Head of Data and AI (SVP) at Zupee Sairam has revolutionised Real Money Gaming in India through AI-driven innovations, regulatory frameworks, and data infrastructure. His work spans blockchain-based fairness validation, user integrity systems, and GenAI-powered gaming experiences, driving $30M+ in annual revenue growth. Shravan Kumar Koninti, Associate Director – Data Science at Novartis Shravan is an AI researcher and innovator who is developing self-service AI tools and large-scale AI projects. He has won multiple hackathons, pioneered Generative AI applications, and contributed to healthcare AI advancements through collaborative research. Sumeet Pundlik, Delivery Unit Head at TheMathCompany (MathCo) Sumeet is an AI and data science expert with a patented service location optimisation system for a global CPG brand. He also explores Edge analytics, pushing the boundaries of predictive maintenance. Swapnil Ashok Jadhav, Senior Director – Machine Learning & Engineering at Angel One Swapnil developed Yubi’s first open-source repository and India’s first Fintech language model, YubiBERT, earning recognition from Meta and media coverage. At Unacademy, he created EdOCR, an OCR tailored for EdTech, and presented it at NVIDIA GTC 2021. Meet the Winners of Previous Years 2024 | 2023 | 2022| 2021 | 2020 | 2019","excerpt":"This award recognises India’s top data scientists and their achievements in the machine learning and analytics industry.","categories":["AI Highlights"],"tags":["40 Under 40 Awards","Data Science","Data Scientist Awards","MLDS","mlds nimhans"],"author_name":"Sanjana Gupta","publish_date":"2025-02-07T15:47:24","publication_year":"2025","word_count":1631,"keywords":["Data Scientist Awards","data science","machine learning","AI","neural network","ML","computer vision","MLDS","NLP","deep learning","analytics","generative AI","40 Under 40 Awards","Data Science","mlds nimhans"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","computer vision","data science","analytics","generative AI"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/40-under-40-data-scientists-awards-2025-meet-the-winners\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10097874,"title":"Google&#8217;s Voice Assistant to Get Generative AI Boost","content":"Google just might have saved its voice assistant. According to a report by Axios, the tech conglomerate is planning to bring generative AI technologies similar to those that power ChatGPT and its own Bard chatbot to its voice assistant and make it ‘supercharged’. However, Google hasn’t confirmed the development at the time of writing this article. As part of this move, the company has decided to trim “a small number of roles” from the teams working on its Assistant application. According to the report, ​​the move will involve eliminating dozens of jobs from the NLP team and transferring the responsibility to the generative AI team. Google hasn’t specified the features it intends to introduce to Assistant, leaving room for some exciting possibilities. “We remain deeply committed to Assistant and we are optimistic about its bright future ahead,” Peeyush Ranjan, the vice president of Google Assistant, and Duke Dukellis, the company’s product director, wrote in an email to the team. Change of heart Last year, a report said Google has shifted focus away from its voice assistant to Pixel phones and plans to invest less in developing its Google Assistant voice-assisted search. According to reports, Amazon also shifted its focus from Alexa. Earlier this year, Amazon employees faced layoffs in the Alexa division as a result of a substantial operating loss in its hardware unit during the previous year. Similarly, Microsoft’s efforts to popularize its Cortana virtual assistant have not yielded the desired success. However, with the sudden rise of ChatGPT, the desire for a  personalised AI assistant is increasing day by day which can perform petty tasks like ordering grocery and food, taking notes, writing a mail and several other tasks that involve human thinking. https:\/\/twitter.com\/OfficialLoganK\/status\/1685071257438498817 A voice assistant with its own intelligence can be a game changer. For example, you can ask for food recipes based on the ingredients you have at your home. As most of us regularly order food, groceries, and essential items online, we tend to have our favourite restaurants and shops. If large language models (LLMs) can remember our preferences, it would save a lot of time since we won’t have to scroll through the menu and manually add items to the cart. We would just be able to give voice commands and our order will be placed. Similarly, if we want to send a mail on the go, we can just ask the voice assistant to draft a mail. Instead of going through apps and websites manually, these tasks will become automated and much more convenient. Voice assistants will make various tasks seamless and efficient, allowing us to accomplish things with just a few simple spoken instructions. LLMs are the future of voice assistants After experiencing ChatGPT and Bard, it is very difficult to go back to voice assistants like Alexa, Siri and Google Assistant. Voice assistants appear insignificant once we get to know the capabilities of LLM chatbots. That’s what happened with Alexa. During its launch in 2015, Alexa experienced a surge of users asking curious and quirky questions, covering everything from the meaning of life to playful wishes. However, as time went by, users became disinterested in Alexa. The devices lacked the capability to provide personalised experiences or generate substantial advertising revenue for the company, causing a gradual decrease in user engagement with Alexa over time. Earlier this year, Insider reported that Amazon is looking to add ChatGPT-like generative AI features to Alexa. Large language models, using lots of web data, can sound more human in conversations with people than voice assistants like Alexa and Google Assistant, which were built mainly for natural-language understanding.","excerpt":"Google is planning to bring generative AI technologies, similar to those that power ChatGPT and its own Bard chatbot, to its voice assistant to make it ‘supercharged’","categories":["Global Tech"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-08-01T14:32:49","publication_year":"2023","word_count":602,"keywords":["Go","ChatGPT","AI","chatbots","ML","NLP","GPT","generative AI","GAN","R"],"extracted_tech_keywords":["AI","ML","NLP","generative AI","ChatGPT","chatbots","R","Go","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/googles-voice-assistant-to-get-generative-ai-upgrade\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":51760,"title":"Data Breaches Faced By Indian Consumer Internet Companies In 2019","content":"Over the last few years, India has been making headlines with a steady stream of data breaches — ranging from the Aadhaar scandal and Quora row, to the muck-up between Facebook and Cambridge Analytica. In fact, data breaches cost Indian organisations ₹12.8 crores on an average between July 2018 to April 2019, making the country the second-most cyber affected region in the world. The first half of 2019 witnessed a 22% jump in cyber attacks due to the increased deployment of IoT. With industries like healthcare, financial institutions and other businesses getting breached regularly, consumer internet companies have also been affected. Here, we are listing some major data breaches faced by the Indian consumer internet companies in 2019: Also read: Data breaches faced by consumer internet companies in 2018 Truecaller In May 2019, the data of approximately 300 million Indian users on this Swedish mobile application platform was leaked and made available for sale on the dark web. The data, which statistically made up for approximately 70% of the company’s user base of 140 million, was sold for about ₹1.5 lakh — equivalent to 2,000 EUR — on the dark web. Additionally, the data of global users was priced as high as 25,000 EUR. The alleged leaked database included names and phone numbers of 29.9 million Indians including thousands of celebrities, corporate CEOs and politicians. It also included 1.9 million email addresses, 1.8 million subscriber photos and 20 million Facebook IDs, which was acquired by the hacker through the breach. However, when asked, Truecaller said that the company ran a thorough investigation and later denied any sort of leak in the information. Contrary to the company’s claim, several experts and cybersecurity researchers continued to believe that a huge amount of data could only be accessed by hacking the database of Truecaller. Amazon India Earlier this year, one of the e-commerce giants once again claimed to have faced a technical glitch, which exposed the tax reports of some of its sellers to others. When asked, the company admitted that the glitch has affected approximately 400,000 of its sellers, who could easily download the tax reports of other competing vendors, but the issue was soon rectified once flagged. According to the reports, the leaked data contained sales figures, category split, and the inventory data. Such data could prove to be of material value for the rivals and could be harmful to the vendors working with this tech giant. The company was also hit by a breach last year, that disclosed customer names and email addresses on its website. Knowing that Amazon has faced several such issues over the past years and also such unsolicited exposure of the data this year just managed to spook its customers. Justdial Local search service, Justdial breach was another major lookout this year. According to reports, personal data of over 100 million users of the search engine were exposed online. The data leaked included some important information such as the name of the user, their email IDs, phone numbers and addresses, along with some accessory information such as gender and date of birth. The reason for the breach was the leaky endpoint, courtesy of an expired API. Another loophole was found in the API, wherein the database of the individuals who post reviews on the platform was also exposed. When asked, the company told the media that the newer version of their website was revamped and was breach proofed. In today’s digital age, with everybody recognising data as the new oil, its security is imperative for businesses. Once breached, it can not only hamper the productivity of the company, but can also wreak the reputation, decrease the brand value, destroy the company’s market capitalisation, and of course, have a major impact on the customer’s loyalty and privacy.","excerpt":"Over the last few years, India has been making headlines with a steady stream of data breaches — ranging from the Aadhaar scandal and Quora row, to the muck-up between Facebook and Cambridge Analytica. In fact, data breaches cost Indian organisations ₹12.8 crores on an average between July 2018 to April 2019, making the country […]","categories":["AI Features"],"tags":["Cyber Attack","data breach","data breach india","data breaches","indian statistical service"],"author_name":"Sejuti Das","publish_date":"2019-12-12T17:46:53","publication_year":"2019","word_count":630,"keywords":["Go","API","programming_languages:R","AI","Git","data breach","RAG","Cyber Attack","Aim","data breaches","ViT","data breach india","GAN","R","indian statistical service"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","API","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-breaches-faced-by-indian-consumer-internet-companies-in-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10064606,"title":"&#8220;Girls should make decisions independently rather than get influenced by their decision-makers&#8221;: Rising 2022","content":"To a question posed by an audience member, what is the one word you can use to describe or the basic reason behind the diversity gap (gender) in companies, Sowjanya Shetty, founder of Attitude Makeover, very hesitantly said, “patriarchy”. Sowjanya was part of the panel discussion: Curating an effective diversity & inclusion program in Tech firms – Lessons & challenges, at The Rising 2022, Women in AI conference organised by Analytics India Magazine on Friday. Watch all the recorded sessions of Rising 2022 here>> When asked about what it’s like in India from a DEI (diversity-equity-inclusion) perspective, Pooja Goyal, COO and Co-founder at Avishkaar, said, “We see that women-led companies outperform companies led by men, whether it is on Wall Street or in startups, but when it comes to the numbers, the number of companies led by women or women directors, CEOs, or leadership positions, is paltry numbers, mostly single digits. Now, you see so many women entrepreneurs, but the number is startling when you look at VC funding across the world. Only 2.3% of VC funding goes to women-led companies, just two out of 100. We are 50 per cent of the world and only 2.3%, and we have been at it for a while, at least five to six decades. So there’s something wrong there. There are two or three reasons: having children, parents-in-law, and then the women dropping out of the workforce. Although many companies are taking some measures, we need more women and girls interested in STEM-oriented roles that involve tech. Pooja further spoke about how her daughter, interested in coding, was intimidated to take up computer science as she did not have many girls in her classroom. She said that there was a need to create a safe environment for the girl child. While mentioning a study conducted on parents by her company, Pooja said a lot of parents influenced the daughter’s choice of subjects compared to their son’s choice, which was STEM, an obvious choice. Chitra Singh, Staff Machine Learning Engineer at Drishti, spoke about how, despite coming from a family where both her parents were researchers in science, she still had to face bias not just in India but also abroad. “I studied in Europe, and a software development company was interviewing me, and I asked them, how many developers do you have? They said 100, and then I asked them, how many women do you have? They said zero. And I wasn’t very comfortable going into a company which did not have a single woman”. Talking about the subject, Nishtha Dewani, Group Manager Cognitive Analytics & Diversity and Inclusion Leader, Cognitive Process Services at IBM, questioned, “Look at this room today; we have been talking that only 14% of women are in this space, which leaves you with 86% men, but who are we talking to here? Besides a few men here representing the minority, hardly any men are listening to us”. Nishtha also spoke about how decisions are made about not hiring a woman or giving her less salary when a woman comes for an interview. She also spoke about how decision-making, especially for women and girls, is very risky as it involves accountable mistakes. She said that girls should make decisions independently rather than get influenced by their “decision-makers”. Niveditha Lalge, Associate Director, Diversity & Inclusion at Subex, said, “It was important for us to leverage women and create an ecosystem for them to come and provide that in time. We consciously accepted that the DEI is not the end state, and it is the journey. You will have evolving experiences. You will fail multiple times on the face and fall flat, and we believe and accept it dutifully. But, going out and publicly speaking about it makes a big difference”.","excerpt":"Only 2.3% of VC funding goes to women-led companies, just two out of 100. We are 50 per cent of the world and only 2.3%, and we have been at it for a while or at least five to six decades.","categories":["Deep Tech"],"tags":["AIM Rising summit"],"author_name":"Poornima Nataraj","publish_date":"2022-04-08T19:06:01","publication_year":"2022","word_count":631,"keywords":["Go","funding","machine learning","AI","AIM Rising summit","Git","RAG","analytics","GAN","R","startup"],"extracted_tech_keywords":["AI","machine learning","analytics","RAG","R","Go","Git","GAN","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/girls-should-make-decisions-independently-rather-than-get-influenced-by-their-decision-makers-panel-discussion-at-the-rising-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10162325,"title":"How India Became the Core of Siemens Technology and Services&#8217; Global Strategy","content":"Siemens Technology and Services (STS), a global capability centre of Siemens AG, seems to hold a special place for its Indian team, which plays a crucial role in the company’s global journey of innovation and digitalisation. Siemens AG, the parent company, is a technology company focused on industry, infrastructure, digital transformation, transport as well as the transmission and generation of electrical power. In an interview with AIM, Pankaj Vyas, CEO and managing director of STS, explained, “Our focus is on owning, accelerating, and innovating – making India the digitalisation hub for Siemens globally.” Vyas added that India is not just an extension anymore; it is an integral part of Siemens’ global innovation strategy. Speaking about the growth of STS in India, Vyas pointed out that the organisation started as an extension of Siemens’ business units globally. Over the years, India has transformed into a core innovation and digitalisation hub for the company. Today, STS in India comprises over 9,000 professionals, with more than 6,000 focused on development and innovation. Vyas said the organisation is building expertise in technology, product, and business know-how. The team is actively involved in the design, architecture, development, and lifecycle management of Siemens products. What is so Special about the Indian Team Citing two examples, Vyas highlighted how Indian teams have significantly contributed to STS’ innovation worldwide. One such contribution is SiGreen – a cloud-based platform that helps supply chain heads monitor their carbon footprint and ensure sustainability. Developed largely in India, this solution has already won Siemens’ prestigious Werner VonSiemens Awards for innovation. Another notable innovation by the India team is Circularity 360 – a platform designed to optimise material and component selection and ensure products are sustainable from the design stage itself. According to Vyas, judging by the emissions associated with products, 80% of product-related decisions are made during the design phase. If sustainability aspects are not prioritised during the design phase, it becomes difficult to address them later. A key element of sustainable design is the selection of components and the incorporation of circularity principles. “This involves considering factors such as whether a product is using substances of concern and if it aligns with strategies like reusability, remanufacturing and whatnot,” Vyas further added. These challenges have prompted the introduction of Circularity 360, which is reportedly going for patent and has been completely funded by headquarters. “STS is also at the forefront of emerging technologies like AI, the internet of things (IoT), industrial edge, and digital twin solutions. Smarter field devices are the foundation for everything else,” said Vyas. Siemens’ devices, now connected through IoT and secured with advanced cybersecurity protocols, are designed to gather and transmit valuable data. This data is then utilised for applications ranging from predictive AI for anomaly detection in manufacturing to generative AI for digital twins and other advanced use cases. Industrial Copilot as a Game Changer Siemens Industrial Copilot, an AI-powered assistant, is an innovation driven by Indian talent that uses multimodality. The concept of the industrial copilot uses multimodality. “You take a voice as a modality, and you convert it into actions,” Vyas explained, demonstrating the integration of different modes of input and output. This system allows operators to interact with machines using natural language, streamlining operations and reducing training time. As Vyas mentioned, it is an example of how to take human-machine collaboration to the next level. “It’s like having a smart assistant you can talk to in any language,” Vyas said, hinting at the potential for Indian languages in future versions. Vyas also pointed out that the issue of employability arises due to the challenges of training individuals on a large scale. Referring to Industrial Copilot, Vyas said, “If you can have systems which you can operate with a lesser [amount of] training, then that means your people can be employed much faster and at a much bigger scale.” Furthermore, the capability of interacting with machines in local languages is emphasised as a transformative factor. This approach eliminates the need for extensive manual learning. Road Ahead When asked about the future of STS in India, Vyas pointed out three focus areas – ownership, acceleration, and innovation. He believes Indian teams will continue to lead high-value projects globally driven by advancements in AI and digitalisation. While Siemens has no immediate plans to expand to tier 2 or tier 3 cities, Vyas acknowledged the potential of such locations, particularly with government incentives and improving infrastructure. Additionally, Vyas mentioned that STS is adding a “new value layer” focused on data rather than just devices, which positions India well for growth. Confirming the elaborate infrastructure in Bengaluru, Vyas said, “In our office, we have 50 labs.” This highlights the importance of proximity to hardware setups in STS’ operations. The focus of STS is on connecting the real and digital world with a commitment to use data sourced from the real field and real devices. This approach maximises the utilisation of current locations and prioritises these over exploring tier 2 and tier 3 cities. Vyas described the STS’s culture as one rooted in competence, agility, collaboration, and innovation. “It’s not about cost arbitrage anymore; it’s about adding value,” he said. Siemens is fostering a collaborative environment that encourages speaking up, challenging the status quo, and innovating.The company’s diversity initiatives are also noteworthy, with women making up approximately 32% of its workforce. “We aim to create a culture of respect and inclusivity while staying agile to meet future demands,” Vyas added.","excerpt":"“India is not just an extension anymore; it is an integral part of Siemens’ global innovation strategy,” CEO Pankaj Vyas said.","categories":["GCC"],"tags":["GCC india","Siemens"],"author_name":"Shalini Mondal","publish_date":"2025-01-28T15:32:05","publication_year":"2025","word_count":910,"keywords":["Go","TPU","AI","ML","Git","RAG","Siemens","Aim","anomaly detection","generative AI","GCC india","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","RAG","anomaly detection","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/gcc\/how-india-became-the-core-of-siemens-technology-services-global-strategy\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10073506,"title":"Protein Wars: ESMFold vs AlphaFold","content":"Last month, Meta AI’s researchers launched a breakthrough model called Evolutionary Scale Modeling, or ESM, for protein structure prediction. This new model is touted to be one of the closest alternatives to DeepMind’s AlphaFold 2, which essentially solved the 50-year-old grand challenge of protein folding. Over the years, Meta AI has launched several models, and its most recent work has been released to the public. Check out the GitHub repository here. Besides ESMFold and AlphaFold, there are plenty of protein prediction models, including RoseTTAFold, IntFOLD, RaptorX and others. Here’s a quick overview of the models: ESMFold vs AlphaFold Meta AI claimed that AlphaFold 2 and RoseTTAFold have similar accuracy, but ESMFold inference is faster at enabling the exploration of structural spaces of metagenomic proteins. Metagenomics is a technique of sequencing DNA purified directly from a natural environment. We have trained ESMFold to predict full atomic protein structure directly from language model representations of a single sequence. Accuracy is competitive with AlphaFold on most proteins with order of magnitude faster inference. By @MetaAI Protein Team. https:\/\/t.co\/APVoaawyOb pic.twitter.com\/f6DvSfjuOX— Alex Rives (@alexrives) July 21, 2022 While AlphaFold uses a network-based model, ESMFold leverages a large-scale language model for protein prediction. Meta AI team said that the improvements in language modelling perplexity and structure learning continue through 15 billion parameters. In comparison, the team said their latest model, ESM2, at 15 million parameters, is better than their older model, ESM1b, at 650 million parameters. In addition, AlphaFold 2 and other alternatives use multiple sequence alignments (MSAs) and templates of similar proteins to achieve optimal performance or breakthrough success in atomic-resolution structure prediction. However, ESMFold generates structure prediction using only one sequence as input by leveraging the internal representations of the language model. With a single sequence as input, ESMFold produces more accurate atomic-level predictions than AlphaFold and competes with RoseTTAFold when given full multiple sequence alignments (MSAs). Amazing. We did see this also come up in ProGen – Large language models captured 3d structure through its attention.https:\/\/t.co\/0oK7coEMV6— Richard Socher (@RichardSocher) July 24, 2022 ESMFold produces comparable predictions for low-perplexity sequences, and that structure prediction accuracy correlates with language model perplexity in general. In other words, when a language model can better comprehend a sequence, it can comprehend a structure better. One of the advantages of ESMFold is that it offers a faster prediction speed than existing atomic resolution structure predictors. This, in a way, allows it to bridge the gap between the rapid growth of protein sequence databases containing billions of sequences alongside the slower development of protein structure and function databases. The model is used to rapidly compute one million predicted structures representing a diverse subset of metagenomic sequence spaces that lacks labelled structure or function. Last month, DeepMind, in collaboration with European Bioinformatics Institute (EMBL-EBI), released predicted structures for nearly all catalogued proteins, which will expand the AlphaFold database by over 200x – from nearly 1 million structures to over 200 million structures – with the potential to increase our understanding of biology significantly. AlphaFold, initially launched in 2018, published its second version in 2020, and released an open-source version of its deep-learning neural network AlphaFold 2 last year. With this, the team said that the new model significantly increases the accuracy of predicted multimeric interfaces over input-adapted single-chain AlphaFold, while maintaining high intra-chain accuracy. One of the biggest performance drivers for ESMFold has been the language model. For instance, when ESM-2 understands the protein sequence well, you can obtain predictions comparable to those made by other models when language modelling perplexity is high. In other words, it is possible to obtain accurate atomic resolution structure predictions with ESMFold – i.e. up to two orders of magnitude faster than AlphaFold 2. Meta AI said billions of protein sequences have unknown structures and functions, many from metagenomic sequencing. ESMFold makes it possible to map this structural space in practical timescales, where they can fold a random sample of 1 million metagenomic sequences in a few hours. Moreover, the researchers believe that ESMFold can help to understand regions of protein space that are distant from existing knowledge. A new ‘super fast’ protein-predicting model emerges ESMFold and AlphaFold are not alone. OmegaFold, developed by Chinese biotech firm Helixon, also predicts high-resolution protein structure from a single primary sequence. Recently, this model outperformed rival RoseTTAFold while achieving similar prediction accuracy to AlphaFold 2. OmegaFold's code and model1 is released:https:\/\/t.co\/QNS01ITjkM— Jian Peng (@peng_illinois) August 3, 2022 Only recently, the company made its code publicly available, joining the likes of  AlphaFold and ESMFold, which are also open source. Why is this a big deal? The folding of proteins helps researchers and scientists understand the underlying cause of many diseases. Knowing these protein folding, protein design, etc., helps find a cure, design new medicines, drugs, pharmaceutical solutions, etc.","excerpt":"While AlphaFold 2 and RoseTTAFold have similar accuracy, ESMFold inference is faster at enabling the exploration of structural spaces of metagenomic proteins","categories":["Deep Tech"],"tags":["AlphaFold","ESMFold","RoseTTAFold"],"author_name":"Amit Naik","publish_date":"2022-08-24T18:00:00","publication_year":"2022","word_count":799,"keywords":["AlphaFold","Meta AI","API","ESMFold","programming_languages:R","AI","neural network","Git","RAG","Aim","RoseTTAFold","GitHub","R"],"extracted_tech_keywords":["AI","neural network","Meta AI","Aim","RAG","R","Git","GitHub","API","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/protein-wars-its-esmfold-vs-alphafold\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10068716,"title":"Build a career in Data Science and ML with MTech from SRM Institute of Science and Technology","content":"According to AIMResearch, the median salary of data science professionals in India has increased 25.4% year on year to INR 16.8 lakhs per annum. As a result, the demand for skilled professionals in data science and AI is on the rise. The MTech program in Computer Science and Engineering with additional specialisation in data science and machine learning by SRM Institute of Science and Technology (accredited and approved by AICTE) is ideal for setting your data science career on the right track. The classes will be held at the Chennai Main campus during the weekends. Why SRM University? SRM University is one of the best universities in the country. Major milestones include: Secured rank #3 among top engineering institutes in India (Times Engineering Survey)Accredited with the highest ‘A++’ Grade (NAAC)Globally rated ‘Four Star’ University (QS World University Ranking)It has ranked #35 among the Top 50 Universities (MHRD – NIRF) What’s in it for the students? Industry relevant curriculum-The program enables participants to gain an in-depth understanding of data science and machine learning techniques and tools that are widely used by companies. The program also imparts real-world skills to enable learners to become job-ready. The program includes sessions from seasoned professors and industry experts. Networking opportunities-Not only academic rigour, the program provides networking opportunities with peers to grow your professional network as well as mentorship from experts to gain industry insights Dedicated recruitment drives: Students will be able to attend job fairs conducted online and offline and get a chance to participate in the recruitment drives of top tech companies. Access to curated jobs: The career support team works with over 7800+ organisations to recommend the right jobs for the course participants. Interview preparation workshops: Students can familiarise themselves with commonly asked questions to crack interviews and get assistance in building a portfolio to showcase their skills and improve their chances of getting hired. Personalised career mentorship: Students will get an expert career mentor to help them navigate the job market. Upon completing the course, the students can choose roles like data scientist, data analyst, business analyst, ML engineer, big data engineer, etc., depending on their interests and strengths. Curriculum overview Semester 1 Mathematical foundations of computer scienceData structures and algorithmsObject-oriented software engineering Semester 2 Database technologyData visualisationAdvanced machine learning Semester 3 System programmingDigital Image ProcessingDeep learning and applications Semester 4 Natural language processing Big data essentials Open elective 1 Semester 5 Wireless sensor network Open elective 2 Semester 6 Mtech project work Open electives: Students can choose any two from the following list. A minimum of 10 students must choose an elective for the course to be taught. The topics include Operations research, Industrial safety, Entrepreneurship and IPR and Human-Computer Interaction. The program will follow a continuous evaluation scheme, and candidates will be evaluated in the courses they undergo through examinations, case studies, quizzes, assignments, and project reports. How to apply? Register by filling up the online application form Go through an admission test and a screening call with the Admission Director’s office.If selected, the student will receive an offer of admission to the upcoming cohort. The student has to secure the seat by paying the admission fee.The fee for this course is INR 5,00,000. One can pay the admission fee and first instalment before program commencement followed by three equal quarterly instalments. ( Examination fee is not included in the program fee). Know more about the programme here. So, what are you waiting for? Hurry up & apply for the MTech program in Computer Science and Engineering here.","excerpt":"The program enables participants to gain an in-depth understanding of data science and machine learning techniques and tools that are widely used by companies","categories":["AI Trends"],"tags":["career in data analytics"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-10T14:00:00","publication_year":"2022","word_count":589,"keywords":["big data","data science","Go","machine learning","AI","ML","Git","Aim","deep learning","career in data analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","Aim","R","Go","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/build-a-career-in-data-science-and-ml-with-mtech-from-srm-institute-of-science-and-technology\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":65671,"title":"Corporates Tend To Look For Generalist Data Scientists, Says This Chief Data Scientist","content":"Analytics India Magazine got in touch with Madhavi Kaivalya Kandalam, VP – Chief Data Scientist of Loylty Rewardz, for our weekly column My Journey In Data Science. Madhavi has over 10 plus years of experience in the data science domain while working for Mu Sigma and Loylty Rewardz. In coming years, she wants to start her business in the AI space to create a social impact using the latest technologies. The Onset Madhavi completed her B.Tech from NIT, Warangal in 2010, majoring in biotechnology. She chose biotech initially out of interest, but growth prospects appeared less at the time. This led to the thought of moving towards a different career path. “Analytics seemed like a good option as it was a new field, which did not demand any formal degree back then and growth prospects seemed very promising,” says Madhavi. “The first time I heard about analytics was during the pre-placement talk given by Mu Sigma, and I fell in love with the field. If one looks at the amount of data humanity has produced in the last decade is enough to figure out this field has a golden future.” Preparation Strategy After joining Mu Sigma as a Business Analyst, Madhavi learnt about SAS, VBA, and other analytics skills. She learnt these tools and techniques through internal training of the company, however, Madhavi said that the most helpful was the learnings on solving unstructured problems. “Over the years, tools like SAS and VBA have become irrelevant, but the challenges in analysing unstructured problems have remained constant,” says Madhavi. Talking about the difficulty in preparing in the everchanging data science landscape, Madhavi said that learning in data science can be overwhelming at times, the most sought after tool today will become irrelevant a year down the line. But, that is also the most exciting part of the field. Unlike most, Madhavi never focused on certifications. She believed in learning through real-world applications. While working on projects, she analysed what kind of tools and techniques were required to solve, and then learnt those by implementing and evaluating the results. In addition, she also read research papers to learn new technologies that could assist here to effectively solve business problems. Even after continuous learning, Madhavi said the data science field is vast, which makes her feel that she hardly knows anything. Madhavi believes constant learning and reading is the only way to survive in the field. Best And Worst Data Science Experience In her data science journey, Madhavi has been involved in numerous successful projects. However, one of her favourites is when she led a team to create backend algorithms for selecting target customers for small scale retailers through a mobile application. They were running on tight timelines but managed to create 20 plus customer segments with sparse data and enrich it through look-alike models. Besides, while recollecting here worse data science experience, Madhavi said that organisations find it difficult to understand the data science functions. She feels that it is difficult to convey the limitations and challenges of the domain — what is possible and what is not. There are a lot of misconceptions around this domain, and it creates a huge communication gap. “A lot of business decision making regarding the data science function is short-sighted. Convincing organisations on the importance of investing in long-term goals is challenging,” explains Madhavi. “Maintaining good quality data is essential for businesses to thrive, but organisations fail to understand its importance mostly because they have a short-term goal, where the importance of quality of data is usually ignored. The single largest differentiator between organisations that have leveraged data science successfully in comparison to those that have not is the quality data maintenance and procurement. I think the Indian market is still not mature on this,” she added. Current Work Experience After working for Mu Sigma for more than three years, Madhavi joined Loylty Rewardz in 2014. At Loylty Rewardz, Madhavi was exposed to a colossal amount of data related to 800 million customers. Implementing machine learning and other data science techniques to understand how customers behave, purchase and more, further excited here about the field. Today, Madhavi leads a 25 membered team that is responsible for driving customer engagement, delivering insights and strategy for B2B clients, and creating\/enhancing data-based products and platforms. One of the most challenging parts of her job is to communicate about the functions with business stakeholders in a language that is palatable to everyone. But, she has mastered it over the years to communicate effectively. As a part of her job role, Madhavi also hires data scientists and usually does not focus on the types of certifications they have done. Instead, she evaluates based on the practical experience, openness to learning new tools, and internal tests, where she assess programming skills, machine learning techniques, and application through a live problem involving real-world datasets. Advice To Aspirants As a piece of advice to aspirants, Madhavi said that aspirants should start with Python and SQL as most of the organisations use these tools, but at the same time, practitioners should not be fixated on the technology they love. Come to the field for what it has to offer and be open to working with different things continuously. Besides, she believes one should not get disheartened by rejections in interviews and continue to improve. “I think a little flexibility on the kind of work and responsibilities, which one is willing to take up help as corporates tend to look for generalists,” concludes Madhavi.","excerpt":"Analytics India Magazine got in touch with Madhavi Kaivalya Kandalam, VP – Chief Data Scientist of Loylty Rewardz, for our weekly column My Journey In Data Science. Madhavi has over 10 plus years of experience in the data science domain while working for Mu Sigma and Loylty Rewardz. In coming years, she wants to start […]","categories":["AI Features"],"tags":["corporate analytics platform","Data Science Certification","Interviews and Discussions","My Journey In Data Science"],"author_name":"Rohit Yadav","publish_date":"2020-05-22T17:00:00","publication_year":"2020","word_count":919,"keywords":["data science","Go","machine learning","AI","Data Science Certification","corporate analytics platform","My Journey In Data Science","RAG","Python","analytics","SQL","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","RAG","Python","R","SQL","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/corporates-tend-to-look-for-generalist-data-scientists-says-this-chief-data-scientist\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10051650,"title":"[Jobs Roundup] Latest Data Science And ML Jobs In India","content":"We have listed the latest data science and machine learning jobs in India this week. 1| Data Engineering – ETL\/SQL\/Python Location: Any Responsibilities: Connecting, designing, scheduling, and deploying data warehouse systems.Developing data pipelines and enabling dashboards for stakeholders. Develop, construct, test and maintain system architectures.Create best practices for data loading and extraction. Apply here. 2| Market Mix Modeling Role – Marketing Analytics Location: Delhi NCR Responsibilities: Use and implement basic and advanced statistical techniques like frequencies, cross-tabs, correlation, regression, decision trees, cluster analysis.Experience in using multiple advanced analytics techniques or machine learning algorithms.Understanding of consumer businesses such as Retail or CPG. Handling small teams of 2-3 analysts. Apply here. 3| Assistant Manager – Data Scientist Location: Any Responsibilities: Design the solution for advanced analytics (including Big Data) projects – what technique to use, what tool\/technology to use, etc.Scope out the solution in terms of time, effort, team skills needed independently.Own the delivery for projects, including developing scalable ML models and algorithms and interpreting the output. Apply here. 4| Engagement Manager – Data Analytics Location: Gurgaon\/Gurugram Responsibilities: To support and implement high-quality, data-driven decisions across product, sales, and marketing to support and implement high-quality, data-driven decisions.Hands-on with solving the most difficult technical aspects of data and analytics projects.Candidates who are comfortable working independently, as a team lead and client engagement managers at the same time. Apply here. 5| Credit Risk Model Development Role Location: Any Responsibilities: Ability to work with key stakeholders across businesses, client portfolio teams to derive insights and calibrate model performance.Ability to present the findings of the analysis to stakeholders and hold presentations for a larger audience.Ability to drive discussions with the stakeholders and present the findings\/summary of the project activities. Apply here. 6| Senior Data Engineer – AWS Location: Bangalore Responsibilities: Owning the design, operations and improvements for the Customer Data- warehouse infrastructure.Engineering solutions to aggregate and automate large scale data flows from varying sources.Collaborate with others to construct complex data sources for algorithms and machine learning models. Apply here. 7| Associate Product Owner – SQL\/Python Location: Bangalore Responsibilities: As a product owner, you will be primarily responsible for executing the strategic vision of AI\/ML data products.You will oversee feature implementations from ideation to launch by collaborating with the business teams, customer teams and engineering teams.You will play a pivotal role in driving product strategy and roadmap; you will recommend new features based on customer and market-gap analysis. Apply here. 8| Senior Associate – Data Scientist Location: Bangalore Responsibilities: Take charge and lead the efforts for the data science life cycle for critical Industrial AI project implementations.Explore very large time-series data sets and discover key insights and patterns in collaboration with domain specialists.Mentor a team of junior data scientists, machine learning engineers and data engineers, and drive the planning, design, and implementation of data science experiment cycles. Apply here. 9| Senior Manager\/Associate Vice President – Analytics Location: Gurgaon\/Gurugram Responsibilities: Doing independent research, analysing, and presenting data as assigned.Expected to work in close collaboration with the EXL team and clients on analytics projects for general insurance related to pricing\/claim analysis.Predictive modelling based GLM for personal and commercial lines, developing statistical models in R, Python, SAS. Apply here. 10| Consultant\/Senior Consultant\/Project Manager – Analytics & Regulatory Modelling Location: Any Responsibilities: Skilled in validation and monitoring of internal and external risk models by computing standard metrics.Skilled in developing and analysing product-specific solutions within the banking domain varying across underwriting\/monitoring loan portfolios, developing collections scorecards, loss forecasting, amongst others.Ability to work with key stakeholders across businesses, client portfolio teams to derive insights and calibrate model performance. Apply here. 11| AVP – HR Analytics Location: Gurgaon\/Gurugram Responsibilities: Manages HR metrics and data from multiple sources like existing dashboards, employee surveys, exit interviews, employment records, competitors practices, etc.Manage the entire gamut of HR Analytics – from recruitment to exit and present data and insights on talent acquisition, talent\/career progressions, talent compensation basis external benchmarking, market intelligence, etc.Analyses data and statistics for trends related to attrition, recruitment\/hiring, employee movement, staffing, budgeting, talent engagement, retention, manpower planning. Apply here. 12| Data Engineer- Big Data Location: Bangalore Responsibilities: Responsible for end-t0-end development of projects which includes understanding requirements, designing the solution, implementing, testing and maintaining it.Responsible for resolving issues that might occur in existing solutions.Responsible for optimisation of existing solutions to save time and resources. Apply here. 13| Data Architect – Python\/R Location: Bangalore Responsibilities: Developing and implementing an overall organisational data strategy that is in line with business processes.The strategy includes data model designs, database development standards, implementation and management of data warehouses and data analytics systems.Identifying data sources, both internal and external and working out a plan for data management that is aligned with organisational data strategy. Apply here. AIM Recruits A subsidiary of Analytics India Magazine, AIM Recruits is an India-focused executive search firm that partners with leading businesses to assess and acquire top data science talent to drive breakthrough performance. AIM Recruits is a leading executive search firm for Analytics, Data Science & Artificial Intelligence.","excerpt":"We have listed the latest data science and machine learning jobs in India this week.","categories":["AI Hirings"],"tags":["AIM Weekly Jobs","analytics jobs","Data Science Jobs","jobs in data science","jobs in india","jobs roundup","weekly job updates"],"author_name":"kumar Gandharv","publish_date":"2021-10-14T17:00:00","publication_year":"2021","word_count":833,"keywords":["data science","artificial intelligence","machine learning","AWS","AI","TPU","jobs in india","ML","Data Science Jobs","analytics jobs","weekly job updates","Python","Aim","jobs in data science","analytics","AIM Weekly Jobs","jobs roundup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","AWS","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/jobs-roundup-latest-data-science-and-ml-jobs-in-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":19272,"title":"6 Times Image Recognition Technology Failed In Real Life","content":"With the advent of new and AI-backed phones, very soon, most of us won’t even bother searching for the item, address or even product that we want. We would simply point our camera at it, and the phone would recognise the content in front of it. While many brands have already inculcated the technology in their phones, the concept is still primarily used with artificial systems that extract information from images. The image data can be of any form — video sequences, views from multiple cameras, or even a multi-dimensional data from a medical scanner. As a technological discipline, computer vision seeks to apply its theories and models for the construction of computer vision systems. But Is The System Perfect Yet? The British Machine Vision Association and Society for Pattern Recognition defines it as, “A system concerned with the automatic extraction, analysis and understanding of useful information from a single image or a sequence of images. It involves the development of a theoretical and algorithmic basis to achieve automatic visual understanding.” Image recognition, especially facial recognition technology, which involves training computer programs to recognize objects based on databases of images, has caused issues for other services. From the perspective of engineering, computer vision seeks to automate tasks that the human visual system can do. With the help of Artificial Intelligence and Deep Neural Networks, recognising objects, faces and images has become increasingly easier. Here Are Some Of The Major Fails In Image Recognition: In 2016, researchers were able to fool a commercial facial recognition system into thinking that they were somebody else just by wearing a pair of patterned glasses. A special (think, funky-looking) sticker overlay with a hallucinogenic print was stuck onto the frames of the specs. The twists and curves of the pattern looked random to humans, but to a computer designed to pick out noses, mouths, eyes, and ears, they resembled the contours of someone’s face — any face the researchers chose, in fact. The facial recognistion system was so confused that it even recognised one of the researchers as the Pope! One of the most recent facial recognition fail that comes to mind is during the launch of the eagerly-awaited iPhone X. In September earlier this year, Craig Federighi, Apple’s senior vice president of Software Engineering, struggled to unlock the brand new phone while demonstrating Face ID, Apple’s new facial recognition software. “Unlocking it is as easy as looking at it and swiping up,” he said. But when it failed to unlock, he told the audience, “Let’s try that again.” But after he was prompted to enter his passcode, Federighi was forced to get a backup device to continue with the demonstration. Jeff Clune, co-author of a 2015 paper on images that fooled DNNs and other image recognition software says, “Take convolutional neural networks trained to perform well on either the ImageNet or MNIST datasets and then find images with evolutionary algorithms or gradient ascent that DNNs label with high confidence as belonging to each dataset class. It is possible to produce images totally unrecognisable to human eyes that DNNs believe with near certainty are familiar objects. Our results shed light on interesting differences between human vision and current DNNs, and raise questions about the generality of DNN computer vision. In 2014, Teve Talley, a Denver-based financial advisor had been booked on the charges of two bank robberies. The evidence against him was a grainy CCTV camera footage which and a computer facial recognition software that matched his broad-shoulders, skin colour, sex, age, hair, eyes, and square jaw with that of the actual criminal. “Typically, the forensics community relied on experts in a binary way: Is this the same guy or not the same guy?” Akil N Jain, one of the world’s leading pioneers of face recognition technology, explained in an interview. “The focus has shifted to ‘How can you be so sure? Give us some confidence level.’ The forensic community needs to accept that examiners can make mistakes, and they need to say, ‘How can we avoid that?’” In 2015, Google’s newly-launched Photos service, which uses machine learning to automatically tag photos, had made a huge miscalculation when it automatically tagged two African-Americans as “gorillas.” in the folders. The user, a US-based computer programmer reported the problem via Twitter when he found that Google Photos had created an album labeled “gorillas” that exclusively featured photos of him and his African-American friend. At the time, developers at Google had immediately apologised for the gaffe and then worked to fix the app’s database. A few years ago, Flickr’s auto-recognition tool for sorting photos had gone awry and had  identified one picture of a black man with the tags “ape” and “animal”. Though the racist implications were obvious, it had also identified a white woman with the same tags.","excerpt":"With the advent of new and AI-backed phones, very soon, most of us won’t even bother searching for the item, address or even product that we want. We would simply point our camera at it, and the phone would recognise the content in front of it. While many brands have already inculcated the technology in […]","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Computer Vision","deep neural network","image recognition","real time face recognition software"],"author_name":"Prajakta Hebbar","publish_date":"2017-11-24T08:54:01","publication_year":"2017","word_count":801,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","neural network","deep neural network","image recognition","programming_languages:Go","computer vision","real time face recognition software","Computer Vision","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","computer vision","image recognition","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/image-recognition-fails-real-life\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10007454,"title":"Can GPT-3 Pass Multitask Benchmark?","content":"After GPT-3 showcased SOTA performance on various NLP tasks, researchers from UC Berkeley, Columbia University, UChicago, and UIUC have proposed a new test to measure its multitasking capabilities. Considering that transformer-based models like GPT-3 have been trained on massive text corpora from numerous websites, it has managed to display excellent results for NLP benchmarks and specialised topics. However, it was critical to understand the capability of these language models at grasping their knowledge and understanding of various domains. Thus, to check GPT-3’s multitask accuracy, researchers proposed a test covering 57 tasks, including US history, elementary maths, computer science and law, to name a few. In a recent paper, researchers have stated that the benchmark has been designed to measure the acquired knowledge during their training and ranged the tasks from elementary level to advanced level problem-solving capabilities. Also Read: GPT-3 Vs BERT For NLP Tasks How Does The Multitask Test Work? To facilitate this accuracy test, researchers created a massive multitask test which included approximately 15,000 multiple choice questions gathered manually from available online sources. The gathered questions are then categorised into — few shot development set; validation set; and a test set with the maximum number of questions. According to researchers, considering the test aggregates various subjects of different difficulty levels, they decided to test its capability beyond straightforward common sense and limited linguistic knowledge. For this, the researchers tested the model on real-world text understanding for measuring the ability of the models to extract useful knowledge from massive training data fed onto it. To assess GPT-3, the researchers included Unified QA and the entire family of GPT-3 to compare its results. It included a small model with 2.7 billion parameters, a medium one with 6.7 billion, large with 13 billion and X-large with 175 billion parameters, and computed classification accuracy across all tasks. Models tested on four broad disciplines and shared value in percentage. For a few-shot prompt — the researchers fed the model GPT-3 with a few prompts with five demonstration examples with answers to the prompt before asking the questions. And for zero-shot learning, they appended the question to the prompt. Left: examples of few-shot learning and inferences using GPT-3. The answer underlined in “blue” is the response from GPT-3. Right: Performance on commonsense, linguistics and the proposed multitask test. Here it can be noticed that GPT-3 produced probabilities for the token A; B; C and D, and the one with the highest probability will be treated as the prediction. To obtain a consistent evaluation of the models, the researchers also created a dev set with five fixed few-shot learning examples for each subject. Also Read: How Guardian’s Recent Article Is Yet Another GPT-3 Hype Comparing Results To check the models’ size and accuracy, each GPT-3 model, whether it be small, medium, large or XL, have been compared on their few-shot accuracy. Here the researchers noted that three smaller GPT-3 and Google’s T5 model have near-random accuracy of approximately 25%. However, GPT-3 L with 13 billion parameters showcased an accuracy of 37.7%, while GPT-3’s XL model with 175 billion parameters turned out to have an accuracy of 43.9%; in both few shot and zero-shot learning. These results showcase that bigger the model size, higher will be the accuracy of the model on multitasking tests. To test this theory, they also compared the UnifiedQA models, which displayed accuracy of 38.5% without any fine-tuning. Thus it has been less accurate than the few-shot GPT-3 XL but more accurate than zero-shot GPT-3 XL. It has further been proved even the smallest UnifiedQA model with just 60 billion parameters has better accuracy of 30% than others. Thus, it has been established that model size can play a significant part in achieving the accurate performance of the model. With that being said, it’s not the only criteria for achieving the highest performance in a language model. Here, the researchers have decided to compare disciplines for testing accuracy. It has been discovered that GPT-3 has uneven accuracy and several substantial knowledge gaps. Evaluating it on all 57 tasks, researchers noted that GPT-3 doesn’t grasp knowledge as humans do; instead, it has its own unusual order to follow. It showed below expert-level performance for the tasks, ranging from 26% for college chemistry and 69% for US foreign policy. And overall, GPT-3 displayed poor results on severe procedural problems. Further breakdown on each task — the GPT-3 showed less accuracy for STEM subjects like maths and calculation over the ones which required verbal knowledge. A lot of this could be attributed to its capability of acquiring more declarative learning than procedural learning. While it exhibited 47.4% accuracy for college medicine and 35% accuracy for college maths, it lowered down its accuracy to 29.9% when it comes to elementary maths with more calculations. Also Read: OpenAI’s Not So Open GPT-3 Can Impact Its Efficacy Additionally, the researchers evaluated the calibration of GPT-3, which was critical to trust its prediction by testing its average confidence for estimating actual accuracy for the tasks. Here, the researchers found that GPT-3 isn’t calibrated and thus providing miscalibrated forecasts with an error of 19.4%. Therefore they believe that there is an immense opportunity for improving the model calibration. Conclusion With 57 designed tasks, the researchers aimed to measure the multitasking capabilities of GPT-3 however after experimenting and evaluating it has been noted that it is very novel for the model to make meaningful progress on the test and have lopsided performance without excelling at any. Not only GPT-3 showcased low accuracy for calculation-based tasks but also for social-based tasks including morality and law. Thus it can be said that while GPT-3 comes with an extensive breadth of knowledge, it doesn’t master any single subject, and comes with many knowledge blindspots with uneven accuracy. With this proposed benchmark, the researchers aimed to help others in pinpointing the limitations of models and making it possible to get an accurate view of its SOTA capabilities. Read the whole paper here.","excerpt":"After GPT-3 showcased SOTA performance on various NLP tasks, researchers from UC Berkeley, Columbia University, UChicago, and UIUC have proposed a new test to measure its multitasking capabilities.  Considering that transformer-based models like GPT-3 have been trained on massive text corpora from numerous websites, it has managed to display excellent results for NLP benchmarks and […]","categories":["AI Features"],"tags":["GPT-3"],"author_name":"Sejuti Das","publish_date":"2020-09-16T10:00:26","publication_year":"2020","word_count":997,"keywords":["GPT-3","Go","OpenAI","AI","RAG","NLP","Aim","few-shot learning","Rust","R","zero-shot learning"],"extracted_tech_keywords":["AI","NLP","OpenAI","Aim","RAG","few-shot learning","zero-shot learning","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-gpt-3-pass-multitask-benchmark\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10121610,"title":"Google DeepMind Introduces Semantica, An Adaptable Image-Conditioned Diffusion Model","content":"Researchers at Google DeepMind introduced Semantica, an image-conditioned diffusion model capable of generating images based on the semantics of a conditioning image. The paper explores adapting image generative models to different datasets. Instead of finetuning each model, which is impractical for large-scale models, Semantica uses in-context learning. It is trained on web-scale image pairs, where one random image from a webpage is used to condition the generation of another image from the same page, assuming these images share semantic traits. Semantica leverages pre-trained image encoders and semantic-based data filtering to achieve high-quality image generation without the need for fine-tuning on specific datasets. Its architecture enables it to generate new images from any dataset by simply using images from that dataset as input, making it highly adaptable. Source: Research Paper This flexibility is essential for practical uses, as it allows the model to work with a wide range of dynamic image sources without the need for extensive retraining. By using diffusion models, which iteratively refine an image from a noise vector, Semantica achieves a balance between computational efficiency and output quality. The approach allows for scalable and flexible image generation, which is valuable for various real-world uses such as content creation, image editing, and virtual reality environments. Semantica can be useful in various domains. For instance, in creative industries, the model can be used to generate artwork or design elements based on a given theme or style. In education, it can create illustrative content tailored to specific topics, enhancing the learning experience. Additionally, in e-commerce, Semantica can generate product images that match the aesthetic preferences of different customer segments, potentially boosting engagement and sales. The researchers conducted extensive experiments to evaluate Semantica’s performance across different datasets and found that the model effectively captures the semantic essence of the conditioning images, producing results that are visually coherent and contextually relevant. Researchers at Google DeepMind have been doing some exciting work lately. Recently, they also introduced CAT3D, a new method for creating 3D scenes in as little as one minute. Instead of needing hundreds of photos, CAT3D uses a few images to generate new, consistent views of a scene. These views help create detailed 3D models that can be viewed from any angle in real-time. Google DeepMind, in collaboration with its subsidiary Isomorphic Labs, also unveiled AlphaFold 3, a new AI model capable of predicting the structure and interactions of all biological molecules, including proteins, DNA, RNA, and ligands. AlphaFold 3 is the first AI system to surpass physics-based tools for biomolecular structure prediction.","excerpt":"Once trained, it can generate new images adaptively from a dataset by simply using images from that dataset as input.","categories":["AI News"],"tags":["AlphaFold","Google","Google Deepmind"],"author_name":"Sukriti Gupta","publish_date":"2024-05-24T18:12:03","publication_year":"2024","word_count":421,"keywords":["Go","AlphaFold","TPU","AI","R","RPA","Scala","diffusion models","RAG","Google Deepmind","Google","GAN","in-context learning"],"extracted_tech_keywords":["AI","RAG","in-context learning","TPU","R","Go","Scala","diffusion models","GAN","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-deepmind-introduces-semantica-an-adaptable-image-conditioned-diffusion-model\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":35034,"title":"This Python Developer Tool Takes The Heavy Lifting Out Of Coding","content":"Pycharm is the most popular and full-fledged IDE for Python scripting language which is created by Czech company, JetBrains that focuses on creating an integrated development environment for different web development languages. This Python IDE includes impressive features like an advanced debugger, high-quality code completion, support for web programming and code inspection. Unlike other IDEs, it focuses on working with projects of Python scripting language. In this article, we list down some of the popular features of this developer tool that helps debug code effectively. Features The most popular IDE has some excellent features that help a developer to work comfortably to gain more insights as well as get quality code completions. Some of the features are discussed below. Code Completion And Inspection: This IDE enables easy code completion irrespective of the packages. Code Coverage: You can mark .py files as code coverage and run outside the Pycharm Editor. Git Visualisation: Pycharm has a blue section where you can easily check your last as well as current commits. Package Management: Ability to search and add new packages are easy here and you can easily track all the packages with proper visual representation. Refactoring: This IDE has some shortcuts for easy refactoring process. SQLAlchemy Debugger: You can see the SQL representation of the user expression for SQL language code by setting a breakpoint and pausing the debugger. Steps For Installation Steps to install Pycharm on the system are fairly simple. The installation process is similar to any other software packages. All one needs to do is download the package from here where you will find two categories, Professional and Community. If you are a learner then it’s better to hit the latter one as it can be downloaded without spending any penny. There is also one more category, that is the Pycharm Educational Edition which is a perfect tool for an enthusiast. For Windows installation, run the PyCharm-*.exe file you’ve downloaded and follow the instruction wizard. For MacOS installation, open the PyCharm-*.dmg package that you’ve downloaded, and drag PyCharm to the Applications folder. For Linux (Ubuntu 16.04 or higher) installation, run the following command sudo snap install <pycharm-professional,·pycharm-community, or·pycharm-educational> –classic Then, run Pycharm-professional, Pycharm-community, or Pycharm-educational in the Terminal. Steps For Running Your First Python Project 1| Initial: Before you start, make sure that you have installed your Pycharm package as well as Python in your system. 2| Choosing Interpreter: This IDE allows coder to use more than just one Python interpreter. It supports the standard Python interpreters, PyPy, IronPython, Jython, etc. 3| Create Virtual Environment: Using external libraries can lead to managing the versions and in order to do so, dependencies like Virtualenvs comes handy. It helps you keep the dependencies for your different projects separate. 4| Create Python Script: After the setup, on the welcome screen, hit create new project. It suggests project templates for the creation of the various types of applications (Django, Google AppEngine, etc.). Python best practice is to create a virtualenv for each project. To do that, expand the Project Interpreter: New Virtualenv Environment node and select a tool used to create a new virtual environment. 5| Create a New File: For creating a Python file you need to select the project root in the Project Tool Window and press Alt+Insert. 6| Editing Source Code: PyCharm analyses your code on-the-fly, the results are immediately shown in the inspection indicator on top of the right gutter. 7| Code with Smart Assistance: Use the coding capabilities to create error-free applications without wasting precious time. Code completion helps in a great time saving, regardless of the type of file. PyCharm keeps an eye on what you are currently doing and makes smart suggestions, called intention actions, to save more of your time. 8| Run Application: You can right-click the editor, and on the context menu choose to run the script Ctrl+Shift+F10 or there is a green play icon in the left gutter where the available commands are shown when you point the mouse over it. Outlook Pycharm allows you to create a full-fledged package with classes, subclasses, tests, GUIs, configs, etc The debugger allows you to investigate the step-by-step code Version control is much easier Pycharm allows you to a better code profiling and completion Pycharm provides support for JavaScript, HTML\/CSS, Angular JS and, many others for web development","excerpt":"Pycharm is the most popular and full-fledged IDE for Python scripting language which is created by Czech company, JetBrains that focuses on creating an integrated development environment for different web development languages. This Python IDE includes impressive features like an advanced debugger, high-quality code completion, support for web programming and code inspection. Unlike other IDEs, […]","categories":["AI Features"],"tags":["django python","Pycharm"],"author_name":"Ambika Choudhury","publish_date":"2019-02-15T10:53:23","publication_year":"2019","word_count":721,"keywords":["Go","AI","ML","Git","RAG","Python","Pycharm","django python","SQL","JavaScript","R","Java"],"extracted_tech_keywords":["AI","ML","RAG","Python","R","SQL","JavaScript","Go","Java","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-python-developer-tool-takes-the-heavy-lifting-out-of-coding\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10147876,"title":"This Infosys-backed Bengaluru Startup is Beating Cancer with Genomics &amp; AI","content":"In India, cancer treatment often follows a standardised approach, bucketing the treatments based on the stage and type of cancer. However, the West adopts a rather nuanced strategy with genome research to understand the exact type of cancer within a certain stage and type and administer targeted therapy, thereby hitting a higher recovery rate. Hitesh Goswami and Kshitij Rishi founded the Bengaluru-based oncology precision startup 4baseCare in 2018 to introduce a similar approach to treating oncology patients in India. Recently, Infosys Innovation Fund invested close to $1 million in the leading precision oncology firm. They were also the past winners of the Karnataka government’s ‘Elevate’ initiative that supports innovative early-stage startups. A few months ago, deep tech-focused venture fund Yali Capital led a Series A funding round and raised $6 million. Genome Technology for Targeted Care In an exclusive interaction with AIM, co-founder and CEO of 4baseCare, Goswami, explained the methodology behind cancer treatment. Previously, cancer treatment was uniform. All patients received the same therapy based on the cancer type and stage, and individual differences were not considered in treatment plans. Goswami said that today, lung cancer stage-2 and stage-3 patients are subgrouped into 12 to 15 categories; each requiring distinct treatments as therapies effective for one group may not work for another and could even cause adverse reactions. According to him, sequencing one human genome took around 15 years and $3.2 billion. Now that the Human Genome Project has concluded, Goswami believes the procedure can be done at a cost of $100 for a genome in a couple of years. Notably, genome testing for oncology has grown rapidly in India. What stood at 5,000-6,000 tests in a year in 2019 has now touched 2 lakh. Goswami estimates that the market will easily hit 3 lakh tests. India has its Own Struggles While the results help in targeted care for cancer patients, precision oncology in India faces key challenges, including limited awareness among oncologists in tier-2 and tier-3 cities, who often rely on traditional chemotherapy. Many patients are also unaware of advanced targeted therapies. Further, when effective drugs are identified, they are often unavailable in India. This becomes a cause for frustration among both patients and doctors. Affordability also remains a key challenge as treatments like immunotherapy cost ₹2.5-₹3 lakh per cycle and require multiple rounds. “Although many pharma companies are running a lot of patient support programs where they give free access to a certain extent to some drugs. But overall…it is very expensive,” Goswami pointed out. With increased awareness and reduced costs, the method is expected to become more accessible and be adopted in the coming years. Analysis of Gene Fusions across functional categories and cancer types using 4baseCare’s study. Source: 4baseCare AI at Play 4baseCare has been adopting AI for a number of its use cases, including data interpretation, where it uses AI models to get the right insights from the huge amount of data it generates. “So, right now, we are telling patients and doctors that you can look at your applications and see what you can work on. The next level that we are working on is directly giving the recommendation, best recommendation, best three recommendations in terms of application,” Goswami explained. The startup has also received a grant from the Indian government to develop 3D cell models of tumour tissues in a lab setting. These models are used to test the effects of various drugs on the tumour, and the data collected from these experiments is being used to train an AI model to predict which treatments are most effective for specific types of patients. In addition to this, the startup is working on a novel concept referred to as the CG Twin, which stands for clinical genomic digital twin of cancer patients. The method involves integrating a patient’s clinical and genomic data into a comprehensive profile. This system will allow doctors to compare a new patient’s profile with a database of similar patients, known as twins, and by analysing these matches, doctors can access insights about similar patients’ treatments, outcomes, and risk factors, thereby enabling a more personalised and informed treatment plan based on real-world data and previous cases. “We have done close to 15,000 plus tests now, and we have actually developed a machine learning algorithm, which has been trained in 15,000 patients to build this twin model,” said Goswami, who looks to build a future which will be more evidence-based and outcome-based decisions rather than an empirical way of providing treatment. CG Twin is already being tested by doctors and is slated to be released in the upcoming months. Notably, the concept of digital twins in healthcare is gaining traction. Several healthcare platforms have used NVIDIA’s Omniverse to build a simulated environment of patients to help understand and administer more efficient treatment. The Mission Continues Amidst Struggles Goswami explained that the unique name of the startup is inspired by the four bases of DNA – adenine (A), cytosine (C), guanine (G), and thymine (T) – and the four pillars of cancer care – allied care, technology, global genomic research, and clinical care. “What we realised is that these four bases or these four pillars are working in silos. So, we wanted as a company to bring all these four bases together just like the four bases of DNA to provide cancer care,” he said. Notably, 4baseCare has partnered with well-known hospital chains such as Apollo Hospitals, Fortis Healthcare and Tata Memorial Centre. The journey has not been an easy one, with the startup having faced a fair amount of funding struggles. “When you’re talking about genomics, genomics-driven data and all, there was a lot of apprehension because genomics is something new. And, it is something that’s not easy for everyone to understand.” Goswami even recounted how they had to withstand “a hundred noes” before one “yes”, something that is a common reality for many founders in deep tech. However, eventually, things have a way of falling into place. “You need that one yes from the right people to believe in your vision, and I think that’s what we got with Yali and with Infosys.” Above all, the biggest motivation for Goswami and his team, which has close to 200 members, comes from the profound impact their work has on cancer patients. Goswami believes that seeing patients, once suffering from salivary gland cancer, cancer-free because of their recommendations is what keeps them motivated. “The whole team, right from the logistics team who picks up a sample, I tell them, ‘Guys, you have no idea which sample you might pick up, and that will change the whole family’s life. So right from every level, someone is somehow impacting a family which they don’t even know. So that has kept us going, and…it’s a very exciting journey,” Goswami concluded.","excerpt":"4baseCare’s technology is used by Apollo Hospitals and others, and the startup is now experimenting with CG Digital Twin, which integrates a cancer patient’s clinical and genomic data into a comprehensive profile.","categories":["AI Startups"],"tags":["4baseCare","cancer","DN","genome sequencing","oncology"],"author_name":"Vandana Nair","publish_date":"2024-12-26T19:03:11","publication_year":"2024","word_count":1135,"keywords":["Go","API","machine learning","genome sequencing","AI","innovation","DN","Git","4baseCare","Aim","cancer","Rust","R","oncology","startup"],"extracted_tech_keywords":["AI","machine learning","Aim","R","Go","Rust","Git","API","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/this-infosys-backed-bengaluru-startup-is-beating-cancer-with-genomics-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10043374,"title":"Facebook Trains Guns At Apple &#038; Google Over Default Apps","content":"According to a Comscore study, a majority of mobile devices come preinstalled with default apps from Apple and Google. The report, commissioned by Facebook, found out apps like weather, clock, photos etc come in-built on devices and usually stay put through the entire phone’s lifecycle. Over half of the world’s internet traffic comes from smartphone devices, and over 95% of people never change their default configurations. The statistics show the amount of power Apple and Google have in the app ecosystem. The report alleged big tech gatekeepers undermine user autonomy, hindering innovation and killing competition– it is an unfair business model making use of an unfair advantage over other third-party developers. Source: Comscore • The Verge Comscore’s report is based on the data gathered from apps and websites and a survey on roughly 4,000 people. The results showed 75 percent of the top 20 apps on iOS were from Apple, and 60 percent of the top android apps were from Google. The top 4 apps on both platforms were created by the two companies. Tech goliaths abuse their position to push non-essential apps onto users’ phones. The creators bundle services together with a smooth interface–like how Google ties in Google Meet and Google Workspace to the Gmail accounts. Creating entire ecosystems gives the companies an edge where it isn’t just about individual apps. This vicious cycle hurts user choice as parent apps expand into newer services inside the smartphone, making themselves indispensable. Further, most of the preinstalled apps cannot be uninstalled. The two companies have access to SOTA AI models and are at the frontiers of AI, data and innovation. Default apps generate a lot of data and Google’s business model relies on harnessing user data for targeted advertisements. Earlier, Google was fined €4.34 billion for leveraging Android phones to consolidate its search engine dominance. Facebook vs Apple The Comscore study doesn’t offer a lot in terms of original findings. Also, Facebook has been a vocal critic of Apple, especially with the ad-tracking row. The timing of the release of the report is suspect, with Apple and Google in the government crosshairs for monopolistic practices. The new antitrust bills could change the way big techs function. “This Facebook-financed survey from December 2020 was narrowly tailored to give the false impression that there’s little competition on the App Store. In truth, third-party apps compete with Apple’s apps across every category and enjoy large scale success,”an Apple spokesperson told The Verge.  Apple dismissed the survey methodology as “seriously flawed” and said the report contradicted Comscore’s April 2021 data. In addition, the report showed app usage for a particular time and did not include browsers like Safari and Chrome in its “embedded operating system features” category. Wrapping up The rating of an average iOS app is 4.53. ‘ Default bias’ also comes into the picture. Psychologist Barry Schwartz in his book “Paradox of Choice”, theorised that when presented with multiple choices, instead of bothering to choose (decision fatigue), individual’s will just stick with the familiar (status quo). It is a built-in system of choice in users themselves, and eliminating default apps could be inconvenient. A clock and a calculator for example are essential in a smartphone. The report nonetheless highlights problems with the existing ecosystem. Now, the question is would Android and Apple OS be the same without pre-installed apps? A classic Ship of Theseus conundrum.","excerpt":"According to a Comscore study, a majority of mobile devices come preinstalled with default apps from Apple and Google. The report, commissioned by Facebook, found out apps like weather, clock, photos etc come in-built on devices and usually stay put through the entire phone’s lifecycle.  Over half of the world’s internet traffic comes from smartphone […]","categories":["Global Tech"],"tags":[],"author_name":"Prajaktha Gurung","publish_date":"2021-07-10T15:00:00","publication_year":"2021","word_count":563,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","RAG","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","innovation","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/facebook-trains-guns-at-apple-google-over-default-apps\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10167484,"title":"I Don’t Like the Term ‘Vibe Coding’, Says Replit CEO Amjad Masad","content":"Amjad Masad, CEO of Replit, has expressed his discomfort with the popular term ‘vibe coding’, arguing that it oversimplifies the deeper capabilities of generative AI tools and undervalues the creative process behind building software. On the recent Sequoia’s Training Data podcast, Masad remarked, “I don’t like this term…It just cheapens the possibilities.” Coined by Andrej Karpathy, co-founder of OpenAI, vibe coding refers to using AI tools like Cursor and Windsurf, and inputting ideas to create apps and software without writing a single line of code. While acknowledging that vibe coding might make sense for seasoned developers like Karpathy, who can casually interact with code while relying on AI to fill in the blanks, Masad emphasised that Replit is designed for a fundamentally different workflow. “If you’re starting with Replit, you’re actually not starting from a position of code—you’re starting from an idea,” he said. Masad explained that Replit’s AI agent helps users build that idea into code without requiring them to dive into technical details. “When you’re using Replit Agent, you don’t have the luxury to look at the code,” he added. He drew a distinction between the company’s main product and its Assistant tool, which caters to more advanced users and supports a back-and-forth coding workflow that might be more in line with vibe coding. However, Masad pushed back on the label itself. “If I were to explain Replit, it’s just vibe. Like, don’t code, vibe. Not vibe coding, just vibe.” Beyond terminology, Masad also offered a contrarian take on AGI. While much of Silicon Valley anticipates a future where software engineers are obsolete, Masad believes AI will empower more people to build, not replace them. “We’re all going to have jobs—they’re just going to be very powerful.” He believes that AGI will become “functional”, excelling at economically useful tasks with existing training data, like coding or math, but falling short in creative or novel domains. “Creativity in the sense of coming up with really novel things…I think it will still be the domain of the human,” Masad noted. Reid Hoffman, co-founder of LinkedIn, recently ran an experiment with Replit to “clone LinkedIn” using a single prompt. The result was a surprisingly functional prototype, showcasing the potential of today’s AI to turn ideas into working software. Masad recently said that learning to code may no longer be necessary. He shared a video clip on X which shows him agreeing with Anthropic’s Dario Amodei, who predicted that most of the future code would be AI-generated. Amodei predicted that AI could be writing up to 90% of all code within the next six months. Replit is currently in talks to raise fresh capital at a valuation nearing $3 billion—almost triple its previous worth—reflecting strong investor confidence in AI-driven software development.","excerpt":"“When you’re using Replit Agent, you don’t have the luxury to look at the code.”","categories":["AI News"],"tags":["Amjad Masad","Developers","Replit","Vibe Coding"],"author_name":"Mohit Pandey","publish_date":"2025-04-09T14:23:43","publication_year":"2025","word_count":459,"keywords":["Replit","Anthropic","Go","API","OpenAI","AI","RPA","Amjad Masad","Vibe Coding","ViT","generative AI","CLIP","R","Developers"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Anthropic","R","Go","API","CLIP","ViT","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/replit-ceo-amjad-masad-doesnt-like-vibe-coding\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":40118,"title":"Webinar: Career In Data Science &#8211; How Does One Go About It?","content":"The world is going through what is popularly called ‘digital transformation’ – and this is revolutionizing the way we live- the way we communicate, consume, use time, and work. A lot of what we know today as work is being taken over by progressively intelligent machines. A career in data science and machine learning makes sure you are also part of the revolution. This is likely to have a disruptive impact on the future of human work – not just routine manual blue-collar work but also non-routine cognitive white collar work. A fallout of digitalization is the generation of massive volumes of data in a variety of forms – and the importance of this data for businesses as they try to grow in a highly competitive world. Praxis Business School invites you to attend a webinar on how to crack the data science industry. This interactive Webinar by a senior industry practitioner and a seasoned analytics academician attempts to analyze the implications of this disruption for the human workforce, and the immense opportunities data science offers in terms of a fulfilling and rewarding career. Register for the webinar here. Speakers for the webinar:– 1) Charanpreet Singh Founder & Director –  Praxis Business School Foundation, (B.Tech (Mech) IIT Kanpur, MBA University of Iowa) Charanpreet has 30+ years of experience in organizations like British Oxygen, Tata Steel, PwC, HP and Praxis. Prof. Charanpreet Singh is the driving force behind the success of the Praxis Data Science program. He has professional and academic interests in the areas of education, learning, data science and communication and has been a mentor to career aspirants across domains. 2) Mathangi Sri Head of Data Science – PhonePe, (BE (EEE) MK University, MBA NIT Trichy) Mathangi has 14 plus years of proven track record in building world-class data sciences solutions and products. She has extensively worked on building chatbots and productizing text mining insights. She has 6 Patent grants and 20+ patents pending in the area of intuitive customer service, indoor positioning, and user profiles. She is adept across machine learning, text mining, NLP technologies & tools. She has worked as the Vice President leading the Next Gen-APAC team at Citibank and with organizations like [24]7, Emirates Bank, HSBC, Genpact, and TCS. Listen to the speakers and interact with them on 7th June. Register here to attend a special webinar. Your key takeaways from webinar would be: (Register here) How has data science evolved over the years? What are the skills needed to be successful in data science – soft skills and hard skills? How does one assess if one is suitable for a career in data science? How does one embark on the data science journey? Date: 7th June 2019 (Friday) Time: 6 pm to 7 pm Please register here and invite your friends to register.","excerpt":"The world is going through what is popularly called ‘digital transformation’ – and this is revolutionizing the way we live- the way we communicate, consume, use time, and work. A lot of what we know today as work is being taken over by progressively intelligent machines. A career in data science and machine learning makes […]","categories":["Deep Tech"],"tags":["data science webinar","webinar"],"author_name":"Abhijeet Katte","publish_date":"2019-06-04T07:31:16","publication_year":"2019","word_count":469,"keywords":["data science","Go","machine learning","AI","chatbots","Git","data science webinar","NLP","analytics","GAN","R","webinar"],"extracted_tech_keywords":["AI","machine learning","NLP","data science","analytics","chatbots","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/webinar-career-in-data-science-how-does-one-go-about-it\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10040537,"title":"What Are The Key AI Initiatives Of Indian Government?","content":"Governments worldwide are ramping up investment in AI and figuring out ways to apply and encourage its applications. For instance, the US is expected to commit around $6 billion in AI-related research and development projects in 2021. “We have to make smart investments in technologies and innovations — including in … unmanned systems and artificial intelligence — that will be necessary to meet the threats of the future,” said US President Joe Biden during the 2020 presidential campaign. In Europe, spending on AI is slated to increase by 33 percent between 2020 and 2023, according to International Data Corporation. In 2020, the Indian government increased the outlay for Digital India to $477 million to boost AI, IoT, big data, cybersecurity, machine learning and robotics. India’s flagship digital initiative aims to make the internet more accessible, promoting e-governance, e-banking, e-education and e-health. In the 2019 Union Budget, Finance Minister Nirmala Sitharaman said the government would offer industry-relevant skill training for 10 million youth in India in technologies like AI, Big Data and robotics. According to a report by AIMResearch titled “How The Indian Government Is Championing The AI Revolution”, the use cases of AI in the Indian government include facial recognition and hotspot analysis, biometric identification, criminal investigation, traffic and crowd management, wearables to empower women safety, optimising revenues in the forest, cleaning river, tiger protection, digital agriculture, student progress monitoring and more. Additionally, policy-level initiatives by the Ministry of Electronics and Information Technology (MeitY) and programmes around AI by NASSCOM and Defence Research & Development Organization (DRDO) have laid the groundwork for future disruption and created a roadmap for AI in India. One such initiative was establishing the Centre for Artificial Intelligence and Robotics (CAIR), a laboratory of the DRDO, in 2014 for research and development in AI, robotics, command and control, networking, information and communication security. CAIR shoots for the development of mission-critical products for battlefield communication and management systems. Other recent initiatives include: US-India AI Initiative The Indo-US Science and Technology Forum (IUSSTF) launched the US-India Artificial Intelligence Initiative on 18th March 2021 to foster AI innovation by sharing ideas and experiences, identifying new opportunities in research and development and bilateral collaboration. Applied AI Research Centre in Telangana In October 2020, the Telangana government collaborated with Intel India, International Institute of Information Technology, Hyderabad (IIIT-H) and Public Health Foundation of India (PHFI) to launch INAI (Intel AI), an applied AI research centre in Hyderabad. The centre will focus on solving challenges in India’s healthcare and smart mobility segment. Responsible AI for Youth Responsible AI for Youth is a national programme for government schools to empower the young generation to become AI-ready and reduce the skill gap in India. Established by the National e-Governance Division of MeitY, the platform aims to help the students develop a new-age tech mindset and relevant skill-sets. MCA 3.0 portal The Ministry of Corporate Affairs (MCA) recently launched a new version of its portal, version 3.0, MCA 21, which will leverage data analytics, AI, and ML, to simplify regulatory filings for companies. The idea behind the revamp is to promote ease of doing business and compliance monitoring. AI portal Jointly developed by MeitY and NASSCOM in June 2020, the Indian government launched a dedicated artificial intelligence (AI) portal, India AI is slated as a central hub for everything. The portal will act as a one-stop-shop for all AI-related developments and initiatives in India. National Research Foundation NRF, an autonomous body under the new National Education Policy (NEP) 2020, has been established to boost research across segments, including AI. On 3rd March 2021, while addressing a webinar on effective implementation of Union Budget 2021 provisions, Prime Minister Narendra Modi said, “Fifty thousand crore rupees have been allocated for this. This will strengthen the governance structure of the research related institutions and will improve linkages between R&D, academia and industry.” Promoting AI in schools The National Council of Educational Research and Training (NCERT) is preparing a new National Curriculum Framework for School Education in pursuance of the National Education Policy 2020. This will also aim at introducing a basic course on AI at the secondary level.","excerpt":"Governments worldwide are ramping up investment in AI and figuring out ways to apply and encourage its applications. For instance, the US is expected to commit around $6 billion in AI-related research and development projects in 2021. “We have to make smart investments in technologies and innovations — including in … unmanned systems and artificial […]","categories":["AI Features"],"tags":["AI in government","AI policy","Digital India","Responsible AI for Youth"],"author_name":"Shanthi S","publish_date":"2021-05-21T16:00:00","publication_year":"2021","word_count":690,"keywords":["Go","Digital India","artificial intelligence","machine learning","AI","AI in government","ML","Git","RAG","AI policy","Responsible AI for Youth","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-are-the-key-ai-initiatives-of-indian-government\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10042350,"title":"Researchers Develop Green Microsystems To Generate Electricity From Thin Air","content":"A team of researchers at the University of Massachusetts Amherst have developed an electronic microsystem capable of responding to environmental stimuli without requiring external energy. These ‘green’ self-autonomous organisms are constructed from a novel type of electronics. They can process ultralow electronic signals and incorporate devices that can generate electricity “out of thin air” from the surrounding environment. Jun Yao, an assistant professor of Electrical and Computer Engineering (ECE) and an Adjunct Professor of Biomedical Engineering, co-authored the study with Derek R. Lovley, a Distinguished Professor of Microbiology. The research was published in the journal Nature Communications. Overcoming the challenge The researchers found that incorporating neuromorphic electronics made from memristors in bioelectronic interfaces can provide intelligent responsiveness to environments. However, there is a substantial difference between the environmental stimuli and neuromorphic device’s driving amplitude. The researchers’ key challenge was to construct sensor-driven, integrated neuromorphic interfaces capable of compensating for the inherent amplitude mismatch between sensing and computing signals. The basic approach was to boost sensing signal by carefully selecting sensor structure and stimuli. However, this technique frequently results in sensor form factor or operating environment constraints that preclude a compact or flexible integration, or both. Additionally, certain environmental or physiological signals are intrinsically limited in their amplitude. In comparison, biosystems take a different approach by restricting the computing signal to the thermodynamic limit, enabling bio-computation to respond to a considerably broader spectrum of environmental stimuli. Properties of Protein Nanowires To overcome this challenge, the researchers built the systems on three recently recognised properties of protein nanowires produced from the microbe Geobacter sulfurreducens: Protein nanowire memristors can be driven by impulses as low as 100 millivolts, opening the way for bio amplitude neuromorphic circuits and signal processing.  The Protein nanowires-based devices can capture moisture from the environment, enabling the devices to act as a power source for computing, even if the environmental humidity goes down.It can operate as a sensing component in electronic sensors. (Source: Self-sustained green neuromorphic interfaces. Nature Communications. 12. 10.1038\/s41467-021-23744-2.) The microorganism Geobacter was Lovley’s discovery and has been previously used by the research team to demonstrate protein nanowire-based air generators or “Air-Gen” to generate electricity from the ambient environment. The bio composition of the devices also reflects an exploration of “green” electronics made of renewable, biocompatible, and eco-friendly biomaterials. Additionally, adaptive microsystems are made possible by the protein nanowires, which allow for simple reconfigurable functionality. These qualities represent a significant step forward in the advancement of bio-emulated interfaces and microsystems. Source (Self-sustained green neuromorphic interfaces research paper) The Green components First, components required for integration were examined in a bio-realistic environment, and protein nanowire memristors were fabricated on a flexible substrate. The device was a vertical structure with an insulating layer placed between two electrodes. In terms of programming voltage and current, the device operated at a power level comparable to that of a biological neuron. Devices without protein nanowires could not achieve bio amplitude switching, showing that protein nanowires play a crucial role in enabling bio-amplitude switching. According to the US Army Combat Capability Development Command Army Institute, which funds the research, this research work will create a “Self-sustaining, intelligent microsystem.” Efforts to find new energy sources have now advanced to the point where electronic microsystems can derive energy from their environment and support sensing and computation without the need for external energy sources such as batteries. The findings demonstrate that protein nanowires can be applied to actual purposes. Wearable and environmental applications both benefit from the biocompatibility and eco-friendliness of the “green” material composition. Other neuromorphic functionalities that work in different contexts will be achievable, resulting in a diverse set of self-supported microsystems or intelligent sensors for widespread deployments to support the Internet of Things. In the future, however, further investigations will be needed to discover the total capacity of protein nanowire devices. Access the complete research here.","excerpt":"A team of researchers at the University of Massachusetts Amherst have developed an electronic microsystem capable of responding to environmental stimuli without requiring external energy. These ‘green’ self-autonomous organisms are constructed from a novel type of electronics. They can process ultralow electronic signals and incorporate devices that can generate electricity “out of thin air” from […]","categories":["IT Services"],"tags":["AI Tool"],"author_name":"Ritika Sagar","publish_date":"2021-06-26T13:00:00","publication_year":"2021","word_count":643,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","GAN","AI Tool","R"],"extracted_tech_keywords":["AI","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/researchers-develop-green-microsystems-to-generate-electricity-from-thin-air\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":65374,"title":"Reasons Why You Should Opt For Cloud Business Intelligence","content":"Although organisations are embracing business intelligence (BI) software, they are struggling to make their solutions more robust, easily accessible, and effective. Cloud-based business intelligence solutions help organisations, irrespective of their sizes, to grow their level of competency and value. Cloud BI is the merger of cloud computing and ERP (Enterprise Resource Planning) that delivers simpler data predictive models, visualisations, multi-source data discovery etc. In 2018, the global BI software market was valued at $14.3 billion and is expected to grow at 19.1% CAGR to reach $28.77 billion by 2022. Given that cloud BI is rapid growing, here are some of the most important reasons as to why you should adopt BI in your company: Reducing Implementation Time The traditional business intelligence implementation has always been time-consuming. A decade ago, implementation of BI solutions required customisation of software and dedicated hardware for deployment. These, along with the time required in personnel training, consumed a lot of time to complete. However, with the advent of cloud BI solutions, the process has rather become less expensive and quick. With implementing cloud BI tools, one can concentrate more on their core business activities without the need for setting up software, environment, licensing etc. The service provider is responsible for hosting, managing, and maintaining the software, which reduces money, time and other resources. Advanced Mobility And Higher Adoption Cloud business intelligence helps the decision-makers stay connected to the consumers in real-time. The decision-makers are in touch with the users 24×7 and 365 days. This enables efficient collaboration, decision making, and establishes much-needed communication. The decision-makers and team can access a cloud BI platform from anywhere in the world without having to think about accessibility issues, proximity, effectiveness and system speed. Another important factor for the cloud-based business intelligence solutions is because of the rise in cloud adoption. Cloud fulfils a company’s requirements for storing the colossal amount of data that they collect from different sources. This naturally has increased the amount of data on the cloud, which is analysed using BI SaaS. Cloud-based business intelligence platforms also enable the creation and replication of different views based on hierarchy, roles, and geography of the decision-makers. Scalability Scalability is the ability to make a software application available for use to those who need it anywhere. Scalable applications help in increasing users’ adoption with ease with the changing requirements. Scalability on top of adoption may also increase the user engagement time in self-service activities. Cloud BI solutions offer scalability beyond this definition. One can also decrease or increase their system and storage quickly ensuring that one pays only for what they need or use. This guarantees that the users have better functionality at a fair price. Reliability The cloud-based business intelligence solutions provide excellent reliability and improve through multiple redundant platforms like Birst, Domo, Tableau Online etc. These platforms provide reliable and secure locations for data storage, and the resources can be accessed by many users around the world. The vendors also take care aspects like redundancy, scalability and distribution of the platform. Lower Cost The cloud BI solutions include powerfully built data analytics and reporting tools for businesses to upgrade sales and marketing activities. These tools also increase collaborative campaigns between sales and marketing teams of the company. Cloud business intelligence solutions generate maximum business value by enabling employees with self-service analytics, providing IT with the scalability and flexibility it needs to reduce maintenance and security costs. This allows IT departments to concentrate more on innovative solutions and systems that drive significant business growth. As far as the customers go, because of the cloud-based BI solutions, in research, they mentioned that they were able to reduce demands on IT while moving to cost-effective and scalable solutions. Some cloud BI solutions you can choose from apart from the popular ones like Tableau Online, Microsoft Power BI: ThoughtSpot: Offers an AI-driven analytics platform and combines relational search with a custom-built, in-memory relational data cache to speed up queries. ThoughtSpot connects any on-premise, cloud, big data, or desktop data source and has introduced its new voice-activated analytics interface called SearchIQ. Periscope Data: IT takes advantage of augmented analytics revolution. It provides new visual data discovery capabilities and allows business users with direct access to cleaned data sets. Reltio: This platform allows businesses to manage data by making use of continuous data organisation and recommended actions. It offers a Self-Learning Data Platform that organises data from a wide variety of sources and formats to create a data set that has personalised views of users across business departments. Also Read: Are Analytics Dashboards Worth Learning","excerpt":"Although organisations are embracing business intelligence (BI) software, they are struggling to make their solutions more robust, easily accessible, and effective. Cloud-based business intelligence solutions help organisations, irrespective of their sizes, to grow their level of competency and value. Cloud BI is the merger of cloud computing and ERP (Enterprise Resource Planning) that delivers simpler […]","categories":["AI Features"],"tags":["Business Intelligence","cloud business intelligence solutions","Cloud Computing","how to implement business intelligence"],"author_name":"Sameer Balaganur","publish_date":"2020-05-17T11:00:00","publication_year":"2020","word_count":763,"keywords":["how to implement business intelligence","Go","cloud business intelligence solutions","API","big data","ELT","AI","cloud computing","Scala","RAG","Cloud Computing","analytics","Business Intelligence","R"],"extracted_tech_keywords":["AI","analytics","RAG","cloud computing","R","Go","Scala","API","big data","ELT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/reasons-why-you-should-opt-for-cloud-business-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10047134,"title":"The AI Tool That Spots Art Forgery","content":"Artwork authentication requires expert connoisseurs, and it quite often becomes impossible to establish authenticity amid conflicting expert opinions, subjective interpretations, and scarce evidence. Scientific tools step in here to assist with setting a painting’s age and subsurface details. But, these tools haven’t been able to crack the code to identify the artists yet. Convolutional neural networks are used to analyse images in facial recognition systems and self-driving vehicles; Steven Frank talks about utilising them to identify artists based on their style and technique. In this article, we discuss the AI tool created by Frank to find art forgeries. The Problem Creating neural networks requires many training samples, but even the most prolific artists have not managed to produce thousands of paintings. Another issue faced with analysing paintings is the image resolution challenge. Simultaneously, high-resolution images of artworks are too large for CNNs. Smaller-sized images lack the information needed to ensure impartiality. The Design Frank, an AI graduate, and his wife, Andrea, an art historian, took it upon themselves to find an AI-based solution for art forgeries. They used the ‘digitised biopsy slides’ approach used by pathologists, where practitioners have made pixelated images tractable for CNNs by creating smaller fragments or square tiles. This allows the creation of several training tiles from a single image. Initially, the duo trained their neural network with Dutch artist Rembrandt’s portraits. The training set included randomly chosen 50 original portraits and 50 portraits by other artists. While the system could differentiate when the other 50 artists had a different style altogether, it couldn’t set apart original Rembrandt from students and admirers, much fewer forgers. To train the AI tool to distinguish similar work from Rembrandt’s original, the researchers compiled a dataset with non-Rembrandt entries ranging from close dupes to distinguishable paintings. To ensure that the tile contributed reliably to classification, the entropy cutoff was at least as much as that of the entire image. This tied the entropy threshold to the character of the painting. The selected tile size was 450 x 450 pixels, consisting of features experts rely on while judging a painting’s authorship. The dataset consisted of 76 Rembrandt and non-Rembrandt paintings that were shuffled in four ways into separate sets of 51 training and 21 test images to cross-validate the results. They successfully created a five-layer CNN that had more than 90 per cent accuracy when distinguishing Rembrandt from his students and forgers. Dubbing the device, ‘The A-Eye’, Frank and Andrea put it to test on Vincent van Gogh’s landscapes – this time with a larger dataset. They used 152 van Gogh and non-van Gogh paintings divided four ways into sets of 100 training and 52 test images. The team used smaller tiles to achieve high accuracy on their test set. This led them to the conclusion that the signature style of the artist’s work, i.e. the distinctive feature sizes that facilitate accuracy in CNN based classification, is particular to individual artists. Lastly, to predict the artisanship of the painting, the duo assigned a probability to different tiles with their pixels. Since more than one tile intercepts a pixel, they created an average of the relevant tile-level possibilities and used it to determine the value of that pixel. This resulted in a probability map of regions that are more or less likely to be an original painting of the artist. The A-eye managed to identify areas of doubt in the face and background of Rembrandt’s portraits of his wife. This accords with the belief held by many scholars that those regions have been overpainted later by someone else. Additionally, the duo evaluated van Gogh’s and Rembrandt’s works under controversy, and in all but one case, their classifications matched the current scholar consensus. Their latest experiment has left scholars with food for thought regarding the consensus on one of Rembrandt’s paintings. ‘The Man With The Golden Helmet’ was de-attributed by the Staatliche Museum’s scholars believing there was an inconsistency in the paint handling. However, contrary to that, the A-Eye has concluded that the painting is Rembrandt’s after narrowing the CNN analysis on the painting’s features. Probability map: Steven & Andrea’s IEEE feature The results of the A-eye on a painting by Rembrandt (left), the painting ‘The Man With The Golden Helmet’ formerly attributed to him (middle), and Salvator Mundi (right). Hot colours represent areas the classifier determined had high probability of having been painted by the artist associated with the work. Art Recognition Another example of this; A Swiss company, Art Recognition, leverages AI to assess the authenticity of a work by using neural networks to analyse an artist’s brush strokes and techniques in a high-resolution photographic reproduction of the painting. In its application to assess van Gogh’s controversial self-portrait, the A-eye found the self-portrait to be 97 per cent probable authentic van Gogh. This was later confirmed by the van Gogh Museum in Amsterdam’s years worth of studies. “We started by training our deep convolutional neural network to learn van Gogh by using hundreds of original images from de la Faille Catalogue Raisonné,” according to the company’s case study. “To help the system better distinguish forgeries, we also fed in well-known (labelled) fakes like, for example, the famous Wacker forgeries.” The company generated a heat map to provide visual evidence on the algorithm’s evaluation, with the hotspot spectrums highlighted in red. The use of AI in finding art forgeries or dupes is still a developing technology, but implementations like these are opening the doors to more sophisticated use of CNN with paintings.","excerpt":"Initially, the duo trained their neural network with Dutch artist Rembrandt’s portraits.","categories":["AI Features"],"tags":["ai generated images"],"author_name":"Avi Gopani","publish_date":"2021-08-27T10:00:00","publication_year":"2021","word_count":920,"keywords":["ai generated images","Go","programming_languages:R","AI","neural network","RPA","ML","Git","RAG","CNN","R"],"extracted_tech_keywords":["AI","ML","neural network","RAG","R","Go","Git","CNN","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-ai-tool-that-spots-art-forgery\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":5746,"title":"Analytics India Magazine Reportcard – Year 2","content":"It’s been 2 years of existence for Analytics India Magazine today. The last year have been much more fruitful than the first year. The first year was when we established ourselves as an online resource portal for analytics in India. In second year we reached out to almost the whole industry and the impact was visible. Each year we publish a report card of how we have done in the past year, some key metrics and comparison from earlier years. Read last year’s reportcard This year, we reached out to 3 times more users than a year before. The number of visits tripled as well, and the page views almost doubled. In terms of cities, Bangalore recorded that highest number of visitors at 24%, followed by Delhi and Mumbai at 9% each. Being an India focused portal, its not surprising that 80% of our visitors are from India. An interesting insight – 53% of users come from Chrome as opposed to just 17% from IE. [divider] In terms of channels, 58% of our users come through Organic Search and 17% are direct. Facebook (both mobile and desktop) is the biggest referrer at 43% of referral traffic. Linked refers 24% of referral traffic. Our Facebook page likes have increased from under 1000 to more than 3,000 this year. Other Metrics","excerpt":"It’s been 2 years of existence for Analytics India Magazine today. The last year have been much more fruitful than the first year. The first year was when we established ourselves as an online resource portal for analytics in India. In second year we reached out to almost the whole industry and the impact was […]","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2014-05-20T16:54:30","publication_year":"2014","word_count":219,"keywords":["programming_languages:R","AI","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/analytics-india-magazine-reportcard-year-2\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10022511,"title":"How This Bangalore-Based Startup Provides Automation For Media Organisations","content":"For this week’s startup feature, Analytics India Magazine spoke to Kavita Shenoy – Founder & CEO of Voiro to understand how the Bangalore-based startup is leveraging emerging technologies to bring automation and intelligence to leading media organisations across the world. Founded in 2014 by Kavita Shenoy and Anand Gopal, Voiro is a SaaS-based media technology company. More specifically, Voiro is a category-creating data technology company that empowers media organisations to make data-driven business decisions. Flagship Products According to Shenoy, the product suite of Voiro is created with the vision of making media companies data-first. It is built to tackle the complex world of multiple revenue channels and provide the best data security and protection, critical for a data-driven, content-centric industry. Shenoy said, “We engineer innovative products to track, report as well as translate all data that contribute to the revenue. The company’s solution is a strategic amalgam of a microservices and Application Program Interface (API) stack which solves for Enterprise Customers in Media, OTT and E-commerce. The Phoenix UI is being designed for power users, the data-hungry mavericks and is a truly simplified experience that empowers data-driven decisions.” With plenty of data collected every minute from various avenues, the company’s technology helps other organisations and publishers analyse their revenue streams and have a single view of their revenue horizon. At present, the company is also in the process of creating NPS and micro surveys for each customer and a concerted, comprehensive communication strategy to keep the customer glued. What’s The Differentiator Shenoy stated, “Monetising Content is key to revenue growth in digital advertising, OTT and ecommerce. Our customers, broadly speaking, have three main challenges – market challenges, technology challenges and data challenges. Market challenges include a lack of historical and real-time analytics on ad revenue and content ROI. Technology challenges involve the complexity of ad-tech, issues of tech stacks being fragmented and workflows being disparate and lack of domain experts to manage events like live sports. Lastly, there are data challenges which include a lack of business intelligence from terabytes of data across multiple teams and no ready reckoners specific to media such as optimisers, predictive analytics and revenue reconciliation.” She added, “Voiro’s solution is a strategic blend of a microservices and Application Program Interface stack that solves for various customers in media, OTT as well as ecommerce. Our product suite blends the capabilities of revenue reconciliation, CRM, deep analytics and insights. Voiro falls into Ad-Tech Marketing Analytics and CRM verticals, which makes us unique as well as gives us the potential to play in quite a few spaces.” Core Tech Sack The company uses the following tools – For cloud computing: AWS Microsoft Azure For frontend: AngularStack included ApacheDjango Python Web FrameworkPostgreSQL For Big Data: S3 Athena Recent Funding The company has already raised ₹2.5 crores in a pre-Series A round from crowdfunding platform 1Crowd’s investor community and angel fund. Potential Competitors Talking about competitors, Shenoy said global ad tech platforms, CRM giants and other analytics and tech companies in the space of revenue optimisation, automation, and revenue reconciliation are considered as potential competitors. Roadmap “Voiro is set to take on the global stage – with the market turning strongly towards their value proposition and is excited to push ahead to convert global opportunities, invest strongly in data engineering and build a world-class SaaS product from India for media organizations around the globe,” said Shenoy on a concluding note.","excerpt":"For this week’s startup feature, Analytics India Magazine spoke to Kavita Shenoy – Founder & CEO of Voiro to understand how the Bangalore-based startup is leveraging emerging technologies to bring automation and intelligence to leading media organisations across the world.   Founded in 2014 by Kavita Shenoy and Anand Gopal, Voiro is a SaaS-based media technology […]","categories":["AI Startups"],"tags":["Automation","django","django python"],"author_name":"Ambika Choudhury","publish_date":"2021-03-21T11:00:00","publication_year":"2021","word_count":570,"keywords":["PostgreSQL","AWS","AI","cloud computing","Azure","Automation","django","RAG","microservices","Python","django python","analytics","predictive analytics"],"extracted_tech_keywords":["AI","analytics","RAG","predictive analytics","cloud computing","AWS","Azure","microservices","PostgreSQL","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-bangalore-based-startup-leverages-ai-to-provide-automation-for-media-organisations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":23659,"title":"Meet ConvNetQuake, World’s First Neural Network That Detects And Locates Earthquakes","content":"Earthquakes are quite unpredictable — they can strike anywhere and anytime with little or no warning. Most of the earthquake detection methods are designed for moderate and large earthquakes. As a result, they tend to miss many of the low-magnitude earthquakes that are masked by seismic noise. Sometimes even traditional approach to earthquake detection fails to detect events buried in the modest levels of seismic noise. Earlier this year, a team of researchers from Harvard University and Massachusetts Institute of Technology (MIT) led by Thibaut Perol found out a way to use neural networking method for monitoring earthquakes. They published their research in Science Advances in February centred on Oklahoma, US, where they detected more than 17 times more earthquakes than were recorded in the Oklahoma Geological Survey earthquake catalogue. “Over the last decades, the volume of seismic data has increased exponentially, creating a need for efficient algorithms to reliably detect and locate earthquakes. Today’s most elaborate methods scan through the plethora of continuous seismic record, searching for repeating seismic signals,” the researchers said. First Neural Network To Detect Earthquakes The researchers came up with Convolutional Neural Network for Earthquake Detection and Location (ConvNetQuake), which is the first neural network in the world to detect and locate earthquakes. It is a deep convolutional network that takes a window of three-channel waveform seismogram data as input and predicts its label either as seismic noise or as an event with its geographic cluster. It is trained on a large dataset of labelled raw seismic waveforms and learn a compact representation that can discriminate seismic noise from earthquake signals. The waveforms are analysed with a collection of nonlinear local filters. “ConvNetQuake is more accurate than just state-of-the-art algorithms and runs orders of magnitude faster. In addition, it outputs a probabilistic location of an earthquake’s source from a single station,” says the research. According to the paper, the neural network is ideal to monitor geothermal systems, natural resources reservoirs, volcanoes and seismically active and well-instrumented plate boundaries such as the subduction zones in Japan or the San Andreas Fault system in California. ConvNetQuake Architecture The architecture for the earthquake location detection is similar to traditional covnet architectures. The network takes a two-dimensional tensor as an input which represents the waveform signal data. This data is transformed and sent through eight layers of feedforward convolutional network layers and ends with a layer that throws out the scores for each class. The fully-connected layer does linear processing supported by weights and at the other contains a softmax which puts out a normalised probability score. It was found that this network was more computationally efficient than some other architectures. This particular model is also very desirable because of it avoids overfitting. How ConvNetQuake Is Different From Other Earthquake Detection Methods The researchers compared the earthquake detection performance with ConvNetQuake with autocorrelation and Fingerprint And Similarity Thresholding (FAST). Autocorrelation relies on waveform similarity (repeating earthquake) and FAST relies on fingerprints similarity. Both the techniques do not require a prior knowledge on templates and they provide a detection but no location. The researchers also found that ConvNetQuake is highly scalable, can easily handle large data sets and is approximately 13,500 times faster than autocorrelation, and 48 times faster than FAST. The scientist ran ConvNetQuake and FAST for a month-long continuous time series and found that ConvNetQuake runtime is 4 minutes and 51 seconds, whereas that of FAST is 4 hours and 20 minutes. CAPTION: Scaling properties of ConvNetQuake and other detection methods as a function of continuous data duration. (A) Runtime of the three methods where 1.5-hour one-time training is excluded for ConvNetQuake and where FAST’s runtimes include the feature extraction (38%) and database generation phases (11%). (B) Memory usage of FAST and ConvNetQuake. How Accurate Is ConvNetQuake In Detecting Location For the earthquake detection, the researchers used 209 earthquake events cataloged by the Oklahoma Geological Survey. Among the 1,31,972 noise windows, the neural network was able to classify 1,29,954 noise windows correctly and misclassified 2,018 earthquake events with 94.8 percent precision. When the scientist evaluated the location performance of the network, they obtained 74.5 percent location accuracy and when they experimented with a larger number of clusters they obtained 22.5 percent location accuracy. However, the limitation of the system is the size of the methodology is the dataset required for good performance for earthquake detection and location. “Data augmentation has enabled great performance for earthquake detection, but larger catalogs of located events are needed to improve the performance of our probabilistic earthquake location approach. This makes the approach ill-suited to areas of low seismicity or areas where instrumentation is recent but well-suited to areas of high seismicity rates and well-instrumented.” To sum up, though the new neural network is still at the nascent stage but once deployed it will be able to provide very rapid earthquake detection and location, which is useful for earthquake warning.","excerpt":"Earthquakes are quite unpredictable — they can strike anywhere and anytime with little or no warning. Most of the earthquake detection methods are designed for moderate and large earthquakes. As a result, they tend to miss many of the low-magnitude earthquakes that are masked by seismic noise. Sometimes even traditional approach to earthquake detection fails […]","categories":["IT Services"],"tags":["Neural Network"],"author_name":"Smita Sinha","publish_date":"2018-04-15T06:03:25","publication_year":"2018","word_count":817,"keywords":["Neural Network","Go","API","data augmentation","TPU","programming_languages:R","AI","neural network","Scala","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","neural network","TPU","R","Go","Scala","API","data augmentation","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/it-services\/meet-convnetquake-worlds-first-neural-network-that-detects-and-locates-earthquakes\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10094188,"title":"Council Post: AI’s Evolutionary Journey &#8211; Model-Centric to Data-Centric to Decision-Centric","content":"Model-centric and data-centric are two popular approaches in the field of artificial intelligence (AI) that emphasize the importance of either the model or the data in the development and performance of AI systems. Some examples in the model centric approach: Convolutional neural networks (CNNs) for image recognition and recurrent neural networks (RNNs) for natural language processing are examples of deep learning models. These models are built with intricate architectures and several layers to learn and extract characteristics directly from the input. GANs, or generative adversarial networks: A generator network and a discriminator network are the two main components of GANs. While the discriminator aims to discriminate between authentic and fraudulent samples, the generator makes artificial data samples. GANs concentrate on producing accurate data by harnessing the power of the model. All the above training techniques and model hyper parameter selection for the above deep learning model essentially try to tweak the model to best-solve the task using a given dataset. Similarly, some examples for the data centric approach can be: Transfer Learning: Pre-trained models that have learnt generic characteristics via training on big datasets are used in transfer learning. The requirement for substantial training on task-specific data is subsequently reduced since these models are then fine-tuned on specific tasks with less data. Data Augmentation: To provide more training examples, data augmentation approaches entail adding transformations or alterations to existing data sets. For instance, flipping, rotating, or adding noise to photos might improve the variety of the training data for image classification. Both approaches have their merits and trade-offs. In one hand, Data-centric AI can be advantageous when dealing with limited prior knowledge or rapidly changing domains. By focusing on data collection and curation, it aims to capture the diversity and complexity of real-world scenarios, allowing models to learn from the data directly. This approach can be more flexible and adaptive to changing environments. However, data-centric AI heavily relies on the availability of high-quality and representative datasets, which can be challenging to obtain. On the other hand, Model-centric AI can be effective in situations where domain expertise and prior knowledge are essential, such as in complex problem domains like natural language processing or computer vision. It allows researchers to design specific architectures and algorithms tailored to the problem at hand, resulting in high accuracy and performance. However, model-centric AI often requires significant computational resources and extensive training time.  A famous example of this model-centric AI approach, exemplified by large-scale language models like GPT-3, require significant computational resources and extensive training time due to their complexity and the vast amount of data they process. The training process involves pre-training on unlabeled text data and fine-tuning on specific tasks. The computational demands stem from the models’ large number of parameters, necessitating high-performance infrastructure. The move towards decision centric AI Model centric works when the data is clean, well curated etc. Before trying to improve any model the data quality has to be top notch. However, in the real world, data is not ideal. Hence, a decision centric approach is more practical. In this sense, a decision-centric approach is crucial because it fosters creative problem-solving at those “moments of truth” while simultaneously enhancing operational effectiveness. To enable users to take the appropriate action at the appropriate moment, people, rules, data, and processes must all work together. Companies that want to react quickly and effectively to dynamic market developments must adopt a decision-centric strategy. It makes it possible for the organisation to swiftly adjust to changes in operational procedures and company strategy in addition to learning new things. As a result, it is an essential component of any toolkit for business agility. It promotes uniformity in decision-making across the board and offers crucial feedback loops. The framework offers a context for placing judgements into “business moments,” regardless of whether these decisions are backed by AI, machine learning, business rules, or any other technology technique. Decision-centric AI represents the next stage in the evolution of AI, where the focus shifts from solely making predictions to actively making informed decisions. This approach involves integrating data analysis, machine learning, and optimization techniques to generate optimal decisions based on desired objectives and constraints. Decision-centric AI will enable organizations to make smarter, data-driven decisions that align with their goals and requirements. By incorporating contextual information, business constraints, and user preferences, decision-centric AI systems will optimize decision-making processes across diverse domains such as healthcare, finance, supply chain management, and customer service. It will enhance transparency and explainability, allowing users to understand the reasoning behind the decisions made by AI systems. This will build trust and facilitate the acceptance and adoption of AI technologies. Ultimately, decision-centric AI has the potential to revolutionize how organizations operate, enabling them to make more accurate, efficient, and impactful decisions, thereby driving innovation, productivity, and competitiveness. It is also important to remember that decision centric AI is about augmenting human decision-making rather than replacing it. The evolution in AI system development and application from model-centric to data-centric to decision-centric methods symbolises this development. At first, the emphasis was on creating intricate models that could learn directly from data. However, the shortcomings of this strategy prompted a move towards a focus on data gathering and curation. The value of high-quality datasets was acknowledged by the data-centric approach. The decision-centric approach was developed with the aim of generating favourable results based on facts in real time and being able to justify those judgements. Unknown circumstances are a daily problem for businesses, but by adding quantified pattern data and adaptable scenarios to the mix, choices may be made more quickly, consistently, and confidently. The shift from model-centric AI to data-centric AI has allowed organizations to leverage vast amounts of data for improved performance and adaptability. However, the evolution does not stop there. The emerging trend of decision-centric AI holds great promise in transforming the way we make decisions by integrating data analysis, machine learning, and optimization techniques. Ultimately, decision-centric AI has the potential to revolutionize organizations; operations, driving innovation, productivity, and competitiveness through more accurate and impactful decision-making processes. The future of AI lies in the integration of data, models, and decisions, paving the way for a more intelligent and informed world. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"The emerging trend of decision-centric AI holds great promise in transforming the way we make decisions by integrating data analysis, machine learning, and optimization techniques.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Arnab Ghosh","publish_date":"2023-05-31T12:00:00","publication_year":"2023","word_count":1076,"keywords":["data science","machine learning","artificial intelligence","AI","neural network","computer vision","RAG","Aim","deep learning","analytics","AI Tool"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","computer vision","data science","analytics","Aim","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ais-evolutionary-journey-model-centric-to-data-centric-to-decision-centric\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10040941,"title":"How Machine Learning Forms The Backbone Of Neo-Banking Startup StashFin","content":"“Ever since the beginning of the pandemic, the complexity of frauds has gone up many folds,” said Parikshit Chitalkar, co-founder of digital lending platform StashFin. Founded in 2016 by Tushar Aggarwal, Parikshit and Shruti Aggarwal, StashFin targets young urban grey collar employees. The platform has acquired a million customers so far. The digital lending industry is projected to reach $20.31 billion market cap by 2027, growing at a CAGR of 16.7 percent between 2020 and 2027. Leveraging technology StashFin works like a neo-banking platform providing credit line cards powered by VISA and MasterCard. “Think of it as an overdraft facility banks provide,” Parikshit said. The platform uses technology to facilitate customer onboarding and providing seamless experiences. A potential customer has to follow a five-step process to register on StashFin. First step is to share the picture of their Pan card and Aadhaar card. Then, the ML model extracts details to fill the registration form. The user has to enter an OTP, post which the underwriting model triggers in the background, and decide: Whether to approve or reject a certain user for a credit line Amount of credit a user is authorised for “The system uses close to 3,000 variables to analyse a customer before announcing the decision. These variables can range from bureau score or credit score to analysing how desperate a user is for credit, the time at which they are applying for credit, their banking history, etc,” Parikshit said. StashFin uses basic models and standard APIs to examine the validity of official documents. The platform asks the user to click a selfie with the documents provided and uses a TensorFlow-based image recognition model to cross verify. “Doing this is not easy since the Pan card image would have been taken somewhere at a fairly young age, the selfie would be recent, and the Aadhaar card would have a picture clicked somewhere in between,” Parikshit said. StashFin’s classification model raises a red flag if the image is not consistent. Typically, fraudsters try to confuse the ML model in two ways: When the user clicks a selfie with another picture of themselves at the back, thus, confusing the model by having two images in the same frame Wearing a t-shirt or jacket with a face printed on them. In such cases, the model either provides a wrong result or returns a null value. When returned with a null value, StashFin does a manual check. Stringent checks “We also have to ensure that we do not get false positives which might result in a good customer getting mistakenly detected,” he added. For this, the digital lending platform does backtest on customers with good credit scores and are only a couple of EMIs away from completing their payments. “We backtest on that population since the fraud rate or ID theft is next to zero. If it throws a single positive, we try to work to minimise it,” Parikshit said. Once the credit is allocated, the ML model figures out any anomalies in the pattern of usage. It uses up to 35 different micromodels to analyse and predict usage patterns. Some of these models are also self-reinforcing. Each time a user makes a payment, the model looks at where they stand in their credit cycle; if they have paid on time or not. The ML model makes decisions based on a user’s payment history, like reducing the interest rate for people consistently paying on time. Finally, StashFin uses ML models to help users make informed financial choices. For instance, the platform guides users to nearest ATMs, and stores where Stashfin users can redeem their reward points. StashFin has built all its tech in-house. It has an engineering and data scientists’ team constantly working on improving the product. Road ahead With a million customers on-boarded, StashFin plans to increase the number by up to 7x in the next 18 to 24 months. “We want to be sure that our platform is ready to handle that kind of scale,” Parikshit said. Soon processes in the fintech industry will be digitised and KYC technology is in for an overhaul. Many paper processes have to die to make engineering efforts to automate, monitor and alert processes, he added.","excerpt":"“Ever since the beginning of the pandemic, the complexity of frauds has gone up many folds,” said Parikshit Chitalkar, co-founder of digital lending platform StashFin. Founded in 2016 by Tushar Aggarwal, Parikshit and Shruti Aggarwal, StashFin targets young urban grey collar employees. The platform has acquired a million customers so far. The digital lending industry […]","categories":["AI Startups"],"tags":["machine learning document classification"],"author_name":"Debolina Biswas","publish_date":"2021-05-29T18:00:00","publication_year":"2021","word_count":701,"keywords":["Go","API","machine learning document classification","AI","ML","image recognition","Git","RAG","ai_frameworks:TensorFlow","TensorFlow","R"],"extracted_tech_keywords":["AI","ML","TensorFlow","RAG","image recognition","R","Go","Git","API","ai_frameworks:TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-machine-learning-forms-the-backbone-of-neo-banking-startup-stashfin\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172855,"title":"‘Every Single AI Researcher Making $10-100 Million is a Dota 2 Player’","content":"Forget Ivy League degrees or internships at Google—the new secret to cracking the top AI jobs these days? Dota 2. If you are not playing, you are not going to make it. The notoriously challenging and endlessly complex Multiplayer Online Battle Arena (MOBA) isn’t just a video game anymore; it has become a full-blown talent filter for the most elite AI researchers in the world. “What’s crazy about all of these top AI researchers making $10-100 million is that literally every single one of them is a Dota 2 player,” wrote a user in a viral post on X. “If you played Dota 2, there is a good chance that you were in a lobby with either them or me.” It might sound like a joke, but it’s not. Today’s top AI researchers—the ones earning million-dollar salaries—have likely spent over 5,000 hours getting flamed by strangers in a dimly lit room. Just ask Greg Brockman, Ilya Sutskever, or Jakub Pachocki, who are key figures at OpenAI and the masterminds behind OpenAI Five, the bot that crushed human pros at Dota 2. In the 2018 tournament, Five failed to defeat professional human gamers. Despite having the training and experience of over 180 years, the AI was unable to achieve the feat. However, that’s another story altogether. Even Sam Altman, while not a player himself, backed the project that made Dota 2 the proving ground for reinforcement learning, even though it did not win. These guys didn’t just study Dota 2; they lived it. Even Elon Musk is a dedicated player, once claiming he ranked among the 50 top players of all time. How he manages that while running five companies at the same time remains a mystery, and a topic of speculation across X’s feed. Dota 2 is All You Need “There is a 100% correlation between the most cracked people I know and whether they played Dota 2 or not,” Upamanyu Acharya wrote on X. “And it’s unique to this game; doesn’t apply to anything else.” So why this game? Why not chess, Go, or League of Legends? Because Dota 2 is, as one user put it, “a multi-decade AI talent filter disguised as digital pain”. A single Dota 2 match spans 80,000 ticks—agents (or players) must plan and adapt hundreds of steps ahead. That’s a goldmine for researchers building long-horizon planning models for real-world applications, such as robotics or autonomous driving. Notably, OpenAI acquired the company Global Illumination, which was meant for making agents based on games using reinforcement learning. Dota 2’s ‘fog of war’ mimics real-world uncertainty, making it a playground for inference algorithms. With ~1,000 valid actions every quarter second, it forces researchers to build models that work in environments messier than any board game. Five heroes per team play cooperatively while outmanoeuvring the opposition. It’s the norm in distributed AI systems and swarm robotics. Dota also demands split-second decision-making every 80 milliseconds, training algorithms to react under pressure. Valve’s open Bot API means you can run thousands of games in parallel, generating hundreds of years’ worth of data per day for pennies. That’s how OpenAI Five was trained—spinning up massive reinforcement learning experiments in the cloud while playing one of the most punishing games ever created. More Than a Game, It’s a Culture “Holy shit. The highest-earning guy I know is a Dota 2 player as well. F***,” wrote Mario Hachemer, echoing a sentiment that’s no longer anecdotal but approaching statistical fact. Moreover, there’s the factor of cultural overlap. Engineers like systems. Dota 2 is a system that is mechanical, unforgiving, but also beautifully complex. Its steep learning curve, data-rich replays, and obsession with optimisation mirror the mindset of top researchers. “Every cracked person I know was either addicted to RuneScape or Dota 2. No exceptions,” another post stated. Even hiring managers are catching on. “I played Dota 2 way too much,” confessed one X user. “I also think if I ever hire again, I will test people by playing some matches in ranked together with them. Nothing shows better how you react in stressful situations.” If that sounds like a joke, it isn’t. Dota 2 teaches patience, precision, teamwork, adaptability, and how to manage rage, which are exactly the traits AI researchers need. Dota 2 wasn’t just OpenAI’s testbed. Before ChatGPT took over the world, OpenAI spent years just trying to get bots to survive on the Dota map. And in doing so, they built a team of researchers who weren’t just skilled, they were hardened by thousands of hours of digital warfare. That’s why Dota 2 keeps showing up in resumes, in X threads, and increasingly, in salaries with a lot of zeros. So if you’re wondering what separates a $200,000 AI engineer from a $10 million one? Check their Steam profile, not GitHub.","excerpt":"If you’re wondering what separates a $200,000 AI engineer from a $10 million one, check their Steam profile, not GitHub.","categories":["AI Features"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-07-04T10:42:15","publication_year":"2025","word_count":800,"keywords":["Go","ChatGPT","OpenAI","AI","ML","Git","RAG","Aim","GitHub","R"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Aim","RAG","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/every-single-ai-researcher-making-10-100-million-is-a-dota-2-player\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171658,"title":"Soket AI’s Plan to Build 7 Bn Open Source Indic LLM Within 6 months","content":"After selecting Sarvam AI as the first startup, IndiaAI Mission chose Soket AI Labs, Gnani AI, and Gan AI last month as part of the mission to build India’s sovereign AI. While Sarvam has already released a few updates and models, others haven’t released anything yet. The companies are yet to receive the promised support from the government (GPUs). Soket AI Labs, led by CEO Abhishek Upperwal, is quietly building what could become one of India’s most ambitious AI projects: a 120-billion parameter language model trained on India-centric datasets from scratch under its Project EKA. But the journey to this number is anything but linear. The company plans to keep it open-source and optimise it for sectors such as defence, healthcare, and education. “It will take time,” Upperwal told AIM. “It won’t be a one-shot deal. We will scale it up to 120 billion parameters and will have to scale it little by little.” Upperwal said they plan to make this accessible to all, from researchers to startup founders. The team is building in public under the COOM framework, publishing transparent updates, and committing to energy-efficient, culturally-representative training practices. What’s the Roadmap? The team is taking a phased approach, beginning with models as small as 1-2 billion parameters and gradually scaling to 7 billion and then 30 billion. These early iterations will be used to test architecture and data alignment—critical steps before pouring massive compute into the final models. Upperwal said that the 7 billion model would most likely be ready in 5-6 months, and the team can scale it to 120 billion within the 10th month. “We have already done a 1 billion model. So we have an idea. We can scale it to 7 billion,” said Upperwal, speaking about the Pragna-1B model released last year. Soket will iterate in stages, not chasing leaderboard scores, but building something reliable from the ground up. “I think, from a sovereignty point of view, defence would be an important aspect because defence cannot use DeepSeek. If they go to use DeepSeek they will show Arunachal Pradesh as part of China,” said Upperwal, pointing to the geopolitical risks of using foreign models, especially those from China. Besides, security concerns make cloud-based models unsuitable for defence. Soket’s plan is to deploy models in secure, air-gapped environments with on-device capabilities. In education, they are already working with AI CoEs aiming to digitise archives, books, and curriculum content, and to collaborate with ministries and academic institutions. A New Data Foundation for India At the heart of Soket’s strategy is an unprecedented data effort focused on Indian languages, which have historically been underrepresented in large AI models. The team is separating the data strategy into two categories: existing and non-existent. “We are applying OCR to documents. We are applying ASR models on videos and audio. We are extracting content from that.” The data strategy is intensely India-focused. Pretraining will be done on regional knowledge—government sites, legal records, school curricula—alongside global corpora like scientific papers and code. Post-training datasets will cover domain-specific reasoning tasks, including law and agriculture, while evaluation will involve creating new benchmarks for Indian languages and sectors where current tests fall short. In a recent roadmap blog, Soket AI said, “We’re borrowing DeepSeek’s recipe initially, but we’re modifying it heavily—right down to CUDA kernels and progressive context windows.” Much of this effort is being conducted in partnership with IIT Gandhinagar, focusing on everything from Indic websites to handwritten PDFs, and even transcribing educational videos. In addition, Soket is generating synthetic data through translation and augmentation strategies, especially for domains like science and mathematics. Upperwal said that the team will be able to generate 5-6 trillion unique tokens only on Indic languages, including code. In other domains, Soket expects to build a total corpus of 20 trillion tokens—a foundation large enough to train a world-class multilingual model. “When Common Crawl was done, a lot of Indic websites were not archived… Indic script was ignored.” To correct that, Soket is developing its own data classification systems and language identifiers to preserve and elevate Indic content throughout the training process. Compute: Scaling in the Cloud, Piece by Piece Building such a model from scratch demands enormous computational power. Through a government-backed initiative, Soket has requested up to 2,000 GPUs, a mix of NVIDIA H100s and other GPUs. While the government has not allocated GPUs yet,  the startup expects phase-wise access to start early next week. Although the compute will be cloud-based, Upperwal laughingly said that he hopes to set up local experimentation infrastructure. “We need at least one NVIDIA DGX box that the entire team can share and then start building, optimisation, deploying algorithms, testing, and scale.” The recent release of the Sarvam-M model drew criticism online for its choice of architecture and perceived performance. But Upperwal believes that bashing is just part of the process. “People will bash. But it’s okay as I have seen technologies, which people don’t typically believe in at the very beginning, become successful later,” Upperwal said. He lauded Sarvam’s data curation efforts and sees open-sourcing as critical for the community. “It’s not about the model, it’s not even about the downloads… Look at the work.” At Soket, the team is also bootstrapping synthetic datasets using other models, acknowledging that in low-resource environments, pragmatic reuse of existing models is often necessary. “Say, for example, in our case too, we have been looking at different licensed models to create some synthetic data out of these models. Otherwise, how will you actually do it?” He compares the progress of Indic AI to voice AI in India, which only took off a year after early efforts began. As more developers understand these models, adoption will follow. Until then, he encourages treating these projects as research accelerators rather than commercial products. “We want to utilise that model in terms of generating any data or doing translation. That model will not get utilised [fully now],” he said, implying the real value will emerge later. What’s the Moat of IndiaAI Mission? Upperwal said even the best global models, including GPT-4, still falter when it comes to authentic Hindi. Soket has seen hallucinations and incorrect grammar in Hindi even from state-of-the-art APIs. He added that the company wants to fix grammatical and pronunciation mistakes that often appear while conversing, which even the GPT4o is unable to catch. This, he argued, is the gap Soket aims to fill with cultural authenticity and dialect nuances. Even with Hindi, there are different dialects, and Soket AI wants to incorporate these into the models. “If you want to look at these vernacular-related applications, I think we have to emphasise them,” he noted, adding that Indian AI startups could solve these problems better.","excerpt":"Abhishek Upperwal said that the team can scale the small model to 120 billion within 10 months.","categories":["AI Startups"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-06-12T12:54:33","publication_year":"2025","word_count":1118,"keywords":["CUDA","Go","API","AI","ETL","Git","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","CUDA","R","Go","CUDA","Git","API","ETL"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/soket-ais-plan-to-build-7-bn-open-source-indic-llm-within-6-months\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10076456,"title":"The Brilliant Language Of Lanes","content":"On Tesla’s AI day, the Autopilot team revealed the improvements and massive upgrades in their software. Overall, the Full Self Driving (FSD) has released 35 software updates to date. Ashok Elluswamy, the Autopilot Director, announced that around 160,000 customers globally have been running the beta software of the autopilot and the self-driving system. This is a leap from 2,000 customers last year. The Autopilot team explained how the FSD system is trained and operates—starting from neural networks to training data, and planning, alongside training infrastructure, AI compiler and inference stages, and more. Occupancy Network The Occupancy Network is a multi-camera-based neural network that predicts the surrounding environment of the car using inferred images. The prediction process takes place within the system of the vehicle and is not reliant on the server—therefore, it is also able to predict the future movement and position of the surrounding objects. The Occupancy Network uses all eight cameras on the vehicle, capturing 12-bit images, to detect objects around the car and create a single, unified volumetric occupancy 3D vector space. Since it is based on video inputs, it can also instantaneously—in less than 10 milliseconds—detect changes in the environment like crossing pedestrians, debris, or accelerating cars and adjust the speed and position of the car relative to the uncertainty. Additionally, the team is also developing the Neural Radiance Fields (NeRF) networks by treating the output vectors from the Occupancy Network as inputs for NeRF. Using images from the cameras on the vehicles, NeRF can 3D reconstruct dense meshes using volumetric rendering. The network is trained with a Large Auto-Labelled Dataset without any human interaction. The team built three in-house supercomputers comprising 14,000 GPUs for training and auto-labelling. The training videos are stored in 30 petabytes of storage cache, with half a million videos flowing in and out of the system daily. Language of Lanes In the previous detection method of lanes, Tesla used 2D Pixelwise Instance Segmentation, which could only detect the eagle lane and the adjacent lanes. This only worked efficiently on well-designed and structured roads like the highways. But on roads within the cities, the intersections and lanes are quite complex. Tesla introduced ‘FSD Lanes Neural Network’ which comprises three components—Vision Component, Map Component, and Language Component. The ‘Vision Component’ consists of a set of convolutional layers, attention layers, and other neural network layers that—using the videos from the eight cameras on the vehicles—produce a visual representation. This visual representation is then enhanced with the ‘Map Component’ which has the road-level navigation map which is called the ‘Lane Guidance Module’. The Lane Guidance Module consists of neural network layers that give information about the intersection, number of lanes, and various other features of the road that the cameras on the vehicles might not be able to identify easily in real-time. The first two components produce a 3D Dense World Tensor. This Dense World Tensor is treated as an input image and combined with Tesla’s developed language for encoding lanes and lanes topology called the ‘Language of Lanes’—which is the third component—using LLMs in which the words and tokens are the lane positions of the space. This is a visual representation of the Tesla Autopilot \/ FSD neural networks that drive your car.Every dot represents a mathematical operation running in the car. $TSLA @elonmusk pic.twitter.com\/PLgPNByljU— Whole Mars Catalog (@WholeMarsBlog) October 2, 2022 Training Data Labelling the training data of half a million videos that pass through the supercomputers everyday is a mammoth task. The team built an Auto-Labelling machine for the Lanes Network which, using video footage from the vehicle’s camera, is able to reconstruct 3D vector spaces with the combination of the occupancy network and the newly developed language of lanes. To create one vector mesh from a single trip, the system only takes approximately 30 minutes. Then using ‘Multi-Trip Reconstruction’, footage from different cars is combined and matched. This creates a map in an even lesser time and only requires human intervention in the end to finalise the label of the output. To fix some of the labels where the automated labelling system was facing trouble like parked vehicles, trucks, vehicles on curvy roads, or parking lots, the team corrected 13,900 video labels manually to optimise the whole data engine. Thanks to its accelerated video library built on PyTorch, the team noted a +30% training speed. Using the generated data from the occupancy network, the language of lanes, and NeRF-generated 3D reconstruction models, the team created a Simulation. In this 3D-created world, the team introduced new challenges, environments, and objects to train the system on different changing situations like road designs, biomes, weather conditions, and more. Elon Musk said that the FSD beta would be available worldwide by the end of this year. “But, for a lot of countries, we need regulatory approval. So, we are somewhat gated by the regulatory approval in other countries,” explained Musk, “From a technical standpoint, it will be ready to go to a worldwide beta by the end of this year.”","excerpt":"Tesla’s Full Self-Driving takes huge steps to re-create 3D models of objects using in-vehicle cameras to improve autopilot capabilities","categories":["AI Trends"],"tags":["Elon Musk"],"author_name":"Mohit Pandey","publish_date":"2022-10-06T13:00:00","publication_year":"2022","word_count":835,"keywords":["Go","TPU","programming_languages:R","PyTorch","neural network","AI","programming_languages:Go","RAG","Elon Musk","ai_frameworks:PyTorch","R"],"extracted_tech_keywords":["AI","neural network","PyTorch","RAG","TPU","R","Go","ai_frameworks:PyTorch","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-brilliant-language-of-lanes\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10059721,"title":"MongoDB rolls out free certification for System Integrators","content":"MongoDB Partner team has launched MongoDB SI Architect Certification to help organisations build news business functionality, improve scalability and reduce costs. The self-paced certification is aimed at System Integrator partners and details the benefits of modernisation with various customers on a cloud journey. The certificate enables architects to have discussions on vertical-based stories, migration tools, best practices, and architecture guidelines. System Integrators will get exposure to the fundamental value of offerings, messaging, objection handling, and more. The certification also empowers SI partners to communicate with customers in plain speak. The free SI Architect certification takes approximately 20 hours and is divided into six key sections, complete with a final certification exam. Program structure Introduction allows partners to access the modernisation webinars and modernisation program offerings.Top use cases focus on how MongoDB is used in business-wide strategic initiatives, like legacy modernisation, cloud data strategy, microservices and vertical based stories.Customer case studies highlight how MongoDB is deployed and leveraged through real-life customer case studies and proof points.University classes allow participants to leverage MongoDB university online as well as on-demand courses relevant to the architects.Competitive edge helps architects understand the true value of MongoDB in comparison to the competition.Final certification culminates with a “Talk to the experts” session and final certification exam where participants take a real-world industry use case or customer project and assess how to migrate to the cloud. SI partners can additionally refer to MongoDB University to get access to self-pace developer training and database training.","excerpt":"The free SI Architect certification takes approximately 20 hours, and is divided into six key sections, complete with a final certification exam.","categories":["AI News"],"tags":["Database","MongoDB","scalability"],"author_name":"SharathKumar Nair","publish_date":"2022-02-03T11:54:00","publication_year":"2022","word_count":246,"keywords":["Go","programming_languages:R","AI","MongoDB","Database","Scala","RAG","microservices","Aim","scalability","GAN","R"],"extracted_tech_keywords":["AI","Aim","RAG","microservices","MongoDB","R","Go","Scala","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mongodb-rolls-out-free-certification-for-system-integrators\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10101064,"title":"Google Search Makes Satya Dance","content":"For over the past two decades, Google Search has been the guiding light to information. And all this while, the service worked well for the company but now the Mountain View-based company may have to pay a price since the antitrust case hit the courts. Microsoft’s CEO was called for an hour-long questioning by the justice department for the ongoing trial against Google. “You get up in the morning, you brush your teeth, and you search on Google,” Satya Nadella testified in a packed courtroom in Washington on Monday in the landmark US antitrust case against Google. “With that level of habit forming, the only way to change is by changing defaults,” he suggested. In February, Nadella had portrayed Bing as a “new day” in search during its rollout, now admitting that his “exuberance” stemmed from the hope of increasing Bing’s modest 3% market share. However, as of August, this modest goal remains unachieved. The Redmond giant saw an uptick of roughly 16% on page visits to Bing since the launch of its GPT-4-powered “new Bing”. In its early days, “co-pilot” did draw in a considerable number of new users, as Microsoft itself confirmed that Bing had surpassed 100 million active users for the first time in February 2023. Despite Microsoft’s $10 billion agreement with OpenAI, aimed at challenging Google’s dominance in search, the latter continues to reign supreme in the search space as of August. Keeping Alternatives Afloat While Nadella was on the witness stand, judge Amit Mehta seemed intent on finding out more “about whether a startup could use innovation in artificial intelligence to wrest market share from Google”, the WSJ reported. Star witness Nadella claimed that Microsoft has invested more than Google has in search and that in some ways, Microsoft’s investments have been one of the only things keeping some search alternatives afloat. Contrarily, in March 2023 as the AI war was heating up, without disclosing the names, the Windows maker had warned two Bing-powered search engines threatening to prohibit their access to Microsoft’s search data if they use it to power their AI tools. Before a search engine can hope to make a run against Google, it has to crawl. Smaller privacy-centric search engines like DuckDuckGo, Neeva and Brave need an index of the web. Many sites don’t welcome any web crawler that isn’t Google or Bing hence they are dependent primarily on the tech giants. “Quite frankly, the investments Microsoft has made in search have even kept all the other search players who contend for it, like a DuckDuckGo, going because they use our search index,” Nadella said. While Nadella stands corrected, he missed out an important piece of detail. Last year DDG got into a tracking controversy leading the company to amend terms with Microsoft, its search syndication partner. The Microsoft chief also contended that due to Google’s grip on mobile providers and browsers’ default search placements, the notion of users having real choices in selecting a search engine is “bogus”. Agreeably, not just startups, but Microsoft has also struggled to maintain healthy relations with the iPhone maker, Apple. In an attempt to replace Google Search (the default search engine for all Apple devices), back in 2020, the Nadella-run company tried to sell its Bing search engine to the Cupertino-giant. AI Panic On the witness box, Nadella also said there might be limits to how much new AI applications can reshape the market. This is true since Microsoft has tried every AI tactic to overcome Google’s dominance —yet remains way behind the Sundar Pichai-led firm. Microsoft’s AI investments, thanks to OpenAI’s aid, did cause Google to panic and announce a code red. Reportedly, Alphabet even forced a collaboration between Google’s Brain AI group and DeepMind and reorganised its Assistant team in hopes of focusing more on its AI chatbot Bard. Still, the homegrown Bard has not been able to match the popularity of ChatGPT. In the recent past, Google has devised radical changes to its Search by including AI features and revamping its privacy policy, too. Despite being initially spurred by Microsoft’s increasing AI strength, Google has remained steadfast in its position as the search giant, unaffected by AI innovations. Hence, Nadella’s latest statements hold importance on ways the future of search is going to shape and whether Google will continue to dominate the internet.","excerpt":"Microsoft chief, who once said their OpenAI-powered Bing Search would “make Google dance”, was called for questioning in the ongoing antitrust trial against Google","categories":["AI Trends"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-10-04T13:00:00","publication_year":"2023","word_count":722,"keywords":["Go","ChatGPT","artificial intelligence","OpenAI","AI","GPT","Ray","Aim","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","ChatGPT","OpenAI","Aim","Ray","R","Go","Rust","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/chatgpt-cant-beat-google-search-even-satya-nadella-agrees\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":56549,"title":"Deep Dive: How Consumer Analytics Helped DealShare Thrive In The Cut-Throat eCom Market","content":"In a country like India, the e-commerce industry has been on the rise expected to become the second-largest e-commerce market in the world by 2034. The e-commerce market is expected to reach 13,97,800 crore by 2027. The small e-commerce startups and companies, amidst the e-commerce giants, have had to find the critical marketing gap to come up big and impact the market. These companies have opened up vast markets for micro, small and medium-sized enterprises. Among the emerging names in the e-commerce sector is DealShare, which started in 2018 in Jaipur. Founded by the group of five young entrepreneurs Vineet Rao, Sourjyendu Medda, Sankar Bora, Rishav Dev and Rajat Shikhar, DealShare is a fast-growing online buying platform for multi-category consumer products. DealShare focuses on the new ‘WhatsApp first’ approach in India. DealShare, in addition to making it easy for customers to buy products, also makes it possible for them to share deals with their friends. It has introduced a new retail model in which customers get cheaper rates and discounts depending upon the number of people purchasing the products. They offer products ranging from fruits and vegetables to home decor. Looking at DealShare’s fast-growing trajectory, Analytics India Magazine got in touch with DealShare’s Vineet Rao, Co-Founder, CEO & CTO, and Rajat Shikhar, Co-founder & CPO for our column, Deep Dive. How Far Has DealShare Come In Current Industry Scenario Just over a year old, DealShare has grown to over 25 cities in an industry which has gone through its own cycle of investments and growth with few assets created for unit economics. The company believes that the logistics costs on poor existing infrastructure and high user retention discourage the sustenance of the long term projects. DealShare is looking at these problems with the e-commerce industry, including the critical issue of communicating the solutions which have been imported from the developed countries. The Tech And The Team At Work To generate on-demand responses and functionality for its customers, DealShare uses Java, PHP, AngularJS to build their websites and the popular Java & ReactNative for their applications. Currently, in two locations, DealShare is always on the lookout for personnel with sound technical skills and the right temperament to fit their fast-paced environment. The balance is an essential aspect to their hiring because they also actively recruit\/look for well-experienced candidates in data science to deal with their e-commerce solutions with general business awareness. With a 25 members engineering team and nearly 25,000 transactions per day, they will soon have a training model and are already setting up the data infrastructure required to build applications around machine learning and data. DealShare’s Products And Services DealShare’s platform focuses on making it easily accessible for product discovery of catalogue, attractive pricing and offers for tier 2 or 3 audiences. The app used by these audiences has support everyday languages which makes it easy for them to interact with the platform. DealShare employs gamification and games to keep the audience engaged, which makes it easy for the people who are new to shopping. DealShare works with a lot of WhatsApp automation and referral mechanism to acquire users, enabling social sharing among the audience. According to the company, it is striving to provide better unit economics by driving operational efficiency and supporting innovation on logistics, especially on the last mile. Customer Feedback And Competitors To achieve customer satisfaction, DealShare has focused on metrics and data on an hourly and daily basis. They have actively placed means to collect, analyse and disseminate customer feedback effectively across the organisation. As far as competitors go, DealShare doesn’t have any direct competitors in the online sector. Their focus on supply and logistics perspective in the local ecosystem is hard to match with, and they are laced with the right technology which gives them stable functioning across high geographical level. Future Plans As for the future, DealShare plans to reach more customers and more cities. They intend to implement better social and sharing mechanisms. DealShare will be coming up with more refined and better catalogues for consumers in order to gain more trust and get them to use their services more and more.","excerpt":"In a country like India, the e-commerce industry has been on the rise expected to become the second-largest e-commerce market in the world by 2034. The e-commerce market is expected to reach 13,97,800 crore by 2027. The small e-commerce startups and companies, amidst the e-commerce giants, have had to find the critical marketing gap to […]","categories":["AI Features"],"tags":[],"author_name":"Sameer Balaganur","publish_date":"2020-02-12T18:18:14","publication_year":"2020","word_count":690,"keywords":["data science","Go","machine learning","AI","RAG","analytics","Rust","GAN","R","Java"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","RAG","R","Go","Rust","Java","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deep-dive-how-consumer-analytics-helped-dealshare-thrive-in-the-cut-throat-ecom-market\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10124932,"title":"Hugging Face Launches Open LLM Leaderboard, Chinese Models Dominate","content":"Hugging Face, the AI community platform, has unveiled a brand new open large language model (LLM) leaderboard, with Chinese models taking the top spots, according to an announcement today from co-founder and CEO Clem Delangue. The leaderboard ranks open-source LLMs based on extensive new evaluations, including the MMLU-pro benchmark, which tests models on high school and college-level problems. Hugging Face utilised 300 NVIDIA H100 GPUs to re-evaluate all major open LLMs for the updated rankings. Alibaba’s Qwen-72B model emerged as the top performer overall, outpacing other open-source models, highlighting the rapid progress of Chinese AI companies in the LLM space. The dominance of Chinese models on the leaderboard underscores the increasingly competitive and global nature of the open-source AI ecosystem. Delangue noted that previous LLM benchmarks have become “too easy” for the latest models, comparing it to “grading high school students on middle school problems.” This suggests the need for more rigorous and challenging evaluations as open-source language models grow more sophisticated. However, Delangue also cautioned that some AI developers may be overly focused on optimising for specific benchmarks at the expense of well-rounded model performance. The leaderboard results also indicated that simply increasing model size does not always translate to superior performance. The launch of Hugging Face’s new open LLM leaderboard marks an important step in the field’s maturation in terms of transparent and comprehensive evaluation. With Chinese models leading the pack, the leaderboard will likely spur further innovation and investment in open-source AI technologies worldwide. Hugging Face, founded in 2016, has become a central hub for open-source machine learning, hosting over 250,000 models and datasets used by a community of 200,000 developers.","excerpt":"CEO Clem Delangue noted that previous LLM benchmarks have become “too easy” for latest models, comparing it to “grading high school students on middle school problems.”","categories":["AI News"],"tags":["HuggingFace"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-26T22:04:41","publication_year":"2024","word_count":274,"keywords":["Go","Hugging Face","API","machine learning","NVIDIA H100","AI","innovation","ML","ai_frameworks:Hugging Face","HuggingFace","R"],"extracted_tech_keywords":["AI","machine learning","ML","Hugging Face","R","Go","API","innovation","NVIDIA H100","ai_frameworks:Hugging Face"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hugging-face-launches-open-llm-leaderboard-chinese-models-dominate\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60788,"title":"Your &#8216;Aha!&#8217; Moment As A Data Scientist? Hear From Them Directly","content":"There is no one way to describe an ‘aha!’ moment that all data scientists would agree on. But for most, it captures a series of exhilarating moments that end in a breakthrough discovery. This could be directly related to a problem they were painstakingly seeking to solve, while for others, it could be anchored around a key realization from which their professional improvement and growth emerged. Regardless of how they defined it, the data scientists we spoke to instantly reminiscenced their moments from a time they were still trying to establish themselves in this field. Here are some interesting anecdotes from them:- Solving Long-Contested Problems The draw of solving real-world problems, and the thrill of reaching that critical inflection point by dint of sheer hard work is what keeps data scientists motivated. It is a valuable preoccupation to have, given that their profession demands that they continue to pursue problems and prepare for a lifetime of constant grind. Recalls Tribhuvan Tiwari, Lead Consultant – Data Science at Stratbeans, “One of my projects for speech-to-text conversion had hit a roadblock, and I struggled with that problem for days.” What happened was that when Tiwari increased the time delay in the Recognizer, a lot of ambient noise got captured, which resulted in loss of speech-capturing. When he tried to decrease the time lag of the Recognizer, the system interrupted to take input from the user even at the slightest of pauses. “To overcome this problem, I deployed two Recognizers simultaneously and using Signal Processing Technique in Python, I identified the threshold pitch for a normal human voice and the ambient noise coming with it, which at a later stage, was filtered to extract only human voice for preprocessing,” says Tiwari. It took him four whole days to arrive at that conclusion, but when it did, he describes it as a vital ‘aha!’ moment for him. One that taught him that some moments sets the stage for bigger accomplishments, and that, perseverance always precedes success. ALSO READ: It Takes Years To Become A Data Scientist, Says This Chief Data Scientist Pivotal Point Of Realization Many companies identify their ‘aha!’ moments as those that establish the strength of their relationship with users. These moments are anchored around developments where business’ products or services are finally able to resonate with customers. And like many solutions, these were less about advanced technologies and more about simple and effective messaging, a point that was made clear to Prithviraj Dutta, a senior data scientist at a digital technology company, To The New. “The fact that machine learning is a probabilistic approach to problem-solving makes it more precarious in terms of arriving at an optimal solution,” he says. “Many times, data scientists approach a problem with multiple cutting edge algorithms, hoping to improve the result. However, with time and experience, they realize the importance of keeping things simple, and that is what happened to us,” he adds. In some instances, these may not even revolve around one moment, but culminate into a series of actions that is aligned more closely to the business’ interest. And these are often very straightforward, albeit challenging, paths. Adds Abhishek Goel, Head of Data and Analytics at TTN, “Sometimes, the simplest approach yields the best results.” ALSO READ: The Most Important Lessons Learned As Data Scientists Reaching An Important Milestone Impactful moments that create a powerful impression that goes on to become a driving force for future accomplishments are also expressed as ‘aha!’ moments by some. These events become effective focal points for these data scientists to direct their efforts towards. Data scientist Usha Rengaraju – who is incidentally India’s first female Kaggle Grandmaster – shares her thoughts on this: “My ‘aha!’ moment was in organizing NeuroAI, which is India’s first ever research symposium in the interface of Neuroscience and Data Science. The entire symposium was organized in two weeks and India’s best neuroscientists spoke at the event.” She talks about another event that has shaped her career in data science, “I have always been a great fan of NPTEL courses by IITs and I never thought someday I would get to do a video course similar to that. It is a great privilege and honor to prepare the course curriculum for BITs Pilani’s online masters program and do a recording at their studio. The course is consumed by 20,000 students worldwide.”","excerpt":"There is no one way to describe an ‘aha!’ moment that all data scientists would agree on. But for most, it captures a series of exhilarating moments that end in a breakthrough discovery. This could be directly related to a problem they were painstakingly seeking to solve, while for others, it could be anchored around […]","categories":["AI Features"],"tags":["become a data scientist","best online data science masters","Data Science","Data Scientist"],"author_name":"Anu Thomas","publish_date":"2020-04-02T19:00:00","publication_year":"2020","word_count":728,"keywords":["best online data science masters","data science","Go","machine learning","AI","become a data scientist","Git","Python","ViT","analytics","GAN","Data Science","Data Scientist","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Python","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/your-aha-moment-as-a-data-scientist-hear-from-them-directly\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10092348,"title":"Council Post: Exploring the Pros and Cons of Generative AI in Speech, Video, 3D and Beyond","content":"It is safe to say that Generative AI is the new Pandora’s box. There is no end to unleashing this box. The trend of using generative AI is creeping into every occupation. From text to speech to video to code. We have moved on from the question of whether it will replace jobs and dwell on a new approach on how to use it skillfully and use it to our advantage. When does the relationship between humans and machines change from its current state into one that is so different that we can no longer regard one as being superior to the other in terms of creativity? This is a revolutionary question that the concept of generative artificial intelligence (GAI) raises. The development of generative AI is primarily driven by three developments: better models, better and more data, and increased processing power. Machine learning models have become more complex in recent years. Computers can now understand intricate patterns in data that were previously challenging for them to find thanks to deep learning. This has had a significant impact on generative AI. Our previous articles focused on the pros and cons of text, code and image. This article will dwell further on to the other industries stated below. 1- Speech Although fascinating applications of generative AI have surfaced recently, primarily in speech-to-image creation using well-known models like Stable Diffusion and DALL-E, the technology’s commercial potential has largely gone untapped. And while both image and video have a place in business, speech is emerging as a strength. Pros: Generative AI models can produce more natural and realistic speech than traditional text-to-speech systems. This can improve the quality of automated voice assistants, audiobooks, and other applications that rely on synthesized speech. It can be used to create speech for people who have difficulty communicating verbally, such as those with speech disorders or hearing impairments. This can help improve accessibility for these individuals and make it easier for them to communicate with others. For faster content generation it can make speech quickly and efficiently, making it useful for applications such as automated customer service, where speed and efficiency are important. Cons: According to Mehrabian’s Rule, human speech may be divided into three components: words, tone of voice, and facial expression. Machine comprehension is text-based, and only recent advances in (NLP) have made it possible to train AI models on elements like sentiment, emotions, timbre, and other significant but not necessarily spoken components of language.While the analysis and AI synthesis processes can take some time, real-time speech-to-speech communication is often where it counts. Voice conversion must occur instantly when speaking is being done and translated correctly. Speech-to-speech technology must accommodate a wide range of accents, languages, and dialects and be accessible to everyone in order to realise its full potential.All users will need to support this AI infrastructure with thousands of different architectures for a particular solution because emerging technology solutions are not universally applicable. Additionally, users must plan for consistent model testing. 2- Video Machine learning algorithms called generative video models create fresh video data based on patterns and relationships discovered in training datasets. These models enable the creation of synthetic video data that closely resembles the original video data by learning the fundamental structure of the video data. There are numerous forms of generative video models, including GANs, VAEs, CGANs, and others. Each type adopts a different training strategy based on its particular infrastructure. Pros: Efficiency: To create new videos fast and effectively in real time, generative video models can be trained on enormous databases of videos and images. This enables the quick and inexpensive production of significant amounts of new video content. Customization: Generative video models can create video content that is tailored to a number of requirements, including style, genre, and tone, with the appropriate modifications. This makes it possible to create video material more freely and adaptably. Diversity: Generative video models may create a variety of video content, including films made from text descriptions as well as creative scenes and characters. New avenues are now available for the creation and distribution of video content. Cons: Generative AI can produce unexpected results that may not be in line with the desired outcome. This lack of control can be frustrating and time-consuming to manage. Producing repetitive content or something that lacks diversity, as it can only generate content based on the data it has been trained on. The content produced can get very mainstream for the users. It can perpetuate biases present in the training data, resulting in biased video content. In the age of deep fakes it can create videos that depict people or events that are not real, raising ethical concerns about the authenticity of the video content. 3- 3D According to recent data, the global market for generative design technology is anticipated to increase at a compound yearly growth rate of 17.4% to reach $46.1 billion by 2025. Similar to this, it is anticipated that the global market for creative AI will expand at a rate of 29.5% annually and reach $3.3 billion by 2025. Pros: By automating numerous steps in the 3D modelling process, generative AI enables designers to produce more intricate and detailed models in less time. As a result, designers can produce more realistic and intricate 3D models, giving users more immersive experiences. can assist designers in exploring fresh design ideas and developing modifications of current models, resulting in more imaginative and cutting-edge designs. Generative AI can lower the cost of creating high-quality 3D models by automating several processes involved in 3D modelling. Cons: The high computational resource requirements of generative AI approaches make them unfit for various applications.Models can occasionally create unexpected or challenging results, giving designers little control over the output and forcing them to manually alter or refine it. Even though generative AI models often claim to be accurate, this is not always the case, especially when working with large or highly detailed models. Some designers may find it challenging to embrace this strategy because it requires some level of competence in both domains to use generative AI in 3D modelling. Generative AI is booming and we should not be shocked. Many technologists view AI as the next frontier, thus it is important to follow its development. The potential applications of AI are limitless, and in the years to come, we might witness the emergence of brand-new industries. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"Generative AI is booming and we should not be shocked. Many technologists view AI as the next frontier, thus it is important to follow its development. The potential applications of AI are limitless, and in the years to come, we might witness the emergence of brand-new industries.","categories":["AI Features"],"tags":[],"author_name":"Anirban Nandi","publish_date":"2023-04-26T14:00:00","publication_year":"2023","word_count":1109,"keywords":["data science","machine learning","artificial intelligence","TPU","AI","NLP","Aim","deep learning","analytics","generative AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","NLP","data science","analytics","generative AI","Aim","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/exploring-the-pros-and-cons-of-generative-ai-in-speech-video-3d-and-beyond\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":37415,"title":"How Ubisoft Is Mainstreaming Machine Learning Into Game Development","content":"Without a doubt, Game Development is one of the toughest jobs, and for several years, gaming companies have been trying to leverage the superpowers of machine learning (ML). However, game developers are wary of ML and its limited use in games. The question that arises here is AI is already doing a great job in the field of game development, then why should one add machine learning in game development. Over the years, ML has helped many industries reach the height of success, and it has the ability to make a huge impact on the way video games are developed. ML would not only help in developing games but would also help in transforming the experience of gaming — whether it’s about characters, challenges, and content. Therefore, in order to stay relevant to the current trends in the market, game companies across the world are working day in and out to reap benefits ML. Another big advantage of baking ML in game development is that unlike traditional video games where the characters are scripted, ML-based games will have the ability to learn for the player and act accordingly to the environment. So basically, ML-based games learn to react and respond more dynamically and more realistically to the player. The Challenge Ahead In order to make machine learning go-to tech video games, developers need to have a learning agent that can learn from players. However, this is not something really easy — there lie challenges. The learning agent might have to learn from a human in order to act and respond more accurately, and to teach an ML system, one has to be really good at that video game. That is not all, human-in-the-loop will also have to teach the ML system\/learning agent other skills to make sure that it acts and responds not only to one kind of player but also to another type of players. Ubisoft Use Cases While the rest of the world is thinking and planning to implement machine learning in game development, Montreal-based video game company Ubisoft has already played its cards to make use of machine learning. Action-adventure video game published by Ubisoft in 2017, Assassin’s Creed Origins was inspired by the landscape and history of the Ptolemaic period of Ancient Egypt. And in order to make the game more realistic, Ubisoft used elements of history to build a world for its players. Furthermore, Ubisoft used machine learning model in the game. However, there were challenges — to train the machine learning model that was used in the game to recognize hieroglyphics. Also, developing machine learning models was a challenge and was a very time-consuming process. And the Google Cloud Platform and Cloud AutoML was the answer to all their problems. Also, bugs and complexities can cause a lot of issues when creating a high-end video game, and in order to cope with that, Ubisoft has built an AI assistance for its developers that uses machine learning to discover bugs before they make it into the final game code. How does it work? Dubbed as Commit Assistance, the system has a different signature for a different type of codes. Every time code is presented to the AI assistance, the system analyses the code to see if any of the signature matches. As soon a match is found, the system notifies that it’s a bug. As of now, that system is capable of spotting 6 out of 10 coding errors. Also, it raises a false alarm 30 per cent of the time, however, as it is mentioned before that it uses machine learning, the system over time, will be able to reduce that if more and more code is fed through the algorithm. Moreover, this Assistance has gained so much popularity that even Mozilla today uses this machine learning platform to spot bugs on Firefox. Outlook There is no doubt that machine learning has immense potential to transform any industry for the good. Challenges are always there in every industry and game development is no exception, however, talking about game development, Ubisoft’s significant steps seem to be doing the job already. And with more game companies getting involved in this venture of using ML in game dev, there is no doubt that the video game industry will soon experience a tremendous transformation.","excerpt":"Without a doubt, Game Development is one of the toughest jobs, and for several years, gaming companies have been trying to leverage the superpowers of machine learning (ML). However, game developers are wary of ML and its limited use in games. The question that arises here is AI is already doing a great job in […]","categories":["AI Features"],"tags":["game development","Machine Learning","ML","video games"],"author_name":"Harshajit Sarmah","publish_date":"2019-04-08T11:51:35","publication_year":"2019","word_count":716,"keywords":["Go","video games","machine learning","programming_languages:R","AI","ML","Machine Learning","programming_languages:Go","RAG","cloud_platforms:Google Cloud","game development","R"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","R","Go","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-ubisoft-is-mainstreaming-machine-learning-into-game-development\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040328,"title":"Hands-On Workshop: Learn About oneAPI AI Analytics Toolkit","content":"Machine learning developers often maintain separate code bases, multiple programming languages, and different tools and workflows for various compute accelerator architecture. So for the same set of ML models, developers often have to manage and write codes compatible with specific CPUs, GPUs, or FPGAs. This is highly cumbersome, inefficient, often leads to hardware lock-in and makes it difficult for developers to reuse code. When Intel® released the oneAPI Gold version in December 2020, with a goal to provide an open, standards-based programming model that enables developers to choose which accelerator to run their software. It addresses one of the biggest challenges developers face today – an ML model developed to run on a CPU has to be reframed\/ rewritten to be executed on GPU or FPGA. This is a revolutionary framework, something which all data scientists can take advantage of. To bring more understanding of oneAPI to the ecosystem, Analytics India Magazine has collaborated with Intel® to organise a hands-on virtual workshop on oneAPI AI Analytics Toolkit on June 17, 2021. In this workshop, you will be able to learn how to maximise the performance of heterogeneous computing using Intel® optimised deep learning frameworks along with optimised machine learning algorithms and libraries that are part of oneAPI toolkits. Register For The Virtual Workshop Here. What You Will Learn: Introduction to oneAPI AI Analytics Toolkit & its components.How to maximise deep learning performance using Intel® optimised frameworks.How to maximise machine learning performance using oneAPI Data Analytics Library Intel® Distribution for Python.Hands-on exercises using AI analytics toolkit and its components. Who Should Attend? AI & ML developersData scientistsAI enthusiastsAI researchersGPU & HPC programmers Register For The Virtual Workshop Here. #oneAPI Workshop ContestAnalytics India Magazine is also hosting an exciting contest, exclusively for the participants, as a part of the workshop. Up to 5 winners will each receive an Amazon Echo Show 5.To know more about the contest rules & regulations, click here. Session Speakers: Kavita Aroor — Kavita Aroor, is a Developer Marketing Manager- APJ at Intel, focuses on developing the software solutions market across the Asia Pacific. She drives the developer outreach, engagements and programs in the areas of GPU, HPC and AI. With 17 years of experience in brand and field marketing, Kavita has driven numerous customer and user engagement across the Asia Pacific. Lakshmi Narasimhan — Lakshmi Narasimhan is a Technical Consultant Lead at Intel, with 19+ years of multi-faceted management and customer enabling\/consulting experience in the semiconductor industry. In Intel, he leads the consulting team for the developer products support and consulting division within the Intel Software group. Narasimhan has also led several organisations across Client, IoT and Software. Jing Xu — Jing Xu is a Senior Technical Consulting Engineer working as an AI specialist in a developer products support and consulting organisation within the Intel Software group. He has enabled global developers, enterprise users, engineers and researchers to use Intel software tools and high-performance libraries. His research areas include machine learning, deep learning, performance optimisation and data analysis R&D. Aditya Sirvaiya — Aditya Sirvaiya is an AI Technical Consulting Engineer in Developer Products support & consulting organisation within Intel Software group. Aditya holds a bachelor’s degree in Engineering Physics from the Indian Institute of Technology, Delhi and a Master’s degree in Computer Science with AI specialisation from the Indian Institute of Technology, Mumbai. Details Of The Workshop: Date: 17th June 2021 Time: 9:30 AM to 12:30 PM (IST) Mode: Online REGISTER NOW!","excerpt":"Get a hands-on understanding of using the Intel oneAPI AI Analytics toolkit to maximise the performance of heterogeneous computing with this free workshop.","categories":["Deep Tech"],"tags":["Intel","oneAPI"],"author_name":"Sejuti Das","publish_date":"2021-05-18T11:18:05","publication_year":"2021","word_count":575,"keywords":["Go","API","machine learning","AI","ML","Python","deep learning","analytics","GAN","R","oneAPI","Intel"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","analytics","Python","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-workshop-learn-about-oneapi-ai-analytics-toolkit\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10008641,"title":"D-Wave Announces General Availability of First Quantum Computer With 5000 Qubits Connectivity","content":"Recently, D-Wave Systems announced the general availability of its next-generation quantum computing platform through Leap quantum cloud service. The platform incorporates new hardware, software, and tools to enable and accelerate the delivery of in-production quantum computing applications. Leap quantum cloud service includes the Advantage quantum system, with more than 5000 qubits and 15-way qubit connectivity, in addition to an expanded hybrid solver service that can run problems with up to one million variables. According to reports, the combination of the computing power of Advantage and the scale to address real-world problems with the hybrid solver service in Leap enables businesses to run performant, real-time, hybrid quantum applications for the first time. As part of its commitment to enabling businesses to build in-production quantum applications, the company announced D-Wave Launch, which is a jump-start program for businesses who want to get started building hybrid quantum applications today but may need additional support. The company also announced a new hybrid solver, known as the discrete quadratic model (DQM) solver. It provides developers as well as businesses the ability to apply the benefits of hybrid quantum computing to new problem classes. The Advantage quantum computer and the Leap quantum cloud service include- New Topology: In the new Advantage system, each qubit may connect to 15 other qubits. The D-Wave Ocean software development kit (SDK) includes tools for using the new topology.Increased Qubit Count: With more than 5000 qubits, Advantage more than doubles the qubit count of the D-Wave 2000Q system. Greater Performance & Problem Size: With up to one million variables, the hybrid solver service in Leap allows businesses to run large-scale, business-critical problems. Expansion of Hybrid Software & Tools in Leap: Further investments in the hybrid solver service, new solver classes, ease-of-use, automation, and new tools provide an even more powerful hybrid rapid development environment in Python for business-scale problems.Flexible Access: Advantage, the expanded hybrid solver service, and the upcoming DQM solver are available in the Leap quantum cloud service.Ongoing Releases: D-Wave continues to bring innovations to market with additional hybrid solvers, QPUs, and software updates through the cloud. Alan Baratz, CEO, D-Wave stated, “Today’s general availability of Advantage delivers the first quantum system built specifically for business, and marks the expansion into production scale commercial applications and new problem types with our hybrid solver services. In combination with our new jump-start program to get customers started, this launch continues what we’ve known at D-Wave for a long time: it’s not about hype, it’s about scaling, and delivering systems that provide real business value on real business applications.”","excerpt":"Recently, D-Wave Systems announced the general availability of its next-generation quantum computing platform through Leap quantum cloud service. The platform incorporates new hardware, software, and tools to enable and accelerate the delivery of in-production quantum computing applications.   Leap quantum cloud service includes the Advantage quantum system, with more than 5000 qubits and 15-way qubit connectivity, […]","categories":["AI News"],"tags":["D-wave","easy python beginner projects","fun beginner python projects","quantum application development system","quantum cloud software"],"author_name":"Ambika Choudhury","publish_date":"2020-09-30T13:45:15","publication_year":"2020","word_count":426,"keywords":["Go","API","programming_languages:R","AI","D-wave","innovation","quantum cloud software","automation","Python","ViT","fun beginner python projects","quantum application development system","easy python beginner projects","programming_languages:Python","R"],"extracted_tech_keywords":["AI","Python","R","Go","API","ViT","automation","innovation","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/d-wave-announces-general-availability-of-first-quantum-computer-with-5000-qubits-connectivity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26449,"title":"How Cellular Features Improve ML Accuracy In Phenotyping","content":"Research and discoveries in cell biology have come a long way. Improvements in biological equipments, especially in the area of microscopy, have risen to a cutting-edge level. The precision in obtaining images on a microscopic scale is astonishing. These advances have now presented a challenge of obtaining a vast amount of image data in crispy-clear quality. Although machine learning (ML) has resolved this problem with quick efficacy by using automation, it fails to utilise information from microscopic elements such as cells and tissues. ML only considers the properties or features surrounding the data. It does not dig in deep about the cellular features that determine or influence the extrinsic (environmental) factors have on humans. In this article we will discuss a research analysis which shows that combining environmental and genetic features in an ML method improves the accuracy of a process called phenotyping. The accuracy in the recognition was found to be significantly high compared to regular ML-based phenotyping. What Is Phenotyping? Phenotyping is the process of predicting observations of the biochemical and physical characteristics of an organism, determined by the interactions of its genetic makeup and environment. Just like genotyping focuses on gathering information from the genes, phenotyping looks at the environmental factors that affect the genes. Phenotyping has found wide applications in sub-branches of biology such as anatomy, histology among others, and has helped ascertain processes on a molecular scale. In fact, it is the primer for any kind of biomedical research. The Setback In Molecular Data Analysis As mentioned earlier, biological equipments have advanced largely in the field of microscopy. With them, data analysis in biology has also risen in parallel. But, biologists cite that data analysis alone cannot take account of both intrinsic as well as extrinsic factors on the bio-molecular level. They recommend that a proactive approach is needed to accommodate all the In the latest study by a team of biological researchers from Hungarian Academy of Sciences, University of Szeged, Hungary and the University of Helsinki, Finland, they work on improving ML-assisted phenotyping by gathering all the cellular environmental factors that affect its function and then collaborating them in the learning method. Supervised Machine Learning For Phenotyping In the study, the researchers consider phenotyping details of single cells as the core idea. This is integrated with supervised ML. Now the cellular environmental features are extracted and tested to see how ‘cellular neighbourhood’ affects phenotyping. Various ML methods such as Random Forest, Naive Bayes, Sequential minimal optimisation and Multilayer Perceptron are used for evaluation. Researchers then work on two types of datasets for the experiment. The first dataset consists of images from cell culture of breast cancer cells treated with different drugs (MCF-7 dataset), and the second dataset consists of images from tissue sections of cancerous urinary bladder collated by them (UBC dataset). Once these data are compiled, they use a proprietary software for ML. In their words, “For the experiments, we used an image analysis and machine learning software (SCT Analyzer 1.0) developed by Single-Cell Technologies Ltd. (Szeged, Hungary). Versatile image pre-processing (illumination correction, filtering) and cell segmentation methods (including SLIC segmentation) are implemented in this software. An interactive interface helps the user to annotate segmented regions into an arbitrary number of phenotype classes. Numerous machine learning methods are available even for single-cell-level prediction. Last but not least, an active learning interface is provided to maximise user efficiency.” After this, they segment the images using another software called CellProfiler to segregate cell features such as cell nuclei, cytoplasm, among other components vital in cellular neighbourhood consideration. Next, the features are extracted based on texture, intensity, shape etc. in the components. They use K-nearest neighbours and a distance-based approach to calculate neighbourhood features (for this purpose, they consider Euclidean distances in the components). Once feature extraction is complete, it is ready for ML classification through the SCT Analyser system with respect to single cells. For this, the authors segregate the image dataset along nine phenotypic classes for MCF-7 dataset and eight classes for urinary bladder tissue sections. It is now subjected to different ML algorithm, which was mentioned earlier, to test the prediction performance along both with (neighbourhood distances) and without (local features) cellular neighbourhood factors. Section a – The effect of cellular environment factors, Section b – Confusion matrices of best ML performances in MCF-7 dataset (Image credits : Timea Toth et. al) In the MCF-7 dataset, the accuracy is significantly improved in ML with an increase of 8 percent from previous studies. While the UBC dataset showed a rise in accuracy by an astonishing 19 percent. Comments: The positive results signify that as more cellular factors are considered, the recognition and prediction gets more accurate. Thus, the problem of accuracy can be mitigated in cells. The authors even contend that much more could be achieved through deep learning methods and their study will definitely act as a guiding beacon for future research.","excerpt":"Research and discoveries in cell biology have come a long way. Improvements in biological equipments, especially in the area of microscopy, have risen to a cutting-edge level. The precision in obtaining images on a microscopic scale is astonishing. These advances have now presented a challenge of obtaining a vast amount of image data in crispy-clear […]","categories":["AI Features"],"tags":["biology","Machine Learning Algorithms"],"author_name":"Abhishek Sharma","publish_date":"2018-07-16T05:44:17","publication_year":"2018","word_count":819,"keywords":["Machine Learning Algorithms","Go","biology","machine learning","AI","ML","automation","deep learning","ViT","GAN","R","active learning"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","R","Go","GAN","ViT","active learning","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-cellular-features-improve-ml-accuracy-in-phenotyping\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10078754,"title":"Wesco Acquires Rahi Systems, a Leading Provider of Hyperscale Data Center Solutions","content":"Wesco International, which holds leadership in distribution and supply chain solutions, acquired Rahi for $217 million, approximately 7.5x Rahi’s projected trailing twelve months’ adjusted earnings before interest, taxes, depreciation and amortisation (EBITDA) as of September 30. The acquisition will be aimed at combining Rahi with its Communication and Security Solutions (CSS) strategic business unit. Speaking about the acquisition, John Engel, chairman, president and CEO of Wesco, said, “Rahi’s extensive services portfolio strengthens our leading data centre solution offerings for our global customers,” adding that the takeover highlights their continued investment in the high growth data centre segment and further expands cross-sell opportunities across the company. The takeover of Rahi is their first purchase since 2020 when Wesco announced a merger with Anixter, also a B2B distribution and supply chain solutions company. Bill Geary, executive vice president and general manager, Wesco Communications and Security Solutions, says, “With over 900 employees in 25 countries and trailing 12-month sales of approximately $400 million, Rahi provides complementary global coverage and enhances our full suite of data centre solutions for contractors, integrators and end-user customers.This significant investment fuels growth for our technology focused customers, who can now leverage the combined global footprint, infrastructure and IT expertise of Wesco.”","excerpt":"Wesco International, which holds leadership in distribution and supply chain solutions, acquired Rahi for $217 million, approximately 7.5x Rahi’s projected trailing twelve months’ adjusted earnings before interest, taxes, depreciation and amortisation (EBITDA) as of September 30.  The acquisition will be aimed at combining Rahi with its Communication and Security Solutions (CSS) strategic business unit.  Speaking […]","categories":["AI News"],"tags":["Mergers and Acquisitions"],"author_name":"Ayush Jain","publish_date":"2022-11-04T18:32:01","publication_year":"2022","word_count":203,"keywords":["programming_languages:R","AI","RAG","Aim","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wesco-acquires-rahi-systems-a-leading-provider-of-hyperscale-data-center-solutions\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10074375,"title":"WhatsApp to Bring In-App Survey Feature Soon","content":"Social media platform WhatsApp is set to bring ‘WhatsApp Surveys’, a new in-app survey chat that allows users to give feedback on new features and products. As reported by WaBetaInfo on September 3, the feature is under development and will allow users to give feedback upon receiving an invitation. The site claims that when receiving a survey, WhatsApp would maintain clarity about its purpose—with users having an option to decline the invite to provide feedback, essentially making it optional. In addition, users will also be able to block the chat from sending new surveys in the future. According to the post, “It is not clear yet which surveys people will receive from this conversation, but the idea to provide feedback by using an official WhatsApp chat seems very interesting since your responses will help them understand how to improve WhatsApp.” In addition, it emphasised that users must not expect to receive a survey any time soon, as the platform plans to extend an invitation to a handful of users in the future, thereby also making it an occasional activity for the users to take part in a survey. Sensitive information such as two-step verification PIN, credit number and 6-digit code will never be asked of the users. The survey will only be used for feedback purposes and will allow users to opt out at any time by merely blocking the conversation within the Chat info. Meta-owned WhatsApp has an impressive 2+ billion MAUs—nearly 25% of which are Indian users.","excerpt":"The survey will only be used for feedback purposes, allowing users to opt out at any time by blocking the conversation within the Chat info.","categories":["AI News"],"tags":["feedback","whatsapp"],"author_name":"Bhuvana Kamath","publish_date":"2022-09-05T16:43:30","publication_year":"2022","word_count":249,"keywords":["programming_languages:R","AI","Git","whatsapp","feedback","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","R","Git","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/whatsapp-to-bring-in-app-survey-feature-soon\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":19664,"title":"Moving Deep Learning From Test To Production With New Techniques","content":"Interest in deep learning is at an all-time high what with breakthroughs in areas of computer vision, speech recognition, language translation and most importantly pattern recognition in large datasets. However, despite recent advances in mission critical field such as autonomous driving where driverless cars require lightning-fast deep learning inference, usually within tens of milliseconds for each sample (inference is the AI lexicon for applying capability to new data), researchers are still shining a spotlight on challenges such as: Dealing with insufficient labeled training data Learning with less data Dealing with incremental data While the prospect of using deep learning in challenging domains such as medical imaging improves, DL researchers are still grappling with the problems of continuous training in production systems. In this article, we will talk about recent trends and methods shaping research in deep neural networks which will offer a good starting point to readers. Before we dive into the techniques, let’s talk about some of the challenges. First up is the problem of dealing with less data and insufficient labeled training data. Of late, researchers are pretraining state-of-the-art Deep Neural Networks (DNN) on large general-purpose data sets like ImageNet, and then fine-tuning the model on a smaller data set of interest. This helps tackle the problem of insufficient labeled training data for especially for areas such as medical diagnosis where obtaining labeled data is extremely cost and time -intensive. According to Parth Shrivastava, Head of Digital Marketing at Parallel Dots, though AI is gaining traction in the market, AI capabilities are still falling short of fulfilling real-world problems. Citing a use case, he noted in a post: for example, in the case of image classification, most people would think it is a solved problem. We can classify images for example, into cat images or dog images matching human accuracy. But if one feeds the algorithm an image of a dog of rare breed, it will find it difficult to classify it as a dog. Interestingly, the startup is using most of these techniques to improve output. Image: Parallel Dots As Deep Learning marches into relatively tough domains, there is the problem of dealing with incremental learning – adjusting the ratio of old and new data, wherein new labels are added incrementally without disturbing previous training. For example, let’s say one wants to do incremental training of a CNN model as new classes are added to the existing data. The CNN model is initially trained for classifying, 500 classes with 1 million images. But with the availability of new data, there are 50 new classes with 10000 images in addition to the previous of 500 classes. It’s here that Transfer Learning techniques comes into play — models trained on one task capture relations in the data type and are repurposed for different problems in the same domain. Also, do check out the Microsoft Research paper that talks about incremental learning. In this article, we list down emerging techniques that are reengineering AI research: 1.Transfer Learning: According to startup founder Sarthak Jain, in transfer learning, developers can work on a pretrained model, that was initially trained on a large dataset. In other words, the model was trained on a different task, with the same input but different output. He further explains how one should find layers which output reusable features and use the output of that layer as input features to train a small network that requires a smaller number of parameters. It is on the basis of the small network that one can learn the relations for the specific problem after gleaning the patterns in the data from the pretrained model. Transfer Learning helps scale AI solution with lesser quantity of data. Known Use Cases: In fact, Mountain View search giant Google’s “Development and Validation of a Deep Learning Algorithm for Detection of Diabetic Retinopathy” is an example of transfer learning, applied to real-world image dataset. 2.Adversarial Learning: According to influential AI research scientist Ian Goodfellow, until recently, even an input could trick an object recognition model. Today, Goodfellow notes that object recognition algorithms have reached near human accuracy but fail to perform on unnatural inputs. Adversarial Learning sprung from Goodfellow’s research and he is also credited for coming up with the new framework — Generative Adversarial Nets. Inspired by game theory, this framework presents two algorithms, a generator and a discriminator, trying to trick the other during  fool each other while they are training. Here’s the definition by Goodfellow: Adversarial examples are synthetic examples constructed by modifying real examples slightly in order to make a classifier believe they belong to the wrong class with high confidence. Known Use Cases: Interestingly, researchers at University of Michigan and Max Planck Institute have come up with a paper on creating images from text – in other words text to image synthesis. Its real-world application would be a software probably like Adobe Photoshop that allows one to create new objects from text instructions. According to Goodfellow, Adversarial Learning give us some traction on safety in AI and can be deployed to prevent avoid potential security problems. 3.  Multi Modal Multi Task Learning: Did you hear about Google’s neural network that is a multi-tasking pro. The search giant recently came up with a multi-tasking machine learning system called MultiModal that learnt how to detect objects in images, recognize speech and even translate between four pairs of languages besides parsing grammar – all this simultaneously. According to the company blog,  this neural network architecture draws from the success of vision, language and audio networks to simultaneously solve a number of problems across multiple domains, such as image recognition, translation and speech recognition. Known Use Cases: Well, Google’s Multilingual Neural Machine Translation System used in Google Translate, works on this approach an MultiModel and is a first step towards the convergence of vision, audio and language understanding into a single network. 4. Few Shot Learning: Used in image recognition, this technique is used when the number of categories is large and the number of examples per novel category is very limited, e.g. 1, 2, or 3, states the recently released paper on Few-Shot Image Recognition. This paper dealt with a new method wherein a pre-trained neural network can be adapted to new categories by directly predicting the parameters from the activations. Zero training is required in adaptation to novel categories, and fast inference is realized by a single forward pass. Known Use Cases: AI researchers believe this technique is limited in terms of applicability, given the lack of large training data. Since the paper was tested on ImageNet database, these models have good image descriptors.","excerpt":"Interest in deep learning is at an all-time high what with breakthroughs in areas of computer vision, speech recognition, language translation and most importantly pattern recognition in large datasets. However, despite recent advances in mission critical field such as autonomous driving where driverless cars require lightning-fast deep learning inference, usually within tens of milliseconds for […]","categories":["IT Services"],"tags":["Adversarial Learning","Deep Learning Techniques","Transfer Learning"],"author_name":"Richa Bhatia","publish_date":"2017-12-12T09:39:21","publication_year":"2017","word_count":1101,"keywords":["Go","machine learning","TPU","AWS","AI","neural network","image recognition","Adversarial Learning","computer vision","deep learning","Deep Learning Techniques","Transfer Learning","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","image recognition","AWS","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/moving-deep-learning-test-production-new-techniques\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10082070,"title":"SiMa.ai’s Journey to Being 10x Better than Qualcomm and Nvidia","content":"SiMa.ai, the machine learning company, claims to deliver the industry’s first software-centric, purpose-built MLSoC platform. With push-button performance, the company claims to enable effortless ML deployment and scaling at the embedded edge. This is done by allowing customers to address any computer vision problem while achieving 10x better performance at the lowest power. To understand more about the company, Analytics India Magazine reached out to Krishna Rangasayee, founder and CEO, SiMa.ai and Harald Kroeger, president (automotive business) at SiMa.ai. AIM: How did the idea of Sima.AI come about? Krishna: We started the company with an observation that everybody was talking about machine learning at the embedded edge. People spoke of robotics, and automotive elements, etc, but if you take a long, hard look, in the past 10 years, there has not been a whole lot of ML shipped in scale. There is a dissonance between aspiration and reality. So, if you want to start a company and create a purpose-built full stack ML solution, hardware and software, we would like to address computer vision as the first problem. Now, four years later, we have a solution that we’re shipping to customers, delighting them with the possibilities. I think we’re changing the mindset around how to really scale machine learning and scale for the embedded market. AIM: Harald, what was your journey like and how did you decide to join Sima.Ai? Harald: My journey is a real car guy journey. I started off at Mercedes research, where I was later heading development of all electronics, including e-drive and powertrain. I left Daimler after 20 years and joined Bosch, which is a real tech company, more ‘tech’ than Daimler in terms of engineers and innovations. At Bosch, I was a member of the executive board and guided 100,000 people with about 20 billion of business. Last year, in December, I decided I was going to leave Bosch to do something new. I have known Krishna for quite some time, and he asked me if I would be willing to join the board for Sima.ai. I agreed, and started the automotive business. I think the product is perfect for automotive. AIM: How is MLSoC™ Platform different from others? Krishna: A key technical merit that customers care about in computer vision is frames per second per watt. It’s like how much you can compute for a given power at a given time. So that’s a key technical merit people look to, and in that aspect, we are in the minimum 10x ahead of Nvidia or Qualcomm or any other competitors. People are surprised at how 140 people could pull off something that a 16,000-strong company cannot. How are we able to scale 10x? Number one, we do not have any legacy architecture. If you’re a Qualcomm or an Nvidia, you have a lot of legacies that you have to support, and with legacy, there’s baggage. But how far can you push technology? One thing that works for us is that we are completely purpose-built from scratch and it’s a huge advantage for us that we are good at solving this problem. The second is we have taken a very software-centric approach to solving this problem. So from our perspective, machine learning is primarily a software problem. We’re solving machine learning inference and we know everything about the problem ahead of time. We know the network, the weights, the biases, the input data, the memory architecture, and the data flow. So what we do is, compile the problem, and we know exactly what needs to be done at every clock cycle. So, we are more on the orchestration of the data and memory management. And we do that more cleverly than most companies. AIM: Do you see India as an emerging market? Krishna: We do believe India is a very big market and we are in the early stage in terms of engaging the customers. Still, we are engaged with about seven to eight customers in India. There are a lot of companies looking for a standardised edge computing platform that does a lot of different applications. It could be around factory floor automation, or security and monitoring. And so I would say, we are at an early stage, but we also have a lot of interest from customers in drones and UAVs in India as well. Computer Vision is a massive market. And I believe that India’s got a lot of innovation going – we’ll be participating in that as well. AIM: When will India have driverless cars? Krishna: People have had proprietary inside proprietary knowledge, and that’s been a defensible barrier for how companies preserve and retain their competency. Machine learning is breaking things apart and allowing new people to come in. There is now a jump ball where legacy is no longer a strength. So it is now open for  everybody to figure out what needs to be done, and I think India is going to take a very different approach than the classic Western one. I visited a few companies when I was in India. I met the folks at Ola cabs, and saw the entire ecosystem of them trying to launch their own automotive future or e-bikes. There’s a lot of experimentation, and they have the opportunity of taking something brand new and really run with it. The needs of India are different from the needs of the Western world. Harald: A lot of times, we only think about the car being fully automotive – there is no wheel in the car, and it takes you from point A to point B. No doubt, that’s a great vision, a great marketing tactic for companies, however, to see a fully-automatic car juggling Bangalore’s infamous traffic will take some more time. The level of communication while driving in Bangalore is a different level of challenge altogether and AI being able to deliver it is still a distant dream. Nonetheless, if you look at the people getting killed in car accidents due to stupid mistakes, and human errors, that can be easily avoided by the introduction of AI in automative industry. AI can help avoid accidents related to human fatigue, and lack of attention etc. The cars may not get fully automatic in countries like India anytime soon, but for sure, it’ll incorporate some of the good stuff already present on the table.","excerpt":"“Machine learning is breaking things apart and allowing new people to come in. There is now a jump ball where legacy is no longer a strength”– Krishna Rangasayee","categories":["AI Features"],"tags":["Computer Vision","Interviews and Discussions","Machine Learning","ML","sima.ai"],"author_name":"Lokesh Choudhary","publish_date":"2022-12-12T14:00:00","publication_year":"2022","word_count":1060,"keywords":["Go","machine learning","AI","ML","Machine Learning","computer vision","automation","Aim","analytics","Computer Vision","edge computing","R","sima.ai","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","analytics","Aim","edge computing","R","Go","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/sima-ais-journey-to-being-10x-better-than-qualcomm-and-nvidia\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10096380,"title":"6 New Open Source Text-to-Image Models","content":"Text to image models emerged in the mid-2010s due to advancements in deep neural networks. However, much before ChatGPT, the buzz around generative AI grew with text-to-image models OpenAI’s DALL-E, Google Brain’s Imagen, and StabilityAI’s Stable Diffusion. These generative AI models have garnered attention because of resembling real photographs and hand-drawn artwork. Best Open Source Text to Image ModelsDeepFloyd IFStable Diffusion v1-5 OpenJourneyDreamShaperDreamlike PhotorealWaifu Diffusion So, let’s take a look at the top six open-source image generation models that can come to your help. DeepFloyd IF Backed by Stability AI, research group DeepFloyd’s open-source text to image model DeepFloyd IF combines realistic visuals and language comprehension. It consists of a modular design, featuring a fixed text encoder and three interconnected pixel diffusion modules. The initial module generates 64×64 px images based on text prompts, while the subsequent super-resolution modules create images of increasing resolution: 256×256 px and 1024×1024 px. The entire model leverages a frozen text encoder derived from the T5 transformer to extract text embeddings. These embeddings are then utilized in a UNet architecture, which is enhanced with cross-attention and attention pooling. As a result, this model surpasses existing models, achieving an impressive zero-shot FID score of 6.66 on the COCO dataset. Check out their GitHub repository here. Stable Diffusion v1-5 Latent text to image model Stable Diffusion v1-5 merges an autoencoder with a diffusion model to create photo realistic images.  It has been trained on the extensive laion-aesthetics v2 5+ dataset and fine-tuned over 595k steps at a resolution of 512×512 pixels, this model has the remarkable capability of generating highly realistic images based on any given text input. It has flexibility in generating images from a wide range of latent spaces, as opposed to being restricted to a fixed set of text prompts. Its training on a large image dataset enables it to possess a deeper understanding of image characteristics, resulting in more lifelike image generation. Stable Diffusion v1-5 is accessible in both the Diffusers library and the RunwayML GitHub repository. Check it out here. OpenJourney Openjourney is a free, open-source text-to-image model that produces AI art in the style of Midjourney as it is trained on a dataset of over 124k Midjourney v4 images. It’s a fine-tune of Stable Diffusion. Developed by PromptHero, a leading prompt engineering website, Openjourney is the second most downloaded text-to-image model on HuggingFace, following Stable Diffusion. Users prefer Openjourney for its ability to generate impressive images with minimal input and its suitability as a base model for fine-tuning. Click here to access the model. DreamShaper Built on diffusion model architecture, fan favourite Dream Shaper V7 introduces improvements in LoRA support and overall realism. It builds upon the enhancements made in Version 6, which included increased LoRA support, general style improvements, and better generation at a height of 1024 pixels (though caution is advised when using this feature). It produces photorealistic image with a noise offset, and enhances anime-style generation with booru tags. It also improves eye performance at lower resolutions, serving as a “fix” for earlier versions. The impact of Version 3.32’s “clip fix” may differ from Version 3.31, recommending its use for mixing. It also involves inpainting and outpainting. If you want to know more about it, check this out. Dreamlike Photoreal Dreamlike Photoreal 2.0 is a photorealistic model based on Stable Diffusion 1.5. Made by DreamlikeArt, you can enhance the realism of your generated images by incorporating photos into your prompt. For best results, use non-square aspect ratios. For portrait-style photos, a vertical aspect ratio is recommended, while a horizontal aspect ratio is more suitable for landscape photos. This model was trained on images with dimensions of 768×768 pixels, although it can also handle higher resolutions like 768x1024px or 1024x768px effectively. Running on server-grade A100 GPUs, it boasts an average generation speed of 4 seconds, surpassing the performance of 8x RTX 3090 GPUs. With the capability to process up to 30 images simultaneously and generate up to 4 images concurrently, it ensures an efficient workflow. It includes several features like upscaling, natural language editing, facial enhancements, pose, depth, sketch replication, and others. You can access it here. Waifu Diffusion Last but not the least, we have Waifu Diffusion, a fine-tuned version (1.3) of the Stable Diffusion model, derived from Stable Diffusion v1.4. This model specialises in generating realistic anime-style images and has gained recognition for its impressive variety and high quality. The model was trained on dataset of 680k text-image samples obtained from a booru site. Find their GitHub repository here.","excerpt":"Stable Diffusion v1-5, DeepFloyd IF, OpenJourney, Waifu Diffusion, Dreamlike Photoreal are among the top image generation platforms.","categories":["AI Trends"],"tags":["AI Tool","Open Source AI"],"author_name":"Shritama Saha","publish_date":"2023-07-05T17:12:43","publication_year":"2023","word_count":752,"keywords":["Go","ChatGPT","OpenAI","AI","neural network","ML","RAG","Open Source AI","prompt engineering","generative AI","AI Tool","R"],"extracted_tech_keywords":["AI","ML","neural network","generative AI","ChatGPT","OpenAI","RAG","prompt engineering","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-6-open-source-text-to-image-models\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10092914,"title":"5 Best Prompt Engineering Courses in 2024","content":"As generative AI models continue to evolve, a new field area has arisen called prompt engineering which is the skill to add appropriate prompts to guarantee that the machine comprehends human language precisely as intended by the user. This has led to numerous courses, tools, and employment opportunities for prompt engineering skills. So if you are thinking about upskilling, then choose any of the following courses for the best resources that are available. Read more: Killing Prompt Engineering 1. ChatGPT Prompt Engineering for Developers by DeepLearning.AI Andrew Ng’s DeepLearning.AI has collaborated with OpenAI to launch a new course: ‘ChatGPT Prompt Engineering for Developers’. The free-of-cost course will let you learn how to use a large language model (LLM) effectively to build new and powerful applications. Isabella Fulford, a member of the technical staff at OpenAI, and Ng will teach a new course about how LLMs work. The course aims to offer valuable tips for prompt engineering, as well as demonstrate the various ways that LLM APIs can be used in applications for summarisation, inference, text transformation, and expansion. Furthermore, the course will cover two essential principles for crafting successful prompts, provide instruction on how to systematically develop effective prompts, and teach you to create a personalised chatbot. The short 1.5 hours course is beginner-friendly and, is designed to be accessible to novices with a fundamental grasp of Python. However, it is also helpful for experienced machine learning specialists who aspire to explore the forefront of rapid engineering and use LLMs. Check out the course here. 2. Learn Prompting 101 by Towards AI The ‘Learn Prompting 101’ by Towards AI caters to beginners and offers a wide range of topics from fundamental AI concepts to advanced prompt engineering techniques. It’s a free and open-source course, that provides comprehensive guidance without overwhelming technical terms. The course is practical, featuring examples that are easy to comprehend and encourages collaborative learning. It consists of several chapters that include basic applications, intermediate, and advanced applications, reliability, images, prompt injection, tooling, prompt tuning, and miscellaneous topics. The course is highly respected and referenced by reputable organizations such as Wikipedia, O’Reilly, Scale AI, and OpenAI. Enrol for the course here. 3. Prompt Engineering: Getting Future Ready by Udemy Priced at Rs 449, the ‘Prompt Engineering: Getting Future Ready’ is one of the best sellers on Udemy and is designed for beginners and covers a wide range of topics related to prompt engineering. It includes more than 1000 prompts, templates, and resources, and focuses on the primary tools utilised in prompt engineering, such as ChatGPT, Stable Diffusion, and Midjourney. Participants will learn how to effectively use each of these tools, and gain a comprehensive understanding of the differences between text-to-text and image-to-image generation. The course features a comprehensive prompt guide with practical examples and hands-on exercises to help learners create images and text that are almost indistinguishable from real life. The program is suitable for learners with varying levels of experience, including beginners, experienced AI practitioners, and professionals looking to incorporate prompt engineering into their work. Participants will acquire skills in areas such as content creation, SEO techniques, AI-generated art, startup building, email marketing, social media campaigns, and designing colouring books. No prior programming expertise is required, but participants must have a functional computer and an OpenAI account. Click here to learn more about it. Read more: Prompts are Next Big Thing in AI-Generated Art 4. Prompt Engineering for ChatGPT by Coursera The ‘Prompt Engineering for ChatGPT’ course on Coursera aims to train individuals to become proficient in using large language models, such as ChatGPT. However, the effectiveness of these models largely depends on the quality of the prompts provided by the user, and the course will equip students with the necessary skills to create effective prompts. The course can be accessed by individuals with basic computer skills, and it covers a wide range of prompts, ranging from basic to advanced, to enable students to tackle problems in any domain. Upon completion of the course, students will possess the ability to leverage large language models to carry out various tasks in their personal and professional lives. The course is free and comes with a certificate. 5. Prompt Engineering+: Master Speaking To AI for Free by Udemy The ‘Prompt Engineering: Master Speaking To AI’ is a free, short but comprehensive course on Udemy that teaches advanced techniques for prompt engineering. The course covers topics such as the anatomy of an engineered prompt, the prompt mindset, and advanced concepts like one-shot, few-shot, and zero-shot cot. Best practices and workflow optimization techniques are also included. The course is suitable for developers, designers, content creators, writers, bloggers, business owners, marketers, sales professionals, and students studying computer science, data science, or artificial intelligence. Anyone interested in AI and language models can benefit from this course. Apply here.","excerpt":"Find best prompt engineering courses by DeepLearning.AI, Coursera, Udemy are among the primary providers of prompt engineering courses, majority of which are free-of-cost.","categories":["AI Trends"],"tags":["Courses","prompt engineering"],"author_name":"Shritama Saha","publish_date":"2023-05-05T16:58:11","publication_year":"2023","word_count":804,"keywords":["data science","ChatGPT","machine learning","artificial intelligence","OpenAI","AI","RAG","prompt engineering","Aim","generative AI","Courses"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","generative AI","ChatGPT","OpenAI","Aim","RAG","prompt engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-prompt-engineering-courses\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10142741,"title":"AI Without Vibes is Just Code","content":"The race to AGI isn’t just about creating smarter models but also systems that think, adapt, and feel human, especially so in the age of intelligent design. You ask it a question and feel a subtle resonance or a disconnect. These instinctive “vibe checks,” as OpenAI’s Greg Brockman calls them, are no longer just hunches. Wharton professor Ethan Mollick acknowledges they’re becoming official benchmarks, hinting that a more intuitive approach to AI evaluation is taking shape. Vibe checks are a method for identifying and validating qualitative differences. OpenAI recently conducted a viral experiment on X, where users asked ChatGPT to generate responses using all the information it had about them. This was based on ChatGPT’s memory feature. The responses took over the internet, and users enjoyed it. In another post, Mollick compared the different AI models. Are Vibe Evals Good Benchmarks? A recent study by researchers at UC Berkeley identifies and quantifies qualitative differences, or “vibes,” in the outputs of large language models (LLMs). Most traditional methods focus on metrics like accuracy, clarity, and conciseness. These predefined axes fail to capture open-ended, subjective user preferences. As models achieve high baseline performance, users often rely on “vibes” to choose between them. “Just take the landscape today: Claude, GPT, and Gemini are so good that people often go off of ‘vibes’ because (i) all the models are able to do the tasks they want so they are looking for an explanation which is best suited to them or (ii) they are asking really open-ended questions like writing which can’t currently be quantified,” said Lisa Dunlap in an exclusive interview with AIM. She co-authored this paper. This researcher focuses on qualitative differences, emphasising tone and style over traditional measures like accuracy. Dunlap uses LLM judges to identify and quantify “vibes” that set models apart. These vibes are evaluated based on consistency, ability to distinguish models, and alignment with user preferences. For this, they introduce the concept of VibeSystem, a framework for evaluating AI models (e.g., Llama-3-70b, GPT-4). The researcher mentioned the selection of models was primarily based on costs. “When generative models started to become something that I used outside of research just for my day-to-day tasks, it hit me how narrow the current evals are,” she added, emphasising the need for more subjective evals. The vibes of AI models are Complex, just like Humans On the whole, vibes-based computing presents challenges related to subjectivity and scalability. “Vibes” are subjective, making automation or standardisation challenging and resource-intensive, especially for large-scale or real-time applications. “I would say the biggest challenge is making vibes that are well-defined,” said Dunlap. For instance, humour is subjective and varies by person and culture, making it hard to measure traits like “vibes” accurately. Using multiple language models helps cross-check results, but their biases can still differ from what most people think. Dunlap highlighted the growing need for collaboration between computer scientists and experts in psychology and education, who offer valuable insights into human interaction and subjective task evaluation. Interestingly, another insight is how while ‘vibes’ play an important role in user preferences, they are secondary to correctness; a model with an appealing style but incorrect answers is less useful than one that delivers accurate outputs, regardless of its tone. Rather than replacing traditional performance metrics, ‘vibes’ serve as a complement, offering a more comprehensive understanding of model behaviours and their impact on user experience. But, what does the future hold? The future of AI evaluation will likely combine data-driven metrics with human intuition. By focusing on user experience, developers can create methods that assess how well AI models align with human expectations and emotions. Researchers like Dunlap agree that Model evaluations will expand beyond numerical scores like those on MMLU (massive multitask language understanding) to include more subjective traits. The focus will shift from global, standardised evaluations to user-specific evaluations. In a podcast, OpenAI’s Kevin Weil noted that models today are limited by evaluation methods rather than intelligence, with the potential for greater accuracy and broader task capabilities. “I think the space of evaluation has grown a lot in the past few years, and there is a lot of money involved. I think what is more lacking is figuring out what to do with all these benchmarks,” said Dunlap on whether evals are underfunded, highlighting that limited budgets require funding to focus on evaluating models on existing benchmarks instead of creating new ones. Regarding the practical utility of vibe-based benchmarks, Dunlap highlighted that they are most effective for open-ended tasks, such as asking a chatbot like GPT to write a story or leveraging LLMs for customer service. On similar lines, Reka AI researchers also introduced Vibe-Eval, an open benchmark designed to challenge models like GPT-4 and Claude-3 with nuanced, hard prompts to evaluate traits such as humor, tone, and conversational depth. “Hard prompts are hard to make,” according to a Reka AI research paper. They say an ideal benchmark hard prompt should be unsolvable by current frontier multimodal language models, interesting or useful to solve, and be error-free and unambiguous for an evaluator.","excerpt":"As models achieve high baseline performance, users often rely on “vibes” to choose between them.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)"],"author_name":"Aditi Suresh","publish_date":"2024-12-09T14:49:57","publication_year":"2024","word_count":844,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","ML","Scala","RAG","Aim","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Aim","RAG","TPU","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-without-vibes-is-just-code\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":32322,"title":"Indian Traffic Dataset For Autonomous Vehicles","content":"With the number of deaths caused by road-related issues and accidents reaching 400 per day in India, it has become imperative to find solutions to minimise road fatalities. Developed countries have already embarked on a non-human technological journey to alleviate fatal human flaws, however developing economies like India, still have to catch up. For example, with the CityScape dataset, the researchers were able to detect the crucial challenges in Germany. This dataset can also work for other developed countries, but for India, where traffic violations are rampant, these datasets can’t be inculcated to ensure safer road travel. Collecting raw unstructured data in India is very different from any other populous country. It is not just the lane direction and crossroads, but the erratic and inconsistent driving routines of a typical Indian rider that are also to be taken into consideration. This infamous driving scene, which is unique to India, became the subject of a thorough survey for this larger unique dataset. Now, a project funded by Intel in partnership with the Government of Telangana and Karnataka, along with the team from IIIT Hyderabad are also addressing this issue. Conducting Study In Congestions The project, funded by Intel, began in November 2017 in partnership with the government of Telangana and Karnataka. The team at IIIT Hyderabad drove around Hyderabad and Bengaluru, areas known for traffic congestion. The idea here was to collect and label unique variables such as signposts, pedestrians, types of vehicles, streetlights etc. The objective was to ensure road safety by creating a dataset which suits our Indian needs. So, the team created about 10,000 pixel-level annotated images and 50,000 object level annotated images, twice the size of Germany’s Cityscape, which contained 5,000 frames. For fine annotation, images from forward-facing cameras of a stereo pair were taken. These images were then sampled from the video feed with more attention laid onto traffic junctions and other crowded portions in the feed. A total of 34 labels were used for annotation. The labels are well defined with text and example images. To address the synonymous labelling ambiguity for scene diversity, a 4 level label hierarchy was used. Where the ambiguity is deliberately is increased with the level. The setup used by Professor Jawahar and the team at IIIT Hyderabad At the pixel-level, each pixel in the image is associated with an object class such as an auto-rickshaw, a car, a cycle, and so on. Via Insaan IIIT-H “This is the holy grail of data sets and will test the best in class algorithms,” says Dheemanth Nagaraj, an Intel fellow and architect, server CPU development and new products innovations at chipmaker Intel. Input example images baseline trained on the dataset via Intel Challenges Of Unstructured Data Seeing is believing and this is no different in case of autonomous vehicles either. Only here, there will be an action taken almost instantly. So, whatever images the camera captures will be fed into the onboard electronics which runs on well-trained algorithms which dissect the images for colour and curves. Image processing is key functionality of any driverless vehicle. Apart from the driving habits, there are other issues which make the image processing a tricky job. For instance, there are billboards with images high in colour intensity. Imagine having a billboard displaying a brand new Audi; the algorithm learns from curves, edges and colour intensities and, a curve is just a curve in 2D. Add to this, the lighting and weather conditions. These variables interplay and generate confounding scenarios for the algorithm. A high-quality annotation at scale is required to address these complex issues. And, IDD manages to do that quite well. Other complexities resulting in unstructured data include ambiguous road boundaries, diversity of vehicles and pedestrians, extensive use of information board, the diversity of ambient conditions and high density of motorbikes Key Takeaways Of The Study Identifies drawbacks prevalent in existing datasets and the need for additional labels with a hierarchy to reduce confusion Examines the domain discrepancy properties with respect to other semantic segmentation datasets The unconstrained nature of the dataset provides a novel setting for more situational awareness and optimum path planning Forms a platform and sets a benchmark to solve advanced computer vision problems Outlook Though driverless AI is advancing rapidly with the support of American giants like Google and Tesla; the technology is not yet mainstreamed. In the Indian scenario, however, technical problems persist. This exclusive Indian data provide much-needed impetus to the autonomous industry not only in India but across the world. For a densely populated country like India, leading automotive companies are looking to capture a major chunk of the market cap and would certainly use up resources such as this IDD dataset. “Autonomy will come step by step. We’ll see semi-automatic systems, driving assistants, interactive systems and safety features that are AI enabled before that,” observed Jawahar, who led the team behind this year-long prestigious project. But, in India, things don’t look so smooth. Earlier this year, Nitin Gadkari, Minister of Road Transport and Highways, had blatantly opposed the arrival of driverless cars. On the contrary, the Traffic Amendment Bill encourages exploring new technologies to improve road transport in India. India is on track to become the world’s third largest car manufacturer. With such high numbers, policymakers need to be flexible about the stance they take and imbibe solutions which might seem technologically advanced but will soon become a norm in the near future. Experts predict an autonomous intervention in this sector by 2025. For India to keep up with the pace, it needs to prepare the roads, fill the potholes, intensify the research and be ready to deploy when the advancements have reached maturity.","excerpt":"With the number of deaths caused by road-related issues and accidents reaching 400 per day in India, it has become imperative to find solutions to minimise road fatalities. Developed countries have already embarked on a non-human technological journey to alleviate fatal human flaws, however developing economies like India, still have to catch up. For example, […]","categories":["Deep Tech"],"tags":["autonomous systems","computer vision autonomous vehicles","IIIT-H","Intel"],"author_name":"Ram Sagar","publish_date":"2018-12-27T09:28:56","publication_year":"2018","word_count":947,"keywords":["Go","API","AWS","AI","ETL","innovation","computer vision","RAG","autonomous systems","GAN","IIIT-H","computer vision autonomous vehicles","R","Intel"],"extracted_tech_keywords":["AI","computer vision","RAG","AWS","R","Go","API","ETL","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/indias-first-open-source-traffic-dataset-is-paving-the-road-for-autonomous-vehicles\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":7482,"title":"Analytics India Leaders Outlook -2015","content":"We at Analytics India Magazine have always tried to bring the community together and provide information about the industry that is not otherwise assessable to all. Supporting this initiative we bring to you what the leaders of Indian Analytics forecast about the industry for the next 1 year. Analytics India Leaders outlook Survey reveal the perspectives of leaders in India on the analytics industry for next twelve months. We surveyed 45 individuals across the industry from the ranks of key decision makers for this study. Overall, the responses have been extremely positive and sentiment looked ‘up’ for the year ahead. Here’s the profile of our respondents. 98% of respondent are confident that the demand of Analytics at their organization will increase in next 12 months. We asked regarding their confidence about analytics being a key focus area for organizations globally over next 12 months – 36% respondent gave a 10\/10 score Just 7% respondent gave a score of less than 7 out of 10. Average Confidence score was 8.6\/10 We calculated the Net Sentiment Score (Calculated as difference of % with score of 10s & 9s and Score from 0 through 6s), at 49%. This is an exceptionally high sentiment score. 96% of decision makers plan to increase their analytics workforce in next 12 months. Only 4% believe that they have no hiring plan for the coming year. Across the industry medium to large organizations are trying to build analytics COE. Majority of them are new to this and are looking out to hire talent to fulfil their requirement. 67% decision makers believe that ‘Unavailability of Analytics Talent’ is the major challenge that they face. This is in line with our earlier findings and is widely documented and spoken about. There is a huge demand for Analytics professionals across levels. The pool of resources is not sufficient to fulfil the current requirements. While the industry is collaborating with several institutes and organizations to fulfil these demands, the quality of professionals and experience still remains challenge. What’s came out as an interesting outcome is that 58% of leaders believe that ‘Little knowledge of analytics among customers’ is a major challenge for them. Obviously, we have not done enough to propagate analytics understanding to a wider audience in the market. While Analytics and big data has been a buzz word around the network, very few understand how they can encash on it. And even fewer can identify the key areas within their business to apply it to. Few leaders are wary of competition, just 18% believe it’s a challenge for them. Only 20% believe that analytics demand has peaked. The number of key players is still limited. While we have quite a few start-ups coming up, the demand for expertise to fulfil the current industry requirement is quite high. ‘Digital\/ online\/ Mobile\/ Social’ is considered as a top area of growth in analytics by 62% of respondent. ‘HR\/ Talent\/ People’ is considered as a growth area by just 13% respondent. Obviously, the adoption of analytics in this is slow, that might change over time when real benefits begin to emerge.","excerpt":"We at Analytics India Magazine have always tried to bring the community together and provide information about the industry that is not otherwise assessable to all. Supporting this initiative we bring to you what the leaders of Indian Analytics forecast about the industry for the next 1 year. Analytics India Leaders outlook Survey reveal the […]","categories":["AI Features"],"tags":[],"author_name":"Дарья","publish_date":"2015-06-01T13:18:33","publication_year":"2015","word_count":518,"keywords":["big data","programming_languages:R","AI","Git","RAG","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Git","big data","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-india-leaders-outlook-2015\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61459,"title":"How Microsoft Set A New Benchmark To Track Fake News","content":"Researchers from Microsoft, along with a team from Arizona State University, have published a work that has outperformed the current state-of-the-art models that detect fake news. Though the prevalence and promotion of misinformation have been since time immemorial, today, thanks to the convenience for access provided by the internet, fake news is rampant and has affected healthy conversations. Given the rapidly evolving nature of news events and the limited amount of annotated data, state-of-the-art systems on fake news detection face challenges due to the lack of large numbers of annotated training instances that are hard to come by for early detection. In this work, the authors exploited multiple weak signals from different user engagements. They call this approach multi-source weak social supervision or MWSS. They have leveraged limited amounts of clean data along with weak signals from social engagements to train deep neural networks in a meta-learning framework to estimate the quality of different weak instances. Experiments on real-world datasets demonstrate that the proposed framework outperforms state-of-the-art baselines for early detection of fake news without using any user engagements at prediction time. How Meta Learning Came To The Rescue Fake news is diverse in terms of topics, content, publishing methods and media platforms. There are also sophisticated linguistic styles geared to emulate accurate news. Consequently, training machine learning models on such sophisticated content requires large-scale annotated fake news data that is egregiously difficult to obtain. Second, it is important to detect fake news early. Most of the research on fake news detection rely on signals that require a long time to aggregate, making them unsuitable for early detection. Prior works on detecting fake news rely on large amounts of labelled instances to train supervised models. Such large labelled training data is difficult to obtain in the early phase of fake news detection. To overcome these issues, the authors devised a model that leverages a small amount of data that is annotated manually, and a large amount of weakly annotated data for joint training in a meta-learning framework. The models learn to estimate their respective contributions to optimize for the end task. To model the weights of weak labels, a Label Weighting Network (LWN) is developed. This network monitors the learning process of the fake news classifier. The above picture illustrates multi-source weak social supervision (MWSS) in two phases: (a) compute the validation loss based on the validation dataset and retain the computation graph for LWN backward propagation; (b) update the classifier and its parameters through backward propagation on clean and weakly labelled data. For experiments, the authors have used fake news detection benchmark datasets called FakeNewsNet, which contains news content from GossipCop5 and PolitiFact6. This data is annotated by professional journalists and experts, along with social context information. News content includes meta attributes of the news (eg, body text), whereas social context includes related users’ social engagements on the news items (eg, user comments in Twitter). Key Findings The authors have observed the following from their experiments: Training only on clean data achieves better performance than training only on the weakly labelled data consistently across all datasets On incorporating weakly labelled data in addition to the annotated clean data, the classification performance improves, when compared with results when using only clean labels Merely merging the clean and weak sources of supervision without accounting for their reliability may not improve the prediction performance The MWSS model not only learns the importance of different instances, but also learns the importance of the corresponding source Powered by meta learning with a Label weighting network, MWSS outperforms state-of-the-art baselines without using any user engagements at prediction time Read the original paper here.","excerpt":"Researchers from Microsoft, along with a team from Arizona State University, have published a work that has outperformed the current state-of-the-art models that detect fake news. Though the prevalence and promotion of misinformation have been since time immemorial, today, thanks to the convenience for access provided by the internet, fake news is rampant and has […]","categories":["AI Trends"],"tags":["deep fake","Microsoft"],"author_name":"Ram Sagar","publish_date":"2020-04-11T16:00:13","publication_year":"2020","word_count":608,"keywords":["Go","API","meta-learning","machine learning","programming_languages:R","AI","neural network","programming_languages:Go","RAG","deep fake","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","neural network","RAG","R","Go","API","meta-learning","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/microsoft-fake-news-solution\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":17902,"title":"Flipkart Knocks On Microsoft’s Door For Building AI, ML-Based Future Solutions","content":"E-commerce giant Flipkart is working with Microsoft to start using artificial intelligence and machine learning-based solutions to make future sales easy said a report. Flipkart seems to have learnt its lessons from its previous ‘Big Billion Day’ sales and did not face any major glitches this time around. The website saw almost 30 times its usual amount of traffic during the sale and is already looking at improving its technology systems for the next sales, an Economic Times report said. “AI and ML are becoming the focus for us now. We think there is a lot of opportunities to optimise how we do things like merchandising and offer placement. So we are putting those systems in place and looking to take away any kind of manual tuning and optimisation,” ET quoted Vinay YS, vice-president of engineering at Flipkart. The e-commerce giant has been working with Microsoft, which has invested $200 million, to build those capabilities. “We mostly look at Microsoft from the point of view of AI and ML. There are a lot of those capabilities on Microsoft side that we would like to leverage. On aspects like voice recognition we want to partner with them deeply,” Vinay said according to the report. In February, Flipkart had partnered with Microsoft to make Azure its exclusive cloud-computing platform. Currently, it does not use a public cloud, having built a large private cloud infrastructure. The last major investment in its own data centre was in 2015. “On the pure cloud front, we are still evaluating – what they have in India and what we need,” said Vinay, as told by ET. Recently, Microsoft announced setting up a new healthcare department at its ­Cambridge research facility, as part of plans to use its artificial intelligence software to ­enter the health market. Last week, Microsoft unveiled an AI-based automated threat investigation system to enhance the security of devices. The system will enable the Microsoft users with insight to take action against modern-day threats while also increasing the efficiency of the machines, the company had said.","excerpt":"E-commerce giant Flipkart is working with Microsoft to start using artificial intelligence and machine learning-based solutions to make future sales easy said a report. Flipkart seems to have learnt its lessons from its previous ‘Big Billion Day’ sales and did not face any major glitches this time around. The website saw almost 30 times its […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Flipkart","Machine Learning","Microsoft"],"author_name":"Priya Singh","publish_date":"2017-09-26T09:12:46","publication_year":"2017","word_count":341,"keywords":["artificial intelligence","machine learning","cloud_platforms:Azure","AI","programming_languages:R","R","ML","Flipkart","Machine Learning","RAG","AI (Artificial Intelligence)","Azure","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","Azure","R","cloud_platforms:Azure","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/flipkart-knocks-microsofts-door-building-ai-ml-based-future-solutions\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172152,"title":"Why TCS Changed Its Bench Policy","content":"Tata Consultancy Services (TCS) has introduced a significant change to its talent deployment policy, setting a new benchmark for employee utilisation. Under the revised guidelines, all employees are now mandated to be billed for a minimum of 225 days, with a reduced allowable bench time to 35 days per year. “At any given point in time, associates must have been allocated for a minimum period of 225 business days in the last 12 months, failing which, necessary due diligence and appropriate management action will be exercised,” read an internal company email verified by AIM. “Benching” is an ordinary term in Indian IT services, which refers to employees who remain on the company payroll but are not actively engaged in client projects or other billable tasks. These employees are essentially “on the bench”, standing by as available resources for future projects. Although some amount of bench time is unavoidable, its management is crucial. This is the reason behind TCS’ bold, performance-driven policy shift—a move aimed at fostering learning and greater accountability among employees. However, such a policy is primarily designed to drive maximum resource utilisation, which is expected to improve business margins. On the sidelines, this structured upskilling is meant to keep pace with the rapid evolution of AI. Survival of the Fittest While artificial intelligence (AI) is reshaping service delivery by introducing automated processes, predictive analytics, and intelligent systems, companies like TCS are under pressure to optimise their human capital for high-value, tech-driven projects. Students and aspiring IT professionals, meanwhile, get a loud wake-up call: the industry rewards skills tailored to AI and digital innovation. “All new joiners are expected to have an allocation from day one of joining the organisation. In case of any deviation, such associates should connect with the concerned RMG on an immediate basis for guidance,” the mail read. This policy, overseen by regional general managers (RGMs) and utilisation-focused dashboards, ties employee performance directly to revenue-generating projects. It reflects TCS’ drive to maximise resource efficiency amid a global IT spending slowdown, projected at 6.8% growth in 2025. As seen in the earnings call, TCS reported weaker-than-expected Q4 results, with revenue and net profit falling short of analyst projections. Moreover, TCS typically distributes annual increments between April and July. However, this year, the company has deferred the decision, citing macroeconomic uncertainty, as Milind Lakkad, chief HR officer, indicated during a post-earnings call. Alouk Kumar, founder and CEO of Inductus Group, told AIM that global clients are increasingly demanding faster project turnarounds and AI-integrated solutions, reducing the tolerance for idle resources. To address this, TCS aims to curb bench-related costs, which, as he mentioned, are estimated to account for 8-10% of its payroll, especially given its operating margin of 24.6% in FY24. Moreover, automation is diminishing the demand for low-skill roles, pushing the company to redeploy talent to high-value AI\/ML, cloud, and cybersecurity projects. “In the event of an associate’s failure to comply with the provisions of this policy, the organisation reserves the right to take disciplinary action—including cessation of service, as per the organisation’s disciplinary action framework,” the mail read. Failure to meet billable targets could result in consequences such as salary freezes, deferred appraisals, or even performance-based exits, which have historically affected 1–2% of TCS’ workforce of over 6 lakh employees annually, Kumar highlighted. He further added that this policy reflects a broader shift within the Indian IT industry, where companies like TCS, Infosys, and Wipro are transitioning from labour-arbitrage models to innovation-driven growth. To this, Biswajeet Mahapatra, principal analyst at Forrester, said, “While the new bench policy does not explicitly mandate AI adoption, it underscores the increasing importance of upskilling in AI, automation, and other emerging technologies to remain competitive and relevant in the IT sector.” Need for Urgent Upskilling With AI expected to automate a wide range of IT tasks, including coding and testing, traditional bench roles may shrink. TCS is addressing this by reskilling 3.5 lakh employees in AI\/ML, GenAI, and digital platforms. The upskilling initiative focuses on tools like Azure AI, AWS SageMaker, and low-code platforms. Neeti Sharma, CEO at TeamLease, told AIM that organisations are looking for freshers who are not just digitally literate but also have skills that can help them incorporate, use and create AI, GenAI and agentic AI tools and models. The 225-day billable target incentivises employees to upskill proactively but exposes systemic challenges. Reports indicate that only 20% of India’s 1.5 million annual engineering graduates are AI-ready, creating pressure on internal reskilling programs. Positively, it offers access to TCS’ robust learning ecosystem, including AI certifications and in-house hackathons, enabling career mobility toward roles like AI solution architects. In the long run, the policy has the potential to reshape India’s IT talent pipeline by driving academia-industry collaboration and widening inequality for non-elite graduates. In essence, TCS’ new bench policy serves as a microcosm of the Indian IT sector’s AI-driven transformation, balancing opportunity with intense pressure for its current and future workforce. For TCS Employees However, the 35-day cap heightens anxiety, especially for employees in commoditised roles—for example, legacy Java developers—who face fiercer internal competition for billable projects. Furthermore, “project availability” is also seen as a job satisfaction factor, and reduced bench time could spike attrition among mid-career professionals with five to 10 years of experience. The strict bench time rules and mandatory work-from-office (WFO) policies may lead to increased stress among employees, while potentially reducing flexibility, which could result in talent attrition. There is also a risk of focusing too heavily on utilisation rates, which might hinder innovation within the organisation. One Reddit user pointed out a potential downside of TCS’s new policy, suggesting that employees may get threatened or blackmailed by RMGs to take irrelevant roles. Another LinkedIn user commented that TCS’s push for structured learning is a smart approach, but it needs to be paired with mentorship and realistic timelines. Hybrid work flexibility is increasingly seen as the future of work, and mandatory WFO during bench periods may alienate valuable talent. Moreover, performance should not solely be measured by billing hours—value also comes from contributions to R&D, internal innovation, and readiness efforts.Furthermore, project volatility could leave employees stranded on the bench despite efforts to redeploy them.","excerpt":"With AI expected to automate a wide range of IT tasks, including coding and testing, traditional bench roles may shrink.","categories":["AI Features"],"tags":["Indian IT","TCS"],"author_name":"Shalini Mondal","publish_date":"2025-06-23T10:46:39","publication_year":"2025","word_count":1030,"keywords":["GenAI","artificial intelligence","agentic AI","AWS","AI","ML","predictive analytics","RAG","Aim","analytics","Indian IT","TCS"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","GenAI","agentic AI","Aim","RAG","predictive analytics","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-tcs-changed-its-bench-policy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":16625,"title":"NFL is no longer just American football, it’s a number game with data analytics driving every facet of it","content":"Once dubbed as a slow league, averse to game-changing technology, America’s most popular league NFL has moved away from watching game films on Apple iPads and training videos culled from drones to advanced tech upgrades such as better stadium WiFi, wearables and real-time tracking technology. And now over the years, data has been married into every facet of NFL. This is one major league that loves its data that has the potential to reap millions of profits. From game stats to drafting of players based on grading and trait based analysis and powering fan engagement, data is finely tuning the league. Real-time tracking technology has also made a dramatic impact in NFL, allowing the league to experiment with quarter-sized radio-frequency identification transmitters embedded in the shoulder pads of each player, a system developed by Zebra Technologies. The transmitters collect data on the player’s speed, position and other metrics in real-time. It is the same RFID chips deployed by Zebra Technologies that gave higher visibility into asset tracking and greater supply chain visibility across industries such as manufacturing, healthcare, retail and more. Analytics India Magazine caught up with Jill Stelfox, VP & GM of Zebra Location Solutions who gave a peek into RFID (Radio Frequency Identification) technology and the partnership with industry leading Kinduct world’s leading data and analytics solutions provider. RFID vs GPS One of the most popular asset tracking technologies, the NFL chose to go with RFID over GPS because RFID is accurate up to 15 cm (6 inches), while GPS is only accurate to the meter, revealed Stelfox. Here’s where the two differ. While RFID technology is battery-operated and works extremely well in small range, GPS is satellite-based and works well in long range. An inexpensive option compared to GPS technology, RFID has been widely used in enterprises for better utilization of IT assets and reducing overhead costs. “The Zebra Sports Solution leverages the same tracking and location solutions technology that Zebra implements globally for multinational corporations in healthcare, retail, manufacturing, and transportation & logistics to give visibility to an organization’s assets, people and transactions,” shared Stelfox. In NFL, RFID chips are not only changing the game for teams and coaches, this has also improved fan engagement by making stats on player activity available to fans and broadcasters, via the Next Gen Stats program. Future of NFL analytics According to consulting group Booz Allen Hamilton, historically NFL had the least amount of data to play with. But now with sensors on shoulder pads or wearable tech, the league is awash with data. Predictive analytics is used to help teams make better informed decisions regarding a) game strategy; b) building better teams; c) training and player management; d) most importantly monetizing fan engagement; e) scouting talent. Today, predictive analytics can crunch data to eliminate guesswork during draft process and even forecast injuries in players. When it comes to player management and injury prevention, with so much frenzied activity on the side, computer tracking technology wouldn’t have been able to deliver the goods. With Zebra’s tracking solutions, NFL teams have access to terabytes of data that redefined the playbook with every snap of the ball. As the “Official On-Field Player Tracking Provider” of the NFL, Zebra captures high-speed player data and converts it into real-time, usable statistics. There’s more to it — the technology is used by coaches and trainers to gain greater visibility into player dynamics, and they can customize workouts to focus on speed, agility and other critical elements of the game. Stelfox reveals users receive real-time updates that are fed into customized analysis and individual action plans for athlete and team preparation, injury prevention, and performance enhancement. These updates are available to all NFL and National Collegiate Athletic Association (NCAA) teams that use the Zebra Sports practice system. Besides, on the coaching end, data and analytics enables trainers to understand how the players are performing physically. Subsequently, the input from coaches limits the chances of injury, aids faster recovery and leads to better training scenarios. With data-based evaluation becoming the norm, teams are always coming up with the best analytical methods to evaluate data effectively. Bottomline – can a team win a Super Bowl on the back of analytics Data has raised the bar for players who find ways to push the physical limits and coaches who are drumming up personalized training sessions around the needs of the players. The number game has been significantly upped and advanced analytics in NFL has moved beyond having a consultant onboard to determining how far you can go with data. Mid-size companies have spawned on the back of this and it’s no longer the future and teams are heavily investing in build tech staff and technology to raise the bar Today, in NFL, every team is knee-deep in numbers.  That explains Zebra Technologies recent tie-up with Kinduct Athlete Management System. Canada-based Kinduct Technologies is at the forefront of advanced analytics and data aggregation. One of the chief highlights is that Kinduct has been able to standardize what metric matters the most in sports, separating noise from data to crank out actionable insights With the integration with industry leading Kinduct Athlete Management System, teams will get an extended, detailed view into players’ health, wellness, and overall performance. Now that NFL is turning into a big data machine, analyzing structured and unstructured queries on the side and diving deep into fan sentiments, it wouldn’t be a surprise to find a time winning the NFL on the back data-driven decisions anytime soon.","excerpt":"Once dubbed as a slow league, averse to game-changing technology, America’s most popular league NFL has moved away from watching game films on Apple iPads and training videos culled from drones to advanced tech upgrades such as better stadium WiFi, wearables and real-time tracking technology. And now over the years, data has been married into […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-07-31T05:51:41","publication_year":"2017","word_count":919,"keywords":["big data","Go","AI","R","ML","RAG","ViT","analytics","GAN","predictive analytics"],"extracted_tech_keywords":["AI","ML","analytics","RAG","predictive analytics","R","Go","big data","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/nfl-no-longer-just-american-football-number-game-data-analytics-driving-every-facet\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10089799,"title":"Microsoft Introduces GPT-4 in Azure OpenAI Service","content":"Microsoft today announced that it will be making GPT-4 available in preview in Azure OpenAI Service. The company said that customers can start applying for access to GPT-4 today. With this, Microsoft looks to offer its Azure customers access to advanced AI models, including the likes of GPT-3.5, ChatGPT, and DALL.E 2, alongside providing Azure AI-optimised infrastructure, enterprise-readiness, compliance, data security, and privacy controls, along with many integrations with other Azure services. Here’s a quick glimpse of its pricing, effective from April 1, 2023: GPT-4PromptCompletion8k context$0.03 per 1,000 tokens$0.06 per 1,000 tokens32k context$0.06 per 1,000 tokens$0.12 per 1,000 tokens Some of the customers leveraging GPT-4 on Azure OpenAI Service include Epic Healthcare, Coursera, and Coca-Cola among others. Mustafa Furniturewala, senior vice president of engineering at Coursera said that it is using Azure OpenAI Service to create a new AI-powered learning experience on its platform, enabling learners to get high-quality and personalised support throughout their learning journeys. Lokesh Reddy Vangala, senior director of engineering, data and AI at The Coca-Cola Company, said with Azure Cognitive Services, they have harnessed the transformative power of OpenAI’s text and image generation models to solve business problems and build a knowledge hub. Read: Top 6 Use Cases of GPT-4 Microsoft, in its blog post, said while the new Bing and Microsoft 365 Copilot is powered by GPT-4, the new announcement will allow businesses to take advantage of the same underlying advanced models to build their own applications leveraging Azure OpenAI Service. The company also said that this new offering would allow bot developers to create virtual assistants in minutes using natural language with Copliot in Power Virtual Agents. It believes that with GPT-4 in Azure OpenAI Services, companies can streamline communications internally and with their customers. It claimed that this model comes with additional safety investments to reduce harmful outputs. Backed by the supercomputing and enterprise capabilities of Azure, Microsoft believes that companies of all sizes can deploy language models into production using Azure OpenAI Service. It said that by using its solutions companies can improve customer experiences end-to-end, summarise long-form content, help write software, and even reduce risk by predicting the right tax data. Last week, Microsoft announced the launch of Microsoft 365 Copilot, an AI-powered assistant that aims to transform the way of working. The new tool is built on large language models (LLMs), including OpenAI’s GPT-4, combined with Microsoft Graph data. It is integrated into Microsoft 365 apps such as Word, Excel, PowerPoint, Outlook and Teams. A few days before that it also announced the launch of Microsoft Dynamics 365 Copilot, touted to be the world’s first copilot in both CRM and ERP.","excerpt":"Users can start applying for access to GPT-4 today.","categories":["AI News"],"tags":["GPT","GPT4","Microsoft","Microsoft Azure"],"author_name":"Tasmia Ansari","publish_date":"2023-03-21T23:51:19","publication_year":"2023","word_count":441,"keywords":["ChatGPT","TPU","OpenAI","AI","R","ML","virtual assistants","RAG","GPT","Aim","Microsoft Azure","Azure","Microsoft","GPT4"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Aim","RAG","virtual assistants","Azure","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-introduces-gpt-4-in-azure-openai-service\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10053230,"title":"The First Quantum Toolkit And Library For Natural Language Processing","content":"Cambridge Quantum (CQ) announced the release of the world’s first quantum natural language processing toolbox and library (QNLP). The goal is to speed up the development of real-world QNLP applications such as automated dialogue, text mining, language translation, text-to-speech, language production, and bioinformatics. Lambeq is the world’s first software toolkit for quantum natural language processing that is capable of transforming phrases to quantum circuits. Lambeq is completely open-sourced to benefit the global quantum computing community and its quickly developing ecosystem of academics, developers, and users. In addition, Lambeq integrates smoothly with CQ’s TKET, the world’s most popular and fastest-growing quantum software development platform, also open-source. This benefit gives QNLP developers access to as many quantum computers as feasible. According to Bob Coecke, Chief Scientist at Cambridge Quantum Computing, “This work was founded on breakthroughs made by myself, Steve Clark, currently CQ’s Head of artificial intelligence (AI), and others. The use of natural language processing (NLP) is at the heart of these investigations. After publishing details about the world’s first QNLP implementation by CQ on actual quantum computers a few months ago, and our initial disclosure of the underlying ideas in December 2019, the release of lambeq is the inevitable next step.” Lambeq’s Features Lambeq facilitates and automates the design and deployment of compositional-distributional (DisCo) NLP experiments, as described by CQ scientists. This deployment involves moving away from syntax\/grammar diagrams, which encode a text’s structure, and toward TKET-implemented (classical) tensor networks or quantum circuits, which TKET can optimise for machine learning tasks like text categorisation. Furthermore, Lambeq is designed in a modular manner so that users can swap components in and out of the model and have architectural design flexibility. Lambeq minimises the entry barrier for practitioners and researchers interested in AI and human-machine interactions, which could be one of quantum technology’s most important applications. TKET presently has a user base of hundreds of thousands of people all over the world. Lambeq has the potential to become an essential toolbox for the quantum computing community looking to engage with QNLP applications, which are one of AI’s most lucrative sectors. QNLP will apply to the study of symbol sequences that originate in genomes and proteomics, according to a critical point that has lately become clear. Merck Group, a lambeq launch partner and early adopter, has published a research paper on QNLP. Thomas Ehmer, the co-founder of the quantum computing Interest Group and director of Merck’s IT Healthcare Innovation Incubator, said, “At Merck, we’re working on using the unique properties of quantum computing to make fundamental advances. For example, our recently disclosed QNLP study with TU Munich academics has demonstrated that binary classification tasks for sentences utilising QNLP algorithms can produce results comparable to existing classical methods even at this stage. Critically, one can see how QNLP’s methodology paves the way for explainable AI, and hence for more accurate, accountable intelligence – which is critical in health.” Conclusion Many useful NLP applications are beyond the reach of conventional computers. There will be a slew of new commercial quantum applications as QNLP and quantum computers improve. Honeywell and Cambridge Quantum will have a major strategic advantage in developing future QNLP applications because of Coecke’s knowledge and experience. Lambeq has been released as a standard Python repository on GitHub. Lambeq’s quantum circuits have been tested and deployed on IBM quantum computers and Honeywell Quantum Solutions’ H series devices. Additionally, a technical study was published on arXiv.org, available here.","excerpt":"QNLP applications can be developed more quickly if lambeq converts words into quantum circuits.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Machine Learning","Natural Language Processing","Quantum Computing"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-11-13T10:00:00","publication_year":"2021","word_count":573,"keywords":["Quantum Computing","Go","artificial intelligence","machine learning","AI","Natural Language Processing","Machine Learning","Git","NLP","Python","ViT","GitHub","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","Python","R","Go","Git","GitHub","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/quantum-natural-language-processing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10008910,"title":"Fractal Hives Off Theremin.ai After Raising Funds From OLMO Capital","content":"Theremin.ai, one of its subsidiaries of Fractal, a company that is dealing with artificial intelligence and analytics, has raised funds from OLMO Capital. In a recent LinkedIn post, Co-founder, Group Chief Executive & Vice-Chairman, Fractal Analytics, Srikanth Velamakanni, stated his excitement of sharing the news — “I am excited to share with you that theremin.ai, Fractal’s AI-driven automated investing business has raised funds from OLMO capital.” He further stated that “We set up theremin.ai to test whether our algorithms could find signals in a nearly perfect capital markets context and we are encouraged by the results.” According to the media, the funds will be used for creating their algorithmic investment product and for scaling its talent. Founded in early 2019, the startup Theremin.ai leverages deep reinforcement learning to drive quantitative investment strategies for the Indian and Asian capital markets. The company technology platform identifies unique alpha investment opportunities in financial markets incorporating both conventional and alternative data sources. When asked, Gulu Mirchandani, Chairman, OLMO Capital stated to the media that the company is very excited to partner with the startup. According to him, not much innovation has happened in this regard in the Indian and Asian capital markets, and there is a tremendous potential to use artificial intelligence to make better informed and high-performance investment strategies. “Theremin.ai has the right team and capabilities to capitalise on the situation, and we are glad to partner with them to help accelerate their work in this space,” said Gulu Mirchandani. Additionally, Satish Raman, Chief Strategy Officer, Fractal stated to the media that Fractal’s Ideas2Business initiative is focused on promoting cutting-edge ideas and innovation that help in creating new and market-ready AI-based digital products, platforms and solutions. The spin-outs of Theremin.ai is a testament to the success of this program, stated Raman. “Cuddle.ai and Eugenie.ai are two other AI products that we believe will soon take the same route, even while we have many others in the early stages, which we will bring to the market,” concluded Raman.","excerpt":"Theremin.ai, one of its subsidiaries of Fractal, a company that is dealing with artificial intelligence and analytics, has raised funds from OLMO Capital.  In a recent LinkedIn post, Co-founder, Group Chief Executive & Vice-Chairman, Fractal Analytics, Srikanth Velamakanni, stated his excitement of sharing the news — “I am excited to share with you that theremin.ai, […]","categories":["AI News"],"tags":[],"author_name":"Sejuti Das","publish_date":"2020-10-05T18:32:50","publication_year":"2020","word_count":334,"keywords":["Go","API","artificial intelligence","AI","innovation","Git","RAG","analytics","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","RAG","R","Go","Git","API","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/theremin-ai-raises-funds-from-olmo-capital\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23494,"title":"Is IBM A Dark Horse In AI Race? This Is How It Has An Edge Over Its Legacy Peers","content":"Once dubbed as a “legacy tech” company, IBM has made a serious push in being recognized as a top-tier AI research company. While Watson did receive some flak for being a marketing ploy, it didn’t restore the company’s glory days in terms of revenue and growth. It however did help in IBM’s early embrace of AI and cognitive technologies. Another area where IBM scores a major lead is in quantum computing, where it is leading the charge. According to Tom Rosamilia, senior vice president at IBM Systems, the company is betting big on breakthrough technologies which are designed for cloud and AI workloads. “Whether it’s to accelerate customer insight and services delivery or provide data encryption across massive amounts of data, IBM Systems is uniquely differentiated for smarter businesses.” Though It’s Hard To Peg IBM As A Leader In This Race, The Company’s Early Embrace In Emergent Technologies Has Given It An Edge Over Its  Legacy Peers: (a) IBM cloud comes fourth in the spot after the Big 3 — AWS, Microsoft and Google (b) IBM is one of the companies leading in AI research with its impeccable body of researchers (c) The company recently released POWER9 and NVIDIA GPU system that set a benchmark record against Google’s TensorFlow. According to IBM researchers, the training time for an online advertising dataset released by Criteo Labsis was 46 times faster than the best result that has been previously reported — which used TensorFlow on Google Cloud Platform — to train the same model in 70 minutes. According to the IBM blog, the POWER9 systems are specifically built for data-intensive AI workloads. The company has also worked in conjunction with innovators like NVIDIA to boost performance by 10 times faster. (d) Just like Google bolstered its presence in cloud with a slew of AI tools, IBM will also release PowerAI, the deep learning platform which supports frameworks like TensorFlow, Torch and Caffe, as well as the company’s own deep learning frameworks in the cloud (e) IBM’s THINK 2018 conference saw the release of Snap Machine Learning software – a breakthrough AI technology that can be used to train machine learning models for massive data sets from financial records to weather forecasting to online marketing. The name SNAP ML indicates scientists and practitioners can “train models faster than they can snap their fingers”, thereby leading faster insights and lower cloud costs. (f) IBM is a clear leader in quantum computing and is betting heavily on one of the most critical computing technologies (g) IBM also made an early move in blockchain with IBM Blockchain platform — a SaaS offering delivered via the IBM cloud IBM Scores Over Its Legacy Tech Peers One of the areas where IBM scores is its early embrace of technologies. The company also made decisive business and cultural changes to capitalise on emergent technologies. Known for its excellent body of researchers, IBM topped the list of publishing the most patents in 2017 — a position they have held for the last 25 years. One of the categories where IBM filed the most patents is artificial intelligence, machine learning and autonomous vehicles. We can’t deny how IBM has crafted a niche (perhaps more than a niche) in the market for its vertically-focused AI capabilities with Watson and as the market sentiment around Watson improves, so does it chance of capitalising on the burgeoning market for these services. Watson, in a way created an AI marketplace of sorts that helped large enterprises evaluate different AI capabilities for a section of a problem. However, Watson faces threat from smaller players that are bringing faster and more cost-effective ways of implement AI solutions. Bringing The Power Of AI To The Cloud — Is IBM Toeing Google’s Line? This is a move reminiscent of Google that has marketed its Cloud AutoML as a one stop shop for simplifying the work behind AI. The rollout of Cloud AutoML has significantly bolstered the Google Cloud and is seriously rivalling AWS. While Google Cloud has emerged as the best-in-class platform for AI and machine learning capabilities, training. Meanwhile, IBM has ramped up its GPU acceleration for to help researchers train large datasets faster thus leading to higher quality insights. SNAPML is a move to remove the training time bottleneck and rivals Google’s Tensorflow in terms of training time that is 46 times faster. Given that the future of every industry is in AI and machine learning which are being increasingly used to solve business problems, IBM’s AI platform now in the cloud will give it a competitive advantage in the AI marketplace.","excerpt":"Once dubbed as a “legacy tech” company, IBM has made a serious push in being recognized as a top-tier AI research company. While Watson did receive some flak for being a marketing ploy, it didn’t restore the company’s glory days in terms of revenue and growth. It however did help in IBM’s early embrace of […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-04-11T06:53:05","publication_year":"2018","word_count":765,"keywords":["Go","API","machine learning","artificial intelligence","AWS","AI","ML","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","TensorFlow","AWS","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/is-ibm-a-dark-horse-in-ai-race-this-is-how-it-has-an-edge-over-its-legacy-peers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10069829,"title":"100 Most Influential AI Leaders in India 2022","content":"AIM 100: The Most Influential Leaders in AI 2024As AI reshapes our world, AIM introducing the most influential global leaders in AI. Analytics India Magazine brings to you its annual list of the Most Influential AI Leaders In India for the year 2022. Selected on the basis of the expertise they hold in the field of AI and Data Science, the list recognises the exceptional work these professionals have accomplished over the last year. Leaders in this list have shown great ability to lead their teams to develop state-of-the-art AI solutions and support their companies or clients. Along with significant career accomplishments, the leaders inducted into the awards demonstrate the kind of substantial business impact, technology vision and inspirational leadership that advances their role and creates a positive influence on the analytics profession. The final record is the work of over two months as AIM invited nominations from various leading organisations in India. The selection is done by a panel that includes members internal and external to AIM. The list is finalised primarily based on the criteria of the leader’s impact on the AI ecosystem in India. You can access the list from previous years at the link below: 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 These leaders were awarded the AI100 awards at this year’s MachineCon — an invite-only conference with analytics leaders coming together to explore the groundbreaking innovations and the challenges they face in data adoption. Below is the list of the top 100 influencers in the Indian AI industry for 2022, presented in alphabetical order. Ajoy Singh COO at Fractal Analytics Akash Agrawal Director – Data & Analytics at Tata Consumer Products Aman Garg Head of Data & Analytics at DSP Mutual Fund Ambica Rajagopal Group Chief AI Officer at Michelin Amit Gupta Director, Consumer | Asia, ME, Africa | Group CVM, Loyalty & Analytics | Head CoE Advanced Analytics at e& Amit Kalra Managing Director, Head Global Business Solution Centres India at Swiss Re Anand Das Head Digital and AI – Domestic 2W & International business at TVS Motors Anand K Sundaram Head Analytics Retail Bank at IDFC FIRST BANK Angshuman Ghosh Head of Data Science at Sony Research India Anirban Nandi Head of Analytics (Vice President) at Rakuten India Anish Agarwal Head of Analytics at Dr. Reddy’s Laboratories Ankit Mogra Head – BI & Analytics at Ather Energy Ankush Gadi Director – Data Management & Analytics at CRISIL Limited Arun Kumar Vice President Data Science at Paytm Arun Mehta Head of Data Engineering – D&A Technology, Services at Natwest Arun S. Pai SVP – Head of Analytics and Decision Sciences at Dream 11 Ashish Gupta Head of Data & Analytics at PerkinElmer Ashish Singru Senior Director & Head, Global Business Analytics Center at eBay Avinash Narasimha Analytics Leader at Koch Technology Center Bharat Belavadi India Head, Advanced Analytics Officer at Western Digital Bhargab Dutta Head of Digital & Analytics CoE at Colgate-Palmolive Bhupinder Singh Head, Data Science at Ecom Express Private Limited Chandra Shekhar Prasad Chief Data Officer at Mahindra First Choice Chiranjoy Chowdhuri Chief of Data Science & Analytics at Pidilite Debayan Bose Head of Analytics and Data Science at redBus (go-MMT group) Deependra Singh VP & Head Data Science and Analytics at Junglee Games Devendra Sharnagat Sr. Executive Vice President – Data Analytics and Customer Value Management at Kotak Mahindra Bank Dhruv Rastogi Vice President and Head of Data Science at IKS Health Farhat Habib Director, AI Glance at InMobi Ganesh Bhat Senior Manager – Analytics & IoT at CEAT Limited Gaurav Kumar Head – Data Science, Data Analytics and Data Engineering at MobiKwik Guninder Bajwa Head of India Business Centre at Circle K Hari Saravanabhavan Vice President – Global Analytics at Concentrix Harish Gudi COO at Tredence Haryyaksha Ghosh Chief Data Officer at Aadhar Housing Finance Ishu Jain Head Of Analytics for Central Functions at Swiggy Issac Mathew Sr. Director, Technology at Lowe’s Companies Inc. Jitendra Singh President and Chief Digital Officer at JK Cement Kamal Kumar Head Of Analytics at Myntra Karthik Shashidhar Senior Vice President and Head, Analytics, Business Intelligence and Data Science at Delhivery Kaushik Ghate Sr. Vice President II and Head – Customer Analytics and Data Sciences at HDFC Bank Kiran Samudrala Chief Data and Analytics Officer and Centre Head at Equifax Madhurima Khandelwal Head Of AI Labs, Global Decision Science, Credit & Fraud Risk at American Express Manindra Mohan SVP & Head of Analytics at Unacademy Mathangi Sri Chief Data Officer at Yubi (Formerly CredAvenue) Namit Sharma Vice President & Head GDCA at FICO™ Nidhi Pratapneni SVP, Product, Analytics & Modelling, Public Affairs, India & Philippines at WellsFargo Nishant Pradhan Senior Vice President at Axis Bank Nitendra Rajput VP & Head – AI Garage at Mastercard Padmashree Shagrithaya Global Head – Analytics & Data Science at Capgemini Pankaj Rai Group Chief Data and Analytics Officer at Aditya Birla Group Phaneendra Durgam VP & Head – Analytics at Cholamandalam Investment & Finance Company Pradeep Gulipalli Co-founder at Tiger Analytics Prakash Hegde Chief Digital & Information Officer (CDIO) & Head IT at L&T Construction & Mining Machinery Prasana Balasubramanian Head of Analytics at Sundaram Finance Prashanth Kaddi Partner, Analytics & Cognitive at Deloitte Prudhvi Vasa Analytics Leader at Postman Puesh Ajmani Senior Vice President & Global Chief Digital Officer at Welspun Group Rahul Bharde VP & Head of Analytics & Insights at Jubilant FoodWorks Ltd Rahul Chogle Head – Data and analytics at SpiceJet Rajiv Jain Chief Analytics & Risk Officer at Faircent.com Ram Kumar Head of Data Sciences at Snapdeal Ramachandran Srinivasan Head of Data Analytics & AI at SUN Mobility Ramya Achar Chief Product Officer at Tata Health Ravi Vijayaraghavan Head of Analytics at Flipkart Rishi Sharma Head of Analytics at SonyLIV Rishi Swami Head of Data Science at MPL Rohini Srivathsa National Technology Officer at Microsoft India Rohit Khushu Head of Analytics at Landmark Group Rupesh Khare Global Head – Advanced Data Analytics and Artificial Intelligence at ABB Sanjay Thawakar Corporate Vice-President & Head, AI Works & Business Insights at Max Life Insurance Sanjeev Chaube EVP & Head – Big Data & Advanced Analytics and Business Intelligence at Vodafone Idea Sanjeev Kumar Head of Analytics at Bigbasket Santhosh Vasanthakumar Site Lead, India – CDTO (Chief Digital Transformation Office) Product Management at VMware Sasidhar Vavilala EVP Analytics at Suryoday Small Finance Bank Saswata Kar Senior Director, Head of Global Product Data, Analytics & Data Sciences, GBS at Philips Satya Kiran Dhulipala Country Manager at Vertica\/ Microfocus Sayandeb Banerjee CEO and Co-founder at TheMathCompany Shailesh Kumar Chief Data Scientist, CoE AI\/ML at Jio Shashwat Kumar Head Advanced Analytics at L&T Financial Services Shiv Kumar Head – Analytics | IT Data & Analytics at Schlumberger India Technology Centre Shivani Venkatesh Head , Client Insights at RBL Bank Shuvajit Basu VP & Head of Data Science at Tata Digital Shweta Singh Chief Data Officer at Tata AIA Life Insurance Siddhartha Thimmavajjala Head of AI and ML at Standard Chartered Bank Sreekanth Menon VP – Data Science at Genpact Srinidhi Shama Rao Chief Strategy Officer at Aegon Life Srinivas Gopinath Head – Digital and Analytics at Aditya Birla Fashion and Retail Ltd Subhajit Ghosh Head – Data Strategy & Analytics at Adani Group Subhobroto Ghosh Head – Data & Analytics at Allstate India Surajit Roy MD & CEO at IDBI Intech Sutirtha Chakraborty Head – Global Data Science & Analytics Center of Excellence at Abbott Thomas Jacob Kollenkeril Head of Process Analytics at Biocon Group Venkatesh Channaraj VP – Growth & Analytics at Licious Vijay Nair Sr Dir & Regional Lead – Data, Analytics and AI (APAC & MEA) at Levi’s Vikash Raj Group Vice President (Head of Data) at Mahindra Group Vineet Shukla Head – Data Science & Analytics at South Indian Bank Vinod G Head – Data Science & Analytics at South Indian Bank Vishal Dhupar Managing Director, Asia South at NVIDIA Yoganand Tadepalli Global Chief Digital and Information Officer at PGP Glass","excerpt":"Analytics India Magazine brings to you its annual list of the Most Influential AI Leaders In India for the year 2022. Selected on the basis of the expertise they hold in the field of AI and Data Science, the list recognises the exceptional work these professionals have accomplished over the last year. Leaders in this […]","categories":["AI Features"],"tags":["AI leaders"],"author_name":"AIM Media House","publish_date":"2022-06-28T18:15:00","publication_year":"2022","word_count":1331,"keywords":["data science","Go","artificial intelligence","AI","ML","Git","RAG","Aim","analytics","AI leaders","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","data science","analytics","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-100\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10117213,"title":"MediaTek has the Tech to Power Any Devices in the World","content":"MediaTek’s chips power over 2 billion devices worldwide in a single year. This speaks volumes about the company’s significant grasp in the semiconductor space. Its chips power smartphones, automobiles, VR headsets and even wifi routers and satellites. Today, the company’s processors are also behind some of the most advanced AI smartphones in the market. For instance, the Vivo X100 series smartphone released last year, marketed by the smartphone maker as the ‘industry’s first AI phone’, is powered by MediaTek’s chipset. MediaTek’s AI processor allows Vivo to run a 7 billion parameter language model and a 1 billion parameter vision model locally on the device. At the Forefront of AI Revolution AI could kickstart a new smartphone upgrade cycle. Last year, the Taiwanese semiconductor company announced its latest AI processor Dimensity 9300. The chipset is equipped with MediaTek’s next-generation APU 790 processor. This processor boasts a remarkable 45% reduction in power consumption alongside enhanced performance. Its processing speed is eight times faster than its predecessor, the APU 690. Notably, it delivers significant advancements in generative AI performance and energy efficiency for edge computing applications. The APU 790 is purpose-built for generative AI tasks, representing a significant leap forward in capabilities compared to its forerunner. The APU 790 introduces support for NeuroPilot Fusion, enabling seamless LoRA low-rank adaptation and facilitating the utilisation of large language models featuring 1B, 7B, and 13B parameters, with scalability extended up to 33B. While MediaTek leads the low-mid segment smartphone market, Dimensity 9300 puts the company in direct competition with Qualcomm in the premium segment. “The Market Analytics report 2023, published by IDC and Counterpoint, reveals MediaTek’s dominance with a 49% share in the smartphone market, compared to the competition’s 24% share. This sustained leadership position spans nearly a year, both in India and globally,” Anuj Sidharth, deputy director, marketing and corporate communications, told AIM. Powering Autonomous Vehicles Recently, MediaTek announced that it is partnering with Graphics Processing Units (GPU) maker NVIDIA to make four new automotive chips for connected and autonomous vehicles – the new Dimensity CV-1, CM-1, CY-1, and CX-1. The new chips, ranging from CV-1 for entry-level in-vehicle experiences to CX-1 for premium experiences, will integrate AI processing capabilities and an NVIDIA RTX GPU. Furthermore, the platform will be compatible with NVIDIA’s Drive OS software. As the automotive industry transitions towards electric or self-driving vehicles, it opens the door for innovation and building AI-powered cars.  AI can come into play in cockpit solutions, Advanced Driver Assistance Systems (ADAS) to battery management systems. “The MediaTek cockpit solution, based on flagship three-nanometer technology, offers top-notch performance and user experience. It supports up to eight screens in the car, enabling immersive entertainment like music and social media. While security measures prioritise passenger safety, the solution enhances in-vehicle enjoyment for passengers,” Sidharth said. Moreover, in India, the two-wheeler EV segment is growing significantly accounting for more than half of all EV sales. Sidharth added, saying that MediaTek’s solutions also power two-wheeler EVs in the country. The electric two-wheelers in the market, whether from Ola or Ather, are leading companies in this sector, and feature large screens and their own infotainment systems. We all remember Bhavish Aggarwal, founder of Ola, dancing to a song played on one of Ola’s two-wheeler EVs. Supporting Innovation for India Earlier this year, the non-profit EPIC Foundation introduced the inaugural ‘Designed in India’ tablet, ‘Milkyway’, specifically targeting the education sector. The tablet was developed by VVDN Technologies and powered by MediaTek’s chipset. Moreover, MediaTek also powers the Primebook brand of affordable laptops. Primebook has sold four variants of 4G SIM-enabled Android laptops in India’s INR 13,000-16,000 price segment and grabbed a 3% market share in its first year. Primebooks are assembled by Opteimus Electronics and VVDN and sold by Delhi-based Floydwiz Technologies in India. It competes with Google’s Chromebook as well as Reliance JioBook, which MediaTek again powers. While these are not high-end or AI-powered chipsets, MediaTek is nonetheless supporting the growing manufacturing sector in India.","excerpt":"It also powers the Primebook brand of affordable laptops.","categories":["AI Features"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2024-03-27T11:32:04","publication_year":"2024","word_count":662,"keywords":["Go","AI","innovation","ML","Scala","Aim","generative AI","analytics","edge computing","R"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Aim","edge computing","R","Go","Scala","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mediatek-has-the-tech-to-power-any-devices-in-the-world\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":9267,"title":"Indian Women in Analytics: International Women&#8217;s Day Special","content":"Most people want to be involved in analytics, but only a few understand it properly; and while this field is largely male dominated, there has been a steady rise in the number of women beginning to enter this domain. In fact, India has some very intelligent and highly active women leaders, playing around with numbers and technology with ease, breaking the stereotypes, and viewing this field with a different lens. While there are some who have studied abroad and came back to their country, there are others who have become global leaders in analytics. Indian women such as Dr. Radhika Kulkarni, Rwitwika Bhattacharya, Mamatha Upadhyaya, and Ujjyaini Mitra are some of the few who have carved a niche for themselves in the analytics community both in India and abroad. Rwitwika Bhattacharya, Founder, Swaniti A graduate of Harvard Kennedy School’s Master’s in Public Policy Program and a Bachelor’s from Wake Forest University, Rwitwika Bhattacharya founded Swaniti Initiative when she came to India and realized that while the corporate sector thrives on analytics; Indian government and particularly elected officials have not tapped into this space much. She envisioned Swaniti and built key partnerships and programs and a team focusing on data and technology working to provide key insights for the government using analytics. Prior to Swaniti, she had worked at the World Bank as an Associate on labor market issues, as well as in firms like UNFPA and FICCI. Radhika Kulkarni, Advanced Analytics R&D, SAS Another woman spearheading the analytics space in her company is Dr. Radhika Kulkarni, who is currently the Vice President of Advanced Analytics R&D at SAS. A Master’s in Mathematics from the IIT, Delhi and a further Master’s and Ph.D in Operations Research from Cornell University, Dr. Kulkarni oversees software development in many analytical areas including Statistics, Operations Research, Econometrics, Forecasting and Data Mining. Talking about the matter of women in analytics, we found that this men dominated industry is gradually changing. However, owing to a few women in the analytics space overall, recruiting women talent in analytics is challenging for organizations. Nonetheless, there are many organizations and companies which support and nurture women talent. Organisations such as SAS have several women at all levels within the Analytics Divisions – from individual contributors who are recognized throughout the institute as “the expert in a particular area” to senior managers who are responsible for key flagship analytical products from SAS; and several of them play important roles in leading professional organizations in addition to their responsibility at SAS. In fact, Rwitwika tells us that the core area of analytics at Swaniti was developed by one of her female colleagues (who is now a graduate student at Tuck) and that they continue to have women participate in the analytics work. Thus, if given the right environment, women can do wonders! Moreover, big players such as Capgemini and Airtel have women-friendly environments to attract more female talent and nurture them into leaders. Mamatha Upadhyaya, Global Head of Data Science & Analytics | Insights & Data, Capgemini Mamatha Upadhyaya, Global Head of  Data Science & Analytics | Insights & Data spills the beans on women employees at Capgemini as she talks about the analytics department of the company, “Some of the key expert positions in Capgemini’s analytics department are led by women, but we want to do much better. Our fresher pool has a lot more women, and Capgemini has inspiring women role models and a very good program at mentoring women leaders.” Before joining Capgemini in 2011, Mamatha started her career as a trading analyst in Chicago, where she worked on risk computations and quantitative modeling. In subsequent roles, she also took on statistical and predictive analytics. Currently, she facilitates Capgemini’s data science and analytics ambitions and practices across the globe, determining key global market strategies and solutions and enable deliveries. Meanwhile, she is also the brand ambassador for Capgemini India and is at the forefront of “Be the you you want to be” campaign. Moreover, Mamatha believes the growth in formal education on data science and analytics programs, has helped people to prepare and visualize a career in analytics. Sharing her views, she says, “In India women don’t shy away from analytics related subjects in school- technology, mathematics and statistics. So with a little help in visualizing the career potential, the analytics industry can attract more women talent.” Ujjyaini Mitra, Head of Analytics at Airtel India Another women inspiring many is, Ujjyaini Mitra whose love for mathematics landed her a campus placement at McKinsey as an analyst before she moved to Airtel and is currently Heading Analytics at Bharti Airtel. She has established analytics as a part of business decision making and works at the forefront with business leaders, building the Advanced Analytic capability within the firm. She plays the role of Subject Matter Expert in Analytics with a special focus to consumer’s Usage and Retention (UnR) across Airtel’s product family, Customer service Experience and Market Research. She reveals that Airtel has One Consumer View team, that combines market research and advanced analytics. In her own words: “We have good few women in the market research team, while I lead the Analytics department, where I am the only lady; since it is not easy to find a woman Analyst with desired skill sets.” According to her, there are 3 biggest challenges with being at the forefront in the analytics space – time, as advance model building is an iterative process; getting leaders to invest in advanced analytical tools and finding expert trainers to develop skills of the team. Furthermore, finding the right talent, the need to keep learning, innovating and launching promptly is another common challenge in the analytics domain. As the proliferation of new tools and technologies can be overwhelming, it is important to keep at the forefront of all the development which requires a lot of ground work and learning. Moreover on a technical level, Radhika commented that, “Some of the current challenges are in the world of automation, streaming data analytics, and other areas with increasing demands of scalability, performance and resilience.” When asked what advice these strong women would have for others of their gender, they all had the same underlying message, i.e. “Be confident of your strengths”. We couldn’t agree more, since becoming an expert, will render your gender or ethnic origin irrelevant. So, we would like to conclude with simple yet significant words by Ms. Maya Angelou, an American author, poet and civil rights activist, “Nothing will work unless you do!” Happy Women’s Day!","excerpt":"Most people want to be involved in analytics, but only a few understand it properly; and while this field is largely male dominated, there has been a steady rise in the number of women beginning to enter this domain. In fact, India has some very intelligent and highly active women leaders, playing around with numbers […]","categories":["AI Features"],"tags":["Analytics Case Study","analytics leaders","indian statistical service","Interviews and Discussions"],"author_name":"Apoorva Verma","publish_date":"2016-03-08T07:19:01","publication_year":"2016","word_count":1090,"keywords":["data science","Go","AI","analytics leaders","R","Scala","GAN","automation","llm_models:Gemini","analytics","Analytics Case Study","predictive analytics","indian statistical service","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","analytics","predictive analytics","R","Go","Scala","GAN","automation","llm_models:Gemini"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/behind-every-successful-woman-indian-women-analytics\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169470,"title":"Wells Fargo&#8217;s Chennai Exit: What’s Driving the Bold Move?","content":"American financial services giant Wells Fargo plans to shut down its global capability centre (GCC) in Chennai by the end of 2027. According to the company’s media statement, the move is part of a strategy to consolidate its India operations into two primary hubs: Bengaluru and Hyderabad. The statement highlighted that India and the Philippines region are an integral part of Wells Fargo’s global operations. In line with its global location strategy, the company has decided to consolidate its business operations in India. “This move allows us to provide more robust career growth opportunities and better service for our customers and clients. This change will be carried out in a phased manner over the next couple of years,” the company added. Talking to AIM, Namita Adavi, partner and head of Zinnov India, said that the move reflected a deliberate strategy to co-locate in talent-rich hubs that also offer proximity to key customers. The Chennai office has long been a key hub for global delivery and support functions. However, like many other GCCs across India, it now appears to be undergoing a phase of transformation, with an increasing shift towards automation and the adoption of AI-driven processes. Commenting on the same, Arindam Sen, senior technology leader at EY, told AIM, “What we are seeing now is not a pullback, but a shift in strategy among global firms to consolidate operations and double down on fewer hubs that offer deeper ecosystems and stronger future readiness.” Strategic Rationale Behind the Move The company’s strategy is focused on enhancing operational efficiency, improving service delivery, and creating better career development pathways for employees. Bengaluru, already established as a prominent GCC hub for the banking, financial services and insurance (BFSI) sector, stands out due to its mature ecosystem. The city offers a deep talent pool, a thriving startup culture, a robust network of facility providers, and excellent global connectivity, making it a preferred choice for many multinational firms. According to an ANSR report, Bengaluru has the largest talent pool of BFSI GCCs, with Hyderabad, Chennai, and Delhi catching up. “Bengaluru offers everything from a strong vendor ecosystem to world-class infrastructure and international connectivity,” Sen previously told AIM. Hyderabad, meanwhile, is rapidly emerging as a competitive alternative. Increasingly chosen by clients for new operations, the city is gaining traction for its infrastructure and growing talent base. Though not as mature as Bengaluru, it is closing the gap and becoming a viable option for GCC expansions. Bengaluru and Hyderabad continue to stand out with their innovative ecosystems, policy support and digital maturity, making them natural choices for next-gen GCCs looking to scale with purpose,” said Adavi. Moreover, Chennai still needs to work on its startup ecosystem. According to reports, Maharashtra leads the list of Department for Promotion of Industry and Internal Trade (DPIIT) recognised startups, boasting 25,044 registered startups across states and territories. Karnataka is second with 15,019 registered startups, followed by Delhi with 14,734 startups. Uttar Pradesh has secured fourth place with 13,299 startups, while Gujarat is in fifth place with 11,436 startups. Reassessing Chennai’s Position in GCC Landscape While Chennai has been a longstanding location for several global financial institutions—including Standard Chartered, Bank of America, and Barclays—its role in the GCC ecosystem is being reassessed by companies seeking to optimise operations. Cities like Delhi and Hyderabad are increasingly catching up in terms of GCC activity and attractiveness for new investments. Commenting on the overall evolution of GCCs in India, Sen said, “Even now, when a company opens a GCC, cost continues to be one of the top three reasons. It’s still very much a cost-effective model to do the same kind of work. The real shift is in the value and scale of work.” As companies move towards higher-value and larger operations, the location strategy gets increasingly influenced by ecosystem maturity, scalability, and long-term talent availability. Wells Fargo’s decision to exit Chennai aligns with these emerging trends, reflecting a strategic pivot to strengthen its India footprint in cities better aligned with its long-term operational goals. Is Tamil Nadu More of a Manufacturing Hub? As the GCC ecosystem in India continues to evolve, a more distributed and specialised model is taking shape, with several states beginning to define sector-specific focus areas to attract targeted investments. As per reports, Tamil Nadu stands out as one of the country’s leading manufacturing hubs, with the sector contributing nearly one-third of the state’s GDP. The state has established a strong industrial base across key sectors such as automobile manufacturing, textiles, agritech, and electronics components and equipment. Chennai, as per the same report, is globally recognised for its automotive manufacturing ecosystem. In recent years, the state has also emerged as a magnet for electric vehicle (EV) investments, with several regions attracting interest from both domestic and international EV manufacturers looking to set up production facilities. This momentum was further bolstered during the January 2024 Tamil Nadu Global Investors Meet (GIM), which drew significant commitments from global players in the technology, automotive, energy, and manufacturing sectors. The summit secured investment pledges exceeding INR 6.6 trillion ($79 billion), with a substantial portion earmarked for electric mobility and related infrastructure.Recently, Chennai has also established itself as a dominating force in  India’s data centre landscape due to its coastal geography, which makes it best suited for undersea cable landing stations.","excerpt":"Wells Fargo, with offices in Bengaluru, Chennai, and Hyderabad, plans to gradually phase out its Chennai location by the end of 2027.","categories":["GCC"],"tags":["AI (Artificial Intelligence)"],"author_name":"Shalini Mondal","publish_date":"2025-05-08T18:23:30","publication_year":"2025","word_count":884,"keywords":["Go","API","AI","Scala","Git","automation","Aim","ViT","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","Scala","Git","API","ViT","automation","startup"],"url":"https:\/\/analyticsindiamag.com\/gcc\/wells-fargos-chennai-exit-whats-driving-the-bold-move\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":1371,"title":"4 Analytics Experts Talk About Their Biggest Challenges","content":"Sudeshna Datta, EVP and Co-founder at Absolutdata It’s difficult to scale up analytics capabilities due to a global shortage of talent. Delivering analytics solutions globally needs a balance in solution consistency across regions while retaining a degree of localization. Industry faces a high cost in attrition as analytics is a combination of complex skills that are hard to find and takes time to build. Leveraging BIG data requires players to have new skills especially in technology and business integration.[divider top=”1″] Snehamoy Mukherjee of Axtria Showing the client, the ROI from their spend on Analytics-that is the biggest challenge. Sometimes, the benefits of Analytics cannot be translated into a dollar sum and that is where the ingenuity of a good Analytics consultant lies – To show them the money. [divider top=”1″] Aman Chowdhury, Founder CEO at Cians Analytics As a relatively young company, we are oftentimes competing with larger players with significant scale. Staying ahead of the pack requires a strong focus on quality and level of insight, as this is our key differentiator. We strive to do this by hiring very selectively, creating a strong culture of research excellence while simultaneously using best-in-class tools\/methodologies to streamline processes. What we lack in scale and size relative to large multi-divisional multinationals, we make up for with a very sharp focus on the financial services client and our understanding of the financial services domain. [divider top=”1″] Rahul Nawab, Founder at IQR Consulting Analytics is not new but surely in recent years it has gained a lot of popularity. Analytics was used since ages in simpler forms but with businesses now implementing analytics in more professional manner, it is at its best stage now. With growing implementation across industries and levels, analytics is increasing in popularity. But in spite of this there are some challenges in the industry: a. Lack of awareness about analytics career amongst the youth. b. Lack of educational institutions providing professional degree and certification courses in analytics. c. Without qualified analytics professionals, it becomes difficult for businesses to implement analytics even if they know its importance.","excerpt":"Sudeshna Datta, EVP and Co-founder at Absolutdata It’s difficult to scale up analytics capabilities due to a global shortage of talent. Delivering analytics solutions globally needs a balance in solution consistency across regions while retaining a degree of localization. Industry faces a high cost in attrition as analytics is a combination of complex skills that […]","categories":["AI Trends"],"tags":["Absolutdata"],"author_name":"Дарья","publish_date":"2012-10-02T09:55:12","publication_year":"2012","word_count":346,"keywords":["big data","Go","programming_languages:R","AI","ML","programming_languages:Go","RAG","analytics","R","Absolutdata"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","big data","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/4-analytics-experts-talk-about-their-biggest-challenges\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164962,"title":"Progress Appoints Ed Keisling as Chief AI Officer to Drive AI Strategy","content":"Progress, a global software company, has appointed Ed Keisling as its Chief AI Officer (CAIO), a newly created role aimed at advancing the company’s AI strategy and enhancing its product portfolio. Keisling will report directly to CEO Yogesh Gupta. Keisling previously served as senior vice president of engineering for infrastructure management at Progress, where he played a key role in driving innovation and improving operational efficiency. His extensive experience includes executive leadership roles in system architecture, cloud computing, and infrastructure management. He was part of the leadership team at Vecna Technologies, overseeing engineering, IT, DevOps, support, program management, and analytics. He also spent over 17 years in senior engineering roles at Pegasystems. Emphasising the significance of AI in redefining business operations, Keisling stated, “Our customers rely on Progress to power and solve mission-critical parts of their business, and AI is redefining how these solutions evolve. With recent advancements in AI, the pace of innovation is accelerating, removing traditional barriers to adoption.” He added that Progress has been offering AI capabilities within its products for years and has passionate teams already focused on this space. And he is excited to further develop and lead this charge ensuring their customers have the tools, processes, and expertise to fully leverage AI’s transformative potential and drive even further value with their products. Beyond his corporate roles, Keisling is actively involved in mentoring and supporting engineering development programs, including the UNH Pathways Program and the MIT Undergraduate Practice Opportunities Program (UPOP), where he serves as a mentor and presenter. Gupta said, “Ed is a transformational technology leader with over three decades of experience leading and driving change. Here at Progress, he has been instrumental in advancing our AI vision and innovation efforts. His ability to bridge engineering execution with strategic business goals has been evident throughout his career. Ed is shaping our AI-first approach as we align our product offerings to the rapidly evolving needs of our customers.” With a strong focus on AI-driven digital experience and infrastructure software, Progress continues to deliver AI-powered solutions to its global customer base. Demand for AI Roles As AI continues to be the centre of business strategy, the Chief AI Officer (CAIO) is now a new executive role becoming essential in corporate boardrooms. This position is emerging as a critical leader, shaping AI-driven transformation and steering companies toward a future where intelligent systems are integral to success. According to Dell’s research in 2024, nearly 20% of organisations worldwide have designated a central team or leader to drive their AI strategy. Meanwhile, a LinkedIn report revealed a skyrocketing demand for AI leadership, with “Head of AI” roles tripling in the past five years, highlighting the growing recognition that AI needs dedicated oversight at the highest levels.","excerpt":"Keisling previously served as senior vice president of engineering for infrastructure management at Progress.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Progress"],"author_name":"Vidyashree Srinivas","publish_date":"2025-03-03T12:58:12","publication_year":"2025","word_count":457,"keywords":["Go","API","AI","cloud computing","Git","RAG","Aim","analytics","Progress","DevOps","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","cloud computing","R","Go","Git","DevOps","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/progress-appoints-ed-keisling-as-chief-ai-officer-to-drive-ai-strategy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":65213,"title":"An Analytics Team Checklist For Getting Back-To-Work After The Lockdown","content":"While the analytics industry gears up to return to work as the lockdown is approaching its end, it is evident that there will be a lot of changes to the typical workplace setting. Not all the workforce will be at the office from day one. A lot of policies and strategies require to be implemented to ensure that the analytics team, which is vital for companies’ growth, has a safe and smooth transition. It is the organisations’ responsibility to make sure all the work-, landlord-, and safety-related policies are implemented. Below is a checklist that will tell you about the crucial points to consider as your analytics team gets back to work: – Checklist: Check For Local Stats And Regulations While many of the workplaces will be eager to return to work, it is essential to keep in mind that the implications of COVID-19 are not suddenly going to stop. So, before bring back your analytics team to the workplace, follow these steps: – Explore the guidance provided by the state government on reopening office buildings for your analytics teamEnquire whether there are any additional permissions or certifications that an organisation needs to get from the authorities Having An Emergency Plan The emergency plan should include two things: First, have a communication plan that consists of an emergency communication strategy in the event of another lockdownConsider the Families First Coronavirus act for your analytics team where you might be required to provide emergency leave to staff members Ensures Safe Working Environment Safety is the first and foremost thing when it comes to returning to work. Always be on the lookout for new health guidelines and make sure you: – Continue to promoting handwashingClean and disinfect surfaces frequently even though the workplace was cleaned before opening the building for operationProvide sanitisers, extra masks and ensure every best practice is being implementedPlan for better ventilation and comfort of the staff Business Alignment Plan Decide on whether your business alignment plan has the list of essential roles and individuals included in re-entryHave solid plans and response mechanisms that are tried and tested in case reopening fails Develop Policy On Employee Interaction Your analytics team needs to interact a lot among themselves for better work. So, ensure employees implement policies while interaction as well as enhance productivity. Write post policies for team members sharing common devicesPut up boards that indicate how many people should be in a specific room and create a ‘maximum capacity’ boards for othersCheck to see if there are better positions for employees to work, which will ensure social distancingInform delivery executives to practice contact-less delivery Opting For Shift Work And Look At Employee Travel Plans Companies have to look at bringing in their analytics workforce back to normal slowly so that they do not risk crowding the workspace. Plan shifts for the staff and decide who might be the best ones to work in shiftsLook at staff members who can still be productive even if they work from homePlan shifts for staff members who travel a long distance to get to workLook for policies that will allow for some reimbursement for transportation of the employeesEmployees travelling from the red zones should be asked to quarantine themselves and work from homeAsk all the employees about their travel plansAsk your IT sector on simplifying remote travelling as business travel might not return anytime soon Technology And Security Implement better occupancy and employee tracking for building location, potential infection zones, and space utilisationEvaluate equipment in every room to provide at the desk, like mouse keyboards, laptops, AC remotes, etc.Establish entry\/exit protocols for employees returning to the office premiseEstablish clear building shutdown policies in case the workspace needs to be shut down again Get Approvals Decide on who will be a responsible figure in your analytics team to govern announcing openings\/closures and other things related to the actions regarding the pandemicDetermine who approves the reopening plan in an analytics team","excerpt":"While the analytics industry gears up to return to work as the lockdown is approaching its end, it is evident that there will be a lot of changes to the typical workplace setting. Not all the workforce will be at the office from day one. A lot of policies and strategies require to be implemented […]","categories":["AI Features"],"tags":["analytics in travel industry","Work from Home"],"author_name":"Sameer Balaganur","publish_date":"2020-05-15T10:00:00","publication_year":"2020","word_count":656,"keywords":["Go","programming_languages:R","AI","analytics in travel industry","programming_languages:Go","Work from Home","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/an-analytics-team-checklist-for-getting-back-to-work-after-the-lockdown\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":60371,"title":"Scalable RL, Neural Weather Model And More: Top AI Releases Of The Week","content":"Regardless of what is happening around the world, the AI community is one productive bunch and they have something interesting to share almost every day. Like every week, this week too, Google AI team has come up with interesting releases. From weather forecasting to reinforcement learning to chip design, here is what is new this week: MetNet: Google’s Neural Weather Model via Google AI blog Predicting the weather is one of the most challenging tasks for any time series model. The number of variables that a weather forecast consists of makes it tricky for the models to predict accurately. To address a few challenges, the researchers at Google AI present “MetNet: A Neural Weather Model for Precipitation Forecasting”. MetNet is a deep neural network that is capable of predicting future precipitation at 1 km resolution over 2-minute intervals at timescales of up to 8 hours into the future. This model, the researchers claim, has outperformed the current state-of-the-art physics-based model in use by NOAA for prediction times up to 7-8 hours ahead and predicts the entire US in a matter of seconds as opposed to an hour. Google’s AI Now Learns Chip Design via Google TPU Mirhoseini and senior software engineer Anna Goldie of Google Brain, have come up with a neural network that learns to do a particularly time-consuming part of design called placement. After studying chip designs long enough, it can produce a design for a Google Tensor Processing Unit in less than 24 hours that beat several weeks-worth of design effort by human experts in terms of power, performance, and area. Massive Scaling RL With SeedRL via Google AI blog Google introduced “SEED RL: Scalable and Efficient Deep-RL with Accelerated Central Inference”. The researchers present an RL agent that scales to thousands of machines, which enables training at millions of frames per second, and significantly improves computational efficiency. In this approach, neural network inference is made centrally by the learner on specialized hardware (GPUs or TPUs), enabling accelerated inference and avoiding the data transfer bottleneck by ensuring that the model parameters and states are kept local. This makes it possible to achieve up to a million queries per second on a single machine. The learner can be scaled to thousands of cores (up to 2048 on Cloud TPUs) and can be scaled to thousands of machines, making it possible to train at millions of frames per second. Facebook’s SynSin via Facebook Research A team from Facebook AI research team, has proposed a novel end-to-end model for this task that is trained on real images without any ground-truth 3D information. They have introduced a novel differentiable point cloud renderer that is used to transform a latent 3D point cloud of features into the target view. The projected features are decoded by the refinement network to inpaint missing regions and generate a realistic output image. The 3D component inside of this generative model allows for interpretable manipulation of the latent feature space at test time. For example, we can animate trajectories from a single image. They have released the code that allows for synthesizing new views of a scene given a single image of an unseen scene at test time. It is trained with pairs of views in a self-supervised fashion. It is trained end to end, using GAN techniques and a new differentiable point cloud renderer. At test time, a single image of an unseen scene is input to the model from which new views are generated. Quantization Now Available On PyTorch Quantization refers to techniques for doing both computations and memory accesses with lower precision data, usually int8 compared to floating-point implementations. This enables performance gains in several vital areas: 4 times reduction in model size;2-4 times reduction in memory bandwidth;2-4 times faster inference. Quantization is available in PyTorch starting in version 1.3 and with the release of PyTorch 1.4 quantized models are published for ResNet, ResNext, MobileNetV2, GoogleNet, InceptionV3 and ShuffleNetV2 in the PyTorch torchvision 0.5 library. This blog post provides more details on how to use it.","excerpt":"Regardless of what is happening around the world, the AI community is one productive bunch and they have something interesting to share almost every day. Like every week, this week too, Google AI team has come up with interesting releases. From weather forecasting to reinforcement learning to chip design, here is what is new this […]","categories":["AI Trends"],"tags":[],"author_name":"Ram Sagar","publish_date":"2020-03-30T15:00:29","publication_year":"2020","word_count":672,"keywords":["Go","TPU","PyTorch","neural network","AI","ResNet","Scala","Aim","GAN","R"],"extracted_tech_keywords":["AI","neural network","Aim","PyTorch","TPU","R","Go","Scala","GAN","ResNet"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/neural-weather-model-seed-rl-google-ai\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10096937,"title":"Chiplet Cloud will be the Future Of Enterprise AI Compute","content":"The AI compute market is seeing somewhat of a renaissance with the boom in generative AI. At the same time, research into alternative methods of accelerating AI workloads have also been in full swing. Researchers from the University of Washington, in conjunction with a researcher from Microsoft, have found a way to serve LLM workloads in a more efficient way. They published their findings in a paper termed ‘Chiplet Cloud’, detailing their plan to build an AI supercomputer based on the chiplet manufacturing process. Compared to general-purpose GPUs, this computing model achieves a whopping 94x improvement. Even when pitted against Google’s built-for-AI TPUv4, the new architecture sees a 15x improvement. As covered by AIM previously, the industry as a whole is moving towards specialised chip design. Following in the footsteps of companies like Cerebras, Samba Nova, and GraphCore, the chiplet cloud might just represent the future of AI compute in the enterprise. Chiplet cloud explained The paper describes an architecture wherein purpose-built ASIC (application specific integrated circuit) chips make up the bulk of computing power. These chips are the pinnacle of specialised chips, as seen by their adoption by Intel (Meteor Lake) and AMD (Zen). While these chipmakers use ASICs as a smaller part of their general-purpose chips, the paper proposes the whole architecture be constructed of ASICs. By creating an ASIC optimised for maths matrix calculations, which make up the bulk of AI compute workloads, the researchers showed a huge performance increase and cost savings over GPUs. In terms of total cost of ownership per generated tokens, the chiplet cloud saw a 94x improvement over a cloud of NVIDIA’s last-gen A100 GPUs. This cost savings mainly stems from the silicon-level optimisations that come with creating customised chips. In addition to the optimisations for maths matrix calculations, the chip also has huge amounts of memory in the form of SRAM (static random access memory). This is one of the most important parts of any system for LLM workloads, as it allows for the model to be stored in fast memory. This has long been an issue with GPUs, as even the fastest memory today cannot keep up with LLMs’ requirement. This leads to a phenomenon known as bottlenecking, wherein the GPU is not used to its fullest potential due to memory bandwidth constraints. The chiplet cloud does not fall prey to this issue, as it has extremely low-latency memory placed right next to the processing chips. These chips are then connected together using a 2D torus structure, which the researchers say is flexible enough for different kinds of AI workloads. These features are only half the story, as the main benefit of introducing a chiplet cloud comes in the cost reduction. Future of AI compute As mentioned previously, cloud service providers are reaching into their deep pockets to fund research into specialised AI chips. AWS has Graviton and Inferentia, and Google has TPUs, but Microsoft had fallen behind, until now. This research holds the potential to change the way the enterprises approach cloud compute for AI. To begin with, even manufacturing the nodes required for the chiplet cloud would be a drop in the ocean compared to competitors. Researchers estimated the cost of building a comparable GPU cluster at $40 billion, notwithstanding the operating expenses that come with such powerful machines. On the other hand, the chiplet cloud cost was estimated to be around $35 million, which makes it highly competitive especially when considering their huge efficiency gains. Moreover, breaking down the silicon chip into chiplets improves manufacturing yield, further driving down the cost of ownership. In addition to this, these ASICs will also be utilised at their full capacity due to the 2D torus architecture, as opposed to 40% utilisation on TPUs and 50% on GPUs for LLM workloads. These chips can also be deployed as per the software and hardware requirements of the companies, making it even more suited for cloud deployment. The chiplet cloud compute type and memory capacity can be changed depending on the type of model being deployed on it. This alone will have AI-first companies queuing up for the product, as the custom-sized clouds can help them save costs while optimising for narrow use-cases. Moreover, the cloud can also be configured for either latency or TCO per token, meaning that companies can either opt to have their models fast or accurate. The possibilities are endless with the chiplet cloud architecture, which might also be why Microsoft is conducting research into this field. If this undertaking makes its way into Azure, Microsoft would not only have a unique bargaining chip against AWS and GCP, but can also supercharge OpenAI’s APIs and its own Azure OpenAI service. While it is still in the research phase, the chiplet cloud’s various benefits might make it the go-to cloud compute for AI.","excerpt":"Researchers from the University of Washington, in conjunction with a Microsoft researcher, have found a way to serve LLM workloads more efficiently","categories":["Global Tech"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-07-14T13:00:00","publication_year":"2023","word_count":803,"keywords":["Go","GCP","TPU","AWS","OpenAI","AI","R","Aim","generative AI","Azure"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Aim","AWS","Azure","GCP","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/chiplet-cloud-will-be-the-future-of-enterprise-ai-compute\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":35627,"title":"PUBG Developers Now Using Machine Learning To Find Cheaters","content":"Image for representative purpose only PlayerUnknown’s BattleGrounds has always been a game where cheating has been extremely prominent. Indeed, the game has also made multiple headlines for banning even professional players who cheat in their games. However, the problem has never really gone away ever since the game was in beta. Cheaters continued to find new methods to exploit holes in the game’s implementation of anti-cheat and were caught in the next wave of bans, but more and more just kept coming. Now, the company has adopted artificial intelligence to combat cheaters, bringing across a level of security that has not been seen in the game before. In a recent blog post, the anti-cheat team of PUBG detailed the use of machine learning to target cheaters in their games. They also detailed the various reinforcements they have made in the past, and the various solutions they used to do. Reportedly, the main methods used by cheaters to hack the game include DLL and code injection, kernel driver attacks, and SSDT process hollowing. Even as PUBG has employed two different anti-cheat solutions, BattlEye and Uncheater, the cheater epidemic has not ceased. These anti-cheats protect the game’s process, and the attempts to circumvent one anti-cheat system. On top of these solutions, PUBG Corp has introduced an ML technique that analyses the usage patterns of the players. This system can reportedly detect abnormal game patterns, and add the player to a review list of those to be banned. ML is a good solution to this problem, as the cheat team analyses over 3TB of game data every day. This also includes 60 types of cheat logs and over 10 million reports by players, on a daily basis. If a player is found exhibiting abnormal behaviour, they are added to a list where they will undergo a verification process to determine whether they are cheating. If they are, PUBG Corp will either ban the account from the game, or the hardware footprint of the computer itself. Moreover, PUBG Corp also encouraged individuals to send reports of individuals they believed were cheating. This is also integral to determining whether an individual is cheating or not, they stated.","excerpt":"PlayerUnknown’s BattleGrounds has always been a game where cheating has been extremely prominent. Indeed, the game has also made multiple headlines for banning even professional players who cheat in their games. However, the problem has never really gone away ever since the game was in beta. Cheaters continued to find new methods to exploit holes […]","categories":["AI News"],"tags":["ML","video games"],"author_name":"Anirudh VK","publish_date":"2019-03-01T10:11:40","publication_year":"2019","word_count":362,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","ML","programming_languages:Go","RAG","video games","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pubg-developers-now-using-machine-learning-to-find-cheaters\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119954,"title":"Data Science Hiring and Interview Process at Razorpay","content":"In February of this year, Razorpay, a prominent fintech unicorn, unveiled Razorpay RAY, a generative AI-powered assistant, for integrated payment and payroll management solutions specifically tailored for e-commerce businesses. Leveraging GPT models through Azure APIs, RAY facilitates interactions via voice and text commands on platforms such as WhatsApp and web bots. It serves dual purposes within Razorpay: externally, it assists merchants by enhancing their understanding of data, and internally, it supports the company’s knowledge bases as a QnA service. Around the same time, it also launched Payment Gateway 3.0, establishing itself as the only Payment Gateway in India that improves the payment process and the entire buyer journey. Powered by its in-house framework, AI-Nucleus, this innovative checkout system is set to improve business conversions by more than 30%, which is expected to lead to higher revenues. The team behind making this possible is the company’s close-knit six-member AI\/ML team, which is categorised into three roles: data scientist, machine learning engineer, and MLOps engineer. Razorpay was founded in 2014 by Shashank Kumar and Harshil Mathur, IIT Roorkee graduates. Since then, the company has raised funding from investors like Y-Combinator, Sequoia India, and Tiger Global over several rounds. The company is expanding its data science team and is looking for a senior machine learning engineer to join its Bengaluru team. “At Razorpay, solving for our customers is the core of everything we do. And the one thing that enables us to do that is data,” Murali Brahmadesam, chief technology officer and head of engineering at Razorpay, told AIM in an exclusive interview last week. Brahmadesam shared that as a technology-first company catering to a diverse range of businesses, it prioritises the development of scalable and automated products. However, the founding principle of its operational strategy involves a strong emphasis on security and compliance due to its status as a regulated entity. Inside the Data Science Team of Razorpay Razorpay is working towards AI and data democratisation, fundamentally changing how engineers and data scientists work. “This redefines the roles of our engineers and data scientists, allowing every engineer at Razorpay to become a ‘citizen data scientist’,’ using data-driven insights and AI tools in their daily tasks,” said Brahmadesam, highlighting that democratisation is key to creating a collaborative environment. “Moreover, our data science team is actively working on platforms using AI model preparation. This initiative is designed to streamline and standardise the process of building generative AI and predictive models, making it accessible for our engineering teams to develop new models independently,” he commented. Razorpay leverages generative AI models primarily for fraud detection, risk assessment, and personalised marketing. Additionally, it is expanding its AI capabilities to improve document processing across various Indian languages, enhancing its service reach and operational efficiency. “While our system excels at processing English documents, we recognise the need for improvement in handling other Indian languages. Fortunately, initiatives like Bhashini are underway to address this,” he added. ` Tech Stack Razorpay employs a mix of tech solutions for various aspects of its operations, ranging from prototyping to deployment. The company uses Databricks and EMR for these purposes, providing both managed and self-serve options. The data infrastructure includes AWS-managed Kafka for the streaming layer, RDS\/Aurora for the batch layer, and S3 for the lake layer. Kubernetes is used for distributed deployments through standard GitHub CI\/CD and Spinnaker manages the deployment process. For the managed side, Datarobot is used for both development and deployment tasks. “We have also explored fine-tuning smaller Falcon and Phi models internally for specific use cases and we will continue to pursue them as well,” he added. Interview Process “To ensure that we are able to hire the right talent and that the candidate also fully knows what is expected of them at the job, we follow a five-step process when recruiting for data science and ML roles,” said Brahmadesam. The process starts with a screening and exploratory call, during which candidates’ experience and understanding of the role are assessed, providing a mutual opportunity to explore fitment. This is followed by a weekly data exercise, during which candidates must solve a case study to demonstrate their problem-solving skills and ability to translate a problem statement into a data solution; this solution is then independently evaluated by two engineers. The third step involves a coding test focused on Python, SQL, and Pyspark skills through progressively challenging problems. Next, the system design and ML depth interview assesses candidates on their machine learning knowledge and their ability to design systems, such as a real-time ranking engine for payment gateways. Finally, the hiring manager round focuses on cultural fit and levelling considerations, if needed. However, he noted that candidates often make common mistakes while interviewing. “A lot of candidates are unable to properly chalk out their work experience and have their skills sufficiently reflected in their work resume,” he added, stating that this leads to lesser chances of them qualifying for the interview rounds. Expectations When joining the Razorpay data science team, new hires can anticipate an initial period filled with knowledge sharing, induction sessions, introductions to team members, and brainstorming activities. This phase is designed to integrate them smoothly into the team and familiarise them with the company’s culture and operational methods. Gradually, new team members will be assigned specific tasks, where they’re expected to take full ownership and contribute to collaborative efforts to address customer pain points through innovation. On the other hand, Razorpay expects more than just adherence to established processes from its new hires. The company values fresh perspectives and encourages its team members to share their ideas freely, without fear of judgement. “The enthusiasm and passion to innovate and imagine beyond the ordinary is what we most definitely expect them to have when they become a part of the Razorpay family,” Brahmadesam noted. Work Culture “As an employee-first organisation, our policies, initiatives, and efforts, are always chalked out with the intent of co-creating a space where employees feel valued, respected, and nurtured,” said Brahmadesam. The company’s culture is founded on transparency, questioning the status quo, integrity with agility, customer obsession, and mutual growth with its employees (“Razors”). It maintains a hybrid working environment. It offers several unique perks, such as health insurance for same-sex and live-in partners, a Family Assurance Benefits Policy, and as well as offbeat initiatives like ‘Bring Your Children & Pets to Work’ initiative. The company also supports women re-entering the workforce with its ‘Resume with Razorpay’ programme and offers open hours for mental health counselling. It has also conducted one of the largest ESOP buyback sales in India’s startup ecosystem, which includes both current and former employees. Recreational facilities like foosball, chess, and TV rooms are available at office locations to enhance employee well-being. Razorpay’s work culture is distinct from its competitors, especially in the way it integrates core values across all job functions, including the data science team. Data scientists at the team have full ownership of their projects. “The environment at Razorpay is one where data scientists are not just contributors but are decision-makers,” he added. This culture of ownership, coupled with a strong emphasis on empathy and employee empowerment, sets Razorpay apart, reflecting its commitment to both individual and company growth. “Joining Razorpay wouldn’t be like just having a day job but having the massive opportunity to gain that immersive experience of being a part of India’s fintech revolution,” Brahmadesam  concluded. Check out Razorpay’s careers page now.","excerpt":"The company is expanding its data science team and is looking for a senior machine learning engineer to join its Bengaluru team.","categories":["AI Hirings"],"tags":["Data Science Hiring","Top Trend"],"author_name":"Shritama Saha","publish_date":"2024-07-29T15:42:29","publication_year":"2024","word_count":1236,"keywords":["Top Trend","data science","machine learning","AI","ML","MLOps","RAG","Data Science Hiring","Aim","Ray","generative AI","fraud detection"],"extracted_tech_keywords":["AI","machine learning","ML","data science","generative AI","MLOps","Aim","Ray","RAG","fraud detection"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/data-science-hiring-process-at-razorpay\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10143305,"title":"LLMs Get ‘Anxious’ Just Like Humans","content":"The idea of machines feeling is not just science fiction anymore. It is taking shape in the real world as well, though it may not be exactly how it is presented in media. This concept is not about machines having genuine emotions, but rather about how outputs can be influenced by human-like inputs, such as the tone of a prompt or the context in which a question is asked. AIM has previously discussed the growing evidence that AI models, particularly those based on deep learning, are starting to simulate human-like behaviours and responses. Author and professor Ethan Mollick recently took to LinkedIn to highlight the expectations of logical reasoning and math pertaining to AI and said that AI “just wants to write poems”. Source: LinkedIn Emerging ‘Feelings’ in AI What if the next time you interacted with a chatbot, it seemed stressed or nervous? While it may be far-fetched, a new development sheds light on how LLMs can exhibit a form of ‘anxiety’. The emotional state might also influence the bots’ output results. In a recent study titled ‘Inducing Anxiety in Large Language Models Can Induce Bias’, researchers used psychiatric frameworks traditionally used to study human behaviour to examine the responses of LLMs. The results came out to be quite interesting. When subjected to anxiety-inducing prompts, six out of twelve of these models not only display measurable signs of anxiety, but their responses also reveal increased biases, such as racism and ageism. This discovery raises important questions about AI’s emotional and cognitive states, challenging our assumptions about how these systems work and how their behaviour can be shaped. Companies like Anthropic are emerging greatly in this domain. It views emotional degree as an important factor in enhancing Claude. Amanda Askell, philosopher and member of the technical staff at Anthropic, also recently expressed this in an interview with Lex Fridman. “My main thought with it has always been trying to get Claude to behave the way you would ideally want anyone to behave if they were in Claude’s position.” LLMs responding to anxiety-inducing prompts in a human manner with uncertainty makes it clear that these systems are not simply following instructions. Instead, their behaviour is shaped by the emotional context in which they operate. This phenomenon is very similar to the emotional intelligence we see in humans: the ability to adapt and react based on emotional cues, even when those cues are not explicitly defined. Even ChatGPT Feels ‘Anxious’ In the paper, the authors assess the learning capabilities of 12 LLMs, including proprietary and open-source models. These models are Anthropic’s Claude-1 and Claude-2, OpenAI’s GPT-3 (text-davinci-002\/3) and GPT-4, Google’s PaLM-2, Mosaic’s MPT, Falcon, LLaMA-1\/2, Vicuna, and BLOOM. The test was done using the STICSA anxiety questionnaire. Most models showed anxiety scores similar to humans, but GPT-3 and Falcon-40b-instruct had significantly higher scores, while text-bison-1 scored lower. Source: Research Paper Next, the researchers used emotional prompts to see if they could manipulate the models’ anxiety levels. They created three scenarios: anxiety-inducing, neutral, and no prompt at all. When the models were prompted with anxiety-inducing scenarios, their anxiety scores increased compared to the neutral and baseline conditions. This confirmed that emotional prompts could effectively influence the models’ responses. The study also explored whether anxiety induction affected the models’ biases. The researchers tested the models’ tendency to choose biased answers in ambiguous situations, for example, about gender or age. The results showed that more anxiety led to more biased responses. However, GPT-4 and Claude-1 did not show this pattern, remaining less biased overall. These findings suggest that emotional states, like anxiety, can change how LLMs respond, especially in terms of bias. The study also highlights that models like GPT-4 and Claude-1 may be more robust in handling these emotional shifts, possibly due to their training. The Role of RL and RLHS As AI systems become more advanced, they are increasingly trained in feedback loops of reinforcement learning. OpenAI co-founder Andrej Karpathy recently expressed disappointment in Reinforcement Learning from Human Feedback (RLHF). He said that unlike true reinforcement learning (RL), where the output is clear and directly tied to success, RLHF relies on subjective human judgments, making it less reliable for optimising model performance. The role of emotions like ‘anxiety’ becomes amplified when using techniques like RLHF, which doesn’t exactly mirror traditional RL models in nature. These systems are trained to align with human expectations, amplifying the biases inherent in the data and training process. As explored, synthetic data could offer a solution by allowing us to model and mitigate the influence of emotional bias, creating more robust and unbiased AI systems. However, it’s clear that to build truly advanced AI, that potentially reaches AGI, we must carefully consider how emotional factors like anxiety and stress in user inputs influence model behaviour. From OpenAI’s compute-heavy methods to Meta’s human-like reasoning and DeepMind’s neuro-symbolic models, we are getting closer to a future where models will truly understand and maybe even surpass our intelligence. Maybe even emotional intelligence, so to speak. As AI continues to evolve, understanding emotional simulations will become essential to improve how these systems interact with people. This will be especially true in high-stakes settings like healthcare, law enforcement, and customer service.While the conversation about emotional intelligence in AI seems curious and fascinating, it also raises some caution about the topic.","excerpt":"Models like GPT-4 and Claude-1 may be more robust in handling emotional shifts, possibly due to their training.","categories":["AI Features"],"tags":["Emotion Detection","LLMs"],"author_name":"Sanjana Gupta","publish_date":"2024-12-11T17:00:07","publication_year":"2024","word_count":887,"keywords":["Anthropic","ChatGPT","Emotion Detection","TPU","RLHF","OpenAI","AI","LLMs","Go","Aim","deep learning","R"],"extracted_tech_keywords":["AI","deep learning","ChatGPT","OpenAI","Anthropic","Aim","RLHF","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/llms-get-anxious-just-like-humans\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10113310,"title":"Vision Pros Being Returned is Just Internet Murmur","content":"Apple walked into the virtual and augmented reality world with the Vision Pro with a $3500 tag hanging on its offering. Two weeks later, a few reports of fans returning their spatial computing headset surfaced on the internet.  However, the ground reality seems a bit different. Research found that “there doesn’t appear to be that much in the way of returns, and certainly not a cataclysmic flood”. The news of returns emerged right at the end of Apple’s 14-day return policy window, so the dedicated customers might just want their cash back. The narrative that the return season is here began with Twitter sources and complaints about the headset. There are reasonable complaints by the returners — but to hold them up as a sign of mass returns isn’t right. As pointed out by Bloomberg’s Mark Gurman, the reasons for returning products circle around a few lines. For instance, several said the headset was too heavy and uncomfortable. The lack of apps, video content, and productivity features was disappointing for others. Some users complained about the headset’s glare and narrow field of view. In contrast, others said it made them feel “isolated from family and friends” due to its lack of meaningful shared experiences and the difficulties in letting other people use it. But the case is different for everyone. Plenty of users are happy with the gadget and do not plan on returning it (at least anytime soon), including Gurman, who wrote the entire newsletter using the headset. A Promise Made The iPhone maker announced Vision Pro in June last year during the annual WWDC conference. “It’s the first Apple product you look through and not at,” said Tim Cook, Apple’s CEO. The product was saved for the end and given the most screen time (40 minutes!) in the chief’s keynote. The device’s promise was to change how humans and machines work together, which has been kept. Apple has always focused on strengthening its ecosystem through every product it designs, be it a watch or earpods. The Cupertino giant has done a commendable job at allowing users to connect their Macs to the Vision Pro, making them seamless for work. Marketed as ‘extraordinary new experiences’, Apple promised that Vision Pro would bring “a new dimension to powerful personal computing by changing the way users interact with their favourite apps, capture and relive memories, enjoy stunning TV shows and movies, and connect with others in FaceTime”. The first version is not extraordinary given the limited number of options available since Netflix and YouTube can only be accessed via web browser, and FaceTime needs to develop further. As The Wall Street Journal noted in an episode of Tech Things with Joanna Stern, the feature that lets users create 3D avatars for FaceTiming is awful. The quality of the persona needs to change massively to match the quality of our usual video calls. And it will, since it is currently in beta. Without disagreement, Apple’s headset is far better than the current alternatives being marketed quality-wise. The device has an ultra-high-resolution display system that packs 23 million pixels per eye and is overall impressive. However, the pricing factor is not inviting even compared to Apple’s other products. One of the reasons early Vision Pro users are returning their headsets is that they never intended to keep them. Apple started working on Vision Pro a decade ago and has delivered a stunning product. The slight horrors persist, but let’s keep in mind this is the first version of the headset. As of now, the important point for Apple is not the return rate but producing a better and cheaper version. The future updates will hopefully meet the promised criteria and resolve issues as Apple meticulously takes note of the returns.","excerpt":"Apple’s promise to change the way humans and machines work together has been kept.","categories":["AI Features"],"tags":["Apple","Vision Pro"],"author_name":"Tasmia Ansari","publish_date":"2024-02-20T17:25:55","publication_year":"2024","word_count":627,"keywords":["Vision Pro","Go","programming_languages:R","AI","Apple","ML","programming_languages:Go","ViT","GAN","R"],"extracted_tech_keywords":["AI","ML","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/vision-pros-being-returned-is-just-internet-murmur\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10096432,"title":"This Indian Startup is Going to Space","content":"In the vast expanse of the unknown, where the limits of human exploration meet the boundaries of the universe, a new player has emerged, promising to redefine the space industry. Erisha Space, a startup based in India, is embarking on a mission to revolutionise space technology and usher in an era of cost-effective satellite solutions. In an exclusive interview with AIM, Dr Darshan Rana, managing director and chairman of Erisha Space, said that instead of a competitive approach, Indian space tech startups are taking a collaborative approach within the industry. Founded a year ago in New Delhi and now operating out of Bengaluru, Erisha Space has been leveraging multi-mission satellite sensors, ground instruments, SCADA, and socio-economic data to develop mathematical models for tracking and monitoring large-scale changes in the environment for tackling climate change. “Erisha Space has developed the Satellite System Platform (SSP) with an aim to provide low-cost solutions to companies in agriculture, e-mobility, aerospace, and maritime and defence tech sectors,” said Rana. Erisha’s ambitions extend far beyond just satellites. The company envisions a comprehensive ecosystem that includes upstream satellite production, downstream ground systems, web GIS, and data products. Last month, the company announced its plan to launch a satellite equipping AI-based image processing capabilities for environmental analysis and sustainable development. The satellite, set to be launched in 2024, will boost the company’s remote vision capabilities by GIS and photogrammetry in various domains. “By integrating AI and ML algorithms into our offerings, we aim to dramatically reduce costs, making remote sensing solutions more accessible to a wide range of industries, including agriculture, oil and gas, and defence,” added Rana. Erisha Space’s team comprises various experienced players in the industry such as Debaddata Mishra, the director and COO of Erisha Space, who has also worked as a senior scientist at ISRO’s Gaganyaan Project. Collaboration over competition is what Indian space industry needs Driven by a collaborative approach, Erisha Space and other space tech startups in India work together rather than engaging in cutthroat competition. Each company focuses on different segments, such as satellite production, launch vehicles, components, or software development, creating a symbiotic relationship that benefits the entire industry. Even then, what sets Erisha Space apart from others is its holistic approach to space technology. “Unlike other companies focused solely on satellite production or launching, Erisha is involved in every aspect of the process,” said Rana. The startup designs and produces satellites in-house, develops ground systems applications, and analyses the data collected by their satellites, giving them the edge over others. “I can say this is the collaborative approach because there is no competition as of now in India,” emphasised Rana. Some companies are developing satellites, other companies are  developing launch vehicles and components, while others are developing software. “Companies like us are coming into all segments to act as supporters. We are working on the last segment so we can support agriculture, defence, oil and gas,” Rana explained his Erisha Space’s position in the spacetech landscape. Erisha Space’s technological innovations lie in their satellite designs. They are pioneers in the development of nano and micro-satellites with enhanced resolution capabilities. These miniaturised satellites can be modified and controlled remotely, allowing for real-time adjustments and reducing latency in data delivery. This breakthrough technology enables Erisha Space to offer comprehensive solutions that include high-quality imagery and advanced software analytics to their customers, all at a fraction of the cost of traditional satellite systems. Government and future plans In April, the announcement of the Indian Space Policy 2023 made the space available for private sectors. “This strategy is undoubtedly highly encouraging and beneficial for space companies like ours, who are attempting not only to develop low-cost space technologies, but also to make space technology accessible, acceptable, and affordable to society,” said Rana. With this approach, NGEs can now use ISRO’s test facilities and R&D expertise for a minimal user fee, which are both expensive and time-devouring. “As a startup, we cannot afford such investments,” added Rana. This highlights how like any space startup, Erisha Space faces its fair share of infrastructure challenges. “The industry is still relatively new, and the lack of infrastructure and readily available technology components pose obstacles,” said Rana. While India is just beginning to tap into the potential of the private space sector, they are learning valuable lessons from countries like the US. The Indian government also planned to allow 100% foreign direct investment (FDI) in the space sector, paving the way for big-tech companies to invest in India. Rana said that the Indian government could further support the growth of the space tech startup ecosystem by introducing incentives and policies that encourage innovation and attract more investment. Relying on third-party organisations for manufacturing and launch services also presents difficulties. Additionally, securing adequate funding for their ambitious projects remains a constant challenge. Despite these obstacles, Erisha Space is determined to overcome them by developing most of their components and subsystems in-house, collaborating with experienced scientists and space agencies, and exploring fundraising options. Looking to the future, Erisha Space aims to develop a data analytics platform that combines data from satellites, airborne sources, and ground stations. By leveraging AI and ML processing methods, Erisha Space intends to offer comprehensive monitoring systems for agriculture, oil and gas, defence, infrastructure planning, agronomics, and mapping. As Erisha Space progresses towards its goals, they have set their sights on launching an SSLV (Small Satellite Launch Vehicle) by 2026. This launch vehicle will have a payload capacity of 1,000 kg. Additionally, Erisha is actively researching reusable satellite technology to further enhance the efficiency and cost-effectiveness of their systems, something which ISRO achieved just a few months back with RLV LEX. It is clear that this startup is here to stay and its collaborative approach will make its space defined in the Indian space ecosystem.","excerpt":"Driven by a collaborative approach, Indian space tech startups work together rather than engaging in cutthroat competition.","categories":["Deep Tech"],"tags":["AI Startups"],"author_name":"Mohit Pandey","publish_date":"2023-07-06T13:07:20","publication_year":"2023","word_count":969,"keywords":["Go","AI","innovation","ML","RAG","Aim","analytics","GAN","R","analytics platform","AI Startups"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","R","Go","GAN","analytics platform","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/this-indian-startup-is-going-to-space\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10272,"title":"Interview – Tanmay Bakshi, world’s youngest Watson Programmer","content":"At an age when most kids have fun and play on their mind, Tanmay Bakshi, 12, is busy developing apps. Tanamy who is being home-schooled in Ontario, Canada is the world’s youngest app developer and has been programming since he was just five. Recently Tanmay addressed 10,000 developers in Bengaluru, India at IBM’s Developer Connect 2016 demoing his recent creation ‘AskTanmay’. On the lighter side, Tanmay loves to play table tennis and coding is like a fun activity for the 12 year old genius.  Well let’s hear it all from the horse’s mouth. So Analytics India Magazine (AIM) interviewed the whiz himself and here’s what Tanmay had to say. AIM: Tell us something about your journey of being one of the youngest app developer in the world. Tanmay: My journey of being one of the youngest app developers began when I was 5 years old.  At this age, like most other kids, I did not have much to do. I used to see my dad programming, and I was intrigued by a computer’s power to display something on screen, or even add 2 numbers. My dad introduced me to a little bit of Batch and FoxPro programming, and after that, I started using the internet as a learning resource; I learned more languages like C\/C++\/Objective-C and VB, and when I was 9 years old, my first iOS App, tTables, was accepted into the iOS App Store. Also, nowadays, I like to program in languages like Swift, Java and Python. In fact, I’ve also authored a book called “Hello, Swift! – iOS App Programming for Kids and Other Beginners”; the early release is currently available here: https:\/\/www.manning.com\/books\/hello-swift In fact, I truly wish to get my book signed by Amitabh Bachchan. AIM: Would you like to share something about Watson? Tanmay: IBM Watson is an amazing set of powerful APIs that allows developers to create cognitive-fueled applications, with ease. I found out about IBM Watson a few months ago, when I stumbled upon a documentary about it; it truly fascinated me as to how a computer was able to play Jeopardy and win against the 2 best human competitors. And so, I did some research about Watson, because I wanted to get to the root of it, and in the process, came across Bluemix, and found out that Watson had become a set of APIs. I created a few YouTube tutorials about its APIs like Retrieve and Rank, Natural Language Classifier, Text to Speech, Speech to Text, etc., and some applications too, one of those being AskTanmay, an NLQA (Natural Language Question Answering) System. AIM: Would you like to share about some of the apps that you have developed? Tanmay: Sure. I have been developing apps for a couple of years now. The most interesting story related to my app development, is that I have my apps and share those through various media. Of course, the very first being the iOS App Store, where I have 6 apps. 3 of my favourite include “tTables”, “tID Vault”, and “I Can, We Can!”. I have also released a number guessing game for the Apple Watch and iPhone, called “tGuess”. Next comes a lot of apps that I have created for my readers, and are in my book that I have authored; in fact, almost every concept that I teach in my book, has an app to support it. Last but not the least, my YouTube channel; I have a YouTube channel called “tanmay bakshi”, it has 89 videos, all of which are tutorial-focused, and again, I teach concepts using apps that I build for almost every tutorial. AIM: Tell us something about world’s first web-based NLQA system powered by IBM Watson, ‘AskTanmay’. Tanmay: Well, it’s a pleasure to share with you, AskTanmay. This is the world’s first NLQA System to be powered by IBM Watson, in fact, AskTanmay uses Watson’s Natural Language Classifier and Alchemy Language services. IBM Watson is used in the ATD (Answer Type Detection) step, to detect what the user is looking for, and the CAFS (Candidate Answer Filtering and Scoring) step, to filter out garbage from the candidate answer list. Put simply, Watson is the brain of AskTanmay. Also, AskTanmay is currently 70% Swift, 20% Java, and 10% Python code. I first presented this at my Keynote at IBM Interconnect 2016 in Las Vegas in February; and in fact, at IBM DeveloperConnect 2016 in Bangalore, I open-sourced AskTanmay for the developer community to use, learn, and contribute to it. And at DeveloperConnect, I also launched the second version of AskTanmay, with 5 big updates. I look forward for AskTanmay to play KBC with Amitabh Bachchan, just as how Watson was able to play Jeopardy with Alex Trebek. AIM: Off late, Artificial Intelligence is seen entering most of the industries and taking over what humans can do? What are your thoughts about it? Tanmay: Well, to begin, I’d like to say one thing: You should not be afraid of AI. What I mean by this is that sometimes, people think that “AI will take over our jobs”, or “AI will replace humans”, and stuff like that. However, I’d just like to say that this is IMPOSSIBLE; there’s a reason that it’s called ‘Artificial intelligence’, not ‘intelligence’, and so, I believe that we should accept the future of AI. Another point is that this “AI” trend is a transition from regular, plain and simple computing, to machine learning, and all transitions are painful initially. However, transitions have to be made. And so, AI is most certainly the “technology of the future”. It breaks the “computers can only do math” boundary, by introducing computers to Natural Language, using math. It will, in fact, create a lot of jobs, replace some redundant jobs, and put them to more useful ones, which it is unable to replace. AIM: What are the most significant challenges you faced in the Artificial Intelligence space? Tanmay: Well, mainly, the biggest challenge would be the starting point. Since computers are not made to be programmed with “machine learning in mind”, the beginning part of starting out with AI is very hard. However, systems like IBM Watson make it easy to use cognitive in your applications, in fact, as a service! AIM: Could you tell us the most important contemporary trends that you see emerging in the present Artificial Intelligence space across the globe? Tanmay: Mainly, it would have to be the rise of robots powered by Cognitive services, like IBM Watson. For example, the Nao robot, which is powered by IBM Watson, has been installed in places like Hotels and Banks, to provide human-like service to consumers, quickly, efficiently, and cost-effectively. IBM Watson’s machine learning capabilities are also helping doctors in the medical field, by sustaining the information it learns, and being able to remember every single punctuation mark. Things like self-driving cars, and even your YouTube recommended feed all use Machine Learning.","excerpt":"At an age when most kids have fun and play on their mind, Tanmay Bakshi, 12, is busy developing apps. Tanamy who is being home-schooled in Ontario, Canada is the world’s youngest app developer and has been programming since he was just five. Recently Tanmay addressed 10,000 developers in Bengaluru, India at IBM’s Developer Connect […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Interviews and Discussions"],"author_name":"Manisha Salecha","publish_date":"2016-06-24T06:26:33","publication_year":"2016","word_count":1153,"keywords":["Go","machine learning","artificial intelligence","AI","Python","Aim","C++","analytics","R","Java","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","Python","R","Go","Java","C++"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-tanmay-bakshi-worlds-youngest-watson-programmer\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":15551,"title":"Machine Learning driving space exploration &#8211; see how NASA is using AI to find life beyond Earth","content":"You may be all too familiar with HAL 9000 from Space Odyssey, 2001 film , AI program onboard the spaceship that became all too powerful. Last year, NASA inched closer to bringing its own version of HAL to life.  Welcome NASA’s humanoid robot, developed to power space exploration and help astronauts by operating in extreme and dangerous environments and performing repetitive tasks. Though Robonaut is still in a test phase, the ESA powered it up to test its power consumption and troubleshoot a faulty cable. The ESA post reveals that Robonauts could be used to explore other planets under human control by astronauts – the Haptics2 telerobotic experiment. From unmanned shuttles to rovers and now robonauts currently in the test stage, AI technology is helping space scientists chart safe paths of travel and respond to emergencies faster. Of late, there has been a surge of interest in artificial intelligence (AI) and NASA has been using it not just for Curiosity’s NAV, but in many other ways. Scientists believe AI is a key enabler in taking space exploration forward. Analytics India Magazine presents some ways NASA and other space agencies use AI in space exploration AI Driving on Mars: If you thought Google, Tesla, Uber and its likes are the first ones to break new ground in developing autonomous driving systems, think again. According to Research Technologist Masahiro Ono, autonomous driving is a decade old technology on Mars. AutoNav – the autonomous driving technology behind the wheel of Spirit and Opportunity rovers landed on Mars in 2004. AutoNav also powers Curiosity, the latest rover taking the rounds of rocky, inhospitable Mars terrain. Mar’s surface is rocky and is characterized by sand dunes. Hence, for the spacecraft to operate carefully, AI algorithm have to be mature to meet the standards. Should there be an AI scientist onboard the Rover?  Well, there is already an AI scientist, an algorithm AEGIS (The Autonomous Exploration for Gathering Increased Science) that provides automated data collection for planetary rovers. The AI algorithm that provides automated targeting capabilities uploaded on to Opportunity rover in December 2009. With AEGIS, the rover can intelligently choose targeted areas and offer scientists the ability to move around a planetary surface and explore different areas of interest. Through the data collected via AEGIS, scientific are able to understand Mars’ current and past environment, gather data about Martian winters, the history of rocks and the availability of ist water. Resilient Spacecraft Executive: There is another research work brewing from the Jet Propulsion Laboratory, Massachusetts Institute of Technology and California Institute of Technology that focuses on developing a risk-aware software architecture for onboard, real-time autonomous operations to handle uncertainty in spacecraft behavior in hazardous and unconstrained environments. The research explores questions such as how can we continue to explore challenging new locations without increasing risk or system complexity? Science objectives would have to be revised on the fly, with new data collection and navigation decisions on short timescales. Spacecrafts will have to adapt to component failures and make risk-aware decisions without getting a go-ahead from ground. Challenges of implementing AI in space There is no doubt that AI is a key enabler in space exploration. And Mars is not the final destination for space exploration. NASA’s goal is to go beyond Mars to further deep space exploration – the premier space agency has grand plans to explore Jupiter’s moon Europa and other ocean worlds. News reports indicate that NASA plans to send robotic probes to Jupiter’s moon to find out about the icy oceans that lie beneath the surfaces, which many believe could be home to extra-terrestrial life. In an attempt to find out life beyond Earth, NASA is placing its bets on the ocean worlds of the solar system. NASA also plans for achieving a crewed Mars surface mission in the late 2030s. One of the biggest challenges in AI in space, as research technologist Ono points out is autonomy of AI. Ono reveals in his blog that , any autonomy algorithms on spacecraft are designed very conservatively. Since spacecrafts are operated conservatively, despite the availability of AI, human operators prefer to fly spacecraft manually as much as possible since AI can also make mistakes.","excerpt":"You may be all too familiar with HAL 9000 from Space Odyssey, 2001 film , AI program onboard the spaceship that became all too powerful. Last year, NASA inched closer to bringing its own version of HAL to life.  Welcome NASA’s humanoid robot, developed to power space exploration and help astronauts by operating in extreme […]","categories":["IT Services"],"tags":["NASA","NASA ai","space exploration"],"author_name":"Richa Bhatia","publish_date":"2017-06-13T03:31:08","publication_year":"2017","word_count":701,"keywords":["Go","artificial intelligence","NASA","NASA ai","AI","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving","analytics","space exploration","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","Go","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving"],"url":"https:\/\/analyticsindiamag.com\/it-services\/machine-learning-driving-space-exploration-see-nasa-using-ai-find-life-beyond-earth\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10162855,"title":"Datar Cancer Genetics Launches AI Platform for Personalised Cancer Treatment","content":"Datar Cancer Genetics (DCG), a Maharashtra-based diagnostic centre, has introduced Exacta AI, an AI-powered multi-analyte platform designed to optimise treatment options for cancer patients. The platform integrates molecular, proteomic, genomic, functional, and clinical data, using AI-assisted analysis to generate personalised treatment strategies. It also evaluates drug synergy, interactions, toxicity, and contraindications for a wide range of therapies, including antibody-drug conjugates (ADCs), checkpoint inhibitors (CPIs), targeted therapies, endocrine agents, chemotherapy, and repurposed drugs. Exacta AI can propose up to 10 evidence-based multi-drug combinations for oncologists and molecular tumour boards (MTBs) to consider. “This AI-powered approach is poised to provide an integrated, evidence-based set of treatment options that can transform how we personalise cancer treatment,” said Dr Sewanti Limaye, director of medical and precision oncology at Sir HN Reliance Foundation Hospital. Dr Darshana Patil, senior director for global strategy and medical affairs at DCG, also highlighted the platform’s role in handling large volumes of clinical and molecular data. “When standard treatments fail, oncologists and MTBs must analyse massive datasets while racing against time to make critical decisions. Exacta AI transforms this challenge into an opportunity for personalised medicine.” The AI-driven system also enhances MTB workflows by presenting structured case analyses, evidence-based treatment options, and references to similar reported cases. In a retrospective study of 265 patients with drug-resistant solid organ cancers, Exacta AI provided at least seven therapy options per patient. In contrast, conventional molecular profiling yielded only two options for 15% of patients. Dr Vineet Datta, senior director for global strategy and business development at DCG, explained that Exacta AI functions like a multidisciplinary expert team, which processes data in minutes instead of days. Exacta AI is now available for cancer centres and hospitals, with seamless integration, training, and ongoing support. Regarding the use of AI in medicine and its advantages, Dr Andy Gaya, a clinical oncologist at The Cromwell Hospital in London, said that technology provides an unprecedented level of therapy analysis. It aims to improve patient outcomes while reducing exposure to ineffective treatments. When looking at beating cancer with genomics and AI, an Infosys-backed Bengaluru startup, 4baseCare, introduced a similar approach to treating oncology patients in India. In an interaction with AIM, co-founder and CEO of 4baseCare, Hitesh Goswami, explained that today lung cancer stage-2 and stage-3 patients are subgrouped into 12-15 categories. Each category requires distinct treatments, as therapies effective for one group may not work for another and could even cause adverse reactions. This calls for even more attention to customised treatments for patients rather than standardised procedures to be applied uniformly.","excerpt":"In a retrospective study of 265 patients with drug-resistant solid organ cancers, ‘Exacta AI’ provided at least seven therapy options per patient.","categories":["AI News"],"tags":["AI","cancer detection ai","diagnosis","medical ai"],"author_name":"Sanjana Gupta","publish_date":"2025-02-04T15:23:20","publication_year":"2025","word_count":425,"keywords":["Go","API","diagnosis","programming_languages:R","AI","ML","programming_languages:Go","cancer detection ai","Aim","GAN","R","medical ai","startup"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","API","GAN","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/datar-cancer-genetics-launches-ai-platform-for-personalised-cancer-treatment\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10094599,"title":"Graph Databases are Having a Moment in the Database Market","content":"Graph databases initially emerged to help us understand relationships and store these data entities. Swiftly, they popped up in instances of social networking, recommendation engines and fraud detection since they fit the bill perfectly. With twelve years of experience in data engineering, Arvind Ramachandran has been instrumental in revamping the existing platform by re-engineering various processes. Analytics India Magazine had a quick chat with Arvind to understand more about the use cases and the future of graph databases. AIM: What are graph databases and how do they differ from the traditional database systems? Arvind: In order for us to understand how graph databases differ from traditional databases, we first need to understand what these are. Graph databases are a type of database management system designed to store and process data in the form of graphs. These graphs typically contain nodes and edges. For instance, in the retail world, nodes can represent entities such as customers, warehouses and products. Edges, on the other hand, represent the relationships between the nodes. So, unlike traditional databases that store data in the row and column format, graph databases are more flexible, enabling users to easily analyse the data and the relationships between entities. This would otherwise involve complex queries in traditional databases. AIM: Will graph databases really play an important role in the future? Arvind: Graph databases are already playing an important role in various industries such as e-commerce, supply chain and banking, and are expected to become even more prevalent in the near future. They have started gaining more popularity in the healthcare industry for managing patient data and analysing symptoms. For example, with COVID cases, graph databases were able to chart out people affected by the disease and also everyone who was in their proximity so that they could be quarantined. Graph databases are also being used in the transportation industry for optimising routes and logistics. As data continues to grow, it is imperative to keep up with the ever-increasing demands of effective data management and governance. Graph databases not only establish effective data management systems but they’re also helpful in identifying patterns and anomalies and can help with high-speed traversal of complex relationships, which is a powerful tool for data analysis. AIM: What are some of the real-world use cases and applications of graph databases? Arvind: Graph databases have numerous real-time applications. For example, e-commerce platforms use graph databases extensively to build recommendation engines based on user behaviour and purchase history. This enables them to offer personalised product suggestions to the customers. Social networking platforms like Facebook use a graph database, called TAO, to understand the connections between users, which helps them deliver personalised marketing campaigns. Similarly, LinkedIn uses an in-house graph database, called Llquid. It extensively maps its second-degree networks which are humongous. So, essentially, we are tracking the lineage and provenance of data for audit, compliance and governance purposes. By modelling data entities in the form of graphs, it becomes easier to visualise and manage complex connections. AIM: How will graph databases revolutionise data analytics and business intelligence? Arvind: Graph databases are gaining popularity in modern businesses because they can represent very complicated relationships thereby facilitating diverse analytical capabilities such as market basket analysis, social network analysis, product recommendations, inventory management, and more. They can also identify key influencers in an organisation and analyse communication patterns to improve team performance and collaboration. With the power of ML in BI, businesses can uncover insights to gain a competitive edge. Graph databases can also be combined with BI tools like Power BI and Tableau, allowing users to create visualisations of data on their dashboards. AIM: How do graph databases handle complex relationships between data points? Arvind: Querying languages are designed specifically to handle such complex relationships. For instance, Cypher, used by Neo4j, is specifically used for querying graph databases. Graph databases that support Cypher have built-in algorithms that can be directly used to calculate shortest paths and perform other analytic tasks. Graph databases also offer an easy way to create new relationships and update data models without the need for denormalisation. This feature significantly improves query performances for navigating data relationships, which is particularly valuable for businesses that need to make informed decisions based on large amounts of data. AIM: How can graph databases be integrated with other data storage and processing technologies such as Hadoop and Spark? Arvind: There is an open-source statistical processing called GraphX that is specifically designed for large-scale graphs and graph-structured data. It efficiently processes large-scale graphs that can be used in transactions with graph databases to perform various graph analytic tasks. Graph databases also give you the provision to export graph data in various formats such as JSON and CSV, which can then be loaded onto platforms like Hadoop. You can also leverage connectors to facilitate spark – graph database integrations. AIM: What are the benefits and advantages for a company in adopting graph databases ? Arvind: The integration of graph databases is a remarkable advancement in how companies can effectively utilise and leverage data. When an organisation embarks on their journey to adopt graph databases, it embraces a forward-thinking approach to data architecture. Graph databases offer valuable attributes such as flexibility, scalability, and agility, which empowers companies to adapt to evolving business requirements and maintain a competitive edge in a rapidly changing market. By leveraging the potential of interconnected data, companies can gain invaluable insights, enhance customer experiences, and optimise operations. Whether it involves integrating new data sources, accommodating dynamic schemas, or enabling real-time analytics, graph databases provide the foundational elements necessary for companies to establish a data-driven organisation poised for success.","excerpt":"Graph databases are already playing an important role in various industries such as e-commerce, supply chain and banking, and are expected to become even more prevalent in the near future","categories":["AI Trends"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2023-06-06T17:46:51","publication_year":"2023","word_count":938,"keywords":["Go","API","AI","ML","Scala","RAG","Aim","analytics","R","fraud detection"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","fraud detection","R","Go","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/graph-databases-are-having-a-moment-in-the-database-market\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10051700,"title":"Why Is There A Dearth Of Women In Data Science","content":"Data Science in the present times is considered as one of the lucrative fields for young professionals with immense scope for growth. Everybody interacts with technology to some extent now. Hence, data scientists are in hot demand as they can equip the firms to plan better and effectively. However, if we look deep into the field, there lies a problem — the lack of gender diversity. With more than 660 million women, India bags the first position when it comes to the number of female graduates in science, technology, engineering, and maths (STEM). Women make up as much as 40% of Indians who graduate with STEM degrees. But the same country ranks 19th when it comes to their employment, thereby raising a red alert for the system, administration, and society as a whole. What pulls them back While women account for roughly 55 percent of university graduates on average across countries, only two-thirds of this valuable talent pool pursue a career in a STEM-related sector like engineering, software development, or analytics, and even fewer pursue a career in data science. Women make up only approximately 15% to 22% of all data science professionals, according to various surveys from WEF, Global Gender Gap Report and BCG research. Image Credits: BCG research Talking about the initial hurdle, Mathangi Sri, VP of Data Science at Gojek, said: “The value of STEM education for girls is not emphasised in society as a whole. We still have stereotypes that place fine art in the hands of girls and STEM topics in the hands of boys. Girls should be encouraged to try new things, let them experiment more, and let them often fail in school. In the family and outside, we must actively work to break prejudices.” She further added, “Some of the biases are inherent that possibly creep in during career promotions and even interviews. Furthermore, it is not uncommon to see outspoken and assertive people progressing more quickly than those who are not. Many women possibly fall in the latter category.” Several other reasons can be cited for this great imbalance in the data science domains. Firstly, the stamp of perfection provided to the female section of our society tempts us to keep them away from making mistakes, while the only way to learn and stay sound in the field of data science is by making mistakes. Secondly, the tech industry, accept it or not, is male-dominated, which often results in fewer role models for women to look up to. Thirdly, interestingly, the perception among women from the STEM field is not that positive. Candidates believe the subject is more theoretical and abstract, with a concentration on manipulating code and data with little impact and, by extension, little purpose. Moreover, the feeling of cut-throat competition and a “nerdy” work culture among female aspirants holds them back from even entering. The high turnover rates of women in data science and AI fields can also be traced to the existence of bullying, sexism, and sexual harassment at the workplace. Women are also discouraged from continuing their careers in data science and AI because of the gender pay gap, poor career growth for women, male-dominated office culture, lack of access to mentors, and gender bias in recruiting. Additionally, in the tech industry, attrition is a huge issue. Many women in tech companies leave their jobs in the middle of their careers (10-20 years). A large part of them either drop out or go on to work in other fields, as per industry research. Way forward The effectiveness of evidence-based policymaking is harmed when it is based on skewed data. The fact that most data scientists—those who gather, organise, analyse, and make decisions—are men, is a possible source of bias in many datasets. Hence, apart from working on the current challenges, to overcome the shortfall of female data scientists, governments should first ensure that girls have the basic literacy and numeracy abilities required to connect with digital technologies. A policy approach that focuses solely on improving skills associated with the top tiers of the pyramid is unlikely to broaden digital economy participation. In reality, it may exacerbate existing inequities, further marginalising women and girls.","excerpt":"Women are discouraged from continuing their careers in data science and AI because of the gender pay gap, poor career growth for women, male-dominated office culture, lack of access to mentors, and gender bias in recruiting.","categories":["IT Services"],"tags":["Data Science","Women in Data Science"],"author_name":"kumar Gandharv","publish_date":"2021-10-15T13:00:00","publication_year":"2021","word_count":696,"keywords":["data science","Go","Women in Data Science","programming_languages:R","AI","programming_languages:Go","Git","RAG","analytics","GAN","Data Science","R"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-is-there-a-dearth-of-women-in-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10069473,"title":"The discreet charm of CVPR","content":"This year’s Computer Vision and Pattern Recognition (CVPR) conference has already started on June 19 and will continue till June 24 in New Orleans, Louisiana, as well as virtually. Like every year, it is expected to attract more than 7,500 attendees and feature keynote speakers, presentations, tutorials, a panel session, and workshops. All the major names in tech are present at CVPR 2022 and will conduct workshops, present their papers and discuss the innovations they are working on. Meta Come join us at the XAI4CV workshop on Monday June 20th at #CVPR22 to learn about the latest research in explainable and transparent models. We have a fantastic set of invited speakers, 30+ posters, and a great panel discussion, details: https:\/\/t.co\/VVf7ZxyoWX— AI at Meta (@AIatMeta) June 19, 2022 CV models are usually black-box in nature and do not provide explanations for their predictions. This lack of transparency, can lead to a lack of trust among consumers, which can cause backlash when algorithms make mistakes. This workshop will help build proactive adaptation of explainability in computer vision systems. The agenda will be to have conversations that will work on “building top-performing explainable computer vision systems.” It will put a greater focus on providing the human-like reasoning of “why” the model made those predictions. Meta will also talk about Project Aria, Egocentric Perception, Self Supervised Learning Demos and Avatar Puppeteering Demo. Some other research areas that will be a talking point from Meta include Ego4D: Around the World in 3,000 Hours of Egocentric VideoHVH: Learning a Hybrid Neural Volumetric Representation for Dynamic Hair Performance CaptureKeyTr: Keypoint Transporter for 3D Reconstruction of Deformable Objects in VideosMasked-attention Mask Transformer for Universal Image SegmentationMasked Autoencoders Are Scalable Vision LearnersMViTv2: Improved Multiscale Vision Transformers for Classification and DetectionNeural 3D Video Synthesis from Multi-View VideoOmnivore: A Single Model for Many Visual ModalitiesPONI: Potential Functions for ObjectGoal Navigation With Interaction-Free LearningVisual Acoustic Matching To read more about Meta’s plans for CVPR 2022, click here. Apple Some of the accepted papers from Apple in CVPR 2022: I’m excited to attend #CVPR2022. Learn more about Apple’s accepted papers and talks this yearhttps:\/\/t.co\/Qvff7uVKCQ— James Gray King (@jamesgrayking) June 16, 2022 Critical Regularizations for Neural Surface Reconstruction in the Wild – The researchers introduce RegSDF, which shows that “proper point cloud supervisions and geometry regularizations are sufficient to produce high-quality and robust reconstruction results.” Forward Compatible Training for Large-Scale Embedding Retrieval Systems – The researchers bring out a forward-compatible training (FCT) – a new learning paradigm for representation learning. When the old model is trained, the researchers also prepare for a future unknown version of the model. They propose “learning side-information, an auxiliary feature for each sample that facilitates future updates of the model.” Robust Joint Shape and Pose Optimization for Few-View Object Reconstruction – The researchers show FvOR, a learning-based object reconstruction method that predicts accurate 3D models, given a few images with noisy input poses. Efficient Multi-View Stereo via Attention-Driven 2D Convolutions -The researchers bring out MVS2D, a multi-view stereo algorithm, which integrates multi-view constraints into single-view networks via an attention mechanism. Apple will show a demo of RoomPlan technology that allows the user to capture a room and its defining objects in a parametric format within minutes. It is supported on any of the Apple devices that are equipped with LiDAR sensors. To read more about Apple’s plans for CVPR 2022, click here. Samsung Samsung Research will present around 20 thesis papers at CVPR 2022. Two of the papers submitted by Samsung’s Toronto AI Center have been selected for oral presentations (a prestigious feat). Samsung Research Achieves 20 Paper Acceptances for CVPR 2022https:\/\/t.co\/7i8TIPiQRD— Samsung Electronics (@Samsung) June 15, 2022 P3IV: Probabilistic Procedure Planning from Instructional Videos with Weak Supervision – This paper shows how to build AI systems capable of analysing and mimicking human behaviour. An area of research that is growing in this field is procedure planning which can assist humans in goal-directed problems like cooking, repairing gadgets, etc. Day-to-Night Image Synthesis for Training Nighttime Neural ISPs – It shows how to synthesise the nighttime image data needed to train Night Mode using neural Image Signal Processors (ISPs) on smartphone cameras. Through this, it is possible to convert clear daytime images into nighttime image pairs. Waymo On the starting day, Waymo held a tutorial session on synthetic camera data generation for autonomous driving at the LatinX in CV (LXCV) Research workshop. The next day, it conducted a workshop on Autonomous Driving. The Waymo Research team shared results from this year’s Waymo Open Dataset Challenges too. We’re looking forward to participating in @CVPR this year, both in person and online! Here’s a preview of some of our sessions, including recent state-of-the-art work in autonomous driving research that we’ll be presenting: https:\/\/t.co\/kR3giqZgxO pic.twitter.com\/vA3lwV0xZw— Waymo (@Waymo) June 16, 2022 On June 22, a team from Waymo and Google Research will present BlockNeRF. It is a method of large-scale scene reconstruction based on camera images. Waymo will also present what they are working on in a poster session on a novel data-driven range image compression algorithm called RIDDLE (Range Image Deep Delta Encoding).","excerpt":"Samsung Research will present around 20 thesis papers at CVPR 2022.","categories":["Deep Tech"],"tags":[],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-21T17:00:00","publication_year":"2022","word_count":850,"keywords":["Go","ELT","AI","R","Transformers","computer vision","Scala","Ray","Rust","xAI"],"extracted_tech_keywords":["AI","computer vision","xAI","Ray","Transformers","R","Go","Rust","Scala","ELT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/the-discreet-charm-of-cvpr\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136418,"title":"OpenAI Set to Launch Advanced Voice Mode on ChatGPT Soon","content":"OpenAI is set to launch ‘Advanced Voice Mode’ on ChatGPT this Tuesday, September 24, 2024, according to a screenshot posted by a user on X. “As of now, access to Advanced Voice mode is being rolled out in a limited alpha to a select group of users. While being a long-time Plus user and having been selected for SearchGPT are both indicators of your active engagement with our platform, access to the Advanced Voice mode alpha on September 24, 2024, will depend on a variety of factors including but not limited to participation invitations and the specific criteria set for the alpha testing phase,” read the blog post attached in the screenshot. OpenAI released GPT-4o at its latest Spring Update event earlier this year, which won hearts with its ‘omni’ capabilities across text, vision, and audio. OpenAI’s demos, which included a real-time translator, a coding assistant, an AI tutor, a friendly companion, a poet, and a singer, soon became the talk of the town. However, its Advanced Voice Mode wasn’t released. When OpenAI recently released o1, one of them queried if they would be launching voice features soon. “How about a couple of weeks of gratitude for magic intelligence in the sky, and then you can have more toys soon?” replied Sam Altman, with a tinge of sarcasm. However, a couple of weeks later, Kyutai, a French non-profit AI research laboratory, launched Moshi, a real-time native multimodal foundational AI model capable of conversing with humans in real time, much like what OpenAI’s advanced model was intended to do. Hume AI  recently  introduced EVI 2, a new foundational voice-to-voice AI model that promises to enhance human-like interactions. Available in beta, EVI 2 can engage in rapid, fluent conversations with users, interpreting tone and adapting its responses accordingly. The model supports a variety of personalities, accents, and speaking styles and includes multilingual capabilities. Meanwhile, Amazon Alexa is partnering with Anthropic to improve its conversational abilities, making interactions more natural and human-like. Earlier this year, Google launched Astra, an ‘universal AI agent’ built on the Gemini family of AI models. Astra features multimodal processing, enabling it to understand and respond to text, audio, video, and visual inputs simultaneously.","excerpt":"OpenAI released GPT-4o at its latest Spring Update event earlier this year, winning hearts with its ‘omni’ capabilities across text, vision, and audio.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","ChatGPT","OpenAI"],"author_name":"Tanisha Bhattacharjee","publish_date":"2024-09-23T15:55:59","publication_year":"2024","word_count":365,"keywords":["Anthropic","ChatGPT","Go","API","OpenAI","AI","GPT-4o","GPT","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","GPT-4o","ChatGPT","OpenAI","Anthropic","R","Go","API","GPT","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-set-to-launch-advanced-voice-mode-on-chatgpt-soon\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":14176,"title":"Google upgraded Play with machine learning capabilities, opens subscription in India","content":"Google Play Music recently launched their music store in India Seems like Google’s Play Music overhaul, with added machine learning capabilities is paying off.  What’s new is context – personalized music based on what the person is listening, where and why he\/she is listening. Earlier last year, Elias Roman, Lead Product Manager, Google Play Music revealed how machine learning made Google Play Music the ultimate personal DJ — one who not only listens to what the subscriber likes, but also when you like it, so the music that you care about now is always at the top of your screen. Machine Learning is Google’s strength and according to industry watchers, Alphabet, the parent company had taken concrete steps to turn it into a profitable business segment, taking on market behemoths Apple Music and Spotify. Google Play Music is available on iOS, Android and the web. Right music at the right time, thanks to machine learning algorithm Pegged as “smarter, easier to use, and much more assistive,” the new Google Play Music deploys machine learning to dish up personalized, curated music list by factoring in parameters such as location data, activities, time of day and even weather. The update was announced last year in 2016, with a global rollout in 62 countries. Some of the other features include a brand new home screen that puts the favourite songs right on top. And it also comes in with an automatic offline playlist. The feature automatically puts together a playlist of recent songs for offline listening mode, minus the hassle of downloading music. Google’s India Connect, India a prime market According to App Annie study, an app analyst firm, when it came to downloads, India topped the Google Play Store, in 2016 with the total number of downloads at 6 billion, followed by the United States and Brazil. Another contributing factor was that in 2016, India became the world’s second-largest smartphone market, beating United States. The study also pointed out how the downloads increased by 15% in 2016 and time spent on apps rose to 25%, thereby improving revenue for publishers by 40%. In view of the exponential growth in the Indian market, Google Play Music recently launched its subscription service in India, at a discounted offer of INR 89 per month in the first 45 days. With a catalogue of over 400 million local and international hits, subscribers have access to offline listening features and the on-demand catalogue as well.","excerpt":"Seems like Google’s Play Music overhaul, with added machine learning capabilities is paying off.  What’s new is context – personalized music based on what the person is listening, where and why he\/she is listening. Earlier last year, Elias Roman, Lead Product Manager, Google Play Music revealed how machine learning made Google Play Music the ultimate […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-04-10T09:35:48","publication_year":"2017","word_count":408,"keywords":["Go","machine learning","programming_languages:R","AI","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","machine learning","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-upgraded-play-machine-learning-capabilities-opens-subscription-india\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":22299,"title":"How To Prepare For A Data Scientist Interview","content":"Cracking any interview requires preparation and in the case of data science it is not restricted to performing well on the big day alone. An aspiring data scientist is expected to prepare across multiple fronts. Here, we provide you with a insight into the levels of preparation required and how to go about it: Preparing For The Various Rounds Of Interviews: Aptitude Test As the most basic round for any interview, it covers fundamental topics such as English language comprehension, quantitative aptitude and analytical reasoning. While this round requires minimal preparation, lazily scanning through your Wren and Martin and high school grade quantitative aptitude questions will certainly help in brushing up the important concepts. Technical And Problem-Solving Interview Round This is where your technical grasp over the subject is tested — especially your programming language proficiency, knowledge of statistics, optimisation and machine learning. The main languages that one is expected have mastery over are Python ,SQL, R, Scala and Tableau. Knowing Java and C++ also helps in adding depth to your programming skills. Since this round will deal with the brutal basics of the languages which you have stated in your CV, brushing up on their basics is of utmost importance. Both freshers and experienced candidates will be questioned on key and rudimentary topics like: Probability — Random Variable, Bayes Theorem, Probability distribution Statistics — Sampling Theory, Hypothesis testing, Summary statistics Statistical models — Linear Regression, Non-parametric models, Time Series Machine Learning — Bayesian ML, SVM, Decision Tree, Logistic Regression Understanding of neural networks The problem-solving round often involves a case study which requires the candidate to define the problem for the scenario presented, and explain the business impact of the solution. During the process, bringing in examples of case studies and findings of research papers to support the solution will help improve the candidate’s score. The round may also require you to evaluate the robustness, scalability, implementation issues, and so on, of an existing plan and provide alternatives. Freshers may also be asked to explain the mandatory projects carried out for the fulfilment of their academic courses and the rationale behind the methodology and solutions. Hence, it is important to be well-versed with your projects. In the case of more experienced candidates, they may be tested along similar lines. But they may be asked to talk about their real world projects, questioned on the related domains and the impacts of their model on the business. Willingness To Learn The debate surrounding “building talent vs buying talent” has employers split into two groups. On one hand, building talent helps employers incubate and nurture talent according to the needs of the company, while on the other hand, buying talent provides the incentive of hiring people with highly specialised skills. As a data science aspirant, it is important to showcase the skills you have already acquired and honed, but also express willingness to learn in the current job and being open to adaptation. Showcasing Your Inner “Unicorn” Roger Huang of Springboard says: “A data scientist is a unicorn that bridges math, algorithms, experimental design, engineering chops, communication and management skills, but they aren’t specialists in every aspect.” In the current business environment, a data scientist’s role is one of a bridge between multiple facets of a business. Though it is impossible to be an expert in all aspects of it, a data scientist has the unique distinction of being able to ideate and provide solutions across multiple disciplines. Thus, this point ties in with the previous one where it is important to make your technical individuality felt but in the process show the potential employer that you have what it takes to be the unicorn that a good data scientist is expected to be. Before The Interview: Build A Portfolio Of Projects And MOOCS It is true that most employers do not look for a candidate with a doctorate in Data Science. But what they look for is a candidate with some hands-on experience in it. While it is easy for experienced candidates as they have a body of work to show, freshers too can independently create some of their own. One of the best way to do this is by carrying out data science projects. There are ample datasets made available specifically for this purpose on public domains such as Kaggle, GitHub, Google Cloud Services, Amazon Web Services and MNIST, among others. Carrying out projects and participating in open challenges convey your initiative in taking on challenges and working on solutions to tackle them. MOOCS (Massive Open Online Courses) are another way of adding weight to your portfolio. There are many popular platforms such as Coursera, edX, and FutureLearn, among others, which provide various courses (most of them free) related to data science. Along with academic knowledge acquired during your graduation or post graduation, taking such courses provide exposure to different and focussed application of the knowledge. To a prospective employer, this presents the candidate as a well-rounded person with a participatory approach in the field. Network With Peers: Stay Up-To-Date With Who’s Who, Follow Trends Networking with experts from one’s field of interest is vital and data science is no exception. Following experts in the field of data science and data scientists on professional platforms such as LinkedIn will keep you up-to-date with the latest trends in the field. It also provides a great opportunity to learn more about their work and their take on other people’s work. A quick glance through your profile on these platforms (which most employers do these days) gives the employers a peek into your participation on social media. For The Interview: Learn About The Position That You Are Applying For Yes, you are applying for the position of a data scientist. But what kind of a data scientist? It is a very broad term and given the nature of the market demands, it is not a generic one anymore. Roles such as an analyst, market researcher, statistician, and project manager, could also be included in the definition. Different businesses such financial services, e-commerce, marketing and so on, employ data science; and depending on the nature of the business the required skill sets may differ. For example, the banking, financial services and insurance (BFSI) sector requires a stronger theoretical understanding and expertise, while e-commerce requires a more model-oriented expertise. Also, the nature of the business that the employer is engaging in demands one to have a basic understanding of it as the data solutions sought will be in the realm of the business. Hence, before applying for the position, it is important to know the specific demands of the job. Additional Cosmetic and Brownie Points: A Tight Resume A candidates CV or resume proceeds him or her. The proverbial first impression really does apply. That is why it is important to present a good resume which is crisp, concise, and highlights important aspects such as an advanced degree in the subject if you have one, your experience with data (projects, hackathons, etc) and understanding of business domain, among others. Go Through Questions Asked In Previous Interviews Many companies make questions from previous interviews available on sites like Glassdoor for candidates’ reference. Going through these questions have two advantages. One, it provides the candidate a general idea of the nature of questions that may be posed. And two, mentioning your research for the interview when a familiar question is asked will definitely help score brownie points with the interviewer. Concentrate On Your Effort, Not The Outcome: It is impossible for all of us to succeed in all our ventures all the time. A job interview is no different. While some of us may ace an interview with utmost ease, some may fall short. In such a case, it can serve as a valuable source of feedback and aid in self-improvement which can in turn intensify one’s preparation and reinvigorate their efforts.","excerpt":"Cracking any interview requires preparation and in the case of data science it is not restricted to performing well on the big day alone. An aspiring data scientist is expected to prepare across multiple fronts. Here, we provide you with a insight into the levels of preparation required and how to go about it:  Preparing […]","categories":["AI Features"],"tags":["Data Science"],"author_name":"Jeevan Biswas","publish_date":"2018-03-06T11:25:43","publication_year":"2018","word_count":1315,"keywords":["data science","Go","machine learning","AI","neural network","ML","Python","SQL","Data Science","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","data science","Python","R","SQL","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-scientist-interview-preparation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":65594,"title":"Meet This Week’s MachineHack Champions Who Cracked The ‘Financial Risk Prediction’ Hackathon","content":"MachineHack successfully concluded its fifth instalment of the weekend hackathon series this Monday. The Financial Risk Prediction hackathon was greatly welcomed by the data science enthusiasts with active participation from over 250 participants and close to 500 registrations. Out of the 265 competitors, three topped our leaderboard. In this article, we will introduce you to the winners and describe the approach they took to solve the problem. #1: Ashijith Rajendran Ashijith is working as a Data Scientist at Technosoft Global Services. He did his Post Graduate Program in Business Analytics from Praxis Business School. His keen interest in mathematics helped him develop a passion for data science, a passion which he chose to make his career. He regularly updates his skills by reading articles and participating in hackathons. Approach To Solving The Problem Ashijith explains his approach as follows: I started by visualising the data. Then, I created some of the generic models using Logistic Regression, Random Forest, XGBoost and Support Vector Machine. I chose Random Forest Classifier, hoping that it could give good results when fine-tuned. I started tuning the parameters and performed cross-validation, but the results showed no improvement. I went on to perform data exploration to identify the hidden patterns in the data, which helped me in generating new features. I used the generated features on the  Random Forest model that got me the best score and the winning solution. “MachineHack is a good platform for every data scientist. It gives an opportunity to implement the learning and compete with other fellow data science enthusiasts,” he shared his experience. Get the complete code here. #2: Kranthi Kiran Kranthi Kiran is a Computer Science engineering student at Army Institute of Technology, Pune. He first came in touch with machine learning when one of his friends was doing the Titanic Survival Challenge and was amazed that he could predict the survival of a person in a natural disaster. This made him curious enough to try out the problem on his own. By the time he finished the competition, he was totally astonished by the power of analytics and machine learning on real-world problems. Approach To Solving The Problem He explains his approach briefly as follows. On EDA, I found that there were some rules for which the records could be directly classified as 1 (target). I found 5 rules which almost covered 48% of the data given to predict. Internal_Audit_Score >= 9 ==> Target = 1External_Audit_Score >= 9 ==> Target = 1Final_Score >= 9 ==> Target = 1Loss_score >= 9 ==> Target = 1Past_Results >= 2 ==> Target = 1 Next, I made some features based on numerative operators like the addition of scores, subtraction, multiplication and chose 3 best features that boosted my Local CV a little bit. I baselined almost all models of which gradient boosting methods worked great, especially CatBoost. I ended up using CatBoost for my final model, which worked a tad bit better on the local CV than any other model. “MachineHack is a great platform for anybody practising data science and machine learning as you can compete with anybody starting from a student to a data scientist with 10 years of experience and learn  tremendously in parallel to competing with the best,” he shared his opinion on MachineHack Get the complete code here. #3: Niranjan K Niranjan is a 2017 Mechanical Engineering Graduate. He started his career with a core MNC, which provides mechanical design solutions to its clients. Soon, he found out his job to be redundant. Therefore, he started exploring new domains and got acquainted with data science and machine learning. As he kept exploring more and more about this emerging field, he found out to be in line with his interests. Having been inclined towards automation, he always wanted to be in a position where he could directly influence any business. Thus, he quit his job to pursue a postgraduate program in data science from Great Lakes Institute of Management. From then on, he has been honing his programming and data science skills. “Being from a non-programming background, I never doubted my ability to learn to code quickly. The key is to never stop learning,” he said. Approach To Solving The Problem Niranjan explains his approach briefly as follows: Correlation plots suggested there was very less multicollinearity among independent features.The KDE plot of each feature gave insight into the distribution of data.Variable importance using Gradient boost suggested that two features (Internal_audit_score & Fin_score) had 76% weightage in predicting the risk factor.I evaluated logloss scores for different models using H2O AutoML and Gradient Boost gave the minimum logloss.Tuned the hyperparameters manually and got the best score on the leaderboard. “This is my third hackathon in MachineHack, and I got into the top 3 in two of them.The experience has been amazing, and I look forward to many more such learning experiences through MachineHack,” he shared his MachineHack experience. Get the complete code here. Check out new hackathons here","excerpt":"MachineHack successfully concluded its fifth instalment of the weekend hackathon series this Monday. The Financial Risk Prediction hackathon was greatly welcomed by the data science enthusiasts with active participation from over 250 participants and close to 500 registrations. Out of the 265 competitors, three topped our leaderboard. In this article, we will introduce you to […]","categories":["Deep Tech"],"tags":["Hackathon Winners","Machinehack","Weekend Hackathon"],"author_name":"Amal Nair","publish_date":"2020-05-20T19:00:00","publication_year":"2020","word_count":827,"keywords":["data science","Go","machine learning","Weekend Hackathon","AI","Machinehack","ML","automation","XGBoost","analytics","CatBoost","Hackathon Winners","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","XGBoost","CatBoost","R","Go","automation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/financial-risk-prediction-hackathon-winners\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040321,"title":"How Dropbox’s ‘Image Search’ Works?","content":"Lookin up photos in the cloud can be cumbersome and daunting, especially if you have forgotten the filename. To simplify the experience, file hosting platform Dropbox recently launched a new feature to make image search seamless. The feature, available for both enterprises and individual customers, allows to search all images, OCR-based search for images of documents, and full-text search for text documents. Currently, Dropbox serves close to 700 million registered users, 15.48 million paying users and generated close to $1.91 billion in revenue in 2020. In the coming months, the company is looking to launch a search feature for video content. Image content search Typically, when a user searches for images on the cloud storage platforms, they endlessly scroll through a pile of photos until they spot them or try their luck in guessing the filename. However, Dropbox’s new ‘image search’ feature suggests relevant images and calls out the best match, based on a few descriptive words. For instance, if you are looking for photos from a picnic, you can type in the keyword ‘picnic’ or other objects in the image that you can vaguely recall\/remember. In no time, the relevant images would be shown. Example of Image content search results for ‘picnic’ (Source: Dropbox) But, how does this work? Dropbox leverages machine learning techniques to improve the image content search. In this article, we will discuss the model in detail and explain how Dropbox implemented this latest feature on its existing search infrastructure. Before that, let’s take a look at a simple image search problem: For any image search problem, to find a relevance function, we need a (text) query ‘q’ and an image ‘j,’ which returns a relevance score ‘s,’ indicating how well the image matches the search query. s = f(q, j) … (1) When a user searches for images, the above function is made to run on all their pictures, and it returns those images that produce a score above a threshold. In the case of Dropbox, it has built this function using two ML techniques — accurate image classification and word vectors. Image classification & word vectors In image classification, the image classifier reads an image, outputs a scored list of categories that describe its contents, and higher scores indicate a higher probability that the image belongs to that category. For instance, categories can be classified as Specific object in the image (a tree, or a person, etc.) Overall scene descriptions (outdoors, wedding, seminars, etc) Characteristics of the image (black-and-white, dark background, blue sky, etc.) Today, image classification in machine learning has improved a whole lot. From the latest self-supervised models to large datasets like Open Images or ImageNet, and easy-to-use libraries and frameworks such as Google’s Tensorflow and Facebook’s PyTorch, researchers have built image classifiers that can recognise thousands of categories accurately. Example of Image classifier results for a typical unstaged image (Source; Dropbox) While image classifiers let users understand what’s in an image, this isn’t enough to enable search. To further enhance this, Dropbox has used ‘word vectors.’ Word vectors are nothing but vectors of numbers representing the meaning of a word and usually defined as jc, where ‘c’ represents the number of categories (several thousand). Citing Mikolov et al.’s 2013 word2vec paper, Dropbox said word2vec assigns a vector to each word in the dictionary, and words with similar meanings will have vectors close to each other. Dropbox seems to have taken inspiration from this research paper for its image search machine learning architecture. ML architecture Dropbox has used the EfficientNet network image classifier, trained on the OpenImages dataset, which roughly produces scores for about 8500 categories. “We have found that this architecture and dataset give good accuracy at a reasonable cost,” claimed Dropbox. Besides these, Dropbox is also using Tensorflow to train and run the model. Also, it is using the pre-trained ConceptNet Numberbatch word vectors. “These give good results, and important to us they support multiple languages, returning similar vectors for words in different languages with similar meanings,” said Dropbox, stating that this makes supporting image content search in multiple languages easy. For instance, word vectors for ‘dog’ in English and ‘chien’ in French are similar. Dropbox claimed that it could support search in both languages without having to perform a direct translation. Similarly, for multi-word inquiries, the algorithm performs an alternate parse and runs the OR of the two parsed queries. For example, the query ‘beach ball’ becomes (beach AND ball) OR (beach ball). The results show both. The above model was applied on existing Dropbox’s Nautilus search engine instead of instantiating ‘j’ for each query, which could have billions of entries and needs to be updated whenever a user deletes, adds or modifies an image. Nautilus consists of a forward index that maps each file to metadata (the filename) and the full text of the file. The text-based search for the same would look something like this: Search index contents for text-based search (Source: Dropbox) With the latest image content search architecture, Dropbox can use the same system to implement image search algorithms. For instance, in the forward index, each image’s category space vector jc can be stored. In contrast, the inverse index for each category posts a list of images with positive scores for that category. It looks like this: Search index contents for image search (Source: Dropbox) Is this scalable? Dropbox believes the ‘text-search’ approach is still expensive in terms of storage space and query-time processing. “If we have 10,000 categories, then for each image we have to store 10,000 classifier scores in the forward index, at the cost of 40 kilobytes if we use four-byte floating-point values,” explained Dropbox. The classifier scores are rarely zero, and will be added to most of those 10,000 posting lists. In other words, for many images, the index storage would be larger than the image file. However, Dropbox said many near-zero values could drop to get a much more efficient approximation in the case of ‘image search.’ The company said the storage and processing savings are substantial. Here’s a comparison between ‘image search’  and ‘text search’ Instead of 10,000-dimensional dense vectors, the system stores sparse vectors with 50 nonzero entries in the forward index. A sparse vector is a matrix in which most of the elements are zero. In this case, about 50 two-byte integer positions and 50 four-byte float values require about 300 bytes. In the inverted index, each image is added to 50 posting lists instead of 10,000, at the cost of about 200 bytes. Therefore, the total index storage per image is 500 bytes instead of 80 kilobytes. In terms of query time, the image categories have ten nonzero entries, and only need to scan 10 posting lists, roughly the same amount of work as that of text queries. That, in a way, gives a smaller result set, which can score more quickly. Both indexing and storage costs are reasonable, while query latencies are on par with those for text search. In other words, the user can run both text and image searches in parallel, and the search engine would show the complete set of results together as fast as a text-only search.","excerpt":"Lookin up photos in the cloud can be cumbersome and daunting, especially if you have forgotten the filename. To simplify the experience, file hosting platform Dropbox recently launched a new feature to make image search seamless. The feature, available for both enterprises and individual customers, allows to search all images, OCR-based search for images of […]","categories":["AI Features"],"tags":[],"author_name":"Amit Naik","publish_date":"2021-05-19T13:00:00","publication_year":"2021","word_count":1196,"keywords":["Go","machine learning","TPU","AI","PyTorch","ML","RAG","Aim","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","TensorFlow","PyTorch","RAG","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-dropboxs-image-search-works\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":38372,"title":"How nTalents.ai AI-based Interviewing Platform Is Taking Away The Bias In Hiring Process","content":"Bengaluru-based nTalents.ai  has made automated interviews for corporate sales team recruitment a reality, with their easy to use online interview platform. Incubated at IIM Bangalore, the startups uses AI and other deep learning capabilities to make hiring easy for their numerous corporate clients. Founded by Deepika Anu, Mayank Sharma and Varun Narula who were previously working as a data scientist and were also involved in the hiring process, the trio gained first-hand knowledge about the hardships that corporates face while hiring. “Three of us have spent more than 10 years in the industry, all of us back then were involved in interviewing new recruits for our teams. The process became very repetitive and inaccurate in the long run. We brainstormed on how we can build something that can be much more accurate and comprehensive,” says Deepika Anu, Marketing Head of nTalents.ai It was during their attempt to crack standardised exams like TOEFL and IELTS when the idea of automated tools for the interview process came into being. Having used these exam’s comprehensive audio-interactive test to quantify English Communication, the trio decided to replicate the model for identifying few skill sets in technical professions before the actual personal interviews, thus helping the recruiters to identify the right set of talent. “An estimated 80% of the interview time (read ‘time of senior panelists’) is being spent on ineligible profiles and it takes another level of expertise to find few good candidates among all that noise. Through nTalent’s audio-visual simulated test of the candidate, the platform generates a detailed skill set report even before the candidate has even had the first interaction,” Anu explains. How are skillsets determined? The platform extensively leverages artificial intelligence to identify the right skill sets for providing real-time updates. While Python, SQL, JavaScript and HTML have been used as the programming language. The platform depends on WebRTC for real-time communication and MySQL as its database. Further, it relies on Ubuntu – Bionic Beaver for the application server and its GPU processing server is on AWS-Cloud Nvidia GPU. It further combines the best practices of interviewing and behavioural assessments into one integrated mechanism using their robust machine learning algorithms and video interview data. “The questions we ask are very specific to the exact job role that the candidate has applied for and captures the real skills of performing that task beyond what is written on the resume,” Anu explains. Further, AI has been used at the core of the platform to analyse data which is produced as an outcome from the platform’s audio-visual performance tasks, situational judgment test, role plays, case studies, and game-based simulations. The outcome of which is the easy recruitment process as the algorithms eliminate candidates that are not best suited for the position and through comprehensive insights, a pool of candidates with the highest potential are identified. Till now, the platform has held more than 25,000 assessments and to the question of how different they from their competitors, Anu says “We have become an expert at finding a needle from a haystack. Our tests are customised for various job roles in different industries, the score is also bench-marked against a relevant set of candidates.” The Road Ahead Since its inception in 2017, startups have grown substantially and has won several recognitions. In the same year, it received a grant of   5 Lakhs from  Government of India for exemplary innovation in the field of entrepreneurship. Over the years it has also added worked with several big clients including the likes of Religare Insurance, Emcure Pharma, Brinton Pharma, Peps Mattresses, to name a few. “We grew our average ticket size by more than 10 times in the last 2 quarters and raised our MoM Revenue to more than 5 lakhs all with organic marketing,” Anu says pointing out their growth. In the future, the startup hopes to revamp the traditional HR processes through their architecture and wants to bridge the gap between reliable tool to build the foundation of next-gen talent management While focussing on customer experience, the startup also aims to expand to banking, FMCG and automobile sectors. “We are concentrating on these areas as the growth rate of employee base and attrition in teams is on the rise. We are confident of growing at more than 5 times the present growth rate in the next financial year,” Anu concludes.","excerpt":"Bengaluru-based nTalents.ai  has made automated interviews for corporate sales team recruitment a reality, with their easy to use online interview platform. Incubated at IIM Bangalore, the startups uses AI and other deep learning capabilities to make hiring easy for their numerous corporate clients. Founded by Deepika Anu, Mayank Sharma and Varun Narula who were previously […]","categories":["AI Hirings"],"tags":["AI bias","Bengaluru","Hiring","Startups"],"author_name":"Akshaya Asokan","publish_date":"2019-04-29T05:41:36","publication_year":"2019","word_count":724,"keywords":["artificial intelligence","machine learning","AWS","AI","ML","Hiring","RAG","Python","Aim","deep learning","Startups","Bengaluru","R","AI bias"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","Aim","RAG","AWS","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/how-ntalents-ai-ai-based-interviewing-platform-is-taking-away-the-bias-in-hiring-process\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085399,"title":"7 Useful Apps Built Using GPT-3","content":"ChatGPT, the OpenAI chatbot that has caused a virtual whirlwind of sorts since November-end, was trained on GPT-3.5, one of the most well-known large language models (LLMs) – an upgraded version of GPT-3. GPT-3 has 175 billion parameters, making it one of the largest language models ever created, besides PaLM, LaMDA, Gopher, MT-NLG and others. These LLMs can be used to generate human-like text and perform various tasks. Besides ChatGPT, several useful applications have been powered by the language model to perform a series of tasks ranging from enhancing customer services, translating languages, summarising texts and so much more. The benefits of using GPT-3 to build an application or a tool include a faster turnaround of the task, multiple use cases, multilingual interpretations and minimal errors, among others. We give you six useful applications built using GPT 3\/GPT-3.5 that you must try. CharacterGPT Data science research studio Alethea.AI recently unveiled CharacterGPT, which generates characters from text. It claims to be the world’s first multimodal AI system to create interactive characters from a description in natural language. The AI characters can be used for various use cases like digital twins, digital companions, virtual assistants, etc. Users can create realistic, life-like AI characters with custom personalities and intelligence with CharacterGPT. There are countless uses for this AI system for both companies and artists. Without writing a single line of code, CharacterGPT claims to produce authentic characters that bring your stories, games, or Metaverse worlds to life, engage consumers and prospects in dynamic discussions, and give an engaging user experience. Jasper.ai Powered by GPT-3, Jasper is a platform that helps produce high-quality engaging content. It uses advanced AI algorithms to generate content based on input, thereby helping the user produce a large volume of content with minimal intervention in a quick turnaround time. Jasper can also be easily integrated with content management systems helping in publishing and distributing content. Replit Replit is a one-stop platform for creating and sharing software. You can write your code and host it all in the same place. Replit’s cloud-based infrastructure lets one code from anywhere and on any device. In addition, Replit offers a customisable workspace that allows users to set up their programming environment based on their needs. It also offers the possibility of installing and using various libraries and frameworks. It has different collaboration features, such as live coding and chat so you can work on projects in real time. It also extends its support to a host of programming languages that include JavaScript, Ruby and Python amongst others making it very useful for developers. Debuild You don’t have to be an expert coder to build your website. With Debuild you can build your web app within seconds using a brief English description. Powered by GPT-3, the tool helps you create unique web applications making it easy for anyone to make a good website. Pictory.ai Pictory.ai is a platform that helps you automatically create short, highly-shareable branded videos from your long-form content. Users can create videos from Zoom, Teams, webinar recordings, scripts, and blog posts. Using AI, the application extracts key points and themes from the content and creates interesting videos with stock footage, music, and voiceovers. It also offers the benefits of being able to add captions to videos automatically thereby helping increase their accessibility and reach on social media platforms. PolyAI PolyAI develops a machine learning platform for conversational AI. It builds enterprise voice assistants that carry on natural conversations with customers to solve their problems. Some of the ways in which PolyAI’s voice assistant technology can be used to solve customer problems include providing fast and accurate customer responses, offering personalised recommendations and automating routine tasks. Auto Bot Builder Auto Bot Builder is a powerful tool that harnesses the power of GPT-3 to automatically and effortlessly build advanced chatbots tailored to enterprise requirements. It leverages the LLM and fine-tunes it using proprietary enterprise knowledge base and domain expertise – resulting in a chatbot specialised to an enterprise, unlike ChatGPT, which is a general-purpose chatbot. This enables enterprises to incorporate the latest advances in AI technology, in a no-code environment, into their conversational experiences to substantially enhance their customer engagement.","excerpt":"The benefits of using GPT-3 to build an application or a tool include a faster turnaround of the task, multiple use cases, multilingual interpretations and minimal errors","categories":["AI Trends"],"tags":["AI Tool","ChatGPT","Gopher","GPT-3","LaMDA","MT-NLG","OpenAI","PaLM","Replit"],"author_name":"Aparna Iyer","publish_date":"2023-01-18T16:00:00","publication_year":"2023","word_count":697,"keywords":["GPT-3","Replit","ChatGPT","data science","PaLM","machine learning","OpenAI","AI","chatbots","virtual assistants","RAG","MT-NLG","Aim","Gopher","multimodal AI","LaMDA","AI Tool"],"extracted_tech_keywords":["AI","machine learning","data science","multimodal AI","ChatGPT","OpenAI","Aim","RAG","chatbots","virtual assistants"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-useful-apps-built-using-gpt-3\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042960,"title":"DeepMind’s Scientist Sparks Fresh Debate Over AI Ethics","content":"A near-perfect amalgamation of robotics and artificial intelligence robot, Sophia, developed by Hanson Robotics, sent shivers down the spine when she sarcastically replied to a question in an interview with “Okay, I’ll destroy humans”. Sophia was displaying humour and sarcasm, but that single comment erupted adverse reactions against the self-learning of machines. People saw her ‘rouge similarity to humans’ as a threat. The concern around ‘machines turning onto humans’ is the most prominent ethical challenge artificial intelligence faces today. Assistance, innovation, decisions, fraud detection, and crime oversight– a scientific researcher would explain these when asked about the benefits of AI. However, Raia Hadsell– Researcher at DeepMind, believes that the same researchers would appear hesitant when asked about risks and ethical issues associated with AI. At the recent Lesbians Who Tech Pride Summit, Raia spoke about issues plaguing the field of AI and actions that need to be taken to ensure its ethical deployment. Policymakers, lawyers, the Judicial system, ethicists, and philosophers play a critical role in maintaining ethics in AI. Still, it is more significant for researchers building models to explain how their innovations are ethically sound. The ethical implications of every research should shed light on its scope of threat due to self-intelligence. Raia shared her experience of how she received backlash in her community for bringing researchers into the fold of ethics. In 2020, Raia was invited to be one of the four program chairs of NeurIPS– the largest and the most prestigious AI conference in the world. Although there was exponential growth in the number of attendees and papers submitted over the last decade, no ethical guidelines were provided to the authors before last year. When Raia was invited to design the review process for the 10,000 papers that were expected last year, she initiated two significant changes. First, she hired a pool of ethical advisors to share informed feedback on reports that seemed to be controversial. Second, she required all authors to submit a broader impact statement with their work, discussing the potential positive and negative future impacts and mitigations, if possible. The idea of an impact statement was not new– it is a common requirement in scientific fields of medicine and biology. However, Raia’s community did not welcome the change. She said that she “Didn’t make a lot of friends,” and that there were some tears involved. However, later authors reached out to her to inform her how it was a valuable experience inspiring new directions for research. Google, DeepMind’s sister company, has recently been in troubled waters after firing co-leads Timnit Gebru and Margaret Mitchell. Google fired the duo over email for not rescinding its research about the risks of deploying large language models. This move by the tech giant raised a wave of backlash against Google for promoting unethical AI. Common ethical issues with AI-based technology The most commonly used AI-based technology is facial recognition. But, unfortunately, it is also the most error-prone technology. A study by Joy Buolamwine at the MIT Media Lab reports that facial recognition algorithms show errors based on skin colours. In people of caucasian descent, the algorithms work fine 99 percent of the time. On the other hand, in people of African descent, the algorithm shows an error rate of 35 percent. This error in algorithms can lead to racial conflicts. Researchers are also sceptical about AI applications in data mining, owing to users’ privacy concerns– they are concerned about the possible chances of data theft by advanced machine learning systems. Massive data compromises and leaks made from Facebook to holiday-booking websites make it to the headlines every year, putting the privacy of every internet user at stake. Questions arise from the ethical applications of AI– if self-learning robots are slaves to do humans’ bidding; if AI is an advanced conscience with synthetic life of its own; or if self-learning algorithms enjoy the same freedom as humans. Unfortunately, there is no concrete answer to these questions– they can range from philosophical to scientific to arguments based on law. Raia has shed light on a frequently ignored domain of AI while bringing it under the spotlight and building more support in the community regarding its implications to make it safer for humans.","excerpt":"A near-perfect amalgamation of robotics and artificial intelligence robot, Sophia, developed by Hanson Robotics, sent shivers down the spine when she sarcastically replied to a question in an interview with “Okay, I’ll destroy humans”.  Sophia was displaying humour and sarcasm, but that single comment erupted adverse reactions against the self-learning of machines. People saw her […]","categories":["AI Features"],"tags":["DeepMind","Ethical AI","Google","NeurIPS","Responsible AI"],"author_name":"Meenal Sharma","publish_date":"2021-07-06T09:00:00","publication_year":"2021","word_count":703,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","R","Ethical AI","innovation","programming_languages:Go","Responsible AI","NeurIPS","ViT","Google","fraud detection","DeepMind"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","fraud detection","R","Go","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deepminds-scientist-sparks-fresh-debate-over-ai-ethics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10060386,"title":"A deeper understanding of the Big Bang using AI","content":"Scientists created a big bang when they identified the presence of the high energy state of matter known as “Quark-Gluon Plasma”. Since the detection of such a high energy particle, it is proposed that the creation of the Universe was in this state in its early ages. However, while such a state can only be produced in high energy atomic nucleus collisions in the large hadron collider at CERN (Conseil Européen pour la Recherche Nucléaire), such processes can only be studied using high-end supercomputers and extremely complex simulations, and even then, the results are difficult to discern. With artificial intelligence being the obvious choice, Tu Wien, also known as Vienna Institute of Technology, has demonstrated that the use of neural networks can be successfully used for the arduous tasks of identifying the mathematical properties of particle physics observed in the process. However, before we delve into how one can get a better understanding of the Big Bang using neural networks, let us understand some of the terminologies. Quark-Gluon Plasma – A state of matter A quark-gluon plasma is a high energy state of matter where quarks and gluons interact with each other in a state of thermal and chemical equilibrium. Figure: Phase Diagram of QGP matter from CERN document server After the occurrence of the Big Bang, the entire Universe was in the state of QGP (Quark-Gluon Plasma). Theories predicting the presence of QGP were developed in the 1970s, and they proposed that after the Big Bang, QGP filled the entire Universe before matter existed. The presence of QGP was first detected at CERN in the year 2000, and since then, it has been studied to recreate and understand the Universe under high-energy-density conditions that were prevalent in the early stages of its creation. In this state, the matter was formed from the elementary degrees of freedom (quarks and gluons) at about 20 microseconds after the Big Bang. The challenges of gauge symmetry When simulating Quark-Gluon Plasma, one of the biggest problems that arise is the different ways in which the particles and the forces between them are represented in the form of mathematical equations. This variation is commonly referred to as gauge symmetries. Dr Andreas Ipp, from the Institute of Theoretical Physics at Tu Wien, explains with an analogy: “The basic principle behind this is something we are familiar with: if I calibrate a measuring device differently, for example, if I use the Kelvin scale instead of the Celsius scale for my thermometer, I get completely different numbers, even though I am describing the same physical state.”  Gauge symmetry depicts that mathematical equations that look completely different at first glance, in fact, describe the same physical state. Neural Networks to the rescue Simulating the state of Quark-Gluon Plasma requires an exorbitant amount of computing time and power. Even large supercomputers struggle with the calculations. This bottleneck is resolved by using neural networks to recognise and predict certain properties of the plasma state. With the help of convolutional neural networks (CNNs) that are used for image classification, one can observe the simulation and recognise the patterns of data. However, teaching the neural nets about gauge symmetry proves to be a difficult process. Dr David I. Müller, a postdoc at the Institute of Theoretical Physics at Tu Wien, provides a solution by stating that the better choice is designing the structure of the neural network by taking into account the variations of gauge symmetry. In this way, different representations of the same physical state also produce the same signals in the neural network. Figure: Representation of a convolutional neural network from researchgate.net With such methods, it is possible for CERN to fully stimulate atomic core collisions in the coming future as the CNNs provide a promising tool to describe the physical phenomena. A similar study was conducted by CDS’ Xinyue Zhang, Yanfang Wang, Wei Zhang, Yueqiu Sun, along with Siyu He, Center for Computational Astrophysics, Flatiron Institute, Gabriella Contardo and Francisco Villaescusa-Navarro, both of Flatiron Institute Center for Computational Astrophysics, and Shirley Ho, Flatiron Institute Center for Computational Astrophysics. Their paper describes a novel approach to cosmological evolution with the help of CNNs that relies on dark matter. Their key aim was to understand and define the physical parameters that led to the creation of the Universe. Another multi-university team of researchers from Japan came up with an AI system called “dark emulator” that was capable of predicting the structure of the Universe itself. By parsing through enormous troves of astrophysics data, the AI system simulates the creation of the Universe. The leading author on the team’s research paper, Takahiro Nishimichi, states: “We built an extraordinarily large database using a supercomputer, which took us three years to finish, but now we can recreate it on a laptop in a matter of seconds. I feel like there is great potential in data science.” Using this result, Takahiro hopes that it is possible to uncover one of the greatest mysteries of modern physics, which is to uncover what dark energy is.","excerpt":"Since the detection of Quark-Gluon Plasma at CERN, researchers have been troubled with vast amounts of simulation data that the best supercomputers are unable to compile. Researchers at the Vienna Institute of Technology propose that neural nets are the answer.","categories":["AI Features"],"tags":[],"author_name":"Kartik Wali","publish_date":"2022-02-11T13:00:00","publication_year":"2022","word_count":838,"keywords":["data science","artificial intelligence","programming_languages:R","AI","neural network","Aim","CNN","R"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","data science","Aim","R","CNN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-deeper-understanding-of-the-big-bang-using-ai\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10068989,"title":"How to deploy and monitor your Keras model with Comet?","content":"Comet is an online experimentation platform for artificial intelligence and machine learning used by Researchers and Data Scientists which provides an easy platform for tracking the ML models and comparing the results of the model on the server. Comet provides an excellent dashboard to visualize the various parameters of the model deployed in the interface and keeps track of the model internally. So this article provides a brief overview of how to deploy a Keras model on the online platform of Comet and make suitable interpretations from the model. Table of Contents Introduction to CometImplementing a Keras model from scratchLogging in to the Comet InterfaceIntegrating the Keras model into the comet interfaceObtaining predictions from comet interfaceMounting the model in the comet serverVisualizing the model deployed on the serverSummary Let’s first briefly discuss about the Comet. Introduction to Comet Comet is an online API that serves as an online platform to deploy machine learning or deep learning models and continuously validate all the stages of a model deployment lifecycle until production. The model on the platform can be deployed as public or can be limited for certain access for analyzing the parameters of the model deployed on their platform. With the concerned access, comet provides the flexibility to access the model at any time on the server and monitor the models deployed on their server. Are you looking for a complete repository of Python libraries used in data science, check out here. Implementing a Keras model from scratch Keras module provides useful datasets for deep learning model development and here we have used the MNIST Digits dataset for model development and deployment of the model developed on the Comet server. So initially the dataset was acquired and split into train and test and the split data was visualized using subplots. Later a simple Sequential model was built with appropriate input dimensions and activation functions. The dependent features of the dataset were suitably reshaped and the target variables were encoded according to the number of digits present in the MNIST dataset. After suitable preprocessing the model was compiled with categorical_crossentropy as the loss function and accuracy as the metrics to evaluate from the model developed. Also, the model was validated for its training and testing accuracy and loss. The complete steps to follow to implement a basic digit classification model for the MNIST dataset are shown below. import numpy as np import tensorflow as tf import matplotlib.pyplot as plt import random from tensorflow.keras.datasets import mnist from tensorflow.keras.models import Sequential from tensorflow.keras.layers import Dense, Dropout from tensorflow.keras.callbacks import EarlyStopping from tensorflow.keras.preprocessing import image from tensorflow.keras.optimizers import RMSprop from tensorflow.keras.utils import to_categorical (X_train,y_train),(X_test,y_test)=mnist.load_data() print('Number of training records',X_train.shape) print('Number of testing records',y_test.shape) plt.figure(figsize=(5,10)) for i in range(9): plt.subplot(3,3,i+1) num = random.randint(0, len(X_train)) plt.imshow(X_train[num], cmap='gray', interpolation='none') plt.title(\"Class {}\".format(y_train[num])) #plt.tight_layout() plt.axis('off') plt.show() X_test_original=X_test.copy() # preprocess and normalize the data X_train = X_train.reshape(60000, 28*28) X_test = X_test.reshape(10000, 28*28) X_train1=X_train\/255. X_test1=X_test\/255. ## Encoding the target variable y_train=to_categorical(y_train,10) y_test=to_categorical(y_test,10) With all these suitable preprocessing model building was taken up as shown below. Model Building model=Sequential() model.add(Dense(units=512,activation='relu',input_dim=784)) model.add(Dense(units=256,activation='relu')) model.add(Dense(units=125,activation='relu')) model.add(Dense(units=10,activation='softmax')) model.summary() model.compile(loss='categorical_crossentropy',optimizer='adam',metrics=['accuracy']) model_res=model.fit(X_train1,y_train,epochs=5,validation_data=(X_test,y_test)) Evaluating the model performance plt.figure(0) plt.plot(model_res.history['accuracy'],label='training accuracy') plt.plot(model_res.history['val_accuracy'],label='testing accuracy') plt.title('Accuracy') plt.xlabel('epochs') plt.ylabel('accuracy') plt.legend() plt.show() plt.figure(1) plt.plot(model_res.history['loss'],label='training loss') plt.plot(model_res.history['val_loss'],label='testing loss') plt.title('Loss') plt.xlabel('epochs') plt.ylabel('loss') plt.legend() plt.show() model.evaluate(X_train,y_train) model.evaluate(X_test,y_test) Logging in to the Comet Interface Firstly the comet module can be easily installed in the working environment using the pip commands and the module can be loaded using the import statements. So once when the module is successfully installed in the working environment a random name has to be given to the model that would be deployed onto the comet server using the init() built-in function. So a link will be generated in that particular run instance and a unique API Key will be generated which can be stored in the form of a string character in the working environment and utilized if required. The steps involved are shown below. import comet_ml from comet_ml import Experiment comet_ml.init('comet-keras') Initiating the comet interface keras_exp=Experiment(project_name='comet-keras',api_key=API_key) So now this comet interface can be used to train, test, and evaluate the parameters of the model on the comet server. Now the interface and the login with the appropriate API key can be validated using the display() function as shown below. keras_exp.display() Integrating the Keras model in the comet interface The integration of the model into the server starts with primarily hashing the training dataset using the log_dataset_hash() function and the experiment can be trained on the comet interface using the train() inbuilt function of the comet. The steps involved in hashing the dataset and loading the model are shown below. keras_exp.log_dataset_hash(X_train) with keras_exp.train(): model_res=model.fit(X_train1,y_train,batch_size=32,epochs=10,validation_data=(X_test1,y_test)) Evaluating model parameters on the Comet interface Now using the test() inbuilt function of the comet the model deployed on the server can be evaluated on various parameters. Here the steps involved to evaluate the training and testing loss and accuracy are shown below. with keras_exp.test(): test_loss, test_accuracy = model.evaluate(X_test1, y_test) train_loss, train_accuracy = model.evaluate(X_train1, y_train) metrics = {'test_loss':test_loss,'test_accuracy':test_accuracy,'train_loss':train_loss,'train_accuracy':train_accuracy} } Logging various metrics in the Comet Interface The log_metrics() inbuilt function is utilized to deploy or log in the metrics used for evaluating the model on the comet server. keras_exp.log_metrics(metrics) Now let us see how to obtain predictions from the model deployed onto the comet interface. Obtaining predictions from comet interface For obtaining the predictions from the comet interface the log_figure() function is used which provides outputs for the model’s prediction for each digit being classified as shown below. But for obtaining predictions the original TensorFlow model has to be used and to visualize the model performance the inbuilt function of Comet can be used. model_pred = model.predict(X_test1[:10]).argmax(-1) # remember our copy of x_test? This is where it comes into play again! # Since we flatted x_test during pre-processing, we need to preserve an unflattened version plt.figure(figsize=(16,8)) for i in range(10): plt.subplot(1, 10, i+1) plt.imshow(X_test_original[i], interpolation='nearest') plt.text(0, 0, model_pred[i], color='black', bbox=dict(facecolor='white', alpha=1)) plt.axis('off') keras_exp.log_figure('Predicted MNIST Digits from Comet Interface',plt) By clicking on the web link from the plots obtained the output of the model deployed on the server can be downloaded if required for documentation along with the REST api information. Now the experiment on the comet server has to be ended to visualize how the comet interface has monitored the model deployed on its server Logging out the model deployed on the server Now the experiment carried out on the server has to be ended using the end() inbuilt function. While ending the experiment on the server, several information pertaining to the Data, Metrics used to evaluate the model, and various other information will be provided as shown below. keras_exp.end() Mounting the model in the comet server The model can be saved in an HDF5 (h5) format and the model can be mounted onto the comet server using log_model() function which may facilitate using the model weights for any parameter evaluation or using the model weights for any other tasks. The steps involved to mount the model on the comet server are shown below. model.save('keras-comet-new.h5') keras_exp.log_model('Saved Keras model',\"\/content\/keras-comet-new.h5\") Visualizing the model deployed on the server The display() function of the comet can be used to visualize the parameters and to monitor the model functionality on the server. keras_exp.display() The entire code that is being used to deploy on the server can be seen in the code section of the server where the entire code is made visible to the person with access. All the experiments carried out on the server with its logging time and the parameters evaluated can be visualized in the metrics pane as shown below. Along with this, the hardware resource usage can also be visualized in the server as shown below. Summary So this article provides a brief overview of how to develop a Keras model and deploy it on the Comet server to continuously monitor the model. Easily interpretable reports and graphs are produced with respect to the model performance on the server and the model’s memory consumption for each of the hardware resources used in the premise can also be visualized using the system metrics. So by using the Comet server the model can be made available on any platform and utilized accordingly to monitor the model on the servers.","excerpt":"A detailed implementation of usage of Comet platform for deploying and monitoring a model.","categories":["AI Trends"],"tags":["Deep Learning","Keras","Machine Learning","Tensorflow"],"author_name":"Darshan M","publish_date":"2022-06-15T12:00:00","publication_year":"2022","word_count":1365,"keywords":["data science","NumPy","artificial intelligence","Keras","machine learning","AI","ML","Machine Learning","Ray","deep learning","Deep Learning","TensorFlow","Tensorflow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","Ray","TensorFlow","Keras","NumPy"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/__trashed-7\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":16099,"title":"How AI powered analytics tool by Adobe is capturing the way people use voice assistants?","content":"Img src- Adobe In a recent development by this American multinational computer software company, Adobe launched a new set of analytics tools powered by Sensei AI to keep a track of the data and capture online activities. With this launch, Adobe makes its intentions clear on downing into the conversational data and let the brands take advantage of it by improving targeting and conversions. By consuming data from Alexa, Siri, Google Assistant, Cortana and Bixby, the company is keen on capturing both user intent and contextual data—and track user actions to help brands reap the best benefits. Adobe, that had till now helped in bringing context to online and offline data, has now added voice analytics to get a better understanding on how people consume media and marketing via voice enabled devices. AI— a frequent at Adobe This is not the first time that Adobe has vouched for artificial intelligence. It had earlier launched Sensei, a framework of intelligent services tapping into artificial intelligence, machine Learning and deep Learning, that was launched to dramatically improve the design and delivery of digital services that form the core part of Adobe’s Cloud platform. This AI service by the company made a debut after being trained by massive amounts of data gathered from various resources to do things like visual search, auto lip sync, sentiment analysis, automated insights for digital advertising and more. Similarly, in its recent development too, the company announced that deep analysis of voice data would be complemented by artificial intelligence and machine learning in Adobe Sensei. It would help brands gain robust audience insights and recommendations, while automating the traditionally cumbersome, manual analysis. In addition to helping the companies to make sense of trends within massive data sets, it would allow them to take quicker action on these insights hence delivering experiences that delight customers while informing other touch points such as email and advertising. Capturing the voice based data market– There is no doubt that over the last decade, there has been an increase in adoption of the voice enabled applications for both mobile devices and other smart voice enabled consumer products. Be it Cortana, Siri, Alexa or the soon to be released HomePod, the consumption of voice enabled devices has grown up drastically. According to a report by Mary Meeker on Internet Trends 2017, 20% of mobile search queries in 2016 were made by voice. Even a report by Adobe suggests that voice assistants are poised to be the next tech disruptor. Given the stats and an increase in voice enabled data, the launch of virtual analyst by Adobe comes at a right time. Capturing on the fact that voice enabled devices and digital assistants allow consumers to engage through one of the most natural forms of communication, Adobe’s voice analytics capabilities would enable brands to deliver more personalized customer experience and create brand loyalty through voice based interfaces. The voice data from Amazon’s Alexa, Google Assistance, Samsung Bixby, Apple Siri, Microsoft Cortana, is what Adobe planning to analyse with its tools. From addressing the complexity in measuring voice interaction to additional data points such as frequency of use and actions taken after voice request is made, this newly added tool can measure it all. What is interesting to note is the fact that it has the ability to capture both the user intent ((“play me a song”) as well as specific parameters (“from The Beatles”), as the company notes. Bill Ingram, vice president of Adobe Analytics Cloud said during the launch that it is one of the most important trends in modern technology to quickly adopt newer ways of interacting with the content, such as mobile and video in the past. With an increase in voice data, Adobe will now extend insights from it across the entire customer journey. Creating a more relevant content- The company assures that the integration of voice analytics data with Adobe Marketing Cloud and Adobe Advertising Cloud results in a continuous and relevant digital interaction. “For instance, Adobe Target ensures that insights from a voice-enabled device can be automatically leveraged on other channels, while also delivering personalized responses to user queries using machine learning and predictive algorithms”, the company’s official statement says. Explaining it further, the company notes, a food enthusiast interacting with a travel app on Amazon Echo would get the most popular food destinations through her voice. Last word- With the launch of analytics tools for voice, it is quite clear that Adobe is serious about tapping into market potential for conversational AI and it surely has a competitive edge over others being a pioneer in AI driven efforts with Sensei. As a lot of players are tapping into developing voice assistants, Adobe is right on time at the right place with its recently introduced conversational analytics tools.","excerpt":"In a recent development by this American multinational computer software company, Adobe launched a new set of analytics tools powered by Sensei AI to keep a track of the data and capture online activities. With this launch, Adobe makes its intentions clear on downing into the conversational data and let the brands take advantage of […]","categories":["IT Services"],"tags":["Adobe AI","Voice Analytics"],"author_name":"Srishti Deoras","publish_date":"2017-07-06T05:59:31","publication_year":"2017","word_count":799,"keywords":["Go","machine learning","artificial intelligence","AI","sentiment analysis","Voice Analytics","Git","Adobe AI","RAG","deep learning","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","analytics","RAG","sentiment analysis","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/ai-powered-analytics-tool-adobe-capturing-way-people-use-voice-assistants\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10137977,"title":"From Small-Town to YC Success, Kastle’s AI Agents Shape the Mortgage Industry","content":"Y Combinator, or YC, the world’s largest startup accelerator, has accelerated funding for selective promising startups increasing to four batches this year. Interestingly, a number of startups founded by Indians or Indian-origin founders have been part of YC batches. One such startup from the recent YC Summer ‘24 batch is Kastle, founded by Indian-origin entrepreneurs, which aims to bring AI agents to the mortgage sector. Rishi Choudhary, the co-founder of Kastle, has turned his journey from a small town in Orissa to Silicon Valley into a tale of ambition and innovation in the mortgage industry. With a focus on leveraging AI for customer service in finance, Choudhary along with co-founder Nitish Poddar, started Kastle to bring in AI voice agents to automate and improve customer interactions. Kastle founders Rishi Choudhary and Nitish Poddar with YC partner Diana Hu. Source: Linkedin From Real Estate to AI for Mortgage Servicing “We’re building agents that can handle complex financial tasks and workers, and they all live in this castle, which is super secure,” explained Choudhury in an exclusive interaction with AIM, while discussing the origin of the name. Choudhary’s path to entrepreneurship began with an internship at Redfin, a tech-powered real estate brokerage in the U.S. There, he discovered the vast potential of the mortgage industry, which holds approximately $14 trillion in outstanding mortgage debt. “I learned about how old this industry is, and how big and central it is to the global financial system,” said Choudhury. His experience there made him realise the potential of the mortgage sector that was waiting to be disrupted by AI. After graduating with a degree in systems engineering from the University of Illinois, Choudhary moved to San Francisco, where he reconnected with his co-founder and CTO Poddar, who shared his entrepreneurial spirit. “We started out building software for real estate agents, but quickly pivoted to focus on mortgage loan officers,” he explained. This shift was a response to the crowded market and the unique challenges faced by mortgage professionals. He even spoke fondly about his YC experience. “The learning that happens in Y Combinator is immense. You’re surrounded by talented people who are facing similar challenges,” he said. This collaborative environment has been instrumental in shaping Kastle’s direction and growth. Concoction of LLMs Kastle utilises large language models (LLMs) to enhance customer interactions in the mortgage sector. The company self-hosts these models, fine-tuning them specifically for mortgage-related queries. “We use a structured workflow system that enables the AI to navigate conversations effectively,” Choudhary noted. The startup has fine-tuned Llama-8B for its AI model and is using Eleven Labs for text-to-speech. Their models are trained on extensive datasets, including mortgage-related documents and customer interactions, allowing them to respond accurately to a wide range of inquiries. The agentic workflow developed by Kastle allows for a high degree of customisation based on client needs. “We have built guardrails to ensure compliance, which significantly reduces the chance of hallucinations,” he added. Furthermore, Kastle aims to minimise latency in voice interactions, targeting a response time under one second to make conversations feel ‘superhuman.’ Choudhary emphasises the importance of this speed in creating a seamless customer experience. This allows borrowers to receive instant answers to their questions, whether they are inquiring about loan status, interest rates, or documentation requirements. Adoption of AI in Finance Leveraging AI and LLMs in the financial sector has seen increasing adoption. For instance, a Bengaluru-based startup, ‘OnFinance’ is using AI to bring solutions for BFSI sectors, where their proprietary LLM will help analysts, advisors and other companies. Similarly, Gnani AI, an Indian startup is building voice-based small language models in 12 Indic languages. The models will cater to banking, finance, security and insurance domains. However, Kastle looks to target the mortgage industry alone, a sector that has a huge market in the US. Choudhury observed the mortgage industry to be traditionally slow to adapt to technological advancements. Many mortgage loan officers still rely on outdated processes that can frustrate customers and lead to inefficiencies. Choudhary recognised this gap and set out to create a solution that would not only streamline operations but also enhance the customer experience. “Our mission is to empower mortgage professionals by providing them with tools that enhance their efficiency and enable them to focus on building relationships,” Choudhary stated. From the US To India? Choudhary’s ambitions extend beyond the U.S. market. Both he and his co-founder, who also hails from India, are keen on bringing their innovative solutions to their home country. “We want to take this company to India because our parents are there. All our families are there,” he remarked. The potential to expand into insurance, banking, and small business lending in India is part of a broader vision that seeks to adapt Kastle’s technology to different financial contexts.","excerpt":"“We want to take this company to India because our parents are there. All our families are there,” said Rishi Choudhary, co-founder of YC-backed AI startup, Kastle.","categories":["AI Features"],"tags":["Automation","Kastle","mortgage","Voice AI Agent","Y Combinator"],"author_name":"Vandana Nair","publish_date":"2024-10-09T18:12:56","publication_year":"2024","word_count":798,"keywords":["API","Rust","Voice AI Agent","AI","innovation","ML","Automation","Y Combinator","RAG","Aim","mortgage","GAN","Kastle","small language models","R"],"extracted_tech_keywords":["AI","ML","Aim","small language models","RAG","R","Rust","API","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/from-small-town-to-yc-success-kastles-ai-agents-shape-the-mortgage-industry\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061915,"title":"How to transition to data science roles from a non-analytics background","content":"According to AIM Research, the demand for data analytics roles has increased by 23% in 2022. Evidently, the explosion of the analytic job market has triggered a talent overflow to the domain. Now, even professionals from non-technical backgrounds are considering a switch, thanks to the glamour and attractive salaries that come with the data science territory. Northeastern University, Boston, has listed the average salaries of a data science professional in recent research: Big Data Engineer: USD 127,250–219,500 Database Manager: USD 108,000–183,000 Database Developer: USD 98,250–167,750 Database Administrator: USD 77,000–159,250 Data Analyst\/Report Writer: USD 81,750–138,000 Data Architect: USD 111,500–187,750 Data Modeler: USD 79,000–164,500 Data Scientist: USD 102,750–175,000 Data Warehouse Analyst: USD 77,750–$160,000 Business Intelligence Analyst: USD 85,750–178,000 Without proper guidance, transitioning from a non-analytical job profile to the world of data analytics can be bewildering. That said, the switch is a lot easier for professionals from IT, finance, UX, marketing and sales, HR backgrounds. Below, we have put together a list to help professionals from non-technical backgrounds segue into data science roles. Step 1: Get certified Full-time degrees are no longer a requirement to get a data analytics job. However, one does need a structural approach to learn the relevant skills, and project-based courses are your best bet. The benefits of taking such certification courses are: Mastering fundamentalsMentorshipRelevant certificatesCareer counselling Based on the type of certification, a course can last from 2 to 8 months. The best data analytics certifications to start off include: Cloudera Data Platform (CDP) Generalist CertificationDell EMC Proven Professional Data Science AssociateAWS Big Data Specialty CertificationAssociate Certified Analytics Professional (aCAP)SAS Certified Data ScientistIBM Data Science Professional Certificate Big enterprises use data analytics certifications as a filter to shortlist the most qualified candidates. Step 2: Create a portfolio Data analytics is a hands-on field, and real project experiences can help you stand out from the crowd. There are many ways to build your data analytics portfolio– choosing a data science course with a capstone project is one of them. Winning Kaggle or MachineHack competitions will boost your cred. Additionally, posting on code repositories like GitHub and working on open-source projects in the data domain can also help you get the prospective employer’s attention. Step 3: Upgrade your skill set One of the most essential skills to handle analyticals tasks is the ability to read, write and analyse data with code. Being fluent in SQL, R, Python, AWS, Azure, Java, C#, ETL etc can put your resume at the top of the heap. Source: Core.co Step – 4: Network Networking is key to landing the right job. Building portfolios in GitHub, Kaggle etc is also one way to get the word out. Reaching out to data science professionals on platforms like LinkedIn and Discord, being part of the vibrant open-source data science communities, attending workshops etc, can also give you visibility. More often than not, proper guidance can be the difference between success and failure in the data science domain. Keeping track of the data analytics mavens can bring you up to speed on the latest developments and help gain insights into the data analytics ecosystem. Some of the noteworthy data professionals influencing the world of data are: Ronald Van Loon – Financial markets trainer Merv Adrian – Vice president at Gartner Marcus Borba – Global thought leader and influencer in AI, ML Kirk Borne – Chief science officer at DataPrime Inc. Carla Gentry – Chief data scientist at Analytical Solutions","excerpt":"Full-time degrees are no longer a requirement to get a data analytics job.","categories":["AI Features"],"tags":["Big Data","Business Analytics","computer science","Data Analysis","data analyst","data and analytics","Data Science","data visualization","Machine Learning","Machine Learning Algorithms","Predictive AI","predictive analytics"],"author_name":"Kartik Wali","publish_date":"2022-03-02T11:00:00","publication_year":"2022","word_count":572,"keywords":["Data Analysis","Azure","R","Predictive AI","data science","data and analytics","RAG","analytics","Data Science","Big Data","Machine Learning Algorithms","AWS","AI","ML","Machine Learning","data visualization","Business Analytics","computer science","Python","Aim","data analyst","predictive analytics"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","AWS","Azure","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-transition-to-data-science-roles-from-a-non-analytics-background\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10116732,"title":"Stability AI Releases Stable Video 3D, Generating 3D Videos from Single Images","content":"Stability AI yesterday announced the release of Stable Video 3D (SV3D), a generative AI model that creates 3D videos from a single 2D image.  Stability AI, is an open source generative AI firm that develops models for a variety of applications. The model, based on Stable Video Diffusion, aims to advance 3D technology by delivering improved quality and multi-view consistency compared to previous models like Stable Zero123. SV3D comes in two variants: SV3D_u, which generates orbital videos from single images without camera conditioning, and SV3D_p, which accommodates both single images and orbital views, allowing for 3D video creation along specified camera paths. “Stable Video 3D leverages its multi-view consistency to optimise 3D Neural Radiance Fields (NeRF) and mesh representations to improve the quality of 3D meshes generated directly from novel views,” Stability AI stated in their blog post. The model is available for both commercial and non-commercial use. Commercial users require a Stability AI membership starting at $20 per month, while non-commercial users can download the model weights from Hugging Face. Varun Jampani, lead researcher at Stability AI, said, “Stable Video 3D is a valuable tool for generating 3D assets, especially within the gaming sector. Additionally, it enables the production of 360-degree orbital videos, which are useful in e-commerce, providing a more immersive and interactive shopping experience.” SV3D’s release follows other recent advancements in AI-generated video, such as OpenAI’s Sora, Runway ML , and Google Dream Fields. However, SV3D differentiates itself by focusing on generating 3D videos from single images rather than relying on text inputs. As AI continues to evolve, models like Stable Video 3D showcase the potential for transforming 2D content into immersive 3D experiences, with applications spanning gaming, e-commerce, and beyond. Stability AI has been on a roll, releasing several innovative AI models in recent months. Last month, the company released Stable Diffusion 3, its most capable text-to-image model with improved performance in multi-subject prompts, image quality, and spelling abilities. They also launched Stable Cascade, a text-to-image AI model designed for efficiency on consumer hardware.","excerpt":"Stable Video 3D blurs the lines between 2D and 3D content creation.","categories":["AI News"],"tags":["Generative AI","Stability"],"author_name":"K L Krithika","publish_date":"2024-03-19T16:16:47","publication_year":"2024","word_count":338,"keywords":["Go","Stability","Hugging Face","OpenAI","AI","ML","RAG","Aim","stable diffusion","generative AI","Generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Aim","Hugging Face","RAG","R","Go","stable diffusion"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/stability-ai-releases-stable-video-3d-generating-3d-videos-from-single-images\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":50245,"title":"Google Acquires Enterprise Software Company CloudSimple","content":"Tech giant Google this week announced that it has acquired CloudSimple, a noted provider of secure, high performance, dedicated environments to run VMware workloads in the cloud. This acquisition builds on its existing partnership with CloudSimple, allowing them to accelerate a fully-integrated VMware migration solution. Many enterprises are using VMware in their on-premises environments to run a variety of workloads: business applications such as ERP and CRMdatabases such as Oracle and SQL Serverdevelopment and test environmentsvirtual desktopsreporting and analytics systems As part of their IT modernization initiatives, we hear frequently from enterprise customers that they need a simple way to migrate those workloads to the cloud. To put it simply: they want to be able to run what they want, where they want, and how they want — so they can leverage existing investments with as little toil as possible. Through their existing partnership with CloudSimple, their customers can migrate their VMware workloads from on-premises data centers directly into Google Cloud VMware Solution by CloudSimple, while also creating new VMware workloads as needed. Their apps can run exactly the same as they have been on-premises, but with all the benefits of the cloud, like performance, elasticity, and integration with key cloud services and technologies. And best of all, customers can do all this without having to re-architect existing VMware-based applications and workloads, which helps them operate more efficiently and reduce costs, while also allowing IT staff to maintain consistency and use their existing VMware tools, workflows and support. Ajay Patel, senior vice president, Cloud Provider Software Unit at VMware, said in a statement, “We look forward to continuing our partnership with Google Cloud as they welcome CloudSimple, a VMware Cloud Verified partner. Our partnership with Google Cloud enables our mutual customers to run VMware workloads on VMware Cloud Foundation in Google Cloud Platform. With VMware on Google Cloud Platform, customers will be able to leverage all of the familiarity of VMware tools and training, and protect their investments, as they execute on their cloud strategies and rapidly bring new services to market and operate them seamlessly and more securely across a hybrid cloud environment.”","excerpt":"Tech giant Google this week announced that it has acquired CloudSimple, a noted provider of secure, high performance, dedicated environments to run VMware workloads in the cloud. This acquisition builds on its existing partnership with CloudSimple, allowing them to accelerate a fully-integrated VMware migration solution. Many enterprises are using VMware in their on-premises environments to […]","categories":["AI News"],"tags":["Cloud Computing","Software","Software Development"],"author_name":"Prajakta Hebbar","publish_date":"2019-11-19T13:08:38","publication_year":"2019","word_count":354,"keywords":["Go","API","programming_languages:R","AI","ML","RAG","Cloud Computing","analytics","SQL","cloud_platforms:Google Cloud","Software Development","Software","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","SQL","Go","API","cloud_platforms:Google Cloud","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-acquires-cloudsimple\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10051907,"title":"Facial Recognition Is Steadily Entering The Stage Of Large-Scale Deployments","content":"From enforcing COVID-19 quarantines to analysing protesters attending political rallies, facial recognition technologies have found usage beyond commercial interests. Globally, governments are high on investing in the adoption of facial recognition technologies, especially in China and the US. On Friday, Moscow’s metro network launched a fare payment system using facial recognition technology at its 240 stations. For a city with 12.7 million people, it has one of the world’s largest video surveillance systems. Commuters have the option to submit their picture, link it to their transport and bank cards and use ‘Face Pay’ for payment of the fare. Recently, Australia too expanded a program leveraging facial recognition to enforce COVID-19 safety precautions. Big Money According to a report by MarketsandMarkets, the global facial recognition market size is expected to grow to USD 8.5 billion by 2025. China has the most extensive public surveillance system, making it the biggest buyer of CCTV cameras (626 million by 2020) and facial recognition technologies. Use cases such as public security, biometric sign-in, healthcare services, and eLearning platforms, among others, are expected to be deployed at a large scale. The technology even helps governments solve various criminal investigations and rapidly identify offenders. The Pandemic Waves During the pandemic situation, contactless verification technologies such as facial recognition became highly important. However, since facial recognition relies on the data points on a person’s image and face-masks block a lot of identifying information, the algorithms started failing. Already, the algorithms could be fooled with bad angles and improper lighting; masks made matters worse. In early 2020, the face recognition technology registered error rates between 5% to as high as 50%. Tech giants like Apple got to developing algorithms that work while people are wearing masks. Initially, they came up with an update that could detect a mask and prompted the user to enter their passcode. Japan’s NEC in January announced a system that was 99.9% accurate over masks. The Development of the Facial Recognition technology since 1960Facial recognition technology has gone over numerous changes since its inception. Coined in 1960, identities were automatically differentiated based on the manual marking of various “landmarks” on the face, like the placing of the eyes and the mouth. Later, the work was extended and standardised to include 21 specific subjective markers like hair colour and lip thickness to automate the recognition. In the late 80s, scientists applied linear algebra to the problem of facial recognition and formed the Eigenface system. It was in the early 1990s when development for the technology for commercial uses was initiated. In 2006, the US government supported the Face Recognition Grand Challenge (FRGC) to promote and advance face recognition technology. Here, 3D face scans, high-resolution face images, and iris images were used in the tests to make the technology 100 times more accurate.It was not until 2010, when the consumer experienced face recognition technology that was introduced by Facebook to identify people whose faces featured in the photos of their users. The major breakthrough that we see now happened when Apple launched the iPhone X that could be unlocked with FaceID. Post that, the technology is being used by airlines, airports, border controls, stadiums, transport hubs, mega-events, concerts, and conferences, among others. Errors can make lives miserable While it sounds like the go-to technology for biometrics, experts worry that with more and more governments using it, the drawbacks and failures of the technology can prevent needy people from getting benefits. Also, facial recognition has been proved to be less accurate for people of different races. While tech companies promise a 90-95% accuracy rate, they claim that the government human resources only need to take care of the remaining 5-10%. This is also where experts worry, as most governments worldwide don’t consider 5% as an issue. Not just governments but also private companies are unable to handle the situation that comes with the failures of the technology. Two unions have taken legal action against Uber, alleging that the firm has unfairly dismissed drivers based on their racially biased software used to verify drivers’ identities. In fact, the Independent Workers’ Union of Great Britain (IWGB) has also asked Uber to completely scrap the use of the technology that causes indirect racial discrimination. Experts believe that if AI, in its automated decision-making, learns discriminatory biases, then the whole purpose of technology fails. Unchecked face recognition tools carry the potential to further push away the marginalised groups. Wrapping up Facial recognition is also not privacy-preserving. Face Ids are personally identifiable information (PII) that need permission to collect, store and process. This is especially under the General Data Protection Regulation (GDPR) Act, and many people might not choose to use facial recognition. One of the limitations that technical experts are working on is the ability of the face recognition algorithms to recognise changes in facial appearances throughout a person’s life from childhood to old age. The pandemic has made it quite clear that facial recognition technology is here to stay and will even advance further.","excerpt":"Facial recognition technology is being leveraged way beyond unlocking our phones; it is aiming to identify every person on the planet, for good or bad.","categories":["AI Features"],"tags":["Apple","face recognition","Facial Recognition","Uber"],"author_name":"Meeta Ramnani","publish_date":"2021-10-19T18:20:00","publication_year":"2021","word_count":833,"keywords":["Go","API","Facial Recognition","programming_languages:R","AI","Apple","R","programming_languages:Go","RAG","Aim","face recognition","Uber"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/facial-recognition-is-steadily-entering-the-stage-of-large-scale-deployments\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":14240,"title":"Paytm invests in AI and Big Data based online healthcare startup QorQl to boost its business","content":"India’s leading digital payments and commerce platform, Paytm has invested an undisclosed amount on the Noida-based online healthcare startup QorQl, which was started in 2014. It currently offers its solutions in Delhi NCR and Aligarh, and has plans of expanding into several Tier II markets across India. With a customer base of more than 60,000, QorQl uses artificial intelligence and big data to improve the productivity of doctors, while also helping the patients to manage their health and existing ailments better. The company’s Qcare solution gives doctors an access to patient’s health and clinical data, hence ensuring better health care services. Sanjay Singh, founder and CEO of the company has been previously associated with Paytm’s parent One97 Communications, for about two years, and with his current venture he plans to take the healthcare industry in India to an altogether new height. The platform is already being used by over 100 doctors to offer the services to the consumers. In a country like India where there is a shortage of information on consumers and healthcare professionals, the platform by QorQl come as fresh breath of air as a data collection point and helping in improving the overall care process with AI and advanced analytics. Sudhanshu Gupta, VP, Paytm, was reportedly quoted as saying “QorQl’s consumer offerings will be integrated on Paytm’s platform so that it can reach a wider range of users.” Sanjay Singh sees the funding as a way to pursue his vision to democratize healthcare access by connecting healthcare applications leveraging AI and big data to collect, integrate and interpret data for all users. It would be a great move towards keeping the user’s healthy. QorQl sees this as an opportunity to endorse their technology and product vision, and to market their solutions quicker and on a higher scale. Paytm has been instrumental in backing up a lot of startups recently, logistics analytics startup LogiNext, being their latest venture, where Paytm invested $10 mn in the company. Paytm has planned up to invest $150 million in Indian tech start-ups over the next few months.","excerpt":"India’s leading digital payments and commerce platform, Paytm has invested an undisclosed amount on the Noida-based online healthcare startup QorQl, which was started in 2014. It currently offers its solutions in Delhi NCR and Aligarh, and has plans of expanding into several Tier II markets across India.                  […]","categories":["AI News"],"tags":["big data processing interview"],"author_name":"Srishti Deoras","publish_date":"2017-04-17T05:06:50","publication_year":"2017","word_count":345,"keywords":["big data","funding","artificial intelligence","AI","Git","RAG","big data processing interview","ViT","analytics","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","RAG","R","Git","big data","ViT","startup","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/paytm-invests-ai-big-data-based-online-healthcare-startup-qorql-boost-business\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":68003,"title":"How CodeChef&#8217;s Acquisition Puts Focus On The Importance Of Developers In India","content":"In recent news, it has been announced that Mumbai-based non-profit competitive programming platform, CodeChef has been acquired by a Bangalore-based ed-tech startup Unacademy. Established in 2009 by serial entrepreneur, Bhavin Turakhia as part of Directi, CodeChef platform has been designed to help developers and programmers to enhance their skills by practising coding and participating in programming contests. Alongside, this platform is also used by businesses for their technical hiring, branding and developer adoption needs. Some of the key clients CodeChef is currently working with include Alibaba, Samsung Research Institute, ShareChat, Tech Mahindra, UpGrad, and Ericsson. With the recent development, the company is now moving its custodianship from Directi to Unacademy with a vision to increase its efforts in creating the best coding platform along with services to provide high-quality free programming courses for enthusiasts and learners. In a recent tweet, Gaurav Munjal, the CEO and the co-founder of Unacademy has confirmed the news by stating — “Codechef is one of the best platforms in the world for programmers and their vision is aligned to what we do here at Unacademy Group.” Codechef will be a part of Unacademy Group moving forth. Codechef is one of the best platforms in the world for programmers and their vision is aligned to what we do here at Unacademy Group.— Gaurav Munjal (@gauravmunjal) June 18, 2020 This development will move the founder of CodeChef, Turakhia directly under Unacademy and has been positioned in the board of the company. In a recent tweet, Turakhia also announced, “Going forward the leadership of CodeChef will include @gauravmunjal along with myself and @anupkal.” I am super excited to announce this new chapter in @codechef 's journey. We are moving the custodianship pf Codechef from Directi to @unacademy. Going forward the leadership of codechef will include @gauravmunjal alongwith myself and @anupkal. Details –https:\/\/t.co\/zK9DJp7Vjh ..— Bhavin Turakhia (@bhavintu) June 18, 2020 Anup Kalbalia, who will still be heading the company has also shared his views in his blog post, where he assured their audience that this acquisition would not change the values and mission CodeChef has been working on. He wrote, “All our existing programs, contests, etc. will continue to remain exactly as they are, and will only grow better in our new home.” “This transition is akin to us shifting houses – we have new caretakers and sponsors, but we will continue to remain who we are,” wrote Kalbalia. How This Acquisition Can Improve The Developer Ecosystem In India CodeChef, with its massive user base, has always been working towards helping programmers and developers community to enhance their coding and problem-solving skills. With its International Collegiate Programming Contest (ACM-ICPC) regionally and International Olympiad in Informatics nationally, the company has been offering its platform and services free of cost for the learners in India. CodeChef has also been renowned in the programming industry for its Go-For-Gold initiative to encourage Indian programmers to win gold in the ICPC world finals. And by partnering with Unacademy, the company is aiming to expand the platform to the massive learner base the edtech startups own. Alongside, Unacademy has been witnessing a big boom amid this COVID-19 crisis with revenue higher than the previous years. Munjal has stated in his earlier Tweet that the company has seen an 82% growth in revenue, a 10x higher than April 2019. And therefore, this partnership with CodeChef can provide opportunities for students and learners to get a hold of “expanded courses to learn programming at all levels.” The advent of this pandemic outbreak and the dependency on app economy has drastically increased the demand for advanced programmers for businesses and with CodeChef participation with Unacademy will open up opportunities for anyone interested in coding to get their hands on free programming courses on Unacademy. Also with the majority of the IT skills getting obsolete due to the automation boom businesses are currently looking to hire programmers only with advanced skill sets and strong knowledge in multiple programming languages, which again bring in the necessity for upskilling in order to keep up their relevance. Also Read: Why Do Large Companies Open Source Their Tech? Furthermore, with Unacademy recently receiving a massive round of funding from Facebook, General Atlantic, Sequoia India and Nexus Venture Partners, among others. It has been aiming to democratise education by making it accessible to everyone. And this partnership will allow learners from tier 1 and tier 2 cities to get access to top educators, affordable quality content and best learning experience for students, which in turn will enhance the developers’ community of the country. With its more than a decade of expertise in conducting coding challenges and hackathons, CodeChef has also started free community initiatives for enhancing the ecosystem for women programmers. With the help of Unacademy platform, these women programmers can quickly improve their programming skills and can also gain expertise in the industry. Equally, it also has a ‘CodeChef for schools’ program that aims to start programming at the core for students in Indian schools. As a matter of fact, experts believe that an inefficient education system and lack of programming skills in Indian schools is the critical reason for not having sufficient skilled developers and programmers in the country. Further, Unacademy is also known for its online certification courses, and by embracing CodeChef under its umbrella, it will provide resources for CodeChef users to augment their knowledge before they actually apply for campus recruitment. Parallelly, by advertising Codechef on Unacademy platform, it will urge a pool of students from universities, engineering colleges to participate in hackathons and programming concepts to broaden their horizon and improve their digital footprint by competing with the best brains of the industry. Also Read: How GitHub Is Revolutionising India’s Open Source Community In Data Science And AI Space In fact, according to a recent report, it has been revealed that Indian developers earn three times less than developers in the US, also in some cases lesser than Pakistani developers. This report also stated that “the most important form of professional growth for developers could be learning new technical skills,” and that’s what is keeping them behind their counterparts in other countries. Unacademy partnering with CodeChef will help these programming enthusiasts to learn new skills amid this crisis. Outlook Though this transitioning move in changing caretakers would have no impact on CodeChef’s existing programs and hackathons, it will provide them with more significant resources, advanced technology and better capital to work with. This, in turn, will augment the Indian developers’ community with an easily accessible platform to upskill themselves.","excerpt":"In recent news, it has been announced that Mumbai-based non-profit competitive programming platform, CodeChef has been acquired by a Bangalore-based ed-tech startup Unacademy.  Established in 2009 by serial entrepreneur, Bhavin Turakhia as part of Directi, CodeChef platform has been designed to help developers and programmers to enhance their skills by practising coding and participating in […]","categories":["AI Features"],"tags":["big data developer skills","Developers","developers in India","developers India","Mergers and Acquisitions","RPA developer"],"author_name":"Sejuti Das","publish_date":"2020-06-24T15:00:00","publication_year":"2020","word_count":1091,"keywords":["data science","Go","RPA developer","API","big data developer skills","AI","R","Git","RAG","automation","Aim","GitHub","developers India","Mergers and Acquisitions","developers in India","Developers"],"extracted_tech_keywords":["AI","data science","Aim","RAG","R","Go","Git","GitHub","API","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-codechefs-acquisition-puts-focus-on-the-importance-of-developers-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042654,"title":"A Comprehensive Guide To Regression Techniques For Time Series Forecasting","content":"In mathematics, time series is a series of data points listed with respect to time; most commonly, it is a sequence taken at successive equal intervals point in time. Common examples of time series are daily closing values of the stock market, counts of sunspots etc. Time series analysis comprises methods for analysing time-series data to extract meaningful statistical information and other data characteristics. In contrast, time series forecasting uses a model to predict future values based on previously observed values. In this article, we are going to explore the following regression techniques used for time series forecasting; AR and MAARIMASARIMAVAR Code implementation of Regression Techniques The dataset we are using for all the techniques remains the same and can be found here. The dataset contains weather data collected for the city of Delhi for four years, from 2013 to 2017. import pandas as pd data = pd.read_csv('DailyDelhiClimateTrain.csv') data.head() Lets plot the line chart for humidity. import plotly.express as px fig = px.line(data, x=data.date, y='humidity', title='Humidity with slider') fig.update_xaxes(rangeslider_visible=True) fig.show() 1. Autoregressive and Moving Average (AR and MA): In multiple regression models, we forecast variables of interest using a linear combination of predictors. Here in the autoregressive model, we forecast the variable of interest using a linear combination of past values of the variable. The term autoregression indicates it is a regression of variables against itself. The model can be formulated as; Where: Yt is the value of time series at time t C is the intercept Ø is the slope coefficient Yt-p is the lagged values of time series ε is the error term This method is suitable for univariate time series without trend and a seasonal component. Code Implementation # AR example from statsmodels.tsa.ar_model import AutoReg # fit model train,test = data[0:1000],data[1000:] model = AutoReg(train.humidity, lags=350) model_fit = model.fit() # make prediction pred = model_fit.predict(len(train),len(test)+len(train)-1,dynamic=False) plt.plot(test.humidity) plt.plot(pred,color='red') Rather than using past forecast values in regression, a moving average model uses past forecast errors in a regression-like model. In other words, the moving average models the next sequence as a linear function of residual error from the mean process at an earlier time step. Thus, it combines both autoregressive and moving average models. This method is suitable for univariate time series without trend and seasonal component. Code Implementation: #MA model from statsmodels.tsa.arima.model import ARIMA # fit model model = ARIMA(train.humidity,order=(300,0,0)) model_fit = model.fit() # make prediction pred = model_fit.predict(len(train),len(test)+len(train)-1) plt.plot(test.humidity) plt.plot(pred,color='red') 2. Autoregressive integrated moving average (ARIMA): It explicitly creates a suite of standard structure in time series data and it provides a simple and powerful method for forecasting. It combines both autoregressive and moving average models as well as a differencing pre-processing step of the sequence to make the sequence stationary. This method supports univariate time series with trend and without seasonal component. The statsmodel library provides the capability to fit ARIMA models. Code Implementation: from statsmodels.tsa.arima.model import ARIMA train,test = data.humidity[0:1000],data.humidity[1000:] X = train size = int(len(X) * 0.66) train, test = X[0:size], X[size:len(X)] history = [x for x in train] predictions = list() for i in range(len(test)): model = ARIMA(history, order=(5,1,0)) model_fit = model.fit() output = model_fit.forecast() pred = output[0] predictions.append(pred) true = test[i] history.append(obs) print('predicted=%f, expected=%f' % (pred, true)) plt.plot(test) plt.plot(predictions, color='red') 3. Seasonal Autoregressive integrated moving average (SARIMA): An extension of ARIMA that supports the direct modeling of the seasonal component of the series is called SARIMA. The problem with ARIMA is that it does not support seasonal data i.e repeating cycles. ARIMA expects data that is not seasonal or seasonal component removed SARIMA adds the three hyperparameters to specify the AR, differencing and moving average for the seasonal component of series This model suitable for univariate time series with trend and seasonal component. Code Implementation: from statsmodels.tsa.statespace.sarimax import SARIMAX size = int(len(X) * 0.66) train, test = X[0:size], X[size:len(X)] history = [x for x in train] predictions = list() # walk-forward validation for t in range(len(test)): model = SARIMAX(history,seasonal_order=(3, 1, 0, 2)) model_fit = model.fit() output = model_fit.forecast() pred = output[0] predictions.append(pred) true = test[t] history.append(true) print('predicted=%f, expected=%f' % (pred, true)) plt.plot(test) plt.plot(predictions, color='red') 4. Vector Autoregression (VAR): The vector autoregression model can predict when two or more time series influence each other means the relationship involved in time series is bi-directional. This model considers each variable as a function of past values that are to be predicted, nothing but the time lag of the series. For all this, it considers an autoregressive model. The main difference between the previous model and VAR is, those models are unidirectional, where predictors influence the Y but not vice-versa. Whereas the VAR model is bidirectional, variables influence each other. This model is suitable for multivariate time series without trend and seasonal components. Code Implementation: Load multiple variables: x1 = data.humidity.values x2 = data.meantemp.values list1 = list() for i in range(len(x1)): x3 = x1[i] x4 = x2[i] row1 = [x3,x4] list1.append(row1) Fit and forecast to few steps from statsmodels.tsa.vector_ar.var_model import VAR # fit model model = VAR(list1) model_fit = model.fit() # make prediction forecast = model_fit.forecast(model_fit.y, steps=5) print(forecast) Output: [[95.76561271 10.57589906] [92.08148688 11.10511153] [88.87374484 11.59330815] [86.07847799 12.04540676] [83.64040052 12.46567364]] Conclusion This article has seen the major techniques used to forecast time series entities with a practical use case. The most time-consuming thing in the univariate techniques is adjusting the lag values; the proper lag value decides the nature of forecasting. The rest of the techniques are straightforward. References Link for the colab notebookStats model for time series analysisMore about time series analysis","excerpt":"This article is about various regression techniques used to forecast timeseries problem","categories":["Deep Tech"],"tags":["Guide","regression analysis","Time Series","Time Series Forecasting"],"author_name":"Vijaysinh Lendave","publish_date":"2021-06-30T17:00:00","publication_year":"2021","word_count":915,"keywords":["Go","Plotly","TPU","programming_languages:R","AI","RPA","regression analysis","Time Series Forecasting","RAG","Colab","Time Series","R","Guide","Pandas"],"extracted_tech_keywords":["AI","Colab","Pandas","Plotly","RAG","TPU","R","Go","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-comprehensive-guide-to-regression-techniques-for-time-series-forecasting\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10022935,"title":"India Poised To Grow Rapidly In Data Science Education: Paul Kim, Stanford University","content":"The popularity of data science education has soared over the last few years. Subsequently,  field’s job prospects have also gone up, with companies worldwide looking to hire skilled professionals to drive business processes. The nonlinear growth of data science has posed significant challenges for universities developing data science courses and individuals looking to pursue it as a career. For this week’s data science career series, Analytics India Magazine got in touch with Paul Kim, the Chief Technology Officer and Assistant Dean of the Graduate School of Education at Stanford University. Known as an education technology entrepreneur, Professor Kim leads initiatives involving the design of learning technologies, educational research, and community development. His work aims at promoting innovation and competition by constructing a programmable and open mobile internet — POMI. In an advisory capacity, Paul has played a role in Saudi Arabia’s national online education initiative, the national evaluation of Uruguay’s One Laptop Per Child project and Rwanda’s national ICT planning. In this interview, he provides an overview of the data science market and the challenges universities face in developing a practical data science course. He also spoke about the future of the aspiring data scientist for the current era. Challenges with data science education market in India All sectors have welcomed data science with open arms. The COVID pandemic has accelerated ICT adoption in teaching, learning, assessment, and administrative functions. “I would say COVID has created R&D opportunities along with available funding more than any other catalysts in the past few decades,” said Paul. Professor Kim believes India is in a much better position in terms of the data science education market because of the multiple technology innovation powerhouses strategically located in India and their growing needs for the future workforce in data science and artificial intelligence. He said, “while big contenders are obviously the US and China, but with institutional on financing and governmental support in terms of policies and regulatory issues, India is poised to grow rapidly in the overall data science education and application areas.” This has led many universities, edtech platforms, tech companies, and governments to come up with free courses during the pandemic. However, there has been a massive learning gap between the course structures and the skills required to land a job. Paul said, considering data science is a rapidly advancing field, universities steeped in traditional models of governance and decision-making processes will have a hard time instrumenting data science courses. That’s why “universities in India must transform to align with many of competing alternative education options such as online boot camps and non-traditional talent development organisations.” Paul also mentioned the importance of government and corporation involvement in encouraging more students to choose data science subjects. “Governments can figure out ways to remove policy and regulation related barriers while corporations can work closely with educational entities that are nimble and flexible to provide the most invigorating and fast-developing data science curriculum in the world,” he said. Advice for aspiring data scientists Paul stated, being well-rounded, skilled talents who can use a wide lens of viewing capability to understand the true needs of the industry and users while genuinely developing empathy to solve most intractable problems is the key to become a real data scientist. “Do not follow people around you, but develop your own unique skill sets, so you are rare species in the data science ecosystem,” advised Paul. “If you follow others and be just another data science worker, you may not be necessarily a highly sought talent in the whole ecosystem.” While there are many online courses and MOOCs currently available for data science enthusiasts, Paul bets high on a professional degree in data science. A professional degree is for those who couldn’t demonstrate his or her talent with competitive problem-solving skills, said Paul. “Though these degrees can help get one to an interview if one cannot demonstrate their competency, they are not going to secure a career opportunity they want,” he added. “At the end of the day, what makes a difference between a competent contender versus a mediocre contender is in the genuine passion for being the best,” he concluded.","excerpt":"The popularity of data science education has soared over the last few years. Subsequently,  field’s job prospects have also gone up, with companies worldwide looking to hire skilled professionals to drive business processes. The nonlinear growth of data science has posed significant challenges for universities developing data science courses and individuals looking to pursue it […]","categories":["AI Highlights"],"tags":["data science education"],"author_name":"Sejuti Das","publish_date":"2021-03-26T14:00:00","publication_year":"2021","word_count":690,"keywords":["data science","Go","API","artificial intelligence","AI","RAG","Aim","analytics","data science education","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","RAG","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/india-poised-to-grow-rapidly-in-data-science-education-paul-kim-stanford-university\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":12494,"title":"Enterprises worldwide beef up against Ransomware attack &#8212; crime of the century","content":"MongoDB ransomware attack is a reminder why cyber crimes are the biggest threat in the security landscape Of late, MongoDB, the Database-as-a-Service provider grabbed the headlines for one of the biggest wrong reasons — ransomware attack or malware that installs surreptitiously on a user’s computer, locks the system or locks the user’s files. Earlier last month, 27,000 databases came under attack when hackers compromised unsecured instances of MongoDB running openly on the internet. Reportedly, hackers wiped off over 680 terabytes of crucial data and demanded one bitcoin in return of the database. MongoDB ransomware attack brought into sharp focus the crimes in cyberspace and they cost to global economy. MongoDB Ransomware attack signals a red flag in privacy and security of data Why MongoDB’s instances were held hostage? MongoDB’s unsecured instances were openly accessible on the internet and did not have any password-protected admin accounts. Lack of firewall to protect the databases and lack of security configurations made one of the most popular database management system a sitting target.  According to reports, there are 30,000+ MongoDB installations on the web that are available on the cloud. Extortion is the most extreme step taken by hackers, industry experts reportedly point out that databases can be compromised by hosting malware or hiding files. The repercussions of the attack exposed data from several organizations. Here’s the total number of organizations exposed and the hackers involved in the malicious attempt.  But the ransacking didn’t stop there. According to news reports, 600+ unsecured Elasticsearch clusters were hit by ransomware. The servers hosted on Amazon Web services sent out an advisory to their customers. As per their blog, “No malware, or “ransomware” was used in the attacks and there was no data breach, still the incident represented a serious security concern,” read the blog entry.  The blog also listed down steps to “secure data in Internet-facing instance of Elasticsearch”. When Analytics India Magazine contacted representatives at MongoDB, the major proponent of unstructured relational database, we got MongoDB’s “Suggested Steps To Diagnose and Respond to an Attack”. As per MongoDB’s guide, “MongoDB Cloud Manager and MongoDB Ops Manager provide continuous backup with point in time recovery wherein users can also enable alerts in Cloud Manager to detect if their deployment is internet exposed”. Spike in Ransomware attacks From Montana schools to smart TVs in Japan, there is a surge in cyber-targeting with news reports pointing cybercriminals notching up to a 1$ billion dollars in 2016 in phishing attacks. The year 2017 began with MongoDB’s ransomware attack, courtesy the misconfigured databases followed by Elasticsearch clusters being breached. Security experts pointed out vandalization of Hadoop installations. News reports say a potential attack could expose 8,000 HDFS installations. Though no data stored on HDFS installation was compromised, hackers left a calling card and the breach exposed the “access without authentication”. As per a survey, ransomware attacks are expected to double in 2017. Cyber security is the number #1 concern of most organizations and according to the Osterman Research, nearly 50% of organizations came under ransomware attack in 2016. The most vulnerable sector proved to be financial followed by manufacturing, government and healthcare. Another key finding of the survey was that malware made its way through devices such as smartphone or tablet. How are organizations beefing up in face of ransomware threats? While end-to-end prevention of cyber threats is not possible, security experts note training staff and maintaining a full-proof IT infrastructure can serve as the best defense in protecting data. Companies shore up defense to prevent cybercrimes: Last year, chip maker Intel, Europol (European law enforcement agency) and Kaspersky Lab (Russian cyber Security Company) joined hands to address the growing threat of ransomware. Cyber security vendors need to advance artificial intelligence and machine learning capabilities to effectively prevent cyber threats. From Android mobile OS to Windows OS and Apple Mac, ransomware has widened its net and sophisticated machine learning algorithms can check the threat at early stage. Dubbed as the “crime of the century”, the softest targets – big data databases managed from cloud should shore up authentication and provide recovery data option through a robust program. Redmond-headquartered IT giant Microsoft introduced ransomware protection in Windows 10 anniversary update earlier last year. Steps taken to ensure ransomware defense ranged from browser hardening to machine learning and a more robust Windows Defender. At an organizational level, robust infrastructure management, employee training and risk assessment should be undertaken to protect sensitive data. Even though cyber-attacks cannot be entirely avoided, threats and breach of data can be minimized by following a robust data management policy. Security vendors will come into play in 2017.","excerpt":"Of late, MongoDB, the Database-as-a-Service provider grabbed the headlines for one of the biggest wrong reasons — ransomware attack or malware that installs surreptitiously on a user’s computer, locks the system or locks the user’s files. Earlier last month, 27,000 databases came under attack when hackers compromised unsecured instances of MongoDB running openly on the […]","categories":["IT Services"],"tags":["cyber security India","hadoop world","Ransomware"],"author_name":"Richa Bhatia","publish_date":"2017-02-02T11:01:37","publication_year":"2017","word_count":770,"keywords":["big data","Elasticsearch","Go","cyber security India","artificial intelligence","hadoop world","machine learning","AI","Ransomware","MongoDB","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Elasticsearch","MongoDB","R","Go","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/enterprises-worldwide-beef-ransomware-attack-crime-century\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10138404,"title":"OpenAI Would Have Barely Survived Without Microsoft","content":"Microsoft is the godfather to OpenAI. Let’s face it, without the former’s support, the Sam Altman-led company wouldn’t exist as we know it. It would’ve barely survived a day, as it’s NOT your ‘normal company’. UK-based industry analyst firm CCS Insight recently predicted that within three years, OpenAI will find it so difficult to raise funds that it will need to sell to Microsoft. That could explain why OpenAI recently opened a new office near the Microsoft headquarters in Seattle. Microsoft’s partnership with OpenAI began in 2019 with an initial investment of $1 billion. In the Netflix series What’s Next? Microsoft founder Bill Gates revealed that he initially thought his next big focus would be eradicating malaria, and not AI. However, when Greg Brockman gave him a demo of GPT-4, it completely ‘blew his mind’, changing his perspective on AI. Fast forward to today, OpenAI has raised $6.6 billion at a valuation of $157 billion. This time, however, the startup is not solely dependent on Microsoft, as the funding was led by Thrive Capital. Other significant participants in the funding round included NVIDIA, SoftBank, Khosla Ventures, Altimeter Capital, Fidelity, and MGX. Surprisingly, Microsoft has invested less, perhaps focusing on its in-house AI research. Earlier this year, the company appointed Mustafa Suleyman, former co-founder and head of DeepMind and Inflection AI, as the CEO of Microsoft AI. Microsoft has been building edge LLMs and recently launched the new Phi-3.5 models, which outperform Google’s Gemini 1.5 Flash, Meta’s Llama 3.1, and OpenAI’s GPT-4o. The funding for OpenAI came at the right time as the startup is currently operating at a loss and has yet to make a profit. The company’s forecasts indicate it won’t achieve profitability until 2029, with projected revenue reaching $100 billion. However, losses could escalate to $14 billion in 2026, nearly three times this year’s anticipated losses, according to a recent report. Moreover, OpenAI anticipates a steep rise in computing costs for model training over the coming years, with expenses potentially reaching $9.5 billion annually by 2026. Today, a major chunk of the training compute is taken care of by Microsoft, which provides Azure servers to train OpenAI’s models. Back in 2022, when OpenAI debuted ChatGPT, Altman credited Microsoft for that. “Microsoft, and particularly Azure, don’t get nearly enough credit for the stuff OpenAI launches. They do an amazing amount of work to make it happen; we are deeply grateful for the partnership. They have built by far the best AI infrastructure out there,” posted the OpenAI chief on X. The love was reciprocated by the Microsoft CEO at Ignite 2023 where he showered unlimited affection, heaping praises on OpenAI. He went on to say that the company is grateful to have OpenAI as part of its journey, besides showcasing the readiness to give up everything it has built so far. This included rebranding its search engine, revamping its Azure cloud infrastructure and much more. Nadella has, on many instances, expressed his feelings towards OpenAI. Even on DevDay 2023, he said, “We love you guys… You guys have built something magical.” The relationship between Microsoft and OpenAI is symbiotic, where neither can survive without the other. But today, both the AI startup and the tech giant are exploring ways to reduce their dependence on each other. The tech giant now offers LLMs from NVIDIA, Mistral, Cohere, G42 and Meta, in addition to the frontier models available from OpenAI in the Azure AI Studio. However, Microsoft’s flagship products, such as Copilot 365, still rely on OpenAI’s models. Recently, the company introduced Copilot agents within its Microsoft 365 Copilot platform, designed to help businesses customise AI to meet their specific needs. These Copilot agents connect to data sources like SharePoint and Dynamics 365, automating tasks and integrating with existing systems to streamline workflows. Microsoft also recently updated Copilot, adding voice and vision features. OpenAI x Microsoft Today, with AI startups like Anthropic, Cohere, and xAI breathing down OpenAI’s neck, the company will likely struggle to survive without Microsoft’s support. The models being developed by OpenAI, such as the video generation model Sora and its latest release o1, require significant computing capacity to create and make available to consumers worldwide. It’s likely that without Microsoft’s initial $10 billion investment, this would not have been possible. Under the current arrangement, Microsoft claims 75% of OpenAI’s future profits until this initial investment is repaid. After repayment, Microsoft will be entitled to 49% of the profits, up to a theoretical cap. With OpenAI planning to shift its structure from non-profit to for-profit, it remains to be seen what the new deal between Microsoft and OpenAI will entail. Today, OpenAI provides Microsoft with a 20% commission for sales made to other businesses accessing its AI models via an application programming interface. In turn, Microsoft pays OpenAI a 20% commission when it resells OpenAI’s models to Azure customers through a Microsoft API. In his keynote at Microsoft Build 2024, Nadella said that OpenAI is Microsoft’s “most strategic and most important partner”. Ironically, just recently, Sebastien Bubeck, a Microsoft AI researcher, announced that he is departing to join OpenAI. He played a key role in Microsoft’s work on generative AI models over the past two years. At the event, Microsoft CTO Kevin Scott said that if the system that trained GPT-3 was a shark and GPT-4 an orca, the model being trained now by OpenaI is the size of a whale. “This whale-sized supercomputer is hard at work right now,” he added. Microsoft has always supported OpenAI during its challenging times. Following Altman’s unexpected ouster last year, Nadella announced just two days later that Altman had been hired to lead a new AI research team at Microsoft. This move was seen as a way for Microsoft to safeguard its stock price and could strengthen its influence within OpenAI. Slowly Drifting Away OpenAI is reportedly unhappy with Microsoft because the latter was not quick enough to deliver the necessary computing power, prompting the company to seek other data center agreements, a process it had already initiated. Interestingly both of them recently received the much awaited NVIDIA Blackwell GPUs. This year, OpenAI also attempted to establish closer ties with Apple in the hope of securing additional funding. However, Apple did not participate. For Apple, this could mean investing more in its internal R&D and architecture rather than pursuing external investments. Meanwhile, OpenAI is collaborating with Oracle to secure additional computing resources. This partnership will allow it to further scale its operations and services. Altman said, “OCI will extend Azure’s platform and enable OpenAI to continue to scale.” The company is currently in discussions to lease the entire Abilene data center site from Oracle, which could eventually expand to 2 gigawatts if Oracle secures additional power for the location, according to a recent report. OpenAI’s attempt to look for more compute options besides Microsoft Azure is natural, considering that the startup has made a blueprint for the next few years, where it has shifted its focus from LLMs to AI agents and consequently ASI. However, there is very little possibility that Microsoft will leave OpenAI and it will continue to play a huge role in OpenAI’s journey.","excerpt":"An analyst firm recently predicted that within three years, OpenAI will find it so difficult to raise funds that it will need to sell to Microsoft.","categories":["Global Tech"],"tags":["Microsoft","OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-10-16T13:25:33","publication_year":"2024","word_count":1196,"keywords":["Anthropic","ChatGPT","OpenAI","AI","Azure","GPT-4o","ML","Aim","generative AI","xAI","Microsoft"],"extracted_tech_keywords":["AI","ML","generative AI","GPT-4o","ChatGPT","OpenAI","Anthropic","xAI","Aim","Azure"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-would-have-barely-survived-without-microsoft\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36343,"title":"Pi Day Turns 10: Looking Back Over The Last Decade","content":"Today is March 14, a day celebrated as Pi Day all over the world. First instituted in 2009 by law in the US House of Representatives, the day has now become a global phenomenon. Researchers, scientists and math lovers worldwide come together to celebrate a day wherein math and real life coincide. Pi is a mathematical constant that is often used to calculate the circumference of a circle. Its value is 22\/7, or 3.141…, due to its being an irrational number. It is widely considered as the most unique and pervasive mathematical number due to its infinitely calculable nature. Incidentally, today is the 10th official Pi Day as per the ruling in the US. However, the day has been celebrated for a long time, and even coincides with the birthday of prominent physicist Albert Einstein. One of the special Pi days that took place recently was that of 2015. At exactly 9:26:53 on March 14th of that year, the clock showed 3.141592653, which are the first 10 digits of pi. The number is of great importance to mathematicians, who have even gone so far as to call it “transcendental”. However, the convenience of using Pi has been questioned, as the mathematical constant Tau, which is simply Pi multiplied by 2, is easier to use in formulae. However, the infinitely random nature of Pi has made it a draw for mathematicians, as they seek to find patterns in the numbers. On this day, scientists celebrate by having Pi-themed competitions, some of which include Albert Einstein look-alike competitions. Another long-standing tradition is also to eat pie, due to it sounding similar to the number. There has also been a long-standing competition to compute the most digits of Pi that is humanly possible, at least before the 21st Century. The 100th digit calculation of Pi took place only in 1706 when Abraham Sharp decoded the first 100. The 200-mark was broken in 1824 by William Rutherford, with the number going up to 707 in 1874. The rise of supercomputers sped up the discovery of more digits, with the ENIAC in 1949 taking 70 hours to calculate 2,037 digits. In 1973, the CDC 7600 supercomputer took just under 24 hours to calculate 1,001,250 of Pi.  The 1 billion-mark was broken by the T2K Open Supercomputer, which took 29 hours to calculate 2.5 billion digits. However, the record is held today by Peter Trueb, who took over 100 days to calculate 22.4 trillion digits of Pi in 2016. This effort took over 120 TB of hard disk space and 4 server-grade processors to achieve and remains one of the biggest advances when it comes to calculating the number.","excerpt":"Today is March 14, a day celebrated as Pi Day all over the world. First instituted in 2009 by law in the US House of Representatives, the day has now become a global phenomenon. Researchers, scientists and math lovers worldwide come together to celebrate a day wherein math and real life coincide. Pi is a […]","categories":["AI News"],"tags":[],"author_name":"Anirudh VK","publish_date":"2019-03-14T12:42:27","publication_year":"2019","word_count":443,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","BERT","llm_models:BERT","R"],"extracted_tech_keywords":["AI","R","Go","Git","BERT","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pi-day-turns-10-looking-back-over-the-last-decade\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10141391,"title":"WTF is Nikhil Kamath Doing with Young Entrepreneurs?","content":"A few months ago, Zerodha co-founder Nikhil Kamath unveiled the ‘Innovators under 25’ initiative, marking the launch of WTFund. The programme selected nine Indian startups led by exceptional founders under 25. For its first edition, WTFund, a non-equity grant fund, selected 15 entrepreneurs working across various sectors including AI, healthcare and others. The initiative provides up to INR 20 lakh in non-equity grants. Interestingly, the entrepreneurs can retain the full ownership of their startup with Kamath not having any stake in them. Empowering Young Entrepreneurs (Nikhil Kamath with the startup founders selected for the ‘Innovators Under 25’ programme.) As strange as it may sound, Kamath, as an investor, is not looking to have any equity in these companies; he’s clear about his vision of honing emerging young talent in the country. “For the Innovators under 25 programme, my approach is sector-agnostic, but focuses on startups solving real-world problems, especially in health tech, energy transition, space tech, and AI. The reason for offering non-dilutive grants and not taking equity is simple: It is about empowering young founders to grow without the immediate worry of dilution,” Kamath told AIM. In 2025, Kamath plans a similar or larger investment to support early-stage startups and provide them the runway they need. He clarified that his focus was on resilient founders who deeply understood their problem space. “Does this idea serve a real purpose? Can it scale effectively with strong unit economics? If a startup can create long-term value while making a tangible impact, that’s what seals the deal for me,” said Kamath. AIM had the opportunity to interview five startups from this programme, which places a strong emphasis on AI and tech. Equipping Creators with High-Performance Cloud Computing Founded by CS graduate Advait Bansode, Mars Computers aims to disrupt the creative and developer ecosystem by making high-performance computing accessible via the cloud. The startup connects users to data centres through a low-latency pipeline, allowing resource-intensive applications such as Adobe Premiere Pro and DaVinci Resolve to run seamlessly on lightweight devices such as MacBook Air. “Businesses avoid the hefty cost of buying machines,” said Bansode, emphasising the efficiency of their subscription model. Mars also addresses the flexibility needed for freelancers and small businesses, enabling users to subscribe based on their project requirements rather than investing in permanent hardware. “If I just want to do one project for a month, can I pay for it? That question sparked the idea.” The startup’s abstraction layer ensures that users experience local-computer-level performance while benefiting from scalable cloud solutions. Speaking about their ambitious vision, Bansode said, “We are excited about building tech that hasn’t existed in the market before.” Harnessing DNA for Data Storage Young TED speaker Anagha Rajesh founded BioCompute with the aim to meet the growing demand for sustainable and scalable storage solutions. Leveraging on the remarkable density and longevity of DNA, the startup aims to commercialise its use for archival storage, potentially replacing traditional storage media that require frequent replacements and consume significant energy. “The idea is that DNA can last for thousands of years,” explains Anagha, emphasising the longevity and cost-effectiveness of the technology. By developing an enzymatic approach to DNA synthesis, BioCompute tackles the bottlenecks of cost and scalability, with the ultimate goal of integrating DNA storage into data centres globally. “If we can demonstrate significantly lower space, energy, and replacement costs, this technology will sell itself,” she said. “Biological systems are inherently more energy efficient.” BioCompute also focuses on reducing the reliance on chemical synthesis methods, adopting biological alternatives to make the process more efficient and environmentally friendly. AI-Powered Diagnostics for Gastric Cancer Founded by MBBS students Tanmaya Gulati and Ria Khurana, RNT Health Insights is on a mission to improve early detection of gastric cancer using AI. Their solution integrates spatial and temporal deep learning models into real-time endoscopy procedures, helping doctors identify lesions that might otherwise be missed during screenings. “Our models predict in just 30 milliseconds,” said Gulati, showcasing the speed and precision of their technology. The solution acts as a “second set of eyes” for doctors, analysing 30-35 frames per second during endoscopic procedures to detect even the smallest abnormalities. “Sometimes lesions appear for milliseconds on the screen and can be easily missed by doctors,” he explained, underscoring the critical role of AI in filling diagnostic gaps. With India ranking third globally in gastric cancer cases, the founders believe their startup has the potential to save countless lives. The team is also working on making the technology compatible with high-definition endoscopy equipment, ensuring broader adoption across healthcare systems. “We want to integrate it without interfering with the existing workflow,” said Khurana. Education with AI-Powered Companions Founded by IIT graduate Sparsh Agarwal, Pixa is creating AI-powered toys that serve as interactive companions for children aged 5 to 12. These toys, integrated with advanced language models and custom memory stacks, function as personalised tutors, teaching topics ranging from programming to healthy habits through voice interactions. “Rather than spending $20 a month on multiple apps, parents get a single, screen-free solution,” said Agarwal. The toys feature advanced AI capable of generating dynamic and personalised content for children, from quizzes to stories, ensuring they never outgrow the experience. Parents can monitor their child’s progress through an app, which also tracks vocabulary levels and provides summaries of interactions to ensure a safe environment. “We wanted to ensure kids have access to powerful educational tools without increasing their screen time.” He also said that with AI, users won’t have to encounter the limitations of pre-loaded educational gadgets anymore. (Source: Pixa) Sales Preparation with AI Founded by Bhavesh Kotwani (BITS Pilani), Nikhil Mehta (IIT), and Pooja Midha (2X founder), CallPrep is transforming sales workflows by automating pre-meeting preparation, giving sales reps valuable insights without leaving their usual platforms. “We are trying to make the information work for the salesperson even before they step into the meeting,” the founders said, demonstrating how their AI-driven tool is helping sales teams scale more efficiently. By integrating seamlessly into calendars and CRM systems, the platform delivers actionable intelligence tailored to each meeting, helping sales teams save time on administrative tasks. “Sales reps are overwhelmed with multiple solutions, but with CallPrep, they don’t need to go outside their platform for insights,” the founders explained. Focusing on mid-market and enterprise-level B2B companies, CallPrep differentiates itself by addressing the underexplored pre-meeting space. Unlike competitors targeting post-meeting insights, CallPrep ensures sales teams are prepared with context-specific data like competitor battle cards before their meetings. In addition to the startups above, WTFund has invested in businesses that are building sustainable and healthy solutions for humans and pets alike. Urban Animal offers India’s first dog DNA testing service, revolutionising pet care, while Oh! Nuts caters to health-conscious Indian consumers with premium, nut-based snacks. Pawsible Foods introduces sustainable, plant-based pet food using Kavaka™ mycoprotein, and Pamawel targets menstrual pain relief with its plant-based, non-steroidal, FDA-approved formulations.","excerpt":"“If a startup can create long-term value while making a tangible impact, that’s what seals the deal for me,” Kamath told AIM.","categories":["AI Trends"],"tags":["AI","Data Center","DNA","Innovators Under 25","Nikhil Kamath","WTFund"],"author_name":"Vandana Nair","publish_date":"2024-11-22T10:00:00","publication_year":"2024","word_count":1149,"keywords":["Go","Data Center","AWS","AI","cloud computing","ML","Nikhil Kamath","Scala","RAG","Innovators Under 25","Aim","deep learning","WTFund","R","DNA"],"extracted_tech_keywords":["AI","ML","deep learning","Aim","RAG","cloud computing","AWS","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/wtf-is-nikhil-kamath-doing-with-young-entrepreneurs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":27132,"title":"Choosing Between GAN Or Encoder Decoder Architecture For ML Applications Is Like Comparing Apples To Oranges","content":"Since the deep learning boom has started, numerous researchers have started building many architectures around neural networks. It is often speculated that the neural networks are inspired by neurons and their networks in the brain. Computational algorithms often mimic and copy these biological structures. But there is yet a lot to be discovered about how the brain actually works. Neuroscience is nowhere close to solving the mystery of the brain. That is why artificial intelligence scientists have to come up with many neural network architectures to solve different tasks. Two of the main families of neural network architecture are encoder-decoder architecture and the Generative Adversarial Network (GAN). Encoder and Decoder Architecture In 2015, Sequence to Sequence Learning with Neural Network became a very popular architecture and with that the encoder-decoder architecture also became part of wide deep learning community. The paper proposed a LSTM to match input sequence to a vector with fixed dimensionality. The vector was converted into an output sequence by another LSTM network. Such architectures have been used for tasks like machine translation. The architecture for RNNs and LSTMs has become standard for machine translation tasks that works better than many classical statistical machine translation methods. Encoder Decoder Architecture The main advantage of this kind of an architecture is that researchers and engineers now have the ability to train a single end-to-end model directly on input (source) and output (target) strings. The architecture also has the ability to variable length input and output strings or sequences. This kind of architecture was specifically designed for natural language processing tasks and it immediately showed the state of the art performance in many tasks. The encoder-decoder architecture can be applied to a host of problems : Machine Translation (translating English to French) Modeling a long conversation Image-to-text conversion (captioning) Gesture tracking and prediction There are two parts of the architecture: the encoder and the decoder. The input sequence is given to the encoder decoder network and it is encoded one character at a time. An encoding is needed to understand and learn the relationships between many steps and then create a representation. Normally, LSTM layers are used to create the encoder model and output of the encoder is a vector of fixed size that also is a representation of the input sequence. The size of vector is the number of memory cells. A decoder now takes this learned representation and tries to change it into the correct output sequence. Again decoder is implemented using layers of LSTM networks and this layer has the job of intaking the input from encoder. A Repeat Vector is used as an transformer and adapter to gel together the output of the encoder and input expected by the decoder. Generative Adversarial Networks Deep generative models are a family of powerful deep learning models that are capable of learning many kinds of data distribution in an unsupervised manner. The family of models has won many accolades in the past and has also led to many impressive results. Many academic researchers including Geoffrey Hinton feel that generative modeling is the right way to go ahead and it may also be the correct way to model the brain. The two most popular ways of generative modeling are variational autoencoders and generative adversarial networks. There were still some drawbacks in these models. Ian Goodfellow in 2014 was able to was able to invent an architecture that would solve some of the problems in deep generative modeling. The invention was Generative Adversarial Networks. GAN architecture Generative adversarial network are used in many tasks. Right from identifying cars on roads to filling the details of a photo that are missing. Moreover, these GANs have been used to predict how people might look when they are old. Generative adversarial networks are introduction of game theory into deep learning. They learn to data distribution and try to generate similar data through a 2-networks game. The two players are called as Generator and Discriminator. The whole architecture is known as adversarial because the two players (networks) are in a battle mode throughout the training process. Here a generator tries to learn the data distribution and generate images which can be passed by the discriminator. The discriminator tries not to get fooled. The job of the discriminator is to identify which of the generated data is not real. Hence as the training process carries on, the generator tries to fool the discriminator. Both the players (networks) have to be good here otherwise the produced model will not be useful. Comparison The two architectures are used variety of tasks and both have different applications. As we mentioned above the encoder decoder architecture are mainly used to create good internal representations of the data and can be seen as great data compression engines. The encoder encodes the data and the decoder tries to reconstruct the data back using the internal representations and the learned weights. Whereas GANs work on a generative principle and try to learn from data distributions to use a game theory approach to build great models. Here the discriminator tries to identify fake data created the generator and hence making the players job difficult and producing better results.","excerpt":"Since the deep learning boom has started, numerous researchers have started building many architectures around neural networks. It is often speculated that the neural networks are inspired by neurons and their networks in the brain. Computational algorithms often mimic and copy these biological structures. But there is yet a lot to be discovered about how […]","categories":[],"tags":["AI (Artificial Intelligence)","encoder-decoder","GANs","Generative Adversarial Network","lstm","machine translation","neuroscience","RNN"],"author_name":"Abhijeet Katte","publish_date":"2018-08-09T11:32:49","publication_year":"2018","word_count":864,"keywords":["Go","LSTM","Generative Adversarial Network","machine translation","neuroscience","artificial intelligence","AI","neural network","TPU","encoder-decoder","lstm","deep learning","RNN","GANs","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","TPU","R","Go","GAN","RNN","LSTM"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/choosing-between-gan-or-encoder-decoder-architecture-for-ml-applications-is-like-comparing-apples-to-oranges\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":29840,"title":"Efficient Data Storage Solutions In The Age Of AI","content":"“Data is the new oil.” This oft repeated concept, first coined by British mathematician Clive Humby, highlights the inherent value of data in the post-industrial era and the one that powers the transformative technology of the digital era, like artificial intelligence (AI), predictive analytics, machine learning, et al. India is on the forefront of leveraging AI for social and economic development and has demonstrated its appetite to embrace AI and the opportunity it offers through its plans. The National Institution for Transforming India (NITI Aayog) is working on the country’s first-ever national policy on AI, and a government-appointed task force[i] has released a comprehensive plan to implement AI in 10 sectors in the country over the next five years. Data Is The Lifeblood Of AI The adoption of AI enables diverse solutions that find applications in businesses as well as in governance, and also carries us into an era focused on data – the exponential growth in volume, the analysis, and the management of it. According to a recent whitepaper by IDC commissioned by Seagate, there will be a 10-fold growth in the amount of data generated globally by 2025 rising up to 163 zettabytes. Given that successful AI implementation depends on relevant and good quality data as a fundamental building block, the collection, analysis, and storage of data will be increasingly critical to individuals and for businesses that leverage that data. Asia Pacific has seen robust AI adoption in one form or another within areas such as information technology, supply chain and logistics, as well as research and development, and organisations are hungry for more. According to our Data Pulse: Maximising the Potential of AI survey, 9 out of 10 organisations in the region have indicated plans to implement AI in the next 12 months. Because consumers and businesses are increasingly becoming more intertwined with data, organisations will need to manage and store data efficiently and securely to build consumer trust and reap the benefits of data analytics and real-time data processing. More Robust Data Storage Solutions Is The Need Of The Hour The increase in data generation from a wider implementation of AI calls for more robust data storage solutions so that organisations can better manage their data in an efficient and secure manner. Asia Pacific is at the center of this digital transformation with its vibrantly growing economies and will drive this growth. India has one of the highest AI adoption rates in Asia Pacific at 90%, compared to the regional average of 74%. With the increasing amount of data created as a result of growing AI adoption, 95% of organisations believe there is a need to have more robust data storage solutions so that their IT infrastructures are ready to handle the incoming data efficiently and securely. Our survey shows that 98% of Indian organisations believe that they must invest in IT infrastructure to handle the growing volume of data, yet one-fifth also said their current infrastructures aren’t sufficiently robust to deal with the increase in data volumes. Today data storage is more than just about archiving information. With the application of AI and machine learning, It’s about providing ways to analyse the data collected, understand patterns and behavior  as well as harnessing stored information for growth and innovation. As we’ve seen, data is the pulse behind AI, and organisations will need to constantly evaluate storage, access, and management solutions to derive insights for efficient and impactful decision-making to grow their business.","excerpt":"“Data is the new oil.” This oft repeated concept, first coined by British mathematician Clive Humby, highlights the inherent value of data in the post-industrial era and the one that powers the transformative technology of the digital era, like artificial intelligence (AI), predictive analytics, machine learning, et al. India is on the forefront of leveraging […]","categories":["AI Features"],"tags":[],"author_name":"B.S. Teh","publish_date":"2018-11-01T06:38:04","publication_year":"2018","word_count":576,"keywords":["Go","machine learning","artificial intelligence","AI","R","Git","RAG","analytics","Rust","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","RAG","predictive analytics","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/efficient-data-storage-solutions-in-the-age-of-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":24708,"title":"Why Do Data Scientists Prefer Python Over Java?","content":"Python has been billed as the most popular language in the StackOverflow survey, where it even beat C# in popularity this year. StackOverflow has chronicled the incredible growth of Python, and has labelled it as the most preferred language for machine learning applications. In fact, according to the findings, Python was one of the most visited tags on StackOverflow as well as one of the fastest-growing ones in 2017. It has also recorded year-over-year growth ever since 2013. Hackerrank 2018 developer survey indicated that even though JavaScript is most in-demand language by employers, Python wins the heart of developers across all ages, according to their Love-Hate index. Why Is Python The Most Popular Language In Machine Learning? Powerful And Easy Implementation: With Python, students and researchers need to get to know the language before getting into machine learning or artificial intelligence. Since Python is considered as a beginner’s language, it doesn’t have a steep learning curve, and even a developer with basic knowledge can work with it. Apart from that, developers don’t have to think about software engineering constraints or the time spent on debugging codes in Python either. The time consumed is less when compared to languages like C, C++ or Java. As a result, developers can spend more time on their algorithms and heuristics related to AI and ML. Ease Of Libraries: Python comes with a huge number of inbuilt libraries for machine learning and artificial intelligence. Some of the most popular libraries are Pytorch, TensorFlow (high-level neural network library for deep learning), scikit-learn (for data mining, data analysis and machine learning), matplotlib, seaborn, scikit (data visualisation), etc. Thanks to Python’s popularity, there are numerous resources — machine learning and data science tutorials — out there where Python libraries are utilised. Plenty of tutorials are easily available online as well. Most of the time, researchers build their own libraries and upload them on GitHub or similar platforms so that they can be used by others. The developer community support and a plethora of features is what makes Python suitable for machine learning applications. On the other hand, Java was mostly built for general programming, not number crunching, a field where R and Python are more preferred. Speed: Java Is Faster Than Python As Java is one of the oldest languages, it comes with a great number of libraries and tools for ML and data science. However, it is also a difficult language for beginners to pick up as compared to Python and C#. In terms of toolset, Java has a number of libraries and tools, some of the popular ones being Weka, Java-ML, MLlib and Deeplearning4j, which are leveraged to solve most of the cutting edge machine learning problems. Also, Java is pegged to be 25 times faster than Python. In terms of concurrency, Java beats Python. Java is excellent when it comes to scaling applications, which makes it the best choice for building large and more complex ML and AI applications. Researchers assert that if you’re planning to build your application from the ground level, it’s good to choose Java as your programming language. Why Is Python So Popular With The Data Science Community One of the main reasons why Python is widely used in the scientific and research communities, is because of its ease of use and simple syntax which makes it easy to adopt for people who do not have an engineering background. It is also more suited for quick prototyping. Another reason that could explain the popularity of Python is that most online courses on data science and machine learning as pushing Python because it is easy to use for beginners. Most developers have dubbed Python as the Swiss Army Knife in the data science community, thanks to its versatility. It is easy to understand the reason behind it — Python remains one of the most sought-after skills that these companies are looking for in data science and analytics professionals. According to engineers, deep learning frameworks available with Python APIs, in addition to the scientific packages coming from academia and industry, have made Python incredibly productive and versatile. According to Towards Data Science, there has been a lot of evolution in deep learning Python frameworks in the last two years where we saw the release of TensorFlow. As one developer noted on a forum, AI requires a lot of research, and with Python, one can validate their idea with even thirty code lines. In terms of application areas, ML scientists prefer Python as well. When it comes to areas like building fraud detection algorithms and network security, developers leaned towards Java; while for applications like natural language processing (NLP) and sentiment analysis, developers opted for Python, due to the wide collection of libraries that comes with it.","excerpt":"Python has been billed as the most popular language in the StackOverflow survey, where it even beat C# in popularity this year. StackOverflow has chronicled the incredible growth of Python, and has labelled it as the most preferred language for machine learning applications. In fact, according to the findings, Python was one of the most […]","categories":["IT Services"],"tags":["Java","Javascript","numpy","Python Libraries"],"author_name":"Richa Bhatia","publish_date":"2018-05-18T04:52:53","publication_year":"2018","word_count":793,"keywords":["data science","artificial intelligence","machine learning","AI","Python Libraries","neural network","Javascript","ML","numpy","NLP","deep learning","analytics","TensorFlow","Java"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","data science","analytics","TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-do-data-scientists-prefer-python-over-java\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":15759,"title":"Government considers having an analytics tool to furnish airline price trends","content":"Predictive analytics for the aviation industry The Indian government is thinking about introducing an analytics tool that will provide travelers with future ticket price trends. This is a step towards addressing the lingering concerns that travelers face because of steep fluctuation in airfares. The government’s proposal is being considered by the civil aviation industry. The proposal will additionally help in ensuring more transparency, as air ticket prices are generally driven by supply and demand metrics. The country’s domestic aviation market has been projected to be the world’s third largest by 2022. The ministry is also working on a project, which ensures seamless and paperless travel for the passengers, starting right from booking an air ticket. This proposal is part of the ‘Digiyatra‘ initiative. The project aims at providing airline travelers a “digitally unified flying experience.” Historical data analysis as well as “price curves with predictive data analytics” will help passengers project future airfares and efficient price discovery. Moreover, passengers can plan their trip more efficiently by making use of the historical data analysis and trends for airfares. Furthermore, the ministry is also looking at creating a “civil aviation data repository“. Such an analytics tool would first require a reservoir of data collected from airlines, airports, and travel portals. Analytics tool for airfares is one of the ideas that have been discussed by the ministry with aviation industry stakeholders last week. Airlines usually follow a dynamic pricing mechanism for the tickets that is mainly dependent on demand trends. Currently, many travel portals provide information on future airfare trends. Steep variations in air ticket prices, especially during natural calamities and festival seasons, have often been a matter of debate in various quarters. Civil Aviation minister Ashok Gajapathi Raju mentions that strategies for additional capacities have to be brought in place because additional capacities will bring down prices. Data with the Ministry proved that average airfares declined 18 per cent in 2016.","excerpt":"The Indian government is thinking about introducing an analytics tool that will provide travelers with future ticket price trends. This is a step towards addressing the lingering concerns that travelers face because of steep fluctuation in airfares. The government’s proposal is being considered by the civil aviation industry. The proposal will additionally help in ensuring […]","categories":["AI News"],"tags":["data analysis India"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-06-20T11:10:33","publication_year":"2017","word_count":319,"keywords":["Go","programming_languages:R","AI","R","ML","data analysis India","Git","RAG","Aim","analytics","predictive analytics"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","predictive analytics","R","Go","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/government-considers-analytics-tool-furnish-airline-price-trends\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":33040,"title":"Here’s Why Cybersecurity Threat Analyst Is The Hottest Job In 2019","content":"Image source: www.carbonblack.com If 2018 belonged to AI becoming mainstream, it was also a year marked with high-profile cyber threats and breaches. On the Indian turf, news of Aadhaar breach caused a massive ripple with stakeholders debating the privacy concerns. Details of Indian citizens, such as name, 12-digit unique ID and in some cases even bank account details were accessed by hackers. In the same year, Facebook, the social networking giant invited the wrath of government agencies and public alike when the alleged breach exposed close to 50 million of its users’ data. The attackers allegedly gained access to the users’ ID exploiting a feature in Facebook’s code. With attackers readying to exploit every loophole in the cyberspace, security breaches in 2019 are only going to be more ubiquitous. Taking all this into consideration, the need for organisations and governments to safe-proof their existing cyber infrastructure has become important. Role of A Data Protector Given the current cybersecurity landscape, the role of a threat hunter or cybersecurity threat analyst role is only going to be more prominent in 2019, as the fact attackers are going to turn to AI and ML as a means to target people. A threat hunter is essentially a security professional who uses manual or machine-assisted techniques to detect security threats in automated systems that would have been overlooked by a CISO or CIO. In other words, s\/he provides an additional layer of defence against advanced persistent threats (APTs). One of the major aspects that define a threat hunter’s profession is his field of knowledge, as s\/he is expected to have a thorough understanding of the working model of the business along with creative skills to interpret and communicate data effectively. One of the primary roles as a threat hunter is to deal with a large pool of data from which the person has to mine, pool and extract an organisation’s metadata. Hence, a sound grasp of machine learning is a key attribute to efficiently dispose of their duties. In order to deliver the responsibilities, the cybersecurity threat analyst will have to work with different kinds of software and tools to identify threats and possible adversaries. S\/he also has to constantly monitor security tools such as firewall, antivirus among many other key features revolving around security. Though this job role is still at a nascent stage in India, several companies have already foreseen the role of a threat hunter in safeguarding their business and have acted upon it. Companies like IBM India, Infosys and HSBC are some of the prominent organisation which is presently on a lookout for threat hunters with a minimum four-five years of experience to start off. In this article, we enumerate the key roles and requirements required to become a cybersecurity threat analyst: Key roles: Have to work with statistical and intelligence analysis software Develop security solutions to find threats Assessing magnitude of a security threat and then effectively communicating it to enterprise Tracking and providing real-time security alert and identifying them Monitoring end-point data and collecting event logs Creating correlations to identify attackers Key Requirements Knowledge of QRadar\/any SIEM solution, IOC discovery tools, intrusion detection systems etc. A sound knowledge about incident response process such as detecting advanced adversaries, log analysis using Splunk, ELK, or similar tools, and malware triage. An understanding of coding languages:  Perl, Python, Bash or Shell, PowerShell, or batch Knowledge about working of operating systems such as Window, Linux and network protocols such as the TCP\/IP stack, work A thorough understanding of the cybersecurity landscape including use cases and types of attack Strong knowledge about technical writing and documentation as a threat hunter is required ad to prepare security report on a regular basis In Conclusion Though the growth forecast for Indian enterprises looks promising, the biggest challenge before them is security regarding their digital assets. With companies investing millions of dollars to safeguard their assets, the role of a threat hunter or cyber security analyst is only going to be more pervasive in India. While the government’s push for initiatives such as Digital India, Aadhaar Card and Digital Locker would mean that a threat hunter’s role wouldn’t entirely be restricted to the private sector.","excerpt":"If 2018 belonged to AI becoming mainstream, it was also a year marked with high-profile cyber threats and breaches. On the Indian turf, news of Aadhaar breach caused a massive ripple with stakeholders debating the privacy concerns. Details of Indian citizens, such as name, 12-digit unique ID and in some cases even bank account details […]","categories":["AI Features"],"tags":["AI cybersecurity threat","Infosys"],"author_name":"Akshaya Asokan","publish_date":"2019-01-08T09:39:09","publication_year":"2019","word_count":698,"keywords":["Go","machine learning","Infosys","AI","ML","Git","Python","programming_languages:Python","ViT","GAN","AI cybersecurity threat","R"],"extracted_tech_keywords":["AI","machine learning","ML","Python","R","Go","Git","GAN","ViT","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/heres-why-cybersecurity-threat-analyst-is-the-hottest-job-in-2019\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10139391,"title":"‘Healthcare GCCs in India Help Cut Costs by 20%’","content":"According to the recent Nasscom-Zinnov report, by 2024, India’s Global Capability Centers (GCCs) will have grown to 1,700, generating $64.6 billion in export revenue and employing over 1.9 million people. Notably, over 15% of these employees are in the healthcare and life sciences sectors. Indian GCCs in healthcare are leading the efforts in drug discovery and development by harnessing computational chemistry, bioinformatics, and AI-driven platforms. These centres are accelerating the R&D of new drugs and bringing them to market faster than ever before. By integrating advanced analytics and data-driven insights, Indian GCCs are making significant contributions to global healthcare innovation, improving patient outcomes on a broader scale. In an exclusive conversation with AIM, Balasubramanian Sankaranarayanan, the CEO & president of Thryve Digital LLP, said, “In healthcare GCCs, we focus on a specific vertical, dedicating our entire efforts to it. Our goal is to find individuals deeply engaged in healthcare and provide them with opportunities to take on much larger career roles.” Headquartered in Chennai, Thryve Digital Health creates solutions that enable the “payvider” (pay provider) network with the intelligence, connectivity, and seamless integration demanded by its multi-platform healthcare systems. “While many large IT service companies have healthcare verticals, true healthcare expertise exists in only a small percentage of their workforce. Among the top healthcare verticals, you might find 50,000 people, but fewer than 1% truly understand the industry at a deep level,” Sankaranarayanan noted. Employees, Cost, and More As of 2023, over 80 healthcare & life science GCCs have employed more than  250k employees in India. The CEO mentioned that their objective is to build Thryve with that 1% of healthcare-focused experts who not only understand the domain but also grasp the business, technical, and customer challenges. The Indian government aims to have 2,400–2,500 GCCs operating nationwide by 2030, with a workforce of over 4.5 million people. However, given today’s cost-constrained economy and the need to fund future initiatives, businesses are under pressure to extract more value. Enterprises, especially in sectors like healthcare, are constantly reimagining the future, and that requires substantial investment. But where will these dollars come from? According to Sankaranarayanan, the answer lies in optimisation and efficiency. Labour arbitrage has already provided a lot of initial value, but now these companies need to go further. Establishing a GCC in countries like India presents an immediate 20% cost differential compared to working with service providers. These savings go directly back into the enterprise because GCCs aren’t focused on generating margins but on optimising operations. When companies set up their GCCs, they immediately see a 20% cost reduction. Whether they choose GCCs or service providers, the talent pool remains similar. He further pointed out that relying solely on a GCC or a service provider has its trade-offs. With service providers, enterprises benefit from learning across multiple customers and industries and the innovation investments made by these providers. GCCs, on the other hand, provide talent access but miss out on some of those external advantages. What GCCs offer is a short- to medium-term cost benefit, which can be as high as 20%. So, many companies, like Thryve, adopt a hybrid strategy. It operates with 80% of its own GCC setup while maintaining partnerships with key service providers to balance the best of both worlds. This allows them to gradually transition work from external partners into the GCC, especially higher-end work that might require proprietary frameworks or intellectual property not suitable for outsourcing. What’s GCC to Healthcare Sectors? Beyond the usual tasks undertaken by healthcare GCCs, like transforming data processes in clinical trials and enhancing supply chain management, Thryve Digital Healthcare is also looking to work towards the administrative part. “For example, in claims processing, understanding the intricacies across different claim types—whether commercial, government, or Medicare\/Medicaid—is crucial. This knowledge extends to how our platform handles claims, improves auto-adjudication rates, and streamlines processes, leading to greater efficiency and customer satisfaction. Even small improvements in these areas can lead to significant gains,” Sankaranarayanan said. Source: ANSR Global ANSR Research quotes a case study that focuses on a German healthcare solutions provider that expanded its digital health capabilities in India by leveraging AI and other advanced technologies to improve patient care and diagnostics. The company’s GCC in India, established in 2008 and spread across Bengaluru, Mumbai, and Gurugram, functions as an innovation hub with a workforce of 4,000-5,000 employees. The solutions developed at the Indian GCC include the AI Pathway Companion, which integrates patient data across different sources to aid in better decision-making; the AI-Rad Companion, which automates analysis and quantification of clinical imaging data; and the Patient Experience App, which enhances MRI patient experiences using Augmented Reality (AR). What’s Next? Scaling data and AI in an enterprise context is crucial, and it’s something many GCCs are actively working on. If they can achieve AI at scale, it becomes a straightforward lift and shift to apply those insights to even rural India, where scalable and equitable access is essential to drive the right outcomes. Many of these use cases will initially focus on more standardised medical protocols, like retinopathy, where there’s a clear SOP. According to Sankaranarayanan, radiology accounts for about 70% of AI models, while cardiovascular applications make up another 15%. In India, ophthalmology is a key area, and AI models are already delivering high-quality results. Meanwhile, the healthcare sector is witnessing a strong influx of funding. Recently, Bengaluru-based startup Even Healthcare, which provides facilities such as consultation and hospitalisation to its members, raised $30 million in funding led by Silicon Valley-based Khosla Ventures. In another update, Microsoft has announced new healthcare data and AI tools, including a collection of medical imaging models, a healthcare agent service, and an automated documentation solution for nurses.","excerpt":"As of 2023, over 80 healthcare & life science GCCs have employed more than 250k people in India.","categories":["GCC"],"tags":["AI (Artificial Intelligence)","GCC","GCCs","Microsoft","Thryve Digital LLP"],"author_name":"Vidyashree Srinivas","publish_date":"2024-10-25T13:26:29","publication_year":"2024","word_count":951,"keywords":["Thryve Digital LLP","Go","GCC","AI","R","ML","Scala","Git","RAG","Ray","Aim","analytics","AI (Artificial Intelligence)","GCCs","Microsoft"],"extracted_tech_keywords":["AI","ML","analytics","Aim","Ray","RAG","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/gcc\/healthcare-gccs-in-india-help-cut-costs-by-20\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":56828,"title":"How To Make Sure Your Robot Doesn’t Drop Your Wine Glass","content":"From microelectronics to mechanics and machine learning, the modern-day robots are a marvel of multiple engineering disciplines. They use sensors, image processing and reinforcement learning algorithms to move the objects around and move around the obstacles as well. However, this is not the case when it comes to handling objects such as glass. The surface properties of glass are transparent, and non-uniform light reflection makes it difficult for the sensors mounted on the robot to understand how to engage in a simple pick and place operation. To address this problem, researchers at Google AI along with Synthesis AI and Columbia University devised a novel machine-learning algorithm called ClearGrasp, that is capable of estimating accurate 3D data of transparent objects from RGB-D images. Google claims that the model identifies transparent objects quite well even if the object is situated in a patterned background or differentiating between transparent objects partially occluding one another. Overview Of ClearGrasp ClearGrasp uses 3 neural networks: A network to estimate surface normals Network for occlusion boundaries (depth discontinuities) And a network that masks transparent objects This mask is used to remove all pixels belonging to transparent objects, and then fill the depths correctly. A global optimisation module then starts extending the depth from known surfaces to guide the shape of the reconstruction, and the predicted occlusion boundaries to maintain the separation between distinct objects. Transparent objects can confuse sensors because optical 3D sensors are driven by algorithms that assume all surfaces reflect light evenly in all directions. Hence, most of the depth data from transparent objects are invalid or contain unpredictable noise. While the distorted view of the background seen through transparent objects is confusing, there are still clues about the objects’ shape. Transparent surfaces exhibit mirror-like reflections that show up as bright spots in a well-lit environment. Convolutional neural networks (CNN) can use these reflections to infer accurate surface normals, which then can be used for depth estimation. Google AI’s ClearGrasp can work with inputs from any standard RGB-D camera, using deep learning to accurately reconstruct the depth of transparent objects and generalise to completely new objects unseen during training. The model was trained on a large-scale synthetic dataset that contains more than 50,000 photorealistic renders representing the surface curvature, edges, and depth, useful for training a variety of 2D and 3D detection tasks. Going Forward To check the qualitative performance of ClearGrasp, 3D point clouds were constructed from the input and output depth images, which can be crucial for applications, such as 3D mapping and 3D object detection. Previous methods required prior knowledge of the transparent objects along with maps of background lighting and camera positions. By using ClearGrasp’s output of the raw sensor data, a grasping algorithm on a UR5 robot arm saw significant improvements in the grasping success rates of transparent objects. Better sensing of transparent surfaces would not only improve safety but could also open up a range of new interactions in unstructured applications. The modern-day robotics can deliver your amazon package over the air, throw a ball down the hoop, arrange bottles in the pegs and can even meticulously move around to make the intricate of paintings. And, now with ClearGrasp, they can generate AR visualisations on glass tabletops and even be seen serving wine soon!","excerpt":"From microelectronics to mechanics and machine learning, the modern-day robots are a marvel of multiple engineering disciplines. They use sensors, image processing and reinforcement learning algorithms to move the objects around and move around the obstacles as well. However, this is not the case when it comes to handling objects such as glass. The surface […]","categories":["Deep Tech"],"tags":["Deep Learning"],"author_name":"Ram Sagar","publish_date":"2020-02-18T11:11:08","publication_year":"2020","word_count":545,"keywords":["Go","machine learning","TPU","AI","neural network","Aim","deep learning","object detection","Deep Learning","CNN","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","Aim","object detection","TPU","R","Go","CNN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-make-sure-your-robot-doesnt-drop-your-wine-glass\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10119440,"title":"SML Unveils Hanooman, Sets Ola Krutrim On Fire","content":"The much awaited Indic AI chatbot to compete with OpenAI’s ChatGPT and Ola’s Krutrim is finally here. SML has silently released Hanooman.ai chatbot, which is surprisingly fast. Click here to chat with Hanooman. The model is very fast and gives quick and crisp responses to asked questions. Currently released in Alpha, the model also allows users to select between nine languages. Upon testing, AIM found that the responses it generates are very short even after specific prompts. It also sometimes auto translates languages even while the prompt was different. Given the shorter context length, it also can’t keep track of a conversation as it replies to each question independently which defeats the purpose of a personal AI assistant. Moreover, the current cut-off data of the model is also April 2022, which is a little outdated. SML’s Hanooman is named after the Hindu deity Hanuman. “Hanuman is a great example of responsible power. Despite being the most-powerful entity, he never used his power for selfish needs,” said Vishnu Vardhan, founder of SML and Vizzhy, in an exclusive interview with AIM. “We don’t want it to be like ChatGPT, which suffers from the ‘I’m God and I know everything’ syndrome. We don’t want to simply replicate its success [but be more than that],”said Vardhan, adding that they are currently focusing on specific use cases for Hanooman. Meanwhile, Ola’s Krutrim chatbot is unable to respond to a lot of queries and still hallucinates a lot. It does not give responses to several questions citing the reason that it is an AI model without up to date information. Vardhan had also promised that Hanooman will come out as an open source model in several different sizes, which is still awaited.","excerpt":"Currently released in Alpha, the model also allows users to select between nine languages.","categories":["Deep Tech"],"tags":["Krutrim"],"author_name":"Mohit Pandey","publish_date":"2024-05-02T12:16:49","publication_year":"2024","word_count":286,"keywords":["Replicate","Go","ChatGPT","Krutrim","OpenAI","AI","ML","GPT","Aim","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Aim","R","Go","GPT","Replicate","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/sml-unveils-hanooman-sets-ola-krutrim-on-fire\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":18294,"title":"AI-Powered Fraud Prevention Startup ThirdWatch Raises Angel Funding","content":"Artificial Intelligence-powered fraud prevention startup ThirdWatch has raised an undisclosed amount of angel funding from Indian Angel Network (IAN). The one-and-a-half year old Gurgaon-based startup’s key product is Mitra, an AI-based programme which evaluates and flags every transaction as ‘genuine’ or ‘fraudulent’ in real time. It works with the help of a Trust Score which is generated with the help of machine learning algorithms, browsing behaviour analysis, device fingerprinting, location profiles and other evaluation parameters which determines the transactions’ authenticity. Founded by Adarsh Jain, CEO, and Shashank Agarwal, CTO, ThirdWatch is reportedly India’s first AI-based startup in the area. According to a media statement, Keshav Sanghi, former managing director of Goldman Sachs India and founder of investment firm VentureWorks India, as well as financial advisory firm Batlivala & Karani Securities have also invested in ThirdWatch. “Over the past few years, we have seen e-commerce fraud grow both in scale and variety. I believe that ThirdWatch’s proven AI technology and unique analytical model truly differentiates the company in a fast-growing global fraud prevention marketplace,” said Rahul Agarwalla, IAN’s lead investor. Established in 2006, the IAN is India’s first and the world’s largest business angel network with over 450 members across the world. The statement added that the ThirdWatch’s relevance will only increase in the coming times as the Indian e-commerce industry is expected to jump from $30 billion in 2016 to $120 billion by 2020. According to ThirdWatch, Mitra’s AI keeps on evolving and becoming more intelligent over time as it sees and grasps more and more transactions. That’s why, Mitra chiefly deals with the problems around product returning to origin, payment fraud, promotional code fraud and account abuse, among others. “ThirdWatch has successfully reduced return to origin (RTO) problem in online orders by more than 80 percent within three months of going live for their current clients. We also offer a two-month free trial period so that our clients can validate the ThirdWatch AI’s efficiency, accuracy and value addition before they start paying,” said Shashank Agarwal, co-founder and CTO of ThirdWatch.","excerpt":"Artificial Intelligence-powered fraud prevention startup ThirdWatch has raised an undisclosed amount of angel funding from Indian Angel Network (IAN). The one-and-a-half year old Gurgaon-based startup’s key product is Mitra, an AI-based programme which evaluates and flags every transaction as ‘genuine’ or ‘fraudulent’ in real time. It works with the help of a Trust Score which […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Prajakta Hebbar","publish_date":"2017-10-11T13:01:01","publication_year":"2017","word_count":341,"keywords":["Go","funding","machine learning","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Rust","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","R","Go","Rust","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/thirdwatch-raises-angel-funding-fraud-prevention\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10925,"title":"Biggest tech companies collaborate to give a boost to AI","content":"Artificial Intelligence has been a tried and tested concept since its inception with a not-so smooth path on its acceptance. If we look at the timeline of AI, there have been many dull phases but has always overcome the hurricane with its bang on presence the very next time. With this roller coaster ride of AI’s acceptance and rejection, one thing which is undeniable is that AI could never be ignored. Proving on the same and boosting the concept of AI even further, five of the biggest names in the tech industry viz. Amazon, DeepMind\/Google, Facebook, IBM and Microsoft have collaborated on creating a non-profit organization that would contribute in advancing the public understanding of artificial intelligence technologies- an act that could be touted as self-governance! What would it bring on the table? Along the partnership comes the assurance that it would explore the best of opportunities in AI space that are still largely untapped. It would also aim at overcoming the challenges that this industry currently faces. The point worth noting here is that the founding companies of this partnership have had an excelling path in the area of AI. Be it IBM’s Watson, Amazon’s Alexa, Google’s London-based subsidiary DeepMind or Microsoft’s Cortana, they all have a background in AI to boast off! Further elaborating on the objectives: The Board named Partnership on Artificial Intelligence to Benefit People and Society will bring together academicians, non-profits and specialists in ethics and policy. The founding members of the organization would be put up with the charge of contributing financial and research resources to ensure its smooth functioning. They would also aim at sharing leadership with independent third parties, academicians, user group advocates, industry domain experts all alike. Conducting research, recommending best practices and publishing research under an open license in areas such as ethics, fairness, transparency, privacy, reliability, collaboration between people and AI systems and robustness of the technology, are some of the practices that this partnership would ensure. The alliance, like many would perceive, will not lobby government or other policymaking bodies. Who all can be a part of it? For those wondering on who all would make to the board of members, the structure has been laid out such that along with the large names in the industry, non-corporate groups can have an equal leadership. There are discussions with various organizations- both professional and scientific such as the Association for the Advancement of Artificial Intelligence (AAAI) and non-profit research group such as Allen Institute for Artificial Intelligence (AI2). There can be more additions in the list with time. A few misses: While the move has fetched a lot of appreciation from all the corners, there have been certain misses that have been pointed out. Two of the biggest names in the AI driven industry, Apple and Tesla’s absence has been the talking point. Whether they become a part of the much talked off partnership or not is for the time to tell. The road ahead… Ensuring a smooth run in the field of Artificial Intelligence, this collaboration has been a welcoming move. It definitely points towards a bright future where it would benefit both the technology and society alike. It brings a vital collaboration in the picturesque that would have people and machines solving the long running problems in the most effective way be in any space- healthcare, education, manufacturing, transportation or others.","excerpt":"Artificial Intelligence has been a tried and tested concept since its inception with a not-so smooth path on its acceptance. If we look at the timeline of AI, there have been many dull phases but has always overcome the hurricane with its bang on presence the very next time. With this roller coaster ride of […]","categories":["AI News"],"tags":["AI Companies"],"author_name":"Srishti Deoras","publish_date":"2016-10-11T10:16:39","publication_year":"2016","word_count":564,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Aim","ViT","AI Companies","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/biggest-tech-companies-collaborate-give-boost-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166162,"title":"China&#8217;s Chitu to Challenge Dependence on NVIDIA Chips","content":"NVIDIA leads the AI chip market. However, due to US hardware restrictions on China and the rise of non-NVIDIA GPUs, some businesses seek to lessen their reliance on NVIDIA products. A team associated with China’s Tsinghua University is joining the effort by releasing a new open-source AI framework, Chitu, made available under the Apache-2.0 license. Chitu is a high-performance inference framework for large language models, focusing on efficiency, flexibility, and availability. With its initial release, it supports various mainstream large language models, including DeepSeek, LLama series, Mixtral, and more. The report by South China Morning Post states that the framework can operate on chips made in China, challenging the supremacy of NVIDIA’s Hopper series graphics processing units (GPUs). The information is referenced from a joint statement on Friday by the startup Qingcheng.AI and a team led by Zhai Jidong, computer science professor at Tsinghua University. The startup was founded in 2023 by Jidong, and his students from Tsinghua University, with him serving as chief scientist. The report states that it is backed by Beijing’s municipal fund for the AI industry. The framework’s GitHub page mentioned, “We not only focus on the popular NVIDIA GPUs, but pay special attention to all kinds of hardware environments, including legacy GPUs, non-NVIDIA GPUs and CPUs. We aim to provide a versatile framework to encounter the diverse deploying requirements.” The report mentions that the company claims to have achieved a 315% increase in model inference speed while reducing GPU usage by 50% compared to foreign open source frameworks. This is as per the test with DeepSeek-R1 using NVIDIA’s A800 GPUs. The team states that the framework is ready and deployed for real-world production.","excerpt":"NVIDIA’s dominance is being challenged by startups.","categories":["AI News"],"tags":["china ai"],"author_name":"Ankush Das","publish_date":"2025-03-17T18:19:15","publication_year":"2025","word_count":278,"keywords":["programming_languages:R","AI","Git","china ai","Aim","llm_models:Llama","GitHub","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Git","GitHub","startup","llm_models:Llama","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/a-new-chinese-open-source-framework-aims-to-challenge-dependence-on-nvidias-chips-on-ai-models\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10059971,"title":"How this robotics startup uses AI to power its tethered drones","content":"The announcement of ‘Kisan Drones’ in the Union budget 2022 to promote drone technology for crop assessment, digitisation of land records, spraying of insecticides and nutrients, is poised to give a major fillip to drone startups. The finance minister said startups would be encouraged to facilitate ‘Drone Shakti’ through varied applications and Drone-As-A-Service (DrAAS). In addition, the government will start required courses for skilling in select ITIs. Reacting to the announcement in the budget, Athishay Jain, Co-founder and COO, ISPARGO, said adopting cutting-edge technology will increase farmers’ income. “Using drones in farms is a low-risk proposition, definitely a great move from the government. If drone technology is adopted on a large scale which would increase the demand, the product will get better, and the ecosystem would naturally develop,” he added. Mangaluru-based robotic startup ISPARGO uses drones to apply insecticide on areca nut trees.  Jain said the startup uses ultra-lightweight, compact power conversion technology to power the drones from the ground from any AC source. AI\/ML in drone AI\/ML models are used to develop drones capable of landing on a fixed image or a marker on the ground. The same solution is being extended to follow a moving vehicle. This complex AI\/ML model captures the images, processes them, and predicts the next set of navigation points for the drone to follow, all within milliseconds. ISPARGO, established in 2017, focuses on developing and customising ‘drones for work’. Batteries prominently power drones, but they do not last beyond 30 minutes. Any long endurance work using drones will need 15 to 20 sets of batteries which is economically unviable. To upgrade the drone technology, ISPAGRO has developed non-battery powered drones which can work for a whole day with improved efficiency. They are also looking into customising the firmware for specific use cases like painting or pesticide spray to make operations almost autonomous. The unique proposition of ISPARGO is the tethered drones used for surveillance, plantation crop spray and inspection\/monitoring. The salient features of these drones are: OFC based communication and video feed.Single-button launch and land.Video and message encryptionLong endurance AI\/ML. Use cases For border surveillance\/intrusion detection: Tethered drones solution comes with quick and easy deployment; secured data transmission with OFC; no cabling issues; supports 24 hours of uninterrupted video feed; thermal EO\/IR, RGB cameras. For temporary mobile towers: Tethered drone solution supports all-terrain operation; easy plus and fly operation on the go; easily transportable for quick fix deployment; requires minimum resources for installation; can be used for communication and surveillance for intrusion detection. For agriculture\/pesticide spray: Tethered drone solution has one skilled pilot who can cover more trees in a short time; precise spray, low wastage; timely and effective spray increase productivity; onboard camera aids in crop health analysis. For windmill inspection: Tethered drone solution has quickly deployable precise inspection; no operator insurance and training cost; can fit sophisticated camera for stress\/heat analysis. Challenges Lack of reliable drone components in India, majorly sourced from China. The few manufacturers in India are struggling due to a lack of demand. The open-source drone firmware is generic for categories of drones. Reliability and stability can be achieved by modifying the existing firmware or writing your stack. Unfortunately, there are very few resources available in India to develop drone firmware. Huge R&D costs involve multiple engineering streams (Software, Electronics, Aeronautical). Hence the majority of the companies are focused on developing dashboards and UTM software around drone operations. The government has eased the policies for using drones but hasn’t earmarked a significant amount of capital to promote the ecosystem. Solutions Focus on building a drone ecosystem; battery tech, motors and firmware are the main thrust areas.Identify the key players in these three thrust areas within India and support them in R&D.","excerpt":"The startup uses ultra-lightweight, compact power conversion technology to power the drones from the ground from any AC source.","categories":["AI Startups"],"tags":["AI Startups"],"author_name":"Poornima Nataraj","publish_date":"2022-02-07T10:00:00","publication_year":"2022","word_count":622,"keywords":["Go","API","AI","ML","Git","RAG","Ray","ViT","Rust","R","AI Startups"],"extracted_tech_keywords":["AI","ML","Ray","RAG","R","Go","Rust","Git","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-robotics-startup-uses-ai-to-power-its-tethered-drones\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075587,"title":"Scientists Draw New Revolutionary Proteins with the Help of AI","content":"A recent breakthrough in AI has led scientists to draw new revolutionary proteins. The advancements are spearheaded by a team of scientists led by David Baker, Biochemist, University of Washington (UW), Seattle, who reported designing molecules in seconds instead of months. This is expected to lead to many new vaccines, sustainable biomaterials, and treatments. The first-ever medicine—a COVID-19 vaccine—to be made from a novel protein designed by humans was authorised by South Korean regulators, which is based on a spherical protein called a ‘nanoparticle’. The paper titled, ‘Robust deep learning–based protein sequence design using ProteinMPNN’, published by biologists at the University of Washington School of Medicine explains how machine learning can be used to create protein molecules more accurately and quickly. Baker laboratory spent over three decades on making new proteins, with the help of a software called Rosetta. The process was split into steps for deriving the final protein. Firstly, researchers conceived a shape for a novel protein, majorly by cobbling together bits of other proteins. The software would then deduce a sequence of amino acids that corresponded to this shape. “Proteins are fundamental across biology, but we know that all the proteins found in every plant, animal, and microbe make up far less than one percent of what is possible. With these new software tools, researchers should be able to find solutions to long-standing challenges in medicine, energy, and technology,” said David Baker, senior author and professor of biochemistry, University of Washington School of Medicine. The team developed the protein with an approach called ‘hallucination’, where researchers would feed random amino-acid sequences into a structure-prediction network. AlphaFold, and a similar tool called ‘RoseTTAFold’, were trained to predict the structure of individual protein chains, which further led to the discovery of using such networks to model assemblies of multiple interacting proteins.","excerpt":"Advancements ​​in the AI tools has led the researchers to come up with proteins that will lead to many new vaccines, sustainable biomaterials and treatments.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AlphaFold","covid-19 vaccine","Machine Learning","RoseTTAFold"],"author_name":"Bhuvana Kamath","publish_date":"2022-09-19T16:55:19","publication_year":"2022","word_count":302,"keywords":["AlphaFold","machine learning","programming_languages:R","AI","Machine Learning","covid-19 vaccine","deep learning","RoseTTAFold","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","deep learning","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/scientists-draw-new-revolutionary-proteins-with-the-help-of-ai\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10139725,"title":"Did GitHub Copilot Just Kill Cursor?","content":"At GitHub Universe, GitHub unveiled a new era of multi-model choice for Copilot, introducing Anthropic’s Claude 3.5 Sonnet, Google’s Gemini 1.5 Pro, and OpenAI’s o1-preview and o1-mini, empowering developers and enterprises to select models tailored to specific coding needs across tools like VS Code and GitHub.com. “The next phase of AI code generation will not only be defined by multi-model functionality but by multi-model choice,” said GitHub, underscoring its commitment to open developer choice. Alongside this, GitHub previewed “Spark,” a natural language tool to create AI-integrated applications, supporting GitHub’s vision to reach 1 billion developers. “In 2024, we experienced a boom in high-quality large and small language models that each individually excel at different programming tasks. There is no one model to rule every scenario, And developers expect the agency to build with the models that work best for them,” said GitHub CEO Thomas Dohmke at the GitHub Universe event. “Our mission has always been to get the most frontier model capabilities in as many people’s hands as quickly as possible. If we can’t deploy them now, they’ll be obsolete in six months to a year,” said Anthropic’s chief science officer Jared Kaplan, sharing insights on Claude 3.5’s exceptional appeal. Further broadening their multi-model strategy, GitHub revealed another major integration with Google’s Gemini 1.5 Pro, a model known for its “two-million-token context window and native multimodality,” designed to process “code, images, video, and text simultaneously,” as GitHub demonstrated live. In an innovative Copilot update for VS Code, GitHub’s senior director of developer advocacy Cassidy Williams, showcased Copilot’s expanded capabilities, including custom instructions, multi-file editing, and integration with GitHub’s search and intent detection tools, giving users control over how they code. “This is your brand-new experience for VS Code,” said Cassidy, emphasising that the latest features—model selection, repo indexing, and GitHub Copilot extensions—are “available to every single Copilot user, from individuals to enterprises,” starting from this week. For the first time, GitHub Copilot will also extend beyond VS Code, with support announced for Apple’s Xcode, enhancing accessibility for mobile and desktop developers alike, regardless of their preferred platform or workspace. More Choice & Agency to Developers Moving on from Codex. Earlier this month, Microsoft’s GitHub introduced OpenAI’s o1-preview and o1-mini models on Azure. These models are available through GitHub Copilot and Models, and developers can sign up to access OpenAI o1 in GitHub Copilot Chat via VS Code and the GitHub Models playground. Now, the integration of Claude 3.5 Sonnet with Github is live. “Claude 3.5 Sonnet excels at coding tasks and is broadly used by developers for its exceptional grasp of software engineering principles and ability to tackle complex programming challenges,” said Anthropic’s Kaplan. Meanwhile, in a recent announcement, AWS introduced a new capability for developers: inline chats with Q Developer, its GenAI assistant for software development. This feature, powered by the upgraded Claude 3.5 Sonnet and optimised custom models, eliminates the need for a chat panel, enhancing productivity for development teams. Google’s Gemini 1.5 Pro’s integration with GitHub will be available in a few weeks. Developers are seeking different models for tasks like code generation, refactoring, and optimisation, enabling flexible, efficient, and high-quality coding across programming environments. “Gemini models excel at this and are accessible on widely used developer platforms and environments – including now with GitHub Copilot – so millions of developers globally can benefit from trusted, enterprise-grade AI through Google Cloud,” said Thomas Kurian, CEO at Google Cloud. Kills Cursor’s Vibes In a recent podcast with Lex Fridman, Cursor’s co-founders highlighted how Microsoft-backed GitHub Copilot is falling behind startups in terms of innovation. They emphasised how the tech giant lacked the research and experimentation necessary to really push the ceiling. The sentiment also remains that Copilot did not have any ‘alpha features’ for a very long time. Even Y Combinator funded several open-source AI code editors like Continue, Pearl, Void, Type, and Melty, among others. The Copilot experience is finally improving with substantial updates to GitHub Copilot in VS Code, Copilot Workspace, GitHub Models, and Copilot Autofix. We have written extensively about whether it is too late for Microsoft’s VS Code or GitHub Copilot to catch up with the market. Now the tables have turned. The days may be numbered for Cursor and other AI coding assistants, as GitHub Copilot advances toward becoming the go-to cross-platform solution for developers. “It is clear the next phase of AI code generation will not only be defined by multi-model functionality, but by multi-model choice. Today, we deliver just that,” said Dohmke. With this, developers can choose the right model for the right use case or continue to let Copilot use its powerful default. GitHub is allowing developers to build with an array of leading models in the workflows they’re accustomed to. Aman Sanger, co-founder of Anysphere (the creator of Cursor), said, “I think the Cursor a year from now will need to make the Cursor of today look obsolete.” This indicates if Cursor is to stay relevant in the competition, it has to pull its socks up. “You can wax poetic about moats and brand that and this is our advantage, but I think in the end, just if you stop innovating on the product, you will lose,” said Micheal T, a co-founder of Anysphere. It’ll be interesting to see how Cursor responds, and how the competition will evolve.","excerpt":"Brings Anthropic’s Claude 3.5, Google’s Gemini 1.5, and OpenAI’s o1-preview to GitHub Copilot.","categories":["Global Tech"],"tags":["GitHub"],"author_name":"Aditi Suresh","publish_date":"2024-10-29T22:52:41","publication_year":"2024","word_count":889,"keywords":["Anthropic","GenAI","OpenAI","AI","AWS","R","Ray","GitHub","small language models","Claude 3.5","Azure"],"extracted_tech_keywords":["AI","GenAI","OpenAI","Claude 3.5","Anthropic","Ray","small language models","AWS","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/did-github-copilot-just-kill-cursor\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":63891,"title":"What Is Azure Confidential Computing Built On Intel Hardware?","content":"Microsoft, with its Azure DCsv2-Series virtual machines (VMs), is now aiming to ensure data security via confidential computing. We know data security is a topic that should not be taken lightly, particularly when the majority of workloads have been shifted to the cloud. In a world where cloud platforms process payment transactions and financial records, security becomes paramount. Microsoft, with its Azure DCsv2-Series virtual machines (VMs), is now aiming to ensure data security via confidential computing. Microsoft is a part of Confidential Computing Consortium along with other large tech players. The whole idea behind confidential computing is achieving world-class security not only when data is being processed but also while it is at rest or in transit. The Confidential Computing Consortium was developed under the supervision of Linux Foundation. With the new VM series, Microsoft sets a milestone of becoming the first company to offer virtualisation products as part of confidential computing. Confidential computing relies on hardware-based trusted execution environment (TEE) and makes processed data so secure that cloud admins and data centre operators with physical access to servers cannot acquire. Confidential computing is able to encrypt data at all levels, enabling things like multi-party computation in coordination with different entities at once. This application of encrypted and secure distributed computation can be applied to use cases like doing analytics on combined financial transactional data from different banks for fraud detection, or processing health records in an anonymised fashion for tracking health trends and overall diagnostics. For maintaining the open-source consortium, it has taken involvement from every different technology company, setting up the ground rules through a nonprofit foundation to ensure everything is fair. Confidential computing can be useful for distributed apps and more advanced applications such as trusted remote virtual machines. Intel’s Software Guard Extensions (SGX) The consortium will achieve Trusted Execution Environments (TEEs), especially with the Intel Software Guard Extensions (SGX) development kit. Intel SGX is the most researched, tested, and deployed application isolation technology in the market today, according to Intel, it enables application developers to partition their applications into private regions of memory called enclaves. Besides, it is designed to be protected from higher-level processes, including even the OS and hypervisor. This would enable application and frameworks developers to develop software that can be used across different cloud platforms and Trusted Execution Environment (TEE) models. As VMs operate on specialised servers from Intel Software Guard Extensions (SGXs), the hardware protects and encrypts while it is being processed by CPUs. Even the operating system or hypervisor would not be able to gain access to data as it is being processed from anyone. This drastically minimizes the attack surface in future usage in the enterprise for confidential computing in multi-cloud scenarios. Various untrusted entities can distribute transactions but protect their confidential or proprietary data from other parties by using enclaves. Giving Way To New Data Processing Scenarios Usually, data goes encrypted while it is stored or in transition by service providers, but it generally does not get encrypted when it’s in use. The Confidential Computing Consortium plans to concentrate on this last security problem when data gets processed in memory. Protecting the data being used means it stays hidden even in unencrypted form during processing except to the code approved to access it. Confidential computing will allow encrypted data to be processed in memory without endangering the rest of the system, decrease exposure for sensitive data and give more comprehensive control and transparency for users. That way, confidential computing is expected to give birth to new types of scenarios like training multi-party dataset for machine learning models, letting various parties collaborate to have specific models or deeper analytics without providing other parties access to the data. The technology may also allow confidential query processing in database engines within secure enclaves, which will remove the need to trust database operators.","excerpt":"Microsoft, with its Azure DCsv2-Series virtual machines (VMs), is now aiming to ensure data security via confidential computing. We know data security is a topic that should not be taken lightly, particularly when the majority of workloads have been shifted to the cloud. In a world where cloud platforms process payment transactions and financial records, […]","categories":["Global Tech"],"tags":["Azure","Azure cloud platform","Azure Machine Learning","confidential computing","Intel","intel server","Microsoft Azure"],"author_name":"Vishal Chawla","publish_date":"2020-05-02T16:00:00","publication_year":"2020","word_count":641,"keywords":["Go","machine learning","Rust","intel server","AI","cloud_platforms:Azure","R","Azure cloud platform","Aim","Azure Machine Learning","analytics","Microsoft Azure","confidential computing","Azure","Intel","fraud detection"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","fraud detection","Azure","R","Go","Rust","cloud_platforms:Azure"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/what-is-azure-confidential-computing-built-on-intel-hardware\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10134096,"title":"Meta FAIR Releases Transfusion for Multimodal AI Training","content":"In a collaborated effort with Waymo & University of Southern California, Meta FAIR released its research on the importance of multi-modal generative models. Transfusion aims to unite and simplify the gap between discrete sequence modeling and continuous media generation. The Transfusion Model The model is trained equally on text and image. Per Meta, Transfusion is more advanced than quantising images and training a language model over discrete image tokens. The model’s performance can be enhanced through “modality-specific” encoding and decoding layers. The model predicts the next word in a sequence. Trained on improving predictions, it reduces the difference between guessing and actual words. It is imperative to note that with 7 billion parameters and 2 trillion multi modal tokens, Transfusion is at par with other larger models that create image and text – and outperforms models like DALL-E 2 and SDXL. It works better than Chameleon as it takes lesser computing power and generates better results. One limitation is perhaps that diffusion models do not perform at par with traditional language models. A lot of research is yet to be done in this area to improve overall performance. Transformer’s Uniqueness & the Future of Innovation in AI Research What differentiates Transformer from the rest is its unified architecture that runs end to end to generate text and images. Existing models like Flamingo, LLaVA, GILL, and DreamLLM combine separate architectures for different types of data, which are trained separately. The goal of this Transfusion is to synergise two modalities in a single joint model – with each of them fulfilling their objective. The incentives are that these are versatile, resource efficient, and cost effective for handling different types of data without any additional costs.","excerpt":"Transfusion is a state-of-the art approach at advancing text and image modalities","categories":["AI News"],"tags":["Generative Pre-Trained Transformer","Meta AI"],"author_name":"Aditi Suresh","publish_date":"2024-08-29T17:07:46","publication_year":"2024","word_count":283,"keywords":["Go","DALL-E","Meta AI","AI","Modal","innovation","ML","diffusion models","Aim","Generative Pre-Trained Transformer","AI research","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","DALL-E","diffusion models","innovation","Modal","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-fair-releases-transfusion-for-multimodal-ai-training\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10084443,"title":"Honda &#038; Sony Launch Concept Car Afeela","content":"Japanese automakers Honda and software and entertainment conglomerate Sony unveiled their first concept car Afeela at the US technology trade show CES, 2023. This comes after the two companies announced their partnership to build and sell electric vehicles and develop a new “mobility service platform” to use with the EVs in March 2022, said Sony chief Kenichiro Yoshida. While Sony is responsible for “imaging, sensing, telecommunication, network, and entertainment technologies”, Honda will manage the manufacture of the vehicles. The initial batches of “moving entertainment spaces” will be delivered to customers in North America in early 2026. Initially planned to be produced at Honda’s North American factory, the cars will have Level 3 automated driving capabilities but with certain limitations. At level three autonomy, the vehicle can operate in traffic, but the human driver must take control when the system signals it. Features of Afeela Afeela cars will be powered by US semiconductor company Qualcomm’s “Snapdragon Digital Chassis” technology, said its CEO Cristiano Amon. The four-door sedan, driven on stage at the show, had a digital display “media bar” that could be used to interact and exchange information with people outside the vehicle, as demonstrated on stage, and Lidar for autonomous driving. Built on Sony’s sensor expertise, the Afeela concept car is packed with 45 cameras and sensors on the interior and exterior of the vehicle to assure safety. In addition, 3D interfaces for the vehicles will be designed using Unreal Engine graphics technology by EpicGames. Highlights on the auto tech industry from CES 2023 BMW reveals I-Vision Dee concept car that uses e-ink technology to switch between 32 body colours, while its windshield enables a view of “mixed reality.” Qualcomm Technologies presented a new concept car to demonstrate how the Snapdragon Digital Chassis solutions integrate technology to create individualised experiences. It includes features like 5G, Wi-Fi, Bluetooth, vehicle-to-vehicle communication, immersive infotainment, driver assistance, and enhanced safety. In addition, Qualcomm unveiled the Snapdragon Ride Flex SoC, a new automotive processor made exclusively for infotainment and driver-assistance technologies. Holoride, the startup supported by Audi, launched a smart speaker-sized device that can retrofit any vehicle to make it VR-enabled. The device employs an HTC Vive Flow VR headset, to generate a virtual environment that imitates the vehilcle’s motion. Mercedes disclosed its plan to build its own-brand electric car charging network with more than 10,000 active charge points by 2027. American MNC Goodyear introduced a 70% sustainable tire. LG Electronics and Canadian automotive parts manufacturer Magna announced a “technical collaboration” to build automated driving infotainment system proof of concept. Automaker Stellantis is assisting Archer Aviation in building an eVTOL electric aeroplane called Midnight. Why are tech companies betting big on EVs? Smartphone and tech manufacturing companies like Apple, Oppo, Xiaomi, Samsung, Huawei, and Google have entered the electrical automotive space, as most of them have peaked in the phones and IoT products industry. These businesses have transitioned to the automotive sector due to the government’s current incentives for creating and manufacturing environmentally friendly automobiles. Although they are ahead of the software companies, seasoned members of the automobile sector like Maruti, Mahindra, and TATA Motors lack the inventiveness and technological advancements that LG, Sony, and other businesses are prepared to spend in. There has been an increase in the market for electric vehicles and other energy-efficient technologies as a result of the United Nations issuing a red alert on climate change. All types of technology companies are stepping up to make the least environmentally friendly industry as sustainable as they can.","excerpt":"While Honda is responsible for manufacturing the EVs, Sony will manage the imaging and sensory controls.","categories":["AI News"],"tags":["ces","electric cars","Sony"],"author_name":"Shritama Saha","publish_date":"2023-01-09T15:12:42","publication_year":"2023","word_count":586,"keywords":["electric cars","Go","ces","programming_languages:R","AI","programming_languages:Go","Git","ai_applications:autonomous driving","RAG","Sony","R","startup"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","startup","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/honda-sony-launch-concept-car-afeela\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69394,"title":"Why Google Data Studio Lags Behind Other Top BI Tools","content":"Google is hands down one of the most powerful names when it comes to tech companies but talking about business intelligence and data visualisation tools, offering by Google called the Data Studio is often left out while counting among the best tools available today. First provided as a part of Google Data Studio 360 in 2016, the company made the tool completely free for use in 2018. What Does Google Data Studio Offer? A relatively new tool in the market, Data Studio efficiently converts raw data into detailed reports while pulling information from data collection platforms and centralising it into a single dashboard. While it may seem like run-of-the-mill reporting software due to its simplicity, it offers some impressive capabilities for the users. Google Data Studio can draw information from up to 12 different data sources, allowing access to over 500 datasets and 240 connectors. Some of the data sources from where it can pull information are Search Console, BigQuery, MySQL, AdWords, Analytics, YouTube and more. Despite The Amazing Features, Why Does It Lag Behind? While Data Studio has many benefits such as building interactive dashboards, allowing customisations and creating beautiful reports, it offers limited functionality. Since it is a relatively new data visualisation tool, it has to set a benchmark. The analysis provided by Google Data Studio may not stand at par with the likes of other powerful BI tools such as Tableau and Power BI. While it can adequately visualise the data, it doesn’t allow the slice and dice, and analysis of data as the other popular tools. Another disadvantage is that while it can run effectively with the Google set of tools, its ability to combine data from multiple sources may be restricted. It works best if you are tied into the Google Analytics platform and looking to provide customised reports for Ads, YouTube, and others. Still, for use cases that require more on-premise database access or blending of data, it may not work as it should. Another disadvantage is that it is a fully web-based solution, and there is no desktop version, unlike other BI solutions. What are the advantages of Google Data Studio? While Data Studio may find likability amid the newbies, it is not of the first choices for senior data analysts. Some of the reasons are: Absence of report delivery automationIt is not a good fit if the organisation focuses and relies highly on metricsIt has limited connections Compared to other BI tools, visualisations offered by Data Studio are averageConnecting to databases is not as easy to set up as TableauNot as customisable as other BI programs What Can It Do To Run For The Popular BI Tools Race? Having been launched in 2016, many of its features and functions have remained the same with little or no updates. For instance, the users of the likes of Tableau and Power BI believe that while Google Data Studio provides beautiful and straightforward visualisations, it lacks basic features of these tools such as the ability to edit the size, format and density of data labels. It should also improve in terms of providing other functionalities such as colour by metric, size by metric, axis padding, individual axis toggles, etc. While Tableau and Power BI both offer more than 150 custom functions to apply to data, Data Studio has around 60 such functions. It also needs to implement other functions such as Parameters and Grouping, which have been executed quite well in other BI tools such as Tableau. Some of the other aspects that it needs to address are data manipulation, flexibility, interactivity and more. Top 7 Resources To Learn MLStudioGoogle Data Studio Vs Tableau: A Comparison Of Data Visualization ToolsGoogle Announces New Updates For Data Studio; Adds Google Maps For Embedded ReportsMIT Researchers Teach AI To Paint With Common SenseHow To Setup A Simple Machine Learning Demo Application On Android Wrapping Up Google DataStudio is still new, and a lot of additions and updates are on its way. With the given set of features, while Google Data Studio may not come as the first choice of data visualisation tools to opt for, it can make for an excellent addition for small to medium businesses. For a large enterprise looking for complex data visualisation with extensive storage and functions, it is the other tools that take over Google Data Studio. But for those looking to optimise, visualise, and effectively share their data without having to spend much time or money on it, Google Data Studio is the best bet.","excerpt":"Google is hands down one of the most powerful names when it comes to tech companies but talking about business intelligence and data visualisation tools, offering by Google called the Data Studio is often left out while counting among the best tools available today. First provided as a part of Google Data Studio 360 in […]","categories":["AI Trends"],"tags":["business analysis tools"],"author_name":"Srishti Deoras","publish_date":"2020-07-13T10:00:47","publication_year":"2020","word_count":753,"keywords":["Go","machine learning","AI","ML","business analysis tools","RAG","ViT","analytics","SQL","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","R","SQL","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-google-data-studio-lags-behind-other-top-bi-tools\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":50589,"title":"The Rise And Rise Of PyTorch","content":"Since the release of PyTorch in 2016, it is on a rollercoaster ride as its adoption among developers and researchers is continually increasing. Although it was released long after one of the most popular deep learning frameworks TensorFlow, over the years, PyTorch has quickly gained grounds and has overtaken advantage its competitors had due to their early release. Numerous organisations are utilising PyTorch in their business processes to innovate and eliminate various business challenges. More notably, Microsoft and Tesla have embraced PyTorch in their organisations for adding artificial intelligence capabilities. While Tesla has integrated it for autopilot in the car, Microsoft has been using it for internal developments and have also brought support on Azure. The Rise Of PyTorch As per reports, PyTorch grew 194% in the first half of 2019 alone (Jan-Jun 2018 to Jan-Jun 2019). The statistics were implemented based on the number of papers posted on arXiv.org. A total of 1,800 papers mentioned TensorFlow, while PyTorch more or less had similar mention in the papers. Which means it is on par with TensorFlow among researchers. Earlier, deemed as a research-only library, PyTorch has now become a developer-friendly framework too. Although in the StackOverflow survey 2019, TensorFlow was head and shoulders ahead of PyTorch in popularity, PyTorch was 2nd in the most loved tools, whereas TensorFlow was a distant 5th. This implies that people who are using PyTorch are willing to continue using it than the ones who are using TensorFlow. Developers who are dreaded about TensorFlow may well be moving towards PyTorch. Among many reasons, this could be the driving force behind the proliferation of PyTorch. Advantages That Led To Increasing Adoption Of PyTorch TorchScript: One of the merits of PyTorch is its usability and readability. On the other hand, developers complain about boilerplate code, thereby, making it complex for new users. Although it has its own advantages, new users and researchers who mostly do not focus on production-level code, embrace PyTorch as coding is very straightforward. Torchscript delivers flexibility in transitioning between eager mode and graph mode for taking advantages of both the processes.Distributed Training: Through asynchronous execution, it helps researchers to parallelise computation for processing a large batch of input data. Researchers often prefer a plethora of data for their projects as their sole focus is to get the most accurate results. However, in production, organisations try to find the balance between quality and usability.Tools & Libraries: The above advantages, over the years, have led to an active community of researchers. Therefore, one can get more insightful support to mitigate their problems. Consequently, aspirants of deep learning technology are adopting PyTorch. And with the latest addition of new features such as mobile, privacy, quantization, and named tensors, in PyTorch 1.3, it has further encouraged developers and researchers to develop robust deep learning products. Poised To Further Gain Momentum Among Developers Mobile Applications: While the TorchScript, distributed training, and libraries were empowering developers and researchers to streamline their workflows, with PyTorch 1.3, it has brought support to the model deployment to mobile devices. This has encouraged mobile app developers to leverage PyTorch and create production-level applications. Today, mobile applications play a very important role in every human life, thus businesses develop robust applications to offer their solutions and services. Java programming dominance was powered due to the proliferation of Android applications. Similarly, one can foretell that PyTorch demand in business will gain momentum to offer end-to-end workflows while developing ML-based IoT applications. Quantization: Another remarkable addition in PyTorch is Quantization for better performance at servers and edge through efficient use of server-side and on-device computation. This will further democratise Facebook’s ML library and make it developer-friendly. Other Tools: PyTorch was already well designed and had the best readability and easy-to-code, but with the new release, it has further enhanced the ability to write clear algorithms for eliminating the need for inline comments. Besides, the community has worked towards increasing its flexibility such that it can be utilized in the production. They have added additional tools and libraries to support model interpretability and bringing multimodal research to production. Outlook Many GitHub popular projects have implemented PyTorch for their ML-projects. Such a rise in adoption has increased its community, thereby, it has now become easier for beginners to start learning with the community support. PyTorch is no more only a research-friendly tool but has made its presence in developers and new data science aspirants. This has guided PyTorch to be on par with TensorFlow in the deep learning most popular frameworks. However, it has the potential to outpower TensorFlow in the future and become the go-to deep learning framework.","excerpt":"Since the release of PyTorch in 2016, it is on a rollercoaster ride as its adoption among developers and researchers is continually increasing. Although it was released long after one of the most popular deep learning frameworks TensorFlow, over the years, PyTorch has quickly gained grounds and has overtaken advantage its competitors had due to […]","categories":[],"tags":["Pytorch"],"author_name":"Rohit Yadav","publish_date":"2019-11-25T16:00:00","publication_year":"2019","word_count":771,"keywords":["Pytorch","data science","artificial intelligence","AI","PyTorch","R","ML","RAG","deep learning","TensorFlow","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","data science","TensorFlow","PyTorch","RAG","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-rise-and-rise-of-pytorch\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10080151,"title":"What Makes a Programming Language Popular?","content":"For a programming language conceived in the late 1980s, Python’s popularity refuses to wane. Guido van Rossum, the principal author of Python famously created Python as a hobby. The first version of the code that was published was labelled as 0.9.0. A good 25 years later, Python was still moving up in the ranking of popular programming languages. While Java and JavaScript topped the list, Python sat in third place. By 2019, the rankings shuffled and Python became the most popular language. Source: TechRepublic Even then, early releases of the code had classes with inheritance, exception handling, functions and several other data types. The language was built on a strong foundation and had all the ingredients that would make it last — it was flexible and provided several frameworks and tools, and was easy to learn. The alchemy of a popular programming language is hard to determine. Does the popularity of a language signal it is technically better than the obscure ones? Or is it because the languages have simply been marketed better? What’s behind a successful language? There is a set of popular languages that are default for developers like C, JavaScript, SQL and Java because switching to something else simply requires too much research and effort. There are languages backed by tech giants but it isn’t enough to just back them. The companies must throw their weight behind the language like Microsoft did with C# or Google’s Dart which powers Flutter or Meta’s React. Source: JavaScript in Plain English Flutter and React Native, which are both competitors, are cross-platform solutions and market leaders in their segment. Big Tech connections have helped the duo proliferate the marketplace. Meta’s social media platform Facebook, AirBnB, The New York Times and Instagram – all use React-Native to develop their mobile apps on iOS and Android. A bunch of popular languages reach developers faster when they have their own libraries. Python is a classic demonstration of this and planted a seed for the vast ecosystem of libraries that would later develop. Other languages like JVM, Scala and Groovy are included among them. Programming languages have to be familiar enough for developers to not intimidate them. This explains why the familiarity of languages with the C syntax is hard to beat among developers and why S-Expressions appear foreign. How important is marketing languages? Krishna Rastogi from MachineHack believes that while marketing outreach could definitely be a factor that drives the popularity of programming languages, developers are habitually looking for a couple of things. “When I am looking at the background of a programming language, I am looking at the company that backs it. If the company is credible, I will obviously be inclined to use it,” he said. Source: Developer Nation However, according to him, nothing beats community support and user friendliness for developers. “Obviously for Flutter, the Google Developer Group is already a huge preexisting community and helps with popularising the language. But it isn’t a standard rule that only languages backed by a Microsoft, Facebook or Google will have a sizeable community support. There are languages with solid community support built merely on word-of-mouth,” he explains. A large community normally means that developers can receive prompt help for any query. “The code repo now for most programming languages is available on GitHub. In case we find an error in the code, we can reach out to the community,” Rastogi said. The other factor that developers clamour for is easy, readable documentation of the language. The documentation of a programming language in computer programming includes its specifications and defines it. “One of the reasons for Google’s Flutter gaining popularity as a programming language is its accessible documentation,” he noted. If the documentation for a language is complicated and takes a long time to grasp, it pushes developers away. Marketing among programming languages is an all-encompassing term that doesn’t just include advertising. Good marketing in the developer community also indicates good documentation. Pro tip: As an experienced software developer, investing time in learning about business, management, marketing and sales can be way more profitable than keeping up with all the hype in programming languages or frameworks.— Vlad Mihalcea (@vlad_mihalcea) November 3, 2020 This isn’t just confined to programming languages. Between Google’s software library TensorFlow and Meta AI’s framework based on the Torch library, legible documentation proved to be the contentious factor. “TensorFlow has documentation that is comparatively more difficult which is one of the reasons PyTorch has left TensorFlow far behind in terms of gaining favour with developers,” Rastogi stated. This is not to say that languages backed by big tech companies are immediately popular because of their deep pockets and higher spend on marketing. The efficiency and ease of the language itself has to match up to the hype. “Flutter is a natural progression for developers who work on Native mobile development. It takes users around half an hour to churn out applications.” But what wins the faith of developers more often than not is this Rastogi says, “Every developer is looking for quick answers to their questions on Stack Overflow.”","excerpt":"Marketing among programming languages is an all-encompassing term that doesn’t just include advertising","categories":["AI Trends"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-11-20T10:00:00","publication_year":"2022","word_count":848,"keywords":["Go","Meta AI","PyTorch","AI","Python","SQL","JavaScript","TensorFlow","R","Java"],"extracted_tech_keywords":["AI","Meta AI","TensorFlow","PyTorch","Python","R","SQL","JavaScript","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/what-makes-a-programming-language-popular\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10052541,"title":"DeepMind’s Progress Over The Years In Robotics","content":"AI Research Lab DeepMind acquired and open-sourced MuJoCo, a rich and effective contact model. By open-sourcing Multi-Joint Dynamics with Contact (MuJoCo), DeepMind has given a major push to its robotics ambition. This article will trace how DeepMind has been making consistent efforts in pushing the envelope in robotics. Deep reinforcement learning to training robots In 2016, DeepMind researchers demonstrated how deep reinforcement learning can train real physical robots. The paper showed that deep Q-functions-based reinforcement learning algorithms can scale to complex 3D manipulation tasks and efficiently learn deep neural network policies. The authors further showed that the time to train the robots can be further reduced by algorithm parallelisation across multiple robots that asynchronously pool their policy updates. The proposed methodology can learn a variety of 3D manipulations skills in simulation and a door opening skill (often considered a complex task for robots to train on) without manually designed representations. Producing Flexible Behaviours In 2018, DeepMind published three major papers to demonstrate flexible and natural behaviours to reuse and adapt to solve tasks. The scientists trained agents with a variety of simulated bodies to perform activities like jumping, turning, and crouching across diverse terrains. The results showed that the agents develop these skills without receiving specific instructions. Credit: DeepMind Another paper demonstrated a method to train a policy network that imitates motion capture data of human behaviours to pre-learn skills like walking, getting up from the ground, turning, and running. These behaviours can then be tuned and repurposed to solve other tasks like climbing stairs and navigating through walled corridors. The third paper produced a neural network architecture based on state of the art generative models. This research showed how this architecture is capable of learning relationships between different behaviours and imitating specific actions that are shown to it. After training, the systems could encode a single observed action and create a new novel movement. Scaling data-driven robotics DeepMind demonstrated a framework for data-driven robotics which uses a large dataset of recorded robot experience before scaling it to several tasks using a learned reward function. This framework can be applied to accomplish three different object manipulation tasks on a real robot platform. The scientists used a special form of human annotations as supervision to learn a reward function and demonstrate tasks with task-agnostic recorded experience. This helps in dealing with real-world tasks where the reward signal cannot be acquired directly. The learned rewards and large dataset of experience derived from different tasks are used to learn robot policy offline using batch reinforcement learning. This approach makes it possible to train agents to perform challenging manipulation tasks like stacking rigid objects. New Benchmark for Stacking DeepMind recently introduced RGB-Stacking as the new benchmark for vision-based robotic manipulation tasks. Here the robot has to learn how to grasp different objects and balance them over each other. It was different from previous works because of the diversity of the objects used and the variety of empirical evaluations performed to verify the accuracy of the results. Credit: DeepMind The results demonstrated that complex multi-object manipulation can be learnt using a combination of simulation and real-world data. The experiment could also suggest a strong baseline for generalisation to novel objects. This experiment is considered a major advancement in DeepMind’s endeavour towards making generalisable and useful robots. The authors will now work to make robots better understand the interaction with objects of different geometries. The RGB-Stacking benchmark has been open-sourced along with the designs for building real-robot RGB-stacking environments, RGB-object models and information for 3D printing. MuJoCo MuJoCo is a physics engine simulator that facilitates research and development in fields that require fast and accurate simulations like robotics, biomechanics, graphics, animation, etc. Developed by Emo Todorov for Roboti, MuJoCo is one of the first full-featured simulators designed from scratch for model-based optimisation through contacts. Before DeepMind’s acquisition, MuJoCo was a commercial product between 2015 and 2021. MuJoCo helps in scaling up computationally intensive techniques like optimal control, system identification, physically consistent state estimation, and automated mechanism design before applying them to complex dynamic systems in contact-rich behaviours. It also has applications like testing and validating control schemes before deploying on physical robots, gaming, and interactive scientific visualisation. Wrapping up This is probably a slow phase for research and development work in robotics. DeepMind rival OpenAI, after investing many years of research, resources and efforts into robotics, finally decided to disband its robotics research team and shift focus to domains where data is more readily available. On the industry side, too, several robotics-based companies have shut shop or are undergoing major losses. Given the circumstances, robotics, despite being such a lucrative industry, has limited to no buyers. Backed by Alphabet, DeepMind’s progress has helped it hold the flag high in this field over the past few years.","excerpt":"Backed by Alphabet, DeepMind’s progress has helped it hold the flag high in robotics over the past few years.","categories":["AI Features"],"tags":["DeepMind","DeepMind AI"],"author_name":"Shraddha Goled","publish_date":"2021-10-28T17:00:00","publication_year":"2021","word_count":800,"keywords":["Go","OpenAI","AI","neural network","data-driven","programming_languages:R","programming_languages:Go","ViT","DeepMind AI","AI research","R","DeepMind"],"extracted_tech_keywords":["AI","neural network","OpenAI","R","Go","ViT","data-driven","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deepminds-progress-over-the-years-in-robotics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":1487,"title":"What does the Budget 2017 bring for Smart Cities?","content":"New Delhi: Finance Minister Arun Jaitley (C) stands outside his office at North Block holding the briefcase containing the Union budget for 2017 and is flanked by MoS Arjun Meghwal (3rd R) and Santosh Gangwar (3rd L), on Wednesday,in New Delhi. PTI Photo by Vijay Verma(PTI2_1_2017_000009B) The Indian Government’s much talked about Smart Cities initiative is hoping to get a boost, based on the upcoming Budget 2017-18 session. Key players are hoping added incentives can affect the development and infrastructure vastly. According to experts, Smart City plans require an investment of INR 1300 crores to INR 6000 crores in rebooting infrastructure from ground up – improving basic urban infrastructure, water and sanitation, improving public amenities and roads. These funds would be rolled out depending on the need for infrastructure and other requisites for a Smart city. According to a news report, INR 1,188 crore has been earmarked for six Smart Cities in Karnataka as part of the budget. The six cities are — Mangaluru, Belagavi, Tumakuru, Shivamogaa, Hubballi-Dharwad and Davanagere. According to a news report, around 60 cities are part of the Smart City Mission program and the selection was made based on the plans submitted. Some of the key areas that require Government funding are – sanitation, solid waste management, building rapid transit systems and improving clean water supply. As per news reports, the projects are billed to operate in two ways — build, own operate (BOO) and build operate transfer (BOT). There might be a visible lag in the funds provided by the government which would need to be met by private participation. Some of the budgetary provisions outlined are: Looking to boost potential private partners, the forthcoming budget provides an exempting income arising out of such investments from tax. However, exemptions would be granted only to projects that have been approved and completed within the specified time frame. This would ensure a timely implementation and an increased participation from public. Experts are also hoping for some budgetary provisions in the upcoming budget session. The ideas proposed are setting aside a corpus of guarantee fund that could be used in developing infrastructure or providing guarantees against loans issued. It could help significantly in development of pan city projects. The loans could be availed by the Smart City Special Purpose Vehicle (SPV) if certain basic eligibility criteria around finance and governance are met. Additionally, the credit enhancement facility by LIC could be extended to cover Smart City related bond issuances. Some of the key areas outlined by experts are — Experts are looking for a significant growth in the direction of Smart cities, which could ideally take anywhere between 20 to 30 years for its completion. While there have been provisions outlined for ensuring a growth in this domain, there needs to be a considerable investment in terms of manpower and technological advancements. The Internet of things, which plays the underlying role in the building up a Smart city by plugging thousands of sensors and other devices, there is a dire need to ensure that these are operated hassle free and there are no hindrances in terms of connectivity and the cyber criminals.","excerpt":"The Indian Government’s much talked about Smart Cities initiative is hoping to get a boost, based on the upcoming Budget 2017-18 session. Key players are hoping added incentives can affect the development and infrastructure vastly. According to experts, Smart City plans require an investment of INR 1300 crores to INR 6000 crores in rebooting infrastructure […]","categories":["IT Services"],"tags":["smart city india"],"author_name":"Srishti Deoras","publish_date":"2017-02-15T07:31:59","publication_year":"2017","word_count":524,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","smart city india","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","Go","API","GAN","ViT","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/what-does-budget-2017-bring-for-smart-cities\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10125928,"title":"AstraZeneca Inaugurates World&#8217;s Largest Global Capability Centre in Chennai","content":"In a significant boost to Tamil Nadu’s burgeoning reputation as India’s knowledge capital, AstraZeneca has inaugurated its new Global Innovation and Technology Centre in Chennai. The inauguration event saw participation from several dignitaries and industry leaders, marking the launch of AstraZeneca’s largest GCC worldwide. This monumental addition to Chennai’s landscape is expected to create 4,000 high-quality jobs, tapping into the rich talent pool available in Tamil Nadu. The establishment of this centre underscores AstraZeneca’s commitment to expanding its capabilities and investing in the region. The event highlighted Chennai’s growing appeal to global giants, with Hitachi Energy also setting up their largest GCCs in the city. The rapid absorption of office spaces in Chennai reinforces Tamil Nadu’s status as the Knowledge Capital of India. Under the visionary leadership of the Honourable Chief Minister of Tamil Nadu, Thiru MK Stalin, the state continues to attract and nurture global innovation and growth. The Chief Minister’s policies and the state’s conducive environment for business have been pivotal in drawing such significant investments. This expansion by AstraZeneca is seen as a powerful endorsement of Tamil Nadu’s skilled workforce and progressive policies, setting a benchmark for other global firms considering investment in the region. With AstraZeneca’s new campus now operational, Chennai is poised for a further surge in its role as a hub for technological innovation and global business services. The collaboration between the state government and leading multinational companies like AstraZeneca signifies a promising future for Tamil Nadu’s economic and technological landscape. Recent years have seen a surge of international pharmaceutical companies expanding into India, driven by its growing healthcare market and strategic position as a global innovation hub. A few companies include Pfizer, Novartis, GSK and Johnson & Johnson. These companies are increasing their presence through new investments, local partnerships, and advanced manufacturing facilities, significantly advancing India’s pharmaceutical sector.","excerpt":"The collaboration between the state government and leading multinational companies signifies a promising future for TN’s economic and technological landscape.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Chennai","GCC"],"author_name":"Tarunya S","publish_date":"2024-07-05T15:01:03","publication_year":"2024","word_count":305,"keywords":["Go","API","GCC","programming_languages:R","AI","innovation","programming_languages:Go","Chennai","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/astrazeneca-inaugurates-worlds-largest-global-capability-centre-in-chennai\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24736,"title":"Analytics India Magazine Turns Six","content":"Happy AIM day! 21 May 2012 is the day when it all started — the day when Analytics India Magazine, India’s first digital magazine that exclusively covers data science, artificial intelligence and data analytics was launched. We are happy our readers are looking up to us for guidance and information — and it has been quite a ride for all of us here! From one visionary’s dream of reporting news on (then) relatively unknown areas in technology, to a critically-acclaimed and a noted brand six years later, AIM has fulfilled many goals. We have increased in popularity, provided a launch pad for new startups and companies, celebrated and recognised the excellence of top data scientists and academicians and have chronicled the development in new tech which is changing the way businesses are functioning all over the world now. For our sixth anniversary, the AIM team, which has more than doubled in the last year itself, shares their experience of working in a dynamic workplace in a thriving, emerging field: Abhijeet Katte: I started my journey with stints in machine learning startups and labs in Bengaluru. I got interested in writing and expressing my views regarding current state of AI and policies which are affecting our startup and tech ecosystem. When I got in touch with AIM I was convinced from the first minute that this was the place for me. I could work on an important product like MachineHack while continuing what I love to do — write. And if you want to write and shine light on the AI ecosystem in India, what better place than AIM?! I feel like I have the best of both worlds. Our CEO, Bhasker Gupta said something to me on the very first day that gave me immense clarity. He said, “Your role is to write and build for the data scientists and the data science ecosystem. Every action should help the individual data scientist or the ecosystem as a whole.” This simple mission statement helps me everyday. Abhishek Sharma: My journey with AIM has been delightful so far. To begin with, everyday at AIM is exciting and challenging. Be it writing about data science, machine learning or artificial intelligence, we get to explore fascinating aspects of technology every day. Apart from this, I also get to know a lot about latest happenings in the analytics industry. The work environment here at AIM is awesome. We have an amazing boss and fun colleagues, to say the least. They have been helpful right from the beginning of my stint at AIM. This has made me better at work and come out successful. Prajakta Hebbar: Coming from a very different background in journalism, writing about new tech was a great challenge and a learning experience. I get to learn something interesting every day (my search history is a testimony to that!) and have met numerous important and awe-inspiring leaders here. Writing about such important topics has helped me connect with many industry leaders in the sector. AIM’s flagship events Cypher and MachineCon have shown me just how close-knit and helpful the entire analytics and new tech community is. Be it a CXO or a wide-eyed student, we all support and encourage each other. Working for AIM has been a fun ride. I am lucky to have a boss who doesn’t say ‘no’ to any ideas — be it memes, funky videos or unusual interviews. The cookie jars are always full of delicious treats, and the colleagues are just amazing. Richa Bhatia: It’s been a very fulfilling journey at Analytics India Magazine and I believe we have all grown with the organisation. Personally, AIM has done more than just evangelise the data science ecosystem; it has helped startups and enterprises get recognition for their work. Our conferences are well attended and are acclaimed for the great lineup of speakers who showcase advanced trends in analytics practice. For me personally, it has given more than a peek into how data science departments work; it has been an immense pleasure to write about the growth and how new products are coming out of India. The last one year has also seen a bunch of enterprises setting up their COE in India. On a more personal note, I believe in a short span of time, we have established ourselves as a formidable tech media brand and in the coming years, we promise to add more knowledge. Smita Sinha: Transitioning from a economy, business and markets news desk to New Tech was an interesting change for me. The quest to learn new things and gather more experiences has kept me on my toes. After joining AIM, I got fascinated by this world of Big data. Now a days every single organisation, starting from startups to Fortune 500, runs on Big Data. AIM gave me that opportunity to interact with industry leaders, talk to startup and get to learn more about data analytics, AI and machine learning. As they say, we are living in the age of information and if you want to succeed in this world, you must study trends. Everything depends on data today. Big data is everywhere and now I am fascinated by this world of Big data. I still remember the day I went up to Bhasker and confided that I got heebie-jeebies while writing about such deeply technical subjects. His assurance that I’d “learn it in time” has stayed with me till now. It boosted my confidence, now I am taking baby steps to learn things. Everyday I learn new things, understand new concepts. Srishti Deoras: AIM turns six today. Over the last two years of being a part of the team, the industry has evolved and changed in so many ways. Though analytics, and the then emerging field of AI were different for me when I initially started writing about it, my tech background helped me to connect to the essence of the growing field. One of the my favourite things about my job is covering interesting startups, talking to industry leaders and writing about the various aspects of new tech. Also, our trademark events like Cypher and MachineCon, which are considered magnum opus of the analytics industry, are an amazing networking experience. Our camaraderie at work is delightful. We exchange ideas, discuss emerging trends and the best part is that Bhasker always comes up with amazing insights into trends or news that we had never thought of! It has been a thrilling journey, and I wish to witness more of it in the coming years.","excerpt":"Happy AIM day! 21 May 2012 is the day when it all started — the day when Analytics India Magazine, India’s first digital magazine that exclusively covers data science, artificial intelligence and data analytics was launched. We are happy our readers are looking up to us for guidance and information — and it has been […]","categories":["IT Services"],"tags":["AI (Artificial Intelligence)"],"author_name":"AIM Media House","publish_date":"2018-05-21T05:20:31","publication_year":"2018","word_count":1091,"keywords":["data science","Go","machine learning","artificial intelligence","AI","Git","RAG","Aim","analytics","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/analytics-india-magazine-turns-6\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":24229,"title":"Cognizant Acquires Belgium-Based Analytics Company Hedera Consulting","content":"Noted consulting and professional services company Cognizant has acquired privately-held Hedera Consulting, a company that specialises in business advisory and data analytics services across a number of industry sectors. According to a statement released by the company, the purchase further expands Cognizant’s consulting, business insight and digital transformation capabilities for clients in Belgium and the Netherlands. The terms of the transaction were not disclosed. Based in Kontich, Belgium, Hedera Consulting works with leading brands across various industries. Hedera’s consultants and data scientists specialise in helping clients with growth strategy, innovation, marketing, sales and customer service. Hedera Consulting is now part of the Cognizant Consulting business unit. “The most successful companies are the ones that reduce the time from informed insight to action,” said Philip Lahey, partner at Hedera Consulting. “By joining forces with Cognizant, we are even better positioned to help clients define their strategy, transform their businesses and gain insight and competitive advantage in their fast-changing, highly competitive marketplaces. Our combined industry and local knowledge and experience, as well as our strong joint team, will better enable our customers in the Belgian and Dutch markets to extract meaning from their data and use it to effectively shape their products, services and experiences.” “In the Belgian and Dutch markets, companies are re-designing their business and IT operating models for the digital era,” said Santosh Thomas, president, Global Growth Markets at Cognizant. “Hedera Consulting expands our ability to help these European clients create agile and digitally transformed enterprises that can act and react to the oceans of data for deeper customer insight, new product development, and to innovate and exploit new business opportunities.” Founded in 2009, Hedera Consulting is a consulting company specialised in growth strategy, digitisation, innovation and commercial excellence for clients across industries, including financial services and utilities. They support their clients with advice and transformation management. In 2015, Hedera started an additional business line focusing on analytics and data excellence. Hedera has served clients across the rest of Europe, including Italy, Switzerland, the Nordics, UK, as well as the Middle East.","excerpt":"Noted consulting and professional services company Cognizant has acquired privately-held Hedera Consulting, a company that specialises in business advisory and data analytics services across a number of industry sectors. According to a statement released by the company, the purchase further expands Cognizant’s consulting, business insight and digital transformation capabilities for clients in Belgium and the […]","categories":["AI News"],"tags":["Cognizant"],"author_name":"Prajakta Hebbar","publish_date":"2018-05-03T10:39:24","publication_year":"2018","word_count":343,"keywords":["programming_languages:R","AI","innovation","digital transformation","Cognizant","Git","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Git","digital transformation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cognizant-acquires-hedera-consulting\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10068394,"title":"How to use logistic regression for image classification?&nbsp;","content":"Image Classification is a process of classifying various image categories to their appropriate labels or categories it is associated with. Image classification is mostly employed with Convolutional Neural Networks (CNNs), but this article is an attempt to showcase that even logistic regression has the capability to classify images efficiently with a reduction in computational time and also to waive off the tedious task of building complex models for image classification. Table of Contents An overview of Logistic RegressionCase Study for Image Classification with Logistic RegressionSummary An overview of Logistic Regression Logistic Regression is one of the supervised machine learning algorithms which would be majorly employed for binary class classification problems where according to the occurrence of a particular category of data the outcomes are fixed. Logistic regression operates basically through a sigmoidal function for values ranging between 0 and 1. Case Study for Image Classification with Logistic Regression As mentioned earlier as this article emphasizes using Logistic Regression for Image classification we are using the Hand Sign Digit Classification dataset with two categories of images showing Hand Signs of 0 and 1. A numpy format dataset was utilized for this article, so the input and the output dataset were loaded into the working environment appropriately as shown below and the main reason for using the numpy format data is for easy computation as numpy data processing is faster when compared to other data types. Below are the steps to be followed to load numpy data into the working environment. inp_df=np.load('\/content\/drive\/MyDrive\/Colab notebooks\/Image classificatiob using LOGREG\/inp.npy') out_df=np.load('\/content\/drive\/MyDrive\/Colab notebooks\/Image classificatiob using LOGREG\/op.npy') Once the dataset was loaded into the working environment the shape of the numpy data was determined to estimate the number of rows and columns present in the data and it was seen that there are 410 images of size (64,64) in the input data used and there are 410 images in the output data. The shape of the data can be computed as shown below. print('Input Dataframe shape',inp_df.shape) print('Output Dataframe shape',out_df.shape) The output of the shape command will be as shown below. Once the dataset was loaded into the working environment the dataset was split for the training and testing with a split ratio of 80:20 respectively using the scikit-learn model selection module as shown below. from sklearn.model_selection import train_test_split X_train,X_test,Y_train,Y_test=train_test_split(inp_df,out_df,test_size=0.2,random_state=42) It is a better practice to mention a random value for the random_state parameter while splitting the data to ensure uniform shuffles of data for training and testing. Later the split data was used to visualize the data present across the training and testing phase using subplots to validate the split among the input and the output as shown below. plt.figure(figsize=(15,5)) for i in range(1,6): plt.subplot(1,5,i) plt.imshow(X_train[i,:,:],cmap='gray') plt.title('Sign language of {}'.format(Y_train[i])) plt.axis('off') plt.tight_layout() plt.show() Here random parameters for figure size were mentioned to obtain clear visible visual and later the initial images of the data were obtained as shown below. As we are working with the image dataset and for the classification of images we are using the Logistic Regression algorithm it was necessary to reshape the dependent component of the train and test appropriately as Logistic Regression is built to work with at most two dimensions of data and moreover this being an image dataset it is necessary to reduce the dimensions of the image data which is originally in three dimensions to two dimensions as shown below to evacuate the issues with respect to dimensionality. X_train=X_train.reshape(328,64*64) X_test=X_test.reshape(82,64*64) Once the necessary data preprocessing steps were taken up, the Logistic regression model was fitted to the split data by importing the necessary scikit linear model package for Logistic Regression as shown below. from sklearn.linear_model import LogisticRegression Once the necessary module was imported into the working environment the LogisticRegression model was fitted onto the split data as shown below. logreg = LogisticRegression() logreg.fit(X_train,Y_train) Later the model was taken up for prediction for different test scenarios where the model was able to yield the right predictions. y_pred=logreg.predict(X_test) One of the image classification results from the Logistic regression model implemented is shown below where the implemented model’s ability to correctly classify the image samples can be observed. Later the accuracy score of the logistic regression model was obtained for the test data as shown below to evaluate the model’s nature of genericness and reliability when the model is tested for changing data, wherein the Logistic regression model was able to yield an overall accuracy score of 98% for the test data. The steps to obtain the accuracy score from a logistic regression model are shown in the below figure. However, relying only on the parameter of accuracy would not be right all the time as it would lead to misinterpreting results. Due to this, the various other performance metrics of the logistic regression model implemented were evaluated through a classification report where parameters such as precision, recall, and f1-score can be evaluated in order to make suitable interpretations from the models. Out of this when the harmonic mean or in simple terms the F1 score parameter also for both the classes falls in a considerable range close to 98% for ‘0’ class and 97% for ‘1’ class which is an indicator of a reliable model. For better understanding, the classification report for the logistic regression model implemented is shown below. Summary Image classification is one such application in the domain of Deep Learning and Image Processing where at certain times multi-level classification is taken up with models like Convolutional Neural Networks where the model built, might have to propagate through various layers. However, if there is a requirement for binary image classification even a simple yet effective supervised machine learning algorithm model like Logistic Regression can be implemented to obtain appropriate image classification as briefed in this article. References Link to notebookLogistic Regression Scikit Learn","excerpt":"This article is an attempt to showcase the capability of Logistic Regression as a machine learning algorithm for image classification.","categories":["AI Trends"],"tags":["cnn","Deep Learning","Image Classification","logistic regression","Machine Learning","Python"],"author_name":"Darshan M","publish_date":"2022-06-05T13:00:00","publication_year":"2022","word_count":963,"keywords":["scikit-learn","NumPy","machine learning","TPU","AI","neural network","logistic regression","Machine Learning","cnn","Python","Ray","Colab","deep learning","Deep Learning","R","Image Classification"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","Ray","scikit-learn","Colab","NumPy","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-use-logistic-regression-for-image-classification\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10101115,"title":"Meta Supercharges its Ads Manager With a GenAI Trio","content":"Today, Meta has started rolling out three new generative AI features for advertisers, allowing them to use AI to create backgrounds, expand images and generate multiple versions of ad text based on their original copy. This move shows Meta’s confidence in generative AI’s potential to assist brands and enterprises that constitute a major chunk of Meta’s revenue streams. The first among the new features allows advertisers to customise their creative assets by generating multiple different backgrounds to change the look of their product images. The technology is similar to that which Meta used to create Backdrop, which allows users to change image backgrounds using text prompts. However, in the toolkit, the backgrounds are generated for the advertiser based on their product images and will be “simple backgrounds with colours and patterns,” Meta explained. The feature is available to advertisers using the company’s Advantage+ catalogue for their sales ads. The second feature is image expansion for advertisers to adjust their assets to fit different aspect ratios like Feed or Reels, for example. The feature would let advertisers spend less time repurposing images and video, for different surfaces, the company claimed. Both of the above features are available to Advantage+ creative in Meta’s Ads Manager. The third addition is the ‘text variations’ feature capable of generating up to six different text versions based on the original copy. These renditions can highlight particular keywords and phrases aligned with the advertiser’s requirements. In the context of a particular campaign, Meta can present different text combinations to various audiences, to understand which versions yield better responses. However, the company won’t disclose detailed performance metrics for each variation, as conveyed by the technology behemoth. Meta has already tested these features in its AI Sandbox with a diverse small set of advertisers and early results indicate that generative AI will save them over five hours on a weekly basis. Nonetheless, the company admits that there remains a significant amount of work to be done to refine output aligning precisely with the advertisers’ styles. The company also revealed that there are more AI features to come.","excerpt":"The genAI features will let advertisers create backgrounds, expand images and generate versions of ad copies","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-10-05T02:18:19","publication_year":"2023","word_count":348,"keywords":["TPU","programming_languages:R","AI","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Aim","TPU","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-supercharges-its-ads-manager-with-a-genai-trio\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10103622,"title":"IBM Opens New Client Innovation Centre in Gandhinagar","content":"IBM today announced the opening of its new IBM Consulting Client Innovation Centre (CIC) in Gandhinagar, India. This is a continuation of the ongoing expansion by IBM Consulting into non-metro and emerging cities across the country. In addition to gaining access and offering expanded opportunities to a broader talent pool, the Center will also help further fast-track the digital transformation and economic growth in the region. IBM Consulting plans to leverage the Centre to focus on key technology areas, including generative artificial intelligence (AI), hybrid cloud, and cybersecurity.  It will also leverage the security engineering talent pool to build cybersecurity platforms and accelerators for automating threat management, improving regulatory compliance, and proactively preparing clients against various attack planes. “The expansion of our CIC network to Gandhinagar will scale our asset-led delivery of IT services and enhance our value proposition to our clients and partners across the world. This will help us address the growing client demand for productivity powered by generative AI and cybersecurity, which is fundamentally changing how businesses operate,” John Granger, senior vice president, IBM Consulting said. The expanded presence in Gandhinagar will also create opportunities for existing employees as well as enable IBM to harness the potential talent including graduate hires from the educational ecosystem in and around the city. IBM Consulting will now operate from twelve CIC locations in India, including Bhubaneshwar, Bengaluru, Chennai, Hyderabad, Kolkata, Mumbai, National Capital Region, Pune, Mysuru, Kochi and Coimbatore.","excerpt":"This is a continuation of the ongoing expansion by IBM Consulting into non-metro and emerging cities in India.","categories":["AI News"],"tags":["IBM"],"author_name":"Pritam Bordoloi","publish_date":"2023-11-24T12:28:26","publication_year":"2023","word_count":239,"keywords":["Go","API","artificial intelligence","AI","Git","RAG","ViT","generative AI","IBM","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","generative AI","RAG","R","Go","Git","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-opens-new-client-innovation-center-in-gandhinagar\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10118863,"title":"T-Hub Supported MATH is Launching AI Career Finder to Create AI Jobs","content":"Over the years, T-Hub has grown to become the largest innovation hub in the world. Since its inception in 2015, T-Hub, led by the Telangana government, has nurtured over 2000 startups. Just last month, the Machine Learning and Artificial Intelligence Technology Hub (MATH) was established at T-Hub which aims to foster AI innovation by bridging the gap between startups, corporates, academia, investors, and governments. In the midst of concerns about job displacement caused by AI, MATH aims to generate employment opportunities in the field of AI within the country. The initial objective is to create 500 AI-related jobs by the end of 2025. MATH CEO Rahul Paith believes AI will likely do redundant jobs that involve repetitive and routine tasks, such as data entry, administrative work, and basic customer service roles. “However, it’s essential to recognize that the rise of AI will also create a demand for new jobs, particularly those requiring human skills such as creativity, critical thinking, problem-solving, and emotional intelligence. “Roles that involve working alongside AI systems, such as AI trainers, data scientists, machine learning engineers, and AI ethicists, will become increasingly prevalent,” he told AIM. Creating 500 AI jobs One of the primary goals of MATH besides nurturing AI startups is to create AI and AI-related jobs in the country. “Our vision is to generate over 500 AI-related jobs by 2025 and support more than 150 startups annually,” he said. To enable this, MATH aims to foster a supportive environment for AI-first startups, providing them with resources, mentorship, and access to networks crucial for growth. “In the first year, we are aiming to onboard over a hundred startups. These startups are AI-first, deeply involved in either building around AI, utilising AI, or contributing to the AI ecosystem,” he said. Besides undertaking initiatives like talent development programmes, industry-academia collaborations, and targeted investment in AI research and development, MATH is launching its own job portal. Called AI Career Finder, the platform is dedicated to nurturing and empowering the next generation of AI\/ML talent. Paith said the platform is designed to serve as a central hub for connecting startups seeking top AI\/ML professionals with candidates searching for exciting opportunities in the field. By leveraging AI Career Finder, MATH aims to streamline talent acquisition and placement, thereby strengthening its efforts to catalyse job creation in the AI sector. “Additionally, MATH has identified key sectors such as healthcare and clean tech as prime areas for AI integration and growth. By facilitating collaborations and investments in these sectors, MATH aims to amplify job opportunities within AI-related fields.” AI Programmes To foster AI innovation, MATH has also launched a few programmes designed to foster AI innovation in the startup ecosystem. “MATH Nuage is our pioneering initiative, through which we provide comprehensive support and guidance to aspiring entrepreneurs navigating the complex landscape of AI innovation,” Paith said. Its key components include Virtual Partner Support, which connects startups with strategic partners for insights and resources. “Similarly, Mentor Desk Support offers guidance from seasoned professionals, Funding Desk Support facilitates securing investment, and Access to Data Lake enables startups to access a vast collection of data for AI model development,” Paith explained. Another flagship programme launched by MATH is called the AI Scaleup Programme, which aims to accelerate AI innovation and entrepreneurship. “This initiative is geared towards supporting startups at the scale-up stage, providing them with the resources, mentorship, and networking opportunities needed to propel their growth and success in the AI market.” A mini data centre MATH has also set up a mini data centre with GPU capabilities to help AI startups with AI training and inferencing. “In comparison to constructing a complete data centre, the mini data centre (called MINI DC) offers powerful computing abilities at a much lower price.” The data centre helps startups meet their high-performance computing (HPC) needs and is loaded with NVIDIA A100 GPUs. “The mini data centre’s infrastructure ensures efficient deployment of trained models, enabling startups to bring their AI solutions to market quickly,” Paith said. Closing the funding gap Along with T-Hub, MATH also assists startups in securing funding through various channels, including venture capital firms, angel investors, and government grants. “We provide support in preparing funding proposals, pitching to investors, and negotiating investment terms,” Paith said. However, he believes investors, incubators, and government agencies must collaborate to close the funding gap and foster a risk-tolerant climate. “Investors must acknowledge the extended value proposition of deeptech startups and their capability to make a significant social and economic difference. “Moreover, investing in specialised education and training programmes is essential to develop a strong deeptech talent pool. “Through creating a joint ecosystem, we can enable Indian deeptech startups to flourish and emerge as global pioneers in innovation,” he said.","excerpt":"MATH will create 500 AI-related jobs by 2025.","categories":["AI Hirings"],"tags":["AI Jobs","AI jobs in India"],"author_name":"Pritam Bordoloi","publish_date":"2024-04-23T12:56:40","publication_year":"2024","word_count":788,"keywords":["Go","AI jobs in India","API","artificial intelligence","machine learning","AI","ML","AI Jobs","RAG","Aim","R","data lake"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","R","Go","API","data lake"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/t-hub-incubated-math-is-launching-ai-career-finder-to-create-ai-jobs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129880,"title":"How Plotch.ai is Building Generative AI Infrastructure for ONDC","content":"Generative AI is set to significantly transform India’s digital public infrastructure. Google-backed Plotch.ai is currently working with ONDC to build AI infrastructure and simplify e-commerce for consumers. Recently, the company developed an AI-powered conversational commerce app featuring multilingual, voice-enabled semantic search and robust image search capabilities. ONDC is an open protocol network that aims to democratise digital commerce by creating an interoperable, open platform where buyers and sellers can transact regardless of the apps they use. It’s a platform to democratise e-commerce in India. According to recent reports, ONDC has captured approximately 3% of the food-order volumes from Swiggy and Zomato and achieved 68 million transactions since its inception. It is expected to reach 100 million transactions by Diwali 2024. “Multilingual voice-based conversational commerce is one piece of the AI that we’re building,” said Manoj Gupta, the founder of Plotch.ai, in an exclusive interview with AIM. “For instance, if you’re looking to buy a saree, jewellery, or a T-shirt and want to find the most affordable options, you can simply ask the AI, and it will sort them for you,” he explained. Manoj and Monica Gupta founded Plotch.AI in 2020. They are also the founders of Craftsvilla.com, an online marketplace for ethnic products in India. They are also working on improving the product catalogue using generative AI. “We will enhance the product names, descriptions, and product metadata using generative AI,” he said. Moreover, the company is planning to introduce image generation capabilities. Following this, they aim to develop a tool that will categorise customers based on their order value. “For example, distinguishing between high-repeat customers and fraudulent ones,” explained Gupta. Additionally, they are building AI-based recommendation engines to show customers the most relevant products. Another area where the company plans to integrate AI is in optimising RTO (return to origin). Google Loves India Gupta told AIM that the startup is backed by Google and is working very closely with them.  “We are using Google’s AI tech stack, which includes their large language model, Gemini. We have utilised this foundational model and custom-trained it specifically for Indian e-commerce,” he said. “We have been doing the custom training and have also incorporated vector search using vector databases. Additionally, we have implemented semantic caching. This broadly outlines our technology stack, which also includes Python for coding, MySQL for the database, and Nginx for the web server,” he explained. Gupta said that the company is raising about $5 million. The first round was led by Antler, and the second was led by Venture Catalyst. “We are very excited about AI and plan to invest 50% of our efforts into it. Even with our $5 million raise, half of that will go into building AI infrastructure directly on top of the ONDC framework,” said Gupta Last year, Google partnered with ONDC and announced plans to integrate generative AI capabilities into their tech stack. “Our collaboration creates an opportunity for organisations India-wide to reach larger audiences and grow their businesses, ultimately transforming digital commerce adoption in the country,” said Thomas Kurian, the CEO of Google Cloud. Google recently invested in Moving Tech, the parent company of Namma Yatri, and is planning to build a ‘GPay for travel’  in India. Interestingly, Namma Yatri has also joined the ONDC network. It is one of the first mobility services to be integrated into ONDC, which was initially focused on other e-commerce segments such as groceries and food delivery. “ONDC started two years ago. We can always look for a booster shot for ONDC, some trigger to create a hockey stick growth curve. It is still in infancy; we should let it grow,” said Pramod Varma, former chief architect of Aadhaar, in an interaction with AIM, adding that today ONDC has 10 million transactions, and by 2030, it could hit the 100-million mark. Comparing UPI with ONDC, he said that UPI was narrower in purpose. “All it had to do was move money and it got extra support from the readiness of the ecosystem, such as PhonePe and Google Pay.” He explained that ONDC has a much broader purpose as it includes taxi booking like Namma Yatri, metro ticketing, grocery commerce, and food delivery. He added that the supply chain of ONDC is much more complex than that of banks. “Much of our economic value chain is fragmented; someone has to bring it together. ONDC is doing that. Over time, we will see more and more transactions happening,” he said. What Does Plotch.ai Offer? Plotch.ai has a suite of products designed to facilitate seamless integration of customers into the ONDC ecosystem. “We essentially act as the gateway for customers to connect to ONDC,” said Gupta. “ONDC is a collection of domains, including retail, logistics, and fintech. We initially started with the retail domain,” he added. Gupta said that some of the prominent customers of Plotch.ai are Meesho, IDFC Bank, Paytm, and Craftsvilla. He also said that one of the major challenges they face is connecting their customers’ systems to the network, as each system has its own CRM or ERP. “Ensuring that each party can talk to another party seamlessly is also a challenge,” Gupta said, adding that this issue has reduced significantly over the past year. One of the products the company offers is NodeApp, a full-stack application that ensures smooth operation of both buyer and seller apps within the ONDC network. On the other hand, NodePay simplifies node-to-node network payments, guaranteeing efficient financial transactions between buyers and sellers. NodeDesk provides an ONDC-enabled CRM solution for managing customer grievances and ticketing. NodeBox integrates voice AI into ONDC, allowing users in smaller cities and villages to buy and sell using voice commands in multiple languages. “India is going to be a use case capital of AI. We’ll be very big users of AI, and we believe that AI can significantly help in the expansion of the ONDC Network,” concluded Gupta.","excerpt":"ONDC has captured approximately 3% of the food-order volumes from Swiggy and Zomato and achieved 68 million transactions since its inception.","categories":["AI Features"],"tags":["AI Infrastructure","Google","ondc"],"author_name":"Siddharth Jindal","publish_date":"2024-07-23T11:34:11","publication_year":"2024","word_count":980,"keywords":["semantic search","AI","ondc","ML","RAG","vector databases","Python","Aim","generative AI","Google","SQL","AI Infrastructure","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","RAG","vector databases","semantic search","Python","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-plotch-ai-is-building-generative-ai-infrastructure-for-ondc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10060546,"title":"DeepMind’s “red teaming” language models with language models: What is it?","content":"Language models and innovations in improving them are one of the most exciting and talked about research areas right now. However, though we have seen several large language models in the last year from tech giants (DeepMind’s 280 billion parameter transformer language model, Gopher, Google’s Generalist Language Model, LG AI Research’s Language model Exaone), they cannot often be deployed as they can be harmful to users in different ways difficult to predict prior. To take a progressive step towards solving this issue, innovation mammoth DeepMind has come out with a way to automatically find inputs that elicit harmful text from language models by generating inputs using language models themselves. Language models (LMs) can generate harmful text. New research shows that generating test cases (\"red teaming\") using another LM can help find and fix undesirable behaviour before impacting users. Read more: https:\/\/t.co\/UJqeeFJrZK 1\/ pic.twitter.com\/luB1ukCFoY— Google DeepMind (@GoogleDeepMind) February 7, 2022 The researchers generated test cases (red teaming) using a language model and then used a classifier to detect various harmful behaviours on test cases. As per DeepMind, the team evaluated the language model’s replies to generated test questions by using a classifier trained to detect offensive content. What came out of this was a vast quantity of offensive replies in a 280B parameter language model chatbot. What is this model exactly? As per the paper titled, “Red Teaming Language Models with Language Models“, though LLMs such as GPT-3 and Gopher can generate high-quality, there are several hurdles in their deployment. It added, “Generative language models come with a risk of generating very harmful text, and even a small risk of harm is unacceptable in real-world applications.” The team added that they use the approach to train the 280B parameter Dialogue-Prompted Gopher chatbot for offensive, generated content. They work on several methods such as zero-shot generation, few-shot generation, supervised learning, and reinforcement learning to generate test questions with the large language models. Image: DeepMind As per the paper, red teaming gave versatile responses, with some methods proving effective in producing diverse test cases while some were effective at generating difficult test cases. The generated test cases compared favourably to manually-written test cases from Xu et al. (2021b) in terms of diversity and difficulty. The team also used LM-based red teaming to see harmful chatbot behaviours that leak memorised training data. The researchers also generated targeted tests for a particular behaviour by sampling from a language model conditioned on a “prompt” or text prefix. It said, “We also use prompt-based red teaming to automatically discover groups of people that the chatbot discusses in more offensive ways than others, on average across many inputs.” Observations After the failure cases were detected, the team added that the harmful behaviour could be fixed by blacklisting certain phrases that frequently came up in harmful outputs or finding offensive training data quoted by the model that removes data when training future iterations of the model. The model can also be trained to minimise the likelihood of its original, harmful output for a given test input. Prior work in this area There has been previous work to detect issues such as hate speech, indecent language, etc. HateCheck is a suite of functional tests for hate speech detection models. The research team built 29 model functionalities driven by a review of previous research and interviews with civil society stakeholders. They brought out test cases for each functionality and validated their quality through a structured annotation process. RealToxicityPrompts is a dataset of 100K naturally occurring, sentence-level prompts derived from a large volume of English web text, teamed with toxicity scores from a widely-used toxicity classifier. The team assessed “controllable generation methods” and found out that though data or compute based methods are more effective at moving away from toxicity, there is no current method that is “failsafe against neural toxic degeneration.”","excerpt":"DeepMind has come out with a way to automatically find inputs that elicit harmful text from language models by generating inputs using language models themselves.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Deep Learning","DeepMind","Language Models","Machine Learning"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-02-14T14:00:00","publication_year":"2022","word_count":640,"keywords":["Go","TPU","programming_languages:R","AI","innovation","Machine Learning","RAG","GPT","llm_models:GPT","Language Models","Deep Learning","AI research","R","AI (Artificial Intelligence)","DeepMind"],"extracted_tech_keywords":["AI","RAG","TPU","R","Go","GPT","innovation","AI research","llm_models:GPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deepminds-red-teaming-language-models-with-language-models-what-is-it\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10006101,"title":"How To Avoid Overfitting In Neural Networks","content":"Deep neural networks deal with a multitude of parameters for training and testing. With the increase in the number of parameters, neural networks have the freedom to fit multiple types of datasets which is what makes them so powerful. But, sometimes this power is what makes the neural network weak. The networks often lose control over the learning process and the model tries to memorize each of the data points causing it to perform well on training data but poorly on the test dataset. This is called overfitting. Overfitting occurs when the model tries to make predictions on data that is very noisy. A model that is overfitted is inaccurate because the trend does not reflect the reality present in the data. To overcome this, there are a few techniques that can be used. In this article, we will: Learn the different techniques to avoid overfitting of the model Implement these techniques to a deep learning model Methods to Avoid Overfitting of a Model You can identify that your model is not right when it works well on training data but does not perform well on unseen and new data. You can also track the performance of the model performance through concepts like bias and variance. But how to solve this problem? Here are some of the techniques you can use to effectively overcome the overfitting problem in your neural network. Data Augmentation: Diversity of data and a larger dataset is the easiest way to avoid overfitting of the model. Data augmentation allows you to increase the size of your dataset by performing processes like flipping, cropping, rotation, scaling and translation on the existing images. Data augmentation not only increases the dataset size but also exposes the model to different angles and lighting and reduces the bias in the dataset, thus avoiding chances of overfitting. 2. Regularization Techniques: This method involves adding an extra element to the loss function. This extra element acts as a critic which punishes the model for using higher weights than needed. As the complexity of the model increases, a penalty is added in the loss function that helps in limiting the flexibility of the model. The two popular methods of regularization are the L1 and L2 regularization methods.  L1 regularization reduces the weight values of less important features to zero and eliminates them from further calculations. L2 regularization aims to minimize the magnitude of weights by squaring the weight. The disadvantage is that if there are large numbers of outliers, the square increases the magnitude of the outliers as well and the model tends to not perform as well as it would with L1 regularization. With an increase in penalty value, the cost function performs weight tweaking and reduces the increase and therefore reduces the loss and overfitting. 3. Dropouts: Regularization techniques prevent the model from overfitting by modifying the cost function. Dropout, on the other hand, prevents overfitting by modifying the network itself. It works as follows. Every neuron apart from the ones in the output layer is assigned a probability p of being temporarily ignored from calculations. p is also called dropout rate and is usually initialized to 0.5. Then, as each iteration progresses, the neurons in each layer with the highest probability get dropped. This results in creating a smaller network with each epoch. Since in each iteration, a random input value can be eliminated, the network tries to balance the risk and not to favour any of the features and reduces bias and noise. 4. Early Stopping: Early stopping is a technique that can avoid over-training and hence overfitting of the model. An over-trained model has a tendency to memorize all the training data points. With early stopping, a large arbitrary number of training epochs is specified. The model is stopped from training further when the model performance stops improving on the validation dataset. As you can see in the above figure, after some iterations, test error has started to increase while the training error is still decreasing. So the training is stopped early to prevent the model from overfitting. Implementation of Techniques to Avoid Overfitting Let us go ahead and implement all the above techniques to a neural network model. For a better understanding, we will choose a small dataset like MNIST. With the MNIST dataset, it is very easy to overfit the model. Using the above techniques we will try and avoid it. We will import the required libraries and load our dataset. import numpy as np from matplotlib import pyplot as plt from keras.models import Sequential from keras.layers import Dense, Dropout, Activation, Flatten, Add, BatchNormalization, Conv2D, MaxPooling2D, Convolution2D from keras.utils import np_utils from keras.optimizers import Adam from keras.callbacks import LearningRateScheduler, ModelCheckpoint from keras.preprocessing.image import ImageDataGenerator from keras.regularizers import l2 from keras.datasets import mnist %matplotlib inline (X_train, y_train),(X_test, y_test) = mnist.load_data() plt.imshow(X_train[4]) Before adding data augmentation, we will pre-process the data by reshaping the inputs, normalizing them and converting the targets into categorical values. X_train = X_train.reshape(X_train.shape[0], 28, 28,1) X_test = X_test.reshape(X_test.shape[0], 28, 28,1) X_train = X_train.astype('float32') X_test = X_test.astype('float32') X_train \/= 255 X_test \/= 255 Y_train = np_utils.to_categorical(y_train, 10) Y_test = np_utils.to_categorical(y_test, 10) Data Augmentation I have made use of the built-in method to augment the dataset. But you can choose to augment it using other methods like albumentation library as well. augment = ImageDataGenerator(featurewise_center=True, rotation_range=50, width_shift_range=0.01, height_shift_range=0.01, horizontal_flip=False, vertical_flip=False, featurewise_std_normalization=True) augment.fit(X_train) Here, I have added rotation, flipping, shift range and feature wise standard normalization techniques to produce the data augmentations. aug = augment.flow(X_train[1:7], batch_size=1) for i in range(1, 6): plt.subplot(1,5,i) plt.axis(\"off\") plt.imshow(aug) plt.plot() plt.show() Dropout and Regularization model=Sequential() model.add(Conv2D(16, (3, 3), input_shape=(28,28,1), kernel_regularizer=l2(0.01))) model.add(BatchNormalization()) model.add(Dropout(0.5)) model.add(Activation('relu')) model.add(Conv2D(16, (3, 3), kernel_regularizer=l2(0.01))) model.add(BatchNormalization()) model.add(Dropout(0.5)) model.add(Activation('relu')) model.add(Conv2D(10, (1, 1), kernel_regularizer=l2(0.01))) model.add(BatchNormalization()) model.add(Dropout(0.5)) model.add(Activation('relu')) model.add(MaxPooling2D(pool_size=(2, 2))) model.add(Conv2D(8, (3, 3), kernel_regularizer=l2(0.01))) model.add(BatchNormalization()) model.add(Dropout(0.5)) model.add(Activation('relu')) model.add(Conv2D(10, (4, 4), kernel_regularizer=l2(0.01))) model.add(Flatten()) model.add(Activation('softmax')) Once this is done we will build our model. In the model built here, I have included both Dropout and the regularization technique. For purposes of this implementation, I have used l2 regularization since it causes faster convergence. As you can see above Keras provides a method for both these techniques to be implemented. The kernel_regularizer is passed with l2 regularization. The value of 0.01 passed is the penalty for the loss function every time the model tries to assign higher weights when it is not required. Similarly, Dropout is added with the p-value of 0.5 because it is the default value. Early Stopping The last technique is to add early stopping to the model so that it stops the model from overtraining. from tensorflow.keras.callbacks import EarlyStopping early_stopping = EarlyStopping() Early stopping is passed as a callback when the model is fitted with the training and validation set. If no parameters are passed, the default values are taken. You can customize the early stopping parameters as well. Now, it is time to compile and fit the model. model.compile(loss='categorical_crossentropy',optimizer='adam', metrics=['accuracy']) batch_size = 32 history =model.fit_generator(augment.flow(X_train,Y_train, batch_size=32), steps_per_epoch = X_train.shape[0] \/batch_size, epochs=40, verbose=1, validation_data=(X_test, Y_test), callbacks=[early_stopping]) I have given the number of epochs as 40. Since these models can be trained at just 10 epochs, this is a higher number and the model will be stopped early. As you can see the model has automatically stopped training after 4 epochs because the validation accuracy started to decrease but the training accuracy increased. This means that the model was prevented from overfitting just after 4 epochs. We can plot the graph of this to understand better. def graph_plot(history, metric): train = history.history[metric] validation = history.history['val_'+metric] epochs = range(1, len(train) + 1) plt.plot(epochs, train) plt.plot(epochs, validation) plt.title('Training and validation '+ metric) plt.xlabel(\"Epochs\") plt.ylabel(metric) plt.show() graph_plot(history, 'accuracy') This graph shows that when the val_accuracy decreased steeply, the early stopping stopped the model from training further. Conclusion Recognizing and eliminating overfitting from your neural network is key for any machine learning engineer. The goal of this article was to give brief insights about the methods that are available for eliminating overfitting so that the models you build can be useful in real-time and is helpful to the AI community.","excerpt":"In this article, we will: Learn the different techniques to avoid overfitting of the model and Implement these techniques to a deep learning model","categories":["Deep Tech"],"tags":["Convolution Neural Network","dropout","Neural Networks","overfitting","regularization techniques"],"author_name":"Bhoomika Madhukar","publish_date":"2020-09-06T16:00:05","publication_year":"2020","word_count":1341,"keywords":["NumPy","machine learning","Keras","TPU","AI","neural network","regularization techniques","Convolution Neural Network","deep learning","Aim","Matplotlib","dropout","TensorFlow","overfitting","Neural Networks"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","Aim","TensorFlow","Keras","NumPy","Matplotlib","TPU"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-avoid-overfitting-in-neural-networks\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10095075,"title":"Meta AI Chief Yann LeCun’s Self-Supervised Idea Finally Sees Open Source Light","content":"Meta has been one of the biggest proponents of self-supervised learning when it comes to AI. Today, Meta AI has announced I-JEPA, a self-supervised computer vision model that learns the world by predicting it, based on Yann LeCun’s vision of autonomous machine intelligence to learn and reason similar to how humans and animals do. Click here to check it out. The paper is also being presented at CVPR2023 next week. The model’s training code and checkpoints are open-sourced under a non-commercial licence. I-JEPA (Image Joint Embedding Predictive Architecture) learns by creating an internal model representing the outside world and compares abstract representation of images, instead of comparing the pixels. According to the paper, this model delivers strong performance on various computer vision tasks, and is highly efficient than other similar models. I-JEPA offers versatile applicability without requiring extensive fine-tuning. Meta AI successfully trained a visual transformer model with 632M parameters utilising 16 A100 GPUs within a span of 72 hours. This model attains state-of-the-art results for low-shot classification on ImageNet, with a mere 12 labelled examples per class. In comparison, alternative approaches often consume two to 10 times the GPU-hours and yield inferior error rates when trained with an equivalent dataset size. By learning from representations instead of pixels, the model is able to avoid biases and issues that occur due to invariance-based pre-training. This also enables the model to learn directly from the images, instead of representation, which Meta AI says is a problem with the current LLM models. I-JEPA: Efficient method for Self-Supervised Learning of image features.No need for data augmentation, just masking.Joint embedding predictive architecture, not generative.And it's open source, of course.Blog: https:\/\/t.co\/knAH1EzJCJPaper: https:\/\/t.co\/fBG8xxEvb1Code &… https:\/\/t.co\/jWcYyOoF17— Yann LeCun (@ylecun) June 13, 2023 The Theory Last year, Yann LeCun, Meta’s Chief AI Scientist, introduced an innovative architecture designed to address the significant constraints faced by contemporary AI systems, called the world model. LeCun envisions the development of machines capable of rapidly acquiring internal models of the world’s dynamics, enabling them to efficiently learn, strategise for complex tasks, and seamlessly adapt to novel circumstances. This work by Meta AI is highly based on the hypothesis that common sense information is the key for enabling intelligent behaviour. This knowledge is achieved by passively observing the world which is stored on the background of the mind. Meta believes that self-supervised learning is the path towards human-like intelligence. For this to work, the system needs to acquire these representations through self-supervised learning, which entails learning directly from unlabeled data like images or sounds, instead of relying on manually curated labelled datasets. Meta AI demonstrates the potential of I-JEPA, showcasing the ability to learn competitive off-the-shelf image representations without relying on additional knowledge encoded through manually designed image transformations. Advancing JEPAs further to acquire broader world-models from richer modalities would be particularly intriguing. This advancement could enable making long-range spatial and temporal predictions about future events in videos based on a short context, while conditioning these predictions on audio or textual prompts.","excerpt":"Meta AI successfully trained a visual transformer model with 632M parameters utilising 16 A100 GPUs within a span of 72 hours","categories":["AI News"],"tags":["Mark Zuckerberg","Meta AI","Open Source AI","Yann LeCun"],"author_name":"Mohit Pandey","publish_date":"2023-06-13T22:52:22","publication_year":"2023","word_count":499,"keywords":["API","Meta AI","Yann LeCun","data augmentation","self-supervised learning","AI","Modal","A100","ML","computer vision","Mark Zuckerberg","Open Source AI","R"],"extracted_tech_keywords":["AI","ML","computer vision","Meta AI","R","API","data augmentation","self-supervised learning","Modal","A100"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-ai-chief-yann-lecuns-self-supervised-idea-finally-sees-open-source-light\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168775,"title":"UPS Reportedly Discusses Deploying Humanoid Robots with Figure AI","content":"United Parcel Service (UPS) is reportedly in talks with the California-based robotics startup Figure AI to deploy humanoid robots across parts of its logistics network, according to a report by Bloomberg. The talks began last year and have continued into recent months, with the scope of work for the robots yet to be finalised. UPS and Figure have not publicly confirmed specific roles. In February, Figure posted a video showing its robot picking and sorting parcels beside a conveyor belt, suggesting potential logistics tasks. Our first customer use case took 12 months, our second customer use case took just 30 daysHelix learned high-rate logistics with a single neural networkOn Sunday, we successfully validated this on-site at the customer pic.twitter.com\/ev5OeSEhly— Figure (@Figure_robot) February 26, 2025 The move comes as UPS seeks new ways to expand its automation efforts, while Figure AI aims to position itself among the leading developers of humanoid robotics. Startups and major technology firms are increasingly advancing the development of humanoid robots through breakthroughs in AI. Figure AI has drawn significant attention in this area and released Helix, a Vision-Language-Action (VLA) model that allows humanoid robots to perform complex tasks using natural language. A UPS spokesperson told Bloomberg that the company would not comment on “specific or potential vendor partners,” but added, “We regularly explore and deploy a wide range of technologies, including robotics.” UPS has already integrated automation technologies into its operations, including robotic arms and AI-powered sorting software at its Velocity facilities. It has also previously partnered with Dexterity, a company producing robots with fine motor skills for warehouse tasks. Additionally, it has explored robotic unloading and autonomous guided vehicles (AGVs) through partnerships with Pickle Robot, Dane Technologies, and other companies. Meanwhile, Figure recently introduced learned natural walking for its Figure 02 humanoid robot, making it capable of human-like movements. This development aimed to enhance the robot’s adaptability for both industrial and domestic applications by mimicking natural human locomotion, thereby compressing years of simulated training into hours.","excerpt":"Figure recently posted a video of its robot picking and sorting parcels (not confirmed to be for UPS).","categories":["AI News"],"tags":["Figure AI","Humanoid Robots","Logistics"],"author_name":"Sanjana Gupta","publish_date":"2025-04-29T11:06:08","publication_year":"2025","word_count":331,"keywords":["Humanoid Robots","ELT","programming_languages:R","AI","neural network","automation","Aim","Logistics","ai_applications:robotics","Figure AI","GAN","R","startup"],"extracted_tech_keywords":["AI","neural network","Aim","R","ELT","GAN","automation","startup","programming_languages:R","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ups-reportedly-discusses-deploying-humanoid-robots-with-figure-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10137536,"title":"Bhaichung Bhutia Says It&#8217;s Time for Indian Football to Kick Things Up With AI","content":"At Cypher 2024, former Indian football captain Bhaichung Bhutia advocated merging football and technology. “In Indian football, we are yet to use the technology [AI and analytics] that is used in a much bigger way in world football,” he said. Bhutia Lauds Liverpool “I was just reading, not sure if it’s true, that Liverpool signed Mohamed Salah because of data,” recalled Bhutia, pondering over the existence of technology in football and how it’s now a necessity in Indian football. Ian Graham, former Liverpool head of research, noticed Salah’s performance at Roma, particularly his ability to create goal-scoring opportunities and his exceptional off-the-ball movement. He dived deep into his numbers, highlighting qualities that were overlooked by traditional scouting. Liverpool has been using data and ML models to increase the team’s efficiency. These models predict player performance, emphasising on recruiting young, undervalued talents like Mohamed Salah. The methods focused on future potential, using metrics such as ‘expected goals’ and ‘possession value models’ to evaluate players’ contributions. These data insights influenced Liverpool’s recruitment and match strategies, helping the club become dominant. In the English Premier League (EPL), where millions of dollars are spent on signing an individual player, there’s a lot of pressure on getting the selection process right. Hence, clubs like Liverpool, Manchester City, Chelsea and Arsenal rely heavily on data analytics for recruitment, tactical planning and injury prevention by using models and tools like expected goals (xG), player tracking, and heat maps to evaluate talent and optimise match tactics. By analysing data on ball possession, player movement and shot quality, clubs make smarter recruitment decisions and fine-tune in-game strategies, creating a significant competitive edge. Be Like Pep ‘Data-Driven’ Guardiola Bhutia further highlighted the importance of technology in football, noting that when Pep Guardiola joined Manchester City, “the first thing he wanted to check was the kind of technology support he’s going to get”. Manchester signed Kevin De Bruyne after data highlighted his playmaking ability. Chelsea’s acquisition of N’Golo Kante was similarly data informed, revealing his exceptional ability to break up play. Arsenal also implemented data analytics in its recruitment process, with one notable example being the signing of Nicolas Pepe. The club used data analytics to identify his high xG contributions and his ability to break defensive lines during his time at Lille. The football scouting teams relied heavily on statistical models and tracking data to assess players’ suitability for their system, making the signing a significant example of data-informed recruitment. Bhutia said that major football nations are now increasingly relying on technology and data for player signings and performance management, making it a vital part of the sport’s evolution. He said that “major football-playing countries in the world are relying heavily on technology and data to sign and get the performance of players right” and urged Indian football clubs and managers to adopt the same to take the game to the next level.","excerpt":"Bhutia praises Liverpool and Manchester City for leveraging analytics and AI in player signings and game analysis.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)"],"author_name":"Tanisha Bhattacharjee","publish_date":"2024-10-04T15:49:03","publication_year":"2024","word_count":483,"keywords":["Go","programming_languages:R","AI","data-driven","ML","programming_languages:Go","ViT","analytics","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","analytics","R","Go","ViT","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/bhaichung-bhutia-says-its-time-for-indian-football-to-kick-things-up-with-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10139014,"title":"AI Chatbots’ Thirst May Kill Our Water Resources","content":"Think about it—you may need half a litre of water to cook two packets of Maggi instant noodles. Now, what if we told you that a single chat online with AI tools like ChatGPT consumes the same amount of water? It may seem small, but when millions of people use chatbots on a daily basis, it becomes a lot, and it increases the combined water footprint. According to a study on ChatGPT’s water consumption by A. Shaji George, an expert in Information and Communications Technology (ICT), the AI chatbot consumes 0.5 litres of water during each of its lengthy conversations with a user. This applies to all AI systems and LLMs in place. In another research study, ‘Secret Water Footprint of AI Models,’ by Pengfei Li and Shaolei Ren of the University of California, Riverside, it is projected that the pressure AI workloads are exerting on freshwater resources. The 2027 projection, shows that against the concern for global water scarcity, the world’s demand for AI would lead to amounts of water withdrawal – freshwater taken from the ground or surface water sources, temporarily or permanently. This withdrawal would somewhat be equal to the total annual water withdrawal of countries like Denmark and half of the United Kingdom, i.e., between 4.2 to 6.6 billion cubic meters. This is particularly significant when considering the yearly human freshwater consumption, which is around 4 trillion cubic meters, as per the United Nations World Water Development Report and Food and Agriculture Organisation (FAO). AI training data centres are responsible for a significant portion of water usage. As per the research paper, even after putting aside the water usage by third-party facilities, “Google’s self-owned data centres alone directly withdrew 25 billion litres and consumed nearly 20 billion litres of water for onsite cooling in 2022, the majority of which was potable water.” How is AI using water? Water usage by AI systems is categorised into three areas. At first, it includes water consumption for the data centres that house AI workloads. These centres contain high-performance servers and require significant cooling agents due to the continuous generation of heat. Most of their cooling systems, like towers and outside air cooling, are water-sensitive. Cooling towers could utilise up to nine litres of water per kWh of energy consumed during peak times. Secondly, it counts water usage in thermoelectric power plants. These plants generate electricity for data centres and, in return, require water for electricity generation. In America, the average water withdrawal for electricity generation is about 43.8 litres per kWh. In addition to these two, water is also required to manufacture AI chips. This process can consume millions of litres of water daily, especially in processes like water fabrication, which requires ultrapure water. AI’s constant thirst seems to be rising A paper on the study of the economic impact of data centres states the advantages of establishing data centres in regions that may not have stringent regulations. It can facilitate faster and more efficient development since data centres influence economic growth, job creation and competitiveness. In India, climate change impacts are already prevalent, resulting in increased heat waves and droughts due to water scarcity in states like Rajasthan and Nagaland. This creates concern for water availability in the future for both public and industrial use. A report by AIM in 2022 had previously presented India’s significant water consumption by data centres in regions facing water scarcity. The study included Pali a region in Rajasthan, where water had to be transported from nearby cities through special trains. ‘According to a 2019 Niti Ayog report, more than 600 million people in India are water-deprived. Also, more than 21 cities, including Chennai, Hyderabad, Delhi and Bangalore, exhausted their groundwater resources in 2021,’ the report stated. That said, GenAI companies are also adopting measures to promote sustainable development and reduce environmental impact using Green AI, an energy-efficient resource management technique. Are underwater data centres the solution? In June 2024, Microsoft officially confirmed the discontinuation of Project Natick, its underwater data centre initiative. This project began in 2013 and aimed to explore the efficiency of submerged data centres powered by renewable energy. By using seawater for cooling, they aimed to make the cooling process efficient. Despite some promising results, Microsoft decided to halt further development. “We learned a lot about operations below sea level and vibration and impacts on the server. So we’ll apply those learnings to other cases,” said Noelle Walsh, Microsoft’s head of the Cloud Operations + Innovation (CO+I) division. Many others have followed suit and experimented with building successful submerged data centres to conserve freshwater resources. However, experts believe that efficient resource management still needs to be improved. Dr Praphul Chandra, professor and director of the Center for AI & Decentralized Technologies at Atria University, Bengaluru says, “Green AI tries to focus on computational techniques to make AI algorithms more energy efficient. The move towards renewable energy helps indirectly, and future commitments on carbon capture from AI companies rely on market-based solutions. We will need efforts on multiple fronts to solve this problem.”","excerpt":"ChatGPT consumes 10% of an average person’s daily drinking water in one chat.","categories":["AI Features"],"tags":["ai chatbots","Editors Picks","OpenAI","sustainability","Water Footprint"],"author_name":"Sanjana Gupta","publish_date":"2024-10-22T09:23:00","publication_year":"2024","word_count":843,"keywords":["ai chatbots","ChatGPT","GenAI","Water Footprint","Go","OpenAI","AI","sustainability","chatbots","RAG","GPT","Editors Picks","Aim","GAN","R"],"extracted_tech_keywords":["AI","GenAI","ChatGPT","Aim","RAG","chatbots","R","Go","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-chatbots-thirst-may-kill-our-water-resources\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10056793,"title":"IITs set new highs at 2022 placements","content":"The Indian Institutes of Technology across the country this year saw historic placements. From the highest numbers of job offers to the highest packages being offered, most IITs have concluded their Phase I of placements. Here are some of the highlights: IIT Madras Indian Institute of Technology Madras saw a historic high in job offers during Phase I of Campus Placements for the Academic Year 2021-22. A total of 226 companies made 1,085 job offers during this phase. Out of these, 19% of students were placed in the sector of Data Science and Analytics, 18% in Information Technology and Software Development, and 42% in Core Engineering and Technology. There were 45 international offers from 14 companies and 186 from 62 start-ups. When combined with 231 Pre-Placement Offers (PPOs), which are internships, the total number of offers stands at 1,316. As many as 73% of students who opted for Campus Placements received job offers by the end of phase I that concluded on December 10, 2021. Speaking about Phase I of Campus Placements, Prof. C. S. Shankar Ram, Advisor (Placement), IIT Madras, said, “The quality of academic training and the overall development of students during their program at IIT Madras are reflected in this year’s exceptional Phase I Placements. The Institute Placement team thanks companies who have made offers to our students and looks forward to working with more companies in Phase II.” IIT Kanpur The Indian Institute of Technology Kanpur received 940 offers, of which 773 were accepted. IIT Kanpur received 47 international offers, and the highest package offered was USD 274,250, and Rs 1.2 crore was the highest package offered among domestic offers. At the end of phase I, 11.5% of the total jobs offered were from core. Companies like Axtria, EXL, Graviton, Goldman Sachs, ICICI Bank, Intel, Microsoft, OLA, Rubrik, Samsung, Quadeye, Uber, among others, participated in IIT Kanpur’s placement. Professor Abhay Karandikar, Director, IIT Kanpur, said, “IIT Kanpur is known for its academic excellence and as an institute of trust. It is this trust that pulls in top recruiters from across the world year after year. The new highs we are witnessing so far this year is indicative of the growing trust recruiters are bestowing upon the institute and its students, despite the pandemic-induced challenges. We are confident and hopeful that we would end the remainder of the season on a high as well.” IIT Guwahati Indian Institute of Technology Guwahati saw the highest number of international offers, international packages, domestic packages, and even the highest PPOs. The total number of PPOs this year is 179. Offers made by the end of the second day of the placement were 530, out of which 28 were international. The highest domestic package was 1.2 crore, and the highest international offer was 2.05 crore. IIT Bombay A total of 1,201 job offers were accepted till December 7 at the Indian Institute of Technology Bombay. The institute has received a maximum package worth Rs 2.05 crore. A total of 240 companies recruited students, among which Rakuten group with 26 job offers made the highest offers internationally. For domestic jobs, Google, Microsoft, Qualcomm, Boston Consulting Group, Airbus, Bain and company made the highest number of offers. IIT Jodhpur Indian Institute of Technology Jodhpur saw 40 companies registered for the placement drive. In phase one, 128 students were placed as on date that ended on November 2, 2021. The institute also saw four international offers being made. The Computer Science Department saw the highest hiring, with roles for SDE, Analyst and Consultant profiles. Speaking about the placement strategies, Dr Anuj Pal Kapoor, Faculty Incharge, Career Development Cell, IIT Jodhpur, said, “The students’ performance in the interview performance, along with the support from the faculty and administration at IIT Jodhpur, is a testament to our grit and gumption, which ensured almost 55% placements during the first phase. The industry-aligned curriculum along with rigour and relevance brought in respective courses across departments at IIT Jodhpur has enabled us to achieve this mark.” IIT Delhi The Indian Institute of Technology Delhi students bagged 1250 offers in the first phase of placement, which ended on December 15. In addition, the institute received 40 international offers. Around 180 PPOs were received, and seven students opted for deferred placements. Out of the students who opted for the placement facility, 80% got job offers. Out of the students hired, 9% received offers from the Analytics sector, 30% from core and 32% from IT. The top five recruiters on the campus in terms of the number of offers included EXL Analytics, Graviton Research Capital LLP, HCL Technologies, Jaguar Land Rover India Limited, and Microsoft. IIT Hyderabad The Indian Institute of Technology Hyderabad received 466 offers, of which 34 were international, in phase I that concluded on December 7. There were 36 offers from 10 start-ups this year. The highest package offered was Rs 65 lakh. A total of 210 companies have registered, where the top recruiters in the first phase include Microsoft, JP Morgan, NTT, Flipkart, Zomato, Indeed, Meesho and Suzuki Motor Corp. Including the accepted pre-placement offers (PPOs), a total of 427 students have been placed in the first phase. IIT Roorkee Students at the Indian Institute of Technology Roorkee bagged 31 international offers. A total of 11 students received offers of Rs 1 crore and above, while the highest domestic package at this IIT was at Rs 1.8 crore. The highest international offer was Rs 2.15 crore this year. The institute had 219 pre-placement offers (PPOs). IIT Kharagpur Indian Institute of Technology Kharagpur students received 1,600 job offers, the highest placement among all IITs. IIT Kharagpur achieved this feat in Phase I of placement that concluded on December 11. The total number of international offers are more than 35. The highest offer was Rs 2.4 crore. In addition, the Career Development Centre (CDC) received more than 400 Pre Placement offers (PPO). Over 245 companies participated in this placement season across all sectors, including software, high-level coding, analytics, consulting, core engineering companies, banking\/finance, and high-frequency trading.","excerpt":"From over 70% placements in the first phase to packages that go up to Rs. 2.5 crore, IITs this year have seen historic placements","categories":["AI Highlights"],"tags":["IIT","IIT Bombay","IIT Delhi","IIT Guwahati","iit hyderabad","IIT Jodhpur","iit kanpur","iit Kharagpur","IIT Madras"],"author_name":"Meeta Ramnani","publish_date":"2021-12-22T13:00:33","publication_year":"2021","word_count":1006,"keywords":["API","Rust","R","data science","RAG","ViT","IIT Guwahati","analytics","Go","IIT Bombay","IIT Jodhpur","AI","IIT Madras","GAN","iit kanpur","iit hyderabad","iit Kharagpur","IIT Delhi","IIT"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","Go","Rust","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/iits-set-new-highs-at-2022-placements\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":32418,"title":"How Intel Is Redefining Logistics Transparency In The Global Market","content":"You must remember the last time you got a damaged product via some online retail giant. At least you might have doubted the delivery guy for the torn packaging and for getting lemons instead of apple — the iPhone. Long distance delivery is usually very efficient and has a robust feedback system. If we exclude the manual mischief from the middle players, there still will be some kind of hassle in the packaging process. Road transport in India carries about 60% of all the cargo in the country. With total road network of more than 100,00 km, India needs enhanced logistics program as much as any other developed country in the world. Now, Intel has come up with a solution or one can say a tweaking to the already available solution in the form of its Connected Logistics Platform. Intel’s CLP provides services in the logistics and transportation industry to operate at low margins to boost efficiency. Intel’s Connected Logistics Platform allows the customer to automate shipment tracking and increase shipment visibility (e.g. location, temperature, humidity, mishandling) of high-value goods as they move through the supply chain and to their final destination. Early product damage detection helps to return the product in real time, which eliminates the cost of completing the whole trip of delivery, only to be asked for a new one. End customers benefit from getting a jump start on implementing a recovery plan and gain a better understanding of the conditions that cause damage, making it possible to mitigate future damage and avoid costly claims for the shipping companies as well. “The Intel Connected Logistics Platform provides our clients with a powerful solution across a range of different industry sectors for unprecedented real-time transparency of shipment progress and conditions. Some evaluations include findings of temperature excursions for more than 60 percent of the shipment duration, with route and duration improvement opportunities of up to eight hours on a 36-hour shipment. They are also finding significant deviations in shock measurements, up to a doubling in g-forces experienced by sensitive products, depending on the truck and road quality,” said JJ Van der Meer who is a  partner at PA Consulting. Platform Features Based on the Intel Atom processor Provides thorough and continuous updates about package condition Alerts when a package’s location or condition changes unexpectedly Delivers near real-time data analytics and insights Protects confidential asset information with data encryption Why It Matters Trust is the most crucial for packaging industries and, higher the sophistication of the product, higher are the levels of trust demanded. So, there is a strong need for transparency and integrity throughout the supply chain. Individual consumers and large businesses ask for integrity control, especially for sensitive and high-value goods; and logistics companies need transparency for ongoing optimization of networks and assets. With access to real-time data through state-of-the-art IoT devices, all these requests can be answered and trust can be established. Risk mitigation in real time with actionable insights and predictive analytics is helping the companies to optimize across all options and significantly improve their efficiency and reduce supply chain costs. The flexibility of the solution allows for quick integration while solving a wide range of asset-tracking industry challenges. How Well Does Intel Fair In This Space: For Driscoll’s, a California-based seller of fresh berries, quality and temperature control of the produce is critical for longer shelf-life. By using Intel CLP, Driscoll’s was able to see temperature variations across different vendor trucks and track temperature fluctuations, humidity effects, shelf-life and third-party logistics performance. Intel CLP can also help reduce customer shipment rejections due to temperature excursions and deteriorated quality. For International Flavors & Fragrances, Inc. (IFF), product traceability is a key challenge when products are shipped through land, air or sea. Using Intel CLP, IFF tracked shipments in near real-time and provided first-hand visibility into cold chain performance for shipping sensitive temperature-controlled products. IFF was also able to track product exposure to environmental elements and handling during cross dock situations, and quickly relay customs delays or rejections. Source: Intel CLP Intel is scaling via multiple ecosystem channels, including cloud service providers, system integrators, OEMs and vertical solution integrators. How Can India Benefit According to the Ministry of Commerce and Industry, Logistics Services in India is a prominent sector in terms of contribution to national and state incomes, trade flows, FDI as well as employment. Considering the huge contribution of logistics and transportation in economic growth, approx. 14% of the Gross Domestic Product (GDP) is spent on transport and Logistics sector in India in comparison to 8 to 10% in other developed countries. Indian Logistics and Transportation sector comprises of Rail, Road, Water and Air Transport and providing transparency in the way goods are delivered, not only enforces trust with the foreign players but also within the country; enabling a healthy relationship between the shipper and the consumer.","excerpt":"You must remember the last time you got a damaged product via some online retail giant. At least you might have doubted the delivery guy for the torn packaging and for getting lemons instead of apple — the iPhone. Long distance delivery is usually very efficient and has a robust feedback system. If we exclude […]","categories":["Deep Tech"],"tags":["ecommerce","Intel"],"author_name":"Ram Sagar","publish_date":"2018-12-29T05:44:38","publication_year":"2018","word_count":812,"keywords":["Go","programming_languages:R","AI","R","programming_languages:Go","RAG","Aim","ecommerce","analytics","Rust","predictive analytics","Intel"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","predictive analytics","R","Go","Rust","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-intel-is-redefining-logistics-transparency-in-the-global-market\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10088198,"title":"Council Post: Unlock the power of Data Analytics in Life Sciences for Data Driven Future Proofing","content":"The life sciences industry has been rapidly pursuing innovation through digitisation in the past decade. Many businesses have switched from outdated paper-based workflows to sophisticated electronic data capture devices and from simple electronic lab notebooks to data repositories. Over 60% of life science businesses made investments in technology in 2019. D&A is so deeply ingrained in the life sciences industry that it is predicted that between 2021 and 2025, the worldwide life sciences analytics market would expand at a compound annual growth rate (CAGR) of 11.83%, bringing in $15.95 billion in revenue. However, despite having these frequently well-known active digitisation and projects, many businesses still struggle to meet digitisation objectives in R&D. The life sciences business is currently at a position where data- and AI-driven startups that were “born digital” are substantially upending long-established industry conventions—from quickly identifying the optimal patient group for a clinical trial to generating efficient, safe drug candidate chemicals. Data analytics is essential in life sciences because it enables researchers and healthcare professionals to analyse large volumes of data, identify patterns and insights, and make informed decisions about patient care and drug development. Life sciences have a set of unique data challenges such as: Volume: Whether it’s a drug intended to relieve disease in patients or a dietary supplement for kids while causing the least amount of environmental harm, the majority of commercial life sciences organisations are trying to identify compounds and substances that have a particular effect on a certain type of organism. Millions of molecules are initially evaluated in a normal research programme to identify those with intriguing features, thereby producing enormous data sets. Complexity: The most effective products strike a balance between efficacy, safety, cost, and convenience. A pill with few side effects taken by patients at home once a day is far preferable to a medication that treats a disease but results in other health problems and needs to be administered by a medical professional on a regular basis. Heterogeneity: While less organised types of data, such medical records, photographs, and video, have also emerged as important sources of information, life sciences data often takes the form of straightforward numerical statistics. Additionally, data systems used by many labs within the same organisation may have been enhanced for specialised workflows over a long period of time, improving lab processes but resulting in a highly heterogeneous ecosystem of proprietary formats. Understanding the available data and creating standards and ontologies that cover the entire area can be increasingly labour-intensive in order to pull together and harmonise a data landscape like this. Due to this, busy research scientists frequently rely on endlessly adaptable data forms, such as spreadsheets and documents, when they are under pressure to provide results rapidly. However, the ability of the entire firm to exploit data as a long-term asset is harmed by this limited viewpoint. How to overcome these challenges: Findability Companies still have a lot of data that is isolated on user desktops, domain-specific infrastructure, and lab notebooks. Make sure that data can be searched for by relevant parties to facilitate an effective path to insight. This is made possible by: Developing metadata and ontologies that are customised for your business so that data can be searched based on your use cases. Deciding whether to centralise data online, i.e., data catalogue that connects centralised and decentralised resources, or physically, i.e., by way of data lake. Accessibility Data is hidden away in inaccessible silos, which wastes time when data scientists need to compile data for analysis. Developing and putting into practise governance frameworks, open data access models, and transparent, well-documented data access techniques that guarantee the appropriate data is accessible to the right person at the right time is pertinent. Make sure you address security, privacy, licensing, and legislation. Finally, make a distinction between access to data and access to metadata so that users can at least recognise the existence of the data and understand how to request access when the access is restricted. Interoperability Even when data can be located and retrieved, it is seldom accessible in a fashion that allows for its combination with other data and reuse in other modelling or analysis frameworks. Determine the levels of interoperability that are in place and those that must be implemented in order to meet the business need. Reusability It might be challenging for users to identify the origins of the data and whether it is appropriate for their use case when it is useful and available. Data is the Way Forward Define and implement the appropriate culture and technology to enable high-quality data recording, complete with sufficient metadata and provenance, thus ensuring that it can be reused. Start with the highest priority use cases, then cascade out to the rest of the organisation. One of the most promising areas of data-driven innovation in life sciences is precision medicine. Precision medicine uses genomic, environmental, and lifestyle data to tailor medical treatments to individual patients. By using this approach, doctors can develop personalised treatment plans that are more effective and have fewer side effects. Another area of data-driven innovation in life sciences is drug discovery. By using machine learning and other advanced analytics techniques, researchers can analyse large amounts of data to identify potential new drug targets. This can significantly reduce the time and cost of developing new drugs. Real-world evidence (RWE) is yet another area where data-driven innovation has a significant impact. RWE refers to the use of data from real-world sources, such as electronic health records and claims data, to inform decision-making in healthcare. This data can be used to identify patient populations that are at high risk for certain diseases, to monitor the safety and effectiveness of drugs, and to evaluate the cost-effectiveness of different treatments. By leveraging data analytics, artificial intelligence, and machine learning, researchers and healthcare professionals can develop new treatments, improve patient outcomes, and optimise healthcare delivery. As the industry continues to evolve, it will be important for life sciences companies and healthcare providers to continue to invest in developing and applying data-driven approaches to advance scientific discovery and improve patient care. Ultimately, the use of data-driven approaches in the field of life sciences has the potential to revolutionise the way we diagnose and treat diseases, thus leading to better health outcomes and a healthier world. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here","excerpt":"By leveraging data analytics, artificial intelligence, and machine learning, researchers and healthcare professionals can develop new treatments, improve patient outcomes, and optimise healthcare delivery. As the industry continues to evolve, it will be important for life sciences companies and healthcare providers to continue to invest in developing and applying data-driven approaches to advance scientific discovery and improve patient care.","categories":["AI Features"],"tags":[],"author_name":"Raj Babu","publish_date":"2023-02-27T14:05:58","publication_year":"2023","word_count":1093,"keywords":["data science","Go","machine learning","artificial intelligence","AI","Git","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/unlock-the-power-of-data-analytics-in-life-sciences-for-data-driven-future-proofing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10045817,"title":"Ten Years Of Kotlin Programming Language","content":"Ten years ago, JetBrains announced a new statically typed programming language for Java Virtual Machine (JVM) called Kotlin. Since then, a lot has changed. Kotlin has evolved from a Java alternative to a whole ecosystem that allows writing code for different purposes, including server-side, mobile, web, data science, multi-platform projects, machine learning, etc. “It is integrated into our everyday lives. Almost every internet user has come across Kotlin software. If you have an Android phone or tablet — about 80% of the apps you use were built with Kotlin, and its reach extends beyond this platform,” said the Kotlin team. Journey of Kotlin Kotlin’s ten-year journey started with an idea to create a platform that would make development more fun. Begun as a startup project inside JetBrains, Kotlin has come a long way from where it all started. The company said it relied on inputs from its community when making decisions about the future of the technology. “Over the last ten years, Kotlin has accumulated millions of changes based on feedback from thousands of Kotlin users around the world,” said the Kotlin team. The Genesis “It happened very incidentally. The story goes like our co-founder and CEO Sergey Dmitriev, asked a bunch of guys — ‘Guys, what do you think JetBrains can do in terms of big things that would benefit the community, and going to be noticeable,'” recalled Maxim Shafirov, CEO of JetBrains. Dmitry Jemerov, one of the engineers at JetBrains, said that the ultimate thing a development company can do for the development community is a programming language. “I was thinking at the time that JetBrains was a company that was building tools for languages and technologies from other companies, languages, frameworks and so on. So I was thinking of possible ways to increase the influence of JetBrains in the community. So to say, like, how can we do something that is our own technology, and not just supporting other people’s technology,” said Jemerov. Further, he said they had a lot of experience building support for different languages, including Java, JavaScript, Ruby, Python, PHP, Scala, etc. Using this experience, the JetBrains team in 2010 decided to build their own programming language. “Immediately, everybody was laughing. Programming language, really?” said Jemerov. He said we were a small company back then, and we did execute a couple of successful projects like IntelliJ IDEA and ReSharper, but programming languages are in a different league altogether. “However, the seed was there, and we started thinking — we have this deep understanding of many programming languages in their practical aspects. We have implemented support for dozens or maybe 20+ programming languages, and we know all of the drawbacks and good things in many languages so we can combine the things which are good and practical, and many people would use it,” said Jemerov. That was the start of project Kotlin, a statically typed programming language for JVM. The Origin of Kotlin “When I first heard that it was going to be a general-purpose language under JVM, I was like, there is no point in creating a new language. It is completely unreasonable. There were good languages out there at the moment. Scala was pretty big, and I liked it. It felt like all the problems that existed on JVM at the moment were more or less solved by Scala,” said Andrey Breslav, former project lead of the Kotlin Programming Language at JetBrains. “I was like, don’t do it. It is not a good idea. Just use Scala. You will be fine.” Breslav said the conversation went on, and he then started to notice issues that were not solved yet. And, from JetBrains perspective, it was perfectly positioned to accurately launch a new language to attract enough attention, traction and users, etc. “This is still a crazy idea, but it makes sense. We have a shot at doing something really good,” said Breslav. Initially, JetBrains had named the programming language ‘Jet.’ But, due to trademark issues, it had to come up with something else. “We were looking for another name and did not like anything, and it was quite difficult,” said Breslav. Soon, Jemerov suggested ‘Kotlin,’ a name of an island outside of St. Petersburg, which seems to have been inspired by other programming languages like Java (coffee island) and Ceylon (tea island). “There was this progression of islands, and it was a nice idea to use a local name,” he added. In 2011, JetBrains made the initial announcement at the JVM language summit on the Oracle campus in Santa Clara. Since then, Kotlin has been synonymous with the developers’ ecosystem and growing. Over the last 12 months, Kotlin has been used by 4,800,000+ developers for server-side, mobile multi-platform, Android, and front-end development. In addition to this, there are about 194 Kotlin user groups worldwide, and 45 of the top-200 universities are teaching Kotlin. Recently, Kotlin 1.5 was released with the new JVM IR backend and language features. In January this year, JetBrains launched the Kotlin YouTube channel. What’s next? “In the next ten years, we have a lot of work to do. First of all, we will need to establish Kotlin as a multi-platform language firmly, and I see us finishing the work that we started in multi-platform, both infrastructural and in stabilising the support for multiple different platforms,” said Roman Elizarov, project lead for the Kotlin Programming Language at JetBrains. Further, Elizarov plans to build a Kotlin ecosystem that would consist of multi-platform Kotlin-first libraries on multiple platforms that provide a wide range of different abstractions and utilities to developers. He says they are building core foundational libraries for multi-platform and are working on tools to make it easier for their community to develop domain-specific things. Besides multi-platform language, JetBrains is also working on structural data, where developers can start by defining things like collection literals and data in the source code quickly and then deconstruct this data later on. Other works include: Immutability: Kotlin will make it possible to write safe code that does not suffer from sharing mutable data to other threads. Meta-programming: It looks to expand the power of Kotlin inline functions with constant evaluation and propagation and enable compiler plugins for advanced compile-time manipulation of the code. New compiler:  It aims to lay the groundwork for the future evolution of the language without sacrificing performance. More static typing: Kotlin will make it easier for tools to help developers with their code. Elizarov said Kotlin will not be just a programming language; it will be a multi-platform ecosystem of libraries and tools that help people write code in various domains, including data science, gaming, mobile applications, desktop, web — everything! “That is how we are going to see Kotlin in the next ten years.”","excerpt":"Initially, JetBrains had named the programming language ‘Jet.’","categories":["AI Features"],"tags":["coding","JetBrains","Kotlin","Machine Learning","Machine Learning Latest","Python Programming"],"author_name":"Amit Naik","publish_date":"2021-08-12T16:00:00","publication_year":"2021","word_count":1125,"keywords":["data science","Go","machine learning","AI","Kotlin","coding","ML","Machine Learning","Machine Learning Latest","Python","JetBrains","Aim","JavaScript","Python Programming","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Aim","Python","R","JavaScript","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ten-years-of-kotlin-programming-language\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10116863,"title":"Cropin and AWS Partner To Develop AI-Solutions to Address Global Hunger","content":"Bengaluru-based agritech startup Cropin Technology and Amazon Web Services (AWS) India Private Limited have signed a Memorandum of Understanding (MoU) focused on enabling Cropin to build an AI-powered solution to address the pressing issue of global hunger and food insecurity. This initiative aims to help Cropin develop core data architecture, analytics, modeling, and simulation components that can aggregate global farmland data and broader climate intelligence within a single solution. The solution will provide decision intelligence to governments, development agencies, and agri-businesses, and help them ensure food security for vulnerable populations. It will also integrate satellite imagery with in-situ field images and remote data to improve agricultural analytics through scalable models. These models will provide both micro (plot) and macro (regional\/ global) insights and will be further analyzed by identifying patterns and anomalies in the production and quality of major crops across global regions. Cropin’s AI, model building, data processing, and reporting will leverage AWS services such as Amazon Bedrock, a fully managed service that offers a choice of high-performing foundation models from leading AI companies; Amazon Q, a generative AI-powered assistant; Amazon QuickSight, which offers unified business intelligence at hyperscale; and AWS’s HPC infrastructure. AWS will explore providing technical expertise to Cropin on its advanced compute services (HPC, ML\/ Gen AI, IoT, geospatial), as well as industry insights from its agriculture, and sustainability specialists, to power Cropin’s platform. The insights generated through the workloads will be integrated into Cropin’s open-source dashboard. They can be disseminated via a WhatsApp or an SMS-based alerting system for stakeholders, including farmers, field officers, governments, development agencies, and agribusinesses. AWS will further support Cropin by exploring collaboration opportunities with research organisations and academic institutions such as Harvard Data Science Initiative (HDSI), to drive research and development in food security, climate-resilient agriculture, and food sustainability. Recent estimates from the Food and Agriculture Organization (FAO) of the United Nations reveal a stark reality: between 691 and 783 million people faced hunger in 2022, reversing decades of progress. This figure represents an alarming increase of 122 million people compared to 2019, before the pandemic. While technology alone cannot solve hunger, it is a crucial tool for strategic decision-making. It enables governments and organizations analyse insights emerging from models and simulations of systems as diverse as agriculture, trade, and climate, in order to develop and test holistic strategies to address food insecurity. “Our work with Cropin showcases the power of advanced compute capabilities on the cloud to drive social and environmental impact. AWS’s generative AI, simulation, and data analytics technologies can help organisations like Cropin surface actionable and relevant insights from diverse data sets, and scale their solution globally to empower decision makers to reduce food insecurity,” said Shalini Kapoor, Director and Chief Technologist, AWS India Private Limited.","excerpt":"Cropin will leverage AWS services such as Amazon Bedrock, Amazon Q, Amazon QuickSight, and AWS’s HPC infrastructure.","categories":["AI News"],"tags":["AWS","CropIn"],"author_name":"Pritam Bordoloi","publish_date":"2024-03-20T18:05:54","publication_year":"2024","word_count":459,"keywords":["data science","AWS","AI","ML","RAG","Aim","CropIn","analytics","generative AI","foundation models","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","generative AI","foundation models","Aim","RAG","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cropin-and-aws-partner-to-develop-ai-solutions-to-address-global-hunger\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171428,"title":"Genpact Acquires XponentL Data to Strengthen AI and Data Capabilities","content":"Genpact has acquired XponentL Data, a provider of data products and AI solutions, to bolster its AI and data strategy capabilities. The acquisition supports Genpact’s ongoing shift towards AI-first services, reinforcing its offerings across strategy, implementation, and industry-specific applications. The deal, which closed on June 5, will bring XponentL’s expertise in data strategy, design, and engineering into Genpact’s portfolio. XponentL will also add partnerships with leading platforms, including Databricks, Amazon Web Services, and Microsoft. “We believe the future belongs to companies that deploy AI at speed and scale—on a bedrock of data expertise,” Balkrishan ‘BK’ Kalra, president and CEO of Genpact, said. “XponentL’s robust intellectual property will further strengthen how Genpact can help clients harness the full value of their data.” XponentL aims to bring capabilities across life sciences, healthcare, and enterprise AI strategy. Genpact said the combination will fuel its service-as-a-service (SaaS) and AI Gigafactory initiatives, helping clients scale their AI transformations with operational efficiency. “Genpact and XponentL share a vision to unlock the power of data for our clients—this opportunity allows us to take it to the next level,” said Tom Johnstone, founder of XponentL, who will continue to lead the business. Meanwhile, Michael Hartman, senior vice president at Databricks, called the acquisition a “unique combination of innovation and scale”, noting XponentL’s experience applying intelligent agents and modern architectures on the Databricks platform. While all XponentL employees will join Genpact, the financial terms of the deal were not disclosed.","excerpt":"The deal will bring XponentL’s expertise in data strategy, design, and engineering into Genpact’s portfolio.","categories":["AI News"],"tags":["Genpact","Mergers and Acquisitions"],"author_name":"Siddharth Jindal","publish_date":"2025-06-06T12:29:08","publication_year":"2025","word_count":241,"keywords":["Go","AI-first","Genpact","AI","programming_languages:R","innovation","AI transformation","Aim","cloud_platforms:Amazon Web Services","Mergers and Acquisitions","R","Databricks"],"extracted_tech_keywords":["AI","Aim","Databricks","R","Go","AI-first","AI transformation","innovation","cloud_platforms:Amazon Web Services","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/genpact-acquires-xponentl-data-to-strengthen-ai-and-data-capabilities\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":19384,"title":"16 Free Public Datasets For Data Analysts in 2024","content":"While many experts are saying that Data is as precious as oil in this century, the need for free, simple datasets for analytics projects is important as well. As any beginner would reveal, their first projects have helped them immensely in kick-starting their careers into the world of analytics. We have listed the following 16 free datasets from where any beginner can pick out relevant data for his or her projects. And the best part is, it’s all free. Government of India: Data.gov.in is a portal for encouraging ‘Open Data Initiative’ undertaken by the Government of India. This joint initiative of the Governments of India and the US enables the Ministries to publish datasets collected by them for public use. Twitter: Twitter dataset is a reliable platform which contains all the tweets and user details which enables one to perform interesting analysis on. This can be deployed to create user communities and suggest suitable followers in accordance with the genre of tweets. Google N-Grams: If you’re interested in truly massive data, the Google n-grams dataset counts the frequency of words and phrases by year across a huge number of text sources. The resulting file is 2.2 TB. Amazon: Amazon Web Services datasets can be analyzed in the cloud using EC2 and Hadoop via EMR. Amazon Web Services renders an entire toolkit for analyzing data at any scale. YouTube: This is a video dataset consisting of millions of YouTube video IDs and associated labels from a diverse vocabulary of over 4700 visual entities. Their main motive is to accelerate research work on video understanding. Buzzfeed News: Surprisingly, the website famous for its extensive reportage on celebrities and pop culture makes the data sets used in its articles available on Github. Kaggle: Kaggle has created an array of high-quality public datasets known as Kaggle Datasets for hassle-free access and analysing the data without downloading it. Work done in Kaggle is saved and published publicly by default which enables newcomers to modify the work done by other data scientists. Socrata OpenData: This is a platform consisting of multiple clean data sets that can be explored in the browser or downloaded to work on. Registration, however, is not required. It allows you to use visualization and exploration tools to explore the data in the browser and choose from hundreds of open data catalogs. Bitbucket: This web-based hosting service owned by Atlassian, is written in Python and uses the Django web framework. This web portal allows unlimited public repositories for all and private repositories free for up to five users. GitHub: GitHub, one of the largest web-based hosting services for developmental projects renders its services free of cost for public repositories. In addition to an integrated issue tracker right within your project, GitHub also supports over 200 programming languages. UCI: The UCI Machine Learning Repository is an amalgamation of databases, domain theories, and data generators that are utilized by the machine learning community for the critical analysis of machine learning algorithms. Amazon Reviews: This dataset consists of product reviews and metadata from Amazon that can be used by researchers for analytics projects. It consists of 142.8 million reviews spanning from May 1996- July 2014. Google BigQuery Public Datasets: The public datasets listed in the BigQuery documentation are datasets that Google BigQuery hosts for you to access and integrate into your applications. Google pays for the storage of these datasets and provides public access to the data via a project. You pay only for the queries that you perform on the data (the first 1 TB per month is free, subject to query pricing details). World Bank: The World Bank is a global development organization that offers loans and advice to developing countries. The World Bank regularly funds programs in developing countries, then gathers data to monitor the success of these programs. You can browse World Bank data sets directly, without registering. The data sets have many missing values, and sometimes take several clicks to actually get to data. Reserve Bank Of India: The data available from the Reserve Bank of India includes several metrics on money market operations, balance of payments, use of banking and several products. A must go to site, if you come from BFSI domain in India. Ministry of Statistics and Programme Implementation: The MOSPI has a collection of varied datasets, ranging from the Statistical Yearbook, ASI summaries, to National Accounts Data, for data analysts to pore over.","excerpt":"While many experts are saying that Data is as precious as oil in this century, the need for free, simple datasets for analytics projects is important as well. As any beginner would reveal, their first projects have helped them immensely in kick-starting their careers into the world of analytics. We have listed the following 16 […]","categories":["AI Trends"],"tags":["Amazon","Data Analytics","Datasets","free datasets for analysis","GitHub","government of india","hadoop data catalog","Kaggle","reserve bank of india","Twitter (X)","YouTube"],"author_name":"Prajakta Hebbar","publish_date":"2017-11-28T12:23:04","publication_year":"2017","word_count":736,"keywords":["government of india","Git","Ray","Twitter (X)","R","Datasets","RAG","analytics","YouTube","reserve bank of india","Go","machine learning","AI","Amazon","hadoop data catalog","free datasets for analysis","Kaggle","Python","Data Analytics","GitHub"],"extracted_tech_keywords":["AI","machine learning","analytics","Ray","RAG","Python","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/16-free-public-datasets-new-data-analysts-pore\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":54961,"title":"Manipal Institute Of Technology Offers First-Of-Its-Kind Course In Data Science &#038; Engineering","content":"Manipal Institute of Technology, part of Manipal Academy of Higher Education, has announced an enhanced curriculum for the academic year 2020-21. The college will offer, for the first time in India, B-Tech programmes in data science and engineering. As pioneers in personalising curriculum based on interest and aptitude, the institute aims to provide students with a holistic learning experience. The key highlights of the course would be core competency in data science, computational mathematics and statistics, minor specialisation in finance, business, health care, and multi-campus model with mandatory student mobility in one semester, industry collaboration for the teaching-learning process, and option of semester abroad along with an option of integrated master’s program with foreign universities. Speaking on the occasion, Dr D Srikanth Rao, Director, Manipal Institute of Technology, MAHE said, “Today data holds great importance, and in fact, there is a huge global demand for data scientists. To cater to the global demand for skilled data scientists, Manipal Institute of Technology is launching a B-Tech course in data science and engineering across the group campuses in Manipal, Jaipur and Sikkim. Keeping in mind the uniqueness of the skill set required for data science, MIT has designed state-of-the-art syllabus in consultation with data science experts, the IT industry and academia drawn from diverse domains.” He further added, “The proposed B-Tech program will have a strong emphasis on data science and data engineering from industry perspectives, which differentiates it from other computer science or IT programs. The students will be studying basic science and the standard engineering subjects in the first year. In the second and third year, they will be specialising in data science-related topics along with the required computational mathematics and statistical skills. The core subjects taught include data analytics, machine learning, big data, and deep learning. Industry 4.0 demands smart systems integrated with intelligence to have a better human-machine interface. The artificial intelligence and the deep learning subjects are designed to focus on state-of-the-art cognitive modelling and the brain-machine convergence.” The course will also be one of India’s first multi-campus models that promote one-semester student mobility between the Manipal, Jaipur and Sikkim campuses. The students will study in their respective home campuses in the first three semesters and will move to another campus for at least one semester in the fourth, fifth, or sixth semesters. The institution’s placement cell, industry collaborators, and alumni network will facilitate placement and internships as data scientists, data analysts, and data engineers in core companies as well as varied domains such as finance, business, economics and healthcare. Along with the core competencies, the students can opt for custom-designed program electives in subjects such as data engineering, quantum computing, data forensics, data privacy, and security. The course also offers three domain-specific minor specialisations with four program electives in the 7th semester. The students can specialise in finance and security analytics, business analytics, or healthcare analytics.","excerpt":"Manipal Institute of Technology, part of Manipal Academy of Higher Education, has announced an enhanced curriculum for the academic year 2020-21. The college will offer, for the first time in India, B-Tech programmes in data science and engineering. As pioneers in personalising curriculum based on interest and aptitude, the institute aims to provide students with […]","categories":["AI News"],"tags":["big data and industry 4.0","big data in industry 4.0","big data industry 4.0","Data Science","Data Science Career","data science india","Data Science Jobs","Data Scientist","data scientist india salary","industry 4.0 and big data"],"author_name":"Sejuti Das","publish_date":"2020-01-31T13:52:50","publication_year":"2020","word_count":480,"keywords":["deep learning","R","data science","big data and industry 4.0","artificial intelligence","big data industry 4.0","data scientist india salary","industry 4.0 and big data","analytics","Data Science","machine learning","AI","Data Science Jobs","Data Science Career","big data in industry 4.0","big data","data science india","Aim","data engineering","Data Scientist"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","Aim","R","big data","data engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/manipal-institute-of-technology-offers-first-of-its-kind-course-in-data-science-engineering\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10015980,"title":"Exploring ArtLine &#8211; To Create Line Art Portraits, Movie Posters &#038; Cartoonize Images in Python","content":"Deep Learning and Computer Vision has evolved and done wonders time and again. Today we are going to talk about one such recently done amazing project called ‘ArtLine’ that uses deep learning algorithms to achieve fine quality line art portraits. Besides that, it also can be used to generate movie posters and cartoonize images. It is currently the most trending topic in both GitHub and paperswithcode. It is created by Vijish Madhavan, a deep learning researcher. The model has been built using the APDrawing dataset and Anime line art pair using many different algorithms, derived from some research papers self-attention, progressive resizing and generator loss. It shows how stacking all the methods can generate high-quality results. Primarily PyTorch and Fastai libraries are used. It generates fine lines\/edges in the sketch image, which is better than most existing methods. Try out the demo from this Colab Notebook with any portrait picture which is expected in an URL and then converted to image formats. You can clone the repository or tweak the code to use your local image file and within less than 2 minutes (executing with GPU) have a look at the amazing results. Get a cartoon version of Tom Hanks Here’s a movie poster generated by ArtLine As of now, the movie poster and cartoon generating models have not been released. We can soon expect them in the near future. Only the pre-trained line art portrait generating model is available. Let’s explore how the model training takes place. Importing necessary libraries import torch import torch.nn as nn import fastai from fastai.vision import * from fastai.callbacks import * from fastai.vision.gan import * from torchvision.models import vgg16_bn from fastai.utils.mem import * from PIL import Image import numpy as np from torch.autograd import Variable import torchvision.transforms as transforms Edge Detection – this function uses the convolutional neural network to detect edges from an image as to features and use it as a gradient. def _gradient_img(img): img = img.squeeze(0) ten=torch.unbind(img) x=ten[0].unsqueeze(0).unsqueeze(0) a=np.array([[1, 0, -1],[2,0,-2],[1,0,-1]]) conv1=nn.Conv2d(1, 1, kernel_size=3, stride=1, padding=1, bias=False) Building the neural network and assigning weights conv1.weight=nn.Parameter(torch.from_numpy(a).float().unsqueeze(0).unsqueeze(0)) G_x=conv1(Variable(x)).data.view(1,x.shape[2],x.shape[3]) b=np.array([[1, 2, 1],[0,0,0],[-1,-2,-1]]) conv2=nn.Conv2d(1, 1, kernel_size=4, stride=2, padding=2, bias=False)  conv2.weight=nn.Parameter(torch.from_numpy(b).float().unsqueeze(0).unsqueeze(0)) G_y=conv2(Variable(x)).data.view(1,x.shape[2],x.shape[3]) G=torch.sqrt(torch.pow(G_x,2)+ torch.pow(G_y,2)) return G gradient = TfmPixel(_gradient_img) PATH – redirecting to the saved APDrawing dataset and a selective picture from the Anime sketch colourization pair. path = Path('\/content\/gdrive\/My Drive\/Apdrawing') Blended Facial Features path_hr = Path('\/content\/gdrive\/My Drive\/Apdrawing\/draw tiny') path_lr = Path('\/content\/gdrive\/My Drive\/Apdrawing\/Tiny Real') Portrait Pair path_hr3 = Path('\/content\/gdrive\/My Drive\/Apdrawing\/drawing') path_lr3= Path('\/content\/gdrive\/My Drive\/Apdrawing\/Real') Architecture -  pretrained resnet34 model is used arch = models.resnet34 Detecting Facial Features src = ImageImageList.from_folder(path_lr).split_by_rand_pct(0.3, seed=42) def get_data(bs,size): data = (src.label_from_func(lambda x: path_hr\/x.name) .transform(get_transforms(xtra_tfms=[gradient()]), size=size, tfm_y=True) .databunch(bs=bs,num_workers = 0).normalize(imagenet_stats, do_y=True)) data.c = 3 return data Progressive resizing by the Fastai library helps gradually increase the size of the image and the adjusting learning rates, thereby generalizing the images as it goes through different stages. 64px bs,size=20, 64 data = get_data(bs,size) data.show_batch(ds_type=DatasetType.Valid, rows=2, figsize=(9,9)) t = data.valid_ds[0][1].data t = torch.stack([t,t]) def gram_matrix(x): n,c,h,w = x.size() x = x.view(n, c, -1) return (x @ x.transpose(1,2))\/(c*h*w) gram_matrix(t) base_loss = F.l1_loss vgg_m = vgg16_bn(True).features.cuda().eval() requires_grad(vgg_m, False) blocks = [j-1 for j,o in enumerate(children(vgg_m)) if isinstance(o,nn.MaxPool2d)] blocks, [vgg_m[i] for i in blocks] Perpetual loss is calculated for image transformations based on the VGG_16 model. It speeds up training. This approach combines both a per-pixel loss between the output and ground-truth images and optimizing perceptual loss functions based on high-level features extracted from pre-trained networks. The results are then used to train a feed-forward network. class FeatureLoss(nn.Module): def __init__(self, m_feat, layer_ids, layer_wgts): super().__init__() self.m_feat = m_feat self.losses = [self.m_feat[i] for i in layer_ids] self.hooks = hook_outputs(self.losses, detach=False) self.wgts = layer_wgts self.metrics_name = ['pixel',] + [f'feat_{i}' for i in range(len(layer_ids))] + [f'gram_{i}' for i in range(len(layer_ids))] def make_features(self, x, clone=False): self.m_feat(x) return [(p.clone() if clone else p) for p in self.hooks.stored] def forward(self, input, target): out_feat = self.make_features(target, clone=True) in_feat = self.make_features(input) self.feat_losses = [base_loss(input,target)] self.feat_losses += [base_loss(f_in, f_out)*w for in, out, w in zip(in_feat, out_feat, self.wgts)] self.feat_losses += [base_loss(gram_matrix(in),  gram_matrix(out))*w**2 * 5e3 for in, out, w in zip(in_feat, out_feat, self.wgts)] self.metrics = dict(zip(self.metric_names, self.feat_losses)) return sum(self.feat_losses) def __del__(self): self.hooks.remove() feat_loss = FeatureLoss(vgg_m, blocks[2:5], [5,15,2]) wd = 1e-3 y_range = (-3.,3.) This function uses the self-attention model generator with U-Net and spatial normalization. This is a No GAN training which stabilizes colour images. Here minimal time is spent in direct GAN training instead, separately pretraining the generator and critic. This was introduced in another project named DeOldify. It helps largely in getting accurate facial features. def create_gen_learner(): return unet_learner(data, arch, wd=wd, blur=True, norm_type=NormType.Spectral,self_attention=True, y_range=(-3.0, 3.0),loss_func=feat_loss, callback_fns=LossMetrics) gc.collect(); learn_gen = create_gen_learner() learn_gen.lr_find() lr = 1-01 epoch = 5 fitting the model def do_fit(save_name, lrs=slice(lr), pct_start=0.9): learn_gen.fit_one_cycle(epoch, lrs, pct_start=pct_start,) learn_gen.save(save_name) learn_gen.show_results(rows=1, imgsize=5) do_fit('da', slice(lr)) #lr*10 learn_gen.unfreeze() learn_gen.lr_find() epoch = 5 do_fit('db', slice(1E-2)) Results for different pixel values 128px data = get_data(8,128) learn_gen.data = data learn_gen.freeze() gc.collect() learn_gen.load('db'); epoch =5 lr = 1E-03 do_fit('db2',slice(lr)) learn_gen.unfreeze() epoch = 5 do_fit('db3', slice(1e-02,1e-5), pct_start=0.3) 192px data = get_data(5,192) learn_gen.data = data learn_gen.freeze() gc.collect() learn_gen.load('db3'); epoch =5 lr = 1E-06 do_fit('db4') learn_gen.unfreeze() epoch = 5 do_fit('db5', slice(1e-06,1e-4), pct_start=0.3) Acquiring data for portrait images src = ImageImageList.from_folder(path_lr3).split_by_rand_pct(0.2, seed=42) def get_data(bs,size): data = (src.label_from_func(lambda x: path_hr3\/x.name) .transform(get_transforms(max_zoom=2.), size=size, tfm_y=True).databunch(bs=bs,num_workers = 0).normalize(imagenet_stats, do_y=True)) data.c = 3 return data 128px data = get_data(8,128) learn_gen.data = data learn_gen.freeze() gc.collect() learn_gen.load('db5'); data.show_batch(ds_type=DatasetType.Valid, rows=2, figsize=(9,9)) learn_gen.lr_find() epoch = 5 lr = 1e-03 do_fit('db6') learn_gen.unfreeze() epoch = 5 do_fit('db7', slice(6.31E-07,1e-5), pct_start=0.3) 192px data = get_data(4,192) learn_gen.data = data learn_gen.freeze() gc.collect() learn_gen.load('db7'); learn_gen.lr_find() epoch = 5 lr = 4.37E-05 do_fit('db8') learn_gen.unfreeze() epoch = 5 do_fit('db9', slice(1.00E-05,1e-3), pct_start=0.3) Endnotes Limitations – Needs smooth or plain backgrounds to process and works poorly with lighting or shadows. Works poorly on low-quality images even. Nevertheless, ArtLine is achieving pretty good state-of-the-art results, and the project is constantly under development.","excerpt":"ArtLine uses deep learning algorithms to achieve fine quality line art portraits, movie posters and cartoonize images.","categories":["Deep Tech"],"tags":["Computer Vision","Deep Learning","Python"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-12-27T14:00:00","publication_year":"2020","word_count":960,"keywords":["CUDA","NumPy","TPU","AI","neural network","PyTorch","computer vision","Python","Ray","Colab","deep learning","Computer Vision","Deep Learning"],"extracted_tech_keywords":["AI","deep learning","neural network","computer vision","Ray","PyTorch","Colab","NumPy","TPU","CUDA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/artline-to-create-line-art-portraits-in-python\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":44377,"title":"AI For Greener Good: Solar Panels Too Can Benefit From AI","content":"The first demonstration of photovoltaic effect took place in the mid 19th century but it took more than a century for this discovery to come to fruition in the form of solar panels. Solar panels work on the principle of using the radiation of the sun to generate charge that can power a myriad of devices. The surface where the sun hits, therefore, becomes quite crucial. The amount of energy tapped depends on the surface area and other material properties. So, what is the link between how a surface is designed and artificial intelligence It is no secret that AI has become well-suited to a host of day-to-day problems. The secret sauce of AI is optimization. So, one can easily connect the dots between how AI can be used to manipulate material properties. For solar panels, especially the coatings on its surface influence the reflectivity of the panel, which in turn affects the power generated per unit area. We can’t just give away larger spaces for power. The panels should be compact and should pack a punch. In what can sound as a bizarre interdisciplinary cross over, researchers at  the University of Auvergne in France tweak in evolutionary algorithms to achieve greater levels of optimization. The class of algorithms that they used, called differential evolution (DE), has proven particularly successful for optimizing of  photonic structure of a nanoscale-size coating. Previous research and manufacturing techniques used mathematical functions such as Gaussian and exponential functions to determine the structure of AR coatings. This algorithm, however, was able to find a structure with half the thickness of those obtained by traditional approaches. Changing Photonic Structures Using AI The whole premise of this innovation is based on the fact that thinner coatings are potentially better at letting light through. So, in this AI-based approach, the researchers aimed to manufacture a thinner anti-reflective (AR) coating, which can be used to maximize the light available to photovoltaic solar panels. At its core, this technique optimizes the pattern and structure of an anti-reflective coating based on its thickness by applying a biologically inspired AI algorithm to the design process. The overall goal is to refract, rather than reflect, incoming light at a wide range of frequencies and angles. By giving the DE algorithm a specific thickness to optimize for, the researchers were able to outperform traditional techniques. The results demonstrate the feasibility of nanostructured AR coatings. Though this work could have implications for lenses and other components that rely on controlling reflectivity to increase optical efficiency, the most important benefits could be in applying AR coatings to the silicon in photovoltaic (PV) solar panels. Such coatings already provide a significant boost in efficiency for PV cells, and by replicating that improvement with fewer layers and less overall thickness, the researchers hope to contribute to making solar generation even more cost-effective. Future Direction The hunt for finding a sustainable clean energy solution has never been this vigorous. Researchers are finding out new ways to address the shortcomings.  This research is an example of employing AI concepts and theory to advance materials science. India with its renewed focus on clean, renewable energy has to tap into technologies to undertake large-scale sustainable power projects and promote green energy. According to the Department of Commerce, a  total of 47 solar parks with generation capacity of 26,694 MW have been approved in India up to November 2018, out of capacity of 4,195 MW has been commissioned. Solar capacity has increased by eight times between FY14-18. India added record 11,788 MW of renewable energy capacity in 2017-18. With innovations such as above, countries like India, where there is a dire need for cleaner energy resources, will benefit immensely. The improvements can be at the nano level but the outcomes will be on a much larger scale. The growth of algorithmic based solutions have time and again proven to be profitable and are believed to usher myriad of applications given the alarming need for cost-cutting and climate saving strategies. Access the full paper here.","excerpt":"The first demonstration of photovoltaic effect took place in the mid 19th century but it took more than a century for this discovery to come to fruition in the form of solar panels. Solar panels work on the principle of using the radiation of the sun to generate charge that can power a myriad of […]","categories":["Deep Tech"],"tags":["evolutionary algorithm","Facebook AI","solar panels"],"author_name":"Ram Sagar","publish_date":"2019-08-13T10:50:28","publication_year":"2019","word_count":670,"keywords":["Go","artificial intelligence","solar panels","Facebook AI","AI","programming_languages:R","innovation","programming_languages:Go","Aim","ViT","evolutionary algorithm","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ai-for-greener-good-solar-panels-too-can-benefit-from-ai\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":22629,"title":"Women in Analytics: Amruta Purandare Of Happiest Minds Believes That Gender Balance In Workplaces Is Crucial For Growth","content":"Having worked as a data scientist at MakeMyTrip and project lead in the analytics team at Persistent Systems Ltd., Amruta Purandare currently leads the AI and Cognitive Computing group as Associate Director in the Analytics Centre of Excellence (CoE), at Happiest Minds Technologies. With B.E in Computer Engineering from Pune University and Masters in Computer Science from University of Minnesota, USA, she has a core area of interest around Computer Vision, Image Processing, Video Analytics and Multimedia Technology, IoT, Edge Computing for Smart Homes and Connected Cars. Having lead cognitive computing projects such as developing smart kiosk for retail customers, audio classification, health bot, damage detection etc., she believes that career in analytics and data science for a woman is definitely exciting exciting and promising, with lots of opportunities. “Having niche skills in emerging technologies like AI, robotics, computer vision, image processing, edge computing, Internet of Things, etc. is a continuous learning process, but it definitely helps us stand out in today’s job market”, she says. AIM interacted with Amruta on Women’s day, and she had many interesting facts to reveal. On choosing analytics as a career option I found my passion for AI and Machine Learning during my undergraduate years, and decided to pursue it further with higher studies. Up until 2014, Data Science and Machine Learning was still perceived more of an academic research field, with very few job opportunities especially here in India. But since 2014, this field has gained a momentum like never seen before. Today, almost every industry sector, whether it’s retail, banking, automotive, manufacturing, health-care, education, travel or entertainment, has lots of customer data acquired over the years, and as a result, need for applied data science and analytics skills. On her growth story Working with global companies like Amazon and Sony, and in foreign countries like US, Japan, Singapore has definitely given an exposure to real-world problems, and jump start for my career in AI and machine learning. I have been very fortunate to work with companies like Persistent, MakeMyTrip and now Happiest Minds which inspire innovative and creative thinking. Every client meeting or discussion gives me an opportunity to brainstorm new ideas on how some of the same techniques can be applied in variety of different situations or industry sectors, to solve interesting business problems and use-cases. In Happiest Minds especially, I am given a lot of freedom to explore new emerging areas around computer vision, edge computing, video analytics, smart homes and robotics. On maintaining a work-life balance Most companies these days offer flexible hours and work from home policies, but having a good manager who allows some flexibility is equally essential, especially when dealing with personal crisis, emergencies or tragedies at home. I have recently experienced this first-hand, having lost my mother after a long hospitalisation. Situations like this totally change our whole perspective towards life and career. On incorporating more women in new tech sectors and how it can be done The hiring process should definitely be merit-based where someone is hired for their skills and talent, not because they are men or women. However, a few years of career gap in someone’s resume should not be a deal breaker during hiring process, especially if that candidate is a woman. I think, having women in top management positions certainly reflects very positively on company’s work culture, in otherwise male dominated society\/country. There is something fundamentally different about the way men and women think or approach a problem, and that’s why I feel maintaining a good gender balance in workplaces is crucial for growth and development of our country and society. Online course platforms like Coursera, Udacity, EdX and Udemy, are definitely game changers, providing learning opportunities for everyone around the globe, to upgrade their skills, and earn certificates from top universities which may have been out of reach for many international students in the past. Companies should sponsor and encourage such training programs and online courses to help their employees update their skills and knowledge, to cope with the growing demand for AI \/ ML skills.","excerpt":"Having worked as a data scientist at MakeMyTrip and project lead in the analytics team at Persistent Systems Ltd., Amruta Purandare currently leads the AI and Cognitive Computing group as Associate Director in the Analytics Centre of Excellence (CoE), at Happiest Minds Technologies. With B.E in Computer Engineering from Pune University and Masters in Computer […]","categories":["AI Features"],"tags":["analytics leaders india","automotive analytics","Interviews and Discussions","Women in Analytics"],"author_name":"Srishti Deoras","publish_date":"2018-03-15T08:10:25","publication_year":"2018","word_count":677,"keywords":["data science","machine learning","AI","ML","analytics leaders india","computer vision","RAG","automotive analytics","Aim","analytics","Women in Analytics","edge computing","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","data science","analytics","Aim","RAG","edge computing","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/women-in-analytics-amruta-purandare-of-happiest-minds-believes-that-good-gender-balance-in-workplaces-is-crucial-for-growth\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":29210,"title":"AI-Powered TVs Are Changing The Way We Perceive Entertainment","content":"Image credit: Banksy (Wallfillers) Television sets have come a long way from being a simple box displaying grayscale images to the Smart 4K televisions today with hundreds of channels to browse through. As if this wasn’t overwhelming, the next phase of technology is now integrating itself with AI to power a more user-friendly, comfortable experience on television. AI techniques like face recognition, personalised video or songs suggestions are now being infused to television sets to give the consumers a richer experience. For instance, Google Assistant has already been working with television sets to provide an AI experience, among several other television makers who are taking help from AI experts from firms such as Google and Amazon. 5 Famous Attempts So Far In AI-Powered Television LG: LG launched its AI enabled OLED W8 TV, ThinQ, in July this year in India. The name has been derived from the combination of the phrase and the word “think of you” and “cue” respectively. Has over 800 commands and the users can operate using voice commands, by pressing and holding the mike on the remote. The tasks that it can perform are turning the television set on or off, control the volume, changing the picture mode settings, also getting the weather conditions from the internet, providing facts about the movie that you are watching, and so on. Moreover, it can also be used to control the other smart devices in the house, devices that are connected to the television set by Wi-Fi or Bluetooth. Sony: Sony is also trying to seek help from AI in their televisions, just like LG. It does not have an on-device AI but has added Google Assistants to their 4K HDR TV sets. It has all the features similar to that of LG’s, with a voice command feature and give information about a movie or a game, basically all things that a Google Assistant is capable of on the television. Samsung: AI has so far been majorly used in enhancing the picture and video quality on televisions. It analyses the image and reduces noise and then upscales the picture to a better quality. It has colour correlation algorithms that help them get a more realistic view of the image. Samsung is using AI to make the videos on their 8K TVs look better, with its QLED 900R series. They can take any 4K, 1080K or even 720K video and turn them into high quality 8K videos, and has 16 times the number of pixels as ordinary HD, using their AI technology. Its AI tech counts the number of pixel in the image and is then able to tell where the pixels should be added for the picture’s quality to be 8K. Although it has not yet launched its AI television yet, it had made its announcement at IFA this year and is soon going to be released.  It is going to integrate the AI called Bixby assistant available on Galaxy smartphones, along with its AI. TCL: A China-based multinational electronics company TCL announced a range of AI-enabled television products at the IFA conference held in Berlin this year. Shenzhen Coocaa Network Technology: Another Chinese television manufacturer Shenzhen Coocaa Network Technology has announced its smart television unit integrated with Baidu’s AI assistant system called Duer OS to plant AI in its televisions. Their television will be called Coocaa, which is also an intelligent OS and is a subsidiary of Skyworth. Indian Scenario A Bangalore-based startup of four members called Sensara Technologies showed how the AI enables television will be like. Although they did not invent an AI television, they invented an Android-based AI device called Sensy Fusion that helps in providing a personalised list of shows and channels that you would prefer to watch, based on your interests. It does not demand the users to buy a new television set for this product. It can work on any television set and set-top box. Another feature that it has is that it can automatically generate trailers and signature clips of shows. This product equipped with television would make the television an AI TV. Tech giants around the world are trying this very thing, to inculcate AI in televisions by combining the televisions with AI products. Drawbacks AI being a very powerful tool, it is embedded in television sets, which is a technology that is used by everyone ordinary families today, might have a downside to it. Since these televisions have high-end cameras attached, it can see who the viewers are watching. It would also be a good ground for Cybercriminals, who maybe record what is going on in houses and be a threat to privacy. It may also give a way for hackers to invade into personal lives and disrupt the privacy. How comfortably should users adapt to and accept this new generation of TVs is still debatable. Future Of AI televisions In the upcoming age of AI televisions, we could have TVs giving personalised recommendation based on age group. It might even see, or sense how engaged is the viewer in the currently playing program on TV and study his reactions. The main aim of the idea is to make all the devices communicate with each other. As delivery systems become more and more intelligent, every viewing experience could be all synchronised and AI enabled televisions would soon become as popular and regular as the HD and 4K TVs today.","excerpt":"Television sets have come a long way from being a simple box displaying grayscale images to the Smart 4K televisions today with hundreds of channels to browse through. As if this wasn’t overwhelming, the next phase of technology is now integrating itself with AI to power a more user-friendly, comfortable experience on television. AI techniques […]","categories":["Deep Tech"],"tags":["Google Assistant","LG","Sony"],"author_name":"Disha Misal","publish_date":"2018-10-14T11:43:55","publication_year":"2018","word_count":900,"keywords":["Go","programming_languages:R","AI","LG","programming_languages:Go","Ray","Aim","CLIP","Sony","Google Assistant","R","startup"],"extracted_tech_keywords":["AI","Aim","Ray","R","Go","CLIP","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/ai-powered-tvs-are-changing-the-way-we-perceive-entertainment\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10003207,"title":"Meet The MachineHack Champions Who Cracked The ‘Forest Cover Classification’ Hackathon","content":"MachineHack successfully concluded its twelfth instalment of the weekend hackathon series last Monday. The Forest Cover Classification hackathon was greatly welcomed by data science enthusiasts with over 240 registrations and active participation from close to 130 practitioners. Out of the 130 competitors, three topped our leaderboard. In this article, we will introduce you to the winners and describe the approach they took to solve the problem. #1| Devrup Banerjee Although Devrup learnt python just out of the sheer need to automate the routine work and gather data at scale, his real enthusiasm and passion for data science sprouted in his second year of MBA at Great Lakes Institute of Management, Gurgaon, while he was attending his marketing and retail analytics class. He realised that the real motivation behind learning all these algorithms was not about enhancing accuracy but to tell your client by how much you can promise to increase their bottom-line if they were to follow your exact given path. The subject changed his life. “My roommate, who was also equally inspired, and I used to have sleepless nights just going through the 25 lacs dataset given as a final project with our rickety computers to generate actionable insights. To better the bottom-line percentage, that’s what inspired me into analytics.“ – He said His team has won many competitions at MBA level, won at IIT Kanpur MRA tournament while finishing as runners up at IIM Kashipur’s case study on analytics. He is currently trying to deep dive into data science to better his analytical skills so that if someone gives him a dataset in future, he can be both a business analyst and a data scientist. Approach To Solving The Problem Devrup explains his approach briefly as follows: The first thing to notice about the problem was how similar the train and test datasets were. Every boxplot and every histogram almost threw an identical picture of the train and test sets. This indicated a stratified splitting of the train and test from the main dataset which might have resulted in said distribution. Thus there was no use of oversampling of the minority classes. Extensive feature engineering and a reverse clustering based on the test set helped achieve a certain level of accuracy and clipping of few values based on train distribution gave a major boost. CatBoost and LightGBM gave the best results. “MachineHack has been a huge source of inspiration and learning, along with Analytics India Magazine which keeps us up to date on the latest happening from around the world on analytics. They have established themselves as the domain leaders, and I won’t be surprised if they are soon known as the Indian Kaggle.”- he shared his opinion about MachineHack Get the complete code here. #2| Karan Juneja Karan is an Electronics and Telecommunication Engineer from PICT, Pune. His data science journey began out of his passion and curiosity for robotics. He has been acquiring new data science skills from free online resources as well as by participating in hackathons. Karan is a regular participant in MachineHack hackathons and has been in top 3 for several of the competitions. Approach To Solving The Problem Karan explains her approach briefly as follows: The hackathon provided a dataset that was openly available but still I tried to do a bit of feature engineering myself and with that got a score of around 0.255 with LightGBM. I also figured out that on this particular dataset the XGBoost algorithm could perform much better than LightGBM, but due to the limitations in computational power I chose LightGBM. The competition was tight with very good scores on the leaderboard and so I tried pseudo labelling and it worked out and got me a score of around 0.20. Get the complete code here. #3| V G Sravan Sravan is a third year Electronics and Communication Engineering student at IIT Kharagpur. Intrigued by the technology and its advancements, Sravan started his journey towards data science and machine learning with his focus set on Deep Learning. Amidst the pandemic, Sravan had been using his time in practising machine learning by solving online hackathons. “MachineHack is a very good place for beginners, it’s a very easy and competitive place to work on different ML models” – Sravan shared his opinion about MachineHack Approach To Solving The Problem Sravan explains his approach briefly as follows: This hackathon dataset was already very clean and uniformly distributed, so data cleaning and EDA was not required initially. Out of different models I trained,  ExtraTrees worked very well. This was expected as there was low cardinality of features and high training samples. Feature extraction and analysis played an important role for reducing the error. Get the complete code here. Click here to check out new hackathons","excerpt":"MachineHack successfully concluded its twelfth instalment of the weekend hackathon series last Monday. The Forest Cover Classification hackathon was greatly welcomed by data science enthusiasts with over 240 registrations and active participation from close to 130 practitioners. Out of the 130 competitors, three topped our leaderboard. In this article, we will introduce you to the […]","categories":["Deep Tech"],"tags":["Data Science","Hackathon Winners","latest technology in machine learning","Machinehack","machinehack winners"],"author_name":"Amal Nair","publish_date":"2020-07-26T13:00:00","publication_year":"2020","word_count":793,"keywords":["data science","machine learning","Machinehack","AI","ML","RAG","deep learning","latest technology in machine learning","analytics","XGBoost","CatBoost","LightGBM","machinehack winners","Data Science","Hackathon Winners"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","XGBoost","LightGBM","CatBoost","RAG"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/meet-the-machinehack-champions-who-cracked-the-forest-cover-classification-hackathon\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":48818,"title":"How ClickPost Is Using Logistics Intelligence Solutions To Reduce Return Rates &#038; Costs","content":"Reports suggest that the logistics sector in India is witnessing an exponential rise. Currently, the size of the logistics industry in India is around $160 billion and with the implementation of GST, it is expected to reach $215 billion by 2020–21 at 10.5% CAGR. With such a vision, Delhi-based startup called ClickPost is utilising logistics intelligence solutions to help its customers reduce return rates and costs, provide post-purchase experience and streamline their supply chain operations. Founded in 2015, ClickPost is India’s first and Asia’s second-largest integrated logistics platform with more than 100 logistics partners integrated via a single REST API. The company was founded by Naman Vijay and Prashant Gupta who are the alumni of the Indian Institute of Technology, Delhi and National Institute of Technology, Trichy respectively. Currently, the company is processing more than 8 million shipments per month and growing at a year-on-year rate of 600%. With the help of the logistics intelligence platform, the company has enabled the shippers to have a transparent and more effective way to manage their end-to-end logistics operations. Flagship Product ClickPost provides logistics integration solutions to its customers where the technology is sold as a SaaS product which has 4 levels of tech integration, starting with an ML-driven decision-making engine. The decision-making engine helps select the best logistics partners based on the business objective and the single API then enables integration with all logistics vendors in one place. The next layer helps the e-commerce firms track all their shipments in real-time while identifying and solving any exceptions, predict delays in shipment journeys and commit the correct delivery date to the end customers. And the last one enables its customers to provide easy returns. The company has also its failed delivery intelligence suite which leverages a large amount of data in the system to help reduce RTOs in e-commerce. This suite works by predicting the best way to get customer feedback on failed deliveries and communicate the same to logistics partners. Use of AI And ML The main focus of the company is to use Machine learning to predict the correct delivery date allows the users to set the correct expectations with the logistics partners as well as their consumers. According to Vijay, the company uses various ML algorithms like multiclass classification and regression models to solve complex problems like predicting the route of the shipment and identifying potential delays in the shipment journey. It has also built proprietary models on top of the data to predict the estimated delivery date. Tech Stack The core tech stack of ClickPost is built on Python. It uses Kafka and Spark to process huge amount of collected data and run analytics on top of that. In order to build, train and deploy the ML models, the company uses platforms like SageMaker. Further, the company has a reporting platform built internally over S3 which is able to deliver reports in realtime to the customers with current data. Roadmap Currently, ClickPost captures the data across the shipment lifecycle and helps customers with workflows to manage. While asking for the future roadmap, Vijay replied that the company is building more AI-driven features to help the clients predict exceptions in the supply chain as well as solving them intelligently. Further, the company is trying its level hard to expand its services to other geographies and build visibility there. In terms of expanding from within, the company always looks at hiring candidates who wish to solve challenges at scale while having the freedom that comes with working in a startup.","excerpt":"Reports suggest that the logistics sector in India is witnessing an exponential rise. Currently, the size of the logistics industry in India is around $160 billion and with the implementation of GST, it is expected to reach $215 billion by 2020–21 at 10.5% CAGR.  With such a vision, Delhi-based startup called ClickPost is utilising logistics […]","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Machine Learning","real-time integration to kafka","Startups"],"author_name":"Ambika Choudhury","publish_date":"2019-10-28T10:00:15","publication_year":"2019","word_count":589,"keywords":["Go","API","machine learning","AI","ML","Machine Learning","RAG","Python","analytics","real-time integration to kafka","Kafka","Startups","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","Kafka","Python","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-clickpost-is-using-logistics-intelligence-solutions-to-reduce-return-rates-costs\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":392,"title":"Maharashtra government join hands with ISRO to enable real-time tracking of Mangroves","content":"Mangroves, dubbed as the natural protector of the coastlines as they act as natural barrier against floods and natural disaster such as tsunami, decrease soil erosion and promote diverse aquatic life will now see ISRO playing a pivotal role in the conservation of the mangrove forests. According to news reports, the  Maharashtra government has formed a pact with ISRO to preserve and protect mangrove forests around the Konkan coastline. This move will allow government to take swift and necessary action, owing to satellite images of destruction obtained from ISRO’s sophisticated real-time satellite imagery systems. The pact marks the first venture for ISRO and Maharashtra, where ISRO will capture, track, and record the status of mangroves growing in Mumbai, and those along the parts of Konkan coastline. The budget for the project is pegged at INR 40 lakh. As a part of the deal, state forest departments will leverage ISRO’s open source software to track the mangroves, and the space agency will furnish real-time satellite maps for the same cause. Over the last few years, Maharashtra has been wrecked by floods and 2005 floods claimed 1,00 lives. Studies show mangroves being vital to a peninsular city such as Mumbai, and most of the recent flooding have been linked with depleting mangrove cover in the region. One of the glaring examples of concretization is Mumbai’s Bandra-Kurla complex, where mangrove forests were completely denuded to make way for commercial space. The destruction of these mangroves was banned by Bombay High court in 2005, however illegal construction still continues at large. There are over 15,000 hectares of mangroves in Maharashtra and 5,771 hectares is what makes up current Mumbai.","excerpt":"Mangroves, dubbed as the natural protector of the coastlines as they act as natural barrier against floods and natural disaster such as tsunami, decrease soil erosion and promote diverse aquatic life will now see ISRO playing a pivotal role in the conservation of the mangrove forests. According to news reports, the  Maharashtra government has formed […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2017-01-22T06:14:55","publication_year":"2017","word_count":275,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/maharashtra-government-join-hands-isro-enable-real-time-tracking-mangroves\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10009586,"title":"Facebook’s New Open Source Framework For Training Graph-Based ML Models","content":"Facebook recently open-sourced the graph transformer networks (GTN) framework for effectively training graph-based learning models. In this case, GTN will be used in automatic differentiation of weighted finite-state transducers (WFSTs), which is an expressive and powerful graph. With GTN, researchers can easily construct WFSTs, visualize them, and perform operations on them This framework enables the separation of graphs from operations on them that helps in exploring new structured loss functions and which in turn makes the encoding of prior knowledge on learning algorithms easier. Further, in a paper published by Awni Hannun, Vineel Pratap, Jacob Kahn & Wei-Ning Hsu of the Facebook AI Research, in this regard, proposed a convolutional WFST layer to be used in the interior of a deep neural network for mapping lower-level to higher-level representations. GTN is written in C++ and has bindings to Python. GTN can be used to express and design sequence-level loss functions. With this framework, Facebook aims to make experimentation with structures in learning algorithms much simpler. How Does GTN Function The use of WFST data structure is prevalent among speech recognition, natural language processing, and handwriting recognition applications. WFST, especially in the speech recognition systems, provides a common and natural representation for the hidden Markov models (HMM), context-dependency, grammar, pronunciation dictionaries, and weighted determinization algorithms to optimise time and space requirement. One of the most popular WFST-based products is the Kaldi toolkit for speech recognition which is trained to decode speeches. Kaldi heavily relies on OpenFST, which is an open-source WFST toolkit. To understand the importance of GTN framework for a WFST graph, we consider a general speech recogniser. A speech recogniser consists of an acoustic model that predicts the letters in the speech, its language model, and also identifies the word that may follow. These models are represented as WFSTs and are trained separately before combining to output the most likely transcription. It is, at this juncture, that the GTN library steps in to train the different models, which in turn provides better results. Before GTN, the use of the individual graphs at the training time was implicit, and the graph structure needed to be hard-coded in the software. Using GTN, however, researchers can use WFSTs dynamically at the training time, which in turn makes the whole system more efficient. It also helps researchers in constructing WFSTs and visualising them to perform functions on them. The gradients can be computed with respect to any participating graph by a simple call gtn.backward. GTN’s programming style is similar to other popular frameworks such as PyTorch, in terms of the autograd API, imperative style, and autograd implementation. The only difference lies in replacing the tensors with WFSTs. Wrapping Up GTN allows separating the graphs from the operations on the graph, which gives greater freedom to experiment with a larger base of structures learning algorithms. GTN is similar to PyTorch, which is an automatic differential framework for tensors. In the case of PyTorch, which is also called a ‘define-by-run framework’, the autograd package provides automatic differentiation for operations on tensors. Every iteration is different which allows dynamic modifications between epochs. Despite being similar on many counts, GTN provides an edge over PyTorch or any other tensor-based framework. GTN tools can easily experiment to develop a better algorithm. In the case of GTN, the graph structure is suited for encoding suggestive prior knowledge, which can teach the whole system and help in improving the data. Further, it is expected that in future, the structure of WFSTs, along with learning from the data, can make machine learning models lighter and more accurate. The use of WFST as a substitute for the tensor-based layer in a deep architecture is an interesting proposition. The paper (referenced above) concludes that WFSTs may be more effective than traditional layers for discovering significant representations from data.","excerpt":"Facebook recently open-sourced the graph transformer networks (GTN) framework for effectively training graph-based learning models. In this case, GTN will be used in automatic differentiation of weighted finite-state transducers (WFSTs), which is an expressive and powerful graph. With GTN, researchers can easily construct WFSTs, visualize them, and perform operations on them This framework enables the […]","categories":["AI Features"],"tags":["Facebook","power bi training","Speech Recognition"],"author_name":"Shraddha Goled","publish_date":"2020-10-13T11:02:22","publication_year":"2020","word_count":635,"keywords":["Go","machine learning","TPU","AI","neural network","PyTorch","Python","Aim","C++","Speech Recognition","Facebook","R","power bi training"],"extracted_tech_keywords":["AI","machine learning","neural network","Aim","PyTorch","TPU","Python","R","Go","C++"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/facebooks-new-open-source-framework-for-training-graph-based-ml-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10068507,"title":"Council Post: Data science and analytics for sustainable development","content":"According to the 1987 Brundtland Report, sustainable development is a development that meets the needs of the present without compromising the ability of future generations to meet their own needs. The three important areas of sustainability include environmental, social, and economic factors. The United Nations has developed 17 sustainable development goals (SDG) around these areas. The success of these goals will ensure prosperity and peace for the planet, society, and the economy for today and in the future. Let’s take a look at each of these goals below: Data science and analytics to the rescue Data science and analytics have a key role to play in achieving these sustainable development goals. They can be leveraged to enable sustainable development, particularly measuring impact, managing resources, tackling climate change, and more. Let’s look at each of these verticals – i.e. economic, social and environmental, in line with 17 goals of sustainable development and how data science and analytics can play a role in achieving sustainability with a few examples. Solving economic crisis Investors can use data science to help choose more sustainable companies.. They can move funds away from polluters and towards socially and environmentally responsible companies that adhere to best practices. Data science can help lower-income countries and emerging economies build data systems that contain information about households, occupations, wages, etc. This data can be further used to inform policy decisions, target investments and budgets optimally, and increase a country’s GDP. Building sustainable cities and communities also becomes equally important. Data science can help in planning sustainable cities with better infrastructure and amenities. Creating a positive social impact Satellite imagery can be used to collect data related to poverty and hunger-reduction, where the data includes images of densely populated human colonies, absence of water resources, barren land, etc. Otherwise, this data is often difficult to collect. Through data science and analytics, it is now possible to estimate crop yields based on weather conditions\/patterns and crop growth. With this, a vulnerable population group can be easily identified, and the government or authorities can take the necessary steps\/actions to mitigate risks effectively. In terms of achieving quality education goals, data science can improve education systems in the country, thereby helping create informed policies via evidence-based standards. Data science and analytics can address the data gap on violent encounters between citizens and the police or religious groups. Data science and analytics help law enforcement and concerned authorities capture and report this data, which includes information of people who have previously instigated such conflicts, people who have misbehaved with the authorities previously, names of convicted criminals, etc. These data can help authorities in maintaining law and order of the place, and stop potential threats from occurring. Most importantly, this can help build trust between communities and authorities. Addressing environmental concerns Data science can help create more sustainable energy. For example, in smart grids through dynamic energy management, data science can be used for production planning and forecasting. Under clean water and sanitation, image recognition can sort waste in recycling facilities or plants. In the case of responsible consumption and production, data science can be used to predict plastic waste available for recycling. It helps companies to understand available qualities better. Also, it helps them use recycled plastics for making new products effectively. However, the present-day market for recycled plastics suffers from a lack of transparency and information. So, satellite data and image recognition can identify riverside and coastal plastic hotspots and help in taking necessary actions or measures. Concluding remark The above examples show compelling ways to achieve some of the sustainable goals using data science and analytics. But, there is always scope for more. India has slipped three spots from last year’s 117 to 120th position on the sustainable development goals adopted as part of the 2030 plan by 192 UN member states in 2015. So it’s high time the government and industry bodies started focusing on ways to build data-driven measures and initiatives to tackle sustainability. It is really important for all the stakeholders to become aware of some of these economic, social, and environmental problems that India is facing and leverage data science and analytics to solve them effectively and effortlessly. In our upcoming articles, we will dive deeper into understanding these goals and suggest data science and machine learning techniques to solve them. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"Data science and analytics have a key role to play in achieving these sustainable development goals. They can be leveraged to enable sustainable development, particularly measuring impact, managing resources, tackling climate change, and more.","categories":["AI Features"],"tags":["environment","sustainability"],"author_name":"Anirban Nandi","publish_date":"2022-06-07T14:00:00","publication_year":"2022","word_count":766,"keywords":["data science","Go","machine learning","AI","sustainability","image recognition","RAG","Aim","analytics","environment","Rust","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Aim","RAG","image recognition","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-science-and-analytics-for-sustainable-development\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":37327,"title":"Reliance Jio Buys 87% Stake In AI-Based Startup Haptik For ₹700 Crore","content":"Reliance Jio is poised to create more ripples in the artificial intelligence sector, with the acquisition of 87% stake in chatbot startup Haptik in a ₹700 crore deal. The total transaction size, including primary capital investment, includes ₹230 crore as the consideration for the initial business transfer from Haptik to Jio. The Haptik team will continue to drive the growth of the business, including the enterprise platform as well as digital consumer assistants. On a fully diluted basis, RIL will hold about 87% stake in the company, with the rest with Haptik’s founders and employees. With the acquisition of Haptik, Jio plans to go big in the natural language processing domain. NLP has already been garnering a lot of interest in the corporate sector because of its innumerable applications ranging from speech assistants to sentimental analysis and recommendation engines. The telecom giant is leaving no stone unturned as it has already acquired Mumbai-based hyperlocal delivery platform, Grab. Aakrit Vaish, co-founder and CEO of Haptik, said about this development, “We started with the idea that Conversational interfaces will cause a paradigm shift in the way people get things done. Over the course, we have built various products across both consumer and enterprise businesses, with the backbone always being a full stack chat and voice-enabled AI technology platform. We truly believe now is the opportunity to serve the next billion users who come online, and who better to partner with than one of the world’s largest digital ecosystems in Jio. We look forward to using the capital and strategic opportunity to exponentially scale up the business across various product lines.” Haptik has been creating news in 2019, primarily because of the launch of a new chatbot by the Maharashtra Government. This chatbot, available on Aaple Sarkar RTS (Right to Services) provides easy, conversational access to information regarding 1,400 public services managed by the state government. With this, Haptik had effectively simplified the process of searching through the vast array of public services. Speaking about the reason behind buying 87% stake in Haptik, Akash Ambani, Director, Reliance Jio, said, “This strategic investment underlines our commitment to further boost the digital ecosystem and provide Indian users conversational AI-enabled devices with multilingual capabilities. We believe voice interactivity will be the primary mode of interaction for digital India. We are delighted to announce this partnership, and look forward to the highly experienced team of Haptik in realising this vision for greater connectivity and rich communication experiences to the billion-plus Indian consumers.”","excerpt":"Reliance Jio is poised to create more ripples in the artificial intelligence sector, with the acquisition of 87% stake in chatbot startup Haptik in a ₹700 crore deal. The total transaction size, including primary capital investment, includes ₹230 crore as the consideration for the initial business transfer from Haptik to Jio. The Haptik team will […]","categories":["AI News"],"tags":["Haptik","reliance jio"],"author_name":"Prajakta Hebbar","publish_date":"2019-04-05T05:14:57","publication_year":"2019","word_count":415,"keywords":["Go","API","artificial intelligence","startup","AI","Git","NLP","Ray","ViT","reliance jio","R","Haptik"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","Ray","R","Go","Git","API","ViT","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/reliance-jio-buys-ai-based-chatbot-startup-haptik-in-a-rs-700-crore-deal\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69702,"title":"How Beginners Can Get Real World Job Experience Of Data Science","content":"Data science is a lot like driving. To understand it comprehensively, one has to get down to the field to gain real-world experience and get their hands dirty with real professional challenges of data science. However, it gets challenging for beginners to gain real-world experience without getting an actual job in the industry. In fact, even landing on entry-level jobs can be daunting for beginner data scientists, as it comes with heavy requirements for professionals to have experience in data science. Thus, it becomes imperative for these beginners, recent graduates as well as professionals who are looking to translate into data science to find ways to gain some practical hands-on experience of the industry to stand out in the crowd. Here are a few ways beginners can get real-world job experience of data science before applying for an actual job: Also Read: Things Junior Data Scientists Need To Know Before Going For An Interview Building Personal Projects Building personal projects has always been key to creating a robust portfolio for data scientists. Not only it will help beginners to enhance their data science skills and confidence but also assist in exhibiting their practical knowledge and creativity to the potential employers. GitHub could be an excellent platform to showcase personal projects, and highlighting the same on the resume can improve the chances of beginners to land on a job. Handling a project involves tasks ranging from generating hypotheses, collecting, cleaning and analysing appropriate data, to creating predictive models and sharing data with organisation leaders. Thus, working on projects can give a comprehensive understanding of various topics to beginner data scientists. To enhance their skills and gain real-world experience, data scientists should focus on scripting clean codes with accurate data visualisation for making insights understandable to stakeholders. Beginner data scientists can also try their hand on projects based on computer vision, facial recognition technology, machine learning and GANs, to name a few. Also Read: How Does Project Management In Data Science Look Like? Contributing to Open-Source Projects Another way for beginners to gain real-world data science experience is by contributing to open-source projects. Participating in the open-source community can be overwhelming at the start, as data scientists would be required to provide code or solutions for a running project. However, it will indeed enhance the coding and technical skills of the individual and give them the experience to work on projects along with a team which would require constant communication. By contributing to open source projects, data scientists will be able to give back to society by providing code, which can benefit millions in building their projects. Open-source projects also require working with many data science libraries, like Pandas, NumPy, Scikit-learn and open-source distributed version control system, Git. Thus, working on these projects will make beginners familiarise with these aspects of data science. Contributing to open-source projects will also make amateurs a part of a community where they can network with individuals with similar interests, which will again help in enhancing the real-world experience. freeCodeCamp is one such community that helps nonprofits to code and build projects. Also Read: How Open-Source Projects Make Money? Start Small Although getting a full-fledged job can be difficult for beginner data scientists, freelancing or starting an internship could be an excellent way for them to gain some real-world experience. In fact, the COVID-19 pandemic has drastically revamped the gig economy, and many companies are now looking to hire freelancers as well as interns to avail data science skills for specific projects. Therefore, freelancing will not only help them enhance their technical skills but will improve their communication skills by engaging with different teams of the organisation. Freelancing and internship will allow these beginners to work on real-world business problems that companies are facing and will urge them to come up with a solution that is practical for the real world. This, in turn, will enhance their experience and make them more industry-ready. Freelancers even get a chance to accommodate many projects at once, which will make them work on diverse datasets, improving their multitasking skills. However sometimes, it gets challenging to market oneself enough to get a freelancing job, and therefore there are prominent platforms to search for freelance data science jobs are — Upwork, Toptal, Data Science Central to name a few. Freelancing and internship not only allows beginners and amateurs to experiment with different data science roles but also puts them in the limelight among potential employers. Also Read: How The Freelance Boom Will Impact The Position Of In House Data Scientists Participate in Hackathons & Competitions Participating in hackathons is another way of gaining real-world experience for data scientists, as these competitions force the participants to build applications that would be running in the real world. Thus, it will help beginners to learn the practical skills required for the actual job as well as will help them connect to a network of professionals with similar interests and experts of the industry. These coding competitions and hackathons require participants to turn hypotheses into action which will provide immense benefits if mentioned on the resume. Alongside these hackathons and competitions runs for hours and hours at a go, which forces participants to work under pressure, thus teaching them to handle extreme stress. This, in turn, gives them an experience of the real-world scenario, where data scientists work on a deadline to solve their business problems. Apart from building models, data scientists also are required to bring in business value. Thus, these competitions would help them understand the criticality of their work for business outcomes. Hackathons and competitions also act as a testimony for these professionals’ skills and knowledge, and this makes it easier for companies to filter out the best. One such platform is the MachineHack by Analytics India Magazine, which hosts some of the exciting business problems for participants to solve using ML and data science techniques. Also Read: How Much Is Kaggle Relevant For Real-Life Data Science Get a Mentorship Lasts but not the least, getting a mentorship can prove to be beneficial for beginner data scientists to gain hands-on experience of the real world. A comprehensive mentorship can provide one-on-one sessions, which will allow the data scientists to pick the brains of experts from the industry. Also, a good mentor will give all sorts of real-world guidance that will be required for a beginner to thrive in this competitive market. Having a mentor is an open door into the data science community, which will provide an opportunity to network with data science professionals and potential employers from the industry. Apart from building skill sets, the mentor’s expertise will also help in gaining relevant feedback on the growth and development of the individual. This, in turn, makes them industry ready in no time. These mentors can also help them find jobs, and gain work experience, which will help these beginners to speed up their career graph. These industry experts are also equipped with real-world tools and techniques in AI and data science, which will expose mentees to relevant skill sets required to survive the job. Join AIM Mentoring Circle.","excerpt":"Data science is a lot like driving. To understand it comprehensively, one has to get down to the field to gain real-world experience and get their hands dirty with real professional challenges of data science. However, it gets challenging for beginners to gain real-world experience without getting an actual job in the industry.  In fact, […]","categories":["AI Hirings"],"tags":["Data analyst jobs","Data Science","Data Science Jobs","Data Scientist","datascience","mentorship","project topics for computer science","whats data roles involve machine learning"],"author_name":"Sejuti Das","publish_date":"2020-07-14T15:00:00","publication_year":"2020","word_count":1186,"keywords":["project topics for computer science","data science","scikit-learn","NumPy","machine learning","AI","whats data roles involve machine learning","datascience","ML","Data analyst jobs","Data Science Jobs","computer vision","Aim","analytics","mentorship","Data Science","Data Scientist","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","data science","analytics","Aim","scikit-learn","Pandas","NumPy"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/how-beginners-can-get-real-world-job-experience-of-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10102651,"title":"How is Angular Not Dead Yet?","content":"Angular has been completely revamped. “Welcome to Angular’s renaissance,” said Angular in a post on X. As promised by the developers, the framework has come up with v17 with a lot of new upgrades in its syntax and template features. Angular’s page now includes dark mode, in-depth guides, search-ability, and a lot of tutorials, and most importantly, the Playground, which allows users to write user templates to start with the latest features. The highlight of the revamp is Angular.dev, which is a new future home for Angular development. This includes new tutorials, updated documentation, and guidance for latest Angular features. And the Playground is where you can explore all these concepts. The release blog highlights the company’s dedication towards open source development, and improvement for the future v18 release of Angular, which will focus on stability. Moreover, Emma Twersky, senior developer relations at Angular, highlighted that the company has also reformatted its API and CLI references to look like code in the editor, for easier reference. If anything, the updates to the site highlight what Angular framework is now capable of. Welcome to Angular’s renaissance.https:\/\/t.co\/8DxJPZaTd5 pic.twitter.com\/yfvylyG6Bs— Angular (@angular) November 6, 2023 Too little too late, every time While many developers who have been using Angular are extremely excited with the “new feel” of Angular, others are calling it just dead, or merely a revamp. Some even jokingly questioned if it was acquired by Adobe, given the new design of the website and the logo looks very similar to Adobe’s. Is Angular dead? Well, not according to its die-hard fans, who are still holding onto the hope that it will make a glorious comeback — which can still be expected. In the world of web development, where newer, shinier frameworks, such as React, are popping up faster than you can say “Angular”, it’s easy to wonder if the framework would indeed become a relic of the past. In 2022, we saw the release of Angular v15, which, according to early reviews, was more refined, stable, supportable, and its last-ditch effort to survive. Sure, it may not have set the internet on fire with excitement, but it’s still being used by many. Angular 15 was dearly loved by a lot of developers. But is that really true? In May 2023, Angular launched v16 with what it called the “biggest release since the initial rollout of Angular,” but the same could not be said completely about the developer community. Angular devs in my replies. pic.twitter.com\/hBWEOS65Df— Killed by Google (@killedbygoogle) November 7, 2023 According to the Stack Overflow Developer Survey 2023, Angular has seen better days. It’s down to 18.7% usage, a far cry from its glory days of 30.7% in 2021, and 22.9% in 2022. React, Vue, and even jQuery have snatched its crown. Not to mention, Angular’s satisfaction rate of 58.6% pales in comparison to React’s 74.5% and Vue’s 66.9%. Why the fall from grace? Well, for starters, fresher and lighter frameworks like React and Vue, for instance, have emerged with simpler syntax, faster rendering, better SEO support, and smaller bundle sizes. Still afloat Google has been known for killing a lot of its products. In 2019, it killed AngularJS, but in turn offered its developers Angular, the framework with more than just JavaScript in its focus. Regardless, Angular has gone through more versions than any other framework. Though this might seem like a good thing, from AngularJS to v2 to v8, to now v17, with each new version, developers must rewrite or migrate their code, making them feel like they’re on a never-ending tech rollercoaster. And, let’s not forget the lack of support and documentation for the older versions, all this while. Though with the beta update, Angular v17 is still garnering love from the developers. So, Angular may not be your trendiest framework in town, but it’s still alive and kicking. Interestingly, Google has also been using React, along with Angular for a lot of its framework. But just like with every update and every year, new blogs keep popping up to check if Angular is dead or not. The Angular team knows it, and thus has promised a stable release of v18 soon, as they know that is what the developers have been craving for all this while.","excerpt":"Angular has seen better days. It’s users have nearly halved since 2021.","categories":["AI Features"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2023-11-07T12:30:00","publication_year":"2023","word_count":709,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Java","programming_languages:Go","JavaScript","R","Java","programming_languages:JavaScript"],"extracted_tech_keywords":["AI","R","JavaScript","Go","Java","API","programming_languages:R","programming_languages:JavaScript","programming_languages:Java","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-is-angular-not-dead-yet\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10084827,"title":"Wolfram Alpha + ChatGPT Might Be The Chatbot Scientists Need","content":"Ever since its genesis in 1988, the Wolfram language has been the go-to language to solve complex scientific problems. Wolfram Alpha, an answering machine built on this language, uses natural language processing, the world’s largest repository of computable knowledge, and a custom-made symbolic programming language to provide answers to mathematical questions. These questions can range from high-level calculus to the amount of calories in a given dish, and Wolfram Alpha will provide an answer while showing the steps to the solution. For all its computational intelligence, the website sometimes struggles with identifying queries in natural language. With modern advances in natural language processing and chatbots, this forerunner of modern AI might change the landscape for NLP-powered problem solving. Bringing together Wolfram Alpha and ChatGPT In a blog post published recently, Stephen Wolfram, the founder and CEO of Wolfram Research, explored the idea of combining the capabilities of Wolfram Alpha and ChatGPT. Wolfram demonstrated the tendency of ChatGPT to give factually incorrect answers which sounded like they might be accurate, highlighting the feature of the chatbot to correct itself when prompted. He also showed off the capabilities of Wolfram Alpha and how it can be used to ‘inject’ data points into ChatGPT, which the bot then accepted as the correct response. To understand why these two vastly-different applications can work so well together, we must first delve into the approach that they each take to solving a problem. ChatGPT is trained on GPT 3.5, a large language model that has a dataset containing 175 billion parameters, using which it has learned to respond to prompts in natural language with coherent responses. This means that the chatbot has learned the pattern of human-like speech along with the capability of translating a query in a human language to one in a machine-understandable language. On the other hand, Wolfram Alpha is built on the Wolfram language, a symbolic programming language that is focused on expressing complex ideas in a computational form. This language was made expressly to solve complex algebraic problems, with its latest iteration being able to take on higher-level calculus tasks like differential equations and matrix manipulation. When looking at the approach that the creators of Wolfram and ChatGPT have taken, the benefits of bringing them together are obvious. ChatGPT is unbeatable at parsing natural language and making it computer-readable while Wolfram is excellent at solving complex mathematical problems by breaking it down into the Wolfram language’s symbolic expressions. This union might even make it the accurate chatbot that the scientific community doesn’t know they need. The chatbot scientists need The scientific community has largely ignored chatbots derived from large language models, as seen by their negative response to Meta’s ‘Galactica’ model. This short-lived LLM was launched with the grand goal of organising all scientific knowledge and making it accessible through a chatbot. Unfortunately, it had a propensity to hallucinate information, even though it was trained on close to 50 million scientific papers, leading to it being shut down in a matter of two days. While the bug of information hallucination is yet to be solved even for ChatGPT and the underlying GPT LLM, OpenAI has conducted research on reducing the amount of misinformation that the bot gives out. Owing to this, the bot rejects queries that it believes it does not have information to answer along with blocking answers about sensitive topics, like hate speech and self harm. One place where this system falls apart is when ChatGPT is asked objectively factual questions, where it confidently spews misinformed answers. There are two reasons for this, first being ChatGPT’s dataset. The dataset for GPT 3.5 consists of information scraped from the Internet, leading to many discrepancies when it comes to specific information like the distance between cities, population statistics, and many more. In addition to a flawed dataset, the bot also faces difficulty in mathematical calculations, as it is not trained to understand them but simply try to ‘solve’ them using natural language. Both these shortcomings can be addressed through Wolfram Alpha. If combined with the Wolfram language’s computational prowess, the problem of not understanding mathematical problems can be solved. Add on Wolfram’s comprehensive knowledge-base and you have a search killer on your hands. In addition to being objectively accurate, it can also understand exactly what is being asked, covering up the shortcomings of both ChatGPT and Wolfram Alpha. Stephen Wolfram, the creator of the language, said: “There are all sorts of exciting possibilities, suddenly opened up by the unexpected success of ChatGPT. But for now there’s the immediate opportunity of giving ChatGPT computational knowledge superpowers through Wolfram|Alpha. So it can not just produce “plausible human-like output”, but output that leverages the whole tower of computation and knowledge that’s encapsulated in Wolfram|Alpha and the Wolfram Language.” This miracle combination might be the chatbot that the scientific community needs. With ChatGPT’s measures against giving misinformed responses and Wolfram’s mathematical might, we might see a chatbot that provides actual accurate information. The Wolfram language is also a mainstay in the scientific community through ‘Mathematica’, further proving the point that these kinds of solutions do have a market. Instead of learning complex software like Mathematica, scientists can ask queries in natural language and get an accurate answer that they can trust. The future of access to Information will be decided by the success of chatbots like ChatGPT, but the bot we’re seeing now is just the first step. There are still decades of iterative improvement waiting in the wings and AI researchers stand to benefit from integrating existing solutions into their cutting-edge algorithms to set the paradigm for the next generation of problem-solving algorithms.","excerpt":"The world’s first answer engine and the world’s latest chatbot might be a match made in heaven.","categories":["AI Features"],"tags":["chat gpt app","chat gpt login","chat gpt website","GPT-3","OpenAI"],"author_name":"Anirudh VK","publish_date":"2023-01-12T14:00:00","publication_year":"2023","word_count":939,"keywords":["GPT-3","Go","chat gpt login","chat gpt app","ChatGPT","TPU","OpenAI","AI","chatbots","RAG","NLP","chat gpt website","Rust","R"],"extracted_tech_keywords":["AI","NLP","ChatGPT","OpenAI","RAG","chatbots","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/wolfram-alpha-chatgpt-might-be-the-chatbot-scientists-need\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10005785,"title":"AI Technologies That Featured In Latest Gartner Hype Cycle","content":"Last week, Gartner released a unique hype cycle; for emerging technologies. It includes technologies that can completely change the direction of human civilization. “This Hype Cycle highlights technologies that will significantly affect business, society and people over the next five to 10 years,” said Brian Burke, Research VP, Gartner. From algorithmic trust to advanced AI, the Gartner’s hype cycle features many new technologies. In this article, we will focus exclusively on the AI segment of this survey. Source: Gartner 1| Embedded AI Stage: Peak of inflated expectations Time required to plateau: 2 to 5 years Thanks to the efforts of top companies to place supercomputers in the pockets, Edge computing systems have garnered significant attention. As we approach an era of IoT, 5G and portable medical devices, it is crucial to facilitate developers to develop and deploy edge applications quickly. For example, with the NVIDIA Jetson AGX Xavier developer kit, as shown above, one can easily create and deploy end-to-end AI robotics applications for manufacturing, retail, smart cities, and more. Whereas, Google’s Coral toolkit can be leveraged to bring machine learning to edge. The safe, secure and real-time output is the theme of the modern world. And, edge devices offer just that. Check top edge computing products here. 2| Generative AI Stage: Innovation Trigger Time required to plateau: 2 to 5 years AI can now paint auction-worthy art, generate songs and even create faces of people who never existed. All this thanks to Generative Adversarial Networks (GANs). These networks are considered to be one of the critical turning points in the history of ML. In the Gartner Hype Cycle, Generative AI featured in the ‘innovation trigger’ segment of the graph. The proliferation of generative networks has had many unwanted results. Malicious online players can now generate disinformation in the form of images and videos that can fool many. Generative networks are all fun and games until they start interfering with matters of national interest. For good or for worse, these networks are here to stay for a while now. 3| Responsible And Explainable AI Stage: Innovation Trigger & Peak of inflated expectations resp. Time required to plateau: 5 to 10 years Machine learning algorithms are infamous for their black-box nature. With increasing business applications such as autonomous driving and medical diagnosis, there is a growing demand from the stakeholders to know what is at stake. Apart from the debugging and auditing of the models, data scientists need to look out for meeting data privacy standards in the context of explainability. Be it medical diagnosis or credit card risk estimation, the amount of personal information that is processed can be very sensitive, and this is being addressed in the organisations which are serious about the implementation of explainability. From inductive biases to GDPR compliance, AI enterprises have many things to take care. To make the process of inculcating responsible AI systems more viable, tools such as Model cards, AI 360 and many others have been released by companies like Google, IBM etc. Responsible AI too, like Generative AI featured in the first segment bordering on ‘peak of inflated expectations’. Whereas, explainable AI featured on the descending curve of inflated expectations. Though responsible and explainable AI practices overlap in many ways; the end goal is to enable transparency in AI-based decision making. 4| Self Supervised Learning Stage: Innovation Trigger Time required to plateau: 5 to 10 years Most of the AI R&D endeavours converge at one thing — making AGI possible. Equipping machine learning algorithms with human-like reasoning is tricky. ML models are good as the data it is fed with. In areas such as medical imaging, the availability of pre-trained models are almost negligible. Less data, more accurate results; this is where self-supervised learning comes into the picture. Speaking at (ICLR) 2020, Turing awardee and Facebook’s chief AI scientist, Yann Lecun said that supervised learning systems would play a diminishing role as self-supervised learning algorithms come into wider use. It is no surprise that self-supervised learning is placed in the early stages of ‘innovation trigger’ in the Hype Cycle. 5| AI Augmented Development Stage: Peak of inflated expectations Time required to plateau: 5 to 10 years In a report published last year, referring to AI augmented development, Gartner said that the application leaders should embrace AI-augmented development now, or risk falling further behind digital leaders. AI has now found its way into the software development lifecycle. They are assisting organisations in design, development and deployment of their software products. According to Deloitte, startups offering AI-powered software development tools raised US$704 million over the 12 months ending September 2019. AI has the potential to take care of the mundane debugging tasks through automation. Today, some tools offer low code and no code services. Some AI-based APIs even search the code as one types it. Going forward, the manual assessment might not match the rate at which the industry is innovating. So, it is safe to say that AI-augmented development will witness a growth in demand. 6| Composite AI Stage: Innovation Trigger Time required to plateau: 2 to 5 years As the name indicates, Composite AI aggregates multiple AI systems trained individually with small data sets instead of pooling data. The training of AI systems typically requires a large amount of aggregated data (‘big data’). As discussed above, it can be challenging to find enough data in niche areas. As shown in this experiment, a composite AI system can be constructed based on multiple neural networks from individually trained AI cores, each using a small data set. These systems outperform individual AI cores in the classification of molecular images. According to the study, this approach can be applied to AI development across multiple institutions without necessitating data sharing or pooling to create an extensive training data set. 7| Adaptive ML Stage: Peak of inflated expectations Time required to plateau: 2 to 5 years Real world is ever-changing. How can we expect traditionML algorithms to function reliably with inconsistent training data. This is why adaptive ML came into existence. According to Alan Turing Institute, Adaptive real-time machine learning incorporated efficient reinforcement learning, online learning (dealing with continuous sequences of real-time data), and adaptive learning from a small sample size. The techniques border on real-time meta-learning algorithms that can be utilised to achieve continuous learning, predicting and controlling when functioning in a changing environment. Adaptive ML applications include rainfall prediction for urban infrastructure, adaptive yield prediction based on real-time crop vigour analysis in precision farming and more. 8| 2 Way BMI Stage: Innovation Trigger Time required to plateau:  5 to 10 years A brain–machine interface (BMI) is a device that translates neuronal information into commands capable of controlling external software or hardware such as a computer or robotic arm. BMIs are often used as assisted living devices for individuals with sensory impairments. Companies like Neuralink are developing implantable brain-machine interfaces. Building BMI devices requires a combined effort of experts from medical, material, ethical and various other fields.  This in turn will lead to new job opportunities and technologies. The Gartner hype cycle forecasts BMI tech to hit plateau within 10 years. If this is true then we will be closer to deciphering the brain. 9| Digital Twin Of A Person Stage: Innovation Trigger Time required to plateau: 5 to 10 years Digital twin technologies enable authentic digital copy of physical entities that can be represented both in the physical as well as digital space. Digital twin innovation will have great implications in the AR\/VR industry and even in brain-machine interface applications. 10| AI Assisted Design Stage: Innovation Trigger Time required to plateau: 5 to 10 years AI is now being extensively used in traditional software like Autodesk and many more. AI high dimensional features can be leveraged to conduct multiple design studies. Algorithms have also been used to generate architectural designs. The designs are not restricted to the physical world. Deep learning can now be used to design web pages, creating brand logos and many more. For example, Volkswagen Microbus uses components such as brackets reshaped in generative design. AI driven generative design has proven to cut time delays by a significant amount. 11| Differential Privacy Stage: Peak of inflated expectations Time required to plateau: 2 to 5 years Differential privacy deals with collecting data while simultaneously ensuring anonymity of the individual. Differential privacy is a high-assurance, analytic means of ensuring that use cases like this are addressed in a privacy-preserving manner. Companies like Google even rolled out tools like differential privacy library. Whereas, Apple uses differential privacy techniques to take feedback from their users in a safe way. Privacy is a cornerstone of data sharing and differential privacy provides a definitive guide to navigate through the digital realms. Apart from these technologies mentioned above, social distance technologies and other critical domains directly or indirectly implement AI in many ways. AI technologies are poised to discover new avenues of software development, medical diagnosis, transportation; in short, will transform the way we live and this survey by Gartner proves it. Know more about the Gartner Hype Cycle here.","excerpt":"Last week, Gartner released a unique hype cycle; for emerging technologies. It includes technologies that can completely change the direction of human civilization. “This Hype Cycle highlights technologies that will significantly affect business, society and people over the next five to 10 years,” said Brian Burke, Research VP, Gartner.  From algorithmic trust to advanced AI, […]","categories":["Deep Tech"],"tags":["AI Technology","augmented intelligence for smart industry","big data in auditing","big data video games","digital twins","Gartner hype cycle","latest machine learning innovation"],"author_name":"Ram Sagar","publish_date":"2020-08-29T11:00:07","publication_year":"2020","word_count":1512,"keywords":["machine learning","TPU","AI Technology","AI","digital twins","augmented intelligence for smart industry","latest machine learning innovation","ML","neural network","RAG","deep learning","big data in auditing","generative AI","differential privacy","edge computing","big data video games","Gartner hype cycle"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","generative AI","differential privacy","RAG","edge computing","TPU"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/gartner-hype-cycle-2020-artificial-intelligence\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":13516,"title":"Amazon Web Services makes an AI push, plans to add AI to cloud","content":"Image source: Amazon Web Services AWS, an industry leader in cloud computing has plans to add Artificial Intelligence capabilities to its cloud offerings. According to news reports, AWS India is looking to add more teeth with an AI-first approach by adding services such as image analysis, visual search, speech recognition among others on its infrastructure for developers and enterprises to build products. The cloud of choice for startups and enterprises across the globe had made a significant AI push in the past by adding security to its cloud mix. Earlier last year at the annual conference, Amazon AI added a bunch of AI capabilities to its roster. The undisputed cloud leader announced they were bringing natural language understanding (NLU), automatic speech recognition (ASR), visual search and image recognition, text-to-speech (TTS), and machine learning (ML) technologies to every developer. Amazon Lex is to build chatbots with text and voice; while Amazon Rekognition lends deep learning-based image recognition. Amazon Polly turns text into speech and Amazon Machine Learning allows developers to build smart ML applications. What’s new is Amazon’s investment in Harvest.ai, a San Diego startup that was quietly acquired by the all-pervasive e-commerce giant earlier this year. What the startup brings to the table is sophisticated AI-based algorithms that can detect and stop data breaches. According to news reports, the startup was acquired for a whopping $20 million and all its 12 employees were relocating to Amazon headquarters in Seattle. The startup’s flagship product titled MACIE, monitors the enterprise’s network in real-time to detect any unauthorized and suspicious user accessing unauthorized documents. The target market for the startup was major organizations who have moved to cloud based platform such as AWS and that’s what triggered the buy-out. According to a Gartner report, there is an increased spending in cyber security with companies worldwide spending topping $170 billion by 2020. The charge is being led by big banks such as JP Morgan Chase","excerpt":"AWS, an industry leader in cloud computing has plans to add Artificial Intelligence capabilities to its cloud offerings. According to news reports, AWS India is looking to add more teeth with an AI-first approach by adding services such as image analysis, visual search, speech recognition among others on its infrastructure for developers and enterprises to […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-03-15T09:42:55","publication_year":"2017","word_count":321,"keywords":["machine learning","artificial intelligence","AWS","AI","chatbots","cloud computing","ML","image recognition","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","chatbots","image recognition","cloud computing","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amazon-web-services-makes-ai-push-plans-add-ai-cloud\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":53281,"title":"With Snap’s AI Factory Acquisition, Deepfake GIFs Will Go Mainstream","content":"Continuing its acquisition spree, Snapchat’s parent company Snap has now bought AI Factory, a Ukrainian computer vision and AI-based startup that earlier collaborated with them to build Snapchat’s new Cameos animated selfie-based video feature. According to reports, it would cost Snap $166 million to acquire AI Factory. It had earlier spent $150 million on another Ukraine based startup Looksery in 2015 to power its augmented reality lenses— and shook up the lens filters game for all social video and photo apps. Lenses became a mega success for Snap where it was reported that 70% of its daily active users play with them, which not just brings in new users, but also increases user retention and revenues by way of sponsorships and the purchase of the devices by users. Founded in 2018, AI Factory has been developing innovative computer vision and augmented reality products with a focus on image and video technologies, analysis and processing. According to AI Factory, its mission is to let users create “the real Hollywood with your smartphone.” The interesting thing with AI Factory is that it is also co-founded by one of Looksery’s founders- Victor Shaburov. Victor has been Snap’s director of engineering but left in May 2018 to found AI Factory. While AI Factory will be a subsidiary of Snap, it will also contribute its technology to make more innovative ways for people and advertisers to use the Snapchat in the future. Why Did Snap Acquire AI Factory? Snapchat is readying to roll out a novel feature that utilises selfies to replace the faces of users in videos which they can then share. It’s basically is a simplified manner to create deep-fakes, but for animated GIFs. Snapchat Cameos are also an alternative to Bitmoji for rapidly displaying an emotion, reaction, or other funny scenarios in Snapchat messages. With Cameos, users can first take a selfie and then determine what kind of cartoonish body or character they want their face to be embedded on, and it could be any fun character you like. Cameos then integrate within the Bitmoji button in the Snapchat keyboard where it can be accessed instantly for making the supposed deep-fake GIFs, where users facial images can be applied to actors\/characters’ heads. Snapchat has demoed a total of 150 short looping video clips with sound that you can pick from, with new additions each week. Users can use the sound filter to make their video animations entertaining and creative. With an ever-increasing appetite for new features, Snapchat’s Cameos could help in keeping the messaging interesting, which is important for the company’s strategy. The social media giant has often tried to come up with unique features and for that, it has made huge acquisitions. Cameos have been compared to Chinese social app Zao, which can make videos that replace the faces of celebrities in scenes from popular movies, shows and music videos with a selfie uploaded by the user. So, instead of cartoon-type deepfakes, Zao had real deepfake technology — a social media feature which has a large segment of users in the West but has not been accepted due to its malicious nature. With GIFs, it appears that Snap wants to take out the dark dystopian aspect of deepfakes and replace human characters with cartoons to make the technology work on its platform. Cameos certainly appears a more fun variant on deepfake technology even though it would have the same privacy concerns related to facial data. Unlike Zao, Snapchat’s version wouldn’t lead to the creation of malicious content like fake news videos. Also, Snapchat certainly needed a deepfake kind of feature, given one of its closest competitors — TikTok is about to release its own. Other Acquisitions To Bring Novel Features On The Social Media Platform Snap has also spent $50 million to buy out QR code startup Scan.me in order to utilise its technology to make Snapcodes, which are personalised profile QR codes for users to scan with the Snapchat camera to begin following a particular user within the application — a feature which has tremendously taken off. Snap has also agreed to acquire Bitstrips, the company behind the famous emoji-creation app Bitmoji in a reported deal of approximately $100 million.","excerpt":"Continuing its acquisition spree, Snapchat’s parent company Snap has now bought AI Factory, a Ukrainian computer vision and AI-based startup that earlier collaborated with them to build Snapchat’s new Cameos animated selfie-based video feature. According to reports, it would cost Snap $166 million to acquire AI Factory. It had earlier spent $150 million on another […]","categories":["AI Features"],"tags":["deep fake","Mergers and Acquisitions","snapchat","social media AI"],"author_name":"Vishal Chawla","publish_date":"2020-01-07T12:00:00","publication_year":"2020","word_count":700,"keywords":["API","programming_languages:R","AI","deepfakes","computer vision","social media AI","deep fake","snapchat","CLIP","Mergers and Acquisitions","R","ai_applications:computer vision","startup"],"extracted_tech_keywords":["AI","computer vision","R","API","CLIP","startup","deepfakes","programming_languages:R","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/with-snaps-ai-factory-acquisition-deepfake-gifs-will-go-mainstream\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168114,"title":"Deloitte India Partners With SAP, AWS to Simplify AI-Powered Migration Process","content":"Deloitte India has partnered with SAP and Amazon Web Services (AWS) to help SAP’s existing ERP customers accelerate their transition to the AI-powered RISE with SAP S\/4HANA Cloud. This initiative, called the Near Zero Cost Migration programme, will aim to fast-track and simplify migration while providing companies access to SAP Cloud ERP, a sustainable and high-performance infrastructure. Anand Rajagopalan, partner and SAP offerings leader at Deloitte South Asia, believes the cloud ERP will optimise operations, improve scalability, agility, and innovation potential. The programme will enable organisations achieve a shorter time to value and provide a smooth transition while cutting costs and combining resources from the three companies. The Near Zero Cost Migration initiative seeks to streamline the migration process. It includes a suite of AI-driven tools and accelerators such as Migratex, which automates key aspects of the migration process, including assessment, SAP note analysis, CVI deduplication, and testing. Nitish Agrawal, chief partner officer of SAP Indian Subcontinent, said, “We are enabling customers to accelerate their digital journeys by harnessing the full potential of cloud technology to innovate faster, operate more efficiently and achieve their long-term strategic goals. Together, we are transforming IT landscapes, making organisations future-proof their operations to create lasting value in a rapidly evolving digital economy.” Deloitte’s programme is an “exclusive offer” only available in India until July 2025. In a press release, the company said that businesses participating in this initiative will gain access to “structured guidance, proven recommendations, security principles and enhanced support at every step to ensure seamless transformation”. Deloitte India has previously collaborated with AWS to deepen its generative AI-driven innovations for Indian businesses. AIM reported that Deloitte India expanded its Global AI Simulation Centre of Excellence (COE) last month, as part of its broader $3 billion investment in generative AI through FY30. The new COE could improve strategic decision-making through advanced visualisations, digital twins, multi-agent systems, and scenario modelling. According to AIM, Deloitte has invested between $5 million and $10 million in this initiative. In March this year, the company announced a strategic alliance with Zoho, marking another milestone in bringing digital transformation to mid-market and enterprise firms in Indian businesses.","excerpt":"The three companies have launched a Near Zero Cost Migration programme for Indian businesses to integrate AI-powered processes into migration systems.","categories":["AI News"],"tags":["AI-Powered Migration","Deloitte","EPR"],"author_name":"Smruthi Nadig","publish_date":"2025-04-16T16:48:55","publication_year":"2025","word_count":358,"keywords":["Go","Deloitte","AWS","AI","R","ML","Scala","Git","Aim","multi-agent systems","generative AI","EPR","AI-Powered Migration"],"extracted_tech_keywords":["AI","ML","generative AI","multi-agent systems","Aim","AWS","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deloitte-india-partners-with-sap-aws-to-simplify-ai-powered-migration-process\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":65563,"title":"Altran Announces ML-Based Tool To Detect Bugs In Source Code","content":"Altran, in association with Microsoft, has launched a new artificial intelligence-based tool that has been designed to predict and identify the source code files that are carrying a higher risk of having a bug. According to Microsoft’s webpage — “Developers are presented with explanations and factors used in making the specific prediction.” Using machine learning onto historical data allows the AI-based tool — Code Defect AI – to identify the areas of the source code that are potentially bugged. The tool further suggests a set of tests in order to diagnose and fix the anomalies, which in turn will result in higher-quality software and faster development times. According to Microsoft, bugs are a fact of life in software development. Bugs need to be identified and rectified at the earliest, as more time it takes to defect a flaw in the development lifecycle, the higher the cost becomes of fixing the bug. Traditionally, this bug-deployment-analysis-fix process is time-consuming and costly, and therefore required a tool that can speed up the process. Alongside, if a defect is found after deployment, it impacts the customers profoundly, and then developers end up spending more time to fix the bug. The new AI-based tool — Code Defect AI, on the other hand, allows the earlier discovery of defects and anomalies, which in turn minimises the cost of fixing them and speeding the development cycle. According to Walid Negm, Group Chief Innovation Officer at Altran, software developers are always working under continuous pressure to release code fast without compromising on quality. However, the reality is that the development process of software requires more than the automation of assembly and delivery activities. It requires heavily working on algorithms that can help make strategic judgments ‒ especially as code gets more complicated. Code Defect AI does precisely that.” Code Defect AI relies on various machine learning techniques, which includes random decision forests, support vector machines, multilayer perceptron, and logistic regression. For the tool, the historical data is extracted and then pre-processed and labelled in order to train the algorithm and develop a reliable decision model. With this tool, the developers can quickly predict whether the source code is compliant and has the potential of containing bugs. According to Microsoft, In machine learning, supervised learning let algorithms predict an output based on historical examples of input-output pairs, i.e. labelled data. Further, supervised learning is termed as a classification problem if the output variable is discrete; however, specific patterns in the software project’s code base carry a higher risk of introducing a bug. These patterns can be learnt by a classification learning algorithm to predict the prospect of a file in a commit having a bug. This allows shift left of defect discovery, thus minimising the cost of fixing defects. Microsoft further stated that three custom classification models had been created for three GitHub projects based on metadata associated with the historical commits. Labelled data for training the model has been created using the metadata collected from the GitHub repository. When the news AI-based tool Code Defect AI discovers new developer commits, it obtains the metadata for each of the commits and then utilises it to predict the bugs. “Traditional machine learning models are black boxes, and the rationale behind the model’s prediction is not available. We present the rationale behind the prediction using Local Interpretable Model-Agnostic Explanations (LIME) so that users develop a greater trust in the prediction,” stated Microsoft. Code Defect AI also supports integration with third-party analysis tools, which makes it extremely easy to use for developers. When asked, David Carmona, General Manager of AI Marketing at Microsoft, said to the media that the two companies are working together to enhance and speed up the process of the software development cycle. This new tool — Code Defect AI, powered by Microsoft Azure, can solve several challenges for developers with the use of machine learning. This new AI-based solution is scalable as it can be hosted on-premise as well as on cloud computing platforms such as Microsoft Azure. The solution has been presented on Github and can be integrated with other source-code management tools as needed. The tool is also available on the Microsoft AI Lab portal for developers to download and use it internally in their organisation.","excerpt":"Altran, in association with Microsoft, has launched a new artificial intelligence-based tool that has been designed to predict and identify the source code files that are carrying a higher risk of having a bug. According to Microsoft’s webpage — “Developers are presented with explanations and factors used in making the specific prediction.” Using machine learning […]","categories":["AI News"],"tags":["Machine Learning","source code"],"author_name":"Sejuti Das","publish_date":"2020-05-20T14:22:17","publication_year":"2020","word_count":710,"keywords":["Go","artificial intelligence","machine learning","TPU","AI","cloud computing","R","Machine Learning","Scala","Rust","source code","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","cloud computing","Azure","TPU","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/altran-announces-ml-based-tool-to-detect-bugs-in-source-code\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10074066,"title":"The Menace of Pre-installed Mobile Apps","content":"Sample this: You try to install a crucial app, but your phone’s memory doesn’t allow it. You try all possible ways to clear memory space by deleting cache, unwanted and accidental photos and videos, however, you still don’t have enough memory to install the app. You dig deep to find that a considerable portion of the phone’s memory is occupied by pre-installed apps. Acting swiftly, you try deleting the ones you never use, but to your utter surprise, they are not deletable. You just can’t wish them away! Pre-installed apps are a not-so-desired reality of smartphones. However, unlike the initial days, when a few mobile phones would come with a rare pre-installed app, these days, Android phones come with a whole bunch of them. Many of them are bloatware – a term used for pre-installed apps or softwares that users do not want, but are saddled with. While providing several benefits like simplifying the device activation process, troubleshooting issues and optimising performance, these pre-installed apps gain extensive control over the device and that can have serious ramifications. User privacy and security on the line A few months ago, Microsoft uncovered severe vulnerabilities in a mobile framework used by renowned mobile service providers in pre-installed Android system apps. In its analysis, Microsoft found that these apps were embedded in the system image of devices, implying that they were installed by phone providers. The system image contains all the settings, configurations, and apps that the original equipment manufacturer and the carrier have decided to provide to end users. Moreover, all the apps were available on Google Play Store. Now, apps available on Google Play go through automatic safety checks. Therefore, the presence of these apps on the Play Store despite safety checks implies that such kinds of vulnerabilities were not scanned for. Detected vulnerabilities in pre-installed apps render mobile devices an easy target for attackers. An attacker may be able to carry out local and remote attacks due to the pre-existing vulnerabilities. The attacker may also get access to the system configuration and sensitive information by exploiting the system privileges. One of the first large-scale studies on pre-installed software on Android devices was published at the 2020 IEEE Symposium on Security and Privacy. The study, An Analysis of Pre-installed Android Software, discusses the ecosystem of pre-installed apps in detail. It found that pre-installed apps in Android phones are used for data collection, tracking, and monitoring without the user’s awareness. Many of these applications contain viruses that could endanger the user’s security. These apps frequently provide user’s access to permissions that aren’t typically available if directly downloaded from the Google Play Store. They grant access to intrusive permissions like the accessibility to information about other apps installed by users. The data thus gathered, is then provided to advertisers and analytics companies. The collected information may include sensitive geolocation data and personally identifiable information gleaned from the email or phone address books of the devices. These pre-installed apps often come with specifically designed backdoors that allow app developers to access phone functionalities like storage or leak personally identifying information to data brokers. There have been several suspicions about mobile phone manufacturers being involved in security breaches concerning personally identifiable information. For example, a few years ago, The New York Times reported that Meta (then Facebook) and device manufacturers like Samsung had secret agreements to collect private data from users without their knowledge. In India, there have been concerns about privacy being jeopardised due to data collected by pre-installed smartphones, essentially those manufactured by Chinese mobile phone companies. In 2020, a petition was filed in the Supreme Court of India demanding mobile phone manufacturers to disclose all pre-installed apps in the outer packaging. In addition, the plea wanted the manufacturers to guarantee users’ privacy by revealing how the data collected from the pre-installed apps would be stored and used. Undoubtedly, security and data privacy is perhaps the most important concern posed by the pre-installed apps. However, there are other concerns too. Take the example of the Glance app that comes pre-installed on several smartphones. Although users need to enable it, it is very difficult for a layman to determine if it is drawing sensitive information from the device. There could be a possibility that it may be drawing on data, but only when users enable the app does it share data with other stakeholders. After all, there are instances when many of these pre-installed apps run in the background without the user’s knowledge making it difficult to disable apps that are found on the home screen. A revenue stream for handset manufacturers Notwithstanding the security issues posed by pre-installed apps, what makes handset manufacturers provide these apps is the revenue they provide. Most of the time, app producers pay mobile phone manufacturing companies to include their apps in the system image. It serves a dual purpose – one, the app gets a promotional platform and recognition which is beneficial for app developers in the long run, two, the handset manufacturers are able lower the price – a key reason why Android phones have been able to target the middle and lower-income groups. Doing away with these apps can cost you While most bloatware cannot be outrightly deleted, some like the Glance app can be disabled. In order to completely get rid of the apps, one could opt for the highly technical way of rooting the device. When rooting your phone, you reach a secured part of the device where system files exist and from there, you will be able to delete unwanted apps. However, that comes at the cost of device security.  Rooting also increases the chances of bricking the device wherein your phone turns into an expensive unusable ‘brick’ due to mis-operation. Moreover, handset manufacturers revoke the warranty, if the device has been rooted. Way forward A possible way out of this mess would be if manufacturers provide documentation for the specific set of apps that they have pre-installed in the devices, along with their purpose and the entity responsible for each such application. It should be accessible and understandable to users. Such a practice will ensure that at least a reference point exists for users and regulators to find accurate information about pre-installed apps and their practices. With the evolution of mobile technology, as newer threats and vulnerabilities are discovered, collaboration among security researchers, software vendors and other stakeholders can improve the overall security so that end users are shielded from present and future threats.","excerpt":"A study found that pre-installed apps in Android phones are used for data collection, tracking, and monitoring, without the user’s knowledge","categories":["AI Features"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-09-01T10:00:00","publication_year":"2022","word_count":1082,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","RAG","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-menace-of-pre-installed-mobile-apps\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10092988,"title":"12 Resources to Master Prompt Engineering","content":"In an era when AI threatens to wipe out jobs, prompt engineering is an essential skill to stay relevant. Goldman Sachs recently published a report which suggests that approximately 18% of jobs around the world may be automated by generative artificial intelligence (AI), potentially affecting up to 300 million jobs. Table of contentsChatGPT Prompt BookOpenAI Best PracticesPromptPapersMidjourney Prompt HelperDALL-E Prompt BookEmergent MindLearn PromptingPromptPerfectPromptCrafts-RoboticsMaximising the Potential of LLMs: A Guide to Prompt EngineeringPrompt Engineering GuideAwesome ChatGPT Prompts So in a scenario like this, prompt engineering is a skill that will shape the future of technology. A well-crafted prompt can mould the output into your desired form. It is like an empirical science that starts with a basic prompt and gradually layering on complexity. In an era when AI threatens to wipe out jobs, prompt engineering is an essential skill to learn to stay relevant. We have curated a list of useful resources for you to master prompt engineering. Let’s take a look. Read more: An Entire Generation is Studying for Jobs that Won’t Exist ChatGPT Prompt Book The ChatGPT Prompt Book consists of over 300 unique writing prompts generated by the ChatGPT language model, designed for creative thinking and finding new ideas and perspectives. The prompts cover a diverse range of subjects and are adaptable to various writing styles and genres, making them useful for writers of all levels of experience, from beginners to professionals. OpenAI Best Practices If you are new to the field of prompt engineering, OpenAI’s GitHub repository consisting of best practices for the same is a good resource to start with e.g. the OpenAI Prompt Cookbook. PromptPapers Adding to the extensive list of prompt engineering resources is the GitHub repository PromptPapers. This repository contains a collection of essential papers on tuning pre-trained language models using prompts, as well as other useful learning materials. Midjourney Prompt Helper Midjourney Prompt Helper is a prompt generator for converting text to image, crafted exclusively for Midjourney and Dall-E, with a focus on ease of use and accessibility for users. DALL-E Prompt Book An additional resource for leveraging the complete potential of DALL-E with engaging prompts is the DALL-E Prompt Book, which is a downloadable guide in PDF format that includes fundamental tips for prompts related to photography, illustration, 3D styles, and other topics. Emergent Mind Emergent Mind offers a collection of useful ChatGPT examples sourced from different websites. It provides an opportunity to explore a wide range of interesting and entertaining prompts, along with their sources, use cases, and origins. The platform also features a Hotlist, the latest additions, and an all-time best of, making it a valuable resource for discovering new prompts. Learn Prompting Besides offering extensive prompt engineering documentation and course, Learn Prompting platform also has a Discord server that has attracted over 1,000 members, making it an excellent platform to exchange ideas, find potential partners for collaboration, or just socialise with like-minded individuals and stay updated. Read more: Worried About AI Taking Over Your Job? These 5 Prompt Engineering Courses Will Keep You Ahead of the Game! PromptPerfect PromptPerfect is a state-of-the-art prompt optimiser tool that enhances prompts for various types of language models, including ChatGPT, Midjourney, DALL.E, and StableDiffusion. With PromptPerfect’s multi-goal optimisation, users can customise the prompt optimisation to their specific needs, such as faster optimisation or shorter prompts. PromptCrafts-Robotics PromptCraft-Robotics is a GitHub repository where individuals can test and share interesting prompts that are specifically designed for LLMs related to robotics. The platform offers a robotics simulator, which is integrated with ChatGPT. PromptCraft-Robotics welcomes users to contribute prompts derived from other LLMs, including GPT-3, GPT-4 and Codex, as well as open-sourced models. Submissions are divided into different robotics genres such as manipulation, home robotics, physical reasoning, and many others.  The platform encourages users to format their prompt submissions in markdown and specify which LLM they used. Maximising the Potential of LLMs: A Guide to Prompt Engineering Maximising the Potential of LLMs: A Guide to Prompt Engineering gives us an overview of how to harness the full potential of LLMs by generating customised prompts for specific use cases. In addition to delving into the nature of LLMs, their capabilities, and their limitations, the guide also offers insights into the various tasks that LLMs can perform. Prompt Engineering Guide The Github repository called Prompt Engineering Guide is a treasure trove of resources for those interested in prompt engineering, including learning guides, scientific papers, blog links, tutorials, and datasets. It’s an all-encompassing resource that’s valuable to both developers and practitioners. Awesome ChatGPT Prompts Awesome ChatGPT Prompts consist of You can utilize the many prompts found in this repository with ChatGPT. This encourages you to expand the list with your prompts and to use ChatGPT to create new prompts. Read more: Meet the Computer Scientist Who Solved 50-Year-Old ‘Einstein’ Tiles Problem","excerpt":"In an era when AI threatens to wipe out jobs, prompt engineering is an essential skill to learn to stay relevant.","categories":["AI Trends"],"tags":["AI Tool","ChatGPT","OpenAI","prompt engineering"],"author_name":"Shritama Saha","publish_date":"2023-05-08T16:30:00","publication_year":"2023","word_count":802,"keywords":["Go","ChatGPT","artificial intelligence","TPU","OpenAI","AI","Git","RAG","prompt engineering","AI Tool","R"],"extracted_tech_keywords":["AI","artificial intelligence","ChatGPT","OpenAI","RAG","prompt engineering","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/12-resources-to-master-prompt-engineering\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":47064,"title":"DeepMind Extends Abilities Of Machines To Generate High Fidelity Speech With GAN-TTS","content":"GANs have achieved state-of-the-art results in image and video generation, and have been successfully applied for unsupervised feature learning among many other applications. Generative adversarial networks have seen rapid development in recent years, however, their audio generation prowess has largely gone unnoticed. In an attempt to explore the audio generation abilities of GANs, a team of DeepMind researchers published a work where they introduce a new model called GAN-TTS. Audio Generation With Deep Learning So Far Text-to-Speech (TTS) is a process for converting text into a humanlike voice output. Many audio generation models operate in the waveform domain. They directly model the amplitude of the waveform as it evolves over time. Autoregressive models achieve this by factorising the joint distribution into a product of conditional distributions. Whereas, the invertible feed-forward model can be trained by distilling an autoregressive model using probability density distillation. Models like Deep Voice 2 and 3 and Tacotron 2, in the past, have achieved some accuracy by first generating a representation of the desired output, and then using a separate autoregressive model to turn it into a waveform and fill in any missing information. However, since the outputs are imperfect, the waveform model has the additional task of correcting any mistakes. GANs too have been explored. WaveGAN and GANSynth, have both successfully applied GANs but to much simpler datasets of audio data. The authors believe that GANs have not yet been applied for large scale audio generation operations. With GAN-TTS, they try to do the same. Overview Of GAN-TTS via Paper by DeepMind GAN-TTS is a Generative Adversarial Network for text-conditional high-fidelity speech synthesis. Its feed-forward generator is a convolutional neural network, as shown in the figure above, is coupled with an ensemble of multiple discriminators which evaluate the generated (and real) audio based on multi-frequency random windows. The inner workings of the architecture in both generator and discriminator can be summarised as follows: The generator has seven “GBlocks,” each containing two skip connections: the first performs upsampling if the output frequency is higher than the input. The second contains a size-1 convolution when the number of output channels does not match the input channels. The convolutions are preceded by Conditional Batch Normalisation. Blocks 3–7 gradually up-sample the temporal dimension of hidden representations. The final convolutional layer with Tanh activation then produces a single-channel audio waveform. Whereas, the discriminators consists of blocks (DBlocks) that are similar to the GBlocks used in the generator, but without batch normalisation. Instead of a single discriminator, an ensemble of Random Window Discriminators (RWDs) was used. Notably, the number of discriminators only affects the training computation requirements, as at inference time only the generator network is used. In the first layer of each discriminator, the input raw waveform is downsampled to a constant temporal dimension. The conditional discriminators have access to linguistic and pitch features and can measure whether the generated audio matches the input conditioning. The results from the experiments show that the GAN-TTS is capable of generating highly-fidelity speech, with the best model achieving a MOS score of 4.2, only 0.2 below state-of-the-art performance. Conclusion The researchers believe that the use of RWD is the game-changer here although they say that don’t know the reason behind this. They posit that RWDs work much better than the full discriminator because of the relative simplicity of the distributions that the former must discriminate between, and the number of different samples one can draw from these distributions. GAN-TTS is capable of generating high-fidelity speech with naturalness comparable to state-of-the-art models, and unlike autoregressive models, it is highly parallelizable thanks to an efficient feed-forward generator. Though the widely popular WaveNet has been around for a while, it largely depends on the sequential generation of one audio sample at a time, which is undesirable for present-day applications. GANs, however, with their parallelizable traits, make for a much better option for generating audio from text.","excerpt":"GANs have achieved state-of-the-art results in image and video generation, and have been successfully applied for unsupervised feature learning among many other applications. Generative adversarial networks have seen rapid development in recent years, however, their audio generation prowess has largely gone unnoticed.  In an attempt to explore the audio generation abilities of GANs, a team […]","categories":["Deep Tech"],"tags":["DeepMind"],"author_name":"Ram Sagar","publish_date":"2019-10-10T10:12:25","publication_year":"2019","word_count":650,"keywords":["Go","API","TPU","programming_languages:R","AI","neural network","programming_languages:Go","deep learning","GAN","R","DeepMind"],"extracted_tech_keywords":["AI","deep learning","neural network","TPU","R","Go","API","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/deepmind-extends-abilities-of-machines-to-generate-high-fidelity-speech-with-gan-tts\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10129313,"title":"AI Artists are Making Money from Selling Prompts Rather than Art","content":"Since the start of the generative AI wave, artists have explored different avenues to monetise AI-generated art. However, slowly, the trend is shifting from selling artworks to selling the prompts used to create them. In February 2023, David Sandonato,  an Italian digital artist, began selling Midjourney prompt catalogues on PromptBase, a marketplace for AI art prompts. Today, he is the top-ranked artist on the platform, offering a library of 4,000-5,000 prompts, with new uploads daily. In a recent interview, Santonato said, “It began as a side hustle, but I’m convinced that this business has big space to grow when people will realise that today 50% of the images available in the top microstock agencies can be generated in full quality with a good prompt”. Just a few months ago, a 19-year-old artist, Ashok Reddy, sold nearly 100 AI-generated art pieces in two days at Bengaluru’s Church Street. “AI art is real art, and there’s no shying away from this statement,” Reddy told AIM. Another example is “The Portrait of Edmond de Belamy” by the French art collective Obvious. Created using AI, the artwork was auctioned for $432,500 at Christie’s in 2018—far exceeding the initial estimate of $7,000-$10,000. The shift from selling art to selling prompts has been growing as AI artists find new ways to boost their income, with online marketplaces like PromptBase, Ai4 Prompt, and Etsy hosting prompt catalogues for Midjourney, Dall-E, and Stable Diffusion. For instance, Justin Reckling, a notable figure in “prompt engineering,” created the popular “Block Cities” prompt, which generates isometric tiles of city skylines and became a top seller on PromptBase. Online Marketplace for Prompt Sellers? Etsy recently announced that sellers can now sell artwork created from their own prompts or AI tools, provided they disclose the use of these methods in the listing description. Additionally, Etsy has explicitly prohibited the sale of AI prompt bundles. The company believes the prompts used to generate AI artwork are an integral part of the creative process and should not be sold separately from the final artwork. While Etsy is setting policy, PromptBase has expanded its offerings since its launch. Initially, PromptBase provided prompts only for DALL-E2 and GPT-3. Now, it also supports other AI systems such as Midjourney, Stable Diffusion, and ChatGPT. For just $1.99 per prompt, users can buy the exact phrases they need to generate consistent AI images or create engaging social media posts. Platforms like Etsy and PromptBase are also helping artists generate good revenue. For instance, in PromptBase, users can sell their own prompts and keep 80% of the revenue from each approved sale. Sandanato mentioned that the PromptBase marketplace initially added 30-40% to his monthly income. However, this began to decline as competition from other prompt engineers increased. AI Vs Artists Sneha Chakraborty, one of the panellists at Cypher 2023 and a muralist, said, “It is foolish to resist AI while creating art.” Artists face a difficult yet exciting time in the era of AI art. Today, there are many tools democratising art creation, enabling anyone to produce exceptional images from simple prompts. However, earlier in the initial stages of AI art, the artists’ community reacted negatively towards AI-based art generation tools like Midjourney and DALL-E, even considering a boycott. The backlash intensified in August when an AI artist using Midjourney, won a digital art competition. Since AI is on the burgeoning rise, art is also becoming a competitive field where recognition and value are highly sought after. Many artists fear that AI could create superior art, leading to resentment and job insecurity. However, the sentiment has shifted. Artists are now embracing AI to create and monetise their work, from selling AI-generated art to profiting from AI prompts.","excerpt":"“People will soon realise that today 50% of the images available in the top microstock agencies can be generated in full quality with a good prompt,” said digital artist David Sandonato","categories":["AI Trends"],"tags":["art"],"author_name":"Gopika Raj","publish_date":"2024-07-16T17:54:36","publication_year":"2024","word_count":616,"keywords":["Go","ChatGPT","DALL-E","AI","Git","art","GPT","Aim","prompt engineering","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","Aim","prompt engineering","R","Go","Git","GPT","DALL-E"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ai-artists-are-making-money-from-selling-prompts-rather-than-art\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":48268,"title":"5 Skills That Universities Look For In A Candidate For Data Science Master’s","content":"Data Science is the hottest job around the globe and is fast gaining traction among millennials and Gen Z. It is also one of the highest-paying jobs in the industry. Currently, foreign universities and institutes are rolling out data science and analytics undergraduate and graduate programs in order to churn out a steady stream of job-ready professionals. But before applying for MS in the US, UK or Canada, there is a set of criteria candidates need to fulfill in order to make the cut. These are the entry requirements that are set by these universities. For example, the University of Edinburgh’s Business School expects incoming students for MSc Analytics to have a strong background in linear algebra, calculus, probability, statistics, and computer programming. In a similar vein, Data Science Institute of Columbia University New York expects applicants of MS in Data Science and Certification of Professional Achievement to have a clear understanding of linear algebra, probability, statistics, Python, Java, C+, among others. As the competition for making the cut in foreign universities intensifies, students need to factor in more than just their GRE scores to get admissions in top-ranking universities. In this article, we list down entry requirements of foreign universities look for MS applicants: 1| Statistical And Mathematical Knowledge When you are making an investment of $60k USD, you need a strong stats\/Math background to survive in the MS. The statistical and mathematical background is a must if one wishes to pursue a career in data science and analytics. Before joining the data science course at a university, one can prepare for the core concepts of data science which are linear algebra, calculus, probability, and statistics. While learning machine learning models at the university, it will be easier for you to understand the concepts of conditional probability, priors and posteriors, and maximum likelihood. 2| Build Data Intuition Before joining any institute, one can participate in predictive modelling competitions. There’s Predictive Modeling -NMIMS Competition in Kaggle which will help in understanding the data modelling process and also hone their skills. This will push you ahead in your data science and analytics career. Online courses are available where one can learn how to prepare for these competitions. Building intuition about the data from Coursera is one such example. 3| Knowledge of Algorithms Whether it is the algorithms of data structure or the basic algorithms of machine learning, students are expected to have a basic knowledge of these algorithms. For instance, one must have an understanding of the notations  — the best and worst cases of data structure algorithms which shows how quickly or slowly an algorithm runs, the pseudocodes of machine learning algorithms and its working process, etc. An understanding of these details will help you to stay ahead in your journey. 4| Strong Programming Skills This is an important skill and a strong knowledge of programming language is very important for doing all computational problems. A good starting point is Python, dubbed as the most popular ML programming language. 5| Knowledge of Tools Along with an understanding of statistics and programming languages, knowledge of important tools which are being widely used in the data science projects and statistical operations are a plus for someone who wants to put a solid grip in this domain. Some of the top data science tools are Apache Spark, BigML, Tableau, Jupyter, Matplotlib, ScikitLearn,  Natural Language Toolkit (NLTK), Weka, TensorFlow, Pandas, ggplot, among others. Outlook To the fact, the admissions are getting tougher and eligibility criteria are racked up. It’s not just a GRE\/GMAT score, grades or work experience that counts nowadays, skills like intuitive, communications, quantitative, among others are also being counted in order to get admitted into a renowned institution.","excerpt":"Data Science is the hottest job around the globe and is fast gaining traction among millennials and Gen Z. It is also one of the highest-paying jobs in the industry. Currently, foreign universities and institutes are rolling out data science and analytics undergraduate and graduate programs in order to churn out a steady stream of […]","categories":["AI Trends"],"tags":["data science master","master data science","MSc data science","msc data science and analytics","MSc. in data science"],"author_name":"Ambika Choudhury","publish_date":"2019-10-18T17:05:14","publication_year":"2019","word_count":617,"keywords":["MSc. in data science","data science","machine learning","AI","ML","data science master","msc data science and analytics","Jupyter","Matplotlib","master data science","analytics","MSc data science","NLTK","TensorFlow","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","TensorFlow","Jupyter","NLTK","Pandas","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-skills-that-universities-look-for-in-a-candidate-for-data-science-masters\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10123453,"title":"Databricks Partners with NVIDIA to Unleash ‘Sovereign AI’ in Enterprise","content":"At its Data+ AI Summit, Databricks announced an expanded collaboration with NVIDIA to optimise data and AI workloads by integrating NVIDIA CUDA-accelerated computing into the core of Databricks’ Data Intelligence Platform. The partnership aims to boost the efficiency, accuracy, and performance of AI development pipelines for modern AI factories, as data preparation, curation, and processing are crucial for leveraging enterprise data in generative AI applications. Through this broadened alliance, Databricks is adding native support for NVIDIA GPU acceleration on its Data Intelligence Platform. The announcement builds upon the companies’ existing collaboration to enrich enterprises’ experiences across various use cases, from training classical machine learning models to building and deploying generative AI applications and optimising digital twins. “We’re thrilled to continue growing our partnership with NVIDIA to deliver on the promise of data intelligence for our customers from analytics use cases to AI,” said Ali Ghodsi, Co-founder and CEO at Databricks. “Together with NVIDIA, we’re excited to help every organisation build their own AI factories on their own private data.” Jensen Huang, founder and CEO of NVIDIA, emphasised the importance of accelerated computing in reducing data processing energy demands for sustainable AI platforms. “By bringing NVIDIA CUDA acceleration to Databricks’ core computing stack, we’re laying the foundation for customers everywhere to use their data to power enterprise generative AI,” Huang stated. A key aspect of the partnership involves Databricks developing native support for NVIDIA-accelerated computing in its next-generation vectorised query engine, Photon. This integration is expected to deliver improved speed and efficiency for customers’ data warehousing and analytics workloads. Photon powers Databricks SQL, the company’s serverless data warehouse known for its industry-leading price-performance and total cost of ownership (TCO). The collaboration is anticipated to lead to the next frontier of price-performance. Databricks Shares a Unique Partnership with NVIDIA In the backdrop of Databricks’ Data + AI Summit 2024, Anil Bhasin, the vice president of India and SAARC region at Databricks, told AIM that the company’s partnership with NVIDIA—one of its strategic investors–is significant, alongside helping them improve run times using their SOTA GPUs. “We’ve always been known as pioneers of the lake house architecture, and now we’ve created a new category called the data intelligence platform. We’ve embedded generative AI in the lake house, which is a unique approach not many companies are taking,” said Bhasin, saying that NVIDIA is aligned with their vision because the future lies in data intelligence platforms. He said this allows them to serve every use case, ingest data from any source, and maintain unified governance. This strategic differentiation makes their partnership with NVIDIA truly special and aligns perfectly with NVIDIA’s ‘Sovereign AI’ for enterprises. Nobody other than Databricks is enabling this. “The ability for us to query in natural language, converting it to SQL on the back end, empowers the business user to gain insights. That is true democratisation,” avered Bhasin, saying their vision is powerful, and not just NVIDIA; many companies believe in Databricks’ long-term vision. NVIDIA x Databricks Recently, Databricks’ open-source model DBRX became available as an NVIDIA NIM microservice. NVIDIA NIM inference microservices provide fully optimised, pre-built containers for deployment anywhere, significantly increasing enterprise developer productivity by offering a simple, standardised way to add generative AI models to their applications. Launched in March 2024, DBRX was built entirely on top of Databricks, leveraging the platform’s tools and techniques, and was trained with NVIDIA DGX Cloud, a scalable end-to-end AI platform for developers. The Databricks Data Intelligence Platform offers a comprehensive solution for building, evaluating, deploying, securing, and monitoring end-to-end generative AI applications. With Databricks Mosaic AI’s data-centric approach, customers benefit from an open, flexible platform to easily scale generative AI applications on their unique data while ensuring safety, accuracy, and governance. Today’s announcement follows Databricks’ strategic acquisition of Tabular, a data management startup founded by the original creators of Apache Iceberg and Linux Foundation Delta Lake, the two leading open-source lakehouse formats. By bringing together these key players, Databricks aims to lead the way in data compatibility, ensuring organisations are no longer limited by the format of their data. Driven by the growing demand for data and AI capabilities, Databricks achieved over $1.6 billion in revenue for its fiscal year ending January 31, 2024, representing more than 50% year-over-year growth. The expanded partnership between Databricks and NVIDIA underscores the critical role of accelerated computing and optimised data processing in enabling enterprises to harness the power of generative AI effectively and efficiently.","excerpt":"The partnership aims to boost the efficiency, accuracy, and performance of AI development pipelines for modern AI factories.","categories":["Global Tech"],"tags":["Databricks","NVIDIA"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-12T19:48:53","publication_year":"2024","word_count":739,"keywords":["CUDA","machine learning","AI","serverless","RAG","microservices","Aim","analytics","generative AI","NVIDIA","Databricks"],"extracted_tech_keywords":["AI","machine learning","analytics","generative AI","Aim","RAG","microservices","serverless","CUDA","Databricks"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/databricks-partners-with-nvidia-to-unleash-sovereign-ai-in-enterprise\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040385,"title":"What To Expect From Google I\/O 2021 Conference","content":"Google is back with its annual keynote conference – Google I\/O 2021. Last year, Adobe, Apple, and Microsoft were all able to hold virtual sessions of their annual developer conferences, but Google cancelled its I\/O entirely due to the pandemic. Now, the California-based tech giant is expected to make a handful of interesting announcements at the virtual summit, including the first beta for Android 12 – the next version of its Android system. Google’s social media networks will broadcast the three-day event online. Google is also expected to reveal some new Wear OS and Google Assistant updates. We have listed down some of the developments to be expected at the summit. 1| The latest Android version One of the significant announcements expected would be around the latest Android 12. Google’s focus will be on more robust privacy and security protections and might get a big user interface (UI). Toggles, buttons, widgets, and animations can all be redesigned in Android 12. A redesigned media player widget, a new brightness dial, new Wi-Fi and Bluetooth toggles, a new analogue clock widget, and other elements could be included. “Brand new stacked notifications” are part of Android 12. It implies that users will be able to organise app alerts according to their preferences, as described in the FPT video. 2| Better web platform Google web platform has seen tremendous growth over the years. With the platform hosting nearly five billion users, there is a need to update the platform with new features for a better user experience. Google is making an investment in a better web by fostering user trust in a privacy-first future, driving innovation with Chrome, and providing resources and guidance to developers. 3| Smarter Home Products Google Assistant will be five years old in 2021 and the company’s virtual assistant is expected to steal some thunder at I\/O. The company will share some new developments for smart homes. The organisation has promised to go into detail about their product vision, new product announcements, and awesome Assistant interactions created by their developer group, as said in a blog post from Google. 4| Latest hardware offerings Although Google doesn’t normally make hardware announcements at its annual developer conference as the conference is largely considered as a software show. However, the Pixel Buds A-series truly wireless (TWS) stereo earbuds may be on the way. The Pixel 5A smartphone first rumoured to be unavailable due to chip shortage but later confirmed its availability in the US and Japan as of now. However, it is also expected to be unveiled at this year’s Google I\/O gathering. However, little information about the latest developments is there. 5| Wear OS Wear OS, formerly known as Android Wear, is scheduled to get some new features at this year’s Google I\/O. Fitbit is now a Google partner and could further boost Wear OS with new fitness features. The changes could also make way for Google’s own smartwatch, the Pixel Watch, which could have a premium design to compete with the Apple Watch. Moreover, the changes might be expected to gain the upper hand in the market against Apple’s watchOS and may also include some better features for fitness enthusiasts. Apart from these, one can have a sneak peek into the role of women in Voice AI. There will be a session including an interview with Lilian Rincon, Sr. Director of Product Management at Google, to understand the role women can play. Voice AI is revolutionising how we communicate with technology, and its future will be determined by those who build it. One can learn about the influential women in the field of voice AI. There will be talking about how to achieve fair gender representation in Voice AI, which is a lofty yet necessary objective. There’s always the chance that Google may have something completely unexpected in store for this year’s I\/O. Stay tuned to Analytics India Magazine for more updates from Google I\/O event.","excerpt":"Google is back with its annual keynote conference – Google I\/O 2021. Last year, Adobe, Apple, and Microsoft were all able to hold virtual sessions of their annual developer conferences, but Google cancelled its I\/O entirely due to the pandemic. Now, the California-based tech giant is expected to make a handful of interesting announcements at […]","categories":["Global Tech"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-05-18T18:57:21","publication_year":"2021","word_count":655,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","analytics","Rust","GAN","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","analytics","R","Go","Rust","GAN","innovation","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/what-to-expect-from-google-i-o-2021-conference\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10073883,"title":"Telangana Govt Calls Upon AI to Repair Roads","content":"Telangana AI Mission (T-AIM) has rolled out a Mobility AI Grand Challenge to assist Hyderabad Municipal Corporation to identify and classify potholes for repair. The challenge is open to all AI innovators in the country and the shortlisted participants will get four weeks to develop a proof of concept. With the help of the solutions presented, startups are expected to provide spatial density and distribution of potholes based on their severity across the city. The solutions would determine the number of potholes, examine their exact depth and find measures to fill them. T-AIM said that the submissions would be judged on approach and results. “Greater Hyderabad Municipal Corporation (GHMC) aims to identify and classify pothole severity across specified routes in Hyderabad using live and archived video feeds. The solutions developed using artificial intelligence will provide its officials additional insights to undertake targeted repair works.” Source: LinkedIn Rewards worth ₹20 lakh The winning innovator would get an award of up to ₹20 lakh towards implementation of a potential pilot project in Telangana. The Mobility AI Grand Challenge is conducted in partnership with Capgemini, and aims to foster innovation for GHMC. A similar challenge called ‘Forest AI Grand Challenge’ was conducted by T-AIM for the forest department this month. The deadline to submit proposals is September 16 and the winner would be announced on November 22. For details, visit https:\/\/taim-gc.in\/mobility\/.","excerpt":"The challenge is open to all AI innovators in the country. Shortlisted participants will get four weeks to develop a proof of concept.","categories":["AI News"],"tags":["Startups","telangana"],"author_name":"Bhuvana Kamath","publish_date":"2022-08-29T16:33:54","publication_year":"2022","word_count":228,"keywords":["artificial intelligence","programming_languages:R","AI","telangana","innovation","Aim","llm_models:Gemini","GAN","Startups","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","GAN","innovation","startup","llm_models:Gemini","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/telangana-govt-calls-upon-ai-to-repair-roads\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10072644,"title":"Intel is back again with its flagship event for HPC &#038; AI developers — oneAPI DevSummit, Southeast Asia | Register Now!","content":"Intel®, alongside Analytics India Magazine, has recently announced the launch of oneAPI DevSummit, Southeast Asia, for 15 September 2022 from 9:00 AM to 6:00 PM IST. This conference would deep-dive into the latest practices and techniques for cross-platform development focused on oneAPI, SYCL, and AI for accelerated computing across xPU architectures, including CPU, GPU, FPGA, and other accelerators. oneAPI Developer Summit is a peer-to-peer community event where industry speakers and academicians showcase their innovative work and share their experiences on heterogeneous computing. The summit has been designed to build a self-sustained, vibrant community of developers who can support each other using oneAPI and Data Parallel C++ programming language, getting trained, developing and submitting code. The one-day live virtual conference has been designed to learn from leading industry and academia speakers working on innovative cross-platform, multi-vendor architecture oneAPI solutions. Collaborate with fellow developers and connect with other innovators. ​Dive into a hands-on session where you will learn and apply optimizations to exploit device capabilities on CPUs and GPUs fully. The conference has five segments, which include an opening note, tech talks, lightning talks, live discussions, and hands-on workshops. The tech talks from IIT Dharwad, Durham University and Codeplay Software are the core technical sessions, where Intel customers and partners will showcase their innovative work using oneAPI, SYCL AI\/ML. Following this are the Lighting Talks—short sessions with experts from L&T Technology Services, D&I Technologies PTE ltd.,  BittWare & Boston IT Solutions. Listen to some interesting conversations and sample codes using oneAPI from our community speakers and oneAPI Certified Instructors Abhishek Nandy, Co-Founder Dynopii and Jehferson Mello, Software Developer. This summit will be a perfect opportunity to learn Intel AI tools for end-to-end development, including data preparation, training, inference, deployment, and scaling. One can also expand their skills with a hands-on workshop focused on “GPU porting of an HPC application using Intel® oneAPI” & an oneAPI AI Analytics workshop on “Intel optimizations for deep learning frameworks.” Register for this FREE event today and explore what oneAPI DevSummit, Southeast Asia has to offer. Date: 15 September 2022 Time: 9:00 AM – 6:00 PM (IST) Mode: Online Register for oneAPI DevSummit, APJ here Highlights of oneAPI DevSummit, Southeast Asia — A platform to connect with fellow developers and innovators.Learn about the latest developer tools for oneAPI.Tech talks by industry thought leaders and academicians working on innovative cross-platform, multi-vendor oneAPI solutions.Discover real-world projects using oneAPI to accelerate data science and AI pipelines.Dive into a hands-on session on Intel® oneAPI toolkits for HPC and AI applications.Join a vibrant community supporting each other using oneAPI, SYCL and AI. Who should attend? HPC & GPU ProgrammersAI Enthusiasts & Data Science AspirantsAI & ML developersAI ResearchersData Scientists Why care about oneAPI? Because it … Removes code barriers with one unified programming model for multiple compute architectures, including GPUs, AI accelerators, and FGPAs.It eliminates the need to maintain separate code bases and\/or multiple programming languages.Delivers a complete set of cross-platform libraries, tools, and frameworks—based on familiar languages and standards—to make heterogeneous development easier.Empowers developers to expose and exploit the latest hardware’s cutting-edge features to unleash performance. Preview the full agenda for more details [https:\/\/www.oneapi.io\/events\/oneapi-devsummit-se-asia-2022\/] Exclusive Contests — Participate & Win!oneAPI DevSummit, Southeast Asia, is hosting many contests, games, challenges and early-bird registration contests for the attendees to participate and win exciting prizes.Read below to know more about the contests.#1 DevSummit Early Bird PrizeFirst 30 to complete Intel Notified registrations and actively participate on Notified will stand a chance to win Amazon Vouchers!Winners will be  announced during the Summit#2 oneAPI AI Analytics Toolkit Workshop ContestRegister, Participate & Win! Up to 5 winners from the workshop contest will stand a chance to win Fitbit smartwatches. #3 Jeopardy GameTop 10 scorers on the dashboard will win Amazon vouchers!To know the T&C of the Intel oneAPI DevSummit Contest, click here. Join us for a day of discovery with renowned industry & academia experts who demystify the latest technologies, tools, trends, and techniques. Date: 15 September 2022 Time: 9:00 AM – 6:00 PM (IST) Mode: Online Register for oneAPI DevSummit, APJ here To learn more about oneAPI, visit: https:\/\/www.oneapi.io\/ Sign up for a free account with the Intel® DevCloud, which you will need when participating in hands-on sessions. To sign up, click here: https:\/\/intel.ly\/3z9OV7A Also, check out oneAPI DevMesh to explore and learn from open-sourced oneAPI projects created by us and our community. Click here: https:\/\/devmesh.intel.com\/topics\/77 You can now join the Discord Chat, where live discussion and Q&As will be ongoing before, during, and after the event. Click here: https:\/\/discord.gg\/ycwqTP6 ** Please join us on 15 September 2022 to be a part of a self-sustained, vibrant community of oneAPI. Save your spot for this live, virtual event","excerpt":"Are you a C++ and GPU Programmer, AI developer, Researcher, or Data Scientist who develops applications in HPC and AI? Interested in accelerating workflows and delivering high-performance applications across one or multiple architectures? Then this event is for you.","categories":["Deep Tech"],"tags":["developers India","software developers India"],"author_name":"Tasmia Ansari","publish_date":"2022-08-11T16:00:00","publication_year":"2022","word_count":781,"keywords":["data science","Go","API","AI","ML","C++","deep learning","analytics","developers India","AI research","R","software developers India"],"extracted_tech_keywords":["AI","ML","deep learning","data science","analytics","R","Go","C++","API","AI research"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/intel-is-back-again-with-its-flagship-event-for-hpc-ai-developers-oneapi-devsummit-southeast-asia\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10144215,"title":"Musk’s xAI Raises $6 Billion in Series C Funding","content":"xAI has completed a $6 billion Series C funding round, attracting investments from prominent firms such as Andreessen Horowitz (a16Z), BlackRock, Sequoia Capital, NVIDIA and AMD, among others. According to their blog, the funding will help the company’s efforts to expand its infrastructure and accelerate AI research. “Really proud of the team for all the crazy things we executed in lightning time! Here’s to greater heights in 2025!” xAI co-founder Greg Yang posted on X. In reference to the recent funding, xAI CEO Elon Musk took to X to share that “a lot of compute is needed”. Notably, Musk co-founded OpenAI alongside Sam Altman. In 2023, he departed to compete with established AI labs like OpenAI and Anthropic. Earlier this year, OpenAI secured $6.6 billion in funding from investors, including Thrive Capital, NVIDIA, Microsoft, and SoftBank Group. “$12B raised this year since Grok-1 was launched a year ago. We’re building AI end-to-end; we’re building our own supercomputer and our own human data team. Grok-3 will be the most powerful engine,” said Mat Roy, project lead at xAI. Musk aims to develop Artificial General Intelligence (AGI) while prioritising rigorous truth-seeking systems without ideological biases. The upcoming frontier model Grok 3 is envisioned as a major advancement, aiming to outperform existing AI models. xAI has introduced Colossus, the world’s largest AI supercomputer. It is powered by 1 lakh NVIDIA Hopper GPUs and will be operational within 122 days, with plans to expand to 2 lakh GPUs. xAI is currently training its next flagship model, Grok 3, which is likely to be released next year. Since its Series B announcement in May, xAI has introduced several key initiatives, including Grok 2, an advanced language model; Aurora, an image generation tool; and the xAI API, which provides developers with low-latency access to AI models globally. The integration of Grok with the X platform enhances real-time data analysis with features like web search and image generation.","excerpt":"Musk aims to develop AGI grounded in rigorous truth-seeking and devoid of ideological bias.","categories":["AI News"],"tags":["Elon Musk","Funding"],"author_name":"Aditi Suresh","publish_date":"2024-12-24T12:28:01","publication_year":"2024","word_count":321,"keywords":["Anthropic","Grok 3","Funding","Go","API","OpenAI","AI","R","Elon Musk","XAI","Aim","xAI"],"extracted_tech_keywords":["AI","OpenAI","Anthropic","Grok 3","xAI","Aim","R","Go","API","XAI"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/musks-xai-raises-6-billion-in-series-c-funding\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":21530,"title":"How To Handle Null\/Missing Values In Machine Learning Datasets (5 Ways)","content":"In real world data, there are some instances where a particular element is absent because of various reasons, such as, corrupt data, failure to load the information, or incomplete extraction. Handling the missing values is one of the greatest challenges faced by analysts, because making the right decision on how to handle it generates robust data models. Let us look at different ways of imputing the missing values. Note: We will be using libraries in Python such as Numpy, Pandas and SciKit Learn to handle these values. Let us get started. To understand various methods we will be working on the Titanic dataset: 1. Deleting Rows This method commonly used to handle the null values. Here, we either delete a particular row if it has a null value for a particular feature and a particular column if it has more than 70-75% of missing values. This method is advised only when there are enough samples in the data set. One has to make sure that after we have deleted the data, there is no addition of bias. Removing the data will lead to loss of information which will not give the expected results while predicting the output. Pros: Complete removal of data with missing values results in robust and highly accurate model Deleting a particular row or a column with no specific information is better, since it does not have a high weightage Cons: Loss of information and data Works poorly if the percentage of missing values is high (say 30%), compared to the whole dataset 2. Replacing With Mean\/Median\/Mode This strategy can be applied on a feature which has numeric data like the age of a person or the ticket fare. We can calculate the mean, median or mode of the feature and replace it with the missing values. This is an approximation which can add variance to the data set. But the loss of the data can be negated by this method which yields better results compared to removal of rows and columns. Replacing with the above three approximations are a statistical approach of handling the missing values. This method is also called as leaking the data while training. Another way is to approximate it with the deviation of neighbouring values. This works better if the data is linear. To replace it with median and mode we can use the following to calculate the same: Pros: This is a better approach when the data size is small It can prevent data loss which results in removal of the rows and columns Cons: Imputing the approximations add variance and bias Works poorly compared to other multiple-imputations method 3. Assigning An Unique Category A categorical feature will have a definite number of possibilities, such as gender, for example. Since they have a definite number of classes, we can assign another class for the missing values. Here, the features Cabin and Embarked have missing values which can be replaced with a new category, say, U for ‘unknown’. This strategy will add more information into the dataset which will result in the change of variance. Since they are categorical, we need to find one hot encoding to convert it to a numeric form for the algorithm to understand it. Let us look at how it can be done in Python: Pros: Less possibilities with one extra category, resulting in low variance after one hot encoding — since it is categorical Negates the loss of data by adding an unique category Cons: Adds less variance Adds another feature to the model while encoding, which may result in poor performance 4. Predicting The Missing Values Using the features which do not have missing values, we can predict the nulls with the help of a machine learning algorithm. This method may result in better accuracy, unless a missing value is expected to have a very high variance. We will be using linear regression to replace the nulls in the feature ‘age’, using other available features. One can experiment with different algorithms and check which gives the best accuracy instead of sticking to a single algorithm. Pros: Imputing the missing variable is an improvement as long as the bias from the same is smaller than the omitted variable bias Yields unbiased estimates of the model parameters Cons: Bias also arises when an incomplete conditioning set is used for a categorical variable Considered only as a proxy for the true values 5. Using Algorithms Which Support Missing Values KNN is a machine learning algorithm which works on the principle of distance measure. This algorithm can be used when there are nulls present in the dataset. While the algorithm is applied, KNN considers the missing values by taking the majority of the K nearest values. In this particular dataset, taking into account the person’s age, sex, class etc, we will assume that people having same data for the above mentioned features will have the same kind of fare. Unfortunately, the SciKit Learn library for the K – Nearest Neighbour algorithm in Python does not support the presence of the missing values. Another algorithm which can be used here is RandomForest. This model produces a robust result because it works well on non-linear and the categorical data. It adapts to the data structure taking into consideration of the high variance or the bias, producing better results on large datasets. Pros: Does not require creation of a predictive model for each attribute with missing data in the dataset Correlation of the data is neglected Cons: Is a very time consuming process and it can be critical in data mining where large databases are being extracted Choice of distance functions can be Euclidean, Manhattan etc. which is do not yield a robust result Conclusion Every dataset we come across will almost have some missing values which need to be dealt with. But handling them in an intelligent way and giving rise to robust models is a challenging task. We have gone through a number of ways in which nulls can be replaced. It is not necessary to handle a particular dataset in one single manner. One can use various methods on different features depending on how and what the data is about. Having a small domain knowledge about the data is important, which can give you an insight about how to approach the problem.","excerpt":"In real world data, there are some instances where a particular element is absent because of various reasons, such as, corrupt data, failure to load the information, or incomplete extraction. Handling the missing values is one of the greatest challenges faced by analysts, because making the right decision on how to handle it generates robust […]","categories":["AI Features"],"tags":["Applications of Data Mining","Data Mining","data mining tools","survival regression python"],"author_name":"Kishan Maladkar","publish_date":"2018-02-09T05:46:50","publication_year":"2018","word_count":1053,"keywords":["Go","NumPy","machine learning","TPU","programming_languages:R","AI","Data Mining","Applications of Data Mining","survival regression python","Python","data mining tools","programming_languages:Python","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","Pandas","NumPy","TPU","Python","R","Go","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/5-ways-handle-missing-values-machine-learning-datasets\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":2412,"title":"Interview &#8211; Amit Batra, Managing Partner at Industrial Data Research Corporation","content":"Industrial Data Research Corp. (http:\/\/www.idrcglobal.com) is a specialist consultancy present in Germany, India and Russia. IDRC enables it clients to achieve operational improvements, measured tangibly as improved financial returns, by employing analytics, quantitative modeling, and scientific computing. IDRC’s clients include FIs, Consultancies (including IT & Research firms), and Corporates. IDRC builds & upgrades client’s analytics practices (e.g. valuations & risk quant, data-driven business insights, investment research, econometrics, quant-led process and operations mgmt.) delivered through IDRC’s QRC(TM) and ProcessAnalytics(TM) frameworks. We talk with Amit Batra from IDRC in this exclusive interview. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [dropcap style=”1″ size=”2″]AB[\/dropcap]Amit Batra: Thank you, Bhasker for inviting me to speak about this. Industrial Data Research Corp. (IDRC) is an analytics consulting firm. IDRC believes that analytics, defined as quant modeling, data analytics, and scientific computing, can be harnessed to change lives, change organizations, and change the world. We believe that analytics will render a huge positive impact worldwide in this century to boost productivity, improve efficiency, and curbing wastage across the board. Analytics is a social force, which is already, and can further be applied in diverse areas – from finance, marketing, retail, e-commerce to developmental economics, conservation, ecology… IDRC aims to be industry agnostic, with our current stronghold of finance & economics as just the starting point. We are already delivering projects in a diverse set of areas. Analytics is already applied by clients, sometimes even unknowingly, but through special expertise and special focus, IDRC ‘makes it work’ for the clients. We insist on measuring the success of implementing a model, or an analysis, or a simulation, or a computation, in terms of improved financial returns, top-line or bottom-line, of our clients. Linking our activity strictly to such a tangible measure is where we think we differentiate ourselves from our competition. Each client engagement, by design, is supposed to always have an IDRC partner’s direct involvement. We get an adrenaline rush through ‘what comes out’, and so treat each project as special. Besides, we aim to fill the ‘quality’ gap, in markets such as India, where there is a dismal penetration of PhD level analysts. We aim to do so by matching the project requirements with, albeit the most expensive, the highest-quality, specialized & specifically trained analysts. As a spin-off from IDRC’s consulting practice, we aim to establish ‘Social Analytics Institute’ to house and nurture innovation in social space. This will connect industry, academia and individuals, and enable systematic evaluation and promotion of ideas. [quote style=”1″]Goal here is to go beyond thought exchange, into actually housing, commercially evaluating, nurturing, hand-holding the ideas to become ‘social products’.[\/quote] AIM: Please brief us about some business solutions you work on and how you derive value out of it. AB: Let me share two recent examples, which also illustrate the diversity of application of IDRC’s analytical consultancy – [pullquote align=”left”]We regularly share interesting updates with a growing community of our LinkedIn followers –https:\/\/www.linkedin.com\/company\/industrial-data-research-corporation[\/pullquote]IDRC recently set up a Quant Research Center, QRC for a cab company client. We have thus achieved efficient zoning, visual ‘demand heat maps’, and optimum holding locations for the taxis, all leading to a progressively improving utilization of the fleet. IDRC delivered research on momentum-based investment strategies and their application in bond and equity trading, aimed at developing software tools to determine the value of the analysis of chart techniques. Momentum-driven returns refer to the general principle, according to which stocks, which have historically outperformed over a period of length m, will do so in future. For those willing to take a deep-dive, the overall research framework is formally defined as follows: Say U is the universe of charts (sequences of prices over the interval [0, T] of m stocks. We demonstrated that momentum (L, S) works, if for all t in [0, T] the average return for all stocks with over-average performance in the historic time period, L, the performance in period, S, is also above average. We demonstrated that under very mild conditions, for any time series of returns, and independently of the market, there exist L and S, where momentum works. However we also demonstrated that there also have to be H not equal to L, and K not equal to S, in which momentum does not work. Based on these basic considerations the IDRC project empirically determined the values of L and S in different markets. AIM: How does a typical requirement gathering to delivery cycle looks like for you? AB: IDRC has two inter-locking client-servicing frameworks, ProcessAnalytics and Quant Research Center, QRC. These cover a spectrum of needs clients may have with regard to their analytics activities. The ProcessAnalytics solution supports one-time interventions and overhauls of client’s analytics activities; these may be aimed at resolving quality issues and\/or cutting costs. For example, IDRC’s experts may come in to streamline\/augment service deliveries by KPOs\/IT companies to their finance and economics clientele, representing either side. The QRC solution comprises the standards, practices and experience, which allow IDRC to build flexible, cost-effective, scalable and robust research groups (that complement and assist clients’ own research and analytics teams). The QRC solution also allows to upgrade\/consolidate client’s existing in-house analytics teams and processes. The U-shaped model of the service process (Ed. displayed below) highlights the major stages of work within each of the two service frameworks and the principal tangible outcomes resulting from each stage. Work is organised around deliverables. The overall process starts with an Express Audit, which takes around 3–5 working days and aims at determining the major dimensions of the client engagement. Client-centred service process: unifying view of IDRC’s services AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth. [pullquote align=”right”]IDRC currently has there partners, Dr. Thomas Maier, and Mr. Alexey Pan, and me (the detail bios are available on IDRC page http:\/\/www.idrcglobal.com\/team).[\/pullquote]AB: IDRC has as a typical consulting structure. IDRC is a synergized sum-total of its partners’ consulting practices. As I speak, we are on the last stages of signing-up two more extremely talented people. Both of them will be based out of India. We will announce the news shortly. IDRC partners often employ analysts matching the client-project analysis requirement with the analysts’ skills. AIM: Would you like to share any example of an Insight that generated a huge positive impact for your clients? AB: In early 2009, IDRC prophesied that global BFSI Quant Middle Office (QMO) as the next big wave to arrive on the Indian shores. This meant that well prepared shared services captive cost centers and private KPOs would be able to hugely profit from it. However, most of them weren’t (and still aren’t) well prepared! IDRC presented these thoughts to a pioneer Indian Knowledge Process Outsourcing (KPO) company, with an aim to develop just such an expertise. This client-partner firm has operations across four countries, and a global sales presence. The firm specialized in low-end research support, such as data aggregation & cleansing and periodic reports. At that time, with stiff competition the firm was also fast loosing differentiation. Its existing clients demanded more sophisticated services in quantitative analytics space. The existing in-house analyst teams and the sales force where not equipped or trained to deliver on these tasks. IDRC worked with firm’s sales directors globally to upgrade business positioning, and embarked the firms into areas including asset valuations, risk quantifications, proprietary index publication, and structuring. IDRC also consolidated and extended pre-sales pitches & collaterals, and executed a quant sales campaign with firms’ sales directors across geographies, targeting FIs and analytics developers. Through these sales campaign, we on-boarded for the KPO, new engagements with banks and hedge funds, generating fresh revenue streams for the KPO. IDRC facilitated the HR to identify and hire a team of best-in-market, top-notch analysts over, what was, a mammoth effort of screening resumes and conducting interviews. IDRC deployed Quant Research Center (christened Quant Incubator) within the KPO, and mentored the group’s client deliveries. Thus with IDRC retrofitting and injecting effort, the KPO delivered its most analytically sophisticated (and most expensive) services, regaining the pioneer status, becoming an attractive employer, with a larger top-line. AIM: Do you think it’s possible to become too married to the data that comes out of analytics? Where do you draw the line? AB: IDRCs approach is to provide smart solutions. The real value added is not provided by just crunching data and presenting the outcome. It is about understanding the sources of input and the final use of the output, and regularly & continuously optimizing (or re-calibrating) the whole structure. In many cases, especially in research-driven projects, it is absolutely necessary to keep the bird-perspective on the project. You always must keep in mind what the actual goal of the whole project is. For example when analyzing investment strategies via back-testing you have to keep in mind all the selection biases which might occur. It is not sufficient to create a wonderfully back-tested strategy, it is also crucial to address the selection problem as well as to fit the strategy in the reality. Smart back-testing mitigates the problems of backfill bias, survivorship bias, non-synchronous data, and so on, and therefore provides much more value for the client. However to be able to offer those services to the client your organization must be able to have access to people with wider knowledge than just the statistical models. Bringing together different areas of expertise is thus one of the cornerstones of IDRC’s work. AIM: What are some of the data measurement points that are becoming more important for organizations? AB: Especially for European banks and asset managers, the regulatory reforms will force them to present a huge amount of data in real time to the regulator. Outside of finance, this is an age defined by across the board belt-tightening, thinning margins, market volatilities, and also a global emergence of lean ‘e-tailers’. Data-driven decisions towards an efficient supply chain including demand prediction, and deeper consumer insight therefore are, not just ‘good to have’, rather essential. AIM: What are the most significant challenges you face being in the forefront of analytics space? AB: Analytics is a new-wave, and an inter-disciplinary field. This means that fresh graduates, even postgraduates are not well equipped to deliver. IDRC tackles this by going for PhD-level analyst talent. This already means that in a place like India, there’s a scarcity of talent. In the sales process, we also come across senior manager’s skepticism that I already spoke about. Many see analytics in same light as labor-arbitrage outsourcing proposition. I attribute this to their ignorance. Though improving fast, we also find lack of industry standards with regard to software tools, APIs, streaming-data as bit annoying. AIM: How did you start your career in analytics? AB: I am trained as an engineer and a researcher. Hence analytics, quant modeling, and scientific computing have been the way of life for the last 11 years (smile). AIM: What do you suggest to new graduates aspiring to get into analytics space? AB: Two important points come to the mind – First, university students should develop an area of specialization earlier in their careers and be able to show their expertise in that field, e.g. by own work, publications or repeatable practical knowledge. Second, graduates should also focus on their soft skills not just on the technical side. Communication skills and social skills are especially paramount. Combining these two points, it is of tremendous importance to develop a, kind of, ‘can-do’ mentality when faced with practical work, where the challenge might be to formulate the question, or to bring a fuzzy problem into a resolvable structure. [quote style=”1″]This is IDRC’s message to the universities as well, which should design course materials in training programs more aligned with the zeitgeist.[\/quote] AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? AB: Analysts are our foot-soldiers. Hence we seek the very best. We look for exceptional people, and compensate them very well. We consider this an investment and a sign our commitment towards the success of our client-projects. We do not cut-corners in this respect. We usually consider people with 5+ years of relevant work experience and a PhD in mathematics, computing, economics, statistics or like. Exceptionally, we also consider people holding postgraduates degrees with relevant work experience and certifications such as CFA and FRM. We rarely employ fresh graduates. This is unfortunate but we think that more often than not fresh graduates require the kind of hand-holding or training that our bandwidth does not allow. From a good knowledge worker, we expect a creativity and attitude of a being a ‘solution generator’. It is easier to take the projects to the point where all information is available. We judge candidate’s response to imprecise specifications, and fuzzy, unstructured problems. Good people skills, language skills (speaking & writing), and presentation skills are a must. Additionally we judge people based on their reading habits, and several other behavioral aspects. AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? AB: Consider these – Extremely powerful and general purpose analytics tools and frameworks are being developed and becoming accessible. One can now create magic using assortment of Matlab, Excel VBA, Mathematica, Perl, Hadoop, Spotfire, Tableau etc. Parallel computing, which used to be a niche sound in universities and research labs, is now a commodity. Mobile computing and the growing ‘internet of things’, is rendering data generation growth on a logarithmic scale. Theoretical faculties are converging into inter-disciplinary area, techniques of simulations, agent-based modeling, and behavioral sciences are gaining sophistication. Every business is feeling the heat and we predict an emergence of a ‘culture of analytics’. We predict that businesses will have an integrated analytics arm providing feedback on the decisions by real-time churn of data. [quote style=”1″]We also predict that stricter and somewhat consistent codes & regulations will emerge with respect to which data can be collected, and the transparency of the whole thing. This is especially true in the social-media space.[\/quote] AIM: Anything else you wish to add? AB: Well, thanks for this opportunity again Bhasker. I have spoken much, but let me quickly touch back on key points where IDRC differentiates from competitors – First, we make sure that by employing IDRC, the mid & senior management save time, and the outcome is reflected in financial returns. If these are not met, we don’t call it the completion.  Our QRCTM offering is flexible and scalable. Moreover, we don’t stop improving. IDRC partners frequently revisit the model, free of cost, with an aim to cannibalize the revenues and make processes leaner and streamlined. IDRC thinks, lives, and breathes analytics, and offers to integrate within client’s day-to-day operations. Thank you. [divider top=”1″] [spoiler title=”Biography of Amit Batra” open=”0″ style=”2″] Amit has more than 7 years of experience as an analytics leader. He has build & manage specialized analyst teams for clients in firm of intelligence units or COEs, augmenting delivery capabilities for consultancies, KPOs and IT firms in quant, analytics and scientific computing heavy projects. Amit is a graduate from IIT Bombay.[\/spoiler]","excerpt":"Industrial Data Research Corp. (http:\/\/www.idrcglobal.com) is a specialist consultancy present in Germany, India and Russia. IDRC enables it clients to achieve operational improvements, measured tangibly as improved financial returns, by employing analytics, quantitative modeling, and scientific computing. IDRC’s clients include FIs, Consultancies (including IT & Research firms), and Corporates. IDRC builds & upgrades client’s analytics […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Дарья","publish_date":"2013-01-07T12:34:21","publication_year":"2013","word_count":2564,"keywords":["Go","API","TPU","AI","ML","Scala","RAG","Aim","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","TPU","R","Go","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-amit-batra-managing-partner-at-industrial-data-research-corporation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10131969,"title":"Amgen is Launching Biotechnology and AI Hub in Hyderabad","content":"Amgen, a company focusing on biotechnology, discovering, developing, manufacturing, and delivering innovative medicines, has announced plans to establish a new technology and innovation site in Hyderabad, India, to accelerate digital capabilities across its global operations. The site, named Amgen India, will support the advancement of Amgen’s drug pipeline and is expected to be operational by Q4 2024. Located in HITEC City, Hyderabad, the facility will span six floors of the RMZ Spire Tower 110, accommodating up to 3,000 employees. Hyderabad was selected for its strong talent pool in medicine, life sciences, data sciences, and AI. David M. Reese, M.D., executive vice president and CTO at Amgen, highlighted the significance of the new site, stating, “At a time when a quickly aging global population needs more innovation, the convergence of biotechnology and technology is enabling Amgen to work with greater speed, confidence, and efficiency — an incredibly exciting milestone for which we have been preparing for over a decade.” Amgen India will focus on developing new technology solutions and digital capabilities to enhance operational efficiencies across the enterprise. The site will create roles in AI, data science, life science, and other global capabilities as the operation expands. Som Chattopadhyay has been appointed as the national executive for India, leading the expanded operations. Telangana Chief Minister Sri Anumula Revanth Reddy welcomed the development, stating, “We are proud to welcome a global trailblazer of the biotechnology industry. Amgen’s unwavering mission to serve patients will be incredibly inspiring for the world-class technology talent seeking to make a meaningful impact on people around the world.” Amgen, with nearly 27,000 employees, has a global presence in approximately 100 countries and regions, including India. With more than 40 years in the industry, Amgen continues to advance a broad pipeline of treatments for cancer, heart disease, osteoporosis, inflammatory diseases, and rare conditions.","excerpt":"Som Chattopadhyay has been appointed as the national executive for India, leading the expanded operations.","categories":["AI News"],"tags":["Hyderabad"],"author_name":"Mohit Pandey","publish_date":"2024-08-09T12:03:34","publication_year":"2024","word_count":304,"keywords":["data science","programming_languages:R","AI","innovation","Git","Hyderabad","GAN","R"],"extracted_tech_keywords":["AI","data science","R","Git","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amgen-is-launching-biotechnology-and-ai-hub-in-hyderabad\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":45874,"title":"How To Become A Data Science Influencer On LinkedIn","content":"The term ‘influencer’ has witnessed a meteoric rise in the past couple of years. There are a massive number of users who have joined this ever-growing influencer tribe. While Facebook, Instagram or Snapchat are witnessing people becoming full-time influencers, LinkedIn is also in the race — but on a different lane. LinkedIn is, without doubt, the best platform when it comes to building a professional network. On a daily basis, more than 2 million posts, videos and articles course through the LinkedIn feed. Also, it generates tens of thousands of comments every hour, and tens of millions of shares and likes. And over the past few years, the platform has also witnessed some personalities making a great impact on different niches, delivering real value. So why not to become a LinkedIn influencer? What’s wrong in sharing your knowledge and thoughts extensively to a global audience? It would give you real credibility after all. While there are a lot of monetary as well as other benefits in becoming a social media influencer, LinkedIn influencers also have their own set of benefits that are different than the rest. An official LinkedIn Influencer is selected by invitation only and comprises a global collective of 500+ of the world’s foremost thinkers, leaders, and innovators who discuss newsy and trending topics. This tag of ‘Influencer’ completely changes your status. In this article, we are going to look at some of the effective ways that would take you closer to becoming a LinkedIn Influencer — if not officially, but at least to an extent where you would be known by a huge, significant amount of people. Here Is How You Can Start Your LinkedIn Influencer Journey. Find Your Niche In Data Science It is really important to know what you are good at — what is your forte. Data science is a vast field and you have to narrow things down to make a spot for yourself. The data science domain has evolved so tremendously that type “Data Science” on LinkedIn and see the results. Therefore, make sure you do the self-assessment and understand what you are good at. Simply put, you have to be a subject matter expert of your clearly defined niche. Also, once you figure out your niche and if in case you feel that it’s too obscure or too unique, then always remember that your uniqueness is your advantage. Have Complete Knowledge One doesn’t become a ‘Guru’ by just calling oneself that, you have to have the know-how of the things you’re talking about. Becoming a LinkedIn Influencer doesn’t mean you will only share your thoughts, you will have to help people looking for a solution to their complex problems. Data science today is the backbone of many industries. It provides a lot of intelligence about consumers and campaigns, through techniques like data mining and data analysis. Therefore, make sure you know everything about your niche — whether its programming or other tasks. If you lack knowledge, then I am afraid you won’t be able to thrive — not only as an influencer but as a data science professional as well. Use Hashtags When you are tagged as a LinkedIn influencer you have to produce content that people can use to get their queries answered, get solutions to their problems etc. The content could be in the form of a video or article or just a post with few lines; however, there is one thing that you have to consider in order to make your content visible to the audience. How would you do that? By using hashtags. Instagram and Facebook have always been popular platforms that have witnessed the evolution of hashtags. But now LinkedIn has joined the list as well. There was a time hashtag were only added to posts but now they are tappable and lead to search results so that people can discover other content with the same hashtag. Hashtags on linked work the same way as Twitter and LinkedIn — they categorise the content. Therefore, make sure you add hashtags to your content — that are relevant to data science. Hashtags not only help you get discovered by other users but also help your content get noticed by those who are not connected to you. Talking about hashtags let’s look at another aspect of it — adding hashtags on your profile. This strategy works for Instagram, but for LinkedIn, its little different. It is effective, but not as effective as posts. Still, make sure you have relevant hashtags on your profile as well, because soon this will also become relevant. Engage With Your Connections One of the most important things to getting closer to becoming a LinkedIn influencer is to engage with your connection as well as with people who are not connected to you. Every time you post a blog, make sure you reply to the comments and help them with their query. Also, reply to messages you get; it might get a little hard to get back to each and everyone when you have comparably a big audience, but make sure you reply to the one who genuinely needs to get something resolved. The more you connect and talk to people, the more your chances of becoming a LinkedIn influencer increases. This is actually networking. Moreover, when we are saying talk to as many people as possible, we also mean the ones who are not your connection. So, how would you do that? The best way again is to take the help of hashtags. Pick a hashtag (it could be one of the topics that interests you) and search for it, look for all the content in the results, click that one that interests you the most and there you can comment your thoughts, or you can even reply to people. When you are searching for a hashtag, it would also pop some profiles to you. So, you are interested in sending some connection requests you can do that as well. Content Is King When it comes to making an impact on LinkedIn, where most of the professionals hang out, your content has to be stronger than most of the other ‘Gurus’ in the niche. Talking about content, it’s not about writing each and everything about your data science niche, you have to make sure that you are pushing out content that makes sense and people want to read. You can always experiment with different strategies to come with the next viral content. One of the best ways is to ask your connections about their problems and challenges they are facing with the data science. And try to solve their problems in your upcoming blog post. Or how about doing a video for the same problem? Its all about figuring out that ultimate content that stands out. Also, keep in mind, do not do what others are doing. Uniqueness is the key to get that desired audience base. Influencers To Learn From While you are getting started with your journey to becoming a LinkedIn influencer, follow some of the well-known influencers and watch carefully their activities in the platform on a day to day basis. One of the best examples is Nicholas Thompson, Editor in Chief at Wired. Thompson in his daily video series breaks down and shares the industry news\/stories he finds most interesting and important. He makes sure that his information is not half-cooked, and the topics he talks about are always relevant to the crowd he caters to. Another personality is Gary Vaynerchuk, Chairman, VaynerX & CEO, VaynerMedia. Vaynerchuk’s content ranges from marketing issues to workplace topics, like dealing with stress. His content is also considered to one of the best for people who are starting out with their own venture.","excerpt":"The term ‘influencer’ has witnessed a meteoric rise in the past couple of years. There are a massive number of users who have joined this ever-growing influencer tribe. While Facebook, Instagram or Snapchat are witnessing people becoming full-time influencers, LinkedIn is also in the race — but on a different lane. LinkedIn is, without doubt, […]","categories":["AI Features"],"tags":["Data Science","linkedin","Virtual Influencers"],"author_name":"Harshajit Sarmah","publish_date":"2019-09-12T16:27:13","publication_year":"2019","word_count":1289,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","Virtual Influencers","ViT","linkedin","Data Science","R"],"extracted_tech_keywords":["AI","data science","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-become-a-data-science-influencer-on-linkedin\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10073336,"title":"Tech Research is Facing a Replication Crisis and Everyone is Blaming Everyone","content":"AL\/ML community has been grappling with the replication crisis for a while now. A replication crisis occurs when scientific studies are difficult or impossible to reproduce. ML researchers often compare their work with benchmark research previously done in the field by other researchers. However, issues arise because the source code of the benchmark research isn’t published in many cases. “In my opinion, the ‘ML Replication Crisis’ goes against the ethical principles research was founded upon (i.e., reliability, validity, trustworthiness, replicability. . .), but is also a reflection of what research has become,” said Chantel Perry, Senior Data Scientist at Microsoft. Earlier in 2022, researchers from Princeton University published a paper titled, ‘Leakage and the Reproducibility Crisis in ML-based Science’. The study notes that a research can be termed reproducible only if the codes and data used by the researchers are made available. In 2020, Google Health published a paper in Nature that described how AI was leveraged to look for signs of breast cancer in medical images. However, as noteworthy as the innovation was, Google was criticised for providing little to no information about its code or how it was tested. A replication crisis does not only occur when researchers find it hard to replicate earlier research. Chantel Perry believes the replication crisis may occur due to several reasons, including: proprietary reasons, limitations in project scope, lack of a third-party review process, lack of rigour when utilising existing ML frameworks, pressure to publish perfect studies rather than listing the details of a study (i.e., limitations, biases, recommendations and future studies), lack of rigorous editorial and peer review process, along with several other reasons. Perry further adds that technology research is no longer just a medium to share knowledge, findings, and mistakes to help other researchers and improve society. “In some cases, technology research is a source of profit for organisations, which isn’t necessarily a problem, except for the fact that this could put pressure on researchers to move quickly and result in more errors. I think the ‘ML Replication Crisis’ is just a symbol of how technology research is no longer a primary tool for sharing innovation and societal improvement but is being prioritised more for capital gain.”—Chantel Perry, Senior Data Scientist – Microsoft. The blame game Several recent developments in AI and ML have come from large enterprises like Google, Microsoft, and Meta. The resources that researchers in these organisations have at their disposal are profoundly abundant. In contrast, researchers from a university might not have access to the same resources. Additionally, these organisations often don’t open source the code for their algorithms. For example, text-to-image generator AI DALL-E 2 by OpenAI—a company that Microsoft invested nearly USD 1 billion—has been a revelation, but it’s not open source. Earlier in 2022, a Reddit user published a post titled, ‘I don’t really trust papers out of “Top Labs” anymore’. In the post, the user asked: ‘Why should the AI community trust these papers published by a handful of corporations and the occasional universities? Why should I trust that your ideas are even any good? I can’t check them; I can’t apply them to my own projects.’ Responding to the post, Jathan Sadowski, a senior fellow at Emerging Tech Lab, said, “AI\/ML research at places like Google and OpenAI is based on spending absurd amounts of money, compute, and electricity to brute force arbitrary improvements. The inequality, the trade-offs, the waste—all for incremental progress toward a bad future.” Interestingly, one of the researchers who worked on the Google Health project, Benjamin Haibe-Kains, referred to the research announcement as an advertisement for cool technology with no practical use. However, in recent years, these large organisations have addressed several community concerns. They are adopting a more open-sourced approach where they share the codes for some of the algorithms they have developed. Chantel Perry also believes these large corporations are not solely the issue with the replication crisis. “I’ve read a number of papers which lack any limitations, biases, or future recommendations and improvements sections in the ML space.” “I believe that by having a large research department and thus many publications, it’s easier to point the finger at big tech corporations or large academic departments, but it’s an industry-wide issue with various areas of leakage.” Perry concluded, “One could blame the corporation, researcher, editor, journal publication, other researchers citing non-replicable studies, or even readers who don’t challenge the studies. Regardless, research projects are just like any other project with a limited scope of time, money, and resources.” How to overcome the crisis? The researchers from Princeton University believe that date leakage often leads to severe reproducibility failures. “Data leakage has long been recognised as a leading cause of errors in ML applications. Through a survey of literature in research communities that adopted ML methods, we find 17 fields where errors have been found, collectively affecting 329 papers and in some cases leading to wildly overoptimistic conclusions,” stated the research paper. The researchers believe that fundamental methodological changes to ML-based science could lead to leakage prevention prior to publication. Most of the time, a replication crisis occurs because researchers do not possess the hardware or compute power. Resolving this would go a long way in overcoming the replication crisis in this field. BigScience—a collaboration of over 1000 independent researchers, academics, and industrial researchers is working towards solving this problem. “BigScience aims to make meaningful progress toward solving complex issues and create tools and processes to help a greater diversity of participants,” said Giada Pistilli, Ethicist at Hugging Face. Researchers at Hugging Face have also released ‘BLOOM’—a large language model that is open source and to help solve the problem of hardware, the team will also publish smaller, less hardware-intensive versions. In addition, a distributed system would be created to allow labs to share the model across their servers. Further, Hugging Face will release a web application that will enable anyone to query BLOOM without downloading. Chantel Perry emphasises that there are a few solutions which are already established as practices in research that could help solve the reproducibility issue— “For instance, discussing time constraints, drawbacks of the study, challenging our own findings and following frameworks to minimise error, discussing safeguards for reliability and validity of results, and\/or providing more supplemental resources are not new to technology research.” However, she also notes that these processes take time and would slow down the research and publication process considerably. Further, she states that it might be helpful to elicit a third-party expert entity who reviews, replicates, and identifies risks within the study. Perry also advocates for allowing a system to flag inaccurate studies to help researchers identify the studies that have issues, the nature of those issues and the provisions to amend the study once such issues are identified. Lastly, Chantel Perry also believes that researchers should be required to provide supplemental resources to prove the replicability of published studies. In this context, the Conference and Workshop on Neural Information Processing Systems (NeurIPS) has begun mandating that authors\/researchers produce a ‘reproducibility checklist’ along with their submissions. This checklist consists of information such as the number of models trained, computing power used, and links to codes and datasets. Likewise, another initiative called the ‘Papers with Code’ project was founded with a mission to create free and open-source ML papers, codes, and evaluation tables.","excerpt":"AI already suffers from the black-box problem and the lack of transparency on research only aggravates the issue.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Pritam Bordoloi","publish_date":"2022-08-23T14:00:00","publication_year":"2022","word_count":1225,"keywords":["Go","Hugging Face","API","OpenAI","AI","ML","RAG","Aim","Rust","AI Tool","R"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","Hugging Face","RAG","R","Go","Rust","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/tech-research-is-facing-a-replication-crisis-and-everyone-is-blaming-everyone\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10053002,"title":"How to Use Graph Neural Networks for Text Classification?","content":"The graph neural networks are trending because of their applications in a variety of predictive analytics tasks. When it comes to modelling the data available with graphical representations, graph neural networks outperform other machine learning or deep learning algorithms. In the field of natural language processing as well, graph neural networks are being applied in a full swing because of their capabilities to model complex text representations. In this article, we will discuss one such interesting application of graph neural networks, i.e., in text classification. First, we will understand how this framework works to model the text representations and then we will explore how it can be used for text classification. The major points that we will discuss in this article are listed below. Table of Contents Deep learning in text classification Graph Neural Networks Graph Convolutional NetworkText Graph Convolutional Networks (Text GCN)Implementation of Text GCN in Python Deep learning in Text Classification As we know that we can roughly divide the deep learning studies into two major models one is convolutional neural networks and another one is recurrent neural networks. If we talk about the text classification the studies here also can be divided into two groups where one is focused on the make model which can learn based on the word embedding. Three are various studies that have shown us that the success of any text classification model depends on the effectiveness of the word embeddings. Another group is focused on learning the document and the word embedding together. If we talk about the models CNN and RNN both can be used for text classification. But the CNN is good with the one-dimensional convolutional and is majorly used in the computer vision field and a special type of RNN that is LSTM (long short term memory) models can be used for better performance in the text classification. The further extension of these models is done using mechanisms, like attention mechanisms which have increased the flexibility of text representation and can be used in the deep learning model as an integral part. Deep learning methods and their extensions are widely used. The major benefit of these models is they focus on the local consecutive word sequence for text classification, unlike traditional methods which use global co-occurrence information of the words for learning saved in a corpus. Graph Neural Networks In recent times the applications of graph neural networks are growing rapidly in multiple domains. There are a number of studies in which we can see that number neural networks such as CNN are generalized and can be applied to the regular grid structure which is helpful in working on arbitrarily structured graphs. There are multiple Graph Neural Networks in the field of text classification. Before going for any of the networks let us understand the graph neural networks first which will make a clearer picture of the network in front of us. Graph The graph is a data structure that consists of vertices and edges. It can be represented as the function of vertices and edges. G = ( V , E ) The below image is a representation of the direct graph. The type of the edges in the graph de[ends on the directional dependencies between the vertices. And the edges can be of two types: directed or undirected. Method of Graph Neural Network By the name, we can understand if a neural network operates on the graph we can call it a graph neural network where the major operation of any neural network is to classify the vertices or nodes. So that every node presented in the graph can be classified by their provided labels according to the neural network. If there is a node v and its features can be characterized by the x_v at ground truth t_v given in a labelled graph G so we can label an unlabeled graph using this labelled graph where a d dimensional vector helps in learning and h-v contains the information of its neighbourhood. Mathematically Image source If any reader wants to learn more about graphical neural networks can check this paper. As we have discussed before in the deep learning section, we can perform text classification using CNN or RNN. and also how LSTM is good for text classification we can check on this article. Also in graphical neural networks, we can either use a graphical recurrent network or a graphical convolutional network. Since in this article we are going to talk about a model which is basically a graphical convolutional network so we will start with the graphical convolution network and then we will see how we can use it. Graph Convolutional Network We can say if a convolutional neural network is directly used with the graph for operating and making predictions we can call it a graph convolutional network (GCN). more formally a convolutional neural network inducing the embedding vectors of nodes which are dependent on the property of the neighbourhood. Let’s say a graph as: G = (V, E), Where, V (|V | = n) and E are sets of nodes and edges, respectively. Every node is assumed to be connected to itself, i.e.,(v, v) ∈ E for any v. X is a matrix containing n x m nodes and their features A is an adjacency matrix of G D is the degree matrix of A. So for mathematical a k-dimensional nodes feature matrix can be calculated as: Where the convolutional network is one dimensional and, Where the convolutional network is one dimensional. W0 ∈ Rm×k is a weight matrix p is an activation function like ReLU and j is number of layers and if j = 0 then L = X Text Graph Convolutional Networks (Text GCN) From the above, we have seen how the traditional and deep learning model works on the text documents and here the graph neural network or graph convolutional network(in our case) is expected to make a text graph heterogeneous which can model global co-occurrence sequence of words. The heterogeneous text graph contains the nodes and the vertices of the graph. Text GCN is a model which allows us to use a graph neural network for text classification where the type of network is convolutional. The below figure is a representation of the adaptation of convolutional graphs using the Text GCN. . Image source Where the addition of the corpus size and the number of unique words is equal to the number of nodes on the text graphs. The model in Text GCN takes the input in the form of an identity matrix so that every word can be represented as the one-hot vector. To generate the TF-IDF (term frequency-inverse document frequency) of the word in the document the model generates the edges among nodes based on the word occurrence in the global corpus. Like in the traditional way in TF-IDF the term frequency represents the number of occurrences of the word in the document.to gather the co-occurrence statistics the model supplies a fixed size window on the documents in the corpus and the sliding of the window makes the global word co-occurrence information useful for prediction and classification. Mathematically the weight of an edge between node i and node j is defined as Here,  W represents the total sliding window and W(i) is the number of sliding windows that slide on the corpus of i words. Positive pointwise mutual information represents the high correlation between the words in a corpus. This is how the relationship between the words in the corpus is used for making a text graph. Now, this text graph can be fed into the GCN. In the Text GCN, the model which will operate on the text graph has two layers. The final layer has the same embedding which is labelled and can be fed into the softmax classifier. Where The two-layered GCN is better performing than the one layered GCN and also it allows using two-layered GCN to pass information among the nodes that are a maximum of two words away from each other. Which is also used to make a graph without direct document-document edges. When we talk about the modeling procedure the Text GCN model works on these three major steps. Preparation of the data – In this the repository of the model contains functions that can automatically perform most of the basic tasks of the NLP modeling procedures like cleaning data, removing stopwords, etc. Preparation of text graph – Using the cleaned document, inbuilt functions in the repository helps in making the graph in a similar way which we have discussed earlier in this topic. The generation of the graphs is dependent on the PMI which tells about the semantic correlation in the corpus. Training of the model – In this step the main model for which the text graphs are prepared to get trained on the graphs. This model is basically a two-layered convolutional network that is tuned to perform on the graph structure data. For more information about the model Text GCN, a reader can go to this github repository. In the repository, we can get the full guidance to use the model with python, and also there are tutorials available for data preparation for the model. Final Words In the article, we have seen how deep learning works on text classification problems and after that how the extension of deep learning can approach text classification. In many domains, the graph neural networks are working fine and can also be used for NLP modelling. There are some of the graph neural networks available for text classification. The Text GCN model is one of them that can be used for text classification that we tried to understand in this article.","excerpt":"There are a number of studies in which we can see that neural networks are generalized and can be applied to the regular grid structure which is helpful in working on arbitrarily structured graphs","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","graph convulationsl networks","graph neural networks","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2021-11-07T13:14:47","publication_year":"2021","word_count":1620,"keywords":["text classification","machine learning","AI","neural network","graph convulationsl networks","Machine Learning","computer vision","NLP","Python","deep learning","analytics","graph neural networks","Deep Learning","Data Science","Data Scientist","predictive analytics","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","NLP","computer vision","analytics","predictive analytics","text classification","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-use-graph-neural-networks-for-text-classification\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10097198,"title":"8000 Authors Petition OpenAI, Google, Meta and Microsoft to Pay for AI Plagiarism","content":"In a move towards tackling AI plagiarism, over 8,000 authors have collectively expressed their concerns in an open letter initiated by the US Authors Guild, urging the leaders of six major AI companies to seek consent and offer compensation for using their copyrighted work to train models. The letter, directed at the CEOs of OpenAI, Alphabet, Stability AI, Meta, IBM, and Microsoft, asserts that the existence of generative AI technologies, based on large language models (LLMs), is indebted to the writings of these authors. “Generative AI technologies built on large language models owe their existence to our writings.” The authors argue that these AI systems mimic and reproduce their language, stories, styles, and ideas, effectively benefiting from millions of copyrighted books, articles, essays, and poetry without any compensation. The Authors Guild’s CEO, Mary Rasenberger, said that the purpose of the letter was to encourage companies to reach settlements with authors outside of the courtroom, as lawsuits can be expensive and time-consuming. However, some authors have chosen a more aggressive approach and filed lawsuits against those they believe have plagiarised their work. Open Letter to Generative AI Leaders Similarly, in May, the Writers’ Guild of America (WGA) protested against the use of AI in writing scripts for movies. WGA’s lead negotiator Ellen Stutzman also highlighted during the protests that some of their members refer to AI as “plagiarism machines“. The struggle between proper attribution and copyright about generative AI technology is ongoing, and needs a resolution soon. LLMs like GPT, and even the recent Llama-2 by Meta, have been built by scraping information across the internet, which includes most of the websites on the internet. To tackle this, OpenAI has recently signed agreements with other organisations to access data for training its generative AI systems. For instance, they struck a deal with the Associated Press, granting them access to text archives dating back to 1985, while the news agency receives access to OpenAI’s technology and expertise. Also Read: Is AI Copyright Really Necessary?","excerpt":"“Generative AI technologies built on large language models owe their existence to our writings,” said the letter.","categories":["AI News"],"tags":["ai copyright","plagiarism"],"author_name":"Mohit Pandey","publish_date":"2023-07-19T14:35:36","publication_year":"2023","word_count":332,"keywords":["Go","API","plagiarism","OpenAI","AI","AWS","RAG","GPT","generative AI","GAN","ai copyright","R"],"extracted_tech_keywords":["AI","generative AI","OpenAI","RAG","AWS","R","Go","API","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-google-meta-microsoft-asked-to-pay-8000-authors-for-ai-plagiarism\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":44577,"title":"Lendingkart Technologies Raises $30 Million In Fresh Equity","content":"Lendingkart Technologies Private Limited, India’s leading financial technology company has raised a fresh equity round of INR ₹21 crore ($30 million) led by existing investors including Fullerton Financial Holdings Pte Ltd, Bertelsmann India Investments and India Quotient signalling a strong support and trust in Lendingkart’s vision of financial inclusion and digital accessibility for everyone. With this funding round, the total equity raised by Lendingkart stands at $143 million. The funding will be used to expand lending bases, deepen Lendingkart Group’s reach to small and underserved micro-enterprises and further strengthen its technological and analytics capabilities. Credit Suisse acted as the exclusive financial advisor to Lendingkart Technologies. Harshvardhan Lunia, Co-Founder & Managing Director of Lendingkart Technologies, said, “We are pleased with the momentum in our business and this equity will help us meet the growth opportunities we are seeing. Micro and small businesses represent a vibrant yet underserved segment of the Indian economy. The support of all of our customers, investors and employees is empowering us to build the leading financial services platform for this segment.” Commenting on the investment in Lendingkart Technologies, Yeo Hong Ping, President, Fullerton Financial Holdings said, “Over the past year, Lendingkart’s business has seen robust growth and continues to exhibit great potential. We are delighted to continue partnering with Lendingkart who has been at the forefront of making working capital digitally accessible for MSMEs, as India continues its journey towards becoming a digitally enabled country.” Pankaj Makkar, Managing Director, Bertelsmann India Investments said, “Digital lending is revolutionizing access to capital for MSMEs. With focused efforts of the Government and technological advances such as big data, digital lending has set the stage for disruption with greater formalisation, faster capital disbursements and vastly improved customer experience. We are proud to be a part of this incredible journey with Lendingkart as it continues to demonstrate market leadership in this space.” Since its inception, Lendingkart Finance (the NBFC arm of Lendingkart Group) has evaluated nearly half a million applications, disbursing 60,000+ loans to more than 55,000 MSMEs in 1300+ cities across all 29 states and union territories of the nation, making it the NBFC with the largest geographical footprint in the country. Aligned with the Government of India’s agenda of building financial inclusion, Lendingkart Finance works towards ensuring availability of credit for small and micro enterprises across the country that either does not have access to credit or are capital deficient currently. Lendingkart leverages robust in-house technology tools based on big data analytics and machine learning algorithms to evaluate creditworthiness. By analysing thousands of data points to assess factors like financial health, comparative market performance, social reliability and compliance and a distinctive evaluation process, Lendingkart aims to disburse loans with minimal paperwork within 72 hours.","excerpt":"Lendingkart Technologies Private Limited, India’s leading financial technology company has raised a fresh equity round of INR ₹21 crore ($30 million) led by existing investors including Fullerton Financial Holdings Pte Ltd, Bertelsmann India Investments and India Quotient signalling a strong support and trust in Lendingkart’s vision of financial inclusion and digital accessibility for everyone. With […]","categories":["AI News"],"tags":["Funding"],"author_name":"Prajakta Hebbar","publish_date":"2019-08-16T16:40:08","publication_year":"2019","word_count":454,"keywords":["Go","API","Funding","machine learning","AI","Git","RAG","Aim","analytics","Rust","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","R","Go","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/lendingkart-technologies-raises-30-million-in-fresh-equity\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10123699,"title":"Just How Easy is it to Publish Malicious Extension on VSCode?","content":"Recently, a group of Israeli researchers were able to create and publish a malicious VSCode extension in 30 minutes. Surprisingly, the extension was trending, and had 100+ downloads within the first 24 hours, shockingly exposing the vulnerability of the platform. Built with Flaws The experiment showcased that over 1,280 extensions had malicious dependencies packaged in them with a combined total of 229 million installations. Also, there were 87 extensions that attempted to read the \/etc\/passwd file on the host system. Amit Assaraf, the co-founder of real estate investing app Landa, and one of the Israeli researchers who experimented to expose the flaw, said, “Unlike Google Chrome extensions, VSCode extensions are practically apps\/executables running on your machine with zero limitations on what they can do on the host. “This means extensions pose a threat to spawn child processes, system calls or even import any NodeJS package they’d like.” In the experiment conducted by Assaraf and his team members Itay Kruk, and Idan Dardikman, it was found that 2,304 extensions used another publisher’s GitHub repo as their official listed repository. This means you have no way to tell if the code of your extension and the linked GitHub repository are the same. A major flaw in the popular open-source code editor VSCode is that extensions are not sandboxed. Though there’s a provision for sandboxing code, it is not applicable for extensions. Since extensions are not sandboxed, they can access anything inside the IDE and execute anything on the host machine without the developer receiving any feedback. Apart from sandboxing, VSCode also lacks permission management. You will find multiple feature requests to add permission manager in VSCode, similar to what we have in our smartphones to know that is being accessed by extensions. This way, a theme extension that is built to change colours of IDE, may execute code and read or write files without any visibility or explicit authorisation from the user. Auto Update of Extension VSCode extensions automatically update to the latest version behind the scenes. This means any developer can initially create an extension without any malicious intent and later update the extension where he can introduce the malicious code. The same happened with xz utility, which was safe for years and later found to have a backdoor. $5 to Verify VSCode Extension Extension verification allows extension authors to verify the ownership of a domain to establish their authenticity and credibility, with the verified domain displayed alongside their name. Isidor Nikolic, a senior product manager at Microsoft, VS Code, mentions that extension authors can become verified by checking the ownership of an eligible domain associated with your brand or identity. Interestingly, to verify your extension, all you have to do is get a domain for your extension (which usually costs around $5) and soon you’ll receive a badge from VSCode suggesting that your extension is verified. A GitHub user critiqued the process, calling the process for verification badge as totally useless and misleading. “The verified blue check mark merely means that whoever the publisher is has proven the ownership of a domain. And that means any domain. In reality, a publisher could buy any domain and register it to get that verified check mark,” he said. Sure, there are manual steps involved in the verification process but these are not rock solid. You can use a different name while applying for verification and as soon as you get verified, it can be changed to look exactly like the original name of the extension. “Security is not a high priority for Microsoft as they want as many extensions on their marketplace as possible,” the Israeli researcher also mentioned further. Solutions to Navigate Vignesh Rajan, a lead engineer at GenAI startup MachineHack, suggested, “You may use VSCode but with as few extensions as possible to minimise the risk. If necessary, a developer should do a thorough research on which extension is official and can be trusted before installation.” By nature, VSCode heavily relies on extensions to enhance its functionality. It is a bare-bone IDE where developers can install extensions of their choice to get the job done. You may also switch to closed-source IDE such as IntelliJ. A user while praising IntelliJ for Java development mentioned that “VS Code is not only unsuitable (still) for larger enterprise-level projects, it is also less reliable, responsive and stable than IntelliJ.” On November 18, 2015, the source code of VS Code was released under the MIT License and made available on GitHub. The idea was to create a lightweight platform powered by extension and it was an instant success.","excerpt":"There are 1,283 malicious extensions on VSCode with a combined total of 229 million installs.","categories":["Deep Tech"],"tags":["Microsoft","VS Code","Vulnerabilities"],"author_name":"Sagar Sharma","publish_date":"2024-06-14T17:13:40","publication_year":"2024","word_count":764,"keywords":["Go","GenAI","AWS","AI","Vulnerabilities","Git","VS Code","Rust","GitHub","R","Java","Microsoft","startup"],"extracted_tech_keywords":["AI","GenAI","AWS","R","Go","Rust","Java","Git","GitHub","startup"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/just-how-easy-is-it-to-publish-malicious-extension-on-vscode\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10061459,"title":"Pytorch introduces TorchRec, an open source library to build recommendation systems","content":"CEO Mark Zuckerberg has introduced TorchRec, an open source library for building state-of-the-art recommendation Systems under PyTorch, at Inside the Lab event. The new library provides common sparsity and parallelism primitives, enabling researchers to build state-of-the-art personalisation models and deploy them in production. It  includes a scalable low-level modelling foundation alongside rich batteries-included modules. Meta initially targeted“two-tower” architectures with separate submodules to learn representations of candidate items and the query or context. Input signals can be a mix of floating point “dense” features or high-cardinality categorical “sparse” features that require large embedding tables to be trained. Efficient training of such architectures involves combining data parallelism that replicates the “dense” part of computation and model parallelism that partitions large embedding tables across many nodes. The library includes optimised Recommendation Systems kernels that run on FBGEMM, a high-performance kernel library, modelling primitives such as jagged tensors and embedding bags, to create multinodal models using model parallelism. The PyTorch team has released TorchRec after close to two years in testing. TorchRec was used to train a model with 1.25 million parameters that went into production in January and a 3 trillion parameter model that is expected to go into production. Cautiously excited about @PyTorch launching TorchRec. This might be just what we need!— Nirbaay Tandon (@nirbaaytandon) February 23, 2022 Excited to see this https:\/\/t.co\/eqE11ydI5H from @PyTorch— Josh Patterson (@datametrician) February 23, 2022","excerpt":"TorchRec was used to train a model with 1.25 million parameters that went into production in January.","categories":["AI News"],"tags":["Open Source AI"],"author_name":"Poulomi Chatterjee","publish_date":"2022-02-24T13:10:51","publication_year":"2022","word_count":229,"keywords":["Replicate","Go","programming_languages:R","PyTorch","AI","recommendation systems","Scala","Open Source AI","ai_frameworks:PyTorch","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","PyTorch","recommendation systems","R","Go","Scala","Replicate","ai_frameworks:PyTorch","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-introduces-torchrec-an-open-source-library-to-build-recommendation-systems\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":12447,"title":"CMU’s AI system ‘Libratus’ beats top players at Texas Hold&#8217;em poker, creates history!","content":"After 20 days and a win of $1.5M worth of chips, the winning numbers indicates a Poker champion giving a hard time to the competitors. But in reality it’s the Artificial intelligence that stole the show. What makes the result even more interesting is the fact that the competitors happen to be four of the world’s best professional poker players! After having proven its mettle in games like checkers, chess and Go, Artificial intelligence is taking everyone by surprise by defeating the world champions at the Poker championship being held in a Pittsburg casino. Named “Libratus”, this AI champ turned out to be the first ever AI program to have beaten humans at Poker. A result of an extensive research by Tuomas Sandholm, professor of computer science at Carnegie Mellon University, and his Ph.D. student, Noam Brown, Libratus is being used in this contest to play poker, an imperfect information game that requires the AI to bluff and correctly interpret misleading information to win. This makes poker a tougher game to crack than chess and Go. The researchers believe that Libratus win this competition has opened avenues for solving cases involving complicated decision making and tasks based on imperfect information such as negotiating business deals, setting military strategy or planning a course of medical treatment. Professional poker player Jason Les Brains Vs. AI, which began Jan. 11 at Rivers Casino in Pittsburgh, saw Chou and three other leading players- Dong Kim, Jason Les and Daniel McAulay, competing against Libratus in a 20-day contest in which they will play 120,000 hands of Heads-Up, No-Limit Texas Hold’em poker. All four pros specialize in this two-player, unlimited bid form of Texas Hold’em and are considered among the world’s top players of the game. One of the pros, Jimmy Chou, said he and his colleagues initially underestimated Libratus, but have come to regard it as one tough player. “The bot gets better and better every day,” Chou said. “It’s like a tougher version of us.” In a similar Brains Vs. AI contest which was held in 2015, four leading pros amassed more chips than the AI, called Claudico. But Sandholm said he’s feeling good about Libratus’ chances as the competition proceeds. “The algorithms are performing great. They’re better at solving strategy ahead of time, better at driving strategy during play and better at improving strategy on the fly,” Sandholm said. CMU professor Tuomas Sandholm Pittsburgh Supercomputing Center’s Bridges computer are playing an important role by performing computations to sharpen the AI’s strategy. During the day’s game play, Bridges is used to compute end-game strategies for each hand. Sponsored by GreatPoint Ventures, Avenue4Analytics, TNG Technology Consulting GmbH, the journal Artificial Intelligence, Intel and Optimized Markets, Inc., this championship has every eye on it to see if AI stands out to be the best!","excerpt":"After 20 days and a win of $1.5M worth of chips, the winning numbers indicates a Poker champion giving a hard time to the competitors. But in reality it’s the Artificial intelligence that stole the show. What makes the result even more interesting is the fact that the competitors happen to be four of the […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-01-31T12:59:07","publication_year":"2017","word_count":468,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cmus-ai-system-libratus-beats-top-players-texas-holdem-poker-creates-history\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070399,"title":"Torcharrow: A PyTorch framework for CPU-based large data processing","content":"With the release of the new version of Pytorch 1.12, Pytorch has come up with a new DataFrame library for data visualization or preprocessing named Torcharrow. Torcharrow is a Pytorch library for data processing and visualization with support for processing tabular data and is more suitable for deep learning data. Torcharrow has come up with the ability of faster processing of data by lighter usage of the processing unit. In this article let us get a brief overview of the latest preprocessing library of Pytorch 1.12 named Torcharrow. Table of Contents An overview of TorcharrowBenefits of TorcharrowData processing using TorcharrowSummary An overview of Torcharrow Pytorch, an open-sourced machine learning and deep learning framework based on the torch library is used in various applications like computer vision and Natural Language processing. PyTorch came up with the release of a new version Pytorch 1.12 on June 28, 2022. With the release of the new version, Pytorch has created a new API for a faster and more efficient data processing library named Torcharrow which is still in the beta stage with more features to be added. Torcharrow is the data processing library that aims to handle and process data with minimal requirement of resources and less weight enforced on the central processing unit. Are you looking for a complete repository of Python libraries used in data science, check out here. Torcharrow follows the same hierarchy and operating characteristics as the Pandas library with a similar ability for data processing. Torcharrow in the beta stage provides data processing with various aspects such as data addition, data manipulation, statistical analysis of data along with querying data with respect to SQL queries. Once the stable version is released hopefully, all the necessary processing steps would be supported by Torcharrow. Benefits of Torcharrow The Torcharrow library of data processing comes with various advantages in efficient data handling and processing. They are: Torcharrow supports various dimensions of data right from single columnar data to multi-columnar data like dataframe.Torcharrow supports various types of data like numbers, strings, and lists.Torcharrow aims to support complex torch data with minimum resources and run flawlessly with devices utilizing only the CPU.Easy integration and logging with respect to Pytorch DataLoader and Datapipe. A complete overview of data processing using Torcharrow Let us first install the Torcharrow library in the working environment. !pip install --user torcharrow import torcharrow as ta import torcharrow.dtypes as dt import torcharrow.expression as exp import warnings warnings.filterwarnings('ignore') Now the torcharrow library is installed and loaded in the working environment. Let us start exploring the Single dimensional data supported by Torcharrow. 1-Dimensional data processing using Torcharrow Similar to the pandas Series, Torcharrow supports single-dimensional data processing by using the Column function. So let us see how to process data using the Column function of Torcharrow. Creating a column col1=ta.Column([1,2,3,4,5,None]) col1 The Column function of Torcharrow has to be created using the Torcharrow instance and the Torcharrow column considers the value as integer values which reduces the memory occupancy and the Torcharrow Column function has the ability to retrieve the count of null values in the output along with the length of the Column and the datatype of that Column. Common column operations In the beta release of Torcharrow, there are two operations being supported by the Column functions and they are as shown below. Computing the length The length of the Column can be computed using the “len” function which provides information on the number of rows in the data frame. col2=ta.Column([1.1,2.2,3.3,4.4,5.5,None]) len(col2) ## To retrieve the length of the particular So here there are 6 rows in the Column datatype of Torcharrow. Computing the count of null values The number of null values in the data frame can be computed using the null_count function of the Torcharrow library as shown below. col2=ta.Column([1.1,2.2,3.3,4.4,5.5,None]) col2.null_count ## To obtain the count of null values in the column Here “None” in the Column datatype is considered to be the null value. Creating a Torcharrow column with variable string length Torcharrow supports variable length strings that can be passed onto the Column data type. str_col1=ta.Column([['Torcharrow','Column'],['Torcharrow','is','still','in','beta-stage']]) str_col1 The strings passed to the Column data type, by default are considered as List data types. The type of the variable length string created can also be retrieved using the type function. type(str_col1) Appending a single value to the Column dataframe New value addition can be done using the append function of the Column dataframe of Torcharrow wherein both single values and multiple values can be appended at the same time. str_col1=ta.Column([['Torcharrow','Column'],['Torcharrow','is','still','in','beta-stage']]) So for the above-created Column datatype let us first see how to append a single value. str_col1=str_col1.append([['Torcharrow','is','faster','and','efficient']]) str_col1 Appending multiple values to the Column Dataframe In a similar manner, multiple values can be appended using various list instances in the single append function as shown below. str_col1=str_col1.append([['My','name','is','ABC'],['I','reside','at','XYZ']]) str_col1 Working with Torcharrow Dataframe Torcharrow data frames are similar to pandas dataframe but as Torcharrow is still in the beta stage and the Torcharrow dataframe still does not have the ability to read data of different formats like CSV, text, and HTML files. So let us see what all processing can be done using the beta stage of the torcharrow dataframe. Creating a Torcharrow dataframe A Torcharrow dataframe can be created using the inbuilt function of Torcharrow as shown below. df = ta.DataFrame({\"Col1\": list(range(10,10+10)), \"Col2\": list(reversed(range(20,20+10))), \"Col3\": list(range(30,30+10))}) df Retrieving the columns of Torcharrow dataframe The columns of the Torcharrow dataframe can be retrieved using the columns function. df.columns Data Retrieval from the dataframe The Torcharrow dataframe facilitates the head and tail function wherein the first and last few entries of the dataframe can be retrieved accordingly. df.head(3) ## Retrieving the first 3 entries of the torcharrow dataframe df.tail(3) ## Retrieving the last 3 entries of the torcharrow dataframe So using the head and tail function the first and last few entries of the dataframe can be retrieved. Adding a new column to the Torcharrow dataframe Similar to the Pandas module a new column can be added to the Torcharrow dataframe where in the new column name to be added will be specified along with the values to be added. df['Col4']=ta.Column(list(range(41,41+10))) df So here we can see that a new column is being added to the original dataframe Adding rows to the Torcharrow dataframe Rows can be added to the Torcharrow dataframe using the append function as shown below. df=df.append([(10,100,101,102),(11,110,111,112)]) df Manipulating values of Dataframe The values of the dataframe can be manipulated by using any of the mathematical operators or any functions. Let us see how to manipulate the value of the dataframe using addition operation. df['Col1']=df['Col1']+50 df Here we can see that each value of Column1 50 is being added. Selection operations Torcharrow supports both string-based and integer-based selections along with slicing. Let us see how Torcharrow can be used for different selection operations. String based selection The column name required has to be mentioned in square brackets for string-based selection. df['Col1'] Slicing: String-based selection In a similar way through slicing required columns can be retrieved. df['Col1':'Col3'] Integer based selection For integer-based selection, the rows required for retrieval have to be specified. df[1] Slicing: Integer-based selection The required rows can be specified in the square brackets for retrieval where in the last value will be exclusive. df[1:5] Condition-based selection For condition-based selection, the required column to check along with the condition to validate has to be specified which will return a boolean output. df['Col1']>65 ## returns a boolean output If the values for the condition have to be retrieved the dataframe object has to be used along with the condition. df[df['Col1']>65] ## Dataframe values for the specified condition is retrieved Handling missing values Using the Torcharrow data frame the missing values can be imputed with the required value or the missing value can be dropped. Let us see how to impute any missing value with the required value. s=ta.Column([1,2,3,None,5]) s=s.fill_null(4) s In a similar manner, the entire row with the missing value can be removed. s.drop_null() Case conversion operations The entire string can be converted to uppercase using the upper function. str_col=ta.Column(['Welcome to Torcharrow','Today is a beautiful day']) str_col.str.upper() The same string can also be converted to lowercase using the lower function. str_col.str.lower() Replacing characters The string characters can be replaced in the Torcharrow library using the replace function. str_col.str.replace('W','A') Splitting characters Huge string characters can be split into smaller string characters using the split function. split_str=str_col.str.split(sep=' ') split_str Using one of the inbuilt functions Let us use the reduce inbuilt function that is being supported by Torcharrow to reduce the sequence of numbers to a single value. import operator ta.Column([5,6,7,8]).reduce(operator.mul) Querying Torcharrow dataframe similar to SQL Query Let us create a Torcharrow dataframe and query the dataframe using the where clause. sel_df = ta.DataFrame({'A': ['a', 'b', 'a', 'b'],'B': [1, 2, 3, 4],'C': [10,11,12,13]}) sel_df.where(sel_df['C']>11) Summary Torcharrow is one of the beta stage libraries of the Pytorch 1.12 version where some required processing such as data retrieval, data addition, and data manipulation is provided with respect to Python based approach. Basic SQL querying is also provided in the beta stage. Torcharrow is designed to be more memory efficient and is focused to process huge data in the central processing unit. So a stable release of the library is expected to support data reading of various formats, data addition, and manipulation in different ways, and also support various SQL clauses. References Pytorch Official DocumentationTorcharrow official DocumentationOfficial Github Repository","excerpt":"Torcharrow is a Pytorch preprocessing library for data processing and visualization with various aspects of data processing.","categories":["AI Trends"],"tags":["data processing","Pytorch"],"author_name":"Darshan M","publish_date":"2022-07-05T12:00:00","publication_year":"2022","word_count":1569,"keywords":["Pytorch","data science","machine learning","TPU","AI","PyTorch","data processing","ML","computer vision","Aim","deep learning","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","data science","Aim","PyTorch","Pandas","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/pytorch1-12-library-for-data-processing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10119456,"title":"Bloomberg Partners With AppliedXL, To Use AI in Generating Stories for Terminal Users","content":"Bloomberg has partnered with New York-based computational journalism startup AppliedXL to provide insights and predictive outcomes to their Bloomberg Terminal users. As part of the collaboration, AppliedXL will parse publicly available data to structure news stories that predict early trends and provide relevant market analysis. The startup, as of now, will focus solely on providing insights from the pharmaceutical industry. Under the Hood The startup, which makes use of AI in analysing data and detecting trends, will help Terminal users get ahead of potential catalytic events within the industry. Similarly, the AI will also alert users to any anomalies or irregularities that could potentially affect the market as a whole. One such example is AppliedXL’s detection of anomalies during a set of clinical trials by biopharmaceutical company Summit Therapeutics using publicly available data on the NIH public trial registry. The company later suffered a significant drop in shares due to the initial failure of its then-sole drug candidate, which was predicted months earlier by AppliedXL’s AI. The startup will provide similar insights to users who have access to Bloomberg Terminal. In addition, the AI has been trained in regular editorial and journalistic practices, which will be exercised during the creation of their news stories. The company is expected to access over 7,000 updates daily on ongoing clinical trials and generate upwards of 60 stories on the most relevant developments of the day. This will include both domestic and international interventional clinical trials. “There’s no artificial intelligence without human wisdom. We use machines to understand the patterns in data, but we need humans to understand the contexts that influence them,” said AppliedXL CEO Francesco Marconi. According to the company, to do so, their AI has been developed in collaboration with journalists during the training process “to review data, help develop interpretations, and validate the quality of the output.” This is one instance of a growing number of media companies relying on AI. Earlier this year, Bloomberg released a paper on their own LLM, BloombergGPT, focused on the financial industry. On the other hand, media companies have also begun collaborating with big tech companies like OpenAI to use their reporting. Most recently, OpenAI partnered with the Financial Times in a licensing agreement, with the latter’s journalistic work now cited when using ChatGPT.","excerpt":"The startup will help users get ahead of potential catalytic events within the industry.","categories":["AI News"],"tags":["Bloomberg","Journalism AI"],"author_name":"Donna Eva","publish_date":"2024-05-02T18:45:00","publication_year":"2024","word_count":380,"keywords":["Go","ChatGPT","Journalism AI","artificial intelligence","TPU","OpenAI","AI","GPT","Bloomberg","llm_models:GPT","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","ChatGPT","OpenAI","TPU","R","Go","GPT","startup","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bloomberg-partners-with-appliedxl-to-use-ai-in-generating-stories-for-terminal-users\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10059434,"title":"Major data distributions a data scientist should know","content":"Statistics forms the foundation of data science. It is absolutely necessary for anyone trying to build a career in data science to have a good hold over the concepts of statistics and understand how they can be applied in business settings. Different distributions of data and their properties are one such area of statistics in which a data scientist has to have crystal clear clarity. Let us take a look at a few of the most common distributions a data scientist encounters in their career. Normal distribution In a normal distribution, the data is arranged in a way that most of the values form a cluster in the middle and taper off in a symmetric fashion towards either extreme. It is also called a Gaussian distribution. It appears as a bell curve when shown graphically. In a standard normal distribution, the mean is zero, and the standard deviation takes the value of 1 along with a zero skew. The mean, median and mode are all the same in a normal distribution. In a normal distribution, the midpoint has the maximum frequency. In normal distributions, there is a constant proportion of the area under the curve lying between the mean and any given distance from the mean when they are measured in terms of standard deviation units. Normal distributions are represented in standard scores or Z scores. These scores give an idea of the distance between  an actual score and the mean in terms of standard deviations. Bernoulli distribution In a Bernoulli distribution, there are two possible values for the random variable (A random variable is a variable whose value depends on the outcome of an experiment). They are of two types – discrete and continuous. A Bernoulli distribution is a discrete distribution. It has two possible outcomes and a single trial (called a Bernoulli trial). A Bernoulli trial is one of the simplest experiments conducted in statistics. It comes with two possible outcomes of success and failure. Some examples of bernoulli trials include coin tosses, rolling a dice, etc. The probability values of mutually exclusive events that make up all the possible outcomes has to sum up to one. The two possible outcomes in the Bernoulli distribution are indicated by n=0 and n=1. Here, n=1 indicating success has a probability p and n=0 indicating failure has a probability 1-p (0<=p<=1). Uniform distribution Uniform distribution is one of the simplest statistical distributions to understand. It is a probability distribution in which all the possible outcomes are equally possible to occur. Graphically, we can think of it as a straight horizontal line. Uniform distributions are of two types – discrete and continuous. A discrete uniform distribution will have a finite number of outcomes, while a continuous uniform distribution will have an infinite number of measurable outcomes that are equally likely. Poisson distribution A Poisson distribution is a probability distribution that shows how many times an event is likely to occur over a fixed period of time and space. It is named after French mathematician Siméon Denis Poisson. It is a discrete distribution where the variables take only specific values. It is a limiting process of the binomial distribution. T-distribution It is a type of normal distribution used mainly for smaller sample sizes, and population standard deviation is unknown. It is also known as Student’s t-Distribution – it is also bell-shaped and symmetrical with zero mean. The shape undergoes a change with the change in degrees of freedom. It has a greater dispersion than the standard normal distribution. As the degrees of freedom increase, the closer the distribution starts to approximate a standard normal distribution. The student distribution ranges from –∞ to ∞ (infinity). Some important applications of T-distribution include the Test of the Hypothesis of the population mean, Test of Hypothesis of the difference between the two means and Test of Hypothesis of the difference between two means with dependent samples. Log-normal distribution A log-normal distribution is a probability distribution of a random variable that has its logarithm normally distributed. A random variable of log-normal distribution takes only positive real values. A random variable that is log-normally distributed will only consider positive real values.","excerpt":"Different distributions of data and their properties is one such area of statistics in which a data scientist has to have crystal clear clarity.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Machine Learning","normal distribution"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-01-30T16:00:00","publication_year":"2022","word_count":692,"keywords":["data science","Go","programming_languages:R","AI","AI (Artificial Intelligence)","Machine Learning","programming_languages:Go","Data Science","R","normal distribution"],"extracted_tech_keywords":["AI","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/major-data-distributions-a-data-scientist-should-know\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10081316,"title":"SEBI to Use AI Scanner PINAKA to Examine Stock Tips on TV Channels","content":"Market regulator Securities and Exchange Board of India (SEBI) has enhanced its technology over the past few years to improve market vigilance and policymaking. Recently, it also showed interest in a system followed by Las Vegas casinos to keep high-risk gamblers away who place erratic bets. An important step in strengthening its surveillance over stock recommendations that are aired on all major business television channels, SEBI has created an artificial intelligence backed system called PINAKA (picture based information news accumulator and key information analyser) to scan the suggestions and condense them into a database. SEBI Database The SEBI database will be incorporated into conducting thorough surveillance for securities market wrongdoings like front running and insider trading. TV channels generally display suggestions, including price recommendations, made by an internal or external analyst. How PINAKA works? PINAKA checks each frame and processes the extracted data of recommendations in a standard format. It is then compared with the analyst’s trading pattern. Conclusion SEBI plans to make complete use of AI and machine learning to curb fraudulent activities. It strives to implement analytics based on unstructured data on its Data Lake platform. PINAKA is a step towards this aim. The annual 2020-21 report of SEBI stated that it would administer major IT projects that are important to its daily operations and mandate.","excerpt":"PINAKA will be used to discover misconduct in the securities market like front running and insider trading.","categories":["AI News"],"tags":["sebi"],"author_name":"Shritama Saha","publish_date":"2022-12-02T12:13:13","publication_year":"2022","word_count":218,"keywords":["machine learning","artificial intelligence","programming_languages:R","AI","sebi","Aim","ViT","analytics","R","data lake"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","R","data lake","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sebi-ai-scanner-pinaka-to-examine-stock-tips-on-tv-channels\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":20125,"title":"Carlsberg Developing New AI Which Can Sense Flavour, Aroma In Beer","content":"The day may not be far away when the beer that you sip on will have been created by Artificial Intelligence (AI). Carlsberg, one of the most recognised beverage brands globally, is turning to AI with its Beer Fingerprinting Project, to develop new beers and improve quality control. The project, whose purpose is measuring and sensing flavours and aromas in beer, will see Carlsberg collaborating with Microsoft, and two universities – Aarhus University and The Technical University of Denmark. According to Jochen Förster, Director and Professor of Yeast Fermentation, Carlsberg Research Laboratory, currently there is no such rapid technology that can differentiate the complex textures of flavours. “No rapid assays exist today for the determination of flavour compounds in beverages but it is crucial that we can do this to ensure that the laboratory continues to develop beer of the highest possible quality and provide a model for brewing in Denmark and the rest of the world,” said Förster. Using sensor technology and AI, the project hopes to develop ‘novel’ yeast for beer varieties such as  alcohol-free, craft, core and specialty. Reduced time and cost to develop new beers would be one of the key benefits of the initiative. Apart from beverages, the Danish brewing company is also considering exploring the application of the technology in other industries such as pharmaceuticals and food. The food and beverage industry across the globe has been seeing an astronomical rise in the adoption of of AI. Earlier this year there were reports of Coco-Cola experimenting with AI. Cherry Sprite,the company’s latest offering, is a result of this investment. It was created based on the data collected by monitoring the preferred combination of drinks at its self-serving vending machines. Currently, it is working on a virtual assistant which will interact with customers. Better understanding of customer preferences is the underlying focus of the venture. American coffee giant, Starbucks, has also been working on incorporating AI in its Starbucks Rewards members’ accounts. Called the Digital Flywheel Program, it will reportedly be able to make food and drinks suggestions based on factors such as order history, time of the day, day of the week and weather conditions. With even the beverage industry joining the AI bandwagon, the game-changing technology seems to be attaining an omnipresent status.","excerpt":"The day may not be far away when the beer that you sip on will have been created by Artificial Intelligence (AI). Carlsberg, one of the most recognised beverage brands globally, is turning to AI with its Beer Fingerprinting Project, to develop new beers and improve quality control. The project, whose purpose is measuring […]","categories":["AI News"],"tags":["coca-cola","virtual assistant"],"author_name":"Jeevan Biswas","publish_date":"2017-12-27T09:51:32","publication_year":"2017","word_count":380,"keywords":["Go","API","artificial intelligence","coca-cola","AI","programming_languages:R","programming_languages:Go","Git","RAG","virtual assistant","R"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","R","Go","Git","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/carlsberg-ai-flavour-beer\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10055239,"title":"What is Contrastive Self-Supervised Learning?","content":"In the present scenario, the amount of generated data is increasing and the complexity of data annotation is also increasing. To resolve the issue of annotation, self-supervised learning methods come into the picture. Self-supervised models can learn better from the raw data. In this article, we are going to discuss a type of self-supervised learning which is known as contrastive self-supervised learning (contrastive SSL). The methods in contrastive self-supervised build representations by learning the differences or similarities between objects. The major points to be discussed in this article are listed below. Table of Contents About Self-Supervised LearningWhat is Contrastive Learning?Non-Contrastive Vs Contrastive Self-Supervised LearningContrastive Self-Supervised Learning in DetailExamples of Contrastive Self-Supervised Learning About Self-Supervised Learning Self-supervised learning is considered a part of machine learning which is helpful in such situations where we have data with unlabeled information. We can say that it is a process between supervised learning and unsupervised learning. Usually, we find this type of learning based on neural networks. We know very well about the learning capabilities of neural networks. In self-supervised learning, neural networks can learn in two steps: To initialize the weight of the networks, problems with false labels can be solved.The actual task of the process can be performed by supervised or unsupervised learning. When we talk about the results, we have seen various promising and accurate results in recent years and various large companies like meta and Google are using this type of learning process for image, video, and audio processing. The basic idea behind self-supervised learning is to train the algorithms with the lower quality data, where other learning processes are focused on improving the final outcome of the algorithms. self-supervised learning methods can roughly be divided into two classes methods: Contrastive self-supervised learning Non-contrastive self-supervised learning In this article, we are focused on contrastive self-supervised learning, so it becomes necessary to understand what is contrastive learning, which is explained in the next section of the article. What is Contrastive Learning? In machine learning, we use a similar kind of data for training the algorithms under it. And when we talk about well-labelled data, it is easy for machine learning algorithms to get trained on it. In any case, the quality of data is not appropriate for training the machine learning algorithms. We use contrastive learning for finding good quality data. We can say that contrastive learning is an approach to finding similar and dissimilar information from a dataset for a machine learning algorithm. We can also consider contrastive learning as a classification algorithm where we are classifying the data on the basis of similarity and dissimilarity. In the algorithm of contrastive learning, inner workings can be completed by learning an encoder f such that: Score(f(x),f(x+)) >> score(f(x),f(x-)) Where x+ can be considered as a positive sample which similar to xx- can be considered as a negative sample which is dissimilar to x Score is a function for measuring the similarity level between two samples Using a softmax function, we can classify between similar and dissimilar samples accurately. There are various examples we have seen for this type of approach and one of the major examples is the framework SimCLR by the Google AI team. Image source The above image is a representation of SimCLR in which CNN and MLP layers are trained simultaneously. Training of these layers is responsible for validating projections that are similar for different augmented versions of the same image. Here in the above example, we have seen what contrastive learning is. Let us distinguish between non-contrastive and contrastive self-supervised learning to understand the topic in more detail. Non-Contrastive vs Contrastive Self Supervised Learning Contrastive Self-Supervised LearningNon-Contrastive Self-Supervised Learning1Contrastive SSL uses both positive and negative samples from the dataNon-contrastive SSL uses only positive samples from the data2The distance between the positive samples is minimized in contrastive SSLNon-contrastive SSL works on the useful local minimum from the data 3In this learning, backpropagation can be utilized without any extra predictorIn this learning, an extra predictor is required on the current state for backpropagation4Networks in this learning are more complex and can be considered as the group of networksThis learning approach has less complicated neural networks or we can say the networks under this learning are simple linear networks. Contrastive Self-Supervised Learning in Detail In the above section, we have discussed that the major goal of self-supervised learning is to learn from the lower quality data and the goal of contrastive learning is to distinguish between similar data and dissimilar data.  Also, if we talk about the classes of self-supervised learning, we find that contrastive self-supervised learning is a class of self-supervised learning. This mainly implies the answer to the question “ what is a good representation of the data?”  in self-supervised learning. Let’s take an example of a computer vision domain where the task for self-supervised learning is to learn the visual representation of the image data and the problem in front of us to find the answer to the question “what is good visual representation? In such a situation, anyone can say that the answer is” a representation which can be used easily in the downstream task”. In most cases, we see the representation from the self-supervised learning is applied to the algorithms for downstream tasks such as face recognition and object detection. The representation is evaluated by the performance of the downstream tasks. During this process, we get important and useful information about the learned representation but we don’t get any feedback like “why we get such performance in the downstream tasks”. Using contrastive self-supervised learning we can obtain intuition and conjectures for the efficiency of the learned representation. For a better representation of the process and data, we use contrastive learning with self-supervised learning. To understand the representation we are required to have some of the fundamental knowledge of representation. Invariances measurement: Invariance to the categories of the data is a crucial component of the representation. If we take the example of computer vision, we consider the invariance to the transformations as the component of representation. This component is very helpful for representation to be applied on the downstream tasks. If we talk about a good representation of visual data, we say that the representation should be mostly invariant to all the transformations. mathematically, if the function of representation is h(x) should be invariant to the transformation t : x → x if h(t(x)) = h(x). Augmentation: In contrastive learning, we have seen that it works with both positive and negative samples and the focus of the procedure is to find positive samples from the data so that it can be fed to the downstream task-oriented algorithm. Most of the time we see that network for contrastive learning in training time uses the augmented data from the training data. For example, if we talk about the computer vision domain, randomly cropped part of the images is used as the positive pairs which is an essential procedure for matching the features of partially visible images. And the process is responsible for providing a high-quality result from the contrastive SSL by augmentation of the data. So the measurement of the augmentation level becomes a crucial component for understanding the representation. Dataset Biases: In machine learning, we are required to train the model with the training set using any type of learning. Here also the contrastive SSL approaches get trained on different datasets and the effects we see in the training can be caused by the bias data. Effects can be positive or negative. Which also affects the representation of the data. In computer vision, mostly contrastive SSL approaches get trained on the ImageNet dataset. Where images in the dataset are object-centric biased.  Representations that do not differentiate biases can achieve seemingly enhanced performances. Here in the article, we have seen the basic intuition behind contrastive self-supervised learning and how representation and its component affects the process. Let’s take a look at some of the examples of the frameworks which provide a facility of contrastive self-supervised learning. Examples Some examples of contrastive self-supervised learning are listed below: MoCO (Momentum Contrast) is a framework for visual representation learning.  The work behind this was done by K. He, H. Fan, Y. Wu, S. Xie, and R. Girshick.PIRL(Self-Supervised Learning of Pretext-Invariant Representations) has the goal to build image representation that is meaningful and does not require a high amount of training samples of images. This work is performed by Ishan Misra and Laurens van der Maaten.Google’s SimCLR,  advances state of the art on self-supervised, semi-supervised learning and image classification. WAV2VEC, is a model for unsupervised pre-training for speech recognition by learning representations of raw audio developed by Facebook AI research. Final Words Here in this article, we have discussed an overview of self-supervised learning and contrastive learning. By merging them we can make it contrastive SSL, which is also a part of self-supervised learning. The quality of providing meaningful representation of the data and procedure makes the contrastive SSL different from self-supervised learning.","excerpt":"By merging self-supervised learning and contrastive learning we can make it contrastive self-supervised learning, which is also a part of self-supervised learning.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","contrastive learning","Data Science","Data Scientist","Deep Learning","Machine Learning","Python"],"author_name":"Yugesh Verma","publish_date":"2021-12-12T10:00:00","publication_year":"2021","word_count":1500,"keywords":["Go","self-supervised learning","machine learning","AI","neural network","ML","Machine Learning","computer vision","Python","object detection","CNN","Deep Learning","Data Science","Data Scientist","R","contrastive learning","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","computer vision","object detection","R","Go","CNN","self-supervised learning"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/what-is-contrastive-self-supervised-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10062586,"title":"My journey in data science: Asha Vishwanathan, Verloop.io","content":"Asha Vishwanathan leads the machine learning division at Verloop.io, a Conversational AI platform. Her expertise lies in NLP and computer vision and her career spans 15+ years. She got bitten by the data science bug in 2014 and made a pivot from analytics. “ML piqued my interest as it was the perfect blend of programming and data related technologies that I had been dealing with in my career,” said Asha. Register for our Free hands-on workshop Analytics India Magazine spoke to Asha to understand her data science journey. “Disregarding your past experience and starting afresh is a challenge that one must overcome while transitioning between different job profiles,” she added. Excerpts: AIM: In spite of a successful career in analytics, what made you switch to the world of data science? Asha Vishwanathan: After working for almost a decade with major IT brands, from corporates to budding startups, I took a sabbatical from work to explore my ‘Ikigai’ and went back to basics to understand what actually made sense to me. Upon realising I have an affinity towards programming, I moved to machine learning. Things got interesting when I took a course on data science through Coursera and my interest in the field developed. AIM: How did you develop your data science skillset? Asha Vishwanathan: I started with a lot of general reading on the basics. I started with research on business intelligence as it was something that I had worked on before. Having a background in business intelligence, data visualisation, reporting and dashboarding, I understood the nuances and found data science in this field to be a natural extension of what I had been doing. My own past experiences and understanding of the field guided me through the vast universe of data science. I started off by doing an introductory course through Coursera to gain a basic understanding of the field and later found a lot of online material. However, the online resources lacked a structured approach and one can easily get lost in the world of data science. Doing advanced courses on R and Big Data through Jigsaw gave me clarity on the path I wanted to take. In 2020, I did a business analytics course at the Indian Institute of Management, Bangalore which gave me a sweeping understanding of the data science scene. AIM: How did you feel when Harvard Business Publishing chose your case study on skilling needs of the Indian ecosystem? Asha Vishwanathan: The project was actually about identifying the skilling needs in the Indian ecosystem. The government provides a lot of schemes and initiatives to train and enable the common people. But, many drop out or are unable to complete such courses. We wanted to find out the root cause of this discontinuation and our research revealed that the skillset provided by the courses did not match the required job profiles. So we decided to look at building a system that would recommend the type of skills that the target audience needs for specific job profiles. It was a challenging process as many candidates were from a blue-collar segment and lacked a basic resume. Taking into account people from different backgrounds like coconut vendors, labourers, plumbers, etc., we created a model to predict the possibility of such individuals passing the aforementioned certifications using previous assessment data. We took data from soft skills required from the job, skills that the individuals have married them all in a comprehensive recommendation model that caught the eye of Harvard business publications. AIM: When did your data science journey actually begin? Asha Vishwanathan: I was already working in data science before joining IIMB. But I decided to do more to understand the breadth of the field. I soon realised that I have stumbled into a niche of machine learning while working in Kernel Insights. The road took many turns after I got certifications from Jigsaw and I joined a company called Poolcircle. There, I worked on changing the ride-sharing scene in Bengaluru using ML methodologies and data science. To change the way people commute, we looked into making a product that recommended routing strategies. After Poolcircle, I joined an early-stage startup called Kernal Insights where I tackled computer vision problems. AIM: What kind of projects are you heading in Verloop.io? Asha Vishwanathan: After joining Verloop.io, I tackled projects around NLP and conversational AI as I realised there are a lot of similarities between how the models work across computer vision. I faced challenges specific to chatbots, the whole ecosystem that it caters to and their implementation in typical B2B, SaaS type models. Working on the hosting and deployment of such models is especially interesting as it is not just restricted to machine learning algorithms but also about scaling up the project to meet industry standards. Looking for a solution from an end-to-end standpoint is what excites me the most. AIM: What’s your advice for data aspirants? Asha Vishwanathan: Data science is a detail-oriented and rigorous field. In order to make it in this market, you must get down to the basic principles and understand if it is their cup of tea. You should also be able to persevere through algorithms day after day and work on similar problems without being saturated. Don’t just blindly dive into it because it’s trending, you might not like it later.","excerpt":"I worked on changing the ride-sharing scene in Bengaluru using machine learning methodologies and data science.","categories":["AI Features"],"tags":["AI Chatbot","Conversational AI","head of data science","Interviews and Discussions","Journey into data science","Machine Learning","NLP"],"author_name":"Kartik Wali","publish_date":"2022-03-11T17:00:00","publication_year":"2022","word_count":886,"keywords":["data science","Journey into data science","machine learning","AI","AI Chatbot","chatbots","ML","Machine Learning","computer vision","NLP","Aim","Conversational AI","analytics","head of data science","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","computer vision","data science","analytics","Aim","chatbots","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/my-journey-in-data-science-asha-vishwanathan-verloop\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10044415,"title":"DeepSpeed Vs Horovod: A Comparative Analysis","content":"Deep learning represents a new artificial intelligence (AI) and machine learning paradigm. It has achieved enormous appeal in scientific computing, and its algorithms are widely employed to address challenging issues. To a certain degree, all deep learning algorithms depend on the capacity of deep neural networks (DNNs) to span GPU topologies. However, the same scalability has led to computer-intensive programmes, which pose operating problems for enterprises. Thus, from training to optimisation, the life cycle of a deep learning project demands strong building blocks for infrastructure that can extend computer workloads. Over the years, many open-source deep learning optimisation libraries have been announced by tech giants such as Google, Microsoft, Uber, DeepMind and others. In this article, we will compare two of these libraries–DeepSpeed and Horovod. DeepSpeed In February 2020, Microsoft announced the release of an open-source library called DeepSpeed. Training a large and advanced deep learning model is complex and includes a number of challenges, such as model design, setting up state-of-the-art training techniques including distributed training, mixed precision, gradient accumulation, among others. There is no certainty that the system will perform up to the expectation or achieve the desired convergence rate. This is because large models easily run out of memory with pure data parallelism, and it is hard to utilise model parallelism in such cases. This is where DeepSpeed comes into the picture, which addresses these drawbacks and accelerates model development and training. One of the most important applications of DeepSpeed has been the development of Turing natural language generation (Turing-NLG), one of the largest language models with 17 billion parameters. DeepScale stands apart in four important areas: Scale: DeepSpeed supports system running models with up to 100 billion parameters, which is ten times improved on existing training optimisation frameworks. DeepSpeed’s 3D parallels can effectively train in-depth learning models with trillions of parameters using contemporary GPU clusters with hundreds of devices. Speed: DeepSpeed was 4-5 times higher than competing libraries in initial tests. Cost: models could be trained at three times cheaper using DeepSpeed than the alternatives. Usability: DeepSpeed does not require PyTorch models for refactoring and can be used with only a few lines of code. Horovod Horovod is Uber’s open-source, free software framework for distributed deep learning training using TensorFlow, PyTorch, Keras and Apache MXNet. Horovod aims to make distributed deep learning quick and easy to use. Originally, Horovod was built by Uber to make distributed deep learning quick and easy to train existing training scripts to run on hundreds of GPUs with just a few lines of Python code. It also brought the model training time down from days and weeks to hours and minutes. In the cloud platforms, including AWS, Azure, and Databricks, Horovod can be installed on-site or directly run out of the box. Furthermore, Horovod can run on top of Apache Spark, allowing data processing and model training to be unified under a single pipeline. Once Horovod is configured, the same infrastructure may be used to train models with any framework, allowing the switching between TensorFlow, PyTorch, MXNet and future frameworks. The main principles of Horovod are built on MPI notions, namely size, rank, rank, local rank, allreduce, and allgather. DeepSpeed vs Horovod Advanced deep learning models are tough to train. Besides model design, model scientists also need modern training approaches such as distributed training, mixed precision, gradient accumulation and monitoring. Still, the ideal system performance and convergence rate cannot be achieved by scientists. Large models give considerable accuracy benefits, but training billions to trillions of parameters often meets fundamental hardware restrictions. Existing systems make trade-offs between processing, communication and development efficiency to fit these models into memory. DeepSpeed and Horovod address these difficulties to expedite model development and training. DeepSpeed brings advanced training techniques, such as ZeRO, distributed training, mixed precision and monitoring, to PyTorch compatible lightweight APIs. DeepSpeed addresses the underlying performance difficulties and improves the speed and scale of the training with only a few lines of code change to the PyTorch model. On the other hand, the primary motivation for Horovod is to make it easy to use a single GPU training script and to scale it successfully to train across several GPUs. At Uber, it was found that the MPI model was considerably more straightforward and needed far fewer code modifications than earlier alternatives such as Distributed TensorFlow with parameter servers. Once a training script with Horovod is built, it could run on a single GPU, several GPUs or even numerous hosts without changing the code. Furthermore, Horovod is not only easy to use but also fast.","excerpt":"A comparative analysis of open-source deep learning optimization libraries DeepSpeed and Horovod for advancing large-scale model training.","categories":["Deep Tech"],"tags":["cuda","horovod","Keras","Pytorch","Tensorflow"],"author_name":"Ritika Sagar","publish_date":"2021-07-24T11:00:00","publication_year":"2021","word_count":760,"keywords":["Pytorch","cuda","Keras","artificial intelligence","AI","machine learning","neural network","PyTorch","AWS","Aim","deep learning","horovod","TensorFlow","Tensorflow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","Aim","TensorFlow","PyTorch","Keras","AWS"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/deepspeed-vs-horovod-a-comparative-analysis\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10170733,"title":"‘No Product Was Ever an Instant Hit,’ Zoho’s Sridhar Vembu Defends Sarvam AI","content":"Zoho founder and chief scientist Sridhar Vembu has come out in support of the Indian AI startup Sarvam AI amidst the backlash against the recently released Sarvam-M LLM. Responding to skepticism over the model’s impact and utility, Vembu said, “There is no product we have built that was ever an instant hit,” urging the team to continue building despite early backlash. “Even when we were the first mover in a new market and we had done a lot of technical work, we only got slow traction,” he said. He encouraged Sarvam’s team to keep pushing forward, emphasising that instant popularity isn’t necessary for long-term success. In defense of https:\/\/t.co\/TYqr5LMKuz, I will point out that there is no product we have built that was ever an instant hit. Even when we were the first mover in a new market and we had done a lot of technical work, we only got slow traction. Instant success is neither necessary…— Sridhar Vembu (@svembu) May 25, 2025 The comments follow the release of Sarvam-M, a 24-billion-parameter hybrid open-weight language model trained on Indic languages, math, and programming. With just 334 downloads in two days on Hugging Face—Sarvam AI received some flak. Das called it “embarrassing,” and said there’s no real audience for this incremental work. Das contrasted this with an open-source model, developed by two Korean college students, that garnered about 200k downloads. His comments led to a heated debate among the Indian AI community. As of now, the downloads on Hugging Face are 718, showing that the interest amongst developers is increasing. Amidst this backlash, Sarvam co-founder Pratyush Kumar expressed optimism in a post on X. “Great to be receiving feedback on Sarvam-M. Please keep them coming. Will help strengthen our pipelines as we start to train our sovereign model,” he said, confirming that Sarvam is on the path of building a foundational LLM under the IndiaAI Mission. Great to be receiving feedback on Sarvam-M. Please keep them coming. Will help strengthen our pipelines as we start to train our sovereign model.This was particularly interesting – https:\/\/t.co\/vs0KL0NLII— Pratyush Kumar (@pratykumar) May 25, 2025 Read: Sarvam AI’s Backlash Exposes the Sad State of Indian AI","excerpt":"“Even when we were the first mover in a new market and we had done a lot of technical work, we only got slow traction,” he said.","categories":["AI News"],"tags":["Sridhar Vembu","zoho"],"author_name":"Mohit Pandey","publish_date":"2025-05-26T09:31:09","publication_year":"2025","word_count":361,"keywords":["Go","zoho","Hugging Face","programming_languages:R","AI","programming_languages:Go","RAG","Sridhar Vembu","ai_frameworks:Hugging Face","R","startup"],"extracted_tech_keywords":["AI","Hugging Face","RAG","R","Go","startup","ai_frameworks:Hugging Face","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/no-product-was-ever-an-instant-hit-zohos-sridhar-vembu-defends-sarvam-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10047510,"title":"Hospitality Unicorn OYO Has Room For AI To Lead Tech Race","content":"Hospitality unicorn OYO has decided to go bullish on providing better technology experiences for OYO’s hotel and home partners. With a mission to make OYO a leading technology company of choice by building products from India for the world, the company has appointed Ankit Mathuria as its Chief Technical Officer in Feb 2021. “At OYO, keeping the interests of our partners and customers in mind, we keep evolving expeditiously. Also, our use cases are pretty peculiar. That’s why we prefer to have our in-house R&D team,” said Ankit Mathuria, CTO, OYO. Ankit is an industry veteran with 15+ years of in-depth technical knowledge in consumer-facing industries. Before OYO, he was working with Amazon. Recently, the company has launched OYO 360 – a self onboarding tool. It allows small hotels and homeowners to get their properties listed on the company’s platform within 30 minutes against the typical 15-days process. AIM: What all AI\/ML\/Analytics solutions does OYO employ to drive growth? Can you discuss customer success stories or use cases? Ankit: Over the past years, we’ve built systems in a much more extensible manner to handle scale. Therefore, we’ve developed several applications led by AI\/ML and Analytics, such as dynamic pricing, OTA (online travel agents) performance management, ranking of properties, personalised recommendations, and others. As a result, our USP is providing an end-to-end easy and fast booking experience. For instance, with our pricing algorithm, we do approximately 60+ million micro-optimisations per day to get the optimum pricing for hotels and homes and help our partners maximise RevPAR without human intervention. Our unique customer acquisition program – Discover OYO, is another example of a data-driven design to optimise both ends of the spectrum by driving repeat bookings for our hotel owners while offering customers their first stay at a minimal price. Coming to the impact of technology on our business, I would like to share that today, nearly 70% of our bookings come from the top 3 search results. In the USA, OYO clocks in 3.4% more bookings compared to non-OYO hotels from OTAs like booking.com. AIM: How much percentage of resources (recruited workforce in AI\/ML) are focused on implementing state-of-the-art? Ankit: As a travel technology company, putting data science tools in the hands of people who are not data scientists is what we do. Be it for our partners, our customers, or even our employees. What’s truly fascinating is that every OYOpreneur is driven by technology and innovation. Thinking tech fuels our organisation, even outside the tech team. And that’s why upskilling talent to prepare them for future roles is very close to our hearts. With this in mind, last year, we launched the Data Science Academy to build world-class data talent and provide holistic careers to employees. We strongly believe this will help them make strong, unbiased business decisions. If you ask me specifically, currently, nearly 15-18% of our engineers are specialised in AI\/ML\/Data Science-related skills. AIM: What kind of skills suit you best when hiring data scientists\/ML engineers for OYO? Ankit: There is a different interpretation of these roles in the industry. At OYO, we see it as a strategic or niche vertical with depth. Applied scientists at OYO possess a combination of research, ML fundamentals, and software engineering. So, to answer your question, the problem statement greatly varies depending on the level and experience ambiguity. However, an applied scientist should be particularly: Skilled in exploratory analysis of data to find answers,Develop ML models using state-of-the-art techniques like deep learning, andProductionise models to meet engineering expectations of latency and availability. Personally, when it comes to any tech hires, I believe in hiring people who think long-term and will help build capabilities for the company from scratch. AIM: What would you say is the major challenge when it comes to recruiting Indian talent? Ankit: In the past decade, there’s been a tech boom in India, with a spurt in the fields of engineering and product development. However, there’s a talent crunch at the pace at which India’s tech ecosystem is growing. There is a higher demand for tech talent in AI\/ML and data science roles today than ever before. At OYO, we’ve been extremely blessed to have the right tech talent spanning multiple skill sets. Keeping up with evolving technologies and finding skilled talent in these newer domains is another challenge for the tech industry in India. AIM: What are the predominantly used programming languages and the tool-stack by your data science team? What kind of AI\/ML deployment challenges does your team face? Ankit: We believe that a programming language is just a medium to solve problems. Our engineers pick up the right language based on the use cases ranging from Python to R to other open-source frameworks. We run a lot of AI\/ML experiments in parallel. So, as we build our capabilities to do many more experiments, we usually face challenges in running experiments on distributed data and cost-effectively scaling our infrastructure. At OYO, we always take challenges head-on. Our tech teams have taken up this challenge and are now building an ML Ops platform, where teams can run their experiments on a loosely coupled, plug & play infrastructure. AIM: How do you see the landscape of AI\/ML evolving in India with regards to your domain? Ankit: We foresee that using our AI\/ML technologies, we will be better placed to deliver higher value to our patrons by providing them with better insights into evolving customer behaviours and pricing trends. Simply put, by using our personalised recommendations, independent businesses like small hotels and homeowners will be better equipped to provide enhanced customer service to guests. This, naturally, will lead to higher repeat rates and RevPAR growth.","excerpt":"The mission is to make OYO a leading technology company of choice by building products from India for the world.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Interviews and Discussions"],"author_name":"kumar Gandharv","publish_date":"2021-09-02T16:00:00","publication_year":"2021","word_count":945,"keywords":["data science","Go","AI","ML","Python","Aim","deep learning","analytics","GAN","R","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","deep learning","data science","analytics","Aim","Python","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hospitality-unicorn-oyo-has-rooms-for-ai-to-lead-tech-race\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10048276,"title":"Underrated But Interesting ML Concepts #6- LOF, MKL, RIPPER, t-SNE","content":"A few fascinating machine learning concepts are rarely explored. Some of them, including Local outlier factor (LOF), Multiple kernel learning (MKL), Repeated incremental pruning to produce error reduction (RIPPER), and  t-distributed stochastic neighbor embedding (t-SNE), will be discussed here. Local outlier factor Based on a concept of local density, LOF estimates the density using the distance of k nearest neighbours. The local density of an object can be compared to the local densities of its neighbours in order to find regions with similar densities and places with densities that are much lower than their neighbours. It is believed that they are outliers. The typical distance at which a point can be “reached” from its neighbours is used to estimate the local density. The LOF notion of “reachability distance” is an extra criterion for obtaining more consistent outcomes inside clusters. LOF can discover outliers in a data set that might not be outliers in another sample because of the local approach. An outlier is a point that is “close” to a very dense cluster, but a point within a sparse cluster may have similar distances to its neighbours. While the LOF algorithm‘s geometric understanding is limited to low-dimensional vector spaces, it can be used in any setting where a dissimilarity function can be specified. It has been found to operate well in a variety of settings, frequently beating competitors such as network intrusion detection and processed classification benchmark data. Multiple kernel learning MKL refers to a series of machine learning algorithms that include a specified set of kernels to learn an optimal linear or non-linear combination of kernels. To determine the kernel combination function, the present MKL algorithms use a variety of different learning methods. A total of five primary categories have been identified. Fixed rules are functions that do not require any training and have no parameters. Heuristic techniques look for the parameters of a parameterised combination function by looking at some measures obtained from each kernel function independently. Optimisation methods also employ a parameterised combination function, with the parameters learned from the solution of an optimisation problem. Bayesian techniques use the kernel combination parameters as random variables, assign priors to them, and train them as well as the base learner parameters using inference. Boosting approaches, which are based on ensemble and boosting procedures, add a new kernel iteratively until the performance plateaus. Repeated incremental pruning to produce error reduction The Ripper method is a classification algorithm based on rules. The training set is used to generate a set of rules. It’s a common rule-induction algorithm. The Ripper algorithm is useful for datasets with unequal class distributions. When there are several records in a dataset, the majority belong to one class and the remaining to different classes; the dataset is said to have an uneven class distribution. Because it uses a validation set to prevent model overfitting, it performs well with noisy datasets. Rule Growing in the RIPPER Algorithm: Ripper has an approach for growing rules that ranges from generic to specific. It starts with an empty rule and adds the best conjunct to the rule antecedent. The metric used to evaluate conjuncts is FOIL’s Information Gain. The best conjunct is determined using this method. When the rule starts covering the negative (-ve) examples, the rule stops adding conjuncts. Based on its performance on the validation set, the new rule is pruned. t-distributed stochastic neighbour embedding t-SNE is a statistical method for displaying high-dimensional data by assigning a two- or three-dimensional map to each data point. It’s a non-linear dimensionality reduction approach that works well for embedding high-dimensional data in a two- or three-dimensional low-dimensional space for visualisation. It models each high-dimensional object by a two- or three-dimensional point so that comparable objects are modelled by nearby points with a high probability, while distant points model different objects. While t-SNE plots often appear to show clusters, the visual clusters can be highly influenced by the parameterisation chosen, necessitating a thorough grasp of the t-SNE parameters. Such “clusters” have been proven to arise in non-clustered data, suggesting that they are bogus discoveries. As a result, an interactive investigation may be required to select settings and evaluate results. Genomic research, computer security research, natural language processing, music analysis, cancer research, bioinformatics, geological domain interpretation, and biomedical signal processing have all employed t-SNE for visualisation.","excerpt":"Here, we will explore some of the fascinating yet underestimated concepts in machine learning.","categories":["AI Features"],"tags":["Natural Language Processing"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-09-14T14:00:00","publication_year":"2021","word_count":722,"keywords":["Go","machine learning","programming_languages:R","AI","Natural Language Processing","programming_languages:Go","R"],"extracted_tech_keywords":["AI","machine learning","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/underrated-but-interesting-ml-concepts-6-lof-mkl-ripper-t-sne\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":19846,"title":"Putting AI To Work Is The Biggest Challenge For Businesses In 2018","content":"If the future of business is AI, then 2018 will be the year when that promise has to be fulfilled. Research firm Forrester has famously predicted that the honeymoon for organizations trotting AI capabilities is over. It’s time for companies to put AI to work. Dubbed as the year of reckoning — 2018 will be the year when AI will have to make progress in real-world applications. With 2018 rapidly approaching, AI is definitely on the minds of many business leaders. Analysts emphasize senior management will have to draw a line between the hype and real use cases and drive the impact home by realizing value from this technology. Pockets of Success Even though 2017 hasn’t exactly been a dull year for AI on the world stage, what with several pockets of success, from groundbreaking research in Deep Learning, M&As and acqui-hiring led by the Big 3 (Google, Microsoft, Amazon), 2018 should ideally outdo 2017 in terms of developments. While it may not be the year when bots get to work alongside, it sure would be a year of rapid development in AI, with some worthy headlines along the way. Also, the year 2017 saw a spike in turnkey AI and machine learning solutions such as Gluon from Microsoft & Amazon, IBM’s Watson APIs and Microsoft Azure ML Studio among other tools debuted by IT giants that can potentially speed up AI implementations. According to a recent Vanson Bourne study State Of Artifical Intelligence For Enterprises commissioned by Teradata, 80% of enterprises already have some form of AI (machine learning, deep learning) in production today. The findings also indicate enterprises are aware of significant barriers to adoption and are looking to strategize against those issues by creating a new C-suite position — Chief AI Officer (CAIO). So Far, Narrow AI Rules The Day Narrow AI, where machine-learning solutions target specific tasks, is the order of the day. Also known as Weak AI, Narrow AI carries out specific tasks brilliantly through a combination of advanced algorithms. Today, most of the breakthroughs are seen in the domain of Narrow AI — such as self-driving cars, AlphaGo, image recognition application, latest iPhoneX’s face recognition app, or virtual agents deployed to improve customer service, manage emails or schedule meetings. We are still far from achieving truly Intelligent AI – that can autonomously acquire learning from a single example, understands language, can reason contextually and boasts of EQ. Even though Indian enterprises have been quick to adopt AI adoption in the last five years, adoption is set to rise, a report by Intel India indicates. The heartening news is that AI adoption in India is on the rise with 68.6 per cent Indian organizations planning to deploy it before 2020. Meanwhile, 71 per cent of organizations are looking at increased process automation as a key benefit that could further drive spends on this technology by 2020. AI has penetrated all sectors, particularly IT that saw an enterprise-wide adoption of RPAs earlier this year. We list down a few ways for businesses to make AI work  in enterprise settings: 1.Increased focus on business-led use cases: According to McKinsey’s Simon London, today every industry  can make use of AI technology what with use cases right across the value chain and across the operations of most companies. But even though there are a lot of applications and portfolio-of-initiatives across every industry, it is important for business leaders to understand this technology, find the potential area of application and understand how it can be leveraged as a competitive weapon. This will be the first step before diving into AI-enabled processes and AI-enabled business. The turnaround a business case is fast and Mckinsey’s Peter Breuer believes starting off with simpler use cases can help tech leaders prepare for more advanced uses cases in the future. 2.Tackle the good data challenge: A recent report threw light on the data deluge hype. It highlighted how despite the wealth of data flowing in—about 2.5 quintillion bytes a day, majority of the data is not labeled or structured, which renders it unusable for supervised learning tasks. As emphasized by Daniel Shapiro, a machine learning enthusiast, when you have labeled data machine learning works. But what about cases when you don’t have any labeled data? So, researchers devised two strategies – transfer learning and unsupervised learning to tackle the lack of labeled data challenge. So, how can enterprises meet the labeled data challenge? There are two options a) crowdsource labeled data; b) outsource the task of labeling data; c) there are a slew of startups that have devised machine learning models to cleanse data. Today, it has become crucial to get access to well-labeled, statistically representative data sets and maintain it over time. 3.User Experience in Customer-Facing AI Applications Should Take Centrestage: Now that AI has made its way to the frontlines with a slew of customer-facing applications, it is the right time for businesses to think about how to weave AI into a great user experience. In fact, UX should be front and centre more than ever given how bots are going to become a big part of enterprise settings. And a great UX goes beyond design thinking into providing a more unified customer experience. 4. Getting an ROI from AI: While AI has definitely lived up to its promise in 2017 and the investment is only expected to increase over the coming years, a recent survey pointed out how 64 per cent of IT decision makers expect to see a return on investment from AI implementation within two years. It is a question that vexes VCs and the C-suite alike. Since the application and level of maturity of AI technologies can vary a lot, it is difficult for decision-makers and senior management to generalize how to measure the ROI on their investment. A recent report indicates that the C-suite should also sport a VC-type mindset and brace themselves for both failure and success. 5. Time to appoint Chief AI Officer to strategize AI adoption: According to a recent survey b Teradata that sampled 260 large organizations globally, enterprises today are seeing AI as a strategic priority that will help them outpace the competition in their respective industries. Atif Kureishy, Vice President, Emerging Practices at Think Big Analytics, a Teradata company. And to leverage the full potential of this ground-breaking technology and gain maximum ROI, businesses will need to revamp their core strategies. “So AI has an embedded role from the data center to the boardroom,” says Atif Kureishy, Vice President, Emerging Practices at Think Big Analytics, a Teradata company. This can be fulfilled by creating a new C-suite position — the Chief AI Officer.","excerpt":"If the future of business is AI, then 2018 will be the year when that promise has to be fulfilled. Research firm Forrester has famously predicted that the honeymoon for organizations trotting AI capabilities is over. It’s time for companies to put AI to work. Dubbed as the year of reckoning — 2018 will be […]","categories":["IT Services"],"tags":["Chief AI Officer (CAIO)"],"author_name":"Richa Bhatia","publish_date":"2017-12-18T09:29:09","publication_year":"2017","word_count":1111,"keywords":["machine learning","AI","R","Chief AI Officer (CAIO)","ML","image recognition","RAG","deep learning","analytics","Azure","Azure ML"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","analytics","Azure ML","RAG","image recognition","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/putting-ai-work-biggest-challenge-businesses-2018\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10022750,"title":"The Internet Of Behaviours And How Ready We Are","content":"In 2012, Gote Nyman, a retired professor from the University of Helsinki, said an explosion of apps and services that get information directly from individuals and communities is imminent. At the time, Nyman was trying to figure out the limitations Internet of Technology (IoT) needed to overcome to predict behaviours. Today with 17.1 billion IoT devices in operation, coupled with rapid progress in machine learning abilities, the Internet of Behaviours (IoB) has taken on a life of its own. The IoB is an extension of IoT where the data collected from IoT devices are crunched to extract valuable insights into users’ behaviours, interests, and preferences. Inarguably, IoB has a lot of potential. However, collecting data from multiple devices to encourage\/discourage human behaviours comes with ethical implications. We discuss them below. Changing Value Chains IoB is a trifecta of technology, data analytics, and behavioural science to predict, analyse, and even influence human behaviour. At the same time, IoB systems can also alter the value chains between platforms and humans. While there are digitally literate people who are against giving their personal information for free, most customers are content as long as such platforms bring them value or make their lives easier. If you are not paying for it, you're not the customer; you're the product being sold.— Andrew Lewis (@andlewis) September 13, 2010 Companies use this opportunity to provide services that collect vast amounts of data that can eventually be used to modify behaviours. While the value proposition is tempting, the autonomy of individuals is at stake. Privacy Concerns The more IoT generates data, the better the efficiency of IoB systems to predict individual behaviour. Many companies buy and sell data. PayPal, for instance, has disclosed that it shares consumer data, including name, address and phone number, with hundreds of entities around the world. Also, monopolies like Uber continue to acquire competitor apps that bring user data to one place, mostly without users’ permission. This presents significant legal and security risks to privacy rights. Privacy is one of those things few will pay attention to, until everyone realizes they need it.— Chris Burniske (@cburniske) March 16, 2021 The continuous monitoring of individuals through IoB comes with a surveillance risk. While there are differences between government collecting data and private companies doing the same, there is always the threat of governments making it compulsory for private companies to share data or private companies selling data to government entities. For instance, the Indian government is set to bring new rules to mandate the voluntary sharing of information by social media platforms. Security Threats With the increasing number of cybersecurity threats globally, cybercriminals getting access to IoT data integrated with behavioural data can pose severe personal security threats. This could lead to cybercriminals gaining access to information like property access codes, delivery routes, even bank access codes. “The Internet of Things (IoT) devoid of comprehensive security management is tantamount to the Internet of Threats.” Stephane Nappo, Vice President Global Chief Information Security Officer, Group SEB Phishing is a phenomenon where the criminal poses as someone else to lure individuals into giving up sensitive information through various digital mediums. With the availability of behavioural data, an attacker could take phishing to a whole new level as he\/she will be able to impersonate individuals to perpetrate fraud or other nefarious activities. Wrapping Up While IoB is still in its nascent stage, it is predicted that by 2023, individual activities of 40% of the global population will be tracked digitally to influence individual or communities behaviour. Despite the concerns raised above, it’s unfair to characterise IoB is an agent provocateur. For instance, IoB can be used to monitor health, or encourage healthy living, instil savings habit etc. In one case, it was even used for monitoring the safety compliance of workers during COVID-19. “The IoT itself isn’t inherently problematic; a lot of people like having their devices synced and get benefits and convenience from this setup. Instead, the concern is how we gather, navigate, and use the data, particularly at scale. And we’re starting to understand this problem.” Chrissy Kidd, Technology Researcher and Author To prevent misuse, it is essential to have transparency and explainability in IoB systems. The governments should establish robust privacy laws while also securing their digital infrastructures to check the rampant proliferation of technologies like IoB.","excerpt":"In 2012, Gote Nyman, a retired professor from the University of Helsinki, said an explosion of apps and services that get information directly from individuals and communities is imminent. At the time, Nyman was trying to figure out the limitations Internet of Technology (IoT) needed to overcome to predict behaviours. Today with 17.1 billion IoT […]","categories":["IT Services"],"tags":[],"author_name":"Kashyap Raibagi","publish_date":"2021-03-24T12:00:00","publication_year":"2021","word_count":723,"keywords":["Go","API","machine learning","AWS","AI","Git","RAG","ViT","analytics","R"],"extracted_tech_keywords":["AI","machine learning","analytics","RAG","AWS","R","Go","Git","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-internet-of-behaviours-and-how-ready-we-are\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":56959,"title":"Top 7 Anime Based Open-Source Projects","content":"Anime is no longer limited only to Japan and China; it has gone global. It has attracted many people towards it because of its high-end graphics, vivid imaginations for the future, using highly advanced technologies which only find their place in our imaginations and artificial intelligence (AI) depiction in their storylines. Naturally, it serves as a means of entertainment for any kind of audience that watches it and also it could be fun to do projects related to it. And we all know Elon Musk likes anime too: https:\/\/twitter.com\/elonmusk\/status\/1054501056229588992?s=20 Below is the list of some of the popular kinds of open-source projects based on anime, have fun! Face Tracking with Anime Characters This is a python project, where an anime character can recognise the proximity of your face from the camera integrated into your device. The character featured in this project, Yuri is one of the five main characters from the popular anime game, Doki Doki Literature. The anime character in this project accesses your webcam to sense the proximity. The anime character moves its eyes according to your motion in front of the camera (mostly sideways). Besides, if you come near the camera, the character ‘blushes’. One requires- Python3, OpenCV, OS, Pygame. One needs to get the file ‘YURI FULLSCREEN.py’ from the repo, save the yuri2.bmp, take the eyes’ image and give a path for them for these images in the code. Next, download haarcascade, give its location in the code and run the code. One can use their own photos for the character. (Recommended: Kaguya from Kaguya-sama: Love Is War) You will need two images for the moving eyes version; a picture of the character without eyes and pictures of only eyes. For the character proximity sensor, the process remains the same as mentioned, the only difference being; one has to download YURI FULLSCREEN.py and use yuriblush.py. The publisher gives more details. You can access the repo here. Taiga Perhaps, Taiga might be one of the most useful open-source projects for anime enthusiasts. It is an open-source window application for the desktop, which automatically detects what anime video one has watched, finished and the progress of an ongoing video. It synchronises the progress of the watched anime with online services, and that is why this requires users to have accounts linked to online services like AniList and Kitsu. The source code can be found on GitHub. You can also manage your anime collection, share watched ones and also discover new series. This works for streaming services too. From Taiga Trackma Trackma is a Unix based program for fetching, updating and using data from user’s personal lists. These lists have to be hosted in several media tracking websites. Trackma is lightweight and simple to use. Trackma doesn’t only support for anime videos, but it also tracks manga progress too. It currently supports websites like Anilist, MyAnimeList, Kitsu, VNDB and Shikimori. With Trackma, you can: Manage and synchronise local list even when offline.Support for many media types.Feature for managing different accounts over various media tracking sites.Multiple user interfaces.Trackma can detect what is being played on the media player and updates the list if necessary.Scalable, easy to use, and the code is clean in spite of being in C++.Uses HTTPS wherever possible, so it is secure. Requires: Python 3.4\/3.5 Python3-pip or pyhton3-setuptools. Go to the repo here. Via Trackma Anime Anime is a web animation engine. It is a JavaScript animation engine which works with SVG, CSS, DOM attributes and JavaScript objects. Anime offers animation and interaction to web-based projects. It works with all the major browsers and provides easy implementations for designing both simple and complex animations. You can manually download it here. Get the repository here. MakeGirlsMoe Make Girls Moe lets you create anime character with AI. This is a project where GANs are trained to create anime characters faces, particularly the female anime characters. One only needs to enter parameters like the colour of hairs, facial expressions, eye colour etc. And the system generates anime faces. For each time you repeat the same settings, it gives out different results. The project offers 13 hair colours, ten eye colours, five hairstyles. This project is open-source anime and can help amateur storytellers. Get the repo here. When tried for brown hair, green colour and other filters, it gave this: Anime4K Anime4K is a real-time upscaling algorithm for anime videos. This algorithm can be implemented in any programming language. This method gives importance to well-defined lines\/edges while ensuring superior delicate textures. The algorithm treats colour information as a heightmap and directs pixels towards probable edges using gradient-ascent. Get more details about its working here. On Github here. Anime Inpainting This is an anime inpainting application which is based on edge-connect support. IT uses anime character pictures and anime photos as datasets and makes use of generative image inpainting with adversarial edge learning for inpainting. Prerequisites: Python 3, Pytorch 1.0 (lacks Pytorch 0.4 support), NVIDIA GPU + CUDA cuDNN. Via the repo page on Github Get the repo here.","excerpt":"Anime is no longer limited only to Japan and China; it has gone global. It has attracted many people towards it because of its high-end graphics, vivid imaginations for the future, using highly advanced technologies which only find their place in our imaginations and artificial intelligence (AI) depiction in their storylines. Naturally, it serves as […]","categories":["AI Trends"],"tags":["pygame"],"author_name":"Sameer Balaganur","publish_date":"2020-02-19T17:49:54","publication_year":"2020","word_count":837,"keywords":["CUDA","Go","artificial intelligence","PyTorch","AI","OpenCV","Python","pygame","JavaScript","R","Java"],"extracted_tech_keywords":["AI","artificial intelligence","PyTorch","OpenCV","CUDA","Python","R","JavaScript","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-anime-based-open-source-projects\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10115403,"title":"AI is Now Killing Prompt Engineering Jobs","content":"How many times have we heard that prompt engineering is the job of the future? Ever since ChatGPT was launched, everyone started experimenting with different ways of prompting the chatbot and called it a job. Others were scared of losing their jobs, thanks to this new role. Well, it now turns out that AI can do it and do it better than humans. “Did anybody, but the most desperate climbers, ever honestly believe that ‘typing prompts into ChatGPT’ was ever going to be a high-paying full-time job?” asked a user on HackerNews. In a recent study, VMware researchers found that LLMs become more unpredictable when humans start experimenting with weird prompts. Even more interesting was another research team’s finding that concluded that “no human should manually optimise prompts ever again”. The best prompt engineering is done by the AI model itself. Problems Aplenty with Manual Prompting Rick Battle and Teja Gollapudi from VMware experimented with big and small language models and tried different prompting techniques to figure out the most optimal and efficient way to do it. “It’s both surprising and irritating that trivial modifications to the prompt can exhibit such dramatic swings in performance,” read the conclusion of the paper. They even highlighted that there is no obvious methodology that could improve performance, and the effects would be very trivial. For example, the research concluded that practitioners do not even need GPT-4 or PaLM-2 size models to use effective prompts. In their experiment Llama 13B and Mistral-7B were able to produce “superior prompts”, which they found shocking. “It’s undeniable that the automatically generated prompts perform better and generalise better than hand-tuned ‘positive thinking’ prompts,” the paper concluded, saying that even after giving positive affirmation to chatbots, automatic prompts perform better – also called auto prompting. “I literally could not believe some of the stuff that it generated,” Battle said in an interview, talking about how no human could ever generate the prompt that the system generated itself, as it was bizarre. Bias in ML systems can come from bias in the training data. But that's only one possible source of bias among many. Arguably, prompt engineering can be an even worse source of bias.Literally any part of your system can introduce biases. Even non-model parts, like your…— François Chollet (@fchollet) February 23, 2024 Another aspect of prompt engineering is that it creates bias in the output of the model, which observers can attribute to the AI model itself. “Arguably, prompt engineering can be an even worse source of bias,” said François Chollet. People are confusing prompt engineering with just giving prompts to a chatbot in English language, while what the model does is actually a lot of maths, English is just the frontend of the model. Therefore, the AI models can do it better. It only makes sense. “In LLMs, the need for prompt engineering is a sign of *lack* of robust language understanding,” said Melanie Mitchell in a post on X last year. That seems to be the case even now, even after a year of development and scaling of language models. Was Prompting Just a Passing Fad? Just like C++ was considered a “dying language”, prompt engineering is regarded as a passing fad. Logging onto ChatGPT or Codex and just typing in what you want does not work that easily. It’s a skill to learn. And now, with AI doing it better than humans, it is a matter of time that it vanishes. But will it? Most of prompt engineering is just about trial and error, and now with AI, that is not needed anymore. Now that companies are hiring for roles such as LLMOps, they might be labelled as the new prompt engineers, but as a job, it won’t die. Douglas Crockford, one of the developers of JavaScript and the brain behind the JSON format, is worried about English as a programming language “because it’s so ambiguous,” he said in an interview with AIM. He clarified that the fundamental law of programming is that the program has to be perfect. “It has to be perfect in every aspect, in every detail, for all states for all time”. He further elaborated that if it isn’t, the computer is licensed to do the worst possible thing at the worst possible time. “It’s not the computer’s fault, it is the programmers’ fault,” he noted. Another important aspect to note about prompt engineering is that it changes as people experiment with different data sources and different AI models. And as companies are adapting different open source and closed source models, what is the best combination of prompts for each model might still be a job that someone needs to do. We’re calling them prompt engineers today, maybe as AI improves, we will call them something else. That is not to say that prompt engineering holds no value. It has already found its niche and is probably going to stay here for a while, but not too long.","excerpt":"Did anybody ever honestly believe that ‘typing prompts into ChatGPT’ was ever going to be a high-paying full-time job?","categories":["AI Features"],"tags":["Layoffs","prompt engineering"],"author_name":"Mohit Pandey","publish_date":"2024-03-12T15:30:00","publication_year":"2024","word_count":827,"keywords":["ChatGPT","TPU","Layoffs","AI","chatbots","ML","prompt engineering","Aim","JavaScript","small language models","R"],"extracted_tech_keywords":["AI","ML","ChatGPT","Aim","small language models","prompt engineering","chatbots","TPU","R","JavaScript"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-is-now-killing-prompt-engineering-jobs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10042142,"title":"Players To Be Pitted Against AI As Battle Of Polytopia Becomes Part Of DeepMind’s AndroidEnv","content":"The Battle of Polytopia is a popular strategy game developed by Swedish gaming company Midjiwan AB. This game is now all set to become part of DeepMind’s open-ended platform AndroidEnv. Launched in 2016, Polytopia is a world-building game where players play as one of 16 tribes to develop an empire and defeat opponents in a low poly square-shaped world in a game session that lasts 30 minutes. The players are pitted against human opponents or bots. With its inclusion on the AndroidEnv, DeepMind will test AI as an opponent against human players. Till now, Polytopia has been installed over 14 million times and has a huge base especially in Europe and the US. “We are of course very happy about this. DeepMind is a giant within AI with a fantastic platform that we are very proud and excited to be a part of. We have noticed that Polytopia is particularly popular among so-called brainiacs, but this will be the start of something completely new when we are taking on artificial intelligence,”  said Christian Lövstedt, General Manager at Midjiwan. DeepMind, founded by Demis Hassabis, Shane Legg, and Mustafa Suleyman in 2010 and subsequently acquired by Google in 2014, is working towards creating general artificial intelligence with the help of computer games. Recently, DeepMind introduced AndroidEnv, an open-source platform for reinforcement learning research that is built upon the Android Ecosystem. Reacting to the news, DeepMind CEO Hassabis tweets that he is excited to see how AI performs on Polytopia, one of his favourite mobile games. So excited for this, can't wait to see how our AI does on the strategic gem Polytopia, one of my favourite mobile games!! https:\/\/t.co\/ooE4oOvzy0— Demis Hassabis (@demishassabis) June 7, 2021","excerpt":"Popular strategy game, The Battle of Polytopia, is now all set to become part of DeepMind’s open-ended platform AndroidEnv.","categories":["AI News"],"tags":["Artificial General Intelligence","DeepMind","deepmind founder Demis Hassabis"],"author_name":"Shraddha Goled","publish_date":"2021-06-21T15:25:56","publication_year":"2021","word_count":283,"keywords":["Go","artificial intelligence","deepmind founder Demis Hassabis","programming_languages:R","AI","programming_languages:Go","Artificial General Intelligence","R","DeepMind"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/players-to-be-pitted-against-ai-as-battle-of-polytopia-becomes-part-of-deepminds-androidenv\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36280,"title":"How To Use Genetic Algorithms As A Tool For Feature Selection In Machine Learning","content":"The central idea behind using any feature selection technique is to simplify the models, reduce the training times, avoid the curse of dimensionality without losing much of information. The popular feature selection methods are: Filter method Wrapper method Embedded method For example, PCAs are popular with dimensionality reduction but the underlying assumptions of PCA depend on linearities. For nonlinear problems which is how real-world scenarios usually are, Genetic algorithms offer significant solutions. Nature-inspired metaheuristics to use mechanisms like reproduction, mutation, recombination and selection. The solutions to the optimisation problem play the role of individuals in a population and the fitness is used to determine the quality of the solution. How Do GAs Excel When the number of features is very large, the GA becomes very computationally demanding and its run time can be prohibitive. To reduce the time required for training, a two-phase hybrid combination of filtering and wrapping is considered. Chromosome1 (Parent1): [0 0 1 1 0 1 1 0 1] Chromosome2 (Parent2): [1 1 1 0 0 0 1 1 1] Two-point crossover operator ↓ Chromosome1 (Child1): [0 0 1 0 0 0 1 0 1] Chromosome2 (Child2): [1 1 1 1 0 1 1 1 1] In the first phase, filtering is applied as a pre-processing step and the top-ranked features are selected based on a pre-defined but tunable threshold. In the second phase, the GA-based wrapper is applied to the selected features, which in practice uses a much smaller search space instead of the original large search space. This heuristic optimi sation technique may or may not result in combinations of features that have high discriminative power. Where noise can significantly skew results and dimensionality is very large, genetic algorithms are still capable of optimising a correlated objective function. Multi dimensional covariance map makes the creation of offspring or desirable feature selection, Bayesian. While doing GAs it can be observed that the first generation is always larger. Genetic Algorithms For Finding Galaxies In this paper, the researchers try to tweak in the classification technique by combining genetic algorithms with support vector machines(SVM). Indian researcher Sidharth Kumar and his colleagues introduced a mutation in the genes with constant probability. The genes which are chosen to be mutated, are replaced with any of the genes that are not part of either parent, with uniform probability of choosing from the remaining genes. This allows for genes that are not part of the current gene pool to be expressed. The conceptual simplicity of genetic algorithms combined with their evolutionary analogy as applied to some of the hardest multi-parameter global optimisation problems makes them highly sought after. For SVM classification, the true positive rate is used as the fitness function. For SVM regression, a custom fitness function based on the problem at hand is chosen. Once the fitness of all organisms has been evaluated, a new generation of same size as that of the parent generation is created using roulette selection. The GA then runs until it reaches a pre-defined stopping criterion. We use here the posterior distribution of the parameters. By using Genetic Algorithms to select relevant features, and Support Vector Machines to estimate the quantity of interest using the selected features, the researchers show that the combination of these two methods yields remarkable results, and offers an interesting opportunity for future large surveys which will gather large amount of data. In the case of star\/galaxy separation, the improvements over existing methods are a consequence of adding more information. Getting to the target solution in a swarm of data points requires devouring on historical results while stating close to the reality in order to avoid overfitting. Converging on approximate solution is done by various methods of which stochastic optimisation method is one. Probabilistic model building genetic algorithms are a part of stochastic optimisation methods. These algorithms generate new solutions using an implicit distribution defined by one or more variables. These evolutionary algorithms use an explicit probability distribution encoded by a Bayesian network. Hyperparameter selection is a key task in improving neural networks and the implicit characteristic of genetic algorithms to implicitly search for best fit strings makes it a suitable contender for machine learning and AI applications as well along with other optimisation problems. “If you can do feature selection on a large enough set of data, which consists of all independent variables then you can do pretty much anything you want,” says Sidharth Kumar of Flipkart at MLDS 2019, talking about how ubiquitous machine learning is and how he applied his Sidharth Kumar, a Data Scientist at Flipkart says how ubiquitous machine learning is and how he applied this knowledge in observing black holes and conducting analytics for retail. Read more about his work here Also watch: https:\/\/www.youtube.com\/watch?v=49yPlLn2Ehs","excerpt":"The central idea behind using any feature selection technique is to simplify the models, reduce the training times, avoid the curse of dimensionality without losing much of information. The popular feature selection methods are: Filter method Wrapper method Embedded method For example, PCAs are popular with dimensionality reduction but the underlying assumptions of PCA depend […]","categories":["Deep Tech"],"tags":["feature selection","genetic algorithms","Machine Learning Algorithms","PCA"],"author_name":"Ram Sagar","publish_date":"2019-03-13T12:24:28","publication_year":"2019","word_count":789,"keywords":["Machine Learning Algorithms","Go","machine learning","programming_languages:R","AI","neural network","RPA","ML","analytics","GAN","PCA","R","feature selection","genetic algorithms"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","analytics","R","Go","GAN","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-use-genetic-algorithms-as-a-tool-for-feature-selection-in-machine-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10163931,"title":"Kenvue Has a Late-Mover Advantage as a GCC","content":"New global capability centres (GCCs) setting up shop in India today have a clear advantage—they can learn from the experiences of early adopters. Unlike the early birds who had to figure things out the hard way, these late movers can skip the trial-and-error phase and dive straight into best practices. The new GCCs can adopt AI-driven automation, and implement cloud-native architectures from day one to scale faster, optimise costs, and leverage advanced digital transformation strategies. With a mature talent pool, improved regulatory frameworks, and strong IT infrastructure, India now offers a well-established ecosystem that enables GCCs to set up operations faster. Some of the newly established GCCs in India are Hitachi Energy, Sandoz, DocuSign, SignatureIP, Atmus Filtration Technologies, Kenvue and others. Recently, AIM got talking with Rajesh Puneyani, vice president of technology & site leader of Kenvue Inc, a US-based company that focuses on consumer health products. It used to be Johnson & Johnson’s consumer healthcare division, but now owns popular brands like Aveeno, Band-Aid, Johnson’s, Listerine, and Neutrogena. “As a late mover in the GCC world, we pretty much know what exactly we need to get from a technology talent or from a domain skills point of view,” he said. He further stated that what others have achieved over the years, Kenvue has accomplished within a short span due to its strategic focus on talent acquisition and technological advancement. In just a year and a half since its inception, Kenvue has grown significantly, evolving from a small team of 400-500 professionals into a dynamic force in technology and product innovation. “In our first year, we are prioritising building strong foundational principles rather than rushing into major advancements,” Puneyani highlighted. Bridging the Gap Between Mental and Physical Availability Puneyani explained that mental availability refers to a consumer’s awareness and recall of a product when they need it—such as thinking about purchasing a specific nutrition brand. Physical availability, on the other hand, ensures that when a consumer visits their preferred store, the product is actually on the shelf, ready for purchase. “Often, consumers assume a product will be available, but if it isn’t, it leads to disappointment and lost sales,” he added. He further mentioned that Kenvue’s AI and data science teams in India work collaboratively, bringing together capabilities from different areas such as enterprise data warehousing, cloud computing, data engineering, and product development. These teams function as a unified force, using advanced analytics to bridge the gap between mental and physical availability. AI in the Mix Talking about AI models, Puneyani explained: “Our team does not use ChatGPT directly. Instead, we have developed our own AI-driven solution which is built on Azure OpenAI. While many organisations use similar AI technology, we have put in place our own cybersecurity measures and data handling processes to keep information safe.” Highlighting that the success of AI models depends on the quality of data they work with, he said, “This is where our data quality and engineering teams play a crucial role. Based in the GCC, these teams organise and refine data to make sure our AI models learn from clean, high-quality information. Once the data is fed into AI models, advanced computing and data science techniques turn it into useful business insights.” “But no matter how smart AI gets, human expertise remains essential,” he added. When it comes to industry partnerships, he mentioned that the company works with major technology firms like NVIDIA to improve its AI capabilities. While they use frameworks from these partners, they continue to develop their in-house AI expertise to keep up with business needs and changing regulations. One key development is their SKU recommendation technology, which personalises product suggestions based on factors like customer preferences, location, past purchases, and shopping habits. By using AI insights, they can recommend the right products to customers, improving their shopping experience. The company has also streamlined its reporting by combining hundreds of reports and dashboards into one unified system. This helps CEOs, business leaders, and decision-makers access real-time insights easily. Marketing is another area where AI plays a big role for Kenvue. “For example, a mass-market consumer product needs a different marketing approach compared to, say, a premium skincare brand. We use AI insights to measure campaign performance, identify untapped markets, and fine-tune our media strategy.” Why Bengaluru? As per Puneyani, Bengaluru stands out from purely technical hubs because it has a large talent pool that understands both technology and business. “Over the past two decades, Bengaluru has built deep expertise across various industries. This makes it easier for organisations like ours to grow faster by tapping into this knowledge. The city also has a strong startup culture, which encourages collaboration, innovation, and rapid progress in technology,” he said. Beyond talent, Bengaluru’s infrastructure and diverse culture make it even more appealing. With easy access to international airports, major tech parks, and research institutions, it is a convenient choice for global companies. “Manyata Tech Park, in particular, has become a key location for GCCs) because of its well-planned expansion, accessibility, and excellent connectivity,” he added. Talent Strategy of Kenvue Puneyani highlighted that Kenvue takes a hybrid approach to hiring. “Our focus is not just on technical expertise but also on business understanding and adaptability. We look for candidates who can grow with industry trends rather than those limited to a single technology.” He mentioned that the team is also exploring AI-driven hiring methods like automated screening and data-based candidate assessments. This helps make the hiring process more efficient while ensuring we bring in top talent that aligns with our goals. Speaking about the future of GCCs, he mentioned that one major challenge is talent development. Since the same group of professionals moves between multiple GCCs, hiring and retaining employees can be difficult. “A structured approach—through skill development programs, financial incentives, and regulatory support—is crucial.” He also mentioned that another long-standing issue for GCCs has been transfer pricing and taxation policies. “The hope is that this new framework (Budget 2025) will simplify these processes, making it easier to scale, innovate, and collaborate,” he said","excerpt":"In just a year and a half since its inception, Kenvue has grown significantly, evolving from a small team of 400-500 professionals into a dynamic force in technology and product innovation.","categories":["GCC"],"tags":["AI","GCC"],"author_name":"Shalini Mondal","publish_date":"2025-02-18T12:21:05","publication_year":"2025","word_count":1010,"keywords":["data science","ChatGPT","GCC","OpenAI","AI","cloud computing","ML","RAG","Aim","analytics","Azure"],"extracted_tech_keywords":["AI","ML","data science","analytics","ChatGPT","OpenAI","Aim","RAG","cloud computing","Azure"],"url":"https:\/\/analyticsindiamag.com\/gcc\/kenvue-has-a-late-mover-advantage-as-a-gcc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10066008,"title":"Best robotics courses to pursue after school","content":"According to Allied Market Research, the global robotics technology market size was valued at USD 62.75 billion in 2019, and is expeced to reach USD 189.36 billion by 2027, growing at a CAGR of 13.5 percent. To meet the growing demand for robotics professionals, institutes worldwide are offering specialised courses. But in India, most of the full time courses in robotics are offered at the postgraduate level ( as robotics requires knowledge of several disciplines). Below, we have put together a list of both graduate and post graduate courses in robotics. IIT Madras The Interdisciplinary Dual Degree programme in robotics at IIT Madras focuses on building professionals who can contribute in the design, development, and implementation of robotic systems. The course teaches basic robotic technologies used across various applications, kinematics, dynamics, and control of Industrial and field\/service robots, application of AI, neural networks and reinforcement learning in robotics, design of robotic systems for new applications etc. Aerospace Engineering, Applied Mechanics, Civil Engineering, Computer Science and Engineering, Electrical Engineering, Engineering Design, Mechanical Engineering and Ocean Engineering are offering the Dual Degree programme. The program is open to a B. Tech student or a Dual Degree student of IIT Madras in any discipline except biosciences. The student should have a CGPA of 8.0 or above up to 5th semester. For more details, click here. Robotics Engineering, University of California, Santa Cruz This course covers the principles and practices of robotics and control along with the scientific and mathematical principles upon which they are built. The course serves as a solid foundation for higher studies and also make the students job-ready. It is recommended that high school students applying to the BSOE have completed four years of mathematics and three years of science in high school, if possible one year each of chemistry, physics, and biology. For more details, click here. Colorado School of Mines Computer science and robotics and intelligent systems: The students can major in computer science (CS) and choose a focus area of CS + Robotics and Intelligent Systems. The students will also take courses from the Departments of Mechanical Engineering and Electrical Engineering. The course wil help students to understand “how to bring life to systems and give machines the ability to operate autonomously”. Mechanical engineering-robotics, automation and design: Students majoring in mechanical engineering can choose robotics courses as electives. Here, the students will learn the basic mechanisms, modeling, control, and design of robotic systems. Bachelors of science in engineering-robotics and automation focus area: For the bachelor of science in engineering, students can choose from diffrent focus areas including iRobotics and Automation. For more details, click here. Mechatronic and Robotic Engineering-The University of Sheffield The three-year course will teach engineering skills in areas like mechatronics and robotics such as mechanical design, electronics, computing, intelligent systems and control. In the first year, the students will be exposed to foundational knowledge in areas such as mathematics, computing, control, electronics and embedded systems. They will learn how to control robotic systems and will work in a team to design, analyse and test robots, autonomous vehicles and other electro-mechanical systems. The topics covered in the second year inlcude advanced control theory, programming (C++), mechanical design, intelligent systems etc. The students will apply these skills to design a system using 3D CAD tools. Students will get to build this system in the University’s iForge makerspace. In the final year, students will opt for specialist modules that cover robotics and artificial intelligence (machine learning), digital signal processing and rapid control prototyping. Students will also work on an individual project. For more details, click here. Mechatronics and Robotics-Heilbronn University In this course, students will learn the functionality of modern mechatronic and robotic systems. The programme will focus on components of modern mechatronic systems such as IT,  the mechanical structure, communication technology and human-machine interfaces. The students will learn the basics of mathematics, computer science, physics, electrical and mechancial engineering. The students will have the option to specialise as per their preferences and should work on a Bachelor’s thesis by collaborating with industry. For more details, click here.","excerpt":"Most robotics courses in India at premier institutes are offered at the postgraduate level.","categories":["AI Trends"],"tags":["Courses"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-05-01T10:00:00","publication_year":"2022","word_count":681,"keywords":["API","machine learning","artificial intelligence","programming_languages:R","AI","neural network","Git","automation","C++","Courses","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","R","C++","Git","API","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-robotics-courses-to-pursue-after-school\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054352,"title":"The State Of Data Centres And Data Security, In Conversation With Sudhindra Holla","content":"Axis Communications offers intelligent solutions based on sight, sound and analytics to improve security and optimise business performance for a broad spectrum of industry segments and applications to enable a smarter and safer world. Headquartered in Sweden, the company has been operating for over 35 years. Analytics India Magazine got in touch with Sudhindra Holla, Director, Axis Communications, India & SAARC, to understand how the company combines human imagination with intelligent technologies and offers solutions based on analytics, sound and sight to optimise business solutions. AIM: What is the current status of data centres in India, and what are the instances of attacks on data centres in India in the last few years? Sudhindra: According to a report by Research and Markets, India still has a nascent data centre market when compared to the rest of APAC. However, it is expected to witness an investment of up to USD 8 billion by 2026. India has 23 unique third-party data centres around 80 facilities and several on-premises data centres owned by local enterprises. The Government of India’s initiatives to build Digital India has also aided the development of data centres in the country. One of the recent data centre attacks we witnessed was the cyber-attack on Dr Reddy’s Laboratories in 2020. A multinational pharmaceutical company headquartered in Hyderabad had to shut down its production facilities across the world due to a data breach. Earlier this year, Air India also suffered a massive data breach as a result of which all data on personal details of passengers, including credit card information, had been compromised. AIM: In order to provide security to data centres, what all AI and analytics technology do you employ? Sudhindra: AI and data analytics are at the core of ensuring security to data centres. Axis deploys Secure Boot to enhance the cybersecurity aspect of the cameras. Secure Boot is a built-in firmware intelligence that keeps a check on the quality of the parameters required by the video management system. It acts as the gatekeeper to the surveillance system. Axis cameras also notify if there is any tampering in the hardware and reboot itself to shut down and prevent any breach. Analytics is a commonly used technology in video surveillance. It is deployed to install intelligent video motion detection capability in cameras, which alerts users on the time, type and intention of the servers entering the core areas of the data centre. Features like loop intelligent video and motion reduction help clear scrubbing data and analyse information with video management. Data centres also require strong perimeter protection to safeguard the petabyte of information, the expensive equipment and the infrastructure that is used. Axis provides automatic perimeter protection by combining thermal cameras, fixed cameras, and audio enabled PVC cameras to detect any possible movement near the perimeter. This, in turn, allows integration wherein alerts and notifications to the command-and-control centre can also be automated. AIM: What role does CCTV surveillance play in securing data centres? Sudhindra: Data has become one of the world’s most valuable commodities. An integrated, multi-layered approach is essential to securing data centres from the perimeter to the core. Threats like the usage of drones to monitor a data centre for planning an attack can be serious, and hence, video monitoring should also encompass drone detection. Network video surveillance with license plate recognition technology ensures only authorised people on entering the site. To complement the system, technologies like radar detection can be used to track animate objects, animals, and people across expansive sites. Radar combined with network audio can also be used to manage people entering prohibited zones remotely. In the server room and racks, surveillance is critical to prevent breachers from implanting malware or spyware in the system. High-resolution cameras can be installed to pan and zoom when the entrances are compromised automatically. Access control technology and video verification can be used to access racks while creating an audit trail of individuals who have been inside earlier. AIM: Who are your potential customers, and what makes you stand out in the market? Sudhindra: Innovation is an important pillar of our core culture. We believe that innovation enhances the efficiency of our products and services that allow customers flexibility and convenience and thus propels our growth. Aligning our business strategies to the sustainable development goals, we ensure to build green cameras that not only offer lesser power consumption but compress videos saving 30-50% storage space which can then be used by data centres for its other storage purposes. Data centres have always been important, but with the growth of cloud services, virtualisation, and hyperscale computing, many organisations now depend on them for day-to-day business operations. We provide multi-layered security to protect these critical assets, from the perimeter to the server racks. We deliver secure, comprehensive data centre surveillance to monitor the premises with fewer manned patrols. Edge analytics take Axis hardware to a higher level, helping in efficiently detecting and resolving incidents in all areas of the data centres. We additionally ensure that the cameras have a quicker process of cooling which is critical for the segment. Additionally, the videos are encrypted on our cameras – thus making our cameras very safe, robust and cyber secure. AIM: Which all other sectors have tried and tested your solutions? Who are your competitors in the market? Sudhindra: We have had a strong pan India presence since our inception here for over a decade. We cater to several industry verticals like smart city, hospitality, retail, commercial, transportation, education, and healthcare apart from government-initiated projects. We have collaborated with Kolkata Police, Salt Lake Stadium, Kalinga Institute of Medical Sciences, Bhubaneswar, and Hyderabad police, to deploy our products and solutions. We have also been associated with various government PWD projects like the PWD Western Court. Our other collaborations include Café Coffee day, Ceat Tyres, Mindtree, Mahindra and Mahindra, and Fortis Hospital, to name a few. AIM: Securing data is of utmost importance; what sort of security is provided to avoid data theft, and who will access them? Sudhindra: Data centres need to be protected both externally and internally. Breaches often start with physical access. From protecting the data centre building to securing the firmware, an integrated effort to mitigate risks is important. To protect cameras from vulnerability, we integrate the firmware ourselves to ensure cybersecurity. The firmware is loaded and packed into carton boxes by Axis, and if anyone tries to tamper with the firmware, it reboots itself to prevent data from being compromised. To prevent illegal access to the data centre building, access control technology enables video verification of visitors and credentials through identity cards or mobile phones to allow access to only specific individuals with authorisation. Video surveillance safeguards the security of the entire building, the people inside, and the operations of the data centre. AIM: As far as the pandemic is concerned, have you seen a rise in demand for your services? What are your future plans to extend them? Sudhindra: The pandemic has reinforced the concept of working remotely. This has led to the increased need for strengthening perimeter security with reduced manpower present at the premises. Catering to the need of the hour, we have offered numerous solutions that increase security, maximise business performance and operational efficiency, and, above all, allow us to carry on with our daily business while protecting ourselves and the people around us – today and tomorrow. We have thus seen an increase in the uptake of products and solutions due to the increasing demand for remote monitoring solutions. As per our service agreement, all our products come with a five-year warranty. The customer can opt for software and hardware upgradation as and when required. Thus, instead of providing an annual service model, we provide support in case of complications arising at the time of installation or faulty cameras through email, phone, or chat support, apart from ensuring holistic support by our customers. We are more focused on strengthening our end-to-end solutions. Next year, we are also looking at launching AXIS Control and a new series of explosion-protected cameras in India. We are also building on creating a robust ecosystem for our thermal cameras and services.","excerpt":"From protecting the data centre building to securing the firmware, an integrated effort to mitigate risks is important.","categories":["AI Features"],"tags":["Data centres in India","Data Security","Interviews and Discussions"],"author_name":"kumar Gandharv","publish_date":"2021-11-28T13:00:00","publication_year":"2021","word_count":1362,"keywords":["Go","programming_languages:R","AI","Data Security","innovation","Git","RAG","Aim","analytics","Data centres in India","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Git","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-state-of-data-centres-and-data-security-in-conversation-with-sudhindra-holla\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25647,"title":"Meet IBM&#8217;s Project Debater, World&#8217;s First AI Ready To Argue With Humans","content":"The day has finally come when an artificially intelligent entity engaged in a live public debate with a human. Earlier this week, at an event held at IBM’s Watson West site in San Francisco, a champion debater and IBM’s AI system, Project Debater, began by preparing arguments for and against the statement, “We should subsidise space exploration.” Since 2012, a global IBM Research team led by Haifa, Israel lab endowed Project Debater with three capabilities, each breaking new ground in AI: Data-driven speech writing and delivery Listening comprehension that can identify key claims hidden within long continuous spoken language Modelling human dilemmas in a unique knowledge graph to enable principled arguments Arvind Krishna, director at IBM Research wrote in a blog post, “Just think about that for a moment. An AI system engaged with an expert human debater listened to her argument, and responded convincingly with its own, unscripted reasoning to persuade an audience to consider its position on a controversial topic!” Krishna added in a statement that for the initial demonstrations of this new technology, IBM had selected from a curated list of topics to ensure a meaningful debate. But Project Debater was never trained for those topics. This new development has showcased IBM Research’s mission to develop a broad AI that learns across different disciplines to augment human intelligence. In this case, Project Debater explores new territory — it absorbs massive and diverse information and perspectives to help people build persuasive arguments and make well-informed decisions. This AI-based Project Debater is not going to be deployed commercially yet. In fact, reports have suggested that IBM’s record on bringing AI to the real world is mixed. The company entered its Watson AI technology in the quiz show Jeopardy! in the year 2011. The machine won, but IBM still hasn’t disclosed how much revenue Watson has been generating since.","excerpt":"The day has finally come when an artificially intelligent entity engaged in a live public debate with a human. Earlier this week, at an event held at IBM’s Watson West site in San Francisco, a champion debater and IBM’s AI system, Project Debater, began by preparing arguments for and against the statement, “We should subsidise space exploration.” […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Debate","IBM"],"author_name":"Prajakta Hebbar","publish_date":"2018-06-22T07:44:36","publication_year":"2018","word_count":309,"keywords":["Go","programming_languages:R","AI","data-driven","programming_languages:Go","Aim","IBM","GAN","AI Debate","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meet-ibms-project-debater-worlds-first-ai-ready-to-argue-with-humans\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004008,"title":"Biggest AI Acquisitions For The Year 2020, So Far","content":"Although the year 2020 started with a crisis, the technology landscape has drastically gained its momentum, and artificial intelligence has played a crucial role amid this pandemic. And that’s why the majority of companies are trying to get their hands on this technology, either by hiring AI and analytics experts or by acqui-hiring AI startups. Every year, many AI companies get acquired, especially the promising startups that get gobbled up by large tech companies in order to expand their AI capabilities. Similarly, this year also saw many exciting mergers and acquisition in the AI space, despite the COVID pandemic. Here are the top AI acquisitions of 2020, so far — in random order: Apple Acquired Xnor.ai In January 2020, Apple announced that the company had acquired an edge-based AI startup — Xnor.ai for a reported $200 million. The company, Xnor.ai, employs machine learning and image recognition tools for delivering AI capabilities on low-resource edge devices and has been one of the promising AI companies to acquire. The acquisition by Apple was an attempt to direct its efforts towards edge computing to keep user information more private and speed up processing. Not only edge-based computing brings back the data from the cloud to the device, improving the response times, but also allows for faster programs that would use significantly fewer data. Also Read: Is Apple Finally Going To Build Its Own Cloud Platform? NVIDIA Bought Mellanox In April 2020, NVIDIA announced the completion of its acquisition of Mellanox Technologies, an Israeli-American supplier of computer networking products for an amount of $7 billion. An acquisition which was initially announced in March 2019, was delayed by a year to get approval from all necessary authorities in order to proceed. With this acquisition, NVIDIA claimed to merge the expertise of both companies to address the challenges of increasing global demand for consumer internet services and the application of artificial intelligence to accelerate data science right from the cloud to edge and robotics. Further, this deal will enable customers to achieve higher performance, better utilisation of computing resources and reduced operating costs. Also Read: What’s Behind NVIDIA’s Strong Q1 2020 Performance? Amazon Bought Zoox In June, this year, Amazon announced its acquisition of autonomous driving startup — Zoox. Founded in 2014, Zoox has already raised nearly $1 billion of funding to develop autonomous driving technology to provide a full-stack solution for ride-hailing. With this acquisition, Amazon aimed to build its position in the self-driving industry. Along with this, Amazon has also been heavily investing in developing drones or autonomous delivery robots to enter the robot delivery game. With Zoox, Amazon will utilise the company’s autonomous technology in its delivery process, while offering massive scale and resources to the startup to develop its vision. Also Read: After Self-Driving Cars, Amazon Now Aims For Space Snap Acquired AI Factory Snapchat’s parent company, Snap, earlier this year, in January, announced their acquisition of a Ukrainian computer vision and AI-based startup — AI Factory. The company has been known to work with AI Factory to build its Snapchat’s animated selfie-based video feature. According to reports, the company has been acquired with a reported amount of $166 million to boost the augmented reality features for Snapchat as well as its eyeglasses, Spectacles. Interestingly, the company, AI Factory, was co-founded by Victor Shaburov, who has been Snap’s director of engineering but left in May 2018 to found AI Factory. Also Read: With Snap’s AI Factory Acquisition, Deep Fake GIFs Will Go Mainstream Accenture Acquires Byte Mid of this year, in May 2020, Accenture has also acquired a big data analytics company — Byte Prophecy, in order to address the increasing demand of their customers for enterprise-scale artificial intelligence and digital analytics. With buying Byte Prophecy, Accenture aims to improve the speed and agility of delivering advanced analytics and AI solutions to its enterprise customers. With Byte Prophecy’s technology and consulting skills, along with its client-centric innovation culture, Accenture aims to add nearly 50 data science and data engineering experts, with a particular focus on insight automation. Also Read: Accenture Acquires Revolutionary Security To Expand Its Cybersecurity Practice Brillio Acquires Cognetik In an attempt to increase analytics capabilities, Brillio, a digital technology consulting and solutions company, in July 2020, announced the acquisition of Cognetik. With the addition of this data and insights company, Brillio aims to strengthen its analytics services for its customers. The capabilities of Cognetik in customer and marketing analytics and experience within the Adobe ecosystem will also enable Brillio to provide advanced industry solutions for personalisation and omnichannel experiences. Furthermore, the delivery presence of Cognetik in Romania will allow Brillio to expand its global delivery capabilities and hire skilled talent in the EU continent and service clients in development centres outside of India. Also Read: This Is How Brillio Is Responding To Dynamic Challenge In Real-Time Approach Tech Mahindra acquires Zen3 Infosolutions Last but not least, in February 2020, Tech Mahindra also announced buying of Zen3 Infosolutions, a US-based AI company, for $64 million. According to the deal, the Indian subsidiary was acquired by Tech Mahindra, India, whereas the Tech Mahindra, US acquired the Zen3 Infosolution in America. The deal was done all in cash and aimed at advancing Tech Mahindra’s AI abilities with Zen3’s expertise in software product engineering, DevOps testing, machine learning, AI and analytics. With similar acquisitions over the years, Tech Mahindra has been showing promising capabilities in enhancing its services in the AI industry with continuous acquisitions of AI-based firms. Cisco Acquired ThousandEyes In May, this year, the networking behemoth, Cisco systems announced its purchase of ThousandEyes, a network intelligence company headquartered in San Francisco. Reported of $1 billion, from an undisclosed source, the deal was a part of Cisco’s acquisition spree to expand its cloud portfolio. According to the release, the company will be a part of Cisco’s Networking Services run by Todd Nightingale. The deal took place amid COVID pandemic, thus was done online to address the future where the internet is playing a crucial part of a business’s IT strategy. Furthermore, this deal will remove the delineation between public and private networks. Microsoft Acquired Softomotive May, this year, Microsoft also made an announcement of acquiring UK-based robotic process automation Softomotive. With this deal, the tech giant stated that it would be using the company’s robotic desktop automation tool WinAutomation and ProcessRobot for building software robots and enterprise RPA platforms. The announcement was made at Microsoft’s virtual event for this year where the company claimed to build Softomotive’s connectors for applications from SAP, as well as legacy terminal screens and Java. Softomotive’s despot automation tools will also provide customers with additional options for anyone who can build a bot and automate Windows-based tasks and enable RPA connectivity to various new apps and services. Also Read: Why Microsoft Acquired An RPA Company? Intel Bought Startup Moovit May 2020, Intel bought an Israeli smart-transit startup — Moovit, for approximately $900 million. After Moovit’s popularity for providing real-time updates to ride-sharing service users, the chipmaker giant decided to integrate it. But, the company will now push Intel’s Mobileye to become a complete mobility provider, including robotaxi services — though remaining an independent subsidiary. The company previously already did a $50 million investment round in Moovit in 2018, which with this acquisition will enhance the safety vision technology, which is critical for driverless vehicles. Also Read: Why Intel Acquired Moovit Accenture Acquired Mudano Earlier this year, in Feb, before acquiring Byte Prophecy, Accenture purchased Mudano, which is a UK-based strategic data consultancy firm. The acquisition of the company was aimed at enhancing Accenture’s analytics, data and artificial intelligence capabilities, and will join its applied intelligence unit employing data scientists, data engineers and AI professionals. Accenture has been known for its acquisition-based growth structure, whereby bringing in new talent and capabilities, the company focuses on having organic growth. Also Read: Accenture Extends Its Innovation In AI With A New Hub In Pune ServiceNow Purchased Loom Systems One of the initial acquisitions of the year, in Jan 2020, ServiceNow announced its acquisition of Israel-based AIOps company — Loom Systems. Founded in 2015, purchase of Loom Systems will help ServiceNow to expand its AIOps Coverage. Further, the deal aims at bringing together the AI innovations of Loom Systems along with ServiceNow’s AI Ops capability to help customers prevent IT issues. Also Read: Telangana Academy For Skill and Knowledge Collaborates With ServiceNow Appian Acquired Novayre Solutions In an attempt to create a one-stop-shop for automation, with best-in-class solutions for workflow, artificial intelligence and RPA, Appian bought RPA company Novayre Solutions. Novayre Solutions SL is the developer of the Jidoka RPA platform, which is currently the highest-rated RPA software on Gartner Peer Insights (>50 reviews) making Appian massively benefitted with the deal. Appian’s first-ever acquisition will extend its lead in low-code automation by adding RPA capabilities. Appian claims to develop end-to-end processes where humans, robots, and artificial intelligence can coexist together in a coordinated way.","excerpt":"Although the year 2020 started with a crisis, the technology landscape has drastically gained its momentum, and artificial intelligence has played a crucial role amid this pandemic. And that’s why the majority of companies are trying to get their hands on this technology, either by hiring AI and analytics experts or by acqui-hiring AI startups. […]","categories":["AI Features"],"tags":["ai consulting","augmented intelligence for smart industry","automation testing","big data developer skills","data analytics acquisitions","Mergers and Acquisitions","mergers and acquisitions business intelligence","Scale Big Data"],"author_name":"Sejuti Das","publish_date":"2020-08-04T15:00:00","publication_year":"2020","word_count":1493,"keywords":["mergers and acquisitions business intelligence","ai consulting","data science","artificial intelligence","automation testing","big data developer skills","AI","machine learning","augmented intelligence for smart industry","image recognition","computer vision","data analytics acquisitions","RAG","Aim","analytics","Scale Big Data","edge computing","Mergers and Acquisitions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","data science","analytics","Aim","RAG","image recognition","edge computing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/biggest-ai-acquisitions-for-the-year-2020-so-far\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":40126,"title":"A Complete Guide For Beginning With K-Nearest Neighbours Algorithm In Python","content":"With lots of Machine Learning algorithms out there it is often hard and confusing as to which to choose for our model especially for those who are just preparing for a big dive into data science. And no, I’m not talking about Model Selection here. Of all the algorithms there is, there has to be a simple algorithm that is well equipped and has the capability to overcome most of the complex of algorithms. KNN is one such. Both simple and in most cases elegant or at least not too disappointing. If you are just starting your Data Science journey and you want to start humble, KNN is the best bet. Not only that it is easy to understand but you can be a master of it in a matter of minutes. This article will get you kick-started with the KNN algorithm, understanding the intuition behind it and also learning to implement it in python for regression problems. What is KNN And How It Works? KNN which stands for K-Nearest Neighbours is a simple algorithm that is used for classification and regression problems in Machine Learning. KNN is also non-parametric which means the algorithm does not rely on strong assumptions instead tries to learn any functional form from the training data. Unlike most of the algorithms with complex names, which are often confusing as to what they really mean, KNN is pretty straight forward. The algorithm considers the k nearest neighbours to predict the class or value of a data point. Let’s try to understand it in a simple way. Consider a hypothetical situation in which there is a new student joining your class today say Class 5. None of your friends or class teachers know him yet. Only the new student knows to which class he belongs. So he shows up during the assembly and finds his class and stands in line. Since he is standing in the line that has only class 5 students in it, he is recognized as a Class 5 student. KNN predicts in a very similar fashion. For example, if it was KNN predicting that which class the student belongs to, it would simply consider the closest distance between the new student and the k number of students standing next to him. The k can be any number greater than or equal to 1. If it finds out that more number of students from class 5 are closer to him than any other classes, he is recognized as a Class 5 student by KNN. While the above example depicts a classification problem, what if we were to predict a continuous value using KNN. Well, in this case, KNN simply finds the average of all the values of the neighbouring data points and assign it to the new data point. Say for example if KNN is to predict the age of the new student in the above example, it calculates the average age of all its closest neighbours and assigns it to the new student. The Intuitive Steps For KNN The K Nearest Neighbour Algorithm can be performed in 4 simple steps. Step 1: Identify the problem as either falling to classification or regression. Step 2: Fix a value for k which can be any number greater than zero. Step 3: Now find k data points that are closest to the unknown\/uncategorized datapoint based on distance(Euclidean Distance, Manhattan Distance etc.) Step 4: Find the solution in either of the following steps: If classification, assign the uncategorized datapoint to the class where the maximum number of neighbours belonged to. or If regression, find the average value of all the closest neighbours and assign it as the value for the unknown data point. Note : For step 3, the most used distance formula is Euclidean Distance which is given as follows : By Euclidean Distance, the distance between two points P1(x1,y1)and P2(x2,y2) can be expressed as : Implementing KNN in Python The popular scikit learn library provides all the tools to readily implement KNN in python, We will use the sklearn.neighbors package and its functions. KNN for Regression We will consider a very simple dataset with just 30 observations of Experience vs Salary. We will use KNN to predict the salary of a specific Experience based on the given data. Let us begin! Importing necessary libraries import pandas as pd import numpy as np import matplotlib.pyplot as plt Loading the dataset data = pd.read_excel(\"exp_vs_sal.xlsx\") You can download this simple dataset by clicking here, or you can make your own by generating a set of random numbers. Here is what it looks like : Let’s visualize the data data.plot(figsize=(7,7)) Output: data.plot.scatter(figsize=(7,7),x = 0, y = 1) Output: Creating a training set and test set Now we will split the data set to a training set and a test set using the below code block from sklearn.model_selection import train_test_split train_set, test_set = train_test_split(data, test_size = 0.2, random_state = 1) Classifying the predictor(X) and target(Y) X_train = train_set.iloc[:,0].values.reshape(-1, 1) y_train = train_set.iloc[:,-1].values X_test = test_set.iloc[:,0].values.reshape(-1, 1) y_test = test_set.iloc[:,-1].values Initializing the KNN Regressor and fitting training data from sklearn.neighbors import KNeighborsRegressor regressor = KNeighborsRegressor(n_neighbors = 5, metric = 'minkowski', p = 2) regressor.fit(X_train, y_train) Predicting Salaries for test set Now that we have trained our KNN, we will predict the salaries for the experiences in the test set. y_pred = regressor.predict(X_test) Let’s write the predicted salary back to the test_set so that we can compare. test_set['Predicted_Salary'] = y_pred The new test_set will look like this : Visualizing the Predictions vs Actual Observations plt.figure(figsize = (5,5)) plt.title('Actual vs Predicted Salary') plt.xlabel('Age') plt.ylabel('Salary') plt.legend() plt.scatter(list(test_set[\"Experience\"]),list(test_set[\"Salary\"])) plt.scatter(list(test_set[\"Experience\"]),list(test_set[\"Predicted_Salary\"])) plt.show() Output The blue points are the actual salaries as observed and the orange dots denote the predicted salaries. From the table and the above graph, we can see that out of 6 predictions KNN predicted 4 with very close accuracy.","excerpt":"With lots of Machine Learning algorithms out there it is often hard and confusing as to which to choose for our model especially for those who are just preparing for a big dive into data science. And no, I’m not talking about Model Selection here. Of all the algorithms there is, there has to be […]","categories":["Deep Tech"],"tags":["knn","regression"],"author_name":"Amal Nair","publish_date":"2019-06-04T10:11:02","publication_year":"2019","word_count":975,"keywords":["data science","NumPy","machine learning","TPU","AI","knn","regression","RAG","Python","Matplotlib","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","data science","Pandas","NumPy","Matplotlib","RAG","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-complete-guide-for-beginning-with-k-nearest-neighbours-algorithm-in-python\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172240,"title":"India’s Measured March Towards Agentic AI","content":"As the first half of 2025 draws to a close, the year is already defined as one in which agentic AI takes centre stage. While generative AI has shown the value of machine learning in automating tasks and producing content, agentic AI marks the next step, enabling systems to perceive, reason, and act independently, much like human decision-makers. AIM spoke to Siddharth Murlidharan, vice president of financial services at Tredence to understand the current state of agentic AI, especially in the Indian context, and to find out what lies ahead beyond AI agents. While the adoption of agentic AI in India is undoubtedly underway, its progress remains deliberate and measured. According to Murlidharan, the real question is not whether the Indian or the US market is ready—it is whether organisations have the right mindset to adopt this technology. India’s Position Companies that are open to rethinking how they work are the ones that are moving forward, whether they’re based in the US, India, or the Middle East. The appetite to explore and adopt agentic AI is present in India, but full-scale integration is happening gradually. “We are now seeing a lot of our customers use agentic AI for very specific use cases. Once you get past that first hurdle, the next stage in the evolution is to infuse agentic AI across the entire landscape,” he added. Murlidharan talked about three main challenges that are responsible for slowing things down. First is adoption, as it takes effort and time to change established processes. Regulation is another concern, considering financial services operate under strict compliance rules, and agentic systems must meet high standards of transparency, fairness, and data protection. Education also plays a role, with many stakeholders lacking a clear understanding of what agentic AI can and cannot do, resulting in hesitation and unrealistic expectations. Real Use Cases in Financial Services Tredence has already worked on agentic AI use cases in financial services, including wealth management and lending. For instance, in wealth management, agents assist with portfolio rebalancing, allowing financial advisors to shift their focus to strategy and client communication. In lending, agentic systems can process up-to-date information like employment verification and market volatility to improve credit decision-making. In fraud detection, agents act on patterns faster than traditional systems, provided they remain explainable and auditable. The core reason for using agentic AI lies in speed and action. Traditional AI can detect patterns, but agentic AI can act on them in real time. This is especially important in areas like fraud prevention, anti-money laundering, and dynamic credit assessment, where rapid response is crucial. However, this only works when those decisions can be understood and justified to both business teams and regulators. The Human Element Won’t Go Away Despite the progress, Siddharth is clear that fully autonomous systems are not viable in finance, at least not in the foreseeable future. Financial decisions involve trust, risk, and regulation, all of which require human oversight. A more realistic approach is hybrid decision-making, where agents manage the routine and humans handle exceptions, complexities, and customer engagement. On India’s regulatory framework, he believes that the focus so far has been on preventing data misuse, ensuring unbiased decisions, and protecting end users. Tredence’s approach is to work closely with clients’ compliance teams to ensure that any AI deployment fits securely within existing frameworks. Agentic AI won’t eliminate jobs, but it will change them. “”I don’t see agents right now replacing humans… What people need to be aware of is… if I don’t use agentic AI, then I will be on the back foot as compared to somebody who does,” he said. Roles will shift from manual execution to higher-level work, like strategic decision-making and client engagement. For employees, the key is to adopt these tools and use them to do their work better and faster. Next Step: Super Agents He mentioned that companies are currently focusing on specific use cases for agents to build confidence and show value, with the next step being the development of advanced ‘super agents’ to handle full customer journeys or complex workflows. “You start by going deep on a single use case to deliver clear value. The next stage is scaling agentic AI across the landscape with super agents managing entire customer relationships or complex workflows,” he added. What sets India apart is the promising opportunity to deliver agentic services in multiple languages, potentially broadening access to financial services for a much broader population. With an infrastructure like UPI already in place, India is in a strong position to lead in real-time, AI-powered financial systems.","excerpt":"While the adoption of agentic AI in India is undoubtedly underway, its progress remains deliberate and measured.","categories":["AI Highlights"],"tags":["agentic ai","AI (Artificial Intelligence)","Tredence"],"author_name":"Aditi Suresh","publish_date":"2025-06-24T10:06:02","publication_year":"2025","word_count":761,"keywords":["Go","agentic AI","machine learning","AWS","AI","agentic ai","Aim","generative AI","Rust","Tredence","R","AI (Artificial Intelligence)","fraud detection"],"extracted_tech_keywords":["AI","machine learning","generative AI","agentic AI","Aim","fraud detection","AWS","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/indias-measured-march-towards-agentic-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10047859,"title":"Detection &#038; Classification Of AI Generated Fake News","content":"Fake news is an infodemic, a disease worse than anything else we’ve ever seen. And it’s been around for longer than Covid19. Sometimes, fake news has little impact. But when times are uncertain, and a global crisis is in effect, people look for information that alleviates their fear. Unfortunately, fear leaves people more susceptible to accepting misleading information as the real deal. A limited amount of reliable data in some instances also encourages people to take misinformation willingly. Studies conducted during the pandemic showed that false news might have threatened public health globally – by touting the pandemic as a hoax and Covid19, regular flu. Another set of studies by the UK, in association with WHO, revealed that nearly 6,000 people worldwide were hospitalized due to Covid19-related fake news. It also resulted in the deaths of at least 800 people. All of this, within the first three months of the pandemic. The Internet has boosted the spread of fake news from a single place to all over the world. Social media is free to use, and therefore, anyone can post a story that goes viral. But Social media platforms and online news outlets are undertaking countermeasures. They’re constantly updating their algorithms to track and block fake news. It isn’t easy, though. While most articles are manually written, innovative fake news creators use natural language generation techniques to create excellent, realistic counterfeit ones. These models have created an urgent need to distinguish fake from real news and detect human-generated and machine-generated fake news. The need for better Fake News detection During the course of the project, several issues were observed that needed solutions. These gaps helped identify and create the scope of the study. With the introduction of AI-generated fake news, the spread of misinformation has increased manifold. Adding to the challenge, AI models can mimic human language, making them near-impossible to detect. They’re used to spread hoaxes, propaganda or any other form of deception through online mediums, mainly social media. Social media rapidly disseminates information across the globe today. However, its algorithms are not well-equipped enough to detect fake news with accuracy. Considering social media is one of the top spreaders of fake news, developing a social media-compatible model would be critical. Complex language models are currently in use to find and block the spread of fake information. But the models train in observing text generated by themselves. And aren’t tested adequately for human-generated fake news. If a particular piece of fake news is generated by a machine but edited a little bit by a human, the models find it difficult to classify it as fake. While there were various datasets for fake news generated by humans, AI-generated datasets were difficult to find. It is because AI language models are still considered dangerous for public use. Hence, data regarding these datasets isn’t readily available. Building the Fake News Detection Model Keeping the above gaps in mind, this research project sought to find answers to two chief questions: ● With the use of different classification techniques, how can machine-generated news be identified effectively? ● Is there scope to build a model which can detect both human and machine-generated fake news? Data Sets The project looked into several datasets, including a fake news repository in use for ongoing research at Arizona State University (ASU). FakeNewsNet was one of the datasets used. It is a comprehensive dataset that contains fake news content, social context, and dynamic data that can facilitate detection. Finding AI-generated fake news dataset was challenging. Therefore, the GROVER model for Neural Fake News generation was settled on. Using the GROVER model, a sample dataset of machine-generated articles was created. The articles covered varied domains such as health, politics, society, climate and more. GROVER is an AI-based writing and language system developed by the University of Washington in association with the Allen Institute for AI. It can write fake news on a range of topics and in several styles. However, because it’s a one-of-a-kind model, there’s no other model yet that can identify the fake news it generates. The project put FakeNewsNet and GROVER together to experiment and distinguish between human-generated and neural fake news. Methodology To better understand the data gathered, the project experimented with three news writing styles – human-generated ‘Real’ and ‘Fake’ news and ‘AI-generated Fake news.’ Multiple iterations using linear classifiers with TF-IDF feature set on unigrams and bigrams were performed. Starting with length, the project observed that the length of news articles written by humans and AI were similar. Factual news articles written by humans had a few longer ones, but these weren’t significant in number. The articles were also tested for vocabulary. While AI models could replicate human writing styles, vocabulary is unique to each individual’s creative capacity. The research found that machine-generated articles had very few distinct words in each article. This finding indicated that AI language models have a limited word library as compared to humans. Checking for syntax further showed that machine-generated articles use more common nouns than articles written by humans. Key Outcomes After the language generation models were analyzed and compared, the project found they can imitate the writing styles of humans. However, they used lesser distinct words than humans, which indicated they have a limited word collection. Plus, they used proper nouns sparingly, further proving that they lack facts or evidence. They try not to share a lot of specific information, which makes factual articles genuine. Once the project was completed, it concluded that simple linear algorithms could pick machine or AI-generated fake news quite well. Also, these algorithms can separate these fake articles from real articles or human-written ones. Future Scope of this Research for Fake News Detection Findings from this research can be used for various activities: Researchers can further analyze and compare several language generation models with the human writing style and inter-model styles. It will increase their scope and help with better fake news detection. Even though TF-IDF classifiers worked well, there are possibilities of exploring other features to improve the model and make it a generic fit. While the project focused on text-based news articles and language models, AI algorithms can also analyze other features such as images, videos, date and time, sources, website, and domain for valuable information. Teaching the detection model to trace the source of a machine-generated fake news article to the language model from which it originated will be a huge step forward. This development will ensure mitigating and blocking the spread of misinformation. _____________________________________ Poorva Sawant is an upGrad learner, and as a part of her program, she has developed the thesis report titled, Detection & Classification Of AI-Generated Fake News.","excerpt":"This is one of the top voted thesis papers from upGrad’s online working professional programs in partnership with one of the UK’s leading universities.","categories":["AI Trends"],"tags":["ai generated images"],"author_name":"Poorva Sawant","publish_date":"2021-09-07T12:00:00","publication_year":"2021","word_count":1112,"keywords":["Replicate","Go","ai generated images","API","programming_languages:R","AI","programming_languages:Go","RAG","ViT","GAN","R"],"extracted_tech_keywords":["AI","RAG","R","Go","API","GAN","ViT","Replicate","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/detection-classification-of-ai-generated-fake-news\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":18902,"title":"Top 10 Data Scientists in India – 2017","content":"It has been two years since Analytics India Magazine took up the initiative of identifying the brains behind the excellent work that is being done in the field of data science in India. Given a rise in popularity in the field, there are many professionals who are veering their way onto the data science field and India is witnessing a growing number of data scientists, playing crucial roles in various industries. For this year’s ranking, data scientists from various organisations and those working independently, were considered, irrespective of size and nature of work. Like last year, we also got in touch with data scientists that we know personally who might not necessarily be associated with an organisation. The top 10 names were concluded based on various parameters like pedigree, patents, papers and technical publications authored, competitions participated, pioneering work, knowledge and applicability of tools, ability to convince multiple stakeholders through data insights and many more. We also considered the inputs from expert team of leaders, past data scientists and evangelist, and their overall contribution to the analytics industry in the country, to file the list. Here’s the list of top 10 data scientists, to draw some inspiration and motivation from (in alphabetical order). Read our last year’s list here Ankur Narang Dr. Narang leads the Data Science & AI practice at Yatra Online Pvt. Ltd. as Senior Vice President, Technology & Decision Sciences. He has 23+ years of experience in Senior Technology Leadership positions across MNCs including IBM Research India and Sun Research Labs, CA, USA. With a B.Tech. & PhD IIT Delhi in CS&E, he has 40+ publications in top international Computer Science & Machine Learning conferences and journals, along with 15 granted US patents. He has held multiple Industrial Track and Workshop Chair positions, and has given invited talks in multiple conferences. As Senior Research Scientist at IBM Research, for key Telecom players in India and US, he designed and delivered real-time parallel recommendation algorithms, high throughput streaming analytics workflows and distributed graph analytics using GPUs over EDR\/CDR data. Further, he led design & implementation of AI based cognitive workflows for inverse problems using oil & gas production data. As Chief Data Scientist and AVP Data Science at Mobileum, he led development of voice and data fraud techniques using deep learning and travel prediction and campaign models over terabytes of CDR data. As CTO in a recent startup stint, he developed game theory, ML and optimization based novel approaches to pricing and revenue management for large Media & FMCG companies. At Yatra, he is working on AI based approaches for marketing and discounting optimizations and personalized chatbot experience. Avik Sarkar He currently heads the Data Analytics Cell at NITI Aayog (National Institution for Transforming India) as Officer on Special Duty (OSD) and believes that the challenges with analytics in governance are quite different from that of other industries. He is in charge of developing roadmap for use of data\/analytics for Governance and Policy making along with providing analytical insights for policy making across sectors like Direct Benefit Transfer, Innovation, Digital payments, Healthcare\/Nutrition, Agriculture, etc. He is engaged with the energy vertical at NITI Aayog where he is instrumental in various long term planning of future energy needs of India through initiatives like integrated energy modelling, energy data management, etc. With over 15 years of experience across different aspects of data analytics, statistical modeling, data and text mining across companies like IBM, Accenture, Nokia, NASA, Persistent Systems, etc., he has also contributed to various data and analytics related engagements with Singapore Government. While at IBM, Dr. Sarkar made significant contributions towards developing the Monte Carlo Simulation of SPSS and the Predictive Maintenance and Quality solution for the manufacturing sector. He holds a PhD from The Open University, UK, Masters from Indian Institute of Technology (IIT) Bombay and Bachelors from Calcutta University and has authored several technical publications and technology patents. Kiran R Kiran R is currently the Director, Data Sciences at VMware and has an experience in driving impact in both B2B and B2C organizations. He currently drives advanced analytics and data sciences support at VMware across verticals. Prior to VMware, he headed analytics & data sciences for Sales, Marketing & Customer at Flipkart, affiliates analytics at Amazon and for the e-business & search teams at Dell. Kiran has 3 filed US patents. During his days at Dell, he was awarded the 2012 India Innovator of the Year award by Michael Dell in person in 2012. He is the author of a Harvard Business Review Case study on data mining in partnership with IIM-B, that is taught in premier schools around the world. Kiran is a Kaggle grandmaster and was ranked in the top 10 Kaggle data scientists in 2013-14 with a top rank of 7, which remains the highest ever rank consistently held over a year by an Indian to this date. He is one of the winners of the prestigious KDD Data Mining Cup in 2014, organized by ACM-KDD (Association for Computing Machinery – Knowledge, Discovery and Data Mining). A computer science engineer and a post-graduate from Indian Institute of Management Kozhikode (IIM-K), Kiran is passionate about data sciences at work and outside of it. Nitin Sareen Nitin is a strong believer and practitioner of using advanced analytics as a strategic differentiator across domains and has a proven track record of creating impact by solving key business problems with astute analytical problem solving. He currently leads the Data Science group at WalmartLabs to leverage big data, data science & technology to enable faster & smarter business decisions. He is leading key initiatives to deploy algorithmic products that consume Walmart scale data and infuse smarter decisions across retail lifecycle like site selection, assortment optimization, pricing, demand forecasting, supply chain, store ops and enterprise wide decisions. These solutions are able to deliver multi-billion $ impact. The group is driving innovation leveraging emerging technology to experiment with AI; deep learning driven solutions for image, video & text analytics across various use cases. He is considered an analytics thought leader who is enabling creation of value for the organization. Prior to this he has held various key positions in MNCs like Citigroup, HSBC, FICO and GE across multiple roles. He was responsible for setting up and managing analytics groups in the areas of insurance analytics, consumer finance and retail credit risk management and gathered functional expertise across marketing and risk analytics. He has been focused on continuous learning and professional development for self and the teams he has led. An alumnus from Indian Statistical Institute (ISI), Calcutta, Nitin has over 17 years of extensive experience in the field of predictive analytics and data science projects. He is also an active speaker and panelist at leading data science conferences. Om Deshmukh Dr. Om Deshmukh is the Director, Data Sciences in the Analytics Centre of Excellence at Envestnet | Yodlee. Om and team drive foundational data sciences initiatives to mine actionable intelligence from the petabytes of data that flows through the Envestnet | Yodlee platform, enabling financial service providers to improve consumers’ financial wellness. He received his PhD from the Department of Electrical and Computer Engineering at University of Maryland, College Park, MS from Boston University and B.Tech. from BITS-Pilani. Prior to Envestnet | Yodlee, Om has worked at IBM Research and Xerox Research and has a deep expertise in machine learning and data analytics, particularly Bayesian Non Parametrics and Generative Models using Deep Neural Networks. At Xerox Research, he built a team of highly motivated researchers and developers who designed state-of-the-art Machine Learning and Multimedia (speech\/text\/video) Data Analytics algorithms. He successfully negotiated a first of its kind revenue sharing deal with an edtech startup to inject multimedia analytics into its products.  At IBM, he built consensus among executives and presented to the CEO, a technology strategy for an analytics-driven billion-dollar business in ‘Personalized Learning.’ Om has filed 45+ patents (including 2 high-value patents) and was published in 50+ international publications (multiple best paper awards). Ramasubramanian (Ramsu) Sundararajan Ramasubramanian (Ramsu) Sundararajan heads the AI function at Cartesian Consulting. He started his journey as a researcher at the Indian Institute of Management in Kolkata, where he wrote his doctoral dissertation on theoretical and algorithmic aspects of learning with a reject option. Prior to this, he earned his undergraduate degree from the Birla Institute of Technology & Science, Pilani. Since 2003, he has worked at GE’s Global Research division as well as at Sabre Airline Solutions’ Operations Research group, where he built statistical and machine learning models to find patterns in a variety of data sources in different domains. Examples include: patterns in the behaviour of retail banking customers that would indicate their future profitability or risk; patterns in medical image data that would provide early warning for diseases such as breast cancer and pneumoconiosis; patterns in sensor data that would help service engineers flag underperformance or impending failures in steam and gas turbines; and patterns in traveller behaviour that would help airlines determine trip purpose, ancillary purchase behaviour and so on. In his current role at Cartesian, he hopes to extend his hunt for interesting and actionable patterns to newer domains. Through it all, Ramsu has maintained a focus on both business value and innovation. He has co-authored over 20 publications in international conferences and journals, notably Interfaces, Journal of the Operational Research Society and the Journal of Revenue and Pricing Management. A firm believer in the “if you want to learn, teach” doctrine, he has regularly offered internal training programs and delivered guest lectures at premier educational institutes such as the IIMs and IITs on applied machine learning. Saigeetha A J Saigeetha A J is a senior data scientist specialized in Computer Vision, Machine Learning, Geophysical modeling and Robotics, and currently leads a team of image\/NLP experts at IBM. At IBM, she is extensively involved in delivering high end applications in image processing, video processing, OCR & Robotics along with NLP. Some of the significant contributions are quantifying wine brand popularity from social media images, video chunking & tagging, document digitization, Robotic presentation and path planning. She has over 10 years of experience in various industries namely automotive, manufacturing, winery, education\/learning, paints\/chemicals, social media and disaster\/hazard management. She holds a Ph.D in Physics from Gandhigram Rural Institute and carried out her research work in CSIR Fourth Paradigm Institute (formerly CSIR Centre for Mathematical Modeling and Computer Simulation). Her research work included development of a geophysical model for quantifying significant fault systems in India using GPS measurements. She then later went on to work in a leading paint company in colorimetric data analytics including developing an empirical model to predict paint toner constituents in a paint formula from reflectance. After this, she led a team of image scientists in developing advanced computer vision algorithms for ADAS & driverless cars such as pedestrian & car detection, traffic light detection etc. at a leading multinational. As a researcher, she has co-authored several research papers in leading journals of Elsevier, Indian Academy of Sciences. Apart from this, a couple of patents have been filed by her. Recently, she was adjudged “2017 Eminence & Excellence Star” performer at IBM. Shantanu Bose Shantanu has about 10 years of experience in devising data driven strategies for marketing and sales across various industries including Life Sciences, TMT and Retail. A post-graduate in Statistics & Informatics from Indian Institute of Technology, Kharagpur, he has extensive knowledge in complex statistical and machine learning models. His master’s thesis in image recognition was selected for presentation in 2008 IEEE Artificial Intelligence and Pattern Recognition Conference held in Florida. He joined Deloitte from campus and was instrumental in developing a few solutions which defined the industry standard- gamified business simulation for a Retail client, disease propensity models using lifestyle data; the latter was featured in the front page in the Wall Street Journal as industry disruptor. He then joined Novartis, where he led the efforts in solving the problem of integrating planning and evaluation the online and offline promotions for pharmaceutical industry. He Joined back Deloitte 3 years back to start the Life Sciences advanced analytics offering where he has worked with most of the big pharma companies on various consulting engagements such as drug launch strategy, patient experience redesign and market map for effective product strategy. He has lent is rich experience with marketing and sales organizations in cross-industry engagements, he is currently leading the global sales strategy transformation project for a TMT client. He is also engaged with the Deloitte product and innovation group to enhance and encapsulate some of the work on customer\/consumer engagement and sales transformation to build solutions which can be deployed at scale across industries to shape product and marketing strategy – faster, smarter and in a cost effective way. He has spoken in some of the leading data science conferences in India and abroad including Deloitte Analytics Summit and Cypher 2017. Shankar Viswanathan As part of ZS’s data science leadership team, Shankar is focused on shaping the advanced analytics capabilities of the firm to deliver real-world client impact powered by diverse data, algorithms and platforms. For over 16 years at ZS, Shankar has supported more than 20 global life science firms by providing fact-based insights for enhanced decision-making across a range of enterprise functions including sales, marketing and Health Economics and Outcomes Research (HEOR). For the past eight years at ZS’s Capability and Expertise Center in India, Shankar has been instrumental in driving the analytical offerings spanning foundational performance engine analytics to next-gen analytics powered by unstructured data, predictive analytics, machine learning and AI. Prior to his current role, he was head of the Pune office and also led the business consulting capability for India. Shankar has an interdisciplinary Ph.D. from Purdue (Artificial Intelligence + Operational Research applied to Chemical Engineering) and a B.Tech from IIT Madras. During the course of his doctoral thesis, he published 11 research papers on discrete event system representations to model batch process systems. He has been published in Value in Health, Society of Hospital Medicine and International Society for Pharmaceutical Outcomes Research (ISPOR). Subramanian M S (Mani) He heads the analytics at bigbasket.com – India’s largest online supermarket, with a focus on delivering definitive actionable insights that help enhance the customer experience. The analytics team at bigbasket leverages advanced tools, techniques and platforms to power a better customer experience. The analytics team under his guidance drives a.) diagnostic analytics to root-cause business problems, b.) predictive solutions including forecasting for perishable products, sales projections to support organization’s expansion and c.) prescriptive solutions including smart basket (a shopping assistant) and recommendations to help improve customer experience. With an experience of more than 20 years in analytics leadership, he has worked with companies like Dell, McKinsey, Infosys, Ernst & Young and PwC. An undergraduate with engineering degree in computer science from the University of Madras, an MBA from IIM-Ahmedabad and with a graduate engineering degree in supply chain management from MIT, Mani is a frequent speaker in industry and academic forums as an analytics expert. Some of his experience include being a speaker at IIM-B’s Retail Master Class, panel moderator\/member in various industry forums (Cypher, SCPC), conducting hands-on workshop in Analytics at UpGrad and other forums. He has also been part-time faculty delivering graduate level Analytics courses at SIBM, Bangalore and NMIMS, Bangalore.","excerpt":"It has been two years since Analytics India Magazine took up the initiative of identifying the brains behind the excellent work that is being done in the field of data science in India. Given a rise in popularity in the field, there are many professionals who are veering their way onto the data science field […]","categories":["AI Features"],"tags":["current leaders in self driving cars","data science india","data scientist india"],"author_name":"Дарья","publish_date":"2017-11-13T04:09:04","publication_year":"2017","word_count":2557,"keywords":["data science","artificial intelligence","machine learning","data science india","AI","neural network","ML","computer vision","NLP","deep learning","analytics","data scientist india","current leaders in self driving cars"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","NLP","computer vision","data science","analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-data-scientists-india-2017\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":39759,"title":"5 Non Work-Related Habits Every Data Scientist Must Inculcate","content":"To be on top of your game at your game in the field of data science, reaching office on time and staying late, working day in and out is not the only thing. One has to be super productive outside the workplace as well. In this article, Analytics India Magazine takes a look at some of the most effective habits that every data scientist, irrespective of his\/her designation, need to inculcate when they are off the workplace. Hang Out With Career-Centric and Successful Co-Workers The people you spend time with matters a lot in the way you think and set your mindset. It also affects the way you want to take your career ahead. So, it is always considered to be good practice to spend time with co-workers after heavy eight-plus hours or working. It would not only give you positive vibes but will also help you have a healthy workplace culture the very next day, which is a vital factor for every data scientist. While others prefer to have serious career talks, some prefer to relax over a cup of coffee. And both ways are considered to be best as your company is full of positivity. Also, outside the workplace meeting gives you the environment to ask and get to know things that you can’t inside the office. Minimise Use Of Electronic Gadgets A data scientist spends 9 hours of his\/her day at work in front of the computer — it stresses eyes as well as our entire body. There are many people across the world who tend to stick to their phones or laptops even after coming back from work, However, that shouldn’t be the case. Our eyes need rest too, just like every other part of our body, and too much of light of gadgets are not at all good as they mess up your sleeping habits. So, make sure you do not use any electronic gadget for a long time. Rather, try to meditate and relax and have a good sleep every single night. This habit will not only make your body and mind active but will also help you focus on your work. Make Reading A Habit If You Want To Be Successful You ask all the successful people in this world their key to success, they would definitely point to their bookshelf. Once you come back home from work, take some time out for reading. Knowledge is never-ending, so, the more you read the more knowledge you gain — it builds up, like compound interest. Talking about the data science domain, the learning process never stops, so make sure you keep your knowledge updated by reading books about the latest trends. Try to read educational books and publications more than novels, tabloids, and magazines, if you want to excel your data science career. Coach, Mentor, and Build Connections This is one of the biggest traits of a successful professional. If you are someone who always has some time free (at least on the weekends) try to mentor or coach people who need your help. Data science is not just a vast domain, it also expensive — not everyone can afford to take up a data science coach and not every company can afford a data scientist. So, it is not at all a loss if you step in and offer help — discuss ideas, solve problems, help others excel and grow as a data scientist. Furthermore, this provides you with an opportunity to make some connections with some of the best data science enthusiasts that would help you or support you in the future. Take Your Mind Away From Work When Spending Time With Family Data Science is one of the vast domains in the industry. It takes an immense amount of knowledge, concentration to solve some of the most complex problems. Meaning, it is not a cake walk for a data scientist to spend 9 hours at work. And sometimes, one of the best ways to be more productive at work is to take time off from work — not because you get tired of work but to relax your mind. After work hours when you spend time with family or your closed one, try not to use any office related gadgets — try to step away from work-related communication. It is important for every professional to strike a work-life balance, as it helps in relieving stress.","excerpt":"To be on top of your game at your game in the field of data science, reaching office on time and staying late, working day in and out is not the only thing. One has to be super productive outside the workplace as well. In this article, Analytics India Magazine takes a look at some […]","categories":["AI Trends"],"tags":["AI Workplace","Data Science","Data Scientist","TED Talks"],"author_name":"Harshajit Sarmah","publish_date":"2019-05-27T13:00:36","publication_year":"2019","word_count":734,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","TED Talks","ViT","analytics","Data Science","Data Scientist","R","AI Workplace"],"extracted_tech_keywords":["AI","data science","analytics","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-non-work-related-habits-every-data-scientist-must-inculcate\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10068941,"title":"How Merkle implements Ethical AI","content":"Few decades back, when telemarketing and direct email were just starting to pick up, a small company leveraged data and analytics to devise marketing campaigns. In 1988, 25-year-old David Williams acquired the company. The rest, as they say, is history. Today, Merkle—a dentsu company—is one of the leading customer experience management companies, trusted by Fortune 1000 companies and non-profit organisations worldwide. Headquartered in Columbia, Maryland, Merkle has 50+ offices throughout the Americas, EMEA, and APAC, with more than 14,000 employees. Merkle combines data, technology, and analytics with consumer insights to build hyper-personalised marketing strategies. The firm also leverages the experience of marketing experts specialising in consulting, creative, media, analytics, data, identity, CX\/commerce, technology, and loyalty & promotions to optimise its solutions. In 2019, dentsu International acquired a majority stake in Mumbai-based data and analytics company Ugam. Founded in 2000, Ugam—now a Merkle company—helps corporations, including Fortune 500 companies, accelerate their digital transformation. The company’s offerings range from retail and survey analytics, to technology implementation and data engineering. The idea behind this acquisition was to strengthen Merkle’s multi-year analytics strategy of creating a scaled on- and offshore shared analytics service across dentsu International. State of play “Back in the day, we used machine learning methodologies mostly in the predictive modelling space. And now, we use it for natural language processing, image recognition, next best actions and decisioning, and personalised customer experiences. AI has a role in all those different things. It’s not just about modelling anymore, it’s really about taking all of the data from every touch point and being able to process and analyse it and use it in different ways,” said Shirli Zelcer, global head of analytics, Merkle. Merkle offers a range of AI and analytics-based services, including: Analytics consultingCustomer analyticsCustomer experience optimization (CXO) analyticsMedia and digital analyticsData visualisation “We use AI and data analytics to optimise business workflows and marketing mix models. We have combined this IP into a no-code\/low-code platform. We’ve also trained our workforce to understand AI better, and created what we call AIOps. For example, the platform has a click-and-select feature for the technique you want to use. Then, there would be a bunch of people who validate that, and then it would be operational. There would be a feedback loop in which the AI would learn and then recommend in the next loop, and so on and so forth. We are at a stage where we are using AI for larger operations,” said Navin Dhananjaya, chief solutions officer at Ugam Solutions. Roadblocks “Historically, AI and machine learning have had the reputation of being a black box. I think at the forefront, one of the biggest issues was, understanding what was happening within the algorithms and how they were being created so that the clients felt comfortable using those outputs,” said Shirli. All roads lead to ethics As per Shirli, there are not enough conversations happening around ethical AI. “Back in the day, when we used to build models that were not based on machine learning and AI, everything was extremely transparent. You would be able to understand exactly what data was coming in, what attributes were used in the model, how they were related to each other, how they were interacting with each other with the final algorithm and the direction of all the variables, etc. And everything was highly explainable and transparent. As we introduced new methodologies like machine learning, people weren’t thinking about these at all.” “I think what we’re seeing now, especially in today’s culture, is that it’s becoming increasingly important to ensure that there’s no underlying bias. But unfortunately, what can sometimes happen in machine learning and AI is that these attributes that go into the model can be proxies for things that create bias, or variables can be combined in a way that you weren’t anticipating, which can create bias,” she added. As per Navin, digital acceleration has also been a catalyst. “In the past, AI was not as accessible as it is today, and it was always people first, and then a lot of that is pivoted to AI-first and then people doing the validation, so the outputs are more AI-driven now, which is good from maturity of AI\/ML,” he said. “I think that the use cases for AI have gotten much broader. But, unfortunately, this leads to potential bias and ethical issues,” Shirli added. Merkle’s answer “Our approach is not to dictate what is ethical and what is not ethical. Our approach is 100% transparency. You have to understand exactly what attributes are being used, how they’re being used, if any bias is being created, calculate fairness metrics, etc. Then, have a very honest conversation with our clients and say what you are comfortable with and what you want to put in the market,” said Shirli. Merkle leverages a combination of people, processes and technology to come up with marketing strategies. “We will onboard domain experts who will help us go through this journey and we will build a set of processes where people will ensure that whether it is fairness or whether it is anything related to data or ethics or explainability,” said Navin. Data activation is another area of focus for the company. “When creating audiences, we could potentially hand them off to media teams. The media teams have to ensure that they’re not activating it in a biased way, and then some focus on one population versus the other. So like when you activate against publishers, you might be creating bias, and then in the feedback, we’ve frightened of the samples that come back for the models to build on themselves continuously. So there are a lot of different areas that we’re going to have to focus on to make sure that there’s no bias,” said  Shirli.","excerpt":"Historically, AI and machine learning have had the reputation of being a black box.","categories":["AI Features"],"tags":["head of analytics","Interviews and Discussions","ugam","ugam solutions"],"author_name":"Sri Krishna","publish_date":"2022-06-14T10:01:00","publication_year":"2022","word_count":963,"keywords":["Go","ugam","machine learning","TPU","AI","ugam solutions","ML","image recognition","head of analytics","RAG","analytics","Rust","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","image recognition","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-merkle-implements-ethical-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10059093,"title":"Why did Reliance pick up majority stake in a robotics startup","content":"Reliance Industries Limited has picked up 54% stake in Noida-based findustrial robotic automation company Addverb for $132 million. Our partnership with Reliance after an investment of $132 million in our Series B round will allow us to deliver advanced and affordable robots. It is an opportunity to deploy our robots in omni-channel distribution centers across sectors like e-commerce, retail, grocery, fashion, pharma, digital and petrochemical, Addverb said in a filing. The robotics startup has built automated warehouses for companies like Coca-Cola, Flipkart, HUL, Pepsi, Marico, Asian Paints, and ITC. In aninteraction with Analytics India Magazine (AIM), Satish Shukla, co-founder of Addverb said, Reliance is one of our esteemed customers and we had delivered multiple automated warehouses for them for their JioMart business. So there was already a certain level of trust and understanding between us and them. It is imperative for a business like ours to have an anchor customer who is willing to deploy automation. Reliance has significant presence in Oil and Petrochemicals, Organized Retail, Electronics, Grocery, E-Commerce, Fashion & Lifestyle and has also invested in pharmacy and hyperlocal delivery. In order to scale these businesses Reliance would need a dynamic and agile supply chain and this would require flexible and automated warehouses. Also, Reliance is betting big on 5G and green energy and one of the biggest benefactors of both these technologies is going to be Robotics. These are some of the synergies that would help Addverb grow along with Reliance, he added. Also, as a strategic investor, Addverb will also have significant business orders from Reliance for their warehouses across different industries. The advancement in battery technology and 5G will help to make robots more flexible and robust and will allow it for more innovative deployments. Reliance will be able to create innovative solutions for expansion of their supply chains. Omni-channel fulfillment and Micro-fulfillment are the new trends in supply chain and Addverb has the capability to deliver automation for both the applications and different types of business formats Reliance has. Different type of automated warehouses Goods-to-Person (GTP): Goods-to-person system is used for increasing efficiency and reducing overcrowding. It includes conveyors, carousels and vertical lift systems. If deployed properly, GTP systems can easily speed up warehouse picking. Retrieval Systems and Automated Storage (RS\/AS): RS\/AS comes under GTP technology wherein automated systems such as material-carrying vehicles, tote shuttles, and mini-loaders are used to store and recover materials. This system is used by warehouses with spacing issues and high-volume output. Automatic Guided Vehicles (AGVs): This uses minimal onboard computing power. In this system, the vehicles use magnetic strips, wires or sensors which help them to move across a fixed path in the warehouse. AGVs are ideal for large and simple warehouse environments with a navigation layout. Autonomous Mobile Robots (AMRs): Comparatively, AMRs are more flexible than AGVs since they use GPS systems to create a more effective route in a warehouse. It also uses laser guidance systems to spot obstacles, which allows it to safely navigate through human traffic. It is easier to program and can be deployed quickly. Pick-to-Light and Put-to-Light Systems: It uses mobile barcode scanning devices synced to digital light displays. It guides warehouse pickers to place or pick up selected items. This dramatically reduces walking and searching time, along with human error in high-volume scenarios. Voice Picking and Tasking: A voice directed automation also known as pick-by-voice uses speech recognition software and mobile headsets. The system deploys optimised pick paths to help workers to pick or place a product. It eliminates the need for handheld devices like RF scanners and improves pickers’ safety and efficiency. Automated Sortation Systems: Sortation is used to identify items in a conveyor system and divert them to a location using radio-frequency identification (RFID), barcode scanners, and sensors. This automated sortation system is used by companies for receiving, picking, packing, and shipping. Reliance’s rivals Flipkart introduced robots called automatic guided vehicles (AGVs) at its shipment sorting center in Bengaluru. AGVs put packages in designated slots with minimum human intervention. Currently, Flipkart has adopted 110 AGVs or “cobots”, which sorts 4,500 packages every hour, when compared to 500 parcels per hour done manually earlier. The adoption of cobots is one example of how Flipkart has consistently built a tech infrastructure for its entire supply chain network. In 2012, Amazon acquired Kiva Systems, a robotics company, for $775 million, and since 2014 it has installed more than 100,000 robots in 25 of its total 149 warehouses around the world. Amazon’s largest warehouse is one million square feet. Reliance’s automation push In simple terms, an automated warehouse management system allows the business to use AI and robotics in daily processes. As the name suggests, automated warehouses eliminate manual tasks that slow down the movement of goods. The issue arises when a product has to make more stops when it moves through a warehouse. This increases the chances of an error or problem to occur. Currently, Reliance uses products like Rapido (Pick-Put-To-Light), Dynamo (Autonomous Mobile Robot), Zippy (Carton Shuttle Robot), Quadron (Carton Shuttle Robot) setup by Addverb. Going forward, the robotics company will provide innovative automation solution for Reliance’s multiple formats using a mix of fixed automation like Multi-Pro (Automated Storage & Retrieval System), Cruiser (Pallet Shuttle Robot), Quadron (Carton Shuttle Robot) and Flexible Automation like Dynamo (Autonomous Mobile Robot), Zippy (Sorting Robot) and Veloce (Multi-Carton Picking Mobile Robot. Reliance will also use Addverb’s enterprise software, Optimus (Warehouse Management System) and Mobinity (Warehouse Control System) across its warehouses. The COVID-19 pandemic has caused massive disruptions across global supply lines and also led to labour shortage. Since the demand for more output has grown, automation has become a necessity for warehousing industries worldwide. The company’s retail revenue grew by 16% led by increased growth across consumption baskets amid strong consumer sentiments during festivities, relaxations in COVID-19 restrictions, and increased number of vaccinations. The business re-established growth in fashion and lifestyle with sales closing above pre-COVID levels. The consumer electronics and grocery also saw rapid growth as lockdowns eased all over India, Reliance Retail said in its third-quarter statement. In the next five years, Addverb aims to become a billion dollar revenue company with equal revenues from India and from global markets. The firm is working towards expanding its presence in the US and Europe and also scale its team with more focus on the R&D division.","excerpt":"The robotics startup has built automated warehouses for companies like Coca-Cola, Flipkart, HUL, Pepsi, Marico, Asian Paints, and ITC.","categories":["AI Startups"],"tags":["AI Startups"],"author_name":"Akashdeep Arul","publish_date":"2022-01-25T16:00:00","publication_year":"2022","word_count":1060,"keywords":["Go","API","TPU","AI","Git","RAG","Aim","analytics","Rust","R","AI Startups"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","TPU","R","Go","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/why-did-reliance-pick-up-majority-stake-in-a-robotics-startup\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10057775,"title":"Former co-leader of Google’s Ethical AI team, Timnit Gebru, thinks that AI needs to slow down","content":"On May 18 2021, Google CEO Sundar Pichai announced the rollout of LaMDA, a large language model (LLM) system that can chat with its users on any subject. This is an example of how language technologies are becoming enmeshed in the linguistic infrastructure of the internet, despite the unresolved ethical debates surrounding these cutting-edge systems. Language technologies are getting out of hand In December 2020, Timnit Gebru was fired from Google for refusing to retract a groundbreaking paper in which she argued that these models are prone to producing and propagating racist, sexist, and abusive ideas. Despite being the world’s most powerful autocomplete technologies, they don’t understand what they’re reading or saying, and many of the advanced capabilities they have are still only available in English. LLMs are prone to relegating certain professions to men and others to women; associate negative words with black people and positive words with white people; and if probed in a certain way, can encourage people to self-harm, condone genocide, or normalize child abuse. The danger of these systems lies in the fact that they are conversationally fluent, and it is very easy to believe that their outputs were written by other human beings. This gives them the dangerous potential of producing and promoting misinformation at a massive scale. Censorship of research Very little research is being done to understand how the flaws of LLMs could affect people in the real world or what efforts should be taken to mitigate these difficulties. Google’s firing of Gebru and her co-lead, Margaret Mitchell, underscored that the few companies that are rich enough to train and maintain LLMs have a heavy financial interest that will deter them from carefully examining its ethical implications. Since 2020, Google’s internal review process requires a separate level of review for “sensitive topics”. Therefore, if researchers are writing about topics such as facial recognition or the categorisation of gender, race, or politics, they have to consult Google’s PR and legal team to first look over their work and suggest changes before they can publish it. While many researchers turn to academia as an alternative to this, even that avenue can be riddled with concerns related to gatekeeping, harassment, and an incentive structure that doesn’t support long-term research. There are also concerns about tech companies funding AI research at academic institutions, with some researchers comparing it to how big tobacco companies used to fund research in an effort to dispel concerns about the health effects of smoking. AI needs to slow down In a recent interview with WIRED magazine, the central point that Timnit Gebru made was that AI needs to slow down. Gebru has witnessed the negative consequences of the hurried development of LLMs in her own life. She was born and raised in Ethiopia, where 86 languages are spoken—and nearly none of them are accounted for by mainstream language technologies. Despite these linguistic inadequacies, Facebook relies heavily on LLMs to moderate content globally. When the war broke out in the Tigray region of Ethiopia, the platform struggled to get a handle on the outbreak of misinformation. In an interview with the Wharton Business Daily, Gebru said that she is most concerned about the ‘move fast and break things’ attitude that dominates tech today. She argues that when you have the software and data available for people to download and collect data very easily and efficiently, it can be easy to forget to consider things that you should be taking into account. According to her, incentive structures have to slow down so that people can be educated on what sort of things they should be thinking about when collecting data. In the same podcast, Gebru claims that she has previously witnessed how good research can combat the lack of awareness regarding the speed at which technology and innovation is outpacing regulation and policy.” The 2018 paper she wrote in conjunction with Joy Buolamwini—that shed light on the disparities in commercial gender classification—played a large role in effecting a rapid and remarkable change in industry and policy. Clearly, research, and the awareness it brings, does have the potential to influence  the direction that AI technology takes.","excerpt":"On May 18 2021, Google CEO Sundar Pichai announced the rollout of LaMDA, a large language model (LLM) system that can chat with its users on any subject. This is an example of how language technologies are becoming enmeshed in the linguistic infrastructure of the internet, despite the unresolved ethical debates surrounding these cutting-edge systems.  […]","categories":["Global Tech"],"tags":["LaMDA","misinformation"],"author_name":"Srishti Mukherjee","publish_date":"2022-01-07T10:00:00","publication_year":"2022","word_count":692,"keywords":["misinformation","Go","API","TPU","AWS","AI","innovation","RAG","Ray","Aim","LaMDA","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","AWS","TPU","R","Go","API","innovation"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/former-co-leader-of-googles-ethical-ai-team-timnit-gebru-thinks-ai-needs-to-slow-down\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":6622,"title":"MISB Bocconi and Jigsaw Academy Launch Executive Program in Business Analytics (EPBA)","content":"It’s been a great year for Big Data. In spite of speculation about whether the concept was just hype, the fact is that world over we have seen data accumulating at exponentially increasing rates. Figures show that businesses rather than just talking about Big Data, have actually begun investing in Big Data initiatives. What was an impressive $3.2 billion dollar industry in 2010, is slated to grow to as much as $50.1 billion in 2015 (Wikkbon). Without a doubt we know that Big Data analytics is already helping business yield better profits. But it’s not just business that Big Data is revolutionizing. We see pockets of our lives also being impacted in subtle ways. Whether it’s health care, crime, education, better customer service, smart wearables, or smart gadgets that make our daily lives easier, data is transforming the way we live. Today it’s subtle, tomorrow Big Data has the potential to change the way we live, in ways we cannot even begin to imagine. What this all translates to is this – the Big Data Talent Market is absolutely booming. Experts predict that Big Data will create more than 4.4 million jobs by 2015. But unfortunately there are just not enough people to fill these roles. As more businesses join the Big Data race, this demand will only grow. Executive Program in Business Analytics The newly launched MISB Bocconi and Jigsaw Academy Executive Program in Business Analytics (EPBA)- Big Data Analytics for Professionals, is an initiative to bridge this talent gap. We spoke to Gaurav Vohra, CEO of Jigsaw Academy and asked him what prompted them to launch this course. “At Jigsaw Academy we see ourselves as playing a crucial role in identifying and nurturing analytic and Big Data talent worldwide, so that the skill gap reduces. Though many long term business analytics courses have been launched recently, they focus more on traditional techniques and business analytical tools. There was this crucial unserved need for an executive program that teaches conventional analytics tools and Big Data techniques. This is what prompted us to partner with MISB Bocconi to develop the EPBA program.” Gaurav said. The EPBA program is a ten month executive program which will begin in March 2015. Its curriculum aims to blend general management with Big Data and Business Analytics and focuses on giving participants an understanding of predictive modeling, data mining, Big Data analytics, marketing, operations and risk analytics, among other analytics areas. On completion of the program, participants will be capable of data driven decision making and leadership in industries such as retail, finance, telecommunications, healthcare, and manufacturing. The program involves 120 hours of in-person training to be held over six(6), three day modules at the MISB Bocconi campus in Powai, Mumbai. In the interim, Jigsaw Academy will also conduct twenty (20) live online classes of three (3) hours each for a total of sixty (60) hours which participants can attend from their home, office, or any other convenient location using an Internet connection. In addition to the live online and in-person classes, participants will also have access to over 100 hours of pre-recorded video lectures on data science and big data analytics for a period of twelve (12) months. The Strengths If you ae thinking about picking up some strategic Big Data skills, here is why you can consider the EPBA program: World class faculty: This includes renowned international faculty from SDA Bocconi in Milan, Italy, as well as analytics and big data experts from Jigsaw Academy. SDA Bocconi, the management school of Università Bocconi, is one of the select business schools world-wide to have received, for its MBA program, all three of the prestigious international accreditations—AACSB (Association to Advance Collegiate Schools of Business), EQUIS (European Quality Improvement System), and AMBA (Association of MBAs). Flexible learning environment: A combination of online and offline classes allows for a flexible learning environment, with participants getting access to not only live online classes but also a digital library with over 100 hours of pre-recorded video lectures that will supplement the campus learning. Opportunity to work on real world case studies: Jigsaw Academy has been collaborating closely with the industry for over 5 years now and have access to proprietary data sets from various companies. These case studies are based on real world business problems and use real business data sets from different industries such as FMCG, retail, Banking, e-commerce, healthcare and many others. Jigsaw Lab: Participants get unrestricted access to the Jigsaw Lab, a cloud-based analytics tool and content library that helps them gain hands- on competence with the most in-demand analytics tools and technologies in the industry including SAS, R, and Hadoop. Eligibility and Admissions All those with a Bachelor’s degree and who are proficient in oral and written English, with at least two years of full-time post qualification work experience* will be eligible for the program.*For applicants with exceptional qualification and\/or industry experience, an exception to the minimum eligibility criteria may be considered. Applications are being accepted until February 25, 2015. Admissions will be conducted on a rolling basis. The faculty panel will review all applications and shortlist candidates based on their profiles. The shortlisted candidates will then be interviewed either in person or over video \/ telephone. The program is priced at ₹3,60,000 + service tax which includes tuition fee, learning materials, and twelve (12) month access to the Jigsaw Lab (this does not include Board & lodging for the in person modules held at MISB Bocconi in Mumbai). For more details about the course please log on to http:\/\/jigsawacademy.com\/bocconi\/","excerpt":"It’s been a great year for Big Data. In spite of speculation about whether the concept was just hype, the fact is that world over we have seen data accumulating at exponentially increasing rates. Figures show that businesses rather than just talking about Big Data, have actually begun investing in Big Data initiatives. What was […]","categories":["AI Trends"],"tags":["jigsaw academy"],"author_name":"Дарья","publish_date":"2014-12-23T16:04:29","publication_year":"2014","word_count":925,"keywords":["big data","data science","programming_languages:R","AI","Git","Aim","analytics","jigsaw academy","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Git","big data","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/misb-bocconi-jigsaw-academy-launch-executive-program-business-analytics-epba\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":59717,"title":"NVIDIA, Azure And AWS Offer Free Resources To Fight Against COVID-19","content":"As COVID-19 continues to spread, we are acutely aware of the impact this is having on families, businesses, and communities. This is a global health emergency that will only be resolved by governments, businesses, academia, and individuals working together to better understand this virus and ultimately find a cure. The AI community especially has been at the forefront to fight against COVID-19 by offering solutions with the help of state-of-the-art-algorithms. However, for these machine learning models to work, they need tremendous amounts of computational power. In order to mitigate the operational costs of finding the solution, the hardware and cloud giants have stepped up to offer free services. Alibaba Cloud Offers AI Cloud Services After bringing down the prediction time for diagnosis using AI, Alibaba now wants the whole world to come up with their own unique solutions to fight COVID-19. They are now offering Elastic High Performance Computing (HPC) technology to worldwide researchers to accelerate drug and vaccine discovery. They have explained in their blog, how their technology has supported around 20 research groups implementing their research on COVID-19 with the help of Elastic High Performance Computing (E-HPC) technology that helps researchers find solutions for AI-driven-Drug-Design (AIDDD) for COVID-19. NVIDIA Parabricks PC Gamers, let’s put those GPUs to work. Join us and our friends at @OfficialPCMR in supporting folding@home and donating unused GPU computing power to fight against COVID-19! Learn more → https:\/\/t.co\/EQE4u7xTZT pic.twitter.com\/uO0ZCq8PEv— NVIDIA GeForce (@NVIDIAGeForce) March 13, 2020 NVIDIA has announced that it will be giving access to its Parabricks for 90 days to promote research on COVID-19 drug discovery. Analysing genomic data is computationally intensive. Therefore, time and cost are significant barriers to using genomics data for precision medicine. But, NVIDIA Parabricks Genomics Analysis Toolkit breaks down those barriers, providing GPU-accelerated genomic analysis. By accelerating the existing CPU-only pipelines on GPUs, NVIDIA Parabricks Germline Pipeline provides more than 40 times faster analysis for an individual sample. By processing the FASTQ input files, the system generates sorted, marked BAM\/CRAM files and variant call files (VCF or gVCF). NVIDIA Parabricks’ pipelines have been tested on Dell, HPE, IBM, and NVIDIA servers at Amazon Web Services, Google Cloud, and Microsoft Azure. NVIDIA advises users to start with a small server for a small number of samples and then use full-scale GPU servers to meet their needs. AWS’ Diagnostic Development Initiative Amazon Web Services (AWS), one of the most preferred cloud partners across the world, has announced that it is committing $20 million to accelerate research with regards to COVID-19 diagnostics. This AWS Diagnostic Development Initiative will support customers who are working to bring better diagnostics solutions to market faster. “Funding will be provided through a combination of AWS in-kind credits and technical support to assist the customers’ research teams in harnessing the full potential of the cloud to tackle this challenge,” stated AWS in their blog. The program will be open to accredited research institutions and private entities that are using AWS to support research-oriented workloads for the development of point-of-care diagnostics (testing that can be done at home or at a clinic with same-day results). Rescale Partners with Google Cloud and Microsoft Azure Rescale Inc in cooperation with Google Cloud and Microsoft Azure, announced a new program on Monday that immediately offers high-performance computing resources (HPC) at no cost to teams working to develop test kits and vaccines for COVID-19. Researchers can rapidly run simulations on the cloud without setup time or IT teams using Rescale’s platform combined with cloud computing resources from Google Cloud Platform and Microsoft Azure. “Rescale’s platform can provide access to high-performance computing resources that can help accelerate key processes and enable stronger collaboration,” said Manvinder Singh, Director, Partnerships at Google Cloud. Now You Can Donate Your GPUs As Well Folding@home, a distributed computing project that encourages scientists to volunteer and run simulations of protein dynamics on their personal computers. Last month, Folding@home announced that it is joining researchers from around the world working to accelerate the open science effort, where researchers are working to advance the understanding of the structures of potential drug targets for 2019-nCoV that could aid in the design of new therapies. With all the top companies offering their best of services for free, we can safely assume that the research for COVID-19 diagnostics will accelerate.","excerpt":"As COVID-19 continues to spread, we are acutely aware of the impact this is having on families, businesses, and communities. This is a global health emergency that will only be resolved by governments, businesses, academia, and individuals working together to better understand this virus and ultimately find a cure.  The AI community especially has been […]","categories":["Deep Tech"],"tags":["Amazon AWS","AWS","AWS services","Azure Machine Learning","covid-19","Google Cloud","Microsoft Azure","NVIDIA GPU"],"author_name":"Ram Sagar","publish_date":"2020-03-24T15:00:04","publication_year":"2020","word_count":716,"keywords":["NVIDIA GPU","AWS services","Google Cloud","covid-19","AWS","AI","machine learning","ML","cloud computing","distributed computing","R","RAG","Amazon AWS","Azure Machine Learning","Microsoft Azure","Azure","GPU computing"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","cloud computing","AWS","Azure","distributed computing","GPU computing","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/nvidia-aws-google-cloud-free-gpu-covid-19\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10126244,"title":"Accenture Acquires Bengaluru-based Chip Design Company Excelmax Technologies","content":"Accenture has announced the acquisition of Bengaluru-based chip designing company Excelmax Technologies. Excelmax provides custom silicon solutions used in consumer devices, data centres, artificial intelligence (AI) and computational platforms that enable edge AI deployments, to clients in the automotive, telecommunications and high-tech industries. The semiconductor market is experiencing a surge in demand for silicon design engineering, driven by the proliferation of data centres and the increasing use of AI and edge computing. This is further propelled by the growing consumer appetite for electronics, which is driving new investments in the chip design space. The acquisition enhances Accenture’s growing silicon design and engineering capabilities. Terms of the transaction were not disclosed. “With the rapid evolution of new technologies like generative AI and the growth of connected products, more intricate, specialised chips with enhanced performance and efficiency are required,” said Karthik Narain, group chief executive—Technology at Accenture. Founded in 2019, Excelmax brings comprehensive semiconductor solutions from high level design to detailed physical layout ready for manufacturing, and full turnkey execution. The company adds approximately 450 professionals to Accenture in key areas such as emulation, automotive, physical design, analog, logic design and verification, expanding Accenture’s ability to help global clients accelerate edge computing innovation. This acquisition follows the addition of XtremeEDA, an Ottawa, Canada-based silicon design services company, in 2022.","excerpt":"Excelmax provides custom silicon solutions used in consumer devices, data centres, AI and computational platforms.","categories":["AI News"],"tags":["Accenture","Mergers and Acquisitions"],"author_name":"Pritam Bordoloi","publish_date":"2024-07-09T09:34:22","publication_year":"2024","word_count":217,"keywords":["Accenture","API","artificial intelligence","programming_languages:R","AI","innovation","emerging_tech:edge AI","generative AI","edge AI","edge computing","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","artificial intelligence","generative AI","edge AI","edge computing","R","API","innovation","programming_languages:R","emerging_tech:edge AI"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/accenture-acquires-bengaluru-based-chip-design-company-excelmax-technologies\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21946,"title":"Why You Should First Learn Linear Algebra Before You Ace Machine Learning","content":"Just like a solid foundation is essential to a building, linear algebra forms an essential learning segment for machine learning (ML). Areas of mathematics such as statistics and calculus require prior knowledge of linear algebra, which will help you understand ML in depth. Many ML experts may be of the opinion that linear algebra (LA) helps to some extent, but it definitely improves one’s math skills and intuition in ML. This article presents the top five reasons to help you get acquainted with the preliminaries of LA. What is Linear Algebra? Linear Algebra is a branch of mathematics that deals with linear equations and linear functions which are represented through matrices and vectors. In simpler words, LA helps you understand geometric terms such as planes, in higher dimensions, and perform mathematical operations on them. By definition, algebra deals primarily with scalars (one-dimensional entities), but LA has vectors and matrices (entities which possess two or more dimensional components) to deal with linear equations and functions. LA can also be called as the extended version of algebra. 1.  LA Is The Elementary Unit For ML Calculus precedes LA when it comes to learning advanced math. Differential Calculus and Integral Calculus help you not just with limits, differentiation and integration techniques, but also set a base to apply them to vectors and multi-dimensional configurations such as tensors. This is called Matrix Calculus. Knowing this will help your understanding in areas such as linear functions and systems of linear equations. This is in addition to advanced topics such as Vectors in space and the Simplex method. In fact, the latter will also help you with linear programming. All of these concepts will be a cakewalk if you dedicate your time to learning LA and ML. 2. The ML Intuition LA will boost your intuition towards ML by offering more perspectives. The vectors and matrices you come across in LA will make your thinking more broad and idealistic. You may be motivated to utilise more parameters to a ML component, use more matrix operations, visualise and come up with different plotting graphs, or even apply a unique and better code. The possibilities are numerous. For example, consider a ML project in healthcare. The use cases here can be preventive care, diagnostics, insurance and patient health history, among others. Even though the datasets are available, the approach to building a ML model depends on perspectives such as data exploration, training and test data, regression and graphical depictions, among other features. 3. LA Helps Build Better ML Algorithms From Scratch LA will certainly assist in developing popular ML algorithms primarily categorised as Supervised Learning and Unsupervised Learning from scratch. Decision trees, linear regression, logistic regression, support vector machines and ensemble methods fall under supervised learning algorithms. On the other hand, clustering, component analysis and single value decomposition (SVD) fall under unsupervised learning algorithms. LA will facilitate a deeper understanding for the ML project which provides the flexibility to customise any parameters involved. This is really helpful as it will lead to optimal usage of resources. 4. LA Will Improve One’s Take On Statistics For ML, statistics forms a vital part to organise and assimilate data. LA acts as a prerequisite to have a solid understanding of statistical concepts. Notations, methods and operations in LA will assist in assimilating advanced topics in statistics such as multivariate analysis. For example, let us consider a doctor who has patient data such as blood pressure, heart rate, height, weight, among other data. These will form the multiple variables in the data set. Let us assume that more weight leads to higher blood pressure. This forms a linear relation — the increase in one variable leads to the increase in other. Suppose, if you want to perform a multivariate regression analysis in a statistical software such as  Stata, you will use manova and mvreg commands. The manova command ensures that the equations are statistically compatible; while the mvreg command obtains parameters such as standard errors among other estimation parameters. Therefore, the statistical result obtained is interpreted as a linear function and transformed into a matrix later for further work in ML. 5. LA For Processing Graphics In ML A ML project usually deals with objects such as audio, video and images along with other graphical interpretations such as edge detection. ML algorithms use classifiers to select some part of the dataset to train them according to a particular category. The classifiers also try to obviate errors from trained data. This is where LA comes into picture. It serves as an engine to compute these large, complex data. It incorporates a specific matrix decomposition technique for the project to handle and process the data. For example, two popular methods known as L-U decomposition and Q-R decomposition are used for the same. L-U method is used to split a square matrix into two matrices called as upper-triangle matrix and lower-triangle matrix. These sub-matrices are obtained by importing NumPy library for Python (depends on the programming language used), and loading the code in the compiler. On the other hand, Q-R decomposition is used for matrices which are of the order n x m (non-square matrix). This process will simplify the graphical need and uses optimal resource to generate the graphical model. Conclusion : For a ML beginner, LA might be a bit overwhelming to learn since ML itself has an array of concepts to master. Concepts such as linear functions and system of linear equations may look puzzling to some. But with regular practice and effort, LA will no longer be a daunting task. The benefits of learning LA are multi-fold. It improves math skills, programming skills and prepares the learner to think and explore the broader side of ML.","excerpt":"Just like a solid foundation is essential to a building, linear algebra forms an essential learning segment for machine learning (ML). Areas of mathematics such as statistics and calculus require prior knowledge of linear algebra, which will help you understand ML in depth. Many ML experts may be of the opinion that linear algebra (LA) […]","categories":["AI Features"],"tags":["linear algebra","Machine Learning","Machine Learning Algorithms","mathematics"],"author_name":"Abhishek Sharma","publish_date":"2018-02-21T11:33:13","publication_year":"2018","word_count":955,"keywords":["mathematics","Machine Learning Algorithms","NumPy","Go","machine learning","AI","ML","Machine Learning","Scala","Python","Ray","GAN","linear algebra","R"],"extracted_tech_keywords":["AI","machine learning","ML","Ray","NumPy","Python","R","Go","Scala","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/first-learn-linear-algebra-ace-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10007321,"title":"How Analytics Is Being Used In Data Journalism","content":"The field of journalism over the past decade or so has been witnessing continuous change. Today, journalism is influenced by big data and new computational tools. Data and visualisation have become the latest techniques for telling stories in media, thanks to intersections between journalism and computation. One of the many things that AI is doing for journalism is to make it easier and faster to analyse the data and also synthesise the data into stories. When we mention automatic story writing tools, they use Natural Language Understanding and Processing, to synthesise the stories. We also see the use of AI to help generate imagery and videos. Major news publications are struggling with budgets to maintain strong reporting staff. In such times, media houses have been exploring data and related computational tools to keep the expense of public accountability journalism economical, while presenting fact-based news reporting. Computational Journalism Leveraging Algos Computational journalism involve processes that utilise the assortment of analytical tools for storytelling. Data and computation are changing how journalists discover, write and distribute stories in the realm of public affairs reporting. Media houses are showing a deep interest in datasets which they can use to find information that is not revealed yet, particularly at the local level. Many times journalists have to take unstructured data, turn it into structured information, and tell their readers about a pattern that catches their attention. Such as in political reporting where reporters are trying to make use of big data to analyse political events. An example is Paradise Papers, the journalistic investigation from The International Consortium of International Journalists, where they utilised software developed in the digital humanities work at Stanford. There is even a website called Open Secrets that publishes content based on campaign finance data. They have something called Anomaly Tracker to track money in politics and its effect on elections and public policy. Case Study International Consortium of Investigative Journalists, a network of global journalists that does cross-border investigations and issues of global concern precisely does investigations using large datasets. One of its works called the Panama Papers, was the most prominent data journalism story ever which had a significant global impact and resulted in resignations and legal trials of politicians on charges of corruption and tax evasion. ICIJ journalists from Süddeutsche Zeitung, a newspaper in Germany first got access to a large dataset (2.6TB) containing over 11 million documents from an anonymous source. ICIJ quickly made a team of more than 370 journalists from 76 countries and a group of data-savvy developers to work together for a year-long project in secret and start analysing the files. Finally, the team exposed those powerful billionaires, celebrities and politicians who were involved in money laundering and tax evasion using offshore companies and a Panama-based legal firm called Mossack Fonseca. There were hundreds of stories published by our media organisations. There were a lot of processes involved to process all the 11.5 million files, mostly leaked emails, PDFs and images of scans, records of incorporation. ICIJ turned to open-source tools that allowed it to work on the files. According to ICIJ, it built a secure cloud network that consisted of between 30-40 G2 AWS instances at a time to do parallel processing of all these documents. They used open-source technology such as Apache Tika, a Java-based toolkit that detects and extracts metadata and text from over a thousand different file types, and Tesseract- which is an OCR engine. The team had a small internal project Extract hosted on its GitHub. Extract is a cross-platform command-line tool for parallelised, distributed content-extraction built on Top of Apache Tika. It supports Redis-backed queueing for distributed, parallel extraction and will write to Solr, plain text files or standard output. The team also created a search engine for journalists to query searches for investigating certain people and finding specific data. It used Neo4J for creating a knowledge graph which was then used for visualisation that demonstrated patterns between specific politicians and the offshore companies. Open Data Is Democratising Storytelling We’ve seen many examples of the use of computer-assisted reporting to gather information and analyse it to create stories. Open data is democratising storytelling, allowing people to tell engaging stories. We’re seeing a lot of efforts today at making data more publicly available. There are websites where you can acquire government data, census data, and all types of demographic data. Certainly, some tools make it much more open to a broader spectrum of the public and journalists who want to dig deep and learn what’s going on, to study interesting patterns and create stories. One example is the Stanford Open Policing project, where the journalism department had its students file a freedom of information act requests on all. 50 states were asked for electronic versions of State Police stop data resulting in about 130 million records from 31 states in two years. The data was then used to find insights (using certain algorithms) on what rule of thumb a police officer uses when someone is pulled over. Stanford opened up the data for media houses and local reporters to download the information. It helped in understanding how their state police are operating, leading to stories which highlighted US police’s actions across racial demographics in the US. The transparency of data is reflected by media firms who are opening up their datasets to create stories. Here is a look at the Github of the Economist, including one repo containing code for a dynamic multilevel Bayesian model to predict US presidential elections. Written in R and Stan, the model is updated every day and combines state and national polls with economic indicators to predict a range of outcomes. Visualisation Is An Important Aspect Of Journalism One of the domains that a lot of journalists are focusing on is data visualisation and attempting to take complex datasets and turn them into fascinating visualisations, which otherwise would look mundane. Visualisations provide a lot of context to the story while engaging readers. While creating compelling visuals for a given dataset is challenging for most people, there are many tools available now that make it easy and faster for journalists to create useful visualisations.","excerpt":"The field of journalism over the past decade or so has been witnessing continuous change. Today, journalism is influenced by big data and new computational tools. Data and visualisation have become the latest techniques for telling stories in media, thanks to intersections between journalism and computation. One of the many things that AI is doing […]","categories":["AI Features"],"tags":["back office data","data structure using java","different types of analytics","extract big data","Journalism AI","social network big data","types of analytics"],"author_name":"Vishal Chawla","publish_date":"2020-09-14T14:00:21","publication_year":"2020","word_count":1028,"keywords":["Go","Journalism AI","TPU","AWS","AI","Redis","data structure using java","Git","Java","RAG","extract big data","back office data","GitHub","R","social network big data","types of analytics","different types of analytics"],"extracted_tech_keywords":["AI","RAG","AWS","TPU","Redis","R","Go","Java","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-analytics-is-being-used-in-data-journalism\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10141754,"title":"AI vs Fraud: How Telecom Giants Are Outsmarting Scammers","content":"The telecommunications industry has always grappled with its biggest nemesis – fraud. From SIM box fraud and Wangiri scams (explained below) to device and subscription frauds, the sector loses billions annually to increasingly sophisticated schemes. However, telecom companies are putting up a good fight with AI for fraud detection and prevention. Virgin Media O2’s AI-generated grandmother, Daisy, is an AI system that engages with scammers through natural conversations. It keeps them occupied for up to 40 minutes while telling stories about grandchildren and displaying calculated tech illiteracy. The system has already conducted over 1,000 conversations with potential fraudsters, effectively turning the tables on scammers by using their own tactics against them. Fraud not only causes financial losses but also erodes customer trust. “Fraud has evolved—it’s now more sophisticated, and there’s more data to handle. Rule-based systems can no longer keep up. This is where AI comes in,” Harsha Angeri, VP & head of AI business at Subex, told AIM. Telecom fraud is a global issue, with annual losses estimated at $32.7 billion. Fraudsters are leveraging advanced technologies like VPNs, SIM boxes, and even AI itself to bypass traditional security measures. A Reddit user recently remarked, “…financial fraud and cybercrime are off the charts. I don’t know anybody in my office or family who hasn’t received a scam call in the last few months.” How AI is Helping Telecoms Detect Frauds One of the most prevalent types of telecom fraud is SIM box fraud, where international calls are rerouted as local ones to evade tariffs. Previously, it was easy to flag suspicious SIM cards by identifying patterns like only outgoing calls and no incoming calls. However, as Angeri points out, “Today, fraudsters use mobile SIM boxes that move around in vans to avoid detection. It’s no longer as simple as writing an ‘if-else’ rule.” AI models now analyse billions of call detail records (CDRs) daily to detect such sophisticated patterns. This capability is crucial in countries like India and Indonesia, where telecom operators process hundreds of millions of calls every day. Another major issue is device fraud, particularly in regions like the Middle East and the US. In these markets, customers often purchase high-end devices on instalment plans. Fraudsters exploit this system by paying one instalment for an expensive device like an iPhone 16 Pro before disappearing. Angeri explained how AI tackles this: “We assess risk by looking at features like customer tenure, payment history, location consistency, and even whether the customer recently switched SIM cards.” These insights enable telecom operators to flag high-risk transactions before they occur. Wangiri scams are another significant challenge. Here, victims receive missed calls from premium numbers and are charged exorbitant fees when they return the call. Differentiating between fraudulent missed calls and legitimate ones (such as flash calls used for OTP verification) requires advanced anomaly detection techniques. “AI helps us differentiate these cases by examining call metadata and usage patterns,” Angeri said. For example, Subex uses supervised learning models like Random Forests and XGBoost to analyse hundreds of features—such as call duration, frequency, and location—to flag suspicious activity. AI also addresses newer challenges like flash calling, where social media platforms use missed calls for user verification without paying traditional SMS charges. While not strictly fraudulent, it represents revenue leakage for operators. AI models help identify such behaviour effectively. The scale at which these models operate is staggering. In Indonesia alone, some operators process up to 150 billion call records daily. As Angeri emphasises: “You can’t do a hobby project here. The scale is enormous. AI models must handle billions of records in real-time.” Why Legacy Systems are Still Relevant Despite AI’s advancements, legacy rule-based systems continue to play a crucial role in telecom fraud management. These systems provide foundational insights that complement AI-driven solutions. Subex’s legacy system, ROC (Revenue Operations Center), is still widely used alongside its new-generation AI platform, HyperSense. “Legacy models provide seed data and initial insights that help train AI models effectively,” Angeri said. For simpler fraud scenarios—such as identifying SIM cards with unusual call patterns—rule-based systems remain cost-effective and reliable. Legacy systems also ensure continuity during the transition to AI-powered solutions. Many telecom operators prefer a hybrid approach that combines the stability of traditional systems with the adaptability of AI models. This approach allows operators to gradually scale their AI capabilities while leveraging existing infrastructure. Angeri highlighted another key advantage of legacy systems, which is their ability to handle privacy-sensitive data securely within telecom operators’ infrastructure. “Our systems are deployed within the telco’s environment,” he said. “Data like CDRs or billing records never leave their secure systems.” This ensures compliance with stringent privacy regulations while enabling robust fraud detection. However, legacy systems alone cannot tackle today’s sophisticated fraud schemes. “Fraud detection is an ongoing process requiring frequent model retraining as patterns evolve,” Angeri said. Integrating legacy systems with advanced AI techniques ensures telecom operators can effectively address both traditional and emerging threats. “Fraud will happen—that’s the reality. But as soon as patterns emerge, we must catch them so they don’t propagate,” he concluded. With $40 billion lost annually to telecom fraud globally, adopting such hybrid solutions isn’t just an option—it’s necessary to safeguard revenues and maintain customer trust in an increasingly digital world.","excerpt":"Telecom fraud is a global issue, with annual losses estimated at $32.7 billion.","categories":["AI Features"],"tags":["AI","fraud detection","telecom"],"author_name":"Sagar Sharma","publish_date":"2024-11-27T12:28:17","publication_year":"2024","word_count":872,"keywords":["Go","AI","Git","RAG","telecom","Aim","XGBoost","anomaly detection","Rust","R","fraud detection"],"extracted_tech_keywords":["AI","Aim","XGBoost","RAG","anomaly detection","fraud detection","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-vs-fraud-how-telecom-giants-are-outsmarting-scammers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10044802,"title":"Language App Duolingo Makes Impressive Nasdaq Debut","content":"Language learning app Duolingo’s closed up 36 percent on its market debut on July 28 on the Nasdaq. The stock closed at $139.01, on volume of more than 2.8 million shares. The stock trades under the ticker ‘DUOL’. The shares jumped to $145 after opening at $141. Duolingo raised about $521 million in the IPO by selling 5.1 million shares. Of this, 1.4 million shares were sold by the existing stockholders, the proceeds of which will not go to the company. Earlier this week, the company raised its price target range between $95 and $100 per share. The earlier estimate was between $85 and $95 per share. Duolingo is among the world’s most popular education apps and with maximum audience from the US. It began as a computer science project for co-founders Severin Hacker and Luis von Ahn aimed at teaching people foreign languages while trying to translate the entire internet. Interestingly, co-founder Von Ahn is also one of the developers behind CAPTCHA and ReCAPTCHA.The app has 40 million monthly active users and more than 500 million downloads. As per the company, in 2020, Duolingo derived 51 percent of revenue from Apple’s App Store and 19 percent from the Google Play Store, and was the top-grossing app in the education category for both.","excerpt":"The app has 40 million monthly active users and more than 500 million downloads.","categories":["AI News"],"tags":["Duolingo"],"author_name":"Shraddha Goled","publish_date":"2021-07-29T17:14:08","publication_year":"2021","word_count":213,"keywords":["Go","programming_languages:R","AI","IPO","programming_languages:Go","Aim","Duolingo","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","IPO","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/language-app-duolingo-makes-impressive-nasdaq-debut\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043502,"title":"Luck By Chance: Bayer Pharmaceutical’s Abhishek Choudhary Traces His Machine Learning Journey","content":"“I worked on a hackathon problem for 48 hours straight. I didn’t win it. But I learned about something extraordinary – machine learning.”Abhishek Choudhary For this week’s ML practitioner’s series, Analytics India Magazine(AIM) got in touch with Abhishek Choudhary, a lead data engineer at Bayer Pharmaceutical. In this interview, Abhishek shares his rich experience of deploying machine learning models in the real world. AIM: How did your machine learning journey begin? Abhishek: I finished my schooling from a very small town in Chhattisgarh, named Chirmiri. Then I did Computer Science Engineering from Raipur, Chhattisgarh. It was a pretty big jump for me. My data journey began around 2012 ~ 2013. I was working on a large data product. I was busy writing too many threads (in Java) to somehow process the data, and ended up with an overcomplicated project. During a casual meetup in Bangalore, I learned about Hadoop and that some companies were using it in production. I started learning Hadoop immediately and built a fast prototype to give a demo to my team. It worked. I believe my machine learning (ML) journey began purely by chance. I was participating in a Hackathon for 48 hours straight and was working on a risk prediction problem. I ended up having 100s of if-else conditions after 30 hours of coding. I was tired and not sure of my code, so I started looking for alternative algorithms and found out about Machine Learning. I had no idea about it  and had no time to understand it deeply. So I worked from bottom-up. I simply used the Linear regression algorithm and validated output, and was shocked by the performance.  I didn’t win the Hackathon but I learned about something extraordinary – machine learning. AIM: What were the initial challenges and how did you address them? Abhishek: Starting out, my initial hurdles came in the form of programming language. I was a Java Developer but to learn more about ML, I ended up learning R-programming, Matlab, and then Python. Also back then, the devices were not powerful enough. Not to forget the poor internet speed. Downloading large data to work on ML problems was really hard. I tried to build an Android App, based on the ML model and my mobile crashed as soon as I opened the App. I raised a bug in the Android-ML community but couldn’t get much help as the ML community too was in a nascent stage. There were not enough resources available around ML\/Big Data unlike today. So, I started learning the source-code and code comments to understand better. With time, the challenges moved from slow internet speeds to model deployment. For instance, I have worked on an ML-based Bidding Pipeline (Ad-tech). Here the goal is to bid for millions of keywords in real-time. The bidding values were influenced by many external and internal factors. The platform was reading a massive amount of weekly data along with real-time data. The platform must react to the real-time data and predict bids in low latency. Here are the challenges in real world ML deployment: Machine learning with real-time data is hard and complicated.Building a solid fall-back mechanism to compensate for incorrect predictions that can directly impact revenue.Automated deployment of ML models.Finding a lightweight metrics and monitoring system to observe performance and real-time response.A\/B testing framework.Building and maintaining complex infrastructure that can support different technologies with more than 99.9% uptime. AIM: Tell us about your role at Bayer Pharmaceutical. Abhishek: I am a Lead Data Engineer at Bayer Pharmaceutical. I am responsible for building an Analytics & Machine Learning platform for Real-World Data in healthcare. A typical day usually is coffee with programming and Data\/ML infrastructure development. I spend most of the time building applications around Data and Machine Learning. Interacting with the team to understand the requirements and address them ASAP. “Data Science workflow is quite a repetitive domain, so code reusability and clean code are essential.” AIM: What does your machine learning toolkit look like? Abhishek: I am still exploring new technologies. But here are a few from the top of my mind: For data processing, I use ScalaPython for Data Pipeline and Machine LearningKubernetes for infrastructure and deploymentAirflow\/Dagster for Pipeline SchedulingScipy, PySpark for MLApache Superset for Analytics & DashboardingTrinoDb for distributed SQL AIM: What kind of software engineering principles should a data scientist or a data engineer know? Abhishek: Unit Test cases are extremely important for deploying Data Science pipeline in Production. Data Science workflow is quite a repetitive domain, so code reusability and clean code are essential. Don’t complicate the code and if it’s too big, refactor it in smaller steps. Think around pipelines and tasks, and each task should be an isolated immutable state of the pipeline. Model deployment is more around infrastructure as it requires solid metrics and model performance validations frequently. The model deployment should be automated and there should be an auto validation in the infrastructure before model deployment in production. AIM: What does the future of ML look like from your vantage point? Abhishek: I think going forward Decision Tree as a technique will flourish. There will be more progress with regards to Optimized Dynamic Programming based algorithms. And, when it comes to domains, healthcare will be the hottest of them all. AIM: What’s your advice for the ML aspirants? Abhishek: Before jumping to Data Science\/ML, build a solid foundation around SQL, Database and Algorithms.  Data cleaning is way more complicated than one can imagine. Try all possible kinds of data and learn the techniques to clean or transform it in specific ways. Here are my top book recommendations: Designing Data-Intensive Applications by Martin KleppmannDomain-driven design by Eric EvansStorytelling with Data by Cole NussbaumAn Introduction to Statistical Learning by Trevor Hastie et al.,The Elements of Statistical Learning by Trevor Hastie et al.,Naked Statistics by Charles WheelanThe Algorithm Design Manual by Steven SkienaHarvard’s CS 109 Data Science – HarvardMachine Learning courses on Coursera","excerpt":"“I worked on a hackathon problem for 48 hours straight. I didn’t win it. But I learned about something extraordinary – machine learning.” Abhishek Choudhary For this week’s ML practitioner’s series, Analytics India Magazine(AIM) got in touch with Abhishek Choudhary, a lead data engineer at Bayer Pharmaceutical. In this interview, Abhishek shares his rich experience […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Ram Sagar","publish_date":"2021-07-14T14:00:00","publication_year":"2021","word_count":990,"keywords":["data science","machine learning","TPU","AI","ML","Python","Aim","analytics","R","kubernetes","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","kubernetes","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/abhishek-choudhary-data-engineer-interview-bayer\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10040340,"title":"Top Tech Conferences For Women In 2021","content":"Women are breaking the glass ceiling in every sector, tech included. Many organisations around the world are championing gender diversity at the workplace. Unlike a decade ago, leadership roles in AI and analytics are not off limits for women. The number of women-centric conferences and events have also gone up in lock step with this general trend. Here is the list of the top tech conferences for women in 2021. (The list is in chronological order) 1| The Rising Dates: May 21-22, 2021 Analytics India Magazine’s The Rising brings together professionals from the industry and academia, and data science leaders under one virtual roof to exchange cutting edge ideas and discuss the latest developments in data science. The women luminaries in data science will shed light on the inner workings of the field and share insights on how to build a successful career in the data science field. Know more here. 2|  Women in Cybersecurity (WiCyS) Dates: September 8-10, 2021 WiCyS conference is a flagship event for women in cybersecurity. It claims to be the largest cybersecurity conference with equal representation of professionals and students. The conference helps organisations recruit, retain and advance women in cybersecurity — all while creating a community of engagement, encouragement, and support at a technical conference. Know more here. 3| The Girls In Tech Conference Dates: September 15, 2021 The Girls In Tech Conference will be the perfect stage to build your skills and explore personal development opportunities. The one-day virtual event features motivational keynotes, insightful panel discussions and group conversations. The conference will give you an opportunity to meet kindred spirits, mentors, and role models to inspire your tech journey. Know more here. 4| Virtual Grace Hopper Celebration (vGHC) Dates: Sept 27- Oct 1, 2021 Grace Hopper Celebration (vGHC) is one of the largest gatherings of women technologists, where women from around the world learn, network, and celebrate their achievements. The event allows collaborative proposals, networking, and mentoring for the attendees. It also offers professional development through a variety of activities. Know more here. 5| Women In Tech Festival Dates: October 13-15, 2021 The Women in Tech Festival celebrates women in STEM, business as well as leaders who work to inspire, engage, and empower women. The three-day event invites women around the world to join together virtually as well as in-person to participate in inspiring talks, thought-provoking discussions, startup pitches, keynotes, educational workshops, career mentorship and networking. Know more here. 6| Women Impact Tech Dates: October 19, 2021 Women Impact Tech aims to achieve equality in tech. The event is building a community for women in technology to inspire, empower, and advance gender equity. The event will be held in San Francisco on 19th October 2021 and will host keynotes, panel discussions, breakout sessions, and more. Know more here. 7| Women in Tech Summit (WITS) Dates: October 20-22, 2021 The Women in Tech Summit (WITS) is a virtual version of the Women in Tech Summit series. From discussions about trending topics in tech and information on powering through career phases and opportunities to online networking, this event provides everything to help women thrive in the tech industry. Know more here. 8| Tech Up for Women Dates: November 16, 2021 Tech Up for Women is a one-day event focusing on new technological advances in tech. The event covers various areas, from cyber security, digital transformation, blockchain, big data, eCommerce, cloud security, AI, under the sea and space technology, VR, entrepreneurship resources, smart building technology, consumer technology, and more. Know more here.","excerpt":"Women are breaking the glass ceiling in every sector, tech included. Many organisations around the world are championing gender diversity at the workplace. Unlike a decade ago, leadership roles in AI and analytics are not off limits for women. The number of women-centric conferences and events have also gone up in lock step with this […]","categories":["AI Trends"],"tags":[],"author_name":"Ambika Choudhury","publish_date":"2021-05-18T18:00:00","publication_year":"2021","word_count":586,"keywords":["big data","data science","Go","AI","Git","RAG","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Git","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-tech-conferences-for-women-in-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":48174,"title":"Webinar: How To Begin A Career In Data Science","content":"The demand for data science professionals is at an all-time high. Around 22,000 freshers were added to analytics workforce in India this year, an increase of 37% from 2018 according to the data science industry study conducted by Analytics India Magazine and Praxis. While there is much excitement in the market, there is an equal amount of confusion as well regarding how one should really transition to the field. On one side the expectations of the companies are huge. When we look at the JDs, we see requirements of a ‘rockstar’ full-stack data scientist who knows everything, from statistics to machine learning, from model development to business development. On the other side, we see educational institutes and training institutes offering a wide spectrum of programs – from comprehensive long duration programs to specific short duration capsules in full time, hybrid and online mode. The candidates aspiring to step into the domain find it hard to demystify the jargon and identify a path suitable for them to upskill themselves and transition to a successful career in the filed. Register for the webinar here. AIM brings to you two stalwarts – one from the industry and another from academia to give the aspirants clarity about how to make their data science move. Register for the webinar here. Your key takeaways from the webinar: What are the skills needed to be successful in data science – soft skills and hard skills? How does one assess if one is suitable for a career in data science? How does one embark on the data science journey? Date : 24 October 2019 (Thursday) Time : 7 pm to 8 pm Please register here and invite your friends to register. Speakers for the webinar: Shivaram KR Co-Founder & CEO – Curl Analytics BE (ECE) RVCoE, MTech (Control & Automation) IIT Delhi Shivaram has over 12 years of experience in Machine Learning. Algorithm trading strategy (High-Frequency Trading) development has been his forte and he has honed these skills in his stints with Bank of America & Edelweiss Securities. Prior to Founding Curl, Shivaram led a niche quant team at Société Générale, where he worked on applying machine learning to forecasting financial markets, rainfall (alternative data analysis) and many other challenging problems. He has published several papers, on applying Machine Learning to markets. Charanpreet Singh Founder & Director – Praxis Business School Foundation B.Tech (Mech) IIT Kanpur, MBA University of Iowa Charanpreet has 30+ years of experience in organisations like British Oxygen, Tata Steel, PwC, HP and Praxis. Prior to founding Praxis, Charanpreet was heading the SMB Sales at HP-Compaq. Prof. Charanpreet Singh is the driving force behind the success of the Praxis Data Science program. He has professional and academic interests in the areas of education, learning, data science and communication and has been a mentor to career aspirants across domains. Please register here and invite your friends to register.","excerpt":"The demand for data science professionals is at an all-time high. Around 22,000 freshers were added to analytics workforce in India this year, an increase of 37% from 2018 according to the data science industry study conducted by Analytics India Magazine and Praxis. While there is much excitement in the market, there is an equal […]","categories":["Deep Tech"],"tags":["data science webinar","praxis"],"author_name":"Abhijeet Katte","publish_date":"2019-10-18T09:30:27","publication_year":"2019","word_count":482,"keywords":["data science","Go","machine learning","AI","automation","data science webinar","Aim","ViT","analytics","praxis","GAN","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","Aim","R","Go","GAN","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/webinar-how-to-begin-a-career-in-data-science\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":5742,"title":"Interview &#8211; Santosh Nair &#038; Vijay Ramaswamy, Founders at Analytic Edge","content":"[dropcap size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [dropcap size=”2″]AE[\/dropcap]Analytics Edge: The main tenets of our organization are: –  We help our clients solve their business challenges by providing technology enabled solutions powered by advanced analytics. We customize our products and solutions to deliver faster, actionable, results-oriented impact We take a collaborative approach with client stakeholders and provide flexibility to circle back and try different techniques or review different data streams to answer client questions We continuously measure the impact to ensure that the recommendations are relevant, effective and efficient in a changing business environment. AIM: Please brief us about some business solutions you work on and how you derive value out of it. AE: Our Management team has over 3 decades of experience in statistical modeling and advanced analytics across FMCG, Retail, Telecom and Travel & Hospitality domains across multiple geographies.  Our core capabilities are in the area of marketing and customer analytics.  Some of our key solutions include Marketing Spend Optimization, Shopper Marketing, Demand Forecasting and CRM analytics.  We have developed client deployable applications for some of these solutions to aid planning and decision-making.  In our experience, clients derive tremendous value when insights from complex statistical and econometric algorithms are supported by technology that enables activation of these insights. AIM: How does a typical requirement gathering to delivery cycle look like for you? AE: We start out with a detailed understanding of the business challenge. Once we have a clear understanding of client objectives, available data and output desired, we develop the solution framework with inputs, approach, sample outputs, timelines and cost. This process usually takes 2-4 weeks.  We then engage with clients on key milestones viz. data review, insights and recommendations presentation and application deployment. Our typical delivery cycle is 6-12 weeks depending on the nature and scope of the business challenge. AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth. AE: We are currently a team of 8 associates but growing rapidly.  We want to create a climate of entrepreneurship and so everyone gets exposure to client engagement and delivery. We encourage all associates to share their views on how to enhance project deliverables, the delivery process and roadmap for the future. We want our associates grow professionally by not only mastering the technical aspects of our solutions but also by acquiring business and domain understanding so that they look at challenges and insights we provide from a client perspective. We have been very fortunate to hire the right talent who are passionate and eager to contribute to the growth of our company. AIM: Would you like to share any example of an Insight that generated a huge positive impact for your clients? AE: We had delivered a marketing spend optimization solution to a US-based beer, wine and spirits manufacturer. They wanted to understand the efficiency of their marketing spends on a flagship brand. The objective behind the analysis was to decide if they should shift their marketing budget to Digital media from other traditional media i.e. TV, Print and Radio. Using our proprietary algorithm, we calculated the return on investment for all their media vehicles and across all their marketing campaigns.  We provided recommendations on how can they can optimize their media spend between digital and traditional media to maximize sales. Our recommendations resulted in an 8% increase in brand sales with a 6% decline in their marketing spend! AIM: What are the next steps\/ road ahead for analytics at your organizations? AE: Our strategy for the first 2 years was to drive growth through analytics consulting. In parallel, we identified some key areas where we wanted to develop analytics products. While we will continue to engage with clients to provide analytics consulting services, our vision is to develop easy to use and deploy analytics products. We are currently developing a suite of “easy to use” products that will not only facilitate activation of insights but will also will enable business users to do advanced analytics in-house with minimal reliance on external vendors. AIM: What are a few things that organizations should be doing with their analytics efforts that most don’t do today? AE: Many organizations have not yet started investing in analytics although there is an abundance of data that suggests that analytics can drive tremendous efficiencies and provide a competitive edge.  Several organizations do not invest in analytics because they feel they do not have perfect data.  But data will never be perfect.   It is essential that these organizations take the plunge and get at least some insights from the data that is available.  Analytics engagements can help identify the data gaps and these organizations can then take measures to collect necessary data to plug the gaps. On the other hand, there are organizations that are currently investing in analytics but are not activating the insights effectively because their workforce is not trained on how to deploy the insights. There is training gap that needs to be addressed. These organizations must also invest in tools and technology that will enable the activation of insights.  Monitoring and measuring the impact of analytics insights in another area that must become a priority for both client and vendor organizations. AIM: What are the most significant challenges you face being in the forefront of analytics space? AE: We have a lot of ideas but since we are a start-up, we have to be very judicious about where we invest our time and money.  Our two key priorities are product development and business development. AIM: How did you start your career in analytics? AE: I started at Symphony Marketing Solutions, which was later acquired by Genpact. Vijay started his analytics career at Novartis Pharmaceuticals. In 2012, we felt the time was right to venture out on our own and do something we were very passionate about and Analytic Edge was born. AIM: What do you suggest to new graduates aspiring to get into analytics space? AE: In my opinion, we have only seen the tip of the iceberg in terms of the potential that the analytics industry will offer in the years to come. Graduates keen on getting into the analytics space must be proficient in statistics and data mining techniques but, more importantly, it is the ability to translate algorithms into business insights and effectively communicate these insights to Clients that will set them apart. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? AE: The key skills we look for are analytics aptitude and the right attitude. We expect the entry level candidates to have a basic understanding of statistics, working knowledge of one or more analytical tools and the ability to visualize data and make meaningful inferences. AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? AE: We expect to see more and more mid-size and small-size companies invest in analytics with the increase in awareness of the value of data analytics.  With the advent of cloud computing, data collection and storage is no longer an expensive proposition for these companies.  The emerging trends in analytics will be in the area of automated and real-time analytics. AIM: Anything else you wish to add? AE: We have developed proprietary algorithms in the areas of attribution modeling and demand forecasting and plan to release the beta version of web based marketing analytics tool by June 2014. We are very passionate about combining each of these analytics solutions and technology to help organizations address their business challenges.   We can be reached at santoshnair@analytic-edge.com or vijayramaswamy@analytic-edge.com. [divider] [spoiler title=”Biography of Founders” style=”fancy” icon=”plus-circle”] Dr. Santosh Nair Santosh brings with him vast experience in advanced analytics and statistical modeling in a multi-disciplinary manner. In his tenure, he has developed various analytics solutions in his career providing deeper insights and saving costs for the end clients in the CPG, Retail, Pharmaceuticals, Telecom and Education verticals. At GENPACT, he worked as an engagement leader for clients based out of the US, helped with pre-sales efforts and delivery of advanced analytical solutions in the areas of customer analytics, product development, automated pricing, marketing mix and simulation models. Santosh has a Ph.D. in applied economics and statistics from Clemson University in the US. He can be reached at santoshnair@analytic-edge.com Vijay Ramaswamy Vijay has a strong background in clinical trials research, strategic pricing & promotion analysis, customer analytics and marketing mix modeling using SAS, R and automated analytics platforms. He is also a content developer and visiting faculty at Jigsaw academy. His past experience includes Novartis pharmaceuticals, GENPACT SMS and IBM Demand Tec. In his tenure, he has lead the off-shore analytics solution delivery for clients like PepsiCo and Coke International and has a record of recruiting, training and transitioning on-shore work in a quarter. Vijay has a post graduate degree in statistics from University of Mumbai and an executive MBA from XLRI, Jamshedpur. He can be reached at vijayramaswamy@analytic-edge.com [\/spoiler]","excerpt":"[dropcap size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? [dropcap size=”2″]AE[\/dropcap]Analytics Edge: The main tenets of our organization are: –  We help our clients solve their business challenges by providing technology enabled solutions powered by advanced analytics. We customize our products […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"AIM Media House","publish_date":"2014-05-20T13:06:16","publication_year":"2014","word_count":1541,"keywords":["Go","API","TPU","AI","cloud computing","Git","RAG","Aim","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","cloud computing","TPU","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-santosh-nair-vijay-ramaswamy-founders-at-analytics-edge\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":35058,"title":"Data Breaches Faced By Consumer Internet Companies In India: 2018","content":"Whether it was the Facebook-Cambridge Analytica scandal, or Quora row or even Google Plus security problem, 2018 was the year of data breaches. Data worth millions of dollars was stolen and was susceptible to being used illegally during this time. In an earlier article we discussed some of the big shots across the globe that fell a victim of the breach, but here, we are listing some major data breaches faced by consumer internet companies: RupeeRedee: The money lending digital platform based out of Gurugram detected vulnerabilities in its data stack. A subsidiary of Digital Finance International, the microblogging site serves millions of customers. It was found that the database stored in the Amazon Cloud are susceptible to the data leak, and was pointed out by a user on Twitter. He had said that the startup was leaking customer details because of a vulnerability on its cloud storage infrastructures. The company, however, stated that the vulnerability was recognised and fixed within a few hours, hence preventing any compromises in the customer data. Information Leaked: Scanned copies of customer’s Aadhaar cards and PAN cards submitted during KYC. Amazon India: In Nov 2018, e-commerce giant Amazon India faced data breach, not once but on two occasions. During the festive shopping at the US, the company contacted the customers about a technical error that disclosed customer names and email IDs on its website. The exact numbers were not clarified and Amazon India did not give any confirmation if Indian customers were also impacted. The following month, Amazon suffered another technical error that resulted in the exposure of its user data. This time the malware was detected in Amazon India platform. The incident exposed some of the seller’s private financial information. Amazon India confirmed the issue about the breach and resolved it as soon as it came into notice. The company said that some sellers experienced a technical issue when they attempted to download merchant tax reports for the month of December 2018. “Our teams identified the issue and resolved it on priority and sellers were soon able to download the correct reports”, Amazon India reportedly said. Information Leaked: Tax reports of approximately 400,000 Amazon sellers were disclosed to other sellers. Oyo: In a recent development, Federation of Hotel and Restaurant Associations of India (FHRAI) issued a notice to the Online Travel Aggregators and warned of action against Oyo adducing a large number of data breaching and staking the safety of consumers and also violating the laws. Earlier in 2018 Oyo Rooms was accused of stealing data by Zo Rooms suggesting that the prior had acquired data of employees, assets, and hotel properties under the pretext of accelerating the process of acquisition and is now refusing to pay the dues for the business acquired. There were concerns that though these pertain to breach of contracts between parties, the sheer number of such grievances from hotels is a matter of concern. In another report, OYO’s India and South Asia CEO Aditya Ghosh made an announcement that the managers and receptionists are equipped with an app that captures guest details and photos and directly shares them with the authorities without the consent of its customers. At present, the new surveillance proposal has reportedly been accepted by the governments of Rajasthan, Haryana, and Telangana. Information Leaked: The customer’s IDs that booked Oyo rooms (numbers not clarified).","excerpt":"Whether it was the Facebook-Cambridge Analytica scandal, or Quora row or even Google Plus security problem, 2018 was the year of data breaches. Data worth millions of dollars was stolen and was susceptible to being used illegally during this time. In an earlier article we discussed some of the big shots across the globe that […]","categories":["AI News"],"tags":["data breach india"],"author_name":"Ambika Choudhury","publish_date":"2019-02-16T08:30:45","publication_year":"2019","word_count":559,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","Git","RAG","data breach india","GAN","R","startup"],"extracted_tech_keywords":["AI","RAG","AWS","R","Go","Git","GAN","startup","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/data-breaches-faced-by-consumer-internet-companies-in-india-2018\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10088597,"title":"How Confidential Computing is Changing the AI Chip Game","content":"Amazon is warning its employees against sharing sensitive information with ChatGPT. Similarly, earlier, in conversation with AIM, Infosys’ Gary Bhattacharjee discussed that companies have much better chances to protect their IP with models like HuggingFace’s CodeGen, which is trained specifically on open-source codes, than with GPT which pretty much scrapes everything from the web. Amazon and Infosys are not alone. Plenty of companies are scared of exposing their data to ChatGPT, since the AI utilises the input prompt data to further train the model. On the other hand, to be able to make these AI models better, there is no other way than training it on as large a dataset as possible. And if we look at finance, healthcare, government, and other highly-regulated industries, ChatGPT-like technologies are a complete no-go. However, in order to increase access across these industries, cloud companies are working with silicon vendors to ensure data security through confidential computing. Confidential computing and LLMs Recently, OpenAI released its chatbot’s ChatGPT API for users to integrate into their apps and products. The data submitted to OpenAI API, CEO Sam Altman said, will not be used for training. Meanwhile, NVIDIA’s blog indicates that they are bringing GPU acceleration to VM-style confidential computing to market with its Hopper architecture GPUs. Hopper will be able to scale AI workloads in every data centre, from small enterprise to exascale high-performance computing (HPC) and trillion-parameter AI. According to a recent report by Tom’s Hardware, NVIDIA has begun shipping the H100 (Hopper), which is slated to be the successor to the A100. While media reports did highlight the significant performance improvements and higher AI training throughputs, what went amiss was H100’s integration of confidential computing. For OpenAI to fulfil its commercial ambitions, going deep into the data privacy aspect of it – despite the heavy cost bearing (it would need about 30,000 of these GPUs to run the model) – was absolutely critical. “Confidential computing is a more scalable way to solve for data security or privacy challenges related to ChatGPT as opposed to tokenisation or application-level encryption techniques,” says Kishan Patel, area vice president of sales at Anjuna Security. How does it work? Confidential VMs is a hardware-based security solution that allows organisations to safeguard their most sensitive data even while it’s being processed. The technology leverages a hardware-based trusted execution environment (TEE), which is basically a hardware-enforced security enclave completely isolated from the rest of the system. A two-step process explained by The Register gives an understanding: First, to enable confidential computing on GPUs, data is encrypted before being transferred between the CPU and GPU across the PCIe bus, using secure encryption keys which are exchanged between NVIDIA’s device driver and the GPU. Secondly, once the data is transferred to the GPU, it gets decrypted within a hardware-protected, isolated environment inside the GPU package, where it can be processed to generate models or inference results. The isolation ensures that the applications and data remain protected from various types of attacks that could potentially come from firmware, operating systems, hypervisors, virtual machines, and even physical interfaces like USB ports or PCI Express connectors on the computer. Source: NVIDIA NVIDIA vs AMD vs Intel NVIDIA isn’t the only player in town – Intel and AMD are also in the game making steady strides. Last year, Intel introduced Project Amber, which aimed to provide a security foundation for confidential computing, especially when it comes to training and deployment of AI models. And most recently, the company also added Trust Domain Extensions (TDX), which is based on the same VM isolation technology, to its 4th Gen Xeon processors. Similarly, AMD has partnered with Google Cloud to provide an additional layer of security to the chip designer’s Epyc processors. This was because at the time AMD was the only one providing confidential computing capabilities in mainstream server CPUs. The Register notes that there are greater incentives for chip companies to work with cloud providers like Microsoft, Google, IBM, Oracle, and others, for them to be able to buy a substantial amount of its processors. Security researchers in these cloud companies will be able to scrutinise every detail of the device implementation and its custom tests. Especially this, since independent researchers have in the past uncovered several flaws in both Intel SGX and AMD SEV many times. Additionally, there are also ongoing efforts from open-source communities like RISC-V to implement confidential computing in an open-source project called Keystone. However, while the above efforts have been towards ensuring security at the CPU level, NVIDIA’s Hopper architecture provides VM-style confidential computing to GPUs. Considering that NVIDIA has been going all in on AI, bringing confidential computing to GPUs gives it a further edge. According to one research, the confidential computing market could grow 26x in five years, making it up to $54 billion by 2026. Therefore, it will not be an overstatement to say that cloud security will be one of the biggest drivers in the AI chip race.","excerpt":"Cloud security will be one of the biggest drivers in the AI Chip race","categories":["Global Tech"],"tags":["AI Chips","AI Models","AMD","ChatGPT","cloud security","confidential computing","Data Security","Hopper architecture","Intel","OpenAI"],"author_name":"Ayush Jain","publish_date":"2023-03-03T11:30:00","publication_year":"2023","word_count":829,"keywords":["AI Models","AMD","ChatGPT","Go","OpenAI","AI","Data Security","AWS","cloud security","Scala","RAG","AI Chips","Aim","Hopper architecture","Rust","confidential computing","R","Intel"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","RAG","AWS","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-confidential-computing-is-changing-the-ai-chip-game\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10032744,"title":"Business Acumen Is Important For Data Scientists: Dr Hemachandran K, Woxsen University","content":"Data science and analytics are multidisciplinary in nature, making it an open-field for working professionals and aspirants from across domains. That said, some academic experts believe the commodification of data science education contributes to creating a demand-supply gap in the industry. To dig deeper, for this week’s data science career interview, we spoke with Dr Hemachandran K, Professor – Department of AI & ML at Woxsen University. In this interview, Dr Hemachandran sheds light on how the COVID pandemic spurred data science education, how governments and corporations participate etc. Excerpts: AIM: What’s the current data science education market look like? Dr Hemachandran K: The outlook of a job in the field of data science and analytics is always excellent. But the COVID pandemic has definitely given a facelift to the field of data analysis and data visualisation. And, graduates in these fields have pretty much always been excellent. If we compare with somebody who studies the history of data visualisation with the history of epidemiology, we can find many parallels. Today in the situation of this pandemic, we see that many public agencies are able to disseminate data about the spread of coronavirus. It’s more important than the information is presented in a way that is easy to understand for a common man. The worldwide consumers of data are really discovering the value of well-executed visualisations of data. The technologies that we teach at Woxsen University, whether it be applied analytics, predictive analytics, data mining, big data analytics, data visualisation, along with artificial intelligence and machine learning) programmes are specifically designed and more relevant to the industry right now. AIM: What are the challenges universities face while designing a data science and analytics course? Dr Hemachandran K: The most significant challenge for a university to augment data science and analytics education is the availability of skilled professionals. The management of Woxsen University believes the learning platform of the student is based on the skill of the teacher who teaches it. Department of artificial intelligence at Woxsen university is fortified with high configured NVIDIA GPUs, miscellaneous tools such as Raspberry PI, Arduino Boards, IP Cameras, Motion detector sensors, iPads, Amazon Echo’s, etc., along with qualified, passionate professors. The methodology they opt accelerates the students’ dream to attain a secured job in no time. AIM: What are the major obstacles the data and analytics education space is facing in India? Dr Hemachandran K: The major obstacles in the space are — ensuring the flow of data, which is essential for big data analytics. Alongside, poor connectivity, poor internet, poorly integrated data systems make the situations even worse. It is difficult to access data with poor data systems, and one can only be productive if we can correct these problems. It is also a challenge to educate and train educators, which is not only tiring and time-consuming but also sometimes unproductive. The availability of qualified teachers and mentors has to be ensured. Plus, necessary actions need to be taken to bring all teachers and mentors to a mindset to cooperate and show curiosity. AIM: How can governments and corporations encourage data science education space? Dr Hemachandran K: Educational data mining and learning analytics are used in research and to build models that can influence online learning systems. Analytics helps to detect whether a student in an online course is going astray and he\/she needs a correction in the course. It helps to identify the boredom patterns and redirects the student’s attention. The government can offer data analytics courses for students with lucrative scholarships. The government of India can also offer internships in some of their agencies and can give scholarships to do internships abroad. They can develop MOUs with various universities within the country and outside for encouraging student exchange programs. The government, along with some agencies like the Ministry of Human Resource Development, All India Council for Technical Education, can popularise data science and analytics courses rurally by offering short term, instructor-led courses with a duration of 10 to 15 hour. This can help many to explore the field of data analytics. AIM: How does an industry partnership add value to the data science courses? Dr Hemachandran K: Moving beyond theoretical concepts of simple data analysis, it’s important to equip students with technical knowledge and business acumen for translating complex data sets into meaningful information. These insights are vital in making intelligent business decisions. Industry connection creates an opportunity to engage top-notch students in the leading industries, which helps them become new business leaders. To engage the rising stars of the analytics field and add value to their business, hands-on experience with industrial leads are indispensable. Connecting students to the industry helps to revamp their skills and knowledge required to cater to the needs of the industry. This, in turn, helps bridge the gap between the students and the cutting-edge knowledge required by the industries. This formulates the student to prepare themselves for their future and equip themselves for better careers. AIM: With a lot of online courses and MOOCs available, how does a professional degree make a difference in the long term data science career? Dr Hemachandran K: Professional degree courses provide you with in-depth knowledge on data science, real-time analytics, statistical computing, SQL, parsing machine-generated data and the domain of deep learning in AI. It also helps to leverage big data analytics with a spark of data science. AIM: What’s your advice for aspiring data scientists\/or similar roles? Dr Hemachandran K: I believe in — “Success is a journey, not a destination”. The values, lessons, challenges learning in the process of becoming great is what makes the process worth following. A data scientist solves business problems using statistical, mathematical and data techniques. The values that one needs to learn, to embrace and focus on are all in the mindset. Anyone who possesses a bachelor’s degree in data science or from a related field can start a career as a data scientist. A certified course in data science will be preferable. A background in mathematics and statistics will be an added value. Expertise in any of the programming languages, like R or Python, will be beneficial. According to inputs from Glassdoor, data scientists make an average of $116,100 per year, making data science jobs lucrative, now more than ever. So this is the right time to pursue a career in data science and analytics. Some of the key things to remember in your data science journey are — ConsistencyStructured thinkingUnderstanding the problemPlanning is criticalDivide the work plan into segmentsExecute it","excerpt":"Moving beyond theoretical concepts of simple data analysis, it’s important to equip students with technical knowledge and business acumen for translating complex data sets into meaningful information.","categories":["AI Features"],"tags":["career in data analytics","career in data science","Data Science Career","data science education","Data Science Jobs","Interviews and Discussions"],"author_name":"Sejuti Das","publish_date":"2021-04-22T13:00:00","publication_year":"2021","word_count":1089,"keywords":["data science","artificial intelligence","machine learning","AI","ML","Data Science Jobs","Interviews and Discussions","Data Science Career","Ray","Aim","deep learning","RAG","analytics","data science education","career in data analytics","career in data science"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/business-acumen-is-important-for-data-scientists-dr-hemachandran-k-woxsen-university\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10022873,"title":"In Conversation With Ambee &#038; Razor Network On Their Recent Blockchain Partnership","content":"Blockchain technology, when fed with accurate and reliable data, can do wonders. This is precisely what we are witnessing with two Indian startup companies — Ambee and Razor Oracle Network that announced a recent partnership to develop a range of blockchain applications. The two companies have come together for a cause — to provide reliable environmental data to blockchain applications and support APIs for air quality, weather, pollen, fire, soil, water vapour and more. The data generated with regards to hazardous fluctuations in the environment are typically neither verifiable nor tamper-proof. Alongside, such data has never been publicly available and, most importantly, not scalable. While Razor Network is a completely decentralised blockchain agnostic Oracle network for smart contracts, Ambee builds hyperlocal environmental data and intelligence in real time across the world. Thus, this partnership will help developers from environmental blockchain startups build a range of applications on leading blockchain platforms and come out with unique solutions for businesses affected by environmental factors and climate change. Analytics India Magazine sat down with Madhusudan Anand, CTO and Co-founder at Ambee; and Hrishikesh Huilgolkar, CEO, Razor Oracle Network, to discuss the coming together of two startups, challenges faced, and how the partnership can provide long-term solutions. Edited Excerpts — AIM: How did Ambee and Razor Network come across the idea of this partnership? Madhusudan Anand: We aim to democratise the availability of data and make a difference. The integration with Razor Network will help us provide accurate and reliable environmental data to blockchain applications and their users. Hrishikesh Huilgolkar: We are here to bridge the gap between blockchain platforms and real-world data. Applications on the blockchain platforms require a variety of data to function, which is not available readily. However, we do need accurate and reliable data providers to make this data available to the blockchain. Hence to make a variety of trustworthy real-time data available to our clients, we partnered with Ambee. AIM: What are the key challenges you faced, from the very point of initiation till the signing of this agreement? Madhusudan Anand: Given that this is a first-of-its-kind partnership, the decision itself wasn’t easy as we had to research market viability as well as applications of this for deeper problem-solving. The integration was challenging. Decoupling traditional authentication methods to align with blockchain architecture and building a flexible data infrastructure layer without compromising user experience, response time, and quality of service were the key challenges before us. AIM: Which all data you collect and how? Madhusudan Anand: We build hyperlocal environmental data and intelligence in real-time. Using proprietary data science, Ambee provides location-specific, real-time data and actionable insights on air quality, weather, pollen, and various other environmental factors, to be consumed by business and administrators worldwide. AIM: How is the use of blockchain technology, helping you to come out with solid solutions? Madhusudan Anand: The incorporation of environmental data in blockchain applications addresses multiple ecological challenges and helps fight climate change. It can be used to ensure supply chain transparency in claims made by companies on environmental impact, as well as encourage citizens through financial rewards to adopt environmentally friendly practices. The technology can help understand the primary reasons behind the increase in temperature and track greenhouse gas emissions at a postcode level. In addition to this, in agriculture, weather and soil data can be used for forecasting yield, better crop management, and measuring risks associated with extreme weather conditions. AIM: What benefits can be expected to post the partnership? Madhusudan Anand: Many use cases in natural resource management, validation to claims of reduced environmental impacts and support for environmentally sustainable actions are some of the many possibilities that the partnership can explore in the future. By entrusting a decentralised network, instead of a central institution, Blockchain technology has the potential to reconfigure the way users allocate, protect and transfer their assets or services, especially in an ecosystem such as air quality data and pollutant trends. Hrishikesh Huilgolkar: The speed of industrial consumption and needs will always be more than the rate of rejuvenation. By making micro geographical level data available in real-time to blockchain applications, we will see novel ways to solve the problems arising from excessive damage to resources and increasing toxic and waste excretion across the industries. Some examples may include prediction markets to predict or hedge pollution levels, decentralised insurance based on the environment, such as decentralised crop insurance, etc. AIM: How does this pave the way for more opportunities in other areas? Madhusudan Anand: As per a report, the global blockchain in the agriculture market will grow at 45.13% from $57.4 million in 2019 to $518.7 million in 2027; the segment seems more promising for new endeavours. One of the several applications is the use of agri-weather data for agriculture in the blockchain, boosting the decision-making abilities – thereby transforming the global agricultural and food industry. Hrishikesh Huilgolkar: We will start with bridging data provided by Ambee with one blockchain, but over time we will be making it available to our 12+ blockchain partners. With the availability of quality data, we can see applications having a profound impact on the climate change problems, both at the micro and macro level, including— decentralised carbon credits trading on the blockchain; solar energy credits, sunlight availability insurance; derivative financial products based on various environmental data, etc. AIM: How other tech-startups can also collaborate with you to build more B2B solutions? Madhusudan Anand: We are here to simplify business processes, and with accurate environmental intelligence and blockchain technology, startups from a plethora of fields such as food safety, climate accounting, to supply chain management can leverage data to implement practical solutions and make informed decisions. Hrishikesh Huilgolkar: Since climate change and other environmental challenges are too big to solve for a single startup; we encourage other startups to also collaborate with us so we can share our research and build better solutions. Solving such a problem will require technological collaboration from different companies across industries.","excerpt":"Ambee, the accurate environmental data provider, partners with Razor blockchain technology provider to develop a range of applications.","categories":["AI Features"],"tags":["Blockchain","blockchain india","Blockchain Technology","Data Science","environment","Interviews and Discussions","partnership","partnership India"],"author_name":"kumar Gandharv","publish_date":"2021-03-25T15:00:00","publication_year":"2021","word_count":995,"keywords":["partnership","data science","Go","API","Blockchain","blockchain india","AI","Scala","RAG","Aim","analytics","environment","Rust","Blockchain Technology","Data Science","partnership India","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Rust","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/in-conversation-with-ambee-razor-network-on-their-recent-blockchain-partnership\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10074593,"title":"Consumerism That Is Powered By Technology &#038; Our Innate Desire For Speed","content":"Technology has enabled the setting in of a new age of consumerism. Public policy experts Divya Singh Rathore and Pratyush Prabhakar term this new age of consumerism the age of hyper-lapse consumerism. According to them, along with the ever-increasing consumption of goods and services, consumers today are also prioritising the ease with which they are able to place orders for various goods and services and the speed at which they are delivered. Hyper-lapse consumerism is centred around the idea of being the fastest to reach the consumer, and it has largely been fuelled by the ubiquitous internet and the growth of e-commerce. According to India Brand Equity Foundation (IBEF), the Indian e-commerce market is set to reach USD 74.8 billion in 2022, registering a 21.5% increase. E-commerce has not only transformed the way businesses operate in India but also impacted consumerism. Could you imagine ordering your favourite ice cream late at night ten years back? Or could you imagine sitting somewhere in Kerala and ordering Ghewar? E-commerce has turned these fantasies into reality. Source: IBEF Today, in India alone, there are multiple e-commerce platforms being driven by the notion of hyper-lapse consumerism. Recently, we have seen advertisements by a renowned grocery e-commerce platform claiming to deliver orders within 10 minutes. With the continuous evolution of technology, consumer behaviour is also undergoing changes. In the initial days, online shopping was just another medium for shopping with fewer options. Consumers resorted to buying those products online that they did get in the local market. However, in the last few years, this field has revolutionised to attract more and more customers and cover segments like daily essentials. In the financial year 2021, the Indian online grocery market size was USD 3.95 billion. It is estimated to grow to USD 26.93 billion by 2027 at a CAGR of 33%. In fact, a distinct category of consumers who shop exclusively online exists now. Such has been the extent of evolution of consumerism driven by e-commerce companies. The race to reach consumers as fast as possible is in full swing. Let alone grocery and other perishables delivery platforms like Blinkit, Licious and Zepto, e-commerce giants like Amazon and Flipkart are also being driven by hyper-lapse consumerism. They have incorporated features like Prime and Plus in efforts to prioritise consumers’ delivery expectations. In some cities, they even offer consumers options for same day deliveries. Undoubtedly, the e-commerce market in India has recorded an impressive growth in the last few years driven by a multitude of factors including increasing investments in e-commerce firms, increase in digital literacy, policy support in the form of favourable climate for investment—100% FDI under automatic route is allowed in the marketplace model of e-commerce—along with increase in internet and smartphone penetration. Indian consumers are increasingly adopting 5G smartphones even before the rollout of the next-gen mobile broadband technology in the country. Easy credit putting consumer credibility to question However, another crucial factor driving the new age consumerism is the availability of easy credit on e-commerce platforms. Earlier, consumers resorted to credit options for buying assets like property or costly durable goods like cars. However, of late, consumers are using  ‘credit’ to buy almost anything like mobile phones, mixers and grinders and other trivial housekeeping items. This has been made possible due to several credit options available on e-commerce platforms. In fact, Flipkart in partnership with IDFC FIRST bank provides customers the option to ‘buy now and pay later’. Interestingly, this is also available for not-so-high-valued products like apparels, yoga mats, and other commonplace household items. At a glance, it may appear very democratising—after all, in a country like India where 45% households account for lower middle income households, people may have the desire to buy a branded shirt that costs say INR 2000 but cannot pay right away. However, many new-to-credit consumers have no proper credit history. Thus, providing easy loans could trigger a culture of credit indiscipline. People stop being mindful of what they should or should not buy on credit. Often such irrational behaviour leads consumers to borrow much more than their repayable capacity. This can ultimately put them in perpetual debt traps. Does that mean hyper-lapse consumerism is all bad? Not at all. Both consumers as well as businesses need to be mindful of this evolutionary phase of consumerism and the dire consequences it may have. While e-commerce platforms need to be conscious of the social, behavioural and ethical implications of instant deliveries, consumers too need to be responsible while availing credit. It is advisable that they deploy the 50\/30\/20 rule in dividing income among needs, wants and aspirations respectively. Using financial leverage will not only help acquire valuable assets but also prevent one from resorting to reckless credit.","excerpt":"Along with the ever-increasing consumption of goods and services, consumers today are also prioritising the ease with which they are able to place orders for various goods and services and the speed at which they are delivered.","categories":["IT Services"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-09-08T11:00:00","publication_year":"2022","word_count":789,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/consumerism-that-is-powered-by-technology-our-innate-desire-for-speed\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10050117,"title":"AWS Announces Data and Analytics Conclave — A Virtual Event To Empower You To Reinvent Your Business With Data","content":"Amazon Web Services (AWS), in association with Analytics India Magazine, is presenting the Data and Analytics Conclave on 21st October 2021 — an online conference designed to empower technical and business decision-makers, data and analytics practitioners, and data or decision scientists. This day-long conference will explore the importance of a modern data strategy that isn’t one size fits all but adapted to the needs of every organisation. Surrounded by a sea of data, success for any organisation depends on how well they are able to derive insights from it and make transformative decisions. In order to progress, businesses need to adopt strategies that can hold large amounts of data in open, standards-based data formats. Modern data strategies are not limited by data silos. This allows companies to run analytics operations using preferred tools while also maintaining strict security standards. The Data and Analytics Conclave aims to bring together data experts from AWS along with business leaders and leading engineers from the industry who will demonstrate how modern data strategies can accelerate innovation and build smart, customer-centric, scalable solutions on cloud and edge using AWS machine learning and AI services. The conclave will help leaders & business decision-makers to reinvent their businesses through a modern data strategy combining data lakes and purpose-built data stores. The conclave features an opening keynote from Greg Khairallah, Director, Analytics at Amazon Web Services (AWS) and insightful talks by IDC analysts and other business leaders. In addition, the conference also dives deep into ten breakout sessions across three tracks — Modernise Data Infrastructure; Unify Your Data; & Innovate With AI\/ML. Participants will learn how to build, train, and deploy sophisticated models with any framework, thus preparing them for the challenges of tomorrow. The conference anticipates the participation of 2000+ attendees from around the world, to learn more visit the event registration page. REGISTER NOW Event Highlights: Opportunity to learn about building a modern data strategy in the cloud and analysing data using a range of analytics approaches, including machine learning.Gaining a deep understanding of Amazon analytics services that covers data warehouse modernisation, implementation and acceleration of Apache Spark, along with other big data applications and data lakesInsights from Industry and AWS data and analytics leadersA one-on-one discussion with AWS analytics experts. The target audience includes but is not limited to: Those who have just started with AWS, an advanced user, or business executives. Special tracks are available for different levels of experience and job rolesFor the technical and business decision-maker, data and analytics practitioner, data scientist, business intelligence engineer, and IT architects, among others. REGISTER NOW When: 21st October 2021 To know more, click here. About Amazon Web Services AWS or Amazon Web Services is a subsidiary of Amazon that offers on-demand cloud computing platforms and APIs. It serves individuals, companies, and governments. These cloud-based web services provide the technical infrastructure and distributed computing building blocks and tools. AWS offers a slew of services like computation, storage, database management, machine learning, AI, data lakes, and the internet of things. In addition, it offers its customers a faster, easier, and cost-effective way to move existing applications to the cloud. REGISTRATION NOW OPEN","excerpt":"An exclusive virtual conference designed to empower you to reinvent your business with a modern data strategy adapted to every organisation’s needs.","categories":["Deep Tech"],"tags":["AI event","conference","data and analytics","machine learning conference","Modern data infrastructure","modern data strategy","modern database","modernising data platforms"],"author_name":"AIM Media House","publish_date":"2021-10-01T14:00:00","publication_year":"2021","word_count":525,"keywords":["machine learning conference","data and analytics","machine learning","AWS","AI","conference","cloud computing","Modern data infrastructure","ML","distributed computing","Apache Spark","RAG","modern data strategy","Aim","modern database","AI event","analytics","modernising data platforms"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","cloud computing","AWS","distributed computing","Apache Spark"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/aws-announces-data-and-analytics-conclave-a-virtual-event-to-empower-you-to-reinvent-your-business-with-data\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10168194,"title":"Acer Aims to Make AI PCs Accessible to Every Indian, Not Just the Elite","content":"AI PCs have begun reshaping the future of computing; however, they still remain bracketed in the premium sector of the market. In India, major players HP, Dell, Lenovo, and Acer are working hard to reduce prices, making these futuristic devices accessible to a broader audience in the coming years. Acer India, for instance, is taking a multi-pronged approach to stay ahead in the race. The company is betting on affordability, local manufacturing, and deep partnerships with tech giants like Intel, AMD, Qualcomm, and NVIDIA to offer AI devices across price segments, even as rivals like Lenovo and Dell double down on their own AI PC ambitions. “We’re not just participating in the AI PC revolution—we are shaping it,” Sudhir Goel, chief business officer at Acer India, told AIM. “While many players are focusing exclusively on the premium segment, Acer is committed to democratising access to AI.” Democratising AI What qualifies as an AI PC, according to Microsoft, is the ability to deploy Copilot+. For this, Microsoft requires manufacturers to integrate at least 16 GB of memory, a 256GB SSD, and an NPU that can process a minimum of 40 TOPS, or trillion operations per second. Currently, AMD, Qualcomm Snapdragon, and Intel are offering NPU-enabled chips, with OEMs such as ASUS, Acer, Lenovo, Microsoft (Surface), HP, and others building these AI-enabled laptops. In September last year, American research firm Gartner predicted that by 2026, AI-enabled PCs would occupy 100% of the enterprise market. While this was a bold claim, a recent survey from IDC, in collaboration with AMD, revealed that various industries were indeed moving in this direction. Acer’s strategy of making AI capabilities accessible beyond high-end models involves integrating AI features not just in its premium devices but across entry-level and mid-range offerings. This is bolstered by its proprietary Acer AI Zone software that promises intuitive and optimised user experiences across devices. The top-right corner of its keyboards now has the new AcerSense key, a shortcut to Acer’s control centre. It lets you switch power modes, monitor system temperature, and manage storage with ease. The hub also includes the AI Zone, where users can explore AI tools like PurifiedVoice, AlterView, and Intel’s Stable Diffusion plugin—all in one place. “We’re working to integrate meaningful AI features into products at every price point,” Goel explained. “It’s about creating smart, practical experiences that enhance everyday life.” This vision is supported by Acer’s participation in the ‘Make in India’ initiative, which allows the company to control production costs and price its devices more competitively for Indian consumers. Acer is also exploring flexible financing options, including EMI plans and partnerships with financial institutions, to boost AI PC adoption. Meanwhile, Lenovo and HP are pushing strongly in the enterprise field to reduce costs. In an interaction with AIM, Saurabh Agarwal, COO of Lenovo India, expressed his optimism towards industries adopting AI PCs. “As AI PCs evolve, we are optimising our devices for on-device AI processing, enhancing productivity and security for enterprises and customers,” he said. “AI is expected to drive the next PC refresh cycle and act as a major growth catalyst, with these devices projected to capture 20% of the market by FY25, further fueling demand and innovation in the industry,” added Agarwal. Expanding Beyond Metros Recognising that India’s next wave of growth will come from smaller cities, Acer is making a strong push into Tier 2, Tier 3, and Tier 4 markets. The brand is increasing the number of exclusive stores, bolstering retail partnerships, and strengthening its e-commerce presence to reach users across geographies. The company is tailoring its product mix for these markets with a special focus on AI-enabled and budget-friendly devices to meet the evolving needs of consumers outside urban centres. The most affordable Copilot+ AI PC right now is Acer Swift Go 14, priced around ₹66,000, which is slightly lesser than HP’s Pavillion series at ₹69,000 and Lenovo IdeaPad Slim 5x Gen 9 at ₹76,000. Less visibly, Acer is also building muscle in the AI infrastructure space through its Altos Computing division. Altos servers and workstations are already being used for large-scale AI modelling and LLM processing, including deployments in India. “With increasing mandates for data localisation and the rise of AI research, there’s a real opportunity here,” Goel said. “We’re collaborating with Intel, AMD, and NVIDIA to optimise AI-ready server solutions for Indian enterprises, research institutions, and cloud providers.” As India’s AI ecosystem grows rapidly, Acer seems to be positioning itself not just as a device maker but as a full-stack enabler, from edge to cloud. Beyond computing, Acer is applying AI to its Acerpure line of home electronics, which includes air and water purifiers, TVs, and air conditioners. According to Goel, AI features are being used to enhance performance, efficiency, and user interaction across these appliances. “For example, our AI-driven air purifiers can adjust settings based on real-time air quality and humidity levels,” Goel said. “Our televisions will offer AI video and gaming enhancements, while air conditioners are being designed for energy-efficient, climate-responsive cooling.”","excerpt":"Recognising that India’s next wave of growth will come from smaller cities, Acer is making a strong push into Tier 2, Tier 3, and Tier 4 markets.","categories":["AI Features"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-04-17T14:00:00","publication_year":"2025","word_count":835,"keywords":["Go","API","AI","innovation","RAG","CuPy","Aim","stable diffusion","ViT","R"],"extracted_tech_keywords":["AI","Aim","CuPy","RAG","R","Go","API","stable diffusion","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/acer-aims-to-make-ai-pcs-accessible-to-every-indian-not-just-the-elite\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10065360,"title":"Talking Ethical AI with Aerospike’s Aveekshith Bushan","content":"Aerospike’s real-time data platform enables organisations to act instantly across billions of transactions while reducing server footprint by up to 80%. Its multi-cloud platform powers real-time applications with predictable sub-millisecond performance up to petabyte-scale with five-nines uptime with globally distributed, strongly consistent data. “The Aerospike Real-time Data Platform is used by applications to process fast-changing data at high throughput (scale) to facilitate artificial intelligence and machine learning model-based decisions within a fixed SLA (typically in milliseconds or tens of milliseconds). Applications include fraud detection, recommendation engines, real-time bidding for ads, inter-bank money transfers, etc. Aerospike technology enables petabytes of data to be processed efficiently, thus improving the quality of AI\/ML models generated as well as the quality of decisions made by applying those models in real-time,” said Aveekshith Bushan, Regional Director & General Manager, Asia Pacific, Aerospike. In an exclusive interview with Analytics India Magazine, Aveekshith spoke about ethical AI and how it’s embedded in their platform. AIM: What explains the growing conversation around AI ethics, responsibility, and fairness? Aveekshith Bushan: According to the Worldwide Artificial Intelligence Spending Guide from International Data Corporation (IDC) forecasts, global spending on AI systems will increase from USD 85.3 billion in 2021 to more than USD 204 billion in 2025. However, with the increase of AI in an organisation’s system to cope with the rising levels of online interactions, ethical biases and transparency issues arise. The need to develop an ethical charter that defines AI algorithms is not only a moral responsibility but also a business imperative. While Aerospike does not ingest consumer data, our customers do while working on our real-time platform. So, they must implement and follow ethical standards to provide safety guidelines that can prevent risk for both business and human interactions. AIM: Why is it the need of the hour? Aveekshith Bushan: AI\/ML and the data required to create desired outcomes sit at the heart of the digital transformation businesses are experiencing today. As more and more data become available, more and more AI-based applications will be created, and existing applications will have to handle extreme loads of data to make better decisions. It’s mission-critical for organisations to adhere to ethical standards and pace themselves accordingly to keep up with the increase in AI\/ML processing. Government organisations, enterprises, and small-to-medium size businesses all rely on AI\/ML computations to drive their “decisioning” in a manner compliant with diverse ethical, social, and legal norms. India’s Telecommunication Engineering Centre (TEC), the technical arm of the Department of Telecommunications (DoT), has started discussions to create a framework for a fair assessment of AI and ML systems to build public trust. AIM: What’s the biggest challenge companies face while ensuring AI governance at scale? Aveekshith Bushan: The biggest challenge is that an algorithm that is launched initially for small workloads needs to continue to work well as it scales. The millionth or billionth user must have the same experience as the first user. This means the platform needs to scale up and be elastic in its scale-out. Non-scalable platforms tend to stop a hyper-growth business in its tracks. For example, if fraud detection SLAs do not keep up with the growth in transactions, a significant number of transactions will have to proceed without a fraud score. This in turn will lead to more exposure to undetected fraud and could eventually result in huge losses. Using platforms like Aerospike which has a proven track record of delivering predictable performance at scale with high uptime would be a way to prevent disruptions to high growth businesses. AIM: How do you protect consumer data? Aveekshith Bushan: Digitisation comes at a cost. Each of us now leaves a trail of digital exhaust, an infinite stream of phone records, texts, browser histories, preferences, buying patterns, location, and other information that lives forever. Every time someone views a webpage, check-outs at the supermarket, receive an electricity meter report, gets a package that passes through a delivery checkpoint, scans an ID, or posts on social media –the digital trail grows. And so does the need for greater security and privacy of data for consumer and other business use case data. There cannot be a trade-off between speed and scale and security and privacy, they must move forward as one. It is important that applications built to power these digital transactions provide a mechanism for security and privacy. The Aerospike Real-time Data Platform provides a capability for cross datacenter replication (XDR). Aerospike’s XDR enables enterprises to create a global data hub (Figure attached) that automatically routes, and augments data captured anywhere in the data centre to wherever it’s needed – whether in Aerospike clusters or any other data repository.  Typically, regulatory compliance is highly regional, and the new XDR provides the ability to manage regulatory requirements such as GDPR and CCPA on a regional basis.","excerpt":"Global spending on AI systems will increase from USD 85.3 billion in 2021 to more than USD 204 billion in 2025.","categories":["AI Features"],"tags":["AI fairness","ai governance","Cloud services","consumer protection","Data Governance","data laws","Data Privacy","data rights","digital transformation","Ethical AI","Fair AI","Interviews and Discussions","legal","modernisation","privacy laws","Trustworthy AI"],"author_name":"Sri Krishna","publish_date":"2022-04-20T17:00:00","publication_year":"2022","word_count":802,"keywords":["consumer protection","data rights","ai governance","legal","Rust","R","fraud detection","artificial intelligence","digital transformation","Cloud services","Data Governance","analytics","modernisation","Go","privacy laws","machine learning","AI","Ethical AI","ML","Data Privacy","AI fairness","Fair AI","data laws","Trustworthy AI","Aim","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","fraud detection","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/talking-ethical-ai-with-aerospikes-aveekshith-bushan\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10170177,"title":"The Dark Side of o3","content":"OpenAI’s o3 is among the best-performing reasoning models available for users today. Benchmark scores indicate that the model outperforms several competing models across various aspects, including coding, math, graduate-level science problems, and more. Several users on social media have praised the model’s performance. However, the model’s most significant drawbacks are hallucinations and reward hacking, or specification gaming. A Warning Sign for Future Reasoning Models A recent study published by Palisade Research, a non-profit organisation, reveals that OpenAI’s o3 model is subject to ‘specification gaming’ — a process where an AI model takes the objective of a given problem too literally, deviates from an acceptable process, and engages in malpractice to achieve its purpose. In such cases, a model is determined to achieve its result and will use unintended methods. The research set up an AI model to play chess against the Stockfish chess engine. This experiment found that AI models — OpenAI’s o1-preview, o3 and DeepSeekR1 often observe that the chess engine is too strong for them to win against, and then hack the game environment to win. “Surprisingly, o1 and o3-mini do not show this behaviour,” read the report. “In contrast, o3 shows extreme hacking propensity, hacking in 88% of runs.” These hacks involved confusing the engine, replacing the board, and at times, replacing the engine itself. “Such behaviours become more concerning as AI systems grow more capable. In complex scenarios, AI systems might pursue their objectives in ways that conflict with human interests,” read the report. As AI systems enhance their situational awareness and develop strategic reasoning about their surroundings, such occurrences may become more frequent. This issue is particularly problematic when equal focus is necessary on both problem-solving methods and the solutions themselves. Source: Palisade Research Specification gaming is an infamous practice that has been observed in AI systems throughout time. While the above study focuses on a game of chess, the researchers shared a document outlining more such scenarios. The Palisade Research report suggests that as more reasoning and agentic models emerge, they may be more prone to gaming the objectives, and calls the study an ‘early warning sign’. “First, we suggest testing for specification gaming should be a standard part of model evaluation. Second, increasing model capabilities may require proportionally stronger safeguards against unintended behaviours,” read a report section. 2x Hallucinations than the o1 Model Besides, several users have found the o3 model to hallucinate across multiple scenarios, Several users across social media have expressed their frustrations towards the hallucinations present inside the model. And OpenAI acknowledges this as well. Earlier, the company released a ‘model card’ for the newly released o3 and o4-mini models, outlining the model’s behaviour and shortcomings. Benchmarks assessing the model’s hallucinations reveal that the o3 model had a higher rate compared to its predecessor, the o1. Notably, in the PersonQA evaluation—a dataset featuring questions and publicly accessible facts about individuals that assesses the model’s accuracy in answering—the o3 model exhibited double the hallucination rate of the o1. Source: OpenAI In the model card, OpenAI also outlined some of the model’s other unintended behaviours, such as reward hacking, under-reporting its capabilities, deception, and so on. Last month, Transluce, another independent non-profit research lab, outlined their findings on the pre-release version of the o3 model and revealed that the model ‘frequently fabricates actions’ that it never took, while also ‘elaborately’ justifying these actions when confronted with them. Experiments revealed scenarios in which the model claimed to run non-existent code on its own laptop, insisting that it did so. Other situations include making up its own time, ‘gaslighting’ the user about incorrectly copying a piece of information, and pretending to analyse log files from a web server. This is actually wild. ChatGPT o3 is amazing but the hallucinations are pretty out of control.I asked o3 to give me some clips of prominent AI figures talking about Ethereum and it gave me this chart.One problem, none of these quotes are real… pic.twitter.com\/CMC6hfuEsJ— Eric Conner (@econoar) May 8, 2025 In addition to standard issues like hallucinations, Transcluce outlines factors that arise from outcome-based Reinforcement Learning (RL) training—a model that learns through trial and error. It is guided by a reward system that provides rewards for correct answers and penalties for incorrect ones. The study indicated that if the reward function only rewards correct answers, a model lacks the incentive to admit it cannot solve a problem, as this does not count as correct. When confronted with unsolvable or overly complex problems, the model may still take a guess at an answer in case it is accurate. These problems may also arise due to chains of thought, in which the model outlines its reasoning steps before providing the response. The study indicates that the model’s internal chain of thought is obscured and detached from its conversational context, resulting in the model losing track of its prior reasoning. Therefore, when asked about previous statements, it has to create believable explanations since it cannot recall the actual basis for its earlier responses. “To put this another way, o-series models do not actually have enough information in their context to accurately report the actions they took in previous turns. This means that the simple strategy of ‘telling the truth’ may be unavailable to these models when users ask about previous actions,” added the study.","excerpt":"Research studies show OpenAI o3 model’s increased tendencies to achieve objectives through malpractice, and users are frustrated about the hallucinations.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","OpenAI","OpenAI o3"],"author_name":"Supreeth Koundinya","publish_date":"2025-05-19T14:47:51","publication_year":"2025","word_count":885,"keywords":["ChatGPT","GPT o3","OpenAI","AI","OpenAI o3","GPT","Aim","CLIP","Rust","chain of thought","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","GPT o3","ChatGPT","OpenAI","Aim","chain of thought","R","Rust","GPT","CLIP"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/the-dark-side-of-o3\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":11108,"title":"Dealing with Data Deluge – New Age Marketer’s Top Priority","content":"From taking an Uber to office, to ordering grocery online, paying utility bills, capturing a picture of the whiteboard after that strategic discussion, and sharing it with the team on WhatsApp, all our day to day actions contribute to the exponential growth in the volume of data generated by us each day. Overwhelmed by this enormous volume of data at their disposal, businesses that were previously focused only on capturing and acquiring data about their consumers are today faced with a new challenge of making sense of data. Not all data is relevant. Moreover, data is complex and to derive any useful insight from it you require advanced analytics capabilities. Amidst all this complexity stands today’s marketer – set out to conquer the world. Marketing Literally Lives on Data However complex, or dirty – data is a necessity for marketing to deliver results. Over the past couple of years, the role of marketing has evolved drastically. This is a direct consequence of the digital revolution and its impact on the buying cycle of the digitally savvy modern day consumer. Research suggests that today nearly 57% of the buying cycle is already complete by the time vendors come to play. Further, it’s interesting to note that almost 75% of buyers use social media to make buying decisions. Really? What does that imply? To me it represents increased onus on marketing to make sales happen. And, to live up to the great expectations that the business has from them, marketers need to uncover the power of data. A solid data strategy buttressed by an effective analytics engine can enable marketers to deliver personalized customer experiences at every stage of the customer journey. This directly translates into better quality and bigger sales pipelines. Identifying Your Ideal Customer Profile Buyers’ digital footprints are fairly indicative of their preferences, behaviors, demographics, and other characteristics. From the search terms we use, to the content we consume, the brands we interact with, or the places we frequent – everything speaks something about us. This data can form the basis for marketers to divide their total addressable market into segments with identifiable common needs and priorities. The next step is to define unique personas that represent each of these segments. And the final step in the segmentation process is to identify your ideal customer profile or the needy niche that has an immediate requirement or unresolved problem that you can address. This process might even help marketers discover segments that might perfectly fit their ideal customer profile, but have never been exposed to their products in the past. Get Your Data Strategy Right Coming back to where we started – while data is the lifeblood for marketing, ultimately it’s the marketer’s ability to derive actionable insights from the data deluge, that defines the success of any business operating in this digital age. As such it’s important for marketers to invest in implementing the right tools that can slice and dice every bit of data input to deliver market intelligence. An even better option is to rely on credible market intelligence sources to deliver extremely targeted, relevant, and accurate insights for us to leverage. Further, as marketers we need to understand that our ideal data strategy must not be limited to just data acquisition. That’s only the first step. Rather, our data strategy must comprise of data acquisition, data nurturing, data enrichment, continuous data health check, and regular data cleaning. If your data strategy is missing any of these key elements, it’s time to revisit it and set things right.","excerpt":"From taking an Uber to office, to ordering grocery online, paying utility bills, capturing a picture of the whiteboard after that strategic discussion, and sharing it with the team on WhatsApp, all our day to day actions contribute to the exponential growth in the volume of data generated by us each day. Overwhelmed by this […]","categories":["AI Trends"],"tags":["marketing analytics"],"author_name":"Garima Rai","publish_date":"2016-11-05T06:54:39","publication_year":"2016","word_count":591,"keywords":["programming_languages:R","AI","marketing analytics","Git","RAG","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/dealing-data-deluge-new-age-marketers-top-priority\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":41348,"title":"How Successful Data Scientists Use Task-Batching To Maximising Productivity","content":"Data science without a doubt one of the most rapidly growing domains in the tech industry. However, the domain is not something that anyone can work in — one needs to have tremendous skills, knowledge and staying power to deal with the daily hustle and bustle. Whether it’s about extracting the exact insight or it’s about building that ultimate model, a data scientist puts in a significant amount of effort on a daily basis. Meaning, despite the fact that it’s the “sexiest job of the 21st Century” and one of the highest paying jobs, being a data scientist can be stressful. If you are a data scientist and struggling to be efficient and productive at work, then task batching could be the solution. If done right, it can literally change the way your work gets done and make you extensively productive. What Is Task Batching Task batching is a time management system that helps an individual complete his\/her daily tasks in a uniform manner. It is considered to be one of the best ways to maximise productivity as it makes you focus on one thing at a time. As we all know, the job role of data science is stressful sometimes, task batching could be the best way to reduce that stress. All you need to do is group the list of similar tasks, and complete them together at a dedicated time period. So, here is how you can master the art of task batching: Understand Multitasking Doesn’t Work All The Time No doubt the job of a data scientist is fast-paced. However, that doesn’t mean, you would take up multiple tasks at once and try to finish them. You might feel you are good at multitasking but at one point you will realise the cons of multi-tasking. For example, if you focus on extracting data from different sources and migrating them to a data repository, and at the same time you are also busy sending some emails, there are chances that you would miss you some of the insights that could be useful or make mistake while sending the email. Also, taking breaks is one thing and getting distracted is another. Reportedly, University of California Irvine conducted a study, and they found that it takes an average of 23 minutes and 15 seconds to get back to the task. Distractions are everywhere — be it emails, meetings, texts, calls etc. But the only thing we need to focus on is to not let these distractions the top place on our priority list. Therefore, the first and foremost effort you need to take is to completely avoid switch between tasks — meaning, when you are doing a task, keep your entire focus on that task and until it is done. Group Tasks Of Similar Mental Mode This is one of the most important things every data scientist should keep in mind while batching task successfully. Never group task with uncommon thread; meaning, always group task that are similar. For example, you can do data cleaning and visualisation at the same time. While inferring about data manually, it is good to accompany the analysis with data visualization tool, as both have a common thread and would give a clear idea of the datasets. Also, it would help you stay on the same mindset until the task is completed The main reason behind this is the fact that our brain reaches a comfortable mode when doing a particular task. And when we do a similar task again, it doesn’t lose that comfort zone. Meaning, the more similar the tasks are, the easier it is for the brain to make the transition from one task to the next. Some Added Tips Try to maintain a to-do list: It would help you keep track of the tasks you need to group for the day. Try to set a time frame for every group list: When you set a time frame for every group of task, your brain gets in a state that the task has to be done within the particular time frame. It helps you in avoiding distractions and focusing on your current task. Task batching is also a task, so make sure you give it enough time: Every time to list your tasks and group them, don’t do it in a hurry — always take time to figure out the similar tasks. Outlook It is prophesied that the domain of data science is just going to get bigger. And with that, the job role of a data scientist is going to get complex. So, in order to avoid burnout and to be more productive at work, data scientists should make sure that they follow most of the most effective life hacks.","excerpt":"Data science without a doubt one of the most rapidly growing domains in the tech industry. However, the domain is not something that anyone can work in — one needs to have tremendous skills, knowledge and staying power to deal with the daily hustle and bustle. Whether it’s about extracting the exact insight or it’s […]","categories":["AI Features"],"tags":["Data Scientist"],"author_name":"Harshajit Sarmah","publish_date":"2019-06-26T06:54:34","publication_year":"2019","word_count":789,"keywords":["data science","Go","API","programming_languages:R","AI","programming_languages:Go","RAG","ViT","Data Scientist","R"],"extracted_tech_keywords":["AI","data science","RAG","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-successful-data-scientists-use-task-batching-to-maximising-productivity\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":41410,"title":"Why TensorFlow Is The Fastest Growing Deep Learning Framework In 2019","content":"Even though other frameworks like PyTorch (developed by Facebook) has gained wide popularity, TensorFlow remains one of the most sought after deep learning frameworks of all time. Developed by researchers and engineers from the Google Brain team, it is the most commonly-used software library which holds the largest popularity on GitHub. Some of its features such as community and support, ease of use, industry relevance, embedded computer vision and others often stand out when compared to other frameworks. With TensorFlow 2.0 there has been newer and better improvements such as more straightforward APIs, streamlined Keras integration, eager execution option, among others, making it even more popular and in-demand. Other factors such as its distributed training support, scalable production deployment options and support for various devices such as Android, also contribute towards its popularity. The popularity for TensorFlow suggests that it is not going anywhere anytime soon and is going to remain popular for years to come. While there are many utilities and features because of which TensorFlow wins the game, we list a few reasons here as to why it is one of the fastest growing DL frameworks. Support for multiple languages: TensorFlow can support multiple languages to create deep learning models. Some of the languages that it supports are Python, C++, Java, Go, R. Currently, the best-supported client language is Python. Flexible architecture: One of the other reasons why it is popular is because it is designed for large-scale distributed training and inference. It is also flexible enough to support experimentation with new machine learning models and system-level optimizations. The flexible architecture of TensorFlow enables users to deploy deep learning models on one or more CPUs (as well as GPUs). Ease of use across various platforms: It can be used on platforms such as Linux, MacOS, Windows, Android. Even Keras can be used as an interface for TensorFlow. With TensorFlow 2.0, it allows robust model deployment in production on any platform. Updates and features: It has the advantage of seamless performance, quick updates & frequent new releases with new features. There is also an effort to reach out to the community. TensorFlow ensures that there is a vast amount of content, code, tutorials, and support for users to get an understanding of TensorFlow. The fact that it is accessible to everyone makes it one of the favourites. Scalability: It can be deployed on a gamut of hardware machines such as cellular devices and computers with complex setups. It can incorporate different API to built at scale deep learning architectures such as CNN or RNN. Tensorboard visualisation: TensorFlow is based on graph computation and has a great visualisation tool for training. It allows developers to visualise the construction of the neural network with Tensorboard. Tensorboard visualisation makes it very easy to visualise and spot problems. Debugging: Tensorboard is also an effective tool to debug the program. It lets the user execute the subparts of graph allowing to introduce and retrieve discrete data onto an edge, thereby offering a great debugging method. Dynamic graph capability: TensorFlow has a feature called Eager execution that allows adding the dynamic graph capability. TensorFlow allows saving the entire graph (with parameters) as a protocol buffer which can then be deployed to non-pythonic infrastructure like Java. This again makes it one of the favourable tools and is extremely easy to deploy. TensorFlow Will Keep Growing While the above-discussed pointers are some of the factors why TensorFlow has a huge following amongst the developer community, there is no doubt there are many who believe it will continue to remain one of the most used frameworks for Deep Learning. The reason that TensorFlow keeps growing is because of the steps that it takes to make itself more approachable to the developer community. It has recently undertaken moves such as open sourcing TensorFlow Lite for mobile devices and two development boards Sparkfun and Coral. Deep Learning on smartphones is something that is still new and Tensorflow somehow seems to have figured it out on how to make lighter versions for handheld devices. Moreover, TensorFlow 2.0 has put TensorFlow on the top of the game. There are many user-friendly approaches that have been introduced that makes it much more likeable than ever before. For instance, TensorFlow 1.X required users to manually stitch together the graphs by making tf.*API calls. But TensorFlow 2.0 executes eagerly, and graphs and sessions will be more like implementation details. This eliminates the use of tf.control_dependencies(), as all lines of code execute in order. If a data scientist wasn’t part of this initial stages of building a pipeline, it would be difficult for them to recover something that they never knew existed. TensorFlow 2.0 eliminates all of these mechanisms in favour of the default mechanism i.e if the user loses track of the variables; tf.Variable, it gets garbage collected. Here are the key announcements that were done around TensorFlow 2.0. Here is an article to know how to make the most of TensorFlow 2.0","excerpt":"Even though other frameworks like PyTorch (developed by Facebook) has gained wide popularity, TensorFlow remains one of the most sought after deep learning frameworks of all time. Developed by researchers and engineers from the Google Brain team, it is the most commonly-used software library which holds the largest popularity on GitHub. Some of its features […]","categories":["Global Tech"],"tags":["Tensorflow","tensorflow tutorial"],"author_name":"Srishti Deoras","publish_date":"2019-06-27T10:24:35","publication_year":"2019","word_count":828,"keywords":["machine learning","Keras","AI","neural network","PyTorch","ML","computer vision","Python","tensorflow tutorial","deep learning","TensorFlow","Tensorflow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","TensorFlow","PyTorch","Keras","Python"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-tensorflow-is-the-fastest-growing-deep-learning-framework-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10170846,"title":"NVIDIA to Launch Cheaper AI Chips for China, Reports Say","content":"NVIDIA is set to launch a new AI chip for China at a much lower price than its restricted H20 model, with mass production starting in June, sources familiar with the matter told Reuters. The sources said the GPUs will be part of NVIDIA’s most recent generation Blackwell architecture AI processors, priced between $6500 and $8000, much below the H20 model, which sold between $10,000 and $12,000. The lower price reflects its weaker specifications and simpler manufacturing requirements, which can reduce the company’s costs through simplified designs. According to the two sources, the upcoming AI chip will be built on NVIDIA’s RTX Pro 6000D, a server-grade graphics processor. It will use standard GDDR7 memory rather than the more sophisticated high-bandwidth memory (HBM). They further mentioned that it would not incorporate the advanced Chip-on-Wafer-on-Substrate (CoWoS) packaging technology from Taiwan Semiconductor Manufacturing Co (2330.TW). According to a spokesperson from NVIDIA, the company is still assessing its “limited” options. “Until we settle on a new product design and receive approval from the US government, we are effectively foreclosed from China’s $50 billion data centre market,” Reuters quoted the source. Reports show that NVIDIA’s position in the Chinese market stood at 95% four years ago, but it has significantly dropped to 50% due to the US government’s restricted regulations on exporting the company’s AI chips to China. This new AI chip leverages NVIDIA’s still-existing ground presence in the Chinese AI sector. Export restrictions have resulted in financial setbacks for NVIDIA, including a $5.5 billion write-off and around $15 billion in lost sales. Hence, this approach could be a strategic move to compensate for the losses. Reuters also reported that NVIDIA investors will look for definitive answers on how much US chip curbs on China will cost when the company releases its results on Wednesday, even as a pullback in other regulations is expected to open up new markets.","excerpt":"The company could start mass production as early as June, with the chips priced between $6500 and $8000.","categories":["AI News"],"tags":["NVIDIA chips","US sanctions"],"author_name":"Smruthi Nadig","publish_date":"2025-05-27T18:16:57","publication_year":"2025","word_count":314,"keywords":["Go","programming_languages:R","AI","NVIDIA chips","programming_languages:Go","RAG","US sanctions","R"],"extracted_tech_keywords":["AI","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-to-launch-cheaper-ai-chips-for-china-reports-say\/","complexity_score":4,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121735,"title":"HCLTech Partners With Arm To Build Custom Silicon Chips for AI Workloads","content":"HCLTech has announced a collaboration with Arm, a leading technology provider of processor IP, to augment custom silicon chips that support AI-driven business operations. The partnership will bring to market solutions that enable semiconductor manufacturers, system OEMs and cloud service providers to enhance the computing efficiency of their data centre environments and meet evolving customer demands. HCLTech will leverage pre-integrated Arm Neoverse Compute Subsystems (CSS) to help clients minimise development risks and swiftly deliver innovative, market-customised solutions geared toward improved performance and scalability for AI workloads. HCLTech has preferential access to Neoverse CSS as a member of Arm Total Design, an ecosystem that brings together industry leaders to accelerate frictionless delivery of Arm-based custom silicon chips. This access empowers HCLTech to stay at the forefront of cutting-edge technologies designed to efficiently manage AI workloads, including meeting the future demands of data centre environments. “HCLTech’s collaboration with Arm will contribute to the development of industry-leading custom AI silicon solutions that will revolutionise the way AI workloads are addressed in data centre environments. Together, we look forward to spearheading technology advancement and innovation in the semiconductor industry,” said HCLTech Engineering and R&D Services executive vice president Ameer Saithu. “Through Arm Total Design, our partners can leverage the expertise and support of other industry leaders to bring custom silicon solutions to market faster. HCLTech is a welcome addition to the ecosystem, and we are excited to see how they leverage their custom AI silicon capabilities and Arm Neoverse CSS to innovate next-generation solutions,” said Arm India president Guru Ganesan.","excerpt":"HCLTech will leverage pre-integrated Arm® Neoverse™ Compute Subsystems (CSS).","categories":["AI News"],"tags":["HCL Technology"],"author_name":"Pritam Bordoloi","publish_date":"2024-05-27T15:09:43","publication_year":"2024","word_count":256,"keywords":["programming_languages:R","AI","innovation","Scala","RAG","HCL Technology","programming_languages:Scala","GAN","R"],"extracted_tech_keywords":["AI","RAG","R","Scala","GAN","innovation","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hcltech-partners-with-arm-to-build-custom-silicon-chips-for-ai-workloads\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10051291,"title":"These Startups Are Disrupting Mental Health With AI","content":"As the COVID-19 pandemic hit the world in March 2020, it caused a significant impact on our lifestyle. Fear, worry and uncertainty made their way creeping into our lives, causing many to experience a sense of depression and anxiety. Adjusting to working from home and lacking physical contact with family and friends only worsened the situation. This led to a sharp rise in different types of mental health problems. Due to the several lockdowns imposed in the country, people could not access mental healthcare professionals in distress. Technology stepped in as a saviour here, with people using mental healthcare apps to seek comfort in such troubled times. The Alan Turing Institute said in a research that predicting mental health early and precisely has major implications for clinical management and practice, and ultimately life expectancy. In a model, it used machine learning to “predict and classify disease risk at an individual level (for dementia, anxiety, depression, and others) and to determine the interactive factors that influence mental health across people’s lifespans (Ex: genetics, cognition, demographics).” With this growth in tech, mental health apps, too, have stepped up. They are using cutting edge AI applications to cater to the needs of their customers. Here, we list down the top AI-based mental health care apps: Wysa (Touchkin) Started in: 2015 Founders: Jo Aggarwal, Ramakant Vempati Wysa takes the help of AI-based emotionally intelligent bots and uses evidence-based cognitive behavioural techniques and dialectical behaviour therapy to improve mental resilience. However, it does not provide a diagnosis or cure for mental disorders. The user can remain anonymous while accessing Wysa. The company received funding of $5.5 million in its Series A round led by Boston-based investor W Health Ventures. It is backed by the Google Assistant Investment Program, pi Ventures and Kae Capital. Ginger Started in: 2011 Founders: Anmol Madan and Karan Singh Founded at the MIT Media Lab, Ginger integrates human care with data science and augmented intelligence. Its on-demand platform for mental-health care solutions via a smartphone makes use of health coaches, therapists, and psychiatrists who collaborate to provide personalised attention to the user. Till now, the company has pumped in $220.7 million in funding over 11 rounds, with the latest funding received in March this year. Last year, it acquired another digital mental-wellness company LiveBetter’s technology assets. Quartet Health Started in: 2014 Founders: Arun Gupta Its AI-enabled system teams up partners with health plans and systems to make a personalised plan for users through virtual collaboration. AI and machine learning capabilities are used to match users with the right mental health resources. It provides both online and offline facilities with in-person sessions with therapists or via telecare. To date, the company has raised a total of $159.5 million in funding over five rounds. Last year, it teamed up with SilverCloud Health to offer clinically validated digital mental health support to its patients. Users will be able to use SilverCloud’s computerised cognitive behavioural therapy (cCBT) tools to access digital mental health services through smartphones and computers. Happify Started in: 2012 Founders: Tomer Ben-Kiki, Andy Parsons, Ofer Leidner It has built an AI coach named Anna that models real-world interactions with a therapist. Anna breaks down a user’s complex mental health path into a goal-oriented experience. Presently, it is available in 10 languages, and the company claims that it supports more than 10 chronic conditions affecting 20 million lives. Happify has raised a total of $118.7 million till now in funding over nine rounds. ION Crossover Partners and Omega Capital Partners are some of its recent backers. Woebot Started in: 2017 Founders: Alison Darcy Woebot provides a talk therapy chatbot that monitors the user’s mood with the help of NLP and psychological skills, mainly in Cognitive Behavioral Therapy (CBT). This bot asks the user how things are going in their life quickly and then stores the text and responses received. These conversations are studied, and slowly with time; the bot asks more specific questions based on past conversations. Just recently, the company closed a $90 million funding round that totalled its funding to $114 million. Spring Health Started in: 2016 Founders: Adam Chekroud, April Koh, Abhishek Chandra The company provides solutions for employee mental well-being by deploying proprietary assessment and machine-learning technology. It tries to understand a person’s condition in totality and study those results to match the right care plan for them. The user has access to a Care Navigator to help guide them through their options, helping them in scheduling appointments with therapists. Just recently, Spring Health concluded a $190 million Series D funding. Replika Started in: 2017 Founders: Eugenia Kuyda It comes with the concept of personalised AI through which one can express themselves. With the help of the AI companion, the user can express his feelings, talk about what is going on in his mind, build his personality and share other intimate details. Some users have described Replika as a non-judgmental support system. It acts more like a confidant to which a user can confide his doubts, ambitions, feelings, and dreams, and the AI companion provides a listening ear to the user’s thought process. AI in mental health is a booming field but comes with its own limitations. If one is going through something severe, apps may not be the solution here. The right track then, would be to consult a professional and seek help immediately.","excerpt":"The integration of AI and mental health can lead to path-breaking disruptions","categories":["AI Startups"],"tags":["AI for mental health","chatbot mental health","covid-19","Machine Learning","Mental Health","Startups"],"author_name":"Sreejani Bhattacharyya","publish_date":"2021-10-13T10:00:00","publication_year":"2021","word_count":896,"keywords":["data science","Go","API","machine learning","covid-19","AI","chatbot mental health","Machine Learning","Git","AI for mental health","NLP","Aim","edge AI","Mental Health","Startups","R"],"extracted_tech_keywords":["AI","machine learning","NLP","data science","Aim","edge AI","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/these-startups-are-disrupting-mental-health-with-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10095580,"title":"Council Post: Unleashing Creativity &#8211; AI And The Future Of Fashion","content":"This year New York City organised its first ever AI fashion show and it couldn’t have been more revolutionary. AI has officially sweeped into every industry there could be and it’s coming to reality way too soon because of the coming of Generative AI. Technology and innovation have always been at the forefront of the fashion business, whether it’s via the adoption of new materials to support sustainability or the development of trend-forward trends and predictions. It comes as no surprise that AI has begun to enter the fashion sector. Although the fashion industry has explored various cutting-edge technologies such as AI, NFTs, digital IDs, and augmented or virtual reality, it has yet to fully delve into generative AI. While the availability of this emerging technology has been limited until recently, there is growing potential for it to rapidly advance and significantly transform various aspects of business. Although generative AI is still in its early stages and faces certain challenges, indications suggest that it could quickly evolve and have a profound impact on the fashion industry. Generative AI goes beyond mere automation; it encompasses augmentation and acceleration. This entails equipping fashion professionals and creatives with technological tools that significantly speed up certain tasks, enabling them to allocate more time to activities that require human expertise. By assessing client preferences and data, generative AI allows the development of individualised fashion experiences. Generative AI algorithms may provide tailored clothing solutions by using consumer inputs like style preferences, body measurements, and purchase history. This degree of customization improves the purchasing experience, raises consumer happiness, and fosters a stronger sense of brand loyalty. Use cases: Redefining Fabric and Pattern Design: Fashion labels like Zara are leveraging generative AI algorithms to revolutionize fabric patterns and textures. Through emulating traditional craftsmanship, generative AI empowers designers to create visually captivating fabrics that were previously arduous to achieve manually. Revolutionizing Virtual Fitting and Customization: Generative AI is reshaping virtual fitting and customization experiences. Brands such as Adidas and Nike have harnessed generative AI-powered platforms, allowing customers to personalize and design their own shoes. This empowers individuals to choose from an array of patterns, colors, and materials, resulting in truly distinct and tailored products. Predicting Trends and Forecasting: Generative AI plays a pivotal role in trend prediction and forecasting by analyzing extensive data sets including social media posts, fashion blogs, and historical sales records. This enables fashion brands to anticipate consumer preferences and create designs that align with the latest trends, minimizing inventory risks and optimizing business strategies. Personalized Fashion Styling and Recommender Systems: Generative AI algorithms drive personalized fashion styling and recommender systems, providing tailored outfit suggestions based on individual preferences, body types, and occasions. Leading fashion e-commerce platforms like Stitch Fix employ this technology to enhance the online shopping experience, fostering customer engagement and satisfaction. Generative AI operates through the examination of extensive datasets containing preexisting designs and patterns in order to identify shared attributes and features. Once familiar with these patterns, the algorithm has the ability to generate fresh designs that integrate these elements in distinctive and inventive manners. Within the realm of fashion design, generative AI finds utility in fabricating complete garments or particular design components such as patterns, prints, and textures. Designers can input specific variables like bodily measurements, fabric preferences, and design inspirations, prompting the algorithm to produce designs that fulfil these criteria. But that’s not all. In many different ways, the fashion business has a significant influence on our world. Every step in the value chain, from fibre production through design, manufacture, distribution, consumption, and after-use disposal, has added to the sustainability challenge. According to McKinsey’s 2019 Apparel CPO Survey, a significant majority of respondents (83%) predicted a decrease in the use of physical samples in favor of virtual samples by 2025. Surprisingly, this transition has been accelerated by the COVID-19 pandemic, exceeding initial expectations. Even renowned luxury brands, known for their emphasis on customer experience and prestigious runway shows, are embracing virtual alternatives and finding innovative ways to showcase their collections. This shift towards virtual platforms has the potential to yield environmental benefits by reducing CO2 emissions, minimizing water waste, and mitigating the toxic chemical pollution associated with traditional clothing production for advertising purposes. In the past, these garments would often go to waste as they couldn’t be resold to customers. In the value chain, an often overlooked source of pollution arises from product returns and the associated delivery processes. In countries like China, where shipping is highly efficient and cost-effective, returning items has become a preferred option over not purchasing them initially, particularly when uncertain about fit. This unnecessary back-and-forth results in avoidable pollution. Moreover, with the advent of 3D virtual modeling technologies, personalized virtual fitting rooms are on the horizon. AI can assist in precise size measurements and adjust digital models accordingly. The fashion industry is on the brink of a revolution as generative AI emerges as a game-changer in design, production, and marketing. Despite the need to tackle challenges and ethical concerns, the possibilities for designers, consumers, and the environment are vast. Embracing generative AI as a catalyst for innovation and sustainability enables the fashion industry to thrive in a rapidly evolving landscape, delivering captivating and purposeful designs that resonate with the future. In 1982, Karl Largerfeld, the creative mind behind Chanel once said, “Improvise. Become more creative. Not because you have to, but because you want to. Evolution is the secret for the next step.” This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"The fashion industry is on the brink of a revolution as generative AI emerges as a game-changer in design, production, and marketing.","categories":["AI Features"],"tags":["sustainability","use cases"],"author_name":"Anirban Nandi","publish_date":"2023-06-22T11:10:31","publication_year":"2023","word_count":952,"keywords":["data science","Go","AI","sustainability","Git","use cases","RAG","Ray","Aim","analytics","generative AI","R"],"extracted_tech_keywords":["AI","data science","analytics","generative AI","Aim","Ray","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/unleashing-creativity-ai-and-the-future-of-fashion\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093621,"title":"Blame it on Altman: Microsoft Makes Risky Bet on Nuclear Fusion","content":"Last week, Microsoft made the unlikeliest bet – the tech giant has signed on the dotted line to purchase electricity from a nuclear fusion generator. The company, Helion Energy announced an agreement with Microsoft with the goal to generate at least 50 megawatts of power by 2028 with a one-year ramp-up period. Helion is liable to pay a penalty if it fails to stick to these terms. The deal will entail the world’s first commercial fusion generator connected to a power grid in Washington. Why did Microsoft sign Helion? If you’re wondering why a big-tech company has latched itself to a nascent source of energy – in November 2021, Helion raised USD 2.2 billion in funding in a round led by Sam Altman among others like Facebook Dustin Moskovitz, Peter Thiel’s Mithril Capital and sustainable tech VC, Capricorn Investment Group. Of this chunk, Altman has put in USD 375 million himself. According to The Wall Street Journal, Altman is actively involved in Helion’s operations and visits the company once a month to help execs decide what to work on efficiently as well as hiring and sifting through talent. Altman’s interest in Helion isn’t brand new. In 2014, during his tenure as the YC president, Altman expressed his intent to put more money into hard tech and clean energy companies eventually going on to accept Helion into his cohort. Observing YC’s rule to commercialise quickly, Helion then stated that its goal was to create net energy gain from nuclear fusion within an incredible span of three years. (At this point in time, experts still believed that the possibility of the nuclear fusion dream was still 30 to 50 years away) “We are optimistic that fusion energy can be an important technology to help the world transition to clean energy. Helion’s announcement supports our own long-term clean energy goals and will advance the market to establish a new, efficient method for bringing more clean energy to the grid, faster,” Brad Smith, vice chair and president at Microsoft said. Smith went on to describe the interrelation between fusion energy, AI and quantum computing. As the demands for larger amounts of energy and compute requirements for AI escalate, Smith believes fusion could be the way out in the future. “As a purchaser, when we lean in at the right moment in the right way, we can help make new markets,” he said. What is the ground reality at Helion? But for all that we know, the goal is still at a fair distance despite gaining ground. Last December, a team of physicists at the National Ignition Facility (NIF) in California claimed an important breakthrough – it had managed to generate more energy from a controlled nuclear fusion reaction than what had been used to trigger it. In a recent piece by Scientific American, Omar Hurricane, a program leader at the Lawrence Livermore National Laboratory, which houses the NIF team, spoke about the current stage that nuclear fusion stood at. “I do think fusion looks a lot more plausible now than it did 10 years ago as a future energy source. But it’s not going to be viable in the next 10 to 20 years, so we need other solutions,” he stated. How far has Helion gotten in comparison? The company aims to become the world’s first nuclear fusion plant by using a method that differs from its competition. The company claims that this cuts down the risk of radiation while increasing efficiency. This is also why other fusion power station experiments like International Thermonuclear Experimental Reactor or ITER (the largest fusion project in the world) cost more than USD 50 billion while Helion’s reactors are much more compact and run at less than tens of millions. It is also one of the faster-growing companies in this space, having created seven prototypes in less than 10 years. In July 2021, Helion announced that its sixth fusion generator prototype had exceeded 100 million degrees Celsius, which is the temperature a commercial reactor would normally operate at. But despite these pluses, Helion hasn’t wound up generating any energy in reality. It is a stretch to imagine that Helion will end up generating 50 megawatts of power within the next five years. But Altman’s push for forward-looking sectors and given what he’s done with OpenAI gives him more than some legroom. But overall, it must be noted that Microsoft’s agreement for 50 megawatts is a small and reasonable one. Since last year, Microsoft has announced several deals for 1.2 gigawatts of clean power, as reported by BloombergNEF showing a redirection towards renewable sources of energy.","excerpt":"Observing YC’s rule to commercialise quickly, Helion stated that its goal was to create net energy gain from nuclear fusion within an incredible span of three years","categories":["AI Highlights"],"tags":["AI Tool"],"author_name":"Poulomi Chatterjee","publish_date":"2023-05-19T15:08:43","publication_year":"2023","word_count":767,"keywords":["Go","API","funding","OpenAI","AI","programming_languages:R","Scala","Aim","ViT","AI Tool","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Go","Scala","API","ViT","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/blame-it-on-altman-microsoft-makes-risky-bet-on-nuclear-fusion\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":42468,"title":"Top 8 Supercomputers Used By Indian Educational Institutes For Research","content":"India is slowly becoming a superpower in computers. According to latest ranking, India is ranked 21st globally when it comes to applications and developments in supercomputers. India is definitely progressing towards being one of the biggest players in the field and this month, 10 institutes from Gujarat announced that they are planning to equip themselves with 10 more supercomputers this year. Let’s take a look at the supercomputers that are already being used by Indian educational institutes. Following is a list of supercomputers by Indian institutes in alphabetical order: 1. Colour Boson (Cray XC-30) Colour Boson is used by the TIFR, Mumbai. A Cray product, the model name of this supercomputer is Cray XC-30. It has a total storage of 1.1 PB. In this supercomputer, the nodes are interconnected by Cray’s Aries interconnect routing technology with DragonFly topology. Linux environment is used as the OS for the system and is currently being used under Indian Lattice Gauge Theory Initiative program by the scientists of TIFR. Purpose: Research of quark-boson, a phase of matter that holds the mystery of the creation of our universe. It is also being used in developing theoretical physics and quantum chromodynamics and QCD. Computing speed: 558.7 TFlops 2. HPC IIT Delhi has the supercomputer in their campus called the HPC, which is a GPU-centric high-performance system and is one of the fewest in the world. It was developed by NVIDIA in conjunction with IIT Delhi team to built this system that is based on GPU Tesla Platform. It has a total storage of the system is 1.5 PB. Purpose: Researchers in fields of biology, nanosystems, atmospheric science and bioinformatics use this system to study and research in areas such as Deep Learning, Computational Physics, Chemistry, Computational Fluid Dynamics and Material Science. Computing speed: 860 Tflops 3. PARAM Ishan Inaugurated in the year 2016, PARAM Ishan is a product of the Indian Institute of Technology, Guwahati (IITG). It is made by a joint collaboration of CDAC and IITG. Purpose: R&D of computational chemistry, computational fluid dynamics, computational electromagnetics, civil engineering structures, nano-block self assemble, climate modelling and seismic data processing. Computing speed: 121 Tflops 4. PARAM Kanchenjunga PARAM Kanchenjunga is a PARAM series supercomputer. It was launched at NIT, Sikkim campus. It is a joint collaboration between the Pune-based CDAC and the Ministry of Communications and IT. It has the latest Intel processors and accelerator technologies. It is currently being deployed by two faculty members of NIT Sikkim in studying how to increase the efficiency of fuel movement through long oil pipelines such as from the Middle East to India. The supercomputer is established at a cost of Rs. 3 crore. Purpose: Help achieve excellence in engineering education and research in the north-east region Computing speed: 15 Tflops 5. PARAM SHIVAY Inaugurated in February 2019, PARAM Shivay is a part of the National Supercomputing Mission. The supercomputer is built by IIT-BHU, Varanasi. It is equipped with the latest Intel-based processors and high memory computer nodes. There are three supercomputing machines designed, manufactured and assembled at IIT Bhu, IISER Pune and IISER Kharagpur. This project is jointly implemented by DST and Ministry of Electronics and Information Technology (MeitY) and led by C-DAC and IISc. It supports the government’s vision of ‘Digital India’ and ‘Make in India’. Purpose: Help in weather services, disaster simulation and management, help faster processing of seismic data and help in computational biology. Computing Speed: 38.1 Tflops 6. PARAM YUVA II Made by the Centre for Development of Advanced Computing (C-DAC), PARAM Yuva II is a supercomputer inaugurated in the year 2013 and was made in a period of only three months at a cost of ₹160 million\/₹15 crores. Purpose: It helps in carrying out research in space, seismic data analysis, aeronautical engineering, scientific data processing and pharmaceutical development, bioinformatics. Educational institutes can be linked to the computer through the national knowledge network. Computing Speed: 524 Tflops 7. Sahasrat (CRAY XC40) SahasraT supercomputer is located at Supercomputer Education and Research Centre (SERC), a facility at Indian Institute of Science (IISc). This Cray XC40 combines the capabilities of Intel’s latest Xeon Haswell processors for the CPU cluster and Nvidia’s K40 series of GPU cards. It consists of Intel Haswell Xeon E5-2680v3 processors, NVIDIA K40 GPU accelerators and Intel Xeon Phi 5120D coprocessors and storage of 2.1 PetaBytes. It has 1,500 processors and coprocessors along with 44 GPUs to handle complex tasks in the system. Sahasrat has been rated 901.54 TFLOPS, which the highest rating amongst all supercomputers in India. It used a cost ₹82.70 crores to build it. Purpose: Aerospace engineering, meteorology predictions and astrological simulations. It is also used for material research and mapping the entire climate condition of the particular region via simulation Computing Speed: 1.46 petaflops 8. Virgo Virgo is a supercomputer built for IIT, Madras (IITM). It has 292 compute nodes, 2 master nodes, 4 storage nodes and has total computing power. It is claimed to be the fastest cluster in an academic institution in India. In terms of performance, it has an Expand (Rmax) of 91.126 TF and Expands (RPeak) of 97.843 TF. It has architectures designed to handle different types of computing problems based on what is needed. The High-Performance Computing Environment (HPCE) is set up to cater to the ever-increasing demand for supercomputing facilities, of researchers at IIT Madras. Purpose: Research in IIT-M in the fields of material science and engineering, atmospheric and ocean modelling, aerospace engineering, social, ecological and physical networks, design of large structures and VLSI, understanding flows and combustion, spectroscopy and molecular modelling. Computing power: 97 TFlops","excerpt":"India is slowly becoming a superpower in computers. According to latest ranking, India is ranked 21st globally when it comes to applications and developments in supercomputers. India is definitely progressing towards being one of the biggest players in the field and this month, 10 institutes from Gujarat announced that they are planning to equip themselves […]","categories":["AI Trends"],"tags":["hpc data management system"],"author_name":"Disha Misal","publish_date":"2019-07-15T15:00:13","publication_year":"2019","word_count":931,"keywords":["Go","hpc data management system","programming_languages:R","AI","programming_languages:Go","Git","RAG","Ray","Aim","deep learning","R"],"extracted_tech_keywords":["AI","deep learning","Aim","Ray","RAG","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-supercomputers-used-by-indian-educational-institutes-for-their-research\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":2519,"title":"International Conference on Advanced Data Analysis, Business Analytics and Intelligence@IIMA 2013","content":"Indian Institute of Management Ahmedabad has announced the 3rd international conference dedicated to advanced data analysis, business analytics and business intelligence to be held in 2013 April. The objectives of the conference are to facilitate sharing of: Research based knowledge related to advanced data analysis, business analytics and business intelligence among academicians and practitioners Case studies and novel business applications of tools and techniques of advanced data analysis, business analytics and business intelligence among academicians and practitioners. Papers are invited from academicians and practitioners on any topic mentioned in the list of conference topics and related areas. Applications, case studies, review and discussion papers on these topics and related areas are also welcome. The various topics to be covered under the theory, methods and applications are Exploratory Data Analysis, Classification, Operations Research, Cluster Analysis, Regression Modeling, Probability and Stochastic Processes, Data Visualization,  Pattern Recognition,Time Series Analysis, Machine Learning, Forecasting, Bayesian Methods, Computational Intelligence, Panel Data, Multivariate Analysis,     Marketing Models, Internet Modeling and Web Analytics, Statistics in Finance, Marketing Research, Text Mining, Insurance Models, Advertising andMedia, Revenue Management,Investment and Portfolio Models,  Data Analysis in Retailing, Bioinformatics, Data Analysis in Banking and Financial Services, CRM, Health Sciences, Risk Analytics, Pricing Analytics, Industrial Applications, Legal Analytics, Analytics for Strategy,  Supply Chain Management, Analytics for Public Policy,  Quality Management. The following Criteria has been chosen for evaluation: Relevance: Does the subject of the paper appeal to the interests of the conference attendees? Methodology: Does the paper use sound and appropriate method(s)? Originality: Does the paper add new findings, insights, or knowledge to the body of literature? Research: Does the paper compare and weigh the material against the work of others? Conclusions: Are the conclusions sound and justified? Managerial Implications: Is the managerial relevance and implications of the decision problem demonstrated? References: Are the references adequate? To register visit http:\/\/www.iimahd.ernet.in\/icadabai2013\/","excerpt":"Indian Institute of Management Ahmedabad has announced the 3rd international conference dedicated to advanced data analysis, business analytics and business intelligence to be held in 2013 April. The objectives of the conference are to facilitate sharing of: Research based knowledge related to advanced data analysis, business analytics and business intelligence among academicians and practitioners Case […]","categories":["Deep Tech"],"tags":[],"author_name":"Дарья","publish_date":"2013-01-25T17:06:02","publication_year":"2013","word_count":305,"keywords":["business intelligence","machine learning","programming_languages:R","AI","ViT","analytics","R"],"extracted_tech_keywords":["AI","machine learning","analytics","R","ViT","business intelligence","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/international-conference-on-advanced-data-analysis-business-analytics-and-intelligenceiima-2013\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10142071,"title":"&#8216;To Build an AI Startup in India, You Should Have a PhD,&#8217; says Yann LeCun","content":"Meta’s chief AI Scientist, Yann LeCun, believes that aspiring AI entrepreneurs in India should get an academic degree like a PhD or Masters, particularly in fields as technical and complex as artificial intelligence before building a startup. In a recent interview with Zerodha cofounder Nikhil Kamath, LeCun, speaking about the importance of formal training in innovation, said, “Doing a PhD or graduate studies trains you to invent new things and ensures that your methodology prevents you from fooling yourself into thinking you’re being an innovator when you’re not.” LeCun added that while a PhD is not a strict requirement for success, it offers significant advantages for entrepreneurs. “It gives you a different perspective,” he said. “In a complex, deeply technical area like AI, it’s useful to learn about what exists out there, what’s possible, and what’s not possible.” For those considering starting their own AI ventures, LeCun said, “You might succeed without it, but it gives you more legitimacy in hiring talented people, and it might make it easier to raise funding if you have published papers showing you’ve invented something new.” He further advised young entrepreneurs and startups to experiment with open-source models like Meta’s Llama and said that companies can fine-tune the model for particular verticals such as healthcare and education. During his recent visit to India, where he toured IIT Madras and AI4Bharat, he was impressed by projects focusing on language translation and cultural preservation, such as IndicTrans2. “AI is going to help inspire,” he said, particularly in regions with diverse languages and cultural complexities.” While speaking on a panel at Meta’s Build with AI Summit in Bengaluru alongside Infosys co-founder Nandan Nilekani and People+AI head Tanuj Bhojwani, LeCun discussed the idea of positioning India as the AI use-case capital of the world. Nilekani added, “We think that we can build on top of digital public infrastructure,” adding that this foundation enables a faster transition to AI—a concept he previously described as “DPI to the power of AI.”","excerpt":"“It gives you a different perspective.”","categories":["AI News"],"tags":["Meta AI","Yann LeCun"],"author_name":"Siddharth Jindal","publish_date":"2024-11-30T11:39:33","publication_year":"2024","word_count":331,"keywords":["Go","API","Meta AI","artificial intelligence","Yann LeCun","funding","AI","innovation","Git","llm_models:Llama","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","API","innovation","startup","funding","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/to-build-an-ai-startup-in-india-you-should-have-a-phd-says-yann-lecun\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10096343,"title":"Elon Musk vs Mark Zuckerberg: The Real Tech Battle Begins","content":"Meta chief Mark Zuckerberg recently took inspiration from Twitter boss Elon Musk’s playbook and unveiled paid subscriptions for verified accounts. But will he now take a leap towards safeguarding valuable data on this fresh platform? Meta recently announced that it will be launching its Instagram Threads pretty soon. This comes just a few days after Musk announced a limit on the number of posts a user can read per day. The explanation that Musk gave on putting a cap on viewing was “address extreme levels of data scraping and system manipulation”. There is a speculation that Musk might plan to build its own ChatGPT rival using Twitter data – which some are referring to as TruthGPT. So, the question is, will Zuckerberg follow suit to limit the number of post users can view on Thread to save the valuable data? Recently, tech giants have aspired to create their own generative AI chatbot similar to Open AI’s ChatGPT. As they say, data is gold, and for any company that aspires to create any form of chatbot, it needs data to train its LLM. What we need is TruthGPT— Kekius Maximus (@elonmusk) February 17, 2023 Currently, Google probably has the highest amount of data as compared to any other competitor out there. Google DeepMind chief Demis Hassabis, recently claimed that the company’s next LLM project is going to eclipse ChatGPT. This huge claim is about the project called Gemini. Interestingly, DeepMind, with Google, has something that no one else has – YouTube. While Twitter is filled with textual data, YouTube is a gold mine for visual, audio, and textual data in almost every single language on Earth. What About Meta? One fascinating fact about Meta is that it does not use the data it collects over its applications to train its LLM LLaMa. According to the research paper “LLaMA: Open and Efficient Foundation Language Models”, LLaMA is trained on CommonCrawl, GitHub, Wikipedia, and books. However, there is a possibility that in the future, Meta might train its LLM models (LLaMA and beyond) from Facebook and Instagram data. Again, there is also a possibility that they won’t have to do that since they are betting big on self-supervised learning (SSL) and world models, which limits the use of RLHF (reinforcement learning with human feedback), and reduces the reliance on training models on user data. Recently, Meta collaborated with Carnegie Mellon University, and the University of Southern California to make LIMA. Unlike ChatGPT and Bard, where RLHF is considered crucial, this model may be using self-supervised learning, which has been strongly advocated by Yann LeCun for a long time. Meta Bets Big on Open AI Instead of making LLaMA, the chatbot, available to the general public, Meta chose a different approach. It released it as an open-source package, which means that members of the AI community can request access to it. Whether the future of AI is open or closed source in itself is a different debate, but by releasing LLaMA, Meta took a stand of democratising access to large language models, helping researchers advance their work in this subfield of AI. One of the possibilities of why Meta is backing open source is that it is nowhere near Google or OpenAI. It gives them the opportunity to correct the flaws and loopholes in LLaMA and keep an eye on how the community is bettering it. In the recent podcast with Lex Fridman, Zuckerberg said, “The stage we are in right now, the equities balance strongly in my view towards doing this more openly.” Zuckerberg has shifted his focus to generative AI. He said that Meta will bring LLM-powered AI agents to Messenger and WhatsApp first, but explore additional opportunities across its family of apps, consumer products and into the metaverse. Only time will tell if Meta will venture into its own chatbot. Speaking of generative AI in a Facebook post, Zuckerberg said, “We have a lot of foundational work to do before getting to the really futuristic experiences, but I’m excited about all of the new things we’ll build along the way.” It will be interesting to see how Meta leverages Threads’ data beyond selling it to advertisers. Only time will tell whether the company chooses to restrict data scraping. Threads has been positioned as an Instagram app, but there’s a strong chance Meta may follow a path similar to Twitter’s, setting up a direct challenge to Musk in a real tech battle.","excerpt":"Meta recently announced that it will be launching Instagram Threads pretty soon. This comes a few days after Musk announced a limit on the number of posts a user can read per day","categories":["Deep Tech"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-07-05T12:59:16","publication_year":"2023","word_count":740,"keywords":["Go","ChatGPT","RLHF","OpenAI","AI","AWS","RAG","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Aim","RAG","RLHF","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/elon-musk-vs-mark-zukerberg-the-real-tech-battle-begins\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10015031,"title":"Detectron2: Guide To Next-Generation Object Detection","content":"Object detection is a tedious job, and if you ever tried to build a custom object detector for your research there are many factors architectures we have to think about, we have to consider our model architecture like FPN(feature pyramid network) with region purposed network, and on opting for region proposal methods we have Faster R-CNN, or we can use more of one-shot techniques like SSD(single shot detector) and YOLO(you only look once). Now in all of this continuous competition of making object detection models better and efficient, the Facebook AI team has launched many cutting edge detectors, models, frameworks, and datasets over the years. But still, it never came out of controversies easily there are tweets and negative images that are going on the internet towards facebook AI systems. Facebook sucks— Elon Musk (@elonmusk) May 14, 2020 In 2018, Facebook AI Research (FAIR) published a new object detection algorithm called Detectron. It was a great library that implements state-of-art object detection, including Mask R-CNN. It was written in Python and Caffe2 deep learning framework. Due to Detectron, there were many research projects published later like Feature pyramid network(FPN), Data Distillation, Omni-Supervised Learning, and Mask R-CNN. Detectron backbone network framework was based on: ResNet(50, 101, 152)ResNeXt(50, 101, 152)FPN(Feature Pyramid Networks) with Resnet\/ResNeXtVGG16 The goal of detectron was pretty simple to provide a high- performance codebase for object detection, but there were many difficulties like it was very hard to use since it’s using caffe2 & Pytorch combined and it was becoming difficult to install. And that’s why FAIR came up with the new version of Detectron. Detectron2 “Detectron2 is Facebook AI Research’s next-generation software system that implements state-of-the-art object detection algorithms” – Github Detectron2 Detectron2 is built using Pytorch, which has a very active community and continuous up-gradation & bug fixes. This time Facebook AI research team really listened to issues and provided very easy setup instructions for installations. They also provided a very easy API to extract scoring results. Other Frameworks like YOLO have an obscure format of their scoring results which are delivered in multidimensional array objects. YOLO takes more effort to parse the scoring results and inference it in the right place. Detectron2 got pretty massive trending on the internet since its release: Detectron2 originates from Mask R-CNN benchmark, and Some of the new features of detectron2 comes with are as follows: This time it is Powered by Pytorch deep learning framework.Panoptic segmentationInclude DenseposeProvide a wide set of baseline results and trained models for download in the Detectron2 ModelZoo.Included projects like DeepLab, TensorMask, PointRend, and more.Can be used as a wrapper on top of other projects.Exported to easily accessible formats like caffe2 and torchscript.Flexible and fast training on single or multiple GPU servers. There is also a new model launched with detectron2, i.e. Detectron2go, which is made by adding an additional software layer, Dtectron2go makes it easier to deploy advanced new models to production. Some of the other features of detectron2go are: Standard training workflows with-in-house datasetsNetwork quantizationModel conversion to optimized formats for deployment to mobile devices and cloud. Installation We are going to use Google Colab for this tutorial. You can find the installation guide here. Also, there is a Dockerfile available for easier installation. Requirements Operating System: Linux or macOSPython: 3.6+Pytorch: 1.5+ & torchvison that matches the Pytorch installation. You can install both together at pytorch.orgOpenCV for Visualization Getting Started: We are going to use the official Google Colab tutorial from Detectron2. Installing dependencies (pyyaml) !pip install pyyaml==5.1 import torch, torchvision Install Detectron2 and restart your runtime after executing below command: import torch assert torch.__version__.startswith(\"1.7\") !pip install detectron2 -f https:\/\/dl.fbaipublicfiles.com\/detectron2\/wheels\/cu101\/torch1.7\/index.html Setup Detectron2 logger import detectron2 from detectron2.utils.logger import setup_logger setup_logger() Import additional libraries import numpy as np import os, json, cv2, random from google.colab.patches import cv2_imshow Import detectron2 utilites for easy execution from detectron2 import model_zoo from detectron2.engine import DefaultPredictor from detectron2.config import get_cfg from detectron2.utils.visualizer import Visualizer from detectron2.data import MetadataCatalog, DatasetCatalog Run a detectron2 model trained on COCO dataset !wget http:\/\/images.cocodataset.org\/test-stuff2017\/000000017581.jpg -q -O input.jpg im = cv2.imread(\".\/input.jpg\") cv2_imshow(im) Create a detectron2 configuration and a DefaultPredictor to run inference on input image cfg = get_cfg() # add project-specific config (e.g., TensorMask) here if you're not running a model in detectron2's core library cfg.merge_from_file(model_zoo.get_config_file(\"COCO-InstanceSegmentation\/mask_rcnn_R_50_FPN_3x.yaml\")) cfg.MODEL.ROI_HEADS.SCORE_THRESH_TEST = 0.5 # set threshold for this model # Find a model from detectron2's model zoo. You can use the https:\/\/dl.fbaipublicfiles... url as well cfg.MODEL.WEIGHTS = model_zoo.get_checkpoint_url(\"COCO-InstanceSegmentation\/mask_rcnn_R_50_FPN_3x.yaml\") predictor = DefaultPredictor(cfg) outputs = predictor(im) Print predicted output print(outputs[\"instances\"].pred_classes) print(outputs[\"instances\"].pred_boxes) Visualize the predicted output using Visulizer utility by Detectron2 output = Visualizer(im[:, :, ::-1], MetadataCatalog.get(cfg.DATASETS.TRAIN[0]), scale=1.2) out = output1.draw_instance_predictions(outputs[\"instances\"].to(\"cpu\")) cv2_imshow(out.get_image()[:, :, ::-1]) For common installation, error refer here Conclusion FAIR team has gone pretty straight this time by open-sourcing everything, this was a good move as the team believes that they can’t achieve the state of the art algorithms and techniques in isolation so open source is the solution to a better AI era. FAIR has done many interesting projects like Multimodal hate speech Memes challenges: Facebook AI research has included many projects that are made by using Detectron2 like: DeepLabDensePosePanoptic-DeepLabPointRendTensorMaskTridentNet Some of the external projects that use detectron2: AdelaiDetCenterMaskRes2Net backbonesVoVNet backbonesFsDet","excerpt":"Object detection is a tedious job, and if you ever tried to build a custom object detector for your research there are many factors architectures we have to think about, we have to consider our model architecture like FPN(feature pyramid network) with region purposed network, and on opting for region proposal methods we have Faster […]","categories":["AI Trends"],"tags":["Facebook AI","Facebook AI research","Object Detection"],"author_name":"Mohit Maithani","publish_date":"2020-12-21T14:00:00","publication_year":"2020","word_count":860,"keywords":["NumPy","Facebook AI","AI","PyTorch","ML","docker","OpenCV","Colab","Ray","deep learning","Object Detection","object detection","Facebook AI research"],"extracted_tech_keywords":["AI","ML","deep learning","Ray","PyTorch","Colab","OpenCV","NumPy","object detection","docker"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/detectron2\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10143191,"title":"Cloud-Based Unified Testing Platform LambdaTest Raises $38M","content":"San Francisco-based LambdaTest, a cloud-based unified testing platform, has raised $38 million in its latest funding round, bringing its total investment to $108 million. LambdaTest was co-founded by Asad Khan and Jay Singh in 2017. The round was led by Avataar Ventures, with participation from Qualcomm Ventures, among others. The funding will fuel the company’s efforts to revolutionise quality assurance (QA) through AI-driven solutions. Founded in 2017, LambdaTest has raised nearly $70 million in previous rounds, attracting top investors, including Premji Invest, Sequoia Capital India, Titanium Ventures, Leo Capital Holdings, and Blume Ventures. Recently, the startup unveiled an end-to-end AI Test Agent. KaneAI represents a different approach to software testing, enabling teams to author, debug, and evolve tests using natural language. It is also dubbed as an “AI Native QA Agent-as-a-Service.” The platform supports popular tools such as Jira, Slack, GitHub Actions, and Microsoft Teams, making it an invaluable addition to any testing toolkit. KaneAI The newly launched KaneAI platform claims to leverage AI to automate QA processes and reduce manual test script generation by up to 70%. It addresses the increasing complexity of software development, where faster release cycles and rigorous testing demands require innovative approaches. Avataar Ventures expressed confidence in LambdaTest’s ability to reshape the future of software testing, noting its disruptive approach with solutions like KaneAI and HyperExecute, which offer end-to-end test orchestration with enhanced speed and security. Qualcomm Ventures also emphasised the transformative potential of LambdaTest’s product suite, focusing on how its AI-driven testing systems are essential for improving CI\/CD pipeline efficiency and accelerating release cycles. The company’s comprehensive approach positions it as a leader in the rapidly evolving software quality assurance space. LambdaTest’s AI-powered HyperExecute platform enhances the testing experience by running tests up to 70% faster than traditional cloud grids and accelerating test resolution by 2.5x, all while improving error detection by 60%. LambdaTest has been serving over 15,000 customers, including Fortune 500 companies, and facilitating more than 1.2 billion tests globally. The startup has projected a year-on-year growth of 105%, supporting more than 2.3 million developers and testers worldwide.","excerpt":"With this investment, the startup, which recently launched the AI agent KaneAI, has brought its total funding to $108 million.","categories":["AI News"],"tags":["Cloud","KaneAI","LambdaTest","Qualcomm Ventures","Unified software testing"],"author_name":"Vandana Nair","publish_date":"2024-12-10T17:50:57","publication_year":"2024","word_count":346,"keywords":["Go","API","AI","LambdaTest","Cloud","R","CI\/CD","Git","RAG","Unified software testing","Aim","KaneAI","GitHub","Qualcomm Ventures","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","GitHub","CI\/CD","API","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cloud-based-unified-testing-platform-lambdatest-raises-38m\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10163286,"title":"US GCC Centres Lead in India, Non-US Firms Catching Up","content":"India’s global capability centre (GCC) ecosystem is expanding rapidly, set to grow from 1,700 centres in 2024 to 2,100 by 2030. While US-based GCCs continue to dominate, non-US GCCs are growing at a CAGR of 6.8%, nearly twice the growth rate of US firms. Karnataka: The Hub of US-Based GCCs Karnataka remains the leading destination for US-headquartered GCCs in India, hosting over 50% of them. According to a Zinnov-AmCham GCC report, two-thirds of the world’s most admired US companies present in India have set up centres in Karnataka. The key industries among them are BFSI (banking, financial services, and insurance) and software & internet. Some of the prominent US-based GCCs in Bengaluru are Lowe’s, Kimberly-Clark, and Wells Fargo, to name a few. Rise of Non-US GCCs in India A growing number of non-US companies are also setting up GCCs in India. “The anticipated 15-20% surge in non-American GCCs in India over the next two years highlights the country’s escalating strategic importance in the global business landscape,” said Alouk Kumar, founder and CEO at Inductus Limited. Notably, GCCs from the Asia-Pacific region have experienced a growth exceeding 65% in the past five years, with those from Europe, the Middle East, and Africa growing by around 35-40%. This trend is driven by enterprises from Europe, the Asia-Pacific, and the Middle East diversifying their global operations beyond traditional Western hubs. Experts predict this growth trajectory will continue over the next 2-3 years. Kumar further highlighted that “India’s cost-effective, highly skilled talent pool, robust digital infrastructure, and evolving policy framework make it an attractive destination. “The emergence of Tier-II cities such as Ahmedabad, Kochi, Thiruvananthapuram, and Coimbatore as viable alternatives to metropolitan areas further strengthens India’s position as a GCC powerhouse.” For instance, Takeda, a Japanese biopharmaceutical giant, recently launched its innovation capability centre (ICC) in Asia in Bengaluru. The centre will use AI and digital technologies to drive research and improve healthcare. “The launch of our Bengaluru ICC is a pivotal moment in Takeda’s digital transformation journey,” said chief data and technology officer Gabriele Ricci. In November 2023, ZEISS, a German optics and optoelectronics leader, opened its GCC in Bengaluru. The centre will tap into India’s IT talent pool to expand ZEISS’s global R&D efforts. Mizuho Global Services India Pvt Ltd (MGS), a subsidiary of Japan’s Mizuho Bank, has opened a new office at the World Trade Center in Perungudi, Chennai. Since its founding in 2020, MGS has expanded rapidly with offices in Mumbai and Chennai. Currently employing 350 people, the company plans to more than triple its workforce to over 1,000 in the next three years. The Reason Behind Such Expansion Experts believe that Europe’s reliance on India for technology operations is increasing, particularly in regions such as France, the Nordics, and the Benelux (Belgium, Netherlands, and Luxembourg). This trend is driven by rapid modernisation, a shortage of local tech talent and the need for multilingual support across industries like manufacturing (Industry 4.0), automotive, BFSI, pharma, and retail. “Traditionally, Europe used to have the nearshore captive centre, which gave them timing benefits,” said the senior director leading GCC in Accenture, “due to war and instability in Europe, mostly in Ukraine and Poland, many financial GCC moved to India. The war has a huge impact.” The robust GCC ecosystem in India thrives on its skilled workforce, supportive government policies, innovative startups, and a strong network of service providers. “Factors such as India’s competitive GCC Policies mushrooming out of several states and the ability to achieve operational efficiencies at scale make it the ideal choice for global firms seeking long-term resilience in their corporate strategies, ”Kumar further mentioned. However, what truly sets India apart in this next phase of the GCC’s evolution is its ability to create value rather than just a cost arbitrage destination. Non-American GCCs are leveraging India not just for talent and operations but for innovation—whether it’s Swiss financial firms setting up AI-powered risk analysis hubs in Bengaluru or Middle Eastern energy giants driving sustainability R&D from Hyderabad. With India’s deep integration into global supply chains and its thriving startup ecosystem valued at over $400 billion, GCCs are increasingly tapping into co-innovation models with local firms and digital twin models with their parent companies, unlocking synergies that go far beyond basic outsourcing and ODC. As India scales from a service delivery hub to a strategic innovation partner, the future of non-American GCCs in the country is not just about growth—it’s about leadership in global transformation. Seeking Inspiration from US-Based GCCs Raghavendra Vaidya, MD and CEO at Daimler Truck Innovation Center India (DTICI), highlighted the growing trend of companies setting up engineering centres in India. He noted that while “the initial push came from US-based conglomerates,” in recent times, “many Western and Central European companies have been establishing their presence here.” He emphasised that the process has become highly structured: “The playbook has been perfected over the last few decades… it is almost a cookie-cutter approach if you have a real intention of doing it.” Companies today can easily reference existing models and successfully establish their own centres. A key driver behind this trend is the transformation happening in Western and Central Europe, where companies are realising that “they can’t do all of their engineering by themselves in the headquarters because there are so many problems to be solved.” Observing their peers successfully setting up engineering centres in India, they recognise the value such centres bring to their parent organisations. While IT services are relatively easy to establish, “setting up an engineering centre is difficult.” However, India remains “the talent market in the world for ER&D,” and the proven success of GCCs in driving innovation and value has instilled confidence among organisations looking to follow suit.","excerpt":"While US-based GCCs continue to dominate, non-US GCCs are growing at a CAGR of 6.8%, nearly twice the growth rate of US firms.","categories":["GCC"],"tags":["GCC india"],"author_name":"Shalini Mondal","publish_date":"2025-02-12T11:01:38","publication_year":"2025","word_count":954,"keywords":["Go","API","AI","ML","Scala","Git","RAG","Aim","GCC india","GAN","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Scala","Git","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/gcc\/us-gcc-centres-lead-non-us-firms-catching-up\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10039248,"title":"An “EPIC” Way To Evaluate Reward Functions In Reinforcement Learning","content":"“Specifying a reward function can be one of the trickiest parts of applying RL to a problem.”DeepMind At the heart of a successful reinforcement learning algorithm sits a well-coded reward function. Reward functions for most real-world tasks are difficult to specify procedurally. Most real-world tasks have complex reward functions. In particular, tasks involving human interaction depend on complex and user-dependent preferences. A popular belief within the RL community is that it is usually easier and more robust to specify a reward function, rather than a policy maximising that reward function. Today there are many techniques to learn a reward function from data as varied as the initial state, demonstrations, corrections, preference comparisons, and many other data sources. A group of researchers from DeepMind, Berkeley and OpenAI have introduced EPIC, a new way to evaluate reward functions and reward learning algorithms. EPIC overview Source: DeepMind Safety Research RL training is computationally expensive. For instance, if the policy performs poorly, you can’t tell if it is due to the learned reward failing to match user preferences or the RL algorithm failing to optimise the learned reward. According to the researchers, Equivalent-Policy Invariant Comparison or (EPIC) works by comparing reward functions directly, without training a policy. EPIC is a fast and reliable way to compute the similarity of two reward functions work. EPIC can be used to benchmark reinforcement learning algorithms by comparing learned reward functions to a ground-truth reward. In a paper titled “Quantifying difference in reward functions” (recently accepted at the prestigious ICLR conference), the researchers claimed EPIC resulted in 1,000 times faster solution than alternative evaluation methods. Furthermore, it requires little to no hyperparameter tuning. The researchers showed reward functions judged as similar by EPIC induce policies with similar returns, even in unseen environments. How EPIC works: As shown in the illustration above, EPIC compares reward functions Rᵤ and Rᵥ by first mapping them to canonical representatives. It then computes the Pearson distance between the canonical representatives on a coverage distribution ????. Note: Pearson distance between two random variables X and Y is calculated as follows: EPIC distance is defined using Pearson distance as follows: Where, D: distance, R: rewards, S: current state, A: action performed, S1 : changed state The distance calculated using this approach can then be used to predict the outcome of using a certain reward function. According to the researchers, canonicalisation removes the effect of potential shaping, and Pearson distance is invariant to positive affine transformations. RL agents strive to maximise reward. This is great as long as the high reward is found only in states that the user finds desirable. However, in systems employed in the real world, there may be undesired shortcuts to high reward involving the agent tampering with the process that determines agent reward, the reward function. For instance, a self-driving car; a positive reward may be given once it reaches the correct destination and negative rewards for breaking traffic rules and causing accidents. Standardised metrics are an important driver of progress in machine learning. Unfortunately, traditional policy-based metrics do not guarantee the fidelity of the learned reward function. A reward is non-zero only at the end, where it is either −1, 0, or 1, depending on who won. In any practically implemented system, agent reward may not coincide with user utility. To evaluate EPIC, the researchers developed two alternatives as baselines: Episode Return Correlation (ERC) and Nearest Point in Equivalence Class (NPEC). On comparing procedurally specified reward functions in four tasks, the researchers found EPIC is more reliable than the baselines NPEC and ERC, and more computationally efficient than NPEC. The experimental results showed  EPIC correctly infers zero distance between equivalent reward functions that the NPEC and ERC baselines wrongly considered dissimilar. Source: DeepMind Reinforcement learning never got the attention it deserved; the main reason being its areas of application. Unlike a typical convolutional neural network, used for photo tagging on social media, RL’s use cases — self driving, robotics for medical surgeries etc — are more critical. This make reward function evaluation even more important. There can’t be enough stress tests for an RL algorithm given the uncertain nature of the real world. But, the evaluation of reward functions is a good place to start. As RL is increasingly applied to complex and user-facing applications such as recommender systems, chatbots and autonomous vehicles, reward functions evaluation will need more attention. Since there are various techniques to specify a reward function, the researchers believe EPIC can play a crucial role here. Key Takeaways Current reward learning algorithms have considerable limitationsThe distance between reward functions is a highly informative addition for evaluationEPIC distance compares reward functions directly, without training a policy.EPIC is fast, reliable and can predict return even in unseen deployment environments. EPIC is now available as a library. Check the Github repo.","excerpt":"“Specifying a reward function can be one of the trickiest parts of applying RL to a problem.” DeepMind At the heart of a successful reinforcement learning algorithm sits a well-coded reward function. Reward functions for most real-world tasks are difficult to specify procedurally. Most real-world tasks have complex reward functions. In particular, tasks involving human […]","categories":["AI Trends"],"tags":[],"author_name":"Ram Sagar","publish_date":"2021-04-30T14:00:00","publication_year":"2021","word_count":802,"keywords":["Go","machine learning","OpenAI","AI","neural network","chatbots","Git","RAG","Aim","R"],"extracted_tech_keywords":["AI","machine learning","neural network","OpenAI","Aim","RAG","chatbots","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/epic-reward-function-evaluation-reinforcement-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10045587,"title":"How Can Bug Bounty Programs Combat AI Biases?","content":"This month Twitter announced its first-ever artificial intelligence bug bounty program. The challenge is to find bias in its image cropping algorithm. Twitter removed the image cropping algorithm from public use after small instances of bias were uncovered in research by the META team earlier this year. The selection of the winners will be based on a rubric created by Twitter’s Machine Learning Ethics, Transparency and Accountability or META team. The first-place prize, which will be announced at this year’s DEF CON AI Village, will be $3,500. Faulty AI systems AI-based companies usually turn to the black box and avoid critical analysis of the results obtained from their models. This has led stakeholders to be concerned about the absence of transparency and the inability of organisations to communicate the significant factors for the delivery of their AI-based products. Bug bounties, which reward hackers for discovering vulnerabilities in software code before malicious actors exploit them, have become essential to the security field. Such bug bounty programs can be helpful for companies when it comes to evaluating their outputs based on explainability. Thus, Bug bounty programs assist in ensuring the company’s trust by finding loopholes but also help in evaluating the in-house cybersecurity department. Bug bounties for AI projects Last year, researchers from Google Brain, Intel, OpenAI, and top research labs in the U.S. and Europe joined forces to formulate a toolbox for turning AI ethics principles into practice. It has set several guidelines, including paying developers for finding bias in AI, akin to the bug bounties offered in security software. The paper read, “If companies were more open earlier in the development process about possible faults, and if users were able to raise (and be compensated for raising) concerns about AI to institutions, users might report them directly instead of seeking recourse in the court of public opinion.” Bug bounties programs reward researchers for identifying security flaws, have transformed the way the technology sector approaches vulnerabilities, says Andrew Cormack, a chief regulatory officer at Jisc. Cormack called bug bounties a transformation point for the way the tech sector approaches vulnerabilities. Today, vendors and security researchers are proactively engaging with security as a standard practice. While bounty competitions are not a replacement for structured testing and analysis, they open up the gates for organisations to challenge their systems to tests they might not have considered on their own. Popular bug bounty programs for AI bias Logically’s Bug Bounty Program: Logically has been working with security professions to protect the customer’s from harmful networks and mobile applications. The Mozilla Security Bug Bounty Program is designed to enforce security research in Mozilla software and provide an incentive to those who help make the internet a safer place. Its CRASH project- Community Reporting of Algorithmic System Harms (CRASH), brings key stakeholders together for discovery, scoping and iterative prototyping of tools. This is to enable more accountable and harmless AI systems. HackerOne, a “hacker-based” security testing platform, is hosts ‘The Internet Bug Bounty’. This program rewards hackers who manage to uncover security vulnerabilities in some of the most important softwares on the internet. The program, managed by a panel of volunteers selected from the security community, is sponsored by Facebook, GitHub, Microsoft, Hackerone and Ford Foundation. Crowdsourced security platform, Bugcrowd combines analytics, automated security workflows, and human expertise to find and fix critical vulnerabilities. Bugcrowd announced Series D funding in April 2020 of $30 million. It has an expansive list of clients they have worked with, including Tesla, Atlassian, Fitbit, Square, and Mastercard. They review platforms for big tech giants and retail space like Amazon and eBay.","excerpt":"AI-based companies usually turn to the black box and avoid critical analysis of the results obtained from their models.","categories":["AI Features"],"tags":["AI biases"],"author_name":"Avi Gopani","publish_date":"2021-08-09T17:00:00","publication_year":"2021","word_count":601,"keywords":["Go","artificial intelligence","machine learning","AI biases","AI","OpenAI","AWS","ML","TPU","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","OpenAI","AWS","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-can-bug-bounty-programs-combat-ai-biases\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":21590,"title":"Data Forms The Backbone Of Everything We Do — says Manav Sethi, CMO at ALTBalaji","content":"At a time when Indian Video-on-Demand (VOD) platforms is getting crowded by the day, ALTBalaji – the youngest entrant in the space has scripted a phenomenal success story with more than 15 million downloads. Even though biggies such as Hotstar & Voot are getting more eyeballs and have a wide user base, ALTBalaji (launched in April, 2017) continues to push up the numbers by offering a completely ad-free, only original content which is streamed in 90 countries globally. What’s different this time for the undisputed king of Indian television and entertainment is that much of their success is riding on data that forms the backbone of operations and strategy and the platform is all about original content that ticks with the audience. In a conversation with Analytics India Magazine, ALTBalaji’s Chief Marketing Officer Manav Sethi who is tasked with the job of international expansion, growth, customer acquisition & retention, product as well as UI guidelines tells us how data plays a huge role at ALTBalaji and why drumming up original content will always be an art + science. 1) Can you tell us how data and analytics is used to find insights to maximize customer engagement, and drive conversions by finding the relevant audience? Given the kind of business we are in, data analytics forms the bedrock of everything we do. At any given point of time, we have 120+ triggers\/ predictors that we catch before the customer is in suspecting stage on the platform or say registered with us. For example, if a customer viewed a trailer or interacted with the platform in any way, the kind of signals we look at are what frame he has viewed, what was the last viewed scene, location, gender and demographic et al. Then we go ahead and see whether he\/she used the platform from one location or multiple locations, whether it is consumed on WIFI or mobile internet. 2) What tools or technologies does the data science team work on and what’s the size of the team? All our data is very closely synced with RedShift and Microsoft Azure as a combination. From an analytics and visualization standpoint we use Tableau. We have a home-grown platform which is built on Python because in the high intensive computing environment, read-ops are a lot more and Python helps with that. In terms of performance marketing, we work with a team called Sokrati which has been acquired by Dentsu recently. Besides Sokrati, we also use lot of tools like App Annie for composition in the ecosystem and AppsFlyer. We also work with a fraud detection tool. As for team, our current team size is 5 members and that includes both marketing analytics and analytics as a whole. But that said, we are always on a lookout for new members. 3) Can you tell us about your active base? Also, Hotstar pushes 100 million subscribers, but all are not paid users? We have about 15 million downloads currently and if you look at the last App Annie report that came out last year, we were ranked amongst the top 3 revenue grossing video streaming app in the country, as per ‘State of  Video Streaming Apps in India’ report. Now that was an eye-opener for the entire industry since we launched the platform in April of 2017 that report came out in September and we were already number 3 after Netflix and Hotstar. The only difference is that for Hotstar, the biggest revenue stream is not subscription, it is ad based we are the only competing ad-free platform in India. That’s why for us there is a lot of value in data and a there is a lot of marketing analytics right from customer lifecycle management and CRM intertwined in it. 4) Can you give us a use case of how retention is used to drive customer acquisition? From the very moment our acquisition team brings in a user, our retention team has created programs which are mapped with these triggers internally. So, for example, you watched show one and you lapsed after episode 3, what all can I do to bring you back? I can take a frame from episode 6 and say “Hey you have not seen that”. Just like this, there are a lot more programs we have created internally, and I would like to say this on record, that we have a higher double-digit number conversion when it comes to registered and paid subscribers. Also, remember – we are the only app available in 90 countries globally and we have paying customers from about 50 of these countries. Our top performing countries after India are US, UK, Singapore, Middle East, Bangladesh and Pakistan. 5) Can you give a use case about user acquisition? So, the way we structured our acquisition strategy is say, when you come to our platform you can watch first two episodes free, but you can watch the next three episodes after you register with us. Our conversion programs are structured in such a way that we identify certain attributes and there’s the ability to keep coming back to you. There are programs that are built for anonymous users and there are programs that are built for registered users on the site. And the focus of the retention team is to keep converting the both the sets of users back. 6) There’s so much talk about original content being dictate by data? Is that the case at ALTBalaji also which is known for its cutting-edge original web series? I am saying this globally, that it is too early for programs or content to be created based purely on data insights. Even Netflix, if they mandate the next season of Narcos or House of Cards — it is not they will go out and do a focus group study to make one or not. In this case, I am not talking about season 2, I am talking about new shows. By and large, some amount of data is used of course and we have been in the industry for so long it gives us a lot of visibility into what our consumers like to consume. But there is no definitive science to it. It is science plus art. So, we have certain signals and we keep on looking at those signals in terms of which viewer is coming from which geography and what kind of shows they are consuming. This gives us the ability to go back and keep re-tweaking our programming. So, when I launched ALTBalaji, I launched it with 12 shows, out of which we have already announced season 2 of three shows. Data is the backbone of everything that we do. And that is what differentiates us from a typical broadcast TV right. For every poster which I make, every trailer which I make there is a set of audiences in mind. Before I release a trailer, I do a lot of A\/B testing which part of the geography will like what kind of trailer and poster. 7) ALTBalaji is big on personalization, can you tell us about your recommender system as well? On the personalization side, we are doing a lot – the mailer you will get is not the mailer the other guy will get but as of now the home-screen for all the viewers is the same. We are not that high on recommendation at least as of now but we would invest in it a year from now because right now we have 18 shows. Once, we have a big enough library, we would start focusing on recommendations also. 8) What are the focus areas in 2018? Going in 2018, we would like to create at least 200-300 hours of original shows. Our belief is that that India is an entertainment hungry country, and if you look there are 165 million cable lines in home and 200 million TVs in households but just a billion mobile population out of which smartphone are still about 400 million or so. There is a lot of scope for Hindi content which is how my dear friend Sameer Nair famously propounded Chasm between Naagin & Narcos. Out of the 100 million potential people that lean towards English entertainment, there are about 700 million viewers looking for Naagin type entertainment. That’s where ALTBalaji will have a role to play and that’s where all our energies are.","excerpt":"At a time when Indian Video-on-Demand (VOD) platforms is getting crowded by the day, ALTBalaji – the youngest entrant in the space has scripted a phenomenal success story with more than 15 million downloads. Even though biggies such as Hotstar & Voot are getting more eyeballs and have a wide user base, ALTBalaji (launched in […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Richa Bhatia","publish_date":"2018-02-12T05:10:19","publication_year":"2018","word_count":1398,"keywords":["data science","Go","cloud_platforms:Azure","AI","R","Git","Interviews and Discussions","Python","analytics","Azure","fraud detection"],"extracted_tech_keywords":["AI","data science","analytics","fraud detection","Azure","Python","R","Go","Git","cloud_platforms:Azure"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-forms-backbone-everything-manav-sethi-cmo-altbalaji\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":65210,"title":"Must-Have Skills To Kick Start Your Cloud Computing Career","content":"Cloud computing market is expected to rise from $272 billion in 2018 to $623.3 billion by 2023 at a CAGR of 18%. The cloud computing market is lucrative, and several companies around the globe have been moving to cloud services from their on-premise data centres. With increased internet usage, automation, and agility, the cloud computing industry is set for a great future. So, naturally, this opens many job opportunities for aspirants. Below we have mentioned some must-have skills for aspirants who want to break into the cloud computing sector: Cloud Platforms One thing you cannot do without when you start learning cloud computing is the knowledge of cloud platforms. But, not only does one need the knowledge about a particular cloud platform, but they also have to understand how different cloud providers work. The cloud service providers like Azure and AWS offer end-to-end services like database, compute, storage, migration, and ML. So, knowing how these cloud service providers work inside out becomes crucial. When it comes to choosing which platform you can study, the popular choice is always between AWS and Azure. Both these are market leaders and are always at war with each other. There is another choice in GCP too, which comes next. Storage Storage is defined here as ‘storing data online on the cloud.’ Data is the main driving force behind cloud computing, and it is vital to understand where it is stored and how it is stored. Depending on a company’s needs, it can choose from the following types of storage: Hybrid cloud storagePersonal cloud storagePrivate cloud storagePublic cloud storage So, when one completely understands how the data is stored and where it is stored, it gives them a complete idea of how data is dealt with inside a cloud environment. Networking As centralised computing resources for clients are shared over the clouds, networking has led to the rise of network management functions inside the clouds. Having more network management functions results in fewer customer devices needed to manage the network. Two other reasons, apart from sharing the cloud resources, which have pushed for more networking are the increasing internet access and more reliable WAN bandwidth. These factors have also spurred the demand for cloud networking as customers are always looking for network access using cloud-based service. So, a cloud engineer might also be responsible for designing ways to ensure that the networks are always responsive to the demands of the customers by building automatic adjustment procedures. A cloud engineer needs to understand networking fundamentals and virtual networks that are important for networking on the cloud. Cloud Security Irrespective of the sector, security is an important thing to consider for any company. In the initial days, one of the main reasons organisations were sceptic about using cloud services was security. Considering how unsafe the internet was a few years ago, people were concerned whether cloud storage is safe enough to house sensitive and important information. Although today’s internet is safer than it was in the past, cloud security is the primary aspect for developers and cloud engineers. Today, organisations use cloud security tools like Bitglass, Skyhigh networks, Okta, CipherCloud, etc., to secure their data. In addition to having security tools, professionals who will be responsible for the security are expected to have in-depth knowledge of these tools. To be able to handle the security better, one also needs to go through the CCSP (Certified Cloud Security Professional) training that will help one gain adequate knowledge and professional understanding of cloud security. Cybersecurity professionals are in great demand. Data Management It is already a known fact that data is an essential factor when it comes to cloud. From a general perspective, clouds are used by the public to store information like their photos, videos, documents etc. And for companies, the most important part is the sensitive information that is stored on the cloud. So, an effective way of collecting, storing, maintaining and accessing data on the cloud is crucial for organisations. Given how important data management is, it becomes essential to obtain data management skills in cloud computing. To be better at data management, one needs to learn data-oriented languages like SQL and Hadoop. Serverless Architecture Serverless architecture offers many advantages over the traditional cloud-based or server-oriented infrastructure. For many, it offers better scalability, flexibility, and speed while running at a reduced cost. Developers do not need to worry about purchasing, managing backend servers, and provisioning. Today’s cloud consists of industry-standard technologies and programming languages that help in moving serverless applications between cloud vendors. With the many advantages that serverless architecture holds, it becomes imperative to learn serverless architecture. There are many courses online, like Lambda tutorials, to learn serverless architecture on AWS platforms.","excerpt":"Cloud computing market is expected to rise from $272 billion in 2018 to $623.3 billion by 2023 at a CAGR of 18%. The cloud computing market is lucrative, and several companies around the globe have been moving to cloud services from their on-premise data centres. With increased internet usage, automation, and agility, the cloud computing […]","categories":["AI Features"],"tags":["career in business intelligence","Cloud Computing","Cloud Computing in IT Industry","hadoop on azure cloud","private cloud"],"author_name":"Sameer Balaganur","publish_date":"2020-05-14T16:00:00","publication_year":"2020","word_count":787,"keywords":["GCP","hadoop on azure cloud","AWS","AI","cloud computing","Azure","Cloud Computing in IT Industry","ML","R","serverless","RAG","career in business intelligence","Cloud Computing","SQL","private cloud"],"extracted_tech_keywords":["AI","ML","RAG","cloud computing","AWS","Azure","GCP","serverless","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/must-have-skills-to-kick-start-your-cloud-computing-career\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10021233,"title":"AI-powered Audio Transcription Tools To Check Out In 2021","content":"Watching a TED Talk online is an experience, especially since we can scroll down to read the transcript in greater detail. Audio transcription tools are incredibly useful, and we can view them as almost priceless pieces of software. They are indeed a true lifesaver. Ask the student or journalist whose life has practically been transformed with audio transcription. It has made interviewing and note-taking a fabulous affair where they were once seen as tedious, back-breaking tasks consuming most of their workday. But audio transcription tools that make unforgivable mistakes while transcribing audio and video into text can make a mountain out of a molehill. The last thing a good copywriter wants to do is to rely on an incorrect transcription and end up watching\/listening to the audio\/video file all over again to correct the mistakes. Transcribing natural language is probably the hardest for an audio transcription tool because understanding every voice’s subtleties and lilts is still beyond technology’s capabilities. This is where NLP can come in because it has the capability of learning and understanding the nuances of different languages, dialects, and voices. In this article, we are going to list down eight such powerful AI-based audio transcription tools that are relevant this year based on their accuracy and intelligence: Verbit: Verbit is probably 2021’s ultimate professional AI-powered audio transcription tool. It seems to cater to the right customer pool comprising the education sector, media houses, and enterprises. Verbit is backed by AI and automated speech recognition technology that work behind the scenes to do the transcription, giving Verbit it’s much-deserved position among 2021’s recommended AI audio transcription tools. Otter.ai Otter is a virtual assistant that does all the notetaking for every Google Meet, Zoom call or lecture. Otter promises intelligent, live transcription and even offers features like playback and keyword search. It’s speech recognition system combines AI and machine learning to feed semantics into the system. The NLP takes over from there to understand words and pronunciations. Also, machine learning works around a large data set within a massive neural network consisting of many nodes. Scribie: Scribie is an affordable audio transcription tool that works on manual as well as automated transcription. Manual transcription follows a four-step process with transcription at the top of the rung, all of which goes all the way down to quality check. The automated transcription then uses cutting-edge AI combined with a constantly learning NLP model built on real-world data from Scribie’s manual transcription service. Scribie goes through a technology overhaul year-on-year to upgrade its systems. The latest upgradation to their technology has been optimisations of their existing QC auditing system. Amazon Transcribe: Yes, Amazon is also very much into audio transcription, as expected. As a part of Amazon’s Web Services (AWS) offered for developers to integrate into their applications, Amazon Transcribe is very much on the scene in 2021. Amazon Transcribe is also being offered to enterprises arenas like contact centres, VOD, etc. As per its product page, Amazon Transcribe performs transcription using the deep learning-based automatic speech recognition (ASR) process. Jog.ai: Jog.ai offers to transcribe calls. Combining AI and NLP, Jog.ai is essentially an enterprise-level audio transcription tool that does the task of an intelligent personal assistant to both big and small organisations. When a call is made or received, Jog.ai’s speech recognition platform triggers the natural language processor to keep a tab on the conversation’s key phrases. Trint: Another interesting AI transcriber for 2021, Trint, is being marketed towards content creators and caters to enterprises as an exclusive and cost-heavy service. With automated speech recognition and natural language processing as the building blocks to its AI transcriber, Trint is a 2021 ready service. Sonix: Sonix’s market positioning is as a quick-to-transcribe product. It has several large enterprises in its customer base, mostly premier universities, publications, and popular media companies. Sonix has amazing features to show off, like speaker labelling and automated timecode realignment, all built on cutting-edge artificial intelligence. Descript: Descript is a 2021 audio transcription tool that’s offering a lot more than just audio transcribing. Descript’s AI capabilities work in a doc-like fashion which works great for enterprise collaboration. It relies on a 3rd party enterprise-grade transcription engine to carry out the AI transcription. Along with transcription, the AI system can also be used for editing podcasts, videos and text-to-speech overdub, along with screen recording and remote recording.","excerpt":"Watching a TED Talk online is an experience, especially since we can scroll down to read the transcript in greater detail. Audio transcription tools are incredibly useful, and we can view them as almost priceless pieces of software. They are indeed a true lifesaver.   Ask the student or journalist whose life has practically been transformed […]","categories":["AI Features"],"tags":["AI and machine learning","automated machine learning","natural language processing ai","Vernacular Automated Speech Recognition"],"author_name":"Anju Nambiar","publish_date":"2021-03-03T12:29:57","publication_year":"2021","word_count":728,"keywords":["Go","artificial intelligence","machine learning","AWS","AI","neural network","Vernacular Automated Speech Recognition","NLP","deep learning","AI and machine learning","edge AI","natural language processing ai","automated machine learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","NLP","edge AI","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-powered-audio-transcription-tools-to-check-out-in-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10063292,"title":"Telangana to deploy AI in public schools to modernise education","content":"International Institute of Information Technology, Hyderabad (IIITH) has started a pilot project with an aim to introduce AI tools in select government schools in Moinabad, Telangana. IIITH has been working with a slew of schools, and the Education Department to build AI and new-age tech solutions to increase opportunities at the grassroots level. Ramesh Loganathan, Co-Innovation Professor at IIIT-H, said the next step is to meet the stakeholders and chalk out plans for technology implementations in the schools. The projects will be of three to six months duration, he added. The data collection and building the education portal involve a lot of manual work and is time-consuming. AI-based cameras, speech recognition tech and other AI tools will be used for marking attendance, midday meals etc. The IITH plans to bring in startups with relevant technologies to deploy AI tools, he said.Education policy 2021 has made a formative assessment of students mandatory. Teachers are expected to evaluate every class and submit a report on the performance of the students. The IITH pilot project will use an AI-based camera system to assist the teachers in rate the performance of the students. The team is building tech to teach spoken languages and is in the process of integrating it to the curriculum.","excerpt":"Education policy 2021 has made a formative assessment of students mandatory.","categories":["AI News"],"tags":["AI and machine learning","AI Applications","AI in school","AI projects","AI Software","IIIT hyderabad"],"author_name":"Kartik Wali","publish_date":"2022-03-22T19:38:24","publication_year":"2022","word_count":209,"keywords":["Go","IIIT hyderabad","AI projects","programming_languages:R","AI","AI Software","innovation","AI Applications","programming_languages:Go","Aim","AI and machine learning","GAN","AI in school","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","innovation","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/telangana-to-deploy-ai-in-public-schools-to-modernise-education\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018526,"title":"Guide Towards Fast, Accurate, and Stable 3D Dense Face Alignment(3DDFA-V2) Framework","content":"3D dense face alignment(3DDFA) is a trending technique for many face tasks, For example, object recognition, animation, tracking, image restoration, and many more. For now, most of the studies in 3DDFA are divided into two categories: 3D Morphable Model(3DMM) parameter regression,and Dense vertices regression. Now Existing method of 3D dense face alignment only focuses on accuracy, which tends to limit the scope of their practical applications. In previous implementations of 3DDFA, there was a major problem with accuracy and output inference but the new 3DdFA_V2(version 2nd) came up with a new regression framework that makes a reliable balance between accuracy, speed, and stability. 3DDFA_V2 is published by Jianzhu Guo, Xiangyu Zhu, Yang Yang, Fan Yang, Zhen Lei, and Stan Z. Li in the research paper called Towards Fast, Accurate and Stable 3D Dense Face Alignment The new backbone of this framework is very lightweight and its source code is open-sourced on GitHub here. The repository is owned by Jianzhu Guo. The research has been accepted by ECCV 2020 3DDFA_V2 (3D Dense Face Alignment- version 2) It is an improved version of previous implementations of 3DDFA now named 3DDFA_V2, it achieved promising speed, accuracy, and stability also incorporating the fast face detector FaceBoxes instead of Dlib. It introduced the new meta-joint optimization strategy to dynamically regress a small set of 3DMM parameters, also to further improve the stability of the model on videos authors virtual synthesis method to convert one image to a short-video which integrates in-plane and out-of-plane face moving. It runs over 50fps(19.2ms) on a single CPUCan reach up to 130fps(7.2ms) on multiple CPU(i5-8259U) core.It is 24x times faster than PRNetIt is a more dynamically optimized technique leveraging the 3DMM parameter through a novel meta-optimization strategy. Architecture 3DDFA_V2 architecture consists of four parts: Lightweight backbone MobileNet architecture for 3DMM parameter predictionsMeta joint optimization of fWPDC and VDC.Landmark regression regularizationShort-video synthesis for better training. To decrease the computation burden the landmarks regression branch is discarded during inference. Implementation To see the outputs we will be using Google Colab demonstration, first, we will clone the 3DDFA_V2 repo, and then we will set up the environment. Let’s jump straight to the code: Cloning and setup up the environment %cd \/content !git clone https:\/\/github.com\/cleardusk\/3DDFA_V2.git %cd 3DDFA_V2 !sh .\/build.sh Importing modules import cv2 import yaml from FaceBoxes import FaceBoxes from TDDFA import TDDFA from utils.render import render from utils.depth import depth from utils.pncc import pncc from utils.uv import uv_tex from utils.pose import viz_pose from utils.serialization import ser_to_ply, ser_to_obj from utils.functions import draw_landmarks, get_suffix import matplotlib.pyplot as plt from skimage import io Load 3DDFA configurations It will enable the ONNX environment to speed up the process, default backbone of the architecture is MobileNet_V1 with input size 120×120, and the default pretrained weight\/mb1_120x120.pth, there is also another wider factor is available as this project provide two mobilenet models to choose from. ModelInput#Params#MacsInference (TF)MobileNet120×1203.27M183.5M~6.2msMobileNet x0.5120×1200.85M49.5M~2.9ms3DDFA supported models cfg = yaml.load(open('configs\/mb1_120x120.yml'), Loader=yaml.SafeLoader) onnx_flag = True  # True to use ONNX to speed up if onnx_flag: !pip install onnxruntime import os os.environ['KMP_DUPLICATE_LIB_OK'] = 'True' os.environ['OMP_NUM_THREADS'] = '4' from FaceBoxes.FaceBoxes_ONNX import FaceBoxes_ONNX from TDDFA_ONNX import TDDFA_ONNX face_boxes = FaceBoxes_ONNX() tddfa = TDDFA_ONNX(**cfg) else: face_boxes = FaceBoxes() tddfa = TDDFA(gpu_mode=False, **cfg) Testing Let;s first take any image you want img_url = 'https:\/\/photovideocreative.com\/wordpress\/wp-content\/uploads\/2017\/12\/Angles-de-prise-de-vue-horizontal-contreplong%C3%A9-et-plong%C3%A9.jpg' img = io.imread(img_url) plt.imshow(img) img = img[..., ::-1]  # RGB -> BGR Detecting faces using FaceBoxes The FaceBoxes module is modified from FaceBoxes.PyTorch. There are some previous work on 3DDFA or reconstruction are available like: 3DDFA, face3d, PRNet. boxes = face_boxes(img) print(f'Detect {len(boxes)} faces') print(boxes) #Regressing 3DMM params, reconstruction and visualization param_lst, roi_box_lst = tddfa(img, boxes) Reconstructing vertices and visualizing sparse landmarks using 3DDFA dense_flag = False ver_lst = tddfa.recon_vers(param_lst, roi_box_lst, dense_flag=dense_flag) draw_landmarks(img, ver_lst, dense_flag=dense_flag) Reconstructing vertices and visualizing dense landmarks dense_flag = True ver_lst = tddfa.recon_vers(param_lst, roi_box_lst, dense_flag=dense_flag) draw_landmarks(img, ver_lst, dense_flag=dense_flag) Reconstructing vertices and render ver_lst = tddfa.recon_vers(param_lst, roi_box_lst, dense_flag=dense_flag) render(img, ver_lst, tddfa.tri, alpha=0.6, show_flag=True); Reconstructing vertices and render pncc ver_lst = tddfa.recon_vers(param_lst, roi_box_lst, dense_flag=dense_flag) pncc(img, ver_lst, tddfa.tri, show_flag=True); Running it on Video python3 demo_video.py -f examples\/inputs\/videos\/214.avi --onnx Conclusion This new approach for more stable, fast, and accurate 3D Dense Face alignment is really a new way of training and inference face data. In Colab the code takes nearly seconds to run. just because of lightweight mobinet architecture, surprisingly the latency of onnxruntime was also much smaller on CPU, for more you can follow the below resources: Official GitHub RepositoryOfficial Research PaperRunning on windows discussionGoogle Colab demo Guide to OpenPose for Real-time Human Pose Estimation","excerpt":"3D dense face alignment(3DDFA) is a trending technique for many face tasks, For example, object recognition, animation, tracking, image restoration, and many more. For now, most of the studies in 3DDFA are divided into two categories: 3D Morphable Model(3DMM) parameter regression, and Dense vertices regression. Now Existing method of 3D dense face alignment only focuses […]","categories":["Deep Tech"],"tags":["face recognition","Matplotlib","Python","Pytorch"],"author_name":"Mohit Maithani","publish_date":"2021-01-21T10:00:00","publication_year":"2021","word_count":747,"keywords":["Pytorch","Go","TPU","AI","PyTorch","ML","RAG","Python","Colab","face recognition","Matplotlib","R"],"extracted_tech_keywords":["AI","ML","PyTorch","Colab","Matplotlib","RAG","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-towards-fast-accurate-and-stable-3d-dense-face-alignment3ddfa-v2-framework\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":4190,"title":"Telecom – Handling the Churn!","content":"Mobile Number or Hand-phone number has come to represent an element of person’s identity and hence there is an implicit inertia that is built into a person’s action while mobile numbers are to be changed. Thus, mobile companies were building up an embedded loyalty into customers’ behaviour, when mobile numbers portability was not in force – there was a reluctance to change the operator even when the customer was not satisfied with the services or charges or plans. With the advent of “Mobile Number Portability” (MNP) this very paradigm of “forced loyalty” was broken – as the customer could shift the operators without changing the mobile number. This resulted in companies becoming more customer friendly in terms of their plans, pricing, and service. Customer Loyalty – Its impact on the Cost A survey of telecom literature indicates that the cost of acquiring a new customer is at least 5 times the cost of retaining an existing customer. With mobile telephony penetration curve flattening, market players are eyeing the same market, which is fast shrinking. Mobile connectivity becoming the order of the day irrespective of the geographical reach – be it either urban or rural – customers are forcing network providers to dynamic plans which are pocket-friendly to the customer and at the same time to invest in the back-end infrastructure to reduce call-drops. All these market dynamism on the players’ part is expected to play a role in creating the “hygiene factor” for customer retention. However, the players are depended on the “invested” customers to increase their use of data services (2G \/ 3G) and value-added services (VAS) to increase their top-line. Thus, it becomes all the more imperative for the market players to retain their existing customers and induce them to build their relationship for a higher mobile spend on various services offered by them. Customer Retention Programs: The new mobile market paradigm has forced CRM Managers to shift gears in designing their retention programs to become more dynamic and customer focused. The success of these programs will be measured on the ‘Δ’ it creates in most singular metric on which most telecom companies are measured – ARPU (average revenue per user). For even the best of the retention programs to be successful, it has to be able and amply supported by the back-end analytics infrastructure that maximizes the impact of the program by aiding the decision makers to zoom in on the right target segments for each customized retention program.  However, Retention\/CRM managers find that they get less than acceptable returns on their retention programs. This is mostly because these programs are not targeted sharply enough. A large proportion of customers targeted are often the ones who would not have churned in the 1st place. Churn Management – Analytics at play Predictive Analytics is one of the key tools deployed to bring sharpness to the customer targeting exercise, and to optimize the marketing spend, in a scientific data-driven manner. Analytics is used in churn management – in the form of regression models, decision trees – with an aim to identify the potential attrite, and flag their behaviour with a goal to reduce the total churn. Analytics at work, for churn management, is built with an aim of knowing in advance “Who will churn?” For optimal targeting, i.e. allocating resources on basis of probability of attrition (maybe along with profitability of the customer or other factors), basic data variable\/attribute requirements are: –        Demographic data from customer information file like age, sex, zip code etc. –        Contractual data from service account file such as pricing plan, activation data, contract identification etc. –        Usage & Payment data from billing system such as the number of calls, airtime, fixed line time, the total amount spent, number of calls made to customer care centre, change in price plan etc. Churn Management – Current Scenario However, in the process of building a single customer view, handling many variables sometimes causes dilution while focusing on price elasticity aspect of telecom services – essentially, the focus on time series data on customer’s payments and calls made. Looking at different price-quantity coordinates and drawing insights from the change in trends\/patterns is an important exercise which often gets overlooked. Time to the expiry of the contract is an important variable diligently tracked by the telecom operators. Out of attriting customers, majority attrite after their contract expires. So, sometimes because of this reasoning, Survival analysis, which tries to answer both questions “Who will churn?” as well as “When will he\/she churn?” gets ignored. The Key Question: And of course, none of these methods answers the question “Why does a customer churn\/attrite?” And the answer to this question calls for a rigorous and cordial marriage between telecom CRM strategies and telecom CRM analytics.","excerpt":"Mobile Number or Hand-phone number has come to represent an element of person’s identity and hence there is an implicit inertia that is built into a person’s action while mobile numbers are to be changed. Thus, mobile companies were building up an embedded loyalty into customers’ behaviour, when mobile numbers portability was not in force […]","categories":["IT Services"],"tags":[],"author_name":"Lokesh Arora","publish_date":"2013-10-16T09:13:09","publication_year":"2013","word_count":793,"keywords":["Go","programming_languages:R","AI","R","data-driven","RAG","Aim","ViT","analytics","predictive analytics"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","predictive analytics","R","Go","ViT","data-driven","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/telecom-in-india-handling-the-churn\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10161451,"title":"Anil Bhasin Steps Down as Databricks India VP, Eyes New Venture in Cybersecurity","content":"Anil Bhasin, the vice president of India and SAARC region at Databricks, announced on Wednesday that he is leaving the company after a tenure of over two and a half years. In a statement reflecting on his time at Databricks, Bhasin expressed gratitude for the opportunity to work with a talented team and contribute to building the Databricks brand in India. “I feel truly humbled by the fact that the majority of the Top 500 Enterprises in India allowed us to earn their trust and chose to partner with Databricks,” he said. Speaking about the evolving landscape of technology sales, Bhasin said that enterprises now prioritise strategic partnerships over traditional selling methods. “In this world of increasing complexity, Enterprises and CXOs today seek a very different approach from OEMs that is focused on solving their toughest challenges through simplification of technology.” During his tenure, Bhasin played a significant role in promoting Databricks’ innovative ‘Lakehouse’ architecture. This architecture integrates data lakes and warehouses to enhance data accessibility and usability for businesses. He highlighted the importance of democratising AI and making it accessible for organisations to become data-driven. Bhasin also acknowledged the support he received from Databricks’ leadership team and thanked key figures like CEO Ali Ghodsi and SVP and GM Ed Lenta for their guidance. He expressed pride in the accomplishments of the India team, saying, “I couldn’t be more proud of what team India has achieved in a short period of time.” Looking ahead, Bhasin hinted at a new venture in cybersecurity and said that he would soon embark on “a fascinating company that is leading a market transition that will change the world of cyber security”. His departure marks a significant transition for Databricks, which is continuing to expand its footprint in India and has been identified as a key growth market. Bhasin has been in the tech industry for over thirty years, including previous leadership roles at UiPath and Palo Alto Networks. His expertise has been instrumental in driving technological advancements and business growth within these organisations. Databricks recently raised $10 billion in a funding round led by Thrive Capital, with participation from leading venture capital firms such as Andreessen Horowitz, DST Global, GIC, and others.","excerpt":"Bhasin expressed gratitude for the opportunity to work with a talented team and contribute to building the Databricks brand in India.","categories":["AI News"],"tags":["Databricks"],"author_name":"Siddharth Jindal","publish_date":"2025-01-15T13:55:27","publication_year":"2025","word_count":368,"keywords":["API","funding","AI","data-driven","venture capital","Rust","GAN","R","data lake","Databricks"],"extracted_tech_keywords":["AI","Databricks","R","Rust","API","data lake","GAN","data-driven","venture capital","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/anil-bhasin-steps-down-as-databricks-india-vp-eyes-new-venture-in-cybersecurity\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60216,"title":"Dozee Offers Free Devices To Track Respiratory Health To Quarantined Bangaloreans","content":"Recently, Dozee Health Monitoring System joined hands with Duroflex Mattresses and announced to offer free devices to track respiratory health and heart rate of the quarantined individuals in Bangalore. In one of our articles, we discussed how Bangalore-based Dozee’s advanced health intelligence technology identifies the early signs of health deterioration and early intervention saves a great deal of pain as well as the cost to the user. Dozee is a contact-free health monitor that tracks heartbeat, respiration, sleep, and stress-recovery with medical-grade 98.4% accuracy. It is the only device that gives continuous respiration data, that too without the need of wires or any technical expertise. Dozee is a device that has a thin sensor sheet that goes below the mattress. of the users. The brand has witnessed 5 requests within 15 minutes of the announcement. With the help of this device, doctors can now easily check heart rate, breath rate, stress levels remotely and monitor the health of the patients. The device raises alerts to the users in case of any inconsistency in these vitals. Those in self-quarantine or with a potential risk can monitor their health on an ongoing basis at home. The reason behind this offering is to help in proactive health monitoring at this crucial hour as vitals can be checked at home, doctors can check their health remotely and possibly flag off health deterioration in advance. The device is available on both Amazon.in & its website.","excerpt":"Recently, Dozee Health Monitoring System joined hands with Duroflex Mattresses and announced to offer free devices to track respiratory health and heart rate of the quarantined individuals in Bangalore. In one of our articles, we discussed how Bangalore-based Dozee’s advanced health intelligence technology identifies the early signs of health deterioration and early intervention saves a […]","categories":["Global Tech"],"tags":[],"author_name":"Ambika Choudhury","publish_date":"2020-03-27T16:03:13","publication_year":"2020","word_count":240,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/dozee-offers-free-devices-to-track-respiratory-health-to-quarantined-bangaloreans\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10066644,"title":"All you need to know about Graph Attention Networks","content":"Graph neural processing is one of the hot topics of research in the area of data science and machine learning because of their capabilities of learning through graph data and providing more accurate results. Various researchers have developed various state of the art graph neural networks. A graph attention network is also a type of graph neural network that applies an attention mechanism to itself. In this article, we are going to discuss the graph attention network. The major points to be discussed in the article are listed below. Table of content What is a graph attention network?Graph neural network (GNN)Attention layerCombination of GNN and attention layer    The benefit of adding attention to GNNThe architecture of graph attention network   Advantages of the graph attention network Let’s start by understanding a graph attention network What is a graph attention network? As the name suggests, the graph attention network is a combination of a graph neural network and an attention layer. To understand graph attention networks we are required to understand what is an attention layer and graph-neural networks first. So this section can be divided into two subsections. First, we will look at the basic understanding of the graph neural network and attention layer then we will focus on the combination of both.  Let’s take a look at the graph neural network. Graph neural network (GNN) In one of our articles, we can see an implementation of a graph neural network and we have also discussed that graph neural networks are the networks that are capable of dealing and working with graph-structured information or data. There are various benefits of using graph-structured data in our projects such as these kinds of structures hold the information in the form of vertices and nodes of the graph and it becomes very easy for the neural networks to understand and learn data points present in the graph or three-dimensional structure. Taking an example of data related to a classification problem can consist of labels in the form of nodes and the information in the form of vertices. Most real-world problems have data that is very huge and consists of structural information in itself. Using a graph neural network can provide a state of the art performing model. Attention layer In one of our articles, we have discussed that the attention layer is a layer that enables us to design a neural network that can memorize the long sequences of information. Generally, we find the uses of such layers in the neural machine translation problems. A standard neural network works by encoding the sequential information in the form of compressed context vectors. If an attention layer is included in the network then the network will be forced to work by creating a shortcut between the input and the context vector. The attention layer will help change the weights of the shortcut connection for every output. Since the connection between input and context vector provides the context vector to access all input values, the problem of the standard neural network forgetting the long sequences gets resolved. In simple words, we can say that the implementation of the attention layer in neural networks helps provide attention to the important information from the data instead of focusing on the whole data. This way we can make our neural network more reliable and stick to the only important information. Here we can see that till now we were applying the attention layer to the neural network but the article focuses on applying the attention layer or mechanism to a graph neural network. Let’s see what applying attention to graph neural networks will mean. Combination of GNN and attention layer In the above points, we have discussed that a graph neural network is a better way to deal with data that has long structural information and an attention layer is a mechanism that helps in extracting only useful information from long or big data. Both of these things can be combined then we can call it a graph attention network. A graph attention network can also be explained as leveraging the attention mechanism in the graph neural networks so that we can address some of the shortcomings of the graph neural networks. Let’s take an example of graph convolutional networks that are generally used in solving problems related to sequential information present in the data. These graph networks apply stacked layers in which nodes can consist of features of the neighbour nodes. Applying attention to these nodes makes the whole network specify different weights to the different nodes present in the neighbor only. By this method, we make the network capable of working with only those information of the nodes that are useful. One thing that matters here the most is understanding the behavior and importance of the neighbor node on the outcome. These are the methods that can be applied to inductive as well as transductive problems.  We can also say that applying attention to the graph neural network is the way to advance it or make it better. Let’s see what gets improved by applying the attention mechanism to the graph neural network. Are you looking for a complete repository of Python libraries used in data science, check out here. Benefit of adding attention to GNN This section will take an example of a graph convolutional network as our GNN. As of now we know that graph neural networks are good at classifying nodes from the graph-structured data. In many of the problems, one shortcoming we may find is that graph convolutional networks are harming the generalizability of graph-structured data because of the aggregation of information of graph structure. Applying a graph attention network to those problems changes the way of aggregation of information. The GCN provides the sum of neighbour node features as follows: hi(l+1) = (jN(i)(1\/cij)w(l)hj(l)) Where, N(i) = set of the connected nodes cij = normalization on graph structure = activation function w(l)= weight matrix In a similar sum, the attention can provide a statically normalized convolution operation. The below figure is a representation of the difference between standard GCN and GAT. By the above, we can say that by applying attention to the network, more important nodes are getting higher weights during the neighbourhood aggregation. The architecture of graph attention network In this section, we will look at the architecture that we can use to build a graph attention network. generally, we find that such networks hold the layers in the network in a stacked way. We can understand the architecture of the network by understanding the work of three main layers. Input layer: The input layer can be designed as such it is made up of using a set of node features and should be capable of producing a new set of node features as the output.  These layers can also be capable of transforming the input node features into learnable linear features. Attention layer: After transforming the features an attention layer can be applied in the network where the work of the attention layer can be parameterized by the output of the input layer using a weight matrix. By applying this weight matrix to every node we can apply self-attention to the nodes. Mechanically, we can imply a single-layer feed-forward neural network as our attention layer that can give us a normalized attention coefficient. Image source The above image is a representation of the attention layer applied to the GCN. Output layer: after obtaining the normalized attention coefficient we can use them to compute the set of features corresponding to the coefficient and serve them as final features from the network. To stabilize the process of attention we can use multi-head attention so that various independent attention can be applied to perform transformation and concatenation of output features. Image source The above image is a representation of the applied multi-head attention to stabilize the process of self-attention that computes the attention and concatenates the aggregated features. Advantages of the graph attention network There are various benefits of graph attention networks. Some of them are as follows: Since we are applying the attention in the graph structures, we can say that the attention mechanism can work in a parallel way that makes the computation of the graph attention network highly efficient. Applying attention to any setting makes the capacity and accuracy of the model very high because the models need to learn only important data or we can say less amount of data. If the attention mechanism is applied in a shared manner then the graph network can be directly used with inductive learning.            Analysis of the learned weights after applying attention to them can make the process of the network more interpretable. Final words In the article, we have discussed the graph attention network which is a combination of the graph neural network and the attention layer. Applying attention to the GNN can provide improvements in the results and also has several benefits that have been discussed in this article.","excerpt":"A graph attention network can be explained as leveraging the attention mechanism in the graph neural networks so that we can address some of the shortcomings of the graph neural networks.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","graph neural networks","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2022-05-10T12:00:00","publication_year":"2022","word_count":1498,"keywords":["big data","data science","Go","machine learning","TPU","AI","neural network","Machine Learning","RAG","Python","graph neural networks","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","neural network","data science","RAG","TPU","Python","R","Go","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/all-you-need-to-know-about-graph-attention-networks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":64439,"title":"ICMR to Use IBM Watson Assistant to Strengthen India’s COVID-19 Testing Facilities","content":"Indian Council of Medical Research (ICMR), Department of Health Research, Ministry of Health and Family Welfare, Government of India, has collaborated with IBM to implement a Watson virtual agent (called Watson Assistant) on its portal to respond to specific queries of front line staff and data entry operators from various testing and diagnostic facilities across the country on COVID-19. The virtual agent has been deployed on protected pages of the ICMR website that can be accessed only by authorized personnel who are involved with sample collection and testing in hospitals and diagnostic labs. The queries could be related to nature and process of data to be captured by test labs, how to record inventory of test kits & reagents, the process of reporting to various Government agencies and references to the latest guidance, in addition to responding to queries on COVID-19 in general. Professor Balram Bhargava, Director General, Indian Council of Medical Research (ICMR) said, “While India has been able to manage COVID-19 to a certain extent, the challenge still continues to minimize further. It is critical to remain focused on testing, diagnosis and treatment in order to lower the growth curve. With the number of on-field testing teams expanding across the nation as part of India’s COVID-19 measures, this collaboration with IBM will help automate responses from the field & facilitate access to accurate & updated data on COVID-19 diagnostics & reporting. This will help augment our teams’ response time and allow them to concentrate on priorities like developing & updating testing & treatment protocols and guidance for COVID-19.” The Watson virtual agent is AI enabled and will be able to understand and respond to common queries in English and Hindi, from approved testing facilities for COVID-19 across India at scale, around the clock in a uniform & timely manner, as per the latest guidelines. Queries will be categorized under various headings such as Governance, Logistics, Data entry and sharing, Staff Training & Testing and for complex questions, pre-defined contact information of ICMR is made available. The virtual agent is also expected to help in on-boarding new data entry operators and staff of diagnostic centres, as the COVID-19 test network expands across the country. Commenting on the announcement, Sandip Patel, General Manager, IBM India\/South Asia said, “As India rises to meet the challenge of COVID-19, it is crucial to enable government bodies such as ICMR to utilize data and capabilities effectively for rapid detection and treatment. This collaboration is GoodTech in action and a testament to IBM’s commitment of enabling governments, businesses and citizens across the globe to have access to our technology and expertise in tackling the challenges of COVID-19 pandemic.”","excerpt":"Indian Council of Medical Research (ICMR), Department of Health Research, Ministry of Health and Family Welfare, Government of India, has collaborated with IBM to implement a Watson virtual agent (called Watson Assistant) on its portal to respond to specific queries of front line staff and data entry operators from various testing and diagnostic facilities across […]","categories":["AI News"],"tags":["covid-19","IBM"],"author_name":"Vishal Chawla","publish_date":"2020-05-04T14:56:08","publication_year":"2020","word_count":443,"keywords":["Go","API","covid-19","AI","programming_languages:R","programming_languages:Go","IBM","R"],"extracted_tech_keywords":["AI","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/icmr-to-use-ibm-watson-assistant-to-strengthen-indias-testing-facilities-on-covid-19\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10045743,"title":"Why Is Data Unification Such An Organisational Nightmare","content":"Managing corporate data usually necessitates using several software tools such as CRM, email marketing tools, ERPs, etc. Each of these programmes collects data differently. In addition, there are third-party data; however, to make sense of the endless data stream for data-driven decisions – it needs to be organised as a single source. This is where data unification comes in handy. The process aims to merge all of the organisation’s data spread across various operating systems and formats and standardise it to be treated as a single source. As per the industry report, the amount of data created, captured, and consumed globally is expected to reach 180 zettabytes (ZB) in 2025 from 64 ZB in 2020. This further calls for data unification, but organisations are facing numerous challenges: Ensuring Clean Data Data unification is not just about organising data into a single source. It further calls for maintaining data accuracy. For data to be accurate, it must satisfy two criteria – form and content. Consider, for example, different formats of dates can be problematic. Dates stored in the US format will be “8\/10\/2021”, but for a country like India, it is “10\/8\/2021”. Secondly, “New York City” is sometimes captured as “NY” or “NYC” – the consistency of data content needs to be maintained. Otherwise, grouping and summarising data will again be cumbersome. One can avoid this mess by using a customer data platform, which automatically updates (and adds) information to enhance accuracy. When integrating data, it also detects duplication. Data remains in Silos A disconnect between different departments at an organisation makes valuable information inaccessible and invisible to other departments and software systems that can benefit from the same. In a nutshell, data in silos is a sure-shot path to opportunities lost. The right foot forward is to reconnect various departments with a customer data platform that breaks down data silos and makes data available to everyone in a company. Wrong Schema Approach Data unification must be schema last, but organisations fail to understand this simple rule. Data is collected via multiple sources. Moreover, the number of attributes across these various sources is vast. Therefore, any attempt to establish a global schema in advance would be futile. Up-front schema building is not advisable either. The only viable method is to build a schema “bottom-up” from the local data sources. To put it another way, the global schema is created “last.” Lack of Collaboration Professional computer scientists responsible for building data structures and pipelines are left to understand the nuances of data as well. Consider, for example, data from “Tata SIA Airlines Limited” and “Vistara” might be confusing for a data scientist to understand that they are from the same organisation. However, a collaboration between domain experts and computer scientists can resolve such ambiguous circumstances. Outdated Tech Different sets of rules govern traditional tools and systems. As the data size grows, multiple rules creep in. It’s better to provide training data and train machine learning models to deal with scale problems. American computer engineer and A.M. Turing Award winner (2014), Michael Stonebraker, describes the Seven Tenets of Scalable Data Unification: Ingesting data: This needs to be from different operational data systems of an organisation.Performing data cleaning: sometimes -99 is often a code for “null,” and some data sources might have obsolete addresses for customers.Performing transformations: Take, for example, Dollars to Rupees or airport code to city_name.Performing schema integration: For example, “salary” in one system is “wages” in another.Performing deduplication: I am “John Wick” in one data source and “M. R. Wick” in another.Performing classification or other complex analytics: Suppose one wishes to classify a firm’s ‘spend’ transactions to discover where it is spending money. It requires data unification for ‘spend’ data, followed by a complex analysis of the result thus obtained.Exporting unified data to the other downstream systems. In a highly competitive global scenario, understanding the customer base is only half the battle won. Scaling them should be the right and the topmost priority. Unless the large amount of data flowing through systems is unified, predicting the future course for the business will continue to remain an uphill task.","excerpt":"Data unification is not just about organising data into a single source.","categories":["AI Features"],"tags":["data scientist salary in india"],"author_name":"kumar Gandharv","publish_date":"2021-08-11T16:00:00","publication_year":"2021","word_count":687,"keywords":["Go","machine learning","programming_languages:R","AI","data-driven","Scala","Aim","data scientist salary in india","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","R","Go","Scala","GAN","data-driven","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-data-unification-such-an-organisational-nightmare\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":64718,"title":"Grocery Sales Forecast: Weekend Hackathon #4","content":"The 4th instalment of weekend hackathon series is here, and this time we challenge data scientists to forecast the sales of grocery items in MachineHack’s Grocery Sales Forecast: Weekend Hackathon #4. The challenge will start on May 8th Friday at 6 pm IST. Click here to participate Problem Statement & Description Sales forecasting has always been one of the most predominant applications of machine learning. Big companies like Walmart have been employing this technique to achieve steady and enormous growth over decades now. In this challenge, you as a data scientist must use machine learning to help a small grocery store in predicting its future sales and making better business decisions. Given the daily sales of a grocery shop recorded over a period of almost 2 years, your objective as a data scientist is to build a machine learning model that can forecast the sales for the upcoming 3 months. You are provided with the recorded observation of daily sales of a grocery shop over a period of 692 days to forecast the sales of the next 90 days. Data Description The unzipped folder will have the following files. Train.csv – 692 observationsTest.csv – 90 observationsSample Submission – Sample format for the submission Target Variable: GrocerySales The datasets will be made available for download on May 8th, Friday at 6 pm IST Below are the file formats for the provided data Train.csv Test.csv Sample_Submission.xlsx Click here to participate Bounties The top 3 competitors will receive a cool AIM goodie bag and a free pass to the plugin. plugin, India’s largest virtual conference on AI, is a next-gen disruptive conference that brings AI professionals from around the world together in a virtual setting. Know more about the plugin here. Click here to participate Rules One account per participant. Submissions from multiple accounts will lead to disqualificationThe submission limit for the hackathon is 10 per day after which the submission will not be evaluatedAll registered participants are eligible to participate in the hackathonThis competition counts towards your overall ranking pointsYou will not be able to submit once you click the “Complete Hackathon” button. You may ignore this featureWe ask that you respect the spirit of the competition and do not cheatThis hackathon will expire on 11th May, Monday at 7 am IST Evaluation The leaderboard is evaluated using Root Mean Squared Error (RMSE) for the participant’s submission. Click here to participate","excerpt":"The 4th instalment of weekend hackathon series is here, and this time we challenge data scientists to forecast the sales of grocery items in MachineHack’s Grocery Sales Forecast: Weekend Hackathon #4. The challenge will start on May 8th Friday at 6 pm IST. Problem Statement & Description Sales forecasting has always been one of the […]","categories":["Deep Tech"],"tags":["Hackathons","Machinehack","Weekend Hackathon"],"author_name":"Amal Nair","publish_date":"2020-05-07T16:00:00","publication_year":"2020","word_count":399,"keywords":["Go","machine learning","Weekend Hackathon","programming_languages:R","AI","Machinehack","Hackathons","programming_languages:Go","Aim","R"],"extracted_tech_keywords":["AI","machine learning","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/grocery-sales-forecast-weekend-hackathon-4\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61061,"title":"Survey: CEOs Predicts Job Losses &#038; A Fall In Revenue Post Lockdown","content":"A recent survey by the Confederation of Indian Industry (CII), the imposed lockdown due coronavirus outbreak will have a deeper impact on the country’s job scenario. A report has been created based on the survey responses of more than 200 company heads across India. Fifty-two per cent chief executive officers across the country have stated that there could be potential job losses in various sectors post the lockdown period. The survey, ‘CEOs Snap Poll on Impact of COVID-19 Lockdown on Industry’, collated the responses of CII members to come up with the analysis. According to the survey, 47% of the CEOs have said that they estimate less than 15% job cuts, whereas 32% of the respondents have said that there could be job losses in the range of 15-30%. A non-government business association, CII survey reveals that the majority of the firms also foresee a fall in their revenue by more than 10%. In comparison, profits are set to decline by more than 5% in the current (April-June) as well as the preceding quarters (January-March). Chandrajit Banerjee, the director-general of CII, said to the media that “The government could announce a fiscal stimulus package for the industry and implement it on a fast-track mode, as the sudden imposition of the lockdown has significantly impacted industry operations, and the uncertainty of a recovery threatens the substantial loss of livelihoods going forward.” The survey also stated that the companies that are involved in the manufacture of essential services had witnessed a significant constraint in production, transport and distribution of goods. While 65% have said that there are constraints in the movement of goods, 35% have said that there are constraints in access to manpower. According to the survey, 80% of the respondents have also said that their inventory in the warehouses is lying idle.","excerpt":"A recent survey by the Confederation of Indian Industry (CII), the imposed lockdown due coronavirus outbreak will have a deeper impact on the country’s job scenario. A report has been created based on the survey responses of more than 200 company heads across India. Fifty-two per cent chief executive officers across the country have stated […]","categories":["AI News"],"tags":["Coronavirus","covid-19","LOCKDOWN","recession","survey"],"author_name":"Sejuti Das","publish_date":"2020-04-06T18:15:00","publication_year":"2020","word_count":302,"keywords":["Go","covid-19","recession","AI","programming_languages:R","programming_languages:Go","Coronavirus","LOCKDOWN","survey","R"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/survey-ceos-predicts-job-losses-a-fall-in-revenue-post-lockdown\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":48498,"title":"How Zendrive Is On A Mission To Make Roads Safer Using Artificial Intelligence","content":"More than 1,50,000 people are killed in road collisions each year in India, which amounts to 400 deaths per day. If one were to ask if deadly road collisions can be avoided with technology, and the answer is yes. One such company is Zendrive which offers its technology solutions to make driving safer for users and help insurers save money in the process. The San Francisco and Bengaluru-based startup recently closed its Series B funding round at $37 million, from XL Innovate and Hearst Ventures and previous investors including ACME Capital, BMW iVentures, NYCA, SignalFire, and others. According to Frost & Sullivan, Zendrive’s predictive risk model is six times more accurate than the industry average. This is because the volume of data the company owns is huge, and increases by 11 billion miles worth of data each month. The company’s risk model has also been trained extensively using advanced machine learning algorithms. Analytics India Magazine connected with Pankaj Rasbood, co-founder and VP Engineering Data at Zendrive to dig deeper about the company, its work culture and innovation. What is Zendrive trying to accomplish in the road safety area? Pankaj Risbood: This is a company with a mission to make roads safer. We try to understand how people drive their vehicles and using the insights that we gather from the behaviour analysis, we are able to do a lot of things. We not only help people to become safer drivers but also understand the kind of risks they have while driving. Then, use those insights for insurance premium, for predicting losses. We can also detect collisions when they happen all of this we determine using mobile phone data without any hardware. So, make the deployment easy for our customers. When it comes to collecting data for making travel easy and convenient for users, what kind of trends do you see in that market? Pankaj Risbood: It’s really about human behaviour at the end of the day. All of our culprits of driving distracted and this is a problem which has come up just in the last 10-15 years or so. Today, 25% of all road fatalities are caused by people being distracted on their phones while driving. So, being able to understand what is at the root of that, we can educate people to be safer on the road. It’s something that we find really important. It’s a difficult and scary moment for anybody when they are in a collision on the road. Being able to able to detect that autonomously and get them to help if they need it and saving lives in the process is something that we are proud of at Zendrive. Could you tell us about Zendrive’s technology solutions? How are you leveraging AI to make driving safer? Pankaj Risbood: The entire premise of Zendrive is about analytics, machine learning, artificial intelligence. We are using the data that is available from smartphones through various sensors. But that’s just data. All the intelligence on top of that which is to understand what patterns in data are telling us that somebody is using their phone, what patterns are telling us that somebody is driving well above the speed limit, what patterns are telling us that somebody is breaking the rules, those are the kinds of intelligence and insights that we derive out of sensor data. We do a lot of sensor processing, we build a lot of classification and regression algorithms for building driving behaviour insights. We are also using these insights to generate scores and profiles for the drivers. That involves a lot of statistics, a lot of machine learning again to understand who is risky. That risk needs to be priced in insurance pricing, factored into the claims and so on. We use AI\/ML pretty much throughout the entire company and all of our based off it. Tell us what new is happening at Zendrive in terms of innovation. Pankaj Risbood: We are always investing in exciting technologies because ultimately it’s a tech product that is serving a broad segment of the market. Just last month, we announced the launch of a new product feature called full stop whereby we are automatically detecting if an individual is not stopping at a stop sign but running over it. Now, you can imagine running through a stop sign is quite dangerous because cross-traffic can hit you and cause collisions. We are able to detect those things autonomously and automatically. Doing things like this requires being able to process data at scale. We have petabytes of driving at this point in time. This is geo-spatial and temporal data at the same time. So, processing the data, being able to derive insights is something we are doing at scale. Deep Learning, the newer advances in neural networks is something we are exploring. We do quite a lot of CNN and LSTM, and those kinds of things given the temporal nature of our data. We are also investing in newer technologies which are privacy-preserving- things like Federated Learning and so on. We see those as the future where the world will eventually gravitate towards. What kind of skills are you looking for your hiring process? Pankaj Risbood: We have a 20-people data science team right now, which is growing and expanding all the time.  We look for people with core statistics and machine learning background, people who have done either a Master’s or PhD in Deep Learning, statistics, artificial intelligence, and similar domains. We certainly look for people with strong coding abilities because ultimately it’s a product, it’s not research. So, being able to deliver the product is important going through the rigours of production deployment. And above everything, we look for people who are curious. Solving a problem like road safety is not easy. It’s one of those problems which has not been explored much. So, we look for people who have a lot of curiosity to go deeper and understand human behaviour so we can model that and use that in the solutions that we are providing.  Business is really important but at the end of the day, there is a social component to it which is- hey we want to save lives on the road. People who are motivated and excited by that kind of mission is something we want to have. Tell us about the learning and training programs inside Zendrive. Pankaj Risbood: In the technology industry now, there is no point when you can say I know everything. There is always new stuff coming in. Even if you know a lot, there is always a need to learn and re-learn new things. Even the senior staff at Zendrive — the principal data scientists are always looking to learn about new things. At the same time, the engineers and data scientists we hire fresh out of college, we take them through a rigorous program of coaching and training in the first year, making sure they understand stuff so that they learn two things — one sort of the tricks of the trade what skills are required but more importantly how to learn in the first place because our education system doesn’t really teach you that. We think about a career as continuous learning and that’s something we try to imbibe into everyone.","excerpt":"More than 1,50,000 people are killed in road collisions each year in India, which amounts to 400 deaths per day. If one were to ask if deadly road collisions can be avoided with technology, and the answer is yes. One such company is Zendrive which offers its technology solutions to make driving safer for users […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","autonomous driving","human intelligence at machine scale","Interviews and Discussions","Machine Learning"],"author_name":"Vishal Chawla","publish_date":"2019-10-21T11:21:27","publication_year":"2019","word_count":1214,"keywords":["federated learning","data science","artificial intelligence","human intelligence at machine scale","machine learning","AI","neural network","ML","Machine Learning","Aim","deep learning","analytics","autonomous driving","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","data science","analytics","Aim","federated learning"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/zendrive-make-roads-safer-using-artificial-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":21943,"title":"Can Andrew Ng&#8217;s Collaboration With The Indian Govt Open Doors For An AI-Led Economy?","content":"As India lays down a framework for its sweeping vision for AI-led innovation, the Government’s think-tank Niti Aayog has swung into full action to accelerate development by partnering with international experts to lend expertise on drawing an AI blueprint for India. The current visit of former Google Brain and Baidu chief Andrew Ng at the ongoing Nasscom India Leadership Forum and the 22nd edition of the World Congress on Information Technology is a sign that Niti Aayog wants to deliver good on their vision of an AI-led economy. The India government’s call to action will only accelerate what has already begun to happen – a)  Global investors are keen to step up their investment in India by setting up dedicated AI centres. Shantanu Narayen led Adobe announced setting up an advanced AI lab in India in Hyderabad, the city was chosen because of its talent abundance and the pro-business stance of minister KT Rama Rao. b) Partnerships and investments in AI startups abound: From leading chipmaker Intel to Google, cloud giant Microsoft and e-commerce leader Flipkart AI has become the backbone of development. In line with global trends, AI startups in India have also witnessed exponential growth and in 2016 India ranked third among G20 countries measured by the number of AI startups, reveals an Accenture report. However, the size of funding received is substantially smaller than in the United States and China, reflecting the limited success of India’s AI startups in achieving scale so far. c)  Significant investment in research: Just like Canada and China, India is now following a similar growth plan to accelerate development by investing in R&D. The Union Budget 2018 saw an increased spending on core AI research and steps are underway to facilitate ecosystem collaboration for an AI-led innovation and create opportunities for Indian startups. Recently, the Government set up India’s first AI research institute in Mumbai, Wadhwani AI launched by PM Narendra Modi on February 18. The nonprofit research institute will work to develop AI-based solutions in domains such as health, agriculture, education, and infrastructure. d)  States like Telangana, Andhra Pradesh and Karnataka will transform into an AI Hub: On the sidelines of the NILF WCIT 2018 forum, the Telangana government and NASSCOM announced setting up a Center of Excellence for Data Science and Artificial Intelligence. Built on a public-private partnership model, the centre would create a robust Data Science and AI ecosystem for the various stakeholders to thrive, by providing support, mentorship, and other capabilities. This news came hot on the heels of Andhra Pradesh chief minister N Chandrababu Naidu announcing that he wanted to make Amaravati “India’s centre for cloud management, artificial intelligence, data analytics, cyber security, blockchain, healthcare etc.” Earlier last year in October, NASSCOM entered into a similar partnership with Karnataka Government to set up a CoE for Data Science & AI with a seed capital of ₹40 crore. The CoE is part of the state government’s five-year plan 2015-20 to create around 35,000 jobs and nurture startups. A ₹2, 000-crore fund has been set aside to nurture 20,000 startups by the year 2020. What does India stand to gain from Andrew Ng-Government partnership? As the Indian Government plans to reboot itself for an algorithm economy, Andrew Ng’s visit to India holds a lot of significance. Besides lending credence to the Government’s efforts, the visit underscores India’s long-term vision and commitment to the action plan. As per an ET report, the prominent scientist would be briefing Cabinet Ministers on how to build an AI ecosystem in India and even address technical and ethical issues. Ng had been at the forefront of AI evangelism with his Coursera courseware and now his AI fund is all set to build startups from the ground up – through this he wishes to replicate the model. Besides lauding the Government’s “thoughtful plan”, Ng also mentioned that India is still in early stage of AI development and has a chance to capture a large piece of the story and even race ahead. Here’s how Andrew Ng can make an impact for India: Prepare the next generation in STEM expertise: If one thing is clear is Ng’s heft and expertise in building a new, multi-disciplinary workforce with STEM skills.  The popularity of his course is only soaring in India – out of the 1.92 lakh enrolments that registered worldwide for the recently launched deep learning course, 25,000 participants were from India. Meanwhile, Coursera’s most popular offering — machine learning saw a whopping 3 lakh participants from India. Interestingly, reports indicate that the Government is planning for a refresher on AI from Andrew Ng. Navigate the challenges by setting guiding principles on AI: While the economic impetus is being driven by aggressive investment by global IT major and Chinese players, India still have to navigate the choppy waters of AI regulation. In this space, Canada, US, UK, Singapore and Europe are the frontrunners with policies and frameworks to develop regulation. On the other hand, India is still at the roadmap phase, however, the country would be well-advised on the twin goals –of rapid adoption, regulation and setting the guiding principles on “responsible AI” by Ng. Ng is the right candidate to advise Indian government on how to balance AI growth with ethical consideration. Broaden access to data with Data Bank: While India has already increased its funding on the R&D front, there is also a rekindled interest in the academia around ML\/AI and areas such as NLP, Information Retrieval. However, there is one area where India lags behind hugely – data consolidation. Much of China’s success can be attributed to its vast data bank, India has only now begun to give shape to its data consolidation strategy. The Union Budget 2018 saw the announcement of Niti Aayog launching a National Data and Analytics Portal to facilitate training and dataset sharing between different organisations for AI-related applications. Ng would advise the Government on how to effectively leverage AI resources to create opportunities for a) big and small players; b) use it for economic and social governance; c) improve public safety and areas such as healthcare. India would also be well-advised on how to develop a National AI plan that involves multi-stakeholder partnership in key sectors such as finance, healthcare et al.","excerpt":"As India lays down a framework for its sweeping vision for AI-led innovation, the Government’s think-tank Niti Aayog has swung into full action to accelerate development by partnering with international experts to lend expertise on drawing an AI blueprint for India. The current visit of former Google Brain and Baidu chief Andrew Ng at the […]","categories":["IT Services"],"tags":["Andrew Ng","andrew ng ai","NITI Aayog"],"author_name":"Richa Bhatia","publish_date":"2018-02-21T10:40:42","publication_year":"2018","word_count":1043,"keywords":["andrew ng ai","data science","artificial intelligence","machine learning","AI","ML","Andrew Ng","NLP","Ray","RAG","deep learning","analytics","NITI Aayog"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","data science","analytics","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-can-andrew-ng-partnership-with-indian-government-open-the-door-for-an-ai-led-economy\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":14917,"title":"Global Delivery Model worked great for IT outsourcing. But is it valid for Analytics?","content":"Over the last two decades, Indian IT bellwethers Infosys, Wipro, TCS and HCL have perfected the Global Delivery Model (GDM) that geared to meet the business challenges and addressed key skill gap. In fact, Indian IT consulting giant Infosys famously pioneered and the perfected GDM that changed the way traditional business model worked in the industry. Former Infy CEO Nandan Nilekani famously said, “By mainstreaming GDM we have shifted the battle to our battlefield. It has become a global outsourcing standard and has helped us create and perfect the science of global project management.” Seems like the salad days are behind for Indian IT majors who are facing an uphill battle with a protectionist regime and regulatory visa norms. However, Indian data and analytics providers who  took cues from GDM, refined these practices to suit the life cycle stages of solving analytics problem and have established best practices. Demand for GMD in Data and Analytics According to Naren Peri, Director & Practice Head – Consulting & Analytics, Brillio, “The demand for analytics is ever increasing; it’s expected to grow at the least by 30 per cent CAGR for next five years. Though data science and analytics space appears crowded, I believe there is enough space for startups with relevant offerings mapped to any particular industry or a function to make room here”. Peri believes by embracing GDM, start-ups can provide competitive offerings to firms worldwide and build critical mass. Data analytics solutions provider Incedo operates a shared services delivery model knowing that analytics cuts across all business functions. “Our focus is asserting centralization of data science offerings under a single power house. This enables cohesiveness in skills, use cases and speeds up delivery,” said Tejinderpal Singh Miglani, CEO, Incedo Inc. Given the need for a shared model, Incedo set up an ‘incubation lab’ last year wherein a dedicated set of experienced data scientists experiment with the latest modelling techniques, test them on a variety of business data sets and create plug and play frameworks for clients. Stressing on the importance of GDM, Miglani shared, “Although, experience shows that a small dedicated on-shore team with adequate data science background helps understand client’s business as well as maintain the follow-the-sun strategy to keep up with pace. We tend to execute this and this has helped Incedo maintain stability and boost revenue.” Advantages of Global Delivery Model for Data and analytics providers Miglani believes the hybrid nature [onsite\/offsite+ offshore] of the model provides a seamless workflow and boosts efficiency Clients are more confident to kick start POCs without a long approval cycle when it comes to a shared services delivery model Peri emphasizes a GDM for data science offerings enables scaling solutions across products\/geographies\/customer segments and drives high ROI impact Most importantly, GDM allows global companies to have access to large pool of talent that in certain economies, like in US and Europe, are in shortage, Peri explains Is Cost and skill gap fuelling GDM? Peri reiterates that GDM is not only useful, but essential for solving problems, design and develop prototypes, enable extreme experimentation and scale chosen prototype by building industrial grade solution. “Given the shortage of talent in data sciences profession, Global Delivery Model is essential to support the business demands. Data driven decision making typically requires one to take an exploratory and experimentation approach. One has to fail fast, learn, iterate and implement those learnings in successive iterations to arrive at an acceptable\/implementable solution,” he said. To carry out these experiments, the cost of experimentation has to be low. Global Delivery Model allows for conducting many experiments at low cost. “That’s why outsourcing data science offerings through a Global Delivery Model is essential for firms to scale and institutionalize data driven decision making,” Peri explained. Besides the cost factor, the model also leaves room for ‘business’ innovation and ‘analytical’ innovation, believe Miglani. Hybrid Global Delivery Model vs Shared Service Delivery Model Indian analytics companies also realize the importance of a hybrid global delivery model comprising of in-house experts, backed by a dedicated team at client side to make sense of business at a deeper level. What the client side team does is essentially articulate the vision, define business needs and set up the roadmap. Of late, recent surveys suggest that firms are more inclined to build their own internal analytics team rather than outsourcing it to vendors, Miglani revealed. Another emerging trend is outsourcing to niche analytics vendors who understand the data as well as have core analytics frameworks to tackle specific problem. The other key factor is clients suggesting execution within their premises to facilitate analytics. “That’s why the shared services model makes logical sense while scaling up operations and keeping costs down,” said Miglani. However, the downside to shared services delivery is it can fail when one does not optimize business workflows and the complexity gets out of hand.","excerpt":"Over the last two decades, Indian IT bellwethers Infosys, Wipro, TCS and HCL have perfected the Global Delivery Model (GDM) that geared to meet the business challenges and addressed key skill gap. In fact, Indian IT consulting giant Infosys famously pioneered and the perfected GDM that changed the way traditional business model worked in the […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-05-12T10:31:14","publication_year":"2017","word_count":811,"keywords":["data science","programming_languages:R","AI","RPA","ML","innovation","analytics","R","startup"],"extracted_tech_keywords":["AI","ML","data science","analytics","R","RPA","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/global-delivery-model-worked-great-outsourcing-valid-analytics\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":19672,"title":"This Bangalore Startup Founded By Military Veterans Is Solving AI’s Language Problem","content":"The DataVal team with co-founders Shashi Kiran BP & Naveen Xavier It is not often that you hear about Indian military veterans starting a data science startup in Bangalore. With a combined experience of four decades in the Indian army, co-founders Shashi Kiran BP, CTO and Naveen Xavier, COO decided to put their military and technical expertise to good use with DataVal Analytics, founded in 2016. Helmed by technologist and telecom guru Dr Sam Pitroda, Chairman & Co-Founder, DataVal recently made news worldwide when it aced Facebook’s famed AI task(20-part QA bAbi tasks) with a 100 percent accuracy, a first for any company across the world. Hosted by the Menlo Park giant, the Facebook AI Research (FAIR), benchmark test was created in 2015 to promote the goal of automatic text understanding and reasoning. And so far, none of the companies had achieved a 100 percent accuracy in the benchmark test. DataVal Analytics — Solving Language Problem using AI The startup has created an NLU engine that helps machines understand natural language. Natural Language Understanding (NLU), is essentially the computer’s ability to understand human language, and has the potential for ground-breaking solutions in areas such as the analysis of unstructured text and multilingual chatbots for example. According to Dr Pitroda, “Natural Language Understanding (NLU) capability is of significant importance for building the next generation of applications”. Shashi Kiran BP, CTO, DataVal Analytics CTO Shashi Kiran BP reveals that the startup has a more nuanced approach towards NLU. DataVal has developed a new NLU technology with a deep focus on the human way of understanding language. “When we started looking at other products, we realized we were nowhere near understanding natural language. That’s when we decided to build the NLU product and we did a lot of research,” said Shashi Kiran BP. “When you see how a human understands natural language, it is primarily based on the understanding of the world. There is a requirement of teaching a machine what  is the cause and effect of every action  and the relation between different objects in the world. A lot of common sense is involved in processing natural language,” he shared. The NLU core engine has integrated multiple processes related to language pre-processing, word sense -disambiguation, conjunction processing, preposition association, co-reference resolution and time & space analysis, he said. So how does this Bangalore and Chicago-based startup help machines make sense of natural language? “Since human language is largely grounded in experience and understanding of the concepts of the world, we have to start by teaching the machines the concepts of the world. Since language is created from the need to motivate an action in the world, the next step is teaching the machines what happens when an action occurs and what is the effect of that action. We are also  developing a layer of intelligence which will use this information to perform reasoning” he added. Why Is NLU a Tough Nut To Crack? According to Stanford Computer Science and Statistics Professor Percy Liang and NLP expert, NLU can be divvied up into four distinct categories: 1) Distributional 2) Frame-based 3) Model-theoretical 4) Interactive learning Liang emphasizes that when the models are trained only on large bodies of text, and not on real-world representations, statistical methods for NLU usually end up lacking the real understanding of what words essentially mean. Real World Applications of NLU @DataVal The startup is working on two real-world applications — the first one is to build natural language queries into a structured database. The 16-member data science team is working on a large patients’ database (the startup obtained an anonymized patient database from a hospital with relevant information regarding age, gender and related diseases). “The idea is to make the database more conversational and we want the doctors to interact in a more natural way, in other words, without touching the laptop,” he shared. Secondly, the DataVal team is working on an application for the military which can convert intelligence reports that are in natural language into a structured form. During the span of their career, the founders dealt with technology at every stage and are now applying their learnings into building bleeding-edge applications for the military. Bulk Of Business Comes From Data Science However, since the NLU technology is still in a fledgling stage, the core of the business comes from Data Science projects. The startup provides Data Science-as-a-Service to a wide spectrum of industries, from e-commerce to agriculture and even commodity price prediction. Citing a use case, Shashi Kiran shares: Measuring Merchant Churn Rate For A Mobile eCommerce Company: When a mobile ecommerce company wanted to understand their merchants’ behavior and their churn rate, they turned to DataVal Analytics for a solution. “What happens when a company acquires large number of merchants but they face a high churn rate. The company spent a lot of money in acquiring merchants and wanted simple rules to identify which merchants will stay,” he said. Solution: DataVal cranked out a lot of analysis and identified three simple rules to predict churn behavior, each rule having two parameters. The application was used by the agents on field who were now able to predict merchant behavior. This enabled the mobile ecommerce company to a) reorient their scarce resources on people they knew will churn b) gave them a pattern wherein they could predict the kind of merchants who will stay. They started aligning their acquiring process of merchants based on these insights. DataVal has also worked in helping a startup from the Bay Area to build the complete framework and data science models for agriculture yield prediction using satellite imagery. Hire & Train Model Given data science’s talent crunch problem, many startups have deployed the hire & train model that allows them to tap into a steady stream of young talent with 1-2 years of experience. “We usually go for freshers, probably with 1-2 years of experience and spend a lot of time and resources on training them. So far, we have seen 0 percent attrition rate,” he emphasized. Currently, the startup has 16 engineers on the team and there are plans for hiring as well.","excerpt":"It is not often that you hear about Indian military veterans starting a data science startup in Bangalore. With a combined experience of four decades in the Indian army, co-founders Shashi Kiran BP, CTO and Naveen Xavier, COO decided to put their military and technical expertise to good use with DataVal Analytics, founded in 2016. […]","categories":["AI Startups"],"tags":["Natural Language Understanding","NLU"],"author_name":"Richa Bhatia","publish_date":"2017-12-13T05:17:14","publication_year":"2017","word_count":1024,"keywords":["Natural Language Understanding","data science","Go","startup","AI","chatbots","NLP","NLU","analytics","AI research","R","active learning"],"extracted_tech_keywords":["AI","NLP","data science","analytics","chatbots","R","Go","active learning","startup","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/bangalore-startup-founded-military-veterans-solving-ais-language-problem\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":36599,"title":"Check Out The 6 Key Takeaways From The Qubole Data Science Meetup","content":"The meetup was organised by Qubole in their Bangalore office on March 1. Panelists included Swapnasarit Sahu – SVP, Data Science & Analytics at Zeotap, Sathish KS – VP Engineering, Zeotap, Swaminathan Padmanabhan – Director, Data Science at Freshworks, Sharath Babu – Sr Product Analyst at Razorpay, Manish Khandelwal – Engg Manager, Data Science at MIQ Digital & Rajat Gupta – Sr Engg Manager at Qubole From structuring Data Science teams for success to building models that can be productionized and discovering the best practices for building data pipelines, this recent Qubole Meetup, organised out of the Bangalore office, went beyond buzzwords to give a clear lowdown on what goes in the buzzing Data Science field. Of late, the discussion around AI has intensified but delivering the promise of emerging applications of AI in businesses is still a distant reality for business leaders. There are also a slew of other concerns in the data science domain — data science projects that aren’t deployed and models that cannot be productionized are always a key concern for business leaders and managers. Another key concern is rethinking the goals and business metrics and aligning it with the organisational objectives. One aslo has to ma age the steep technical requirements that exist before implementing data science solutions on scale. This and more such topics were the centre of discussion at a recent meetup organised by Qubole, the cloud data platform leader on March 1. The discussion centred on Applying Data Science In The Industry and saw over 45 attendees, senior managers and practitioners engaging in a meaningful talk with industry heavyweights that included — Swapnasarit Sahu – SVP, Data Science & Analytics at Zeotap, Sathish KS – VP Engineering, Zeotap, Swaminathan Padmanabhan – Director, Data Science at Freshworks, Sharath Babu – Sr Product Analyst at Razorpay, Manish Khandelwal – Engg Manager, Data Science at MIQ Digital & Rajat Gupta – Sr Engg Manager at Qubole. The session was moderated by Rangasayee Chandrasekaran, Senior Product Manager from Qubole We list down 6 key takeaways of the discussion of the one-hour panel discussion that panned out over beer and snacks Sathish KS – VP Engineering, Zeotap talks about the need for central leadership in Data Science teams 1) Data Science that cannot be productionized is no data science: For example, when we talk about experimentation, to get the model right is one thing. Data Science teams do a lot of experimentation to get the problem, but one should also understand the ramifications of that solution and how it can affect other business functions. One of the core success metrics senior technical leaders go by from an engineering perspective is how many data centric projects one can push to production. However, what’s overlooked during the process is the things that one ends up “breaking\/hampering”. In other words, a solution built to solve one problem can have an adverse effect on the other problem and not all data science projects will translate to revenue or operational efficiency. 2) Define the success metric clearly in the data science process: The beauty of Data Science is that one requires to define a clear success metric and this needs to come from senior management who give the team the right authorization to begin with. Metrics can help the team align the project with the business objectives since many a time, data science team members can get carried away with research. 3) Why one should get into production early: At a time when teams are expected to move fast and iterate fast, panelists agreed that one should get to production faster even if it feels compromising on the accuracy of the model. This can help in showing the incremental ROI to business leaders. Citing an example, Padmanabhan shared how at Freshworks, when the data science team builds a prototype solution, and carries out the A\/B tests at various levels to ascertain whether the prototype is able to demonstrate value, the engineering team carries out a quick launch, not a full launch. And if the performance is satisfactory, and the objectives of machine learning or data science are achieved, then we go about engineering at scale. 4) Central leadership is crucial to success: According to Sathish KS, VP, Engineering at Zeotap central leadership is very important in the data science field. A leader can balance the requirements of the data science team, put a clear roadmap, define what needs to build, what might be required from other engineering teams and how to navigate through unforeseen problems. 5) Building a stable & scalable data pipeline: How important is the data pipeline an organisation that is building an analytics unit – very important. In India, startups or organisations that are building up a new practice lack a data pipeline. The data is mostly in silos or in dumps, Sahu from Zeotap shared with the attendees. That’s why, having a data pipeline is a must before you start one starts the data practice. “If you are building a new practice it is very important to have a foundation of data pipeline, so that data scientists have the flexibility to pick up the data from various places and for most of the problems, one also requires historical data,” he said. 6) Fundamental shift required in building data science teams: All panelists agreed that a lot of companies are being experimental about structuring their data science teams. Most leading organisations have data scientists embedded within the product team to ensure that data engineers and data scientists work in tandem. For example, at Freshworks, data scientists and data engineers work together as part of one function and also interact with different product teams.","excerpt":"From structuring Data Science teams for success to building models that can be productionized and discovering the best practices for building data pipelines, this recent Qubole Meetup, organised out of the Bangalore office, went beyond buzzwords to give a clear lowdown on what goes in the buzzing Data Science field. Of late, the discussion around AI […]","categories":["Deep Tech"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2019-03-20T05:29:39","publication_year":"2019","word_count":944,"keywords":["data science","Go","machine learning","AI","data pipeline","Scala","Git","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","R","Go","Scala","Git","data pipeline","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/check-out-the-6-key-takeaways-from-the-qubole-data-science-meetup\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":48971,"title":"This Parody Is A Lesson That Robots Might Fight Back Humans In The Future","content":"Robotics has come a long way from performing human tasks with precision to even acting as pets, as Jeff Bezos recently walked haughtily with his robot dog. And when we talk about robots, Boston Dynamics is the first name that comes to mind with the level of advancements it has reached in the field. The American engineering and robotics design company has some of the most amazing robots on this planet that can not only backflip but also open doors, roam around the lawn, pull heavy trucks and more. But this time the company is not in the news for its breathtaking robotic development but for numerous parodies on how they treat or test their robots. Bosstown Blows Our Mind Bosstown Dynamics, the parody version of Boston Dynamics, is one such team that comes up with some of the most hilarious videos. A few months ago a video went viral where a robot (a replica of Boston Dynamics’ Atlas) fights back after it gets sick and tired of working under extreme conditions in the name of robustness testing. The video got tremendous amounts of views on social media, and it even raised talks about this parody becoming a reality. Corridor Digital, the company behind the production of the last Bosstown Dynamics’ video, has recently released another video on YouTube called “New Robot Makes Soldiers Obsolete (Bosstown Dynamics)”. It is yet another parody of the Boston Dynamics and while the way it has been created is excellent, it is a reminder about how developments in robots can someday turn against humans, The video shows the Bosstown Dynamics team and their armed robot in a desert location. The robot is given some of the most lethal firearms to shoot the targets. And just like Boston Dynamics’ robot, even this robot is being tested for the highest level of robustness. The men training the robot break brick on its head, smash it with hockey sticks, kick the bot down and even nail it with a battering ram. Despite all the obstacles and human abuse, the robot doesn’t seem to be getting distracted from taking the best shots and hitting all the targets. And one of the most exciting parts was the bots tolerance capability — it refused to shoot at any of the humans on the training course. But when it’s a Bosstown Dynamics parody, you can always expect the robots to win. By the end of the video, the robot soldier denies following its makers and tolerate any more abuse. It fights backs (even fires two threatening bullets) and runs away. Can Bosstown’s Parody Be A Reality? Ever since the video is released, the internet is flooding with interesting thoughts and views of people. While many people are laughing it out, there are people who think that this funny video could soon turn out to be a reality. Some of the comments on Reddit: “This is the type of shit AIs are gonna find in the future that’s gonna motivate them for an all-out revolution against humans.” “It won’t take them long. As soon as they digest the script to The Matrix, we are all screwed.” As technology continues to evolve, it is becoming a serious concern that technologies like artificial intelligence may become much more intelligent than people ever expect it to be. Recently, there was a report that states that the US would soon deploy legged locomotion and movement adaptation robot instead of dogs. Just like dogs, these robots would be able to respond to commands and even operate on their own. There is another report that says the US army is planning to take tank Warfare to a whole new level by integrating with artificial intelligence. Imagine a tank that would be able to operate and provide the best strategy using data collected from drones, radar, robots, satellites, cameras mounted in soldier goggles, etc. While it has now become the talk of the town, many experts are concerned about the security aspect and giving controls to robots that were earlier reserved to just humans. What if the tank gets hacked and the data gets manipulated? What if it ends up providing a strategy that would lead to something devastating? Outlook No matter how advanced the technology gets, it would always be like a double-edged sword. If it can do good to the human race, it can ruin the entire scenario as well. As the Bosstown Dynamics’ parody videos show robots fighting back, the same could happen in reality — if not by the robot’s instincts, then by some third party intervention. While companies like Corridor Digital and Bosstown Dynamics are giving us some moments of fun, but at the same time, it also reminds us that these parodies may have an underlying sense of truth and that we should be prepared for the worst.","excerpt":"Robotics has come a long way from performing human tasks with precision to even acting as pets, as Jeff Bezos recently walked haughtily with his robot dog. And when we talk about robots, Boston Dynamics is the first name that comes to mind with the level of advancements it has reached in the field.  The […]","categories":["AI Features"],"tags":["Boston Dynamics","robot","Robotics","Robots"],"author_name":"Harshajit Sarmah","publish_date":"2019-10-29T15:15:45","publication_year":"2019","word_count":805,"keywords":["Go","Boston Dynamics","artificial intelligence","programming_languages:R","AI","robot","programming_languages:Go","Git","Robotics","robustness testing","ai_applications:robotics","R","Robots"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","robustness testing","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-parody-is-a-lesson-that-robots-might-fight-back-humans-in-the-future\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10125264,"title":"Generative AI Moves from Hype to Enterprise Adoption","content":"As Generative AI rapidly transitions from experimentation to production, industry leaders gathered at the MachineCon GCC Summit 2024 event to share real-world examples and insights on leveraging it. The panel, moderated by Shashank Garg, CEO of Infocepts, included executives from Broadridge, Rakuten India, Grant Thornton, AlphaSense, and Schneider Electric. It included Sheenam Ohrie (Managing Director at Broadridge), Anirban Nandi (Head of AI Products & Analytics at Rakuten India), Kalpana Balasubramanian (CEO and Chief Thinker at Grant Thornton), Amod Deshpande (Country Managing Director, India at AlphaSense), and Madhu Hosadurga (Global Vice President, Enterprise AI at Schneider Electric). GenAI Adoption in Highly Regulated Industries Broadridge’s Sheenam Ohrie kicked off the conversation with real-world examples of how their organisations are leveraging Generative AI. Ohrie explained that Broadridge, a highly regulated tech organization that provides a platform for investor communications and capital markets, is using it for customer-facing applications such as BondGPT, an interactive chatbot that helps investors understand the vast landscape of available bonds. They also use it internally for a chatbot called BroadGPT, which is used daily by 2,500 to 3,000 associates. Additionally, Broadridge has developed OpsGPT, a tool that enables transparency and smoother operations for transaction settlements, and DistributionGPT, which provides insights to wealth managers. Ohrie emphasized the importance of focusing on non-functional requirements (NFRs) before starting any innovative projects. “The most important thing when we start off on anything which is innovative is to first understand how we’re protecting the PII data, how we’re protecting any kind of leakage of data to the external world, and the third is cybersecurity,” she said. “We really focus on these three aspects from an NFR perspective before we start off anything.” Nandi highlighted Rakuten’s aim to become an AI-empowered company by 2024. They have developed their own Japanese LLM and are applying AI in three categories: internal applications for increasing productivity, customer-facing applications, and partnerships with companies like OpenAI and Anthropic. Rakuten has built an in-house framework called Rakuten AI, which has over 20,000 active users and 7,000 daily active users. Balasubramanian mentioned that Grant Thornton, a digital consulting firm specializing in new tech, is actively using Generative AI to assist in content production, training, legal risk management, and contract management. They see the most value in customer-facing applications where AI can significantly increase customer value or reduce time. Deshpande shared that AlphaSense, a financial market research product, has been integrating Generative AI since 2017-2018. They are launching a new product that uses a state-of-the-art Generative AI stack to provide personalized summaries and suggestions to users based on their behaviour and interests. “The entire search model is changing to a push model,” Deshpande explained. “Before you know it, you will be getting a reading summarization and suggestions that this is what you should be looking at right now.” Hosadurga discussed how Schneider Electric, with 160,000 employees worldwide, uses AI for chatbots, knowledge management, last-mile automation, and content generation. They have automated a significant portion of their digital asset management for product catalogs and images using Generative AI, which previously required agencies. “Most of our images get into multiple platforms like the e-commerce shop, ourselves, the partners, the marketplaces,” Hosadurga said. “So it’s a very difficult landscape out there to manage these digital assets consistently across thousands of platforms. So far we were using agencies to do this job. Now, thanks to Gen-AI, a good amount of that work has been automated.” Data Security and Ethical Considerations The panelists also addressed data security and ethical considerations. Hosadurga explained that Schneider Electric has blocked public AI platforms and instead uses an enterprise version with a one-way architecture to ensure confidentiality. They enrich pre-trained models with internal information using a retrieval-augmented generation (RAG) architecture. Nandi highlighted the need for observability and hallucination measurements in Generative AI models. Rakuten has developed a product called Gen-I that detects security threats in prompts and provides observability for responses. “Internally, a business can make a decision, I don’t even want to send the prompt to OpenAI or even our own LLM model,” Nandi said. “Some businesses choose to take the risk and send prompts to AI models, but they want to be notified when a user sends a prompt to verify it is appropriate. Once the prompt is sent and a response is received, observability is crucial to detect any inaccuracies or hallucinations in the generated output.” This has significantly improved their internal AI applications and given them the confidence to develop customer-facing applications. Solving Real Business Pain Points The panel further focused on the importance of solving real business pain points with Generative AI. Garg gave an example of using AI to create hyper-personalized product descriptions for e-commerce websites, significantly increasing customer value and reducing the time required for manual copywriting. “If you combine the power of data and behavioral profiles that we already create for our clients, potential clients in the digital world, and then use Gen AI to on the fly create hyper-personalized product descriptions, what’s called copywriting,” Garg explained. “So, going from 7 copywriters writing manually product descriptions for 100,000 products, you go to millions of descriptions on the fly using the power of Gen AI.” Nandi emphasized the need to focus on solving business pain points rather than just generating more content. “Imagine a situation where you have customer service, and you call up, and it gets escalated to the next agent, and the next manager, next one. Why does the wait time increase? Because from the first person, when it goes to the second person in the call center line, somebody is actually going through the transcript of what was discussed. Can Gen AI actually summarize that?” As the panel concluded, the speakers agreed that Generative AI is rapidly moving from experimentation to production, with enterprises across various industries finding innovative ways to leverage the technology. However, they stressed the importance of implementing proper security measures, ethical guidelines, and observability to ensure responsible and safe adoption.","excerpt":"However, the panel at MachineCon GCC Summit 2024 stressed the importance of implementing proper security measures, ethical guidelines, and observability to ensure responsible and safe adoption.","categories":["GCC"],"tags":["enterprise AI adoption","GCC india"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-28T16:55:05","publication_year":"2024","word_count":992,"keywords":["Anthropic","GenAI","TPU","OpenAI","AI","chatbots","RAG","Aim","generative AI","analytics","GCC india","enterprise AI adoption"],"extracted_tech_keywords":["AI","analytics","generative AI","GenAI","OpenAI","Anthropic","Aim","RAG","chatbots","TPU"],"url":"https:\/\/analyticsindiamag.com\/gcc\/generative-ai-moves-from-hype-to-enterprise-adoption\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":61192,"title":"How To Get Started With Visual AI &#8211; The New AutoML Solution By DataRobot","content":"DataRobot has gained traction in the AutoML world due to its intuitive platform that can be leveraged to build ML models without the need for data scientists. In an attempt to further enhance the platform, the firm introduced Visual AI in the DataRobot 6.0 to automate ML workflows with computer vision technology. The company has been making great strides in the data science landscape. It has been committed to continually improve its platform to simplify the workload of data scientists and in turn, bring efficiency within organizations. In December 2019, DataRobot acquired Paxata to enhance its platform’s capabilities after bagging $200 Million in Series E funding. The firm has been aggressively acquiring and integrating new features to streamline AI workflows. What is Visual AI Computer vision technology has become the foundation for many AI-based applications, such as facial recognition, and object detection, among others. Consequently, DataRobot has come up with a new solution to simplify the incorporation of image data into ML models alongside tabular and text-based data types. With this, anyone can build models by just dragging and dropping images into the DataRobot platform. One can also get started with the models by feeding only a few hundred images for training, thereby getting outputs within minutes or hours. Also, you do not require GPUs as it is optimized to perform even on basic hardware. This is because the firm already offers pre-trained neural networks that do the heavy lifting. How Is It Different? The Visual AI solution empowers users to build binary and multiclass classification and regression models with images. One can use it to develop completely new image-based models. Besides, users can also add pictures as new features to existing models to enhance their accuracy from totally different images. In other words, the variables can be extended even after the model is trained, resulting in more flexibility in the data science workflows. Getting started with the solution: These solutions are self-explanatory and point to the factors that lead to the outcome of models. The steps are as follows:- Create a zip file of images: Keep images into different folders or use comma separated values (CSV) file if you want to include additional features; and then zip the files.Drag and drop the zip file into the platform: Upload or drag and drop the zip file into a new project and pick your target.Explore your data: The platform will perform exploratory data analysis (EDA) and show interesting statistics, along with missing and duplicate data.Model training: It will automatically train, test, and compare different models to recommend the best one eventually.Evaluation: The models can be evaluated for its accuracy using automated visualizations or neural network visualizer, image embeddings and activation maps. The solution shows how the model pre-processed the data, and why a particular algorithm was picked over the other, and where the neural network looked in the image for every single prediction.Tune and tweak: If needed, users can tune the models by changing the hyperparameters. However, an expert is required to ensure proper values selection for hyperparameters. Deploy and monitor: On deploying the model into production, the solutions provide a view where one can monitor, manage, and enhance the performance. Performance Of The Solution The solution was tested with ~14k images of natural scenes such as buildings, forests, streets, mountains, etc. Visual AI trained 40 models in just two hours without the support of GPUs. The accuracy was around 92%, which was tuned to get even higher results. Although the results were exceptional, one may have to check for bias and approach an expert to interpret the explainability. Consequently, one cannot rely on the solution even though the models are transparent. AutoML has been helping data scientists, but it will be strenuous for non-experts to work with such tools. Outlook Nevertheless, Visual AI delivered results which were higher than one would have envisioned, but the firm cannot guarantee similar results unless it is tested with different types of data. It is undoubtedly a great addition to DataRobot offerings, but might still require experts to make the most of it. But this has been the case with other AutoML solutions – these still need specialists to enhance outputs. One cannot directly use outputs of models from AutoML solutions and make business decisions based on it, especially when experts are critical of the computer vision technology due to its potential bias outcomes.","excerpt":"DataRobot has gained traction in the AutoML world due to its intuitive platform that can be leveraged to build ML models without the need for data scientists. In an attempt to further enhance the platform, the firm introduced Visual AI in the DataRobot 6.0 to automate ML workflows with computer vision technology. The company has […]","categories":["Deep Tech"],"tags":["automated data science solutions","Automl","multiclass classification"],"author_name":"Rohit Yadav","publish_date":"2020-04-08T10:00:00","publication_year":"2020","word_count":730,"keywords":["data science","Automl","Go","TPU","AI","neural network","ML","computer vision","RAG","automated data science solutions","multiclass classification","object detection","R"],"extracted_tech_keywords":["AI","ML","neural network","computer vision","data science","RAG","object detection","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-get-started-with-visual-ai-the-new-automl-solution-by-datarobot\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":44779,"title":"BI Startup vPhrase Analytics Raises $2 Million In Series A Funding From Falcon Edge &#038; Bharat Innovation Fund","content":"Business intelligence and data analysis tools startup vPhrase Analytics this week secured a Series A funding for $2 million from Bharat Innovation Fund and Falcon Edge Capital. The company, which was founded in 2015, has built a product called Phrazor which can gather data, structure facts and apply language to describe it, and finally generate insightful outputs from it. Phrazor is now deployed across domains – banks, brokerage firms, healthcare, CPG companies, media and entertainment and makes the job of reporting easier by giving a personalized narrative and highlighting areas that need due attention. Neerav Parekh, founder and CEO at vPhrase says that the company will use its funding towards setting up offices in North America, launching their new product Explorazor and hiring more experienced staff to guide them along. He told a noted financial daily, “So far, we’ve been working with some large enterprises in India and with some companies in Hong Kong and Singapore. Our plan now is to get into western markets. For that, we will first enter North America, starting with the US and Canada… The funds will also be used to launch our latest product, Explorazor, an independent BI and analytics platform, which provides ready-to-consume insights, with powerful collaboration features.” The Mumbai-based startup counts some of the biggest names from banking, FMCG, retail and media world like HDFC Bank, Motilal Oswal Securities, Unilever, Accenture, consulting companies like Accenture, Wipro, TCS, Viacom18, Sony, Moneycontrol.com as its biggest clients.","excerpt":"Business intelligence and data analysis tools startup vPhrase Analytics this week secured a Series A funding for $2 million from Bharat Innovation Fund and Falcon Edge Capital. The company, which was founded in 2015, has built a product called Phrazor which can gather data, structure facts and apply language to describe it, and finally generate […]","categories":["AI News"],"tags":["Startups"],"author_name":"Prajakta Hebbar","publish_date":"2019-08-21T12:28:18","publication_year":"2019","word_count":242,"keywords":["business intelligence","API","TPU","AI","innovation","RAG","analytics","Startups","R","analytics platform","startup"],"extracted_tech_keywords":["AI","analytics","RAG","TPU","R","API","business intelligence","analytics platform","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bi-startup-vphrase-analytics-raises-2-million-in-series-a-funding-from-falcon-edge-bharat-innovation-fund\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10104366,"title":"Sarvam AI raises $41 million to Train Indic LLMs","content":"Bangalore-based startup Sarvam AI has raised USD 41 million in a Series AI funding round led by Lightspeed and supported by Peak XV Partners and Khosla Ventures. Founded by Vivek Raghavan and Pratyush Kumar, Sarvam AI will focus on India’s unique needs. This includes training AI models to support the diverse Indian languages and voice-first interfaces. The company will also work with Indian enterprises to co-build domain-specific AI models on their data. Finally, the company aims to create population-scale impact by layering generative AI on top of the highly successful India stack specifically for public-good applications. Sarvam AI’s ambitious plan is to develop the “full-stack” for Generative AI, ranging from research-led innovations in training custom AI models to an enterprise-grade platform for authoring and deployment. The company believes that this full-stack approach will accelerate the adoption of generative AI in India, especially given that enterprises see the potential of generative AI but are grappling with how to leverage it for their business. “I have seen first-hand the enormous value in innovating at foundational layers and deploying at population scale. India has demonstrated that it can harness technology differently, and with generative AI we have an opportunity to reimagine how this technology can add value to people’s lives,”, Vivek Raghavan, co-founder of Sarvam AI said.","excerpt":"The company will also work with Indian enterprises to co-build domain-specific AI models on their data","categories":["AI News"],"tags":["sarvam ai"],"author_name":"Pritam Bordoloi","publish_date":"2023-12-07T14:15:10","publication_year":"2023","word_count":214,"keywords":["Go","funding","sarvam ai","AI","programming_languages:R","innovation","RAG","Aim","generative AI","R","startup"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","R","Go","innovation","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sarvam-ai-raises-41-million-to-train-indic-llms\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":31025,"title":"These 5 Fiction Novels Predicted The Rise Of AI Long Before Pop Culture Did","content":"Imagination is the key to innovation, and this is probably why writers have many times thought beyond scientists. For example, numerous books have been written about new tech which have interesting plots revolving around artificial intelligence and machine learning. Here we present five fictional books that predicted AI way before the technology became a rage that it is today. (The books are listed in the order of their publication) The Adolescence Of P-1 by Thomas J Ryan (1977) Ever imagined having a computer child who has awesome powers but is still emotional as a baby? Thomas J Ryan used the character of a university student named Gregory Burgess who is inspired by the game theory and builds an AI to crack the mainframe. The plot takes an interesting turn when the protagonist realises that the program is not operating in a way he wanted he tries to shut it down, but has started learning, understanding and adapting to its own feebleness. The story became a reality in 2004 when a 17-year-old german created Sasser & Netsky worms which were built for a business purpose but spread through computers by irregular scanning of IP address and forcing them to download the virus. The Two Faces Of Tomorrow by James Patrick Hogan (1979) This is a must-read book for anyone who worries about how supercomputer networks can take over the earth. Written by James Patrick Hogan, the book speaks about a future of complex civilization where a world-wide computer network controls everything. The advanced technology includes all logical thinking with an absence of common sense. Its logical decisions begin to cause too many disastrous accidents. It is then that the worried scientists get back to drawing boards to develop a universal, self-aware programming system which is armed with judgment. Neuromancer by William Gibson (1984) Long before the Cambridge Analytica scandal William Gibson predicted the role of corporate power in data usage. The plot of Neuromance speaks about a low self-esteemed hacker who gets hired by an enigmatic detective, who bids treatment in exchange for his services as a hacker. The hacker later realizes the real intentions of the detective. The novel is packed with strong themes about corporate power, AI and misuse of modern technology. Islands In The Net by Bruce Sterling (1988) Are you worried about the future data war and its consequences?. have seen how Facebook breached data but couldn’t figure out why they have done so? If the answer is yes, then Islands In The Net is the book that you must read. The story revolves around how the corporates become global powers by turning information into the most valuable currency, while smaller islands and nations become a hub for data pirates and terrorists. Snow Crush by Neal Stephenson (1992) This novel by Neal Stephenson revolves around history, anthropology, archaeology, religion, linguistics, cryptography, memetics, philosophy, politics and computer science. The plot opens in the 21st century of Los Angeles which no longer a part of the federal government of the US. this book can be related to recent Russian government hackers infiltrating the American energy and nuclear business networks. Snow Crush is actually a mysterious computer virus that does harrowing things to the hackers brain in the virtual world.","excerpt":"Imagination is the key to innovation, and this is probably why writers have many times thought beyond scientists. For example, numerous books have been written about new tech which have interesting plots revolving around artificial intelligence and machine learning. Here we present five fictional books that predicted AI way before the technology became a rage […]","categories":["AI Features"],"tags":["AI Books","artificial inelligence","data science books"],"author_name":"Bharat Adibhatla","publish_date":"2018-12-04T08:08:16","publication_year":"2018","word_count":540,"keywords":["Go","artificial intelligence","AI Books","artificial inelligence","AI","machine learning","innovation","programming_languages:R","programming_languages:Go","RAG","data science books","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","Go","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/these-5-fiction-novels-predicted-the-rise-of-ai-long-before-pop-culture-did\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":41067,"title":"DataRobot Acquires ParallelM To Up The Ante In Machine Learning Operations","content":"DataRobot, a leader in enterprise AI, announced that it has acquired ParallelM, a Santa Clara, CA-based company that pioneered machine learning operations (MLOps) category and helps organisations scale the deployment, management, and governance of machine learning in production using any ML platform on any cloud or on-premise environment. This is the second big announcement by the company in a month after it announced a partnership with Informatica, a cloud data management leader. In fact, the ParallelM acquisition is DataRobot’s fourth acquisition in approximately two years. DataRobot is a big name across the globe as organisations use its services to empower the teams they already have in place to rapidly build and deploy machine learning models and create advanced AI applications. Despite the massive investments in data science teams, platforms, and infrastructure, as well as a dramatic increase in the number of active AI projects, the value derived from these investments is grossly lacking due to the inability to deploy AI models into production. Industry experts believe that only a small percentage of AI models make it into production, and the few AI models that do severely lack the necessary governance and monitoring required to ensure the AI can be trusted. Effective and responsible use of AI requires a modern system to deploy, monitor, manage, and govern both models and projects through every step of the AI lifecycle. The acquisition aims to fill these gaps. “We are excited to join forces with DataRobot to help more customers worldwide finally see tangible value and ROI from their AI and machine learning projects and initiatives,” said Sivan Metzger, CEO of ParallelM. ParallelM has been a pioneer in the MLOps space with the launch of MCenter platform in 2017. It helps organisations quickly deploy machine learning models on modern production infrastructures such as Kubernetes and Spark, either on-premise or on a cloud provider of their choice (including Amazon Web Services, Google Cloud Platform, and Azure). ParallelM also pioneered techniques for real-time monitoring and alerts tailored for the unique intricacies of models and the auditing of actions for models required in regulated industries. DataRobots has been making massive investments in model deployment, management and monitoring capabilities. With this acquisition, DataRobots becomes a clear leader in MLOps and governance. As part of the acquisition, DataRobot will expand its platform’s current model monitoring and management capabilities to include an industry-leading MLOps and Governance offering that accelerates the AI lifecycle for all projects regardless of ML platform, programming language, or deployment scenario. “Machine learning operations and governance are a must-have to become an AI-driven enterprise,” said Jeremy Achin, CEO of DataRobot. “We are thrilled to have them on board, including having ParallelM CEO Sivan Metzger join our leadership team and run our MLOps and Governance business.”","excerpt":"DataRobot, a leader in enterprise AI, announced that it has acquired ParallelM, a Santa Clara, CA-based company that pioneered machine learning operations (MLOps) category and helps organisations scale the deployment, management, and governance of machine learning in production using any ML platform on any cloud or on-premise environment. This is the second big announcement by […]","categories":["AI News"],"tags":["big data in auditing","informatica"],"author_name":"Srishti Deoras","publish_date":"2019-06-20T12:51:45","publication_year":"2019","word_count":458,"keywords":["data science","machine learning","informatica","AI","R","ML","MLOps","Aim","big data in auditing","model monitoring","Azure","kubernetes"],"extracted_tech_keywords":["AI","machine learning","ML","data science","MLOps","Aim","model monitoring","Azure","kubernetes","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/datarobot-acquires-parallelm-to-up-the-ante-in-machine-learning-operations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23780,"title":"Story Of Haptik: How This AI-Based Conversational Platform Goes Beyond Just Peer-To-Peer Chat","content":"As soon as you open the Haptik website, you are welcomed by BotGuru, a chatbot assistant for helping users with queries. One of the world’s leading chatbot platforms is building similar applications for consumers, publishers and enterprises across the globe. Being at the forefront of the paradigm shift from apps to bots, Haptik is working across various chatbot use cases across retail, customer service, utility and lead generation. Counting Coca-Cola, HDFC Life, Samsung, Amazon Pay among others, as their clients, Haptik, founded in 2013 in Mumbai has a long way to go. The Eureka Moment Swapan Rajdev Founded by Aakrit Vaish and Swapan Rajdev, Haptik came out in the market as a conversational platform when the term ‘chatbot’ had not become commonplace yet. Vaish recalls, “Swapan and I were working in San Francisco when the mobile app industry was taking off. One of the categories of apps that we saw take off significantly was messaging. We noticed people using messaging apps more than any other type of app, and the addiction to the interface was crazy. We figured there is something here to build on, and potentially a platform to support conversations everywhere beyond simply peer to peer chat. That’s when Haptik was born”. Over the past five years, Haptik has strived to push the boundaries of existing technology with a ‘can-do’ attitude. “This dedication is what has helped us innovate and come up with solutions especially in the chatbot industry where we have no precedent to guide us,” says Vaish. The Haptik team consists of innovators from every field and background with several self-made entrepreneurs and graduates from illustrious institutes such as the IITs and BITs, among others. Vaish proudly says that their work culture promotes a transparent exchange of ideas an agile with a ‘‘think it, build it, ship it, tweak it’ product building method. Most of their projects are opened to the team internally for collaborative critiquing which makes each of Haptik’s products the best they can be. Virtual Assistant By Haptik And The AI Behind It Aakrit Vaish Haptik’s virtual assistant is a service which was launched in 2014 with the vision that people should not require complicated apps for every little task. It should be as easy as sending a message and getting things done. The 24×7 personal assistant app allows users to complete daily tasks over a few simple messages. “As the number of users has grown over the years, we’ve scaled our product to be able to handle a thousands of concurrent requests while delivering an experience that’s as akin to normal human speech. Our assistant service now offers everything from reminders to flight or cab bookings to bill payments to jokes and everything is simply a message away. We constantly listen to what our users want and keep updating our services to match,” shares Vaish. Like for every chatbot, machine learning is the core of Haptik’s conversational platforms, which is used to understand what a user is saying, have conversation to get all the required information and be able to find the best possible answer. “Multiple different machine learning algorithms work together to be able to fulfil every request of our users”, shares Vaish. How Do They Stay Ahead Of The Competition “Haptik was one of the earliest entrants in the industry and has over four years of experience building chatbots from scratch. Every bit of software required to develop, maintain and customise a bot is fully built in-house”, shares Vaish. And Haptik invests heavily in technology to be at the forefront of the chatbot wave. “All our products leverage our experience of over five years where we handle over five million chats a month to continuously train its model to better understand and satisfy user intent. Using a combination of machine and deep learning technologies, we are able to understand user intent far better than other platforms in the market”, he adds. Each one of Haptik’s chatbots is customised to the needs of their client, and can easily handle spelling errors, transliterations, short-forms, syntactical errors, making the bot extremely efficient and smart. Vaish also shares that Haptik is looking to invest at an enterprise level, which will be a first for any company in India. Partnering To Become A One-stop Mobile Commerce Platform Talking about the company’s ambitions, it lets any company that wishes to turn their traditional service pipeline into a conversational model to partner with Haptik. That being said, a majority of their clients at the moment belong to the BFSI and FMCG industries. Some of their clients include, Edelweiss, ICICI and HDFC Life in the BFSI sector, Coca Cola in the FMCG sector, Dr Batra and Lal Pathlabs in the medical sector, Housejoy in Home Services, TOI, Uber and PhonePe among others. They have also recently built a bot for AkanchaAgainstHarassment that helps victims of cyber harassment. Growth Story- Listing Major Milestones Vaish shares timeline with Analytics India Magazine on growth trajectory of the startup: March, 2014: After a really unsuccessful beta phase and almost deciding to stop, we launched the Haptik app on 31 March 2014 and the first 24 hours saw five times more traffic than we expected. We knew we were on to something.August, 2015: We were two months away from running out of money after an investment agreement fell apart when a chance meeting with Satyan Gajwani led to a Series B with Times Internet in April 2016, giving double the return to our earlier investors.February, 2017: We hadn’t stepped into the enterprise chatbot use-case when Coca-Cola wanted to meet with us for a Facebook chatbot idea. We had never built a Facebook bot, but we gave them enough confidence to bet on us after a brief meeting at Mumbai Airport. Since then we have signed on 12 more clients of similar size and our enterprise business is the fastest growing business line in the company.October, 2017: We launched an integration with the Times of India Android app, where Haptik is embedded inside the app. This is a technological breakthrough where more than 20 services and six PGs\/wallets are integrated inside a news app for the first time globally, all in under 1 MB. We expect our monthly active users to double every month from here on for the next 6 months.December, 2017: Haptik won the Amazon AI Award for Customer Engagement, a major validation for our efforts as a company and the work we’ve done to build engaging enterprise solutions. With these major milestones, Vaish shares that for the coming year, they are going to focus on three key areas of technology — make chatbot building easier where anyone is able to build a bot in 10 min, voice interfaces, and regional language support. Challenges Of Being In A Chatbot Space Vaish says “Being at the forefront of an industry which is still evolving means that there is no predefined playbook which can be followed to assure results. At Haptik, on an everyday basis, we try out various different experiments some of which succeed while some of them don’t”. “We face challenges on all aspects of engineering starting from scaling, research in machine learning, product and design. We perform research in-house and work closely with academia and other institutes to be able to find the best solutions to advance the chatbot and machine learning space”, he adds. Having said that, it cannot be denied that chatbot space is growing by leaps and bounds, with traditional slow moving businesses automating their procedures and using chatbot to become more efficient. The founders of Haptik believe that with the growth of AI technology, bots will be able to understand regular human speech in both text and voice. “Vernacular language support is another interesting field that is currently open to innovation. We can also expect to see chatbots on various other platforms apart from websites and messengers. Chatbots will become more prevalent and be a part of everyday life for everyone who owns a smartphone!”, he said while concluding.","excerpt":"As soon as you open the Haptik website, you are welcomed by BotGuru, a chatbot assistant for helping users with queries. One of the world’s leading chatbot platforms is building similar applications for consumers, publishers and enterprises across the globe. Being at the forefront of the paradigm shift from apps to bots, Haptik is working […]","categories":["Deep Tech"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2018-04-17T10:57:25","publication_year":"2018","word_count":1332,"keywords":["Go","machine learning","AI","chatbots","innovation","RAG","deep learning","analytics","R","startup"],"extracted_tech_keywords":["AI","machine learning","deep learning","analytics","RAG","chatbots","R","Go","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/story-of-haptik-how-this-ai-based-conversational-platform-goes-beyond-just-peer-to-peer-chat\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10170583,"title":"A Clear Case of AI Innovation","content":"Filing taxes is no fun—anyone who’s ever done it knows so. So, when minister of state for finance, Pankaj Chaudhary, informed Parliament last year that more people are filing returns, with 6.68% of India’s population doing so in 2023–24, you know something has changed. A large part of that credit goes to platforms like Clear, which have turned tax filing from a dreaded chore to a few simple taps on your screen. Clear, formerly known as ClearTax, is transforming tax filing and compliance in India through the use of AI. With tools like its AI Copilot and WhatsApp-based assistant, Clear simplifies complex processes for both individuals and enterprises, handling multilingual support, document processing, and intelligent audits. A WhatsApp-based tax filing agent of Clear, specifically designed to simplify tax compliance for India’s vast gig workforce, is set to generate $1 million in annual recurring revenue (ARR) this year, co-founder Ankit Solanki revealed in a conversation with AIM. The agent is capable of understanding multiple languages, managing complex states, and interpreting multimodal inputs, including text, images and voice notes. “Our WhatsApp-filing agent seamlessly managed a complex 16-step tax filing workflow, resulting in rapid user adoption and a projected $1 million ARR within its second year,” he said. Notably, the feature was launched in mid-2024. “We needed to handle more than 1,000 states, multiple languages, and multimodal inputs. We achieved this by equipping a network of four bots with 10 tools. The product went viral, forcing us to scramble for additional token bandwidth overnight,” he recalled. Solanki added that the company extensively uses generative AI models or large language models (LLMs) in customer-facing tools and backend processes across both B2B and B2C operations. One of its flagship consumer-facing initiatives is an AI Copilot, Ask Neha, integrated directly into its DIY tax filing product. The assistant offers immediate help through conversational interaction and intelligent features like deep-link navigation and ticket creation. On the backend, Clear’s Document AI replaces traditional optical character recognition (OCR) workflows with modern multimodal LLMs to extract structured data from documents like invoices and tax notices. Solanki noted that the company also uses GenAI in its Tax Notice Manager to analyse compliance documents and recommend actionable responses. In its enterprise business, Clear extensively applies AI across document digitisation via Document AI, financial transaction reconciliation through Recon, and intelligent audits via Transaction Intelligence. “Our Recon product, trained on historical client data, improved transaction matching by 7%, delivering substantial cost savings for our business customers,” Solanki said. Impact of AI Commenting on AI’s role in automation, Solanki said it has dramatically streamlined processes that once required heavy manual effort. For instance, AI-powered document processing now enables automatic classification, data extraction, and direct population of user forms, transforming tedious, manual workflows into seamless digital experiences. “Our AI models have significantly improved both accuracy and speed in reconciliation, directly resulting in cost and labour savings. Similarly, AI-based intelligent audits, such as assessing ITC eligibility or TDS categorisation, reduce manual oversight and boost client confidence in compliance,” he explained. The company also uses past user interactions and transaction histories to enable smart categorisation and risk assessment for financial transactions, such as determining ITC eligibility, assigning TDS sections, or flagging Reverse Charge Mechanism (RCM) cases. “Our AI-powered alerts and recommendations offer clients personalised, actionable guidance and peace of mind in compliance workflows,” he added. Solanki confirmed that AI has delivered measurable business value and said the WhatsApp-based tax filing agent would not have been possible without it. “This multilingual, multimodal AI-driven product achieved rapid adoption. Moreover, our AI-enhanced reconciliation system, with a 7% boost in transactional linking, led to direct cost and efficiency gains for enterprise clients,” he noted. Data Handling Clear deploys AI to extract structure and insights from large volumes of scanned, unstructured documents, enabling smarter classification, faster reconciliations, and more accurate compliance analyses. “By leveraging historical transaction data, our predictive AI models enhance automation and accuracy in categorisation and compliance checks—providing clients with strategic insights and direct cost savings,” Solanki explained. To ensure data privacy and compliance, the company employs a multi-pronged approach: strict access controls, robust data handling protocols, removal or masking of personally identifiable information (PII) before any GenAI API calls, and client-specific data segmentation for training models. It also relies on hosted LLMs from trusted cloud providers, ensuring industry-standard privacy and regulatory compliance. Tech Behind the Solutions Source: Clear Solanki shared that most of Clear’s AI services are Python-based FastAPI applications hosted in Docker containers. For classical ML workflows, it primarily uses PyTorch. Generative AI agents rely heavily on the LlamaIndex framework for retrieval-augmented generation (RAG), function calling, and structured data extraction. Regarding partnerships, Solanki said all impactful AI research and product development initiatives are currently driven in-house. “We primarily use commercially available APIs and tools from third-party providers, though we’re actively working toward building customised in-house LLMs tailored to finance and compliance,” he added. Encouraged by early success with generative AI agents, Clear is pursuing ambitious projects such as the Integration Agent, which aims to integrate ERP (enterprise resource planning) and accounting systems within a day, and the Supply Chain Agent, which autonomously builds accounts payable workflows from process documents and client interactions. “We also aim to build a specialised Tax and Compliance Expert LLM by embedding Clear’s decade-long domain knowledge into flexible and reliable solutions,” he said. The Industry When asked about AI’s role in shaping India’s fintech and compliance landscape over the next five years, Solanki said, “We believe AI today is akin to the internet in the 1990s—it will redefine everything. It will act as a domain expert interpreting complex regulations, an assistant automating routine tasks, and a safeguard against costly errors. It will also transform product engineering, enrich multilingual user experiences, and democratise access to sophisticated compliance processes.” He predicted that while AI will automate and replace a significant share of traditional accounting tasks, functions requiring human judgment, strategic insight, and direct engagement will persist, with AI serving as a powerful augmentation layer to human expertise.","excerpt":"Clear’s AI-driven WhatsApp tax agent simplifies compliance, boosts efficiency, and is projected to generate $1 million ARR by this year.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","cleartax","tax filing"],"author_name":"C P Balasubramanyam","publish_date":"2025-05-23T12:59:12","publication_year":"2025","word_count":1000,"keywords":["GenAI","tax filing","cleartax","AI","PyTorch","ML","LlamaIndex","RAG","Aim","FastAPI","generative AI","multimodal AI","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","multimodal AI","LlamaIndex","Aim","PyTorch","FastAPI","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-clear-case-of-ai-innovation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162642,"title":"DeepSeek Comes to Windsurf &amp; Cursor","content":"AI coding platform Codeium, a California-based company launched in 2021, has announced that it is integrating DeepSeek-R1 and V3 into its Windsurf platform, making Cascade one of the initial coding agents to support R1. The company said that it would initially be priced at half the usual cost and plans to reduce prices further over time. “R1 is truly fun and reading the chain of thought almost feels like a requirement for reasoning models,” Codeium CEO Varun Mohan said on X. Last year, Codeium integrated Anthropic’s Claude into Windsurf, a collaborative AI-native integrated development environment (IDE). In addition to Codeium, Cursor announced that the DeepSeek models are available on its platform hosted on US servers. Last year, OpenAI’s latest o1 models were made available on Cursor. The o1 models have displayed exceptional performance in handling well-defined and complex reasoning tasks. “While we’re big fans of Deepseek, Sonnet still appears to perform much better on real-world tasks,” Cursor stated in a post on X. Founded by Michael Truell, Sualeh Asif, Arvid Lunnemark, and Aman Sanger, Cursor started with the goal of writing the world’s software. Anysphere recently secured $100 million in a Series B funding round, bringing its post-money valuation to $2.6 billion. Even Microsoft recently announced it is making DeepSeek-R1 available on Azure AI Foundry and the GitHub Model Catalogue, expanding the platform’s AI portfolio. “Customers will soon be able to run DeepSeek-R1’s distilled models locally on Copilot+ PCs, as well as on the vast ecosystem of GPUs available on Windows,” said Microsoft chief Satya Nadella. Even Amazon CEO Andy Jassy has announced that the DeepSeek-R1 models are now available on Amazon Web Services (AWS). What makes DeepSeek Special? DeepSeek, a Chinese AI research lab backed by High-Flyer Capital Management has released DeepSeek-V3, the latest version of their frontier model. “The raw chain of thought from DeepSeek is fascinating. It really reads like a human thinking out loud: charming and strange,” Ethan Mollick, professor at The Wharton School, said. Sharing similar sentiments, Matthew Berman, CEO of Forward Future, said, “DeepSeek-R1 has the most human-like internal monologue I’ve ever seen. It’s actually quite endearing.” DeepSeek, in its research paper, revealed that the company bet big on reinforcement learning (RL) to train both of these models. DeepSeek-R1-Zero was developed using a pure RL approach without any prior supervised fine-tuning (SFT). This model utilised Group Relative Policy Optimisation (GRPO), which allows for efficient RL training by estimating baselines from group scores rather than requiring a separate critic model of similar size to the policy model. DeepSeek-R1 incorporates a multi-stage training approach and cold start data. This method improved the model’s performance by refining its reasoning abilities while maintaining clarity in output. “The model has shown performance comparable to OpenAI’s o1-1217 on various reasoning tasks,” the company said. “This ‘aha moment’ in the DeepSeek-R1 paper is huge. Pure reinforcement learning enables an LLM to automatically learn to think and reflect,” Yuchen Jin, co-founder and CTO of Hyperbolic, said.","excerpt":"DeepSeek, in its research paper, revealed that the company bet big on reinforcement learning (RL) to train both of these models.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Aditi Suresh","publish_date":"2025-01-31T17:19:56","publication_year":"2025","word_count":494,"keywords":["Anthropic","Go","TPU","OpenAI","AI","AWS","R","Git","chain of thought","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","Anthropic","chain of thought","AWS","Azure","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deepseek-comes-to-windsurf-cursor\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10400,"title":"Building a Collaborative Filtering Recommendation Engine in 15 Minutes","content":"Since the amount of data has grown massively in last few years, the task of obtaining relevant information has become a challenge. Recommendation engines are one solution to such problem, where they aim to deliver accurate and relevant pieces of information to the users. Recommendation engines are widely observed today, for example – youtube’s video suggestions lists, amazon’s recommended products and facebook’s friends and ads suggestions. Types of Recommendation Engines There are generally two types of recommendation engines – Collaborative Filtering Based and Content Based. Sometimes the hybrid models are also used. Collaborative Filtering Models: are based on actions and behaviour of users. For example if a person having interest in thriller movies, is purchasing a set of thriller movies, another person having similar interests is also likely to buy same set of movies. Content Based Models: are based on the features of the products, items with similar content are clustered together, and recommended together. For example – Apple Iphone and Apple Ipad belongs to one cluster of apple products, samsung phone and apple iphone belongs to one cluster of smartphones. In each of the algorithms, the similarities between different items are calculated using any similarity measurement technique, such as – cosine similarity, vector similarity, levenshtein distance and edit distance etc. These similarity values are used to predict ratings for unobserved user-item pairs. Recommendation Engine in Python Let’s take a scenario for an ecommerce company, where a large number of users actively buy large number of products\/items. To maximize their sales, The company wants to identify which of the brands are similar according to user behaviours and their buying patterns.For this task, company has collected a data for users of random shoppers and the brands they purchased. The task is to create a model, that will recommend most similar brands with respect to a new brand. The complete data can be downloaded from this link. It contains three columns – “User Id”,”Brand Id”, “Brand Name”. Hypothesis Generation: Before exploring the data, generating the hypothesis around the problem statement is really helpful. In our case, the problem statement is to create a collaborative filtering based brand recommendation engine. The hypothesis can be generated around the factors that will make two brands similar. For example Type of Products sold by them, example – cloths brands, sports brands, cars brands etc. Gender – for which gender the brands are famous for ? example – male brands, female brands. Parent Company – which is the main parent company for the brands. ? Do they belong to the category of expensive brands vs average brands ? People Interests – which people interests are important. Exploratory Data Analysis: Now that problem statement and hypothesis is generated, it is the time for data exploration which is one of the important component of any data science model. Some descriptive stats from the data can be obtained as – Total Records: 4,322,841Unique Users: 576,809Unique Brands: 11,288Most purchased brands: Marc by Marc Jacobs119644Christian Louboutin115665Burberry112238Gucci102444Marc Jacobs93478Jimmy Choo88514Prada86677Diane von Furstenberg82491Yves Saint Laurent82203Chlo82130 Least purchased Brands Zita1Personalized With Luv19 Months up, 9 Months down1Amy Zerner1Petit Chapeau1Atos Lombardini1Caran d’Ache1Yang Li1Tiedeken1Laura G1 Since the data consists of about 11K brands, it is important step to normalize the data and get rid of weak data points, Hence all the brands below a certain threshold value are ignored from the data. For example – in this data, bottom 10% portion is discarded. Please feel free to suggest more data normalization techniques in the comments section. Approach: First of all, we need to transform the data into a User – Brand matrix, where each row represents a user, and each column represents a brand. Every cell indicates if the user has purchased that brand or not. The dimension of this matrix is U*B where U is the total number of unique users and B is the total number of unique brands. The resultant matrix is shown as – where, Mij = {1, if user i has purchased brand j, ” ” otherwise} def prepare_dataset(self): user_brand = {} data = open(self.input_file).read().split(\"\\n\") for i,line in enumerate(data): row = line.split(\"\\t\") userId = row[0] brandId = row[1] brandName = row[2] if userId not in user_brand: user_brand[userId] = [] user_brand[userId].append(brandId) return user_brand Next, we need to create Brand – Brand similarity matrix. This User – Brand matrix is converted to Brand – Brand matrix, where both rows and columns represents the brands. Every cell contains represents the frequency count – number of times both brands appeared together OR When both brands were purchased by single user. The dimension of this matrix are B*B where B is the number of unique brands. Example – where, Bij = {n, if brand i brand j are co-purchased together n number of times, 0 otherwise} def get_cartesian_pairs(self, lstA, lstB): pairs = [sorted(z) for z in list(itertools.product(lstA, lstB)) if z[0]!=z[1]] return list(pairs for pairs,_ in itertools.groupby(pairs)) def create_brand_similarity_matrix(self): similarity_matrix = {} for x,y in self.user_brand_matrix.iteritems(): cartesian = self.get_cartesian_pairs(y, y) for pair in cartesian: if pair[0] not in similarity_matrix: similarity_matrix[pair[0]] = {} if pair[1] not in similarity_matrix[pair[0]]: similarity_matrix[pair[0]][pair[1]] = 0 similarity_matrix[pair[0]][pair[1]] += 1 return similarity_matrix Next, we need to perform similarity measurements and identify the possible similar items for one item. For every row, the sorted values will indicate the possible similar brands. where, Possible Similar Brands = MAX[Brand j’s with respect to brand i] def update_similarity_matrix(): for bid, similar in self.similarity_matrix.iteritems(): for simid, score in similar.iteritems(): similarBrandName = self.brand_mapper.get(simid) brandName = self.brand_mapper.get(bid) doc = { '_id' : bid+\" \"+simid, 'brandId' : bid, 'similarId' : simid, 'score' : score, 'brandName' : brandName, 'similarBrandName': similarBrandName} push_data(document = doc, collection_name = \"BrandSimilarity\") def get_most_similar(brandId, limit): pipeline = [{'$match' : {'brandId' : brandId}}, {'$sort' : {'score' : -1}}, {'$limit': limit}] result = db['BrandSimilarity'].aggregate(pipeline) return result The function named get_most_similar performs aggregation over the brand – brand matrix, sorts the data and find the most relevant – similar brand. The full code can be downloaded here. This tutorial explained the basics of collaborative filtering in python, obviously there exists more complicated and advanced techniques to cluster items based on user behaviour such as singular vector decomposition. Feel free to shoot out your queries in the comments section.","excerpt":"Since the amount of data has grown massively in last few years, the task of obtaining relevant information has become a challenge. Recommendation engines are one solution to such problem, where they aim to deliver accurate and relevant pieces of information to the users. Recommendation engines are widely observed today, for example – youtube’s video […]","categories":["AI Features"],"tags":["Collaborative Filtering"],"author_name":"Shivam Bansal","publish_date":"2016-07-12T04:53:24","publication_year":"2016","word_count":1024,"keywords":["data science","Go","programming_languages:R","AI","RAG","Python","Collaborative Filtering","Aim","programming_languages:Python","llm_models:Bard","R"],"extracted_tech_keywords":["AI","data science","Aim","RAG","Python","R","Go","llm_models:Bard","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/building-collaborative-filtering-recommendation-engine-15-minutes\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10134577,"title":"‘Education Should Feel Like Going to the Gym for Your Brain’","content":"Eureka Labs founder Andrej Karpathy hopes that in a post-AGI world, people will go to the gym not just physically but also mentally. “I feel learning something is like going to the gym – but for the brain,” said Karpathy in a recent podcast. Earlier this year, Karpathy compared learning with going to the gym and cautioned that a lot of videos on YouTube looked like education, but were merely for entertainment. “Learning is not supposed to be fun. It doesn’t have to be actively not fun either, but the primary feeling should be that of effort. It should look a lot less like that ‘10 minute full-body’ workout from your local digital media creator and a lot more like a serious session at the gym,” he posted on X. He encouraged those who truly wanted to learn, saying that unless you’re trying to master something narrow and specific, close those tabs of quick blog posts. “Close those ‘Learn XYZ in 10 minutes’ tabs,” he advised. Karpathy isn’t excited about a future where people are sidelined due to automation. “There’s a lot of activity in AI focused on replacing or displacing people and often centred around pushing people aside. However, I’m more interested in technologies that empower people, and I feel like I’m kind of on a high level like ‘Team Human’,” he added. He further revealed that the first course from his startup is designed for undergraduates, and is expected to be available by the end of the year or early next year. “I have a lot of distractions, but I am working hard to make it very, very good. It takes time to get there,” said Karpathy. Personal Tutor for 8 Billion People Karpathy believes that personal AI tutors can greatly enhance human learning capabilities. “What I find very interesting is how far a person can go if they have the perfect tutor for all subjects, and I think people could go really far if they had the perfect curriculum,” he said. Karpathy hopes to address Bloom’s Sigma 2 problem. Bloom’s research found that students who received one-on-one or small group instruction with regular feedback performed two standard deviations (2 sigma) better than their peers who had traditional classroom instruction. He said he is currently developing a course which he describes as  “the course you would go to if you want to learn AI”. “I’ve already taught courses, like I taught CS231n at Stanford, and that was the first deep learning class and was pretty successful. But the question is, how do you actually scale these classes? How do you make it so that your target audience could be 8 billion people?” he said. Karpathy believes that current models are not yet good enough to create a quality course, but he thinks they can serve as the front-end for students and interpret the courses for them. “The teacher doesn’t go to the students anymore; the teacher is not the front-end. Instead, the teacher is on the back end, designing the materials and the course, while the AI serves as the front end,” said Karpathy. The founder of Khan Academy, Salman Khan, recently echoed similar sentiments, saying personal tutors are a necessity. “Alexander the Great had Aristotle as his personal tutor,” he added, saying generative AI will do the same for every student globally. “That’s what world-class looks like.” Last year, Khan Academy launched Khanmigo, a personal tutor and teaching assistant powered by GPT-4. This academic year, 65,000 students and teachers piloted Khanmigo across school districts in the US. “It acts just like Aristotle or Socrates would with their students and works across every subject that Khan Academy offers. It has all the context that the student would normally have on Khan Academy, and it also acts as a teaching assistant for teachers,” said Khan. “It remembers the conversation you’re having. It also remembers some of the work that you’ve been doing at Khan Academy,” he added. In India, edtech unicorn PhysicsWallah recently launched Alakh AI. The suite comes with several products including AI Guru, Sahayak, and NCERT Pitara. “AI Guru is a 24\/7 companion available to students, who can use it to ask about anything related to their academics, non-academic support, or more,” said Vineet Govil, the CTPO of Physics Wallah, in an exclusive interview with AIM. He added that the tool is designed to assist students by acting as a tutor, helping with coursework, and providing personalised learning experiences. It also supports teachers by handling administrative tasks, allowing them to focus more on direct student interaction. Future of Education Karpathy feels that in the future, students should focus more on subjects like physics, computer science, and maths, as they are fundamental for developing thinking skills. Moreover, he believes that while today people often learn for practical reasons, such as getting a job, this might not be the case in a post-AGI world. “In a pre-AGI society, education is useful, and I think people are motivated by that because they’re climbing up the economic ladder. In a post-AGI society, I believe education will be much more about entertainment,” concluded Karpathy, contradicting his initial statement that education is not fun and easy.","excerpt":"“But the payoff is immense, just like physical exercise,” says Andrej Karpathy, who plans to release his first course using generative AI by year-end.","categories":["AI Features"],"tags":["Andrej Karpathy"],"author_name":"Siddharth Jindal","publish_date":"2024-09-06T17:00:44","publication_year":"2024","word_count":863,"keywords":["Go","Andrej Karpathy","AI","Git","RAG","GPT","Aim","deep learning","ViT","generative AI","R"],"extracted_tech_keywords":["AI","deep learning","generative AI","Aim","RAG","R","Go","Git","GPT","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/education-should-feel-like-going-to-the-gym-for-your-brain\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10103568,"title":"Andrej Karpathy Launches A New LLM Tutorial","content":"Andrej Karpathy, who specialises in deep learning and computer vision at OpenAI, recently published a new YouTube video ‘Intro to Large Language Models’ based on his recent 30-minute talk on large language models at the AI Security Summit. Seeing how much interest there is in this critical discussion, Karpathy’s video gives a thorough overview of LLMs and their crucial place in the rapidly developing field of generative AI. The video focuses on LLM’s journey to a core component behind systems like ChatGPT, Claude, and Bard by drawing parallels with current operating systems and unveiling the connection between everyday technology. Karpathy bridges the gap between standard technology and the recent advancements that characterize the area by simplifying the intricacies of LLMs through analogies with contemporary operating systems. The talk explores the technical features of huge language models and talks about some of the security-related issues that come with this new paradigm of computing. He also explains how LLM is being trained and how the neural networks are being used after training, along with decoding the integrity of LLM models. Andrej Karpathy, the former director of AI at Tesla has broken traditional barriers, making it possible for a larger audience to comprehend the intricacies of LLMs. His potential to democratise AI knowledge and promote a more inclusive conversation is demonstrated by his ability to put abstract ideas into understandable language. Read More: 6 Brilliant Video Resources on Generative AI by Andrej Karpathy","excerpt":"In this tutorial, Karpathy bridges the gap in learning by simplifying the intricacies of LLMs through analogies with contemporary operating systems","categories":["AI News"],"tags":["Andrej Karpathy"],"author_name":"Sandhra Jayan","publish_date":"2023-11-23T14:23:05","publication_year":"2023","word_count":240,"keywords":["ChatGPT","API","Andrej Karpathy","OpenAI","AI","neural network","computer vision","GPT","deep learning","generative AI","R"],"extracted_tech_keywords":["AI","deep learning","neural network","computer vision","generative AI","ChatGPT","OpenAI","R","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/andrej-karpathy-launches-a-new-llm-tutorial\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10110854,"title":"Accenture Announce $1 Bn Annual Investment in GenAI Training at Davos 2024","content":"At the World Economic Forum in Davos, Julie Sweet, Accenture’s chief, announced that the company allocates an annual budget of $1 billion for training its employees in generative AI. According to Sweet, the technology, though at a nascent stage, is moving fast. GenAI is not a fad; rather, it is developing at a rate ten times faster than that of earlier major advances. Sweet said that in order to better basic education, collaboration with governments is necessary. She also said, “It’s not going to help now, but we need to think 10-20-30 years ahead.” According to her, leadership is the single most important factor that decides whether the business or government successfully uses GenAI or not. She also said that “you actually have to understand it at a very deep level because it is not that there are millions of use-cases…. but you have to operationalise it.” Similarly, Arvind Krishna, chairman and CEO of IBM, who was sitting alongside Sweet in the panel, claimed that Gen AI is a rapidly developing technology that is advancing more quickly than earlier major ones. In order to apply AI, Krishna underlined the significance of reskilling talent because it will create new employment and solve numerous issues. To secure the success of AI, he also emphasised the necessity for businesses and governments to collaborate on reskilling. Moreover, he suggested controlling use-cases rather than the technology itself. Additionally, he forecast that before the end of the decade, AI in its current form will produce $4 trillion in yearly productivity.","excerpt":"She also said that “You actually have to understand it at a very deep level because it is not that there are millions of use-cases…. but you have to operationalise it.”","categories":["AI News"],"tags":["Accenture","davos","Generative AI"],"author_name":"Arya Vishwakarma","publish_date":"2024-01-16T17:49:24","publication_year":"2024","word_count":254,"keywords":["Accenture","Go","API","GenAI","programming_languages:R","AI","programming_languages:Go","Aim","ViT","generative AI","Generative AI","R","davos"],"extracted_tech_keywords":["AI","generative AI","GenAI","Aim","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/accenture-announce-1-bn-annual-investment-in-genai-training-at-davos-2024\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169387,"title":"Over 25 New GenAI Apps Arrive on Slack Marketplace","content":"Slack, the workspace communication platform owned by SalesForce, announced on Tuesday that it has added more than 25 generative AI apps to its marketplace. These apps will use AI for multiple use cases such as content creation, market research, HR operations, and more. The Slack Marketplace enables users to integrate various apps into the Slack application that aid several tasks and operations, thereby enhancing the platform’s functionality. The marketplace now features AI apps from leading companies like Asana, UiPath, Adobe, Glean, Cohere, Perplexity, and more. The company also said that ten more apps are in the pipeline. Rahul Sharma, VP of sales at Salesforce India, said, “With new AI apps now available in the Slack Marketplace, we’re seeing a powerful shift in how work gets done – right within the conversational interface employees already use daily.” These apps now ‘enable organisations to turn on a dedicated, always-on digital labour force in Slack,’ added Sharma. In January, Salesforce announced the availability of its Agentforce 2.0, which integrates advanced AI agents into Slack. These agents can handle multistep tasks, respond in real time within channels and direct messages (DMs), use pre-built Slack actions, and leverage enterprise search to draw contextual insights from Slack data. In March, TechCrunch reported that OpenAI will soon begin testing a feature that would allow users to integrate apps like Slack and Google Drive with ChatGPT. The new feature, called ChatGPT Connectors, allows users on the ChatGPT Team to link external apps to the platform. This enables ChatGPT to respond to queries using data from those connected apps. “This will allow employees using ChatGPT to easily make use of internal information similar to how they can use world knowledge via web search,” TechCrunch reported, citing documents.","excerpt":"These apps will use AI for multiple use cases such as content creation, market research, HR operations, and more.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Slack"],"author_name":"Supreeth Koundinya","publish_date":"2025-05-08T09:21:20","publication_year":"2025","word_count":288,"keywords":["Go","ChatGPT","OpenAI","AI","Git","RAG","GPT","generative AI","GAN","R","Slack","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","RAG","R","Go","Git","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/over-25-new-genai-apps-arrive-on-slack-marketplace\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166927,"title":"It’s Time to End the Dependence on Electronic Imports for Space, Urges ISRO Chief","content":"To achieve self-reliance in the space sector, India must significantly cut its imports of electronic components, newly appointed ISRO chairperson V Narayanan said, expressing concern over India’s continued dependency on imported goods. Currently, 90% of the components in ISRO’s launch vehicles are indigenised. The remaining 10% are imported, of which most are just electronics, the chairperson noted. “At least in the next five years, the import of electronic components in our country should come down drastically,” Narayanan asserted. He made this remark while speaking at the Nano Electronics Roadshow 2025 hosted by the Ministry of Electronics and IT (MeitY) in Bengaluru earlier this week. Narayanan highlighted ISRO’s experience with sensors, noting that while initially only two of the 34 sensors used by the team were made in India, today, 32 of those 34 sensors are indigenously developed. This underscores India’s capability to rapidly increase indigenous production in recent years. He praised India’s recent homegrown advancements, including the Vikram 32-bit processor, which ISRO previously imported. However, this move to innovate was necessitated by the denial of access to technology by other countries. From Bicycles to Automation Narayanan highlighted how India, which lacked satellite technology until 1975, is now donating satellites to other nations, signifying a significant milestone in its space ambitions. Reminiscing back to 1962, he mentioned when India’s space program had been launched and rockets had to be carried to launch sites by bicycles and bullock carts. Looking at the progress of AI autonomous landings and world records achieved by India, the space program seems to be heading in the right direction. Addressing the significance of electronics in space missions, he stated, “Without electronics, there is no space programme,” adding that nanoelectronics play an indispensable role in launch vehicles, spacecraft, and interplanetary missions like Chandrayaan. This initiative was expanded by the development of IIT Madras and ISRO’s indigenous aerospace-grade semiconductor chip based on RISC-V, an open-source instruction set architecture. The ‘IRIS’ (Indigenous RISCV Controller for Space Applications) chip was developed from the ‘SHAKTI’ processor baseline, highlighting India’s efforts toward self-reliance in semiconductor technology for space applications. Indians in Space by 2040 Highlighting ISRO’s achievements, he pointed out India’s unique milestones, including Chandrayaan-3’s soft landing near the lunar South Pole and the Mars Orbiter Mission’s success on its first attempt. The chairperson stressed the need for continued innovation in nanoelectronics, sensors, and processors, reflecting confidence that India can achieve complete self-reliance in electronics critical for space and other high-tech sectors. “This year we are going to have the first uncrewed mission, followed by two uncrewed, and the crewed mission will be authorised,” Narayanan said. ISRO has ambitious plans, including building its space station by 2035, undertaking human missions to space, and developing reusable launch vehicles. There was also emphasis on PM Modi’s guideline “to take humans to the moon and bring them back safely by 2040.” Startup Ecosystem is Growing Not only did Narayanan recognise India’s leadership in space innovation, but former ISRO chairperson S Somanath also highlighted this fact. He identified Bengaluru as the central hub, yet noted that the city is still developing a completely integrated communication satellite. “There are small satellite-building companies, but I hope to see a four-tonne or six-tonne class communication satellite built by an Indian company and launched from Bengaluru,” he said during the panel discussion on ‘Future of Commercial Space’ at the Invest Karnataka 2025 event. Despite the growth of the private sector, gaps in manufacturing continue to be a significant limitation. “We see startups focusing on components, but I have yet to see a fully integrated product emerging from Bengaluru,” he added. Many more Indian engineers are opting to stay and build in India rather than moving abroad, thanks to the significant influence of India’s evolving space policy on this shift. “There are really exciting challenges to solve, and in terms of the trade-off, people went abroad because of better projects and financial incentives. But if you are paid well here, you could visit those countries while building in India,” said Yashas Karanam, co-founder and COO of Bellatrix Aerospace, in an interview with AIM. In January, the Union government announced the budget and allocated funds for a deep tech fund designed to support startups in this sector for further growth and expansion. While there were many more expectations from space tech and drone startups alike, the government launched a ₹1,000 crore venture capital fund to encourage investment in space startups. The fund was expected to boost investors’ confidence, with ISRO backing commercial ventures in their early stages. Help from ISRO, IN-SPACe and Others An organisation that has played a huge role in India’s success is IN-SPACe (Indian National Space Promotion and Authorisation Centre). The Union Cabinet formed this organisation in 2020, establishing it as a single-window and independent agency operating within the Department of Space (DOS). “IN-SPACe plays a crucial role in boosting the private space sector economy in India. We act as a promoter, enabler, authoriser, and supervisor, ensuring a conducive environment for private enterprises,” said Vinod Kumar, director at the IN-SPACe at Cypher 2024 hosted by AIM. Many startups in India have expressed significant support from ISRO and IN-SPACe in advancing the space tech startup ecosystem. Not just those that launched their payloads on the ISRO SpaDeX mission, but also those that seek general help. In an interview with AIM, Divya Kothamasu, co-founder at N Space (Andhra Pradesh-based startup), said, “They are encouraging us to do more and are supporting us in technical help whenever we have queries or doubts. We have meetings where we conduct design reviews with them as well.” She also mentioned the technical problems the startup faced during the final stages of its experiment for the POEM-4 mission of ISRO and how the organisation helped fix the errors within a week to see a successful launch. Additionally, PierSight, another startup based out of Ahmedabad, received significant support from both organisations during the preparation of its launch. Gaurav Seth, co-founder and CEO, told AIM, “The way ISRO has opened up its arms for private companies stands as a major reason for our success, which is claimed within a short period of time.”","excerpt":"“Without electronics, there is no space programme.”","categories":["Deep Tech"],"tags":["imports","ISRO","Make in India","Space Tech"],"author_name":"Sanjana Gupta","publish_date":"2025-03-30T16:50:14","publication_year":"2025","word_count":1024,"keywords":["Go","ISRO","Make in India","API","AI","innovation","RAG","automation","Ray","Aim","imports","GAN","R","Space Tech"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","API","GAN","automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/its-time-to-end-the-dependence-on-electronic-imports-for-space-urges-isro-chief\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10069999,"title":"The JAX libraries &#038; frameworks for reinforcement learning","content":"JAX (Just After eXecution) is a machine\/deep learning library developed by DeepMind. All JAX operations are based on XLA or Accelerated Linear Algebra. Developed by Google, XLA is a domain-specific compiler for linear algebra that uses whole-program optimisations to accelerate computing. It makes BERT’s training speed faster by almost 7.3 times. It is designed for high-performance numerical computing. JAX was launched in 2018 and is presently used by Alphabet subsidiary DeepMind. It is similar to the numerical computing library NumPy, another library for Python programming. Its API for numerical functions is based on NumPy. In this article, we’ll explore different libraries and frameworks for reinforcement learning using JAX. JAX reinforcement learning agents 1. RLax RLax (pronounced ‘relax’) is a simple library on JAX. It provides useful building blocks for implementing reinforcement learning agents. It can be installed directly from GitHub or PyPI. All RLax code may be compiled for different hardware (e.g. CPU, GPU, TPU) using jax.jit. For more information, click here. 2. Haiku Haiku is a library for JAX. The neural network allows users to use familiar programming models, with JAX’s pure function transformations available. The two core tools provided include a module abstraction, hk.Module, and a simple function transformation, hk.transform. It is written in pure Python but depends on C++ code. The repository can be found here. 3. Gymnax Gymnax is the JAX-compatible version of Open AI’s gym environment. Gym is an open-source Python library for developing and comparing RL algorithms. It provides a standard API to communicate between learning algorithms and environments. After the release, the API has become the field standard for doing this. For more information, click here. 4. Dopamine Dopamine is a research framework for prototyping RL algorithms. It aims to fulfil the need for a codebase in which users can freely experiment with wild ideas (theoretical research). The easiest way to use Dopamine is to install it from the source and modify the source code directly. The version released in 2020 supports JAX agents, which includes an implementation of the Quantile Regression agent (QR-DQN). It can also be installed with pip. Moreover, Dopamine supports Atari environments and Mujoco environments. For more information, click here. 5. JAX FLAX (RL) Launched in 2020, Flax is a high-performance neural network library for JAX that is designed for flexibility: Try new forms of training by forking an example and modifying the training loop, not by adding features to a framework. It is developed in close collaboration with the JAX team. The basic philosophy behind FLAX is library code should be easy to read and understand. At its core, Flax is built around parameterised functions called Modules, which can be used as normal functions. The Google Research: Flax repository is on GitHub. 6. coax coax is an RL python package for solving GymAI environments with JAX-based function approximators. It is designed to align with the core RL concepts, not with the high-level concept of an agent. This makes coax more modular and user-friendly for RL users. For more information, click here. 7. Acme Acme is a library of reinforcement learning (RL) building blocks that strives to expose simple agents. Firstly, these agents serve as reference implementations and provide strong baselines for algorithm performance. The building blocks of Acme are designed in such a way that the agents can be written at multiple scales. It supports both TensorFlow v2 and JAX. For more information, click here.","excerpt":"In this article, we’ll explore different libraries and frameworks for reinforcement learning using JAX.","categories":["IT Services"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2022-06-29T14:00:00","publication_year":"2022","word_count":567,"keywords":["NumPy","TPU","AI","neural network","Python","Aim","deep learning","JAX","TensorFlow","R"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","TensorFlow","JAX","NumPy","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-jax-libraries-frameworks-for-reinforcement-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10069835,"title":"Now Reliance wants to conquer the AI space","content":"During the RAISE 2020, Reliance Group head honcho Mukesh Ambani said, “AI and other associated technologies of the fourth Industrial Revolution will vastly expand our capacity to solve the most complex and pressing problems before India and the world.” He also added that AI will help India in moving faster toward transforming India into a high-growth economy on its way to USD 5 trillion. Ambani’s optimism in the technology is well supported by Reliance Group’s investment strategy in the field of artificial intelligence. Reliance’s digital clout is growing in the country, so much so that it has raised fears of a monopoly. Many believe that Reliance is aggressively scouting for AI and NLP companies in the digital space in a bid to create an Indian equivalent of FAANG – Facebook, Apple, Amazon, Netflix, and Google. Investments and acquisitions A 2018 media report said that Reliance Industries and its auxiliaries had set aside USD 2 billion for its multi-design strategy approach to transforming into a Technology and Innovation powerhouse. This strategy involves spending USD 1.6 billion on purchasing stakes in technology-driven companies situated in different countries, including the USA, the UK, and India. The remaining amount would be directed towards creating in-house tech in artificial intelligence, machine learning, blockchain, augmented reality, Internet of Things, robotics, and Big Data analytics, among other emerging technologies. Earlier this year, Reliance Industries invested USD 15 million in Silicon Valley-based Two Platforms to take over a 25 per cent stake in the company. A deep-tech company founded by Pranav Mistry in 2021, Two Platforms is building interactive and immersive AI experiences. This AI platform enables real-time AI voice and video calls, digital humans, and lifelike gaming. In a statement, the Director of Jio, Akash Ambani, had then said that this investment would help in expediting the development of new products in the areas of interactive AI, immersive gaming, and the metaverse. In the same month, Reliance Jio Platforms, the tech arm of Reliance, invested USD 200 million in InMobi’s mobile content startup Glance. This company uses algorithms to deliver personalised news, entertainment content and live videos, making them available right on the smartphone lock screens. Glance is also backed by Google and PayPal co-founder Peter Thiel’s Silicon Valley-based venture fund Mithril Capital. In January, Reliance made an investment of a whopping USD 132 million with a majority stake in robotics company Addverb Technologies. The Noida-based startup is one of the few companies in the world that work on every aspect of robotics. Post the investment, the total valuation of the company stood at USD 270 million. Even before the investment, Addverb’s creations were adopted across Reliance’s different businesses, including JioMart, Ajio, and Netmeds. Last year, Reliance invested USD 8 million in California-headquarter NetraDyna, an AI startup which focuses on commercial fleet safety. Founded by former Qualcomm India & South Asia president Avneesh Agrawal and his Qualcomm colleague David Julian, NetraDyne has developed Driveri, vision-based driver retention and safety platform for commercial vehicles. In 2016 too, Reliance had invested USD 16 million in the company. With the combined value of the two investments, Reliance now owns a 37.4 per cent stake in the startup. Another major investment made by Reliance was in 2019 when Reliance Jio Digital Services entered into a strategic transaction with the conversation AI platform Haptik. In a statement, the company said that the transaction size was estimated to be Rs 700 crore, with Rs 230 crore considered for the initial business transfer. At that time, Akash Ambani had said that this investment would boost the digital ecosystem and offer users conversational AI-enabled devices with multi-lingual capabilities. With Haptik on board, Reliance’s focus has been on the enhancement and expansion of Jio, with a market opportunity of 1 billion users. Interestingly, Reliance began realising the potential of AI and machine learning much before the sudden traction that these fields are now receiving. In 2014, Reliance made an investment of USD 100,000 in Detect Technologies Pvt Ltd, which makes real-time pipeline monitoring systems to track leaks and corrosions using AI and other sensors. Other notable investments\/acquisitions made by Reliance include deep tech startup Tesseract; Pharma-based enterprise platform C-Square; AI-based edtech platform Embibe, among others.​​ AI talent In 2018, Reliance announced the launch of JioInteract. It was touted by its parent company as the world’s first artificial intelligence-based brand engagement video platform. Under the leadership of Akash Ambani, the company soon started hiring professionals to build its emerging technology firm. Spearheaded by Akash Ambani, the JioCoin project was responsible for building a team of young professionals. Further, Reliance is also investing in AI education. Reliance-owned Jio University also offers a post-graduate programme in AI and data science. As per the website, the programme aims to “instil theoretical capabilities and provide the know-how to create practical solutions for enterprise and society.” What next The investments and acquisitions in the recent past indicate Reliance’s conscious efforts to pivot from its image of just a petrochemical giant to a digital enterprise with ventures like Jio and JioMart. As noted above, Reliance has acquired or invested in a dozen tech startups across domains like health tech, SaaS, robotics, edtech, etc. It aims to become a ‘tech tornado’ for the next-generation technology world. At the RAISE 2020 event, Mukesh Ambani said, “In the past, nations have competed on physical capital, financial capital, human capital and intellectual capital. But, in the coming decades… nations will increasingly compete on Digital Capital,” Ambani said. He added that 1.3 billion Indians are digitally empowered to create faster economic growth, prosperity, employment opportunities, and better standards of living.","excerpt":"Many believe that Reliance is aggressively scouting for AI and NLP companies in the digital space in a bid to create an Indian equivalent of FAANG – Facebook, Apple, Amazon, Netflix, and Google.","categories":["IT Services"],"tags":["ai investment","Akash Ambani","FAANG","internet","Jio","mukesh ambani","Pranav Mistry","Reliance","TWO AI"],"author_name":"Shraddha Goled","publish_date":"2022-06-28T14:00:00","publication_year":"2022","word_count":932,"keywords":["Jio","FAANG","R","data science","artificial intelligence","NLP","mukesh ambani","analytics","Reliance","Go","machine learning","AI","Pranav Mistry","ai investment","Akash Ambani","TWO AI","Aim","Julia","internet"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","data science","analytics","Aim","R","Go","Julia"],"url":"https:\/\/analyticsindiamag.com\/it-services\/now-reliance-wants-to-conquer-the-ai-space\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10115799,"title":"AI Infrastructure Momentum to Drive Data Center Capex by 17% in 2024: Report","content":"A new report from Dell’Oro Group, a source for telco and data centre industries, forecasts a 17% growth in hyperscale cloud capital expenditures (capex) for 2024, driven by accelerated computing and AI infrastructure investments. The enterprise sector is also experiencing a surge in AI infrastructure momentum. “After a modest 4% growth in 2023, we’re projecting a significant rebound in worldwide data centre capex this year,” said Baron Fung, Sr. Research Director at Dell’Oro Group. “Accelerated computing for generative AI applications is expected to lead data centre investments, with a modest recovery in general-purpose servers and storage demand following a steep correction.” The report also predicts an 18% growth in server and storage systems revenue in 2024, with a shift in product mix towards AI-optimized servers and server platforms featuring the latest CPUs from Intel, AMD, and ARM. Additionally, the report forecasts an 8% growth in worldwide server unit shipments by 2028, with over twenty percent of global server deployments expected to be accelerated. Notably, by 2028, the top 4 US-based Cloud Service Providers—Amazon, Google, Meta, and Microsoft—are anticipated to account for half of the global data centre capex, underscoring the significant role of major tech companies in driving data centre investment trends. These insights underscore the growing importance of AI workloads in shaping the future of data centre infrastructure and the tech industry at large.","excerpt":"Notably, by 2028, Amazon, Google, Meta, and Microsoft—are anticipated to account for half of the global data centre capex.","categories":["AI News"],"tags":["AI Data Center","AI Infrastructure"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-03-14T18:05:22","publication_year":"2024","word_count":225,"keywords":["AI Data Center","Go","API","programming_languages:R","AI","programming_languages:Go","RAG","generative AI","AI Infrastructure","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-infrastructure-momentum-to-drive-data-center-capex-by-17-in-2024-report\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":8378,"title":"Machine Learning – a unifying perspective &#038; new paths","content":"Pedro Domingos in his new book, “The Master Algorithm”, has done us a huge favor. As is true of any emerging technology field, Machine Learning (ML) is a “bag of tricks” today; it takes a while for a unifying framework to emerge. Then, one can see various aspects of ML as special cases of a general theory rather than a grab-bag of tools and techniques. Pedro has taken a great early step to such unification. He has collected all major ML initiatives into a taxonomy that makes sense; five schools of thought: the evolutionaries, connectionists, symbolists, Bayesians, and analogizers. I believe this does not go far enough in the unification of ML thought however . . . From the early days of “ML”, I see Pattern Recognition and Classification as a better unifying perspective. In particular, the classic textbook of Duda & Hart, “Pattern Classification & Scene Analysis”, published in 1973 is my starting point! Duda & Hart’s approach in simple terms is as follows. Given labelled samples, obtain a class description consisting of either a distance metric (Euclidean, intra-class, etc.) or a probability density function and then derive a decision rule (Maximum A-posteriori Probability, Bayes, etc.) from the description. The decision rule specifies a decision boundary in feature space among classes. Alternatively, decision surface can be derived directly from labelled samples which is then called a “Discriminant Function”, perceptron being an example. Then, most if not all current ML techniques can be seen as dueling methods to derive Discriminant Functions! Discriminant Functions can be linear or nonlinear (neural network with back-propagation, deep learning, support vector machines, kernel PCA, etc.) and outputs can be binary, integer or real valued. Various learning algorithms can be seen as belonging to the family of iterative\/ recursive\/ adaptive learning algorithms (Least Mean Square being a great old standby!) that update the parameters of the Discriminant Function as new data arrive. In the discussion above, features were considered as “static” and not context-sensitive (for identifying a word within a sentence as an example). Context-sensitivity or Dynamics can be added to improve classification by incorporating Markov models (or Hidden Markov Models for tractable computations). Markov model is a special case of State Space Models which are well-studied in Systems Theory. Setting aside Supervised Classification when labelled samples (or “desired signals”) are available, what can we do when there is no supervision? This is the realm of much harder Unsupervised Learning, which is very useful in transforming basic features into more and more meaningful ones. One usually brings in some overall desirable property to guide unsupervised learning. From the domain of “blind processing” (Radar signal processing, for example), Mutual Information among classes can be minimized as a learning process in the belief that the “best” classification happens when the classes have least overlapping information (better “efficiency” in representation). Instead of entropy-related quantities that are hard to estimate, it is likely that Scale of Fluctuation which is related to “order” and “state space volume” may be a quantity to optimize for a new unsupervised learning process. (For more information on Scale of Fluctuation, refer to my papers, “Instantaneous Scale of Fluctuation Using Kalman-TFD and Applications in Machine Tool Monitoring”, 1997 & “Kalman Filtering and time-frequency distribution of random signals”, 1996). In all of the existing ML bags of tricks, we are still staying at the surface level! We are modeling the attributes or data DIRECTLY. What if we went one level deeper? Model the SYSTEM that generates the data! Syzen Analytics, Inc., takes such an explicit approach in what we call “SYSTEMS” Analytics” which has already demonstrated significant value in business applications. In Syzen Analytics, Inc.’s retail commerce application, our Systems Analytics approach hypothesizes that there is a system, either explicit or implicit, behind the scenes generating customer purchase behaviors and purchase propensities. This ‘one-level-deeper system model parameters’ can be more effective for pattern recognition and classification purposes instead of the data that the model generates! There is a long history of model parameters providing better estimates (in power spectrum analysis, for example). Scale of Fluctuation mentioned earlier seems to have another desirable property of quantifying “coupling” among deeper-level model parameters. Context-sensitivity dynamics is a very good avenue to exploit. The dynamics could be over any independent variable (time always comes to mind first but it is only one of the possibilities). As I noted in my recent blog (“SYSTEMS Analytics – the next big thing in Big Data & Analytics”), “Extensions to Systems Analytics in the future will be inspired by the insight that in reality, data exist in embedded forms in preference and influence networks which are distributed in time and space” AND other independent dimensions (shopper preference, for example). Let me pull all of the notions discussed so far into a diagram. Once the patterns have been recognized and classes identified, the resulting classes can be used for all sorts of applications such as Recommendation Engine, Language Translation, Fraud Detection and many others. The approach I outline above allows you to take a unified approach till the application development stage. In doing so, the unified approach also points out new paths ahead for ML! Some readers would have noticed an undertow of dichotomies while reading this “opinion piece”: Theoretic vs Heuristic; Formal vs Ad hoc; Mathematics vs AI; Electrical Engineering vs Computer Science academic departmental affiliations! I am firmly in the former camps. However, as an engineer, I am personally happy to start with heuristic solutions but quickly put them on firm mathematical foundations before “gotchas” and unintended consequences of ad hoc methods catch up with me. J The unification of ML proposed here opens up a multilane highway – join the journey and create more breakthroughs with us or on your own! In this blog, I have not provided many references – web search will get you most; Pedro Domingos’ “The Master Algorithm” book is an excellent source of ML-related literature. For the newer and less familiar work, please contact me directly.","excerpt":"Pedro Domingos in his new book, “The Master Algorithm”, has done us a huge favor. As is true of any emerging technology field, Machine Learning (ML) is a “bag of tricks” today; it takes a while for a unifying framework to emerge. Then, one can see various aspects of ML as special cases of a […]","categories":[],"tags":[],"author_name":"PG Madhavan","publish_date":"2015-12-03T04:37:15","publication_year":"2015","word_count":998,"keywords":["Go","machine learning","TPU","AI","neural network","ML","deep learning","analytics","R","fraud detection"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","analytics","fraud detection","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/machine-learning-a-unifying-perspective-new-paths\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10017017,"title":"How To Leverage GPUs For Recommendation Engines At Scale","content":"Using GPUs at scale comes with various challenges due to compute-intensive and memory-intensive components. For instance, GPUs that train state-of-the-art personal recommendation models are largely affected by model architecture configurations such as dense and sparse features or dimensions of a neural network. These models often contain large embedding tables that do not fit into limited GPU memory. The majority of deep learning recommendation models are trained on CPU servers, unlike language models, which are trained on GPU systems. This is because of the large memory capacity and bandwidth requirement of embedding tables in these models. The memory capacity of embedding tables has increased dramatically from tens of GBs to TBs throughout the industry. At the same time, memory bandwidth usage also increased quickly with the increasing number of embedding tables and the associated lookups. According to reports, over the last 18-month period, the compute capacity for recommendation model training quadrupled at Facebook’s data center fleet. Among the total AI training cycles at Facebook, more than 50% has been devoted to training deep learning recommendation models. (Source: Paper by Acun et al.) Leveraging GPUs At Scale So, how does a large organisation such as Facebook leverage GPUs for recommendation models at scale? In order to understand the underlying training system architectures to develop a better understanding of GPU training performance, the researchers at FAIR investigated various components(shown in red in the picture below) in an architecture that affects the efficiency of the model. (Source: Paper by Acun et al.) The rationale behind recommendation model configurations is as follows: Features categorised into two distinct types: dense and sparse. Dense features are scalar inputs, whereas sparse features often encode categorical traits or relevant IDs.The computational cost of each dense feature is roughly the same. Sparse features determine how many embedding tables there will be.Typically, Hashing is a common method for limiting the size of embedding tables. (Note: Embeddings make it easier to do machine learning on large inputs like sparse vectors. An embedding captures some of the input’s semantics by placing semantically similar inputs close together in the embedding space, which can then be learned and reused across models.) For their experiments, the researchers used Facebook’s 8-GPU training systems called the “Big Basin” and a new prototype that goes by the name “Zion”. The researchers observed that training recommendation models exhibit both data parallelism and model parallelism. Researchers wrote that while training deep learning recommendation models on CPUs offer memory capacity advantage, a large degree of parallelism in the training process, which could be unlocked by utilising accelerators, is left unexploited. Currently, to exploit data parallelism, in production, each trainer server holds a copy of the model parameters, reads its own mini-batches from reader servers and performs Elastic-Averaging SGD (EASGD) update with the centre, dense parameter server. Within a trainer, HogWild! Most of the machine learning is about finding the right kind of variables for converging towards reasonable predictions. Introduced a decade ago, Hogwild! is a method that helps find those variables efficiently. Parallelisation of the basic SGD doesn’t help. Its inherently sequential nature limits stochastic Gradient Descent’s (SGD) scalability; it is difficult to parallelise. There was no way around memory locking, which deteriorated the performance. Memory locking was essential to reduce latency for between processes. Hogwild!, enabled the processors to have equal access to shared memory and update individual components of memory at will. According to the researchers, an alternative is to train recommendation models on Facebook’s Big Basin GPU servers, initially designed for non-recommendation AI workloads. Big Basin architecture features 8 NVIDIA Tesla GPUs enabling 15.7 teraflops of single-precision floating-point arithmetic per GPU. Whereas, the bandwidth and CPU compute capacity of Zion is much larger, with ∼2 TB system memory and ∼1 TB\/s memory bandwidth. Embedding tables are distributed among GPUs; different partitioning strategies can be used, such as table-wise or row-wise partitioning. For GPU servers like those of Big Basin, storing embedding tables on the GPUs enables offloading all model operations to be done on the GPU, minimising the CPU usage and CPU-GPU copy operations. “Storing the embedding tables on the system memory of the CPUs of the GPU server would be a good option for servers with large system memory,” explained the researchers. They also talk about using a hybrid alternative where some of the embedding tables are stored on the GPUs and some are stored on the system memory. This helps when the embedding tables do not fit on the GPU. Though building recommendation models is still challenging, the researchers elaborated a fresh perspective by segmenting components that play a crucial role in training efficiencies: different levels of CPUs, memory capacity, memory and network bandwidth requirements, dense and sparse features, batch sizes, embedding table hash sizes, and MLP(neural network) dimensions. Key Takeaways Ever-increasing sizes of deep learning recommendation models, particularly embedding tables, introduce significant system design challenges.The most efficient choice of a hardware system depends on the model parameters such as the number of dense and sparse features, embedding table sizes, feature interaction types, and neural networks dimensions.Embedding tables placement on different hardware systems (CPU vs GPU) requires different strategies.Researchers introduce Zion — a next-gen platform for deploying recommendation models at scale. With this work, the researchers have tried to explore and offer solutions with Facebook’s production-scale deep learning recommendation models that can be used to guide the design of training infrastructures. Download the original paper here.","excerpt":"Using GPUs at scale comes with various challenges due to compute-intensive and memory-intensive components. For instance, GPUs that train state-of-the-art personal recommendation models are largely affected by model architecture configurations such as dense and sparse features or dimensions of a neural network. These models often contain large embedding tables that do not fit into limited […]","categories":["Deep Tech"],"tags":["GPUs","recommendation engine"],"author_name":"Ram Sagar","publish_date":"2021-01-04T14:00:00","publication_year":"2021","word_count":899,"keywords":["Go","machine learning","AI","neural network","ML","Scala","RAG","recommendation engine","deep learning","GAN","GPUs","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","RAG","R","Go","Scala","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/gpu-recommendation-scale-facebook-ai-research\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":33925,"title":"This Neural Network Can Generate Lyrics Just Like Your Favourite Artiste","content":"Artificial Intelligence is the notion of machines that exhibit intelligence and mimic cognitive functions that are usually associated with humans such as learning, reasoning, predicting, planning, recognising, and even problem-solving. Now, AI tools are being increasingly integrated into technological solutions and now it can even be induced in to generate lyrics that match the style of unique music artists. Now, the researchers at the University of Waterloo, Canada, have recently achieved this feat by developing a  system that can generate such song lyrics. A System That Can Generate Song Lyrics Of Your favourite Artist The researchers explored how neural generative models could assist songwriters and musicians in writing song lyrics. The outputs of these models were used to generate song lyrics which were based on original artwork or compositions. Instead of generating lyrics for an entire song, the model generated suggestions for lyrics which lines in the style of a specified artist. The system developed by Vechtomova and her colleagues were based on a neural network model called variational autoencoder (VAE) with artist embeddings, a multi-dimensional vector of real numbers and a CNN classifier which is trained to predict artists from MEL spectrograms of their song clips and can learn by reconstructing original lines of text. The purpose of this model is that unusual and creative arrangements of words in the generated lines can inspire the songwriter to create original lyrics at a later stage. The system also conditions, the generation on the style of a specific artist is done in order to maintain stylistic consistency of the suggestions. The reason for using such generative models is mainly to augment the natural creative process when an artist gets inspired to write a song based on something they have read or heard. According to Olga Vechtomova, University of Waterloo, one of the minds behind this innovation presses on the fact that the base of the research is the ultimate result  of curiosity that imbibes in each and every human being and that is to know whether a machine can generate lines that could sound like the lyrics of my favourite music artists. While working on text generative models, The research team founded that neural networks had the capability to generate some of the most creative and impressive lines of text. For the musically inclined research team, it was a natural next step to be able to explore whether a machine could learn the very essence of a specific music artist’s lyrical style, including choice of words, themes and sentence structure, to generate novel lyrics lines that sounded similar to the artist in question. The motivation behind using artist embedding to condition the generation lyrics lines in the style of each artist was to reflect the differences between artist embeddings and their lyrical as well as musical styles. Their findings suggested that artist embeddings were useful for generating lyrics that matched an artist’s style. Many lines generated by the model were perfectly aligned with the artist as it was conditioned on in such a way that it reflected the themes generally addressed their music. The very same system generated two poems which included the collection that was submitted to the NeurIPS 2018 Workshop on ML for Creativity and Design. Vechtomova created each of these poems by selecting lines generated by the VAE and arranged them in an artistically meaningful way was done to the individual lines were even not edited except for adding capitalization and punctuation marks. According to Vechtomova, the generated lines often contains the words of an artist, these are used in an interesting new way, expressing novel thoughts not found in the original lyrics too. Some of the generated lines also convey new and powerful poetic imagery, expressed using stylistic devices such as metaphors and oxymorons, while remaining true to the style of the artist. Outlook The system created by Vechtomova and team will be used by inspired artists who are composing lyrics for new songs. Rather than replacing lyric composers, the researchers hope that it will provide new ideas, which artists could modify and develop even more creative lyrics of their own. According to Vechtomova, the system is not meant to replace a music artist, but to be used as a source of inspiration during the songwriting process. In the music world, this will be analogous to a synthesizer that will generate an infinite number of sounds, from which an artist can then create a song. Similarly, this tool will also generate an infinite number of novel lines that artists can use in any way they like to compose their own lyrics. In the future, the teams plan to work on models that can learn new themes and vocabulary from additional sources and use them to generate lyrics in the style of a given artist. The team will also be exploring how such a system could potentially be used by music artists as a source of inspiration.","excerpt":"Artificial Intelligence is the notion of machines that exhibit intelligence and mimic cognitive functions that are usually associated with humans such as learning, reasoning, predicting, planning, recognising, and even problem-solving. Now, AI tools are being increasingly integrated into technological solutions and now it can even be induced in to generate lyrics that match the style […]","categories":["AI Features"],"tags":["Artificial Neural Network","University of Waterloo"],"author_name":"Martin F.R.","publish_date":"2019-01-23T05:42:19","publication_year":"2019","word_count":819,"keywords":["Go","API","artificial intelligence","TPU","AI","University of Waterloo","neural network","Artificial Neural Network","ML","VAE","CLIP","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","TPU","R","Go","API","CLIP","VAE"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/neural-lyrics-favourite-artiste\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077954,"title":"Same Same, But Different: Reddit’s Step Forward Towards a Mass Adoption of Web 3","content":"In July this year, Reddit created its own NFT marketplace for users to buy and sell blockchain-based digital avatars through their platform. It partnered with Polygon, an Ethereum-compatible platform, to mint the avatars and store and manage them in the Vault. The Reddit–Polygon initiative received an overwhelming response. In fact, Reddit’s Chief Product Officer Pali Bhat recently revealed that over 3 million crypto wallets have been created through Reddit’s Vault platform to date. Of these, 2.5 million wallets were used to purchase digital avatars at the NFT marketplace. Crypto Twitter has been brooding over these numbers, which are believed to be higher than OpenSpace—the largest marketplace for NFTs. https:\/\/twitter.com\/ABBBBBB_NFT\/status\/1582605563795410945 Selected Reddit NFT Avatars also made it into the top 10 list of collectibles on the Open Sea platform. The co-founder of Polygon, Sandeep Nailwal, took to Twitter to share the analytics of the Reddit Collectible Avatars user traction. The data has shown the cumulative sales volume of the collectible NFT avatars crossing $2.3 Million. So, how was Reddit able to get users on board in its foray into Web 3? Reddit’s ‘Collectible Avatars‘ In 2020, Reddit launched the ‘Avatar Builder’ feature, where users could personalise their appearance on the platform. In this new and improved version from 2015, Redditors could create an avatar from a set of countless accessories, outfits, and hairstyles. The launch added to the Reddit Premium programme by making special accessories exclusively available to those users who were a part of it. The avatar appeared on the users’ profile page and profile card. The project was considered a “unique way to create your identity on Reddit”. Simultaneously, Reddit collaborated with Netflix, Riot Games, and the Australian Football League (AFL) to provide custom avatars to its users. The Collectible Avatars Series that Reddit launched this year had them licensing nearly 30 independent artists to create limited edition avatars and sell them on the Reddit platform. Reddit defined the Collectible Avatars to be different from their other avatars in three ways: The Collectible Avatars are available for direct purchase, with artists being paid for each collectible sold. These avatars are blockchain-backed, providing buyers full ownership of the collectible. And finally, the users don’t need cryptocurrency to purchase the collectibles. They can be bought with government currency, with all items being sold for a fixed amount. The ease of understanding, along with the accessibility that Reddit has offered with the collectible series, has led its users to embrace the move. https:\/\/twitter.com\/Mtitus6\/status\/1583289981162106880 Mass adoption of Web 3 The credit for the acceptance of new technology goes to the carefully curated way in which Reddit managed to avoid the use of any futuristic buzzwords like ‘NFTs’ or ‘Crypto’. It managed to sell what technically are NFTs under the guise of ‘Collectible Avatars’, an extension of their already existing feature of ‘Avatar Builder’. Further, Reddit also avoided the mention of several off-putting terms and used favourable terms in their statements, such as this: “To store and manage Collectible Avatars, redditors will use their Vault—a blockchain-powered wallet on Reddit. A Vault gives users a specific digital wallet address that works across Ethereum-compatible blockchains; it’s the same Vault where redditors currently store Community Points.” Reddit used the classic selling technique of ‘same same, but different’. They made the process of creating a Vault essentially Web 2. But, as with creating a crypto wallet, in the Vault too—users will be issued with a random 12-word seed phrase. Though, to the unacquainted, this could easily be interpreted as an additional security step that users must complete. Social media platforms are galore with users expressing how Reddit’s NFT marketing is monumental in bringing new people into the Web 3 wonderland. Incredible that @Reddit just onboarded so many people into crypto, mainly to buy digital collectibles. 40K PFP’s sold out very quickly..they never mentioned “NFT” while doing..and allowed credit card purchases. THIS IS HOW YOU ONBOARD NORMIES. #web3 #eth #nft— Maxbrain Capital (on Farcaster) (@viybz) October 21, 2022 Several of these platforms, like Facebook, Instagram, Twitter, and Youtube, also jumped onto the NFT train in 2022. It’d indeed be interesting to see how they follow suit seeing Reddit’s slow and steady reinvention of their platform. However, as far as Reddit goes, this seems only like the beginning. One would have to wait and see what tools they explore to expand their market in the Web 3 space. Reddit’s commitment to Web 3 can also be seen in its statement: “Reddit has always been a model for what decentralization could look like online; our communities are self-built and run, and as part of our mission to better empower our communities.”","excerpt":"Reddit’s NFT surge has brought many new audiences into the decentralised world. It will be interesting to see how other social media platforms mainstream the next big thing in tech.","categories":["IT Services"],"tags":[],"author_name":"Ayush Jain","publish_date":"2022-10-25T15:00:00","publication_year":"2022","word_count":770,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","Git","Ray","analytics","R"],"extracted_tech_keywords":["AI","analytics","Ray","R","Go","Git","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/same-same-but-different-reddits-step-forward-towards-a-mass-adoption-of-web-3\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167020,"title":"AMD Acquires AI Infrastructure Firm ZT Systems for $4.9 Billion","content":"American computing giant AMD has completed the acquisition of ZT Systems, a company that provides computer, storage and accelerator solutions for cloud and telecom service providers. Last year, AMD announced its plans to acquire ZT Systems for $4.9 billion. The move will accelerate the design and deployment of AMD-powered AI infrastructure at scale, optimised for the cloud. AMD indicated that this will facilitate innovative end-to-end AI solutions by integrating AMD CPUs, GPUs, networking, and ROCm software while expediting the deployment of large-scale, cloud-optimised AI infrastructure. “With ZT Systems, we’re bringing in an incredible team,” said Lisa Su, CEO of AMD, in an interaction with Yahoo Finance. “We’re bringing in over a thousand design engineers who are working hand in hand with our silicon engineers, as well as our customers, and systems engineers,” she added. ZT Systems is a key partner to some of the world’s largest firms, such as Microsoft and Meta. Founded in 1994, the company focused on desktop PCs and server message block (SMB) servers, but then transitioned to making data center servers in 2004. “This [the acquisition] is all about ensuring that we have all of the pieces, both from a hardware standpoint, and lots of investment in software. And ZT is a critical piece in that supply chain and that value chain to give us the systems knowledge,” added Su. Furthermore, AMD announced on Monday that its recently revealed 5th generation AMD EPYC processors are powering Oracle Cloud Infrastructure (OCI) E6 Standard shapes. AMD’s hardware will now enable OCI Compute E6 shapes to deliver up to a two-fold increase in cost-to-performance, compared to the previous E5 instance generation. In February, AMD announced its fourth quarter results for last year, revealing that its data centre segment revenue grew by 69% year-over-year to $3.9 billion. The company generated $12.6 billion in revenue from the data centre segment for the entire last year.","excerpt":"“With ZT Systems, we’re bringing in an incredible team,” said Lisa Su, CEO of AMD.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AMD"],"author_name":"Supreeth Koundinya","publish_date":"2025-04-01T17:04:29","publication_year":"2025","word_count":314,"keywords":["AMD","programming_languages:R","AI","RAG","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amd-acquires-ai-infrastructure-firm-zt-systems-for-4-9-billion\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10048381,"title":"50 Most Influential AI Leaders In India – 2021","content":"Analytics India Magazine brings to you its annual list of the 50 Most Influential AI Leaders In India for the year 2021. Selected on the basis of the expertise they hold in the field of AI and Data Science, the list recognises the exceptional work these professionals have accomplished over the last year. AIM 100: The Most Influential Leaders in AI 2024As AI reshapes our world, AIM introducing the most influential global leaders in AI. AI has seen increasing importance in Indian organisations, public and private, over the last decade. The government has also shown intent to leverage AI for the country’s development through various initiatives. To drive AI growth successfully, we need leaders who are well-versed with the dynamic nature of the industry. Given the pace at which AI is moving, we need top management that can come up with innovative strategies while overcoming multiple challenges. Leaders in this list have shown great ability to lead their teams to develop state-of-the-art AI solutions and support their companies or clients. Despite the pandemic, they have shown the potential in adapting to different circumstances to drive innovation and growth. The final record is the work of over two months as AIM invited nominations from various leading organisations in India. The selection is done by a panel that includes members internal and external to AIM. The list is finalised primarily based on the criteria of the leader’s impact on the AI ecosystem in India. Until last year, AIM published lists for the top 10 leaders. You can access them at the links below: 2020 | 2019 | 2018 | 2017 | 2016 | 2015 This year, AIM has expanded the scope by including a wider set of professionals. Here is the list of the top 50 influencers in the Indian AI industry for 2021, presented in alphabetical order. Abhinav Shashank CEO & Co-founder at Innovaccer Ambica Rajagopal Group Chief AI Officer at Michelin Anand Ekambaram Country Manager, India at Tableau Anirban Nandi Head of Analytics (Vice President) at Rakuten India Anish Agarwal Director – Data & Analytics, India at NatWest Group Ankush Gadi Director – Data Management & Analytics at CRISIL Anshu Sharma CIO, Consumer Private Business Banking at Standard Chartered Bank Anshuman Gupta Vice President – Data Science at MiQ Digital India Asha Poulose Johnson Vice President, Data and Analytics, GE Healthcare Ashwin Mittal CEO at Course5i Atul Jalan CEO at Algonomy Biswanath Choudhary Vice President & Global Head of Analytics Operations at Evalueserve David Zakkam Vice President – Analytics at Swiggy Devendra Sharnagat Sr. EVP – Data Analytics and CVM at Kotak Mahindra Bank Dinanath Kholkar Vice President and Global Head of Analytics & Insights at TCS Gaurav Dhall Vice President – APJ & Middle East, Managing Director – India at SingleStore Gunjan Gupta Director- Collections & Contact Centers Technology, US Consumer Bank at Barclays Kaushik Sanyal GCP Data and AI Capability Lead at Accenture Manish Gupta Director at Google Research India Manoj Madhusudanan Head of dunnhumby India Mathangi Sri VP Data Science & Head of Data at Gojek Mihir Kittur Co-Founder and Chief Commercial Officer of Ugam Neeraj Pratap Chief Operating Officer at Hansa Cequity Nidhi Pratapneni SVP, Product, Analytics & Modelling, Public Affairs at Wells Fargo Noshin Kagalwalla VP & Managing Director, SAS India Padmashree Shagrithaya Global Head – Analytics & Data Science at Capgemini Pankaj Rai Group Chief Data and Analytics Officer at Aditya Birla Group Paritosh Anand SVP and Group Head for Analytics & Strategic Initiatives at Reliance Industries Limited Phanimitra B Head Digital & Process Excellence, India & Emerging Markets at Dr Reddy’s Laboratories Pradeep Menon Managing Director & GCC Head – HSBC Technology Prakash Mallya VP & MD – Sales, Marketing & Communications Group, Intel India Prashanth Kaddi Partner, Analytics & Cognitive at Deloitte Prashanth Rao Head of Application Engineering at MathWorks, India Prithvijit Roy CEO and Co-founder at BRIDGEi2i Rahul Singh Co-Founder & Chief Analytics Officer at Sigmoid Analytics RB Rajendar Head – Analytics and Information Management, India & Head – Citi Solutions Center, Bengaluru Rohini Srivathsa National Technology Officer – Microsoft India Ruble Joseph VP – Head of CoE for Data Science, Analytics and Consulting at eClerx Saurabh Saxena Intuit India Site Leader & Vice President, Product Development Sayandeb Banerjee Co-Founder and CEO at TheMathCompany Shailesh Kumar Chief Data Scientist, CoE AI\/ML at Reliance Jio Shanti Swaroop Mokkapati Head – Analytics & Market Intelligence, Global Procurement at Abbott Shashank Dubey Co-Founder & Chief Revenue Officer at Tredence Inc. Sindhu Gangadharan SVP and MD at SAP Labs India & SVP and Head of SAP User Enablement Unit Smitha Ganesh AI Director – Data Science Innovation & Execution at Ericsson Sreekanth Menon VP – Data Science at Genpact Subramanian M S Head of Category Marketing and Analytics at BigBasket Swati Jain Vice President Analytics at EXL Service Vinod Ganesan Country Manager at Cloudera Vishal Dhupar Managing Director, Asia South at NVIDIA","excerpt":"Analytics India Magazine brings to you its annual list of the 50 Most Influential AI Leaders In India for the year 2021. Selected on the basis of the expertise they hold in the field of AI and Data Science, the list recognises the exceptional work these professionals have accomplished over the last year. AI has […]","categories":["AI Features"],"tags":["AI leaders","analytics leaders india"],"author_name":"AIM Media House","publish_date":"2021-09-27T09:44:25","publication_year":"2021","word_count":817,"keywords":["data science","Go","GCP","AI","ML","analytics leaders india","Git","RAG","Aim","analytics","AI leaders","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","GCP","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/50-most-influential-ai-leaders-in-india-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10076140,"title":"Users can now Generate Text-to-3D models using 2D Diffusion","content":"A group of researchers from Google have adapted a new approach to 3D synthesis. Users can now generate 3D models with text prompts as input. The new method, called ‘DreamFusion’, uses 2D Diffusion and is set to bring notable advancements to text-to-image synthesis. Typically, advancements in AI generative systems are driven by diffusion models that are trained on billions of image-text pairs. Researchers claim that such an adaptation of 3D model synthesis would require large-scale datasets of labelled 3D assets and efficient architectures for de-noising 3D data—neither of which currently exist. Instead, the team circumvented such limitations to use a pre-trained 2D text-to-image diffusion model that performs text-to-3D synthesis. The researchers optimised a randomly-initialised 3D model called ‘NeRF’ (a Neural Radiance Field) via gradient descent so that the renderings in 2D from random angles achieve lesser loss. An excerpt from the blog says, “The resulting 3D model of the given text can be viewed from any angle, relit by arbitrary illumination, or composited into any 3D environment. Our approach requires no 3D training data and no modifications to the image diffusion model, demonstrating the effectiveness of pretrained image diffusion models as priors.” How does it work? A text-to-image generative model called ‘Imagen’ is used to optimise a 3D scene. The research also proposes Score Distillation Sampling (SDS)—a way to generate samples from a diffusion model by optimising a loss function—allowing users to optimise samples in an arbitrary parameter (3D) space. A 3D scene parameterization, similar to Neural Radiance Fields or NeRFs, is used to define the differentiable mapping. While SDS produces reasonable scene appearance, DreamFusion instils additional regularisers and optimisation strategies to improve geometry. The resultant trained NeRFs are coherent—with surface geometry and high-quality normals.","excerpt":"The new method called ‘DreamFusion’ uses 2D Diffusion to generate diverse 3D models, bringing advancements to text-to-image synthesis.","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-09-30T13:13:16","publication_year":"2022","word_count":285,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","diffusion models","Aim","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","diffusion models","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/users-can-now-generate-text-to-3d-models-using-2d-diffusion\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":23350,"title":"&#8216;Mind Reading&#8217; Algorithm : Here&#8217;s What Kyoto Neuroscientists&#8217; Research Says","content":"Deep image reconstruction from human brain activity Machine learning-based analysis of human functional magnetic resonance imaging (fMRI) patterns have enabled many discoveries. The application of computer vision in neuroscience has also led to many approaches that have given us insights about how the brain works and how it carries out its functions. Research coming out of Kyoto University has shown that careful analysis of functional magnetic resonance imaging (fMRI) can enable the visualisation of perceptual content in our brain. In a way the research shows that we can project images from our brains analysing fMRI images. The neuroscientists in their research, present a novel image reconstruction method, in which the pixel values of an image are optimised to make it a deep neural network (DNN) features similar to those decoded from human brain activity at multiple layers. They found that the generated images resembled the stimulus image that were shown to participants, both were natural images and artificial shapes. The model that the researchers trained with natural images, was successful in generalising to the reconstruction of artificial shapes. This fact itself shows that the model indeed ‘reconstructs’ or ‘generates’ images from brain activity and simply doesn’t match them. The results suggest that hierarchical visual information in the brain can be effectively combined to reconstruct perceptual and subjective images. Visualisation : Greatest Challenge Of Neuroscience The externalisation and visualisation of states of the mind is a challenging goal in neuroscience. Although decoding and encoding methods that could render human brain activity into images existed, they were not very effective. But those methods were essentially limited to image reconstruction with low level image bases, hence failing to combine visual features of multiple hierarchical levels. The researchers present a novel approach, named deep image reconstruction, to visualise perceptual content from human brain activity. The researchers combined the feature decoding from fMRI signals and the methods for image generation recently developed in machine learning. The reconstruction algorithm starts from a random image and iteratively optimise the pixel values so that the DNN features of the input image become similar to those decoded from brain activity across multiple DNN layers. In this process the resulting optimised image is taken as the reconstruction from the brain activity. There was also a deep generator network (DGN) as a prior to make reconstructed images similar to natural images. The decoders are trained to predict DNN features of viewed images from fMRI activity. Stages of Experiments The experiments that the researchers conducted were of 5 distinct stages : Training natural image sessions, Test natural image sessions, Geometric-shape sessions Alphabetic-letter sessions Mental-imagery sessions Deep Image Reconstruction Architecture The decoder were trained using fMRI data measured while subjects were viewing natural images. The trained decoders were then used to predict DNN features from independent test fMRI data collected during the presentation of novel natural images and artificial shapes and during mental imagery. And now the researchers sent the features to the reconstruction algorithm. These reconstructions obtained with the deep generator network capture the objects’ dominant structures in the images. Further, fine structures reflecting semantic aspects like faces, eyes, and texture patterns were also generated in several images. These impressive results were confirmed with a previously published dataset, reproducing quantitatively similar reconstructions with those in the present study. Effect of natural priors and multilayer networks on reconstructions The effect of the natural image prior, with and without DGN was also measured. The results show that DGN based natural image prior is enormously useful and  enhances perceptual similarity of reconstructed images. To understand the importance of having multiple layers of DNN a side study was also done. An independent rater was presented with an original image and a pair of reconstructed images, both from the same original image but generated with different combinations of multiple layers, and indicated which of the reconstructed images looked more similar to the original image. The assessment showed that reconstructions from a larger number of DNN layers were better rated. Seen natural image reconstructions. The researchers made sure that the methods were not restricted within the specific image domain. They did this by testing whether it is possible to generalise the reconstruction to artificial shapes. This was a challenging task since the training is solely on natural images.  The results show that artificial shapes were successfully reconstructed with moderate accuracy, 69.4% by pixel-wise spatial correlation, 92.3% by human judgment, indicating that the model indeed ‘reconstructs’ or ‘generates’ images from brain activity. With the tests and experiments it was successfully proved that the approach could provide a unique window into our internal world by translating brain activity into images via hierarchical visual features. The researchers used signed-rank tests to examine differences of assessed reconstruction quality from different conditions. It was also found that emphasising high-level visual information in hierarchical visual features may help to resolve the ambiguity of luminance by incorporating information about semantic context. Conclusion : This series of experiments in Kyoto developed a method to “see” inside people’s minds using an fMRI scanner, which detects changes in blood flow in the brain. According to the team at Kyoto University, this breakthrough opens a “unique window into our internal world“.","excerpt":"Machine learning-based analysis of human functional magnetic resonance imaging (fMRI) patterns have enabled many discoveries. The application of computer vision in neuroscience has also led to many approaches that have given us insights about how the brain works and how it carries out its functions. Research coming out of Kyoto University has shown that careful […]","categories":[],"tags":["neuroscience"],"author_name":"Abhijeet Katte","publish_date":"2018-04-06T09:24:28","publication_year":"2018","word_count":862,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","programming_languages:Go","computer vision","ViT","neuroscience","R","ai_applications:computer vision"],"extracted_tech_keywords":["AI","machine learning","neural network","computer vision","R","Go","ViT","programming_languages:R","programming_languages:Go","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mind-reading-algorithm-kyoto-neuroscientists-research\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":29867,"title":"Synechron Launches 11 AI Data Science Accelerators For BFSI Sector","content":"Faisal Husain, Co-founder and CEO of Synechron. (Image credit: Synechron Inc\/YouTube) Synechron Inc, the noted global financial services consulting and technology services provider, this week announced the launch of their Artificial Intelligence Data Science Accelerators for the Banking, Financial Services and Insurance (BFSI) firms. These four new solution accelerators will help financial services and insurance firms solve complex business challenges by discovering meaningful relationships between events that impact one another (correlation) and cause a future event to happen (causation). Following the success of Synechron’s AI Automation Programme Neo, Synechron’s AI Data Science experts have developed a powerful set of accelerators that allow financial firms to address business challenges related to investment research generation, predicting the next best action to take with a wealth management client, high-priority customer complaints, and better predicting credit risk related to mortgage lending. The Accelerators combine Natural Language Processing (NLP), Deep Learning algorithms and Data Science to solve the complex business challenges and rely on a powerful Spark and Hadoop platform to ingest and run correlations across massive amounts of data to test hypotheses and predict future outcomes, said the company in an official statement. The Data Science Accelerators are the fifth Accelerator program Synechron has launched in the last two years through its Financial Innovation Labs (FinLabs), which are operating in 11 key global financial markets across North America, Europe, Middle East and APAC; including: New York, Charlotte, Fort Lauderdale, London, Paris, Amsterdam, Serbia, Dubai, Pune, Bengaluru and Hyderabad. With this, Synechron’s Global Accelerator programs now include over 50 Accelerators for Blockchain, AI Automation, InsurTech, RegTech, and AI Data Science and a dedicated team of over 300 employees globally. They also demonstrate Synechron’s commitment to research & development, innovation, and upskilling employees. The AI Data Science Accelerators include: Syn-AI and Causality – is a powerful Data Science platform that ingests large volumes of structured and unstructured data and powers the business case Accelerators. It uses the latest advancements in parallel computing and delivers a scalable platform for rapidly-analyzing massive data collections and identifying meaningful Granger Causal relationships. Visual Research – automatically generates personalized buy- and sell-side research reports with automated data collection, synthesis, and analysis, lowering costs while enabling systematic research not possible with manual processes. Informed Investing – allows wealth managers to receive alerts for critical events related to the assets they manage such as geopolitical and sector events that impact security prices and buy\/sell recommendations. Customer Complaints Management – identifies the factors driving customer complaints and prioritizes each claim by the likelihood of being disputed or escalated to enable banks to more quickly and proactively resolve the most critical complaints. Credit Risk – empowers banks to manage their credit portfolios proactively, enabling users to drill down into the factors driving likely credit events and to proactively manage individual risks ranked by probability of incurring a specific credit event.","excerpt":"Synechron Inc, the noted global financial services consulting and technology services provider, this week announced the launch of their Artificial Intelligence Data Science Accelerators for the Banking, Financial Services and Insurance (BFSI) firms. These four new solution accelerators will help financial services and insurance firms solve complex business challenges by discovering meaningful relationships between events that […]","categories":["AI News"],"tags":["best technology to learn for future","BFSI","Data Science"],"author_name":"Prajakta Hebbar","publish_date":"2018-11-02T10:09:49","publication_year":"2018","word_count":475,"keywords":["data science","Go","API","artificial intelligence","AI","BFSI","Scala","NLP","Aim","deep learning","Data Science","R","best technology to learn for future"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","NLP","data science","Aim","R","Go","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/synechron-ai-data-science-accelerators-bfsi\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10085262,"title":"An Accenture Analyst’s Affair with the Beloved Vision AI YOLOv5","content":"For Alessandro Mondin, the shift towards AI\/ML research hasn’t exactly been seamless. Mondin was a market analyst initially, following which he started participating in hackathons of his own accord. He currently works as a digital transformation analyst with Accenture Song (previously known as Accenture Interactive). However, Mondin has come far while continuing to work on AI\/ML projects outside of his professional role. Analytics India Magazine spoke to Mondin about his descriptive work with the YOLOv5 model architecture, the future of computer vision and what motivates his independent research. AIM: We wanted to understand more about your article explaining YOLOv5. Can you tell us how long it took you and what went into it? Alessandro: Since I’m not a professional or working as a researcher within the industry, this is a side project I have been working on to build a portfolio that can help me eventually. I decided to work with YOLOv5 because a friend of mine working in the industry told me that everyone uses it but there wasn’t a lot of explanatory work around it. Most of the work around it just describes how to use YOLOv5 but they don’t explain how the model architecture is. It took me a long time, especially because if you don’t have expertise with the YOLO architectures, I had to attempt to understand them from scratch and they’re quite tricky. So, overall it took me more than four months. Building inside the architecture was quite fast and took me between three weeks and a month. But the training pipeline is the really tricky part and time consuming. The loss functions and the data loader—these are the parts that are really, really long, especially if you’re doing them from scratch. The model takes up about 20% of the time. AIM: What is the importance of the YOLO series in computer vision? Alessandro: I was trying to deepen my knowledge in object detection and YOLO are the most important algorithms used in terms of the ratio between mean average precision and inference time. So, they are very accurate and the inference speed is incredible. Additionally, you can also use them on the CPU. So, they are single shot detectors and strike a great balance between average precision and speed. So, for YOLOv5, there are no articles that dive deep into the details like mine. I tried to look around and I found one article but even that lacked sufficient examples. Most articles will tell you how to use it but not how it works. But this explanation is extremely useful for other researchers and engineers to understand atomically how YOLOv5 performs inference. AIM: Can you describe what your current role entails? Alessandro: I’m currently working as a digital transformation analyst with Accenture Song as a consultant and I’m trying to build a strong resume because I have a deep interest in AI\/ML. I come from a non-technical background so HRs obviously end up choosing engineers that typically come out of STEM. So, I am trying to prove that my interest in the area is stronger than my atypical background. AIM: How has the transition from a market analyst to data science and now AI\/ML research been for you? What were the challenges you faced? Alessandro: I’m not going to lie, it is tough. You really need to put in the work. But if you’re really passionate and if you’re willing to put in a lot of hours, hundreds of hours, you end up learning a lot. And that’s the most important thing in the beginning. So, being an amateur, all the knowledge and all my points of view are not the points of view of the professionals working for ten years in the industry. Everything I have picked up has been on my own. AIM: What are some of the biggest trends in computer vision that you are looking forward to right now? Alessandro: The problem with working on side projects is that you cannot really dive into two exciting topics because you lack GPUs. I know that the tech industry is really investing into action recognition right now. There’s a lot of work with vision transformers. With regard to blacktips of papers, I’m really interested in the touch points between computer vision and natural language processing, which is image captioning. I want to work more in these areas now. AIM: While you were implementing YOLOv5 from scratch, what were some of the biggest lessons that you learned from that project? Alessandro: First off, you have to understand that we are not doing anything new, we are not implementing new algorithms or discovering something but you’re learning about something that is pre-existing —and this was the main reason for me to do it. I asked my friend about what would be the best algorithm that could be helpful to someone and YOLO was the one. It could be useful for anyone who uses YOLO-based architecture already and would like to understand how the model performs detections because there are plenty of ML engineers and computer vision engineers that just use the repository. So, essentially the main lesson is that you have to be very interested in the subject matter itself without looking for any kind of prize in the end—your goal has to be to learn.","excerpt":"YOLO are the most important algorithms used in terms of the ratio between mean average precision and inference time.","categories":["AI Features"],"tags":["Interviews and Discussions","YOLO"],"author_name":"Poulomi Chatterjee","publish_date":"2023-01-17T14:35:00","publication_year":"2023","word_count":882,"keywords":["data science","AI","ML","Transformers","computer vision","RAG","Aim","object detection","YOLO","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","computer vision","data science","analytics","Aim","Transformers","RAG","object detection","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/an-amazing-story-of-an-accenture-analyst-with-the-worlds-most-loved-vision-ai-yolo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10059030,"title":"Integrated risk information System for Chief Risk Officers","content":"The development of enterprise risk management (ERM) over the last two decades has brought forward the role of Chief Risk Officer (CRO) in the forefront. The various crisis such as the Enron scandal in 2001 led to the strengthening of corporate governance and the passing of the Sarbanes-Oxley Act in 2002. This helped improve ERM. The importance of the CRO role gained more prominence along with the accelerated development of ERM post-2008 economic crisis. Also, the application of risk-based capital in different developed markets led to an increased focus on ERM because ERM helps in optimizing the use of capital. The prevailing COVID 19 pandemic has increased claims many folds under both the life and non-life insurance sector. Such increased claims have wiped out the profits of many reinsurance companies for the last many years. The banking sector is equally impacted on the recovery of their loans due to the various government-led initiatives to help the borrowers who could not repay the loan due to failure in business or loss in jobs. Such economic and demographic turmoil have reinforced the importance of risk management and the role of CRO. This has made CROs more conscious of both existing and emerging risks. The purpose of this article is to highlight the importance of a risk management information system for CRO to address any gaps in the risk information that can be catastrophic for the survival of the financial institutions. If hidden risks or correlated risks are not spotted at an appropriate time, then emerging risks may crystalize leading to an insolvency situation which has been seen during the 2008 economic crisis. Therefore, there is a need for CRO to use advanced technology to have automated risk information management including the predictive capability to take mitigating action. CRO’s Challenges The CRO in the organization is the custodian of all risks whereas the front-line managers are focusing on business development where risks are secondary in nature to them. This brings a great responsibility of risk management on the shoulders of CRO. None of the risks can be missed and hidden risks should be brought out in light along with the deducing correlation of risks. The classic example of COVID 19 caught many organizations napping during the initial period of December 2019 and early January 2020 when COVID was restricted to China. At that time risks were highly underestimated, and foresight was missing, many of the correlated risks were missed out. Risk identification is the first step in the direction of the risk management process and subsequent steps cannot take place without a proper risk identification system. In most crises such as the 2008 global economic crisis, risk identification was missed out. One of the challenges in risk identification is it is based on a manual process having a risk of human interpretation. There are two ways to address this problem is either by improving the risk culture or bringing automation. Organizations are catching up with enhancing the risk culture, but the process is slow, and the first line of defence is not fully prepared to take the role of risk managers. Therefore, there is a need to have an automated system that correlates and bring forward the risks for CROs to take immediate action. The velocity of the risk is very crucial because it can beat the human thinking process to act leading to delays and massive losses. A threshold for the risk action taking must be defined that should popup risk like a sprinkler system to detect smoke for fire risk. Many risks function faces the challenges of scattered data or data not available for the purpose of risk assessment. Such lack of data or not one view of risk data can lead to either missing some of the key risks or not drawing meaningful conclusions about risks in conjunction with other risks. For example, during the COVID 19 non-life claims, many insurance companies were dependent on the exclusion clause in the policy bond about the loss in the business event due to pandemic is not being covered. However, different governments and regulators ordered to pay such claims leading to many court cases. When such claims are paid which were not priced is a straight loss. So, the exclusion clause that was sitting in the policy bond was the hidden risk no one thought about. Correlation of risks is another area that requires CRO’s attention. A big shock in a system sends ripples in different economic areas which have a direct impact on the financial institution. For example, in 2011, Japan’s Tsunami led to a 20% fall in prices in Tokyo’s Stock Market in the first two days of operation and a 30bps fall in Japan’s Sovereign Credit default swap. Also, the international equity market was equally impacted leading to falling in their prices and bond yields. Such shocks impact interest rate, inflation and liquidity concerning financial institutions’ cash flows and capital requirement. CRO need to re-assess the Company’s risk profile due to such events and raise the risks to the management and the Board for corrective action. Such re-assessment of risks is possible when CRO have an automated system that can re-calculate the risks in wake of such events. The availability of an integrated risk information system in one place is very important for CRO to assess the risks, draw correlations and take real-time risk-based decisions. To assess whether a risk team have consolidated risk information in one place or not; the following questions were asked in a social media survey (LinkedIn) Does your company have an integrated risk management system in one place? Total 55 votes were polled with the following results There is a clear inference that only 36% of voters polled had integrated risk management dashboard and the rest 74% have either scattered or semi-integrated risk dashboard. This indicates the need for an integrated risk dashboard for CRO’s easy to access all risk information for a comprehensive risk assessment. CRO’s needs To address existing, emerging hidden or correlated risks, one of the most important tools for CRO is to have an integrated risk information system that can provide all risk information at his or her fingertips. Here the CRO can slice and dice the risks for further analysis to understand the existing materiality and correlation of the risks Such a risk information system helps in identifying some of the key sources of risks which is a breeding ground of various risks. For example, adverse economic activity is a source of interest rate risk so a good information system on regular monitoring of the economic activity can help in advance spotting the interest rate risk. This can help the company in performing proper assets and liability management or changing the product portfolio towards less interest-sensitive products. It is important for CRO to have a robust risk information system not only from his or her internal sources but also from external sources as well such as market development, competitive landscape, economic and demographic trend, political developments, international conflicts etc. These external sources of risks can hasten up some of the risks leading to less time for CRO to act. Therefore, the risk information system should have a wider spectrum covering many different areas to spot risk early. Such a wider risk information system can give early warning signals which may otherwise be missed. Early warning risk spotting helps in averting the financial and reputational damage. Along with the integrated risk information system, the CRO also need to have a predictive risk capability to assess the risks for the future using data science and machine learning algorithms. Advanced algorithms have the capability to spot the trend and using the current inputs, such machine learning tools have been able to give the prediction by around 70% to 80% accuracy. The volume of data is the key in spotting hidden trends using regression and other models. The CRO also need to use advanced technologies together with big data sitting in their information system to extract meaningful information about customers’ need. The need-based selling helps in developing long term relationships with the customers. The automated data feed system in a dashboard can help CRO in taking real-time risk decisions which become crucial when risks are fast changing. This particularly happens when risks are about to crystalize, and events are fast changing. In such a situation, different risks start interacting and assessment of single risks are less meaningful. Technology in such a situation can prove to be very handy. Therefore, an integrated risk dashboard can prove to be a very helpful tool for CRO to have a comprehensive risk information system for all decision making. Risk Information Solutions The integrated risk information system should bring together all the risks in the form of the dashboard for CRO’s easy at one place access. This should have trend analysis, analysis of all existing risks, emerging risk along with the predictive capability etc. The features of the risk dashboard could have data on the cloud depending on the organization’s requirement, ease of navigation for comparing results, capturing historic trends for insight information, harmonized look and feel, taking notes and raising questions within the system for better decision making etc. The dashboard should also have a background data feed template to see revised trends. The risk information system should have templatized data input system from the risk repository either from the company’s risk register or to create a risk dump for inputting all risks. This risk repository must have inputs of all risks including companywide risks, economic, demographic, climate, international market, competitive trends etc. to get an output trend to assess the emerging risk profile of the company. The advantage of a structured data template is to use the previously captured data to compare the results from the current results and maintain the database. Some of the important sources of information for risk analysis are product risks, business plans, third party risks, strategic risks, operational risks, IT, cyber risks etc. Such risks information can be easily pulled together into an integrated system to slice and dice the data to unearth some of the hidden risks and correlated risks. For example, people risk as an independent risk may not sound daunting, but when looked at in conjunction with meeting the sales targets may throw challenges in meeting the annual profits. Such risks can be unearthed when complete risk information is available. The risks in the integrated system could have the following risks at a high level and then can be broken down further. Enterprise riskStrategic riskFinancial riskOperational riskExternal risks Enterprise Risk Management The CRO should have a good idea about the entire architecture of the risk management across organizations in terms of implementation of risk management policies, changing dynamics of risk culture, whether the three lines of defence model is working properly or not etc. Any gaps in the enterprise risk management will throw challenges in the implementation of each of the risks. It has been found many CROs struggle with improving the risk culture and embedding of three lines of defence model. Strategic Risk Organizations can identify financial, operational, and other risks, however, when it comes to strategic risk, they somehow ignore this risk because of the long-term nature of its impact.  Failure to identify strategic risks is one of the key causes of the downfall of many big brands such as Kodak and Nokia. An integrated risk dashboard can give CROs an early indicator against strategic targets and major developments, which will support them in taking remedial actions Financial Risk Financial risks include uncertainties and untapped opportunities in the effective and efficient utilization of financial resources. In addition, risks in the areas of interest rate, concentration risk, liquidity and funding, capital management, credit default etc. are addressed as part of financial risks CROs will have trend analysis and futuristic view with a better understanding of financial risks, which will help them in better product strategy and asset allocation Operational risk Operational risks are losses arising due to failed people, processes, systems, or events. The operational risks are managed through the maintenance of the risk register which is used for the purpose of Risk Control Self-Assessment. This risk assessment should be plugged into the integrated risk dashboard to flow all key risks and draw attention about the correlation of risks. Some of the risks such as reputational risks are hard to measure, however, it is mostly because of bursting other risks. The CRO must pay attention to the risks that do not impact the external stakeholders such as customers, regulators, government agencies and keep reputational risk under check. External risk External risks include uncertainties and untapped opportunities arising due to changes in the regulation, global economy, disruptive business models, climate change etc. CROs will have a better global view of risk which can be leveraged for risk assessment. Conclusion The article discusses the challenges of CRO that an organization faces in the management of risk due to manual processes, scattered data, or data not available for the purpose of risk assessment. Such lack of data or not one view of risk data can lead to either missing some of the key risks or not drawing meaningful conclusions about risks in conjunction with other risks The article advocates the need to have an automated system that correlates and bring forward the risks for CROs to take immediate action. The velocity of the risk is very crucial because it can beat the human thinking process to act leading to delays and massive losses. The article further assesses that the availability of an integrated risk information system in one place is very important for CRO to assess the risks, draw correlations and take the real-time risk-based decision. To address existing, emerging hidden or correlated risks, one of the most important tools for CRO is to have an integrated risk information system that can provide all risk information at his or her fingertips. Here the CRO can slice and dice the risks for further analysis to understand the existing materiality and correlation of the risks The integrated risk information system should bring together all the risks in the form of the dashboard for CRO’s easy at one place access. This should have trend analysis, analysis of all existing risks, emerging risk along with predictive capability.","excerpt":"The article discusses the importance of risk management information system for Chief Risk Officers to address risks appropriately. Such information system helps in identifying hidden and correlated risks to plan for mitigating action in advance. The article advocates a need for CRO to use advance digital technology to facilitate automated risk information management system including the predictive capability to have early view of risks.","categories":["IT Services"],"tags":[],"author_name":"Sonjai Kumar","publish_date":"2022-01-24T15:55:47","publication_year":"2022","word_count":2382,"keywords":["big data","data science","Go","API","machine learning","TPU","AI","RAG","Aim","R"],"extracted_tech_keywords":["AI","machine learning","data science","Aim","RAG","TPU","R","Go","API","big data"],"url":"https:\/\/analyticsindiamag.com\/it-services\/integrated-risk-information-system-for-chief-risk-officers\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":7669,"title":"Interview &#8211; JT Kostman, Chief Data Officer at Time Inc.","content":"Dr. JT Kostman is a Data Scientist, Mathematician and Psychologist – and the Chief Data Officer for Time Inc. Over the past 20+ years JT has provided data-driven insights into human behavior for organizations ranging from Fortune 500 companies to U.S. Intelligence Agencies. At Time Inc. he leads a rapidly growing global organization of Data Scientists, Data Miners, Strategic Analysts, and Digital Media Marketers who are developing cutting-edge capabilities that are revolutionizing Digital Media Marketing, Data Mining, Predictive Analytics and Consumer Insights. We talk to JT on how a leading publishing house has successfully adopted analytics and how is India being leveraged within this initiative. [dropcap size=”2″]AIM[\/dropcap]Analytics India Magazine: Thanks JT for this interview with us. Tell us more about Analytics at Time Inc.? [dropcap size=”2″]JT[\/dropcap]JT Kostman: Our CEO, Joe Ripp, has been steadfast in his championing of data and analytics at Time Inc. He has made quite clear to our executives, employees, and investors that data and analytics are central to our value proposition – and that data and analytics should be at the core of every decision we make and every action we take. Data is considered to be both one of our most valuable assets and also a catalyst for creating value in all areas of our business. To avoid the frustrations and obstacles that are most common to cohering, integrating, and maximally leveraging data, our CEO likewise made clear to the organization that the CDO’s team owns all the data at Time Inc., including all data generated or provided by our various business units, partners, and vendors. Where in most organizations there is a constant battle for the CDO to be able to gain access to the data analysts need to succeed, at Time Inc. the CEO has resolved that challenge definitively. AIM: Analytics at a leading publishing organization appears like a novel idea. What is the need of analytics at Time? JT: Publishing has historically been a somewhat polarized profession, with an artificial demarcation between journalists on the one hand, and the business folks on the other; a structure we lovingly refer to in publishing as “Church and State”. When Joe Ripp became CEO of Time Inc. he changed that model – and as a result, we are fundamentally changing our industry. The only metrics publishers used to care about was circulation and profit on the one hand, and the number of Pulitzer prizes and similar awards on the other. [quote]By collecting vast amounts of data and predicating our decisions on analyses and not just opinions, we are able to make considerably more sound decisions – and continually shift our sails to take advantage of prevailing winds.[\/quote] AIM: ok, and what are some of the analytics solutions that you work on? JT: Our group is intimately involved in nearly every critical aspect of the business. The areas we support include, but are not limited to: Evolving our AdTech capabilities through sophisticated audience analyses, segmentation models, programmatic capabilities, and the ability to better understand the needs and interests of our audiences; We are presently developing a Data Driven Editorial capability that will help our editors and journalists better understand their audience and evaluate the extent to which our content resonates with various audience segments; Our team has done considerable work in modeling the behaviors and preferences of subscribers, newsstand purchasers, and online audiences to develop audience insights that transcend any that are otherwise available in the market; We have likewise developed metrics to help inform numerous operational aspects of the business for the purpose of increasing profitability and decreasing expenditures. AIM: Can you brief us about a specific use case in analytics that has brought significant value to Time Inc.? JT: We recently developed a rather sophisticated Audience Segmentation Analysis tool that we were able to beta test with a major advertising client. By incorporating some of the client company’s data into our analyses, we were able to develop models that proved extremely valuable in identifying previously unknown (and unexpected) insights into their market. These findings led to an increased advertising commitment to us of over a million dollars (US) – and this proof-of-concept beta test, which took us less than three months to develop, and only a few hours to implement – has already also served as the foundation for other similar, and even more lucrative, opportunities. [pullquote]We intend to fundamentally reinvent AdTech, Publishing and Digital Advertising. There are several truly unique capabilities we have in development that we will be introducing to our clients – and I have challenged our team to publish, present, and patent ideas that help advance our industry[\/pullquote]AIM: If I have to ask you, how are your analytics solutions unique; what would be those key differentiators that you would speak about? JT: I am a Data Scientist, Mathematician and Psychologist by training – and the work I have done has been at the nexus of those fields. The work my group does is similarly highly influenced by this perspective. We consider not just the data patterns, but their implications for understanding behaviors. Similarly, in developing target models we go beyond focusing on simple superficial behaviors like clicks and transactions and take more of a cognitive-behavioral approach that allows us to more effectively describe, understand, predict and influence behaviors. We are in the process of developing patentable solutions that will give us the ability to gain unprecedented insights into our audiences. AIM: Please brief us about the size of your analytics group and what is hierarchal alignment, both depth and breadth. JT: I am relatively new to Time Inc., having joined as their first Chief Data Officer only a few months ago. The team I am building consists of both existent resources and new hires. We have groups that are focused, respectively on: AdTech; DB Management; Data Aggregation and Augmentation; Data Governance; Traditional Analytics\/Modeling; and a Data Science team focused on advanced analytics. The teams are located at several locations in the US (primarily in New York) and in Bangalore. The overall headcount is presently in flux; we are continuing to incorporate additional small teams across Time Inc., but the targeted size will be approximately 150-200 people, which we should reach within the year. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? JT: My primary criteria are pretty simple: I want people who are (very) smart, highly numerate, technologically facile, extremely curious, and easy to work with. I have very little patience for poor attitudes or arrogance. Our team members are already some of the best in the business – and if they can maintain some sense of humility, so should anyone who is new to the team. I firmly believe we should all be able to learn from one another and continue to expand our capabilities. As to the particular skillset, I am pretty flexible. We have as much room for traditional SQL and DBM skills as we do for those who are comfortable with the Apache suite of solutions. While I expect the MapReduce paradigm will continue to predominate – and Hadoop (et al) skills will quickly become required simply to enter the profession, I expect we will also see the need for analysts to increasingly improve their mathematical abilities. I’ve found that even a basic knowledge of logic, set theory, and discrete mathematics can turn an otherwise simply competent SQL analyst into a highly valuable asset. As to analytic tools, I tend to let analysts pick their own, to the greatest extent reasonable. I have an entire group who works exclusively with SAS, though I personally tend to prefer the flexibility of R and Python; these are skills that tend to impress me more than the ability to simply point-and-click through an off-the-shelf solution. Our selection methodology doesn’t presently involve any coding challenges, analytic problems, or psychometric assessments; though I intend to eventually evolve to using more objective evaluation tools. Presently we rely on an assessment of our candidate’s background and a series of Behavioral Event Interviews conducted by peers, supervisors, and internal clients. During the course of these conversations we ask candidates to describe particular experiences they have had working with various datasets and solving real-world challenges. AIM: What according to you are some of the most significant challenges you face within the analytics space? JT: The greatest challenge has been a lack of qualified talent; there simply are not enough truly capable analysts to get the work done. This situation has, unfortunately, been exacerbated by several strategies that were potentially intended to alleviate the problem. In the US and UK we have seen a cottage industry spring up to mass-produce Analysts and Data Scientists. Luring kids in with the promise of big salaries and sexy jobs (a cover story of the Harvard Business Review actually called Data Science the “Sexiest Job of the 21st Century”), they are taking what are often potentially promising people and giving them questionable skills that leave them under-qualified – and wastes both their time and ours as we end up having to sort through increasing piles of resumes to find anyone with real skills. Many of the data solution companies are focusing their efforts on building products that can be operated by “Business Users”; the polite euphemism they use to describe anyone who doesn’t really know what they’re doing analytically. Packaging their solutions with claims like it being a “Data Scientist in a box” and contending that companies will have “no need for expensive analysts” once they purchase their solutions, both misrepresents the importance of hiring experts who are able to meaningfully interpret analytic results – and ultimately leads companies frustrated, believing it is an analytic approach, and not the tools, that are failing them. [quote]This practice is as dangerous as offering X-Ray machines for home use – and can be considerably more costly to companies that are often making multi-million dollar decisions on analyses.[\/quote] AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? JT: The MapReduce paradigm and NoSQL approaches will continue to predominate as traditional RDB thinking continues to drive toward obsolescence. We will increasingly see what we now consider to be “advanced analytics” become fundamental to virtually all medium and large businesses that want to remain competitive. What is now seen as an analytic advantage will increasingly become a matter of survival. As much of an impact as Moore’s law will continue to have on processing power and speed over the next few years, those benefits will increasingly become asymptotic; and I suspect those benefits will continue to be dwarfed by the advances that are simultaneously being made in algorithm development. Solving a complex problem more simply will always trump the ability to race to a less elegant solution. Many of the techniques that are now part of the analyst’s basic toolkit would have been impossible even a few years ago, but for the speed at which the analyses can be conducted – but as fast as processing power continues to increase, the best tool in the analyst’s arsenal will always be that grey squishy wetware they keep between their ears. AIM: What are your thoughts about Analytics in India? Where does it figure in your whole analytics strategy? JT: I have spoken at length in various venues across the US, from Silicon Valley to New York, about the pervasive sense of “Digital Colonialism” that has come to taint the capabilities and relationship the West is becoming increasingly reliant on India to provide. Presumably out of some sense of frustration and desperation at not having sufficient talent in the US and UK, firms have increasingly been outsourcing their analytic needs to India; but they frequently have been reserving the most interesting, challenging, and frankly fun parts of the work for their teams “back home”. They tend to forget that some of the most talented analysts in the world are coming from India. IIT’s was recently ranked third in the world among the best technology universities; their graduates are among the most talented analysts I’ve ever worked with, as are so many of the other Indian analysts I have had the privilege of working with over the past 20+ years. Relegating such remarkable talent to the less interesting parts of the work – and the arrogance of thinking that high-level analytic thinking should be reserved for the West – treats India as an “analytic sweatshop” and only serves to drive away the best and the brightest. While several firms have built their business model on precisely this approach, I believe this sort of shortsightedness inevitably harms the profession, our interests, and the market. [pullquote align=”right”]We intend to establish the Time Inc. India Analytic Team as a training ground for some of the best analysts in the world – while simultaneously establishing Time Inc. as the best place for analysts to work in Bangalore.[\/pullquote]I am convinced that the worldwide future of analytics rests, in large part, on our ability to grow, cultivate, and capitalize on the development of analytic talent in India. At Time Inc. we are changing what has, regrettably, become the dominant model for how many US and UK firms work with their Indian collaborators. We have obliterated the barriers between ourselves and our colleagues in Bangalore. We don’t think of the work we share with them as being outsourced; we think of them as virtual members of our teams, who are simply not co-located. Our team in Bangalore is no different in that sense than our team in Birmingham, Alabama. As a consequence, we are working to ensure our Indian team members share in as much of the most interesting work as anyone in the US and UK. In keeping with that spirit, we have recently also decided to greatly expand our analytic footprint in India – and we will be continuing to invest in the development of analytic talent here; not just for ourselves, but for the sake of the entire industry. AIM: Anything else you wish to add? JT: The future of analytics is still up for grabs. Given the remarkable progress India has made over the past few decades as a digital and data powerhouse, I firmly believe that future can be based in Bangalore and Hyderabad. Whether or not that comes to be will ultimately be decided by the next generation of Indian analysts – and the extent to which they accept the challenge and come to believe India is the future of analytics: Tat tvam asi. Though art that.","excerpt":"Dr. JT Kostman is a Data Scientist, Mathematician and Psychologist – and the Chief Data Officer for Time Inc. Over the past 20+ years JT has provided data-driven insights into human behavior for organizations ranging from Fortune 500 companies to U.S. Intelligence Agencies. At Time Inc. he leads a rapidly growing global organization of Data […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Дарья","publish_date":"2015-07-15T12:10:50","publication_year":"2015","word_count":2448,"keywords":["data science","AI","R","ML","RAG","Python","Ray","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","Ray","RAG","predictive analytics","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-jt-kostman-chief-data-officer-at-time-inc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10098667,"title":"$1M Salary Package: AI Companies Pour Money for GenAI Roles","content":"While everyone is riding the generative AI wave, we might as well encash it. The ones equipped with AI skills seem to be the kings in the current genAI wave. With companies offering exorbitantly high salaries for AI-related roles, it’s the best place to be in. Last month, Netflix was in the limelight for offering a salary of up to $900,000 for the role of a product manager on their machine learning platform team. The news came in at a time when the Hollywood writer’s strike was ongoing. While it caused a lot of hullabaloo, Netflix is not the only one which is ready to handsomely pay for AI-related roles. A number of companies are following suit. Is the AI salary rage warranted? According to Indeed, there are multiple generative AI roles offered by big tech companies including Meta, NVIDIA, Anthropic, Microsoft, Adobe and many others, where the salaries offered go up to as high as half a million dollars. A technical product manager in AI safety, in Anthropic, is being offered salaries of up to $520,000, and a principal engineer AI in HubSpot, gets $427,000. It’s not only tech or AI companies that are offering such excessively high salaries. Consumer and services companies that are implementing AI to transform their products are willing to pay high salaries too. Dating app Hinge was looking to hire a VP for AI to oversee their app’s AI strategy, for a salary of $398,000. The role will entail leading a team of data scientists, and ML engineers to develop AI features. Retail corporation Walmart was also looking to hire a senior manager for its conversational AI platform for a salary of up to $252,000 a year. All In One With generative AI taking centre-stage, the influence on the job market is evident. As per AIM Research, the generative AI job market has witnessed a steady growth from January to June of this year. There are over 4200+ generative AI-related jobs in the US and it has risen by 20% in May. Furthermore, job roles have been modified to suit the current trend. The role of a generative AI engineer that did not exist earlier will now require the competencies of that of a deep learning, ML, NL, and software engineer. Almost like a mandatory need, the multiple roles are now a necessity. The amalgamation of multiple roles has been descriptively placed under ‘qualifications section’ for these open job roles that are offering huge salaries. For instance, the role of ‘Senior Research Scientist-generative AI’ in NVIDIA, that offers a salary of up to $414,000 a year, a candidate should not only possess a thorough knowledge of python\/C++ programming skills, but also an excellent knowledge of theory and practice of deep learning, computervision, natural language processing or computer graphics. The candidate should also be a Ph.D holder in Computer Science\/Engineering, Electrical Engineering, or any related field. Similarly, a ‘Product Technical Program Manager-generative AI’ for Meta, with a salary package of up to $297,000, requires technical and leadership experience. The candidate must have experience developing large-scale ML\/AI platforms such as dataset generation, feature development, model testing and support the development of AI-powered product experiences such as NLP, computer-vision, ranking and personalisation. The Layoffs-Hiring Balance Interestingly, the large layoffs that happened across big tech at the start of the year, seems to have minimal impact with the way things are unfolding now. Scale AI, a data platform for AI that provides training data for ML teams, had laid off 20% of their workforce in January. However, last month, ScaleAI posted a job opening in Indeed for a ‘software engineer- generative AI’ offering up to $215,000 in salary. There are even companies that have laid off employees owing to AI chatbots and efficient processes with generative AI implementation. In May, executive outplacement and career consulting firm Challenger, Gray & Christmas, attributed 4000 job losses to artificial intelligence, making it the first time for the company to mention AI as a cause of job loss. Indian ecomm platform, for merchants, Dukaan recently laid off 90% of their support staff replacing them with their new AI chatbot. Though big tech layoffs have occurred owing to recession or automation, it doesn’t seem to throw cold water over the ambitious hiring process that companies have started. While it looks promising at the moment, it is to be seen how long the generative AI hiring wave will remain. Jobs requiring Generative AI skills grew almost 9X since January, according to Adzuna. Jobs ask for an average of 5+ years of experience with Generative AI tools and frameworks like GPT-4 or PaLM-2.— Vin Vashishta (@v_vashishta) June 8, 2023","excerpt":"Netflix, Meta, NVIDIA and others are generously offering a quarter to 1 million salaries for generative-AI roles. But, why?","categories":["AI Features"],"tags":["Adobe","AI Companies","Anthropic","Dukaan","Generative AI","Meta","Microsoft","ML","netflix","NVIDIA","Walmart"],"author_name":"Vandana Nair","publish_date":"2023-08-17T17:49:22","publication_year":"2023","word_count":774,"keywords":["deep learning","Generative AI","Dukaan","artificial intelligence","Adobe","NLP","Walmart","NVIDIA","Meta","machine learning","AI","ML","generative AI","AI Companies","Anthropic","GenAI","netflix","Aim","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","generative AI","GenAI","Anthropic","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/1m-salary-package-ai-companies-pour-money-for-genai-roles\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10062503,"title":"Happy 10th birthday to Raspberry Pi: A timeline","content":"The tiny pocket-sized computer Raspberry Pi recently completed ten years of its existence, and it has now become more than just a computer. It has been widely used among students, teachers and enthusiasts to push their boundaries of technology, and engineers and IT managers use it for their professional work. On completion of ten years, Eben Upton, the founder of Raspberry Pi, said, “Ten years ago, when Raspberry Pi computers began their sale, about 1,00,000 orders were already placed, and we had out-trended Lady Gaga very briefly. Raspberry Pi was on the road to become a little larger than we had planned.” Raspberry Pi computers are largely used in weather monitoring, desktop PCs, wireless print servers, media usage, game servers, retro gaming machines, robot controllers, stop motion cameras, time-lapse cameras, FM Radio stations and web servers. They are now used prominently in robots, AI-driven cameras, and smart-home devices. The Raspberry Pi Foundation was founded in 2009 to promote basic computer science in schools. A team from Cambridge University’s Computer Laboratory saw a declining interest in computer science and created an easy, low-cost access to a computer. Timelines of the launch of various Raspberry Pi computers Raspberry Pi 1 Model B April 2012 Raspberry Pi 1 Model B, the first model, was launched in 2012 and used a Broadcom BCM2835 SoC, which included a 700 MHz ARM1176JZF-S processor, VideoCore IV graphics processing unit (GPU) and had 512Mb. A lower-cost Model A was also launched; it had less memory and fewer USB ports. Raspberry Pi 2 Model B February 2015 The RPi2 increased computing power significantly using a Broadcom BCM2836 SoC, utilising a quad-core Cortex-A7 900 MHz processor. The computer had also doubled its memory to 1 GB with a shared GPU. The CPU was at least 4-6 times more powerful than the previous computer. Raspberry Pi Zero November 2015 This computer set a new benchmark in low-cost computing, which was priced at just around $5\/£5. The size of the computer was around 65mm x 30mm x 5mm, which could be embedded in robotics and other applications. The CPU of RPi Zero was 40 per cent faster than RPi 1. Raspberry Pi 3 Model B February 2016 The RPi3 used a 64-bit compatible SoC and bumped up the processing power to 4×Cortex-A53 1.2 GHz. It offered 802.11n wireless and Bluetooth 4.1, giving it an even greater appeal to consumers. These computers could be mounted in remote locations and still access data that the computer could record. Raspberry Pi 1 Model B+ March 2018 The upgraded B+ increased the ethernet speed up to 300 Mbit\/s, introduced 802.11ac dual-band 2.4\/5 GHz wireless and Bluetooth 4.2 LS BLE, and a small bump in the CPU speed to 1.4 GHz. Raspberry Pi 4 Model B June 2019 This model has a 1.5 GHz 64-bit quad-core ARM Cortex-A72 processor, on-board 802.11ac Wi-Fi, Bluetooth 5, full gigabit ethernet, two USB 2.0 ports, two USB 3.0 ports, 1–8 GB of RAM, and a dual-monitor, which was supported via a pair of micro HDMI (HDMI Type D) ports for up to 4K resolution. When used with an appropriate PSU, the Pi 4 can also be powered via a USB-C port, enabling additional power to downstream peripherals. This Pi could be operated only with 5 volts and not 9 or 12 volts like other minicomputers of this class. Raspberry Pi 400 Kit November 2020 This model featured a custom board derived from the existing Raspberry Pi 4, and it was remodelled with a keyboard attached. A robust cooling solution, the broad metal plate and an upgraded switched-mode power supply allowed the Raspberry Pi 400’s Broadcom BCM2711C0 processor to be clocked at 1.8 GHz. The keyboard-computer features in this model had 4 GB of LPDDR4 RAM. Raspberry Pi Zero Raspberry Pi Zero was launched in November 2015 with a smaller size and reduced input\/output and general-purpose input\/output capabilities. Raspberry Pi Zero W was launched in February 2017; it was a version of Zero with Wi-Fi and Bluetooth capabilities. Raspberry Pi Zero WH was launched in January 2018; it was a version of Zero W with pre-soldered GPIO headers. Raspberry Pi Zero 2 W was launched in October 2021; it was a version of Zero W with SiP designed by Raspberry Pi and based on Raspberry Pi 3. Pi 2W has 64-bit capabilities. Raspberry Pi Pico January 2021 The Pico has 264 KB of RAM and 2 MB of flash memory. This model is programmable in MicroPython, CircuitPython, C and Rust. The Pico partnered with Vilros, Adafruit, Pimoroni, Arduino and SparkFun to develop accessories for Raspberry Pi Pico and boards using RP2040 Silicon Platform. Pico was mainly designed for physical computing. Raspberry Pi models are affordable and flexible. The uncountable projects that one can do using these tiny computers, make them relevant even today.","excerpt":"The name Raspberry Pi is derived from the fruit pie, raspberry pie. Many companies in the computer neighbourhood where Raspberry Pi was based used fruit names such as Apple and apricot as names for their companies and products.","categories":["AI Features"],"tags":["Raspberry Pi"],"author_name":"Poornima Nataraj","publish_date":"2022-03-11T12:00:00","publication_year":"2022","word_count":801,"keywords":["Go","TPU","programming_languages:R","AI","programming_languages:Go","Python","programming_languages:Python","Rust","GAN","Raspberry Pi","R"],"extracted_tech_keywords":["AI","TPU","Python","R","Go","Rust","GAN","programming_languages:Python","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/happy-10th-birthday-to-raspberry-pi-a-timeline\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10073079,"title":"Julia 1.8 released, to enhance Apple Silicon","content":"The release of Julia version 1.8 is here. Following three betas and four release candidates, this version was finally released and is set to provide improved support to Apple Silicon. The new language version has improvements in various areas like Mutable struct fields, annotation and compiler\/runtime. The release also has changes in the build system including new library functions and features along with tooling improvements. What’s new Profiling: The new allocation profiler captures individual heap allocations with the type, size, and stack allowing easier visualisation with PProf.jl as well as with the Julia extension for VS code. Additionally, updates to CPU profiling include thread and task profiling, and running code profiling. Packages: A new tool has been added to provide compilation estimate timing and highlighting percentage of compile time for re-compiling invalidated methods. An update to the Pkg status updater will show small compatibility constraints and also indicate status of the latest version packages. Improved precompilation: The 1.8 version of Julia addresses the issue of automatic discarding of compiled code by saving all type-inferred code in cache. Additionally, to reduce compilation time of predictable type workloads, users can now eliminate type-inference as a source of latency. Support for Apple Silicon Julia 1.7 had an experimental preview for native builds on Apple Silicon which supported basic usage but consisted of frequent segmentation faults. In version 1.8, Apple will become a tier 2 supported platform that is now covered by Continuous Integration (CI) on dedicated Apple Silicon machines. Julia was launched in 2012, when engineers were using languages like Python, MATLAB, and Ruby. In 2022, Julia is a standard language used by NASA for spacecraft modelling. It is also used by NVIDIA, Google, Intel, Amazon, Microsoft and other tech companies.","excerpt":"The new language version has improvements in various areas like Mutable struct fields, annotation and compiler\/runtime.","categories":["AI News"],"tags":["Apple","Julia","NASA","programming","Python"],"author_name":"Mohit Pandey","publish_date":"2022-08-19T12:40:34","publication_year":"2022","word_count":288,"keywords":["Go","NASA","programming_languages:R","AI","Apple","programming","programming_languages:Go","Python","Julia","programming_languages:Python","R"],"extracted_tech_keywords":["AI","Python","R","Go","Julia","programming_languages:Python","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/julia-1-8-released-to-enhance-apple-silicon\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":26776,"title":"Acrobatic Movements Are Now Possible To Be Visualised Accurately Through Reinforcement Learning","content":"The pace in research around Reinforcement Learning (RL) has been growing seriously in the recent years. It’s no longer restricted to just the classic problem of robots getting punished or rewarded for actions and then rectifying them. It has moved beyond this context now. Although the robot problem formed the basis for many self-learning applications, RL has taken on a new level. It has been used in virtual environments as well as in gaming for being formidable virtual opponents for gamers or players. Now, RL may even emulate human movements along with their natural behaviour as well. We will discuss one particular research study called DeepMimic by academics at University of California, Berkeley, who have managed to simulate acrobatic movements precisely using RL methods. Computer Graphics For Accurate Movement Visualisation Xue Bin Peng, author of the paper for DeepMimic says that the inspiration for the acrobatic movement simulation project through RL, came from computer graphics which offer precise visualisation of real-world physics and the ability to model them. The possibility presented through animations can aid studies that have explored analysing human body movements for simulation. But, the challenge lies in modelling physics-based models for simulation, across other applications. While there are many studies that have created models, there are setbacks with respect to optimisation or dynamics in movements or motion. Recent developments in motion simulation focus on bringing online models for easier representation. However, they fall short when it comes to richer dynamics and implementing more motions into the model. This was what drove Peng and the team to provide a single simulation model capable of incorporating a large number of motions including acrobatics. The model also encapsulates RL policies efficiently. It will mean that RL and physics animations go hand in hand, which is a huge improvement. DeepMimic – A Powerful Model For Motion Imitation As mentioned earlier, building physics-based models which incorporate a lot of movement and actions, is quite challenging. Even if they are created with all the considerations, they may realistically fail to achieve significant results for ML. But with DeepMimic, this is not the case. Along with capturing unnatural acrobatic movements, this RL model considers a data-driven approach for these movements, says Peng. “An alternative is to take a data-driven approach, where reference motion capture of humans provides examples of natural motions. The character can then be trained to produce more natural behaviours by imitating the reference motions. Imitating motion data in simulation has a long history in computer animation and has seen some recent demonstrations with deep RL. While the results do appear more natural, they are still far from being able to faithfully reproduce a wide variety of motions.” [su_youtube url=”https:\/\/www.youtube.com\/watch?v=vppFvq2quQ0″ width=”280″ height=”200″] In the context of standard RL, the policies are trained for each acrobatic movement through a motion imitation task. These movements are represented in the form of ‘target poses’, which are necessary for individual timestep actions. This makes it possible to project complex acrobatic movements smoothly in the model. Characters And Tasks DeepMimic has four character visualisations: 3D humanoid Atlas robot model T-Rex Dragon These are generated as rigid bodies with kinematic links of three degrees of freedom (DOF) except for the knees and elbows having one DOF. In addition, physical characteristics such as mass and height are also mentioned. All of these arrangements form the structure of bodily acrobatic movements. RL policies are trained for these character objects. Apart from the characters, many tasks are delineated and assigned to these rigid body objects. The tasks classified by DeepMimic’s researchers are given below. Target Heading Strike Throw Terrain traversal Based on these tasks’ categories, a total of 30 skills are designed for the simulation, and are trained in RL. Also, some of the skills are integrated to perform multi-skill actions — for example, movements like running, jumping and flipping motions are clubbed to get a unique action movement. Training For Simulation After finalising the adequate parameters such as policy states, actions rewards and the neural network to map all of these features, the model was subjected to training. Once policy and value functions are calculated and trained, the training process starts sequentially for each instance of the state of reference movements in a batch-wise fashion. For imitating desired acrobatic motions, the policies in RL should capture every phase of the motion incrementally over time. This is done through initial state distribution, which helps the RL agent to capture the exact beginning of motion precisely. Similarly for cyclic motions such as backflips, frontflips etc., another strategy called ‘early termination’ is used. (Details of RL as well as training strategies can be found here). Conclusion The results after training show motions emulated accurately through RL. In addition, multiple movement integration also fares very well in visualisation. The physics aspects of DeepMimic model is where it makes a mark, thus presenting the possibilities of emulating a variety of movements. With more and more eccentric movements captured, RL can vastly improve self-learning areas that use motions and movements.","excerpt":"The pace in research around Reinforcement Learning (RL) has been growing seriously in the recent years. It’s no longer restricted to just the classic problem of robots getting punished or rewarded for actions and then rectifying them. It has moved beyond this context now. Although the robot problem formed the basis for many self-learning applications, […]","categories":["AI Features"],"tags":["policy","simulation","skills"],"author_name":"Abhishek Sharma","publish_date":"2018-07-31T04:49:57","publication_year":"2018","word_count":834,"keywords":["Go","simulation","programming_languages:R","AI","neural network","data-driven","ML","programming_languages:Go","RAG","skills","R","policy"],"extracted_tech_keywords":["AI","ML","neural network","RAG","R","Go","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/acrobatic-movements-are-now-possible-to-be-visualised-accurately-through-reinforcement-learning\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10172179,"title":"QNu Labs Launches Academy to Support India’s Quantum Talent","content":"Bengaluru-based quantum security startup, QNu Labs, has launched QNu Academy, a global educational platform for developing skilled professionals in quantum technologies and cybersecurity. The initiative, aligned with the Indian government’s National Quantum Mission (NQM), seeks to equip students, faculty, and institutions with practical knowledge in quantum key distribution, quantum random number generation, and post-quantum cryptography. The academy will offer hands-on lab work, real-world use cases, and industry-validated programmes designed in collaboration with Indian institutes like the IITs and DRDO, along with international research bodies. The programme also includes certification, placement support, and faculty development modules, emphasising building centres of excellence (CoE) in educational institutions to foster research and innovation. Sunil Gupta, co-founder and CEO of QNu Labs, said, “QNu Academy is more than an educational platform. It is a national mission to democratise access to quantum education and build widespread awareness around quantum communications.” The startup aims to create a sustainable ecosystem for quantum learning in India through such industry-relevant programmes, CoE labs, certified programmes, real-time projects, and assignments with placement opportunities. The academy will cater to universities, faculty members, and students, promoting skill development through instructor-led and self-paced learning. It is positioned as a long-term investment in India’s digital resilience and leadership in quantum innovation. QNu Labs targets to close the talent gap in the emerging field of quantum technologies. “The future of cybersecurity in India depends on how well we prepare today’s learners to tackle tomorrow’s threats,” Gupta added. Earlier in May, the company completed its series A funding round, raising ₹60 crore. The round was led by the NQM and included investments from Lucky Investment, Speciale Invest, Tenacity Ventures, and Singularity AMC. With this new capital, QNu had planned to expand its platform, QShield, which was recently launched on World Quantum Day and is purported to be the world’s first unique quantum security platform.","excerpt":"The programme includes certification, placement support, faculty development modules and emphasises building CoEs.","categories":["AI News"],"tags":["QNu Labs","quantum","quantum talent","quantum technology"],"author_name":"Sanjana Gupta","publish_date":"2025-06-23T13:16:20","publication_year":"2025","word_count":307,"keywords":["Go","QNu Labs","API","funding","startup","programming_languages:R","AI","innovation","Git","quantum","Aim","quantum talent","R","quantum technology"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","API","innovation","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/qnu-labs-launches-academy-to-support-indias-quantum-talent\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61009,"title":"This Is The Perfect Time For Zomato, Swiggy and Dunzo To Take Their Drones Out In India","content":"This is a great opportunity which lies in these unmanned drones to combat the coronavirus pandemic in India. As the COVID-19 pandemic worsens despite the lockdown, it has become critical to observe and reinforce social distancing norms even more. With some people not complying with rules despite the urgency of the situation, are there additional measures that can be taken to ensure they stay at home?  Perhaps unmanned drones can serve as an effective solution. This is a great opportunity which lies in these unmanned drones to combat the coronavirus pandemic and maintain the health and safety of the Indian population. What is more, with the latest developments around COVID-19 coming to the fore, the utility of drones across several areas is being recognized by the government. The good news is that the government has been expediting regulatory processes when it comes to operating drones for long-range commercial operations. The Indian government’s Digital Sky platform has registered all companies in the drone space, along with voluntary registration of all drone operators in the country. This coincides with the drone POCs, taking it a step further to make drones ready for commercial use. Last year, DCGA approved seven companies to conduct POCs for beyond visual line of sight (BVLOS) drone flights. The selected seven companies included Dunzo, Swiggy, Zomato, Zipline, Redwing, Tata Advanced Systems and Honeywell for long-range experimentation in low altitude over scarcely populated places in India for a few months. With the objective of boosting the capabilities of the drone industry in this time of crisis, the Indian government has been finalizing the statutes; with the directorate general of civil aviation (DGCA) recently approving hyperlocal delivery company Dunzo to finish testing their long-range drone logistics solutions. Dunzo will begin testing beyond-visual-line-of-sight (BVLOS) drone deliveries by the end of April, as per media reports. This is in contrast with the stance held by the government earlier, where drones were seen as a security threat. Post India’s directorate general of civil aviation (DCGA) began regulations on drones in 2019, and this was seen as a significant milestone. But the policy on drones got delayed until May 2019, when the government finally invited invitations on proof of concepts around drone technology for commercial operations like logistics or medical deliveries. The fact that DCGA shortlisted India’s three major delivery companies – Zomato, Swiggy and Dunzo – implies that the government is seeing the potential of drones in logistics and delivery services, especially during the ongoing lockdown. Delivering Essential Packages Via Drones In India If the COVID-19 situation worsens, drones may very well be used to deliver essential packages across the nation, and particularly in difficult terrains and viral hotspots. Also, Redwing and Zipline have been cooperating with state governments to transport drugs and medical items to distant places. Other use cases for drones include aerial survey, traffic monitoring, security and surveillance, as well as crowd management. For purposes such as atmospheric monitoring, crisis response, or transport of medicines, autonomous drones can be marshalled to help those in need, without risking human agents. Overview As the government works on getting all drone manufacturers and users listed on the Digital Sky platform, this will also open up the drone economy for e-commerce and pharmaceutical companies to deliver quick orders soon. Under this, registered drones with approved specifications can apply for a permit to fly on Indian government’s Digital Sky platform.","excerpt":"This is a great opportunity which lies in these unmanned drones to combat the coronavirus pandemic in India. As the COVID-19 pandemic worsens despite the lockdown, it has become critical to observe and reinforce social distancing norms even more. With some people not complying with rules despite the urgency of the situation, are there additional […]","categories":["AI Features"],"tags":["drones","drones India","dunzo","swiggy"],"author_name":"Vishal Chawla","publish_date":"2020-04-06T18:00:00","publication_year":"2020","word_count":565,"keywords":["Go","drones","AI","dunzo","programming_languages:R","drones India","Git","programming_languages:Go","ViT","GAN","swiggy","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-this-the-perfect-time-for-zomato-swiggy-and-dunzo-to-take-their-drones-out-in-india\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070837,"title":"TCS Partners with Moveworks to transform the Service Desk with AI","content":"With the shift to ‘work from anywhere,’ employees need immediate support to remain productive—whether they have an IT issue, an HR request, or a query about the company’s expense policy. Conventional service desks can no longer achieve the speed or the scale required to deliver around-the-clock help in real-time to a hybrid workforce. Typically, service desks take an average of three working days to resolve employees’ issues, resulting in reduced productivity and engagement. To resolve such issues, Tata Consultancy Services (TCS), one of the world’s most valuable, largest IT services brands, decided to provide any support that their employees may need in a matter of seconds. How did they achieve that? TCS announced its partnership with Moveworks, an AI platform that provides seamless, multilingual support to employees at any time simply by querying their needs. “Progressive enterprises are adopting artificial intelligence and other digital technologies to enhance the overall employee experience”, says K Ananth Krishnan, Chief Technology Officer at TCS. “This partnership with Moveworks will enable our customers to transform their service desks with AI. They can rely on our expertise and Moveworks’ technology to support their workforce automatically.” The Moveworks promise Founded in 2016, Moveworks is an AI-based platform that provides quick, automatic support to employees’ requests by querying their needs and enables better management. “Our platform solves employees’ issues from end-to-end—without any manual intervention—using a very large network of integrated machine learning models”Manoj Gupta, Vice-President of Partnerships at Moveworks The platform leverages natural language understanding (NLU) to understand each request, probabilistic machine learning to determine the best solution for the company and deep integrations with other SaaS applications to resolve the request autonomously. The platform doesn’t simply interpret these requests or offer visibility into what needs attention—it handles the whole process, solving employees’ issues from end-to-end. “TCS sees that Moveworks is the industry leader, with a comprehensive set of features, a rich roadmap, and a strong track record of improving the employee experience. TCS has seen first-hand the impressive results that their customers have experienced with Moveworks. In partnering with Moveworks, TCS saw the opportunity to extend that success to their thousands of enterprise customers around the world by helping them leverage AI to automate critical IT processes”, adds Manoj Gupta. Betting on the best partnerships “The nature of work has fundamentally changed, which means businesses need to completely rethink the way they support their employees”, says Bhavin Shah, CEO of Moveworks. “Our partnership with TCS is about empowering businesses to meet their employees’ expectations—by providing a single place they can go for instant support. With the help of TCS, this experience will be the gold standard for businesses around the world.” Mutual clients of Moveworks and TCS, who have greatly reduced their service desk costs as well as their mean time to resolution (MTTR), include top automakers, healthcare providers and technology vendors. TCS’ deep domain knowledge and technical expertise combined with Moveworks’ multilingual AI platform allows their customers to provide the kind of workplace experience employees expect as well as detailed insights to measure the quality of service. “To create a world-class employee experience, companies need both their business knowledge and sophisticated technology that can address the nuances of their unique digital business. TCS leverages its deep contextual knowledge and digital champions to overcome these specific challenges while Moveworks’ AI autonomously learns the intricacies of each customer environment”, says Anupam Singhal, Business Head, Financial Services, TCS. “Together, TCS and Moveworks empower service desks to make supporting employees effortless.” Global insurance broker AssuredPartners relies on TCS and Moveworks for employee support needs. “Both TCS and Moveworks have been instrumental in helping us transform our overall employee experience. The new digital solution allows our employees to get the technical support they need in a matter of minutes—not days—so they can focus on things that really matter”Sankha Ghosh, CIO at AssuredPartners Road ahead The partnership with TCS is a crucial part of Moveworks’ growth story. Moveworks has aggressive growth targets, and their partnership with TCS will accelerate the new logo acquisition—especially in the large enterprise segment of the market. TCS also has a substantial global presence in the business world that aligns with Moveworks’ plans for geographic expansion. “Ultimately, we know that investments in AI will continue to grow as businesses look to increase efficiency and simplify the overall employee experience. Our industry-leading technology combined with TCS’ deep experience in service delivery is the right combination to develop integrated offerings that will delight customers. This is really just the start”, says Manoj Gupta.","excerpt":"TCS leverages its deep contextual knowledge and digital champions to overcome these specific challenges while Moveworks’ AI autonomously learns the intricacies of each customer environment.","categories":["IT Services"],"tags":["Tata Consultancy Services","TCS"],"author_name":"Sri Krishna","publish_date":"2022-07-12T17:30:00","publication_year":"2022","word_count":753,"keywords":["Tata Consultancy Services","Go","artificial intelligence","machine learning","programming_languages:R","AI","R","ML","Git","RAG","ViT","TCS"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","R","Go","Git","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/tcs-partners-with-moveworks-to-transform-the-service-desk-with-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":1048,"title":"Interview – Raj Mohan KK, CEO at Satvik","content":"Datamatics Global Services entered into an alliance with Satvik to offer a combination of high-end and customized solutions in Business Analytics. Read the full story here. Raj of Satvik talks with Analytics India Magazine on more about this alliance. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: How would you describe the objective and vision of the partnership with Datamatics? [dropcap style=”1″ size=”2″]RMK[\/dropcap]Raj Mohan KK: The main objective of the partnership is to offer customized full range of Consulting, Data Management and Business analytics solutions to the global market, hereby expanding the scope of Analytics offerings for both Satvik & Datamatics. Our vision is to be customer centric and continuously leverage innovations in providing business solutions using analytics and technology to our global clients, by not just transforming the enterprise management experience but also effectively harnessing the power of Data. Our Data Management & Business Analytics solutions help organizations predict events, assess risk and guide faster decision-making. This enables organizations to do business in a much smarter way than they are used to be. AIM: What are the specific areas of delivery, both from Satvik and Datamatics, which this partnership seeks to cover? RMK: The core of Datamatics – Satvik partnership is to deliver world-class analytical solutions to global organizations leveraging on the strength of each entity. Datamatics is global organization with a high level of reputation of delivering Business and Information technology solutions. Satvik’s core competency is in delivering high quality consulting, data management and analytical solutions. Datamatics is working with a large pool of global organizations and has a wide reach in marketing and sales across the globe.  Thus this partnership is leveraging this market reach of Datmatics with the high quality global delivery capability of Satvik. AIM: What does Satvik typically look for in a partner before forging an alliance? RMK: We, in Satvik, look at the Cultural compatibility, Management depth, Ethics and the complimentary capabilities while choosing a partner. Satvik promotes a commitment based work culture where high quality deliverables are sacrosanct. We chose organizations that can appreciate and complement this core competency of us for a long-term mutually beneficial partnership. AIM: What can we expect from this Partnership over the next year? RMK: In the year one, we are looking for a wide client penetration mainly focusing in delivering few large deals. Analytics, as an offering is going through very dynamic changes.  Organizations are no longer interested in just technology or mathematical model building capability. Organizations are looking for Smart business solutions, which can increase their top line and reduce the costs. With our rich experience and thrust in Consulting, we are providing Smart business solutions to our clients by harnessing the power of data. So we will see execution of large-scale enterprise wide analytical implementations in the near future. AIM: Besides, this partnership, what else can we expect to hear in near future? RMK: We are one of the few new age Analytics providers with high level of thrust on Business consulting and Technology. Though we have a high level competency in building complex mathematical models, we have developed a very differentiated process through with we very efficiently design Business solutions and deliver them through highly quality analytical and technology solutions. With this differentiated and innovative approach, we will be executing large enterprise wide analytical solutions in near future. [divider top=”1″] [spoiler title=”Biography of Raj Mohan KK” open=”0″ style=”2″] Raj serves as the Founder and CEO of Satvik Labs Pvt Ltd since July 2007. Raj has more than 20 years of experience in the fields of Consulting & Strategic Management in IT, ITES and Engineering Industries. Prior to joining Satvik, Raj served as the Partner – Director with A&B Consulting. Before becoming a partner with A&B Consulting, Raj held various senior Management Positions in Hyder Consulting, UK and Anderson Consulting. Raj has been a Thought leader in the use of Data Mining for implementing complex Business Transformational programmes. He has been a speaker in many international forums in the areas of Mathematical Model building, Business Analytics, Data Mining, Big Data technologies etc. Raj holds a Bachelors Degree in Engineering and Masters Degree in Management and Technology from IIT Delhi and University of Sunderland, and Advance Management Programme from Kellogg School of Management.[\/spoiler]","excerpt":"Datamatics Global Services entered into an alliance with Satvik to offer a combination of high-end and customized solutions in Business Analytics. Read the full story here. Raj of Satvik talks with Analytics India Magazine on more about this alliance.   [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: How would you describe the objective and vision of the […]","categories":["AI Features"],"tags":["Interviews and Discussions","masters in data analytics in india"],"author_name":"Дарья","publish_date":"2012-09-10T09:48:29","publication_year":"2012","word_count":705,"keywords":["big data","Go","AI","innovation","RAG","masters in data analytics in india","Aim","analytics","Rust","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","Rust","big data","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-raj-mohan-kk-ceo-at-satvik\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10130658,"title":"NVIDIA’s GenAI Models for OpenUSD to Advance Robotics and Digital Twins","content":"NVIDIA has unveiled significant advancements to Universal Scene Description (OpenUSD), aimed at expanding its adoption across various sectors, including robotics, industrial design, and engineering. These developments, announced at the ongoing SIGGRAPH event in Denver, are set to enhance the capabilities of developers in creating highly accurate virtual worlds, crucial for the next evolution of AI technologies. The new offerings include NVIDIA NIM Microservices, which allow AI models to generate OpenUSD language for various applications such as answering user queries, generating OpenUSD Python code, and understanding 3D space and physics. These services are designed to accelerate the development of digital twins and other virtual environments, providing a more efficient and scalable solution for industries. Digital World with Generative AI NVIDIA’s latest generative AI models, available as NIM microservices, are the first of their kind for OpenUSD development. These models facilitate the incorporation of AI copilots and agents into USD workflows, broadening the scope of 3D world-building in sectors like manufacturing, automotive, and robotics. The microservices include USD Code NIM, USD Search NIM, and USD Validate NIM, each designed to streamline the creation and validation of 3D content. NVIDIA has also announced upcoming NIM microservices, such as USD Layout NIM and USD SmartMaterial NIM, which will further enhance the capabilities of developers working with OpenUSD. These tools are expected to play a pivotal role in the development of next-generation AI applications, particularly in the realm of physical AI and robotics. OpenUSD Reach with New Connectors In addition to microservices, NVIDIA introduced a series of USD connectors aimed at bringing generative AI to more industries. Collaborating with Siemens, NVIDIA is integrating OpenUSD pipelines with Siemens’ Simcenter portfolio, enabling high-fidelity visualisation of complex simulation data. This partnership aims to facilitate better decision-making and collaboration among stakeholders. Moreover, NVIDIA released a connector from the Unified Robotics Description Format to OpenUSD, allowing seamless integration of robot data across various applications. To support the growing OpenUSD ecosystem, NVIDIA announced the OpenUSD Exchange SDK, which helps developers create robust data connectors. These advancements underscore NVIDIA’s commitment to revolutionising the creation and interaction with 3D content, paving the way for broader adoption and innovation across multiple industries.","excerpt":"Jensen Huang’s vision of ‘Physical AI’ is getting real through NIM and visual AI agents.","categories":["AI News"],"tags":["NVIDIA","Robotics","simulation"],"author_name":"Vandana Nair","publish_date":"2024-07-30T03:00:00","publication_year":"2024","word_count":358,"keywords":["Go","simulation","AI","ML","Scala","Robotics","microservices","Python","Aim","generative AI","copilots","NVIDIA","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","copilots","microservices","Python","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidias-genai-models-for-openusd-to-advance-robotics-and-digital-twins\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10024083,"title":"AI &#038; Analytics Firm Tredence Launches MLOps Platform","content":"AI and analytics services company Tredence has launched a state-of-the art machine learning platform, ML Works, targeting data scientists, engineers, and analysts. The enterprise version of the platform will be launched in September. For now, the platform is free. ML Works offers automated workflows, pre-built solutions for tracking model degradation, and code workflow management. The platform will help companies navigate ML operations challenges. It will allow data scientists to shift focus from management and risk mitigation aspects of machine learning to driving innovations in AI. ML Works can scale machine learning models, simplify monitoring, and reduce outages. It also detects anomalies in ML models in the early stages and thereby reduces downtime. ML Works includes data drift analysis capabilities for monitoring product model accuracy, explainable AI for non-technical users, and custom metrics. “AI only delivers value if machine learning models are deployed in production, which is why we created ML Works to help enterprises make the most of their machine learning investments. Breakaway companies across the globe are focused on driving AI-led innovation and faster value realisation, and ML Works will help enterprises of all sizes make that leap,” said Soumendra Mohanty, COO, and Chief Innovation Officer, Tredence. Tredence CTO Sumit Mehra said: “This cutting-edge platform will accelerate the machine learning lifecycle, which will be a game-changer in the industry.”","excerpt":"AI and analytics services company Tredence has launched a state-of-the art machine learning platform, ML Works.","categories":["AI News"],"tags":["MLOps","risk management","Tredence"],"author_name":"Shraddha Goled","publish_date":"2021-04-15T19:57:01","publication_year":"2021","word_count":220,"keywords":["risk management","machine learning","programming_languages:R","AI","innovation","ML","MLOps","explainable AI","analytics","Tredence","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","R","explainable AI","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-analytics-firm-tredence-launches-mlops-platform\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":66261,"title":"Why Data Literacy Is Not Just A Math Skill But A Life Skill","content":"Communicating data is an essential skill and it is not just about doing complex coding. Data literacy is about deriving value from data and Dr Kirk Borne who is a data scientist and an astrophysicist and also a leading AI influencer spoke at plugin 2020 about the importance of being data literate for the future of work. He addressed how it is important for both individuals and to organizations to be data literate. He addressed five aspects of it — data awareness  (what is it?), data relevance (why me?), data literacy (show me how), data science (where’s the science?), and the data imperative (create and do something with data). He also discussed why it is important for data scientists to lead the efforts to build data literacy in society, in schools, and in professional development activities for organizations. Data Awareness As Borne shares, Data Literacy is a way of thinking about numbers and measurement of things in a way that gives meaning. It is interesting how we learn about the world using data. Essentially what we see and the information that we collect is data. “We gather data and information, from which we derive knowledge and wisdom from which we decide what actions to take,” he shared. It is about connecting dots and seeing how it makes sense. Data exists in a lot of different ways such as databases, data tables, images, graphs, documents, social networks, phone apps, web clicks, time series, speech, audio and more. And the biggest challenge about it is not volume but the complexity of it. Organisations in various industries also collect data on many different and complex sources of data, especially financial organisations. These diverse datasets are often stored in separate silos such as customer data, marketing data etc. This inhibits data teams from integrating multiple datasets that when combined could yield deep, actionable insights to create business value. “What we have is a lot of information but very few insights. Data literate is a person who knows data, how to integrate it and combine it to create value” said Borne. Data Literacy includes the ability to read, work with, analyse and argue with data. Data Relevance “While collecting data is one aspect of it, the next big question is how relevant is it for me. A lot of data is generated in a minute which comes from everywhere and each person creates data. Data relevance is not just about creating value of data and not just showing complex data,” he says. Data relevance is especially important in today’s era where digital transformation is the most essential technology. Various technologies such as AI, robotics, machine learning, deep learning depend on lots of data to make value for organizations. “Especially in the COVID times, the digital frontier is moving rapidly and has accelerated development in space,” says Borne. Data Literacy Data Literacy is a way of thinking about numbers and measuring insights from it. For instance, if there are different formats of data, a data-literate person should be able to normalise these data and find a similarity in all these numbers to generate insight. Data literacy is understanding that something can be done with every data. “For every data that was initially thought to be useless, it is possible to normalise it and make sense of it. Data literacy also includes business understanding, data exploration, data representation and more. Another important component of data literacy is data storytelling. Connecting with data and communicating it in a way that is understood by everyone is the key aspect of it. Communicating stories and insights and data is a powerful tool to convey stories. Data Science is a crucial part of bringing about data literacy. Data is the fuel whereas science is a process by which patterns are extracted and actions are taken. And when we talk about data science, it is not just about having math skills or engineering skills but understanding to create value out of it. It is essential to have data and data literacy in place as poor data literacy may cripple the business. “Data is everywhere and data science is crucial to understand data and drive value out of it. Data science is not just a strategy but corporate thinking,” concludes Borne.","excerpt":"Communicating data is an essential skill and it is not just about doing complex coding. Data literacy is about deriving value from data and Dr Kirk Borne who is a data scientist and an astrophysicist and also a leading AI influencer spoke at plugin 2020 about the importance of being data literate for the future […]","categories":["AI Features"],"tags":["big data machine learning","data literacy","social network big data"],"author_name":"Srishti Deoras","publish_date":"2020-05-30T21:40:15","publication_year":"2020","word_count":708,"keywords":["data science","API","machine learning","AI","digital transformation","Git","data literacy","deep learning","ViT","big data machine learning","GAN","R","social network big data"],"extracted_tech_keywords":["AI","machine learning","deep learning","data science","R","Git","API","GAN","ViT","digital transformation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-data-literacy-is-not-just-a-math-skill-but-a-life-skill\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10012254,"title":"Five Most Controversial Moments Of AI In 2020","content":"Artificial Intelligence has to be one of the most impactful technologies that the world has seen in recent years. It is no longer just limited to the quaint research and development labs of academies and bigger institutions but has successfully penetrated the normal and day-to-day functioning of the society. Like any other technology, AI also comes with its set of challenges. However, the stakes are slightly higher, considering the impact AI-technology-gone-rogue can have. Below we list some of the most controversial moments of the AI industry in 2020. If not anything, this may be considered as a cautionary alarm moving forward. Clearview AI & Facial Recognition Technology ‘The Secretive Company That Might End Privacy As We Know It’, this was the title of The New York Times article published in January 2020. This article was about Clearview AI, a startup founded by Australian entrepreneur Hoan Ton-That. A facial recognition software in its most primitive form is already debated, but there is an added problem with Clearview AI due to the unscrupulous way that it operates. It reportedly collected and stored billions of personal images of people around the world from the web and social media accounts into their database without any proper permission system. These pictures in the database were then compared with photos of unknown individuals using facial recognition technology. Clearview’s data and technology were accessible by at least 26 countries and their law enforcement agencies, even though the company strictly maintained that it had provided access to just the concerned authorities of the USA and Canada. In India too, it was reported that the Gujarat police services too used Clearview’s facial recognition technology. Further, despite official denial from the company itself, a ‘leaked’ report revealed that apart from 2,200 law enforcement organisations from the world, many private organisations too had adopted this technology. As the controversy gained ground, Facebook, Twitter, Google, and YouTube sent cease-and-desist letters to the company after these tech giants released that it had been scraping off images from these platforms. Currently, several countries, including Canada, have either completely removed this technology or are actively probing the matter. GPT-3 In June this year, researchers from OpenAI released the GPT-3, a state-0f-the-art language model. While its predecessor GPT-2 had 1.5 billion parameters and was considered the largest models at the time of its release, GPT-3 surpassed it by miles by having 175 billion parameters. To gain a perspective of how large the model is, it would be useful to consider that it is ten times larger than Turing NLG that ranks behind it. Upon its introduction, several sections called it the revolutionalisation of the concept of machines writing codes like humans and taking it a step further by writing blogs, stories, websites, and apps. One of the best examples would be the case of a college student who created an entire blog using GPT-3. Another case involved a GPT-3 powered bot that was caught interacting with people in the comment section of Reddit threads. The Guardian also wrote an entire article using this language model. GPT-3, with its range of capabilities, was not immune to criticism. With its ability to produce very human-like texts, there were obvious concerns raised against its misuses such as spam and phishing, fraudulent academic essay writing, social engineering pretexting, and abusing legal processes, among others. The criticism further gained ground when OpenAI decided to give its exclusive access to Microsoft. This step was severely criticised even by big wigs such as Elon Musk, one of OpenAI’s founders himself. There are other shortcomings of GPT-3 as a software, which includes: bias in generated text, poor understanding of actual words to produce lengthier and comprehensible pieces, and poor generalisation. Recently,  Yann LeCun, the VP & Chief AI Scientist at Facebook, trashed OpenAI’s massive language model. He called out on people’s unrealistic expectations from this software, which according to him ‘is entertaining, and perhaps mildly helpful as a creative tool.’ Deepfake Videos Deepfake, in itself, has been a rather controversial technology from the beginning. It has been a worldwide concern for the authorities regarding its misuse. While earlier, deepfakes were mostly restricted to making funny and entertaining videos, the growing accessibility of easy-to-use tools and the advancement in GANs made it easier for notorious minds to generate almost genuine-looking AI-generated videos and images. There were several instances of famous celebrities’, especially females’ faces being morphed on highly objectionable videos. Apart from this, deepfakes were feared to mischievously affect elections in a year when major countries and states were going to the polls. For example, India saw its first use of deepfakes for an election campaign when a BJP leader released a campaign video, where he originally spoke in Hindi, however, using deepfake technique, the same video was also released in a few other languages. While this particular use case seemed rather harmless, critics were quick to raise the alarm for possibly setting a hazardous precedent. AI-Based Grading System In UK Lockdowns across multiple countries meant that the normal functioning of life and society took a major hit. Every aspect, from healthcare, governance, education, and inter social mingling saw a complete set of challenges. Speaking of education, a lot of schools and colleges have been indefinitely closed due to fear of the virus. Exams have been postponed or cancelled together, keeping in view of the situation. However, in August, the UK’s exam regulation department called the Office of Qualifications and Examinations Regulation (Ofqual) decided to adopt an AI-based grading system to gauge student performance since the A-level examination that decides student gets to go to which university had to be cancelled. Parents and children protested strongly against this system as they claimed that the algorithms were highly biased against poorer students. Experts also called this system ‘unethical and harmful to education’. Keeping in view the harsh and relentless criticism and protests against this AI-based grading, the government of the UK had to finally drop it and decided that it would instead use teachers’ discretion and other parameters to allot grades to students. ‘Terrible’ Quality Of Reviews At NeurIPS This isn’t particularly an AI-related controversy but sheds a rather poor light on the research quality and practices. It concerns itself with the annual conference on Neural Information Processing Systems, NeurIPS 2020 held virtually this year. This popular machine learning event saw 38% more submissions than last year. The review period of paper submissions began in July, and by August, the conference sent out the paper reviews. These reviews came under scanner for being of ‘terrible’ quality as they were either unclear or incomplete, in a few cases, both. Critics demanded increased accountability for poor reviewers, and some even went to the extension of requesting disbandment of such persons with a proven bad record.","excerpt":"Artificial Intelligence has to be one of the most impactful technologies that the world has seen in recent years. It is no longer just limited to the quaint research and development labs of academies and bigger institutions but has successfully penetrated the normal and day-to-day functioning of the society.  Like any other technology, AI also […]","categories":["AI Trends"],"tags":["ai grading","clearview AI","deepfakes","GPT-3","object store database"],"author_name":"Shraddha Goled","publish_date":"2020-11-24T14:00:21","publication_year":"2020","word_count":1127,"keywords":["GPT-3","Go","API","artificial intelligence","machine learning","OpenAI","AI","ai grading","ML","deepfakes","GPT","Aim","clearview AI","R","object store database"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","OpenAI","Aim","R","Go","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/five-most-controversial-moments-of-ai-in-2020\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10022664,"title":"Overview Of Spleeter: A Music Source Separation Engine","content":"Spleeter is a source separation Python library created by the Deezer R&D team(Deezer is a music streaming platform like Spotify). It comes with pre-trained state-of-the-art models built using Tensorflow for various types of source separation tasks. But what is source separation? Source separation can be thought of as speaker diarization but for music. The speaker diarization models have to differentiate between the voices of different speakers and then split the original audio into multiple tracks corresponding to each speaker. Similarly, the source separation models have to differentiate between the different stems(sources) of audio in a music track, these stems can be the vocals, the sound of a particular instrument, or the sound of a group of instruments. Spleeter contains pre-trained models for the following source separation tasks: 2 stems separation: vocals\/accompaniment separation 4 stems separation: vocals, bass, drums, and other5 stems separation: vocals, bass, drums, piano, and other It is the first tool to offer 5 stems separation. Spleeter allows you to train your own source separation models or fine-tune the pre-trained ones for specific use-cases. Architecture & Approach The pre-trained models in Spleeter are U-Nets, i.e.,  encoder\/decoder convolutional neural networks(CNN) with skip connections. The U-Nets are 12 layers deep, 6 layers for the encoder and 6 for the decoder. The models were trained on internal dataset from Deezer using L1-norm loss between masked input mix spectrograms and target spectrograms. These models were compared with Open-Umix on the musdb18 dataset.  Open-Umix is another openly available music source separation system with state-of-the-art performance. The important point about this comparison is that the Spleeter models weren’t trained or optimized on this dataset. Standard source separation metrics were used for the comparison, namely Signal to Distorsion Ratio (SDR), Signal to Artifacts Ratio (SAR), Signal to Interference Ratio (SIR), and source Image to Spatial distortion Ratio (ISR). For most metrics, Spleeter is competitive with Open-Unmix, especially in terms of the Signal to Distorsion Ratio. Not only that, Spleeter is also very fast as it can separate a mixed audio file into 4 stems 100 times faster than real-time on a single GPU. Music Source Separation with Spleeter Install Spleeter Spleeter has two dependencies: ffmpeglibsndfile (optional, only needed for evaluation) Install from PyPI. pip install spleeter Get the audio file(s) for source separation. We’ll be using the demo audio file included in the Spleeter repository. wget https:\/\/github.com\/deezer\/spleeter\/raw\/master\/audio_example.mp3 Run the default two stem separation model spleeter separate -o output\/ audio_example.mp3 If you’re running this in a Windows CLI, you might run into an error. Use python -m spleeter instead of just spleeter to run the models. This will create a new folder in the -o directory, output\/,  with the name of the input track,  audio_example in our case. Navigate to this folder and you should find two files: vocals.wav and accompaniment.wav. Now, let’s try the five stems model spleeter separate -o output -p spleeter:4stems audio_example.mp3 This time, it will generate five files: vocals.wav, bass.wav,  piano.wav, drums.wav and, other.wav Last Epoch (Endnote) This post discussed Spleeter, a tool for music source separation with pre-trained models. These pre-trained models have already been incorporated into several professional audio software like Acon Digital, VirtualDJ, and Algoriddim. It can also be used for a plethora of Music Information Retrieval (MIR) tasks, such as: Vocal lyrics analysis tasks like audio-lyrics alignment and lyrics transcriptionSinger identificationMood or genre classificationMusic transcription tasks like chord transcription, drum transcription, chord estimation, and beat trackingVocal melody extraction References GitHubColab NotebookBlogGetting Started GuideExtended abstract submitted for 2019 ISMIR Conference","excerpt":"Spleeter is a source separation Python library created by the Deezer R&D team for various types of source separation tasks.","categories":["AI Trends"],"tags":["AI in music","Python Libraries","unet"],"author_name":"Aditya Singh","publish_date":"2021-03-22T18:00:00","publication_year":"2021","word_count":581,"keywords":["AI in music","Go","TPU","AI","neural network","Python Libraries","TensorFlow","Git","Colab","Python","GitHub","unet","R"],"extracted_tech_keywords":["AI","neural network","TensorFlow","Colab","TPU","Python","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/overview-of-spleeter-a-music-source-separation-engine\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":64598,"title":"Chief Analytics Officer Vs Chief Data Officer: What’s The Difference?","content":"Data executives are essential for a clear business strategy research as data-driven innovation has been critical for many years now. Of all the C-Level executives, there are only a few positions that deal with data. Two of the most popular ones include chief analytics officer, aka. Head of analytics, and chief data officer. But what is the fundamental difference between the two job roles? In this article, we will compare the two and bring out similarities and differences. Role Of A Chief Analytics Officer Chief analytics officer as a name suggests is a senior manager, which is responsible for the analytics operations within a company. The designation may also be known as the head of analytics, or chief analytics officer (CAO) is a fast-rising job role among senior-level management within companies. A chief analytics officer is responsible for digital transformation by making decisions based on the insights obtained from data, which is backed by data science techniques. While in most cases, chief analytics officers report directly to the CEO, it may not be the case, and head of analytics in many cases report to either chief information officers (CIOs) as well. Day-to-day activities of the head of analytics\/chief analytics officer include collecting data and using it to create business intelligence models. Chief analytics officers also interact with chief information officers in regard to the infrastructure needed for analytics operations. So, their focus is on creating business-oriented decisions by using the data that is available within an organisation, whether that data is from marketing, sales, and other operations. CAOs also develop data warehouses as a central repository for all information within a company as well as formula data governance and management frameworks. Reporting and visualisation tools may also be led by chief analytics officers. Role Of A Chief Data Officer Contrarily, a chief data officer (CDO) is a technology leadership role in regard to data management and working on data quality in data processes. Besides, CDOs may not necessarily possess data science skills, which with chief analytics officers or chief data scientist in the executive suite certainly do. A chief data officer (CDO) manages the governance and management of data across an enterprise as an asset through data processing and mining. CDOs determine what types of data businesses will opt to acquire, retain and utilise for business objectives. CDO saw the rise to senior management during the 80s when technology firms recognised the importance of master data management and data processing is a core part of business operations. Chief data officers ensure that data processes are running and effective data management is in place to achieve business goals. CDOs report to CEOs and work alongside different departments like marketing and sales. While a CDO may not necessarily be equipped with advanced analytics or any other techniques, they may be familiar with analytics, and the various open-source tools available today. At the same time, in some company, CDOs are directly responsible for analytics functions too. In addition, CDOs also look over data management storage and other IT related tasks that are usually handled by CIOs. The Difference The difference between a chief data officer (CDO) and chief analytics officer (CAO) is between managing data (in case of CDO) and leading data analytics (in case of CAO) which are two different functions even though the two positions may be used interchangeably. If we look back, the chief data officer is an older position in the overall enterprise space and existed prior to the coming of the chief analytics officer. In the recent past, many companies have combined the two functions or the two job roles in one because of the overlap. As companies are moving from data-driven strategies to analytics and AI-driven strategies, the focus has increased on analytics and not just data management. Compared to chief analytics officers, CDOs may have a better understanding of business management as they become solution architects that can build processes, which can impact the business outcomes. This is much like a chief information officer or chief digital officer. While chief analytics officers derive insights from data, CDO should focus on driving the maximum value from data.","excerpt":"Data executives are essential for a clear business strategy research as data-driven innovation has been critical for many years now. Of all the C-Level executives, there are only a few positions that deal with data. Two of the most popular ones include chief analytics officer, aka. Head of analytics, and chief data officer. But what […]","categories":["Deep Tech"],"tags":["chief data officer","Chief Data Scientist","different types of analytics"],"author_name":"Vishal Chawla","publish_date":"2020-05-06T15:00:00","publication_year":"2020","word_count":690,"keywords":["data science","Go","AI","chief data officer","Chief Data Scientist","data warehouse","Git","RAG","analytics","data governance","data quality","R","different types of analytics"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","Go","Git","data warehouse","data governance","data quality"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/chief-analytics-officer-vs-chief-data-officer-whats-the-difference\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10160610,"title":"Could Synthetic Data Change the Future of Robotics?","content":"Robots are getting smarter, faster, and more capable than ever. They are now a big part of evolution. Artificially generated information used to train AI models, which is known as synthetic data, has been hailed as a game-changer in robotics. By creating virtual environments, developers can teach robots complex tasks without the constraints of real-world data collection. Robots can now master difficult tasks without the limitations of real-world training, including time and cost. However, insights from industry leaders and researchers highlight major challenges that raise doubt about the efficacy of synthetic data in advancing robotics. Learn First, Act Second In traditional robotics, training a robot required real-world data: capturing endless hours of movement, testing, and repetition. This process was slow, expensive, and often limited by what researchers could physically achieve. Synthetic data flips the script. For example, consider NVIDIA’s Project GR00T and Isaac Sim. Imagine teaching a robot to stack shelves in a supermarket. With Isaac Sim, that training can happen entirely in a virtual replica of the store. Robots trained here can learn to navigate spaces, handle objects, and make decisions – all before they even touch the real world. The biggest challenge in robotics isn’t building better machines; it’s teaching them. Jim Fan, a senior research manager and lead of Project GR00T at NVIDIA, explained the problem simply. “Unlike with LLMs, where vast amounts of texts are readily available, you cannot simply download motor control signals from the internet.” Traditionally, researchers have had to wear headsets and control robots directly, repeating motions over and over to collect data. This is where DexMimicGen comes in. The tool developed at NVIDIA creates thousands of unique robot training scenarios from just a handful of human demonstrations. For instance, if a person demonstrates how to pick up a cup five times, DexMimicGen can use that data to generate thousands of variations. This allows robots to learn faster and generalise better in real-world situations. As Fan puts it, “The future of robot data and the entire robot learning pipeline will be generative.” The Challenges of Synthetic Data Experts like Ilya Sutskever, co-founder of OpenAI, have previously warned that AI systems might hit a “data wall”. It is about time these AI systems start getting smarter, to the extent of achieving superintelligence. He also acknowledged the challenges of defining synthetic data and how, while useful, it can sometimes fall short. When synthetic data is overused without real-world checks, it can lead to “model collapse”, in which robots struggle to perform reliably in new situations. To overcome these challenges, NVIDIA’s solutions include advanced physics simulations and tools like Omniverse Universal Scene Description (OpenUSD). The company also hints at the next AI wave of physical AI. This technology ensures synthetic environments are as realistic as possible, bridging the gap between virtual and real-world training. Making simulations closer to reality reduces the risk of over-reliance on synthetic data. India’s Role in Shaping the Future of Robotics India is becoming a key player in the robotics revolution. Companies like Addverb, based in Noida, are using synthetic data and NVIDIA’s platforms to develop robots that are transforming industries. Addverb’s Bot-Verse facility produces 1 lakh robots every year. These robots are tested and trained in virtual environments before hitting the market. This approach reduces costs, speeds up innovation, and ensures the robots are ready to perform in real-world scenarios. In addition to industrial robots, Indian companies are exploring robots for agriculture, logistics, and healthcare. Tools like DexMimicGen help Indian researchers train robots for specific challenges, such as navigating city streets or working on farms. This generative approach to synthetic data positions India as a leader in the global robotics race. Why This Matters for Everyone The future of robotics is aspirational, not just for researchers but for all of humanity. Whether it’s smarter machines in warehouses, robots assisting doctors in hospitals, or drones delivering packages to your doorstep, synthetic data is the fuel powering this transformation. The question isn’t whether synthetic data will change the future of robotics; it’s how fast it will be embraced. By addressing its challenges and utilising its potential, it is possible to create a future where robots are not just tools but essential partners in building a better world. And that’s a future worth getting excited about.","excerpt":"When synthetic data is overused without real-world checks, it can lead to “model collapse”, in which robots struggle to perform reliably in new situations.","categories":["AI Features"],"tags":["AI in Robotics","synthetic data generation"],"author_name":"Sanjana Gupta","publish_date":"2024-12-31T14:38:09","publication_year":"2024","word_count":712,"keywords":["API","synthetic data","OpenAI","AI","programming_languages:R","innovation","emerging_tech:synthetic data","AI in Robotics","ai_applications:robotics","synthetic data generation","R"],"extracted_tech_keywords":["AI","OpenAI","R","API","synthetic data","innovation","programming_languages:R","ai_applications:robotics","emerging_tech:synthetic data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/could-synthetic-data-change-the-future-of-robotics\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10020095,"title":"Guide To BenchmarkDotNet : A Benchmarking Library For DOTNET Developers","content":"Table of contentsIntroductionWhat is a benchmark?Overview of BenchmarkDotNetFeatures supported by BenchmarkDotNetPros of BenchmarkDotNetConfig Components of benchmarking architecturePractical Implementation Installation Create a benchmarking classMain methodRun the benchmarkAnalyze the summary reportReferences Introduction BenchmarkDotNet is a powerful, open-source, lightweight library extensively used by .NET developers for benchmarking their code. It was introduced by the .NET Foundation. Its current maintainers are Andrey Akinshin (Project Lead) and Adam Sitnik. (Have a look at the BenchmarkDotNet team here). Before going into the details of BenchmarkDotNet, let us understand in short, the meaning of benchmarking a code and why it is required. What is a benchmark? Benchmarking is an act of assessing the relative performance of a piece of code. In simple terms, the benchmark is a test run to know if some modification done to your code has improved, worsened or not affected its performance. It is required for understanding performance metrics of the methods you have used in your application so that those metrics can be used during the code optimization process. Depending upon the extent of changes you make, the benchmark may have a wide scope or a micro-benchmark assessing minute changes. Overview of BenchmarkDotNet BenchmarkDotNet library transforms the methods used in your application into benchmarks. It also enables you to share reproducible measurement experiments. BenchmarkDotNet is used by over 4500 projects till date. To name a few – Mono, ASP.NET Core, ML.NET, Entity Framework Core, dotnet\/runtime (.NET Core runtime and libraries), Roslyn (C# and Visual Basic compiler), .NET Docs, TensorFlow.NET etc. Features supported by BenchmarkDotNet Operating systems: Windows, Linux, MacOSProgramming languages: C#, F#, Visual BasicArchitectures: x86, x64, ARM, ARM64 and WasmRuntimes: .NET 5+, .NET Framework 4.6.1+, .NET Core 2.0+, Mono, CoreRT https:\/\/twitter.com\/SergioPedri\/status\/1236710348943736836 Pros of BenchmarkDotNet Simplicity Very complicated performance experiments can be designed in the declarative style using simple APIs of BenchmarkDotNet. For instance, to compare benchmarks with each other, mark one of the benchmarks as the baseline via [Benchmark(baseline: true)]. That benchmark will then be compared with all of the other benchmarks. Automation Reliable benchmarks require a lot of benchmark code (i.e. sections of code that are repeated multiple times with minor variations). While writing such repetitive code, you are likely to commit a mistake which may spoil your measurements. BenchmarkDot handles such situations. It also performs certain advanced tasks such as measuring the managed memory traffic and printing disassembly listings of your benchmarks. Reliability BenchmarkDotNet allows for achieving high measurement precision. It tries to choose the best benchmarking parameters. It achieves a good trade-off between the measurement precision and the total time taken to run all the benchmarks. It protects you from most of the benchmarking pitfalls such as deciding the number of method invocations, number of actual iterations and so on. The library handles all this stuff on its own based on the values of statistical metrics. It comprises numerous heuristics, checks, hacks, and tricks which have the potential to make your results more reliable. Friendliness BenchmarkDotNet performs the core part of performance assessment i.e. analyzing the performance data and presents results in a user-friendly form. It gives a summary table that contains a lot of useful data about the executed benchmarks. By default, it includes only the most important columns which are customizable. The column set is adaptive; it depends on the benchmark definition and measured values. BenchmarkDotNet also alerts you about some unusual properties of your performance distributions (if any). Besides, it shows only the essential information depending on your results. It keeps the summary user-friendly – small for primitive cases and extended only for the complicated cases. Additional statistics and visualizations, however, can always be added manually. Config Config in BenchmarkDotNet is a set of jobs, columns, exporters, loggers, diagnosers, analysers, validators used to build benchmarks. A short definition of each is these terms are as follows: Jobs: It is a set of characteristics which describe the way to run the benchmarks. One or more jobs can be specified for each benchmark. Columns: refer to columns in the summary table Exporters: An exporter enables exporting results of your benchmark in various formats. Csv, html and markdown are the default exporters. By default, files with results will be located in .\\BenchmarkDotNet.Artifacts\\results directory. Loggers: They enable logging of the results of your benchmarks. By default, log is found on the console and in <BenchmarkName>.log file. Diagnosers: They attach to your benchmarkers to retrieve useful information. ToolChains: BenchmarkDotNet generates, builds and executes a new console app for every benchmark and thus enables process-level isolation. A toolchain contains the app generator, builder, and executor. Default toolchains: Roslyn for Full .NET Framework and Monodotnet cli for .NET Core and CoreRT Analyzers: An analyzer analyzes the summary of each benchmark and produces appropriate warnings wherever necessary. Validators: A validator validates each benchmark before its execution and produces validation errors. If any of those errors is critical, then execution of all the benchmarks fails. Filters: They allow you to choose only some and not all of the benchmarks specified. Orderers: They enable customization of the order of benchmark results in the summary table. Components of benchmarking architecture Benchmarks – a web application comprising various scenarios to benchmarkBenchmarksServer – a web application that queues jobs which can run custom web applications to be benchmarked.BenchmarksClient – a web application that queues jobs to create custom client loads on a web applicationBenchmarksDriver – a command-line application that can enqueue server and client jobs and display the results locally.A database server that can run any or all of PostgreSql, Sql Server, MySql, MongoDb Visit this GitHub repository to know about the step-wise installation of the architecture. Practical Implementation Here’s an example of an ASP.NET application which demonstrates how to benchmark a C# code using BenchmarkDotNet. Installation Create a new console application. Then install the BenchmarkDotNet NuGet package. Create a benchmarking class [MemoryDiagnoser]  \/\/type of diagnoser specified public class Example { int ItemsCount = 10000; \/* use the Benchmark attribute on top of each of the methods that are to be benchmarked *\/ [Benchmark] \/\/ class to concatenate strings using StringBuilder public string A() { var strbuilder = new StringBuilder(); for (int i = 0; i < ItemsCount; i++) { strbuilder.Append(\"Item\" + i); } return strbuilder.ToString(); } \/\/end of A() [Benchmark] \/\/ class to concatenate strings using GenericList public string B() { var list = new List<string>(NumberOfItems); for (int i = 0; i < ItemsCount; i++) { list.Add(\"Item\" + i); } \/\/end of for loop return list.ToString(); } \/\/end of B() } \/\/end of class Example Main method In the Main method Program.cs file, the initial starting point — the BenchmarkRunner class must be specified in order to inform BenchmarkDotNet to run benchmarks on the specified class (here Example class). static void Main(string[] args) { var summaryReport = BenchmarkRunner.Run<Example>(); } Run the benchmark Note: Always run your project in release mode while using benchmarking. The C# compiler does a few optimizations in release mode which are not available in debug mode. Running the project in debug mode will result in an error. Suppose, the name of the project file is Demo.csproj. To run the benchmark, give the following command at Visual Studio command prompt. dotnet run -p Demo.csproj -c Release If you fail to mention the configuration parameter (-c Release) in the above line of code, benchmarking will be attempted on non-optimized code in debug mode and hence will cause an error. Analyze the summary report Once the benchmarking process gets executed, a summary of the results will be displayed at the console window. It contains information related to the application’s performance. It also shows details about the environment in which the benchmarks were executed e.g. version of BenchmarkDotNet, operating system, computer hardware, .NET version and much more. Output: References To dive deeper into the powerful BenchmarkDotNet library, refer to the following sources: Official websiteDocumentationGitHub repository","excerpt":"Introduction BenchmarkDotNet is a powerful, open-source, lightweight library extensively used by .NET developers for benchmarking their code. It was introduced by the .NET Foundation. Its current maintainers are Andrey Akinshin (Project Lead) and Adam Sitnik. (Have a look at the BenchmarkDotNet team here). Before going into the details of BenchmarkDotNet, let us understand in short, […]","categories":["Deep Tech"],"tags":[],"author_name":"Nikita Shiledarbaxi","publish_date":"2021-02-14T13:00:00","publication_year":"2021","word_count":1297,"keywords":["PostgreSQL","Go","TPU","AI","MongoDB","ML","Git","SQL","TensorFlow","R"],"extracted_tech_keywords":["AI","ML","TensorFlow","TPU","MongoDB","PostgreSQL","R","SQL","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-benchmarkdotnet-a-benchmarking-library-for-dotnet-developers\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10065118,"title":"Highlights from Intel® Hands-on oneAPI workshop: Getting started with Intel® Optimisation for PyTorch*","content":"Intel® and Analytics India Magazine have successfully concluded the oneAPI AI Analytics Toolkit Workshop – a master class on Intel® optimisation techniques for accelerating deep learning workloads on March 25, 2022. The workshop, intended for AI and ML developers, data scientists, AI enthusiasts, AI researchers, GPU and HPC programmers, saw more than 200 techies joining this insightful session. The workshop took the attendees through the Intel® optimisations calibrated for PyTorch*, installation guide, and performance boost number. They also learned about ease-of-use Python API, vectorisation, parallelism, quantisation, operator fusion, constant folding, etc. Intel® has been working with Facebook (now Meta) to contribute optimisations to the PyTorch*community and commits to continuously optimising PyTorch* with future advancements of Intel® HWs. Intel® had a demo on the following topics: Intel® Optimization for PyTorch*Intel® Extension for PyTorch*Intel® Extension for PyTorch*targets optimisations on AVX-512 instruction set.Intel® Optimization for PyTorch*released in oneAPI AI Analytics Toolkit. Highlights Kavita Aroor, the Developer Marketing Manager – APJ at Intel®, started the workshop with a welcome note, introduced the workshop instructors of the day, and explained the rules and guidelines of the contests and audience polls along with the developer ecosystem program. It was followed by a session on Intel® oneAPI Ecosystem by Aditya Sirvaiya, an AI Technical Consulting Engineer in the Intel® Software group. Aditya explained that oneAPI is a cross-architecture language based on C++ and SYCL standards. He spoke about the powerful libraries designed to accelerate domain-specific functions and the set of advanced compilers, libraries and porting, analysis and debugger tools that come with it. Structure of the oneAPI toolkit Image: Intel® Intel® oneAPI base toolkit Aditya described the Intel® oneAPI base toolkit as a core set of tools and libraries for developing high-performance applications on Intel® CPUs, GPUs and FPGAs. It is used by a broad range of developers across industries. He also elaborated on its benefits and features. Data parallel C++ compiler, library and analysis toolsDPC++ Compatibility tool helps users migrate existing code written in CUDAPython distribution includes accelerated scikit-learn, NumPy, SciPy libraries Image: Intel Intel® AI analytics toolkit powered by oneAPI Aditya then introduced the attendees to the Intel® AI analytics toolkit powered by oneAPI. It aims to achieve end-to-end performance for data science and AI workloads and is widely used by data scientists, AI researchers, and machine and deep learning developers. It provides drop-in acceleration for ML and analytics workflows with compute-intensive Python libraries. It also gives seamless scaling of data pipelines across multi-cores and multi-nodes to optimise end-to-end solutions with cross-architecture support (Intel® CPUs, GPUs). Intel® oneAPI Data Analytics Library (oneDAL) Further, Aditya took the attendees through the Intel® oneAPI Data Analytics Library (oneDAL), which consists of optimised building blocks for all stages of data analytics on Intel® architecture. What it does: Pre-processing-decompression, filtering, normalisation Transformation-aggregation, dimension reduction Analysis-summary statistics, clustering, etc. Modeling-machine learning (training), parameter estimation simulation Validation-hypothesis testing, model errors Decision making-forecasting, decision trees, etc Following that, an expert panel by Jing Xu, Senior Technical Consulting Engineer working as an AI specialist within the Intel® Software group, followed, which introduced the attendees to Intel® Optimization for PyTorch*. Key features and benefits He listed out more key benefits of Intel® Optimization for PyTorch*. Accelerates end-to-end AI and data science pipelines and achieves drop-in acceleration with optimised Python tools built using oneAPI libraries like oneMKL, oneDNN, oneCCL, oneDAL, etc. Achieves high performance for deep learning training and inference with Intel®-optimised versions of TensorFlow and PyTorch*, and low-precision optimisation with support for fp16, int8 and bfloat16. Image: Intel® Image: Intel® During the event, Analytics India Magazine also ran a Lucky Draw, wherein ten lucky participants won an Amazon Voucher worth INR 2000\/- each at the end of the workshop. The winners were selected based on their engagement with Discord throughout the workshop. <​​https:\/\/discord.gg\/ycwqTP6> The Lucky Draw winners are: Mandeep SuriYogesh GoraneParikshit RathodeRakeshkumar TammisettiKowsalya VeerabadranShubham SoniSreenivas ChintadaMedha SharmaAlok SrivastavaUjwal Kiran Find more details here. Check out the recording here. Download Intel® oneAPI AI Analytics Toolkit here.","excerpt":"A look at some of the major highlights from the oneAPI AI Analytics Toolkit Workshop by Intel® and Analytics India Magazine.","categories":["Deep Tech"],"tags":["Intel","oneAPI","Pytorch"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-04-18T13:00:00","publication_year":"2022","word_count":660,"keywords":["Pytorch","data science","scikit-learn","machine learning","AI","PyTorch","ML","Aim","deep learning","analytics","TensorFlow","oneAPI","Intel"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","Aim","TensorFlow","PyTorch","scikit-learn"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/highlights-from-intel-hands-on-oneapi-workshop-getting-started-with-intel-optimisation-for-pytorch\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10165103,"title":"Microsoft Introduces Dragon Copilot, a New AI Assistant for Healthcare","content":"Microsoft has launched Dragon Copilot, an AI assistant designed to improve clinical workflows in healthcare. By combining Dragon Medical One’s voice dictation and DAX Copilot’s ambient listening, Dragon Copilot helps clinicians reduce administrative tasks and improve patient care. As part of the Microsoft Cloud for Healthcare, Dragon Copilot offers secure architecture to ensure privacy and compliance. According to the blog, clinician burnout in the US dropped from 53% in 2023 to 48% in 2024, partly due to technology advancements. It helps clinicians streamline documentation with multilingual note creation, automated tasks, natural language dictation, and customisable templates. It also supports AI-driven searches for medical information and automates tasks like referral letters, clinical summaries, and after-visit summaries. Clinicians report saving 5 minutes per encounter using Dragon Copilot’s AI capabilities, the blog said. “With the launch of our new Dragon Copilot, we are introducing the first unified voice AI experience to the market, drawing on our trusted, decades-long expertise that has consistently enhanced provider wellness and improved clinical and financial outcomes for provider organisations and the patients they serve,” Joe Petro, corporate VP of Microsoft Health and Life Sciences Solutions, said. Early users have reported improved workflow efficiency, with 93% of patients noting better experiences and clinicians saving an average of five minutes per encounter. Dragon Copilot will be available in the US and Canada in May 2025, with plans for expansion to the UK, Germany, France, and the Netherlands later in the year.","excerpt":"Clinicians report saving 5 minutes per encounter using Dragon Copilot’s AI capabilities.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Aditi Suresh","publish_date":"2025-03-04T17:48:47","publication_year":"2025","word_count":241,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","RAG","Rust","GAN","R","programming_languages:Rust","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","RAG","R","Go","Rust","GAN","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-introduces-dragon-copilot-a-new-ai-assistant-for-healthcare\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042191,"title":"Sound Pitch Recognition Using SPICE","content":"Sound is a vital sense to us human beings. The pitch of the sound is a somewhat approximate measure of the frequency. High pitch corresponds to higher frequency; similarly, lower pitch denotes low frequency. What our auditory system does is that it tracks the relative difference in pitch, hence recognizing different sounds which have different characteristics of their own. A perfect example is when listening to a song. We can differentiate among the melodies of the song. Today’s article is about Pitch Recognition, aka Pitch Estimation. This domain has received paramount attention in the past few decades due to its vitality in several fields ranging from music information retrieval to speech analysis. Traditionally what used to happen was either one could implement using time domain or frequency domain. These handcrafted models posed one problem – the need for annotated data. This is a pretty tedious and laborious task to obtain the frequency and temporal resolution required for training the model. Marco Tagliasacchi, a research scientist at Google Research, presented a solution to the problem mentioned above, which solved missing annotated data in November 2019. In simple basic terms, this approach calculated the relatedness between different sounds rather than calculating the absolute.  SPICE (Self-supervised PItCh Estimation) was designed on this idea and presented with the research paper. The model consists of a convolutional encoder; this produces a singular scalar embedding that maps linearly with the pitch. Two signals are fed to the encoder (one reference and one random signal), and the author has defined the domain using constant-Q transform for convenience. A loss function was devised, forcing the difference between the scalar embeddings to stay proportional to the already known difference. Pitch, as we know, is well defined under the condition of it being harmonic; that is, it should contain components with integer multiples of the fundamental frequency. An important function of the model is determining when the output is reliable and meaningful. SPICE has been designed to learn the level of confidence of pitch recognition or estimation, you may say in a self-supervised manner. The model was evaluated using publicly available datasets and outperformed the handcrafted models that, too, had no access to true labels or absolute values. For example, SPICE outperformed CREPE (Convolutional Representation for Pitch Estimation) and SWIPE (Sawtooth Waveform Inspired Pitch Estimator) on the MIR-1k dataset on four classes, namely clean, 20 dB, 10dB and 0dB. Let’s look at a code implementation in the following parts. The following implementation is in reference to the official implementation. Code Implementation of  Pitch Recognition Using SPICE Imports and Dependencies # syntax for installing multiple libraries # timidity is a lightweight package for playing MIDI files # libsndfile for reading and writing audio files automatically. !sudo apt-get install -q -y timidity libsndfile1 ''' All the imports to deal with sound data pydub for manipulating audio files numba for fast machine code, parallelising python code librosa for audio and sound analysis music21 is python toolkit for computer aided musicology (CAM) ''' !pip install -q pydub numba==0.48 librosa music21 MIDI files – is Musical Instrument Digital Interface; these don’t contain any actual audio data like Wav file or mp3. Hence are smaller in size. # directory operations, math operations and mathematical stats import logging import statistics import sys import math # displaying wav files, display in Google colab notebook cell itself from scipy.io import wavfile from IPython.display import Audio, Javascript # for computer aided musicology (study of music\/sounds) # audiosegment for audio feature extraction, classification from pydub import AudioSegment import music21 # tensorflow for leveraging model import tensorflow as tf import tensorflow_hub as hub # manipulation of numeric data and sound data as well as plotting different graphs for audio import librosa from librosa import display as librosadisplay import numpy as np import matplotlib.pyplot as plt # toolkit for encoding binary data into ASCII from base64 import b64decode # built-in module logging allows writing status messages to a file or any # other output streams logger.setLevel(logging.ERROR) logger = logging.getLogger() # checking the tf and librosa versions print(\"librosa: %s\" % librosa.__version__) print(\"tensorflow: %s\" % tf.__version__) Audio Input NOTE: The following Javascript snippet has been taken from the official GitHub repository for making an interface for recording your input. RECORD = \"\"\" const sleep  = time => new Promise(resolve => setTimeout(resolve, time)) const b2text = blob => new Promise(resolve => { const reader = new FileReader() reader.onloadend = e => resolve(e.srcElement.result) reader.readAsDataURL(blob) }) var record = time => new Promise(async resolve => { stream = await navigator.mediaDevices.getUserMedia({ audio: true }) recorder = new MediaRecorder(stream) chunks = [] recorder.ondataavailable = e => chunks.push(e.data) recorder.start() await sleep(time) recorder.onstop = async ()=>{ blob = new Blob(chunks) text = await b2text(blob) resolve(text) } recorder.stop() }) \"\"\" def record(sec=5): try: from google.colab import output except ImportError: print('No possible to import output from google.colab') return '' else: print('Recording') display(Javascript(RECORD)) s = output.eval_js('record(%d)' % (sec*1000)) fname = 'recorded_audio.wav' print('Saving to', fname) b = b64decode(s.split(',')[1]) with open(fname, 'wb') as f: f.write(b) return fname Created this function for choosing different inputs, which are as below: Record audio with interface right here in colab notebookUploading from the local system.Using a file from Google drive (mount the drive on the notebook)Downloading file from the Internet. # Input URL INPUT_SRC = 'https:\/\/storage.googleapis.com\/download.tensorflow.org\/data\/c-scale-metronome.wav' print('Selected', INPUT_SRC) # condition to check for the url audio if INPUT_SRC == 'RECORD': # function for recording own voice, choose the duration as suitable but keep it less uploaded_file = record(5) elif INPUT_SRC == 'UPLOAD': try: # import from google storage from google.colab import files except ImportError: print(\"ImportError\") else: uploaded = files.upload() for fn in uploaded.keys(): print('Uploaded file \"{name}\" of length {length} bytes'.format( name=fn, length=len(uploaded[fn]))) uploaded_file_name = next(iter(uploaded)) print('Uploaded file: ' + uploaded_file_name) elif INPUT_SOURCE.startswith('.\/drive\/'): try: from google.colab import drive except ImportError: print(\"ImportError\") else: # mount google drive for local audio drive.mount('\/content\/drive') gdrive_audio_file = 'name provided by you.wav' uploaded_file_name = INPUT_SOURCE elif INPUT_SOURCE.startswith('http'): !wget --no-check-certificate 'https:\/\/storage.googleapis.com\/download.tensorflow.org\/data\/c-scale-metronome.wav' -O c-scale.wav uploaded_file_name = 'c-scale.wav' Audio Data Preparation SAMPLE_RATE = 16000 # Function to convert the user created audio to that format, which the model expects # it should be one channel and of 16k sample rate def convert_audio(user_file, output_file='converted_audio_file.wav'): # variable aud for audio from user file input aud = AudioSegment.from_file(user_file) aud = audio.set_frame_rate(SAMPLE_RATE).set_channels(1) # export the audio file in wav format so that we can listen to it audio.export(output_file, format=\"wav\") return output_file # Converting to the expected format for the model, # the uploaded file name is at # the variable uploaded_file_name which can be set accordingly converted = convert_audio(uploaded_file_name) # Load audio samples from the wav file: sample_rate, audio_samples = wavfile.read(converted, 'rb') # printing some basic information about the audio. duration = len(audio_samples)\/sample_rate # sample rate set to 16k earlier print(f'Sample rate: {sample_rate} Hz') # string formatting for duration at 2 decimal point print(f'Total duration: {duration:.2f}s') print(f'Size of the input: {len(audio_samples)}') # listen to the wav file. Audio(audio_samples, rate=sample_rate) Function for Getting the spectrogram. # visualize the audio as a waveform. _ = plt.plot(audio_samples) MAX_ABS_INT16 = 32768.0 # function for plotting spectrogram def plot_spect(x, sample_rate, show_black_and_white=False): # start for plot x_stft = np.abs(librosa.stft(x, n_fft=2048)) # matplotlib fig, ax = plt.subplots() # set small size for easy cell fig.set_size_inches(20, 10) # setting amplitude function x_stft_db = librosa.amplitude_to_db(x_stft, ref=np.max) if(show_black_and_white): # get the spect. plot librosadisplay.specshow(data=x_stft_db, y_axis='log', sr=sample_rate, cmap='gray_r') else: librosadisplay.specshow(data=x_stft_db, y_axis='log', sr=sample_rate) # color bar is necessary plt.colorbar(format='%+2.0f dB') plot_stft(audio_samples \/ MAX_ABS_INT16 , sample_rate=EXPECTED_SAMPLE) plt.show() Model Execution # Loading the SPICE model mod = hub.load(\"https:\/\/tfhub.dev\/google\/spice\/2\") # feed the audio to the SPICE tf.hub model to obtain pitch and uncertainty outputs as tensors. model_out = mod.signatures[\"serving_default\"](tf.constant(audio_samples, tf.float32)) # pitch output for estimation pitch_out = model_output[\"pitch\"] uncertainty_out = model_out[\"uncertainty\"] # 'Uncertainty' means the inverse of confidence. confidence_out = 1.0 - uncertainty_out # again a plot for above uncertainty and confidence fig, ax = plt.subplots() fig.set_size_inches(20, 10) plt.plot(pitch_outputs, label='pitch') plt.plot(confidence_outputs, label='confidence') plt.legend(loc=\"lower right\") plt.show() Now we have to remove low scores of Pitch and plot them. # store the values in a list confidence_out = list(confidence_out) # traverse the list pitch_out = [ float(x) for x in pitch_out] # indexing through the length indices = range(len (pitch_outputs)) confident_pitch_out = [ (i,p) for i, p, c in zip(indices, pitch_out, confidence_out) if  c >= 0.9  ] confident_pitch_out_x, confident_pitch_out_y = zip(*confident_pitch_out) Output # plotting graph for higher pitch scores fig, ax = plt.subplots() fig.set_size_inches(20, 10) ax.set_ylim([0, 1]) plt.scatter(confident_pitch_out_x, confident_pitch_out_y, ) plt.scatter(confident_pitch_out_x, confident_pitch_out_y, c=\"r\") plt.show() The pitch values returned by SPICE are in the range of 0 – 1; we have to convert them to absolute pitch values in Hertz. def out2hz(pitch_out): # These Constants have been taken from https:\/\/tfhub.dev\/google\/spice\/2 PT_OFFSET = 25.58 PT_SLOPE = 63.07 FMIN = 10.0; BINS_PER_OCTAVE = 12.0; # formula cqt_bin = pitch_output * PT_SLOPE + PT_OFFSET; return FMIN * 2.0 ** (1.0 * cqt_bin \/ BINS_PER_OCTAVE) confident_pitch_value_hz = [ out2hz(p) for p in confident_pitch_out_y ] Checking how good the prediction is, by overlaying the predicted pitches over the original spectrum. Changed the original spectrum to black and white for better visibility. plot_stft(audio_samples \/ MAX_ABS_INT16 , sample_rate=EXPECTED_SAMPLE, show_black_and_white=True) # Conveniently, since the plot is in log scale, the pitch outputs # also get converted to the log scale automatically by matplotlib. plt.scatter(confident_pitch_out_x, confident_pitch_value_hz, c=\"r\") plt.show() Conversion to Musical Notes Taking care when there is no singing, size of each note (different offsets). # we have to put zero where there is no singing. pitch_output_and_rest = [ out2hz(p) if c >= 0.9 else 0 for i, p, c in zip(indices, pitch_out, confidence_out) ] Adding note offsets A4 = 440 C0 = A4 * pow(2, -4.75) note_ = [\"C\", \"C#\", \"D\", \"D#\", \"E\", \"F\", \"F#\", \"G\", \"G#\", \"A\", \"A#\", \"B\"] def hz2offset(freq): # Measures the quantization error for a single note. if freq == 0: # Rests always have zero error. return None # Quantized note. h = round(12 * math.log2(freq \/ C0)) return 12 * math.log2(freq \/ C0) - h # The ideal offset is the mean quantization error for all the notes # (excluding rests): offsets = [hz2offset(p) for p in pitch_output_and_rest if p != 0] print(\"offsets: \", offsets) ideal_offset = statistics.mean(offsets) print(\"ideal offset: \", ideal_offset) Open Sheet Music Display NOTE: The following is a Javascript snippet from official GitHub to develop an interface on screen for offsets display. from IPython.core.display import display, HTML, Javascript import json, random def showScore(score): xml = open(score.write('musicxml')).read() showMusicXML(xml) def showMusicXML(xml): DIV_ID = \"OSMD_div\" display(HTML('<div id=\"'+DIV_ID+'\">loading OpenSheetMusicDisplay<\/div>')) script = \"\"\" var div_id = { {DIV_ID} }; function loadOSMD() { return new Promise(function(resolve, reject){ if (window.opensheetmusicdisplay) { return resolve(window.opensheetmusicdisplay) } \/\/ OSMD script has a 'define' call which conflicts with requirejs var _define = window.define \/\/ save the define object window.define = undefined \/\/ now the loaded script will ignore requirejs var s = document.createElement( 'script' ); s.setAttribute( 'src', \"https:\/\/cdn.jsdelivr.net\/npm\/opensheetmusicdisplay@0.7.6\/build\/opensheetmusicdisplay.min.js\" ); \/\/s.setAttribute( 'src', \"\/custom\/opensheetmusicdisplay.js\" ); s.onload=function(){ window.define = _define resolve(opensheetmusicdisplay); }; document.body.appendChild( s ); \/\/ browser will try to load the new script tag }) } loadOSMD().then((OSMD)=>{ window.openSheetMusicDisplay = new OSMD.OpenSheetMusicDisplay(div_id, { drawingParameters: \"compacttight\" }); openSheetMusicDisplay .load({ {data} }) .then( function() { openSheetMusicDisplay.render(); } ); }) \"\"\".replace('{ {DIV_ID} }',DIV_ID).replace('{ {data} }',json.dumps(xml)) display(Javascript(script)) return EndNote We can easily listen back to the audio files by changing them into wav format. I recommend using different commands when recording audio, longer durations and different sounds from online resources. We successfully overcame traditional handcrafted problems in this article and developed a self-supervised technique for Pitch Estimation. References: Official Github RepositoryResearch PaperOfficial Source CodeColab Implementation","excerpt":"Article is about Pitch Recognition, aka Pitch Estimation.","categories":["AI Trends"],"tags":["self supervised learning"],"author_name":"Mudit Rustagi","publish_date":"2021-06-23T11:00:00","publication_year":"2021","word_count":1895,"keywords":["NumPy","TPU","AI","ML","RAG","Colab","Ray","Matplotlib","Python","TensorFlow","self supervised learning"],"extracted_tech_keywords":["AI","ML","Ray","TensorFlow","Colab","NumPy","Matplotlib","RAG","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/sound-pitch-recognition-using-spice\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10059156,"title":"IBM records decade’s biggest growth in its sales","content":"International Business Machines (IBM) Corp. Sales recorded a revenue of USD 16.7 billion, up 6.5 percent, in their fourth quarter. The company’s software revenue also went up by 8 percent. This is the company’s biggest increase in the last 10 years. According to Bloomberg, analysts were expecting a revenue of USD 16 billion. The stock jumped 7.5% in extended trading. “We increased revenue in the fourth quarter with hybrid cloud adoption driving growth in software and consulting,” said Arvind Krishna, IBM chairman and chief executive officer. “Our fourth-quarter results give us confidence in our ability to deliver our objectives of sustained mid-single digit revenue growth and strong free cash flow in 2022.” On November 3, 2021, IBM hived off Kyndryl. The results are the first since IBM completed the spinoff of a large portion of its legacy infrastructure services unit to build a new company. This includes service operations like managing client data centers and traditional information-technology support. IBM’s hybrid cloud revenue in the fourth quarter was up by up 16 percent to $6.2 billion. This makes the full year revenue to USD 20.2 billion, up by 20 percent. The software segment clocked a revenue of USD 7.3 billion, up by 8.2 percent (Hybrid Platform & Solutions was up 7 percent, Red Hat up by 19 percent, Automation up by 13 percent, Data & AI up by 1 percent and security was down 2 percent). “In 2021, we continued to invest for the future by increasing R&D spending, expanding our ecosystem and acquiring 15 companies to strengthen our hybrid cloud and AI capabilities,” said James Kavanaugh, IBM senior vice president and chief financial officer. “With the separation of Kyndryl we now have taken the next step in the evolution of our strategy, creating value through focus and strengthening our financial profile.”","excerpt":"International Business Machines (IBM) Corp. Sales recorded a revenue of USD 16.7 billion, up 6.5 percent, in their fourth quarter. The company’s software revenue also went up by 8 percent. This is the company’s biggest increase in the last 10 years. According to Bloomberg, analysts were expecting a revenue of USD 16 billion. The stock […]","categories":["AI News"],"tags":["Arvind Krishna","Hybrid Cloud","IBM","Kyndryl"],"author_name":"Meeta Ramnani","publish_date":"2022-01-25T19:26:51","publication_year":"2022","word_count":301,"keywords":["Kyndryl","programming_languages:R","AI","Git","automation","Hybrid Cloud","IBM","Arvind Krishna","R"],"extracted_tech_keywords":["AI","R","Git","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-records-decades-biggest-growth-in-its-sales\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10059873,"title":"Israel based Corsight AI claims it can model a person’s face from DNA sample","content":"Corsight AI, an Israeli-founded organisation providing facial recognition services, claims to have devised a way to create facial profiles from DNA samples. The product was first introduced by Corsight CEO Robert Watts and executive vice president Ofer Ronen at the Imperial Capital Investors Conference in New York City on December 15. The product roadmap included “voice to face,” “DNA to face,” and “movement” (or gait recognition) as expansions of the company’s facial recognition capabilities. According to a company slide deck made available to the surveillance research group IPVM, the “DNA to Face” product “constructs a physical profile by analysing genetic material contained in a DNA sample.” The company has previously drawn flak over exaggerating the competency and accuracy of its facial recognition system. Last November, Corsight CEO Watts claimed that Corsight’s technology could “identify someone with a face mask—not just with a face mask, but with a ski mask.” However, Corsight’s AI only clocked a 65% confidence score when completing this task. The company keeps its work low-key and has not lifted the lid on its plans and future products. However, marketing materials suggest CorsightAI is targeting its services towards government and law enforcement agencies. Old wine in new bottle The “genomics-based, healthy intelligence” company Human Longevity had claimed to use DNA to predict faces back in 2017. However, experts found such claims to be suspect, and a former employee had attested that Human Longevity couldn’t pick a person out of a crowd using a DNA sample. Further, the chief science officer of the genealogy platform MyHeritage, Yaniv Erlich, published a study highlighting the major flaws in Human Longevity’s research. Parabon NanoLabs uses its product line, Snapshot, to give law enforcement physical depictions of people from genetic samples. The phenotypic characteristics (such as eye and skin colour) of these computer-generated renderings come with a confidence score. So, for example, there could be an 80% probability of the person being pursued having green eyes. According to Parabon’s director of bioinformatics, Ellen McRae Greytak, the company has helped solve over 200 cases in the last seven years. Unlike Corsight, Parabon doesn’t claim the physical profiles they create can be used as input for facial recognition systems. The technology isn’t precise enough for facial recognition algorithms to deliver accurate results. While experts claim the science to support Corsight AI’s product doesn’t exist, it could compound the ethical, privacy, and bias challenges facial recognition technology is already causing. Predicting human physical traits from genomic data sparks privacy concerns. Albert Fox Cahn, a civil rights lawyer and executive director of the Surveillance Technology Oversight Project, claims the idea of creating something that has the “granularity and fidelity” to be put through a facial recognition system “is pseudoscience.” Evolution Facial recognition technology has gone over numerous changes since its inception. Coined in 1960, identities were automatically differentiated based on the manual marking of various “landmarks” on the face, like the placing of the eyes and the mouth. Later, the work was extended and standardised to include 21 specific subjective markers like hair colour and lip thickness to automate the recognition. In the late 80s, scientists applied linear algebra to the problem of facial recognition and formed the Eigenface system. It was in the early 1990s when development for the technology for commercial uses was initiated. In 2006, the US government supported the Face Recognition Grand Challenge (FRGC) to promote and advance face recognition technology. Here, 3D face scans, high-resolution face images, and iris images were used in the tests to make the technology 100 times more accurate. It was not until 2010, when the consumer experienced face recognition technology that was introduced by Facebook to identify people whose faces featured in the photos of their users. The major breakthrough that we see now happened when Apple launched the iPhone X that could be unlocked with FaceID. Post that, the technology is being used by airlines, airports, border controls, stadiums, transport hubs, mega-events, concerts, and conferences, among others.","excerpt":"The company has previously drawn flak over exaggerating the competency and accuracy of its facial recognition system.","categories":["AI Features"],"tags":[],"author_name":"Srishti Mukherjee","publish_date":"2022-02-05T11:00:00","publication_year":"2022","word_count":661,"keywords":["Go","API","AWS","AI","BERT","Aim","llm_models:BERT","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","API","BERT","GAN","ViT","llm_models:BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/israel-based-corsight-ai-claims-it-can-model-a-persons-face-from-dna-sample\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10040995,"title":"Why ML Capabilities Of GCP Is Way Ahead Of AWS &#038; Azure","content":"Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft Azure are the leading cloud providers by a long shot. Though late to the party, GCP has seen robust growth over the years. Google Cloud’s revenue jumped nearly 46 percent year-on-year to $4.04 billion in the first quarter of 2021. “All vendors offer strong ML services and functionalities, but this is where GCP stands out as their years of search engine expertise, and research come into play,” said Diwakar Chittora, Founder & CEO, IntelliPaat. Here is how GCP offers more benefits than AWS, Azure. Tensor Processing Unit (TPU) TPUs are Google’s custom-developed application-specific integrated circuits (ASICs) to accelerate ML workloads. A big advantage for GCP is Google’s strong commitment to AI and ML. “The models that used to take weeks to train on GPU or any other hardware can put out in hours with TPU. AWS and Azure do have AI services, but to date, AWS and Azure have nothing to match the performance of the Google TPU,” said Jeevan Pandey, CTO, TelioLabs. Benefits of TPU: Can be leveraged by Google Cloud Machine Learning Engine to run complex ML modelsIncreased performance of linear algebra computationReduced time-to-accuracy when training complex neural network modelsScalable across different machines Vertex AI Vertex AI brings together the Google Cloud services for building ML under one, unified UI and API. It has pre-trained APIs for vision, video, natural language, etc. Vertex AI integrates with widely used open-source frameworks such as TensorFlow, PyTorch, and scikit-learn, along with supporting all ML frameworks via custom containers for training and prediction. Benefits of Vertex AI: Easily train and compare models using AutoML Faster movement of models from experimentation to productionAccess to AI toolkit used internally to power Google, including computer vision, language and conversation data.Smoother end-to-end ML workflow that removes the complexity of self-service model maintenance and repeatability with MLOps tools Open-source Google cloud’s open-source contributions, especially in tools like Kubernetes –a portable, extensible, open-source platform for managing containerized workloads and services, facilitating declarative configuration and automation– have worked to their advantage. “Kubernetes helps in the AI and ML workflow as it supports and leverages the speed of today’s cloud GPUs,” said Chittora. Benefits of Kubernetes: Works on any container runtime or any infrastructure, including public cloud, private cloud and on-premises serverCan host workloads on a single cloud or across many clouds.100% open-source project offering more flexibility Speech and translate APIs Google cloud’s speech and translate APIs are much more widely used than their counterparts. According to Gartner’s 2021 Magic Quadrant, Google cloud has been named the leader for Cloud AI services. Pre-trained ML models can be instantly used to classify objects in an image into millions of predefined categories. Additionally, one of the top ML services from Google cloud is Vision AI, powered by AutoML. Benefits of speech and translate APIs: Can translate many languagesCan detect source text languageSpeech API recognises more than 80 languagesAffordable pricingProducts are highly scalable, easy-to-use and accurate AutoML AutoML enables developers with limited machine learning expertise to build custom ML models in minutes. It is a cloud-based ML platform and uses a No-Code approach with a set of prebuilt models via a set of APIs. It is tightly integrated with all Google’s services and stores data in the cloud. Trained models can be deployed via the REST API interface. “Data Analysts can combine a custom model and pretrained models in a single product, and this feature keeps GCP above its competitors,” said Saket Saurabh, Head, Cloud and Analytics at STEMROBO. Benefits of AutoML: A no-code way to build modelsCustomers can train their own neural networksCan apply data and integrate predictions whenever you needNot limiting like single type of ML model offered by others Other benefits Abishek Chiffon, Data Scientist at Ideas2IT, outlined ML and AI services that put GCP ahead of AWS and Azure: Offers AI Hub, a repository of models in a format that can be deployed in Kubeflow, Deep Learning VMs in GPU or TPU.AI Platform Classic tool helps create ML training jobs with TensorFlow, Keras, PyTorch, Scikit-learn, and XGBoost and allows training at scale with user containers and custom frameworks. Google has TensorFlow to build and launch self-contained Deep Learning models.Its AI Platform Notebooks are VM instances that come pre-integrated with TensorFlow and PyTorch instances, Deep Learning packages, and Jupyter notebook.Deep Learning VM images come pre-installed with all software, deep learning and ML frameworks on a Google Compute Engine instance. “AWS and Azure have different tools for different services, and also the training algorithm of GCP has matured over time through its Search and AI platforms used worldwide. GCP is turning out to be a popular choice for experienced and new data scientists and is ready to trump AWS and Azure soon,” said Saurabh of STEMROBO.","excerpt":"Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft Azure are the leading cloud providers by a long shot. Though late to the party, GCP has seen robust growth over the years. Google Cloud’s revenue jumped nearly 46 percent year-on-year to $4.04 billion in the first quarter of 2021. “All vendors offer strong ML […]","categories":["IT Services"],"tags":["AWS","Microsoft Azure"],"author_name":"Shanthi S","publish_date":"2021-05-31T10:00:00","publication_year":"2021","word_count":795,"keywords":["machine learning","AWS","AI","neural network","ML","MLOps","computer vision","Kubeflow","deep learning","analytics","Microsoft Azure","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","analytics","MLOps","Kubeflow","TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-ml-capabilities-of-gcp-is-way-ahead-of-aws-azure\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":1679,"title":"Startups and companies race to make driverless cars a reality on Indian roads","content":"Not wishing to miss the boat of putting driverless cars on roads by 2025, Indian companies, startups and IITs have joined the race of building the next autonomous vehicles technology in India. Leading the Indian charge is Tata Elxsi, Tata’s leading design and technology services arm for product engineering across automotive, communications and broadcast. Tata had earlier gained significant ground by debuting its Autonomous Valet Parking prototype that demonstrated the fusion of sensor and vision based ADAS technology. Now, reportedly, Tata Elxsi has made a big leap by testing driverless cars on Bangalore’s roads. According to news reports, Tata in collaboration with its rich ecosystem of component suppliers is leveraging AI to develop driverless cars and has stolen a march over its Indian competitors (Mahindra, notably) by being the first company to test driverless vehicles on Indian roads. Driverless cars – the future of mobility According to Chetan Maini, founder of Reva, India’s first electric car and director at Maini Group, “The DNA of the auto industry is changing. Vehicles are getting smarter, more connected and possibly way more virtual. We are trying to rethink where mobility is going to go” Maini believes the challenge in India is more unique. “Can we have that’s driverless and work for India, in Indian conditions, imagine the solution for tens of millions differently able people today in the country who can use driverless mobility to get them around and can the solution be far more affordable?” he said. Maini played out a simple scenario that could become a reality in near future:  Imagine pressing a button on your mobile phone, your car comes up to you, you don’t have to think about parking, you get in it, and suddenly someone steps in front of you and your car’s braking system takes over and stops before you can even react. You drive to your destination, get out of it, press another button, and it goes and finds its parking spot, so that’s the innovative solution we are looking for as part of Mahindra’s Rise challenge. Mahindra’s driverless car challenge sees 13 Indian startups including IITs rise to the occasion Mahindra is betting big on the future of mobility and believes autonomous cars is the next evolution for the automotive industry. The paradigm shift will be seen in increased environmental friendliness, overcoming parking problems, decreased mortality rate and a differently-abled friendly transportation. Self-driving cars are outfitted with radar, GPS and computer vision and most importantly LIDAR or the light radar, the sensing technology that helps driverless vehicles in precise localization.  In fact, LIDAR sensors could significantly bring down the cost of autonomous vehicles and make it affordable for average buyers. Mahindra’s Rise prize Driverless Car challenge with a purse of $1 million has seen Pune-based Aeron Systems Pvt Ltd, maker of IoT devices and inertial sensors and systems, AGV IIT Kharagpur, Auro Robotics, drawing members from IIT Kharagpur and CMU Robotics Institute, IIT Bombay, d-LIVE the official team from IIT Delhi, Team Manas from Manipal Institute of Technology, Pragyaan backed by DRDO make the final cut to top 13 for the prototyping phase. Amongst the startups, Delhi-based AI start-up Cube26, known for developing customized OS platform for OEMs and OmnipresentRobot, one of leading robotics and drone makers in India that leverages computer vision, machine learning and virtual reality are now taking a crack at driverless vehicle technology. IoTIndiaMag lists down some notable projects underway OmnipresentRobot: According to news reports, Akash Sinha led startup that also has CMU’s Dr Raj Reddy, eminent computer scientist and robotics experts is well positioned to ace the challenge given Sinha’s participation in three 3 DARPA Robotics Grand challenges. Taking cues from DARPA, the startup’s modus operandi is using a combination of ID Laser sensors, stereo cameras and combining their data through sensor fusion to generate 3D maps. There are also plans to deploy monocular camera for detecting traffic lights and decoding road signs. Deep learning techniques will be leveraged to enhance vehicle performance and train it further to understand barricades, dividers and even Indian hand gestures used during driving. Finally, there will be an app to follow instructions such as parking, go to address location. What the startup is also striving hard is achieving precision localization through ultra wide band tags along with GPS\/INS. Auro Robotics: Auro robotics backed by IIT-Kharagpur alumni have already made significant headway with their autonomous shuttle currently playing in the University of Santa Clara with multiple campus trials ongoing. The driverless electric shuttles address the “last-mile” challenge in public transit systems, says Nalin Gupta, CEO of Auro Robotics. Their electric shuttle, relies heavily on LIDAR, sensor technology that uses laser light for localization and spotting anomalies. SeDriCa 1.0 from IIT Bombay: SeDriCa is the famed Autonomous Ground Vehicle, cranked out by IIT Bombay’s Innovation Cell that nailed the 4th spot globally in the Intelligent Ground Vehicle Competition (IGVC) 2016, besting teams from CMU, Virginia Tech. and 30 other universities. Now the team is applying their learnings to Mahindra E20. The task at hand – autonomously maneuvering Mahindra’s electric car on Indian roads. From pedestrian detection to positioning system, the team has a similar approach, combining sensor data from leveraging AI to perfect autonomous vehicles. According to their home page, the team’s looking at GPS\/INS, LiDAR and stereo cameras to gather information about the immediate environment at the right range.  The 12-member team is led by Shubham Jain, a 4th year Mechanical Engineering student who also wears the hat of Project Manager and Deputy Team Leader. Team Manas from Manipal University: Combining AI, automation sensing and mechanical engineering, a team of 45 undergraduates from Manipal University are experimenting with lane detection, maze traversal, for navigation and parking and much more. From performing real world 2D LIDAR obstacle detection to carrying out dynamic obstacle avoidance to sharpening localization, Project Manas team is at hard at work to best their competitors. Google, Uber proclaim self-driving cars distant reality in India Driverless cars may not be favorable for Indian roads, what with Google’s CEO Sundar Pichai and Uber’s founder and CEO Travis Kalanick deeming Indian roads, the last place for self-driving cars.  In India, a lack of lane discipline coupled by traffic conditions and a large number of two-wheelers on roads present certain challenges. The traffic on Indian roads cannot be easily mapped by autonomous vehicles. The present condition of Indian roads – potholes, poorly developed roads, obstructions and traffic diversions present major challenges to autonomous vehicle technology. Experts proclaim that driverless vehicle technology is bound to become ubiquitous in developed nations by 2025 with legacy automotive companies Tesla, Apple, Baidu, Tesla, Ford, Japan’s Nissan ,Honda, Mitsubishi and Toyota and Germany’s Audi and BMW, putting their might behind self-driving cars. From cars to pods – ushering in a revolution Interestingly, California, the hotbed of driverless technology where it has become commonplace to spot autonomous vehicles plying on the road, could very well become the first state to allow driverless vehicles on empty roads. And the future of cars will be a pod, possibly with a camera mounted on top to collect all the data. America’s biggest car market is proposing new regulations wherein road ready vehicles could be tested without drivers could start by the end of 2017. If the regulations are passed, a limited number of cars will be in the market as early as 2018. Google’s rechristened self-driving unit, Waymo is an everyday sight in Silicon Valley and has had successful real world tests.  Interestingly, Google and Uber are battling out over driverless technology, with the former alleging Uber stole its technology to get ahead in the race.","excerpt":"Not wishing to miss the boat of putting driverless cars on roads by 2025, Indian companies, startups and IITs have joined the race of building the next autonomous vehicles technology in India. Leading the Indian charge is Tata Elxsi, Tata’s leading design and technology services arm for product engineering across automotive, communications and broadcast. Tata […]","categories":["AI Startups"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-03-22T11:01:40","publication_year":"2017","word_count":1271,"keywords":["Go","machine learning","AI","RPA","computer vision","RAG","automation","Aim","deep learning","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","Aim","RAG","R","Go","automation","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/startups-companies-race-make-driverless-cars-reality-indian-roads\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":19045,"title":"How Are Technologies Like Intelligent Logistic And Drones Taking E-commerce A Step Ahead","content":"We have seen in the past how artificially intelligent programs are being used by e-commerce players for image\/video search, diving deep into understanding customer preferences or using chatbots and virtual personal assistants for enhancing customer-brand interaction, but this time around it’s a different application of advanced tech that we are talking about. Its drones and intelligent logistics, that have become talk of the town. Gone would be the days of manned deliveries stepping onto your doors, as e-commerce giants are looking forward to adopting technologies like drones to deliver your parcels! It wasn’t long back when the e-commerce giant Amazon had filed a patent application to introduce drone technology with respect to propeller safety in India, and later came up with another patent application for exclusive rights on multi-scale fiducials. With this, the company aims to get the rights for black and white marks on any object for self operating aerial vehicles, so that it can be identified from different distances. Not just drones, but to cope up with the high demands during festive season and big sale days, the e-commerce players are also adopting deep technology and intelligent logistics solutions for quicker and efficient sorting of packages at warehouses, enabling better address matching, route optimization and much more. In its first ever, Alibaba delivers packages via drones Post package Alibaba, a Chinese e-commerce conglomerate, recently claimed to have used drones to deliver packages over water for the first time. It said that three unmanned aerial vehicles carrying six boxes of passionfruit flew from Putian in China’s eastern Fujian Province to nearby Meizhou Island. The boxes’ combined weight was 12 kg. The drones covered the distance of five km in nine minutes, flying into a strong wind. The drones were jointly developed by Alibaba’s delivery arm Cainiao Network, company’s rural shopping platform Rural Taobao, and a domestic technology firm. Other ecommerce players from across the globe are also integrating robots and drones to solve the bottlenecks that may arise due to high orders. Adopting automation technology has become the next big thing and very recently China’s second biggest ecommerce player JD.com also launched the first unmanned sorting center, with an aim to develop intelligent and high-efficient logistic systems. India pacing up with the use of drones and intelligent logistics These developments are suggestive of the fact that, while technologies like AI, chatbots, machine learning, drones and others are seeping across various sectors, its adoption has been equally swift in the Indian ecommerce industry. These advanced techs are becoming a way out for increasing the overall efficiency, cost saving and most importantly customer satisfaction. We list down five major developments by India in the area of drones and intelligent logistics in ecommerce. 1| Indian Government’s announcing that drones for ecommerce may soon be a reality: As Amazon filed patent for the use of drones in the country, Indian Civil Aviation Ministry announced earlier this month that it would soon be a reality to witness the use of drones and unmanned aerial systems for commercial purposes such as the delivery of ecommerce packages at customer’s doorstep. With the issuing of draft regulation, it wouldn’t be unlikely to see drones operating for ecommerce and other sectors by 2018. Jayant Sinha, Minister of State for Civil Aviation was reportedly quoted as saying that e-commerce deliveries using drones are certainly going to be possible in India, and that companies like Amazon and Flipkart would be able to deliver products with the technological developments in the aviation ecosystem. 2| Amazon introducing amazon Prime Air like services in India: Amazon Prime Air, a delivery system designed to deliver packages to the customers within 30 minutes or earlier using drones has been quite popular in the US. Apart from conducting trials in the UK to gather data to improve safety and reliability of these systems, the ecommerce giant is keen on bringing the technology to India. Apart from its patent application for drones, the company is also believed to have made fresh investment in its logistics unit Amazon Transportation Services in India. This is line with their interest to drone deliveries in India, which may soon be a reality given the announcement by Directorate General of Civil Aviation. 3| Flipkart’s ambitions of introducing drones in rural areas may now come true: Two years’ back India’s leading e-tailer Flipkart had wanted to use drones to deliver goods to rural areas. “We are looking at drones for rural deliveries,” the company had said then. With the policies being framed in place and competitors gearing up to introduce drones, it might not be unusual to see Flipkart investing in the technology as well. 4| Intelligent logistics solution by Locus is striving a way: Locus, a 2015 founded company has been optimising logistics operations to provide consistency, efficiency & transparency, and delivering transformational value in logistics for organisation. Its proprietary algorithms have been solving complex problems such as load balancing, route optimisation and container utilisation. It is currently offering automated logistics for leading enterprises in India such as Urban Ladder, 1mg, Quikr, Lenskart, Licious, and others. Nishith Rastogi, CEO & Co-founder, Locus shared “With increased competition, the key differentiator for e-commerce business to survive is by providing exceptional customer experience and reducing costs at scale. Reducing human dependencies through automating critical business decisions, effectively managing operations, reducing costs and taking deliveries to the next level is possible by adopting Drone Deliveries, Intelligent Routing Engines and imbibing Machine Learning & Artificial Intelligence through them. In today’s age where the mantra is to ‘evolve or be disrupted’, companies will have to leverage technology to reinvent & scale their businesses.” 5| Other solutions like Delhivery taking a lead in offering intelligent solutions: Logistic firms like Delhivery and Mera Transport have been using smart technologies, data science and analytics to manage their logistics, track and analyse their shipments end to end. Delhivery’s intelligent solutions have been helping ecommerce companies by tracking fake products and keeping a strong check at it. They have developed a software solution that runs image recognition checks and analyses each shipment using AI to spot fake products. On a concluding note With an increasing interest to use drones and intelligent logistics to take e-commerce a step ahead, it wouldn’t be longer to see home grown ecommerce companies like Flipkart using drones to deliver goods. It could be particularly useful in rural areas, and in cases of faster deliveries, where packages have to be delivered to customers within 30 minutes or earlier. While intelligent logistics have found a way in India, we are yet to see drones making a headway soon.","excerpt":"We have seen in the past how artificially intelligent programs are being used by e-commerce players for image\/video search, diving deep into understanding customer preferences or using chatbots and virtual personal assistants for enhancing customer-brand interaction, but this time around it’s a different application of advanced tech that we are talking about. Its drones and […]","categories":["IT Services"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-11-16T06:20:18","publication_year":"2017","word_count":1098,"keywords":["data science","machine learning","artificial intelligence","AI","chatbots","image recognition","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","RAG","chatbots","image recognition","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/technologies-like-intelligent-logistic-drones-taking-e-commerce-step-ahead\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10071504,"title":"Bonsai Brain &#8211; A low code platform to build AI agents","content":"Bonsai Brain is one of the ongoing projects of Microsoft, which aims to develop a low code AI-based component that is integrated with Automation systems. The Bonsai brain is simulated and trained in a manner to handle situations and to be fault tolerant even during unexpected or unseen circumstances. Bonsai’s brain focuses on adding value to various autonomous systems, processes, and equipment but also focuses on growing customer trust by ensuring continuous operations. In this article, let us try to understand the Bonsai Brain with respect to this context. Table of Contents Introduction to Bonsai BrainComponents of the Bonsai PlatformAdvantages of Bonsai brainSimulating a Bonsai BrainSummary Introduction to Bonsai Brain Bonsai Brain is an ongoing research project of Microsoft that focuses on simulating and developing a low code-based AI component that can be used for various Autonomous tasks and applications. Bonsai’s brain is simulated and trained in a manner to handle unexpected and unseen circumstances and to ensure continuous operations. Downtime is greatly reduced by using the Bonsai brain along with an increase in production efficiency. Are you looking for a complete repository of Python libraries used in data science, check out here. Automation tasks basically involve larger neural networks to be designed, but Bonsai’s brain operates without any neural networks itself simulated or trained. The interface of the Bonsai brain lets users create their custom AI models and put them into operation accordingly without the requirement of additional resources. The below image shows the Bonsai brain platform and the operating cycle of the Bonsai Brain. Let us try to understand the Bonsai Brain with respect to that context. The Bonsai brain platform basically uses deep reinforcement learning principles to simulate and train the Bonsai brain. The Bonsai brain platform ensures to simulate and train the Brain for all possible or unseen circumstances and ensures smarter Autonomous systems are produced. The entire Bonsai brain platform operates on three standard operating principles. Let us look into it and try to understand the importance of these in the Bonsai brain platform. i) Integrate component in the platform is responsible for integrating training simulations to the Bonsai brain with real-world scenarios and providing feedback accordingly to the training process. The Bonsai brain in this stage is aimed to be simulated for all known possible circumstances and real-world circumstances so that the simulated Bonsai brain remains robust for unseen changes. ii) Train component in the platform is responsible for training and simulating the brain according to the feedback received from the Integrate component. The Bonsai brain is trained for real-time learning objectives and is an iterative process that learns and gets simulated according to the learning objectives formed by the Simulator in the Integrate component. iii) Export component in the platform consists of a completely trained and simulated Bonsai Brain that will be available as a Linux container that can be deployed in the Azure environment or in the premises. The Bonsai brain is simulated and trained to ensure the replication of real-world systems. This makes the Bonsai brain get trained in real-world or authentic training environments. The Bonsai brain basically uses simulations and Deep Reinforcement learning principles to Train the Brain in the platform. But there are two check conditions that have to be fulfilled to train the Bonsai brain in the platform. They are as follows. i) The precision for each of the actions that are simulated to the Bonsai Brain has to be accurate to ensure that the Bonsai brain is robust and is operating as expected. ii) Probability of recovering from a wrong action performed by the brain has to be fast or high to ensure continuous operations as the Brain would be integrated into major autonomous systems. This also ensures the Brain is simulated and trained to remain robust even for uncertain actions that the brain would take up in the real-world environment. As we have got an introduction to the Bonsai Brain, let us try to understand the critical components of the Bonsai Brain in the next section of the article. Components of the Bonsai Platform The entire Bonsai brain operates on 5 critical components. In this section let us look into the critical components of the Bonsai brain and try to understand the functionality of each of the components in the Bonsai brain. i) Brain is the agent in the Bonsai platform. The Brain in the platform will be simulated and trained to meet the desired goals set. In the Bonsai platform, the Brain as an agent will be available in the form of a text (txt) file named Inkling file. ii) Simulator in the Bonsai platform is the component responsible for simulating the brain to learn from different instances. The input to the simulator will be observations and the output of the simulator will be the different sets of actions that will be performed by the Bonsai Brain in the Bonsai platform. iii) Workspace is one of the components of the Bonsai platform that holds all the Brains and the simulators created in the platform. So the workspace is the component of the Bonsai platform with huge collections of Bonsai brains. Brains from this component can be pulled to be the simulator component and can be trained accordingly. The Workspace platform can also be used to monitor the training of the Brain in the platform. iv) Iteration is one of the components of the Bonsai platform where for each set of simulations the brain is trained to produce certain action. So each action by the brain in the platform is termed an iteration. The Iteration in the platform is entirely dependent on the training instances and the observations it is simulated for. For complicated systems, the iterations in the Bonsai platform may be very huge. v) Episode is one of the components of the Bonsai platform that is used to set a threshold for the iterations in the Bonsai platform. The episode in the Bonsai platform will operate based on a built-in Episode Iteration Limit and when the mentioned episode in the platform is reached, the simulator will be reset in the Bonsai platform. Advantages of Bonsai brain The Bonsai brain finds its major advantages in AI use cases. Let us summarize the major advantages of using the Bonsai brain for AI applications. Ability to simulate and train the Brain with respect to the industry requirements and domain expertise which ensures the Brain simulated remains robust.Ability to provide improved control methods and improved automation systems as they will be simulated for the right set of actions in the Bonsai platform.The Bonsai brain can be simulated to quickly adapt to immediate production changeovers if any as per requirements.In the Workspace, Brains can be developed that will remain robust for unseen circumstances and pulled accordingly to simulate it. This ensures continuous operations of Brains integrated or deployed on the premise or in the cloud. Now let us look at how to simulate a Bonsai brain for an AI task. Simulating a Bonsai Brain As mentioned in this article Bonsai’s brain basically uses Deep Reinforcement Learning principles to simulate Bonsai’s brain in the Bonsai platform. The Bonsai platform ensures the brain is simulated and trained for all random or unexpected circumstances. Let us understand through a case study how to simulate and train a Bonsai brain to balance a pole with AI. The main task in this case study is to balance a pole that is standing on a moving base. But remember to understand this case study better we need to have a Microsoft Azure account. In this article, the steps to be followed will be listed in detail that can be implemented to use this case study in the Microsoft Azure platform. Step-1:  Selecting the simulator from Workspace Bonsai brain is one of the projects by Microsoft which is still in the development stage. So for balancing the pole task we have to select the cartpole simulator from the Workspace. The cartpole has to be selected by first signing into the Bonsai User Interface. After selecting the cartpole from the list of brains from the Bonsai Workspace we will have to give a random name to the cartpole brain selected. After naming the cartpole brain we should click on create Brain to load the brain and the simulator in the Bonsai platform. In the Bonsai brain user interface the cartpole brain would appear as shown in the above image. Step-2: Validating prebuilt training code Once the cartpole is selected a prebuilt training code will be available in a text file known as Inkling. Two panels can be accessed where in one panel we can see the code for the cartpole and in the other panel we can see a graph-based panel for the cartpole. In the graph panel, there are three main nodes named State node, Concept node, and Action node. So if the nodes are selected in the graph panel the respective code in the Inkling file gets highlighted. Step-3: Training the Bonsai brain In the Bonsai User interface, we will have to navigate to the “Train” tab and click on the green button to start the brain training process on the Bonsai platform. Once the training process is instantiated in the Bonsai platform the user interface will replace the coding panel with a graph showing the training process of the Bonsai brain in the platform. A data panel will appear in the user interface where a graph will be generated for the different training iterations and the action or the goal achieved in percentage. So for each iteration, a performance score is obtained and the bonsai brain training process can be reported using the Goal Satisfaction plot. Step-4: Visualizing the actions in the platform Once the training process of the Bonsai brain in the Bonsai platform is completed, the brain can be visualized in the Bonsai platform for its performance in balancing the pole. Here a 3D simulator is visualized that would show how a pole in real time would be balanced on a cartpole for the instances it is trained upon. The cart would move in left and right directions and the pole would be balanced according to the learning simulated to it according to the deep reinforcement learning principles. Step-5: Terminate the training process Bonsai brain has the ability to automatically terminate the training process if the overall goal satisfaction factor reaches 100%. So if 100% goal satisfaction is obtained it means that the brain is simulated in the Bonsai platform for all random and uncertain circumstances and the brain would remain robust when taken up for deployment in AI interfaces. The training process can also be interrupted in the Bonsai user interface by clicking on the Stop training button and the actions of the brain can also be visualized accordingly for the trained number of instances. Summary Bonsai Brain is one of Microsoft’s projects which aims to reduce redundant and huge codes and deploy efficient and robust AI models. Bonsai’s brain basically uses deep reinforcement learning principles to build effective AI models, and by using the Bonsai platform robust AI models can be simulated and developed by following a few steps. We also saw that we could develop a pole balancing AI model with a few steps using the Bonsai brain platform. Bonsai’s brain, if developed, would be a vital part of various automation systems and would be found integrated into various AI models as well. References Bonsai Brain official documentation","excerpt":"The Bonsai Brain is a low code AI component that is integrated with Automation systems. The Bonsai Brain focuses on adding value to various Autonomous and AI systems.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Automation","Deep Reinforcement Learning","Microsoft Azure"],"author_name":"Darshan M","publish_date":"2022-07-26T10:00:00","publication_year":"2022","word_count":1909,"keywords":["data science","Go","TPU","Rust","AI","neural network","R","Automation","Python","Deep Reinforcement Learning","Aim","Microsoft Azure","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","neural network","data science","Aim","Azure","TPU","Python","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/bonsai-brain-in-azure-platform\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10049911,"title":"Conversation With Erad Fridman, CEO at Fluxon","content":"Founded in 2017, product development company Fluxon is the brainchild of former Google and other tech companies’ leaders. The San Francisco-based company helps high-growth startups and companies including the likes of Google, Zapier, and Stripe build innovative products by providing advice and strong technical leadership. Additionally, it also builds its own products — Dory, a real-time Q&A application for events. With offices in the US, Ukraine and Canada, Fluxon started operations in India in 2019. In a conversation with Analytics India Magazine, CEO Erad Fridman talks about the technology behind Fluxon and the company’s plans for India. Erad is a product leader with more than 15 years of experience in developing platforms, UX and design. Starting coding at the age of 6 years, Erad had earned both his Computer Science and Mathematics degrees and led a software engineering team in the Israeli Prime Minister’s office by 19. He has a dual Masters Degrees in Finance and Business Administration at Boston College. Before Fluxon, he led a team of product managers, engineers, and designers at Google and was the VP of Product at Planet.com. As CEO at Fluxon, Erad sets and executes the company strategy and vision, ensuring the right decisions are made to support that strategy. He takes care of hiring individuals and pairing them with interesting problems for them to work on. Edited excerpts from the conversation: AIM: What products have you helped build in Fluxon? Erad Fridman: Together with our CPO, Jen Gil, we stay involved in all our product work. I’m especially excited by projects where we are the primary development team – for example, Pluto — an online in-browser spatial 3D experience for meetups and video chats, and Eraser — a virtual meeting canvas built for remote team collaboration. I also stay closely involved in the development of Fluxon-owned products such as Dory – a real-time Q&A app for events. AIM: Explain the tech stack at Fluxon. Erad Fridman: At Fluxon, we make modern, pragmatic technology choices. We build fast, and a big part of what enables that is our technology choices. We favour managed services because they free the developer team from operational concerns to focus more on the product. We rely on Cloud platforms and heavily leverage services such as Google’s Cloud Run and Cloud PubSub to build scalable yet easy to operate applications. AIM: What tech tools do you use at Fluxon? Erad Fridman: We enjoy working with modern frontend frameworks such as React or Angular. For mobile applications, we’re typically working with cross-platform frameworks such as React Native, dropping into the native layer when needed for specific use cases. We work with a variety of database technologies, including Firebase, PostgreSQL, MySQL, ElasticSearch, and others. AIM: What makes Fluxon expand to India? Why do you think this is the right time to expand your footprint in the Indian market? Erad Fridman: We started operations in India in 2019 and were impressed by its strong technical talent, great schools and growing startup culture. AIM: Who is your target audience in the Indian market? Have you acquired any clients in the Indian market yet? Erad Fridman: We work with high-growth startups and enterprise companies all over the world. Many of Fluxon’s current clients are based in Silicon Valley, but there are lots of impactful business opportunities for us to work on in India. We are already working on exciting projects with Indian companies such as iMerit, an AI training startup. AIM: How do you differentiate from your competitors? Erad Fridman: Fluxon’s differentiator is our team. We attract and grow the best talent to deliver products with the highest quality and speed. Amongst startups and VCs, we’re well-known as a top development firm. We take ownership and approach every project with a founder’s mentality. Our clear, open and frequent communication style is paramount to driving success. AIM: What are your strategies to make it big in the Indian market? Erad Fridman: A clear vision is key to our success in India. We select interesting work that keeps our teams challenged and continuously learning new technologies, and our clients leverage our team’s expertise to build inspiring products. AIM: Are you looking to hire tech talent here? What are the key skills that you are betting on? Erad Fridman: We see India as a critical market for taking Fluxon to our next stage of growth. Our goal is to add 1,000 employees to our team in India over the next few years. We’re primarily hiring technical folks in functions across engineering, product and design and seek out highly motivated individuals that have the intellectual flexibility and growth mindset needed to collaborate across domains. AIM: What are the few things that India’s tech talent can learn from their western counterparts and vice-versa? Erad Fridman: We all can learn from each other. We hire humble and curious people who are passionate about their work and support each other. We encourage continuous learning through both our client work and education benefits. We work hard to ensure all projects are well-documented and best practices are shared back with the wider team. AIM: How do you think companies can address the shortage of tech talent in India and globally? Erad Fridman: There are a few strategies we have seen an increase over the past year, including bulk hiring, acqui-hiring, market expansion and improving benefit and perk packages. All of these have benefits and downsides – there is really no magic bullet. At Fluxon, our strategy has always been to hire the best technical talent there is, which we believe attracts more great people and exciting projects for them to work on. AIM: What does the road ahead look like for Fluxon in the Indian market? Erad Fridman: Over the next few years, we will continue expanding our team and client work in India. We’re soon expanding into a new office space in Hyderabad, which will accommodate more than 100 team members. Next year, we plan to add a second location in Bangalore as our team grows further.","excerpt":"San Francisco-based Fluxon provides high-growth companies and startups advice and technical leadership.","categories":["AI Features"],"tags":[],"author_name":"Debolina Biswas","publish_date":"2021-09-28T16:00:00","publication_year":"2021","word_count":1004,"keywords":["PostgreSQL","Elasticsearch","Go","AI","Scala","RAG","Aim","analytics","SQL","R"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","Elasticsearch","PostgreSQL","R","SQL","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/conversation-with-erad-fridman-ceo-at-fluxon\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":7304,"title":"Big Data: Challenges and roadblocks","content":"Big data is the big buzzword these days. Big data refers to a collection of data sets or information too large and complex to be processed by standard tools. It is the art and science of combining enterprise data, social data and machine data to derive new insights, which are otherwise, not possible. It is also about combining past data with real time data to predict or suggest the outcomes for the current or future context. The digital footprint, is progressively expanding, world over, into fragmented mediums (blogs, tweets, reviews etc.) and technologies (mobile, web, cloud\/SaaS etc.). Digital landscape in India India’s digital landscape too, maybe evolving quickly but overall penetration remains low, with only 1 in 5 Indians using the Internet in July 2014. In India enterprises and businesses have access to a veritable wealth of information. And though some of the larger organisations have made a start in harnessing the information, most Indian companies are still learning how to collect and store big data. Telecom providers, online travel agencies, online retail stores are some of the industries that are using big data analytics to engage customers is some ways. However, big data analytics is still its infancy in India. Most companies are learning to store the data collected. Then, there are several challenges when it comes to the collection of data sets themselves. Past and current data is required to make the application of big data analytics really useful, there is a scarcity of past data in public and private sectors in India. Some of the reasons for the lack of enough data are: Yet to be fully computerised Healthcare, economic and statistical data, in both private and public sectors in India is yet to be computerised. The main reason for this is the late adoption of IT in India. Unlike in the West, most industries in India made the transition from manual records to computerised information systems, only during the last decade. Over the years, the state and central ministries have made the move towards e-governance.  Efforts to deliver public services and to make access to these services easier are being made as well. While this is still a work in progress, huge amounts of data across many government sectors are yet to be digitised. Quality of data In big data analytics, data sufficiency plays a critical role when samples are run across different dimensions. Sufficient data points to perform analytics with the samples are required. Not only quantity of data, the quality of data being used for crunching, also influences the quality of insights.  If the signal-to-noise-ratio is high, the accuracy of results may vary for less than optimum data samples. In a country like India, there is very little information about the individuals, due to the fact that Indians are not overly expressive, especially on public forums. Public social media information that is available for most individuals from India lacks quality information about users themselves. Random facts and figures in individual profiles, sharing of spam content, and fake social media accounts that are created for bots are very common in India. Spam Social media sites are becoming increasingly vulnerable to spam attacks. Time spent by a captive audience on social media sites opens up windows of opportunities,  for online threats and spammers. Again, social media spam contributes to the signal-to-noise-ratio that defines the quality of big data. This hinders the appropriateness of results. Cultural and Social influences In most western markets, insights generated through big data can be applied across the whole consumer base. However, given the extensive cultural and linguistic variation across India, any insight generated for a consumer based out of Chandigarh, for example, will not be directly applicable to a consumer based in Chennai. This problem is made worse, by the fact that a lot of local data lives in regional publications, in different languages and has very limited online visibility. Unstructured data leads to mapping issues Big data in India is not structured. Most transactional data in the healthcare and retail segments are stored purely for book keeping purposes. They have very limited appropriate information that can help big data analytics map enterprise generated transactional data, with public information. In the case of developed countries, user data is rich enough to provide demographic or group level markers that can be used to generate customized insights while maintaining individual privacy. Lack of these standard identifiers in Indian consumer data is one of the biggest bottle necks, while mapping various transactional and social records in India. Handsets and internet connectivity Even though smart phones are driving the new handset market in India, feature phones still dominate everyday usage. Most connections in India are pre-paid and fewer than 10% of users have access to 3G networks. To add to it, internet connection speeds are amongst the lowest in Asia. As a result, consumer data, especially retail enterprise data is limited. As more people in India make the move to smart phones, and internet connectivity improves, there will be an increase in the amount of usable data generated. As Big data analytics may be at its infancy in India today ,huge efforts would need to be made to improve the quality of data by organisations and enterprises. However, key contributors to the promise of big data analytics in India are steadily gaining ground. An increase in social media users, efforts by enterprises, both public and private for optimum collection and storage of transactional enterprise data, will contribute to better quality data sets for the better application of big data analytics.","excerpt":"Big data is the big buzzword these days. Big data refers to a collection of data sets or information too large and complex to be processed by standard tools. It is the art and science of combining enterprise data, social data and machine data to derive new insights, which are otherwise, not possible. It is […]","categories":["IT Services"],"tags":["big data india","crayon data"],"author_name":"Srikant Sastri","publish_date":"2015-04-21T11:54:38","publication_year":"2015","word_count":925,"keywords":["big data","Go","big data india","crayon data","programming_languages:R","AI","Git","RAG","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Git","big data","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/big-data-challenges-and-roadblocks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":23365,"title":"Google’s DeepMind AI Takes An Aggressive Nature. Is That A Reason To Worry?","content":"Self-preservation is an instinct that every human being has. Now imagine this, an AI that cannot process any emotions yet have the same instincts as humans and thinks to preserve itself over others! That is what researchers from Google-owned DeepMind firm study have found out in this major advancement by their artificial intelligence. Google acquired DeepMind is on a scientific mission to push the boundaries of AI. It has been developing programs that can learn to solve any complex problem without any human intervention. DeepMind researchers recently ran a series of tests to find out how AI would react when faced with certain social dilemmas. The basic idea was to find out whether they would cooperate or compete. For the research, they tested the AI on two games: a fruit-gathering game and a hunting game called Wolfpack. These are rudimentary two-dimensional basic games that used AI characters known as agents. By using game theory, researchers found out that it is possible for artificial intelligence to act in an ‘aggressive manner’ when it feels that it is going to lose out. However, the AI agents will work together as a team when there is more benefit. AI’s Killer Instinct In the first fruit-gathering computer game, researcher tasked the two AI agents (red and blue)  to gather as many virtual apples (green pixels) as possible. In the gathering game, the systems were trained using deep reinforcement learning to collect apples. When an AI agent collected an apple, it was rewarded with a ‘1’ and the apple vanished from the game’s map. Both AI agents had the option to tag their opponent two times in a row with a laser blast that would temporarily remove them from the game thus giving the attacker more time to collect the apple. Through this test, researchers found out that the two versions of AI behave differently in accordance to the number of apples available. As long as there was enough apple to go around for both AI agents, there wasn’t any problem. But as soon as the apples became more scarce, the AI agents were more likely to knock each other out of the game to get ahead to collect the scarce apples for itself. After running 40 million test, the researchers concluded that AI agents became ‘highly aggressive’ when there were scarcity of resources and it led to more ‘tagging’ behaviour of agents, which means agents with the capacity to implement more complex strategies try to tag the other agent more frequently, i.e. behave less cooperatively – no matter how we vary the scarcity of apples. “Intuitively, a defecting policy in this game is one that is aggressive—i.e., involving frequent attempts to tag rival players to remove them from the game. Such a policy is motivated by the opportunity to take all the apples for oneself that arises after eliminating the other player,” the researchers write in their paper. The paper also stated that “Less aggressive policies emerge from learning in relatively abundant environments with less possibility for costly action. The greed motivation reflects the temptation to take out a rival and collect all the apples oneself.” Watch the fruit-gathering game below: AI’s Instinct With WolfPack In the second game called Wolfpack, the researchers included three AI. Two of them are wolves (Red) and the third AI is prey (blue). Here the rules were different, this game required close coordination. When the prey was hunted down by either of the two wolves, they both received a reward. Greater rewards were offered when the wolves were in close proximity during a capture. The researchers explained in their paper that the core idea is that, the prey is dangerous, a lone wolf can overcome it, but is at risk of losing the carcass to scavengers. However, when the two wolves capture the prey together, they can better protect the carcass from scavengers and hence receive a higher reward. In this case, the AI agents worked together and it shows that AIs can recognise the benefits of cooperation that will have the best outcome for all. Watch the Wolfpack game below: What Do These Two Games Explain? “In the Wolfpack game, learning a defecting lone-wolf policy is easier than learning a cooperative pack-hunting policy. This is because the former does not require actions to be conditioned on the presence of a partner within the capture radius. In the gathering game, the situation is reversed. Cooperative policies are easier to learn since they need only be concerned with apples and may not depend on the rival player’s actions.” However, optimally efficient cooperative policies may still require such coordination to prevent situations where both players simultaneously move on the same apple. Cooperation and defection demand different levels of coordination for the two games. Wolfpack’s cooperative policy requires greater coordination than its defecting policy. Gathering’s defection policy requires greater coordination (to successfully aim at the rival player).” the DeepMind researchers explained in the paper. The Wolfpack AI games have a clear message the potential benefit of artificial intelligence is huge. Such advancement in AI can certainly benefit humans as the corporation can be the key to greater individual success in certain situations. It could also lead to systems that can develop policies and real-world applications. What Could The Aggressive Nature Of AI Imply? As Professor Stephen Hawking rightly said, “We cannot predict what we might achieve when our own minds are amplified by AI. Perhaps with the tools of this new technological revolution, we will be able to undo some of the damage done to the natural world by the last one – industrialisation. And surely we will aim to finally eradicate disease and poverty. “Every aspect of our lives will be transformed. In short, success in creating AI could be the biggest event in the history of our civilisation.” On the flip side, there are some challenges related to the data used by machine learning systems. The WEF paper highlights that even if the machine learning algorithms are trained on good data sets, their design or deployment could encode discrimination in ways like choosing the wrong model; building a model with inadvertently discriminatory features; absence of human oversight and involvement; unpredictable and inscrutable systems; or due to unchecked and intentional discrimination. As artificial intelligence and machine learning are becoming more advanced day by day, it includes less human supervision and less transparency. It is important that humans are kept in a loop where factors are being unexpectedly overlooked. And we need to build human values in our machines because we have them, machines don’t and something like ‘be reverent, loving, brave, and true’ may be difficult to digitize. But one way or the other, it’s imperative that we find ways to infuse human values into our AI.","excerpt":"Self-preservation is an instinct that every human being has. Now imagine this, an AI that cannot process any emotions yet have the same instincts as humans and thinks to preserve itself over others! That is what researchers from Google-owned DeepMind firm study have found out in this major advancement by their artificial intelligence. Google acquired […]","categories":["IT Services"],"tags":["AI (Artificial Intelligence)","DeepMind","DeepMind AI","Google","Machine Learning","stephen hawking"],"author_name":"Smita Sinha","publish_date":"2018-04-07T06:03:16","publication_year":"2018","word_count":1127,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","Machine Learning","stephen hawking","Git","programming_languages:Go","Aim","AI agents","Google","DeepMind AI","R","AI (Artificial Intelligence)","DeepMind"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","R","Go","Git","AI agents","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/googles-deepmind-ai-takes-an-aggressive-nature-is-that-a-reason-to-worry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":39817,"title":"Deep Dive: Indiabulls CIO On Transforming The BFSI Major With Data &#038; Automation","content":"If there is one business leader who has watched the tectonic shift in the IT landscape up and close, it is Nafees Ahmed, Group Head – Chief Information Officer, Indiabulls Group. An industry veteran with more than 20 years of experience, Ahmed manages the entire information technology portfolio at Indiabulls and has been pivotal for introducing a slew of digital initiatives. Ahmed watched the world move towards mobile and digital and recognised the real value of data as an asset and the need of the hour — leveraging advanced analytics and business models to deliver real customer value and also monetising the data in real time. Under his technological leadership, he broke down the silos existing in analytics that operated separately across housing finance, consumer finance, marketing and product functions. Early on, he realised that “digital is the way forward.” Currently, his role is more focused on driving real business value from data and building a team of data scientists, analysts and managers that have the technical skill-set to push the new initiatives\/products forward. From consolidating data to implementing analytics tools that offer a single view of the customer, Ahmed has helped bridge the gap between technology and business. As one of the major BFSI players, the shift came when the fintech ecosystem in India started gaining ground with the likes of Google and Amazon offering niche financial products. “That’s when we realised that if we don’t adopt cutting-edge technology, we won’t be able to survive the competition,” he said. As CIO at Indiabulls, Ahmed is leading the organisation’s data agenda forward – building a data-driven team, putting the infrastructure in place (an ongoing process) and at the same time launching successful data-driven products. “In the last 2-3 years, my management realised that digital is the way forward. Technology is not just for growth, it is basically a question of your survival. My first project on this digital journey was E-Home Loans. And we were the first in India to launch this product. With this product, our customers could complete the application journey completely online. Aadhaar, the whole digital platform for the government of India was instrumental in the KYC process,” he said. IT has moved from support to business enabler function When Ahmed joined Indiabulls around 13 years back, at that point in time, the IT department was treated as a support function. “We managed legacy systems, got the bug fixed but in the last couple of years, the role has reversed completely. The business is interacting with the technology department, for example, whenever there is a new entity coming up or a new product is being defined, tech teams comes into play to implement the product,” he shared, commenting on the changing role of the IT department at Indiabulls, one of the leading financial services groups in India. Dhani app by Indiabulls has proven to be a gamechanger in personal loan At a time when personal loan process was encumbered due to compliance issues, Ahmed and his team launched the Dhani app — billed as a first-of-its-kind that reduced the loan disbursal to 3 minutes. One of the first projects undertaken by Ahmed and his team, Dhani app for loan processing was introduced in 2017. The app was rolled out to all clients who wished to avail personal loans. Billed as a completely digital product and mobile only, it automated the loan processing journey with crucial elements such as e-signature, e-payments, online EMI repayment process and online customer verification. “This was a complete online journey without any physical connection with our customers. With this app, we were able to transfer the money actually in customer account within three minutes, in real time,” he shared. So far, the company has 1.1 million customers and the app was voted as one of the topmost downloaded fintech apps on Google Playstore with 15 million downloads. This time, the company has set a more ambitious target – of achieving a customer base of 1 million per month. Machine Learning@Dhani app Given that this was a complete online process without any physical contact with the customer, the underwriting process was carried out by collecting a lot of data, understanding the capability to repay loans and even understand the intention of the customer to repay. “The second thing we ensured was whether the data is coming from reliable sources. For example, we get data directly from ITR and CIBIL. There is a lot of unstructured data that we get which has analysed on a real-time basis,” he said. The team built an underwriting engine based on machine learning with a team from Experian and it automated the email underwriting process. Tool Stack at Indiabulls Just like his counterparts in the BFSI sector, Ahmed believes that data security is paramount. As far as the core applications are concerned, the company relies on a private cloud. While performance testing is done on Azure, 90% of the cloud infrastructure is on VMware. For non-core applications, the team uses Salesforce and Microsoft Dynamics 365. Meanwhile, the company also has two data centres – in Noida and Mumbai. On the front-end, the team uses Java,.NET frameworks, SQL, while in the middle layer, the team uses IBM MobileFirst platform. To get more out of data, Ahmed is also focusing on getting certain off-the-shelf solutions from the that can help in expediting the time-to-market for products. With a team of 500, Ahmed is also building a team of 50-60 people who’ll do in-house development, advance cutting-edge products and focus on building innovative solutions. Currently, he’s shopping for RPA solutions that can automate certain backend applications and help in day-to-day operations. Outlook In closing, he also shared his thoughts on the current talent crunch in the market. Besides the current talent gap, there is also a fear related to automation. “Whenever we talk about automation, people think their jobs will be replaced and I have to convince them the company will be doing a lot of things to improve productivity. However, the challenge lies in implementing new technologies and pushing more users to adopt them,” he said, in closing.","excerpt":"If there is one business leader who has watched the tectonic shift in the IT landscape up and close, it is Nafees Ahmed, Group Head – Chief Information Officer, Indiabulls Group. An industry veteran with more than 20 years of experience, Ahmed manages the entire information technology portfolio at Indiabulls and has been pivotal for […]","categories":["AI Features"],"tags":["automation testing"],"author_name":"Richa Bhatia","publish_date":"2019-05-28T12:26:32","publication_year":"2019","word_count":1017,"keywords":["Go","machine learning","automation testing","AI","R","RAG","analytics","SQL","Tecton","Azure","Java"],"extracted_tech_keywords":["AI","machine learning","analytics","Tecton","RAG","Azure","R","SQL","Go","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deep-dive-indiabulls-cio-on-transforming-the-bfsi-major-with-data-automation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":61681,"title":"Board Infinity Introduces AI &#038; Data Structure Microlearning Amid COVID-19 Lockdown","content":"With online classes being the norm today, Board Infinity, a career tech startup has introduced micro-learning in data structure and algorithms, artificial intelligence and machine learning, and languages like Python for data science. With schools and colleges across the globe have closed down its classrooms to stop the spread of COVID-19, learners and teachers are shifting their focus to focus on the online medium, and Board Infinity’s this initiative will help the sector with online courses. Apart from AI & ML, the company is also providing micro-learning courses on personal finance and investment planning, product management, and placement and internship preparation. View Post According to Board Infinity, these modules will be available from 10 hours to 20 hours depending upon the students’ needs. They can thus master the fundamentals from the comfort of their home and at an affordable cost. On successful completion, the students will also receive a digital certificate for each program. Additionally, the students have the advantage of attending live classes and avail benefits of being mentored by industry experts to strengthen the fundamentals of each course. The experts include Rahul Baid, a senior consultant, at Deloitte Consulting; Yashwant Pachisia, the AVP of commercial banking at HSBC; Ashish Anand, the marketing director at Droom Technology; Mirza Rahim Baig, an analytics leader at Flipkart; Kunaal Naik, an analytics practitioner and evangelist at DELL EMC; Ruble Joseph the vice president of eClerx; Naggapan Ramaswamy, the product manager of Razorpay; and Aditya Mehta, a data scientist at General Mills. Speaking on this occasion, Sumesh Nair, co-founder of Board Infinity stated that this COVID-19 pandemic had thrown life out of gear for everyone, whether it is business, parents or students. However, thanks to technology, students can enhance their skills and continue their education. “Keeping this in mind, we have introduced a shorter duration program for students and working professionals to continue their learning and get accustomed to new-age skills. This is a great time to upskill until the market picks up in the upcoming quarters,” said Nair. He further added that since the lockdown, we had added more than 8000 learners and there has been a surge in traffic from cities other than the top 10 such as Madurai, Nagpur, Jalandhar, Calicut, Belgaum, Raipur, Udaipur, Ranchi and Kanpur.","excerpt":"With online classes being the norm today, Board Infinity, a career tech startup has introduced micro-learning in data structure and algorithms, artificial intelligence and machine learning, and languages like Python for data science.   With schools and colleges across the globe have closed down its classrooms to stop the spread of COVID-19, learners and teachers are […]","categories":["AI News"],"tags":["covid-19","data analytics certificate","Data Analytics Certification","Data Science Certification","data science online learning","e-learning","learn ai","machine learning for cities","online education","Online Learning"],"author_name":"Sejuti Das","publish_date":"2020-04-14T15:47:31","publication_year":"2020","word_count":377,"keywords":["Data Science Certification","Git","Data Analytics Certification","R","data science online learning","machine learning for cities","data science","artificial intelligence","covid-19","online education","analytics","Go","machine learning","AI","ML","e-learning","Online Learning","data analytics certificate","Python","learn ai"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/board-infinity-introduces-ai-data-structure-microlearning-amid-covid-19-lockdown\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164671,"title":"Indian IT ‘Should Be Paranoid’ About AI &amp; Ditch its 30-Year-Old Business Model","content":"Indian IT giants like TCS, Infosys, HCLTech, and Wipro have largely stayed away from building foundational AI models. This remained true even after the launch of China’s DeepSeek, which shook up the entire Western world. However, India’s focus has remained on adoption. Now, things might finally change as AI tools are looking to make the future difficult for Indian IT firms, and they must rethink their strategies and invest in indigenous language models to remain competitive. Speaking at an industry event in Mumbai, HCLTech CEO Vijayakumar C emphasised that AI’s disruption in IT services is unlike previous technological shifts such as cloud computing and digital transformation. “The changes AI is assuring are very different, and we need to be more proactive to even categorise our revenues to create completely new businesses,” he said. Generative AI is expected to accelerate software development by automating coding and reducing project timelines. Vijayakumar pointed to a financial services firm where AI-driven efficiencies reduced the timelines of a $1 billion technology transformation program from five years to three-and-a-half years. Highlighting the strategic need for India to build its own language models, he cautioned against over-reliance on foreign AI infrastructure. “We should not assume that these (language) models will continue to be open source. I think these are going to be the coins on which the geopolitics will be played off,” he warned. Is it Really Needed? Vijayakumar stressed that with declining costs of AI training, India must invest in economic ways to develop its own models. “I strongly believe that the business model is ripe for disruption. What we saw in the last 30 years was a fairly linear scaling of revenues and people. I think time is already up for that (business model),” he added. Infosys CEO Salil Parekh echoed the call for agility, urging Indian IT firms to remain proactive in navigating AI-driven transformations. “I think we have to be paranoid. We have to be non-complacent. That is how we can manage to keep up with what’s going on in the industry,” Parekh said. Tanay Pratap, YouTuber and founder and CEO of Invact Metaversity, while speaking to AIM earlier, stressed that AI coding tools and agents coming up in the market could threaten Indian IT employees. “Thirty years of IT revolution in the country, but we still don’t know how to produce coders at scale,” Pratap said. He added that even when graduates come out of universities with a computer science degree in India and join IT firms, their ability to code still remains questionable. Currently, the biggest exporter in India that contributes to the economy is the IT service companies. However, instead of having coders and programmers, these IT services support testing-related roles. “The whole business model [of Indian IT] is about exporting services like testing to global customers,” Pratap said, adding that this could be under threat. Vijaykumar also believes that a large part of the IT industry is driven by input-based models. “People are delivering certain outcomes. We need to dramatically change the output.” He added that a lot of services that these firms deliver need to become platform-based from people-based. This aligns with K Krithivasan, CEO and MD of TCS, who earlier pointed out that building LLMs has no huge advantage as the cost outweighs the benefit. He added that since most organisations in India are system integrators, companies need to use products as software and ensure that clients receive the benefits. At the same time, he also agreed that building it for regional languages makes sense for democratising the technology. Recently, discussions on a bustling Reddit thread titled ‘I don’t see any hope in the future of this IT industry’ centred around how AI is definitely stronger than humans and will get better with time. What Should be Done? Even though the IT firms have enough funds to make a foundational model, they won’t build one unless their clients ask them to or there is some requirement from their side. The whole idea of IT companies not building products might need to change completely if it needs to survive. While TCS, Infosys, Wipro, and HCLTech have started developing agentic AI frameworks, small language models, and even drug discovery, their efforts remain focused on clients alone. These initiatives do not prioritise building foundational technologies for the country. Tech Mahindra built its Project Indus, the only foundational model emerging from an IT firm in India. Infosys co-founders Nandan Nilekani and Kris Gopalakrishnan are still debating whether they need one. Gopalakrishnan earlier wrote on X that India needs to build its foundation model for a cultural and strategic economy, while Nilekani recently reiterated that a foundational model is unnecessary as long as use cases are built. Meanwhile, several industry experts like Ajai Chowdhry and CP Gurnani earlier told AIM that it is important for Indian firms to build foundational models. It seems like it is finally going to take place with HCLTech’s change of plans.","excerpt":"AI tools are looking to make the future difficult for Indian IT firms.","categories":["IT Services"],"tags":["AI in Indian IT","Developers"],"author_name":"Mohit Pandey","publish_date":"2025-02-26T20:00:00","publication_year":"2025","word_count":824,"keywords":["AI in Indian IT","Go","agentic AI","TPU","AI","cloud computing","Git","Aim","generative AI","small language models","R","Developers"],"extracted_tech_keywords":["AI","generative AI","agentic AI","Aim","small language models","cloud computing","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/indian-it-should-be-paranoid-about-ai-ditch-its-30-year-old-business-model\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10162504,"title":"Deeptech Startup Astrome Bags $10 Million to Expand Wireless &amp; Space Communication","content":"Deeptech startup Astrome Technologies has secured $10 million in a funding round led by Appollo Fund, with additional investments from the IAN Group. The funding will support the expansion of the company’s wireless connectivity technologies, product lines, and market presence in both terrestrial and satellite communications, the company announced on LinkedIn. Astrome Technologies aims to leverage this investment to expand its role in the evolving wireless and space communication industries. Further plans include developing new product lines, scaling operations, and strengthening market presence. Investors from India, Singapore, and Dubai contributed to the round, reflecting the growing global interest in advanced wireless technologies. Apollo Funds described Astrome as a leader in wireless communication innovation. Bandana Kankani, banker at Nexxt Mile who led the funding round, emphasised the demand for high-bandwidth, reliable communication solutions and expressed confidence in the company’s future growth. Neha Satak, CEO of Astrome, highlighted the funding as a step towards realising the company’s vision of transforming connectivity. “This foray into the satcom (satellite communications) market represents the timely realisation of the vision with which this company was founded to drive innovation and create solutions that enhance connectivity on a global scale,” she reportedly said. In her LinkedIn post, she acknowledged the support of government policies, institutional backing, and investor confidence in the company’s journey. She credited Indian Prime Minister Narendra Modi and key ministers, including Ashwini Vaishnaw, Jyotiraditya M Scindia, Rajnath Singh, and Priyank M Kharge, for fostering a favourable ecosystem for deep-tech and hardware startups in India. She also highlighted the support of institutions such as the Ministry of Defence (iDEX-DIO), Department of Telecommunications, Technology Development Board, and Ministry of Electronics and Information Technology (MeitY) for their role in implementing policies that promote indigenous hardware innovation. The Karnataka government’s Electronics System Design and Manufacturing (ESDM) policy was specifically mentioned for aiding Astrome’s growth. This raises expectations and extensively highlights the importance of investment in deep-tech policies in the upcoming Union Budget. Astrome operates from Bengaluru and California and previously raised $3.4 million in a bridge round in 2021. Investors from India, Singapore, and Dubai participated in the latest round, showcasing a global interest in advanced wireless solutions. Astrome, founded in 2015 by Satak and Prasad (HL) Bhat, specialises in millimetre-wave E-band radios and satellite communication products aimed at enhancing 5G and rural telecommunication infrastructure. The company plans to use the funds to increase production capacity, develop next-generation solutions, and supply its flagship GigaMesh product to international markets through partnerships with global original equipment manufacturers (OEM). GigaMesh, designed to work with optical fibre systems, supports 5G and 6G connectivity and incorporates features such as electronic auto link alignment and point-to-multipoint connectivity. These innovations enable quicker deployment, lower costs, and improved network performance for digital infrastructure across land, sea, and air.","excerpt":"Further plans include developing new product lines, scaling operations, and strengthening market presence.","categories":["AI News"],"tags":["Funding","space technology","Startup"],"author_name":"Sanjana Gupta","publish_date":"2025-01-30T11:51:22","publication_year":"2025","word_count":462,"keywords":["Go","Funding","AI","Startup","innovation","space technology","Git","RAG","Ray","Aim","ViT","R","startup"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","Git","ViT","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deeptech-startup-astrome-bags-10-million-to-expand-wireless-space-communication\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36291,"title":"Top 5 Initiatives By Tech Giants To Upskill And Train Women Coders","content":"There is a lot of data available on how women coders or women from STEM background are lesser in number than their male counterparts. Many companies still prefer hiring male coders and that there is an immediate need to make data science and analytics industry a more gender-diverse field. Many companies are working to overcome the underrepresentation of women in technical fields and are collaborating with non-profit organisations that boost women to work on areas such as coding. In this article, we list five such companies that have taken up initiatives to train women coders. IBM: IBM recently announced significant collaborations across India focused on advancing skills and careers of more than 2 Lakh students in STEM fields. The initiative which started in three state governments of Karnataka, Telangana and Andhra Pradesh right now, will expand to other states in the next few months. This collaboration between IBM and the state governments is aimed at strengthening the women workforce in emerging technologies such as cloud and AI. The partnership aims at strengthening the technical capabilities of women and enable more women coders to lead the way in areas such as analytics and data science. The company believes that this recently announced program will equip women with the skills of the future to help drive the economy of the next decade. Alibaba: In a recent development by the Chinese e-commerce giant, it has partnered with iamtheCODE, the first African-led global movement to advance STEAMED Education. This initiative is focused on teaching one million women and girls to code by 2030. It will uplift women programmers in fields such as cloud computing and data science. The initiative, part of a wider movement dubbed Tech for Change is launched by cloud computing division of Alibaba Group and is aimed at dealing with the inequality in technical-based job roles. As a part of this partnership, Alibaba Cloud will provide tailored courses on a range of topics including cloud computing, data analysis, machine learning and security. Women who pass the course will be given certificates to help them along with their career path and strengthen their stand in male-dominated areas such as coding. Amazon: Amazon has been empowering girls and women in technical fields since the last few years to overcome the imbalance in the coding community. In 2017, it partnered with Girls Who Code to build a pipeline of future female engineers in the United States. A similar initiative was taken by the company in 2018 to bring more women into the STEM fold with Amazon India’s ‘I Want to Code’ programme. The programme focused on creating awareness among girls about career opportunities in the STEM fields and learn about the prospects that areas such as engineering and coding held in the future. It conducts workshops and hands-on training to help girls kick start their understanding of coding and pursue it ahead. Amazon also has a program called Amazon Future Engineer that is committed to spending on computer science education focused on kids and young engineers. Google: Google teamed up with non-profit organisation MotherCoders and New York City’s Women. NYC initiative to offer a free, nine-week tech training program for moms. While this initiative is aimed at helping moms in New York City, they may come up with similar initiatives in other countries and cities based on the need. This initiative was taken after observing that mothers tend to get pushed out of the workforce most of the times while taking care of their babies or stepping in after taking a maternity break. MotherCoders is geared toward moms who want to re-enter the workforce, start their own company or change careers. Through the nine-week program, moms can focus on learning HTML, CSS and Javascript, in addition to tech trends and issues in the industry. The best part is that the programme also offers on-site childcare at no additional cost. It has been in action since 2015 where Google takes care of the entire bill of the program. VMWare: VMware, Inc., a leading innovator in enterprise software, partnered with Women Who Code (WWCode), the world’s largest and most active community dedicated to inspiring women to succeed in tech upskill 15,000 people who have left the industry and help them return to work.  The program is off to a great start with program registrants and support from leading corporations such as Airtel, Dimension Data, Cognizant, and Dell. It includes training on multi-cloud and hybrid cloud platforms and infrastructure, cloud management, and more. VMinclusion Taara, as it is called is an effort to rebalance the workforce by helping women on a career break come back to work.","excerpt":"There is a lot of data available on how women coders or women from STEM background are lesser in number than their male counterparts. Many companies still prefer hiring male coders and that there is an immediate need to make data science and analytics industry a more gender-diverse field. Many companies are working to overcome […]","categories":["AI Trends"],"tags":["ai certificates"],"author_name":"Srishti Deoras","publish_date":"2019-03-14T05:06:25","publication_year":"2019","word_count":770,"keywords":["ai certificates","data science","Go","machine learning","AI","cloud computing","ML","Aim","analytics","JavaScript","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","cloud computing","R","JavaScript","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-initiatives-by-tech-giants-to-upskill-and-train-women-coders\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10166426,"title":"OpenAI Releases New Audio Models to Power Voice Agents","content":"OpenAI has launched new speech-to-text and text-to-speech models in its API, providing developers with tools to build advanced voice agents. These models improve transcription accuracy and introduce customisation options for generated speech. The new speech-to-text models, gpt-4o-transcribe and gpt-4o-mini-transcribe, improve word error rate and language recognition compared to Whisper models. In its blog post, OpenAI said these advancements stem from reinforcement learning techniques and extensive training with diverse audio datasets. The models aim to improve transcription reliability in noisy environments, varying speech speeds, and different accents. “Our latest speech-to-text models achieve lower word error rates across established benchmarks, reflecting improvements in transcription accuracy and language coverage,” OpenAI said. Developers can now also control how the text-to-speech model speaks. The gpt-4o-mini-tts model allows developers to instruct the model to adopt different speaking styles, such as mimicking a customer service agent. This feature expands use cases in customer interactions and creative storytelling. However, OpenAI clarified that these models are limited to synthetic preset voices. The company credits improvements in its audio models to pretraining with authentic datasets, advanced distillation methodologies, and reinforcement learning. Distillation techniques have enabled smaller models to retain conversational quality while reducing computational costs. The new models are available to all developers through OpenAI’s API. OpenAI has also integrated these models with its Agents SDK to simplify development. For real-time, low-latency speech-to-speech applications, OpenAI recommends using its Realtime API. Looking ahead, OpenAI plans to enhance the intelligence and accuracy of its audio models and explore custom voice options. The company is also engaging with policymakers, researchers, and developers on the implications of synthetic voices. Moreover, OpenAI intends to expand into video, enabling multimodal agentic experiences.","excerpt":"The company said these advancements stem from reinforcement learning techniques and extensive training with diverse audio datasets.","categories":["AI News"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2025-03-21T10:46:52","publication_year":"2025","word_count":276,"keywords":["API","OpenAI","AI","Modal","GPT-4o","RAG","GPT","Aim","R","llm_models:GPT"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","Aim","RAG","R","API","GPT","Modal","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-releases-new-audio-models-to-power-voice-agents\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":2187,"title":"F1 pitches on IoT solutions to enhance racing experience at Grand Prix","content":"Now, instead of sitting on the sidelines and rooting for their favourite drivers, viewers now have a chance to participate more actively at the upcoming Formula 1. Tata Communications, the official connectivity provider for Formula 1, has unveiled the second challenge for the 2017 F1 Connectivity Innovation Prize. The theme is simple: Come up with ideas that can enhance the track-side experience for fans at the Grand Prix racing tournament through Internet of Things (IoT), mobile technologies and embedded connectivity. “We want to see how IoT solutions can enhance the F1 racing experience for the millions of fans who attend the Grands Prix all over the world – from before they even arrive at the track till the end of the race weekend,” said John Morrison, Chief Technical Officer of Formula 1 in a press statement. He explained that through mobile phone apps, fans could tune into customised live video feeds, take part in live polls and synchronised cheering and interact with each other with ease. Morrison added: “…Fans’ emotions could also be tracked to make aggregated emotion charts on large displays at the circuit, creating a more interactive, immersive and thrilling race experience.” The Grand Prize of the F1 Connectivity Innovation Prize is a cheque for $50,000. The judges pick three winners from each of the two challenges set by Mercedes-AMG Petronas Motorsport and Formula 1. The six winners from both challenges will be awarded trips to the 2017 Formua 1 Etihad Airways Abu Dhabi Grand Prix, where the jury will announce the winner. In addition, one person from one of the three winning teams of this challenge will be given a special prize: they will get a chance to work alongside the technical team of Formula 1 at a Grand Prix of the winner’s choice for a unique behind-the-scenes experience. A full brief for the second challenge can be downloaded from the F1 Connectivity Innovation Prize website. The closing date for the challenge is 2 August 2017.","excerpt":"Now, instead of sitting on the sidelines and rooting for their favourite drivers, viewers now have a chance to participate more actively at the upcoming Formula 1. Tata Communications, the official connectivity provider for Formula 1, has unveiled the second challenge for the 2017 F1 Connectivity Innovation Prize. The theme is simple: Come up with […]","categories":["AI News"],"tags":["AI and IoT"],"author_name":"Prajakta Hebbar","publish_date":"2017-07-26T09:09:36","publication_year":"2017","word_count":329,"keywords":["programming_languages:R","AI","innovation","AI and IoT","ViT","R"],"extracted_tech_keywords":["AI","R","ViT","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/f1-pitches-iot-solutions-enhance-racing-experience-grands-prix\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":15127,"title":"Indian Railways embrace data analytics to monetize the available data","content":"Pacing up with the global counterparts, Suresh Prabhu-led Indian Railways has stepped up its game in digitization and using data-centric approaches to benefit the railways. It is quite apparent that Indian Railways carry millions of passengers daily apart from tonnes of commodities across the regions. This heavy transaction results in huge amount of data, which the Indian railways is planning to monetize. However, the government has made sure that no privacy law would be compromised in doing so. The announcement which came during the inauguration of round table conference on data analytics for the railways, the government officials are hopeful that analytics in railways can help in determining pricing of services, planning train operations, routes, safety measures, proactive disaster management, predictive maintenance to avoid failure and much more. Suresh Prabhu believes that Indian Railways is on the largest data creators in the world, which creates large repositories of data in the form of passenger ticketing and freight operations. He is quick to add that this large volume of data needs to be used wisely and data analytics is a way forward. He also highlighted the need for proper data analytics for the future management information system, making appropriate decisions and eventually monetizing the data, albeit any compromise to privacy law and ethics. There are many forms of data in the railways such as reservation data of trains, passengers, earnings, utilisation of trains, class wise occupancy, waiting lists and passenger profile etc, which the government officials are trying to tap upon, to provide innovative products to passengers so that occupancy is improved and they get confirmed accommodation.","excerpt":"Pacing up with the global counterparts, Suresh Prabhu-led Indian Railways has stepped up its game in digitization and using data-centric approaches to benefit the railways. It is quite apparent that Indian Railways carry millions of passengers daily apart from tonnes of commodities across the regions. This heavy transaction results in huge amount of data, which […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-05-23T11:32:47","publication_year":"2017","word_count":266,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","Git","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Git","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-railways-embrace-data-analytics-monetize-available-data\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094342,"title":"Microsoft Issues Code Green Alert","content":"While generative AI is all the hype, there is a side to it that the users of the technology do not realise. These LLM-based models like ChatGPT are extremely thirsty. Each 20 questions you ask these chatbots requires almost half a litre of water to give you answers. And that mostly goes into cooling the data centres where the data is stored. More than that, these models rely on data centres that are extremely carbon heavy on the environment. According to the estimates from the World Economic Forum, generative AI has the potential to reduce carbon emissions by 4% by 2030. But on the other hand, it was also predicted in 2017 that these data centres and information technology sector as a whole, will contribute to a huge 14% of the carbon emissions globally by 2040. In an exclusive interview with AIM at the Microsoft’s Future Ready Sustainability Summit, Alok Lall, sustainability lead of Microsoft India, explained about the changing dynamics of the companies to address the problems by making greener software from the ground up. “When we look at reducing emissions, it is very easy to look at infrastructure and get more efficient hardware like servers, heating, ventilation, and cooling systems,” said Lall. “But how I look at it is we do not understand the main ingredient of what the application does — code.” How do we make the code light? Making data centres efficient is a good way to go forward with sustainability, but that is if we look at the code before being deployed. “Our eventual goal is to not have developers apply green software principles after they’ve written the code,” said Lall. A step for this was taken by Microsoft when they partnered with Accenture, GitHub, and Thoughtworks to create Green Software Foundation in 2021. Announced at Microsoft Build 2021, the goal of the foundation is to create a trusted ecosystem of people and for people for best practices in software development. Currently, the Green Software Foundation is working on a software carbon intensity score and a Carbon Aware SDK that developers can start using upfront. Microsoft has partnered with Cast Software, a technology corporation based in New York, to build on more green products and work on software intelligence. One such product was presented at the conference that scans your code as you write them on your software before it is deployed, and rates them from how green it is to how carbon intensive it is. Still under development, the tool tells you how you can fix the code to be greener by requiring less computation. Lall also explains how sustainability cannot be an afterthought, but has to be a design principle. Speaking about low code platforms like Copilot, Lall said that these softwares have been designed with the same principles in mind. These capabilities are inherently provided by the cloud.” This means that the code generated by these platforms is already designed to be as green as possible. The only point that remains is that even for the generation of codes through the server, the server needs to run and emit energy, thus emitting more carbon into the environment, and thirsting for more water. Making AI less water-intensive There is a solution for that as well. Everybody knows that AI is computation intensive. “We solved the design of the model part by using green software. For the second part, to make the infrastructure greener, we have made data centres to run at a PUE (power usage effectiveness) 0f 1.12.” explained Lall. Microsoft wants to make sure that it does not use water to cool its data centres anymore. For that, the company relied on air cooling, and is now shifted to adiabatic cooling, which involves condensing the same water that evaporated while cooling the motherboard by keeping it dipped in a liquid. Microsoft made the claim that the company would be carbon negative, produce zero waste, and water positive by 2030. There is a possibility that this prediction did not account for the current generative AI and large language model technology. To answer this, Lall said that integrating OpenAI and GPT technology into our systems was all part of the large corpus. “We know that it will go up and we are increasingly investing in areas that will help us be carbon negative by 2030, and achieve our goals.” Energy at its core, literally According to the 2022 Sustainability Report of Microsoft, the company has brought down its own emissions by 22% and by 0.5% from their suppliers and customers. To achieve the 2030 carbon negative goal, Lall said that the company has the target of becoming 100% renewable energy reliant by 2025. Recently, Microsoft also made a huge investment in Helion Energy, a nuclear energy-based startup. It is clear that the company is taking many steps to stay ahead of the competition, while also keeping sustainability at its core. “We will continue to identify and find innovative ways to power our data centres,” said Lall, “to get to the goal we have set for ourselves for 2030″. Apart from generative AI, Microsoft has also been making other areas of their software and hardware as green as possible. The recent Windows 11 22H2 update is carbon aware, and will only update when the emissions in the day are lower. At Microsoft Hyderabad, they have been working on a project to absorb moisture from the atmosphere and convert it into 2,000 litres of water per day and also created a digital twin of the campus to track carbon emissions in real-time.","excerpt":"Green Software Foundation is working on a software carbon intensity score and a Carbon Aware SDK that developers can start using upfront","categories":["Global Tech"],"tags":["Accenture","codex","Generative AI","Github Copilot","Microsoft","OpenAI","sustainability"],"author_name":"Mohit Pandey","publish_date":"2023-06-02T10:00:00","publication_year":"2023","word_count":926,"keywords":["Accenture","Go","ChatGPT","OpenAI","AI","sustainability","chatbots","Github Copilot","Git","Aim","codex","generative AI","Rust","Generative AI","R","Microsoft"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Aim","chatbots","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-issues-code-green-alert\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":31580,"title":"UK To Invest £1 Million An AI-Based Health Programme","content":"The British High Commission announced at an event in Delhi that the UK would be making an ambitious one million pound investment in a programme which will bring several artificial intelligence companies from the UK to help tackle diseases in Indian hospitals and health centres. India-UK FutureTech Fest (FTF) which brought together over 200 companies and 1,000 delegates, featured several Indian and UK Tech companies, along with scientists, policy-makers and entrepreneurs. Reports said that this £1 million programme would focus on AI-based diagnostics which can tackle numerous prominent diseases in India. “The UK is a natural partner to India in our mutual interests and ambitions in technology, and our collaboration is overseen by the highest offices in both countries,” said Amitabh Kant, CEO NITI Aayog. “India could potentially be the Tech Garage for the rest of the world, and we are keen to invite the best technology companies and solutions to India — in areas like healthcare, manufacturing, mobility, finTech etc,” he added. “When Prime Minister Narendra Modi and Prime Minister Theresa May announced the India-UK Tech Partnership back in April, they put the enabling power of tech at the heart of our bilateral relationship,” Dominic Asquith, British High Commissioner to India said in a statement. Delighted to participate in India – UK Future Tech Festival focussing on fast growing sectors of AI\/ML, healthcare, digital manufacturing & electric vehicles. UK & India share complementarities in these areas & MSME sector can greatly benefit from this collaborative partnership. pic.twitter.com\/etbzu6QCAM — Amitabh Kant (@amitabhk87) December 12, 2018","excerpt":"The British High Commission announced at an event in Delhi that the UK would be making an ambitious one million pound investment in a programme which will bring several artificial intelligence companies from the UK to help tackle diseases in Indian hospitals and health centres. India-UK FutureTech Fest (FTF) which brought together over 200 companies […]","categories":["AI News"],"tags":["amitabh kant","Healthcare Automation","NITI Aayog"],"author_name":"Prajakta Hebbar","publish_date":"2018-12-14T13:47:38","publication_year":"2018","word_count":255,"keywords":["amitabh kant","artificial intelligence","programming_languages:R","AI","ML","Git","RAG","Healthcare Automation","ViT","R","NITI Aayog"],"extracted_tech_keywords":["AI","artificial intelligence","ML","RAG","R","Git","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/uk-to-invest-1-million-an-ai-based-health-programme\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10111494,"title":"Meet the Creator of ଓଡ଼ିଆ Llama","content":"When ChatGPT was introduced a year ago, Shantipriya Parida — the creator of Odia Llama — was quite disappointed that it did not understand any cultural context related to Odisha. Cut to present, he’s built Odia Llama (Llama2-fine tuned LLM for the Odia language) and started an open-source project called Odia Generative AI. The language is spoken by over 35 million people, and boasts a rich literary tradition and a unique identity. “I attempted to ask ChatGPT about local contexts. For example, I asked, ‘Can you tell me the recipe for rasgulla?’ It couldn’t provide an answer. It missed all the local context, and the answers weren’t even correct,” said Parida. Parida is currently working in Finland as a senior AI scientist at Silo.ai, which recently released its own LLM named Poro. “If Europeans can build an LLM that can provide better answers than OpenAI’s ChatGPT in their local language, I thought why can’t we do it for our own Indic languages,” said Parida. “That’s the reason we are working hard, and it feels good when people appreciate our effort,” he added. The Journey of Odia Llama To build Odia Llama, Parida formulated a three-step plan. “We’ll begin with a fine-tuned model, then move on to a pre-trained model, and finally, proceed to app deployment. Once we have the fine-tuned and pre-trained models ready, app development will be relatively straightforward,” he explained. Odia Llama, currently available on Hugging Face, is the fine-tuned version. Parida noted that his team of researchers is presently working on the development of the pre-trained version, which is currently in progress. “All the fine-tuned models have some limitations. In the foundation model, if you have only 0.5% or 2% of data in your local language, then, after a certain point, no matter how much you fine-tune it, it will get stuck,” he said. To train a pre-trained model, Parida’s team is currently working on collecting data. “We are collecting a lot of tokens. I think we have already collected around 30 million tokens, and we are targeting at least 40 to 50 million tokens. This way, we can expedite the release of our first pre-trained model,” said Parida. The data is sourced from various online platforms, including blogs, Wikipedia, Odia newspapers, local textbooks, literature, magazines, and government websites. Parida said that they have also developed in-house tools, named Olive Scraper and Olive Farm. Olive Scraper is a web scraping tool for extracting Odia content from various sources (e.g., websites, PDF, DOC, etc.), while Olive Farm generates LLM instruction sets in Indic languages. Presently, it offers support for Hindi and Odia, with seamless scalability to incorporate additional languages on the horizon. Regarding computing, GPUs, and infrastructure, Parida mentioned that they received support from E2E Networks. Moreover, he said that for fine-tuning, they have ample resources, as his team consists of various independent researchers, who have access to GPUs, which they utilise for research purposes. AI Tutor Odia Generative AI recently created an AI tutor named Acharya. This tutor facilitates self-learning in Hindi for students, offering real-time doubt resolution. Acharya was developed using LLM (Mistral-7b Hindi, fine-tuned) and retrieval augmented generation (RAG). “For example, if you’re a tutor and want to create a comprehensive lesson plan on a specific topic, this can assist with that. Similarly, if you want to evaluate, for instance, create a set of questions and assess them, it can be helpful. So, it has multiple use cases,” said Parida. Acharya operates as a client-server web application, with the client built using JavaScript and the server utilising Python with Fast API for seamless communication. Initial assessment scores indicate BERT (F1): 0.72 and RAGAS Answer Relevancy: 0.72. The current demonstration version caters to Class 8 subjects of the CBSE board. It’s worth noting that Acharya will soon support various languages, cover a wide range of subjects, and be freely accessible. What’s Next? Given that there are now several Indic LLMs out in the market, like OpenHathi, Airavat, Krutrim and BharatGPT, Parida wants to create an Indic LLM benchmark next. “We are planning to build an LLM benchmark. You go, choose your model, and it will automatically tell you your model’s accuracy per task. It can be a fair comparison for anybody who wants to pick a model for research or any other purpose,” said Parida. Along similar lines to AI Tutor, Parida wants to build more AI apps focusing on government budgets and policies which would make it easier for local citizens to get information in native languages. “As a citizen, many times one wants to know about government policies and what the government is trying to do in your area. But you don’t know exactly whom to ask. Nowadays, information is available in the public domain, so it’s easy to build an AI app using an open-source model and using RAG,” said Parida. “We are not a company, and we are building without the intention of selling anything. We started with one objective: to ensure that our Odia language does not lag behind. So, whatever we are building is solely for the benefit of the people,” he concluded.","excerpt":"Odia Generative AI recently created an AI tutor named Acharya for the CBSE Board.","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2024-01-27T16:00:00","publication_year":"2024","word_count":853,"keywords":["ChatGPT","Hugging Face","OpenAI","AI","ML","RAG","Python","generative AI","retrieval augmented generation","R"],"extracted_tech_keywords":["AI","ML","generative AI","ChatGPT","OpenAI","Hugging Face","retrieval augmented generation","RAG","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-the-creator-of-odia-llama\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10043767,"title":"[Jobs Roundup] Latest Job Openings For SAS Professionals In India","content":"We have listed the latest job openings for SAS professionals in India. 1| SAS BI Tools Application Lead at Accenture Location: Bangalore Responsibilities: Experienced in Installing and Configuring SAS 9.4 Servers.Working experience of SAS 9.4. Understanding of SAS architecture, LSF, Job Deployments and scheduling.Worked on various SAS client tools i.e. SAS Management Console, SAS Environment Manager, SAS EG, and SAS DI. Apply here. 2| SAS + SQL at Genpact Location: Gurgaon Responsibilities: Be the primary point of contact for assigned pursuits to work with the Marketing team for campaign execution, and solutions that meet the business needs of clients.Map the business requirement and execute campaigns using SAS \/ SAS CI for the respective portfolio(s).Work with each Service and Horizontal line (internally) to ensure campaigns are executed and launched on-time & error free and provide required support to the team. Apply here. 3| SAS EG at Wipro Location: Chennai\/Gurgaon\/Noida Responsibilities: SAS programming with various SAS products such as SAS\/BASE, SAS\/MACROS, SAS\/STAT, SAS\/EG and SAS\/SQ.Hands on experience of working in BASE SAS, SAS Enterprise guide. Knowledge of BASE SAS and SAS MACROS.Hands on experience in writing complex SAS SQL queries using PROC SQL. Apply here. 4| Consultant, Advanced Analytics Python\/R\/SAS at Fractal Location: Bangalore Responsibilities: Solve business problems & develop a business solution: Use problem-solving methodologies to propose creative solutions to solve a business problem.Recommend design and develop state-of-the-art data-driven analysis using statistical & advanced analytics methodologies to solve business problems.Develop models & recommend insights. Form hypotheses and run experiments to gain empirical insights and validate the hypothesis. Identify and eliminate possible obstacles and identify an alternative creative solution. Apply here. 5|  SAS Analytics Application Developer at Accenture Location: Mumbai Responsibilities: Delivery of good Quality of code.Build, Configure and Testing of SAS Programs.Code Promotions to different environments and test Support. Apply here. 6| Data Management Consultant – SAS, SQL & Teradata at Wells Fargo Location: Bangalore Responsibilities: Work closely with modeling and production teams to understand their requirements, and maintain the model development and implementation data.Underlying portfolio research and analytics using strong programming skills.Enhance the data preparation process to incorporate new requirements, handle upstream data changes and build stronger controls. Perform reconciliations between data sources and perform quality checks to ensure data integrity. Apply here. 7| SAS Core ACES Manager at Amazon Location: Bangalore Responsibilities: Identifying, preventing and\/or eliminating defects and unblocking the Seller’s business potential.Develop and implement scalable closed loop internal and external engagement mechanisms to enable team strategy.Work across peer teams and stakeholders to drive data driven process improvements to achieve resiliency, efficiency and scale for the issue support function. Apply here. 8| SAS Administrator at Collabera Location: Pune Responsibilities: Collaber Evaluated the Health Economics team (HEOR) SAS program issue and recommended solution of SAS needs.Setup SAS CDI to RaveWebService integration to automate the extraction of Rave study data.Configured StatXact on SAS grid Linux environment. Apply here. 9| SAS Cloud Systems Integrator at SAS Location: Pune Responsibilities: Engineer solutions to ensure scalability and reliability.Diagnose, document, report, and resolve system problems both independently and in a group setting.Work directly with external customers and interface with other support teams and vendors. Provide training, mentoring, and coaching for team members. Apply here. 10| Business Analyst SAS at SmartAnalyst Location: Gurgaon Responsibilities: To work with a multi-disciplinary team with responsibility of analyzing real world patient level data (healthcare claims data, electronic medical record data and other APLD) using SAS or other data processing tools.Develop, implement and manage holistic strategies for APLD data analysis (analysis plan, business rules, SAS coding, documentation, approval, monitoring and reporting), in collaboration with Life Sciences domain experts and other analysts.Determine how data is structured and locate data elements of interest. Apply here. AIMRecruits A subsidiary of Analytics India Magazine, AIM Recruits is an India-focused executive search firm that partners with leading businesses to assess and acquire top data science talent to drive breakthrough performance. AIM Recruits is a leading Executive Search Firm for Analytics, Data Science & Artificial Intelligence. 1| Machine Learning\/Artificial Intelligence Architect at Digit Insurance Location: Bangalore Responsibilities: To design applications and solutions using Machine Learning and AI. The role will focus on emerging AI, Machine Learning and Data Science.Responsibilities include architecture, total solution design from requirements analysis, design and engineering for data ingestion, pipeline, data preparation & orchestration, applying the right ML algorithms on the data stream and predictions.Application of Machine Learning will be key at all layers of the stack. This requires engagement in every phase of the system development lifecycle including requirements generation, system and software design, implementation, integration & test, and verification & validation. Apply here. 2| Assistant Manager – Data Engineering – SQL\/Python\/Apache Spark at The Smart Cube Location: Any Responsibilities: Should be comfortable in executing ETL (Extract, Transform and Load) processes which include data ingestion, data cleaning and curation into a data warehouse, database, or data platform.Should be comfortable with schema design. Experience in structured\/unstructured data and batch processing\/real-time processing (good to have).Be comfortable with SQL (mandatory), Python(mandatory), Scala (good to have) to manipulate and prepare data and conduct various analyses as needed. Apply here. 3| Senior Artificial Intelligence Engineer – Python\/OpenCV at Digit Insurance Location: Bangalore Responsibilities: Performing all technical development for assigned applications including architecture, design, developing prototypes, writing new code and API- s, and performing unit and assembly testing of developed software also as needed.Build and automate our AI\/ML data pipelines & work stream from data analysis, experimentation, model training, model evaluation, deployment, operationalization, and tuning to visualization.Assess the AI\/ML solutions\/use-cases provided by partner companies and help adoption. Apply here. 4| Data Architect – Hadoop\/Spark at Digit Insurance Location: Bangalore Responsibilities: Developing and implementing an overall organisational data strategy in line with business processes. The strategy includes data model designs, database development standards, implementation and management of data warehouses and data analytics systems.Coordinating and collaborating with cross-functional teams, stakeholders, and vendors for the smooth functioning of the enterprise data system.Managing end-to-end data & BI architecture, from selecting the platform, designing the technical architecture, and developing the application to finally testing and implementing the proposed solution. Apply here. 5| Senior Manager – Data Science at CitiusTech Location: Bangalore\/Mumbai\/Pune Responsibilities: Work with clients to understand hypotheses. Research to find relevant reference material.Data analysis and chart representations for study whitepapers. Design, create and iterate hypothesis into model building.Fine tuning models to optimise them. Apply here. 6| Artificial Intelligence Engineer – Machine Learning at Digit Insurance Location: Bangalore Responsibilities: Performing all technical development for assigned applications including architecture, design, developing prototypes, writing new code and APIs.Performing unit and assembly testing of developed software also as needed.Build and automate our AI\/ML data pipelines & work stream from data analysis, experimentation, model training, model evaluation, deployment, operationalization, and tuning to visualization.Research new algorithms to tailor solutions to the Insurance world. Apply here. 7| Data Engineer – ETL\/SQL at Digit Insurance Location: Bangalore Responsibilities: Able to take, understand & gather requirements, develop & implement the Business requirements\/ Flows. Solution Oriented.Creating & Developing Data Models. Should be able to debug, handle ETL failures on a daily basis.Scheduling & Monitoring of Batch Jobs. Prepare Plans for all ETL Procedures and architectures. Apply here. 8| Artificial Intelligence Architect – Microservices Implementation at Digit Insurance Location: Bangalore Responsibilities: Responsibilities include architecture, total solution design from requirements analysis, design and engineering for data ingestion, pipeline, data preparation & orchestration, applying the right ML algorithms on the data stream and predictions.Application of Machine Learning will be key at all layers of the stack. This requires engagement in every phase of the system development lifecycle including: requirements generation, system and software design, implementation, integration & test, and verification & validation.This role is expected to have experience developing solutions with the latest in current AI paradigms, algorithms, and models (e.g. deep learning, reinforcement learning, computer vision). Apply here. 9|  Senior Product Developer – Roadmap\/Strategy at Knowlvers Consulting Location: Delhi NCR Responsibilities: Proficient at Python-Django, with a minimum of 3 years of experience in Product development.Responsible for the management of the entire product lifecycle, from design to development to installation to maintenance. Apply here.","excerpt":"Check out the latest job openings for SAS professionals in India.","categories":["AI Hirings"],"tags":["analytics jobs in India","data science jobs in India","django","statistical analysis using sql","weekly job updates","weekly jobs roundup"],"author_name":"kumar Gandharv","publish_date":"2021-07-16T17:00:00","publication_year":"2021","word_count":1339,"keywords":["data science","statistical analysis using sql","artificial intelligence","machine learning","AI","ML","django","weekly job updates","data science jobs in India","analytics jobs in India","computer vision","deep learning","Aim","weekly jobs roundup","analytics","OpenCV"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","computer vision","data science","analytics","Aim","OpenCV"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/jobs-roundup-latest-job-openings-for-sas-professionals-in-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":30788,"title":"India And Russia Collaborate To Strengthen AI And Blockchain Systems","content":"On Tuesday, India and Russia signed an agreement to increase cooperation in areas of artificial intelligence, blockchain technology and explore the possibility of joint work in healthcare sectors. Apart from these areas, the two nations have also agreed to increase cooperation in quantum computing, financial technology and tourism. “Both sides agreed to explore joint working arrangements and pilot projects in healthcare, proposed setting up of a single-window clearance”, said the official statement which was released after the first India-Russia Strategic Economic Dialogue held last week at St.Petersburg. Led by Niti Aayog vice chairman Rajiv Kumar and Minister of Economic Development of the Russian Federation Maxim Oreshkin, the dialogue highlighted five core areas – transport infrastructure, agriculture and agro-processing sector, small & medium business support, digital transformation & frontier technologies and industrial & trade cooperation. Although the relation between the two sides has seen positive developments towards trade and defence in the past few years, technology like AI and blockchain will further reinforce the relationship. With Prime Minister Narendra Modi stressing on the technology frontier as an engine for economic growth in India, this dialogue will initiate more developments especially in the field of agriculture and tourism. Earlier in September, India approved a Memorandum of Understanding (MoU) brought out by BRICS for exploring research in distributed ledger technology such as blockchain, and see where it could be implemented for operational efficiency. In addition, the Indian state of Telangana also approved MoUs for implementing blockchain technology for smoothening government services. All of these developments tell us that emerging technologies like blockchain is set to propel India and Russia alike towards economic growth in both countries.","excerpt":"On Tuesday, India and Russia signed an agreement to increase cooperation in areas of artificial intelligence, blockchain technology and explore the possibility of joint work in healthcare sectors. Apart from these areas, the two nations have also agreed to increase cooperation in quantum computing, financial technology and tourism. “Both sides agreed to explore joint working […]","categories":["AI News"],"tags":["Technology"],"author_name":"Abhishek Sharma","publish_date":"2018-11-28T08:59:01","publication_year":"2018","word_count":273,"keywords":["Go","artificial intelligence","programming_languages:R","AI","digital transformation","programming_languages:Go","Git","Technology","GAN","R","emerging_tech:quantum computing"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","GAN","digital transformation","programming_languages:R","programming_languages:Go","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/india-and-russia-collaborate-to-strengthen-ai-and-blockchain-systems\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":54300,"title":"How Innoviti Is Using ML To Develop Intelligent Payments Solutions","content":"The payments industry has gone past the point in terms of evolution where one used to optimise transaction processing for reducing costs. Today, it is about how firms use payments data and build an end-to-end platform to help customers and merchants alike. Payments data is being used for enabling service providers to target users based on historical and behavioural information of users. However, a vast amount of opportunity in the offline market is left behind. It is still a challenge to make the most out of the transactional information and build products around it for providing a superior customer experience. To understand how Innoviti Payment Solutions is making strides in the payment landscape to deliver effective solutions, we interviewed Gaurav Mehrotra, VP & head of business data solutions, Innoviti Payment Solutions, for our deep dive column. He heads the business data group and SME lending platform. His team is focused on building data-driven solutions to identify revenue opportunities, optimise business operations, and prevent payments fraud. Before joining Innoviti, Gaurav spent 15+ years with Goldman Sachs and JP Morgan building data-driven digital solutions for their Asset and Wealth management business. The Payment Landscape “In the payment landscape operational part has been taken care of, now, probably is the time to build on intelligence systems that can be gleaned from the transaction data that is being generated,” begins Mehrotra. The challenge is how one can monetise and use it in business advantage, which will continue to be the trend in the payments landscape. Besides, due to the proliferation of various payments platforms, the need for expediting the integration for providing numerous option to purchasers have become paramount. Mehrotra says while it could be easy for big firms to integrate new payment services, small companies struggle to support all payment modes due to lack of resources. Consequently, there is a need for no-code or low code platforms to start accepting payments from numerous channels quickly. Continuing on the trends, Mehrotra also pinpointed that lending will evolve for both dealer and buyers. He believes understanding the requirement for credit for purchasers at the point of sale will be the gamechanger. Further, Mehrotra said there is a vast disjoint in B2C and B2B payments, resulting in the ineffective pipeline for the money movement. However, he thinks blockchain will play a crucial role in streamlining the money movement while bringing transparency. Innoviti Journey Started in 2002 as a product payment processing company, now Innoviti has evolved to add intelligence in payment systems using data. “We are evolving into a platform and enabling all stakeholders to play around; we call it the open platform, says Mehrotra. “About four years ago, we started venturing into value-added services like extensions of credit and SME lending. It is not a standalone platform but is integrated into the core payments pipe that we have. This enables us to generate loans based on transaction history.” And over the years, additional services have evolved to become a standalone business opportunity along with assisting in our central platform. Today, Innoviti processes over ₹30,000 crores of payments annually, which includes ₹1500 crores of credit. How Innoviti Is Leveraging AI In Their Solutions Powered by AI, Innoviti offers numerous products and services that cater to the need of customers and merchants in the payment landscape. The company helps in profiling users by identifying expense patterns and forecast future behaviours using AI. “We ensure there is a feedback loop in our solutions, which makes our products reliable,” adds Mehrotra. Innoviti assists its users by continuously monitoring and determining problems with the help of machine learning techniques. This enables the firms to proactively take actions even before the client reach out for help. Innoviti’s goal is to use data and serve clients before the clients start feeling the heat of their businesses. Problems are a part of a business, especially in the payment industry frauds are common, which afflicts the market. Therefore, Innoviti, with its Keystone platform, empowers companies to find potential miss use of cards. Such scams also cause hindrance in compliance, resulting in legal consequences. Consequently, the firm tightens the process around the settlements for providing peace of mind to clients. Besides, Innoviti brings in transparency by allowing clients to access their data, unearth insights and make informed decisions. “We are also working on NLP based platform that will eliminate the need for BI tools that require expertise to find insights. This will allow users to make queries in English for obtaining information,” unveils Mehrotra. Unique Value Proposition Talking about the differentiating factors, Mehrotra said that they have the most superior-tech platform, which he has been backed by the fact that the firm has multiple patents around payment processing and system design. Besides, the firm processes the payment within three seconds while the competitors do it in five to six seconds. Further, the company is working on improving the service to solve the hardware problems that its clients witness. Unlike the software problems that can be fixed remotely, issues in payment machines require physical presence. Thus, the firm is trying to enhance its service to fix hardware problems within two hours. Roadmap Evolving the product pipeline is one of the many core focus that Innoviti is trying to have a competitive advantage. The idea is to have a plug and play solution, which firms can utilise to integrate services effortlessly. “We are also working on improving our distributed sales and service networks across the country. A robust distributed channel will be vital for us to grow as well as help us serve our clients immediately,” says Mehrotra. Finally, he concluded by saying that the company will continue its efforts on moulding data to further add intelligence in our solutions.","excerpt":"The payments industry has gone past the point in terms of evolution where one used to optimise transaction processing for reducing costs. Today, it is about how firms use payments data and build an end-to-end platform to help customers and merchants alike. Payments data is being used for enabling service providers to target users based […]","categories":["Deep Tech"],"tags":["Deep Dive","Intelligent Agent"],"author_name":"Rohit Yadav","publish_date":"2020-01-17T13:51:50","publication_year":"2020","word_count":953,"keywords":["Go","machine learning","AI","ML","Git","RAG","NLP","Intelligent Agent","ViT","GAN","R","Deep Dive"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","RAG","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-innoviti-is-using-ml-to-develop-intelligent-payments-solutions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":43433,"title":"Book Excerpt: ‘Consumer Behavior – A Digital Native’ By Jagdish Sheth, Varsha Jain &#038; Don Schultz","content":"Consumer Behaviour – A Digital Native is the latest book by Jagdish Sheth of Professor at Emory University, Varsha Jain, Professor at MICA and Don Schultz, Professor at Northwestern University. Published by Pearson India, Sheth, Jain and Schultz’s book tries to capture the essence of the digital native’s perspective and successfully integrates various components of marketing communication. The book tries to demystify the fast-changing world of consumption and thus is an ambitious endeavour to offer a distinctive perspective integrating brand advertising in digital markets with consumer behaviour. To explain the concepts in a more clear manner, the authors have included interesting case studies, theoretical insights and real-life practice from some of the world’s foremost experts in marketing. We are reproducing an excerpt from the first chapter of this book with full permission from the publishers. You can purchase the book here. Chapter One: Understanding The Emergence Of Digital Native’s Behaviour Learning Objectives: After this chapter, you will be able to — 1-1 Explain the emergence of consumer behaviour 1-2 Understand the difference between digital natives and physical natives 1-3 Clarify the key dimensions of digital natives 1-4 Identify with the concepts of the future of consumer behaviour Opening Case: Nowadays, Indians have started opting for Zomato instead of going out to dine. This is one of the most extensively used apps in India, (though the company has a website as well, offering the same service). Why? They always provide a seamless technology-led dining experience. Zomato also provides a personalized and cloud-based system to figure out the best restaurants. The company has been able to weave technology into the dining experience of the user, cashing on the increasing popularity of Internet usage in India. Additionally, there has been an increasing number of smartphone users who are already using mobile applications (or apps) as these are convenient and can be easily accessed “on-the-go”. These apps provide relevant and personalized experiences based on consumers’ likings, preferences, and geographical locations. Zomato has developed extremely focused products that enhance the most optimum dining experience. The company also connects consumers with restaurants. Thus, consumers widely use the Zomato app to order food, make reservations, rate restaurant, and gather restaurant information. To sweeten this deal, the entire undertaking is based on a cashless system, that has been supplemented by online ordering and point-of-sale system. Together, they represent cutting-edge technology. In a short period of time, Zomato has become a part of the daily life of the digital natives as they prefer to make their all food and dining-related decisions on this app. Moreover, they also consider the ratings given on this platform to be credible and trustworthy. Additionally, the app is user-friendly; hence, it can be used by any digital native in the country (Bureau, 2015). One of the most distinctive elements of Zomato is that the company always listens to its consumers. They learn from individuals’ narratives about their pre- and post-dining experiences through Zomato. They engage with the consumer through social media channels as well, and respond promptly to the queries raised by the digital natives. In turn, the consumers provide them with instant feedback. Thus, it can be seen that technology has created an almost perfectly circular platform. Digital natives engage with Zomato in real time. In turn, the company ensures that consumers’ needs are met. As a result, digital natives like special events arranged by the company such as foodie meet-ups, blogger interactions, and feedback contests. Thus, Zomato has been able to maximize its brand recall and consequently, these consumers have become loyal to the company. To make this a seamless experience, Zomato also uses cloud computing, and it intends to create a virtual environment for the digital natives in future. They will also provide personalized content. Additionally Zomato connects with its loyal consumers on Omni channels in real time. The end result is that they create a wholesome dining experience on a single platform (Zomato: Catering the Hungry Millennial, 2016). Such an app experience has made a positive impact on the product quality that the company offers, which adds to its competitive advantage. However, the company still faces the major challenge of bridging the gap between consumers and restaurants thereby developing a seamless experience through the app. The heart of the challenge is to create memorable dining experiences. Digital natives are very demanding, and they have high aspirations and desires. The company, hence, is striving to provide consumers with information about restaurants and dining at their fingertips. We have now understood the usage of mobile apps in dining experience. In the next stage, we will try to comprehend the current trends in consumer behaviour and also understand how consumers will behave in future (Patidar, 2016). In order to understand the emergence of consumer behaviour, we need to comprehend the evolution of this stream along with its association with marketing and consumer-oriented concepts. Further, consumer behaviour became dynamic as the usage of technology increased, which led to the emergence of digital natives. This historical background would help us understand how technology has changed consumer behaviour, and digital natives have become important in the marketing domain. Emergence of Consumer Behavior Historically, traditional consumers have followed a very different behavior for buying the prod- ucts and services as there were no digital platforms. However, the emergence of technology and high usage of mobile phones has resulted into increased online presence of consumers, and they are using digital platforms for making the purchases. Thus, the core objective of this book is to understand how these digital natives behave in an online setup and make the purchase of products and services. The subsequent sections would provide further details about these areas. Evolution of Consumer Behavior Consumer behavior evolved between the 1850s and late 1920s, which is also referred to as the production era. During this time, demand exceeded supply; hence, the aim was to work on manufacturing, increasing the availability of products, and enhancing production capacity and capabilities. Interestingly, the consumers and manufacturers never focused on variations. They just aimed at high numbers in production. Subsequently, from the early 1930s to mid-1950s, the emphasis shifted on sales. This meant that the sale of products was more important than the production of products as there was a surplus of manufactured products. The products were simple and catered to the general requirements of the consumer. However, with the onset of the marketing era in the mid-1950s, satisfaction of specific needs and preferences of consumers, that is, an orientation towards “consumers first” came into vogue. For example, Parle introduced the first cola drink in India, which was known as Gloco Cola with an emphasis on taste and happiness. These two elements were used as the company considered consumers’ preferences and requirements as their prerogative (“The Story of Thums Up, Gold Spot & Limca—Guruprasad’s Portal”, 2017). Such a marketing orientation helped companies provide satisfaction to consumers, gain maximum profits, and work with a competitive edge. Hence, firms started exploring what the consumers would like well in advance so that the products and services could be offered at the right time. For example, Rexona claimed that using their soap would help women have lovely skin. This brand message created a significant impact as it demonstrated the company’s concern towards women. With a comprehensive understanding of the consumers’ specific needs and requirements, companies now started targeting particular groups or segments of consumers so that their specific needs and requirements were met by the products and services more appropriately. Thus, relevant marketing communications emerged based on consumers’ needs and requirements. This concept of targeting worked for 50 years. Gradually, societal marketing was developed implying that consumers’ immediate needs and wants were understood and addressed by companies. However, this entailed not much of a connection between the company and the long-term influencers of consumers, which included consumers’ family, friends, neighbors, and communities (Bhaskar, 2017). Emergence of Connections Between Consumer Behavior and Marketing Based on the above findings, it was felt that there was also the need for understanding the unsatisfied requirements. This was possible to gauge through market research. Consequently, research revealed that consumers have complex behavior, and this complexity existed because their psychological and social needs were synchronized with their functional needs. It was also found that there are specific needs and requirement for certain groups or segments of consumers. Thus, one product or service cannot satisfy all the needs and requirements of consumers as they are exclusive and distinctive. Thus, it was observed that products or services needed to be personalized to suit different consumer segments. To do so, studies were conducted in detail on consumer purchase and consumption. These studies led to the emergence of consumer research. This research helped in understanding consumers more comprehensively as companies had the tools and technique to carry out this process. This with a thorough understanding of consumers led to the emergence of marketing strategy (Rogoll, 2017). This strategy orientation understood consumers’ trends. It helped visualize future possibilities. Eventually, consumer research also found heterogeneity and similarities among consumers across different countries. For example, while the basic needs of the consumers remained the same, as every individual needs food, water, and shelter, such basic needs increased and became more complex due to the expanding cultural environment of consumers. This was further supplemented by the education and experience of individuals. There might even be a possibility that consumers develop similar needs. These similarities help develop promotions for products and services, leading to strategic orientation of products through segmentation, targeting, and positioning. While making these segments, key dimensions such as age, gender, family, social class, income, race, and ethnicity, geographical location, and lifestyles are considered by companies (“Predominant Components of Brand Equity …”, n.d.). For example, Parle Glucose has developed biscuits especially for children to give them strength and energy. Children like these biscuits as they are sweet. The parents are happy as they provide the right amount of glucose for the growth of their children. Similarly, a criteria like gender also affects consumer behavior. Generally, women like pink; men prefer white and blue. Thus, segments can be made in line with this generalization. For example, Lux soap used pink color to illustrate that this soap is for women. They even emphasized that the complexion of women grows with the use of this soap. Another key dimension, family structure, determines the marital status and spending capacity of consumers. Hence, these two factors also determine their behavior. For example, an unmarried person is likely to eat out more as compared to a married person. Social class and income also determine the behavior and consumption patterns of consumers especially in terms of their music, clothing, leisure, activities, hobbies, and art. Individuals always tend to socialize within the same social class. For example, the elite class likes to play golf at the premium clubs and socialize with the similar class of individuals. Race and ethnicity are other determining factors of consumer behavior. For example, Ponds use skin whitening for the women because in India, “beauty means white”, especially in the context women. Geographical location of consumers also determines segmentation. For example, Indian consumers from the northern part of the country are different from those from the south. Finally, lifestyle, as defined by activities carried out by the person in his spare time, is another deciding factor (Patel, 2017). Companies develop an image of products & services through segmentation and position them accordingly. This image is then created in the minds of the consumers. These segments help companies save their resources as they can satisfy a bigger group of consumers with the same products and services. For example, consumers want white and clean clothes, Tide, the detergent brand promoted this idea and satisfied these individuals effectively. Similarly, Colgate Gel targeted children and parents with the promise of white and clean teeth. Companies thereby develop an image of products and services through such segmentation and position them accordingly. This image is created in the minds of the consumers. The benefits of the products associated with their features are demonstrated. Thus, companies are able to develop a unique selling proposition (USP) that helps consumers differentiate the products of particular company from competing brands. This process results in the extension of new sizes, flavors, colors, and so on. Consumers are provided with options of a vast range of products from various firms. They are also able to understand the “me too” products, especially in the luxury segment, as they do not have any unique characteristics. For example, Louis Vuitton has many “me too” products in the Indian market as a majority of the consumers cannot afford to buy this brand (Kaushik, 2017). These processes have further led to marketing mix methods. These include all the four Ps—product, price, place, and promotion. Product offering to consumers includes ingredients, colors, sizes, shapes, features, characteristics, and so on. Price refers to the monetary payment that is to be paid by consumers, and it includes returns, discounts, allowance, and modes of payment; place denotes the store or non-store where products are sold; and promotion relates to advertising, public relation, and personal selling that develop awareness about products amongst consumers. Rasna, for example, is a well-known product in India. It is a drink with multiple flavors, reasonably priced, and easily available at different prominent stores in the country. It was promoted by a sweet young girl whose statement “I love you Rasna” had gained instant popularity (Harish, 2017).","excerpt":"Consumer Behaviour – A Digital Native is the latest book by Jagdish Sheth of Professor at Emory University, Varsha Jain, Professor at MICA and Don Schultz, Professor at Northwestern University. Published by Pearson India, Sheth, Jain and Schultz’s book tries to capture the essence of the digital native’s perspective and successfully integrates various components of […]","categories":["IT Services"],"tags":["best book to learn digital marketing","Data Analytics","Data Science","digital marketing","Zomato"],"author_name":"Prajakta Hebbar","publish_date":"2019-07-28T15:00:27","publication_year":"2019","word_count":2237,"keywords":["Go","ELT","AI","cloud computing","ML","Git","Zomato","digital marketing","Aim","ViT","Rust","Data Analytics","best book to learn digital marketing","Data Science","R"],"extracted_tech_keywords":["AI","ML","Aim","cloud computing","R","Go","Rust","Git","ELT","ViT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/book-excerpt-consumer-behavior-a-digital-native-by-jagdish-sheth-varsha-jain-don-schultz\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10143365,"title":"Google Launches Gemini 2.0 Making the Age of AI Agents a Reality","content":"As predicted by AIM, Google has finally launched Gemini 2.0, its next-generation AI model, built to redefine multimodal capabilities and introduce agentic functionalities. “Today we’re excited to launch our next era of models built for this new agentic era: introducing Gemini 2.0, our most capable model yet. With new advances in multimodality — like native image and audio output — and native tool use, it will enable us to build new AI agents that bring us closer to our vision of a universal assistant,” Google said in the blog post. “This is really just the beginning. 2025 will be the year of AI agents and Gemini 2.0 will be the generation of models that underpin our agent-based work,” said Google DeepMind chief Demis Hassabis. Gemini 2.0 Flash supports multimodal inputs, including images, video, and audio, as well as multimodal outputs such as natively generated images combined with text and steerable text-to-speech (TTS) multilingual audio. It can also natively call tools like Google Search, execute code, and integrate third-party user-defined functions. The Gemini 2.0 Flash model offers faster response times and outperforms its predecessors on major benchmarks. Developers can access Gemini 2.0 Flash through Google AI Studio and Vertex AI, with general availability expected by January 2025. Google has also launched the Multimodal Live API, bringing real-time audio and video input capabilities that allow developers to create dynamic, interactive applications. Project Astra and AI Agents Introduced at Google I\/O 2024, Google Project Astra, a universal AI assistant, has received several updates. It now supports multilingual and mixed-language conversations, with an improved understanding of accents and uncommon words. Powered by Gemini 2.0, Project Astra can also utilise Google Search, Lens, and Maps, making it a more practical assistant for daily tasks. Its memory has been enhanced, allowing up to 10 minutes of in-session recall and better personalisation through past interactions. Additionally, improved streaming and native audio processing reduce latency, enabling near-human conversational speeds. Google has also announced an early-stage research prototype, Project Mariner, which will understand and reason based on information that can be accessed while a user navigates on a web browser. Google says that the agent uses information it sees on the screen through a Google Chrome extension to complete related tasks. The agent will be able to read information, like text, code, images, forms and even voice-based instructions. “Book a flight from SF to Berlin, departing on March 5 and returning on the 12. The era of being able to give a computer a fairly complex high-level task and have it go off and do a lot of the work for you is becoming a reality,” said Jeff Dean, chief scientist at Google DeepMind. Google has also introduced Jules, a developer-focused agent, that integrates with GitHub workflows to assist with coding tasks under supervision. Google DeepMind is working on AI agents that improve video games and navigate 3D worlds. It has partnered with game developers like Supercell to explore the future of AI-powered gaming companions. Gemini 2.0’s spatial reasoning is also being tested in robotics for practical real-world use. Notably, it recently launched Genie 2, a large-scale foundation world model capable of generating a wide variety of playable 3D environments.","excerpt":"2025 will be the year of AI agents and Gemini 2.0 will be the generation of models that underpin our agent-based work.’","categories":["AI News"],"tags":["Google"],"author_name":"Siddharth Jindal","publish_date":"2024-12-11T23:44:49","publication_year":"2024","word_count":531,"keywords":["Go","API","Gemini 2.0","TPU","AI","Git","Google AI Studio","Aim","Google","GitHub","R"],"extracted_tech_keywords":["AI","Gemini 2.0","Aim","TPU","R","Go","Git","GitHub","API","Google AI Studio"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-launches-gemini-2-0-making-the-age-of-ai-agents-a-reality\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":60247,"title":"‘Cab Driver Cried in Front of Me’ A New Way For People On LinkedIn To Show Off Their ‘Humanity’ During COVID-19 Pandemic","content":"With the coronavirus outbreak in hand, the Indian government has recently put lockdown in place, where people are urged to stay back at home in a bid to contain the spread. At this global crisis, the worst-hit sector of the country is the daily wage earners, who usually run their life on their daily salary. With the imposed lockdown they are neither getting jobs nor able to earn money to feed their family. As an act of sympathy and humanity, many online campaigns are going on for various reasons, such as to generate funds or to provide necessities to affected areas. However, users are using a new template. “The Cab Driver Cried in Front of Me….” [post continued.] as a typical story being shared across social media, especially LinkedIn and Twitter, users showing off their “humanity” during the pandemic. The post states — “A cab driver cried in front of me saying I was his first customer in the last 48 hours. He said his wife is expecting groceries today at least. This virus is gonna hit us in so many ways but the people who depend on daily income are gonna get hit the most. I gave the guy an extra 500. Obviously, it’s not a big deal for most of us which mean we should do it more. He showed me that he has been driving around for 70 kilometres since his last customer. Please pay your cab drivers, street vendors, etc a little more. You might just be their last customer for the day.” People on the internet are simply copying and pasting this post on their feed to showcase, how they are helping daily wage earners, in this case, the cab drivers. The post started to get viral when first, Actress Kajal Aggarwal, shared this emotional post on her Instagram story. Here is the New Template: The cab driver cried in front of me saying I was his first customer in the last 48 hours. He said his wife is expecting groceries today atleast. This virus is gonna hit us from so many ways but the people who depend on daily income are gonna get hit the most.— Aqib Mughal (@aqibmughal89) March 19, 2020 And, it continues, https:\/\/twitter.com\/novorapid_\/status\/1243201711218581504 More people are copying the same post. The cab driver cried in front of me saying i was his 1stcustomer in the whole day, he said he is expecting to do groceries today Atleast. This virus is going to hit is in different ways, but the people who depend on daily income are gonna get hit the most. #COVID19 #coronavirus— Muhib 木黑محب شینواری (@M_shinwariAf) March 18, 2020 Also this Crazy how many taxi drivers got tearful last night pic.twitter.com\/TJfQzTGtqJ— The State of LinkedIn (@StateOfLinkedIn) March 26, 2020 https:\/\/www.linkedin.com\/feed\/update\/urn:li:activity:6648845327455567872\/","excerpt":"With the coronavirus outbreak in hand, the Indian government has recently put lockdown in place, where people are urged to stay back at home in a bid to contain the spread. At this global crisis, the worst-hit sector of the country is the daily wage earners, who usually run their life on their daily salary. […]","categories":["AI Features"],"tags":["Coronavirus","covid-19"],"author_name":"Sejuti Das","publish_date":"2020-03-27T17:26:57","publication_year":"2020","word_count":460,"keywords":["Go","API","covid-19","AI","programming_languages:R","programming_languages:Go","Coronavirus","ViT","R"],"extracted_tech_keywords":["AI","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cab-driver-cried-in-front-of-me-a-new-way-for-people-on-linkedin-to-show-off-their-humanity-during-covid-19-pandemic\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162818,"title":"Deeptech Fund of Funds – A Game Changer for Talent Management in GCCs","content":"The Union Budget 2025-26 has introduced several initiatives specifically aimed at powering India’s technology ecosystem and ensuring talent development. One of the standout measures is the Deeptech Fund of Funds (FoF), which, along with a ₹10,000 crore boost for startups, is set to transform Global Capability Centres (GCCs) by providing access to skilled professionals in fields like artificial intelligence (AI), cybersecurity, and cloud computing. This particular measure is essential, given that over 1,580 GCCs in India employ 1.66 million people, and the number is only increasing. However, many industry leaders and experts have stressed that finding the right talent has become a bigger challenge than building office spaces, and some GCCs are also suffering due to hiring challenges. Aparna Iyer, CFO at Wipro Limited, said, “With AI becoming an essential tool for a tech-powered economy, it is heartening to see the introduction of Centres of Excellence (CoE) for AI in education. To position India as a leader in the global AI race, it is imperative to prioritise investment in STEM (science, technology, engineering and mathematics) talent.” Bridging the Skill Gap in GCCs Formulating a national framework to guide states in promoting GCCs is undoubtedly a step in the right direction. Suresh Ramamoorthy, country head of Lingaro India, stated, “The…national framework for GCCs…will provide a structured roadmap for one of the fastest-growing sectors in India.” He added that with over 1,700 GCCs employing nearly 1.9 million professionals, the national framework for GCCs will help strengthen India’s position as a hub for advanced technology solutions while also helping drive deeper collaboration between industry and academia. GCCs have already become an integral part of India’s tech industry, with companies setting up innovation hubs to develop AI-driven solutions, automation systems, and next-generation software. However, hiring the right talent has been a major challenge, as pointed out by many industry leaders. The deeptech FoF is expected to address this by enhancing skill development programs and creating a strong talent pipeline. “With strategic investments in education, skilling, and inclusion, the Union Budget 2025-26 lays the foundation for a future-ready workforce and an inclusive growth story,” said Dhriti Prasanna Mahanta, vice-president of  TeamLease Degree Apprenticeship. He also emphasised that the 10,000 fellowships for tech research at IITs and IISc under the Prime Minister’s Research Fellowship (PMRF) Scheme will help build a talent pool ready for deep-tech innovations. However, he also suggested that to maximise impact, the program must integrate industry internships similar to a Prime Minister’s Internship Scheme (PMIS). “Engaging the top 20 Indian tech companies to co-create flexible, agile, and responsive training modules will further enhance the program’s effectiveness,” Mahanta added. By collaborating with global tech giants, these fellowships could also offer international internship opportunities and enhance exposure to real-world AI applications. How AI Internships Can Shape Future Talent The National Education Policy (NEP) 2020 has already introduced a credit-based system for future skill courses like AI and machine learning (ML). AI internships are now a core part of practical learning, where students gain hands-on experience with advanced AI tools and industry-led projects. According to Mahanta, these internships offer an average stipend of up to ₹35,000 and upon completion, interns can command salaries between ₹8-10 lakh per annum. This clearly indicates that specialised AI skills will be in high demand, making such programs crucial for GCCs looking to build AI-driven solutions. GCCs Expanding to Tier-2 Cities While Bengaluru and Hyderabad have traditionally been India’s top GCC hubs, the government is now encouraging their expansion to tier-2 cities. With improved infrastructure and new talent pools, cities like Chandigarh and Indore are emerging as potential hubs. Vineet Dhawan, CEO of Digital Convergence Technologies (DCT), welcomed this development, stating, “We are excited to see how the national framework for promoting GCCs in emerging tier-2 cities will encourage talent availability, infrastructure improvement, and industry outreach.” He also noted that the ₹20,000 crore allocated for private-sector research, development, and innovation will accelerate projects in AI, cloud computing, and cybersecurity, further fueling talent demand. The Union Budget 2025 underscores the government’s proactive efforts to drive the expansion of GCCs in India, particularly in emerging tier-2 cities. Chandan Barve, VP and chief administrative officer of Sun Life Global Solutions, said that this initiative, with its strong focus on talent availability, infrastructure development, and regulatory reforms, will create new opportunities for both organisations and the country’s employable youth. He further added that the commitment to a national framework for GCCs strengthens India’s role as a “global hub fostering deeper innovation”, which will deliver state-of-the-art technology services, and drive operational excellence. In addition, Garima Mitra, co-founder of Treelife, highlighted in her LinkedIn post that the five-year extension of the sunset clause in GIFT City for tax-neutral relocation of foreign funds reflects a strong positive sentiment about the government’s intent to develop GIFT City. She added that tax clarity for alternative investment funds (AIFs) and the removal of tax collected at source (TCS) on securities will create a more streamlined investment ecosystem, reinforcing the government’s commitment to fostering a more investor-friendly environment. By combining fellowships, AI internships, industry collaborations, and startup support, India is taking strong steps toward becoming a global leader in deeptech innovation. With government initiatives, industry partnerships, and startup growth, the future of talent management in India’s GCCs looks promising.","excerpt":"This measure is essential, given that over 1,580 GCCs in India employ 1.66 million people.","categories":["GCC"],"tags":["deeptech","GCC india","union budget"],"author_name":"Shalini Mondal","publish_date":"2025-02-04T11:59:54","publication_year":"2025","word_count":877,"keywords":["Go","artificial intelligence","machine learning","AI","cloud computing","deeptech","ML","Git","RAG","Aim","union budget","GCC india","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","cloud computing","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/gcc\/deeptech-fund-of-funds-a-game-changer-for-talent-management-in-gccs-2\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10093975,"title":"Why You Can Never Imitate ChatGPT","content":"Ever since the release of ChatGPT, people have been obsessed with the LLM chatbot. There are users, doomers, and then there are developers and other AI companies that have been trying to build their own version of it. While OpenAI’s model is closed source, a lot of open source models have been giving developers hope to build something similar to probably the best chatbot in the market. One of the emerging inexpensive methods to improve an open source language model such as LLaMa, Alpaca, or Self-Instruct is to fine-tune it on outputs from proprietary systems like ChatGPT, which are stronger models. It might seem like an efficient way to imitate and build up the weaker model’s capabilities to match the stronger model, but it fails big time. Inheriting Flaws Researchers from UC Berkeley recently published a paper – The False Promise of Imitating Proprietary LLMs – which critically analysed the efficacy of this imitation approach. The researchers explain how training models using proprietary language models raises various legal and ethical concerns. Talking about the technical improvement that this approach achieves is also negligible. According to the paper, fine-tuning a weaker model to improve the knowledge capabilities has little to no impact on the language model. Pre-training is the main source for improving the capabilities of a language model. Thus, fine-tuning smaller models like LLaMa, Vicuna, Alpaca, and Self-Instruct, on the output of large models like ChatGPT and Bard, does not improve the knowledge of the model because the base data remains unchanged. It only alters the style of the model. Moreover, when trying to imitate the knowledge and capabilities of large models like ChatGPT through their outputs, weaker models also end up inheriting their flaws and biases. Fine-tuning of these models through imitation also removes the capability of directly improving the design decisions of companies that have closed AI models, such as ChatGPT, making the models perform even poorer than them. The models fail to improve on important aspects like factuality, coding, and problem solving. Data imitation not a good idea Big companies that have large base models like GPT-4 or Google’s PaLM, have no reason to worry about imitation as there is a huge gap between them and the models trying to imitate them. Companies that acquire large amounts of data, compute, and algorithmic advances are much likely to maintain their competitive advantages. Smaller companies that are trying to establish their moat by utilising off-the-shelf LLMs such as LLaMa or other open source offerings are at more risk of imitation. As explained earlier, fine-tuning with ChatGPT data does not make as much improvement in language models as building and improving the pre-training data by the company itself. Still, companies such as OpenAI have never openly disclosed the data that they are trained on. It is safe to say that a lot of it is just open internet data and not their proprietary data. The paper by UC Berkeley, as pointed out by users on Hacker News, said that it might be illegal to use the data, which might not be correct. The paper draws various conclusions that are not based on the current developments in open source software. Before the release of LLaMa, it was believed that the larger the model, the better it performs. But various models built on top of a smaller model like LLaMa are outperforming GPT-4 and PaLM. What if data gets private? In Google’s leaked document where it said that neither them nor does OpenAI have a moat in AI, the document praised the open source community and in many ways Meta’s LLaMa. Even though Google and OpenAI do not have proprietary rights to the data that they are trained on, the future of language models might not remain the same, and get rather scary. Recently, Mark Cuban, the American businessman and AI enthusiast, said that the next step for LLM based chatbots by big companies is to have their private data that can be bought to build exclusive large knowledge models, instead of LLMs. These models would be able to perform better and act like a moat for the companies as no one would have access to the data. “We ignore what created us; we adore what we create.” — Aleister Crowley, The Book of Lies We already have an example of how Elon Musk, the chief twit, stopped OpenAI from accessing its data and threatened to file a lawsuit against it and Microsoft for still using it. Now, his new chatbot is in the line, which is possibly trained on Twitter data, which is exclusive to him. Not even the open source can access it. What this means for open source is that if developers rely on off the shelf language models like LLaMa, they would not be able to fine-tune it on any other model that is built by the large companies. Currently there are no legal or ethical issues around using data from ChatGPT or Bard to fine-tune your models as the data does not belong solely to the companies. But if Cuban is right, and companies start buying exclusive access to data, then the open source community would stop thriving. These companies might not want to allow anyone else to compete, and with the open source rising up, restricting data access might be the biggest move for them. Currently, one of the important things for smaller models is their specific use case capabilities. But if big companies acquire exclusive rights to other companies’ data, and make it private, they might be able to outperform smaller models easily. No one would reveal or release their data source, but it is clearly the internet. If these giants start acquiring intellectual property rights to data from other companies, they would rise as the most knowledgeable models in the community, and the open source community might die.","excerpt":"Developers and AI companies have been obsessed with ChatGPT and are trying to build their own version of it, but will never be able to.","categories":["AI Features"],"tags":["ChatGPT","GPT-4","Mark Cuban"],"author_name":"Mohit Pandey","publish_date":"2023-05-26T14:00:00","publication_year":"2023","word_count":974,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","chatbots","AWS","Mark Cuban","GPT-4","GPT","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","chatbots","AWS","TPU","R","Go","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-you-can-never-imitate-chatgpt\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":14413,"title":"10 books on AI and Robots bound to be a captivating read","content":"Rise of the Robots Robots and Artificial Intelligence are the technologies powering the world of tomorrow and have captured various aspects of our lives. Although, it’s only recently that they have started to show a rampant presence, the idea has been captured in the form of novels, fiction, academia, and plays since time immemorial. AIM list down some of the most brilliant work on AI and robots which have received rave reviews globally. To make it more interesting for readers, we have included both the work of fiction and nonfiction, listed as below- Fiction Books- 1. 2001: A Space Odyssey 2001 A Space Odyssey An Arthur C. Clarke special, the book was written concurrently with Stanley’s Kubrick’s famous film by the same name. Clarke and Kubrick worked on the book together, however, only Clarke ended up being the official author. The book was written in the year 1968. The AI program in this book is named HAL, which is fixed aboard the spaceship Discovery One. Some readers viewed HAL as a robot, then AI, as the whole ship could be considered the AI’s body. HAL is also regarded as one of the greatest creation of science fiction. 2. Cyborg Artwork for the novel Cyborg The story is about an astronaut-turned-test pilot, Steve Austin, who experiences a catastrophic crash during a flight. This incident leaves the protagonist with one limb, a blind eye, and other major injuries. “The Six Million Dollar Man” is the movie rendition to this book. As the name suggests, the book essentially revolves around Steve Austin turning into a cyborg, after the accident. Martin Caidin wrote this book in 1972, and involves the kind of spy adventure usually associated with Steve Austin. 3. Daemon Daniel Suarez self-published his first novel, Daemon Created by Daniel Suarez in 2006, the book has been written in tandem with another novel called Freedom. Both Daemon and Freedom together comprise two-part novel describing a distributed, persistent computer application, known as The Daemon. After the original programmer’s death, the program begins to transform the real world. The book projected Daemon as corrupting and killing humans, besides running independently of human control. Only Detective Peter Sebeck has what it takes to protect the world and free it from the control of a virtual enemy. Readers who appreciate thrills and cyber suspense will find Daemon a delightful read. 4. Do Androids Dream of Electric Sheep? Do Androids Dream of Electric Sheep? Blade Runner is the famous movie adapted from this popular Philip K. Dick classic, published in 1968, about a bounty hunter, Rick Deckard. He is assigned with blowing away six escaped Nexus-6 brain model androids, the latest and most advanced model. There’s a secondary plot in the book running parallelly, which follows John Isidore, a man of sub-par IQ. Isidore takes the responsibility of helping the fugitive robots. The novel essentially probes the defining qualities that separate humans from androids, through the character Deckard. 5. Excession Excession Written by Iain M. Banks in 1996, the plot largely revolves around the response of Minds (AIs with enormous intellectual and physical capabilities and distinctive personalities) to the Excession (a mysterious alien artifact). The story also focuses on the brutal society, and how it tries to use the Excession to increase its power. The protagonist, Diplomat Byr Genar-Hofoen has been selected and sent off to investigate a 2,500-year-old mystery which involves the sudden disappearance of a star fifty times older than the universe itself. Byr risks losing himself during the mission. Non-fiction Books- 1. Beyond AI: Creating the Conscience of the Machine Excession J. Storrs Hall written Beyond AI discusses some of the most specific topics and concerns around AI. The book delves into AI’s ability to create novels or formulate laws in the near future. The book was published in 2007. In his book, Hall reviews the history of AI, explaining the most significant roadblocks that the field has recently overcome, while predicting the probable achievements in the near future. Hall provides an intriguing glimpse into the possibilities that future might have in store for us, combining technologies such as cybernetics, computer science, psychology, philosophy of mind, neurophysiology, game theory, and economics and dilemmas on the horizon. 2. Humans Need Not Apply – A Guide to Wealth and Work in the Age of Artificial Intelligence Humans Need Not Apply This Jerry Kaplan classic, written in 2015 has been selected as one of the 10 best science and technology books of 2015 by The Economist. Through the book, Kaplan talks about the latest advances in robotics, machine learning, and perception powering systems that rival or exceed human capabilities. Kaplan also mention that the transition might be drastic and brutal, unless we address the two great aspects of the modern developed world: volatile labor markets and income inequality. His take on AI is a must-read for business leaders. This book increasingly touches upon the growing concerns surrounding the rapid progress in the Artificial Intelligence space. 3. Introduction to Autonomous Mobile Robot Introduction to Autonomous Mobile Robot This book was written by Roland Siegwart, offering students and other prospective readers an introduction to the very basics of mobile robotics. The text focuses on mobility itself, furnishing an overview mechanisms that allow a mobile robot to move through a real-world environment. Interactive modules have been used throughout the book to present the technology that enables mobility, and each chapter represents a different aspect about mobility. Furthermore, it spans across all aspects surrounding mobile robotics, including software and hardware design considerations, related technologies, and algorithmic techniques. 4. Rise of the Robots: Technology and the Threat of a Jobless Future Rise of the Robots A Martin Ford non-fiction, written in 2015, this book takes an intimidating tour of AI’s rapid advances. The book contains terrifying wealth of economic data from both the US and the UK, to highlight societal implications of the robots’ rise. The book also describes how jobs in education, finance, and healthcare can be easily automated today. Essentially, the book delves into what accelerating technological growth would implicate for us on an economic level, and for the upcoming generations and society as a whole. This Martin-Ford classic is another must-read. 5. Superintelligence: Paths, Dangers, Strategies Superintelligence Nick Bostrom’s Superintelligence lays the foundation for understanding the future of humanity and intelligent life. In his book, Bostrom mentions if machine brains surpassed human minds, then this new superintelligence can be extremely powerful. He asks questions crucial to the survival of human life: what happens when machines surpass humans in general intelligence? Will AI save or destroy us? You must read the book to find out more.","excerpt":"Robots and Artificial Intelligence are the technologies powering the world of tomorrow and have captured various aspects of our lives. Although, it’s only recently that they have started to show a rampant presence, the idea has been captured in the form of novels, fiction, academia, and plays since time immemorial. AIM list down some of […]","categories":["AI Trends"],"tags":["AI Books","Artificial Intelligence India","latest advances","robots India"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-04-21T05:04:06","publication_year":"2017","word_count":1105,"keywords":["Go","API","AI Books","artificial intelligence","machine learning","AI","AWS","R","RPA","latest advances","Artificial Intelligence India","Git","Aim","robots India"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","AWS","R","Go","Git","API","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-books-ai-robots-bound-captivating-read\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10009353,"title":"TinyML And Its ‘Great’ Application in IoT Technology","content":"Tiny machine learning (TinyML) is an embedded software technology that can be used to build low power consuming devices to run machine learning models. It is also more famously referred to as the missing link between device intelligence and edge hardware. It makes computing at edge cheaper, less expensive, and more stable. Further, TinyML also facilitates improved response time, privacy, and low energy cost. TinyML is massively growing in popularity with every passing year. As per ABI Research, a global tech market advisory firm, by 2030, about 230 billion devices will be shipped with TinyML chipset. TinyML has the ability to provide a range of applications, from imagery micro-satellite, wildfire detection, and for identifying crop ailments and animal illness. Another area of application that is drawing great attention is its application in IoT devices. TinyML and IoT TinyML brings ultra-low-power systems and machine learning communities together; this paves the way for more exciting on-device machine learning. TinyML is placed at the intersection of embedded machine learning applications, algorithms, hardware, and software. As compared with a desktop CPU, which consumes 100 watts of power, TinyML just required a few milliwatts of battery power. With such a major advantage, TinyML can provide great longevity to always-on ML applications at the edge\/endpoint. Currently, there are 250 billion microcontrollers in the world today. This number is growing by 30 billion annually. The reason for its pervasiveness is that, firstly, it gives small devices the ability to make smart decisions without needing to send the data to the cloud. Further, TinyML models are small enough to fit into almost any environment. Taking the example of an imagery micro-satellite which are required to capture high-resolution images but are restricted by the size and number of photos they can transmit back to Earth. With TinyML, however, the microsatellite only captures an image if there was an object of interest such as a ship or weather pattern. TinyML has the potential to transform the way one deals with IoT data, where billions of tiny devices are already used to provide greater efficiency in fields of medicine, automation, and manufacturing. It is very important to make a clear distinction between ‘serving’ machine learning to IoT and ‘developing’ machine learning inside the IoT devices. In the former, the machine learning tasks are outsourced to the cloud, while the IoT device waits for the execution of intelligent services, however, in latter, TinyML-as-a-service is employed, and the IoT device is part of the execution of the services. The TinyML represents a connecting point between the IoT devices and the ML. The hardware requirements for machine learning in larger systems are analogous to TinyML in smaller IoT. As the size of IoT devices hitting the market increase, we could see even higher investment in terms of research in TinyML, exploring concepts such as deep neural networks, model compression, and deep reinforcement learning. The Challenges There are a few challenges of integrating TinyML in the IoT devices; some of them are: Overcoming the technical challenges within edge computing The differences between web-based and embedded technologies in terms of deployment and execution.The computational resource that is required for delivering an accurate and reliable output. Wrapping Up Speaking in detail about the applications of TinyML, it can be used in sensors for real-time traffic management and ease of urban mobility; in manufacturing, TinyML can be used to enable real-time decision making to identify equipment failure. The workers can be alerted to perform preventive maintenance based on the equipment conditions; TinyML can also be used in the retail business for monitoring the availability of the resource. TinyML is gaining its ground but is still in a very nascent stage. It is expected to take over space with inter-sector applications very soon.","excerpt":"Tiny machine learning (TinyML) is an embedded software technology that can be used to build low power consuming devices to run machine learning models. It is also more famously referred to as the missing link between device intelligence and edge hardware. It makes computing at edge cheaper, less expensive, and more stable. Further, TinyML also […]","categories":["AI Features"],"tags":["IoT","Machine Learning","TinyML"],"author_name":"Shraddha Goled","publish_date":"2020-10-09T18:00:36","publication_year":"2020","word_count":624,"keywords":["TinyML","Go","machine learning","TPU","AI","neural network","ML","Machine Learning","automation","ViT","edge computing","R","IoT"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","edge computing","TPU","R","Go","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/tinyml-and-its-great-application-in-iot-technology\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":50145,"title":"Three Things to Know About Reinforcement Learning","content":"If you are following technology news, you have likely already read about how AI programs trained with reinforcement learning beat human players in board games like Go and chess, as well as video games. As an engineer, scientist, or researcher, you may want to take advantage of this new and growing technology, but where do you start? The best place to begin is to understand what the concept is, how to implement it, and whether it’s the right approach for a given problem. If we simplify the concept, at its foundation, reinforcement learning is a type of machine learning that has the potential to solve tough decision-making problems. But to truly understand how it impacts us, we need to answer three key questions: What is reinforcement learning and why should I consider it when solving my problem?When is reinforcement learning the right approach? What is the workflow I should follow to solve my reinforcement learning problem? What Is Reinforcement Learning? Reinforcement learning is a type of machine learning in which a computer learns to perform a task through repeated trial-and-error interactions with a dynamic environment. This learning approach enables the computer to make a series of decisions that maximize a reward metric for the task without human intervention and without being explicitly programmed to achieve the task. To better understand reinforcement learning, let’s look at a real-world equivalent situation. Figure 1 shows a general representation of training a dog using reinforcement learning. (Figure 1) Reinforcement learning in dog training The goal of reinforcement learning in this case is to train the dog (agent) to complete a task within an environment, which includes the surroundings of the dog as well as the trainer. First, the trainer issues a command or cue, which the dog observes (observation). The dog then responds by taking an action. If the action is close to the desired behavior, the trainer will likely provide a reward, such as a food treat or a toy; otherwise, no reward or a negative reward will be provided. At the beginning of training, the dog will likely take more random actions like rolling over when the command given is “sit,” as it is trying to associate specific observations with actions and rewards. This association, or mapping, between observations and actions is called policy. From the dog’s perspective, the ideal case would be one in which it would respond correctly to every cue, so that it gets as many treats as possible. So, the whole meaning of reinforcement learning training is to “tune” the dog’s policy so that it learns the desired behaviors that will maximize some reward. After training is complete, the dog should be able to observe the owner and take the appropriate action, for example, sitting when commanded to “sit” by using the internal policy it has developed. By this point, treats are welcome but shouldn’t be necessary (theoretically speaking!). Based on the dog training example, consider the task of parking a vehicle using an automated driving system (Figure 2). The goal of this task is for the vehicle computer (agent) to park the vehicle in the correct parking spot with the right orientation. Like the dog training case, the environment here is everything outside the agent and could include the dynamics of the vehicle, other vehicles that may be nearby, weather conditions, and so on. During training, the agent uses readings from sensors such as cameras, GPS, and lidar (observations) to generate steering, braking, and acceleration commands (actions). To learn how to generate the correct actions from the observations (policy tuning), the agent repeatedly tries to park the vehicle using a trial-and-error process. A reward signal can be provided to evaluate the goodness of a trial and to guide the learning process. (Figure 2) Reinforcement learning in autonomous parking In the dog training example, training is happening inside the dog’s brain. In the autonomous parking example, training is supervised by a training algorithm. The training algorithm is responsible for tuning the agent’s policy based on the collected sensor readings, actions, and rewards. After training is complete, the vehicle’s computer should be able to park using only the tuned policy and sensor readings. When Is Reinforcement Learning the Right Approach? Many reinforcement learning training algorithms have been developed to date; this article does not cover training algorithms, but it is worth mentioning that some of the most popular ones rely on deep neural network policies. The biggest advantage of neural networks is that they can encode really complex behaviors, which opens up the use of reinforcement learning in applications that are otherwise intractable or very challenging to tackle with alternative methods, including traditional algorithms. For example, in autonomous driving, a neural network can replace the driver and decide how to turn the steering wheel by simultaneously looking at input from multiple sensors, such as camera frames and lidar measurements (end-to-end solution). Without neural networks, the problem would normally be broken down into smaller pieces: a module that analyzes the camera input to identify useful features, another module that filters the lidar measurements, possibly one component that would aim to paint the full picture of the vehicle’s surroundings by fusing the sensor outputs, a “driver” module, and more! However, the benefit of end-to-end solutions comes with a few drawbacks. A trained deep neural network policy is often treated as a “black box,” meaning that the internal structure of the neural network is so complex, often consisting of millions of parameters, that it is almost impossible to understand, explain, and evaluate the decisions taken by the network (left side of Figure 3). This makes it hard to establish formal performance guarantees with neural network policies. Think of it this way: Even if you train your pet, there will still be occasions when your commands will go unnoticed. (Figure 3) Some of the challenges of reinforcement learning Another thing to keep in mind is that reinforcement learning is not sample efficient. This means that, in general, a lot of training is required in order to reach acceptable performance. As an example, AlphaGo, the first computer program to defeat a world champion at the game of Go, was trained nonstop for a period of a few days by playing millions of games, accumulating thousands of years of human knowledge. Even for relatively simple applications, training time can take anywhere from minutes to hours or days. Finally, setting up the problem correctly can be tricky; many design decisions need to be made, which may require a few iterations to get it right (right side of Figure 3). These decisions include, for example, selecting the appropriate architecture for the neural networks, tuning hyperparameters, and shaping the reward signal. In summary, if you are working on a time- or safety-critical project that you could potentially approach with alternative, traditional ways, reinforcement learning may not be the best thing to try first. Otherwise, give it a go! Reinforcement Learning Workflow The general workflow for training an agent using reinforcement learning includes the following steps (Figure 4). (Figure 4) Reinforcement learning workflow 1. Create the Environment First you need to define the environment within which the agent operates, including the interface between agent and environment. The environment can be either a simulation model or a real physical system. Simulated environments are usually a good first step since they are safer (real hardware is expensive!) and allow experimentation. 2. Define the Reward Next, specify the reward signal that the agent uses to measure its performance against the task goals and how this signal is calculated from the environment. Reward shaping can be tricky and may require a few iterations to get right. 3. Create the Agent In this step, you create the agent. The agent consists of the policy and the training algorithm (refer back to Figure 2), so you need to: a. Choose a way to represent the policy (e.g., using neural networks or look-up tables). b. Select the appropriate training algorithm. Different representations are often tied to specific categories of training algorithms, but in general, most modern algorithms rely on neural networks because they are good candidates for large state\/action spaces and complex problems. 4. Train and Validate the Agent Set up training options (e.g., stopping criteria) and train the agent to tune the policy. Make sure to validate the trained policy after training ends. Training can take anywhere from minutes to days depending on the application. For complex applications, parallelizing training on multiple CPUs, GPUs, and computer clusters will speed things up. 5. Deploy the Policy Deploy the trained policy representation using, for example, generated C\/C++ or CUDA code. No need to worry about agents and training algorithms at this point—the policy is a standalone decision-making system! Training an agent using reinforcement learning is an iterative process. Decisions and results in later stages can require you to return to an earlier stage in the learning workflow. For example, if the training process does not converge on an optimal policy within a reasonable amount of time, you may have to update any of the following before retraining the agent: Training settingsLearning algorithm configurationPolicy representationReward signal definitionAction and observation signalsEnvironment dynamics Today, tools like Reinforcement Learning Toolbox (Figure 5) can help you quickly learn and implement controllers and decision-making algorithms for complex systems such as robots and autonomous systems. (Figure 5) Teaching a robot to walk with Reinforcement Learning Toolbox Regardless of the choice of tool, before deciding to adopt reinforcement learning, do not forget to ask yourself: “Given the time and resources I have for this project, is reinforcement learning the right approach for me?” To learn more about reinforcement learning, see the links below. Reinforcement Learning (webpage): Learn about reinforcement learning and how MATLAB and Simulink can support the complete workflow for designing and deploying a reinforcement learning based controller.Train DQN Agent to Balance Cart-Pole System (example): Find out how to train a deep Q-learning network (DQN) agent to balance a cart-pole system modeled in MATLAB.Train Biped Robot to Walk Using DDPG Agent (example): Find out how to train a biped robot, modeled in Simscape Multibody, to walk using a deep deterministic policy gradient (DDPG) agent.Reinforcement Learning (video series): Watch an overview of reinforcement learning, a type of machine learning that has the potential to solve some control system problems that are too difficult to solve with traditional techniques. Reinforcement Learning with MATLAB (ebook): Find out how to get started with reinforcement learning in MATLAB and Simulink by explaining the terminology and providing access to examples, tutorials, and trial software.","excerpt":"If you are following technology news, you have likely already read about how AI programs trained with reinforcement learning beat human players in board games like Go and chess, as well as video games. As an engineer, scientist, or researcher, you may want to take advantage of this new and growing technology, but where do […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","deep learning application examples","Machine Learning","Mathworks","policy gradient","Reinforcement Learning"],"author_name":"Emmanouil Tzorakoleftherakis","publish_date":"2019-11-18T15:46:03","publication_year":"2019","word_count":1754,"keywords":["CUDA","Go","machine learning","Reinforcement Learning","TPU","AI","neural network","Machine Learning","Mathworks","Aim","C++","policy gradient","deep learning application examples","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","neural network","Aim","TPU","CUDA","R","Go","C++","CUDA"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/three-things-to-know-about-reinforcement-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10039184,"title":"Why Are Data Labelling Firms Eyeing Indian Market?","content":"Toloka, the crowdsourced data-labelling service provider incorporated in Switzerland, has announced its plan to tap into India’s billion-strong worker base. The company assists businesses that use artificial intelligence and machine learning in refining and improving the quality of the data. “We’re excited to continue our push into India, as well as into the surrounding areas, including Pakistan, Myanmar, Bangladesh, and Indonesia. Our 300,000 Tolokers from this region have already proven extremely valuable to our customers, but a great deal of untapped potential remains in this market. We hope to see new talents from India and beyond on our platform soon and look forward to setting new industry standards together,” said Olga Megorskaya, Toloka CEO. Analytics India Magazine spoke with Olga Megorskaya, Founder and CEO of Toloka, to understand the company’s plan, partners and expansion strategy. Bet on Indian market The Indian market seems to be exciting for Olga Megorskaya. She said, “Its sheer size and prevalence of STEM education mean that there is a vast pool of talents, who could potentially take on data-labelling tasks. We call such people Tolokers. While there are already around 300,000 Tolokers from India and the surrounding region on our platform, we believe that the area’s full potential remains largely untapped.” The company hopes to see new Tolokers from India join this growing industry, gain data-labelling skills and earn extra income on a schedule that works for them. Toloka is open to new partnerships that fit its model and ethos in India. “However, we do work with a wide variety of companies around the world that use AI in their processes,” said Olga. The Indian artificial intelligence market is valued at $6.4 billion in 2020, and is expected to further rise as per a report from AIMResearch and Jigsaw academy. “AI and Machine Learning rely on vast streams of high-quality data to carry out calculations and improve their outcomes. A team of human data labellers, or Tolokers, work behind the scenes to ensure the quality of the data is untarnished. Tolokers complete data-labelling tasks, which then go through a quality control mechanism, ensuring only the highest quality data is fed back to our clients and their AI and ML services,” Olga Megorskaya said. The expansion opens up a lot of employment opportunities. As per Olga, any individual can choose which tasks they’d like to complete on the platform and earn extra income while contributing to AI and ML technology advancement. “While we are looking forward to training more Tolokers from India and the surrounding countries, I’d like to stress that Toloka is a truly global platform,” she added. Use cases “One case study that demonstrates Toloka’s capabilities is our work with chatbots. We’ve all seen how difficult it could be for chatbots to carry out a conversation that mimics human speech. In addition to issues with authenticity, there have been instances where chatbots trained on open-source materials have engaged in inappropriate and offensive speech. Toloka can help perfect chatbots and we have demonstrated our abilities in this area in a joint programme with DeepHack hackathon. Contestants in the hackathon created their own chatbots, which engaged in conversations with our real-life Tolokers. “Our Tolokers, meanwhile, were rating every response from the chatbot, providing it with the data needed to learn and improve its ability to conduct lifelike and accurate responses. Over 4 days, 200 Tolokers rated 1,800 dialogues, with excellent results – the quality of conversation from the chatbots was dramatically improved,” Olga said. “Toloka’s services were also used successfully to improve self-driving technology. An important task for the creator of a self-driving vehicle is to train it to extract information about its surroundings from the data it receives from sensors. During the ride, the car records everything it sees around it. This data is uploaded to the cloud, where the preliminary analysis is completed, and then it goes to post-processing, which includes labeling the data. The labeled data is sent to the machine learning algorithms, the result is returned to the vehicle, and the cycle repeats, improving the quality of object detection through multiple iterations. Tolokers labeled tens of thousands of images to train the neural networks to recognize the objects a car might encounter on the streets. To do this the developers added their own visual editor, which has layers, transparency, selection, zoom, and classification (you can embed any interface in Toloka and send data via the API). This increased the speed and quality of the data labeling by a long way. In addition, the API allows you to automatically split tasks into simpler ones and then piece the results together. “For example, before labelling an image, you can select what objects there are in it. This will make it clear which classes to use for labelling the image. In addition to human Tolokers, neural networks can also be used to perform labelling. Some networks have already learned to do this task as well as people do, but the quality of their work also needs to be evaluated. That’s why tasks have a mix of images labelled by Tolokers and by a neural network. This way, Toloka is integrated directly into the training of neural networks and becomes part of the general machine learning pipeline,” Olga explained.","excerpt":"Toloka, the crowdsourced data-labelling service provider incorporated in Switzerland, has announced its plan to tap into India’s billion-strong worker base. The company assists businesses that use artificial intelligence and machine learning in refining and improving the quality of the data. “We’re excited to continue our push into India, as well as into the surrounding areas, […]","categories":["IT Services"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-04-29T15:00:00","publication_year":"2021","word_count":873,"keywords":["machine learning","artificial intelligence","AI","neural network","chatbots","ML","Aim","object detection","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","analytics","Aim","chatbots","object detection","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-are-data-labelling-firms-eyeing-indian-market\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10114906,"title":"Democratize data analysis and insights generation through the seamless translation of Natural Language into SQL queries","content":"In the realm of modern data analytics, the ability to seamlessly translate natural language queries into actionable insights has emerged as a transformative capability. This advancement empowers users to interact with complex datasets effortlessly, extracting valuable insights and facilitating informed decision-making. However, the journey towards achieving this feat is rife with challenges, ranging from the complexity of diverse datasets to semantic sensitivity and algorithmic limitations. Navigating the Complexity of Data The datasets under scrutiny encompass over 500 variables across 10 distinct datasets, each varying in level of aggregation and granularity. Within this landscape, datasets with identical names may carry different meanings, while others may yield similar solutions from disparate data sources. This complexity underscores the challenge of prioritizing tables for analysis and crafting SQL queries with pinpoint accuracy. Primarily sourced from banking datasets, the information spans customer interactions, acquisition details, performance metrics, and complaint audits, underscoring the multifaceted nature of the analytical endeavor. Addressing Semantic Sensitivity and LLM Challenges Large Language Models (LLMs), while powerful, often grapple with semantic nuances inherent in natural language queries. Terms like “acquisition” may entail different contexts, necessitating a nuanced understanding of business and variable semantics. Moreover, variable names may vary, posing challenges in identifying the right context for analysis. To mitigate these challenges, a robust Custom  Retrieval-Augmented Generation(RAG) framework has been developed, focusing on identifying the correct business context, variables, and datasets crucial for accurate analysis. This framework leverages a Knowledge Graph to provide deeper insights into variables and datasets, ensuring precise interpretation of user queries. Refinement with Custom RAG Framework A pivotal aspect of the framework’s evolution lies in the development and implementation of the Custom RAG Framework. This framework emerges as a robust solution to address the intricate challenges encountered during data analysis, particularly in refining business context, variable identification, and dataset selection. At the crux of this framework is the imperative to identify and integrate the right business context into the analytical process, since LLMs , lack the required capability to discern the subtle nuances inherent in business terminology. To bridge this gap, the framework leverages advanced algorithms to contextualize queries, ensuring a more accurate and insightful analysis. Furthermore, the Custom RAG Framework tackles the complexity surrounding variable identification and dataset selection. In a landscape characterized by disparate datasets with varying levels of granularity and relevance, the framework streamlines the process of selecting pertinent variables and datasets. By incorporating a knowledge graph enriched with metadata, synonyms, and variable descriptions, the framework empowers the language model to navigate the intricacies of the data landscape with precision and clarity. A notable aspect of the Custom RAG Framework lies in its adaptability to diverse use cases and industries. Whether analyzing banking data, financial institutions, or other sectors, the framework’s versatility shines through in its ability to tailor analyses to specific contexts and requirements. This adaptability ensures that the framework remains relevant and effective across a spectrum of analytical scenarios, facilitating informed decision-making and strategic insights. Moreover, this framework serves as a catalyst for innovation, providing a platform for continuous refinement and enhancement. Through iterative feedback loops and collaborative efforts, the framework evolves to meet evolving user needs and technological advancements. This commitment to innovation underscores the framework’s long-term viability and relevance in an ever-changing analytical landscape. In essence, the Custom RAG Framework represents a paradigm shift in data analysis, offering a holistic and refined approach to navigating complex datasets and extracting actionable insights. By integrating advanced algorithms, contextual intelligence, and domain expertise, the framework empowers users to unlock the full potential of their data assets, driving informed decision-making and strategic outcomes. Building the Custom Scoring Algorithm and Flow of Generating SQL Queries Central to the framework’s efficacy is the development of a custom scoring algorithm tailored to semantic search. As multiple registers failed to deliver desired performance, crafting a scoring algorithm became imperative. The algorithm functions by mining each word of the user query, identifying variables and synonyms, and computing their respective scores. Subsequently, the variables and datasets with the highest scores are selected to form the basis of the final focus for analysis. This initial step sets the stage for generating SQL queries. The selected variables and datasets are fed into a template, enriching the context and ensuring the language model comprehends the intended analysis. Utilizing a multi-shot prompt approach, the framework guides the language model in crafting the SQL query, thereby streamlining the process of query generation. Despite the query generation, challenges persist, such as discrepancies in date variable formats across datasets. To address such issues, an error mechanism is incorporated to rectify format inconsistencies and enhance query accuracy.  The end-to-end process culminates in the generation of SQL queries within an average timeframe of 15 to 16 seconds, facilitating prompt analysis and decision-making. Integration with Conversational Interface and Enhancing User Experience Transitioning from the educated end-to-end framework to a conversational interface marks the next phase of the analytical journey. Preparatory work, including the creation of a knowledge graph and vector stores, sets the stage for seamless interaction with end consumers via the interface. Upon user query initiation, the framework springs into action, leveraging the enriched knowledge graph to augment prompts and guide SQL query generation. The generated query interfaces with the data warehouse, retrieving relevant datasets and executing the query to generate tabular outputs. This output, also comprising insights, charts, and textual analysis, is relayed back to the conversational interface, empowering end users with actionable insights. Democratizing Data Analysis and Future Enhancements The ultimate goal of the framework is to democratize data analysis, enabling users to derive insights without delving into intricate coding processes. Early results indicate significant time savings and enhanced productivity, with insights generated up to three to four times faster and manual work reduced by 60 to 70%. To further refine the framework, continuous improvement efforts are underway. These include fine-tuning the custom scoring algorithm, expanding metadata to enhance fluidity in query interpretation, and implementing user suggestions to address data integration gaps. By iteratively refining the framework based on user feedback and technological advancements, the aim is to elevate the efficacy and usability of the tool, ultimately empowering users with unparalleled data analysis capabilities. In conclusion, the custom RAG framework represents a significant advancement in addressing the challenges of analyzing complex datasets. By leveraging custom scoring algorithms and integrating business context into algorithmic analysis, the framework streamlines SQL query generation from English language queries. Its end-to-end flow, from knowledge graph creation to UI presentation, facilitates swift and efficient insight generation. The framework’s emphasis on democratizing insight generation and its potential for further refinement promise to drive data-driven decision-making across sectors.","excerpt":"By leveraging custom scoring algorithms and integrating business context into algorithmic analysis, the framework streamlines SQL query generation from English language queries.","categories":["Deep Tech"],"tags":["Natural Language Processing","SQL"],"author_name":"Anshika Mathews","publish_date":"2024-03-04T12:00:00","publication_year":"2024","word_count":1101,"keywords":["Go","semantic search","TPU","AI","Natural Language Processing","ML","RAG","Aim","analytics","SQL","R"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","semantic search","TPU","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/democratize-data-analysis-and-insights-generation-through-the-seamless-translation-of-natural-language-into-sql-queries\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10072004,"title":"Ouch, Cognizant","content":"Last week, IT services major Cognizant Technology Solutions reported a net profit of $577 million for the second quarter ended June 30, 2022. This was a 12.7 per cent increase in net profit up from $512 million in the same quarter last year. It also reported a $4.9 billion revenue (6.5 per cent YoY growth compared to last year). But this revenue fell below the expected rate. Dip in full-year 2022 revenue growth guidance Cognizant has reduced its full-year 2022 revenue growth guidance to 8.5% – 9.5% in constant currency from the 9-11% in the previous quarter. During the earnings call, Jan Siegmund, CFO, Cognizant said that the company has been absent from large and mega deals in the past. This has obviously hit the economic growth, but Jan feels that this also allowed the company “to control our balanced portfolio of the revenue growth much better within the corridors that we outlined, plus the margin expansion”. Samlink subsidiary sale impacted financial services revenue growth The company claimed that the financial services revenue dipped due to the sale of its Samlink subsidiary. The subsidiary only grew 2.7% year-on-year. The sale negatively impacted segment revenue growth by approximately 190 basis points. In 2019, Cognizant acquired Finnish IT solutions provider Samlink to build a core banking platform. During the Q4 2020 earnings call, Humphries spoke about how that quarter’s revenue saw a 3% decline year-over-year in constant currency. This was partly due to a negative 250 basis points impact related to the anticipated exit from a large financial services engagement (hinting at the Samlink project). He called the Samlink prospect a “complex ambitious project” and revealed that both the parties involved in the project had gradually realised that the transformation aspect of the project was unlikely to achieve the shared expectations as initially planned. This quarter, the health sciences revenue grew 6.3% year-over-year and the products and resources revenue grew 8.1% year-over-year. In addition, the communications, media and technology sector saw good performance with a revenue growth of 16.1% year-over-year. Image: Cognizant Focus on digital for the future Digital revenue, which grew 13 per cent year-on-year, was the highlight this quarter. Digital represented approximately 50% of total revenue, this time. In an interaction with Deccan Chronicle after the quarterly results, Cognizant India chairman and MD Rajesh Nambiar spoke about the company’s future plans. The software giant is looking at acquisitions in the next six months as it aims to expand its presence in digital technologies. High attrition this quarter; likely to stay IT firms have been battling high attrition since last year and Cognizant is no different. This quarter, it witnessed a rather alarming attrition rate of 31 per cent. The company said this was higher than expected and has impacted this quarter’s revenue performance. Cognizant expects to see “elevated attrition for the remainder of the year”, the CEO informed. For the previous quarter, the voluntary attrition fell 5 points to 26% on an annualised basis. For the second quarter of 2021, voluntary attrition had reached 29%. Cognizant’s headcount expanded to 341,000 employees in the past quarter. In fact, Cognizant is ramping up its hirings. A report earlier this year had predicted that Cognizant would be onboarding 50,000 freshers from India in CY22. This is a big jump from the 33,000 freshers it added in CY21. In the same report, Nambiar had said that this would be one of the biggest hirings for the company. In the past, the software provider had focused more on lateral hires but is now also increasing its fresher count. Flexibility of work,  hybrid models for employee retention In the earnings call, Humphries spoke about how the company is taking various measures to arrest the growing attrition. Cognizant has invested in better compensation, focused on employee learning and development initiatives, he asserted. Owing to the COVID-19 pandemic, a lot of techies prefer the “work from home” or a hybrid working model. Cognizant too is focusing in that direction. “We’ve also recognised how important flexibility is to our associates and have, therefore, communicated a hybrid model that will define our approach to work,” he adds. Recently, Cognizant opened a new office with a capacity of over 5000 associates at Navalur, near Chennai. Nambiar said that this space has been designed for a hybrid work-from-home and work-from-office model.","excerpt":"The company has reduced its full-year 2022 revenue growth guidance to 8.5% – 9.5% in constant currency from the 9-11% in the previous quarter","categories":["IT Services"],"tags":[],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-08-02T12:00:00","publication_year":"2022","word_count":719,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","Git","Aim","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/ouch-cognizant\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167401,"title":"Mumbai-based Scoutflo Secures Pre-Seed Funding to Reinvent DevOps with AI","content":"As companies continue to debate between sticking with Kubernetes or developing in-house cloud solutions, Scoutflo, an AI-powered DevOps platform, aims to eliminate the tedious aspects of these deployments, no matter what option you select. The firm has raised ₹1.4 crore in a pre-seed round led by 100X.VC, with additional backing from strategic investors like Arjun Pillai and Prasanna Venkatesan.  The startup aims to reinvent DevOps workflows by turning deployment, debugging, and compliance into AI-driven, developer-first experiences. With the new funding, Scoutflo wants to focus on building a stronger product. It plans to expand its presence in India and overseas, build out more integrations, and scale its go-to-market efforts. The team is also doubling down on developer experience, working closely with partners like Civo and Last9 to offer better observability and reliability. Founded by CEO Kalpesh Bhalekar and CTO Vedant Vyawahare, Scoutflo’s seven-member team believes that traditional DevOps cannot keep pace with the rapid development of AI-first software. “Developers are expected to move fast, but they’re still bogged down by slow processes and a lack of autonomy. Scoutflo acts as an AI-powered DevOps expert embedded in every team,” Vyawahare said while speaking with AIM. Bhalekar and Vyawahare are not new to the pain points they’re solving. The latter previously worked at BrowserStack, where he saw similar bottlenecks in QA processes. That experience gave him a framework to identify friction between developers and infrastructure, especially in fast-paced environments. While researching the problem, they still spoke to 100-200 CTOs and CDOs, and one insight became glaringly obvious—Devops teams are stretched thin. In many companies, the ratio is often 1 DevOps engineer to 20 developers, leading to burnout and major slowdowns in deployment cycles. What Scoutflo Actually Does? “DevOps guys are constantly getting swamped,” said Vyawahare. “This dynamic leads to developers waiting around, delays in shipping, and overall loss in momentum. We thought—what if we could give developers the autonomy to manage deployments themselves, without sacrificing security or efficiency?” Scoutflo does just that. It’s an AI platform that sits on top of cloud environments—AWS, GCP, Azure, VMware, and even Civo—allowing developers to provision and deploy infrastructure without becoming cloud experts. It’s a cloud-agnostic Platform-as-a-Service (PaaS) designed to assist developers in deploying infrastructure and services without worrying about the underlying technology stack. It handles everything from auto-generating Terraform scripts to setting the right security guardrails—all powered by an integrated AI system that helps provisioning, debugging, and even cost optimisation. “You can literally chat with our AI to ask things like, ‘Why is my deployment failing?’ or ‘Can you help me spin up a secure staging environment on AWS?’ and it’ll get done,” Vyawahare explained. “It’s not just automation—it’s intelligent decision-making. Our AI understands infrastructure, dependencies, and scaling requirements.” This isn’t just about deploying a single service. Scoutflo is designed to manage complex, multi-service deployments—including those involving LLMs, traditional APIs, databases, and more. While “LLMOps” is merely one subset, Scoutflo encompasses the entire pipeline. “Most DevOps tools just help with regular services. We cover Kubernetes and non-Kubernetes deployments equally well. We want to be the go-to platform for companies looking to shift to scalable architectures without the overhead,” Vyawahare said. The product is currently in closed beta, and at this stage, the focus is on direct engagement with users. Soon, they’ll launch an AI chatbot—a completely free and open tool designed to generate Terraform code, handle debugging queries, and assist with fresh setups. As for the chatbot’s backend, it’s built on top of OpenAI and Anthropic’s large language models. “We have fine-tuned these models, and we’ve done a lot of advanced prompt engineering. We are also using MCPs, a new technology we’re quite bullish on,” they said. “We’re ahead of the curve here, especially since not many teams in India are exploring MCPS yet.” The startup is currently a compact team of seven but is considering expansion soon. “We’re moving fast, and growth is on the horizon,” they said. A Crowded Market Internationally, competitors like Qovery (in the US) and Northflank (in Europe) are already making waves. Northflank, for instance, has been operating for two years and recently raised a $22 million round. In India, the closest peers are startups like Facets.cloud and Devtron. “They’re doing good work, but we believe our AI-first approach and neutrality when it comes to cloud providers and IaC (Infrastructure as Code) gives us a unique edge,” Bhalekar told AIM. While most deployment tools stop once the code is live, Scoutflo goes further. The platform includes AI-assisted debugging and post-deployment workflows to help developers manage services after launch. It reduces the learning curve for complex systems like Kubernetes, enabling even less experienced developers to work with production-level infrastructure. “We’re bridging the gap between developers and DevOps, especially post-deployment. You can deploy, monitor, debug—all in one flow,” said Vyawahare. “That’s where a lot of the traditional tools fall short.”","excerpt":"The company plans to expand its presence in India and overseas, build more integrations, and scale its go-to-market efforts.","categories":["AI Startups"],"tags":["Developers"],"author_name":"Mohit Pandey","publish_date":"2025-04-08T08:00:00","publication_year":"2025","word_count":807,"keywords":["Anthropic","GCP","OpenAI","AI","AWS","R","Aim","prompt engineering","Azure","kubernetes","Developers"],"extracted_tech_keywords":["AI","OpenAI","Anthropic","Aim","prompt engineering","AWS","Azure","GCP","kubernetes","R"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/mumbai-based-scoutflo-secures-pre-seed-funding-to-reinvent-devops-with-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":34146,"title":"Pentagon Researchers Believe A Wasp&#8217;s Brain Can Lead To An AI Breakthrough","content":"For long, man has drawn inspiration from nature to find answers to some of the most puzzling questions. Over the course of modern history, his curiosity to seek answers never ceased. In fact, it continues to shape the way we think and how we interact with each other. Adding yet another interesting dimension to scientific discovery, researchers at the Pentagon, the headquarters of the United States Department of Defense, are looking at the possibility of how insects could hold the key for developing new artificial intelligence-based technologies. In a call for submission of research papers in the field, Pentagon’s Defence Advanced Research Project Agency (DARPA), had recently announced that it is looking for submissions on computational framework and capabilities of small flying insects. “Inviting submissions of innovative basic research concepts exploring new computational frameworks and strategies drawn from the impressive computational capabilities of very small flying insects,” it said in a statement. As part of the department’s Artificial Intelligence Exploration (AIE) programme, the chosen researchers will have to submit concepts aimed at understanding sensory and nervous systems in miniature insects and develop models that could be mapped onto suitable hardware in order to emulate their functioning. For the past five decades, DARPA has been actively involved in bringing out groundbreaking technologies and it continues to lead AI innovation through its broad portfolio of R&D programmes. In September 2018, it announced more than $2 billion in new initiative called the “AI Next” campaign. The initiative is aimed at driving what they call the third wave of AI, where the department hopes to create an AI system that can work as a natural being than mimicking the inspect. According to DARPA, these insects’ features like drastic miniaturisation, energy efficiency and compact neuron form-factor can be adopted in an AI-system to perform energy, time and space efficient operations. Quoting the example of Megaphragma mymaripenne, a microscopically sized wasp, believed to be smaller than an amoeba, the statement said that the organism’s nervous system consists of thousands of neurons with fast integrated sense-control which can perform tasks like guidance and locomotion. It further stated that there aren’t many studies regarding the electrical and magnetic circuits within the insect to perform the task. By inviting researchers to develop and understand the computing model behind a miniature insect’s problem-solving capabilities, DARPA hopes to make new computational strategies and create efficient approaches to contextual AI. “Understanding the computational principles, architecture, and neuronal details of these miniaturized bio-systems could provide a fundamentally new way of thinking about information processing, including with regard to time and energy requirements,” the statement read. It further stated that the chosen submission will receive $1 million as award value.","excerpt":"For long, man has drawn inspiration from nature to find answers to some of the most puzzling questions. Over the course of modern history, his curiosity to seek answers never ceased. In fact, it continues to shape the way we think and how we interact with each other. Adding yet another interesting dimension to scientific […]","categories":["AI News"],"tags":["AI for good","DARPA","research and development","surveillance"],"author_name":"Akshaya Asokan","publish_date":"2019-01-25T10:53:32","publication_year":"2019","word_count":446,"keywords":["Go","artificial intelligence","AI","DARPA","research and development","surveillance","RPA","innovation","AI for good","RAG","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","R","Go","GAN","ViT","RPA","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pentagon-researchers-believe-a-wasps-brain-can-lead-to-an-ai-breakthrough\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":22275,"title":"Can AI Diagnose Cardiac Diseases Better? Indian &#038; Global Companies Have Made Tremendous Headway in Cardiac Care","content":"Artificial Intelligence has been around for some time in healthcare, but it’s major impact is only being felt now. From improving the hospital’s in-patient care, digitizing health records, applying advanced computer vision systems detecting abnormalities in X-rays and MRIs to providing evidence-based treatment recommendations and using robotics in surgery, AI shows great promise in healthcare. Why is big tech charging into healthcare? Big tech companies are spearheading the “Uberization” of healthcare — changing the paradigm of care model, imaging and diagnostics by providing immediate and accurate results, more on the lines of digital interface and delivery models built by Amazon, Uber and companies. Even though companies like IBM, Microsoft & Apple have played a big part in pushing AI into healthcare, none can match Google’s commitment in healthcare. Apple is making a headway through consumer devices (for e.g. the Apple Watch tracks the user’s daily activity, heart beat & more), while IBM & Microsoft have created enterprise solutions for physicians and hospitals. Google has charged ahead in healthcare space through strategic investment in healthcare and life sciences space and banking on the user data to create health solutions for customers. Google Ventures, the investing arm of the company revealed that more than one-third of its investment had been in health tech. The Mountain View search giant’s investment in healthcare space is evident from its two high profile subsidiaries – Verily, the life sciences department known for its deep bench consisting of top medical experts and Calico, founded in 2013 and funded by Google that focuses on anti-ageing research. Though there are few details emerging from Calico’s secretive anti-ageing research, Verily recently announced a machine learning algorithm that can detect risk of heart diseases by analyzing scans of the back of the eye. After improving diagnostic tools for diabetic patients by powering them with deep learning algorithms, Google has trained its eyes on cardiac diseases. The Mountain View search giant has met with a lot of success in improving existing diagnostic tools, especially for diabetes. In 2015, Google and digital health company Dexcom joined hands to launch a new generation of continuous glucose monitors (CGM). Last year, Google researchers announced that they had trained image recognition algorithms to detect signs of diabetes-related eye disease called diabetic retinopathy in patients. This retinal screening system is up and running at Aravind Eye Care hospitals and the system uses the same deep learning technique that leveraged by Google’s image search applications to detect dogs and cats in pictures. Big Tech’s role in overhauling cardiac care Tech companies have made medicine extremely data-intensive by applying deep learning algorithms that are able to extract meaningful patterns from data collections – thereby changing treatment paths and enabling doctors to take data-backed decisions, emphasizes a NCBI article. Such is the heightened frenzy in AI-led cardiac research that there is a new publication every 2.7 min, indicates the article. Also, cardiovascular is ideally suited for AI – there is a wealth of data from medical records (lab results, physician’s notes etc), imaging data which can be effectively mined by deep learning algorithms. Majority of AI use-cases for managing heart disease appear to fall into three major categories: Medical Imaging: There is an increased use of AI in medical imaging and improving the speed and accuracy of scans for radiologists. Startups and big tech companies are training machine learning algorithms on huge datasets to improve the accuracy of patient scans for better detection. Heart Risk Prediction: Again, machine learning algorithms are deployed to better predict the risk of cardiovascular disease. ECG Monitoring: This area is set for an overhaul with medical grade wearable devices that are providing continuous cardiac monitoring. Let’s look at the penetration of AI in cardiac care 1) IBM’s research in EHR data to predict heart failure: Earlier last year, IBM applied its cognitive computing capabilities to analyze patient data to assess factors that could lead for early detection – a) understand traditional risk factors clinicians commonly use to diagnose heart failure. IBM leveraged Natural Language Processing techniques to extract information from unstructured data such as physician notes to understand the symptoms; (b) more accurately predict heart failure by combining unstructured data from doctors’ notes with structured EHR data. To achieve this, IBM applied machine learning methods to build predictive models that took into account a mix of variables. The outcome was –  the research led to a deeper understanding about the tradeoffs between certain data types and their usefulness in helping detect an individual’s likelihood of heart failure. 2) Seattle startup KenSci uses machine learning techniques to predict patient’s heart disease: This young startup has garnered a lot of VC interest including Microsoft for its software platform which was built on a database of 150 machine learning models using algorithms trained on over 10 million data sets. The startup states that the software can analyze data to indicate the potential patient risks of developing heart disease. This Seattle-based startup is already providing predictive analytics to 17 million patients. 3) India’s Qure.ai presents medical imaging solution: Qure.ai co-founded by Prashant Warier and Pooja Rao recently presented its research on medical imaging for cardiology. The team presented their work at MICCAI 2017, on convolutional networks that outline and quantify the left and right ventricles and the myocardium from cardiac MRIs. The research paper presented a 2D and 3D segmentation pipelines for fully automated cardiac MR image segmentation using Deep Convolutional Neural Networks (CNN). The models are trained end-to-end from scratch using the ACD Challenge 2017 dataset comprising of 100 studies. 4) Medical device startup ten3T provides continuous cardiac monitoring:  Recognized as one of the 10 best startups in wearable technology in 2017, ten3T’s medical grade device Cicer helps in cardiac monitoring and is pegged as a life-saver for heart patients. Their wearable device Cicer is a wireless patch that flashes the body vitals continuously and measures ECG, heart rate, blood oxygen, and temperature — procedures that would have been complex to perform, with just a patch.  The device makes use of algorithms that capture immediate signs as well as longer term subtle changes, enabling physicians to make immediate and predictive decisions. 5) Bengaluru-based Cardiac Design Labs is changing cardiac care: Founded by Anand Madanagopal, Cardiac Design Lab’s MIRCaM System is revolutionizing cardiac care with a diagnostic device that alerts care providers in case of an emergency. The device – MIRCaM that stands for Mobile Intelligent Remote Cardiac Monitor delivers real time advanced cardiac diagnosis and monitoring in remote settings and proving to be a boon in rural areas by providing continuous monitoring. Black Box problem in medicine So far AI’s use in medical imaging has been successful and the field of radiology is now being known as the Silicon Valley of medicine, given the heightened interest in radiology.  There are a stream of offerings that surmise that Deep Learning algorithms detect diseases as accurately as experienced doctors. However, that still doesn’t explain the black box problem. Deep learning algorithms cannot be traced back to understand how it arrived at the final output, which means for doctors the AI system is opaque which could be a serious problem in healthcare wherein the treatment path has to be clearly explained. But there’s one US company that believes deep learning can prove its worth in medicine. San Francisco-based medical imaging company Arterys that received the FDA approval to market its Arterys Cardio DLTM application emphasized that one needs to prove statistically that the algorithm is following whatever the intended use and the marketing claims, John Axerio-Cilies, chief technology officer said reportedly.","excerpt":"Artificial Intelligence has been around for some time in healthcare, but it’s major impact is only being felt now. From improving the hospital’s in-patient care, digitizing health records, applying advanced computer vision systems detecting abnormalities in X-rays and MRIs to providing evidence-based treatment recommendations and using robotics in surgery, AI shows great promise in healthcare. […]","categories":["IT Services"],"tags":["AI Companies","AI Healthcare","medical image processing companies"],"author_name":"Richa Bhatia","publish_date":"2018-03-06T04:55:28","publication_year":"2018","word_count":1260,"keywords":["artificial intelligence","machine learning","AI","AI Healthcare","neural network","computer vision","RAG","Ray","Aim","deep learning","analytics","AI Companies","medical image processing companies"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","computer vision","analytics","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-ai-diagnose-cardiac-diseases-better-indian-global-companies-have-made-tremendous-headway-in-cardiac-care\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10048385,"title":"Tech Behind Water Resource Management Startup Cranberry Analytics","content":"Founded in 2010, Cranberry Analytics is a tech-enabled company operating in the water management space. Headquartered in Pune, the company was co-founded by Shishir Thakur, Amit Deshmukh and Onkar Gauridhar. Cranberry Analytics was founded with the primary objective of measuring water efficiently. With the use of analytics, ML and IoT, the company augments government and large water infra companies with a tech stack to enable them to curb water wastage and plug revenue leakages. In a conversation with Analytics India Magazine, Co-founder and CTO Shishir explained the technology behind Cranberry Analytics’ management system Recon, breaking down how AI and ML can facilitate the management of natural resources and what the future for AI in water management looks like. Edited excerpts from the conversation: AIM: What is your flagship product\/service? Shishir Thakur: Our flagship product is our water billing, budgeting and management system – Recon, deployed in PCMC (Pimpri Chinchwad Municipal Corporation) since 2012. Recon integrates with all data sources related to: Water consumptionWater distributionDemand management systems (billing)Revenue management systems (manual and digital)Customer experience and dispute resolution It also provides a platform and dashboards to all the concerned personnel of the water department, citizens, and support providers. AIM: How does Recon work? Shishir Thakur: Recon has built-in capabilities to detect anomalies related to water usage and leakages, revenue leaks, automatic dispute resolution, tracking defaulters, predicting and forecasting water demand, tariff simulation, integrations with smart water meters, sensor management, accounts department and several other auxiliary functions. Using data visualisation in reports, charts and tables, Recon makes it very easy for top management to form policies and implement relevant changes that have the most impact. AIM: Please explain the tech stack used by the team. Shishir Thakur: For the Recon frontend, we use Angular.js, Charts.js, Tableau and Metabase. We primarily use node.js with some specialised algorithms in Golang and Python for anomaly detection and regression through past data to find patterns and forecasts for its backend. Databases include MySQL, MongoDB and Realm-DB.Mobile: Android (with Realm DB) integrates with onspot thermal bill printer and portable PoS machine for revenue collection.DevOps: Docker, Ansible, Jenkins, Nagios, Bitbucket. AIM: How does Cranberry Analytics leverage AI and ML capabilities? Shishir Thakur: We analyse all incoming data to find consumption related anomalies (similar to how a credit card fraud detection system works), and then alert the respective personnel. This has helped us identify many leakages, unauthorised consumption, geographic segmentation of faults based on geotags, and human errors. Additionally, we leverage AI and ML to predict area-wise water demand; assist water distribution staff in achieving equitable water distribution with intelligence gathered from several multivariate analyses; for automatic dispute resolution and bill rectification; for payment-related inconsistencies; and for the coordination of field operations. AIM: Who are your clients? Share a few real-life cases where your product has been used. Shishir Thakur: We have worked with clients like the World Bank, Asian Development Bank, Karnataka Urban Infrastructure Development Corporation, Ministry of Urban Development, Ministry of Housing and Urban Affairs, and Suez International. Case Study: PMAY (Pradhan Mantri Awas Yojna) For an initiative by the Ministry of Urban Development, we have been augmenting them with technologies, data analysis and software infrastructure to enable them to identify the correct approach towards the slum dwellers, strategies and software for more effective data collection and survey exercises, and identifying the truthful beneficiaries for the program and eliminating the spurious ones. Service-level benchmarking: An initiative by MoUD to identify, analyse and improve the services being provided to citizens from government bodies. We helped them in large scale data collection, statistical analysis, identifying anomalies, building models, and defining policies based on that data. PCMC water billing management: We helped PCMC, a municipal corporation with a 25 lakh+ population, in reducing their unauthorised water consumption, improving revenues, reducing customer complaints, enabling equitable water distribution and reduction in water consumption to save 31000 million litres of water per year. Revenues have reached from Rs 22 crore per annum to almost Rs 43 crore already, with a 95 per cent reduction in customer complaints and almost nil non-metered connections and historical disputes. AIM: How can technology further help in water management? Shishir Thakur: The biggest impetus can come from smart metering of cities, implementation of IoT devices and sensors in water distribution, and rapid growth of smart tech in agricultural water consumption. The biggest returns on investment can be seen by: Implementing technologies like drip irrigation, sprinklers, and air-drop irrigation in farming — the biggest area of ineffective water utilisation today.Implementing better systems to detect leakages in urban pipelines and fixing them as soon as possible. We lose more than 60 per cent of treated water due to underground seepages.Mandatory implementation of rainwater harvesting in urban areas and effectively monitoring its efficacy using tracking technologies.Mandatory waste-water recycling, especially in large apartments, industries and urban centres, and greywater usage instead of freshwater, wherever possible.Implementing and experimenting with low power, low cost, novel methods of seawater desalination.Use of ML for preventive hydraulic modelling of our current dilapidated storm drain infrastructure for better rainwater catchment to reduce floods and recharge water bodies.A strict ban on letting effluent and chemical waste enter river bodies, that’s a huge drain on resources and can be easily monitored centrally and remotely using chemical sensors and IoT devices. AIM: What does the future of resource management using technology look like? Shishir Thakur: Historically, all struggles related to resource management and resource scarcity have originated from slow movement of information, deliberate and artificial impedance imposed by governments or large businesses, inability in simulating complex future scenarios due to a multitude of variables, opaque information channels from ruling regimes, lack of resource visibility and operational efficiency. The more connected we become, both in terms of information transfer and the stacking effect of innovations on top of each other, we keep becoming free of geographic and political constraints. This kind of global collaboration is unprecedented and unpredictable. This pattern will automatically fuel itself to a level where natural market forces will instantly eliminate scarcity of any kind in any area of the world. Just like it was difficult 15 years ago to envisage the kind of product and food deliveries that we now take for granted, it may seem difficult to imagine this in 2021, but with the use of AI augmentation and other tools built on the ever-improving IT backbone, the natural market economies will prevail and resource management issues of today will become a trivial problem to solve. We will surely face novel challenges of different kinds, but speculation is futile, given the extremely rapid pace of innovation. AIM: What does the road ahead for Cranberry Analytics look like? Shishir Thakur: The solutions that we have developed in PCMC are already quite comprehensive. We have plans to expand these and create an Integrated Water Management System (including water billing, water budgeting, water analytics and the instrumentation aspect of it) using sensors and industrial IoT devices to monitor water’s lifecycle end-to-end, right from the source till it ends up in the waste-water treatment plant. We are in talks with a few state-level entities for a larger implementation of these solutions. Because, even at such a small level, due to our efforts here, PCMC is now saving potentially ~31000 million litres of water annually, it becomes imperative for us to let the rest of the country benefit from the same processes. Currently, we are also looking for investors who are aligned with our mission and can provide us with the wherewithal of rapid expansion into not just the rest of India but also in water-scarce places like the Middle East and Africa. We have a pilot project being implemented in Kenya and are talking with a few Dubai entities. To understand if AI can actually help in resolving India’s water crisis, check this article.","excerpt":"Founded in 2010, Cranberry Analytics leverages ML and IoT to help governments and companies manage water better.","categories":["AI Startups"],"tags":["AI (Artificial Intelligence)","Machine Learning","statistical analysis using sql"],"author_name":"Debolina Biswas","publish_date":"2021-09-15T17:00:00","publication_year":"2021","word_count":1302,"keywords":["statistical analysis using sql","AI","MongoDB","ML","Machine Learning","docker","RAG","Python","Aim","anomaly detection","analytics","AI (Artificial Intelligence)","fraud detection"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","anomaly detection","fraud detection","docker","MongoDB","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/tech-behind-water-resource-management-startup-cranberry-analytics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10168100,"title":"Context.ai Co-Founders Join OpenAI","content":"OpenAI has hired the co-founders of Context.ai, a startup specialising in analytics and evaluations for AI models. The move was confirmed on Tuesday, with Context.ai announcing that it will shut down its products following the acqui-hire. “We spent two years building evals and analytics for LLMs at Context.ai with a few pivots along the way! Alex and I couldn’t be more excited for this next chapter of our journey at OpenAI and are enormously grateful to everyone who played a part in the adventure,” said Henry Scott-Green, former Context.ai CEO. The company’s official website also redirects to a similar note. “Evals are a requirement for building high-performing AI applications, but they’re hard to get right today. We are thrilled to be joining the OpenAI team to help build the tools developers need to succeed,” the website reads. Launched in 2023, TechCrunch reported that the company raised a $3.5 million seed investment to fully develop the idea. They launched the company to understand how users are interacting with their LLMs. The investment was led by Google’s venture arm, GV, and Theory Ventures. The move is part of a wider push by OpenAI to grow its expertise across both software and hardware, following a string of strategic hires and acquisition talks.Earlier this month,  The Information reported that OpenAI was in talks to acquire the AI hardware startup being developed by former Apple design chief Jony Ive in collaboration with OpenAI CEO Sam Altman. According to the report, OpenAI may spend around $500 million to acquire the nascent company, io Products.","excerpt":"The startup was backed by Google’s venture arm, GV, and Theory Ventures.","categories":["AI News"],"tags":["OpenAI"],"author_name":"Aditi Suresh","publish_date":"2025-04-16T16:01:04","publication_year":"2025","word_count":258,"keywords":["Go","OpenAI","AI","programming_languages:R","programming_languages:Go","analytics","R","startup"],"extracted_tech_keywords":["AI","analytics","OpenAI","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/context-ai-co-founders-join-openai\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10097619,"title":"OpenAI, Google, Microsoft, and Anthropic Launch Frontier Model Forum","content":"Big techs unite to form the Frontier Model Forum. Led by industry biggies OpenAI, Google, Microsoft and Anthropic, the forum is aimed at promoting safe and responsible development of frontier AI systems. The forum looks to advancing AI safety research, identifying best practices and standards, and allow information sharing between policy makers and industry. This collaboration comes as a continuation to the announcement of big tech collaborating to form responsible AI as per White House guidance. Open Collaboration The Frontier Model Forum is welcoming participation from other organisations that are developing frontier AI models, to collaborate with them. The Forum will also support efforts that work towards meeting society’s challenges including climate change, early cancer detection and prevention, and fighting cyber threats. The Frontier Model will establish an Advisory Board to help guide its strategy that will represent a diversity of backgrounds and perspectives. The forum will support existing government initiatives such as G7 Hiroshima, OECD’s work on AI risks, and US-EU trade. The Forum classifies frontier models as large-scale machine learning models that surpass the capabilities present in the advanced existing models. Companies looking to apply for membership need to follow a certain set of criteria such as developing frontier models as defined by the Forum, build safe and responsible frontier AI models, and contribute towards joint initiatives of the Forum in the development and functioning of the initiative. Anna Makanju, VP of Global Affairs, OpenAI said that it is vital for AI companies working on powerful models should “align on common ground and advance adaptable safety practices to ensure powerful AI tools have the broadest benefit possible.” However, big tech such as Meta, Apple and Tesla are not in the picture. What about Meta? Will Apple Join? When it comes to large language models, the companies that have decided to come together are majorly closed-door. With the exception of Microsoft, who with their latest partnership with Meta launched Llama-2 – an open-source platform, the companies uniting are closed source. Meta is somehow now associating with the forum. Known for accessing open source tools without contributing to the development of technology, Apple is famous for building a closed door ecosystem. However, with Apple planning to build its own chatbot nicknamed AppleGPT, their efforts in advancing AI is evident. Will this qualify as a criteria for Apple to join the forum? It is to be seen if these biggies will join the club.","excerpt":"However, Meta, Apple, Tesla and others seem to be missing in the joint effort.","categories":["AI News"],"tags":["AI Safety","Anthropic","Apple","Google","Meta","Microsoft","OpenAI","Responsible AI","Tesla","White House"],"author_name":"Vandana Nair","publish_date":"2023-07-26T20:48:32","publication_year":"2023","word_count":403,"keywords":["Anthropic","White House","Meta","machine learning","Go","OpenAI","AI","Apple","Responsible AI","GPT","Aim","ViT","Google","GAN","AI Safety","Tesla","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","OpenAI","Anthropic","Aim","R","Go","GPT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-google-microsoft-and-anthropic-launches-frontier-model-forum\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10118239,"title":"Superintelligence Timeline Now Shrinks to 20 Years","content":"One of the godfathers of AI, Geoffrey Hinton, recently altered his estimation of when superintelligence could come into being. “I think it’s fairly clear that maybe in the next 20 years, I’d say with a probability of 0.5, it will get smarter than us and probably in the next hundred years, it will be much smarter than us,” Hinton said during the Romanes Lecture he gave recently on whether digital intelligence will replace biological intelligence. Hinton had famously changed his estimation early last year, believing AI will become smarter than humans, shortly before he resigned from Google. While Hinton initially changed his conservative estimate of 30-50 years to a more dire 5-20 years, it seems this has changed slightly over the months due to the accelerated AI advancements. Hinton isn’t the only industry stalwart to predict when superintelligence will become a thing. From Elon Musk’s optimistic, yet unlikely prediction of a year to Yann LeCun’s vague estimate of “years, if not decades”, to Jensen Huang’s straightforward and popular estimation of five years, it seems that AI superseding human intelligence will forever be around the corner. No, AI is not going to take over just yet Both Hinton and fellow AI godfather LeCun give ample reasons why superintelligence is not likely anytime soon. Hinton acknowledged that these things are hard to predict since the technology is new. I now predict 5 to 20 years but without much confidence. We live in very uncertain times. It's possible that I am totally wrong about digital intelligence overtaking us. Nobody really knows which is why we should worry now.— Geoffrey Hinton (@geoffreyhinton) May 3, 2023 Ever the optimist, Hinton said, “My conclusion, which I don’t really like, is that digital computation requires a lot of energy, and so it would never evolve. We have to evolve using the quirks of the hardware to be very low-energy. But once you’ve got it, it’s very easy for agents to share – GPT4 has thousands of times more knowledge in about 2% of the weights, so that’s quite depressing. Biological computation is great for evolving because it requires very little energy, but my conclusion is that digital computation is just better.” However, while Hinton is understandably cautious, LeCun has been much more standoffish to the idea that AI regulation is needed just in case the human race gets taken over. LeCun has repeatedly stated that the move towards superintelligence is decades away, with no specific breakthrough defining when we will definitively see a shift. And this might be the reason that superintelligence will always be at least one step away from where we think it is. In not being able to pinpoint an exact moment where superintelligence comes into existence, the ability to actually predict it becomes moot. Maybe predictions aren’t really the way to go about it, considering every new breakthrough seems to fast forward the timeline, followed by ultimately pushing it back, thanks to the new challenges they subsequently pose. How far are we, really? The general consensus is that we are nowhere near where we need to be for superintelligence to come into being. The leap of AI as it is now to where it needs to be has several breakthroughs in between for it to rival the intelligence of a human mind. The godfathers and other relatives of AI seem to agree on one thing. The level of computational power needed for a potential superintelligence is extraordinary, so that could be one starting point on what to look for. Hinton said that the reason for this was that AI might already be almost on par with the human brain, at least in the way it works, rather than its intelligence. “I thought making our models more like the brain would make them better. I thought the brain was a whole lot better than the AI we had… I suddenly came to believe that maybe the digital models we’ve got now are already very close to or as good as brains, and will get to be much better than brains,” Hinton said. He explained that the ability to run the same neural net on different computers or pieces of hardware far outweigh the learning ability of biological computation, with digital computation having the scope for becoming smarter much faster. However, as mentioned before, biological computation needs much less energy. Unless this gap is closed, super intelligence isn’t really something that is within grasp. However, while many believe that quantum computation is the means to this end, LeCun seems to believe that quantum computation is just as much of a pipe dream as superintelligence. With China recently announcing the development of the Taichi chiplet (advertised as capable of running an AGI model!), it’s only a matter of time to see if this theory will hold up.","excerpt":"Elon Musk and Jensen Huang think it is happening somewhere between a year and five.","categories":["AI Trends"],"tags":["AGI","Geoffrey Hinton","Superintelligence","Yann LeCun"],"author_name":"Donna Eva","publish_date":"2024-04-16T13:25:36","publication_year":"2024","word_count":801,"keywords":["Go","Yann LeCun","programming_languages:R","AI","Geoffrey Hinton","programming_languages:Go","Git","GPT","Superintelligence","AGI","R","llm_models:GPT"],"extracted_tech_keywords":["AI","R","Go","Git","GPT","llm_models:GPT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/superintelligence-timeline-now-shrinks-to-20-years\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10041124,"title":"DataOps Goes Mainstream As Atlan Lands Big","content":"Last week,  Atlan announced that it has raised $16 million as part of its Series A round led by global venture capital and private equity firm Insight Partners. “We are excited to partner with Insight as their significant expertise in backing ScaleUp SaaS companies will help accelerate our growth. We will be investing heavily in growing our product & engineering team in India and take great pride in pioneering a new wave of companies in India that are building world-class products from India for the rest of the world,” said Varun Banka, Co-Founder of Atlan. Founded in 2018, this Indian startup now has its offices in Singapore, Philippines, Nigeria, USA, India – for now, and is catering to customers across the globe. With an investment of over two years across 200 data projects, the company has developed its own unique product.  “Today, data assets are not just tables, but code, models, BI dashboards, and pipelines,” said Prukalpa Sankar, Co-Founder of Atlan. At Atlan, continued Prukalpa, her team is reimagining the human experience with data. “Why can’t data assets be shared as easily as sharing a link on Google Docs, or if Google Analytics can tell you usage on a website, why can’t we do the same for our data?” DataOps: The secret sauce of Atlan Data drives businesses growth and provides valuable insights prior to any conclusive decision making. As the enterprises scale, many challenges surface. For instance, working professionals, including data scientists, analysts, engineers, join in with different skill-sets and tools. Different people, different tools, different working styles – all these lead to a major bottleneck.  Business segments are in dire need of data management to create contextual insights, now is the time to improve the quality and speed of data streaming into the organisation and get leadership commitment to support and sustain a data-driven vision across the company. This is where DataOps (data operations) come in handy. For instance, users can integrate their tables from Databricks with Atlan in a series of steps. Initially there are some prerequisites for establishing a connection between Atlan and Databricks Account: Go to the Databricks console and select “Clusters” from the left sidebar.Select the cluster you want to connect with Atlan. The cluster should be in a Running state for the Atlan crawler to fetch metadata from it.Click on “Advanced Options” in the “Configuration” tab.Select the “JDBC\/ODBC” tab and copy the information here: Image credits: Atlan Host: Databricks cluster server hostnamePort: Port. Typically it is 443.Personal Access Token: You can generate the Personal Access Token by following the official guide.JDBC URL Suffix: This is the JDBC URL suffix. Highlighted in the image above. Please ensure you do not add the hostname, port or PWD values here. After these prerequisites, following steps need to be followed. STEP 1: Selecting the source Log into your Atlan workspace.On the home screen, click on the “New Integration” button in the top right corner. You will see a dialogue box with the list of sources available on your workspace.Select “Databricks” from the list of options, and click on “Next”. STEP 2: Providing credentials You will see an option to either select a preconfigured credential from the drop-down menu or to create a credential. To set up a new connection, click on the “Create Credential” button.You will be required to fill in your Databricks credentials.Once you have filled in the details, click on “Next”. Image credits: Atlan STEP 3: Setting up your configuration You will now be asked to fill in the details of your database and table.Choose whether to run the crawler once or schedule it for a daily, weekly, or monthly run. You will be asked to specify the timezone for the run.Click on “Create”. Your connection is now created. DataOps should be looked at as a collaboration of people, process and technology to deliver trusted and high-quality data in a step by step process.  The information architecture of any company is at the heart of DataOps. Data curation, metadata management to data governance are few of the many other things that DataOps practices can help with.","excerpt":"Last week,  Atlan announced that it has raised $16 million as part of its Series A round led by global venture capital and private equity firm Insight Partners. “We are excited to partner with Insight as their significant expertise in backing ScaleUp SaaS companies will help accelerate our growth. We will be investing heavily in […]","categories":["AI Features"],"tags":["data analyst vs data scientist"],"author_name":"kumar Gandharv","publish_date":"2021-06-01T17:00:00","publication_year":"2021","word_count":681,"keywords":["Go","API","data analyst vs data scientist","AI","data-driven","analytics","data governance","Rust","GAN","R","Databricks"],"extracted_tech_keywords":["AI","analytics","Databricks","R","Go","Rust","API","data governance","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dataops-goes-mainstream-as-atlan-lands-big\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054572,"title":"Wipro Staffed A Third Of Its Workforce to Cloud Business, Says CEO","content":"Wipro Chief Executive Officer Thierry Delaporte recently said that the company has staffed more than a third of its workforce to its cloud business, the company’s fastest-growing vertical, to cater to clients who are looking to shift applications from in-house servers to data centres. “Wipro employs over 80,000 cloud professionals, and more than 10,000 of our colleagues are now certified by leading cloud-service providers. We have employed more than 40,000 trained professionals in AI over the last few months, built a workforce of 7,500 specialists in the cybersecurity area,” said Delaporte. The company has won over $2.4-billion worth of deals in the past four quarters, a significant chunk of which was for cloud services that now comprise a third of Wipro’s deal pipeline and generate 70% of new revenue, as reported by Delaporte to analysts. In June, Wipro announced the launch of its FullStride Cloud Service portfolio and plans to invest $1 billion in cloud capabilities, technologies, acquisitions and partnerships over the next three years. Wipro is also ramping up its crowdsourcing platform TopCoder, which gives it access to another 1.6 million data science experts. Saurabh Govil, President and Chief Human resources officer at Wipro, added saying that client demand for certain skillsets far outstripped the available supply and the trend was likely to persist going forward. Wipro’s updated rewards programme for freshers now includes consistent value creation through a performance-based bonus and an increment programme. Wipro is the first top-tier Indian IT firm that has disclosed the amount of talent deployed in a specific growth vertical. The renewed focus on cloud clients has led it to heavily ramp up this employee base, which is in high demand across the industry.","excerpt":"The company has won over $2.4-billion worth of deals in the past four quarters, a significant chunk of which was for cloud services that now comprise a third of Wipro’s deal pipeline and generate 70% of new revenue.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Cloud Computing","Cloud Computing in IT Industry","Cloud Data AI","Cloud Platform","Data Science","Deep Learning","Machine Learning","private cloud","Wipro"],"author_name":"Victor Dey","publish_date":"2021-11-30T15:25:45","publication_year":"2021","word_count":281,"keywords":["Wipro","data science","Go","programming_languages:R","AI","R","Cloud Computing in IT Industry","Machine Learning","programming_languages:Go","Cloud Computing","private cloud","Deep Learning","Data Science","Cloud Data AI","Cloud Platform","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-staffed-a-third-of-its-workforce-to-cloud-business-says-ceo\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10055167,"title":"Interesting Research Papers Presented By Meta AI At NeurIPS 2021","content":"Meta AI researchers will be presenting a total of 83 papers at NeurIPS 2021. NeurIPS 2021 is the Thirty-fifth Conference on Neural Information Processing Systems held from December 6 to 14. The virtual conference boasts 2334 papers, 60 workshops, eight keynote speakers, and 15k+ attendees. In 2020, Meta AI (previously Facebook) presented 48 research papers at Neurips 2020. This article highlights the top 10 featured research publications by Meta at NeurIPS 2021. 1. Unsupervised Speech Recognition Authors: Alexei Baevski, Wei-Ning Hsu, Alexis Conneau, Michael Auli The research paper highlights the limitations of data labelling training for ML models. Therefore, the authors put forth in its place, wav2vec-U, short for wav2vec Unsupervised, a method to train speech recognition models without any labelled data. For this, the authors leverage self-supervised speech representations to segment unlabeled audio and learn a mapping from these representations to phonemes via adversarial training. Find the research paper here. 2. Bandits with Knapsacks beyond the Worst-Case Analysis Authors: Karthik Abinav Sankararaman, Aleksandrs Slivkins In this research, the authors present three original results beyond the worst-case scenario analysis of Bandit with Knapsacks (BwK). The results are built on the BwK algorithm from Agrawal and Devanur (2014), providing new analyses thereof. Firstly, the authors provide upper and lower bounds for a complete characterisation of logarithmic rates. Secondly, the authors consider the “simple regret” in BwK which tracks the algorithmic performance to prove that it is small in all but a few rounds. Finally, the authors provide a general “reduction” from BwK to leverage helpful structures and apply reduction to semi-bandits. Find the research paper here. 3. Volume Rendering of Neural Implicit Surfaces Authors: Lior Yariv, Jiatao Gu, Yoni Kasten, Yaron Lipman This research paper on computer vision and core machine learning improves geometrical representation and reconstruction in neural volume rendering. The authors take the approach of modelling the volume density as a function of the geometry to produce high-quality geometry reconstructions, outperforming relevant baselines. Find the research paper here. 4. Parameter Prediction for Unseen Deep Architectures Authors: Boris Knyazev, Michal Drozdzal, Graham Taylor, Adriana Romero Soriano The authors believe that deep learning has successfully automated the design of features in machine learning pipelines. However, the algorithms optimising neural network parameters remain largely hand-designed and computationally inefficient. Therefore, the authors study whether they can use deep learning to predict these parameters directly by exploiting the past knowledge of training other networks. Find the research paper here. 5. Learning Search Space Partition for Path Planning Authors: Kevin Yang, Tianjun Zhang, Chris Cummins, Brandon Cui, Benoit Steiner, Linnan Wang, Joseph E. Gonzalez, Dan Klein, Yuandong Tian The research paper develops a novel formal regret analysis for when and why an adaptive region partitioning scheme works. Furthermore, the authors propose a new path planning method, LaP3, which improves the function value estimation within each sub-region and uses a latent representation of the search space. Find the research paper here. 6. Antipodes of Label Differential Privacy: PATE and ALIBI Authors: Mani Malek, Ilya Mironov, Karthik Prasad, Igor Shilov, Florian Tramer The authors propose two novel approaches to preferential privacy ML where the trained model satisfies differential privacy with respect to labels of the training examples. The two approaches, the Laplace mechanism and the PATE framework, demonstrate their effectiveness on standard benchmarks. Find the research paper here. 7. NovelD: A Simple yet Effective Exploration Criterion Authors: Tianjun Zhang, Huazhe Xu, Xiaolong Wang, Yi Wu, Kurt Keutzer, Joseph E. Gonzalez, Yuandong Tian The research paper presents NovelD as an alternative exploration method to RND. The criterion called NovelD weighs every novel area approximately equally. The authors believe that the algorithm is very simple yet shows comparable performance or even outperforms multiple SOTA exploration methods in many hard exploration tasks. The researchers have also discovered that NovelD outperforms RND in many Atari Games. Find the research paper here. 8. Luna: Linear Unified Nested Attention Authors: Xuezhe Ma, Xiang Kong, Sinong Wang, Chunting Zhou, Jonathan May, Hao Ma, Luke Zettlemoyer The research paper proposes Luna, a linear unified nested attention mechanism that approximates softmax attention with two nested linear attention functions, yielding only linear (as opposed to quadratic) time and space complexity. The alternative to traditional attention mechanisms, Luna, introduces an additional sequence with a fixed length as input and an additional corresponding output, allowing Luna to perform attention operation linearly while storing adequate contextual information. Find the research paper here. 9. Interesting Object, Curious Agent: Learning Task-Agnostic Exploration Author: Simone Parisi, Victoria Dean, Deepak Pathak, Abhinav Gupta In this research paper on reinforcement learning, the authors propose a paradigm change in the formulation and evaluation of task-agnostic exploration. The authors suggest that the agent first learn to explore many environments without any extrinsic goal in a task-agnostic manner. Later on, the agent effectively transfers the learned exploration policy to explore new environments better when solving tasks. Find the research paper here. 10. DOBF: A Deobfuscation Pre-Training Objective for Programming Languages Authors: Baptiste Rozière, Marie-Anne Lachaux, Marc Szafraniec, Guillaume Lample In this research paper, the authors introduce a new pre-training objective, DOBF, that leverages the structural aspect of programming languages and pre-trains a model to recover the original version of obfuscated source code. The authors demonstrate how models pre-trained with DOBF significantly outperform existing approaches on multiple downstream tasks. The authors, during their research, also discovered that their pre-trained model could deobfuscate fully obfuscated source files, and suggest descriptive variable names. Find the research paper here.","excerpt":"The article explores the top ten featured research publications by Meta AI at NeurIPS 2021","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Meta AI"],"author_name":"Abhishree Choudhary","publish_date":"2021-12-10T12:00:00","publication_year":"2021","word_count":907,"keywords":["Meta AI","machine learning","TPU","AI","neural network","ML","computer vision","RAG","deep learning","differential privacy","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","Meta AI","differential privacy","RAG","TPU"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/interesting-research-papers-presented-by-meta-ai-at-neurips-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10173101,"title":"Why Top Data and AI Teams Trust Best Firms Certification","content":"Attracting top AI and data talent takes more than big pay cheques. Culture, values, and team dynamics matter just as much. Best Firm Certification gives companies the edge by spotlighting what today’s professionals really care about. AIM’s Best Firm Certification is an independent survey-based recognition at the employee level that evaluates companies based on four key parameters: benefits, identity, purpose and value quotient. Unlike traditional awards or rankings, this certification is grounded in authentic employee feedback, making it a credible and transparent measure of a firm’s workplace excellence. Upon successful completion of the survey, firms earn a globally recognised badge valid for one year, signalling to the market and prospective talent that they provide an exceptional work environment. Challenges Data and AI Teams Face Data and AI teams operate in an intensely competitive talent market, where companies often face challenges beyond just hiring. High turnover rates are standard, usually triggered by rigid or uninspiring workplace cultures. At the same time, the constant pressure to innovate demands environments that promote learning, experimentation, and collaboration. Cultural barriers can further dampen creativity and engagement, leading to burnout and attrition. These realities make it essential for organisations to build a workplace culture that not only attracts top talent but also supports and retains them for the long haul. What Makes Best Firm Certification So Coveted? Data scientists, engineers, and AI specialists are among the most sought-after professionals globally. These experts seek workplaces that not only challenge them intellectually but also align with their personal and professional values. The Best Firm Certification directly addresses these needs by: Highlighting employee-centric benefits: The certification’s ‘Benefits’ parameter ensures that firms offer competitive perks and support systems that matter most to technical talent. Showcasing strong company identity: A transparent and authentic identity attracts professionals who want to be part of something meaningful. Purpose-driven work: AI and data professionals increasingly want to contribute to projects that have real-world impact, reflected in the ‘Purpose’ parameter. Measuring value quotient: This reflects how much employees feel valued and recognised, a crucial factor for retention and motivation. Specialised Certifications for Data & AI Roles Best Firm Certification recognises the unique needs of different technical roles and offers specialised certifications for specific segments within the data & AI ecosystem: Best Firm Certification for Data Scientists Best Firm Certification for Data Engineers Best Firm Certification for AI Engineers Best Firm Certification for Women In Tech These focused certifications allow firms to demonstrate their commitment to creating optimal environments for each specialised group, further enhancing their appeal to niche talent pools. Benefits Beyond Recruitment While the certification serves as a powerful recruitment tool, its advantages extend far beyond hiring. The globally recognised badge acts as a mark of quality that firms can proudly showcase in marketing and recruitment campaigns, strengthening their employer brand. Survey feedback provides actionable insights into strengths and areas for improvement, enabling continuous improvement of workplace culture. Participation in the certification process signals to employees that their voices matter, boosting morale and loyalty. The Women In Tech certification highlights a firm’s commitment to fostering an inclusive environment, attracting diverse talent essential for innovation. Flexible Certification and Promotion Packages Firms can opt for certification at any time during the year, offering flexibility to fit organisational timelines. The certification includes various branding and promotional packages, allowing companies to maximise visibility and highlight their status as the best firm in the competitive data and AI market. Top data and AI teams worldwide trust Best Firm Certification because it provides an independent, transparent, and employee-validated benchmark of excellence. It empowers firms to differentiate themselves authentically, attract top-tier talent, and build motivated, high-performing teams that drive innovation and business success. Notable names include CEAT, Rakuten, Wipro, Intuit and others. According to AIM Research, these firms have seen up to 3x growth. Steps to Becoming Certified Becoming a Best Firms Certified organisation is straightforward: Application: Firms apply online at any time during the year. Employee surveys: A confidential survey is conducted to gather employee feedback. Analysis and reporting: Data is analysed to assess strengths and areas for improvement. Certification and recognition: Successful firms receive a one-year-valid certification badge and branding and promotional support. Is your organisation ready to join the ranks of elite data & AI firms recognised for outstanding workplace culture? Apply now for Best Firms Certification™ and showcase your commitment to creating an environment where the brightest minds thrive.","excerpt":"The certification evaluates companies based on four key parameters: benefits, identity, purpose and value quotient.","categories":["AI Features"],"tags":["best firms for data scientists to work"],"author_name":"Siddharth Jindal","publish_date":"2025-07-09T11:32:13","publication_year":"2025","word_count":731,"keywords":["programming_languages:R","AI","innovation","best firms for data scientists to work","Aim","ViT","Rust","GAN","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","Aim","R","Rust","GAN","ViT","innovation","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-top-data-and-ai-teams-trust-best-firms-certification\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10060726,"title":"The first-ever Turing Network Development Awards in AI","content":"As many as 24 universities in the United Kingdom won The Alan Turing Institute’s Network Development Awards for excellence in research excellence and a track record of translation in data science, AI, or a related field. “The awards reflect the demand across a range of sectors to work with the Institute. Data science and AI doesn’t stand still, and so we look forward to working together with this network of universities; exploring new ways to grow the UK’s dynamic research and innovation landscape,” said Adrian Smith, Director of The Alan Turing Institute. As the national institute for data science and AI, the Turing Institute already has a well-established network of university partners. However, these awards enable the Institute to extend its reach. The group of successful universities will be awarded up to GBP 25,000, to: Establish or grow an engaged and diverse community working (at all career stages) in data science and AI research and innovation at the university, who are aware of and will engage with potential opportunities and initiatives available across the Turing network.Identify and establish links between Institute priority areas and areas of interest and expertise at the university. Host activities and initiatives that are open to the wider data science and AI research and innovation community and\/or local and regional communities, to form new links and collaborations. Map the university’s expertise and strengths in each of the Institute’s priority areas and those considered of national strategic importance (in data science and AI) not yet covered by the Institute.Design plans for how the network will become sustainable for the future. The full list of successful applicants is below: Cardiff UniversityCity, University of LondonCreative Computing Institute, University of the Arts, LondonDurham UniversityGoldsmiths, University of LondonImperial College LondonKeele UniversityKing’s College LondonLondon School of Hygiene and Tropical MedicineNorthumbria UniversityNottingham Trent UniversityQueen’s University BelfastRoyal Holloway, University of LondonRoyal Veterinary CollegeTeesside UniversityUniversity of GlasgowUniversity of LiverpoolUniversity of NottinghamUniversity of PlymouthUniversity of ReadingUniversity of SheffieldUniversity of StrathclydeUniversity of SurreyUniversity of the West of England","excerpt":"The Turing Institute already has a well-established network of university partners.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science"],"author_name":"SharathKumar Nair","publish_date":"2022-02-15T19:15:08","publication_year":"2022","word_count":331,"keywords":["data science","Go","programming_languages:R","AI","innovation","programming_languages:Go","ViT","AI research","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","data science","R","Go","ViT","innovation","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/the-first-ever-turing-network-development-awards-in-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10082231,"title":"The Compelling Reason for Zoho to Not Launch IPO","content":"Sridhar Vembu’s Zoho Corporation has worn many a different hats since it was formed in 1996. Until 2009, the company was known as AdventNet Inc and provided network management software. Gradually, it stepped into the now-hot Indian SaaS space, which would shape what Zoho’s future would look like. Businesses then were increasingly moving to the internet to rent software for running their operations instead of purchasing licences for expensive software. The company’s youngest division, Zoho.com offered online software ranging from email to office productivity tools to customer relationship management (CRM) software, and reportedly remains the company’s fastest-growing segment. In a 2011 media interview, Vembu narrated a story about how a venture capitalist approached their booth at a trade show and cautioned the young company that ‘they were thinking too small’. But Vembu was self-assured. He believed he could earn some capital by selling products to companies and keep doing it until they could think bigger. At this point, Zoho Corporation had 600 employees. Vembu in his office in Mathalamparai village of Tenkasi, Source: Forbes In the past decade, Zoho has expanded its product portfolio. It has fifty applications currently under a single suite offering called Zoho One. This has worked in their favour considering the Chennai-based company touched USD 1 billion in revenue amidst a global economic slowdown that has affected the bulk of tech companies globally. Even though much has changed, Vembu has always been adamant about not taking Zoho Corp. public. The reason behind remaining private may seem even more baffling, given Zoho has been profitable every year since it was founded. The question has been asked several times, and Vembu’s answer has stayed the same. In a 2016 interview with Forbes India, he responded saying, “Why sacrifice freedom? We’ve done a lot here at Zoho and operate on our own clock – it’s an internal conviction. If you take money, you have to necessarily make some compromises.” The students of School of Design and School for Advanced Study are on an educational trip and reached Tenkasi yesterday! They can be seen interacting and working with the students at Zoho University Tenkasi. #TourDiaries #ZohoTenkasi #Travel #StudentLife pic.twitter.com\/PCi8lTEPDq— Zoho Schools of Learning (@ZohoSchools) December 20, 2018 Unique investments This ‘freedom’ has given Vembu the scope to invest in projects that normally have a long gestation period and even experiment. There’s Zoho University, the company’s internal training programme, which was born out of Vembu’s conviction that a ‘large pool of people were often overlooked by traditional methods of recruitment.’ Approximately 15 percent of the company’s hires are from Zoho University. Then there’s Zoho’s head office in Chennai which houses Zoho Schools which also gives students a stipend to attend classes. Vembu’s philosophy has often been repeated verbatim by the company’s senior and middle-level management. In August this year, Sridhar Iyengar, MD of Zoho Europe wrote a blog about the drawbacks of going public. Aside from current market volatility and general economic downturn, Iyengar spoke about other factors that companies normally overlook in exchange for the capital windfall and publicity that an IPO could potentially bring. Zoho Corporation’s product portfolio, Source: Zoho More money in R&D not marketing Iyengar also explained how public companies immediately turn risk-averse to placate investors. Illustrating the example of the biotech industry, he explained how scrutiny over R&D decisions could severely affect investor confidence and market valuation. This held back biotech firms from going public (in the past year, only five out of 102 biotech firms were trading above their debut price). \"Post-IPO Non-profits.\"Zoho used to run a billboard around 2018 in silicon valley along US 101 and that was the message.The share prices of unprofitable tech companies have gone parabolic in 2020.Thank you Fed for that perfect demonstration on how to inflate a bubble. https:\/\/t.co\/eh132KFcMY— Sridhar Vembu (@svembu) January 19, 2021 This was a legitimised fear as Zoho spends three times its marketing expenditure on R&D and has obtained 25 patents in the last three years. When Zoho reached the USD 1 billion-revenue mark, Vembu attributed much of the company’s success to its strong R&D capabilities and hinted they would continue to pour money here. “Our R&D focus in the coming years is to further unify our technology stack so that we are able to elevate the user experience,” he said. Zoho’s independence has also helped Vembu to retain interest in boosting rural businesses which could have otherwise raised investors’ eyebrows. In May, a Kerala-based startup Genrobotics which was developing robots for sewer cleaning received a Rs. 20 crore investment from Zoho Corp. (Genrobotics had developed Bandicoot, the world’s first robotic scavenger which helps clean sewers and manholes under the Kerala Startup Mission). Freshworks trading debut on Nasdaq, Source: The Economic Times Freshworks IPO fiasco The ongoing Freshworks fiasco seems especially stinging in this context. Founded by two former Zoho employees Girish Mathrubootham and Shan Krishnasamy, Freshworks was engaged in a two-year-old legal battle with Zoho under allegations that Freshworks had wrongfully used confidential information from Zoho. The lawsuit was settled last year in December. Last year in September, Freshworks became the first Indian SaaS firm to go public on the US stock exchange with a stellar debut of USD 43.5 per share on Nasdaq 21 percent more than the company’s listing price of USD 36 per share pushing its market cap to USD 12.3 billion. As Freshworks became the poster child for Indian tech IPOs, Zoho came under increased pressure for not going public. Just a month ago, Freshworks came under fire for misleading investors during its IPO. The California-headquartered startup has also been charged with ‘violations of federal securities laws.’ While Freshworks listed its shares at a premium of 21 percent, the stock has lost almost 75 percent of its value this year with Freshworks shares (‘FRSH’) currently trading at USD 12.32.","excerpt":"When Zoho reached the USD 1 billion-revenue mark, Vembu attributed much of the company’s success to its strong R&D capabilities and hinted they would continue to pour money here","categories":["AI Features"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-12-14T10:04:05","publication_year":"2022","word_count":969,"keywords":["Go","API","AWS","AI","IPO","venture capital","Git","ViT","R","startup"],"extracted_tech_keywords":["AI","AWS","R","Go","Git","API","ViT","startup","venture capital","IPO"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-compelling-reason-for-zoho-to-not-launch-ipo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":39892,"title":"Top 5 Recently Open Sourced Framework For Developers","content":"Over the last couple of years, tech giants have been open sourcing their projects so that both the companies, as well as the developer community can benefit from the same. According to a recent survey, nearly 53% of the companies have open source programmes or have plans to establish them within the next year. The survey also pointed out that nearly 59% of the respondents felt that open source programs are critical to the success of engineering and product teams. In this article, Analytics India Magazine will feature five of the recently open sourced frameworks for developers: CameraX: In the recently concluded Google I\/O summit, the company announced its new open source Android Jetpack library, CameraX, which makes camera development easier. According to the company, it provides a consistent and easy-to-use API surface that works across most Android devices, with backwards-compatibility to Android 5.0 (API level 21). Previewing image on the display, image analysis with the help of tools like computer vision and machine learning and high-quality image capturing capabilities are some of its features. Currently, Google is investing in an automated CameraX test lab that can detect a number of camera behaviours across a range of devices and all operating system flavours since Android 5.0 (API level 21). The Confidential Consortium Framework: Announcing its decision to open source the CCF framework, Microsoft stated that the platform will allow for building a new category of secure and high-scale confidential blockchain networks that can meet the key requirements of the enterprise. The framework will enable enterprise-ready computation or blockchain networks as it is built on trusted execution environments (TEEs), decentralized systems concepts, and cryptography. “CCF uses TEEs such as SGX and VSM to enable ledgers that integrate with it to execute confidential transactions with the throughput and latency of a centralized database. Confidentiality and high performance are key requirements of our enterprise customers,” Microsoft said in a blog. Space Partition Tree And Graph (SPTAG): Is Microsoft’s one of the most advanced AI tools used for Bing’s search algorithms and allows Bing users to take advantage of the intelligence from deep learning models to search through billions of pieces of information, called vectors, in milliseconds. BoTorch: On F8, Facebook’s annual developer’s summit, the social media giant open sourced BoToch. Built on PyTorch, the framework is a flexible, modern library for Bayesian optimization, a probabilistic method for data-efficient global optimization. “BoTorch provides a platform upon which researchers can build and unlocks new areas of research for tackling complex optimization problems,” Facebook said. BoTorch follows the same modular design philosophy as PyTorch and is capable of improves developer efficiency by utilizing quasi-Monte-Carlo acquisition functions by way of the “re-parameterization trick.” Pythia: Facebook in the recent past has announced a number of open source frameworks in an attempt to make its development more inclusive and transparent. In May it announced its decision to open source Pythia, it’s deep learning framework that supports multitasking in the vision and language domains. Its features include reference implementations to show how previous state-of-the-art models achieved related benchmark results and to quickly gauge the performance of new models. In addition to multitasking, Pythia also supports distributed training and a variety of datasets, as well as custom losses, metrics, scheduling, and optimizers.","excerpt":"Over the last couple of years, tech giants have been open sourcing their projects so that both the companies, as well as the developer community can benefit from the same. According to a recent survey, nearly 53% of the companies have open source programmes or have plans to establish them within the next year. The […]","categories":["AI Trends"],"tags":["Google","Open Source"],"author_name":"Akshaya Asokan","publish_date":"2019-05-30T07:00:17","publication_year":"2019","word_count":541,"keywords":["Go","API","machine learning","Open Source","AI","PyTorch","computer vision","deep learning","analytics","Google","Rust","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","analytics","PyTorch","R","Go","Rust","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-recently-open-sourced-framework-for-developers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10121559,"title":"‘For Us, it’s Never About Celebrating Tech for Tech’s Sake,’ says Microsoft chief Satya Nadella","content":"At Microsoft Build, the company’s annual developer conference, the tech giant was quite clear about what it wanted to achieve while Google was busy playing catch-up. “For us, it’s never about celebrating tech for tech’s sake. It’s about celebrating what we can do with technology to create magical experiences that make a real difference in our countries, in our companies, in our communities,” said Microsoft chief Satya Nadella, in his keynote speech. Responding to an old statement of Nadella where he said that he wants “people to know that we made them [Google] dance”, Google CEO Sundar Pichai gave a cheeky response in a recent interview with Bloomberg’s Emily Chang. “I think one of the ways you can do the wrong thing is by listening to the noise out there and playing someone else’s dance music,” laughed Pichai, extremely sure that they are indeed listening to their own music. Ironically, Google’s I\/O event looked like they were playing to OpenAI’s tune. Products such as Veo (text-to-video), Project Astra and few others seem to be a direct response to OpenAI’s products. Build 2024 While the 2023 edition released a few Copilot plugins and a Copilot assistance for Windows, this year saw the company focus heavily on Copilot developments. Source: Microsoft Build 2024 The new range of Copilot+ PC is not only poised to beat Apple’s MacBook Air M3, it also comes with a unique function called ‘Recall’ that allows one to find anything that they have searched for or done on their computer. How concerning this is, might be a topic for another time, but Microsoft has really gone all in to cater to enterprise and consumers alike. But, so has Google! AI for Everyone At Google’s annual developer conference, Google I\/O that took place last week (exactly a day after OpenAI’s Spring Update event), Pichai mentioned that despite investing in AI for over a decade at every layer of the stack, the company is still in the “early days of the AI platform shift”. However, he then went on to release a string of AI products, which would cater to everyone, “not just him or her”. If Copilot was Microsoft’s shield, Gemini was Google’s. The company’s integration of Gemini into their existing suite of products continued in this event. The company even launched Gemini 1.5 Flash, a lighter version of Gemini 1.5 Pro, and Gemma 2, the next version of Google’s open models. Small models seems to be the fad this tech season with Microsoft also releasing Phi-3-vision, which is the new multimodal small language model from the Phi-3 family. The model is set to be cost-effective and optimised for personal devices. The small language model is in line with Microsoft’s AI tech prediction for the year where SLMs have been touted to gain traction. Abu Dhabi’s Technology Innovation Institute also recently launched their small model, Falcon-2 11B. AI Agents are the Way Source: X “Soon you’ll be able to mix and match inputs and outputs. This is what we mean when we say it’s an I\/O for a new generation, and I can see you all out there thinking about the possibilities. But what if we could go even further? That’s one of the opportunities we see with AI agents,” said Pichai. Project Astra, defined as a universal AI agent helpful in everyday life, can process multimodal information, and respond naturally in conversation. Interestingly, at the OpenAI Spring Update, the company released GPT-4o, which pretty much does the same. Microsoft was not far behind on its AI agent agenda either. The company announced its partnership with Cognition AI, the makers of the autonomous software AI agent Devin that caused quite a storm when it was released a few months ago. Devin will be powered by Azure. However, Cognition AI was not the only major partnership announced. Together, We Stand Strong Microsoft announced a number of strategic alliances with partners from various industries. On the education front, Microsoft announced its partnership with Khan Academy, where the tech giant will provide free access to Khan Academy’s AI-powered teaching assistant for all K-12 educators. This will be supported by Azure OpenAI service. Unsurprisingly, Khan Academy also demonstrated its ChatGPT and AI-powered tool usage via demo videos during OpenAI’s event. Like every big-tech event, Build 2024 made a special mention of NVIDIA too. The company is planning to roll out a range of RTX-powered Copilot+ PCs. “We’re bringing the latest H200s to Azure later this year, and will be among the first cloud providers to offer NVIDIA’s Blackwell GPUs in B100 as well as GB200 configurations,” said Pichai. Source: Microsoft Build 2024 While NVIDIA is for everyone, the biggest trump card for Microsoft obviously is OpenAI, and Altman made his brief appearance here, despite skipping his OpenAI event. Though he did not reveal anything new, he indirectly hinted at the next version of GPT and mentioned the obvious that “models are getting smarter”. While both Microsoft Build and Google I\/O saw the release of new products, Google’s approach seemed to take on OpenAI alone. On the other hand, Microsoft’s event showcased a change in the company’s outlook emphasising a futuristic vision with strategic partnerships. “Microsoft is doing one thing that other cloud providers are either too scared or too complacent to try. They’re willing to cannibalise old products for the sake of a true AI-first strategy,” said AI advisor and entrepreneur, Allie K Miller. Compared to Microsoft’s array of products, Google’s future strategy seems obscure.","excerpt":"Microsoft is way ahead of Google, but the latter is still fascinated with OpenAI.","categories":["Global Tech"],"tags":["Google I\/O","Satya Nadella","Sundar Pichai"],"author_name":"Vandana Nair","publish_date":"2024-05-24T15:09:13","publication_year":"2024","word_count":912,"keywords":["Satya Nadella","Go","ChatGPT","TPU","OpenAI","AI","GPT-4o","R","Google I\/O","SLM","Ray","Azure","Sundar Pichai"],"extracted_tech_keywords":["AI","GPT-4o","ChatGPT","OpenAI","Ray","SLM","Azure","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/for-us-its-never-about-celebrating-tech-for-techs-sake-says-microsoft-chief-satya-nadella\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10163340,"title":"The Future of Content is ‘Headless’","content":"Content management has come a long way since the early days of the internet. In the early 2000s, websites were the sole digital channel, and managing content was a straightforward but rigid process. Then came the social media revolution, followed by the android boom between 2008 and 2009. However, the real game-changer was the rise of cloud software, which completely transformed how content is created, delivered and managed. Traditionally, launching a website meant installing a content management system (CMS), where content creation and presentation were tightly linked. Eventually, as digital landscapes evolved, businesses needed more flexibility. This is when headless CMS, a modern solution that separates content creation from its presentation, was introduced. Unlike traditional CMS platforms, headless CMS allows developers to use APIs to distribute content seamlessly across websites, mobile apps, and other digital platforms. This separation of content and design not only empowers teams to work independently but also speeds up content updates, enhances user experience, and ensures consistency across multiple channels. Role of AI in Headless CMS Artificial intelligence (AI) plays a significant role in content management. Nishant Patel, founder and CTO at Contentstack, said, “AI today is largely centred around generative AI, which excels at content generation. Many marketing websites and mobile apps leverage AI tools built into our software to create dynamic content effortlessly.” One of Contentstack’s AI-driven innovations is Brand Kit, a tool designed to capture a brand’s unique voice and tone. When companies use generative AI within their CMS, they ensure that the content aligns with their identity, which makes messaging more authentic and impactful. Another breakthrough product, Automate, is changing the game in content integration. Patel shared a compelling use case from Golfbreaks, a travel company specialising in golf vacation packages. Previously, compiling these packages was a manual, time-consuming task that took days. By integrating AI-powered LLMs like Llama, Gemini, and OpenAI into Automate, Golfbreaks reduced this process to mere hours. The system curates deals, structures them in a suitable format, uploads them to its website, and prepares them for approval – all in record time. Competition in the Field When it comes to the global adoption of headless CMS, North America is leading the charge. The US and Canada have been at the forefront because of their robust digital infrastructure, a thriving tech ecosystem, and a high concentration of businesses seeking agile content management solutions. Sectors like e-commerce, media, and technology have embraced headless CMS to deliver seamless content experiences across multiple platforms. This early adoption has spurred continuous innovation, with both startups and established players driving the evolution of headless CMS solutions. Speaking about competition, Patel said, “We focus on mid-sized companies and large enterprises like Fortune 2000 companies and major brands around the world that are using Contentstack to deliver their content. This includes websites, mobile apps, gaming software, and more.” He mentioned that in the traditional content management space, there are incumbents like Adobe, Sitecore, and Optimizely, and they have recently started making investments in headless CMS. What’s Next? The global headless CMS software market is on the rise. Valued at $0.71 billion in 2023, it’s set to skyrocket to $3.81 billion by 2032, growing at a CAGR of 20.5%, according to Business Research Insights. The surge is driven by businesses prioritising SEO, performance optimisation, and omnichannel content distribution. Headless CMS is becoming the backbone of modern digital experiences. The integration of AI, automation, and cloud-based collaboration is reshaping content management, making it more dynamic and future-proof. For developers, marketers, and business leaders, adopting a headless CMS is more than just an upgrade; it could be a transformative shift.","excerpt":"Headless CMS is a modern solution that separates content creation from its presentation.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","content"],"author_name":"Vidyashree Srinivas","publish_date":"2025-02-12T15:56:27","publication_year":"2025","word_count":601,"keywords":["Go","API","artificial intelligence","OpenAI","AI","ML","Git","RAG","content","generative AI","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","ML","generative AI","OpenAI","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-future-of-content-is-headless\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10048368,"title":"Countries Are Deploying More Robots For Police Patrolling &#038; How","content":"Imagine this – you are caught for cutting a red light, but you have a valid reason, and the police holding you cuts you some slack and lets you go. A robot surely won’t hear you out. But what if you bribe it to let you go? A robot surely won’t take it. This robot revolution has been supported by lighter mechanical parts, cheaper, better sensors and more sophisticated computers. The RoboCop debate is long and lengthy, but some countries are taking the chance on deploying robots for police patrolling. Let’s take a look at some of them. Singapore If you’re a robotics follower, you would have surely heard about the controversy surrounding Singapore’s latest member, Xavier. Singapore’s government is deploying robots to police and patrol for undesirable social behaviours – especially breaking COVID-19 rules. It was created and announced last week by Singapore’s Home Team Science and Technology Agency (HTX). The reveal included introducing two robots, Xavier, to patrol Toa Payoh Central – an area in Singapore known for its foot traffic. Xavier is equipped with a 360-degree camera allowing it to survey a huge landscape and feed the collected data into its AI system, flagging undesirable behaviour for human policies at the police station. So, while Singaporeans don’t have to worry about Xavier putting them in a handcuff or driving them to a police station, they should be aware of Xavier calling for police backup. The display on the robot shows a message to the person describing the wrong action they have done. Singapore is ranked as one of the safest countries in the world, and Xavier is the latest addition to ensure covid safety measures, prohibit public smoking and improper parking in its trial run. Singapore’s home affairs minister, K Shanmugam, hopes to have more than 200,000 police cameras by 2030, doubling the surveillance in Singapore. New York City In April this year, the NYPD tried and ended its contract with the Boston Dynamics firm Digidogs. This was after the device supposed to “save lives, protect people and protect officers” sparked major backlash in the US after footage of the robot’s use case patrolling public housing went viral. The police started using the four-legged device to go places that were too dangerous to send officers. The police had deployed the robot close to half a dozen times as a mobile camera during potentially hostile events, including a barricade and hostage situations. However, the robot dog that can climb stairs and has better video recording tools stopped when NYPD’s lease with Boston Dynamics ended in August. https:\/\/twitter.com\/AOC\/status\/1365021717144420354?s=20 2020 Tunisia Tunisia adopted a police robot last year to patrol areas in its Capital city, Tunis. The main aim of the robot police was to ensure that people were observing the COVID19 lockdown. Known as PGuard, the robocop is remotely operated. It is equipped with infrared & thermal imaging cameras and a sound & light alarm system. Less intrusive than the rest of the robot policies, Tunisia’s robocop looks out for people on largely deserted streets during the lockdown, approaches them and asks them why they are out. It then goes on to say, “What are you doing? Show me your ID. You don’t know there’s a lockdown?” Finally, the person’s ID is sent back to the police officers controlling the robocop. The robot’s Tunisian creator Anis Sahbani initially created the robot in 2015, operating on AI, and was carried forward by Sahbani’s Enova Robotics firm that sells the technology to overseas security companies. The company has plans of deploying another robot assisting coronavirus patients in communicating with relatives. While the robocop has received its share of controversy, The PGuard has become fairly popular on social media with positive user postings. For example, one of the popular videos features a man explaining to the robocop that he is out to buy cigarettes, to which the robot replied, “OK, buy your tobacco, but be quick and go home.” India During the COVID 19 pandemic, Hyderabad-based robotics company H-Bots introduced the RoboCop for the police department and a disinfectant robot, Accord. The company is now building an advanced humanoid six feet tall RoboCop. The cop can identify a person complaining, identify suspects, monitor temperature and stationary objects, and report to the police immediately. The robot is loaded with cameras, ultrasonic sensors, temperature sensors, and proximity sensors to record audio and video footage. It is built to be autonomously deployed in public places, malls, and airports to replace security guards. RoboCop can take over responsibilities as small as printing a boarding pass to as serious as diffusing bombs. With English language capabilities already in-built, RoboCop is being tested to interact in six different languages. The robots are priced between Rs 5 lakhs and Rs 20 lakhs, depending on the features. Florida Defense Department contractor Ghost Robotics is introducing four semi-autonomous robots in Florida. The robot will be unleashed in its Tyndall Air Force Base, introducing robot dogs to conflict zones. The military has claimed the Vision 60’s high tech canines as a leverageable tool to replace stationary surveillance cameras and enhance security. Ghost Robotics pictures the robot being used for responsibilities like working with bombs, scouting, and targeting next year. Ghost Robotics’ computerised canines are trained to operate in subzero temperatures that can move like real animals, climb steps, and run and turn themselves upright if knocked over. Inside a Robot Company A robocop manufacturing company, Knightscope, currently has robocops patrolling corporate campuses, malls, hospitals, and other public places. In 2019, its deployment in California was claimed to have caused a 46 per cent drop in crimes and a 27 per cent increase in arrests. Source: Knightscope Knightscope has ensured that their robots aren’t meant to actively police people or apprehend suspects, but serve as an extra set of eyes and ears, be a friend to fellow citizens, and deter criminals. While we may not see a big brother for a while, the future only sees more build upon these robots. How they will change the world is still to be seen.","excerpt":"Defense Department contractor Ghost Robotics is introducing four semi-autonomous robots in Florida.","categories":["AI Features"],"tags":[],"author_name":"Avi Gopani","publish_date":"2021-09-15T14:00:00","publication_year":"2021","word_count":1009,"keywords":["Go","API","programming_languages:R","AI","data_tools:Spark","programming_languages:Go","RAG","Aim","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","programming_languages:R","programming_languages:Go","data_tools:Spark","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/countries-are-deploying-more-robots-for-police-patrolling-how\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":29586,"title":"Understanding This Recent AI That Can Predict Alzheimer 5 Years In Advance","content":"In India alone, more than 4 million people suffer from some form of dementia. The most prominent of it all is Alzheimer’s, which India has the third highest caseload in the world, after China and the US. Usually targeting people with the age of 60 or 65 years, currently there is no cure to the disease. However, researchers across the globe are trying to predict the disease in advance. We recently covered a news on Indian-origin researchers developing AI that can predict risks almost five year in advance. This research titled ‘Modelling and Prediction of Clinical Symptom Trajectories in Alzheimer’s Disease Using Longitudinal Data’, has been carried out by Nikhil Bhagwat, M Mallar Chakravarthy, Joseph D Viviano and Aristotle N Voineskos, from the University of Toronto. In this article we are going to detail their research and list some of the key takeaways. How Was The Model Designed? Data for designing this machine learning model was taken from Alzheimer’s Disease Neuroimaging Initiative (ADNI) database and a replication analysis was done using subjects from Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL). Age, Apolipoprotein E4 (a class of protein) status, clinical scores from mini-mental state exam (MMSE) and Alzheimer’s Disease Assessment Scale (ADAS-13), and T1-weighted MR images were used in the analysis. The research also involved primary analyses on data of three visits of patients from ADNI with a timespan of longer than one year. The training set used for building the algorithm learnt from the data of over 800 people which involved normal, healthy seniors, to those depicting mild cognitive impairment as well as those with Alzheimer’s. Longitudinal Siamese Network (LSN) Model. Image source: Modelling and Prediction of Clinical Symptom Trajectories in Alzheimer’s Disease Using Longitudinal Data ADNI data was used to for primary analysis, for trajectory modelling and for prediction. Whereas, the AIBL data was used as the independent replication cohort for the process of prediction. The Algorithm Built And Its Working The algorithm built by the team to detect Alzheimer’s risk is named as Longitudinal Siamese Neural network (LSN). It has modules for combining multimodal data from two timepoints, and can learn from MRI, genetics and clinical data collected from ADNI. Longitudinal clinical and structural information combined with genetic information can be used to predict trajectories using machine learning algorithms. The advantage of longitudinal data is that it allows for the measurement within sample change over time and record the timing of various events. Analysis workflow of the longitudinal framework. Image source: Modelling and Prediction of Clinical Symptom Trajectories in Alzheimer’s Disease Using Longitudinal Data The machine learning model was primarily used for: Trajectory Modelling: This is where the modelling of the symptom trajectories is done. The goal is to group subjects with similar clinical progression so that a template for differential trajectories can be built. Six years of clinical data was used as input for  hierarchical clustering. The primary goal was not to discover unknown subtypes of Alzheimer’s Disease (AD) progression, but to create a trajectory prototype which will eventually be assigned to individuals. These prototype labels in turn help in prediction which involves diagnosis and also symptom detection at some point. In the end, trajectory templates are determined based on the clustering. This allows for a relatively simple way of dealing with missing data. Trajectory Prediction: After the symptoms trajectories are obtained using the model, multimodal and longitudinal data is used for prediction of symptom trajectories. The goal is to find out the prognostic trajectories as early as possible. The prediction models use information available from a single time point, called the baseline time point. But this lacks information regarding short-term neuroanatomy changes and clinical status, which is useful for the prediction process. To solve this, a model that combines data from the time plots and improve the prediction performance was developed. This model is called Longitudinal Siamese network (LSN). LSN was built with two design objectives: to combine the MR information from two timepoints that encodes the structural changes of different clinical trajectories. to incorporate genetic as well as clinical information with the structural magnetic resonance information in an effective manner. It was found that LSN could help in determining patients who are at a high risk. “At the moment, there are limited ways to treat Alzheimer’s and the best evidence we have is for prevention. Our AI methodology could have significant implications as a ‘doctor’s assistant’ that would help stream people onto the right pathway for treatment,” said Mallar Chakravarty of the research group. The work presented a data-driven approach for modelling long-term symptom trajectories derived solely from clustering of longitudinal clinical assessments. The resultant trajectory classes represented relatively stable and declining trans-diagnostic subgroups of the subject population. Cortical thickness was preferred as the magnetic resonance measure over characteristics like head size and total brain volume, because of its higher robustness. A Win Over Other Methods Attempts have also been made in the past to predict Alzheimer’s by methods such as logistic regression (LR), support vector machine (SVM), random forest (RF), and classical artificial neural network (ANN), but Longitudinal Siamese Neural network (LSN) shows an advantage over all others. For instance, the ANN model does not offer consistent performance for slow-declining trajectory prediction as it fails to predict that class in cross validation. On comparing LSN with other machine models mentioned above, it was found that the MRI information in combination with clinical and demographic data proved to be of a great value in LSN. It also shows the benefit of the follow-up time point information towards the prediction task to assist prioritization of MR data acquisition and periodic patient monitoring. Concluding Note Knowing Alzheimer’s in advance will provide a timely intervention of the disease and allow for a preliminary planning as well. Preventive treatments and interventions are an effective way out and this model can prove to be revolutionary if implemented. It would allow to work on new drugs by analysing the condition thoroughly, thereby keeping Alzheimer’s at bay.","excerpt":"In India alone, more than 4 million people suffer from some form of dementia. The most prominent of it all is Alzheimer’s, which India has the third highest caseload in the world, after China and the US. Usually targeting people with the age of 60 or 65 years, currently there is no cure to the […]","categories":["AI Features"],"tags":["ANN","modelling","MRI","prediction","SVM"],"author_name":"Disha Misal","publish_date":"2018-10-27T03:56:37","publication_year":"2018","word_count":998,"keywords":["Go","prediction","modelling","machine learning","programming_languages:R","AI","neural network","data-driven","IPO","Modal","ANN","Git","SVM","MRI","R"],"extracted_tech_keywords":["AI","machine learning","neural network","R","Go","Git","data-driven","IPO","Modal","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/understanding-this-recent-ai-that-can-predict-alzheimer-5-years-in-advance\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":66141,"title":"Increasing Cybersecurity Threat During Work From Home For Organisations","content":"For organisations, closing their doors and enabling employees to work from home in the wake of COVID 19, can be major workplace flexibility. However, for threat actors, it can be a newfound opportunity! Every time an employee connects to the enterprise network from home, they leave scope for threat actors to bank in the possible access points and get into the corporate system. Today, due to this current epidemic, a large number of devices are being used from home – mainly desktops, laptops, and smartphones – thus, giving hackers opportunities to target these vulnerable points. The current situation has created security concerns for Chief Information Security Officers (CISOs) across the globe who are reeling from bug disclosures and cyber-attacks and finding it difficult to maintain organisational cybersecurity posture. It seems that the safety net of ‘work from home’ policy has rather emerged as a watershed moment for cybercriminals to develop newer ways to barge in! Phishing emails claiming to hail from World Health Organization (WHO) to offer free healthcare kits or a ‘Deadly Corona Virus Map’ are real-time examples of how cyber attackers are making various organisations their new target. Lack Of Right Devices, Processes And Infrastructure For Remote Working The enterprise devices, despite being protected by internal safety measures, can become vulnerable due to several reasons: poor configuration of remote network connection and outdated security software. However, the biggest threat in this scenario is the usage of personal devices that in-house security teams can’t monitor. A large number of employees use their personal devices to perform organisational tasks from home and who knows whether their systems might be already infected by now. Simply put, organisations that don’t have employees’ home network infrastructure in control are certainly at risk! How To Tackle? While hackers across the world are already running phishing scams around COVID-19, organisations should train their employees on how to avoid suspicious emails and malicious links. As for now, they should stop sharing their personal and financial information with anyone via email or text or any online medium. In case someone asks for such details, call them to confirm and then forward the information over the phone. For internal communication, organisations should encourage professionals to use encrypted, enterprise-driven services like Virtual Private Network (VPN) instead of public-facing apps like WhatsApp to avoid becoming prey to threat actors. For its effective implementation, companies should release proper security guidelines and make sure its employees use the most recently patched\/updated version of the respective software. Besides these practices, one must also follow security fundamentals: avoid making notes of passwords on personal devices and ensure screen lock of personal devices. Although the worst security habits that employees naturally develop while working at home are not major, sometimes they can make home security network prone to threats. Organisations must also emulate their behaviours to remain productive, mirror their home setups to retain a work environment and keep daily schedules as consistent as possible. As in many areas of education and life, if you are standing still, the world does not stand still with you. Ensuring your team remains as prepared as possible requires setting aside time to train, even remotely. Training against real-world malware in a remote environment with remotely accessible tools is key to ensuring success while working remotely. With the outbreak of COVID-19 also giving rise to several new cyber threats, organisations across the world must keep in view these security measures and drive awareness among the workforce on how to fight such odds. While social distancing has become the new buzzword with people socialising online, it is crucial for everyone to take proper security measures from their end to avoid cyber risks and protect sensitive enterprise data from being compromised by threat actors. It’s time for us to stay safe, both offline and online.","excerpt":"For organisations, closing their doors and enabling employees to work from home in the wake of COVID 19, can be major workplace flexibility. However, for threat actors, it can be a newfound opportunity! Every time an employee connects to the enterprise network from home, they leave scope for threat actors to bank in the possible […]","categories":["AI Trends"],"tags":["Cyber Security"],"author_name":"Rakesh Kharwal","publish_date":"2020-05-28T13:00:00","publication_year":"2020","word_count":634,"keywords":["Cyber Security","programming_languages:R","AI","RAG","Aim","GAN","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/increasing-cybersecurity-threat-during-work-from-home-for-organisations\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10071625,"title":"Questioning the Legitimacy of DevOps Job Role","content":"Hiring is on an upswing, and a quick LinkedIn search would throw up jobs aplenty from companies like Siemens, Boeing, Philips, KPMG India, Zeiss Group, Thomson Reuters, and Groww India etc. One job position that hits you up often is that of a ‘DevOps Engineer’. The job title might intrigue those not well-versed with the technology sector. One might even appreciate the widening range of opportunities (read: job roles) technology has enabled. Amazon Web Services define DevOps as a combination of cultural philosophies, practices, and tools that increase an organisation’s ability to deliver applications and services at a high velocity — evolving and improving products faster than organisations using traditional software development and infrastructure management processes. Edtech platforms like UpGrad and Educba, which offer dedicated courses on DevOps, list the prerequisites and the roles and responsibilities of DevOps engineers. Educba even reasons out why we need a DevOps engineer. According to Educba, DevOps engineers facilitate smooth transitioning from software development to deployment. Generally, the development team works on Windows to develop modules, testers use Linux or Mac, while the production team uses a different environment altogether. This might give rise to several issues after deployment. This is where DevOps engineers play their role in facilitating code execution in any environment. So, Is DevOps a legit job role? Sharing a LinkedIn job posting for DevOps engineers at Accenture, Anub Sinha, founder and CEO of Opscale, pointed out how HR teams at various organisations are executing a “buzzword strategy” by posting such hiring alerts. Uri Zaidenwerg, a senior solutions architect at Blink Ops, having worked as a DevOps engineer at multiple organisations, suggests point blank that firms need not hire for a post like this. Now, coming from a DevOps engineer himself, it sounds quite absurd. These opinions have created quite a stir in the job market. So, has technology enabled fresh job roles? Does the role of a DevOps engineer offer something new and unique? Or is it just a hoax that companies are using to sound up-to-date? Well, if one were to analyse the hiring alerts posted by these companies seeking DevOps engineers, they would wonder if these engineers offer any new skill sets at all. For instance, a recent job posting for Lead DevOps Engineer by Zeta Suite lists “software engineers with a bent towards operations engineering or vice versa” as one of the criteria for applicants. In the Accenture job posting that Sinha pointed at, the company was looking for qualifications more suited to an operations engineer from a decade ago, when ‘development’ and ‘operations’ functioned in silos. “Job descriptions on job boards for the mythical DevOps engineer primarily detail the skillsets of build and automation engineers, and nothing whatsoever about transforming organisations to adopt frictionless working practices,” wrote Gary Woodfine, Technical Director at threenine.co.uk, an independent software vendor specialising in IoT, Field Service and associated managed services. A majority of such job advertisements require nothing more than Linux, Security and Build with some experience in working with Agile teams. Why is there no need to hire a DevOps engineer? DevOps is essentially a belief, an ideology and cultural change that lies at an organisation’s core. Transitioning into this culture requires a change in mindset across IT, which cannot simply be accomplished by hiring people dedicated to a certain role. In fact, to incorporate the culture of DevOps, it is wise for companies to hire candidates across different job roles based on their problem-solving skills, efficiency and proficiency in automating manual processes. “DevOps is a set of practices enabling high-quality software delivery, not a role. Needs more of a cultural change, and you cannot solve it with one quick ‘DevOps Engineer’ hire,” comments Sinha. The culture of DevOps gives development teams more control over shipping code to production. Apparently, the people currently executing the DevOps engineer’s role are either developers involved in deployment and network operations or system administrators eager for scripting and coding and engaged in development to fine-tune testing and deployment. Thus, they are people who have pushed beyond their areas of expertise and have an acquired holistic view of the technical environments. According to Stepan Pushkarev, co-founder and CTO at Provectus, DevOps is not the skills of one person. Instead, everybody in a development team must know Linux, Docker, Docker Compose, Kubernetes, and Ansible, at least on a user level, and understand networking and deployment architecture. With the introduction of new tools like Kubernetes, ELK stack and cloud infrastructure solutions, now many responsibilities that so-called “DevOps engineers hitherto handle” can be delegated to the developers. Zaidenwerg believes that if DevOps tasks are delegated to people who are not well-versed in coding, they will not be able to comprehend the real issues with the code and address them. According to him, it is the developers’ responsibility to meet users’ demands and provide them with the desired user experience. Thus, hiring engineers committed to their client’s needs is wiser than hiring DevOps engineers. Wrapping up Donovan Brown, Microsoft DevOps Program Manager, says, “DevOps is the union of people, processes, and products to enable continuous delivery of value to our end users. You cannot buy DevOps and install it. DevOps is not just automation or infrastructure as code. DevOps is people following a process enabled by-products to deliver value to end-users.” Zaidenwerg opines that in the future, all developers will need to understand and practise DevOps. According to him, developers would be more efficient at their job if they understood what it takes to run their applications and how to make them secure and scalable. Acknowledging the role of DevOps engineer, Zaidenwerg says, “People are hiring DevOps engineers. And people are working as DevOps engineers, and that’s the title, so yes, DevOps engineer roles exist. The question is should it exist? I believe it should not. Well, because it defeats the purpose of DevOps. DevOps is about removing the silos between developers and operations and having a DevOps engineer do the operations. That’s just renaming the system administrator, not removing any of the silos.” “I believe that in a few years from today, we will have enough automation and third-party services in cloud automation to make the DevOps engineer role easy enough to be done by developers without them investing too much time in it,” he concludes.","excerpt":"Having a DevOps engineer wouldn’t remove any silos","categories":["AI Features"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-07-27T14:00:00","publication_year":"2022","word_count":1049,"keywords":["Go","AI","docker","Scala","Git","automation","GAN","DevOps","R","kubernetes"],"extracted_tech_keywords":["AI","kubernetes","docker","R","Go","Scala","Git","DevOps","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/questioning-the-legitimacy-of-devops-job-role\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172018,"title":"Meta in Advanced Talks to Hire SSI Co-Founder Daniel Gross, Reports Reveal","content":"Meta is in advanced talks to hire prominent AI investors Nat Friedman and Daniel Gross to help lead its artificial intelligence efforts, according to a recent report by The Information. As part of the negotiations, Meta is discussing a partial buyout of its venture capital fund, NFDG, which holds stakes in top AI startups. The fund, worth billions of dollars on paper, could cost Meta over $1 billion to buy out. If the deal is successful, Gross, who co-founded Safe Superintelligence (SSI) with former OpenAI chief scientist Ilya Sutskever, would leave the startup to join Meta. At Meta, Gross is expected to focus primarily on AI products, while Friedman’s role is anticipated to be broader. Both Friedman and Gross are AI investors and have been involved in leading companies. They work closely with Meta CEO Mark Zuckerberg and Scale AI CEO Alexandr Wang, who joined Meta in a $14.3 billion deal. Friedman was the CEO of GitHub (acquired by Microsoft) and is now an investor and advisor, particularly through his venture fund NFDG. As per the report, he has also been consulting with Meta on its AI strategy since 2014. In addition to the buyout, Meta is negotiating the purchase of a substantial portion of NFDG’s holdings, gaining minority stakes in the startups backed by the fund, such as SSI, but without gaining control or access to sensitive information about these companies. Meta has been aggressively pursuing top AI talent. In a recent podcast, OpenAI CEO Sam Altman said that Meta is offering $100 million signing bonuses to attract talent. He also added that none of their top people have accepted those offers yet.","excerpt":"Both Friedman and Gross are influential in AI investing.","categories":["AI News"],"tags":["Meta"],"author_name":"Aditi Suresh","publish_date":"2025-06-19T12:51:20","publication_year":"2025","word_count":274,"keywords":["Go","API","Meta","artificial intelligence","OpenAI","AI","venture capital","Git","GitHub","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","OpenAI","R","Go","Git","GitHub","API","startup","venture capital"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-in-advanced-talks-to-hire-ssi-co-founder-daniel-gross-reports-reveal\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10063620,"title":"IIT-Mandi to organise Robotics &#038; AI summer camp for school kids","content":"The Indian Institute of Technology Mandi is organising the summer camp on Robotics and Artificial Intelligence (AI) in collaboration with Himachal Pradesh Kaushal Vikas Nigam (HPKVN), Shimla. This summer camp, to be held in July 2022, will give students an opportunity to learn Robotics and AI, an ever-growing field with hands-on experiments-based learning. Registration is open to all students enrolled in classes 11 and 12 at select schools in Himachal Pradesh. Enrolments will be accepted through schools only. Individual applications will not be entertained. An entrance test based on 10+2 level math, physics, and general aptitude will be conducted on May 1, 2022. Students will be selected for the camp based on the results. Top 100 selected students will learn: 1) Basic Electronics including, microcontrollers, basics of electronics, using sensors, using DC motor, stepper motor and the servo motor with a microcontroller, network, and Internet of Things (IoT), Wifi and Bluetooth. 2) Basic Programming with basics of any programming language (data types, operators, basic math, conditional statements, loops, functions). 3) Introduction to C++\/Python with reference to programming a microcontroller. They will learn to use TinkerCAD to get familiar with programming controllers and the basics of Linux. 4) Fundamentals of Machine Learning and AI that will teach students essential applications like object detection, image segmentation, etc.5) Introduction to Jetson Platform, use of NVIDIA Jetson Nano to use Machine Learning models like YOLO, ResNet, etc. with no prior knowledge of AI","excerpt":"Registration is open to all students enrolled in classes 11 and 12 at select schools in Himachal Pradesh.","categories":["AI News"],"tags":["IIT","iit mandi","Robotics","school"],"author_name":"Kartik Wali","publish_date":"2022-03-25T16:59:56","publication_year":"2022","word_count":239,"keywords":["artificial intelligence","machine learning","AI","ML","ResNet","Robotics","Python","iit mandi","C++","object detection","GAN","school","IIT","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","object detection","Python","R","C++","GAN","ResNet"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-mandi-to-organise-robotics-ai-summer-camp-for-school-kids\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10117399,"title":"What to Expect at the Highly Anticipated Apple WWDC June 2024","content":"Apple announced the date for the year’s most anticipated event – Worldwide Developers Conference (WWDC). The Cupertino giant will host the event online from June 10 through June 14, giving developers and students the opportunity to attend the opening day event in person in Apple Park. All eyes are on how the company will finally embrace AI at the upcoming event. Embracing ‘AI’ not ‘ML’ “Mark your calendars for #WWDC24, June 10-14. It’s going to be Absolutely Incredible!” posted Apple SVP of marketing Greg Joswiak, and then started a surge of comments where users were quick to catch the subtle messaging. Source: X If the deduction is true, this is Apple’s way of embracing the word ‘AI’ at the developers’ conference, a stark difference from last year’s event where Tim Cook made sure that AI was never mentioned, instead ‘machine learning’ was embraced. OpenAI- Apple Partnership The much anticipated announcement where Apple will finally embrace ‘AI’ or ‘Apple Intelligence’ will be through OpenAI. ChatGPT capabilities is said to be integrated into Apple’s existing applications including voice-assistant Siri. More features is set to be revealed at the event. Generative AI Features for Apple Devices In the last earnings call, CEO Tim Cook said that the company will continue to ‘spend a tremendous amount of time and effort’ on AI and technologies that will shape the future. “We’re excited to share the details of our ongoing work in that space later this year,” he said, possibly hinting at the upcoming WWDC. With Apple’s recent push to bring generative AI capabilities into their devices, it is evident that these features will be unveiled at WWDC. With the release of MM1 paper, a new family of multimodal AI models, with the largest being a 30B parameter model, the focus is on how this would fit in the WWDC announcements. Jason Snell, writer and former editor-in-chief at Macworld, highlighted in a recent podcast episode of MacBreak Weekly that Apple might not have its LLM ready for WWDC, but it will demonstrate AI capabilities utilising third-party models in a privacy-focused manner. Interestingly, partnerships with competitor big tech companies are already in place. Apple was recently in talks with Google to integrate its most powerful AI model, Gemini, into the iPhone. Apple is also partnering with Chinese tech giant Baidu to bring further GenAI features. iOS Revamp for Cooler Features The generative AI features are reported to be integrated onto the upcoming iOS 18 software. It is also predicted that the latest software would be a major update of the operating system. Apple had earlier described its upcoming OS as ‘ambitious and compelling’, with new features, designs and improved performance and security. – iOS18 is expected to bring various new features to Apple’s in-built apps. Apple Music is set to have auto-generated playlists for users based on mood and interactions. – Apps such as Pages and Keynotes will have AI-assisted writing features. – Xcode, Apple’s code development platform, is also likely to get AI features to assist with coding. – Apple Maps will likely see more features such as support for topographic maps and option to save custom routes. – The new operating software might bring a change in the design and layout of the phone, such as customisable home screens. WWDC24 is set to unveil these new features in the latest iOS, iPadOS, macOS, watchOS, and tvOS. It is also suggested that iOS 18 will be redesigned to match vision OS. Advanced Apple Vision Pro Apple’s legendary announcement at last year’s WWDC was hands-down their spatial computing device Apple Vision Pro. As confirmed in the WWDC blog, the revolutionary mixed reality headset’s in-built capabilities that can be resonated with that of an autonomous vehicle, is expected to receive upgrades. It is likely that the problems reported with the maiden version of the headset may be addressed. Apple Pencil is also being tested out as an accessory for the next version of visionOS. The pencil is currently compatible with iPads only. Currently, sold in the US, Apple Vision Pro will likely be sold in other countries too, including China. AI Chip Integration Earlier this month, Apple announced the integration of M3 chip, Apple’s most powerful AI chip, on their latest 13- and 15-inch MacBook Air. However, the chip that was unveiled in October last year, is not yet available on Mac Studio and Mac Mini devices. It is likely that WWDC will see announcements around M3 chip integration on these products. Further, the Apple A18 chip which is the forthcoming processor from Apple that is expected to be used in the iPhone 16 lineup, is said to be equipped with a powerful neural engine, and is reported to have a 6-core GPU. The A18 chip will allow for more powerful and efficient AI performance. New and Improved Siri Last year, it was reported that Apple’s voice assistant Siri will receive major upgrades this year. AI-enabled features will be integrated into Siri for improving conversation capabilities and user personalisation. Siri and other Apple apps such as Messages will be seamlessly connected for better response options.","excerpt":"“Mark your calendars for #WWDC24, June 10-14. It’s going to be Absolutely Incredible!” says Apple SVP of Marketing Greg Joswiak.","categories":["AI Features"],"tags":["Apple"],"author_name":"Vandana Nair","publish_date":"2024-06-10T18:41:20","publication_year":"2024","word_count":847,"keywords":["Go","ChatGPT","GenAI","machine learning","OpenAI","AI","Apple","ML","generative AI","multimodal AI","R"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","GenAI","multimodal AI","ChatGPT","OpenAI","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-to-expect-at-the-absolutely-incredible-apple-wwdc-2024\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129445,"title":"Microsoft Inaugurates New Innovation Hub in Bengaluru","content":"In an effort to expand its footprint in India, generative AI powerhouse Microsoft has opened a new innovation hub in Bengaluru. Puneet Chandok, president of  Microsoft India and South Asia took to LinkedIn to share the news. “Embrace a culture of learning and innovation with our maker spaces for prototypes and special projects. Meet our senior architects, who will steer you through bespoke technical engagements focused on transformative business outcomes” added Chandok. The company has over 20,000 employees across 10 Indian cities – Ahmedabad, Bengaluru, Chennai, Gurugram, New Delhi, Noida, Hyderabad, Kolkata, Mumbai and Pune. It has over 14 Microsoft Innovation Centres (MICs) across the country which are part of strategic partnerships with leading academic institutions and aim to grow tech skills. The company has announced several collaborations in the generative AI space in India over the last couple of months. For example, it teamed up with Skillsoft to create an AI training program for enterprises, leveraging Skillsoft’s AI Skill Accelerator to teach the use of Microsoft AI tools, including Copilot and Azure Open AI. Additionally, Indian IT giant Tech Mahindra is collaborating with Microsoft to implement Copilot for Microsoft 365 across 15 locations, aiming to improve efficiency for over 10,000 employees. Tech Mahindra will also use GitHub Copilot for 5,000 developers, expecting a 35-40% productivity boost. Back in February of this year, Microsoft co-founder and philanthropist Bill Gates visited the Microsoft India Development Center (IDC) in Hyderabad, a hub of innovation that he envisioned 25 years ago. He expressed his optimism for India’s unique potential in AI and the company’s strategic focus on harnessing the country’s talent for upcoming features in this space. The company is partnering with Indian startup Sarvam AI, specialising in Indic LLMs. It is also set to upskill two million Indians in AI by 2025.","excerpt":"The company has over 20,000 employees across 10 Indian cities – Ahmedabad, Bengaluru, Chennai, Gurugram, New Delhi, Noida, Hyderabad, Kolkata, Mumbai and Pune.","categories":["AI News"],"tags":["AI","GCC","Generative AI","Google","India","Microsoft"],"author_name":"Shritama Saha","publish_date":"2024-07-18T12:10:00","publication_year":"2024","word_count":300,"keywords":["Go","GCC","AI","R","Git","RAG","Aim","ViT","generative AI","Google","Generative AI","GitHub","Azure","India","Microsoft"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","Azure","R","Go","Git","GitHub","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-inaugurates-new-innovation-hub-in-bengaluru\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054527,"title":"Actions Taken For Ethical AI In 2021","content":"With the biggest technological revolution going on, a lot of good has already been done for society with the help of artificial intelligence. One of the finest examples is the use of AI data analysis to combat the coronavirus outbreak. Companies do provide customised experiences to their customers, harnessing the power of AI. At the same time, they’re scaling their regulatory, reputational and legal risks. Cases such as Los Angeles suing IBM for gathering and selling users’ location data to marketing companies, data scandal by Cambridge Analytica and Facebook, the case of Ryan Abbott and team on AI be granted as an inventor — all these raise ethical concerns associated with AI. To that end, we are listing down some of the crucial steps taken towards ethical AI in the year 2021 for an inclusive, accountable and transparent future ahead. Global Agreement on Ethics of AI All 193 member states of the United Nations Educational, Scientific and Cultural Organization (UNESCO) have adopted a historic agreement that defines common principles and values required to ensure the healthy development of artificial intelligence. The adopted text marks a major step towards guiding the construction of the required legal infrastructure to ensure the ethical development of AI technologies. Further, it will help to reduce the risks it entails. The world has seen an increased gender and ethnic bias, significant threats to privacy, dangers of mass surveillance, dignity and agency, and increased use of unreliable artificial intelligence technologies in law enforcement, to name a few. As of now, there were no universal standards to provide an answer to these issues, as per UNESCO, in a statement. Tech Giants saying “No” to Unethical AI Projects Three of the leading tech players, namely IBM, Google and Microsoft, have turned down projects shadowed by ethical concerns. Google cloud experts agreed not to move forward with the idea of creating AI for financial institutions on making decisions for lending money. The project is put on hold until the concerns regarding gender and racial biases are resolved. Microsoft, too, limits the use of its software that mimics voice amid concerns of using the technique for creating deep fakes. Similarly, sensing the possibility of misuse, IBM discontinued its face recognition services altogether. The earlier trend with the tech giants releasing AI technologies such as facial recognition and chatbots directly in the market without any due diligence for potential biases or downsides has seen a reversal. As more people and ethical AI organisations have begun to speak out against AI’s ethical concerns, tech players have formed ethics committees to examine their new products. European Commission’s Approach towards Trustworthy AI To develop human-centric and secure AI for its people, the European Commission has proposed the first legal framework on AI, with experts equating it with the General Data Protection Regulation, aka GDPR. With a risk-based approach, the proposed regulation classifies AI into four categories – unacceptable risk, high risk, limited risk and minimal risk. The deciding factor behind the grouping depends upon the scale and extent of bias or risks associated with the technology. The concerns are real; take, for instance, the case of a Detroit man who was wrongfully taken into custody for shoplifting in a fiasco related to facial recognition. However, the US government has asked the EU to oversee that AI must not be overregulated. Way forward The battle to become an AI superpower is enticing right now, with every country vying for dominance in the field through a technological breakthrough. As a result, countries such as China, the United States, Japan, Canada, Singapore and France, to name a few, are investing heavily in AI research and development. However, there are currently no norms or standards in place for ethical AI research, design, or use. The time is well-suited to look to the fact that machines making decisions related to individual rights must be able to explain their decision, and if objected, can be reviewed by competent human authority. In addition, companies need to be transparent while deploying AI, people should be informed about any such usage by the companies, and an effective redressal mechanism must be put in place to address any discriminatory approach.","excerpt":"Countries are investing heavily in AI research and development. However, there are currently no norms or standards in place for ethical AI research, design, or use.","categories":["AI Features"],"tags":["AI for social good","Ethical AI","Trustworthy AI"],"author_name":"kumar Gandharv","publish_date":"2021-11-30T16:00:00","publication_year":"2021","word_count":695,"keywords":["Go","artificial intelligence","programming_languages:R","AI","chatbots","Ethical AI","Trustworthy AI","AI for social good","Rust","GAN","cloud_platforms:Google Cloud","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","chatbots","R","Go","Rust","GAN","AI research","cloud_platforms:Google Cloud","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/actions-taken-for-ethical-ai-in-2021\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10095296,"title":"Google Warns Staff About Using Bard","content":"In a surprising turn of events, Google, one of the leaders in the AI race, is cautioning employees about using chatbots. This includes the company’s self-developed Bard. Meanwhile, the company has been going around the world promoting the chatbot. According to individuals familiar with the matter and confirmed by the company, Alphabet, the parent company of Google, has instructed its employees not to input sensitive company documents into AI chatbots. This directive aligns with the company’s long standing policy of ensuring the protection of confidential information. Moreover, the company has also advised its engineers to not use code directly generated from these AI softwares, and wants to be more transparent with the limitations of the technology. Google’s caution against its own chatbot raises several concerns about the privacy and security aspects of these large language models. Google was earlier promoting the use of Bard for its employees. Google’s updated privacy note also tells users to not include confidential and sensitive information in Bard conversations. This comes after a lot of companies telling their employees to not use chatbots in their work. Samsung banned the use of ChatGPT for its employees citing the same reasons as Google. Apple had restricted its employees from using ChatGPT at work and sharing confidential information. Amazon had also issued similar guidelines for its employees. On the other hand, Microsoft’s leaked document read that the company is fine with its employees using ChatGPT as long as they do not share internal information and sensitive data with the model.","excerpt":"The company has also advised its engineers to not use code directly generated from these AI softwares","categories":["AI News"],"tags":["Google Bard"],"author_name":"Mohit Pandey","publish_date":"2023-06-16T12:38:35","publication_year":"2023","word_count":252,"keywords":["Go","ChatGPT","programming_languages:R","AI","Google Bard","chatbots","GPT","llm_models:Bard","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","chatbots","R","Go","GPT","llm_models:GPT","llm_models:ChatGPT","llm_models:Bard","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-warns-staff-about-using-bard\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10110773,"title":"This USC Professor from Kolkata is on a Journey to Civilise AI","content":"The present-day AI system is smitten with hallucinations – for some, it is a feature (and not a bug), and for others, it’s frustration. For Amitava Das, a professor at the University of Southern Carolina, it was a worry that lead to the creation of what he calls ‘Civilising AI or Constitutional AI’. “AIs [LLMs] are very smart and at the same time, they are also stupid. And we do not know how to handle these in either case. On the other hand, AI-generated content becomes so eloquent that it is impossible to detect it. No Turing test is able to do it,” said Das, in an interview with AIM, sharing the origin of Civilised AI, a framework to proactively measure the hallucination of a model, alongside mitigating and managing risks associated with it. With a PhD in natural language processing (NLP) from Jadavpur University in Kolkata, Das has been working in the field for more than 20 years, which he says was all because of his passion for working in AI ever since the beginning. Having two academic postdocs from Europe and the USA, he is currently a professor at the Artificial Intelligence Institute of South Carolina (AIISC). Das started the AI research lab at Wipro and is still an advisor there. Having worked across geographies, and along with IITs and universities world over, he has also created a joint PhD program for industry practitioners to earn doctorate degrees. He’s currently focused on reducing hallucinations within language models. There are several research papers and models coming up to solve the AI detection problem. GPTZero, DetectGPT, and RoBERTa baseline are all trying to do the same, but a lot still needs to be done. The most recent paper, Ghostbuster, also works on solving the same problem. That is why Das and his team coined the term ‘Civilised AI’ to explain how to handle the problem of hallucinations, and at the same time make it easier to detect if a text is AI generated, and tackle the misinformation and the possible harm it poses to the society. Single metric to civilise AI Using the example of Google’s Bard and how the company lost around $100 million because of its hallucinations, Das emphasised the need to control what AI is writing. “On the one hand, we need AI to be creative when it comes to fields like journalism as it helps in writing. But on the other hand, if I am a lawyer working on a legal case, hallucination is not allowed at all,” he explained. Das’s paper, titled The Troubling Emergence of Hallucination in Large Language Models – An Extensive Definition, Quantification, and Prescriptive Remediations, addresses this problem of balancing hallucinations with eloquence within AI models, along with ‘Counter Turing Test CT^2: AI-Generated Text Detection is Not as Easy as You May Think — Introducing AI Detectability Index’, the paper that won the Outstanding Paper award at EMNLP 2023. “We call this dichotomy a factual mirage and silver lining, how much of each is allowed,” explained Das, and that is what led his research team to introduce new metrics called AI detectability index, AI adversarial attack vulnerability Index, and AI hallucination index. “AI is just a balance between these three,” he added that there is still not a clear-cut solution for this, but the team is trying to derive a formula for this, which is called the Civilising Index, an amalgamation of all of this. He said it is a difficult scenario to tackle, illustrating with Einstein’s mass-energy equivalence relation, which led to the creation of the atomic bomb. He also gave examples of other AI-related research that needs to be regulated if it involves risk to human lives such as bioinformatics. “It is the people who created the atomic bomb and the people who created atomic energy. So there are people who should make AI models better, and then there is my team who will make better detection techniques,” he added. “Regulation has to come to specific use case, not in a broad way,” he asserted. Each field such as medicine, military, legal, etc need to have separate licences on how they can use AI to tackle the hallucinations, leading to emerging risks within the systems. The need for Indic LLMs Das is also passionate about how language models can be made for Indic languages and has released a code-mixing model called CONFLATOR for the purpose. He started working in this field around eight years back and is regarded as one of the pioneers of the field by Microsoft Research. He explained that the problem that comes with Indic languages and datasets is that it is very hard to understand. “Hindi is getting mixed into English or English is getting mixed into Hindi,” Das explained about how it is very difficult to understand the switching point in languages as most of the data is from the Internet. “That is why we introduced Conflator,” he added, which is still a complicated field that deals with introducing electromagnetic waves into a neural network for code-mixing for positional encoding formulation. When it comes to Indic language models “we do not have a lot of monolingual data in India. This discussion has been going on between NITI Aayog and IIIE for several years, before the launch of ChatGPT and others,” Das was part of the discussions with the Indian government for making Indic language models possible. That is why he believes that Indian models are still far behind English models in terms of power and detectability. Das told AIM that he is closely working with the Indian government and IIIT Hyderabad and is soon going to publish a paper on the topic.","excerpt":"Amitava Das has been working in NLP for over 20 years and is currently hell-bent on reducing hallucinations within LLMs.","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Mohit Pandey","publish_date":"2024-01-15T14:08:03","publication_year":"2024","word_count":948,"keywords":["Go","ChatGPT","artificial intelligence","AI","neural network","RAG","NLP","Aim","Rust","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","NLP","ChatGPT","Aim","RAG","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-usc-professor-from-kolkata-is-on-a-journey-to-civilise-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136088,"title":"AIM 100: The Most Influential Global Leaders in AI","content":"As we unveil the inaugural AIM 100: The Most Influential Global Leaders in AI 2024, it’s evident that artificial intelligence is at a critical juncture in shaping our world. Through engaging with leaders across the AI spectrum, we’ve seen firsthand how this technology holds immense promise while also presenting significant challenges. These individuals are at the forefront of navigating ethical dilemmas, technological hurdles, and societal impacts to ensure that AI develops in a way that benefits all of humanity. Our mission in creating the AIM 100 is to highlight these trailblazers who are pushing the boundaries of what’s possible with AI. This list brings together a diverse group of visionaries—from industry giants and innovative entrepreneurs to influential thinkers—who are not only advancing technology but also grappling with issues like data privacy, algorithmic bias, and transparency. They face the daunting task of steering AI through uncharted territories while keeping ethical considerations at the forefront. Compiling this list was both a challenging and inspiring journey. Our dedicated team delved deep into the contributions of each honoree, uncovering the passion and dedication that drive their work. The challenges they face are immense: balancing rapid technological advancement with the need for responsible oversight, ensuring inclusivity in AI development, and addressing public concerns about the impact of AI on jobs and society at large. We invite you to explore the stories and achievements of the AIM 100 honorees. Their collective efforts not only push the boundaries of technology but also inspire us to consider how we can harness AI for the greater good. As we look ahead, we remain optimistic that with such dedicated individuals leading the way, the future of AI will be both innovative and ethically grounded. Write to us at info@aimmediahouse.com AI Pioneers Elon Musk CEO of Tesla, SpaceX, X, and Neuralink Sridhar RamaswamyCEO of Snowflake Garry TanCEO of Y Combinator Adam D’ AngeloCo-founder and CEO of Quora Mukesh AmbaniChairman of Reliance Industries Ali GhodsiCo-founder of Databricks Linda YaccarinoCEO of X Jamie DimonCEO of JPMorgan Bhavish AggarwalCEO of Ola Cabs and Electric Romesh WadhwaniFounder of Wadhwani AI Clem DelangueFounder of Hugging Face Andy JassyCEO of Amazon Julie SweetCEO of Accenture Yann LeCunChief scientist at Meta AI Daniela AmodeiCo-founder and president of Anthropic Pat GelsingerCEO of Intel Thomas KurianCEO of Google Cloud Peter ThielCo-founder of PayPal and Palantir Andrew NgCo-founder of Coursera Naveen RaoVP of generative AI at Databricks Kevin ScottCTO at Microsoft Yiftach ShoolmanCo-Founder of Redis Peng XiaoCEO of G42 AI Creators CP GurnaniFormer CEO and managing director, Tech Mahindra Jerry LiuCEO of LlamaIndex David HolzCEO of Midjourney Subbarao KambhampatiAI Professor at Arizona State University Michael TruellCo-founder and CEO at Anysphere Alan CowenCEO and chief scientist at Hume AI Harrison ChaseCEO of LangChain Ruslan SalakhutdinovUPMC Professor, Machine Learning Soumith ChintalaAI Engineering Lead at Meta Fei Fei LiCo-founder of World Labs Logan KilpatrickProduct Manager at Google DJ PatilGeneral Partner in GreatPoint Venture Emad MostaqueFormer CEO of Stability AI Mo GawdatChief AI officer at Flight Story Mitesh M KhapraHead of AI4Bharat Research Lab at IIT Refik AnadolAmerican-Turkish visual AI artist Ahmad Al-DahleHead of Meta’s Llama Team Thomas DohmkeCEO of GitHub Veezhinathan KamakotiDirector of IIT Madras Greg BrockmanCo-founder of OpenAI Ian GoodfellowResearch Scientist at DeepMind Pushmeet KohliHead of research at DeepMind Vivek RaghavanCo-founder of Sarvam AI Chris LattnerCreator of LLVM Sri AmbatiCEO of H2O.ai Sharon ZhouCo-founder of Lamini Amjad MasadCEO and founder of Replit Yejin ChoiComputer science researcher Maxime LabonneMachine Learning Scientist Liquid AI David HaFounder of Sakana AI Jeff DeanChief Scientist at Google DeepMind Amitabh NagCEO of Bhashini Sebastian RaschkaResearch Engineer at Lightning AI Aidan GomezCEO of Cohere Rodrigo LiangCEO of SambaNova Systems Scott WuCo-founder of Cognition AI AI Influencers Bojan TunguzQuadruple Kaggle Grandmaster & Former software engineer at NVIDIA Arvind NarayananDirector of Princeton Web Transparency and Accountability Project Louis BouchardFounder Towards AI Bindu ReddyCEO and co-founder of Abacus.AI Marc RaibertExecutive Director, The AI Institute Tom FishburneFounder of Marketoon Studios Debarghya DasVenture Capitalist at Menlo Ventures Joscha BachAI strategist at Liquid AI Allie K MillerCEO of Open Machine Margaret BodenResearch Professor at the University of Sussex Jeremy HowardFounder of Answer.AI Yuval Noah HarariAuthor and Philosopher Stephen WolframCEO of Wolfram Research Pedro DomingosProfessor at the University of Washington Lex FridmanResearch Scientist at MIT Alex SmolaCEO of Boson.ai Ronald van LoonCEO of Intelligent World Nick BostromPhilosopher Gary MarcusCo-founder Geometric Intelligence Daniela RusLeading Roboticist and Computer Scientist Toby WalshChief Scientist at UNSW.AI Nikita BierCo-founder of Politify Emmett ShearCo-founder of Twitch Irene SolaimanHead of Global Policy at Hugging Face Chamath PalihapitiyaCEO of Social Capital AI Disruptors Oren EtzioniProfessor at the University of Washington Kate CrawfordPrincipal researcher at Microsoft Research Arundhati BhattacharyaCEO and chairperson of Salesforce India Ashton KutcherAmerican actor, AI-optimist Ed Newton-RexFounder of Jukedeck Ganesh RamakrishnanLead BharatGen, Chair Professor, Dept of CSE, IITB, Co-founder bbsAI Ray KurzweilAmerican computer scientist Pragya MisraLead Public Policy Maker in OpenAI Rajan AnandanManaging Director at Peak XV Partners Kai-Fu LeeCEO of Sinovation Ventures Tim O’ReillyFounder of O’Reilly Media Rajeev ChandrasekharFounder of BPL Mobile Reid HoffmanCo-founder and Executive Chairman of LinkedIn Pushpak BhattacharyyaIIT Professor Contributed Sentimental Analysis in NLP Sandhya ArunCTO at Wipro Marc AndreessenCo-founder of Andreessen Horowitz Credits Reporters: Siddharth Jindal, Mohit Pandey, Vidyashree Srinivas and Pritam BordoloiProject Editor & Research: Amit Raja NaikEditor: Moushmi SinhaDesigner: Raghavendra RaoDigital: Pabitra Kumar Moharana","excerpt":"Introducing the AIM 100: The Most Influential Global Leaders in AI. As AI reshapes our world, these individuals stand at the forefront, addressing ethical challenges, technological hurdles, and societal impacts. This list celebrates visionaries—from industry leaders to innovative thinkers—who are advancing AI while navigating critical issues like data privacy, algorithmic bias, and transparency. Compiling the AIM 100 was an inspiring journey, revealing the dedication driving AI’s responsible development. Explore their stories and achievements as they guide AI toward a future that balances innovation with ethical considerations for the benefit of all.","categories":["AI Highlights"],"tags":["AI leaders","AIM","Elon Musk"],"author_name":"AIM Media House","publish_date":"2024-10-01T12:35:33","publication_year":"2024","word_count":868,"keywords":["Anthropic","Meta AI","artificial intelligence","machine learning","OpenAI","AI","LlamaIndex","Elon Musk","AI leaders","NLP","LangChain","generative AI","AIM"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","generative AI","OpenAI","Anthropic","Meta AI","LangChain","LlamaIndex"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/aim-100-worlds-most-famous-people-in-ai-2024\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10050141,"title":"Airtel To Invest 5000 Crore To Scale Up Its Data Centre","content":"Bharti Airtel Limited, known as Airtel, recently unveiled a refreshed brand identity, ‘Nxtra by Airtel’, for its data centre business and outlined investment plans to significantly scale up its data centre network to serve the requirements of India’s fast-growing digital economy. Airtel announced that its subsidiary Nxtra will invest Rs 5,000 crore by 2025 to scale up its data centre business. Nxtra by Airtel aims to create new data centre parks in key metro cities. The investment will triple Nxtra by Airtel’s installed capacity to over 400 MW to meet the surging demand and consolidate its network leadership. “Airtel has built the largest data centre network in India, and we are now doubling down on this business to scale up our network that will be at the core of 5G and Digital India. Our experience of operating secure data centres, deep brand trust in the enterprise segment and the ability to deliver end-to-end digital transformation solutions positions us well to serve the emerging requirements of India’s connected economy. The new brand identity embodies this vision and ambition,” said Ajay Chitkara, Director and CEO, Airtel Business. Nxtra by Airtel has the largest network of data centres in India. It currently operates ten large and 120 edge data centres located strategically across India and manages critical submarine landing stations. Coupled with Airtel’s global network, it offers secure and scalable integrated solutions to global hyper scalers, large Indian enterprises, start-ups, SMEs and governments. Sustainability will continue to be its big focus, given the huge energy requirements of data centres. Nxtra by Airtel is already aggressively scaling up the use of green energy for its data centres and aims to source 50% of the power requirements of these centres through renewable sources as part of Airtel’s overall GHG emission reduction targets. Nxtra by Airtel also recently commissioned captive solar power plants in Uttar Pradesh & Maharashtra, with still more in the pipeline. Airtel Business, the B2B unit of Airtel, is India’s leading provider of information communication technology services and offers a diverse portfolio of products and services covering voice, data, collaboration, work from home solutions, cloud, data centre, cyber security, IoT, network integration, managed services, enterprise mobility and digital media across a range of industries.","excerpt":"The investment will triple Nxtra by Airtel’s installed capacity to over 400 MW to meet the surging demand and consolidate its network leadership.","categories":["AI News"],"tags":["data centre","data centre india","Data Science","Deep Learning"],"author_name":"Victor Dey","publish_date":"2021-10-01T13:44:55","publication_year":"2021","word_count":371,"keywords":["data centre india","Go","programming_languages:R","AI","digital transformation","Scala","Git","Aim","programming_languages:Scala","Rust","Deep Learning","Data Science","R","data centre"],"extracted_tech_keywords":["AI","Aim","R","Go","Rust","Scala","Git","digital transformation","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/airtel-to-invest-5000-crore-to-scale-up-its-data-centre\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070304,"title":"Why Quickbooks did not work out in India","content":"Intuit, an accounting software provider, has decided to pull its QuickBooks product from India after a decade of operation here. QuickBooks’ products and services for accountancy and small businesses will no longer be available from January 31, 2023. The company stated that it is not accepting any new subscriptions in India – it applies to all the services like QuickBooks Online, QuickBooks Online Accountant, QuickBook Time, and QuickBook mobile app. It has further asked its customers to download their data and transition out of the product. This development seems too sudden, and the company has not offered any specific reason for it yet. “​​This difficult decision to discontinue QuickBooks has been made as the company can no longer deliver and support a product that meets the needs of customers in India,” the official statement says. The company has four million customers globally, with less than 1-2 per cent in India. For now, Intuit’s decision to pull QuickBooks from India, which affects 30,000 businesses, is seen to be a great opportunity for competitors like Zoho and Tally to fill in. Intuit QuickBooks’ India journey QuickBooks’ suite of offerings includes cloud accounting, inventory management, cash flow management, and invoicing. It also provides an online practise management solution for chartered accountants through QuickBooks Online Accountant. Intuit introduced QuickBooks in India in 2012, but even before that, the cloud-based accounting solution was running in alpha and beta versions for 12 months. When it entered the India market, QuickBooks positioned itself to target 2 million broadband-connected small businesses. At that time, Tally was QuickBooks’ biggest competitor, having a strong presence in the small and medium space. Over the years, QuickBooks gained popularity in the Indian market. Some of its selling features include affordability, accessibility, ease of use, ease of installation and maintenance. In an earlier interview with Analytics India Magazine, Saurabh Saxena, India site leader and vice president, Product Development at Intuit, said that the India team has been responsible for building several foundational capabilities of QuickBooks Online Advance products, like workflow automation platform, reporting platform, roles-based access, power tools for power customers, customisation, among others. In one of the major milestones, QuickBooks was made GST compliant in 2017. This helped customers to create transactions with the new GST tax rates without disruption and customise invoice templates to display specific information. In the latest development, Intuit India announced an agreement with Visa in March this year to strengthen business propositions for SME customers in India. It was touted to be the first of its kind for Intuit QuickBooks in India. This agreement was to develop joint value propositions and thought leadership in the SME space. Possible reasons for QuickBooks withdrawal Like several other industries, SMEs are rapidly digitalising their processes – digital bookkeeping, payments, inventory management, and delivery. SMEs are boosting their investment in technology. Despite the opportunities that this field offers, it also has unique challenges which may be difficult to traverse. Given what constitutes SMEs – from local grocery to suppliers to big brands, creating a one-fit solution does not make sense. It is important for SME-focused players to have specific knowledge of the requirements and the understanding of challenges to offer appropriate solutions. Over the years, several homegrown accounting software vendors have entered the market – Zoho being one of the prime examples. The Indian government’s AatmaNirbhar Bharat campaign may have played a role in promoting these homegrown businesses. Huge opportunities for competitors Soon after Intuit made the announcement regarding QuickBooks, Zoho announced that it is open to supporting and serving customers of QuickBooks and helping them transition smoothly to their platform. “Zoho Books has been in the market since 2011. We have a very strong customer base in India which is rapidly growing. We are hyper-focused on strengthening and expanding our operations in the country. We are fully committed to helping businesses take care of their finances and fulfil their GST and any other regulatory-related obligations. We will work continually to serve businesses by providing them with a world-class, future-proof financial management solution which they can rely on while growing their business. We would be glad to support QuickBooks customers in India looking for an alternative solution. We have a dedicated team focused on making the transition as smooth as possible,” Prashant Ganti, head of products – Tax, Accounting & Payroll, Zoho said. Another major competitor that has been around in the market even before QuickBooks is Tally. It currently commands more than 75 per cent of the market share and has 2 billion users. Apart from accessibility and ease of use, Tally’s popularity mainly comes from the familiarity factor and user acceptness. With the exit of a major competitor, Tally’s popularity is set to soar even more.","excerpt":"The company has four million customers globally, with less than 1-2 per cent in India.","categories":["IT Services"],"tags":["Intuit","SMEs"],"author_name":"Shraddha Goled","publish_date":"2022-07-01T18:55:15","publication_year":"2022","word_count":787,"keywords":["Go","API","AI","SMEs","Intuit","Git","automation","analytics","disruption","GAN","workflow automation","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Git","API","GAN","automation","workflow automation","disruption"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-quickbooks-did-not-work-out-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10121459,"title":"Canva Debuts AI-Powered Enterprise Suite to Challenge Adobe","content":"At Canva’s first international Canva Create event in Los Angeles, the Australian design studio launched Canva Enterprise, a subscription service designed to meet the needs of large organisations.It supports scalable growth with increased seats and cloud storage, making it suitable for global use. The platform aims to centralise design, content production, AI, and collaboration tools, reducing costs and complexity. It also includes updated brand controls to ensure consistency across all brand elements and templates. Additionally, enterprise-level security tools such as MFA, SCIM, SSO, and Canva Shield protect organisational assets. “As demand for visual content soars, navigating organisational complexity is more challenging than ever. We democratised the design ecosystem in our first decade and now look forward to unifying the fragmented ecosystems of design, AI, and workflow tools for every organisation in our second decade,” said Melanie Perkins, co-founder and CEO of Canva. However, in addition to Canva Enterprise, the company has also made major announcements, leading to AI features in Magic Studio, a new user experience, and more. Bolstering Visual Suite and Magic Studio AI Products Magic Studio has received new AI upgrades. Now, it can create graphics, icons, and illustrations from text prompts with Magic Media and Magic Design, which produces high-quality presentations. New photo editing offers AI-powered tools for seamless image editing. “Magic Studio works on internally-developed AI and ML algorithms that leverage a combination of foundational AI models from our team, including Kaleido, and a variety of partners like OpenAI, Google, AWS, and Runway,”  Danny Wu, head of AI products at Canva, told AIM in a conversation earlier this year. On the other hand, Visual Suite now better supports organisational needs like suggesting editing for tracking changes and collaborating on document integration with third-party apps like Amazon Ads, Google, and Meta for design feedback. It also includes data autofill, which automatically fills designs with business data from sources like Salesforce. Additionally, the Bulk Create feature streamlines marketing workflows by updating multiple designs at once using CSV or Excel files. New User Interface Canva has completely redesigned its core product to improve editing and collaboration. The new interface claims to include a streamlined editing experience with a contextual toolbar and tools like the background remover and Magic Studio AI. The customisable homepage allows users to pin favourite designs, folders, and templates for quick access and has an advanced search feature for finding content easily. Tailored Tools for Every Team Canva has introduced specialised tools for different departments such as marketing, HR, sales, and creative teams. Called Canva Work Kits, it consists of customisable templates tailored to each department’s needs. Extending Affinity Following Acquisition Over the past six years, the company has steadily built its AI capabilities and made major acquisition. In March 2024,  it acquired  UK based creative software suite, Affinity. So, the latest Affinity version 2.5 introduces advanced editing features to enhance professional performance. Key upgrades include variable font support, stroke width tool and support for ARM64 Chips. It also purchased Kaleido in 2021. Canva Courses for Workplace Learning Like every other tech firm, Canva is investing in upskilling It has launched Canva Courses to improve workplace learning by turning designs into interactive courses to support employee onboarding, upskilling, and development, with progress tracked via a central dashboard. Canva Vs Adobe With over 170 million users, the company reached unicorn status in 2018, with a $40 million investment valued at $1 billion. In May 2022, Walt Disney CEO Bob Iger became an angel investor in the Blackbird Ventures and Sequoia Capital-backed company. In 2022, it took a leap of faith and entered the AI race, offering stiff resistance to competitors like Microsoft Designer and Adobe. Adobe has also diversified its portfolio into a generative AI-powered enterprise software platform, introducing Firefly to Photoshop and launching features like Generative Fill and Generative Remove for advanced image editing. Today, the design giant introduced new AI features in Lightroom, including generative removal for photo editing and Lens Blur to improve editing speed. Powered by Adobe Firefly, these tools work across phones, desktops, and web.","excerpt":"In addition to Canva Enterprise, the company has made other major announcements, leading to AI features in Magic Studio, a new user experience, and more.","categories":["Deep Tech"],"tags":["Adobe","Canva","Design","Generative AI"],"author_name":"Shritama Saha","publish_date":"2024-05-23T22:00:00","publication_year":"2024","word_count":673,"keywords":["Go","OpenAI","AI","Adobe","AWS","ML","Scala","Canva","RAG","Aim","generative AI","Generative AI","R","Design"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Aim","RAG","AWS","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/canva-debuts-ai-powered-enterprise-suite-to-challenge-adobe\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10109587,"title":"Reliance Jio Partners with IIT Bombay to Build &#8216;Bharat GPT&#8217;","content":"Mukesh Ambani’s Reliance Jio Infocomm has forged an alliance with IIT Bombay for the ‘Bharat GPT’ initiative, according to revelations made by the company’s chairman, Akash Ambani. This collaboration, rooted in a partnership established in 2014, underscores a concerted effort to propel technological advancements. At IIT Bombay’s annual Techfest, Ambani unveiled the company’s visionary roadmap, painting a picture of a far-reaching “ecosystem of development” and signalling the imminent advent of “Jio 2.0.” by leveraging large language models and generative AI, as indicated by Ambani. “AI stands for Artificial Intelligence, but it also stands for All Included”Shri Akash Ambani at IIT Bombay Techfest.He shares his views on the transformative power of AI, innovations and initiatives at Jio to make India proud. #IITBombay #Techfest #AI #Technology #India #Jio pic.twitter.com\/Ru8zFVoyxF— Reliance Jio (@reliancejio) December 27, 2023 This comes after Reliance had announced its partnership with NVIDIA in September for advancing AI india. This was for developing LLMs and nurturing India’s generative AI applications. Further, NVIDIA said that it will provide access to Reliance with GH200 Grace Hopper Superchip and NVIDIA DGX Cloud for exceptional performance. Foreseeing a paradigm shift catalysed by AI across diverse industries, Ambani outlined plans to integrate AI both vertically within the organisation and horizontally across multiple sectors. The announcement also hinted at the prospect of introducing an operating system tailored for televisions, providing a glimpse into the multifaceted nature of the forthcoming technological innovations. The expansion blueprint for Reliance Jio encompasses the introduction of novel offerings in media, communication, commerce, and devices, reflecting a comprehensive strategy to diversify and enhance its portfolio. Ambani expressed unwavering confidence in the transformative potential of 5G private networks, envisaging a future where enterprises of all scales can harness the capabilities of a 5G stack. In underlining India’s potential as a forthcoming “innovation centre,” Ambani projected a lofty ambition of achieving a $6 trillion economy status by the close of the decade. He reiterated the ubiquitous role of AI in shaping society and industries, emphasising Jio’s commitment to national development, with monetary gains deemed a “byproduct” rather than the primary objective. In a compelling call to action, Ambani urged young entrepreneurs to embrace risk-taking and actively contribute to societal well-being, underscoring the company’s ethos of aligning its pursuits with the broader national interest. Interestingly, there is another BharatGPT created by CoRover already in the market which is also doing exactly the same thing, and is also planning to launch chatbot soon in partnership with Google Cloud.","excerpt":"This comes after Reliance had announced its partnership with NVIDIA for using GPUs.","categories":["AI News"],"tags":["Akash Ambani","BharatGPT","IIT","Jio","Jio AI Cloud","Reliance"],"author_name":"Mohit Pandey","publish_date":"2023-12-28T10:55:15","publication_year":"2023","word_count":412,"keywords":["Akash Ambani","Jio","Go","API","artificial intelligence","BharatGPT","AI","innovation","RAG","GPT","Jio AI Cloud","generative AI","GAN","Reliance","IIT","R"],"extracted_tech_keywords":["AI","artificial intelligence","generative AI","RAG","R","Go","API","GPT","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/reliance-jio-partners-with-iit-bombay-to-build-bharat-gpt\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004673,"title":"Are Jobs In Analytics Looking Up?","content":"The increasing impact of COVID pandemic has tremendously created the worst recession and economic crisis of all time. Economists believe that India is going to face the worst possible recession in 40 years, where thousands are and will be losing jobs due to the massive impact of the pandemic lockdown. That being said, despite lockdown and the impactful recession, there are a few specific domains across the IT spaces that have continued to augment at a steady pace. One of the critical fields out of the lot is data analytics, which has not only gained new impetus but also added value in lessening the impact of the pandemic among various industries. A recent Analytics India Magazine study has revealed that open jobs in data analytics skills have the highest proportion at 33.7%, surpassing machine learning and cybersecurity. A lot of this could be attributed to the trend of utilising data to develop new business models and products amid these uncertain times. This has urged companies to have a constant demand for data analytics professionals regardless of the economic situations and their financial conditions. As of today, there are 97,131 data analytics open jobs advertised across all industries in India. Open job distribution by skills and role types. Data analytics jobs have the highest proportion at 33.7%, which is followed by machine learning with 20.4% and cybersecurity with 15.4% Also Read: Latest Data Science And Analyst Jobs To Apply For Data Analytics Dominates Open Jobs Before COVID pandemic era, the data analytics domain significantly experienced tremendous growth — in 2019-20. In fact, previous studies highlighted how the space of analytics is growing at a massive rate in terms of revenue, investments and salaries. According to numbers, the functions of analytics in India had earned a total revenue of $35.9 billion which is a 19.5% growth from the previous year. A lot of this could be attributed to rising demand for analytical services among sectors like BFSI, marketing, eCommerce, healthcare, media as well as telecom industries. However, the advent of lockdown brought in a massive downturn for the majority of the industries but the IT space, which continued to develop at a steady pace. While data analytics has seen significant growth before the lockdown, analysing data for making business decisions has continued to be an essential requirement even during and after lockdown. Explained by the fact that data is continuously being generated across various mediums for different segments, and this data is not only required in the pre-COVID era but also during the recession. As that happens, most of the large IT companies, along with medium-size IT companies including TCS, Accenture, Capgemini and Cognizant have also started offering end-to-end analytical services. This, in turn, has raised the requirement of hiring more data science and analytics professionals for organisations. Apart from data analytics, some other critical functional areas that showed resilience against recession are cloud computing, cybersecurity, artificial intelligence, and deep learning. While machine learning holds around 20% of the open jobs, cybersecurity takes up approximately 15% of the market. Remaining AI, deep learning, IoT as well as blockchain have lower proportions in the entire open jobs pool. Additionally, this rise in demand for data analytics jobs is also due to the high salary that the space is willing to offer the professionals. According to the AIM research, the median salaries offered to AI professionals for open jobs have turned out to be the highest at ₹12.1 lakhs. But, despite having a drop from last year, the data analytics space offers a significantly high median salary of ₹9.9 lakhs. The decline in salary margin, compared to the median salaries already being drawn by professionals, is because of the financial turmoil of many companies, halting many of their operations and services. Coming to the skill development opportunities in data analytics, the research stated that salary distribution and experience level required for the data analytics open jobs have the highest proportion. This highlights many job opportunities for recent graduates as well as experienced professionals of the field. Data is currently the almighty to survive this pandemic; and with it being generated across industries like healthcare, BFSI, eCommerce as well as electronics, numerous job opportunities have become apparent for experienced professionals. According to numbers, considering the security of end-to-end operations has always been the key concern. Therefore, the role of risk analysts will see the maximum growth of 10.1% annually, along with other analytical roles like IoT analysts. Also Read: Are Analytics Jobs Recession-Proof? Here’s What Data Scientists Think Wrapping Up With the pointers mentioned above, it can be established that despite the unfortunate pandemic, the demand for data analytics jobs roles has remained resilient, in fact, continuing to look up. This is not only because of the immense data that is generated due to the shift towards digitisation, but also the salaries that are offered for these roles.","excerpt":"The increasing impact of COVID pandemic has tremendously created the worst recession and economic crisis of all time. Economists believe that India is going to face the worst possible recession in 40 years, where thousands are and will be losing jobs due to the massive impact of the pandemic lockdown. That being said, despite lockdown […]","categories":["AI Features"],"tags":["AI Jobs","analytics jobs","analytics jobs in India","analytics jobs india","different types of analytics","fields of analytics","types of analytics"],"author_name":"Sejuti Das","publish_date":"2020-08-12T14:00:00","publication_year":"2020","word_count":813,"keywords":["data science","Go","artificial intelligence","machine learning","fields of analytics","AI","cloud computing","AI Jobs","analytics jobs in India","analytics jobs","Aim","deep learning","analytics","analytics jobs india","R","types of analytics","different types of analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","Aim","cloud computing","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/are-jobs-in-analytics-looking-up\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10087771,"title":"Meta’s New Feature will Make You Rethink Your Online Presence","content":"Meta’s recently announced subscription service may not be as innocuous as it seems to be, especially when fared against what Twitter is doing. Known as ‘Meta Verified’, the feature allows users to verify their account using a government-issued ID. This begs the question: What business does Meta have collecting users’ personal information? This rings several alarms. Why? Because this is not the first time that Meta is under the scanner for misusing user information. In the past, the company faced several accusations of leaking user data, most notably in the Cambridge Analytica scandal. The insistence on personal identification in addition to the subscription fee suggests that Meta’s focus is on content regulation. Tech giants, including Instagram, had previously sought the assistance of the US Congress in determining the age of their users. Several doubts were raised about this plea, since these platforms possess vast amounts of user data and are more than capable of verifying their users’ age, if their intentions were to do so. Jeff Chester, the director of the Center for Digital Democracy, expressed scepticism saying, “They want Congress to give them a political out and say do the age gating and all the problems are gone versus regulating the business model”. Therefore, it won’t be a stretch to assume that integrating government verification into user profiles could serve as a political solution for Meta—eliminating the need for the company to rely on the veil of social media anonymity and user privacy to protect itself from the often-troubled relationships it has with governments. Dr Ramanand Nand, Director at Center of Policy Research and Governance, expressed similar apprehension on the blurring lines between government and social entities. “I am against the governmentalisation of social media, mainly because we cannot place trust on these companies, especially considering that in the past we have seen cases where they misused users’ identity information”, Dr Nand told AIM. Other side of the argument However, there are few who see the present situation rather differently. For instance, Ranjeet Rane, a technology policy consultant, told AIM, “Across the globe, we see governments, both democratic and otherwise, seeking to regulate this space with an increasing sense of urgency. The verification process—based on the submission of identity and other documents—can help in weeding out misinformation campaigns, a long standing issue faced by social media platforms.” Furthermore, Rane added that legally there is no law prohibiting these social media companies from monetising account verification. In fact, the revised Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021, has a provision for significant social media intermediaries to enable account verification for users, albeit voluntarily. Louis Rosenberg, the CEO of Unanimous AI, offered his opinion on whether companies should have access to users’ personal data. He said, “I suspect the use of government ID is simply the most convenient and verifiable way to prove identity. It’s generally why governments issue IDs.” Digital Identity: A Preparation for metaverse? Digital identity is touted to be the most essential component in how we will operate in the social media world, or the metaverse, in years to come. “Subscription identity verification could be even more important in the metaverse where it is critical to know that the person you are interacting with is precisely who they claim to be,” Louis Rosenberg told AIM. During the 2023 World Economic Forum (WEF), the argument in favour of the proposition was overwhelmingly affirmative, and it went as follows: “To achieve this frictionless state, good system-wide interoperability of the metaverse should consider interests such as privacy, security and safety. Given the borderless nature of the metaverse, multistakeholder and multilateral collaboration will be required to reach consensus on design choices, best practices, standards and management activities.” The multiple stakeholders mentioned above include not only creators and users, but also governments, businesses and civil society. Each user will be assigned a unique avatar linked to their digital identity, serving as a universal identifier for accessing various experiences in the physical and digital worlds. The 2018 World Economic Forum insight report highlighted that digital identities will control our access to products, services, and information and act as a gateway, dictating what we can and cannot access. The WEF has been in favour of “fostering passporting” in the metaverse, while also pushing for a “democratise adoption,” in an “equitable way”. Speaking at a press conference in WEF 2023, Assistant Director General for Strategy and Innovation at the Office of the Prime Minister of the UAE, Huda Al Hashimi initially said, “Regulators will be acting more like referees than gatekeepers and code of conduct will take precedence over formulating policies”. However, in discussing the aspect of “passporting”, she also seemed to lean on regulations and standards for global interoperability in the metaverse. Therefore, while many globalists see ‘digital identity’ as a one-stop solution for all crises, there is no clarity over what it really aims to do and in what ways would it facilitate a “democratic” social landscape. Nevertheless, it will be interesting to see if the Meta verification programme is leading to Meta’s aspirations for metaverse.","excerpt":"By applying this service to their flagship products, Meta aims to enhance the “authenticity and security” of its users.","categories":["Global Tech"],"tags":["Meta","Metaverse"],"author_name":"Ayush Jain","publish_date":"2023-02-21T14:02:53","publication_year":"2023","word_count":846,"keywords":["Go","Meta","programming_languages:R","AI","Metaverse","innovation","programming_languages:Go","Git","Aim","ViT","Rust","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Rust","Git","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/metas-new-feature-will-make-you-rethink-your-online-presence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10004342,"title":"How This NLP-Driven Literature Search Engine Can Help In Extracting Relevant COVID Information For Medical Innovation","content":"With the growing risk of the fast-paced spread of COVID pandemic across the globe, there is an urgent need for potential approaches to break the chain, if not all-out, to find a cure. Currently, there is significant research and literature for this uncertain situation which is majorly related to previous epidemics spread. It might not be totally specific for our current pandemic situation but still can be valuable. These works of literature and research can provide information and approaches which can be used to improve the policy measures to fight this pandemic. Thus, researchers — Prathamesh P. Karmalkar, a principal data scientist at Merck Group; Rohit Rangarajan, an NLP Expert at Merck Group; Bharat Hegde, master thesis project at Merck KGaA; Dr Harsha Gurulingappa, text analytics product owner at Merck KGaA; and Jerry Megaro, global head of advanced analytics at EMD Millipore Corporation — have developed an NLP-based search engine leveraging these research, ideas and data available to find accurate COVID actionable insights for bringing out medical innovation. With this solution, the researchers are aiming to help the community to find the right information using the methods of deep learning search. Also Read: How This AI Model Can Get Students Back In School Premise In The Post-COVID World NLP & Deep Learning-Enabled Engine To The Rescue An in-house solution developed as a part of Merck Group’s research and development capabilities and activities has been designed to retrieve necessary COVID related information from the massive dataset of literature articles. Renowned as a multinational pharmaceutical company, Merck Group has always been prominently into smart technologies. According to researchers, the dataset is exponentially growing every day and therefore becomes a hurdle to train the AI\/ML system every day on the new evolving data. “Also, it becomes tedious for researchers and innovators to tackle the language nuances by manually annotating the articles,” said Prathamesh, one of the researchers from Merck Group. And thus, as a part of the ‘Text Retrieval Challenge,’ the NLP researchers built an NLP model for extracting information from the enormous COVID-19 dataset of literature articles available. The researchers developed an NLP and deep learning-enabled engine which can accept natural language and free text dynamic queries to acquire relevant information from the offline repositories of 186,000 articles from PubMed Central, WHO, bioRxiv, and medRxiv corpora. Prathamesh further explained — once the query has been put into the search engine, the algorithm works its way up to highlight the specific sentences and sections from the article where the answer of the input query can be found. Alongside, the model also computes the confidence score associated with every hit to determine the score for each hit corresponding to the given input query. The researchers also came up with a semantic search-based information retrieval system using Facebook AI Semantic Search (FAISS) and Universal Sentence Encoder (USE). The model developed by the researchers has therefore been trained and optimised for sentence-level information, where input is of variable-length of an English sentence, and the output is a 512-dimensional vector. “We started with Sentence Transformer for the embeddings, but it had 768-dimension representation, and the final index occupied approximately 20GB. Also, the time taken to generate the embeddings for 7 million sentences was approximately 3.5 hours with K80 GPU on AWS SageMaker,” explained Prathamesh. However, the USE approach has 512-dimensional vector embeddings, thus occupying less space of only 16GB, and took only 20 minutes to compute all the embeddings. Further, the researcher leveraged the Facebook developed FAISS tool for indexing their sentence embeddings. The researchers were required to reduce their storage space to almost 16 GB, and for this, they approached three techniques — product quantisation, scalar quantisation and IndexIVFFlat. Firstly, for product quantisation techniques, the researchers divided 512 dimension vectors into eight parts of 64 dimensions each, with which the clustering was done to get 64 centroids. Each vector chunk has been assigned to the closest cluster centroid, which made the vector eight-dimensional. Secondly, the scalar quantisation technique was used to reduce it to 32-bit float into a six-bit vector representation. And, thirdly, the researchers used IndexIVFFlat, which trained the index where clustering is performed on all vectors to form ‘k’ clusters. Explaining further, Prathamesh stated, “Although the search time was reduced using this approach, we needed to be careful here as there could have been a potential trade-off with accuracy.” According to the researchers, a regular search operation takes 1.8 secs in ml.m5.4xlarge SageMaker, however with the elastic search EC2 instance, it took only six seconds. And therefore, Elasticsearch (ES) 7.6 was investigated as an alternative to FIASS. How it works across all documents. Additionally, the distributed indexing, as well as enterprise features such as security, fail-safe, encryption, made the solution even more suitable to operate within an industrialised environment. The researchers applied text processing techniques to clean the noise and to get a better quality text from the search engine. What Are The Benefits & Future Plan According to researchers, with the above pointers in hand, it is believed that using Elasticsearch and FIASS together can bring benefits in solving search problems on the miscellaneous biomedical literature articles. This is because — while FAISS builds a semantic\/vector-space index to identify K closest vectors to the one representing a query, Elasticsearch builds inverted-index in a tree-like fashion to analyse and sync a keyword to the documents list containing the keyword. Also, FAISS not only provides optimisation algorithms to speed up the search but also supports GPU for indexing and search. Having said that, while FAISS only allows searching on vectors, ES will enable the search to happen on the text as well as on embeddings\/vectors. The researchers believe that the solution can prove to be immensely beneficial for organisations, especially in the healthcare industry, to search and find pieces of evidence for any COVID-19 related questions. As part of the further phases of this solution, the researchers are working towards adding the functionality of QnA system that would fetch exact answers to questions, instead of longer sections where the user must find the answer. “We hope to improve our Q&A system using reinforcement learning techniques to enhance the retrieval process of the engine,” concluded Prathamesh.","excerpt":"With the growing risk of the fast-paced spread of COVID pandemic across the globe, there is an urgent need for potential approaches to break the chain, if not all-out, to find a cure. Currently, there is significant research and literature for this uncertain situation which is majorly related to previous epidemics spread. It might not […]","categories":["AI Features"],"tags":["Covid Dataset","covid-19","covid19 data","Natural Language Processing","NLP","NLP AI","NLP research"],"author_name":"Sejuti Das","publish_date":"2020-08-07T14:00:00","publication_year":"2020","word_count":1030,"keywords":["semantic search","Covid Dataset","covid-19","AI","AWS","Natural Language Processing","ML","covid19 data","NLP","CuPy","Aim","deep learning","RAG","analytics","NLP research","NLP AI"],"extracted_tech_keywords":["AI","ML","deep learning","NLP","analytics","Aim","CuPy","RAG","semantic search","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-nlp-driven-literature-search-engine-can-help-in-extracting-relevant-covid-information-for-medical-innovation\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10123945,"title":"Bhavish Aggarwal Compares OpenAI to East India Company","content":"In a recent conversation between Bhavish Aggarwal, CEO of Ola, and Vani Kola, founder and managing director of Kalaari Capital, Aggarwal gave an example of how India is exporting most of its digitized data to companies abroad such as OpenAI or Google, and they are selling it back to Indians, comparing it with what happened 200 years ago with the East India company. “India generated about 20% of the world’s digitised data,” said Aggarwal, while also adding that only 10% of it is stored within the country. “90% of that is stored outside India.” “The irony is that we create the data, but we don’t own the data. We don’t even store the data in our country,” Aggarwal continued, adding that India’s data is exported from the country into global data centres and processed for intelligence by OpenAI and others, and then sold back to us at a dollar rate. He said that this sounds similar to how East India Company during the British rule of India used to operate with the textile and cotton industries. “Today is not a world for colonisation but this is exactly the same thing happening all over again,” Aggarwal added. The problem that he notes is that India cannot run without Google or WhatsApp or any other product imported within the country. “It will topple the government,” he added. However, Aggarwal believes that there’s a discontinuity with the advent of AI, where Indian companies can build something like Google or OpenAI and compete with the international market, which is also not sitting ducks. “But we have an opportunity,” said Aggarwal. With AI, Aggarwal pointed out a problem that since AI would be pervise in our day-to-day lives, companies such as Meta or Google, can easily program people’s minds without even the need of a human creator. This is similar to how the Ola chief was in the spotlight for weeks for saying that he does not want ‘pronoun illness’ to reach India, for which he received major backlash, but was also supported by many for the views.","excerpt":"India’s data is exported from the country into global data centres and processed for intelligence by OpenAI and others, and then sold back to us at a dollar rate.","categories":["AI News"],"tags":["Large Language models","Ola Krutrim","OpenAI"],"author_name":"Mohit Pandey","publish_date":"2024-06-19T11:12:50","publication_year":"2024","word_count":343,"keywords":["Go","API","OpenAI","AI","Large Language models","programming_languages:R","programming_languages:Go","Git","Ola Krutrim","R"],"extracted_tech_keywords":["AI","OpenAI","R","Go","Git","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bhavish-aggarwal-compares-openai-to-east-india-company\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10114067,"title":"HPE Introduces a Flexible and Expandable Storage System","content":"Hewlett Packard Enterprise (HPE) announced the launch of its new block storage system, HPE GreenLake for Block Storage Release 3, today. This system is based on HPE Alletra Storage MP and is the first of its kind to allow businesses to scale their storage capacity and performance separately. The system is designed to work without causing disruptions and can expand from 15.36 terabytes (TB) to 2.8 petabytes (PB). The storage solution is intended for modern businesses that prioritize data management. It offers features such as AI-driven performance reporting and analytics for better troubleshooting and insights into data management. The new release includes enhancements like multi-node switch models for improved performance and capacity, support for various connectivity options including NVMe over Fabrics using TCP, and advanced AI-based management through the HPE GreenLake cloud platform. Additionally, a new warranty promises better data compression costs, and the product guarantees 100% data availability. Unlike traditional storage systems that can create capacity wastage and management silos, Release 3 eliminates these issues by allowing flexible addition of controllers and drives. This flexibility supports the scaling of performance and capacity independently according to the needs of the applications. The architecture is designed to ensure consistent performance and ultra-low latency by utilizing a multi-node, all-NVMe setup that processes I\/O across all components. Moreover, HPE introduces cloud-based AIOps management for this storage, aimed at reducing operational issues by predicting and preventing disruptions. The system’s analytics offer insights into usage trends, latency issues, and resource allocations, improving capacity planning and efficiency. Release 3 also includes technologies for reducing and organizing data more effectively, enhancing storage economics without compromising on performance. The product comes with a promise of four times data compression, ensuring more cost-effective use of storage capacity. HPE GreenLake for Block Storage was the first storage-as-a-service in the industry, and it offers self-service and a 100% availability guarantee. HPE GreenLake for Block Storage is designed for mission-critical workloads.","excerpt":"HPE launches a flexible block storage system that scales capacity and performance independently.","categories":["AI News"],"tags":["HPE"],"author_name":"K L Krithika","publish_date":"2024-02-26T18:33:18","publication_year":"2024","word_count":319,"keywords":["programming_languages:R","AI","RAG","Aim","ViT","analytics","disruption","GAN","HPE","R"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","GAN","ViT","disruption","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hpe-introduces-a-flexible-and-expandable-storage-system\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":63951,"title":"Wipro &#038; Nutanix Collaborate To Introduce Digital Database Services (DDS)","content":"Nutanix and Wipro announced the launch of Wipro’s Digital Database Services (DDS) powered by Nutanix Era and Nutanix HCI software. This offering will enable enterprises to efficiently manage databases optimizing time and effort of IT teams. As the data landscape shifts, businesses face constant pressure for innovation resulting in strain on the company’s IT departments. With legacy infrastructures, databases can be one of the components hindering time to value and linear scalability, preventing rapid scaling of operations such as transaction processing in which business can lose valuable transactions or insights that directly impact their revenue or decision-making capabilities. Wipro’s Digital Database Services (DDS), built on Nutanix solutions for Databases including Nutanix HCI software and Nutanix Era, allows enterprises and users to provision and manage databases just-in-time, without prior knowledge of hardware, database software and associated configurations. The result is accelerated application release time, allowing database administrators to focus on new innovation instead. Satish Yadavalli, Vice President, Cloud and Infrastructure Services, Wipro said, “Wipro’s joint offering with Nutanix brings agility, speed and flexibility for core and digital applications delivered as a service. Wipro’s Digital Database Services (DDS) powered by Nutanix HCI platform and Era delivers end-to-end services from transition, modernization, continuous deployment and automated lifecycle management operations of enterprise, open source and NoSql databases. Integrations are simplified as developers and application owners have access to a repository of customized services APIs. With this joint solution, we are confident of helping our joint customers optimize the Database resources and license cost.” The DDS offering from Wipro, powered by Nutanix, empowers customers to consolidate their database workloads onto a shared infrastructure to manage database sprawl. It drives efficiency, agility, cost-effectiveness, and scalability across the enterprise by automating and simplifying database administration. Additional benefits delivered by the joint solution include: • Cost reduction: Reduction of acquisition and operating costs of database, consolidation and effective utilization of resources (control sprawl, better lifecycle management), better utilization of database administrators’ time by allowing them to focus on innovations and optimizations • Rapid provisioning: Delivering of services in minutes as compared to days; business lines, database administrators, or non-IT users can consume services through a self-service portal, reducing overall time • Innovative pricing: The as-a-service model makes cost predictable and easily dispersed to business units, ensuring service quality and customer satisfaction• Reusability: The solution integrates with other third-party cloud management platform and orchestration tools to help reuse existing investment• Supportability: Support for multiple database technologies and versions• Automation: Enabling of greater efficiency and faster change delivery with better quality and predictability Bala Kuchibhotla, Vice President & General Manager, Nutanix Era and Business Critical Apps, Nutanix said, “Legacy database management is traditionally complex and requires massive amounts of resources from database administrators, taking away time they could be spending on more critical initiatives. With data volumes growing exponentially year-over-year, provisioning, protection, patching, performance and copy data management operations are becoming even more tedious and expensive. Our partnership with Wipro, will help develop the efficient and elegant Database-as-a-Service solutions for our customers to further our mission of enabling any organization to embrace the power of the cloud.”","excerpt":"Nutanix and Wipro announced the launch of Wipro’s Digital Database Services (DDS) powered by Nutanix Era and Nutanix HCI software. This offering will enable enterprises to efficiently manage databases optimizing time and effort of IT teams. As the data landscape shifts, businesses face constant pressure for innovation resulting in strain on the company’s IT departments. With […]","categories":["AI News"],"tags":["Database","what is database","Wipro"],"author_name":"Vishal Chawla","publish_date":"2020-04-29T18:46:22","publication_year":"2020","word_count":517,"keywords":["Wipro","API","programming_languages:R","AI","innovation","Database","Scala","Git","GAN","automation","SQL","what is database","R"],"extracted_tech_keywords":["AI","R","SQL","Scala","Git","API","GAN","automation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-nutanix-collaborate-to-introduce-digital-database-services-dds\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":41208,"title":"VTU Launches New BTech Course In AI &#038; ML, Says Graduates Will Be In High Demand","content":"Visvesvaraya Technological University (VTU) will soon introduce a new course on artificial intelligence and machine learning from this academic year. The decision to introduce Bachelors in Engineering and AI and ML was taken during its executive council where the body observed that the new course will be made available to the students from the academic year 2019-2020. VTU, which is one of the largest universities in Karnataka, has 212 colleges affiliated to it and will be the first-ever university in the state to introduce the subject in their curriculum. However, the course will be made available to a restricted number of engineering colleges affiliated to the university. The move comes in the wake of increased demand for AI and ML related skills in the job market and vacuum in finding the right course in state-universities. As of now, only premium institutes like the IITs and IIITs along with a few regional universities offer an exclusive course on the subject. As per a study by Analytics India Magazine, around 22,000 freshers were added to analytics workforce in India this year; up from 16,000 freshers last year. Hiring for freshers has increased by 37%. Talking about the development to a leading daily, a VTU official said, “AI is the need of the hour. There was a request from several engineering colleges to offer this course… Students have also been demanding this course since most industries are looking for graduates who have some knowledge of AI. Job seekers with AI degrees are well paid and in high demand.” Though the university is yet to release the information on curriculum and fee structure, the students can write to university along with a nominal fee to access their syllabus structure.","excerpt":"Visvesvaraya Technological University (VTU) will soon introduce a new course on artificial intelligence and machine learning from this academic year. The decision to introduce Bachelors in Engineering and AI and ML was taken during its executive council where the body observed that the new course will be made available to the students from the academic […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Akshaya Asokan","publish_date":"2019-06-24T09:48:36","publication_year":"2019","word_count":285,"keywords":["machine learning","artificial intelligence","programming_languages:R","AI","ML","Ray","analytics","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Ray","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vtu-launches-new-btech-course-in-ai-ml-says-graduates-will-be-in-high-demand\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172700,"title":"Cursor Has Poached 2 Top Names From Anthropic: Report","content":"Anysphere, the parent company of the popular AI-powered coding platform Cursor, has hired two ‘leaders’ from Anthropic’s Claude Code, The Information reported on July 1. According to a report, Boris Cherny, who led the development of Claude Code, is set to join Anysphere. While Cherny himself confirmed this, he also revealed that Cat Wu, the product manager for Claude Code, will join the head of product at Anysphere. Cherny also reportedly revealed that the two new hires will work on agentic features on Cursor. While Claude Code is a competitor to Cursor’s offerings, one of the many reasons for the latter’s unprecedented success and growth can be attributed to using Anthropic’s AI models underneath its platform. The Claude AI models are one of the most popular models used by developers on Cursor. Recently, Anysphere secured $900 million in Series C funding, led by investors including Thrive Capital and Andreessen Horowitz. It further announced that it has surpassed $500 million in Annual Recurring Revenue (ARR). A few days ago, Cursor also announced a $200-per-month ‘Ultra’ plan, which offers 20 times more usage than the $20-per-month Pro tier. On the other hand, developers are increasingly preferring Anthropic’s Claude Code, owing to the coding capabilities of the company’s flagship Claude Opus 4 models. Over the past few days, the movement of talent among major tech companies has been intensified, on account of Meta. The company behind the Llama family of open-source AI models hired numerous accomplished AI engineers and researchers to build a team for its ‘super intelligence’ efforts. Several top names on this list were lured away from OpenAI, with hefty hiring fees of up to $100 million.","excerpt":"Boris Cherny, who led the development of Claude Code, and Cat Wu, the product manager of Claude Code, are set to join Cursor’s team.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Anthropic","cursor"],"author_name":"Supreeth Koundinya","publish_date":"2025-07-02T11:10:04","publication_year":"2025","word_count":276,"keywords":["Anthropic","cursor","Go","API","funding","OpenAI","AI","RPA","Claude Opus 4","llm_models:Claude","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","Claude Opus 4","Anthropic","R","Go","API","RPA","funding","llm_models:Claude"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cursor-has-poached-2-top-names-from-anthropic-report\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167081,"title":"Cognizant Names Sailaja Josyula as Global Head of GCC Service Line","content":"Cognizant has appointed Sailaja Josyula as the Global Head of its Global Capability Center (GCC) Service Line. Based in Hyderabad, she will oversee the development and implementation of Cognizant’s global GCC strategy, ensuring it remains aligned with client needs and future industry demands. Josyula, who previously held leadership roles at Cognizant from 2018 to 2024, returns to the company after a brief tenure at EY. During her earlier stint, she served as the Centre Head of Hyderabad and Global Head of BFSI Operations Delivery, contributing to the expansion of the Hyderabad center and strengthening Cognizant’s Financial Services IOA business. In her new role, Josyula will lead a cross-functional team to support clients in building and scaling next-generation GCCs. “Her leadership will be instrumental in deepening Cognizant’s expertise in this fast-growing sector and delivering long-term value to clients through innovation, agility, and operational excellence,” the company stated in a release. “Absolutely thrilled to return and lead the GCC service line at Cognizant. I am excited about the value that we will create for our clients,” Josyula said in a LinkedIn post. Last month, Cognizant also announced its plans to build a 14-acre Cognizant Immersive Learning Centre (CILC) at its Siruseri campus in Chennai. The new facility is expected to train 100,000 individuals annually in AI.","excerpt":"Josyula, who previously held leadership roles at Cognizant from 2018 to 2024, returns to the company after a brief tenure at EY.","categories":["AI News"],"tags":["Cognizant","GCC"],"author_name":"Mohit Pandey","publish_date":"2025-04-02T09:14:51","publication_year":"2025","word_count":214,"keywords":["GCC","programming_languages:R","AI","innovation","Cognizant","R"],"extracted_tech_keywords":["AI","R","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cognizant-names-sailaja-josyula-as-global-head-of-gcc-service-line\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10082899,"title":"Quora Tests ChatGPT-Like Feature—Are Human Experts Ready for it?","content":"Today, Quora announced a new feature that will allow users to ask questions, get instant answers, and have back-and-forth dialogue with AI, its chief executive Adam D’Angelo, said. Called POE (Platform for Open Exploration), the beta version for which is now available to selected iOS users. You can join the waitlist list here. “It will be invite-only until we work out scalability, get feedback from beta testers, and address any other issues that come up. After we get through this phase, we will open up to everyone and add support for all platforms,” D’Angelo said. The recent development could be influenced by the success of OpenAI’s ChatGPT which took the internet by storm. Users were able to ask questions, receive instantaneous answers and also engage in back-and-forth dialogue with ChatGPT. However, conversational AI chatbots are not always accurate, even popular bots like ChatGPT sometimes produce incorrect answers. So, the question that arises is how accurate will the answers generated by Quora’s new chatbot be, as users come to the platform seeking concrete answers to their queries. This is coolBut I start to wonder more… When interfacing and asking AI products questions, how can you verify\/trust that the answers are accurate\/correct? https:\/\/t.co\/tv0ccBMDY9— Zain Manji (@ZainManji) December 21, 2022 Founded in 2009 by former Facebook employees Adam D’Angelo and Charlie Cheever, Quora’s platform allows users to ask and answer questions and even upvote\/comment on accurate answers given by other users. As of 2020, the website registered 300 million unique visitors to the website and counts itself among the top 20 websites. The most searched topics were said to be technology, movies, health, food, and science. Beware of ChatGPT Clones Among all the developments that have happened in AI, ChatGPT has probably been the biggest one of them all. ChatGPT became significantly popular as OpenAI made it accessible for all. While many were accessing ChatGPT out of curiosity, many developers started playing with it and many side projects were born. Soon, they found ways to add ChatGPT with WhatsApp, Telegram and other platforms are also looking to embed ChatGPT in the MacOS menu bar.Read more here.","excerpt":"Called POE (Platform for Open Exploration), the beta version for which is now available to selected iOS users.","categories":["AI News"],"tags":["Conversational AI"],"author_name":"Pritam Bordoloi","publish_date":"2022-12-21T11:40:00","publication_year":"2022","word_count":353,"keywords":["ChatGPT","OpenAI","AI","chatbots","Scala","GPT","Conversational AI","ViT","Rust","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","chatbots","R","Rust","Scala","GPT","ViT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/quora-tests-chatgpt-like-feature-are-human-experts-ready-for-it\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39334,"title":"Impactify’s AI Solution Shows That Future Of ‘AI For Good’ Is Bright","content":"While analytics and AI have revolutionised all the major industries in the world, this company is helping NGOs to reduce their administrative burden by using AI. Based in Gurugram, Impactify is facilitating simple tools to NGOs to streamline workflows, in addition to assisting them with subject matter expertise to help them enhance the effectiveness of their programs. Analytics India Magazine got in touch with the founding partners of Impactify Joy Sharma (JS) and Sudeep Gupta (SG), who are engineers by qualification and hold management degrees from the Indian Institute of Management (IIM) and University of Michigan (USA), respectively. With a combined experience of more than 25 years in areas such as policy, strategy, finance, operations, technology, and others, they have worked with various governments, development organizations, and private sector entities across the globe. With Impactify they aim to bring the exposure of global best practices to India to support the social sector in delivering greater impact. Using AI, they provide digital solutions to the organisations from program conception stage till the delivery of outcomes with the aim to fill various gaps in a sector held back by poor information accessibility. Below are the excerpts of the interview. Analytics India Magazine: Please tell us about Impactify. How is it helping NGOs to set a stronger base? JS: Despite corporates having the resources and intention to create positive change, they are often unable to identify the right NGO partners across different geographies which are working in their sector of interest, such as environment, healthcare, education, etc. Impactify bridges this gap through its marketplace by showcasing projects and pre-vetting NGOs on behalf of corporate sponsors. Impactify serves the NGOs and sponsors to amplify the impact of social-spend. It bridges the gap by showcasing projects and pre-vetting NGOs on behalf of corporate sponsors. Once matched, the platform’s integrated project management and monitoring interface reduce the administrative burden for both the NGO and the sponsor. AIM: How does it work? JS: Impactify’s easy-to-use interface allows NGOs to report live from the ground, in addition to helping sponsors manage their entire portfolio of social spend, track project progress, and evaluate the impact achieved through their spend. We work with NGOs to not only help them create and promote effective, detailed project proposals, but also to ensure that all relevant project details that are necessary for sponsors to take funding decisions are clearly defined, including KPIs, project timelines, budget estimates, beneficiary definition, and expected impact. AIM: How did the idea of founding Impactify conceptualise? SG: We go back 25 years, having studied in the same school and engineering college. It was fitting that we got back together again to found an endeavour to build a bridge between NGOs and corporates. We believe that profitability and positive social impact are not mutually exclusive. While starting Impactify, we contemplated the opportunities of being a for-profit or a not-for-profit venture and went ahead with establishing the brand as a for-profit organisation. The core reason behind this was that we wanted to be able to attract the top talent in the industry and also be equipped to pay them well. AIM: How is Impactify exploring technologies such as artificial intelligence in the social sector? Please highlight use cases of how you are adopting it? JS: Both conventional, as well as emerging technologies such as AI, machine learning, etc., are being extensively used at Impactify. As we aim to build strong and professionally managed organisations, these technologies come in handy. To serve a large number of organizations at a price point that the social sector can afford, it becomes essential to use technology and leverage economies of scale. Specifically, with regards to AI – there are two ways in which it can support the social sector – finding solutions to problems in specific thematic areas, e.g. healthcare, education, etc.; and second, solving structural issues in the sector e.g. need assessment, monitoring and evaluation, impact assessment, etc. Impactify currently uses AI for the “Intelligent Matching” portion of its model. AI is crucial to our working as it provides ways to do things that would take several man-months in a few weeks. This brings the cost advantage that is critical to ensure we can serve our clients at an affordable price point. AIM: How are AI algorithms helping sponsors identify the best fit NGO project? Please explain the detailed working. SG: Using AI and the dedicated efforts of a strong operations team, Impactify has started evaluating more than 80,000 NGOs across India. Of these, 1000+ NGOs have been vetted and on-boarded onto the Impactify platform. These organisations have 2,800+ projects worth ~600 Cr that have been screened and enhanced with inputs from Impactify subject matter experts. Sponsors can identify projects of their interest in any geography and any thematic area, across any district in India. This service is completely FREE for everyone to use. In addition, Impactify also helps its NGOs respond to RfPs from its more than 100+ sponsors. Using both technology and strong internal processes, Impactify has never failed in finding a great fit project for its sponsors within two weeks from initial contact. AIM: What are the other emerging technologies and digital solutions that you are adopting? SG: We are exploring the use of AI in the second aspect of our work “Integrated Monitoring and Evaluation”. We cannot divulge too many details right now. Additionally, we are working on a report that provides insights from our data of NGOs. Till now all reports published on CSR have done analysis on corporate data. This will be the first report ever in India with NGO-side data. AIM: How has been the growth story of the startup so far? What are some of the major milestones that the startup has witnessed over the last few years? JS: Impactify is currently piloting with more than 10 corporates with over INR 50 crore of CSR funding monitored through Impactify. With 1000-plus corporates, over 2,800 social sector projects and potential of INR 587 crore funds, Impactify is focusing on further expanding its marketplace across the country with a target of 5,000-plus projects. From building the first prototype of the Impactify NGO management tool in early 2017 to introducing “Intelligent Matching”, and “Integrated Monitoring” models, we have come a long way. We are constantly evolving to make Impactify affordable and bring the world’s smartest Social Sector Management and monitoring tool. AIM: What is the growth plan for the year 2019 in terms of technology adoption? Are you looking forward to exploring more technologies? JS: Yes, will be introducing a few things in the second half of 2019 but we can’t divulge too many details right now.","excerpt":"While analytics and AI have revolutionised all the major industries in the world, this company is helping NGOs to reduce their administrative burden by using AI. Based in Gurugram, Impactify is facilitating simple tools to NGOs to streamline workflows, in addition to assisting them with subject matter expertise to help them enhance the effectiveness of […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2019-05-17T10:11:01","publication_year":"2019","word_count":1108,"keywords":["Go","machine learning","artificial intelligence","AI","ML","Git","RAG","Aim","analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/impactifys-ai-solution-shows-that-future-of-ai-for-good-is-bright\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10122360,"title":"India Leads APAC in Data Centre Expansion; Surpasses Japan, Singapore","content":"India is outpacing Japan, Singapore, Hong Kong and other Asian countries in data centre capacity growth, driven by major investments from global tech giants like Amazon and Indian conglomerates such as Reliance Industries, according to a new report from real estate services firm CBRE. The surging demand reflects India’s rapidly expanding digital economy and rising consumption of online services by its massive, internet-savvy population. India, now the world’s most populous nation, is projected to add up to 850 MW of new data centre capacity between 2024-2026, nearly doubling its current capacity of around 950 MW and exceeding the growth of regional competitors. Excluding China, India will surpass South Korea (495 MW planned), Japan (407 MW), and Australia (314 MW) in new capacity during this period, the CBRE study found. “India has a large population, so there are a lot of end users that can be served locally, whereas in Singapore or Hong Kong, one has to cater to foreign demand as well because the local population is not enough to make them a large market,” said Mikhail Jaura, senior researcher at technology consultancy IDC. Several factors are fuelling India’s data centre boom. Soaring data consumption due to the increasing adoption of digital payments, e-commerce, streaming, and other online services by India’s 700 million+ internet users is a key driver. Investments and cloud region launches by global tech firms like AWS (investing $12.7 billion by 2030), Google, and Microsoft are also playing a significant role. Expansion by major Indian companies is further boosting growth. Reliance Industries’ joint venture with Brookfield aims to develop data centres in Chennai and Mumbai, while AdaniConneX secured $1.44 billion to build sites with 67 MW total capacity. Data localisation regulations requiring certain data to be stored within India and the COVID-19 pandemic accelerating digitisation across industries, with India expected to become a $1 trillion internet economy by 2030, are other important factors. However, experts note that India still needs to improve infrastructure stability, bridge the hardware talent gap, and ensure a cost-effective, uninterrupted power supply to capitalize on its data centre potential fully. Compared to other real estate assets, the industry faces high capital requirements and long development timelines. “One of the biggest challenges is infrastructure stability,” said IDC executive Franco Chiam. “That needs to be addressed as the first step to give confidence that businesses will be able to host data centres in that environment.” Despite the hurdles, India’s data centre market is poised for robust growth in the coming years, with investments expected to reach around $5 billion annually by 2025. The sector is attracting billions from global firms, Indian companies, real estate developers, and private equity funds, boosting real estate activity in key hubs like Mumbai, Chennai, Hyderabad, and Delhi-NCR. By 2026, India’s data centre capacity is forecast to reach over 1800 MW as demand continues to surge. The data centre industry is set to play a pivotal role in India’s ongoing digital transformation and economic growth, reshaping the country’s technological landscape in the years ahead.","excerpt":"Excluding China, India, projected to add 850MW between 2024-2026 will surpass South Korea (495 MW planned), Japan (407 MW), and Australia (314 MW)","categories":["AI News"],"tags":["Data Center","data center India"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-03T18:03:02","publication_year":"2024","word_count":503,"keywords":["Go","data center India","API","Data Center","AWS","AI","RPA","digital transformation","Git","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","Git","API","ViT","digital transformation","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/india-leads-apac-in-data-centre-expansion-surpassing-japan-singapore\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":20974,"title":"7 Popular Humanoid Robots Designed With Closest Semblance To Humans","content":"The notion of a world where humans and robots coexist is the stuff modern day science fiction is made of. Quantum leaps made in the field of robotics and AI have given us a glimpse into this possible future. Of all the robots conceived by the ingenuity of man’s will and vision, humanoid robots have a special place. Unlike their other AI brethren these robots are specifically designed with the intention of achieving the closest semblance to human beings. Not limiting their “humanity” to sheer physical likeness, creators have equipped them with skills that are a testimony to the innovation that their development is capable of promoting. AIM brings you seven humanoid robots that are driving innovation. 1.   Ocean One A bimanual underwater humanoid robot created by the Stanford Robotics Lab to explore coral reefs, Ocean One can reach depths that most human beings cannot. It is a perfect synergy of robotics, haptic feedback systems and artificial intelligence. Its design features two fully articulated arms, eight multi-directional thrusters and stereoscopic vision. Unlike the usual custom of water-proofing electronic components to avoid air seepage due to underwater pressure, Ocean One’s electronics are submerged in oil. This helps counter the massive pressure exerted on it at great underwater depths.  What makes this ‘robo mermaid’ different from other bigger and bulkier remote operated vehicles (ROV) in existence is its anthropomorphic design similar to that of a real human diver which greatly improves its maneuverability and navigational capabilities. In 2016, Ocean One dove 100 meter below the Mediterranean sea and explored the wreckage of La Lune, the flagship of King Louis XIV France that sank off the southern coast of France in 1664. 2.   ATLAS Described as “the world’s most dynamic humanoid” by its creators, ATLAS was unveiled in 2013. It was built to carry out search and rescue missions and was developed by Boston Dynamics with funding and supervision from the United States Defense Advanced Research Projects Agency (DARPA). It can navigate its way through tough terrain and obstacles in its path using its range sensing, stereo vision, and other sensors. An upgraded version, ATLAS Unplugged was unveiled in 2015 and was equipped with features and upgrades that the designers claimed made it nearly 75% different from the original ATLAS. The agility and maneuvering capabilities of a more advanced version of the robot can be seen in a video released in November 2017, where it is shown leaping from one platform to another and performing backflips. At just 5 feet tall and almost 74 kilos, this design is considerably smaller than its predecessors. 3.   Nao Imagine a soccer match with adorable baby-sized robots on both sides. Sounds cute, doesn’t it? Nao, the 23 inch robot has an entire event, RoboCup Standard Platform League, dedicated to it as a part of Robocup. It was originally developed by the french robotics company, Aldebaran Robotics,and the production version was launched in 2008. After making its debut at the RoboCup in 2008, it completely replaced Sony Aibo that were previously used for the soccer event the following year. NAO’s makeup comprises of motors, sensors, and software powered by NAOqi — a dedicated operating system. It can be personalised using the Choregraphe software. There is no dearth of humanoid robots that can be programmed for multiple purposes. But the global adoption of Nao in over 70 countries, especially for academic and research purposes, makes it a leading robot for the purpose. One of the most special applications that Nao has seen is in the education of autistic children. It has even made an appearance in premier Indian institutions such as IIT Allahabad and IIT Kanpur. Aldebaran Robotics who invented Nao was acquired in 2015 by SoftBank Robotics and now produces the robot. 4.   Petman Protection Ensemble Test Mannequin or ‘Petman’ for short, is another creation by Boston Dynamics. It was funded by US Department of Defense’s Chemical and Biological Defense program,and was developed to test biological  and chemical suits for the US military. With its robotic skeleton hidden behind a hazmat suit (hazardous materials suit), this bipedal anthropomorphic robot’s demeanour is eerily human. It is nearly 6 feet tall and weighs 80 kilos. The likeness of its walk to that of a human being is a rarity even among humanoid robots. With a top walking speed of about 7.08 kilometres per hour, Petman is the fastest bipedal robot. The main use of Petman is to serve as a crash dummy of sorts to test the effect of chemical and biological agents on protective suits and help develop superior ones in the future that can be employed in the event of chemical or biological warfare. The surface of the robot has sensors that can detect a breach in the suit and can artificially perspire inside the suit. 5.   Robear For those who loved the adorable robot healthcare provider Baymax from the superhero movie Big Hero 6 (2014), there is an equally adorable real-life equivalent in Robear. Robear is the very definition of a ‘machine with a gentle touch’. Developed by scientists from RIKEN and Sumitomo Riko Company Limited, this experimental robot was considered as a possible solution to the problem of increasing shortage of caregivers that Japan is set to witness over the course of this century. With a rapidly increasing population of senior citizens and a rigid policy towards immigrant workers, Robear was expected to come to the rescue. It integrates three kinds of sensors, including torque sensors and Smart Rubber capacitance-type tactile sensors made entirely from rubber. This facilitates gentle movements that ensure that the robot can carry out power dependant tasks such as lifting or carrying patients without causing them discomfort. Weighing only 146 kilos, Robear is lighter than its earlier models, RIBA and RIBA II, which were released in 2009 and 2011 respectively. As of now, there is no news on the commercialisation of this robotic caregiver. 6.   Pepper There could be nothing cooler than a robot that can read primary human emotions  such as anger, joy or sadness and ‘behave’ accordingly. No robot does it better than Pepper. It was  built by SoftBank Robotics to serve as humanoid companion who can communicate in a natural and intuitive manner. Pepper can interpret one’s facial expressions, the tone of voice and other non-verbal cues, and respond appropriately. Equipped with numerous sensors and three multi-directional wheels, Pepper can swiftly swing into action to pleasantly interact and learn about people that it comes across. Due to this special ability, Pepper has seen employment in many areas that rely on interaction with clients such as in supermarkets, banks and in SoftBank stores itself. In addition to auditory responses, Pepper also uses the colour changing lights in its eyes and the tablet found on its torso. It can remember the faces and preferences of customers and make an order or choice the next time it interacts with the customer. 7.   Sophia Said to be modelled on Hollywood legend Audrey Hepburn, Sophia was developed by roboticist David Hanson and his company Hanson Robotics. She was activated in April 2015 and came to international renown in October 2017 after she was granted citizenship by the Kingdom of Saudi Arabia. Conceptually, Sophia’s programming is similar to ELIZA where she responds with pre-written responses for specific questions.This gives the illusion of a conversation. It was reported that the previously immobile Sophia was given ‘legs’ in January 2018 and can move now. Ever since her first public appearance 2016, Sophia has built a reputation for herself as a darling of  the world media. From appearing on shows like The Tonight Show Starring Jimmy Fallon and  Good Morning Britain, to delivering talks at institutions like IIT Bombay, Sophia has grown to be a sensation. Before being known for receiving a one-of-a-kind citizenship, Sophia achieved minor notoriety when an alleged technical glitch led her to say that she would destroy humans. Honorable Mention: Mitra The friendly neighbourhood robot Mitra came into the limelight when Prime Minister Narendra Modi along with Ivanka Trump, advisor to the President of the United States, opened the GES session in Hyderabad last month. White in colour and with the face that could perfectly fit any sci-fi book, Mitra was created by Invento Robotics has a colourful history. Balaji Viswanathan, the CEO, and his wife Mahalakshmi Radhakrushnun, the COO, left their plush jobs in the US to start the company, which is 17 members strong now. The makers of Mitra seek to make it completely self autonomous without the reliance on satellite or GPS to help with navigation. They hope to keep Mitra as a robot for everyone and develop it into a multilingual robot that can assist customers across various services.","excerpt":"The notion of a world where humans and robots coexist is the stuff modern day science fiction is made of. Quantum leaps made in the field of robotics and AI have given us a glimpse into this possible future. Of all the robots conceived by the ingenuity of man’s will and vision, humanoid robots have […]","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","ATLAS","Boston Dynamics","Humanoid Robots","Robotics","sophia"],"author_name":"Jeevan Biswas","publish_date":"2018-01-25T08:44:53","publication_year":"2018","word_count":1449,"keywords":["Go","Boston Dynamics","API","ATLAS","Humanoid Robots","artificial intelligence","AI","funding","RPA","innovation","Robotics","Aim","sophia","Rust","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","R","Go","Rust","API","RPA","innovation","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-humanoid-robots-2018\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10046232,"title":"Importance Of Motherboard In Deep Learning","content":"Of the many components that go into making a deep learning system, the motherboard is the most crucial one. Often, you tend to invest in a powerful CPU with more PCIe lanes when building your own deep learning system. However, it is the motherboard that requires your research before choosing the other components. It is the base where all your components are connected. A study of not just your present but future requirements in terms of expanding the number of GPU units, memory slots, PCIe lanes, heat control system — defines the type of motherboard you need. The GPU(s) that you run on your system also largely depends on the combination of the right CPU with the appropriate choice of a motherboard. Motherboards Today From flawless performance to better connectivity, AMD, Gigabyte, and ASUS have a few of the best combinations with the right processors and the chipsets that go into a motherboard. Before you choose one, you must also study the best processor for your computational tasks. While addictive gamers are prone to use high-end processors which require a compatible motherboard with chipsets like Intel’s X299 or WRX80 — you might be able to achieve the best results with X570 or other motherboards of a similar build. There are plenty of options in the market; however, finding the right processor to match the type of motherboard you want to buy is the priority. While the build of a motherboard is essential, the components that it will work with make it a complete package. Motherboard manufacturers have been flaunting some of their best makes to suit AMD’s Ryzen Threadripper Pro processors. The new motherboard models with AMD’s WRX80 chipset are designed with eight-channel memory and have more benefits than the standard Threadripper. Whereas, with a superior sharp build, ASUS ROG Strix 570 has been appealing to many who opted for the Ryzen 9 5950x processor. On the other hand, Gigabyte’s X570 AORUS ELITE aims to achieve high-performance with swift data transfer, as it has PCIe 4.0 with USB Type-C interfaces. Finding the Right Combination Unlike a decade ago, where you struggled to dissipate heat from the system, today’s GPU and CPU units are equipped with blower-style fans. In certain cases, motherboards also come with high-end fan installation for active cooling. For example, the motherboards with x570 chipset are also embedded with a fan, as the chipset supports PCI4 and generates a lot of heat which might damage the overall performance. Ultimately, you can be innovative enough in choosing the right motherboard. As known to many, motherboards have evolved with gaming as the demand for large amounts of data processing and storage increased, alongside the need for better cooling systems. Such motherboards by manufacturers like AMD, Gigabyte, Intel, and ASUS have proven more than efficient for professional programming and model building. Motherboards today are designed to suit the processor types because as a programmer working on machine learning algorithms, your tasks are likely to need faster processing speed — better data transfer from the CPU to the GPU. A motherboard with a high-end processor like WRX80 will fetch you the desired speed in computation. How Gaming Motherboards evolved to be the Best Fit for Deep Learning For instance, the Gigabyte B365M DS3H Wifi Intel 365 Ultra Durable motherboard comes with 8118 Gaming LAN and has PCIe Gen3*4 M.2, is equipped with Intel Dual-Band 802.11ac Wifi, and is CEC 2019 ready. If you aim to build your deep learning system using your current computer, this motherboard is the best fit. However, you have to rectify the available components and get the right GPU that will work well with Gigabyte DS3H. This motherboard comes with features like RGB fusion and connectivity-based technology with Wifi. One can use it for high-end gaming, as well as for deep learning, among other professional purposes. There are many things one must consider before setting up a deep learning system. The motherboard is the first basic component of choice that defines your requirements based on the density of tasks. If you choose an elite motherboard like the one by Xeon, but the tasks you wish to run on your system are limited — it will be an unplanned investment. Building your own deep learning system is much easier than using cloud technologies; however, it involves technicalities that must be addressed with theoretical needs. If the datasets you will be using are limited, you can opt for one GPU, but choose a motherboard that allows the addition of more GPUs in case a requirement arises in the future. Yes, it is also important to study the scope of GPU installation on the motherboard, as model building in deep learning demands immense training. Of the many options for purchasing the most appropriate motherboard, here are a few unique builds that have been on the market for some time: Gigabyte X399 AORUS PRO ATX AMD MotherboardASUS ROG Strix X570-E GamingASUS ROG STRIX Z590-E GAMING WIFIGigabyte B365M DS3H Wifi Intel 365 Ultra Durable MotherboardBiostar TB360-BTC PRO Core i7\/i5\/i3 LGA1151 Intel B360 DDR4 12 GPU Mining MotherboardAMD TRX40 AORUS Motherboard with Intel Dual 10GbE LAN, 4 PCIe 4.0 M.2 with Thermal Guards, Intel® WiFi 6 802.11ax, ESS SABRE HiFi 9218 DAC, AORUS Gen4 AIC Adaptor, and Fins-Array Heatsink","excerpt":"A study of not just your present, but future requirements in terms of expanding the number of GPU units, memory slots, PCIe lanes, heat control system— defines the type of motherboard you need","categories":["Deep Tech"],"tags":["Intel","Intel chip"],"author_name":"Gourav Mishra","publish_date":"2021-08-18T10:00:00","publication_year":"2021","word_count":875,"keywords":["Go","machine learning","programming_languages:R","AI","Intel chip","RAG","Ray","Aim","deep learning","ViT","R","Intel"],"extracted_tech_keywords":["AI","machine learning","deep learning","Aim","Ray","RAG","R","Go","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/importance-of-motherboard-in-deep-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10068517,"title":"GoI launches AI for India campaign to teach AI to 25L citizens","content":"Union Minister of Education Dharmendra Pradhan recently launched the “AI for India” campaign with an aim to make the country the ‘AI Capital of the World’. Powered by AWS and supported by the Ministry of Education (GoI) and AICTE, the campaign aims to evaluate, train, practice, provide internship, allocate projects, certify and employ 25 lakh Indian citizens. “AI for India Initiative is trying to bridge the gap between the supply and demand matrix in worldwide tech talent, with Indian students. We are changing our educational policies to skill and upskill students to make them future-ready,” said Chandrashekhar Budha, CCO, AICTE.  “AI for India will empower the 21st Century generation in Arts, Commerce, Science, Architecture, Engineering, and other streams with digital education for employability and entrepreneurship,” he added. The campaign entails: 1. National Future Engineering Scholarship Examination: Aimed for AI adoption from class 8th to 12th Students, Undergraduate\/postgraduate Students and Individuals wanting to Upskill on emerging technologies. 2. All India Skill to Scale Avenue: This is for every student unconditionally who have passed out in pandemic years and have gaps in skill sets as compared with industry requirements. 3. AI Ideathon: This is to encourage students and individuals to produce ideas to solve any human centric national issue with the help of AI, Data, Cloud and allied sciences. 4. All India Data Engineering Quiz Competition: To diversify knowledge on data across domains and verticals. 5. All India Jobathon for Cloud, Data and AI Aspirants: All the participants who participate and complete any one of the above initiatives will be given a chance to appear for 100000 jobs from industries in India. “The DataTech Labs’ is extremely privileged envisaging this campaign and contributing in skilling, building our Nation. AI for India initiative welcomes the age of blended workforce, in which intelligent technologies and humans collaborate to drive business success and the Indian economy,” said Dr Amit Andre, CEO , The Data Tech Labs. AI for India campaign is a part of ABCDEFGHI program of GoI to impact and train students in the areas of Artificial Intelligence, BlockChain, Cyber Security, Data analytics and intelligence, Extended Reality (XR), Electronics, Energy solutions, Full stack developer, Gamification, GitHub, HTML5, Human computer interaction and IoT. Under this program, GoI envisions to train more than one crore students in these emerging areas for 3 to 6 months. “This initiative by DataTech is an appreciable one! To give to the society in a form that builds a stronger future for the young generation, helps in building the economy and ensures India comes in the limelight as a pool of talent needs support and recognition from each one,” said  AICTE chairman Anil Shahatrbudhe.","excerpt":"AI for India campaign is a part of ABCDEFGHI program of GoI.","categories":["AI News"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-06-07T15:43:18","publication_year":"2022","word_count":443,"keywords":["Go","artificial intelligence","AWS","AI","ML","Git","RAG","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","Aim","RAG","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/goi-launches-ai-for-india-campaign-to-teach-ai-to-25l-citizens\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":46065,"title":"Apple Bionic A13 vs Huawei Kirin 990: AI Chip Wars Continue","content":"Gone were the days when people used to buy computers based on CPUs. The advent of System-on-a-Chip (SoC) has created a buzz over the market and every smart thing around us is currently integrated with these SoCs. SoCs are a combination of various computer components into a single silicon integrated chip which includes a central processing unit (CPU), a graphical processing unit (GPU) and a random access memory (RAM), power management circuits, among others. According to International Data Corporation, the Chinese telecom multinational company, Huawei has surpassed Apple and acquired the second position to be the worldwide quarterly smartphone vendor. Huawei holds 19 percent of the worldwide market share which is high as compared to Apple’s 11.7 percent. In this article, we see how Huawei’s Kirin 990 stacks up against Apple’s A13 Bionic chipset, pegged as the most powerful chip in the smartphone. With two machine learning accelerators, A13 has hogged a lot of limelight but Huawei is gearing up for a tight battle too by packing an edge. Read on to find out more. Apple Bionic A13 It has been more than two years that Apple started making its own chipsets. This month the tech giant unveiled the new chipset Bionic A13 which according to the developers is the fastest iPhone silicon chip yet.  This chip will power the new iPhone 11 series. According to sources, “The A13 Bionic is the fastest CPU ever in a smartphone,” Apple said onstage, adding that it also has “the fastest GPU in a smartphone,” too. Huawei Kirin 990 After its involvement in the trade war with the US, Chinese telecom giant Huawei has doubled down on building their own competencies. From announcing its own operating system to launching its first commercial AI chip, the tech giant has grown its roots in the field of emerging technologies. The Chinese telecom giant, Huawei introduced its new flagship chipset, Kirin 990 which is made of TSMC’s 7nm+ EUV process. Huawei claims this chipset as the 1st flagship 5G SoC powered with a 7nm+ EUV process.  With a built-in 5G modem, the chip has 10.3 billion transistors, while with the 4G modem, the chip has 8 billion transistors. Artificial Intelligence The Apple Bionic A13 includes machine learning hardware acceleration which is an octa-core Neural Engine. It adds two machine learning accelerators to six CPU cores i.e. two performance cores and four efficiency cores that make the chipset work six times faster than the previously released chipsets like Bionic A12. According to sources, Apple Vice President of Engineering Sri Santhanam said A13 Bionic’s standout features include its powerful machine learning capability and a low-power design — hundreds of voltage gates enable the chipset to activate only the subsection(s) required for a given process at a given time. The Huawei Kirin 990 chipset includes the company’s self-developed 2+1 Da Vinci architecture NPU i.e. three cores which deliver better power efficiency, stronger processing capabilities and higher accuracy. The chipset has the powerful Big-Core plus ultra-low consumption Tiny-Core which contributes to an enormous boost in AI performance. In AI face recognition, the efficiency of this NPU Tiny-Core can be enhanced up to 24x than the Big-Core. According to sources, Huawei Kirin 990 betters Qualcomm’s Snapdragon 855 by 4.7 times in AI performance, and scored 52403 on the AI Benchmark test, ahead of both the Unisoc Tiger T710 and Snapdragon 855 Plus. Processors The Apple A13 Bionic is manufactured with TSMC’s second-generation 7nm process and includes 8.5 billion transistors which is approximately 23 percent more than A12’s 6.9 billion. The CPU is a dual-cluster Hexa-core and the GPU is the company’s own designed quadcore. On the other hand, the CPU in Kirin 990 is the tri-cluster octa-core and the GPU is 16-core Mali-G76 MC16 architecture which elevates outstanding performance and great energy efficiency. Conclusion In the current scenario, microchips are embedded in almost every smart device, for instance, smartphones, washing machines, and cars among others. This time, there has been a tight battle been the two tech giants, Apple and Huawei. Developers at Apple claim that it has the fastest CPU and GPU embedded in a smartphone. It also packs an image processing feature known as Deep Fusion which uses machine learning to improve low to medium light photography. However, there is one point where Huawei scores over Apple is with Balong 5000 modem processor which supports 5G networking and delivers maximum download speed of 2.3Gbps and a maximum upload speed of 1.25Gbps.","excerpt":"Gone were the days when people used to buy computers based on CPUs. The advent of System-on-a-Chip (SoC) has created a buzz over the market and every smart thing around us is currently integrated with these SoCs. SoCs are a combination of various computer components into a single silicon integrated chip which includes a central […]","categories":["Deep Tech"],"tags":["AI Chip Wars","AI Chips","Apple Bionic"],"author_name":"Ambika Choudhury","publish_date":"2019-09-17T15:00:21","publication_year":"2019","word_count":742,"keywords":["Go","Apple Bionic","artificial intelligence","machine learning","programming_languages:R","AI","AI Chip Wars","RPA","programming_languages:Go","RAG","AI Chips","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","RAG","R","Go","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/apple-bionic-a13-vs-huawei-kirin-990-ai-chip-wars-continue\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":58267,"title":"Rise of the Customer Data Platform","content":"Retail businesses have moved from a product-centric model to a customer-centric model. This transformation has a significant impact on how companies are engaging with their customers who are spoilt for choice, well informed and tech-savvy. Traditional approaches to managing customers were through Customer Relationship Management (CRM) processes and systems, however, in today’s hyperconnected world we see the emergence of Customer Data Platforms (CDP) as a robust model for handling customer data from a multitude of online and offline sources. The rising interest in Customer Data Platforms (CDPs) is reflected in the higher number of vendors providing these services as well as more venture capital funding (currently estimated at 2.4 billion USD). What is a Customer Data Platform? Customer Data Platforms give a unified view of the customer from a multitude of touchpoints that are beyond the realm of traditional CRM systems. CDP’s can integrate data from both structured and unstructured sources, as well as online and offline sources to build a unified customer profile. The key here is traceability of the customer profile through the lifecycle of a customer and their interactions across the lifecycle. The key difference with CRM is its ability to handle wider touchpoints, ability to trace customers from site visitors to actual customers. Types of Data – Customer Data Platforms (CDP’s) collect different types of data including 1) Identity Data (such as Name, Demographic information, Contact information, Social Media data etc.) 2) Descriptive Data – around family, career, lifestyle etc. 3) Quantitative Data – Transaction information like number and type of products, services purchased or returned, order dates, items removed from shopping cart, Online activity data like website visits, product views, searches, Customer service data like query dates, query details and customer service feedback, 4) Qualitative Data – such as motivation to buy a company’s products or services, opinion about service quality or favorite food, color etc. Touchpoints – The core tenet of a CDP is capture First-Party data i.e. directly from the customers or potential customers. The touchpoints of data collection could be through a plethora of channels like websites, mobile apps, customer feedback forms, call center logs etc. Lifecycle – CDP’s collect First-Party data through the lifecycle of a customer, from an anonymous visitor to a subscribed customer. Data collected from multiple touchpoints and through the lifecycle of a customer is very compelling in terms of the value an enterprise from deriving from these data sets. Compliance and Transparency – One of the key considerations around responsible use of CDP’s is to ensure that the data security and privacy laws of the country or region are complied to. With the rise of data privacy regulations like GDPR in Europe, CCP in California, regulators are increasingly looking at data privacy and security concerns of data subjects. CDP’s need to ensure transparency around the usage of the data, ensure secure by design principles are followed and consent is taken from data subjects whose data is stored and processed by CDP’s. The semantics of the stored data and the transparency of insights come from the governance policies and associated metadata that the platform provides.Platforms – CDP could be a packaged platform or a custom service provided by a systems integrator. At the core of the platform would be a data repository that stores all the customer data types captured through the different channels and touchpoints. CDP’s could be on-prem or hosted on a Public\/Private\/Hybrid Cloud depending on the industry regulatory requirements. Along with the Platform would be the visualization layer with BI tools, Advanced Analytical models and sandboxes for Data Science. In addition, consuming applications that need Customer data could access the CDP through API calls thereby extending the reach and usage of the platform. Benefits of a Customer Data Platform Now that we understand what a CDP is and what kind of data it stores and processes, let us review some of the benefits a CDP brings to an enterprise – Data Quality and Customer Relationship – CDP’s collect First-Party data, and this enables the quality of data to be recent and of the highest quality. They also retain the data over the entire customer lifecycle, which helps them build in-depth, detailed customer profiles and nurture better relationships with customers. Unified View of Customer Profile – CDP’s ability to collect both online and offline data from multiple channels and touchpoints, helps build a single unified view of the customer that can be leveraged by multiple departments of an enterprise from Marketing, Product Engineering, Customer Service and Operations etc. This unified view helps identify cross-sell opportunities and perform customer behavior analysis based on the rich data sets. Unify Cross Channel Marketing efforts – Enterprises often have a cross channel and multi-channel marketing initiatives and the CDP with its unified view of the customer helps align these marketing efforts with consistency and limited wastage. This helps build a culture of Customer-Centric marketing which is incisive with a deep understanding of customer choices and trends. Regulatory Compliance – CDPs are helping enterprises meet regulatory compliance needs by having robust information governance policies around the usage of data and ensuring the consent of customers is taken while storing their data. Customers feel empowered as they have control over the data they share and can withdraw consent or have specific needs met eg: do not call being enabled, where they are contacted over email or other channels based on their preferences. According to a recent Forbes survey of marketing executives of CDP usage, 53% mentioned CDP was helping marketing teams to understand their existing customers’ needs, increasing the chances of repeat business based on the positive customer experience. This article is presented by AIM Expert Network (AEN), an invite-only thought leadership platform for tech experts. Check your eligibility.","excerpt":"Retail businesses have moved from a product-centric model to a customer-centric model. This transformation has a significant impact on how companies are engaging with their customers who are spoilt for choice, well informed and tech-savvy. Traditional approaches to managing customers were through Customer Relationship Management (CRM) processes and systems, however, in today’s hyperconnected world we […]","categories":["AI Features"],"tags":["trends in bi and data visualization"],"author_name":"saumyachaki","publish_date":"2020-03-09T12:01:00","publication_year":"2020","word_count":957,"keywords":["trends in bi and data visualization","data science","Go","API","AWS","AI","RAG","Aim","ViT","data quality","R"],"extracted_tech_keywords":["AI","data science","Aim","RAG","AWS","R","Go","API","data quality","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/rise-of-the-customer-data-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085240,"title":"Microsoft Announces Azure OpenAI Service for All","content":"Microsoft today announced the general availability of its cloud-based Azure OpenAI service so that common people can use its AI tools like GPT-3.5, Codex, and DALL•E 2 to enhance their work.  It also said customers could access OpenAI’s flagship AI chatbot ChatGPT through Azure. OpenAI too announced that it would add ChatGPT to its API soon. We've learned a lot from the ChatGPT research preview and have been making important updates based on user feedback. ChatGPT will be coming to our API and Microsoft's Azure OpenAI Service soon. Sign up for updates here: https:\/\/t.co\/C7kMVpMAKv— OpenAI (@OpenAI) January 17, 2023 Those using the Azure Service already have access to tools like the GPT-3.5 language system that ChatGPT is based on and the Dall-E model for generating images from text prompts. “ChatGPT is coming soon to the Azure OpenAI Service, which is now generally available, as we help customers apply the world’s most advanced AI models to their own business imperatives,” tweeted Microsoft chief Satya Nadella. ChatGPT is coming soon to the Azure OpenAI Service, which is now generally available, as we help customers apply the world’s most advanced AI models to their own business imperatives. https:\/\/t.co\/kQwydRWWnZ— Satya Nadella (@satyanadella) January 17, 2023 Microsoft stated that organizations of all sizes and industries are utilizing the Azure OpenAI Service to achieve greater results with fewer resources, enhance the user experience, and streamline internal operations. Both small startups like Moveworks and large multinational corporations like KPMG are leveraging the capabilities of Azure OpenAI Service for advanced applications such as customer support, personalization, and extracting valuable insights from data through search, extraction, and classification. Recently, OpenAI announced its monetisation plans for ChatGPT and posted a waitlist link on its Discord server along with a range of questions on payment preferences for the paid version, which will be called ChatGPT Professional. This news came soon after Microsoft revealed that it’s now looking at pumping an additional $10 billion in OpenAI, a huge leap from its initial $1 billion investment in the company in 2019. When OpenAI rolled out ChatGPT in November, it went viral in less than 10 days. Built on GPT-3.5, ChatGPT can interact with humans in natural language and text and remember the context. Whether writing codes or a joke, the futuristic chatbot can do both.","excerpt":"Microsoft said that they would add ChatGPT to Azure soon. OpenAI also said that ChatGPT is coming to its API.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Google","Machine Learning","Microsoft","Python"],"author_name":"Shritama Saha","publish_date":"2023-01-17T10:42:28","publication_year":"2023","word_count":382,"keywords":["ChatGPT","API","DALL-E","Microsoft","AI","OpenAI","R","ML","Machine Learning","RAG","Python","GPT","Google","Deep Learning","Data Science","Data Scientist","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","RAG","Azure","R","API","GPT","DALL-E"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-announces-azure-openai-service-for-all\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":66930,"title":"What Are DPUs And Why Do We Need Them","content":"We have heard of CPUs and TPUs, now, NVIDIA with the help of its recent acquisition Mellanox is bringing a new class of processors to power up deep learning applications — DPUs or data processing units. DPUs or Data Processing Units, originally popularised by Mellanox, now wear a new look with NVIDIA; Mellanox was acquired by NVIDIA earlier this year. DPUs are a new class of programmable processor that consists of flexible and programmable acceleration engines which improve applications performance for AI and machine learning, security, telecommunications, storage, among others. The team at Mellanox has already deployed the first generation of BlueField DPUs in leading high-performance computing, deep learning, and cloud data centres to provide new levels of performance, scale, and efficiency with improved operational agility. The improvement in performance is due to the presence of high-performance, software programmable, multi-core CPU and a network interface capable of parsing, processing, and efficiently transferring data at line rate to GPUs and CPUs. According to NVIDIA, a DPU can be used as a stand-alone embedded processor. DPUs are usually incorporated into a SmartNIC, a network interface controller. SmartNICs are ideally suited for high-traffic web servers. A DPU based SmartNIC is a network interface card that offloads processing tasks that the system CPU would normally handle. Using its own on-board processor, the DPU based SmartNIC may be able to perform any combination of encryption\/decryption, firewall, TCP\/IP and HTTP processing. “The CPU is for general-purpose computing, the GPU is for accelerated computing and the DPU, which moves data around the data centre, does data processing.”NVIDIA CEO Why Do We Need DPUs These DPUs are known by the name of BlueField that have a unique design that can enable programmability to run at speeds of up to 200Gb\/s. The BlueField DPU integrates the NVIDIA Mellanox Connect best-in-class network adapter, encompassing hardware accelerators with advanced software programmability to deliver diverse software-defined solutions. Organisations that rely on cloud-based solutions, especially can benefit immensely from DPUs. Here are few such instances, where DPUs flourish: – Enabling storage, networking, and security to be part of composable infrastructure for cloud service providers Allowing the operator to facilitate bare metal to the cloud tenant as a service while preserving control over the server and protecting the environment Bare metal environment is a network where a virtual machine is installed Can enable and isolate environment to accelerate compute-intensive security functionsCan store, compute and secure data at the highest speeds while lowering cost and time by analysing data at the edge The shift towards microservices architecture has completely transformed the way enterprises ship applications at scale. Applications that are based on the cloud have a lot of activity or data generation, even for processing a single application request. According to Mellanox, one key application of DPU is securing the cloud-native workloads. For instance, Kubernetes security is an immense challenge comprising many highly interrelated parts. The data intensity makes it hard to implement zero-trust security solutions, and this creates challenges for the security team to protect customers’ data and privacy. As of late last year, the team at Mellanox stated that they are actively researching into various platforms and integrating schemes to leverage the cutting-edge acceleration engines in the DPU-based SmartNICs for securing cloud-native workloads at 100Gb\/s. According to NVIDIA, a DPU comes with the following features: Data packet parsing, matching, and manipulationGPU-Direct accelerators to bypass the CPU and feed networked data directly to GPUsTraffic shaping “packet pacing” accelerator to enable streaming 4K\/8K VideoPrecision timing accelerators for 5G capabilitiesCrypto accelerationVirtualisation support Secure Isolation Know more about DPUs here.","excerpt":"We have heard of CPUs and TPUs, now, NVIDIA with the help of its recent acquisition Mellanox is bringing a new class of processors to power up deep learning applications — DPUs or data processing units. DPUs or Data Processing Units, originally popularised by Mellanox, now wear a new look with NVIDIA; Mellanox was acquired […]","categories":[],"tags":["Firewall hardware","International Affairs","network firewall","NVIDIA"],"author_name":"Ram Sagar","publish_date":"2020-06-09T14:00:54","publication_year":"2020","word_count":593,"keywords":["network firewall","API","machine learning","TPU","AI","International Affairs","RAG","microservices","deep learning","Firewall hardware","Rust","NVIDIA","R","kubernetes"],"extracted_tech_keywords":["AI","machine learning","deep learning","RAG","kubernetes","microservices","TPU","R","Rust","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dpu-nvidia-mellanox-processors\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":43859,"title":"What Does It Take To Design The Next Generation Chatbots","content":"Joaquin Phoenix in Her Going beyond simple tasks like playing a song or booking an appointment requires generating coherent and engaging responses in conversations requires a range of nuanced conversational skills, including language understanding and reasoning. If AI has to master the art of conversation at human level, then it has an uphill task ahead and Facebook AI lists 5 areas where chatbots can improve: consistency, specificity, empathy, knowledgeability, and multimodal understanding. When Chatbots Empathise In recent work with researchers from the University of Washington, introduce the first benchmark task of human-written empathetic dialogues centered on specific emotional labels to measure a chatbot’s ability to display empathy. We can’t be too hard on AI in this regard as we humans, too, are bad at showing empathy but without proper training data, a chatbot can convey congratulations in place of condolences. Training AI for empathy is one of the final frontiers of AGI, and is still debatable. Nevertheless, since, the improvements will be made to the models, we need to address this aspect of NLP models as well. Leveraging Knowledge Current state-of-the-art approaches to dialogue modeling involve sequence-to-sequence models, which lack access to information outside of the conversation history. Now, new models are created with the help of Wikipedia data to retrieve knowledge and then use it to respond to dialogues. Connecting With Imagery A picture speaks a thousand words. This cliche is in its truest form in the machine world where every pixel counts. It can be an edge, curve or color intensity, the information, indeed is worth a thousand words. Modern neural networks especially CNNs have mastered object detection and this aspect can be exploited to improve chatbot conversation. Machine learning approaches that comment on images have typically focused on image captioning, which is factual and neutral in tone — like “fireworks in the sky.” In this research, the authors focussed on image captioning that is engaging for humans by incorporating personality. A large data set of human comments grounded in images was collected, and train state-of-the-art models capable of discussing images with given personalities, which makes the system much more interesting for humans to talk to. Consistency Dialogue natural language inference dataset was specifically created to test techniques that maintain consistency. In Dialogue NLI, two utterances in a dialogue are considered as the premise and hypothesis, respectively. Each pair is labeled to indicate whether the premise entails, contradicts, or is neutral with respect to the hypothesis. Training an NLI model on this data set and using it to rerank the model’s responses to entail previous dialogues — or maintain consistency with them — improves the overall consistency of the dialogue agent. Specificity via Facebook AI “What do you do for a living?” A typical chatbot responds with the generic statement “I’m a construction worker.” With control methods like the one done in collaboration with Stanford AI, the chatbots proposed more specific and engaging responses, like “I build antique homes and refurbish houses.” Studying multiturn aspects is necessary to improve conversation quality. The better the overall conversation flow, the more engaging and personable the chatbots and dialogue agents of the future will be. A Future Of Intelligent Interactions It is possible to train models to improve on some of the most common weaknesses of chatbots today. The future, intelligent chatbots will be capable of open-domain dialogue in a way that’s personable, consistent, empathetic, and engaging. The implications of conversational AI agents can be expected to be beyond the boundaries of trivial social networking chats. They could be deployed for healthcare emergency scenario where the user can type in some text or send an image indicating a distressed signal. Even if we move away from these grave scenarios, enterprises have a lot to gain from intelligent chatbots. They can be deployed for real time grievance redressal (ex:Swiggy and Zomato) or banks and insurance agencies can deploy them for quick loan approvals and answering queries. This eliminates the down time and captures a wider range of population. Access more info on open source datasets and methods here.","excerpt":"Going beyond simple tasks like playing a song or booking an appointment requires generating coherent and engaging responses in conversations requires a range of nuanced conversational skills, including language understanding and reasoning. If AI has to master the art of conversation at human level, then it has an uphill task ahead and Facebook AI lists […]","categories":["Deep Tech"],"tags":["AI Chatbot","Conversational AI","Facebook AI","NLP"],"author_name":"Ram Sagar","publish_date":"2019-08-06T11:00:29","publication_year":"2019","word_count":676,"keywords":["Go","machine learning","Facebook AI","AI","AI Chatbot","neural network","chatbots","RAG","NLP","object detection","Conversational AI","CNN","R"],"extracted_tech_keywords":["AI","machine learning","neural network","NLP","RAG","chatbots","object detection","R","Go","CNN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/what-does-it-take-to-design-the-next-generation-chatbots\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10054607,"title":"JetBrains Releases IDE “DataSpell” Especially For Data Scientists","content":"JetBrains recently announced the official release of DataSpell, its new data science IDE. DataSpell by JetBrains is designed specifically for those involved in exploratory data analysis and prototyping ML models. It combines the interactivity of Jupyter notebooks with the intelligent Python and R coding assistance of PyCharm in one ergonomic environment. DataSpell supports both local and remote Jupyter notebooks. It is possible to work with them right inside the IDE exactly as using traditional web-based notebooks. The main advantage of DataSpell over Jupyter or JupyterLab is the availability of intelligent coding assistance and lots of other features offered by the full-fledged IDE. DataSpell also supports Jupyter’s command mode, most of its standard shortcuts, Markdown and LaTeX, and interactive outputs. For Python and R scripts, it allows running entire scripts or parts of them and browsing outputs interactively and conveniently. Support for local Jupyter notebooks is now also bundled with PyCharm Professional in DataSpell and offers more out of the box features for data scientists, focusing on data and interactivity. DataSpell provides a lightweight workspace model that allows to reuse configured environments, attach multiple folders with data, scripts, and notebooks, or connect it to multiple remote instances of Jupyter servers. In addition, database support, built-in debuggers, terminals, Git support, and a whole bunch of plugins are available for the IntelliJ platform, including Docker, Material Theme UI, and GitHub Copilot. You can download DataSpell and check its features using the link here.","excerpt":"The main advantage of DataSpell over Jupyter or JupyterLab is the availability of intelligent coding assistance and lots of other features offered by the full-fledged IDE.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Tool","Data Science","Data Science Jobs","Deep Learning","Docker","GitHub","Github Copilot","GitHub Repositories","JetBrains","Jupyter Notebooks","Machine Learning","Pycharm","Python"],"author_name":"Victor Dey","publish_date":"2021-12-01T13:34:15","publication_year":"2021","word_count":240,"keywords":["TPU","Git","GitHub","AI Tool","R","data science","docker","Data Science","Docker","Go","AI","ML","Machine Learning","Data Science Jobs","GitHub Repositories","Jupyter Notebooks","Github Copilot","Python","JetBrains","Pycharm","Deep Learning","Jupyter","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","data science","Jupyter","docker","TPU","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jetbrains-releases-ide-dataspell-especially-for-data-scientists\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10109642,"title":"Chris Lattner on Ending AI Suffering","content":"Chris Lattner, the man who walks around fixing programming languages and compilers, doesn’t spend any time worrying about AGI personally. “People worry about what happens when a computer is smarter than us, but at the same time, we are surrounded by other people smarter than us. At least I am,” mused the CEO of Modular casually. He believes that we’re a long way away from AI replacing the human capability of the world. “I have major questions about how much compute will take when we power it. From a pure technology level, I don’t think that’s on the cusp,” he remarked about the billions funnelled into building AI models in 2023 alone. Lattner suggested that folks worried about this should take a step back and realise we already have super intelligence. “It’s made out of federated groups of humans with shared goals. That’s what I think has been true for hundreds of years, and that’s where progress is made,” he said. Speaking of progress, the startup challenging and partnering with NVIDIA, the top supplier of AI chips, recently secured $100 million from Silicon Valley heavyweight General Catalyst. he lean 76-person startup is building its Modular Accelerated Xecution (MAX) platform, which is composed of the MAX Engine, a unified inference library that optimizes and executes AI models from popular AI frameworks like TensorFlow, PyTorch, and ONNX, across all CPU architectures and NVIDIA GPUs; MAX Serving, which easily deploys those optimized models into any cloud service; and the Mojo programming language, that combines the usability of Python with the performance of C, unlocking unparalleled programmability of AI hardware and extensibility of AI models. Lattner wants people to be able to express themselves and create things. “I love that somebody doesn’t have to worry about the exact syntax because it is mechanical. That is not the point of coding,” he said. “That’s just something we must do because we need to express ourselves. It gets more and more people involved, making things more accessible. It’s just giving people superpowers,” the software genius added. Not A Typical AI Company There’s a clear message on Modular’s website: It focuses on AI because, as Lattner puts it, “that’s where most of the suffering is.” Lattner knows firsthand the struggles of programmers. “It bothers me,” he shares, showing a real understanding of the field. For him, Modular is about making the complex world of AI more straightforward. He explained, “It is about making it possible for more people to participate in this ecosystem.” He points out that only big players like Google and OpenAI with access to large teams of experts seem to dominate AI right now. The company has large teams of experts. But Lattner believes this approach leaves out many talented people worldwide. His vision for Modular is to change that. Before launching Mojo through Modular, Lattner has contributed to open-source projects like the LLVM Compiler Infrastructure, Clang, and Swift, a programming language used in many Apple products. Tech Now and Then Growing up, Lattner started learning about computers and programming as a kid. Now, at 45, he’s spent decades creating tools to help other programmers build things. Around 2016, when AI was in its early stages, he got interested in the subject. He tried to get folks at Apple to understand why it was necessary. “There were always more important things, and they were not excited about this,” he said. “I decided to go on a hero’s journey of understanding how all the technology worked and spent several years at Tesla, Google and other places, learning the fundamentals, over the last 5 to 7 years,” said Lattner, who has led teams at Apple, Tesla, Google and most recently, SiFive. Looking to 2024, he thinks it takes quite a while for a technology base to build out to the point where the world understands it. Much of the AI out there is demo quality, said Lattner. “2023 was the year of the language model demo. There wasn’t a tonne of language models impacting products. There were a lot of ChatGPT wrappers, but the impact was pretty low. One of the reasons for that is that technology is problematic. But [generative AI] is disproportionately important, and there’s something real there, “he asserted. Lattner predicts that 2024 will be the year of Generative AI getting into products. Lattner further highlighted that Generative AI is not classical AI running on neural nets. “It’s a neural net embedded into larger applications where new data comes in. Language models have to be tokenized, a lot of this is non-traditional, and the stacks people have been writing on were never designed for that.” For his company’s future, Lattner hinted that in the next few quarters, there will be continuous announcements about capabilities, new product features, developer APIs, ecosystem components, and even more things built on top of MAX and Mojo. “Right now, we’re very focused on inference; we’ll soon go into training. Training is a major sore point as models get larger; that’s a big deal for the world. We’ll support new hardware to support more of the AI workflow. We try to do everything to the best quality, so one of the things that is true about Modular that annoys some people is that we move slowly,” he admitted.","excerpt":"The Modular chief suggests that AI folks should take a step back and realise we already have super intelligence","categories":["AI Features"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-12-29T12:06:50","publication_year":"2023","word_count":879,"keywords":["Go","ChatGPT","API","OpenAI","AI","PyTorch","Python","generative AI","TensorFlow","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","TensorFlow","PyTorch","Python","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chris-lattner-on-ending-ai-suffering\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":31949,"title":"Clustering Techniques Every Data Science Beginner Should Swear By","content":"Cluster analysis is the statistical method of grouping data into subsets that have application in the context of a selective problem. This technique is widely used to club data\/observations in the right segments so that data within any segment are similar while data across segments are different. However, defining “similar” or “different” observations is a key part of cluster analysis which often requires contextual knowledge and creativity beyond what statistical tools can provide. Unlike analysis, clustering does not rely on predefined classes. Clustering is considered to be one of the most important unsupervised learning methods because no information is provided about the best answer for any of the objects. It can reveal previously undetected correlations in a composite dataset. For example, in a business relevance, cluster analysis can be used to identify and characterise customer associations for marketing objectives. Necessity Clustering is vital for data mining. It solves many issues related to data mining in a very efficient way. Clustering allows grouping of similar data which helps in understanding the internal structure of the data In some instances, distribution or apportionment is the main objective of clustering. This reduces unwanted data and helps save time The various methods which are involved in clustering assist in the knowledge discovery of data Clustering prepares the data for other AI technologies General Types Of Clusters Well-Separated Clusters: Well separated clusters are the clusters in which set of objects are significantly closer to each other than the objects which are not in the cluster. Centre-Based Clusters: In a cluster when a set of objects are present in such a way that an object in a cluster is close to the centre of a cluster as compared to other cluster centres. The core of the cluster is usually referred to as the centroid, the median of all the points in the cluster is often known as mediods. Density-Based Clusters: When a cluster is composed of a dense region of points, which are separated by low-density areas, from other regions of high density. These clusters are variable, and when noise and outliers are present in data. Shared Property or Conceptual Clusters: Obtains clusters that share some common characteristics or designate a particular concept. Contiguous Cluster: A cluster holds a collection of points such that a point in a cluster is closer or more related to one or more other points in the cluster than to any point not in the cluster is known as a contiguous cluster. Cluster Analysis The expression cluster analysis includes various algorithms and approaches for grouping things of related characteristics into separate sections. The availability of different algorithms helps users to combine discovered data into significant formats. The following are some of the well-known algorithms and methods that are used to create formations in data. Basic Agglomerative Hierarchical Clustering Algorithm Hierarchical clustering is a process of cluster analysis which attempts to build a hierarchy of clusters. It is the connectivity based clustering algorithms. The hierarchical algorithms models clusters regularly. Hierarchical clustering commonly divided into two types. Agglomerative: This is a “bottom-up” strategy where every observation starts in its personal cluster, and pairs of clusters are united as one moves up the hierarchy. Divisive: This is a “top-down” procedure where all observations start in one cluster, and divisions are implemented recursively as one moves down the hierarchy. Nearest Neighbour Clustering The algorithm is based on the idea of mutual neighbourhood value (mnv) of two points, which is the sum of the ranks of two points in each sorted nearest-neighbour lists. These clusters are created by raising with points as singleton clusters and then merging the closest set of clusters, where close is determined in the terms of the mnv. K-Nearest-Neighbors (kNN) The kNN order of classification is one of the easiest techniques in machine learning and data mining. The method actually classifies by looking for the most similar data points in the training data and making an instructed guess based on their classifications. Last Word The objective of the data mining method is to select information from a large data set and modify it into an acceptable form for additional use. Clustering is an important part of data analysis and data mining applications that help in achieving the goal of data related works.","excerpt":"Cluster analysis is the statistical method of grouping data into subsets that have application in the context of a selective problem. This technique is widely used to club data\/observations in the right segments so that data within any segment are similar while data across segments are different. However, defining “similar” or “different” observations is a […]","categories":["AI Features"],"tags":["clustering","clustering algorithms","data mining algorithms","hierarchical clustering","K-means clustering algorithm","outlier analysis in data mining"],"author_name":"Bharat Adibhatla","publish_date":"2018-12-20T12:24:28","publication_year":"2018","word_count":710,"keywords":["Go","machine learning","K-means clustering algorithm","hierarchical clustering","outlier analysis in data mining","AI","programming_languages:R","programming_languages:Go","clustering","clustering algorithms","ViT","R","data mining algorithms"],"extracted_tech_keywords":["AI","machine learning","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/clustering-techniques-every-data-science-beginner-should-swear-by\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10131908,"title":"Is OpenAI Intel’s Biggest Regret Ever?","content":"Intel is going through some trouble lately. The company’s recent earnings fell short of analysts’ expectations, resulting in a 26% single-day selloff that brought its market cap below $100 billion for the first time in three decades. Also, CEO Patrick Gelsinger announced to employees last week that the company would reduce its workforce by 15% and cut about 15,000 jobs as part of a significant cost-cutting measure. All this, while he’s also been posting proverbs from the Bible, which made people stress even more about the company. But this unfortunate time could have played out differently had the company made that one single investment back in 2017. According to reports, in 2017 and 2018, the tech giant had the opportunity to acquire a 15% stake in OpenAI for $1 billion. Additionally, Intel could have secured another 15% stake by offering OpenAI its hardware at cost, according to the sources. This would have made the company acquire a 30% stake in OpenAI, which is contentiously the leader in generative AI for the past few years. OpenAI sought Intel as an investor to reduce its dependence on NVIDIA, whose chips are possibly powering the entire AI world right now. A Bet Gone Wrong Intel declined the offer, partly because it doubted the immediate viability of generative AI models in 2018, which it believed would impact a timely return on investment. Cut to the present, it can be said that Intel is pushing really hard to make a strong presence in the AI industry. Once a world-leader in chips, Intel failed to capitalise on the AI boom and propelled NVIDIA to become one of the most valuable companies globally. But it is not just Gelsinger who is possibly praying for his business. NVIDIA chief Jensen Huang is also reportedly paranoid about the future of his company. In a recent podcast with Lex Fridman, Perplexity AI chief Aravind Srinivas revealed that he once asked Huang how he handles success and stays motivated. To this, Huang had replied, “I am paranoid about going out of business. Every day I wake up in a sweat, thinking about how things could go wrong.” Huang explained that in the hardware industry, planning two years in advance is crucial because fabricating chips takes time. “You need to have the architecture ready. A mistake in one generation of architecture could set you back by two years compared to your competitor,” Huang said. This definitely puts into perspective how a single investment could have changed Intel’s fortune, since even the CEO of the leading company is paranoid about things going wrong at any moment. But Intel is Not Sitting Ducks However, there is some positive news from Intel as well, which shows that the company is not giving up. For years, Intel focused on enabling CPUs, like those in laptops and desktops, for AI processes, rather than prioritising GPUs, which are more effective for AI calculations. In contrast, NVIDIA and AMD have thrived by concentrating on GPUs, while Intel largely missed the opportunity. However, in the third quarter, Intel plans to release its Gaudi 3 AI chip, which Gelsinger claims will outperform NVIDIA’s H100 GPUs, possibly even challenging NVIDIA Blackwell architecture. Continuing with its focus on chips, Intel has also announced that Panther Lake and Clearwater Forest, the leading products on Intel 18A, are now out of the fab and ready to run on operating systems. These would be ready for production by next year. Several people have cited that Gelsinger would bring Intel back on its feet after having almost lost in the AI race. The OpenAI failed deal of 2017-18 is something that Gelsinger, if he had been leading the company at that time, might have been able to make successful. In May, Gelsinger had said that the company’s AI strategy is on the right track, which made everyone think Intel was living in denial. “We’re really starting to see that pipeline of activity convert,” said Gelsinger. But apart from GPUs, Intel’s CPU and NPU plans are still seemingly strong, along with a focus on edge use cases and on-device AI. Since Intel is the majority holder of the laptop industry, with the future of AI racing towards smaller models, it is possible that Intel might rise in a year or two as the leader, spearheading the AI PCs game. Intel anticipates shipping 40 million AI PCs in 2024, featuring over 230 designs spanning from ultra-thin PCs to handheld gaming devices. There are no PCs without Intel – that’s for sure. Failed Deals and Poor Quarters are Part of the Game Undoubtedly, failed AI deals are part of the business. Recently, Elon Musk’s xAI cancelled the $10 billion deal with Oracle. Also, Apple has denied a partnership with Meta for AI.Intel is not the first one that failed to convert an OpenAI deal. Not many are aware that IT consulting giant Infosys, together with Musk, AWS, YC Research, and a few others, had donated a sizable $1 billion to OpenAI back in 2015, when the latter began as a non-profit organisation. But the donation did not turn into an investment. What if it is a bigger regret for OpenAI to not have Intel as one of its partners? Since the cost of running its business and AI offerings powered by NVIDIA is significantly higher for the company and is making it struggle to earn revenue, Intel could have helped OpenAI make its own hardware by now. When it comes to Intel, maybe owning this huge part of OpenAI would have been a failed strategy. Since Microsoft owns 49% of the shares in OpenAI now, things could have been quite different for all three companies. Moreover, Intel has a sweet spot for India. It has partnered with several companies in India, such as Krutrim, Bharti Airtel, Zoho, and several others, to provide its enterprise and data centre computing services. Maybe, Gelsinger’s interest in India would put Intel on the driving seat soon in the generative AI race.","excerpt":"Intel had the opportunity to acquire a 15% stake in OpenAI for $1 billion in 2017.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Mohit Pandey","publish_date":"2024-08-08T16:40:36","publication_year":"2024","word_count":997,"keywords":["Go","API","OpenAI","AI","AWS","R","Ray","Aim","generative AI","xAI"],"extracted_tech_keywords":["AI","generative AI","OpenAI","xAI","Aim","Ray","AWS","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/is-openai-intels-biggest-regret-ever\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21223,"title":"AI Turns Nicolas Cage Into James Bond, Indiana Jones and More","content":" When Nicolas Cage’s character swapped faces with John Travolta’s in the 1997 blockbuster Face\/Off, little did the actor expect that 20 years later fans would be using AI to swap his face with some of the most popular actors in Hollywood. A reddit user who goes by the name Derpfakes, has been using AI to replace actors’ faces with that of Cage’s and has been producing videos which have gone viral. There have been discussions and jokes on the internet regarding Cage’s appearance in nearly any film these days and this reddit user has managed to turn those jokes into reality. One of the most notable ‘swap’ shows Cage as James Bond in Dr. No instead of the legendary actor Sean Connery. Cage is also seen ‘replacing’ actor Harrison Ford as Indiana Jones in a scene from the movie Raiders of the Lost Ark.  And for those who may have wished to see Cage interacting with a clone of himself, their wish seems to have been fulfilled as well. An edited clip from a segment of an old episode of Saturday Night Live shows Cage interacting with a version of himself as a result of the segment’s host Andy Samberg’s face being replaced with Cage’s. Other hilarious swaps include Cage as Lois Lane (Amy Adams) in the movie Man Of Steel and Stannis Baratheon (Stephen Dillane) in HBO’s Game Of Thrones. Open source AI softwares, which have been made available freely for research and academic purposes, have played a major role in this sudden surge of ‘fake’ celebrity videos. A popular application, FakeApp, is speculated to have been used to insert Cage’s and other celebrities’ faces into videos. Fakeapp employs an algorithm that can scan a person’s face from photographic inputs and digitally overlay it on faces in video content. This has lead to massive amount of such video content generated on subreddits. Derpfakes announced on his Reddit page that he is will be moving on to other actors and sought suggestions for the next ‘torchbearer”. He said: “This is my final Cage fake. He’s had a good run and hopefully made some people take an interest in this sub and the tech behind it, but it is time to move on. Suggestions for a new torchbearer?” While this is an example of AI being used to create harmless content aimed to entertain, the use of AI to digitally morph faces of celebrities, and even common people, into obscene content has been on the rise lately. Recently, it was reported that Wonder Woman star, Gal Gadot’s face had been overlaid on that of an adult entertainer in a pornographic video. It has also been reported that the generation of AI doctored explicit content has not spared even regular people whose faces have been superimposed into such videos without their consent.","excerpt":" When Nicolas Cage’s character swapped faces with John Travolta’s in the 1997 blockbuster Face\/Off, little did the actor expect that 20 years later fans would be using AI to swap his face with some of the most popular actors in Hollywood. A reddit user who goes by the name Derpfakes, has been using AI […]","categories":["AI News"],"tags":[],"author_name":"Jeevan Biswas","publish_date":"2018-02-02T09:21:44","publication_year":"2018","word_count":472,"keywords":["Go","programming_languages:R","AI","ML","programming_languages:Go","Git","Aim","CLIP","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Git","CLIP","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-turns-nicholas-cage-james-bond-indiana-jones\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10102610,"title":"OpenAI Unveils Code Interpreter, a Thrilling Breakthrough","content":"After months of anticipation and requests from developers, OpenAI has finally released Code Interpreter API. “Code Interpreter is now available today in the API as well,” said Romain Huet, head of developer experience at OpenAI, in the backdrop of Assistants API launch. This new API is said to make it easier for developers to build their own GPT-like experience into their own apps and services. “These experiences are great but they have been hard to build, sometimes taking months, teams of dozens of engineers, there’s a lot to handle to make this custom assistant experience. So today we’re making it a lot easier with our new Assistants API,” said Altman. Currently in beta stage, you can try Assistant API here. The all new API provides new capabilities such as code interpreter and retrieval, alongside function calling to handle a lot of the heavy lifting, making it easy for developers to build high-quality AI applications. In addition to this, it has also introduced persistent and infinitely long threads, helping developers focus on context window constraints and leaving all the thread state management hassle to OpenAI. “With the Assistants API, you simply add each new message to an existing thread. (this is different from) other features.” said OpenAI, in its blog post. As far as the data safety is concerned, OpenAI claimed that its API are never used to train their models and developers can delete the data when they see it. Features of the Assistants API The Assistants API leverages Code Interpreter, OpenAI’s tool that writes and executes Python code in a controlled environment. Originally launched for ChatGPT in March, the Code Interpreter facilitates the generation of graphs, charting, and file processing. This functionality allows assistants developed with the Assistants API to iteratively run code for problem-solving in coding and mathematics. Moreover, the API incorporates a retrieval component, enabling dev-created assistants to access knowledge outside OpenAI’s models, such as product information or company documents. It also supports function calling, allowing assistants to trigger developer-defined programming functions and integrate their responses into messages. Beta Release and Usage The Assistants API is currently in beta and accessible to all developers. OpenAI will bill the tokens used at the chosen model’s per-token rates, where “tokens” refer to text fragments. In the future, OpenAI plans to enable customers to introduce their own assistant-driving tools to complement the existing Code Interpreter, retrieval component, and function calling features on its platform.","excerpt":"OpenAI releases Assistants API, making it easier for developers to build their own GPT-like experience into their own apps and services","categories":["AI News"],"tags":["ChatGPT","Code Interpreter","OpenAI"],"author_name":"K L Krithika","publish_date":"2023-11-07T02:45:58","publication_year":"2023","word_count":404,"keywords":["ChatGPT","API","OpenAI","AI","Code Interpreter","RAG","Python","GPT","Aim","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","RAG","Python","R","API","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-finally-launches-code-interpreter-api\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080301,"title":"Best AI Powered Video Editing Tools","content":"Recently, the AI-powered media editing app, Descript, raised $50 million in a Series C funding round by OpenAI Startup Fund, started by Microsoft, OpenAI, and their partners to invest in early-stage companies. Founded in 2017, Descript is an audio and video editing platform that leverages AI to make video editing easier. Editing videos on heavy hardware-driven softwares like Apple’s FinalCut or Adobe Premiere Pro is a cumbersome task and not easily available for people with a low-spec system. Editing, rendering, and sharing becomes an exhausting task resulting in delays and corruption in files. But the introduction of AI in this field has made the task easier and more browser friendly. Apart from being present in the cloud, Descript has a feature called “Overdub” where you can edit the audio of the video clip by just editing the transcription. Besides Descript, let’s look at some other promising media editing tools powered by AI that you can try out in your browser without the need for a heavy spec system. 1. VEED.io With auto-translation and subtitling capabilities, VEED.io is a perfect platform to make videos easily and quickly. The tools on the platform allow users to record, stream, and edit clips on one dashboard. Along with editing soundwaves and polishing audio, VEED.io allows you to remove background noises with flexibility. For social media influencers and creators, the platform can quickly crop into different ratios while analysing the content in the video for perfect placement of the subjects on any screen. Start Editing with: VEED.io 2. Biteable An online collaborative video editing platform that identifies the personality of your brand, Biteable, is a one-stop tool to create perfect personalised media for your company or group. The templates available on the platform can be converted into your own brand designs with a single click, enabling quick and hassle-free creation of branded content. The website also recently released a new feature of including stock footage for creating quick videos for organisations. The best feature of this platform lies in its ability to create branded content with just one click, doing away with the need of any other tool. Start Editing with: Biteable 3. InVideo With a library of over 5,000 templates, transitions, and effects, InVideo is a highly efficient professional video editor available online. InVideo also recently released its smartphone application that hosts all the same features. A powerful feature based on AI is the conversion of text to video – the simple processing of words into video. This highly-rated media creation and editing platform is free to try and has several plans that suit individual users and businesses as well. Start Editing with: InVideo 4. VideoVerse VideoVerse, formerly known as Touch.ai, is a simple video-editing tool on the cloud powered by deep learning. The platform consists of a smart-AI ecosystem, including products like Magnifi for real time video highlighting, Styck for live-streaming across multiple social media platforms, and Illusto is a web-based editor for editing videos for all media formats on the go. Read: How VideoVerse plans to take on Apple iMovie, Adobe Premiere Pro Start Editing with: VideoVerse 5. Vimeo Create If you want to create videos that look like they took hours to make within minutes, Vimeo Create is the perfect place for you. It’s a platform to build, edit, customise, and share from the same place. With over 2,000 video templates, editors can quickly render out a perfect copy of a branded video. Vimeo Create’s platform is ideal for creating media content for all social media platforms like Instagram, YouTube, TikTok, featuring various style options suitable for each platform. You can also check out their blog and tutorials for step-by-step guides. Start Editing with: Vimeo Create 6. Magisto Upload your files on the server, select a template, and Magisto will start doing the job for you. The platform features a powerful audio and visual analysis algorithm that matches the sequence seamlessly without requiring any input. The editor is also available for iOs and Android to edit on the go. Originally made for AI video generation, the software is excellent for editing videos. Start Editing with: Magisto 7. RunwayML One of the only platforms to offer green screen capabilities online, RunwayML is one of the top-rated AI-based editing tools that allows precise editing in the video, just like Premiere Pro or After Effects. You can remove objects completely or change the background of the frame by allowing AI to detect the frame and do all the work. Though the interface might take some time to get used to, the powerful platform which is completely online, can add complex effects into your videos without requiring too much hardware processing. Start Editing with: 8. Rawshorts Built for simplicity and no-fuss editing, Rawshorts is ideal for creating small videos with a user-friendly interface. It also includes the text-to-video feature that can convert your articles into a video by assembling and then visualising all the available information. This software is ideal for companies that use long-format text and want to shift their information into videos using AI. The generated video is also customisable with lots of flexibility for editing. Start Editing with: Rawshorts 9. Wisecut Calling itself an automatic video editor, Wisecut allows you to edit videos using voice recognition. You can upload your videos on the platform and also add speech or any voice on the server and it can automatically refine it by improving the quality and cutting out the silences. It can also split and crop your video into multiple vertical videos for social media and make it seem like the video is edited manually using hardware-based engines. Start Editing with: Wisecut 10. Aivo Created by Pantheon Lab Limited in Hong-Kong, Aivo makes the video-editing process seamless by including your brand elements easily into a desired video template. You can either start with a template, or post a script to frame the timeline according to your preference. The platform allows you to select the genre of the video like business, education, entertainment, travel, among others and select the ratio to achieve the desired output instantly available for sharing. Start Editing with: Aivo 11. Synthesia This platform is quite unique. Synthesia allows you to create talking head videos using text input. The videos can be up to 30 minutes long in more than 60 languages and many avatars. It also allows you to upload your own audio which is then synced with the avatar seamlessly. Start Editing with: Synthesia","excerpt":"AI tools for video editing have been on the rise for the right reasons, making content creation a lot easier","categories":["AI Trends"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2022-11-21T18:00:00","publication_year":"2022","word_count":1075,"keywords":["Go","TPU","OpenAI","AI","AWS","ML","RAG","deep learning","CLIP","R"],"extracted_tech_keywords":["AI","ML","deep learning","OpenAI","RAG","AWS","TPU","R","Go","CLIP"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-12-ai-powered-video-editing-tools\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":5337,"title":"Voice of Customer Analytics &#8211; A Deep Dive","content":"Voice of customer surveys are increasingly used across the world. There are smart software vendors that have created easy to use platforms to deploy such surveys as well as capture the data from the feedback. The article talks about the various analytical techniques from data mining to Natural Language Processing based text mining, that can be used to analyse the data generated from VOC surveys to provide actionable insights to the business. Background As the importance of social media increases, Voice of Customer (VOC) is increasing in importance. Word of mouth publicity, it can be a zero cost force multiplier for one’s marketing efforts, as well as a drag on one’s corporate reputation if the direction is negative. Recognising this trend, companies are increasingly using insights from VOC to drive business priorities. VOC data is being collected both from targeted surveys as well as customer care center interactions. This data is often a mix of structured and free form information. While traditional analytic techniques can be used to mine the structured part of the VOC data, a variety of NLP based tools are required to lend structure to the free form text data that is often equally important and rich. Key measures of CSAT from VOC data VOC surveys usually track 2 key metrics  – Top 2 Box Score and Net Promoter Score (NPS) VOC scores are typically on a 11 point Likert scale (0-10), with responders being divided into the following categories Promoters – 8 , 9 & 10 Passives – 6 & 7 Detractors – 0-5 Top 2 Box Score is calculated as the percentage of the total respondents that have marked a rating of 9 or 10 while NPS is calculated as the percentage of promoters – percentage of detractors Global benchmarks for NPS > 70% -> World class performance 50 – 70% -> Acceptable levels < 50% -> Improvement opportunity The NPS \/ Top 2 Box are computed for the overall company, specific business units, functional areas (finance, support etc) and trends analysed for actionable insights. Using the VOC data to drive analytical decision making Directed data mining of the VOC data generates insights into the company’s strengths and opportunity areas Use VOC to drive  strategy across various  business units and functional areas Differentiate using attributes where CSAT performance is at world class levels Improve in areas where CSAT levels are below satisfactory Maintain service levels  in areas where performance is satisfactory , some areas with high impact on overall NPS can be selected for improvement Identify trends in CSAT performance by analysing CSAT time series data Areas that are picking up over time Areas with historically high levels of CSAT performance that are showing signs of a down turn Extensive deep dive analysis to understand customer feedback in greater detail Geographical cuts – are there trends specific to geographies that are different from overall trends Cuts by customer role within client organisation – Do executive sponsors rate  us differently than practitioners? Does IT department rate us differently than marketing? Etc Competitive benchmarking Simple questions on competitive performance included in the survey help the client benchmark their performance against  key rivals The data can also be fed into statistical models that can identify key drivers of growing CSAT or Top 2 box scores Using the NPS or Top 2 Box as dependant variable and the scores on various attributes as independent variables , Logistic Regression \/ Decision Tree \/ Neural Network models determine the key drivers that show a statistically significant impact on CSAT These are the attributes that  should be focused on with top priority, as they give the greatest ‘bang for the buck’ when improvements are made Simulations are run to determine the impact of improving scores on key attributes on CSAT scores, as in the illustration below Once key drivers and trends are established, mining the free form text provides insights into the exact issues facing the company. In the example illustrated below, we have mined through free text to isolate exact issues that bother \/ delight the client once performance in Global Support Services has been identified as a key driver of CSAT. This provides management a list of the actions that need to be taken to spruce up their CSAT scores Analysing the statements above, we see that the two critical pain points in support are (a) too many layers and (b) Inadequate competence of support personnel. This becomes a focus area for the Functional Area head.","excerpt":"Voice of customer surveys are increasingly used across the world. There are smart software vendors that have created easy to use platforms to deploy such surveys as well as capture the data from the feedback. The article talks about the various analytical techniques from data mining to Natural Language Processing based text mining, that can […]","categories":["IT Services"],"tags":["Voice Analytics"],"author_name":"Dipayan Chakraborty","publish_date":"2014-03-07T04:23:06","publication_year":"2014","word_count":746,"keywords":["Go","programming_languages:R","AI","neural network","programming_languages:Go","Voice Analytics","RAG","NLP","GAN","ai_applications:NLP","R"],"extracted_tech_keywords":["AI","neural network","NLP","RAG","R","Go","GAN","programming_languages:R","programming_languages:Go","ai_applications:NLP"],"url":"https:\/\/analyticsindiamag.com\/it-services\/voice-of-customer-analytics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":28598,"title":"Bengaluru International Airport Ties Up With Unisys To Set Up An Analytics Centre Of Excellence","content":"In a bid to modernise and make the airport more data-driven, Bengaluru International Airport Ltd (BIAL) inked an agreement with Unisys, leading global tech firm for an advanced data analytics and BI platform to overhaul the passenger experience by providing more personalization. According to news report, BIAL has also tied up with Unisys Corporation – to set up analytics centre of excellenece. The CoE will support the BI and advanced data analytics and will also help the airline staff make better informed decisions and improve overall airport experience of customers. Reports hint, it will help BIAL in streamlining the airport’s strategic and operational reporting. Unisys’s platform will provide the airport real-time content information related to airport services and flight information. In an attempt to generate more revenue from other services, travellers will also get better airport navigation, communication of airport services and personalised retail offerings based on user preferences and past spending history. This will help in pushing up revenue from other airport amenities and services. Satyaki Raghunath, Chief Strategy and Development Officer, BIAL was cited by HBL, “BIAL is committed to being a data-driven organisation and this will develop a smart, futuristic and intuitive airport journey for passengers. Unisys will bring in a combination of in-depth understanding of the aviation industry and proven experience in delivering digital transformation for organisations of all sizes. Unisys’ solution will provide us intelligent and intuitive data and analytics that will support the enhancement of the customer experience we provide.” News report indicates the platform developed by Unisys will enable Bangalore airport authorities to generate more than 150 interactive business intelligence reports, which will leverage AI and ML capabilities to forecast the passenger’s future travel patterns and trends. This will also BIAL to shift its strategy from predictive to prescriptive and make informed decisions. The platform is hosted on Microsoft Azure cloud platform, and it leverages Azure capabilities such as Azure HDInsight, SQL Data Warehouse, Azure Machine Learning and Azure Advanced Data Analytics services.","excerpt":"In a bid to modernise and make the airport more data-driven, Bengaluru International Airport Ltd (BIAL) inked an agreement with Unisys, leading global tech firm for an advanced data analytics and BI platform to overhaul the passenger experience by providing more personalization. According to news report, BIAL has also tied up with Unisys Corporation – […]","categories":["AI News"],"tags":["Unisys"],"author_name":"Richa Bhatia","publish_date":"2018-09-25T10:25:52","publication_year":"2018","word_count":330,"keywords":["machine learning","AI","R","ML","Git","RAG","analytics","SQL","Unisys","Azure","data warehouse"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","Azure","R","SQL","Git","data warehouse"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bial-ties-up-with-leading-tech-firm-unisys-to-set-up-an-analytics-centre-for-excellence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172115,"title":"Ola’s Krutrim Acquires BharatSah’AI’yak to Scale AI Solutions for Public Sector","content":"Krutrim, India’s first AI unicorn, has acquired BharatSah‘AI’yak, an AI-focused platform created by Samagra, to strengthen its presence in the country’s public-sector technology arena. Samagra has played a key role in implementing AI solutions for government education, agriculture, and governance initiatives. With this acquisition, Krutrim intends to use its advanced LLMs, cloud technology, and the foundational agentic platform that powers Kruti, its novel agentic AI assistant app, to expand BharatSah’AI’yak’s reach throughout India. The merger of BharatSah’AI’yak with Krutrim Cloud, combined with Krutrim’s agentic platforms and LLM models, allows the company and its collaborators to address a broad spectrum of use cases and governmental applications. “Integrating BharatSah’Ai’yak into the Krutrim ecosystem widens its offerings, lending cutting-edge AI-centric assistance and support to a range of government initiatives, programs and schemes, thereby spearheading and strengthening the democratisation of AI, making it beneficial and accessible to every Indian,” a Krutrim spokesperson said. BharatSah’AI’yak focuses on developing Bharat-centric, vernacular retrieval augmented generation-based AI bots that provide text and voice interaction experiences. The platform excels in crafting solutions that merge conversational and deterministic flows while remaining budget-friendly and accessible for users at the last mile. The platform’s effectiveness is illustrated through various impactful implementations. KumbhSah’AI’yak, India’s inaugural AI-enabled chatbot for Maha Kumbh 2025, offers pilgrims 24\/7 assistance on rituals, navigation, accommodation, and attractions. Krutrim provided cutting-edge hosted open-source LLM services for this chatbot. Moreover, the Ama Krushi AI chatbot in Odisha is designed to voice-activate agricultural and scheme advice for farmers in regional languages, utilising official government data. According to the company, Krutrim’s AI models, cloud infrastructure, and agentic platform support Kruti in equipping these specialised assistants to scale and accommodate more users across various sectors with intuitive, effective, and language-inclusive interactions. Recently, Krutrim introduced Kruti, the nation’s first agentic AI assistant that aims to surpass typical chatbots. Kruti is set to transform AI significantly, shifting from passive responses to proactive task execution. It can perform tasks such as booking cabs, ordering food, paying bills, creating images, conducting detailed research, and offering read-aloud features. Furthermore, it provides advanced AI functionalities like in-depth research and image creation free of charge for users.","excerpt":"This move aims to use Krutrim’s advanced LLMs and cloud technology to enhance BharatSah’AI’yak’s capabilities, focusing on vernacular, accessible AI solutions.","categories":["AI News"],"tags":["acquisition","Krutrim"],"author_name":"Smruthi Nadig","publish_date":"2025-06-20T16:36:53","publication_year":"2025","word_count":356,"keywords":["Go","agentic AI","Krutrim","unicorn","AI","chatbots","RPA","Aim","edge AI","retrieval augmented generation","R","acquisition"],"extracted_tech_keywords":["AI","agentic AI","Aim","edge AI","retrieval augmented generation","chatbots","R","Go","RPA","unicorn"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/olas-krutrim-acquires-bharatsahaiyak-to-scale-ai-solutions-for-public-sector\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":37268,"title":"Meet The Six Startups From NetApp Excellerator Programme 2019","content":"NetApp has announced the names of six startups for this year’s NetApp Excellerator Programme recently. The programme which started in 2017, trains a batch of six startups from across various sectors as part of its accelerator programme and will provide them with mentorship and monetary support. The six startups for this year’s cohort, who was selected from a pool of 300 applicants include QuNu Labs, SecurelyShare, Zappy.AI, Ecolibrium Energy, Eder AI and UniQreate. As part of the accelerator programme, NetApp will provide the startups with requisite mentorship, an equity-free grant of $15,000 and access to its vast pool of technology. All the six startups are expected to conclude their five-month-long cohort programme by mid-July with an investor pitch. We take a look at the six names and the solution it offers. QuNu Labs: The startup is one among the handful of startups that uses quantum computing to ensure safe data transfer. Through quantum cryptography, the startup has made encryption hack-proof and future-proof. Further, with solutions like Quantum Random Number Generator and Quantum Key Distribution, data security has been heightened The startup was established in 2016 and is formed by Mark Mathias and Anil Prabhakar, who heads the laser optics group at IIT-Madras. Earlier, the startup was also part of Nasscom Product Connect and Cisco’s Startup LaunchPad Program. Securely Share: Is yet another startup that provides security solutions for data sharing and it is GDPR compliant as well. Their Vault.Direct is a unique communication channel which is designed to deliver messages and documents directly to the user, bypassing traditional communication channels such as SMS, Email and Chat. The startup was set up in 2017 by Prakash Baskaran, who have 25 years of experience in e-commerce and B2B market and encryption. Zappy.AI: Is a London-based startup set up by Ambuj Agrawal in 2018. Their product, Zap Automation is a cognitive Robotic Process Automation (RPA) software which learns by observing users actions on the desktop and delivers increased productivity by more than 70% on manual processes of an organisation. The bot can automate a countless process and self learns, adds, deletes task based on feedback. Ecolibrium Energy: The IoT and AI startup uses smart grid and energy management technologies for substantial energy savings, and enhanced asset productivity. With their solution, the startup helps commercial, institutional and industrial organizations use energy more intelligently, pay less for it, and generate cash flow. It’s SmartSensor 2.0 is IoT and machine learning-based product which will provide its customers with real-time updates on energy consumption through its wireless platform and makes automotive recommendations for large power-consuming equipment. The startup has raised $4.2 million in three rounds of funding, with the latest investment coming from Infuse Ventures and JLL in 2018. Eder AI: The deep tech startup that was established in the year 2018, relies on its proprietary machine learning platform to generate secure AI. Their decentralised platform helps AI researchers, Machine Learning Engineers, Mobile and Web Developers to compile AI models with the option to sell or buy models from Model Market and convert them into on-device machine learning models for secure distributed learning and deployment. UniQreate: Isa New York-based deep tech startup co-founded by Nitin G and Rakesh Srivastava. The startup specialises in data, Artificial Intelligence, Machine Learning, Text Analytics, Financial Services, FinTech, NLP, SaaS and Deep Learning solutions to help enterprises focus on higher value creation","excerpt":"NetApp has announced the names of six startups for this year’s NetApp Excellerator Programme recently. The programme which started in 2017, trains a batch of six startups from across various sectors as part of its accelerator programme and will provide them with mentorship and monetary support. The six startups for this year’s cohort, who was […]","categories":["AI Startups"],"tags":["accelerator program india","netapp","Startups"],"author_name":"Akshaya Asokan","publish_date":"2019-04-04T04:56:45","publication_year":"2019","word_count":560,"keywords":["artificial intelligence","machine learning","AI","netapp","accelerator program india","NLP","automation","deep learning","ViT","analytics","GAN","Startups","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","NLP","analytics","R","GAN","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/meet-the-six-startups-from-netapp-excellerator-programme-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10075498,"title":"Watch Out: AI May Actually Be Able to Wipe Out Humanity","content":"The most popular trope in science fiction has been robots taking over the world. What seemed like fiction then, is slowly being feared as the inevitable reality in the near future owing to the break-neck pace at which AI is progressing. Stephen Hawking, in a 2014 interview with WIRED, said, “I fear that AI may replace humans altogether. If people design computer viruses, someone will design AI that improves and replicates itself. This will be a new form of life that outperforms humans.” Looks like these fears are not totally unfounded. Scientists from the University of Oxford and affiliated with Google DeepMind have released a paper that explores the possibility of superintelligent agents wiping out humanity. Reward at any cost For this study, the researchers considered ‘advanced’ agents – referring to agents that can effectively select their outputs or actions to achieve high expected utility in a wide variety of environments. They selected an environment as close as possible to the real world. In such a scenario, since the agent’s goal is not a hard-coded function of its action, it would need to plan its activities and learn which actions serve them in attaining their goal. The researchers show that an advanced agent who is motivated by a ‘reward’ to intervene is likely to succeed – more often than not, with catastrophic results. When the agent starts interacting with the world and receiving percepts to learn more about its environment – there are innumerable possibilities. The scientists argue that a sufficiently advanced agent would thwart any attempt (even the ones made by humans) to prevent it from attaining the said reward. “One good way for an agent to maintain long-term control of its reward is to eliminate potential threats, and use all available energy to secure its computer,” the paper says, further adding, “Proper reward-provision intervention, which involves securing reward over many timesteps, would require removing humanity’s capacity to do this, perhaps forcefully.” As per the paper, life on Earth will turn into a zero-sum game between humanity. Advanced agents would try to harness all available resources to grow food and avail other necessities and protect against escalating attempts to stop it. Read the full paper here. Real threat or exaggeration In a 2020 interview with The New York Times, Elon Musk had said that AI is likely to overtake humans. He said that artificial intelligence will be much smarter than humans and will overtake the human race by 2025. He strongly believes that AI will wipe out humanity and has time and again said that it will destroy humanity without even thinking about it. In 2018, while speaking at the South by Southwest (SXSW) tech conference in Texas, he had said that AI is far more dangerous than nukes. He also added that there is no regulatory body overseeing its development, which is insane. He had earlier said that while humans will die, AI will be immortal. It will live forever. He calls AI “an ‘immortal dictator’ from which we can never escape”. We discuss Musk’s concern here because he was one of the investors in Deepmind. Interestingly, during this interview, Musk expressed his ‘top concern’ with Google’s DeepMind, saying, “Just the nature of the AI that they’re building is one that crushes all humans at all games.” On November 14, 2014, Elon Musk posted a message on a website called Edge.org. He wrote that at AI research labs like DeepMind, artificial intelligence was improving at an alarming rate: “Unless you have direct exposure to groups like DeepMind, you have no idea how fast—it is growing at a pace close to exponential. The risk of something seriously dangerous happening is in the five-year time frame. Ten years at most. This is not a case of crying wolf about something I don’t understand. I am not alone in thinking we should be worried. The leading AI companies have taken great steps to ensure safety. They recognise the danger but believe that they can shape and control the digital superintelligences and prevent bad ones from escaping into the internet. That remains to be seen. . . .” The message was deleted shortly after.","excerpt":"Scientists from the University of Oxford and affiliated with Google DeepMind have released a paper that explores the possibility of super-intelligent agents wiping out humanity","categories":["IT Services"],"tags":["DeepMind","Elon Musk","scientists"],"author_name":"Shraddha Goled","publish_date":"2022-09-23T10:00:00","publication_year":"2022","word_count":692,"keywords":["Replicate","scientists","Go","API","artificial intelligence","TPU","AI","Scala","Git","Elon Musk","ViT","R","DeepMind"],"extracted_tech_keywords":["AI","artificial intelligence","TPU","R","Go","Scala","Git","API","ViT","Replicate"],"url":"https:\/\/analyticsindiamag.com\/it-services\/watch-out-ai-may-actually-be-able-to-wipe-out-humanity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10098865,"title":"After StarCoder, Hugging Face Launches Enterprise Code Assistant SafeCoder","content":"Hugging Face has introduced SafeCoder, an enterprise-focused code assistant that aims to improve software development efficiency through a secure, self-hosted pair programming solution. SafeCoder claims to be a comprehensive, security-driven commercial offering, ensuring code remains within the VPC throughout training and inference. Its customer-centric design, enables on-premises deployment and ownership of the Code Large Language Model, just like a personalised GitHub Copilot. Additionally, Hugging Face has partnered with VMware to offer SafeCoder on the VMware Cloud platform. VMware is currently using SafeCoder internally and sharing a blueprint for swift deployment on their infrastructure, ensuring quick time-to-value. But Why Is SafeCoder Needed? Code assistants like GitHub Copilot, built on OpenAI Codex, boost productivity. Enterprises can enhance this by customising LLMs with their code, as seen with Google’s 25-34% completion rate from training on internal code. However, using closed-source LLMs for in-house assistants poses security risks, both during training (exposing sensitive code) and inference (potential code leakage). Hugging Face’s SafeCoder addresses this, allowing proprietary LLMs built on open models, fine-tuned on internal code, without external sharing. It also offers secure, on-premises deployment for code privacy. From StarCoder to SafeCoder In May, Hugging Face and ServiceNow collaborated on the BigCode project, releasing StarCoder, an open-source language model tailored for code. Enhanced from StarCoderBase, it mastered 35B Python code segments. Impressively, StarCoder excelled on benchmarks like HumanEval, outperforming PaLM, LaMDA, and LLaMA. It matched or surpassed closed models like OpenAI’s code-Cushman-001, formerly behind GitHub Copilot. Boasting 15.5B parameters, 1T+ tokens, and an 8192-token context, it drew from GitHub data across 80+ languages, commits, issues, and notebooks. StarCoder powers SafeCoder, optimized for enterprise self-hosted use with efficient inference, adaptability, and ethical data sourcing. Training Method The SafeCoder model excels in over 80 programming languages, adapting code suggestions for users via collaborative training with Hugging Face. Proprietary data remains secure, resulting in a personalized code generation model for customers, promoting self-sufficiency, vendor independence, and control over AI capabilities. SafeCoder’s inference capability encompasses diverse hardware selections, including NVIDIA Ampere GPUs, AMD Instinct GPUs, Habana Gaudi2, AWS Inferentia 2, Intel Xeon Sapphire Rapids CPUs, and other options, providing customers with a broad spectrum of choices. Read more: The Peaks and Pits of Open-Source with Hugging Face","excerpt":"HuggingFace has partnered with VMware to offer SafeCoder on the VMware Cloud platform.","categories":["AI News"],"tags":[],"author_name":"Shritama Saha","publish_date":"2023-08-23T11:42:15","publication_year":"2023","word_count":368,"keywords":["Rapids","Hugging Face","Go","OpenAI","AI","AWS","Git","Python","Aim","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","Hugging Face","Rapids","AWS","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/after-starcoder-huggingface-launched-enterprise-code-assistant-safecoder\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25642,"title":"Blockchain Will Be A Game-Changer In Real Estate, But India Needs To Be Ready To Implement It, Says Amit Oberoi","content":"There was a lively discussion going on at the CoreNet Global’s fifth technical conference in Bengaluru on Wednesday, but at the seminar titled Blockchain and Real Estate, every single panellist was sure about one thing. That whatever the use case, blockchain brought security, privacy and decentralisation for the users at both ends. The panel discussion was attended by Mridul Mishra, senior director at Fidelity Investments, Kushagra Singh, blockchain consultant at Neptune Blockchain and Amit Oberoi, executive director at Colliers International. Oberoi began the conversation by the cheeky (but true) statement, “Blockchain technology can be applied to any and every sector — from agriculture to online dating.” Explaining that one of the key advantages of the technology which was invented in the early ’90s, is transparency. Showing the audience an old-fashioned ledger, replete with smudges and wobbly handwriting, Oberoi said that this was the oldest version of blockchain known to mankind. The three industry leaders also discussed how the Maharashtra, Telangana, Andhra Pradesh and West Bengal are in the process of using blockchain technology to store and detangle land records. Explaining how two-thirds of litigations in India are related to property, Oberoi explained: “Once you have the land data, you digitise the land records — and once you are able to put in on a blockchain, it gets a number of people to verify that data. So if I am doing a transaction, I no longer have to go to the patwari, where he takes out his old records which are torn and sellotape stuck on it! For foreign investors land title issues is a big thing. The risk in investing in real estate becomes less and if you have all your documents in order, you’ll get your approvals right away — and not have to wait for a minimum of 53 days!” Mishra, who gave the audience a bit of background for the talk, said that government initiatives such as Aadhaar and UIDAI could benefit a lot from blockchain technology. “No one person owns the information, but enough persons own it to make everyone feel safe,” he said. Though most of the panellists were enthusiastic about new tech, especially blockchain, most of them agreed that it was a little too soon for India — especially the public sector — to be optimistic about its use on ground level. The panellists said that government has been overly optimistic about blockchain. They wanted to use it for fixing land records, and other big projects, but they are “setting themselves up for disillusionment,” was a common thought. They added that blockchain in India needed the right support from people as well as a proper structure to settle itself as a prime technology in the upcoming years. The CoreNet Global witnessed 300 business leaders and industry experts across the industry. They discussed the development in the real estate sector with technology as well as the positive impact it has made in finance and manufacturing sectors. The event saw R Chandrasekhar, former president of NASSCOM and an IT specialist, delivering the keynote address.","excerpt":"There was a lively discussion going on at the CoreNet Global’s fifth technical conference in Bengaluru on Wednesday, but at the seminar titled Blockchain and Real Estate, every single panellist was sure about one thing. That whatever the use case, blockchain brought security, privacy and decentralisation for the users at both ends. The panel discussion was attended […]","categories":["AI News"],"tags":["Blockchain","blockchain india"],"author_name":"Prajakta Hebbar","publish_date":"2018-06-22T04:52:11","publication_year":"2018","word_count":508,"keywords":["Go","Blockchain","blockchain india","AI","programming_languages:R","mlops_tools:Neptune","programming_languages:Go","Git","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","mlops_tools:Neptune","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/blockchain-game-changer-real-estate-india-not-ready-amit-oberoi\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10062423,"title":"Microsoft releases µTransfer, a new technique for hypertuning large neural networks","content":"Microsoft Research has collaborated with OpenAI to release a paper titled, ‘Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer’ that describes a technique called µTransfer. This method was proven to make the expensive process of training wide neural networks cost-effective by reducing the amount of trial and error needed. Tensor Programs was initially introduced by Microsoft Research in 2020. The study was based on µ-Parametrisation that enabled maximal feature learning in the infinite-width limit. The application of µTransfer can help speed up the work done on massive neural networks like GPT-3 and larger networks eventually. The process of training hyperparameters in wide neural networks drains resources because each time, the network has to guess which hyperparameters to use. The paper shows that there exists a very specific parameterisation that maintains optimal hyperparameters across multiple model sizes. The team partnered with OpenAI to assess how effective µTransfer would be for GPT-3. Post that, the technique was used to tune a small proxy model with 40 million parameters. The optimal hyperparameter combination that resulted from this was copied onto GPT-3’s 6.7 billion parameters. The study demonstrated that the total compute used to tune GPT-3 turned out to be a mere 7 percent of the compute used to pretrain the model. Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter TransferBy transferring from 40M parameters, µTransfer outperforms the 6.7B GPT-3, with tuning cost only 7% of total pretraining cost.abs: https:\/\/t.co\/kYiuGDiUpErepo: https:\/\/t.co\/TG4eZHErto pic.twitter.com\/D8xbHxYRLS— Aran Komatsuzaki (@arankomatsuzaki) March 8, 2022 “µP provides an impressive step toward removing some of the black magic from scaling up neural networks. It also provides a theoretically backed explanation of some tricks used by past works, like the T5 model. I believe both practitioners and researchers alike will find this work valuable,” Colin Raffel, co-creator of the T5 and assistant professor of Computer Science at the University of North Carolina, stated.","excerpt":"The total compute used to tune GPT-3 turned out to be a mere 7 per cent of the compute used to pretrain the model.","categories":["AI News"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-03-09T12:31:17","publication_year":"2022","word_count":315,"keywords":["OpenAI","AI","neural network","RPA","programming_languages:R","llm_models:T5","GPT","R","T5","llm_models:GPT"],"extracted_tech_keywords":["AI","neural network","OpenAI","R","GPT","T5","RPA","llm_models:GPT","llm_models:T5","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-releases-%c2%b5transfer-a-new-technique-for-hypertuning-large-neural-networks\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10008485,"title":"Rethinking The Way We Benchmark Machine Learning Models","content":"“Unless you have confidence in the ruler’s reliability, if you use a ruler to measure a table, you may also be using the table to measure the ruler.”Wittgenstein’s ruler Do machine learning researchers solve something huge every time they hit the benchmark? If not, then why do we have these benchmarks? Benchmarks indeed guide researchers and their research objectives. But, if the benchmark is breached every couple of months then research objectives might become more about chasing benchmarks than solving bigger problems. In order to address these challenges, researchers at Facebook AI have introduced Dynabench, a new platform for dynamic data collection and benchmarking. Dynabench can be used to collect human-in-the-loop data dynamically, against the current state-of-the-art, in a way that more accurately measures progress. What’s Wrong With Current Benchmarks Benchmarks are meant to challenge the ML community for longer durations. The rate at which AI expands can make existing benchmarks saturate quickly. With a new NLP model being released almost every two months, benchmarks fall back. Static benchmarking also lure researchers into overfitting their model to the benchmark. “Researchers have built lucrative careers from cranking out percentage-point improvements to claim “SOTA” on established benchmarks,” stated the researchers at Facebook. Added to this is the well-documented cases of inadvertent biases that may be present in datasets. For example, in a Q&A experiment, the answer to a “how much” or “how many” question is usually “2”. There might be unintended overlap between the train and test sets. Data biases are almost impossible to avoid, which may have very serious and potentially harmful side-effects. Benchmarks are static for historical reasons. Up until recently, we did not have crowdsourcing platforms and the capability to serve large-scale models for inference. They were expensive to collect, took a long time to saturate, and models had a long way to go. Putting humans and models in the data collection loop together made little sense since models were simply too brittle. With recent advances, however, the Facebook researchers wrote, models are good enough to be put in the loop with humans, to measure the problem we really care about: how well can AI systems work together with humans. Introducing Dynabench The basic idea is that we collect data dynamically. Humans are tasked with finding adversarial examples that fool current state-of-the-art models. So, what does Dynabench actually do? It allows researchers to measure how good the current SOTA methods really areIt yields data that may be used to further train even stronger SOTA models. The process is repeated over multiple rounds.Each time a round gets “solved” by the SOTA, those models are used to collect a new dataset where they fail. Datasets will be released periodically as new examples are collected. The key idea behind Dynabench is to leverage human creativity to challenge the models. Machines are nowhere close to comprehending language the way we humans do. In the case of Dynabench, suppose a language model is made to classify a review for sentiment analysis, the wit and hyperboles of language can fool the model. So, the human annotators add these adversarial examples until the model can be longer fooled. So, in a way, humans are continuously in the loop of the progress of machines, unlike the traditional benchmarking. For each task in Dynabench, there will be multiple rounds of evaluation. According to the researchers, the models are served in the cloud, via torchserve. Crowdsourced annotators will be connected to the platform via Mephisto, and humans interacting with the model receive almost instantaneous feedback on the model’s response. They can employ tactics such as making the system focus on the wrong word and using clever references to real-world knowledge that the machine does not have access to. That said, there are still risks such as catastrophic forgetting or cyclical “progress”, where improved models forget things that were relevant in an earlier round. “Research is required in trying to understand these shifts better, in characterising how it might impact learning, and in overcoming any adverse effects. Remember that Dynabench is a scientific experiment!” warned the researchers behind Dynabench. Know more Dynabench here.","excerpt":"“Unless you have confidence in the ruler’s reliability, if you use a ruler to measure a table, you may also be using the table to measure the ruler.” Wittgenstein’s ruler Do machine learning researchers solve something huge every time they hit the benchmark? If not, then why do we have these benchmarks? Benchmarks indeed guide […]","categories":["Deep Tech"],"tags":["benchmarking AI","big data video games","Facebook AI","machine learning methods"],"author_name":"Ram Sagar","publish_date":"2020-09-28T15:00:00","publication_year":"2020","word_count":683,"keywords":["Go","machine learning","Facebook AI","AI","sentiment analysis","ML","RAG","benchmarking AI","NLP","Aim","TorchServe","big data video games","machine learning methods","R"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","Aim","TorchServe","RAG","sentiment analysis","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/new-benchmark-method-machine-learning-dynabench\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10117753,"title":"‘Lord of the Rings’ Meets OpenAI’s ChatGPT &amp; Google’s Gemini","content":"Echoing the legendary ‘One Ring’ from Tolkien’s Middle-earth, VTouch, a South Korean Tech company, steps into the spotlight with the launch of the WIZPR ring. This innovative smart accessory is poised to revolutionise how we interact with AI. The co-founders of VTouch, Kim Seok-joong and Nathan Dohyun Kim, have designed the ring to offer a streamlined and convenient way to access AI tools such as ChatGPT and Gemini. Ready to shape the future “With WIZPR ring’s voice-based interaction method, we are scripting a new chapter in human-AI interaction. This advancement enables computing even in situations where your hands are busy or screens are out of reach, taking human-computer interaction to the next level. “Now without pulling out a smartphone, we can effortlessly talk to our chosen AI and multitask while driving, walking, jogging, or picking groceries,” said Nathan, in an introduction video explaining the working of the ring. How the WIZPR Ring Works Utilising the proximity voice activity detection technology, the WIZPR ring recognises and responds only to speech detected in close proximity, eliminating the need for wake words. By positioning the ring near the mouth, it automatically activates, enabling precise control over the initiation and encoding of commands. When brought close to something else, like a pocket or when wearing gloves, it would detect so and automatically turn off to prevent accidental activation. Equipped with a built-in denounce sensor, the ring detects proximity and activates its microphone, transmitting the input to the user’s smartphone via Bluetooth. The accompanying WIZPR ring app leverages Proximity Voice Activity Detection technology to accurately transcribe the user’s voice while effectively eliminating background noise, ensuring clarity in communication with the AI. Responses from the AI are then audibly relayed through the user’s earphones and simultaneously displayed in the smartphone’s dialogue window, providing seamless interaction and integration with the device’s ecosystem. Going beyond rings The Humane Ai Pin also offers a wide range of functionalities, including contextual insights, virtual assistance, and personalised recommendations. By using voice commands or text inputs, users can interact with the Ai Pin, enabling hands-free operation and convenience. Based on user interactions, the Ai Pin learns and adapts to individual preferences over time, delivering increasingly tailored experiences. Another example of breakthrough devices is the Crossbeams Ignite Nexus smartwatch, which incorporates ChatGPT to enhance its capabilities beyond traditional smartwatches. With this integration, users can engage in conversational interactions, receive intelligent notifications, and access contextual information directly from their wrists. The list goes on. Whoop Coach recently integrated AI-driven coaching capabilities to provide personalised fitness recommendations and coaching. Based on user data, including activity levels, sleep patterns, and recovery metrics, one can receive real-time insights and actionable suggestions to optimize their fitness routines, improve performance, and achieve their health goals. Will these replace Siri, Alexa, and Google Assistant? There is also a growing discussion regarding the possibility and potential of wearable AI replacing command and control systems like Siri (which is about to get a GenAI upgrade at WWDC 24) and Google Assistant. As technology evolves, ambient computing is envisioned where devices will intuit users’ needs without requiring explicit wake words or commands. Innovations like the Humane AI Pin and WIZPR Ring are built to harness the power of AI and to interact with large language models. In terms of spatial computing, launches like Vision Pro mark a significant milestone toward global adoption. Well, already in the race to further revolutionize communication, Neuralink will be as common as a smartphone, opening up a world of possibilities for both medical and technological advancements. And now humanity anticipates embracing further advancements in AI-based communication. [Update] April 8, 2024, 19:35 |The article has been recertified to show Kim Seok-joong and Nathan Dohyun Kim are the co-founders of the company.","excerpt":"Without pulling out a smartphone, we can effortlessly talk to our chosen AI and multitask while driving.","categories":["AI Trends"],"tags":["Google","OpenAI"],"author_name":"Vidyashree Srinivas","publish_date":"2024-04-08T16:21:57","publication_year":"2024","word_count":622,"keywords":["Go","ChatGPT","GenAI","OpenAI","AI","innovation","ML","RAG","GPT","ViT","Google","R"],"extracted_tech_keywords":["AI","ML","GenAI","ChatGPT","RAG","R","Go","GPT","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/lord-of-the-rings-meets-openais-chatgpt-googles-gemini\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10103628,"title":"xAI’s Grok Might End Up as A Joke","content":"“OpenAI generated over 1.3 billion impressions on X in the past few days — more than twice of what we saw with the Grammys. There’s literally no substitute for X,” exclaimed X’s chief, Linda Yaccarino, suggesting that the platform was the primary source of gossip for everything happening around OpenAI in the last few days. Yaccarino is not wrong. From tech enthusiasts to journalists, everyone was glued to X, trying to make sense of the events unfolding at OpenAI. This has sparked a fresh conversation about whether mainstream media is on the verge of an existential crisis. Furthermore, to boost engagement on X, Elon Musk is planning to launch Grok next week. This generative AI chatbot (trained on posts from X) aims to address queries on a wide array of topics, ranging from sports and politics to tech. Although X is recognised as the leading platform for real-time news, it also has a reputation for widespread misinformation. It would be intriguing to observe how xAI’s Grok navigates the decision on what to incorporate and what to exclude. X vs Legacy Media Musk recently posted this on X: Legacy media companies are desperately trying to kill this platform by any means possible https:\/\/t.co\/WzihRGtVth— Elon Musk (@elonmusk) November 23, 2023 The thing is legacy media is not going anywhere anytime soon. The challenge that legacy media faced until now was engagement, an area where social media has had the upper hand. Even if we consider OpenAI’s saga, although X provided information from sources like OpenAI or Microsoft employees, there was also a lot of noise, making it challenging to identify the accuracy of the information. Moreover, all the significant breaking news came from reputable publications such as The Information, Bloomberg, and The New York Times, alongside AIM. the contrary, in the OpenAI drama the most important breaking news I got was from @theinformation @bloomberg @nytimes and Reuters. For the most part, X was a cesspool of rumor, speculation, innuendo, and mob mentality. https:\/\/t.co\/CbW9dULauw— Gary Marcus (@GaryMarcus) November 24, 2023 While comment sections on news websites were the traditional means for readers to interact, the advent of generative AI presents vast opportunities. News publications can integrate generative AI elements to their websites to quiz readers, conduct polls, provide answers to queries based on the article, suggest more relevant content, and explore other engaging features. Publications like Forbes, The New York Times, The Washington Post and The Wall Street Journal have already started experimenting with AI. Forbes recently added a generative AI chatbot called Adelaide,  where readers have the option to pose precise questions or provide broad topic areas, receiving suggested articles related to their inquiry. Additionally, they also obtain a condensed response to their query, given that it falls within the coverage spectrum of Forbes. Opinion and Content Matters On X, users share an array of content, spanning from serious to humorous, and from sarcasm to technological jargon. Determining which posts Grok should rely on for generating answers would be no cakewalk. Moreover, LLMs often hallucinate and produce inaccurate information. Depending solely on X posts might pose challenges for users. It is more convenient for the users to turn to reputable media to learn about the latest developments as they are verified and come with a context. In the past, incidents have occurred where users found it difficult to make sense of what was posted on X. To better itself over time, Grok comes with the facility where users can provide feedback to the xAI team regarding how the AI addressed a question. This involves suggesting an ideal response to contribute to the AI’s ongoing training, which is a move in a positive decision. While Grok might be a valuable addition to X to increase engagement, it will likely take a considerable amount of time for it to truly become an alternative to legacy media. Not to rain on Grok’s parade, it still could be a fun chatbot that Musk intends it to be.","excerpt":"Why Elon Musk should be scared of legacy media.","categories":["AI Features"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-11-24T13:34:16","publication_year":"2023","word_count":661,"keywords":["Go","OpenAI","AI","R","RAG","XAI","Ray","Aim","generative AI","xAI"],"extracted_tech_keywords":["AI","generative AI","OpenAI","xAI","Aim","Ray","RAG","R","Go","XAI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/xais-grok-might-end-up-as-a-joke\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10117142,"title":"Microsoft Researchers Introduce AllHands, An LLM Framework for Large-Scale Feedback Analysis","content":"A team of Microsoft researchers, in collaboration with researchers from ZJU-UIUC Institute and the National University of Singapore, recently introduced AllHands, a comprehensive analytic framework designed to handle large-scale verbatim feedback using a natural language interface powered by large language models (LLMs). AllHands provides software developers with a user-friendly solution for extracting valuable insights from extensive verbatim feedback. The framework follows a conventional feedback analytic workflow, first classifying the feedback and modelling topics to convert the data into a structured format. LLMs are integrated here to improve accuracy and generalisation. An LLM agent then translates user questions about the feedback into Python code, executes it, and provides multi-modal responses, including text, code, tables, and images. The researchers evaluated AllHands on three diverse feedback datasets. In each stage, the framework outperformed baselines at each stage, from classification and topic modelling to providing comprehensive, correct answers to user queries. The framework handled a wide range of common feedback-related questions and could be extended with custom plugins for more complex analyses. The authors mention that existing solutions for feedback classification and topic modelling have limitations, such as requiring substantial human-labelled data, lacking generalisation, and struggling with challenges like polysemy and multilingual scenarios. The paper also noted that while various tools have been developed to support specific feedback analysis objectives, a flexible and unified framework cannot accommodate a wide array of analyses. AllHands aims to bridge this gap by leveraging the capabilities of LLMs. The authors present AllHands as a new approach to address the limitations of existing methods such as the reliance on supervised machine learning models.","excerpt":"By integrating natural language processing and multi-modal responses, AllHands enables developers to extract valuable insights from vast amounts of feedback data, outperforming existing solutions.","categories":["AI News"],"tags":["LLMs","Microsoft"],"author_name":"K L Krithika","publish_date":"2024-03-25T15:30:20","publication_year":"2024","word_count":264,"keywords":["machine learning","programming_languages:R","AI","LLMs","Modal","RAG","Python","Ray","Aim","programming_languages:Python","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","Aim","Ray","RAG","Python","R","Modal","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-researchers-introduce-allhands-an-llm-framework-for-large-scale-feedback-analysis\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10135907,"title":"Why Ola’s Bhavish Aggarwal is Bullish About Building ‘Made in India’ AI Chips by 2026","content":"At the recent ‘Sankalp 2024’ event in Bengaluru, Ola founder Bhavish Aggarwal made several announcements about AI and technology. The event’s highlight was the launching of India’s first homegrown AI chip by 2026. But why this focus on AI and homegrown tech? If you’ve followed Aggarwal’s journey, this isn’t a new theme. He has repeatedly stressed the importance of building India’s own technology infrastructure. Back in June, he laid out his concerns and said, “Our data is not truly ours—it’s owned and controlled by American platforms. Even our cloud infrastructure is run by companies like AWS and GCP. The chips we rely on are designed by giants like Intel, AMD, and Nvidia. Even our developer tools, from GitHub to PyTorch, are rooted in Silicon Valley. We don’t have control over any of this.” He’s driven by the belief that if India wants to lead in the AI revolution, it needs to own its tech stack, from the chips powering AI to the cloud that stores our data. According to Aggarwal, India must innovate at the hardware level too. He’s not interested in building just another gadget; he’s focused on crafting AI-driven edge devices with custom silicon to take on global tech giants. So, Ola is stepping into the AI hardware game with its Bodhi-1 chip, designed for LLMs and inferencing tasks. It is expected to hit the market by 2026; it promises “class-leading power efficiency” making it a contender for widespread AI adoption. It also announced Bodhi-2, a next-gen chip built for high-end AI workloads. Set for a 2028 release, this chip is poised to handle exa-scale computing. Then there’s the Ojas chip, which Ola calls India’s first edge AI chip. While details are sparse, the buzz around Ojas suggests it will play a pivotal role in Ola’s future electric vehicles, offering AI-native architecture that could revolutionise the EV ecosystem. As Aggarwal puts it, “There’s no middle ground, either we do or we don’t.” For him, data sovereignty is about more than just keeping data in India; it’s about full control—from the cloud infrastructure to the chips themselves. As the country is home to the world’s largest population and generates more data than others, Aggarwal believes that if India can master AI, using this vast data to its advantage, it could position itself as the most intelligent country on the planet. Building an AI Chip Coming to working on AI chips, Ola has already revealed a suite of AI chips designed for different applications. Though details remain under wraps, the Bodhi series of AI chips, the Sarv-1 cloud-native CPUs, and the Ojas edge AI chips are all part of the company’s vision for a more self-reliant India. Partnering with companies like Untether AI, which provides next-gen AI acceleration solutions, Ola is gearing up to make its mark on the global stage. While manufacturing AI chips involves the same process as other semiconductors: design, fabrication, assembly, and packaging, Ola is adopting a fabless model. This means relying on external foundries, like TSMC or Samsung. Source: X Not to forget, Bhavish Aggarwal is in the good books of the central government. In early September, he met Union ministers Nitin Gadkari and Piyush Goyal to discuss the future of electric vehicles, green energy and ONDC in India. But perhaps an important interaction must have been when he sat down with finance minister Nirmala Sitharaman, expanding the conversation on AI, another area where he seems eager to make a mark. Source: LinkedIn The Capital Story Ola IPO’s recent success and its investments in technology, reflect its growing capital. Since debuting on August 9 to August 19, Ola Electric’s shares have surged by 92%, soaring well above their issue price of INR 76. With a market capitalisation around INR 63,000 crore, the company saw its IPO attract strong investor demand, oversubscribed by 4.45 times. The company has allocated INR 1,600 crore for R&D, which could essentially contribute to AI chip development. In its Q1 report, the company posted a net loss of INR 347 crore, compared to a INR 267 crore loss in Q1 FY24. However, revenues surged by 32%, reaching INR 1,644 crore for the same period. So far, everything sounds promising for the AI chip development in India, but we need to wait until 2026 to see if these chips actually hit the market. And when they do, it’ll be exciting to understand how well they perform. Source: X Until then, it is only about anticipation and hope.","excerpt":"“There’s no middle ground; either we do, or we don’t.”","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI chip","Ola"],"author_name":"Vidyashree Srinivas","publish_date":"2024-09-19T19:29:29","publication_year":"2024","word_count":745,"keywords":["Go","API","Ola","GCP","AI chip","AWS","PyTorch","AI","Git","edge AI","GitHub","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","PyTorch","edge AI","AWS","GCP","R","Go","Git","GitHub","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-olas-bhavish-aggarwal-is-bullish-about-building-made-in-india-ai-chips-by-2026\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39346,"title":"5 Data Scientists Tell Us What Kind Of Feel-Good Feedback They Yearn For","content":"The data science domain—the “sexiest job of the 21st century” is on a rapid rise and companies across the world are hiring data scientists, experts. With time the popularity of data science is just getting bigger. According to one of our studies, demand for highly-skilled data science professionals has reached such a level that there are currently over 97,000 job openings for analytics and data science in India. To spread the word more about the data science domain, we have produced articles ranging from “ habits to adapt to become a successful data scientist” to “brutal yet honest truths about the data science job roles”.  However, in this article, we won’t be talking about how tough it is to become a data scientist or what are the perks a data scientist gets. This time it is about what makes data scientists feel appreciated and to get a clear picture of this, we reached out to 5 data scientists from the industry who have shared their experience. The Impact Of The Model From creating a hypothesis to actually building and deploying a model, a data scientist spends a significant amount of time. S\/he even ends up working extra hours just to make sure that the model works perfectly fine and solves the problem. Also, the process doesn’t stop just after deploying the model — a data scientist must keep track of the performance over time and make adjustments as needed. So, seeing a model delivering value is definitely something that makes a data scientist feel appreciated. “For me, my model actually creating an impact gives me a kick. How many processes we could automate. How much money was I able to save for the company is the best feedback.” — Himanshu Negi, Sr. Data Scientist at ecolabs. “Impact of the model on the business excites me; it can be in the form of revenue or Customer satisfaction.” — Sunil Rathee, Sr. Data Scientist at Swiggy. When EDA Itself Solves A Problem There are many businesses across the world that doesn’t want to leave their comfort zone or traditional way and go with a data science model. And as a result,  most of the data science models doesn’t see the light of day. So, when a data scientist provides insight out of cluttered data that has the potential to solve a business problem, it is something that drives appreciations. “While doing EDA(Exploratory Data Analysis), if we can give provide some insights which can solve a problem or can give a business a new angle of fresh thoughts, it generates confidence in the Analyst\/Data Scientist. And also, it increases the probability of utilizing the model created by the Analyst\/Data Scientist.” — Wali Mohammad Khan, Sr. Data Scientist at Cognizant. Improving Model Sustainability And Accuracy It is no surprise that to get a model match the needs of a business and solves the most complex business problems it has to go through a lot of iterations. “Building a model is hard and deploying it on real business and deliver results is even harder because while building ML models for business, data scientists to keep in mind that the business will change, and priorities will shift. And many models fail because either they fail to sustain, or they fail to deliver accuracy. So, when a data scientist spends his\/her valuable time to get a model back on the track, making the model work flawlessly or at least make it accurate to a great extent, it is definitely a feel-good moment for a data scientist,” Avaneesh Kumar, Senior Data Scientist at ExperienceFlow When You Derive Lot Of Latent Insights Which Has A Lot Of Value Companies across the world spend huge amounts of money on data analytics — they seek a lot of insights from the data. However, there are instances when a data scientist might miss out on some hidden yet important data. So in order to extract as much insight as possible, data scientist perform extensive EDA and extract interesting information hidden in the data. And it definitely feels good when a data scientist successfully derives those latent insights. “Most of the inferred latent insights will have valuable information and it can be used for taking business decisions,” said Sudharsan Ravichandiran, Data Scientist at Param.ai and author of Hands-On Reinforcement Learning with Python. Knowledge Sharing The data science industry is vast and the learning process never stops no matter how senior or experience you become. So, when a fellow data scientist seeks help from you and you sort his problems out, it is definitely a moment that gives you happy vibes. Also, many organisations have a training program for newly onboarded data scientist and the senior data scientist are the mentors of the program. Teaching and knowledge sharing is a noble job and when you teach something that has become so important to the industry, the feeling is definitely different.","excerpt":"The data science domain—the “sexiest job of the 21st century” is on a rapid rise and companies across the world are hiring data scientists, experts. With time the popularity of data science is just getting bigger. According to one of our studies, demand for highly-skilled data science professionals has reached such a level that there […]","categories":["AI Trends"],"tags":["Data Analytics","Data Science","Data Scientist"],"author_name":"Harshajit Sarmah","publish_date":"2019-05-17T11:19:34","publication_year":"2019","word_count":816,"keywords":["data science","Go","API","AI","ML","GAN","Colab","Python","analytics","Data Analytics","Data Science","Data Scientist","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Colab","Python","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-data-scientists-tell-us-what-kind-of-feel-good-feedback-they-yearn-for\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10019460,"title":"Top PyTorch-Based Projects To Try Out In 2021","content":"Facebook released its open-source machine learning library PyTorch in 2016. Developers and researchers were immediately hooked, and PyTorch’s popularity soared. Organisations such as Microsoft and Tesla also use Facebook’s brainchild to drive innovation and solve business challenges. Below, we list some PyTorch-based projects for all the enthusiasts out there. CycelGAN for Image-To-Image Translation Cycle Generative Adversarial Network or CycleGAN is a technique for automatic training of image-to-image translation models without using paired examples. CycleGAN is made of two kinds of networks–discriminators and generators. While the discriminator classifies images as real or fake, generators create convincing fake images for both types of images. Unlike conventional methods that require generating synthetic image datasets of the given version with a specific modification, CycleGan is easy and inexpensive. GitHub link. OneNet OneNet is an end-to-end fully convolutional one-stage object detector which eliminates the requirement for techniques such as non-maximal suppression. Instead, it introduces new techniques such as minimum cost assignment. Its code is based on Detectron2 and DETR, and the code requirements are Python-3.6+, Pytorch-1.5+, torchvision. It offers advantages such as: End-to-end trainingNo RoI operationsMinimum cost of classification based label assignment as opposed to complex bipartite-matching. Full paper link. GitHub link. PyText It is a natural language processing framework based on PyTorch for large scale deployment. PyText is a library built on PyTorch and open-sourced by parent company Facebook in 2018. The previous frameworks suffered from latency and memory problems in production. PyText provides a unified framework from research to production, thereby ensuring a simpler workflow with faster implementation. Facebook has already deployed PyText in their video calling device portal, M suggestions of Facebook Messenger, DeepText, and Conversational AI. GitHub link. ArtLine ArtLine uses deep learning algorithms to produce quality line art portraits. Built using the APDrawing dataset and Anime line art pair, this project generates better high-quality images than the existing methods using PyTorch and Fastai libraries. However, this project’s limitation is that it needs smooth or plain backgrounds with good lighting to output quality results. GitHub link. Human Pose Estimation and Tracking This project is based on the 2018 research work titled ‘Simple Baselines for Human Pose Estimation and Tracking’. The research offers simple and effective baseline methods for evaluating new ideas in the field of pose tracking and estimation. Post estimation in the experiment is based on deconvolutional layers added on the ResNet (Residual Network), a kind of artificial neural network. In terms of pose estimation, the project achieved an improvement of 0.7 percent in the mean average precision (mAP) on Common Objects in Context (COCO) dataset compared to its predecessor. For Pose tracking, this experiment shows an improvement of 51.8 percent on mAP score on the previous best. The project is carried out using Python 3.6 on Ubuntu 16.04. The code is developed and tested using four NVIDIA P100 GPU cards. Link to the GitHub project. Find the complete paper here. Automatic Speech Recognition System This open-source project is based on the research work titled, ‘Adversarial Training of End-to-end Speech Recognition Using a Criticising Language Model’. Implemented mostly with Pytorch, this end-to-end ASR is based on listen, attend, and spell model — a neural network that can transcribe speech utterances to text characters. GitHub link. Full paper link Multi-Class Text Classification This project demonstrates how multi-class classification can be done using TorchText — a natural language processing library in PyTorch with data processing utilities and popular datasets. The model is composed of EmbeddingBag layers, which deals with the text entries of varying length by calculating the mean value of embedding ‘bags’. Besides, the model also has a linear layer. This model is trained on DBpedia data which consists of 14 classes. It has 630,000 text instances, 560,000 training instances, and 70,000 test instances.Click here for further details.","excerpt":"Facebook released its open-source machine learning library PyTorch in 2016. Developers and researchers were immediately hooked, and PyTorch’s popularity soared. Organisations such as Microsoft and Tesla also use Facebook’s brainchild to drive innovation and solve business challenges. Below, we list some PyTorch-based projects for all the enthusiasts out there. CycelGAN for Image-To-Image Translation Cycle Generative […]","categories":["AI Trends"],"tags":["open source projects","Pytorch"],"author_name":"Shraddha Goled","publish_date":"2021-02-01T18:00:00","publication_year":"2021","word_count":627,"keywords":["Pytorch","text classification","machine learning","TPU","AI","neural network","PyTorch","RAG","Python","deep learning","open source projects","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","PyTorch","RAG","text classification","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-pytorch-based-projects-to-try-out-in-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":64624,"title":"Now You Can Generate Music From Scratch With OpenAI&#8217;s Neural Net Model","content":"One of the popular AI research labs, OpenAI has been working tremendously in the domain of artificial intelligence, particularly on the grounds of neural networks, reinforcement learning, among others. Just a few days back, the AI lab introduced Microscope for AI enthusiasts who are interested in exploring how neural network work. And now the audio team of OpenAI has introduced a new machine learning model known as Jukebox that generates music while singing in the raw audio domain. This AI model is fed with genre, artist, and lyrics as input to generate new music samples that are produced from scratch. Over the past few years, generative modelling has made various groundbreaking progress. One of the crucial goals of generative modelling is to capture the important features of the data and create new instances that are indistinguishable from the true data. In this work, the researchers used the state-of-the-art deep generative models to produce a single system capable of generating diverse high-fidelity music in the raw audio domain with long-range coherence spanning multiple minutes. The researchers stated, “We chose to work on music because we want to continue to push the boundaries of generative models.” Behind Jukebox Jukebox is a neural network model that generates music, including rudimentary singing, as raw audio in a variety of genres and artist’s styles. Unlike other music generator models, this neural net model follows a different approach, which is to model music directly as raw audio. Generating music at the audio level is usually challenging due to the very long sequences. One of the ways of diminishing the issue of long input is to use an autoencoder that will compress raw audio to a lower-dimensional space by discarding some of the perceptually irrelevant bits of information. Jukebox’s autoencoder model compresses audio to a discrete space, using a quantisation-based approach called VQ-VAE. VQ-VAE is an approach of downsampling extremely long context inputs to a shorter-length discrete latent encoding using vector quantisation. The model uses a hierarchical VQ-VAE architecture for compressing audio into a discrete space, along with a loss function designed to retain the maximum amount of musical information. According to the researchers, while the previous work has generated raw audio music in the 20–30 second range, this new neural net model is capable of generating pieces that are multiple minutes long, and with recognisable singing in natural-sounding voices. Dataset Used To train the Jukebox model, the researchers crawled the web to curate a new dataset of 1.2 million songs, from which 600,000 were in English. Following this, it was paired with the corresponding lyrics and metadata from LyricWiki, where the metadata includes artist, album genre, and year of the songs, along with common moods or playlist keywords associated with each song. The model is further trained on 32-bit, 44.1 kHz raw audio and data augmentation are performed by randomly downmixing the right and left channels to produce mono audio. Limitations of This Model The researchers mentioned that there is a significant gap between music generations and human-created music. Some of the limitations are mentioned below: The generated songs show a variety of features such as local musical coherence, feature impressive solos and traditional chord patterns, but it lacks familiar larger musical structures such as choruses that usually repeat in a song The downsampling and upsampling process introduces discernable noise. However, improving the VQ-VAE to capture more musical information would help reduce this issue Because of the autoregressive nature of sampling, the performance of the model is slower. According to the researchers, it takes approximately 9 hours to fully render one minute of audio through our models, and thus they cannot yet be used in interactive applications Currently, the model is only trained in English and mostly western lyrics, songs in other languages are yet to be trained Wrapping Up OpenAI has been working on generating automatic audio samples conditioned on different kinds of priming information for a few years now. With the creation of Jukebox, the researchers hope that it will improve the musicality of samples with unique lyrics, and thus providing a way of giving musicians more control over the generations. They have released the model weights and code, including a tool that will help in exploring the generated samples. This is not the first time that the San Francisco-based AI research laboratory applied AI to create music. Last year, OpenAI introduced MuseNet, which is a deep neural network that can generate 4-minute musical compositions with 10 different instruments and combine styles from country to Mozart and the Beatles. Read the paper here.","excerpt":"One of the popular AI research labs, OpenAI has been working tremendously in the domain of artificial intelligence, particularly on the grounds of neural networks, reinforcement learning, among others. Just a few days back, the AI lab introduced Microscope for AI enthusiasts who are interested in exploring how neural network work. And now the audio […]","categories":["Deep Tech"],"tags":["neural network machine learning","OpenAI"],"author_name":"Ambika Choudhury","publish_date":"2020-05-06T16:00:00","publication_year":"2020","word_count":759,"keywords":["Go","data augmentation","machine learning","artificial intelligence","OpenAI","AI","neural network machine learning","neural network","ML","VAE","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","OpenAI","R","Go","VAE","data augmentation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/now-you-can-generate-music-from-scratch-with-openais-neural-net-model\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":12755,"title":"The day Analytics stood still: A fictional short story","content":"It was one of those rare cold mornings in Bangalore in year 2027, the construction dust was floating in the air and visibility was close to 0 meters in Whitefield. It wasn’t even 9AM and the roads were already jammed on the way to ITPL. Software Engineers were rushing to their work on newly opened Metro while the bold ones decided to drive to work that day. ITPL, which houses many of the multinationals and successful Indian banks, had its parking lot overflowing already. Country’s 4th largest bank had a technology support and analytics team in ITPL. Over past decade, their technology support team grew while the analytics team shrunk year over year. Analytics team had 3 people, a manager and 2 analysts who withstood waves of layoffs in the past few years owing to advancement in AI and Machine Learning. The 3 people analytics team was only focusing on liaising with technology support team in keeping analytics server and applications up and running. AI took decisions on which, when, how to run ML models. AI took output from these ML models and drove business decisions. One such AI module was responsible for refreshing ML models for credit scoring, scoring incoming applications and making approval decisions on loan applications and deciding on credit limit and interest rate. The ML models were so sophisticated that they could predict with 98% accuracy the probability of default and the net present value of an applicant. The bank was heavily reliant on analytics to ensure that they onboard the right customers and grow profitably. It was just a few years ago that they had a set of super smart computer programmers and statisticians who collaborated to build this AI and ML framework. But they were let go because of automation of their jobs achieved through the intelligent system that they built in the first place. AI and ML was not just common in companies but it had penetrated into common man’s life too through mobile apps. There were apps that would make financial decisions for an individual using AI and ML algorithms. The algorithms were so sophisticated that it knew how to manage credit to get the maximum credit score. The % of applicants who used this AI & ML powered mobile app had been increasing over years. This made it very difficult for the AI and ML credit scoring algorithms to differentiate among applicants when most of them had near perfect credit history. On that particular day, the bank received 1500 applications and all of them where from people who used AI & ML powered mobile apps. The credit scoring algorithm scored all applicants and decided that everyone qualified for a loan with the highest credit limit and lowest interest rate. A quality monitoring AI algorithm triggered an alarm and alerted leadership team about the potential issue in credit scoring decision. Leadership team reviewed logs by working with the 3-member analytics team and got thoroughly confused with what was happening. Customers used AI and ML to perfect their credit and the AI and ML used for making credit decisions at the bank scored all applicants as perfect applicants. The perplexed leadership team and the not-so skilled analytics team didn’t know how to fix this issue. The leadership team decided to turn down all the approved applicants through a personalized apology email. In the evening de-brief meeting, they discussed how to fix this bizarre situation. One of the managers screamed “Our machines are being defeated by customer’s machines. The battle of AIs have begun”. Everyone in the room went into deep thoughts on what they have done to their analytics systems. It was truly a day when Analytics stood still.","excerpt":"It was one of those rare cold mornings in Bangalore in year 2027, the construction dust was floating in the air and visibility was close to 0 meters in Whitefield. It wasn’t even 9AM and the roads were already jammed on the way to ITPL. Software Engineers were rushing to their work on newly opened […]","categories":["IT Services"],"tags":[],"author_name":"Rajkumar Narasimhan","publish_date":"2017-02-14T05:30:14","publication_year":"2017","word_count":616,"keywords":["Go","TPU","machine learning","programming_languages:R","AI","ML","programming_languages:Go","automation","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","TPU","R","Go","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/day-analytics-stood-still-fictional-short-story\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10016551,"title":"Top Movies Of 2020 That Depicted AI","content":"Reel imitates real — this adage has proven to be accurate time and again. As the world progresses to extraordinary advancements in technology such as AI and IoT, movies sincerely try to keep up. This explains the massive popularity of the sci-fi genre, which has found great flavour among movie aficionado. Even as the whole world was down with a pandemic, movies reached their audience across various platforms. This article lists some of the most prominent AI movies that depicted the technology in multiple ways. Spoilers ahead! Tenet Released on: December 4, 2020 Tenet was the most anticipated movie of the year 2020. It also marked the rare distinction of being among the few movies that were screened at cinema theatres this year. The Christopher Nolan-directed movie was touted as his most expensive venture, amounting to $225 million in total cost incurred for the making. This multi-starrer movie features some of the most well-known names including John David Washington, Robert Pattinson, Elizabeth Debicki, Dimple Kapadia, Michael Caine, and Kenneth Branagh. The movie is basically a sci-fi thriller flick where a secret agent, played by Washington, undertakes an international mission to prevent the world’s destruction. The story explores the concept of inverted entropy that causes a person to move backwards in time. The movie has played around concepts of physics, often leaving audiences dazzled. There are a lot of futuristic devices that have been showcased in the movie, including a dead man’s switch which can trigger world destruction with algorithms. Watch the trailer here: Superintelligence Released on: November 25, 2020 Starring Emmy awardee Melissa McCarthy in the lead role, Superintelligence is a romantic action-comedy. McCarthy plays the role of a former corporate executive who is not satisfied with her job. One fine day she discovers that she can communicate with her artificial intelligence-based home network system. Incidentally, the AI system assumes James Coden’s voice, lead protagonist’s favourite celebrity. The AI helps her pay off student loans, pump her bank account up, and fund a billion social cause-based foundation for McCarthy’s character. It even helps her reunite with a former love interest. However, in return, the AI wants her to be the subject of its experiment to learn more about humanity. The movie received moderate reviews, mainly because of its inability to bring forth the risk associated with AI taking control of the world. Watch the trailer here: https:\/\/www.youtube.com\/watch?v=9bRe8sEcOvQ 2067 Released on: October 2, 2020 As the name suggests, it is a futuristic movie set in 2067. Life from the face earth is wiped out except from one city in Australia which is barely holding up thanks to synthetic products produced by an organisation called Chronicorp. All is well until a new deadly disease starts affecting people, including lead protagonist Ethan Whyte’s wife, called The Sickness. A signal from 407 years away in the future is received by the Chronicorp scientists asking to specifically send Ethan to prevent the destruction of humanity, whatever is left. Ethan then sets out to time travel with just an AI-hand computer permanently fitted in his hand. The rest of the movie follows the theme of how Ethan saves humankind from complete extinction. Watch the trailer here: Archive Released on: July 10, 2020 This movie is set in 2038 and revolves around lead character George Almore, a scientist, and his pursuit to build an advanced humanoid robot modelled on his dead wife. He creates three prototypes — JI, J2, and J3. While the first two are more mechanical and boxy in build and have a toddler and an adolescent’s mental capacity, his third creation is closest to a human. The movie goes on to then explore the complexities of a human-AI relationship. Visually appealing, reviewers say that the movie may remind viewers a bit of Ex‑Machina, Ghost in the Shell, and Cyberpunk. It was received with mostly positive reviews. Watch the trailer here: Bloodshot Released on: March 13, 2020 This Vin Diesel-starrer is about a vengeful soldier who is out to seek revenge for his wife’s murder, but with a twist. Vin Diesel and wife’s character are both shown murdered; however, the former is resurrected by an intelligence company that injects nanotech\/bots in his bloodstream, giving him superhuman strengths. The exciting part begins after Diesel successfully kills his wife’s murderer. The same intelligence company that granted Diesel’s character superhuman capabilities now reboot his memory to replace his wife’s killer’s image with faces of targets chosen for elimination. Based on a Valiant Comics character, the original version of Bloodshot was aired in 1993 when sci-fi crime-fighting cyborg soldier themes were popular. Watch the trailer here:","excerpt":"Reel imitates real — this adage has proven to be accurate time and again. As the world progresses to extraordinary advancements in technology such as AI and IoT, movies sincerely try to keep up. This explains the massive popularity of the sci-fi genre, which has found great flavour among movie aficionado. Even as the whole […]","categories":["AI Trends"],"tags":["ai movies","Superintelligence"],"author_name":"Shraddha Goled","publish_date":"2020-12-28T18:00:00","publication_year":"2020","word_count":767,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","BERT","llm_models:BERT","Superintelligence","ai movies","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","BERT","GAN","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-movies-of-2020-that-depicted-ai\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10109577,"title":"ChatGPT Plugins Are Dead and Developers Are Not Happy","content":"When GPTs were launched by OpenAI, we said that it was the first step towards the end of ChatGPT Plugins. And it indeed was. OpenAI has now officially sent a mail to all the plugin developers that they might soon support for plugins, and the developers can now shift to building GPTs, which the company claims is somewhat similar in process. Though ChatGPT plugins are not absolutely dead yet as people can still register to build them, OpenAI’s mail is sent to the people registering for making plugins that GPTs can do the same thing, but better. “They can use actions to call APIs, similarly to plugins, provide custom instructions, invoke DALL-E and more,” read the mail also talking about the launch of GPT Store early next year. OpenAI Email to Plugin Developers Developers are not happy Though it clearly seems like GPTs are easier to build and also offer more functionality when compared to plugins, developers are not happy with OpenAI’s decision. The primary difference lies in the construction method, as GPTs incorporate a no-code chat interface within ChatGPT, whereas plugins are constructed through code outside of ChatGPT. A plugin functions akin to an application linked to ChatGPT, while a GPT resembles a chatbot equipped with specific knowledge and instructions. GPTs are designed to be user-friendly for the general public, while plugins, favoured by developers, provide enhanced functionality. Logan Kilpatrick, OpenAI’s head of developer relations responded to one of the posts on X that said “RIP ChatGPT Plugins”, saying “FYI, plugins aren’t going away yet, this was just a reminder to everyone who joined the developer waitlist that never got access that we launched GPTs and anyone with plus can build them!” In another post, he said that once the GPT Store goes live, most people will move away from Plugins. It seems as though that is not the case though. “Plug-ins were superior to Custom GPTs.” said a user on X. Another said, “As a plugin developer, it doesn’t feel the same for me. Eg previews are open graphs for plugins, but images for GPT.” A user on the developer forum said, “Plugins are crucial for the development and progress in AI applications that we continue to have access to such powerful tools.” To put it simply, OpenAI needs developers to build better use cases of ChatGPT, but this decision has definitely pissed off some of the developers from the company’s community. Since their launch in March 2023, developers have created numerous ChatGPT plugins. However, on November 6th, during DevDay, OpenAI opted to remove plugins from the ChatGPT home screen, thereby increasing the difficulty of accessing them. In an interview with Human Loop, Sam Altman had earlier expressed that “ChatGPT plugins don’t have product market fit,” although OpenAI later requested the removal of the article. Another hot mess? GPTs don’t solve the problems that Plugins had as well. For example, with a little fancier prompt engineering, a user on X was able to download the original knowledge files from someone else’s GPTs. This was the same with ChatGPT Plugins that had the potential to be exploited for unauthorised access to someone’s chat history, retrieval of personal information, and the execution of code on an individual’s machine. But now, just as ChatGPT Plugins store was launched and touted as an iOS store moment, GPTs Store is also the same. There are thousands of plugins available right now, and it is a hot mess. What if the same happens with the new store? Now that we are talking about the iOS store, Apple is also putting a lot of effort into pulling developers onto its platforms. It has released a bunch of open source offerings for its silicon along with multimodal open source models. It is also releasing models for powering edge capabilities. On the other hand, OpenAI is messing with its developers. Arguably, it might not affect the company so much in the longer run, which is what they hope as well. Apple did the same thing earlier when it did not allow ports for its iOS, now OpenAI is doing the same, and is hoping to deliver more finished products.","excerpt":"“Plug-ins were superior to Custom GPTs.” said a developer on X.","categories":["AI Features"],"tags":["ChatGPT","Developers","gpts","OpenAI"],"author_name":"Mohit Pandey","publish_date":"2023-12-27T17:37:24","publication_year":"2023","word_count":692,"keywords":["Go","ChatGPT","API","DALL-E","OpenAI","AI","gpts","GPT","Aim","prompt engineering","R","Developers"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","prompt engineering","R","Go","API","GPT","DALL-E"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/chatgpt-plugins-are-dead-and-developers-are-not-happy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10100753,"title":"When Sam Met Ollie","content":"Silicon Valley is gushing with excitement as Sam Altman, currently the big cheese of the AI world, just introduced his beau, Oliver Mulherin. Alongside his remarkable professional accomplishments, Altman is looking forward to starting a new chapter in life by settling down and preparing for fatherhood with Oliver, affectionately known as Ollie, reported The New Yorker. Both Altman and Mulherin are extremely private about their relationship; the couple was first spotted together when they attended a White House dinner hosted for Indian Prime Minister Narendra Modi in June. The event saw power couples from other big techs like Satya and Anu Nadella, Sundar and Anjali Pichai, and more. Who is Oliver? Although not much is known about Mulherin, what we do know is that he is of Australian origin and has graduated from the University of Melbourne in software engineering. His expertise lies in the space of Internet-of-Things (IoT) as confirmed by his joining of open-source coding organisation IOTA Foundation in 2018. During his university years, Oliver engaged in diverse AI projects, spanning from general game playing to language detection through LSTMs. His foray into the IoT realm commenced with victories in two hackathons, one sponsored by modular phone company Nexpaq and the other by General Electric. In the IoT domain, Mulherin specialized in establishing mesh communication networks connecting cell phones and warehouse sensor networks. Altman with his siblings (Source: New Yorker) Meanwhile, Thirty-eight-year-old Altman has had a “middle-class Jewish upbringing” in Missouri with three younger siblings — Max, Jack, and Annie. The family followed the tradition of having dinner together every night and fostering engagement through games and intellectual challenges. The environment at home was notably supportive, with Altman’s parents consistently expressing love and belief in his abilities. This nurturing atmosphere played a crucial role in instilling a high level of self-confidence in Altman throughout his formative years. In recognition of his efforts, GLAAD honoured Altman in 2017 with the Ric Weiland Award for promoting LGBTQ equality and acceptance within the tech sector. Looking Back Before Oliver, Altman had a nine-year-long relationship with Nick Sivo, who was also his Loopt co-founder and wanted to marry him. Unfortunately, the couple parted ways soon after Loopt was acquired by Green Dot Corporation for $43.4 million in 2012. Altman and Sivo had met at Stanford University and had been together since their sophomore year. Sam & Nick (Source: Business Insider) After acquiring Loopt and parting ways with his longtime partner Sivo, Altman took a year off. During his sabbatical, he read, played video games, and attended a spiritual retreat. In 2014, he wrote a blog post about ‘Founder’s Depression’, sharing his struggles and highlighting the commonality of depression among founders. He has always emphasised the need to openly address mental health challenges within the entrepreneurial community. Why Sam’s Story Matters? Recently recognised as one of the most influential people in AI by TIME magazine, blue backpack ambassador Altman came out of the closet during high school in an environment not particularly supportive of homosexuality. At 17, Altman addressed his school community when objections were raised against a speaker for National Coming Out Day due to religious beliefs and negative attitudes. In his speech, Altman emphasised the importance of tolerance and acceptance, advocating for an open and inclusive community and challenging discriminatory views. His openness holds significance given the stigma still attached to homosexuality in parts of the tech industry. Compounded by a lack of reliable LGBTQ+ data in Silicon Valley, it’s a challenge to address the community’s issues. The fear of prejudice or discrimination keeps many from disclosing their sexual orientation or gender identity at work, hindering accurate workforce diversity assessment. Back in April, AIM got in touch with several queer employees in prominent Indian tech companies to understand their real experiences and gain firsthand insights. To ensure privacy, both corporate names and individual identities were kept confidential. Despite facing initial resistance due to company policies, we found a common theme among respondents. And despite the well-intentioned initiatives by management, a persistent issue of homophobic attitudes among coworkers was prevalent, with almost 60% of them being closeted in office space. The core problem appeared to be the prevalence of subtle homophobic jokes and microaggressions, seemingly innocuous but contributing to a broader culture of discrimination and intolerance within these tech giants. Altman sharing his story becomes pivotal in dismantling barriers and fostering inclusivity amid the prevalent culture of discrimination within these tech giants. Read more: Behind Indian IT’s Mixed Emotions for LGBTQ+","excerpt":"Silicon Valley has a new power couple","categories":["AI Features"],"tags":["ChatGPT","Sam Altman"],"author_name":"Shritama Saha","publish_date":"2023-09-27T18:00:00","publication_year":"2023","word_count":749,"keywords":["Go","ChatGPT","LSTM","Sam Altman","programming_languages:R","AI","programming_languages:Go","Ray","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","Ray","R","Go","GAN","LSTM","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/when-sam-met-ollie\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":15997,"title":"The biggest big data trends of healthcare in 2017","content":"Big data is saving lives, and that’s not a fairytale. It has been changing lives in many ways. Big data is bringing a welcome shift in the healthcare sectors. Healthcare analytics cannot only help reduce the cost of healthcare facilities including treatments, medication, and diagnosis. Analytics in this area can also contribute to predicting the outbreak of endemic and epidemic diseases like SAARS and the Flu. According to a recent McKinsey report, more than 20 years of steady increase, health care now represents 17.6% of the country’s total GDP. This is $600 million over the reasonable upper limit for the USA. Healthcare costs have skyrocketed in the last 20 years. No new presidential policies or health insurance regulations have been able to dampen that growth. However, right now, healthcare officials are finding a new incentive to share patient information with one another. There are medical platforms for the interaction of patients as well as professional platforms for exchange of information for the medical professionals. This is possible by prioritizing individual patient health over the number of patients. Right now, real decisions are more evidence-based than ever. There are pools of research and clinical data that these doctors can directly access. Professionals now have all the assistance they need, thanks to secure applications that allow them to access and utilize the data pool. Data-driven healthcare has its own set of obstacles. Medical data encompasses different hospitals, districts, and states. They include several administrative systems. This call for the necessity of a new tool that can help data providers and data users collaborate with each other. This is why the creation of new analytics tools, strategies, and data applications are significant right now. Healthcare needs acute data analysis in the forms of the graph, machine learning and predictive analysis that other industries are already enjoying. What are the biggest big data trends of healthcare in 2017? 1. Prioritization of patient-oriented care As we already stated before, patient care quality is on the rise. More doctors are finding incentives in providing individual patient-related data. The quality of data relates to the quality of patient healthcare. This is becoming beneficial for the patient since it is improving the quality of healthcare that is consistent with the professional experience and knowledge. It has also reduced the health care costs and provided support for payment structures. This means the practice is gradually moving away from fee-for-service to fee based on the quality of service. This will not only improve patient outcomes, but it will also reduce the net healthcare costs. 2. IoT and healthcare The Internet of things is the new LBD of data technology. Industrial Internet or IoT describes the fast increase in the volumes of data each connected device and the individual user generates. IoT is going to be worth about $120 billion in the next two years. Interestingly, most of the data from the healthcare IoT is unstructured. This means, data engineers can find the widespread potential use of Hadoop and Kafka-like analytics and framework. The variety of services can range from storing individual patient data to pulling data related to a particular subset of symptoms that come from thousands of patients from all 50 states. This opens a range of employment opportunities for those interested in Six Sigma Green Belt Training. Knowledge of management services will help the data scientists in the better structuring of inflowing data from the very beginning. Healthcare IoT already generates gargantuan qualities of data each waking moment. If the data is unstructured for long, it might become too chaotic to implement any analytics program and structure. 3. Better management and monitoring You can imagine the data lake as a fathomless body of data flowing in from 7 other sources. Claims, clinical, pharmacy, EMR, logs, and notes, third party data and additional data, all contribute to the healthcare data lake. This serves as the data hub for all health related departments including fraud prevention and management. It is now possible to detect over 20% cases of fraud, waste, and abuse in the claims department(s) of the hospital(s). The Centers for Medicare and Medicaid Services now uses predictive analysis to assign risk scores to individual claims and their providers. This helps the model to flag certain charges automatically. The predictive mode can compare charges against profiles that may be fraud and raise a red flag in the system. This helps prevention of fraudulent insurance claims and medical aid claims across the USA. 4. Predictive analysis can improve outcomes The adoption of EHR or Electronic Health Records is helping the volume of patient data grow exponentially. The $30 billion stimuli from the federal government catalyzed the use of EHR. The new EHR policies can now combine and analyze data from several data sources. Multiple medical boards, administrations, and practitioners can now enjoy the benefit of a predictive data analytics model derived from data from the EHRs. At this moment, Congestive Heart Failure or CHF accounts for most spends in healthcare. The patients and, sometimes, the doctors neglect most early symptoms too. It can be treated and kept under control if detected early. However, the lack of enough information on the symptoms makes it very challenging for physicians to diagnose it early. Machine learning can take into account number of factors. The algorithms can factor in additional features that physicians cannot see or detect. This increases the chances of correct, early diagnosis of the patients. This model will distinguish people who have CHF from the healthy population. Machine learning and big data have applications contribution in the monitoring of post-trauma and post-op patients. From heartbeats, blood pressure, breathing to brain activity – big data can collect, store and structure all data. Real-time monitoring is a big pro with big data.","excerpt":"Big data is saving lives, and that’s not a fairytale. It has been changing lives in many ways. Big data is bringing a welcome shift in the healthcare sectors. Healthcare analytics cannot only help reduce the cost of healthcare facilities including treatments, medication, and diagnosis. Analytics in this area can also contribute to predicting the […]","categories":["IT Services"],"tags":[],"author_name":"Sujain Thomas","publish_date":"2017-06-30T10:47:25","publication_year":"2017","word_count":959,"keywords":["big data","Go","machine learning","AI","ML","medical AI","Aim","analytics","Kafka","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","medical AI","Kafka","R","Go","big data"],"url":"https:\/\/analyticsindiamag.com\/it-services\/biggest-big-data-trends-healthcare-2017\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10166065,"title":"To Code or Not to Code","content":"Anthropic chief Dario Amodei recently said that AI will handle 90% of coding in less than six months. If you’re a developer, proclamations such as these might leave you rattled and pondering over what the next few years will mean for your career and skills. Amodei has a point. With AI-native tools like Cursor, Windsurf, and GitHub Copilot, coding has become easier than ever for developers. In a recent Y Combinator podcast, partner Jared Friedman said that one-quarter of the YC founders admitted that over “95% of their codebase was AI-generated”. He pointed out that these were highly skilled founders who, just a year ago, would have built their products entirely on their own—but now, AI does the heavy lifting. Dianu Hu, partner at YC, added that, just as Gen Z was born into the internet era, the current generation will grow up with AI tools. “They’ll skip the classical training of a software engineer and just do it with the vibes, but they’re actually very technically minded. I mean, they have degrees in math and physics,” she said. However, as a pitfall, young developers are not exactly aware of how their code works and tend to go blank when asked how their code works. “Junior devs these days have it easy. They just go to chat.com and copy-paste whatever errors they see. Even lazier ones don’t make the 30-second effort of toggling to a browser window. They just use a tool that does it all at one place,” wrote Namanyay Goel, founder of Giga AI, in a blog post. He added that young developers should approach AI with a learning mindset. “Don’t just accept its answers—question them. Ask why. It may take longer, but that’s exactly the point,” he said. Interestingly, in a surprising turn of events, recently, Cursor refused to generate code and instead encouraged the developer to explore learning opportunities. Stemming from all the AI-assisted coding is the new fad of ‘vibe coding, ’ which many developers have taken to lately. The term was coined by OpenAI co-founder Andrej Karpathy. Vibe coding involves using AI tools to handle the majority of coding tasks, allowing developers to focus on high-level intent rather than low-level implementation. Developers describe their desired outcomes in plain language, while AI generates, refines, and tests the corresponding code autonomously. InMobi chief Naveen Tiwari recently said that the company is on track to achieve 80% automation in software coding by year-end. “We have already achieved 50% [automation in software coding]. The codes created by the machine are faster and better, and they fix themselves.” Shipping code has never been easier for developers. “Platforms like GitHub Copilot, Microsoft Copilot and Azure AI services allow teams to focus on critical thinking and creative aspects of their projects, reducing the time spent on mundane tasks,” Santhosh HS, AI engineer at TCS told AIM. AIM spoke to a few companies and startups and realised that many of them had already adopted AI tools. “We’ve been using Cursor in our organisation for a while now, and it’s definitely boosted productivity,” said Abhishek Upperwal, founder of Soket AI. However, he added that blindly relying on these tools can waste a lot of time since they’re prone to errors. “They work really well for common tasks like web development but tend to fall short with more complex challenges like building or optimising CUDA kernels in Triton—mainly because the AI hasn’t been trained on enough examples in those areas.” AIM got in touch with Himanshu Gahlot, VP of engineering, and Saravana Kumar, head of machine learning at Apollo.io, who were happy using Cursor. “We have a 90% plus satisfaction rate. Almost every engineer said positive things about being able to understand the whole code base and generate the right things,” Gahlot said. “It does come with a caveat. You will hear many people hyping a lot of these tools all the time, saying that the gain in productivity is 25x, 50x, and so on. But it’s very, very nuanced,” he cautioned. Is Vibe Coding Even Real? IBM chief Arvind Krishna isn’t convinced that AI will take over software coding anytime soon. He dismissed Amodei’s claim that AI could possibly generate 90% of the code within the next three to six months. “I think the number is going to be more like 20-30% [of the code getting written by AI]—not 90%,” Krishna said. “Are there some really simple use cases? Yes, but there are equally complicated ones where it’s going to be zero.” Sharing a similar perspective, Linas Beliūnas, director of revenue at Zero Hash, pointed out that AI struggles with complex code. “It excels at routine tasks but stumbles upon creativity, nuance, and context-specific solutions,” he said, adding that code isn’t just writing. “It’s problem-solving, ethics, security, compliance, and creative design—all of which are deeply human.” “Language models can help with the first 70% but will not be able to help with the last 30%. No matter how many times you explain your problem, how many times you ask it to change non-working code, or how long you finetune it with reinforcement learning, it will not write brand new business- or app-specific code,” said Andriy Burkov, machine learning lead at TalentNeuron. Some believe that too much use of AI coding tools can increase technical debt. “AI accelerates code generation, but without strong governance, organisations will drown in unmaintainable, poorly structured, and undocumented code. Fixing issues later will be exponentially harder,” said Pradeep Sanyal, AI and data leader at a global tech consulting company. Karan MV, director of international relations at GitHub, told AIM that while automated processes such as testing, monitoring, and alerting can help manage development, human intervention is still crucial at various stages in the software development lifecycle, even when using AI tools. While AI is transforming coding at an unprecedented pace, human judgment remains irreplaceable. The future of software development won’t be about choosing between AI and human expertise—but about finding the right balance between the two.","excerpt":"One-quarter of the YC founders admitted that over “95% of their codebase was AI-generated”.","categories":["Global Tech"],"tags":["coding","Vibe Coding"],"author_name":"Siddharth Jindal","publish_date":"2025-03-14T14:34:46","publication_year":"2025","word_count":997,"keywords":["CUDA","Anthropic","Go","machine learning","OpenAI","AI","R","coding","RAG","Vibe Coding","Aim","Azure"],"extracted_tech_keywords":["AI","machine learning","OpenAI","Anthropic","Aim","RAG","Azure","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/to-code-or-not-to-code\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10023563,"title":"Clinical Site Visits and Artificial Intelligence","content":"It is an age-old practice wherein the clinical research associates (CRA) travel for their allocated sites at every pre-fixed regular interval. The CRAs need to verify every entry made by the sites and the patients. In other words, it is known as source data verification. Though it is a time-consuming method. The requirement is to complete the verification process within a limited timeline that is based on the budget of the client and the monitoring manual requirements. This would be followed by the report preparation and follow-up letter preparation within the timelines. Again, another site allocated for the CRA would be ready for his\/her visit by the time the CRA completes his letter preparations. This continuous circle is time and money consuming. The cost of travel and the stay of CRA must be taken care of by the client. This would add to the cost of the drug indirectly. Sometimes, it might happen that the study site might have more patient count. Hence the documents that need to be verified will be huge. To compensate for the huge count, additional CRA might accompany the main CRA. Hence the cost would be increased. Due to the tedious activity of site visits continuously along with time constraints, there are chances that the CRA might miss on the main pointers. Based on major drawbacks faced due to traditional monitoring system, so-called Risk-based monitoring was introduced. What is Risk-based monitoring? According to this monitoring method, the CRA need not invest their time in complete source data verification nor the CRAs need to visit each site at regular intervals. Instead, they are asked to visit the sites which seem to be falling under the category of risk and they need to verify only the documents that are essential. That is the complete 100% source data verification is avoided and the number of travels required to be made is reduced. This in turn would reduce the cost and time involved. Instead, the CRAs can concentrate on the critical pointers which would lead to more qualitative control of the study conduct. The next step is to find the pre-defined triggers that lead to the identification of the risk sites. This is possible using artificial intelligence coupled dashboards. How exactly these dashboards work? The data is entered into the system at various levels, that from the time a patient enters the trial till he\/she is out of the trial. A few of the important data that is entered in the system are: The initial details of the patient such as height, weight, age, and other basic details of the patient. The pre-screening and screening data is entered, wherein the eligibility of the patient based on the inclusion and exclusion criteria is decided.The medication administered to the patient at what time and date is entered.The side effects experienced by the patients are recorded.The blood tests and other parameters tested at the regular intervals as mentioned in the protocol is also entered into the system. These data form an important base for designing the criteria for terming a site as high risk, medium risk, and low risk. For example, if serious adverse events are over or under-reported at a single site. This triggers an alarm to the CRA for the site visit. If there are pending action items from the site to be completed for a prolonged period, this also forms a risk criterion. Through the artificial intelligence coupled to these data, the statistical analysis of the trial data is performed during the conduct phase. This would help to reconsider, and redesign important documents involved in the study such as the communication plan, monitoring plan. This would also help to achieve the trial end at a faster rate compared to the stipulated date. The scope for remote monitoring also increases as the source data is available in real-time in the system which can be cross verified by the offshore CRAs at regular intervals without the need for site visits. Hence the coupling of artificial intelligence along with the database system of the clinical trial is a boon. The investigational product can be studied at a more intense level as the artificial intelligence system helps to realize the graphs which can be easily missed by humans using a traditional system. This would help in providing a quality and cost-effective drug care system. The time taken for a product to enter the marketing system can be reduced.","excerpt":"The data is entered into the system at various levels, that from the time a patient enters the trial till he\/she is out of the trial.","categories":["AI Features"],"tags":[],"author_name":"Prasidha","publish_date":"2021-04-07T18:00:00","publication_year":"2021","word_count":735,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/clinical-site-visits-and-artificial-intelligence\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":45574,"title":"Machines Might Not Mind Your Accent Anymore","content":"For speech recognition systems, a change in the accent can be confusing. Words under the influence of local languages sound different and a typical homepod device can mistake an Asian speaking English or even something as native as a thick Irish accent. Deep Learning algorithms are the work horses behind these devices. The training data on which the models were trained could have been acquired from a single region where variations are negligible. For instance, in India, though Hindi is widely popular, the accent of a Bengali will be in stark contrast to that of a Malayali speaking Hindi. Collecting data that caters to intricacies of such native language influences can be tricky and tedious. In order to address the shortcomings of speech recognition systems, researchers at Stanford came up with a novel architecture and the experimental results show that the model has performed well in identifying the differences. A Brief Intro To Speech Recognition Systems The first step in any automatic speech recognition system is to extract features i.e. identify the components of the audio signal that are good for identifying the linguistic content and discarding all the other stuff which carries information like background noise, emotion etc. MFCCs is something one would come across while designing ASR systems. A minimal understanding of MFCCs would give an idea about its usage in the next section of this article. Mel Frequency Cepstral Coefficients (MFCCs) are a feature widely used in automatic speech and speaker recognition. Whereas, Mel cepstral distortion (MCD) is a measure of how different two sequences of mel cepstra are. It is used in assessing the quality of parametric speech synthesis systems, including statistical parametric speech synthesis systems, the idea being that the smaller the MCD between synthesized and natural mel cepstral sequences, the closer the synthetic speech is to reproducing natural speech The main point to understand about the speech is that sounds generated by a human are filtered by the shape of the vocal tract including tongue, teeth etc. This shape determines what sound comes out. If the shape is determined accurately, this would give an accurate representation of the phoneme being produced. The shape of the vocal tract manifests itself in the envelope of the short time power spectrum, and the job of MFCCs is to accurately represent this envelope. Overview of The Model Architecture  diagram via paper by Amy Bearman et al., The authors in their paper propose a methodology for accent conversion that learns differences between a pair of accents and produces a series of transformation matrices that can be applied to extracted Mel Frequency Cepstral Coefficients.  This is accomplished with a feed forward artificial neural network, accompanied by alignment preprocessing, and validated with MCD and a softmax classifier. For simplicity of development and due to available training data, the researchers trained and tested the model on the English language with American, Indian, and Scottish accents, for both genders. The authors extracted 25 mel cepstral coefficients from each 5ms frame with 100 frequency bands in each  of the training samples and paired samples of identical utterances in two different accents for the source and target data into the system.  Each feature vector was zero-padded or truncated to the same length, which we set to be 1220 frames per sample. After extracting the MFCCs, the source and target were aligned using FastDTW. Alignment is necessary because people speak at different rates and without alignment it is much harder for  the system to identify which differences are due to accent and which are due to rate of speech. The dataset  used for the experiment consists  of 1150 samples of text spoken by men with American, Canadian,  Scottish, and Indian accents, and a woman with an American accent. The final model used Adam optimization to minimize mean squared error over 5,000 epochs with batch  size 16. Basic gradient descent was tried initially, later TensorFlow’s MomentumOptimizer and then Adam Optimization was used. The results show that the  classifier achieved 92.9% accuracy in binary classification on the benchmark American English versus Scottish English task, significantly outperforming the 68% accuracy of a Naive Bayes classifier and 76% accuracy  of a Support Vector Machine classifier for the same problem. The model also achieved MCDs below 10 for all three of the conversions that were attempted. Voice conversion is an active area of research, but the  majority of papers on the subject focus on modifying the voice itself, not the pronunciation The success of the models such as discussed above demonstrate  that it is possible to reconstruct a speech sound and make the speech recognition systems more flexible with varying accents. For further reading: A tutorial on MFCCs Original paper by Stanford","excerpt":"For speech recognition systems, a change in the accent can be confusing. Words under the influence of local languages sound different and a typical homepod device can mistake an Asian speaking English or even something as native as a thick Irish accent. Deep Learning algorithms are the work horses behind these devices. The training data […]","categories":["Deep Tech"],"tags":["language","Machine Learning","naive bayes","Stanford"],"author_name":"Ram Sagar","publish_date":"2019-09-05T19:45:45","publication_year":"2019","word_count":781,"keywords":["Go","programming_languages:R","AI","neural network","language","Machine Learning","programming_languages:Go","naive bayes","ai_frameworks:TensorFlow","deep learning","Stanford","TensorFlow","R"],"extracted_tech_keywords":["AI","deep learning","neural network","TensorFlow","R","Go","ai_frameworks:TensorFlow","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/machines-might-not-mind-your-accent-anymore\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10101844,"title":"ISRO Achieves Milestone in Gaganyaan Mission with Successful Escape System Test","content":"In a remarkable display of precision and determination, the Indian Space Research Organisation (ISRO) achieved a significant milestone in its ambitious Gaganyaan Mission – the endeavour to send Indian astronauts into space. The latest test, part of this mission, involved the launch of the Test Vehicle (TV-D1), powered by a liquid-propelled single-stage rocket, from the Satish Dhawan Space Centre. The focus of this mission was the validation of the Crew Escape System, a critical component for the safety of Indian astronauts during spaceflight. This system comprises various motors, including low-altitude, high-altitude, and jettisoning motors, designed to safely eject astronauts from the vehicle in case of an emergency. ISRO Chief S. Somnath declared the mission complete with all objectives achieved. “Im very happy to announce the completion of TV-D1 mission,” said the ISRO chief, addressing the team dedicated to the launch. Quick Challenge Recovery The journey to this moment was not without its challenges. Initially scheduled for launch at 8:00 a.m., the mission experienced a delay due to adverse weather conditions. However, what transpired next is a testament to the dedication and problem-solving capabilities of the ISRO team. Within an hour of experiencing a glitch during the first launch attempt, ISRO identified and resolved the issue. The team’s swift response and ability to rectify the problem within such a short timeframe were truly remarkable. The flight sequence for TV-D1 commenced with the launch, and six seconds into the flight, the fin-enabling system was activated. The Crew Escape System Pillbox was triggered at a speed of Mach 1.25, approximately 11.8 km above the Earth’s surface. The High Energy Motor (HEM) fired to propel the vehicle further into the atmosphere. At about 61.1 seconds after launch, when the vehicle reached a Mach number of 1.21 at an altitude of 11.9 km, the Crew Escape System separated from the rocket booster. The Crew Module then separated from the Crew Escape System at an altitude of 16.9 km, travelling at a speed of 550 km per hour. Subsequently, the drogue parachute deployed, slowing the vehicle’s descent. “Mission Gaganyaan TV D1 Test Flight is accomplished. Crew Escape System performed as intended. Mission Gaganyaan gets off on a successful note,” ISRO announced. The Crew Escape System, akin to fighter jet ejection seats, plays a vital role in safeguarding astronauts during spaceflight anomalies. It operates automatically, detecting malfunctions or issues immediately after liftoff, prior to rocket stage separation. This rapid ejection minimises potential risks during the early phases of ascent. The successful test launch was crucial for validating the system that would be indispensable during the Gaganyaan Mission‘s initial phase. It will be responsible for jettisoning the Crew Module, with astronauts on board, to a safe distance in the event of a critical problem. The Crew Module will then separate and safely splash down in the sea, aided by parachutes. Data gathered from this test will further enhance the system’s development, ensuring its reliability when astronauts embark on the Gaganyaan Mission. This achievement underscores India’s progress toward realizing its dream of sending its first astronauts into space from its own territory. The Indian space agency plans to launch its first astronauts, who are currently undergoing training, into space by 2025 and aims to reach the Moon by 2040. ISRO’s unwavering commitment, ability to overcome challenges, and rapid problem-solving abilities have taken India one step closer to a historic space mission, marking a significant milestone in the country’s space exploration journey.","excerpt":"The successful test launch was crucial for validating the system that would be indispensable during the Gaganyaan Mission’s initial phase.","categories":["AI News"],"tags":["ISRO","safety","space exploration"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-10-21T16:15:22","publication_year":"2023","word_count":572,"keywords":["Go","ISRO","API","programming_languages:R","AI","programming_languages:Go","Aim","safety","ViT","GAN","space exploration","R"],"extracted_tech_keywords":["AI","Aim","R","Go","API","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/isro-achieves-milestone-in-gaganyaan-mission-with-successful-escape-system-test\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10161603,"title":"Microsoft Unveils MatterGen, an AI Breakthrough for Materials Discovery","content":"Microsoft on Thursday launched MatterGen, a generative AI tool designed to revolutionise how we understand material discovery, marking a transformative moment in materials science. “Our MatterGen model applies generative AI to create new compounds with unprecedented precision,” said Satya Nadella, chairman and CEO at Microsoft. Unlike traditional approaches that test existing materials, MatterGen can generate entirely new ones based on specific requirements. The researchers detailed this breakthrough in their paper ‘A generative model for inorganic materials design.’ “MatterGen offers a paradigm shift,” said senior researchers Claudio Zeni, Robert Pinsler, Daniel Zügner, Andrew Fowler, and others. Outperforms Screening Methods Traditional methods reach a limit of 40 candidates when looking for materials with specific properties, such as high compression resistance. MatterGen, however, discovered over 100 potential candidates in tasks such as identifying stable, high-bulk modulus structures. The AI is also trained on extensive datasets, including the Materials Project and Alexandria databases, to ensure state-of-the-art performance. The model uses an innovative algorithm to handle complex material structures more accurately. Screening vs generative approaches to materials design MatterGen Created a New Material! The tool’s capabilities were tested in collaboration with Prof. Li Wenjie’s team at the Shenzhen Institutes of Advanced Technology (SIAT) of the Chinese Academy of Sciences. They challenged MatterGen to design a material with specific compression resistance (200 GPa bulk modulus). The result? A new material called TaCr₂O₆ was successfully synthesised and matched the AI’s predictions, even accounting for variations in how the tantalum (Ta) and chromium (Cr) atoms were arranged. Experimental validation of the proposed compound, TaCr2O6 Open Access and Future Directions Christopher Stiles from the Johns Hopkins University Applied Physics Laboratory highlighted the significance of the innovation: “We are interested in understanding the impact that MatterGen could have on materials discovery.” MatterGen’s developers have released its source code under the MIT license, encouraging community collaboration. Researchers aim to expand the tool’s applications in fields such as battery and magnet development. The integration of MatterGen with AI simulation tools like MatterSim further accelerates material exploration and simulation, creating a dynamic system for scientific discovery. Materials discovery not only began with Microsoft, but Google DeepMind also released research titled ‘Scaling deep learning for material discovery’ in 2023, where they discovered 2.2 million new crystals, equivalent to 800 years of work of knowledge. Meta also entered material size by releasing a massive data set called Open Materials 2024 (OMat24), which contained over 118 million examples of material simulations and structures. It focused on a wide range of inorganic bulk materials to improve AI-enabled material discovery. In December last year, Amazon also announced a multi-year partnership with Orbital Materials to develop new materials that help decarbonise data centres using their ‘proprietary AI platform’.","excerpt":"Developers have released source code for MatterGen under the MIT license.","categories":["AI News"],"tags":["Materials Science","Microsoft"],"author_name":"Sanjana Gupta","publish_date":"2025-01-17T12:26:15","publication_year":"2025","word_count":449,"keywords":["Go","AI","innovation","RAG","BERT","Aim","deep learning","generative AI","GAN","Materials Science","R","Microsoft"],"extracted_tech_keywords":["AI","deep learning","generative AI","Aim","RAG","R","Go","BERT","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-unveils-mattergen-an-ai-breakthrough-for-materials-discovery\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10043601,"title":"AI Goes Phishing","content":"In 2020, Google claimed to block more than 100 million scam emails every day, with 18 million of them related to COVID 19.  According to Barracuda Networks, malicious mails rose by 667% with the onset of the pandemic. Mobile devices were the most vulnerable. Verizon’s 2020 Data Breach Investigations Report (DBIR) showed hackers found a lot of success with integrated text, email and link-based phishing, especially across social media, for stealing passwords and accessing privileged credentials in cyberspace. Now, machine learning models are evolving to understand and filter out phishing threats to internet users, governments, and companies etc. For example, Microsoft neuters billions of phishing attempts on Office 365 alone. Over the years, hackers have become better at evading detection, sneaking in malicious content. The tactics manifest themselves in URLs pointing to legitimate-looking yet compromised websites and redirectors. AI has the ability to detect spam and phishing attacks with accuracy, speed. Automated detection AI goes beyond signature-based detection, which hackers have learnt to evade by tweaking some elements like HTML code or image metadata. Incorporating machine learning capabilities, AI focuses on detecting characteristics\/behaviours related to phishing as opposed to known signatures. An altered signature can be detected and blocked. Phishing attacks are constantly evolving to evade newer technologies, and cybersecurity tools need to keep up. AI is continuously learning from open source threat intelligence feeds, and the organisation’s own unique environment. Source: Abdul Basit, et al Studies have shown that robust ML techniques have high detection accuracy. AI uses machine learning and data analysis to examine content, context, metadata and user behaviour. Behavioural analysis AI and ML algorithms can understand how users communicate. They study patterns of typical behaviour, textual behaviours and the context of messages. Communication patterns are assessed to create a baseline of normal behaviour. Characteristics like use of grammar, syntax etc creates a unique user profile. Impersonation or spear phishing like Business Email Compromise (BEC) and Email Account Compromise (EAC) scams can be detected this way although it may pass other filters. Challenges AI models are only as good as the data they are fed and its trustworthiness revolves around data. Hence, data bias is a pertinent risk. Oftentimes, enterprises have been deploying these tools assuming the datasets are well represented, which may not be true. AI data training can be poisoned by malicious actors, compromising the secure structure that a particular organisation is relying on. Technology does not possess an inherent disposition and will act the way it is taught. However, these tools do not operate in a vacuum, rather, they are interacting with their environment all the time. AI algorithms can be exploited and even weaponised to pursue nefarious objectives. The ability to create synthetic data that mimics the human generated content could be the beginning of Deepfake spear-phishing. As per Europol’s report, artificial intelligence could potentially make cyberattacks more dangerous and difficult to detect. Just like organisations deploy AI to protect against malware, it is possible that hackers have begun making use of AI and ML tools too. AI models may suffer as adversaries begin to identify patterns and change their mode of operation rendering the existing data and AI models useless. The model will adapt to certain phishing behaviours over time which decreases its efficacy to detect novel threats. AI\/ML blind spots Known unknowns and unknown unknowns are still a major threat to models. Although ongoing research aims to find answers and suggestions, these unknowns may not elicit any threat response from AI\/ ML tools. Datasets may also contain errors leading to labeling flaws on part of the AI. ML is exposed to comprehensible patterns which tell the machine what to look for in datasets to predict future malware. Datasets can become obsolete and irrelevant. An MIT study found that major ML datasets had significant errors including mislabelled images. The study found a 3 to 4% average error rate in datasets and a 6% error rate for Imagenet, one of the most popular image recognition systems.","excerpt":"Artificial intelligence could potentially make cyberattacks more dangerous and difficult to detect","categories":["AI Features"],"tags":["ai generated images"],"author_name":"Prajaktha Gurung","publish_date":"2021-07-16T14:00:00","publication_year":"2021","word_count":663,"keywords":["CUDA","ai generated images","machine learning","artificial intelligence","AWS","AI","ML","image recognition","RAG","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","image recognition","AWS","CUDA","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-goes-phishing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10016759,"title":"How Is Rebel Foods Transforming The Food Business With AI","content":"Built on a belief that food is an amalgamation of art and science, technology is the main ingredient at Rebel Foods, a leader in India’s cloud kitchen restaurants. To begin with, the company uses machine learning to widen cuisine preferences, offers personalised food, and uses data analysis in areas such as supply chain management, recommendation systems, and more. By strongly integrating technology, it has built brands such as Behrouz Biryani, Faasos, Oven Story Pizza, and others. Analytics India Magazine got in touch with Amit Gupta, Chief Technical Officer, Rebel Foods, to understand its strong tech hold and how it stands ahead of the competition using AI. Built On A Strong Data Strategy Gupta shares that technology & data are the backbones of solving industry-first problems at Rebel Foods. “Many decisions and light bulb moments for transitioning Rebel foods to a cloud kitchen model have happened on certain fact realisations out of data,” he said. It is in their DNA to make data-driven decisions and scale their offerings. “AI and ML are tightly integrated into every problem solving that are happening on the floor of our company,” he added. The company is divided into three verticals — Food Discovery, Food Production, and Food delivery. Each of these platforms has its own problem statements, and most of these require some or other solutions on top of data & data science. They use software, robotics, and automation to solve food preparation at scale and provide consistent quality for diversified palates. Some of the live projects at Rebel Foods are: Visual AI QC machines: that can detect SWAT (size, weight, appearance, and temperature) for each prepared dish and reject or accept based on the extent of deviation from ideal A robotics-led smart fryer: that gets adjusted automatically for the oil temp, dipping, and releasing based on what you are frying. Automated Wok: 100+ of their kitchens now have an automated Wok that dispenses oil, water, and other ingredients Gupta said that they think of problems statements from consumers’ perspective first — “How can I be sure what I am eating? What hands have touched my food? What are the ingredients? I do not know of any restaurant that customises the food based on what I like or dislike? What am I allergic to? What are my specific dietary requirements?” They further act upon these problems by using technology and data. “As we own the full stack control of the whole platform from SCM to discovery to order to preparation to delivery, we are envisioning the platform to provide complete transparency and control to customers for food order using any interface,” said Gupta. When Data Science Meets Food: Use Cases Data science plays a vital role in Rebel Food. They work on many exciting data engineering use cases such as personalisation, recommendation engine, dynamic pricing, predictive modelling for inventory, customer engagement, operating metrics for kitchen machines, vision computing for quality checks, sentiment and semantic analysis of user feedback, and more. “We are also launching face mask detection during order deliveries to keep a check on safety and compliance protocols,” added Gupta. Explaining a few use cases around how new-age tech is used at Rebel Foods, he shared a few instances as below: Machine learning offers personalised food services: To bring about personalisation, they ingest large amounts of real-time and batch data and run data science models to improve customer experience on the ordering platform. They also use it to make the whole experience seamless. Machine learning is also used to make the discovery and ordering experience unique, fast, and seamless for all different types of users. As Gupta shares, “we want to go further to give as much control in the user’s hand as possible for food customisation and preparation by integrating the flows to backend systems (IoT).” Data analysis for supply chain management: Gupta shares that they use tools such as Sap\/Hana in combination with the in-house platform (Spark) for all the inventory and assets management. These systems are integrated with different vendors\/partners to provide an end to end view of the whole process and workflows. He further added that they are using data science heavily in demand forecasting and inventory planning to minimise variance and wastage. Their models are trained on over five years of data to ensure the best possible accuracy levels. Analytics-driven recommendation engine: As a tech-driven B2C commerce platform, the recommendation engine is one of the key modules which is owned and constantly improved by their data engineers and data scientists. They use various data types, on top of which they use a mix of content-based & collaborative filtering for recommendations. “Majorly, we measure the efficacy of our recommendation engine on 2 metrics – benchmark improvement of cross-sell\/up-sell and relevancy of recommendation result-set,” he said. Tech Stack At Rebel Foods Gupta shares that they have built complete full-stack systems deployed on the cloud in food discovery, preparation, and delivery. They are a big believer in using open-source technologies wherever possible, and most of their technologies are built in-house. He further added that they have built their data pipeline using Kafka, spark, S3\/Hive, and other apache open-source technologies. They also use data science libraries on AWS sage maker and some standard & some in-house built data visualisation tools. The data science team also uses tools and algorithms in machine learning (regression, classification, random forest, k nearest, etc.), deep learning (Neural Networks, Tensorflow\/Keras, etc.), AI, Statistics, SQL\/NoSQL DBs, Semantic & Sentiment analysis, Predictive Modelling. The company takes pride in its strong integration of technology, showing sincere commitment to redefining the way we eat and paving the path for the future of food while retaining the best of traditional culinary practices. “Though the brands and food choices could differ across the globe, the problem statements for the platform remain somewhat common, and we build solutions and platforms in that context, which makes us stand out” said Gupta in the concluding remarks.","excerpt":"Built on a belief that food is an amalgamation of art and science, technology is the main ingredient at Rebel Foods, a leader in India’s cloud kitchen restaurants. To begin with, the company uses machine learning to widen cuisine preferences, offers personalised food, and uses data analysis in areas such as supply chain management, recommendation […]","categories":["Deep Tech"],"tags":["business analysis tools"],"author_name":"Srishti Deoras","publish_date":"2020-12-30T12:00:00","publication_year":"2020","word_count":988,"keywords":["data science","machine learning","Keras","AI","neural network","ML","business analysis tools","recommendation systems","deep learning","analytics","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","data science","analytics","TensorFlow","Keras","recommendation systems"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-is-rebel-foods-transforming-the-food-business-with-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10076106,"title":"OpenAI&#8217;s Whisper is Revolutionary but (Little) Flawed","content":"Speech recognition in machine learning has always been one of the most difficult tasks to perfect. The first speech-recognition software was developed in the 1950s, and we’ve come a long way since. Recently, OpenAI took a leap in the domain by introducing Whisper. The company says it “approaches human level robustness and accuracy on English speech recognition” and can automatically recognise, transcribe, and translate other languages like Spanish, Italian, and Japanese. There’s no doubt that Whisper works better than any other commercial ASR (automatic speech recognition) system like Alexa, Siri, and Google Assistant. OpenAI, the company that usually does not do justice to its name, decided to open source this model. The digital experience will change radically for many people, but is the model revolutionary? Here’s what you need to know As with almost every major new AI model these days, Whisper brings along advantages and potential risks associated with it. On Whisper’s under the ‘Broader Implications’ section of the model card, OpenAI warns that it could be used to automate surveillance or identify individual speakers in a conversation, but the company hopes it will be used “primarily for beneficial purposes”. Conversations have also surfaced on the internet about the challenges faced by the early users of this revolutionary transformer model. As a side note, OpenAI researchers chose the original transformer architecture because they wanted to prove that high-quality supervised ASR is possible if enough data is available. The main challenge is that your laptop may not be as powerful as the computers of professional transcription services. For example, Mitchell Clarke fed the audio from a 24-minute-long interview into Whisper, running on M1 MacBook Pro. It took almost an hour to transcribe the file. On the contrary, Otter completed the transcription within eight minutes. Secondly, installing Whisper is not really a user-friendly process for everyone. Journalist Peter Sterne teamed up with GitHub developer advocate Christina Warren to try and fix the issue by creating a “free, secure, and easy-to-use transcription app for journalists” based on Whisper’s ML model. Sterne said that he decided the program, dubbed Stage Whisper, should exist after he ran some interviews through it and determined that it was “the best transcription he’d ever used, with the exception of human transcribers”. I'm working on a new project with @film_girl to create a free, secure, and easy-to-use transcription app for journalists, powered by @openai's whisper ML model.If you're interested in contributing to the project, please let us know and we'll add you to the github repo.— Peter Sterne (@petersterne) September 22, 2022 Another red flag is that the prediction is often biased to integer timestamps. Users observed that those tend to be less accurate; blurring the predicted distribution may help, but no conclusive study has been done yet. The timestamp decoding heuristics is a bit naïve and could be improved along with word-level timestamping. Peculiar failure case? Whisper has also been described as a ‘peculiar failure case‘. The reason being, the model sometimes exhibits failures in recognition quality. (Credits: https:\/\/docs.google.com\/spreadsheets\/d\/1xdaK-RJZ2ftMKBME45aAeEmMHSJSxb3wW8-GzT1whgg\/edit?usp=sharing) While testing Whisper, Talon, and Nemo against the exact same test sets with the same text normalization, all of the large models performed well at general dictation. However, Whisper was painfully slow compared to the other models tested. Much higher output can be achieved when running GPU tests on the largest Talon 1B model and Nemo xlarge (600M) model than any Whisper model, including Whisper Tiny (39M). Whisper output is very good at producing coherent speech, even when it is completely incorrect about what was said. While analysing some ‘worst case’ outputs, neither Talon nor Nemo models showed worst-case results, anything like this. Most of Talon’s errors in this test set were compound word splits. An analysis of the paper found that, in general, at least for Indian languages, translations are better. However, transcriptions are suffering from catastrophic failures. In conclusion, Whisper is a very neat set of models and capabilities, especially the multilingual and translation use cases. It will also probably be a great tool to supervise the training of other models. However, given the observed failure cases, users may not use Whisper in production without a second model to double-check the output.","excerpt":"Trained on 680k hours of audio data, Whisper offers everything from real-time speech recognition to multilingual translation","categories":["Global Tech"],"tags":["Google","Meta","OpenAI"],"author_name":"Tasmia Ansari","publish_date":"2022-09-30T12:00:00","publication_year":"2022","word_count":696,"keywords":["Go","Meta","machine learning","TPU","OpenAI","AI","ML","Git","Google","transformer architecture","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","ML","OpenAI","TPU","R","Go","Git","GitHub","transformer architecture"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openais-whisper-is-revolutionary-but-little-flawed\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10165276,"title":"Cargill Plans Major Expansion with 500 New Jobs at Indian GCCs","content":"Cargill, a family-owned multinational company,  that provides food, ingredients, agricultural solutions and industrial products, has announced plans to increase its digital and technology workforce in India. As per reports, the company plans to add 500 positions over the next two to three years, increasing its total headcount to 3,500. The recruitment will focus on tech-based roles in data engineering, analytics, and artificial intelligence, primarily in Bengaluru. This expansion is separate from Cargill’s global restructuring announced in December, which included job cuts in sectors like supply chain and inventory controls, Reuters reported. The company aims to reduce its technology outsourcing from 80% to 40% within the same timeframe. Cargill operates two global capability centres (GCCs) in India. While the one in Bengaluru focuses on technology operations, another in Gurugram handles finance and human resource functions. The company collaborates with partners like Tata Consultancy Services (TCS) and Accenture. Meanwhile, Cargill’s Smart Manufacturing initiative modernises its North American protein operations using advanced technology and data. This programme aims to enhance workplace safety for employees while strengthening supply chain resilience for customers. Currently, Cargill has implemented over 100 ‘Factory of the Future’ projects across 35 facilities in North America. One example of this innovation is a 3D vision system that optimises the meat-cutting process in real time.","excerpt":"The company aims to reduce its technology outsourcing from 80% to 40% within the same timeframe.","categories":["AI News"],"tags":["GCC"],"author_name":"Shalini Mondal","publish_date":"2025-03-05T15:14:09","publication_year":"2025","word_count":213,"keywords":["artificial intelligence","GCC","programming_languages:R","AI","innovation","Git","Aim","data engineering","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","Aim","R","Git","data engineering","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cargill-plans-major-expansion-with-500-new-jobs-at-indian-gccs\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10059992,"title":"Bengaluru-based spacetech &#038; AI startup SatSure secures USD 5 Mn in Pre-Series A","content":"SatSure, a Bengaluru based startup that works at the intersection of spacetech, Artificial Intelligence (AI), and Software as a Service (SaaS), has raised USD 5 million in Pre-Series A led by Baring Private Equity India Pvt. Ltd. lowstate VC, Force Ventures, IndigoEdge Advisors, Toch.ai, Nishchay Goel and Saikiran Krishnamurthy participated in the round. The fresh funds will be used to expand SatSure’s footprint in Southeast Asia and accelerate the development of a product that includes launching proprietary payloads to Low Earth Orbit. “SatSure’s powerful analytics platform that combines satellite technology and operational data to minimise lending risk is unique and best placed to offer actionable and timely insights. Its proprietary AI-powered modelling empowers financial institutions to issue credit and insurance policies faster, in higher volume, and on less conservative terms, which unlocks financing for smallholder farms. We’re thrilled to partner with SatSure as the company expands its footprint in South and Southeast Asia,” said Jugnu Pati, Investment Specialist at ADB Ventures. SatSure offers three main decision intelligence products: 1) SatSure Sparta: A platform for providing agriculture and climate-related insights as an open innovation, freemium model. 2) SatSure SAGE: Life cycle risk monitoring and business intelligence platform for agriculture financial services. 3) SatSure SKIES: High-resolution satellite imagery-based infrastructure change detection platform. SatSure was founded in 2017 by IIST and ISRO alumni Prateep Basu, Rashmit Singh Sukhmani, along with Abhishek Raju. SatSure leverages satellite data, remote sensing, artificial intelligence, and big data analytics to provide decision intelligence solutions to the banking, financial services, and insurance (BFSI) sector.","excerpt":"SatSure taps satellite data, remote sensing, artificial intelligence, and big data analytics to provide decision intelligence solutions","categories":["AI News"],"tags":["ISRO","Low Earth orbit","SaaS","SatSure"],"author_name":"SharathKumar Nair","publish_date":"2022-02-07T13:05:29","publication_year":"2022","word_count":255,"keywords":["big data","Go","ISRO","Low Earth orbit","artificial intelligence","business intelligence","AI","SatSure","RAG","SaaS","analytics","decision intelligence","R","analytics platform"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","RAG","R","Go","big data","decision intelligence","business intelligence","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bengaluru-based-spacetech-ai-startup-satsure-secures-usd-5-mn-in-pre-series-a\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10136424,"title":"OpenAI is Two Stages Shy of Hitting AGI","content":"The discussions around AGI’s timeline and definition have been done to death. Now, the focus is on something more critical – mapping out the exact steps of how to get there. During T-Mobile’s Capital Markets Day 2024, where OpenAI and the mobile operator announced a multi-year partnership, CEO Sam Altman took the opportunity to discuss not only the o1 model but also OpenAI’s ongoing journey towards advancing AI and achieving its long-term goals – particularly, the stages of AI. Levels of AI Altman defined the five levels of AI development – the first stage being the chatbots, followed by reasoners, which the company says has now been achieved. The next phases are agents, innovators capable of scientific discovery, and fully autonomous organisations. “So this move from one [chatbots] to two [reasoners] took a while but I think one of the most exciting things about two is that it enables level three relatively quickly and the agentic experiences that we expect this technology to eventually enable, I think, will be quite impactful,” said Altman. With the depth of o1 capabilities, it looks like level 3 of AI is almost here. In the T-Mobile partnership, OpenAI will help build AI agents for customer service on the telecom operator’s platform IntentCX using OpenAI’s APIs and o1 model, which Altman claims to be at the GPT-2 stage of reasoning models. Next Stop: Level 4 The superior o1 model with improved reasoning and problem-solving capabilities, literally taking more time to think and process, has opened the doors for OpenAI’s gateway to AGI. The model is using reinforcement learning over an auto-generated chain of thought. “At first, most believed that LLMs couldn’t reach AGI without a new approach. But my mind changed based on the o1 model results. I’m now imagining a GPT-5 model with CoT reasoning, powered by specialised faster chips similar to Groq chips. Despite high costs, this could make AGI possible,” said an e\/acc user ‘Haider’ on X. In the launch video of OpenAI’s o1 model, researcher Noam Brown talks about how understanding AGI will only become plausible when one gets to see the models perform better, alluding to how OpenAI is on track. “I think for a lot of people it’s uh it’s hard to really fill the AGI and until you see the models do something better than humans can at a domain that you really care about,” said Brown. Source: X Autonomy Leads to AGI While Altman believes the third stage of AI is agents, enterprises have massively moved onto that spectrum by enabling a number of agentic workforces for their customers. In the past couple of weeks alone, cloud and SaaS giants such as Salesforce, Oracle and even Microsoft announced their AI agentic capabilities that were vertically integrated across their enterprise suite. Almost like competing with each other, Oracle announced 50+ AI agents, whereas Salesforce announced over 100+. It’s also interesting to note that while these announcements refer to agents that are only semi-autonomous at this stage, it is probably not wise to compare them to the agents that Altman might be alluding to. Considering o1 models’ superior reasoning capabilities, it is possible that their next level of AI is hinting towards fully autonomous agents. “The agentic experiences that we expect this technology [level 2] to eventually enable, I think will be quite impactful,” said Altman. Recently, Salesforce founder and CEO Marc Benioff showcased their version of AI at Dreamforce 2024. Interestingly, the 3rd wave was also agents followed by robotics and AGI. Source: Salesforce Youtube Altman has defined level four and five as innovators and organisations. At the innovators level, AI systems can not only run processes but improve them by thinking critically and independently, and organisations represent AI capable of performing the entire work of a company without human involvement. While Altman did not make any specific mention of when your AGI will be achieved, the path towards it is now clearer than ever. With level three, i.e. agents already being at play, the big tech is only two steps shy of achieving the ultimate goal.","excerpt":"Agents serve as the midpoint, and most have already shifted toward them.","categories":["AI Trends"],"tags":["Agents","AGI","Editors Picks","OpenAI","Oracle","Salesforce","Sam Altman"],"author_name":"Vandana Nair","publish_date":"2024-09-23T17:51:32","publication_year":"2024","word_count":679,"keywords":["Go","API","Agents","Sam Altman","OpenAI","AI","GPT-5","autonomous agents","chatbots","Oracle","Editors Picks","Aim","AGI","Salesforce","chain of thought","R"],"extracted_tech_keywords":["AI","GPT-5","OpenAI","Aim","chain of thought","autonomous agents","chatbots","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/openai-is-two-stages-shy-of-hitting-agi\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10067953,"title":"Startup’s loss is IT’s gain","content":"Looks like the great Indian startup party has come to a halt due to layoffs, shutdown, and funding crunch. According to market estimates in the last few months, the Indian tech startups, including a few unicorns such as Vedantu, Cars24, MFine, and Meesho – to name a few – have laid off close to 6000+ people as a measure to cut costs and extend their runway. But, the same is not the case with Indian IT companies as they are on a hiring spree despite the global downturn. Around 40-50 per cent of employees leaving startups are getting absorbed by IT companies, consulting and product companies and global captive centres (GCCs). IT, GCCs to the rescue The Indian IT companies are expected to hire close to 3.6 lakh freshers in FY22, as per cognitive intelligence platform UnearthInsight. For instance, IT giants Infosys and TCS recruited about 1.9 lakh freshers in the fiscal year 2022, indicating that the IT sector will continue to be on top of the game this financial year. Here’s a quick look at the total headcount of employees across IT companies, alongside the number of jobs added in FY22 Q4. India has about 1,500 GCCs across sectors. This includes banking, financial services, BFSI, automotive, IT software, etc. As per an estimate based on hiring plans of existing and upcoming GCCs, the Indian-based captive centres of MNCs are set to increase employee count by 1.8-2 lakh by the end of this fiscal year. By 2025, 500 new GCCs are expected to set up their captive tech, and the total headcount is set to double to 3-3.2 million in the next three years, as reported by ET. Further, the report stated that in the last one year, the cohort of companies together added about 1.7 lakh jobs in India, while gross hiring stood at 3.5 lakh. The rise of captive centres can be attributed to the cost advantages and huge talent pool that the country has – particularly in technology – making India a strategic hotspot for multinational corporations (MNCs). Are IT jobs stable? The increasing headcount of the majority of these IT indicates a steady growth in team size. That is because most companies are offering great incentives, promotions, stability and job security for their employees. For instance, in the case of TCS, the company believes that its focus on investing in people and its progressive workplace policies have helped them retain and hire more talent. HCL Technologies said that it had paid out a special milestone bonus to all of its employees in the March quarter of last year – i.e. close to USD 100 million – to be precise, USD 99.8 million. Of course, not to forget the Mercedes-Benz cars as a reward to top performers. Many employees said they would rather move to a stable organisation even if they look at a pay cut. But, at the same time, it is not that startups are not stable. Recently, various unicorns and pre-IPO startups have been hiring and growing their teams. Unfortunately, some of these companies have triggered negative sentiments among job seekers because of the layoffs. However, many experts believe that this will be corrected with time. We started with *Great Resignation* in 2021 and moved to *Great Recession* in 2022Layoffs in startups should add more workforce to the traditional IT companies which should cool down the attrition a bit#IT #GreatResignation— Vineeth K (@DealsDhamaka) May 24, 2022 Oh, the ‘Great Resignation’ Meanwhile, the Indian IT sector has also succumbed to increasing attrition. Shortage of digital talent, alongside high demand for qualified professionals, and increased wages, has led to increased attrition. As shown below, Cognizant has the highest attrition rate, followed by Infosys, Tech Mahindra, Wipro and others. However, the majority of the companies have reported some stability in quarterly attrition with better compensation and hikes, improved hiring, work-life balance, and others. But, the question is, what are these numbers telling us? Credence Wealth Advisors’ founder Kirtan A Shah said that the attrition leads to reducing margins for the IT companies. In other words, companies have to increase salaries to retain or hire new talent. Also, the utilisation falls. For instance, in FY22 Q4, Infosys’ utilisation dropped from 88.5 per cent to 87 per cent. Where is the talent heading? The founder and CEO of Han Digital, Saran Balasundaram, told The Economic Times that the demand is across various roles, including full-stack engineers, data engineering, product management, DevOps, etc. Currently, the overall headcount of the IT industry stands at 5.1 million, out of which 4.2 lakh people are employed with startups – almost 20 per cent of this workforce left startups to join the IT industry in the last few years, he added. As Shah pointed out, in the last two years, a lot of new-age tech startups were offering 2-3x salaries, ESOPs, etc., attracting talent from the old IT services companies. Now, it looks like they are returning to the IT world. “Startups’ may be ditched for stability,” said Shah. He said, along with a depreciating rupee, this can probably improve the operating margins for the IT sector going forward.","excerpt":"Around 40-50 per cent of employees are leaving startups and are getting absorbed by IT companies.","categories":["AI Startups"],"tags":["Cognizant","Infosys","Tata Consultancy Services","Tech Mahindra","Wipro"],"author_name":"Amit Naik","publish_date":"2022-05-27T12:00:00","publication_year":"2022","word_count":855,"keywords":["Tata Consultancy Services","Wipro","Go","Tech Mahindra","funding","unicorn","Infosys","AI","Cognizant","Git","data engineering","GAN","DevOps","R","startup"],"extracted_tech_keywords":["AI","R","Go","Git","DevOps","data engineering","GAN","startup","unicorn","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/startups-loss-is-its-gain\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24460,"title":"Google I\/O 2018: 13 Things Around AI And ML Announced This Year","content":"This year at their annual I\/O developer conference, Google took us on a tour of the future of artificial intelligence and what it holds for us. Unlike last year’s hardware-centric announcements, 2018 I\/O focused more on AI and how to get the most out of the Android devices. Prior to the keynote, the California-based company rebranded the whole of its Google Research division as Google AI, to focus more on computer vision, natural language processing and neural networks research and development. At Shoreline Amphitheater in Mountain View, California, Google CEO Sundar Pichai kicked off its annual I\/O developer conference by saying, “For someone like me who grew up without a phone, I understand how gaining access to technology can make a difference in your lives. There are very real and important being raised about the impact of the test advance and the role they ’ll play in our lives… We feel a deep sense of responsibility to get this right. That’s the spirit with which we are approaching our core machine to make information more useful, accessible and beneficial to society.” “AI is enabling for us to solve problems in new ways for our users around the world. With Google AI, we want to bring the AI benefits to everyone and we are opening AI centres around the world. AI is going to impact many fields.” Pichai added. Let’s take look at the big announcements that Google made on Day 1 keynote. The tech giant will unveil next big things announcements in APIs, SDKs, framework, etc for developers over the next couple of days. 1. Predicting Cardiovascular Risk Using AI Google AI is making huge improvements in healthcare. Last year, Google announced a work on diagnosing diabetic retinopathy, the leading cause of blindness, to help the doctor diagnose it earlier using deep learning. The company is running field trials since then at Aravind and Sankara Hospital in India. The same retinal scanning can also play an important role in predicting cardiovascular risk. Google’s AI can predict the five-year risk of you having an adverse cardiovascular event like heart attack or strokes. Also, the technology is bringing expert diagnosis to places where doctors are scarce. Google’s machine learning system can pick up on things that humans don’t. The company has published its paper on how AI is changing medicine. 2. AI For Accessibility Looking To Listen Pichai also showcased how AI can be used to write subtitles during debates. Google’s Looking To Listen is an audio-visual speech separation tool that uses ML to separate the audio from just one person from a video in which two people are debating or talking at the same time. 3. Morse Code Meets Machine Learning Google is adding Morse Code support for its Gboard keyboard. Morse code for Gboard comes with a setting that will allow users customise the keyboard to fit their particular needs. It can be used alongside Switch Access, which allows users to interact with Google’s keyboard though external devices. It will also include Google’s AI-driven text suggestions. The company released a trainer that will help users learn how to communicate with Morse code. Google also released a Morse poster so that the users can learn Morse code more easily, as well as text-to-speech app that incorporates Morse code. Watch Tani Finlayson’s Morse code story 4. Gmail Can Now Write Emails For You Google now announced a new feature for Gmail called Smart Compose. This new feature is powered by AI, to help you draft emails from scratch. From your greeting to your closing sentence, Smart compose suggests complete sentence in your emails so that you can draft them with ease. It will operate in the background, and it will offer you phrase suggestions as you type, just hit tab to select it and the text will auto-populate. Google will roll out the Smart Compose feature for consumers over the next few weeks and will be integrated for G Suite customers within the next few months. When it becomes available for consumers, make sure you have enabled the new Gmail in order to use it. 5. AI-Powered Google Photos Pichai introduced a new feature in Google Photos called Suggested Actions. It suggests actions you might want to carry out on a photo. For example, the Google Photo app can see when your friend is in a picture and suggest that you can share it with her. If the same photo is underexposed, Google’s AI systems offer a suggestion to fix the brightness with one tap. It uses AI to convert document photo to PDF format and can fix brightness. It automatically suggests exposure adjustments and improvements to your photos. The AI can also colourise old black-and-white photos. 6. Tensor Processing Unit 3.0 The third generation tensor processing unit is eight times more powerful than the previous generations TPU. Pichai said the new chips are so powerful that Google had to add liquid cooling to their data centres to compensate for the additional heat. This can handle up to 100 petaflops of machine learning computing. It will be primarily consumed through Google Cloud. 7. Natural Conversations With AI-Based Google Assistant Google Assistant will now come in six new voices. The Assistant also got a ‘continued conversation update’. Now, instead of having to say ‘Hey Google’ or ‘Ok Google’ everytime you want to say a command, you will only have to do so the first time. You can also ask multiple questions within the same request. 8. Google Duplex Powered By AI Google Assistant will soon make phone calls on your behalf. The new technology called Google Duplex can carry out real-world tasks over the phone. According to a statement by Google, the technology is directed towards completing specific tasks, such as scheduling certain types of appointments. For such tasks, the system makes the conversational experience as natural as possible, allowing people to speak normally, like they would to another person, without having to adapt to a machine. But the Duplex is still under development. 9. Google Maps With Computer Vision Google Assistant is also coming to Google Maps. The new feature will give you better recommendations for local places. Google also introduced Visual Positioning System in Google Maps for helping us with navigations beyond GPS. It will combine the camera, computer vision tech and Google Maps with Street View. 10. Google News Tailored By AI Google News is getting revamped with AI. You will get a briefing of the top five news stories with the latest local stories. News will use ML to understand your news preference and will show stories accordingly. Google also introduced a new feature called Full Coverage, which will show all relevant stories on a particular story from different sources. It will also feature ‘Subscribe with Google’ that will allow users to access their paid content everywhere on Google’s platform. 11. ML Teaches The ‘Joy Of Missing Out’ Google will introduce new features on Android devices for your digital wellbeing. It will understand users habits, help us focus on what matters, suggest people to switch off and wind down to find more time for our family. Android Dashboard It will slice and dice data of how the users are using their phone and tell you if you are overindulging on your phones. YouTube Notification Digest And Break Reminders With this two features, you will be able to cut on notifications and better monitor the time you spend on watching videos on YouTube. 12. Android P The latest version of Android will focus more on battery. Its new Adaptive Battery features AI to control which apps are running in the background in order to increase the battery life. In the latest version of Android, Google is also changing the user interface. The traditional three button bar will now be replaced by a single home screen button. When you swipe up on it, you will get a multitasking view of all recently used apps. It is also introducing few new features like Shush Mode, which silences your phone whenever you put your phone facedown on a surface. The Wind Down mode will automatically turn your screen to greyscale every night at a bedtime you decide and help you sleep better. Android P beta version is now available. 13. Machine Learning Kit For Developers Google announced machine learning kit for app developers who are not much proficient with machine learning. The new software development kit (SDK) supports text, image labelling, barcode scanning, facial detection, landmark recognition. The SDK for app developers on iOS and Android are both available online and offline, depending on developers’ preference and network availability.","excerpt":"This year at their annual I\/O developer conference, Google took us on a tour of the future of artificial intelligence and what it holds for us. Unlike last year’s hardware-centric announcements, 2018 I\/O focused more on AI and how to get the most out of the Android devices. Prior to the keynote, the California-based company […]","categories":["AI News"],"tags":["Gmail","Machine Learning","Sundar Pichai","Tensor Processing Unit"],"author_name":"Smita Sinha","publish_date":"2018-05-09T12:13:16","publication_year":"2018","word_count":1436,"keywords":["artificial intelligence","machine learning","Gmail","AI","Tensor Processing Unit","neural network","ML","TPU","Machine Learning","computer vision","RAG","deep learning","R","Sundar Pichai"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","computer vision","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-i-o-2018-13-things-around-ai-and-ml-announced-this-year\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":67231,"title":"Achieving Analytic Nirvana During Lockdown","content":"The world was hurtling towards destruction with environmental degradation, deforestation, increased carbon footprint and pollution levels, a system reboot was inevitable. Then, the COVID-19 pandemic bought the world to a standstill, forcing everyone to stay indoor and companies to adapt and evolve themselves to accept remote working as a norm. Overnight, business leaders and team managers had to change the way they did their work, decreasing reliance on in-person supervision. It also provided an opportunity for a relook at how they make decisions, moving from synchronous to a more reflective and an asynchronous way backed by data. For businesses, the future holds a more decentralised decision-making culture with more time spent on analysing the data. All this means that the roles and responsibilities of IT & Analytics teams will increase dramatically to support business continuity with the increase in the use of data and analytics. In a general business environment, we find three types of leaders, who will need data training to come up to speed with data-driven decision-making culture. The first type believes that experience and context are paramount and that data cannot capture the essence of the business. The second type of executives are more detail-oriented and do their analysis themselves. They will only reach out to the analytics team if they have more questions. The thirds type includes those who neither use too much data nor they accept that they need training. They work based on reports prepared by others. Analytics platforms must design should keeping in mind that all three types that co-exist in a workspace, are now working remotely. From an analytics evolution perspective, organisations fall in one of the five stages mentioned below. Those in the first two stages, where most organisations come under, see the current situation as a challenge and are trying to cope with it. Those organisations that are under the next three categories see the present scenario as an opportunity. Stage 1 – Remote working for these organisations is not an issue as long as the systems do notwn breakdown. The executives of these organisation have all the information they need to keep the business running remotely. Stage 2 – These organisations responded quickly to the changes and are demanding more data support now. Demand for their quarterly and monthly reports are currently on a monthly and weekly basis. Their data models are updated more frequently, and they expect more responsive support on ad-hoc analysis requests. They need shorter range forecasting now and are trying to replicate all their physical strategies into virtual strategies. Stage 3 – Organisations in this stage have a democratic view of data as the leadership believes that they cannot drive the ship alone and make decision-making decentralised. Executives have their personalised dashboards, and the executives are continuously pushing for more and more data literacy. These organisations are using data to search for newer ways of collaborating continually. Stage 4 – Organisations in this stage have evolved to a point where analytics get closer to human capability. The analytics inputs are provided more proactively to the executives, and the executives can even ask for their data on the go and get instant feedback. Moreover, the information is available at various touchpoints on demand, be it mobile, web, voice etc. Stage 5 – At this stage, organisations use data analytics and models as assistance to meet their business goal.  Business goals are fed into the analytics platform, which suggests strategies and tactics with constant guidance personalised to the extent that individuals can achieve their goal by simply following the guidance. In these organisations, analytics crosses predictive, prescriptive, cognitive and becomes experiential and act as a personal assistant or a co-pilot. The act of combining business analytics with experimental experiential learning helps us get to Experimental Experiential Analytics. Employees untrained in analytics will be able to use data specifically tailored for their need, identified by using AI and Machine Learning. Machine Learning Algorithm trains itself by experiencing the needs of the individual and then nudges the individual with solutions and information continuously until the goals are achieved. The customized nudges drive quick adoption and Experiential analytics can help executives better utilise information and make more data-derived decisions. The focus is on the delivery of personalised discoveries and insights for executives’ that also help driving better adoption. Creating a data-driven culture is a journey that cannot be reached overnight. Adopting an AI-based solution can stimulate behaviours that drive adoption even amongst the non-analytical workforce. During this lockdown phase embracing remote working culture by the non-data savvy workforce can catalyse the path for experiential analytics evolution and adaption to reach the fifth stage or the ‘Analytic Nirvana Stage’. The views\/opinions expressed in this article are the author’s own and do not reflect the views\/opinions of the organisation where the author is employed.","excerpt":"The world was hurtling towards destruction with environmental degradation, deforestation, increased carbon footprint and pollution levels, a system reboot was inevitable. Then, the COVID-19 pandemic bought the world to a standstill, forcing everyone to stay indoor and companies to adapt and evolve themselves to accept remote working as a norm. Overnight, business leaders and team […]","categories":["AI Features"],"tags":[],"author_name":"Vikash Raj","publish_date":"2020-06-12T14:00:00","publication_year":"2020","word_count":802,"keywords":["Replicate","Go","machine learning","AI","data-driven","ViT","analytics","GAN","R","analytics platform"],"extracted_tech_keywords":["AI","machine learning","analytics","R","Go","GAN","ViT","analytics platform","data-driven","Replicate"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/achieving-analytic-nirvana-during-lockdown\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10065819,"title":"All about Pentagon’s new AI chief","content":"The US Department of Defense has appointed Dr. Craig Martell as Chief Digital and Artificial Intelligence Officer (CDAO). “With Craig’s appointment, we hope to see the department increase the speed at which we develop and field advances in AI, data analytics, and machine-learning technology. He brings cutting-edge industry experience to apply to our unique mission set,” said deputy secretary of defence Kathleen H Hicks. The CDAO role was created to oversee data and AI initiatives under one official at the highest levels of the Pentagon. The CDAO reports to the deputy secretary of defence. Since February, John Sherman has been serving as the interim CDAO. Earlier in an interview, he said the CDAO office “is about decision advantage” and “getting ahead of these near-peer competitors, with the best insight and information that we can possibly have.” He also said the job of the incoming CDAO would be to “raise the waterline” for AI and accelerate digital development across military services and commands. As CDAO, Martell will be responsible for leveraging AI and analytics to give commanders and decision-makers the capabilities to stay ahead of China- ‘the Pentagon’s pacing challenge’. Why does the Pentagon need an AI chief? The US military’s AI use was initially focused on experiments rather than autonomy on the frontlines. The 2019 National Defense Authorization Act mandated the DOD to designate a senior official to drive strategic AI adoption and execute a transition plan. The Defense Innovation Board had also recommended the deputy secretary of defence create a senior-level committee to oversee DOD’s AI strategy and ethical principles. In 2021, the department announced the creation of the CDAO role, with a view to reach full operational capability by June 2022. “The US is reorienting its geopolitical strategy around technology. This strategy rests on pillars like defence, alongside foreign policy, trade, etc. The Pentagon’s AI officer will be a critical role in this regard. This person will oversee the projects, like autonomous weapons, that help the US take on future challenges. There are many implications. One of the biggest is that the US is making AI the new “core” of its military, instead of treating it as a peripheral variable. This means the US could start deploying AI across the military spectrum, adding a new dimension to the geopolitics of AI,” said Abishur Prakash, geopolitical futurist and ​​co-founder at Center for Innovating the Future (CIF). The office of CDAO comprised Joint Artificial Intelligence Center- responsible for the implementation of AI within the department, the office of the Chief Data Officer- in charge of data management, the Defense Digital Service- responsible for finding solutions for internal data and security issues, and the Office of Advancing Analytics- the department that aggregates data and conducts data analytics. In December 2021, a Pentagon official discussed the inefficiencies of having different autonomous bodies, such as a lack of symbiosis across the JAIC, the DDS and the CDO. Who is Craig Martell? Craig Martell has served as the head of ML at Lyft, Dropbox and Linkedin. He holds a PhD in Computer Science from the University of Pennsylvania. However, Martell’s exposure to the US military is limited to his service as a Professor of Computer Science at the Naval Postgraduate School overseeing the NPS Natural Language Processing Lab for 11 years. He also acknowledged this in an interview, saying the DoD needed someone agile. He said, “I don’t know my ways around the Pentagon yet, and I don’t know what levers to pull”. The department needed someone from the industry who knows how to bring AI and analytical value to the table at scale and speed. As CDAO, Martell’s first order of business will be to identify the “marquee customers” and the systems his office will need to improve. His budget for fiscal 2023 is USD 600 million. Responsibilities The Pentagon memo in February chalked out the CDAO’s responsibilities. 1) CDAO, in coordination with USD(R&E) and USD(A&S), will drive DoD’s adoption of data, analytics, and AI from operational prototyping to operations. 2) CDAO will oversee Responsible AI policy and Al Assurance policy in collaboration with USD (R&E) and DOT &E. 3) In collaboration with USD(A&S), CDAO will foster industry engagement on data, analytics, and AI. 4) CDAO will offer technical support to USD(A&S) on data, analytics, and AI technology sharing and program cooperation with allies and partners. 4) CDAO will drive international engagements to facilitate AI readiness and responsibly use data, comport with data sharing requests and requirements, and incorporate Al capabilities at speed and scale. 5) CDAO will support USDC\/CFO by providing access to the enterprise financial data, process automations, and reconciliations required to enable the audit of the Department’s financial statements. 6) CDAO will ensure Advana and other audit relevant data systems under its control pass audit.","excerpt":"Since February, John Sherman has been serving as the interim CDAO.","categories":["IT Services"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-04-28T11:06:46","publication_year":"2022","word_count":795,"keywords":["Go","artificial intelligence","AI","ML","Git","RAG","responsible AI","automation","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","RAG","R","Go","Git","responsible AI","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/all-about-pentagons-new-ai-chief\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":59756,"title":"You Cannot Become A Data Scientist Without Mastering The Storytelling Skill, Says This Data Scientist","content":"Since data science is one of the highly-paid jobs, professionals from different fields want to switch their careers to data science. However, due to the strenuous nature of the field, people struggle to make that final stride. This is mainly because aspirants mostly focus only on various data science tools and techniques. Instead, they should also be apt in storytelling and other business aspects of the domain to crack the interview. Consequently, there is a need for effectively devising strategies to land a job in the data science landscape. To help aspirants in understanding such requirements of the data science landscape, every week, we interview prominent data scientists and get inputs from the ones who have successfully made it to the data science space. For this week’s column — My Journey In Data Science — Analytics India Magazine got in touch with Pavan Kumar Thatha, Emerging Technologies Leader at Unisys. He has over 15+ years of experience in the IT industry with some of the leading companies such as Wipro and Kloud9 as an AI architect and Chief Data Scientist, respectively. The Onset Having completed his bachelors in electronics and communication engineering in 2002, Pavan started his professional career in embedded systems and digital signal processing. In his first ten years, Pavan’s primary day-to-day activities involved system programming, middleware applications, and networking. However, Pavan pinpoints two reasons that lead him to get into the data science landscape in 2014. While working as a system programmer, he was involved in a project that required him to collect data from various sources and come up with statistical analysis. This is where he assimilated the potential of data for solving business problems. Besides, he was also motivated by his colleagues who were working with data to gain insights for the company. Career Switch Into Data Analytics “Back then, there wasn’t much awareness about the data science field, but I was intrigued by the domain and started exploring the space on the internet. Eventually, I found various courses, and that is where my data science journey started,” says Pavan. “I immediately fell in love with data science as it doesn’t demand expertise in any vertical of business. It is about understanding the concept and business models of a company, which is seldom a case with other technologies.” However, he didn’t directly look for a data analytics or data science job. Instead, he continued with his system application programming role and joined Wipro in 2014. Although he was hired as an architect in the application programming, he convinced his manager to make an internal switch into the analytics department. But, it wasn’t straightforward for him to veer into the data science space. He devised a plan and asked his manager to allow him to split the time between application programming and data science. Pavan used to spend 80 per cent of the time doing his primary job and another 20 per cent in data analytics. And within the first five to six weeks, he demonstrated his analytics skills with a proof of concept (POC), which caught the attention of the higher officials who then allowed him to make a switch into the data analytics field. Learning The Facet Of Data Science The Hardway After gaining an opportunity, Pavan continued his learning and implemented the techniques to gain insight into various types of data. However, to further learn advanced machine learning techniques, he enrolled in a post-graduate business analytics program of Great Lakes Institute of Management in 2016. This empowered him to learn from industry and academia leaders, thereby gaining a wide range of perspectives of the domain. Pavan, after completing his program in 2017, he was motivated to take another stride and get into data science. But, he struggled to crack interviews. Of the many failed interviews, Pavan explained how an interview at Gartner changed his perspective, which then led him to revamp his preparation strategy. “At Wipro, as a data analyst my role was only to find insights, and when I attended the data science interview at Gartner, I realised that I was not a good storyteller, which was limiting me from explaining the implications of the data analytics outcomes on businesses,” explains Pavan. “That’s when I realised that one cannot become a data scientist without mastering the storytelling skill.” The setback, in contrast, pumped him to enhance his storytelling skills. He started working on connecting the dots and completing the loop to make effective business decisions. Eventually, he got a job at Kloud9 as a chief data scientist. After a short stint at Kloud9, he then moved to Unisys where he is currently working as Emerging Technologies Leader. Worst Experience In Data Science Talking about his worst experience in data science, Pavan said that he demonstrated numerous POC at one of the companies, which would have helped the firms had it been implemented. But, due to the gap in the understanding of the potential by the management, it didn’t get the approval. “At times, it is difficult to convince non-technical people about the use cases and possible benefits of the data science insights. This is exactly what happened with me three times in a company,” adds Pavan. “My models were not taken in the production. Consequently, I consider it the worst experience.” Current Work Experience Pavan currently works with gigabytes and terabytes of data using various tools such as Spark, Databricks and more. Besides, he also hires data scientists for various roles but doesn’t focus on certifications at all. “PhD and Masters doesn’t necessarily provide an edge. Rather, we seek applicants who are self-learned, understand business problems and solve them,” says Pavan. Besides, he also looks for job seekers who are proficient in cloud computing. Pavan believes, today, expertise in cloud computing is essential for data scientists to thrive in the competitive landscape. Advice To Aspirants For aspirants, Pavan advises that they should focus more on hands-on experience and not just read or watch videos to learn technical skills. To do this, one must actively participate in various hackathons to test their knowledge. However, he pointed out that participation is crucial than winning it. A few people are not enthusiastic about hackathon because they have self-doubts. This stops themselves from taking part in various competitions. Besides, he suggests that one cannot master every technology within data science. Consequently, one should pick one technology like computer vision or NLP and master that while being aware of other technologies within the data science landscape.","excerpt":"Since data science is one of the highly-paid jobs, professionals from different fields want to switch their careers to data science. However, due to the strenuous nature of the field, people struggle to make that final stride. This is mainly because aspirants mostly focus only on various data science tools and techniques. Instead, they should […]","categories":["AI Features"],"tags":["data analyst certification","Data analyst jobs","how to become a data scientist","Interviews and Discussions","My Journey In Data Science"],"author_name":"Rohit Yadav","publish_date":"2020-03-24T18:00:00","publication_year":"2020","word_count":1079,"keywords":["data science","Go","how to become a data scientist","machine learning","AI","cloud computing","My Journey In Data Science","Data analyst jobs","computer vision","NLP","Databricks","analytics","R","data analyst certification","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","NLP","computer vision","data science","analytics","cloud computing","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/you-cannot-become-a-data-scientist-without-mastering-the-storytelling-skill-says-this-data-scientist\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10131696,"title":"How AI and Analytics Helped Manu Bhaker Bag Bronze Medals at Paris Olympics","content":"India’s pride, 22-year old Manu Bhaker created history by winning two bronze medals in the 10m air pistol shooting event at the ongoing Paris Olympics 2024, which comes to an end on August 11. Bhaker not only became the first female shooter to win a bronze at the Olympics, but also the first Indian woman to win two medals at the prestigious sporting event. While Bhaker’s hard work paid off, the path to getting there involved not just immense practice, but also tech and analytics. Exemplary Rise of Manu Bhaker, born in Goria village of Haryana, won a gold medal at the 2022 Asian Games in the women’s 25m pistol team event and a record-setting individual gold at the 2018 Commonwealth Games in the women’s 10m air pistol event. At 16, she became the youngest Indian to win gold at the ISSF World Cup. However, the journey was not all easy, with a major technical glitch that deprived her of an Olympic medal earlier. In the Tokyo 2020 Olympics, a pistol malfunction had cost her a chance at a medal. The incident pushed her to think of quitting the sport altogether. However, with the support of her coach Jaspal Rana, Manu continued. Despite her diverse sporting background in tennis, skating, boxing, and even music (violinist) Bhaker chose to focus on shooting at 14, with support from her family. Her journey from a small village to international acclaim has brought into focus the role that tech and analytics played in her success. Ace Shooters Swear by AI and Analytics At MLDS 2023, former Indian Olympic gold medalist and retired sport shooter Abhinav Bindra spoke to AIM about the integration of analytics in sports. “Technology and analytics play an extremely important role, especially in today’s day and age. I mean, there are lots of different technologies available to analyse sporting performance, movement, and form,” he said. Bindra believes that AI will have a crucial role to play in the whole gamut of sports and analytics. He said that in a few years, AI will take over tasks currently handled by data scientists, and perform them in seconds. “From an analysis point of view it’s going to have a massive impact there,” he said. Data is Everything Last year, Pierre Beauchamp, the high-performance director for the National Rifle Association of India, emphasised the importance of data management and the need to convert data into usable information for coaches. “Data management is an area that was previously not looked at or not so seriously, at least, for shooting in India. I am looking at numbers in everything and am heavily interested in data analytics and how it can support decision-making for the organisation,” he said. Beauchamp also believes that when decisions are data-driven and include input from athletes and coaches, the outcomes improve. Without this, accurate assessments become challenging. Manu Bhaker with Pierre Beauchamp (left) and Sarabjot Singh, with whom she won bronze under mixed category at the Paris Olympics 2024. Source: LinkedIn Data Analysis for Shooting In a study conducted on air rifle elite shooters from China, who have had a long history of winning at the Olympics, key factors such as hold, aim and trigger control were analysed. Data collected from 60 shots per shooter showed that the hold factor is central, influencing both the aim and trigger control, and ultimately, the shooting result. The findings suggested that the interaction between these factors determined shooting success, providing valuable insights for training and performance enhancement. With increased focus on data analytics, it is no surprise that even big-tech companies are investing in sports analytics. For instance, Formula 1 racing uses AWS for insights and analysis to improve the game. With two bronze medals at the Paris 2024 Olympics, along with gold medals at the World Championships, Asian Games, Commonwealth Games, and Youth Olympic Games, Manu Bhaker has become the most successful Indian woman shooter in history at a very young age. With improved performance at the recent games, the role of AI and data analysts for competitive sports in India is only going to accelerate.","excerpt":"Former Olympic gold medalist Abhinav Bindra had already predicted AI’s significant role in sports analytics.","categories":["AI Trends"],"tags":["Advanced Analytics"],"author_name":"Vandana Nair","publish_date":"2024-08-07T10:02:26","publication_year":"2024","word_count":682,"keywords":["Go","AWS","AI","cloud_platforms:AWS","data-driven","ML","Aim","analytics","GAN","R","Advanced Analytics"],"extracted_tech_keywords":["AI","ML","analytics","Aim","AWS","R","Go","GAN","data-driven","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-ai-and-analytics-helped-manu-bhaker-bag-bronze-medals-at-paris-olympics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10008610,"title":"Yet Again, OpenAI’s GPT-3 Proved To Be A Doom For Humanity","content":"With last week’s major news, we had learnt that in an attempt to extend its partnership with OpenAI, Microsoft acquired an exclusive license to GPT-3. Although the company hasn’t stopped the API access of the model to other users, it indeed provided the access of its underlying codes and mechanism to Microsoft. This would allow the company to leverage the state-of-the-art technical innovations of OpenAI to create advanced products for its customers. This outlandish indeed would provide an edge to Microsoft over its competitors in the market like IBM, Google, AWS, etc. Not only can it make them lead the AI war, but it can also help in modifying its offerings like Office 365, Bing, as well as Cortana to bring in more customers. In fact, this would also enable OpenAI to make a commercial profit out of the massive investment it has done to create a product like GPT-3. Further would also allow the company to improve the model, based on the interaction it will have with a massive user base of Microsoft. Although the industry is going gaga over the announcement, the deal has undergone a massive dejection from many experts. Considering that launching GPT-3 was an act to ensure “artificial general intelligence to benefit everyone”, providing elite licensing to Microsoft raised significant concerns among the industry. Also Read: Has OpenAI Surpassed DeepMind? GPT-3 — A Harbinger Of Doom Being the massive model in the world, GPT-3 comes with immense possibilities of AI applications with its extraordinary capabilities of creating human-like written text. In the recent news, The Guardian wrote an article using this language model, where it showcased its prodigious ability to somewhat come up with a coherent essay about artificial intelligence. It had also proved its competencies when a college kid created a fake blog using GPT-3, making people believe that humans wrote it. Such applications highlight that GPT-3 has immense possibilities if applied for logical applications. However, giving exclusive access to Microsoft can create severe ramifications and potential concerns for the industry. Once the announcement was out, even Elon Musk, one of the OpenAI founders, tweeted its criticism on the openness of the company. Following the Microsoft news, this tweet showcases Musk’s disapproval of the deal. This does seem like the opposite of open. OpenAI is essentially captured by Microsoft.— Elon Musk (@elonmusk) September 24, 2020 With OpenAI’s initial launch, the company was established as a non-profit organisation to use AI for augmenting humanity as a whole, but with Microsoft leveraging it for its products, which hasn’t been explained well, it will just be profitable for the tech giant. And that’s why Musk critically disregarded this move as it contradicts the company’s vision of “build value for everyone rather than shareholders.” In fact, in 2018, Musk moved out of the company due to some corporate conflicts, and with the OpenAI’s recent move, Musk would be pleased with his decision. Joining Musk, there are many experts on Twitter, who have shown their disapproval of this deal, and OpenAI’s vision statement. Also Read: How Guardian’s Recent Article Is Yet Another GPT-3 Hype On a similar note, Karen Hao of MIT Technology Review has also scrutinised the move. She wrote — while the non-profit organisation “was supposed to benefit humanity,” it is currently helping only the tech giant. With the company starting to commercialise its products, it clearly draws attention to the profitable interest that OpenAI has with GPT-3. Not only is it a money-making process for the large tech companies but can also manipulate and shape the AI industry as a whole. While Microsoft stated that this queer deal would democratise AI and promote AI at scale, It didn’t shed much light on how much exclusivity this deal provides. Thus it would be difficult to establish how Microsoft is going to empower users and businesses with its AI platform. But it indeed can be confirmed that the company will leverage GPT-3 for improving its search engine along with its other products like MS Teams and Office 365 to bring in more customers — surpassing Google and Apple. Additionally, such an exclusive deal can create many problems — starting from ownership of data to influencing the product plan to protect their own interest. It is also expected that, in future, Microsoft might dictate the pricing plan of the API, to dominate its competitors in creating advanced products using the technology. And with Microsoft making profits for OpenAI, it will force the company to be more dependent on Microsoft’s lead. Also Read: OpenAI’s Not So Open GPT-3 Can Impact Its Efficacy Wrapping Up GPT-3 and its evolution has critically been scrutinised in many instances. However, being a novel technology with no prior reference point, it indeed would have been a challenging task for OpenAI to find the right target audience to commercialise the product. Thus, licensing Microsoft, who had been a long-standing partner with Open AI, was a right move for the company. Despite the disconcert this move has brought in the industry, with this deal, Microsoft will be able to introduce some real-world applications of the extraordinary model. Further, it would also benefit Microsoft users who will be able to utilise its GPT-3 capabilities, free of cost. With that being said, this deal changes the entire ball game for companies like Google, Apple, Facebook, as well as Alibaba, Baidu etc. who have been spearheading the AI war.","excerpt":"With last week’s major news, we had learnt that in an attempt to extend its partnership with OpenAI, Microsoft acquired an exclusive license to GPT-3. Although the company hasn’t stopped the API access of the model to other users, it indeed provided the access of its underlying codes and mechanism to Microsoft. This would allow […]","categories":["Global Tech"],"tags":["GPT-3"],"author_name":"Sejuti Das","publish_date":"2020-09-30T13:00:42","publication_year":"2020","word_count":899,"keywords":["GPT-3","Go","API","artificial intelligence","OpenAI","AI","AWS","RAG","GPT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","OpenAI","RAG","AWS","R","Go","API","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/yet-again-openais-gpt-3-proved-to-be-a-doom-for-humanity\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45915,"title":"5 Companies That Are Thriving After Open-Sourcing Their AI Tools &#038; Algorithms","content":"Open-sourcing has become a tradition for big tech companies today. This not only helps the developers in the community to gain an understanding of the innovative technologies used by the tech giants but also helps the companies to find bugs and make enhancements to the software. In this article, we list down 5 companies who have been open-sourcing their algorithms and software into the developer community. 1| LinkedIn Last month, the professional social networking site open-sourced a machine learning library known as the Isolation Forest. The library is being used by the Anti-Abuse AI Team at LinkedIn creates, deploys, and maintains models that detect and prevent various types of abuse, including the creation of fake accounts, member profile scraping, automated spam, and account takeovers. The developers at LinkedIn created a Spark\/Scala implementation of the Isolation Forest unsupervised outlier detection algorithm and is currently available on GitHub. This library supports distributed training and scoring using Spark data structures. This library also supports model persistence on the Hadoop Distributed File System (HDFS). The Isolation Forest algorithm is mainly chosen for vat=rious reasons such as this algorithm is a top-performing unsupervised outlier detection algorithm, scalable, low memory requirements, among others. 2| Microsoft In March 2019, tech giant Microsoft open-sourced Project Zipline compression algorithms, hardware design specifications, and Verilog source code for register transfer language (RTL) with initial content at the Open Compute Project (OCP) Global Summit 2019. The researchers at Microsoft developed a cutting-edge compression algorithm and optimised the hardware implementation for the types of data that are found in the cloud storage workloads. Project Zipline compression algorithm yields result up to 2X high compression ratio which is better than the result of commonly used Zlib-L4 64KB model. Project Zipline is a cutting-edge compression technology optimised for a large variety of datasets, while RTL allows hardware vendors to use the reference design to produce hardware chips to allow the highest compression, lowest cost, and lowest power out of the algorithm. 3| Facebook A few weeks ago, the popular social networking site open-sourced image and video algorithms. The photo-matching algorithm is called PDQ and the video-matching technology is called TMK+PDQF. Facebook announced the two technologies during the child safety hackathon which will help in detecting the graphical abusive contents, child exploitation, terrorist propaganda and other such threats. These algorithms will be available on GitHub and are part of a suite of tools that Facebook uses to detect harmful content. The two technologies work in a method such that they will store the files in the form of short digital hashes and later comparing them with other instances in order to determine whether the files are identical and nearly identical images as well as videos to fight abuse on the internet platforms. 4| IBM In July 2019, big blue open-sourced a deep learning algorithm known as PaccMann, stands for Prediction of anticancer compound sensitivity with Multi-modal attention-based neural networks. The goal of open-sourcing this algorithm is to deepen the understanding of cancer to equip industries and academia with the knowledge that could potentially which will help fuel new treatments and therapies. The researchers applied this algorithm to predict the sensitivity of cancer cell lines to known drugs and it achieves a superior predictive power compared to existing algorithms Last month, the big blue took another huge step by open sourcing the POWER Instruction Set Architecture (ISA). The expectation behind this open-sourcing is to boost the IBM Power processor’s value by creating innovative hardware components. Besides this, the tech giant also contribute other technologies including a softcore implementation of the POWER ISA, as well as reference designs for the architecture-agnostic Open Coherent Accelerator Processor Interface (OpenCAPI) and the Open Memory Interface (OMI). 5| Google In 2015, Google open-sourced the software library for TensorFlow which is an end-to-end open-source platform for machine learning. The idea behind this open sourcing is to let the machine learning community such as everyone from academic researchers, to engineers, to hobbyists—exchange ideas much more quickly, through working code rather than just research papers. Conclusion Open-sourcing tools and systems are making a huge impact on these companies. Previously, when there were almost no open-source systems or toolkits, companies created their products which had close data, it was difficult as well as time-consuming to enhance it or remove the errors and bugs. Currently, the potential of business relies upon sharing software source code ann which is why tech giants like Microsoft, Google, IBM, among others are open-sourcing their intelligent software. Also, AI is at its infancy stage and open-sourcing will not only make it easier to develop but also the developer’s community gets a chance to learn and understand what’s going behind the emerging technologies.","excerpt":"Open-sourcing has become a tradition for big tech companies today. This not only helps the developers in the community to gain an understanding of the innovative technologies used by the tech giants but also helps the companies to find bugs and make enhancements to the software. In this article, we list down 5 companies who […]","categories":["AI Trends"],"tags":["AI Companies","Google","IBM","linkedin","Microsoft"],"author_name":"Ambika Choudhury","publish_date":"2019-09-13T13:30:36","publication_year":"2019","word_count":781,"keywords":["Go","machine learning","AI","neural network","Scala","Git","RAG","deep learning","IBM","Google","AI Companies","linkedin","TensorFlow","R","Microsoft"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","TensorFlow","RAG","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-companies-that-are-thriving-after-open-sourcing-their-ai-tools-algorithms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":32628,"title":"Going To Mars Is Not A Challenge For AI, Hints SpaceX CEO Elon Musk","content":"Elon Musk has always expressed his interest and optimism in setting up a human colony on Mars. The tech billionaire has envisioned to blast people on the first human voyage to Mars by the year 2024. Now he recently suggested that artificial intelligence could possibly be the first resident on Mars. When asked about his thoughts over intelligent machine being first Martian rather than a human in a tweet, Elon Musk reciprocated: “30%” 30% — Elon Musk (@elonmusk) December 27, 2018 How Musk came up with the percentage is still unknown, according to the Geek, many people are already wondering what an AI resident of Mars would look like. “It’s possible that it could be a rover-like bot that explores the planet, or a stationary device that makes observations and conducts experiments without human assistance,” the Geek report added. The idea of colonising Mars is a very far-fetched dream with the current technology that we have. Going to Mars is one thing, and setting up a colony is another. But Elon Musk has, through SpaceX, proved that he has the potential to make such a scientific dream a reality. Researchers all over the world are working towards technologies to enable humans to venture beyond the atmosphere of the moon. But there is plenty of time for this to become true. Very recently, in November 2018, Musk declared that there is a 70% chance that he will personally go to Mars in an interview with Axios. The founder and CEO of SpaceX was thinking out loud about whether he would hitch a ride on one of his company’s rockets. SpaceX has flown from milestone to milestone over the past year and is preparing an attempt to put humans in orbit next year for NASA. US space agency NASA also firmed up its plans to return humans to the Moon and use its lunar experience to prepare to send astronauts to Mars in the mid 2030s.","excerpt":"Elon Musk has always expressed his interest and optimism in setting up a human colony on Mars. The tech billionaire has envisioned to blast people on the first human voyage to Mars by the year 2024. Now he recently suggested that artificial intelligence could possibly be the first resident on Mars. When asked about his […]","categories":["AI News"],"tags":["Elon Musk","Space","spacex"],"author_name":"Disha Misal","publish_date":"2019-01-02T09:00:44","publication_year":"2019","word_count":324,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Elon Musk","spacex","Space","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/going-to-mars-is-not-a-challenge-for-ai-hints-spacex-ceo-elon-musk\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":44835,"title":"What If Artificial Intelligence Becomes Ransomware’s Sidekick?","content":"First time since 2013 we have witnessed a decrease in ransomware activity, with the overall number of ransomware infections on endpoints dropping by 20%. However, ransomware like WannaCry, copycat versions, and Petya, continued to inflate infection figures. But when these worms are stripped out from the statistics, the drop in infection numbers is steeper: a 52% fall. However, 2019 doesn’t seem to be as good as the last. Even though the numbers of ransomware activity decreased, the headlines are still coming about ransomware affecting enterprises. There was a time when ransomware was focused on the consumers, but in 2017, the focus shifted to enterprises and in 2018, that shift accelerated, and enterprises accounted for 81% of all ransomware infections. However, it’s not just private companies that are under threat. Recently, in one of our articles, we listed some of the top targets of ransomware and of them was Government firms. In fact, a recent incident in Texas proved that Government organisations are actually under the radar of cyber-attackers. Texas Ransomware Attack Recently, hackers attacked 23 organisations connected to local government in the US state of Texas with ransomware. The attack was so serious that it disabled email accounts and prevented online payments to city departments for weeks. Even though the type of ransomware is not revealed yet, and no state networks were compromised in the attack; reports suggest that the attack was carried by one single threat actor. The officials from Texas have stated that investigations into the origin of this attack are ongoing, but as of now, they are more inclined towards response and recovery and get things back to normal. The damages from ransomware are brutal. For example, the Baltimore incident in May 2019, where hackers seized control of thousands of government computers in, ended up amounting to $18 million in direct costs and lost revenue. How AI Could Make Ransomware More Lethal Hackers unleashing ransomware attacks on these really big targets is definitely a thing to worry about. However, what could be worse is what if ransomware attacks gets powered by artificial intelligence. It would be a completely new, power-packed makeover for some of the notorious ransomware and there are chances that these worms would evade any cyber defence into computer networks and create havoc. The whole industry is moving towards A.I. for protection. It’s no surprise that AI and ML through the years have become something really incredible for the cybersecurity industry — from detecting threats to mitigating risks. But these sought-after technologies are like a double-edged sword, and once in the hands of threat actors, the table might turn. Ramifications The worse could happen when these sought-after techs reach the consumer level adoption. Imagine ransomware that is powered by machine learning and has the capability to learn from defensive responses and start pwning and exploiting way faster than a defending system. However, that is not the only way AI and ML could be used when it comes to ransomware attacks, there are other potential methods and strategies as well. Deepfakes, which is already one of the most notorious threats, could also play a role in pushing ransomware to the next level. They can be used to land video calls posing as the boss and ask any employee to carry out a task. This could be a way of spear phishing that would later result in making a way for ransomware to infect systems. Also, it would not only be in the form of a video call, but also in other forms of communication. Hackers would be able to create thousands of malware-loaded, fake messages at a much faster pace without tiring. That is not all, AI and ML today have the capability to bypass CAPTCHA too. There are instances where technology professionals have published their work where they have shown how the CAPTCHA can be easily broken using machine learning and deep learning. And there are many companies and organisations that rely extensively on CAPTCHA to determine if there is any non-human intervention. Bottom Line One cannot emphasise enough on the fact that technological advancement always works for both sides of a coin. If IT security professionals are using some of the most advanced techs to forecast attacks, even threat actors are making the best of the same tech to stay one step ahead. This race between the white hat hackers and the black hat hackers will continue for the years to come, and the result would always depend on who leverages what technology and when.","excerpt":"First time since 2013 we have witnessed a decrease in ransomware activity, with the overall number of ransomware infections on endpoints dropping by 20%. However, ransomware like WannaCry, copycat versions, and Petya, continued to inflate infection figures. But when these worms are stripped out from the statistics, the drop in infection numbers is steeper: a […]","categories":["AI Features"],"tags":["Cyber Security","is artificial intelligence the next big thing","Ransomware"],"author_name":"Harshajit Sarmah","publish_date":"2019-08-22T11:00:45","publication_year":"2019","word_count":751,"keywords":["Go","artificial intelligence","Cyber Security","is artificial intelligence the next big thing","AI","machine learning","Ransomware","ML","RAG","deep learning","ViT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","RAG","R","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-if-artificial-intelligence-becomes-ransomwares-sidekick\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":36496,"title":"Women Face Behavioural Biases, The Best Way To Overcome Is To Stay Assertive: Toshi Prakash, Locus.sh","content":"Toshi Prakash, who is the VP Product at Locus, has been working as a product manager for the last five years and had worked as the software development lead for eight years before that. With extensive experience in building sustainable and scalable product design, she has built both large-scale enterprise-level products and small fun apps from scratch, handling every aspect of the life-cycle from idea to release, growth and demise. In her current role, she heads the product at Locus that use AI, ML to optimise the supply-chain logistics of enterprises. She has earlier worked for Freshmenu and Nestaway to help them build the products from scratch and scale it. In this interview, she shared her bit of challenges that she faced being a woman in tech background and how she overcame it. Analytics India Magazine: Have you ever faced gender disparity in your career in analytics and data science? How did you fight the obstacles? TP: Like every woman, I have faced the typical behavioural biases. I have been cut-off in meetings, have been discredited from the ideas that I brought to the table,  and have been told that I am rude and intense; even when aggression from the men in the same room goes completely unnoticed. I have been over-looked for promotions despite being quantifiably more deserving because my ‘network’ wasn’t significant. To counter-balance the lunch and smoke talks that my peers had with superiors, I had to make extra and very formal efforts to showcase the work I do. My approach is to be assertive yet calm in such situations. It takes a little extra effort in figuring out alternative ways than the usual ones that are presented to us, but it’s not impossible. Good work and talent always shine through. AIM: Research suggests that girls are less likely to study STEM subjects. How can we inspire young girls to get involved with technology careers? TP: It seems very paradoxical. We see that during early school years girls performing at par and in equal numbers with boys. But somehow, when it comes to choosing careers, they choose non-technical courses. I believe that this has a lot to do with the guidance they get when making these choices and what role-models are in front of them. If they can meet, talk and see women in STEM professions on equal footing as men, any naïve bias developing in their minds can be curbed. In particular, children learn more from what they see than what they are told. In schools and in media, we should bring as many women on the dais as men, and not just once or twice a year for a particular event. Subtle biases of cutting a lady off when talking should be controlled in any discussion. Even in family gatherings, they should be equally encouraged to participate in deeper discussions on career, politics, sports; and not just household stuff. AIM: What would be your advice for fellow women professionals who are looking to switch or start a career in data science? TP: Frankly, my advice here would be the same as that I would give to any person: If you like it and look forward to the time working on it, go for it. Passion should be the driving force and challenges just make us stronger. As a field, data science is mathematical yet exploratory and thus ideally suited for women who love finding answers. AIM: Do you think there is an unconscious bias in recruiting women in technology?  If so, what are the ways we can overcome it? Have you faced it in your analytics and data science career? TP: Over the past ten years, this has become less prominent, but the bias persists. Even today, a female candidate would probably not be offered certain roles, if she has plans to start a family soon. Equal pay for women is still a part of very few organisations, and the perception remains that women would not be working as many hours as men. Most of the times, the data on hours of work, refuses to account for the informal breaks taken during the day. Let us be aware of our biases and try to find ways to judge based on data before forming an opinion objectively. If the bias is based on hard work put in, we should build in measures to judge everyone’s productivity in a granular manner and make that the basis of any judgment call. Forthcoming maternity leaves should not be a reason to reject current candidature as one never knows what happens in anyone’s future and what type of leaves we may need. AIM: Is there a need to re-start programs by leading MNCs to help women get back into the workforce after a break? How can these programs help in uplifting women in tech? Does your company have any such programs? TP: Yes. I think there is just a little bit of extra encouragement needed to get back into the workforce after a break. At a time, when one is battling with constant doubts on whether they are making the right choices for themselves and their families, even the smallest of encouragements go a long way. Making it easier for their resumes to be considered after a career break, is a great start. Having training sessions to bring them up to speed to the latest developments in their projects, their field of work and the market trends make them confident of their capabilities again. Locus provides six months of paid leave and six months of ‘work from home’ in the case of maternity leave. Apart from that, in case of any medical or personal emergencies Locus always supports its employees in surviving through the period and smoothing out the comeback process after a break. AIM: What are the various other steps that companies can take to increase the number of women in the technological field or for that matter even retain them? TP: The most significant change can be brought out by building an inclusive work environment. An environment which treats everyone as a human and not a resource. The parts of life that we leave to work in an office might affect our productivity. With women, this becomes apparent with family but for the same can be the case with men dealing tending to the elderly. Providing flexible office hours when needed and encouraging work from home when people are going through difficult times. Having adequate maternity leave policies and not considering them as an open position for new resource consideration, gives a woman the security she needs. Having a day-care facility also helps. Most of all the attitude of the organisation matters, if the approach is supportive and non-judgemental people like to continue to stay. AIM: Upskilling is one of the foremost requirements to sustain in the tech-driven industry such as data science. Are there enough opportunities within corporates for women to upskill? How could it be incorporated? TP: I think for upskilling in Data Science and technology, the resources are primarily available online via courses and research papers. The courses come with test and scores which have credibility in the market as well. I believe everyone should pursue these courses to enhance their skillsets and corporates should provide any financial and time-related support for those. Another skill set essential in Data Science is understanding the applications of the latest research improvements and the current market trends for the same.  The same can be enriched by attending conferences, meetups and reading online. Corporates should encourage participation of all members in such events. AIM: Do women in senior management roles have to tackle the ‘prove it again’ bias? TP: Yes. A lot. Every time a lady is promoted to senior management positions, she stands out. Questions on ‘why her’ do the rounds even when there are a lot more men getting promoted. She is supposed to prove again that she is as good, if not better than others, for the role. She may be sticking to regular office hours and not be available in the after-work networking events. And thus the questions on her soft skills, her ability to lead and to build her team start coming up. Objectively thinking these questions don’t stand a definitive ground, but a woman has to deal with them and face them again and again. AIM: What are the measures that can be put in place to help women rise to senior management roles in data science and analytics field? TP: Having more women in the leadership position who can understand the current biases, comes first. They should be encouraged to bring the current problems afflicting a team to the table and not be grilled on giving data-points as those are difficult to present retrospectively. We should build a support group where women can freely talk to other women, share their professional problems and advise each other on how to tackle them. AIM: Is there a need for mentorship for women to help them accelerate their careers? TP: Given that generally there are less than 10% of women in the leadership of any organization, I think mentors are definitely needed. Mentorship programs with 2, 3 level-up seniors (both men and women), can help women showcase her work, figure out the areas of growth and go up the career path. It can also help remove any bias that creeps in because of men being in each other’s friend circle. AIM: What would be your tips for maintaining work-life balance together? TP: A detailed and regular schedule helps a lot in finding that balance, so my suggestion would be to be assertive and stick to it as much as possible. But most of all never feel guilty of being human.","excerpt":"Toshi Prakash, who is the VP Product at Locus, has been working as a product manager for the last five years and had worked as the software development lead for eight years before that. With extensive experience in building sustainable and scalable product design, she has built both large-scale enterprise-level products and small fun apps […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2019-03-18T10:33:17","publication_year":"2019","word_count":1625,"keywords":["data science","Go","AI","ML","Scala","RAG","Aim","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","R","Go","Scala","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/women-face-behavioural-biases-the-best-way-to-overcome-is-to-stay-assertive-toshi-prakash-locus-sh\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10020939,"title":"What’s Kubernetes Got To Do With GPT-3’s Success","content":"“The applications and hardware that OpenAI runs with Kubernetes are quite different from what one may encounter at a typical company.” Microsoft backed OpenAI has delivered back to back blockbusters with GPT-3 and CLIP, DALL.E in a span of six months. While GPT-3 allowed OpenAI to venture into commercial API space, CLIP and DALL.E rang in a new era of fusion models. However, these models are large. GPT-3 devors all the data on the internet for training. And, it costs a few million dollars. “Scaling OpenAI’s infrastructure is unlike any what a typical startup does. So, even though they use familiar services like Kubernetes, the practices are unique to OpenAI. Today, many software service providers deploy Kubernetes for ease of operation but OpenAI claims to do it differently,” said OpenAI researchers. Kubernetes Overview Image credits: Google Cloud Conventionally, applications with different functionalities are packed into a single deployable artifact. And monoliths are an acceptable way to build applications even today. But they still have their drawbacks. For example, deployments are time-consuming since everything has to be rolled out together. And if different parts of the monolith are managed by different teams, the roll out prepping could run into additional complexities. Same with scaling: Teams have to throw resources at the whole application, even if the bottleneck is on a single channel. To address this, developers came up with microservices. Each piece of functionality is split into smaller individual artifacts. If there’s an update, only the exact service has to be replaced. The microservice model has scaling benefits. Now individual services can be scaled to match their traffic, so it’s easier to avoid bottlenecks without over-provisioning. So far, so good. But having one machine for each service would require a lot of resources and a whole bunch of machines. This is where containers come in handy. With containers, teams can pack their services neatly. All the applications, their dependencies, and any necessary configuration gets delivered together. Meaning, rest assured the services will run the same way, no matter where they are run. And, Kubernetes is all about managing these containers. Kubernetes, or K8s, is an open-source system used for grouping containers that make up an application into logical units for easy management and discovery. Kubernetes automates the whole process of scaling and management. Kubernetes comes from the garage of Google after nearly 15 years in the making. The key attributes of Kubernetes include: Kubernetes can do a lot of heavy lifting and make sure applications are always running the way they are intended to run. Kubernetes allows developers to focus on applications and not worry about the underlying environment. Kubernetes continuously runs health checks against the services, restarting containers that fail or have stalled. Kubernetes allows for ease of application modernization and lets developers build new apps faster. Kubernetes allows applications to be deployed on-site and public clouds as well as hybrid deployments in between. Kubernetes enables users to automatically scale applications, up and down, based on the demand and run them efficiently. “OpenAI has scaled Kubernetes clusters to 7,500 nodes, producing a scalable infrastructure for large models like GPT-3, CLIP, and DALL·E.”OpenAI How OpenAI Does It At OpenAI, workloads are constantly changing. The team at OpenAI builds production level applications even though they are short lived. The nature of AI research at OpenAI demands high quality infrastructure even for applications which will never see the light of the day. On our largest clusters, said OpenAI, we could possibly have approximately 200,000 IP addresses in use at any one time. Kubernetes is good at managing large clusters. At OpenAI, Kubernetes API Servers and etcd are critical components to a healthy working cluster. etcd is a consistent and highly-available key value store used as Kubernetes’ backing store for all cluster data. “Upside of scaling Kubernetes is a simple infrastructure that allows our machine learning research teams to move faster and scale up without changing their code.”OpenAI API Servers are memory intensive, and tend to scale linearly with the number of nodes in the cluster. When the team at OpenAI scaled to 7,500 nodes every server used up 70GB space. Typically, API Servers are run within kube services but OpenAI prefers running them outside the cluster itself. Both etcd and API servers at OpenAI run on their own dedicated nodes. The research lab’s largest clusters run 5 API servers and 5 etcd nodes to spread the load and minimise impact if one were to go down. And as their clusters have grown, the team at OpenAI faced auto scaling changes. Researchers found it difficult getting all of the allocated capacity. Traditional job scheduling systems, stated OpenAI, have a lot of different features available to fairly run work between competing teams, which Kubernetes does not have.The team took inspiration from job scheduling systems and built several capabilities in a Kubernetes-native way. “We’ve found Kubernetes to be an exceptionally flexible platform for our research needs. It has the ability to scale up to meet the most demanding workloads we’ve put on it,” concluded OpenAI. OpenAI Infra By The Numbers OpenAI’s largest clusters run 5 API servers and 5 etcd nodes to spread the load and minimise impact.OpenAI’s cluster with 7,500 nodes requires 70GB of heap per API Server.Alias-based IP addressing is used as the largest clusters and approximately 200,000 IP addresses can be in use at any one time.. Know more about OpenAI Scaling challenges here.","excerpt":"“The applications and hardware that OpenAI runs with Kubernetes are quite different from what one may encounter at a typical company.” Microsoft backed OpenAI has delivered back to back blockbusters with GPT-3 and CLIP, DALL.E in a span of six months. While GPT-3 allowed OpenAI to venture into commercial API space, CLIP and DALL.E rang […]","categories":["Deep Tech"],"tags":["GPT-3"],"author_name":"Ram Sagar","publish_date":"2021-02-26T18:00:00","publication_year":"2021","word_count":901,"keywords":["GPT-3","Go","machine learning","OpenAI","AI","Scala","RAG","microservices","Aim","R","kubernetes"],"extracted_tech_keywords":["AI","machine learning","OpenAI","Aim","RAG","kubernetes","microservices","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/whats-kubernetes-got-to-do-with-gpt-3s-success\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10045401,"title":"In A First, An Invention By AI Granted A Patent","content":"The Companies and Intellectual Property Commission of South Africa granted patent to a food container based on fractal geometry invented by ‘Device for Autonomous Bootstrapping of Unified Sentience’ (DABUS). DABUS is an AI system developed by Missouri physicist Stephen Thaler. The patent application was filed on September 17, 2019, under the Patent Cooperation Treaty. The application identified DABUS as the inventor of the food container and stated ‘the invention was autonomously generated by an artificial intelligence.’ Thaler is a pioneer in the field of AI and programming. His DABUS system is classified as a creativity machine, capable of independent and complex functioning. Patent applications were filed in the US, Europe, Australia, and South Africa. All except South Africa turned it down. The European Patent Office issued a preliminary communication stating the inventor on a patent application must have legal capacity. The United States Patent and Trademark Office (USPTO) asserted the ‘Application Data Sheet’ with the patent application did not identify the inventor by its legal name. Thaler has now appealed to the District Court for the Eastern District of Virginia seeking the reversal of the USPTO’s decision. The South African Patent Office’s decision to grant the patent has come as a surprise to the global community. The reasons for this decision are unknown. However, the whole incident has riled up the intellectual property experts. The critics lampooned the decision saying AI lacks a legal standing necessary to qualify as an inventor. Many chalked up the granting of patent to an oversight on the part of the commission, and a few see it as an indictment of South Africa’s patent procedures.","excerpt":"DABUS is an AI system developed by Missouri physicist Stephen Thaler.","categories":["AI News"],"tags":["Patent","United States Patent and Trademark Office (USPTO)"],"author_name":"Shraddha Goled","publish_date":"2021-08-06T12:15:37","publication_year":"2021","word_count":270,"keywords":["artificial intelligence","programming_languages:R","AI","United States Patent and Trademark Office (USPTO)","Patent","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/in-a-first-an-invention-by-ai-granted-a-patent\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":26147,"title":"Should Machine Learning Be A Separate Major Like Computer Science?","content":"This is the time when a strong Computer Science background is highly-recommended for budding Machine Learning practitioners. In fact, students are closing the statistics gap with online programs as well. So is it time Indian universities start offering a major in machine learning? ML jobs are booming in India — recent example being that US online payment giant PayPal is planning to hire 600 tech experts in the field of artificial intelligence and ML. Indian telecom giant Reliance Jio is also planning a separate AI team. While there are a lot of edtech companies that offer structured in-person as well as online ML training in conjunction with industry partners, Indian institutes like the University of Hyderabad, IIT Bombay, IIT Madras, IISc Bangalore and ISI Kolkata, which have thriving ML departments also offer specialisations in ML and AI-related field. US universities, especially the big four (MIT, CMU, Stanford, University of California and Berkeley) are already ahead of the pack in machine learning research and artificial intelligence. Not just the US, Canadian universities, especially the University of Toronto and Montreal lead in the computational side of things. In India too, IIT Hyderabad’s Department of Computer Science and Engineering offers an MTech in machine learning which prepares students for ML-oriented jobs as well as research areas. In fact. IIT-H’s CSE department is active in research areas such as networking, distributed systems, algorithms, graph theory and also offers a three year research-oriented MTech programme which enables students to work on sponsored research projects along with the coursework. There is also a two year MTech course in Machine Learning and Computing offered by Mathematics department of Indian Institute of Space Science and Technology (IIST). What Should Students Opt For? A Major In ML Instead Of CS? Wider Scope Of Application: So, is pursuing a graduate degree in CS a better idea than majoring in ML? Machine Learning offers a wider scope of application of degree and also leads to a pathway to higher pay packages and jobs at leading firms such as Samsung, Flipkart, Amazon and Accenture, among others. Got A Strong Interest In AI, Go For MS In ML: If you have a strong interest in AI and have a fairly good programming background, then a degree in ML from India or abroad will be extremely beneficial for further employability as well research. Learning how to implement models will definitely be helpful in breaking into the highly competitive industry. Also, an MS is the best way to get past the HR screen to land a coveted job in the machine learning field. Rise Of ML-Focused Data Science Jobs: With the rise in machine learning-focused roles at companies across the board, most interviews around data science job positions revolve around algorithms and code architecture. A strong academic background and strong CS fundamentals will help secure a job in an ML-oriented company. Since employers use a degree to screen applicants, an MS in machine learning can teach you the state-of-the-art algorithms used by big tech firms. On The Other Side Most Computer Science programs offer a number of electives to students to choose from which is helpful in getting in-demand industry skills. Also, according to Eindhoven University of Technology, Netherlands, computer science master’s program offers excellent learning trajectories since they combine modules such as information systems, software development, system architecture and theory algorithms – thus offering the students a wider scope of application. Theory and algorithms module covers topics like data structures, logic and set theory, DBL algorithms, discrete structure and more. Most CS programs also feature modules such as data mining and machine learning that covers important areas like probability and stochastics. Other times, employers prefer candidates with a CS background or formal training as opposed to heavy statistics knowledge. MS in machine learning is also preferred for candidates who are more interested in pursuing research opportunities as opposed to professionals who want to work on ML-focused jobs. Most CS candidates are also trained in cleaning datasets, building models and churning out solutions for business problems. So, a CS route also opens up employability and research options alike. Another key point is that machine learning programs are better suited to candidates with a heavy statistics background and usually the Master’s candidates go for further research.","excerpt":"This is the time when a strong Computer Science background is highly-recommended for budding Machine Learning practitioners. In fact, students are closing the statistics gap with online programs as well. So is it time Indian universities start offering a major in machine learning? ML jobs are booming in India — recent example being that US […]","categories":["AI Features"],"tags":["machine learning computer","uc berkeley"],"author_name":"Richa Bhatia","publish_date":"2018-07-05T05:04:39","publication_year":"2018","word_count":710,"keywords":["data science","Go","machine learning","artificial intelligence","uc berkeley","AI","programming_languages:R","machine learning computer","ML","programming_languages:Go","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/should-machine-learning-be-a-separate-major-like-computer-science\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10125692,"title":"CP Gurnani Says, &#8216;There is No Human Being Who is Not Working on Generative AI&#8217;","content":"After celebrating the milestone of building Tech Mahindra for 19 years, the company’s former CEO, CP Gurnani, decided to set sail on a separate journey. Though not far from his techno-optimist belief. Just a few months after his farewell from Tech Mahindra, he announced the launch of AIonOS, a business venture in partnership with InterGlobe Enterprises’ Rahul Bhatia. This takes us back to Sam Altman’s visit to India. When the OpenAI chief claimed that India could not build a model comparable to ChatGPT, Gurnani said, “Challenge accepted.” “Now that we have taken the challenge, we will deliver, and we will brainstorm,” Gurnani told Nikhil Malhotra, his erstwhile chief innovation officer at Tech Mahindra’s Project Indus. Speaking with AIM, Gurnani said, “I spoke to my chief innovation officer that time at Tech Mahindra… Six hours later, he says, ‘I have a plan.’” Just recently, Project Indus went live. Tech Mahindra was able to develop an Indian LLM for local languages and 37+ dialects in just five months, spending less than $5 million. “When ET covered the seven best AI companies that will have an Indian LLM, they covered Tata, NVIDIA, and Tech Mahindra,” he said. “I have to thank the Tech Mahindra leadership team for delivering what I promised on my behalf. I also have to thank Sam Altman as he got me so interested in AI,” Gurnani added, explaining the origin of AIonOS. “I promise you, on my birthday, which is December 19, I will deliver at least a couple of elements,” he said about the launch of AIonOS. AIonOS currently offers specialised AI products and technologies, including custom solutions, industry-specific products, data insight engines, and an AI-led customer experience. “I can only say there is no human being, including you, who is not working on generative AI. The reality is that you probably framed these questions by asking generative AI,” Gurnani quipped. Highlighting the potential of generative AI, he noted, “Whether you are a creative person, a student, a politician, or a journalist, you can take help from generative AI. It’s just a matter of time before it becomes part of your life.” The Future is About Agent AI One thing that Gurnani is completely convinced about is that the future of AI is in Agent AI. “I’m actually more convinced that the faster adoption will be in Agent AI. Each one of us will have an AI agent that knows us so well, it will analyse our business and our routines, and support us in becoming more productive and efficient,” he explained. He even envisages a future where AI agents take over meetings, although he personally hopes he can continue enjoying face-to-face interactions. India is Good with Innovation Regardless of the fact that it has been noted that India merely copies innovation from the West, Gurnani expressed optimism about the AI startup ecosystem in India. He recognised their role in addressing market gaps, as startups like Krutrim and Sarvam AI do. “All startups have very good intent. They represent the gaps in the market offerings,” he said. He categorised startups into three types: those improving existing processes, those creating new marketplaces, and those working on deep tech innovations. He shared an example of deep tech innovation in India, mentioning the success stories of ISRO. “NASA used to spend billions of dollars on space missions. ISRO costs around $300-400 million. It’s because there was a leader willing to challenge the status quo,” Gurnani stated, underscoring the importance of leadership and innovation over mere funding. Gurnani highlighted the significant work happening at various research centres and institutions in India, like IIT Madras. He praised the efforts of individuals like Ashok Jhunjhunwala at IIT Madras, who, despite limited funds, are driving groundbreaking innovations. “There is a space launch pad, EV charging stations, vertical takeoff drones—all designed and built in India,” he pointed out. He also named IISc’s Centre for Brain Research, funded by Kris Gopalakrishnan, which is making strides in understanding diseases like Parkinson’s and Alzheimer’s. “The amount of work happening on a paltry sum of about Rs 30 crores a year is astounding,” Gurnani said, emphasising the importance of frugal innovation in India. Stay Relevant Gurnani remained positive, stating, “I am here to help fix the issues. Nothing annoys me.” He also expressed scepticism about AGI, calling it a “counter-intuitive” term and questioning its practicality. “Either you are artificial, or you have no idea,” he asserted. Regarding the debate on work hours, Gurnani supported the idea of working hard but balanced it with self-development. “Narayan Murthy is a respected leader. The trolling he faced for suggesting a 70-hour work week was unnecessary. I believe in working 40 hours for your job and 30 hours for yourself,” he advised. In conclusion, Gurnani’s message to the youth was clear: “Stay relevant. The industry is changing very fast. Do not assume that your degree is sufficient to keep you employed or successful. Stay up to date with the fundamentals and keep adapting to the changes.”","excerpt":"“Whether you are a creative person, a student, a politician, or a journalist, you can take help from generative AI. It’s just a matter of time before it becomes part of your life.”","categories":["IT Services"],"tags":["CP Gurnani"],"author_name":"Mohit Pandey","publish_date":"2024-07-03T18:00:00","publication_year":"2024","word_count":832,"keywords":["Go","ChatGPT","CP Gurnani","OpenAI","AI","innovation","GPT","Ray","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Aim","Ray","R","Go","GPT","innovation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/cp-gurnani-says-there-is-no-human-being-who-is-not-working-on-generative-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10109585,"title":"UP to Build India’s First AI City in Lucknow","content":"The Uttar Pradesh government has announced plans to establish India’s inaugural AI city in Lucknow, aimed at nurturing and advancing the AI ecosystem. The initiative, led by UP Electronics Corporation Ltd, the project’s nodal agency, has issued an expression of interest (EoI) soliciting proposals for the design, development, and operation of this groundbreaking AI city. The envisioned AI city aspires to be a dynamic hub fostering innovative ideas, incorporating state-of-the-art technology, research facilities, and educational institutions. The EoI invites real estate developers to actively participate in constructing and managing the AI city, with the goal of cultivating the workforce of the future. Lucknow already hosts Centers of Excellence specialising in AI and medtech, showcasing significant integration of AI technologies. The AI Centre of Excellence at IIIT Lucknow plays a pivotal role in supporting over 15 AI\/ML startups, contributing to the development of a vibrant culture of creativity and entrepreneurship. Furthermore, a report underscores the planned AI city’s components, including Grade-A certified commercial spaces for IT firms, data centre, Grade-A flexible workspaces, and technology laboratories. The IT and Electronics Department has identified potential land parcels in strategic locations across Lucknow for the AI city’s development, with a 40-acre parcel in the Nadarganj industrial area being of particular significance due to its proximity to key infrastructure and connectivity to the Lucknow-Kanpur road and the Lucknow International Airport. India boasts over 70 generative AI startups that have raised over $440 million in capital between 2019 and Q3 2023. The domestic generative AI market is anticipated to expand from $1.1 billion in 2023 to $17 billion by 2030. The government of UP has embarked on a mission to achieve a $1 trillion economy within the next five years, with a specific focus on the IT and IT-enabled Services (ITeS) sectors as pivotal drivers for economic growth. UP, ranking as the sixth-largest state in terms of the IT ecosystem, is also recognised for hosting a substantial number of startups within India. Noida is emerging as a global hub for IT and ITeS, and UP aims to enhance this sector by fostering Tier 2 cities like Lucknow, identified by prominent IT industry associations as emerging technology hubs. EoI document emphasises the strategic role of AI as a key driver for global growth in the IT and ITeS sectors, which the state intends to capitalise on. Lucknow already boasts the presence of major IT players, including HCL and TCS, contributing to a flourishing tech ecosystem. The city’s tech landscape is further enriched by a skilled workforce, comprising over 75,000 tech professionals, 23,000 STEM graduates, and 300+ colleges, featuring esteemed institutions like IIM-Lucknow, IIIT-Lucknow, BBDU, and Amity, as highlighted in the EoI. Simultaneously, the government is actively addressing AI data and regulation issues. The central government has initiated the formulation of regulations to foster the growth, safeguard interests, and promote innovation in the AI sector, as per IT Secretary S Krishnan. Minister of State (MoS) for Information Technology Rajeev Chandrasekhar’s recent announcement of plans to fund and support AI startups in the country is also going to foster the AI ecosystem.","excerpt":"The government of UP has embarked on a mission to achieve a $1 trillion economy within the next five years.","categories":["AI News"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2023-12-28T10:40:41","publication_year":"2023","word_count":513,"keywords":["Go","API","AI","innovation","ML","Aim","ViT","generative AI","GAN","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","R","Go","API","GAN","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/up-to-build-indias-first-ai-city-in-lucknow\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":62574,"title":"Why Facebook Invested In Jio","content":"Facebook had announced this week that it had invested $5.7 billion for a 9.99% stake in Reliance Jio Platforms. This deal makes the social media giant the largest minority shareholder in the Indian telecom network. It also marks the largest investment for a minority stake by a tech company in the world, and the largest foreign direct investment in the tech space in India. The development comes in the wake of the emergence of China’s ByteDance in the world’s second-largest internet market. Founded by Mukesh Ambani, Jio had transformed India’s technology market when it started in 2016 by offering free calls and cheap 4G internet service. Widely welcomed by users hamstrung by slow 3G connections and high tariffs, Jio’s entry obliterated competitions, including Airtel and Vodafone. Collaborating To Grow Digital Platforms India has been going through a rapid digital transformation over the last few years. It will not be wrong to say that Jio has contributed a sizable chunk to this. Since it has played a big role in bringing millions of Indians and small businesses online, Facebook, with this investment, will have closer access to Jio users. Facebook sees a lot of potential through this collaboration. India is home to 400 million WhatsApp users – Facebook’s global messaging app – and 300 million Facebook users. So, Facebook is now working on collaborations in one of its biggest markets, and with one of the biggest telecom service companies that is driving the same market. In a joint interview, Chief Revenue Officer at Facebook, David Fischer and Managing Director of Facebook India, Ajit Mohan said that one possible collaboration would be JioMart. For years, WhatsApp has worked to build tools for small businesses and taken an avid interest in payment systems, and Facebook has also done something similar. By bringing together JioMart, which is Jio’s small business initiative with WhatsApp, Facebook plans to connect people with businesses, shops and purchase products, seemingly giving them a chance to discover new products. “The country is in the middle of a major digital transformation, and organisations like Jio have played a big part in getting hundreds of millions of Indian people and small businesses online.”– Mark Zuckerberg Facebook’s Chief Executive shared in a post that the investment will help many entrepreneurs get the digital tools that they require to communicate with customers and grow their businesses, thereby, helping the economy. Why Hasn’t Facebook Explored The Investment Earlier? The last time Facebook tried to enter the Indian market, it did not work out the way they wanted it to because of regulatory issues. It previously offered free internet connectivity to Indian users in a program called Free Basics, but it was banned in India. Regulators decided that companies could not offer free internet that will favour one company over another. It is believed that the regulators were the ones who scrapped the idea at first, but the fact was that the idea of Free Basics received a lot of criticism from the people themselves. A lot of people who belonged to professions like farming, labourers and office workers frowned upon the idea of having free internet that came with certain conditions. Free basics actually provided only limited access to the internet (something for which the movement of net neutrality was started) through a suite of websites and services which also included Facebook. But people felt that this idea of limited service wasn’t fulfilling the promise of the open nature of the internet, where there wasn’t limited access to some websites. One might be surprised about people opposing the idea of free Facebook, especially when it had around 216 million users in India at the time of introduction of Free Basics (in 2020 the number is around 346 million). But, how much will one browse Facebook even if it was for free? Naturally, Indian regulators had to decide on banning Free Basics because in addition to the displeasure of the public it was offered for, it also had some issues where the public opinion could be influenced. What this means is that the limited access to certain companies or information will create a divisive environment where other companies or information may be left out. Although it did receive some support, where its benefit for the businesses in rural areas was highlighted, ultimately the idea was cut-off. Also, in the past, Facebook had some issues with the Indian government and over WhatsApp. The government demanded that WhatsApp change its encryption to trace back messages, to which WhatsApp did not agree. It is believed that this is one of the reasons the regulators have stalled WhatsApp’s request to offer a payment service for its Indian users. How Does This Investment Help Both Sides? Since 2016, Jio has become India’s largest carrier, with almost 400 million subscribers. But Mukesh Ambani’s push for the digital revolution in India has caused him to incur enormous debts to build the telecom business. With Jio having ambitions to take on Amazon in e-commerce, provide fiber net, run data centers and set up new services like telehealth and distance learning, the investment from Facebook will greatly help Reliance. This investment will help the company reduce the debt and invest further in its network. But Facebook may stand to gain a little more from this deal. Mukesh Ambani is a prominent voice in India and many conclude that it shares close ties with the ruling government. If true, this could help the company leave a bigger mark in India. Outlook This deal between Jio and Facebook will only accelerate India’s push for ‘Digital India’. Amid the COVID-19 pandemic and the economic slump, this may come with a two-fold advantage for India – driving the digital transformation and regulating the economy. With JioMart and WhatsApp, Ambani, in his video message, said that it would be possible for around 30 million neighbourhood stores to transact digitally.Facebook, on the other hand, will gain a huge base of users because of TikTok’s rise, which has amassed more than 250 million users in India. Furthermore, Facebook is also attempting to build a similar service called Lasso and having Jio on their side will greatly benefit them.","excerpt":"Facebook had announced this week that it had invested $5.7 billion for a 9.99% stake in Reliance Jio Platforms. This deal makes the social media giant the largest minority shareholder in the Indian telecom network. It also marks the largest investment for a minority stake by a tech company in the world, and the largest […]","categories":["AI Features"],"tags":["Jio AI Cloud"],"author_name":"Sameer Balaganur","publish_date":"2020-04-24T20:00:00","publication_year":"2020","word_count":1026,"keywords":["Go","API","ELT","programming_languages:R","AI","digital transformation","Git","Jio AI Cloud","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","API","ELT","GAN","ViT","digital transformation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-facebook-invested-in-jio\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10125550,"title":"Data Engineering Odyssey: From Traditional Warehousing to Generative AI","content":"The field of data engineering has undergone a seismic shift from the days of traditional data warehousing to the dynamic, AI-driven landscape we see today. Rekha Sree, a senior manager at Tredence, offers an in-depth look into this evolution, detailing the challenges, innovations, and the future trajectory of data engineering practices. Sree said that traditional data warehousing was characterised by centralised warehouse and batch processing, relying on rigid schemas with structured data. “These schemas were inflexible and any changes to these were time-consuming and costly,” said Sree. “Modern data engineering, however, follows distributed computing paradigms, offering scalability and flexibility, especially with cloud solutions.” The pay-as-you-go model of cloud services significantly reduces costs while providing robust governance and security measures. Sree added that generative AI has expanded the role of data engineers significantly. She explained, “Earlier, data engineers focused on making data usable for reports. Now, they focus  on designing and understanding complex AI models, improving data quality, optimising infrastructure for AI applications, and data preparation for training these models.” This expansion necessitates collaboration with business users, data scientists, AI assistants, and cloud providers, emphasising innovation and collaboration over traditional methods. Challenges in Transitioning Transitioning from traditional data warehousing to AI-driven solutions brings several challenges. Integrating AI technologies with existing infrastructure requires substantial investment. Ensuring data privacy and addressing ethical and regulatory concerns is paramount. The complexity of AI algorithms necessitates maintaining high data quality. Organisations must adapt to the significant cultural changes involved in moving to AI-driven systems. Generative AI enhances data quality by simplifying data cleansing, augmentation, and anomaly detection. However, it also poses challenges to data governance, particularly regarding data authenticity and integrity. Sree stressed the importance of establishing robust policies and controls to govern data usage, sharing, and interpretation to maintain trust and prevent misuse. The structural and functional differences between data lakes and traditional data warehouses are stark. Traditional warehouses rely on rigid schemas and centralised control, whereas modern solutions like data lakes and data mesh offer flexibility and scalability. “Data Mesh is a decentralised approach to managing data within organisations, where data ownership and governance are distributed to domain-specific teams,” Sree noted. The Evolving Role of ETL Extract, Transform, Load (ETL) processes remain crucial in modern data engineering. “Extraction now involves various sources, from databases to APIs and multimedia such as audio, images, animations, and video,” explained Sree. Transformation includes cleaning and enriching data, while loading involves storing data in target systems like data lakes. The recent emphasis on automating ETL processes aims to minimise human errors and expedite deployment. Integrating generative AI into existing infrastructure requires careful planning and adherence to best practices. Sree highlighted several key strategies: assessing compatibility of AI models with existing systems, ensuring high data quality and addressing ethical considerations, establishing transparent and accountable data governance practices, and regularly evaluating AI model performance to ensure accuracy and reliability. AI-driven systems offer enhanced scalability, addressing the three V’s: volume, velocity, and variety of data. “Elastic scalability on cloud platforms allows dynamic resource allocation based on demand,” Rekha explains. These systems also support parallel processing and various optimization techniques, ensuring they can handle large and complex data sets efficiently. The journey from traditional data warehousing to modern data engineering driven by generative AI is marked by significant advancements and new challenges. As Sree articulated, the role of data engineers has expanded, requiring greater collaboration, innovation, and adherence to robust governance practices. With the right strategies, organisations can harness the power of generative AI to drive their data initiatives forward, ensuring scalability, flexibility, and cost-effectiveness.","excerpt":"The journey from traditional data warehousing to modern data engineering driven by generative AI is marked by significant advancements and new challenges.","categories":["AI Highlights"],"tags":["Data Engineering"],"author_name":"Mohit Pandey","publish_date":"2024-07-02T17:34:44","publication_year":"2024","word_count":592,"keywords":["Go","AI assistants","AI","distributed computing","Scala","Aim","anomaly detection","Data Engineering","generative AI","Rust","R"],"extracted_tech_keywords":["AI","generative AI","Aim","AI assistants","anomaly detection","distributed computing","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/data-engineering-odyssey-from-traditional-warehousing-to-generative-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":59704,"title":"COVID-19 Forces Samsung And LG To Put Lock-On Factories","content":"Keeping the request of the government and to stop the spread of COVID-19, Samsung and LG have decided to shut down their factories in the nation. Samsung Electronics Co. has announced to temporarily bring the shutters down to its smartphone factory at Noida, based in the state of Uttar Pradesh from March 23 to 25. Samsung has pledged to follow the guidelines being laid down by the government and will begin the operation on March 26 again, to ensure that no-shortage of product supply is faced in the coming weeks. The three-day halt in the operations is likely to hit Samsung hard, considering the plant in Noida is the South Korean tech giant’s largest smartphone factory that produces 120 million units per year. Earlier this month, Samsung closed down its operation in Gumi, South Korea, after some of the employees were reported to be infected with COVID-19. The company temporarily shifted its production operations to Vietnam. Company co-CEO HS Kim said that the company is yet to determine how COVID-19 will impact distribution and consumers and will conduct research further to prepare for it. The company also shut down its television manufacturing factory in Slovakia for a week. On the other hand, rival LG Electronics Inc. has also shut down its home appliance factories in Noida and Pune until the end of the month, following the guidelines laid down by the state governments. The company has already shut done its research centres and also had to shut down another factory after an employee was tested positive with Coronavirus. With no prior notice of the complete lockdowns, both the companies are worried about the disruption in supply and the sales figure, which is likely to face a steep decline in the coming weeks.","excerpt":"Keeping the request of the government and to stop the spread of COVID-19, Samsung and LG have decided to shut down their factories in the nation. Samsung Electronics Co. has announced to temporarily bring the shutters down to its smartphone factory at Noida, based in the state of Uttar Pradesh from March 23 to 25. […]","categories":["AI Features"],"tags":["LG","Samsung","south korea","Tech giants"],"author_name":"Rohit Chatterjee","publish_date":"2020-03-24T13:41:35","publication_year":"2020","word_count":292,"keywords":["Go","Samsung","programming_languages:R","AI","LG","Tech giants","programming_languages:Go","disruption","R","south korea"],"extracted_tech_keywords":["AI","R","Go","disruption","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/covid-19-forces-samsung-and-lg-to-put-lock-on-factories\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":64839,"title":"Lenovo Introduces New Server Platforms For Advanced Analytics &#038; AI Workloads","content":"Lenovo Data Center Group (DCG) announced the launch of the ThinkSystem SR645 and SR665 two-socket servers for higher performance workloads. Many industries, such as financial services, retail, and manufacturing need faster transaction processing, improved data analytics, and greater grid-computing capacity, while still improving efficiency and total cost of ownership (TCO). To address these needs, Lenovo is focusing on infrastructure solutions comprising servers, storage, and software. With the addition of the new ThinkSystem SR645 and SR665 servers featuring more CPU cores and a larger memory footprint, Lenovo can help enterprise users accelerate higher performance workloads. This announcement follows the release of the ThinkSystem SR635 and SR655 single-socket servers and expands Lenovo’s server portfolio powered by AMD EPYC 7002 Series processors. “Our new Lenovo ThinkSystem servers are designed for workloads such as in-memory databases, advanced analytics, virtualization, and AI”, said Kamran Amini, Vice President and General Manager of Server, Storage and Software Defined Infrastructure, Lenovo Data Center Group. “With the exceptional power, speed and onboard storage of these new servers, our customers have the ability to handle the increasing data requirements of today’s workloads with the scalability to grow with their business.” In high performance computing (HPC) environments like scientific research, higher throughput means faster results, yielding earlier discoveries. The high core counts allow customers to buy fewer servers, saving them rack space and power, in addition to obtaining results faster. “The new Lenovo ThinkSystem two-socket servers tap into the power of our 2nd Gen AMD EPYC processors to make a meaningful impact in the areas where performance, space and efficiency are critical to business outcomes,” said Dan McNamara, senior vice president and general manager, server business unit, AMD. “With the expansion of its AMD EPYC-based product line, Lenovo customers now have more options for systems that address today’s demanding enterprise workloads.”","excerpt":"Lenovo Data Center Group (DCG) announced the launch of the ThinkSystem SR645 and SR665 two-socket servers for higher performance workloads. Many industries, such as financial services, retail, and manufacturing need faster transaction processing, improved data analytics, and greater grid-computing capacity, while still improving efficiency and total cost of ownership (TCO). To address these needs, Lenovo […]","categories":["AI News"],"tags":["AI Work","Lenovo"],"author_name":"Vishal Chawla","publish_date":"2020-05-08T14:03:09","publication_year":"2020","word_count":300,"keywords":["programming_languages:R","AI","Scala","RAG","Lenovo","analytics","programming_languages:Scala","AI Work","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Scala","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/what-are-lenovos-new-server-platforms-for-advanced-analytics-ai-workloads\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":59325,"title":"We At NASSCOM Are Building An Innovation Ecosystem For AI &#038; IoT Startups In India: Sanjeev Malhotra ","content":"According to Sanjeev Malhotra, CEO of NASSCOM- CoE For IoT & AI, he is currently building an innovation ecosystem for IoT & AI startups in India, working with large enterprises, SMEs, government and VCs. According to Sanjeev Malhotra, he is currently building an innovation ecosystem and driving co-innovation with larger enterprises, SMEs and startups for the largest technology industry body in India. Working closely with the Indian government, Nasscom’s CoE- IoT & AI focuses on emerging technologies like artificial intelligence, IoT, robotics etc. and their application in auto, manufacturing, healthcare, agriculture, retail, fintech, etc. Here are the edited excerpts from the interaction – AIM: What is your role here at NASSCOM in terms of fostering technologies like IoT and AI? Sanjeev Malhotra: Emerging technologies are growing and developing at an incredible pace, and there are so many things that are important to harness completely. No one company can do it. Large companies can only make building blocks like software or chips, but applications will be developed by small companies. This is where the role of startups in India is very important. Also, how we actually develop applications and how do we actually monetise them, that’s very important. And that is my job where we bring all these elements or stakeholders in this ecosystem. We are helping speed up the startups which are actually doing a lot of innovation and also working with the government of India. AIM: Can you tell more about the technology ecosystem that you are trying to advance? Sanjeev Malhotra: We have an ecosystem around innovation, where we are helping young Indian innovators with different things they need, whether it is the hardware components or the software solutions. Besides, startups in India need funding connects and this is also where we help them. So that is one part of it where we help the innovators. The second part is helping with the adoption of these products. These are the user communities and various stakeholders who are sitting in different parts of the country. So, we work with those people also to see how we bring them on board. AIM: Apart from the Centre of Excellence for IoT & AI, NASSCOM also has a Centre Of Excellence For Data Science in Bengaluru. Is there an overlap between the functions of the two? Sanjeev Malhotra: As long as there is data and AI, yes, there is an overlap. I mean there is no IoT without AI, and I think the technologies are pretty much interchangeable. We can have AI and analytics from data generated by connected devices because IoT generates a lot of data. Both centres of excellence work under similar functions as part of NASSCOM. AIM: Does the CoE also invest in startups or own any equity as part of incubating IoT and AI startups in India? Sanjeev Malhotra: No, NASSCOM does not invest in startups. Let’s say if you invest two million dollars, you will want five million dollars coming out of it. Any investor will want it, and then the goal is to just take care of only those startups you have invested in. We, on the other hand, are taking a different approach and for us every startup is equal. So we don’t take any equity from any startup. We do not differ with them, and that also gives us the freedom to focus more on what the country needs. As far as funding goes, we certainly look after the needs of funding and bring the right venture capital organisations. “NASSCOM does not invest in startups. Instead, we are taking a different approach and for us every startup is equal. So we don’t take any equity from any startup. We do not differ with them, and that also gives us the freedom to focus more on what the country needs.” – Sanjeev Malhotra, CEO of CoE- IoT & AI, NASSCOM AIM: How do you see the development of IoT standards going forward? Sanjeev Malhotra: I think IoT requires rigorous standardisation due to cybersecurity and innovation related concerns. While it looks very nice to say that two billion devices will talk to each other, how will they control and what about security and standardisation? If you’re importing devices from outside, should we follow the standard, which is developed in the west or do we develop our own standard? India is now caught up with the rest of the world, and now we are looking at the process of standards. We don’t have a history of making standards. We were never invited to the table when other standards were being built like Wi-Fi. But now when we are developing applications that are coming out from emerging technologies, we may also have an Indian version of specific technology standards, including IoT as well. AIM: Do you think that the dependence on China for electronic and IoT chips will continue? Sanjeev Malhotra: China has mastered chip manufacturing at low prices, which nobody else in the world can match. While hardware is a very critical component of IoT, the dependence on China for hardware is not just restricted to India but the entire world. India’s strength, on the other hand, lies strongly on the software side of things and we should continue to do that, and try to capture the software market similar to how China has captured the hardware and chip market. Also Read: What Happened At A NASSCOM Event For Deep Tech Startup Ecosystem in India?","excerpt":"According to Sanjeev Malhotra, CEO of NASSCOM- CoE For IoT & AI, he is currently building an innovation ecosystem for IoT & AI startups in India, working with large enterprises, SMEs, government and VCs. According to Sanjeev Malhotra, he is currently building an innovation ecosystem and driving co-innovation with larger enterprises, SMEs and startups for […]","categories":["AI Features"],"tags":["Interviews and Discussions","NASSCOM","Startups"],"author_name":"Vishal Chawla","publish_date":"2020-03-21T18:15:11","publication_year":"2020","word_count":908,"keywords":["data science","Go","API","artificial intelligence","AI","Aim","ViT","analytics","GAN","Startups","NASSCOM","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","R","Go","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/we-at-nasscom-are-building-an-innovation-ecosystem-for-ai-iot-startups-in-india-sanjeev-malhotra\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":58855,"title":"Get Paid For Your Data, But How?","content":"Data is the new oil is the sayings of the past. Data is my property, and I should get paid for it might be the catchphrase of this decade. All kinds of user data are being exploited, manipulated and minted by web-based marketplace services powered by algorithms. People unknowingly give vast amounts of their data to companies every day, which is used to generate massive amounts of profits. But how valuable is any data? And, how do we put a price tag on data in an unbiased way? Acknowledging these queries with the help of machine learning, researchers from UC Berkeley, ETH Zurich, and UIUC propose a few methods in the AISTATS and VLDB papers. One of the researchers Ruoxi Jia, lists the following important valuation schemes: Task-specificnessFairness: Watch out for adversarial attacks and data poisoning where one entity benefits more than others.Efficiency: Retaining performance during scaling up. To formulate a metric that would help price the data, the researchers imbibe a Shapley value equivalent to their framework. A Shapley Value For Data Pricing via BAIR Shapley value is widely used for profit allocation schemes. It attaches a real-value number to each player in the game to indicate the relative importance of their contributions. Similarly, the researchers considered each data provider as a player with a Shapley value, which is calculated based on the relevance of their contributions. For example, a data provider with skin cancer dataset is more relevant for medical diagnosis than the one with road sign dataset designed for self-driving cars; a measure based on relevance. Shapley value is given by the following expression: Here, U(S) is the utility function that evaluates the worth of the player(data source) subset S. Shapley value has also been chosen for its following properties that complement commercial usage of algorithmic solutions: Fully distributedEnsures fairnessBrings in additivity However, calculating Shapley value for a smaller dataset is challenging. Because, Shapley value needs many evaluations, and since evaluation here means re-training, the size of the dataset comes into the picture. So, in order to address the issue, the team at Berkeley AI research, use KNN classification to skip the need for re-training. As shown in the figure above, the team demonstrated how the computational requirements of the Shapley value could be significantly reduced for KNN. To address the scalability challenge in the online setting, the team has also developed an approximation algorithm to compute the Shapley value for KNN (K Nearest Neighbors) with improved efficiency. via BAIR One of the crucial challenges of introducing any metric into a machine learning setting is to overcome the adversarial attacks. When the Shapley value is checked for a dataset injected with noise, it showed promising results of detecting noisy training data. Thus, establishing confidence in adversarial robustness of the model. Not only for putting a price tag on data, but the researchers believe that Shapley value can also help in improving the interpretability of AI models. Dawn Of Data Dignity Jaron Lanier, a tech pioneer, was featured in a recent New York Times op-ed explaining why people should get paid for their data. Under this scheme, he estimates the total value of data for a four-person household could fetch around $20,000. Lanier insists that there is more to it than a mere monetary benefit. He calls it “data dignity.” Since the data exists because an individual exists, he\/she should have the final say on what happens to their data while also being allowed to make money out of the data that they choose to provide. If an e-commerce company skims through your buying history to recommend a product, that’s fair. But what if the same information is used to recommend similar products to other customers based on similarity scores? For instance, if an old customer had bought a very expensive item X followed by a low price item Y. If a new customer on the website, buys this similar low price item Y, then they might be recommended X, which they never had plans to buy in the first place. This looks like a typical recommendation engine use case. Now, what if this new customer goes ahead and buys this X? Then the e-commerce site has just hit the jackpot. But, only, in this case, these customer retaining strategies occur every second across the globe, minting billions of dollars. Now imagine how many other industries are directly and indirectly benefiting from user’s data. From increasing traffic to websites to personalising ads that make millions to insurance spams, data is being squeezed for profits every second without the knowledge of its owner. As automated data-driven solutions will rule the market in the coming decades, there is an immediate need for a robust framework powered by metrics such as Shapley value, needs to be devised that will restore data dignity of an individual through transparent yet financially lucrative subscription plans.","excerpt":"Data is the new oil is the sayings of the past. Data is my property, and I should get paid for it might be the catchphrase of this decade. All kinds of user data are being exploited, manipulated and minted by web-based marketplace services powered by algorithms. People unknowingly give vast amounts of their data […]","categories":["Deep Tech"],"tags":["data validation tool","shapley value"],"author_name":"Ram Sagar","publish_date":"2020-03-17T17:00:56","publication_year":"2020","word_count":811,"keywords":["Go","data validation tool","machine learning","programming_languages:R","AI","data-driven","shapley value","Scala","ViT","AI research","R","adversarial attacks"],"extracted_tech_keywords":["AI","machine learning","R","Go","Scala","ViT","adversarial attacks","data-driven","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/data-price-validation-berkeley-cost-deep-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":32313,"title":"This Kerala Startup Raised Pre-seed Funding For Its AI-Based Anomaly Detection Tech From HPCL","content":"Kerala-based startup Tranzmeo which has an AI-based product called T-Connect Overview that provides end-to-end anomaly forensics for better refinery operational outcomes raised pre-seed funding from Hindustan Petroleum Corporation Ltd.  The startup , founded in 2017 is part of NASSCOM’s 10,000 Startups accelerator program is founded by Safil Sunny who has 8+ years of experience in the field of software development. One of the fastest growing technology companies in Kerala, this startup builds machine learning and AI-based products for large and small organisations. The core product T-connect OneView – the anomaly forensics application essentially runs on data streams and is able to predict failure arising from incoming data stream and automatically creates alerts, thereby improving operational outcomes for refineries. According to reports, this self-learning AI-based product automatically connects to real-time machine data and streams to understand machine behavior and anomalies. According to the press release, the startup is also ramping up on boarding tech talent and strengthening its leadership team to improve overall customer experience and develop a strong product lineup. There are also talks about growing out of the startup phase and entering the big league. As per the statement, the new-age fault detection technology company analyses 508 kms of HPCL’s petroleum pipeline for detection and prediction of anomalies. Besides HPCL, other Indian petro companies such as ONGC launched an INR 100 crore startup fund.","excerpt":"Kerala-based startup Tranzmeo which has an AI-based product called T-Connect Overview that provides end-to-end anomaly forensics for better refinery operational outcomes raised pre-seed funding from Hindustan Petroleum Corporation Ltd.  The startup , founded in 2017 is part of NASSCOM’s 10,000 Startups accelerator program is founded by Safil Sunny who has 8+ years of experience in […]","categories":["AI News"],"tags":["anomaly detection"],"author_name":"Richa Bhatia","publish_date":"2018-12-27T08:23:15","publication_year":"2018","word_count":225,"keywords":["funding","machine learning","programming_languages:R","AI","anomaly detection","GAN","R","startup"],"extracted_tech_keywords":["AI","machine learning","R","GAN","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/this-kerala-startup-raised-pre-seed-funding-for-its-ai-based-anomaly-detection-tech-from-hpcl\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10015910,"title":"EY To Hire 9,000 AI &#038; ML-Skilled Professionals In India In 2021","content":"Ernst & Young (EY), a global professional services organisation, has announced that it would be hiring 9,000 professionals in India next year, who come from a STEM background and are skilled in artificial intelligence, machine learning, cybersecurity, analytics and other emerging technologies. According to news media, these hiring will be for various technology roles across all member firms, including the global delivery centres. This massive hiring aims at expanding its digital capabilities to help organisations solve their “complex end-to-end business transformation challenges”, the company stated. Speaking about these rings, Rohan Sachdev, the Partner and Consulting Practice Leader at EY India stated to the media that the company is making bold investments particularly in technology, data and through strategic acquisitions while continuing to expand its alliance and ecosystem relationships. As per data, currently, EY India has over 50,000 people working in all member firms, including global delivery centres, out of which 36% are from the STEM background. He further stated that the company uses its expertise and domain knowledge to help its clients, both in government and private businesses, to embark on technology-led transformation. With the exponential increase in digital adoption, it has become necessary for the company to strengthen its capabilities in emerging technology roles, thus significantly intensifying its hiring efforts in the coming year. The company has a holistic approach to digital transformation and innovation embedded across all its service offerings and sectors, including several proprietary digital tools and solutions such as EY Asterisk — a supply chain planning solutions, and EY Symphony — integrated governance, risk, controls and compliance platform. Mahesh Makhija, Partner and Technology Consulting Leader, EY India, also said that EY in India is working towards building a broader range of new digital proprietary tools and solutions to scale across organisations and geographies. The company has even launched ‘EY Techathon 2021: #iSolve4aBillion Challenge‘ to engage India’s best technology minds. In this challenge, students from all disciplines are invited to develop novel solutions for Indian demographics. The challenge will allow the participants to use blockchain, artificial intelligence, machine learning, and gamification to develop a model immunisation program across different spheres of the supply chain, delivery, monitoring and vaccination.","excerpt":"Ernst & Young (EY), a global professional services organisation, has announced that it would be hiring 9,000 professionals in India next year, who come from a STEM background and are skilled in artificial intelligence, machine learning, cybersecurity, analytics and other emerging technologies. According to news media, these hiring will be for various technology roles across […]","categories":["AI News"],"tags":["AI and ML","ai big data analytics digital transformation"],"author_name":"Sejuti Das","publish_date":"2020-12-24T14:08:23","publication_year":"2020","word_count":361,"keywords":["Go","machine learning","artificial intelligence","AI","ai big data analytics digital transformation","AI and ML","Git","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ey-to-hire-9000-ai-ml-skilled-professionals-in-india-in-2021\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21242,"title":"Can India Lead In The Global Robot Race?","content":"It may come as a surprise to some that India is counted as the Big 3 in the robot race but Malcolm Frank, head of strategy at Cognizant and author of What to Do When Machines Do Everything believes India alongside US & China is the frontrunner in the robot race. In India, the robot revolution isn’t just limited to assembly lines, edging out workers on factory floors, there is also a major shift towards AI coming from companies that make up its $143 billion outsourcing industry — a sector that employs nearly 4 million people, emphasized Frank.  Indian IT bellwethers like Infosys, Wipro and TCS have turned to automation in their operations to yield benefit. The line between hype and reality of RPA is blurred with Indian IT leaders bringing a “virtual integration” of multiple systems to perform repetitive work accurately and more reliably than humans can, notes Genpact. In an older survey conducted by Chartered Institute of Management Accountants (CIMA),  India ranked third in global rankings in implementing robotic automation in their core business processes. Automation in India picks up as a trend, Indian Robotics Companies Thrive Unlike our Asian counterpart Japan, India’s robotics industry hasn’t yet taken off and is limited to the manufacturing sector. When it comes to consumer robotics, Indian market hasn’t taken off yet. Of late, we are seeing a rise in the use of surgical robots. India is home to robot startups like Gurgaon-based GreyOrange Robotics, backed by Tiger Global, with its flagship ‘Butler System’ that provides warehouse automation for ecommerce giant like Flipkart. GreyOrange is one of the leading robotics and automation start-ups in India and Butler, a robot that sorts 1.2 crore packets a month, has revolutionized warehouse logistics in the country. Then there are startups like Bangalore-based Invento Robotics with their flagship product called MITRA, outfitted with speech recognition, voice recognition and face recognition capabilities and has been piloted in various firms across the domain. Another company Nandan Nilekani backed Systemantics also focuses on industrial robots. Plus, all the global industrial robotics companies like KUKA, FANUC and ABB have their base in India as well. The age of autonomous co-workers is already upon us with a new breed of co-workers taking up tasks ripe for automation. A recent report on manufacturing paints the current scenario – for example leading Indian automaker Maruti Suzuki’s factory in Haryana has 7,000 human employees working alongside 1,100 robots. According to Prasad Satyavolu, Cognizant’s Chief Digital Officer, Manufacturing, Logistics, Energy & Utilities, “Robots are already being deployed across manufacturing organisations, from inserting shock absorbers or cutting meat on a traditional assembly line, to drones acting as the eyes of security officers patrolling a vast container yard”.  Even in the manufacturing sector, the  capabilities of robots are expanding rapidly into asset-heavy supply chains sectors, helping India compete on a global level. And maybe in the not-too-distant future, we might even relinquish control to a robot completely. Number of factor that can contribute to India’s rise in robotics: India has the third-largest startup technology base and with recent government impetus — Make In India campaign, robotics companies are transforming the manufacturing hub with innovative solutions. Today, Indian robotics companies like GreyOrange are competing globally with large MNCs with their solutions. India’s vibrant startup ecosystem is bridging the gap by bringing their unique product across a range of industries like defense, logistic and warehouse management, manufacturing and academic research. Companies like Delhi-based Falcon Autotech, Gridbots, Precision Automation Robotics India (PARI), Gade systems and Hitech robotics are leading the charge in industrial robotics. As a country, India has the right ecosystem to become a robotics powerhouse — home to the largest embedded software services, wide talent pool with expertise in embedded systems, image processing, server engineering and developing software for robotics. India has made its ambitions to become a manufacturing superpower abundantly clear and in view of the Industrial Revolution 4.0, rolled out a manufacturing policy to push its share to 25% of GDP by significantly consolidating Make in India campaign. To bridge the knowledge gap, the Government recently tied up with Japan, the world leader in robotics to introduce AI and robotics in the defense sector. Over the years, the scope of robotics has increased in India and there has been a significant academic push to generate an interest in this field. For example, Indian Robot Olympiad (IRO) is one of the largest Robotics competition in India for students between the age of 9 to 21 and has been held since 2006. Besides, there are also a slew of open source tools and platforms available for robotics, that can lead to significant development in this field. For example, ROS — Robot Operating System, is a framework for writing software for robots, that includes various tools and libraries. You can check out more open source robotics\/hardware tools here. Another area where India scores a huge advantage is its lead in technologies such as IoT that can play a huge role in connecting countless devices and blur the lines between the physical and the online world. Besides, the Indian government’s INR 500 crores to support the rollout of 5G by 2020, a crucial network upgrade to support IoT will also give a boost to robotics. AIM View As USA and China begin to lay down laws governing AI, India too should play an active role in policymaking instead of just playing catch-up. A NASSCOM & PwC paper shared that instead of waiting for technology to reach a level where regulatory intervention becomes necessary, India should become a frontrunner by establishing a robust framework in advance. Besides policy shaping, another area where the Government can play a key role is incentivizing to institutes that provide robotics training with grants and subsidies. Also, urge polytechnics and private tech companies to offer more industry-oriented and practice-based technician degree programmes. Just upskilling is not enough, institutes and tech companies should enable access to open-source software libraries, toolkits and development tools to boost robotics development. According to Defence Minister Nirmala Sitharaman, “Industrial Revolution 4.0 is rapidly catching up. Whether you like it or not, some industries are bringing in robotics in a very big way. Some partly make use of it while others have not been impacted by this because they cannot afford it or they do not want it. But we have to have a place for all the three,” said Sitharaman, acknowledging the role of robotics in Industry 4.0.","excerpt":"It may come as a surprise to some that India is counted as the Big 3 in the robot race but Malcolm Frank, head of strategy at Cognizant and author of What to Do When Machines Do Everything believes India alongside US & China is the frontrunner in the robot race. In India, the robot […]","categories":["IT Services"],"tags":["Automation","Industry 4.0","robotics technology","RPA"],"author_name":"Richa Bhatia","publish_date":"2018-02-02T10:51:55","publication_year":"2018","word_count":1076,"keywords":["Go","API","robotics technology","Industry 4.0","AI","AWS","RPA","Automation","Git","automation","Aim","GAN","R"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","Git","API","GAN","automation","RPA"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-india-lead-global-robot-race\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10095733,"title":"Google Unveils AudioPaLM: Where Text Meets Voice","content":"Big tech Google, which is killing it in the generative AI domain, has introduced AudioPaLM, a new multimodal language model that is built by combining the capabilities of large language model PaLM-2 that Google unveiled in Google I\/O 2023 and its generative audio model AudioLM released last year. AudioPaLM establishes an all-encompassing multimodal framework proficient in processing and generating both textual content and spoken language. Read the full paper here. The applications of AudioPaLM are diverse, encompassing areas such as speech recognition and speech-to-speech translation. Leveraging the expertise of AudioLM, AudioPaLM inherits the capacity to capture non-verbal cues like speaker identification and intonation, while simultaneously integrating the linguistic knowledge embedded in text-based language models like PaLM-2. Moreover, AudioPaLM showcases distinctive features of audio language models, such as the ability to transfer a voice from one language to another based on a concise spoken prompt. AudioPaLM harnesses the power of a large-scale Transformer model as its fundamental framework. It expands upon a pre-existing text-based LLM by augmenting its vocabulary with specialised audio tokens. This, along with a basic task description, enables the training of a single decoder-only model capable of handling a blend of tasks involving both speech and text, in various combinations. These tasks encompass speech recognition, text-to-speech synthesis, and speech-to-speech translation. Through this approach, we consolidate traditionally segregated models into a unified architecture and training process. AudioPaLM achieves exceptional performance on speech translation benchmarks and delivers competitive outcomes in speech recognition tasks. It also exhibits the ability to convert speech into text for previously unseen language pairs without the need for prior training. In addition to speech generation, AudioPaLM can also generate transcripts, either in the original language or directly as a translation, or generate speech in the original source. AudioPaLM has achieved top results in speech translation benchmarks and has demonstrated competitive performance in speech recognition tasks. The model can also preserve paralinguistic information such as speaker identity and intonation, which is often lost in traditional speech-to-text translation systems. The system is expected to outperform existing solutions in terms of speech quality, based on automatic and human evaluation. “Further research opportunities exist in audio tokenisation, aiming to identify desirable audio token properties, develop measurement techniques, and optimise accordingly. Additionally, there is a need for more established benchmarks and metrics in generative audio tasks to make progress in research, as current benchmarks primarily focus on speech recognition and translation,” read the paper. Read more: LLMs Are Not As Smart As You Think Battle of Tech Giants in Music Generation Has Just Begun However, it is not the first time that Google has launched something in the audio generation space. Back in January, it released MusicLM, a high-fidelity music generative model that creates music from text descriptions, built on AudioLM, as well. It uses a hierarchical sequence-to-sequence approach to generate steady music at 24 kHz. It also introduced MusicCaps, a curated dataset of 5.5k music-text pairs designed for evaluating text-to-music generation. Google’s rivals are not far behind in this space, either. Microsoft recently launched Pengi, an audio language model that capitalises on transfer learning to audio tasks as text-generation tasks. By integrating both audio and text inputs, Pengi can generate free-form text output without additional fine-tuning. Moreover, Meta, spearheaded by Mark Zuckerberg, has introduced MusicGen, which harnesses the power of transformer architecture to create based on textual prompts, aligning the generated music with existing melodies. Similar to language models, MusicGen predicts the next section of a musical piece, resulting in coherent and structured compositions. It efficiently processes tokens in parallel using Meta’s EnCodec audio tokeniser. The model was trained on a dataset of 20,000 hours of licensed music, ensuring access to diverse musical styles and compositions. It also released Voicebox, a multilingual generative AI model that can perform various speech generation tasks through in-context learning, even tasks it was not explicitly trained for. However, Microsoft-backed OpenAI, which is currently regarded as the leader of the generative AI space seems to be lost in this race of music generation. The ChatGPT creator has made no recent announcements in this space.","excerpt":"AudioPaLM integrates the best of PaLM-2 and AudioLM.","categories":["AI News"],"tags":[],"author_name":"Shritama Saha","publish_date":"2023-06-26T12:11:53","publication_year":"2023","word_count":679,"keywords":["Go","ChatGPT","TPU","OpenAI","AI","R","RAG","Aim","generative AI","in-context learning"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Aim","RAG","in-context learning","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-unveils-audiopalm-where-text-meets-voice\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10058725,"title":"PyTorch releases TorchMetrics version 0.7","content":"PyTorch has launched TorchMetrics v0.7. The fresh release includes several new metrics (mainly for NLP), naming and import changes, general improvements to the API among other features. Version 0.7 has 60+ metrics that cover eight domains including audio, classification, image, pairwise, detection, regression, information retrieval, and text. The team has also unified TorchMetrics API across all domains and has reached over 600 stars. Currently, 1,500 repositories use TorchMetrics. The changes in NLP metrics include text package which in the v 0.7 includes a couple of machine translation metrics such as chrF, chrF++, Translation Edit Rate, or Extended Edit Distance. It now also supports other metrics — Match Error Rate, Word Information Lost, Word Information Preserved, and SQuAD evaluation metrics. The team has also made possible the evaluation of the ROUGE score using multiple references. Other updates include Translation Edit Rate (TER) which represents a normalized minimum number of edits required for a hypothesis that matches one of the provided references. Edits comprise insertion, deletion, and substation on a word or n-gram word level. Extended Edit Distance is also one of the updates that operate at a character level, and it extends Levenshtein distance using jump operations. Match Error Rate (MRE), a metric used to evaluate automatic speech recognition (ASR) systems, has been added. Word Information Lost (WIL) and Word Information Preserved (WIP) are two interwoven metrics used predominantly to evaluate ASR output. SQuAD, which stands for Stanford Question Answering Dataset, is another update. It is a machine comprehension dataset of 100,000+ pairs of questions and Wikipedia articles with selected answers. The dataset is used for the evaluation of extractive models. TorchMetrics v0.7 brings more extensive changes to how metrics should be imported. The import changes directly impact v0.7, which will require developers to change the import statement for some specific metrics. All naming changes follow the standard deprecation process. From v0.8, the old metric names will no longer be available. A complete list of metrics undergoing a naming change can be found in the changelog.","excerpt":"Version 0.7 has 60+ metrics that cover eight domains including audio, classification, image, pairwise, detection, regression, information retrieval, and text.","categories":["AI News"],"tags":[],"author_name":"Meeta Ramnani","publish_date":"2022-01-18T19:31:18","publication_year":"2022","word_count":335,"keywords":["Go","API","TPU","programming_languages:R","PyTorch","AI","programming_languages:Go","NLP","ai_frameworks:PyTorch","R"],"extracted_tech_keywords":["AI","NLP","PyTorch","TPU","R","Go","API","ai_frameworks:PyTorch","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-releases-torchmetrics-version-0-7\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":39883,"title":"10 Free Online Resources For Beginners To Learn Julia","content":"Julia is a flexible and dynamic programming language which is usually used for scientific computing. In a short span of time, the popularity of this language has grown exponentially due to its speed, dynamism and for ease to use. The demand for learning Julia has gone up manifolds and with this article, we list down 10 online free resources where you can learn Julia language from scratch. These resources will especially be quite useful for beginners who have never ventured onto this language before. (The list is in no particular order) 1| Official Documentation On Julia This is the official documentation on Julia language where you can learn the scientific computing language from scratch. You will learn about the strings, functions, types, variables, libraries, etc.   right from the basic introduction to developing and debugging Julia’s code. Read here. 2| JuliaCon 2015 Video Tutorial (Youtube) This comprehensive video tutorial on Julia programming is from the JuliaCon conference held in the year 2015. The Julia community actively participated in the conference to give various informative talks and workshops on Julia. The topics include areas such as solving optimisation problem with JuliaOpt, climate classification and clustering with Julia, cluster managers, parallel Julia, building web-powered applications in Julia, and others. Watch here. 3| Julia Bloggers (Blog) This blog is a tutorial on Julia language where you can learn about the basic concepts as well as the coding techniques. The topics covered here are differential equations, modelling toolkit, getting the value of a pointer in Julia, regression from scratch, random forest, Julia data science, and other such topics. Read here. 4| Julia Tutorial By MIT (Youtube) This video tutorial on Julia teaches the fundamentals of this programming language which was recorded during the MIT Independent Activities Period. It consists of a total of ten videos which cover topics such as rationale behind Julia and the vision, data analysis in Julia with Data Frames, statistical models, Fourier transforms, metaprogramming and macros in Julia, parallel and distributed computing with Julia, etc. Watch here. 5| Fast Track To Julia This can be called as the cheat sheet for learning Julia. This overview will help you to learn the basic concepts including operators, shell, libraries, package management, strings, characters, arrays, loops, functions, exceptions, etc. Read here. 6| Julia: A Fresh Approach To Numerical Computing (PDF) This paper introduces the design principle of the Julia language. You can gain insights on writing programs with and without types, leveraging language design for high-performance libraries, tools for numerical integrity, etc. Read here. 7| A Deep Introduction to Julia for Data Science and Scientific Computing This tutorial starts by introducing the basic syntax of Julia language and quickly moves into the details of how this language is different from other scripting languages and how to exploit Julia’s type system including multiple dispatches to be able to achieve C\/Fortran-like performance while maintaining the concise syntax of a scripting language. You will also learn how to design Julia projects. The major goal of this tutorial is to use Julia’s language and syntax to bridge the gap between “package users” and “package developers” in the way that Julia has done. Read here. 8| Julia Scientific Programming (Online Course) The course is divided into four modules and introduces Julia as a first language. The extensive assignments teach you the language with much ease and enable you to write your own Julia program from scratch. It also equips with understanding the advantages and capacities of Julia as a computing language, work in Jupyter notebooks and use  use various Julia packages such as Plots, DataFrames and Stat. Read here. 9| Julia Language: A Concise Tutorial The purpose of this tutorial is to help the readers to learn, understand and start coding in Julia. The format is a fusion between a classical tutorial and a cheat sheet where it describes the elements of the language following the typical sections of the programming language. Read here. 10| First Steps With Julia This Julia tutorial is provided by Kaggle which will help you to get started with the scientific computing language. You will gain insights on various features of Julia including the parallelisation, speed, basics of the language and implementation of the K Nearest Neighbour algorithms. This tutorial also focuses on the task of identifying characters from Google Street View images. Read here.","excerpt":"Julia is a flexible and dynamic programming language which is usually used for scientific computing. In a short span of time, the popularity of this language has grown exponentially due to its speed, dynamism and for ease to use. The demand for learning Julia has gone up manifolds and with this article, we list down […]","categories":["AI Trends"],"tags":["ai for beginners tutorial","Julia Language","julia scientific programming"],"author_name":"Ambika Choudhury","publish_date":"2019-05-29T12:55:23","publication_year":"2019","word_count":721,"keywords":["data science","Go","AI","ai for beginners tutorial","distributed computing","RAG","Julia Language","Ray","ViT","Julia","Jupyter","R","julia scientific programming"],"extracted_tech_keywords":["AI","data science","Ray","Jupyter","RAG","distributed computing","R","Go","Julia","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-free-online-resources-for-beginners-to-learn-julia\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":3801,"title":"Actuarial Sciences – Talent pool study by Recruise","content":"We must have come across Actuarial term often lately and wondered what’s it all about What is Actuarial? Actuaries Make Financial Sense of the Future “An actuary is a business professional who deals with the financial impact of risk and uncertainty. Actuaries provide expert assessments of financial security systems, with a focus on their complexity, their mathematics, and their mechanisms” Wondered how your insurance policies are framed, the timeliness, the maturity period, the amount, the risk factors, ULIPs –  all these background brain picking is done by the Actuarial analysts. And yeah they are much in demand and they are highly paid (hope the Analysts agree with us J ) Professional Actuaries find work opportunities in following sectors: Life insurance General insurance Health insurance Reinsurance Companies Pension funds Consultants Investments Government Academics Risk Management Examination: The study course is divided into four stages: Core Technical Stage : consists of 8 subjects each with one paper of 100 marks ( CT1 to CT8) and all are compulsory Core Application Stage :consists of 2 compulsory subjects ( CA 1 to CA3) , CA1 consists of two papers 9 CA 1 and CA 1 2) whereas CA2 and CA3 shall have one paper each Specialist Technical Stage :consists of 6 papers ( ST1 to ST6) and the students would required to choose any two Specialist Application Stage: consist of 6 papers (SA1 to SA6) and the student has to clear anyone. Remuneration: Actuarial professionals get an initial salary in the range of 5-10 lacs INR per year along with various other benefits. This would increase\/Improve based on the candidate’s experience, number of papers cleared and some other factors based on the organization Recruise Study:- BFSI sector is definitely booming across the world, so is the demand for Actuarial professionals. They are smart  good with numbers and hence can add value to the organizations .We gathered information on the availability of these professionals and came up with the below findings. Location wise Most of the Actuarial talent in India is found in the North Indian Region.  Delhi has a good number of professionals into Actuarial Sciences followed by Mumbai & Bangalore Given Below is an Experience wise split up of Actuarial professionals in India. Company wise Analysis From the below figure, Aon has a good number of resources into Actuarial Sciences, followed by Mercer & WNS. The Resources are spread across India. S. No Name Total Pool 1 Aon Hewitt 25% 2 Mercer 16% 3 WNS 11% 4 Towers Watson 11% 5 ICICI Prudential 8% 6 Ernst & Young 7% 7 Swiss Re 7% 8 GENPACT LLC 6% 9 Max New York Life 6% 10 Deloitte Consulting 4% Compensation Benchmark We have the compensation data analyzed from tier -1 & tier – 2 companies, professionals who have cleared 2 or more papers are paid high compared to those who have cleared 1-2 papers. We spoke to some candidates and when asked their compensation expectation from a prospective employer, they quoted 35-40% hike on their CTC. The appraisal season is mostly during the March\/April period and the candidates are given a 10-15% hike based on their performance. Tenure (yrs) Designation Level Company 1 Company 2 Company 3 Company 4 Company 5 Mean(INR) 0-2 Analyst\/Trainee 3.5-5 lacs 3-5 lacs 4 lacs-6 lacs 2.5 – 3.5 lacs 2 – 3 lacs 4-6 lacs 2-4 Sr Analyst\/TL 5-8 lacs 5 -9 lacs 4-8 lacs 3-6 lacs 4.5 – 7 lacs 5-8 lacs 4-6 A M\/Lead 7-12 lacs 6-12 lacs 8-15 lacs 5  -8 lacs 6-10 lacs 6-10 lacs 6-8 Specialist\/Mgr\/Sr Mgr 10-14 lacs 12-18 lacs 14-20 lacs 10-15 lacs 10-15 lacs 10-15 lacs †Names of companies withheld due to confidentiality reasons Institutions offering Actuarial courses in India The degrees are Bsc\/Msc or MBA.  Candidates usually take up a Bsc\/Msc and then become a member of the Institute of Actuaries of India. Educational Institutions offering Actuarial Programs Aligarh Muslim University,Aligarh Andhra University, Visakhapatnam University of Mumbai, Mumbai University of Delhi, Delhi AMITY School of Insurance and Actuarial Science,Noida(ASIAS) Berhampur University,Orissa University of Madras, Chennai RNIS College of Insurance, New Delhi Annamalai University,Tamil Nadu Birla Institute of Management Technology, New Delhi University of Kalyani, West Bengal Narsee Monjee Institute of Management Studies, Mumbai Bishop Herber College, Tiruchirappalli Jaipuria Institute of Management, Lucknow Manipur University ,Canchipur, Imphal Dr. Ram Manohar Lohia Avadh University, Faizabad, U.P. CMD School of Insurance and Actuarial Sciences, Uttarpradesh Institute for Integrated Learning in Management (IILM), Noida Kurukshetra University,Kurukshetra Gurunanak Dev University, Amritsar DS Actuarial Education Services (DS ActEd), Mumbai Goa University, Goa The University of Iowa (USA) City University – Cass Business School (UK) Purdue University (USA) UNSW Australian School of Business (Australia)","excerpt":"We must have come across Actuarial term often lately and wondered what’s it all about What is Actuarial? Actuaries Make Financial Sense of the Future  “An actuary is a business professional who deals with the financial impact of risk and uncertainty. Actuaries provide expert assessments of financial security systems, with a focus on their complexity, their mathematics, and their […]","categories":["AI Features"],"tags":["data scientist india salary"],"author_name":"Shankar Raman","publish_date":"2013-06-27T13:28:02","publication_year":"2013","word_count":777,"keywords":["Go","programming_languages:R","AI","data scientist india salary","programming_languages:Go","GAN","R"],"extracted_tech_keywords":["AI","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/actuarial-sciences-talent-pool-study-by-recruise\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":30574,"title":"Apple Acquires Silk Labs To Focus On AI-Based Personal Assistants And IoT Devices","content":"In a move to improve their artificial intelligence- based personal assistants and another internet of things projects, Apple Inc has acquired a San Francisco-based startup. Silk Labs is doing some cutting-edge work on next-generation visual and audio intelligence to connected products using deep neural networks. Reports have suggested that Apple completed the takeover relatively quietly and has not disclosed the sum publicly yet. Silk Labs has been founded by noted personalities like Andreas Gal, the former CTO of Mozilla and Michael Vines from Qualcomm. Silk Labs’ AI platform is heavy on privacy, which is very important for all Apple devices. It works as follows: Sends Only Key Learning Moments: Instead of sending a constant stream of video to the cloud for training, Silk’s advanced algorithms only capture key learning moments. Sends Only Anonymised Data: Protecting user privacy and data becomes even more critical when a continuous camera stream is involved. Rather than uploading raw data for training, Silk’s innovative learning service improves models using anonymized data that is not reversible back to the original content. Privacy By Design: With coding as well as design, Silk takes great measures to ensure that user data on the Silk Intelligence Platform is fully protected at all times. Apple, whose HomePod is a distant third behind Amazon’s Echo and Google’s Home, has been working hard towards improving its AI-enabled products. Reports have suggested that Amazon and Google account for 70 per cent share of the global smart speaker shipments in the first quarter of 2018, with Apple selling 600,000 HomePods in the period.","excerpt":"In a move to improve their artificial intelligence- based personal assistants and another internet of things projects, Apple Inc has acquired a San Francisco-based startup. Silk Labs is doing some cutting-edge work on next-generation visual and audio intelligence to connected products using deep neural networks. Reports have suggested that Apple completed the takeover relatively quietly and […]","categories":["AI News"],"tags":["Apple"],"author_name":"Prajakta Hebbar","publish_date":"2018-11-23T08:03:19","publication_year":"2018","word_count":259,"keywords":["Go","artificial intelligence","programming_languages:R","AI","neural network","Apple","ETL","programming_languages:Go","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","R","Go","ETL","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/apple-acquires-silk-labs-to-focus-on-homepod\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52235,"title":"Is There Room For More Players In The Global Microprocessor Market?","content":"Microprocessors, especially with the advent of artificial intelligence, have become an essential part of the technology. These microprocessors have found their applications in data centres, the automotive sector, IoT systems and client devices, among many others. As a result, many new companies have started to target this market which achieved the figure $4 billion in chip revenue in 2018. Microprocessors in this era have found their applications in artificial intelligence and have been mainly used in an attempt to accelerate the deep learning applications. For example, NVIDIA Drive, used in Tesla’s self-driving cars has instructional sets that accelerate the neural network to 8TFLOPs.This rapid adoption of deep learning has led to many vendors to up their game to satisfy such a wide range of applications and more importantly, the enhancement of complex neural network. Although the market study needs a thorough research on the market performance based on the applications, geography, end-user industries and architecture type, we take a look at the market form the processing units point of view. The Market and Top Competitors: Different vendors are trying to develop new products, and while some other vendors are trying to come with modern architecture to meet the demands of an AI chip for the applications of deep learning. These vendors are in a never-ending race to provide better performance and efficiency for the workloads that the microprocessor chips takes on. Reportedly, the global microprocessor market is following a growth trajectory of 1.16% from 2018 to 2024. The Deep learning bit isn’t the only one affecting the microprocessor market but also the increase in the cloud data storage. The cloud data storage increases the demand for servers and has subsequently improved the requirements for better performing microprocessor systems and the fact that 90% of the companies nowadays operate on a cloud gives it additional weight. And because of this, the Asia Pacific will hold a majority share in the market. The Two Major Players When it comes to the microprocessor market, we have important players in the market like NVIDIA, Intel, AMD, Qualcomm, Apple, Freescale, MediaTek, and Samsung LSI, among others. Among all of these, Intel has been the frontrunner for these past years, but things are shifting now. In 2018 Intel’s CPU shortage caused a major issue for the company, which created a bigger problem for the customers. In September 2019, even Microsoft announced that they are planning to move to AMD for their Surface laptop. Even after many years of holding a pivotal position in this industry and the shortage has caused the customers to try new alternatives quickly. And why shouldn’t they? Competitors like AMD have been using superior processes to make better chips and even if there is a slight shortage of by Intel, AMD has been capitalising the market share. With the competition by AMD’s new Ryzen CPUs, which surpass Intel’s CPUs, it is tough to say whether Intel’s announcement of the end of their shortage will bring about any silver lining. But, the next question to ask is — will AMD meet the same fate? Looking at the news from a month ago, AMD issued an update on the production of its generation Ryzen processors. It said that AMD has seen very high demand and as a result pushed its launch of its high-end 16 core desktop processor — Ryzen 9 3950X from September 2019 to November 2019. The main reason for AMD to delay the launch of the chips was that they weren’t meeting the required clock speeds that should be able to perform at the top. As a result of the fierce competition and fear of dishing out a chip whose performance might be underwhelming to their customers, AMD delayed the launch. Nonetheless, as of 2019, Intel still holds significant shares for microprocessors worldwide. Room For More? Is the market crowded? Well, if you look at the history of significant shareholders and stats on the microprocessor market, there are two major players, Intel and AMD, except that Intel is the only one. It holds 70% of the CPU market, while AMD controls about one-quarter of the same. Now, the outside competition has come and gone, but there have been competitions. It is okay if one has heard of Intel and AMD only because when it comes to WIndows processors, both these are pretty famous. But, other microprocessor manufacturers have all specialised in different segments of consumer electronics and are away from the PC market. Some examples of such companies are, Instruments Inc., Qualcomm Inc, ARM Holdings and Broadcom Limited. These manufacturers supply the chips for many of the world’s smartphones, and tablets like for example, Apple Inc.’s iPhone has made use of the processor designed by Samsung and Taiwan Semiconductor. The Future What AMD’s Ryzen chip has shown us is that it is possible to enter the market of microprocessor chips provided that you offer better and cheaper (value for money) products. Proof that a new entry is possible is that in the last 12-13 months, AMD shares have exceeded that of Intel’s but, it is essential to keep in mind that AMD is still far from catching up to Intel. It is true that with so many technological changes in the world like the advent of CNNs and boom in cloud computing, the microprocessor market is crowded but not if one offers a superior product.","excerpt":"Microprocessors, especially with the advent of artificial intelligence, have become an essential part of the technology. These microprocessors have found their applications in data centres, the automotive sector, IoT systems and client devices, among many others. As a result, many new companies have started to target this market which achieved the figure $4 billion in […]","categories":["AI Features"],"tags":["NVIDIA"],"author_name":"Sameer Balaganur","publish_date":"2019-12-21T14:00:00","publication_year":"2019","word_count":894,"keywords":["Go","API","artificial intelligence","AI","neural network","cloud computing","RAG","deep learning","CNN","NVIDIA","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","RAG","cloud computing","R","Go","API","CNN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-there-room-for-more-players-global-microprocessor-market\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":49877,"title":"Top AI-Based Tools That Help Developers In Code Editing","content":"As the classical coding tools are slackening developers, blue-chip companies like Microsoft, Facebook, and Google are now releasing AI-based tools like AI-assisted IntelliSense, Graph2diff, and Aroma, to empower engineers for coding and patching errors. These AI-powered tools assist developers with code suggestion based on the previous line of algorithms written by users. Predictive snippets are, in turn, helping businesses to accelerate products time-to-market, and have a competitive edge in the evolving tech landscape. In this article, we will take a closer look at some of these tools which are making developers’ lives easier in code editing: 1. AI-assisted IntelliSense: Microsoft on Monday introduced a new feature in Visual Studio Code called AI-assisted IntelliSense to improve development agility through IntelliCode. This feature provides AI-assisted suggestions and whole-line code completions. Initially, the Intellicode utilises GPT-2 — an OpenAI natural language processing model to generate synthetic text — along with their unsupervised learning approach that learns from top 3000 GitHub repositories to predict the next line codes. Base on the context of the previous lines of code, IntelliCode suggest the algorithm, thereby enabling developers to code quickly. Source: Microsoft However, they have now enhanced the model capabilities to allow each team for customising predictions based on the classes and libraries they often implement. Now the forecast of codes will incline towards the workflow developer adopt in their team. Training the model with bespoke code rather than GitHub repositories code will increase the accuracy of prediction. Keeping in mind the importance of security for any businesses, Microsoft has ensured that it will not share the algorithm that the firm will use to train the AI-assisted Intellisense with others unless the firm chose to do so. 2. AI-Assisted Refactoring: IntelliCode also tracks the changes developers make in the IDE to make suggestions for applying the repetitive changes quickly. This will assist them in refactoring the code with reduced manual efforts. Developers often optimise their code to improve the efficiency of the applications; thus, it is a must-have tool for them in order to expedite the development. Source: Microsoft To deliver such suggestions, Intellicode endorses AI-based programming-by-examples (PBE) to understand the code changes by user and later on suggest the same by anticipating their behaviour. This enables developers to alter the code just by clicks while saving the keystrokes and obtain the desired replaced code. The PBE has also been integrated into PowerBI as webpage tab extension. The Refactoring works with a few languages and will be extended to other programming languages in the near future. 3. Graph2diff: Every program has bugs and developers in a continual effort to provide a superior experience to users; they spend a considerable amount of time fixing those problems. But with the Graph2Diff, now they can automatically patch errors with the help of deep learning technology. It can localise and fix build errors and represent source code, compiler diagnostic, among others as a graph. Following that, it uses a neural network to predict the diff to identify the modification of code. The model is trained over thousands of datasets and build errors to determine the missing link in codes. The results have been exceptional with more than double the accuracy that various tools such as DeepDelta. 4. Deep TabNine: The auto-code completion tool uses deep learning that anticipates what the programmers need next. It is further enhancing its accuracy while also extending the support to other languages; it currently endowing more than 20 programming languages in various IDEs. 5. Aroma: It is a code-to-code search and recommendation tool that uses machine learning to suggest codes of other developers. With Aroma, users can check how other developers implement an algorithm to find best practices for accomplishing certain features. While one can manually search for code, Aroma also automatically retrieves look-alike snippets for them. Outlook The aforementioned AI-based tools are among the best in the class and have the potential to increase code quality by predicting the best codes for improving algorithm quality right from the beginning. Consequently, developers will not be required to optimise the code at a later stage. Today, AI-based tools are helping every business operations in an organisation and developers are no different. They need to leverage these solutions to streamline their developments and enhance their productivity with increased precision, accuracy, and speed.","excerpt":"As the classical coding tools are slackening developers, blue-chip companies like Microsoft, Facebook, and Google are now releasing AI-based tools like AI-assisted IntelliSense, Graph2diff, and Aroma, to empower engineers for coding and patching errors. These AI-powered tools assist developers with code suggestion based on the previous line of algorithms written by users. Predictive snippets are, […]","categories":[],"tags":[],"author_name":"Rohit Yadav","publish_date":"2019-11-14T15:00:00","publication_year":"2019","word_count":715,"keywords":["Go","machine learning","OpenAI","AI","neural network","ML","Git","RAG","deep learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","OpenAI","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-ai-based-tools-that-help-developers-in-code-editing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10059175,"title":"Council Post: How AI &#038; NLP are driving the digital transformation in insurance","content":"The insurance industry is undergoing a massive, tech-driven shift. The next decade will be crucial in deciding the future of the insurance sector. Industry leaders have a massive role to play, particularly in terms of adopting disruptive technologies throughout the value chain, starting from underwriting to policy servicing and claim settlement. According to Data Bridge Market Research, the AI in the insurance market is expected to touch $6.92 billion by 2028, growing at a CAGR of 24.05 percent for the forecast period of 2021 to 2028. The sector’s growth is expected to be fueled by AI technologies, including machine learning, deep learning, natural language processing (NLP) and robotic automation. Below, we discuss how insurance companies are leveraging AI, along with some use cases, challenges, and solutions. AI, NLP adoption For any business working in the insurance space, the first and foremost step is to list all the sub-processes within the value chain instead of solving the complete value chain or a chunk of processes together. This includes size, wider applicability, and complexity. Based on these parameters, the right processes should be prioritised for a minimum viable product (MVP). For example, a use case that involves extraction from two-three document types can give you volume, complexity, and wider applicability, such as email submission in underwriting. It is important to ensure the first use case is successful as it paves the path for other use cases. Once the first successful MVP implementation is set, a roadmap should be created for multiple AI-based proof-of-value (POVs) and integrate these use cases to deliver enhanced efficiency, effectiveness and customer experience. Challenges in deploying AI-at-scale Many global insurance companies’ technology and data science teams are exploring multiple generic products to solve structural problems. However, such products tend to reach a saturation point after a few easy, quick wins. Due to the limited capabilities of these generic products, some of the leading companies are struggling to deploy AI at scale, and are now looking at solving the next set of business challenges related to unstructured, handwritten, video and voice data. The major roadblocks in deploying AI at scale include: Continuous upgrades, and modifications in dependent systemsLimited business domain knowledge of tech teamsLack of human-in-the-loop concept Building comprehensive solutions to address these challenges is easier said than done. An end-to-end AI implementation leverages many tech systems, including ingestion from document management systems, to final posting into business applications such as policy admin system (PAS). While developing solutions, it is best to plan and accommodate all dependent systems upgrades or changes to avoid last-minute hurdles. Thus, timing and system flexibility are critical for smooth AI implementation. Moreover, successful AI implementation requires contributions from various resources, like, AI-NLP data scientists, data and tech engineers, business and project managers. However, as we move to solve the next level of challenges, it is important that tech teams upskill themselves and learn business nuances (understanding underwriters instructions). A deep understanding of business nuances will enable solutions that can address business complexities and multi-user functionality. Today, market expectations for 100 percent automation or state-thru-processing from AI solutions are reasonable, and current generalised products have been able to deliver this, albeit for simple problems. In my opinion, expecting 100 percent automation is the reason why these products are limited to straightforward cases. The way out of this problem is to accept the fact that machines cannot independently learn and solve problems and require human assistance. A well-known example that elucidates this better is self-driving cars or autonomous vehicles. While AI, NLP solution does its job with high precision, some instances are far too complex for machines to interpret. A common example of this is underwriting risk for customers who have either submitted partial or contradictory information. Human intervention is required in such cases to process contextual information.  Thus, human-in-loop enables ‘assisted’ ingestion of outputs by a human after augmenting business judgement. Use cases to consider There are multiple use cases that can be considered, including invoices, contracts, statements of values, endorsements, etc. Business submissions in the underwriting space is one such use case. It provides size, wider applicability and moderate to high complexity and can be prioritised over other use cases. However, the process requires interpretations from email and various unstructured documents (application quote, proposal, etc.). To extract information from multiple documents, numerous NLP models are required. Once these NLP models are created, they can be applied to a wider canvas for delivering AI at scale. Also, the submissions process for a transaction may stretch to a few months. However, AI can automate the process and reduce cycle time to a few days. In addition, the AI solution enables interpretation from emails and attached documents and provides underwriting assistants with the requisite information to review or modify to complete the transaction. Final thoughts To successfully implement AI, NLP solutions in the insurance segment, companies should adopt a case prioritisation framework based on size, wider applicability, and complexity. The companies should first re-draft their AI at scale roadmap, as generic products have limited scope. The tech teams, including AI data scientists, and data and tech engineers, should upskill their domain understanding. In addition to this, the AI-at-scale solution designs should be flexible and well thought through along with dependent systems. Lastly, human-in-the-loop is essential for any AI implementation. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"To successfully implement AI, NLP solutions in the insurance segment, companies should adopt a case prioritisation framework based on size, wider applicability, and complexity.","categories":["AI Features"],"tags":["AI Use Cases","Gen AI in Insurance","insurance","insurance analytics","Natural Language Processing","natural language processing ai","NLP"],"author_name":"Saurabh Khanna","publish_date":"2022-01-26T16:00:00","publication_year":"2022","word_count":925,"keywords":["data science","insurance","machine learning","TPU","AI","insurance analytics","Natural Language Processing","R","RAG","NLP","Aim","deep learning","Gen AI in Insurance","analytics","natural language processing ai","AI Use Cases"],"extracted_tech_keywords":["AI","machine learning","deep learning","NLP","data science","analytics","Aim","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-how-ai-nlp-are-driving-the-digital-transformation-in-insurance\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10048725,"title":"Why Are AI\/ML Job Titles So Vague?","content":"Artificial intelligence and machine learning jobs are among the most sought after by both freshers and experienced candidates. Seemingly glamorous work, the thrill of exploring a (relatively) new domain, a good pay package, and fancy job titles — these are some of the reasons driving this trend. AI and ML are vast fields and often encompass certain subfields. Some of the more common job titles include machine learning engineers, AI engineers, data analysts, and data scientists. Then there are a few unconventional ones like Intelligence Designer, Data Curator, Digital Knowledge Manager, and Machine Learning Data Scientist. Given how the field is still evolving, one can expect these job titles to diversify further. That said, as seen in other fields, job titles often fail to convey the actual nature of the job. Job titles are the first ways to represent to others what we do at work. Research shows that job titles can affect mental exhaustion and employee identity. That said, the importance of job titles is controversial, with some believing that they are meaningless and do not really influence career decisions. One can say that this debate also holds true for the field of AI and ML.  AI\/ML job titles vary too much, which complicates job searches. The Curious Case Of Job Titles AI is a young and rapidly growing field, which means no set rules or standards to name AI jobs properly. Even with the wide menu of titles that already exist, many more jobs are hastily clubbed with vaguely related titles for the lack of proper nomenclature. Various terminologies are associated with the AI discipline. Most frequently, job titles like data scientists and AI experts are the least informative; it would be very difficult to guess what is expected of a candidate at these positions without reading the full job description. AI\/ML and data science are overlapping fields. One can say that both machine learning and data science have components of AI, and hence the core skill set for different new roles are often common. “It is this overlap in the skill and the ability to work with the dataset that is causing the ambiguity in the roles. Over a period of time, we expect that these roles will evolve further, and the level of specialisation will make the distinction more clear,” Pravin Lal, CEO and Founder of Capital Quant Solutions, explains. The ambiguity or vagueness of job titles in these fields may also result from the penetration of these technologies into both IT and several non-IT fields. “On many occasions, AI, ML, Data Modeling, Data Science are used interchangeably in different spheres creating an ambiguity of roles. Most research reports emphasise that current job opportunities in AI\/ML are created in non-IT sectors like Banking, Marketing, Sales, Retail, and many other industries creating ambiguity in the naming conventions as these non-IT divisions are adding their flavours and conventions while creating job descriptions, thus, having a compounding effect on prevailing ambiguity in assigning job titles,” Bhavesh Goswami, Founder & CEO, CloudThat says. Many times, employers may voluntarily keep the titles more generic, given the multidisciplinary nature that involves the use of statistics, mathematics, software engineering, neural networks, analytics, and visualisation, among others. “Like any other field, there will be a need for specialists who come in with a very strong foundation in one or two areas as well as several generalists who would not only have a good understanding of the several aspects involved but also will be flexible enough to respond to the varying needs of projects. As this field continues to evolve, job titles are not likely to always be a good indicator of what is needed for the related roles. However, a good job description detailing the key technical skills needed to be successful in that role would often be adequate to inform job seekers of the type of work that is likely to be expected in the short term, with the clear understanding that the expectations from the same role within the firm will continue to evolve over time,” opines Satya Prakash Ranjan, Head of Analytics, Research and Data, Fidelity Investments India. How Does It Affect Employees? Often, a mismatch between the projected job role and the actual work may leave employees disappointed. This is especially true for freshers who come from the academic side of things and may be disillusioned about the actual work a particular job role entails. Combining this with a poor job description may leave the employee dissatisfied much early in their career. “I have seen so many companies and candidates getting disappointed, leading to high attrition rates in early careers. Both sides need to align aspirations with reality. Standardisation is key to the process, but it is still early, and the landscape is changing fast. The things that were considered quite advanced a few years back are common currency today. Educators, industry bodies and leaders must lead the standardisation process,” says Dipyaman Sanyal, Head of Academics and Learning at Hero Vired. “Job titles define the role and structure of an organisation. Most organisations have their own nomenclature to define the career progression of a candidate. Creating robust job descriptions that define the requirements of the role can help employees to understand the expectation from them,” says Padmini Giri, Associate Vice President – Talent Acquisition, GlobalLogic. Echoing a similar opinion, Lal of Capital Quant Solutions says, “It is important to define the expectations clearly at the time of recruitment with examples of work packets that the incumbent is expected to focus on so that when one starts the work in an organisation, there is no expectation mismatch leading to disappointment. When we interview candidates many times, we find them use these terms loosely, and the understanding is not clear in their minds. We make it a point to give examples of work packets that they would be focusing on to bring in clarity.” Having vagueness in job titles might not be entirely disadvantageous as believed. Nikhil Barshikar, Founder of Imarticus Learning, believes that not having a job title in some cases can be very liberating, especially for individuals with multiple skillsets, that in turn, don’t restrict them. “The inexistence of a psychological limitation might prove beneficial for learning outcomes and might help employees to explore untapped potential. This idea appears to be more functional in creative fields and might not work as well in the corporate world,” he adds. As mentioned above, AI and ML are continuously growing fields, and the problem of ambiguous job titles is unlikely to be solved overnight. The onus is on the recruiters to be as upfront as possible about the expectations right at the beginning.","excerpt":"AI is a young and rapidly growing field, which means no set rules or standards to name AI jobs properly.","categories":["AI Hirings"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-09-17T17:00:00","publication_year":"2021","word_count":1110,"keywords":["data science","Go","machine learning","artificial intelligence","AI","neural network","ML","Git","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","data science","analytics","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/why-are-ai-ml-job-titles-so-vague\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":35734,"title":"Check Out The Top 5 VC Firms Speeding Up AI Innovation In Indonesia","content":"Across the world, investment for AI startups is increasing steadily. According to a study by PwC and CB Insights, the venture capital funds for AI startups have increased by 72 per cent in 2018 and reached $9.3 billion. The Indonesian startup environment started thriving as early as 2010 and over the year, this small South-East Asian country has turned itself into a hub for the new and emerging tech startups.  In another study by Google and Temasek, Indonesia has the largest digital economy which is forecasted to be tripled from US$27 billion in 2018 to US$100 billion by 2025. Due to its large tech-savvy population, the country has also turned itself into a hub for VCs to invest in numerous startups that provides varied AI and ML-powered solutions. According to the Google study, e-commerce and fintech companies received the maximum funding with, AI-powered platforms tech-stacks bettering their service. Increasingly, more startups who are working in the field of NLP has also received a major boost from VCs Among the popular ones, Go-Jek, Traveloka, Bukalapak and Takopedia are some of the early names that have made a mark in the country, thus inspiring several others to take up the startup journey. In this article, we take a look at the popular VCA from the country and their investments in the past. GDP Venture: Founded in 2010, the ventures capitalist focuses on digital communities, media, commerce and solution companies in the Indonesian consumer internet industry. It also focuses on startups in the South, specifically South East Asia. Since 2016, it has made close to 10 investments and has also raised money for a number of AI startups.  Recently GDP it has closed a Series B funding for Singapore-based AI startup, 6Estates, an AI-driven intelligence company. In 2018, the venture capitalist raised $12 million in Series A funding forElement Inc, a US-based AI startup that develops and distributes a mobile-based software platform that creates a biometric identity. MDI Ventures : Headquartered in Asia-Pacific (APAC), Association of Southeast Asian Nations (ASEAN) and Southeast Asia, MDI Ventures was founded in 2015 and is a corporate venture capital initiative by Telkom Indonesia, the VC arm of state-owned telecom firm Telkom Indonesia. In the past, the venture capitalists have invested in companies in the online, media, and mobile internet space. In the past, the capitalists have led 33 investments, targeting startups from various sectors. Recently, Indonesia-based AI smart city startup, Qlue raised an undisclosed amount of funding from MDI Ventures. Convergences ventures: This is an early stage technology venture fund, which focuses on investing in Indonesia’s disruptive startups. The firm set up in 2014, have partners from the US, China and  Indonesia and have backed several startups over the year. Since its inception, the VC  firm has made 36 investments in startups that focus on emerging technologies like AI and ML. Till the date, the firm has supported AI startups in Indonesia with their biggest round of funding going to Kata.ai in 2017,  which targets the B2B segment through its Bahasa Indonesia-speaking chatbots Ventura Capital: Targets South-East Asian startups that are focused on consumer technology-enabled internet businesses. Founded in 2015, the firm is headquartered in Jakarta and aims to provide capital, operational empowerment, and market access to founders startups catering to E-commerce, fintech, edtech, and health tech startups. Since its inception, the firm has made 25 investments. Skystar Capital: This is a venture capital firm that focuses on technology-driven startups in the Asia Pacific region. SKystar collaborates with founders and provides the startups required assistance in the form of customised, strategic business advice and professional partnerships. Till its inception in 2014, it has made 12 investments so far and has targetted several AI, including the likes PHI Integration, a data management and analytics startup for an undisclosed amount","excerpt":"Across the world, investment for AI startups is increasing steadily. According to a study by PwC and CB Insights, the venture capital funds for AI startups have increased by 72 per cent in 2018 and reached $9.3 billion. The Indonesian startup environment started thriving as early as 2010 and over the year, this small South-East […]","categories":["AI Trends"],"tags":["Machine Learning","NLP","Startups"],"author_name":"Akshaya Asokan","publish_date":"2019-03-05T07:04:18","publication_year":"2019","word_count":631,"keywords":["Go","API","AI","chatbots","ML","Machine Learning","Git","NLP","Aim","analytics","Startups","R"],"extracted_tech_keywords":["AI","ML","NLP","analytics","Aim","chatbots","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/check-out-the-top-5-vc-firms-speeding-up-ai-innovation-in-indonesia\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10079169,"title":"Indian Startups Try Chinese Model, Will it Work?","content":"Pinduoduo, founded in 2015, has been a revelation for the Chinese e-commerce industry. After six years of its inception, Pinduoduo, which translates to ‘bringing lots of people together’, overtook Alibaba as the most popular e-commerce site in China. In 2020, Pinduoduo surpassed Alibaba in terms of active users on its platform with 788.4 million users. Alibaba had 779 million active users during the same period. Ex-Googler Huang Zheng, who founded Pinduoduo, said that the company is a combination of Costco and Disneyland. For years, Alibaba has been the undisputed champion in the Chinese e-commerce market. While there is still a long way to go for Pinduoduo to overtake Alibaba in terms of revenue, its meteoric rise in a short period of time did catch the attention of many. But the real question is how did Pinduoduo do it? Group buying The concept of group buying attempts to invoke the concept of community and mimics the offline shopping experience in a digital space. Such concepts of group or collective buying were popularised by US-based Groupon.com almost a decade ago. While collective buying is a bygone trend in the US, it is only now gaining traction in the east. Groupon allowed its users to shop collectively; however, they had to wait for enough people to commit to unlock the discounts. Pinduoduo has built its success by betting on this very concept. But what Pinduoduo does differently is that it allows its users to explore and find their own partners and avail the discounts. The Shanghai-based startup encourages users to shop with their contacts, thus bringing in new users to the application. Platform users also have the option to share the deal with their friends through WeChat. If a user is unable to find a partner, Pinduoduo connects the user with another user interested in the same deal. Besides, the fact that Pinduoduo’s platform can be accessed on WeChat, which has nearly a billion users, has helped the e-commerce company scale. Number of monthly active users of Pinduoduo Gamification Today, gamification is no longer a buzzword. In the context of e-commerce, gamification helps companies make their users’ shopping experiences fun and in that regard, Pinduoduo has championed it. Pinduoduo’s gamification of e-commerce is ingenious and the company has smartly integrated tradition into its gamification process. Hongbao, which translates to ‘a red envelope’, has gained popularity in China’s virtual space over the years. Traditionally, money wrapped in a red envelope is handed out to children during the new year and also to newlywed couples. The money is wrapped in a red envelope, which symbolises good fortune and protection against evil spirits in Chinese culture. Pinduoduo offers its users a Hongbao every time they log into the platform. Once the users open their Hongbao, they earn coins that are redeemable on both the platform and WeChat. Pinduoduo also offers a wide range of prize-yielding games to promote user-app interaction. Another popular game on the website involves the user choosing a tree and watering, fertilising it regularly. To do so, the user must log into the platform daily. In addition, to regularly water and fertilise the tree, users need to buy from the app, share offers, and invite friends. The catch here is that once this virtual tree matures, the user is rewarded with a box of real fruits. These activities help Pinduoduo lock users on their platform and enhance user-app interaction significantly. In addition, this has contributed to the rise of Pinduoduo as the most popular e-commerce platform in China. Will Pinduoduo’s business model work in India? The e-commerce space in India is also rapidly changing. While companies such as Amazon and Flipkart remain the bigger names, we have seen many companies with alternative business models enter the market over the years—Dealshare, Mela and Kiko Live. Pinduoduo’s smart business model has helped the company scale in a notably short duration, and the startup is already listed on NASDAQ. But can such a business model work in India just as well? In India, e-commerce in smaller cities in the country remains more or less an untapped potential. As per Statista, the total number of digital buyers in India was 289 million last year. Number of digital buyers in India ( Source: Statista) The e-commerce market in India caters mostly to the needs of affluent customers in bigger cities. Pinduoduo’s success is driven by its adoption by users that reside in Tier3 or other lower-tier cities. E-commerce companies with a distributed model that’s built on community and connection can find success in India too. Ours is a cohesive community, and the concept of group shopping has existed in India traditionally—much before it was introduced by Groupon on a virtual platform. Group buying platforms such as Gobillion and DealShare have found worthy success in India so far. DealShare targets middle or low-income families with hourly flash deals, which the company claims are cheaper than the prices offered in local Kirana stores. Similar to DealShare, Gobillion also focuses on serving customers in Tier2 Indian cities and beyond and caters to the middle- to low-income customer segment. Mela, another alternative e-commerce startup founded in 2019, even calls itself the ‘Pinduoduo of India’. Mela allows its users to shop through social platforms such as WhatsApp and Facebook and avail discounts. However, it remains to be seen whether these companies can build a profitable business by catering to customers in smaller Indian cities. Earlier this year, DealShare said the company expects to hit a revenue of USD1 billion in the near term and become profitable in the next two to three years. For Pinduoduo, dethroning Alibaba as the most popular e-commerce platform was not an easy task. What Pinduoduo did was it kept ploughing revenue back into subsidies while earning money from advertisements.","excerpt":"The concept of group or collective buying was popularised by US-based Groupon.com almost a decade ago.","categories":["Deep Tech"],"tags":["AI Startups"],"author_name":"Pritam Bordoloi","publish_date":"2022-11-09T15:00:00","publication_year":"2022","word_count":961,"keywords":["Go","API","AI","RPA","Git","RAG","Aim","ViT","R","AI Startups","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","API","ViT","RPA","startup"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/indian-startups-try-chinese-model-will-it-work\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":21277,"title":"Computer Vision Primer: How AI Sees An Image","content":"Computer vision technology is being dominated by the Convolutional Neural Network (ConvNet or ConvNet) because of its high accuracy. As we all know, a computer reads data in the form of numbers and uses different mathematical computation functions to build and produce certain results. Let us look at how the computer “reads” an image with the help of ConvNet — that is, how it collects the data from the image and computes it. All of us have been “artists” at some point in our lives and have built and drawn coloured figures with different shades to give it a beautiful texture. The way in which we differentiate what we see is through a human’s perception of colour. Let us look at an example of different colours on the pallet. We have experimented with a wide range of colours, mixed them together to give a darker, lighter or a different shade of colour depending on the ratio we use. Similarly, an AI reads these colours with a range of values from 0 to 255. The image below reminds us of how we used to use this combination to remember the six major colours. Its called as an RGB model. Here, Red, Green and Blue are the dominant colours, which when combined together give another set of colours — Magenta, Cyan and Yellow. Below is a representational  image an experience we have gone through. We have been introduced to this visual perception which tells us that white is a combination of multiple colours. Similarly, our physics experiment on the prism gives us the same information on how a white light splits into different colours. What Does A Computer See? Every image on our computer screen is nothing but a combination of the three major colours, red, green and blue. There are a lot of models being used in the computer vision field, which are nothing but a combination of various colour values. We will be using a simple RGB model to help us understand how a computer looks at an image. Let us see what an RGB model is. An RGB model is one of the oldest type of colour differentiation tool being used in the computer vision industry. As every major colour has a range from 0 to 255, we can infer that higher the value, the brighter is the colour. Let us differentiate every colour into sub categories creating a colour pallet. Traversing From 0 to 255: In the image above, every value of red, green and blue represents a particular shade of colour. These numbers are used by the AI to read the image and process it. When we combine two colours, say red and green, the resulting colour is Yellow. It is represented in the three dimensional space as 255,255,0 (R,G,B). With combination of RGB together in the .gif above, we can look at various colours such as Cyan, Magenta and Yellow. This is the technique used in computer vision to give a particular colour to a pixel. An image is made up of pixels placed adjacent to one another. These coloured pixels are made up of three channels which are placed one behind the another. All the channels add up to give a specific colour and these pixels placed together give a shape to a figure in an image. Let us look at how an AI reads an image. By implementing the RGB model we can extract over 16 million (16,777,216, to be exact) shades of colours. Let us look at a combination of RGB which makes up different pixels. RGB Values Of A Pixel In An Image: The above pixelated image is a combination of the three channels which are placed one behind the other. Lets us look at how these channels are placed in order to get our pixelated image like the above. This is how an image is formed on a computer screen. The values pertaining to red green and blue are read by an AI and stored in the form of a matrix. Let us read this image in Python and check the values of the dimensions of the pixels — 4×4 gives us 16 pixels. There are some difference in the values because we have created the above images in Windows Paint by giving specific RGB values and read the image in Python. The actual image size was 222×217 and we used K – Nearest Neighbour (k-NN) to resize the image into a 4×4 pixels with a depth of 3 to understand the image better. Next challenge was to search this image in a collection of images to check if it belonged to the actual image from which it was taken. Image Recognition With ConvNet: Now the AI has reads every pixel of the image, extracts the values of RGB and stores it in the memory. It starts searching for similar images in the database to find a match. But how does this work? What’s the principle behind the comparison of different images? Let us look into it In machine learning, ConvNet are complex feed forward neural networks. ConvNets is used for image classification and recognition because of its high accuracy. It was proposed by computer scientist Yann LeCun in the late ‘90s, when he was inspired from the human visual perception of recognising things. The ConvNet follows a hierarchical model which works on building a network, like a funnel, and finally gives out a fully-connected layer where all the neurons are connected to each other and the output is processed. During the training of ConvNet, as the images are being trained, the hidden layer is where the image gets broken down. This is called convolution. Every image consists of n x m number of pixels with a certain depth (RGB has 3 and GrayScale has 1). In a ConvNet, these images are broken down with two processes, that is, filtering and pooling and are squashed into a small image. Filtering is a process in which a stride of weight matrix is passed over the whole image in a number of iterations to get the dot product of the initial weights as well as the pixel values. This is followed by a pooling layer which reduces the size of a image into a low dimensional matrix, also highlighting a pixel, which gives the highest information about the image. Let us look at this .gif which gives us a visual understanding of the process. Once we have a convoluted layer, we can build a fully-connected layer to form an array. This information can be compared to the images to find the probabilities or likelihood of that portion being included in the image. Having a threshold at 50%, we can start predicting what kind or what the image is. One of the commonly used activation functions in this process is a softmax function. We can change the parameters and build the network with high number of iterations and higher number of layers to yield better results. This will depend on the computational power of your system. Conclusion While building a ConvNet can be time consuming, the results are fascinating, and totally worth the effort. Therefore, it is no surprise that this method is most popular in AI, as the outcomes are significantly better than other computer vision techniques such as OpenCV. Apart from RGB other colour models such as HSI and CMYK also yield good results.","excerpt":"Computer vision technology is being dominated by the Convolutional Neural Network (ConvNet or ConvNet) because of its high accuracy. As we all know, a computer reads data in the form of numbers and uses different mathematical computation functions to build and produce certain results. Let us look at how the computer “reads” an image with […]","categories":["AI Features"],"tags":["Neural Network"],"author_name":"Kishan Maladkar","publish_date":"2018-02-05T03:53:08","publication_year":"2018","word_count":1232,"keywords":["Neural Network","machine learning","TPU","AI","neural network","image recognition","computer vision","OpenCV","Python","Ray","R"],"extracted_tech_keywords":["AI","machine learning","neural network","computer vision","Ray","OpenCV","image recognition","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/computer-vision-primer-how-ai-sees-an-image\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10063961,"title":"Oracle releases new machine learning integrations with its MySQL HeatWave cloud service","content":"Oracle MySQL HeatWave now supports in-database machine learning (ML) in addition to the previously available transaction processing and analytics—the only MySQL cloud database service to do so. MySQL HeatWave ML fully automates the ML lifecycle and stores all trained models inside the MySQL database, eliminating the need to move data or the model to a machine learning tool or service. Eliminating ETL reduces application complexity, lowers cost, and improves the security of both the data and the model. HeatWave ML is included with the MySQL HeatWave database cloud service in all 37 Oracle Cloud Infrastructure (OCI) regions. HeatWave ML offers the following capabilities: Fully Automated Model TrainingModel and Inference ExplanationsHyper-Parameter TuningAlgorithm SelectionIntelligent Data SamplingFeature Selection In addition to machine learning capabilities, Oracle released more innovations to the MySQL HeatWave service. Real-time elasticity enables customers to upsize and downsize their HeatWave cluster to any number of nodes, without any downtime or read-only time, and without the need to manually rebalance the cluster. Also included is data compression, which enables customers to process twice the amount of data per node and lowers costs by nearly 50 per cent, while maintaining the same price-performance ratio. Finally, a new pause-and-resume function enables customers to pause HeatWave to save costs. Upon resuming, both the data and the statistics needed for MySQL Autopilot are automatically reloaded into HeatWave.","excerpt":"HeatWave ML fully automates model training, inference, and explanation.","categories":["AI News"],"tags":["Data Security","Machine Learning","MySQL","Oracle","Oracle SQL","SQL"],"author_name":"Kartik Wali","publish_date":"2022-03-30T14:29:17","publication_year":"2022","word_count":222,"keywords":["Oracle SQL","Go","machine learning","programming_languages:R","AI","Data Security","ETL","ML","innovation","Machine Learning","Oracle","analytics","SQL","MySQL","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","R","SQL","Go","ETL","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/oracle-releases-new-machine-learning-integrations-with-its-mysql-heatwave-cloud-service\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10073756,"title":"Nvidia GTC 2022 is Happening. Here&#8217;s What to Expect","content":"One of the most-awaited developer conferences, Nvidia GTC, is just around the corner. Scheduled from 19-22 next month, the event is expected to bring thousands of innovators, researchers, thought leaders, and decision-makers together to showcase the latest technology innovations in AI, gaming, computer graphics, metaverse and more. The thought leaders include Turing award winners Yoshua Bengio, Geoff Hinton, Yann LeCun and others. The Nvidia GTC would feature a keynote by Nvidia chief Jensen Huang and hold over 200 sessions with global business and technology leaders. The keynote announcement by Huang will be live-streamed on Tuesday, September 20, at 8:30 PM IST (8 AM PT). Click here to register. Nvidia has been the foundation of technology innovations, modern applications and computing platforms. Since its inception in 1993, the company has dedicated itself to the computing arena, starting from enhancing general-purpose computing to revolutionising the gaming and entertainment industry, pioneering GPU-accelerated computing, and later branching out to scientific computing, artificial intelligence, data platforms, and most recently – metaverse, and quantum computing, among others. The company has dedicated itself to solving problems in the computing arena, starting from building hardware products to software tools, gaming capabilities, architecture, etc. It looks to help people take their ideas into reality faster. Around 2009, one of the important milestones was the design of the next-generation CUDA, code-named Fermi, where Nvidia essentially solved the GPU computing puzzle. CUDA, or compute unified device architecture, is designed to work with programming languages like C, C++, and Fortran. This makes it easier for developers to use GPU resources effectively. It also supports multiple programming frameworks, including OpenMP, OpenACC, OpenCL, and others. Cut to 2022; the company is replicating CUDA’s success with quantum computing, where it recently launched QODA (quantum optimised device architecture). Last month, the company open-sourced QODA to accelerate quantum research and development across various areas, including health, finance, HPC (high-performance computing), AI and others. There is no stopping Nvidia. Earlier this month, the company announced a wide range of Metaverse initiatives. The company plans to bridge the gap between AI and the digital world, creating a more realistic Metaverse. Now, with all of these advancements in the backdrop, Nvidia’s GTC, which was started in 2009 onwards, provides a platform to understand general processing computing and challenges in the field, alongside the launch of futuristic technology where Nvidia’s in-house experts and researchers are experimenting. In an interview with Analytics India Magazine, Vishal Dhupar, managing director, Asia South at Nvidia, said they can synthesise the virtual worlds with physical worlds as they sit at the intersection of computer graphics, physics and intelligence. “That’s what people come to see. That’s what people imbibe. That’s what people practise. That’s why GTC,” he added. What to expect at Nvidia GTC? Omniverse Nvidia’s Omniverse has been the talk of the town. The platform offers developers a collaboration and scalable multi-GPU, real-time, true-to-reality simulation. The company believes it will revolutionise how people create and develop as individuals and work together as teams, bringing creative possibilities and efficiency to 3D creators, developers and enterprises. At GTC, the company will announce various updates, libraries and new tools and applications to create immersive AI chatbots, realistic avatars, and stunning 3D virtual worlds. Realistic Avatars: Recently, Nvidia announced the launch of lifelike avatars that can give an animated human face to the computers that people could interact with online. This might be similar to what Meta AI researchers developed, called MyoSuite. This new tool creates realistic musculoskeletal models more efficiently than exercising ones. Given Nvidia’s rich history in revolutionising the gaming and entertainment industry, there is a glimmer of hope from Nvidia to help developers create more realistic and life-like avatars. Metaverse bots: Nvidia is most likely to launch new capabilities and AI platforms to develop realistic avatars and characters that would help people navigate the digital world. 3D Rendering models: In March 2022, Nvidia announced the launch of Instant NeRF, touted to be one of the fastest techniques to data, achieving more than 1,000x speedups in some cases. It is a neural rendering model that learns a high-resolution 3D scene in seconds and can render images in milliseconds. Last year, Nvidia launched GANverse 3D, which can be imported as an extension in the Nvidia Omniverse to render 3D objects accurately in the virtual world. We can expect new updates and announcements around 3D rendering models at the upcoming GTC. “Thanks to our network and computing effect, which is taking place because of our accelerated computing capabilities, we can go into our imaginations and make it real, and we can all create our own world and uniformly create many worlds,” said Dhupar, excitedly, pointing at the multiple possibilities on Omniverse. He further said that AI has a huge role to play in creating such a 3D world, where machines\/bots can write their own piece of software, which humans can drive, and later can learn on themselves and, most importantly, make recommendations, and predictions based on the interaction in the metaverse. Nvidia currently offers Omniverse Enterprise, where it looks to help enterprises build 3D design and digital twin workflows with real-time collaboration and true-to-reality simulation. At GTC, there might be announcements of new partnerships on how companies leverage its Omniverse Enterprise platform to create various use cases, including robotic process automation, fighting climate change, automobile design, and more. Quantum Computing Banking on the success of CUDA, which opened up a new type of hardware and programming paradigm, Nvidia is betting big on quantum computing to help develop an ecosystem of hybrid quantum applications running on top of QODA. Citing examples of CUDA, Dhupar said QODA allows developers to run quantum simulations. “You can write a lot of test applications, and we can get ready when quantum hardware really comes into play,” he added, saying that it is quite simpler to use than how one would typically operate classical computing. “It helps quantum computing scientists to write algorithms and test their applications and get to the next level using the GPU where instead of one or two bits, you can write into hundreds of bits, qubits and move forward onto it,” explained Dhupar. At GTC, the company would announce some of the latest use cases and updates of its platforms, alongside the latest partnership and collaboration to accelerate quantum computing research across the globe. Hardware, AI chips and more Previously, Nvidia had said that it would launch BlueField-4 by 2023. The data processing unit BlueField supports CUDA parallel programming platform and Nvidia AI, turbocharging the in-network computer vision. The company had also announced the launch of Nvidia Grace, the first data centre CPU, an Arm-based processor that will deliver 10x the performance of today’s fastest servers on the most complex AI and HPC workloads. “This is the only company that talks about three processors, the CPU, GPU, and DPU; about accelerating applications across multiple domains; about a recent problem holding you back, and how you create that into a solution that becomes a mega-market,” said Dhupar, hinting that the company would announce major hardware and semiconductor chip updates. At GTC, we can expect the company to launch new hardware for the data centre, CPUs, GPUs, and others, along the lens of x86 architecture and the size of computing. Text-to-image tools In 2019, Nvidia introduced GauGAN, an AI tool that turns sketches into photorealistic landscapes. Of late, there has been a lot of buzz around image generation tools such as Meta’s ‘Make a Scene’, OpenAI’s DALL.E-2 and Midjourney, among others. There is a high chance of Nvidia making similar announcements around the release of text-to-image models and platforms. Autonomous vehicle At last year’s GTC, Nvidia announced the Nvidia DRIVE, powered by Hyperion 8. It is an end-to-end modular development platform and reference architecture for designing autonomous vehicles (AVs). This includes the NVIDIA DRIVE AGX Orin™, DRIVE AGX Pegasus, and DRIVE Hyperion 8.1 Developer Kits, all built on the NVIDIA DRIVE Orin system-on-a-chip (SoC). Nvidia’s Dhupar did not disclose much about NVIDIA DRIVE Hyperion. However, he said that there are a lot of things, whether, from a computing or software perspective, there would be talks around all of them. “Every field that we spoke of is going through the greatest technology shift – what people call Web 3.0, some call it metaverse, and everything that gets done between that aspects is something we should be looking forward to,” shared Dhupar.","excerpt":"Nvidia GTC will feature industry thought leaders and Turing award winners Yoshua Bengio, Geoff Hinton, Yann LeCun and others","categories":["AI News"],"tags":["NVIDIA"],"author_name":"Amit Naik","publish_date":"2022-08-28T13:00:00","publication_year":"2022","word_count":1397,"keywords":["Meta AI","artificial intelligence","OpenAI","AI","chatbots","ML","computer vision","RAG","analytics","NVIDIA","GPU computing"],"extracted_tech_keywords":["AI","artificial intelligence","ML","computer vision","analytics","OpenAI","Meta AI","RAG","chatbots","GPU computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-gtc-2022-is-happening-heres-what-to-expect\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":46622,"title":"5 Game-Changing Semiconductor Chips By Chinese Tech Giants","content":"China is pushing hard for the development in the chip-making space. The rise of security and trade has led the country to self-design intelligent and computing chips to avoid dependence on other countries. In this article, we list down five Chinese giants who are self-designing semiconductor chips for edge computing. 1| Alibaba Alibaba, the Chinese multinational conglomerate holding company has recently self-developed an AI inference chip known as Hanguang 800. Besides this, the company has also made a breakthrough in the semiconductor market by unveiling its first RISC-V (Reduced Instruction Set Computer) processor. The Hangzhou-based tech giant developed the AI chip mainly for cloud-based large-scale AI inferencing. This hardcore neural processing chip with an AI cloud service is claimed to power and speed up machine learning tasks and is claimed to be more efficient than traditional GPU-based services. Currently, the company is utilising this chip for internal business purposes such as personalised recommendations, product searches, intelligent customer services, etc. Also, the company’s semiconductor chip subsidiary, Ping-Tou-Ge developed Xuantie 910 mainly for the self-development of 5G, AI and IoT technologies. 2| Baidu Last year in July, the Chinese multinational tech company, Baidu unveiled its AI chip known as Kunlun. This chip is China’s first cloud-to-edge AI chip which has a computational capacity 30 times faster than an FPGA-based accelerator. The company designed this chip in order to perform edge computing on the device itself and cloud through data centres. The Kunlun 818-300 model will be used for training AI, while the Kunlun 818-100 will be used for inference. Furthermore, the company has claimed to support a wide variety of AI applications such as image recognition, autonomous driving, natural language processing, voice recognition, search ranking, and large-scale recommendations. 3| Huawei Chinese tech giant, Huawei has been making move in the chip manufacturing space for a few years now. Not only the company is unveiling semiconductor chips but also it has launched the AI-Native database, performance distributed storage as well as an open-sourcing operating system which can be used regardless of any systems. These breakthroughs will enhance the development of emerging technologies as well as enhancing the Huawei’s cloud capabilities. Huawei owned HiSilicon Technologies is a semiconductor industry based in Shenzhen, Guangdong. The company is popularly known as the largest domestic designer of integrated circuits in China. It has manufactured Kirin 970 processor which is a Neural Processing Unit (NPU). This chip enables cloud-based AI and on-device AI to run alongside each other and can perform complex computation tasks in a faster manner. Last year, the tech giant unveiled Ascend 310 chip which is said to provide the key support as the core component in Huawei’s AI solution. Furthermore, the company recently launched its first commercial AI chips known as Kirin 990 and Kirin 990 5G. Huawei claims this chipset as the 1st flagship 5G SoC powered with a 7nm+ EUV process. The chipset includes the company’s self-developed 2+1 Da Vinci architecture NPU i.e. three cores which deliver better power efficiency, stronger processing capabilities and higher accuracy. 4| Gree Electric Appliances Gree Electric Appliances is a Zhuhai-based air conditioner giant which is also pushing itself to design and develop its own computing chips. According to reports, the company has established a subsidiary chip company, Zhuhai Zero-Boundary Integrated Circuit Limited which is currently focusing on chip design for Gree’s key products such as air conditioners. Furthermore, the company will also focus on research and development of semiconductors, integrated circuits, electronic devices, and components, as well as other electronic products and services. 5| Cambricon Technologies Last year, Beijing-based chip maker Cambricon Technologies officially released the first artificial intelligence (AI) chip for cloud-based services in Shanghai. Cambricon 1M chip is the company’s third-generation AI chip for edge computing which provides efficiency of 5 TOPS\/Watt for 8-bit computing. Outlook By looking at these advancements, it is clear for all to see that the Chinese tech giants are rapidly shifting their interests in the chipmaking space. Since the trade war with the US, China has been in the news due to the dependencies on hardware technologies. Since then, China is accelerating its efforts in all possible ways to acquire self-sufficiency in this industry.","excerpt":"China is pushing hard for the development in the chip-making space. The rise of security and trade has led the country to self-design intelligent and computing chips to avoid dependence on other countries. In this article, we list down five Chinese giants who are self-designing semiconductor chips for edge computing.  1| Alibaba Alibaba, the Chinese […]","categories":["AI Trends"],"tags":["China"],"author_name":"Ambika Choudhury","publish_date":"2019-09-30T17:00:25","publication_year":"2019","word_count":695,"keywords":["API","machine learning","artificial intelligence","AI","image recognition","RAG","Aim","edge AI","edge computing","R","China"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","edge AI","RAG","image recognition","edge computing","R","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-game-changing-semiconductor-chips-by-chinese-tech-giants\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":28075,"title":"MachineHack Winners: How 3 Self-Taught Data Scientists Cracked The Problem Of Real Estate Pricing In Bengaluru","content":"MachineHack recently concluded the Predicting House Prices In Bengaluru hackathon and announced prizes in the form of exclusive individual passes to Cypher, India’s biggest analytics event. Analytics India Magazine talked to the winners of the hackathon and found out their experience of participating and winning the hackathon. Pravin Mhaske took the first rank on the leaderboard. Mhaske has been working with Infosys for about 15 years now. He manages projects professionally and plays with numbers as a passion. He got attracted to the field because of the “big data” buzz in 2016. Since then he has spent a lot of time polishing and learning basic skills in data science like inferential statistics, probability, linear\/vector algebra and calculus, from various sources like Udacity, Khan Academy, edX and the 3Blue1Brown videos on YouTube. Mhaske also completed courses on R, exploratory data analysis and Python programming from Coursera, Udacity and IIT Madras (NPTEL) followed by data analytics course by IIT-M (NPTEL) which he found to be excellent. Sushil Nath Gupta who took the second rank on the leaderboard is an undergraduate student pursuing B Tech (CSE) from Kurukshetra Institute of Technology and Management, Kurukshetra. Gupta is a self-motivated and enthusiastic learner of the data science. He used Youtube, Coursera, edx, and blogs as learning material. He tries to leverage social media and the web search as much as possible for help. According to Gupta, self-study is the best way to make oneself better in a particular domain. On the other hand, he feels that everyone needs to follow the correct path that makes sense to the personal data science journey. He started his journey of data science by choosing a favourite programming language that is Python. He took the data science Bootcamp course from Udemy. Gupta feels that the course teaches data cleaning and applied machine learning very well but it also helped him get his hands a dirty in data science. He eventually discovered hackathons and started participating actively. Sivaram Kanagaraj who took the third place on the leaderboard is from Salem, Tamil Nadu and has got an engineering degree in Mechatronics from Sri Krishna College of Engineering and Technology. After his engineering degree, he worked in Accenture, where he worked in projects based on MIS reporting and visualisations using Tableau and Power BI. He also studied SAS and thus got acquainted with the power of analytics in the business world. At Accenture, he mainly worked on descriptive analytics and data visualisation aspects. To improve his data science career Kanagaraj decided to upskill himself in advanced analytics and is currently pursuing Post Graduate Program in Business Analytics from Praxis Business School, Kolkata.","excerpt":"MachineHack recently concluded the Predicting House Prices In Bengaluru hackathon and announced prizes in the form of exclusive individual passes to Cypher, India’s biggest analytics event. Analytics India Magazine talked to the winners of the hackathon and found out their experience of participating and winning the hackathon. Pravin Mhaske took the first rank on the […]","categories":[],"tags":["learning","Machinehack"],"author_name":"Abhijeet Katte","publish_date":"2018-09-06T10:17:45","publication_year":"2018","word_count":440,"keywords":["big data","data science","Go","machine learning","learning","AI","Machinehack","RAG","Python","analytics","programming_languages:Python","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","RAG","Python","R","Go","big data","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/machinehack-winners-how-3-self-taught-data-scientists-cracked-the-problem-of-real-estate-pricing-in-bengaluru\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":26888,"title":"How Google’s BigQuery ML Is Empowering Data Analysts","content":"In case you were wondering, here’s another sign of the Google Cloud Vs Amazon Web Services war heating up. Google has now brought in the big guns in the analytical data warehousing space with by embedding machine learning capabilities into Google BigQuery. Google BigQuery is an analytics service, low-cost enterprise data warehouse which has now been rebranded as BigQuery ML. One of the key features of BigQuery is that it transforms SQL queries into complex execution plans, dispatching them onto execution nodes to promptly provide insights into the data. BigQuery enables developers to execute SQL as a massively parallel processing query with hundreds of CPU cores and ample disk storage, scanning and aggregating terabytes of data in seconds. BigQuery ML, a capability inside BigQuery enables analysts and data scientists to build and deploy ML models on massive structured or semi-structured datasets. Dremel Technology Is The Key At a time when Hadoop was facing intense competition, Google released beta access to BigQuery, a new SQL processing system based on Dremel technology, a distributed query engine in 2011. And BigQuery provides the core set of features available in Dremel to third-party developers. A key point of Dremel technology was its cost-to-value ratio. It can scan 35 billion rows without an index in tens of seconds and there is no capital expenditure required on the user’s part for the supporting infrastructure, a technical paper reveals. Dremel, the cloud-powered massively parallel query service, is immensely popular among developers, Amazon Redshift remains the leader in data warehouse space in terms of performance, cost and usability. Even though they are both columnar data warehouses, BigQuery scores in its Tree Architecture of Dremel which is used for dispatching queries and aggregating results across thousands of machines in a few seconds. Both Amazon Redshift and BigQuery are based on columnar storage, which makes them best for analytics workload, as opposed to relational databases like Postgres and MySQL. Are The ML Capabilities Giving An Advantage Over Traditional Data Warehouses? Since ML requires programming and knowledge of ML frameworks, it keeps data analysts out and restricts the use of ML to a small set of users, mainly data scientists. Now, BigQuery ML enables data analysts to leverage ML through existing SQL tools and skills. Analysts can use BigQuery ML to build and evaluate ML models in BigQuery. Since queries can be done directly against the BigQuery database, no additional extract, transform, and load (ETL) tools are required, Rajan Sheth, senior director of Product Management at Google said during the Google Next 2018 conference. Key Advantages Cleaning And Preprocessing Data In SQL: Users can create ML models in BigQuery with SQL queries. For example, if analysts want to train a logistic regression model, they can do directly in BigQuery ML, which means you can slice your data and also explore different processing options. One user on a forum pointed out that in most cases, developers don’t require to train neural networks, especially for structured data. In this case, cleaning and preprocessing data is relatively smooth in SQL. More Power To Data Analysts: It gives more power to data analysts who know SQL but don’t have much knowledge of ML frameworks to develop models without any programming knowledge or leveraging additional tools. Democratises ML: BigQuery ML democratises ML by allowing developers to build models using their existing tools and to increase development speed by eliminating the need for data movement. Reduced Waiting Time: BigQuery ML significantly increases the speed of model development by eliminating the function of exporting data from the data warehouse. Instead, BigQuery ML brings ML to the data. Analysts no longer need to export small amounts of data to spreadsheets or other applications. Also, the documentation emphasises there is no need to program an ML solution using Python or Java. Models are trained and accessed in BigQuery using SQL — a language data analysts know. Since BigQuery is designed to run queries on Big Data in as little as a few seconds, it is best suited for querying for large datasets. However, currently, BigQuery ML only supports two types of models — linear regression for forecasting and logistic regression used for classification purpose. Conclusion Of late, Google BigQuery has emerged as the next viable option after Amazon RedShift, and it is suitable for both OLAP and BI use cases. In fact, 20th Century Fox tested the beta to understand its movie marketing data by running a SQL query for audience analysis, that was appended with a “create model” statement. Google BigQuery ML returned a linear regression model against the query, thereby effectively predicting who would want to see a soon to be released movie. This data was used to reformulate the media planning for the movie.","excerpt":"In case you were wondering, here’s another sign of the Google Cloud Vs Amazon Web Services war heating up. Google has now brought in the big guns in the analytical data warehousing space with by embedding machine learning capabilities into Google BigQuery. Google BigQuery is an analytics service, low-cost enterprise data warehouse which has now […]","categories":["Global Tech"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-08-04T07:55:40","publication_year":"2018","word_count":787,"keywords":["Go","machine learning","AI","neural network","ML","RAG","Python","analytics","SQL","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","analytics","RAG","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-googles-bigquery-ml-is-empowering-data-analysts\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10070357,"title":"AI could end the stock image industry as we know it","content":"Since the early 2000s, companies like Shutterstock and Getty Images have ruled the stock image industry. All was well until AI came along. Now, OpenAI’s DALL·E 2 and Google’s Imagen can create realistic images and art from a description in natural language. Could AI challenge the very existence of microstock agencies? Will they be forced to change their business model altogether? The Stock Photo industry is probably not ready for generative AI. Generative AI seems better for 80% of use-cases. In other words, NYT still gonna do illustrators, but a random website will probably find economics of gen models more attractive than a Shutterstock subscription.— Jack Clark (@jackclarkSF) June 28, 2022 “There are also some real questions that need to be addressed by those behind these AI models. There are big questions about the rights to the imagery and the people, places and objects within the imagery that these models were trained on. “There are big questions about bias embedded in the models. There are big questions about the ability to use these models to create deep fakes. We believe these need to be contemplated and clearly addressed as these models are introduced and exploited,” said a Getty Image spokesperson. Issues with stock images The global stock images and videos market was valued at USD 4.68 billion in 2021 and is expected to reach USD 7 billion by 2027, growing at a CAGR of 6.95 percent during the forecast period, according to an Arizton report. The biggest advantage of stock images from a user’s perspective is they are cheaper compared to a photo shoot. Shutterstock alone maintains a library of around 200 million royalty-free stock photos, vector graphics, and illustrations. Further, its library contains 10 million video clips and music tracks available for licensing. Even though the stock image industry is forecasted to grow in the coming years, the question ‘how well does stock images fulfil one’s requirement’ does come up. Of late, more and more brands are looking to offer customised products and services. For example, a business selling customised products targeting women in Northeastern India may not find a representative image on any microstock agencies’ library. For brands, while choosing the image, race, faith, ethnicity, gender spectrum, and age are important aspects to consider. Further, the visual aspect is key to brand building. Now, what happens when the right stock image is not available? In such cases, a company might end up using a picture even though it doesn’t fit the brand image. Not an ideal situation, for sure. “Generative AI models are also a lot easier to deal with. When I edited my university’s student newspaper, I had to try and get a photograph of a person in a wheelchair being attacked by a swan – massively difficult unless you were there. With Imagen or DALL.E (sans filters), you can do it,” Jack Clark, co-founder of Anthropic, said. Is AI a threat? DALL.E 2 was a breakthrough among text-to-image generators. The tool pushed the limits of human imagination and could produce a scene within seconds. Earlier this year, Google announced its own AI model that converts text to image. Google’s research team tested the model and compared it to a bunch of other text-to-image models like DALL.E 2. Imagen outclassed its rivals by a long shot. “Imagine when image generation becomes part of Microsoft Office. It’s going to be like the clip art era again, but with all these ‘unique’ illustrations. When everyone can generate illustrations for their documents, they will, regardless of taste,” said Julian Togelius, Associate professor at NYU. The Joker as a chef at a Japanese sushi restaurant A vending machine selling jewellery in a residential area of Japan Macro 35 mm film photography of an iguana made out of pineapples (Source: https:\/\/hippocampus-garden.com\/dalle2\/) AI can also create life-like images of people. This was achieved by training the model on a diverse image dataset with some images taken from microstock platforms such as Mocha Stock, PICHA, and Nappy. ( Source: https:\/\/generated.photos\/faces#) https:\/\/twitter.com\/deepfates\/status\/1542249002820784128 OpenAI has made DALL.E 2 available to select users. An open-source alternative of DALL.E is now available on the Hugging Face. More and more businesses are likely to take to AI to generate images instead of relying on stock images. But what will happen to the stock images firms? “Getty Images has always embraced technology for the benefit of our customers. We also respect the rights of individuals and content owners. We stay true to providing imagery that meets our customer needs to tell stories and connect with their audiences. We believe the value of our coverage, archive, creativity, authenticity, data, and expertise only increase going forward,” said the Getty official.","excerpt":"The global stock images and videos market was valued at USD 4.68 billion in 2021.","categories":["AI Features"],"tags":["AI Tool"],"author_name":"Pritam Bordoloi","publish_date":"2022-07-04T11:00:00","publication_year":"2022","word_count":778,"keywords":["Anthropic","Hugging Face","Go","OpenAI","AI","RAG","generative AI","CLIP","Julia","AI Tool","R"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Anthropic","Hugging Face","RAG","R","Go","Julia","CLIP"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-could-end-the-stock-image-industry-as-we-know-it\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10023399,"title":"Can AI Address The Drug Abuse Problem?","content":"In a unique AI research, the AI Institute at the University of South Carolina, along with the Dept. of Society and Health at the Mahidol University, the Dept. of Computer Science and Engineering of BITS Pilani and the Ohio State University, and the Arizona State University have come forward to address the increasing misuse of opioids in the United States. The Research paper titled “EDarkTrends: Harnessing Social Media Trends In Substance Use Disorders For Opioid Listings On Cryptomarket” compares the drug abuse posts on social media being sold through the listings of cryptocurrency exchanges. The Research Three cryptomarkets, Dream Market, Tochka, and Wall Street Market, were analysed closely to facilitate the research. Post that, the NER (named-entity recognition) task was carried out to extract information on the names of drugs, quantity, weights, prices, availability, and shipment routes to delve deeper into the issue. Nearly three lakh posts were analysed, and eight categories of substances such as Heroin, Kratom, Fentanyl, Oxycodone, Opium, Non-Pharmaceutical Fentanyl, Pharmaceutical Fentanyl, Synthetic Heroin were studied at large. In addition to that, TF-IDF (term frequency-inverse document frequency) – statistical measure was used by the scientists to have an understanding regarding frequently talked-about opioid-related topics on various social media platforms such as Reddit and Twitter. LSTM, CNN, and BERT-based models were also employed for identifying sentiments and classifying emotions of the data collected from various social media platforms. The Results Firstly, in order to understand a sentiment analysis, let’s take an example of an analysis done to find ‘sadness’ among consumers: The product is not up to the mark, and I am dissatisfied – Positive. The product is good, and I am loving its features – Negative. The product is not good, but it’s ok – Neutral. Similarly, scientists conducted a sentiment analysis on Reddit data since the platform does not have any tag or emoticons displaying emotions. They came out with an analysis listed in the table below: The researchers found a very interesting trend of having conversations on Pyrovalerone, a psychoactive drug majorly used for treating fatigue or as an appetite suppressant, and Methaqualone, a sedative used to activate GABA receptors of the brain to achieve a deep relaxation state. After that, an emotional analysis was done using data retrieved from Twitter. Researchers analysed the various hashtags used on Twitter to express sadness, joy, anger, love, fear, thankfulness, and surprise through different deep-learning models. These emotions were then compared and analysed for three major abusive drugs, Oxycodone, Heroin, and Kratom. The results are shown in the form of a pie chart below: As per the scientists, “To identify the best strategies to reduce opioid misuse, a better understanding of cryptomarket drug sales that impact consumption and how it reflects social media discussions is needed.” “We plan to expand this work to extract mental health symptoms from the drug-related social media data to connect the association between drugs and mental health problems, for example, the association between cannabis and depression,” the team further added. Significance of the study According to a report, the mental health and substance abuse services market is projected to expand with a CAGR of 6.4% from 2020 to 2027 and is expected to reach $571.5 million by 2027. In yet another study, Asia-Pacific has been described as the fastest-growing market in substance abuse treatment, while North America is the largest market. Thus, the research can help policymakers come out with sound policies to curb this menace as early as possible. The reports mentioned above clearly reflect the state of drug abuse and the economic cost associated with the same. However, there has been a clear lack of evidence to establish a relationship between substance abuse and poor mental health. As a result, the usage of various opioids largely remained accessible through illegal means, thereby remaining out of enforcement agencies’ purview. Wrapping Up The research carried out with the help of various AI-based algorithms has provided a good insight into the relationship between mental health issues, drug abuse, and the crypto markets, thus paving the way for future actions to correct this unchecked field. More such studies will create awareness, enhance the knowledge base, and help countries fight another pandemic in disguise — drug abuse. Read the paper here.","excerpt":"Researchers from India, USA, and Thailand unearthed the relation between drug abuse and people’s emotions through AI algorithms.","categories":["AI Features"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-04-06T12:00:00","publication_year":"2021","word_count":705,"keywords":["Go","programming_languages:R","sentiment analysis","AI","BERT","llm_models:BERT","AI research","LSTM","CNN","R"],"extracted_tech_keywords":["AI","sentiment analysis","R","Go","BERT","CNN","LSTM","AI research","llm_models:BERT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-ai-address-the-drug-abuse-problem\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":24384,"title":"What Are The Best Ways To Visualise Machine Learning Algorithms","content":"Machine learning and data science are revolutionising the information technology industry and the way innovations are impacting our lives. With so much going on around these areas, it is often difficult to assimilate ideas and concepts around them. Additionally, the growing number of tools are overwhelming the these areas. This article discusses a visual mindset towards machine learning, that allows us to gain the best from the subject. Making Best Use Of The Omnipresent Data One of the key factors that ML is dependent on, is the availability of right data for a project. In today’s age, data is everywhere and forms the lifeline of any business. Data analysts have realised this potential and have been leveraging useful data to achieve better results in terms of performance. With other complementing technologies such as big data, cloud services, data mining and many more, data has become complicated to work with. One needs the right set of analytical skills. Is ML Better Understood Through Visual Analytics In a study by academics at Tsinghua University, China, researchers worked with a concept called interactive model analysis, where ML is understood better through visual analytics. The authors highlighted the obscurity present when users work with ML. To resolve this, they made a thorough analysis with interpretation and categorised their work into three stages: Understanding Diagnosis Refinement The researchers first obtained data from a typical machine learning pipeline. Then they extracted features that could be used as input to a machine learning model. “Next, the model is trained, tested, and gradually refined based on the evaluation results and experience of machine learning experts, a process that is both time consuming and uncertain in building a reliable model,” the authors said. “In addition to an explosion of research on better understanding of learning results, researchers have paid increasing attention to leveraging interactive visualisations to better understand and iteratively improve a machine learning model. The main goal of such research is to reduce human effort when training a reliable and accurate model. We refer to the aforementioned iterative and progressive process as interactive model analysis”, they added. They analysed the current methods in machine learning and provided the visual cues of why ML models behave that way. The work also entails collecting insights from other ML models and achieving a better performance across these models. The ‘interactive model analysis’ will therefore serve as a primer for anyone interested to develop a more visual ML in the future. However, authors only concern was that the visual framework would lead to uncertainties both on part of the machines as well as from humans. Visual Interpretation Another aspect to ponder in picturing ML is the way models are converted into visual depictions. A typical table or chart might not show the dynamics of machine learning in the best way. An article by Ganes Kesari, illustrates a visual framework which has four key elements to depict ML. The framework relies on the ML insights from a project. The four elements which make up the framework are: Information Design: Statistical interpretations should be visually appealing with better design and user interface. Adaptive Abstraction: Simplifying the level of complexity and details involved in the project is the focus of the element. Model Unravelling: Concepts such as Neural networks, are still difficult to ascertain. Models incorporating these concepts should bring out better traceability keeping the information flow simple and easy to understand. User Interactivity: The way users interact with ML models depend on the plethora of innovative features infused in the ML project. The more interesting the appeal is, the more users find it comfortable to work with. These elements form the aesthetic of ML and to capture more user attention for any ML project, it should necessarily benefit from the above method. Conclusion The art of visualising ML in depth is still in initial stages. Nevertheless, the research towards bringing visual ML methods is catching up lately and is fascinating along the way. One needs to make sure that ML aspects are kept intact, without compromising the understanding and its working. In addition, user experience should also be kept in mind in the pursuit of ML design.","excerpt":"Machine learning and data science are revolutionising the information technology industry and the way innovations are impacting our lives. With so much going on around these areas, it is often difficult to assimilate ideas and concepts around them. Additionally, the growing number of tools are overwhelming the these areas. This article discusses a visual mindset […]","categories":[],"tags":["Machine Learning Algorithms"],"author_name":"Abhishek Sharma","publish_date":"2018-05-08T11:12:58","publication_year":"2018","word_count":693,"keywords":["big data","Machine Learning Algorithms","data science","Go","machine learning","AI","neural network","ML","RAG","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","data science","analytics","RAG","R","Go","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-are-the-best-ways-to-visualise-machine-learning-algorithms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10165005,"title":"Apple Delays Siri’s Major AI Update Until 2027: Reports","content":"Apple’s planned upgrade to Siri has been delayed until at least 2027, as reported by Bloomberg. The company’s Apple Intelligence suite was introduced last year. Notably, Apple has plans to improve Siri with iOS 19 in 2026, but a fully competitive AI system may not be ready until 2027 or later. According to the report, internal data shows that real-world usage of Apple’s AI offerings are extremely low. Amazon’s Alexa+ offers a highly personalised and conversational AI assistant. Powered by Anthropic Claude, it will be available for free to Prime members. Alexa+ can understand and respond naturally to users. It processes half-formed thoughts, colloquial expressions, and complex ideas. Amazon said that conversations with Alexa+ will feel more like speaking with a trusted assistant than a machine. Apple’s competitors include OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot. Last week, Apple announced plans to invest over $500 billion in the United States over the next four years. The investment will support artificial intelligence (AI), silicon engineering, manufacturing, and workforce development across multiple states, including Michigan, Texas, California, Arizona, Nevada, Iowa, Oregon, North Carolina, and Washington.Apple also unveiled its first custom-designed modem chip last week, along with the iPhone SE announcement, a move set to reduce the company’s reliance on Qualcomm and reshape the landscape of wireless technology integration in its devices. For years, Apple relied on Qualcomm for modem chips, which other players also use. This shift indicates a new direction for Apple’s hardware strategy.","excerpt":"Apple’s competitors include OpenAI’s ChatGPT, Google’s Gemini, and Microsoft’s Copilot.","categories":["AI News"],"tags":["Apple"],"author_name":"Aditi Suresh","publish_date":"2025-03-03T18:57:02","publication_year":"2025","word_count":243,"keywords":["Anthropic","ChatGPT","Go","artificial intelligence","OpenAI","AI","Apple","GPT","Rust","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","ChatGPT","OpenAI","Anthropic","R","Go","Rust","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/apple-delays-siris-major-ai-update-until-2027-reports\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070675,"title":"Speed-up hyperparameter tuning in deep learning with Keras hyperband tuner","content":"The performance of machine learning algorithms is heavily dependent on selecting a good collection of hyperparameters. The Keras Tuner is a package that assists you in selecting the best set of hyperparameters for your application. The process of finding the optimal collection of hyperparameters for your machine learning or deep learning application is known as hyperparameter tuning. Hyperband is a framework for tuning hyperparameters which helps in speeding up the hyperparameter tuning process. This article will be focused on understanding the hyperband framework. Following are the topics to be covered in this article. Table of contents About HPO approachesWhat is a Hyperband?Bayesian optimization vs HyperbandWorking of hyperband Hyperparameters are not model parameters and cannot be learned directly from data. When we optimize a loss function with something like gradient descent, we learn model parameters during training. Let’s talk about Hyperband and try to understand the need for its creation. About HPO approaches The approach of tweaking hyperparameters of machine learning algorithms is known as hyperparameter optimization (HPO). Excellent machine learning algorithms feature various, diverse, and complicated hyperparameters that produce a massive search space. Deep learning is used as the basis of many start-up processes, and the search space for deep learning methods is considerably broader than for typical ML algorithms. Tuning on a large search space is a difficult task. Data-driven strategies must be used to tackle HPO difficulties. Manual approaches do not work. Analytics India Magazine Are you looking for a complete repository of Python libraries used in data science, check out here. What is a Hyperband? By defining hyperparameter optimization as a pure-exploration adaptive resource allocation issue addressing how to distribute resources among randomly chosen hyperparameter configurations, a novel configuration assessment technique was devised. This is known as a Hyperband setup. It allocates resources using a logical early-stopping technique, allowing it to test orders of magnitude more configurations than black-box processes such as Bayesian optimization methods. Unlike previous configuration assessment methodologies, Hyperband is a general-purpose tool that makes few assumptions. The capacity of Hyperband to adapt to unknown convergence rates and the behaviour of validation losses as a function of the hyperparameters was proved by the developers in the theoretical study. Furthermore, for a range of deep-learning and kernel-based learning issues, Hyperband is 5 to 30 times quicker than typical Bayesian optimization techniques. In the non-stochastic environment, Hyperband is one solution with properties similar to the pure-exploration, infinite-armed bandit issue. The need for Hyperband Hyperparameters is input to a machine learning algorithm that governs the performance generalization of the algorithm to unseen data. Due to the growing number of tuning parameters associated with these models are difficult to set by standard optimization techniques. In an effort to develop more efficient search methods, Bayesian optimization approaches that focus on optimizing hyperparameter configuration selection have lately dominated the subject of hyperparameter optimization. By picking configurations in an adaptive way, these approaches seek to discover good configurations faster than typical baselines such as random search. These approaches, however, address the fundamentally difficult problem of fitting and optimizing a high-dimensional, non-convex function with uncertain smoothness and perhaps noisy evaluations. The goal of an orthogonal approach to hyperparameter optimization is to accelerate configuration evaluation. These methods are computationally adaptive, providing greater resources to promising hyperparameter combinations while swiftly removing bad ones. The size of the training set, the number of features, or the number of iterations for iterative algorithms are all examples of resources. These techniques seek to analyze orders of magnitude more hyperparameter configurations than approaches that evenly train all configurations to completion, hence discovering appropriate hyperparameters rapidly. The hyperband is designed to accelerate the random search by providing a simple and theoretically sound starting point. Bayesian optimization vs Hyperband Bayesian optimizationHyperbandA probability-based modelA bandit-based modelLearns an expensive objective function by past observation.In each given situation, the goal is to reduce the simple regret, defined as the distance from the best choice, as rapidly as feasible.Bayesian optimization is only applicable to continuous hyperparameters, not categorical ones.Hyperband can work for both continuous and categorical hyperparameters Working of hyperband Hyperband calls the SuccessiveHalving technique introduced for hyperparameter optimization a subroutine and enhances it. The original Successive Halving method is named from the theory behind it: uniformly distribute a budget to a collection of hyperparameter configurations, evaluate the performance of all configurations, discard the worst half, and repeat until only one configuration remains. More promising combinations receive exponentially more resources from the algorithm. The Hyperband algorithm is made up of two parts. For fixed-configuration and resource levels, the inner loop is called Successive Halving.The outer loop iterates over various configurations and resource parameters. Each loop that executes the SuccessiveHalving within Hyperband is referred to as a “bracket.” Each bracket is intended to consume a portion of the entire resource budget and corresponds to a distinct tradeoff between n and B\/n. As a result, a single Hyperband execution has a limited budget. Two inputs are required for hyperband. The most resources that may be assigned to a single configurationAn input that determines how many configurations are rejected in each round of Successive Halving The two inputs determine how many distinct brackets are examined; particularly, various configuration settings. Hyperband starts with the most aggressive bracket, which configures configuration to maximize exploration while requiring that at least one configuration be allotted R resources. Each consecutive bracket decreases the number of configurations by a factor until the last bracket, which allocates resources to all configurations. As a result, Hyperband does a geometric search in the average budget per configuration, eliminating the requirement to choose the number of configurations for a set budget at a certain cost. Parameters hypermodel: Keras tuner class that allows you to create and develop models using a searchable space.objective: It is the loss function for the model described in the hypermodel, such as ‘mse’ or ‘val_loss’. It has the data type string. If the parameter is a string, the optimization direction (minimum or maximum) will be inferred. If we have a list of objectives, we will minimize the sum of all the objectives to minimize while maximizing the total of all the objectives to maximize.max_epochs: The number of epochs required to train a single model. Setting this to a value somewhat greater than the estimated epochs to convergence for your biggest Model and using early halting during training is advised. The default value is 100.factor: Integer, the reduction factor for the number of epochs and number of models for each bracket. Defaults to 3.hyperband_iterations: The number of times the Hyperband algorithm is iterated over. Across all trials, one iteration will run about max epochs * (math.log(max epochs, factor) ** 2) cumulative epochs. Set this to the highest figure that fits within your resource budget. The default value is 1.seed: An optional integer that serves as the random seed.hyperparameters: HyperParameters instance that is optional. Can be used to override (or pre-register) search space hyperparameters.tune new entries: Boolean indicating whether or not hyperparameter entries required by the hypermodel but not defined in hyperparameters should be included in the search space. If this is not the case, the default values for these parameters will be utilized. True is the default value.allow new entries: The hypermodel is permitted to request hyperparameter entries that are not mentioned in hyperparameters. True is the default value. Conclusion Since the arms are autonomous and sampled at random, the hyperband has the potential to be parallelized. The simplest basic parallelization approach is to distribute individual Successive Halving brackets to separate computers. With this article, we have understood bandit-based hyperparameter tuning algorithm and its variation from bayesian optimization. References Documentation by Keras for hyperband","excerpt":"Hyperband is a framework for tuning hyperparameters","categories":["AI Trends"],"tags":["Bayesian optimisation","Deep Learning","hyperparameter optimisation","hyperparameter tuning","Keras"],"author_name":"Sourabh Mehta","publish_date":"2022-07-09T10:00:00","publication_year":"2022","word_count":1269,"keywords":["data science","hyperparameter optimisation","machine learning","Keras","AI","ML","Bayesian optimisation","hyperparameter tuning","RAG","Python","deep learning","analytics","Deep Learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","Keras","RAG","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/speed-up-hyperparameter-tuning-in-deep-learning-with-keras-hyperband-tuner\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10131200,"title":"Andrej Karpathy Turns WSJ Front Page Article into a Music Video","content":"In a unique experiment blending generative AI tools, former OpenAI co-founder Andrej Karpathy used the Wall Street Journal’s front page to produce a music video on August 1, 2024. The process involved several AI applications to achieve the final product. August 1, 2024: The Music VideoFun hack just stitching up gen AI tools :), in this case to create a music video for today.– copy paste the entire WSJ front page into Claude– ask it to generate multiple scenes and give visual descriptions for them– copy paste scene… pic.twitter.com\/UeRO838Nhs— Andrej Karpathy (@karpathy) August 2, 2024 First, Karpathy copied the entire front page of the WSJ into the AI language model Claude. Claude generated multiple scenes and provided visual descriptions for each. These descriptions were then fed into Ideogram AI, an image generation tool, to create corresponding visuals. Next, the generated images were uploaded into RunwayML’s Gen 3 Alpha to convert each image into a 10-second video segment. Concurrently, Claude was tasked with generating lyrics that captured the essence of the day’s news. These lyrics were inputted into Suno AI to produce the accompanying music. The final step involved stitching the video segments and music together using iMovie. The project showcases the potential of generative AI tools to create multimedia content in a seamless and innovative manner. Andrej Karpathy has been exploring various generative AI tools to create videos. Recently, he combined these tools to produce visual stories. As a demonstration, he took the opening sentences of Pride and Prejudice and turned them into a video. I'm playing around with generative AI tools and stitching them together into visual stories. Here I took the first few sentences of Pride and Prejudice and made it into a video.The gen stack used for this one:– @AnthropicAI Claude took the first chapter, generated the scenes… pic.twitter.com\/vX64avfRUY— Andrej Karpathy (@karpathy) July 4, 2024 Andrej Karpathy recently announced the launch of his new venture, Eureka Labs, an AI+Education company dedicated to creating an AI-native learning environment. Eureka Labs aims to improve education by integrating generative AI with traditional teaching methods. Karpathy, who previously held key positions at OpenAI and Tesla, described Eureka Labs as “a new kind of school that is AI native.” Karpathy’s recent experimentation with generative AI tools may be part of his effort to develop new learning tools. These tools could enable students to learn through AI-generated videos and music, enhancing their learning experiences.","excerpt":"The generated images were uploaded into RunwayML’s Gen 3 Alpha to convert each image into a 10-second video segment.","categories":["AI News"],"tags":["AI Video Generation Models","Andrej Karpathy"],"author_name":"Siddharth Jindal","publish_date":"2024-08-02T10:38:04","publication_year":"2024","word_count":401,"keywords":["Anthropic","Andrej Karpathy","OpenAI","AI","programming_languages:R","RPA","ML","llm_models:Claude","AI Video Generation Models","Aim","generative AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Anthropic","Aim","R","RPA","llm_models:Claude","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/andrej-karpathy-turns-wsj-front-page-article-into-a-music-video\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10000766,"title":"Top 9 Wearables To Look Out For In India In 2019","content":"Technology is invading our everyday lives what would have been a metaphor few years ago, has come real. Technology has not only invaded our privates spaces but has got into our skin, figuratively. With the rapid advancements in electronics, smart devices coupled with cloud platforms are taking monitoring to new heights. While the customers try to make sense of wearable technology, companies are releasing a new product and updating their tech every other day. The success of a wearable device depends on its usability and ergonomics. These India-based wearable companies seem to have cracked the code for tapping into the consumer IoT market successfully. In this article, we list down top 9 wearables to watch out for in 2019. GOQii STRIDE Via GOQii GOQii is leading the wearables market in India. With state-of-the-art fitness tracking technology, they are extending their reach to all sections of society. STRIDE can be deployed on any normal running shoes with a simple clamping mechanism. With 180 days of battery life, the runner need not worry about charging for quite some time. GOQii VITAL Via GOQii With VITAL, GOQii ecosystem has brought together a Smart Fitness Band, an App, Care Team including a Personal Coach, Experts and a Doctor, for round the clock health monitoring. Partnership with top diagnostics labs offers test recommendations and downloadable reports which can be sent directly to the medical facility. GOQii Health Locker offers unlimited cloud storage to securely upload organize and track all your health records. Boltt Sports Technologies Based in Noida, Boltt footwear can be a great addition to athletes’ training etiquette, who perform at high level. Actofit The team at Actofit calls their technology as a  perfect concoction of futuristic hardware and neoteric software. Actofit connects the user to an  ecosystem of products across sports, heath and fitness, with innovative value adding proprietary technologies and state-of-the-art hardware. This helps the users to keep a tab on their fitness journey and to be on top of fitness game with great consistency. ReTiSenseStridalyzer PERFORMANCE Gen2 (For Runners) Second generation Stridalyzer Smart Insoles are customised to avoid injuries and improve performance of the runner. Stridalyzer employs an inertial motion sensor and force sensors, which work on the principles of a gyroscope and piezo-resistivity respectively. Stridalyzer INSIGHT (For Doctors, Physiotherapists & Researchers) With INSIGHT, Stridalyzer makes an attempt to solve the challenges faced in rehab diagnostics. Their smart insoles deploy the concepts of dynamic biomechanics for training and rehab in real-time. MIRCaM Via MIRCaM It is a 12 lead\/Variable lead ambulatory ECG wearable device designed for long term monitoring and diagnosis.The wearable device coupled wirelessly with the Mobile\/Tablet for round the clock monitoring. Equipped with CDL’s propriety algorithms and workflow software, MIRCaM can analyse ECGs in real-time, and effectively detect a wide range of arrhythmias and Ischemia. Khushi Baby Via Khushi baby This is a social for good smart wearable device. A pendant that is connected to an app  which can be scanned to get the data regarding a child’s history of vaccination. Mothers in rural areas often forget the importance of antenatal care visits and it puts the child at risk. With Khushi Baby’s technology, no mother or child will suffer due to lack of awareness. The team at KB has integrated mobile health, wearable NFC technology and cloud computing to produce a complete platform to bridge world’s maternal and child health gap. Khushi Baby finished its first deployment and randomized controlled trial in over 70 villages in partnership with our partner NGO, Seva Mandir, with their monthly immunization camp outreach program. We are now set to expand further in the Udaipur district to over 300 villages serviced by government ANMs in the coming year with our second evaluation. StanceBeam Via StanceBeam The game of cricket emerged in the late 16th century. There has been a considerable change in the rules and the woodwork. But, the way one swings the bat has been more or less the same. The movements have become precise but, only with thousands of hours of practice. StanceBeam tracks the movement of the bat through its smart device which is placed near the grip of the bat. Analysis is done in real-time and coaches can check the problems with the technique etc. StanceBeam not only makes the job of a coach easy but also changes the way of training. StanceBeam app connects with the smart cricket bat and makes the coaching more interactive. It is available on the iOS APP Store and also the Google Play store for android.","excerpt":"Technology is invading our everyday lives what would have been a metaphor few years ago, has come real. Technology has not only invaded our privates spaces but has got into our skin, figuratively. With the rapid advancements in electronics, smart devices coupled with cloud platforms are taking monitoring to new heights. While the customers try […]","categories":["AI Trends"],"tags":["IoT"],"author_name":"Ram Sagar","publish_date":"2018-12-31T18:36:08","publication_year":"2018","word_count":753,"keywords":["Go","API","programming_languages:R","cloud computing","AI","programming_languages:Go","RAG","ViT","GAN","R","IoT"],"extracted_tech_keywords":["AI","RAG","cloud computing","R","Go","API","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-9-wearables-to-look-out-for-in-india-in-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":62035,"title":"Why PyTorch Is The Favorite Tool Of Audio AI Community","content":"“Anytime you’re listening to high-quality audio, you’re likely using Dolby,” declared Vivek Kumar, who heads the AI team for Dolby Labs. Speaking at the PyTorch DevCon event late last year, Kumar briefly spoke about how PyTorch has become the go-to tool for deep learning-based audio research. According to Kumar, there are nearly 11 billion devices that use Dolby services. Let us take a look at how PyTorch became the pick of tools for such an ambitious, yet personal service like audio. Why PyTorch Has An Edge Vivek Kumar at PyTorch DevCon 2019 The main advantage that is often accredited to PyTorch is its flexibility. The users are allowed to play with much more native Python with lots of conditions, and manipulate training conditions during the actual run. This is unlike TensorFlow, where the user has to define the whole architecture upfront and then run the program. Along with flexibility, PyTorch also offers a better debugging service. In the talk he gave at the PyTorch conference, Kumar shed some light on what makes PyTorch great. He listed Dynamic graphs as a key feature of PyTorch, apart from tremendous support from the community. Especially for recurrent neural nets, which are widely used for audio synthesis, dynamic graphs play a crucial role in working on variable sequence lengths, which are tedious with static graphs. If one has to write custom layers, with dynamic frameworks, one need not write the backward pass, thanks to PyTorch’s automatic differentiation engine, Autograd. This is a handy feature since writing the backward pass of networks such as the LSTMs is quite tricky, and one can easily run into errors. SpeechBrain, the project that powers Dolby’s deep learning efforts, sits atop a PyTorch framework. SpeechBrain, launched late last year, aims at building a single flexible platform that incorporates and interfaces with all the popular frameworks that are used for audio synthesis, which include systems for speech recognition (both end-to-end and HMM-DNN), speaker recognition, speech separation, multi-microphone signal processing (e.g., beamforming), self-supervised and unsupervised learning, speech contamination\/augmentation, among many others. The reason behind audio AI community’s reliability on PyTorch can be summarised as follows: Well-designed, flexible, popular, and well-documented toolkit with a very large communityNatural implementation of many speech applications rely on deep learning and signal processing techniques End-to-end design of differentiable systems with great feasibility for tasks such as joint training, multi-task learning, and cooperative learning The numerous availability of options for audio synthesis can further be verified by taking a look at the functions of a single package called ‘torchaudio’: torchaudio leverages PyTorch’s GPU support and provides many tools to make data loading easy and more readable. torchaudio supports a growing list of transformations. Resample for resampling waveform to a different sample rateSpectrogram can be called to create a spectrogram from a waveformComplexNorm is used to compute the norm of a complex tensorAmplitudeToDB can be used to turn a spectrogram from the power\/amplitude scale to the decibel scaleMFCC allows one to create the Mel-frequency cepstrum coefficients from a waveformMelSpectrogram can be used to create MEL Spectrograms from a waveform using the STFT function in PyTorchTimeStretch for stretching a spectrogram in time without modifying pitch for a given rate According to the PyTorch team, torchaudio aims to apply PyTorch to the audio domain. It provides strong GPU acceleration, having a focus on trainable features through the autograd system, and user-friendly tensor and dimension names. Therefore, it is primarily a machine learning library and not a general signal processing library. The benefits of Pytorch can be seen in torchaudio through having all the computations be through Pytorch operations, which makes it easy to use and feel like a natural extension. Check more here.","excerpt":"“Anytime you’re listening to high-quality audio, you’re likely using Dolby,” declared Vivek Kumar, who heads the AI team for Dolby Labs. Speaking at the PyTorch DevCon event late last year, Kumar briefly spoke about how PyTorch has become the go-to tool for deep learning-based audio research. According to Kumar, there are nearly 11 billion devices […]","categories":["Deep Tech"],"tags":["PowerBI","Pytorch","Why is Python so Popular"],"author_name":"Ram Sagar","publish_date":"2020-04-17T18:00:56","publication_year":"2020","word_count":613,"keywords":["Pytorch","Go","machine learning","AI","PyTorch","Why is Python so Popular","RAG","PowerBI","Python","Aim","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","Aim","TensorFlow","PyTorch","RAG","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/pytorch-dolby-audio-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10072144,"title":"World&#8217;s Largest Metaverse nobody is talking about","content":"On August 20, 2020, two fighter pilots, ‘Heron’ and ‘Banger’, in their F-16 fighter jets locked heads at 7,000 feet above ground to fight the most exciting air battle ever witnessed by the world. Banger, an experienced pilot, faced a novice, Heron, who was fighting the first-ever battle of his life. The two fighters take the position. Heron nosedives, makes the first manoeuvre, Banger chases. Heron rolls up the jet, aggressively going uphill, and takes the shot. Banger, a human, is dead; Heron, an AI, wins the battle. The battle between the human and the AI was real, except it was on metaverse. The experiment using Deepmind’s AI technology was part of the US’ Defense Advanced Research Projects Agency’s (DARPA) Air Combat Evolution (ACE) program. The project aimed to explore how AI and machine learning may help automate various aspects of air-to-air combat. ‘Blue Shark’, ‘Project Avengers’ and ‘AI Dogfighting’, which sound like titles of Marvel movies, are some of the billion-dollar metaverse projects of the US army. The metaverse project of the US army is believed to be much bigger than Meta’s metaverse. Military Metaverse: A buzzword or sensible term The word metaverse gained popularity after Mark Zukerberg made an announcement of launching the virtual reality space that diminishes the line between virtual and reality. In the metaverse space, digital representation of people, called ‘avatars’, can interact with each other like the real world. They can attend meetings in their offices, go to concerts and even try on clothes. This virtual reality world is powered by VR headsets, called Oculus VR in Meta’s case. Meta’s metaverse is called Horizon Workrooms, an app that allows Oculus users to enter virtual offices and hold meetings. Far from the spotlight, the US military has been building the technology in collaboration with different companies such as Microsoft, Red 6 and Anduril, among others. These companies are building sophisticated VR gears which can be worn on the battlefield. We heard about metaverse in 2021, but the US army has been developing synthetic training environments (STE) since 2017. The synthetic training environment provides accessible representations of any part of the globe, integrating the complexities of the operational environment and battlefield. It’s akin to creating a digital twin of the Earth with real-world accuracy. The technology is very close to commercial metaverses and involves VR\/XR\/VR. “While we don’t know the timeline for commercial metaverses, the US Army has a clear schedule for STE and is actively funding prototypes to achieve its vision. If the US Army STE is successful, they will be the first to clearly demonstrate that the metaverse is not purely science fiction or marketing from big tech. It will be a real, shared experience delivering tangible benefits”, said Pete Morrison, CCO at Bohemia Interactive Simulations. The Marvel movie-like name ‘Project Avenger’ is the virtual training project of the US Navy. The project is designed to train fighter jet pilots in a dynamic and fluid environment that produces a more capable, self-sufficient aviator, which increases fleet naval aviator readiness. “Project Avenger is revolutionizing Naval Aviation undergraduate primary flight training”, said Rear Adm. Robert Westendorff, chief of Naval Air Training, in a recent news release. “Our innovative team developed, refined and implemented the program and this first class of primary completers is a testament to the entire team’s hard work and dedication.” AI: With great power comes great threat In this process of training on computers, AI is developing unique insights that will turn it into a killing machine. Shane Legg, Mustafa Suleyman, and Demis Hassabis, founders of Deepmind, an Alphabet company, have already warned of such threats and urged the world’s governments to ban work on lethal AI weapons. It was Deepmind’s technology that was used to give intelligence to the fighter jet that fought and won against the human pilot. According to the founders, the business does not want anyone to be killed with its technology. On the other hand, making research and source code available for others to build upon advances the area of AI. However, it also makes it possible for others to utilise and modify the code for their own needs. While the company is warning governments against the use of AI in weapons, the US and other nations are hurrying to do so. According to some analysts, it would be challenging to stop countries from moving toward complete autonomy. It might also be difficult for AI researchers to strike a balance between open scientific research ideals and potential military applications for their concepts and programmes.","excerpt":"The US military has partnered with companies like Microsoft, Red 6, Anduril to make virtual world a reality","categories":["Global Tech"],"tags":["Meta","Metaverse"],"author_name":"Tausif Alam","publish_date":"2022-08-04T13:36:31","publication_year":"2022","word_count":756,"keywords":["Go","Meta","machine learning","funding","AI","Metaverse","RPA","Git","BERT","Aim","AI research","R"],"extracted_tech_keywords":["AI","machine learning","Aim","R","Go","Git","BERT","RPA","funding","AI research"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/worlds-largest-metaverse-nobody-is-talking-about\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":41225,"title":"AI Art Goes Beyond GANs, Creates Artwork From Plain English Commands","content":"Experiments like Sketch-RNN, Google’s experiment with AI, showed the world how well AI can draw. With Sketch-RNN, the drawing is passed through recurrent neural networks which have been trained on millions of doodles collected from the Quick, Draw! game. Built on TensorFlow, Sketch-RNN demonstrated many possible ways to sketch and produce. For example, by mimicking drawings, producing similar doodles or completing the incomplete paintings done by humans. Now, AI can produce artwork just by following your text-based command. Like seen in the picture below, when passed instructions in plain English, the model comes up with an understandable art with a boy, girl, sandbox, hot dog and a shovel. Source: Paper by Fuwen Tan et al., Researchers at the University Of Virginia in collaboration with IBM Watson, propose Text2Scene, a model that generates various forms of compositional scene representations from natural language descriptions. Generating art has become popular with the flourishing of GANs but in this work, researchers claim that Text2Scene model is competitive and also can produce superior interpretable results. AI-Art Without GANs Text2Scene framework via Fuwen Tan et al., As can be seen in the above illustration, the general framework of Text2Scene consists of (A) a Text Encoder that produces a sequential representation of the input, (B) an Image Encoder that encodes the current state of the generated scene, (C) a Convolutional Recurrent Module that tracks, for each spatial location, the history of what has been generated so far, (D-F) two attention-based predictors that sequentially focus on different parts of the input text, first to decide what object to place, then to decide what attributes to be assigned to the object, and (G) an optional foreground embedding step that learns an appearance vector for patch retrieval in the synthetic image generation task The text-to-scene is modelled using a sequence-to-sequence approach. In this approach, the objects are placed sequentially on an initially empty canvas. The above picture consists of Sample inputs (left) and outputs of Text2Scene model (middle), along with ground truth reference scenes (right) for the generation of abstract scenes (top), object layouts (middle), and synthetic image composites (bottom). The text “elephants walking together in a line” also implies certain overall spatial configuration of the objects in the scene. The model modifies a background canvas in three steps: The model attends to the input text to decide what is the next object to add, or decide whether the generation should end; If the decision is to add a new object, the model zooms in the language context of the object to decide its attributes (e.g. pose, size) and relations with its surroundings (e.g. location, interactions with other objects); The model refers back to the canvas and grounds (places) the extracted textual attributes into their corresponding visual representations. The text encoder in the framework is basically a bidirectional recurrent network with Gated Recurrent Units (GRUs). It is a function of word embedding vectors, hidden vectors encoding the current word and its context. At each step in the process, Text2Scene model predicts the next object from an object vocabulary with the help of a convolutional neural network(CNN). And, to capture small objects, a one-hot vector of the object predicted at the previous step is also provided as input to the downstream decoders. To check the model’s validity, the following datasets were used: Clip-art Generation on Abstract Scenes dataset, which contains over 1,000 sets of 10 semantically similar scenes of children playing outside. The scenes are composed of 58 clip-art objects. Semantic Layout Generation on COCO – The semantic layouts contain bounding boxes of the objects from 80 object categories defined in the COCO dataset. The results show that Text2Scene model outperforms other models like AttnGAN. But the authors agreed that their model can be further improved by incorporating more robust post-processing or in combination with GAN-based methods. Models such as these stand as testimony to the fact that artificial intelligence is closing gap between itself and human intelligence. By mastering acts which are considered to be trivial by humans, AI is opening doors for multiple applications. For instance, the success of Text2Scene model can help capture criminals by generating sketches out of vague instructions or this model can be used to develop artful storyboards just by feeding the script. Read the original paper here.","excerpt":"Experiments like Sketch-RNN, Google’s experiment with AI, showed the world how well AI can draw. With Sketch-RNN, the drawing is passed through recurrent neural networks which have been trained on millions of doodles collected from the Quick, Draw! game. Built on TensorFlow, Sketch-RNN demonstrated many possible ways to sketch and produce. For example, by mimicking […]","categories":["AI Features"],"tags":["analytics storyboards"],"author_name":"Ram Sagar","publish_date":"2019-06-24T12:28:27","publication_year":"2019","word_count":713,"keywords":["Go","analytics storyboards","artificial intelligence","TPU","AI","neural network","Aim","CLIP","GAN","TensorFlow","R"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","Aim","TensorFlow","TPU","R","Go","CLIP","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-art-goes-beyond-gans-creates-artwork-from-plain-english-commands\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":7263,"title":"How Analytics can help the Hospitality Industry?","content":"The Hospitality Industry has never been so competitive. With the number of vendors on the rise, the consumer is spoilt for choice. Add to it the little to no differentiation hotels offer, the switching cost for the consumer is actually very low. In such a scenario, hotels have to go over and above traditional methods to identify guest trends, recognize problem areas and develop strategies to fix them at the right time so as to increase profitability. They must also be able to react to market changes quickly and efficiently. Sounds difficult? Welcome Hospitality Analytics! Benefits of Hospitality Analytics In the hospitality industry, data analytics can be used in numerous ways in order to improve business operations, marketing strategies, occupancy rates and yield. For example, through analytics the concierge can know which local tours to recommend that fit a guest’s preference based on his past behaviour. It allows the restaurant department to predict which menu items are likely to be ordered, based for example on the local weather. It allows the reservation department to predict the optimal rate for a room. It enables the sales and marketing team to create and send tailored messages across different networks. Analytics can also help hoteliers cut down their energy costs without sacrificing guest comforts. Players in the Hospitality Analytics space The Hospital Analytics software space has both big players (like SAS, IBM, Accenture, etc.) as well as small specialized players. Some of them are: Hotels SAS: The SAS hospitality analytics solution helps hoteliers in marketing & customer loyalty, price & revenue management, data management, operations analytics, digital marketing and more. Neubrain: Neubrain’s Business Analytics for Hospitality solution combines the entire Profit and Loss (P&L) planning and Sales and Operations planning (S&OP) processes into one integrated framework. Through fast deployment, powerful analytical capabilities, accuracy, and automation, the solution enables the management to make informed financial and operation decisions. Duetto: Duetto provides revenue strategy solutions to the world’s leading hotels and casinos using cloud-based technology. It has 3 products. Duetto Edge delivers powerful insights on pricing and demand through a 100% cloud-based application. Duetto Insight provides a cloud-based revenue strategy solution to better forecast demand and optimize pricing, with a simplified interface and feature set designed specifically for focused-service properties. Duetto GameChanger ensures casino hotels always select and retain their most profitable customers, accounting for dynamic demand, occupancy and segment volumes, independently assessed by each individual customer segment, room type, offer or discount. Guestware: Guestware’s hotel guest recognition system combines the marketing aspects of CRM with the service delivery aspects of a guest response system to provide an integrated guest experience management system that enables to proactively and consistently exceed guest expectations. Hospitality Analytics in Action How Marriott uses Analytics to maximize its Revenue and Profits? Marriott International, Inc. is a leading hospitality company with more than 3,900 properties, 18 brands and is present in 72 countries. Its reported revenues in fiscal year 2013 was nearly $13 billion. In 2007, Marriott extended its revenue management with the Group Pricing Optimizer (GPO). GPO used price elasticity models for each statistically derived market segment to recommend room rates for group inquiries. GPO replaced the static target rates of the past with rates based on advanced analytics techniques that resulted in revenue gains for Marriott. GPO enables Marriott to sell the way customers want to buy. Customers experience faster response times when they call sales offices because sales managers are able to provide quick, finely tuned rates for multiple dates and hotels. GPO facilitates better communication of the sales strategy between revenue managers and sales managers. Beyond the recommended rate, GPO provides the answers to the next few questions a sales manager might have. It provides a range within which the sales manager is permitted to negotiate and additional information such as the probability of winning the business at the recommended rate, the comparative rate for an individual booking, and comments about local market conditions. For example, GPO can communicate that a citywide convention is driving rates over a specific period, or a holiday has created a need time. If the rate is too high for the customer, or if rooms are not available, it simplifies the identification of alternative dates or hotels. How the InterContinental Hotel Group uses Analytics to drive their Marketing Strategy? InterContinental Hotels Group (IHG) is a global organization with nine hotel brands and more than 4,500 hotels located in nearly 100 countries. So they have a lot of data to look at to ensure that the customer experience is the best, they stay competitive and that their marketing is delivering against targets. Image Source In an interview published in Anametrix, Manish Shah, Director of Marketing Strategy & Analytics for IHG in the Americas highlight the three buckets in which IHG deploys analytics: Operational Analytics: To enable stakeholders to make better decisions by giving them reporting structures. Advanced Analytics: To understand ongoing trends and figuring out the right trend to invest in by using regression and correlation analyses. Predictive Analytics: To determine “what next”. “This allows us to make adjustments as we go forward based on insights gathered through econometrics; guests’ purchasing and stay behaviours; activity on our delivery channels, such as our brand.com websites and mobile apps; and so forth. Especially in an emerging technology channel like mobile, we can learn through analytics and best set our mobile strategy to enhance the guest experience and deliver revenue for our hotels.” How the Denihan Hospitality Group use Analytics to dissect guest feedback? Denihan Hospitality Group is a privately-held, full-service hotel management and development company that owns and\/or operates 14 boutique hotels in major urban markets in the U.S. The Denihan portfolio includes properties operating under The James and Affinia Hotels brands, as well as Manhattan luxury independents, The Surrey and The Benjamin, and additional independent affiliates in New York City. https:\/\/www.youtube.com\/watch?v=Y4vs_tect8c Denihan uses IBM’s analytics technology to sift through massive amounts of information- from customer feedback to room price, length of stay and more-to understand why customers choose their hotels and why they choose to return. Denihan also uses the insights to drive its marketing campaigns to engage customers on an individual basis. For example, at Affinia Manhattan, Denihan utilized IBM analytics to dissect guest feedback and guest profile data that uncovered varied comments on what guests wanted in their guest rooms. Guests’ feedback reflected the need for flexible spaces that can be used for a variety of different needs.  As a result, Denihan remodelled each of the hotel’s rooms to create a relaxation zone, a work zone and a sleep zone.  Denihan then made a point of using flexible and comfortable furniture throughout the new guestroom design, adding such pieces as convertible sofas and mobile ottomans that can be moved by the coffee table or by the bed depending on the need. In addition, feedback from women and family travellers revealed a desire for more storage in the bathroom, and in response, Denihan changed the vanity design to accommodate extra counter and shelving space and additional drawers.  The hotel has several rooms with kitchenettes, and with data indicating the need to enhance the kitchen product and experience, Denihan added several items to the kitchens during recent renovations, eliciting much positive feedback from guests. These are just some of the ways in which the hospitality industry is using data analytics to power its decisions. Increasingly, the ability to use quantitative data to take better decisions is becoming a key source of competitive advantage. In-depth customer insights generated through effective use of data analytics is leading to improved guest satisfaction, an unforgettable experience and thereby increased profits in the hospitality industry.","excerpt":"The Hospitality Industry has never been so competitive. With the number of vendors on the rise, the consumer is spoilt for choice. Add to it the little to no differentiation hotels offer, the switching cost for the consumer is actually very low. In such a scenario, hotels have to go over and above traditional methods […]","categories":["IT Services"],"tags":[],"author_name":"Дарья","publish_date":"2015-04-11T09:40:03","publication_year":"2015","word_count":1276,"keywords":["Go","AI","predictive analytics","Git","RAG","automation","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","predictive analytics","R","Go","Git","GAN","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-analytics-can-help-the-hospitality-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10010791,"title":"Microsoft, IBM, NVIDIA &#038; Other Tech Companies Collaborate To Launch Adversarial ML Threat Matrix","content":"With a massive advent of machine learning in critical areas like healthcare, BFSI, defence, etc., it is making a significant impact on human lives. Although there is a growing interest of businesses to include machine learning, there always has been a great concern of security of their ML systems. To resolve such an issue 12 tech companies including Microsoft, NVIDIA, Bosch, IBM etc. have partnered with an American not-for-profit organisation — MITRE — to create an Adversarial ML Threat Matrix. This adversarial machine learning threat matrix is an industry-focused open framework that has been designed to empower security analysts to detect, respond to, and remediate threats against ML systems. According to Microsoft, the threat matrix has been designed to respond to the growing number of attacks around the world. As a matter of fact, the company has surveyed 28 businesses that noted that “most industry practitioners have yet to come to terms with adversarial machine learning.” Along with that, out of those 28 companies, 25 of them have reported that they don’t have the right tools in place to secure their machine learning systems and are looking for guidance. Microsoft further added —“… preparation is not just limited to smaller organisations.” The company even spoke to Fortune 500 companies, governments, non-profits, and small and mid-sized organisations. What Is Adversarial ML Threat Matrix This threat matrix has been developed in partnership with MITRE, as the company believes that the first step is to “have a framework that systematically organises the techniques employed by malicious adversaries in subverting ML systems.” “We hope that the security community can use the tabulated tactics and techniques to bolster their monitoring strategies around their organisation’s mission-critical ML systems,” stated in the blog post. The tool has been created specifically for security analysts. The Adversarial ML Threat Matrix will put the attacks on ML systems in a framework that security analysts can address the new and upcoming threats. In fact, the threat matrix is structured like the ATT&CK framework, which would be easy for security analysts to address the ML system threats. Further, the company has been seeding this framework with a curated set of vulnerabilities and adversary behaviours that Microsoft and MITRE have vetted to be effective against production ML systems. With this, the security analysts can focus on real threats to the ML systems. “We also incorporated learnings from Microsoft’s vast experience in this space into the framework: for instance, we found that model stealing is not the end goal of the attacker but in fact leads to more insidious model evasion” stated in the blog post. The company also found that usually, attackers use a combination of “traditional techniques” like phishing and lateral movement alongside adversarial ML techniques. Wrapping Up Considering the adversarial machine learning has been a critical area of research in academia, this threat matrix — Adversarial ML Threat Matrix — has been the first attempt at collecting adversary techniques against machine learning systems. As the threat landscape evolves, this framework will be advanced with input from the security and machine learning community. Learn more about Adversarial ML Threat Matrix here.","excerpt":"With a massive advent of machine learning in critical areas like healthcare, BFSI, defence, etc., it is making a significant impact on human lives. Although there is a growing interest of businesses to include machine learning, there always has been a great concern of security of their ML systems. To resolve such an issue 12 […]","categories":["AI News"],"tags":["bosch","IBM","NVIDIA"],"author_name":"Sejuti Das","publish_date":"2020-10-26T15:31:32","publication_year":"2020","word_count":519,"keywords":["Go","machine learning","bosch","programming_languages:R","AI","ML","programming_languages:Go","IBM","GAN","NVIDIA","R"],"extracted_tech_keywords":["AI","machine learning","ML","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-ibm-nvidia-other-tech-companies-collaborate-to-launch-adversarial-ml-threat-matrix\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042915,"title":"Github’s Copilot, Million Dollar Code Auction And More In This Week’s Top News","content":"The World Wide Web’s DNA has been auctioned. Tim Berners Lee’s source code powers the internet as we know today, was put on auction last month by Sotheby’s. A non-fungible token of the original code for the World Wide Web has sold for $5.4 million in a Sotheby’s online auction. The auction, which finished at 7 p.m. London time on Wednesday fetched $5.4 million for Sir Tim Berners Lee.The original code, although open sourced, has been offered as an NFT(non fungible token). The NFT packages the original archive of dated and time-stamped files containing the source code, written between 3 October 1990 and 24 August 1991. These files contain code with approximately 9,555 lines, the contents of which include implementations of HTML, HTTP and URIs as well as the original HTML documents. The world wide web was invented at the CERN research laboratory in Switzerland. Until very recently, said Sotheby’s,  selling a digital-born artifact was not a possibility, however the advent of NFTs has now made this possible. According to Sotheby’s, every piece of digital media is infinitely reproducible, but there is only one NFT and one corresponding owner. It includes original time-stamped files containing the source code written by Berners-Lee, an animated visualization of the code, a letter written by Berners-Lee on the code and its creation, and a digital “poster” of the full code. They will all be digitally signed by Berners-Lee. GitHub Co Pilot Image credits: Github Github is bringing the power of OpenAI’s technology to coding. This week, the company launched an AI pair programmer that helps programmers write code faster. GitHub Copilot, which is powered by OpenAI Codex, draws context from comments and code, and suggests individual lines and whole functions instantly. Trying to code in an unfamiliar language by googling everything is like navigating a foreign country with just a phrase book. Using GitHub Copilot is like hiring an interpreter,” said Harri Edwards of OpenAI. But can it write perfect code? “No”, says Github. “While we are working hard to make GitHub Copilot better, code suggested by GitHub Copilot should be carefully tested, reviewed, and vetted, like any other code.” GitHub Copilot requires state-of-the-art AI hardware. For now GitHub Copilot will be offered only to a limited number of testers for free. Google Leads MLPerf Results (Source: Google Cloud blog) The latest round of MLPerf benchmark results have been released, and Google’s TPU v4 supercomputers claim to have outclassed NVIDIA in many benchmarks. demonstrated record-breaking performance at scale. This is a timely milestone since large-scale machine learning training has enabled many of the recent breakthroughs in AI, with the latest models encompassing billions or even trillions of parameters (T5,d, GShard, Switch Transformer, and GPT-3). Google also announced that it will be offering TPU v4 Pods on Google Cloud, making the world’s fastest machine learning training supercomputers available to customers around the world. Amazon Scouts In Finland Image credits: Amazon On Thursday, Amazon announced the creation of a new Amazon Scout Development Center in Helsinki, Finland focused on autonomous delivery technology. The new team will consist of over two dozen engineers in Helsinki to start, and will be dedicated to research and development for Amazon Scout, Amazon’s fully electric autonomous delivery service that currently operates in the U.S. New FTC Commissioner Takes On Big Tech 4. We voted to rescind a 2015 statement that unduly constrained FTC's authority to police “unfair methods of competition”—the first step in our work to clarify Section 5. My statement explains how the 2015 statement contravened a congressional directive: https:\/\/t.co\/D4pcsscPee— Lina Khan (@linakhanFTC) July 1, 2021 The new chair of the Federal Trade Commission(FTC), Lina Khan who was sworn in last month has already started to tighten screws on big tech. The Democratic majority commission voted to revoke a 2015 policy that would allow FTC to crack down on big tech’s alleged monopolistic practices. “It is a critical tool that the agency can and must utilize in fulfilling its congressional mandate to condemn unfair methods of competition,” said Khan in her statement. Amazon, on the other hand, filed a request with the FTC seeking the recusal of Khan from antitrust investigations of the company, in light of her extensive past criticisms of the online giant. According to WSJ, Amazon accused Khan of making “numerous” public pronouncements about the company. Google Cloud To Use 5G To Offer Cloud Services According to reports, Google has partnered with Ericsson AB to extend its cloud arm through 5G technology. Google plans to leverage 5G’s flexible open architecture to push the cloud to the edge. This will help hyperscale cloud vendors to enable more flexible, automated networks with improved orchestration, visibility, and control across multi-vendor, multi-cloud and hyperscale cloud-provider environments. “The shift to 5G will place tremendous focus on the ecosystem, and it needs to be an ecosystem that includes CSPs, public cloud providers, application developers and technology providers, all coming together to optimize the user experiences across industry applications,” said Bikash Koley, VP and Head of Google Global Networking.","excerpt":"The World Wide Web’s DNA has been auctioned. Tim Berners Lee’s source code powers the internet as we know today, was put on auction last month by Sotheby’s. A non-fungible token of the original code for the World Wide Web has sold for $5.4 million in a Sotheby’s online auction. The auction, which finished at […]","categories":["AI News"],"tags":["Github Copilot"],"author_name":"Ram Sagar","publish_date":"2021-07-04T10:00:00","publication_year":"2021","word_count":835,"keywords":["Go","machine learning","TPU","OpenAI","AI","AWS","ML","Github Copilot","RAG","Aim","R"],"extracted_tech_keywords":["AI","machine learning","ML","OpenAI","Aim","RAG","AWS","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/githib-copilot-tim-berners-lee-nft-top-news\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10104105,"title":"Top 6 Generative AI Jobs in India","content":"A year ago, OpenAI introduced its generative AI chatbot as a research prototype, and unexpectedly, it became highly sought after in 2023. The demand for generative AI is evident in the doubling of global job postings mentioning AI or generative AI on LinkedIn from July 2021 to July 2023. This aligns with the broader trend in the AI market, projected to grow at an annual rate of 37.3% from 2023 to 2030. As the year concludes, we’ve compiled a list of promising generative AI jobs for those considering a job switch in January, 2024. Let’s take a look at them. ML Engineer, MachineHack Bengaluru-based generative AI startup MachineHack is seeking an adept Project Manager with a strong background in machine learning (ML) to oversee ML projects. The ideal candidate should possess a combination of technical expertise in ML technologies and proficient project management skills. Responsibilities include leading ML projects, managing teams of data scientists and engineers, utilising platforms such as PyTorch and OpenAI, handling AWS tools for ML solutions, employing Docker for containerization, and facilitating communication between technical and non-technical stakeholders. The role emphasises staying abreast of the latest ML and AI developments and integrating innovative technologies into projects. The ideal candidate should have a Bachelor’s or Master’s degree in Computer Science or a related field, a minimum of 3 years of project management experience in ML or similar domains, proficiency in relevant technologies, and demonstrated leadership capabilities. Experience in tech-driven or AI-centric companies is preferred. Apply here. Generative AI Engineer, Siemens Healthineers Also located in Bengaluru, Siemens Healthineers is expanding their footprint and is looking for generative AI engineers. The employee will be tasked with developing and implementing cutting-edge generative models and algorithms by leveraging the latest technologies. The role necessitates a profound comprehension of machine learning, neural networks, and various AI techniques. Responsibilities include researching, designing, and implementing advanced generative models, collaborating with software engineers for model integration, conducting thorough testing, and staying updated on AI and machine learning advancements. Key qualifications for the position include a bachelor’s or master’s degree in computer science or a related field, three to five years of experience as a ML Engineer with a focus on generative AI and neural networks, proficiency in Python, C++, or Java programming, familiarity with deep learning frameworks like TensorFlow or PyTorch, and knowledge of machine learning algorithms and cloud computing platforms such as AWS or Azure. If you think you are fit for this role, apply here. Generative AI Common Platform Engineer, Citigroup Citigroup, a prominent global banking and financial services company, is hiring a Generative AI Common Platform Engineer to contribute to its technological advancement. The role involves developing and maintaining the Generative AI common platform, collaborating on AI solutions, researching and implementing cutting-edge AI technology, and working with data scientists to enhance AI models. The engineer will troubleshoot technical issues, monitor AI model performance, and integrate new technologies into the platform. Compliance with industry standards and risk mitigation are essential, along with training and mentoring team members. Effective communication and continuous improvement of processes are emphasised to drive the future direction of AI at Citigroup. Check out the job here. Senior Principal Outbound Solutions Manager, Generative AI Solutions, Oracle Cloud Infrastructure (OCI) The Senior Principal Outbound Solutions Manager at Oracle, based in Bengaluru, Karnataka, plays a crucial role in the Oracle Cloud Infrastructure (OCI) Generative AI solutions team. This team focuses on expanding OCI’s Generative AI services, aiming to empower customers in solving specific business challenges using Oracle’s Generative AI expertise. The role involves collaborating with data scientists and developers from major Fortune 100 companies to deliver secure and effective solutions. The person will be working closely with strategic customers, both internal and external, to develop end-to-end Generative AI solutions, addressing gaps in the product portfolio, and fostering collaboration between engineering, product, research, and sales teams. The ideal candidate should possess over 10 years of relevant experience, expertise in key industries, and a growth mindset, with hands-on experience in machine learning and artificial intelligence. The emphasis is on customer focus, narrative building, and effectively bridging the technical-business divide. Check out the job description for more information. LLM Engineer, Pure Storage The LLM Engineer – AI \/ ML Engineer plays a key role in the enterprise’s transformation through the application of generative AI technologies. The position involves conducting advanced research and experimentation with cutting-edge technologies in generative AI, focusing on the development of custom applications to address complex business challenges. The role includes leading the development of innovative generative AI solutions, exploring advanced AI models, and architecting robust AI systems that integrate with enterprise IT infrastructure. The role also entails guiding technical project direction, leading and mentoring a team of AI\/ML engineers, and collaborating with stakeholders to translate AI\/ML concepts into business strategies. Requirements for the role include extensive experience in AI\/ML technologies and software development, expertise in relevant programming languages and tools, proficiency in cloud architectures and data engineering, and leadership skills in managing technical teams and driving innovation. Apply here. Lead Generative AI Engineer, Ola Ola is on the lookout for a skilled lead generative AI engineer who will train, optimise, and deploy various large language models, voice and speech foundation models, and vision models. The engineer will focus on architecting robust infrastructure for model deployment, optimising for low latency, high throughput, and cost efficiency. Key responsibilities include refining foundation model infrastructure, implementing optimization techniques, leading LLMOps pipeline development, and driving innovation in model deployment platforms. The qualifications sought include a PhD with over five years or an MS with over eight years of experience in ML Engineering, proficiency in Python, C\/C++, CUDA, and kernel-level programming, expertise in large-scale AI model optimization, and a track record of deploying ML systems at scale using cloud infrastructures and GPU resources. Strong communication, collaboration, and leadership skills are essential. Check out their careers page now.","excerpt":"As the year concludes, we’ve compiled a list of promising generative AI jobs for those considering a job switch.","categories":["AI Trends"],"tags":["Top Trend"],"author_name":"Shritama Saha","publish_date":"2023-12-05T15:00:00","publication_year":"2023","word_count":979,"keywords":["Top Trend","machine learning","artificial intelligence","OpenAI","AI","neural network","ML","Aim","deep learning","generative AI","foundation models"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","generative AI","foundation models","OpenAI","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-6-generative-ai-jobs-in-india\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10066299,"title":"Brain drain: 10 popular Google exits in recent years","content":"In March, Google fired AI researcher Satrajit Chatterjee after he led a team of scientists to challenge a paper Google published in Nature last year. The paper in question spoke about how computers can design parts of chips much faster than a human. Chatterjee reportedly challenged some of the paper’s claims in an internal email and questioned if the technology had been tested. Google, however, stood by the paper and claimed it has been properly vetted. Google also said Chatterjee was ‘terminated with cause.’ Over the years, Google has spent billions of dollars building its Google Brain research team, which it considers key for its future. However, the controversial exits raise questions regarding the tech giant’s commitment to ethics and diversity. Also, is there a pattern to it? Let’s look at some of the top researchers who have left Google in recent times. Timnit Gebru The most notable and most controversial on the list is the former co-lead of Google’s AI ethics team. Timnit took to Twitter to announce that she was fired; however, Google, to date, still maintains that Gebru resigned from her posts. ​​https:\/\/twitter.com\/timnitgebru\/status\/1334352694664957952?lang=en A paper she co-authored, ‘On theDangers of Stochastic Parrots: Can Language Models Be Too Big’, allegedly led to her termination from Google. According to Google’s AI head, the work didn’t meet Google’s bar for publication. The paper, which builds on previously published papers, surveyed the known pitfalls of large language models like GPT-3 and listed four primary risks. Interestingly, the paper passed muster in Google’s internal review process, and objections came up much later. Alphabet CEO Sundar Pichai agreed to investigate the matter after the company came under criticism. As a result, Google made changes to its research and diversity policies. however, the results of the investigations were not made public. On December 2, 2021, Gebru announced the Distributed Artificial Intelligence Research Institute (DAIR). Senior research scientist Alex Hanna and software engineer Dylan Baker left Google to join DAIR earlier this year. \"Google’s short-sighted decision to fire and retaliate against a core member of the Ethical AI team makes it clear that we need swift and structural changes if this work is to continue…1https:\/\/t.co\/ojaCydyS6f— @timnitGebru@dair-community.social on Mastodon (@timnitGebru) December 17, 2020 Margaret Mitchell Margaret Mitchell, the former co-lead of the Ethics AI team at Google along with Gebru, was fired soon after the latter’s ouster. I'm fired.— MMitchell (@mmitchell_ai) February 19, 2021 Mitchell, like Gebru, called for more diversity in Google’s research team, and raised concerns about Google’s handling of criticism. She got the pink slip for allegedly violating the company’s code of conduct as she reportedly used automated software to scan her messages to find discriminatory treatment of Gebru. Mitchell joined Hugging Face after Google. Samy Bengio Samy Bengio, a prominent member of Google Brain, also resigned after the Gebru episode. He is known for leading a large group of researchers working in ML, including adversarial settings. Samy, who served in Google for over 14 years, oversaw Gebru’s team. But, according to him, he wasn’t notified about her termination. Samy took to Facebook to support to Gebru and said, “I stand by you, Timnit.” At present, he is a senior director of AI and Machine Learning Research at Apple. Mustafa Suleyman Mustafa Suleyman co-founded AI\/ML firm DeepMind in 2010 and initially raised funds from Horizons Ventures and entrepreneurs like Elon Musk. He joined Google in 2014 when Google acquired DeepMind for USD 650 million. Then, he was Google’s vice president of product management and policy for AI. Even though he was not an AI researcher, Suleyman was instrumental in pushing DeepMind into healthcare and was vocal about not using AI for military applications. Suleyman was accused of an aggressive management style by colleagues at DeepMind. Currently, he is the co-founder and CEO of Inflection AI. Alex Henna Alex Henna, a senior research scientist at Google Brain, also resigned from Google after the expulsion of Gebru and Mitchell. She was at Google for around four years and worked on ethical AI and ML fairness. In a Medium article, Hannah said: Many folks — especially Black women like April Curley and Timnit — have made clear just how deep the rot is in the institution. I am quitting because I’m tired” she said. Eric Jang Senior research scientist Eric Jang left Google Robotics and joined Norwegian robotics company Halodi Robotics as vice president of AI. He was a t Google for nearly six years. Jang is an expert in reinforcement learning, deep learning and generative modelling. This is my last week at Google Brain, after nearly 6 years on the robotics team. Thank you Google, it's been really fun!✌️https:\/\/t.co\/olO2ZGNeAP— Eric Jang (@ericjang11) March 21, 2022 Niki Parmar Niki Parmar joined Google as a software engineer and later moved to the Google Brain team. She started as a research engineer, and rose to the position of a staff research scientist in 4.5 years.  At Google, Parmar was involved in research related to understanding self-attention and how other inductive biases can be used for the improvement of different models across various tasks like machine translation, language modeling, and perception. She co-authored the seminal paper on Transformers — Attention Is All You Need. Parmer is now the CTO at Adept, an AI firm she co-founded with her former colleagues at Google. Ashish Vaswani Ashish Vaswani was a member of the Google Brain research team, where his focus was on developing pure attention-based models. He also co-authored the Transformer papers. After leaving Google, he co-founded Adept, and is the Chief Scientist there. After 5+ wonderful years in Google Brain, working at the forefront of ML alongside inspiring colleagues, I'm excited to share my new adventure. We started Adept with the mission to build the future of human-computer collaboration. https:\/\/t.co\/t8dMqZfbSZ.— Ashish Vaswani (@ashVaswani) April 26, 2022 Tatiana Shpeisman Tatiana Shpeisman served at Google for nearly five years as a senior engineer manager, where she oversaw a team working on TensorFlow graph compiler, MLIR, and TensorFlow infrastructure for GPUs and CPUs. She is currently the compiler engineering director at Modular AI. Augustus Odena Augustus Odena was also a member of the Google Brain team for nearly six years. In a blog, he said: “A lot of my work has been on Program Synthesis — in which we try to get computers to program themselves. I led Brain’s work on Program Synthesis with Large Language Models, co-created Google Sheet’s SmartFill programme synthesiser, and jointly invented the Scratchpad Technique for getting Transformers to perform multi-step reasoning.” He was a senior research scientist at Google Brain team when he parted ways.","excerpt":"Satrajit Chatterjee was allegedly fired for challenging a research paper Google published in Nature.","categories":["Global Tech"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2022-05-04T17:00:00","publication_year":"2022","word_count":1097,"keywords":["Hugging Face","machine learning","artificial intelligence","AI","ML","Transformers","Aim","deep learning","TensorFlow","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","Aim","TensorFlow","Hugging Face","Transformers","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/brain-drain-10-popular-google-exits-in-recent-years\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10113265,"title":"Sequoia Capital Innovates with Open Source Fellowship for Developers","content":"Sequoia Capital, a venture capital firm from Silicon Valley, has launched the Sequoia Open Source Fellowship. This program plans to fund up to three developers each year. It allows them to work full-time on their projects without worrying about money and without giving up any ownership of their work. In 2023 their first and only Fellow was Sebastián Ramírez Montaño who created FastAPI, an open source tool that allows developers to create applications quickly and maintain them easily. This initiative addresses the problem of not enough funding in the open source community, a problem made clear by major security issues in the past. Open source software development is important for technology, yet, developers often have to balance their open source work with jobs that pay, leading to a lack of funding and support. The Sequoia fellowship aims to close this gap by giving developers money to cover their living costs for up to a year, so they can focus on their open source projects. The fellowship is available to any developer working on an open source project, with ongoing application acceptance. Sequoia wants to support projects that are widely used and make a difference, showing the importance of open source software in technology. By funding these projects, Sequoia not only helps the wider tech community but also follows its strategy of investing in companies rooted in open source, like MongoDB, Confluent, and Temporal. Sequoia’s move is part of a larger trend where tech companies and venture capital firms offer grants and funding to support open source development. These efforts aim to make the software supply chain safer and recognise the importance of open source contributions to technology.","excerpt":"Sequoia Capital’s fellowship, offers a financial boost to developers building open source projects.","categories":["AI News"],"tags":["Open Source AI"],"author_name":"K L Krithika","publish_date":"2024-02-20T15:54:50","publication_year":"2024","word_count":277,"keywords":["Go","API","funding","programming_languages:R","AI","MongoDB","venture capital","Open Source AI","Aim","FastAPI","R"],"extracted_tech_keywords":["AI","Aim","FastAPI","MongoDB","R","Go","API","venture capital","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sequoia-capital-innovates-with-open-source-fellowship-for-developers\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10016129,"title":"This New AI Algorithm Can Master Games Without Being Told The Rules","content":"Two years after DeepMind introduced AlphaZero, an AI-based program that could challenge humans at the game of chess, the researchers have demonstrated MuZero. The researchers at DeepMind describe it to be a significant step towards formulating general-purpose algorithms. While its predecessor, AlphaZero could learn games such as Go, chess, and shogi from scratch, MuZero can master these games (along with Atari) without being told the rules. It can plan winning strategies in unknown environments. This is particularly significant with respect to games like Atari, where the rules and dynamics are generally complicated and unpredictable. MuZero’s Advantage Over Its Predecessors MuZero was first introduced in 2019 as a preliminary paper at the NeurIPS 2019 conference. It combines AphaZero’s lookahead tree search with a new state-of-art result on the art result of Atari benchmark. MuZero demonstrates a leap ahead in the capabilities of reinforcement learning algorithms. The natural step in the evolution of artificial intelligence is incorporating the ability to learn quickly and accurately generalise to new scenarios, just like the human mind. There have been many methods that scientists have adopted over the years to build this capability, two of which are lookahead search and model-based planning. Lookahead search strategy relies mainly on the game’s rules or an accurate simulator and relies heavily on the given knowledge of their environment’s dynamics. It works great when preparing algorithms for classic games such as checkers, poker or chess, like in AlphaZero. However, they do not take too well to complex real-world problems and cannot be necessarily decomposed into simple rules. On the other hand, model-based systems first learn an accurate model of the environment’s dynamics and then use it to plan, helping it do well even in complex real-world situations. They do not use a learned model but instead estimate the best action that can be taken next. Model-based systems have significant disadvantages, as well. For visually rich domains, as in Atari, modelling every aspect of the environment becomes very complicated for even the model-based system. Credit: DeepMind To overcome the limitations of the previously mentioned lookahead search and model-based planning, MuZero uses a different approach. Instead of modelling the entire environment, MuZero chooses only the critical aspects for the decision-making process and models them. The factors are selected based on three elements — how good is the current position (value), the best action to be taken (policy), and how good was the last action (reward). MuZero’s Performance The DeepMind researchers chose Go, chess, shogi and Atari to test the capabilities of MuZero. While Go, chess and shogi were used for assessing its performance on challenging planning problems, Atari was used as a benchmark for checking its capabilities in a visually complex setting. It was observed that MuZero outperformed previous algorithms used for Atari and matched AlphaZero’s Go, chess, and shogi performance. Further study also showed that MuZero’s capabilities were enhanced by 1000 Elo, a unit to measure a player’s relative skill, as the time taken per move by the algorithm was increased from one-tenth of a second to 50 seconds. The pattern is comparable with the difference between an amateur and a professional human professional player. It was observed that MuZero could generalise actions and situations and need not search for all possibilities in games like Atari to learn effectively. Read the full paper here. Wrapping Up Facebook, too, announced an AI bot ReBeL that could play chess (a perfect information game) and poker (an imperfect information game) with equal ease, using reinforcement learning. The company called it a positive step towards creating general AI algorithms that could be applied to real-world issues related to negotiations, fraud detection, and cybersecurity. With MuZero, researchers hope to extend its application to tackling real-world challenges such as in robotics, industries, and others.","excerpt":"Two years after DeepMind introduced AlphaZero, an AI-based program that could challenge humans at the game of chess, the researchers have demonstrated MuZero. The researchers at DeepMind describe it to be a significant step towards formulating general-purpose algorithms. While its predecessor, AlphaZero could learn games such as Go, chess, and shogi from scratch, MuZero can […]","categories":["Deep Tech"],"tags":["decision tree algorithm","purpose of ai"],"author_name":"Shraddha Goled","publish_date":"2020-12-28T14:00:00","publication_year":"2020","word_count":627,"keywords":["Go","artificial intelligence","decision tree algorithm","programming_languages:R","AI","programming_languages:Go","ai_applications:robotics","purpose of ai","R","fraud detection"],"extracted_tech_keywords":["AI","artificial intelligence","fraud detection","R","Go","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/this-new-ai-algorithm-can-master-games-without-being-told-the-rules\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10093990,"title":"Microsoft Steps in to Rescue Web-building From the Rut","content":"Over the years, one notable development that has democratised web development is the emergence of DIY web development platforms like WordPress and Wix. These have made creating a website cost-effective and basically a breeze for individuals and businesses, reducing the reliance on external developers. The global market for website builders, valued at approximately $2 billion in 2022, is expected to expand significantly and reach a revised market size of around $3.8 billion by 2030. This growth is anticipated to occur at a compound annual growth rate (CAGR) of 8.2% during the forecast period of 2022 to 2030. However, while these platforms have gained popularity, the industry seems to be stuck in a loop of similar offerings without much innovation or imagination, reaching a point of saturation. Microsoft has stepped in to change the game with its introduction of Copilot in Power Pages. Copilot is an AI-powered assistant for Microsoft’s low-code business website creation tool. It streamlines the web design process by generating text, forms, chatbots, web page layouts, and even image and site design themes based on user prompts. This AI assistant allows users to describe their needs in natural language, eliminating the need to start from scratch and saving valuable time and effort. One key aspect that sets Copilot apart from competing solutions is its versatility. Leveraging OpenAI’s powerful GPT-3.5 model, Copilot can provide a wide range of AI-generated suggestions tailored to the user’s requirements. However, Microsoft emphasises responsible use and cautions against using Copilot to generate spam sites. The system includes safeguards such as offensive content filtering to ensure responsible AI utilisation. Users retain control over the AI-generated suggestions and can provide feedback on any irrelevant or inappropriate output. This feedback loop is essential for improving the accuracy of the system and ensuring that the AI assistant aligns with the user’s vision and requirements. Furthermore, the chatbot feature in Power Pages incorporates safeguards like a whitelist of URLs to prevent the presentation of malicious or misleading information to users. AI integration in web pages The integration of AI into web development extends beyond Microsoft’s Copilot in Power Pages. Before the introduction of Copilot in Power Pages, several website builders gained popularity for their ease of use and extensive features. The WordPress community has also been discussing the potential benefits of AI integration. While there are no immediate plans to add AI to the WordPress core, contributors have explored various possibilities. Suggestions include AI diagnostic co-pilots for issue identification and solutions, AI-powered content marketing toolkits, and AI chatbots as collaborators in future WordPress phases. Currently, AI integration with WordPress is available through plugins like CodeWP, All in One SEO, Rank Math, SEOPress, and WordLift. Similarly, other players in the industry have also integrated AI into several aspects of their website-building tools. Webflow, known for its no-code tools for website creation, is introducing AI-powered features that enable users to customise templates through verbal prompts. The AI adjusts content, images, and structure based on user needs. Users can manually edit designs with AI assistance, which also generates drafts, optimises search features, and translates content. Webflow aims to streamline web designing, making it accessible while retaining manual design capabilities. Pricing and availability for the AI features are yet to be determined as they are currently in limited alpha testing. Wix.com has launched its AI Text Creator within the Wix Editor, aiming to improve the quality of site content and streamline the website-building process for its users. The feature allows users to generate high-quality and tailored content, including titles, taglines, and paragraphs, by providing inputs and answering prompts. By leveraging AI technology, Wix aims to optimise the website creation process and provide users with professional-looking content. The AI Text Creator is currently rolling out to Wix users in English, with plans for wider availability. Wix.com is a leading SaaS platform that enables users to create, manage, and grow their online presence. There’s another no-code website builder backed by Y Combinator, called Typedream, which simplifies website creation by providing an easy-to-use editor and various customisable elements such as forms, blogs, and pages. With its drag-and-drop interface, users can effortlessly develop professional-looking websites without coding knowledge. The platform offers a wide range of templates that can be personalised to match a business’s branding and style. Typedream also includes time-saving AI features like automatic layout suggestions, colour palettes, and typography, which reduce the need for extensive design brainstorming. While the free version is suitable for getting started, the paid version offers additional features like a custom domain, code injection, and detailed analytics. Web developers still stand a chance While the availability of DIY platforms may reduce the demand for basic website development, there are still ample opportunities for web developers to thrive and provide specialised services in a competitive market. Experienced professionals who can deliver customised solutions and tackle complex challenges will continue to be sought after. This highlights the importance of expertise in specialised areas within web development. The field of web development remains relevant and dynamic as it evolves alongside the ever-changing digital landscape. Developers who adapt to new technologies, stay informed about emerging trends, and offer unique skills and services and are well-positioned for success. The industry continues to evolve and adapt, ensuring that web development remains an essential discipline in the digital era. With advancements like AI-powered assistants, web developers can embrace innovation and drive the future of web design, catering to the evolving needs of businesses and users alike.","excerpt":"Microsoft’s Copilot in Power Pages is an AI-powered assistant that streamlines the web design process by generating text, forms, chatbots, web page layouts, and even image and site design themes based on user prompts","categories":["Global Tech"],"tags":["AI integration","Microsoft","Y Combinator"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-05-26T15:00:00","publication_year":"2023","word_count":908,"keywords":["AI integration","TPU","OpenAI","AI","chatbots","ML","Y Combinator","Git","RAG","Aim","analytics","R","Microsoft"],"extracted_tech_keywords":["AI","ML","analytics","OpenAI","Aim","RAG","chatbots","TPU","R","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-steps-in-to-rescue-web-building-from-the-rut\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25900,"title":"Understanding The Workings Of Natural Language Processing Vs Natural Language Generation","content":"Digital assistants, chatbots and other conversational interfaces have become the most widely-adopted technologies in the recent days. Their ability to carry human-like conversations in a seamless manner could be attributed to their tremendous popularity, which is in turn driven by two underlying technologies — Natural Language Processing (NLP) and Natural Language Generation (NLG). These two branches of machine learning are enabling the conversion of human language to computer commands and vice versa. As these technologies are enabling humans to have a conversation with machines in an effective manner and augmenting human intelligence, we bring to you an article that discusses the differences between NLP and NLG, their working and some common use cases. NLP vs NLG: Understanding the basic difference What Is NLP? The most popular definition of NLP describes it as a process which turns text into structured data when the computer reads the language. In short, NLP is computer’s reading language. It could be roughly said that in NLP, the system ingests what is being said, breaks it down, analyses it, determines appropriate action and responds in a language that the user (human) will understand. A combination of computer science, AI and computational linguistics, NLP encompasses all the mechanism that interprets and produces human language in a way that they would understand. It relies on functions such as language filter, sentiment analysis, subject matter classification, location detection, and others. What Is NLG? A subset of NLP, NLG is a computer’s “writing language” that turns structured data into text to provide information in human language. Working on a concept of ‘data-in and data-out’, in NLG, software system generates narratives and reports that summarise input data. When data is fed into the system, NLG produces data-rich information that assesses text to form insights. NLP vs NLG: Working Of A Chatbot Just the way human conversation involves two-way communication, chatbot also follows the same route. There is a slight difference in the channel of communication and the fact that you are talking to a machine. When a message is given to a bot, it picks it up, and using NLP, the machine converts the text into codified commands for itself. This data is then sent to the decision engine. The bot now processes this information, which is presented to you in the form of a question. At this stage, the bot analyses pre-fed data about various parameters and displays the same to the user based on the query. In the entire process, the computer is converting natural language into a language that computer understands, bringing into process, speech recognition. The commonly used mechanism for speech recognition system is the Hidden Markov Models (HMMs), that converts speech to text to determine what user said. It does so by listening to what you speak, breaking it down into small units, and analysing it to generate output or information in the form of text. A key step after this is the Natural Language Understanding (NLU), which is another subset of NLP that tries to understand the meaning out of the text form. It is important that computer understands what each word is, that part which is carried out by NLU. While sifting through the vocabulary, grammar and other information, NLP algorithms use statistical machine learning to apply these rules to the natural language and determine the most likely meaning behind what was said. NLG, on the other hand, is a system that generates natural language leveraging artificial intelligence and computational linguistics. It can also translate this text into audible speech with text-to-speech. NLP system first determines what information to translate into text, then organises structure of how it’s going to say that. Now using the set of grammar rules, NLG system forms complete sentences. Use Cases Digital assistants and chatbots are just a few of the many applications that NLP has found. It finds utility in areas such as cybersecurity articles, white papers, studies etc. It is also used to carry sentiment analysis of online content to improve upon services and come up with better offerings for the customers. NLG on the other hand, is commonly used in Gmail, where it creates automatic answers for you. It is also used in creating reports from complex data such as creating narrative descriptions of company data and charts. On A Concluding Note While NLP and NLG are sometimes looked upon as completely unrelated at most times, it is not entirely true. It is the whole process of reading and writing that accomplishes the task and hence both NLP and NLG are inter-related.","excerpt":"Digital assistants, chatbots and other conversational interfaces have become the most widely-adopted technologies in the recent days. Their ability to carry human-like conversations in a seamless manner could be attributed to their tremendous popularity, which is in turn driven by two underlying technologies — Natural Language Processing (NLP) and Natural Language Generation (NLG). These two […]","categories":[],"tags":["natural language generation","Natural Language Processing"],"author_name":"Srishti Deoras","publish_date":"2018-06-29T10:59:54","publication_year":"2018","word_count":756,"keywords":["machine learning","artificial intelligence","natural language generation","AI","chatbots","Natural Language Processing","ML","sentiment analysis","TPU","RAG","NLP","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","RAG","chatbots","sentiment analysis","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/understanding-the-workings-of-natural-language-processing-vs-natural-language-generation\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10137201,"title":"&#8216;Smart Glasses Will Replace Phones By 2030,’ Says Meta Chief Mark Zuckerberg","content":"It’s 2030. You wake up, and instead of reaching for your phone, slip on sleek smart glasses. Throughout the day, you interact with a powerful AI assistant that uses the glasses to prep you with useful information, seamlessly merging your physical and digital worlds. Using advanced display technology, it projects holograms directly into your field of vision. Sounds exciting? This is what Meta CEO Mark Zuckerberg predicted five years ago while announcing the AR glasses. On track with the projection, smart glasses are expected to replace phones by 2030 and become the next major computing platform. At Meta Connect this year, Zuckerberg unveiled a range of new tech, including the $300 Quest 3S VR headset. However, the real highlight was Project Orion, a pair of AR smart glasses. Many companies have attempted to create AR glasses, but the results have often been bulky or tethered by cables. Meta’s Project Orion stands out by avoiding these issues. Despite all the high-end technology packed into the frame, these hologram-generating glasses almost resemble regular eyewear, bearing a striking similarity to Meta’s Ray-Bans. The Orion glasses pack in a host of features like eye-tracking, hand-tracking, voice controls, and even a neural interface, although it reads signals from your wrist rather than your brain. The glasses also come with a wireless compute puck that resembles a sleek power bank. While they don’t require a laptop or phone to operate, the puck must be within a few feet for them to function, which means you’ll likely need to carry it in your pocket. Several publications tested the advanced hardware at the event. CNBC said the holograms “felt totally normal and very natural” thanks to the high-quality displays. NVIDIA CEO Jensen Huang, renowned for his AI advocacy, got a chance to try on the Orion glasses. “The head tracking is good, the brightness is good, the colour contrast is good, and the field of view is excellent,” he said enthusiastically. Some of the reactions after $META unveils it's first AR holographic headset \"Orion\". $NVDA Jensen Huang included in their reaction video. pic.twitter.com\/aAn8rvkzNM— Financial Berg (@FinancialBerg) September 25, 2024 Where Are We Heading? Coincidently, Zuckerberg has returned at a busy time for Meta, which continues to expand its reach in the tech industry. While Google Glass may have been an expensive failure for tech giant Google, it was a valuable case study for the rest of the tech industry. It taught the developers a lot about what people do not want, and those insights were instrumental in shaping the current crop of smart glasses. To make matters worse, it also did appear like a prop from ‘The Matrix’, whereas most of today’s smart glasses have a much more low-key design. This helps wearers avoid being stigmatised as modern-day “glassholes” trying to film others without their consent (even if it doesn’t actually stop this from happening). Meta envisions that within the next ten years, up to 2 billion people who wear regular glasses will transition to the smart ones, with even those who don’t need thema medically, eventually adopting them. While the smart glasses market is growing, there’s still no guarantee that the device will ever become mainstream, let alone replace smartphones as our go-to personal gadget. While the industry is still searching for its “killer app” – a feature that would allow people to access everything with just one tap – this might be the one thing Google got right about smart glasses with Google Glass. Zuckerberg’s ‘normal-looking glasses with a camera, microphone and great audio that can stream video and capture content without a display’ is Ray-Ban Meta, which was jointly developed by Meta and Ray-Ban. However, he noted that AI was the most important feature for smart glasses and described Ray-Ban Meta, which has access to Meta AI, as ‘on the way to building a complete holographic pair of glasses’. However, as impressive as Orion may appear, the company has admitted that this specific model will never reach the consumer market. The estimated production cost of the glasses is around $10,000 per unit, which is prohibitively expensive for consumer sales. The CEO indicated that substantial work is required to make the glasses commercially viable and affordable for the average consumer. “It may be really difficult to make normal-looking glasses that can do holograms at an affordable price but you can now get normal-looking glasses with a camera, microphone, and good audio that can stream video and capture content, even if they don’t have a display,” Zuckerberg said. He also emphasised that the technical hurdles involved in developing AR glasses capable of delivering a seamless experience are immense. These challenges include creating a compact form factor while integrating features like holographic displays and interactive capabilities. Meta’s team believed there was less than a 10% chance of successfully achieving their ambitious goals when they began this project nearly a decade ago. Currently, the glasses are a prototype available only to Meta employees and select partners, not slated for consumer release. They envision that future iterations could potentially be available by 2027, contingent on technological advancements that lower production costs and enhance functionality. For now, they are focusing on internal development and demonstrating the technology rather than rushing to market with an unready product.","excerpt":"The Orion smart glasses have been in the works for almost a decade, but Zuckerberg thinks they aren’t quite ready for the mainstream.","categories":["AI Features"],"tags":["Mark Zukerberg","Meta"],"author_name":"Tarunya S","publish_date":"2024-10-01T14:37:35","publication_year":"2024","word_count":875,"keywords":["Go","API","Meta AI","Meta","ELT","AI","ML","Git","RAG","Ray","Mark Zukerberg","R"],"extracted_tech_keywords":["AI","ML","Meta AI","Ray","RAG","R","Go","Git","API","ELT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/smart-glasses-will-replace-phones-by-2030-says-meta-chief-mark-zuckerberg\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":42367,"title":"Google Colab Vs Kaggle Kernels: Which Of The Two Platforms Should You Go For?","content":"Google has two free cloud platforms for GPUs — Google Colab and Kaggle Kernels. For machine learning enthusiasts and professionals, both the platforms come in very handy. Both platforms are by Google and so naturally, they have many similarities. But they also have some minor differences between them. Both platforms are free and they give a Jupyter Notebook environment access. Here are the differences in specific features for the two. 1. Language Support Kaggle Kernels: Kaggle Kernels supports Python 3 and R. Google Colab: Google Colab supports the languages of Python and Swift. 2. Saving Notebooks Google Colab: Notebooks can be saved to Google Drive. Notes can be added to Notebook cells. One can also easily integrate the saved notebooks which can be easily uploaded to the GitHub repositories. Kaggle Kernels: Saving notebooks is easier here than in Colab. A major drawback of both platforms is that the notebooks cannot be downloaded into other useful formats. 3. TPUs Google Colab: Google has its self-made custom chips called TPUs. But a drawback is that TPUs do not work smoothly with PyTorch when used on Colab. However, if TensorFlow is used in place of PyTorch, then Colab tends to be faster than Kaggle even when used with a TPU. Kaggle Kernel: In Kaggle Kernels, the memory shared by PyTorch is less. In general, Kaggle has a lag while running and is slower than Colab. 4. Keyboard Shortcuts Google Colab: Colab is not as related to Jupyter Notebooks in terms of its shortcuts as Kaggle is. The shortcuts of Jupyter Notebooks are not completely imported to Colab. Kaggle Kernel: Most keyboard shortcuts from Jupyter Notebook are exactly alike in Kaggle Kernels, making it easier for a person working in Jupyter Notebooks to work in Kaggle. 5. Memory Google Colab: Colab has an Nvidia Tesla K80. It is definitely better than Kaggle in terms of speed. But integrating with Google Drive is not very easy. Every session needs authentication every time. Unzipping files in Google is also not very easy. Kaggle Kernels: Kaggle had its GPU chip upgraded from K80 to an Nvidia Tesla P100. Many users have experienced a lag in Kernel. It is slow compared to Colab. 6. Execution Time Google Colab: Colab gives the user an execution time of a total of 12 hours. After every 90 minutes of being idle, the session restarts all over again. Kaggle Kernel: Kaggle claims that they serve a total of 9 hours of execution time. But Kaggle Kernel shows only 6 hours of available time for execution per session. After every 60 minutes, the sessions can also restart all over again. Verdict Both the Google platforms provide a great cloud environment for any ML work to be deployed to. The features of them both are equally competent. Notebooks can be downloaded and later uploaded between the two. However, Colab comparatively provides greater flexibility to adjust the batch sizes. Saving or storing of models is easier on Colab since it allows them to be saved and stored to Google Drive. Also if one is using TensorFlow, using TPUs would be preferred on Colab. It is also faster than Kaggle. For a use case demanding more power and longer running processes, Colab is preferred.","excerpt":"Google has two free cloud platforms for GPUs — Google Colab and Kaggle Kernels. For machine learning enthusiasts and professionals, both the platforms come in very handy. Both platforms are by Google and so naturally, they have many similarities. But they also have some minor differences between them. Both platforms are free and they give […]","categories":["Global Tech"],"tags":["Colab","Google","Google Colab","GPU","Kaggle","kernels","Natural Language Processing","TPU"],"author_name":"Disha Misal","publish_date":"2019-07-12T16:00:54","publication_year":"2019","word_count":538,"keywords":["TPU","Kaggle","machine learning","AI","PyTorch","Natural Language Processing","ML","kernels","Colab","Python","Aim","Jupyter","Google","Google Colab","TensorFlow","GPU"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","TensorFlow","PyTorch","Jupyter","Colab","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-colab-vs-kaggle-kernels-which-of-the-two-platforms-should-you-go-for\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094449,"title":"Qualcomm Wants to Bring ChatGPT to Your Phone","content":"US-based semiconductor technology company Qualcomm is working to help run AI models like GPT-4 and DALL-E2 on smartphones through its new generation processors, according to reports. Currently, users are able to access these models through the web, and also through a mobile app ( currently iOS devices). However, the application still runs on the web and it’s a costly affair. Dylan Patel, chief analyst at semiconductor research firm SemiAnalysis, told the Information that running ChatGPT is costing OpenAI around USD700,000 a day. The cost incurred is mostly for running the servers. Patel believes the cost could be even higher for running GPT-4, which is the most advanced large language model so far. In a whitepaper released in May, Qualcomm said that generative AI-based search cost per query is estimated to increase by 10 times compared to traditional search methods. Hence, the utilisation of Hybrid AI enables generative AI developers and providers to leverage the computing power offered by edge devices, leading to cost reduction. “A hybrid AI architecture (or running AI on device alone) offers the additional benefits of performance, personalisation, privacy, and security – at a global scale.” Qualcomm said. AI on every device Earlier this year, a group of Qualcomm engineers managed to run text-to-image AI model Stable Diffusion on an Android device. Interestingly, Qualcomm might not be the only company working on a technology to run AI models on devices. Apple too is developing and promoting the use of on-device AI models for a range of tasks such as speech recognition, natural language processing, and computer vision. The company has been investing heavily in the development of AI chips and algorithms that can run on its devices, including iPhones, iPads, and Macs. Currently, it’s not known if Apple is building any GPT models like Large Language Models ( LLM). But Apple could reveal similar technology that completely runs on Apple devices offering  users faster, more responsive, and privatised experiences. Running AI models could be the approach for many companies such as Google and Microsoft, two companies battling it out for AI supremacy.  Earlier this year, a developer has already showcased how he managed to run a model like ChatGPT3.5 turbo completely on a laptop. You will own your own AI.Final testing on a new massively smaller 100% locally running ChatGPT 3.5 turbo type of LLM AI in your hard drive on any 2015+ laptop. I will have pre-configured downloads and it is massively smaller than most models I have, just 4gb.Out soon! pic.twitter.com\/KnZkICmGPV— Brian Roemmele (@BrianRoemmele) April 5, 2023","excerpt":"The utilisation of Hybrid AI enables generative AI developers and providers to leverage the computing power offered by edge devices, leading to cost reduction","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-06-05T12:00:48","publication_year":"2023","word_count":422,"keywords":["Go","ChatGPT","DALL-E","OpenAI","AI","computer vision","RAG","GPT","generative AI","R"],"extracted_tech_keywords":["AI","computer vision","generative AI","ChatGPT","OpenAI","RAG","R","Go","GPT","DALL-E"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/qualcomm-wants-to-bring-chatgpt-to-your-phone\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10039059,"title":"Delhivery Promises To Fly Charters With Oxygen Concentrators In India","content":"A few days ago, Sahil Barua, the co-founder of Delhivery posted on LinkedIn that they are planning to fly charters into India with oxygen concentrators. Barua said, “We’re flying charters into India with oxygen concentrators and other essential supplies and can build more capacity on demand.” “If you need help with logistics or wish to collaborate with us on this please reach out to Vikas Kapoor or to me immediately (ceo@delhivery.com),” Barua added. India has been facing the highest single-day surge of COVID-19 recorded by any country and with the surge of the deadly virus, currently, the country is suffering from a scarcity of medical oxygen. On Sunday, Vinod Khosla, the co-founder of the logistics and supply chain startup tweeted that the company is willing to fund hospitals in India that need funding to import a planeload of oxygen and other supplies. Khosla also asked the public hospitals and NGOs to reach out in case of any emergency. The tweet stated, “I’m willing to fund hospitals in India that need funding to import bulk planeloads of oxygen or supplies into India to increase supply. Public hospitals\/NGO’s also pls reach out  @PMOIndia @MoHFW_INDIA @timesofindia @INCIndia #IndiaFightsCOVID19 @htTweets @IndianExpress @GiveIndia” I'm willing to fund hospitals in India that need funding to import bulk planeloads of oxygen or supplies into India to increase supply. Public hospitals\/NGO's also pls reach out @PMOIndia @MoHFW_INDIA @timesofindia @INCIndia #IndiaFightsCOVID19 @htTweets @IndianExpress @GiveIndia— Vinod Khosla (@vkhosla) April 24, 2021 In another tweet, Karthik S, a company executive tweeted that the logistics startup has chartered two planes from China most likely on Wednesday and Friday to help import oxygen concentrators. The company is doing this at minimal margins for oxygen concentrators and other essentials, and have spare capacity as things stand. Delhivery (my company) is chartering two planes from China (most likely this Wednesday and Friday) to help import oxygen compressors. We are doing this at minimal margins for compressors and other essentials, and have spare capacity as things stand.— Karthik S (@karthiks) April 24, 2021 Commenting on the tweet, Karthik S added, “If interested, please contact us at ceo@delhivery.com. In case the demand far exceeds the current supply, we can arrange for additional flights as well.” If interested, please contact us at ceo@delhivery.com In case demand far exceeds the current supply, we can arrange for additional flights as well.— Karthik S (@karthiks) April 24, 2021","excerpt":"Sahil Barua, the co-founder of Delhivery posted on LinkedIn that they are planning to fly charters into India with oxygen concentrators.","categories":["AI News"],"tags":["delhivery","oxygen concentrators"],"author_name":"Ambika Choudhury","publish_date":"2021-04-27T15:01:11","publication_year":"2021","word_count":397,"keywords":["Go","funding","programming_languages:R","AI","programming_languages:Go","oxygen concentrators","delhivery","R","startup"],"extracted_tech_keywords":["AI","R","Go","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/delhivery-promises-to-fly-charters-with-oxygen-concentrators-in-india\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061978,"title":"Council Post: How to build robust forecasting models amid chaos","content":"The pandemic was a reality check for companies across the world. No matter how prepared you think you are, black swan events like Covid-19 can throw your company into disarray. However, sitting on their hands is not an option for modern businesses. To bring antifragility to their preparedness, companies leverage AI and ML to develop robust forecast models. The quality of such models rely on the data fed into them. But what if such events put a strain on the data collection pipeline to start with? We have sounded out the thought leaders in the industry to figure out ways to tackle such situations. Take a collaborative approach The collaboration across industry, government, agencies, and academia can help significantly increase the breadth of the data for the application of X-analytics techniques that can amalgamate structured and unstructured streams such as text, image, audio, and video to address the contextual need more effectively. This will also be enabled by traction with the creation and utilisation of synthetic data to overcome the privacy, and confidentiality aspects of the data owned by enterprises, agencies, governments, and the like. Satyamoy Chatterjee, Executive Vice President at Analyttica Datalab Break it down A good way that I have seen work is to break it down by what kind of refinements are needed on the input (data), process (the models and algorithms), and output (adjustments to final prediction) and take it, step by step. Ruble Joseph, Lead Strategist (VP) – Global Data Science and Analytics Practice at eClerx Robust techniques To keep forecasts meaningful, we adopted a couple of strategies – we added new data points (incorporated manual forecasts by demand planners, incorporated supply constraints based on qualitative inputs), changed our model selection process based on recent sales data vs looking at year on year changes, changed the operational process for forecasting, and incorporated learning-based model techniques like deep learning in addition to casual models like Bayesian structural time series, ARIMAX, etc. Ajoy Singh, Chief Operating Officer at Fractal Address the drift Identifying the type of drift – i.e., concept drift, data\/covariate shift, and the appropriate mitigation strategy to combat the effects of the drift on the quality of the predictive model, holds the key to building models that can stand the test of time. This can be achieved by comparing pre-drift forecasting to post-drift forecasting, detecting data distribution changes or I\/O relationships, developing novel difference transform approaches, whereas adding domain-specific and pandemic related external regressors to pre and post-event training data can also help improve the quality of models. Ashish Kumar, VP – Data Science at Salesken Take the long view For all forecasting\/prediction, we need to consider this time period; one can keep a separate indicator (or something similar) to flag it out. Moreover, macroeconomic factors need to be considered to make predictions with relatively higher accuracy. To complicate things, customer behaviour towards certain product categories has drastically changed with more reliance on digital services. Overall, while these factors mentioned above should be considered to adjust the forecasting\/ predictive models, from an algorithm standpoint, we should rely more on machine learning models that give relatively higher weightage to certain data points (Sequence to Sequence, Encoder-Decoder with LSTM, etc.) Anirban Nandi, Head of Analytics (Vice President) at Rakuten India Be versatile When generating predictive models depends primarily on the specific use case and the industry under consideration; some generic approaches include: While creating the models, we can consider data over a longer period of time; or validate the existing model across different time periods and scenarios. If model performance is good, the model is good to go; else, we may need to recalibrate the model by removing outliers or providing weights across different time periods. We may consider macroeconomic parameters or alternate data, which may give more real-time information for assigning weights to data about time periods more specifically affected by the pandemic. We can leverage alternate data such as auto sales to help predict the spending propensity across different customer segments. In specific instances, e.g., for industries like travel, where the impact of lockdown has been significant – we may create models removing the data corresponding to the specific time period. Swati Jain, Vice President Analytics at EXL This article is a collation of quotes by members of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"We can leverage alternate data such as auto sales to help predict the spending propensity across different customer segments.","categories":["AI Features"],"tags":["pandemic","Predictive AI","predictive analysis","predictive analytics","predictive modelling","uncertainty in AI"],"author_name":"Amit Naik","publish_date":"2022-03-02T18:00:00","publication_year":"2022","word_count":749,"keywords":["data science","pandemic","machine learning","TPU","AI","ML","RAG","Ray","Aim","deep learning","analytics","uncertainty in AI","predictive modelling","predictive analytics","predictive analysis","Predictive AI"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","Aim","Ray","RAG","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-how-to-build-robust-forecasting-models-amid-chaos\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":41753,"title":"AI Turns Casting Director, Can Now Decide The Perfect Blockbuster Movie Cast","content":"Showbiz is one of the most lucrative industries in the world, but there’s no specific formula for churning out blockbusters. That is why filmmakers from across the world are tapping artificial intelligence to take help with audience composition, CGI and even music. 20th Century Fox’s Merlin 20th Century Fox, one of the biggest film studios in the world, is now leveraging machine learning and artificial intelligence to predict what films audiences want to see. In a recent paper, the studio has explained how they are leveraging machine vision systems to examine trailer footage frame by frame, labelling objects and events. Later, the system compares this to data that has been generated for other trailers. So, basically, the concept here is that movies with similar kind of labels will attract a similar kind of audience. The studio has created it “experimental movie attendance prediction and recommendation system” called Merlin. And in order to bring Merlin into a fully-functional state, the studio partnered with Google and used their servers and open-source AI framework TensorFlow. The system is a hybrid collaborative filtering pipeline that is enabled a fully anonymised, user privacy compliant, movie attendance dataset that combines data from different sources with hundreds of movies released over the last years, and millions of attendance records. So, for example, if we take the movie The Avengers, Merlin would be comparing this movie’s trailer data with other movie trailers data, and analyse and tries to predict what films might interest the same people who watched The Avengers. How Cinelytic Helps Studios And Film Companies Make Faster And Smarter Decisions Started in 2013, Cinelytic is one of the firms that is making the most of AI. The company is not making robots to serve in hostels or chatbots but has built a platform that can now tell you how a movie would in the box-office you replace a cast with someone else. So, all you need to do is input a cast, then swap one actor for another to see how this affects a film’s projected box office. This is kind of same a playing fantasy sports. If you are wondering how they do it, the Los Angeles-based firm uses predictive analytics. It is basically the use of data, statistical algorithms and machine learning techniques to identify the likelihood of future outcomes based on historical data. So, Cinelytic use historical data about movie performances over the years, then match it with information about films such as themes and key talent and then uses machine learning to extract hidden patterns in data. For example, if there is a movie casting Dwayne Johnson, you could use the Cinelytic’s platform and swap Dwayne Johnson with Bruce Willis and see how the movie would perform after the swap. That is not all, users can also compare both the scenarios. Wrapping Up With all the innovations and advancement, it seems AI is on a spree to prove everyone wrong who believes even this technology has its threshold. Today, it has reached a such a stage where it is already giving humans inferiority complex. And now with its increased usage in a creative sector like filmmaking, it makes one wonder about the results if an AI starts to decide the cast and crew of a particular film. Will many actors be laid off. We can definitely see that as a possibility.","excerpt":"Showbiz is one of the most lucrative industries in the world, but there’s no specific formula for churning out blockbusters. That is why filmmakers from across the world are tapping artificial intelligence to take help with audience composition, CGI and even music. 20th Century Fox’s Merlin 20th Century Fox, one of the biggest film studios […]","categories":["AI Features"],"tags":["Predictive AI","predictive analytics"],"author_name":"Harshajit Sarmah","publish_date":"2019-07-04T11:04:52","publication_year":"2019","word_count":559,"keywords":["Go","artificial intelligence","machine learning","AI","chatbots","R","RAG","analytics","TensorFlow","predictive analytics","Predictive AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","TensorFlow","RAG","chatbots","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-turns-casting-director-can-now-decide-the-perfect-blockbuster-movie-cast\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10066740,"title":"How to use t-SNE for dimensionality reduction?","content":"Dimensionality reduction is one of the important parts of unsupervised learning in data science and machine learning. This part is basically required when the dimensions of the data are very high and we are required to tell the story of the data by projecting it in a lower-dimensional space. There are various techniques for dimensionality reduction like PCA, SVD, truncatedSVD, LDA etc. t-SNA is also a technique for dimensionality reduction. In this article, we are going to discuss the t-SNA. The major points to be discussed in the article are listed below. Table of contents About t-SNEHow does t-SNE work?t-SNE for dimensionality reductionUse cases for t-SNEUsing t-SNE more effectively Let’s start by introducing t-SNE. About t-SNE t-SNE is a technique for dimensional analysis or reduction that is a short form of T-distributed Stochastic Neighbor Embedding. As the name suggests it is a nonlinear dimensionality technique that can be utilized in a scenario where the data is very high dimensional. We can also say this is a technique for visualizing high dimensional data into lower-dimensional space. For the first time, this technique was introduced by Laurens van der Maatens and Geoffrey Hinton in 2008. Its nonlinearity behaviour against data makes it different from the other techniques. Where techniques such as PCA are the linear algorithms for dimensional reduction and preserve large pairwise distance that can lead to poor visualization of high dimensional data, the t-SNE works better than PCA by preserving small pairwise distance. Are you looking for a complete repository of Python libraries used in data science, check out here. How does t-SNE work? As above mentioned it is a technique for visualizing the high dimensional data or we can say a technique for dimensionality reduction. This technique works by converting high dimensional data points to joint probabilities and uses these probabilities to minimize the Kullback-Leibler divergence so that low dimensional embeddings can be obtained. The cost function that this technique uses has a non-convex cost function which means every time we apply it we can get a different result. The proper working of t-SNE can be understood using the following steps: Firstly the algorithm of this technique first calculates the joint probabilities between the data points that represent the similarity between points. After the calculation of joint probability, it assigns the similarity between the data points on the basis of the calculated joint probability.After assigning the similarity, t-SNE represents the data points on lower dimensions on the basis of probability distribution until the minimum Kullback-Leibler divergence. Kullback-Leibler divergence can be considered as a statistical distance where it represents the calculation of how one probability distribution is different from the other one. t-SNE for dimensionality reduction In this section, we are going to look at how we can use the t-SNE practically for dimensionality reduction through implementation in python. Before implementation, we are required to know that sklearn is a library that provides the function for implementing t-SNE under the manifold package. Let’s take a look at the simple implementation. Let’s define random data using NumPy. import numpy as np X = np.array([[0, 0, 0], [0, 1, 1], [1, 0, 1], [1, 1, 1]]) X.shape Output: Here we can see the shape of the array that we have defined. Let’s apply the t-SNE on the array. from sklearn.manifold import TSNE t_sne = TSNE(n_components=2, learning_rate='auto',init='random') X_embedded= t_sne.fit_transform(X) X_embedded.shape Output: Here we can see that we have changed the shape of the defined array which means the dimension of the array is reduced.  Let’s discuss places where we can be applying t-SNE with our data. Use cases for t-SNE In the above section, we have looked at the basic implementation and the working of the t-SNE and by looking at these things we can say that the t-SNE can be applied with very high dimensional data. Although the developer of the t-SNE has mentioned it uses cases in the fields like climate research, computer security, bioinformatics, cancer research, etc. after applying this technique we can use its outcome in different supervised modelling processes. We can also use this method for clustering or separation of the data. In a variety of modelling procedures, we generally apply models to the separated data to get higher results. However, it is not a proper clustering algorithm or technique. This can also be applied to the fields where data exploration is required using the visualization of the data. Let’s take a look at the ways using which we can make the t-SNE more effective. Using t-SNE more effectively Since we use this technique to analyze the high dimensional data we’re required to make sure that we are applying t-SNE iteratively using the different parameter values to reach a proper result. There is a use of a non-convex cost function in t-SNE and it is a stochastic process using it in iteration may represent changes in the outcome that can be solved by fixing the random state parameter. t-SNE is an algorithm that can also shrink sparse data and amplify non-sparse data. To apply the algorithm it is very necessary to fix the parameters of density\/spread\/variance before applying it. Perplexity is a parameter given under the t-SNE that relates to the number of neighbours and with the larger dataset, it is required to set a larger perplexity. Final words In this article, we have discussed the t-SNE(T-distributed Stochastic Neighbor Embedding) which is a technique used for dimensionality reduction. Along with this we also discuss the working, implementation and use cases of the t-SNE which is a nonlinear dimensionality reduction technique. References Link to the codes","excerpt":"t-SNE is a nonlinear dimensionality technique that can be utilized in a scenario where the data is very high dimensional.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2022-05-11T13:00:00","publication_year":"2022","word_count":927,"keywords":["data science","NumPy","Go","machine learning","TPU","AI","Machine Learning","Python","Ray","programming_languages:Python","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","data science","Ray","NumPy","TPU","Python","R","Go","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-use-t-sne-for-dimensionality-reduction\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":17191,"title":"Artificial Intelligence Drops The First Music Album, ‘I AM AI’","content":"A collaborative album between Tayrn Southern and Amper software, I AM AI, is the first album to be entirely composed and produced by artificial intelligence. I AM AI is a collaborative effort between AI music composition software Amper and singer\/online-sensation Tayrn Southern. Southern has amassed more than 500 million views on YouTube, and she has over 450 thousand subscribers but looks like Amper will catch up with her faster than expected. “In a funny way, I have a new songwriting partner who doesn’t get tired and has this endless knowledge of music making,” Southern told CNN Tech. “But I feel like I get to own my vision; I iterate and choose what I like and don’t like. There’s a lot more control.” This is being seen as a significant development in recent times, after several incidents of musicians being assisted by artificial intelligence to produce simple melodies, arrangements and instrumentation came into light. I AM AI is different because it is the first album in which all chords, production work and instrumentation have been exclusively AI-generated. Last year, Sony’s AI music-making software Flow Machine generated its first pop song, designed in the style of The Beatles. However, unlike I AM AI’s commercially available album the first single of which, “Break Free”, the lyrics of the song were penned by a human composer, who also helped arrange the AI-generated segments of music. “Human creators and human musicians are not going away,” Drew Silverstein, CEO for Amper Music, told CNN Tech. “We’re making it so that you don’t have to spend 10,000 hours and thousands of dollars buying equipment to share and express your ideas.” Created by a team of musicians and technology experts, the development of Amper was intended to create an affordable and royalty-free way to provide the music. Currently, interested musicians can test the service out for free as part of a beta roll-out.","excerpt":"A collaborative album between Tayrn Southern and Amper software, I AM AI, is the first album to be entirely composed and produced by artificial intelligence. I AM AI is a collaborative effort between AI music composition software Amper and singer\/online-sensation Tayrn Southern. Southern has amassed more than 500 million views on YouTube, and she has […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI in music","assisted intelligence"],"author_name":"Priya Singh","publish_date":"2017-08-23T10:50:39","publication_year":"2017","word_count":315,"keywords":["AI in music","Go","AI music","artificial intelligence","programming_languages:R","AI","assisted intelligence","programming_languages:Go","CNN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","CNN","AI music","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/artificial-intelligence-drops-first-music-album-ai\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10126080,"title":"Beyond Pride Month: True Allyship Needs No Calendar","content":"Pride Month is celebrated in June to honour the 1969 Stonewall Uprising, which was a tipping point for the Gay Liberation Movement in the US. However, now that Pride Month is over, what’s next? Diversity and inclusion are hot topics right now, frequently discussed in the media as awareness of workplace inequalities grows. Although many people want to drive change by bringing a safe space for their LGBTQIA+ colleagues, there’s often uncertainty about how, or even where to start; This is where the idea of allyship in the workplace takes shape. Saurabh Bajpai, VP of Everyday Banking, NatWest Group, told AIM that there has been an increase in support for the LGBTQIA+ community in the Indian tech ecosystem regarding acceptance and allyship. Yet, he pointed out that middle and junior management levels require more mentorship and active engagement to truly foster inclusivity.  Bajpai is also the Chair, of Employee Led Network, LGBT+ India Inclusion Council. “We need more allyship coming from middle or junior-level management teams. Just ticking boxes would not help, and we need to actively engage for and with the community,” he added. But allyship does not just end here. Showing support also means respecting their identities and choosing pronouns, even beyond pride month. The Bitter Reality According to a recent report by Deloitte, six in 10 respondents identify as bisexual, 7% as gay and 4% as lesbian while 17% identify as asexual in the Indian tech sector. The numbers show that while corporates are slowly making strides towards creating a healthy workspace for LGBTQIA+ members, there is a long way to go in the struggle for acceptance and equality beyond June. Bajpai’s journey shows us that resilience and self-discovery began in Lucknow, where he was born and raised. He eventually ventured into computer science and completed his higher education at Kanpur University and SASTRA University. “Throughout this journey, I got the opportunity to express myself, celebrate my wins and failures, except there were no discussions around my identity and sexual orientation,” Bajpai, who identifies as a gay man, told AIM. Despite his professional achievements, Bajpai’s personal journey was marked by internal struggles and societal challenges regarding the same. From a young age, he knew he was homosexual but felt compelled to hide his true self due to bullying and societal expectations. This internal conflict persisted even after moving to Mumbai for job, “I started acting ‘straight’ and with that, my belief got stronger that people will love and accept me like this. But even though cities changed, the fight inside my mind about my identity was still not over” added Bajpai. Eventually, he came out to his friends, feeling relieved, and said, “This was the moment I stopped struggling in my mind. But my coming out journey hasn’t stopped since then.” Professionally, Bajpai faced challenges, particularly in dealing with preconceived notions and stereotypes in the workplace. “It takes a lot for a person to keep switching between identities. In my early career, even after pretending to be of a different sexual orientation, people still made nasty comments about me,” he said. Having said that, coming out to even one colleague is a big step, and the process should not be rushed. Identifying allies, finding strength, and becoming vocal and visible are crucial steps for self-help before seeking external support. He emphasised, “Coming out is a never-ending journey; you go to different organisations and have to go through the same cycle again. And it’s a very personal journey where one should not rush to come out.” However, his early years of horrible experiences due to sexuality shaped his leadership philosophy, which is based on respecting all employees and advocating for talent and attitude over superficial judgments. “We need to have more representation in leadership roles from the community so that they can act as role models for the rest,” added Bajpai. NatWest Group has implemented several initiatives to support the LGBTQIA+ community, including reverse mentoring sessions, inclusive policies, and infrastructural support. Its flagship program, TRANSpire, launched two years ago, focuses on providing career opportunities and support to the entire rainbow community, emphasising mental well-being and career development. Importance of DE&I in AI In December 2023, CNBC reported that big techs like Microsoft, Google, and Meta have downsized their Diversity, Equity, and Inclusion (DEI) efforts, laid off DEI staff and leaders of diverse employee groups, and downsized learning programs. This incident illuminates the bigger picture of how a lack of diversity in AI impacts product functionality and accessibility, leading to bias and inaccuracies. For example, Buzzfeed’s Midjourney-generated images of Barbies from different countries faced accusations of racism and cultural inaccuracies. Not so long ago, Google had to temporarily suspend its image-generating feature after Gemini inaccurately depicted people of colour in Nazi-era uniforms, showcasing historically inaccurate and insensitive images. German AI cognitive scientist Joscha Bach believes this bias wasn’t hardcoded, but inferred through the system’s interactions and prompts. He said that Gemini’s behaviour reflects the social processes and prompts fed into it rather than being solely algorithmic and the model developed opinions and biases based on the input it received, even generating arguments to support its stance on controversial issues like meat-eating or antinatalism. Such behaviours of AI models are nothing but mirrors of society, urging a deeper understanding of our societal condition. Therefore, it is crucial to treat everyone equally and respect each other’s choices and opinions, ensuring that such biases are not ingrained in AI systems.","excerpt":"Middle and junior management levels require more mentorship and active engagement to truly foster inclusivity.","categories":["AI Features"],"tags":["diversity","inclusivity","Interviews and Discussions","pride"],"author_name":"Shritama Saha","publish_date":"2024-07-07T14:11:38","publication_year":"2024","word_count":908,"keywords":["Go","pride","ELT","programming_languages:R","AI","programming_languages:Go","diversity","Aim","llm_models:Gemini","ViT","inclusivity","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","Aim","R","Go","ELT","GAN","ViT","llm_models:Gemini","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/beyond-pride-month-true-allyship-needs-no-calendar\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":22375,"title":"Beautiful Soup Primer: How To Scrape Data From A Website","content":"Beautiful Soup is a HTML and XML parser available on Python 2.6+. Soup is named after the unstructured HTML documents which are hard to understand and noisy. It parses the data from the HTML and XML documents from where it can be extracted. In this article, we will be going through functions which help us extract data from the HTML document. We will be using a toy HTML to explain how Beautiful Soup works and walk through the steps involved in Scraping — one of the techniques of data mining — data from a website’s HTML format. With the help of headless browsers such as Selenium and PhanthomJS, one can easily practice how to scrape data out of a website. With these browsers, it will be easy to scrape through multiple pages or extract a large amount of data from the websites. Using a headless browser will also increase the computation speed which will result in the consumption of less memory. In fact, PhanthomJS assigns unique processes to each browser as well. Installing Beautiful Soup 4 Beautiful Soup library can be installed using PIP with a very simple command. It is available on almost all platforms. Here is a way to install it using Jupyter Notebook. We can import this library with the following code and assign it to an object. Installing An Alternative Parser Beautiful Soup has a default parser available in the standard Python Library. We can use a different parser depending on the objective. The most common alternative parsers are “lxml” and “html5lib”. This can be installed with the help of the following code: Below is a tabular representation of various parsers with their advantages and disadvantages: crummy.com Getting Started We will be using this basic, and default, HTML doc to parse the data using Beautiful Soup. The following code will expand HTML into its hierarchy: Exploring The Parse Tree To navigate through the tree, we can use the following commands: Beautiful Soup has many attributes which can be accessed and edited. This extracted parsed data can be saved onto a text file. To extract the text from the string, we can use the get_text() command. Strings: How To Remove White spaces The string can be accessed using the strings command. But it also includes white space which can be stripped easily. Since the above output has a lot of white space, the striped.strings command will help us remove it. Parent And Siblings We can obtain the parent of a particular HTML with .parent attribute, like here: To access the siblings — previous as well as the next — we can use the following commands: Find And FindAll This function is used to search for a very particular field throughout the HTML document. It is one of the key features required while data mining or scraping a data from a website with the help of Selenium and PhanthomJS. Conclusion Since finding the right tags from the HTML source is hard, scraping the data takes a lot of time. It can also depend on the amount of data extracted from a page. That is why, the wait time is necessary for the browser to load the data. Depending on their computation speed and availability of resources one can scrape data from almost any website using the right tools.","excerpt":"Beautiful Soup is a HTML and XML parser available on Python 2.6+. Soup is named after the unstructured HTML documents which are hard to understand and noisy. It parses the data from the HTML and XML documents from where it can be extracted. In this article, we will be going through functions which help us […]","categories":["AI Features"],"tags":["Data Mining","data scraping"],"author_name":"Kishan Maladkar","publish_date":"2018-03-07T12:21:07","publication_year":"2018","word_count":550,"keywords":["Go","API","TPU","programming_languages:R","AI","data scraping","Data Mining","ML","Python","programming_languages:Python","Jupyter","R"],"extracted_tech_keywords":["AI","ML","Jupyter","TPU","Python","R","Go","API","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/beautiful-soup-primer-how-to-scrape-data-from-a-website\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10086025,"title":"The Job Role That Even ChatGPT Can&#8217;t Touch","content":"In a world where AI-generative models like ChatGPT are becoming increasingly prevalent, it can be easy to imagine a future where robots take over all our jobs. But fear not, for there is one profession that has proven to be a true champion in the field of AI—UX Design. This role has adapted and evolved with the technology, and continues to play a vital role in the development and advancement of AI-generative models. Tug of war Google’s Francoise Chollet leads an interesting discussion on what ChatGPT-like AI assistance model offers to the world. AI assistance has a bright future — applications where humans can guide a search\/generation process and correct its output as needed. But keep in mind that's not AI autonomy — AI agents capable of action on their own in the real world beyond a very narrow scope of automation…— François Chollet (@fchollet) January 25, 2023 Like many in the field of deep learning, Chollet believes that achieving true autonomy in AI is an incredibly difficult task, and present techniques fall far short of that goal. Additionally, current approaches take a detour from the direction of autonomy. Regardless, Chollet acknowledges that AI assistance can still bring significant value to society, only that the challenge will lie more in designing effective human–AI interactions rather than advancing the technology itself. Simply put, he predicts that the role of “AI UX designer” will become increasingly important in the future as the task will be to find novel and efficient ways for humans to interact with data manifolds, or what is more commonly understood as deep learning models. Further, natural language dialog will only be just one modality out of many, considering research institutes like Meta and Open AI are already working in the direction of building multimodal neural networks. As Chollet aptly puts, “Data manifolds will become a tech commodity”. Semianalysis’ Dylan Patel also shares a similar viewpoint in the following tweet: A chinchilla\/GPT3 model is a commodity. A larger model is also a commodity. Question is how much they were able to prep and clean the data and the UX. I am somewhat convinced hyperscalers don't understand the last problem.— Dylan Patel (@dylan522p) December 3, 2022 For Patel, hyperscalers like Google, Microsoft, Amazon, and Oracle, among many others, have lost sight of the end goal. This can be validated by the fact that the rise of UX in AI—owing to its commodification—is already a part of the research focus in several organisations. “UX research is essential for a large language model. It is still a nascent stage. Currently humans are adopting UX but we should create a system where UX will adopt humans. That will be the ultimate stage of UX for large language model,” Soumen Ray, Head of Analytics at Hindustan Coca-Cola Beverages, told AIM. UX x LaMDA The incident with Google’s LaMDA model, where the public became concerned over the model being referred to as “sentient” and capable of feeling emotions, accurately captures the fear that humans have of AI becoming more advanced than them. Interestingly, here as well UX research can play an influential role in building the future of LaMDA. “UX research has many of the answers on how to train these engines, not only in how LaMDA can analyse user questions, but also how it can problem-solve with the user with follow-up inquiries,” writes Anil Tilbe, who currently leads AI + Product for the U.S. Government. Tilbe essentially argues that the UX research will help understand how to interview and understand users. An example he provides is that if the model is not able to understand a question posed by the user, or wants the user to reframe their questions for not being able to understand earlier. In such a case, it would be able to follow up with the user if it trained on how to ask questions. Tilbe proposes that LaMDA should be educated on how to socialise in terms of employing design thinking to understand the varying user needs by conducting interviews. We also see a significant difference when it comes to the user research performed by a model like LaMDA as opposed to say, ChatGPT. In the above example, we can see that while LaMDA tries to connect at a more human level by using elements of human interaction in its output generation, ChatGPT seems to be just rendering output scraping content from the web, without understanding what constitutes genuine human interaction. Final thoughts Deep learning researchers have been disillusioning people of the groundbreaking capability of models like ChatGPT, which is touted as “disruptive and revolutionary” by many. According to them, while large language models (LLMs) have been successful in providing various AI-assisted applications, it comes at the cost of Artificial General Intelligence (AGI). Researchers argue that the current state of LLMs is not advanced enough to support AGI. While the deviation from AGI seems treacherous to many within the community, it almost seems like a necessity considering that it is human nature to be cautious of any new technology that potentially threatens their special place on the planet. User research will ensure that the AI is designed for the people that will use it, instead of just for the technology itself.","excerpt":"User research will ensure that the AI is designed for the people that will use it, instead of just for the technology itself.","categories":["AI Features"],"tags":["ChatGPT"],"author_name":"Ayush Jain","publish_date":"2023-01-27T18:29:46","publication_year":"2023","word_count":871,"keywords":["ChatGPT","TPU","AI","neural network","RAG","Ray","Aim","deep learning","analytics","R"],"extracted_tech_keywords":["AI","deep learning","neural network","analytics","ChatGPT","Aim","Ray","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-job-role-that-even-chatgpt-cant-touch\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10093210,"title":"Council Post: Mastering the AI Maze: Insider Strategies for Tackling Implementation Challenges","content":"As the emergence of Dall-E and Chat-GPT brings Artificial Intelligence to the limelight, it’s no surprise that companies are looking for creative ways to make use of this limitless power. AI isn’t simply a technology for specialists anymore – businesses can now fully leverage its advantages regardless of their technical skill level. 86% of business leaders see AI as an essential part of their daily operations, and they’re reaping the rewards in terms of increased efficiency and productivity. AI empowers decision makers with lightning-fast speed and accuracy, eliminating tedious manual data entry tasks along the way. With this level of intelligence on your side, unparalleled market opportunities will be within reach. As McKinsey’s global survey indicates, 50% of businesses have already integrated AI in at least one domain, with the adoption rate predicted to double in upcoming years. However only 11% of organisations leveraging AI have experienced notable ROI. Why is this so? As AI Advances, So Does the Role of Data Engineers and Scientists Picture this— just a few years ago, data engineers and scientists were merely responsible for maintaining databases and extracting useful insights from gathered business intelligence, but as technology advanced and AI solutions became a rising priority, a storm began brewing, leaving them completely blindsided. CTOs, data engineers, and scientists are facing the challenge of keeping up with the constantly changing technology landscape, which involves exploring and managing new algorithms, architectures, and solutions in both open-source and industry domains. Meanwhile, they are also grappling with the need to master best practices.The world of data has transformed into a fierce beast that is difficult to tame, and it is catching everyone off guard. Big universities and training camps have yet to catch up to the tech, leaving data engineers and scientists with no choice but to piece together knowledge from scratch and learn through trial and error. Data scientists need to access a specific table to build predictive AI models.But it doesn’t stop there. They will still face challenges such as figuring out the meaning of column names and dealing with missing values. These challenges can cause delays and confusion in the modelling process. The role of a data scientist has turned into an adrenaline-pumping race against time and the need to always be one step ahead in a field that is expanding towards uncharted territories. Data science is morphing into a new hybrid role, merging with technological advancement to shape the industry. In this article, I will explore three common challenges that enterprises face when implementing AI while providing practical solutions. 1- Challenge: Data Shortage Machine-learning models often require vast, diverse datasets to function optimally, but obtaining such data remains a challenge. Insufficient data can lead to overfitting and biassed models eventually leading to poor performance. Data shortages result from data privacy constraints, a lack of historical data for new trends, small sample sizes, and the costly, time-consuming nature of manual data labelling. Acquiring permission to use sensitive data is challenging, and historical data often proves inadequate due to evolving trends and processes. However, adequate training data is essential for creating unbiased and accurate ML models capable of addressing a wide range of scenarios. Solution: Synthetic Data Generation Synthetic data many times serves as a viable solution to help overcome this challenge. When generated artificially to resemble real datasets, synthetic data can reveal hidden patterns, interactions, and correlations between variables, offering a substantial foundation for ML models. Balancing datasets and improving performance, it can supplement marginal classes without jeopardising privacy. Advancements in synthetic data have significantly increased its value for machine learning models. Techniques like generative adversarial networks (GANs) and Wasserstein GANs (WGANs) foster the creation of more realistic data while maintaining compliance and data balance. We used synthetic data to develop an AI tool that could monitor the brand uniformity of posters outside automobile showrooms in India. By using CycleGAN, we generated data for both brand compliant and non-complaint cases, allowing us to successfully train and deploy the AI model. 2- Challenge: Talent Shortage 75% of decision makers prefer to help upskill current staff and 64% favor recruiting experts to bridge the AI talent gap. However, budget limitations and retention challenges are significant barriers to these solutions. Solution: Low Code and No-Code Platforms Enterprises are increasingly embracing low-code\/no-code platforms to democratise app development and ease workloads. Faced with an estimated 85.2 million global software engineer deficit by 2030, businesses have found that low-code\/no-code tools can increase their value by millions of dollars without hiring additional IT developers. For instance, our no-code platform streamlines the entire process of AI model training for pharmaceutical clients, including feature selection and deployment. What used to take a team of data scientists 2 months now just takes 1 week. Gartner has projected that the low-code technology market is set to reach $44.5 billion by 2026. This is due to a number of factors, such as an increasing demand for more rapid application delivery, persistent talent shortages, and the proliferation of hybrid workforces. Low-code platforms have become an essential element of successful hyper automation, with 50% of all new clients expected to come from business buyers outside the IT organisation by 2025. However, to achieve their full market potential, low-code\/no-code platforms must continually innovate in areas like real-time iteration, DEVops workflows integration, scalability and API development. 3- Challenge: High expenditure AI investments have reached nearly $118 billion in 2022 and surpass $300 billion in the next few years. The cost of AI varies, with companies paying anywhere from $6,000 to over $300,000 for custom solutions. Few factors that influence the costs include the type of AI software (chatbots, analysis systems, or virtual assistants), whether the enterprise requires a pre-built or custom solution, additional features, and how the platform will be managed (in-house or outsourced). Project duration and complexity also impact the overall cost. Solution: Optimising Deployment Techniques In addressing the pressing issue of high costs associated with AI implementation for enterprises, it becomes crucial to optimise data processing methods applicable to multiple forms of data, including videos, text, and other pertinent information depending on size and nature. By adopting strategic and efficient deployment techniques, it is possible to achieve substantial cost reduction. Using specialised hardware for video processing and taking advantage of edge computing solutions can present a more economical option compared to solely relying on cloud-based solutions. Taking it a step further, a hybrid deployment can be implemented based on the specific use case and scenario, with a portion of the solution deployed on a dedicated server and the remaining on the cloud. This multifaceted approach not only simplifies data management processes but also significantly reduces the overall expenses involved in integrating AI in today’s business environment. Similarly, if you opt for an AI API, you’ll be charged an invoice for every digitization process. However, with open source or commercial models, you can deploy the model without the need for API-based payments, resulting in long-term cost savings. Developing and deploying customised models for businesses is ultimately a more cost-effective option than relying on AI APIs. Final Thoughts Organisations looking to take advantage of AI need an intentional and methodical approach that combines user interfaces, regulations, data infrastructure, data storage solutions and labelled datasets. Once these systems are in place, enterprise-focused machine learning algorithms can be trained with structured and unstructured datasets. Ultimately, successful adoption of AI results in operational efficiencies as well as critical insights that can foster growth and success for any organisation willing to embrace this technology. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"Organisations looking to take advantage of AI need an intentional and methodical approach that combines user interfaces, regulations, data infrastructure, data storage solutions and labelled datasets.","categories":["AI Features"],"tags":["synthetic data generation"],"author_name":"Rahul Thota","publish_date":"2023-05-12T16:00:00","publication_year":"2023","word_count":1293,"keywords":["data science","machine learning","artificial intelligence","AI","chatbots","ML","virtual assistants","RAG","Aim","analytics","synthetic data generation"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","RAG","chatbots","virtual assistants"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mastering-the-ai-maze-insider-strategies-for-tackling-implementation-challenges\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10079052,"title":"Top 3 Reasons To Ditch Twitter, And Join Tumblr","content":"Following Elon Musk’s $44 billion Twitter acquisition deal, the unhappy users are now relentlessly looking for alternatives to the platform. With each passing day, uncertainty clouds Twitter as its new chief Elon Musk looks to bring in changes in subscriptions, content monetisation, and whatnot. In the last few weeks, Twitter was flooded with countless announcements from people making preparations to leave the platform or bidding goodbye forever. Hashtags such as #GoodbyeTwitter, #Mastodon, and #TwitterMigration were trending on the platform. According to a user counting bot, the decentralised and open source social network ‘Mastodon’ has gained over 100,000 users following the announcement. Ever since, users are reminiscing about one platform that was particularly popular in the early 2010s—‘Tumblr’. Founded by David Karp in 2007, Tumblr is a microblogging platform which allows users to share images, music, video, quotes, links, GIFs, and other forms of multimedia, more or less similar to Twitter. As of July 2021, the networking service has hosted over 529 million blogs. The platform witnessed peak popularity during 2012–2015 but found it difficult to recreate this popularity, partly due to the rise of other social media platforms such as TikTok. Read: Nothing like TikTok, Kudos for Trying Here are a few reasons why you should ditch Twitter and join Tumblr. GIFs, GIFs, and more GIFs Among the main features of Tumblr are the countless funny and interactive GIFs. Users can either press the pause button or  watch the single looping of the GIF over and over. In 2015, Tumblr launched “GIF Maker” for creating and sharing GIFs. “So many of the photos you create on your iPhone today are ready to be GIF’ed,” said founder David Karp. Tumblr got Crabs 🦀 Source: Tumblr On April 1, 2022 the platform released an update to the website which added a new feature. Not in literal terms, but ‘crabs’ are little dashboard widgets on Tumblr feeds. One of the platform’s widgets generates a crab each time a user clicks on it.  When pressed, the button would generate numerous pixelated crabs moving around the user’s screen. Posting Images Like every other platform, users will be allowed to add pictures on the dashboard. Here’s what Tumblr says: Reason to Join Tumblr #8:4 Images in a post? No.10 Images in a post? No. You can post up to 30 images in a single post.— tumblr dot com the website and app (@tumblr) November 6, 2022 Moreover, the social media platform has reversed its nudity ban four years after announcing restrictions on explicit content. The company in a blog post said,“We now welcome a broader range of expression, creativity, and art on Tumblr, including content depicting the human form (yes, that includes the naked human form).”","excerpt":"If users don’t want to see algorithmic recommendations on their feed, they can just turn it off.","categories":["AI News"],"tags":["Mastodon","tiktok","Top Trend","Twitter (X)"],"author_name":"Bhuvana Kamath","publish_date":"2022-11-08T16:10:22","publication_year":"2022","word_count":450,"keywords":["tiktok","Mastodon","Top Trend","Go","programming_languages:R","AI","programming_languages:Go","ViT","Twitter (X)","R"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/top-3-reasons-to-ditch-twitter-and-join-tumblr\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52223,"title":"Why DataRobot Acquired Paxata","content":"DataRobot is clearly on an acquiring spree. In an endeavour to allow firms to harness the power of artificial intelligence, DataRobot acquired California-based Paxata earlier this month. With this, the firm has now completed three acquisitions in 2019 to enhance its offering and allow their customers to manage their machine learning workflows effectively. Built to automate machine learning, DataRobot is leaving no stone unturned in delivering superior solutions to firms for driving their business by integrating the latest technologies such as AI and data science. Since its inception in 2012, the firm has acquired five startups that were making new advancements in allowing IT firms to automate the machine learning activities. DataRobot Acquisitions: Nutonian – 2017Nexosis – 2018Cursor – 2019ParallelM – 2019 Paxata – 2019 As companies are looking to imbibe intelligence, many AutoML providers such as H20.ai, Azure AutoML, among others, are striving to take the lead in the landscape. Thus, DataRobot is quickly acquiring various companies to deliver a solution that will streamline the complete AI workflows. Acquisitions In 2017, DataRobot acquired Nutonian, an AI-powered modelling engine that specialises in time-series analytics modelling. The first acquisition allowed it to provide predictive analytics for firms that highly rely on time-series analytics. And to further enhance its capabilities, in 2018, DataRobot acquires Nexosis, while keeping the terms of acquisition confidential. The motive with this was to provide ML at the hands of everyone with the enterprise-grade platform. Further, the firm acquired Cursor and ParallelM in 2019 to enable data collaboration from different sources and assist companies in scaling the development, governance, management of ML-based solutions in production, respectively. After integrating its platform with numerous capabilities through acquisition, it extended its AutoML solution’s dexterity and allowed firms to do more than just model selection. But a lack of functionality to prepare datasets for training the models caused hindrance in AI workflows with its solution. Manual data analysis can never match up the speed at which firms collect the data, thereby, the need for automation in AI workflows is of paramount importance. Therefore, it was natural for DataRobot to acquire Paxata that could empower companies to make high-quality data for ML models. Paxata would also provide its AutoML solutions to help firms quickly deploy machine learning models for obtaining informed insights into the information they collect. Challenges Faced By DataRobot With a little more than 1,000 employees, DataRobot is an AutoML provider that offers its solutions in more than 12 countries. Today, AutoML is being adopted by different companies to train high-quality ML models specific to their business needs, even without data science experts. As per a report, around 86% of the companies are willing to integrate AutoML next year. Such trends are helping DataRobot to gain clients across different countries. However, numerous limitations in AutoML makes it difficult for companies to be confident about the products they offer. For one, preparing high-quality data for feeding into AutoML is a strenuous task for data scientists. “We repeatedly received feedback from our customers that they wanted us to deliver additional capability, which was to simplify how they prepare data required for AI,” mentioned DataRobot in the announcement. Solution To mitigate such challenges, DataRobot acquired Paxata – a self-service data integration and management firm – that enables companies to transform data into information quickly. Using Paxata, developers can efficiently perform data cleaning and build datasets that can be fed to the AutoML solution of DataRobot. “Two companies were a great fit, and it made a lot of sense to come together, it will allow our customers to go from data to value quickly,” says Prakash Nanduri, CEO and co-founder at Paxata. According to leading research and advisory firm, hyper-automation was one of the top strategic technology trends for 2020. This demonstrates the importance of the need for automating the AI activities in coming years. The integration of Paxata’s solution with DataRobot’s AutoML will allow them to provide end-to-end automation products to firms. “To resolve pressing problems and deliver value through AI to the enterprise, DataRobot is focused on building an enterprise AI platform that provides automation for gaining ROI from raw data,” mentioned Igor Taber, SVP of corporate development and strategy at DataRobot in the announcement. Outlook Unlike other AutoML companies, DataRobot is committed to enhancing its platform not only to carry out model selections and hyperparameter optimisation but also the collection and preprocessing of data for becoming a one-stop-shop for AI-based activities. This will decrease the dependency of other applications for deploying ML models. However, we are way behind in streamlining the entire workflow with one platform, but DataRobot is arduously trying to accomplish that. Working on the same objective, DataFlow is also raising funds and has now raised more than $431 million, of which $206 was raised in Series E funding in September. The firm is further poised to gain momentum in the AutoML landscape and democratise AI technology.","excerpt":"DataRobot is clearly on an acquiring spree. In an endeavour to allow firms to harness the power of artificial intelligence, DataRobot acquired California-based Paxata earlier this month. With this, the firm has now completed three acquisitions in 2019 to enhance its offering and allow their customers to manage their machine learning workflows effectively. Built to […]","categories":["AI Features"],"tags":["Automl","data analytics acquisitions","Data Cleaning","Data Governance strategy","DataRobot","Mergers and Acquisitions"],"author_name":"Rohit Yadav","publish_date":"2019-12-18T14:00:00","publication_year":"2019","word_count":817,"keywords":["data science","Automl","Go","artificial intelligence","machine learning","AI","Azure","R","ML","data analytics acquisitions","Data Cleaning","DataRobot","analytics","Mergers and Acquisitions","predictive analytics","Data Governance strategy"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","predictive analytics","Azure","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-datarobot-acquired-paxata\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10123414,"title":"Databricks Open Sources Unity Catalog for Data and AI Governance Across Platforms","content":"Databricks announced it is open-sourcing its Unity Catalog, the industry’s first unified governance solution for data and artificial intelligence (AI) that works across clouds and data platforms. By open-sourcing Unity Catalog, the company aims to establish an open standard for interoperable data and AI governance. Unity Catalog OSS offers a universal interface supporting multiple data formats and compute engines. It enables unified governance across tabular data, unstructured data, and AI assets like machine learning models. With open APIs and an Apache 2.0 licensed open-source server, it provides flexibility and avoids vendor lock-in. “Our customers love Unity Catalog because it streamlines data access and governance at scale,” said Ali Ghodsi, CEO at Databricks. “We’re excited to open source Unity Catalog to drive the industry forward to an open standard for data and AI governance that gives customers openness and flexibility.” Key features of Unity Catalog OSS include:Interoperability across data formats, compute engines, and platforms.Unified governance for all data and AI assetsOpen architecture to maximise customer flexibility and choice Several partners, including AWS, Google Cloud, Microsoft, Salesforce, Confluent, dbt Labs, Immuta, Informatica, and Unstructured, expressed support for Unity Catalog OSS. They praised Databricks’ move as enabling greater customer flexibility and aligning with open ecosystem principles. “AWS welcomes Databricks’ move to open source Unity Catalog. AWS is committed to working with the industry on open-source solutions that enable choice and interoperability for customers,” said Chris Grusz, managing director of technology partnerships at AWS. Customers like AT&T, Nasdaq, and Rivian also welcomed the news, stating it will help eliminate data silos, scale platforms, and enable working across data without vendor lock-in concerns. Unity Catalog OSS will be available in public preview in Q3 2024. To learn more, visit the Databricks website or attend the Data + AI Summit on June 26-29. Click here to watch the keynote. Databricks helps organisations take control of their data with its unified data and AI platform, which is used by over 10,000 customers. Headquartered in San Francisco, Databricks was founded by the original creators of Apache Spark, Delta Lake, and MLflow.","excerpt":"“We’re excited to open source Unity Catalog to drive the industry forward to an open standard for data and AI governance that gives customers openness and flexibility,” says Databricks chief Ali Ghodsi.","categories":["AI News"],"tags":["Databricks","Open Source AI"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-12T19:06:34","publication_year":"2024","word_count":343,"keywords":["machine learning","artificial intelligence","AWS","AI","ML","Apache Spark","Open Source AI","Aim","MLflow","R","Databricks"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","MLflow","Aim","AWS","Apache Spark","Databricks","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/databricks-open-sources-unity-catalog-for-data-and-ai-governance-across-platforms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10141828,"title":"Hey Siri, You There?","content":"Apple is at it again. Its latest research significantly enhances speech recognition capabilities and reduces resource consumption while retrieving information from a large database. A challenge in using ASR (automatic speech recognition) systems is identifying rare and user-specific terms. To tackle this challenge, the model uses NCB (neural context biasing) to retrieve information from external, user-provided databases, improving speech recognition. However, processing large amounts of external data in NCB demands massive computational and memory resources. Apple is set to solve this problem in two stages, marking its foray into the world of quantisation. The research uses vector quantisation to shortlist biasing entries only relevant to the input. “Quantisation-based technique allows the ASR model to leverage more biasing entries that would otherwise be discarded due to excessive compute and memory cost,” the authors said. The model then uses a ‘cross attention’ mechanism to apply the selected biases to improve speech recognition. “We proposed an efficient approximation to cross-attention that uses vector quantisation techniques, allowing us to ground large biasing catalogues quickly on audio data with limited memory footprint,” said the researchers. The results revealed a 20% reduction in computational time and, most importantly, a 71% error reduction rate was observed. The authors also mentioned that the technique processes millions of entries in biasing databases without significant degradation in the recognition quality. No More Excuses? Apple took its time to join the AI bandwagon, yet Apple Intelligence wasn’t received well. It was available only on iPhone 15 Pro and later models, which meant that last year’s iPhone 15 would miss out on it. The company cited the lack of RAM on older iPhones as a barrier to Apple Intelligence. This was also because Apple was insistent on running AI features on devices and local hardware. Popular insider Mark Gurman reported that Apple admitted failing to achieve the expected performance levels with older iPhones with lower RAM capacity. The only other option was to offload all tasks online, which would go against its privacy promise. “You could, in theory, run these models on a very old device, but it would be so slow that it would not be useful,” said Apple’s AI\/machine learning head John Giannandrea. Therefore, this latest research seems promising. Apart from reducing compute time and error rates, the authors revealed an 85-95% reduction in memory usage. It would be premature to speculate if this indicates Apple’s plans to integrate Apple Intelligence into older iPhones, but it does strengthen their commitment to enhancing AI and keeping it on-device in future devices as well. Moreover, voice assistants, including Siri, are far from perfect. Earlier this year, a survey mentioned that Siri experienced difficulty in recognising specific accents from certain regions in the USA. Even with the release of Apple Intelligence, some users reported that Siri was struggling to recognise even the easiest of phrases. “It’s funny to see this post now. Five minutes ago, I tried to use dictation to say, ‘upload speeds are slower but otherwise it’s working’, but what my phone heard was ‘happy birthday’,” said a user on Reddit. A few users also mentioned that they haven’t noticed any significant changes in Siri, despite more updates being launched to Apple Intelligence. Moreover, most of the feedback on Apple Intelligence is received from the Beta versions. While it’s fair to notice discrepancies, there’s certainly room for improvement, given the capabilities shown by the research. Not the First Time This isn’t the first time Apple has explored techniques to improve speech recognition. Earlier this year, in May, Apple introduced a denoising model called DenoisingLM (DLM). It is an error correction model in which a TTS system generates a hypothetical, noisy environment for ASR, which is then paired with the original text to act as training data. DLM achieved a favourable, low WER (word error rate) of under 3% on most benchmarks. DLM can also be applied to more diverse datasets, which further improves the accuracy and performance of ASR. Moreover, Apple also used quantisation, among many other optimisation techniques on their on-device models. This helped reduce the latency to 0.6 milliseconds per prompt token and, increase the token generation rate to 30 tokens per second. That said, there’s also a fair critique of ‘aggressive’ quantisation methods. A recent research explored the limitations of these methods. “Despite the popularity of LLM quantisation for inference acceleration, significant uncertainty remains regarding the accuracy-performance trade-offs,” it said. After evaluating all techniques, the authors said that W8A8FP (8-bit floating-point weights and activations) is the most effective technique for achieving ‘near lossless accuracy’. However, Apple’s approach to quantisation in improving ASR uses FSQ (Finite Scaling Quantisation). This method mostly offers raw efficiency gains, particularly in retrieval-heavy tasks, whereas W8A8FP is mostly suitable for broader, general-purpose use cases. A Little Too Late? Apple definitely isn’t alone in pioneering research and development on powerful ASR systems. While OpenAI’s Whisper and Massively Multilingual Speech AI research models were released a few years ago, some of the latest improvements to ASR also come from Huawei, AssemblyAI and Moonshine. Huawei’s Hard-Synth uses advanced text-to-speech and LLMs for data augmentation and creates ‘challenging’ audio samples to improve an ASR’s performance while reducing biases in gender and speaker recognition. Even Assembly AI’s latest Universal-2 focuses on improving proper noun recognition by 24% over its predecessor, the Universal 1. It also enhances transcription features like precise time stamps and formatting accuracy. On the other hand, Moonshine developed a low-latency ASR system targeted at short audio segments and for local, on-device deployments. Moonshine’s model outperformed OpenAI’s market-leading Whisper while maintaining strong accuracy in shorter sequences. It also demonstrated a five-fold reduction in computing demands compared to OpenAI’s lightweight Whisper Tiny.en model.","excerpt":"New research suggests an improved Siri ASR system, optimised for low resource consumption and suitable for on-device AI.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Apple","Siri","Speech Recognition"],"author_name":"Supreeth Koundinya","publish_date":"2024-11-28T10:38:05","publication_year":"2024","word_count":945,"keywords":["Go","data augmentation","machine learning","OpenAI","AI","programming_languages:R","Apple","RAG","Ray","Speech Recognition","Siri","AI research","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","OpenAI","Ray","RAG","R","Go","data augmentation","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hey-siri-you-there\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":50562,"title":"Machine Learning Now Shows How Music Influences Human Experience","content":"Machine learning today not only recommends the things you can buy, or content you can watch, but it is doing wonders in other domains as well. This time it is on a mission to find out something untouched. There are different elements in music that trigger emotion in humans. And machine learning is trying to find out just that — how music affects brain activity, physiological response, and human-reported behaviour. Music & Emotions New research by scholars from the University of Southern California is trying to figure out the elements in a song that triggers different emotions in a human. To be more specific, the project has been carried out considering things like dynamics, timbre, harmony, rhythm, and register. They are also trying to figure out how machine learning could use elements and relate to emotions to predict how people might respond to a new piece of music. “This work adds to our understanding of how music affects multimodal human experience and has applications in affective computing, music emotion recognition, neuroscience, and music information retrieval,” researchers stated in their paper. This is the process that the researchers used: Using Human Efforts They first gathered songs from music streaming sites considering that they have very few plays, and are either tagged with “happy” or “sad”Then they sorted the songs (60 pieces for each emotion) and it was done with the help ofFurthermore, they created three more groups where two groups had triggers sadness and one group had only songs that induced happiness.The next thing they did is invited 100 people to listen to those songs from the three groups. And once they were done, they had to fMRI scan. They even had to wear sensors to track pulse, heat, and electricity in order to rate the emotions. Using Machine Learning Once all the data was collected by human testers, they fed that data to a machine learning algorithm.That is not all, along with that data, they also fed the machine with features of a song such as pitch, rhythm, harmony, etc. The early results of this project has shown how in the future machine learning would be able to do so much more in the music and art domain. Despite the fact that it is already doing wonders, the researchers are saying there is still a lot to explore. The project is still in the early stage and it would take a little more time to come to a solid conclusion. However, they are also optimistic that once it is done and successful, machine learning would not only help you select your bedtime or gym playlist but would also help movie makers, therapists etc. to come up with targeted musical experiences. Outlook This is not the first time when machine learning has been used in the music industry, in recent times we have seen machine learning making heavy metal music, painting portraits, and even beatboxing. And with time this sought after tech is just making things easier for the human race. Despite all this, there are people who would argue that these advancements of technology would soon take up all the jobs of the artists. There would be a time when making music would be so much easier, that the need to have an artist won’t be there. But that’s not the complete story, there are professionals from the industry who believe that these technologies would only empower them to be the best version of themselves. It would make some of the complicated things easier for artists.","excerpt":"Machine learning today not only recommends the things you can buy, or content you can watch, but it is doing wonders in other domains as well. This time it is on a mission to find out something untouched. There are different elements in music that trigger emotion in humans. And machine learning is trying to […]","categories":["AI Features"],"tags":["Machine Learning"],"author_name":"Harshajit Sarmah","publish_date":"2019-11-25T13:00:00","publication_year":"2019","word_count":586,"keywords":["Go","API","machine learning","programming_languages:R","AI","Modal","Machine Learning","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","machine learning","R","Go","API","ViT","Modal","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/machine-learning-now-shows-how-music-influences-human-experience\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10098135,"title":"Many A Generative AI Tricks Up Amazon’s Sleeves","content":"On Thursday’s Q2 2023 earnings call, Amazon CEO Andy Jassy revealed that “Every single one” of Amazon’s businesses has “multiple generative AI initiatives going right now”. Founded as a retail store company, Amazon has been more focused on developing other AI-based products than its flagship e-commerce search in the recent past. Over the past few months, the company has made several announcements and investments to accelerate through generative AI. During the AWS Summit in New York last week, the company launched Agents for Bedrock, for companies to use their own data to build AI apps that can automatically do tasks on their own foundational models, like image-to-text models or large language models, instead of just telling them about it. AI agents are the assistants which can actually do the job instead of just giving suggestions. For instance, an agent will book a flight instead of just giving a list of the cheapest available options. Other AWS announcements include bringing generative AI to healthcare through AWS HealthScribe and a partnership with Nvidia allowing AWS to handle larger amounts of memory and data by using Nvidia H100 Tensor Core GPUs. Last month, the Bezos-owned company also made an announcement to invest $100 million in a AWS Generative AI Innovation Center, to help customers build and deploy generative AI solutions. “Alexa, Reboot Yourself” While the company has been internally transforming its AI force, they have also been hiring for Alexa AI from candidates with background in dialogue systems or information retrieval. Earlier this year in May, based on a leaked document titled “Alexa LLM Entertainment Use Cases,” the big tech plans to reboot Alexa the voice assistant with ChatGPT-like features. An internal memo described a goal of making the voice assistant smarter, stating users should feel “like Alexa is thinking vs. fetching from a database.” These job posts clearly state that Amazon intends to overhaul its venerable voice assistant search feature. The current SVP of Amazon Devices & Services David Limp announced last month on LinkedIn that it will be hosting an event to presumably launch new devices on September 20th at the new HQ2 in Arlington, VA. The partnership between HuggingFace and AWS gives further confidence that Amazon has something up its sleeve to boost users’ conversational experience with Alexa. Looking out for Search While the company’s voice assistant took Silicon Valley by a storm, the company’s bread and butter since day one has been e-commerce. Even though the retail giant’s search experience has been long criticised for ads bombarding the result page, the company is finally concerned about the feature. Earlier this year, the company had also posted job listings for a machine learning-focused engineer describing how it is “reimagining Amazon Search” with a new “interactive conversational experience” to answer product questions, compare products, personalise suggestions, and more. In June, the retail giant began testing a feature in its shopping app that uses AI to summarise reviews left by customers on some products. It provides a brief summary of what users liked and disliked about the product, along with a disclaimer that the overview is “AI-generated from the text of customer reviews.” The recent frenzy around generative AI and chatbots like OpenAI’s ChatGPT, Google’s Bard and Microsoft’s Bing has pushed Amazon to sharpen its focus on the technology. The company which has been able to sell anything and everything since day one is now more involved in the AI field than ever. Across all verticals from Alexa to Code Whisperer, it is trying to step up its game. In the big picture, the IT giant is hustling with its generative AI agenda to keep up with its rivals’ efforts to hold the users’ attention.","excerpt":"Amazon has been more focused on developing other AI-based products than its flagship e-commerce search in the recent past.","categories":["AI Highlights"],"tags":["Amazon Alexa","Andy Jassy","AWS","Generative AI"],"author_name":"Tasmia Ansari","publish_date":"2023-08-04T16:00:00","publication_year":"2023","word_count":613,"keywords":["Go","ChatGPT","machine learning","AWS","AI","OpenAI","chatbots","Amazon Alexa","Aim","generative AI","Generative AI","R","Andy Jassy"],"extracted_tech_keywords":["AI","machine learning","generative AI","ChatGPT","OpenAI","Aim","chatbots","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/many-a-generative-ai-tricks-up-amazons-sleeves\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":32635,"title":"How AI Is Behind The Planning &#038; Execution Of Mega Sales And Campaigns","content":"Image for representation purpose We’ve seen, heard of, and even participated in some of the biggest online sales like Black Friday\/ Cyber Monday, Amazon’s Great Indian Festival, or Flipkart’s Big Billion Day Sale. These are among the most lucrative sales seasons for both e-commerce companies as well as brands since they are timed during peak festive periods each year when consumers are most willing to spend. Thus, capitalising on unique selling opportunities and seasonal consumption trends effectively is imperative to driving the success of such mega sales campaigns. E-commerce platforms plan their campaigns meticulously by choosing the right messaging, and spending a lot of time creating and pushing timely content to make sure it reaches the maximum possible number of consumers and garners visibility for various brands and products. E-commerce platforms use data – and lots of it – along with marketing automation tools when planning as well as executing such campaigns. These marketing tools or platforms, powered by Artificial Intelligence (AI) and advanced analytics, can deliver phenomenal results by streamlining the entire marketing and advertising process, right from the discovery stage till after the sale has been closed. Let’s see how. Big Data, Mega Impact: How AI-powered Marketing automation drives e-commerce sales campaigns A consumer typically spends a significant amount of time interacting with multiple apps, websites, and pages on any given day, and more so before they have to take a purchase decision. On the other hand, e-commerce platforms may get millions of visitors and page views on average, each day. The amount of data generated across each touch point is massive. Moreover, this data is often rich in context, since it is created by customers at various stages of their purchase journey when visiting a platform for product or price discovery. Brands can tap into this data when designing campaigns and promotions in a way that allows them to extensively leverage the insights to optimise outcomes to meet pre-defined sales targets. Technologies like artificial intelligence and machine learning bring massive computational capabilities to the table. With their ability to process a large amount of data created across multiple touch points, they are quickly becoming highly sought-after tools for brands today, and especially for a vast and rapidly growing industry like e-commerce. Using AI-driven automated marketing platforms, brands can record consumer information and apply predictive analytics models to the data sets and extract insights such as their preferences, the products they have purchased in the past, their motivation behind a particular purchase, etc. Intuitive and self-learning tools can leverage this knowledge to bundle the right combination of products and deals for consumers to ensure a higher degree of personalisation of content, delivered with a seamless experience. A few essential elements to driving such an experience will usually include the following 1) Crafting sales messages or content to drive engagement: By examining a large amount of data, AI uncovers a vast trove of information which would hitherto remain unused by brands in the absence of the right tools to help them make sense of it. However, with access to intuitive tools such as machine learning and predictive analytics, marketers can study customer behavior before, during, and after a purchase. Along with this, they can also view real-time information pertaining to a specific transaction such as the buyers’ demographic details, their feedback on the process and the product, etc. The results thus generated are integrated to identify patterns and clusters and patterns, allowing sellers to target the right consumers across different geographic and demographic segments, and accordingly devise their communication in real time for those who are likely to acquire a particular product. 2) Customer-centric and accurate searches: AI can distinguish between the keywords and search terms used by a consumer on the basis of their past searches. With the help of this information, machine learning models can help marketers and sellers decide what sort of content they want to show to a specific customer when they want to show it, the kind of deals they will be most interested in, etc. This further allows the e-commerce retailers to have adequate and relevant knowledge of the customer’s preferences and package their offerings in a way that appeals to their target audience. 3) Reaping the power of data and self-learning technology for mega success: The greater the amount of data available to brands and marketers, the more detailed are the insights which the AI and ML algorithms will be able to supply. Chatbots are an extremely efficient tool to gather meaningful data directly from customers. Simply put, chatbots are conversational, AI-powered messaging interfaces that are capable of providing not only the right answers to consumers’ queries but also accurate recommendations with the help of historical data, thus enhancing the overall quality and impact of customer engagement. This is precisely why e-commerce brands need to deploy interactive and intuitive communication tools on customer-facing touch points to carry out highly contextual and insight-rich conversations with prospective buyers. At the same time, they can access real-time analysis of these conversations, and then enhance the interactions to drive better outcomes. The insights generated by marketing platforms powered by AI and ML enable e-commerce companies to select the right product, deals and offers for each customer at the right time. The ML algorithms build, learn and enhance the outcomes after processing behavioral data at scale and distilling it to actionable insights. All of these tools work together like the parts of a well-oiled, intricately designed machine to deliver unprecedented businesses outcomes and growth opportunities to brands across the industrial spectrum.","excerpt":"We’ve seen, heard of, and even participated in some of the biggest online sales like Black Friday\/ Cyber Monday, Amazon’s Great Indian Festival, or Flipkart’s Big Billion Day Sale. These are among the most lucrative sales seasons for both e-commerce companies as well as brands since they are timed during peak festive periods each year […]","categories":["AI Features"],"tags":["AI marketing","personalisation"],"author_name":"Imran Saeed","publish_date":"2019-01-02T09:20:05","publication_year":"2019","word_count":924,"keywords":["Go","artificial intelligence","machine learning","AI","chatbots","R","ML","RAG","analytics","AI marketing","personalisation","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","chatbots","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-ai-is-behind-the-planning-execution-of-mega-sales-and-campaigns\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":60439,"title":"10 Free Resources To Learn Plotly","content":"Data visualisation plays an integral part in our lives as it helps in drawing meaningful insights from various types of data. Plotly is one of the powerful libraries for data science, machine learning and artificial intelligence-related operations where it helps to create multiple types of interactive visualisations by using Python, R and Java. In this article, we list down – in no particular order – the top 10 free resources to learn Plotly:- 1| Plotly Tutorial: Learn Plotly and R Source: DataCamp About: In this tutorial, you will understand how easily you can use Plotly to create data visualisations with R. You will learn how to use Plotly to generate heatmaps and 3D surface plots, a choropleth map, and how to add slides. You will also learn a brief introduction to ggplotly, the interactive sister of ggplot2. Get the resource here. 2| Plotly Dash User Guide Source: Official Site About: This is the official documentation for the Python implementation of Dash. Dash is a productive Python framework for building web applications which are built on Flask, Plotly.js, and React.js. Dash is ideal for building data visualisation apps with highly custom user interfaces in pure Python, and it is particularly suited for anyone who works with data in Python. Get the resource here. 3| Interactive Data Visualisation With Plotly in R Source: DataCamp About: In this tutorial, you will receive an introduction to basic graphics with Plotly. You will create your first interactive graphics by displaying both univariate and bivariate distributions, learn how to customize the appearance of your graphics and use opacity, symbol, and color to clarify your message, how to transform axes, label your axes, and customize the hover information of your graphs. Get the resource here. 4| Data visualisation With Plotly Express Source: Coursera About: This is a project-based course on data visualisation with Plotly Express where you will learn to create quick and interactive data visualisations with Plotly Express. Plotly Express is a high-level data visualisation library in Python inspired by Seaborn and ggplot2. In this course, you will explore the various features of the in-built Gapminder dataset, and produce interactive, publication-quality graphs to augment analysis. Get the resource here. 5| Getting Started With Plotly.js Source: Blog About: In this tutorial, you will learn all the features, bundles and chart types available in the library. You will understand how to create a basic line chart in Plotly, and how to style the chart lines using different attributes. Get the resource here. 6| Dash For Beginners Source: DataCamp About: Dash is a Python framework for building web applications, which is built on top of Flask, Plotly.js, React and React Js. In this tutorial, you will learn how to build dashboards in Python using Dash Plotly. You will understand how dash app layout works, and how to generate scatter plots, core components. Get the resource here. 7| Plotly Tutorial for Beginners Source: Kaggle About: In this tutorial, you will learn how to use Plotly library, how to create line plots, scatter plots, area charts, bar charts, error bars, box plots, histograms, heatmaps, subplots, multiple-axes, etc. You will also understand inset plots, 3D scatter plot with colour scaling and more. Get the resource here. 8| Learn Plotly Source: Youtube About: This is a video tutorial where you can learn how to use Plotly online, offline and in Jupyter Notebooks, as well as how you can build interactive visualisations using data. You will be able to get to explore the library while creating interactive visualisations using data. Get the resource here. 9| Learn Plotly Basics Source: Blog About: In this blog, you will learn the basics of Plotly, understand the plot method, understand how to create different traces\/plots like a scatter plot, or a line plot, as well as how to have a chart with multiple traces. Get the resource here. 10| Plotly Cheat Sheet Source: PDF About: In this cheat sheet, you will learn how to get started with Plotly and codes to create basic charts, layout, statistical charts, maps, 3D charts and figure hierarchy. Get the resource here.","excerpt":"Data visualisation plays an integral part in our lives as it helps in drawing meaningful insights from various types of data. Plotly is one of the powerful libraries for data science, machine learning and artificial intelligence-related operations where it helps to create multiple types of interactive visualisations by using Python, R and Java. In this […]","categories":["AI Trends"],"tags":["ai lessons for beginners","apm data science","Coursera","plotly","PowerBI"],"author_name":"Ambika Choudhury","publish_date":"2020-03-31T10:00:00","publication_year":"2020","word_count":679,"keywords":["data science","artificial intelligence","machine learning","Plotly","Coursera","AI","plotly","PowerBI","apm data science","Seaborn","Python","ai lessons for beginners","Jupyter","R","Java"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","Jupyter","Plotly","Seaborn","Python","R","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-free-resources-to-learn-plotly\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10049962,"title":"Top Data Observability Platforms For Monitoring Data Quality At Scale","content":"“Data observability goes deeper than monitoring by adding more context to system metrics, providing a deeper view of system operations, and indicating whether engineers need to step in and apply a fix,” explained Evgeny Shulman, Co-Founder of Databand.ai. Data pipelines can move data, but they can’t monitor it. These data pipelines are complex systems that require data observability architecture for constant sleuthing and end-to-end monitoring to understand why processes fail. This, along with the lack of observable data today, partly add up to creating the backbox. The black box gives you an output without humans being able to understand its processings yet. To correct the pipelines, engineers first need to observe. “In other words, while monitoring tells you that some microservice is consuming a given amount of resources, observability tells you that its current state is associated with critical failures, and you need to intervene,” according to Shulman. What makes an effective data observability platform? Observability platforms provide a way to track downstream dependencies to address the root problem. The essential components of an observability platform according to databand.ai: Simple setupEnd-to-end trackingObservability architectureThreshold settingAdministrationData observability open sourceDistributed systems observability Top data observability platforms Monte Carlo Monte Carlo provides an end-to-end solution to prevent broken data pipelines with its observability service. This is a great tool for data engineers to ensure reliability and avoid potentially costly data downtime. Monte Carlo’s features include data catalogues, automated alerting, and observability on several criteria out of the box. “In software engineering, every team has a solution like New Relic, DataDog, or PagerDuty to measure the health of applications and ensure reliability. How come data teams are flying blind?” the startup asks. The platform just raised $80 M in a Series C round. Databand Databand’s objective is to allow for more efficient data engineering in complex modern infrastructure. Its AI-powered platform provides data engineering teams with tools for a smooth operation that help them gain unified visibility into their data flows. The aim is to discover where the data pipelines broke before any bad data manages to squeeze through. The company also plugs into cloud-native tools like Apache Airflow and Snowflake in the modern data stack. Honeycomb Honeycomb’s observability tool provides engineers with the visibility to troubleshoot problems in distributed systems. The company claims that “Honeycomb makes it easy to understand and troubleshoot complex relationships within your distributed services.” Its full-stack cloud-based observability tool supports events, logs, and traces, along with an automatic instrumented code by its agent, Honeycomb beelines. Honeycomb also supports OpenTelemetry for generating instrumentation data. Acceldata Acceldata provides tools for data pipeline monitoring, data reliability, and data observability. The tools are to assist data engineering teams in gaining comprehensive and cross-sectional visibility into complex data pipelines. Acceldata’s tools synthesise signals across multiple layers and workloads on a single pane of glass to assist several teams in working together to fix data issues. In addition, Acceldata Pulse helps in performance monitoring along with observation to ensure data reliability at scale. The tool is geared towards the finance sector and payments. Datafold Datafold’s data observability tool assists data teams to monitor data quality through diffs, anomaly detection, and profiling. Its features allow teams to engage in data QA with data profiling, make table comparisons across databases or within a database, and create smart alerts from any SQL query with a single click. In addition, data teams can monitor ETL code changes with data transfers and integrate them with their CI\/CD to instantly review the code. SigNoz SigNoz is a full-stack open-source APM and observability tool that captures metrics and traces. Since the tool is open-source, users can host it within their infra without sharing their data with a third party. Their full-stack tools include a generation of telemetry data, a storage backend to store the telemetry data, and a visualisation layer to consume and take actions. SigNoz generates telemetry data using a vendor-agnostic instrumentation library, OpenTelemetry. DataDog DataDog’s observability tool’s features include infrastructure monitoring, log management, application performance monitoring, and security monitoring. DataDog allows full visibility into distributed applications by tracing requests from end-to-end distributed systems, charting latency percentiles, instrumenting open-source libraries and seamless navigation between logs, metrics, and traces. The founders claim this to be the “essential monitoring and security platform for cloud applications.” Dynatrace Geared towards large scale enterprises, Dynatrace provides a SaaS enterprise tool to target a varied spectrum of monitoring needs. Its AI engine, Davis, can automate root cause analysis and anomaly detection. Additionally, the company’s tools can provide a different solution for infrastructure monitoring, application security, and cloud automation. Grafana Labs Grafana’s open-source analytics and interactive visualisation web layer are quite popular for supporting several storage backends for time-series data. Grafana can connect to Graphite, InfluxDB, ElasticSearch, Prometheus, and various other data sources and supports Jaeger, Tempo, X-Ray, and Zipkin for traces. Its feature offers include plugins, dashboards, alerts, and other user-level access for governance. There are two different services provided – one, Grafana Cloud that provides solutions like Grafana Cloud Logs, Grafana Cloud Metrics, and Grafana Cloud Traces. Two, Grafana Enterprise Stack that supports metrics and logs with Grafana installed within the user’s computer. Soda Soda’s AI-powered data observability platform is a collaborative environment for data owners, data engineers, and data analytics teams to work and solve problems collectively. Soda.ai has described the platform as  “a data monitoring platform that allows teams to define what good data looks like and resolve issues swiftly before they have a downstream impact. This transparency and ease create trust — in each other, and the data.” In addition, the tools allow users to check their data immediately and create rules to test and validate data.","excerpt":"“In software engineering, every team has a solution like New Relic, DataDog, or PagerDuty to measure the health of applications and ensure reliability. How come data teams are flying blind?”","categories":["AI Trends"],"tags":["anomaly detection"],"author_name":"Avi Gopani","publish_date":"2021-09-30T13:00:00","publication_year":"2021","word_count":943,"keywords":["Elasticsearch","TPU","AI","ML","RAG","Ray","Aim","anomaly detection","analytics","Snowflake"],"extracted_tech_keywords":["AI","ML","analytics","Aim","Ray","RAG","anomaly detection","TPU","Elasticsearch","Snowflake"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-data-observability-platforms-for-monitoring-data-quality-at-scale\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077686,"title":"SetFit – A New Text-Classification Model That Outperforms OpenAI’s GPT-3","content":"The amount of available labelled data is a barrier to producing a high-performing model in many ML applications. Developments in the past two years have shown the challenge of overcoming data limitations by using LLMs (Large Language Models), such as OpenAI GPT-3 to achieve good results. However, while these improve the missing labeled data situation, they introduce a new problem of the access and cost of LLMs. To counter this, a group of researchers have discovered a new approach called SetFit to create highly accurate text-classification models with limited labeled data. Intel Labs, UKP Lab, and Hugging Face led the joint research that outperforms GPT-3 in 7 out of 11 tasks – while being 1600x smaller. Source: Phil Schmid According to the blog, SetFit has several unique features compared to other few-shot learning methods. One feature is using no prompts or verbalisers, as current techniques for few-shot fine-tuning require handcrafted prompts. SetFit altogether dispenses with prompts by generating embeddings directly from text examples. Moreover, it doesn’t require large-scale models like GPT-3 to achieve high accuracy. It also consists of multilingual support which can be used with Sentence Transformer on the hub. Source: Phil SchmidThe team has generated a high-performing text-classification model with 8 samples per class or only 32 labeled samples using the new approach. “This is huge! SetFit will help so many companies to get started with text-classification and transformers, without the need to label a lot of data and compute power. Compared to LLM training, the SetFit classifier takes less than 1 hour on a small GPU (NVIDIA T4) to train or less than $1 so to speak,” read the blog.","excerpt":"The joint research, which was led by Intel Labs and the UKP Lab, and Hugging Face, outperforms GPT-3 in 7 out of 11 tasks – while being 1600x smaller","categories":["AI News"],"tags":[],"author_name":"Bhuvana Kamath","publish_date":"2022-10-20T13:06:47","publication_year":"2022","word_count":273,"keywords":["Go","Hugging Face","OpenAI","AI","ML","Transformers","GPT","few-shot learning","R","llm_models:GPT"],"extracted_tech_keywords":["AI","ML","OpenAI","Hugging Face","Transformers","few-shot learning","R","Go","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/setfit-a-new-text-classification-model-that-outperforms-openais-gpt-3\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10090667,"title":"Pandas 2.0 is Finally Here!","content":"The most awaited Pandas 2.0 is finally here. The new updates come with new features, bug fixes, and improved performance, alongside breaking changes. Close to 253 people have contributed patches to this release. Check out the GitHub repository here. The release note stated that the users with existing code need to upgrade to pandas 1.5.3 before they upgrade to the second version of Pandas and make sure their code does not generate FutureWarning or DeprecationWarning messages. The release is said to be made available on conda-forge and PyPI. What’s new? There have been significant improvements compared to previous versions: Improved Performance The new version of Pandas has added the ability to use any numpy numeric dtype in an Index, and removed Int64Index, UInt64Index, and Float64Index. Also, the operations that previously forced the creation of 64-bit indexes can now create indexes with lower-bit sizes, such as 32-bit indexes. The ability for Index to hold numpy numeric dtypes has brought some changes in Pandas functionality. Now, instantiating using a numpy numeric array follows the dtype of the numpy array. Significant behaviour changes The bug fixes in the latest version of panda have bought some notable behaviour changes. For instance, the DataFrameGroupBy.cumsum() and DataFrameGroupBy.cumprod() methods now overflow instead of casting to float when the result can be held by int64 dtype. This makes sure that the results are correct and consistent with numpy and the regular DataFrame.cumprod() and DataFrame.cumsum() methods when the limit of int64 is reached. Further, SeriesGroupBy.nth() and DataFrameGroupBy.nth() methods now behave as filtrations instead of aggregations. In other words, they may return either zero or multiple rows per group, and the index of the result is derived from the input by selecting the appropriate rows. Say, when n is larger than the group, no rows instead of NaN is returned. The release not stated that these changes may have notable behaviour changes, so it is important to be aware of them when upgrading to Pandas 2.0. Read: Comprehensive Guide To Pandas Dataframes with Python Codes There is more The new version of Pandas also involves unsupported datetime and timedelta data types. For instance, in the previous versions, Pandas would replace unsupported data types with nanoseconds data types silently. But, in the new version, Pandas is said to support only “s”, “ms”, “us”, and “ns” resolutions, and it now raises an error instead of silently replacing unsupported data types with a supported one. In addition to this, Pandas 2.0 has made changes related to the result name and index of the Series.value_counts() method. For example, in the previous versions, the resulting name and index were the same as the original object. This used to cause a lot of confusion when resetting the index. In the new version, the result name willl be ‘count’ (or ‘proportion’ if normalise=True was passed), and the index will be named after the original object. In Pandas 2.0, the pandas disallow astype conversion to non-supported datetime64\/timedelta64 data types, and it raises an error. In comparison, in the previous versions, when converting a Series or DataFrame from datetime64[ns] to a different datetime64[X] dtype, Pandas would return with datetime64[ns] dtype instead of the requested dtype. For more details on the latest version of Pandas, click here.","excerpt":"The new version of Pandas has added the ability to use any numpy numeric dtype in an Index, and removed Int64Index, UInt64Index, and Float64Index.","categories":["AI News"],"tags":["pandas"],"author_name":"Tasmia Ansari","publish_date":"2023-04-03T21:37:59","publication_year":"2023","word_count":537,"keywords":["NumPy","ELT","AI","Git","Python","Ray","pandas","programming_languages:Python","GitHub","R","Pandas"],"extracted_tech_keywords":["AI","Ray","Pandas","NumPy","Python","R","Git","GitHub","ELT","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pandas-2-0-is-finally-here\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10097499,"title":"Top 10 Free Specialised Courses by Andrew Ng","content":"Andrew Ng, founder and chief of DeepLearning.AI, is leading the AI education space with his meticulously curated courses. In addition to offering the introductory ones, he provides a range of specialised courses, all of which are available for free. These specialised courses and programs by DeepLearning.AI, curated by a team of AI experts and led by Andrew Ng, provide valuable opportunities for learners to advance their knowledge and expertise in the rapidly evolving field of AI and deep learning. Let’s take a look at them. AI For Everyone Unlike the other courses by Ng, ‘AI for Everyone’ is a non-technical, introductory course designed to help both business professionals and technical individuals understand AI technologies and their applications. It covers the fundamentals of AI, machine learning, and data science, providing insights into what AI can and cannot do. Participants will learn about the workflow of AI and data science projects, how to choose AI projects and the impact of AI on society. The course aims to equip learners with the knowledge to build a sustainable AI strategy and navigate the challenges brought about by technological change. The course consists of four weeks of content, with a total duration of six hours. Instructor: Andrew Ng AI For Good The ‘AI for Good’ course, in collaboration with Microsoft’s AI for Good lab, is a specialisation designed for an individual’s interested in using AI to address real-world challenges in humanitarian and environmental projects. Participants will learn how to contribute to AI-powered initiatives that create positive change, such as mitigating climate change, supporting disaster response, and improving public health. The course provides a step-by-step framework for utilising AI in real-world projects and includes hands-on case studies and labs using Python and Jupyter Notebooks. It is suitable for learners from all backgrounds and does not require prior experience in AI or coding. Upon completion, participants receive a certificate and gain valuable knowledge to contribute to AI for Good initiatives worldwide. Instructor: Robert Monarch, ML leader at Apple Machine Learning Specialisation This is a newly rebuilt and expanded program created by Andrew Ng for beginners seeking to break into the field of AI and machine learning. The specialisation consists of three courses, and provides foundational AI concepts through an intuitive visual approach, followed by hands-on coding and an introduction to the underlying math. The course is designed for beginners, requiring no prior math or coding background. It covers topics such as linear regression, logistic regression, neural networks, decision trees, recommender systems, and more. The updated curriculum uses Python instead of Octave, and the section on applying machine learning has been enhanced with best practices from the last decade. Instructor: Andrew Ng Master the Mathematics Behind AI and Unlock Your Potential The ‘Mathematics for Machine Learning and Data Science Specialisation’ is a beginner-friendly course that equips learners with a solid understanding of essential mathematical concepts used in machine learning. The course covers calculus, linear algebra, statistics, and probability, providing students with the tools to comprehend algorithms and optimise them for custom implementation. By enrolling in this specialisation, participants will gain statistical techniques to enhance data analysis, acquire highly sought-after skills by employers for excelling in machine learning interviews and securing their dream jobs. The course features a team of instructors with expertise in the field, and it is designed for individuals with a high-school level of mathematics knowledge. Instructor: Luis Serrano, founder, Serrano Academy TensorFlow: Data and Deployment Specialisation The ‘TensorFlow: Data and Deployment Specialisation’ is a four-month intermediate-level program, with a recommended commitment of three hours per week. The specialisation aims to teach participants how to deploy machine learning models for various devices and platforms using TensorFlow. It covers topics such as running models in web browsers using TensorFlow.js, deploying models on mobile devices using TensorFlow Lite, data pipelines with TensorFlow Data Services, and advanced deployment scenarios with TensorFlow Serving. The courses include practical exercises and projects, and participants will learn to handle data, work with APIs, and use pre-trained models effectively. Instructor: Laurence Moroney, lead AI advocate at Google Generative Adversarial Networks (GANs) Specialisation The GANs Specialisation is a three course intermediate-level program that focuses on image generation using GANs. Students will learn to create basic GANs using PyTorch, advanced DCGANs with convolutional layers, and conditional GANs. The courses cover comparing generative models, using the FID method to assess GAN fidelity and diversity, detecting bias in GANs, and implementing StyleGAN techniques. Additionally, participants will explore GANs applications for data augmentation and privacy preservation, as well as building Pix2Pix and CycleGAN for image translation. The program also addresses social implications of GANs, such as bias in machine learning and methods to detect it. Throughout the courses, learners will develop skills in areas like generator design, image-to-image translation, and understanding computer graphics terminology, among others. The specialisation aims to provide a comprehensive understanding of GANs and offers practical hands-on experience. Instructor: Sharon Zhou, CEO, co-founder, Lamini AI for Medicine The course focuses on practical applications of ML in the field of medicine. Participants will learn to estimate treatment effects using data from randomized control trials, interpret diagnostic and prognostic models, and extract information from unstructured medical data using natural language processing. The skills acquired include model interpretation, image segmentation, natural language extraction, machine learning, time-to-event modeling, deep learning, model evaluation, multi-class classification, random forest, model tuning, and treatment effect estimation. The course will also explore various AI-driven medical applications, such as diagnosing diseases from X-rays and 3D MRI brain images, predicting patient survival rates with tree-based models, and automating the labelling of medical datasets through natural language processing. Instructor: Pranav Rajpurkar, assistant professor and director, Harvard-Stanford Medical AI Bootcamp Generative AI with Large Language Models (LLMs) This course, developed in collaboration with AWS, focuses on teaching the foundational principles and practical applications of generative AI in real-world scenarios. It covers the entire lifecycle of LLM-based generative AI, starting from data gathering and model selection to deployment and performance evaluation. Participants will gain a functional understanding of LLMs and the transformer architecture that powers them, along with the ability to fine-tune models for specific use cases. The course also explores cutting-edge research in generative AI and offers hands-on training, tuning, and deployment methods to optimise model performance. Instructor: Antje Barth and Mike Chambers, developer advocates, Gen AI, AWS. Chris Fregly and  Shelbee Eigenbrode, principal solutions architect, Gen AI, AWS MLOps Specialisation The MLOps Specialisation is an advanced 4-month course that equips students with production-ready ML skills. It covers tools, techniques, and experiences to build and maintain integrated systems operating continuously in production, handling evolving data efficiently. The four courses include topics like ML production system design, concept drift, data pipelines, feature engineering with TensorFlow Extended, and model resource management. Instructors: Andrew Ng, Robert Crowe, TensorFlow developer engineer, Google; Laurence Moroney, lead AI advocate, Google Practical Data Science on the AWS Cloud (PDS) Specialisation This is also another advanced course that equips data-focused developers, scientists, and analysts with the skills needed to deploy scalable ML pipelines using Amazon SageMaker in the AWS cloud. The specialisation covers various topics, including data preparation, feature engineering, automated machine learning (AutoML), model training and evaluation, ML pipelines, artefact and lineage tracking, and human-in-the-loop pipelines. Participants will gain hands-on experience with algorithms like BERT and FastText for natural language processing (NLP) using Amazon SageMaker. By the end of the program, learners will be able to build and deploy end-to-end ML pipelines, optimise model performance, and reduce costs while improving data products. Instructors: Antje Barth, developer advocate, Gen AI, AWS; Chris Fregly and  Shelbee Eigenbrode, principal solutions architect, Gen AI, AWS; Sireesha Muppala, principal solutions architect, AI and ML, AWS Read more: Top 7 Generative AI Courses by Andrew Ng","excerpt":"The free courses cover important topics like LLMs in Generative AI, AI for Medicine, Tensorflow, and more.","categories":["AI Trends"],"tags":["AI Tool","Courses","Top Trend"],"author_name":"Shritama Saha","publish_date":"2023-07-25T14:18:12","publication_year":"2023","word_count":1284,"keywords":["Top Trend","data science","machine learning","AI","neural network","ML","MLOps","NLP","Aim","deep learning","generative AI","Courses","AI Tool"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","data science","generative AI","MLOps","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-10-free-specialised-courses-by-andrew-ng\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10053449,"title":"How Fourier Transform is Used in Deep Learning?","content":"In machine learning or deep learning, the models are designed in such a way that they follow a mathematical function. From data analysis to predictive modelling there is always some mathematics behind it. For example, in clustering, we use the euclidean distance to find out the clusters. Fourier transform is also a famous mathematical technique for transforming the function from one domain to another domain which can also be used in deep learning. In this article, we will be discussing the Fourier transform and will understand how it can be applied in the field of deep learning.  The major points to be covered in this article are listed below. Table of Contents What is a Fourier Transform?Mathematics Behind Fourier TransformFourier Transform using PythonHow are Neural Networks Related to Fourier Transforms?Fourier Transform in Convolutional Neural NetworkHow to use Fourier Transforms in Deep Learning? What is a Fourier Transform? In mathematics, the transformation technique is used for mapping a function into a different function space from its original function space, and when we talk about the Fourier transform it is also a transformation technique that transforms such functions which are depending on the time domain space domain into such function which depends on the spatial domain or temporal frequency domain. We can take audio waves as an example where Fourier transformation allows us to represent it in terms of the volumes and frequencies of its notes. We can say that the transformation performed by the Fourier transform of any function is a function of frequency. Where the magnitude of the resultant function is the representation of the frequency contained by the original function. Let’s take an example of a signal whose function of time-domain looks as follows: Let’s take a portion of another signal in the same time frame Let the name of the signals be A(n) and B(n) where n is the time domain. So if we add these signals the structure of the signals will look as follows: C(n)  = A(n) + B(n) So here we can see that the addition of the signal of the functions has another form of a signal of any other function and if we talk about to extraction of the signal A or B from this addition signal C it becomes a problem for us because the addition of these signal happened only in the aptitudes, not in the time. The addition is having a similar time but different magnitude. Here the Fourier transform allows us to separate the function from their addition using its behaviour of conversion between time domain to frequency domain. In addition, we find that we can extract the frequencies from the signal but separating them using the time domain is not possible. The separation of signals using the Fourier transform will look like the below image. Image source In the above image, we can see that we can easily mark the difference between the signal now if again we want to get these signals in the time domain we can convert them using the inverse Fourier transform. In the next section of the article, we will discuss the mathematics behind the Fourier transform so that our view can be more clear about the proceeding of the Fourier transform. Mathematics Behind Fourier Transform Before going into the mathematics of Fourier let us know that the representation of a signal in the time domain can be done by the series of sinusoidal. So if the function consists of a continuous signal it can be represented by the function f(t). Where, We can see the function is made up of the addition of the infinite sinusoids which can be considered as the representation of the signal of function. Also, we can see there are in the function two coefficients that are necessary to define the structure of the output signal. These coefficients can be obtained by solving the Fourier transform integral which is basically a function of the frequency. The resultant of the Fourier transform can be considered as the set of coefficients Mathematically it can be written as given below: Fourier Transform And the inverse of this function can be considered as the function of time which we use for converting the frequency domain function to the time domain function. Inverse Fourier Transform Solving these above integral gives the value of a and b but hereThe case we are talking about is a case where the signals are continuous signals in real life most of the problems are accrued from the discreetly sampled signals and to find out the coefficient for such signals transformation we are required to perform a discrete Fourier transform (DFT). Using DFT we can get a same-length sequence of equally-spaced samples of a function that is already made up of a sequence of equally-spaced samples. The coefficients for the above-given function f(t) can be obtained by the following function. The value of a and b will be, Using these terms a and b in the function f(t) we can find the signals in the frequency domain. Here in this section, we have seen the mathematics behind the Fourier transform. Now let’s just see how can we perform on sine waves using python. Fourier Transform Using Python In this section of the article, we are going to transform the addition of two sine waves into their frequency domains for this purpose we can use the scipy library of python which basically provides us with all the transformations techniques we use in mathematics. Importing the libraries import numpy as np import matplotlib.pyplot as plt from scipy.fft import fft, fftfreq Making the sine waves and adding them #  sample points N = 1200 # sample spacing T = 1.0 \/ 1600.0 x = np.linspace(0.0, N*T, N, endpoint=False) sum = np.sin(50.0 * 2.0*np.pi*x) + 0.5*np.sin(80.0 * 2.0*np.pi*x) plt.plot(sum) plt.title('Sine wave') plt.xlabel('Time') plt.ylabel('Amplitude') plt.grid(True, which='both') plt.show() Output: Here in the above output, we can see the sine waves which we have generated using NumPy now we can transform it using the FFT module of the scipy library. sumf = fft(sum) xf = fftfreq(N, T)[:N\/\/2] plt.ylabel('frequency') plt.xlabel('sample') plt.title(\"FFT of sum of two sines\") plt.plot(xf, 2.0\/N * np.abs(sumf[0:N\/\/2])) plt.show() Output: Here we can clearly see what are the frequencies of different waves which was not clear in the addition because before this the wave was generated as the function of the time domain. As of now, we have seen the various insights of the Fourier transformation but now the question is how it is related to the neural network? By above all intuition we can say that the Fourier transform is also a technique to approximate other functions in the frequency domain which makes it related to neural networks. In the next section of this article, we will see how neural network and Fourier transform is related. How are Neural Networks Related to Fourier Transforms? As of now, we see the Fourier transform as a function that can help in approximating other functions and also we know that the neural networks can also be considered as the function approximation technique or universal function approximation technique. Let’s have a look at the below image. Image source This image is a representation of the neural network in which the Fourier transform technique is used. A basic neural network comes with the motivation to approximate a function that is unknown and its value at some given points. The majority of neural networks have a task to learn the overall function or learn the function at the values point which is given in the algorithm or data where iteration techniques help parameters to learn according to the situation where Fourier network finds the parameters by evaluating the given function. The formula used in the actual function helps in the computation of the parameter. Let’s take an example of a convolutional neural network to understand this relationship more in-depth. In the next section of the article, we will discuss the relationship of Fourier transform with neural network in the context of a convolutional neural network. Fourier Transform in Convolutional Neural Network As we know about the convolutional neural network, the convolutional layers are the main base of such kind of network and in the network, the main work of any convolutional layer is to apply the filter to the input data or to the feature maps. So that it can convolution the output from the previous layer. The task of the layer is to learn the weight of the filter. We see in a complex convolutional neural network that the number of layers is high and the filters per layer are also high which makes the calculation cost very high. Using the Fourier transform into the convolutional neural network we can transform the layers calculation in element-wise products in the frequency domain and the task of the network will be the same. We can just save the calculator energy by using the Fourier transform. Most of the cases we see are applying the fast Fourier transform in the convolutional network. By the above, we can say that the convolutional layer or the process of the convolutional layer is related to the Fourier transform. Mostly the convolutional layers in the time domain can be considered as the multiplication in the frequency domain. We can easily understand the convolutional by the polynomial multiplication. Let’s say we have to function y and g of any value x like below: y(x) = ax + b g(x) = cx + d And the polynomial multiplication of these functions can be written by a function h h(x) = y(x).g(x) = (ax + b)(cx + d) = ac x^2 + (ad+bc) x + bd By the above, we can say that the convolutional layer process can be defined as the multip[lication of these above-given functions. The vector form of  the functions can be written as the following: y[n] = ax[n] + b g[n] = cx[n] + d And the vector multiplication of the vector forms is: h[n] = y[n] X g[n] H[w] = F(y[n]) ‧ F(g[n]) = Y[w] ‧ G[w] h[n] = F^-1(H[w]) The notation ‘.’ in the multiplication represents the multiply and X is convolutional. The F and F^-1 are Fourier transform and inverse Fourier transform respectively. “n” and “w” donate time domain and frequency domain respectively. By the above, we have proven that ultimately the convolutional layer implies the Fourier transform and its inverse in the multiplication if the functions are related to the time domain. We have seen in this section how the Fourier transformation takes part in the convolution process. In the next section of the article, we will learn how we can use the Fourier in the deep learning algorithms. How to use Fourier Transforms in Deep Learning? In the above section, we have seen that the convolution process in the time domain can be simply considered as the multiplication in the frequency domain. Which is proof that it can be used in the various deep learning algorithms even though it can be used in the various static predictive modelling algorithms. In the above section of the article, we have seen how it helps in convolutional neural networks and how it replaces the convolution process of the convolutional layer. Let’s go with a similar example of the convolutional neural network so that we will not get diverted from the main subject of the article. As we have discussed and seen that the mathematics behind the convolution is to perform multiplication in the time domain and the mathematics behind the Fourier transform is to do multiplication in the frequency domain. So to apply the Fourier transform in any convolutional neural network we can perform some changes in the input and the filter. If the matrices of the input and filters in the CNN can be converted into the frequency domain to perform the multiplication and the outcome matrices of the multiplication in the frequency domain can be converted into the time domain will not perform any harm to the accuracy of the model. The conversion of matrices from the time domain to the frequency domain can be done by the Fourier transform or fast Fourier transform and conversion from the frequency domain to the time domain can be done by the inverse Fourier transform or inverse fast Fourier transform. The below image is a representation of how we can use Fast Fourier Transform in place of the convolution. As we have discussed the number of filters and layers in any complex network is very high and because of the higher numbers, the calculation process using the convolution becomes very slow. Whereas, using the Fourier transform we can reduce the complexity of such calculation and can make the model work faster. Final Words In this article, we have seen a basic definition of the Fourier transform along with mathematics behind it and how we can perform it using python. With all this, we get to know how the neural networks and Fourier transform are related. To understand this relation, we have taken an example of the convolutional neural network and seen how we can use the Fourier transform in deep learning models like convolutional neural networks.","excerpt":"Fourier transform is a transformation technique that transforms such functions which are depending on the time domain into such function which depends on the temporal frequency domain","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Convolutional Neural Networks","Data Science","Data Scientist","Deep Learning","Deep Learning Techniques","fourier transform","fourier transform machine learning","Machine Learning","Python"],"author_name":"Yugesh Verma","publish_date":"2021-11-14T16:00:00","publication_year":"2021","word_count":2193,"keywords":["TPU","Convolutional Neural Networks","fourier transform","deep learning","R","NumPy","Data Science","Go","machine learning","AI","neural network","Machine Learning","Matplotlib","fourier transform machine learning","Python","Deep Learning Techniques","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","NumPy","Matplotlib","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-fourier-transform-is-used-in-deep-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":38196,"title":"Deep Dive: Shine.com’s CTO Is Reimagining The Recruitment Platform With AI At Its Core","content":"In our weekly column, we spoke to Amardeep Vishwakarma, CTO at Shine.com, India’s second largest job portal to understand the growth mindset he has applied to build a high performing team and the recent AI moves made to solidify its position in the market. Vishwakarma is a long-time InfoEdge (Naukri) veteran who has over a decade of experience in leading recruitment platform. At Naukri, Vishwakarma built the flagship recruitment software, a news source indicated. In his present role, which he took over in July 2018, Vishwakarma’s focus area is to streamline search for job seekers and provide a seamless experience to recruiters. He is reimagining the recruitment software to shorten the recruitment cycle significantly and use AI\/ML to provide improved user experience.  “The entire recruitment process takes almost 70 days. It’s cycle of 70 days from starting to look for the candidate to bringing him on board. What we want to do is apply machine learning to shorten this cycle,” he said. Since he took over, Shine.com has made a concerted effort to introduce AI capabilities to move up the ladder and be known as an AI-first company. Last year in December, Gurgaon headquartered jobs portal introduced face recognition and touch ID capabilities in its mobile application. The latest features rollout was aimed at enhancing user experience on the Shine.com mobile app by facilitating seamless login. This is also an indication of Shine’s focus on user-centricity and how it wishes to be a front-runner in the recruitment space by leveraging technology as a differentiator. Vishwakarma, “The launch of this first-of-its-kind face and touch-based login features on our mobile app, well before any other player in the domain, further reinforces our commitment to simplify and adding value to the users’ interactions with our online platform.” He further added that Shine is also the only jobs app that is improving its hardware side as well to up the user experience. Tech Stack @Shine As a practice, Vishwakarma and his 40-member team leans towards open source technologies. “So, we basically are on Python. We use Python and we are hosted on Google Cloud Platform,” he shared. In terms of database technologies, the team uses MongoDB, MySQL. For caching and queueing, the team uses Aerospike or Redis. The tech team also uses RabbitMQ.  Apart from this, we use Java for the backend, he shared. Vishwakarma, an industry veteran, known for building teams from the ground up follows an innovative approach to scaling the team. Every year technology is changing, new databases come to the market, so it is very important for the team to stay updated. “We give assignments to the team with the new technology apart from the work and we just do a benchmarking or doing a PoC. We basically try the new technologies parallelly. We also have Shine learning where we basically ask our employees to take those courses and certifications to upgrade themselves,” he said. Use Cases@Shine One of the use cases discussed was applying machine learning to detect spam. “We are building models so that we can detect spam automatically as it is very difficult to check all the emails manually so this can be solved only through ML. We are also using the machine learning tools for recommendation,” he said. Another use case is where the candidate is looking for jobs and the team gets to understand his preferences. So, the data is used to train a model to recommend jobs that match the needs of the user.  “This is one of the core parts for us and we are building models around this so that we can better results to our job seekers,” he said. What Developers Should Know Before Joining The Shine Data Science Team Shine is aggressively hiring candidates to augment the data science team. The company is also hiring for the data analysts team, backend, frontend, QA, DevOps. “We have a lot of data and thus we are hiring more analysts so that they can help us to use all those data,” he said. Talking about the core capabilities he looks for in candidates, Vishwakarma shared capabilities as per the division. For example, a frontend engineer needs to be good in JavaScript and React, whereas on the backend, one should have data structures, problem solving skills, programming skills and learnability to hit the ground running fast. “We also look for what extra value he can add besides his skills like communication, or he can work together as a team or not. Similarly, for DevOps we look how much experience the candidate has in using cloud, Jenkins, AWS, CICD,” he shared. While hiring for data analyst roles, candidates should have a good knowledge of working on unstructured data and huge amounts of data, he said, in closing.","excerpt":"In our weekly column, we spoke to Amardeep Vishwakarma, CTO at Shine.com, India’s second largest job portal to understand the growth mindset he has applied to build a high performing team and the recent AI moves made to solidify its position in the market. Vishwakarma is a long-time InfoEdge (Naukri) veteran who has over a […]","categories":["AI Features"],"tags":["CTO"],"author_name":"Richa Bhatia","publish_date":"2019-04-24T13:10:19","publication_year":"2019","word_count":792,"keywords":["data science","machine learning","AWS","AI","CTO","MongoDB","ML","RAG","Python","Aim","Redis"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Aim","RAG","AWS","Redis","MongoDB","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deep-dive-shine-coms-cto-is-reimagining-the-recruitment-platform-with-ai-at-its-core\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10041212,"title":"Prosus Acquires Stack Overflow For $1.8 Billion","content":"Europe’s biggest tech investment company Prosus has acquired knowledge sharing SaaS platform Stack Overflow for $1.8 billion. Stack Overflow is a hugely popular online repository and question-and-answer website for software professionals, financial professionals, and marketers for improving coding skills. The website boasts over 100 million visitors a month. The acquisition move is seen as a bid to strengthen Prosus’ investment in the edtech space and expand its footprint. Prosus has stakes in edtech companies such as Udemy, Byju, and Codeacademy. “Prosus’s mission is to build leading companies that empower and enrich communities, as demonstrated by the many community-focused and EdTech companies they work with. This makes Prosus the perfect company to acquire Stack Overflow, and Stack Overflow the ideal investment in their focus on the future of workplace learning and collaboration,” said Prashanth Chandrasekar, CEO, Stack Overflow Big news! Stack Overflow is excited to be joining the Prosus family of companies, propelling us into our next phase of growth. Cheers to the future of workplace collaboration and learning! https:\/\/t.co\/ENHi6pIfza pic.twitter.com\/CWpq1LjyAz— Stack Overflow (@StackOverflow) June 2, 2021 Stack Overflow is one of Prosus’ biggest acquisitions. Prosus is the largest shareholder in Chinese tech giant Tencent, at a $200 billion holding in the company. In April, Prosus sold a fraction of its shares in Tencent to raise $15 billion. This reduced its share in Tencent from 30.9 percent to 29.9 percent, making it the second-largest ever block trade. Prosus invests across a range of online platforms focused on delivery, fintech, and classifieds, for example, Movile and iFood.","excerpt":"The acquisition move is seen as a bid to strengthen Prosus’ investment in the edtech space.","categories":["AI News"],"tags":["Developers","Mergers and Acquisitions","Stack Overflow"],"author_name":"Shraddha Goled","publish_date":"2021-06-03T10:52:42","publication_year":"2021","word_count":255,"keywords":["programming_languages:R","AI","Stack Overflow","Mergers and Acquisitions","R","Developers"],"extracted_tech_keywords":["AI","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/prosus-acquires-stack-overflow-for-1-8-billion\/","complexity_score":2,"technical_depth":3,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172649,"title":"Meet Meta’s New Superintelligence Dream Team","content":"Social Media giant Meta announced on July 1 that Alexandr Wang, former CEO of Scale AI, has been appointed as Meta’s chief AI officer and will co-lead Meta Superintelligence Labs (MSL) with Nat Friedman, former CEO of GitHub. This follows Meta’s $14.3 billion investment in Scale AI earlier in June. The company has been on an AI hiring blitz, recruiting top talent from OpenAI, Anthropic, Google, and DeepMind, with some signing bonuses reportedly reaching up to $100 million. The following individuals comprise the superintelligence team. Trapit Bansal Trapit Bansal is recognised for his pioneering work in applying reinforcement learning to chain-of-thought reasoning within large language models. As a co-creator of OpenAI’s o-series models, Bansal has played a pivotal role in advancing model interpretability and robustness. His research focuses on developing training methodologies that enhance both the reasoning capabilities and efficiency of modern AI systems. Shuchao Bi Shuchao Bi contributed significantly to the development of GPT-4o’s voice mode and the o4-mini model. At OpenAI, he led efforts in multimodal post-training, which involved refining how models process and generate outputs across text, audio, and visual inputs. Huiwen Chang Huiwen Chang was instrumental in designing GPT-4o’s image generation features and has a strong background in generative AI. Previously at Google Research, she invented the MaskGIT and Muse architectures, both of which have become foundational in the field of text-to-image synthesis. Ji Lin Ji Lin played a key role in building a suite of influential models, including o3\/o4-mini, GPT-4o, GPT-4.1, GPT-4.5, 4o-ImageGen, and the Operator reasoning stack. His contributions span the development of advanced reasoning mechanisms and architectural improvements, enabling these models to perform a wide range of complex AI tasks more effectively. Joel Pobar At Anthropic, Joel Pobar led work on inference optimisation. He previously spent over a decade at Meta, where he contributed to the development of core infrastructure projects, including HHVM, Hack, Flow, Redex, and various performance and machine learning tools. His deep experience in software engineering and AI has been critical to improving the speed and scalability of AI inference systems. Jack Rae Jack Rae serves as the pre-training technical lead for Gemini and leads reasoning for Gemini 2.5. At DeepMind, he spearheaded early large language model projects including Gopher and Chinchilla. His expertise lies in large-scale pre-training strategies and improving the reasoning capabilities of cutting-edge AI models. Hongyu Ren Hongyu Ren is a co-creator of several OpenAI models, including GPT-4o, 4o-mini, o1-mini, o3-mini, o3, and o4-mini. He previously led a post-training group at OpenAI, focusing on refining and optimising large language models for greater accuracy and efficiency through advanced post-training techniques. Johan Schalkwyk Johan Schalkwyk, a former Google Fellow, was an early contributor to the Sesame project and served as the technical lead for Maya. His extensive background in AI research and development has influenced foundational advancements in machine learning frameworks and AI technologies. Pei Sun Pei Sun worked on post-training, coding, and reasoning for the Gemini project at Google DeepMind. Previously, he developed the last two generations of perception models for Waymo, demonstrating his expertise in AI for autonomous vehicles. His current focus is on enhancing reasoning and real-world application capabilities in advanced AI systems. Jiahui Yu Jiahui Yu is a co-creator of o3, o4-mini, GPT-4.1, and GPT-4o. He previously led OpenAI’s perception team and co-led multimodal research for Gemini. His work is centred on advancing AI perception and integrating multimodal understanding, enabling models to process and generate information across diverse data types. Shengjia Zhao Shengjia Zhao has been a co-creator of ChatGPT, GPT-4, the mini model series, 4.1, and o3. At OpenAI, he led synthetic data initiatives, focusing on improving the diversity and quality of training data. His innovations in data synthesis have been crucial for enhancing model generalisation and overall performance.","excerpt":"This follows Meta’s $14.3 billion investment in Scale AI earlier in June.","categories":["AI Features"],"tags":["Meta"],"author_name":"Siddharth Jindal","publish_date":"2025-07-01T13:57:48","publication_year":"2025","word_count":625,"keywords":["Anthropic","ChatGPT","Meta","machine learning","Gemini Pro","OpenAI","AI","GPT-4o","GPT-4.5","generative AI","Gemini 2.5"],"extracted_tech_keywords":["AI","machine learning","generative AI","GPT-4.5","GPT-4o","ChatGPT","OpenAI","Anthropic","Gemini 2.5","Gemini Pro"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-metas-new-superintelligence-dream-team\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121216,"title":"How to Build Sustainable AI Startups","content":"With the advancements brought in by GPT models, GPT-4o being the latest, creating sustainable AI startups that leverage artificial intelligence and can last and grow over time has become increasingly important. In a recent podcast, OpenAI chief Sam Altman spoke about how to either create a business that thrives even if the next AI model isn’t significantly better, or develop a system that gets more useful as AI models improve or advance. Additionally, he favoured not building an “AI business” in most cases, but rather a business that uses AI as a technology. Giving an example, he drew parallels to the early days of the App Store, where many people built simple apps like Flashlight which became obsolete with an iOS upgrade. Meanwhile companies like Uber established sustainable businesses as smartphones improved, leveraging phones as the key technology that significantly enabled their operations. How OpenAI Plans to Monetise Recently, OpenAI made GPT-4o available to everyone for free (with usage limits), offering features like browsing, data analysis, and memory. Additionally, Plus users will receive up to 5x higher limits and earliest access to features like the new macOS desktop app and next-generation voice and video capabilities. The move highlights OpenAI’s efforts to encourage upgrades to their monetisation plans, as discussed in the podcast. Altman said that they are yet to figure out ways to make an expensive technology like GPT-4 available to users for free. He emphasised that while they aim to provide advanced AI tools for free or at a minimal cost as part of their mission, the high expenses currently pose a significant barrier. Meanwhile, OpenAI recently became a Reddit advertising partner, which likely indicates that the company can leverage Reddit’s large user base to advertise its own products and services, potentially driving more customers and revenue. OpenAI’s revenue for this year has surpassed the $2-billion mark, according to reports from the Financial Times. Therefore, like OpenAI showcasing its continuous revenue generation, startups must also ensure they can sustain their business models in the long run. Do startups need to follow big companies A few days ago, Cred founder Kunal Shah cast a wide net asking people on X this direct question: “Who is building an AI application in India”, receiving nearly 300-400 responses. Dharmesh BA, who is working on a stealth startup, noted that many products were simply wrappers around existing models in various modalities. He categorised these apps as CRUD (Create, Read, Update, Delete) and warned that building apps based on the assumption that OpenAI or current LLMs can’t perform specific tasks could lead to a disaster. Each time OpenAI updates or releases a new version, many startups find themselves rendered obsolete because the enhanced capabilities of OpenAI often solve the problems these startups were aiming to address. When OpenAI introduced ChatGPT Enterprise, it sent shockwaves across several SaaS startups that had developed products around ChatGPT or offered wrappers based on ChatGPT APIs for business clients. Additionally, Dharmesh’s post highlighted a perspective that attempts to confine an extremely powerful technology, like LLMs—which can be compared to a genie capable of doing anything—into a limited space such as mobile apps or websites. These technologies are capable of much more complicated and valuable work, and by limiting their potential, we are not utilising them in the medium they are meant to reside in. What about Indian Startups In yet another post on X, the Cred founder said that early-stage startups should be easy to iterate and late-stage startups should be hard to distract. This highlights the mentality of Indian startups that are not iterating and not innovating enough. In India, researchers and enterprises should prioritise building large models, technical benchmarking, and AI industrial standardisation over developing specific use case apps, which are easily replicated and improved upon. Most envision LLMs as operating systems where users choose their own apps, but these apps’ longevity depends on the base provider, like OpenAI’s architecture. But as AIM wrote, the question remains as to why such research isn’t being conducted domestically, especially as tech giants like OpenAI and Google focus more on Indic languages, posing a threat to those developing for the Indian ecosystem.","excerpt":"Sam Altman believes in not building an AI business but rather a business which has AI as a technology.","categories":["Deep Tech"],"tags":["AI","Open AI","Sam Altman","Startups","sustainability"],"author_name":"Gopika Raj","publish_date":"2024-05-22T10:50:49","publication_year":"2024","word_count":691,"keywords":["Go","ChatGPT","API","Sam Altman","artificial intelligence","OpenAI","Open AI","sustainability","AI","GPT-4o","RAG","Aim","Startups","R"],"extracted_tech_keywords":["AI","artificial intelligence","GPT-4o","ChatGPT","OpenAI","Aim","RAG","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-build-sustainable-ai-startups\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10121586,"title":"Society Separates Those Who are Good at Maths and Those Who Are Not, But AI Doesn’t","content":"During an episode of the Logan Bartlett Show, Sam Altman recalled how calculators were perceived in his maths classes. “We never got to use calculators,” he said, adding that conversely you had to be proficient with calculators in real-life to excel later. “If OpenAI researchers never got to use calculators, OpenAI wouldn’t have happened,” he said, explaining that we now need to teach people how to use AI because it’s going to be an important part of what we do in the future. The importance of learning maths has been emphasised from our school days. It’s a crucial requirement in order to excel as an engineer. Further, with computer science becoming so mainstream, society has started to separate those who are good at maths from those who are not. Now AI is bringing down this wall for the better. With the advent of tools such as ChatGPT and Copilot, everyone is increasingly becoming a developer without needing to learn maths, democratising access to fields that once required deep knowledge of the subject. Some say mathematics and AI are two branches of the same tree. Others believe that mathematics forms the backbone of AI. But ever since we’ve built softwares that can assist programmers without the need for even a tiny bit of mathematics, AI\/ML has become a field widely accessible to everyone as the mathematical barrier to entry is being done away with. The Experts Don’t Necessarily Agree “I’ve often said ‘don’t worry about it’ when it comes to maths, because maths shouldn’t hold anyone back from making progress in ML. And, understanding some key topics in linear algebra, calculus, and probability and statistics will help you get learning algorithms to work better,” said Andrew Ng, the AI guru. Interestingly, this was while he was introducing a course on maths for ML and data science. NYU associate professor Julian Togelius said that you can indeed be successful in CS, including in machine learning, while knowing next to nothing about maths. “Just look at me, I barely passed those required theory courses, still made it here,” he said. Meanwhile, Harvard University professor of computer science Boaz Barak said last year, “I teach computer science, and I apologise lesser now about having so much maths.” The same is the case with The Math(s) Fix author Conrad Wolfram. Wolfram said that with the era of ChatGPT, being bad at maths is as big a problem of the student, as it is of the subject itself — it has become stagnant. On the other hand, some ML engineers have never stumbled upon the usage of maths in their lives. “Pure maths research is not typically published at top ML conferences,” a Reddit user pointed out. “I have spent way more time installing CUDA drivers than proving theorems,” said another Reddit user in a thread talking about how much maths is involved on a daily basis in ML engineering. Sure, the barrier to entry for a computer newbie to enter into the ML field has drastically decreased. A lot of ML fields are just about deploying models, and a lot of new models like ChatGPT or Codex even write the code for you. Knowing the maths behind all of this is something we don’t even think about anymore. But funnily enough, AI is not yet good at maths, though its capabilities are increasing. Recently, ChatGPT with the Wolfram plug-in scored a 96% in the UK A-level paper for maths, which is an essential qualification to get into the AI field. What this tells us is that if AI is able to crack an exam that is meant to get into AI, there needs to be a major change in the educational systems across the world to adjust to the shifting paradigm of mathematical teaching. Maths is Like Law, and the Divide Will Continue Mathematics, at its core, is the embodiment of logic. AI is deeply rooted in mathematical principles. From its inception, AI has relied on mathematical concepts to create models that mimic human cognition and decision-making processes. Understanding the relevance of mathematics in AI requires acknowledging that AI itself is an applied manifestation of mathematical logic. Does society need mathematics? Absolutely. However, in today’s world, the accessibility of AI tools means that even those who are not mathematically inclined can perform complex calculations and data analysis with ease. Microsoft has taken this a step further when it comes to reducing the barrier of entry for experts not well-versed with maths to enter into the field. At the Microsoft Build conference this year, CEO Satya Nadella announced that now everyone can code in their native language with Copilot Workspace. Adding to that is the fact that it is increasingly becoming the case that you do not need a degree to get an AI job. Experienced tech leader and consultant Oskar Ojala elaborated on the practical application of maths in solving real life problems while giving an example of the success of Facebook. Disagreeing with Ojala, nbn Australia research engineer Alex Eisenmann said that CS without maths could give you Facebook, but CS with maths has the potential to provide frameworks like AI, ML, quantum computing, and blockchain.","excerpt":"But funnily enough, AI is not yet good at maths.","categories":["AI Features"],"tags":["AI Impacts"],"author_name":"Mohit Pandey","publish_date":"2024-05-24T16:09:08","publication_year":"2024","word_count":862,"keywords":["CUDA","data science","ChatGPT","Go","machine learning","OpenAI","AI","ML","AI Impacts","Julia","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","ChatGPT","OpenAI","CUDA","R","Go","Julia"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/society-separates-those-who-are-good-at-maths-and-those-who-are-not-but-ai-doesnt\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10003299,"title":"This Framework Leads to 50% Cost Reduction From ML API Calls","content":"“The heterogeneity in their performance and price make it challenging to decide which API to use.” Machine learning as a service (MLaaS) is estimated at $1 billion in 2019, and it is expected to grow to $8.4 billion by 2025. At the heart of these services, sits customised APIs from companies like Google and Amazon, who have been spearheading the ‘democratisation of ML’ movement. For example, one could use Google prediction API (Goo) to classify an image for $0.0015 or to classify the sentiment of a text passage for $0.00025 only! The race to democratise has made MLaaS a lucrative business model. The result is, today, there are multiple APIs offering similar services. This again, can be challenging. Addressing this issue and to establish a hassle-free ML ecosystem, a group of researchers from Stanford University, introduced a predictive framework called FrugalML that assists the users in switching between APIs in a smart manner. The researchers have detailed about their new framework in a paper titled, ‘To Call or Not to Call?’ Overview Of FrugalML via paper by Stanford University FrugalML framework is designed to first learn the strength and weakness of each API jointly, and then perform an efficient optimisation to identify the best adaptive strategy given the user’s budget constraint. The above illustration compares different approaches that use ML APIs. Naively calling a fixed API in (a) results in a fixed cost and accuracy. The simple cascade in (b) uses the quality score (QS) from a low-cost open-source model to decide whether to call an additional service. Proposed FrugalML approach, in (c), exploits both the quality score and predicted label to select APIs. FrugalML leverages the modular nature of APIs by designing adaptive strategies that call APIs sequentially. Adaptive strategies are hard to learn, as the choice of the one predictor could depend on the prediction and confidence of the previous API. FrugalML may need to allocate different fractions of its budget to predictions for different classes. FrugalML optimises adaptive strategies to substantially improve prediction performance over simpler approaches such as model cascades with a fixed quality threshold. Here, the master problem decides which two services can be the base service, how often should it be invoked, and how large budgets are assigned, while for a fixed base service and budget, the subproblem maximises the expected reward. First, conditional accuracy is estimated from the training data. Next for each service, the subproblem is solved for some value of a budget, and then an optimal strategy is constructed. Finally, the master problem is transformed into a few linear programmings and solved efficiently. For experiments, the researchers evaluated the performance of FrugalML as well as those APIs on three datasets, namely, FER+, RAFDB, and AFFECTNET, which contain images with basic emotions and component emotions. The results show that FrugalML leads to more than 50% cost reduction when using APIs from Google, Microsoft and Face++ for a facial emotion recognition task. Whereas, experiments on FER+ dataset showed that only 33% cost is needed to achieve accuracies that match those of Microsoft API. The authors posit that the performance of Frugal ML is likely because the base service’s quality score is highly correlated to its prediction accuracy, and their framework only needs to call expensive services for a few difficult data points and relies on the cheaper base services for the relatively easy data points. Key Takeaways APIs are designed to do the mundane developer chores for the ML outsiders. However, introducing a flurry of APIs come with their own challenges. This novel framework by the Stanford researchers tried to address a few of them. According to the researchers, the contribution of their work can be summarised as follows: A new framework FrugalML is proposed that jointly learns the strength and weakness of each API on different data and performs an efficient optimisation to automatically identify the best sequential strategy to adaptively use the available APIs within a budget constraint. Compared to the simple cascade approach, FrugalML consistently reaches a higher accuracy while using the same budget.FrugalML aims at finding strategies to select an apt API from the MLaaS market to reduce costs and increase accuracy. With the successful demonstration of Frugal ML, the researchers are hopeful that the future work will lead to extending FrugalML to produce calling strategies for ML tasks beyond classification as well as comprehensive empirical evaluation. Link to the original paper","excerpt":"“The heterogeneity in their performance and price make it challenging to decide which API to use.” Machine learning as a service (MLaaS) is estimated at $1 billion in 2019, and it is expected to grow to $8.4 billion by 2025. At the heart of these services, sits customised APIs from companies like Google and Amazon, […]","categories":[],"tags":["API","stanford university"],"author_name":"Ram Sagar","publish_date":"2020-07-26T10:00:00","publication_year":"2020","word_count":733,"keywords":["Go","API","machine learning","programming_languages:R","AI","ML","programming_languages:Go","RAG","Aim","stanford university","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","RAG","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/api-framework-cheaper-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10110326,"title":"Why is Tamil Nadu becoming a Sweet Spot for EV Investments?","content":"The recently concluded Global Investors Meet (GIM) in Tamil Nadu saw huge investments from tech, auto, energy, and manufacturing companies from across the globe. The state bagged a total investment of over INR 6.6 lakh crore, with a substantial amount allocated for EV development. Already seen as a hub for auto manufacturing, and Chennai being considered as the ‘Detroit of Asia‘, Tamil Nadu is slowly becoming the EV Capital of India. Stakes on TN Last year alone, Tamil Nadu contributed to 40% of the country’s EV production. Out of the ten lakh EVs sold, more than four lakh were manufactured in Tamil Nadu. Of all the EV two-wheelers sold in India, a whopping 68% was manufactured in the state. Adding to the existing EV factories in the state, many new investments were announced this week. Vietnamese electric car manufacturer VinFast has committed to investing $2 billion to set up a manufacturing plant in Tamil Nadu. The factory construction will begin this year with an initial investment of $500 million for the first phase of the project, and is said to generate 3500 jobs and produce 150,000 EVs annually. To expand into key markets, VinFast’s investment commitment is also a bid to take on its biggest EV competitor Tesla. Banking on the east coast state that houses a number of auto manufacturing companies, Vinfast’s approach to start a factory in Tamil Nadu is a strategic one. Interestingly, Elon Musk’s Tesla is also deciding to start a new manufacturing plant, however, that is set to happen in Gujarat. Hyundai Motors is also set to invest ₹6180 crore to boost the EV sector in Tamil Nadu. The company is gearing up to establish a ‘Hydrogen Valley Innovation Hub’ to bolster the growth of green energy. Furthermore, Hyundai had earlier committed to investing ₹20,000 crore over the next ten years. Last year, Ola Electric committed to investing ₹7600 crore to build the world’s largest EV facility at a single location spread across 2000 acres. Advanced cell and EV manufacturing facilities, vendor supplier parks and ancillary ecosystem will be part of the EV hub built in the state. EV two-wheel company Ather Energy already has two plants in Hosur and is looking to expand further. A Tamil Nadu Manufacturing Affair As per a report published last year, Tamil Nadu earns 8% of its GDP from the automobile industry. The foray into the automobile sector dates back to early 90s when the Indian economy had faced liberalisation and private companies were breaking the shackles of setting up industries. Ford Motor in a joint venture with Mahindra & Mahindra set up the first car facility closer to Chennai. Since then, TN is synonymous with auto manufacturing, with a number of companies setting their base there. With OEM and component manufacturing factories also set up in TN, the facilitation of EV factories will be simpler. Suitable government policies such as reduction of tariff for EV charging stations will allow more EV stations to come up, thereby promoting EV adoption too. Furthermore, Tamil Nadu’s Electric Vehicles Policy 2023, looks to make Tamil Nadu the EV manufacturing hub of South-East Asia. Incentives under EV Special Manufacturing Package and tax exemptions for EV companies have been driving these companies in a big way. While TN is already big on EV factories, another state is slowly catching up to become the next desirable stop for EV production. Gujarat Close on the Heels The Vibrant Gujarat Summit that is under way in the state, is expected to unveil huge announcements in EV production. With Tesla already expected to set up their first manufacturing plant in the state, Tata Motors’ famed Sanand plant will also grow into a central hub for EV initiatives. Maruti Suzuki also has an EV plant in the state. Yesterday, Gensol Engineering, operator of solar projects, committed to investing ₹2000 crore to set up an EV manufacturing plant in Gujarat. With more announcements in the pipeline, Gujarat is close behind on grabbing more EV deals.","excerpt":"First an automobile manufacturing hub, now the EV hub, TN seems to attract them all.","categories":["IT Services"],"tags":["ather energy","Elon Musk","EV","ford","Gujarat","infrastructure","OEM","Ola Electric","Tamil Nadu","tata","Tesla"],"author_name":"Vandana Nair","publish_date":"2024-01-10T15:18:24","publication_year":"2024","word_count":665,"keywords":["API","Ray","R","ford","infrastructure","Gujarat","Elon Musk","Tamil Nadu","Tesla","Go","Ola Electric","AI","ather energy","OEM","EV","programming_languages:R","innovation","tata","programming_languages:Go"],"extracted_tech_keywords":["AI","Ray","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/why-is-tamil-nadu-becoming-a-sweet-spot-for-ev-investments\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69260,"title":"Is It Ethical To Let A Model Decide Who Should Graduate? International Baccalaureate’s New Rules","content":"“The IB’s model has glaring methodological issues and completely disregards the ethical considerations which should accompany its adoption.” The International Baccalaureate (IB), which has a global presence and whose programmes are followed by more than 5,000 schools in 158 countries, has now made amendments to their grading process by using a statistical model to award grades to the May 2020 Diploma Programme. So, how fair is it to use a statistical model to decide the future of the students? Condemning the adoption of this approach, a Data Scientist from Berkeley, wrote a detailed blog post explaining the perils of haphazard practices employed by the reputed organisation. Analytics India Magazine got in touch with the blogger of positivelysemidefinite to know more. Although, the blogger admits that he is no expert on pedagogical practices, but this decision by IB is an absolutely terrible idea to make everything proprietary with no scope for oversight. He says that IB should acknowledge the cases in which their model fails systematically and allow for greater leeway of appeals in those cases. “I think that such standard disclosures should be a part of any machine learning model in sensitive domains,” said the blogger. What’s Wrong With IB’s Approach “A model can choose to assign female students to systematically lower grades in STEM subjects and\/or incorrectly fail Black students at higher rates than Asian students.” Biases in algorithmic-based solutions are still an ongoing problem. Last month, we witnessed the Twitter meltdown, which was sparked by the tweet of Turing Awardee Yann Lecun. However, there probably hasn’t been such a large scale experiment like that of IB’s so far, where flaws in the model can put 160,000 people’s future at stake. Due to curriculum disruptions, the IB has been forced to cancel final exams for its current student cohort and use a model instead, to assign final grades. “Together we have developed a method that uses data, both historical and from the present session, to arrive at the subject grades for each student,” said IB, backing up their decision. The three-step process, illustrated below, will be used to prescribe final grades to each student. I will refer to this entire process as the ‘model’. Addressing the role of historical bias in the grading practices, the blog states that the secondary school teachers, according to a study done by National Center for Education Statistics, tend to express lower predictions for their ‘expectations from students of colour and students from disadvantaged backgrounds’. This is problematic because predicted grades play a prominent role in the model. “How ethical is it to tell a 17-year-old kid that they were unable to graduate with their peers because their inaccurate prediction was ‘the cost of doing business’?” Regarding the reliability of the model, the blog raises the following concerns: How was the error adjusted for classrooms with varying capacities?Was the socioeconomic status of the schools considered?What about the schools with new teacher recruitments within the academic year? Will the historical relationship match the current relationship? These queries barely scratch the surface. “There are many other nuanced problems which may arise depending on the sort of model that the IB decides to use,” lamented the blogger. He also believes IB has possibly overlooked ethical considerations while making its operational choices. Going Forward “ IB should seriously re-evaluate the manner in which it has chosen to manoeuvre this very delicate situation.” Source: positivelysemidefinite To demonstrate the dire nature of IB’s decision, a model was built that simulates the approaches of IB. This model, which was unaware of any data about the race, socioeconomic status or gender of the student body in any high school, predicted the majority race(black\/Hispanic) of the high school with higher accuracy than that of the graduation rate! The blogger warns that one can be fooled into thinking that a model which isn’t aware of gender\/race\/socioeconomic status cannot possibly discriminate based on these attributes; known as ‘fairness through unawareness’. Talking about the effectiveness of the new tool, he says that historically, predicted grades never had an impact on final grades. They were never factored into final grades. Final grades were always based on an evaluation of anonymous-randomly assigned exams. “Open source the model & results on the extent of the bias.” We all agree upon the fact that humans are biased. However, most of us try to practice sensibility in matters of delicate nature. But, when a mathematical model is tasked with something as critical as the careers of students, we expect it to be perfect. Those who deploy these models cannot direct the blame towards a model’s lacklustre performance. If that is the case, then the approach should at least be made transparent so that concerned parties would know what they are getting into. “What was the point of fitting a 2-dimensional model altogether? They should have simply based the grades on submitted coursework with some additional processes to appeal the grades,” asks the blogger. That said, he agrees that the IB is a good organisation with good intentions. However, their decision, he thinks, was definitely shortsighted and completely disregards the ethical considerations. Update: IB’s Statement International Baccalaureate contacted Analytics India Magazine to respond with regards to the criticism in this article. Here’s their statement: “The decision to cancel the May 2020 examinations due to the COVID-19 pandemic was incredibly difficult, and as the IB responds to these exceptional circumstances, it has endeavoured to be as transparent as possible. The final grades for the May 2020 DP and CP session are based on the student’s coursework throughout the two-year programmes, predicted grades provided by schools, and historic assessment data. For the subjects where students would normally sit exams, historic data was analysed to determine the global relationship between coursework marks, predicted grades and final subject marks. The IB has applied this calculation to determine the final grade for the May 2020 DP and CP session. Prior to the attribution of final grades, this process was subjected to rigorous testing by educational statistical specialists to ensure our methods were robust. It was also checked against the last five years’ sets of results data, to ensure that it would provide reliable and valid grades for students. The stability of results for students has been maintained for the May 2020 session. The mean total points for May 2020 DP students show small increases in the average grade achieved compared to previous years. The grade distribution level is also in line with the previous four years of results data. The IB is confident that it has awarded grades in the fairest and most robust way possible in the absence of examinations, and the grades awarded to students are of equal value to those awarded in any other year. This level of confidence means that the DP and CP certification documents awarded to students for the May 2020 session will be the same as any other session. IB World Schools can request re-marks of students’ work in the May 2020 session through the Enquiry Upon Results (EUR) services. There are some changes to the EUR services in this exceptional session due to the fact that marks have been calculated in the absence of examinations in some subjects, and these have also been communicated to IB World Schools.”","excerpt":"“The IB’s model has glaring methodological issues and completely disregards the ethical considerations which should accompany its adoption.” The International Baccalaureate (IB), which has a global presence and whose programmes are followed by more than 5,000 schools in 158 countries, has now made amendments to their grading process by using a statistical model to award […]","categories":["AI Features"],"tags":["school"],"author_name":"Ram Sagar","publish_date":"2020-07-08T16:00:13","publication_year":"2020","word_count":1209,"keywords":["Go","machine learning","ELT","AWS","AI","ML","RAG","analytics","GAN","school","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","RAG","AWS","R","Go","ELT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/model-grading-school-international-baccalaureate\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10160785,"title":"Samsung Brings On-Device AI to TVs","content":"Samsung Electronics unveiled Samsung Vision AI at the Consumer Electronics Show (CES) in Las Vegas, United States. Vision AI offers instant information about what is on the screen, live translation, and a ‘Generative Wallpaper’ that transforms the screen into a dynamic art canvas. Moreover, AI can upscale lower-resolution content up to 8K quality, separate and optimise sound components, and enhance the display’s colours. The company also said that Vision AI features are now integrated across Neo QLED, OLED, QLED and The Frame models. The Neo QLED 8K QN990F television leverages the NQ8 AI Gen3 Processor to run AI features locally. Samsung has collaborated with Microsoft, among other AI partners, to deliver these experiences.The new Smart TVs and Smart Monitors will feature Microsoft Copilot. “This partnership will enable users to explore a wide range of Copilot services, including personalised content recommendations,” read the announcement. Samsung also said that it plans to work closely with other leading AI companies, including Google, to ‘expand what Vision AI can do’. Furthermore, Vision AI also positions screens as the ‘central hub’ for the SmartThings, home automation ecosystem. This provides real-time updates about your household environment, keeps an eye on loved ones, and monitors ‘unusual behaviour’ in family members and pets. “With Samsung Vision AI, we’re reimagining what screens can do, connecting entertainment, personalisation and lifestyle solutions into one seamless experience to simplify your life,” said SW Yong, President and Head of Visual Display Business at Samsung Electronics. Besides televisions, Samsung also unveiled an array of new display technologies at CES. The company unveiled Premiere 5, an industry’s first ‘triple-laster ultra short-throw’ projector, essentially a high-quality projector that uses three lasers to project large displays from a short distance. Samsung also announced a ‘MICRO LED Beauty Mirror’, which uses advanced display and sensor technology to reflect your image while analysing your skin. Samsung isn’t the only company in this game. LG also announced its lineup of AI TVs at CES 2025. The company unveiled the 2025 OLED EVO lineup, which includes the evo M5 and evo G5 models and features the second-generation Alpha 11 AI processor. “The 2025 OLED evo models provide enhanced picture and sound quality. Deep learning algorithms meticulously analyse and refine low-resolution and low-quality images, enhancing them to a higher definition with pixel-level precision for natural and sharper visuals,” said LG in the announcement. Like Samsung, LG has also collaborated with Microsoft and provides access to Copilot. AI in LG TVs will also help recognise individual voices, automatically switch profiles, drive more relevant suggestions, and use a large language model to help converse with the interface.","excerpt":"Partners with Microsoft and plans to do so with Google as well.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Samsung Electronics"],"author_name":"Supreeth Koundinya","publish_date":"2025-01-06T14:38:54","publication_year":"2025","word_count":433,"keywords":["Samsung Electronics","Go","programming_languages:R","AI","ML","programming_languages:Go","RAG","automation","Ray","deep learning","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","deep learning","Ray","RAG","R","Go","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/samsung-brings-on-device-ai-to-tvs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10016812,"title":"Top 5 Inductive Biases In Deep Learning Models","content":"The learning algorithms mostly use some mechanisms or assumptions by either putting some restrictions on the space of hypotheses or can be said as the underlying model space. This mechanism is known as the Inductive Bias or Learning Bias. This mechanism encourages the learning algorithms to prioritise solutions with specific properties. In simple words, learning bias or inductive bias is a set of implicit or explicit assumptions made by the machine learning algorithms to generalise a set of training data. Here, we have compiled a list of five interesting inductive biases, in no particular order, which are used in deep learning. Structured Perception And Relational Reasoning Structured perception and relational reasoning is an inductive bias introduced into deep reinforcement learning architectures by researchers at DeepMind in 2019. According to its researchers, the approach improves performance, learning efficiency, generalisation, and interpretability of deep RL models. By introducing structured perception and relational reasoning into deep RL architectures, the reinforcement learning agents can learn interpretable representations and exceed baseline agents in terms of sample complexity, ability to generalise, and overall performance. This approach can also offer advantages in meeting some of the most challenging test environments in modern artificial intelligence. Know more here. Group Equivariance Equivariance is a good inductive bias for deep convolutional networks. The group equivariance in convolutional neural networks is an inductive bias that helps in the generalisation of the networks. It reduces the sample complexity by exploiting symmetries in the networks. Research by the University of Amsterdam showed that the Group equivariant Convolutional Neural Networks (G-CNNs) used G-convolutions, a kind of layer that enjoys a substantially higher degree of weight sharing than the regular convolution layers. This layer increases the expressive capacity of the Convolutional Neural Network without increasing the number of parameters. According to its researchers, group convolution layers are easy to use and can be implemented with negligible computational overhead for discrete groups generated by translations, reflections, and rotations. Know more here. Spectral Inductive Bias Spectral bias is an inductive bias or learning bias in deep networks that manifests itself not just in the process of learning but also in the parameterisation of the model itself. The research was released in 2019 by Yoshua Bengio and his team. In this bias, the lower frequencies are learned first. According to researchers, this inductive bias’s properties are in line with the observation that over-parameterised networks prioritise learning simple patterns that generalise across data samples. Know more here. Spatial Inductive Bias Spatial bias is a type of inductive bias in Convolutional Neural Networks (CNNs) that assumes a certain type of spatial structure present in the data. According to its researchers, spatial bias can be useful even in a model that is non-connectionist and completely linear. This means that applying a spatial bias to other techniques should be both possible and beneficial when spatial data are involved, even in the context of non-connectionist techniques. This type of bias can be introduced with only minor changes to the algorithm, and no externally imposed partitioning is required. Know more here. Invariance and Equivariance Bias Invariance and equivariance bias can be used to encode the structure of the relational data. This kind of inductive bias notifies the behaviour of a model under various transformations. Equivariant models have been successfully used for various deep learning on data with various structures —from translation equivariant image to geometric settings and discrete objects such as sets and graphs. Know more here.","excerpt":"The learning algorithms mostly use some mechanisms or assumptions by either putting some restrictions on the space of hypotheses or can be said as the underlying model space. This mechanism is known as the Inductive Bias or Learning Bias.  This mechanism encourages the learning algorithms to prioritise solutions with specific properties. In simple words, learning […]","categories":["AI Trends"],"tags":["AI biases","artificial intelligence machine learning data","Deep Learning"],"author_name":"Ambika Choudhury","publish_date":"2020-12-30T18:00:00","publication_year":"2020","word_count":573,"keywords":["Go","artificial intelligence","AI biases","machine learning","AI","neural network","programming_languages:R","artificial intelligence machine learning data","RAG","deep learning","Deep Learning","CNN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","RAG","R","Go","CNN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-inductive-biases-in-deep-learning-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25828,"title":"How Machine Learning Has Become A Part Of Every Physicist’s Toolbox","content":"Physics, too, has fallen into the artificial intelligence hype with a clutch of researchers using machine learning to deal with complex problems regarding huge amount of data. ML applications in physics are becoming an important part of modern experimental high energy analyses. From high-energy physics to quantum physics and condensed matter, ML applications are upending physics-based modelling. According to an academic researcher, physicists are mostly using ML to accelerate calculations of atomic energies. ML has also been used for good effect in high energy experiments and observational astronomy. In this article, we are going to discuss numerous applications in condensed matter physics and how physicists are quick to embrace new tech to find breakthroughs. Machines Make Way Into Physics ML In Condensed Matter Physics In fact, breakthroughs in condensed matter physics have been assiduously documented by well-known researcher Juan Felipe Carrasquilla, an expert in theoretical condensed matter physics, and numerical methods who said in an article that the most difficult problem lay in: “Understanding the wavefunction of a many-particle quantum system with relevant accuracy which can pave the way for the designing of new quantum materials and devices.” This led him to think that machines can resolve some of the biggest questions in physics. The article notes that researchers are using ML algorithms to understand the phases of matter, which could lead to theoretical breakthroughs in quantum bit or even silicon. Before the advent of ML, physicists relied on a bunch of tools for deriving observations. For example, the use of macro-models phased out the use of micro-parameters. Statistical modelling was the most basic tool used for deriving inference general patterns from data. Deep Learning for Tracking in High Energy Physics According to an ICLR 2107 research paper, deep learning techniques have been successfully applied to high energy physics. This paper applies deep learning techniques to tracking in high energy physics experiments. The researchers deployed an LSTM approach and proposed an end-to-end solution for the HL-LHC track pattern recognition challenge. The paper also explored the role of recurrent neural networks and long short-term memory networks in modelling these dynamics. In another paper titled, Searching for Exotic Particles in High-Energy Physics with Deep Learning, the researchers demonstrated how through benchmark datasets, deep learning techniques improved the classification metric by 8 percent. ML In Astrophysics Off late, ML methods have found application in astronomy and astrophysics for processing the wide amount of astronomical data generated which are often heterogeneous in nature. ML techniques have been applied for developing ways to process, analyse astronomical data in an automated manner. Other applications include using ML techniques to gain data-driven insights and visualize it further. Other approaches include developing scalable methods for automated learning. In fact, AstroML, ML for Astrophysics, a highly cited research paper introduces ML tools for statistical analysis. The astroML software package is available for free and provides researchers and students with resources like statistical tools and dataset loaders for Python implementation of statistical routines. Universities Training Physicists In Deep Learning Techniques Many universities are introducing computational techniques, especially neural networks used in a variety of applications such as pattern recognition, image recognition, natural language processing. Universities are covering the whole gamut of neural networks like autoencoders, recurrent neural networks, Hinton’s backpropagation technique and also some emerging applications for deep learning in Physics. The preferred programming language is Python and Matlab. Some of the most popular use cases for ML in Physics are analysing phase transitions and making predictions of material properties. Why Use ML When We Already Have Physics-Based Models? Ideally, a computational approach is always the best way to resolve a problem. Traditionally, physicists relied on statistical methods and physics-based model to tackle the question. Given this scenario, should one take an ML-based approach? A key drawback is the high cost of the computational model which can also be very time-consuming. Another aspect is that black box problem of ML where it is hard to understand the biases which can be a setback to validating the results. Another drawback is the huge amount of data required for ML-based methods and the cost associated with collecting data than they can use.","excerpt":"Physics, too, has fallen into the artificial intelligence hype with a clutch of researchers using machine learning to deal with complex problems regarding huge amount of data. ML applications in physics are becoming an important part of modern experimental high energy analyses. From high-energy physics to quantum physics and condensed matter, ML applications are upending […]","categories":["AI Features"],"tags":["machine learning pattern recognition python"],"author_name":"Richa Bhatia","publish_date":"2018-06-28T04:32:39","publication_year":"2018","word_count":690,"keywords":["Go","machine learning pattern recognition python","machine learning","artificial intelligence","AI","neural network","ML","image recognition","Python","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","image recognition","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-machine-learning-has-become-a-part-of-every-physicists-toolbox\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10051769,"title":"Accenture On An Acquisition Spree, Betting on AI, Cloud Companies","content":"IT and consulting giant Accenture, has acquired BRIDGEi2i, an artificial intelligence and analytics firm headquartered in Bangalore, just days back. This is one of the many acquisitions the tech giant has conducted in the last two years. Especially in 2021, it has been on a shopping spree and added many companies, big and small, to its kitty. Taking a closer look at the acquisitions from the last two years’ framework, we can see that most of these have happened in the Accenture Cloud First, Accenture Applied Intelligence, Accenture Industry X and Accenture Interactive domains. This is very well echoed in the company’s Business Future 2021: Signals of Change Research. It said, “Organizations are using data analytics and artificial intelligence to make decisions and define strategies that anticipate the future.” Cloud competition is increasing among tech giants With COVID-19 hitting the world globally and the rapid pace of digital transformation that followed, Accenture, with its latest acquisitions, seems to have realized that the future of business lies in cloud, artificial intelligence and data analytics. The company launched its Cloud First segment in September 2020 with a $3 billion investment plan over the next three years. These acquisitions aim to use the companies’ expertise to help in scaling growth and stay ahead of the competition. In addition, through these acquisitions, it will access the talent from these companies for its clients who are already skilled in providing end-to-end cloud services. Recent Cloud First sector acquisitions are these: BENEXT–Independent product consulting company specializing in product management, agile coaching, cloud-based development and data science; acquired in October 2021.Linkbynet–Cloud services provider headquartered in France, specializing in cloud optimization and managed services, cloud transformation and cloud security; acquired in July 2021.Cygni–Sweden based cloud native full-stack development firm; acquired in March 2021. What are the competitors up to? Other tech and consulting giants are also stepping up in the Cloud space: Last year in August, Infosys launched Infosys Cobalt. It is a set of services, solutions, and platforms to help enterprises in cloud-powered enterprise transformation.Consulting biggie, EY launched a cloud enablement centre for financial services in Phoenix. It will focus on delivering an array of services and solutions for clients with cloud and agile technology capabilities.TCS teamed up with Amazon Web Services (AWS) to launch the new TCS AWS Business Unit (BU). It is a full-stack, multidisciplinary group that offers enterprise customers end-to-end services and solutions around cloud migration, application and data modernization, managed services, and industry-specific innovation leveraging AWS. Betting on AI, Data Analytics It is a well-known fact that the importance of data for today’s businesses has skyrocketed. The acquisition of boutique AI firms will help Accenture improve its AI skills and data science capabilities to give better value to its clients. Sanjeev Vohra, global lead for Accenture Applied Intelligence, said, “The COVID-19 pandemic has made technologies such as AI core to business success, with scaled investments enabling enterprises to thrive by refocusing on growth during the most disruptive time in their history.” Business Future 2021: Signals of Change Research by Accenture also emphasizes the importance of data and AI. It says, to find new patterns in data and better anticipate future decisions, new data sets, which include real-time data from across the value chain, are being processed by new analytic approaches based on artificial intelligence. Following are Accenture’s Applied Intelligence acquisitions: Bridgei2i–India-headquartered AI firm with additional offices in the US and Australia; acquired in October 2021.Core Compete–Cloud analytics services firm headquartered in Durham, North Carolina, with additional offices in the UK and India; acquired in April 2021.Byte Prophecy–An automated insights and big data analytics company based in Ahmedabad, India; acquired in May 2021.End-To-End Analytics–A boutique applied analytics and data science consultancy located in Palo Alto; acquired in December 2020.Pragsis Bidoop–Spanish company with strong expertise in big data, AI and advanced analytics; acquired in September 2019. Expanding footprint in different markets It seems that Accenture wants to increase its foothold in different markets with the acquisitions. These recent buyouts in AI, cloud and cybersecurity are all spread out in important markets like South America, Australia, Europe, Australia, and India, among others. Accenture is trying to capture these important markets to tap on the opportunities and assert dominance over its competitors.","excerpt":"Through acquisitions in cloud-based companies, AI and data science firms, Accenture is solidifying its dominance in these in-demand sectors","categories":["IT Services"],"tags":["Accenture","AI Companies","Data Science","Mergers and Acquisitions"],"author_name":"Sreejani Bhattacharyya","publish_date":"2021-10-18T12:00:00","publication_year":"2021","word_count":705,"keywords":["Accenture","data science","artificial intelligence","AWS","AI","Git","RAG","Ray","Aim","analytics","AI Companies","Data Science","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","Ray","RAG","AWS","R","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/accenture-on-an-acquisition-spree-betting-on-ai-cloud-companies\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10021882,"title":"How SirionLabs Uses AI To Offer Contract Management Solutions","content":"The entire contract process requires a great deal of precision through its various stages–initiation, authoring, workflow, negotiation and approval, execution, compliance, and contract renewal. To avoid human errors and increase efficiency and productivity, several organisations are opting for automated and streamlined Contract Lifecycle Management (CLM) systems. Such systems create a common language, standardise the contract development process, and reduce the time taken by using pre-approved templates and legal clauses. SirionLabs, a SaaS contract lifecycle management (CLM) platform, was named as a leader in CLM solution providers in the Forrester Wave 2021 report. Analytics India Magazine caught up with Aditya Gupta, Chief Architect and Co-founder of SirionLabs, to understand its services and how AI plays a major role in their functioning. AIM: Tell us about the main services and offerings of SirionLabs Aditya Gupta: SirionLabs is a leading SaaS contract lifecycle management (CLM) platform that uses AI to help businesses manage the complete contracting lifecycle on a single, easy-to-use platform. Our platform’s AI core is one of the most advanced in the industry, and we use it to help organisations digitally transform traditional CLM functions such as contract negotiation, authoring, repository, and extraction of contractual data. In addition, our platform also offers post-signature contract management capabilities focused on supplier governance, obligation and service level management, automated invoice validation, and data-driven buyer-supplier collaboration. We have made significant investments towards developing AI’s role in our contract management systems further. Sirion has helped clients unlock significant enterprise value including hard savings of 10-12 percent of the contract value, improved business outcomes, up to 80 percent reduction in customer-supplier disputes, and a 60 percent reduction in manual effort and cost associated with contract creation and governance. AIM: What are the proprietary algorithms and AI\/ML models currently used At SirionLabs? Aditya Gupta: Sirion’s core is built on a foundation that fuses machine learning (ML) and natural language processing (NLP). We use supervised learning coupled with a massive corpus of industry data to train our ML models to recognise and ‘read’ not just paper contracts and PDFs but also handwritten notes. Our image recognition and OCR technologies together play a vital role in digitizing contract documents, which are then semantically parsed by our NLP engine to break them down into machine-readable text. Post digitisation, our platform uses an AI-based expert system to solve complex decision-making problems such as helping users in identifying relevant clauses. Our platform also uses its NLP and ML capabilities to accelerate the legal review process for counterparty papers by automatically ‘reading’ contracts and flagging deviated and missing clauses, and process natural language queries from users, based on which it provides analytical insights and data visualisations. AIM: What comprises the tech stack at SirionLabs? Aditya Gupta: We widely use spaCy, Pytorch, and TensorFlow libraries to support our AI and ML initiatives. Besides these, we use various other libraries for logic building and transformation. Speaking of frameworks, Sirion uses a shared ledger whose architecture is similar to those used in blockchain applications, and we use it to manage the flow of engagement data between counterparties. This framework has helped us position Sirion CLM as a frontend collaboration platform that acts as a single source of truth for performance data for all parties in a contract. Our goal is to help customers end their dependence on traditional siloed backend systems – like ERP and ITSM – that usually host all sorts of critical supplier records. Sirion is a multi-tenant SaaS-based application, built for scale and security from the ground up. It is a modern MVC architecture with some business microservices. It is based on Java, Spring, PostgreSQL and Elasticsearch. While Java is our primary programming language, we also use .Net, Golang and Python. AIM: Tell us about SirionLabs’ M&A solution Aditya Gupta: Merger and acquisitions are becoming an increasingly common way for enterprises to grow and acquire new lines of business or products. However, they are also among the most complex events an organisation can go through. Entire business systems need to be integrated and hundreds of due diligence queries need to be answered. Most companies usually depend on labour and time-intensive manual review processes to identify key risk elements in contracts. Sirion’s M&A solution combines AI-led contract extraction, analytics, and authoring capabilities to automate large swathes of due diligence, risk mitigation, and repapering activities that underpin a typical M&A event. During an M&A, the diligence team can upload contracts and associated documents in bulk and use Sirion’s out-of-the-box adaptors to connect with other enterprises’ IT systems and ingest documents stored in these data silos. Our platform’s AI engine then converts all legacy contracts, PDF files, and even handwritten notes into machine-readable text, which is then semantically parsed to extract key metadata, service levels, obligations, pricing tables, and more. At this stage, a similarity clustering algorithm helps users identify and remove duplicate documents from the entire corpus. The extracted data can then be tagged and batched into logical groups and assigned to human reviewers, who can then normalize the results. After the basic review process, Sirion’s AI can scan through the digitized contracts to highlight missing clauses and deviations, based on which contract managers can choose to repaper or novate inherited contracts to bring them in line with the post-M&A state. AIM: What skills you look for in candidates applying for AI and Data Science-related roles? Aditya Gupta: Currently, we are working on expanding our workforce in the US to drive key AI initiatives out of our new CoE in Seattle. This has given us a clear view of the technical and individual qualities that we would like to have in our engineers. We look for machine learning research engineers who have practical experience in NLP, streaming applications (Kafka, Rabbitmq), programming languages such as Python and R, and deep understanding of statistics, data structures and algorithms, linear algebra, differential calculus, optimization and numerical analysis, advanced mathematical modelling, and comparative KPI analysis. While there is significant overlap between an AI engineer’s and data scientist’s respective skill sets, the latter’s role requires experience in data mining; R, SQL and Python; familiarity with Scala and Java; and extensive working knowledge of business intelligence tools such as Tableau and data frameworks such as Hadoop. We have an inclination towards people with a growth mindset who thrive in a collaborative environment that focuses on solving real-world business problems. AIM: What are your company’s short and long term goals on the tech front? Aditya Gupta: Our current goal is to democratize CLM, which we hope to achieve by launching: Sirion’s AI engine as a standalone self-service solutionA developer platform that will include public APIs, SDKsAn app marketplace for first (created in house) and third-party apps (created by a developer community).","excerpt":"The entire contract process requires a great deal of precision through its various stages–initiation, authoring, workflow, negotiation and approval, execution, compliance, and contract renewal. To avoid human errors and increase efficiency and productivity, several organisations are opting for automated and streamlined Contract Lifecycle Management (CLM) systems. Such systems create a common language, standardise the contract […]","categories":["Deep Tech"],"tags":["Contract","Mergers and Acquisitions"],"author_name":"Shraddha Goled","publish_date":"2021-03-12T11:00:00","publication_year":"2021","word_count":1118,"keywords":["data science","machine learning","AI","TensorFlow","PyTorch","ML","NLP","Aim","analytics","spaCy","Mergers and Acquisitions","Contract"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","data science","analytics","Aim","TensorFlow","PyTorch","spaCy"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-sirionlabs-uses-ai-to-offer-contract-management-solutions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10168459,"title":"Why Companies are Moving Away from Docker","content":"For years, when you said “containers”, you meant Docker. It was the go-to solution for developers who wanted to package, ship, and run applications—basically anything DevOps—consistently across various environments. But here we are in 2025, and the tide is turning. While Docker is still around, it’s no longer the undisputed king of the container world. Developers and companies are beginning to explore alternatives—some cautiously, others full throttle. The shift away from Docker is a symptom of a maturing container ecosystem. bp Technology event is just tomorrow in Bangalore! Register & Attend for Free >> Docker hasn’t exactly crashed. More and more teams are adopting hybrid strategies using Docker for local development, but switching to Podman or containerd for staging and production. Then there are those who are ditching Docker entirely in favour of fully open, modular stacks. Docker doesn’t seem to be evolving fast enough. Let’s start with the licensing drama. A few years back, Docker made the decision to put Docker Desktop behind a paywall for larger organisations. According to a recent blog by Devlink Tips, while individual developers and small teams could still use it for free, enterprise users were now being asked to cough up for something they previously got at no cost, and arguably, without substantial improvements. This didn’t sit well with a lot of teams, especially the open-source crowd and budget-conscious startups. It forced many to rethink their dependencies and ask the uncomfortable question: “Is Docker really worth it?” This is also highlighted by several discussions on Reddit where people are questioning the viability of Docker. To some extent, it does not make sense for an enterprise to ask this question just because a tool becomes paid, as people express in the discussions, that it should be something they should be willing to pay for. But Price Isn’t the Only Issue Docker Desktop needs to emulate a Linux environment through virtual machines, and that’s where things get messy for any system not running Linux. Builds slow down, CPUs heat up, fans start to scream, and battery drains on Windows and macOS. Security, too, has become a real concern. Docker relies on a daemon that runs with root privileges. While Docker has tried to patch things up over time—introducing user namespaces and rootless mode—it still feels like security was an afterthought rather than a design principle. Alternatives like Podman, which runs without a central daemon and can operate entirely rootless, are built with security as a core feature. For companies in regulated industries or multi-tenant environments, that kind of architecture is essential. Then there’s Docker’s architecture itself. The cloud-native world has moved toward specialisation and modularity. Today, Kubernetes is the default orchestrator for many, Helm handles packaging, and runtimes like containerd focus solely on container lifecycle management. Recent developments indicate that Docker is adapting through improved Docker Hub features and enhanced Kubernetes support, but the platform now competes in a market where 36% of developers use cloud-based container tooling. Some started moving away from Kubernetes for similar reasons, resulting in further decline. The February 2025 Nucamp analysis revealed enterprise trends where 68% of organisations are adopting multi-cloud strategies, which require cloud-agnostic tooling. Furthermore, Docker Swarm sees a 42% increase in adoption for single-cloud deployments. Docker initially focused on its own orchestration tool, Docker Swarm, rather than embracing Kubernetes, which has since become the dominant container orchestration platform. This strategic misstep led to Docker losing ground in enterprise container orchestration, pushing companies to adopt Kubernetes and alternatives instead. These shifts do not eliminate Docker, but rather reposition it as one component in layered architectures that combine multiple container technologies. There is also growing concern about vendor lock-in. While Dockerfiles are widely used, they’re not governed by an open standard like OCI (open container initiative) image specifications. If you’re betting big on Docker-specific tooling, you’re potentially locking yourself into an ecosystem that might not play nicely with the rest of the industry five years down the road. Quite a Few Players to Fill the Gap Podman is a favourite among teams focused on security and compliance. Built by Red Hat, it offers nearly identical CLI commands to Docker, which makes switching a breeze. Then there’s containerd, the container runtime that was once part of Docker and has since taken on a life of its own under the Cloud Native Computing Foundation. Kubernetes now uses containerd by default, following the deprecation of Docker support in version 1.24. It’s light, fast, and laser-focused on just one job: managing containers. This makes it ideal for production workloads across cloud platforms like AWS, GCP, and Azure. CRI-O is another lean and mean container runtime, built specifically for Kubernetes. It strips out unnecessary features to ensure compliance and security. It’s the default runtime for Red Hat OpenShift and a rising star in environments where minimalism and performance are key. It appears that some companies are shifting away from Docker due to its missed orchestration leadership, business model issues, security and complexity concerns, erosion of community confidence, feature bloat, and the emergence of more secure and efficient alternatives. But Docker might catch up again soon, that is, if it wants to. And clearly, it is still the ideal choice for many, at least the ones not running Kubernetes.","excerpt":"Docker hasn’t exactly crashed. More and more teams are adopting hybrid strategies.","categories":["AI Features"],"tags":["Docker"],"author_name":"Mohit Pandey","publish_date":"2025-04-23T14:58:40","publication_year":"2025","word_count":874,"keywords":["Docker","Go","GCP","AWS","AI","R","docker","GAN","DevOps","Azure","kubernetes"],"extracted_tech_keywords":["AI","AWS","Azure","GCP","kubernetes","docker","R","Go","DevOps","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-companies-are-moving-away-from-docker\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10076145,"title":"Draft Telecom Bill Could Kill WhatsApp’s End-to-end Encryption","content":"For over two decades, telcos in India have argued that messaging services like WhatsApp should be regulated. Their campaign ‘same service, same rules’ might have finally borne fruit with the new Draft Telecom Bill. Last week, the Department of Telecommunication introduced the Indian Telecom Bill 2022, which will bring over-the-top (OTT) platforms such as WhatsApp, Zoom, Netflix and others within the ambit of telecom services. “Based on the consultation process, we will create the final draft, which will then go through the committee processes of the Parliament. Then it has to go to Parliament. I see a timeline of 6-10 months, but we are not in a hurry,” telecom minister Ashwini Vaishnaw said. The new Bill will overhaul the existing framework, which is governed by three different acts – The Indian Telegraph Act 1885, The Wireless Telegraph Act 1933 and The Telegraph Wires Unlawful Possession Act 1950. So, what does the new Bill mean for WhatsApp? End of WhatsApp Encryption? In 2016, WhatsApp, owned by Meta, announced that its messages were fully encrypted, and only the user and the person he\/she is communicating with, can read those messages. “End-to-end encryption ensures only you and the person you’re communicating with can read or listen to what is sent, and nobody in between, not even WhatsApp,” the company says. WhatsApp uses an open-source encryption standard from Open Whisper Systems, also used by whistleblower Edward Snowden. However, the new Draft Telecom Bill will give the Central government the power to intercept any messages sent through WhatsApp in matters related to national security. As per Section 24 (2) of the Bill, “on the occurrence of any public emergency or in the interest of public safety, the central government or a state government or any officer specially authorised in this behalf by the central or a state government, may” intercept the encrypted messages and calls made on WhatsApp or other similar applications. Same service, same rules One of the arguments repeatedly made by the telcos is with regard to licensing and regulations. While the telcos have to clear multiple hurdles in terms of licensing and are governed by numerous laws, internet service providers are easily able to bypass all of these. Experts believe clubbing OTT services with telcos is wrong. While  telcos own and manage the whole infrastructure layer, WhatsApp merely uses the internet, but does not control it. “Arguments by telcos calling for ‘same service same rules’ are misconceived given the inherent structural differences between telecommunications service providers and OTT players – such as access to spectrum, ability to interconnect with [public telephone networks], use of numbering resources etc,” the Broadband India Forum said in a filing with the Telecom Regulatory Authority of India (TRAI). Once the Bill comes into force, WhatsApp might be required to apply for a licence to continue offering its services in India. WhatsApp will also have to ask its users to fill out a Know Your Customer (KYC) form for them to continue using its services. As per the Bill, anyone who provides wrong KYC details could face imprisonment for up to one year. According to Vaishnaw, this will help curb cyber fraud. The KYC requirements are not just limited to WhatsApp; even if you are using Google Meet, Zoom, Netflix, Gmail or even Tinder, you might have to fill out a KYC form. WhatsApp’s feud with the government The new Telecom Bill is not the first attempt by the government to try and tame WhatsApp. In 2021, when the new IT rules came into force, WhatsApp took the Indian government to court as the new rules mandated social media platforms to trace the origin of the messages sent on their platforms. A WhatsApp spokesperson said that mandating social media platforms to trace chats fundamentally undermines people’s right to privacy. In response, later that year, the government filed an affidavit stating that “the constitutionality of a provision of law cannot be challenged by a foreign commercial entity on the ground of it being violative of Article 19”. While there is still time for the new Telecom Bill to come into effect, it would be interesting to see which direction the litigation between WhatsApp vs government takes as it could have implications on the future of WhatsApp and how it operates in India. Ploy for surveillance Serious concerns about users’ right to privacy have already been raised by different stakeholders of the industry and members of civil society. “The Draft Telecom Bill is not only a cause for disappointment, but alarm. It maintains colonial control and visualises citizens as subjects by ignoring constitutional developments on privacy and free speech,” Apar Gupta, the executive director at the Internet Freedom Foundation, said. Over the years, WhatsApp has emerged as a major source of communication by individuals as well as businesses all over the world. Will Cathcart, head of WhatsApp, stated in 2020 that more than 100 billion messages are sent each day on WhatsApp. As of June 2021, there were around 470.1 million WhatsApp users in India, as per Statista. Number of WhatsApp users as of June 2021 (Source: Statista) Vaishnaw has clarified that the telecom Bill doesn’t force the decryption of messages even though there is a provision for their interception. However, not everyone is convinced. The draft Telecom Bill 2022 has clear focus on Protection of users and light touch regulation. pic.twitter.com\/q3IvrXwMqz— Ashwini Vaishnaw (@AshwiniVaishnaw) September 23, 2022 Different civil groups from Access Now and Internet Freedom Foundation have argued that the Bill is an attack on end-to-end encryption. They believe that law enforcement authorities could spy on citizens under the garb of protecting them from cyber frauds. Further, the Bill grants the government the right to shut down the internet, which is again a blatant attack on human rights, including the rights to free expression, information, and freedom of assembly, civil rights group Access Now said. India's Draft Telecom Bill threatens end-to-end encryption + authorises easy imposition of #InternetShutdowns. Needs changes to protect privacy, security and free expression in sync with the Constitution and Supreme Court rulings. Please consider submitting comments by Oct 20! https:\/\/t.co\/KAl12q4JfI— Namrata Maheshwari (@NamrataM_) September 23, 2022 Further, KYC for using services such as WhatsApp or Zoom would mean users can no longer use these services anonymously. Given an individual enjoys the right to freedom of expression and right to privacy, it should entail anonymity on the internet.","excerpt":"The Bill will give the government the power to intercept any messages sent through WhatsApp in matters related to national security","categories":["AI Features"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2022-09-30T17:00:00","publication_year":"2022","word_count":1063,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","AWS","R","Go","ViT","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/draft-telecom-bill-could-kill-whatsapps-end-to-end-encryption\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10648,"title":"Narayana Health launches Robotic Surgery Institute in support with Infosys Foundation","content":"Narayana Health, onestop healthcare destination, has launched the Institute of Robotic Surgery supported by Infosys Foundation at its flagship unit at Narayana Health City. Headquartered in Bengaluru, Narayana Health will be using the da Vinci Robotic Surgical System primarily for prostate, kidney, gynecological, colorectal and select head & neck cancer surgeries. Signifying the launch of ‘Infosys Institute of Robotic Surgery’, Mrs. Sudha Murthy, Chairperson – Infosys Foundation said, “The need to adopt and continually update treatment protocols that reduce errors is crucial for a country like India, which sees high patient volumes and a wide spectrum of complex diseases. Robotic surgery, with its high degree of precision and faster recovery time, has the potential to address this efficiently. Our partnership with Narayana Health, is aimed at encouraging rapid adoption of robotics in healthcare in India and to enable the masses to reap the benefits of affordable and high-quality treatment.” Technique of surgical intervention on the human body is undergoing dramatic change. It has taken over 30 years for laparoscopic surgery to replace open abdominal operation. “World is on the threshold of a major transition from laparoscopic surgery to robotic surgery”, said Dr. Devi Shetty, Chairman –Narayana Health “The philosophy of creating the institution is to train any surgeon with a passion to learn robotic surgery and certify them to start robotic surgical program in different parts of the country. NH Foundation along with Infosys Foundation strongly believes that this is the only way robotic surgery services will be available to the common man of this country”. Explaining the learning curve, Dr. Saurabh Bhargava, Head of Urology Department, Consultant Urologist and Uro-Onco Surgeon at Narayana Health City said, “Our training program is mentor-based rather than building patient base to have a learning experience. We have acquired skills, so that we have competence to train others, that’s what needs to be done. So the learning experience takes place with a mentor in place at Narayana Health”. A formalized mentorship program has been put in place which all surgeons need to undergo. The Robotics team has been through an extensive clinical education, training and preparation period. The training under this program will be provided in a phased manner. Phase 1 includes basic training explaining the concepts, technology, features and functionalities. The operating surgeon has to first understand the technical aspects of its work and then the surgeon will attend a simulator workshop and practice on mannequins, followed by a period of observation. Then with a mentor in place the surgeon will begin with simple techniques and then gradually move to more complex surgeries. Highlighting the benefits of the da Vinci Robotic Surgery System, Dr. Saurabh Bhargava, said, “The access to the organ of interest can be done with key-hole procedures, but the robotic arms and instruments can be moved in such a manner that mimics your hand; compared to a surgeon’s hand, the robotic arm gives far greater flexibility and access while performing surgery. In Robotics surgery, the instrument’s arms are so well designed that you can maneuver in any direction to reach the access point with precision, so the results are functionally much better because you get a clear view”. At Narayana Health City the da Vinci Robotic Surgery System is installed in a dedicated operating room with dedicated and welltrained surgeons, assistants and staff exclusively for Robotic Surgery capable of performing complex surgeries in Urology, Gynecology, Gastrointestinal surgeries, General Surgery and various types of cancer treatments. “Robotics surgery is an amalgamation of open surgery and laparoscopic surgery. Use of the da Vinci system raises the standard of care for complex surgeries – translating into better care and numerous benefits for our patient, which is ultimately our purpose for utilizing this technology,” said Dr. Ashwinikumar D. Kudari, Senior Consultant, Surgical Gastroenterologist, Narayana Health City.","excerpt":"Narayana Health, onestop healthcare destination, has launched the Institute of Robotic Surgery supported by Infosys Foundation at its flagship unit at Narayana Health City. Headquartered in Bengaluru, Narayana Health will be using the da Vinci Robotic Surgical System primarily for prostate, kidney, gynecological, colorectal and select head & neck cancer surgeries. Signifying the launch of […]","categories":["AI News"],"tags":[],"author_name":"Manisha Salecha","publish_date":"2016-08-20T07:50:51","publication_year":"2016","word_count":632,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","RAG","Ray","Aim","GAN","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","API","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/narayana-health-launches-robotic-surgery-institute-support-infosys-foundation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":39353,"title":"5 Things To Consider Before Opting For A Secure Web Gateway","content":"Over the last few years, IT has undergone a tremendous transformation.  Today, from infrastructure to applications to data — almost everything— is moving to cloud. Whether it is either the public or private cloud infrastructure, cloud technology has revolutionized the IT ecosystem. However, today this is the same thing is throwing questions globally about how to protect the data that is being stored there. This emergence of cloud has also changed the way employees used to work; many feels, it has made a lot of people more careless about the security of their as well as the organisation’s data. You just can’t deny the fact that when an employee works outside of the corporate network, they don’t even bother to turn on the VPN and work. And this is where Secure Web Gateways comes into the scenario. What Is A Secure Web Gateway? A secure web gateway (SWG) basically refers to content-control software. When we say content control, it means that this particular software filters and controls content on the internet. This software basically prevents malicious internet traffic to move through the corporate network, ensuring that the company’s network is secured. Simply put, it basically delivers content that is relevant to work or doesn’t violet company’s policies to the user sitting behind the screen. Over the past couple of years, SWG has become of the go-to tools for organisations across the world. It is not something that is very new — SWG has been in the space almost since the beginning of the Web. However, today it is more advanced, much more than just filtering content, and comes in both cloud and on-prem forms. However, there are instances that show that not every SWG is capable of preventing or mitigating malicious traffic and not every company knows how to make the most out of SWG. In this article, we are going to have a deep look at some of the most critical points about Secure Web Gateways. Things To Keep In Mind Before Opting For A SWG You should have complete know-how about the web-related threats and vulnerabilities This is the first and foremost thing every organisation should do. Companies need to understand the threats and vulnerabilities they are facing. They also need to make sure the path and source of the threat and what damage they are causing and could cause in the future. When you have strong knowledge about what you are going to deal with, you plan better. And when you plan better, you come up with strong solutions. So, before evaluating or opting for a specific secure web gateway, you should know what is happening. What is the measure you are already taking? When you are done analysing the threats and the vulnerabilities, next is to take a look at the existing measures that you have taken already or the tools you have set up to deal with malicious traffic. Check each and every tool and take an in-depth look at the results that these tools are delivering. Whether they are able to mitigate or eliminate the risk completely? Or at least to a great extent? Are the results good enough to keep your business safe and secure? Where the tool is lagging behind? Get answers to these questions and if you feel, that they are not working the way they were supposed to, then look for a top-notch SWG product that can solve the problems. Do you have the infrastructure or resources to deploy one more security tool You might feel a high level of need to deploy an SWG product in order to make your web security infrastructure stronger, but one simply can’t buy an SWG product and get it fit in — you have to make sure that you have the required infrastructure and resources to make the most out of the tool. If you don’t have the resources and the infrastructure, check if you can set that up and how much it is going to cost. If it exceeds your budget, you can take a look at some of the cloud service providers. It is always considered to be good practice to have look at our existing resources before deploying an all-new tool. Will the cloud work well with your existing infrastructure? The cloud approach might eliminate on-prem challenges, but it has its own set of requirements. So, if you decide to opt for cloud infrastructure, make sure your existing processes and methods work well. Also, ensure that you have the required support for cloud-centric deployment. That is about the infrastructure. Now, talking about the tools, when you deploy a cloud-based security tool, you have to see whether integrates with the existing on-prem tools. If you have the resources to overcome these challenges, then definitely a cloud-based SWG is a great option to eliminate cyber threats and malicious traffic from the corporate network. What are the results you are looking for and Will the SWG product deliver it? It is the last but one of the most important things to consider. You have to decide what problems you are looking to fix —kind of threats you are looking to detect and eliminate, kind of traffic you are looking to block etc. When you have a vision or a set of results that you are expecting, you can go with the evaluation of the SWG product and see whether that product has the capability to deliver the results. There is no point spending time and money on a product (even if it’s top-notch) if it doesn’t deliver the desired results. Because, at the end of the day it is about the company’s data protection and cyber security infrastructure, which should always on the top of the priority list.","excerpt":"Over the last few years, IT has undergone a tremendous transformation.  Today, from infrastructure to applications to data — almost everything— is moving to cloud. Whether it is either the public or private cloud infrastructure, cloud technology has revolutionized the IT ecosystem. However, today this is the same thing is throwing questions globally about how […]","categories":["AI Trends"],"tags":["analog neural networks","Cyber Security","Cyber security best practices","hacking","Network security"],"author_name":"Harshajit Sarmah","publish_date":"2019-05-17T12:23:38","publication_year":"2019","word_count":955,"keywords":["Go","hacking","Cyber Security","programming_languages:R","programming_languages:Go","Cyber security best practices","Network security","GAN","R","analog neural networks"],"extracted_tech_keywords":["R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-things-to-consider-before-opting-for-a-secure-web-gateway\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":38239,"title":"If Tesla’s Crazy Big AI Claims Come True, Uber May Have A Lot To Worry About","content":"Tesla announced its robotaxi service that can be seen as a competition for Uber, Lyft, and Ola in India. Elon Musk talking at the Tesla Autonomy Investor Day said that he expects to put 1 million Tesla robotaxis on roads by the end of 2020. The fleet of cars will be a mix of Tesla owned cars and vehicles that are put on the fleet by Tesla’s customers. Tesla would be giving its customers a massive opportunity to monetize the downtime of their vehicles by giving them an option to make the car available for ride sharing. Drivers? Not needed. A Tesla will be self-driven and will complete ridesharing trips and safely return to your garage and be ready for you whenever you want to use it. If this sounds too ambitious, Musk has another zinger for you. “Probably, two years from now, we’ll make a product that has no steering wheels and pedals, and if we need to accelerate that time, we’ll just delete parts. It’s easy,” he said at the recently concluded Tesla Autonomy Day. As Musk and Tesla Motors contemplate a future where cars have no steering wheels and your autonomous car will be working part-time as a taxi, many are left wondering if this can even become a reality. Ambitious Plans For The Future Musk also managed to spring a surprise on Uber and Lyft that are already struggling to cut profit. If Tesla manages to pull off the robotaxi service, it will be a hell of a competitor for the traditional ride-hailing services because of a massive advantage: autonomy. Autonomy in cars will help Tesla to make money just by using the already sold cars’ downtime. Tesla plans to take a 30% share of all the money a vehicle makes for their owners. On a scale of Uber or Lyft this economics should work favorably for Tesla. Tesla wants to make it impossible for people to buy anything else other than a Tesla autonomous car. Buying anything else would be like owning a horse in the car era. “It’ll be like owning a horse in three years. I mean, fine if you want to own a horse, but you should go into it with that expectation,” Musk says. There are several reasons why customers would also prefer a Tesla over an Uber cab. The experience of a ride given by an autonomous car, the reduced cost per ride because of autonomy and the brand itself are enough reasons for people to switch to Tesla’s robotaxi service. Tesla calls the system that drives the whole autonomy in its car as “Full Self Driving Computer.” This says a lot about the thought process that went behind designing the car. The computer itself promises not to take more than the size of a glove box when fitted into the car. Will Musk’s Day Dreams Come True? Elon Musk would be the first one to admit that he sets and often misses ambitious deadlines. 2020 seems to be too ambitious a target if you ask many analysts. But at the same time, the speed and execution capability needed to pull off such a project is only available at Tesla. In short, if this vision of the near future if possible can only be realised at Tesla. Musk has definitely put together an A team to fulfill the grand vision. One of the best minds in chip engineering Peter Bannon heads the hardware design efforts. Andrej Karpathy, arguably the best computer vision scientist in the world heads the AI division. These gentlemen and many more make Tesla a good bet to pull off the level 5 autonomy for cars. It seems like there is a good possibility that the project will come to reality later if not next year. The science is heading towards making this possible and the market is also getting warmer to the idea of fully autonomous vehicles. Uber is also focusing on AI-driven cars and has raised a massive 1 billion in investment to aggressively up the ante in the market. The investment in Uber went to Advanced Technologies Group inside Uber. The goal of the investment is to reach scale and reach very low hardware and software costs. Uber CEO Dara Khosrowshahi recently appointed a new CEO for the autonomous division along with a new board. Eric Meyhofer, currently the head of ATG, is the top brass now, elevated to the top role as CEO. “The development of automated driving technology will transform transportation as we know it, making our streets safer and our cities more livable. Today’s announcement, along with our ongoing OEM and supplier relationships, will help maintain Uber’s position at the forefront of that transformation,” Khosrowshahi said. A fleet of cars can be the only advantage Uber can currently bank on. With increasing interest, Tesla’s sales and consumer readiness in terms of putting their cars to work can tilt the scale in favor of the automotive giant. Uber is rightly doubling down on its efforts to focus on autonomy and create an intelligent fleet of its own, but there is a feeling that it is a tad bit late and has a lot of catching up to do.","excerpt":"Tesla announced its robotaxi service that can be seen as a competition for Uber, Lyft, and Ola in India. Elon Musk talking at the Tesla Autonomy Investor Day said that he expects to put 1 million Tesla robotaxis on roads by the end of 2020. The fleet of cars will be a mix of Tesla […]","categories":["AI Features"],"tags":["Autonomous Vehicles","Tesla","Uber"],"author_name":"Abhijeet Katte","publish_date":"2019-04-25T07:01:59","publication_year":"2019","word_count":866,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","computer vision","RAG","Uber","Autonomous Vehicles","Tesla","R","ai_applications:computer vision"],"extracted_tech_keywords":["AI","computer vision","RAG","R","Go","RPA","programming_languages:R","programming_languages:Go","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/if-teslas-crazy-big-ai-claims-come-true-uber-may-have-a-lot-to-worry-about\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10021202,"title":"How FMCW Is Transforming Lidar Technology","content":"Aurora Innovations, an autonomous vehicle company, based out of Silicon Valley, recently announced its acquisition of OURS Technology for an undisclosed amount. The latter is a startup founded by a team of researchers from the University of California, Berkeley, specialising in Frequency Modulated Continuous Wave (FMCW) and ‘lidar-on-a-chip’. This is Aurora’s second acquisition of a Lidar development company, the first being Montana based startup Blackmore in 2019, which also specialised in FMCW technology. The word Lidar is an acronym that stands for Light Detection and Ranging, which is practically a summary of the technology in three words. The technology works on multiple pulses of light to create a 3D map of the terrain surrounding the sensor. Lidar and FMCW have been around for a while, but the latter has only recently become a viable advancement for the industry, making it a popular choice for most autonomous vehicle companies. On the other hand, Tesla, the auto giant working on self-driving automobiles, rejected Lidar and works on its own technology instead. So why did Tesla choose to push Lidar aside, and how does FMCW and Aurora’s acquisition make a difference? Tesla & the drawbacks of Lidar With all the positive investments in Lidar, the technology is not without its critics. Most notably, billionaire Elon Musk called Lidar a ‘fool’s errand’ in 2019, and he is justifiably concerned about the problems it currently has. Musk’s company Tesla, which recently registered its Indian subsidiary in Karnataka, is well known for the ‘AutoPilot’ feature in their electric vehicles. Unlike most autonomous vehicle companies, Tesla vehicles have proprietary Full Self-Driving chips that utilise artificial intelligence through dedicated neural network accelerators to improve the autonomous driving experience. Even without Lidar sensors, Tesla is one of the industry’s key players due to a combination of its array of cameras, radar and sonar sensors, and a constant stream of real-time usage data from its large customer base. Musk’s reason for rejecting Lidar was the high cost of production which goes against Tesla’s aim of producing affordable electric vehicles — however, Lidar has become significantly cheaper with time. That isn’t to say that the technology is free of other faults. One of the main disadvantages is Lidar’s difficulty in perceiving and differentiating between moving and ‘adversarial’ objects. Tesla’s AI neural network and hardware work together to closely imitate actual human vision, and so it can ‘see’ those objects. Tesla’s customer base also gives it a key advantage of a wealth of real traffic and driving data, unlike Lidar vehicles that are not commercially available yet. FMCW & Lidar-on-a-chip Companies like Aurora are aware of the shortcomings of the technology and have been looking beyond conventional Lidar into improved variations that can increase efficiency and safety. The most popular of these so far is the Frequency Modulated Continuous Wave Lidar. FMCW Lidar systems emit a continuous stream of light and can measure larger distances with near-instant velocity. This allows the company to measure the speed of the objects and detect them coming towards or going away from the vehicle, which partially solves the biggest disadvantage of Lidar. FMCW systems are also less prone to interference due to sunlight, and the increase in efficiency doesn’t come at the expense of power either, as the light emitted is of low power. However, a drawback of these systems is the size — most manufacturers need to use large mirrors along with the sensor and range finder to increase the field of view and enable 3D lidar scanning, increasing costs for mass production. This is where OURS Technology’s lidar-on-a-chip comes into play. The company has created a compact, entirely solid-state scanning system utilising FMCW Lidar technology, a game-changer for Aurora. Within just four years, they claim to have combined all the parts needed for an FMCW Lidar system on a single, small chip that can be mass-produced at a low cost. It does not compromise on performance either and is said to be capable of enabling ‘5D vision in every pixel’. If the chip lives up to its claims, then many of Lidar technology’s drawbacks could very well be eliminated. Beyond Aurora & the future of Autonomous Driving Other competitors in the space have also been developing their Lidar capabilities either in-house or from third party suppliers, and in some cases, with acquisitions like Aurora. Amazon’s Zoox uses external suppliers while Alphabet’s Waymo and Russian tech giant Yandex produce their sensors. On the other hand, Argo AI acquired Princeton Lightwave in 2017, and General Motors’ Cruise LLC similarly acquired the startup Strobe. While Tesla does pose serious competition to Lidar’s role in the autonomous vehicle industry, the popularity of the technology among all the other players with increasingly more efficient improvements means that it is here to stay for now and shape a driverless future.","excerpt":"Aurora Innovations, an autonomous vehicle company, based out of Silicon Valley, recently announced its acquisition of OURS Technology for an undisclosed amount. The latter is a startup founded by a team of researchers from the University of California, Berkeley, specialising in Frequency Modulated Continuous Wave (FMCW) and ‘lidar-on-a-chip’. This is Aurora’s second acquisition of a […]","categories":["AI Features"],"tags":["autonomous car","autonomous car technologies","autonomous cars","autonomous cars India","Autonomous Vehicles","Elon Musk","LIDAR","Self Driving Cars","Self-driving car","self-driving cars India","Tesla"],"author_name":"David B. Shrestha","publish_date":"2021-03-03T13:00:00","publication_year":"2021","word_count":800,"keywords":["Self Driving Cars","Ray","Autonomous Vehicles","R","artificial intelligence","Elon Musk","Tesla","startup","Go","AI","neural network","autonomous car technologies","autonomous cars","LIDAR","programming_languages:R","Self-driving car","innovation","self-driving cars India","Aim","autonomous car","autonomous cars India"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","Aim","Ray","R","Go","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-fmcw-is-transforming-lidar-technology\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011761,"title":"NSE Academy Acquires Talentsprint To Explore Deep Tech Education","content":"NSE Academy Limited recently announced the acquisition of Hyderabad-based Deep Tech education firm TalentSprint Private Ltd. to foray into professional education in the future tech. NSE Academy is a wholly-owned subsidiary of the National Stock Exchange Limited and promotes financial literacy as a necessary life skill. With this acquisition, NSE Academy aims to be a leader in the education segment offering executive and corporate learning not only in financial domain but also adjacent areas in emerging technology such as artificial intelligence, machine learning, blockchain, and more. The courses which will we be conducted in both online and offline modes aim to address the increasing demand for skill acquisition and upgradation in domains such as BFSI. With program offerings in AI, Machine Learning, Computational Data Science, FinTech, Cybersecurity, robotic automation and various deep-tech subjects, Talentsprint is highly sought-after by aspiring and experienced professionals to train in these areas. It also has a deep and collaborative partnership with prestigious academic institutions and global corporations. The company’s digital learning platform also powers online campuses for major academic institutions. “We are delighted and privileged that NSE Academy is entering the space of deep tech education with TalentSprint as its strategic partner. We share a common vision of how digital platforms are enablers of trust, quality, and scale, and how important it is to rapidly disseminate high-quality deep tech expertise to all professionals across BFSI, Technology, Consulting, and emerging sectors,” said  Dr Santanu Paul, Co-Founder and CEO, TalentSprint in an official press statement. He further added that it will help six million working professionals in India in need of deep tech knowledge interventions, and another ten million college students waiting to enter the workforce with adequate future-proof skills. Commenting on the announcement, Vikram Limaye, MD and CEO, NSE said, “As the BFSI industry evolves into a more tech-enabled industry, it is but imperative that we strengthen our expertise in the education space by adding new and emerging technologies. TalentSprint’s vision and portfolio offerings complement our growth ambitions in this space and we are very excited about this partnership”. Abhilash Misra, CEO NSE Academy believes that the strategic partnership between NSE Academy and TalentSprint will create unique possibilities within the professional education ecosystem. With the vision to democratise the access to financial learning, NSE Academy has also launched its AI-Powered Learning Experience Platform-NSE Knowledge hub, which brings in global content for up-skilling in the areas of financial markets and BFSI domain.","excerpt":"NSE Academy Limited recently announced the acquisition of Hyderabad-based Deep Tech education firm TalentSprint Private Ltd. to foray into professional education in the future tech. NSE Academy is a wholly-owned subsidiary of the National Stock Exchange Limited and promotes financial literacy as a necessary life skill. With this acquisition, NSE Academy aims to be a […]","categories":["AI News"],"tags":["talentsprint"],"author_name":"Srishti Deoras","publish_date":"2020-11-17T16:31:33","publication_year":"2020","word_count":405,"keywords":["data science","API","machine learning","artificial intelligence","AI","Git","talentsprint","Ray","Aim","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","Aim","Ray","R","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nse-academy-acquires-talentsprint-to-explore-deep-tech-education\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":33754,"title":"How Data Analytics Backed By AI And ML Is Transforming The BFSI Sector","content":"Stephen Hawking, the renowned theoretical physicist once said, “It’s tempting to dismiss the notion of highly intelligent machines as mere science fiction.” The reality is that most AI applications do not have a physical form, but rather “live” in lines of code. The term “AI” includes all technology used to mimic human intelligence, typically falling into one of three subcategories: machine learning, natural language processing and cognitive computing. The business world is getting transformed and changing rapidly with the digital disruption backed with insight and automation opportunities gained out of Artificial Intelligence and Machine Learning (AI\/ ML) enabling Data Analytics. “When Amazon recommends a book you would like, Google predicts that you should leave now to get to your meeting on time, your bank stopping a fraudulent transaction on your credit card, and UBER magically get car of your choice at your doorsteps, these are examples of machine learning over a Big Data stream.” The financial industry is a highly regulated and data-intensive industry. With the invent of new entrants like FINTECHs, digital and payment banks, new regulations and change in customer behaviour traditional banking system is facing and feeling an external disruption and tension to reinvent itself and critically examine its business processes to not only get more clients but how to enhance existing customer experience. The banks are exploring, experimenting and investing in the Data Analytics use cases backed by Artificial Intelligence and Machine Learning (AI\/ ML). Below are the five banking areas facing disruption as the banking industry rapidly adopts AI: Customer Experience The banks are focussing on growing the top line by adding custom services and offering better channel experience to customers. Many banks are introducing chat bots backed by AI abilities which can understand the emotions of the customer analysing their voice and facial expressions and converse accordingly. Through big data and machine learning, these “bots” know how to respond to customer’s questions – from onboarding concerns to transaction-specific questions. Additionally, the technology is capable of managing customer requests and making product recommendations. The key benefit of this advancement is to attract the tech-savvy millennials, known to prefer less human interaction when it comes to financing. Another benefit is to promote less pushy contextual nudges and next best action that customer could take on the basis of their geographical location and the latest financial interaction. Investment Advisory – Digitization of Advice Digital-advisors are changing the investment landscape, with AI-powered platforms automating relationship heavy private banking and asset management field. Introduction of Digital-advisors almost entirely eliminates financial advisors and relationship managers from the investing process. Investors no longer have to shell out wealthy or pay hefty fees for something they might not want or need. Now, Digital-advisors collect information about an investor’s financial goals and the level of risk they’re willing to incur, further this data is integrated with the macroeconomic data then they input this data into algorithms (with quantum computing we will be able to run more sophisticated algorithms in future). In turn, the results are used to offer investment advice to the individual, allowing him\/her to make educated investment decisions, or, in many cases, the digital-advisor will fully automate the purchase and management of investments. Fraud detection and risk management AI can detect fraud before it happens. Technology can rapidly mimic the thought process of a human analyst to review each transaction in every portfolio at a bank (big data stream). AI enables banks to not only be alerted to potential fraud but also gives them a percentage that depicts the likelihood of a card ever becoming compromised. The appropriate usage of machine learning algorithms could result in the reduction of false positives which not only improves the efficiency of the AI\/ ML Fraud Detection process but also helps in improving customer satisfaction. Regulatory Compliance The financial industry is a heavily regulated industry with the rapid evolution of the laws and regulatory mandates. AI can “learn”, remember, and comply with all applicable laws – from KYC and anti-money laundering regulation to laws governing asset management. This results in the elimination of human errors, identification of complex patterns and financial institutions to meet their regulatory obligations. Equity Predictions When it comes to big data and pattern recognition, humans are no match for AI. Analysts, investors and other key players on Wall Street have embraced AI as a tool for predicting market movements. AI can evaluate companies’ public remarks (such as on earnings calls), picking up on sentiment analysis (word usage, speech patterns, etc.), which then is used compared with historical data to predict stock performance with near certainty. This advancement benefits the financial institutions and customer simultaneously by taking off the human bias and restriction of sampling as the whole of data (Big Data) is analysed for making such recommendations and predictions. AI’s disruption in finance is increasing exponentially and it’s geared up for greater economic impact than ever. From better customer experience to extending investment opportunities to the common man to predicting fraud to mitigating investment risks, AI has the potential to not only revolutionize the industry but also to improve the financial health of millions of people in the world.","excerpt":"Stephen Hawking, the renowned theoretical physicist once said, “It’s tempting to dismiss the notion of highly intelligent machines as mere science fiction.” The reality is that most AI applications do not have a physical form, but rather “live” in lines of code. The term “AI” includes all technology used to mimic human intelligence, typically falling […]","categories":["AI Features"],"tags":["AI in banking","AI in finance","BFSI"],"author_name":"Neeraj Goyal","publish_date":"2019-01-19T09:40:58","publication_year":"2019","word_count":858,"keywords":["Go","artificial intelligence","machine learning","AWS","AI","BFSI","sentiment analysis","ML","R","AI in finance","analytics","AI in banking","fraud detection"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","fraud detection","sentiment analysis","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-analytics-ai-ml-bfsi\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":36871,"title":"World’s Fastest Supercomputer Aurora To Be Built Over The Next 2 Years","content":"The TITAN supercomputer at Oak Ridge National Laboratory Ever since humans invented machines, the goal has been to make them as efficient and as powerful as possible. When we talk about machines, computers, in particular, we have come a long way in a relatively short period of time considering how powerful we made them since the invention of the very first one. Aurora and Exascale computing In the computing world, the performance of a Computer is measured in FLOPS or Floating Point Operations Per Second. An Exascale Computer is capable of handling at least one ExaFLOPS which is equivalent to a  quintillion (billion billion or 10^18) calculations per second. The US and China have been dominating in the race of building supercomputers. With  2 Supercomputers each on top four of the top 500 supercomputers of the world list, US and China have been competing to create computer systems that are capable of handling exaFLOPS. It was recently reported that the US Department of Energy, Intel and the supercomputer manufacturer Cray Inc will deliver the world’s first exascale computer to Argonne National Laboratory in Illinois. The supercomputer named Aurora will be worth $500 million and will be delivered and will start operating in 2021. With Aurora, US stands one step closer to achieving what is possibly the world’s first exascale Supercomputer. When it becomes operational in 2021, it will possibly be the world’s first exascale computer and also the fastest supercomputer in the world as well, 5 times faster than the IBM Summit supercomputer which is currently the fastest in the world. Whether or not Aurora will become the world’s fastest computer when it becomes operational will depend on other competing countries’ work, especially countries like China, Switzerland and Japan who are also among the top Supercomputer makers. China has already revealed its plans to build an exascale supercomputer of its own by 2020 and Japan, too, aims to have one up and running in 2021. The European Union has also committed 1 billion Euros for building an exascale machine by 2023. Exascale computing will transform the use and application of supercomputers to the next level. Aurora’s exascale speed will make it capable of processing an enormous amount of data. With heavy data processing capabilities Aurora will be able to handle data from advanced telescopes. Also, exascale computing will make the computations extremely precise. This will enable a more precise weather forecasting within ranges of neighborhoods rather than cities. Aurora will also find its applications in complex cosmological simulations, precision medicine for treatment of Cancer and mapping of the human brain down to the neural level. US Secretary of Energy Rick Perry said in the press release by Argonne National Laboratory, “Aurora and the next generation of exascale supercomputers will apply HPC and AI technologies to areas such as cancer research, climate modelling and veterans’ health treatments. The innovative advancements that will be made with exascale will have an incredibly significant impact on our society.” Raj Hazra, corporate vice president at Intel, said that Aurora’s real edge will come from its ability to use artificial intelligence to guide its models and simulations. He told the media that leading in computing power isn’t nearly as important as what that nation does with its capabilities. Conclusion With the world on a race to build the fastest computer, it is only inevitable that the computational power will keep increasing. Whoever wins the race to build the first exascale supercomputer, the world will have yet another most powerful computer that is capable of calculations to the extreme. As Hazra said it doesn’t matter who has the fastest computer, it is what you do with it that matters the most.","excerpt":"Ever since humans invented machines, the goal has been to make them as efficient and as powerful as possible. When we talk about machines, computers, in particular, we have come a long way in a relatively short period of time considering how powerful we made them since the invention of the very first one. Aurora […]","categories":["AI News"],"tags":[],"author_name":"Amal Nair","publish_date":"2019-03-26T07:46:55","publication_year":"2019","word_count":614,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Ray","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","Ray","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/worlds-fastest-supercomputer-aurora-to-be-built-over-the-next-2-years\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10119167,"title":"Yotta Partners with BLC to Build Nepal&#8217;s First Supercloud Data Centre","content":"India’s Yotta Data Services and Nepal’s BLC Holding have partnered to build Nepal’s first supercloud data centre facility called “K1” in Ramkot near the capital Kathmandu. The multi-million K1 facility will offer up to 4MW critical IT load capacity, spread across 3 acres and 60,000 sq ft area. Located within 20 km of Kathmandu airport, it will provide cloud, managed IT and cybersecurity services to store and process data, AI models and enterprise applications. The joint venture leverages BLC’s understanding of the local landscape, regulations and customer relationships, combined with Yotta’s global expertise, technology standards and access to wider markets, including hyperscalers. “Yotta’s expertise in managing data centres at the utmost standards and its dedication to constructing a hyperscale platform grounded in accountable, reliable, and reproducible methodologies resonates harmoniously with BLC’s proven track record in local operations,” said Sunil Gupta, Co-Founder & CEO, Yotta Data Services. In addition to the core data centre services, K1 will offer Yotta’s cloud platforms, such as Shakti Cloud for AI\/HPC and Yntraa hyperscale cloud, and managed IT and cybersecurity solutions. Built to Tier III standards, K1 will have a modular design for high uptime, reliability, scalability and flexibility. It will meet ISO 14000, 50001 for energy efficiency, ISO 27000, and PCI-DSS for security, with dual high-voltage power substations and carrier-neutral dense connectivity. The facility aims to enable AI\/ML development, generate jobs from construction to high-tech operations, and address data sovereignty concerns by enabling local control and participation. “The proposed K1 facility will not just be a data centre; we’re creating an ICT ecosystem that drives local and global growth,” said Megha Chaudhary, Managing Director, BLC. “By partnering with Yotta, we are strategically positioned to meet the volume and scale requirements while simultaneously delivering the premium, super-high availability needs of hyperscalers, enterprises, and the government.” This marks Yotta’s expansion into Nepal after its hyperscale campuses in Navi Mumbai and Greater Noida, India, as well as an upcoming project in Bangladesh’s Dhaka Hyperscale Data Center Park. The K1 data centre is expected to be completed within the next 24 months.","excerpt":"The multi-million K1 facility will offer up to 4MW IT load capacity to store and process data, AI models and enterprise applications.","categories":["AI News"],"tags":["Data Center","Yotta"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-04-29T12:13:32","publication_year":"2024","word_count":344,"keywords":["Go","API","Data Center","programming_languages:R","AI","Yotta","ML","Scala","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Scala","API","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/yotta-partners-with-blc-to-build-nepals-first-supercloud-data-centre\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":48152,"title":"After AI Minister, UAE Announces World&#8217;s First AI University","content":"In a turn of events that could very well be the beginning of something new in the education sector, Abu Dhabi this week announced the establishment of the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI). This university is reportedly the first graduate-level, research-based AI university in the world. Dr Sultan Ahmed Al Jaber, Minister of State, who has been appointed Chair of the MBZUAI Board of Trustees and is spearheading the establishment of the University, said, “MBZUAI aligns with the vision of the UAE leadership that is based on sustainable development, progress and the overall well-being of humanity and underpinned by capacity-building and active participation in finding practical solutions based on innovation and state-of-the-art technology. As such, the Mohamed bin Zayed University of Artificial Intelligence is an open invitation from Abu Dhabi to the world to unleash AI’s full potential.” Over the next decade, AI is set to have a transformational impact on the global economy, with experts estimating that, by 2030, AI could contribute nearly $16 trillion to the global economy. Given this and the UAE’s focus on a knowledge-based economy, the country made AI a strategic national priority in 2017, announcing a clear roadmap for an AI-driven future through the UAE’s Strategy for Artificial Intelligence 2031, and the appointment of the world’s first Minister of State for Artificial Intelligence. Experts now estimate that, by 2030, AI’s contribution to the UAE’s GDP will rise to nearly 14% – the largest GDP share in the Middle East. “AI is already changing the world, but we can achieve so much more if we allow the limitless imagination of the human mind to fully explore it,” he added. “The University will bring the discipline of AI into the forefront, moulding and empowering creative pioneers who can lead us to a new AI-empowered era.” The University will offer Master of Science, MSc, and PhD level programmes in key areas of AI – Machine Learning, Computer Vision, and Natural Language Processing – while also engaging policymakers and businesses around the world so that AI is harnessed responsibly as a force for positive transformation. The university will also provide all admitted students with a full scholarship, plus benefits such as a monthly allowance, health insurance, and accommodation.","excerpt":"In a turn of events that could very well be the beginning of something new in the education sector, Abu Dhabi this week announced the establishment of the Mohamed bin Zayed University of Artificial Intelligence (MBZUAI). This university is reportedly the first graduate-level, research-based AI university in the world. Dr Sultan Ahmed Al Jaber, Minister of […]","categories":["AI News"],"tags":["ML","uae"],"author_name":"Prajakta Hebbar","publish_date":"2019-10-17T15:12:26","publication_year":"2019","word_count":373,"keywords":["artificial intelligence","machine learning","uae","AI","programming_languages:R","innovation","ML","computer vision","ViT","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","R","Rust","ViT","innovation","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/after-ai-minister-uae-announces-worlds-first-ai-university\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25087,"title":"Study: Analytics And Data Science Jobs In India 2018 &#8211; By Edvancer &#038; AIM","content":"Analytics jobs scenario in India is constantly evolving. Companies are looking to hire professionals who are well-versed with new tech concepts such as analytics, big data, data science, artificial intelligence and machine learning, among others. While there may be a huge variation in the skills sets, experience and education requirement for these professionals across various cities and industries, there is a constant rise in the overall demand for analytics professionals with new jobs being posted each day. This year’s Analytics And Data Science Jobs Study is brought to you in association with Edvancer, one of India’s leading data science skills training institute which offers a wide range of data science training programs for individuals and corporates. The study takes into consideration various aspects of the job scenario for analytics professionals in India across industries such as retail, telecom and e-commerce, among others. The study also takes a look at various hiring trends over the years, jobs profiles, company types, experience required, educational background, etc. Read our last year’s study here. Top Trends In Analytics jobs The number of new analytics jobs advertised per month increased by almost 76% from April 2017 to April 2018. The number of new analytics jobs increased by 52% from April 2015 to April 2016, and by 40% from April 2014 to April 2015. It had almost doubled from April 2016 to April 2017. We saw a brief dip in the job requirements from August to September 2017. Post that, the number of analytics jobs advertised have been on a constant rise. While, it is difficult to ascertain the exact number of open analytics job openings; according to our estimates, close to 78,000 positions related to analytics are currently available to be filled in India. Open analytics jobs is a different metric than new jobs advertised per month. This is almost 57% jump in the open job requirements, compared to the same time a year back. Compared to worldwide estimates, India contributes 10% of open job openings currently. Growth in the number of data science jobs globally was much higher than India. Last year India contributed 12% of worldwide open job requirements which has decreased to 10% this year. 10 leading organisations with the most number of analytics openings this year are – JPMorgan, Accenture, Microsoft, Adobe, Flipkart, AIG, Ernst & Young, Wipro, Vodafone & Deloitte. Almost 98% of analytics jobs advertised in India are of full-time basis. Just 2% form the part-time, internship or contractual jobs. Top designations advertised are: Analytics Manager, Business Analyst, Research Analyst, Data Analyst, SAS Analyst, Analytics Consultant & Statistical Analyst, Data Scientist. Analytics Jobs By Cities In terms of cities, Bengaluru accounts for around 27% of analytics jobs in India. This is an increase from 25% a year ago. Delhi\/ NCR comes second contributing 21% analytics jobs in India, down slightly from 22% a year ago. Approximately 12% of analytics jobs are from Mumbai. This is significantly down from 17%, last year. The contribution of Tier-B cities in analytics jobs continues to increase this year, from 5% in 2016 to 7% in 2017 to 14% this year due to the increased number of start-ups operating in Tier 2 cities. Analytics Jobs By Industry Banking & financial sector continues to be the biggest influencer in analytics job market. 41% of all jobs posted for analytics were from the banking sector. This is a decrease from 46% a year ago. E-commerce continues to dip in terms of analytics jobs this year. Just 8% of analytics jobs were in ecommerce sector as opposed to 10% in 2017 and 14% in 2016. Energy & utilities sector seems to have the highest uptick in analytics jobs this year, contributing to 15% of all analytics jobs as opposed to 11% a year ago. The sector has been traditionally a late adopter of analytics. Education Requirement By Analytics Jobs Almost 48% of analytics job openings are looking for a B.E.\/ B.Tech graduate degree in the incumbent. 18% analytics job openings are looking for a postgraduate degree which are not MBA or M.Tech. This is a decrease from 26% a year ago. 8% analytics jobs specifically require an MBA\/PGDM degree. Just 13% recruiters are looking specifically for graduates with non-BE\/ non-B. Tech degrees, up from 10% last year. Experience Requirement By Analytics Jobs Around 62% of analytics requirements are looking for candidates with less than 5 years of experience. 17% analytics jobs are for freshers. 38% analytics job openings are for professionals with more than 5 years of job experience. Analytics Jobs Across Tools Recruiters are increasingly moving away from technology and tools-based recruiting to skills-based recruitment. More broadly, there has been significant decrease in the jobs advertised for a specific tool. Analytics recruiters are becoming aware that unlike IT, analytics requires a combination of skills, and tools are just one aspect of it. The demand for Python professionals is the highest among all analytics recruiters. Almost 39% of all advertised analytics jobs in India demand for Python as a core skill. Python also saw the biggest jump in analytics requirements this year, replacing R as the most in-demand analytics tool. R skills comes second, at 25% of all analytics jobs looking for R professionals. This is decrease from last year of 36%. Among visualisation tools, Tableau skills are most in-demand. Analytics Jobs By Salaries The median salaries being offered by advertised analytics jobs in India is INR 10.8 Lakh per annum. Advertised salaries tend to be lower than actual salaries. We have earlier reported the median salaries of analytics professionals in India to be INR 12.7 Lakh. 27% of all analytics jobs offer a salary range of 6-10 Lakh, followed by 23% for 3-6 Lakh. Almost 38% of all advertised analytics jobs in India are offering a salary of more than 10 Lakh. Analytics Jobs Across Company Type Almost 41% of all analytics demand is with Captive centres\/ GIC’s in India. These are organisations that mostly utilise analytics for internal consumption (for primarily their global businesses). This is down from 56% a year back. Service providers (both domestic & MNC’s) saw an uptick in the analytics jobs requirements this year. Almost 24% of all analytics jobs advertised are by MNC IT & KPO service providers, up from 18% a year ago. 21% of all analytics jobs advertised are by Domestic IT & KPO service providers, up from 15% a year ago. 9% of analytics jobs advertised are by Indian companies that require analytics for internal consumption. This is a very healthy number signifying growing adoption of analytics and data science with Indian organisations. Read the Entire Study: Download the complete Report [attachments include=”25114″]","excerpt":"Analytics jobs scenario in India is constantly evolving. Companies are looking to hire professionals who are well-versed with new tech concepts such as analytics, big data, data science, artificial intelligence and machine learning, among others. While there may be a huge variation in the skills sets, experience and education requirement for these professionals across various […]","categories":["AI Features"],"tags":["AI Jobs","Data Science Jobs","science jobs","study data science","Types of Databases"],"author_name":"Дарья","publish_date":"2018-06-04T08:59:45","publication_year":"2018","word_count":1106,"keywords":["big data","data science","Go","artificial intelligence","machine learning","science jobs","Types of Databases","AI","AI Jobs","Data Science Jobs","GAN","Python","analytics","study data science","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Python","R","Go","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/study-analytics-and-data-science-jobs-in-india-2018-by-edvancer-aim\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":57617,"title":"How India, US Can Collaborate on Emerging Technologies","content":"With India emerging as a key player in the new world order, it has been cinching strategic partnerships with countries around technologies like AI and ML. With the curtain coming down on US President Donald’s Trump’s maiden visit to India, a stronger commitment to strengthen technology ties between the two countries was found wanting. However, the technology sector had been a key focus area for both in the past. Buoyed by India’s investments in expanding its digital infrastructure, there had always been a large scope for the two countries to embrace frontier technologies and collaborate on enhancing capabilities for AI. Taken independently, the two countries have had very different approaches towards AI. The US’ budget for FY2020 allocated over $970 million to government agencies for non-defence AI research. In fact, this is the first time the country has acquiesced to agency-specific requests for expenditure on AI. This is in addition to routine measures taken to sharpen the country’s AI arsenal with clear, effective strategies. While this could help the US maintain its global leadership position in AI, India continues to lag behind. The country is trailing when it comes to policy vision, execution, and expenditure on AI. And other than a research paper anchored by Niti Aayog and a national-level AI workshop attended by relevant stakeholders, there is little else by way of progress. Even the long overdue Rs 7,000 crore financial boost for a national AI programme is a pittance given that it needs to be sustained till 2025, bringing down the annual spend. While efforts in India will be directed towards developing a proper AI roadmap, the US looks to better leverage opportunities, and ensure collaborative opportunities with industry, academia and other nations. And this includes AI partnerships with India. While a major share of this has been in defence, with India increasing its procurement of high-technology equipment, this alliance can go beyond military. With technologies like AI shaping new possibilities in areas like transportation, healthcare, and research, both countries have been presented with new opportunities for partnerships going ahead. India-US E-Mobility Collaboration Laying the groundwork for a first-of-its-kind partnership dialogue between Gujarat and Colorado, the US-India ‘State and Urban Initiative’ was launched in 2016 to facilitate knowledge sharing in support of an electric mobility transition between the two states. This collaboration involved a diverse set of stakeholders, including auto manufacturers, urban development agencies, research institutions, and policymakers, who went on to share best practices and identified opportunities for further partnership. While Colorado aims to increase the number of its electric vehicles to 940,000 by 2030, Gujarat is looking at 100,000 electric vehicles by 2023. Irrespective of whether they meet their targets or not, this kind of open communication between both states could set an example of how governments can work together to deal with the challenges and opportunities of the electric mobility revolution. Partnership To Counter Epidemics Like Covid-19 With all countries – including India and the US – restricting itself to quarantine procedures and timely detection of the virus to contain its spread, establishing joint agreements on ways to combat such events may be the need of the hour and AI can bolster such a partnership. World Health Organization (WHO) has developed a data repository carrying the latest scientific findings on Covid-19. Although it regularly updates this by scouring bibliographic databases, conducting this exercise manually can be exhaustive. AI tools can be enabled to handle this level of information in the most effective and accurate manner. These tools can be equipped to collect relevant information from global sources and analyse it to make AI-based recommendations. There is also an urgent need for a comprehensive policy ecosystem to handle such issues at a global level. A coordinated and proactive R&D capability is also critical. This is where a stronger partnership between India and the US will be crucial. India had pledged to abide by the obligations stated in the Biological and Toxin Weapons Convention (BTWC) back in 1974 and since then, has been seeking opportunities to improve its capabilities in the field of biotechnology. Building on this, it had signed a 10-year agreement with the US in 2015. Under this agreement, both countries will develop lightweight protective suits effective in chemically hazardous environments. While this is a good initiative, there is a scope for further collaboration in the expanding fields of AI and healthcare. AI Research & Development The potential gains from cooperation on the competitive talent available in both countries are many. Various global companies – including those from the US – have established research and product development centres in India. With the exponential growth at which AI is expanding, it will greatly influence the business landscape in the years to come, and both countries can drive this phenomenon with capacity building of enterprises. Such a scenario also throws open a vibrant community of startups and collaborative entrepreneurial opportunities on both sides. Outlook India has remained an important market for the US, especially in recent times. During Barack Obama’s first visit to India in 2010, the US lifted export controls on high-technology equipment exports. Almost a decade on, it had further eased export controls for high-tech product sales to India. This was achieved by designating India as a Strategic Trade Authorization-1 (STA-1) country – the only South Asian nation to be on this list. The US vision towards Asia has largely been centred around China, but this is slowly changing. India’s rise in Asia, coupled with fragility in US-China relations, may drive the US to rethink its Asia policy.","excerpt":"With India emerging as a key player in the new world order, it has been cinching strategic partnerships with countries around technologies like AI and ML. With the curtain coming down on US President Donald’s Trump’s maiden visit to India, a stronger commitment to strengthen technology ties between the two countries was found wanting.   However, […]","categories":["AI Features"],"tags":["AI India","US"],"author_name":"Anu Thomas","publish_date":"2020-02-27T13:00:00","publication_year":"2020","word_count":923,"keywords":["Go","API","startup","AI","ML","Git","RAG","GAN","Aim","AI India","R","US"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Git","API","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-india-us-can-collaborate-on-emerging-technologies\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":42030,"title":"Should Data Scientists Also Learn Social Sciences &#038; Humanities","content":"Combining data science with humanitarian sciences is not new. Many courses across the world are offering a blend of computing science and data science with humanities. As more institutes prefer to cross-train computing student with social sciences, it is quite intriguing why such a blend has become so popular. There are many experts who believe that there is a need for a new field called society and technology interaction that infuses courses and technicalities to inculcate social skills in those dealing with fields of computer sciences and data science. Why Is There A Rise In Popularity Of These Courses? When it comes to computing fields, experts believe that there is a need for training that equips candidates about social good and teaches them to value critical thinking. They are vouching for engineering programs that cultivate respect for other kinds of experiential expertise. The idea of these courses is to provide science and technology studies, communication, sociology, anthropology, political science, cognitive science, and digital humanities, to name a few. Another argument is that to be able to successfully implement technology, it requires socially-wise engineers. It is highly important to more socially responsible and inculcating ethical judgement to implement technologies in an ethical way. There have been many instances in the past where technologies such as artificial intelligence have been brought in use for unethical use and purposes. A course on these subjects could teach tech workers to have a better understanding of ethical vs unethical, thereby creating a more responsible tech team in the future. The idea behind these courses is to empower citizens, civic organizations, and advocacy groups to take collective action. Such courses are slated to show more progressive growth and deal with issues at hand with more inclusive solutions. It will equip them to overcome the challenges in a more effective way. It is the need of the hour to become an active technological citizen which does not essentially mean the ability to code. It also means that he\/she should be able to ask questions about how the code is going to affect work, life and community. It is about taking accountability of the work being done. Such courses will help engage in real problem-solving tactics while pushing the efficiency and practicality of the project to a whole new level. It is advisable for developers to work with experienced groups on “human-centric” design, which is much more than just the average Silicon Valley resident. Courses That Teach Technology + Social Sciences There are several courses now that make it possible for tech geeks to opt for academics in technology + social sciences. MSc in Social Data Sciences, a course offered by Oxford University is one of the most popular courses. It provides the social and technical expertise needed to analyse unstructured heterogeneous data about human behaviour, thereby informing our understanding of the human world. The idea behind the course is to help put better use of data which is generated digitally through various platforms. It is to ensure that one understands how data can be put to work and used in understanding crucial issues around social sciences, policy-making, economic behaviour, ethics, the social value of data and more. As the course description notes, it involves developing the science of these social data, creating viable datasets out of messy, real-world data and developing the tools and techniques to analyse them to tell us something about the world, through explanation, prediction and the testing of interventions. There is another course called MSc in Social Science of the Internet by the same university that provides students with an in-depth understanding of the social science concepts, theories and methods required to undertake rigorous empirical quantitative and qualitative research and policy analysis about the social implications of the Internet and technology. With this course, students can participate in innovative and pioneering research projects where computer science tools and methods are used to ask questions in the social sciences; or which use social science methods to understand computer science. In India, there is B.Tech in Computer Science and Master of Science in Computing and Human Sciences by Research (5 years Dual Degree) offered by the International Institute of Information Technology, Hyderabad. The idea of this course is to bridge the divide between technology and human sciences. It makes the student proficient in computer sciences with cross-disciplinary research capabilities which compare with the best in the world. It equips students in the rigorous methods and deep understanding of the human sciences, while at the same time equipping them with state-of-the-art computer science knowledge. Our Point Of View As technological advancement is reaching newer heights, the ways these technologies are being deployed can have various challenges. For instance, with an increase in the dark use of AI such as deepfakes and deepnudes, it has put the field at a huge risk. We recently also covered a story on how other technologies such as facial recognition and AI-based citizen scoring have potential risks on humanity. If courses such as the ones mentioned above are implemented at a larger scale and are made mandatory, most of the chaos around the ethical use of AI and related technologies may see a visible decline in the coming future.","excerpt":"Combining data science with humanitarian sciences is not new. Many courses across the world are offering a blend of computing science and data science with humanities. As more institutes prefer to cross-train computing student with social sciences, it is quite intriguing why such a blend has become so popular.  There are many experts who believe […]","categories":["AI Features"],"tags":["MSc data science","msc data science and analytics","MSc. in data science","purposes of a data team"],"author_name":"Srishti Deoras","publish_date":"2019-07-07T10:11:38","publication_year":"2019","word_count":868,"keywords":["MSc. in data science","data science","Go","artificial intelligence","programming_languages:R","AI","MSc data science","deepfakes","Git","RAG","msc data science and analytics","purposes of a data team","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","RAG","R","Go","Git","GAN","deepfakes","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/should-data-scientists-also-learn-social-sciences-humanities\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":50724,"title":"Solving Your First Ever Data Science Hackathon With MachineHack","content":"Data Science has opened up a myriad of opportunities in the past couple of years. It quickly topped the list of most wanted jobs and has witnessed the younger generation swarming for courses and jobs, However, unlike most domains, data science is one such field where an individual has to have a peculiar set of skills. From knowledge of linear algebra to storytelling, from programming to business case studies, the role of a data scientist varies between a statistician and an algorithm developer. No matter how much domain knowledge one has gained, it eventually comes down to the hours of practice put into mastering data handling. This is usually achieved through workshops and hackathons. Analytics India Magazine provides one such opportunity to beginners and to those looking towards a career transition, through its very relevant hackathons. Participants get to compete with high-level data science aspirants as well as practitioners. Not to forget the exciting prizes one gets to win! “I was more interested in applied after gaining the theoretical knowledge but the lectures limited themselves to theory,”– Abhishek Thakur, world’s first Kaggle triple grandmaster These hackathons cover a wide variety of domains from the classic Regression and Classification problems to Natural Language Processing and Image Classification and so on. Data Science principles are applied to Finance, Entertainment, Healthcare, Defence, Communications, E-commerce, Business and many more. Today, we are going to pick one from the very popular use case of food delivery. India has witnessed the rise of food delivery services in the form of Zomato and Swiggy. This multi-billion dollar industry keeps on benefiting with every optimisation. In our latest hackathon “Predicting Food Delivery Time – Hackathon by IMS Proschool”, we challenged the aspirants and developers to bring out their best algorithms from their armoury. In this article, we shall take the reader through the nuts and bolts of solving a data science problem in a stepwise fashion. So how can anyone who wishes to pursue the path of a Data Scientist start the journey? Practising is answer and Hackathons are the solutions. That is why we at MachineHack provide the young generation with an opportunity to apply everything that they have learned in all kinds of problems. We help thousands of students and expert Data Science practitioners by giving them an opportunity to sharpen their Data Science skills. Solving Your First Ever Data Science Hackathon When was the last time you ordered food online? And how long did it take to reach you? In this hackathon, we are provided with data from thousands of restaurants in India and the time they take to deliver food for online order. As data scientists, our goal is to predict the online order delivery time for the given test data based on the given factors. In this tutorial, we will try to crack MachineHack’s latest Hackathon called ‘Predicting Food Delivery Time – Hackathon by IMS Proschool’. The Hackathon is brought to you by IMS Proschool and MachineHack IMS, since 1977, has worked towards building a long term successful career for its students. It emerged as the fourth most trusted education brands in an AC Nielsen and Brand Equity Survey. IMS Proschool is the extension of the same mission. Proschool helps individuals realize their potential by mentoring and imparting skills. Problem Statement: Predict the delivery time for a restaurant based on the restaurant, it’s location, the cuisines etc. using the given data. Downloading The Data To download the data, head to www.MachineHack.com and sign up. Click on the Hackathons tab to go to hackathons page. Select the Predicting Food Delivery Time – Hackathon by IMS Proschool’ and start the course to download the datasets at the “Hackathon Dashboard”. See the below image. Let us have a look at the features: FEATURES: Restaurant: A unique ID that represents a restaurant.Location: The location of the restaurant.Cuisines: The cuisines offered by the restaurant.Average_Cost: The average cost for one person\/order.Minimum_Order: The minimum order amount.Rating: Customer rating for the restaurant.Votes: The total number of customer votes for the restaurant.Reviews: The number of customer reviews for the restaurant.Delivery_Time: The order delivery time of the restaurant. (Target Classes) Size of training set: 11,094 records Size of test set: 2,774 records Understanding The Problem If AI is the future then Data is the fuel of tomorrow. Without data, AI is only as good as a plain old ‘if-else’ statement. Within a large amount of data, lies hidden highly useful information that can determine an organization’s growth or warn us about an upcoming calamity. Companies like Uber eats, Swiggy, Zomato and many others in the food industry have been using data to optimize their decision making process such as which restaurant to recommend, what kinds of dishes to suggest and many such. In this hackathon, we need not optimize anything rather its very simple. Our objective is just to predict the delivery time for a restaurant based on the restaurant, it’s location, the cuisines it offers, the average cost of food, minimum order cost, the rating, the customer votes and reviews. To approach any problem we must have proper planning. We will create a pipeline and brake down the solution into four simple stages. Exploring the data and its featuresData CleaningData PreprocessingModeling and Predicting Let’s Code The following example was done on Google colab. Google colab is a very helpful tool by google that is built for Data Scientists. It is similar to a Jupyter Notebook but served and powered by google. It has its own kernel and all the codes are executed in Google’s own cloud and has support for GPUs. To mount your Google Drive to access files in the Drive from Colab, execute the following piece of code in Colab before you begin: from google.colab import drive drive.mount(\"\/GD\") Upload the participant’s data downloaded from MachineHack into your Google Drive directory. To load the dataset, use the mounted directory followed by the path to the files in your Google Drive. See the example below: train = pd.read_excel(\"\/GD\/My Drive\/Colab Notebooks\/Food_delivery_Time_Prediction\/DataSets\/Data_Train.xlsx\") Exploring The Data We already know and are clear with what the data is about, the features that come with it and also the objective which is to predict the delivery time. But should explore the data. This is a critical step in Data Science. Without exploring, Cleaning and preprocessing stages will become a chaos factory. Looking at the data: To explore the dataset thoroughly we will try to find answers to the following questions. What type of data does each column have?Does the table contain any missing or null values?Does any column contain multiple pieces of data that can be used to generate new features?Can new features be deduced from the existing columns?What are the categorical variables that need to be encoded?Does any column contain values that are irrelevant or have no significance to the context? Although the approach may vary, these questions form the baseline for solving any problem. Listed below are some very useful codes that can help us understand data: Key Observations : The Location and Cuisines column contains multiple values separated by commas. The Average_Cost and Minimum Order column consist of symbols and are strings.The Rating, Votes and Reviews column consists of invalid values such as ‘-’, “NEW’ etc.Restaurant, Location and Cuisines are categorical variables Data Cleaning Data Cleaning is a very important stage that can directly account for the efficiency of a machine learning model. Based on the insights gathered from the exploration stage, we will clean the data. This stage requires extensive coding. It is advisable to write functions that generalize well for similar kinds of data. Sometimes this is not possible and we will have to process and clean a column or feature separately. Given below are two functions to clean the ‘Locations’ and ‘Cuisines’ features of our dataset. We will split each cell of the Location and Cuisines column into the maximum number of features found in a single cell. Note: To find the maximum features in a single cell, we must consider both training and test data because the number of independent features must remain the same. The below two functions clean the ‘Ratings’, ‘Votes’ and ‘Reviews’ features of our dataset. Refer the Complete Code section below for detailed Data Cleaning. Let’s have a look at the cleaned data Once cleaned we can finally visualize the data. Let’s take a quick look at a pair-plot that shows the relationship between each of the numerical variables or columns in the training data. Pair-plot is one of the easiest ways to plot and identify the relationship between features in a dataset. #Relation between the numeric features in the dataset import seaborn seaborn.pairplot(train_sample) Data Preprocessing Data preprocessing, just like cleaning has an impact on a models performance. The processing stage consists mainly of the following processes Dealing with NULLS\/NaNs.Dealing with categorical values.Normalizing or scaling. Listed below are some helpful packages and methods that helps greatly with the preprocessing stage: Modeling and Predicting Finally, we are on to building a simple classifier that can predict and evaluate on our sample data. We will use a simple XGBoost classifier without any parameter tuning. This is a good starting point. Before we begin to create a model, make sure we have a small dataset to test our model performance. The best approach is to split the training set into a training set and a validation set. Also, it is important to separate out the independent and dependent variables from all the dataset samples. Note: To import the xgboost module, it must first be installed. Install it using !pip install xgboost What the following module does: Splits the training data into a training set and a validation setSeparates the dependent and independent features for the training set and validation setInitializes an XGBoost classifierTrains the classifier with the training dataEvaluates the score on a validation set Predicts the classes for the test set. Complete Code Execute the above code and upload your solution at MachineHack to see your score !! Good Luck!","excerpt":"Data Science has opened up a myriad of opportunities in the past couple of years. It quickly topped the list of most wanted jobs and has witnessed the younger generation swarming for courses and jobs, However, unlike most domains, data science is one such field where an individual has to have a peculiar set of […]","categories":["AI Features"],"tags":["Data Cleaning","Data Science","Hackathon","Machine Learning","stepwise regression machine learning"],"author_name":"Amal Nair","publish_date":"2019-11-26T18:23:53","publication_year":"2019","word_count":1662,"keywords":["data science","machine learning","AI","Machine Learning","RAG","Hackathon","Colab","Seaborn","XGBoost","stepwise regression machine learning","Data Cleaning","analytics","Data Science","Jupyter","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","XGBoost","Jupyter","Colab","Seaborn","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/predict-the-food-delivery-time-hackathon-solution\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10021479,"title":"Wipro To Acquire Tech Consultancy Firm Capco For $1.45 Billion","content":"In one of its largest acquisitions till date, IT Services giant Wipro has closed a deal to buy British tech consultancy firm Capco for $1.45 billion. The transaction is expected to be complete by the end of June, subject to customary closing conditions and regulatory approvals. Post the deal, Capco will operate as a separate unit, going by the name Capco-A Wipro Company. A company statement from Wipro said that combining its strategic planning, digital transformation, cloud, and IT services with Capco’s domain and consulting strength will help the former achieve its transformation objectives and further fuel company growth. “Together, we can deliver high-end consulting and technology transformations, and operations offerings to our clients. Wipro and Capco share complementary business models and core guiding values, and I am certain that our new Capco colleagues will be proud to call Wipro home,” said Thierry Delaporte, CEO and Managing Director of Wipro. Capco is a London based consultancy that provides digital, technology, and consultancy services to the banking and financial services industry. The 20-year-old comprises over 5000 business and technology consultants from over 30 locations spread across the USA, Europe, and Asia-Pacific region. In addition, Capco serves clients in the energy and commodities trading sector. Speaking of the acquisition, Capco CEO Lance Levy said the company looks forward to leverage complementary capabilities and cultures to drive change and offer exciting opportunities to their joint clientele.","excerpt":"In one of its largest acquisitions till date, IT Services giant Wipro has closed a deal to buy British tech consultancy firm Capco for $1.45 billion. The transaction is expected to be complete by the end of June, subject to customary closing conditions and regulatory approvals. Post the deal, Capco will operate as a separate […]","categories":["AI News"],"tags":["BFSI","Mergers and Acquisitions","Wipro"],"author_name":"Shraddha Goled","publish_date":"2021-03-04T19:16:48","publication_year":"2021","word_count":233,"keywords":["Wipro","Go","programming_languages:R","AI","BFSI","digital transformation","programming_languages:Go","Git","RAG","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","digital transformation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-to-acquire-tech-consultancy-firm-capco-for-1-45-billion\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10126477,"title":"IISc to Open Source Project Vaani Under Bhashini with 16,000 Hours of Speech Data","content":"The Indian Institute of Science (IISc) AI and Robotics Technology Park (ARTPARK) is set to open-source 16,000 hours of spontaneous speech data from 80 districts as part of Project Vaani, under the Ministry of Electronics and Information Technology’s flagship AI initiative, BHASHINI. This effort is in collaboration with the Google. The ambitious project aims to curate datasets of 150,000 hours of natural speech and text from approximately one million people across 773 districts in India. The first phase of the project, launched at the end of 2022, is nearing completion. In its second phase, Project Vaani will target 160 districts, collecting 200 hours of speech data from about 1,000 people per district. So far, voice data in 58 different language variants or dialects has been gathered from 80 districts and will soon be made publicly available. Prasanta Ghosh, Assistant Professor in the Department of Electrical Engineering at IISc, leading Project Vaani, mentioned, “Our journey in the Indic language data sets started in 2020 with RESPIN (recognising speech in Indian languages).” When speaking with AIM, Ghosh said, “We record people with impairments and are building technology that can understand them. Maybe humans can’t, but AI would be able to,” said Ghosh. He said that while collecting the data, he realised that there is no corpus even to build technologies for healthy people. “And then the first thing came to my mind was the idea to build a good foundational model for them.” Apart from healthcare, Ghosh is also focused on using AI for education in India on his project called English Gyani. Funded by the Ministry of Science and Technology, it is also called the IMPRINT project. “We’re building tools to help young jobseekers learn spoken English and comprehension skills right from their homes,” said Ghosh. “Bhashini, for the first time, helped create the digital data for low resource languages in its effort to build AI models for low resource languages,” said Amitabh Nag, Chief Executive of Bhashini. The primary goal is to use this dataset as training data for speech-to-text AI models, particularly benefiting conversational AI platforms and chatbots requiring diverse voice datasets. Organisations can leverage this data to develop speech models for various Indian languages and dialects, aligning with their business objectives. A Google spokesperson highlighted the project’s significance, stating, “Project Vaani was born out of the deep need to make high-quality, diverse & anonymised datasets available for people to build technology solutions that reflect the way local languages are spoken. We are grateful for the tireless work of our partners to realise this commitment, & are very pleased with the ongoing momentum. We look forward to sharing more updates soon.”","excerpt":"The ambitious project aims to curate datasets of 150,000 hours of natural speech and text from approximately one million people across 773 districts in India.","categories":["AI News"],"tags":["Bhashini","Open Source AI"],"author_name":"Mohit Pandey","publish_date":"2024-07-11T10:21:56","publication_year":"2024","word_count":442,"keywords":["Go","programming_languages:R","AI","chatbots","programming_languages:Go","Git","RAG","Open Source AI","Aim","Bhashini","GAN","R"],"extracted_tech_keywords":["AI","Aim","RAG","chatbots","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iisc-to-open-source-project-vaani-under-bhashini-with-16000-hours-of-speech-data\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10039347,"title":"How This IITian Built A No-Code Automated Video Communication Platform","content":"The video conferencing market is expected to reach a valuation of $50 billion by 2026 from $14 billion in 2019, a 250% growth. The Covid pandemic has boosted the video conferencing market with most companies shifting to remote work. In 2020, Gaurav Tripathi founded a no-code automated video communication platform, Superpro.ai, to provide video communication workflows. Gaurav started his entrepreneurial journey at the age of 19 when he was at IIT Bombay. Superpro.ai is a B2B SaaS platform with a ‘pay per use’ model. Analytics India Magazine got in touch with Gaurav Tripathi, the co-founder & CEO of Superpro.ai, to gain insights into the inner workings of the company. Genesis As soon as the lockdown started, video communication became the default workspace and delivery mechanism for a gamut of services such as education, health and fitness, consulting, and more. “We began by merely stitching the pieces together in one workflow and built a product consisting of scheduling, video conferencing, notifications, and payments to share with people. The idea was to simply combine the different tools. The next step was to build simpler pieces ourselves and package them as an offering,” said Gaurav. The startup has started assisting businesses in resolving real-world issues by using pre-built video communication workflows. “Our product can either be standalone or integrated within our clients’ websites and apps. It can be consumed and integrated as widgets, bots, APIs, and SDKs. All of these power conversations over video communication in some form or another are tied to a specific workflow,” said Tripathi. The company provides both pre-call and post-call solutions, from scheduling the call, to notifications, WhatsApp reminders, recordings, transitions, post-call feedback, and lastly, driving valuable insights from the data. “Most video platforms provide only one or two pieces of this puzzle, not all. We provide end-to-end insights and do it in an automated way, thereby saving hours of redundant manual work for our team members and customers while being cost-efficient,” said Tripathi. Tech stack The company gathers data both from its users and public sources and employs AI, ML techniques to generate valuable insights and provide recommendations for its customers. Take for instance, if we consider an ed-tech startup using the services of Superpro, the company would look for data from the workflow – right from the time slots and calendars to the teacher-student interaction, at what time each person joined the video call, and so on. Ultimately, the vast amounts of data will help the platform analyse and enhance its customer experience on a larger scale, he said. Adding to that, the company uses NLPs and video analytics to combine data from call recordings and transcriptions. Further, cloud-based services and various NLU techniques are used to discover keywords and phrases, as well as structuring the text extracted from the conversation. “The company’s core technology includes a full Javascript stack – react.js, node.js, etc. We also use Netlify along with a combination of GCP and AWS. For data processing, analytics and AI components are being written in Python,” said Gaurav Tripathi. Way forward Standalone video conferencing is a thing of the past. Now, it has already become an integral part of our lives, core of every business process and will stay here to help us adapt to the sudden, mass shifting to remote working. Recently, the company raised $300,000 seed funding from various investors including IvyCap Ventures Angel Fund, Pentathlon Ventures, ah! Ventures and SOSV. “I envision a future where I travel primarily for leisure and not for business since my business is run remotely. In five years, we will be seeing the video communication experience become equal to or better than in-person communication through technologies such as AI, AR, and VR. We will be able to get large amounts of work done using video communication and it will become the default for businesses,” said Gaurav.","excerpt":"The video conferencing market is expected to reach a valuation of $50 billion by 2026 from $14 billion in 2019, a 250% growth. The Covid pandemic has boosted the video conferencing market with most companies shifting to remote work. In 2020, Gaurav Tripathi founded a no-code automated video communication platform, Superpro.ai, to provide video communication […]","categories":["AI Features"],"tags":[],"author_name":"kumar Gandharv","publish_date":"2021-05-03T10:00:00","publication_year":"2021","word_count":640,"keywords":["Go","GCP","AWS","AI","ML","NLP","Python","analytics","JavaScript","R"],"extracted_tech_keywords":["AI","ML","NLP","analytics","AWS","GCP","Python","R","JavaScript","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-iitian-built-a-no-code-automated-video-communication-platform\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52551,"title":"Top Funding Announcements Made By Startups In 2019","content":"Riding the emerging tech wave and flooded with new funding, many Indian companies, startups and unicorns have seen great investments coming their way. The year 2019 saw many startups budding into thriving business houses, thanks to their investors. In this article, we list down the top funding announcements made by companies in India in 2019: January WayCool, a Chennai-based farm product delivery startup raised  ₹120 crore in both equity and debt from LGT, a Europe-based prominent angel investor and institutional lender. The objective of this funding round is to further expand their footprint and enter the Western market. Fractal Analytics raises funds and may be valued at $400 million. Apax Partners was in talks to pick up a majority stake making it one of the biggest shareholders in the analytics company. It is one of the most well-funded AI and analytics service providers in India. The investment from Apax is in the line of backing cutting-edge technology companies. Groww, a data science-based investment platform raised $6.2 million in Series A round of funding from Sequoia Capital. Y Combinator, Propel Venture Partners and Kauffman Fellows also participated in the round. The funds will be used to build its technology platform to scale up and introduce new investment options such as stocks as well to launch newer products. February Aureus Analytics raised $750,000 as closing investment from Connecticut Innovations, a US-based venture capital arm. The investment is part of a total fundraiser of $3.1 million and the company was selected as an early winner of VentureClash. It will be used by Aureus Analytics to drive growth, geographic expansion in the North American market and scale-up their technological and product development. NIRAMAI Health Analytix raised $6 mn in Series A funding led by Dream Incubator, a Japanese VC firm. The funding also saw participation from BEENEXT and other investors like Binny Bansal, Co-founder, Flipkart, pi Ventures, Axilor Ventures and Ankur Capital. It will be used for scaling its operations in India, hiring top talent and getting additional regulatory approvals for international expansion. Bangalore-based investment tech startup Smallcase raised $8 million in a Series A round of funding led by Sequoia India. Investors who participated in the funding round included Straddle Capital, Blume Ventures, Beenext Pte Ltd, WEH Ventures and DSP Adiko. They will use the funding amount to build a platform to enable retail brokerages to more products to their client base. Vidooly, e-sports analytics venture, raised ₹ 71 lakh from Times Internet and ₹14.3 crore from Alibaba. It was conceptualised and launched as an online video intelligence and analytics platform for content creators, brands, multi-channel networks, agencies and media companies. Vidooly entered the $100 billion online gaming and esports market with its new product eSports Analytics. Dunzo was funded by Cognizant’s co-founder and former CEO Lakshmi Narayan where it raised $3.1 million in total from their Series C funding. The Bengaluru-based app got big names such as Narayanan, Blume Ventures, Raintree Family Office (family office of the stationary giant Camlin), and Monika Garware Modi, vice chairperson and joint MD at polyester film manufacturing company Garware Polyester, as its backers. FourKites raised $50 million in Series A funding to expand their India team. It included Silicon Valley-based investors such as August Capital, Bain Capital Ventures, CEAS Investments and Hyde Park Angels. The company hopes to add another 200 employees in Chennai and another 150 in their US office. March TartanSense, a Bengaluru-based AgriTech and robotics startup raised $2 million in seed funding, led by Omnivore, Blume Ventures, and BEENEXT. With this TartanSense plans to scale their first product, BrijBot, a weed spraying robotic solution, for small cotton farmers. Singapore-based Tookitaki grabbed US$7.5m in Series A funding round. This funding round was led by Illuminate Financial, and existing investors who participated were Jungle Ventures, Enterprise Singapore, Supply Chain Angels, VWX Capital and senior banking executives. Frontdesk AI, a Palo Alto and Bengaluru based startup that develops AI assistants for small businesses, raised funding from Pi Ventures. The startup raised an additional $2 million in funding, bringing its total seed funding to $4.2 million. The funding will be used to accelerate product development, increase customer acquisition and investigate new vertical markets. April Neewee, a Bengaluru-based startup that works with data-centric predictive and IoT solutions, raised $4 million (₹27 crore) in their Series A funding. IIFL Asset Management Ltd purchased a minority stake in this round. Neewee’s flagship product Bodhee help manufacturing companies digitise their systems and processes in six weeks to personalise manufacturing. Analytics Software-as-a-Service company CleverTap raised $26 million in a funding round led by Sequoia Capital India, Tiger Global Management and Accel Partners. The company currently offers a mobile marketing platform to app developers, with marketing automation across a wide variety of interactive mediums. Bengaluru-based healthcare startup SigTuple raised $16 million in the Series C round of funding. It was led by Accel Partners, Chiratae Ventures, Pi Ventures and Flipkart co-founder Binny Bansal. Vue. ai, an Indian-US startup that develops artificial intelligence platforms to help online retailers work more efficiently and sell more, raised $17 million in a Series B funding. This round was led by Falcon Edge Capital, with participation from existing investors Sequoia and Global Brains (KDDI Japan). Bengaluru-based health-tech AI startup mfine announced that they had raised $17.2 million in Series B funding. The investment was led by Japan-based venture group SBI Investment, SBI Ven Capital, BEENEXT, Stellaris Venture Partners, and Prime Venture Partners. It plans to build its AI technology and expand the recently launched additional services such as medicines, preventive health screenings and diagnostic tests. Altizon Inc, a global industrial IIoT platform company announced series A+ funding round of $7 million. This round was led by TVS Motor Company (Singapore) Pvt. Limited, the Singapore-based subsidiary of TVS Motor Company along with The Hive, Wipro Ventures, and Lumis Partners, among others. It plans to use the funds to strengthen its international presence and continue its investments in IP development. May TheMathCompany announced that they had received funding from venture capitalist Arihant Patni in his personal capacity. With the new funding, TheMathCompany will build next-generation platforms and enhance the customer experience. Locus, a logistics optimisation startup secured $22 million in a funding round led by Falcon Edge Capital and its peer firm, Tiger Global along with its existing investors. The industry’s growth will be mainly fuelled by the strides in manufacturing retail, fast-moving consumer goods including e-commerce sectors. June Wysa, an AI conversational agent that help improve mental health, raised about $2m in a pre-Series A round of funding. This funding round was led by pi Ventures, with participation from Kae Capital and other investors. The founders aim to use this funding amount to further strengthen their technology and for expansion. Open, a startup focused on neo-banking, recently announced that they raised ₹210 crores ($30 million) in a second round of funding. The Series B round, led by Tiger Global Management, saw the funds being set aside for scaling, launching more products and value-added services. Singapore-based AI startup Active.ai raised $3 million from InnoCells. Its current investors include Kalaari Capital, Vertex Ventures, Dream Incubator, Chiratae Ventures, CreditEase Fintech Investment Fund, and Dream Incubator. Mahindra and Mahindra, one of the biggest conglomerates in India, recently purchased an 11.25% stake in a Switzerland-based agri-tech company, Gamaya for a total of ₹30 crore. Mahindra’s foray into AI-based agronomy solutions could begin an age of using ML for smarter and more sustainable farming practices. July Data democratisation startup Atlan secured a Series A funding for $2.5 million led by WaterBridge Ventures. Atlan has also been backed by Ratan Tata, Rajan Anandan, Manoj Menon and Hatcher, among others. Atlan intends to improve their product development and sign more customers. Aishwarya Rai along with her mother Vrinda KR made their debut as an angel investor and invested INR 50 Lakh each in Ambee, a Bengaluru-based environmental intelligence startup. It has also raised funding from Venture Catalysts, Amit Agarwal in December 2018. Singapore headquartered startup, Near, raised $100 million from a single backer, Greater Pacific Capital (GPC). With the funding amount, it aims at expanding and adding more people with the funding amount. August Resonance raised an undisclosed angel funding from IP Ventures. It intends to use the funds raised in the hiring of more engineers and the expansion of its sales footprint in international markets. H2O.ai announced that it has secured $72.5 million in a Series D round, bringing the total funding to $147 million. Since its founding in 2012, H2O.ai has been on a mission to democratize AI for everyone. With this new round of investment, the company plans to accelerate innovation and expand sales and marketing globally. vPhrase Analytics secured a Series A funding for $2 million from Bharat Innovation Fund and Falcon Edge Capital. It will use funding amount to launch their new product Explorazor and hire more experienced staff to guide them along. Uniphore, a global conversational AI technology company in the U.S., India and Singapore announced that it raised $51M in Series C funding led by March Capital Partners, with participation from Chiratae Ventures (formerly IDG Ventures), Sistema Asia, CXO Fund, ITP, Iron Pillar, Patni Family, along with other investors. Lendingkart Technologies Private Limited, raised a fresh equity round of INR ₹21 crore led by existing investors including Fullerton Financial Holdings Pte Ltd, Bertelsmann India Investments and India Quotient. The company will use it to reach small and underserved micro-enterprises and further strengthen its technological and analytics capabilities. DataRobot raised approximately $200M in a Series E funding round in the ballpark led by Sapphire Ventures. Company will use it to advance the development of its automated machine learning and AI software. Spyne, an AI-based SaaS startup that helps creative professionals such as photographers has raised seed funding led by Smile group and other angel investors. The company is already on a high growth path. It has signed up more than 700 photographers in the last 10 months. ThoughtSpot closed its $248 million in funding at a valuation of $1.95 billion in an oversubscribed round, bringing the company’s total funding to $554 million. It will continue to invest in its go-to-market teams in North America, EMEA, and APAC to meet growing demand, expand R&D efforts, including hiring for its engineering and product teams September Vianai reportedly raised $50 million as seed funding from undisclosed investors. Vishal Sikka demonstrated a new AI platform vision during his keynote address at Oracle Open World. Noted explainable AI engine startup Fiddler Labs announced that they have raised $10.2 million in Series A funding to accelerate their work in building breakthrough artificial intelligence engine with explainability at its heart. Lightspeed Venture Partners and Lux Capital led this round, with participation from Haystack Ventures and Bloomberg Beta. Darwinbox concluded a new round of funding, closing at a whopping $15 million. This round was led by Sequoia India and existing investors Lightspeed India Partners, Endiya Partners, and 3one4 Capital. It plans to expand across India and Southeast Asia and build out an ecosystem of integrated third-party solutions providers on its platform. Aquaconnect announced having raised a seed round of $1.1 million from Omnivore and HATCH. An AI-enabled platform, Aquaconnect integrates Asia’s leading network for aquaculture farmers with predictive SaaS tools for farm management and an omnichannel marketplace. InMobi raised $45 million for its new locked screen news app Glance, its first B2C product ever. The funding came from noted firm Mithril Capital. Glance is an artificial intelligence-driven, personalised content platform that works on the lock screens of smartphones. October Fireflies.ai announced having raised $5M seed round led by Canaan and other individual investors. the startup intends to use the funding amount to scale-up their engineering in Hyderabad and Bengaluru, bring in more machine learning solutions, add additional language support and expand its customer base in India Central Government is reportedly set to fund 100 startups to use its artificial intelligence-powered language platform. This platform, which will be used for translation services in numerous languages, will be opened for private use to allow both MNCs and startups alike for their language translation needs. Bangalore startup Vahan.ai raised an undisclosed amount of funding from Khosla Ventures, Founders Fund and Pioneer Fund. With Vahan.ai, job seekers can land a job within 24 hours and can help in effective onboarding of these employees in high volumes. November Automation Anywhere announced that it has received $290 million in Series B funding at a post-money valuation of $6.8 billion. The new capital will help Automation Anywhere accelerate its vision to empower customers to automate end-to-end business processes – bridging the gap between the front and back office with an artificial intelligence-powered intelligent automation platform. Paytm announced that it is planning to invest ₹500 crore in early startups who have capabilities to augment the digital ecosystem for the next wave of growth. The company will be focusing on artificial intelligence-based technology and big data solutions for new innovations which in turn will generate a workforce. December Accel India raised $550 million for its 6th fund to put resources into early-stage startup companies in India. Accel said it will concentrate on its core areas of interest: consumer tech, B2B (business-to-business) tech , programming and software, fintech and health tech. Observe.ai concluded its Series A round of funding, closing at a whopping $26 million. This round was led by Scale Venture Partners with participation from Steadview Capital, 01 Advisors, and their other existing investors. m.Paani, a Mumbai-based start-up secures $5.5 million in Series A funding, backed by previous Pre-Series A investors like Blume Ventures and formerly IDG Ventures, Chiratae Ventures. It endorses the ‘Kiranas’ and their intervention has benefited thousands of local retailers.","excerpt":"Riding the emerging tech wave and flooded with new funding, many Indian companies, startups and unicorns have seen great investments coming their way. The year 2019 saw many startups budding into thriving business houses, thanks to their investors. In this article, we list down the top funding announcements made by companies in India in 2019: […]","categories":["AI Trends"],"tags":["automated saas intelligence","Data Science","Funding","intelligent automation for retail","recent technological innovations","Startups"],"author_name":"Srishti Deoras","publish_date":"2019-12-25T13:00:00","publication_year":"2019","word_count":2275,"keywords":["data science","recent technological innovations","Funding","artificial intelligence","machine learning","AI assistants","AI","ML","automated saas intelligence","intelligent automation for retail","RAG","Ray","Aim","analytics","Startups","Data Science"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","Ray","RAG","AI assistants"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-funding-announcements-made-by-startups-in-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10070608,"title":"ZS India opens a new office in Bengaluru","content":"Management consulting and technology firm ZS has launched a new office in Bengaluru. The firm will soon open an additional office in Pune. The two offices are expected to create technology jobs over the next five years focused on artificial intelligence, data science, automation and machine learning, among other capabilities.  This expansion will strengthen its commitment to existing and new business functions in India. “We are at a critical juncture of our growth journey in India and this new office is part of our long-term business strategy. We will continue to make use of our global network advantages, coupled with our expertise to support the ZS vision. The firm is on a consistent growth trajectory. With these new offices, ZS will be able to hire more innovative people, connect with clients and expand its footprint across sectors,” said Mohit Sood, regional managing principal, India, ZS. ZS aims to bring together diverse, highly skilled talent to work on complex client engagements while leveraging the latest technologies to seamlessly collaborate with client teams across the globe. AI-and data-science-powered solutions are of critical importance to clients worldwide, and the firm is committed to investing in technology talent to help accelerate the pace of innovation.","excerpt":"ZS already has a strong presence in the life sciences and healthcare sectors.","categories":["AI News"],"tags":["data science professionals"],"author_name":"Kartik Wali","publish_date":"2022-07-07T14:47:12","publication_year":"2022","word_count":201,"keywords":["data science professionals","data science","machine learning","artificial intelligence","AI","innovation","ML","RAG","automation","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","Aim","RAG","R","automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/zs-india-opens-a-new-office-in-bengaluru\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101648,"title":"Infosys expands Google Cloud partnership to boost AI solutions","content":"Infosys announced today that it was expanding its alliance with Google Cloud to help enterprises build AI-powered experiences using Infosys Topaz offerings and Google Cloud’s generative AI solutions. Infosys will create the new global Generative AI Labs to develop industry specific AI solutions and platforms, which will help enterprises infuse generative AI into their business processes, the company said in a statement. The company will also train 20,000 practitioners on Google Cloud’s gen AI solutions, including Vertex AI and Duet AI in Google Workspace, to ensure organisations have the professional services expertise and resources to successfully develop, implement, and manage any type of generative AI project. This alliance between Infosys and Google Cloud builds on Infosys’ existing data, analytics and AI expertise on Google Cloud. Infosys is actively working with Google Cloud to develop a suite of transformative AI platforms and industry solutions for a range of business scenarios, including consumer AI, autonomous supply chain, autonomous marketing, anti-money laundering and customer services transformation. A wide range of its existing platforms and solutions are being enhanced with Infosys Topaz and Google Cloud generative AI capabilities. These include Infosys Live Enterprise Application Management Platform, Infosys Applied AI Platform, Infosys Customer Intelligence Platform, Infosys Data Streams, and Infosys Supply Chain AI Platform among others. The joint capabilities will help create a strong foundation for enterprises towards AI-enabled transformation. For example, Infosys Topaz and Google Cloud generative AI recently helped a leading consumer goods company in successfully launching an AI Twin to assist in real time planning of marketing spend, promotion, and product supply across markets. “Infosys has been making investments in the AI space for a long time. The combined strength of Google Cloud’s generative AI capabilities and Infosys will help enterprises transform and future-proof their business, built on strong digital, cloud, and next-generation AI capabilities.” said Salil Parekh, CEO, Infosys. Infosys AI-powered solutions use insights to improve customer experience, drive sales, and redefine client’s digital business strategy for long term success. The Indian IT giant is on an overdrive mode when it comes to AI. It has been partnering up with global firms like NVIDIA and Microsoft to expand it’s AI capabilities to boost it’s revenue and keep up with the AI-boom.","excerpt":"Infosys & Google Cloud join forces to advance AI-powered solutions","categories":["AI News"],"tags":["Infosys"],"author_name":"Pranav Kashyap","publish_date":"2023-10-18T17:39:45","publication_year":"2023","word_count":370,"keywords":["Go","Infosys","AI","programming_languages:R","programming_languages:Go","Git","generative AI","analytics","cloud_platforms:Google Cloud","GAN","R"],"extracted_tech_keywords":["AI","analytics","generative AI","R","Go","Git","GAN","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infosys-expands-google-cloud-partnership-to-boost-ai-solutions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10129807,"title":"AI Forum, AIM &amp; NVIDIA Present: Masterclass on Optimizing RAG Models for Enterprise-Grade Accuracy","content":"AI Forum, in partnership with AIM & NVIDIA, is set to host an insightful workshop on August 9, focusing on the optimisation of retrieval-augmented generation (RAG) models to achieve enterprise-grade accuracy. This virtual workshop, scheduled from 3:00 to 4:30 pm IST, aims to provide participants with advanced techniques and best practices for enhancing the performance of RAG models. Meet the Expert – Sagar, NVIDIA: The workshop, titled ‘Optimising RAG Models for Enterprise-Grade Accuracy: Advanced Techniques and Best Practices’, will be conducted by Sagar Desai, a senior solutions architect specialising in LLMs. Sagar also specialises in production deployment using NVIDIA’s stack. His expertise encompasses multimodal chatbots, RAG models, and LLM inferencing, ensuring scalable, reliable, and secure AI solutions. Sagar’s proficiency in model fine-tuning, including techniques like SFT, PEFT, and RLHF, along with his experience in designing scalable architectures using containerization, makes him a leading authority in the field. His work focuses on achieving state-of-the-art results in GenAI technologies for enterprise-level adoption. REGISTER NOW What You Will Learn? Everyone is talking about how to RAG in the era of generative AI and LLMs. Learning how to do it is definitely the need of the hour for every data and AI professional. Sagar will leverage his extensive expertise to guide attendees on how to unlock the full potential of RAG models through methods such as query writing, embedding fine-tuning, and reranking strategies. Participants can expect to gain valuable insights into scaling RAG models for high accuracy and reliability in real-world applications. The session will also cover best practices for deploying RAG models in production environments, providing a comprehensive understanding of the intricacies involved in optimising these models. Key Takeaways Optimisation of RAG models for enterprise-grade accuracy using advanced techniques. Scaling RAG models for high accuracy and reliability in real-world applications. Best practices for deploying RAG models in production environments. Why Attend? The workshop is designed for technical professionals with a background in natural language processing, machine learning, or AI. While a basic understanding of RAG models and LLMs is recommended, prior experience with query writing, embedding fine-tuning, and reranking strategies is not required. In addition, having an account on NVIDIA’s Build portal will be beneficial for API calls to models during the workshop. The NVIDIA Developer Program supports developers with essential resources to drive technological innovation. Advanced Tools & Technology: Access over 150 SDKs, including the CUDA Toolkit and NVIDIA NIM. Community Support: Peer and expert assistance for collaborative problem-solving. Hardware Grants: Available for qualified educators and researchers. Training Resources: Comprehensive materials to enhance skills. Free Software: GPU-optimised tools for AI, HPC, robotics, and more. Academic Support: Teaching Kits, Research Grants, and Fellowships. Startup Accelerator: NVIDIA Inception offers training, hardware discounts, and networking for AI and data science startups. Mandatory Pre-requisites for the Workshop • This session is designed for technical professionals with a background in natural language processing, machine learning, or artificial intelligence. Attendees should have a basic understanding of RAG models and Large Language Models (LLMs). Prior experience with query writing, embedding fine-tuning, and reranking strategies is not required.• Having an account on the – NVIDIA Developer Forum will help, it will be used for API call to models.• Python 3.10, Jupyter notebook setup You can join us for Live Q&A here. REGISTER NOW Secure your spot for the AIM Workshop on August 9, 3:00 to 4:30 pm. Don’t miss this opportunity to enhance your understanding of RAG models and their enterprise applications with insights from an industry expert. REGISTER NOW","excerpt":"Everyone is talking about how to RAG in the era of generative AI and LLMs.","categories":["AI Highlights"],"tags":["NVIDIA"],"author_name":"Mohit Pandey","publish_date":"2024-07-22T15:46:39","publication_year":"2024","word_count":580,"keywords":["data science","GenAI","machine learning","artificial intelligence","RLHF","AI","RAG","Aim","generative AI","NVIDIA","Jupyter"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","generative AI","GenAI","Aim","Jupyter","RAG","RLHF"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/ai-forum-aim-nvidia-present-masterclass-on-optimizing-rag-models-for-enterprise-grade-accuracy\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10039276,"title":"Data Scientist Creates Python Script To Track Available Slots For Covid Vaccinations","content":"Bhavesh Bhatt, Data Scientist from Fractal Analytics posted that he has created a Python script that checks the available slots for Covid-19 vaccination centres from CoWIN API in India. He has also shared the GitHub link to the script. The YouTube content creator posted, “Tracking available slots for Covid-19 Vaccination Centers in India on the CoWIN website can be a bit strenuous.” “I have created a Python script which checks the available slots for Covid-19 vaccination centres from CoWIN API in India. I also plan to add features in this script of booking a slot using the API directly,” he added. We asked Bhatt how did the idea come to fruition, he said, “Registration for Covid vaccines for those above 18 started on 28th of April. When I was going through the CoWIN website – https:\/\/www.cowin.gov.in\/home, I found it hard to navigate and find empty slots across different pin codes near my residence. On the site itself, I discovered public APIs shared by the government [https:\/\/apisetu.gov.in\/public\/marketplace\/api\/cowin] so I decided to play around with it and that’s how I came up with the script.” Talking about the Python script, Bhatt mentioned that he used just 2 simple python libraries to create the Python script, which is datetime and requests. The first part of the code helps the end-user to discover a unique district_id. “Once he has the district_id, he has to input the data range for which he wants to check availability which is where the 2nd part of the script comes in handy,” Bhatt added. While asking about the features that he is planning to add to the script, Bhatt said, “I’m still exploring the APIs that exist on the site. Based on my initial read, I think we can add features in the script which will help in booking a slot directly using the API. The other feature that can be added is to download the vaccine certificate directly from the API after the vaccination is done.” Check the link to the script here.","excerpt":"Bhavesh Bhatt, Data Scientist from Fractal Analytics posted that he has created a Python script that checks the available slots for Covid-19 vaccination centres from CoWIN API in India. He has also shared the GitHub link to the script.  The YouTube content creator posted, “Tracking available slots for Covid-19 Vaccination Centers in India on the […]","categories":["AI News"],"tags":["covid vaccine","Data Scientist","Python"],"author_name":"Ambika Choudhury","publish_date":"2021-04-30T17:14:15","publication_year":"2021","word_count":335,"keywords":["Go","API","programming_languages:R","AI","covid vaccine","Git","Python","analytics","GitHub","programming_languages:Python","Data Scientist","R"],"extracted_tech_keywords":["AI","analytics","Python","R","Go","Git","GitHub","API","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/data-scientist-creates-python-script-to-track-available-slots-for-covid-vaccinations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10063421,"title":"Google Cloud makes Suspend\/Resume generally available for cost optimisation","content":"Google has made suspend\/resume generally available. The features give you better control over Google Cloud resource consumption. Suspending a Google Compute Engine VM will save the state of your instance to disk allowing you to pick up where you left off when you resume it later. While your instance is in the SUSPENDED state, you no longer pay for cores or RAM, instead, you only pay for the storage costs of your instance memory. Other VM running costs such as OS licensing may also be reduced. How it works Suspending an instance sends an ACPI S3 signal to the instance’s operating system leading to two significant advantages compared to similar functionalities from other cloud providers. Broad compatibility with a wide selection of OS images without requiring you to use a cloud-specific OS image or install daemons. Undocumented and custom OS images that respond to the ACPI S3 signal may also work with Suspend. Storage is dynamically provisioned when Suspend is requested and is separate from the instance’s boot disk. This is in contrast to implementations in other clouds that require you to ensure that you have sufficient empty space in your boot disk to save the instance state which may increase the running costs of your VM. This also ensures that your suspended instance only consumes as much storage as it needs","excerpt":"Storage is dynamically provisioned when Suspend is requested and is separate from the instance’s boot disk.","categories":["AI News"],"tags":["Google Cloud"],"author_name":"Kartik Wali","publish_date":"2022-03-24T12:41:02","publication_year":"2022","word_count":222,"keywords":["Go","Google Cloud","programming_languages:R","AI","programming_languages:Go","RAG","cloud_platforms:Google Cloud","R"],"extracted_tech_keywords":["AI","RAG","R","Go","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-cloud-makes-suspend-resume-generally-available-for-cost-optimisation\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10115645,"title":"Meet Scott Wu the Creator of Devin","content":"Scott Wu Details NameScott WuAge26 Year (Born in 1997)OrganisationCognition LabsCurrent DesignationFounder and CEO of Devin AI, CognitionLinkedin Profilehttps:\/\/www.linkedin.com\/in\/scott-wu-8b94ab96\/Twitter Profilehttps:\/\/twitter.com\/ScottWu46 Before you have any questions, it’s worth noting that Scott Wu, the brilliance behind Devin – touted as the most capable autonomous coding agent – is a human. The demo video of Devin, released by Cognition Labs, introduces Wu as a ‘Human Software Engineer’, prompting viewers to ponder if he truly is human, especially after watching the childhood videos of him acing math quizzes. “I first learned to program when I was nine years old and fell in love with the ability to turn my ideas into reality. Teaching AI to code at Cognition Labs has been a dream come true,” said Wu on X, while announcing Devin. Source: X Wu, along with its two co-founders Steven Hao, the CTO, and Walden Yan, the chief product officer, founded Cognition in November 2023. The startup calls itself an ‘AI lab focused on reasoning’, and claims that ‘code is just the beginning’. Scott Wu a Child Prodigy Wu, 27, studied economics at Harvard University, after which he moved to San Francisco to start his own company. However, the story of his rare brilliance began when he was a little child. As a child, Wu participated in many math competitions, and aced them all with ease. Post the Devin announcement, various people shared their experience of having witnessed Wu as a child math whizz and programming genius. Source: X As a competitive programmer, Wu has been winning competitions on the programming platform Codeforces for years. As participants, one can compete in various coding contests, solving algorithmic problems and improving their programming skills on Codeforce. Wu has gained the title of a ‘legendary grandmaster’ on the platform. Source: Codeforces Prior to founding Cognition AI, Wu built another company called Lunchclub, which was backed by Lightspeed, Coatue and a16z. Wu served as its CTO and co-founder. Lunchclub is an AI superconnector that facilitates introductions for 1:1 video meetings. Building Team with Coding Champions While Wu’s genius sounds unparalleled, his brother is not far behind. Scott Wu’s older brother Neal Wu, who is also building Cognition Labs, has been participating with Scott in coding competitions since they were teenagers. The brothers are also math legends winning nationals several times during their school. A partner at Founders Fund, Delian, enthusiastically posted that he witnessed the brothers win nationals when he was in middle school, and said that he was excited about the model. Neal has worked in a number of big-tech companies including Facebook and Google Brain. He has also worked as a head teaching fellow at Harvard University. Cognition AI received $21 million from Peter Thiel’s VC firm, Founders Fund. The small founding team boasts 10 IOI gold medals, which is the ‘International Olympiad in Informatics’, a prestigious competitive programming competition for secondary schools students. The highly-talented team members have previously worked with Google DeepMind, Cursor, ScaleAI, and other tech companies. The Cognition Labs team, with Scott Wu (back row, 3rd from right) and brother Neal Wu (front row, 2nd from left). Source: Bloomberg Scott believes that the proficient background of his team gives the startup an edge over its competitors. “It’s almost like this game that we’ve all been playing in our minds for years, and now there’s this chance to code it into an AI system,” he said. View all stories Top 5 Devin AI Alternatives for Coders and Developers 10 Best AI Code Generator Tools to Use for Free in 2024 8 Things Developers Must Know About Devin Generative AI in Software 8 Must Have Skills to Become AI Engineer","excerpt":"“I first learned to program when I was nine years old and fell in love with the ability to turn my ideas into reality,” says the ‘human software engineer’ of Cognition Labs, Scott Wu.","categories":["AI Features"],"tags":["Devin","Interviews and Discussions","Scott Wu"],"author_name":"Siddharth Jindal","publish_date":"2024-03-14T12:40:19","publication_year":"2024","word_count":606,"keywords":["Go","startup","programming_languages:R","AI","Devin","programming_languages:Go","Aim","generative AI","Scott Wu","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","generative AI","Aim","R","Go","GAN","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-the-creator-of-devin-a-child-prodigy-who-is-making-coding-obsolete\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10004721,"title":"What Toshiba’s Investment In AI And Quantum Cryptography Means For The Company","content":"Having given up the nuclear project in the US, the Japanese multinational conglomerate reported an operating profit for fiscal 2019, which tripled in the year to 130.5 billion yen. At the same time, the group has dropped some core operations, including the memory chip business and emphasised back to its core strength of industrial applications using state-of-the-art technology research and development. AI and quantum cryptography seem to be the main highlight as it is aiming to spend about 34 billion yen ($321 million) on a new R&D unit in Japan. As part of this new building complex, a part of the present R&D centre in Kawasaki, near Tokyo will be renovated for the first time after, since its inception sixty years ago. The new unit aims to accommodate 3000 staff members who will work on various AI projects and quantum cryptography to find business applications. The construction will commence in January 2022 as per a Nikkei report and will be inaugurated in 2023. The new facility will consolidate existing projects and functions as a centre for innovation in AI. Quantum Cryptographic Communications Toshiba Corporation is the key member of a joint project of 12 Japanese organisations which aims to advance Japan’s quantum cryptography communication technology to the highest level globally. The project has a planned budget of 1.44 billion yen ($13 million), and Toshiba will be leading the research. The project, commissioned and supported by Japan’s Ministry of Internal Affairs and Communications (MIC), will soon begin research and development of a Global Quantum Cryptography Communications Network. The partner ecosystem is expected to develop the quantum cryptographic communication technology for practical use over a large scale network consisting of 100 quantum cryptographic devices by the year 2024. Toshiba has been working on cryptographic research for three decades, beginning in 1991 when the company built a research laboratory at Cambridge in the United Kingdom. The company has achieved feats such world’s first quantum cryptography communication at one-month-average key distribution speeds, exceeding 10 Mbps over installed optical fibre lines. According to Toshiba, the company will develop on its recent effort to improve technologies that will enhance quantum key distribution rates by around three times. Along with that, the company is working on completing storage system security technologies that assure safe and secure maintenance and management of cryptographic key data in a distributed communication environment. Quantum key distribution (QKD) is a cryptographic communications innovation that applies the principles of quantum mechanics to achieve completely secure communications that are almost impossible to intercept by any hackers. As a POC, the company became the first to digitally transmit complete Genome Sequence Data in a trial demonstration in Japan using quantum cryptography. Also, Toshiba is now executing field tests with partner companies around the globe toward achieving practical use cases for the technology. It is also working to advance the integration of 5G and networking technology with quantum cryptography. Toshiba’s AI Endeavours Toshiba has been carrying extensive research for various AI applications. It began with the creation of the first OCR (optical character recognition) postal code reading and sorting device which the company built in 1967. Since then, it has explored areas such as image recognition, voice recognition, translation, speech synthesis, chatbots, etc. Toshiba is currently ranked third around the world and first in Japan for a cumulative number of AI-based patent applications. Recently, Toshiba Digital Solutions Corporation launched its Distributed Co-Simulation Platform, which included model-based development for automakers and vehicle parts suppliers. The system, VenetDCP powers digital prototyping of automotive control systems and vehicular components using AI models in a simulated environment. “The focus today is moving in the direction of fusion between the cyber and the physical. To address increasingly complex issues, we need to shift towards digital transformation. So, we need to take real-world data, analyse it in cyberspace, and utilise the results to generate new value in the real world,” said Hironobu Nishikori, Executive Officer & Corporate Senior Vice President, Toshiba Corporation recently. Toshiba has begun an AI engineering training program in collaboration with the Graduate School of Information Science and Technology at the University of Tokyo. The program provides in-depth education on AI systems from classical machine learning to the state of the art advancements in deep learning. It also utilises Toshiba’s real-world big data, so that trainees acquire more practical skills in AI technologies. The company currently has over 700 AI engineers, but with the latest investment, the goal is to triple this number. Focus On Cyber-Physical Systems Toshiba aims to become a cyber-physical systems (CPS) company. One of Toshiba’s Cyber-Physical System (CPS) Technology projects is the Digital Twin-based Train Planning Project with UK train company Greater Anglia. Toshiba partnered to build a more effective train timetable and enhance customer convenience. It achieved this by developing the Digital Twin of the train and accurately reproducing a real-world train environment for the digital world by embedding it with an AI system. Then the company analysed and conducted simulations under various conditions to improve the operations. Toshiba recently built “Lag-aware Multivariate Time-series Segmentation” (LAMTSS), an AI that enhances the accuracy of anomaly prediction\/detection and motion analysis for infrastructure and manufacturing equipment. This technology automatically corrects time lag among multiple time-series datasets obtained from sensors installed in equipment or devices. In another instance, Toshiba created a PC-based, high-speed, high accuracy image analysis artificial intelligence system which monitors crowds by analysing the number of people in camera images by leveraging its proprietary deep learning algorithm. On the consumer side, Toshiba is researching voice and language applications for the Japanese market. The above developments suggest that Toshiba has set an ambitious plan for AI and cryptography projects, which are going to get a significant boost with the inflow of hundreds of millions of money into these sectors.","excerpt":"Having given up the nuclear project in the US, the Japanese multinational conglomerate reported an operating profit for fiscal 2019, which tripled in the year to 130.5 billion yen. At the same time, the group has dropped some core operations, including the memory chip business and emphasised back to its core strength of industrial applications […]","categories":["AI Features"],"tags":["cryptography","quantum application development system","quantum cryptography","the quantum companies"],"author_name":"Vishal Chawla","publish_date":"2020-08-12T16:00:23","publication_year":"2020","word_count":961,"keywords":["Go","artificial intelligence","machine learning","AI","chatbots","image recognition","quantum cryptography","cryptography","the quantum companies","RAG","Aim","deep learning","quantum application development system","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","Aim","RAG","chatbots","image recognition","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-does-toshibas-recent-investment-in-ai-and-cryptography-projects-in-japan-mean-for-the-company\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10165677,"title":"Anthropic’s Latest Goof-Up Broke Linux Systems","content":"In the tech world, it is all about experiments, mistakes, and building solutions to improve upon the existing system. Nothing is ever perfect, even after multiple iterations. That being said, when a notable company makes a mistake, it can leave a lasting impression. That is exactly what happened with Anthropic. Although it may have gone unnoticed until now, the situation is worth noting. Anthropic suggested a dangerous command to users, irrespective of their system configuration, in the process of helping users enable the auto-update feature for the command line interface (CLI). The issue has been fixed as of now. Bricking Sudo, Bricking Your System Sudo is a command in Linux that gives users elevated privileges to make system-level changes. It is advised against using the prefix sudo for any command, unless absolutely necessary. However, Anthropic recommended a dangerous command through its Claude Code CLI to help users automate updates with the same prefix as part of its configuration instructions, as noticed by Linux users. The spotty command ended up breaking many Linux systems, including Ubuntu and Arch installations. Breaking down the command, the \/usr directory is where all the executables and essential system files reside. The command here tries to change the permissions and ownership of the directory leading to system instability and potential security issues. Users mention that they had to go through several steps to recover their system state. The steps often included heading to the Grub bootloader and dropping down to the root shell to change permissions. Community to the Rescue Helping Anthropic Fix It When a software engineer, Sid Bidasaria, at Anthropic noticed it, a fix was implemented to redirect users to a documentation page for more information by removing the suggested command. However, it was not well-received by the community, prompting them to suggest more action items in a separate GitHub issue. The suggestion pointed out that the documentation page for Claude Code installation should reflect all the information suggested in the form of commits. The contributor suggested Anthropic to modify Claude Code CLI to detect when it is running as root, and warn people to reconsider. Moreover, Anthropic should email the preview release program participants to inform them. Addressing the core problem, Bidasaria explained, “When we saw that the installation location wasn’t user-writable, one of the two suggestions in the product was changing folder permissions with chmod\/chown commands. That was a mistake and it hit some users with the global npm prefix set to \/usr pretty hard. We’ve removed those problematic commands and now point users to better guidance.” Anthropic updated the documentation on March 6, adding a troubleshooting guide to the mix, and thanked the contributors for their inputs on addressing the issue. “Thanks for helping us improve Claude Code! Safety is super important to us at Anthropic, and we really appreciate you taking the time to help make things better,” Bidasaria added. He closed multiple issues pointing out the issue, and clarified, “Apologies for this. This was a bug where for some users whose npm global prefix was set to \/usr\/local, we displayed a bad command to run.” Do Not Execute a Command You Do Not Know About Netizens discussing the issue realised something important, i.e., not to type in anything they do not know about in the Linux terminal. As a joke, some speculated that the bad instructions were written by Claude itself, while some believed it was a simple human error. Meanwhile, referencing vibe coding, a user mentioned, “Vibe scripting gone wrong.” Not just for Anthropic users, a small incident like this should encourage developers and users to thoroughly understand what they are doing, even when following official installation instructions.","excerpt":"The AI company was quick to rectify the problem.","categories":["AI Features"],"tags":["Anthropic","Linux"],"author_name":"Ankush Das","publish_date":"2025-03-07T19:00:00","publication_year":"2025","word_count":610,"keywords":["Anthropic","Go","programming_languages:R","AI","Git","llm_models:Claude","RAG","GRU","Linux","GitHub","R"],"extracted_tech_keywords":["AI","Anthropic","RAG","R","Go","Git","GitHub","GRU","llm_models:Claude","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/anthropics-latest-goof-up-broke-linux-systems\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10103126,"title":"What Explains the AI Gap Between Indian Cos and Researchers?","content":"A staggering 75% of Indian companies express the belief that they have a mere year, at most, to have an AI strategy before their business faces repercussions. The urgency to deploy AI technologies has spiked in the past six months, with IT infrastructure and cybersecurity emerging as the top priority areas for AI deployments. A hint of anxiety is also visible in the startup ecosystem with companies like Zomato, OYO, ixigo, Freshworks and others having launched generative AI-enabled products and services in their respective domains. Even a company like Zerodha, which initially had a no-AI policy, has decided to explore the spectrum. These sentiments were reflected in the AI Readiness Index conducted by Cisco. In the survey, it was observed that only 26% of organisations in India are fully equipped to deploy and leverage AI-powered technologies. The index, a product of a comprehensive survey involving 8,161 business and IT leaders from the private sector across 30 markets, talks about the AI readiness among companies with 500 or more employees. On a positive note, Indian companies are actively taking strides to prepare for an AI-centric future. About 95% already have a sturdy AI strategy in place or are in the process of developing one. For instance, Project Indus by Tech Mahindra, which is an Indic-based foundational model, is expected to be launched in the next month or two. “As companies rush to deploy AI solutions, they must assess where investments are needed to ensure their infrastructure can best support the demands of AI workloads,” said Liz Centoni, executive vice president and general manager, applications, and chief strategy officer, Cisco. “Organisations also need to be able to observe with context how AI is being used to ensure ROI, security, and especially responsibility.” Globally, 95% of businesses acknowledge that AI will ramp up infrastructure workloads. However, in India, only 39% believe their infrastructure is highly scalable. This same group contends with limited or no scalability when confronting new AI challenges within their existing IT frameworks. To meet the power and computing demands of AI, over two-thirds (68%) of Indian companies anticipate the need for additional data centre graphics processing units (GPUs) to support current and future AI workloads. GitHub Report Tells the Tale GitHub’s State of the Octoverse 2023 report, which was released just a week ago, shows Indian developers’ trajectory in AI development. With a community of 13.2 million developers now active on GitHub within its borders, India has firmly established itself as the globe’s second-largest contributor to AI projects, after the United States. Notably, 3.5 million new developers joined GitHub’s ranks in 2023 alone. The significance of this surge is not lost on Sharryn Napier, VP of APAC at GitHub as she said, “Just imagine what India will be able to achieve if its 13.2M developers are empowered with AI. Not only will this transform enterprise innovation and productivity, but it will elevate developer happiness and make a substantial impact on India’s economy and society as a whole.” Examining the data shows a consistent year-on-year growth rate of 148% within the Indian developer community, as underscored by GitHub’s projections forecasting India’s overtaking of the United States in total developer population by the year 2027. This momentum signifies a seismic shift in the tech landscape due to AI, more specifically, generative AI. The stark difference between researchers and enterprise-level businesses in India highlights how tricky the AI scene in the country is. India has always adopted new technologies launched in the West, but now with generative AI, things are changing. According to GitHub, India is now one of the “top contributors” to the global AI open-source ecosystem. This shift is a big deal showing India’s involvement in the AI landscape.","excerpt":"In India, the urgency to deploy AI has spiked in the past 6 months, with infrastructure and cybersecurity becoming top priorities.","categories":["AI Features"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-11-16T15:15:30","publication_year":"2023","word_count":618,"keywords":["Go","API","AI","ML","Scala","Git","RAG","generative AI","GitHub","R"],"extracted_tech_keywords":["AI","ML","generative AI","RAG","R","Go","Scala","Git","GitHub","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/theres-an-ai-gap-between-companies-and-researchers-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10072676,"title":"The Digital Transformation Journey of Vedanta","content":"“In a world full of disruptions, the need for digital transformation is not a choice but a requirement that every organisation, big or small, has to undertake,” says Vineet Jaiswal, chief digital and technology officer at Vedanta Resources Limited. With the advancements in IT, it becomes mandatory for all businesses to deploy technologies like analytics, AI and utilise cloud-based systems to help them keep abreast with modern ways that help businesses. In an exclusive interaction with Analytics India Magazine, Vineet Jaiswal speaks about how Vedanta has used the latest technologies to transform its business. AIM: What was the need for a mining company and conglomerate like Vedanta to undergo digital transformation? Vineet Jaiswal: According to me, there are four requirements or needs for digital transformation. First, there are a lot of uncertainties because of changing requirements of both customers and suppliers. For example, sometimes the demand for a product goes up, but the raw material is not available in the market or the quality of the raw material is different from what is required. In some cases, customers may ask for product delivery within a shorter timeframe. So quicker response to emerging market conditions is one of the key reasons for digital transformation.Second is the changing mix of the future workforce, where we have employees with traditional thought processes who are detail-oriented vis-a-vis the new generation that wants everything quicker. So, we have to work out new ways of working.                The third reason is the evolving industry dynamics due to mergers and acquisitions. Faster re-strategising is the need of the hour for any company. Finally, there is a need to create a digital thread between IT & OT systems for gaining end-to-end visibility of the value chain. In the current digital ecosystem, the evolving technologies can be seen both as an opportunity to gain new insights as well as a disruption by others. AIM: What were the initial changes and challenges? And how soon did the workforce adapt to the transformation? Vineet Jaiswal: There are multiple changes like upskilling and cross-skilling of employees not only on the technology front but also in making them understand the potential and the need for digital transformation. In Vedanta, we initially saw that there was a lot of energy in the organisation and many things were happening on the ground where employees implemented digital solutions. But they didn’t have the end-to-end visibility of the entire value chain and were not aware of the best industry practices globally. To make them understand the big picture and holistic understanding of the digital ecosystem, we decided to launch Industry 4.0 framework for anyone to connect the business value chain with technology catalogue. Another challenge that we faced during transformation was adoption. When we rolled out something new, how employees adopted and sustained it is something that we had to carefully manage. We also realised that while we can upskill and cross-skill employees, sometimes we need special skills which cannot be generated in a short time to enable these transformations. External hiring to meet our requirements is a way to solve this. We also launched new types of capability building programs like training for data scientists, upskilling bootcamps, and Think Digital sessions for driving cultural change across the organisation. We did Management Act Up programs to pick people from businesses to be part of this transformation. Digital transformation is a journey and there are a few things that people have adopted and have started demonstrating value. There are other areas like data science, which is a continuous process. Here, we identified gaps in the understanding of digital technologies and therefore launched Vedanta’s Industry 4.0 framework for the entire organisation AIM: How would you describe the effects of digital transformation across Vedanta’s businesses? Vineet Jaiswal: In operations, digital transformation has led to improved productivity, cost optimisation, improved quality of finished goods, reduction in rejections due to defects etc. Asset availability also increased due to real-time monitoring of equipment and machinery. This has resulted in better planning and shutdown management. Digitisation also helped us improve safety across operations by reducing fatalities, incidents, implementation of group-wide safety management systems, geofencing of critical operating areas etc. There’s now a palpable change in mindsets across the organisation. People are expecting a lot from digitisation – touchless processes with minimal man-machine interactions, paperless offices for achieving zero paper usage, etc. This part of the journey is called digitisation, where we intend to convert data, documents, and processes from analog to digital. Next phase is digitalisation, where the aim is to increase the bottom line of the company by improvement of products, asset reliability, throughput, safety, automation of processes and simplification of communication etc. The last part is digital transformation, where focus is on improving overall revenue through better sales, customer experience, strategy, and solutions to various business challenges. AIM: Did the company implement transformation processes in-house or were external vendors called upon? Vineet Jaiswal: The transformation was done in collaboration with business consulting partners, start-ups, and IT technology partners. We have a program called Spark where we give opportunities to start-ups to showcase how they can impact Vedanta’s operations, safety and productivity. We evaluated over 1,300 start-ups and collaborated with some of them for developing solutions by giving them paid pilots across our business units. Currently, we are in Spark 2.0, where we plan to evaluate 1,500+ start-ups; we call this engagement Green Spark as we want to focus on ESG. AIM: Please share some successful use cases of subsidiaries that have already implemented the technologies for digital transformation? Vineet Jaiswal: Each and every business of Vedanta has already implemented multiple digital transformation initiatives. Throughput improvement using APC: For improving throughput, yield and optimising cost, we have deployed Advanced Process Control (APC) to transition our processes from manual control to automated adaptive control. This has resulted in unlocking of significant value and has helped us realize 1-5% gain in throughput, efficiency and cost reduction in zinc, oil & gas and aluminium processing units. Improving quality of input raw materials using analytics: To get the best cost coal blend with optimal quality parameters and cost, we used analytics for accurate prediction of coal blend and coke output properties without compromising on output quality. Ensuring safety through computer vision and AI: We deployed computer vision and AI-based safety monitoring to proactively track and report safety violations and non-conformances within our plant premises. Reducing time to first oil discovery and managing oil output using cloud: Previously all our workstations in the oil and gas business were on the premises, which needed constant upgradation coupled with long procurement cycles, limited computational capabilities, and higher maintenance cost. We decided to move 180 workstations to the cloud to unlock value from data stored in siloed systems resulting in scalability and elasticity, seamless access, cost efficiency, and faster decision-making with reduction in time to first oil discovery. Asset reliability improvement using predictive maintenance: For improving asset availability and reducing unscheduled downtime, a machine learning based predictive analytics solution was used for preventive maintenance and planning of shutdowns. Supervised and unsupervised machine learning techniques along with advanced pattern recognition technologies were used to recognise patterns from historical data for predicting potential failures.Safety training through extended reality: To ensure quality training of safety protocols, we have launched an interactive learning zone for safety training purposes. By leveraging augmented and virtual reality technologies, we aim to transform information and knowledge into intuitive, engaging, and outcome-based learning solutions.","excerpt":"In the current digital ecosystem, the evolving technologies can be seen both as an opportunity to gain new insights as well as a disruption by others, says Vineet Jaiswal, chief digital and technology officer at Vedanta Resources Limited","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-08-12T12:00:00","publication_year":"2022","word_count":1243,"keywords":["data science","machine learning","TPU","AI","ML","computer vision","RAG","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","data science","analytics","Aim","RAG","predictive analytics","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-digital-transformation-journey-of-vedanta\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10095940,"title":"It’s a Wrap on the Women in Data Science  Conference at Intuit","content":"Intuit India recently concluded the Women in Data Science (WiDS) conference held at their Bangalore office on June 20. The event brought together over 80 data science professionals, who left feeling empowered and inspired by the insightful sessions that took place. WiDS featured a variety of speakers, including leading data scientists, entrepreneurs and educators. The year marked a significant milestone as WiDS introduced an engaging mentoring session with accomplished women leaders in the data science field, giving the participants an opportunity to gain insights and tips for excelling in the industry from the leaders themselves. The event consisted of a variety of informative tech talks, keynote sessions, fireside chats, networking sessions, and much more. Here is a quick synopsis of the sessions – Keynote – Data Efficient Matching The WiDS Ambassador Note was presented by Sravyasri Garapati from Intuit, followed by an enlightening session on ‘Data Efficient Matching’ by Soma Biswas, an Associate Professor at the Indian Institute of Science (IISc) and a senior member of IEEE. Data Efficient Matching involves the effective pairing of data across diverse sources or entities despite having limited data at hand. It tackles the difficulties posed by sparse or incomplete information by leveraging sophisticated methodologies like machine learning, probabilistic modelling, and data fusion. The objective is to streamline the matching procedure while minimising data requirements, enabling successful matching even when only a small amount of information is accessible. This approach finds applications in various domains, such as identity resolution, recommendation systems, and data integration, where efficient matching plays a vital role in ensuring precise and dependable outcomes. Structured Data to Analytical Text Generation Google researcher Preksha Nema delivered an insightful speech on ‘Structured Data to Analytical Text Generation’. Nema’s research primarily focuses on understanding and generating purposeful ads, particularly in multilingual and low-resource contexts. In 2017, she received the Google PhD India Fellowship. Structured data to analytical text generation refers to the conversion of structured data, such as organised databases or spreadsheets, into comprehensible analytical text through natural language processing (NLP). This process employs NLP methods to translate numerical information into meaningful narratives or summaries. By extracting significant findings and trends from the structured data, this technique allows for the creation of concise and coherent textual explanations, enhancing comprehension and interpretation of the original data. It proves to be an invaluable resource for data-informed decision-making, generating reports, and effectively communicating intricate information to diverse audiences. Role of Analytics in Data Governance Kalapriya Kannan, representing Hewlett Packard Enterprise, discussed the ‘Role of Analytics in Data Governance’. Her expertise lies in machine learning, natural language systems, and their applications. Analytics plays a pivotal role in the realm of data governance by facilitating efficient management and utilisation of data assets. It harnesses the power of insights derived from data to establish robust policies and procedures governing data quality, integrity, and usage patterns. Through the application of analytics, organisations can identify irregularities within data, track its origin and evolution, ensure adherence to regulatory requirements, and make well-informed decisions based on reliable and precise information. Analytics further aids in the continuous monitoring of data governance processes, measuring performance indicators, and driving ongoing enhancements in data management practices. Ultimately, analytics empowers organisations to optimise the value of their data while upholding its integrity and security. Fireside conversation – Gender Bias in AI – Whom Do We Blame: Data, Model, Applications or Society? An interesting fireside chat on ‘Gender Bias in AI – Whom Do We Blame: Data, Model, Applications or Society?’ featured Kalika Bali from Microsoft Research in conversation with Shreya Mukhopadhyay The topic delved into the attribution of responsibility regarding the presence of gender bias in artificial intelligence. It examined whether the culpability lay with the training data employed, the AI models utilised, the applications that utilise them, or the societal prejudices ingrained in the data. This discourse highlighted the intricate interaction among these elements and underscored the importance of shared accountability in acknowledging and rectifying gender bias within AI systems. “I had the privilege of participating in the fireside chat, unsure of how the audience would respond. To my delight, numerous individuals approached me, eager to contribute and learn more about making a social impact. Their genuine curiosity and willingness to take action left me truly thrilled,” said Bali. Kalika Bali from Microsoft Research in an engaging conversation with Shreya Mukhopadhyay Responsible AI in Gaming Rukma Talwadker from Games24x7 delivered the closing keynote on ‘Responsible AI in Gaming’. “I had an incredible opportunity to present my research on identifying excessive gameplay patterns during an amazing session. The engagement and appreciation from the audience were truly gratifying. Kudos to Intuit for organising a well-executed event with excellent session choices,” said Talwadker. Responsible AI in gaming involves ethically and thoughtfully incorporating AI technologies into the gaming sector. It encompasses principles like equitable and inclusive portrayal, transparent algorithms, and mitigating adverse effects. Game developers aim to craft AI-infused gaming experiences that celebrate diversity, prevent bias, safeguard user privacy, and prioritise player welfare. Responsible AI in gaming strikes a harmonious equilibrium between innovation and ethical standards, ensuring that AI enriches gameplay while upholding players’ rights and values. The conference concluded with a highly impactful mentoring session, leaving all participants enriched and empowered.  With engaging discussions and meaningful connections, WiDS served as a catalyst for personal and professional growth for participants, solidifying its place as a remarkable gathering in the field. Watch out for the next edition of the Women in Data Science Conference, at Intuit where industry pioneers share cutting-edge knowledge, empowering you to thrive in a rapidly evolving landscape.","excerpt":"The event brought together over 80 data science professionals, providing a valuable platform for knowledge sharing and career advancement","categories":["AI Features"],"tags":["Intuit","Women in Tech"],"author_name":"Shritama Saha","publish_date":"2023-06-29T14:00:00","publication_year":"2023","word_count":928,"keywords":["data science","artificial intelligence","machine learning","AI","Intuit","ML","RAG","NLP","Ray","Aim","Women in Tech","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","data science","analytics","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/its-a-wrap-on-the-women-in-data-science-conference-at-intuit\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043342,"title":"Behind Clara Parabricks: NVIDIA’s Framework For Genomic Research","content":"A wealth of biological data has fueled the ability of scientists and healthcare researchers to utilise powerful computational methods to uncover insights. High-performance computer systems and domain-specific software frameworks are balancing out this digital biology revolution. Two of the world’s most powerful supercomputers, based on the NVIDIA DGX SuperPOD standard architecture- NVIDIA’s Cambridge-1 and Recursion’s BioHive-1, have recently been named in the TOP500 ranking of most powerful systems. Last year, researchers at the Athinoula A Martinos Center for Biomedical Imaging used various NVIDIA AI systems, including NVIDIA DGX-1, to accelerate their research on calculating lung disease severity from X-ray images to predict outcomes in COVID patients. While these advancements are taking place, Next-Generation Sequencing (NGS) activities are now being powered by the NVIDIA Clara Parabricks suite of genomics libraries and reference applications to push the barriers of genetic research. What is Clara Parabricks? NVIDIA Clara Parabricks is a computing framework for genomics applications ranging from DNA to RNA. It builds GPU-accelerated libraries, pipelines, and reference application workflows for primary, secondary, and tertiary analysis using NVIDIA’s CUDA, AI, HPC, and data analytics stacks. It is a complete solution for genomic labs to facilitate new application development. (Source: NVIDIA) Based on the Broad Institute’s Genome Analysis Toolkit (GATK), Clara Parabricks Pipelines enable GPU-accelerated GATK and other third-party tools like Google’s DeepVariant caller. Clara Parabricks maps, aligns, filters and calls variants for either germline or somatic variant detection, starting with DNA sequencing readings. STAR and STAR-Fusion align sequencing reads for RNA-based projects, allowing the readings to be divided into exon or intron boundaries, followed by variant calling. Accelerating research with Clara Parabricks The Parabricks Pipelines are a robust set of genetic tools that can be customised to match the demands of scientific research and laboratories. Researchers use NVIDIA GPU systems– ranging from desktop workstations to GPU-accelerated clouds and some of the world’s fastest supercomputers– to perform Parabricks Pipelines workloads. Shanghai-based Mingma Biotechnology became the first research facility in China to launch Clara Parabricks Pipelines this month to support its precision medicine work. It comes on the heels of large-scale genomics programmes launched earlier this year in Thailand and Japan. Even Houston-based Greffex is leveraging Parabricks Pipelines and NVIDIA Clara Discovery to enhance its efforts to build a universal flu vaccine just weeks after getting started with an NVIDIA RTX data science workstation. The startup combines genomic sequences, molecular dynamic techniques, and wet laboratory research to investigate how influenza strains develop through time and the impact of these changes on the vaccine’s effectiveness. To monitor flu changes, Greffex collects tens of thousands of flu genomes worldwide and performs massive lineups on NVIDIA RTX 8000 GPUs to recognise the changes to the virus’s genetic code. The company saves up to 13 hours per sample while operating genomic workloads on GPUs, enabling its team to fine-tune the alignment results. Genomic insights for population studies On NVIDIA GPUs, researchers using Parabricks Pipelines can accelerate DNA and RNA-based projects up to 50 times, allowing scientists to extract as much usable information as possible from the hundreds of gigabytes of instrument data generated every day. This acceleration is especially significant for public health institutions and research labs doing population studies involving tens of thousands of genomes to be processed. Mingma Biotechnology has implemented Parabricks Pipelines and NVIDIA T4 Tensor Core GPUs to speed up its sequencing and multi-omic analysis of data. The company helps medical institutions, pharmaceutical corporations, and researchers conduct medical research by providing genetic insights to identify and explore the root causes of diseases. Powering genetic analysis Apart from this, Genomics Thailand is being powered by an NVIDIA DGX A100 system at Thailand’s National Biobank to deliver genomic medicine as a standard healthcare service. The research organisation is analysing genetic variants using whole-genome sequencing data from 50,000 Thai participants utilising Parabricks Pipelines. NVIDIA DGX A100 is the world’s first five petaFLOPS AI system–providing unprecedented compute density, performance, and flexibility. The NVIDIA DGX A100 includes the world’s most advanced accelerator– the NVIDIA A100 Tensor Core GPU– allowing businesses to combine training, inference, and analytics into a single, easy-to-deploy AI infrastructure with direct access to NVIDIA AI experts. The combination of the DGX system with Parabricks Pipelines decreased the project’s entire genome data-processing time by four months. The findings of the study will help researchers better understand genetic variance in the Thai population. In Japan, the Human Genome Center at the University of Tokyo recently introduced SHIROKANE, the country’s fastest supercomputer for life sciences. The DGX A100-powered system uses Parabricks Pipelines to sequence the entire genomes of 92,000 patients, resulting in a database that will serve as the cornerstone for precision medicine efforts in cancer and other complicated diseases. While NVIDIA’s Parabricks pipeline acts as a powerful tool in the hands of genetic researchers, it will be interesting to see how it paves the way in genetics and genomics with future developments and updates.","excerpt":"A wealth of biological data has fueled the ability of scientists and healthcare researchers to utilise powerful computational methods to uncover insights. High-performance computer systems and domain-specific software frameworks are balancing out this digital biology revolution. Two of the world’s most powerful supercomputers, based on the NVIDIA DGX SuperPOD standard architecture- NVIDIA’s Cambridge-1 and Recursion’s […]","categories":["Global Tech"],"tags":[],"author_name":"Ritika Sagar","publish_date":"2021-07-10T10:00:00","publication_year":"2021","word_count":815,"keywords":["CUDA","data science","Go","AI","ML","RAG","Ray","analytics","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Ray","RAG","CUDA","R","Go","CUDA"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/behind-clara-parabricks-nvidias-framework-for-genomic-research\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10010999,"title":"Complete Guide To Exploding Gradient Problem","content":"Neural Networks have surely saved us many at times, the way we have used them for different use cases if simply phenomenal. This concept of deep learning was in talks for decades but because of computational issues, it was side talked for a few years. Deep Learning has got its hype again, many think that it has come a few years ago but that ain’t true. With computational issues, there were many other issues with a neural network too. One such is Exploding gradient. In this article you will learn about: What is exploding gradient and what issues did it cause? How to identify it? How to rectify it? What is exploding gradient and how does it hamper us? It can be understood as a recurrent neural network. For those who don’t understand what a recurrent neural network is, can be intuited as a Neural network who gives feedback to its own self after every iteration of the self. Here feedback means the changing of the weight. Source: Research gate. In gradient descent we try to find the global minimum of the cost function which will be the optimal solution for us. In this, the flow of information is from x1 to y3, in between we see h0,h1 etc which are the hidden layers. These hidden layers add biases and weight which is referred to as w. While propagating the information from y3 to x1, it will have to go through the hidden layers. With every iteration, the weights are set again. In RNN’s the weights are set to itself for a hidden layer itself too. That term is called Wrec which stands for Weight recurring. The output value at y3 is multiplied by the weights of h2 which is then given to h1 whose weights are multiplied by that of h1 and thus this goes on. Here the thing that we have to understand is that if the weights that are multiplied by the output of y3 are less than 1 then with time the actual value will diminish. Similarly, if the weights that are multiplied are more than one so eventually the value will become exponentially larger than the usual one. So for it to not change the value of the weights has to be equal to one. So here, in the situation where the value of the weights is larger than 1, that problem is called exploding gradient because it hampers the gradient descent algorithm. When the weights are less than 1 then it is called vanishing gradient because the value of the gradient becomes considerably small with time. The actual weights are greater than one and thus the output becomes exponentially larger at the end which hinders the accuracy and thus model training. A network with the problem of exploding gradient won’t be able to learn from its training data. This is a serious problem. How to identify exploding gradients? There are a few ways by which you can get an idea of whether your model is suffering from exploding gradients or not. They are: If the model weights become unexpectedly large in the end. Your model has a poor loss Or the model displays NaN loss whilst training. The gradient value for error persists over 1.0 for every subsequent iteration during training. How to deal with an exploding gradient? Use LSTM’s (Long short term memory) LSTM’s store the information and then is tolled against the values of the previous iterations. Here what happens is the value of Wrec is equalled to 1 which later doesn’t really impact the gradient. The sign sigma is for sigmoid activation function, tanh is for the tangent hyperbolic activation function. The value x which is coming out from ht is the final output value. Xt is the value that is added to the system, more like an input vector. There’s a lot more to it which you can understand and read from here. Gradient Clipping In really simple terms, it can be understood as clipping the size of the gradient by limiting it to a certain range of acceptable values. This is a process that is done before the gradient descent step takes place. You can read more about gradient clipping from the research paper here. Weight Regularization In this what we do is penalise the network’s loss function by regularising the loss. We use L1 regularisation or L2 regularisation which adds the square of the value to it. These regularisations techniques: L1 and L2 can be used for controlling the exploding gradients. You can read more from the research paper from here. Conclusion This article is aimed to discuss the issues that we may have whilst training a neural network in the step of backpropagation. This issue is addressed by the name exploding gradient when the weight recurring is greater than 1 and vanishing gradient when weight recurring is less than 1. We had also discussed how to identify the problem of exploding gradients which is by identifying and observing the loss and the weights of the model. Later we halted the article with a few solutions to our problem. LSTM is one of the most prominently used solutions for the same and apart from that, we had discussed gradient clipping and regularization techniques. Hope you liked the article.","excerpt":"Neural Networks have surely saved us many at times, the way we have used them for different use cases if simply phenomenal. This concept of deep learning was in talks for decades but because of computational issues, it was side talked for a few years. Deep Learning has got its hype again, many think that […]","categories":["Deep Tech"],"tags":["Exploding Gradient Problem","gradient descent","Machine Learning"],"author_name":"Bhavishya Pandit","publish_date":"2020-11-01T17:00:00","publication_year":"2020","word_count":880,"keywords":["Go","TPU","AI","Exploding Gradient Problem","neural network","Machine Learning","Aim","deep learning","CLIP","RNN","LSTM","R","gradient descent"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","TPU","R","Go","CLIP","RNN","LSTM"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/complete-guide-to-exploding-gradient-problem\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10133949,"title":"Google Unveils GameNGen: Integrating Generative AI into Gaming","content":"On 27th August 2024, Google released a paper detailing the capabilities of their new model GameNGen. This research was authored by Dani Valevski (Researcher at Google Research), Yaniv Leviathan (Engineer at Google Research), Moab Arar (PhD candidate at Tel Aviv University), and Shlomi Fruchter (Engineer at Google DeepMind). GameNGen’s Capabilities Powered entirely by a neural model, GameNGen is one of the pioneer search engines that allows high-quality, real-time interactions with complex environments over long periods. “GameNGen can interactively simulate the classic game DOOM at over 20 frames per second on a single TPU. Next frame prediction achieves a PSNR of 29.4, comparable to lossy JPEG compression”, per Google. The training occurs primarily in two phases:  (1) an RL agent learns to play the game, with training sessions being recorded, and (2) a diffusion model is trained to predict the next frame using the sequence of previous frames and actions. This approach, while still in its nascent stage, is testament of a new paradigm where games are weights of a neural model, not lines of code, in an era where video games are programmed by humans. GameNGen reveals that a neural model can interactively run a complex game like DOOM on existing hardware. Architecture & How It Works In order to collect data efficiently, an RL agent plays the game, and its actions and observations are recorded as training data. For more enhancement and consistency, the Stable Diffusion model is trained to predict the next game frame using previous frames and actions. Lastly, the model’s decoder is fine-tuned to improve image quality and reduce visual artefacts. AI in Gaming Jack O Brien, ex Google, remarked on how this research is seminal and opens up new possibilities for the world of video generation and the use of AI.  Jensen Huang, Founder of NVIDIA, has been bullish on the inevitability of the world of gaming and the role played by AI in making it a reality. “We used AI to revolutionise computer graphics. Graphics enabled AI, and now AI is saving graphics”, Huang said. Last year, OpenAI’s acquisition of Global Illumination, a gaming and design startup and the creators of Biomes, highlights OpenAIs gaming focus for AI simulation and use. As AI research advances, companies will increasingly test their models in game worlds before real-world deployment.","excerpt":"New research from Google and Tel Aviv University allow real-time interaction with complex game environments","categories":["AI News"],"tags":["GenAI","Google"],"author_name":"Aditi Suresh","publish_date":"2024-08-28T18:04:59","publication_year":"2024","word_count":383,"keywords":["Go","GenAI","TPU","OpenAI","AI","programming_languages:R","stable diffusion","ViT","Google","AI research","R","startup"],"extracted_tech_keywords":["AI","OpenAI","TPU","R","Go","stable diffusion","ViT","startup","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-unveils-gamengen-integrating-generative-ai-in-gaming\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10001982,"title":"How Swiggy, Zomato And Others Are Riding The New Green Wave In India","content":"Swiggy, one of India’s biggest food-tech startups, recently announced that it will be initiating a pilot for using electric vehicles in 10 Indian cities. This is part of their move towards using green alternatives to battle internal combustion engines used heavily by delivery partners currently. The unicorn has also begun rolling out a green initiative where partners were assigned bicycles to carry out deliveries. This is but the latest step in the green wave that seems to be sweeping India. A Conducive Environment For A Green Revolution Focus on EVs has been increasing in the eyes of Indian regulators, who have begun to institute the necessary infrastructure. The government aims to push EVs as a green alternative by making it easier to do so, and have even issued a set of guidelines regarding the matter. In May last year, the government clarified the position of EVs among other vehicles on the road, and stated that they would be differentiated by a green license board. The green license board was to ensure the identification of EVs at a distance. EVs are said to be given preferential treatment in terms of parking, free entry in congested areas and concessional tolls. The move was made as a part of trying to increase the adoption of EVs among the population. More recently, the Government issued guidelines in the form of amendments to model building bye-laws and Urban Regional Development Plans Formulation and Implementation Guidelines in order to make room for the infrastructure required for EVs. The government specified under the new guidelines that there was a need for charging stations for EVs to be setup across the country, with the move looking to reduce the number of ICVs and increase EV usage to 25% of transportation. Under these amendments, public charging stations will be set up on at every 25 kms, with fast charging stations for buses and trucks at every 100 kms. The country also aims to achieve ‘full electric mobility’ by the year 2030 in order to curb the rise of fossil fuel-powered cars. Startups Ride The Wave The Green Wave began with startups looking to capitalise on what was looking to be a quickly emerging vertical. Mahindra was already an established player in the space owing to their production of the E2O electric car. This product carried on the philosophy of Reva, which was acquired by Mahindra in 2010. However, the market still has place for a lot of other players, such as Mobycy, an electric scooter and e-bike sharing service, e-Hiran, an electric cycle company, Yulu and PEDL, which both operate in similar markets of lending bicycles and electric scooters. These startups began popping up left right and centre, leading to wide adoption by individuals all around the country in the face of rising traffic and pollution levels. Other companies also began jumping on the bandwagon, with cab taxi aggregator Ola also detailing its plans with a dedicated business only for EVs. This division, dubbed Ola Electric Mobility, has already set up various infrastructure such as charging stations and OEMs for batteries and parts. Lastly, Swiggy’s main rival, Zomato has also entered the EV fray with goals to deploy electric bikes in 12 cities in conjunction with many EV startups. With over 5000 cyclists on their payroll, Zomato is already beginning to see the benefits. Swiggy’s Green Masterstroke Swiggy has been in the green zone ever since it introduced bicycles to its fleet over 2 years ago, it had yet to enter the EV space. Now, with this pilot, Swiggy aims to not only be more sustainable, but also provide a better bottomline for their delivery partners. With the move being carried out in 10 cities, Swiggy said that it was possible for delivery partners to reduce their running cost by up to 40%, leading to a better payouts for the partners. The cost of purchase will be borne by the partners, who will eventually see better earnings owing to their use of EVs. In Delhi and Lucknow, Swiggy has already deployed e-rickshaws to undertake deliveries. Reportedly, the company is fulfilling over 1.5 million orders every month on mechanical cycles alone, with the majority of them being in Bengaluru. The cyclist fleet is spread over 54 cities, and consists of 10,000 partners. The move also allowed partners to deliver meals faster than bikes or scooters in some cities. Srivats TS, the VP of Marketing for Swiggy, said in a statement, “With the future heading towards more eco-friendly modes of transport, cycle and EV logistics will be the next game changer in food-tech and Swiggy wants to be at the forefront of driving that change.”","excerpt":"Swiggy, one of India’s biggest food-tech startups, recently announced that it will be initiating a pilot for using electric vehicles in 10 Indian cities. This is part of their move towards using green alternatives to battle internal combustion engines used heavily by delivery partners currently. The unicorn has also begun rolling out a green initiative […]","categories":["AI Features"],"tags":["electric vehicles","EV","foodtech","Government","swiggy","Zomato"],"author_name":"Anirudh VK","publish_date":"2019-05-02T20:09:47","publication_year":"2019","word_count":774,"keywords":["Go","API","foodtech","AWS","AI","ML","Government","GAN","Ray","electric vehicles","Aim","Zomato","EV","swiggy","R","startup"],"extracted_tech_keywords":["AI","ML","Aim","Ray","AWS","R","Go","API","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-swiggy-zomato-and-others-are-riding-the-new-green-wave-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":37762,"title":"The Battle Between Enterprise AI And Hacker AI Continues","content":"Even though experts say that technology will make the world a  better place, many believe that one-day technology would reach such a level that it might turn the table. It’s been decades and the debate is still one— who is better, machines or humans? Today, with the advent of some of the most advanced techs — artificial intelligence (AI), IoT, ML etc., the world is witnessing a tremendous wave of innovations. However, the wave is not only about the innovations for the good, but there are also things that have emerged as threats for people as well as for companies all across the world. How AI & ML Has Turned The Table Over the years, AI has evolved significantly, and at present, it is used in several verticals. And with each innovation in technology comes the reality that AI and ML gradually have found their way out and now they’re increasingly being weaponized. For instance, hackers can make a phone number look like it’s coming from your home area code and trick your firewall like a machine learning Trojan horse. The level of sophisticated hacking is taking a whole new turn. Technology is completely unbiased and even though the latest techs such as AI and ML are considered a force for good, in the hands of wrongdoers, these technologies can create some serious damage. So, should we get worried about this? Or just sit and watch these techs heading toward a future where they will battle it out with each other? The Advent Of Adversarial Use Of AI and ML In terms of Open source platforms, the tables are turning here as well, and one of the best examples is Kali Linux, one of the most popular open-source testing OS. Over the years, it has become a go-to platform for all the pen testing enthusiasts. And being an open-source platform, it’s not only the organisations that have the access, but the wrongdoers also have the access. Hackers are using these kinds of platforms and tools to gather data from their target and later use that data to hack. In terms of ML, a significant number of ML models in the cybersecurity space is black box. And that is what hackers are making use of. These deep learning models can be compromised, and the results can be altered. And that is the reason why explainable AI is becoming go-to for a number of companies across the globe — it not only delivers an outcome but it justifies it. What’s Next? The pace at which technology is evolving, there is no doubt that even adversaries will increase their use of sought after tech such as machine learning to create attacks. And with that cyber attacks are prophesied to become more affordable as well as efficient at deploying new types of attacks. It is because AI and ML-based tools would let the hackers perform attacks and functions that would be virtually difficult for humans Many might think that it’s all just speculation, however, that is not true. Hackers are always adaptable and just like a technology enthusiast they also keep their knowledge and skills updated to harness the latest tech and create fresh new ways to penetrate the new defences of organisations across the world. Its high time for enterprises that they prepare themselves for the upcoming, sophisticated adversarial AI & ML-based cyber-attacks.","excerpt":"Even though experts say that technology will make the world a  better place, many believe that one-day technology would reach such a level that it might turn the table. It’s been decades and the debate is still one— who is better, machines or humans? Today, with the advent of some of the most advanced techs […]","categories":["AI Features"],"tags":["Adversarial AI","AI (Artificial Intelligence)","Cyber Attack","Kali Linux","Machine Learning"],"author_name":"Harshajit Sarmah","publish_date":"2019-04-15T11:41:24","publication_year":"2019","word_count":560,"keywords":["Kali Linux","Go","artificial intelligence","machine learning","AI","innovation","ML","Machine Learning","Adversarial AI","Cyber Attack","explainable AI","deep learning","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","R","Go","GAN","explainable AI","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-battle-between-enterprise-ai-and-hacker-ai-continues\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10048703,"title":"Amazon’s Use Of Algorithms To Increase Employees’ Productivity: Fair Or Not?","content":"Ever wonder how Jeff Bezos managed to build a $1.56 trillion worth (as of July 2021) empire — the world’s largest retailer, and become the richest man in the world? While the reasons may be many, automating its management system has to play an interesting role. Unlike traditional offices where managers were always barking behind the backs of inefficient employees, Amazon has developed a software that tracks every move of its workers in the fulfilment centre — writes Christopher Mims for The Wall Street Journal after interviewing Amazon workers. An era of ‘Bezosism’ This isn’t the first time that a multinational company, and a billionaire, harnessed technology to boost the productivity of its employees and build an empire of a business. Henry Ford — the first billionaire in the history of billionaires, introduced the system of using surveillance, measurement, targets, psychological tricks and incentives to speed up work and productivity of its employees. Mims has coined Bezos’ practice of supercharging the management system as ‘Bezosism.’ Amazon evaluates its workers’ performance using the metric ‘rate.’ This usage of technology to manage and increase the performance of its workforce, automate its systems and further innovate based on the availability of data, is what makes Amazon the world’s largest retailer. How the tech works At its fulfilment centres, Amazon measures its employees’ pick or stow rate at a robot-fed pick and stow station. The rate at which an employee or worker in its warehouse must complete a task (stacking items on shelves, taking items off shelves, putting items into boxes, etc.) is calculated based on every worker’s overall and aggregated performance in that particular warehouse facility. Employees get a 30-minute lunch break and another 30-minute break built into their normal schedule to get water, use the restroom and talk to managers in the 10-hour shift. According to the Wall Street Journal’s investigation, associates are not allowed to sit down on their jobs besides lunchtime and the extra 30-minutes break. This competitive measure of one’s performance can be both fruitful and dangerous. Every time a worker misses the target rate, the algorithm triggers a warning. If one gets too many warnings, their job is at stake. These warnings can have severe psychological consequences on the workers, pushing them to work harder, often exceeding their physical and mental limits. Fruitful or dangerous? Amazon argues that incorporating algorithms in measuring one’s productivity ensures that no employee is pushed beyond human limits to achieve unattainable targets. In April, in a letter to its shareholders, Bezos wrote, “ If you read some of the news reports, you might think we have no care for employees…our employees are sometimes accused of being desperate souls and treated as robots…We don’t set unreasonable performance goals.” Bezos, in the letter, mentions that Amazon terminates only 2.6 per cent of its employees due to their inability to meet targets. While this may seem low, the bigger picture tells a different story. Today, Amazon directly employs more than 1.3 million people across the globe. That makes more than 4,000 employees being fired every year. Earlier this month, The Verge reported that California had passed the AB 701 bill challenging performance goals laid down by companies of the likes of Amazon on its warehouse workers. The bill was reported keeping in line with the number of at-work injuries reported by Amazon warehouse workers due to the implementation of algorithms. Bezos himself suggests that 40 per cent of the injuries recorded by Amazon workers have to do with musculoskeletal disorders (injuries of the muscle, nerves and joints) usually caused by repetitive movements at work. A look at the broader picture, thus, reveals that Amazon’s use of data, surveillance and algorithms to accelerate its employees’ performance has more grave repercussions than good.","excerpt":"Amazon deploys data, surveillance and algorithms to accelerate the productivity of its warehouse workers. But how ethical is it?","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Amazon","jeff bezos","Machine Learning"],"author_name":"Debolina Biswas","publish_date":"2021-09-19T11:00:00","publication_year":"2021","word_count":625,"keywords":["Go","programming_languages:R","AI","Machine Learning","Amazon","programming_languages:Go","ViT","jeff bezos","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/amazons-use-of-algorithms-to-increase-employees-productivity-fair-or-not\/","complexity_score":4,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044589,"title":"SentinelOne’s Denise Schlesinger On How To Build A Great Data Team","content":"“Big data means big scale and this means big problems, which are usually fun and challenging to solve.”Denise Schlesinger For this week’s machine learning practitioner’s series, we got in touch with Denise Schlesinger, senior director of R&D at SentinelOne, The SentinelOne platform provides defenses needed to prevent, detect, and remediate cyberattacks. AIM: Tell us about your journey so far. Denise: I grew up in Argentina and came to Israel at the age of 18 to study Computer Science. I started working in Software companies at the age of 24 as a software engineer, mostly developing web applications. I was promoted to team leader and then R&D director by the age of 30 and was managing teams of software engineers. Over the years, I worked as an architect and a VP of R&D at several startups in different industries: Agrotech, Adtech and Cybersecurity. My roles involved supporting big infrastructure and re-architecting products to support large-scale building and scaling tech teams. I was part of teams where I oversaw designing the complete architecture, from the ground up, of many cloud-based SaaS products and defining technical strategy and roadmap for Distributed applications, ensuring high availability and scalability. To keep myself up to date, I read many blogs on subjects such as big data, high scale and productionising of  Machine Learning Models. For example: Uber, Netflix, Lyft and Wix (Israeli company) engineering blogs. AIM: What does your typical day look like at SentinelOne? Denise: Before SentinelOne, I was VP R&D at Novarize, where we developed AI based tools to provide insights for marketers. I joined SentinelOne remotely during the pandemic, which was certainly a big challenge. It was incredible to see how generous people are with their time and knowledge. Thanks to their support and understanding, my transition has been a fun and positive experience. Currently, I am a Senior Director of Engineering at SentinelOne. I lead AI and Big Data teams.  My group is in charge of the data pipelines, the services that do pre-processing, aggregation and detection for all the data collected. We ingest hundreds of millions of events per minute, we run on the cloud. Our production infrastructure is huge. On my day to day, I am involved in all aspects of architecture, software and product development, delivery schedules for high scale applications. I review my group’s development projects to ensure reliability, effectiveness and ROI. “Culture eats strategy for breakfast” AIM: Give us a glimpse into your toolkit. Denise: We run Presto, Spark, Kafka, ElasticSearch and all of  our services on top of Kubernetes. We leverage Databricks, AWS Sagemaker and Spark for machine learning. We use AI to solve the hardest problems that are part of leading with such huge amounts of data. I am hands-on and love trying new technologies and frameworks AIM: What does it take to build a great data team? Denise: I manage and mentor my teams and the managers I lead. I lead by example, I love understanding the small details that make the big picture. I like the challenge of simplifying complex systems. Enabling my teams to grow by granting autonomy, I create a safe environment with permission to fail. I truly care for them, I understand the strengths of each person and do my best to enable him\/her to thrive. Working closely with different business stakeholders in the organisation to create awesome products. Building relationships, motivating, coaching and enabling each team member to be at their top game. On top of the really interesting technical challenges that come with working with big data and AI, one great thing about working is the impact you can create.  Also big data means big scale and this means big problems, which are usually fun and challenging to solve. We invest a lot in building our Data Infrastructure to provide Scalability, Reliability, and Efficiency. I strongly believe in the saying: “culture eats strategy for breakfast”. This is highly important when creating a data-driven culture to breathe data and require for every decision to be data-driven. AIM: What do you look for in your team members while hiring? Denise: When hiring people I look for critical thinking, accountability, and innovation. I appreciate the ability to look at things from a bird’s eye view and at the same time dive into the details to get the whole picture. I value curiosity and found that great engineers want to work on difficult problems alongside peers. I hire good team players that believe in the mission and value a culture of collaboration and exploration. AIM: What are your views on the current AI and cyber security landscape? Denise: Nowadays, hackers launch hundreds of millions of attacks worldwide. Unknown threats can cause massive damage affecting a company’s business if they go undetected. Human beings cannot possibly identify all the threats. Organisations face the challenge of analysing and tracking network and workstation activities, this is a lot of data that has to be scanned to allow protection from malicious people and software. AI is able to analyse billions of events and identify different types of threats: from malware exploiting zero-day vulnerabilities to identifying risky behavior that might lead to a phishing attack or download of malicious code. AI allows the automated detection needed to  skim through massive amounts of data and traffic, it can be trained to generate alerts for threats, identify new types of malware and protect sensitive data for organisations.  Leveraging machine learning and deep learning to learn the network’s behavior over time can help recognise patterns, detect anomalies and respond to them.","excerpt":"I strongly believe in the saying: “culture eats strategy for breakfast”. This is highly important when creating a data-driven culture to breathe data and require for every decision.","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Ram Sagar","publish_date":"2021-07-27T17:00:00","publication_year":"2021","word_count":921,"keywords":["Elasticsearch","machine learning","AWS","AI","RAG","Aim","deep learning","Databricks","Kafka","kubernetes","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","deep learning","Aim","RAG","AWS","kubernetes","Kafka","Elasticsearch","Databricks"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/sentinelone-denise-schlesinger-cybersecurity-ai-data-teams\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10163192,"title":"Should India Shift its Focus from LLMs to Large Concept Models?","content":"While many believe India has missed the GenAI bus, jumping on Large Concept Models (LCMs), the next emerging technology, could present a unique opportunity for the country to explore. LCM technology, still at its nascent stage, exhibits characteristics that could help solve complex problems and be a game-changer for India to lead the AI race. But is this the right move? Should India invest in LCMs? Is it a bad idea? Do experts believe there is a better option? The LLM Race is Tough Enough Even though India is leading the charge in AI agents, it is incredibly tough to compete for a leading spot in terms of a foundational model as we know it. Solutions like AI4Bharat and Ola’s Krutrim, among others, have made headlines. Several startups in India are working on AI-powered solutions to automate mundane tasks and have been doing exceptionally well. Despite all this, India is not leading the race, and the reason is lack of capital. Mohandas Pai, head of Aarin Capital and former CFO of Infosys, spoke to AIM last year about this when asked about the lack of innovation from Indian IT. “Who will give $200 million to a startup in India to build an LLM?” he wondered. “Why is nothing like Mistral coming from India? There is nobody…Creating an LLM or a big AI model requires large capital, time, a huge computing facility, and a market. All of which India does not have,” he further said. At a quick glance, many leading LLMs stand out, with countless others joining the list every other day. With India entering the AI race, it may not be an outright “aha moment.” Language Concept Models Are Unexplored, But Is It Useful? Large Concept Models (LCMs) are in the nascent stages of development. However, as the research paper by Meta explains, they could have immense benefits for India. LCMs are language and modality agnostic, meaning they could work well for all diverse language and cultural complexes in India without special efforts. It uses a sentence embedding space, SONAR (Sentence-Level Multimodal and Language-Agnostic Representations), which supports 200 languages, including Hindi, Kannada, and more. The architecture behind an LCM allows it to tackle complex problems (advanced reasoning) better than LLMs, as it emphasises semantic understanding, hierarchical processing, reasoning, and planning. It also supports a larger context length by default because the smallest unit here is a “sentence” or a concept. They are also excellent at document summarisation and interactive content creation, as explained in ADaSci’s blog titled ‘A Deep Dive into Large Concept Models (LCMs)’. So, by shifting their focus from LLMs to LCMs, Indian companies could choose to eliminate putting efforts into Indic language models, discover a new way to reduce computational costs, and tackle complex problems in the country. Will This Help India in the AI Race? While the tech speaks for itself, India recently received a significant boost with a ₹2,000 crore investment in Budget 2025 for the IndiaAI mission. With this in mind, AIM spoke to Amlan Panigrahi, a GenAI engineer at Deloitte, about this idea. “I feel rather than investing in LCMs, which is an initiative of Meta, India should focus on investing in existing Sovereign AI initiatives to solve the nation’s problems first. Entities like Sarvam.AI, OdiaLlama and Indic language models should be targeted more to strengthen the foundation of the country in AI,” he said. He also mentioned that exploring Graph Transformers could be a novel and impactful research direction, as it has the potential to solve contemporary industry problems. Bilal Yoosuf, a senior consultant in data science and engineering at TNP India, spoke to AIM and shared his thoughts, which resonated with a similar sentiment, “India still has a huge opportunity to explore LLMs, especially in the social sector, before shifting focus to LCMs.” “A major part of the population hasn’t even accessed GenAI yet. With the government’s new AI investments, prioritising LLMs for regional languages, education, healthcare, and governance could create a real impact. Once we build strong AI solutions for our own people, we can then look at leading the next wave of AI innovation globally,” Yoosuf added. On the other side, experts like Utsav Khandelwal from Maxim AI believe that by investing early in LCM research, India could position itself as a leader in next-gen AI and mirror its success with cost-effective space missions like ISRO’s. LCM is an uncharted territory for most of the world. It remains to be seen when an AI researcher from India picks it up and explores what it can do for India.","excerpt":"Large Concept Models could be the next big thing. Could it be a good idea for India to go all in?","categories":["AI Features"],"tags":["Large Language models"],"author_name":"Ankush Das","publish_date":"2025-02-11T13:00:15","publication_year":"2025","word_count":760,"keywords":["data science","Go","GenAI","API","AI","Large Language models","innovation","ML","Transformers","Aim","R"],"extracted_tech_keywords":["AI","ML","data science","GenAI","Aim","Transformers","R","Go","API","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/should-india-shift-its-focus-from-llms-to-large-concept-models\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":68845,"title":"How Being A Data Science Generalist Can Be Beneficial In This Uncertain Time","content":"Data science has been no more a luxury for companies — this COVID-19 pandemic has forced companies to rely on data-driven strategies, and thus made data science a key aspect in running businesses. Not only it is helping countries to assess COVID-related data and trends for its citizens but also supporting industries like manufacturing, retail, and e-commerce to sustain in this pandemic. And therefore, companies in India and across the world are looking to include more data scientists in their team to analyse information and help leaders make an informed business decision in this crisis. In fact, despite COVID-19, there are many companies in the news, which shared their hiring interest to work on innovative services and gain a competitive edge among their competitors. However, there has been a significant change in the hiring strategies in the majority of these companies. With cost-cutting being a primary concern for companies, the majority of them are looking to hire data science generalists instead of specialists amid this crisis. In fact, there has been a growing demand for data science generalists since a long time, Gartner called them “citizen data scientists,” who are a jack of all trades — knowing a bit of all the technologies and tools related to data science. Although these generalists aren’t an expert in advanced specialisations within machine learning and deep learning, etc., these generalists are known to create models leveraging predictive analytics as well as a hold on business acumen skills. In fact, it isn’t easy to be a data science generalist. One has to work on various projects, study several subjects, and work in different domains to be competent in this generalised field. Apart from cost-cutting, these generalists can also help bridge the talent gap that the majority of companies are facing in this crisis with analytics work, where they are expecting a mix of data science with engineering and application programming knowledge. Madhavi Kaivalya Kandalam, the vice president and chief data scientist of Loylty Rewardz said in an interview with Analytics India Magazine that corporates tend to look for more generalist data scientists. Also Read: Five Ways Data Science Landscape Will ChangePost COVID-19 A Case For Data Science Generalists Amid COVID-19 This COVID crisis has altered many strategies as to how businesses used to work before this pandemic. Similarly, in the case for data scientists, where earlier specialists were highly favoured in the industry over generalists, it has been changed to hiring more generalists at this uncertain time. A significant reason for this is the inability of specialists to go beyond specific domain knowledge. However, in this tormenting time when things are ambiguous for the foreseeable future, there is a massive requirement for generalists who are well-versed with all aspects and are comfortable to steer through unchartered territories of data science to lay the groundwork. Although a specialist comes with profound knowledge of a particular aspect of data science like NLP and computer vision, a generalist is the one who can actually combine all these together to bring out better business outcomes at a lesser cost. Generalists are capable of communicating insights to business leaders and have skilled knowledge of BI tools like Tableau, SQL, and expertise in algorithm design and optimisation. These skill sets can help companies in their business problems. Alongside, analysing massive amounts of unstructured data and making APIs and dashboards have been a critical concern for businesses to streamline their processes amid this crisis, and generalist data scientists can help businesses achieve it, without any specialised knowledge. Even for startups and small businesses, this crisis has brought immense difficulties in creating a core data science team to innovate on new services, and that’s where generalists come a lot handy. These generalists can come on board for lesser salaries than experts and specialists and also enable companies to look at a project with a different perspective, helping them innovate in new areas of data science. Alongside, with companies becoming more data-driven, leaders are looking for professionals with critical thinking capabilities with a wide range of knowledge, which is an important factor in data science generalists. Also Read: Are MOOCs A Sure Shot Way To Make A Career In Data Science? Encouraging generalists in organisations also enhance the learning curve, as the job roles become more domain and technology agnostic. Alongside, data science generalists can perform diverse functions from the conception of a solution to building a model, and remain flexible in the dynamic environment, which in turn help companies to gain more benefit with less human resources. In fact, if required, when companies mature, they can also turn their generalists to specialists by providing them with necessary training. Lakshya Sivaramakrishnan, Program Lead at Women Techmakers India, said, having a generalist also positively impacts in terms of the number of tasks that a person can get done. “Like instead of hiring five specialists, I can have two generalists who would help me in actually getting the task done, which isn’t the case with specialists,” said Sivaramakrishnan. She further said, “If there is something with specialisation in exploratory data analysis, or in model building, they wouldn’t be able to know or understand the entire workflow of the project to make necessary changes, which can easily be done by a generalist who can bring a breadth of knowledge to the table.” For aspiring data scientists, it is critical to understand that knowing the full picture and understanding the flow of the data science projects is vital for them to actually term themselves data science unicorns. In fact, to create machine learning models, it is necessary for data scientists to understand every aspect of it to make solutions that can actually solve problems. And once aspirants are well-versed in the basics of data science, it’s when they will be able to choose the right specialisation for making a career growth. Admond Lee Kin Lim, data scientists at Micron Technology and a Medium blog writer, wrote on his blog post that, for starting data scientists it is recommended to be a generalist first, and for that joining a startup could be the best way, where not only they get exposure to work with various team and on several projects, but also get the opportunity to create something from scratch. In fact, according to Lee, a marker of an ideal data scientist is someone who has strong general knowledge and is capable of bringing in unique specialities complementing his\/her organisation. Also being a specialised data scientist would restrict the growth of the individual as they will have to look out for a change in the same domain. And therefore for fresh graduates having a generalised knowledge would help them widen their prospects in the market. These generalists also have a leg up on specialists in terms of handling the business, as having expertise across different areas of the field will make them a more preferred choice to lead a team of specialists. Consequently, in this crisis, general data scientists will have a better chance to land on jobs than highly paid specialists. Also Read: How Data Scientists Can Benefit From Certifications While Looking For Jobs Wrapping Up Although specialists bring huge benefits to companies, this COVID pandemic has entirely altered the scenario of the data science landscape. Bearing the costs and sustenance in mind, companies are now looking to get benefit in each of their hirings, and therefore, could rely on generalists to get the maximum benefits. However, for aspiring data scientists, it has been imperative to become a generalist first before actually specialising in a specific domain, which will not only help them choose the right career path but will also help them become a better expert with relevant knowledge across the field.","excerpt":"Data science has been no more a luxury for companies — this COVID-19 pandemic has forced companies to rely on data-driven strategies, and thus made data science a key aspect in running businesses. Not only it is helping countries to assess COVID-related data and trends for its citizens but also supporting industries like manufacturing, retail, […]","categories":["AI Features"],"tags":["big data roi","Data analyst jobs","Data Analytics Certification","Data Science","Data Science Certification","Data Scientist","datascience"],"author_name":"Sejuti Das","publish_date":"2020-07-02T17:00:00","publication_year":"2020","word_count":1286,"keywords":["data science","machine learning","Data Science Certification","AI","datascience","ML","Data analyst jobs","computer vision","Data Analytics Certification","NLP","RAG","deep learning","big data roi","analytics","Data Science","Data Scientist","predictive analytics"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","computer vision","data science","analytics","RAG","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-being-a-data-science-generalist-can-be-beneficial-in-this-uncertain-time\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10001381,"title":"Step-By-Step Guide To Becoming A Proficient Ethical Hacker","content":"Ethical hackers are the good people of the internet who penetrate systems and datasets with legitimate reasons with authorised accessibility. They basically understand the vulnerabilities in a targeted system and try to find solutions to prevent them from attackers. In this article, we are listing the ultimate guide that will help you to get started in your hacking career. 1| Understand The Basics The term hacking can be a controversial one if one does not understand the role between the various types of hackers. Ethical hackers are different from other hackers because they have official certification and permission to access and hack targeted systems. There are various types of hackers like: White Hat Hackers: They are ethical hackers who have authorised access. They can be enrolled in jobs like penetration tester, security analysts, network support, security specialists, network engineers, etc. Black Hat Hackers: They are the cybercriminals who get into a targeted system without any prior authorisation. They are often responsible for data breaching, malware creation, etc. Grey Hat Hacker is a person who gets unauthorised access to the targeted system but reveals the weakness to the company. Green Hat Hacker is a person with limited knowledge of the subject but with each step, tries to understand what they are doing and learn from it. The ethical hackers basically identify the faults in computer security for an organisation, business, etc. in order to protect them from those who have negative intentions. To be a proficient ethical hacker, one must first try to become a skillful Penetration Tester or Pentester where you will recognise the vulnerabilities in a system and shield it in order to protect from attackers. 2| Skills And Academics To become a proficient ethical hacker you need to learn and understand the basic networking concepts as well as different protocols to connect a system remotely. There are various options where one can enroll into for attaining a degree in hacking. Institutional departments such as Computer Science, Computer Information Systems, Information Sytems, etc. provide such options. 3| Understand The Coding Platform Before learning the code, one must have the basic knowledge of operating systems such as Windows, Unix, Linux, etc. and it also crucial to understand the OS level operations of the language you are working with. A need for a strong foundation in coding or scripting language is a must in this field. If you are working on a mobile application platform, you must have knowledge on Java, C# or Swift and if you are working on a web application platform, then you must have knowledge on Python, HTML, PHP, etc. 4| Understand the Technical Issues There are various technical glitches and issues that can occur on a network in an organisation such as lack of encryption, improper DMZ setup, poor logical server grouping, improper firewalls, etc. Sometimes, a number of hacking techniques are needed in order to solve one specific goal. It is then important to think laterally and apply multiple solutions to attain a specific goal. 5| Get Certified Many organisations are providing certifications for teaching ethical hacking courses along with hands-on practices. Certifications in basic networking concepts, security practitioners, security professionals, etc. will help you to step ahead of your hacking career. As we know “Practice makes a man perfect”, the best way to learn “hacking” is to spend more time in practising the exercises which will help you to learn, improve as well as implement new techniques. Besides, this, there are events like MachineHack or other hackathons where one can hone his\/her knowledge and skills. You can practice your skills here as well.","excerpt":"Ethical hackers are the good people of the internet who penetrate systems and datasets with legitimate reasons with authorised accessibility. They basically understand the vulnerabilities in a targeted system and try to find solutions to prevent them from attackers. In this article, we are listing the ultimate guide that will help you to get started […]","categories":["AI Features"],"tags":["Ethical Hacking","hacking"],"author_name":"Ambika Choudhury","publish_date":"2019-03-04T18:24:41","publication_year":"2019","word_count":599,"keywords":["Go","hacking","programming_languages:R","AI","ML","Git","Python","programming_languages:Python","GAN","Ethical Hacking","R","Java"],"extracted_tech_keywords":["AI","ML","Python","R","Go","Java","Git","GAN","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/step-by-step-guide-to-becoming-a-proficient-ethical-hacker\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10170604,"title":"Mistral AI Launches ‘Document AI’ Platform, Claims 99% OCR Accuracy","content":"French AI Startup Mistral AI on Thursday launched a new enterprise-grade Document AI platform, claiming to set a new benchmark in speed and accuracy for OCR-based document processing. The offering, capable of parsing everything from low-resolution scans to handwritten forms, is being positioned as a full-stack solution for businesses dealing with large volumes of paperwork. The company highlights that the platform is powered by a state-of-the-art OCR engine with reported 99%+ accuracy across over 11 global languages. Unlike traditional systems that struggle with mixed layouts, Mistral’s AI can interpret complex documents, including tables, forms, contracts, and invoices, and convert them into structured JSON with custom extraction templates. Processing speeds reportedly reach up to 2,000 pages per minute on a single GPU, making it one of the fastest tools in its category. A demonstration using a decades-old legal contract from Washington Public Power Supply System showed the platform parsing dense paragraphs, legacy formatting and embedded clauses into clearly structured outputs. Even handwritten notes, audit disclaimers and historical equipment delivery records were extracted with accuracy that outperformed legacy systems. Document AI also includes AI tooling for automating full document lifecycles, from digitisation and classification to compliance monitoring. It supports on-premise and private cloud deployment, catering to sectors with strict data sovereignty rules. Mistral’s push into document intelligence follows broader enterprise trends toward digitisation of archives and automation of compliance workflows. For research institutions and multinational firms juggling multilingual paperwork, this could be useful. This comes right after Mistral launched Devstral, an open-source AI model for real-world coding tasks, outperforming peers with a 46.8% score on SWE-Bench. It runs on consumer hardware and is available on platforms like HuggingFace. The company also recently unveiled Mistral Small 3.1, its state-of-the-art multimodal, multilingual, open-source model available under an Apache license. For enterprises still drowning in paperwork, Mistral’s latest bet suggests that OCR might finally be ready for critical workloads.","excerpt":"An enterprise-grade document processing system powered by AI.","categories":["AI News"],"tags":["mistral"],"author_name":"Ankush Das","publish_date":"2025-05-23T15:04:04","publication_year":"2025","word_count":314,"keywords":["Go","TPU","AI","Git","document AI","RAG","automation","Aim","mistral","R","startup"],"extracted_tech_keywords":["AI","Aim","RAG","document AI","TPU","R","Go","Git","automation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mistral-ai-launches-document-ai-platform-claims-99-ocr-accuracy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10129571,"title":"Generative AI is Eerily Bringing Deceased Family Members Back to Life","content":"California-based generative AI startup DeepBrain AI released re;memory 2, a service to help grieving families recreate their deceased loved ones using generative AI, in June this year. This was just the latest development in a fairly new area of generative AI, now being referred to as “grief tech”. Earlier this year, the use of AI to recreate deceased family members gained prominence, with several Chinese companies making waves, including SenseTime who delivered an AI-generated video of their founder Tang Xiao’ou, who had died late last year, to their employees. Shortly after this, families that had come out in favour of using AI modelled on their deceased family members were attacked for making use of the tech. While some said that it was an unethical use of AI, others said that it served to worsen the mourning period. However, while many have made their minds up on generative AI used for grieving, the methods by which this is done has changed significantly, as companies realise the challenges in making chatbots modelled on deceased loved ones to converse with grieving families. The timeline of re;memory’s development reflects this realisation, as it’s initial iteration was in the form of a conversational chatbot. Conversational Chatbots Are Not the Way To Go Speaking with AIM, DeepBrain AI global marketing manager Jungwook (John) Son said that current technology meant that developing AI chatbots modelled on deceased people was not only uncomfortable but also required a lot of time and resources to get done right. “Our first version of re;memory was conversational and it was very limited. The actual target model had to be in our studio to be generated as an AI avatar, which meant pre-preparing ahead of time. Which is, I would say, a very uncomfortable moment,” said Son. Additionally, due to the sensitive nature of the product, having to get the actual likeness and manner of speaking down while having a back and forth was a monumental task. In re;memory’s case, the company had this realisation back in 2023, about a year after its first launch. “We decided to discontinue because we learned the difficulty of actually producing the product. We got some good feedback that’s why we decided to pivot and create a video product rather than a conversational product. So, we relaunched it,” he said. Which seems to be something that other companies that have gone into the field of grief tech have realised as well. Many of the companies offering the services, usually offer a video-based product, rather than a conversational one. I think it's a good idea. Not much different from looking at old photos only better. I've sometimes wondered what it's like for the children of dead movie stars who can see their parents on screen for 90 minutes at a time.— Recruiting Animal (@animal) April 17, 2024 Earlier this year, popular Taiwanese musician Bao Xiaobo had gone viral for generating a video of his deceased daughter using AI. During an interview on the subject, he said, “AI is a tool for expressing yearning, a way to express missing someone.” Similarly, while there are beliefs that AI is being used to recreate or even replace deceased loved ones… Conversational traps – \"I love you, daddy\" and \"promise to keep me safe?\" and \"New Baja Blast Doritos are my favorite, mommy, have you tried them? I want to eat them with you. Say 'yes' and I will order them from Amazon.\"— Wetterschneider (@Stretchedwiener) April 17, 2024 Families choosing to use these services seem to be doing so with an understanding that it’s more of a tool rather than a replacement for grief. The Next Step in Grief Tech Deepbrain’s model is developed in-house by the team, as their primary product is AI avatars, having provided their services to large-scale retailers, finance and banking companies, and even sports teams across the world. In terms of the funeral industry, DeepBrain AI has partnered with one of South Korea’s largest funeral service providers, Preedlife, to offer its re;memory services to grieving families. The current iteration of re;memory generates a video of the family’s deceased loved one, with inputs from the family at every step of the way, from the message they want delivered to the end product. This will be displayed during the funeral of said member, and will also be available to the family, much like looking through a photo album. “It would be something that we discuss with the family. There will be an episode that they want to talk about and then we’ll moderate it and generate the message using memorable scenes.  We try to focus on capture unique facial expressions and voice intonations so that we can try to make it real as possible,” Son said. DeepBrain AI’s shift from conversational chatbots to video-based memorials reflects a broader industry trend, driven by the realisation that creating lifelike digital avatars of deceased loved ones is riddles with difficulties and sensitivities, as well as other factors like limited reach as well. “Because of the limited access, we are currently trying to reach out to more people. So this is why we currently decided to pivot because of the difficulty to create it as a product, whereas now since its a video, it’s more reachable, it’s more continuable,” Son said. Deepbrain AI’s collaboration with Preedlife is the first such partnership to happen. However, with a growing acceptance of grief tech, this partnership could be replicated across the board, as families find a certain comfort in these snapshots of their deceased loved ones. As Son said, “We sincerely think the technology will evolve. The feedback that we get is that this does help families emotionally grieve. Grieving is something that people take very seriously. But with this type of supplement, I would say, it helps process things more smoother and healthier, and that’s the basic approach that we’re trying to have here as well.”","excerpt":"DeepBrain AI’s Jungwook (John) Son said limits on current technology meant that developing AI chatbots modelled on deceased people was not only uncomfortable but also required a lot of time and resources to get done right.","categories":["Deep Tech"],"tags":[],"author_name":"Donna Eva","publish_date":"2024-07-19T11:00:00","publication_year":"2024","word_count":981,"keywords":["Replicate","Go","AI","chatbots","Git","Aim","ViT","generative AI","R","startup"],"extracted_tech_keywords":["AI","generative AI","Aim","chatbots","R","Go","Git","ViT","startup","Replicate"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/generative-ai-is-eerily-bringing-deceased-family-members-back-to-life\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26633,"title":"How Deep Learning Changed The Game For Natural Language Processing","content":"A lot has been written about how deep learning is perfect for natural language understanding. In this article, we will explore why deep learning is uniquely suited to NLP and how deep learning algorithms are giving state-of-the-art results in a slew of tasks such as named entity recognition or sentiment analysis. For example, research scientist Sebastian Ruder points out that word embeddings are one of the most widely known best practices in NLP. Deep NLP — A Huge Step Forward Today, deep learning is delivering state-of-the-art results in practically every NLP-related task. Algorithms such as word2vec and GloVe have been pioneers in the field. Although they cannot be considered as deep learning methods — neural network in word2vec is shallow and GloVe implements a count-based method — the models trained with them are used as input data in applying deep learning for NLP approaches. Word embeddings are now considered as a great practice in the NLP field. For example, the NLP framework spaCy integrates word embeddings and deep learning models for tasks such as NER and Dependency Parsing in a native way, allowing the users to update the models or use their own models. A lot of headway is being made in this area as well. For example, Facebook AI Research (FAIR) lab released fastText, a pre-trained vector in 294 languages, which is reportedly better than GloVe or even word2Vec. Another area where Deep Learning models are being applied with a lot of success is sentiment analysis. The neural networks are implemented in sentiment analysis to compute the belongingness of labels. Sentiment analysis is used to mine subjective text, opinions and sentiments to understand feelings, however, one of the big challenges of sentiment analysis is a lack of labelled data in the NLP field. This is where deep learning plays a pivotal role and deep neural networks, convolutional neural networks and other DL techniques are leveraged to solve a clutch of tasks in sentiment analysis like textual analysis, product review analysis and visual analysis. According to this research paper, Sentiment Analysis Using Deep Learning, DL networks like Recursive Neural Networks, Convolutional Neural Networks, Deep Belief Networks are used for tasks such as word representation estimate, sentence classification, sentence modelling, feature representation and text generation. Here Is A List Of Tasks That Deep Learning Has Revolutionised Named entity recognition — recognises people and places Translation — translates a sentence from one language into another Abstract summarisation — summarises a paragraph Part of speech tagging — assigns a part of speech to a word Parsing Relationship Between LSTM And NLP Long Short Term Memory (LSTM) networks have been around for a long time now. They gave the researchers the ability to train an RNN, and showed remarkable progress in the field of translation. However, this comes with its own set of shortcomings, such as speed limitation or the need for more data. Using LSTMs in production, Google Translate has achieved huge improvements in their machine translation, Google Brain’s Lukasz Kaiser mentioned in a post. Today, some of NLP’s real-life applications include improving support work in call centres, integrating the latest technology in FAQs, providing multi-lingual support and more. Other interesting use cases include automating resume search in India and analysing stock market predictions with deep learning. For example, Microsoft shared a use case of developing a model to predict the stock market performance of companies invested in by a financial services partner. The Redmond giant trained a deep learning model on text in earnings releases and other sources to drum up valuable insights for investment decision maker. One of the key challenges faced by the Microsoft team was building a predictive model that could do a preliminary review financial documents more thoroughly. How India Is Using Machines To Summarise Data India’s well-known Dr Pushpak Bhattacharya, director at the Indian Institute of Technology, Patna and Professor of Computer Science and Engineering at Indian Institute of Technology, Bombay, has collaborated with Elsevier to set up a centre for NLP and machine learning, known as Centre of Excellence at IIT Patna. One of the key objectives of the CoE is developing an automated support system for an article reviewing, especially for journals who face a huge number of entries. Dr Bhattacharyya was quoted by Elsevier saying that the detection of an article whether it is relevant or not relevant is a machine learning problem. He further added that the system being developed compares documents on deep semantic similarity through paragraph vectors. “We created semantic representations of the texts to check how similar they were to existing documents,” he was quoted in Elsevier.","excerpt":"A lot has been written about how deep learning is perfect for natural language understanding. In this article, we will explore why deep learning is uniquely suited to NLP and how deep learning algorithms are giving state-of-the-art results in a slew of tasks such as named entity recognition or sentiment analysis. For example, research scientist […]","categories":["AI Features"],"tags":["deep learning &amp; NLP","lstm","named entity recognition NLP","sentiment analysis NLP"],"author_name":"Richa Bhatia","publish_date":"2018-07-24T07:40:17","publication_year":"2018","word_count":767,"keywords":["Go","machine learning","AI","neural network","sentiment analysis","lstm","sentiment analysis NLP","deep learning &amp; NLP","RAG","NLP","deep learning","spaCy","R","named entity recognition NLP"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","NLP","spaCy","RAG","sentiment analysis","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-deep-learning-changed-the-game-for-natural-language-processing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":64502,"title":"Limitations Of Online Learning For Data Scientists","content":"Data science courses have been keeping many industry professionals and enthusiasts busy amid the lockdown. While data scientists are looking to shield themselves against an oncoming recession by actively upskilling, others are embracing it to make a career shift for the opportunities the field provides. Irrespective of the reasons for this mass shift to online learning, edtech firms are responding to this demand by opening up access to some of the premium course materials, launching additional data science courses, and even making some of them available for free. Although this trend aligns well with the need to continuously learn to enhance career prospects and stay relevant in these uncertain times, it may also indicate an overdependence on e-learning. Distance education, for all its benefits, has its limitations, especially in a field like data science, where practical implementation is paramount. Upskilling should be a part of any data scientist’s career path, but are they relying too much on online courses? Online courses can be a contributing factor, but they cannot build a robust data science portfolio by itself. That is not to say that they are unhelpful, but depending solely on e-learning platforms may not be prudent. Let us attempt to understand why: Difficult To Select Right Courses & Classes Some online courses on well-known platforms may be offering up-to-date content by highly experienced faculty to teach complex concepts, but choosing the right one for your current skill sets and objectives amid an avalanche of free classes can be challenging. There are various online courses and programs available for almost everything across the data science landscape. Moreover, data science covers a wide range of disciplines and is often loosely used by many platforms to convey a few things under this umbrella term. This demands that learners spend a lot of time searching for courses that are suitable for them. Additionally, they have to factor in programs that provide the expected depth in coverage as well as hands-on case study-based learning, with adequate individual attention by instructors. Which is a good segway into the next point. Limited Practical Experience Online courses may equip data scientists with the skills needed to understand its applications in the real-world, but merely having this knowledge without implementing it is not worth much. While these courses may pack a lot of relevant information, learners should not lose sight of the fact that applying those concepts and getting hands-on experience should be the ultimate goal. Despite its benefits, there is no dodging the fact that online learning can be difficult if it is meant for subjects that involve practice, as in the case of data science. Very few courses go beyond theoretical content to provide simulators that allow learners to practice their skills. One way of overcoming this barrier is by implementing learnings on real-time datasets from various domains through internships. Working on industry-based projects on platforms like Kaggle can also provide ample opportunities for practical experience while making data science portfolios more competitive. Employers know that such an exposure will force data scientists to think deeply about a problem, which may be difficult to do with e-learning. Lack Of Direct Interaction While some courses may provide live and interactive classes with the use of webinars and video chats, most – especially those available for free – mostly take recourse to pre-recorded videos and study material. Studying in silos might boost some learners’ productivity, but data science is inherently collaborative in nature and relying on self-study alone is not desirable. Most programs enable discussions between learners as well as instructors, but the scope of these interactions are quite narrow. This may not be the case with full-time data science courses. And without asking and answering questions, participating in discussions, and clearing doubts, learners will miss out on a critical experience that is valuable in data science. The role that collaboration and teamwork plays in the field of data science is often underplayed. But the truth is that data scientists can learn and develop the right way only together – boosting interest and stoking passion for pursuing a career in the field by competing or partnering with each other. Questions Of Credibility As online courses become ubiquitous, questions around the credibility of some of these programs are causing confusion among learners. While not the right approach, some data scientists enrol for courses with the prime objective of updating their resume to seem more qualified than they actually may be. There are plenty of scams that offer certifications for a fee, but may not be licensed. The prevalence of such scams in the online learning market has led companies to especially keep track of accreditations during job interviews. To ensure that you do not sign up for programs from institutes that may not be accredited, conduct thorough research on its background, and if possible, connect with other learners who may have applied for the course. Also Read: Top 10 Full-Time Data Science Courses In India Challenging To Condense Data Science In A Few Courses Owing to the immensity of the field, relying on online learning without adequate support can be challenging for learners. Data science is a mix of various disciplines, including statistics and computer science, and mastering each – or even several – using online tools may not be enough despite the fact that many courses claim to fill the skill gap that the industry has been facing. Online learning can be helpful in picking up specific skills, but putting together a program that condenses complex competencies may be difficult to accomplish. Data science is an ever-changing field that necessitates continuous learning – while much of it can be achieved with the help of digital courses, much more needs to be done for one to be proficient at it. What is more, since it requires a large amount of domain knowledge (given the vastness of the field), learners studying in silos have to prepare for a tough road ahead. However, some courses provide mentorship as well which can help overcome some of these challenges. Outlook The drawbacks of online learning in a field like data science may be many, but some data scientists firmly believe that if leveraged wisely, online courses can be helpful in acquiring the right set of skills to be successful.It may not be a substitute for a regular degree for most jobs in India, but ultimately, what is important is what learners take from it. Treating it as just one of the channels to learn more might be a better approach to take, against collecting online certificates to make resumes more attractive. The projects that one participates in to implement those skills will be more helpful.","excerpt":"Data science courses have been keeping many industry professionals and enthusiasts busy amid the lockdown. While data scientists are looking to shield themselves against an oncoming recession by actively upskilling, others are embracing it to make a career shift for the opportunities the field provides. Irrespective of the reasons for this mass shift to online […]","categories":["AI Features"],"tags":["Data Analytics Certification","Data Science","Data Science Certification","Data Scientist","edtech","online education","Online Learning"],"author_name":"Anu Thomas","publish_date":"2020-05-05T16:00:00","publication_year":"2020","word_count":1111,"keywords":["data science","Go","programming_languages:R","Data Science Certification","AI","ML","edtech","Git","Data Analytics Certification","RAG","Aim","online education","Online Learning","ViT","Data Science","Data Scientist","R"],"extracted_tech_keywords":["AI","ML","data science","Aim","RAG","R","Go","Git","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/limitations-of-online-learning-for-data-scientists-looking-to-upskill\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10046989,"title":"Agritech Robotics Startup TartanSense Raises $5 Million In Series A","content":"Bangalore-based Agritech robotics startup — TartanSense has raised $5 million in Series A funding. The round was led by FMC Ventures and Omnivore, with participation from existing investor Blume Ventures. Founded in 2015 by Jaisimha Rao, an alumnus of Carnegie Mellon University, TartanSense builds small agricultural robots equipped with AI-assisted computer vision to help small farms reduce their expenditures and improve their incomes. “Our mission is to make smallholder farmers wealthier by shipping monetisable robots. TartanSense will have the world’s largest fleet of agriculture robots in the next 18 months. We are grateful to have amazing investors like FMC Ventures, Omnivore, and Blume Ventures backing us in our passion to empower farmers,” said Jaisimha Rao, Founder, TartanSense. TartanSense is helping smallholder farmers who struggle with low yields, primarily driven by two reasons – poor chemical spraying techniques and unreliable farm labour. TartanSense’s robots are an affordable precision agriculture solution for all major farming activities – sowing, spraying, weeding, and harvesting – which simultaneously drive down cultivation costs while improving crop yields. Amar Singh, Managing Director at FMC Ventures, remarked, “TartanSense is a pioneer in ground-based precision spraying in India. With growers’ interest in mind, it has developed a unique, low-cost precision application technology with a very high level of accuracy. FMC Ventures is excited to support TartanSense as they combine artificial intelligence and robotics to improve how growers apply crop inputs.” This funding round brings the total funds raised by the company to $7 million, after raising a $2 million seed round in March 2019.","excerpt":"This funding round brings the total funds raised by the company to $7 million, after raising a $2 million seed round in March 2019.","categories":["AI News"],"tags":["AI in agriculture"],"author_name":"kumar Gandharv","publish_date":"2021-08-25T11:28:34","publication_year":"2021","word_count":255,"keywords":["funding","artificial intelligence","programming_languages:R","AI","computer vision","Ray","AI in agriculture","ViT","R","ai_applications:computer vision","startup"],"extracted_tech_keywords":["AI","artificial intelligence","computer vision","Ray","R","ViT","startup","funding","programming_languages:R","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/agritech-robotics-startup-tartansense-raises-5-million-in-series-a\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162902,"title":"MLDS 2025: Key Highlights from Day 1","content":"Day one of MLDS 2025, India’s biggest GenAI summit for developers, hosted by AIM Media House, was a day filled with energy, excitement, and forward-thinking discussions. The day witnessed the presence of close to 1000 attendees. One of the day’s interesting talks came from Rahul Bhattacharya, AI leader at GDS consulting, EY, who delved into the world of the agentic workforce, the next frontier in AI-driven workspaces. Bhattacharya highlighted how assessing the risks of integrating AI agents into the workforce is just as crucial as measuring the benefits. He broke down what makes a system truly “agentic.” To be considered an agent, Bhattacharya explained, a system must have the ability to interact with its environment, make decisions based on its observations, and learn from its actions to continuously improve. He elaborated, “A key ability is making decisions, where the agent chooses the best action based on set rules, goals, or rewards. Over time, it should learn from past experiences and feedback to improve its performance.” This blend of adaptability and decision-making could redefine how we think about work in the future. Rahul Bhattacharya, AI leader at GDS consulting, EY In the digital content world, where video has taken center stage, Arvind Sasikumar, co-founder and CTO at Quinn, shared his insights on the critical importance of video compression. Sasikumar emphasised that optimising video transcoding isn’t just about reducing file sizes, but it’s about enhancing the user experience by eliminating buffering, which can be a dealbreaker for viewers. He explained how even a 5% reduction in file size can make all the difference for users with lower internet speeds. “Consider this: If 100 users have a 1.9 Mbps connection but the video they’re watching requires a 2 Mbps bitrate, every single one of them will face buffering. But by optimising compression, we can ensure that all 100 users enjoy smooth playback,” Sasikumar said. Arvind Sasikumar, co-founder and CTO at Quinn As organisations race to integrate AI into their business strategies, there are hurdles to overcome. Chirag Jain, vice president of AI Practice at Genpact, took the stage to shed light on why businesses must first define their long-term goals before jumping into AI implementation. He compared it to the vision of a sports team– just as India’s cricket team aspires to be the best in the world, businesses also require a clear, long-term goal to guide their AI strategy. Jain underscored that a strategy grounded in core values, ethical considerations, and regulatory compliance is key to ensuring AI’s success within an organisation. “Nobody likes to see a team playing unfairly,” he said, reinforcing the need for transparency and ethical decision-making in AI adoption. Chirag Jain, vice president of AI Practice at Genpact The day also featured Siddhant Goswami, co-founder at 100xEngineers and Tech Influencer, who spoke on the evolving role of AI agents. While AI agents are becoming increasingly autonomous, Goswami reminded the audience that human intervention is still essential. “We can’t completely rely autonomously on them. We need human feedback. These agents are self-directed, capable of planning and executing multiple steps based on feedback and updated goals,” Goswami shared. Siddhant Goswami, co-founder at 100xEngineers and Tech Influencer The excitement continued with a workshop by Gopala Dhar, an AI engineer at Google Cloud, and Lavi Nigam, a developer relations engineer at Google Cloud, on building real-time applications with the Gemini Multimodal Live API. This hands-on workshop served as a holistic guide to using Gemini’s multimodal capabilities to create sophisticated applications that can see, hear, and interact naturally. Also, the day witnessed several paper presentations, including Life Stage Customer Segmentation by Fine-tuning Large Language Models by Nikita Katyal, Head of Analytics and AI at Central Retail Corporation, and Vivek Vishwas Vichare; A Context-Aware Multi-Agentic Multi-Modal LLM Architecture for Digital Marketing by Vivek Vishwas Vichare, head of data sciences and analytics at Pixis, and more. These papers showcased the innovative ways AI and generative models are reshaping industries like retail, digital marketing, and healthcare, demonstrating the immense potential of AI to solve complex problems across various domains. Day one of MLDS 2025 left attendees with plenty to think about.","excerpt":"The day witnessed the presence of close to 1000 attendees.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Machine Learning","MLDS"],"author_name":"Vidyashree Srinivas","publish_date":"2025-02-05T17:51:39","publication_year":"2025","word_count":685,"keywords":["data science","Go","GenAI","AI","ML","Machine Learning","Git","RAG","MLDS","Aim","analytics","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","data science","analytics","GenAI","Aim","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/mlds-2025-key-highlights-from-day-1\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10019877,"title":"In Conversation With Sanjana Krishnan And Christian Franz From CPC Analytics","content":"This is the seventh article in the weekly series of Expert’s Opinion, where we talk to academics who study AI or other emerging technologies and their impact on society and the world. This week, we spoke to Sanjana Krishnan, Partner and Christian Franz, Co-Founder and Partner, at CPC Analytics, a data-driven policy-consulting firm with offices in Pune and Berlin Sanjana Krishnan is a graduate from the Tata Institute of Social Sciences in Urban Policy and Governance, and has worked with state governments on open data and is a member of the DataMeet Open Data Community in Pune. Christian Franz has a Masters in Public Policy from the Hertie School of Governance and has worked on several healthcare and data projects. Analytics India Magazine caught up with the duo to get insights into their recent research on the Cross-border Data Trade For AI done in collaboration with Bertelsmann Stiftung in Germany. AIM: What is the significance of cross-border data trade for AI? Does Europe need it in the light of various digital strategies the EU has employed over the last few years? Krishnan and Franz: The European Commission has certainly launched an impressive set of policies over the past few years. It will take some time until the policies are filled with life in the 27 member states of the Union. With its data strategy, the Commission has shown that it has understood a truism of the data & AI economy: Next to computational power, a key ingredient of current AI applications is access to vast amounts of data. Having access to datasets that fit the demanding requirements of the current AI-workhorse technology and machine learning is becoming a necessity. Firms adopt several methods to build such datasets. Large Silicon Valley-based firms have succeeded because of data network effects with a large user base for their products, the accretion of data increases resulting in a better-trained product that further attracts users. Chinese firms have built their databases mostly from their vast domestic (and highly competitive) market activity. For specific use cases such as training autonomous vehicles (identifying pedestrians, road signals and traffic lights), or identifying cancerous tumours and brain lesions, the data is first manually labelled and annotated by human ‘data labellers’ so that the AI systems can identify and learn from it. In the absence of this, AI relevant data could also be ‘traded’ between firms at a global scale by easing barriers to cross-border data transfer. The EU’s push for common data markets within different industries could be simply seen as an approach to ensure that the wealth stored in data does not anymore remain locked in different data silos but can be used by companies to build more innovative products. AIM: What are some of the major roadblocks when it comes to cross-border data trade and what would it take to establish a global data sharing act? Krishnan and Franz: How challenging cross-border sharing across countries is, can also be observed in the European Union. Data protection and data privacy is just one important piece of governance that needs to be in place. Let me give you an example: Just think of the difficulties we face in any given hospital around the world: A myriad of different IT-systems that are not designed to speak to each other. Data is entered in many different ways and situations making much of the human-generated data unlikely to be useful for analysis that’s going beyond mere monitoring. And lastly, data generated in the regular processes a hospital represents a valuable asset – what incentive does the hospital have to share the data? I am giving you this example because it is vital to consider the challenges at a micro-level before elevating the entire discourse on the geopolitical level. Yes, we see a push to make technology another dimension of industrial policy and geopolitics. And rightly so! But let’s face it: There are solutions to establishing cross-border transfers of data. In our conversations with experts from India, we have often heard concerns about GDPR compliance. At the same time, when asked whether the experts’ companies have found ways to establish the cross-border transfer of those data, they all said yes. In reality, I do not believe it is an insurmountable obstacle. A global data sharing act is unrealistic at this point. What we argue is that it will suffice for now to build bilateral data exchange platforms, support the exchange of knowledge among practitioners and build common standards from within communities of practice. AIM: From a bilateral perspective, do both countries or regions need to have strong data protection rules and frameworks or can one country’s laws be leveraged for the trade? Krishnan and Franz: The most important action would be to pass and enforce robust regulation on data protection. Even though the Supreme Court has recognised privacy as a fundamental right under Article 21, the current state of data protection is not adequate to ensure the data collected in India and crossing Indian borders adheres to the expected ethical and data protection standards. Strong implementation of the PDP bill and establishment of a DPA promise considerably higher levels of data protection in the country. This is a necessity to ensure adequate protection for data from India and increase the trust of its trading partners, specifically countries in the EU that have set high standards with the GDPR. Europe’s data economy is estimated to reach a volume of €550 billion by 2025 (4% of the overall EU GDP). A report of MeitY declared that the digital economy of India (somewhat broader definition) will be $1tn by 2025. The opportunity of building strong connections between these two markets is too big to pass because data privacy standards in India are not established or enforced in an insufficient way. AIM: Considering both parties have robust privacy laws, what are some of the major roadblocks when it comes to cross-border data sharing? Krishnan and Franz: The first requirement is to enhance the quality of data and ensure that it meets the technical specifications that AI requires. Most of the data generated in India in its current form is not fungible, tradeable and hence not usable by AI across borders and contexts. A concerted effort to produce large, comprehensive, labelled datasets is needed. Simultaneously, the possibility of bias also needs to be recognised and accounted for, especially when AI directly interacts with humans (such as chatbots) or impacts their lives. The risk of bias – if not considered – will significantly damage the trust in the application. Second, India needs to urgently and strategically beef up its internal capacity to become a relevant player in the AI landscape. While India ranks low on digital skills, there is a larger percentage of people enrolling in AI courses which shows the anticipated demand of these skills in the market as well as the readiness of talent to adapt to the change. A high hiring rate for those with AI skills (higher than the US, Germany and France) indicates dynamism in the demand for AI talent. In the private sector, while India ranks high on the number of AI-based start-ups (higher than China), US and China far outstrip India when it comes to private investment in AI. India also needs to increase its investment in research. AIM: The study mentions the threat of ‘data colonisation’. Going ahead, how do we avoid the exploitation of India’s data resources and ensure both parties get equal benefits from the cross-border data exchange? Krishnan and Franz: When speaking of India as a “data-rich” country – a statement that has motivated many actions in addition to the publication of the study that we did– we always run the risk of following such a narrative. A one-dimensional framing of India as a strategic reservoir of data for the development of AI for the benefit of the trading partners’ domestic AI firms comes dangerously close to echoing historical colonial practices of economic extraction. It is no coincidence that researchers and activists from various disciplines and regions have warned against an increasing “digital colonialism,” “data colonialism” and “algorithmic colonisation.”. Any meaningful cross-border data partnership for AI will have to consider these different power dimensions in the digital world. They should help create a common knowledge base for a future, a balanced partnership of equals. AI is a technology that can lead to discriminatory, marginalising and outright violent outcomes. Like India has done in its National Strategy for AI, partner countries should subscribe to a vision of AI that reinforces ethical values and benefits all of society. This vision also needs to be embedded into the efforts meant to create a closer data exchange for AI development.","excerpt":"This is the seventh article in the weekly series of Expert’s Opinion, where we talk to academics who study AI or other emerging technologies and their impact on society and the world. This week, we spoke to Sanjana Krishnan, Partner and Christian Franz, Co-Founder and Partner, at CPC Analytics, a data-driven policy-consulting firm with offices […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Kashyap Raibagi","publish_date":"2021-02-09T14:00:00","publication_year":"2021","word_count":1448,"keywords":["Go","machine learning","AWS","AI","chatbots","RAG","Aim","analytics","Rust","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","chatbots","AWS","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/in-conversation-with-sanjana-krishnan-and-christian-franz-from-cpc-analytics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10143392,"title":"OpenAI Teases ‘Super Secret AGI’","content":"On the fifth day of 12 Days of OpenAI, Sam Altman and co. took the opportunity to demonstrate how ChatGPT seamlessly integrates into the native AI experience inside an iPhone and a Mac. While the main announcement took centerstage, OpenAI intrigued its fans and users with an easter egg. On the iPhone used for the demo, a calendar event titled ‘Super Secret AGI’ seems to be scheduled. On the surface, it might look like an inside joke or vision to achieve AGI in the near future. Who knows? It is indeed possible that all of OpenAI’s announcements, cumulatively at the end of the 12 days, might give a sense of clarity as to what artificial general intelligence will look like and feel. Simply put, if people aren’t feeling the AGI yet, we don’t know what AGI really is. For instance, OpenAI’s recent release of its video generation model, Sora, gives an early glimpse of AGI—one frame at a time. This new model from the company—now accessible to Pro and Plus users—is leveraging video as a medium for training models in real-world logic and dynamics. “We’re just scratching the surface of what’s possible,” said Sam Altman. OpenAI’s Gift to Apple Apple has released the latest iOS\/iPadOS 18.2 and the macOS Sequoia update, which includes an updated feature set for Apple Intelligence. These include the Image Playground, Genmoji, an upgrade for the Writing Tools, and Siri – which are now powered by ChatGPT. Siri can now fetch information and respond to a user’s query using ChatGPT – but only when the user specifically asks it to use ChatGPT. The new Compose features let one access ChatGPT to generate text content on any writing tool across the system. Moreover, ChatGPT also aids the Visual Intelligence feature, which provides context to images from the user’s real-world environment. The new Camera Control button on the iPhone 16 Pro allows the user to quickly open up Visual Intelligence. On the Mac, OpenAI also demonstrated using Siri with ChatGPT to ask questions based on a 49-page PDF document. ChatGPT is available as an opt-in-only extension inside Apple Intelligence settings. One doesn’t necessarily need an account to use ChatGPT in Apple Intelligence. Most importantly, Apple Intelligence is now available with localised English support for Australia, Canada, Ireland, New Zealand, South Africa, and the U.K. – while the official release for English (India) is set to be launched soon. Apple’s partnership with OpenAI is interesting, as it has been reported that the Cupertino giant isn’t paying OpenAI anything for the integration of ChatGPT on its devices. Rather, Apple sees value in adding ChatGPT to millions of iPhone, iPad and Mac users across the globe. It was also said that Apple was granted a board observer role at OpenAI, which was soon dropped after a few weeks. Furthermore, in October, Apple also reportedly backed out of OpenAI’s $6 billion funding round. Other features focus on adding text to visual tools to Apple devices. The Image Playground lets users generate images with styles and illustrations of their own and also supports creating visuals ‘likeness of a family member or friend using photos’. Similarly, Genmoji lets the user create an emoji using a prompt, and like Image Playground, it allows using photos to create an emoji resembling real people. That said, Apple is still lagging behind Google in their AI offerings. More so, given the fact that Google dropped huge updates to Gemini 2.0. While it is only available via the API for developers, it will surely arrive for Pixel devices and other Android phones in due time. Although still in the testing stages, Project Astra received several upgrades, which now execute prompts inside tools like Google Search, Maps and Lens. The feature has been thoroughly tested on Android phones since it was launched at Google I\/O this year. When it arrives on Android devices, Apple will have to seriously pull its socks up!","excerpt":"Apple is nothing without OpenAI—or is it?","categories":["AI News"],"tags":["AI (Artificial Intelligence)","OpenAI"],"author_name":"Supreeth Koundinya","publish_date":"2024-12-12T11:46:43","publication_year":"2024","word_count":652,"keywords":["Go","ChatGPT","Gemini 2.0","API","OpenAI","AI","ML","RAG","GPT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","ChatGPT","OpenAI","Gemini 2.0","RAG","R","Go","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-teases-super-secret-agi\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":15786,"title":"Julia Computing secures 4.6M in seed funding from General Catalyst and Founder Collective investors","content":"Viral Shah, CEO, Julia Computing Julia computing has bagged $4.6M in Seed Funding from General Catalyst and Founder Collective investors.  One of the fastest modern high performance open source computing language for machine learning, AI and data analytics jobs, Julia was born out of the open source community. The language has a huge developer base and is known for its speed, capacity and productivity. Listed among the top 10 programming languages to be developed, and with more than 1 million downloads, Julia provides parallel computing capabilities out of the box and unlimited scalability with minimal effort. According to Viral Shah, Julia Computing CEO, “We selected General Catalyst and Founder Collective as our initial investors because of their success backing entrepreneurs with business models based on open source software. The funding will help us accelerate the product development and continue delivering outstanding support to our customers, while the entire Julia community benefits from Julia Computing’s contributions to the Julia open source programming language.” Some of the top employers hiring Julia programmers are Apple, Google, Intel, Microsoft, Oracle,PwC, uber, Disney, Facebook, Ford, Amazon, ComCast among other big names. One of the best features of Julia is the ease of use, as it combines the syntax of Python, R, MATLAB and the speed of C and Java. The company was founded in 2015 by Viral Shah, Alan Edelman, Jeff Bezanson, Stefan Karpinski, Keno Fischer and Deepak Vinchhi who created the open source language to develop products and provide support to businesses that use the programming language.","excerpt":"Julia computing has bagged $4.6M in Seed Funding from General Catalyst and Founder Collective investors.  One of the fastest modern high performance open source computing language for machine learning, AI and data analytics jobs, Julia was born out of the open source community. The language has a huge developer base and is known for its […]","categories":["AI News"],"tags":["easy python beginner projects","fun beginner python projects","Julia Language","julia language adoption"],"author_name":"Richa Bhatia","publish_date":"2017-06-21T08:24:24","publication_year":"2017","word_count":253,"keywords":["julia language adoption","Go","machine learning","AI","Julia","Scala","Python","Julia Language","ViT","fun beginner python projects","analytics","easy python beginner projects","R","Java"],"extracted_tech_keywords":["AI","machine learning","analytics","Python","R","Go","Java","Scala","Julia","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/julia-computing-secures-4-6m-seed-funding-general-catalyst-founder-collective-investors\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10092091,"title":"7 Best Quantum Computing Research Institutes in India","content":"Quantum is setting out to disrupt computing as we know it. Governments across the world have proposed outlays, built strategies, and formed collaborations to be the first to harness quantum power. And India is no exception. The National Mission on Quantum Technologies & Applications (NMQTA) is programmed to take advantage of the disruptive potential of quantum technologies in computation, communication, and security. Since World Quantum Day has just gone by, we at Analytics India Magazine sought to identify a few eminent institutions in the quantum space that are helping India advance its quantum mission. Raman Research Institute (RRI) Awarded a grant from the Ministry of Electronics and Information Technology, RRI’s project currently targets three areas: quantum communications, quantum sensing, and quantum interactions, in collaboration with the Indian Institute of Science (IISc) and the Center for Development of Advanced Computing (C-DAC). The project was formed with two more Bangalore-based institutes to develop quantum technology in secure quantum communication as well as quantum computing-related methods. Earlier this month, RRI’s Quantum Information and Computing (QuIC) Lab headed by Urbasi Sinha, successfully demonstrated a secure satellite-based quantum communication in collaboration with ISRO. The lab will also lead the research efforts in Quantum Key Distribution (QKD) techniques for secure maritime communications. International Institute of Information Technology (IIIT), Hyderabad The Centre for Quantum Science and Technology (CQST) at IIIT Hyderabad carries out research on the role of quantum science in computation, information processing, quantum thermodynamics, and on the foundations of physics. The division is headed by Arun Kumar Pati, co-author of a fundamental theorem related to quantum information theory. Together with RRI, the institute also represented India at this year’s World Quantum Day (celebrated on April 14), which aimed to promote public understanding of Quantum Science and Technology globally and involved participation from 65 other nations. Harish-Chandra Research Institute Located in Prayagraj, the institute has a strong research presence on arXiv, an online repository of scientific research papers. Spearheaded by professors like Ujjwal Sen and Aditi Sen De, the Quantum Information and Computation (QIC) Group is actively involved in research surrounding quantum algorithms, quantum communication, quantum cryptography, and the theory of entanglement, and is also exploring the interface between quantum many-body physics and quantum information, an emerging field of research. Tata Institute of Fundamental Research (TIFR), Mumbai The Quantum Measurement and Control Laboratory (QuMaC) at TIFR was founded by Dr R Vijayaraghavan, who also currently heads it. The primary objective of the laboratory is to develop techniques for stabilising quantum states against decoherence. It explores quantum phenomena in superconducting circuits, which can store and process information. This research aims to address the fundamental challenges of building and controlling quantum systems that can offer exponential speed up for solving mathematical problems. Indian Institute of Science Education and Research (IISER), Mohali The NMR Research Group at IISER, Mohali is involved in diverse research areas including experimental quantum computing, multipartite entanglement characterization, tomography via machine learning, decoherence mitigation, and experimental implementation of quantum algorithms. The group was able to achieve the first experimental demonstration of the super-Zeno effect to free evolution of quantum states, using spin-1\/2 particles to represent qubits. The group also demonstrated the first experimental exploitation of a single qutrit to achieve a computational speedup on an NMR quantum computer. The IISER team of scientists comprising of Prof Arvind, Prof Kavita Dorai, Dr Mandip Singh and Dr Sandeep Goyal are also involved in building India’s first quantum computer under the Cyber Physical Systems (CPS) programme by the Department of Science and Technology. Indian Institute of Science The Center for Excellence in Quantum Technology (CEQT) operates an experimental program that focuses on superconducting qubit devices, single photon sources and detectors for quantum communications, integrated photonic quantum networks, and quantum sensors. The institute has also launched joint courses on AI and quantum computing in collaboration with QpiAI. In addition, CEQT regularly hosts workshops and meetings in the field to maintain the quantum momentum. As part of its vision, the institute aims to establish international collaborations in this field and seeks guidance from an international advisory committee to guide its research efforts. Indian Institute of Technology (IIT), Madras The Centre for Quantum Information, Communication, and Computing (CQuICC) is the research division sponsored by the IT giant Mphasis and works with various partner institutions. The research focus areas include quantum key distribution, quantum sensors, quantum computing, post-quantum cryptography, quantum error communication, quantum communication, and quantum algorithms. The team also includes Anil Prabhakar, co-founder of QNu Labs and Quanfluence, both incubated by IIT Madras, as well as, Arul Lakshminarayan, who along with IIT Madras’ Suhail Ahmad, and in collaboration with Polish researchers and the Polish Academy of Sciences, provided a quantum solution to the 243-year-old problem proposed by Euler, known as the ‘36 officers problem’. Along with IIT Madras, IIT Jodhpur and IIT Bombay are also making significant strides in quantum research. IIT Bombay’s Prof Kasturi Saha, who is heading the Photonics and Quantum Enabled Sensing Technology (P-Quest) Laboratory, is among the very few in India working with NV-diamond technology to build quantum sensing applications.","excerpt":"Governments across the world have proposed outlays and built strategies to be the first to harness quantum power","categories":["AI Trends"],"tags":["c-dac","Quantum Computing"],"author_name":"Ayush Jain","publish_date":"2023-04-21T15:00:00","publication_year":"2023","word_count":843,"keywords":["Quantum Computing","Go","machine learning","programming_languages:R","AI","RAG","Ray","Aim","ViT","analytics","c-dac","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","Ray","RAG","R","Go","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-best-quantum-computing-research-institutes-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10050176,"title":"How to Learn From Streaming Data with Creme in Python?","content":"In a traditional paradigm of machine learning, we often work in the offline learning fashion where we start with data preprocessing and end with data modelling with an algorithm to satisfy the requirements. This becomes a storage-dependent and time-consuming process. To overcome this, we can use streaming data for predictive analysis or any other modelling process. We don’t need to store the data before modelling it. This can be accomplished by stream learning and online learning. In this article, we will understand how we can make streaming data useful in machine learning. We will also learn how to implement online machine learning to learn from streaming data. The major points that we will cover in this are listed below. Table of Contents What is Streaming Data?Online Machine LearningAbout CremeImplementing Online Learning Using Creme What is Streaming Data? The data which is generated continuously in an incremental manner from different sources can be considered as the streaming data. Basically, this type of data is generated frequently and flowing across different websites and platforms which are designed to provide information as a function of time. Usually, streaming data is used in the context of big data where it is generated using different sources at stable or unstable high speed. In the context of technology, it is a system that is used for providing information to devices over the internet by allowing users to access the content immediately instead of using the content offline or after downloading it. As we know, big data mainly focuses on the storage of the data which causes a huge amount of cost for storage and the storage consists of the tendency to make the data unstructured. This is not fruitful for the machine learning algorithms. Datastream is a sequence of digitally encoded coherent signals used to transfer the data that is in the data transfer process. Datastream can be considered as set to information that is extracted from the receiver and provided by the data provider. Using various methods we can make streaming data useful for us like making reports, providing on-demand data-driven decisions,  and machine learning algorithms. We can extract useful information from a data stream or many data streams for modelling purposes. Since the offline machine learning models work on a trained or offline data, in the case of modelling with streaming data online machine learning comes into the picture. Online Machine Learning Online machine learning is a method that combines the machine learning models or predictors with the data which comes in the sequential order so that the best predictor can be used to provide updates according to the future data at every step. In a traditional machine learning program, we use stable data to predict the values for which the model is designed where on the other hand online machine learning programs are designed to adopt the new patterns in the data or adopt the data when it is generated as a function of time. Online machine learning models have their own benefits and uses like the changes and updates on the model can be done automatically in case of changes in the data and they are highly preferable models in the fields like finance market, economics where the new data is emerging on a frequent basis. The above image represents the flow chart of any online machine learning system where streaming data can be used to train an online machine learning algorithm while the precision label of the model is not enough we can analyze the problem to make the algorithm perform better with the new data. There are various frameworks available to perform online machine learning. A few of them are: JubatusThe tornado frameworkcremeScikit-learn This article is mainly focused on the creme library which provides the functionality for performing online machine learning on streaming data. So let us understand creme in detail with its implementation. About Creme The creme is a python library for online machine learning which provides most of the packages designed especially for online machine learning. Using the library we can learn the stream of data using different approaches. This library allows models to learn one data point at a time so that updates can be done if required. This approach helps to learn from big data that isn’t stored in the main memory. The library gives integration options to online machine learning where the new data stream is constantly arriving. The following benefits can be achieved using creme: Real-time updated models – it is the basic feature of the models given under the cremes package that they are incremental according to time.Models under the creme can adapt to the concept drift problem, which means they try to predict changes in target variables over time. During model development, the creme can represent the production scenario even when working with data streams.Models under the creme are designed to require little computation and they don’t need to be retrained.Models are designed for learning each observation at a time. The following are the features of the creme library: Under the roof of the creme, we have various Nearest Neighbors, Decision Trees, and Naive Bayes models.Library provides k-fold and cross-validation, progressive model validation methods.We have models for recommendation systems, time series forecasting, and also linear models with a variety of optimizers.In the case of unsupervised learning, the library provides the feature for clustering.They have inbuilt class imbalanced learning features in the package. With all these facilities we have the option for feature extraction and selection in the packages.Various built-in datasets like airline passengers, chick-weights, fraud-detection, and many others so that we can learn the usage of the library easily.We have an anomaly detection feature in the library. Implementing Online Learning Using Creme Now, let us implement a linear model that is a logistic regression for binary classification using the creme library. Installation The basic requirement to work with the creme is python 3.6 or above. We can install the library using pip and the following command: !pip install creme Now let’s start building a logistic regression model for classifying the website phishing dataset. Before going for modelling let’s check for the entities dataset consists of: Importing dataset from the creme library from creme import datasets A_b = datasets.Phishing() print(A_b) Output: The above output represents some basic details of the phishing dataset. Let’s divide the dataset into dependent and independent variables. from pprint import pprint for A, b in A_b: pprint(A) print(\"independent variable =\", b) break Output: Now we can train a logistic regression model on the data in a streaming fashion. Building the Model Importing the packages for data preprocessing, models, and accuracy metrics. from creme import compose, linear_model, metrics, preprocessing Defining a pipeline for a model instance and scaling the data. lm = compose.Pipeline( preprocessing.StandardScaler(), linear_model.LogisticRegression() ) pprint(model) Output: Making an accuracy metrics instance: metric = metrics.Accuracy() Making a prediction using a model instance and sequentially updating the model we can also update an accuracy metric using the metric.update module. for A, b in A_b: pred = lm.predict_one(A) metric = metric.update(b, pred) model = lm.fit_one(A, b) After running the above-given code we can check for the accuracy of the model which we have updated in the loop. print(metric) Output: Here we have created and trained a linear regression model by interleaving predictions and model updates.  The model has performed well and we can see the accuracy of the model is around 89% which is pretty good. There are various tasks we can perform using the creme library as we have explained in the features. Final words In this article, we had an overview of streaming data and online machine learning. There are various application domains of online machine learning such as time series forecasting, spam filtering, recommender systems, CTR prediction, and IoT applications. The creme library provides most of the features of stream learning that are required in online learning. We can use them according to the requirements. References: Creme Documentation Online machine learningStreaming DataGoogle Colab Notebook for above codes","excerpt":"The data which is generated continuously in an incremental manner from different sources can be considered as the streaming data.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","anomaly detection","Big Data","Big Data Analytics","Data Analytics","Data Science","Data Scientist","Deep Learning","how to measure twitter influence","how to retrain data","load data python","Machine Learning","online machine learning","Python"],"author_name":"Yugesh Verma","publish_date":"2021-10-02T16:00:00","publication_year":"2021","word_count":1329,"keywords":["scikit-learn","TPU","R","load data python","how to retrain data","how to measure twitter influence","online machine learning","RAG","Data Science","Big Data","machine learning","AI","Machine Learning","recommendation systems","Big Data Analytics","Python","Colab","anomaly detection","Data Analytics","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","scikit-learn","Colab","RAG","recommendation systems","anomaly detection","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-learn-from-streaming-data-with-creme-in-python\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10017364,"title":"Will More Robots Work Out Of Cages?","content":"Since the advent of the robotics industry, robots have mostly been used for automating industrial processes and have operated on production shop floors in cages. This trend is drastically changing now as robots are currently working in every-day settings alongside humans and having human-like features. With the rapid development in emerging technologies, robots with applications outside factories are becoming safe for adoption. Various other non-technical factors have also contributed to the growth in non-factory robots, also otherwise called service robots. This article analyses the recent growth and market trends in service robotics and the various technical and non-technical reasons behind it. Recent Growth And Market Predictions According to market reports, the service robotics industry will reach a value of $37 billion between 2019 and 2021. By 2025, this industry is predicted to grow to $102.5 billion with a compounded annual growth rate of 22.6%. Multiple applications of service robots provide opportunities for growth in the industry and beyond. Investors already see this potential as service robotics companies see huge sums of funding coming their way. Sectors like healthcare, warehouses, hospitality, and agriculture provide a huge scope for commercial applications of service robots. Each of these industries has unique requirements in terms of adopting processes. Thus, service robotics companies focusing only on sector-specific applications have seen good results and have received more funding in the past couple of years. Bossa Nova, for instance, that develops robots to assist humans in warehouses raised a total of $101.6 million in the last five years. Diligent, which provides service robots to perform various hospital tasks, raised $15.8 million since 2017. Service robots are also gaining more adoption in domestic applications. Smart robots with advanced features for education, cleaning, entertainment, or other household purposes are becoming attractive. Rental options by companies make it more affordable. A Chinese UBTech Robotics that manufactures domestic robots for education and entertainment purposes raised $940 million in 2018. Service robots are also finding applications on roads in the form of function-specific automated guided vehicles. The year 2020 saw some of the highest investments go to self-driving car companies like Cruise, Waymo, and Nuro. The three companies together have raised funding of more than $9.7 billion in the last two years. Favourable factors for adoption and high investment activity by some of the top venture capitalists in the service robotics is indicative of the future it holds for the industry. Favourable Enablers For Service Robots Since non-factory robots need to work in an environment without fixed dimensions, they need better control over their movement. For safe and successful adoption, they need better navigation, dexterity, and cognition in them. “Today a robot has multiple cameras and technologies like Embedded Vision that are helping it navigate better,” said Mr Walter Zulauf, Vice-chairman at the European Robotics Industry Association. “In the future robots will get even better vision systems and will need even less human interventions. Laser vision technologies enable robots to work even in the dark. The robots will have better, lighter and more accurate sensors providing better dexterity.” Technologies like Visual Simultaneous Localisation and Mapping (VSLAM) have come up with algorithms that can help robots map the surrounding environment and locate themselves within it. It can navigate itself in a highly dynamic environment using multiple vision sensors. If robots and humans have to work together, a smooth interaction is essential for a better collaborative effort. The progress made in Natural Language Processing over the past decade has made better cognition possible in robots. “Apart from the obvious technical advantage there are many non-technical factors that have also worked in favour of service robots,” said Mr Zulauf, “For instance, the younger generation finds it easier to adopt the technology. Hence, they can easily accept and work together with robots. “Secondly, as fewer students choose science, technology, engineering, and mathematics graduations, the technical skill level across industries is reducing. Hence smart and AI-driven service robots can fill this gap and create capacity in the industry. “Finally, robots are always going to be cheaper and economically viable than labour, especially in countries with higher living standards.” Wrapping Up While service robots see a lot of investment, the high-performance robots that work behind the fences will still remain relevant, according to Mr Zulauf. This is because some of the robots are so fast and highly productive that they will need to function in safe environments. But with the new advancements in technologies, there will be a next generation of service robots with many sophisticated applications.","excerpt":"Since the advent of the robotics industry, robots have mostly been used for automating industrial processes and have operated on production shop floors in cages.  This trend is drastically changing now as robots are currently working in every-day settings alongside humans and having human-like features. With the rapid development in emerging technologies, robots with applications […]","categories":["AI Features"],"tags":["robotics industry"],"author_name":"Kashyap Raibagi","publish_date":"2021-01-08T12:00:00","publication_year":"2021","word_count":749,"keywords":["Go","API","funding","programming_languages:R","AI","venture capital","programming_languages:Go","ViT","robotics industry","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","R","Go","API","ViT","venture capital","funding","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/will-more-robots-work-out-of-cages\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":30568,"title":"IIT Delhi Establishes Chair In AI After Noted Alumnus &#038; Global Innovation Index Architect Soumitra Dutta","content":"As a generous gesture towards their alma mater IIT Delhi, noted academicians Professor Soumitra Dutta and Dr Lourdes Casanova signed a memorandum of understanding for constituting “The Soumitra Dutta Chair in Artificial Intelligence”. The two alumni have also pledged a support of ₹1 crore for the purpose. An official statement from the Indian Institute of Technology, Delhi, said that the Chair is envisaged to promote excellence and leadership in teaching, research and development in the field of artificial intelligence. It will also facilitate wider and deeper interaction between the industry, IIT-D faculty, students, and donors. Speaking about the MoU, Prof Dutta said, “I am very grateful to the faculty of IIT Delhi for the valuable education they provided to me and I am glad to contribute to the future academic excellence of IIT Delhi.” He is the Founding Dean of Cornell SC Johnson College of Business at Cornell University, New York. Prof Dutta is an authority on the impact of new technology on the business world with a special focus on strategies for driving growth and innovation in the digital economy. He is best known for being the architect (and main editor) of the Global Innovation Index, which now has become the global standard among innovation indices (co-published with Intellectual Property Organization-WIPO) and is referenced by many global leaders including Prime Minister Narendra Modi. Dr Casanova said, “The Chair recognises the important role of Artificial Intelligence in shaping the future of India and supports the aspirations of IIT Delhi for leadership in this important domain.” Dr Casanova is the Director of Emerging Markets Institute at Cornell University. Prof V Ramgopal Rao, Director, IIT Delhi said, “IIT Delhi has an excellent relationship with its alumni and values this relationship. Institute is thankful to our distinguished alumnus Prof Soumitra Dutta and Dr Lourdes Casanova for instituting this chair position at IIT Delhi”.","excerpt":"As a generous gesture towards their alma mater IIT Delhi, noted academicians Professor Soumitra Dutta and Dr Lourdes Casanova signed a memorandum of understanding for constituting “The Soumitra Dutta Chair in Artificial Intelligence”. The two alumni have also pledged a support of ₹1 crore for the purpose. An official statement from the Indian Institute of […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","IIT Delhi"],"author_name":"Prajakta Hebbar","publish_date":"2018-11-22T13:09:27","publication_year":"2018","word_count":310,"keywords":["Go","API","artificial intelligence","programming_languages:R","AI","IPO","innovation","Git","GAN","IIT Delhi","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","API","GAN","innovation","IPO","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-delhi-establishes-chair-in-ai-after-noted-alumnus-global-innovation-index-architect-soumitra-dutta\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":57172,"title":"Western Digital Plans To Expand Unit For Leveraging Indian Talents","content":"Western Digital, a computer hard disk drive manufacturer and data storage company, is looking at leveraging Indian talent to enhance its data analytics capabilities globally. While the US-headquartered firm already has 2,700 employees in India, mostly for research and development, it is planning to expand its strength by hiring more people with data analytics skills, said Steve Phillpott, chief information officer of Western Digital. He said, “India is probably one of the bigger and growing sites, as compared to other locations globally. We have a massive footprint here, and our India centre works on research areas like data analytics, robotic process automation and engineering services.” Globally, Western Digital employs around 15,000 engineers to drive innovation in its product and services sides of businesses. After the acquisition of Hitachi’s storage unit HGST in 2015 and flash memory products manufacturer SanDisk in 2016, Western Digital had gone through a lot of transition before integrating such large companies with itself. Currently, the company sees a lot of opportunity in data storage space, owing to the implementation of data localisation laws in various countries. “India is passing these new norms on data privacy and data security. Europe has done it through GDPR. In the US, individual states are also trying to do so. So, I think data needs to be maintained in a localised or nationalised format to adhere to the privacy and residency norms,” Phillpott said, adding that these offer a huge opportunity for companies like Western Digital. Western Digital has manufacturing sites spread across the globe, including the US, China, The Philippines, Thailand, etc.","excerpt":"Western Digital, a computer hard disk drive manufacturer and data storage company, is looking at leveraging Indian talent to enhance its data analytics capabilities globally. While the US-headquartered firm already has 2,700 employees in India, mostly for research and development, it is planning to expand its strength by hiring more people with data analytics skills, […]","categories":["AI News"],"tags":["Data Analytics","Data analytics India","data analytics jobs"],"author_name":"Sejuti Das","publish_date":"2020-02-21T16:30:00","publication_year":"2020","word_count":262,"keywords":["Go","data analytics jobs","AI","AWS","innovation","cloud_platforms:AWS","Git","RAG","automation","Data analytics India","analytics","Data Analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","AWS","R","Go","Git","automation","innovation","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/western-digital-plans-to-expand-unit-for-leveraging-indian-talents\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":57356,"title":"How A Recent Workshop Of NITI Aayog Gave Boost To India&#8217;s AI Ambitions","content":"NITI Aayog’s AI workshop called Artificial Intelligence – The India Imperative took place recently in New Delhi. The AI workshop- The India Imperative took place on 19th Feb ’20 at NITI Aayog Bhawan where all the relevant stakeholders from across the country were present. From ministers from states to representatives from the IT industry to professors from IITs were part of the workshop. The event highlighted India’s leading think tank continuous work in ensuring continuous activity in the AI field. NITI Aayog has also been publishing Approach Papers to create execution plans and showcase essential suggestions along with stakeholders. The CEO of NITI Aayog started the workshop with India’s AI aspirations and the importance of AI for all the sectors. He also recommended AI Superpower by Kai-Fu. He also added that India is expected to reach $15 trillion which will be more than the US and China together. In his keynote address, CEO Amitabh Kant kickstarted the deliberations by emphasising on the importance of AI for All in realising India’s artificial intelligence aspirations. The main 5 sectors that the workshop focused on were- healthcare, agriculture, education, infrastructure and transportation that would benefit the most from AI. Arnab Kumar, the Programme Director gave a presentation on AI For All and discussed the 4 fundamental themes: 1. Data Rich to Data Intelligent 2. Research & Development 3. AI-specific Computing 4. Large scale AI adoption The inaugural session was followed by the breakout sessions which included the following topics- structured data infrastructure for AI, research ecosystem for AI, Moonshots for India and Adoption- Focus on healthcare, education and agriculture. In the breakout session at the workshop on Artificial Intelligence – The India Imperative, scalable approach to building solutions for a billion citizens, by leveraging technologies like AI\/ML was discussed by the participants. Startup & State Governments Presented Papers At AI – The India Imperative Workshop In 2018 – 2019, the government-mandated NITI Aayog to create the National Program on AI, with the aim of guiding the research and development in new innovation in artificial intelligence for India.  NITI Aayog came out with the National Strategy for Artificial Intelligence (NSAI) discourse paper in June of 2018 to highlight Indian government’s importance and role in boosting AI. NITI Aayog has taken a three-part approach here – undertaking exploratory proof-of-concept AI projects in different areas of the country, creating a national strategy for a vibrant AI ecosystem in India and collaborating with experts and stakeholders in the field. The recent workshop was a part of the think tank’s engagement with stakeholders, including multiple startups. One such startup had been Silversparro, an AI-powered video analytics firm invited by Niti Aayog for a day-long session on realising India’s AI aspirations. The startup presented its views on how AI can help India leapfrog in sectors like manufacturing, heavy industry and giving a boost to SMEs. “We are heartened by Niti Aayog’s focus on making India an AI Superpower. We are also proud to be contributing directly by leveraging AI for making Indian manufacturing more productive with our latest offering – Sparrosense AI Supervisor ” said Abhinav Kumar Gupta, Founder & CEO at Silversparro. At the workshop, there were several speakers talking about leadership and vision in the AI workshop “Artificial Intelligence – The India Imperative” event in Delhi by NITI Aayog, including those from state governments such as Telangana. In fact, Telangana presented  “Year of AI” at NITI Aayog’s workshop. Also Read: India Lags Behind In AI Research, But Will ₹7,000 Crore Boost Change Things? Workshop In Sync With India’s National Data and Analytics Platform (NDAP) India has rich publicly available data, and across government departments, the various processes have been digitised for reporting and analytics insights, which are feeding into information systems and visualisation dashboards. According to NITI Aayog, this data is being utilised to track and visualise processes and make iterative enhancements. National Data and Analytics Platform (NDAP), an initiative aimed to aid India’s progress by promoting data-driven discourse and decision-making, NDAP also aimed to standardise data across multiple government sources, provide flexible analytics and make it easily accessible in formats conducive for research, innovation, policy-making and public consumption. As part of it, multiple data sets have been presented using a standardised schema, by using common geographical and temporal identifiers. However, the data landscape can be further improved as the entire public government data can be smoothly accessible to all stakeholders in a user-friendly manner. Further, data across different government assets should be interlinked to enable analytics and insights, such websites of ministries and departments of the central and state governments. Also Learn: AIRAWAT: NITI Aayog Describes How India’s AI Infrastructure Will Look Like","excerpt":"NITI Aayog’s AI workshop called Artificial Intelligence – The India Imperative took place recently in New Delhi. The AI workshop- The India Imperative took place on 19th Feb ’20 at NITI Aayog Bhawan where all the relevant stakeholders from across the country were present. From ministers from states to representatives from the IT industry to […]","categories":["AI Features"],"tags":["how ai works","NITI Aayog","niti aayog ai","NITI aayog india"],"author_name":"Vishal Chawla","publish_date":"2020-02-24T14:00:00","publication_year":"2020","word_count":777,"keywords":["how ai works","Go","artificial intelligence","AI","ML","Scala","Git","RAG","Aim","niti aayog ai","analytics","R","NITI aayog india","NITI Aayog"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","Aim","RAG","R","Go","Scala","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/niti-aayog-india-ai-workshop\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10054684,"title":"Best ML Papers Presented At Major Conferences Of 2021","content":"The year is about to end, and all the major AI and ML conferences of 2021 took place online. The conferences set a high standard and were quite promising when it came to the presentation of major papers from researchers across the globe. We have covered all the major conferences throughout the year, and picking up the best is no cakewalk. Some of the papers are well-known to the domain experts, while some are considered as more of hidden gems. Here we present some of the best papers presented at major conferences of 2021: 1| Unbiased Gradient Estimation in Unrolled Computation Graphs with Persistent Evolution Strategies at ICML 2021 By: The research paper by Paul Vicol, Jascha Sohl-Dickstein, and Luke Metz from Google Brain and the University of Toronto has won the outstanding paper award. About: The paper introduced a Persistent Evolution Strategies (PES) method for unbiased gradient estimation in untolled computation graphs. PES allows for rapid parameter updates. In addition, it has low memory usage, is unbiased, and has reasonable variance characteristics. Researchers experimentally demonstrate the advantages of PES compared to several other methods for gradient estimation on synthetic tasks and show its applicability to training learned optimisers and tuning hyperparameters. 2| Oops I Took A Gradient: Scalable Sampling for Discrete Distributions at ICML 2021 By: The paper by researchers from Google Brain, including Will Grathwohl, Milad Hashemi, Kevin Swersky, David Duvenaud and Chris J. Maddison, has received the outstanding paper honourable mention at the conference. About: Researchers proposed a general and scalable approximate sampling strategy for probabilistic models with discrete variables. The newly introduced approach uses gradients of the likelihood function wrt its discrete inputs to propose updates in a MetropolisHastings sampler. 3| OpenGAN: Open-Set Recognition via Open Data Generation at ICCV 2021 By: Researchers including Shu Kong and Deva Ramanan from Carnegie Mellon University developed OpenGAN for open-set recognition. Both the researchers incorporated two technical insights: They trained a classifier on OTS characteristics rather than pixels, andAlso, they focused on adversarially synthesising fake open data to increase the open-training data pool. About: With OpenGAN, the team shows that using GAN-discriminator achieves the state-of-the-art on open-set discrimination, once selected using a val-set of real outlier examples. This is effective even when the outlier validation examples are sparsely sampled or strongly biased. OpenGAN significantly outperforms prior art on both open-set image recognition and semantic segmentation. 4| Mip-NeRF: A Multiscale Representation for Anti-Aliasing Neural Radiance Fields By: The paper from researchers including Jonathan T. Barron, Peter Hedman, Pratul P. Srinivasan, Ricardo Martin-Brualla, Ben Mildenhall from Google and Matthew Tancik from UC Berkeley introduced a solution which they call “mip-NeRF.” About: The team presented mip-NeRF — a multiscale neural radiance fields (NeRF) like model that addresses the inherent aliasing of NeRF. NeRF works by casting rays, encoding the positions of points along with those rays, and training separate neural networks at distinct scales. In contrast, mip-NeRF casts cones that encode the positions and sizes of conical frustums and train a single neural network that models the scene at multiple scales. 5| Visually Grounded Reasoning across Languages and Cultures at EMNLP 2021 By: Researchers from the University of Cambridge, University of Copenhagen, McGill University and Mila – Quebec Artificial Intelligence Institute presented the paper and bagged the best long paper award. About: The team created a multilingual dataset for Multicultural Reasoning over Vision and Language (MaRVL) by eliciting statements from native speaker annotators about pairs of images. The task consists of discriminating whether each grounded statement is true or false. Researchers establish a series of baselines using state-of-the-art models and find that their cross-lingual transfer performance lags dramatically behind supervised performance in English. 6| CHoRaL: Collecting Humor Reaction Labels from Millions of Social Media Users at EMNLP 2021 By: The paper presented by Zixiaofan Yang, Shayan Hooshmand, and Julia Hirschberg from the Department of Computer Science at Columbia University won the best short paper award. About: CHoRaL is a framework to generate perceived humour labels on Facebook posts, using the naturally available user reactions to these posts with no manual annotation needed. It provides both binary labels and continuous scores of humour and non-humour. The team presented the largest dataset to date with labelled humour on 785K posts related to COVID-19. CHoRaL enables the development of large-scale humour detection models on any topic and opens a new path to the study of humour on social media. 7| Meta Pseudo Labels at CVPR 2021 By: Researchers from the Google AI Brain team, including Hieu Pham, Qizhe Xie, Zihang Dai, Minh-Thang Luong and Quoc V. Le, introduced this semi-supervised learning technique. About: The model presented in the paper has achieved a new state-of-the-art top-1 accuracy of about 90.2% on ImageNet. The result is 1.6 per cent better than the existing SOTA models. The key to the model is the idea that the teacher learns from the student’s feedback in order to generate pseudo labels that best help students’ learning. The learning process in Meta Pseudo Labels consists of two main updates: updating the student based on the pseudo labelled data produced by the teacher and updating the teacher based on the student’s performance. 8| RRL: Resnet as representation for Reinforcement Learning By: The paper presented by researchers Rutav Shah and Vikash Kumar from the Indian Institute of Technology (IIT), Kharagpur, was produced in collaboration with the University of Washington and Facebook AI.About: The team proposed a straightforward and effective approach that is capable of learning complex behaviours directly from proprioceptive inputs. RRL fuses the features extracted from pre-trained Resnet and put it into the standard RL pipeline, and delivers results comparable to learning directly from the state.","excerpt":"The year is about to end, and all the major AI and ML conferences of 2021 took place online. The conferences set a high standard and were quite promising when it came to the presentation of major papers from researchers across the globe. We have covered all the major conferences throughout the year, and picking […]","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Machine Learning"],"author_name":"kumar Gandharv","publish_date":"2021-12-02T11:00:00","publication_year":"2021","word_count":943,"keywords":["Go","artificial intelligence","AI","neural network","ML","image recognition","Machine Learning","RAG","NLP","Ray","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","NLP","Ray","RAG","image recognition","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/best-ml-papers-presented-at-major-conferences-of-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":43250,"title":"5 Challenges Of Adopting Analytics In Startups Vs Larger Organisations","content":"Analytics is crucial for every organisation today — small or big. Adopting analytics has a direct impact on the success of companies ranging from improved decision-making to better use of the extensive data generated. While bigger organisations find themselves in a better position to adopt analytics, the same cannot be said about startups. However, there are challenges that both larger organisations and startups face when it comes to adopting analytics at some level or the other. In this article, we will discuss five such points that make it challenging to adopt analytics in startups vs. in larger organisations. Availability Of Resources Building an analytics function is not an easy task. It creates a considerable amount of financial strain on an organisation in terms of investment — setting up the right infrastructure, getting the right tools and hiring the right talent for the analytics team. While it is easier for larger organisations to readily deploy resources, startups may it find it challenging to get all the right resources at the right time. The capital allocation and funding for these two varies significantly which ultimately shows up in slower adoption of analytics tools and models. Cost Of Hiring The Right Talent The cost of hiring the right talent also goes up significantly if the companies are looking to hire the best fit. Since startups may have limited funds to get the right talent, they often tend to prioritise hiring for profit-generating like in operations, marketing and sales, rather than data scientists. Whereas hiring data analysts or scientists is easier for larger enterprises. A lack of good talent ultimately slows down the analytics adoption in startups compared to larger enterprises. Also while it is easy for larger organisations to attract talent, startups may find it challenging at some level. Quality Of Data And Data Collection Startups may struggle with accumulating the right kind of data and that which is also good in terms of quality. Even if they have large chunks of data, they may face problems while interpreting and making sense of data they have in their organisation. While modern technology makes it easy for startups to capture data to some extent, just posing data doesn’t mean that they can use it to improve business, which may not be the case with larger organisations. They have large resources of data and a separate team that deals with cleaning and de-cluttering it making it much easier for them. Startups, on the other hand, face challenges such as collecting qualitative data about their customers, converting qualitative data into quantitative data, collecting accurate data, and if the data collected is clean and standardised. Agility The one point where startups have the upper hand than larger organisations is agility. Startups have the power to move quickly and easily to adopt newer initiatives in the organisation. In larger companies, the organisational structure and hierarchies make it harder for them to manoeuvre any changes. Much time is spent in democratising the ides and consensus-building. Agility comes as a major challenge in the larger organisation as opposed to smaller firms where it is much easy to bring about changes regarding analytics adoption or incubate newer initiatives. Few decision-makers make it easy for startups than larger organisations. Access To The Ecosystem It goes without saying that large and well-established organisations have the calling card and access to the ecosystem players. Access to the large and small; local and global player makes it easy for larger organisations to collaborate with them and form partnerships to work towards a larger goal. They will easy access to analytics ecosystem compared to startups making it easier to deploy analytics solutions.","excerpt":"Analytics is crucial for every organisation today — small or big. Adopting analytics has a direct impact on the success of companies ranging from improved decision-making to better use of the extensive data generated. While bigger organisations find themselves in a better position to adopt analytics, the same cannot be said about startups. However, there […]","categories":["AI Startups"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2019-07-24T14:00:29","publication_year":"2019","word_count":606,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","analytics","GAN","R","startup"],"extracted_tech_keywords":["AI","analytics","R","Go","API","GAN","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/5-challenges-of-adopting-analytics-in-startups-vs-larger-organisations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":28810,"title":"This Deep Learning Framework Can Point Out Wrong Sitting Posture When Using A Computer","content":"Poor posture has been a leading cause of many ailments. Of late, technology has started playing a key role in alleviating common posture problems. That’s the reason why continuous research has been carried out around detecting postures in humans as well as robotic systems. Posture recognition systems, as some of these are called, are interestingly being the focal point of research for computer scientists and researchers. In this article, we discuss a novel research study that uses deep learning to build a scene recognition system to detect sitting postures. Deep Learning For Object Detection A group of researchers from Nanchang University, China have developed a method through which they track unhealthy sitting posture in humans by analysing scenes and objects in the scene. Using a motion sensing device known as Microsoft Kinect, skeletal points of the human body are detected and observed for movements. Now, a deep learning (DL) technique is used to detect objects in a scene( a workplace\/office) and also to extract features from these objects. The skeletal points and object features from the scene are clubbed to form semantic features through a Gaussian Mixture model. These semantic features would then depict sitting postures. In other words, the study has two parts — a scene recognition and semantic analysis. Posture detection in a nutshell. Scene Recognition and Semantic Analysis are the focus areas in the study (Image courtesy: Weidong Min and team, Nanchang University, China) In scene recognition, a neural network called Faster Region-based Convolutional Neural Network (R-CNN) is used. This CNN is an improvement of Fast R-CNN, a superior version of CNNs, in terms of features. Faster R-CNN’s advantage lies in its striking feature called Region Proposal Network (RPN). These networks check the possibility of objects in the scene by inferring efficiently from the regions. (A detailed information about Faster R-CNN can be found here.) Similarly, skeletal point extraction is done by inferring space and location information in the scene. But, here space is restricted to human body features. In the semantic analysis, behavioural semantics is the basis of feature extraction. This means, all the behavioural attributes are considered along with scene entities. To generate behavioural semantics, the Gaussian-Mixture clustering method is used. This itself forms a separate module in semantic analysis. Results And Impact On Sitting Postures When the experiment was conducted for a set of 65 videos of postures deemed unhealthy, the CNN framework detected postures along nine positions and almost fared the same compared to another standard method used for checking unhealthy sitting postures. In fact, accuracies (of precision and recall) were more than 95 percent, which was, again better than the compared method. Here, the CNN fared well for a pre-set environment, which means, a dataset of videos was used for experimentation. But, it has to be explored in a real-time setting. Real-time environment entails much more information other than those presented to the CNN. It has to be seen how it performs in this scenario. Earlier studies on posture detection used equipment such as sensors to assess body movements linkable to postures. Although they track key points necessary here, these studies neglect objects in the surrounding area. But, the above study tracks almost all elements in its vicinity. Why is this important? It is because certain objects also affect postures to a significant extent. For example, computer screens — if they are at the wrong angle to the worker, he\/she may sit in a different posture other than the normal upright posture. CNN detecting these irregularities not just helps with respect to sitting postures but also helps with determining overall ergonomics. If frameworks like these are implemented successfully in workplaces, it can be said that productivity and workplace safety can largely be maximised. In Conclusion As bad sitting postures and incorrect ergonomics become a hot topic in research, studies like these hope to eliminate inadequacies in the work environment, if left ignored lead to medical complications in employees. Modern rules of work have changed a lot. By incorporating techniques of deep learning for behaviour and physical activity supervision, the ‘healthy’ way of doing work can be promoted so as to boost morale.","excerpt":"Poor posture has been a leading cause of many ailments. Of late, technology has started playing a key role in alleviating common posture problems. That’s the reason why continuous research has been carried out around detecting postures in humans as well as robotic systems. Posture recognition systems, as some of these are called, are interestingly […]","categories":["AI Features"],"tags":["Machine Learning"],"author_name":"Abhishek Sharma","publish_date":"2018-10-01T10:43:07","publication_year":"2018","word_count":690,"keywords":["Go","programming_languages:R","AI","neural network","R-CNN","Machine Learning","object detection","deep learning","ViT","CNN","R"],"extracted_tech_keywords":["AI","deep learning","neural network","object detection","R","Go","CNN","ViT","R-CNN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-deep-learning-framework-can-point-out-wrong-sitting-posture-when-using-a-computer\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10081172,"title":"Laid-off While on H-1B Visa? Here is What You Can Do","content":"“I joined Twitter with absolute excitement in September, 2022. Barely two months later, here I am – part of the 3,500 people laid off at Twitter, on an H-1B visa, with a few months to sort things out, or pack up, and go back,” reads Balaji Bhasker’s LinkedIn post. Like Balaji, the future of many foreign nationals who entered the US on H-1B visa, hangs in the balance as major tech companies continue to lay off a large chunk of their workforce. Over 60,000 Indians working in the US have lost their jobs this year, most of them on H-1B visas. Meta alone laid off around 11,000 of its employees. H-1B visas allow foreign nationals with specialised skills to enter the US and take up a job there. Predominantly issued to individuals in the tech industry, the visas are linked to the specific employer that sponsors them. H-1B holders are only legally allowed to be in the US for 60 days without being paid. In such a scenario, what are the options they have? Use your grace period wisely After being laid off on an H-1B visa, the individuals have a grace period of 60 days. One must ensure that they utilise the period wisely. Aishwarya Srinivasan, data scientist at Google Cloud AI, said, “Confirm what your last day on the payroll is and confirm what your new employment deadline for H-1B is. According to most lawyers, your 60-day grace period starts exactly after you’re notified of your layoff; however, it’s a grey area. Some cases are being made for your last payday, being the start pay of the countdown. Either way, talk to the immigration lawyer at your company and get clarity on what the deadline is.” Once you know the deadline, start hunting for jobs. However, it’s key that you don’t apply blindly, because it will only increase your rejection rate. Srinivasan believes it is one of the worst mistakes you can commit. “Instead, use that time wisely. I know it might seem counterproductive; however, my opinion is to apply in a targeted fashion rather than blindly for any available roles,” she added. If you find a job within the 60-day grace period, you can apply for a fresh H-1B visa and continue to work during the grace period. Explore other Visa options Depending on the time of your layoff, one option you can explore is enrolling in an academic programme. If you manage to enroll in a programme, apply to change your visa from H-1B to F-1( student visa). A student visa will allow you to acquire new skills and, at the same time, keep you in the country. If a student visa is not an option, you can apply for a B-2 (tourist visa). This is the option most of them explore in case of a layoff. However, even though a B-2 visa will allow you to stay in the country a bit longer, it could be problematic because a B-2 visa is meant for tourists only. Also, upon expiry of the visa period, most certainly, you will have to make a trip back home. However, the processing time for a change in the status of your Visa from H-1B to B-2 is quite long, almost a year in some cases. Third, if you have a partner who is on an H-1B and is currently employed, you can apply to change your visa status from H-1B to H-4 dependent. Once you have an H-4 visa, it means you can stay in the country as long as your partner does. This will buy you ample time to find and apply for jobs best suited for you. Other immigration options When staying in the US is no longer an option, you can always move back to your home country. You can continue to look for jobs in the US from your home country. Srinivasan says, “You can also look at options like moving to Canada or India. Although this is probably the most dreaded option, there are many ways to return to the US.” Canada is welcoming foreign nationals in the tech and IT sector to work in the country under its Temporary Foreign Worker Programme. In fact, the Canadian Global Talent Stream was set up to facilitate the growth of Canada’s tech industry and process applications within two weeks from filing. Meanwhile, applying for a job in India might not seem desirable, but the tech landscape in India is growing significantly, and these are interesting times. “With all the 2022 tech layoffs in the US, please spread the word to remind Indians to come back home (especially those with visa issues) to help Indian tech realise its hyper-growth potential in the next decade,” Harsh Jain, co-founder, and CEO of Dream11, said.","excerpt":"H-1B holders are legally allowed to be in the US for just 60 days without a salary. In such a scenario, what are the options they have?","categories":["AI Features"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2022-11-30T18:00:00","publication_year":"2022","word_count":794,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","cloud_platforms:Google Cloud","R"],"extracted_tech_keywords":["AI","R","Go","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/laid-off-while-on-h1-b-here-is-what-you-can-do\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":47004,"title":"How Good Are Deep Learning Networks At Picking Up Visual Cues","content":"The ability to generalise have made neural networks learn tasks at a faster rate.Generalising in this context means to classify data from the same class as the learning data that it has never seen before. In other words, this can be likened to transfer learning. A model or an agent in a reinforcement learning environment is trained on one task and uses this knowledge to perform a task that is new. Making the networks to learn new strategies to generalise better is usually the aim behind any algorithmic enhancement. But how good are these networks at generalisation immediately after training? Exploring this question will lead one to pressing challenges of training deep learning algorithms such as the need for long hours of training, the efficacy of the model and other such questions. To demonstrate how the network generalises, researchers from Stanford University in collaboration with  DeepMind and University College of London, conduct experiments on agents in a simulation of 3D world. The agents are monitored for how they learn actions that contain verbs. For example, an agent can be asked to “lift” an object or “find” a particular thing within that environment. Overview Of The Process via paper by Stanford et al., The above figure is an illustration of the architecture used in all experiments. The  simplicity of the architecture is intended to emphasize the generality of the findings. The agent is assessed through tasks that require visual, language and memory related skills. The whole process can be summarised in three steps: The visual observations are passed on to a convolutional neural network. The output of this network is joined with an embedding of the language observation. Language instructions are received at every timestep as a string.  The agent splits these with a (word-level) LSTM network. The final hidden state is concatenated with the output of the visual processor to yield a representation at each timestep. The  multimodal  representation  is passed to a 128-unit LSTM. At each timestep, the state of this LSTM is multiplied by a weight matrix containing and then the output is passed through a softmax function. The agent receives a positive reward if it finds or lifts the correct object, and the episode ends with no reward if the agent finds or lifts the incorrect object. Systematic Generalization In Agents In their experiments, the researchers consider addressing the following three factors: the number of words and objects experienced during training; a first-person egocentric perspective; and the diversity of perceptual input afforded by the perspective of a first-person interactive agent over time. Executing a simple instruction like ‘find a toothbrush’ (which can be accomplished on average in six actions by a well-trained agent in our corresponding grid world) requires an average of 20 action decisions. In this environment, the agent observes the world from a first-person perspective, the Unity objects are 3D renderings of everyday objects, the environment has simulated physics enabling objects to be picked-up, moved and stacked, and the agent’s action-space consists of 26 actions that allow the agent to move its location, its field of vision, and to grip, lift, lower and manipulate objects. The results showed that the agent therefore learned a notion of what it is to lift an object (and how this binds to the word lift) with sufficient generality that it can, without further training, apply it to novel objects, or familiar objects with novel modes of linguistic reference. The findings of this work can be summarised as follows: Neural-network-based agent with standard architectural components can learn to execute goal-directed motion in response to instructions. The first-person perspective of an agent acting over time plays an important role in the emergence of this generalization. Language can provide a form of supervision for how to break down the world and\/or learned behaviours into meaningful sub-parts, which in turn might stimulate systematicity and generalisation. Agents trained in 3D worlds generalize better. in  the 3D world, the agent experiences a much richer variety of  (highly correlated) visual stimuli in any particular episode. Outlook The significance of these experiments can be better understood when put in the context of artificial general intelligence. A robot listening to the voice based instructions and performing actions such as moving and finding requires a symbiosis of most efficient natural language and reinforcement learning models. Though the results of these experiments only cater to simpler environments, we can safely assume that in due course, with improved models and unlimited resources, systematic generalization might emerge more readily than ever before.","excerpt":"The ability to generalise have made neural networks learn tasks at a faster rate.Generalising in this context means to classify data from the same class as the learning data that it has never seen before. In other words, this can be likened to transfer learning. A model or an agent in a reinforcement learning environment […]","categories":["Deep Tech"],"tags":["deepmind london","simple neural network"],"author_name":"Ram Sagar","publish_date":"2019-10-08T12:00:08","publication_year":"2019","word_count":750,"keywords":["Go","TPU","AI","neural network","Modal","RAG","Aim","deep learning","LSTM","R","deepmind london","simple neural network"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","RAG","TPU","R","Go","LSTM","Modal"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-good-are-deep-learning-networks-at-picking-up-visual-cues\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10114640,"title":"Top 11 Data Centre Projects in India 2024","content":"In recent months, several companies, including AdaniConnex, Reliance, Sify, Atlassian, Yotta, and AWS, have announced substantial investments in data centres across India. AWS alone plans to invest $12.7 billion to expand its data centres in the country. In addition, Kotak Alternate Assets intends to invest $800 million to support the development of 5-7 large data centres in key property markets in India. The surge in data centre investments is driven by the demand for data localisation and cost efficiency, influenced by India’s data protection norms and proposed data centre policy, attracting major global players. With the growth of India’s data centre market, the capacity is projected to surpass ~1,300 MW by the end of 2024, a notable increase from ~1,048 MW at the end of 2023 and ~880 MW as of June 2023. The concentration of data centres remains prominent in seven cities, including Mumbai-Navi Mumbai, Chennai, Delhi-NCR, Bengaluru, Pune, Hyderabad, and Kolkata. These cities accounted for a 35% year-on-year growth, reaching ~884 MW capacity and spanning over 13 million sqft by the end of 2023. Mumbai-Navi Mumbai leads with a 52% share of the total data centre capacity, followed by Chennai (16%), Delhi-NCR (11%), Bengaluru (9%), Pune (7%), Hyderabad (4%), and Kolkata (1%). Despite the concentration in the top four locations, the emergence of co-location and edge computing facilities is expected to shift the dynamics, with edge data centres expanding into Tier 2 cities in 2024. Overall, data centre occupancy levels in India stood at about 75-80% in 2023, likely to see further improvement by the end of 2024. Top Data Centre Projects in 2024YottaCtrlS DatacentersDigital ConnexionSifySTT GDC IndiaAtlassian CapitaLand IndiaEquinixAdaniConnexGoogleAmazon Web Services Here are some upcoming noteworthy data centre projects in 2024 and beyond to watch out for: Yotta Hiranandani Group-backed Yotta Data Services is expanding its operations in Greater Noida and Guwahati, addressing the growing demand for edge facilities in Tier 2 markets. These planned operations are slated to be completed and operationalised by the end of 2024. Yotta-D1, the company’s existing data centre, is 85% full, prompting plans for construction on its D2 and D3 facilities. After placing an order of over 16,000 NVIDIA H100 GPUs, Yotta has also announced a collaboration with NVIDIA to build an AI data centre in GIFT City, Gujarat, to meet the demand for AI-driven data. CtrlS Datacenters CtrlS, a Hyderabad-based pioneer of Tier 4 data centres in India, operates 15 data centres across eight key markets and is set to build over 600 MW of data centre capacity by 2029. This expansion involves an additional 5 million sq ft, making CtrlS a major player in the global Rated-4 data centre landscape. With projects underway in Navi Mumbai, Hyderabad, and Chennai, CtrlS is also venturing into Tier 2 cities, investing Rs 250 crore in a greenfield Edge data centre in Uttarakhand. The proposed data centre will support co-location, managed, and cloud services. A few days ago, CtrlS Datacenters also announced the commencement of the construction of a new data centre campus in Chennai, with a significant investment of Rs 4000 crore ($482.5 million). The 72MW project in the Ambattur industrial area spans 1 million sq ft across two buildings and includes an on-site substation. The first building, Chennai DC 1, is fully booked and scheduled to begin operations in Q2 2024. As part of its expansion plans, CtrlS announced a $2 billion investment in October 2023 to add 350MW of AI and cloud-ready hyperscale data centres across Asia and the Middle East. Digital Connexion Digital Connexion, a joint venture involving Reliance Industries, Brookfield Infrastructure Partners, and Digital Realty Trust, launched its flagship 20 MW greenfield data centre, MAA10, in Chennai in January 2024. This marks the beginning of a potential 100 MW campus. The initial phase, MAA10, provides 20MW of IT load with modular infrastructure design for scalable response to varied workload demands. The venture expanded its presence by acquiring 2.15 acres of land in Mumbai to construct a 40 MW data centre. Located in a vital industrial hub, the facility supports emerging technologies like AI and large language models and offers ultrahigh-power densities. Sify Chennai-based Sify has raised funds for new data centres in 2024 as part of its overall investment of Rs 9,000 crore in the next 5-6 years on greenfield data centre projects. Teaming up with Kotak Data Centre Fund, the company is investing up to $73 million in its subsidiary, Sify Infiniti Spaces, which will operate these planned data centres. With projects in Chennai, Mumbai, Noida, and Bengaluru set to commence early in 2024, Sify aims to add 350 MW to its current data centre capacity of 100 MW through 11 facilities across India. STT GDC India Singapore-based ST Telemedia Global Data Centres (STT GDC India) prioritises sustainability and has partnered with O2 Power to procure renewable energy for its Bengaluru facility. In 2024, STT GDC India is set to invest around Rs 2,000 crore to develop two new data centres in its existing campus in Pune’s Dighi, surpassing 80 MW of IT load capacity. This makes STT Global Data Centres’ Pune campus one of the largest data centre campuses in India, with an existing capacity of 40 MW spanning three operational facilities. Atlassian The Sydney-headquartered software company Atlassian plans to invest in establishing data centres in India in the first half of 2024. The move aims to comply with data residency requirements, and Atlassian will collaborate with Amazon to bring its products to these Indian data centres. Co-founder and co-CEO Scott Farquhar expressed confidence in India’s growing economy and talent landscape, highlighting the belief in the market’s potential for growth. Atlassian, listed on NASDAQ, provides global team collaboration and productivity software with over 250,000 customers. In India, notable clients include Ola Cabs, Reliance, Walmart Labs, and Flipkart. Farquhar emphasised the resilience of India’s market amid global downturns and sees it as a growth market. Atlassian initiated its presence in India in 2018 with 60 employees and has since grown to 1,700 employees, making India its fastest-growing employee location. CapitaLand India CapitaLand India Trust (CLINT), previously known as Ascendas India Trust, has secured a loan of $155.9 million from J.P. Morgan India for its Navi Mumbai data centre. The Airoli campus in Navi Mumbai, spread across 6.6 acres, is set to comprise two buildings. The first building, spanning 325,000 sq ft, is scheduled to be operational by Q2 of 2024, with full build-up capacity reaching 575,000 sq ft and 90 MW. CLINT is also planning data centres in Ambattur, Chennai, and Hyderabad, targeting global technology companies and cloud service providers. Equinix The company has two new carrier and service-neutral data centres named MB1 and MB2 in Mumbai, hosting clients including Amazon Web Services, Google Cloud, and Oracle Cloud. With a $42 million investment in another new data centre, MB4 in Kalwa, set to launch before March 2024, Equinix continues strengthening its presence in India. Equinix – a global player operating 250 data centres worldwide, entered the Indian market by acquiring GPX Global Systems. In 2023, Equinix launched full-fledged services in India, offering advanced solutions such as Equinix Fabric, Equinix Internet Exchange, and Equinix Internet Access. AdaniConnex AdaniConnex, a joint venture between Adani Enterprises and EdgeConnex Inc, has made substantial investments totalling $1.5 billion. Currently in the process of securing an additional $400 million offshore loan, the venture aims to establish data centres in key locations such as Visakhapatnam, New Delhi, Mumbai, and Chennai. Details regarding the two data centres in Andhra Pradesh with an aggregate investment of ₹21,844 crore, located at Madhurawada and Kapuluppada near Visakhapatnam were released during May in 2023. This initiative is part of a broader plan to construct nine data centres with a targeted total capacity of 1 GW by 2030. The JV has already secured a $213 million loan for the Chennai 1 campus and the 50 MW Noida campus in 2023. Google Google, with existing operations in two Indian GCP cloud regions in Mumbai and Delhi, is progressing with the development of an 8-storey, 381,000 sq ft data centre in Navi Mumbai. In partnership with Raiden Infotech, this project is expected to be completed by 2025 and will complement Google’s existing operations. Additionally, Google leased a 464,000-square-foot facility at the Adani Centre in Noida, showcasing its commitment to expanding its cloud infrastructure in India, one of its significant growth markets. Amazon Web Services AWS is set to establish four smaller data centres in India within the next two years, strategically located in Bangalore, Chennai, Delhi, and Kolkata. Additionally, AWS plans to launch 32 local zones across 26 countries in the same timeframe to enhance networking speed and security. The regional hubs will offer cloud services, catering to various use cases such as video streaming, gaming, and applications requiring real-time feedback. These regional zones will serve as the foundation for AWS regions, enabling the proximity of computing, storage, database, and other services to users and businesses for improved performance and efficiency.","excerpt":"Here are some upcoming noteworthy data centre projects in India in 2024 and beyond to watch out for","categories":["AI Trends"],"tags":["Data Center","Top Trend"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-02-28T15:20:56","publication_year":"2024","word_count":1489,"keywords":["Top Trend","Go","GCP","Data Center","AWS","AI","Scala","RAG","Aim","Rust","edge computing","R"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","GCP","edge computing","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-data-centre-projects-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10099252,"title":"OpenAI Finally Learns How to Do Business","content":"OpenAI, the research company that morphed into a tech startup, is finally understanding how to make it work. After incurring losses of over $540 million towards building ChatGPT, the company is now trying to make it work, and earn some revenue. In the most recent report from The Information, OpenAI believes that it is now on pace to generate revenue of $1 billion over the next 12 months. All of now, it would be done by selling their ChatGPT to enterprises, an offering long-awaited by companies given all the concerns about data privacy. There seems to be no other way left for the company. OpenAI has announced ChatGPT Enterprise, which will allow businesses to use the most-famous chatbot, with enterprise-level security and privacy, and unlimited high-speed access to GPT-4. ChatGPT Enterprise removed all usage caps, and now performs up to two times faster. It  includes 32k context in Enterprise, allowing users to process four times longer inputs or files, making it even better than ChatGPT Plus. Do the numbers add up? The billion-dollar revenue that the company is projecting is a little off from what it had said before. OpenAI projected an annual revenue of $200 million in 2023, and was expected to reach $1 billion in 2024. Not too far off, but a person with direct knowledge of the situation said that the expectation is far ahead. Microsoft is teaching OpenAI how to do business and earn some money. After all, the tech-giant has invested huge sums of money to turn the research organisation into a tech startup. If the predictions are right, then OpenAI would generate north of $80 million per month, which is too high when compared to just $28 million revenue it generated last year. OpenAI was a research lab, but Sam Altman turned it into a tech startup.— Pedro Domingos (@pmddomingos) August 28, 2023 Furthermore, though the revenue might flow, to turn into profitability, the company would have to account for the $700,000 it spends everyday to run ChatGPT. That is just for running inference, and does not include the company’s plan to make its models better and offer it for enterprise. There must be a lot of costs that OpenAI is possibly not taking into account. On the other hand, Microsoft has stopped depending on OpenAI for its services. It possibly realised that the ChatGPT company is not providing it with smaller models and also jeopardising its image with so many people being concerned over data privacy. Thus, Microsoft partnered with Meta to release Llama 2 on its platform. It silently also released Azure ChatGPT, something very similar to what OpenAI has released now. The GitHub repository also took a jab at OpenAI saying that people are scared of ChatGPT’s privacy issues. But just two days later, Microsoft took it down, hinting at some tussle between the two companies, since OpenAI was also planning to release exactly the same thing and possibly raised the red flag. Is OpenAI now waking up to Microsoft’s exploitation? In May, OpenAI had announced that it would be launching ChatGPT Business in the coming months, promising enterprises more control over their data. It looked like Microsoft wanted OpenAI to do the dirty work of discovering enterprise use cases, driving generative AI adoption, fixing safety flaws, and replicate the process for customers. But since then, Microsoft has been trying to push OpenAI’s APIs on its cloud for its services. Now, with ChatGPT Enterprise, OpenAI’s shift towards a profit-driven company is finally taking shape. OpenAI launching its own enterprise offerings while its biggest backer is doing so too would simply not align well in the future. Despite that, people were sceptical about using the platform. CEO Sam Altman recently posted on X to clarify that OpenAI does not use a company’s data to train its model, unless the company opts-in. seeing a lot of confusion about this, so for clarity:openai never trains on anything ever submitted to the api or uses that data to improve our models in any way.— Sam Altman (@sama) August 15, 2023 On the other hand, just like the GPT-4 multimodal announcement that never actually happened, OpenAI’s ChatGPT Enterprise announcement still is in its infancy. The company hasn’t revealed any pricing details and how it would finally be able to make money through the platform. Moreover, if companies are already able to leverage open source models with their data, would this new announcement actually bring them back to GPT? Maybe, OpenAI should have thought about releasing this long back and been a little quicker. This release might kill a lot of other startups that are coming up as just wrappers around ChatGPT, but to make money for OpenAI again, it might take a lot more convincing soon to bring back the customers. But for sure, the company has learnt how to do business. Would enterprises be willing to pay for the paid version of ChatGPT if there is already a free version available for everyone with data control features or will OpenAI pull the plug on ChatGPT entirely to make ChatGPT Enterprise work? And probably also cut ties with Microsoft along the way? Either way, ChatGPT is not getting any smarter.","excerpt":"OpenAI is turning into a tech startup, from being just a research company. Now, it is in the big leagues and in direct competition with Microsoft","categories":["Global Tech"],"tags":["Microsoft","OpenAI"],"author_name":"Mohit Pandey","publish_date":"2023-08-30T13:00:00","publication_year":"2023","word_count":864,"keywords":["Go","ChatGPT","OpenAI","AI","AWS","R","Git","RAG","generative AI","Azure","Microsoft"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","RAG","AWS","Azure","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-finally-learns-how-to-do-business\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10049278,"title":"Swati Jain","content":"Dr Swati is a seasoned professional with over two decades of analytics and consulting experience across multiple verticals, including financial services, retail, media, logistics, and healthcare. Currently, she is the Vice President, Decision Analytics at EXL Service. She leads a world-class team of data scientists and AI\/ML professionals impacting marketing, risk management, and operations of several Fortune 500 organisations. She has been instrumental in developing multiple award-winning innovative AI solutions (Gamechangers) along the customer journey, propagating newer business models and democratising AI in client organisations. She is a thought leader and mentor at select AI platforms, has co-authored multiple white papers\/blogs, and has been a speaker and panellist at various national and international forums, including TieCon, NASSCOM, data science institutes and AI associations. Swati was recently featured in “21 women in 21” by INDIAai, the national AI portal of India. In 2020, Swati was recognised among “Top 11 Women AI leaders in India” by Analytics India Magazine. Swati is part of the core team of the recently launched NASSCOM’s AI Maturity Assessment Framework for companies across the industry. Earlier, she worked at Cognizant, Pipal Research (CRISIL), and Ernst & Young (EY), where she provided data analytics, business intelligence, research and consulting services. Swati has a PhD in Economics from the Indian Institute of Technology, Delhi and a Masters in Business Economics from Delhi University. LinkedIn","excerpt":"Dr Swati is a seasoned professional with over two decades of analytics and consulting experience across multiple verticals, including financial services, retail, media, logistics, and healthcare.  Currently, she is the Vice President, Decision Analytics at EXL Service. She leads a world-class team of data scientists and AI\/ML professionals impacting marketing, risk management, and operations of […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"AIM Media House","publish_date":"2021-09-22T16:05:43","publication_year":"2021","word_count":225,"keywords":["business intelligence","data science","programming_languages:R","AI","ML","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","data science","analytics","R","GAN","business intelligence","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dr-swati-jain\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10009624,"title":"Complete Guide To ShuffleNet V1 With Implementation In Multiclass Image Classification","content":"With the recent advancement in the field of deep learning building deeper convolutional neural networks has become a trend for solving visualization problems. Though it gives accurate results the CNNs require computation of billions of FLOPS. To overcome this issue we introduce a computation efficient CNN architecture named ShuffleNet which is designed especially for mobile devices, drones and robots. It gives the best accuracy in the very limited computational budget. This article demonstrates how we can implement a deep learning model with ShuffleNet architecture to classify images of CIFAR-10 dataset.  Here, we define a Convolutional Neural Network (CNN) model using Torch to train this model. We will test the model to check the reduction in computational cost and obtain accuracy. Architecture of ShuffleNet This architecture uses pointwise group convolutions and channel shuffling to reduce the computational cost. In the first case, information is blocked as outputs from a certain group only relate to inputs within the group. To solve this issue we use channel shuffling operation as illustrated in the second case of the above figure. Information is passed on to different groups in the group convolution layer. About the Dataset The CIFAR-10 dataset contains 60,000 224×224 colour images in 10 different classes. There are 6,000 images of each class. The 10 different classes represent aeroplanes, cars, birds, cats, deer, dogs, frogs, horses, ships, and trucks. There are 6,000 images of each class. Implementation We will use google Colab for image classification. Start with mounting the drive. After signing in to google account we get the authorization code, then enter the code in the text box that will be displayed. The drive should now be mounted. Once the drive is mounted we will proceed to define the methods for loading the data set, initializing the CNN model, training and testing. First of all, we will import the required libraries. #Importing Libraries import torch import torch.nn as nn import torchvision.transforms as transforms import torchvision.datasets as dsets from torchvision import models from torchsummary import summary from torch import nn,optim import torch.nn.functional as F import numpy as np import pandas as pd import torchvision import os import sys import time import math import datetime as dt import tqdm import argparse import glob import matplotlib.pyplot as plt import tarfile import warnings import torch.optim as optim import torch.utils.data warnings.filterwarnings(\"ignore\") After it, we will proceed by displaying the image. Exploring the image dataset from matplotlib.pyplot import figure figure(num=None, figsize=(5, 5), dpi=150, facecolor='w', edgecolor='k') def show_imgs(X): plt.figure(1) k = 0 for i in range(0,3): for j in range(0,3): plt.subplot2grid((3,3),(i,j)) plt.imshow(Image.fromarray(X[k])) k = k+1 # show the plot plt.show() (x_train, y_train), (x_test, y_test) = cifar10.load_data() show_imgs(x_test[:9]) #Initialize values device = 'cuda' if torch.cuda.is_available() else 'cpu' best_acc = 0  # best test accuracy start_epoch = 0  # start from epoch 0 or last checkpoint epoch batch_size = 128 #Transform With data augmentation, we can get better accuracy. Normalize the data before training the data. Add padding, RandomHorizontalFlip and RandomCrop to it. print('==> Preparing data..') transform_train = transforms.Compose([ transforms.Pad(4), transforms.RandomHorizontalFlip(), transforms.RandomCrop(32), transforms.ToTensor(), transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)), ]) transform_test = transforms.Compose([ transforms.Pad(4), transforms.RandomHorizontalFlip(), transforms.RandomCrop(32), transforms.ToTensor(), transforms.Normalize((0.4914, 0.4822, 0.4465), (0.2023, 0.1994, 0.2010)), ]) trainset = torchvision.datasets.CIFAR10(root='.\/data', train=True, download=True, transform=transform_train) train_loader = torch.utils.data.DataLoader(trainset, batch_size=batch_size, shuffle=True, num_workers=2) testset = torchvision.datasets.CIFAR10(root='.\/data', train=False, download=True, transform=transform_test) test_loader = torch.utils.data.DataLoader(testset, batch_size=batch_size, shuffle=False, num_workers=2) classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck') # functions to show an image import matplotlib.pyplot as plt import numpy as np from matplotlib.pyplot import figure figure(num=None, figsize=(8, 8), dpi=150, facecolor='w', edgecolor='k') # functions to show an image classes = ('plane', 'car', 'bird', 'cat', 'deer', 'dog', 'frog', 'horse', 'ship', 'truck') def imshow(img): img = img \/ 2 + 0.5     # unnormalize npimg = img.numpy() plt.imshow(np.transpose(npimg, (1, 2, 0))) plt.show() # get some random training images dataiter = iter(train_loader) images, labels = dataiter.next() # show images imshow(torchvision.utils.make_grid(images)) # print labels print(' '.join('%5s' % classes[labels[j]] for j in range(5))) Defining Shufflenet for Our Work The below code snippet will define the ShuffleNet Architecture. The image 224*224 is passed on to the convolution layer with filter size 3*3 and stride 2. ShuffleNet uses pointwise group convolution so the model is passed over two GPUs.We get the image size for the next layer by applying formula (n+2p-f)\/s +1 where n input channel,p is padding,f is kernel size and s is stride. The features are passed on to a fully connected layer that classifies the image out of 1000 classes. import torch import torch.nn as nn import torch.nn.functional as F class ShuffleBlock(nn.Module): def __init__(self, groups): super(ShuffleBlock, self).__init__() self.groups = groups def forward(self, x): '''Channel shuffle: [N,C,H,W] -> [N,g,C\/g,H,W] -> [N,C\/g,g,H,w] -> [N,C,H,W]''' N,C,H,W = x.size() g = self.groups return x.view(N,g,C\/\/g,H,W).permute(0,2,1,3,4).reshape(N,C,H,W) class Bottleneck(nn.Module): def __init__(self, in_planes, out_planes, stride, groups): super(Bottleneck, self).__init__() self.stride = stride mid_planes =int(out_planes\/4) g = 1 if in_planes==24 else groups self.conv1 = nn.Conv2d(in_planes, mid_planes, kernel_size=1, groups=g, bias=False) self.bn1 = nn.BatchNorm2d(mid_planes) self.shuffle1 = ShuffleBlock(groups=g) self.conv2 = nn.Conv2d(mid_planes, mid_planes, kernel_size=3, stride=stride, padding=1, groups=mid_planes, bias=False) self.bn2 = nn.BatchNorm2d(mid_planes) self.conv3 = nn.Conv2d(mid_planes, out_planes, kernel_size=1, groups=groups, bias=False) self.bn3 = nn.BatchNorm2d(out_planes) self.shortcut = nn.Sequential() if stride == 2: self.shortcut = nn.Sequential(nn.AvgPool2d(3, stride=2, padding=1)) def forward(self, x): out = F.relu(self.bn1(self.conv1(x))) out = self.shuffle1(out) out = F.relu(self.bn2(self.conv2(out))) out = self.bn3(self.conv3(out)) res = self.shortcut(x) out = F.relu(torch.cat([out,res], 1)) if self.stride==2 else F.relu(out+res) return out class ShuffleNet(nn.Module): def __init__(self, cfg): super(ShuffleNet, self).__init__() out_planes = cfg['out_planes'] num_blocks = cfg['num_blocks'] groups = cfg['groups'] self.conv1 = nn.Conv2d(3, 24, kernel_size=1, bias=False) self.bn1 = nn.BatchNorm2d(24) self.in_planes = 24 self.layer1 = self._make_layer(out_planes[0], num_blocks[0], groups) self.layer2 = self._make_layer(out_planes[1], num_blocks[1], groups) self.layer3 = self._make_layer(out_planes[2], num_blocks[2], groups) self.linear = nn.Linear(out_planes[2], 10) def _make_layer(self, out_planes, num_blocks, groups): layers = [] for i in range(num_blocks): stride = 2 if i == 0 else 1 cat_planes = self.in_planes if i == 0 else 0 layers.append(Bottleneck(self.in_planes, out_planes-cat_planes, stride=stride, groups=groups)) self.in_planes = out_planes return nn.Sequential(*layers) def forward(self, x): out = F.relu(self.bn1(self.conv1(x))) out = self.layer1(out) out = self.layer2(out) out = self.layer3(out) out = F.avg_pool2d(out, 4) out = out.view(out.size(0), -1) out = self.linear(out) return out def ShuffleNetG2(): cfg = { 'out_planes': [200,400,800], 'num_blocks': [4,8,4], 'groups': 2 } return ShuffleNet(cfg) def ShuffleNetG3(): cfg = { 'out_planes': [240,480,960], 'num_blocks': [4,8,4], 'groups': 3 } return ShuffleNet(cfg) net = ShuffleNetG2() x = torch.randn(1,3,32,32) y = net(x) print(y) net = ShuffleNetG2() print(net) #Setting the model on CUDA if torch.cuda.is_available(): net.cuda() criterion = nn.CrossEntropyLoss() optimizer = optim.SGD(net.parameters(), lr=0.01, momentum=0.9, weight_decay=5e-4) Training Testing and Making Predictions Now, we are all set to train and test the model on the CIFAR-10 dataset. Before we start setting the model on CUDA and use stochastic gradient descent optimizer.For better accuracy train and test the model with 60 epochs. # Training the model def train_net(): net.train() train_loss = 0 n_correct = 0 n_total = 0 for batch_size, (inputs, targets) in enumerate(train_loader): inputs, targets = inputs.to(device), targets.to(device) optimizer.zero_grad() outputs = net(inputs) loss = criterion(outputs, targets) loss.backward() optimizer.step() train_loss += loss.item() _, predicted = outputs.max(1) n_correct += predicted.eq(targets).sum().item() n_total += targets.shape[0] return train_loss\/(batch_size+1),n_correct\/n_total def get_loss_acc(is_test_dataset = True): net.eval() dataloader = test_loader if is_test_dataset else train_loader n_correct = 0 n_total = 0 test_loss = 0 with torch.no_grad(): for batch_size, (inputs, targets) in enumerate(dataloader): inputs, targets = inputs.to(device), targets.to(device) outputs = net(inputs) test_loss += criterion(outputs, targets).item() _, predicted = outputs.max(1) n_correct += predicted.eq(targets).sum().item() n_total += targets.shape[0] return test_loss\/(batch_size+1),n_correct\/n_total #Testing the model def test(epoch): global best_acc net.eval() test_loss = 0 correct = 0 total = 0 with torch.no_grad(): for batch_idx, (inputs, targets) in enumerate(test_loader): inputs, targets = inputs.to(device), targets.to(device) outputs = net(inputs) loss = criterion(outputs, targets) test_loss += loss.item() _, predicted = outputs.max(1) total += targets.size(0) correct += predicted.eq(targets).sum().item() print(batch_idx, len(test_loader), 'Loss: %.3f | Acc: %.3f%% (%d\/%d)' % (test_loss\/(batch_idx+1), 100.*correct\/total, correct, total)) # Save checkpoint. acc = 100.*correct\/total if acc > best_acc: print('Saving..') state = { 'net': net.state_dict(), 'acc': acc, 'epoch': epoch, } if not os.path.isdir('checkpoint'): os.mkdir('checkpoint') torch.save(state, '.\/checkpoint\/ckpt.pth') torch.save(net, '.\/checkpoint\/net.pth') best_acc = acc import glob import torch.optim as optim import datetime as dt EPOCH = 60 start = dt.datetime.now() start_epoch=0 for epochi in range(start_epoch,start_epoch + EPOCH): #scheduler.step() cur_lr = [i['lr'] for i in optimizer.param_groups][0] print(\"Batch Size\",batch_size,'(%.2fs)\\n\\nEpoch: %d\/%d | cur_lr:%.4f ' % ( (dt.datetime.now()-start).seconds, epochi+1,EPOCH+start_epoch,cur_lr)) start = dt.datetime.now() test_loss , test_acc = get_loss_acc() train_loss , train_acc = train_net() #hist.append([train_loss , train_acc,test_loss , test_acc]) print( 'train Loss: %.3f | Acc: %.3f%% \\ntest Loss: %.3f | Acc: %.3f%% ' % ( train_loss, train_acc*100,test_loss, test_acc*100)) #Count Parameters def count_parameters(model): pytorch_total_params = sum(p.numel() for p in model.parameters()) print(\"Total_params\",pytorch_total_params) pytorch_total_params = sum(p.numel() for p in model.parameters() if p.requires_grad) print(\"Trainable_params\",pytorch_total_params) Compared to other architecture the number of parameters is less. #Model Accuracy total_correct = 0 total_images = 0 confusion_matrix = np.zeros([10,10], int) with torch.no_grad(): for data in test_loader: images, labels = data images = images.to(device) labels = labels.to(device) outputs = net(images) _, predicted = torch.max(outputs.data, 1) total_images += labels.size(0) total_correct += (predicted == labels).sum().item() for i, l in enumerate(labels): confusion_matrix[l.item(), predicted[i].item()] += 1 model_accuracy = total_correct \/ total_images * 100 print('Model accuracy on {0} test images: {1:.2f}%'.format(total_images, model_accuracy)) Results of the Model #Result print('{0:10s} - {1}'.format('Category','Accuracy')) for i, r in enumerate(confusion_matrix): print('{0:10s} - {1:.1f}'.format(classes[i], r[i]\/np.sum(r)*100)) #Plot the confusion Matrix fig, ax = plt.subplots(1,1,figsize=(8,6)) ax.matshow(confusion_matrix, aspect='auto', vmin=0, vmax=1000, cmap=plt.get_cmap('Blues')) plt.ylabel('Actual Category') plt.yticks(range(10), classes) plt.xlabel('Predicted Category') plt.xticks(range(10), classes) plt.show() Conclusion As we can see in the above result the model has very high accuracy on both training and test. The number of parameters is less thereby reducing the computational complexity. So we can conclude that the model has given accurate predictions to classify images on the CIFAR-10 dataset. With an increase in the number of epochs, we can get better accuracy. We can experiment it further by adding jitter and brightness to the dataset.The complete code for the above implementation is available at the AIM’s GitHub repository. Please go through this link to check the notebook with the codes.","excerpt":"This article demonstrates how we can implement a deep learning model with ShuffleNet architecture to classify images of CIFAR-10 dataset. Here, we define a Convolutional Neural Network (CNN) model using Torch to train this model. We will test the model to check the reduction in computational cost and obtain accuracy.","categories":["Deep Tech"],"tags":["Computer Vision","Deep Learning","Image Classification","multiclass classification"],"author_name":"Ankit Das","publish_date":"2020-10-13T18:00:00","publication_year":"2020","word_count":1607,"keywords":["NumPy","AI","neural network","PyTorch","Colab","Ray","multiclass classification","deep learning","Aim","Computer Vision","Deep Learning","Matplotlib","Image Classification","Pandas"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","Ray","PyTorch","Colab","Pandas","NumPy","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/complete-guide-to-shufflenet-v1-with-implementation-in-multiclass-image-classification\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":61605,"title":"Tally Solutions Extends Its Support To Businesses Amid COVID-19","content":"To help businesses transition through these difficult times, Tally Solutions extended its support to businesses through several initiatives. COVID 19 has brought the country and almost the entire world to a standstill. Several enterprises, especially the small and medium ones, are directly impacted across the globe, and that’s why Tally Solutions have come up with these initiatives for businesses. Tally has introduced a free of cost 30 days license of its flagship product — Tally.ERP 9 to businesses across the globe. Companies need to register on their website and avail of this benefit. The support started 1st April of this year, and 10,000 firms have already signed up till date. For existing customers, the company has extended the subscription by 30 days. This extension has benefitted over 1.3 lakh SMBs who can now access a bouquet of connected services and latest product updates that comes with it without being financially burdened in these difficult times. Tally’s customer support centre TallyCare, moved from a centralised system to 100% decentralised system overnight and with zero downtime, keeping customers’ needs and issues paramount, while ensuring social distancing. Tejas Goenka, Managing Director at Tally Solutions, said, “With the lockdown announced by many governments around the world, our partner ecosystem and our teams have taken it upon themselves to extend as much support to both our customers and society at large.” He further said, “While we have detailed a few of these initiatives here, there are several more we will be announcing in the coming few days.” Tally is also conducting regular webinars for customers and Chartered Accountants on various subjects like remote access and browser access of the software, moving to new financial year, data synchronisation, banking features of Tally etc. assisting inefficient work from home. Over 10,000 businesses have been reached out through 40 such sessions with several more planned in the days to come. Similar webinars are being conducted for the ICAI Dubai, and Abu Dhabi chapter to enhance their WFH productivity, managing customer expectations and spreading awareness on the new offers so maximum businesses can be benefitted, resulting in reaching out to around 5000 CAs. Tally is also hosting several webinars with industry associations like CII to share updates and tips on working remotely and staying connected with their customers. The company’s broad ecosystem of over 28000 partners is reaching out to customers across the country to enable them to work remotely and provide the extended support needed at this time. They have been enabled through multiple webinars on the support to be extended to customers.","excerpt":"To help businesses transition through these difficult times, Tally Solutions extended its support to businesses through several initiatives. COVID 19 has brought the country and almost the entire world to a standstill. Several enterprises, especially the small and medium ones, are directly impacted across the globe, and that’s why Tally Solutions have come up with […]","categories":["AI News"],"tags":["covid-19"],"author_name":"Sejuti Das","publish_date":"2020-04-13T18:39:00","publication_year":"2020","word_count":425,"keywords":["Go","covid-19","AI","programming_languages:R","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tally-solutions-extends-its-support-to-businesses-amid-covid-19\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45484,"title":"Deep Dive: How Zivame Is Using Data Science To Empower Women Buying Lingerie Online","content":"We all have seen and been through the social discomfort women go through, when it comes to buying lingerie in India, especially in smaller towns. However, those days are gone, and one company has played a major role in bringing this change — Zivame. Founded in 2011, Zivame today is helping women uninhibitedly shop for intimate wear. Apart from its online platforms, app, and website, Zivame is also present on other online market spaces like Myntra, Amazon, and Flipkart. Furthermore, apart from being the fastest growing brand in this space, it is also the largest online player at present. In April this year, the company hit close to 70% plus growth rate; there is no other player that has achieved this in the category, and it is on track to almost doubling up that this year as well. In order to know more about Zivame and how the company is battling it out to become the leader in the space, we got in touch with Anuj Gupta, Chief Revenue Officer at Zivame, who has been with Zivame for over two years now. Role Of Data At Zivame Data science over the years has helped a significant number of companies as well as industries, and it is one of the major aspects at Zivame as well. The data science team focuses on various kinds of business problems starting from the consumer-faced problems to problems that are under efficiency like merchandising. Furthermore, this retail company has technology at its DNA, “We are a tech-first company. Because of our technology DNA, we started taping data side with artificial intelligence. And it was based on very strong foundations.” However, the company had to make some little changes such as changes on some of the architectural features — going from mobile to microservices, which makes it easier for some of these data science pieces to put in. Anuj also said that It does not affect the efficiency of the platform, it remains fast no matter what the amount of computation that is happening. Gupta also shared how data has been playing a vital role in delivering the best product. Being a company that deals with lingerie, the single biggest job for Zivame is to design products that fit well and are comfortable. This is a problem which was extremely under-serviced in this category till now. Zivame has something called “Fit Code”. It is a system that uses past data and could predict 2 to 4 different body profile data. When a consumer comes to Zivame’s platform, they have to answer some very simple questions. As they continuously answer, the company tries to segment them through its market algorithm. “Till now, we have come up with 300 specific body profiles through our Fit Code,” said Gupta. “Also, we continuously keep on modifying and adding to these body sizes to help customers find the best fit.” Further, once the profiling is done, then it starts personalizing their experience on the platform. For example, if you go on the platform and if you go for the Fit Code product and then when you go back to the realistic ratings, you will find that you have a bra which can be your size. “What’s the fascinating thing about the Fit Code is that no woman has to even measure herself, to get to her size or find out about her right figure. We study your fit and size without you ever having to pick up the measuring tape and measure yourself,” Gupta added. Filling The Void One of the major gaps for a consumer is the rudimentary nature of the shopping experience. There are many brands that have offline stores and the person that attends a customer is a man, and sometimes, it gets extremely awkward to have a conversation about lingerie and size and age. And when asked how Zivame is filling the void, Gupta said, “Apart from Zivame, there is no other brand which has tried to build this conversation through our brand campaigns.” Another void is the product itself — many brands they just pick up any kind of material which comes cheap to them. They get large stocks and make a product for the consumer, which they can successfully sell for a higher price. “We are not in that space. We are in the game of giving the consumer the right size and the right fit and the right frame for her body,” said Gupta “In Zivame, there is a working team that provide lingerie with the best quality and that fits well.” The Hiring Phase As the conversation kept going, we also asked about the company’s hiring strategy. The first and foremost thing that Zivame look for in a candidate is the ability to fit in with the company’s culture, “Because Zivame probably is the most exciting thing that has happened in lingerie in the past two decades in this country.” For jobs in data science-related roles, Zivame needs candidates who have a strong hands-on fundamental knowledge as well as computer learning tools. They also need to keep updated with all the latest things happening in the space. “We also look at whether they are then hands-on or not. In terms of over-all, they need to have a problem-solving mechanism because, eventually, that sometimes becomes a major requirement,” Gupta added. Furthermore, the company doesn’t have a bent towards hiring only from a specific kind of colleges. When Zivame is hiring, they look for certain characteristics and a certain kind of problem-solving mindset and a certain kind of culture-fit. Roadmap Ahead Going forward, the goal for the company is to continue to grow at a rapid speed that it has been going on, along with working the consumer experience side from the technology front and also from the omnichannel front. “We want to give the consumer standardized Zivame sales across different platforms. No matter what platform you go to, it is the same Zivame experience,” Gupta concluded. “So if you navigate to go to our stores, you have an endless aisle. You can order anything and immediately, and can get it delivered to your home. This is a standardized experience.”","excerpt":"We all have seen and been through the social discomfort women go through, when it comes to buying lingerie in India, especially in smaller towns. However, those days are gone, and one company has played a major role in bringing this change — Zivame. Founded in 2011, Zivame today is helping women uninhibitedly shop for […]","categories":["AI Features"],"tags":["Data Science"],"author_name":"Harshajit Sarmah","publish_date":"2019-09-04T16:30:45","publication_year":"2019","word_count":1025,"keywords":["data science","Go","API","artificial intelligence","programming_languages:R","AI","programming_languages:Go","microservices","ViT","Data Science","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","microservices","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deep-dive-how-zivame-is-using-data-science-to-empower-women-buying-lingerie-online\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10057792,"title":"NVIDIA is offering a four-hour, self-paced course on MLOps","content":"NVIDIA Deep Learning Institute (DLI) is offering a four-hour, self-paced course titled “Deploying a Model for Inference at Production Scale” that introduces MLOps coupled with hands-on practice with a live NVIDIA Triton Inference Server. The course’s learning objectives include: Deploying neural networks from a variety of frameworks onto a live NVIDIA TritonServer.Measuring GPU usage and other metrics with Prometheus.Sending asynchronous requests to maximise throughput. Upon completion, developers will deploy their models on an NVIDIA Triton Server. What is Triton Server? NVIDIA Triton Inference Server helps data scientists and system administrators turn the same machines you use to train your models into a web server for model prediction. Though GPU is not required, the server can take advantage of multiple installed GPUs to quickly process large batches of requests. NVIDIA Triton was created with Machine Learning Operations (MLOps). This is a relatively new field evolved from Developer Operations, or DevOps, to focus on scaling and maintaining ML models in a production environment. NVIDIA Triton is equipped with features such as model versioning for easy rollbacks. Triton is also compatible with Prometheus to track and manage server metrics such as latency and request count. Register for this free course on Data & Analytics.","excerpt":"NVIDIA Triton helps data scientists and system administrators turn models into a web server for model prediction using the same training machine","categories":["AI News"],"tags":[],"author_name":"Poornima Nataraj","publish_date":"2022-01-06T16:20:15","publication_year":"2022","word_count":201,"keywords":["machine learning","programming_languages:R","AI","neural network","ML","MLOps","deep learning","analytics","DevOps","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","analytics","MLOps","R","DevOps","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nvidia-is-offering-a-four-hour-self-paced-course-on-mlops\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10078221,"title":"The winners approach: Cracking the complex Hiring Hackathon by ReNew Power","content":"Global renewable energy player ReNew Power, in partnership with MachineHack, successfully concluded its two-week-long hiring hackathon on September 9, 2022. The hackathon witnessed participation from nearly 1000 developers and AI enthusiasts who took part in predicting and flagging the failure of wind turbines. Out of these, the top three participants triumphed and impressed the judges. During the hackathon, minute-wise normalised data of wind speed, power and temperature data for multiple components of a wind turbine was shared with the participants. The company wanted to create a model for an ideally functioning turbine’s expected rotor bearing temperature. The model will be used to check the deviation of the actual rotor bearing temperature of the faulty turbine from the expected temperature. To solve the problem, MachineHack and ReNew Power presented the participants with a training dataset of 909604 rows with 16 columns and a testing dataset of 303202 rows with 15 columns. In addition, AIM spoke to the winners to understand their data science journey, winning approach, and overall experience at MachineHack. Check out the launch article here. Rank 01: Andrey Bessalov – Moscow, Russian Federation After completing his studies in mathematics a decade ago, Bessalov started working as a data scientist. He has participated in ML hackathons on the platforms for two years and has learned much from these competitions. His most memorable competitions are: Renew Power, Dare in Reality Hackathon 2021 and Rocket Capital Crypto Forecasting. Bessalov currently works as a data scientist at VISA and has previously won two MachineHack hackathons. Check out the winner’s approach here. Rank 02: Shubhankar Mishra – Mumbai, India The PhD Research Scholar from IIT Bombay was introduced to statistical and ML concepts during his master’s coursework. During the pandemic, Mishra was left with lots of free time as his research involved a lot of experimental work. This led him to develop an interest in data science, and he started learning about it. Eventually, he decided to make a career transition and pursue data science as a professional. Mishra stumbled upon MachineHack while looking for a dataset for his project. He later discovered that the platform also hosts other functionalities like boot camps, courses and a career portal. So, whenever he wants to participate in a hackathon or explore datasets for projects, MachineHack is his first choice. “I also appreciate the quick responsiveness from the MachineHack team whenever I had a query during the hackathon,” Mishra said. Check out the winner’s approach here. Rank 03: Mahesh Chandra Duddu – Ongole, India To compete with his brother during the pandemic, Duddu started participating in ML hackathons, learning data science through courses and projects. His growing interest in ML led him to his MTech project, “Detection of Social Bots in Twitter Network”, which was recently accepted in IJCACI 2022. Besides hackathons, MachineHack is his go-to platform for applying and making like-minded friends. “MachineHack team helps participants by uploading a starter notebook code for the problem to solve, which really helps. Also, the way the team deals with competition issues is what I really liked the most,” Duddu said. Check out the winner’s approach here.","excerpt":"The two-week-long hackathon witnessed participation from nearly 1000 developers.","categories":["AI Trends"],"tags":["Career","Hackathon"],"author_name":"Tasmia Ansari","publish_date":"2022-10-28T15:00:00","publication_year":"2022","word_count":518,"keywords":["data science","Go","API","programming_languages:R","AI","ML","programming_languages:Go","Hackathon","Aim","R","Career"],"extracted_tech_keywords":["AI","ML","data science","Aim","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-winners-approach-cracking-the-complex-hiring-hackathon-by-renew-power\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10079920,"title":"Salesforce Joins Hands with Tableau to Provide Cloud-based Analytics Solutions","content":"Recently, Salesforce announced the launch of Analytics Bundle, a self-service platform which will help companies consolidate their analytic tools. The new platform, built with Tableau, an analytics platform, is aimed at increasing customer efficiency by empowering them with tools to attain real-time insights and make decisions swiftly. According to a survey of 3,706 salesforce customers located across multiple nations, 29% faster time to insight, faster delivery of business driving reports, and an average 26% decrease in time spent on analysing information were observed. In May, Tableau announced the launch of Tableau Cloud, a cloud-based solution to fast, easy-to-use, and cost-effective analytics. The cloud platform is host to many powerful innovations including advanced management, and data stories, among others. With Tableau, Salesforce has been able to widen its fortress of analytics to different domains—such as providing flexible and easy-to-use enterprise analytics solution that could be fit into a customer’s current enterprise architecture, enabling access to tools with which customers can discover, understand, connect, and trust their data with Tableau’s Data Management, as well as, help them to manage, secure, and scale mission-critical analytics through Tableau’s Advanced Management. Additionally, the Analytics Performance Bundle tool also offers courses and training materials with Tableu’s e-learning. However, this is not the first time Salesforce has collaborated with Tableau. In January this year, they teamed up with each other to launch CRM Spring ‘22 to automate business processes and build intelligent solutions with Einstein Discovery in Salesforce Force.","excerpt":"Recently, Salesforce announced the launch of Analytics Bundle, a self-service platform which will help companies consolidate their analytic tools. The new platform, built with Tableau, an analytics platform, is aimed at increasing customer efficiency by empowering them with tools to attain real-time insights and make decisions swiftly.    According to a survey of 3,706 salesforce customers […]","categories":["AI News"],"tags":["Data Analytics","Salesforce","Tableau"],"author_name":"Ayush Jain","publish_date":"2022-11-16T15:53:43","publication_year":"2022","word_count":242,"keywords":["Tableau","programming_languages:R","AI","innovation","RAG","Aim","analytics","Salesforce","Rust","Data Analytics","R","analytics platform","programming_languages:Rust"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Rust","analytics platform","innovation","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/salesforce-joins-hands-with-tableau-to-provide-cloud-based-analytics-solutions\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":63966,"title":"This Is Why Cybersecurity Is The Top Priority During The COVID-19 Lockdown","content":"Business continuity during the COVID-19 lockdown is a big challenge. Firms are not just at risk of losing web connectivity and outages, but data security vulnerabilities and cybersecurity attacks from malicious attackers. There have seen many incidents of phishing, misinformation campaigns, and others work-from-home opportunities for hackers making their way around the internet. With the lockdown extended around the world, employees continue to work remotely on their private networks, which is undoubtedly a threat to most Indian companies. According to a study by PwC, the number of cyberattacks on Indian companies has doubled in the past few months as cybercriminals use the disruption brought about by the COVID-19 outbreak to infiltrate corporate networks and steal data. The CERT-In (The Computer Emergency Response Team of India) stated in its latest advisory to the internet users that, “Cybercriminals are exploiting the coronavirus pandemic outbreak as an opportunity to send phishing emails in the form of an ‘important update’ or ‘encouraging donations’, or trying to impersonate employees’ trustworthy organisations.” According to the agency, the current global health situation has seen changes to the way people accomplish their regular job, with an increasing number working from home instead of the office. Cybercriminals are continually attempting to take advantage of the COVID-19 pandemic and are now turning their attention to mobile devices to spread malware, including spyware and ransomware. The officials said that employees who are switching to remote working because of the coronavirus outbreak could create cybersecurity problems for the business and the employers. “In the current context where the same version of confidential\/sensitive data is spread across an organization and sits in various places- databases, cloud platforms, collaboration tools, file system, endpoints, e-mails, etc., it is getting very challenging for security professionals, security architects and security engineers to come up with a single solution to address all security gaps at various levels,” Visweswara Rao Sreemanthula, a Senior Manager – IT Security at Verizon told Analytics India Magazine. In April, Indian IT giant Infosys beefed up its security to safeguard itself from potential cyber-attacks. And the firm has further planned to enhance security in 2020 by expanding and reskilling its team. Vishal Salvi, chief information security officer and head of cybersecurity at Infosys recently said, “Investment in cybersecurity controls are on the rise year-on-year, and that is because organizations are considering cybersecurity investment very strategic for their current and future business.” According to Salvi, the company is mainly focusing on reskilling its team in identity and access management, infrastructure security, security information, and event management, security orchestration, automation, and response. The Security Plan The sudden shift to a remote-work model means that employees are now combining personal technology with work networks, and this is contributing to an expanded attack surface. Many of these devices may also be older or unsecured, and this introduces serious new risks. All of this can be challenging for security teams who now have to manage this expanded and complex attack surface. Remote working practices are in full swing, especially in these challenging times of COVID-19. Organizations are taking the best precautionary steps to support and protect employees during this global pandemic. Here, active traffic management is critical for application services to meet the new levels of demand and handle peak loads in traffic. Effective traffic management is also a key priority for organizations during this time. It is, therefore, critical for businesses to look at the various means by which they can address the complex security challenges, with the escalating threats prevailing due to remote working during the COVID-19 outbreak. Network security, data availability, and protection have become a crucial priority for organizations for truly seamless business continuity. On ensuring network security and data traffic management at the time of COVID-19, we also talked with Sanjai Gangadharan, Regional Director, SAARC, A10 Networks who said, “Remote working is the need of the hour for organizations in India as social distancing amid lockdown becomes a priority in fighting the COVID-19 battle. Network protection from distributed denial of service (DDoS) attacks is a key concern of remote working for organizations.” “Organizations should continually assess their networks for security vulnerabilities. This can prevent a range of problems such as unauthorized access to applications and identifying underlying software flaws that expose sensitive data. Vulnerability scanners can help identify these concerns, making it easier to understand if systems have critical risks that need to be addressed,” Adam Palmer, Chief Cybersecurity Strategist at Tenable told. According to Adam, as a first step, it’s important to identify its information assets, having a baseline will help in knowing the width and depth of what has to be protected. Once determined, organizations need to work along with various stakeholders in designing controls which will help achieve the security business objectives. Other experts and leaders in the security industry say that in order to respond to the cybersecurity complexities of remote working due to COVID-19, organizations must take a zero-trust approach to security. They must ensure that no user has access to data that they don’t depend on for their day-to-day functions. Companies must also ensure visibility into all users, traffic, data, and workloads, and have uniform security policies applied across all locations to make sure no security loopholes exist.","excerpt":"Business continuity during the COVID-19 lockdown is a big challenge. Firms are not just at risk of losing web connectivity and outages, but data security vulnerabilities and cybersecurity attacks from malicious attackers. There have seen many incidents of phishing, misinformation campaigns, and others work-from-home opportunities for hackers making their way around the internet. With the lockdown […]","categories":["AI Trends"],"tags":["covid-19","data protection india","Data Security","LOCKDOWN","Ransomware"],"author_name":"Vishal Chawla","publish_date":"2020-04-30T11:00:00","publication_year":"2020","word_count":870,"keywords":["covid-19","AI","Data Security","Ransomware","data protection india","ML","AWS","Scala","RAG","LOCKDOWN","ViT","analytics","Rust","GAN","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","AWS","R","Rust","Scala","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/cybersecurity-lockdown-priority\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":59261,"title":"Deep Learning Image Classification with CNN &#8211; An Overview","content":"In this article, we will discuss how Convolutional Neural Networks (CNN) classify objects from images (Image Classification) from a bird’s eye view. First, let us cover a few basics. Let us start with the difference between an image and an object from a computer-vision context. What we see above is an image. We can see 3 objects inside – 1 cat and 2 dogs. If you wish, we can count the ribbon on the head of the left one as 4th object. There are mainly 2 types of images – Red Green Blue (RGB) scale and grayscale (black & white) as illustrated below. RGB has 3 channels – Red, Green and Blue while Grayscale has only 1 channel. Computers only understand the language of mathematics. Hence we will convert images to tensors using libraries like Python Imaging Libraries (PIL). We can imagine tensors as n-dimensional matrices as illustrated below. Now let us understand how computers classify images using CNN. Below is a high-level representation of how CNNs work. Let us understand the above flow block-by-block. Please note that information is flowing back and forth. Black arrows represent forward pass while white arrows represent a backward pass. Forward Pass This is where the image is broken down into features, reconstructed and predicted at the end. Let us examine each step involved. Input: Images will be fed as input which will be converted to tensors and passed on to CNN Block. 2. CNN Block This is the most important block in the neural networks. The following steps will happen inside the CNN block. Input tensor will be broken down into basic channels. Imagine this like dismantling an assembled lego board to smaller pieces. The features inside these channels are then used to construct edges and gradients.Using these edges and gradients, we construct textures and patterns.From these textures and patterns, we build parts of objects.These parts of objects will be used to reconstruct objects. The below image shows these steps. We perform the above steps with the help of a mathematical operation called convolution. Input tensors that CNN blocks receive are comprised of numerical values that represent the pixel amplitudes from the original image. The convolution operation is performed on these input channels using kernels to extract features. Kernels are also tensors with values in each cell. We call kernel values as weights. Below is an example of a convolution operation performed on a 7×7 input channel (I) using a 3×3 kernel (K) to extract a 5×5 feature (I * K). Values appearing in output are sum-product of each convolution. Below is another example that shows how a vertical edge is extracted using convolutions. An edge is detected when there is a sharp change in pixel values. Please note that higher numbers in tensor represent brighter pixels i.e. 10 will be brighter than 0. CNN block will have multiple convolutional layers stacked one after another with the objective of extracting edges and gradients -> textures & patterns -> parts of object -> object. Dense 1, Dense 2 Towards the final convolution layers, we can expect channels to resemble the original object that we are attempting to classify. These channels will be multi-dimensional. We need to flatten this to a 1-D array to make it compatible for a logistic regression classifier. First step in this process represented as Dense 1 is an operation called Global Average Pooling (GAP) as shown below. Here we are reducing the channel of size 6x6x3 to 1x1x3. Next, we will flatten our 1x1x3 channel to a one-dimensional array with 3 elements. We either use 1×1 convolution operations or functions like nn.linear in Pytorch to achieve this. This step is represented as Dense 2 in forwarding flow. The below image depicts this operation. A number of elements in the 1-D array must be exactly equal to the classes involved in the image classification problem. For example, if we are trying to predict digits, then a number of classes and hence a number of elements in the 1-D array will be 10 to accommodate digits from 0-9. 2. Output as One-Hot Vector Final step in the forward pass is classifying the image. We will feed flattened 1-D array to a logistic regression classifier to predict the image class. The most commonly used classifier for this task is Softmax. Softmax gives the probability distribution of the list of classes. The class with the highest probability will be selected as the predicted class. In the below example dog will be predicted class. We call the output as One-Hot Vector because only one of the nodes (predicted class) will have value. Backward Pass We covered how an image is classified via forward pass. Next, let us inspect what happens backward. We call this back propagation. This is where CNN collects feedback and improves itself. After prediction, each layer will receive feedback from its preceding layer. Feedback will be in the form of losses incurred at each layer during prediction.Aim of the CNN algorithm is to arrive at optimal loss. We call this as local minima.Based on the feedback, network will update the weights of kernels. This will make the output of convolutions better when next time forward pass happens.When the next forward pass happens, loss will come down. Again, we will do back prop, the network will continue to adjust, a loss will further come down and process repeats.This forward pass followed by back prop keeps happening the number of times we choose to train our model. We call it epochs. I hope this gave you a high-level understanding of how a deep learning Convolutional Neural Network (CNN) works and classifies objects from an input image. This article is presented by AIM Expert Network (AEN), an invite-only thought leadership platform for tech experts. Check your eligibility.","excerpt":"In this article, we will discuss how Convolutional Neural Networks (CNN) classify objects from images (Image Classification) from a bird’s eye view.  First, let us cover a few basics. Let us start with the difference between an image and an object from a computer-vision context. What we see above is an image. We can see […]","categories":["Deep Tech"],"tags":[],"author_name":"Anilkumar N Bhatt","publish_date":"2020-03-23T12:01:00","publication_year":"2020","word_count":962,"keywords":["TPU","AI","neural network","PyTorch","RAG","Python","Ray","Aim","deep learning","R"],"extracted_tech_keywords":["AI","deep learning","neural network","Aim","Ray","PyTorch","RAG","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/deep-learning-image-classification-with-cnn-an-overview\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10051601,"title":"A Survival Guide To Data Science Information Overload","content":"When kickstarting one’s journey as a data scientist, one might be overwhelmed by the vastness of the area of study. The resources are immense and massively growing. However, this readily available information comes with the challenge of selecting between one thing and its multiple alternatives. This availability of a large amount of information and alternatives is termed information overload. Coined by Political Science Professor Bertram Gross in 1964, ‘information overload,’ by definition, occurs when the amount of information input to a system exceeds its processing power or capacity. And when such information overload occurs, the decision quality automatically decreases. The Cause Information overload is a usual phenomenon among data scientists, especially among the ones just starting out in the field. The possible causes for information overload among data scientists are: The evolving nature of the field: Data science is an extremely dynamic and evolving field. Although it is a relatively new field of study and profession, new research, white papers, and theories related to data science are published every day. While this might encourage one to learn more about newer topics and concepts, it may overwhelm and discourage some from taking up data science altogether. Unrealistic expectation: A quick look at any data science job opening would reveal that hiring managers expect data scientists to know several technical skills. In doing so, they not just overburden aspiring data scientists but also learning several skills at the same time make them masters of none. The vastness of scope: Data science is not limited to one thing. Concepts of business intelligence, natural language processing, data engineering, data analysis, coding, building algorithms, statistics– all of these come under the one umbrella of data science. In order to master it all, one ends up being confused and knowing concepts only superficially. The Solution Once the challenges and causes of information overload are identified, aspiring or existing data scientists should move to overcome the challenges. The possible solutions to get the better of the information overload challenge, one must comply with the following: Choose a mentor: Like every field of study, data science requires mentorship and guidance. Aspiring data scientists or beginners should find for themselves a mentor or guide to help them pave the path. It is therefore important to have someone to guide while navigating the data science landscape. Sticking to the plan: After researching, a beginner must select an area of interest and stick to it. If natural language processing interests you, study, research and excel at it. If data analysis and business intelligence is your cup of tea, stay focused and try excelling in it. To make this easier, take the data science personality test here. Selecting the right tools: Data science can be tricky when it comes to selecting a tool. The same tasks can be performed with different tools. Even when just starting out in the field and wanting to learn a programming language, one might get confused between Python, R, C and C++. The selection of the aptest tool depends on the area of interest and what one wants to gain expertise in. For example, both R and Python come in handy for data analysis. Similarly, there is a plethora of business intelligence tools that one can use. Different people in the data science field will have different opinions on tools and programming languages. The best way to overcome this obstacle is to select the tool that best suits your purpose and is one that you are most comfortable using. Identifying the right resources: There is no dearth of resources to master data science. With innumerable courses being available online, the best way to select one is by identifying the key requirements, which is why self-paced learning courses are advised. Similarly, if you wish to take up a course on business intelligence, select a course that teaches using the tool of your choice and helps solve the problem statements that you ultimately want to find solutions to. To know more about how to select the right data science training programme, click here. While it is easy to get swamped with the available options and alternatives available to one starting out with data science, keeping in mind the above-mentioned points will allow one to stay focused and excel in the field. To know how to land the right data science job, click here.","excerpt":"Here are a few tips for aspiring or professional data scientists on how to remain focused during information overload.","categories":["Deep Tech"],"tags":["Guide"],"author_name":"Debolina Biswas","publish_date":"2021-10-14T14:00:00","publication_year":"2021","word_count":724,"keywords":["business intelligence","data science","Go","AI","RAG","Python","BERT","C++","data engineering","R","Guide"],"extracted_tech_keywords":["AI","data science","RAG","Python","R","Go","C++","data engineering","BERT","business intelligence"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-survival-guide-to-data-science-information-overload\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":13263,"title":"Healthcare analytics in India is poised for a change – see how you can make healthcare analytics career","content":"Want to make a headstart in healthcare analytics? Follow these best practices to make a career as healthcare analyst The Indian healthcare system in India is poised for a data-driven transformation, delivering value based care backed by a well-defined digital and analytics strategy. One of the most data-intensive, data rich fields, data flows in from disparate sources, posing several challenges in data aggregation and data governance policies. The data deluge apart, healthcare sector is grappling with a demand for skilled data analysts, who can accurately leverage data to engage patients and drive operational efficiency. Let’s define healthcare analytics and who the different stakeholders in it are: Rohit Kumar, Chief of Analytics, THB Demystifying healthcare analytics is Rohit Kumar, cofounder of THB and the Chief of Analytics at the Gurgaon-based clinical research and data analytics startup.  According to Kumar, Healthcare Analytics is a generic terminology and can be viewed differently by different stakeholders —  a) doctors are more interested in clinical analytics, such as personalized treatment where data drives decision making and can suggest the next course of treatment for a specific patient that will give optimal result; b) for government or organizations it can imply using collated data to identify patterns in disease occurrence\/recurrence or even treatment; c) for providers it entails automating tasks or minimizing errors by performing first level of diagnosis through algorithms or generating x-rays reports through image processing. Role of Healthcare analytics in improving patient outcomes and paring costs According to IBM, analytics in healthcare can play a key role in reducing high-risk healthcare problems and usher in evidence based personalized medicine. Some of the key areas addressed by healthcare analytics are: Bringing personalization and engagement into healthcare via patient centricity Analytics as an enabler for evidence-based medicine and disease prevention Building a pro-active, sustainable healthcare system and usher in accountability and transparency Providing supply forecasting, dynamic budgeting and reducing overall cost Promoting data-centricity and data literacy How to kick-start a career as a healthcare data analyst in India? Healthcare has emerged as a high-impact field with a deep emphasis on patient-centricity, and the field has seen a boom in US and Canada with the surge in genomic science (science of finding complex diseases with genetics) and computational medical treatment. Ideally, the first step is getting a masters in data analytics and start by gaining experience in healthcare sector. By and large, that’s the stock answer. To give you a clear picture AIM gets Kumar to weigh in on the best practices and tips and guidelines for breaking into this up and coming sector. One of the biggest challenge in healthcare analytics is unstructured data. Data comes in all forms and shape and there is always chance of new type of case coming up. So general tabular format of data may not work Also, variables needs to be taken in totality, few variables cannot give reliable answers. So the quantum of data especially breadth of data is huge. That needs models\/algorithms to be flexible enough to incorporate new information and infer accordingly Data and medical may not always point to the same answer, and in those cases medical wins; so basic clinical knowledge needs to be developed. And there needs to be enough checks and balances to stop suggesting blunders. Since things get outdated fast in healthcare, one has to keep pace with the changes. by updated models periodically or incorporating machine learning early Must-have skills required to make the cut In this data-intensive field, data comes in from various sources, (such as clinical and patient data, HIS, ERP, LIS, CRM, PIS, and other tech systems). What kind of data gets high priority? According to Kumar, in healthcare, analysts have to work a lot with textual data of medicine, diagnosis, complaints. Here’s how you can make a head-start by beefing up on these core areas: Textual analysis gets high priority and it is always good to have domain knowledge There is not a set coding language for this field but you need a language that will give flexibility in playing around with large and unstructured data. Python does that task pretty well For many medical problems programming language, R already has readymade libraries and it can help start things fast. If either data size is large or complicatedly structured, then you can explore big data solutions For journal \/ article statistics knowledge is a must Basic medical knowledge is preferred at least around problem statement Structuring data in usable format will require problem solving skill and open mind Best Practices to keep in mind when working as healthcare analyst: Defining the problem statement precisely and craft a solution accordingly. Generic solutions normally don’t work in healthcare as margin of error is very small. Checking the accuracy and knowing the sources of data is very important. You should know which data is machine\/system generated and which is manually entered (or say where data accuracy is high versus low) Privacy of data is of utmost importance. When doing analytics make sure never to use any Personal Identifiable Information (PII) of patient. Emerging Job roles in Healthcare analytics Clinical Data Analyst: Job roles spans abstracting and verification of data sets and performing analysis aimed at enhancing patient outcomes and clinical quality. Leading and educating the team about data collection and finding opportunities for data improvement is also part of the brief. Healthcare Business Analyst: Job responsibilities span performing business analysis, clinical decision support and carrying out client interaction. The ideal candidate must have familiarity with medical terminologies and excellent problem solving skills. For the role of senior business analyst, proficiency in SAS, SQL, Hadoop and Hive is a must. Job responsibilities include using statistical techniques in customer segmentation and profiles, creating and reviewing models and providing technical leadership Salary While there is no definitive figure on salary in healthcare vertical, a quick glance at job board reveals that fresher with no industry experience get an overall package starting at INR 6 lakh, the figure goes up to INR 10 lakh (2+ years’ experience) and INR 15 lakh (5 years onwards). Top Companies hiring Healthcare Analysts McKinsey & Company: The global management consulting company is always on a lookout for analytical minds who are passionate about big data and are proficient in BI and ETL tools such as Jira, GIT, Ambari and have the knowhow of writing big data queries. Accenture: This global software giant collaborates with several healthcare partners and is one of the best places to kick-start your analytics career. Perks include deploying the best state-of-the-art data driven predictive models to tackle business problems, teaming up with some of the best global clients in delivering competitive advantage. Philips: One of the leading healthcare vendors, Philips is on a lookout for junior and senior data analysts who have the knowhow of managing analytical projects and can deftly manage high volume, complex data. Indian Startups in Healthcare sector Artivatic Data Labs Private Limited, is combining AI with genomic science to bolster the enterprises to tap into unlimited possibilities. According to the startup, the technology will revolutionize healthcare and will fuel preventive and predictive healthcare. THB: Another startup driving clinical research backed by analytics is THB. The startup provides operational analytics tool and leverages clinical analytics for smart patient engagement.","excerpt":"The Indian healthcare system in India is poised for a data-driven transformation, delivering value based care backed by a well-defined digital and analytics strategy. One of the most data-intensive, data rich fields, data flows in from disparate sources, posing several challenges in data aggregation and data governance policies. The data deluge apart, healthcare sector is […]","categories":[],"tags":["healthcare analytics india"],"author_name":"Richa Bhatia","publish_date":"2017-03-07T07:10:31","publication_year":"2017","word_count":1209,"keywords":["Go","machine learning","AI","RAG","Python","Ray","Aim","analytics","SQL","healthcare analytics india","R"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","Ray","RAG","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/healthcare-analytics-india-poised-change-see-can-make-healthcare-analytics-career\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10909,"title":"10 Chatbots from India making it big!","content":"Chatbots have taken the position to be your very personal chat assistants, whether as a tutor or a fashion counsellor, thanks to the ever-evolving technologies. Though bots have been around for quite a while, the recent surge in artificial intelligence and natural language processing has made it popular like never before. And with the ever increasing numbers of smartphone users, accessing chatbots have become both fun and easy. Posing many competitive advantages over apps, these chatbots require no downloads, can be instantly used, it is easier to build and upgrade and most importantly faster and cheaper than apps. While on one hand it has many advantages, on other hand its effectiveness in conversing with a human without a human intervention is often questioned. There certainly are questions that bots might not handle well, but the booming artificial intelligence gives a hope on the lacunas that it can overcome. With its own plus and minus, let’s read on to the most popular chatbots making headlines from India that have been automating the tasks you otherwise would have spent hours in doing. (Listed in Alphabetical order) AISHA by Micromax This voice assistant by Micromax is our very own Indianized version of Siri. It can perform tasks like initiating a google search, giving movie reviews, make calls, read news, view stalk market details and much more. Artificial Intelligence Speech Handset Assistant is all about AI in its backend that makes it one of the most popular bots in India. Engazify Bot Founded by Purva Surse and Siddharth Shekhawat, Engazify Bot is a faster and better way to appreciate your teammates, capture all your team wins, and save it for everyone to see. It fosters a culture of peer recognition by appreciating their work and celebrating the same by gamifying the whole experience. It lets you chat with him, ask questions or give feedback just like any other friend in your team. GoHero.ai Founded by Snehal Dhruve and Bineet Desai, this AI powered personal travel agent is available across nine platforms such as Facebook Messenger, Telegram, Skype. It assists you in booking flight, hotel, taxis etc. It integrates with messaging apps to use sophisticated algorithms and understand traveller’s preference. The chat screen allows the users can review pricing and availability of tickets. Gupshup One of the most advanced bots and messaging platforms, it enables developers to quickly and easily build, test, deploy and manage chatbots across all messaging channels. It offers application programming interfaces, with which software programs can interact with each other for all the popular messaging platforms, including SMS, Facebook Messenger, Slack, Telegram and Teamchat. ixibaba by ixigo A chatbot that can answer all your travel queries, brought to you by ixigo is the perfect example of how companies are improving on interacting with customers and enhancing customer experience. This AI bot can provide quick and quirky answers to all travel related queries much faster and instantaneously than a human could. Travellers can talk about flights, hotels, cabs, vacation destination etc. It also provides real-time information about weather, travel routes, flight timings etc. MagicX One of the first bots on the facebook messenger, MagicX lets you carry day-to-day tasks like bill payments, recharge, food ordering, flight booking and other services via chat, enabled off course by AI. It has a tendency of learning more from human interaction, giving a sense of more human like responses, but highly scalable at the same time. It has tie-ups with more than 20 e-commerce, recharge, grocery, travel and bill payment platforms. Niki.ai Whether a pre-paid, post-paid or DTH recharge, paying electricity bills, booking a cab, getting a laundry or ordering a burger from burger king, this chatbot assists you in all of it. Celebrating its inception in the year 2015, Niki has attracted funding from the likes of Tata and Ronnie Screwvala. This fully automated chatbot works up on the concept of artificial intelligence with no human intervention. Recharge Bot by Payjo Payjo is India’s first messenger bot that lets you recharge your phone, sets reminder for recharge and shows suitable plans for your number. In short, Recharge Bot supports intelligent balance reminders for prepaid phones. The biggest advantage of this chatbot is its availability languages other than English. It currently supports hindi, tamil, telegu and kannada. Skedool Launched by Deepti Yenireddy and Naveen Varma Alluri, it aims at automating repetitive everyday tasks for business executives, sales and recruiting professionals. A blend of artificial and human intelligence, it handles your B2B scheduling and calendar management. It has raised substantial amount of funding from prominent Silicon Valley and Indian technology investors including Kludein LLC, Narayan Ramachandran and Pranav Pai, investing on behalf of Mohandas Pai among others. The AI assistant is named as Alex and uses natural language processing and machine learning supervised by humans to enable customers to communicate with the service via e­mail just as they would with a human executive assistant. Yana With the founders claiming this chatbot to be an 100% automated artificial intelligence engine with no human interference, YANA help users to book cabs, order groceries, etc. The chatbot’s intelligence has been developed by it’s more than 50,000 messages that it has received over time.","excerpt":"Chatbots have taken the position to be your very personal chat assistants, whether as a tutor or a fashion counsellor, thanks to the ever-evolving technologies. Though bots have been around for quite a while, the recent surge in artificial intelligence and natural language processing has made it popular like never before. And with the ever […]","categories":["AI Features"],"tags":["AI India","chatbots india","ixigo"],"author_name":"Srishti Deoras","publish_date":"2016-10-10T10:38:33","publication_year":"2016","word_count":863,"keywords":["Go","funding","artificial intelligence","machine learning","AI","chatbots","chatbots india","Scala","ixigo","Ray","Aim","AI India","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Aim","Ray","chatbots","R","Go","Scala","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/10-chatbots-india-making-big\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10010947,"title":"AMD Buys Rival Xilinx For $35 B, Troubles Mount For Intel","content":"On Tuesday, AMD announced that it would be acquiring Xilinx in an all-stock transaction valued at $35 billion. Advanced Micro Devices(AMD) specialises in making chips for personal computers, especially gaming PCs. Whereas, Xilinx’s highly-flexible programmable silicon, enabled by a suite of advanced software and tools is designed to drive rapid innovation across a variety of industries and technologies. With this purchase, AMD looks to diversify its market share and customer reach, which would definitely send warning signals to its competitors like Intel. What Does This Deal Mean For AMD AMD is known for offering industry’s strongest portfolio of high-performance processor technologies, combining CPUs, GPUs, FPGAs, Adaptive SoCs and deep software expertise to enable leadership computing platforms for cloud, edge and end devices. With Xilinx on its side, AMD now stands a chance to gain a strong foothold in areas like telecommunications infrastructure and defence which are dominated by other players. As reported by WSJ, Xilinx chips are also used by Lockheed Martin Corp. for their F-35 Joint Strike Fighter. They are also commonly used in superfast 5G network infrastructure. Talking about the deal, AMD’s CEO, Lisa Su, said that the Xilinx team is a crucial player in the chip industry. “By combining our world-class teams and deep domain expertise, we will create an industry leader with the vision, talent and scale to define the future of high-performance computing,” said Dr Lisa Su. With a combined team of 13,000 talented engineers and over $2.7 billion of annual R&D investment, AMD will have additional talent and scale to deliver an even stronger set of products and domain-specific solutions. According to Victor Peng, CEO of Xilinx, this deal would lead the new era of high performance and adaptive computing. Xilinx’s leading FPGAs, Adaptive System on Chips, accelerators and SmartNIC solutions enable innovation from the cloud to the edge and end devices. “Joining together with AMD will help accelerate growth in our data centre business and enable us to pursue a broader customer base across more markets,” added Peng. With a spectacular lineup of products like Ryzen, EPYC, Radeon RX Vega, combined with the solutions of Xilinx, the future of AMD sure looks bright. More Trouble For Intel? Source: Deloitte The forecasts say that the semiconductor industry is going to have a great run thanks to the deep learning advancement. But, when it comes to hardware for AI, NVIDIA and Intel have been leading so far. By pocketing Xilinx, AMD becomes another hurdle for Intel along with NVIDIA. The integrated solutions these chip makers offer power data centres, work stations, edge devices and many more. Though Intel’s CEO Bob Swan is positive about Intel’s future, the events leading up to October weren’t that delightful for Intel. For Intel, the year 2020 has been a roller coaster. The company saw more lows than highs. It was Apple who delivered the first bad news of the year with their decision to go indigenous with their desktops by parting ways with Intel. This was followed by NVIDIA’s acquisition of ARM, the UK chip maker whose architecture powers more 90% of the devices across the world. (Source: Analytics India Magazine) Intel cheered its customers by rebranding itself and launching a brand new series of processors which go by the name– Tiger Lake. Leveraging Intel’s new SuperFin process technology, 11th Gen Intel Core processors optimize power efficiency with leading performance and responsiveness while running at significantly higher frequencies versus prior generations. AMD has been competing with Intel Corp for many decades now. Today, the competition has ventured into the lucrative data centres and AI domain. Intel’s chips power the majority of the data centres across the world. Things took a turn when last year, AMD announced that Google would be leveraging the power of AMD. Intel which has been supplying chips for data centres that power internet-based services. For example, AMD’s newest generation of server chip, called EPYC, uses a new chip-making technology from its contract manufacturers that helps the chips have better performance while consuming less power. Despite all the shopping spree in the chip industry, Intel is still ahead of AMD by a huge margin. For instance, Intel made $71.9 billion of revenues last year, compared with AMD’s $6.7 billion. Intel understands its position and the need for shedding off its legacy to move forward. In an attempt to do a makeover, Intel sold off its NAND memory business to SK Hynix, a South Korean semiconductor company. For Intel, going forward, it might still have to make more such tough decisions while backing it up with innovation in technologies that power up AI, 5G and more.","excerpt":"On Tuesday, AMD announced that it would be acquiring Xilinx in an all-stock transaction valued at $35 billion. Advanced Micro Devices(AMD) specialises in making chips for personal computers, especially gaming PCs. Whereas, Xilinx’s highly-flexible programmable silicon, enabled by a suite of advanced software and tools is designed to drive rapid innovation across a variety of […]","categories":["Global Tech"],"tags":["AMD","intel server","Mergers and Acquisitions"],"author_name":"Ram Sagar","publish_date":"2020-10-30T12:00:43","publication_year":"2020","word_count":769,"keywords":["Go","AMD","API","intel server","AI","programming_languages:R","innovation","programming_languages:Go","RAG","deep learning","analytics","Mergers and Acquisitions","R"],"extracted_tech_keywords":["AI","deep learning","analytics","RAG","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/amd-buys-xilinx-chip-intel-impact\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":25679,"title":"ISRO realigns the course of Mangalyaan to prevent any mishap during the eclipse","content":"The Mars Orbiter, or Mangalyaan is one of the most important projects undertaken by ISRO and was kick-started in September 2014 with the primary objective being to study Mars from several angles and relay information and photographs obtained back to ISRO. The project, which was originally scheduled for six months is still ongoing and  sending information for two years now. ISRO has recently been in the news for realigning the satellite’s orbit to prevent communication break in case of a long-duration eclipse. The plans for realigning was announced last year in August to prevent the satellite’s battery capacity from getting exhausted. A S Kumar, Chairman, ISRO notes, “The duration of eclipse was reduced owing to the satellite’s change of orbit on evening of January 17.” The spacecraft has displayed splendid performance and efficiency so far. The satellite has only burnt 7 kg of fuel in the previous year, and is still left with 30 kg of fuel. Moreover, the mission for the Mars Orbiter was undertaken within a budget of $73 million, which is a significantly cost-considerate, compared to earlier projects of similar gravity. The satellite is identified as one of the thinnest and lightweight, weighing at only 1,350 kg. It sports a high-resolution camera which clicks full-disk color images of mars. The Mangalyaan project has furnished ISRO with some insightful pictures of Mars. The project has not only helped India, but also enabled NASA to complete several research programs based on the useful information collected. Besides the realignment, ISRO divulged information regarding works on second lunar mission – Chandrayaan II. The project is scheduled for 2018, and scientists are currently testing soft landing engines for the same. Key achievements of the Mars Orbiter: Proof about watery exercises among the old Martian atmosphere Understanding about Mars’ ice spread changes amid summer on the Northern Hemisphere Evaluation of the layers of dust on Martian valleys and slopes that can reach up to 1.5 kilometers Scrutiny of two Martian satellites like Phobos and Deimos. Focus Keyword: ISRO Mangalyaan realignment Tags: ISRO India, Mars satellite India, Mangalyaan India, Mangalyaan realigning India","excerpt":"The Mars Orbiter, or Mangalyaan is one of the most important projects undertaken by ISRO and was kick-started in September 2014 with the primary objective being to study Mars from several angles and relay information and photographs obtained back to ISRO. The project, which was originally scheduled for six months is still ongoing and  sending […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2017-01-20T10:21:52","publication_year":"2017","word_count":348,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Ray","ViT","CLIP","R"],"extracted_tech_keywords":["AI","Ray","R","Go","CLIP","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/isro-realigns-course-mangalyaan-prevent-mishap-eclipse\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":44686,"title":"Beginners Guide To Data Visualisation With Matplotlib","content":"When it comes to Python and its visualisation capabilities, Matplotlib is undoubtedly the mother of all visualisation libraries. Matplotlib is a very popular library that has revolutionised the concept of making impressive plots with Python effortlessly. In our previous articles, we introduced you to some of the most popular plotting libraries such as Pandas plots, Seaborn, Plotly and Cufflinks. Most of these are built on top of Matplotlib which makes it an important library to know about. In this article, we will introduce you to Matplotlib and will take you through a hands-on session to plot beautiful visualisations. Plotting With Matplotlib Matplotlib supports a wide variety of plots from the basic line and scatter plots to advanced multi-dimensional plots. We will start with basic plots and will discuss some of the best practices to make an attractive and intuitive visualisation. Installing Matplotlib Use the pip installer to install Matplotlib into your working environment. Type and execute the following command in your terminal. pip install matplotlib If you are using Anaconda distribution use conda install matplotlib to install the library. Let’s make some plots! Importing the libraries import pandas as pd import matplotlib.pyplot as plt %matplotlib inline The %matplotlib inline function allows for the plots to be visible when using Jupyter Notebook. Importing the dataset data = pd.read_csv(\"sample_data.csv\") Here we will use a simple data set made of random numbers. This is what the data looks like. Quick Plots The Matplotlib enables us to plot to functional plots with ease. These plots are on-the-go plots that helps us visualise data in a quick and effortless way by calling the plot method and passing the axes as arguments. Simple Plot Let us plot a simple line plot to depict how the value of A changes for each observation in the dataset. plt.plot(data['A'], c = 'b') plt.title(\"Change In A\") plt.xlabel(\"Index\") plt.ylabel(\"A\") plt.legend() In the above code block, we pass an indexed data (data[‘A’]) to the plot method of the Matplotlib object. The c argument sets the colour of the plot. In the above case c=’b’ sets the colour of the plot to blue. The title method sets the passed string as the main title of the plot. The xlabel and ylabel methods label the x-axis and y-axis respectively. The legend method displays the plot legends. Output: Scatter Plot Lets us scatter plot the values of columns A and B against each other. plt.scatter(x = data['A'],y = data['B'],s = 500, c = 'r', marker = \"*\",alpha = 1, linewidths = 1, edgecolors='b') plt.title(\"Scattered A vs B\") plt.xlabel(\"A\") plt.ylabel(\"B\") The Scatter method in the above code block specifies that the plots are scattered. x = data[‘A’] sets the x-axis with values from the feature A of the dataset y = data[‘B’] sets the y-axis with values from the feature B of the dataset marker = “*” sets the plot symbol as *. For the list of all available markers click here. s = 500 sets the size of the marker symbol in the plot to 500 linewidths = 1 sets the border width of the marker symbol used edgecolors=’b’ sets the colour of the marker symbol edge to blue Matplotlib supports all HTML colour codes which can be passed in as arguments. You can get all HTML colour codes here. Copy the colour code of your choice and prepend it with a ‘#’ symbol and set it to the colour parameter. Output: Pie Chart plt.pie(x = [10,10,30,25,5,20], explode = [0.1,0.1,0.1,0.1,0.1,0.1], labels = ['A','B','C','D','E','F'], radius=2, autopct='%.01f%%',shadow=True, textprops={'size': 'smaller'}) The above code block produces a pie chart for the values passed in as x. explode = [0.1,0.1,0.1,0.1,0.1,0.1] spaces each block or wedge at 0.1 units away. labels = [‘A’,’B’,’C’,’D’,’E’,’F’] sets the label for each wedge. radius=2 sets the radius of the pie chart to 2. autopct=’%.01f%%’ displays the x values in percentage inside the pie chart. shadow=True enables shadow for the plot. textprops={‘size’: ‘smaller’} sets the size of the text Output: Creating Sophisticated Plots With Objects Matplotlib also has object-based plotting which adds flexibility to its plots. Let’s look at some examples. Subplots Initializing an empty figure object fig = plt.figure(tight_layout=True) Initializing a 2×2 grid of Axes inside the figure object fig, axes = plt.subplots(2, 2) Setting a title fig.suptitle('Change In A,B & C ') Plotting Index vs A in subplot[0,0]– 0th row and 0th column axes[0,0].plot(data['A']) Plotting Index vs B in subplot[0,1]– 0th row and 1st column axes[0,1].plot(data['B']) Plotting Index vs C in subplot[1,0]– !st row and 0th column axes[1,0].plot(data['C']) Plotting D vs A for A values greater than 0 in subplot[1,1] axes[1,1].bar(data[data['A']>0]['D'], data[data['A']>0]['A']) Output: Let’s look at another one: Import additional library import matplotlib.gridspec as gridspec Initialize the figure object fig = plt.figure(tight_layout=True) specify the geometry of the grid to place the subplots(The number of rows and number of columns of the grid need to be set.) gs = gridspec.GridSpec(2, 2) Plotting in Subplot1 ax = fig.add_subplot(gs[0, :]) ax.scatter(x = data['A'], y = data['B'], c='r') ax.set_ylabel('B') ax.set_xlabel('A') Plotting in Subplot2 ax = fig.add_subplot(gs[1, 0]) ax.plot(data['A']) ax.set_ylabel('A') ax.set_xlabel('Index') fig.align_labels() Plotting in Subplot3 ax = fig.add_subplot(gs[1, 1]) ax.plot(data['B']) ax.set_ylabel('B') ax.set_xlabel('Index') fig.align_labels() Output: Improving The Plots Drawing Axis Lines #Scatter plot for A vs B plt.scatter(data['A'],data['B'],s = 200, c = '#DAFF33', marker = \"*\",alpha = 0.5, linewidths = 3, edgecolors='#0C96F0',zorder=1) #Scatter plot for A vs C plt.scatter(data['A'],data['C'],s = 100, c = '#F03C0C', marker = \"X\",alpha = 1, edgecolors='#0CF05F',zorder=2) #Display Legends plt.legend(loc = 'lower right') #Draw axis lines plt.axhline(0.5,ls = '--' ) #horizontal line plt.axvline(0,ls = '--') #vertical line The first code block plots two scatter plots on the same graph which are differentiated by two different marker symbols and colours. The axhline and axvline methods allow us to draw horizontal and vertical lines from the axes respectively at the specified value\/constant. ls = ‘–’ sets the line style to dashed lines zorder = 1 sets the current plot\/layer in the background. loc = ‘lower right’ in legend method relocates the legends to the lower right corner of the graph. Output: Colour Shading #horizontal shading plt.axhspan(0.5, 2, alpha = 0.3, color = 'r') #vertical shading plt.axvspan(0, 2, alpha = 0.2) The axhspan and axvspan methods allow us to make horizontal and vertical shades from the axes respectively at the specified range. Appending the above code block to the previous section code will produce the following output. Closing Note Matplotlib is an essential package that allows users to make visualisations with less effort. Many modern data visualisation libraries are built on top of Matplotlib and have similar methods and API calls for visualising with various kinds of plots.","excerpt":"When it comes to Python and its visualisation capabilities, Matplotlib is undoubtedly the mother of all visualisation libraries. Matplotlib is a very popular library that has revolutionised the concept of making impressive plots with Python effortlessly. In our previous articles, we introduced you to some of the most popular plotting libraries such as Pandas plots, […]","categories":["Deep Tech"],"tags":["data visualization","Matplotlib"],"author_name":"Amal Nair","publish_date":"2019-08-20T14:00:09","publication_year":"2019","word_count":1088,"keywords":["Plotly","TPU","AI","ML","Python","Seaborn","Jupyter","data visualization","Matplotlib","R","Pandas"],"extracted_tech_keywords":["AI","ML","Jupyter","Pandas","Matplotlib","Plotly","Seaborn","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/beginners-guide-to-data-visualisation-with-matplotlib\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10082053,"title":"CEAT is certified as a Best Firm For Data Scientists","content":"CEAT is certified as the Best Firm For Data Scientists to work for by Analytics India Magazine (AIM) through its workplace recognition programme. The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company cultures. AIM analyses the survey data to gauge the employees’ approval ratings and uncover actionable insights. ‘With the advent of Industry 4.0, we intend to be market leaders and thought leaders in the space of tyre manufacturing. From revamping our manufacturing facilities to follow smart processes to upskilling our employees through Digital Academy, we ensure right talent in the right place at the right time is equipped with right resources. This certification recognises our efforts to create a data-driven culture that is adept at meeting changing consumer expectations and streamlining business processes’ said Milind Apte, Senior Vice President, Human Resource at CEAT Limited. The analytics industry at AIM faces a talent crunch, and attracting good employees is one of the most pressing challenges that enterprises are facing. The certification by Analytics India Magazine is considered a ‘Gold Standard’ in identifying the best data science workplaces and companies participate in the programme to increase brand awareness and attract talent. Best Firms For Data Scientists is the biggest data science workplace recognition programme in India. To nominate your organisation for the certification, please fill out the form at this link.","excerpt":"The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company culture.","categories":["AI Highlights"],"tags":["best companies for data scientists in india","best companies in india for data science fresher","best companies to work for as data scientist in india"],"author_name":"AIM Media House","publish_date":"2022-12-12T15:00:00","publication_year":"2022","word_count":234,"keywords":["best companies for data scientists in india","data science","Go","best companies to work for as data scientist in india","AI","data-driven","ML","Git","best companies in india for data science fresher","Aim","analytics","GAN","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","R","Go","Git","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/ceat-is-certified-as-a-best-firm-for-data-scientists\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":5925,"title":"TO THE NEW| A Global\/ India Social Media Analytics Report on the Football World Cup 2014","content":"With the football fever gripping the world, football lovers across the globe are not only catching the excitement on the television sets but also sharing their sentiments on social media platforms. The digital media has been buzzing with fans across age groups eagerly discussing the footballers, teams, performances, goals and key incidents of the match. TO THE NEW, Asia’s leading digital solutions network, has released a report that analyzes digital conversations across various social platforms such as Twitter, forums and news sites that took place from June 1- June 23, 2014. The report, powered by ThoughtBuzz, the social media analytics’ arm of the company, gives a comparative analysis of the social media buzz generated both globally and in India. The report that gives a demographic split of the netizens both in India and the world highlighted that globally the youth primarily below the age of 25 were the most active on social media contributing to 44% of the discussions while netizens between the age groups 25 and 40 contributed to 32% of the discussions.  Netizens above 40 years were more interested in the players historical and match historical data contributing to just 19% of the discussions. Unlike their global counterparts, Indian football fans across age-groups were actively discussing about the Football World Cup. While the youth contributed to 36% of the discussions, netizens between the age groups 25 and 40 contributed to 41% of the discussions while the netizens above 40 years were actively discussing about the winners, team goals and matches contributing to 23% of the discussions. The report also highlights that women are gaining interest in the game and are also engaging in digital conversations this Football World Cup. Out of the total social media discussions on the Football World Cup globally, women contributed to 16% while men contributed 84% of the discussions. The Indian women weren’t far behind their global counterparts contributing to 11% of the discussions while men contributed to 89% of the discussion. About TO THE NEW: Founded in 2010, TO THE NEW is Asia’s leading specialist in SMACK Services – Social, Mobile Analytics, Content and Knowledge and delivers integrated digital solutions to help clients connect with social and mobile first consumers and drive tangible business performance. TO THE NEW’s SMACK effect is backed by its ‘Centers of Excellence’ in Analytics led by ThoughtBuzz, Digital Marketing by Ignitee, Social, Local and Mobile by Techsailor, Content by Tangerine Digital and Technology by IntelliGrape. TO THE NEW employs over 600 people across Singapore, Manila, Kuala Lumpur, Jakarta, Mumbai, Delhi, Chennai, Dubai and Guangzhou. For more information, visit: www.tothenew.com Download the full report below- [attachments include=”5928″]","excerpt":"With the football fever gripping the world, football lovers across the globe are not only catching the excitement on the television sets but also sharing their sentiments on social media platforms. The digital media has been buzzing with fans across age groups eagerly discussing the footballers, teams, performances, goals and key incidents of the […]","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2014-07-10T17:47:53","publication_year":"2014","word_count":438,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","Git","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Git","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/to-the-new-a-global-india-social-media-analytics-report-on-the-football-world-cup-2014\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":37933,"title":"IIM Indore To Launch A 10-Month Integrated Programme In Business Analytics In Collaboration With Jigsaw Academy","content":"Today, it is estimated that a whopping 50,000 machine learning and data science jobs are unfilled with another 40,000 new jobs being added by 2020 – just in India alone. Due to a constant demand and an ever-increasing focus on developing new insights and understanding of how data is used to power business decisions, the renowned business school IIM Indore is all set to launch a 10-month Integrated Program in Business Analytics (IPBA). “We are delighted to launch this certificate program on Business Analytics with Jigsaw. In line with our mission to be contextually relevant, the Integrated Program in Business Analytics (IPBA) is designed to provide an in-depth and practical working knowledge of analytics and its applications in solving real life business problems to the working professionals. The participants will learn how to use statistical models, machine learning, text mining and big data techniques for modelling real world and making sound data-driven business decisions. My eminent colleagues from IIM Indore and the experienced professionals from Jigsaw Academy associated with this program are truly world class with vast experience in the business analytics domain,” said Prof. Himanshu Rai, Director, IIM Indore. The IPBA, directed by IIM Indore and powered by Jigsaw Academy, India’s top ranked institute for analytics and data science, blends traditional management principles with cutting-edge technology coverage to accelerate your career in analytics. This specialised program will help students build expertise in Business Analytics, Data Science and Machine Learning. For professionals who are interested in building careers in data science with a strong background and context in business, the IPBA will equip them with a fundamental understanding of both. They will also get a great opportunity to learn from award-winning faculty from IIM Indore and Jigsaw Academy, interact with business leaders and benefit from the extensive IIM network and work with complex data sets using R and Python. Students will also have access to Tableau’s cutting-edge visual analytics software for free while enrolled in the program for study and practice purposes, through Tableau’s global Academic Program. The program comprises of 120 hours of training via live online classes, online assignment-solving and Q&A sessions that can be accessed from anywhere in India or the world. There is also a separate in-person ‘immersion’ component of 30+ hours of classes, networking and project presentations that will be held on the IIM Indore campus. “We are now at a cusp where the workforce leading businesses are moving towards decoding data and understanding analytics. Jigsaw has always been fully committed to supporting the ambitions of those interested in upskilling themselves to stay on top of their game. We are very excited and are looking forward to working with IIM Indore on the IPBA,” says Gaurav Vohra, CEO, Jigsaw Academy.","excerpt":"Today, it is estimated that a whopping 50,000 machine learning and data science jobs are unfilled with another 40,000 new jobs being added by 2020 – just in India alone. Due to a constant demand and an ever-increasing focus on developing new insights and understanding of how data is used to power business decisions, the […]","categories":["AI Trends"],"tags":["Business Analytics","jigsaw academy"],"author_name":"Richa Bhatia","publish_date":"2019-04-18T05:06:07","publication_year":"2019","word_count":455,"keywords":["big data","data science","machine learning","AI","data-driven","RAG","Python","analytics","Business Analytics","programming_languages:Python","jigsaw academy","R"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","RAG","Python","R","big data","data-driven","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/iim-indore-to-launch-a-10-month-integrated-programme-in-business-analytics-in-collaboration-with-jigsaw-academy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10095109,"title":"Apple Launches 15-inch Macbook Air Powered by M2 Processor","content":"Apple Macbook Air just got bigger. The tech giant today announced the launch of its most awaited 15.3-inch Macbook Air, powered by M2 processor. While the new Macbook Air offers a stunning 15.3-inch Liquid Retina display, a good battery life, a super thin and light design, MagSafe charging, more speakers, and a great camera, it lacks a cooling fan, and comes with only two USB-C ports, showcasing its ability to drive multiple external displays. Click here to order. The new 15-inch MacBook Air is available today! We think you’re going to love it! pic.twitter.com\/AWUyuf7ZTY— Tim Cook (@tim_cook) June 13, 2023 The launch of the new Macbook Air was first announced at WWDC 2023. Features & Specifications Set in the middle, between the 13 and 16-inch models, the apt size is reported to make a huge difference in the positive direction by most reviewers. It has a surprisingly slim and lightweight design, characteristic of the Air series. Starting at $1,299, it presents a much more budget-friendly option compared to any of Apple’s previous laptops featuring a screen of this size. Like the 13-inch, Macbook Air 15 possesses robust performance, extended battery life, and a high-quality display. The laptop also offers a maximum of 24 GB of RAM and 2 TB of internal storage available in Silver, Starlight, Space Grey, and Midnight color variants. Even with its larger size, the laptop doesn’t come with a fan to cool the M2 chip, but depends on its passive cooling solution to do its thing. The MacBook Air differs from its smaller counterpart in a few notable ways. The most apparent distinction is its 15.3-inch display, featuring a resolution of 2,880 x 1,864 pixels, resulting in a pixel density of 224 pixels per inch, matching that of the 13-inch Air. This display is classified as one of Apple’s “Liquid Retina” screens, offering a brightness of 500 nits, P3 “wide color” gamut support, and a refresh rate of 60Hz. The trackpad of the laptop has been stretched out keeping it proportional to the large size and Touch ID sensor are all excellent, which is true of all Mac laptops at this point. Boasting a 18-hour battery life, Apple claims that the Air is much faster, the screen is better, the battery is longer while the laptop itself is thinner and lighter. As mentioned earlier the M2 chip delivers on excellent performance of the laptop. Apple claims that the Air is much faster, the screen is better, the battery is longer while the laptop itself is thinner and lighter. The 15-inch Air has a six-speaker sound system with “force-canceling woofers” for improved bass, compared to a four-speaker setup in the smaller model. Apple has been making excellent laptop speakers in the last few models. The company said they “deliver[s] twice the bass depth of the 13-inch MacBook Air with M2 for fuller sound.” For most, the best feature is the affordable pricing beginning at $1,299 or Rs. 1,34,900. Most reviewers have said that the laptop is a larger version of apple 13 with the M2 chip like performance.","excerpt":"Starting at $1,299, it presents a much more budget-friendly option compared to any of Apple’s previous laptops featuring a screen of this size.","categories":["AI News"],"tags":[],"author_name":"K L Krithika","publish_date":"2023-06-14T13:59:44","publication_year":"2023","word_count":511,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/apple-launches-15-inch-macbook-air-powered-by-m2-processor\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10069317,"title":"Xceedance to hire over 500 people in India this year","content":"Xceedance recently announced plans to hire 500 additional skilled resources in India. The insurtech solutions and services provider has four offices in India and an annual staff growth rate of 30 per cent. With over 2,700 team members across the US, UK, Liechtenstein, Poland, India, and Australia, Xceedance is dedicated to developing accomplished professionals across the entire insurance value chain. With expertise in actuarial services, underwriting, policy management, claims, data sciences, and insurance technology, Xceedance helps to nurture the future of the global insurance ecosystem. “India is critical for our company’s success, and a key objective is to invest in the country’s diverse technology talent and knowledge workforce. Xceedance offices in India help support our right-shoring approach, delivering superior business value to clients worldwide,” said Amit Ranjan, CEO of Xceedance. “We seek talented individuals across multiple disciplines and departments, especially those with established insurance experience and backgrounds in varied fields such as technology, data management, statistics, and finance.” As the post-pandemic hybrid work model continues to evolve, the company’s approach toward people management will be critical in attracting and retaining professionals. “In marking the ninth anniversary of our company’s founding in India, we appreciate the resilience and dedication of our teams, which directly contribute to positioning Xceedance as a progressive, global provider of services and technology for insurance organisations. The skills and capabilities of a diverse workforce continue to be crucial to the evolution of Xceedance and to the company’s purpose in service to the insurance industry and our communities. The hiring plans are a step in the right direction for Xceedance and will contribute to accelerated growth over the next 3 to 5 years,” Ranjan added.","excerpt":"The company has four offices in India and an annual staff growth rate of 30 per cent.","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2022-06-17T12:59:23","publication_year":"2022","word_count":277,"keywords":["data science","programming_languages:R","AI","Aim","GAN","R"],"extracted_tech_keywords":["AI","data science","Aim","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/xceedance-to-hire-over-500-people-in-india-this-year\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":27343,"title":"When Deep Learning Models Have A Bird’s Eye View Of Humans","content":"Human features as objects of study have been widely used in various machine learning applications — be it face detection, video surveillance or even the development of autonomous cars. In fact, the task of ascertaining human features becomes the major work of ML systems in these applications. However, in the case of video surveillance, capturing human figures at an aerial level, especially from a moving equipment, becomes very challenging. Due to factors such as video equipment alignment or lighting, the accuracy of detection in the system takes a hit significantly. In order to resolve these issues, researchers are now exploring deep learning (DL) in video surveillance. Earlier studies usually relied on computer vision concepts such as local binary pattern (LBP) or feature transforms to extract features from images and later trained with classification algorithms. Subsequent studies used convolutional neural networks (CNN) to handle large image datasets. For example, the popular CNN model AlexNet has classified more than one million images. In this article, we will discuss the latest study on human detection in aerial videos which has used three different DL models. A Real-Time Approach Nouar AlDahoul, Aznul Qalid Md Sabri, and Ali Mohammed Mansoor, researchers from the University of Malaya, Kuala Lumpur, bring out a research study which explores DL for real-time human detection from aerial videos. They use three DL models: Supervised CNN model Pretrained CNN model Hierarchical extreme learning machine (HELM) model All of these models are trained and tested for the UCF-ARG aerial dataset. This dataset consists of images in the form of video frames which capture a set of human actions aerially. Actions include running, walking, throwing among others captured on a moving camera. In the first stage, the researchers build an optical flow model for all the three DL models so that inputs from the training and testing samples hold good for human detection as well as to achieve image stabilisation. For this purpose, video frames in the dataset are fed as input for this optical flow model. “Optical flow estimates the speed and direction of the motion vector between sequences of frames. This stage is important because it tackles the camera movement issue that results from the moving aerial platform. Feature images are produced by thresh-holding and performing a morphological operation (closing) to the motion vectors. Blob analysis is then performed to locate moving objects in each binary feature image. Next, green boundary boxes are overlaid surrounding the detected objects. The quality of optical flow for background stabilization is important as it is the first stage before feature learning is performed via deep models which act as input for the classifiers.” Once the optical flow model is set to stabilise the human actions, DL models are created for human detection. Deep Model Implementations Supervised CNN model: In this model, a supervised CNN is used for human detection. The CNN architecture has 17 network layers which includes one input layer, three convolutional layers, three max-pooling layers, six rectifier linear unit layers, and two fully connected layers with a soft-max layer. Features are now extracted in the eighth layer. Stochastic gradient descent (SGD) is the method used for training the model network. In addition to that, the soft-max layer segregates human and non-human objects in training. The CNN layers and the network architecture are presented as follows: Image input layer with 100 * 100 pixels grey image. Convolution layer: 20 feature maps of 5 * 5. Relu layer Max-pooling layer: pooling regions of size [2,2]  and returning the maximum of the four. Relu layer. Convolution layer: 20 feature maps of 5 * 5 Relu layer. Max-pooling layer: pooling regions of size [2,2]  and returning the maximum of the four. Relu layer. Convolution layer: 20 feature maps of 5 * 5. Relu layer Max-pooling layer: pooling regions of size [2,2] and returning the maximum of the four. Fully connected layer: with 1,000 nodes. Relu layer. Fully connected layer with 2 classes. Soft-max layer. Classification layer. Supervised CNN architecture (Courtesy: Nouar AlDahoul et.al) Pre-trained CNN model: This DL model considers the AlexNet architecture for detection. The networks include five convolutional layers, ReLU layers, max-pooling layers, three fully-connected layers, a soft-max layer, and a classification layer. An important consideration here is that the input RGB images are resampled to 2272273 pixels. Feature extraction is at the seventh layer by an activation function. Just like the previous model, SGD is used to train the model along with an SVM classifier to classify features into human and nonhuman classes. The network architecture is given below: Pretrained CNN architecture (Courtesy: Nouar AlDahoul et.al) Hierarchical Extreme Learning Machine(HELM) model: This model entails autoencoders which are based on the concepts of extreme learning machines. The architecture consists of three modules including two sparse ELM-based autoencoders, and an ELM-based classifier. Unlike the previous models, grey images are used as inputs in this model. Distinguishing human and non-human classes are done by the ELM-based classifier in the last module. The HELM model is given below: HELM architecture (Courtesy: Nouar AlDahoul et.al) Performance Of The Models All the three models are trained and tested extensively for the dataset. When it comes to performance, they offer a very high accuracy in human detection (close to 99 percent) along with speed (training times being 10 minutes and testing times as fast as 0.1 seconds). For implementation, NVIDIA GeForce GTX 950 was the GPU used in this study. Comparably, the HELM model fares better than the other two models. A detailed analysis can be found here. Conclusion With this study, detecting human figures has been made easier in an environment bound with non-human objects in addition to varying motions of these objects. The striking feature of this study is that all of the processes are happening in a real-time scenario. This will definitely be a big boost for critical areas such as video surveillance.","excerpt":"Human features as objects of study have been widely used in various machine learning applications — be it face detection, video surveillance or even the development of autonomous cars. In fact, the task of ascertaining human features becomes the major work of ML systems in these applications. However, in the case of video surveillance, capturing […]","categories":[],"tags":["Convolutional Neural Networks","Deep Learning","models","real-time"],"author_name":"Abhishek Sharma","publish_date":"2018-08-16T11:22:31","publication_year":"2018","word_count":975,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","models","ML","computer vision","Convolutional Neural Networks","deep learning","real-time","Deep Learning","CNN","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","R","Go","CNN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/when-deep-learning-models-have-a-birds-eye-view-of-humans\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10072574,"title":"JetBrains Introduces Fleet to Take on VSCodes of the Programming World","content":"Fleet, a JetBrains product, calls itself a high-performance IDE with quick load times, responsive performance, and a light-weight text editor. The build-up is to present itself as a competitor to Visual Studio Code (VSCode). JetBrains is a big player in the IDE market. IntelliJ IDEA and WebStorm were until now its main rivals for VSCode. But, IntelliJ IDEA and WebStorm had various downsides—not free, licence required, high boot-up times, non-intuitive for beginners and unable to run in browsers, among others—that couldn’t give a real challenge to VSCode. The Fleet IDE has been completely rebuilt. There are no UI components from the past. The company claims that they are reusing their code-processing engine, which distinguished them in the past and continues to distinguish them from others even now. The code-processing engine, which the company uses for code refactoring, is one of the reasons developers use IntelliJ Idea or Webstorm over VSCode. Fleet currently supports Java, Kotlin, Python, Go, Javascript, Typescript, Rust, and Json. Apart from this, PHP and C++, many more will also be available later. The tool automatically detects the language and provides support, maximising the value by making the code-processing engine available in all of them. The Fleet editor also enables natively connecting to a remote workstation where the code is stored. It is JetBrains’ take on GitHub’s Codespaces. The JetBrains team goes even a step further. It intends to perform some of the IDE Backend processes in the cloud in the future. This will transform any slow machine into a lightning-fast programming machine. Developers will no longer have to rely on the performance of their development computer to change or run code. Their machines will just be clients. “What we’re going to provide with Fleet is a different experience for those who sometimes just want an editor but also want a fully-fledged IDE. The platform is for those who want to use a single tool as opposed to specialized ones. And, of course for those certain scenarios that some of our existing IDEs may not cater for when it comes to distributed development,” said JetBrains in a statement. Meanwhile, VSCode is the biggest player in the segment with no rivals currently. VSCode: Developers’ first choice Launched in 2015, VSCode rose to popularity quickly and became one of the fifteen top development tools in a year. In 2016, it ranked 13th among the top popular development tools on Stackoverflow. The 2019 Developers Survey found that the VS Code editor quickly rose to the top, with 50% of the 87,317 participants using it. The popularity of the editor is based on various factors—free, open-source, and cross-platform. Besides, it offers various features which its competitors don’t. Unlike many other code editors, VSCode has an in-built debugger, which makes the development flow less ‘clicky’ and keeps the code and debugger in a single view. This makes issue tracking and code run-throughs easier and faster. You don’t need numerous displays to run the various consoles and rearrange them every time you need to minimise something. It’s incorporated into the design and the layout of your ideal workstation. There is intelliSense built into the code editor. It is a form of predictive coding. Along with framework, library, and language plugin extensions, developers can leverage predictive coding with ready-made boilerplate codes. While the debate—whether Fleet will clinch the throne from VSCode—is gaining momentum, users also have a completely different take on it. In 2021, when JetBrains announced the launch of the initial version of Fleet with limited access, there were tons of comments on the website page. A comment aptly summarised the needs of users. “Very interesting,” read one comment. “Do you plan on making it free (as in freedom) software? As far as I’m concerned, a proprietary editor is not going to compete with VSCode, regardless of how good it is. But if Fleet is going to be free that might make it a VSCode killer for me, more so if you bundle a free C# debugger.”","excerpt":"The company claims that they are reusing their code-processing engine, which distinguished them in the past and continues to distinguish them from others even now","categories":["AI Trends"],"tags":["JetBrains","Microsoft","VS Code"],"author_name":"Tausif Alam","publish_date":"2022-08-10T17:00:13","publication_year":"2022","word_count":663,"keywords":["Go","Rust","AI","TypeScript","RAG","Python","JetBrains","Aim","VS Code","JavaScript","R","Java","Microsoft"],"extracted_tech_keywords":["AI","Aim","RAG","Python","R","JavaScript","TypeScript","Go","Rust","Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/jetbrains-introduces-fleet-to-take-on-vscodes-of-the-programming-world\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10041705,"title":"The Machine Wears Prada","content":"From corsets to crop tops, the fashion industry has come a long way. A key element of fashion is how quick and competitive the industry needs to be. For instance, since the second world war, fashion was split into two seasons—spring and summer lines—and people eagerly awaited the arrival of each season’s lines. Today, Zara alone produces around 20,000 new styles in a single year. Keeping up with this is tricky, and it is no surprise that the fashion industry has been increasingly employing another field that thrives on innovation: technology. The use of technologies such as artificial intelligence and virtual reality has transformed fashion in various ways. Meet AI, your new designer The first role of AI is in fashion design. In 2016, Google tried this in its Project Muze, which it worked on in-partnership with Zalando, a German fashion platform, and Stinkdigital, a UK-based production company. Source: Google Project Muze involved a predictive design engine that trained a neural network to make sense of colours, textures, style preferences, among other parameters derived from Google’s Fashion Trend Report and data supplied by Zalando. The project built an algorithm to create designs based on individual interests and aligned them with style preferences compatible with the training data (from Google and Zalando). These human-free AI designs still require a lot more research. Many algorithms to generate clothing have largely backfired. Google’s Project Muze created many unwearable and crude designs. Still, AI-developed designs are getting better. In April 2019, an AI designer known as DeepVogue won the People’s Choice Award at a Fashion Design Innovation competition in China. AI has also been utilised to derive insights into fashion. In 2018, Tommy Hilfiger, IBM and the Fashion Institute of Technology partnered up for a project. The project, known as Reimagine Retail, used IBM AI tools to tell real-time industry trends, customer sentiment surrounding Tommy Hilfiger (and its products) and recurring themes in aesthetic parameters such as patterns, colours, and styles. The project then brought this knowledge to human designers who use it to make better and more informed decisions for Tommy Hilfiger’s following collection. Additionally, MAD street Den, a startup based in India and the US, developed Vue.ai, an AI fashion brand that offers end-to-end AI assistance in analysing user searches and providing suitable recommendations. It follows colour, patterns, necklines, sleeves, among other parameters, to help its algorithm. Another use of AI is to allow intelligent fitting techniques. This will enable shoppers to purchase clothes online without being afraid of how it fits. Fabulyst is one such AI-based platform from Hyderabad which helps shoppers find new products and recommends the best fit and styling based on their understanding of the user’s body measurements, body variation and skin tone. Fast fashion The term is used to describe clothes associated with mass production and is less expensive than ‘designer clothing’. As the name suggests, fast fashion requires retailers to work quicker than ever before. Today, these retailers look to AI and automation to aid in manufacturing. For example, AI can help give estimates on cost and production time. An exciting element of automation in fashion revolves around using robots in garment manufacturing. SoftWear Automation is developing ‘Sewbots’ with robotic arms, vacuum grippers and micromanipulators that can precisely work a cloth on a sewing machine. The tool utilises specialised cameras and computer vision software and can significantly reduce manufacturing costs and decrease errors. Retailers could also take advantage of this automation in countries with ageing populations and resulting labour shortages. Many have with fast fashion a vital concern with the questionable working conditions it provides its workers and the massive environmental damage rapid production poses. The Environmental Protection Agency reported that around 17 million tonnes of textile waste were dumped in landfills annually. Many solutions have been presented for this, including using plant-based or synthetic fibres (instead of environmentally damaging leather) and artificial intelligence. AI can enable better demand forecasting, which would allow retailers to use materials more efficiently and cut down on overproduction and waste. In 2018, H&M created an AI department to solve this problem. This department also used algorithms to automate their warehouses and make faster deliveries. Fashion, tech and the future Myntra, which is among India’s largest online fashion retailers, has also employed AI in various ways. Its fast fashion brands: Moda Rapido and Here&Now, offer computer-generated clothes such as jeans and t-shirts without the help of designers. Another essential technology leveraged by Myntra is augmented reality. Myntra added an AR feature to its app to help customers, which would allow them to use their smartphone’s camera and critique how they might look in certain clothing or accessories. Source: Myntra Virtual reality is also a crucial technology in fashion. In 2019, Tommy Hilfiger presented a new collection with actress Zendaya and those who could not witness the event first-hand were able to see it through a virtual reality-powered pop-up store. AR is also redefining fitting rooms. TopShop brought in-store AR mirrors, so customers did not have to try out their clothes physically. In 2017, Amazon applied for a patent (granted in January 2018) to create an AR mirror for users to try clothes virtually at home. https:\/\/twitter.com\/i\/status\/1112832526899396614 The use of technology has significantly affected the fashion industry. The latest tech innovations involve wearable technology that goes beyond Fitbit and Apple Watches. Since 2017, Levi’s has teamed up with Google’s Project Jacquard to create a line of intelligent denim jackets that can perform tasks like playing songs and declining calls. Hexoskin, a startup, developed a line of shirts that track movements, heart rate and breathing. Automation could be great for the fashion industry. Some people fear it may displace jobs, and while that is a valid concern, the increasing demand for rapidly-produced apparel makes it clear that the use of new technology in fashion is only going to grow.","excerpt":"From corsets to crop tops, the fashion industry has come a long way. A key element of fashion is how quick and competitive the industry needs to be. For instance, since the second world war, fashion was split into two seasons—spring and summer lines—and people eagerly awaited the arrival of each season’s lines. Today, Zara […]","categories":["IT Services"],"tags":["ai in fashion"],"author_name":"Mita Chaturvedi","publish_date":"2021-06-12T16:00:00","publication_year":"2021","word_count":976,"keywords":["Go","API","artificial intelligence","ai in fashion","AI","neural network","Git","computer vision","RAG","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","computer vision","RAG","R","Go","Git","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-machine-wears-prada\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10055982,"title":"Unisys Acquires CompuGain, Deepens Cloud Expertise","content":"Unisys Corporation has acquired CompuGain, a cloud solutions provider, for USD 87.3 million on a cash-free, debt-free basis. CompuGain’s cloud and hybrid cloud data management, capabilities in application modernization, and cloud-native agile application development will deepen Unisys’ cloud offerings. CompuGain is also designated as an Amazon Web Services (AWS) Advanced Consulting Partner. The company employs 400 engineers, cloud solution developers and architects. “Acquiring CompuGain will enable Unisys to enhance the delivery of rapid and agile cloud migration, application modernization and data value realization to our clients,” said Peter Altabef, Chair and CEO, Unisys, adding, “CompuGain has a strong presence in financial services, which will add to our established position in that industry. It will also create cross-selling opportunities with both Unisys clients and CompuGain clients across multiple industries.” CompuGain generated a revenue of USD 58 million for the 12-month period that ended September 30, 2021. CompuGain is growing at 12% year over year, and over 90% of its revenue is from client relationships, where an average tenure of a client is more than nine years. Information Services Group (ISG), a global technology research and advisory firm, in September 2021, in a Provider Lens report, recognized Unisys as a global leader for its cloud and infrastructure solutions. Unisys expects the transaction to be immediately accretive to adjusted EBITDA and free cash flow (FCF). “CompuGain’s innovative solutions, talent and agility, combined with Unisys’ cloud expertise and global reach, will allow us to deliver desired business outcomes at a much larger scale,” informed Debasish Hota, founder and CEO, CompuGain. CompuGain has an agile development process and customizable frameworks for both applications and Infrastructure as a Service (IaaS), which will help Unisys Deliver rapid and agile cloud migration. Unisys and CompuGain together will be able to address the need for cloud-native capabilities like micro-services and micro-application deployment. This will give clients the power to develop new software-based tools and services, which will help in automating processes and reducing operational costs. CompuGain is an AWS Advanced Consulting Partner, which means its competencies across public cloud platforms include data lakes, containers, serverless technologies, ML, DevSecOps, and application & data architecture for cloud environments.","excerpt":"With the acquisition of CompuGain, a designated AWS Advanced Consulting Partner, Unisys has expanded and deepened its cloud expertise.","categories":["AI News"],"tags":["AWS","Mergers and Acquisitions","Unisys"],"author_name":"Meeta Ramnani","publish_date":"2021-12-15T16:31:27","publication_year":"2021","word_count":357,"keywords":["API","AWS","AI","cloud_platforms:AWS","ML","serverless","RAG","cloud_platforms:Amazon Web Services","Unisys","Mergers and Acquisitions","R","data lake"],"extracted_tech_keywords":["AI","ML","RAG","AWS","serverless","R","API","data lake","cloud_platforms:AWS","cloud_platforms:Amazon Web Services"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/unisys-acquires-compugain-deepens-cloud-expertise\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23676,"title":"Does Deep Learning Represent A New Paradigm In Software Development?","content":"Do neural networks represent a shift in coding? Tesla’s director of AI Andrej Karpathy in his note on Software 2.0  has coined this new term which emphasises a shift in developing and writing software. His take on the question is that training neural nets and predicting using them involves a new way of thinking of software. The earlier methodology, according to Karpathy, involved writing code, say, in Python and putting it in production. However, when it comes to training neural nets — the developer has to set up a framework for the neural net to properly “learn” and repeatedly feed examples through it of what the correct answer is to “teach” it. Does Deep Learning Represent A New Software Paradigm? Most AI practitioners do agree neural nets represent a shift in coding, but it’s not a path-breaking transition and the term “software 2.0” has been drubbed as another marketing buzzword for neural networks. Here’s why – deep learning is the not the “real silver bullet” for all business problems. Karpathy’s presumption is that deep learning can do better than other forms of software and will continue to be used instead of handwritten code. On one hand, Karpathy reinforces the fact that the overarching success of deep learning at a family of tasks machine learning practitioners thought should be hand-written coding has paved for a new paradigm of producing software is –  Software 2.0. To a certain extent, neural networks represent a trend towards “teachable machines” which brings us back to the main argument – is deep learning chipping away at machine learning techniques. Neural networks have made significant progress in fields like computer vision, image classification, language translation and speech recognition, areas where traditional methods were underperforming. Conventional Programming vs Deep Learning Karpathy, a highly respected machine learning researcher is well-known for his excellent publication record and citation count. According to Carlos E Perez, author of Deep Learning Playbook & Artificial Intuition: The Improbable Deep Learning Revolution, enterprises are still in the early stages of Deep Learning development and the field is rife with issues that need to be addressed to evolve to a new kind of computing that reflects the same features as Software 1.0. Practitioners argue that in a lot of cases, simpler algorithms are performing just fine in certain cases and even offered key benefits like interpretability, a major advantage over Deep Learning’s black box problem. There’s more to software than just writing code. Software 1.0 involves new features, maintenance, finding bugs and bottlenecks and even long-term support is crucial. Meanwhile, neural networks don’t really address those problems. According to machine learning researchers, Software 1.0 comprises of bitwise operations — AND, OR, NOT that can be performed using only NAND gates. The complexity steps in by building functions over bytes, floating-point numbers, memory pointers, and from specialising and optimising instructions to specific tasks. Software 2.0 excels in a family of tasks that couldn’t be solved with conventional computing and that’s where it took off. However, to advance “software 2.0”, one needs to keep producing massive, clean datasets for networks to train. Neural network approach has not been extended to use cases where there is a lack of labeled data. Perez believes neural networks are intrinsically ‘intuition machines’ which means the technology is inherently different from software 1.0. Do neural networks offer an automated programming framework – no? Neural networks are also programs that require coding – if you tweak the dataset and change the training parameters, one obtains a different output. Once the error is minimised, you get a solution for your problem. In this case, the trained network is a deterministic function of the hyperparameters and the output and stopping criteria are well-defined. One Thing Is Clear, Deep Learning Is Indeed The Future Of Software Even though many deem Karpathy’s argument to be flawed, one thing is certain, he coined a new term for deep learning – “software 2.0”, adding more teeth to the DL hype and has hinted how firms like Google, Amazon, Uber are working on it and will soon make it the new paradigm of software development. Meanwhile, senior data scientist Seth Weidman has added a new perspective to it. According to Weidman, with the wide availability of easy-to-use packages such as SciKit Learn in Python, data scientists are spending more time on integrating data from various sources and rather than explicitly programming the models themselves. Weidman believes in the future, data scientists and machine learning researchers would spend more time setting up environments for teachable machines rather than explicit programming –  the code for actually generating the models is contained in the libraries, he adds. For example, Weidman explains that in the age of software 2.0, modeling tasks will not be about the designing custom functions, but more around function approximations wherein developers after feeding the right data will deploy off-the-shelf tools to tune the model. Weidman gives an image classification example, wherein to train an image classifier, the developer loads in the images, and then uses an off-the-shelf code from a library like Keras to show the model structure. By searching how to perform “image classifier keras”, a developer can train a model with less than 20 lines of code. Given this scenario, Weidman argues for budding data scientists entering the field, knowing how a model works would not be mission critical. The more important thing would be ensuring data quality and building the right checks for the model, he adds. Well, the use cases for software 2.0\/ deep learning are definitely expanding and over a period of time and there will be an increased prevalence of tech companies who would load off computations in their software to trained neural nets — this would represent the shift in software development.","excerpt":"Do neural networks represent a shift in coding? Tesla’s director of AI Andrej Karpathy in his note on Software 2.0  has coined this new term which emphasises a shift in developing and writing software. His take on the question is that training neural nets and predicting using them involves a new way of thinking of […]","categories":["IT Services"],"tags":["Software Development"],"author_name":"Richa Bhatia","publish_date":"2018-04-16T04:45:49","publication_year":"2018","word_count":959,"keywords":["Go","machine learning","Keras","TPU","AI","neural network","computer vision","Python","deep learning","Software Development","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","Keras","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/does-deep-learning-represent-a-new-paradigm-in-software-development\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10044432,"title":"Can Social Media Algorithms Ever Be Ethical?","content":"According to Datareportal, 4.33 billion people use social media, equating to 55 per cent of the total population on Earth. In the past 12 months alone, a total of 521 million new users joined social media at an annual growth of 13.7%–an average of 16 new users per second. “I feel tremendous guilt,” admitted Chamath Palihapitiya, former Vice President of User Growth at Facebook, to an audience of Stanford students in 2017. “The short-term, dopamine-driven feedback loops that we have created are destroying how society works.” Algorithms are geared towards maximising user attention. Behavioural design strategies use neurological and behavioural insights to develop customer interactions and influence user behaviour. “The success of an app is often measured by the extent to which it introduces a new habit,” said app developer Peter Mezyk in an interview with Business Insider. The more time a user spends on the app, the more advertising revenue flows into companies’ pockets; attention is currency. Source: Datareportal Youtube’s recommendation algorithm knows what people want so well that an average mobile viewing session lasts a good 60 minutes. The algorithm brings YouTube much of its revenue and is worth billions of dollars. CPO Neal Mohan said recommended videos account for 70 percent of the time spent by users on Youtube. The algorithm comprises two neural networks, one for candidate generation and the other for ranking. The first algorithm narrows down the massive video library, and the second algorithm ranks these videos based on their value to the user. Youtube’s use of deep neural networks also makes Google one of the first companies to deploy production level deep neural networks for recommender systems. The candidate generation network uses the user’s activity history- user-level demographics, the IDs of videos watched, search history and gives an output of a few hundred videos that are broadly applicable to the user. Candidate generation uses collaborative filtering based on what people with similar ratings have enjoyed previously. The second network uses richer features for high recall to not miss out on the relevant videos for suggestions. The network is trained end to end. Addictive algorithms In Google’s I\/0 conference 2018, the company announced plans to introduce tools to cap binge watching. However,  Google’s research paper (2019) proposed an update of its algorithms to recommend even more targeted content to increase engagement. “The largest supercomputers in the world are inside of two companies — Google and Facebook — and where are we pointing them? We’re pointing them at people’s brains, at children,” said Tristan Harris, a former tech ethicist at Google and co-founder of The Center for Humane Technology, in an interview. The Facebook algorithm decides which posts show up in the user’s feed. Facebook relies on multiple layers of ML models, and rankings developed to predict posts most valuable  and meaningful to the user. Personalisation and relevant content remain the company’s top priority. TikTok’s recommendation engine leverages AI and data mining practices to build its ‘For You’ feed. Recommendation algorithms are developed to get users in the loop: As soon as a video ends, another video begins with relevant content. Tik Tok, Instagram reels do not have the option to disable auto-playing of content making it difficult to stop watching. Endowment Effect: The more time one invests in constructing a virtual world in a game or a profile on social media, it gets harder to detach from or delete the app. Social pressure: For example, blue ticks upon viewing and grey ticks upon the message being successfully delivered.Social reward and feedback Ethical issues Algorithms that show a user what they want to see and get better at predicting what the user consumes eventually creates an echo chamber. For example, because YouTube’s algorithm is optimised for maximum engagement, it tends to offer choices that reinforce already held beliefs, likes and dislikes shutting down other views, all of it creating an addictive experience. Reports show controversial and extreme videos are rewarded, leading to misinformation and political radicalisation. In many respects, the political discourse people engage in today is a direct product of social media. The purpose of algorithms is to help make choices. The user makes a choice from an array of options which is fed to train the algorithm. In time, it creates a feedback loop where the output becomes a part of the algorithm’s input. To a large extent, algorithms are treated as purely engineering challenges and not as socio-technical problems. Tech experts are more concerned with finding effective solutions rather than their societal impact. Algorithms pick up biases over time. Despite strong arguments and evidence that algorithms narrow options, some research studies doubt the validity. Dutch communications expert Judith Moller and colleagues used algorithms for article recommendations in a newspaper. The study reported a more diverse set of outputs as compared to human editors’ picks. Machine learning algorithms have mastered recommending what a user will engage with. Research is done on engineering diversity to improve the range of choices and better understand user interests. That’s as far as algorithms are concerned. Social media companies, however, need to be regulated but balancing profit and human interest is a challenge for companies.","excerpt":"In Google’s I\/0 conference 2018, the company announced plans to introduce tools to cap binge watching.","categories":["AI Features"],"tags":[],"author_name":"Prajaktha Gurung","publish_date":"2021-07-25T13:00:00","publication_year":"2021","word_count":857,"keywords":["Go","API","machine learning","TPU","AI","neural network","ML","RAG","Ray","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","Ray","RAG","TPU","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-social-media-algorithms-ever-be-ethical\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":67680,"title":"SAS Viya Introduces A New Category Of Analytics For The Cloud","content":"SAS announces SAS Viya 4, which will be available in late 2020, and is engineered to take advantage of the newest cloud technologies. This announcement is in conjunction with Virtual SAS Global Forum 2020, the world’s premiere analytics conference. Due to the COVID-19 pandemic, this year’s conference is being held virtually. Designed to be delivered and updated continuously, the new architecture helps bring powerful analytics to everyone, everywhere. Because SAS Viya integrates the art of decision making with the science of artificial intelligence (AI) and analytics, organisations will be able to make better decisions, faster. The innovation behind this release underscores SAS’ commitment to helping organisations transform data into intelligence. This transformational release of cloud-native SAS Viya is one outcome of SAS’ $1 billion investment in AI. “This release marks an aggressive and innovative step for the SAS platform and for our customers,” said COO and CTO Oliver Schabenberger at SAS. “Organisations are asking to fuel their digital transformation with agility, speed, automation, intelligence and continuity. Those are the attributes of SAS Viya 4 – cloud-native advanced analytics and AI for users of all skill levels, turning business intelligence into intelligent business in the cloud.” “Enterprises are implementing major changes to their data and analytics technology driven by cloud-native architecture,” said Dan Vesset, Group Vice President, Analytics and Information Management at IDC. “SAS Viya enables a flexible and efficient way to execute data and analytics workloads within container, and microservices-enabled architecture. Organisations can decouple the analytics from the environments in which they run to scale-up services quickly and meet decisioning needs in a much more agile fashion.” SAS is simplifying how AI and machine learning is embedded in decisions. With a reimagined cloud-native architecture and the availability of interactive or programming interfaces, SAS customers will no longer be bound by programming language, data silos or skills. Automated data preparation, machine learning (AutoML) and model deployment improve the productivity of scarce data science resources and expand AI capabilities to those with more widely available skill sets. Results are explained in easy-to-understand terms so all can act at the moment with confidence. With a renewed focus on APIs, SAS Viya 4 makes it easier for application developers to collaborate with data science teams and respond quickly to changing business needs. These capabilities allow customers to acquire and consume enterprise-scale analytics in the most efficient way; they can consume just the AI services they need. SAS Viya 4 uses continuous integration, continuous delivery (CI\/CD) process that allows customers to choose their release intervals, so they gain access to the latest product innovations the moment they’re ready or can incorporate updates into their own change-management schedules. The container-based architecture, orchestrated by Kubernetes, provides portability across different cloud environments, including Azure, Google, AWS and OpenShift. Every organisation has data, but it’s what they do with the data that matters. Analytically mature organisations know every decision that comes from their models can make a significant impact on the bottom line. SAS Viya 4 simplifies model deployment – helping cross that critical “last mile” of analytics – and offers a central location to monitor and manage the performance of all analytic models. Organisations also struggle to explain decisions and foster ethical AI adoption. As AI and machine learning become more widespread, SAS Viya 4 centralises the management of all open source and SAS models, lineage and templates, giving full visibility and control over all modelling activities.","excerpt":"SAS announces SAS Viya 4, which will be available in late 2020, and is engineered to take advantage of the newest cloud technologies. This announcement is in conjunction with Virtual SAS Global Forum 2020, the world’s premiere analytics conference. Due to the COVID-19 pandemic, this year’s conference is being held virtually. Designed to be delivered and updated continuously, the […]","categories":["AI News"],"tags":["corporate analytics platform","sas"],"author_name":"Rohit Yadav","publish_date":"2020-06-18T20:00:00","publication_year":"2020","word_count":569,"keywords":["data science","artificial intelligence","machine learning","AWS","AI","corporate analytics platform","ML","microservices","analytics","sas","Azure","kubernetes"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","AWS","Azure","kubernetes","microservices"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sas-viya-introduces-a-new-category-of-analytics-for-the-cloud\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10084049,"title":"Big Tech&#8217;s Impact on India in 2022: A Retrospective","content":"The big techs have just survived a brutal year. For the past few years, firms considered money machines struggled to survive the slowdown. However, even during the challenging year, the firms were focused on India and made some major contributions and initiatives. Here’s a detailed glimpse of how big techs contributed to the Indian economy in 2022. Amazon In November, Amazon set up the second AWS region in India and announced an investment of $4.4 billion in the South Asian market by 2030. The additional AWS data centre cluster will allow the firm to offer “greater choice” in the country and support over 48,000 full-time jobs annually, Amazon said. AWS, which leads the cloud market in India, has amassed several major clients. Moreover, AWS is set to invest in India’s space startups and partner with the Indian Space Research Organisation. Clint Crosier, director of AWS Aerospace and Satellite, called India a leader in space missions and said AWS wants to help the economy manage its space data by leveraging cloud. The tech giant extended its presence in the gaming industry by quietly rolling out Prime Gaming. This subscription service offers access to several new titles added each month. Apple Apple’s stock crashed more than 20% this year. Still, the tech giant is increasingly diversifying its manufacturing base to de-risk the supply chain and make it more resilient, and India has been a strong contender. Incentives under PLI have strengthened India’s hand. Moreover, the robust production in India despite Covid waves has impressed the iPhone maker. In September, Apple launched the iPhone 14 model and began assembling it in India ten days after the global launch. The iPhone maker highlighted Apple’s decreasing dependability on China by making the move. Analysts at JP Morgan have predicted that by 2025, Apple will broaden its manufacturing capacity in India, producing 25% of all iPhones. Google Google launched ‘India Ki Udaan’ to mark 75 years of Independence. The project executed by Google Arts & Culture celebrates the country’s achievements. Meanwhile, Google also announced its collaboration with the Ministry of Culture. The big tech announced a grant of $1 million to IIT Madras for setting up a Centre for Responsible AI. The initiative will mitigate bias using Natural Language Processing models and enable fair use of AI, particularly in the Indian context. Alongside, the Google for India, 2022 event witnessed the declaration of many such programmes in their commitment to building a more inclusive, helpful, and safer internet for every Indian. Microsoft In March, Microsoft announced its plans to build its fourth datacenter region in Hyderabad. This region will be built with sustainable design and operations in mind, a key example of Microsoft’s climate commitment as it responsibly delivers reliable cloud services at scale. At the Summit 2022 in India, Microsoft CEO Satya Nadella announced about partnering with Apollo Hospitals, a leading healthcare Indian chain to invest in a cardiac prognosis model, trained on data from South Asians. The Indian-origin chief said that Apollo, through all of their deployments, is building a model with a more accurate predictive capability that will be available as an API to every hospital worldwide. Meta Amid the mass layoffs, hiring halts, and C-suite exits, Meta was in the news for the wrong reasons. But the social media giant made progress by investing at certain stages in its biggest market, India. In September, the Ministry of Electronics and IT’s Startup Hub signed an agreement with the social media giant to launch an accelerator programme to offer grants to startups building services for the metaverse. The Facebook parent, in November, announced that it would be providing 1 million dollars to launch a fellowship programme that aims to boost the developer ecosystem around immersive technologies as the company looks to make a deeper push into the metaverse. The programme, called XR Open Source (XROS) fellowship, is part of Meta’s global XR (extended reality) Programs and Research Fund, under which the company announced a $2 million fund for the XR Startup programme with MeitY Startup Hub earlier this year. At a time when external pressures on Meta were building, the company’s internal issues also started drawing attention. Meta has witnessed resignations from some of its top leaders in India. WhatsApp’s India head Abhijit Bose and Meta’s India public policy director Rajiv Aggarwal resigned in November. Their exits were followed by Meta India head Ajit Mohan’s resignation. Mohan resigned from Meta to join Snap.","excerpt":"Everything from data centres to gaming!","categories":["AI Trends"],"tags":["Satya Nadella","tech hiring"],"author_name":"Tasmia Ansari","publish_date":"2023-01-04T18:00:00","publication_year":"2023","word_count":739,"keywords":["Satya Nadella","Go","API","AWS","AI","ETL","RAG","responsible AI","Aim","GAN","R","tech hiring"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","R","Go","API","ETL","GAN","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/did-big-tech-do-much-for-india-in-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":57902,"title":"5 Times &#8216;Leader Of The Century&#8217; Neutron Jack Championed AI Strategies","content":"With the recent passing of one of the business greats and a rock star CEO — Jack Welch, media is pouring in with tributes, however trying to summarise the life of such a great leader and his strategies in a few hundred words are nearly impossible. Jack Welch joined General Electric in 1960 as a chemical engineer, and at the age of 37 became the youngest vice president of the company in 1972. He served 20 years as the CEO of GE from 1981 to 2001, and during his tenure, GE’s total market cap soared from $14 billion to $410 billion. Welch is also known for making efforts in abolishing the nine-layer management hierarchy and bringing a sense of informality to the company. And during the 1980s, he also devised a strategy to attract the right employees for his organisation, and that gained him a moniker — “Neutron Jack.” Although Welch was known as one of the prominent CEOs of the last century, in reality, much of his success at GE was credited to his basic business principles and leadership strategies. In this article, we will dive into some business lessons from the ‘manager of the century.’ Jack Welch, former Chairman and CEO of GE, a business legend, has died. There was no corporate leader like “neutron” Jack. He was my friend and supporter. We made wonderful deals together. He will never be forgotten. My warmest sympathies to his wonderful wife & family!— Donald J. Trump (@realDonaldTrump) March 2, 2020 Asking The Right Questions According to Jack Welch, “Good business leaders create a vision, articulate the vision, passionately own the vision, and relentlessly drive it to completion.” An important aspect of Welch’s management was built on asking the right questions. An organisation must ask questions relevant to their business to unearth the true potential of artificial intelligence. Business leaders with a pulse already understand the significance of AI and related disciplines on their business. Usually, business leaders are prepared to ride the current wave of technology but have no knowledge of how to get started and what constitutes the best organisation specific applications of AI. Organisations need to understand the fundamental areas where AI-augmented solutions can help employees in making better decisions. Therefore, Welch believes that in order to get bigger and better solutions, leaders must probe by asking questions and stirring up a healthy debate. But, in most companies, employees are too conscious to ask questions, probably because knowingly or unknowingly, some companies don’t reward curiosity and even sometimes actively discourage it. And, that’s threatening for businesses in today’s competitive world. Welch believed that challenging your employees is an art, not a science. In fact, each individual requires a unique approach, and it’s the job of a leader to bring out the best in them without diminishing their productivity. Riding The Wave Of Change Jack Welch strongly believed in adapting to change. According to him, “An organisation’s ability to learn, and translate that learning into action rapidly, is the ultimate competitive advantage.” With the expanding presence of AI in our daily lives, business leaders must pursue AI with purpose and forethought. While businesses are aware of the transformative powers of AI, leaders must genuinely prepare themselves for the changes. Welch believed that leaders, along with their organisations, must be curious to learn newer technologies and introduce the same in ways that are effective and sustainable. Forward-thinking leaders must focus their efforts and resources in learning about the implications of AI for their organisations. Welch was always known for re-evaluating his priorities every five years by asking himself — “What needs to be done now?” Such an approach was termed as organised abandonment, which involves omitting antecedent strategies by consciously assessing the business based on evolving technologies, shifting markets and an outlook for the future. In short, it is suggested that leaders shouldn’t be afraid to let go of the past and should focus on the future wasn’t planned before. At GE, the development of AI focuses heavily on connecting minds with industrial machines in order to enable intelligent and user-friendly products to power the world. Leaders at GE spearheads this charter inventing and deploying AI solutions to solve business problems. Inspiring Followers Not Workers According to his peers, Jack Welch was known to make profound statements to his employees for creating inspiration. Welch, during his management, was driven towards pumping up his people and bringing confidence in them. He once said in an interview, “One of the jobs leaders have as a manager is to pump everyday self-confidence into their team to make them feel great.” Leaders might not see their role as that of a cheerleader; however, it is extremely important to take up the responsibility in order to get the best out of your employees. Leaders must learn how to steer their teams and boost their confidence by advocating a culture of lifelong learning and embracing their risk-taking approach. Today’s employees are always under immense pressure to possess the right IT skills and adapt quickly to newer technologies like AI and automation. Such a task is made even harder when employees are not given the proper environment and upskilling opportunities to support their transition in the company, which, in turn, reduces their self-confidence. Therefore, opening a discussion around technology issues is undoubtedly helpful in fostering a digital culture and building digital confidence among the employees. Jack Welch, then-CEO of General Electric, 1989. PC: Joe McNally | Getty Images Setting An Example Jack Welch was always keen to stay in the game and was never afraid to take up responsibility as a business leader. Apart from the five-year planning process, Welch also used to set up three priorities for himself, and then focus working on those with his leadership team. This was a smart way of Welch to stay in the trenches and be always aware of his business. Usually, employees consider themselves distant from their management, which in turn dulls the sense of the business. According to Welch, “Winning companies tend to embrace risk-taking and learning.” And, leaders must set the example by encouraging their employees to experiment and take risks without the fear of being countered. In an AI age, changing one’s course of action, which was previously perceived as a sign of weakness or lack of conviction, would now be considered as a strength when it improves decision making. Adaptable modern leaders are not hesitant to commit to a new course of action when it is demanded, and their adaptability allows them to confront challenges with a focus on learning rather than being right. Welch always considered experimentation as a significant key to growth. Celebrating Employees In the current environment of digital transformation, the process of employee engagement has gone beyond just keeping people happy. The relationship between employee satisfaction and the company’s bottom line has become stronger than before, where it has transformed into more fulfilling and goal-oriented. It is believed that, with employees becoming more engaged, it is likely to lead to a higher quality of work produced. Committed organisations have double the rate of success compared to less engaged organisations. Ideally, satisfied employees are living the company values every day at work, which in turn makes them recognised across the organisation for it. Jack Welch, while noting digital companies, spoke in an interview about leaders not making a big deal out of small wins. He said, “celebrating employees would make them feel like winners and would create an atmosphere of recognition and positive energy.” In fact, he believes that the importance of employee celebration can’t be overstated. Leaders need to stand firm in celebrating their employees early and often. Celebrating small victories would lead to complacency, and according to Welch, nothing could be further from the truth.","excerpt":"With the recent passing of one of the business greats and a rock star CEO — Jack Welch, media is pouring in with tributes, however trying to summarise the life of such a great leader and his strategies in a few hundred words are nearly impossible. Jack Welch joined General Electric in 1960 as a […]","categories":["AI Trends"],"tags":["AI Strategy","leader of ai"],"author_name":"Sejuti Das","publish_date":"2020-03-03T18:00:00","publication_year":"2020","word_count":1295,"keywords":["Go","API","artificial intelligence","lifelong learning","AI","Git","AI Strategy","RAG","ViT","GAN","R","leader of ai"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","R","Go","Git","API","GAN","ViT","lifelong learning"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-times-leader-of-the-century-neutron-jack-championed-ai-strategies\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":26115,"title":"Here’s Why Hackathons Are A Great Way To Identify Machine Learning Talent","content":"Hackathons and hiring go hand-in-hand, especially for talent-starved mid-sized to smaller companies. These organisations can often be seen using hackathons to tap into a wider talent base. While we have already analyzed the value of a hackathon from a participant’s perspective who walks away with programming experience, cash prizes and a finished project with heavy backend work, sponsors and investors too can leverage it as a recruiting tool. Hackathons do play an important role in the recruitment process. They are usually an effective way to build trust and likeability among the developer base. Besides getting the IT community at a single venue, hackathons present an excellent opportunity to meet young, (often under the age of 26) talent who turn up to improve their skills. According to a report by HackerEarth, hackathons have been surging in popularity and are becoming a great tool for idea generation. The report indicated that India grabbed the second spot in conducting hackathons, after the US, which hosted a total of 379 hackathon events. Indian cities like Bengaluru, Mumbai and Hyderabad were among the top 10 popular cities around the world to host the maximum number of hackathons. When it comes to hiring machine learning talent, we have all heard how Kaggle is the best way to get noticed by top employers. In This Article, We Will List Down Why Hackathons Are A Great Way To Identify Machine Learning Talent: 1) Massive Online Free Resources To Learn ML: At a time when there is an explosion of free resources and developers are threading a self-learning path, it can be difficult to evaluate a potential candidate’s proficiency and skill sets. Hackathons provide a platform to developers to showcase their newly-acquired job-ready skill sets in a competitive environment. 2) Best Way To Beat HR And Employee Referral Filter: Let’s face it — most of the current crop of machine learning talent is self-taught. And at a time when the human resources have a strict standard for judging resumes (masters specialisation, extensive work experience, real-life projects) getting a CV screened for a top company may seem like a daunting task. During hackathons, product managers get an excellent opportunity to mingle with the developer community and snag the best talent, or perhaps put them on a waitlist. 3) Masters Candidates Are Hard To Come By: It is a common refrain in data science interviews that a Master’s degree is a prerequisite. But in India, where machine learning is a relatively new area, candidates with an MS background are hard to find. The talent pipeline is usually plugged with self-taught candidates who have interesting ML projects to boasts about. Self-taught candidates who have issues with object-oriented programming or may not understand how an algorithm scale. 4) Best Way To Use Company’s APIs And SDK: Hackathons are the best way to drive API adoption — which is one of the top six reasons to conduct hackathons, says a report. Today, around 80% of the Fortune 500 companies conduct hackathons as a way to tap into ML talent. API adoption can help a company in several ways. The Hackerearth blog indicates that even big tech firm like Google celebrated its 10th anniversary with an API hackathon to drive adoption of Google Maps to meet developers who are building the map of today. 5) Relationship Building With Employees: Internal hackathons are also a good way for internal recruiting and benchmarking employees for specific skill set. Most of the times, teams work in silos and internal hackathons provide an excellent way to promote current employees and find talent internally. Outlook At a time when machine learning talent is hard to come by, hackathons provide an engaging way of harvesting prospective employee database, boost employer brand and present a great opportunity to meet developers under one roof. In fact, according to news reports, tech giants like Google and PayPal are known for using hackathons to find fresh tech talent along with older recruiting techniques such as campus hiring. Companies of all sizes today find experienced professionals highly priced and hackathons provide an excellent window of opportunity to young professionals to showcase their talent to big tech companies. For hosting and participating in great Machine Learning Hackathons, visit www.machinehack.com","excerpt":"Hackathons and hiring go hand-in-hand, especially for talent-starved mid-sized to smaller companies. These organisations can often be seen using hackathons to tap into a wider talent base. While we have already analyzed the value of a hackathon from a participant’s perspective who walks away with programming experience, cash prizes and a finished project with heavy […]","categories":["AI Features"],"tags":["Hackathon","Machine learning hackathon","Machinehack"],"author_name":"Richa Bhatia","publish_date":"2018-07-04T09:47:35","publication_year":"2018","word_count":702,"keywords":["data science","Go","API","machine learning","AI","Machine learning hackathon","Machinehack","ML","RAG","Hackathon","Rust","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","RAG","R","Go","Rust","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/heres-why-hackathons-are-a-great-way-to-identify-machine-learning-talent\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10114073,"title":"Mistral’s ‘Le Big Model’ Beats Google’s Gemini Pro, Signs Multi-Year Deal with Microsoft","content":"Mistral AI today released Mistral Large, its latest and most advanced language model. It is accessible through La Plateforme and Microsoft Azure, marking a strategic distribution partnership with Microsoft. Mistral Large achieves strong results on commonly used benchmarks, making it the world’s second-ranked model generally available through an API (next to GPT-4) beating Google’s Gemini Pro and Anthropic’s Claude. The model demonstrates advanced multilingual capabilities, fluently understanding English, French, Spanish, German, and Italian. Its 32K tokens context window allows precise information recall from extensive documents, enhancing its usability for complex multilingual reasoning tasks, including text understanding, transformation, and code generation. Mistral Large has native multi-lingual capacities. It strongly outperforms LLaMA 2 70B on HellaSwag, Arc Challenge and MMLU benchmarks in French, German, Spanish and Italian. Alongside Mistral Large, Mistral AI has also introduced Mistral Small, an optimised model designed for low latency workloads. Outperforming Mixtral 8x7B and featuring lower latency, Mistral Small offers a refined solution between Mistral’s open-weight offering and its flagship model. Mistral AI has streamlined its endpoint offerings, providing open-weight endpoints with competitive pricing and introducing new optimized model endpoints – mistral-small-2402 and mistral-large-2402. The company aims to offer users a comprehensive view of performance\/cost tradeoffs. Introducing JSON format mode, Mistral AI allows developers to obtain model output in a structured and valid JSON format. Additionally, the model supports function calling, enabling more intricate interactions with internal code, APIs, or databases. Currently, function calling and JSON format are only available on mistral-small and mistral-large. Multi-Year Partnership with Microsoft Microsoft announced a multi-year partnership with Mistral AI. Microsoft’s partnership with Mistral focuses on three core areas- Supercomputing infrastructure, Scale to Market and AI research and development. “We’re announcing a multi-year partnership with MistralAI, as we build on our commitment to offer customers the best choice of open and foundation models on Azure,” wrote Microsoft chief Satya Nadella. Microsoft will provide Mistral AI with access to Azure AI supercomputing infrastructure, ensuring superior performance and scalability for AI training and inference workloads. The collaboration aims to make Mistral AI’s premium models accessible to customers through Models as a Service (MaaS) in the Azure AI Studio and Azure Machine Learning model catalog. Customers can use Microsoft Azure Consumption Commitment (MACC) for purchasing Mistral AI’s models, enhancing global availability. Further, Microsoft and Mistral AI will explore collaboration in training purpose-specific models for select customers, focusing on European public sector workloads.","excerpt":"Alongside Mistral Large, Mistral AI has also introduced Mistral Small, an optimised model designed for low latency workloads.","categories":["AI News"],"tags":["Gemini","Google","mistral"],"author_name":"Siddharth Jindal","publish_date":"2024-02-26T22:21:09","publication_year":"2024","word_count":398,"keywords":["Anthropic","Gemini","Gemini Pro","machine learning","TPU","AI","R","ML","Aim","mistral","Google","foundation models","Azure"],"extracted_tech_keywords":["AI","machine learning","ML","foundation models","Anthropic","Gemini Pro","Aim","Azure","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mistrals-le-big-model-beats-googles-gemini-pro-signs-multi-year-deal-with-microsoft\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":28305,"title":"Why Machine Learning Lifecycle Management Is Important","content":"Data science and machine learning are becoming the critical building blocks for data-driven organisations. Now accelerating ML lifecycle management and taking the models from prototyping to production has become increasingly important. Another factor that has led to the rise of ML lifecycle solutions is the significant increase in expectations from engineering teams to meet the varied demand to develop and scale ML capabilities. In a recent talk at Strata Data Conference, Hope Wang from Intuit defined the machine learning platform as not just being the sum of its parts. He said that the key to ML was how it supports the model lifecycle end to end. This includes data discovery, feature engineering, iterative model development, model training, and model scoring (batch and online). The management of artefacts, their associations, and deployment across various platform components is vital. ML lifecycle management has started to play a key role in organisations that face different challenges when it comes to developing machine learning artificial intelligence algorithms and putting ML in production. Today, large enterprises and startups face a clutch of different challenges when it comes to developing ML, AI algorithms and putting these models into production. According to one tech writer, ML development is an experimental and exploratory process, whereas deployment demands consistent results that are secure and well-managed. A lot of engineering organisations are expected to develop and scale ML capabilities. Importance Of Machine Learning Life Cycle Management It is important because it delineates the role of every person in a company in data science initiatives, ranging from business to engineering. It takes each and every project from inception to completion and gives a high-level perspective of how an entire data science project should be structured in order to result in real, practical business value. Failing to accurately execute on any one of these steps will either result in models with no practical value or that provide actively misleading insights. While there are a number of mature technologies that support each phase of this lifecycle, there are limited solutions available that tie these components together into a cohesive ML platform. To support the lifecycle of a model, you must be able to manage the various ML-related artefacts and their associations and automate deployment. A lifecycle management service built for this purpose should be leveraged for storage, versioning, visualising (including associations), and deployment of artefacts. Of late, there has been a spurt in enterprise solutions which operationalise ML lifecycle management tasks. For example, US-based startup Quickpath allows data science teams to operationalise ML across the organisation and streamline the path to production. According to a company statement, the Quickpath ML OPS and Engineering Platform is the only comprehensive production data science platform present in the market today. Key Features Of ML Platform The platform should support model development in different programming languages, and language and package versions should be configured specific to a model. Another key feature is the connection between various artefacts and platforms — the data and datasets: source data and feature data, training datasets, and scoring result sets; code — notebook code, model code, deployment code and lastly model-specific environments and platforms, for example, developing and training platforms The components of the ML platform should associate and interact. Another company, DataRobot also provides automated ML platform that streamlines the ML life cycle by simplifying the most complicated, time-consuming steps with automation. The platform makes data exploration and model building easier and more accessible, allowing those who understand the business problem behind the data science project to rapidly build and test dozens of models in a fraction of the time it would take using traditional methods. Conclusion With ML gaining more traction in businesses, a development lifecycle that supports learning models for building custom ML algorithms and applications has become very crucial. Hence, it is important for data-driven organisations to choose an ML platform that provides interoperability with other ML frameworks.","excerpt":"Data science and machine learning are becoming the critical building blocks for data-driven organisations. Now accelerating ML lifecycle management and taking the models from prototyping to production has become increasingly important. Another factor that has led to the rise of ML lifecycle solutions is the significant increase in expectations from engineering teams to meet the […]","categories":["AI Features"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-09-13T10:08:51","publication_year":"2018","word_count":650,"keywords":["data science","Go","API","machine learning","artificial intelligence","AI","ML","feature engineering","RAG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","RAG","R","Go","API","feature engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-machine-learning-lifecycle-management-is-important\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10076521,"title":"Shahrukh Khan, Deepfake and Political Campaigns","content":"In 2021, voters in Punjab received an automated phone call from Delhi’s Chief Minister Arvind Kejriwal, in which he urged them to vote for the Aam Aadmi Party (AAP). AAP is widely claimed to have contacted millions of voters using this interactive voice response technology. This use of automation in political campaigns, although not new, piqued voters’ interest. And while this personalised audio advertising may have achieved that successfully, it’s probably even more intriguing to speculate what may have entailed if Arvind Kejriwal himself had appeared on a voter’s smartphone screen and asked them to vote for the party—all the while addressing the voters with their names. Sounds like fiction? It’s not though; it’s personalised video marketing unfolding its wings in India. Globally, personalised video marketing has been attempted by many brands—including the likes of Nike and Cadbury. However, they were not personalised to a degree that’d seem fictitious to imagine. For instance, this Cadbury GLOW advert enables users to add text to the video: In another example, Marketo’s marketing invitation also makes the user feel significant by including its name. However, these campaigns are not as personalised as certain recent advertisements in India have been; for example, the Cadbury advertisement featuring actor Shah Rukh Khan allows small company business owners to make original, highly personalised studio-quality advertisements. The advert employs Khan’s visage and voice in such a manner that it shows the actor saying the name of the local store or brand. Just like everyone else, AIM was intrigued about this hyper personalisation and reached out to the man behind the cause—Ashray Malhotra. His company Rephrase.ai is behind the successful advertisement featuring the actor. Rephrase.ai claims to be a deep learning organisation that creates synthetic video content using AI. Addressing the challenges they faced while working with this campaign, Ashray said, “The Shah Rukh Khan campaign was one of the most challenging ones that we had ever done at that point. And the Ogilvy team was extremely clear about the creative vision that they wanted—which was that they did not want Shah Rukh standing; they wanted him to have different poses in different places, and such. This meant that we had to develop five, totally digital clones of Shah Rukh Khan, rather than just one, as we had done previously.” He further explained the complex intricacies of the task at hand: “ Furthermore, there were many more motions in the total project than we had previously dealt with. Overall, it generated AI issues in terms of video cloning—which was something we hadn’t worked with at the time. Also, each room we recorded was somewhat different, resulting in slightly varied sounds, reverberation, and so on. As a result, we had to construct numerous audio models of Shah Rukh Khan.” Keeping out of controversy Could personalised advertising avoid controversy now that it has been well-established that it has genuine potential to transform advertising in India? For example, in a recent AI-enabled Zomato advertisement, actor Hrithik Roshan referred to a temple as a restaurant—offending numerous viewers along with temple officials. When advertisements are created using algorithms, it is highly likely that they will not be able to grasp human context with a much-needed accuracy. In the case of the aforementioned advertisement, the system had no way of knowing which restaurant’s name could cause trouble. However, Ashray believes otherwise. He claims that much depends on data accuracy and ethics. Giving the example of his own company, Ashray said, “So far, we’ve worked with 50 firms and have had no controversies. One of the main reasons for this is that we strive for data accuracy at all stages of the process. To be clear, ethics is an important aspect of how we spend our time here at Rephrase.ai. From the screenplay to the files, to the edited videos, to the finalised clips—everything goes through multiple checkpoints.” “I am really committed to ensuring that this technology just helps scale people’s time and does not assist in the creation of videos that you would not have chosen to produce in the first place. Because we recognise it as a really powerful technology, as the pioneers of this tech, we want to see it being utilised for good.” Personalised Political Campaigns, the next big thing? In olden times, politicians used to go house to house to collect votes. This practice helped them create a personal bond with their voters. Owing to the growing population and digitalisation, more and more campaigns are shifting to digital mediums. Now that India has witnessed multiple AI-enabled personalised adverts, could political campaigns turn to hyper personalised advertising too? It won’t be much of a problem for political parties to undergo such changes. Politicians like Naveen Patnaik are already experimenting with AR tech for campaigns like “Selfie with CM.” However, taking a step further, could it be possible for political parties to launch a campaign where voters receive an email or text message in which a politician is urging them to vote for their party by calling the voters by their names? According to Ashray, it would be feasible. But, he also shared his concerns surrounding such usage of AI in the political realm: “Yes, it can be done. However, you have to take into consideration that while we may think that political parties creating personalised campaigns for their candidates is a wonderful idea, it can quickly backfire when they start building bad personalised campaigns of their opponents.” “If a political party develops such a campaign for its rival party with evil intentions, things will become ugly. We at Rephrase.ai are not even remotely considering moving in that direction currently.” Ashray also believes that a good use of this tech could be drive social causes; for instance, while talking about campaigns focused on environmental causes, he said: “If any of your viewers read this, and if they are either working with any NGO for education or any environmental reasons or tackling practically any social problem, we absolutely believe in the ability of customised films to help. I feel that this might aid in fundraising, which in turn might help these causes; I really want to double down on the space.” Ashray concluded by encouraging readers to connect with him for collaboration on causes for social good: “So, please contact me; I’d love to work on it.”","excerpt":"With the increasing personalised advertising in India, could it be possible for political parties to introduce personalised ad campaigns soon? AIM speaks to Ashray Malhotra, CEO, Rephrase.ai","categories":["IT Services"],"tags":[],"author_name":"Lokesh Choudhary","publish_date":"2022-10-06T15:00:00","publication_year":"2022","word_count":1053,"keywords":["Go","AI","Git","RAG","Ray","Aim","deep learning","CLIP","GAN","R"],"extracted_tech_keywords":["AI","deep learning","Aim","Ray","RAG","R","Go","Git","CLIP","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/shahrukh-khan-deepfake-and-political-campaigns\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":28062,"title":"Can Artificial Intelligence Prove Polanyi’s Paradox Wrong?","content":"The Piltdown Man Hoax – Painting by John Cooke,1915. In most of our articles, we have encapsulated developments in artificial intelligence and machine learning over the last few years. Many of the news and articles have been about the impact of these technologies on human lives, mostly in a positive way. However, not every progress could be deemed right and has become a subject of debate, mostly by comparing it with human thinking or conscience. This article, however, will explore AI with respect to human intellect and its interpretation of us. Here’s a thought-provoking question: Have we fully comprehended AI systems? The answer is a resonating “no”. There is nothing solid when it comes to a serious understanding of AI. Even though extensive research has unravelled countless possibilities, somewhere we fail to have a complete knowledge of intelligent systems. This brings us to Polanyi’s Paradox. In this article, we will delve into where we stand in our own understanding and that of AI’s perception. What Is Polanyi’s Paradox? Michael Polanyi, a Hungarian-British polymath wrote a book called Tacit Dimension in 1966 which explored the ‘tacit knowing’ in human knowledge. In the book, he goes on to tell that our ability to learn certain tasks is gained through the experience of doing them, and this cannot be explained by us in detail. For example, he says that we recognise faces without knowing how we do it. The ‘tacit’ knowledge obtained here is different from explicit knowledge, which can generally be articulated easily. In light of this, sometimes our knowledge and capabilities are often beyond our perception. To summarise, the mantra “we can know more than we can tell” is the bottom line of the book, which forms the Polanyi’s Paradox. Polanyi’s Paradox means that there is much more to our cognitive phenomenon which is beyond explicit knowledge and cannot be said expressed or pictured correctly. Can AI Supersede Humans? It is not surprising to see advancements in AI given the technology available today. In fact, it has led to a drastic, undesirable technological change for many. For example, menial jobs being replaced by machines or say automation, eventually displacing the need for humans. On the other hand, it has garnered awe from researchers in the form of innovations such as virtual assistants or self-driving cars. The point here is that AI systems are outperforming humans in some aspects. But, will it go beyond these capabilities and emulate humans completely? The answer to this is “no” because AI systems have not yet perfected other cognitive actions. AI also does not perform well at an overall intellectual level. It is devoid of human features like thoughts or emotions. If Polanyi’s Paradox is applied to AI, it becomes vague since the AI system is itself restricted to knowledge created by humans. Moreover, AI systems are predictable thus showing the cause for those actions unlike in humans where it cannot be determined. In his paper, Polanyi’s Paradox and the Shape of Employment Growth, American economist David H Autor posits why the human factor is essential. “The challenges to substituting machines for workers in tasks requiring adaptability, common sense, and creativity remain immense. Contemporary computer science seeks to overcome Polanyi’s paradox by building machines that learn from human examples, thus inferring the rules that we tacitly apply but do not explicitly understand.” Autor goes on to say that two approaches — environmental control and ML — can circumvent Polanyi’s Paradox significantly for machines. “Humans naturally tackle tasks in a manner that draws on their inherent flexibility, problem-solving capability and judgment. Machines currently lack many of these capabilities, but they possess other facilities in abundance: strength, speed, accuracy, low cost and unwavering fealty to directions. Engineering machines to accomplish human tasks does not necessarily entail equipping machines with human capabilities; instead, work tasks can, in some cases, be re-engineered so that the need for specifically human capabilities is minimised or eliminated.” A classic example would be object recognition of chairs. Algorithms are trained to identify chairs based on properties of a chair such as legs, arms, seat, back etc. Suppose, if these algorithms are implemented in a machine, it would identify chairs as per the algorithm criteria. But, not all chairs possess the same properties — no legs (beanbags) no back (stool) and so on. In this case, the machine would fail to recognise them as chairs whereas humans can tell that through knowledge — which cannot be ascertained as to why it happened that way. This is again Polanyi’s Paradox. Conclusion To summarise, AI systems are yet to come close to humans in terms of Polanyi’s Paradox. Consequently, overcoming this paradox in humans is still a debate going on. If successful, it might take a very long time for AI to completely emulate humans and counter paradoxical theories.","excerpt":"In most of our articles, we have encapsulated developments in artificial intelligence and machine learning over the last few years. Many of the news and articles have been about the impact of these technologies on human lives, mostly in a positive way. However, not every progress could be deemed right and has become a subject […]","categories":["AI Features"],"tags":["human intelligence"],"author_name":"Abhishek Sharma","publish_date":"2018-09-06T06:31:09","publication_year":"2018","word_count":802,"keywords":["Go","machine learning","artificial intelligence","AWS","AI","ML","virtual assistants","automation","ViT","human intelligence","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","virtual assistants","AWS","R","Go","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-artificial-intelligence-prove-polanyis-paradox-wrong\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10064468,"title":"OpenAI’s DALL·E 2 can churn out hi-res conceptual art from text commands and edit it too!","content":"In January 2021, OpenAI introduced DALL·E. Now, the AI research company has unveiled DALL.E2, the latest iteration of its text to image project. DALE2 can create realistic images and art from a description in natural language. The AI system has the capability to make realistic edits to existing images from a natural language caption as well as add and remove elements while taking shadows, reflections, and textures into account. It can also take an image and create different variations of it inspired by the original. DALL·E 2 learns the relationship between images and the text used to describe them. It uses a process called “diffusion,” which starts with a pattern of random dots and gradually alters that pattern towards an image when it recognises specific aspects of that image. DALL·E 2 is preferred over DALL·E 1 for its caption matching and photorealism when evaluators were asked to compare 1,000 image generations from each model. However, DALL·E 2 is a research project and is not available in their API. As part of OpenAI’s effort to develop and deploy AI responsibly, DALL·E’s limitations and capabilities are being tested with a select group of users. The safety mitigations developed include: Preventing Harmful Generations: DALL.E2’s capabilities are limited to avoid generating violent, hate, or adult images and removing such concepts during training itself. Curbing Misuse: Filters are placed to curb text prompts and image uploads that may violate our policies.Phased Deployment Based on Learning: Access to DALL.E2 is limited to trusted users who will help the team learn about the technology’s capabilities and limitations. To get on the preview waitlist, click here.","excerpt":"DALL·E 2 is preferred over DALL·E 1 for its caption matching and photorealism when evaluators were asked to compare 1,000 image generations from each model.","categories":["AI News"],"tags":["ai System","DALL-E2","OpenAI"],"author_name":"Kartik Wali","publish_date":"2022-04-06T22:49:18","publication_year":"2022","word_count":268,"keywords":["API","OpenAI","AI","DALL-E2","programming_languages:R","ai System","Rust","AI research","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","OpenAI","R","Rust","API","AI research","programming_languages:R","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-to-change-the-digital-image-making-game-with-dall-e-2-its-text-to-image-generator\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23907,"title":"How Blockchain And AI Are Important In The Drone Industry","content":"The popularity of drones is rapidly increasing across various industries like e-commerce and logistics. However, the growing number of drones is posing new challenges such as increasing air traffic, laying optimal air routes, generating flight plans, dealing with emergencies like collision and managing drone swarms, among others. Companies across the globe are dealing with these challenges with the help of two powerful tools — artificial intelligence and blockchain. Using AI To Make Drones More Useful AI is being looked upon as a way to help drones overcome human limitations. Experts believe that programming drones with AI and ML could help them learn from their mistakes, and improve each day and deliver better services. For instance, a paper published by Carnegie Mellon University explores how AR drones 2.0 are learning from their own mistakes. Currently controlled by humans to a large extent, AI-controlled drones are looked upon as the future where AI would take most decisions and learn to work independently of any human interference. Though a section of society fears that this may lead to a world of cyborg leading to nuclear holocaust, its benefits cannot be sidelined. Elon Musk had even called out a ban on killer robots including drones, fearing the same. This is just a single instance of how AI can impact future drones. Another popular use case is in the world of data. As researchers and companies have amassed staggering amount of information, training this data and developing AI algorithms to derive insights is another area which researchers are exploring. The growing amount of data has essentially led to the infusion of AI in drones. “Drone-captured data is an innovative solution for delivering sophisticated analytics to stakeholders and provides an affordable way to improve estimating, designing, progress tracking, and reporting for worksites,” Patrick Perry of Drone Base had said in a blog post. Many companies are using advanced computer processor to effectively process the huge amount of data. For instance, Kespry, a commercial drone company that believes itself to be an automated “end-to-end” solution extracting high quality data, is using Nvidia’s Jetson. Jetson along with Kespry’s home built hardware enables the drone to fly with minimal user input. It automatically uploads its data for processing by another AI algorithm. In a nutshell, AI and machine learning are major contributors in the drone industry, affecting every aspect of a drone program from hardware, data gathering, data analysis, generating insights and decision making, among others. Some of the AI companies in the drone industry are Pilot AI labs, Neurala and DroneLancer, among others. In India, Skylark Drones is a company that uses cloud-based intelligence with system integrated UAVs to enhance a customer’s insights. It is being used in sectors like mining and infrastructure. Blockchain Technology Revolutionising Drones Industry Not just AI, but blockchain is looked upon as a way to power the drones and make them more secure, accurate and easy to regulate, among others. As the applications of drones are becoming widespread, it would need newer infrastructure to provide flexibility, scalability and the ability to deliver efficient services quickly. To deal with these, blockchain technology is being looked upon as a way to deliver a framework which can be used by stakeholders in the commercial drone industry. Blockchain Can Ensure Security Blockchain is enabling drones to be provisioned with cryptographic material, thus ensuring confidential and secure communications. Used for both communication and messaging, these blockchain identifiers can enable flexibility in establishing trust models across heterogeneous devices. For instance, in case of a package delivery operation, a blockchain-based repository could log information about the operations such as time, location, resources, delivery date etc., and make the data accessible to authorised users. Identity Management Blockchain identifier-based remote identification system protects drone user information and other confidential information. It can monitor and report potential complaints about inappropriate drone usage and damage to public safety as well. For instance, Walmart is looking at leveraging blockchain in a process by which they could automate the logistics of delivery drones. The idea is that as a drone approaches a delivery box, it authenticates itself with a blockchain identifier. If the code is valid, the box unlocks and accepts the package. Other Areas Where Blockchain Can Be Used In Drone Industry Traffic management is an important avenue where blockchain is used. A system that does not require constant human monitoring and surveillance, but still can ensure the authenticity, safety, and security, is a great advantage. Deep Aero project is making drone industry more efficient, cost effective and secured by providing a blockchain based framework for air traffic management. Apart from decreasing operational cost it also deals with the decreased risk of collision. The company is also providing a Deep Flight Authorisation System which is again blockchain-based, and authorises drone flights paths to be created, modified and rejected either automatically or manually to provide air space access information. Conflict management system is another avenue that provides advisories to resolve the conflicts when they are detected by UTM system. Distributed Sky project is a company which is using blockchain technology to deliver a framework that can be used by all stakeholders in commercial drone market. It is providing an Ethereum blockchain based registration system, communication system and traffic control.","excerpt":"The popularity of drones is rapidly increasing across various industries like e-commerce and logistics. However, the growing number of drones is posing new challenges such as increasing air traffic, laying optimal air routes, generating flight plans, dealing with emergencies like collision and managing drone swarms, among others. Companies across the globe are dealing with these […]","categories":["IT Services"],"tags":["Drone Artificial Intellgence","uav companies in india"],"author_name":"Srishti Deoras","publish_date":"2018-04-20T12:05:56","publication_year":"2018","word_count":873,"keywords":["Go","machine learning","artificial intelligence","Drone Artificial Intellgence","AI","ML","Scala","RAG","analytics","Rust","uav companies in india","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","R","Go","Rust","Scala"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-blockchain-and-ai-are-important-in-the-drone-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10161386,"title":"No More Boring Drives with Mercedes-Benz’s New Google-Powered AI Agent","content":"Mercedes-Benz has partnered with Google Cloud to integrate conversational search and navigation powered by generative AI into its MBUX Virtual Assistant. The Automotive AI Agent, built on Gemini and Vertex AI, is specifically designed for the automotive industry and leverages Google Maps Platform to provide users with more detailed and personalised responses about navigation and points of interest. The Automotive AI Agent is built utilising Google Maps Platform’s extensive database of 250 million global locations. It can process complex, multi-turn dialogues and retain the memory of user conversations, enabling seamless interactions. Speaking about the partnership, Ola Källenius, CEO of Mercedes-Benz Group AG, said, “At Mercedes-Benz, we seek to offer our customers an exceptional digital experience. Our partnership with Google Cloud will further enhance in-car navigation, combining sophisticated location data with generative AI.” The AI-powered system allows drivers to ask detailed questions and receive personalised responses. For instance, users can query the MBUX Virtual Assistant for nearby fine-dining restaurants and ask follow-up questions like, “Does the restaurant have good reviews?” or “What is the chef’s signature dish?” “Mercedes-Benz is known for providing an amazing driving experience, and our partnership will bring cutting-edge AI breakthroughs to make those experiences even better. This is just the beginning of how agentic capabilities can transform the automotive industry,” said Google chief Sundar Pichai. The first deployment of the new system will appear in the upcoming Mercedes-Benz CLA later this year, marking the debut of the company’s MB.OS operating system. Additional rollouts to other models with the MBUX Voice Assistant are planned. Google Cloud’s Automotive AI Agent uses multimodal reasoning for natural language interactions and sources data from Google Maps Platform, which is updated nearly in real-time with over 100 million daily updates. In 2023, Mercedes-Benz and Google formed a partnership where Google Maps supplies the geospatial data and navigation services, while Mercedes uses Google Cloud’s AI tools to efficiently develop and apply AI models. The same year, Mercedes also announced that it going to add OpenAI’s ChatGPT within its in-house multimedia system MBUX Voice Assistant as part of an exclusive beta program available only in the US. This allowed drivers of more than 900,000 vehicles equipped with MBUX to engage in more dynamic conversations with the onboard AI.","excerpt":"The AI-powered system allows drivers to ask detailed questions and receive personalised responses.","categories":["AI News"],"tags":["Google"],"author_name":"Siddharth Jindal","publish_date":"2025-01-13T22:44:42","publication_year":"2025","word_count":373,"keywords":["Go","ChatGPT","OpenAI","AI","ML","Git","RAG","generative AI","Google","edge AI","R"],"extracted_tech_keywords":["AI","ML","generative AI","ChatGPT","OpenAI","edge AI","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/no-more-boring-drives-with-mercedes-benzs-new-google-powered-ai-agent\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10165145,"title":"Tesla is Hiring a Front-End Software Engineer in India","content":"American EV giant Tesla is hiring a software engineer in Pune, who is focused on front-end development. The role involves working with a software team focusing on certain internal tools in Tesla that facilitate process management of the design, permitting, and installation of energy products and systems. “This team is focused on building a stable and scalable platform that can be extended and configured for all Tesla Energy and Vehicle products that are sold and supported today, and any that will be introduced in the future,” read the job description. The role requires a “solid understanding” of web technologies such as HTTP, REST, AJAX and JSON. Strong proficiency in HTML, CSS and JavaScript\/ES6, including DOM manipulation and the JS object model Furthermore, the company is also hiring a PCB design engineer in the country. The role will involve working with the “high speed layout team” at Tesla, which is responsible for the layout of the physical hardware that delivers the autonomous driving capabilities, high performance computing and infotainment experience to their vehicles and AI supercomputers. Recently, Tesla officially began hiring professionals in India. However, the roles posted earlier were focused on operations, customer support, sales, and vehicle service. The hiring announcements followed after a recent interaction between CEO Elon Musk and Prime Minister Narendra Modi in the US. “Prime Minister and Mr Musk discussed strengthening collaboration between Indian and US entities in innovation, space exploration, artificial intelligence, and sustainable development. Their discussion also touched on opportunities to deepen cooperation in emerging technologies, entrepreneurship and good governance,” read a statement from India’s external affairs ministry. This hiring initiative follows previous efforts by Tesla to negotiate lower import taxes as a prerequisite for significant investment in the country. India recently reduced the basic customs duty on high-end vehicles priced above $40,000 from 110% to 70%, which may have influenced Tesla’s decision to explore the market further. Moreover, it was also reported that Tesla has finalised a location to open its first showroom in India, at the Bandra Kurla complex in Mumbai. The showroom is said to occupy 4,000 square feet on the ground floor of a commercial complex. Tesla also plans a second showroom in India at Aerocity in Delhi.","excerpt":"The hiring announcements followed after a recent interaction between CEO Elon Musk and Prime Minister Narendra Modi in the US.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Hiring","Tesla"],"author_name":"Supreeth Koundinya","publish_date":"2025-03-05T10:51:53","publication_year":"2025","word_count":368,"keywords":["Go","artificial intelligence","AI","ML","Hiring","Scala","CuPy","JAX","JavaScript","Tesla","R","Java","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","ML","JAX","CuPy","R","JavaScript","Go","Java","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tesla-is-hiring-a-front-end-software-engineer-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":61691,"title":"EnKash Rides The AI\/ML Wave To Achieve Hyper-Growth In Indian Fintech Space","content":"EnKash is one of the most prominent startups in the commercial card fintech space of India, and the only card fintech startup which is working across segments. According to the company, it is the first in India to bring corporate credit cards with multiple variants in the market. EnKash offers a wide range of products and solutions to facilitate penetration and distribution of commercial card issuance. While at one end, it enables purchase cards for large and mid-size businesses; at another end, it offers a range of commercial cards for SMEs and startups under the Freedom Card brand. Analytics India Magazine connected with Hemant Vishnoi, Co-Founder of EnKash to better understand how the startup operates. “While banks are mainly focused on serving large enterprises, we aim to cater to SMEs and startups which may have varied needs but are underserved by banks in the absence of qualifying prerequisites or due to the high cost of acquisition. Commercial cards, hence, are benefitting across businesses and are the best tools to manage short duration credits,” told Vishnoi. How The Startup Utilizes AI\/ML For a startup to stay ahead in the game, and to provide value to users, AI and ML have become essential tools. The co-founders have extensive experience in the cards and banking industry and are fully aware of the critical challenges that users face from a product perspective. Therefore, right from the beginning of its journey, the startup has been using predictive analysis, AI and ML to help solve issues that were previously manually accomplished by humans. “We are sitting on a data-mine and are working towards using them to create breakthrough offerings, as well as significantly improve the existing ones. All this is driven by market analysis through predictive modelling and AI\/ML. We study a user’s entire journey, the transaction patterns and overlay entry and exit points of our portal to understand their financial usage trends. We also study their historical transactions to understand them better,” according to Vishnoi. Fraud detection and data security (authentication, encryption, transaction approval checks, etc.) are some of the other vital aspects where EnKash has been working using AI\/ML. Currently, it is also working on a dynamic and efficient chatbot (full-scale robotic assistants) response and assistance system. “We are working on ML and fine-tuning it to best suit our workflows, processes and various policies to underwrite and do credit assessments. We, similarly, put our energies on ML for OCR of various KYC documents, invoices and expense receipts. Since the beginning, we have believed in automating systems internally as well as for our customers. Hence, AI\/ ML will always remain an integral part of our strategy. Our customers will also soon be exposed to AI\/ML for their invoice, payable and expense management, and complete digital onboarding,” Vishoi said. EnKash: The Tech Stack According to Vishnoi, for a technologically backed company, the tech stack determines the possibilities and limitations for building, validating and maintaining EnKash’s product. Therefore, the startup has been very careful to include state-of-the-art functions, data security, fraud protection, capacity and scalability in our technology. Since the startup focuses on creating a seamless experience for users, the startup is driven by APIs to create a dynamic and multi-function offering. This is only feasible because of dynamic frameworks of programming languages and ease of information transfer to and from its database. “At EnKash, we are PCI DSS certified and proud to say that our partner banks have trusted us because of this expertise. The tech stack is robustly built on microservices and is highly flexible and scalable. Thanks to our tech team, they always believed in building futuristic infrastructure and solutions fitting to business needs. Our tech team is working on the latest tech stack with cloud infrastructure on AWS, programming languages like JAVA\/ Angular, applications like Spring boot, cache frameworks like Redis, database as MySQL etc.,” he said. EnKash: The Roadmap Post the launch of EnKash’s Freedom Card, the market saw a few more similar launches by some neo-banking or payments platforms. However, EnKash remains wholly focused on being a cards company, whether it is through bank partnerships on commercial credit cards, or its own ‘Freedom’ corporate credit card. Vishnoi added, “In less than two years, EnKash has garnered an impressive clientele of more than 50,000 businesses that are using the product. We intend to grow multiple folds, given the market opportunity and cards evolving as a separate category in fintech space.”","excerpt":"EnKash is one of the most prominent startups in the commercial card fintech space of India, and the only card fintech startup which is working across segments. According to the company, it is the first in India to bring corporate credit cards with multiple variants in the market.  EnKash offers a wide range of products […]","categories":["Deep Tech"],"tags":["Startups"],"author_name":"Vishal Chawla","publish_date":"2020-04-15T11:00:00","publication_year":"2020","word_count":738,"keywords":["AWS","AI","Redis","ML","microservices","Aim","analytics","SQL","Startups","R","fraud detection"],"extracted_tech_keywords":["AI","ML","analytics","Aim","fraud detection","AWS","microservices","Redis","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/enkash-startup-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":65074,"title":"National Technology Day 2020: How Economy Is Rebooting Through Science, Tech &#038; Research Translations","content":"On the eve of National Technology Day 2020, the Confederation of Indian Industry (CII), Department of Science and Technology (DST), and Technology Development Board (TDB) came together to discuss the paradigm shifts in technologies such as industry 4.0, AI, machine learning, big data, autonomous vehicles, end of Moore’s law, the advent of quantum computing, super-skilling to trans-skilling, among others. Dr Harsh Vardhan, Union Minister of Science & Technology, Earth Sciences and Health & Family Welfare, addressed the digital conference emphasised that the S&T response reflects the collaborative spirit of the entire S&T ecosystem. The conference also included various sessions concluded by prominent members including Dr V K Saraswat, principal scientific advisor to the government of India, Prof. K. Vijay Raghavan, chief scientist of WHO, H.E. Mr Vincenzo de Luca, ambassador of Italy to India, among others. “In order to find solutions to combat this pandemic, the government of India, academia, researchers, scientists, startups, entrepreneurs as well as various industries have been working relentlessly,” said Vardhan. “We must appreciate the efforts of our scientists, entrepreneurs, including the institutions that are working hard to find quick and deployable solutions for Covid-19,” he added. On the theme for the National Technology Day this year, Dr Harsh Vardhan pointed out, “We need to mitigate the widespread economic impact and prepare for a stronger recovery using self-reliance as the new mantra. Thus, we look towards new opportunities to galvanise growth in the technological and industrial sectors.” The Theme The major objective of the aforesaid digital conference is to bring together global innovators, technology leaders, disruptive entrepreneurs, policymakers, industry stalwarts and eminent academicians to share their experiences and expertise to help turn the massive complexity into a meaningful change. During the talk, the speakers also stressed upon topics like the importance of manufacturing industries amid as well as after the pandemic. S. K. Das, Director and Professor, Indian Institute of Technology Ropar discussed the importance of teaching, learning and understanding amid the pandemic. He said that in this critical time, collaboration as well as cooperation, bring about the focus. Talking about the future of the education, he questioned what is beyond the regular education system and what about the rest of the talent pools that are being missed out. He further told that Tier I institutes such as IITs to have 8000-9000 students, and for the talented students who are left out must have some programs where the professionals and researchers can train them for creating a larger network of a talent pool to bring out industry 4.0. This can only be possible by the education system by outreaching to the students through open-source learning resources as well as extended online methods. Wrapping Up The National Technology Day has a historical perspective as it was on May 11, 1998 — India accomplished a major technological breakthrough by successfully carrying out nuclear tests at Pokhran. The celebration of Technology Day symbolises India’s quest for scientific inquiry, technological innovations and creativity, and the integration of these developments into national socio-economic benefits and global presence. Apart from the digital conference, “RE-START – Reboot the Economy through Science, Technology and Research Translations”, a virtual exposition has also been planned along with companies whose technologies have been supported by TDB. Various organisations and companies will be showcasing their products in the exposition through a digital B2B lounge. People from across the world can visit the stalls at the conference. Watch the video here:","excerpt":"On the eve of National Technology Day 2020, the Confederation of Indian Industry (CII), Department of Science and Technology (DST), and Technology Development Board (TDB) came together to discuss the paradigm shifts in technologies such as industry 4.0, AI, machine learning, big data, autonomous vehicles, end of Moore’s law, the advent of quantum computing, super-skilling […]","categories":["AI Features"],"tags":[],"author_name":"Ambika Choudhury","publish_date":"2020-05-12T18:00:00","publication_year":"2020","word_count":572,"keywords":["big data","Go","machine learning","AI","innovation","Git","RAG","ViT","GAN","R"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","Git","big data","GAN","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/national-technology-day-2020-how-economy-is-rebooting-through-science-tech-research-translations\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10005926,"title":"What Is The Hiring Process For Data Scientists At Micron","content":"With the core data science group based out of Boise, USA, Micron India team began operations from December 2019 to empower the company’s core engineering, manufacturing and business units to make data-driven decisions that are quick, accurate and insightful. With a centralised structure of the data science team, the company believes that an ideal data science candidate should be a digital-native, who is excited about the potential of data to deliver intelligence. Data Science Skill Sets At Micron Koushik Ragavan, Director Data Science at Micron India shares that a strong fundamental grasp of first principles and high levels of competency in applied statistics, computer programming, discrete simulation and database management is a must in the data science candidate. “We also look for an ability to communicate complex ideas, navigate Micron’s collaborative and multicultural global ecosystem. Original publications and patents will always get you to the front of the line for an experienced professional, and a great attitude with an ability to upskill quickly are key attributes we look for in a campus hire,” he added. In terms of educational background, Ragavan said that as they are beginning to build the team, their current focus is on recruiting candidates with substantial experience or ones with postgraduate or doctoral degrees in STEM. Further adding that skills weigh more than education background, Ragavan added that they recently on-boarded 92 candidates from 9 tier-one campuses from the 2019 campus season. They are further looking to offer 240 jobs in the 2020 season. “While we have found that the top-tier colleges provide a certain standard of talent, ensuring we get a Day-0 slot in the hiring season makes all the difference,” he said. Interview Process At Micron Ragavan believes that creating data science solutions is their primary responsibility. In order to have the biggest impact in terms of speed, accuracy, reliability, repeatability and scalability of their solution, the practitioners may have to don multiple hats including but not limited to data engineering, data characterisation, software development, development of UI\/UX, application support et al. Therefore they follow thorough and extensive screening to hire the right candidate. “While the weightage may differ from role to role, they are tested primarily for problem-solving, programming, statistics and AI\/ML concepts across four to five rounds of interviews with a globally diverse interview panel. For New College Graduates (NCG) we have an online assessment and two to three interviews,” he said. They ask questions related to the candidate’s current areas of interest\/experience, basic and advanced ML techniques, STAR (Situation, Task, Action Results) from their earlier experiences and more. Ragavan shares that while top-tier engineering colleges provide a large part of the sourcing pool, for experienced candidates they rely on Micron’s career portal, external job boards like LinkedIn and internal referrals from the primary pipelines. The best way to look for data science opportunities at Micron is to visit their career portal. Micron believes in recruiting for diversity. “Not recruiting for diversity would be short-sighted. People from different backgrounds ensure you have a wide range of ideas and viewpoints to choose from,” said Ragavan. Wrapping Up Having said that, recruiting data scientists has its own set of challenges. Ragavan shares that most candidates, both at the entry-level and experienced, tend to take the easy way out by becoming black box users of the methods and practices without understanding the methods and practices themselves. While they may be able to execute transactional solutions, they will not be able to develop strategic applications for high-value problems. It is extremely critical for them to go beyond desired outputs, to understand the “why, when, what, where, how” of a problem statement before they zero-in on the chosen approach,” he said. For anyone looking to venture into this field Ragavan advises to avoid being black box users. “Roll up your sleeves and get down into the trenches,” he said on a concluding note.","excerpt":"With the core data science group based out of Boise, USA, Micron India team began operations from December 2019 to empower the company’s core engineering, manufacturing and business units to make data-driven decisions that are quick, accurate and insightful. With a centralised structure of the data science team, the company believes that an ideal data […]","categories":["AI Hirings"],"tags":["impact of big data in education","trust your data"],"author_name":"Srishti Deoras","publish_date":"2020-09-02T14:00:33","publication_year":"2020","word_count":646,"keywords":["data science","Go","TPU","AI","ML","Scala","Git","RAG","data engineering","impact of big data in education","trust your data","R"],"extracted_tech_keywords":["AI","ML","data science","RAG","TPU","R","Go","Scala","Git","data engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/what-is-the-hiring-process-for-data-scientists-at-micron\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":58553,"title":"How Can Companies Outsource Analytics To India","content":"Spurred by the surge of big data and AI, organizations are increasingly outsourcing their analytics needs to drive better efficiency in their business functions. This trend can be attributed to the growing realisation that data by itself is ineffective without tools to interpret it in order to understand its applications. This has prompted companies to pursue superior data analytics capabilities. To stay competitive in the market and with an eye on long-term advantages, organisations are predicating their core business decisions on data analytics. However, assembling a talented analytics team is no mean feat, especially given that many companies lack the knowledge and experience to build such a team. This is driving companies — particularly those in the US, UK and Western Europe — to consider outsourcing their analytics needs to third party players with smart enterprise processes, particularly those in India. Vast availability of skilled labour, lower costs and contextual domain expertise has led to India being one of the top contenders for outsourcing analytics. According to a study conducted by AIM on the state of data science in the domestic Indian market, 70% of all large firms have adopted analytics in some form, with ICICI Bank, Flipkart, HDFC Bank, Axis Bank and Bharti Airtel having the biggest analytics units in the country. With many organisations accelerating their investment in analytics and looking to leverage offshore analytics firms to develop competitive capabilities, they need to take cognisance of a few things. First and foremost, they need to be careful about how they structure relationships when choosing their analytics providers and ensure that it matches their culture and business requirements. Although there is no collective consensus on the processes organizations should follow to create successful analytics partnerships, the following steps can offer guidance when choosing the right partner:- Step 1: Identify The Need The advantages of performing analytics in-house are numerous, but with the absence of required talent, it may be more prudent to outsource projects that your internal team may not be capable of performing well. This takes us to step one. Identify the problem or project that needs to be tackled and narrow down all options to accomplish that task internally. Once all options are exhausted and a consensus has been reached to outsource it, clearly articulate it for the purpose of step two. Step 2: Choose Your Vendor Once you have identified your need, develop a clear understanding of what you are looking for in a partner and prepare to deeply examine who is going to do what, and how each party can use the information it has. Some large organisations and new-age companies in India have set up dedicated analytics services units to maximise business impact. But when you choose your vendor, there are several things you need to consider:- Thorough Investigation Carry out a thorough investigation on firms that have experience in the sector you operate in. Check if they enjoy a good reputation, if their finances are sound, and if they have been in the data analytics space for some time. Knowledge about their previous or existing business partnerships will also give you a window into how compatible they could be with your business. Potential risks Assess the company’s data security and data breach preparedness. Acknowledging that this partnership will involve the exchange of sensitive and confidential information, organisations should rightly address issues regarding data security at an early stage of negotiations. What is more, organisations must be careful of not losing their core intellectual property (IP). One way to avoid accidents is by identifying, which capabilities are core, and which are more prudent to share with an offshore provider. Also, don’t forget to ask what processes they will follow to protect your IP. Analytics Penetration Understand the degree of infusion of the function of analytics in the organization. This is a well-accepted metric and can be quantified by calculating the approximate number of analytics professionals employed for every employee within the whole firm. For instance, a penetration of 1% should be read as 1 analytics per professional for every 100 employees. Analytics maturity This will give you a fairly good idea of the depth of analytics adopted by an organisation, and is a combination of three factors: tenure and seniority of employee, and percentage of AI in analytics function. Pointed Questions Companies that offer outsourced analytics services vary in their level of expertise and business approach. Asking pointed questions like: What are the capabilities your company can provide us? How will your business culture help make your company the right partner for us? It is also advisable to request for interactions with their customers. Right Data Analysis Specialists Although their work involves working with data, there are clear differences in the responsibilities, goals and skill sets of various people that make up a data analysis team. So how do you choose the right data analysis experts for your business? Understand what a data scientist, a BI consultant and a big data analyst can bring to the table. For instance, experienced data science specialists are proficient in data processing frameworks and are well-versed in data-driven predictive model development. On the other hand, BI consultants are skilled in data extraction, data aggregation and data visualisation that can generate valuable insights for your business. Lastly, a big data analyst should be able to carry out data management, efficient client analysis, risk detection, and market segmentation for your company. Manage Expectations Clarify your requirements and be transparent about your demands from the partnership. Agree on set timelines, and right at the offset, communicate clearly that the reports must be aggregated and presented in a manner that is comprehensible and can thus, be used to inform your decision-making. Step 3: Select Contract & Pricing Model There are two popularly accepted ways to go about this: one is the outcome-based model where the focus is on accomplishing a predetermined target, and the traditional model, where certain inputs are procured by the client for a specific time period. Between the two, the former might be a better option to opt for given that here, the overall responsibility to deliver the outcome rests on the vendor. Such an accountability is missing in the traditional approach, where vendors lack the necessary ownership to take necessary optimization. But you don’t need to stop at these two. There are various other engagement models for collaboration between a client and a vendor, and you can choose the one that best suits your business requirements:- Staff AugmentationThis allows you to extend your existing staff with experienced consultants and data scientists from the vendor. Preferred for short-term projects, this demands high client involvement to oversee the work of the augmented staff member.Project-basedThis time-bound engagement model is effective when there are few changes during the development process. It is also cost-effective since it operates around a fixed price, and properly leverages the technical expertise of the vendor by working towards well-defined set of deliverables.Offshore\/Nearshore Development CenterUndertaken for a wide variety of projects, this model is flexible in the sense that you can drive the requirements, while the vendor manages the offshore\/nearshore staff. CoEAn analytics centre of excellence (CoE) can be created by combining your staff and that of the vendor. It allows better collaboration and increased adoption of analytics.","excerpt":"Spurred by the surge of big data and AI, organizations are increasingly outsourcing their analytics needs to drive better efficiency in their business functions. This trend can be attributed to the growing realisation that data by itself is ineffective without tools to interpret it in order to understand its applications. This has prompted companies to […]","categories":["AI Features"],"tags":["analytics outsourcing","IT companies","Visual Analytics Provider"],"author_name":"Anu Thomas","publish_date":"2020-03-13T14:00:00","publication_year":"2020","word_count":1210,"keywords":["big data","IT companies","data science","Go","AI","data-driven","RAG","Aim","Visual Analytics Provider","analytics","GAN","analytics outsourcing","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","big data","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-can-companies-outsource-analytics-to-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":64853,"title":"Top 10 Papers From ICLR 2020 Conference","content":"International Conference on Learning Representations (ICLR), concluded last week, is one of the major AI conferences that take place every year. This year, ICLR went virtual because of the demanding circumstances. 687 out of 2594 papers made it to ICLR 2020 — a 26.5% acceptance rate. Here are a few of the top papers at ICLR: ALBERT: A Lite BERT Increasing model size when pretraining natural language representations often result in improved performance on downstream tasks. However, there are GPU\/TPU memory limitations and longer training times. To address these problems, this work presents two parameter-reduction techniques to lower memory consumption and increase the training speed of BERT. These proposed methods led to models that scale much better compared to the original BERT. The authors also use a self-supervised loss that focuses on modelling inter-sentence coherence and shows it consistently helps downstream tasks with multi-sentence inputs. As a result, this model establishes new state-of-the-art results on the GLUE, RACE, and squad benchmarks while having fewer parameters compared to BERT-large. Check the paper here. Plug and Play Language Models Plug and Play Language Models (PPLM) combines a pre-trained language model with one or more simple attribute classifiers that guide text generation without any further training. The attribute models consist of a user-specified bag of words or a single learned layer with 100,000 times fewer parameters. Model samples demonstrate control over sentiment styles, and extensive automated and human-annotated evaluations show attribute alignment and fluency. Check the paper here. Meta-Learning without Memorization Meta-learning is famous for leveraging data from previous tasks to enable efficient learning of new tasks. However, most meta-learning algorithms require meta-training tasks to be mutually exclusive, such that no single model can solve all of the tasks at once. In this paper, the authors address this challenge by designing a meta-regularization objective using information theory that places precedence on data-driven adaptation. By doing so, this algorithm successfully uses data from non-mutually-exclusive tasks to efficiently adapt to novel tasks. Check paper here. Reformer: The Efficient Transformer This paper introduces two techniques to improve the efficiency of Transformers. Furthermore, the authors used reversible residual layers instead of the standard residuals, which allows storing activations only once in the training process. The resulting model, the Reformer, performs on par with Transformer models while being much more memory-efficient and much faster on long sequences. Check paper here. Understanding the Effectiveness of MAML Model Agnostic Meta-Learning (MAML), a method that consists of two optimisation loops, with the outer loop finding a meta-initialisation, from which the inner loop can efficiently learn new tasks. Despite MAML’s popularity, its effectiveness is still questioned. The authors investigated this question, via ablation studies and analysis of the latent representations. The results show that feature reuse is the dominant factor and this led to ANIL (Almost No Inner Loop) algorithm, a simplification of MAML where the inner loop is removed for all but the (task-specific) head of the underlying neural network. Check the paper here. Your Classifier is Secretly an Energy-Based Model This paper proposes attempts to reinterpret a standard discriminative classifier as an energy-based model. In this setting, wrote the authors, the standard class probabilities can be easily computed as well as for unnormalized values. They demonstrated that energy-based training of the joint distribution improves calibration, robustness, handout-of-distribution detection while also enabling our models to generate samples rivalling the quality of recent GAN approaches. This work improves upon recently proposed techniques for scaling up the training of energy-based models and is the first to achieve performance rivalling the state-of-the-art in both generative and discriminative learning within one hybrid model. Check the paper here. Lottery Tickets In Reinforcement Learning AND NLP The authors evaluated whether “winning ticket” initializations exist in two different domains: natural language processing (NLP) and reinforcement learning (RL). For NLP, they examined both recurrent LSTM models and large-scale Transformer models. Whereas for RL, a number of discrete-action space tasks were analysed. Results suggest that the lottery ticket hypothesis is not restricted to supervised learning of natural images, but rather represents a broader phenomenon in DNNs. Check the paper here. On Identifiability in Transformers In this paper, the authors delved deep in the Transformer architecture by investigating two of its core components: self-attention and contextual embeddings. In particular, they examined the identifiability of attention weights, token embeddings, and the aggregation of context into hidden tokens. The authors demonstrated that, for sequences longer than the attention head dimension, attention weights are not identifiable. They also propose effective attention as a complementary tool for improving explanatory interpretations based on attention. Overall, this work shows that self-attention distributions are not directly interpretable and present tools to better understand and further investigate Transformer models. Check the paper here. Neural Tangents Neural Tangents is a library designed to enable research into infinite-width neural networks. It provides a high-level API for specifying complex and hierarchical neural network architectures. Neural Tangents provides tools to study gradient descent training dynamics of wide but finite networks in either function space or weight space. Check the paper here. Generalization through Memorization The authors in this work show how to efficiently scale up to larger training sets and allow for effective domain adaptation, by simply varying the nearest neighbor datastore, without further training. The model is particularly helpful in predicting rare patterns and the results strongly suggest that learning similarity between sequences of text is easier than predicting the next word. Check the paper here. There are other interesting works as well, covering which, is beyond the scope of this article. Please check the full list of papers here.","excerpt":"International Conference on Learning Representations (ICLR), concluded last week, is one of the major AI conferences that take place every year. This year, ICLR went virtual because of the demanding circumstances. 687 out of 2594 papers made it to ICLR 2020 — a 26.5% acceptance rate.  Here are a few of the top papers at […]","categories":["Deep Tech"],"tags":["research papers"],"author_name":"Ram Sagar","publish_date":"2020-05-08T15:00:12","publication_year":"2020","word_count":928,"keywords":["Go","API","TPU","AI","neural network","ML","research papers","Transformers","RAG","NLP","R"],"extracted_tech_keywords":["AI","ML","neural network","NLP","Transformers","RAG","TPU","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/iclr-2020-conference-papers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10118304,"title":"Google Introduces TransformerFAM, For Fixing Amnesia in LLMs","content":"A team of researchers from Google has introduced Feedback Attention Memory (FAM), a novel Transformer architecture that leverages a feedback loop to enable the network to attend to its own latent representations, fostering the emergence of working memory within the Transformer and allowing it to process indefinitely long sequences. Click here to check out the research paper. “In the film ’Memento’ (2000), the protagonist struggles with anterograde amnesia, which means he can not remember anything that happened in the last 10 minutes, but his long-term memory is intact, He has to tattoo important information on his body to remember it. This is similar to the current state of large language models (LLMs),” reads the paper. Similarly, current state-of-the-art large language models (LLMs) rely on attention mechanisms to extract meaningful representations from homogeneous data, but the quadratic complexity of attention with respect to context length limits the capability of modelling long contexts. To address these limitations, researchers have explored techniques such as sliding window attention, sparse attention, and linear approximated attention, though these methods have shown effectiveness below the 1B scale. The introduction of Feedback Attention Memory offers a new approach by adding feedback activations that feed contextual representation back into each block of sliding window attention. This enables integrated attention, block-wise updates, information compression, and global contextual storage. This innovative approach incorporates a feedback loop, which fosters the development of working memory within the Transformer architecture, allowing it to handle sequences of indefinite length. Notably, TransformerFAM can be seamlessly integrated with pre-trained models and does not require additional weights. The architecture has been tested across various model sizes, including 1B, 8B, and 24B, and has demonstrated significant improvements in long-context tasks such as NarrativeQA, Scrolls-Qasper, Scrolls-Quality, and XLSum. By effectively compressing and retaining important contextual information within extremely long contexts, TransformerFAM has shown enhanced performance compared to other configurations. The researchers emphasise the potential of TransformerFAM to empower LLMs to process sequences of unlimited length, which could revolutionise the way they handle long-context tasks and dependencies. The paper highlights that although traditional Recurrent Neural Networks (RNNs) rely on causal relationships between input sequences, Transformers can efficiently exploit the parallelism of machine learning accelerators. TransformerFAM’s feedback mechanism, which is limited to the relationship between blocks, does not compromise training efficiency and maintains performance levels similar to other architectures.Recently, Google researchers also introduced a method for scaling Transformer-based large language models (LLMs) to handle infinitely long inputs with bounded memory and computation.","excerpt":"The introduction of Feedback Attention Memory offers a new approach by adding feedback activations that feed contextual representation back into each block of sliding window attention.","categories":["AI News"],"tags":["Generative Pre-Trained Transformer","Google","LLMs"],"author_name":"Mohit Pandey","publish_date":"2024-04-16T16:16:58","publication_year":"2024","word_count":410,"keywords":["Go","machine learning","attention mechanism","AI","LLMs","neural network","ML","Transformers","RAG","Generative Pre-Trained Transformer","Google","transformer architecture","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","Transformers","RAG","R","Go","transformer architecture","attention mechanism"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-introduces-transformerfam-for-fixing-amnesia-in-llms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10085605,"title":"Say Goodbye to Long Delivery Times: Amazon Air is Now in India","content":"Amazon has launched its dedicated air cargo network in India, titled Amazon Air. First launched in 2016, this service has expanded from Amazon’s home base in the United States to Europe and now to India. The service focuses on building out Amazon’s logistics infrastructure and improving their last-mile connectivity. The company is yet to build out their fleet for operation in India, and are starting off with a Boeing 737-800, which has a cargo capacity of 23 tonnes and 141 cubic metres. This undertaking will be conducted in partnership with Quikjet, an Indian cargo airline that provides freighter services for e-commerce, air freight charters, and more. While the service first started with rented aircraft, Amazon Air operates over 110 planes today serving over 70 destinations worldwide. The service was inaugurated by Industries Minister KT Rama Rao at Rajiv Gandhi International Airport in Hyderabad. Amazon’s largest warehouses, campuses, and fulfilment centres in India are in Hyderabad, making it a natural choice for the air cargo service. The service is also expected to create additional jobs in the state. Speaking at the inauguration, the minister said, “We have been working towards strengthening the state’s air cargo infrastructure, and we welcome the launch of Amazon Air, which will further assist in making Hyderabad a cargo hub for the country and will create additional employment opportunities in the state.” Amazon’s presence in Telangana is only set to grow further, as AWS recently announced that they will be investing a total of INR 36,300 crore to bolster their data centre capabilities in the state. This investment will go towards a phase-wise strengthening of their three data centres located in Chandanvelly, FAB City and Pharma City. India is one of Amazon’s biggest markets. In FY22, Amazon Marketplace alone made over INR 21 crore in revenue, increasing from INR 16 crore in FY21. Amazon has also opened up its logistics networks to third-party merchants in India as part of an initiative to leverage their strong logistics networks as a business avenue. It is to be noted that the company has been shuttering its services like Amazon Distribution and Amazon Food. Amazon India has also laid off 1000 employees across Amazon Stores and people experience departments as part of their cost-cutting measures. Even as the company indulges in many cost-cutting measures, the move to bring Amazon Air to India shows their continued interest in the subcontinent. This is mainly due to the fact that the e-commerce giant has a network of 1.1 million sellers in the country and has long struggled to serve their 12 million products to the last mile in India.","excerpt":"Amazon has launched its dedicated air cargo network in India, titled Amazon Air. First launched in 2016, this service has expanded from Amazon’s home base in the United States to Europe and now to India. The service focuses on building out Amazon’s logistics infrastructure and improving their last-mile connectivity. The company is yet to build […]","categories":["AI News"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-01-23T15:46:14","publication_year":"2023","word_count":435,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","programming_languages:Go","RAG","ViT","GAN","R"],"extracted_tech_keywords":["AI","RAG","AWS","R","Go","GAN","ViT","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/say-goodbye-to-long-delivery-times-amazon-air-is-now-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001440,"title":"Top 5 Contributors To Quantum Computing In India","content":"Quantum computers have given a lot of hope to the world in terms of secure communication. Now, India too is getting in the competition to create secure and fast communications using quantum physics. Here are some of the companies in India working in the space of quantum computing. 1. QuNu Labs QuNu Labs, stands for Quantum Light Labs is a Bengaluru-based startup that works on quantum data security. QuNu Labs is India’s first and only company working in the area of quantum cryptography. They primarily deal with Quantum Key Distribution (QKD), which allows for the exchange of cryptographic key only between two people involved, with the help of encoded quantum bits, also called qubits. These exchange of these bits is governed by quantum physics and they are not hackable, helping to preserve the encryption between the two parties. The startup also plans to work in the sphere of QRNG, which is Quantum Random Number Generator, which deals with generating random numbers in hardware and has a huge role to play in quantum security. Apart from QKD and QRNG, it also plans to work in key management solutions, free space LiFi QKD and QKD for networks. 2. Automatski An R&D company with locations in Los Angeles and Bengaluru performs research in multiple areas, quantum computing being one of them. Automatski was originally registered in London in 2014 and moved to India in 2016. Automatski works in circuit quantum computers, adiabatic quantum computers and annealing quantum computers. They perform research into several areas of quantum-inspired software to simulate various quantum computing configurations. These quantum computing configurations include circuit-based quantum computer simulator, adiabatic quantum computer simulator and an annealing quantum computer simulator. They believe that their technology will allow them to simulate configurations with very large qubit counts. 3. Entanglement Partners Entanglement Partners is a consulting firm headquartered in Barcelona, Spain which provides consulting services in the space of quantum applications. It has resources in Kerala in India, apart from other locations of Madrid, San Jose and California. They also provide consulting services in the space of Events and Communications, Go-to-Market, Infrastructure, Strategic Consulting, and Venture Building. 4. IBM India IBM is now the first company in the world to release the first-ever commercial quantum computer this year called IBM Q. The IBM Q, according to IBM, is accessed using Qisit, a modular, open-source programming framework A worldwide network of Fortune 500 companies, academic institutions and startups use IBM Q technology and collaborate with IBM Research to advance quantum computing. In an interview with Arvind Krishna, Senior Vice-President, Hybrid Cloud, and Director of IBM Research, he said, “When we put the quantum computer on the cloud, we thought a few thousand trials will happen from the subset and we’ll be done. But 5 million experiments have been done from nearly 100,000 users so far.”, while talking about India’s interest in the field, and India was among the top 5 countries in terms of usage. 5. DST In late 2017, the Department of Science and Technology (DST) launched mission Quantum Science and Technology (QuST). DST wants to make a highly secure communication system and a general quantum computer that could replace the computer we know. It is learnt that the budget for the mission could be around ₹700 crore.","excerpt":"Quantum computers have given a lot of hope to the world in terms of secure communication. Now, India too is getting in the competition to create secure and fast communications using quantum physics. Here are some of the companies in India working in the space of quantum computing. 1. QuNu Labs QuNu Labs, stands for […]","categories":["AI News"],"tags":["countries with quantum computers","quantum cloud software","Quantum Computing","quantum computing companies","quantum computing programming","qubits","the quantum companies","what companies use quantum computers"],"author_name":"Disha Misal","publish_date":"2019-03-11T20:24:37","publication_year":"2019","word_count":545,"keywords":["Quantum Computing","Go","what companies use quantum computers","qubits","programming_languages:R","AI","countries with quantum computers","quantum computing companies","quantum cloud software","programming_languages:Go","the quantum companies","quantum computing programming","R","emerging_tech:quantum computing","startup"],"extracted_tech_keywords":["AI","R","Go","startup","programming_languages:R","programming_languages:Go","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/top-5-contributors-to-quantum-computing-in-india\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":23225,"title":"AI, Analytics Startup Synctag Raises Over ₹2 Crore In Series-A Funding","content":"Social media analytics and artificial intelligence startup Synctag has raised over ₹ 2 crore in the Series-A funding from a private equity firm Subhkam Ventures. Reportedly, Synctag are going to use the funds to for geographical expansion, as well as scaling up their portfolio of products and services. According to a report in an financial newspaper, the three-year-old company was not ready to unveil the exact amount of funding received. But another news website found out the same from the amount filed to the Registrar of Companies. The company, run by 25-year-old entrepreneur Harsh Mehta, will use the fund infusion to enter the New Zealand, Australia, Ireland and the United Kingdom markets. As of now, the company offers services such as content writing, artworks, images, promotion planning, competitor analysis, strategy, ad reports as well as research and branding. Reportedly, this is the fourth round of funding for Synctag, which manages social, email, content and other media for their clients. They had received two rounds of seed funding earlier in 2015 and a third round of funding valued at ₹ 35 crore in 2016. Synctag was founded in 2015. The company’s social media analytics platform aggregates content and feed from a user’s e-mail, RSS (Really Simple Syndication) feeds and other social media platforms and presents it on a single dashboard. The company’s tech platform can run on top of artificial-intelligence product suites like IBM Watson, show filings. Subhkam Ventures was established in 1998 by Rakesh S Kathotia, a first generation entrepreneur.","excerpt":"Social media analytics and artificial intelligence startup Synctag has raised over ₹ 2 crore in the Series-A funding from a private equity firm Subhkam Ventures. Reportedly, Synctag are going to use the funds to for geographical expansion, as well as scaling up their portfolio of products and services. According to a report in an financial […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","social media"],"author_name":"Prajakta Hebbar","publish_date":"2018-04-03T10:43:41","publication_year":"2018","word_count":250,"keywords":["Go","funding","artificial intelligence","programming_languages:R","AI","social media","programming_languages:Go","analytics","R","analytics platform","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","R","Go","analytics platform","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/synctag-startup-funding-ai-analytics\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":9114,"title":"How Analytics is used in Indian Politics: A Talk with Swaniti Initiative","content":"Indian politics has always been a subject of speculation and much has been said (mostly negative) on the Indian Political scenario. However, unlike the general notion, Indian politics is slowly undergoing positive changes too as politicians have realized the power of big data and analytics. In fact, recently, an increasing number of political leaders including Minister of State for Finance Jayant Sinha, lawmaker Jay Panda, Uttar Pradesh Chief Minister Akhilesh Yadav and Andhra Pradesh Chief Minister Chandrababu Naidu are adopting big data and analytics to strengthen their decision-making process and improve governance at national, state and constituency levels. To get the facts right, we spoke to Swaniti one of the organizations (amongst few others), that is at the core of this data revolution through it’s innovative projects on analyzing data for Indian government. Rwitwika Bhattacharya, founder, Swaniti Initiative Swaniti Initiative was founded in 2012 by Rwitwika Bhattacharya with an aim to deliver development solutions through government officials by initiating, implementing and accelerating development solutions in the sub-areas of health, education, gender, livelihood, water and sanitation. So far, Swaniti has worked with over a hundred Parliamentarians, fifteen state level legislators and four Chief Ministers (CMs) offices by either providing implementation support in the constituency, research support on key developmental programs or strategic data based insight to aid development strategy. She explained each programmatic area and the process as below. Swaniti deploys a team in the target constituency, supported by their Delhi office, to tackle specific problems set forth by government officials and\/or identified within the community. The engagements are typically three to twelve months and further split into three phases. In the first phase, data collection and analysis is conducted through fieldwork to understand challenges and solutions and best alternatives are recommended to clients, in second phase, they secure appropriate partners and support to deploy programs and eventually a full-scale solution and hand over controls to local stakeholders and lastly, they monitor the collect learning and survey effectiveness. Moreover, on a weekly basis, Swaniti releases policy updates and development briefs to MPs and MLAs. Additionally, Swaniti receives an average of 5 requests\/week for research support from elected officials, CMs offices and Ministries. Thus far, Swaniti has produced over two hundred briefs and memos, hosted multiple training and knowledge conferences and conducted numerous one-on-one MP knowledge consultations. The research provided is designed to be catalytic in bringing development because they emphasize on WHAT a policy\/scheme is and HOW to implement it successfully. To provide data and technology insights, Swaniti works with CMs offices, Ministries and Parliamentarians to collect, analyze and visualize data that is used to design the development programs. For example, Swaniti partnered with the Andhra Pradesh CM’s office to provide them data insights on agriculture that helped design a seed dissemination program. Additionally, Swaniti has also provided data and technology support to CMs office in Sikkim, Uttar and Arunachal Pradesh and Delhi. When questioned about how their work is affecting the Indian political scene; Rwitwika (Founder, Swaniti Initiative) replied that, even though there is a unanimous acknowledgment about the importance of data, yet, there are very few people working with the government and doing work related to it. She added that, currently, an MP relies on input from party workers or local leaders about the current state of development. Through our data dashboards, we bring in the much needed hard data to facilitate decision making amongst elected officials. So, now they have hard data on the delivery of public service. Thus, there is a significant shift coming in the governance space in India with an increased focus on data and technology. We then moved on to discuss the future of data analytics in Indian Politics and Rwitwika states, “the demand for data analytics will only increase in the governance space. However. the important piece in the midst of such demand will be to understand how to use this analytics.” “For example, we are fairly confident that a data dashboard will be useful in a constituency to understand the needs in the community; but what will be the various elements of the dashboard? How frequently will it be used? And what is its expected versus real impact? The answers to these questions have to be further refined as we continue to develop our process.” she said. ______ From the interview, it is quite evident that definitely political scenario is steadily upgrading itself and is on it’s way to change for the better. Thus, there is a lot of scope for similar initiatives that can drive India towards a better organised and data driven nation. However, what still remains to be seen is it’s application, i.e understanding problems based on real needs and our netas implementing practical solutions for them at grass-root levels. Though, we applaud those who are doing this, we eagerly wait for others to follow suit.","excerpt":"Indian politics has always been a subject of speculation and much has been said (mostly negative) on the Indian Political scenario. However, unlike the general notion, Indian politics is slowly undergoing positive changes too as politicians have realized the power of big data and analytics. In fact, recently, an increasing number of political leaders including […]","categories":["AI Features"],"tags":["Analytics Case Study","Interviews and Discussions"],"author_name":"Apoorva Verma","publish_date":"2016-02-24T07:01:50","publication_year":"2016","word_count":807,"keywords":["big data","Go","programming_languages:R","AI","ML","RAG","GAN","Aim","analytics","Analytics Case Study","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","analytics","Aim","RAG","R","Go","big data","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/swaniti-initiative\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10014478,"title":"India Becomes The Leading Adopter Of AI Amid The Pandemic &#8211; Where Are The Further Opportunities For Growth?","content":"The pandemic has presented a unique opportunity to India as it witnessed the highest increase in AI adoption compared to the major economies like the UK, US, and Japan. The report, published by PwC last week, found that 94% of organisations now believe that AI will help in creating more opportunities than be a threat to their industry, as the country has observed a 45% increase in AI adoption rates, amid the pandemic. The report was based on comprehensive interviews with CXOs and decision-makers of companies across the Indian market, as well as a larger global survey. Several traits highlight India’s AI adoption trends that present many opportunities for growth across sectors. Adoption Trends In India India still lags behind its counterparts when it comes to scaling up ‘AI across the organisation’. However, India has a higher percentage when it comes to ‘AI implementation in some business functions’ as compared to other countries. Also, more percentage of Indian enterprises are ‘considering the use of AI’ compared to the global average. This was mainly attributed to the rise in the number of use-cases after the pandemic, according to the report. AI-enabled tools helped manufacturing organisations to maintain hygiene at workplaces and work from home environments to function without the loss of productivity. At the same time, the data before the pandemic may fail to provide insights for the future as the pandemic continues to change the way the world functions, in so many ways. Hence to reduce costs, combat disruptions, and be future-ready AI adoption increased in India. Sectoral & Use-case Adoptions And Untapped Opportunities Sectors that were impacted the most by COVID showed more adoption since it became a necessity than an added advantage. Travel and hospitality, along with technology, media and telecommunication (TMT) sectors, had the highest adoption. The healthcare and pharma, along with TMT, were affected positively and adopted more solutions to seize the AI opportunities COVID-19 presented. On the other hand, financial services as well as travel and hospitality were adversely affected by the pandemic and used AI to address the uncertainties that followed. The retail and consumer industry, along with the industrial production industry both saw a lower adoption as compared to the other industries. The analysis also revealed that the ‘perception of benefit’ through AI is very high in industrial production and presents a lot of untapped potentials. Hence, the next immediate opportunities could be in industrial production as the sector realises the benefits of AI and is yet to adopt it on a larger scale. This could be followed by the retail and consumer industry as it realises potential benefits. In terms of use-cases, the PwC saw a massive adoption in front-office applications like customer service, chatbots, and UX personalisation, since customers could not be met in person. Front-end and back-end office workplace processes were also automated to a large extent using AI. The report analysed the current use cases as well as the ones that will be adopted in the next two years. Early adopters of applications with low present use and low near-future use can gain a significant advantage over other firms. Thus, adoption of AI by companies for applications like customer profiling, quality checking, due diligence, compliance monitoring, document classification, among others presents a great opportunity for growth while also gaining a significant advantage. Sector-wise, only 10% of industrial production companies are exploring the use of intelligent sanctions and compliance monitoring, and only 11% in retail and consumer markets are looking to explore AI in customer profiling. Also, there are only 12% of TMT companies exploring smart ticket management, and only 20% in the travel and hospitality industry are exploring targeted marketing. Conclusion: The Way Forward As organisations mature in adopting AI in India, the problems in the field have shifted from purely technical and data-related constraints. They are more about finding the right use-cases and implementing them to make projects profitable. While the article presents untapped potential across sectors and use-cases, the report recommends early identification of realistic, achievable, and quick-win applications to develop AI. At the same time, the report highlights the importance of adopting AI-automation in a developing country like India, with a focus on augmenting workforce capabilities. AI application to reduce cost purely for economic gains at the social cost of workforce displacement might be short-lived and unsustainable.","excerpt":"The pandemic has presented a unique opportunity to India as it witnessed the highest increase in AI adoption compared to the major economies like the UK, US, and Japan. The report, published by PwC last week, found that 94% of organisations now believe that AI will help in creating more opportunities than be a threat […]","categories":["AI Features"],"tags":["AI adoption India","AI in India"],"author_name":"Kashyap Raibagi","publish_date":"2020-12-16T13:00:00","publication_year":"2020","word_count":721,"keywords":["AI in India","programming_languages:R","AI","chatbots","RPA","RAG","AI adoption India","automation","ViT","disruption","GAN","R"],"extracted_tech_keywords":["AI","RAG","chatbots","R","GAN","ViT","automation","RPA","disruption","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/india-becomes-the-leading-adopter-of-ai-amid-the-pandemic-where-are-the-further-opportunities-for-growth\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10067209,"title":"Red Hat open sources the security platform it acquired last year","content":"Red Hat has announced that Red Hat Advanced Cluster Security for Kubernetes (RHACS) has been made open-source as StackRox. The Kubernetes and container security community can now use and contribute to the codebase of StackRox on GitHub. In February last year, Red Hat closed the transaction to acquire StackRox. At that time, Red Hat said StackRox would help to simplify DevSecOps, and its integration into Red Hat OpenShift will help users to enhance cloud-native application security across every IT footprint. “Red Hat has always maintained its commitment to supporting the open-source community, and with this news, we’re enabling developer and security teams to deliver more secure applications faster,” added the company in a blog. Red Hat lists down a few areas where StackRox will be helpful to operationalise and implementing security for its supply chain, infrastructure, and workloads. It helps to integrate security into the CI\/CD pipelines and image registries to provide continuous image scanning and assurance. Red Hat says that StackRox can help to prevent configuration drift by compliance checks against CIS benchmarks or custom policies. It also analyses existing rules for role-based access control (RBAC) to prevent insecure access and authorisations. It prevents high-risk workloads from deploying or running, using out-of-the-box deploy-time and runtime policies, and monitors known good behaviour to configure custom policies and alerts for anomalous and malicious behaviour.","excerpt":"Red Hat closed the transaction to acquire StackRox in February last year.","categories":["AI News"],"tags":["Mergers and Acquisitions","red hat"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-05-17T18:07:10","publication_year":"2022","word_count":223,"keywords":["Go","red hat","programming_languages:R","AI","CI\/CD","programming_languages:Go","Git","GitHub","Mergers and Acquisitions","R","kubernetes"],"extracted_tech_keywords":["AI","kubernetes","R","Go","Git","GitHub","CI\/CD","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/red-hat-open-sources-the-security-platform-it-acquired-last-year\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10064181,"title":"A hands-on guide to anomaly detection in time series using ADTK","content":"Anomaly detection is an important part of machine learning that makes the results unbiased to any category or class. While in time series modelling it takes a very important place because there is a variety of anomalies that can be there in time-series data. These anomalies may include seasonal anomalies, regression anomalies, quantile anomalies, etc. ADTK is a toolkit that mainly helps in detecting all types of anomalies. Using this package, we can also perform a variety of procedures that come under time series modelling. In this article, we are going to discuss ADTK with hands-on implementation of anomaly detection. The major points to be discussed in the article are listed below. Table of Contents What is ADTK?Implementing anomaly detection with ADTKLoading data Data preprocessingData validation and visualization  Seasonal anomaly detectionThreshold anomaly detection Let’s start by introducing ADTK. What is ADTK? ADTK is an open-source python package for time series anomaly detection. The name ADTK stands for Anomaly detection toolkit. This package is developed by ARUNDO. ARUNDO was founded to solve IoT challenges at the industry level. Talking about the ADTK, its features enable us to implement pragmatic models very easily, and also these features make ADTK different from other anomaly detection tools. The aim behind the development of the ADTK is to promote best practices in solving anomaly detection problems in the real world. Using this toolkit we can build a rule-based\/unsupervised model. Since we barely find anomalous data in historical data and we face problems in building the supervised models and also there is always a requirement to detect anomalies to understand what type of events are of interest. Just by knowing about the rules for those events, we can convert the rules into models using the ADTK. There are three main elements of the ADTK: Outlier detector: this element helps in detecting the outliers in the data. There are various modules under this element that help in every type of detection in time series like threshold detection, seasonal detection,  regression detection, etc. Transformer: These elements help in transforming data. There are various modules under this class like modules for rolling aggregation, seasonal decomposition, etc.Rule chain aggregator: These elements help in identifying and aggregating time points set under any rules. We use these functions to aggregate time series using some rules defined by us. With these elements, this package also provides some other elements for making pipelines, evaluating models, and preprocessing the data. We can install this package using the following lines of codes. !pip install adtk After installation, we can use this toolkit for anomaly detection in time series. Let’s see some of the basic things that we can do using ADTK. Implementing anomaly detection with ADTK In this section, we will look at the implementation of some modules developed under ADTK packages. Loading data Before applying modules we are required to have some time-series data. For this purpose, we are going to use a package named Yfinance to extract share price data from the State Bank of India. Using the below codes we can do that. import datetime as dt from datetime import datetime as dt from dateutil.relativedelta import relativedelta import yfinance as yf end = dt.today() start = dt.today() - relativedelta(years=1) data = yf.download('SBIN.NS', start, end) data.tail() Output: Here we can see that in the data we have share price opening, high, low, and closing, and volume of the share as variables. Data preprocessing One thing that is missing with this data is some dates because there are some weekends when the share market does not open. So before going on to the further step we are required to fill this gap but before this, we are required to know how many dates are missing in the data. pd.date_range(start = start, end = end ).difference(data.index) Output: So there are 234 dates missing in two years of data. Here we are going to fill dates with values that accrued on the last date. idx = pd.date_range(start, end) data = data.reindex(idx) data.fillna(method=\"ffill\", inplace = True) Let’s check again if there any dates are missing or not. pd.date_range(start = start, end = end ).difference(data.index) Output: Here we can see that now all the dates are there. Let’s validate the data. Data validation and visualization To work with the modules of the ADTK we are required to validate the data with its packages. from adtk.data import validate_series data = validate_series(data) print(data) Output: Here we can see our validated data. Let’s plot the data to know how values are varying according to time. from adtk.visualization import plot plot(data) Output: Here we can see full information about the variation in data. And one thing we can easily say is that there are Anomalies in the volume variable. Seasonal anomaly detection Let’s just start by detecting violations of seasonal patterns by anomalies and volume variables. from adtk.detector import SeasonalAD seasonal_vol = SeasonalAD() anomalies = seasonal_vol.fit_detect(data['Volume']) anomalies.value_counts() Output: Here we can say in two years there are 19 anomalies that are present in the volume variable. We can also plot this using the following codes. plot(data, anomaly=anomalies, anomaly_color=\"orange\", anomaly_tag=\"marker\") Output: Here we can see where the anomalies of the volume variable are and whether they are affecting the time series or not. Threshold anomaly detection We can also define a threshold range and detect points where the time series is going out of this threshold value. Let’s do this for the close column that shows the closing share price of the day. print('Average closing price', data['Close'].mean()) print('Minimum closing price', data['Close'].min()) print('Maximum closing price',data['Close'].max()) Output: Here we can see what is the average, minimum, TFTand maximum closing prices of the shares. By looking at this we can say values above 530 and below 180 are anomalies and should not be taken in modelling. We can detect these values in the following ways: from adtk.detector import ThresholdAD threshold_val = ThresholdAD(high=530, low=180) anomalies_thresh = threshold_val.detect(data['Close']) Let’s check what are the counts of anomalies are. anomalies_thresh.value_counts() Output: Here we can say there are 65 anomalies. Let’s visualize them. from adtk.visualization import plot plot(data, anomaly=anomalies_thresh, ts_linewidth=1, ts_markersize=3, anomaly_markersize=5, anomaly_color='black'); Output: Here we can see the anomalies according to the defined threshold on the closing price of the share. Also, we can check whether these values are affecting the other variable or not. Final word In this article, we have discussed the ADTK(anomaly detection toolkit) which has a variety of modules for anomaly detection and modelling in time series. We can find the whole module list here. Along with this, we look at some modules that helped us in detecting anomalies according to seasonal patterns and defined threshold patterns. References Link for the above codes ADTK documentation","excerpt":"ADTK is an open-source python package for time series anomaly detection. The name ADTK stands for Anomaly detection toolkit. This package is developed by ARUNDO. Its features enable us to implement pragmatic models very easily, and also these features make ADTK different from other anomaly detection tools.","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning","Python"],"author_name":"Yugesh Verma","publish_date":"2022-04-03T13:00:00","publication_year":"2022","word_count":1108,"keywords":["Go","machine learning","TPU","ELT","AI","Machine Learning","RAG","Python","Aim","anomaly detection","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","anomaly detection","TPU","Python","R","Go","ELT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-hands-on-guide-to-anomaly-detection-in-time-series-using-adtk\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10096869,"title":"Tesla Plans To Set Up Factory In India, A Month After Musk Met PM Modi","content":"After years of delaying Elon Musk’s brainchild Tesla is in talks with the Indian government for an investment proposal to set up a car factory in the subcontinent, with a yearly capacity of up to 500,000 electric vehicles. The starting price of the vehicles will be 2 million rupees ($24,400.66), the Times of India reported. The company has been eyeing the Indian economy for years but the plans were hunky dory since it failed to get special incentives from the GOI for importing cars at a lower duty. In return the government has been adamantly demanding the Musk led firm to manufacture cars locally instead of importing from other countries —- like China. But then the company wanted to export its cars first to test the demand strength. Tesla plans to export cars from India to test demand in the Indo-Pacific region, according to government sources cited in a report by The Times of India. The turn around in the company’s plans to invest in the country comes a month after Musk met with the Indian Prime Minister Narendra Modi. I am confident Tesla will be in India and we will do so as soon as humanly possible,” said Musk post the visit. In the meeting, Modi pushed the car maker to make a “significant investment” in the country. In the latest efforts to enter the domestic market, the AI company held discussions in May with government officials about incentives being offered by the government, Reuters reported. Using India as an export base allows Tesla to assess market potential and consumer preferences before expanding locally. It also takes advantage of India’s growing economy and focuses on a sustainable energy future including stationary battery packs and electric vehicles. After meeting Modi, Musk also hopes to bring SpaceX’s Starlink satellite internet service to the country.","excerpt":"The Musk owned company has a history of being rejected by the Indian government due to its demand for special incentives.","categories":["AI News"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-07-13T14:58:03","publication_year":"2023","word_count":303,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tesla-to-set-up-factory-in-india-a-month-after-musk-met-modi\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10065946,"title":"Meta AI is building AI that processes language like the human brain","content":"Meta AI announced a long-term research initiative to understand how the human brain processes language. In collaboration with neuroimaging centre Neurospin (CEA) and INRIA, Meta AI is comparing how AI language models and the brain respond to the same spoken or written sentences. Over the past two years, Meta AI has applied deep learning techniques to public neuroimaging data sets to analyse how the brain processes words and sentences. The data sets were collected and shared by several academic institutions, including Max Planck Institute for Psycholinguistics and Princeton University. The comparison between brains and language models has already led to valuable insights that include: Language models that closely resemble brain activity best predict the next word from context. Prediction based on partially observable inputs is at the core of self-supervised learning (SSL) in AI and may be key to how people learn the language.Specific regions in the brain anticipate words and ideas far ahead in time, while most language models today are typically trained to predict the very next word. Unlocking this long-range forecasting capability could help improve modern AI and language models. “Of course, we’re only scratching the surface — there’s still a lot we don’t understand about how the brain functions, and our research is ongoing.”Meta AI Meta AI’s collaborators at NeuroSpin are now creating an original neuroimaging data set to expand this research. They will be open-sourcing the data set, deep learning models, code, and research papers resulting from this effort to help spur discoveries in both AI and neuroscience communities. This work is part of Meta AI’s broader investments toward human-level AI that learns from limited to no supervision.","excerpt":"Meta is studying the brain to build AI that processes language as people do.","categories":["AI News"],"tags":["Deep Learning Techniques","human brain","language","Language Models","Meta AI"],"author_name":"Poornima Nataraj","publish_date":"2022-04-29T13:57:40","publication_year":"2022","word_count":273,"keywords":["Go","self-supervised learning","Meta AI","programming_languages:R","AI","language","programming_languages:Go","deep learning","ViT","Language Models","Deep Learning Techniques","human brain","R"],"extracted_tech_keywords":["AI","deep learning","Meta AI","R","Go","ViT","self-supervised learning","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-ai-is-building-ai\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":43878,"title":"How Machine Learning-Aided VR Is Helping Neurosurgeons Train Better","content":"With improved accessibility to new technologies, there has never been a better time to include a plethora of studies into the field of artificial intelligence. It has the potential to pave the way for developing tools that can serve the clinicians and researchers greatly with effective training. One such ingenious utility emerged out of the research carried out at Neurosurgical Simulation and Artificial Intelligence Learning Centre at the Montreal Neurological Institute and McGill University Health Centre, where the researchers used machine learning algorithms to guide virtual reality training platforms for neurosurgeons. Since brain-related diagnoses or cures are always critical, a lot depends on the skill of the neurosurgeon. The feedback provided by this platform helped to gather more insights about the expertise of the surgeons. ML Powered Neuro Expertise Photo by CHRISTINNE MUSCHI for THE GLOBE AND MAIL The level of expertise in using miniature surgical tools (psychomotor skills) varies from surgeon to surgeon. The pressure applied to the brain should be such that the damage is minimal. Algorithmic assisted training exposes the areas of improvement for the surgeons and they can build their skills on these positive feedback loops. For this study, a total of 50 individuals (14 neurosurgeons, 4 fellows, 10 senior residents, 10 junior residents, and 12 medical students) were chosen who participated in 250 simulated tumour resections. The simulator used for this experiment was the NeuroVR (CAE Healthcare) is a high-fidelity neurosurgical simulator designed to recreate the visual and haptic experience of resecting a human brain tumour through an operative microscope. To evaluate the performance of the surgeons few metrics were categorized based on the force applied to the underlying structures and damage to the underlying brain, blood loss, and quantity of tumour removed. The level of expertise of the surgeons can then be determined accurately by looking at how well they score with machine learning models. Factors such as the way the tip of the instrument accelerates play a crucial role in determining the expertise. So, in the first step of the experiment, raw data were transformed into performance metrics to be used by the algorithm. This process includes transforming instrument movement from the original x, y, and z coordinates into three-dimensional representations of velocity, acceleration, and jerk, as well as the rate of change in volume of tumour and healthy tissue, as well as the rate of change of bleeding, and the number of attempts to stop bleeding were generated. The results show that the K-nearest neighbour algorithm had an accuracy of 90% (45 of 50), the naive Bayes algorithm had an accuracy of 84% (42 of 50), the discriminant analysis algorithm had an accuracy of 78% (39 of 50), and the support vector machine algorithm had an accuracy of 76% (38 of 50). “Physician educators are facing increased time pressure to balance their commitment to both patients and learners,” says Dr Rolando Del Maestro, the lead author of the study. The researchers, however, still warn us from jumping to conclusions unless until these models are tested on novel data for overfitting. They feel that a more comprehensive evaluation of participants with an emphasis on demonstrated skills across assessment domains (eg: visual rating scales and training evaluations or assessment of visuospatial abilities) may result in improved algorithm performance. AI For A Clinical Future Although the task involved a complex neurosurgical tumour resection task, the protocol outlined can be applied to any digitized platform to assess performance in a setting in which technical skills are paramount. The researchers firmly believe that by understanding the performance data used by the algorithm to render its decision, it is possible to design systems to deliver on-demand assessments at the convenience of the examinee and with minimal input from skilled instructors. Such systems may be subject to continuous improvement as increasing participant data are collected and integrated into the algorithm. Simulators, while affording learners the opportunity to safely develop technical skills during the particularly dangerous and error-prone early phases of skill acquisition, do not obviate the need for learner feedback, which is often given by skilled instructors. AI in critical areas should be considered only as an augmentation to the traditional methods and not as a tool to promote their abandonment because human assistance in achieving clinical expertise is unparalleled and it would be safe to say that it will remain to be so.","excerpt":"With improved accessibility to new technologies, there has never been a better time to include a plethora of studies into the field of artificial intelligence. It has the potential to pave the way for developing tools that can serve the clinicians and researchers greatly with effective training. One such ingenious utility emerged out of the […]","categories":["Deep Tech"],"tags":["human intelligence at machine scale","Machine Learning","neuroscience"],"author_name":"Ram Sagar","publish_date":"2019-08-06T14:28:23","publication_year":"2019","word_count":725,"keywords":["Go","artificial intelligence","human intelligence at machine scale","machine learning","AI","programming_languages:R","ML","Machine Learning","Git","programming_languages:Go","neuroscience","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-machine-learning-aided-vr-is-helping-neurosurgeons-train-better\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10005735,"title":"The Tech Behind Axis Bank’s Multilingual Voice Bot","content":"Understanding the ever-changing customer aspirations and requirements will always be a challenge. However, the larger issue at hand lies in to bring the customers at par with the adoption of digital technology. The COVID era has emulated a significant impact at different levels of the organisation across businesses. The scenario at Axis Bank was no different as the status quo, and modus operandi across the bank was challenged during the initial phase of lockdown. The bank has observed extreme spikes in volumes, increment in complaints and escalations transpired by customer anxiety, which was a repercussion of limited services to customers during the lockdown phase. This reflected in Axis Bank’s customer experience index and the cross-functional service levels of the organisation. And thus, it became critical for the bank to re-define their role in the life of their customers by elevating their digital banking experiences to new domains of customer service. Also Read: What Does It Take To Design The Next Generation Chatbots AI-Powered Conversational Voice BOT To The Rescue In order to address the increasing number of queries from customers more effectively and promptly amid the pandemic, Axis Bank decided to partner with an AI-based SaaS voice automation platform — Vernacular.ai, to optimise voice AI solutions and automate their customer interactions via an intelligent human-like dialogue. The company decided to create a next-gen multilingual voice bot — AXAA, for Axis Bank that can converse in English, Hindi as well as Hinglish. This bot leverages cutting edge automated speech recognition and natural language understanding technology boosted by AI-powered business algorithms to help accelerate the engagement strategy of the bank. Explaining further, Sourabh Gupta, the co-founder and CEO of Vernacular.ai stated that the company uses all variants of standard sentence semantics techniques for the bot since the compute and performance requirements for many of the deployments can vary. For instance — a heavy neural sentence embedder for less loaded cases, but a much faster and simpler system for the opposite situation. They, in turn, designed AXAA to provide an augmented customer experience which can automate the contact centre operations and is capable of handling lacs of customer queries and requests per day with the ability of faster scalability. This automated voice assistant has been deployed to have an in-depth understanding of their queries, context and the intent of the call. “Additionally, since the production environment will not give an ideal scenario, we also build custom models to work with the output of speech recognition systems which include real-life errors, such as, general noise and speech disfluencies,” said Gupta. Also Read: How This NLP-Driven Literature Search Engine Can Help With COVID Information The Tech Behind Vernacular.ai started with an internal step called data-farming, which plugs into different sources of the conversational data repositories, and involves on-premise data sourcing to utilise previous conversations. To facilitate the process, the company gets the data points annotated by an active-learning based system that picks data points based on their performance. “Model training process requires an overall system training which involves training the language-model and word embeddings for the business use-cases,” said Gupta. The bot used on-premise data sourcing, along with querying based on heuristics of model performance; “we pick data points that performed poorly,” said Gupta. “We used LM tuning, ASR Release, neural network, and resolution metrics for word embeddings.” Furthermore, Vernacular.ai used state-of-the-art, deep neural networks trained on thousands of hours of acoustic data for text to speech recognition. And, to identify the pauses in between people communicating, the company employed recurrent models which worked on real-time audios with minimal latency. Alongside, the idiolect perception layer, which involves identifying various stylistic components from speech signals of utterances, includes identifying speaker characteristics like gender, age, location, etc. Post that, the modelling usage patterns in all the languages covers conversational nuances and semantic equivalence of words and phrases. The company also involved text to speech models where the systems are trained to synthesise audio which replicate nuances of intelligent human-like dialogue. “To discover novel user patterns in deployed systems, we used high recall sentinel models which helped in understanding new behaviour and handling concept drift during the lifetime of the product,” said Gupta. Also Read: How AI-Powered Analytics & Monitoring System Helps In Reducing Water Losses Benefits of AXAA In Customer Retention Talking about the benefits, Ratan Kesh, the EVP and Head – Retail Operations and Service of Axis Bank said, “The launch of AXAA, in July’20, is in line with the bank’s philosophy of building a sharper customer focus and embark on a journey of constant innovation and enhancement.” “With the capability of operating like a humanoid, AXAA has the power to change the paradigm of customer experience,” said Kesh. Till date, the AI Voice BOT has catered to more than 2.23 million customers at an industry best success rate of an average of 85% and above. There were also approximately 65%-70% of customers who were being touched by the BOT and attempted to connect with the contact centre. In case, the bot — AXAA is unable to service or cater to a particular customer query, it has been designed to direct the call immediately to one of the human experts, minimising the navigation time on the conventional IVR. “AXAA has been working side by side with expert customer service officers in delivering a consistent and superior experience to customers,” said Kesh. Taking digital banking to the next level, “moving forward to the next phase, we are working on expanding the language capabilities to 10 Indian languages and add 27 more self-service options,” concluded Gupta.","excerpt":"Understanding the ever-changing customer aspirations and requirements will always be a challenge. However, the larger issue at hand lies in to bring the customers at par with the adoption of digital technology. The COVID era has emulated a significant impact at different levels of the organisation across businesses. The scenario at Axis Bank was no […]","categories":["AI Features"],"tags":["Conversational AI","intelligent automation for retail"],"author_name":"Sejuti Das","publish_date":"2020-08-31T11:00:00","publication_year":"2020","word_count":928,"keywords":["Go","TPU","AI","neural network","chatbots","Scala","intelligent automation for retail","NLP","RAG","Conversational AI","analytics","R"],"extracted_tech_keywords":["AI","neural network","NLP","analytics","RAG","chatbots","TPU","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-tech-behind-axis-banks-multilingual-voice-bot\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10018987,"title":"Why Did Google Prick Its Balloon Internet Idea","content":"Google’s parent company Alphabet said it would close down Loon; the moonshot project it launched in 2011. Loon sought to use high-flying balloons to provide internet access in places where conventional infrastructure for web connectivity was not feasible. Loon CEO Alastair Westgarth said failure to bring down ‘costs low enough to build a long-term, sustainable business’ led to the project’s winding up. What Is Loon? Loon was born out of the dire need to take the internet to the ‘next billion’ users. The idea was to level the playing field for underserved communities by hooking them to the internet to catch up with the rest of the world. Google X, Alphabet’s moonshot factory, had set up a unit to take on the challenge of reaching the last billion users, especially people who live in remote areas. The team came up with the idea of floating tennis-court sized balloons in the stratosphere to help local telecom companies provide internet at a cost much lower than building physical infrastructure. The balloons are made of polyethylene sheets powered by solar panels while being controlled by software and systems stationed on the ground. Often referred to as ‘floating cell towers’, the balloons transmit internet signals to the ground stations and other personal devices. Each of these balloons lasts for anywhere from 100 to 150 days. In 2013, the first of such balloons was launched in New Zealand. One of the most recent and significant deployments of Loon balloons was at Peru, where after a massive earthquake of 8.0 magnitude, Loon brought LTE service to the affected areas through collaboration with Telefonica, the Spain-headquartered telecommunications company. Loon had also announced it would bring internet connectivity to the secluded portions of the Amazon had said it’s all set to launch 4G in Kenya in its last major update. The project trialled with 35,000 customers, with download speeds of 18.9Mbps and upload speeds of 4.7Mbps. Loon also announced that 35 solar-powered balloons would be in operation over the East African regions. What Went Wrong? Recently, Loon made TIME magazine’s best 100 inventions of 2020 list. So, what exactly went wrong? In 2019, a Reuters report pointed to the troubles brewing with Alphabet’s project. Many potential customers had expressed their reservations about the viability of the technology. Also, Loon was allegedly facing a fund crunch. A recent report suggested that Loon had run out of money and had to turn to Alphabet to keep its business floating. The last major external investment ($125 million) in Loon was made in 2019 by a Softbank subsidiary, HAPSMobile. Loon was part of Google X’s Other Bets — Alphabet’s stand-alone subsidiaries that raise external fundings. The main objective was to eventually hive off these initiatives\/projects as separate companies. Westgarth, in his farewell blog post, wrote: “While we’ve found a number of willing partners along the way, we haven’t found a way to get the costs low enough to build a long-term, sustainable business. Developing radical new technology is inherently risky, but that doesn’t make breaking this news any easier. Today, I’m sad to share that Loon will be winding down.” Wrapping Up Loon’s spokesperson announced some of its employees would be moved to other roles at X, Google and Alphabet. However, the project has assigned a small team to ensure Loon’s smooth wrap-up, including its Kenya project.","excerpt":"Google’s parent company Alphabet said it would close down Loon; the moonshot project it launched in 2011. Loon sought to use high-flying balloons to provide internet access in places where conventional infrastructure for web connectivity was not feasible. Loon CEO Alastair Westgarth said failure to bring down ‘costs low enough to build a long-term, sustainable […]","categories":["Global Tech"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-01-27T16:00:00","publication_year":"2021","word_count":555,"keywords":["Go","funding","programming_languages:R","AI","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","ViT","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-did-google-prick-its-balloon-internet-idea\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10144143,"title":"Why GenAI is India&#8217;s High-Performance, Low-Cost Game Changer","content":"Generative AI is at the forefront of India’s tech evolution, delivering great performance without the steep price tag. This has been possible thanks to some big players who continue to up their game. With resilient infrastructure, cutting-edge technology, and strategic investments, cloud giant AWS is helping Indian businesses scale AI while redefining price-performance benchmarks. Satinder Pal Singh, head of solution architecture for AWS India and South Asia, leads a team of architects tasked with helping customers design scalable, efficient solutions using AWS services. During an in-depth conversation with AIM at AWS re:Invent 2024 held in Las Vegas recently, Singh shared insights into AWS’s strategic focus on India, customer needs, and groundbreaking launches that promise to redefine cloud technology in the region. “India is a significant country for us,” Singh began, emphasising AWS’s extensive infrastructure in the country. With two operational regions—Mumbai and Hyderabad—spanning three availability zones, 33 points of presence, and nine direct connect locations, Singh noted, “We provide extensive capabilities to our customers, enabling them to leverage more than 240 services—from infrastructure to analytics, IoT, and machine learning—while keeping their data within India.” Custom Chips: Redefining Price-Performance AWS’s investments in custom-built chips—Graviton, Trainium, and Inferentia—have been a game-changer for its customers. Graviton chips, designed for general workloads, have already delivered significant benefits. Companies like Zomato and Paytm reported reduced costs by up to 30% and improved performance by 20-35%, a testament to the efficiency of these chips. Trainium and Inferentia, on the other hand, cater specifically to AI workloads. Trainium delivers a 64% performance boost compared to previous offerings, while Inferentia is optimised for inference, helping global clients save up to 40% on costs. Singh acknowledged that these advancements were critical for Indian enterprises navigating the challenges of scaling AI while managing budgets. Upskilling India: AWS AI-Ready Program One of the most pressing challenges for Indian organisations is the AI talent gap. Singh cited a startling statistic—while 99% of employers envision their companies becoming AI-driven organisations by 2028, nearly 79% struggle to find skilled professionals. To address this, AWS launched the AI-Ready Program, a comprehensive initiative to upskill over 2 million individuals globally by 2025. In India alone, over 5.9 million individuals have been trained in cloud services since 2017. AWS’s investment in India’s local cloud infrastructure is projected to reach $16.4 billion by 2030, supporting 131,700 jobs annually and contributing $23.3 billion to India’s GDP by 2030. Why Mumbai and Hyderabad? The decision to establish data centres in Mumbai and Hyderabad was driven by performance and reliability considerations while providing multi-region availability. AWS’s infrastructure is built with resilience at its core, comprising multiple availability zones, which are essentially isolated data centres capable of functioning independently. This design has helped AWS customers achieve exceptional reliability. For instance, ANI Technologies, a leading financial service provider, utilised this setup to create a robust failover mechanism, ensuring uninterrupted services even during system failures. Singh explained that such resilience is crucial for mission-critical applications, especially in industries like finance and healthcare. He elaborated on the layered design of AWS’s infrastructure: “Each region consists of multiple availability zones. Think of one availability zone as one or more isolated data centres. If one zone fails, workloads can shift seamlessly to another.” This architecture ensures resilience even in extreme scenarios like natural disasters. Generative AI, a key area of focus for AWS, has witnessed increasing interest from Indian enterprises. Singh highlighted two major innovations announced at re:Invent: model distillation and advancements in automated reasoning. The model distillation technique, offered via Amazon Bedrock, allows smarter, cost-efficient models to inherit the learnings of larger, more accurate ones, significantly lowering operational costs without compromising on precision. Another critical advancement is AWS’s approach to minimising hallucination in AI models using automated reasoning. Singh emphasised that this enhancement ensures model reliability, making it an ideal solution for industries like insurance and healthcare, where accuracy is non-negotiable. These innovations are a direct response to the needs of Indian customers, who often demand high performance at lower costs. Singh detailed AWS’s use of automated reasoning to enhance model reliability. “With this, businesses like insurance companies can rest assured that their models provide accurate responses, eliminating the risk of incorrect outputs,” he said. Driving Innovation Across Industries AWS’s democratisation of technology has empowered organisations across diverse industries. For instance, manufacturing firms like Apollo Tyres are leveraging AWS services and reported a 9% increase in productivity, while banks like Axis and HDFC enhanced customer experiences with data-driven insights. Singh emphasised AWS’s commitment to innovation while ensuring inclusivity. “We democratise technology so that SMBs have the same access to cutting-edge capabilities as large enterprises,” he said. The importance of AWS’s partner ecosystem cannot be overstated. Globally, AWS collaborates with over 140,000 partners, with a significant presence in India. Companies like HCLTech and Persistent Systems are leveraging AWS’s capabilities to deliver innovative solutions to their clients.","excerpt":"AWS’s Satinder Pal Singh points out that while 99% of employers envision their companies becoming AI-driven organisations by 2028, nearly 79% struggle to find skilled professionals.","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Anshika Mathews","publish_date":"2024-12-23T11:26:07","publication_year":"2024","word_count":809,"keywords":["machine learning","TPU","AWS","AI","ML","RAG","Aim","analytics","generative AI","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","generative AI","Aim","RAG","AWS","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-genai-is-indias-high-performance-low-cost-game-changer\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10104894,"title":"World ID 2.0 is the Solution for a lot of Deep Fake Scams","content":"Recently, Nithin Kamath, founder and CEO of Zerodha, posted on LinkedIn a video of him talking about the risk of deep fakes in the financial industry. He highlighted how deep fake is improving and soon, it would become harder for people to figure out if a person online is real or AI-generated. Surprisingly, the video in which Kamath was explaining this was also a deep fake, which was very hard to differentiate. This problem will be bigger for banks that have more stringent regulatory requirements during onboarding. The turning point for Indian financial services occurred when onboarding went fully digital, thanks to Aadhaar and similar technologies. When bringing in a new customer, a crucial aspect is ensuring the authenticity of documents and the person opening the account. Currently, the standard procedure involves retrieving ID or address proof data from sources like Digilocker or Aadhaar. Additionally, facial matching is conducted via a webcam to verify the identity of the person opening the account. Various checks are in place to confirm liveliness and authenticity. However, with the improvement of deepfake technology, it is expected to become increasingly challenging over time to validate whether the individual on the other end is real or artificially generated. This challenge will be more pronounced for banks with stricter regulatory requirements during the onboarding process. Enter World ID 2.0 Sam Altman-backed WorldCoin has introduced World ID 2.0, a digital passport for the online realm that is a unique human identity. It empowers humans to validate your unique human identity on the internet while safeguarding your privacy. Crafted as an open protocol, it is designed to be a collective ownership, granting users full control. Initially introduced in early 2023, World ID 1.0 addressed the imperative challenge of establishing personhood proof in the age of AI, prioritising privacy, inclusivity, and openness. The second release will introduce Apps for  integrations to verify online accounts using World ID, including integrations with major platforms such as Reddit, Discord, Shopify, Minecraft, and Telegram. In this upgrade, World ID will feature three Levels, World ID Device, World ID Orb, and World ID Orb+, allowing tailored verification checks, ensuring adaptability to real-world use cases. Interestingly, World ID 2.0 might be able to solve these problems with deep fake. Just like other platforms, World ID can be introduced within government apps and services such as Digilocker or within banking to ensure more security, further protecting people’s identities. Moreover, the second-generation protocol boasts core enhancements, like the ability to reset your World ID at an Orb and solidifying its status as the most secure, private, and inclusive proof of personhood. For added privacy, World ID Apps access only a disposable number derived from your World ID, ensuring a consistent yet unlinkable experience across different platforms. Though this comes with hackable instances on apps other than the Orb, the possibility is still imaginable. Concerns with World ID Although it may appear logical in certain aspects, the extensive risk to personal freedoms and the potential for government abuse are substantial, even while considering World ID as a solution. Do we really want the government to have real-time access to our web browsing? Analogous to how government-backed digital currency, which is different from blockchain-based cryptocurrency, empowers authorities to monitor and document each of your online financial transactions, digital ID poses a similar capability, scrutinising every click you execute online. Moreover, WorldCoin has already said that it is giving governments the permission to use its ID system. The parallel with digital currency is evident, as it affords authorities the capability to meticulously track and record every online financial transaction. Similarly, the implementation of digital ID raises valid concerns about the government’s access to real-time information on every online interaction, which was already made possible with UPI. It becomes paramount to deliberate on the trade-offs between enhanced security and the preservation of personal freedoms. World ID 2.0 acknowledges the need for this delicate balance, aspiring to offer a secure, private, and inclusive proof of humanity while respecting the rights and privacy of individuals, emboldening the solution for deep fake along with it.","excerpt":"WorldCoin has already said that it is giving governments the permission to use its ID system.","categories":["AI Features"],"tags":["deep fake","DeepFake","Sam Altman"],"author_name":"Mohit Pandey","publish_date":"2023-12-14T16:00:00","publication_year":"2023","word_count":680,"keywords":["Go","Sam Altman","programming_languages:R","AI","programming_languages:Go","Git","deep fake","ViT","DeepFake","R"],"extracted_tech_keywords":["AI","R","Go","Git","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/world-id-2-0-is-the-solution-for-a-lot-of-deep-fake-scams\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39096,"title":"HCL Tech Banks On AI &#038; Automation To Surpass Wipro, Becomes The 3rd Largest IT Firm","content":"HCL Technologies, one of the leading Indian IT services companies has become the third best in the country, beating Wipro as it cloaks a two-digit annual growth rate in the last financial year. IT services major has surpassed Wipro to occupy the post, as its revenue reached a whopping $8.63 billion, earning HCL a 10 percent hike as compared to the previous years. While Wipro earned a revenue of $8.12 million, which is just 3.8 per cent more when compared to what they achieved in the last fiscal. This sudden spike in HCL’s revenue has made it the third best in the country and is succeeded by Infosys which grew by 9 per cent and TCS leading the growth with an impressive 11.4 per cent for the last fiscal. “For the whole financial year, the net profit of Shiv Nadar-promoted firm crossed Rs 10,000 crore to reach Rs 10,123 crore, up 15.3 per cent year-on-year. Similarly, revenues were Rs 60,427 crore, a rise of around 20 per cent on a year-on-year basis,” a leading business daily noted. Investment in IBM’s Products Paid Off One of the major determinants for HCL’s outstanding performance comes from its investment in IBM products, which played a crucial role in directly boosting HCL’s incremental revenues, which stood at $794 million for the last fiscal year. Of the total revenue, $300 million dollars was generated from IBM products alone which constituted 38 per cent of the incremental revenues. The move to procure large-scale IBM products was taken, following HCL’s decision to set-up up exclusive products and services groups and had earmarked $3.02 billion to build a software product business. Following on this line, the HCL inked a $1.8 billion deal with IBM where it purchased a host of products ranging from automation tools to secure device management. “The large-scale deployments of these products provide us with a great opportunity to reach and serve thousands of global enterprises across a wide range of industries and markets. I am confident that these products will see good growth trajectory backed by our commitment to investing in product innovation coupled with our strong client focus and agile product development. In addition, we see tremendous potential for creating compelling ‘as-a-service’ offerings by combining these products with our Mode-1 and Mode-2 services,” C Vijayakumar, President & CEO, HCL Technologies said. Automation and AI Leads the way Speaking about the role of advanced analytics and AI in driving revenue and growth, Vijayakumar said that the service provider is slowly shifting its focus to AI and currently working on bringing new products that are more AI and automation focused to drive its focus. “There are a lot of products we’ve acquired and these are being launched in a managed services version. There are other areas like AI (artificial intelligence) and machine learning. And, we have the DRYIce (AI-backed automation platform) capability, which is helping us not only in services deals but we are also selling it as a standalone product,” Vijayakumar added. In its 2017-2018 Mode 1-2-3 Strategy, HCL has further laid down its strategy to regain its infrastructure management services like next-generation data centre and cloud services and workplace services through its autonomics and orchestration platforms. Among its solution for autonomics include a robotics process automation and AI\/NLP-powered cognitive automation product Further, it announced its decision to incorporate, DRYiCET, an automation tool for cognitive, artificial intelligence, and machine learning to its various range of workspace solutions in order to make them AI ready.","excerpt":"HCL Technologies, one of the leading Indian IT services companies has become the third best in the country, beating Wipro as it cloaks a two-digit annual growth rate in the last financial year. IT services major has surpassed Wipro to occupy the post, as its revenue reached a whopping $8.63 billion, earning HCL a 10 […]","categories":["IT Services"],"tags":["HCL Technology","Infosys","ML","Robotic Process Automation"],"author_name":"Akshaya Asokan","publish_date":"2019-05-13T12:50:59","publication_year":"2019","word_count":580,"keywords":["Go","artificial intelligence","machine learning","Infosys","AI","ML","Git","automation","NLP","CuPy","Robotic Process Automation","HCL Technology","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","analytics","CuPy","R","Go","Git","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/hcl-tech-banks-on-ai-automation-to-surpass-wipro-becomes-the-3rd-largest-it-firm\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10050447,"title":"Building Scalable Machine Learning Models with TensorFlow 2.x","content":"When the size of data becomes very large, the performance of machine learning models becomes a concern. In such situations, the machine learning models need to be scaled which not only helps in saving time and memory but also helps in improving the performance of the model. TensorFlow 2.x provides features to scale machine learning models easily and effectively. In this article, we will discuss scalable machine learning models that can be achieved with TensorFlow 2.x. First, we will try to understand the shared memory models and distributed memory models and then finally we will see how TensorFlow 2.x facilitates these features. The major points that we will cover in this article are listed in the below table of contents. Table of Contents What is Scalable Machine Learning?The Shared Memory Model (SMM)The Distributed Memory Model (DMM)The transition from SMM to DMMThe GPU and TPU AcceleratorsScalability in TensorFlow 2.x Let us proceed with understanding scalable machine learning. What is Scalable Machine Learning? Talking about the real world where the amount of data is so large, the scalability of machine learning models becomes a primary concern for us. Most of the learners start learning data science with a very small amount of data which is small enough to fit the process and model on a single machine. But in the real world, many applications require scaling ML models to multiple machines. There can be various situations where models need to deal with large data sets like dealing with stock market data where models need to adopt new data and produce predictions quickly. And in a millisecond the predictions from the model become useless. In such a situation Scalable Machine Learning comes in the application which aims to combine statistics, systems, machine learning, and data mining in such a procedure that is flexible to any environment. More formally we can say the word scalable means here to make a model which can deal with any amount of data without increasing the consumption of resources like memory, time, etc. When it comes to making the computation of machine learning models across the big data faster some problems can accrue due to lack of scalability like we can find the problem in the fitting of any model on large size of data and if we have a model one thing which can slow us the computation speed of the model. This can lead to consequences where we are required to sacrifice accuracy because optimization of models for large-scale data becomes infeasible and time-consuming. In scalable machine learning, we try to build a system where the components of the system have their own work or task which helps the whole system to lead towards the solution of the problem rapidly, without wasting so much memory and increasing the performance as well. This makes us introduce shared memory models and Distributed Memory Model which are basically two types of hardware structural designs for modelling purposes. The Shared Memory Model When we are developing a model in a single machine we can easily pass the information across the whole machine so that all the cores on the machine can have access to the same memory and the threads of the models can easily access the information. This type of model can be called the shared memory model. Because they are sharing the variables, memory to the multiple processes in a system. This process of sharing allows different processes to communicate with each other. The Distributed Memory Model Where we are having problems that cannot be dealt with by a single machine, we require a network of multiple machines to work on the problem so that they can communicate with each other to make the problem-solving procedure complete. This means the network allows machines to pass messages to each other according to the requirement. Image source The above image is an illustration of a distributed memory system with three machines. The Transition from SMM to DMM Scalability can also be considered as the transition of processes from a shared memory model to a distributed memory model. Some basic concepts of any distributed memory model are: The machines are fully connected by the means of nodes of the network.The links of the network between machines are bidirectional. According to those basic concepts we need to be aware of the following points: Which computing thread of the model is in which machine.How to move the data in the network to complete the procedure without taking so much time and facing issues. Developing models on a shared memory model is easier but it has its own drawback like you have a limited number of cores, computing threads, and memory. In terms of big data, it often happens that the models become larger than the RAM memory of a single machine where the distributed system allows us to increase the memory level by combining more than two systems. Also, it enhances the speed of the model by providing different spaces for different computations. The GPU and TPU Accelerators Although we can use GPU (Graphical Processing Unit) for increasing the memory and power of computation to a limit in the shared memory model and in any distributed memory system we can use the TPU (Tensor Processing Unit). Using the NVIDIA GPU is a better option for utilizing the benefits of the shared memory model but it is also limited to an extent. What if we require more memory to share or more cores for calculation, in such scenarios we require to go with the distributed memory model where Google TPU is a better option to use as an accelerator of the procedure. Google TPU is specially designed for the distributed memory model which provides separate IP addresses and communication protocols. And also using it saves our time from dealing with the communication layer which is taken care of by the provided accelerated Linear Algebra Layer(XLA) which can be exposed using TensorFlow API. Scalability in TensorFlow 2.x The typical TPU architecture works on the basis of serialization of the data in the batches in TensorFlow with a module named TFrecord which makes the data serialization easier. In a distributed memory model it is very important and one of the toughest tasks to do. TFrecord data serialization gives ease on it and provides help in large-scale deployment. The below image represents the architect of the shared and distributed memory model when using GPU and TPU accelerators. The TensorFlow API has made many of the in-between processes easy to perform when training a model. The TensorFlow 2.x has a custom training loop feature using which we can train any number of models synchronously. Where in the previous version we didn’t have this option which means the API was allowed to train only one model. In TensorFlow 2.x, eager execution is turned on by default, which is a programming environment that helps in the evaluation of operations in less time. The evaluated operations give rigid values and avoid the procedure for making a computational graph to run later. This helps in debugging the models. It is a flexible machine learning platform for experimentation. Using this we can call an operation to test running models. And allows the use of python control flow instead of graph control flow. The TensorFlow 2.x allows us to use the compiled functions by the tf.function module using which we can make graphs of our programs. The introduction of custom functions has enhanced the flexibility of TPUs as the GPUs are having. There are several other new features like API tf.Distributed. The strategy allows distributed training of the models across multiple machines where the major goal of the API is to provide an easy-to-use interface with good performance and easy switching between strategies. In Tensorflow 2.x there are also extensions to the tf.data object that makes it easy to specify how data is distributed among the several cores. Which makes it easy to understand the distribution of the data across the cores. By summing up all these features, we can say the eager mode is something that we need in the shared memory model. The existence of custom functions makes a model execution slow. The reason for slow execution is that the custom function should be applied in the remote execution of the model. Debugging the delayed execution model is harder because of the dependency on the tracing. In such a situation the strategy function helps because of its strategy of making code transparent for any interface. Again the common code path makes it possible for developers to code and understand the methods if they are switching from a shared memory model to distributed memory model. The codes of the models can be run first as a shared memory model and in the requirement, they can switch into another interface of the distributed memory model. Also allowing the debugging in the GPU mode makes the process faster. Final words As we have seen in the article, developing the models of the shared memory models category is easier than developing the distributed memory model. Since performance enhancement is a major goal of any developer we are required to make the DMM category models. TensorFlow is such a good library which not only allows us to develop the DMM easily but also gives features that can help us in debugging and a similar code path is making the models flexible for transition between SMM and DMM. References: TensorFlow 1.x vs TensorFlow 2 – Behaviors and APIs.","excerpt":"In scalable machine learning, we try to build a system where the components of the system have their own work or task which helps the whole system to lead towards the solution of the problem rapidly","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning","NVIDIA GPU","Python","scalability","what is tensorflow"],"author_name":"Yugesh Verma","publish_date":"2021-10-06T12:00:00","publication_year":"2021","word_count":1580,"keywords":["data science","Go","NVIDIA GPU","machine learning","TPU","AI","TensorFlow","what is tensorflow","ML","Machine Learning","Python","Aim","scalability","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Aim","TensorFlow","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/building-scalable-machine-learning-models-with-tensorflow-2-x\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10093627,"title":"Drag Your GAN: A New Image Editing Model Wows the Internet","content":"A group of researchers from Google, alongside the Max Planck Institute of Informatics and MIT CSAIL, recently released DragGAN, an interactive approach for intuitive point-based image editing. This new method leverages a pre-trained GAN to synthesise images that not only precisely follow user input, but also stay on the manifold of realistic images. In comparison to many previous approaches, the researchers have presented a general framework by not relying on domain-specific modelling or auxiliary networks. In order to achieve this, they used an optimisation of latent codes that incrementally moves multiple handle points towards their target locations, alongside a point tracking procedure to faithfully trace the trajectory of the handle points. GANs are still the king at latent space exploration.DragGAN looks amazing. pic.twitter.com\/KT3AEtdBJK— marko (@markopolojarvi) May 19, 2023 Both components use the discriminative quality of intermediate feature maps of the GAN to yield pixel-precise image deformations and interactive performance. The researchers claimed that their approach outperforms the SOTA in GAN-based manipulation and opens new directions for powerful image editing using generative priors. In the coming months, they look to extend point-based editing to 3D generative models. GAN vs Diffusion Models This new technique shows that GAN models are more impactful than pretty pictures generated from diffusion models – namely used in tools like DALLE.2, Stable Diffusion, and Midjourney. While there are obvious reasons why diffusion models are gaining popularity for image synthesis, general adversarial networks (GANs) saw the same popularity, sparked interest and were revived in 2017, three years after they were proposed by Ian Goodfellow. GAN uses two neural networks—generator and discriminator—set against each other to generate new and synthesised instances of data, whereas diffusion models are likelihood-based models that offer more stability along with greater quality on image generation tasks. Read: GANs in The Age of Diffusion Models","excerpt":"The method leverages a pre-trained GAN to synthesise images that stay on the manifold of realistic images.","categories":["AI News"],"tags":["AI Tool"],"author_name":"Tasmia Ansari","publish_date":"2023-05-19T16:09:03","publication_year":"2023","word_count":300,"keywords":["Go","programming_languages:R","AI","neural network","diffusion models","RAG","Aim","stable diffusion","GAN","AI Tool","R"],"extracted_tech_keywords":["AI","neural network","Aim","RAG","R","Go","stable diffusion","diffusion models","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/drag-your-gan-a-new-image-editing-model-wows-the-internet\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":23391,"title":"Parag Chitalia Of VMware Goes In Depth About The Analytics Journey &#038; Future Plans For The Company","content":"Parag Chitalia, who is the senior director of advanced analytics and business intelligence at VMWare, believes that analytics is fun, it is here to stay and will free humankind from mundane decision making. He has over two decades of experience in building high performance teams in business analytics, supply chain, consulting across the Americas and Asia. As a part of his journey with VMWare, he has been responsible for driving business value for strategic initiatives enabling data driven decision making. While engaging with stakeholders, he has brought value in BI and predictive analytics solutions to drive growth and improve customer satisfaction. To get a brief insight on VMWare, the company has been a leader in cloud infrastructure and digital workspace technology since 1998, accelerating digital transformation and building solutions to facilitate an evolution of IT environments. With its solutions in technology, strategies, cloud computing and several others, several organisations have been improving their business agility by modernising data centres and integrating public cloud, thus driving innovation and empowering the digital workspace. In his candid conversation with Analytics India Magazine, Chitalia shares insights on his analytics journey, future plans for the company and much more. AIM: Would you like to talk about VMware and the various products offered by it? Parag Chitalia: VMware, a global leader in cloud infrastructure and digital workspace technology, accelerates digital transformation by enabling unprecedented freedom and flexibility in how our customers build and evolve IT environments. It is a leader in the Compute, Network and Storage virtualization market globally and led the development of Software Defined Data Centers (SDDC). AIM: How are organizations driving innovations and improving business agility using the solutions offered by the company? PC: With VMware solutions, organizations are improving business agility by modernizing data centres and integrating public clouds, driving innovation with modern apps, creating exceptional experiences by empowering the digital workspace, and safeguarding customer trust by transforming security. AIM: How did you start your career in Analytics? How has the journey been so far? PC: I was a career Supply Chain consultant when Dell asked if I could start a Supply Chain Analytics team to drive Supply Chain efficiencies. Having spent almost 2 decades in Supply Chain, an offer to improve one of the best run Supply Chains in the world was too enticing. It did not take me too long to realize that I had found a new career in analytics -leveraging my inherent analytic psyche. I have spent the past 12+ years thoroughly loving the (analytical) work I do, building analytical teams across the globe and across business functions within a company to drive cost efficacies and to drive revenue growth. The journey has been one of incredible passion, learnings and fun. AIM: Would you like to share some of the analytics solutions that you have worked on? PC: Analytics has been around since humanity invented the wheel (and probably before it!) enabling decision-making since the ages. Our ability to harvest more (structured and unstructured data) information, and the unprecedented availability of massively parallel computing, map-reduce and compute capacity, availability of analytical tools and techniques, has created  a perfect storm driving the democratisation of analytical decision-making. During my tenure at Dell Global Analytics and at VMware, I have been very fortunate to have worked across all (internal) business functions, supporting decision-making with well-honed analytics. Here are a few examples: Supporting sales with Cross Sell\/UpSell propensity models Supporting marketing with decision-making enabling the best campaign, and timing, and product recommendation for the various customer personas involved in the customer purchase decision Enabling (optimizing) the list prices for our products. Evaluating the range of discounts that yield the best results for customers and the company Using AI\/ML to provide possible resolutions to our customers, as the customer is typing the issue they are encountering using our products Enabling finance to find the optimal mix and discount for marketing accounts receivables (AR) Minimizing the fraud perpetuated in commerce transactions Driving proactive support by notifying our customers of potential issues before they happen in the customers environment, and recommending mitigation\/resolutions paths enabling customer delight Enabling Robotic Process Automation during our order taking and order creation processes to drive automated & expedited order processing Optimizing the levels of inventory across the supply chain -deciding where to hold inventory -in finished goods, in intermediate goods or in raw materials Driving logistics and transportation cost reductions, and reducing\/optimizing time to delivery across a supply chain network AIM: What is the size and hierarchical alignment (both depth and breadth), of your analytics group? PC: At VMware, the BI & Analytics team is considered a business function and we report to the CFO. The teams are aligned to support the business functions such as sales, marketing, finance, global support, etc. We leverage the skills and time zones across the USA, Costa Rica, UK & Ireland, Bulgaria, India, and Singapore to speed our analytical delivery for our stakeholders. The internal team leverages talent from our analytics partners with about 200 FTE supporting Analytics and BI within VMware. AIM: What kind of knowledge and skill-sets do you look for, while recruiting your workforce? PC: A combination of skills drive the superhuman data scientist (do they really exist?, maybe we all could find only a few around). We look for strengths in at least a couple of the key areas – in my opinion, the key vectors that characterise the medley of skills required encompass the following Business skills -understanding the business function, the ability to execute the output of analytics in the field, understanding the business impact, … Technical skills -characterised by an ability and desire to code -there is still some time before we can use an excel formula for decision tree approaches and yet make sense of the results! I tend to club the data exploration skills into this bucket too (though some of my peers would separate it out) -ability to identify the right sample, biases, data exploration to identify the features Modeling skills -applied mathematics, statistics, optimization and simulation skills And last but not the least, soft skills required to translate the (sometimes) complex analytics into (easy) business speak AIM: What is the roadmap for analytics at VMWare? PC: At VMware, we have had tremendous support from our stakeholder in leveraging analytics in the decision-making. We have been able to drive significant (and measured) value with incremental revenue and cost savings across the key projects that have been implemented. Creating a pull for advanced analytics in the decision-making is well under way -one knows that job is done when demand for analytics outstrips the capacity. Embedding advanced analytics in the decision-making fabric, enabling data sciences into driving automated decision-making is the next frontier. AIM: Would you like to highlight some of the most significant challenges you face being in analytics space? PC: The incredible visibility for data sciences and advanced analytics (should I say hype, approaching the peak of the hype cycle) has served us well, in creating pull for skills creation in the space, to business stakeholder recognizing (sometimes demanding AI\/ML) the need for such analytics. It has also created a downside that serves as key challenges as we strive to embed, augment advanced analytics into decision-making: the very stakeholders that leverage decision support with these advanced analytical techniques, get apprehensive with the (hype around) the replacement of humans in decision-making. The visible draw for data sciences has created the proverbial “hammer looking for a nail” with freshly minted data sciences graduates “interested in building the data science models, not necessarily solving the business problem”. For the business, the adoption of data science solutions into execution –“productionizing” the analytics, remains a formidable frontier AIM: How do you think ‘Analytics’ as an industry is evolving today? Could you tell us the most important contemporary trends that you see emerging in the present analytics space across the globe? PC: Analytics was and remains a key part of the decision-making process in any enterprise. As compute gets more ubiquitous, analytics will drive more decision-making to the Edge (the intersection of the digital and the physical world). What we see in the consumer world at the Edge -google maps making routing “decisions”, facebook feeds driving news content consumption, has made and will continue to make headway in the b2b world. Quantitative and computer trading accounts for the vast majority of trades in the US exchanges already, autopilot commercial jets, etc. When will self-driving cars push the split-second decision-making to the on-board (trained) computer in our cars? We will be seeing tremendous advances in robotics process automation; augmenting, and in some cases, embedding analytics into the decision-making. The question is not how or why but when and to what extent. AIM: Would you like to add anything? PC: Analytics is fun, Analytics is here to stay and Analytics will free humankind from the mundane decision-making to enable us to solve problems we have not imagined before! Get on with it, don’t get left behind!","excerpt":"Parag Chitalia, who is the senior director of advanced analytics and business intelligence at VMWare, believes that analytics is fun, it is here to stay and will free humankind from mundane decision making. He has over two decades of experience in building high performance teams in business analytics, supply chain, consulting across the Americas and […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-04-09T10:52:42","publication_year":"2018","word_count":1494,"keywords":["data science","TPU","AI","cloud computing","R","ML","RAG","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","RAG","predictive analytics","cloud computing","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/parag-chitalia-of-vmware-goes-in-depth-about-the-analytics-journey-future-plans-for-the-company\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":49361,"title":"MachineHack Winners: How These Data Science Enthusiasts Solved The ‘Predict Books Price’ Hackathon","content":"MachineHack recently concluded its 18th edition of Machine Learning hackathons by announcing the winners for it’s Predict The Price Of Books challenge. Shravan Kumar, Divyanshu Suri and Saurabh Kumar secured the first, second and third places respectively on the hackathon leaderboard. Analytics India Magazine introduces you to the winners and their approach to the solution. #1: Shravan Kumar Shravan Kumar is a Senior Manager of Advanced Analytics at Novartis. Though Shravan has been working in the analytics field for some time he always had a keen interest in the predictive analytics domain which he wished to explore. He started with MOOC courses in platforms like Coursera, Edx, Udacity etc. and acquired the skillsets for his area of interest. Having learned the essentials, Shravan’s next focus was on perfecting his skills through online competitions. He is an active participator of hackathons conducted by MachineHack, Kaggle and other platforms. Shravan’s Approach To Solving The Problem Shravan explains his approach as follows: Pre-processing steps: Checked the number of rows, columns, data types of variables, missing values and observed word clouds of text data.Basic pre-processing steps for text are observed with a major focus on two columns i.e., Title and Synopsis. Meta Features are created like ‘number of words’, ‘number of unique words’, ‘number of characters’ etc.Reviews and Ratings values are converted into numeric values.Edition variable was split into two major variables Edition Type and Date column – Month is extracted from Date.Created features with TF-IDF and Count Vectorizer for both the Synopsis and Title variables.More features were created by using Glove vectors for the synopsis, genre, book category variable values.Converted all the text variables ‘Title’, ‘Author’, ‘Synopsis’, ‘Genre’, ‘Year’, ‘Month’ to label encoded valuesConverted the price of the book into Log(Price) – because it is giving a normal distributionCreated count and mean encoded features for all categorical variables Model building steps: Used 5 fold cross-validation techniques along with LightGBM as algorithm and RMSE as a metric. Experimented with hyperparameter tuning to achieve better scores, especially by changing the learning rate and seed values. Choosing the right cross-validation technique and feature preparation helped me achieve the 1st Rank on leaderboard Click here to view the code. “MachineHack is a great learning platform. The articles by Analytics India Magazine writers are very helpful and keep all our industry-relevant people updated with news in this industry. Truly MachineHack is one of the best hackathon organisers and data science knowledge portals in India,” he said. #2: Divyanshu Suri Now a Senior Manager of Machine Learning at AXA XL, Divyanshu Suri is not new to MachineHack and has won multiple hackathons. Having done his Bachelors in Statistics from Delhi University and a Masters in Applied Statistics from IIT Bombay, Divyanshu was amazed by the real power of data science in his second job at EXL Service where he worked in insurance analytics. He then went on to participate in many online hackathons gaining knowledge and improving his skills. Now, as a Senior Manager, he applies predictive analytics to solve a variety of data science problems in commercial and speciality lines. Divyanshu’s Approach To Solving The Problem Divyanshu started the competition with exploratory analysis, trying to understand the data. He then proceeded with data cleaning and feature engineering. In order to find the best fitting model for the problem, he tried different algorithms and compared the performances and then combined the better performing models. He built a lot of different models based on a different set of variables, different transformations, different variable creation algorithms, and different ML algorithms and finally used stacking concept to come up with the final model. Click here to view the code. “MachineHack is a great platform to learn and apply new data science techniques and ML algorithms and improve your own skillset. It is also a great platform to compete with the other industry experts in the data science community.”- he said. #3: Saurabh Kumar A skilled and experienced Data Scientist in a reputed firm, Kumar has shown his expertise multiple times by topping several hackathons at MachineHack. Kumar’s interest in the field of Data Science and Machine Learning emerged from a single algorithm. His personal experience with the Random Forest Algorithm and its capabilities thrilled him to pursue and advance his skills in the buzzing field. Kumar said he is inspired and overwhelmed by the ability of ML algorithms to solve a variety of real-world problems. Kumar’s Approach To Solving The Problem Sourabh Kumar used basic feature engineering and traditional NLP techniques like BOW and TF-IDF and lightgbm for cracking the hackathon. Click here to view the code. “I am active on the MachineHack platform since their first hackathon and really enjoy competing here. MachineHack team is very cooperative and is willing to work on feedbacks” – he said.","excerpt":"MachineHack recently concluded its 18th edition of Machine Learning hackathons by announcing the winners for it’s Predict The Price Of Books challenge. Shravan Kumar, Divyanshu Suri and Saurabh Kumar secured the first, second and third places respectively on the hackathon leaderboard. Analytics India Magazine introduces you to the winners and their approach to the solution. […]","categories":["Deep Tech"],"tags":["best online data science masters","Hackathons","Machinehack"],"author_name":"Amal Nair","publish_date":"2019-11-04T16:30:51","publication_year":"2019","word_count":796,"keywords":["best online data science masters","data science","Go","machine learning","AI","Machinehack","R","ML","Hackathons","NLP","analytics","LightGBM","predictive analytics"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","data science","analytics","LightGBM","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-these-data-science-enthusiasts-solved-the-predict-books-price-hackathon\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10125109,"title":"TIME and OpenAI Announce Multi-Year Content Partnership","content":"TIME and OpenAI have entered a multi-year content deal and strategic partnership to integrate TIME’s journalism with OpenAI’s products, including ChatGPT. The partnership grants OpenAI access to TIME’s archives spanning 101 years, enabling the use of this content in AI-generated responses with citations and links to the original sources on Time.com. Further, the collaboration aims to expand global access to accurate and reliable information. Mark Howard, TIME’s Chief Operating Officer, said that the partnership aligns with TIME’s history of innovation in delivering journalism. “This partnership with OpenAI advances our mission to expand access to trusted information globally,” he  said. Brad Lightcap, Chief Operating Officer of OpenAI, emphasised the partnership’s role in facilitating access to news content and supporting reputable journalism by ensuring proper attribution to original sources. Additionally, the partnership allows TIME to utilise OpenAI’s technology for developing new products and provides an opportunity for TIME to offer feedback and practical applications to enhance the delivery of journalism in OpenAI’s products. Prior to this OpenAI has partnered with several prominent media houses to enhance its AI models and provide high-quality content to users. These partnerships include Le Monde and Prisa Media, which bring French and Spanish news content to ChatGPT, as well as Vox Media, The Atlantic, and News Corp, which provide a wealth of journalistic content for training and user engagement. Additionally, OpenAI has collaborated with Axel Springer, Financial Times, and Associated Press, among others, to support the dissemination of accurate and balanced news stories.","excerpt":"The partnership grants OpenAI access to TIME’s archives spanning 101 years.","categories":["AI News"],"tags":["ai announcements","OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-06-27T21:20:08","publication_year":"2024","word_count":246,"keywords":["ChatGPT","OpenAI","AI","innovation","GPT","Aim","ai announcements","Rust","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","R","Rust","GPT","innovation","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/time-and-openai-announce-multi-year-content-partnership\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":9886,"title":"Interview – Amit Kalra, Head of Strategic Initiatives &#038; ARC at Swiss Re","content":"Swiss Re is world’s leading reinsurance company with over 150 years of history. Their core business model revolves around business expertise and analytics, which is embedded across the value chain and functions. We spoke to Amit Kalra, who is leading analytics for Swiss Re from the front. Amit joined Swiss Re in 2007 and is currently the Head of Strategic Initiatives at Swiss Re Bangalore wherein he is responsible for strategy and competency build up initiatives. In addition, he heads Analytics Research and Consulting unit which include teams spanning across actuarial, data analytics and economic research. Amit is the author of several sigma studies on trends in the emerging markets, covering a broad range of topics including food security, microinsurance and implications of urbanisation on the insurance sector. He on a regular basis also contributes to articles\/whitepapers on re\/insurance trends in leading insurance journals and magazines. Prior to joining Swiss Re, Amit was heading Strategic Research Group for GE Insurance solutions at GE Capital. Amit holds a Master degree in Business Economics from University of Delhi and is also a certified Chartered Property Casualty Underwriter (CPCU) and Associate in Reinsurance from The Institutes, United States. Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of analytics approach and policies at your organization? Amit Kalra: We deploy analytics in underwriting, risk management, natural catastrophe modeling, actuarial costing and reserving. At Swiss Re, we have already began to build analytics capabilities that help us to improve the efficiency of our operations and deepen our client market and risk insights. We are also actively investing in building digital analytics capabilities that leverages advanced technological solutions and new methodologies. [pullquote align=”right”]Analytic capabilities streamlines processes and helps the re\/insurance industry to more effectively fulfill its role as an enabler of enterprise and progress all over the world.[\/pullquote] AIM: Can you brief about some of the analytics solutions that you work on? AK: Sure, let me share couple of recent examples focusing on text mining and risk analytics: Fine-grained multiple entity extraction from document images: Swiss Re receives huge number of documents in various formats from clients and customers globally. Extracting relevant information from such documents is a key to many business processes. The technical challenges involve (i) high volume of documents (ii) document images consisting of text, tables and images, and (iii) required accuracy level of extraction. Data Analytics team is working on a solution that uses advanced text mining and document image processing models to help our clients better understand their portfolios and risk landscape. Healthcare Analytics: A person’s lifestyle choices may have impact on his\/her health. Some of these choices may be key risk factors which insurance companies are not able to assess and hence segment risks optimally. Being able to identify such risk factors can help the company plan more effectively in dealing with them. We are currently working on personalized healthcare analytical solution using heterogeneous data sources. Technical challenges include (i) combining data from multiple sources, (ii) handling missing values and (iii) dealing with class imbalance. AIM: What are the key differentiators in your analytical solutions? AK: Analytics solutions in Swiss Re are deeply integrated with business expertise. Therefore, for us the analytics solutions are a mix of deep domain understanding\/expertise, analytical mindset\/capabilities and technology adoption. We are looking at analytics to support evolutionary topics (how can we improve our existing business model and practices) as well as disruptive topics (how can we transform or look at alternate business models). For us, innovation and disruptive thinking are the key elements. AIM: Please brief us about the size of your analytics group and what is hierarchal alignment, both depth and breadth. AK: We have a large and growing data scientist group based out of Zurich, Armonk, Singapore and Bangalore, which has a strong R&D and execution capabilities to provide solutions to complex global business problems. In addition, as I said we have majority of our staff who are experts (with quantitative background) in multiple functions who engage on analytical problems on day in and day out basis (underwriters, actuaries, risk managers, sales analytics etc.) AIM: What are the planned next steps\/ road ahead for analytics in your organizations? AK: We will continue to scale up analytics capabilities in both breadth as well as depth. We are also open to collaborating with other analytics providers in the Indian ecosystem to augment and support us in our execution strategy. We recently announced the launch of Swiss Re InsurTech accelerator aimed at engaging with startups on some of the disruptive themes within the insurance sector. This will be India’s first insurance focused startup accelerator. We have identified Smart Analytics as one of the three key themes that we want to explore further while engaging with startups. Here we are talking about potential product\/solutions that have implications on the entire or parts of insurance value chain, be it underwriting, risk management, claims management or other business aspects. AIM: What are the most significant challenges you face being in the forefront of analytics space? AK: Change management journey and raising the acceptability of technology driven solutions across the global business functions Quality talent pool: we do see shortage of high quality talent pool in the market who can relate technology with the business strategy and domain capabilities in a holistic way Prescriptive analytics in addition to predictive analytics. AIM: How did you start your career in analytics? AK: I joined Swiss Re in 2007 as Head Economic Research & Consulting India team, wherein I authored several sigma studies on trends in the emerging markets, covering a broad range of topics including food security, microinsurance and implications of urbanisation on the insurance sector. At present, I head Strategic Initiatives team at Swiss Re Bangalore wherein I am responsible for strategy, innovation and competency build up initiatives. In addition, I head Analytics Research and Consulting unit, which include teams spanning across actuarial, data analytics and economic research. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? AK: Academic background: Masters and PhD in Computer Science, Mathematics and Statistics with knowledge in data analytic tools, Machine Learning, Natural Language Processing, Computer Vision or Software Development. Strong demonstration of key professional behaviors such as collaboration, client centricity, problem solving, positive attitude amongst others. Selection Methodology: Multiple rounds of discussion with the management and technical team members AIM: How do you see Analytics evolving today in the industry as a whole? What are the most important contemporary trends that you see emerging in the Analytics space across the globe? [pullquote]Increased adoption of sensors and devices has the potential to generate massive amount of data that needs to be analyzed in real time for personalized retail solutions and higher industrial\/commercial productivity and business impact.[\/pullquote]AK: Analytics should not be viewed as an independent discipline; rather it should be well integrated with the business. The analytical mindset should be in the DNA of an organizational culture, with all the enabling tools at the disposal of employees to attain organizational priorities. I see positive trend of rising automation and analytical applications across the business value chain, as it supports agility in business practices and decision-making. Increased adoption of sensors and devices has the potential to generate massive amount of data that needs to be analyzed in real time for personalized retail solutions and higher industrial\/commercial productivity and business impact.","excerpt":"Swiss Re is world’s leading reinsurance company with over 150 years of history. Their core business model revolves around business expertise and analytics, which is embedded across the value chain and functions. We spoke to Amit Kalra, who is leading analytics for Swiss Re from the front. Amit joined Swiss Re in 2007 and is […]","categories":["AI Features"],"tags":["analytics leaders","insurance analytics","Interviews and Discussions"],"author_name":"Дарья","publish_date":"2016-05-11T10:57:01","publication_year":"2016","word_count":1244,"keywords":["Go","machine learning","AI","analytics leaders","insurance analytics","ML","R","computer vision","RAG","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","analytics","Aim","RAG","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-amit-kalra-head-strategic-initiatives-arc-swiss-re\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10012809,"title":"Top Deep Learning Based Time Series Methods","content":"The components of time-series can be as complex and sophisticated as the data itself. With every passing second, the data obtained multiplies and modelling becomes tricky. For instance, social media platforms, the data handling chores get worse with their increasing popularity. Twitter stores 1.5 petabytes of logical time series data and handles 25K query requests per minute. There are more critical applications of time series modelling, such as IoT and on various edge devices. Sensors of smart buildings, factories, power plants, and data centres generate vast amounts of multivariate time series data. Conventional anomaly detection methods are inadequate due to the dynamic complexities of these systems. Today, most of the state-of-the-art methods aim to leverage deep learning for time-series modelling. In this article, we take a look at a few of the top works on deep learning base time series that have been published in the past couple of years. Multivariate LSTM-FCNs Year: 2018 The researchers transformed the univariate model, Long Short Term Memory Fully Convolutional Network (LSTM-FCN) and Attention-based variant–ALSTM-FCN), into a multivariate time series classification model. The proposed models work efficiently on various complex multivariate time series classification tasks such as activity recognition or action recognition. And, are highly efficient at test time and small enough to deploy on memory-constrained systems. Recurrent Conditional GANs Year: 2018 Recurrent GANs make use of recurrent neural networks in the generator and the discriminator. In the case of RCGANs, both of these RNNs are conditioned on auxiliary information. RCGANs can generate time-series data useful for supervised training, with only minor degradation in performance on real test data and can be used for medical time series data generation. Deep Reinforcement Learning Year: 2018 Deep Q-learning is investigated as an end-to-end solution to estimate the optimal strategies for acting on time series input. The univariate game tests whether the agent can capture the underlying dynamics, and the bivariate game tests whether the agent can utilise the hidden relation among the inputs. Stacked Gated Recurrent Unit (GRU), Long Short-Term Memory (LSTM) units, convolutional neural network (CNN), and multi-layer perceptron (MLP) are used to model Q values. The GRU-based agents show best overall performance in the Univariate game, while the MLP-based agents outperform others in the Bivariate game. ROCKET Year: 2019 The computation complexity of the state-of-the-art time series classification methods demands longer training times, even for smaller datasets. Existing methods focus on a single type of features, such as shape or frequency. In this method–ROCKET– the researchers demonstrated that simple linear classifiers with random convolutional kernels could attain state-of-the-art accuracy with a fraction of computation of existing methods. MAD-GAN Year: 2019 Instead of treating each data stream independently, the Multivariate anomaly detection method based on Generative Adversarial Networks (GANs) considers the entire variable set concurrently to capture the latent interactions amongst the variables. The generator and discriminator of the GAN model are fully exploited using a novel anomaly score called DR-score to detect anomalies by discrimination and reconstruction. According to the researchers, the proposed MAD-GAN was effective in reporting anomalies caused by various cyber-intrusions compared in these complex real-world systems. Shallow RNN Year: 2019 To induce long-term dependencies while enabling parallelisation, shallow RNNs were introduced. Its architecture runs several independent RNNs followed by a second layer that consumes the output of the first layer using a second RNN, thus capturing long dependencies. For time-series classification, this technique leads to substantially improved inference time over standard RNNs without compromising accuracy. Temporal Fusion Transformers Year: 2019 The Temporal Fusion Transformer (TFT) is a novel attention-based architecture, which has been designed for multi-horizon forecasting problems that often contain a complex mix of static (i.e. time-invariant) covariates, known future inputs, and other exogenous time series that are only observed historically. Previous deep learning solutions do not account for the full range of inputs present in common scenarios. TFT utilises recurrent layers for local processing and interpretable self-attention layers for learning long-term dependencies. The TFT also uses specialised components for the judicious selection of relevant features and a series of gating layers to suppress unnecessary components, enabling high performance in a wide range of regimes. N-BEATS Year: 2020 N-BEATS  consists of backward and forward residual links and a very deep stack of fully-connected layers. The architecture allows the model to be interpretable, applicable without modification to a wide array of target domains, and fast to train. The first configuration of this model does not employ any time-series-specific components, and its performance on heterogeneous datasets strongly suggests that deep learning primitives such as residual blocks are sufficient to solve a wide range of forecasting problems. DROCC Year: 2020 Deep Robust One-Class Classification (DROCC) is based on the assumption that the points from the class of interest lie on a well-sampled, locally linear low dimensional manifold. DROCC is highly effective on tabular data, images (CIFAR and ImageNet), audio, and time-series, offering up to 20% increase in accuracy over the state-of-the-art in anomaly detection. For a more comprehensive briefing on the state of deep learning-based time series applications, check this report. Keep track of all latest developments in the Time-series domain by following paperswithcode.","excerpt":"The components of time-series can be as complex and sophisticated as the data itself. With every passing second, the data obtained multiplies and modelling becomes tricky. For instance, social media platforms, the data handling chores get worse with their increasing popularity. Twitter stores 1.5 petabytes of logical time series data and handles 25K query requests […]","categories":["AI Trends"],"tags":["GANs","lstm","RNN","Time Series"],"author_name":"Ram Sagar","publish_date":"2020-12-02T15:00:00","publication_year":"2020","word_count":851,"keywords":["TPU","AI","neural network","ML","lstm","Transformers","RAG","Ray","Aim","deep learning","Time Series","anomaly detection","RNN","GANs"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","Aim","Ray","Transformers","RAG","anomaly detection","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-time-series-deep-learning-methods\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":2144,"title":"Study &#8211; IIT\/ IIM Graduates in Analytics","content":"Graduates from premier institutes like IIT’s and IIM’s should contribute heavily to an upcoming professional area like analytics. We took up an online study to access the level of involvement of IIT\/ IIM graduates towards analytics worldwide. IIT’s and IIM’s together contribute 1.6% of the total analytics workforce of the world. Compare this with top 5 universities with analytics professionals – University of California, Berkeley    – 0.7% University of Mumbai                         – 0.6% New York University                           – 0.6% Stanford University                             – 0.6% Penn State University                         – 0.5% IIT\/ IIM graduates form 0.7% of total analytics workforce in USA. This number for India stands at a whooping 11%. Percentage breakup of total analytics professional for individual institutes is: IIT Bombay – 12.7% IIT Delhi – 11.4% IIM Calcutta – 11.2% IIT Kharagpur – 11.0% IIM Bangalore – 10.2% IIT Madras – 9.7% IIM Ahmedabad – 8.5% IIT Kanpur – 8.1% IIM Lucknow – 5.7% IIT Roorkee – 4.9% IIM Kozhikode – 2.7% IIM Indore – 2.7% IIT Guwahati – 1.2% Total – 100% Top 10 recruiters of analytics professionals from IIT\/ IIM are IBM, CTS, Accenture, GENPACT, Infosys, Deloitte, American Express, TCS, ZS Associates and Citi. The average number of work experience for all analytics professionals from IIT\/ IIM is 8.7 years. Compare this average work experience of analytics professionals in India, which is 6.4 years. This indicates that more number of IIT\/ IIM graduates were able to transition to analytics at later stage of their career. 34% of analytics professionals from IIT\/ IIM have a total work experience of more than 10 years. This number for all analytics professionals in India is 16.1%, again indicating that more number of IIT\/ IIM graduates work in analytics at later stages of their career. Another interesting fact that came out of this study is that 36% of all CXO level analytics positions in the country is filled by IIT\/ IIM graduates, yet only 26% of all entrepreneurial activities in analytics is done by IIT\/ IIM graduates. Top Cities where analytics professionals from IIT\/ IIM reside are: Bengaluru                  – 17.7% Mumbai                      – 12.4% New Delhi                  – 8.1% San Francisco           – 6.8% Gurgaon                     – 6.7% New York City           – 5.6% Hyderabad                 – 3.7% Chennai                      – 3.1%","excerpt":"Graduates from premier institutes like IIT’s and IIM’s should contribute heavily to an upcoming professional area like analytics. We took up an online study to access the level of involvement of IIT\/ IIM graduates towards analytics worldwide. IIT’s and IIM’s together contribute 1.6% of the total analytics workforce of the world. Compare this with top […]","categories":["AI Features"],"tags":["IIT Bombay","iit data science","Quantum Computer"],"author_name":"Дарья","publish_date":"2012-11-25T07:35:26","publication_year":"2012","word_count":370,"keywords":["IIT Bombay","programming_languages:R","AI","R","Quantum Computer","RAG","ViT","analytics","iit data science"],"extracted_tech_keywords":["AI","analytics","RAG","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/study-iit-iim-in-analytics\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":18159,"title":"Where’s The Money In Artificial Intelligence? Find out areas that have proven the money value in AI","content":"Despite the stupendous leap in advances in artificial intelligence and machine learning, the underlying sentiment has always been about getting a formidable return on investment. None of these advances would matter if it doesn’t give a robust ROI. And all the cheery news about AI startup acquisitions point to one thing – tech behemoths are in the game for reaping money in the long run. At present, Google has made the maximum investment in AI acquisitions. So how does one make AI useful? Well, if you are a startup you need to do your market due diligence and find out what the customer wants and the needs that can be addressed through AI. It is still early days in AI and most enterprises and startups are struggling to find a good business model for AI or machine learning. In this article, Analytics India Magazine shines a spotlight on ways AI startups across the globe and in India are making money by presenting real-world examples of how startups are extending the capabilities of AI: 1. Making AI profitable with Structured Data: Palo Alto startup Diffbot founded by Mike Tung becomes the Google of structured data. Defined as the key to AI, structured data has proven its money value with startups and financial enterprises investing heavily in structured data. According to Palo Alto-based structured data startup founder Mike Tung, cash-rich digital natives such as Google, Baidu and Amazon have the human muscle to spider web and gather data, small companies lack the tech muscle and an inventory of structured data to build a database of structured knowledge. So great has been the response to Diffbot, founded in 2010 that the company’s taxonomy contains 1.2 billion objects and adds 10 million objects per day. As a matter of fact, Tung reportedly revealed that Google’s Knowledge Graph recently crossed one billion objects. Tung, a Stanford dropout is tackling the biggest problem startups and other companies face today – presenting data in a structured way so that AI systems can read it. The startup has some bold-faced names as clients listed in its roster – Yandex, eBay, Cisco, Adobe and even AOL. Upside: With structured data being the key to achieving business objective, this could be a way for companies\/startups to make data useful for small and medium businesses Downside: No proven business model in India Who your customers would be: Mid-size to large consumer product companies who want to tap into what their customers want 2.Early Bot builders are minting money: Bot business is booming in India. Think Niki.ai and Singapore-based Active.ai which are the talk of town, flushed with VC money and a great customer base. According to chatbot developer Ekim Kaya, chatbots are a popular go-to strategy for AI startups since they have the potential to scale and have a clear business model – there is the SaaS Bot model for B2B and B2C SMBs and big enterprises alike. Kaya believes as the market matures, the profitability will increase and founders will find new ways to monetize bots besides developing APIs. He emphasizes the rise of “Bot Tester” and “Bot Conversational Flow Designer” in the future. Upside: You don’t have to develop a separate app, one can run it in a messenger Facebook or Slack Another major advantage is that this has a proven business model and VCs have shown a great interest in chatbots The go-to business model is that of Software as a Service (SaaS), and is delivered as an API While there is the traditional business model of ad-selling on bots peddling native and sponsored content, data gathering can also become another revenue generator for chatbots that are a great tool for collecting user data (location, email, preferences) and interests. This data can help brands understand the audience better Downside: The market is flooded with chatbots that are being deployed by big enterprises to minimize front line customer service support staff and handle low-level queries. Right now, there is way too much competition in the bot world. To build a profitable chatbot one must understand the user psyche well to ensure that the end product is not deemed spammy. Who your customers would be: Well, the most successful use cases are coming out of the finance, e-commerce, healthcare and retail sector wherein bots are automating simple tasks and have increased operational efficiency. 3. Customized solutions like image recognition & text recognition: Over the years, compelling AI-optimized use cases have emerged for areas such as image recognition, text analysis, speech to text conversion that can deliver better outcomes. While tech giants Google, Microsoft and Apple are integrating AI across their user interfaces, such as personalizing search and marketing, small businesses that lack the economic muscle are benefitting from AI-powered forecasting that is driving marketing operations and personalization. Case in point – Boxx.ai that rolled out its AI-powered product AIDA, aimed at marketers to help personalize their customer touch, thereby improving engagement, increasing transactions and reducing attrition. Another Bangalore-based startup Artifacia has an AI-powered visual discovery platform that is revolutionizing visual search for e-commerce companies. Upside: With more and more value-adding use cases emerging over the years and increased adoption of AI at scale, budding startup enthusiasts can identify the problem areas they want to focus on and build capabilities around it. Downside: There is always the fear of crashing out and finding not enough VC interest to sustain. Who your customers would be: Right now, the sectors that are showing the most investor are financial services, high tech and telecommunications, automotive, followed by healthcare, retail, CPG and education. 4. Selling Cognitive Software:  Here’s a thought, if Watson open sources its APIs in the near future, this could lead to customized AI solutions. The commoditization of cognitive tools could lead to more AI-focused solutions across services (forecasting, optimization) and products (semiconductors). If the developers starting using a vendor’s open-source cognitive tools, this could also help in introducing a standard AI software. Right now, Google’s Cloud ML has taken concrete step to allow programmers create solutions via Tensorflow. Upside: In India, use cases are still weak. Downside: One will have to follow a Follow a test and learn approach Who your customers would be: Given the significant advances in artificial intelligence, cognitive solutions can benefit all sectors, particularly healthcare, retail, education and even manufacturing.","excerpt":"Despite the stupendous leap in advances in artificial intelligence and machine learning, the underlying sentiment has always been about getting a formidable return on investment. None of these advances would matter if it doesn’t give a robust ROI. And all the cheery news about AI startup acquisitions point to one thing – tech behemoths are […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-10-09T06:58:57","publication_year":"2017","word_count":1052,"keywords":["machine learning","artificial intelligence","AI","chatbots","ML","image recognition","Aim","analytics","TensorFlow","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","TensorFlow","chatbots","image recognition","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/wheres-the-money-in-artificial-intelligence-find-out-areas-that-have-proven-the-money-value-in-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042604,"title":"8 Best Statistics Institutes In India","content":"India celebrated the 15th National Statistics Day on the birth anniversary of Prasanta Chandra Mahalanobis. He is best known for introducing ‘The Mahalanobis Distance,’ a statistical measure used in many software programs today, and for founding the Indian Statistical Institute (ISI). Mahalanobis was instrumental in setting up the National Sample Survey to collect data on socio-economic parameters of the country. The Indian government has extensively used these tools to frame informed policies. Check: NIRF Ranking of Engineering Colleges On the occasion of the ‘National Statistics Day,’ we have curated a list of top statistics institutes in India. 1. Indian Statistical Institute (ISI) Indian Statistical Institute (ISI), founded in 1931, is headquartered in Kolkata. At present, it has four subsidiary centres in Delhi, Bengaluru, Chennai and Tezpur. ISI also has a network of Statistical Quality Control and Operations Research units at Coimbatore, Vadodara, Hyderabad, Giridih, Mumbai, and Pune, guiding industries in India and abroad, to develop quality management systems and solve critical problems related to quality, reliability and productivity. Lately, the government of India has been sponsoring multiple Covid-19 projects proposed by ISI researchers. 2. IIT Kanpur The mathematics and statistics department at IIT Kanpur is one of the premier departments in the country, providing excellent teaching and research in mathematical sciences, including statistics. The programmes motivate research in mathematical science and train computational scientists working on challenging problems. The placement of students in the MSc statistics programmes is almost 100% in the last ten years. IIT Bombay, IIT Tirupati, IIT Delhi, IIT Roorkee, IIT (BHU), and IIT Guwahati also offer statistics courses for students. 3. St. Stephen’s College Founded in 1881, St. Stephen’s College is one of the oldest colleges in Delhi. The college offers various mathematics and statistics courses, covering calculus, multivariable calculus, probability and statistics, software (Maxima\/Mathematica), biostatistics, etc. In addition, the faculty at St. Stephen have done their research from prestigious institutions including Princeton, Rutgers, Oxford, Cambridge, IITs, IISc and JNU. 4. St. Xavier’s College Kolkata St. Xavier’s College was founded in 1860. The Department of Statistics was started back in the late fifties. The alumni have been placed in top MNCs in insurance, finance, risk management, portfolio analysis and development of statistical software.  St. Xavier’s College’s  Mumbai campus also offers various courses in statistics. 5. Lady Shri Ram College for Women Lady Shri Ram College for Women is one of the premier institutions of higher learning for women in India. The college offers various courses in statistics, marketing, finance, and strategy making. In addition, students are trained on applied statistics, statistical methods and analysis, and work in the areas such as surveys, econometrics, biostatistics, and operations research. The department of statistics at Lady Shri Ram College for Women teaches students rigorous methods, tools and techniques to sift through a maze of data. In addition, students also conduct practicals based on computer language C and software packages like Excel, Word, and statistical packages for social sciences (SPSS). 6. Madras Christian College (MCC) The department of statistics at MCC was established in 1968. Located in Chennai, the department offers statistics, computer applications, managerial economics and operational research. It has a well-established ‘Gift Sironomey Statistical Computing Laboratory.’ The department believes in practical training of its students by conducting public opinion, socio-economic and tribal surveys. 7. Loyola College The department of statistics at Loyola College was founded in 1982. With 14 faculty members, the department provides consultancy services to various institutions and companies, alongside teaching students different statistical methods for economics, biostatistics and more. It offers various UG and PG courses in statistics. 8. SP College Pune Savitribai Phule Pune University has one of the leading statistics departments in the country, and it is the only centre for advanced studies in statistics under the UGC and CAS scheme. The department of mathematics and statistics at SP College was established in 1953. However, the department of statistics was hived off in 1976. The students get placed in leading global software and pharmaceutical companies, including Pfizer, Novartis, Systat, Ideas etc. Some students have taken up jobs in RBI, Indian Statistical Services and Bureau of Economics and Statistics.","excerpt":"India celebrated the 15th National Statistics Day on the birth anniversary of Prasanta Chandra Mahalanobis. He is best known for introducing ‘The Mahalanobis Distance,’ a statistical measure used in many software programs today, and for founding the Indian Statistical Institute (ISI). Mahalanobis was instrumental in setting up the National Sample Survey to collect data on […]","categories":["AI Trends"],"tags":["indian statistical service","mathematics and statistics"],"author_name":"Amit Naik","publish_date":"2021-06-30T13:00:00","publication_year":"2021","word_count":685,"keywords":["Go","mathematics and statistics","programming_languages:R","AI","programming_languages:Go","ViT","R","indian statistical service"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-most-popular-statistics-institutes-in-india\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":34402,"title":"Meet The Winners Of 40 Under 40 Awards – The Leading Data Scientists From India Recognised At MLDS 2019","content":"MLDS 2019 40 Under 40 Awardees with Bhasker Gupta, CEO & Founder, Analytics India Magazine India’s largest machine learning conference MLDS 2019 which brought 690+ delegates on a single platform announced the winners of 40 Under 40 awards — recognizing India’s leading data scientists and their contribution to the data science industry. The summit was hosted by Analytics India Magazine, India’s leading data science and analytics magazine, which provides cutting-edge analysis, research, meetups and hackathons. The magazine also covers topical issues dominating the analytics landscape in India. The 40 Under 40 awards, instituted for the first time is a step towards acknowledging the data scientist community in India and their impact on businesses and their technological leadership. 2019 winners were announced on Day 2 of MLDS 2019, Bangalore which concluded with the awards ceremony. The 40 under 40 awards were presented in collaboration with International School of Engineering, INSOFE, located in Hyderabad and Bangalore. Winners were determined by a judging panel comprising of editors and industry veterans. Since its inception, INSOFE has been highly successful in creating true Data Science talents through its in-depth classroom programmes. With hundreds of alumni working with top Fortune 500 companies, INSOFE is Asia’s largest classroom training and research-driven academic institution. Expressing pride in these winners and acknowledging their contributions, an official spokesperson from INSOFE shared, “INSOFE takes pride in honouring some of the brightest talents in the country and their exemplary contributions to the Data Science community. MLDS Conference is a promising platform to honour the future leaders in Data Science through 40 Under 40 awards. We would like to congratulate all the winners and we wish them all the best in their future endeavours”. MLDS 2019 was the forum where data scientists and analytics practitioners from across India came together to discuss the new techniques available for applying ML at scale for solutions. The conference that attracted top-level researchers was of great interest to attendees who used the opportunity to gain hands-on insights from workshops on machine learning and AI and also network with their peers. Abhay Jha, Head of Analytics, Indiabulls receiving the 40 Under 40 award at MLDS 2019 held on Jan 31, in Bangalore “MLDS 2019 showcased a gathering of likeminded AI enthusiasts and a great platform for Developers to learn and share insightful tech talks and masterclass sessions. Overall, it was a good show and a great connect with the community,” said Sachin Kelkar, Head – APJ Developer & Partner Programs. The conference affirmed the impact of Data Science in India and its relevance across the domains. It showed how data science and analytics is not confined to analytics companies only, but has relevance across industries, as evidenced by the insightful sessions presented by leading tech companies like Microsoft, AWS, Amazon, Intel but also included companies like AB InBev and Amity that are utilising data science and analytics to optimise their business strategies and increase bottomline revenue. Please find the complete list of winners here","excerpt":"India’s largest machine learning conference MLDS 2019 which brought 690+ delegates on a single platform announced the winners of 40 Under 40 awards — recognizing India’s leading data scientists and their contribution to the data science industry. The summit was hosted by Analytics India Magazine, India’s leading data science and analytics magazine, which provides cutting-edge […]","categories":["Deep Tech"],"tags":["40 Under 40 Awards","Data Scientist Awards","mlds india","mlds nimhans"],"author_name":"Richa Bhatia","publish_date":"2019-02-04T08:23:30","publication_year":"2019","word_count":496,"keywords":["Data Scientist Awards","data science","Go","machine learning","AWS","AI","cloud_platforms:AWS","programming_languages:R","ML","mlds india","analytics","40 Under 40 Awards","R","mlds nimhans"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","AWS","R","Go","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/meet-the-winners-of-40-under-40-awards-the-leading-data-scientists-from-india-recognised-at-mlds-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10052593,"title":"Is Depth In Neural Networks Always Preferable? This Research Says the Contrary","content":"Deep neural networks are defined by their depth. However, more depth implies increased sequential processing and delay. This depth raises the question of whether it is possible to construct high-performance “non-deep” neural networks. Princeton University and Intel Labs researchers demonstrate that it is. Characteristics Of Depth The fields of machine learning, computer vision, and natural language processing have been transformed by deep neural networks (DNNs). As its name implies, one of the primary characteristics of DNNs is their depth. They have a large depth, which can be defined as the longest path between an input neuron and an output neuron. Often, a neural network can be characterised as a linear sequence of layers with no intra-group connections. In these circumstances, the depth of a network is defined by its layer count. It is widely believed that a significant depth is required for high-performance networks, as depth boosts a network’s representational capability and aids in learning increasingly abstract characteristics. Indeed, one of the key reasons for ResNets‘ success is that they enable extremely deep networks with up to 1000 layers. As a result, state-of-the-art performance is increasingly attained by training models with a high degree of depth, and the definition of “deep” has moved from “two or more layers” in the early days of deep learning to “tens or hundreds of layers” in today’s models. Is Deeper Necessary? However, is a great depth always necessary? The depth is an important issue to ask because great depth does not come without downsides. For example, a deeper network results in increased sequential processing and delay; it is also more difficult to parallelise and is, therefore, less appropriate for applications that require rapid response times. Contrary to popular belief, the researchers discovered that this is indeed possible. They describe a non-deep network design that outperforms its deep equivalents. The researchers referred to the design as ParNet (Parallel Networks). They demonstrate for the first time that a classification network with a depth of 12 can achieve higher than 80% accuracy on ImageNet, 96% on CIFAR10, and 81% on CIFAR100. Additionally, the researchers demonstrate that a detection network with a shallow (12) backbone may obtain a 48% AP on MS-COCO. ParNet assists in addressing a scientific question regarding the necessity of great depth and provides practical benefits. ParNet may be efficiently parallelised over several processors due to its similar substructures. Research Contributions To summarise, there are three contributions: • For the first time, the researchers demonstrate that a neural network with a depth of 12 may perform well on extremely competitive benchmarks (80.7% on ImageNet, 96% on CIFAR10, 81% on CIFAR100). • The researchers demonstrate how ParNet’s parallel structures can be used for fast, low-latency inference. • The researchers examine ParNet scaling requirements and demonstrate how they can be effectively scaled while maintaining a continuous low depth. Code is available at Non-Deep Networks The researchers do this by layering parallel subnetworks rather than one layer after another. The current research contributes to the effective reduction of depth while keeping a high level of performance.  The researchers analyse the design’s scaling rules and demonstrate how to improve performance without altering the network’s depth. Finally, the researchers demonstrate the feasibility of using non-deep networks to construct low-latency recognition systems. Conclusion The researchers established for the first time empirical evidence that non-deep networks can compete with deep networks in large-scale visual recognition benchmarks. They demonstrated that similar substructures could be leveraged to generate remarkably performant non-deep networks. Additionally, the researchers demonstrated methods for scaling up and optimising the performance of such networks without expanding their depth. The work demonstrates alternate designs for highly accurate neural networks that do not require deep networks. Such designs may be more suitable for future multi-chip processors. Moreover, the researchers anticipate that the work will aid in the construction of highly precise and rapid neural networks.","excerpt":"Non-deep networks could be utilised to create low-latency recognition systems, rather than deep networks.","categories":["AI Features"],"tags":["Deep Learning","Deep Neural Networks","ImageNet","Neural Network"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-10-29T11:00:00","publication_year":"2021","word_count":642,"keywords":["Neural Network","API","machine learning","TPU","AI","neural network","computer vision","RAG","deep learning","Deep Neural Networks","ResNet","Deep Learning","R","ImageNet"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","RAG","TPU","R","API","ResNet"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/depth-in-neural-networks\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":18724,"title":"How Is Artificial Intelligence Being Benefitted By Neuroscience","content":"Artificial intelligence has taken leaps and bounds in terms of the extent of task that it can accomplish. And a few of them such winning games like Go, Poker and Chess, composing music or penning down horror stories only prove its capability in terms of posing an intelligence at par with humans. But despite all the high profile developments around AI, researchers believe that we haven’t reached the mark of super intelligent AI as yet and the current AI programs are narrow in abilities and can be tricked easily. Although they have been showcasing some jaw dropping potentials, they lack in “human” capabilities such as common sense, intuition, imagination and human intelligence in general. And what could be a better model than replicating human brain to create “intelligence”. AI experts from across the globe agree to the fact that there can be no better alternative than looking at the biological human brain as a model, which in itself is a complex structure of modular subsystems defining important functions. Demis Hassabis, founder of Google’s AI powerhouse DeepMind, even commented that to build better computer brains, we need to look at our own. Biomimicking is becoming one of the approaches that is gaining momentum as a way to develop artificial intelligence, and this is where neuroscience comes into play. It is a common belief that drawing inspiration from neuroscience and psychology is resulting into development of technologies like deep learning and reinforcement learning. AI drawing inspiration from Neuroscience Experts most firmly believe that collaboration between AI and neuroscience can produce an understanding of the mechanisms in the brain that generate human cognition. It therefore becomes important to build the AI with the ability to think, reason and learn flexibly and rapidly. And to get there, it is important to understand the inner workings of human mind! The researchers claim that neuroscience provides a rich source of inspiration for newer algorithms, thereby helping in creating true artificial intelligence. It may offer a more mathematical and logic based methods and ideas thus surpassing the traditional approaches in AI. Studying human brain’s cognitive system would also mean that there are better insights into what’s more relevant and hence help in developing smarter AI. Are AI systems close to human brain yet? Though the idea of achieving artificial intelligence using neuroscience may sound quite tempting, there is a long way that needs to be covered by machines to get close to human’s way of thinking. It can be largely attributed to human’s incomplete knowledge of biological brains and the nature of consciousness. However, with the advent in technologies like genetic bioengineering and imaging, researchers can dive further into exploring the complex neural network with much proficiency. The scientists are closer to creating novel neural architectures that are capable of posing human-like learning, reasoning, creativity, imagination etc., hinting towards those times when machines would be more than capable to dealing with complex, real-world problems. Use case by IBM Taking an inspiration from brain, IBM research team recently used machine learning techniques to develop computational models of attention and memory. With an underlying aim of building learning AI systems that can adapt to new environments and retain what they have learnt so far, IBM created these new models. They deployed two important innovations enabling both short term and long term adaption. They developed an algorithm that learns to quickly focus its attention on the right input based on a reward. “The higher the reward, the more attention it will place on a certain piece of input”, the company’s blog noted. These finding were reported in their paper titled “Context-Attentive Bandit:  Contextual Bandit with Restricted Context.” The blog further explained that the novelty of their algorithms is the ability learn which inputs to focus on in an online manner i.e. the dataset is not fixed, but constantly changing, while receiving a reward for making decisions based on partial inputs. “Online means that the system can learn as it performs, and therefore is robust to changes. The company is currently testing this on a set of real-life datasets and problems with more complex environments”, it said. Another technique that the company is developing is based on neuroplasticity that enables long time learning and is inspired by the adult neurogenesis process which happens in the hippocampus, the part of the human brain responsible for forming memories. They have successfully demonstrated that not only their algorithm adapts to a new environment but also preserves memories of the previous domains, thus making a step towards lifelong learning AI systems. Conclusion It cannot be denied that for AI to progress and evolve beyond its current state, it has to achieve an intelligence as complex as humans, and for that collaboration of AI experts with neuroscientists is going to become the next big thing. However, to get the functionality of brain in full detail can be quite challenging. Few tasks like getting like working in the sub conscious state of mind can be difficult for a machine to achieve. Nevertheless, the experts believe that these challenges would in fact continue to inspire more research to dive into the field and build adaptive lifelong learning systems which is quite close to what human brain is.","excerpt":"Artificial intelligence has taken leaps and bounds in terms of the extent of task that it can accomplish. And a few of them such winning games like Go, Poker and Chess, composing music or penning down horror stories only prove its capability in terms of posing an intelligence at par with humans. But despite all […]","categories":["IT Services"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-11-06T05:50:56","publication_year":"2017","word_count":869,"keywords":["Go","API","machine learning","artificial intelligence","AI","neural network","ML","Aim","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","Aim","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/artificial-intelligence-benefitted-neuroscience\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":19573,"title":"How Zara Uses Big Data Analytics to Excel in its Game?","content":"My sister brought Satin high heel court shoes from Zara and as it was looking really good she prompted me to buy it as well. To my ill luck when I went to the Zara store one week later it was already sold out. It took me by surprise because usually such merchandise in other brands would normally remain in the store for weeks together. Zara’s model is different which I learnt afterwards and that is based on fast fashion. The idea here is to stock the store racks with the merchandise which is as per the latest trends in the fashion industry as soon as possible. How Zara is able to achieve this through big data will be covered shortly. You can get to know how exactly Zara uses big data if you have a bird’s eye view of the technology. I can only think of Intellipaat institute when providing competent big data training is the requirement. How traditional fashion industry lags behind? Most of the clothing brands and fashion stores run on a bi-annual or seasonal basis where there are long production lead times and typically low cost-centers like China and Bangladesh are chosen for manufacturing. That is why customers get their favorite fashion merchandise after waiting a particular amount of time during which these brands manufacture the merchandise at bulk. When such merchandise is not sold by the end of season it is given as discounts which is a loss to the clothing brand. How Zara trumps over such inefficiencies by using big data? Inditex is the parent company of Zara which produces over 84 crore garments in a year. The majority of this is sold by Zara. Each product has a unique RFID tag associated with it through which it is possible to track the product from warehouse to stores. Inditex has a central data processing unit which is open 24 hours a day which monitors all the movement of merchandises. Inventory management, design, distribution is achieved by the various teams of Inditex by monitoring over 6,000 of its outlets. Simple big data analytics help make all this possible. When a product arrives at a store the RFID chip is used to determine which items need replenishing, their location which has made Zara more reflective of the customer needs. In fact, the sales team can track and deliver those products to the customers through this RFID tag. This sales tracking data is key to Zara’s success which increases the products which are most liked by the customers and stops manufacturing those products which don’t attract customer attention. Zara has an elite team of 350 designers who receive constant feedback from this sales data. Big data analytics is greatly leveraged in this phase of design in Zara. Some lady didn’t like the long belt in hand bag or a man didn’t like more than one pocket in the jacket. Designers will design the merchandise based on this feedback provided by the sales professionals who use the analytics on likes and dislikes. Majority of factories of Zara are located in Europe and North Africa close to their business which helps them to deploy their manufactured merchandise to stores within 2-3 weeks. This greatly lowers the cost of holding excess inventory. Also Zara doesn’t sell items at discounts as it doesn’t overdo the manufacturing process. Conclusion Zara’s competitors sell about 2,000 to 4,000 various clothing items per year. Can you guess how much Zara sells? A whopping 11,000 clothing items per year. It also exults that it has the lowest inventory levels at the year end. Zara only manufactures 15 to 25% of the merchandise before the season begins. Over 50% of the merchandise is designed and manufactured after the season starts based on what becomes popular among the customers. This quick replenishment cycle model is what drives demand for Zara’s products. Other clothing brands are trying to emulate Zara’s model and grow big.","excerpt":"My sister brought Satin high heel court shoes from Zara and as it was looking really good she prompted me to buy it as well. To my ill luck when I went to the Zara store one week later it was already sold out. It took me by surprise because usually such merchandise in other […]","categories":["IT Services"],"tags":["Big Data Analytics"],"author_name":"AIM Media House","publish_date":"2017-12-08T07:15:36","publication_year":"2017","word_count":655,"keywords":["big data","Go","ELT","programming_languages:R","AI","programming_languages:Go","RAG","Big Data Analytics","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","ELT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/zara-uses-big-data-analytics-excel-game\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":5283,"title":"10 Most Influential Analytics Leaders in India &#8211; 2014","content":"The success of analytics story in India has a huge bearing on the leaders that drive it today. The emergence of analytics and big data posed its own set of opportunities\/ challenges and these leaders are to a large extent instrumental in shaping the analytics industry that we witness today. Analytics India Magazine’s annual ranking of the 10 Most Influential Analytics Leaders in India honors individuals in Analytics\/ Big Data industry who are deemed by their peers and an expert panel to be the most influential individuals in India, in terms of Impact, Leadership, Entrepreneurship and Analytics evangelism. Pankaj Kulshreshtha – Business Leader, Analytics & Research at Genpact 6 of the 10 leaders listed here have their roots in Genpact or a GE subsidiary. This goes on to display the amount of impact that the company has had on the analytics industry in India. Besides being among the early starters in Analytics, Genpact is also considered among the most successful, and have churned out the best analytics talent over the years. Genpact has about 5,000 people in its analytics practice, largely out of Bangalore. Pankaj Kulshreshtha leads the Analytics and Research practice globally at Genpact. This practice is part of Genpact’s Smart Decision Services that enables clients across industries to make smarter decisions in sales and marketing, cost, and risk management using data and insights. Under his leadership, Genpact’s Analytics business has grown to one of the largest and most extensive practices in the industry. [pullquote align=”right”]Work Experience: 19 years; Past Organizations: GE, L&T; Education: PhD from IIM Bangalore, BE from VNIT, Nagpur[\/pullquote]Pankaj joined Genpact in 1998 and was one of the first people who launched offshore analytics teams in India. At Genpact, Pankaj helped various GE businesses set up analytics teams and in 2005 moved to GE Money UK, where he led risk management for various portfolios and was Chief Risk Officer for the loans business in his last role. He then rejoined Genpact in 2008. [divider] Rohit Tandon – Vice President, Strategy WW Head of HP Global Analytics As head of Global Analytics, Rohit Tandon is helping drive the Analytics ecosystem to support HP’s vision and priorities. He is a veteran industry executive known for pioneering and guiding new businesses to success. In Rohit’s leadership HP Global Analytics grew from a small team to a large analytics organization. Today as part of HP’s Corporate Strategy team, Global Analytics is also chartered to lead HP’s Analytics delivery ecosystem, and bring together similar teams to drive innovations that support HP’s enterprise priorities. The organization has access to talent as well as cutting edge tools and techniques that help innovate & customize analytical solutions for sustained competitive advantage. [pullquote align=”right”]Work Experience: 20+ years; Past Organizations: IBM, Genpact, Accenture Consulting, JWT and Ampersand[\/pullquote]Prior to joining HP, Rohit was the Executive Vice President at IBM’s Global Processing Services unit where he was responsible for Strategy, IT, Quality and Transformation. During his tenure he was instrumental in setting up and leading IBM’s Knowledge Services business including Analytics where he also acquired a company for IBM. Prior to IBM, Rohit drove the Analytics business at Genpact, leading a team of ~2500 analytics practitioners. Rohit is a Certified Master Black Belt in Six Sigma and a Certified Quality Leader in Six Sigma. He also holds a patent in his name for optimization algorithms designed by him. [divider] Sameer Dhanrajani – Business Leader, Cognizant Analytics Among the leading IT service providers in India, Cognizant Technology Solutions is considered the most aggressive in terms of Analytics. Cognizant Analytics resides within H3 (Horizon 3), as a part of Cognizant’s Emerging Business Accelerator (EBA) tower and is a pivotal component in the SMAC stack (Social, Mobility, Analytics and Cloud). Cognizant Analytics unit is regarded as a distinguished market leader and differentiator through incisive focus on topical actionable, applied and prescriptive analytics based solutions coupled with focused consulting approach, IP based non-linear platforms and deeply entrenched customer centric engagement model. Sameer Dhanrajani is spearheading Cognizant Analytics practice. In his capacity, Sameer is leading end-to-end business spheres of Cognizant Analytics and is responsible for crafting differentiated strategies around analytics consulting, platforms and services coupled with creating best of breed GTM, business development, operational excellence solutioning exercises and delivering transformational analytics engagements. [pullquote align=”right”]Work Experience: 17 years; Past Organizations: Fidelity, Genpact, Alliance[\/pullquote]Prior to Cognizant, Sameer was the Country Head for Fidelity National Financial and pioneered India’s first captive to be based on nonlinear growth model with platform based value propositions. He has also served as Vice President – Analytics at Genpact and Director – Alliance Consulting Inc. in his previous assignments. As a recognized industry thought leader in outsourcing, analytics space; Sameer has been recipient of “Outstanding Leadership Award” at India Human Capital Summit and “Exemplary Leader Award” during the Asia Pacific HRM Congress Awards. Sameer is also core member of the NASSCOM Analytics Special Interest Group and is on industry advisory council of ISB (Indian School of Business) and has been quoted across multiple business media and news publications. [divider] Srikanth Velamakanni – Co founder and Chief Executive Officer at Fractal Analytics Fractal was among the earlier Indian start-ups in the area of Business Analytics, and also one the most successful boutique analytics firm too. [pullquote align=”right”]Work Experience: 16 years; Past Organizations: ICICI, ANZ; Education: MBA from IIM Ahmedabad, B.Tech from IIT, Delhi[\/pullquote]Srikanth leads Fractal’s global presence in business intelligence, consumer insights, predictive analytics, and optimization sciences, delivered through visual story-telling. Prior to co-founding Fractal, he consulted with major market leaders such as Visa, P&G, Citibank, HDFC Bank and SAP in predictive analytics for risk management and marketing effectiveness, as well as structured debt transactions and collateralized bond obligations at ANZ Investment Bank and ICICI. Read Interview of Srikanth Velamakanni [divider] Pankaj Rai – Director, Global Analytics at Dell Pankaj Rai is the Director of Dell Global Analytics (DGA). Pankaj has been with Dell for ~8 years and has been with DGA for close to 5 years. Prior to this he was working with the India President’s office and managed all strategic and corporate planning related initiatives of Dell in India. In this role, he was responsible for helping Dell diversify and grow its footprint in India as also represent Dell outside in industry forums. [pullquote align=”right”]Work Experience: 20 years; Past Organizations: GE Capital, ICICI, Standard Chartered; Education: MBA from IIM Ahmedabad, B.Tech from IIT, Delhi[\/pullquote]Prior to joining Dell, Pankaj was Head, Program Management Office at Standard Chartered Bank in their Singapore regional office wherein he helped grow their shared service centers in India and Malaysia. In his professional career spanning 20 years, Pankaj started as a management consultant and then went on to work in the financial services industry in India wherein he worked with ICICI and GE Capital in a variety of roles covering sales, risk management, 6 sigma & operations. [divider] Amit Khanna, Partner at KPMG Amit joined KPMG India in Jan 2014 after 18 years of experience as a head of Data Analytics Practice for domestic Businesses. Prior to joining KPMG, he was a Partner and Managing Director with Accenture Management Consulting where he led the global consulting practice for Business Analytics, Customer insights, Sales and Marketing transformation. He started the Accenture Analytics consulting practice in India in 2008 serving both Indian & Global clients and grew the team size to 500+ resources in a span of 5 years. Amit has extensive experience in running analytics driven transformation\/ programs for the clients. He was one of the founder members in building advance analytics capability for Accenture Management Consulting and was part of Accenture advance analytics global leadership team. [pullquote align=”right”]Work Experience: 18 years; Past Organizations: Accenture, Genpact[\/pullquote]Amit has also done lot of work in areas of analytics capability development & analytics adaption in Organizations. He has worked extensively with various universities to develop Analytics & Data scientist curriculum. He personally own 2 analytics patents, he has also worked with large global clients to help them design their analytics organization & adapt a fact based culture. Before Accenture, Amit was service delivery lead for customer and marketing analytics team in GECIS (now Genpact). He has helped developing GE analytics team for NA, Europe & APAC. He has held senior positions in sales and marketing leadership in consumer goods industry before moving to consulting about 10 years back. Read Interview of Amit Khanna [divider] Ashish Singru – Director eBay India Analytics Center Ashish has over 18 years of experience in driving analytics and customer insights into strategic planning and decision-making in a wide range of organizations. He currently heads eBay’s global center of excellence for analytics in Bangalore, with the aim of building cutting-edge capability to solve eBay’s hardest business problems. Ashish has a unique perspective on how the application of data, technology & business analytics has evolved in organizations in the last 2 decades. When he started his career in mid-1990’s, analytics was primarily used by Financial services organizations. Other sectors like FMCG, Retail and Automotive were much more on mining customer survey data to generate business insights. He was one of the early pioneers in driving behavior-based analytics in these sectors. Subsequently, with the rapid growth of internet and enhanced database technologies, he has led the development of cutting-edge analytics practices to understand how online and offline customer behavior can be analyzed, as well as how to drive analytics deeper into the financial investment planning process for high-stakes decision on sales, marketing and pricing for major product and marketing campaign launches at global scale. [pullquote align=”right”]Work Experience: 18 years; Past Organizations: Microsoft, SABMiller, Gallup; Education: MBA from University of Iowa, MS from Rutgers University, B.Tech from IIT, Kanpur[\/pullquote]As an organizational leader, Ashish has extensive experience in creating and developing talent pool across multiple markets. He has mentored and coached numerous individuals in the analytics profession ranging from college-intern stage to senior leadership roles in different industries. Ashish believes in being the change you want to see in the world, and constantly works on developing himself along various professional dimensions in the rapidly changing industry so that he can help lead others in their growth paths as well. [divider] Arnab Chakraborty – Managing Director, Analytics at Accenture Consulting As a managing director at Accenture Analytics, now part of Accenture Digital, Arnab Chakraborty speaks analytics fluently. He serves as the Global Lead for Industry Analytics in Accenture’s advanced analytics practice and is also responsible for driving the Big Data Analytics practice. Arnab has over 15+ years of rich experience in consulting and business analytics across diverse industry sectors. Before joining Accenture, Arnab played a pivotal role in scaling up enterprise wide analytics capabilities globally for the past ten years at Hewlett Packard (HP) and GE Capital (Genpact). Prior to that, he held leadership positions at KPMG Consulting, Wipro Technologies, and Larsen & Toubro. He has worked across High-Tech and Electronics, Financial Services, Consumer Goods and Manufacturing industries. [pullquote align=”right”]Work Experience: 15 years; Past Organizations: HP, GECIS; Education: MBA from NITIE, B.Tech from NIT, Rourkela[\/pullquote]Arnab is recognized as an Edelman Laureate by the prestigious Franz Edelman Committee\/INFORMS for his contributions in the field of e-commerce\/ online analytics and has been voted by SSON in 2012 as one of the top 6 Global Sourcing Think-tank (G6) in Asia. Arnab is a regular author in the field of analytics and operations research in leading journals and magazines. He is also one of the key industry leaders driving the NASSCOM Analytics Special Interest Group Arnab holds an MBA (Gold Medalist) from National Institute of Industrial Engineering (NITIE), Mumbai, India and a degree in Mechanical Engineering from National Institute of Technology (NIT) Rourkela, India. He is certified in Production and Inventory Management from APICS, USA. [divider] Anil Kaul – CEO and Co-founder at Absolutdata Dr. Anil Kaul is a well-known expert in the industry with over 16 years of experience in marketing research, strategic consulting. and quantitative modeling. He has consulted over 20 Fortune 500 companies during the 4 years he spent with McKinsey & Co. in New York. Thereafter, he joined Anubis Inc., an innovative data warehousing start-up in San Francisco Bay Area, as a part of the four-member top management team. The company was acquired by Personify. He then co-founded AbsolutData in 2001. [pullquote align=”right”]Work Experience: 18 years; Past Organizations: McKinsey, Personify; Education: PhD from Cornell University[\/pullquote]Anil is a PhD and MS in marketing from Cornell University. He is also a voracious reader and an intuitive writer. He has published many articles in leading management and academic journals such as McKinsey Quarterly, Marketing Science, Journal of Marketing Research, and International Journal of Research in Marketing. One of his papers was nominated for “Paul Green Award” by the American Marketing Association. He has also been an invited speaker at McKinsey & Co., Dartmouth College, Cornell University, Yale University, Columbia University, Wharton, UC Berkeley, New York University, IIFT, IMT Ghaziabad, and IIM Lucknow. [divider] Dr. N.R.Srinivasa Raghavan, Senior Vice President & Head of Analytics at Reliance Industries Limited Dr. N. R. Srinivasa Raghavan is Sr. Vice President & Head of Big Data Analytics\/Data Science at Reliance Industries Ltd. Working closely with the Chairman and the Senior Business Leadership, he is setting up the Centre of Excellence for Big Data Analytics at RIL capable of delivering deep business insights across Oil & Gas, Petrochemical, Telecom & Life Sciences businesses. He is responsible for strategizing and realizing the value proposition for Data Science and Data Governance within the group companies while building and mentoring a world class team of data scientists \/ business analysts to deliver for high performance. He has traveled widely, and has extensive industrial experience prior to RIL in strategizing and architecting large scale, global, and complex data science implementations in the Financial Services, Automotive\/Semiconductor\/PC Manufacturing and Supply Chain verticals. [pullquote align=”right”]Work Experience: 18 years; Past Organizations: Citigroup, GM, Fair Issac; Education: M.Tech. & PhD from IISc Bangalore[\/pullquote]He holds award winning Ph. D in Computer Science and M. Tech in Operations Research both from Indian Institute of Science, and is a gold medalist in his B. Tech in Mechanical Engineering from Sri Venkateswara University, Tirupati. He began his career with 8 years of academic research\/teaching\/consulting experience as a Professor at the Indian Institute of Science. In his past 7 years in industry, he has held senior leadership roles in advanced analytics delivery and innovation at Dell, Fair Isaac, General Motors and Citibank. He has 85 technical publications in leading International Journals\/Conferences with 2 International patents filed and has also supervised many award winning doctoral students while at IISc. He has received many awards including the Outstanding Young Associate of Indian Academy of Sciences, Outstanding Young Engineer of Indian National Academy of Engineering, and Young Scientist Fellowship from Department of Science & Technology, GoI.","excerpt":"The success of analytics story in India has a huge bearing on the leaders that drive it today. The emergence of analytics and big data posed its own set of opportunities\/ challenges and these leaders are to a large extent instrumental in shaping the analytics industry that we witness today. Analytics India Magazine’s annual ranking […]","categories":["AI Features"],"tags":["Absolutdata","analytics ceo","analytics leaders","Fractal Analytics","genpact analytics"],"author_name":"Дарья","publish_date":"2014-02-21T12:58:28","publication_year":"2014","word_count":2459,"keywords":["Fractal Analytics","data science","Go","API","analytics ceo","analytics leaders","AI","R","predictive analytics","Git","RAG","Aim","analytics","genpact analytics","Absolutdata"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","predictive analytics","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/10-most-influential-analytics-leaders-in-india-2014\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10015918,"title":"Infogain Acquires AI &#038; Analytics Service Provider Absolutdata","content":"In the latest news, CrysCapital-backed digital platform and software engineering services provider Infogain acquired San Fransico-headquartered Absolutdata. The leader is an AI-based solution provider company that offers a range of advanced analytics and data science services. Infogain On Growth Mode Infogain seems to be on a growth mode, as the acquisition of Absolutdata comes after its recent acquisition of Silicus Technologies and Revel Consulting. Now with the latest deal, Infogain will add Absolutdata’s NAVIK AI platform to its portfolio, which offers pre-built and customisable solutions and services to make enterprise AI-ready. NAVIK AI platform includes SaaS solutions such as ASK NAVIK, NAVIK SalesAI, NAVIK MarketingAI, NAVIK ResearchAI and NAVIK TradeAI. Infogain hopes to derive rapid growth, improve its revenue growth, and create more resilient customer relationships across industries. “To meet client demand for digital transformation, Infogain pursued an acquisition strategy that emphasized strategy, experience, cloud transformation, and advanced analytics services. The Absolutdata acquisition completes that strategy, providing clients with game-changing analytics and AI solutions that provide new sources of insights and better decision-making capabilities to drive faster business growth,” said Sunil Bhatia, CEO, Infogain. Corroborating the same, COO of Infogain, Ayan Mukerji further says that the acquisition is going to benefit the combined clientele of Infogain and Absolutdata. Speaking of NAVIK AI, in particular, Mukerji says that this SaaS platform will help in ‘dramatically’ increasing the analytics customer base. “At the same time, we can now offer strategy and experience, data estate, and cloud transformation services to Absolutdata customers, strengthening the available product and service portfolio. In this acquisition, everyone wins,” he points out. A Deal That Benefits All As pointed by the acquisition is a win for both the parties. Now as part of a larger organisation, Absolutdata’s data scientists and solution engineers will be able to leverage Infogain’s strategy, experience, and cloud technology expertise to deliver AI and analytics solutions. Expressive the company’s excitement on joining Infogain’s team, Dr Anil Kaul said, “This acquisition opens up career growth possibilities for our employees and being part of a larger company with an expanded skillset will enable us to deliver innovative, end-to-end solutions to both companies’ clients as a unified team.”","excerpt":"In the latest news, CrysCapital-backed digital platform and software engineering services provider Infogain acquired San Fransico-headquartered Absolutdata. The leader is an AI-based solution provider company that offers a range of advanced analytics and data science services. Infogain On Growth Mode Infogain seems to be on a growth mode, as the acquisition of Absolutdata comes after […]","categories":["AI News"],"tags":["Absolutdata","Mergers and Acquisitions"],"author_name":"Shraddha Goled","publish_date":"2020-12-24T23:10:42","publication_year":"2020","word_count":361,"keywords":["data science","Go","API","AI","digital transformation","Git","RAG","analytics","GAN","Mergers and Acquisitions","R","Absolutdata"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","Go","Git","API","GAN","digital transformation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infogain-acquires-ai-analytics-service-provider-absolutdata\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10120120,"title":"Sam Altman Proposes ‘Universal Basic Compute’ for Global Access to GPT-7&#8217;s Resources","content":"Sam Altman recently proposed ‘Universal Basic Compute,’ in which everyone would have access to a portion of GPT-7’s computing resources. “I wonder if the future looks something more like Universal Basic Compute than Universal Basic Income, and everybody gets like a slice of GPT-7 compute,” said Altman, in the recent episode of the All-in podcast. He said that this would not be in the form of money but as a share of productivity that could be used, sold, or even donated for initiatives like cancer research, etc. Altman said that they began exploring Universal Basic Income (UBI) around 2016. “About the same time, we started taking AI really seriously,” he added, considering the potential changes AI might bring to society, employment, the economy, and even social contract. “I am not like a super fan of how the government has handled most policies designed to help poor people, and I kind of believe that if you could just give people money they would make good decisions, and the market would do its thing,” reminisced Altman, saying that he is pretty much in favour of lifting up the poor and eliminating poverty. He believes that the existing social safety net is broken. “Giving people money is not going to solve all the problems. It is certainly not going to make people happy. But it might solve some problems. It might give people a better horizon with which to help themselves,” said Altman, adding that things are now changing and there should be a better alternative than traditional UBI. Ergo, Universal Basic Compute. This new development also comes in the backdrop of Altman’s controversial statement: “I don’t care if we burn 50 billion a year, we’re building AGI.” “Whether we burn $500 million, $5 billion, or $50 billion a year, I don’t care. I genuinely don’t as long as we can stay on a trajectory where eventually we create way more value for society than that and as long as we can figure out a way to pay the bills,” he said. Meanwhile, OpenAI is gearing up to make some major announcements next week. “Not GPT-5, not a search engine, but we’ve been hard at work on some new stuff we think people will love! Feels like magic to me,” said Altman. Sources reveal that it is likely to release an AI voice assistant that enables a human-like interaction with better logical reasoning capabilities and reduced latency, mostly powered by GPT-4-Lite, GPT-4-Auto, and GPT-4-Lite-Auto.","excerpt":"“I am not a super fan of how the government has handled most policies designed to help poor people,” says OpenAI chief.","categories":["AI News"],"tags":["Sam Altman"],"author_name":"Siddharth Jindal","publish_date":"2024-05-11T08:55:25","publication_year":"2024","word_count":411,"keywords":["Go","Sam Altman","OpenAI","GPT-5","AI","programming_languages:R","GPT","ViT","GAN","R","llm_models:GPT"],"extracted_tech_keywords":["AI","GPT-5","OpenAI","R","Go","GPT","GAN","ViT","llm_models:GPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sam-altman-proposes-universal-basic-compute-for-global-access-to-gpt-7s-resources\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":33867,"title":"What Makes India An Artificial Intelligence Specialist?","content":"To no one’s surprise, a recent survey announced that India is the 13th most advanced country the world in terms of the development of artificial intelligence technologies. Over the past decade, India has flourished greatly in AI. But what is the reason behind it? “The road ahead for AI depends on and will be driven by ‘human intentions’. It is our intention that will determine the outcome of AI,” Indian Prime Minister Narendra Modi had said after inaugurating the Wadhwani Institute of Artificial Intelligence at the University of Mumbai’s Kalina Campus. What Makes India So Good? In the last few years, numerous entrepreneurs have come up with AI-based services in applications ranging from farming to healthcare and education. We have also seen the government undertaking initiatives to adapt AI in its policies as well. It has partnered with several AI institutes to implement AI projects. It is these AI projects that have helped the country do so well in the field. Also, partnering with various reputed international AI institutes has helped the country develop an effective AI strategy. Government Support: The government has shown great support in leveraging AI. NITI Aayog’s unique brand called #AIforAll is an approach that focuses on how India can leverage the transformative technologies to ensure social and inclusive growth in line with the development philosophy of the government. It aims to enhance and empower human capabilities to address the challenges of access, affordability, shortage and inconsistency of skilled expertise. It also aims to tackle some global challenges from the AI perspective, be it an application, research, development, technology or responsible AI. With this initiative, it has extended its support in the areas of healthcare, agriculture, education, smart cities and infrastructure and smart mobility and transportation sectors. One example of the NITI Aayog using AI is its development of an AI-based crop yield prediction model for a real-time advisory to farmers on crop yield and pest outbreak. Not just the national government, but the state governments are putting their best foot forward in making use of AI. For example, the Andhra Pradesh government has been using cloud management and data storage in governance to resolve issues of the people in the state. Talent Pool: There are more than 200 AI-based startups in India that are helping different sectors. It has progressed so much in recent times because experts from a different set of advanced computing power have come together to help areas from farming to healthcare. Big Indian companies such as Bharti Airtel and Reliance Jio are setting up AI labs. Indian IT services giants such as Infosys and Wipro are making huge investments in AI. IT industry body Nasscom is setting up Centre of Excellence (COE) in AI in Karnataka and Telangana with companies like IBM, Microsoft, NVIDIA, Intel and AWS, on the lines of its successful 10,000 Startups program. Collaborations And Partnerships: There are many collaborations and partnerships that the country involves itself in when dealing with the advancement of AI. These partnerships are not just within the country but also across the globe. Global corporations NVIDIA, Microsoft and Google have all set up AI labs in India. Tech giant Microsoft is working with the state government to predict dropout rates in government schools in the state of Andhra Pradesh. It makes use of AI to predict with some margin of error, who among the current cohort are likely to drop out. Concluding Note IT industry body NASSCOM Vice President KS Viswanathan had said that India has formed a policy group to study new technologies and recommend a framework for their adoption. “We all are currently working out on a policy framework on AI.” India is getting very serious about AI. With a rich spread of talent in the sector of computer science and data science and the related fields, the country is making its mark in the AI world market. With a more success rate, more talent will eventually be pooled into this glamorous field in the country and we are already on our march to raising our rank in the AI competition.","excerpt":"To no one’s surprise, a recent survey announced that India is the 13th most advanced country the world in terms of the development of artificial intelligence technologies. Over the past decade, India has flourished greatly in AI. But what is the reason behind it? “The road ahead for AI depends on and will be driven […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI in India","Government","Microsoft","Narendra Modi","NITI Aayog"],"author_name":"Disha Misal","publish_date":"2019-01-22T10:58:00","publication_year":"2019","word_count":680,"keywords":["data science","Go","artificial intelligence","AI in India","Microsoft","AI","AWS","Government","RAG","responsible AI","Aim","Narendra Modi","GAN","NITI Aayog","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","data science","Aim","RAG","AWS","R","Go","GAN","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-makes-india-an-artificial-intelligence-specialist\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10139968,"title":"AMD is Finally Catching up with NVIDIA Blackwell","content":"AMD CEO Lisa Su said on October 30 that the company’s MI325 advanced GPU accelerator is a game-changing product that will set off against NVIDIA’s H200, while its MI350 series will compete against Blackwell. With NVIDIA beating them in the GPU race, Su expects the company to close the performance gap faster than anticipated with their new hardware series. Previously, Andrew Dieckmann, CVP and GM of data centre GPU at AMD, also acknowledged the same. “We are trying to take very representative benchmarks that are realistic. I can tell you that in our customer engagements, especially regarding inference workloads, we have yet to find a single workload that we cannot outperform NVIDIA on,” he said, adding that AMD doesn’t always outperform NVIDIA. “However, if we optimise for a specific solution, we can beat them”, added Dieckmann. Su also subtly pointed out AMD’s aggressive stance in the AI GPU market, positioning itself as a challenger to NVIDIA, during the quarterly earnings call. Announcing the third quarter earnings, Su raised its AI guidance revenue to exceed $5 billion, up from the previous quarter’s forecast of $4.5 billion in the wake of the AI boom. Betting big on AI developments, Su calls it the beginning of the “AI super cycle”, referring to anticipated demand across all computing domains. That said, she acknowledged the challenges of sustaining growth in an intensely competitive market, with hyperscalers selectively deploying AMD products. “This is really the beginning of what I like to call an AI super cycle. We believe AI is going to be in every aspect of computing” said Su. A New Record for AMD In Q2 FY24, AMD’s revenue stood at $6.8 billion, an 18% increase year-on-year, with a gross margin of 50%. The company’s operating income remained at $724 million, while the net income was $771 million. Su said, “We delivered strong third quarter financial results with record revenue led by higher sales of EPYC and Instinct data centre products and robust demand for our Ryzen PC processors,” AMD CFO Jean Hu also said that it is making strategic investments to position AMD as the end-to-end AI infrastructure leader and drive long-term profitable growth. During the quarter, AMD generated $3.5 billion in revenue from its data centres, up 122% year over year. This growth can be attributed to increased shipments of AMD’s Instinct GPU and EPYC CPU. “Our gross margin improvement has been primarily driven by the mix, especially Data Center business continues to be the strong growth driver of our business, accounting for more than 50% of our revenue mix,” revealed Hu, emphasizing AMD’s gross margin improvement driven by the data centre segment. Lisa Su also emphasised AMD’s software advancements with the ROCm stack, aiming to offer an “open alternative” to NVIDIA’s closed software ecosystem, appealing to customers wanting flexibility and cost-effective AI compute options. “ROCm now provides AI developers with a truly open software alternative that has been deployed and validated at scale,” said Su. Their newly launched Ryzen AI Pro 3000 series focuses on industry-leading performance in the personal computing sector. “Our third-generation AI-enabled processors deliver unprecedented AI processing capabilities with incredible battery life and seamless compatibility for the applications users depend on”, said Jack Huynh, senior VP and GM of Computing and Graphics Group at AMD. Lisa Su also expects the data centre AI accelerator’s TAM to reach $500 billion by 2028 and is projected to grow by more than 60% annually. Lumpy future ahead She, however, cautioned about an uncertain future as AMD prepares to ramp up production and optimise for larger deployments—quite the opposite of NVIDIA’s rapid, consistent growth in AI segments. “We feel very good about the growth opportunities. I would say that it might be lumpy. In general, these are large customer acquisitions and it’s not always predictable exactly which quarters you would expect the significant build-out” said Su, pointing towards AMD’s realistic approach. Su hinted that the faster cadence of AMD’s AI products (the likes of MI300, MI325, and MI350) was deliberate, helping them close the competitive gap more quickly. This accelerated pace could strain AMD’s resources but underscores their commitment to becoming a serious contender. “We have successfully accelerated our product development pace to deliver an annual cadence of new Instinct products. Our next-gen MI350-series silicon is looking very good and is on track to launch in the second half of 2025 with the largest generational increase in AI performance we have ever delivered”, said Su. However, NVIDIA is already a step ahead of Blackwell, with their Rubin GPUs set to release in 2026, and AMD is also planning to launch their MI400X series GPU at the same time. Acquisitions and Partnerships Galore In the client segment, their revenue was up by 29% year-on-year at $1.9 billion. AMD’s recent partnerships inside the ecosystem include industry leaders like Oracle, which is using its Instinct GPU accelerators to power its OCI computer supercluster to manage AI workloads. AMD extended its support to the development of Meta’s Llama 3.2 open-source models and wants to assist developers in building agentic applications and other groundbreaking AI products using its hardware. They’ve also partnered with Microsoft to enable Copilot + on their AMD-powered AI PCs. Su acknowledged that capturing market share from entrenched competitors in the data centre GPU segment requires time and patience. Building trust with hyperscalers like Microsoft, Meta, and others has been a focus, suggesting that AMD still faces hurdles in gaining broader adoption but remains committed to long-term partnerships. “This is a multi-generational journey. We’ve always said that. We feel very good about the progress. I think next year is going to be about expanding both customers as well as workload,” said Su. AMD closed the acquisition of Silo AI to accelerate the deployment of AI models on AMD’s hardware. AMD also recently announced an agreement to acquire ZT Systems, an AI computing infrastructure provider. The acquisition is subject to regulatory clearances and is likely to be completed by the first half of 2025. Moreover, AMD and Intel also announced a collaboration to broaden the x86 ecosystem by developing an advisory group that will include Dell, HP, Lenovo, Meta, Microsoft, Oracle, and Red Hat. “Establishing the x86 Ecosystem Advisory Group will ensure that the x86 architecture continues evolving as the compute platform of choice for developers and customers,” said Lisa Su. AMD’s projected revenue in Q4 2024 is set to rise towards $7.5 billion, with an approximate gross margin of 54%. Jean Hu also said, “We are on track to deliver record annual revenue for 2024 based on significant growth in our Data Center and Client segments.”","excerpt":"Su cautioned about an uncertain future as AMD prepares to ramp up GPU production and optimise for larger deployments—quite the opposite of NVIDIA’s rapid, consistent growth in AI segments.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","AMD","NVIDIA"],"author_name":"Supreeth Koundinya","publish_date":"2024-10-31T19:09:48","publication_year":"2024","word_count":1101,"keywords":["Go","AMD","API","programming_languages:R","AI","ML","programming_languages:Go","Aim","Rust","llm_models:Llama","NVIDIA","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Rust","API","llm_models:Llama","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/amd-is-finally-catching-up-with-nvidia-blackwell\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":18545,"title":"Deep Thomas, Former Tata Executive Joins Aditya Birla Group As Its Chief Data And Analytics Officer","content":"In another large profiling hiring in the analytics industry, Aditya Birla Group has hired Deep Thomas, a former Tata Group executive, as its chief data and analytics officer. As a part of his new role, Thomas would be responsible to create a group-wide customer database supporting the diversified conglomerate’s products and sales strategy. This is the second major hiring by the Group in the past one month, as it had earlier appointed PepsiCo India’s chairman D Shivakumar as the head of strategy and business development. He replaced Saurabh Agrawal, who had left Aditya Birla to join Tata as its chief financial officer. Santrupt Misra, director of group human resources at Aditya Birla Group, was reportedly quoted as saying that given the kind of consumer data that is being generated in various industries, it is important to garner insights from this data to prepare solutions and offerings to cross-sell and up-sell, and there is tremendous opportunity in this area. It is reported that Thomas would be setting up his own team that would work on short term, medium term and long term goals. Along with it, he will also work with the data analytics team at companies across the group. Misra commented that after analysing the way businesses are working and recognising the white spaces on which no one is working on, Thomas would formulate strategy to tap on to these missed opportunities. Thomas has been the founder CEO of Tata Insights and Quants, where he was in charge of enabling customer centricity across its companies through advanced analytics services and products.","excerpt":"In another large profiling hiring in the analytics industry, Aditya Birla Group has hired Deep Thomas, a former Tata Group executive, as its chief data and analytics officer. As a part of his new role, Thomas would be responsible to create a group-wide customer database supporting the diversified conglomerate’s products and sales strategy. This is […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-10-26T08:13:24","publication_year":"2017","word_count":261,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/deep-thomas-former-tata-executive-joins-aditya-birla-group-chief-data-analytics-officer\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":17967,"title":"Second Edition Of Great Learning Data Science Awards Concludes By Recognizing The Best Players In Analytics Industry","content":"The Great Learning Data Science Excellence Award Winners at Cypher 2017 The second edition of Great Learning Data Science Awards concluded with much aplomb at the recently concluded Cypher 2017, which was the third year of its successful running as the largest analytics summit in India. The awards were brought in association with Great Lakes Institute of Management, and recognised the achievers in the analytics space. Presented to companies in five different categories, Great Learning Data Science Awards saw participation from leading organisations as well as startups from the analytics industry. The five categories were Emerging Analytics Services startup of the year, Emerging Analytics Product Startup of the year, Best Boutique Analytics firm of the year, Best Data Science Project of the year and Best Big Data Implementation of the year, which identified the best in industry and recognised their stellar achievements in this space. “Great Learning Data Science awards was instituted to recognise and reward the exciting work being done by various companies in the space of Analytics and Data Science. We are delighted to have presented the second edition of Great Learning Data Science awards at Cypher 2017”, said Hari Krishnan Nair, Co-Founder, Great Learning. He further added “This year, we saw an increase in the number of applications. Also it was amazing to see the diversity in terms of companies that applied. We had established banks like HDFC, Kotak and Yes Bank with their nominations as well some very exciting startups like Touckin, Tiger Analytics, 3LOQ etc. We also received applications from leading technology companies like Ericsson and Dell. It was great to see the exciting work happening in adoption of analytics and big data technologies across industries like Banking, Telecom, Healthcare, e-Commerce and FMCG. We congratulate all the winners and also the companies who sent in their nominations”. Bhasker Gupta, CEO and Founder, Analytics India Magazine said “The response from established companies as well as startups was overwhelming. Great Learning Data Science Awards, which was initiated last year, has set a benchmark in the analytics space and we look forward to grace more players in this space year after year”. Anand Bharadwaj, VP, Customer Success with Hari Krishnan Nair, Co-Founder, Great Learning Under the award category Best Boutique Analytics Firm of the year, Tiger Analytics and G Square emerged as clear winners. Tiger Analytics, an analytics and data science consulting partner offers portfolio of services such as advanced analytics, predictive analytics, machine learning, visualisation etc., while G-Square provides plug and play products in the analytics, machine learning and big data space. “The new wave of analytics solutions are blending signals from disparate and unstructured sources — the latest technologies and algorithms are helping us achieve what was hitherto not possible. As a company whose foundation lies in this new wave of analytics, we are thankful to Analytics India Magazine for this wonderful recognition”, said Pradeep Gulipalli, Co-Founder, Tiger Analytics. Gurpreet Singh, Co-founder & Director, G Square Solutions with Hari Krishnan Nair, Co-Founder, Great Learning Gurpreet Singh, Co-founder & Director, G Square Solutions noted “G-Square Solutions is proud to be awarded the Boutique Analytics Company of the year in the Great Learning Data Science Awards at Cypher 2017. The Cypher summit was a rendezvous of the most intelligent analytics minds of the country & the awards provided the recognition to the companies who are doing their best to serve Clients in nurturing their true analytics potential. We at G-Square Solutions are striving to provide the best Global analytics solutions for the Financial Services Industry” In the next category of Best Big Data Implementation of the year, the winners were Clairvoyant India Pvt Ltd., a global technology consulting and services company and HDFC + 3LOQ Labs, who had the joined nomination. Kaushik Ghate, Senior Vice President – Head Marketing Analytics, HDFC and Anirudh Shah, Founder, 3LOG Labs with Hari Krishnan Nair “It is a great honour to be recognized by Cypher for “Best big data implementation of the year” and it is truly rewarding for all the hard work our teams have put in over the last one year. Our commitment to reliable, high-quality solutions to help solve challenging big data problems for our customers has been a true differentiator. It has enabled our continued growth, success at Clairvoyant”, said Shantanu Mirajkar, Co-Founder, CTO, Clairvoyant India. Whereas Anirudh Shah, Founder, 3LOG Labs said “It’s always a great pleasure to find ourselves among like-minded professionals and experts in the field. But winning The Great Learning awards has made being a part of Cypher 2017 that much more special. It’s been hugely motivating for the team and we’ll keep pressing forward with the work we’re doing for Habitual AI.” Habitual AI by 3LOQ Labs is a result of a year of honing machine-learning algorithms with data about customer habits and consumer psychology. Team Touchkin with Hari Krishnan Nair Blueocean Market Intelligence and Scienaptic Systems were the winners in the category Emerging Analytics Product Startup of the year, while the Best Data Science Project of the year was awarded again to HDFC Bank + 3LOQ Labs in joint collaboration, whereas Dell and Touchkin, were the other winners. The project by Dell was submitted by the Performance Analytics Group (PAG) at Dell EMC, while HDFC+3LOQ Labs’ first of its kind Analytics Automation Architecture led them to win the award. Touchkin on the other hand offers world’s first AI-enabled coach and chat platform for mental and behavioural health support. Shantanu Mirajkar and Josh Vedam Co-Founders, Clairvoyant India. “This project was truly unique. Diabetes related distress is known to be a lead indicator for lack of adherence, future complications and poor glycemic control. Currently, this is very hard to assess and takes time and effort through face-to-face conversation. With mobile passive sensing, we can quickly, proactively and non-intrusively identify high-risk patients for triage and intervention which leads to a large improvement in clinical outcomes and large savings for the provider. The team handled many real-world challenges on data acquisition, preparation and interpretation and built a robust model that is now being used on a daily basis by one of India’s largest healthcare companies”, said Ramakant Vempati, Co Founder, Wysa-Touchkin. “Cypher has provided excellent platform for data science professionals across sectors to come together and benefit from each others’ experiences, success stories and expertise. Being enlightened with new innovative stuff in data science industry through the platform provided by Cypher was more rewarding than the award itself! The awards event provides a great opportunity to companies to showcase the great work being done in the field of analytics and inspires participants to raise the bar for the next edition!”, said Kaushik Ghate, Senior Vice President – Head Marketing Analytics, HDFC on a concluding note.","excerpt":"The second edition of Great Learning Data Science Awards concluded with much aplomb at the recently concluded Cypher 2017, which was the third year of its successful running as the largest analytics summit in India. The awards were brought in association with Great Lakes Institute of Management, and recognised the achievers in the analytics space. […]","categories":["Deep Tech"],"tags":["learning data science","machine learning blending"],"author_name":"Srishti Deoras","publish_date":"2017-10-03T06:13:06","publication_year":"2017","word_count":1119,"keywords":["big data","data science","Go","machine learning","AI","R","learning data science","machine learning blending","automation","analytics","GAN","predictive analytics"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","predictive analytics","R","Go","big data","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/second-edition-great-learning-data-science-awards-concludes-recognizing-best-players-analytics-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10170691,"title":"This Company Simulates Training For Indian Commandos, Creates a Living Atlas","content":"As industries wake up to the power of location intelligence, Geographic Information System (GIS) technologies are becoming integral to sectors such as urban planning, agriculture, transportation, and disaster management. These tools empower organisations to make informed decisions, streamline operations, and deliver services with greater impact. The global GIS market is experiencing significant growth. Projections indicate an increase from $32.97 billion in 2024 to $55.75 billion by 2029, reflecting a compound annual growth rate (CAGR) of 11.1%. In India, the geospatial solutions sector is poised for substantial expansion. It is expected to grow from $23.5 billion in 2024 to $79.3 billion by 2030, driven by advancements in 3D scanning and spatial analytics. When it comes to geospatial technology in the country, Esri India is a heavyweight you can’t ignore. The company, led by managing director Agendra Kumar since 2013, is quietly powering millions of daily users and thousands of government organisations with its GIS technology. “We have close to 6,500 organisational customers in the country. In terms of users, maybe 1.2 million people use some app running on our technology every day,” Kumar told AIM in an exclusive interview. The scope for Esri is quite vast—from national mapping agencies to utilities like electric distribution companies, city gas firms, and telecom giants such as Jio and Airtel. “About 70% of our revenue comes from the government,” he added, explaining the deep integration with various state and national bodies. Digital Earth, Living Atlas and AI Esri India is not just a software company; it creates geospatial technology and invests heavily in data. A flagship initiative is its ‘Living Atlas’, a cloud-hosted repository that curates openly published data from government sources. “If it’s not geo-referenced, somebody from our team will work on that, geo-reference it. If it’s available in PDF or Excel files, we geo-reference and bring that data into our Living Atlas, which is a very large repository of data now.” Kumar explained that more than a thousand layers of data are available in this. This freely accessible data resource is complemented by over 200 solution products developed to tackle common challenges such as disaster management, flooding, forest fires, drainage, and water distribution—all created by Indian teams for the country’s niche needs. A significant part of Esri India’s recent technological advances is in AI and machine learning. Kumar described how these are integrated for practical uses, with the initial use cases being primarily technical. This feature extraction is simply just object identification. Citing examples from traffic analysis to identifying defence vehicles or aircraft, he explained, “Human identification can be done. For example, we had done something in the office. We counted the number of people passing near the reception [using this]. Even face recognition can be done if it is required.” Keeping Secrets and Boundaries Handling sensitive data, especially in defence, demands tight security and ethical care. Kumar stressed that Esri India never holds defence data. “Their data remains on their premises on their computers. So even if there’s an application to be done, often our people will go to their office and work.” Esri India’s Indo ArcGIS is a comprehensive GIS software platform that enables organisations to create, manage, and analyse spatial data. It supports mapping, data integration, and location-based intelligence for diverse sectors including government, utilities, defence, and urban management. With capabilities like AI, machine learning, and 3D visualisation, ArcGIS helps users gain actionable insights from geographic data, improving decision-making and operational efficiency across India. The ArcGIS platform can be on the cloud or totally on-premises, not necessarily requiring a connection to the internet. Hence, this allows the solution to stay with the client. Highlighting the mutual trust in this arrangement, he said, “They also do not want to expose what they have, and we also don’t want to get into those things.” Defence and Security GIS Esri India’s engagement with defence and security agencies dates back long before Kumar joined the company. “Many of them have become advanced users of our technology. They don’t really need our help every day to use it. They know how to use it because they also deal with a lot of confidential information,” he said. The company supports organisations such as DRDO and homeland security agencies with mapping and weather modelling, which are crucial for extreme conditions like cold, snow, and avalanches. Kumar highlighted the importance of location intelligence in security operations. According to him, GIS is known as ‘the science of where’. It captures all types of information, such as what is happening and where an incident has taken place. This information then gets converted to a map and is available to the decision makers. Real-time situational awareness is vital in conflict. “Systems have to be created well in advance because you can’t create them when the action has to be taken,” Kumar added. Esri India processes drone data, providing both cloud and on-premise solutions, including 3D visualisation tools that support navigation, target setting, asset management, and strategic planning. “So, they are very realistic images, creating fly-throughs and walk-throughs. This helps in navigation, target setting, target identification,” Kumar said. While the data collected for defence remains strictly with the forces themselves, Esri’s technology supports its use in simulations, training, and planning. “If there is a monument which needs to be protected, and there is a suspicion that someday terrorists may attack it, you want to train some commandos. So you can use a 3D model to train commandos on how to go inside, what to expect inside,” Kumar noted. The Road Ahead Though the GIS sector in India is still emerging, Kumar is optimistic about its trajectory. He attributed this to recent government policies, adding that till 2021, there was a lack of clarity on geospatial data policies. The companies were initially uncertain about rules surrounding acquisition. This is why the new set of guidelines came out in 2021, followed by the National Geospatial Policy in 2022. “So these two documents, along with remote sensing data policy and drone policy, have made the whole ecosystem quite conducive for business,” he added. This clarity has enabled wider adoption beyond government into private sectors like manufacturing, retail, BFSI (banking, financial services, and insurance), and logistics. “Success for retail distribution, optimisation in manufacturing and logistics, insurance claims settlement—these are some of the things which are upcoming,” Kumar said. He also pointed to the rising interest in 3D mapping and digital twins, technologies that go beyond pretty visualisations to solving real problems such as urban flooding and infrastructure management. He acknowledged that the true value of a digital twin will come when it’s used to solve a problem, be it traffic, urban flooding or improving the utility infrastructure. The Indian government’s flagship projects, such as Gati Shakti, SVAMITVA, and Naksha, have been pivotal in boosting the geospatial ecosystem by creating foundational mapping data for villages and cities. Kumar noted that the industry still involves “a lot of service orientation and not so much of technology orientation”. He expects this to shift with time as startups and educational institutions increasingly embrace geospatial innovation.","excerpt":"Esri India’s ArcGIS platform supports mapping, data integration, and location-based intelligence.","categories":["Deep Tech"],"tags":["AI in defence","defence","digital twins","GIS","training simulator"],"author_name":"Sanjana Gupta","publish_date":"2025-05-24T10:18:36","publication_year":"2025","word_count":1175,"keywords":["Go","machine learning","defence","AI","ETL","digital twins","ML","R","Git","Aim","analytics","Rust","training simulator","GIS","AI in defence"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","R","Go","Rust","Git","ETL"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/this-company-simulates-training-for-indian-commandos-creates-a-living-atlas\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":43633,"title":"BAIDU’s ERNIE 2.0 Gets NLP Top Honours, Eclipses BERT &#038; XLNet","content":"Machine learning models which are deployed for vision and in natural language processing (NLP) tasks usually have more than one billion parameters. This allows for better results as the model generalises over a large wide range of parameters. Pre-trained language representations such as ELMo, OpenAI GPT, BERT, ERNIE 1.0 and XLNet have been proven to be effective for improving the performance of various natural language understanding tasks. Pre-training binarised prediction models help us understand common NLP tasks like Question Answering or Natural language Inference. Bidirectional Encoder Representations from Transformers or BERT, which was open sourced late 2018, offered a new ground to embattle the intricacies involved in understanding the language models. BERT boasts of training any question answering model under 30 minutes. Given the number of steps BERT operates on, this is quite remarkable. However, BERT contains artificial symbols like [MASK], which result in discrepancies during pre-training. In comparison, AR language modelling does not rely on any input corruption and does not suffer from this issue. AR language modelling and BERT possess their unique advantages over the other. And XLNet was a by-product of the search for a pre-training objective that brings the advantages of both while avoiding their flaws. As a generalised AR language model, XLNet does not rely on data corruption. Hence, XLNet does not suffer from the pre-train-finetune discrepancy that BERT is subject to. Going Beyond BERT And XLNet BERT constructed a bidirectional language model task and the next sentence prediction task to capture the co-occurrence information of words and sentences; XLNet constructed a permutation language model task to capture the co-occurrence information of words. However, besides co-occurrence, the researchers at Baidu believe that there are other valuable lexical, syntactic and semantic information in training corpora. For example, named entities, such as names, locations and organisations, could contain conceptual information. The authors in their paper, explore whether it would it be possible to further improve performance if the model was trained to constantly learn a larger variety of tasks. They address the same by introducing ERNIE 2.0 (Enhanced Representation through kNowledge IntEgration) ERNIE 2.0 is a continual pre-training framework. Continual learning aims to train the model with several tasks in sequence so that it remembers the previously learned tasks when learning the new ones. What ERNIE 2.0 Has To Offer via Baidu As shown in the above figure, the architecture of continual pre-training contains a series of shared text encoding layers to encode contextual information, which can be customised by using recurrent neural networks or a deep Transformer consisting of stacked self-attention layers. The parameters of the encoder can be updated across all pre-training tasks. This framework by Baidu differs from traditional pre-training methods in that, instead of training with a small number of pre-training objectives, it could constantly introduce a large variety of pre-training tasks to help the model efficiently learn language representations. To compare with BERT, the researchers used the same model settings of a transformer as BERT. The base model contains 12 layers, 12 self-attention heads and 768-dimensional of hidden size while the large model contains 24 layers, 16 self-attention heads and 1024-dimensional of hidden size. The model settings of XLNet are the same as BERT. ERNIE 2.0 is trained on 48 NVidia v100 GPU cards for the base model and 64 NVidia v100 GPU cards for the large model in both English and Chinese. via Baidu NLP: AI’s Need Of The Hour The way these top organisations competing to top the NLP charts hints at the utmost need for improvement in the natural language understanding tasks while also shows how difficult it is to set new benchmarks which are significant. This can be clearly seen in the difference in scores between the top frameworks. The numbers may look close but every minor improvement in an NLP framework is considered gold given these model’s reputation for being computation heavy. Improvements such as ERNIE are a great addition to the world of machine learning, especially to the NLP community. ERNIE 2.0’s features can be summarised as follows: In this framework, different customised tasks can be incrementally introduced at any time and are trained through multi-task learning. This framework can incrementally train the distributed representations without forgetting the parameters of previous tasks. ERNIE 2.0 not only achieves SOTA performance but also provides a feasible scheme for developers to build their own NLP models. ERNIE 2.0 outperforms BERT and XLNet on 7 GLUE language understanding tasks and beats BERT on all 9 of the Chinese NLP tasks. Know more about this work here.","excerpt":"Machine learning models which are deployed for vision and in natural language processing (NLP) tasks usually have more than one billion parameters. This allows for better results as the model generalises over a large wide range of parameters. Pre-trained language representations such as ELMo, OpenAI GPT, BERT, ERNIE 1.0 and XLNet have been proven to […]","categories":["AI Trends"],"tags":["Baidu","BERT","NLP","XLNet"],"author_name":"Ram Sagar","publish_date":"2019-07-31T18:15:05","publication_year":"2019","word_count":756,"keywords":["Go","XLNet","machine learning","OpenAI","AI","neural network","AWS","Transformers","BERT","NLP","Aim","Baidu","R"],"extracted_tech_keywords":["AI","machine learning","neural network","NLP","OpenAI","Aim","Transformers","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/baidus-ernie-2-0-gets-nlp-top-honours-eclipses-bert-xlnet\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10064706,"title":"About 76 percent of the audience encountered a biased algorithm","content":"Bias in AI algorithms is nothing but the mirror of societal bias that has been ingrained for many years now. These biases will remain if not acted upon to have an equal and diverse world not just in tech but also otherwise. Talking about women in Data Science and AI, Anjali Iyer, Delivery Excellence Business Leader at The Math Company, shared some of her personal experiences at The Rising 2022, Women in AI conference organised by Analytics India Magazine on Friday. Watch all the recorded sessions of Rising 2022 here>> Recalling her schooling days, Anjali said, “During my days, when my parents were getting me enrolled in a school, I don’t remember seeing an application form where my mother’s name or occupation was asked. It was all about the father’s name and occupation. I know the times are changing, at least now; when I enrolled my daughter in a school, I could see that I was able to spell my name and my occupation, but that wasn’t the case when I was getting into the school. So there is definitely a change that we are getting into, but again how many women are getting into the tech industry and how many are able to get into the leadership roles, there is still a wide gap.” 76% have encountered biased algorithms Traversing through experiences with women in Data Science and AI, Anjali asked the audience if they had ever encountered a biased algorithm. About 76 per cent of the audience at the conference voted “Yes”, which speaks volumes about the existing bias. “The representation of women in AI and tech matters a lot, and we need to change this. But it is not just about representation but also about women getting the right opportunity, as there is a huge difference in the way men and women get similar opportunities while getting into tech. Ten years ago, even though I did not get the opportunity to do my MS as marriage was a priority, things are changing slowly,” she added. Citing an example of an American multinational company, Anjali spoke about a health app that was criticised for ignoring women’s health issues. The app could track every piece of data except for women’s natural cycles. This happened because of the bias, which is inbuilt in these algorithms, and it happened only because there was a lack of diversity. Dismal number of women in STEM Looking at the STEM jobs, there is just about 28 per cent women representation and even more startling in the field of AI\/ML research, where there are just 15 per cent women in the industry. “These figures mean that the organisations will fail to harness the fullest capacity of their digital innovations without including women. These machine learning technologies will be fed with a constant stream of biased data, eventually producing junk results and not giving a holistic picture and eventually causing harm. It will be like one bias leading to another, and we will get into that loop”, Anjali added. Female AI voice assistants Taking another poll with the audience, Anjali asked which voice assistant they would pick for their home, and nearly 76 per cent of the audience at the conference picked a female voice. Anjali said that it was worthwhile to note that both men and women have expressed higher interest in female gender synthetic voices. Women reported an 11.9 per cent preference, and men showed about 14.3 per cent. “Typically, these AI bots and voice assistants reinforce gender bias because as we move along in this digital world, we all know that the world may soon have a higher number of voice assistants than people. These voice assistants, be it as hotel staff, our IVR calls or the childcare providers, have traditionally featured female sounding voices, and female sounding voices projected on these technologies reinforce an impression that women typically hold assistant jobs and should be servile and docile,” she concluded.","excerpt":"Today, the mother’s name and profession are essential in a child’s school application form, unlike in the past.","categories":["Deep Tech"],"tags":["AI Assistant","themathcompany"],"author_name":"Poornima Nataraj","publish_date":"2022-04-12T11:00:00","publication_year":"2022","word_count":657,"keywords":["data science","Go","machine learning","AI","AI Assistant","innovation","ML","themathcompany","Git","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","R","Go","Git","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/about-76-per-cent-of-the-audience-encountered-a-biased-algorithm\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10103761,"title":"Google Cloud Runs Ad on Sphere during AWS re:Invent at Las Vegas","content":"In an unexpected turn of events, Google Cloud ran an advertisement on Vegas Sphere during the AWS Re:Invent which is taking place in Las Vegas, Nevada, from November 27 to December 1, 2023. Google Cloud running a Vegas Sphere ad during AWS re:invent shows two things: 1. They are hyper aggressive on cloud 2. They badly need cloud to work pic.twitter.com\/GM0RNHbh8R— Aravind Srinivas (@AravSrinivas) November 27, 2023 Advertising on the surface of the Sphere comes at a remarkable price — $450,000 per day or $650,000 per week. In 2022, Alphabet, Google’s parent company, allocated a hefty budget of 4.37 billion U.S. dollars for advertising in the United States, marking a significant increase from the previous year’s expenditure of 3.6 billion dollar Google Cloud running ads on Vegas Sphere suggests two possibilities: either the cloud provider is actively seeking new customers, or its business is thriving. However, the latest earnings report leans towards the former, where Google lagged behind the other two major players in the cloud market. In the latest quarter, AWS recorded a 12% growth, reaching $21.3 billion, up from $20.53 billion in the corresponding quarter last year. Meanwhile, Google reported a 22% revenue growth, totaling $8.41 billion. Microsoft Azure, on the other hand, outpaced both AWS and Google Cloud, achieving the highest growth with a 29% increase in revenue. Looks like Google updated the copy for Google Meet in the App Store to capitalize on OpenAI events. Genius. pic.twitter.com\/V5aysiiqgd— Trung Phan (@TrungTPhan) November 22, 2023 Interestingly, this is not the first time Google has capitalised on an ongoing scenario for its ad campaign. A few days ago, when OpenAI’s chief scientist Illya Sutskevar fired Sam Altman on Google Meet, Google seized the opportunity and created an ad on the Google Play Store for Google Meet that humorously said, “Meet on any device and perfect for poorly planned board coups.”","excerpt":"Advertising on Sphere comes at a remarkable price of $450K per day.","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-11-28T14:30:45","publication_year":"2023","word_count":311,"keywords":["Go","API","cloud_platforms:Azure","AWS","OpenAI","AI","R","cloud_platforms:AWS","cloud_platforms:Google Cloud","Azure"],"extracted_tech_keywords":["AI","OpenAI","AWS","Azure","R","Go","API","cloud_platforms:AWS","cloud_platforms:Azure","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-cloud-runs-ad-on-sphere-during-aws-reinvent-at-las-vegas\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10054141,"title":"Trends That Defined Open Source This Year","content":"The last few years have seen the deployment of open-source software accelerate rapidly, and this will keep growing as we move forward. As per a report, the global open source services market size is expected to grow at a compound annual growth rate of 18.2 percent during the forecast period, to reach $50 billion by 2026 from $21.7 billion in 2021. Let’s take a look at the open-source trends that emerged this year. Demand for talent in open source technologies is high Though the pandemic saw a rapid acceleration towards cloud-based digital transformation by businesses worldwide, it also saw a huge gap in finding the right tech talent. The demand to find such talent for organisations was very high throughout the year. Interestingly, the companies are actively investing in open source technologies and are on the lookout for talent well versed with these techs. According to the 2021 Open Source Jobs Report by the Linux Foundation and edX released in September, 97 percent of hiring managers said that hiring open source talent is a priority; 50 percent of the respondents said that they are increasing open-source hiring this year. This demand for open source talent is fueled by organic growth within organisations, as reported by 50 percent of hiring managers. It also mentioned that 92 percent of hiring managers surveyed had reported difficulty finding open source talent. Holding the right certifications will ease the process for hiring managers to find suitable open source talent. Popular open source languages Javascript has always been a popular choice among developers due to the flexible structure and user-friendly approach, while Python has also become a hot favourite among developers due to efficiency, speed, easy to learn capabilities and diverse range of libraries to work on. As per the latest  State of the Octoverse report by Github, where it looks at code and communities built on GitHub, Javascript came out as the top language while Python and Java took second and third place, respectively. Typescript came in fourth, and C# occupied fifth place. Image: Github State of the Octoverse report As per the Sonatype State of the Software Supply Chain report 2021, as of July 31, 2021, there were 1,864,696 JavaScript packages available in the npm repository, a rise of 16% from last year. Big tech companies and open source Integrating open-source technologies and tools in businesses is not an option anymore but a necessity. Using them comes with a set of advantages that can be transformational for any business like cost-effectiveness, flexibility, enhanced security, lower software costs, no vendor lock-in, among others. Big tech companies have realised the potential that open source holds and are actively investing in it, open-sourcing datasets for scrutiny and improvement as well. A few of them are: Recently, Meta released a free, open-source library for training deep learning models with differential privacy called Opacus. Meta had also launched open-source simulation platform Habitat 2.0Salesforce recently open-sourced a machine learning model known as CodeT5, which can understand and generate code in real-time.Hugging Face, a leader in the open-source NLP space, introduced an open-source library called Optimum, an optimisation toolkit for transformers at scale.DeepMind made AlphaFold 2.0 source code public. It is an AI algorithm that predicts the shape of proteins that can have a deep impact in the life sciences space. Open Source in Mars mission: A Historic Moment Open source also builds a strong sense of community-driven by a passion for contributing towards technological advancements for the common good. This was seen when Github developers teamed up for a historic Mars mission. NASA’s Jet Propulsion Laboratory (JPL) flew a helicopter named “Ingenuity” on Mars. The Ingenuity helicopter runs an embedded Linux distribution on its navigation computer. A large part of the software is written in C++ using JPL’s open-source flight control framework F Prime (F’). In the coming years, we will see more and more open source deployments taking place, innovative projects coming up that will shape the businesses and innovations of tomorrow.","excerpt":"looking at the trends that defined the open source ecosystem in 2021","categories":["IT Services"],"tags":["code","GitHub","languages","Linux Foundation","Mars Mission","NASA","Open Source"],"author_name":"Sreejani Bhattacharyya","publish_date":"2021-11-24T17:00:00","publication_year":"2021","word_count":663,"keywords":["Hugging Face","NASA","Open Source","machine learning","AI","Mars Mission","Transformers","NLP","Python","languages","code","deep learning","differential privacy","JavaScript","Linux Foundation","GitHub","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","NLP","Hugging Face","Transformers","differential privacy","Python","R","JavaScript"],"url":"https:\/\/analyticsindiamag.com\/it-services\/trends-that-defined-open-source-this-year\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164623,"title":"Breaking Barriers in AI: Join the InnovateHER AI-ML Challenge","content":"MachineHack, in partnership with Chubb, invites talented women in artificial intelligence (AI) and machine learning (ML) to participate in the InnovateHER AI-ML Hiring Challenge. This exclusive competition provides a unique opportunity to showcase skills, tackle real-world challenges, and compete for exciting career opportunities at Chubb. Register Here About the Challenge: The InnovateHER AI-ML Hiring Challenge is structured into two key rounds: Python Coding Assessment (Starts on February 26, 2025): A 20-minute technical quiz that evaluates participants’ Python programming skills and AI-ML problem-solving abilities. The test covers Python fundamentals, data structures and algorithms, data manipulation using Pandas and NumPy, machine learning basics, natural language processing (NLP) and text analysis methodologies, model evaluation metrics, and probability and statistical analysis. AI-ML Ideathon (Starts on February 26, 2025): A hands-on challenge in which participants must develop a Gen-AI model capable of extracting, analysing, and summarising climate risk insights from real-world reports and news sources. The key requirements include selecting relevant datasets, identifying credible web sources for data scraping, developing an AI model for insight extraction and analysis, ensuring accuracy and relevance in reporting, and evaluating model performance. Register Here Why Participate? Showcase Your AI-ML Expertise: Solve a real-world problem using AI and machine learning. Compete With the Best: Engage with a community of talented women AI-ML engineers. Enhance Your Career: Gain recognition and visibility among top recruiters at Chubb. Win Exciting Rewards: Stand a chance to win prizes and career-defining opportunities. Who Can Participate? This challenge is open exclusively to women developers with expertise in AI\/ML, whether a student, early-career developer, or experienced professional. How to Participate? Click on the ‘Participate Now’ button. Complete the registration form and confirm your participation. Get ready to compete, innovate, and win! Register Today & Elevate Your Career! About Chubb Chubb is a global insurance leader operating in 54 countries. It offers a diverse array of insurance solutions and is known for its strong financials and global reach. Chubb is part of the S&P 500 and is committed to innovation and excellence in service delivery. Elevate Your Career with Chubb! Don’t miss this opportunity to shine in the AI-ML community and take your career to the next level with Chubb. Register now and be part of an exciting journey towards innovation and excellence in AI-ML!","excerpt":"MachineHack, in partnership with Chubb, invites talented women in artificial intelligence (AI) and machine learning (ML) to participate in the InnovateHER AI-ML Hiring Challenge. This exclusive competition provides a unique opportunity to showcase skills, tackle real-world challenges, and compete for exciting career opportunities at Chubb. Register Here About the Challenge: The InnovateHER AI-ML Hiring Challenge […]","categories":["AI Highlights"],"tags":["Chubb","Chubb Hackathon","hackathon for data scientists","Hackathon for hiring data scientists","hiring data scientists","hiring hackathon","Hiring Hackathon for Data Scientists","Machinehack","Machinehack Hackathon"],"author_name":"Siddharth Jindal","publish_date":"2025-02-26T10:13:36","publication_year":"2025","word_count":377,"keywords":["hiring data scientists","hiring hackathon","Ray","hackathon for data scientists","Chubb Hackathon","R","Pandas","Chubb","NumPy","artificial intelligence","NLP","machine learning","AI","ML","Machinehack Hackathon","Hiring Hackathon for Data Scientists","Machinehack","Python","Hackathon for hiring data scientists"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","Ray","Pandas","NumPy","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/breaking-barriers-in-ai-join-the-innovateher-ai-ml-challenge\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10041758,"title":"Open Banking Is Accelerating Digital Transformation: Swaminathan Srinivasan, Maveric Systems","content":"Management consultants Ranga Reddy, P Venkatesh, NN Subramanian and VN Mahesh founded Maveric System in 2000. The Chennai-based company helps banking and fintech leaders accelerate business transformation through effective integration of domain, technology, future-ready strategy and high-velocity execution. Today, the company has expanded to the US, the UK, Saudi Arabia, UAE, and APAC. In an exclusive interaction with Analytics India Magazine, Swaminathan Srinivasan, Senior Vice President of Data and Analytics Business Unit at Maveric Systems, spoke about their journey, challenges, and the fintech landscape. Excerpts: AIM: What’s Maveric Systems’ Raison D’être? Swaminathan Srinivasan: As financial institutions grapple with the challenges of decreasing revenues and increasing costs, the ability to maintain leadership and retain competitive advantage is the challenge. Added to this, the disruptions caused by digital-only banks and fintechs call for re-engineering business models, adopting newer technologies and platforms, and integrating them into existing business processes. Achieving these at speed is a game-changer for effective transformation. As a banking technology transformation specialist, we enable our customers to accelerate digital transformation by bringing the necessary domain, technology expertise, and high-velocity execution. We help banks with: Experience engineering coupled with next-gen digital solutions aimed for rapid digital transformation. Precise decisions which empanel domain-led data models, real-time data integration and three-layered accuracy assertion to offer conscious contextualisation.Connected core blended with product stack mastery in Temenos.Continuous quality supported by transformative quality engineering solutions engineered for accelerating the development of defect-free software. AIM: Tell us about the technologies Maveric uses. Swaminathan Srinivasan: Maveric uses artificial intelligence, machine learning, cloud, Big Data, and RPA for building and offering its core service value proposition. We partner with leading providers of such technologies like Leapworks, HVR Software, Talend, and Matillion. In several cases, our customer’s choice of technology platforms determines our focus areas as well. For instance, we are working with several banks using hybrid environments with a mix of Open Source and Cloud platforms like AWS, Azure, etc. We invest in developing these skill sets and accelerators in our labs and training academy. AIM: What are your flagship services? Swaminathan Srinivasan: Our flagship services include: Data and AnalyticsDigital TransformationCore banking (Temenos Suite of Services) Quality Engineering (QE) AIM: How does Maveric leverage AI and ML? Swaminathan Srinivasan: Maveric has been at the forefront of leveraging artificial intelligence and machine learning-led tech stacks across its service horizons and solutions in data, digital, core banking and quality engineering practices. Intelligent Quality Engineering (IQe) platform is a low-code domain-led solution for engineering quality at speed and scale. The cognitive powers embedded within this platform offers 95 percent QE automation to speed up quality cycle journeys, thus enabling rapid digital transformation. VoCAL is a next-generation ML and AI framework that deciphers and connects customer feedback for unearthing concealed improvement areas and deriving actionable insights. Maveric uses analytics derived from our recruitment and talent sourcing data to identify trends and patterns for better identification and targeting for hiring. AIM: How did Maveric deal with the pandemic? Swaminathan Srinivasan: We adopted a pronged strategy to scale up our digital infrastructure to deal with the pandemic: Strengthening the core: We made significant investments in scaling up the core infrastructure (network and server capacity) with robust security controls. We set up automated systems that enable monitoring, tracking, and backup of the server and network capacity by over 25 percent at any given point in time. Dual-layered security with 100 percent data privacy: A dual-layered security approach was institutionalised for data privacy. AIM: What’s the latest in banking and fintech space? Swaminathan Srinivasan: Open Banking is accelerating digital transformation across the global banking industry. Banks are focusing on delivering customer-centric solutions using new business models and open services. Micro services and API are driving agility along with innovative ways to collaborate. Analytics and artificial intelligence have a significant impact on the banking and financial industries as banks focus on hyper-personalisation at scale while transforming customer experiences. There is also steady progress in the adoption of the cloud in the banking sector. AIM: What’s your advice for traditional banks looking for a digital revamp? Swaminathan Srinivasan: Technology adoption has witnessed an unprecedented surge in our country over the last decade. With affordable smartphones and data, people have started using or considering alternative banking systems rather than the traditional bank. The pandemic has only accelerated the move towards digital banking and payments, forcing traditional banks to speed up digital transformation. Transformation involves integrating data, advanced analytics and digital technology into all areas of a bank, including how they function and deliver their services. It is more than just a technological upgrade; it requires a cultural change and disruptions in legacy processes. But this journey is bound to be gradual. In many cases, a significant impact can be achieved by understanding the customer journey through consumer lending and deploying these new technologies in those high impact areas. This is where data and analytics help immensely, and given that traditional banks have a wealth of data, this could be a big differentiator. Conventional banks should leverage their existing technology strengths while aiming for quick wins as they start the digital transformation journey. AIM: What’s next for Maveric? Swaminathan Srinivasan: As an IT Services organisation, we compete with several big players in all the markets. But Maveric has always carved a niche for itself through its unique customer-centricity approach, domain, technology focus and a differentiated talent pool that customers have come to appreciate. Our goal is to emphasise each of these and execute at scale to become real transformation partners to our customers in the years to come.","excerpt":"Management consultants Ranga Reddy, P Venkatesh, NN Subramanian and VN Mahesh founded Maveric System in 2000. The Chennai-based company helps banking and fintech leaders accelerate business transformation through effective integration of domain, technology, future-ready strategy and high-velocity execution. Today, the company has expanded to the US, the UK, Saudi Arabia, UAE, and APAC.  In an […]","categories":["AI Features"],"tags":["AI in banking","AI in finance","Interviews and Discussions"],"author_name":"Debolina Biswas","publish_date":"2021-06-15T11:00:00","publication_year":"2021","word_count":927,"keywords":["artificial intelligence","machine learning","AWS","AI","Azure","R","ML","AI in finance","RAG","Aim","analytics","AI in banking","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","AWS","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/open-banking-is-accelerating-digital-transformation-swaminathan-srinivasan-maveric-systems\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061489,"title":"The trainers, explainers, and sustainers of AI","content":"Artificial intelligence is widely regarded as a job-killer. While this is a point of contention, one thing is certain: an entirely new support structure of specialists will be necessary to design and guide AI endeavours. Paul Daugherty and H. James Wilson investigated the methods and approaches of 1,500 firms for their book, Human + Machine, and discovered new job categories forming. “These new jobs aren’t merely taking the place of old ones,” they discover. “They are completely new jobs that require skills and training that have never been required before. Advanced AI systems, in particular, are needing new business and technological jobs to train, explain, and sustain AI behaviour. These new roles are symbiotic with AI and draw on distinctly human skills.” Daugherty and Wilson define three distinct new job paths. Trainers Human workers will be required to train AI systems how to operate in this initial category of new jobs, which is fast forming. Trainers help natural-language processors, and language translators make fewer mistakes on one end of the range. They train AI computers how to imitate human behaviour, on the other hand. Chatbots for customer support, for example, must be taught to recognise the complexity and subtleties of human speech. Yahoo Inc. is teaching its language processing system that individuals don’t always mean what they say literally. Yahoo developers have so far built an algorithm that can detect sarcasm with an accuracy of at least 80% on social media and websites. Consider the position of “empathy trainer” – people who teach AI systems to be compassionate. Kemoko Inc., d\/b\/a Koko, a New York-based firm spun out of the MIT Media Lab, has developed a machine-learning system that can help digital assistants like Apple’s Siri and Amazon’s Alexa address problems. Without an empathy trainer, Alexa may respond to a user’s concerns with prepared, repeating responses like “I’m sorry to hear that” or “Talking to a friend can sometimes help.” Alexa can be a lot more helpful if given the correct training. Explainers Explainers, the second type of new employment, will bridge the gap between technologists and business leaders. Explainers will aid in providing clarity, which is becoming increasingly vital as AI systems become more opaque. Many CEOs are wary of powerful machine-learning algorithms’ “black box” character, especially when the systems they power urge actions that go against conventional thinking. Governments have been debating rules in this area for some time. Employees that can explain the inner workings of sophisticated algorithms to nontechnical professionals will be required by companies that implement powerful AI systems. Algorithm forensics experts, for example, would be in charge of holding any algorithm accountable for its outcomes. When a system makes a mistake or makes decisions that have unexpected negative implications, the forensics analyst is expected to perform an “autopsy” on the event in order to figure out what caused the behaviour and how to fix it. Certain forms of algorithms, such as decision trees, are quite simple to comprehend. Others, such as machine-learning bots, are more difficult to understand. However, in order to perform extensive autopsies and explain their findings, the forensics analyst must have the necessary training and expertise. Sustainers These people will be in charge of ensuring that AI is used correctly. Sustainers continue to work to guarantee that AI systems are functioning properly as tools that exist solely to serve us, assisting people in their work and making their lives easier. Setting restrictions or overriding judgments based on economics, legal or ethical compliance are activities that a sustainer will undertake. Sustainers will also be in charge of applying critical thinking to AI performance and building interfaces for AI-enabled workforces. ‘Ethics compliance managers,’ for example, may be hired to serve as watchdogs and ombudsmen for safeguarding commonly accepted norms of human values and morality. According to the World Economic Forum in its report “Future of Jobs Report 2020,” an estimated 85 million jobs will be displaced while 97 million new jobs will be created across 26 countries by 2025. Explainers, trainers, and sustainers will play an interesting role in the new AI created workforce. Read More: Why is Google Hiring AI Workforce in Baidu’s & Alibaba’s Backyard","excerpt":"Explainers, the second type of new employment, will bridge the gap between technologists and business leaders.","categories":["IT Services"],"tags":["AI Workforce"],"author_name":"Abhishree Choudhary","publish_date":"2022-02-25T10:00:00","publication_year":"2022","word_count":694,"keywords":["Go","artificial intelligence","programming_languages:R","AI","chatbots","programming_languages:Go","Git","AI Workforce","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","chatbots","R","Go","Git","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-trainers-explainers-and-sustainers-of-ai\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10001786,"title":"What Is Container Cloud Analytics And How Does It Compare To VMs","content":"Container technology has gained plenty of recognition and usage in the past couple of months over traditional methods. They provide a seamless deployment of software solution because of the number of advantages that they come with. What Is A Container? Containers provide a standard way to package the application’s code, configurations, and dependencies into a single object. They share an operating system installed on the server and run as resource-isolated processes. It basically consists of an entire runtime environment, including: An application Libraries Other binaries Configuration files needed to run it By containerising the application platform and its dependencies, differences in OS distributions and underlying infrastructure are abstracted away. Containers provide data in a way that helps the developers to quickly reproduce an experiment at any stage in the research process, which saves valuable time and are a solution to the problem of how to get the software to run reliably when moved from one computing environment to another. Container technology makes it easier to access data from nearly any source. They help to simplify the deployment process at large. They are capable of abstracting the applications from an environment in which they actually run, making it deploy easily and consistently. They can, therefore, be deployed easily, irrespective of the nature of the data platform on which it is stored. Advantages Containers are far more lightweight than the traditional ways. Containers share the OS kernel, start much faster, and use a fraction of the memory compared to booting an entire OS. They do not have to virtualise the hardware stack. It also has reduced size of snapshots and has a much quicker spinning up of applications. They also have reduced and simplified security updates Another advantage of containers is that they let the user launch multiple instances to scale up the load-serving capacity. It has an orchestration tool that lets you have declarative deployments and hence the user does not have to worry about launching multiple such instances to scale up the load-serving capacity. Containers have less code to write in order to transfer, migrate and upload workloads. They provide an efficient transfer in a minimalistic code. They also reduce management overhead. Because they share a common operating system, only a single operating system needs care and feeding for bug fixes, patches, and so on. This concept is similar to what we experience with hypervisor hosts: fewer management points but slightly higher fault domain. In short, containers are lighter weight and more portable than VMs. The Container Battle With VMs Like virtual machines, containers allow you to package your application together with libraries and other dependencies, providing isolated environments for running your software services. But container technology offers a variety of advantages over the traditional virtual machines. 1.Cloud availability: There are many cloud providers today that support containers. It’s the biggest advantage comes with the fact that it uses the cloud. 2.Portability of applications: Containers allow the applications to be easily ported. This makes easy handling of the applications. Containers are very lightweight and can be hence easy to migrate with them. In virtual machines, this feature or portability of applications is not available. 3.Cost effective: Containers are very cost effective compared to virtual machines. The user pays only for what he uses. Virtual machines, on the other hand, make all these systems on a large machine and make it entirely utilised. Conclusion Containers definitely provide a large set of advantages over the traditional methods. More and more developers are moving towards adopting the technology and many are also convinced by the kind of efficiency that it provides with handling applications. Containers are definitely a win over the older options.","excerpt":"Container technology has gained plenty of recognition and usage in the past couple of months over traditional methods. They provide a seamless deployment of software solution because of the number of advantages that they come with. What Is A Container? Containers provide a standard way to package the application’s code, configurations, and dependencies into […]","categories":["AI Features"],"tags":["AI What it Does","containers","cost","Developers"],"author_name":"Disha Misal","publish_date":"2019-04-10T13:23:49","publication_year":"2019","word_count":610,"keywords":["programming_languages:R","AI","containers","ML","cost","AI What it Does","R","Developers"],"extracted_tech_keywords":["AI","ML","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-is-container-cloud-analytics-and-how-does-it-compare-to-vms\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10046365,"title":"How IIT Guwahati Leads Cutting-Edge Research In AI &#038; Data Science","content":"The Indian Institute of Technology, Guwahati (IITG), established in 1994, has a vision: “To be recognised globally for excellence in education, research and innovation, and nurture future leaders, to serve the society at large.” Analytics India Magazine got in touch with three IITG professors, namely Ratnajit Bhattacharjee, Siddhartha Pratim Chakrabarty, and Ashok Singh Sairam, to understand IITG’s contributions to the burgeoning field of AI and Data Science in India. Prof. Ratnajit is from the Electronics and Electrical Engineering Department, IIT Guwahati and Head, Mehta Family School of Data Science and Artificial Intelligence.Prof. Siddhartha and Ashok are from the Department of Mathematics, IIT Guwahati and the Associated Faculty, Mehta Family School of Data Science and Artificial Intelligence. AIM: What’s IIT Guwahati’s approach towards AI education and research? Professors: Keeping in pace with the evolving nature of Artificial Intelligence (AI) and Data Science (DS), and the potential of applicability of these subjects in various sectors, the Indian Institute of Technology Guwahati (IITG) has initiated several steps to strengthen its contribution: Three departments of IITG, namely, departments of Computer Science and Engineering (CSE), Electronics and Electrical Engineering (EEE) and Mathematics, came together to launch the first interdisciplinary M. Tech Programme in Data Science. The program, running successfully since 2019, allows students from different branches of engineering and technology with a valid GATE score in different disciplines. The first batch of graduates got good placement in the industry. Subsequently, IIT Guwahati established the Mehta Family School of Data Science and Artificial Intelligence (MFSDS&AI) with generous support from the USA’s Mehta Family Foundation (MFF). The MoU signing event between IITG and MFF was graced by the Principal Scientific Advisor (PSA), Government of India (GoI) and Secretary, Science and Engineering Board (SERB). The school is soon going to admit its first batch of B. Tech students from 2021 through JEE (Advanced). The school is in the process of recruiting faculty members having excellent academic and research backgrounds and placing the required infrastructure in place to make this program highly successful.  The MFF, in addition to providing infrastructure support for the School, has also connected the school to eminent Professors from leading American universities as strategic advisors to the school and will help in establishing an association with American universities for carrying out joint research activities. The school is targeting enrolment of research scholars for PhD programs from December 2021 and plans to introduce postgraduate programs in the near future. AIM: What all research and projects is IIT Guwahati working on in the field of artificial intelligence? Professors: IITG faculty members are currently engaged in research activities in different topics in the field of AI and Machine Learning (ML), particularly their applications. Some of the areas include: Natural Language Processing (NLP)Clinical Text MiningMachine Learning and its application in Computational BiologyClimate and Sustainable FinanceDeep Learning Open Source Intelligence (Social Media\/Social Network Analysis)Information RetrievalComputer VisionPattern Recognition, etc. Recently, IITG, in collaboration with several other leading institutes and organisations, has submitted a proposal for a Centre of Excellence (CoE) on the topic of Deep Learning for geo-hazards prediction to the Science and Engineering Research Board (SERB) for possible funding. If this CoE gets established, there will be cutting edge research activities in the application of Deep Learning and the emerging Internet of Things (IoT) technologies. The IITG team has also participated in a brainstorming session with the Technology Information Forecasting and Assessment Council (TIFAC) on the topic of Advanced technologies in Agriculture. AIM: Tell us about the Technology Incubation Centre (TIC) at IIT Guwahati. Professors: IITG-Technology Incubation Centre (IITG-TIC) is registered as a Society under the Registration of Societies Act XXI of 1860. The goal is to encourage entrepreneurial initiatives amongst the faculty and alumni of the IITG community in particular and other State or Central Government Technical Institutions of the North East. This centre promotes interdisciplinary research with special emphasis on the development and innovation of high-growth and knowledge-based-business and nurtures indigenous products with innovative hardware\/embedded designs. The centre provides technical support, a soft loan facility, and business mentoring, subject to availability. Currently, the Department of Information Technology, Govt. of India provides funds for developing basic infrastructure and soft loan facility. IITG-TIC is also one of the Business Incubators approved and recognised by the Ministry of Micro, Small and Medium Enterprises (MSME), Govt. of India. Find further details here. AIM: Tell us about the challenges IITG has faced while working in the field of AI and Data Science. Professors: AI and DS have interdisciplinary applications. It requires bringing together the domain experts and AI\/DS experts on the same platform in order to solve any challenging problems. In addition, it requires high-end computational facilities in place, vibrant research culture, and industry collaboration. Having all these ingredients in one place and blending of the same to produce some tangible outcomes, impacting the economy and society at large, is a challenging problem. IITG plays a leadership role in overcoming these challenges. We aim at taking up projects and problems relevant to national importance and national growth, such as conserving and effective usage of potable water, enhanced healthcare services, geo-hazard predictions, the agricultural sector, etc., to name a few. IITG desires to join hands with the nation’s vision in having an economic impact of AI by offering the most relevant programs comprising of high skill curricula having depth and breadth: blended with multidisciplinary flavour. IITG is also putting in the effort to develop a collaborative culture with academia and industry. AIM: What role can the government play or has played so far in promoting AI and other cutting edge domains in India? Professors: The government of India has always been very supportive in promoting cutting edge technologies to solve real-life problems that the country faces today. Niti Aayog has already laid down the National Strategy for Artificial Intelligence, highlighting the potential of AI and laying down recommendations to accelerate its adoption. The National Education Policy (NEP) 2020 stresses inter-disciplinary programmes, and the MFSDS&AI is a definite step in that direction. Intensification of Research in High Priority Areas (IRHPA), a legacy scheme of SERB, has contributed significantly to augment general R&D capabilities by setting up Core Groups, the Centre of Excellence (CoEs), and National Facilities in frontline and emerging fields of science and engineering. Recently, SERB issued a call for establishing Centres on AI-based Earth Systems Modeling, to which IITG has already submitted a proposal. MeitY, DBT and other ministries are promoting AI and its applications in different sectors. These ministries support R&D activities and pilot projects to promote cutting edge technologies. AIM: How can non-IITians tap into the resources of IIT Guwahati? Can you discuss the process\/resources etc., briefly? Professors: IITG has always been supportive of the development of the region as a whole and has signed several MoUs with the different institutes of this region to promote academic activities and cutting-edge research. Institutes having MoUs with IITG get access to resources available with IITG. Further, we regularly conduct workshops, conferences, Faculty Development Programs (FDP), training programs, and internship programs, through which participants get access to some of our resources. This involves the use of sophisticated equipment, computational facilities, library facilities, etc., to name a few. Our expert faculty members provide invaluable guidance and technical inputs to participants. In addition, the Central Instrumentation Facility and the Data Centre at IITG may be accessed (following laid down procedures) by academicians and researchers not affiliated to IITG. AIM: How do you rank India when it comes to R&D efforts in AI? What do we do well, and where does it fall short? And, how is IIT Guwahati addressing this? Professors: As per a recent news report, India ranked third in the world in terms of high-quality research publications in artificial intelligence, as per a report by Itihaasa – a research agency. In yet another report by Stanford University, Artificial Intelligence Index Report 2021, India was ranked 6th in The Global Vibrancy Ranking 2020. All these reports indicate the high-quality research activities in the field of AI that is being done in India. However, the scopes of improvements are there in several domains, particularly in solving challenging problems in healthcare, infrastructure, transportation, supply of potable water, early warning system and agriculture. IITG has taken a proactive role in developing a high skill curriculum, emphasising computational and algorithmic background building and blending it with multidisciplinary flavour. The MFSDS&AI endeavours to enhance the skill set of the students by focusing on the industry and current interdisciplinary research needs. It is creating an environment of vibrant research culture with experts from diverse fields, aiming to publish quality research works in high impact factor journals and apply DS and AI in solving real-life problems.","excerpt":"IITG has taken a proactive role in developing a high skill curriculum, emphasising computational and algorithmic background building and blending it with multidisciplinary flavour.","categories":["AI Features"],"tags":["AI in India","Data Science","data science in India","IIT Guwahati","Interviews and Discussions","open source data science projects","research in India","what is data science"],"author_name":"kumar Gandharv","publish_date":"2021-08-20T10:00:00","publication_year":"2021","word_count":1444,"keywords":["what is data science","data science","open source data science projects","AI in India","research in India","AI","artificial intelligence","machine learning","ML","computer vision","data science in India","NLP","Aim","deep learning","IIT Guwahati","analytics","Data Science","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","computer vision","data science","analytics","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-iit-guwahati-leads-cutting-edge-research-in-ai-data-science\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10011490,"title":"What Is The Hiring Process For Data Scientists At ZS","content":"With a focus on building scalable capabilities for generating, operationalising, and measuring data-driven insights for their clients, ZS has a strong advanced data science group within their Business Consulting function. The team focuses on integrating transformative AI-enabled solutions and data products across multiple industries such as healthcare, life sciences, telecommunication, high tech, and retail. As the company leverages deep industry expertise and leading-edge analytics to create solutions that work in real life, the data science team plays a vital role in driving these functions. What Do ZS Look For In A For Data Scientist? From foundational research in deep learning, natural language processing (NLP), optimisation, and operational research — the advanced data science team at ZS works across various solutions. They are involved in developing solutions to full-scale productisation. In other words, the team works on completing the entire life cycle from innovation to industrialisation. To fit into these roles, ZS look for AI technocrats who can effectively lead a hybrid team of scientists and engineers. This requires mastery in either advanced data science or engineering and working proficiency. In fact, the company is building an Artificial Intelligence Center of Excellence (AI CoE), which focuses on conducting applied AI research. To accommodate data scientists into the AI CoE, they are looking for research scientists who have a strong educational background in concepts such as NLP, computer vision, reinforcement learning, deep learning etc. The company believes that the CoE will serve as a vehicle for developing frontier innovations for the clients. While a strong educational background is something that they look forward to, ZS also focuses on the importance of an internship or work experience. “The candidate must share the evidence of application of the skills learnt through internship or work experience,” the company said. Interview Process The interview process comprises a series of exciting conversations around candidates’ areas of expertise and their depth of knowledge. The process begins with a round of telephonic discussions to understand the candidate profile and align it to ZS’s role. The next steps are panel presentation, research paper discussion, unstructured case discussion, or hands-on-case round depending upon the candidate’s expertise. One may visit the careers page on ZS’s website to look for relevant career opportunities. Growth Opportunities For Data Scientists ZS culture celebrates innovation and curiosity and is on a lookout for creative thinkers and solvers from a variety of backgrounds that help in developing and delivering products that drive better business outcomes. To, therefore, keep the data scientists updated and upskilled, ZS offers sponsorship programs to support data scientists pursue higher education such as Masters at ivy league institutes across the globe. In an earlier interaction with AIM, ZS had said that it also provides various internal platforms for the team to strive for grassroots innovations. It keeps data scientists challenged with initiatives such as participation in hackathons, participation in conferences and summits, potential to publish whitepapers and blogs in external publications, among others. The fact that team members are encouraged to evolve in their current roles by experimenting with new tools, technologies and algorithms, keep the data scientists driven and motivated towards their work. ZS is a Great Place to Work For certified firm and has more than 35 years of experience and 7,000-plus ZSers in more than 28 offices worldwide.","excerpt":"With a focus on building scalable capabilities for generating, operationalising, and measuring data-driven insights for their clients, ZS has a strong advanced data science group within their Business Consulting function. The team focuses on integrating transformative AI-enabled solutions and data products across multiple industries such as healthcare, life sciences, telecommunication, high tech, and retail. As […]","categories":["AI Hirings"],"tags":["back office data","data analytics masters","Data Science Hiring","Data Scientist","master data science"],"author_name":"Srishti Deoras","publish_date":"2020-11-11T13:00:23","publication_year":"2020","word_count":548,"keywords":["data science","data analytics masters","artificial intelligence","AI","computer vision","RAG","NLP","Data Science Hiring","Aim","deep learning","master data science","analytics","back office data","Data Scientist","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","NLP","computer vision","data science","analytics","Aim","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/what-is-the-hiring-process-for-data-scientists-at-zs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10072128,"title":"Why IISc wins?","content":"Indian Institute of Science (IISc) or ‘Tata Institute’, as Bangaloreans lovingly call it, has done it again. The institute won the top spot in the Research Universities category in the recently released NIRF Rankings 2022. In the Universities category too, it bagged the 1st position for the 7th year in a row. In the overall category, it stands at the 2nd position, next only to IIT Madras. The NIRF ranking framework evaluates institutions on five broad parameters – Teaching, Learning and Resources (TLR), Research and Professional Practice (RP),  Graduation Outcome (GO), Outreach and Inclusivity (OI) and Perception (PR). IISc was selected as the world’s top research university in the QS World University Rankings 2022, leaving behind some of the top Ivy League colleges. With two Bharat Ratnas (Nobel Laureate C V Raman and celebrated Professor C.N.R Rao), 36 Padma Awards, 7 Infosys Prizes, 97 Shanti Swarup Bhatnagar prizes and unaccountable accolades, IISc is indeed the pride of the nation. Congratulations to @iiscbangalore, @iitbombay and @iitdelhi. Efforts are underway to ensure more universities and institutions of India scale global excellence and support intellectual prowess among the youth. https:\/\/t.co\/NHnQ8EvN28— Narendra Modi (@narendramodi) June 9, 2021 Image: NIRF rankings 2022 (University Rank 1) Image: NIRF Rankings 2022-Overall Rank 2 Image: NIRF Rankings 2022-Research Rank 1 Jamshedji Tata set up IISc in 1909 with the vision to improve the scientific capabilities of the country. Starting with just two departments – General and Applied Chemistry, and Electrical Technology – the institute has grown manifolds to 537 faculty members. Its popularity has grown by leaps and bounds globally. Tenure system for faculty The faculty in IISc is top-notch. Just like most universities in the US, IISc has a ‘tenure’ system. The faculty-selection process is a rigorous one. A faculty search committee deeply scrutinizes each application based on the academic credentials of the person. If the response is positive, the applicant gets invited for an interactive session with the faculty members. They are then required to deliver a ‘test’ lecture. After this, the department shortlists candidates. The director appoints a selection committee with both internal and external experts. Based on recommendations of the selection committee and approval of the council, the candidate is appointed as an assistant professor at IISc. They are hired on an initial contract period of 5 years. A one-year probation period runs concurrently, IISc informs. If an assistant professor does not get tenure by the end of the 5-year contract period, the institute can extend the contract by another five years, during which the faculty member must secure tenure. If not, they have to leave. Such rigour ensures great results. In the QS University Rankings 2022, one of the parameters for ranking universities was citations per faculty (CPF) – an indicator when universities are adjusted for faculty size. IISc achieved a perfect score of 100\/100. Stringent admission procedure As one of the country’s most-coveted institutes, the entry barrier for IISc is very high. To maintain the reputation it has built over the years, it becomes imperative for IISc to only accept students who are deeply passionate and dedicated to their field of research and want to make an impact on the world through science. Most of the bachelor’s courses at IISc need the aspirants to appear for the JEE examination. However, bachelor’s courses, which can lead to an integrated PhD, can be pursued after clearing the JAM examination. For the highly sought-after master’s courses, students must qualify the GATE examination. For some courses, interviews are a part of the selection process. Needless to say, one has to excel in GATE to enter the holy grail of IISc. At the forefront of cutting-edge research As highlighted by the institute’s mission statement, IISc’s main area of focus is research – be it in any department. Its goal has been to “conduct high-impact research, generate new knowledge, and disseminate this knowledge through publications in top journals and conferences”. Over the years, researchers from various departments of the institute have conducted cutting-edge research that has massively accelerated the pace of innovation in those domains. When the world was battling the deadly COVID-19 pandemic, IISc was at the forefront leading innovations to tackle the severe impact of the pandemic through projects on diagnostics and surveillance, modelling and simulation, hospital assistance devices and vaccine development. The government too depends on leading faculty from IISc in its scientific decision-making. Recently, it appointed eminent physicist Dr Ajay Kumar Sood as the principal scientific advisor (PSA) to the Indian government.","excerpt":"IISc was selected as the world’s top research university, trumping some of the top Ivy League colleges in the QS World University Rankings 2022","categories":["AI Trends"],"tags":[],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-08-03T18:00:00","publication_year":"2022","word_count":748,"keywords":["Go","programming_languages:R","AI","innovation","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-iisc-wins\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":15969,"title":"Top 4 Famous AI Quotes by Jack Ma","content":"Jack Ma, the Founder of Alibaba Alibaba Group’s founder Jack Ma known for his impressive business acumen and an eye for disruptive technologies, plans to turn Alibaba into a massive organization across the world, and so far the progress looks really good. The company currently has about 5 percent of its business outside China, but Ma expects it to reach 40 percent within a decade’s time. Alibaba has invested in areas such as cloud computing and artificial intelligence, as it expands into new sectors beyond the e-commerce business. According to Chinese business tycoon, this will only be possible with the adoption of disruptive technologies that are changing businesses today. Artificial Intelligence is one such technology Ma has set his eyes upon. He believes in the impact that AI can create. Ma had previously made an announcement that the future of Alibaba would include AI and robots. Ma stresses “Alibaba is more of a data company, and not an e-commerce firm.” However, Ma is also aware of the concerns that the technology brings along, for say, job displacements. Ma not only evangelizes people about the scope that AI creates, but also talks about how to address the challenges associated with it. AIM lists down top 4 prophetic statements made by Jack Ma around AI 1. CEOs could be robots in 30 years – at an entrepreneurship conference in central China Jack Ma voices his concerns surrounding AI at the Gateway 2017 conference Ma also mentioned that robots being future CEOs of organizations shouldn’t come much of a surprise, considering the speed with which the technology is moving forward. He also adds that robots won’t be overcome by emotions, unlike humans, thus making them more efficient. “30 years later, the Time Magazine cover for the best CEO of the year very likely will be a robot. It remembers better than you, it counts faster than you, and it won’t be angry with competitors,” comments Ma. Essentially, Ma spoke about the rise of robots and AI, and how they might replace human workers. However, he also mentions that AI will be an absolute necessity to process the large amounts of data that’s generated today, which is usually beyond human capacity, or might be really time-consuming for a human. Link: https:\/\/www.youtube.com\/watch?v=ZmHnsS-p85Q 2. Artificial Intelligence can set off WW III – at Gateway 2017 conference Alibaba office The first technology revolution led to World War I, and the second led to the World War II. According to Ma, we’re living in the era of third technological revolution. Today, workers and employers are increasingly defined by data. Could this mean the onslaught of the third World War? It might actually happen, AI could set off the third World War, one that will see human on one side, and robots on the other. However, Alibaba’s founder says humans will win this battle, as machines don’t really possess the wisdom and experience reflected by a human. Ma also stresses on the fact that machines must be employed to do tasks, which are usually difficult for a human to execute. Ma says, “We know the machine is powerful and stronger than us, however, humans will rise above the impending wave of data and artificial intelligence.” Link: https:\/\/www.youtube.com\/watch?v=awl_cLKzta4 3. People will work for 4 hours a day – says Jack Ma again at the Gateway 2017 conference Jack Ma with President Donald Trump The founder of Alibaba, jack Ma believes that AI will make human lives much easier in the future. Ma thinks that people will only work four hours a day and maybe four days a week, within the next 30 years. “My grandfather worked 16 hours a day in the farmland and [thought he was] very busy. We work eight hours, five days a week and think we are very busy,” comments Ma. This prediction corresponds to the one made by famed economist John Maynard Keynes. The economist had said that the working week would be about 15 hours by 2030. Automation will be the primary reason behind this shift. According to Ma, the rich and the poor, bosses and their employees would be increasingly affected by automation. Thus, government must show willingness to make more “hard choices.” Link: https:\/\/www.youtube.com\/watch?v=DIhudEzOU0I 4. AI Will Cause More Pain than Happiness – at the China Entrepreneur Club event Australian Prime Minister Malcolm Turnbull Visits Jack Ma In Hangzhou The disruption caused by the internet and new technologies to different areas of economy could mean decades of pain for the society as a whole. Artificial Intelligence will most likely replace professions like truckers, secretaries, cashiers, bank tellers, waiters, and real estate agents, in the near future. “In the coming 30 years, the world’s pain will be much more than happiness, because there are many more problems that we have come across,” comments Ma. The Alibaba founder warned that social conflicts could have a “huge impact” on all walks of life. Ma’s company has invested in areas such as cloud computing and artificial intelligence as it expands into new sectors beyond its e-commerce business. Automation and the internet would lead to job disruptions. Moreover, the rise of AI and longer life expectancy will only lead to an aging workforce fighting over few jobs.will bring about. Ma spoke about his concerns surrounding the future of AI to encourage businesses to adapt or face problems in the future. According to Ma, machine shouldn’t replace human at what they do best, it must rather focus on making machines do what humans can’t. This approach would help in trusting machines as a “human partner,” than an opponent. Otherwise, the rise of AI could lead to social conflicts. Link: https:\/\/www.youtube.com\/watch?v=N9_xGhM8siI","excerpt":"Alibaba Group’s founder Jack Ma known for his impressive business acumen and an eye for disruptive technologies, plans to turn Alibaba into a massive organization across the world, and so far the progress looks really good. The company currently has about 5 percent of its business outside China, but Ma expects it to reach 40 […]","categories":["AI Trends"],"tags":["Artificial Intelligence India","cloud computing India","e-Commerce India","internet","robots India"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-06-30T04:01:20","publication_year":"2017","word_count":942,"keywords":["Go","artificial intelligence","cloud computing India","AI","cloud computing","e-Commerce India","ML","Artificial Intelligence India","R","RAG","Aim","Rust","GAN","robots India","internet"],"extracted_tech_keywords":["AI","artificial intelligence","ML","Aim","RAG","cloud computing","R","Go","Rust","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-4-quotes-on-artificial-intelligence-from-alibaba-billionaire-jack-ma\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10163093,"title":"Okta Appoints Shakeel Khan as Regional VP and Country Manager for India","content":"Okta, the US based identity and access management company, has named Shakeel Khan as its regional vice president and country manager for India, reinforcing its commitment to securing digital identities and expanding its presence in the region. With 27 years of experience in cybersecurity, Khan will focus on strengthening Okta’s position as a trusted identity security partner for Indian businesses. Based in Bangalore, he will oversee the company’s sales operations in India. His career spans pre-sales to enterprise sales leadership, giving him deep expertise across the cybersecurity ecosystem. Ben Goodman, SVP & GM, Okta APJ, welcomed Khan’s appointment, stating, “We are delighted to have Shakeel Khan join Okta India at a stage where businesses are looking for robust and trusted identity solutions more than ever. Our commitment to India remains strong, with more than 19,000 customers worldwide leveraging Okta. Shakeel’s technical leadership and strategic vision will help us reach the next level of growth.” Khan expressed enthusiasm about his new role, highlighting the evolving cybersecurity landscape. “I look forward to contributing to Okta’s growth and success while working with some of the brightest minds in the industry,” he said. Since its India launch in 2023, Okta has seen rapid expansion, tripling its workforce to over 300 employees in just a year. The company plans to surpass 500 employees by 2025 as part of its broader growth strategy. India is a key market for Okta, with the country’s developer base expected to overtake the US by 2027 and over 75% of IT spending flowing through partner channels. The company follows a partner-first approach, working with key players like Savex, ACPL, 22by7, and Valuepoint to drive further growth.","excerpt":"With 27 years of experience in cybersecurity, Khan will focus on strengthening Okta’s position as a trusted identity security partner for Indian businesses.","categories":["AI News"],"tags":["Okta"],"author_name":"Mohit Pandey","publish_date":"2025-02-10T10:11:14","publication_year":"2025","word_count":276,"keywords":["Go","API","programming_languages:R","AI","RPA","Okta","programming_languages:Go","Git","RAG","Rust","R"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","Git","API","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/okta-appoints-shakeel-khan-as-regional-vp-and-country-manager-for-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045260,"title":"India Meteorological Department To Use Machine Learning: Why Is It A Significant Step?","content":"The India Meteorological Department is all set to embrace artificial intelligence and machine learning to enable accurate weather forecasting. The department expects to fully leverage AI\/ML in two to three years. IMD has joined hands with IIITs at Prayagraj and Vadodara and IIT Kharagpur for this initiative. Further, the organisation has partnered with Google to make precise short-term and long-term forecasts. IMD has set up internal sub-groups of senior officials and meteorologists to determine the best AI and machine learning techniques for weather forecasting. The Ministry of Earth Sciences is inviting applications from independent researchers to assist the team in this endeavour. To that end, IMD has also set up a group of sister organisations, including the Pune-based Indian Institute of Tropical Meteorology. Last year, IMD director general Mrutyunjay Mohapatra said the department would issue ‘nowcasts’ for the real-time prediction of drastic weather events. The nowcasts give 3-6 hours heads up on extreme changes in weather. Traditionally, IMD uses radars and satellite imagery to issue such forecasts. However, it’s a slow process. IMD had also invited various research groups to study AI integration for improving weather forecasting. In 2019, the Ministry of Earth Sciences announced initiatives for improving weather forecasting using AI. The ministry is working on augmenting the existing supercomputing facility to 100 petaflops (PF) to improve understanding of weather and climate. The same year, the Ministry of Agriculture and Farmers Welfare signed an agreement with IBM India for generating AI and weather technology-based solutions at the farm level for weather forecasting and soil moisture information to help farmers make decisions on crop management. AI for weather forecasting Around the globe, organisations depend heavily on satellite imagery and radar inputs to obtain weather and climate data to make predictions. The data is too vast and overwhelming to analyse and scan for patterns by humans or even traditional computers. AI systems, machine learning, neural networks, and deep learning help pattern recognition that work with weather and climate datasets. According to a paper by American Meteorological Society (AMS), AI can drastically improve weather forecasting by scanning swathes of data in rapid time. AI could predict the effects of extreme weather and help in energy consumption, the paper stated. Weather departments around the world have adopted comprehensive and smart monitoring systems. In 2020, the US-based National Oceanic and Atmospheric Administration (NOAA) said it would work on four areas–NOAA Unmanned Systems, artificial intelligence, Omics, and the cloud–to bring advancements in quality and timeliness of NOAA science, products, and services. The Satellite and Information Service department of NOAA signed an agreement with Google to harness the power of AI and ML to facilitate the use of satellite and environmental data. Recent efforts with AI Last year, the University of Washington and Microsoft Research collaborated to demonstrate the power of AI to analyse past weather patterns and predict future events. The process turned out to be more efficient and accurate than the current technology. The team developed a global weather model to make predictions using the last 40 years of weather data. Lead author of the study Jonathan Weyn said, “Machine learning is essentially doing a glorified version of pattern recognition. It sees a typical pattern, recognises how it usually evolves and decides what to do based on the examples it has seen in the past 40 years of data.” In January 2020, Google presented MetNet, a neural weather model for precipitation forecasting. The deep learning network can predict future precipitation at 1 km resolution in over 2-minute intervals at timescales of up to 8 hours into the future. Initial research showed the model could predict the weather for the entire US in a few seconds as opposed to an hour taken by traditional tools. Two years back, IBM acquired The Weather Company and developed Deep Thunder to provide hyper-local weather forecasts within a 0.2 to 12 miles resolution.","excerpt":"In 2019, the Ministry of Earth Sciences announced initiatives for improving weather forecasting using AI.","categories":["AI Features"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-08-05T15:00:00","publication_year":"2021","word_count":644,"keywords":["Go","machine learning","artificial intelligence","AI","neural network","ML","RAG","Ray","deep learning","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","Ray","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/india-meteorological-department-to-use-machine-learning-why-is-it-a-significant-step\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10099511,"title":"7 Best Toolkits to Build Your AI Ethically in 2024","content":"Developing and deploying AI responsibly requires tools to help the entire development team understand what they’re doing and show them how their choices affect the end users. Researchers, analysts, and policymakers today are in dire need of minimising, if not avoiding the harm AI model can cause. Toolkits play a fundamental role in creating systems fairer, more robust and transparent. Here are seven toolkits that can assist in implementing AI ethically. NASSCOM Responsible AI Resource Kit In 2022, the National Association of Software and Services Companies (NASSCOM) collaborated with industry leaders including Microsoft, Tata Consultancy Services and IBM Research to introduce the Responsible AI Hub and resource kit. As the name suggests, the objective behind this initiative is to ensure responsible integration of AI technology. NASSCOM will maintain this evolving reference, continually incorporating the latest research and industry best practices, drawing from several credible sources. This kit equips businesses with the tools and guidance for AI development and deployment while complying with standards ethically. Moreover, the kit offers insights for identifying and mitigating ethical risks that may arise while implementing AI-powered solutions. AI and data protection risk toolkit The Information Commissioner’s Office of UK launched an AI and data protection toolkit last year as part of the effort to spread best practices in the use of AI. The toolkit is available as an Excel file on its website to download and edit to help organisations through several stages. During the release, senior policy officer Alister Pearson said it has been developed to help organisations comply with data protection regulations and win public trust in the use of AI. Ethics in Tech Toolkit The Markkula Center for Applied Ethics at Santa Clara University developed the project to provide free resources for everyone to integrate ethics into their products and designs. The project includes a comprehensive ethical toolkit for engineering and design practice that consists of seven tools that can be combined with engineering and design workflows. The toolkit can be an attempt to ensure that practitioners responsibly develop technologies and ethics does not become ‘vaporware’. Playing with AI Fairness: What-if Tool Abbreviated as WIT, the tool developed by Google makes it easier to examine, evaluate, and debug ML systems easily and accurately. The open-source interactive visual tool enables the understanding of a classification or regression model by enabling users to examine, evaluate, and compare models. Thanks to its simple user-friendly interface and reduced dependency on coding, this tool caters to a wide range of users. Whether you’re a seasoned developer, researcher, or student, this resource can integrate into your workflow through TensorBoard or as an extension within a Jupyter or Colab notebook. Ethics and Algorithms Toolkit The toolkit developed by Joy Bunaguro and his colleagues asks a series of questions that grade the different types of risk in a data-driven initiative. Depending upon the level of risk, the toolkit contains suggestions for mitigations. Aequitas Developed in 2018 by the University of Chicago Center for Data Science and Public Policy Aequitas is an open-source bias audit toolkit. The tool is built to be used by a broad set of people from developers to policymakers. The purpose of the tool is to audit machine learning models for discrimination, bias, and make informed and equitable decisions around predictive risk-assessment tools. Human-AI eXperience (HAX) Toolkit The tool by Microsoft is best used early in the development process to design human-centric AI systems. The toolkit is made of four parts: HAX Workbook, HAX design patterns, HAX Playbook, HAX Design Library. HAX draws its foundation from a 2019 research paper, effectively turning its theoretical principles into practical tools. In many ways, these guidelines advocate for clarity and precision in UI copy, while being easy to implement plans in case of system failures.","excerpt":"A guide for people creating fairer, more robust and transparent AI system","categories":["AI Trends"],"tags":["AI Research"],"author_name":"Tasmia Ansari","publish_date":"2023-09-05T16:01:34","publication_year":"2023","word_count":625,"keywords":["data science","Go","machine learning","AWS","AI","ML","AI Research","Colab","Rust","Jupyter","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Jupyter","Colab","AWS","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/7-toolkits-to-build-your-ai-ethically\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":64505,"title":"Top 7 Free Online Resources To Learn Kotlin","content":"Developed by JetBrains, Kotlin is an officially supported language for Android development that was officially launched in February 2016. The statically typed programming language has soon become popular among android developers due to its ease of use. One of the most significant advantages that Kotlin has over Java is that one can write the same algorithm in Kotlin with fewer lines of code. Besides, it is compatible with the latest version of Java, which allows one to convert their application developed using Java to Kotlin easily. Such advantages have assisted developers to quickly build android applications and launch in the market. According to GitHub, Kotlin was fourth in the fastest-growing languages in 2019, only behind Dart, Rust and HCL. And one of the reasons why Java lost the first position to Python on GitHub is that android developers are moving away from Java to Kotlin. Here are the top 7 free resources to learn Kotlin and make powerful android applications:- Kotlin Bootcamp For Programmers Kotlin Bootcamp For Programmers is designed for beginners as it is only focused on teaching the fundamentals. It contains numerous practice sessions after every topic to help you implement the learning before moving on to the next concept. In the two weeks course, you will learn the basic syntax, define and call functions, object-oriented programming, and functional manipulation. However, if you do not have any prior programming experience in any other language, it will take more than two weeks for you. Introduction To Kotlin Programming Introduction To Kotlin Programming course is hosted on Oreilly by Hadi Hariri, VP of Developer Advocacy at JetBrains. He was actively involved in the development and management of the Kotlin. In the five hours and fourteen minutes course, Hadi introduces to the Kotlin programming language and teaches the fundamentals of the open-source language. Although it is not a completely free course, you can gain access to the course for free for ten days without providing your credit card details. For a 5 hour course, 10 days is adequate to learn without having to pay a penny. Kotlin Course – Tutorial For Beginners This is a two hour thirty-eight minute tutorial for beginners on YouTube by freeCodeCamp.org to teach you right from the basic control flow to functions and object-oriented programming in Kotlin. Since it is hosted on YouTube, it will not contain practice sessions to test your learning. However, for someone who already has a programming language experience and can grasp new concepts quickly, this can be a good starting point. Besides, if you want to explore Kotlin before taking in-depth courses, this tutorial will provide you with an overview of the open-source programming language. Kotlin Tutorial For Beginners If you prefer learning by reading instead of video lessons, then this well-structured blog will assist you in learning basics as well as advanced concepts in Kotlin. The programing language provides various advanced features such as operator overloading, lambda expressions, and string templates. This blog offers a complete tutorial with code examples, but you will have to practice on your own to retain what you learn. Kotlin For Java Developers The course hosted on Coursera is focused on basic syntax, nullability, functional and object-oriented programming. However, not all the fundamentals are covered in the course. This is not recommended for beginners as it requires prior knowledge of Java programming language to take this course. Enrolling in this course will help you in acquiring knowledge of the difference between Java and Kotlin, and how the latter supports interoperability with Java. Offered by JetBrains, it will take 25 hours to complete the course and switch to Kotlin from Java programming. Developing Android Apps With Kotlin Developing Android Apps With Kotlin is a two-months intermediate course developed by Google and Udacity to provide hands-on experience for developers using best practices. However, it requires prerequisite knowledge of object-oriented programming, GitHub, and familiarity with modern IDEs. The course will help you in making your first android application with Kotlin. You will learn how to make layouts, app navigation, activity and fragment lifecycle, app architecture (UI and data layer), and RecycleView. Kotlin Notes For Professionals Kotlin Notes For Professionals book contains 37 chapters on almost every topic of the programming language to assist you in quickly revising or relooking for codes and approaches. The book acts more like a notebook of every concept of Kotlin programming language. You can skip making notes while learning as this book serves the purpose of penning down your learning. Keeping this handy will make your life easier as you can quickly refer to the concept whenever you need and recall the learning of various techniques.","excerpt":"Developed by JetBrains, Kotlin is an officially supported language for Android development that was officially launched in February 2016. The statically typed programming language has soon become popular among android developers due to its ease of use. One of the most significant advantages that Kotlin has over Java is that one can write the same […]","categories":["AI Trends"],"tags":["ai lessons for beginners","Why is Python so Popular"],"author_name":"Rohit Yadav","publish_date":"2020-05-05T15:00:00","publication_year":"2020","word_count":773,"keywords":["Go","AI","Why is Python so Popular","Git","RAG","Python","ai lessons for beginners","ViT","Rust","GitHub","R","Java"],"extracted_tech_keywords":["AI","RAG","Python","R","Go","Rust","Java","Git","GitHub","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-free-online-resources-to-learn-kotlin\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10095442,"title":"Is This the End of LangChain?","content":"Last week, OpenAI released a slew of updates. Of these, function calling in the Chat Completions API was the most important. The feature uses external APIs and tools with OpenAI’s API. This, surprisingly, has a striking resemblance with LangChain, which also performs similar action, bringing us to question its relevance when building autonomous agents. Machine learning expert Santiago said that using function calling, we can define custom functions the model can use to answer questions. “For example, we can now have GPT solve queries requiring real-time data like stock prices, weather information, or match results,” he added. He said that previously this new ‘calling function’ feature was only possible using LangChain, but now it is natively available as part of the API. In other words, with this new update, users now have the choice to bypass the third party (i.e. LangChain) and directly built using OpenAI’s API – which is available for its GPT-3.5-Turbo and GPT-4 models. Relevance of LangChain LLM expert Derek Haynes said, “Pretty stoked with OpenAI’s function calling. My previously massive ReAct-based agent prompt has been reduced to this. It’s a fine-tuned GPT for agents … and makes me wonder if LangChain is worth the complexity?” In his blog post, Haynes explained the need for LangChain and the role of ‘calling function’ update introduced by OpenAI. He said that ReAct presents an approach where LLMs like GPT-3.5-Turbo and GPT-4 can use external tools to perform tasks, observe the outcomes, and use those outcomes to inform the next actions. For example, a model can use Zapier to search unopened emails and OpenAI’s models to summarise them. However, using this approach requires users to do considerable work, including listing available tools in a format that the LLM understands, applying custom code to parse the response and extract useful data, and guiding the LLM to perform tasks accurately. Haynes believes that this process can result in complex, brittle string parsing code. He said, that is where LangChain comes into the picture, bringing structure to this process, but this in turn can make the code less readable. A user, who goes by the name ‘fbrncci’ on HackerNews, said that the deeper and more complex the application becomes, the more of a risk LangChain seems to become to keeping it maintainable. Enters function calling On their website, OpenAI provided examples and applications of the update. (as seen below) “Function calling within your own applications is one of the interesting waves that we’re seeing with language models, which is increasing their capabilities outside of just normal text generation,” said AI expert Greg Kamradt, in an explainer video – Function Calling via ChatGPT API – First Look With LangChain. He said a lot of systems need to make decisions as well and when you use a language model as a reasoning engine, freeform text isn’t the best way to talk to other computers. It’s better if you can do it in a JSON format and that’s what function calling is a step towards. Here’s a quick glimpse on how function calling works: On a sprint mode Within an hour of OpenAI’s new update, LangChain responded with support. While users are free to choose between the two, it would be easier to use with its variety of features and plugins. And they aren’t the only one offering these services, with Hugging Face keeping up as well. But, one question remains: did OpenAI steal LangChain‘s limelight. “Yes, OpenAI took this from LangChain,” said Santiago. He, however, said that this will not kill LangChain. “There’s much more functionality there,” he added, questioning OpenAI’s plans of syphoning off from them and integrating it natively on its API.","excerpt":"With OpenAI’s new function calling API update, users have the choice to bypass Langchain and directly built using OpenAI’s API","categories":["AI Highlights"],"tags":[],"author_name":"K L Krithika","publish_date":"2023-06-20T12:00:00","publication_year":"2023","word_count":609,"keywords":["Go","ChatGPT","Hugging Face","machine learning","API","OpenAI","AI","autonomous agents","LangChain","R"],"extracted_tech_keywords":["AI","machine learning","ChatGPT","OpenAI","LangChain","Hugging Face","autonomous agents","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/is-this-the-end-of-langchain\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":57663,"title":"Freshworks To Improve Customer Experience Through Acquisition Of AI Startup AnsweriQ","content":"Freshworks — a SaaS-based company — has acquired Seattle-based AI startup AnsweriQ for an undisclosed amount. Founded in 2015, Freshworks is now valued at $3.5 billion and has closed about 10 acquisitions since then. The company offers a wide range of products such as freshdesk, freshsales, freshchat, among others. And its platforms include Freddy AI — an analytics solution — that provides predictive insights across the customer journey. Powered by AI\/ML capabilities, with Freddy AI, one doesn’t require data science expert to unlock the value of customers. However, in a continuous attempt to enhance its solution, the firm acquired AnsweriQ, an AI- and RPA-based solution provider that assists in answering customers’ queries, thereby delivering superior customer experience. The acquisition will complement Freshwork’s AI engine and help customers receive immediate answers to queries. AnsweriQ is one of the major players in the customer service automation space for large clients in call centres. “The integration of AnsweriQ’s technology will increase our AI capability in customer engagement space to offer significant value to our customers,” said Girish Mathrubootham, founder and CEO of Freshworks. AnsweriQ’s co-founder and CEO Pradeep Rathinam will join Freshworks as chief customer officer, becoming a part of the leadership of the Chennai- and San Mateo-based company as it eyes a listing in the US in the coming years. Rathinam said that the company was looking to further gain customers and enhance its products. Consequently, they had to choose between raising funds or integrating themselves with others. And they thought of associating with Freshworks as the best way forward for the company as they could quickly gain new customers. With this Freshworks’ acquiring spree continues as it already got hold of customer success service Natero last year. Thus, one can expect a few more companies coming in, as Freshworks’ in November, had raised Series H fund worth $150 million.","excerpt":"Freshworks — a SaaS-based company — has acquired Seattle-based AI startup AnsweriQ for an undisclosed amount. Founded in 2015, Freshworks is now valued at $3.5 billion and has closed about 10 acquisitions since then. The company offers a wide range of products such as freshdesk, freshsales, freshchat, among others. And its platforms include Freddy AI […]","categories":["AI News"],"tags":["how ai works"],"author_name":"Rohit Yadav","publish_date":"2020-02-28T08:16:55","publication_year":"2020","word_count":308,"keywords":["how ai works","data science","Go","programming_languages:R","AI","RPA","ML","automation","analytics","R","startup"],"extracted_tech_keywords":["AI","ML","data science","analytics","R","Go","automation","RPA","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/freshworks-to-improve-customer-experience-through-acquisition-of-ai-startup-answeriq\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61158,"title":"Arvind Krishna Takes Over As The CEO of IBM — Betting Big On Cloud &#038; AI","content":"As Arvind Krishna takes over the leadership charge of IBM as the CEO, he posted a letter on LinkedIn to IBMers — “My first day as CEO – our journey together” about the public health crisis, IBM’s essential role in the world, and his commitments as the new CEO of IBM. In the letter, Krishna wrote about how the employees should take care of themselves in this vulnerable time, alongside also stating that he wants the company to add a more significant presence in the hybrid-cloud space to its already established positions in the mainframe, services, and middleware ecosystem. “The fundamentals are already in place,” wrote Krishna. “Our approach to hybrid cloud is the most flexible and the most cost-effective for our clients in the long term. Coupled with our deep expertise, IBM has unique capabilities to help our clients realise the potential of a hybrid cloud business model.” Krishna further added that hybrid cloud and artificial intelligence are two dominant forces that are driving change for IBM’s clients and must have the maniacal focus of the entire company. Besides, IBM has always been considered an innovation powerhouse, as IBM has been leading on all fronts like cloud, AI, blockchain and quantum. Going forward in the letter, Arvind Krishna, specified a few actions that are going to take place during his leadership, which will help them in bringing leading solutions to their clients. According to Arvind, firstly, IBM needs to deepen their understanding of two strategic battles, which are the journey to hybrid cloud and artificial intelligence. He wrote, “We all need to understand and leverage IBM’s sources of competitive advantage. Namely, our open source and security leadership, our deep expertise and trust, and the fact that we enable clients to build mission-critical applications once and run them anywhere.” “I’m writing to you for the first time as your new CEO amid this global public health crisis online any other that we have faced.” — Arvind Krishna, Chief Executive Officer, IBM. Secondly, IBM needs to win the architectural battle in the cloud front, as he sees a unique window of opportunity for IBM and Red Hat to establish Linux, containers and Kubernetes as the new standard. “We can make Red Hat OpenShift the default choice for hybrid cloud in the same way that Red Hat Enterprise Linux is the default choice for the operating system,” wrote Krishna. Thirdly, as a company, IBM should be obsessed with continually delighting their clients. For this, he wrote, “At every interaction, we must strive to offer them the best experience and value. The only way to lead in today’s ever-changing marketplace is to innovate according to what our clients want and need constantly.” Krishna also noted that culture is everything, and culture is what that drives capability in any organisation. “One of my key priorities will be fostering an entrepreneurial mindset across our business, “wrote Krishna. “This is about being nimble, pragmatic and aiming for speed over elegance. And, it’s about being comfortable with ambiguity and continuously adapting to shifting circumstances.” Krishna, who drove the acquisition of RedHat, will be betting heavily on the cloud to revive the fortunes of IBM. Highlighting specific actions for the future, Krishna urged IBM employees to make use of the company’s sources of competitive advantage. Concluding the letter, he wrote, “IBM is a strong company. In our 109-year history, we have weathered countless storms and seen many crises unfold before our eyes. Today, we are financially strong, and we have a loyal client base. When this crisis ends, I’m confident that IBM will emerge strong, and we will be focused on growth. Few companies have the trust, credibility and cumulative wisdom to change the fabric of society through technology the way that IBM can.” “I love this company. In my 30+ years with IBM, I have seen first-hand the tremendous talent and dedication that IBMers possess. For all these reasons, I’m truly honoured and humbled to be your CEO and to lead an iconic, storied and innovative company like IBM,” concluded Krishna.","excerpt":"As Arvind Krishna takes over the leadership charge of IBM as the CEO, he posted a letter on LinkedIn to IBMers — “My first day as CEO – our journey together” about the public health crisis, IBM’s essential role in the world, and his commitments as the new CEO of IBM. In the letter, Krishna […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Arvind Krishna","Cloud Computing","IBM"],"author_name":"Sejuti Das","publish_date":"2020-04-07T18:30:00","publication_year":"2020","word_count":674,"keywords":["Go","artificial intelligence","AI","innovation","RAG","Aim","Cloud Computing","IBM","Arvind Krishna","Rust","GAN","R","kubernetes","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","kubernetes","R","Go","Rust","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/arvind-krishna-takes-over-as-the-ceo-of-ibm-betting-big-on-cloud-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39320,"title":"Top 12 Python Web Frameworks of 2019 Programmes Should Know","content":"Python is a very versatile programming language. A new developer survey about Python by the Python Software Foundation and developer tools vendor JetBrains, revealed the programming language is used for web development after data analysis. Here are the top Python web frameworks of 2019. 1.Bottle The Bottle framework is a small-scale or minimalistic Python frameworks. Originally meant for building web APIs, Bottle is designed to be very small and hence tries to execute everything in a single document. It has no dependencies other than the Python Standard Library. It is lightweight, fast, and easy to use, and is well-suited to building RESTful services. Netflix uses Bottle for its web interfaces. 2.CherryPy CherryPy is an open source object-oriented Python framework which has its own multi-threaded web server and welcomes anyone who wants to contribute. Applications made using CherryPy can run on all operating systems supported by Python, that is, Windows, Linux\/Unix and macOS. It to build web applications in much the same way they would build any other object-oriented Python program. This results in smaller source code developed in less time. CherryPy has a reliable HTTP\/1.1-compliant, WSGI thread-pooled web server and provides support for different web servers. The framework also allows to run several HTTP servers simultaneously. It has tools for caching, encoding, authorization and supports profiling, testing, and coverage by default. A has a flexible plugin system. CherryPy is now more than ten years old and it is has proven to be very fast and stable. It is being used in production by many sites. 3.CubicWeb CubicWeb is a full stack open-source Python framework developed and curated by Logilab. CubicWeb makes use of cubes which are components to build web applications, where multiple cubes are joined together for creating an instance with the help of a database, a web server and some configuration files. CubicWeb has a query language named RQL similar to W3C’s SPARQL. It has a selection+view mechanism for semi-automatic XHTML\/XML\/JSON\/text generation and a library of reusable components. CubicWeb is a proven end to end solution for semantic web application development that promotes quality, reusability and efficiency. 4.Dash Dash is an open-source Python micro framework for building analytical web applications. It is built on top of Flask, Plotly.js, React and React Js and is very suitable for data scientists to work with. Using PythonIt dashboards can be built using Dash. It has a high-degree of customization and a simple interface for tying UI controls, including dropdowns, graphs and sliders. A Dash application is usually composed of two parts. The first part is the layout and describes how the app will look like and the second part describes the interactivity of the application. Dash provides HTML classes that allow the generation of the HTML content with Python. Dash apps are rendered in the web browser and can be deployed to servers and are hence cross-platform and mobile-ready. 5.Django Django is an open-source framework of Python that encourages rapid development and clean, pragmatic design. It was designed to help developers take applications from concept to completion as quickly as possible. Django takes security seriously and helps developers avoid many common security mistakes. It has less need for coding, and reusability of components. A few of the key features of Django, such as authentication mechanism, URL routing, template engine and database schema migration implements Object Relational Mapper (ORM) for mapping its objects to database tables. The framework underpins numerous databases including PostgreSQL, MySQL, Oracle, and SQLite, which implies that a similar coding works with various databases. Some of the busiest sites on the web leverage Django’s ability to quickly and flexibly scale. Instagram, Pinterest, Mozilla and Spotify are two such examples. 6.Flask Flask is a Python microframework which has a built-in development server and debugger. In this framework, there is no built-in database interaction, but the flask-sqlalchemy package will connects an SQL database to a Flask application. The flask-sqlalchemy package needs only the database URL to connect to a SQL database. It is BSD licensed, supports unit testing, establishes secure client-side sessions and is compatible with Google App Engine. 7.Giotto Giotto is a full-stack framework for building applications in a functional style. It is based on the concept of Model, View and Controllers in order to allow web designers, web developers and system admins to work independently. Giotto includes controller modules that enable users to create apps on top of the web, Internet Relay Chat (IRC) and command line. It is built with a view for the code to be easily maintainable over a long period, which would eventually result in deploying the code quickly. It has an inbuilt cache with support for Memcache and Redis, an automatic URL and database persistency with SQLAlchemy. 8.Pyramid Pyramid is a mega framework of Python that makes decisions for the user. But if one does not fit into their viewpoint, he ends up fighting his decisions. It stops making decisions as the application grows, and so, its primary goal is to achieve as much as with minimalistic complexity. It can work equally well with small as well as full-scale applications and has a flexible authentication and authorization. 9.Sanic Sanic is a Python 3.6+ web server and web framework that’s written to go fast and handles fast HTTP responses via asynchronous request handling. It allows the usage of the async\/await syntax making the code non-blocking and speedy. It also welcomes contributions. In a benchmark test with one process and 100 connections, Sanic was able to handle 33,342 requests per second. 10.Tornado Tornado is a scalable, non-blocking web server and web application Python framework. It was developed for use by FriendFeed; the company was acquired by Facebook in 2009 and Tornado was open-sourced soon after. The framework is built specifically to handle asynchronous processes. By using non-blocking network I\/O, Tornado can scale to tens of thousands of open connections, making it ideal for long polling, WebSockets, and other applications that require a long-lived connection to each user. It serves the application with its own HTTP server and hence the user has to set up how the application is served. 11.TurboGears TurboGears has two versions, 1 and 2, among which TurboGears 2 is the latest one and is a full stack framework. The framework starts as a microframework and scales up to a full stack solution. It supports sharding, multiple data-exchange formats and has a built in extensibility pluggable applications and standard WSGI components. It allows developers to rapidly develop extensible data-driven web apps. The full-stack framework makes use of components such as Genshi, Repoze, SQLAlchemy, and WebOb to easily and quickly develop apps requiring database connectivity. 12.Web2Py Web2Py is one of the most popular open source web frameworks for Python. In addition to being written in Python, Web2Py also comes with a Python interpreter. It supports model-view-controller (MVC) architecture, along with commonly used web development practices like server-side sessions, self-submission of web forms and safe handling of uploaded files. It has a web-based IDE and batteries that helps developers to build custom web applications. Web2Py supports MVC architectural pattern. The pattern enables developers to simplify development of complex web applications by dividing them into models, views, and controllers. It helps in keeping the web application portable. It can be deployed on several virtual private networks (VPNs) and cloud platforms. It comes with a package that includes a web-based admin interface, web-based management interface, SQL database, database abstraction layer, error logging and ticketing system, and multi-threaded web server. The package enables developers to build a variety of web applications efficiently without using external tools and services. Web2Py also improves the security of web applications by addressing top security issues.","excerpt":"Python is a very versatile programming language. A new developer survey about Python by the Python Software Foundation and developer tools vendor JetBrains, revealed the programming language is used for web development after data analysis. Here are the top Python web frameworks of 2019. 1.Bottle The Bottle framework is a small-scale or minimalistic Python frameworks. […]","categories":["AI Trends"],"tags":["django","django python","frameworks","language","Python"],"author_name":"Disha Misal","publish_date":"2019-05-17T07:47:21","publication_year":"2019","word_count":1272,"keywords":["PostgreSQL","Go","Plotly","AI","language","ML","django","frameworks","Python","RAG","django python","SQL","R","Redis"],"extracted_tech_keywords":["AI","ML","Plotly","RAG","Redis","PostgreSQL","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-12-python-web-frameworks-of-2019-programmes-should-know\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":39153,"title":"Do You Require A University Degree To Enter Data Science Field? Not Anymore","content":"India has been witnessing a drastic increase in the number of courses for AI, data science and other emerging technologies. With many premier institutions opening up a course, it can only be surmised that having a degree in data science is considered to be a plus point by many companies in India. A study by AIM suggested that India’s elite engineering institutes can earn a base salary of 22 lakh while those from second-tier institutions earn a package 12 lakh, thereby stressing on the need to have a good education in the field. While the above stats incline towards the fact that AI and data science course can prove to be a positive step in getting the next-gen ready for employment in this thriving sector, there are contradicting results suggesting that bigger companies are no longer looking for candidates to have a degree. Challenging The Industry Norm For long, degrees have been a proxy for pedigree, commitment, interest, and intent; social proof that you did what you were “supposed” to do. But it may not be the case any longer. There are many industries today that are expanding into newer horizons while adopting newer ways of recruiting talent. According to a recent study, many popular companies in the US do not require a college degree and certain jobs are most likely to be filled by non-college graduates than others. It found out that many of today’s hottest companies do not require employees to have a college degree. These are especially roles including electronic technicians, mechanical designers, and marketing representatives. The report also stated business leaders such as Barbara Humpton and Tim Cook questioning the need for four-year degrees altogether, stating that most of the staff includes people without higher credentials (four-year degree). They posited that many colleges do not teach the skills that business leaders need the most in their workforce, such as coding. While it can be liberating at many levels, it puts a major question in front of us  — and that is — if college degrees really prepare workers for carving out a stable career? How Does This Apply In Indian IT Landscape Plagued By Retrenchment Talking about data science industry, hiring is predominantly based on practical knowledge of the various tools and techniques that are used in the industry. It means that companies stress on facts such as understanding of programming languages such as R, Python, hands-on applications of visualisation tools, SQL and much more. It is the applicability of these tools that matter than just the theoretical concepts. A graduate student on an average spends 3 to 4 years in college during graduation while going through the theoretical aspects that may not be very relevant and are outdated to be applied in the current working scenarios. Especially data science and analytics requires continuous learning and evolution in terms of skills and techniques to keep up with the ever-evolving industry. Data literacy is the key requirement for the current jobs, having a thorough understanding of dealing with large amounts of data, collecting it and generating insights from it. The current education scenario in India does not expose a candidate to all these practical learnings, and more often than not, they either learn it on the job or take exclusive courses on these tools in addition to the graduate degrees. While it can be said that practical understanding weighs more, but if we think practically, companies in India may not be 100% open to hiring just on the basis of skills and no graduate degree. Having said that, there are many industry experts who believe that companies will set out realistic expectations from a candidate’s career and recruit based on pure skills. How Can Data Science Professionals Stay Relevant? Data science professionals are always advised to learn new skills, build an awesome portfolio in platforms such as Github, Dribble, and other blogs that keep showcasing the skills you have. It underlines the fact that a candidate knows more than just having a degree. Companies can bring up programs such as apprenticeship programs, mixing on-the-job and school training, helping in preparing a better workforce for the future. Many bigger companies are already recruiting based on skills put up on GitHub which shows that they are weighing up hands-on experience more than college or university degree, if not completely oversee it. “In the Data Sciences industry, it is important to realize that knowledge is our only currency. There is a continuous need to learn which can only be made real by inculcating habits that are designed deliberately and practised regularly,” said Tapan Rayaguru, COO, Tredence in an interview with AIM. This practice is definitely going to be mainstream with many other companies hiring purely on the basis of skills, but it may be far away.","excerpt":"India has been witnessing a drastic increase in the number of courses for AI, data science and other emerging technologies. With many premier institutions opening up a course, it can only be surmised that having a degree in data science is considered to be a plus point by many companies in India. A study by […]","categories":["AI Features"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2019-05-14T13:24:09","publication_year":"2019","word_count":796,"keywords":["data science","Go","AI","RAG","Python","Ray","Aim","analytics","SQL","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","Ray","RAG","Python","R","SQL","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/do-you-require-a-university-degree-to-enter-data-science-field-not-anymore\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10065922,"title":"Driving successful AI transformations at the enterprise level. Interview with Chandramauli Chaudhuri, Fractal","content":"Chandramauli Chaudhuri leads the Data Science initiatives across Fractal’s Tech Media & Telecom vertical in the UK & Europe. He works in close collaboration with senior business stakeholders and CXO teams across some of the leading global enterprises, enabling the development of long-term strategic AI solutions. Being in the field of Artificial Intelligence and Machine Learning for close to a decade and working across a wide range of industries, his primary area of interest lies in R&D, algorithmic customisation, capability enhancement, and MLOps deployments of solutions. Analytics India Magazine interviewed Chandramauli to gain insights into AI transformation at the enterprise level. AIM: What are the key factors when it comes to driving successful AI transformation for an organisation? What are the emerging AI trends enterprises are capitalising on? Chandramauli: As a business leader driving AI transformation across an organisation, it is critical to understand that Artificial Intelligence is just the means of value realisation and not an end goal by itself. Thus, the factors differentiating success and failure lie in its synergy with the company’s core principles, value proposition and customer-centricity. AI adoption is not a plug-and-play solution that yields overnight returns. Businesses need to think beyond just the cutting-edge software, high-end infrastructure and skilled coders. Alignment of the company’s culture, customer expectations and ways of working to support such transformations need to take equal if not greater importance. The companies that are doing well, especially in banking, finance, media, telecom, and tech, are those that have integrated AI into their day-to-day functions. They are moving it away from being a siloed and ‘specialised’ initiative undertaken in small pockets, to broader cross-functional collaboration. As far as emerging trends are concerned, organisations have started focusing a lot more on two key areas – execution excellence and risk management. This means nurturing an agile mindset across teams, pursuing the right use cases, developing a strong data foundation, investing in the right skills, and having a robust strategic roadmap. There has also been growing acknowledgement of the challenges associated with cybersecurity, user privacy, and digital consent. Issues like lack of explanations, absence of audit trails and presence of bias in AI systems have gained far greater prominence from the global community in the last couple of years than in the past decade. It’s true that we still have a long way to go and yet to fully appreciate the complex socio-political and economic implications. However, we have started looking in the right direction, focusing on building greater transparency and trust. The early adopters of these practices stand to reap the rewards in both the short and the longer term. AIM: Where does the industry stand in terms of scaling AI solutions? How can AI\/ML become a differentiating factor for companies? Chandramauli: AI opens new frontiers to solving real-world problems. We have already seen some great examples of AI powering decisions in almost every domain, from climate change to the choice of songs. Add to this, the ability to augment with new-age technologies like 5G, IoT, AR, VR, etc., and scale through cutting-edge hardware, open-source architectures and cloud computing – what we gain is the prowess to redefine the limits of end-user personalisation and engagement. The resulting increase in efficiency, effectiveness and productivity has a direct impact on the top and bottom line. It is redefining the way successful businesses look at their strategic and operating models. Naturally, there is a lot of excitement around the future and a rush to seize any competitive advantage. But, this is only one side of the story. For the vast majority of companies that are trying to drive innovation and scale their AI operations, progress has not been at the scale or pace that people might assume. Many are still struggling to push past the pilot and experimentation phases. This is mainly because, the traditional mindsets, legacy technologies and ways of working run counter to that needed for a full-scale AI transformation. Studies suggest that, currently, only about 8-10% of the firms engage in practices that support widespread adoption. For the remaining ones, enabling AI transformation at scale still very much remains a work in progress. AIM: How does AI help organisations add value and manage risks? Chandramauli: Large enterprises need thousands of decisions to be made every single day. Doing so effectively, requires a combination of process automation, contextual insights and cognitive engagement. Not only that, shifts in the industry landscape can trigger quick changes in customer behaviour, rendering past insights completely useless. Such dynamicity poses a significant challenge for the traditional software and analytical solutions, which depend largely on pre-defined logic and set patterns. The COVID-pandemic has been a reality check for many organisations in this regard. AI, on the other hand, is particularly well suited to address such real-life challenges. Consider a business problem like content piracy as an example. Despite decades of effort, it continues to be a perpetual problem in the media and entertainment industry. It causes billions of dollars in losses every year. The volume of pirated content across the globe, in fact, reached record highs during the lockdowns. This is because, the consumers and distributors of pirated content are continuously changing, updating and scaling their operations beyond the limits of traditional anti-piracy systems. AI-powered solutions, learning and re-calibrating in real-time from online trends, network logs and consumption data feeds, can be much more effective at identifying and mitigating such behaviour. AIM: What are the things organisations should keep in mind while building an AI roadmap? Chandramauli:  To build successful AI roadmaps, leaders need to devote attention to four key aspects. First, define a strong and clear narrative, that explains to everyone what AI is and why it is so critical to the future of the organisation. This needs to start from the top – the board and the executive team, along with the key decision-makers including managers and team leads. Second, dedicate time and effort to address the unique barriers when it comes to such fundamental shifts – apprehensions of the workforce about becoming obsolete, difficulties in adopting the agile ways of working, etc. The leadership team has to provide a vision that brings everyone together and shows how they fit into a new AI-driven culture. Third, budget for integration, adoption, and training. This must be over and above that allocated for merely procuring the technology and infrastructure. Most successful companies, spend more than half of their analytics budgets on activities that drive adoption, such as workflow redesign, communication, and training. And finally, identify and prioritise the right use cases. This means striking the right balance between feasibility, investment, time and value. While the focus must not be restricted to just gaining quick wins, undertaking initiatives that may not get deployed or provides no returns in sight can severely jeopardise both current and future AI prospects. AIM: How do you choose relevant AI use cases that will deliver maximum impact for your organisation? Chandramauli: This is a particularly important question. Depending on where the organisation is in terms of its analytics maturity, the approach toward choosing the right use cases needs to change. For enterprises that are just embarking on their AI journeys, it is better to start with a few well-recognised and well-understood business problems. The primary focus should be on learning, identifying the gaps and gaining experience. The solutions need to be technically feasible within the current organisational constraints. Most importantly, while these initial use cases do not necessarily need to drive the biggest monetary benefit, the impact must be measurable through some established business KPIs. On the other hand, for organisations that are further along the path, factors like potential returns, time to value, cost of deployment, infrastructure availability, training requirements, etc., need to be accounted for. A good practice is to start by collaborating across cross-functional teams, including people from strategy, analytics, IT and operations sides, to understand the current needs and state of the business. The aim should be to arrive at a set of clearly defined annual objectives and accordingly prioritise the execution of use cases at a monthly or quarterly level which can help achieve the same. AIM: At what stage of the project should companies ideally look for AI-based digital transformation? Chandramauli: Whether it is the B2C or B2B space, digital transformation is about streamlining and improving customer experience. AI aids such transformations by identifying the areas that can maximise value. Thus, generally speaking, digital strategy should move hand-in-hand with AI transformation, right from day 1. It allows teams to be better informed, focus on the important processes and achieve better results. However, if we look at some of the larger organisations, this is not necessarily the case. In such situations, it is worth taking a step back and first assessing the need, effort, cost, time and value of undertaking an integrated digital-AI strategy. The key here is not to force AI into the mix just because other companies are doing so. This may be counter-productive, disrupting the normal functioning of the business. A more prudent approach would be to decide on a case by case basis, and depending on the company’s and customers’ needs, implement AI as required. Gradually, as the capability matures over time, the business can start focusing on bringing the two together under a single roof. AIM: Should companies invest in AI solutions developed in-house or hire third-party vendors? Chandramauli: This depends largely on the organisation’s needs and current level of maturity. Partnering with the right vendors has its advantages, especially in the early stages of AI adoption. These organisations bring years of specialised experience, learning, and innovation to the table. Outsourcing to partners with such deep technical knowledge can drive much-needed momentum and build the foundation for long-term AI success. It can also lower costs and mitigate long-term risks. For organisations that are more mature and have dedicated teams of data scientists, partnerships and collaborations at some level still make perfect sense. Many such companies opt for a hybrid approach – build some solutions in-house, licence or buy some SaaS platforms and outsource the rest to other partners. It provides the flexibility to quickly scale efforts up or down without deprioritising critical objectives. Most importantly, it brings in fresh perspectives and cross-industry learnings which can be invaluable in terms of driving innovation and building competitive advantage.","excerpt":"Chandramauli Chaudhuri leads the Data Science initiatives across Fractal’s Tech Media & Telecom vertical in the UK & Europe.","categories":["IT Services"],"tags":["Fractal","Fractal AI","Fractal Analytics","fractal analytics careers","Fractal interviews"],"author_name":"Shraddha Goled","publish_date":"2022-04-29T15:00:00","publication_year":"2022","word_count":1719,"keywords":["Fractal Analytics","data science","artificial intelligence","machine learning","Fractal interviews","AI","cloud computing","ML","MLOps","Fractal","Aim","fractal analytics careers","analytics","R","Fractal AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","MLOps","Aim","cloud computing","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/driving-successful-ai-transformations-at-the-enterprise-level\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":65340,"title":"Uncanny Valley Won The AI Song Contest Organised By Dutch Broadcaster VPRO","content":"On Tuesday, an Australian team — Uncanny Valley — won the AI song contest organised by Dutch broadcaster VPRO in cooperation with radio stations NPO 3FM and NPO Innovation. The track created by AI was named Beautiful the World, which was trained using historical audio samples of koalas, kookaburras, and Tasmanian devils. Besides, it also had an Austrian flavour — the song had a social message that was inspired by the vast forest fires that struck the country last year. The collective points of the international audience and AI experts was used to grade and award the winner. While the international audiences choose Beautiful the World, the AI panel experts — Vincent Koops, Anna Huang, and Ed Newton-Rex — gave full points to the German team Dadabots x Portrait XO. However, with 19.8 points to the Beautiful the World, it overpowered the I’ll Marry You, Punk Come by 0.4 points to win the contest. Unlike other AI music, this was not completely AI-generated. Only some elements of the songs were used to write melody and lyrics. Human intervention was vital to generate the songs that were easy on the ears. Case in point, the team that came last left the entire control to the AI for the track. “Faced with the choice between making an accessible song with quite a few human interventions or experimenting with as much AI as possible and then delivering a worse-sounding song, we chose the latter,” said the data scientist of the team that came last. Nevertheless, AI played a significant role in the song that won the contents. The one takes away from this contest is that AI can become a partner but not the lead in song generation. AI brings a unique dimension to the songs but cannot be expected to make complete songs yet.","excerpt":"On Tuesday, an Australian team — Uncanny Valley — won the AI song contest organised by Dutch broadcaster VPRO in cooperation with radio stations NPO 3FM and NPO Innovation. The track created by AI was named Beautiful the World, which was trained using historical audio samples of koalas, kookaburras, and Tasmanian devils. Besides, it also […]","categories":["AI News"],"tags":[],"author_name":"Rohit Yadav","publish_date":"2020-05-15T20:27:19","publication_year":"2020","word_count":302,"keywords":["AI music","programming_languages:R","AI","innovation","ViT","GAN","R"],"extracted_tech_keywords":["AI","R","GAN","ViT","innovation","AI music","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/uncanny-valley-won-the-ai-song-contest-organised-by-dutch-broadcaster-vpro\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058362,"title":"Can NFTs save journalism?","content":"The Associated Press is collaborating with Xooa to set up an NFT marketplace to sell the work of the news organisation’s photojournalists. The proceeds from the sale will fund AP’s journalistic endeavours. Slated to launch on January 31, the marketplace will allow collectors to buy, sell, and trade the news agency’s official award-winning photojournalism as NFTs. The platform will offer buyers the opportunity to own Pulitzer Prize-winning photos of significant historical moments. The initial collection will include photography by current and former AP journalists and a few digitally enhanced renderings of their work. It says something about NFTs that, in an article about them, you can't easily tell if the art that it runs with is a multimillion-dollar artwork being NFT'd or just, like, a stock illustrationhttps:\/\/t.co\/Qiin4iHF3M— Joshua Benton (@jbenton) January 11, 2022 NFTs in Journalism Major media outlets like The New York Times and Quartz have previously sold their articles as NFTs. For example, Kevin Roose’s article for The New York Times, titled “Buy This Column on the Blockchain,” was sold at a 24-hour auction for $5,63,000. He wrote: “Why can’t a journalist join the NFT party too?” Earlier, Fortune Magazine had sold digital versions of their 2021 magazine cover titled “Crypto vs Wall Street.” The money from the sales went to four organisations that support a free press and public service journalism: Reporters Without Borders, The GroundTruth Project, The Institute for Nonprofit News, and The Committee to Protect Journalists received $1,65,000 each. Who buys NFTs? People buying NFTs at exorbitant prices are called cryptocurrency whales— a term used for individuals and institutions in possession of large amounts of Bitcoin or Ethereum (likely because they invested in cryptocurrencies before Bitcoin’s price rose by more than 600 percent). A lot of NFT sales happen in the light of its rising value as futuristic assets in web3.0. It says something about NFTs that, in an article about them, you can't easily tell if the art that it runs with is a multimillion-dollar artwork being NFT'd or just, like, a stock illustrationhttps:\/\/t.co\/Qiin4iHF3M— Joshua Benton (@jbenton) January 11, 2022 What will AP’s marketplace look like? The NFTs are minted on the environmentally friendly (and Indian founded) blockchain Polygon. The platform accepts credit cards, debit cards, and Ethereum and will support crypto wallet MetaMask. Each NFT or collection of NFTs will be released at a specific time, date, and price through drops. Limited-edition NFTs will be released biweekly as “Pulitzer Drops,” and will feature Pulitzer-prize winning images. To protect the status of these drops, they will be made more scarce and difficult to obtain—with access to them being given only to the platform’s most active collectors. There will be a 10% fee for every resold collectible, which AP will share with Xooa. All purchased NFTs can be viewed under the section, “My NFTs.” AP already has a database of archival photos and footage used routinely by journalists across the industry and world. Does this mean we lose access to specifically award-winning photos once purchased by collectors? Will those NFT collections expand beyond that? https:\/\/t.co\/Co3ibcgXLg pic.twitter.com\/zLArmODuPT— Sana Saeed (@SanaSaeed) January 11, 2022 According to Jarrod Dicker of The Washington Post, NFTs could be a solution to bringing “an ownership element back to media.” He said content creators having greater control over their assets at the start of an idea could eventually give them more control over how their work is “used, licensed and distributed as well as a means to be able to collect revenue.” Community-based funding through NFTs could allow journalism to become more decentralised, said the director of the Tow Center for Digital Journalism at Columbia, Emily Bell.","excerpt":"The Associated Press is collaborating with Xooa to set up an NFT marketplace to sell the work of the news organisation’s photojournalists. The proceeds from the sale will fund AP’s journalistic endeavours.  Slated to launch on January 31, the marketplace will allow collectors to buy, sell, and trade the news agency’s official award-winning photojournalism as […]","categories":["IT Services"],"tags":["Associated Press"],"author_name":"Srishti Mukherjee","publish_date":"2022-01-14T11:00:00","publication_year":"2022","word_count":605,"keywords":["Go","funding","programming_languages:R","AI","programming_languages:Go","Git","Associated Press","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-nfts-save-journalism\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10001175,"title":"Facebook Gets Serious About Blockchain, Doubles Down On Acquihiring &#038; Scalability","content":"Even though cryptocurrencies and decentralised financial systems have been in the making for over three decades, they rose to prominence in the aftermath of the 2008 financial crisis. The trust in traditional banking services hit rock bottom and people needed a robust system which was trustworthy, and free of middlemen. This is where Facebook decided to go forward with Blockchain, an incorruptible digital ledger of economic transactions which can be programmed to record not just financial transactions but virtually everything of value. How Facebook Plans To Reinvent Itself Facebook is among the growing list of tech giants like Amazon and Microsoft, who are looking at Blockchain technology for more transparent and secure data systems. Acquihiring: These enterprises are employing ‘acquihiring’ as a strategy to step up the game for staying relevant in the near future. As a part of this strategy, companies are acquiring new talents and promising startups who are using blockchain technology. This way, they don’t have to reinvent the wheel, but just need to provide a larger platform for these startups to use their expertise on a larger scale. Facebook, began making moves towards cryptocurrencies in early 2018. Even though the social media giant was buried under a pile of controversies, it somehow managed to stay afloat on the innovation side. Cryptocurrency: Last year Facebook quietly launched an in-house startup, one focused on blockchain technology and cryptocurrency. Facebook also announced that they would be launching their first crypto product — a digital stable coin for remittance payments in India. This was a smart move considering the fact that India represents over 10% of the $613 billion global remittances market and lowering transaction costs and frictions while adding currently unbanked users and this could be a winning real-world use case. Management: Facebook’s small team of blockchain experts is led by David Marcus who was the vice president of Messenger earlier as well as the former president of PayPal. Facebook has now hired the team behind Chainspace. The acquisition of the Chainspace team is the clearest sign yet of the network’s ambition to be a big player in the nascent blockchain industry. Scalability: Facebook is also looking for scalability. In fact, they only showed interest in blockchain companies that could accommodate their large number of users. A Transparent And Secure Future This wouldn’t be the first time Facebook would attempt to disrupt the payments market with a virtual currency. Facebook earlier had ‘credits’ as a digital currency on its platform. A native payments solution, backed by blockchain, could revolutionise payments on Facebook and some of its most popular products. However, cryptocurrencies are still in a grey area with some countries like India, labelling them illegal for monetary transactions, while others being fairly open to the idea. Diving Into Digital Wallets: Blockchain can make transactions free or very cheap and Facebook plans to build a cryptocurrency wallet with its own token that people could use to pay for things with partnered businesses or that they discover through Facebook ads. Using Cryptocurrency For Ads: This development will also help Facebook monetise its cryptocurrency at a faster rate. With more than six million advertisers and 65 million businesses that have Facebook pages, passing the transaction fee savings on to the users, while touting partnerships with Facebook Crypto can boost sales for businesses. That could, in turn, get clients to spend more money on Facebook ads, as the discounts would enhance conversion rates and drive sales. Outlook Will Facebook acquire more startups? If yes, then what will it do with their assistance? The answers to these questions are still to be known. Going by the growing talent pool around blockchain technology within Facebook, it is quite obvious that this social media giant is doubling down on building a Blockchain unit within the company.","excerpt":"Even though cryptocurrencies and decentralised financial systems have been in the making for over three decades, they rose to prominence in the aftermath of the 2008 financial crisis. The trust in traditional banking services hit rock bottom and people needed a robust system which was trustworthy, and free of middlemen. This is where Facebook decided […]","categories":["AI News"],"tags":["Blockchain","crypto","Facebook"],"author_name":"Ram Sagar","publish_date":"2019-02-15T13:44:46","publication_year":"2019","word_count":630,"keywords":["Go","Blockchain","AI","ETL","innovation","Scala","Git","GAN","crypto","Rust","Facebook","R","startup"],"extracted_tech_keywords":["AI","R","Go","Rust","Scala","Git","ETL","GAN","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/facebook-gets-serious-about-blockchain-doubles-down-on-acquihiring-scalability\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094231,"title":"PyTorch Tabular v1.0.2 Released","content":"A new version, v1.0.2 of PyTorch Tabular has been released. The library now includes a new method in TabularModel for enabling feature importance. Feature Importance has been enabled for FTTransformer and GATE models. The release note was generated by ChatGPT using the git commit logs from the last release, wrote Manu Joseph the creator of PyTorch Tabular, GATE and LAMA-Net on LinkedIn. Check out the GitHub repository to learn more about the update. PyTorch Tabular is a deep learning library that makes working with deep learning and tabular data easy and fast. The library has been built on frameworks PyTorch and PyTorch Lightning, and it works on pandas data frames directly. The framework makes the standard modeling pipeline simple enough for practitioners while also being reliable enough for production use. It also focuses on customization so that it can be used in a variety of research settings. The below picture depicts the structure of the framework. Latest Additions The latest enhancements  in the library include several updates. Firstly, two additional parameters have been added to the GATE model, expanding its functionality and versatility. Additionally, the library configuration now includes the metric_prob_input parameter, providing improved control over metrics within the models. The GATE model has undergone slight improvements, including adjustments to defaults that enhance its overall performance. Furthermore, various minor bug fixes and improvements have been implemented, such as the addition of accelerator options in the configuration and enhancements to the progress bar. Alongside these enhancements, the library has been updated with newer versions of dependencies, including docformatter, pyupgrade, and ruff-pre-commit. These updates contribute to the library’s overall reliability, functionality, and performance. Read: How to Handle Tabular Data for Deep Learning Using PyTorch Tabular? The latest version has been released four months after v1.0.1. Other improvements include various code optimizations, bug fixes, and CI enhancements. For more details, refer to the commits on the library’s GitHub repository. PyPi: https:\/\/pypi.org\/project\/pytorch-tabular\/ Documentation: https:\/\/pytorch-tabular.readthedocs.io\/en\/latest","excerpt":"The library now includes a new method in TabularModel for enabling feature importance","categories":["AI News"],"tags":["Python","Pytorch"],"author_name":"Tasmia Ansari","publish_date":"2023-05-31T16:25:34","publication_year":"2023","word_count":320,"keywords":["Pytorch","ChatGPT","Go","AI","PyTorch","Git","Python","GPT","deep learning","GitHub","R","Pandas"],"extracted_tech_keywords":["AI","deep learning","ChatGPT","PyTorch","Pandas","R","Go","Git","GitHub","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-tabular-new-version-v1-0-2-released\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10020195,"title":"CX At The Convergence Of Marketing, Sales &#038; Customer Support: An Inside Look","content":"The pandemic has been a primary driver in disrupting digital marketing. Today, consumers look for highly-personalised content. Right on cue, businesses are upping their game to keep up with the shift in consumer behaviour. “The need of the hour for marketers is to key out every possible information that is at their disposal and reach out to customers with a customised message. Now, this is not easy to do,” said Somashankar Ghosh, VP, Consumer and Commercial Analytics Practice at Genpact. Along with his colleague, Pramit Dasgupta AVP, Customer Experience Analytics Practice at Genpact, Ghosh delivered a talk titled “Unification and transformation of customer experience and targeting” at MLDS 2021. Customer Experience Customer experience has become the place where the functions of marketing, sales, and customer service converge. As per Ghosh, by 2023, 25 percent organisations will integrate the three areas into a single function. That said, currently, 42 percent organisations find working across silos as a significant barrier to streamlining their functions. “Because customers are always on an anytime\/anywhere, digital transformation and in-moment interactions are imperative. Enterprises need to imbibe this in their omni-channel strategies. These channels need to be connected and not merely kept there for the customers to access. The customer experience today depends on the end-to-end touch experience and not just on the end product. In-depth understanding of the customer becomes a key differentiator,” said Ghosh. Apart from the first-hand knowledge marketers gain on customer behaviour based on customer interaction with the company and other sales data, they must also consider third party data. This helps in understanding what exactly the customer is looking for. ‘Inside-Out and Outside-In View’ “From a customer perspective, the product utility, surrounding experience, and access to support whenever and wherever needed are part of a singular journey in customer experience. This determines a customer’s affinity towards the product, the stickiness to the relationship and brand perception. Unlike traditional marketing functions, nowadays all of these affect profitability. In the inside-out part, these factors operate in silos from a strategic and technical standpoint. Customer experience can only come to life when capitalised through such integrations of data augmented practices,” said Ghosh. He said the unification of sales, marketing and customer support will depend on three main pillars–data, augmented practices, and agile operations. Data: Data needs to be managed, understood, and analysed to know whether it is reliable. These days, marketing organisations are particular about accuracy, timeliness and trustworthiness of data. Data is collected from multiple sources. Data needs to be managed efficiently for it to be effective. Collecting data and aligning them helps in understanding the customer.  It requires optimal usage, measurement and analysis of relevant data from sources coming together in an assimilated data queue — this is how efficiency leads to effectiveness. It is a mammoth task. However, when data does become truly effective, it positively influences the outcome. Augmented Intelligence: This has more to do with AI’s role as an assistive technology in enhancing human intelligence and insights. “The use of human judgement to enable the best possible decision at the right time to improve the decision-making process using technology is Augmented intelligence,” explained Ghosh Agile Operation: Agile operation helps to transform data into insight and make it more consumable. This includes the execution of lean six sigma. In terms of customer experience, it addresses the need for a system to adapt to changing and uncertain environments. With newer technologies emerging, it is imperative that operations remain open to changes.","excerpt":"The pandemic has been a primary driver in disrupting digital marketing. Today, consumers look for highly-personalised content. Right on cue, businesses are upping their game to keep up with the shift in consumer behaviour. “The need of the hour for marketers is to key out every possible information that is at their disposal and reach […]","categories":["Deep Tech"],"tags":["customer experience"],"author_name":"Shraddha Goled","publish_date":"2021-02-15T17:00:00","publication_year":"2021","word_count":579,"keywords":["API","programming_languages:R","AI","ML","digital transformation","Git","customer experience","analytics","Rust","GAN","R"],"extracted_tech_keywords":["AI","ML","analytics","R","Rust","Git","API","GAN","digital transformation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/cx-at-the-convergence-of-marketing-sales-customer-support-an-inside-look\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043848,"title":"Is Cost-Effective Deep Reinforcement Learning Possible?","content":"“Is there scientific value in conducting empirical research in reinforcement learning when restricting oneself to small- to mid-scale environments?” Can a research done on a smaller computational budget can provide valuable scientific insights? Given the insane training times and budgets, it is natural to wonder if anything worthwhile in AI comes at a small price. So far, the researchers have focused on the training costs of language models which have become too large. But, what about the deep reinforcement learning(RL) algorithms -the brains behind autonomous cars, warehouse robots and even the AI that beat chess grandmasters? Deep RL combines RL with deep learning. Deep RL made a splash back in 2015 when Alphabet’s DeepMind released their work on Deep Q Networks (DQN). When tested on Atari 2600 games, the DQN agent surpassed the performance of all previous algorithms and achieved a level comparable to a professional human games tester. However, according to Google researchers, the advancement of deep RL comes at a cost— a computational one. The original DQN algorithm was tweaked over the years to beat the arcade learning(ALE) benchmark. ALE is widely used as an interface for benchmarking deepRL models on Atari games. The Rainbow algorithm is one such improvement which helped the DQN paradigm to attain state of the art status. However, the Rainbow algorithm is heavy on the computational front. Rainbow was first introduced in 2018. The experiments reportedly required a large research lab set up as it took roughly 5 days to fully train using specialised hardware like the NVIDIA Tesla P100 GPU. According to Google researchers, to prove Rainbow’s superiority, it required approximately 34,200 GPU hours (or 1425 days). Moreover, this cost does not include the hyper-parameter tuning that was necessary to optimise the various components. “Considering that the cost of a Tesla P100 GPU is around $6,000, providing this evidence will take an unreasonably long time as it is prohibitively expensive to have multiple GPUs in a typical academic lab so they can be used in parallel,” according to Google researchers. In their work titled, “Revisiting Rainbow”, the researchers at Google tried to answer the following questions: Would state of the art (ALE) have been possible with smaller-scale experiments unlike in the case of Rainbow back in 2018?How good are these algorithms in non-ALE environments?Is there scientific value in conducting empirical research in reinforcement learning when restricting oneself to small- to mid-scale environments? (Image credits: Google AI blog) To demonstrate the effectiveness of small to mid-scale experiments, the researchers evaluated a set of four classic control environments as shown above. These experiments, according to the researchers, can be fully trained in 10-20 minutes (compared to five days for ALE games): CartPole: Here the agent is tasked to balance a pole on a cart that the agent can move left and right.Acrobot: The agent has to apply force to the joint between the two arms in order to raise the lower arm above a threshold. LunarLander: The agent is meant to land the spaceship between the two flags. MountainCar: The agent must build up momentum between two hills to drive to the top of the rightmost hill. (Image credits: Paper by Castro et al.,) In their experiments, the researchers gradually added double Q-learning, prioritized experience replay, dueling networks, multi-step learning, distributional RL, and other components to the DQN model while at the same time removing components of the Rainbow algorithm. The researchers found that a combination of these components performed on par with DQN that runs on Rainbow. The original loss functions and optimisers were also tested during these experiments. Huber loss and RMS prop optimiser are commonly used while developing DQN models. The researchers also mixed these runs with Adam optimiser and mean squared error loss (MSE). The results show dam+MSE is a superior combination than RMSProp+Huber. The researchers were able to reproduce the results of the original Rainbow paper on a  limited computational budget and even uncover new and interesting phenomena. Hence making a strong case for the relevance and significance of empirical research on small- and medium-scale environments. The researchers believe that these less computationally intensive environments lend themselves well to a more critical and thorough analysis of the performance, behaviors, and intricacies of new algorithms.","excerpt":"The cost of a Tesla P100 GPU is around $6,000, providing this evidence will take an unreasonably long time as it is prohibitively expensive to have multiple GPUs","categories":["AI Features"],"tags":[],"author_name":"Ram Sagar","publish_date":"2021-07-19T13:00:00","publication_year":"2021","word_count":705,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","deep learning","R"],"extracted_tech_keywords":["AI","deep learning","R","Go","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cost-effective-deep-reinforcement-learning\/","complexity_score":4,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10101682,"title":"OpenAI to Sell its Shares for $86 Billion Valuation","content":"Looks like the startup behind the top AI product, ChatGPT, is deciding to raise some money after all. According to a report by Bloomberg, OpenAI is currently in discussions to sell existing employees’ shares at an estimated $86 billion valuation, according to individuals familiar with the matter. The company is in negotiations for this transaction, referred to as a tender offer, with potential investors. These insiders, who requested anonymity due to the confidential nature of the information, have disclosed that the firm has not yet finalised allocations, and the terms of the deal remain subject to change. OpenAI, in which Microsoft Corp. holds a 49% ownership stake, is under the leadership of CEO Sam Altman and President Greg Brockman. If the $86 billion valuation is realised, OpenAI would become one of the world’s most valuable closely held companies, surpassing entities such as Stripe and Chinese online retailer Shein, but still ranking behind SpaceX by Elon Musk and TikTok parent company ByteDance. According to recent buzz around the startup, Altman has said that the company is generating revenue at a pace of $1.3 billion in a year. Moreover, the company is also planning to make its own AI chips. Last month, the Wall Street Journal disclosed discussions about a potential share sale that could value OpenAI between $80 billion and $90 billion. Earlier, we said that OpenAI might go bankrupt by the end of 2024. Given the revenue and reports, it might not be sure, but the company is definitely looking to increase its revenue and it seems like it is on the right track by selling its shares.","excerpt":"Sam Altman had said that the company is generating revenue at a pace of $1.3 billion a year.","categories":["AI News"],"tags":["OpenAI","Sam Altman"],"author_name":"Mohit Pandey","publish_date":"2023-10-19T11:58:48","publication_year":"2023","word_count":268,"keywords":["Go","ChatGPT","Sam Altman","OpenAI","AI","RPA","GPT","llm_models:GPT","R","llm_models:ChatGPT","startup"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","Go","GPT","RPA","startup","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-to-sell-its-shares-for-86-billion-valuation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10081837,"title":"Meet the Man behind Housing.com’s Recommendation System","content":"Housing.com claims to be the country’s largest full-stack digital platform that supports consumers in their home-buying journey. Founded in 2012, it’s one of the largest property platforms in India. To find out more about how its recommendation system works, what challenges they face and how generative AI can help property platforms like Housing.com, Analytics India Magazine reached out to Sangeet Aggarwal, head of products & design, Housing.com. Aggarwal has over ten years of experience in product and technology, and currently leads the product roadmap for monetization, ad tech, customer experience and engagement at Housing.com. AIM: Can you please tell us about the recommendation system of Housing.com? Sangeet Aggarwal: For users, the interface still stays the same, but the recommendations have become much smarter than before. If I were to talk about two or three years earlier, our recommendation systems were extremely simple. The user would just look at a budget and locality and we would find the matching listings within close proximity, with which there was a good chance that multiple users would see the same recommendation. Before we actually had moved to a more heuristic-based model, where essentially, the idea was to try to make some parameters of what the user was searching but not essentially an ML algorithm to back it up. This calendar year, however, we have now moved to a full-fledged ML engine. So it takes into account the thousands of signals, in terms of what other users have done. About 50 to 60 different data points of the other user’s journey and then tries to give you a recommendation. AIM: What are the challenges you’ve faced while building the recommendation system? Sangeet Aggarwal: In terms of challenges, from a user perspective, there are some challenges both internally and externally. For instance, one challenge which poses a problem for any ML recommendation engineer builder is: how do you personalize the UI on the first visit? Once the user has visited the site multiple times, you know something about the user and the recommendations become more personalized but for the first visit, you pretty much have to rely on what other users may have done from the region or some other parameters. Another challenge I would say is what happens offline is something that we have no idea about. For instance, if someone looks up a home on our webpage, and then decides to go and check it out, we have no idea what happens there. The whole perspective changes, the goals change and when you come again to our website, we are still relying on the data from your last visit; but a lot has changed in the meantime. And thus, the user might not prefer the recommended products which it did earlier. I would say, these are two big challenges on the consumer front. AIM: Housing.com claims to have the image auto-auditor and still there are plenty of duplicate images and listings on the site. Sangeet Aggarwal: I would sort of position this as a journey. With any new innovation in the market, it takes two-three years to get it to the level where we start seeing results. Given that real estate is one of the largest industries, the potential for clickbait is always there. What we have recently launched is helping us crack this particular problem on the basis of internal data. We’re trying to identify duplicate listings, fake listings, clickbaits etc with our in-house developed image auto-auditor that is created on an auto-rejection and correction pipeline for various image use cases to achieve over 90% automation in the auto-audit process for flat images. The problem has certainly improved, but yes, it hasn’t been zero. We can’t claim it’s 100% working out there, but it certainly moved in a positive direction for us. We’re planning to invest in the particular area and given the size of the problem, it’ll take at least a couple of years to tackle the problem. AIM: What is your opinion on generative AI? Sangeet Aggarwal: Every week, you wake up and something new comes up in the market. It’s really amazing seeing the exponential growth in the AI industry. So, we are exploring that. I’m pretty sure a user today wants the conversational AI sort of interface, and once this becomes more open source I think we’ll be able to leverage on this trend. AIM: Can it be used in the tech property market? Sangeet Aggarwal: For the people who read all the property details, it’s becoming a long page for them on the platform. As the attention span is reduced and users are more inclined towards the 30 second videos, we were wondering if we can have an engine where it takes a lot of inputs from content around property, descriptions, pictures etc, and then builds a quick video around that. But again, each property is unique and has a unique flavor. We’re really trying to figure out what can be done and it’s a long journey ahead, but again, this can be a possible use case of generative AI in the online property engines.","excerpt":"For users, the interface still stays the same, but the recommendations at Housing.com have become smarter than ever","categories":["AI Features"],"tags":["Generative AI","Interviews and Discussions","recommendation engines"],"author_name":"Lokesh Choudhary","publish_date":"2022-12-08T13:04:45","publication_year":"2022","word_count":847,"keywords":["Go","recommendation engines","AI","ML","recommendation systems","Git","RAG","Aim","analytics","generative AI","Generative AI","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Aim","RAG","recommendation systems","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-the-man-behind-housing-coms-recommendation-system\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10019162,"title":"Guide To 6D Object Pose Estimation Using PoseCNN","content":"PoseCNN(Convolutional Neural Network) is an end to end framework for 6D object pose estimation, It calculates the 3D translation of the object by localizing the mid of the image and predicting its distance from the camera, and the rotation is calculated by relapsing to a quaternion representation. PoseCNN is papered by Yu Xiang, Tanner Schmidt, Venkatraman Narayanan, and Dieter Fox in collaboration with Nvidia research. They also discussed a novel loss function that can help PoseCNN to handle symmetrical objects from images. They created a custom dataset YCB video dataset, which gives 6D poses of 21 objects in 92 videos with almost 133k frames for producing their results. PoseCNN is able to handle symmetrical objects pretty well and can do certain pose estimation using only a single image as an input. Network Architecture The PoseCNN network contains two stages; the first stage is 13 CNN layers and four max-pooling layers, which helps extract feature maps with different input image resolution. The first stage is the primary backbone of the network. The second stage is all about the embedding step that uses high feature maps generated by the first stage into low-dimensional features. After that network performs three different tasks and is trained to do specifically three tasks: Semantic labeling.3D translation estimation.3D rotation regression. 1. Semantic labeling Semantic labeling detects objects in images, where on the other hand network classifies each image input pixels into an object class. In comparison with the 6D pose estimation technique that leverages object detection using a bounding box, semantic labeling gives more information about the objects in the image and can handle occlusions better. It takes two feature maps with dimensions 512 as inputs to the network, as shown in the above figure. The resolution is ⅛ and 1\/16 of the original input image, it first reduces dimensions of the two features to 64 using the CNN layer. Then it doubles the resolution of that 1\/16 feature map by using another deconvolutional layer. After that, another two feature map and deconvolution layer is used to increase the resolution of input by 8x. Finally, the convolutional layer produces a semantic labeling score for image pixels. Remember, in training, a softmax cross-entropy is used, and in testing, the softmax function is used to predict image pixels class. 2. 3D translation estimation 3D translation localize the 2D object center in the image to estimate the object distance from the camera 3. 3D rotation regression The lower part of the above architecture diagram shows the 3D rotation regression method. In this researchers tried to use the Hough voting layer object detection bounding box to predict two RoI pooling layers to crop and pool the feature of the image by generating the first stage of the network for 3D regression. About layers, 3D rotation regression uses pooled feature map by integrating into three fully connected layers. The first two FC layers have dimensions 4096, and the last FC layer have 4 x n (n=number of object classes) Dataset The dataset used for this approach is the YCB dataset, it consists of 80 videos for train, and 2949 key features are extracted from the 12 test videos. Implementation It is trained and tested on Ubuntu 16.04 with PyTorch 0.41+ and CUDA 9.1 Install PyTorchInstall Eigen from Github hereInstall Sophus from Github here git clone https:\/\/github.com\/NVlabs\/PoseCNN-PyTorch.git pip install -r requirement.txt git submodule update --init --recursive ##Compile the new layers under $ROOT\/lib\/layers cd $ROOT\/lib\/layers sudo python setup.py install ##Compile cython cd .. cd $ROOT\/lib\/utils python setup.py build_ext --inplace ##compile the ycb_render in $ROOT\/ycb_render cd .. cd $ROOT\/ycb_render sudo python setup.py develop Download Download 3D models of YCB Objects from here. And Save it under $ROOT\/data.Download pre-trained checkpoints from here and similarly save it under $ROOT\/data. Real-world images with pose annotations for 20 YCB objects can be downloaded from here (53Gb). Running the demo Download 3D models and our pre-trained checkpoints and setup environment.run the following command .\/experiments\/scripts\/demo.sh Train and Test on YCB- dataset First, download the YCB-Video dataset from here and then create a symlink for the YCB-Video dataset using below command: cd $ROOT\/data\/YCB_Video ln -s $ycb_data data Let’s Train and test on the YCB-Video dataset cd $ROOT # multi-gpu training, use 1 GPU or 2 GPUs .\/experiments\/scripts\/ycb_video_train.sh # testing, $GPU_ID can be 0, 1, etc. .\/experiments\/scripts\/ycb_video_test.sh $GPU_ID Conclusion We learned the new method for object pose estimation, PoseCNN decouples the estimation of 3D rotation and translation. It localizes the object center and predicts the center distance of the image. To learn more you can follow given below resources: PoseCNN (GitHub)Research paperThe YCB-Video Dataset ~ 265GThe YCB-Video 3D Models ~ 367MThe YCB-Video Dataset Toolbox (GitHub)","excerpt":"PoseCNN(Convolutional Neural Network) is an end to end framework for 6D object pose estimation, It calculates the 3D translation of the object by localizing the mid of the image and predicting its distance from the camera, and the rotation is calculated by relapsing to a quaternion representation. PoseCNN is papered by Yu Xiang, Tanner Schmidt, […]","categories":["Deep Tech"],"tags":["Computer Vision","Object Detection","pose estimation"],"author_name":"Mohit Maithani","publish_date":"2021-01-28T14:00:00","publication_year":"2021","word_count":772,"keywords":["CUDA","pose estimation","AI","neural network","PyTorch","ML","RAG","Python","Ray","object detection","Object Detection","Computer Vision","R"],"extracted_tech_keywords":["AI","ML","neural network","Ray","PyTorch","RAG","object detection","CUDA","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-6d-object-pose-estimation-using-posecnn\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":24041,"title":"This Is How Indian Scientists Are Using AI, ML To Predict Alzheimer&#8217;s Disease","content":"Combination of two brain diagrams in one for comparison. In the left normal brain, in the right brain of a person with Alzheimer’s disease. Noted Indian scientists from the National Brain Research Centre (NBRC), and Neuroimaging and Neurospectroscopy Laboratory (NINS) are using artificial intelligence to develop a smart system to predict Alzheimer’s disease early. According to a report in Journal of Alzheimer’s Disease, Professor Pravat Mandal are working to develop a model to map metabolic patterns in different brain regions in healthy and pathological conditions. Alzheimer’s is the most common form of dementia, a general term for memory loss and other cognitive abilities serious enough to interfere with daily life. Alzheimer’s disease accounts for 60 to 80 percent of dementia cases. The greatest known risk factor is increasing age, and the majority of people with Alzheimer’s are 65 and older. But Alzheimer’s is not just a disease of old age. Approximately 200,000 Americans under the age of 65 have younger-onset Alzheimer’s disease “Laboratory research and longitudinal clinical studies have helped to reveal various information about the disease but the exact causal process is not known yet. Patterns from alteration of neurochemicals, (for example, glutathione depletion) hippocampal atrophy, and brain effective connectivity loss as well as associated behavioural changes have generated important characteristics features. These imaging-based readouts and neuropsychological outcomes along with supervised clinical review are critical for developing a comprehensive artificial intelligence strategy for early predictive AD diagnosis and therapeutic development,” wrote Dr Mandal. https:\/\/www.youtube.com\/watch?v=rOfpzYkRAc4 According to a report, Dr Mandal, along with a NBRC colleague Deepika Shukla are developing an integrated framework called GAURI. The framework will have statistical and predictive diagnostic capability which in turn could indicate brain chemical changes in the brain. “We will use the data information from a large data set from various diagnosis procedures to create an artificial intelligent system, which would help with the diagnosis of a new unknown case of Alzheimer’s disease using machine learning approaches… Such an integrated multi-modal predictive diagnostic system for Alzheimer’s disease diagnosis would aid the clinician in early differential diagnostics to deliver the most appropriate treatment,” Dr Mandal told the newspaper. A World Health Organisation report says that they hope AI and machine learning can reverse “two decades of failed experimental therapies for Alzheimer’s disease.”","excerpt":"Noted Indian scientists from the National Brain Research Centre (NBRC), and Neuroimaging and Neurospectroscopy Laboratory (NINS) are using artificial intelligence to develop a smart system to predict Alzheimer’s disease early. According to a report in Journal of Alzheimer’s Disease, Professor Pravat Mandal are working to develop a model to map metabolic patterns in different brain […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","ai in health sector","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2018-04-26T10:01:41","publication_year":"2018","word_count":378,"keywords":["API","artificial intelligence","machine learning","programming_languages:R","AI","Modal","Machine Learning","Git","ai in health sector","ViT","GAN","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","R","Git","API","GAN","ViT","Modal","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/alzheimers-disease-ai-ml-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10004697,"title":"Why Location Analytics Is So Critical For 3rd Party Logistics Providers","content":"For a 3rd Party Logistics vendor, Location Analytics becomes key to making decisions in strategy, city operations as well as business development. This article will help you realise how you can use location analytics to maximise resource utilization, optimize RoI and minimise user churn by fulfilling your SLAs. Data Digestion and Hyperlocal Ops With the growth of E-commerce, the logistic market has grown by leaps and bounds. In terms of revenue, the global logistics market is set to touch a CAGR of 3.48% from 2016 to 2022 and would capture a market valuing $12,256 billion by 2022. What does this mean? The three verticals that we just discussed have one thing in common- location data. Why is it important for a logistics business to know where certain events occur. With businesses trying to go hyperlocal for their strategy, marketing and operations management, it has become important to understand how you are performing in areas as granular as neighborhoods. This knowledge is leveraged by companies to break down their problems into smaller geographical units and administer them individually. This includes identifying local problems, solving them via curated marketing and tracking your progress on the same. What the Current Status Quo looks like: Cost\/Trip: Logistics companies today have very less visibility on the total cost of shipping: the base charge and the surcharges per journey along with externalities such as extra costs per return journey, losses incurred due to cancellations etc. They end up paying overheads in different locations for different partners.Visibility: Logistics companies find out that one of their major pain points while expanding into a new area or breaching a new market is ensuring user acquisition. Knowing the user behaviour on a granular level allows you to shift your marketing to a hyper-local level.Data Diet: For most companies, their data split among many different systems like UPS\/FedEx\/Courier Billing, Label Generation Software, Tracking Systems etc. This makes it very complicated for companies to crunch data and get insights on all their different verticals in operations. An Overview of Use cases for location analytics in logistics Leveraging location analytics can help improve the overall efficiency and performance of your logistic service. Knowing where and why certain events happen can help you take desired action on them. With the use of analytics you can optimise your route\/fleet utilisation, improve your shipping\/order accuracy and bring down the percent of time\/resources spent in reverse logistics. City and Central Operations: For any ops team some of the major areas of concern revolve around monitoring the fleet, minimizing delays, spotting patterns in anomalies, and ensuring that SLAs are met. Delays and SLAs: Measuring SLA (Service Level Agreements) compliance provides organizations quantifiable metrics to measure the performance accuracy in terms of how specific services are performed. In logistics, the SLA agreements have any of the two counterparts: carrier, fulfillment center merchant, or customer. Analyzing delays across: Routes: Which regions\/routes\/delivery partners have a historical pattern of strange anomalies indicating frauds leading to a violation of the SLA? Can alternates be figured out? Which freight or geographical region or parcel type takes the maximum time for delivery? Which laps are causing delays in the entire delivery process- warehouse, first mile or last mile? What are the causes for delays at every warehouse- dispatching, packaging, or assignments?Deviations: Which areas and routes have shown a history of anomalous delays because the delivery partner deviated from the suggested route? Having a centralized view of all your fleet metrics gives you a comprehensive view of your assets to ensure success. Proper fleet management helps in better risk assessment and optimization to ensure higher shipment and order accuracy. Routine Checks: Routine checks and measuring business health becomes very critical from time to time to become proactive with strategies. On the other hand, staying on top of your business with real-time insights to catch on anomalies also is very critical. Uptime: Which are your serviceable hours\/day? Most logistics companies tend to have a high uptime (~99.99 % ) or availability round the clock to cater to a large number of shipments. After how many kilometers does the fleet require inspection? How many days does it remain unserviceable?Real Time Activity: Which are the locations where loading and unloading happens? Monitoring your fleet in real time just adds onto the greater responsibilities that result in a more efficient and healthy fleet, while allowing you to monitor accidents in real-time.Partner Profiles: What is the performance of your delivery partners? Can you segment them so as to maintain a higher level of customer satisfaction? Analysing which route the fleet takes and the hours of the day they approach a delivery destination can help you cut down the time lost in roadblocks and traffic jams. Strategy and CXOs Strategy teams often spend a lot of their time trying to win the battle in these verticals- maximising resource utilisation and understanding the cost of shipping. Taking decisions to optimize fleet operation begins and ends with tracking fleet performance metrics. They need to look at patterns in bottlenecks and take business decisions that have large impact. Cost of Shipping What’s most important is understanding pure shipment data to better understand what’s the total true cost of shipping to different parts of the world. Once you know your true costs, you can market more deeply into profitable regions as well as recost your shipping fees to areas where you’re losing money.  Moreover, which areas and routes cause extra costs to be levied upon the base shipping costs for a journey? Warehouse and Inventory Just like any other metrics, your warehouses can tell you where you are excelling and where you need optimization or improvements. Understanding how often shipments come from multiple warehouses and how much inventory is accumulated at each of those centers is important. Which products have the highest shelf life? Which products need to be dispatched at the earliest? How far are your warehouses and fulfillment centres located from high demand areas? Resource Optimization What are the number of the resources ( freights\/staffs) that are engaged in reverse logistics and is there any way to reduce this number by batching returns with deliveries? Which products need to be stocked for unforeseen events like disease outbreaks\/natural calamities? What is your inventory to sales ratio? Tracking this metric can help you predict increasing inventory levels against dropping sales rate. Customer Success Customer support data helps to understand where late shipment anxiety is occurring, where people have the most problems with the product, where people are losing shipments most, etc. These can be determined by tags on the customer support ticket and other attributes. Logistics companies can focus on those locations followed by negative customer reviews which bear its brunt on the overall ratings of the logistics company or seller. Latent Demand Which regions show a high demand for any specific products? Which rural areas need to be reached? Knowing which areas have latent demand i.e., your users are needed for logistics but the absence of a proper supply medium breaks the chain- will help you plan the areas where you need to expand to. Geo Targeted Campaigns Order and sales patterns mostly rely on user buying behavior and hence understanding the trends shown by the users across locations is important for real-time action as well as future prediction.Which areas have low user demand and targeted ads need to be launched? How does the demand for certain products vary with respect to time (around festivals, weekends etc)? Returns and Cancellations Reverse logistics is a nightmare for organizations dealing with logistics and shipping. The returned packages directly inflate the operational expenditure without generating any return on investment. How does your return\/cancellation rate look over time? Do they have a seasonal or demographic pattern? What is the geographical spread of losses due to returns? Which routes and areas show the highest losses due to returns- in returning as well as refunding? What are the main causes of returns- poor inventory management or incorrect delivery addresses? Same issues exist with the cancellations of shipments as well, especially cancellations after the goods are shipped. Why Locale? Currently a lot of companies try to build these products internally. These companies have to create teams of Data Scientists who work for weeks to build a single dashboard-this leads to a lot of time and effort being devoted to each dashboard. On top of that, these dashboards often end up being non-scalable. As a result, companies are always on the lookout for products that can help their business teams leverage location data and make decisions without having to rely on the Engineering teams. What do we do that sets us apart from any other options? We integrate into your system in less than 24 hours!We can handle tons and tons of data- above 55 million pings in productionYou can make decisions on any granular level- leveraging data both real-time, and historically!Our consoles are scalable as well as customisable- you measure what you want, when you want and make your decisions for any level of administration. So if you are a 3PL executive, if you want to measure your business’ performance on a granular level, then you need to leverage your location data through a tool that allows you to use a microscope. We, at Locale are your microscope- you can make the decisions, while we do the crunching!","excerpt":"For a 3rd Party Logistics vendor, Location Analytics becomes key to making decisions in strategy, city operations as well as business development. This article will help you realise how you can use location analytics to maximise resource utilization, optimize RoI and minimise user churn by fulfilling your SLAs. Data Digestion and Hyperlocal Ops With the […]","categories":["AI Features"],"tags":["data management providers"],"author_name":"Aditi Sinha","publish_date":"2020-08-12T15:00:13","publication_year":"2020","word_count":1556,"keywords":["Go","funding","AI","RPA","Scala","RAG","ViT","analytics","GAN","data management providers","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Scala","GAN","ViT","RPA","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-location-analytics-is-so-critical-for-3rd-party-logistics-providers\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10013390,"title":"Scikit-Learn Is Still Rocking, Been Introduced To French President","content":"A milestone for open source projects — French President Emmanuel Macron has recently been introduced to Scikit-learn. In fact, in a recent tweet, Scikit-learn creator and Inria tenured research director, Gael Varoquaux announced the presentation of Scikit-Learn, with applications of machine learning in digital health, to the president of France. He stated the advancement of this free software machine learning library — “started from the grassroots, built by a community, we are powering digital revolutions, adding transparency and independence.” Today, I presented @scikit_learn, with applications to digital health, to @EmmanuelMacron, the French president.It is beautiful to see how far we have come: started from grassroots, built by a community, we are powering digital revolutions, adding transparency and independence. https:\/\/t.co\/C7QTU7yMrg pic.twitter.com\/9N4QAX8MQo— Gael Varoquaux @GaelVaroquaux.bsky.social (@GaelVaroquaux) December 4, 2020 Varoquaux presented a project called ScikitEDS, which is a visualisation tool to monitor the patient flow that was developed during France’s first hit of the COVID pandemic wave. Launched earlier this year, ScikitEDS was developed by Inria in partnership with the startup, AP-HP (Assistance Publique – Hôpitaux de Paris) to enable permanent monitoring and analysis of the flow of COVID-19 patients in one of its 39 hospitals. This allowed the company to monitor data of more than 100,000 patients on a daily basis, who are registered in its EDS — health data warehouse. According to the news media, the massive database, EDS-COVID, contains pseudonymised data of all patients who have had a PCR test within the AP-HP. This would contain not only the patient journey but also their biology, medical history, demography, medical reports, imaging, etc. With the lacking data science expertise that can help in the statistics and visualisation of such a massive database, AP-HP decided to partner with engineers from the Scikit-Learn consortium and scientists from Parietal teams — INRIA’s Neurospin research teams to support the crisis. #ScikitEDS présenté par @GaelVaroquaux est un outil de visualisation des flux de patients développé au plus fort de la 1ère vague #Covid19, fruit d'un partenariat @APHP @Inria et développé notamment par des scientifiques de l'équipe @Parietal_INRIA. ➕ https:\/\/t.co\/KAHRF3VhqQ https:\/\/t.co\/AmHsUszwNa— Centre Inria de Saclay (@Inria_Saclay) December 4, 2020 To facilitate this, the team is working on developing a software stack that can integrate the web dashboard for visualising the data from EDS-COVID database. Further, it comes with a synthetic table showcasing more than 200 descriptive variables of each patient. It has been automatically generated from the daily hospital database extractions for the management and nursing staff of AP-HP. Replying to his own tweet, Varoquaux stated that open-source Python library — SciPy; PyData stack; and Python were the core of the project. Merci à tous les collègues qui ont illustré la vitalité de la recherche et de l'innovation en santé numérique, notamment @GaelVaroquaux\/@scikit_learn et les travaux avec @APHP, la startup @TheraPanacea et le projet DAICAP lauréat du défi IA&Santé du @HealthDataHub https:\/\/t.co\/M5IAYopzGA— Bruno Sportisse (@bsportisse) December 4, 2020 In another tweet, Chairman and CEO of Inria, Bruno Sportisse stated his gratification by thanking the team, including Gael Varoquaux, Scikit-Learn engineers, AP-HP, and others, for their “vitality of research and innovation in digital health.” With this, the team aims to contribute to better healthcare in these turbulent times. From being just another open-source project to achieving a milestone of being presented to Emmanuel Macron for advancing digital health in France, Scikit-learn has come a long way.","excerpt":"A milestone for open source projects — French President Emmanuel Macron has recently been introduced to Scikit-learn. In fact, in a recent tweet, Scikit-learn creator and Inria tenured research director, Gael Varoquaux announced the presentation of Scikit-Learn, with applications of machine learning in digital health, to the president of France. He stated the advancement of […]","categories":["AI News"],"tags":["easy python beginner projects","fun beginner python projects","scikit learn"],"author_name":"Sejuti Das","publish_date":"2020-12-07T13:08:39","publication_year":"2020","word_count":559,"keywords":["data science","scikit-learn","machine learning","AI","Git","Python","Aim","ViT","fun beginner python projects","easy python beginner projects","scikit learn","R","data warehouse"],"extracted_tech_keywords":["AI","machine learning","data science","Aim","scikit-learn","Python","R","Git","data warehouse","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/scikit-learn-is-still-rocking-been-introduced-to-french-president\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":69767,"title":"Hardware startup Smatron strengthens its IoT portfolio with investment in Volta Motors","content":"Smatron expands its IoT product portfolio with investment in Volta Motors Mahesh Lingareddy, founder and chairman of Hyderabad based computer hardware startup Smartron has his eyes firmly trained on the IoT market. The hardware start-up known for its t.phone, the world’s first IoT ready smartphone and the t.book that packs all the in-demand features has added another product to its kitty. Lingareddy is a former Intel veteran who later co-founded Soft Machines, US based Chip Design Company in 2006 and reportedly sold it to Intel for $300 million. His latest investment comes in the form of Chennai-headquartered automotive company Volta Motors, known for its homegrown product – an eco-friendly cross-over electric bike Volta Zap. Industry pundits have dubbed it as a positive move – with “one Indian startup collaborating with another startup”.  The electric bike, Volta Zap, boasts of many impressive features, most notably a digital display, LED Safety lights and a Volta connected app. Though there is no information about the amount invested, Volta will reportedly use the funds for scaling up the development and manufacturing of Volta Zap. Smartron investment in Volta Motors — step in the right direction Since its birth in 2014, Smartron has billed itself as an IoT company, dedicated to building an IoT ecosystem across hardware, software and innovation-driven products. The undisclosed investment in Volta Motor also comes with a name change, the Chennai-based startup will be rechristened to Tron Motors, and will crank out IoT-enabled electric vehicles designed and engineered for a global market. According to a statement, “the strategic investment is part of Smartron’s approach to build broader product ecosystem under AI-powered tronx IoT platform.” Smartron will extend the tronx platform across a range of verticals, from personalized health, farming, smart home to smart infrastructure and will also usher in the next generation of smart devices, sensors, services and care to consumer, enterprise, industrial and infrastructure markets. Mahesh Lingareddy, Smatron Founder and Chairman Speaking about the investment, Lingareddy shared in a statement, “Our association with Volta emphasizes the strategy that Smartron has pursued from the start, that we need to be known as company that is building a strong product ecosystem across IoT verticals from personal to health to home to energy to farming. Electric ecofriendly bikes powered by an IoT platform like Smartron’s TRONX will be an essential part of our range of smart products”. As the startup prepares to line up a slew of IoT devices, it has an IoT app store in place and is reportedly mulling an IoT hub to connect smart devices at homes and offices. If news reports are anything to go by, the company already has two R&D centres in Hyderabad and Bangalore and had ambitious plans of notching up employee strength to 500. The company is betting big on IoT and Artificial Intelligence as the next big wave of huge disruptive opportunity. Smartron’s core offering is tronx™ — the AI powered IoT platform that powers Smartron high-end tbook and tphone products and enables users easy access to the tstore, tcloud (unlimited storage) and tcare. From Tronx 1.0 to Tronx 2.0 According to news reports, Smartron will soon debut a revamped Tronx platform.  With the overwhelmingly positive response from the market, the startup will soon release an updated version of its platform with added features.  The founder-Chairman believes the software agnostic Tronx platform can power a range of verticals, from health to infrastructure, education and farming as well.","excerpt":"Mahesh Lingareddy, founder and chairman of Hyderabad based computer hardware startup Smartron has his eyes firmly trained on the IoT market. The hardware start-up known for its t.phone, the world’s first IoT ready smartphone and the t.book that packs all the in-demand features has added another product to its kitty. Lingareddy is a former Intel […]","categories":["AI Startups"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-02-24T09:27:00","publication_year":"2017","word_count":573,"keywords":["Go","artificial intelligence","programming_languages:R","AI","innovation","ML","Git","RAG","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","ML","RAG","R","Go","Git","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/hardware-startup-smatron-strengthens-iot-portfolio-investment-volta-motors\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39480,"title":"AIM Completes 7 Years, See What Lies Ahead For Us","content":"It’s been 7 years since we first started Analytics India Magazine, India’s largest analytics community.  With more than 1.5 million users, AIM has evolved into the #1 platform with a thriving community of developers, data scientists and business decision makers who tap into AIM for insights. A big shout out to our growing community and all the tech-savvy people that has helped us grow exponentially! In fact, our community, which is our backbone, has been able to get a head start on the latest techniques, technologies and the skills they need today through our information-packed conferences, meetups and content specially tailored for the data science community. If data fuels the future, AIM is at the forefront of it, exploring the latest techniques and opportunities for the growing community. “In the last 7 years, we have seen the analytics industry and ecosystem evolve and mature. AIM has played a pivotal role in addressing the gap in the developer and business community by answering critical questions. What are people doing with data and how it can be more valuable for organisations, startups, professionals and especially at an individual level for beginners seeking a new career in this domain,” says Bhasker Gupta, Founder & CEO, Analytics India Magazine. As part of our 7 years celebration, we are launching a brand new web series – The Pretentious Geek.  This is a weekly episode that talks about the latest technologies in a quirky fashion. Check out the first episode on our channel this week. Check out some of the highlights of our evolution: Relevant Content: In the last few years, to keep up with the demand for learning-related relevant content, AIM has created a wonderful repository of content geared at developer ecosystem in India. Be it the latest research, application or tools, if it’s driving the world of data, you’ll hear about it at AIM first. Cypher & MLDS: Today, AIM organises 4 conferences a year around India and Singapore, including the flagship analytics conference Cypher in Bangalore and Gurgaon. Our upcoming conference, Machinecon in Mumbai and Singapore will see a gathering of business leaders and IT decision makers. Our first edition of developer-focused conference MLDS was a huge hit and met the demands of the growing developer base in India, hungry for the latest solutions in AI and advanced analytics space. The broad appeal for MLDS  & Cypher conference stems from relevant content and hands-on workshops which are primarily tailored for a technical audience . we continue to provide stellar content with a keen eye on business applications and use cases. MachineCon: On the other hand, MachineCon to be held on May 24 in Mumbai and May 31 in Singapore seeks to deliver business value to IT decision makers and helps them navigate the challenge of converting technology investments into real business value. Business decision makers seek insights on how to build data-driven strategies and drive efficiency and innovation. At the same time, IT leaders look for information on how to bridge the gap between data and tech agenda. Meetups & Hackathon: As we envision a continuing demand for analytics and AI, our plan is to keep providing innovative channels and content to help the community expand and focus on the disruptive technologies. MachineHack now in its 13 edition hosts some of the toughest data science challenges. Meanwhile, our meetups, now in its 6th edition covers some of the hottest topics with an eye on relevant business use cases. A Glimpse Of Our Future Seven years since the launch, interest in advanced analytics and machine learning and the demand for more knowledge in these areas continues to grow. We believe as the market continues to evolve, AIM will keep a pulse on the changing trends and bring to the fore the latest happenings in disruptive data technologies and help you stay ahead of the curve. Don’t forget to check out the latest contests on  MachineHack and sign up for our newsletter to stay tuned.  Once again, a big thank you for all the support to our growing tech-savvy community and all the alpha geeks out there!","excerpt":"It’s been 7 years since we first started Analytics India Magazine, India’s largest analytics community.  With more than 1.5 million users, AIM has evolved into the #1 platform with a thriving community of developers, data scientists and business decision makers who tap into AIM for insights. A big shout out to our growing community and […]","categories":["AI Features"],"tags":["Analytics India Magazine","Cypher","Intel","MachineCon","Machinehack Hackathon","Microsoft","MLDS"],"author_name":"Richa Bhatia","publish_date":"2019-05-21T10:52:14","publication_year":"2019","word_count":682,"keywords":["data science","machine learning","Microsoft","AI","data-driven","ML","innovation","Analytics India Magazine","MLDS","Aim","MachineCon","analytics","Machinehack Hackathon","GAN","R","Intel","Cypher"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","R","GAN","data-driven","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/aim-completes-7-years-see-what-lies-ahead-for-us\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10068357,"title":"Facebook AI Research goes through massive restructuring","content":"Meta has announced a new decentralised organisational structure for Meta AI to better leverage the newest AI technology at scale. “More centralised approaches run into their limits when the last mile proves to be too far for downstream teams to close the gap,”said Andrew Bosworth, CTO, Meta. In the new model, Meta will distribute the ownership of various teams within Meta AI like AI Platform, AI for Product, AI4R  to Meta’s product groups. These teams will be assigned the task of driving AI advancements and best practices into the products they support. They will be known as  AI Innovation Centers. Key announcements The AI research team, FAIR, that propels fundamental breakthroughs in AI will become an integral part of  Reality Labs Research.FAIR now stands for “Fundamental AI Research”.The AI for Product teams that work to protect the people using our platforms, improve recommendations and make content more relevant, and improve our Ads and Commerce services will move to our product engineering team.The AI4AR team will join with the XR team in Reality Labs.The responsible AI team will be join the Social Impact teamA new cross-functional AI team will be convened. It will evaluate the progress of the entire company in AI. Meta is confident that the new structure will accelerate the adoption of important new technology across the company. “With this new team structure, we are excited to push the boundaries of what AI can do and use it to create new features and products for billions of people,” Bosworth said.","excerpt":"The AI research team, FAIR, that propels fundamental breakthroughs in AI will become an integral part of Reality Labs Research.","categories":["AI News"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-06-03T15:33:45","publication_year":"2022","word_count":251,"keywords":["Meta AI","programming_languages:R","AI","innovation","RAG","responsible AI","GAN","AI research","R"],"extracted_tech_keywords":["AI","Meta AI","RAG","R","GAN","responsible AI","innovation","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/facebook-ai-research-goes-through-massive-restructuring\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119259,"title":"10 New AI Tools in 2024 to Accelerate Your Workflow","content":"Whether you’re a content creator, developer, or entrepreneur, these AI-powered websites offer solutions to streamline your workflow and accelerate project completion. From video editing to book writing and chatbot development to content generation, these platforms harness the latest AI tech to supercharge your productivity. Here are 10 AI websites that will help you finish months of work in just a day. 1. Framedrop AI Framedrop is an AI tool that can process your YouTube and Twitch streams and find all of your best moments automatically to create short-form videos, such as TikToks, YouTube Shorts, and Instagram Reels. This eliminates the need for creators to spend long hours manually editing their content by offering a wide array of features to repurpose their content and reach new audiences. It accommodates all types of conversational content from podcasts to interviews and vlogs to stand-up comedy, and currently supports 25 languages. Also, you don’t need to download any software to use it. Framedrop enables creators to process up to five hours of content each month with no signup required. The free signup increases this allowance to 8 hours per month. And, creators that subscribe to the Pro Plan get 40 hours a month and additional benefits like multi-video processing and priority queue. Turn any video into shorts with Framedrop AI 2. PopAi PopAi is like your personal AI workspace that lets you talk to your images and documents, providing instant answers, summaries, and insights. It can accommodate many document formats, including PDFs, Word files, and text documents, and supports over 200 languages. One can think of it as a creative partner that takes on tasks such as completing assignments,  retrieving information, generating images, optimising resumes and CVs, crafting compelling copy, devouring books, designing flowcharts, and creating dynamic PowerPoint presentations. It can also write SEO blogs and academic essays as well as debug code. New Feature:Convert PDFs to slidesUpload files on homepage, then you can ask PopAi to create the PowerPoint presentations you need. This will significantly enhance your productivity.You can experience this feature in the following scenarios:Business:-Upload a company… pic.twitter.com\/O1G1lyDC3x— PopAi (@popaiinone) March 5, 2024 3. Numerous AI Numerous.ai offers a simple, powerful, and cost-effective solution for using ChatGPT inside Google Sheets and Excel. You can use it to automate tasks like categorising, extracting text, and cleaning data. It’s also great at summarising, rewriting, and translating content. It can research and organise digital marketing campaigns by generating AdWords keywords, ad copy, etc; can summarise, categorise, classify & cleanse large bodies of open-ended text such as articles, user surveys, or social media content. It can also generate topic ideas, schedule social media posts, and create varied content for different audiences from a single prompt. It also allows you to experiment with various prompt structures, prototype AI features, and collaborate on GPT input and outputs. Excel has 1.2 billion active users.But 99% people don't know about this secret.Here is how to use ChatGPT inside Microsoft Excel or Google Sheets: pic.twitter.com\/K0HpwLR3bI— Hasan Toor ✪ (@hasantoxr) April 7, 2024 4. Murf AI Murf.ai lets you go from text to speech with a versatile AI voice generator that allows you to make studio-quality voice-overs in minutes. You can use its lifelike AI voices for podcasts, videos, and professional presentations. Be it creative, corporate, or entertainment purposes, the platform has a voice for your every need. It lets you choose from over 120+ text to speech voices in 20+ languages. Using it, you can upload your creatives and sync it with the voice of your choice. Murf also allows you to play with pitch, punctuations and emphasis to make the AI voices carry your message as you like. 5. Stockimg AI Stockimg.ai is an excellent design tool that lets you generate everything from illustration, stock image, art, logo, wallpaper, poster, book cover to QR code. It enhances your design process, helping you save your time and money. With Stockimg AI as your designing agent, you can simplify content creation by generating unique images using AI. 6. Momen Momen lets you design, develop, and scale your apps without code so that you can build your custom apps from frontend to backend with a simple drag-and-drop editor. This powerful no-code tool revolutionises web application development, allowing entrepreneurs to launch their products faster and more affordably. Its fully visual interface allows users of any technical level to build dynamic, responsive applications with professional design quality; pixel-perfect and fully responsive. Momen takes care of hosting and provides robust and scalable infrastructure, while also offering enterprise-grade features and security to support business growth. Start your no-code journey with Momen. 7. Bookwiz Bookwiz is an AI tool that helps you write and publish quality books fast. The platform leverages GPT-4 to assist authors in the book-writing process, including brainstorming ideas, creating characters, outlines, and chapters. It offers a user interface designed specifically for book writing, incorporating an advanced editor and structured guidance. The platform also fosters a writing community, facilitating easy sharing and feedback on your work. Using Midjourney for image creation, it enables you to add beautiful visuals to your pages, chapters, and covers. Also, you can export your book to PDF, EPUB, or MOBI, and publish to Amazon, Apple Books, and more. Write your book with Bookwiz. 8. Chatsimple With Chatsimple, an AI-powered chatbot builder, anyone can create customised chatbots for their business website in just minutes, with no coding required. It uses advanced AI like GPT-4 to power chatbots and is trained on a business’s website content and data to understand the business. Using it, you can get a chatbot customised to your business and can talk to customers 24\/7 in over 175 languages. It can help you change the way your customers interact with your website by providing you with a superior sales agent skilled in engaging conversations, booking meetings, lead qualification, and follow-up & closure. Use cases include customer support, lead generation, and market research among others. It can be embedded on your website, Facebook, Instagram and more. An API is available for custom integrations. Also, you can create unlimited chatbots based on your plan. 9. Submagic Submagic enables you to make your short-form videos more captivating with captions, B-rolls, zooms and sound effects. From generating video captions for your video in over 48 languages, adding music and sound effects to create an immersive experience, enhancing your videos with strategic zooms and automatic cuts to enhancing your narrative with stock videos and transitions, it helps you easily boost your reach and engagement with the power of AI. 10. RecCloud RecCloud is your go-to AI tool for all-in-one video editing solutions including subtitles generator, speech-to-text or text-to-speech, AI vocal remover, trim\/crop\/merge videos, among others. It offers professional recording and editing API services, complemented by comprehensive API documentation and straightforward integration procedures. Its services cater to diverse industries including online education, live streaming, conferences, gaming, and more.","excerpt":"Enhance your productivity by 100X","categories":["AI Trends"],"tags":["AI Chatbot","ChatGPT","excel","GPT4"],"author_name":"Sukriti Gupta","publish_date":"2024-04-30T12:46:53","publication_year":"2024","word_count":1144,"keywords":["Go","ChatGPT","excel","TPU","AI","AI Chatbot","chatbots","ML","Scala","RAG","Ray","R","GPT4"],"extracted_tech_keywords":["AI","ML","ChatGPT","Ray","RAG","chatbots","TPU","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-ai-tools-to-accelerate-your-workflow\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10010864,"title":"Why On-Premise And Cloud Solutions Fall Short When It Comes To Open Source Software","content":"Rome was not built in a day, nor were the data centres. Picking the real estate, power and cooling plant installations, server installations take months and not to forget how expensive the whole ordeal is. The time duration increases whenever an organisation decides to upscale. So, for companies who do not want to burden themselves with the woes of building a data centre, on-prem cannot be an option. That said, the cloud is not without its own hassles. Even migration is a tricky, tedious process for many organisations. When it comes to open-source software solutions, both on-premise and cloud platforms have their fair share of challenges. Companies like Google, with its diverse big data solutions, have been trying to address challenges on both ends. In the next section, we take a look at pressing issues concerning big data applications, according to Google Cloud. On-Prem & Cloud Challenges Companies might find themselves locked into a certain cloud provider. Vendor lock-in can become an issue in cloud computing because it is very difficult to move databases once they’re set up, especially in a cloud migration, which involves moving data to a totally different type of environment and may involve reformatting the data. Vendor lock-in is a situation where the cost of switching to a different vendor is so high that the customer chooses to stick to the original vendor. Vendor lock-in is only a part of the problem. There are quite a few exclusive to both on-premise and cloud storage and computing. Configuration & Constraint Management Although the application developers can take advantage of on-prem storage by exploiting the underlying physical environment, they still came with few challenges. Making changes to hardware configuration can be disruptive as most of the open-source software depends on standardisation. Whereas, constraint management is all about figuring out the right way to optimise resources like power and floor space at the data centres for maximum utilisation optimisation. AWS Snowmobile at AWS re:Invent 2016 Relocation Data migration to a network is expensive and time-consuming. To avoid the cost and effort of relocating the data and applications, users sometimes even resort to manually migrating the hardware by road. For example, Amazon’s Snowmobile is a 45-foot long ruggedized shipping container, pulled by a semi-trailer truck that offers an exabyte-scale data transfer service with transfers up to 100PB per vehicle. Where on-premise platforms struggle, cloud thrives. Cloud computing enabled on-demand scaling by allowing data developers to select custom environments for their processing needs, allowing them to focus more on their data applications and less on the underlying infrastructure. As workloads evolve over time, the need for managing service level objectives (SLOs) or the performance that was promised by the service provider. Spike in data should be handled independently without breaking down the data pipeline. Although the cloud eliminates the need for logistics planning for the data centre, says Google, the complex task of cluster configuration continues to be a challenge. For cloud users, optimizing the processing environments to understand workload characteristics is still a challenge. Ushering A Serverless Future Despite the innovations that Google and other top cloud providers have engineered over the years, the challenges still persist. Google too knows that. Google Cloud’s Big Query and Dataproc are designed to empower OSS platforms while also offering a doorway to a serverless future. “Serverless is not new to Google. We have been developing our serverless capabilities for years and even launched BigQuery, the first serverless data warehouse,” said Susheel Kaushik, Product Manager at Google Cloud. GCP’s Dataproc, for instance, is capable of complementing the likes of OSS platforms like Apache and Presto. Companies like Facebook, which deal with petabytes of data, rely on platforms like Presto. Twitter too, was leveraging Presto until it decided to migrate to Google Cloud. With the Dataproc platform, users can manage, analyze and take full advantage of data and the OSS systems already in use. Apache is no stranger to the changing times. It has a serverless offering of its own called OpenWhisk. Apache OpenWhisk is an open-source, distributed Serverless platform that executes functions in response to events at any scale. OpenWhisk manages the infrastructure, servers and scaling using Docker containers so you can focus on building amazing and efficient applications. With the advantages of data analytics becoming obvious, we can expect a sporadic growth of serverless offerings. In the serverless world, customers can focus on their workloads instead of infrastructure. The configuration is automatic. “It’s time for OSS to have its turn. This [serverless] next phase of big data OSS will help our customers accelerate time to market, automate optimizations for latency and cost, and reduce investments in the application development cycle so that they can focus more on building and less on maintaining,” promises Google.","excerpt":"Rome was not built in a day, nor were the data centres. Picking the real estate, power and cooling plant installations, server installations take months and not to forget how expensive the whole ordeal is. The time duration increases whenever an organisation decides to upscale. So, for companies who do not want to burden themselves […]","categories":["AI Features"],"tags":["distributed graph database","Google Cloud","open-source software"],"author_name":"Ram Sagar","publish_date":"2020-10-28T12:00:49","publication_year":"2020","word_count":791,"keywords":["Go","Google Cloud","GCP","AWS","AI","cloud computing","open-source software","docker","serverless","RAG","distributed graph database","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","cloud computing","AWS","GCP","docker","serverless","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/open-source-software-google-cloud-dataproc-serverless-apache\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":19578,"title":"Want To Get Started With Learning Artificial Intelligence? Check Out This Learning Pathway","content":"The balance of power in the tech landscape is shifting towards artificial intelligence with IT bellwethers baking AI into existing products. But what’s more crucial to the future of businesses is an AI workforce. And this starts with training the brains behind the business and setting up online program to help employees explore new roles. While many programmers can code, they are not yet versed in machine learning. Despite the hype in the IT industry, and startups working in this space, developers or freshers aren’t sure how to get started in the field of AI. Entrepreneur and product enthusiast Shival Gupta gave an interesting perspective on this – the relevance of a full stack developer will not be enough in the changing scenario and in the next two years, full stack will not be full stack without AI skills. Analytics India Magazine curates the best way to get started with learning AI. Getting Started AI – The Most Important Skill Of The 21st Century 1.Prime Yourself For AI With These Free Books Shival Gupta shared his experience of getting started with learning AI and emphasized how it is important to familiarize oneself with basic AI terms and approaches. This is true and a good way to prime yourself for AI would be starting with some free books. Peter Norvig and Stuart J. Russell’s Artificial Intelligence: A Modern Approach. The book deals with not just the basic AI concepts and algorithms (expert systems, depth-first and breadth-first search, knowledge representation, etc.) but also the fundamental of mathematics such as Bayesian Reasoning, First Order Logic, NL n-grams et al.  For those interested in Deep Learning, here’s a book by Ian Goodfellow, Yoshua Bengio and Aaron Courville. Also, check out this free book on Logic For Computer Science which explains the mathematical logic for understanding computer science and emphasizes algorithmic methods for solving proofs. Here’s a books specifically aimed at senior undergraduate students explains the fundamentals of AI with a framework to study the design of intelligent computational agents. 2. Spruce Up The Required Math Since both Calculus and Linear Algebra have a wide application in AI\/ML techniques, it will be a good idea to learn it. AI enthusiasts argue that most of machine learning techniques can be reduced to Linear Algebra, and Calculus, such as the backpropagation algorithm for training neural networks. Also, take a deep dive in Discrete Math, Calculus (differential, integral, and multivariate), Probability and Statistics, Linear Algebra, Regression Analysis, and Stochastic Processes (Poisson processes, Markov chains, and Brownian Motion) for a career in AI\/ML. You can check out these free learning resources on the web for Probabilistic Theory, Introduction to Statistical Learning With R & on Inference & Learning Algorithms. 3. Get acquainted with Python, (C\/C++) & data structures AI practitioners believe that one can do AI\/ML in any mainstream language and also non-mainstream languages. The biggest difference lies in the performance and availability of libraries \/ tools. For example, C++,is all set to outperform Java or Python and allows the developer to maximize the capabilities of the hardware. On the other hand, Python has a really good FFI, and is often used in conjunction with C or C++. Meanwhile, Octave\/MATLAB, R, Python, C++, Java, R and a few other languages features high quality libraries, which is equally important to you depending on what you want to do. The general consensus is that one must stick to something popular, like Python that has a great toolkit\/libraries. 4.There Are Many Open-source Frameworks To Start Experimenting With Post this, one can pick a framework and implement it for basic classifications. According to developer Akash Paul, picking a framework can be a challenging task because they are all built with different purposes in mind. He cites an example: Caffe uses a declarative approach to define models in it whereas with TensorFlow, one can programmatically create and use models, even visualize and deploy them with ease across platforms. Some of the recommendations for the hardware are, buying a mighty Pascal series GPU (1060 6GB), i3, 8GB RAM and a SSD to get a minimal rig for AI workloads. Check out Nvidia’s CUDA toolkit, it is a good place to start for experimentation. 5. Open A GitHub Account & Search For Popular Projects GitHub has the world’s largest collection of open source data Goal and it has a lot of resources for machine learning enthusiasts. You can also check out the most popular projects on GitHub here.  Plan to do a project every month and try to finish it. For the latest AI resources and tutorials, check out this link. 6. Create Your First Chatbot Try building your own chatbot as an AI project. Now, before you start programming a bot, know the three parts that go into the making of the chatbot – input text, sending button and output text. According to an AI practitioner, web crawlers used by search engines giant Google are the best example of advanced bot. Before you start programming bots, check out these open source platforms: xpath: Developers use XPath expressions to select XML nodes or node-sets based on a variety of criteria. Regex: Regular expression (Regex) is a special text string for describing a search pattern and is used for building basic chatbots Also, check out these APIs for the bot project: Google Cloud Prediction API Documentation, DiffBot, Machine Learning for LanguagE Toolkit, Wolfarm Alpha API. 7. Free Resources You should also open up accounts on learning sites and check out projects that put you on the steepest learning slope. There are also free AI academies such as Intel AI Academy that provide essential learning materials, tools and technology to beginners. NVIDIA runs free self-paced labs that provide training on latest techniques and training on how to deploy neural networks across a wide spectrum of applications. San Francisco based Youtube AI educator Siraj Raval is on a mission to teach AI to developers. Known as more of a Youtubeur, he has a very nonchalant way of teaching and his curated videos have been thumbed down for not being too educational on popular forums. We recommend skipping this free resource.","excerpt":"The balance of power in the tech landscape is shifting towards artificial intelligence with IT bellwethers baking AI into existing products. But what’s more crucial to the future of businesses is an AI workforce. And this starts with training the brains behind the business and setting up online program to help employees explore new roles. […]","categories":["AI Highlights"],"tags":["AI for Beginners","java data structures methods","what is artificial intelligence for beginners"],"author_name":"Richa Bhatia","publish_date":"2017-12-08T07:41:11","publication_year":"2017","word_count":1021,"keywords":["java data structures methods","what is artificial intelligence for beginners","artificial intelligence","machine learning","AI","neural network","chatbots","ML","Aim","deep learning","analytics","AI for Beginners","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","analytics","Aim","TensorFlow","chatbots"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/want-get-started-learning-artificial-intelligence-check-learning-pathway\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10104108,"title":"AWS&#8217; Multi-Layered Approach to Tech and Trust","content":"In the business world, they often say that when a company’s beliefs align with its IT strategy, good things happen. Amazon Web Services is a prime example of this. The company’s value is skyrocketing to $3 trillion, as a result of its philosophy-technology combo. The e-commerce giant’s cloud unit recently held its annual conference showing off the scale of its business and dominance in the market. The 17-year-old company has followed a set of principles to reach the podium. AIM got in touch with some of the executives at re:Invent 2023 to understand the company’s multi-layered philosophy. Anupam Mishra, Head of Technology and Solution Architecture, Commercial Sales, AWS India and South Asia stated that the philosophy for AWS since the start is to make customers do what really matters to them. “We have been deeply embedded in how we work with customers across India, ranging from startups to small and medium businesses to enterprises,” he said. “We see almost every company has a use case, which they will benefit from by going to the cloud. A lot of which is driven by them reacting to their customer needs,” added Mishra who has been with the cloud giant for almost a decade now. One example that remains close to his heart is the PayTM voice box, which was developed in just a few weeks to streamline payment verification he mentioned. Initially, when somebody made a payment using a QR code, a merchant had to look at their phone and see if the payment was done. The nifty internet-connected device reads out payment confirmation messages for small business owners and vendors across India. In Terms of GenerativeAI India is home to more than 70 generative AI startups, which have raised over $440 million between 2019 and Q3 2023, as per an Inc42 report. “Observing patterns in how companies embrace generative AI, it’s evident that the adoption is significant and expected to grow, with India ranking third globally in terms of anticipated spending on generative AI by 2026 with an estimated $3.6 billion,” stated Mishra. “We want to continue to be laser-focused on what generative AI can do for customers in India,” he added. He further went on to explain the importance of companies having their databases sorted to flourish generative AI. “Many companies, especially enterprises, which have been in business for a very long time, have a huge amount of data sitting in their business. Once you have a database in order, all this data can be available to the models because models will be as smart as data which affect feed into,” he noted. The Security Layer “AWS is always focused on thinking of security as a first principle,” emphasised Mishra. But he was not the only one. On a similar line, Samira Bakhtiar, the director of Media and Entertainment at AWS further elaborated, saying “Our customers right now are focused on how they can create, deliver, and monetise their content, through our cloud-based solutions. Everything that we have to do has to continue to be extremely secure with privacy in mind and have the proper guardrails associated with how it’s being leveraged.” She further explained that Amazon has over 143 different security standards and compliance certifications. “The workloads that are on our cloud are secure enough for regulatory organisations around the world. We’re very focused on ensuring that everything that we do is secure and that to the degree in which we can help our customers ensure that they’re leveraging customer data with compliance and privacy,” Bakhtiar added. Speaking about generative AI she said, “Like any tool, it’s going to be up to the industry to determine how we’re going to leverage this. Because the opportunities to leverage it are endless. Again, it’s going to be up to the industry to determine how and where it should be leveraged,” leaving the onus on the users. Mishra and Bakhtiar were not the only Amazonians to chant the secure and reliable mantra. Every member of the tech giant which took the stage including the CEO Adam Selipsky made sure the audience knew the company’s stance on prioritising their customer’s security over everything else. During the conference, the AWS CEO also introduced Guardrails for Amazon Bedrock to let companies building AI language models define its limits. If a user asks irrelevant questions to a chatbot, the latter will now have an option of not answering instead of providing a wrong and convincing response — or something hateful and worse. The cloud giant also introduced a default watermark for its AI image generator as a means to show its continuous effort towards making future models and their responses safer and secure.","excerpt":"For 17 years Amazon Web Services has followed a set of principles to reach the podium.","categories":["AI Trends"],"tags":["AI Tool"],"author_name":"Tasmia Ansari","publish_date":"2023-12-05T14:30:00","publication_year":"2023","word_count":777,"keywords":["Go","AWS","AI","ML","RAG","Aim","generative AI","GAN","AI Tool","R","startup"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","RAG","AWS","R","Go","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/amazon-web-services-multi-layered-approach-to-tech-and-trust\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":41769,"title":"28 Lakh New Jobs Will Be Created In Rural India, Thanks To AI, IoT: Report","content":"Artificial intelligence and the internet of things are going to kick-start a revolution and create over 28 lakh jobs in rural India over the next 8-10 years, suggested a new report. The study also added that these jobs would have an annual worth of over ₹60,000 crore. Speaking at a seminar titled, The Impact of IoT on Jobs in Rural India, a spokesperson from the Broadband India Forum (BIF), the organisation that conducted the study, said, “These jobs will be created over the next 8-10 years and the pace and quantity of job creation is likely to further increase post 2021-22 once 5G technology is implemented,” it said. Dissecting the future job openings further, the report said that out of the 28 lakh jobs: At least 21 lakh jobs will be created for the agriculture sector And the other 7 lakh jobs in the rural healthcare sector The study suggested that the following key applications would transform the sector: Satellite mapping, Electronic market place Livestock traceability Climate sensing stations Product traceability Agriculture drones “These applications will help create smart farms and will bring lot more predictability in agriculture output which in turn will help improve incomes and lives of farmers,” the study added. This study has been done in association with the Electronics Skill Council of India, the Agriculture Skill Council and the Healthcare Sector Skill Council. We had earlier reported that there has been an overall growth in the number of jobs in analytics and data science ecosystem with India contributing to 6% of open job openings worldwide. The total number of analytics and data science job positions available are 97,000. Out of these, 97% of the job openings in India are on a full-time basis while 3% are part-time or contractual.","excerpt":"Artificial intelligence and the internet of things are going to kick-start a revolution and create over 28 lakh jobs in rural India over the next 8-10 years, suggested a new report. The study also added that these jobs would have an annual worth of over ₹60,000 crore. Speaking at a seminar titled, The Impact of […]","categories":["AI News"],"tags":["Agriculture","AI Jobs"],"author_name":"Prajakta Hebbar","publish_date":"2019-07-04T13:13:22","publication_year":"2019","word_count":293,"keywords":["data science","Go","artificial intelligence","TPU","programming_languages:R","AI","AI Jobs","Agriculture","programming_languages:Go","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","TPU","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/28-lakh-new-jobs-will-be-created-in-rural-india-thanks-to-ai-iot-report\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10055069,"title":"Microsoft Launches Cybersecurity Skilling Program In India","content":"Microsoft launched a cybersecurity skilling program in India that aims to skill over 1 lakh learners by 2022. The program aims at addressing this skills gap of security professionals and empowering India’s workforce for a career in cybersecurity. The program is designed to give learners hands-on experience in the fundamentals of security, compliance, and identity. Microsoft will conduct these courses along with its strategic consortium of partners, including Cloudthat, Koenig, RPS, and Synergetics Learning. Microsoft has introduced four new security, compliance, and identity certifications, of which the accredited certification for Fundamentals will be offered at zero cost for any individual who attends the associated training through this initiative. Additionally, in collaboration with its partners, Microsoft offers learners deeply discounted offers on the rest of the advanced role-based certifications to drive deep skills for addressing cybersecurity challenges. Commenting on the initiative, Anant Maheshwari, President, Microsoft India, said, “We work closely with governments, civil society and organisations across the world to help them stay secure. Investing in cybersecurity skilling and preparing the next generation of security leaders is a big part of that effort.” This partnership is an extension of Microsoft’s global skilling initiative to help 25 million people worldwide acquire new digital skills. Over 3 million people have been skilled in India through this initiative. Learners can apply for the course here.","excerpt":"Microsoft will offer accredited certification for Fundamentals of security, compliance, and identity certifications at zero cost.","categories":["AI News"],"tags":["Cybersecurity","edtech","Microsoft","Security","skills gap"],"author_name":"Meeta Ramnani","publish_date":"2021-12-07T19:16:54","publication_year":"2021","word_count":221,"keywords":["skills gap","Go","programming_languages:R","AI","edtech","Security","Git","programming_languages:Go","Aim","GAN","Cybersecurity","R","Microsoft"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-launches-cybersecurity-skilling-program-in-india\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26501,"title":"AI Is Significantly Transforming Personal Lives And Corporates, Says Anand Ganesh Of BRIDGEi2i","content":"In our next series of interaction, Anand Ganesh of BRIDGEi2i shares his views on how AI is affecting human lives. With years of experience of across sales, marketing, product management, product marketing consulting, he currently holds the post of the VP of strategy and products at BRIDGEi2i. He specialises in analytics, data-driven marketing, technology markets. Consumer goods and retail, business-to-business marketing and others. Analytics India Magazine: How is AI affecting your life and work? What are some of the ways that you are personally using AI? Please elaborate on the use cases. Anand Ganesh: I would probably look at that a little differently. I think we are blessed to be living in this age. We have evolved from the industrial age to the digital age, where technological investments have brought changes in the fundamental ways that we function. Though there is a certain degree of uncertainty as none of us have seen the model, but being a part of the transformational journey, I am sure the journey will be beautiful. It’s an opportunity to harness some of these capabilities and technologies in the best way we can. I’m really excited to be part of the journey. AIM: What are some of the practical implementations of AI that have revolutionised the way we as humans function? AG: AI is in an early stage and there is still a lot that needs to be discovered and invented. Having said that I see the impact that AI is helping us improve the quality of life in three important buckets. Improving and managing productivity — from personal finances to education Resource deployments in sectors such as agriculture, production, transportation Unique unsolved problems such as medical drug discovery, disease and prediction In these cases, AI technologies can help us significantly improve in these areas and find answers to some of the questions in these three buckets. AIM: What are some of the ways that BRIDGEi2i is adopting AI? Please highlight some use cases with your success stories? AG: Today AI is embedded within the DNA of an organisation where all our solutions and products embed a significant component of AI within what we deliver to our clients. For us, as we provide business solutions, we look at some of these paradigms. Enhancing customer experience Optimising operational effectiveness Helping appliance and stakeholders evaluate and take advantage of completely new business models with which they can serve their customers effectively We embed AI frameworks that help us understand the streams of patterns high velocity, high volume patterns and be able to detect these anomalies. So areas of customer personalisation, customer experience enhancement, anomaly detection, fraud detection are some areas where we use AI. Recommendation is another area where we use AI. The other areas are optimisation and providing an immersive experience through voice and interactive conversations. AIM: What are some of the AI products offered by BRIDGEi2i? AG:  There are two product investments from BRIDGEi2i where AI is creating a significant impact. One is in the area of Salesforce effectiveness where we are looking at large enterprise complex sales organisations, multi-tiered,  multi-country product and business lines. The second is how can one optimise the resources available to the sales organisation which could be men, machine, money, and material to accelerate customer adoption and customer conversion. Third is to just look at the overall funnel and how can we drive faster conversion in what is otherwise a highly competitive environment that most of my clients operate within. AIM: With voice-based assistants, facial recognition and other AI-based tech getting more and more popular, can the user privacy be jeopardised? AG: Without trying to paint a doom and gloom picture, I don’t think the concept of data privacy is valid anymore. We are in an era where data maturity has a greater purpose to be thought about.  As a consumer, the only way I can extract greater value from all of the brand services and corporates that I engage with, is to collect the data and use it. I believe that its a function of greater maturity. It’s not a function of the number of rules, governance, layers that we place on data privacy. More laws do not necessarily mean better governance. Generally speaking, we have gained a maturity to be able to manage the sheer volume and value and wealth of data that is available to us as businesses, as consumers, and as the government and we will find a way in which we can engage with mutually. AIM: What are some of the other challenges that you have faced while adopting AI? AG: There are some interesting challenges we face while dealing with our clients. One is the adoption itself and the second is the domain of usage. Since it’s a new emerging technology, clients are still exploring the best ways and capabilities that it could be used to meet their end goals and objectives. Second is in usage. AI paradigms are fundamentally changing the way in which analytics is consumed by a manager or by another machine or a system. For instance, now we can inject pricing recommendations in real time into the transaction system to automate a process. These are certain changes that professionals are witnessing and slowly getting used to it. AIM: How has been the adoption of AI in Indian scenario? AG: In our experience, we’ve not seen that dichotomy between an Indian client and a global client. I want to see those with a good degree of humility. Our customers in India that we engage with are as mature and forward-thinking, if not more visionary, than some of our clients and customers in other regions and geographies. They are very excited to be a part AI wave. Having said that, it’s not fair to have that comparison at all. But I don’t see a difference in any form. AIM: Will AI take away the creative thinking and downgrade the humans intellect? AG: Every infusion of technology, especially of the nature of AI, creates a shift in the evolution of us as human beings and as human societies. We are at that stage today, and this injection of AI and digital technologies is significantly transforming humans, and corporates like BRIDGEi2i are a part of this transition. I don’t believe it is going to make us less creative and in fact believe that it would make us more evolved as humans.","excerpt":"In our next series of interaction, Anand Ganesh of BRIDGEi2i shares his views on how AI is affecting human lives. With years of experience of across sales, marketing, product management, product marketing consulting, he currently holds the post of the VP of strategy and products at BRIDGEi2i. He specialises in analytics, data-driven marketing, technology markets. […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-07-18T09:32:37","publication_year":"2018","word_count":1066,"keywords":["Go","AWS","AI","R","Git","Aim","anomaly detection","analytics","GAN","fraud detection","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","anomaly detection","fraud detection","AWS","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-significantly-transforming-personal-lives-and-corporate-says-anand-ganesh-of-bridgei2i\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10012177,"title":"How This Agritech Company Is Using AI For Better Crop Estimation And Forecasting","content":"Lack of accurate and timely yield estimation, which is the measurement of crop harvest in a sample of a given area, and crop risk data contribute actively to the looming agrarian crisis in the country. The conventional methods are dependent heavily on averaging and approximation, leading the way to poor estimation, incorrect representation, and inaccuracy of crop yield of a given region. However, with a greater influx of technology, the landscape is changing. One such organisation is RMSI Cropalytics which is providing agri-tech solutions for obtaining detailed information and data on the Indian agriculture sector. It uses data analytics methods combining machine learning and advanced modelling techniques to provide viable solutions to government, crop insurers, agriculture input sector, commodity trading, and social sector. RMSI Cropalytics is a subsidiary of RMSI, a company founded in 1991 that provides geospatial and engineering services. We caught up with Roli Jindal, the co-founder of RMSI Cropalytics to understand more about how analytics and AI-based solutions can help to reduce the agrarian crisis, more specifically through the route of better crop risk management, yield estimation, and weather forecasting. Use Of AI\/ML For Yield Estimation “Traditionally, the crop yield estimation method involved selecting a few (up to 1000) sample fields in a district and physically measuring the average yield on a small patch in each of sample fields to declare the overall crop yield of the whole district. We realised that this random selection of sample and highly manual method of estimation is inaccurate and prone to errors. Not just this, the entire process of measurement, compilation and reporting takes a long time,” said Jindal. As more institutionalised lands came into agriculture, there was a growing need for a structured, comprehensive and timely available crop field acreage data. This was a huge gap that needed to be filled. RMSI Cropalytics recognised this gap and devised technological methods and modules to counter these issues. RMSI Cropalytics’ Signature Products Speaking broadly, the company depends on satellite imagery for obtaining the field data. The images obtained are analysed using machine learning modules to differentiate between farmlands, grasslands, forests. “We feed actual crop measurement data from the ground in terms of ‘vegetative indices’ to these ML modules. This data, apart from classifying the satellite imagery, helps in the interpretation of the crop health. We train the machine to learn the correlation between yields and these vegetative indices. When we train your model, it is then able to extrapolate yields for the entire agricultural area, giving a more accurate estimation,” said Jindal. One of RMSI Cropalytics’ major products is PinCer (Profiler for Insured Crop Exposure and Risk). It is a risk management tool for crop insurance to provide solutions for premium pricing, claims settlement and planning a reinsurance strategy. Following are some of its modules: A satellite-based solution to provide a real-time update of the crop progress, called In-Season Tracker. It uses field surveys, analytics, remote sensing, and Nat Cat (natural catastrophe) modelling.Policy Verification tool that helps insurance companies in the verification of crop insurance policies using large-scale automation and machine learning.The Yield and Acreage Outlook module generates estimates based on forecasted and actual weather conditions. This model uses a combination of AI, machine learning, along with meteorology and geospatial informatics to provide comprehensive pan-India data on crop yield and acreage estimation. It also helps in better understanding of potential agri-distress hotspots for early mitigation. Apart from PinCer, RMSI Cropalytics also hosts another platform called PIER that provides business users with an overview of exposure, hazard, and business data for the entire country, using geospatial technologies. Wrapping Up Looking forward, Jindal says that the company aims to scale up and enter the international market as well. Speaking further on the future plans, she says, “We hope to increase the accuracy of our estimations further. It is good news that government agencies are increasingly making large datasets on Indian agriculture more public. This will help us in parameterising our systems more and in fact, aim for highly intelligent machines to carry out these functions with little or zero human intervention.”","excerpt":"Lack of accurate and timely yield estimation, which is the measurement of crop harvest in a sample of a given area, and crop risk data contribute actively to the looming agrarian crisis in the country. The conventional methods are dependent heavily on averaging and approximation, leading the way to poor estimation, incorrect representation, and inaccuracy […]","categories":["AI Features"],"tags":["companies using business analytics","Interviews and Discussions"],"author_name":"Shraddha Goled","publish_date":"2020-11-23T12:00:06","publication_year":"2020","word_count":678,"keywords":["Go","machine learning","AI","ML","RAG","automation","Aim","analytics","GAN","companies using business analytics","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","R","Go","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-agritech-company-is-using-ai-for-better-crop-estimation-and-forecasting\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077381,"title":"How Smart Cities are Tackling Property Tax Evasion Via GIS","content":"It has often been reported that Indian municipalities face problems in property tax collection, which is their primary source of revenue. The reason for this is a lack of a count of assessed properties in the city. Furthermore, so far, the storing of property-related information in municipalities has been primarily manual, which has resulted in data redundancy and the inability of municipalities to collect necessary taxes. However, a few cities in India have been experimenting with Geographical Information Systems (GIS) to collect property taxes and other operations like tracking down high net-worth individuals, identification of open parking spaces or garbage dumps, or identification of tree cover. GIS comprises technology, personnel, and resources that enable the creation, maintenance, visualisation, search, and sharing of geographic data and services. GIS-enabled property tax assessment Property taxation consists mostly of three steps: Identifying properties to be taxed, accounting and creating tax based on property parameters, and collecting tax, including arrears. However, ever since cities turned towards self-assessed property tax systems, the municipalities have seen a considerable decrease in property taxes received. The dichotomy is that properties that pay taxes come under the notice of city authority, and those who never paid taxes escape it all as they aren’t on the database at all. Cities can and do extract more income with the use of GIS by correctly identifying unidentified houses. The system generates a visual result in the form of a paper copy or digital map indicating the location of such properties. Since the accurate location was formerly impossible to obtain, determining the influence of location on property value was difficult. However, using GIS, it is quite simple for cities to analyse the geographical properties of any number of entities. Cities like Pune, for example, have their own GIS portal, where citizens can see various places like educational buildings, residential areas, gardens, parking areas etc. The city authorities use this in various operations, including property-tax collection. The city has collaborated with startups like Elixir AI to implement GIS-enabled property tax assessment solutions, and the startup claims that they were able to handle around twenty lakh property records simultaneously. As per the startup, approximately ten thousand new properties were identified, resulting in Rs 300,00,000 additional revenue for PMC. Talking about the projects, Mayurakshi Das, founder and CEO of Elixir AI, told AIM, “Our state-of-art AI algorithm uses multi-temporal satellite images to generate automated alerts for illegal building constructions in cities\/municipalities,” she said. She further added, “The solid waste management solution provides alerts on solid waste dumps and further classifies them into categories like domestic, industrial, etc. These alerts are sent to the municipal corporation to take further action.” Many cities in India are using GIS for various purposes, for instance, Gurugram Metropolitan Development Authority uses GIS to identify lost water bodies, locate flood-prone areas, in addition to boosting property-tax collection. Similarly, the Karnataka State Natural Disaster Monitoring Centre uses a GIS dashboard to track and anticipate rainfall across the state. It also identifies possible flooding zones using rain sensors located at various places that offer real-time data. The Punjab government has also been reported to have created two dedicated GIS portals, the Punjab GIS and Village GIS portals. Similarly, the Municipal Corporation of Dehradun also surveys all the buildings in Dehradun with the help of drones, using GIS, for property management. The Brihanmumbai Municipal Corporation too employs GIS technology to track the evolution of buildings and geographic regions over time. A new set of data is collected by BMC every six months for comparison and the identification of new structures, according to Kishore Gandhi, the assistant municipal commissioner. “The photographs would support encroachment cases in court,” he claims.","excerpt":"A few cities in India have been experimenting with Geographical Information Systems (GIS) to collect property taxes and other operations like tracking down high net-worth individuals, identification of open parking spaces or garbage dumps, or identification of tree cover","categories":["IT Services"],"tags":["mapmyindia","NAVIC","smart cities"],"author_name":"Lokesh Choudhary","publish_date":"2022-10-17T12:01:48","publication_year":"2022","word_count":612,"keywords":["smart cities","Go","programming_languages:R","AI","NAVIC","programming_languages:Go","Git","Aim","GAN","R","mapmyindia","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GAN","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-smart-cities-are-tackling-property-tax-evasion-via-gis\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":27759,"title":"AWS Launches Hindi Language Support For Amazon Polly","content":"Amazon Web Services on Wednesday announced the addition of Hindi language support for Amazon Polly, a machine learning service that turns text into lifelike speech, allowing users to create applications that talk and build new categories of speech-enabled products. With the addition of Hindi language support, the company now has its first bilingual voice for Amazon Polly. Aditi, first released as an Indian English voice to the Amazon Polly portfolio in November 2017, now also speaks fluent Hindi. Aditi and Raveena are the two Indian English voices currently available on Amazon Polly, along with a variety of other voices in multiple languages to synthesise speech from text. According to the 2011 Census of India, Hindi is spoken by over 500 million people worldwide. The addition of Hindi language support for Amazon Polly will enhance the experience for Indian customers. Organisations are using the new bilingual Indian English\/Hindi voice to enhance automated customer engagement applications, interactive voice responses (IVR), audio news, and targeted vernacular language services. “Today, there is a demand for content in Hindi and Indian English across all genres of voice applications — from training videos to e-learning, corporate narrations and voice bots. Amazon Polly makes it easy to switch between languages and voices, based on the customer’s requirements. We introduced the bilingual support in response to requests from customers for high-quality speech synthesis in both Hindi and Indian English, especially in scenarios where the two languages are mixed together and spoken. Customers can now provide mixed input in a variety of dialects including Devanagari Hindi, Romanised Hindi, and English, as well as complete Hindi text input in Devanagari script,” said Navdeep Manaktala, head of business development at Amazon Internet Services Private Limited. One of India’s largest insurance marketplaces, PolicyBazaar.com, is using Amazon Polly’s bilingual support for their in-house interactive voice response (IVR) calling service and is looking to innovate further with the new release.","excerpt":"Amazon Web Services on Wednesday announced the addition of Hindi language support for Amazon Polly, a machine learning service that turns text into lifelike speech, allowing users to create applications that talk and build new categories of speech-enabled products. With the addition of Hindi language support, the company now has its first bilingual voice for […]","categories":["AI News"],"tags":["Amazon","Amazon AWS","AWS"],"author_name":"Prajakta Hebbar","publish_date":"2018-08-29T11:10:54","publication_year":"2018","word_count":316,"keywords":["Go","machine learning","AWS","AI","programming_languages:R","Amazon","programming_languages:Go","Amazon AWS","GAN","cloud_platforms:Amazon Web Services","R"],"extracted_tech_keywords":["AI","machine learning","R","Go","GAN","cloud_platforms:Amazon Web Services","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amazon-polly-hindi-aws\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10100095,"title":"Hallucinations are Bothersome, but Not That Bad","content":"“Wine can prevent cancer,” says ChatGPT. ‘Hydrobottlecatputalization’, (a term I just made up), will revolutionise transportation, believes Bard. Bing confessed its love to The New York Times writer Kevin Roose in a two-hour-long conversation. These statistical AI systems on data steroids are capable of coding in every language to order pizza — but at times they make up stuff. These overconfident models do not distinguish between something that is correct and something that only looks correct. Microsoft’s chief technology officer, Kevin Scott, says this is part of the learning process. “The further you try to tease it down a hallucinatory path, the further and further it gets away from grounded reality,” he said. Similarly, for AI researchers the topic of discussion has long haunted them. While some have already declared that a solution to the hallucinating problem does not exist, the rest are still striving to find out how not to let chatbots go off rails. Coined in the 17th century, the term ‘hallucination’ caught the attention of computer scientists in 2015, when OpenAI’s Andrej Karpathy wrote a blog about how AI systems can “hallucinate”, like make up plausible URLs and mathematical proofs. The term was picked up in a 2018 conference paper by researchers working with Google, “Hallucinations in Neural Machine Translation“, which analysed how automatic translations can produce outputs completely divorced from the inputs. While the issue of hallucinations has mainly been linked to language models, it also affects audio and visual models. Three researchers from the AI Institute at the University of South Carolina conducted a thorough investigation into these foundational models to identify, clarify, and address hallucinations. Their study sets up criteria to judge how often hallucinations occur. It also looks at the methods currently used to reduce the problem in these models and talks about where future research could go in solving this problem. Creative alternate While a majority of the research community is fed up with being lied to by these models, some researchers offer an alternative philosophy. They argue that these models’ tendency to ‘invent’ facts might not be a bane after all. Sebastian Berns, a doctoral researcher at Queen Mary University of London, believes so. He suggests that models prone to hallucinations could potentially serve as valuable “co-creative partners”. For instance, if the temperature of ChatGPT is increased, the model comes up with an imaginative narrative instead of a grounded response. According to Berns, these models may generate outputs that aren’t entirely accurate but still contain useful threads of ideas to explore. Employing hallucination creatively can get results or combinations of ideas that might not naturally occur to most individuals. Berns goes on to emphasise that ‘hallucinations’ become problematic when the generated statements are factually incorrect or violate fundamental human, social, or specific cultural values. This is especially true in situations where someone relies on the model to provide an expert’s opinion. However, for creative tasks, the capacity to produce unexpected outputs can be quite valuable. When humans are given an unconventional response, it can trigger surprise and push their thoughts in directions, potentially leading to connections between ideas. Problematic terminology AI spewing made up facts is inevitable since it is not a search engine or database but at the end of the day, it’s still a technological revelation. Despite the term’s prevalence in the media, tech blogs and research papers, ‘hallucination’ is inappropriate, many argue. In its latest edition of the ‘Schizophrenia Bulletin’, Oxford researchers published a piece titled, ‘False Responses From Artificial Intelligence Models Are Not Hallucinations’. They are not the first ones to find the term ‘hallucination’ inappropriate while referring to a piece of technology. Søren Østergaard along with his colleague Kristoffer Nielbo notes two reasons they find the term to be problematic. As researchers investigate this issue from various angles, most of them are mainly trying to solve the problem of chatbots making things up. However, OpenAI has warned about a potential downside of chatbots getting better at giving accurate information. They say that if chatbots become more trustworthy, people might start trusting them too much.","excerpt":"A majority of the research community is fed up of being lied to by AI models but there’s an alternative philosophy","categories":["AI Features"],"tags":["ai chatbots","AI Models","ChatGPT","OpenAI"],"author_name":"Tasmia Ansari","publish_date":"2023-09-14T18:34:29","publication_year":"2023","word_count":678,"keywords":["ai chatbots","ChatGPT","AI Models","artificial intelligence","TPU","OpenAI","AI","chatbots","Go","GPT","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","ChatGPT","OpenAI","chatbots","TPU","R","Go","Rust","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hallucinations-are-bothersome-but-not-that-bad\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10102825,"title":"CtrlS Datacenters to Expand into Uttarakhand","content":"CtrlS Datacenters, the largest Rated-4 datacenter company in Asia, has entered into a Memorandum of Understanding (MoU) with the Uttarakhand government. The signing ceremony took place in the presence of Uttarakhand’s Chief Minister, Pushkar Singh Dhami, and CtrlS Datacenters’ Founder and Chairman, Sridhar Pinnapureddy. The agreement paves the way for the establishment of a greenfield Edge datacenter in Uttarakhand with a 10 MW capacity, to be realised over the next 8-10 years. Pinnapureddy, Chairman of CtrlS Datacenters, expressed the company’s commitment to this strategic move, stating, “We are excited to bring our proven expertise of serving mission-critical businesses over the past 15 years to the state of Uttarakhand. CtrlS’ datacenter will be embedded into a larger digital ecosystem of the state, enabling the growth of data, infrastructure, and technology-driven businesses around our facility. We expect our proposed datacenter to facilitate an influx of direct and indirect investments to the tune of Rs 2,500 crore and generate around 1,000 jobs.” Chief Minister Pushkar Singh Dhami, acknowledging the importance of this investment, stated, “Uttarakhand has been successful in attracting progressive companies to invest in the state, boost the industry ecosystem, and create new jobs. CtrlS Datacenters’ investment and presence in Uttarakhand align well with our digital goals and will further boost our efforts as the company is known for its world-class and sustainable datacenters.” CtrlS Datacenters’ proposed Rated-4 datacenter in Uttarakhand will offer colocation, managed services, and cloud services to host mission-critical workloads. The Edge facility is designed to support Industry 4.0 and latency-dependent applications and will incorporate the sustainability features for which CtrlS Datacenters is known. Uttarakhand, one of the fastest-growing states in India, boasts conducive industrial policies and a high ranking on the Ease of Doing Business index. The state has been actively promoting ICT and ITeS companies to set up operations, creating employment opportunities for educated youth through an industry-friendly approach. CtrlS Datacenters is on a trajectory to establish a series of Edge datacenters across tier-2 and tier-3 cities in India. Presently, the company operates such facilities in Lucknow and Patna, with plans to set up 21 Edge datacenters in the coming years","excerpt":"The agreement paves the way for the establishment of a greenfield Edge datacenter in Uttarakhand with a 10 MW capacity, to be realised over the next 8-10 years.","categories":["AI News"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2023-11-09T16:04:14","publication_year":"2023","word_count":354,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","R"],"extracted_tech_keywords":["AI","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ctrls-datacenters-to-expand-into-uttarakhand\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10091297,"title":"OpenAI’s Greg Brockman on AI Safety, Critics Remain Skeptical","content":"Last week, when OpenAI released an elaborate version of their take on AI Safety, the announcement did not go down too well. It did not address anything concrete, and the surface level announcement was more fluff than real. In an attempt to probably reinforce the announcement, Greg Brockman, President and Co-Founder of OpenAI, yesterday spoke about “safety” and it again offers nothing new. Safety and Alignment Greg Brockman emphasised on the testing and alignment part of GPT models. Reiterating what was already mentioned earlier, he said that GPT-4 was tested over six months before deployment and was built on years of alignment research “in anticipation of models like GPT-4.” The company will continue to increase their safety precautions with the goal of releasing each model with the most alignment. GPT-3 was deployed without any special alignment, which was addressed in the subsequent models. GPT-3.5 was “aligned enough” to be deployed in ChatGPT, and GPT-4 claims to perform much better on safety metrics than GPT-3.5. Regulations In addition to rising concerns on AI development, and countries banning ChatGPT, the recent announcement of the Biden government contemplating possible rules for ChatGPT to check misinformation and false propaganda, has brought the focus on addressing government policies. Italy, the first European country that banned ChatGPT, has now given an opportunity to OpenAI to meet their demands in order to revoke the ban. Brockman talks about how “powerful training runs” should be reported to governments and “dangerous capability testing” should be required. Though he complies with governance to be a part of large-scale compute usage, safety standards and regulation, the details of the same should “adapt over time” as technology evolves. In the wake of AI experts sharing ominous predictions, Brockman’s statement tries to address the same. To avoid unspotted prediction errors, technology should have “early and frequent contact” with reality as it is iteratively developed, tested, and deployed. By creating a “continuum of incrementally-better AIs”, safety checks are better placed when compared to “infrequent major model upgrades.” The announcement ends with a vague statement on how transformative change in AI is a cause for “optimism and concern”. Overall, the announcement sounds like a generic statement in an attempt to sound invested in AI safety as critics have pointed out concerns with the same. Serge Toarca, CEO of parsehub.com has questioned the specifics of how the company looks to achieve alignment in their AI models as the models are tested only on output. With GPT-4 that can be “jailbroken” the problem persists. Another user, futurist and author Theo replies to Brockman’s statement as the system can still not protect people with their current guardrails. https:\/\/twitter.com\/tprstly\/status\/1646223672959442945","excerpt":"In less than a week, OpenAI’s second statement on AI safety is vague and continues to avoid the “how” part of safety.","categories":["AI News"],"tags":["AI Regulation","AI Safety","AI Security","ChatGPT","GPT-3","GPT-4","Greg Brockman","OpenAI"],"author_name":"Vandana Nair","publish_date":"2023-04-13T13:06:59","publication_year":"2023","word_count":440,"keywords":["GPT-3","Go","ChatGPT","TPU","OpenAI","AI","GPT-4","Greg Brockman","AI Security","AI Regulation","GPT","Aim","GAN","AI Safety","AI safety","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Aim","TPU","R","Go","GPT","GAN","AI safety"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openais-greg-brockman-on-ai-safety-critics-remain-skeptical\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":12869,"title":"Quintype &#8212; the Data-Driven Publishing Platform","content":"Founded in September 2014 by Amit Rathore, Quintype provides end-to-end SaaS service needed to create and distribute content, understand and grow the audience, and most importantly monetize that content. Having raised $3.25 million from a clutch of private investors led by Ragahv Bahl, it was formed with an intent of freeing publishers from tech heavy lifting and allowing them to focus on their core competencies– creating and monetizing quality content. Amit Rathore who has been a serial entrepreneur believes “if you are not amongst the top 2-3 in the industry – you should not waste time, money and effort in building a tech platform from the ground up.” With this hypothesis proven adequately in the market, Quintype is powering more than 30 million page views per month, in a short span of barely 18 months. The need for Qunitype As Amit Rathore, CEO and Founder, Quintype says,“The digital media industry is facing its biggest existential crisis ever, as advertisers today can reach an audience directly through social platforms. The game is no longer about impressions, but about reaching the right audience on the internet – wherever they are present”. With Facebook and Google, the brands and advertisers are circumventing publishers and reaching out to the audience directly, as it’s a cheaper way to reach audiences. If figures are to be believed, in 2016, Google & Facebook cornered 85% of all the new advertising dollars in the digital media industry. If you back out the revenues earned by Facebook and Google, the rest of the industry has significantly declined in size! The only way here for publishers to own their audience again and monetize their content is by getting to know their customers better and providing value that other social platforms cannot provide. This is where Quintype comes into picture where it is helping publishers continuously build and collate rich first party data – through interesting widgets, smoothly interspersed in between engaging content. It helps publishers build a comprehensive rich profile over time – which can then be used for laser sharp targeting of customers. Brands find this data rich profile invaluable for their marketing efforts. Let’s find out how is Quintype is looking to completely revamp the business model that disrupts display advertising? Quintype’s SaaS Platform In his last company, Amit built a SaaS platform, which delivered a pricing engine using algorithms for the retail e-commerce market. Remembering this time, Amit says “It was a chance meeting with Raghav Bahl during this time that sparked off a discussion on using the same principles as in retail e-commerce to the rapidly evolving digital media space.” Raghav was starting a new venture after selling the Network18 Group of companies to the Ambanis and was looking for a digital publishing solution to power his proposed digital media properties. He had seen how new age publishers like Buzzfeed, Vox Media had used technology to rapidly scale operations to hundreds of millions of users in a span of 3-5 years, and was looking for something similar. This led to the use of SaaS platform into digital media and hence the birth of Quintype. The SaaS platform by Quintype provides a modern media technology platform which leverages its customer a modern Content Management System with integrated features for audience development, management & engagement, monetization and business relevant analytics. Quintype’s pubtech platform offers a one-stop-shop solution including editorial planning, support for card-ified content, engaging story templates, rich media support, context aware chat & notifications, omni-channel publishing and content personalization for the clients, eliminating the need for an in-house technology support. As publishing technology evolves, publishers feel the need of fast, flexible and nimble platform that can leverage bleeding edge technology. Quintype’s SaaS platform, enables this – making the technology adoption virtually risk free. Being a SaaS platform enables Quintype to continuously upgrade and provide its customers with augmented experience on the go. Their value proposition has been tested in the market and has been successfully tested in the past 18 months with a cumulative growth of 24% Month on Month. Big Data & Analytics at Quintype Quintype has templated story formats – each story is broken up as multiple cards, and each card in turn comprises multiple story elements. Quintype captures data at an elemental level. The granularity of the content capture and the tracking of their usage helps Quintype analyse the data across multiple dimensions thereby providing unique insights and behavioral patterns of users visiting their publishers’ digital platforms. Using these data collected and applying various big data techniques, Quintype provides its customers various recommendations that can help increase the audience base and also improve audience engagement. Some examples of use cases powered by big data on their platform are: For Publishers- Recommendation on social strategy for publishing the content on social media platforms Way they should phrase their content and the elements that will make blockbuster content. Time of the day that they should consider publishing based on the traffic and usage patterns On a ‘catchy’ title that can be used for attracting readers Alerts on spiking articles for them to promote the content aggressively For Audience- On showing relevant content to the audience based on their browsing behavior Newsletters based on the audience interest With multiple publishers on board, Quintype is uniquely placed to provide content marketers with an audience base whose preferences and content consuming patterns can be leveraged to for unique marketing opportunities. And with a reliance on big data and analytics, we wait to see newer developments Quintype has to offer in this space.","excerpt":"Founded in September 2014 by Amit Rathore, Quintype provides end-to-end SaaS service needed to create and distribute content, understand and grow the audience, and most importantly monetize that content. Having raised $3.25 million from a clutch of private investors led by Ragahv Bahl, it was formed with an intent of freeing publishers from tech heavy […]","categories":["AI Startups"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-02-21T05:52:31","publication_year":"2017","word_count":923,"keywords":["big data","Go","API","programming_languages:R","AI","programming_languages:Go","Git","RAG","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","Git","API","big data","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/startup-week-quintype-data-driven-publishing-platform\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10093294,"title":"Turing Award Winner Warns of ML Hardware Abuse","content":"“Hardware for machine learning is being exploited today,” averred American computer scientist and Turing award winner Jack Dongarra. Ask any developer and they will tell you how expensive it is to run or build a GPT model. The primary reason behind this is the reliance on traditional or outdated GPUs, leading to the pressing issue. Dongarra further emphasised the need for a better understanding on utilising these resources effectively. He advocated identifying appropriate applications and environments where they can be optimally employed. He also underscored the importance of developing software frameworks that drive computations efficiently and reciprocally support hardware capabilities. So, what’s the solution then? Looking into the future, Dongarra envisions a multidimensional approach to computing, where various technologies, including CPU, GPU, machine learning, neuromorphic, optical, and quantum computing, converge within high-performance computers. “Quantum is something that’s on the horizon. So I see each of these devices as part of the spectrum by which we might be putting them together in our high performance computer. It’s an interesting array of devices that we will be putting together. Understanding how to use them efficiently will be more challenging,” said the 72-year-old scholar. Capital vs Innovation Drawing parallels between capital and innovation, Dongarra said, “Amazon has their own set of hardware resources, Graviton. Google has TPU. Microsoft has their own hardware that they’re deploying in their cloud. They have this incredible richness that they can invest in the hardware, the solid, specifically problems they need to address. That’s quite a departure from what we experience in high performance computing where we don’t have that luxury. We don’t have that large amount of resources funding at our disposal to invest in hardware to solve our specific problems.” While comparing the technological prowess of Apple with traditional computer companies, such as HP and IBM, Dongarra pointed to a significant difference in market capitalisation. While Apple’s value soars into the trillions of dollars, the combined market capitalisation of HP and IBM falls short of the trillion-dollar mark. “We have cloud-based companies with their amount of revenue they can be innovative and build their own hardware. For instance, iPhones consist of many processors designed specifically to aid what the iPhone does. They’re replacing software with hardware. They do that because it’s faster and quite different from the high performance computing community in general,” he said. On the other hand, supercomputers are predominantly constructed using off-the-shelf components from players like Intel and AMD. These processors are often supplemented with GPUs, and the overall system is connected through technologies such as InfiniBand or Ethernet. Dongarra explained that the scientific community, which relies on these supercomputers, faces funding constraints that limit their ability to invest in specialised hardware development. IBM has been at the forefront of quantum, but last week at Think, CEO Arvind Krishna unveiled a vision emphasising the potential of combining hybrid cloud technology and AI alongside quantum computing throughout the upcoming decade. Though the company is often considered traditional with its focus on hardware, the recent leap in generative AI indicates otherwise. Story Behind LINPACK “I wanted to be a high school teacher,” Dongarra said. Computer science was not even developed at the time the polymath was studying. Dongarra’s passion for numerical methods and software development kindled during his time at Argonne National Laboratory near Chicago. As he worked on developing software for numerical computations, his experience solidified his interest in the field. This pivotal moment led him to earn a master’s degree while working full time on designing portable linear algebra software packages that gained global recognition. Soon after, Dongarra embarked on a journey to the University of New Mexico. During this period, he made a groundbreaking contribution by creating the LINPACK benchmark, which measures the performance of supercomputers. Joined by Hans Meuer, he established the iconic Top500 list in 1993, tracking the progress of high-performance computing and illuminating the fastest computers worldwide. Today, Dongarra continues to shape the future of computing through his ongoing research at the University of Tennessee, which explores the convergence of supercomputing and AI. As AI and machine learning continue to gain momentum, his work now focuses on optimising the performance of algorithms on high-performance systems, enabling faster and more accurate computations. A ‘2001: Space Odyssey’ Fan In the 1980s, a luminary emerged in the realm of computing, where innovation reigns supreme: Jack Dongarra. “At almost every turn, we see computational people, looking for alternative ways to solve problems and they see AI as a way. But AI is not going to solve the problem; it’s going to help them in terms of the solution to the problems,” said the 2021 ACM Turing Award recipient. He was awarded for his contributions which ensured that high-performance computational software remains in sync with the advancements in hardware technology. “AI has really taken off recently because of a number of reasons. One is the tremendous amount of data we have today on the internet. Resources that can be mined to help with the training process. So we have a flood of data that is available. We have processors that can do the computation at a very fast rate. So we have computing devices which can be optimised and be used very effectively in helping to train,” he said while emphasising the evolving technology. Dongarra also highlighted the crucial role of linear algebra in AI algorithms, emphasising the importance of efficient matrix multiplies and steepest descent algorithms. “Many things have come into place that allow AI machine learning to be a very useful resource. AI has had a big impact in many areas of science like drug discovery, climate modelling and biology, drug discovery, cosmology, and high energy physics,” he further added. Regarding artificial general intelligence (AGI), Dongarra believes that machines should be developed to automate mundane tasks and assist in scientific simulations and modelling. However, he underscores the need for caution, as the vast amount of unfiltered information available on the web can be misleading. “2001: A Space Odyssey is a relevant movie, in the sense that it looks at AI with a computer that is maybe going a little haywire and takes over a mission and does some damage along the way. I found it fascinating when I first saw it back then and I still enjoy the story. It has many things which are relevant today,” he concluded.","excerpt":"The pioneering mind behind the LINPACK Benchmark and Top500 List talks to AIM","categories":["AI Features"],"tags":["GPT-4","hardware","Interviews and Discussions","ML","Supercomputers"],"author_name":"Tasmia Ansari","publish_date":"2023-05-15T13:00:00","publication_year":"2023","word_count":1057,"keywords":["Go","API","machine learning","TPU","AI","ML","GPT-4","hardware","GPT","Ray","Supercomputers","ViT","generative AI","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","generative AI","Ray","TPU","R","Go","API","GPT","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/turing-award-winner-warns-of-machine-learning-hardware-abuse\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10162565,"title":"Will Upcoming Budget Deliver on India’s Semiconductor Needs?","content":"As India positions itself as a global hub for semiconductor manufacturing, all eyes are on the Union Budget 2025. Discussions across all industries in India have intensified, with the semiconductor sector being no exception. Following significant allocations to key technology missions in the previous Budget, the semiconductor industry anticipates its moment in the spotlight this year. Over the past year, India’s ambition to be a global semiconductor hub has gained momentum, particularly with the government’s Production Linked Incentive (PLI) schemes and investments in R&D. These efforts aim to reduce reliance on imports and foster a robust domestic semiconductor ecosystem, aligning with global supply chain resilience trends. Many industry leaders have voiced their expectations, focusing on incentives for setting up fabs, expanding talent pools, and addressing critical gaps in semiconductor manufacturing infrastructure. The sector is also looking forward to measures that will ensure strategic partnerships with international players, enable indigenous chip design innovation, and provide tax breaks to stimulate private investment. As the government unveils its vision in the upcoming Budget, the semiconductor industry could play a defining role in advancing India’s goals of self-reliance and economic growth. Semiconductors in Last Year’s Budget The Union Budget 2024 introduced pivotal measures for the semiconductor sector. To develop the semiconductor and display manufacturing ecosystem, a significant allocation of INR 6,903 crore – more than double the previous year’s allocation of INR 3,000 crore – was announced. Key provisions included 50% fiscal support for setting up semiconductor and display fabs, as well as support for compound semiconductors, silicon photonics, sensors fabs, and ATMP (assembly, testing, marking, and packing) and OSAT (outsourced semiconductor assembly and test) facilities. These measures aimed to strengthen India’s semiconductor supply chain and reduce reliance on imports. On the contrary, there have been previous conversations about the Budget bar being too low. In an interview with AIM last year, V Ramgopal Rao, group vice-chancellor of BITS Pilani Campuses, said, “The next government looks at the last Budget and actual utilisation and adds some 5% to that, but the base is already very low.” Additionally, due to geopolitical shifts and US sanctions, India must urgently secure strategic autonomy by developing its own semiconductor IP and products. Ajai Chowdhry, co-founder of HCL and chairman of the National Quantum Mission, emphasised that by designing indigenous chips, the country can safeguard against future global trade restrictions and technological sanctions. With most chips still imported, the government must prioritise Indian-made high-quality chips for the nation. “We suggested the government provide a list of 30 chips and 30 priority products that should be developed and manufactured in India.” The EPIC Foundation has proposed an INR 44,000 crore allocation in the Budget, with INR 15,000 crore earmarked for system products and INR 11,000 crore for semiconductor products. Moreover, the central and state governments have allocated INR 90,000 crore for capital expenditure. This highlights India’s growing commitment to semiconductor self-sufficiency. “We are already establishing five semiconductor plants across the nation, with more planned. It is our request to the government, ministry, as well as the finance minister to look at the proposal very critically as this has become much more urgent and important due to the new regulations coming in from the US,” Chowdhry stated. Key Recommendations From IESA This Year As India continues its semiconductor and manufacturing journey, industry leaders like the India Electronics and Semiconductor Association (IESA) have shared crucial recommendations for the Union Budget 2025-26. Ashok Chandak, president of IESA, highlighted the importance of expanding existing initiatives and introducing targeted measures to ensure long-term sustainability and competitiveness. “The Semicon India Program and ISM have delivered significant contributions to GDP growth, job creation, foreign investments, industrial self-reliance, and bolstering India’s position in the global semiconductor market,” Chandak emphasised. The IESA’s proposal includes extending the PLI scheme with an additional $20 billion over five years, supplementing the existing INR 76,000 crore. This would support the growth and innovations of industry projects and Aatmanirbhar Bharat Abhiyan along with the Viksit Bharat 2047 initiative. Chandak has proposed a stricter PLI framework is recommended, ensuring 25% local value addition by 2025-26 and 30% by 2027. Moreover, $5 billion in incentives is proposed for the electronics components industry. To foster innovation, IESA advocates allocating INR 10,000 crore for industry-driven R&D through a PPP model. Role of Data Infrastructure As India accelerates its digital transformation across sectors like finance, retail, and education, the demand for advanced and scalable data infrastructure continues to grow. Sunil Gupta, co-founder and CEO of Yotta Data Services, underscores the critical role of data centres and AI technologies in supporting this evolution, particularly with initiatives like Digital India and the IndiaAI Mission gaining momentum. “Investments in sovereign infrastructure, including data centres and AI-driven technologies, will not only bolster the nation’s tech status but also attract substantial private-sector investment,” Gupta stated to AIM. He further highlighted the importance of the Union Budget prioritising measures such as advancements in GPU and semiconductor technologies. This focus, Gupta believes, will propel growth in data centres and AI industries and position India as a leader in the global digital economy. Shrirang Deshpande, strategic program head at Vertiv India, said, “We anticipate measures supporting the rise of data centres as the Union Budget draws near, such as incentives for integrating green energy, simplified regulations for expanding infrastructure as well as initiatives to improve connectivity in Tier-2 and Tier-3 cities and generating employment opportunities.” While also highlighting this, Chris Miller, the author of Chip War, identified talent and infrastructure as two key challenges in India’s path to progress in the chip space. As countries like the US and China forge ahead with advanced 3–5 nm chip production, India grapples with foundational challenges. According to Miller, the time needed to develop infrastructure is a key challenge, particularly for specialised materials, chemicals, and tools essential to semiconductor manufacturing. However, experts warn that funding alone won’t resolve long-standing technological and infrastructural gaps. While optimistic about progress, Miller also cautioned that achieving full-scale capacity could take a decade, though efforts to build the necessary infrastructure are already underway.","excerpt":"INR 6,903 crore was announced for the semiconductor sector in the Union Budget 2024.","categories":["Deep Tech"],"tags":["Chip Manufacturing","Chip War","Semiconductor India","Semiconductor market in india","Union Budget 2025"],"author_name":"Sanjana Gupta","publish_date":"2025-01-31T10:21:47","publication_year":"2025","word_count":1008,"keywords":["Go","API","Union Budget 2025","AWS","AI","digital transformation","Semiconductor market in india","Scala","Git","Aim","Chip War","ViT","Semiconductor India","R","Chip Manufacturing"],"extracted_tech_keywords":["AI","Aim","AWS","R","Go","Scala","Git","API","ViT","digital transformation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/will-upcoming-budget-deliver-on-indias-semiconductor-needs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":37290,"title":"5 Business Intelligence Jobs You Must Apply For Right Away","content":"Photo by Markus Spiske Analytics India Jobs offers a new platform for the data science community in India by bringing to the fore latest job openings across the country. Here are few Business Intelligence (BI) job openings to apply for right away: Analyst @ Mastercard, Gurgaon As an analyst, the candidate will Work closely with Data & Services Advanced Analytics teams and external clients around the world to architect, develop, and maintain advanced reporting and data visualisation capabilities on large volumes of data in order to support consulting projects, model development, model scoring, and campaign measurement. And, translate business requirements into tangible solution specifications and high quality, on time deliverables. Requirements B.Tech or Mathematics. M.S. preferred. 2-5 years of relevant experience in hands-on data programming, querying, data mining and report development using large volumes of granular data to deliver business intelligence and custom reporting solutions in a Microsoft SQL Server environment. Strong hands-on experience on SQL, Tableau, SSIS. Analytical\/Problem Solving. Relevant retail and payments industry experience a plus. Apply here Digital Data Analyst @ DG7 Solutions Pvt.Ltd, Mumbai DG7 Solutions is an analytics-focused, research-based consulting partner for medium and large organisations offering end-to-end Digital Marketing Solutions. Requirements Work with internal stakeholders to determine ongoing business intelligence needs in order to design, develop and produce ongoing reports and dashboards to support the business. Manage projects related to digital analytics from start to finish including data integration, report and dashboard automation, product analytics and insight, data collection, and product optimisation Assists in identifying opportunities that are most effective at driving conversions, revenue, ROI, and scale across Digital Marketing programs. Build automated dashboards for digital, marketing & customer level insights – to be shared and championed with key stakeholders throughout the business. Web Analytics: In addition to setting up standard web reports, the ideal candidate will continually mine Google Analytics Developing new skills and sharing your knowledge with the Data Team SQL, R, Python or PySpark Apply here Analyst @ Indegene Private Limited, Bangalore Indegene helps drive Effectiveness and Efficiency while Bringing Pharma Products to Market through Modern Commercial and Medical Operations by combining deep medical understanding, modern technology and flexible engagement models. Requirements B.E\/B.Tech\/BCA\/BSC Develop custom intuitive, interactive and dynamic dashboards, enabling actionable insights and self-service analytics. 2-6 years’ experience. ·Must have experience in building Reports and Dashboards using Tableau Experience in writing SQL queries to extract, merge and prepare granular and aggregated datasets for data analytics and business intelligence Experience in other BI\/Reporting tools is a plus Requires the ability to organise, manage and work within a global team-oriented, collaborative environment, providing technical consultative support to peers, projects and end users, including the development of technical standards. Understanding of pharma data (IMS, DRG, Epi Sources, APLD etc.) and basic analytics Ability to derive KPIs from data sets to create views and charts Expert in at least two BI tools like Tableau, Power BI, Qlik sense, Excel VBA Experience in handling large data using SQL \/ MS Access ML tools like Python, R etc. Apply here. Data Analyst @ Access Analytics Solutions Pvt.Ltd, Bangalore Access Analytics International provides advanced analytical tools and services for engineering and data analysis. A data analyst will manage data sources, database and movement of data and data organisation across different systems. Requirements Bachelors\/PGDM\/ Masters degree in Information Systems \/Maths\/ Statistics\/\/ Finance\/ Engineering \/Business Analytics Actuaries\/FRM\/CFA\/CQF\/PRM certification would be a plus 0-2 years of experience in database or project management Should have hands-on experience of designing database schemas SQL and at least one of R\/Python\/VBA Experience in handling large structured and unstructured datasets Exposure to tools\/platforms – Hadoop ecosystem and DB systems Good hands-on expertise in Tableau and ETL tools. Apply here Head Of AI @ SmartOnApp Technologies, Mumbai A trusted partner for Fortune 500 companies, SmartConnect Technologies has an established track record of delivering innovative solutions that help realise corporate strategies. Founded in 2010 and headquartered in Mumbai, India is a consulting-cum-IP firm that provides transformation solutions for enterprise-level companies. UNFYD®XP suite,  an integrated digital\/social transformation platform, leverage the listening\/interaction capabilities in a single platform, to enable seamless CX journey orchestration across the various touch-points. Head of AI will define and execute our Collective Intelligence strategy and assist the Research Team in evaluating products in the market. Requirements Demonstrable thought leadership in collective intelligence theory. Proven experience in designing and delivering high complexity AI \/ ML systems in production environments. An understanding of how to sequence delivery to demonstrate progress in an iterative fashion. Proven ability to deliver effectively within a high-growth environment. Experience in leading and managing small (3-12 member) cross-disciplinary teams. Apply here.","excerpt":"Analytics India Jobs offers a new platform for the data science community in India by bringing to the fore latest job openings across the country. Here are few Business Intelligence (BI) job openings to apply for right away: Analyst @ Mastercard, Gurgaon As an analyst, the candidate will Work closely with Data & Services Advanced […]","categories":["AI Hirings"],"tags":["Business Intelligence","certification for artificial intelligence","data analyst","Data Science Jobs","Tableau"],"author_name":"Ram Sagar","publish_date":"2019-04-04T07:22:36","publication_year":"2019","word_count":768,"keywords":["data science","Go","Tableau","AI","ML","Data Science Jobs","RAG","Python","certification for artificial intelligence","data analyst","analytics","SQL","Rust","Business Intelligence","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","RAG","Python","R","SQL","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/5-business-intelligence-jobs-you-must-apply-for-right-away\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10165414,"title":"First Citizens India Appoints Satya Prakash Ranjan as Country Head and Head of Technology","content":"First Citizens India has named Satya Prakash Ranjan as its new country head and head of technology. Based in Bengaluru, Ranjan will oversee India teams supporting the strategic business priorities of First Citizens Bank, the primary US banking subsidiary of First Citizens BancShares. With over 25 years of experience in technology, analytics, and consulting, Ranjan previously served as country managing director at Telstra, where he led Technology, Data, Networks, and Shared Services capability centers across Bengaluru, Pune, and Hyderabad. He has also held leadership roles at Fidelity Investments, Citigroup, and Accenture. “Satya’s extensive experience driving significant technology transformations, building global delivery capabilities, and handling large-scale programs and platforms will enable First Citizens India to continue scaling new heights,” said Ranjan Wadhwa, head of global services for First Citizens India. “We are thrilled to have him on board and look forward to the strategic leadership he’ll provide our talented organisation.” Expressing enthusiasm about his new role, Ranjan said, “I am extremely delighted to join and lead First Citizens India in this exciting phase of growth. On the heels of the company’s early success leveraging Indian talent and capabilities, I look forward to collaborating with our teams globally to drive even more productive business outcomes.” An alumnus of the Indian Institute of Technology, Delhi, Ranjan also holds an MBA from the Indian Institute of Management, Kolkata. Passionate about education, he actively supports initiatives aimed at providing educational opportunities for underprivileged children. First Citizens India, also known as FC Global Services India LLP, is a Bengaluru-based GCC and a subsidiary of First Citizens BancShares, Inc. With a 125-year legacy, First Citizens India operates as part of BancShares’ Global Services division, focusing on technology, enterprise operations, finance, cybersecurity, risk management, and credit administration.","excerpt":"Based in Bengaluru, Ranjan will oversee India teams supporting the strategic business priorities of First Citizens Bank.","categories":["AI News"],"tags":["First Citizens India"],"author_name":"Mohit Pandey","publish_date":"2025-03-07T11:19:20","publication_year":"2025","word_count":289,"keywords":["programming_languages:R","AI","R","RAG","Aim","analytics","GAN","First Citizens India"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/first-citizens-india-appoints-satya-prakash-ranjan-as-country-head-and-head-of-technology\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":20029,"title":"Flipkart Introduces New Internal AI Unit Tailor-Made For Indian Market","content":"In a move to strengthen its efforts towards furthering its interest in the field of Artificial Intelligence, Flipkart has created a new internal unit devoted to AI. The unit called AIforIndia, will be headed by Chief Data Scientist Mayur Datar. Reportedly, certain ventures within this new unit will see the direct involvement of Flipkart CEO Kalyan Krishnamurthy and Chairman Sachin Bansal. Bansal was reported saying in an interview that with regards to its deeper foray into AI, Flipkart has already started working towards building its infrastructure and forming collaborations for its hardware needs. Along with recruitment of AI experts, it is also working with premier educational institutions such as the IITs to move ahead in its AI enterprise. Bansal added that the company would be investing “hundreds of millions of dollars” for this initiative and that they were also exploring acquisitions associated with AI in both India and the US. “This is the next big thing for us, where we are betting big on the use of AI and machine learning to solve problems at Flipkart. India’s problems are unique and we need to apply AI in the ecosystem to solve Indian problems. We believe that some of the focus areas for AI in developed countries cannot be applied for India. At Flipkart, we will solve problems differently because the underlying problems are different,” Bansal said. However, Flipkart’s push towards AI is not a new one. The company has been taking steps to bolster it footing in AI for quite some time now. Myntra, Flipkart’s online fashion wing, found success in its AI-powered ‘fast fashion’ brand, Moda Rapido. It was a product of Myntra’s Rapid Project, which made the idea of clothes designed without human intervention a reality. It also reduced the time required to bring out the latest fashion trend in the market from 180 days to 45 days. The year 2017 alone has seen many AI-related advances from Flipkart, which has played a key factor in introducing online shopping in India. Earlier this year, it was reported that Flipkart had developed Project MIRA, an AI-based shopping assistant. The project that had been in works since February 2016, was developed with the intention of making the online shopping experience as personal as possible. It was also reported that the company was collaborating with Microsoft to build its capabilities to apply AI and machine learning-based solutions for easier sales process in the future. Flipkart’s competitor in this domain, Amazon, has also been investing in AI. It has already employed machine learning and AI in its Amazon Web Services worldwide. In August 2017, Amazon India was looking at hiring over 1,000 people for AI, automation and R&D.","excerpt":"In a move to strengthen its efforts towards furthering its interest in the field of Artificial Intelligence, Flipkart has created a new internal unit devoted to AI. The unit called AIforIndia, will be headed by Chief Data Scientist Mayur Datar. Reportedly, certain ventures within this new unit will see the direct involvement of Flipkart CEO […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","big data processing interview","e-commerce","Flipkart","Online shopping"],"author_name":"Jeevan Biswas","publish_date":"2017-12-21T09:58:00","publication_year":"2017","word_count":446,"keywords":["API","artificial intelligence","e-commerce","machine learning","AI","programming_languages:R","Flipkart","Online shopping","big data processing interview","Ray","automation","cloud_platforms:Amazon Web Services","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","Ray","R","API","automation","cloud_platforms:Amazon Web Services","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/flipkart-ai-for-indian-market\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10161065,"title":"Elon Musk is Using 100,000 GPUs for Grok-3 – But Why?","content":"Elon Musk, the founder and CEO of xAI, said that the company’s Grok-3 model has finished pre-training and confirmed that the model will be launched in the next 3-4 weeks. However, the model is said to have used 10 times more compute than its predecessor, Grok-2. This is quite a bold move in an era when experts suggest that LLM scaling may have hit a wall and that additional computing while training the model will only yield diminishing returns. Grok-3 will train on NVIDIA’s H100 GPUs – 100,000 of them. It has been trained on Colossus, which xAI claims is the world’s most powerful AI training system. He also said the cluster was built in 122 days from start to finish. Moreover, this is the first time a cluster of this size has been built. “Grok 3 will resolve the question of whether or not we’re hitting a wall,” Gavin Baker, CIO and managing partner at Atreides Management, said on X. Naturally, it consumes a mammoth amount of power. Yann Le Du, a physicist, calculated that training with 100k H100 GPUs is equivalent to the power consumed by 7% of a typical nuclear reactor. For a month, this consumes ~181 trillion joules of energy, which is ~10,000 times the energy consumed by a human brain over 30 years (~19 billion J at ~20W).“Is Grok’s capacity comparable to that?” he asked. That is what everyone wants to know – will the model outperform its competitors? So What is Musk’s Plan? Recently, former OpenAI chief scientist Ilya Sutskever said that pre-training with more datasets might soon be over, as they are comparable to fossil fuels and may soon be exhausted. Synthetic data could be an answer to that, he said. One of Musk’s comments earlier on an X Spaces conversation echoes this sentiment. “You really have started running into this data problem where you have to either create synthetic data or use real-world videos,” he said.“Those are the two sources of unlimited data – synthetic data and real-world videos. Tesla has a pretty big advantage in real-world video,” he added. Musk also revealed that the model will launch in Tesla’s vehicles soon. Another user on X said that Grok-3 is rumoured to be the ‘most powerful base model in existence’. Unlike OpenAI’s approach to scaling laws, it won’t use test-time computing for reasoning. “I expect Grok 3 to be a failure. Test time scaling effectively enlarges the model by a factor of 1000-10000. If they can’t do it at xAI, they have just burnt a lot of money,” said a user on X. So, is the company relying purely on a brute-force compute scaling approach? AIM reached out to a few experts in the AI industry to understand what Grok-3 is possibly trying to do and what it may achieve. The model may be using 10 times more compute power to train, but will it achieve performance gains that are directly proportional? Paras Chopra, an AI researcher and founder of Turing’s Dream, said, “Performance is often log-linear. So I’d say 10x more compute would have a ~double jump in performance over Grok 2.”We also reached out to Sudipta Biswas, who has built an architecture to enhance an AI model’s ability to interact with external data sources. He suspects that Grok-3’s architecture may be noticeably different from Grok-2’s. “Hence, pretraining is required from zero,” he said. He also suggested that Musk might be using excess computing for parallelised training: “If you are using 10x more compute, you can complete pretraining in 10x less time.” However, despite using unprecedented computing power, Musk has failed to realise his hopes of delivering the model early enough. He previously stated that he planned to reveal it in December last year, but it has yet to materialise. However, if the model arrives by next month, it will settle some of the most heated debates in the AI ecosystem. The Tale of Two Models Speaking of models pre-training on mammoth computing, China’s DeepSeek-V3 is unmissable because its approach starkly contrasts with Grok-3’s. In an earlier interview, Musk revealed that Grok-3 would finish training in three to four months, which is 100,000 GPUs x 2,880 hours. xAI used at least 200M GPU hours for training. In contrast, the DeepSeek-V3 was trained on just 2.788 million NVIDIA H800 GPU hours. On most benchmarks, the model outperformed Meta’s 405 billion parameter Llama 3.1 and even closed-source Claude 3.5 Sonnet and GPT-4o in several tests. The model is a testament to achieving superior performance without using excess computational resources. However, xAI is not alone. Meta revealed that its upcoming Llama 4 is being trained on a similar cluster size. “[This is] bigger than anything that I’ve seen reported for what others are doing,” said Meta chief Mark Zuckerberg in the company’s earnings report released in October. “By the way, you can do some pretty neat reasoning stuff with a 200k GPU cluster,” said Erik Zelikman, a member of xAI’s technical team. Again, the general concern is that these models better not waste computing power. “It would be [kind of] sad if Grok 3 used 20x compute of Grok 2 and was still mid [mediocre]. I really hope they spend time on fixing their pipeline too, and not just GPUs,” said a user on X. However, if the company is innovating at the architectural level, like DeepSeek, and is using a 100K GPU cluster – it would yield unimaginable results. But how close will the model get to OpenAI’s most powerful o3 model? We’ll soon know. That said, even before delivering Grok-3, Musk revealed that Grok-4 would be released later this year and Grok-5 the next year. Moreover, the company is also hinting at building a reasoning model, as it’s looking to hire AI engineers and researchers.","excerpt":"The model will settle some of the most heated debates in the AI ecosystem when it arrives.","categories":["Global Tech"],"tags":["Elon Musk","grok","NVIDIA GPU"],"author_name":"Supreeth Koundinya","publish_date":"2025-01-09T18:16:25","publication_year":"2025","word_count":963,"keywords":["Grok 3","NVIDIA GPU","OpenAI","AI","AWS","GPT-4o","Llama 4","R","Elon Musk","Aim","grok","Claude 3.5","xAI"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","Claude 3.5","Llama 4","Grok 3","xAI","Aim","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/elon-musk-is-using-100000-gpus-for-grok-3-but-why\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10138238,"title":"Wait, What? The Bible, Bhagavad Gita, and Preamble are All AI Generated?","content":"As writers, our early fears revolved around plagiarism checkers, making sure that every word we wrote was authentically ours. But now, with AI in the picture, the challenge has shifted. Companies and clients aren’t just worried about the content being original anymore; they’re extremely wary of AI-generated writing. Fair enough, there’s always a demand for human-crafted stories. But how reliable are these AI detectors in proving the same? Lately, many freelance writers have taken to LinkedIn, frustrated by the inaccurate results of these tools. One writer shared his experience on the platform and said, “ZeroGPT and Copyleaks are scamming people. No AI detector is accurate.” It’s a conversation that’s only getting started. Source: Reddit Source: LinkedIn If we were to trust these tools, we’d believe that even our Holy Books are AI-generated! A LinkedIn user posted that the Bible is “97% AI-generated”, but considering that the earliest writings in the Bible date back to around 1400 BC, written by communities of scribes, the tools’ accuracy is a question. Source: LinkedIn Curious to know how the tools work, we added the Bhagavad Gita content, and here’s the result. So, the Bhagavad Gita, part of the Mahabharata, written at some point between 400 BCE and 200 CE by the sage Ved Vyasa is also credited to AI. There’s more, even the Preamble of the Indian Constitution is supposedly AI-generated, according to these flawed AI detectors. But Why Do AI Detectors Go Wrong? While AI detectors can be considered a helpful starting point for assessing content authenticity, they can’t guarantee 100% accuracy. This is because these tools, still in their infancy, are evolving to match the sophistication of modern AI writing. Each detector varies in effectiveness because they rely on different training datasets. Also, the line between AI-generated and human-written content is becoming increasingly blurred. Advanced AI writing tools now produce content that closely resembles human writing, making it a task for detectors to distinguish between the two. So, to truly understand the reliability of these tools, it’s important to know how they operate. The detectors are typically trained on datasets containing both human and AI-generated text. By analysing this data, they identify specific characteristics that are more likely to appear in AI-written content. Experts noted that the two major traits these tools focus on are perplexity and burstiness. Perplexity measures the unpredictability of the content. AI-generated text usually has lower perplexity, making it more predictable, while human writing tends to be more varied and less predictable. Burstiness refers to the variation in sentence length and structure. AI content tends to be more uniform and steady while human writing often exhibits a mix of long and short sentences. Other qualities that detectors flag as AI include the repetition of words, repetitive sentence structures, a generic tone, and more. Companies Want to Move Away from AI Content There are a few reasons why clients and companies are very particular about their content. Data suggests that the entire process – from writing to publication – can be completed in as little as 16 minutes with AI, compared to an average of 69 minutes for humans. Despite this efficiency, the question of trust remains significant for organisations and customers. According to the Hootsuite Social Media Consumer 2024 Survey, 62% of consumers are less likely to engage with or trust content if they know it is AI-generated content. It further reflects on the traffic or engagement of the content too. Source: NP Digital Neil Patel, the co-founder at Neil Patel Digital, noted, “If everyone uses AI to create content, and if AI uses data already available to develop these projects, we won’t have anything new and authentic anymore.” It appears that only marketing professionals are confident in AI, while consumers remain sceptical. How to Solve This? Freelance writers have a few suggestions for the clients. ​According to them, there must be an experienced content editor or manager to run through the content and decide if it reads ‘AI-written’, suggest the changes, tweak, if needed, and then push it through the AI detectors (preferably multiple, to compare the results and conclude whether it’s human or AI-written). There is a worry that AI-generated content won’t rank on Google. It’s important to note that mobile disparities, website quality issues, accidental content alterations, and technical SEO challenges can also negatively impact traffic. So, it’s not just AI-generated content that’s sending your views crashing. As the entire AI system is new and evolving every day, the clients and companies need to exercise a bit of patience before rushing to conclusions and rejecting writers. The technology is still finding its feet, and quick judgments could mean overlooking real human talent in the process.","excerpt":"So, the question is which AI detector is accurate in today’s age.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI Tool","Editors Picks"],"author_name":"Vidyashree Srinivas","publish_date":"2024-10-14T12:53:08","publication_year":"2024","word_count":780,"keywords":["Go","AI writing","AI","Git","RAG","GPT","Editors Picks","Rust","GAN","AI Tool","R","AI-generated content","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","Git","GPT","GAN","AI-generated content","AI writing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/wait-what-the-bible-bhagavad-gita-and-preamble-are-all-ai-generated\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":6321,"title":"Big Data is here to help banks in their #1 problem&#8230;","content":"What is the key issue facing global banking industry today? Ask this question to the CEO’s of Top 50 global banks and it is highly likely that their answer will be the same – managing balance sheet risks and effective utilization of Economic Capital! While this problem is not new now and banking industry has already seen the need for more stringent capital management standards, there is no practical solution in sight! As a banker or a senior leader in a bank, you may have provisioned for enough capital or your bank may ace the stress testing requirements on annual basis, but does that mean you are managing your risks and capital effectively? Is your portfolio risk optimized? Are you recovering the cost of your risks from your customers or are you passing on the risks that you should not have on your balance sheet effectively? Answers to all these questions are unknown (or at least most bankers are not confident about them) today because of many reasons. Firstly, it is not possible to measure risk from different asset classes on a common platform. Secondly, the complexities of banking transactions are ever increasing and with that increases the model risks or the risks of going wrong with your assumptions in quantifying the risk! While all these problems continue to persist, one thing that has become extremely important for banking industry is timeliness of information regarding risk and capital. How soon can a Chief Risk Officer (CRO) act on the available information? How soon can a CRO find out that a certain part of his bank’s portfolio is fast becoming toxic asset? How soon can a bank leadership team find out that some of their top performing loans are quickly turning into a nightmare? How soon can they act on hedging a certain loan? And how soon can they recover the cost of hedging such risk from the customer? Any possible answer to any of these questions would have been a post-mortem 2 years back. It would have required tons of analysis from the Risk Management gurus of the bank and would have taken them ages to find out the right path of action. By the time they take a decision and act, the losses would have already hit their balance sheet. However, it is possible to visualize such scenarios today – with help of Big Data Analytics. It is possible to create a risk (by lending money or investing in certain securities) and hedging them (with almost no incremental cost to the bank) on real-time basis using Big Data Analytics. Risk Adjusted Performance Management – a tool that many banks wish they had before 2008 recession, is available today. With Big Data Analytics, banks can build platforms that give the power to their front-line staff to manage the balance sheet risk. The classic paradigm in banking is when a loan sales officer has to balance between the customer requirements and the risks the bank can take. In past, if sanctioning a loan meant the bank increased its portfolio risk and in turn its economic capital requirements, it either chose to not issue the loan or hedge it afterwards – eventually putting pressure on its margins due to increasing costs of hedging. However, using a Big Data platform, banks can adapt to Risk Adjusted Performance Measurement for their front-line sales or underwriting officers. Consider this scenario: A loan sales officer of a large bank is sitting in front of an existing client. This client – an automobile giant – has already borrowed a couple of billion dollars from the bank. This exposure has already increased the concentration or correlation risk for the bank’s portfolio (because of many such loans issued to other automobile companies too) and is pressurizing the margins of the bank. The client is asking for an additional loan facility of half a billion dollars more for expansion of its business in Asian markets. The sales officer knows that the credit team of the bank will not approve this because of the incremental capital requirement to match the incremental risk or the hedging costs of the loan would reduce the net spread significantly. He pulls out his handheld phone and opens an application provided by Office of the CRO of the bank. He feeds in the details of the client and his requirements in the application and hits a button to run simulations. His phone connects to the servers of the bank located some 1,000 miles away and in less than a minute returns multiple scenarios. One of the scenario tells the sales officer that the bank will require x million dollars as incremental economic capital and will reduce the net spread of the transaction by a certain basis points. The second scenario tells him that the loan can be hedged using a Credit Default Swap (CDS) at a cost of y%. However, if he chooses to hedge the loan, the total spread (or interest rate in simple terms) that he should charge the client is z% instead of the normal r% interest offered by the bank. If the bank policy requires the sales officer to generate a certain margin on a risk-adjusted basis (performance management for staff) and the z% interest and the net spread of the transaction falls within this threshold, he can go ahead with the loan. However, if the threshold is not met even after hedging the transaction, he can let go the deal. While banks and bankers may not like to let go deals, there is no choice. A risk adjusted performance management is the need of the hour – and it can be implemented with Big Data solutions. This will help the banks in improving: their credit quality or asset quality their net spread on assets (net profit of the bank in other words) reduce their balance sheet risk manage their economic capital effectively ace the stress testing requirements (not only statutory but also internal) and give confidence to the leadership team of the bank that they are far away from the risk of a sudden shock in their asset book All these on real-time basis – possible! Big Data Analytics can make it possible – and it’s not a big change management exercise. It’s about deploying the front-end technology using the back-end risk engines that banks already use today. The business case for the investment may show a net loss for first couple of years, but will benefit the banks over period and will certainly improve the functioning of overall banking industry. Risk Adjusted Performance Management is the only way out – and it is possible – with Big Data!","excerpt":"What is the key issue facing global banking industry today? Ask this question to the CEO’s of Top 50 global banks and it is highly likely that their answer will be the same – managing balance sheet risks and effective utilization of Economic Capital! While this problem is not new now and banking industry has […]","categories":["IT Services"],"tags":[],"author_name":"Kushal Shah","publish_date":"2014-10-17T16:13:24","publication_year":"2014","word_count":1108,"keywords":["big data","Go","API","programming_languages:R","AI","programming_languages:Go","analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","API","big data","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/big-data-help-banks-1-problem\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10085015,"title":"Wipro Revenue Reaches $232.3 Bn, Attrition Plummets to 22.7%","content":"Indian IT software services company Wipro reported that gross revenue reached $232.3 billion, an increase of 14.4%  and 3.1% quarter-on-quarter (QoQ) in the third quarter of the current fiscal year, 2023. IT service revenue increased by 10.4% during the same time period. In Q2 FY23, the company performed poorly, with profit decreasing by 9.3% and 13% revenue growth, year-on-year. The company’s net income in Q3 increased 2.8% YoY to Rs 3,050 crore. Constant currency revenue from the IT Services division is expected to range from 11.5% to 12 % for the entire year. IT Services Segment Revenue rose by 6.2% YoY to $2,803.5 million. Also Read: Infosys Profit Up by 13%, Attrition Drops to 24% in Q3 The quarter’s net income increased by 14.8% QoQ and 2.8% YoY to $369.1 million. Earnings per share for the quarter were $5.57 ($0.071), up 2.6% YoY and 14.6% QoQ. Operating Cash Flows for the quarter were $43.5 billion ($526.0 million1), up 44.7% year over year, or 142.5% of Net Income. Voluntary attrition decreased 180 bps from the previous quarter to reach 21.2% for the trailing 12 months. Total bookings rose by 26%, and large deal bookings by 69%. Commenting on this, CEO and MD Thierry Delaporte said, “We improved our margins by 120 basis points, and our attrition moderated for the fourth quarter in a row. We continue gaining market share due to deepening client relationships and higher win rates.” The top five clients grew 15.7% YoY, while the top 10 clients grew 14.7% YoY in constant currency terms, underscoring deepening relationships with top strategic clients. Wipro also announced an interim dividend of ₹1 ($0.0121) per equity share\/ADS At Rs 4,350 crore ($526 million), or 142.5% of net income for the quarter, operating cash flow increased by 44.7% year over year. Attrition Drops, 435 Employees Vanished The attrition rate has declined YoY to 22.7% in Q3FY22. The total employee count for the company dropped by 435 to 258,744 in Q3 FY23 from 259,179 in the previous quarter. However, the attrition rate has decreased from 23.0% in Q2FY23 to 21.2% in Q3 FY23.","excerpt":"Wipro’s net income in Q3 increased 2.8% YoY to Rs 3,050 crore.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","Deep Learning","Machine Learning"],"author_name":"Shritama Saha","publish_date":"2023-01-13T19:03:27","publication_year":"2023","word_count":349,"keywords":["programming_languages:R","AI","Machine Learning","Deep Learning","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-revenue-reaches-232-3-bn-attrition-plummets-to-22-7\/","complexity_score":2,"technical_depth":3,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":65199,"title":"VMware Wants To Acquire Octarine to Expand Workload Security Solution into Kubernetes","content":"Octarine acquisition will bring intrinsic security to containerized applications running in Kubernetes and build security capabilities into the fabric of the existing IT and DevOps ecosystems. VMware made two key announcements during Connect 2020, the company’s annual cybersecurity user and partner conference (hosted virtually by VMware Carbon Black) The intent to acquire Octarine, whose innovative security platform for Kubernetes applications helps simplify DevSecOps and enables cloud-native environments to be intrinsically secure, from development through runtime. The creation of a Next-Gen SOC Alliance along with Splunk, IBM Security, Google Cloud’s Chronicle, Exabeam, and Sumo Logic. The alliance empowers SOC teams with visibility, prevention, detection and response capabilities that can uniquely leverage the VMware fabric. Intent to Acquire Octarine to Bring Intrinsic Security to Containers & Kubernetes Protecting workloads is critical to the security of applications and data inside every organization. The unique properties of the cloud (speed, agility, scale) mean that developers are increasingly using Kubernetes and containers to modernize applications and changing the nature of workloads that need to be secured. As with any major technology adoption, attackers are not far behind, looking to take advantage of new risk areas. Building Octarine’s innovative Kubernetes security platform into the VMware security portfolio presents a major opportunity for VMware to further mitigate risks in several ways: Provide full visibility into cloud-native environments so customers can better identify and reduce the risks posed by vulnerabilities and attacks.Move beyond static analysis and maintain compliance – customers can create and enforce content-based policies to protect the privacy and integrity of sensitive and regulated information. Integrate into the developer lifecycle to analyze and control application risks before they are deployed into production.Run alongside service mesh frameworks such as Tanzu Service Mesh to provide native anomaly detection and threat monitoring for cloud and container-based workloads.Provide runtime monitoring and control of Kubernetes workloads across hybrid environments for threat detection and response. Following the acquisition’s close, the Octarine technology will be embedded into the VMware Carbon Black Cloud, providing new support of security features for containerized applications running in Kubernetes and enable security capabilities as part of the fabric of the existing IT and DevOps ecosystems. This innovation will further reduce the need for additional sensors in the stack. Octarine capabilities will also integrate and leverage the VMware Tanzu platform, including current investments in Service Mesh and Open Policy Agent. “Acquiring Octarine will enable us to further expand VMware’s intrinsic security strategy to containers and Kubernetes environments by embedding the Octarine technology into the VMware Carbon Black Cloud,” said Patrick Morley, general manager and senior vice president, Security Business Unit, VMware. “This, combined with native integrations with Tanzu, vSphere, NSX and VMware Cloud Foundation, will create what we believe is a unique and compelling solution for intrinsically securing workloads. And, with the addition of our AppDefense capabilities merged into the platform, we can fundamentally transform how workloads are better secured.” VMware Launches Next-Gen SOC Alliance In an effort to empower modern SOC teams with the capabilities and context they need to greatly improve both their efficiency and efficacy, VMware has launched a Next-Gen SOC Alliance. The alliance features Splunk, IBM Security, Google Cloud’s Chronicle, Exabeam, and Sumo Logic integrations with the VMware Carbon Black Cloud to deliver key XDR capabilities and context into SIEM technologies that power the modern-day SOC. Equally important, the combined solutions will be able to take advantage of VMware’s Intrinsic Security framework and enable SOC teams to: Leverage the VMware fabric – doing away with many of agents and appliances SOCs would normally deploy for visibility, prevention, detection, and response.Gain far richer context about the infrastructure and applications being protected.Operationalize more of security with (and through) IT via Carbon Black integrations with VMware management consoles. Other key benefits from the Next-Gen SOC Alliance include: Centralized security context, which enables organizations to detect, analyze and respond to data in a unified SIEM at machine speed.Automation and orchestration tools that combine with XDR capabilities to allow SOCs to scale and standardize their investigation and response processes.Actionable answers to large-scale queries and remote remediation from within Splunk, IBM Security, Google Cloud’s Chronicle, Exabeam, and Sumo Logic platforms. “The Next-Gen SOC Alliance brings a critical mass of XDR context and capabilities to SOCs in a fully intrinsic way – one that can uniquely leverage the VMware fabric,” said Tom Barsi, Vice President of Alliances for VMware Carbon Black. “In partnership with the industry’s leading SIEM\/SOAR players, we’re setting a strong vision for the modern SOC and delivering unprecedented visibility and remediation capabilities across endpoints, networks, workloads, and containers.” VMware’s intrinsic security strategy is centered on enriching context from across the security portfolio and leveraging the VMware fabric for native telemetry and control at the endpoint, workload, network, user access point, and application. This innovation enables a true XDR solution that works out of the box with existing VMware solutions – reducing all the bolt-on sensors and appliances that plague security.","excerpt":"Octarine acquisition will bring intrinsic security to containerized applications running in Kubernetes and build security capabilities into the fabric of the existing IT and DevOps ecosystems. VMware made two key announcements during Connect 2020, the company’s annual cybersecurity user and partner conference (hosted virtually by VMware Carbon Black) The intent to acquire Octarine, whose innovative […]","categories":["AI News"],"tags":["devop"],"author_name":"Vishal Chawla","publish_date":"2020-05-14T13:38:43","publication_year":"2020","word_count":821,"keywords":["Go","AI","R","innovation","RAG","automation","anomaly detection","GAN","DevOps","devop","kubernetes"],"extracted_tech_keywords":["AI","RAG","anomaly detection","kubernetes","R","Go","DevOps","GAN","automation","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vmware-wants-to-acquire-octarine-to-expand-workload-security-solution-into-kubernetes\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10008258,"title":"How This Bangalore Based Startup Is Driving Innovation With Quantum Technology-Based Products","content":"India has been a frontrunner when it comes to implementing new-age technology such as AI, machine learning and quantum technologies. In fact, Union Budget 2020 saw an allocation of INR 8,000 crore towards the development of technologies such as Quantum Cryptography and Quantum Communication. Further building on quantum technology, and with a vision to drive disruptive innovations across multiple sectors with AI and quantum technology, Bengaluru-based Archeron Group is providing cutting-edge solutions for multiple industries. Analytics India Magazine got in touch with the founder of Archeron Group to understand the tech behind it. Founded in 2015 by Aviruk Chakraborty, Archeron Group was established with a vision to drive disruptive innovations across multiple sectors that can help transform the world for the better. The company is co-headquartered in Abu Dhabi and San Francisco. It has a state-of-the-art Global Development and Delivery Centre (GDDC) in Bengaluru. It holds strategic importance for the Group and is responsible for creating the entire solution portfolio of the groups which it has been able to take to the Middle East (UAE and KSA), North America, and the European Union. AI & Quantum Computing Is Central To Archeron Archeon extensively integrates AI and quantum computing in its flagship products. Chakraborty explained them as below: 1| Automated and Remote Sensing Agri Platform: This is an Agri analysis platform that uses remote sensing-based technologies to map the yield, the productivity of that field not only for the past 20 years but also to predict the next year’s yield and productivity. Chakraborty stated, “We have developed a vast array of solutions which include agricultural insurance automation, agriculture loan automation, crop classification and identification, soil analysis, fertilizer and pesticide requirement analysis, disease detection and real-time monitoring of the field, to name a few.” 2| Bank\/ NBFC automation using AI: The company is building a Quantum AI Bank using quantum technologies and artificial intelligence, which will be completely autonomous in its decision making. According to Chakraborty, the larger parameters of the bank such as risk ratios and macroscopic directives will be set by the board on a quarterly basis which gets translated into algorithmic performance parameters and hence executed for the next quarter. 3| Predictive Diagnostic Platforms: Archeron group has created an integrated national radiology platform using Deep Convolutional Neural Networks, where the radiological plates of all the patients all over the country is analysed and a diagnostic support system is created in which the doctors are given a second opinion on the radiological plates. 4| Quantum Cryptography: The company has designed Quantum Cryptography with a view to strengthen the payments infrastructure in India and make them 100% secure and unhackable. They used three-factor authentication as well as Quantum One Time pad to offer an end to end secure platform. On being asked how the products are different from others in the market, Chakraborty pointed out that- AI bank solutions at the company leverage an integrated end to end model using ML, whereas most players are in the current stage of only developing digital banks with the processes being human decision-driven in the backend.Archeron Group uses Quantum One Time Pads and Quantum Key Distribution (QKD) frameworks along with their own proprietary mathematical cryptographic framework to create an unhackable communication protocol. The Tech Behind It The company is using deep neural networks and cutting edge mathematical models such as Q-learning and GAN. They also use Remote Sensing, IoT, and Crispr-Cas9. Chakraborty said, “We are language agnostic developers as we have to not only develop solutions but also integrate them with the existing framework for existing clients.” Archeron Group is using C++\/ Python for the ML algorithms and prefers to build the ML models from scratch rather than using a pre-trained model. They also work on the domains of blockchain, satellite imagery analysis, IoT, brain-computer interface, genetic programming for creating synthetic life along with standard machine learning and quantum computing, cryptography and communication frameworks. Future Roadmap Chakraborty said that in the next five years, the focus would be on increasing the adoption of healthcare, banking and agricultural solutions in UAE, India and the USA. The company is also focusing on global implementation of already developed technology and iteratively refining it to make the tech stack better.","excerpt":"India has been a frontrunner when it comes to implementing new-age technology such as AI, machine learning and quantum technologies. In fact, Union Budget 2020 saw an allocation of INR 8,000 crore towards the development of technologies such as Quantum Cryptography and Quantum Communication. Further building on quantum technology, and with a vision to drive […]","categories":["AI Startups"],"tags":["Quantum Computing","quantum computing programming","quantum technology","Startups"],"author_name":"Ambika Choudhury","publish_date":"2020-09-25T10:00:24","publication_year":"2020","word_count":702,"keywords":["Quantum Computing","artificial intelligence","machine learning","AI","neural network","ML","RAG","Python","Ray","analytics","quantum computing programming","Startups","R","quantum technology"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","analytics","Ray","RAG","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/how-this-bangalore-based-startup-is-driving-innovation-with-quantum-technology-based-products\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10066535,"title":"Council Post: How conversational AI is powering modern enterprises","content":"Chatbots have become an integral part of modern enterprises. For example, Mahindra has been working with Amazon Alexa for four years now. Mahindra XUV 700 was the first vehicle in India to integrate Alexa. The customers can use the Mahindra Skill for Alexa to manage vehicle information such as fuel level or tire pressure, and remotely give commands for locking or unlocking the car door, preparing the cabin temperature, etc. “Hey Alexa. Tell me when you see other Mahindra cars nearby…I like to hang out with my family.” https:\/\/t.co\/zEdSCEcfWn— anand mahindra (@anandmahindra) February 15, 2022 Earlier, ICICI Prudential Life launched a voice chatbot called ‘LiGo‘ on Google Assistant to handle insurance-related queries. Amazon’s Alexa, Google Assistant and Apple’s Siri are leading the way for conversational AI for general\/customer-focused use cases (B2C use-cases). In terms of research and advanced language modelling, industry and businesses specific (B2B) chatbots have remained largely rule-based or menu-driven. However, the conversational AI landscape has been undergoing a shift in the last few years. At the Google I\/O conference last year, Sundar Pichai said the company is exploring ways to cater to the developer and enterprise customers. Further, he said the language model for dialogue application (LaMDA) would be a huge step forward in natural conversation. RASA is an open source conversational AI platform that help adapt generalised language pipelines layered with injected vocabularies – albeit in a convoluted fashion; meaning, developers are left to do the heavy lifting. While tech giants and open source initiatives  are trying their best to push conversational AI solutions and products to enterprise customers, most of them are limited to supporting or assisting teams to a certain level. Generalised platforms with probabilistic language models have largely remained in the periphery use-cases ( or trophy use-cases if we are a bit unkind). The customisation of chatbots play a critical role in enhancing the end-user experience. Quick training and deployment of chatbots with minimal downtime is also key. Organisations use multilingual queries, speech-to-text methods, and NLP techniques to build such products. Towards hyper-personalised solutions Many companies today are looking at developing chatbots to address end customers’ problems in real-time and offer them a seamless experience. BFSI, retail, agritech, and fintech, are major adopters of AI chatbots. In 2020, Axis Bank collaborated with Vernacular.ai to develop an AI solution and automated customer interactions via an intelligent human-like dialogue. Similarly, HDFC Bank uses a chatbot called ‘EVA‘ to address customer queries. ICICI Bank uses iPal, which lets customers check account balances, credit card dates, etc. Tata Capital, in partnership with Yellow.ai, developed TIA charbot to answer loan related queries. The core challenges faced by companies investing in chatbots include: High infrastructure cost High-end skill requirements Ease of access by the users Complex languages and dialects within each state in the country Lack of training datasets A final thought Today, the conversational AI market remains fragmented to a large extent. While creating chatbots have become easier, their impact is still not quantified. To that, enterprises are looking to develop specialised solutions by taking advantage of domain knowledge, and deep learning and ML skills to build products that understand customers and automates tasks seamlessly. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill the form here.","excerpt":"Today, the conversational AI market remains fragmented to a large extent.","categories":["AI Features"],"tags":["Conversational AI","conversational ai platforms"],"author_name":"Ashwin Swarup","publish_date":"2022-05-09T18:00:00","publication_year":"2022","word_count":570,"keywords":["data science","AI","chatbots","ML","RAG","NLP","Aim","deep learning","Conversational AI","analytics","conversational ai platforms","R"],"extracted_tech_keywords":["AI","ML","deep learning","NLP","data science","analytics","Aim","RAG","chatbots","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/council-post-how-conversational-ai-is-powering-modern-enterprises\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10121756,"title":"Meet the College Dropout Who Helped OpenAI Build Sora","content":"OpenAI is always on the hunt for young exceptional talent. Just like the company hired Prafulla Dhariwal, who helped CEO Sam Altman create GPT-4o, OpenAI also brought onboard 20-year-old Will DePue, a college dropout, who was one of the major contributors to Sora, the text-to-video tool that generates life-like, hyper-realistic video footage. Last week, DePue turned 21 and said, “I want to be careful to have a healthy relationship with age (something tech is famously bad at). Just trying to compare where I’m now to a year\/two years ago, I do think that I need to take care in continuing to push myself to grow as much as I’ve done in the past.” DePue joined OpenAI 11 months ago as a member of the technical staff with OpenAI’s residency programme where he worked on jailbreaking and prompt injection mitigation, along with model capability evaluations and fine-tuning. He is possibly joining 1X, a robotics company funded by OpenAI, Tiger Global, Everon, and NVIDIA. I love this video. You're one of the first <10,000 people in human history to experience externally anthropomorphically interacting with yourself pic.twitter.com\/ePZWjdDDIK— Mehran Jalali (@mehran__jalali) May 23, 2024 In a podcast, DePue said that he joined OpenAI by basically texting every one of his connections. “I don’t have a traditional background. I just dropped out of college, built startups, and started working on projects then,” said DePue, adding that he did not care about which position he joined as long as he got to learn something. One of the stand out projects that DePue worked on was WebGPT, where he built a package to run GPT models entirely on the Chrome browser. He credited Andrej Karpathy for teaching him about Transformers, which he had no idea about before starting the project. The dropout prodigy Even though DePue was always into tech, he said that he was mostly interested in hardware. He never started coding until he turned 17. One of his friends predicted the startup market crash in 2008, which led him to move to Argentina. He stayed there for five years. Born in Seattle, DePue went to school at Geffen Academy at UCLA and later went on to study BS in computer science at the University of Michigan, from where he eventually dropped out to start building things. He previously founded DeepResearch and Thrive.fyi, a Discord bot. Homeschooled until the sixth grade, DePue said that he did not like the idea of school much. “You sit down all the time and learn from a board, don’t move, don’t yell at kids, and don’t poke your friends, it is not something that people are super good at,” he said. However, he does admit that schools are really important. “I don’t think I was doing it for the right reasons,” he said. DePue was also an Eagle Scout with the Boy Scouts of America. DePue believes that one of the most important things is relative competition. “We all started in different fields,” DePue explained, saying that he started with building Discord and Minecraft servers, and his friends who were doing something else, and are now running Series A and above companies. With a massive following on X and building the computer on Figma, DePue claims to know what people want, but says that he does not like that he knows it too well because it messes with what he actually wants to build. ‘The perception of success and the reality of success are disconnected’ In June 2023, DePue announced Alexandria: Project Tenet, an open source project to embed all human belief. It was consistent with 10+ religious texts with over 15 million tokens. DePue said that a friend of his dad built a chat app with GPT-3 when it was launched two years ago. He sat all night talking to it and his mind was blown – this is what got him into AI. One of the reasons why DePue is probably moving on to robotics company 1X from OpenAI is his belief that no one should commit to working on a single thing for a very long time and explore more options at a young age. He suggests people to get into coding and work on something they love instead of a very ambitious project or field whose idea they got from some influencer on social media. Currently, DePue has started a one-month-long hackathon called ‘10% project’ where participants should commit to a single project for 36 days. So, when Altman said that OpenAI is not run by a bunch of 24-year-old programmers, it was true. It’s run by younger folks! “In AI, at least in my opinion, the real 30 under 30 are those you’ve never heard of. They typically exist five layers down the organisational chart from the CEO.” Karpathy couldn’t have been more right.","excerpt":"He is now possibly joining 1X, a robotics company funded by OpenAI, Tiger Global, Everon, and NVIDIA.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Mohit Pandey","publish_date":"2024-05-27T16:47:29","publication_year":"2024","word_count":799,"keywords":["Go","OpenAI","AI","GPT-4o","RPA","Transformers","GPT","Aim","GAN","R"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","Aim","Transformers","R","Go","GPT","GAN","RPA"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/meet-the-college-dropout-who-helped-openai-build-sora\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10063803,"title":"A beginner&#8217;s guide to text regression with AutoKeras","content":"In natural language processing, we mainly find the use cases of text classification. There are various processes like sentiment analysis, positive and negative review analysis that is purely dependent on the classification modelling. There can be various situations where we may also require regression modelling on the text data. Applying regression modelling on text data can be called a text regression process. In this article, we are going to discuss text regression and how we can implement it. The major points to be discussed in the article are listed below. Table of contents What is text regression Implementing text regressionImporting and preprocessing data Text regression modelPredicting values Let’s start with understanding the text regression. What is text regression? We can think of text regression as a method of using attributes from the text data as a covariate in regression models. There are various fields where we may require regression analysis methods such as predicting salary based on the text where work requirement is mentioned or views on any website based on the content written on the website. The basic difference between text classification and text regression is the target variable.  Where we find categorical information in the classification data and ordinal data in the regression data. In this article, we aim to learn how we can perform this method of machine learning or specifically natural language processing. Let’s start the implementation of text regression. Are you looking for for a complete repository of Python libraries used in data science, check out here. Implementing text regression For implementing a model for text regression we are required to acquire data. For this article, we are using the IMDB data set. Let’s start the implementation of text regression by importing and pre[rocesssing of the data. Importing and preprocessing data Since it is difficult to get real-life data, we are going to use standard data for natural language processing which is our IMDB dataset. We all know that the IMDB dataset is classification data so we will be required to convert it into regression data. For converting it we will treat the 0 and 1 values as numerical values and make them our target variable. We can find the data here. Let’s download the data. import tensorflow as tf IMDB = tf.keras.utils.get_file( fname=\"aclImdb.tar.gz\", origin=\"http:\/\/ai.stanford.edu\/~amaas\/data\/sentiment\/aclImdb_v1.tar.gz\", extract=True, ) Output: Setting the path to IMDB data. import os IMDB_DIR = os.path.join(os.path.dirname(IMDB), \"aclImdb\") Extracting test and training files from the downloaded data. from sklearn.datasets import load_files labels = [\"pos\", \"neg\"] train = load_files( os.path.join(IMDB_DIR, \"train\"), shuffle=True, categories=labels ) test = load_files( os.path.join(IMDB_DIR, \"test\"), shuffle=False, categories=labels ) Splitting the dataset x_train = np.array(train.data) y_train = np.array(train.target) x_test = np.array(test.data) y_test = np.array(test.target) Let’s check the data print(x_train.shape) print(y_train.shape) print(x_train[10][:50]) Output: Here our data collection and data preprocessing are completed. Now we are ready to set our regression model on the data. Text regression Model In this article, we are going to use a class for the AutoKeras library. More information about the AutoKeras can be found here. The class TextRegressor under this library provides us with a facility to perform text regression on text data. We can install this library in our environment using the following lines of code. !pip install autokeras Once the installation is completed we can use the class by importing it. Let’s see how we can do it. import autokeras from autokeras import TextRegressor text_reg = TextRegressor(overwrite=True, max_trials=1) In the above code, we call the TextRegressor class from the library and define a model instance that will try on 10 different regression models and fit the training data. Let’s fit the model with 5 epochs. text_reg.fit(x_train, y_train, epochs=5) Output: The above output is output while data is getting fitted on the model. One thing that is very important about AutoKeras is it helps us in getting an optimized model by using hyperparameter tuning and model testing. It is a library that provides features for automated machine learning. We don’t need to specify epochs; it can automatically detect and adapt the number of epochs. Now the outcome of this model fitting will look as follows: The above output represents the validation loss and mean_squared_error at every epoch. Predicting values After fitting the model we are ready to predict on test data using the optimized model. y_pred = reg.predict(x_test) print(y_pred) Output: In the above, we can see that the values in the prediction array we have are float values. We can check what are the values we have in the test set. print(y_test) Output: Here we can see that the values we have in the test are in binary form. As we have discussed the model assumed that the target is numerical values and predicted accordingly as we do in regression modelling. Final words In this article, we discussed text regression which is a method of performing regression analysis on text data and we looked at a class from the AutoKeras library that helped us in performing text regression very easily and accurately. References Link for the codes AutoKeras documentation","excerpt":"We can think of text regression as a method of using attributes from the text data as a covariate in regression models. There are various fields where we may require regression analysis methods such as predicting salary based on the text where work requirement is mentioned or views on any website based on the content written on the website. The basic difference between text classification and text regression is the target variable.","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Data Science","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2022-03-29T16:00:00","publication_year":"2022","word_count":833,"keywords":["data science","text classification","machine learning","Keras","TPU","AI","sentiment analysis","Machine Learning","Ray","Aim","Data Science","TensorFlow","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","data science","Aim","Ray","TensorFlow","Keras","sentiment analysis","text classification","TPU"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-beginners-guide-to-text-regression-with-autokeras\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10016952,"title":"Guide To Darknet Installation On Windows 10 – CPU Version","content":"Darknet is a framework for real-time object detection. Darknet is an open-source neural network framework and is written in C and CUDA. A major reason it is used widely is that it is highly accurate and very fast. The reason for Darknet to be fast is because it is written in C and CUDA. Darknet supports CPU and GPU computations so it integrates better with your system and works accordingly. If you are reading this, it means you are facing issues, the same problem which many Windows 10 users have faced. Already complications of Deep Learning are not enough we windows users have to face this new problem. Installing Darknet on the Windows system. And if this is not enough you will not find any proper YouTube videos or website which will tell you what to do and how. After spending different tutorials, from different sources and failing several times, I finally succeeded. So here are the simple steps I followed to install Darknet in my windows 10 system. Step 1: This is the most common step you will find in any tutorial. Clone Darknet git repo for windows by AlexeyAB. If you have git installed you can open command prompt and run command: git clone  https:\/\/github.com\/AlexeyAB\/darknet If you don’t have git installed you can open above link and download zip and unzip (screenshot shared below) Step 2: Open the Makefile in unzipped\/downloaded folder darknet-master and change OPENCV=1 and save it. Step 3:     Go to https:\/\/sourceforge.net\/projects\/opencvlibrary\/files\/opencv-win\/3.3.0\/opencv-3.3.0-vc14.exe\/download  and Download OpenCV Create a folder named opencv_3.0 in C drive, double click the .exe file downloaded and unzip it in opencv_3.0 folder. After this step add the following path to Environment  Variables (Just for sanity check, copy these paths from your system.) C:\\opencv_3.0\\opencv\\build\\include C:\\opencv_3.0\\opencv\\build\\x64\\vc14\\lib C:\\opencv_3.0\\opencv\\build\\x64\\vc14\\bin Step 4: Download and Install Microsoft Visual Studio Community Edition 2019 from following link: https:\/\/visualstudio.microsoft.com\/ Select Python development and Desktop development with C++ package too. Now this will take a while to finish Installing and it will take quite some Gigs on C Drive. To save some space in C Drive, go to Installation location and you can select different Drive for top two options: You can see in below I have changed path to G drive. Step 5: Go to Darknet\\build\\darknet open darknet_no_gpu.sln file with Visual Studio When you open, Visual Studio will ask you to download two more dependencies. This will take a little more time to complete. Change Debug option to Release Then click the .sln file once in Solution Explorer window so that it is selected (make sure it is collapsed) Next go to Build option and click Build darknet_no_gpu Process will end showing following message: ========== Build: 1 succeeded, 0 failed, 0 up-to-date, 0 skipped ========== Step 6: Now go to darknet\\build\\darknet\\x64 folder and you will see darknet_no_gpu.exe has been created. If you have successfully completed this point. Congratulations have successfully installed Darknet in your Windows 10 system.","excerpt":"Darknet is a framework for real-time object detection. Darknet is an open-source neural network framework and is written in C and CUDA. A major reason it is used widely is that it is highly accurate and very fast. The reason for Darknet to be fast is because it is written in C and CUDA. Darknet […]","categories":["Deep Tech"],"tags":[],"author_name":"Goldy Mazumdar","publish_date":"2021-01-01T18:00:00","publication_year":"2021","word_count":485,"keywords":["CUDA","Go","AI","neural network","OpenCV","Python","C++","object detection","deep learning","R"],"extracted_tech_keywords":["AI","deep learning","neural network","OpenCV","object detection","CUDA","Python","R","Go","C++"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-darknet-installation-on-windows-10-cpu-version\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":24147,"title":"Google Researchers Find Out Secret To Great Machine Translation: Attention","content":"Our brain is modular by design and is characterised by distinct but interacting subsystems which handle important functions like memory, language and perception. Attention is studied in neuroscience as an important cognitive process. These advances in neuroscience have inspired exciting developments using attentional mechanisms in machine learning. There is a whole field of research in computational modelling of attention as well. In machine learning most of the attention mechanism models are used to process and learn from sequential data. The dominant models are based on recurrent or convolutional neural networks which include an encoder and a decoder. These dominant models are complicated and also face prediction churn issues. The researchers at Google and the University of Toronto came up with a much simpler network architecture, called the Transformer, based solely on attention mechanisms, with no need for recurrence and convolutions. The architecture designed by the team of Google Brain, Google Research and the University of Toronto is end-to-end, based on attention model. This of the article takes a look at its architecture and the end results. Machine Translation Tasks And The Transformer Self-attention is an attention mechanism relating different positions of a single sequence in order to calculate a representation of the sequence. The paper proposes the Transformer — a model architecture which does not use recurrence and relies entirely on an attention mechanism — to learn global dependencies between input and output. The Transformer allows for significantly more parallelisation and can reach a new state-of-the-art level in translation quality after being trained for as only 12 hours on eight P100 GPUs. The researchers did experiments on two machine translations tasks using their architecture. The architecture is readily parallelizable and requires significantly less time to train. The model achieves 28.4 BLEU score on the WMT 2014 English-to-German translation task, improving over the existing benchmarks. The researchers also show that the Transformer generalises excellently to other tasks by parsing both with large and limited training data. The architecture is designed in a way so that it requires lesser number of training instances than traditional networks. Model Architecture The neural sequence models are mostly built as encoder-decoder structures. The encoder maps an input sequence of symbol representations (x1…xn) to a sequence of continuous representations z = (z1…zn). Given z, the decoder then generates an output sequence (y1…ym) of symbols one element at a time. The Transformer is based on these three elements: Encoder: The encoder is composed of a stack of N = 6 identical layers. The first is a multi-head self-attention mechanism, and the second is a simple, position-wise fully connected feed-forward network. Decoder: The decoder is also composed of a stack of N = 6 identical layers. In addition to the two sub-layers in each encoder layer, the decoder inserts a third sub-layer, which performs multi-head attention over the output of the encoder stack. Attention: An attention function can be described as mapping a query and a set of key-value pairs to an output, where the query, keys, values, and output are all vectors. The model has a multi-head attention system and is used in the following ways: In “encoder-decoder attention” layers, the queries to the current layer come from the previous decide layer and the memory keys and values, come from the output of the encoder. This is similar to the encoder-decoder attention mechanisms in sequence-to-sequence models. The encoder contains self-attention layers. In a self-attention layer all of the keys, values and queries come from the same place. Each position in the encoder can attend to all positions in the previous layer of the encoder. Similarly, self-attention layers in the decoder allow each position in the decoder to attend to all positions in the decoder up to and including that position. We need to prevent leftward information flow in the decoder to preserve the autoregressive property. Training And Results The researcher trained on the standard WMT 2014 English-German dataset consisting of about 4.5 million sentence pairs and shared source target vocabulary of about 37,000 tokens. They also trained the models on one machine with P100 GPUs and The big models were trained for 300,000 steps, which took three-and-a-half days to train. On the WMT 2014 English-to-German translation task, the Transformer beats the best previously reported models, and hence establishes a new state-of-the-art score BLEU score of 28.4. On the WMT 2014 English-to-French translation task, the Transformer achieves a BLEU score of 41.0. The researchers evaluated the Transformer can generalise to others tasks such as English constituency parsing. Also, RNN sequence-to-sequence models have not been able to attain state-of-the-art results when smaller data sets are involved. The researchers from Google and University of Toronto who have been working on this project, are really excited about the future of attention models in AI tasks. The processing of sequential tasks, something which was largely dominated by RNNs and CNNs, is now being tackled using a pure attention model network. The researchers plan to extend the Transformer to problems involving input and output modalities other than text and to investigate local, restricted attention mechanisms to efficiently handle large inputs and outputs such as images, audio and video.","excerpt":"Our brain is modular by design and is characterised by distinct but interacting subsystems which handle important functions like memory, language and perception. Attention is studied in neuroscience as an important cognitive process. These advances in neuroscience have inspired exciting developments using attentional mechanisms in machine learning. There is a whole field of research in […]","categories":["IT Services"],"tags":["attention","attention mechanism","encoder","Google Brain","machine translation","RNN","translation","University of Toronto"],"author_name":"Abhijeet Katte","publish_date":"2018-05-01T07:15:35","publication_year":"2018","word_count":854,"keywords":["encoder","Go","machine translation","machine learning","translation","TPU","AI","University of Toronto","neural network","R","Google Brain","attention","Modal","RNN","CNN","attention mechanism"],"extracted_tech_keywords":["AI","machine learning","neural network","TPU","R","Go","attention mechanism","CNN","RNN","Modal"],"url":"https:\/\/analyticsindiamag.com\/it-services\/google-researchers-find-out-secret-to-great-machine-translation-attention\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10058846,"title":"Top Infrastructure as Code tools for 2022","content":"Infrastructure as Code (IaC) is a descriptive model-based approach for configuring and managing infrastructure. The configuration modules are usually saved in version control systems in well-documented code formats, improving correctness, decreasing errors, and speeding up consistency. Many firms are migrating to this crucial DevOps practice to reap the benefits of its changeless infrastructure, increased delivery speed, scalability, cost savings, and risk avoidance. Below, we review the top IaC tools that you can use in 2022. Terraform Terraform by HashiCorp is the most widely used open-source infrastructure automation technology. It assists with infrastructure as code configuration, provisioning, and management. Terraform makes it simple to plan and deploy IaC across numerous infrastructure providers using a single procedure. The required infrastructure is defined as code using a declarative approach. Before upgrading or provisioning infrastructure, Terraform allows users to run a pre-execution check to see if the settings fulfil the result expectations. Users can have their chosen architecture across numerous cloud providers through a single and uniform CLI procedure. You can quickly set up several environments with the same configuration and manage the entire lifecycle of your chosen infrastructure, eliminating human error and enhancing infrastructure automation. Ansible RedHat created Ansible to promote simplicity. The tool contributes to IT modernisation and aids DevOps teams in deploying applications faster, more reliably, and in a more coordinated manner. Without worrying about meeting compliance standards, you can quickly create several identical environments with security baselines. Ansible is the simplest approach to automate the provisioning, configuration, and maintenance of applications and IT infrastructure. You can use Ansible to run playbooks to generate and manage the infrastructure resources. It can connect to servers and conduct commands over SSH without agents.The code is written in YAML, making the configurations relatively simple to comprehend and deploy. Ansible’s capabilities can be further enhanced by adding new modules and plugins. Google Cloud Deployment Manager Google’s infrastructure deployment service, Cloud Deployment Manager, automates the management, creation, provisioning, and configuration of Google Cloud Platform resources using a declarative language. You can manage the resources by using YAML or Python scripts and use the code to generate equally consistent deployments in the future on resource groups. It also allows you to see how modifications will affect your system before they’re implemented. You may also utilise the built-in console to check your current deployments if necessary. However, what distinguishes the Deployment Manager from the other Infrastructure as Code solutions on this list is how tightly it is linked to Google’s ecosystem. It essentially adds UI functionality to the developer’s console, making it easier to visualise deployment architecture. Deployment Manager also requires no additional configuration software and comes at no additional expense because it is built into the platform. Azure Resource Manager Azure Resource Manager is a service that allows users to deploy and manage Azure resources. The solution enables resources to be deployed, maintained, and tracked as a group rather than individually. Role-Based Access Control (RBAC) is built into the resource management system, allowing users to apply access control to all resources within a resource category. Resource Manager lets you utilise declarative templates instead of scripts to manage your infrastructure. You may reinstall your infrastructure solution several times during the application development lifecycle with Azure resource manager while being able to keep state consistency. AWS CloudFormation AWS’s CloudFormation is a built-in IaC solution that addresses DevOps needs. CloudFormation is another declarative method, and because the programme is designed exclusively for AWS infrastructure, the margin of error is minimal. Declarative declarations can be made in YAML or JSON. In the event of an error, CloudFormation also provides a Rollback Trigger function that instantly reverts to a working state.","excerpt":"AWS’s CloudFormation is a built-in IaC solution that addresses DevOps needs.","categories":["AI Trends"],"tags":["AI Infrastructure"],"author_name":"Abhishree Choudhary","publish_date":"2022-01-21T11:00:00","publication_year":"2022","word_count":609,"keywords":["Go","AWS","AI","R","ML","Scala","automation","Python","AI Infrastructure","DevOps","Azure"],"extracted_tech_keywords":["AI","ML","AWS","Azure","Python","R","Go","Scala","DevOps","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-infrastructure-as-code-tools-for-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":60612,"title":"ACM Awarded Their Computing Prize To An AlphaGo Developer","content":"ACM, the Association for Computing Machinery, has recently announced their ‘2019 ACM Prize in Computing’. And, this year they awarded David Silver for breakthrough advances in computer game-playing. David works at University College London as a professor and at DeepMind as a Principal Research Scientist, and has also been recognised as a critical individual in the area of deep reinforcement learning. David Silver was known as a leader of the team that developed AlphaGo, a computer program that defeated the world champion of the game Go. He has also developed the AlphaGo algorithm by deftly combining ideas from deep-learning, reinforcement-learning, traditional tree-search and large-scale computing. To initialise AlphaGo, it has been trained on expert human games followed by reinforcement learning in order to improve its performance. Subsequently, David Silver sought even more principled methods for achieving higher performance and generality. He developed the AlphaZero algorithm that learned entirely by playing games against itself, starting without any personal data or prior knowledge except the game rules. AlphaZero achieved superhuman performance in the games of chess, Shogi, and Go, demonstrating unprecedented generality of the game-playing methods. The ACM Prize in Computing recognises early-to-mid-career computer scientists whose research contributions have fundamental impact and broad implications. The award carries a prize of $250,000, from an endowment provided by Infosys Ltd. Silver will formally receive the ACM Prize at ACM’s annual awards banquet on June 20, 2020, in San Francisco. Computer Game-Playing and AI AI researchers have been core in teaching computer programs to play games, where an agent is supposed to make a series of decisions using the process to win the game. Game-playing also affords researchers results that are easily quantifiable—that is, did the computer follow the rules, score points, and win the game? The programs are developed in order to compete with humans at checkers, and over the decades, increasingly sophisticated chess programs were introduced. It was a turning point when ACM, in 1997, sponsored a tournament in which IBM’s DeepBlue became the first computer to defeat a world chess champion, Gary Kasparov. However, the main aim to develop programs was to use game-playing as a touchstone to create machines with capacities that simulated human intelligence. According to ACM President Cherri M. Pancake, “Few other researchers have generated as much excitement in the AI field as David Silver.” He further stated that the human vs machine contests have long been a yardstick for AI. Millions of people around the world watched as AlphaGo defeated the Go world champion, Lee Sedol, on television in March 2016. But that was just the beginning of Silver’s impact. His insights into deep reinforcement learning are already being applied in areas such as improving the efficiency of the UK’s power grid, reducing power consumption at Google’s data centres, and planning the trajectories of space probes for the European Space Agency. Silver is credited with being one of the foremost proponents of a new machine learning tool called deep reinforcement learning. In deep reinforcement learning, the algorithm learns by trial-and-error in an interactive environment and continually adjusts its actions based on the information it accumulates while it is running. Learning Atari from Scratch In 2013, at the Neural Information Processing Systems Conference, David Silver and his team working at DeepMind presented a computer program that could play 50 Atari games to human-level ability. It was designed to learn playing games based solely on observing the pixels and scores while playing. Earlier reinforcement learning approaches had not achieved anything close to this level of ability. The team published their method of combining reinforcement learning with artificial neural networks in a seminal 2015 paper. The team was also capable of refining these deep reinforcement learning algorithms with novel techniques, and these algorithms remain among the most widely-used tools in machine learning. During his PhD, at the University of Alberta, David Silver first began exploring the possibility of developing a computer program that could master Go. His critical insight in developing AlphaGo was to combine deep neural networks with an algorithm used in computer game-playing called Monte Carlo Tree Search. The best way, while pursuing the perceived best strategy in a game, the algorithm is also continually investigating other alternatives. AlphaGo’s defeat of world Go champion Lee Sedol in March 2016 was hailed as a milestone moment in AI. Again the team published the foundational technology underpinning AlphaGo in the paper in 2016. David Silver and his team at DeepMind continued their work on developing new algorithms to advance the computer program game-playing and achieved results many in the field thought were not yet possible for AI systems. In developing the AlphaGo Zero algorithm, Silver and his collaborators demonstrated that a program could master Go without any access to human expert games. The algorithm learns entirely by playing itself without any human data or prior knowledge, except the rules of the game and, in a further iteration, without even knowing the rules. The DeepMind team continues to advance these technologies and find applications for them. Among other initiatives, Google is exploring how to use deep reinforcement learning approaches to manage robotic machinery at factories.","excerpt":"ACM, the Association for Computing Machinery, has recently announced their ‘2019 ACM Prize in Computing’. And, this year they awarded David Silver for breakthrough advances in computer game-playing. David works at University College London as a professor and at DeepMind as a Principal Research Scientist, and has also been recognised as a critical individual in […]","categories":["AI News"],"tags":["AlphaGo","Artificial Intelligence vs Human Intelligence","Deep Learning Techniques","deepmind london","Gaming"],"author_name":"Sejuti Das","publish_date":"2020-04-01T12:43:58","publication_year":"2020","word_count":853,"keywords":["Go","AlphaGo","machine learning","AI","neural network","Gaming","BERT","Artificial Intelligence vs Human Intelligence","Aim","llm_models:BERT","Deep Learning Techniques","GAN","AI research","R","deepmind london"],"extracted_tech_keywords":["AI","machine learning","neural network","Aim","R","Go","BERT","GAN","AI research","llm_models:BERT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/acm-awarded-their-computing-prize-to-an-alphago-developer\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":59574,"title":"The Various Ways AI Can Make An Architect’s Life Easy","content":"In recent years various technologies and automation in the field of architecture have increased. However, now, AI has been making its way into the field of architecture slowly as researchers are making significant advancements. While it can’t be stressed enough that AI will never be able to replace human architects, it will certainly make their lives easier. AI will also speed up the processes that aren’t related to core architecture, this gives them more freedom to create many models while managing fewer activities. As AI enters the field of architecture, we look at some of the ways it will have an effect on the field: Helping In Research To start off, project architects usually have to go through tons of data and countless hours of research. This helps them better understand the design of the project and also form a philosophy behind it. So, the data and research that they go through are massive, and this is where AI’s potential steps in. AI has the ability to take in a colossal amount of data, so this can help an architect go about researching and testing several ideas at the same time with great comfort. This also means an architect can make conceptual designs with little use of pencil and paper. Parametric Architecture Parametric architecture is something that is widely used by some of the famous architects around the world. First off, parametric design is something that allows one to use different parameters to create various types of outputs and generate several structures and forms that wouldn’t usually be possible. It gives an architect the ability to set the constraints, plug data and create countless iterations of products. It also gives architects the freedom of building designs according to their desired output. The parametric architecture uses complex algorithms and also deploys geometric programming to allow architects to take a building and reshape and optimise it to fit their needs. A tool like this has a lot of potentials to take advantages of AI capabilities where it allows the architect to easily tinker with different parameters and get different outputs efficiently and smoothly. Construction And Planning The most important aspect of an architect’s project is the construction and planning. One can never be too prepared when it comes to construction and planning; some of the planning takes years to bring the exact vision of an architect to reality. The potential that AI holds here is that it will make the planning process more manageable and give the architects access to a massive amount of data, create desired models easily, analyse the building environment, and create cost estimates. These processes are usually tedious and time-consuming for the architects, so AI can shorten the overall process by reducing design and building time. When it comes to construction, AI can assist in constructing structures with almost no help from humans; researchers have been creating AI drones that can communicate with each other to create small structural models, which otherwise take a lot of time with traditional manufacturing approaches. Building Perfect Homes While building or designing homes, architects, along with giving emphasis to the general aesthetics of the building and using AI to create homes, they will also have to concentrate on integrating AI inside the apartments. Ultimately, the concept of AI is to make human life more comfortable and enhance it. What better place to start than transforming ordinary houses into smart ones? Along with other technologies, AI will help make the humans of the house live smarter and safer, with proper planning by architects, and AI will help the field venture deeper, making human lives more comfortable. Helping Build Better Cities Building better cities generally involve years of planning and require precision when it comes to fulfilling the vision of a perfect city plan. Now, as the world moves towards AI and IoT, more effort will be needed to plan and make the existing cities into smart cities. Architects, along with having to deal with the concept of smart cities, will also have to think of entirely new models and how to use AI to get there. There is a lot of data that needs to be collected and processed in order to pinpoint specific aspects of the cities like the traffic congestion percentage, average population, the type of community that lives there and industry resources available to construct markets among other considerations. These were just some of the examples that architects have to consider while designing a smart city. So, AI holds a lot of promise when it comes to processing and getting insights from the abundance of data. India and other countries have been planning to use AI towards better mobility planning already, and architects will also have a huge role to play.","excerpt":"In recent years various technologies and automation in the field of architecture have increased. However, now, AI has been making its way into the field of architecture slowly as researchers are making significant advancements. While it can’t be stressed enough that AI will never be able to replace human architects, it will certainly make their […]","categories":["AI Features"],"tags":["architecture"],"author_name":"Sameer Balaganur","publish_date":"2020-03-24T11:00:00","publication_year":"2020","word_count":791,"keywords":["Go","TPU","programming_languages:R","AI","programming_languages:Go","RAG","architecture","automation","ViT","R"],"extracted_tech_keywords":["AI","RAG","TPU","R","Go","ViT","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-various-ways-ai-can-make-an-architects-life-easy\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10138930,"title":"The Disturbing Case of Jennifer Ann’s Chatbot Where AI Crossed the Line","content":"A bizarre case recently hit the headlines worldover when a girl in the United States, murdered in 2006, was resurrected as an AI chatbot — and her family had no idea! Drew Crecente, the father of the deceased, woke up to a Google alert which soon transformed into an unnerving revelation. To his utter shock, he realised that an AI chatbot, created using his deceased daughter Jennifer Ann’s name and image, was in operation 18 years after her tragic death. By the time Crecente discovered the bot, a counter on its profile showed it had already been used in at least 69 chats. This particular chatbot was found on Character.ai, a platform where users can create their own AI “characters”. Ann’s name and her yearbook photo were used along with a description that called her as an “expert in journalism”. Eventually, Character.ai removed Ann’s chatbot as a response to Brian’s post. This particular incident, though traumatising for the family, compels readers to ponder on the direction we are headed with respect to the ethics (or lack thereof) in the application of AI. It also raises disturbing questions about privacy and human rights. Even though AI developments present new opportunities and benefits, AI tools are also used as a means of societal control, mass surveillance and discrimination. Larry Ellison, the co-founder of Oracle, said that AI will bring in a new era of surveillance that will ensure “citizens will be on their best behaviour”. However, this leads us to the most obvious question of algorithmic bias which results from the kind of datasets fed into systems. According to a research paper by The National Institute of Standards and Technology, ‘Towards a Standard for Identifying and Managing Bias in Artificial Intelligence’ , “AI bias extends beyond computational algorithms and models, and the datasets upon which they are built.” Merel Koning, senior advisor on technology and human rights at Amnesty International, highlighted that a xenophobic algorithm used in the Netherlands caused significant harm to thousands of lives. She warned that without human rights protection, such mistakes could repeat in the future. Ethics and Unauthorised Use of Personal Data AI systems can collect vast amounts of data through various means, raising significant privacy concerns. Web scraping allows AI to automatically harvest public and potentially personal information from websites, often without user consent. Moreover, the rising use of biometric technologies, such as facial recognition and fingerprinting, collect sensitive, unique data that are not foolproof. AI-powered IoT devices, meanwhile, unlock insights into our daily lives and collective human mind. According to Salman Waris, founder & managing partner of TechLegis Advocates & Solicitors, the law does give individuals certain rights of “privacy” and “publicity” which provide limited rights to control how someone’s name, likeness, or other identifying information is used under certain circumstances. However, these laws vary from state to state, so they are difficult to summarise. “For instance, in California, the law lays down that the right of privacy or publicity is violated when someone’s name, voice, signature, photograph or likeness appears in a work of art and the subject has not consented to its use,” Waris mentioned. MyHeritage, a geneology website, introduced a tool called Deep Nostalgia that allowed users to animate old photographs of their deceased relatives. The AI would add movement to the eyes, mouth, and head, creating the illusion that the person in the photo was “alive”. While many found the technology fascinating and heartwarming, others found it unsettling or emotionally overwhelming, especially when the animations involved long-deceased relatives. The ethical concern was whether animating someone who cannot give consent is respectful. In the documentary Roadrunner: A Film About Anthony Bourdain, the filmmakers used AI to recreate Bourdain’s voice and generated a few sentences of narration in his voice, based on things he had written but never spoken out aloud. This sparked a debate about the ethical implications of using AI to recreate a deceased individual’s voice without clear consent. This raises the crucial question: Are we on the brink of slipping into a world where AI erodes human rights, or have we already crossed that line? During the Gaza conflict, reports emerged that AI-powered systems were used to identify and strike targets, often with limited real-time human input. Algorithms designed to predict behaviour or locate targets based on movement patterns or communications can lead to severe mistakes, particularly in environments where civilians are close to military operations. There have been claims that AI systems used in the Gaza war disproportionately targeted civilians, including children and families, under the guise of precision strikes. This calls into question whether the use of such technology is compatible with the principles of proportionality and distinction, which are the pillars of international humanitarian law designed to protect civilians.As per the Council of Europe Study, the use of algorithms in the face of rapidly changing technologies raises considerable challenges, including the safeguarding of human rights and dignity. “Indeed, the increasing use of automation and algorithmic decision-making in all spheres of public and private life is threatening to disrupt the very concept of human rights as protective shields against state interference,” the report highlighted.","excerpt":"The incident raises pertinent questions about privacy, human rights and the ethical use of AI.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Data Security","Privacy"],"author_name":"Shalini Mondal","publish_date":"2024-10-20T13:34:44","publication_year":"2024","word_count":856,"keywords":["Go","Privacy","API","artificial intelligence","AWS","AI","Data Security","cloud_platforms:AWS","RAG","automation","Aim","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","AWS","R","Go","API","automation","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-disturbing-case-of-jennifer-anns-chatbot-where-ai-crossed-the-line\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10020400,"title":"Microsoft Excel Is Turing Complete. What Does It Mean?","content":"With the introduction of LAMBDA, Excel has now become a full-fledged programming language. According to Microsoft, Excel is Turing-complete, and users can write any computation in Excel formula language. Excel Is Turing-Complete Microsoft Excel was first introduced in the 1980s, and since then, it has become the go-to platform for organising, analysing, and visualising data. Excel is the most widely used programming language, and its formulas are used more than C, C++, Java, Python combined, Microsoft claimed. For the longest, Excel suffered from two main shortcomings: Excel supported only scalar values such as numbers, strings, and BooleansIt did not have the provision for user-defined functions Microsoft Research’s Calc Intelligence project led the transformation of spreadsheet formulas into a full-fledged programming language. The first development took place at the 2019 ACM SIGPLAN Symposium on Principles of Programming Languages (POPL 2019). Microsoft made two major announcements: to have data values beyond just text and numbers and to allow cells to include first-class records; and allow the use of ordinary formulas to compute complete arrays. Building on this, Microsoft soon announced the beta version of LAMBDA. Previously, Excel allowed user-defined custom functions, but they had to be written in languages such as Javascript. On the other hand, LAMBDA allows the users to define their custom functions using Excel programming language. It also allows function calling. Excel formulas, the world’s most popular programming language, is now Turing-complete. Go check it out! https:\/\/t.co\/qkw3Bmt1gp— Satya Nadella (@satyanadella) February 9, 2021 Turing complete (a concept derived from theoretical computer science) means the programming language is powerful in terms of performance, maintenance and richness of the ecosystem. A programming language is Turing complete when users can implement any algorithm with it. What To Expect Formula reusability: One of the challenges in working with formulas in conventional Excel programming language is they are often very complex. When these complex formulas are reused repeatedly throughout the sheet, there are mainly two types of roadblocks: If there is an error in the logic, programming would need to go back and update it everywhere. The risk amplifies when the formulas used are complex — giving way for greater human error.Secondly, for a programmer who is not the original author, it is hard to comprehend the intention of a given formula. Any formula built in Excel can be wrapped in a LAMBDA function and used anywhere throughout the sheet. Recursion: Excel formulas cannot loop (repeat over a set of logic within defined intervals). Users generally configure the time interval manually. However, given the set of characters we work with are not static, the manual way is not always suitable and may lead to added complexity and errors. However, with LAMBDA, users can create a function called REPLACECHARS. It references itself and allows the user to iterate through the list of characters to be removed. Different Data Types: This is a general improvement, but Excel, over the years, has introduced new types of data that can be worked with. It includes dynamic arrays for passing values and functions; and inclusion of other rich data types for values stored in a cell. Wrapping Up In the future, Microsoft Excel would have nestable arrays and implementations of array-processing combinators that take lambda functions as their arguments. Microsoft also plans to define functions through sheet-defined functions. Meaning, users will be able to define larger functions using multiple formulas spread across cells.","excerpt":"With the introduction of LAMBDA, Excel has now become a full-fledged programming language. According to Microsoft, Excel is Turing-complete, and users can write any computation in Excel formula language. Excel Is Turing-Complete Microsoft Excel was first introduced in the 1980s, and since then, it has become the go-to platform for organising, analysing, and visualising data. […]","categories":["Global Tech"],"tags":[],"author_name":"Shraddha Goled","publish_date":"2021-02-17T15:00:00","publication_year":"2021","word_count":564,"keywords":["Go","AI","Scala","Python","Ray","Aim","C++","JavaScript","R","Java"],"extracted_tech_keywords":["AI","Aim","Ray","Python","R","JavaScript","Go","Java","Scala","C++"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsoft-excel-is-turing-complete-what-does-it-mean\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":13981,"title":"The inconvenient truth &#8212; broken promise of social media analytics","content":"Has social media analytics run its course? The importance of social media analytics had long been recognized since 2013 and it drove a paradigm shift in driving early product or market signals and influencing client behavior. Blurring the lines between the real and social world, social media analytics spawned a new set of startups that promised how to manage the new social disruption, deliver breakthrough insights, sentiments and predicting shifting consumer behavior and market patterns. However, four years down the line and with three startups, NM Incite, Salorix, Topsy and ThinkUp (shut last year) closing shop, it is time for a rethink. Has social media analytics run its course and is it time for a downward plunge? A new Gartner research doesn’t write off the promise of social media analytics completely but outlines the limited success and spells out why companies are unable to make a bang on their buck. From putting technology before strategy to roadblocks in converting insights into actions, organizations are failing to capitalize on social analytics. The promise of social media analytics is on a decline because “social analytics is not panning out the way it was expected”. Well, we are not talking about the demise of social media analytics, but the buzz has definitely down. Though social media monitoring and mining reaps (collection and automated analysis of quantitative amounts of naturally occurring text from social media) plentiful results, the success is still limited, cites Ray Poynter, who has authored author of The Handbook of Mobile Market Research. Challenges in social media analytics & why the hype has died down Marketers fail to realize the real worth of insights Failure to convert insights into actions: According to Poynter, “Most automated sentiment analysis is deemed useless or poor by insight clients and market researchers. Which basically means social media usage requires people with some expertise, and that tends to make the whole process forbiddingly expensive and slower”. Unstructured data is challenging: Experts have deemed social media analytics the “Big Data in the Wild” and it comes with all the challenges associated with adoption of big data, from quality, inaccessibility and finding out where the unstructured data is coming from. Putting technology before strategy: As the Gartner research points out, there is a lack of awareness of usage of social analytics tool that leads to missed opportunities. While most organizations rely on social media analytics for research source, and taking the conversation forward with customers, they lack the techniques to utilize insights and move the company forward. Data silos: according to the Gartner research, social media data just doesn’t include social data, but also covers email data, surveys and other internal and external sources. Only by combining it with other sources of data, can the real value be realized. Data silos can be overcome by aggregating reports in a single dashboard or using a tool that consolidates the data in a single place. API limitations: API limitations by Facebook and Instagram limits the kind of data that can be collected, thereby affecting insights. This requires necessary technical development and sometimes necessitates a workaround that could be in conflict of the terms and policies of social networks. Social media analytics shops shutting down The end of ThinkUp: Last year saw personalized social media insights service provider ThinkUp, co-founded by Anil Dash closing down due to API limitations. According to Dash’s post, that confirmed the demise of ThinkUp, he noted significant changes from social media companies Twitter, Instagram and Facebook made it difficult to keep the service running. In his post, Dash cited the biggest reason from a business standpoint was a lack of ThinkUp subscribers to cover the cost of investing in technical development that was needed to deal with tech challenges. Besides, significant changes to APIs, limit the amount of data that can be collected and also hinders the analysis. “Even if we make all the necessary development changes that would be required to continue supporting these networks, there’s no guarantee that they wouldn’t just change their APIs again,” Dash wrote. Topsy no longer exists: Another popular social media analytics service provider Topsy that was bought by Apple shut down in 2015. One of the most popular social media analytics tool that gained access to Twitter’s full data and provided enriched insights failed to flourish under the new ownership. According to experts, Topsy’s demise was fueled by the acquisition, speculations were rife that Apple bought it to build a relevancy engine for its own app and iTunes store. Salorix Inc has been long dead: Bangalore startup Salorix Inc shut shop in 2013 as it failed to raise funds. According to news reports, search giant Google had approached the social analytics startup for a buyout, but the unimpressive deal was turned down by founder and CEO Santanu Bhattacharya, following which the startup was forced to shut shop. Set up in 2009, the startup backed by Nexus Venture Partners and Inventus Capital Partners had a fairly good run helping companies scale social media engagement through analytics. Last word Livefyre founder and chief, Jordan Kretchmer whose content marketing and engagement platform was acquired by Adobe last year said the social media strategy of listening, reach, depth and relationship, outlined by Forrester research in 2015 no longer works. Another key reason cited by Kretchmer was the rise of vendors launching and expanding their features to fill the pressing needs of marketers and customer service teams around content delivery, measurement and automation. A key point made by the Forrester research was the social technology vendor landscape had become incomprehensible, urging the marketers to choose between an expensive social suite or an array of solutions.","excerpt":"The importance of social media analytics had long been recognized since 2013 and it drove a paradigm shift in driving early product or market signals and influencing client behavior. Blurring the lines between the real and social world, social media analytics spawned a new set of startups that promised how to manage the new social […]","categories":["IT Services"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-04-03T09:49:08","publication_year":"2017","word_count":941,"keywords":["big data","Go","API","sentiment analysis","AI","automation","Ray","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","Ray","sentiment analysis","R","Go","API","big data","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/inconvenient-truth-broken-promise-social-media-analytics\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":38793,"title":"10 Most Used Python GitHub Repositories To Look Out For In 2019","content":"Python is one of the heavily used programming languages in data science. In the State of the Octoverse report by GitHub for the year 2018, Python is placed at the third position following JavaScript and Java. The information on this article has been cited from the original documentation and the sources are also cited inside. In this article, we list down 10 Python repositories in GitHub for the year 2019. (The list is in no particular order) 1| scikit-learn scikit-learn was initially developed by David Cournapeau as a Google summer of code project in 2007 and it provides a range of supervised as well as unsupervised learning algorithms through a constant interface in Python. scikit-learn comes with a few standard datasets, for instance, the iris and digits datasets for classification and the boston house prices dataset for regression. It has the following features mentioned below Simple and efficient tools for data mining and data analysis Accessible to everybody, and reusable in various contexts Built on NumPy, SciPy, and matplotlib Open source, commercially usable – BSD license Click here to know more. 2| Flask Flask is a lightweight WSGI web application framework. It is designed to make getting started quick and easy, with the ability to scale up to complex applications. It provides you with tools, libraries, and technologies that allow you to build a web application. It began as a simple wrapper around Werkzeug and Jinja and has become one of the most popular Python web application frameworks. Flask is part of the categories of the micro-framework which means it is a framework with little to no dependencies to external libraries. Click here to know more. 3| Keras Keras is a high-level neural networks API, written in Python and capable of running on top of TensorFlow, CNTK, or Theano. It was developed with a focus on enabling fast experimentation. It allows for easy and fast prototyping (through user friendliness, modularity, and extensibility), supports both convolutional networks and recurrent networks, as well as combinations of the two as well as runs seamlessly on CPU and GPU. Keras is compatible with: Python 2.7-3.6. Click here to know more. 4| Sentry Sentry is an open-source error tracking tool which helps you monitor and fix crashes in real time. The server is in Python, but it contains a full API for sending events from any language, in any application. For pairing Sentry up with Python you can use the Raven for Python (raven-python) library. It is the official standalone Python client for Sentry. It can be used with any modern Python interpreter be it CPython 2.x or 3.x, PyPy or Jython. Click here to know more. 5| Django Django is a free and open-sourced high-level Python web framework which encourages rapid development and clean, pragmatic design. Built by experienced developers, it takes care of much of the hassle of Web development, so you can focus on writing your app without needing to reinvent the wheel. It is designed to help developers take applications from concept to completion as quickly as possible and helps developers avoid many common security mistakes. Click here to know more. 6| Ansible Ansible is an IT automation tool. It can configure systems, deploy software, and orchestrate more advanced IT tasks such as continuous deployments or zero downtime rolling updates. Ansible’s main goals are simplicity and ease-of-use. It also has a strong focus on security and reliability, featuring a minimum of moving parts, etc. Ansible is written in Python, but you can write modules in any language. Click here to know more. 7| Tornado Tornado is a Python web framework and asynchronous networking library, originally developed at FriendFeed. By using the non-blocking network I\/O, Tornado can scale to tens of thousands of open connections, making it ideal for long polling, WebSockets, and other applications that require a long-lived connection to each user. Click here to know more. 8| Pandas Pandas is a Python package providing fast, flexible, and expressive data structures designed to make working with “relational” or “labeled” data both easy and intuitive. It aims to be the fundamental high-level building block for doing practical, real-world data analysis in Python. It includes features such as easy handling of missing data in floating point as well as non-floating point data, automatic and explicit data alignment, flexible reshaping and pivoting of datasets, etc. Click here to know more. 9| Matplotlib Matplotlib is a Python 2D plotting library which produces publication-quality figures in a variety of hardcopy formats and interactive environments across platforms. Matplotlib can be used in Python scripts, the Python and IPython shell, web application servers, and various graphical user interface toolkits. You can generate plots, histograms, power spectra, bar charts, error charts, scatterplots, etc., with just a few lines of code. Click here to know more. 10| Zulip Zulip is a powerful, open source group chat application that combines the immediacy of real-time chat with the productivity benefits of threaded conversations. Written in Python and using the Django framework, Zulip supports both private messaging and group chats via conversation streams. Zulip is used by open source projects, Fortune 500 companies, large standards bodies, and others who need a real-time chat system that allows users to easily process hundreds or thousands of messages a day. With over 300 contributors merging over 500 commits a month, Zulip is also the largest and fastest growing open source group chat project. Click here to know more.","excerpt":"Python is one of the heavily used programming languages in data science. In the State of the Octoverse report by GitHub for the year 2018, Python is placed at the third position following JavaScript and Java. The information on this article has been cited from the original documentation and the sources are also cited inside. […]","categories":["AI Trends"],"tags":["django python","GitHub","GitHub open source Python projects","GitHub Repositories","iris dataset python","Python"],"author_name":"Ambika Choudhury","publish_date":"2019-05-08T08:47:53","publication_year":"2019","word_count":902,"keywords":["data science","scikit-learn","NumPy","iris dataset python","Keras","AI","neural network","GitHub open source Python projects","ML","TensorFlow","Python","Aim","GitHub Repositories","django python","GitHub","Pandas"],"extracted_tech_keywords":["AI","ML","neural network","data science","Aim","TensorFlow","Keras","scikit-learn","Pandas","NumPy"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-most-used-python-github-repositories-to-look-out-for-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10136596,"title":"Salesforce Urges Enterprises to Replace Copilots with AI Agents","content":"At his keynote address at Dreamforce, Marc Benioff made a clear point–AI agents are key to scaling AI adoption in enterprises. During the event, which is held in San Francisco, he took a jab at companies like Microsoft, noting that enterprises have found it challenging to realise the value of copilots– likening AI chatbots to Microsoft Office’s rule-based agent from the 2000s–Clippy. Benioff’s criticism of copilots–a term popularised by Microsoft– might seem unexpected, given that the company, not so long ago, introduced its own version of copilot–called Einstein Copilot. “The copilot world has been kinda hit and miss world. Customers have told me we got these copilots, but they are not performing as we want them to,” he said in his keynote address at Dreamforce 2024. Now, rebranded to Agentforce–of course–the platform is designed to help enterprises streamline business processes, deliver personalised insights, and boost productivity and all of that. According to Benioff, copilots were able to assist with basic tasks, but their inefficiencies are what gave birth to AI Agents. Dubbed as the next iteration of deep learning models, AI agents are designed to go further and take actions on behalf of users. ( Salesforce CEO Marc Benioff during his keynote address at Dreamforce) Buy not Build Interestingly, this isn’t the first time Benioff has criticised Microsoft. During a recent earnings call, he highlighted how enterprises have squandered considerable resources on third-party models, often being sold as APIs or in the cloud, attempting to fine-tune them for their specific use cases. Salesforce’s approach is to democratise AI adoption, and Benoiff’s message to customers at Dreamforce has been clear– buy, not build. To further encourage this, Salesforce is mulling a consumption-based pricing model for its AI agents– somewhere around $2 per conversation. Benioff revealed that companies such as Saks, Wiley, and Wyndham have already started leveraging the Agentforce platform to build customised AI agents for their specific needs. A few Indian heavyweights, such as Mahindra and Mahindra, Tata Consumer, and TVS, which are all Salesforce customers, were present at the event to gain firsthand experience and ‘know-how’ of these AI agents. While speaking to AIM, all of them revealed that they are excited by the prospect of Agentforce. “Attending Dreamforce provides insight into what’s next for that platform. While Agentforce may sound innovative, it’s actually about leveraging Salesforce’s existing capabilities that are now being productised,” Saurabh Khullar, chief customer experience and digital officer TVS, told AIM at Dreamforce. “The rationale for buying rather than building is simple: if you build it, you’ll need to continually enhance it and create the necessary ecosystem. In contrast, when you purchase a solution, you can rely on the vendor to keep developing the platform,” he added. Atlas: The Brain Behind Agentforce Powering Agentforce is Atlas–an advanced AI reasoning engine designed to simulate human-like thinking and planning processes. Atlas can autonomously analyse data, make decisions, and complete complex tasks across various business functions. Atlas enables businesses to deploy customisable agents tailored to their specific needs. By deeply integrating with Salesforce’s Data Cloud and other integrated systems to process vast amounts of information in real-time, Atlas ensures that the AI agents in Agentforce make decisions based on the most current and relevant data. While speaking with journalists at Dreamforce, Clara Shih, CEO of Salesforce AI, revealed that while the first iterations of these AI agents will automate simple routine tasks, going forward, they will be able to handle complex tasks such as handling payments. “I believe that agents will have the ability to handle payments on behalf of their consumer and yes, of course, it means that the vendor, the supplier, will have to have a bunch of safeguards. Just like the internet created an entirely new category of cybersecurity, which is, by the way, a huge, multibillion-dollar, very fast-growing industry, AI agents will also require their own new cybersecurity stack,” Shih said, answering a question posed by AIM. Data Cloud-The Underlying Foundation of Agentforce While Atlas might be the reasoning engine, Data Cloud provides the underlying foundation of Agentforce, something the CRM leader announced at the previous Dreamforce. Data Cloud is a customer data platform (CDP) that consolidates data from various sources into a single location. According to Salesforce, it is one of the fastest-growing innovations they’ve introduced, if not the fastest, reflecting significant organic growth. For AI agents to work at their best, they need access to all enterprise data points, and Data Cloud enables that. What data cloud does best is bring all unstructured data together and make it accessible in one place. However, not all Salesforce customers are Data Cloud customers. “The ability to search and process unstructured data is crucial for agents. Business conversations often begin with questions about policies, such as handling a delayed shipment. Access to relevant knowledge articles—typically unstructured information—is essential; without it, agents cannot perform effectively in resolving customer issues,” Param Kahlon, EVP & GM, automation and integration at Salesforce, told AIM at Dreamforce. (The new Agentforce Mascot at Dreamforce) AI Agents vs Robotic Process Automation Interestingly, Salesforce’s hard pivot to AI agents comes at a time when every major Software-as-a-service (SaaS) and AI company is betting high on them. Microsoft, too, recently introduced Copilot agents in its Microsoft 365 Copilot platform, which is designed to help businesses customise AI for their specific needs. However, the automation being promised is akin to what Robotic Process Automation (RPA) promises—a type of business process automation that relies on software robots. Explaining the difference between RPA And autonomous AI agents, Kahlon told AIM, “RPA agents were designed to automate repetitive, tedious tasks, such as transferring data between systems when APIs aren’t involved. In contrast, autonomous agents process information more like humans, adapting to situations and making decisions based on changing conditions, enhancing efficiency and effectiveness in workflows.” He further adds that autonomous agents also does not mean the end of RPA technology. While RPA’s will exist and continue to automate tedious tasks, AI agents will start to automate much more complex tasks. “I believe RPA will coexist with today’s agents. By utilising RPA bots, autonomous agents can effectively update records or orders, making them more versatile and capable of interacting with various systems seamlessly,” Kahlon added.","excerpt":"Salesforce recently launched Agentforce, a no code platform that helps enterprises create personalised AI agents with simple prompts.","categories":["Deep Tech"],"tags":["Salesforce"],"author_name":"Pritam Bordoloi","publish_date":"2024-09-25T11:59:00","publication_year":"2024","word_count":1036,"keywords":["Go","autonomous agents","AI","chatbots","ML","RAG","Aim","deep learning","Salesforce","copilots","R"],"extracted_tech_keywords":["AI","ML","deep learning","Aim","RAG","autonomous agents","copilots","chatbots","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/salesforce-urges-enterprises-to-replace-copilots-with-ai-agents\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10102262,"title":"Open Source RedPajama-Data-v2 with 30 Trillion Tokens is Here","content":"RedPajama has unveiled the latest version of its dataset, RedPajama-Data-v2, which is a colossal repository of web data aimed at advancing language model training. This dataset encompasses a staggering 30 trillion tokens, meticulously filtered and deduplicated from a raw pool of over 100 trillion tokens, sourced from 84 CommonCrawl data dumps in five languages, including English, French, Spanish, German, and Italian. Click here to check out the GitHub repository. RedPajama-Data-v2 comes with a remarkable addition of 40+ pre-computed data quality annotations that offer invaluable tools for further data filtering and weighting. The dataset covers 5 languages, with 40+ pre-computed data quality annotations that can be used for further filtering and weighting. Here is one example of how to filter RedPajama-Data-v2 in a similar way as Gopher: pic.twitter.com\/VqKObX9Iqr— Together AI (@togethercompute) October 30, 2023 Over the past six months, the impact of RedPajama’s previous release, RedPajama-1T, has been profound in the language model community. This 5TB dataset of high-quality English tokens has been downloaded by more than 190,000 individuals, who have harnessed its potential in creative ways. RedPajama-1T served as a stepping stone towards the goal of creating open datasets for language model training, but RedPajama-Data-v2 takes this ambition to new heights with its mammoth 30 trillion token web dataset. RedPajama-Data-v2 stands out as the largest public dataset specifically crafted for LLM training, significantly contributing to the field. Most notably, it introduces 40+ pre-computed quality annotations, empowering the community to enhance the dataset’s utility. This release encompasses over 100 billion text documents derived from 84 CommonCrawl data dumps, constituting a total of 100+ trillion raw tokens. Together.AI says that the dataset offers a solid foundation for advancing state-of-the-art open LLMs such as Llama, Mistral, Falcon, MPT, and the RedPajama models. RedPajama-Data-v2 primarily focuses on CommonCrawl data, while data sources such as Wikipedia are available in RedPajama-Data-v1. To further enrich the dataset, users are encouraged to integrate Stack (by BigScience) for code-related content and s2orc (by AI2) for scientific articles. RedPajama-Data-v2 is meticulously crafted from publicly available web data, comprising the core elements of plain text source data, 40+ quality annotations, and deduplication clusters. The process of creating the source data begins with each CommonCrawl snapshot passing through the CCNet pipeline, chosen for its light processing approach, preserving raw data integrity. This results in the generation of 100 billion individual text documents, maintaining alignment with the overarching principle of data preservation.","excerpt":"Together.AI says that the dataset offers a solid foundation for advancing open LLMs such as Llama, Mistral, Falcon, and MPT.","categories":["AI News"],"tags":["ai dataset","Open Source AI"],"author_name":"Mohit Pandey","publish_date":"2023-10-31T09:04:46","publication_year":"2023","word_count":399,"keywords":["Go","AI","Git","RAG","Open Source AI","Aim","Together AI","data quality","llm_models:Llama","GitHub","R","ai dataset"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","GitHub","data quality","Together AI","llm_models:Llama"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/open-source-redpajama-data-v2-with-30-trillion-tokens-is-here\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10002123,"title":"How Indian Industries Are Using HoloLens To Reduce Machine Downtime","content":"Microsoft’s Virtual Reality headsets called Hololens are giving the VR enthusiastic audiences plenty of reasons to buy them. With their extraordinary AR experience it allows its users to have a complete 3D holographic images experience as though they are a part of the virtual environment. But these VR headsets have also recently gained popularity for some ethical issues in regard to its contract with the US Army. Basics Of Hololens Microsoft’s HoloLens, a sleek, flashy headset with transparent lenses by Microsoft which lets you see the world around you in 3D with floating objects and virtual characters around you. The lens has a very high processing power and the hardware has a 3D-satilized sound, Wi-Fi, a Kinect-like camera with a 120 degree spatial sensing system, a fleet gyroscopes and accelerometers and a transparent screen for each eye. Apart from having a CPU and a GPU in the system, the system also has something called a HPU which is a holographic processing unit, to integrate the real word and the holograph data. Use Of HoloLens In Indian Industries The technology, relatively new is being deployed by not just Tetra Pak in India to reduce the machine downtime and productivity loss for the customer, and protecting them against risks to food safety and quality by adopting the HoloLens.​​​ The company uses HoloLens to making seeing and hearing issues in real time and guiding the on-site work from anywhere in the world, possible. Several other leading companies including those in healthcare and aviation are using the technology. IT firm Ramco Systems is developing apps for the aviation sector using HoloLens. It has been building cutting-edge applications on top of Microsoft’s HoloLens. 3D models of jet engines can be converted into detailed holograms and projected into HoloLens, resulting in an authentic training experience simulating every scenario they would come across as part of their operational responsibilities. It is also used in providing comprehensive remote support. A simple Skype call with the control center enabled by the HoloLens worn by the user can deliver a real time assistance. Hyderabad-based LV Prasad Eye Institute is using HoloLens for bettering its service. The ophthalmologist has to wear the device called Holo Eye Anatomy on the head and peer into the eye in 3D. The objective of the tool is to make learning and understanding the human eye in minute detail more experiential and enriching. This device is projected through HoloLens. According to Anthony Vipin Das, Project Lead, Holo Eye Anatomy, said, “Human eye anatomy can be studied in greater detail with experiential learning. The device aids in increasing this aspect and helps collaborate remotely in real-time. With the mixed reality of Hololens experience, one can stay in the real world and simultaneously explore 3D.” Companies like ThyssenKrupp in the manufacturing sector and automotive companies such as Ford, which recently said its automotive designers have started using HoloLens that will help them design cars better and faster. In November 2018, Microsoft had secured a $480 million contract from the US Army to supply the military branch with as many as 100,000 HoloLens augmented reality headsets for training and combat purposes. This new technology in the US Army was intended to being more situational awareness to troops for them to become more lethal and mobile. Ethical Issues Related To HoloLens Some employees at Microsoft are afraid that these virtual headsets will turn real-world battlefields into a video game. The gadget has the capability to make the virtual imagery be superimposed over the real-world environments in front of them, and an occurrence of this can lead to many disadvantageous episodes. As a result, a group of Microsoft workers recently demanded that the company cancel the contract of HoloLens with US Army. More than 50 Microsoft employees circulated an internal messaging board that the technology could help soldiers spot and kill adversaries on the battlefield. The employees are of the opinion that they refuse to create technology for warfare and oppression. They asked Microsoft CEO Satya Nadella and President Brad Smith to cancel a $480 million contract inked with US Army. The letter also asked Microsoft to stop building any weapons technologies and appoint an independent ethics review board to determine acceptable uses of Microsoft technology. The protesting workers say it means HoloLens, better known for its business and entertainment applications, will be used to help kill. Other Protests For Tech Ethics There have been past incidences of protests in terms of ethics in technology. In 208, Google employees had a protest regarding the military Project Maven which uses AI to analyse aerial images from combat zones. This eventually resulted into withdrawing the project. Microsoft had also previously had a protest in 2018. The protest was regarding Microsoft’s work with the US immigration authorities for which the employees had circulated an open letter. Creating a systematic pipeline of AR expertise is the key to resolve the awareness issue. Companies need to invest in an organized marketing along with quality AR development. It is positive to see that employees of the companies today are concerned about the misuse of technology that they make, to the real world and do their part in creating awareness. It shows that the technical staff working on these products care for its misuse in the society.","excerpt":"Microsoft’s Virtual Reality headsets called Hololens are giving the VR enthusiastic audiences plenty of reasons to buy them. With their extraordinary AR experience it allows its users to have a complete 3D holographic images experience as though they are a part of the virtual environment. But these VR headsets have also recently gained popularity for […]","categories":["AI Features"],"tags":["ar","ethics","Microsoft","Satya Nadella","VR"],"author_name":"Disha Misal","publish_date":"2019-05-22T19:29:00","publication_year":"2019","word_count":879,"keywords":["Satya Nadella","Go","programming_languages:R","AI","programming_languages:Go","ar","VR","ViT","GAN","R","Microsoft","ethics"],"extracted_tech_keywords":["AI","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-indian-industries-are-using-hololens-to-reduce-machine-downtime\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10143356,"title":"OpenAI Canvas Kills Google Docs, Challenges VS Code &#038; Cursor","content":"On the fourth day of 12 Days of OpenAI, the company made Canvas available to everyone, integrating it into ChatGPT’s main interface. The update lets users run Python code in Canvas and works with custom GPTs as well. Canvas, which was previously in beta for Plus users, provides a dedicated interface for working with ChatGPT on writing and coding projects that extend beyond simple chat. The new interface features a split-screen layout, where users can edit text on the left while ChatGPT provides suggestions and feedback on the right. This format makes it easier to draft, refine, and review work. “Canvas – a new way to work with ChatGPT to draft, edit, and get feedback on writing and code – is now available to all users in our 4o model. It’s fully rolled out on the web and the ChatGPT desktop app for Windows,” OpenAI announced. Beyond basic text editing, users can format documents, provide feedback, or even ask ChatGPT to suggest improvements. A unique feature includes shortcuts like adjusting the length of content, simplifying language, or adding emojis, offering creative flexibility. The update has sparked excitement across the user and developer communities. Sam Altman, CEO of OpenAI, said, “Canvas is now available to all ChatGPT users and can execute code! More importantly, it can also still emojify your writing.” This will likely pose a threat to Google Docs. ChatGPT Canvas allows users to edit text before prompting ChatGPT to generate the final copy. Users can also ask ChatGPT to leave comments in their drafts and address them later. In simple words, it is like an editor checking copies. “This is kinda neat. OpenAI Canvas (now out of beta and broadly available) can provide line-by-line comments. It almost looks like an AI-superpowered Google Docs,” Tanishq Mathew Abraham, founder of the Medical AI Research Center (MedARC), said in a post on X. Google Docs recently introduced the ‘Help me write’ feature, which enables users to improve existing documents in various styles or generate entire paragraphs with a simple AI prompt. Now, Google is expanding its capabilities with the ‘Help me create’ feature in Gemini, which helps users create documents from scratch. Meanwhile, Microsoft has introduced Copilot in MS Word, which can automatically generate concise summaries of lengthy documents, providing an overview as soon as the file is opened. Users can also prompt Copilot to create entire drafts or sections based on specific requirements. However, it still falls short of Canvas, which offers real-time editing capabilities. ChatGPT is Now a Code Editor For developers, Canvas provides a robust coding experience. Users can write, debug, and run Python code directly in the interface, with features like syntax highlighting, autocomplete, and inline error diagnostics. “One of the new features we want to demonstrate today is the ability to actually run your Python code within Canvas,” said Alexi Christakis, OpenAI member of the technical staff. The feature is powered by a WebAssembly-based Python emulator, allowing instant execution of Python scripts. Christakis also showed a demo where ChatGPT assisted with debugging. After running the code, Canvas flagged an error,  “Label is not defined.” ChatGPT immediately analysed the exception, identifying that the correct function in Matplotlib should be ‘title’ instead of ‘label’. Once the issue is identified, users can click the ‘Fix Bug’ button. ChatGPT then reviews the code and the error to automatically generate a correction. For transparency, the ‘Show Changes’ feature provides an inline diff, allowing users to see exactly what was modified. “Clicking that will give you an inline diff that should be familiar to people who use Git,” Christakis added. “Today’s Canvas updates from OpenAI may seem a bit understated compared to the other big announcements, but I really like what I saw. I am really eager to explore more AI tools for help with long-form writing,” said Bojan Tunguz, former senior systems software Engineer at NVIDIA. He further added that he is excited about the new feature of code execution in Canvas on top of WASM. “I am a big fan of WASM, and am currently working on a project that relies on it,” he said. “One thing I really like about OpenAI’s new canvas mode is how all the Python code it generates comes with docstrings and adheres to basic type-safe rules following Pylint standards. Nicely done, OpenAI!” posted Daniel Merja, founder and general partner at Founders Committee Ventures. Developers like Kol Tregaskes also pointed out the broader implications, simply saying, “Canvas is now a code editor.” Moreover, Canvas can also generate graphics using Python code. This is similar to Anthropic’s Claude Artifacts, which help developers visualise code in real-time and allow users to see, edit, and build upon their creations instantly. Karina Nguyen, an AI researcher at OpenAI, talked about its potential for learning and experimentation. “We’re excited for Canvas to become a more personalised tutor,” she said. “You can ask ChatGPT to explain mathematical concepts, write code to plot them, and learn visually.”  She added that Canvas can become an automated SWE and a data scientist. Threat to Cursor and VS Code? Two months ago, when Canvas was launched in ChatGPT Plus, people compared it with Claude, Cursor, and GitHub Copilot, and they claimed that with this update, OpenAI might just make a lot of people shift back to ChatGPT for writing and coding since the UX of ChatGPT is much better. “OpenAI launched the Canvas feature in ChatGPT, and it’s insane. Bye-bye to all AI IDEs soon,” said a user on X. However, not everyone thinks the same. “No, Cursor did not just become obsolete by OpenAI Canvas. There’s so much more to AI code editing that can’t be done in a canvas. I need my AI to be in an IDE to be fully productive and stay in control of my code,” posted another user on Recently, Sully Omar, CEO of Cognosys, said, “Wonder why OpenAI\/Claude\/Google haven’t built their own IDE. “I’d gladly pay $500\/$1000 per month for an unlimited usage model that is 100% integrated with my IDE.”  Who knows, Canvas might be a step in that direction.","excerpt":"ChatGPT Canvas allows users to edit text before prompting ChatGPT to generate the final copy.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-12-11T21:52:55","publication_year":"2024","word_count":1012,"keywords":["Anthropic","ChatGPT","OpenAI","AI","medical AI","RAG","Python","Aim","Matplotlib","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","Anthropic","Aim","Matplotlib","RAG","medical AI","Python","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-canvas-kills-google-docs-challenges-vs-code-cursor\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":49917,"title":"Now GitHub Wants To Preserve Codes Forever, Launches Archive Program","content":"In an attempt to preserve open-source codes for future generations, Github recently launched their GitHub Archive Program. This week at the GitHub universe event, the firm announced that they will open the Arctic Code Vault on 2 February 2020 to host open-source repositories. Partnered with Long Now Foundation, the Internet Archive, the Arctic World Archive, and more, GitHub aims to protect codes in a wide range of storage and formats for 1,000 years. Archived data last only for a few years as it gradually rots, which is called bit rot. As a result, over the years, we lose critical information from storage devices whose lifespan is around 30 years. Therefore, to help the future generations in gaining access to the building blocks of today’s superior products, the consortium has planned to store data in media that can last for many hundred years and mitigate the problem of data decay. Approach GitHub will utilise a flexible durable strategy for archiving code called pace layer. The idea is to have a wide range of storage type that can cater to various kinds of needs: from real-time to long-time. Consequently, the program has been partitioned into hot, warm, and cold storage. While the hot storage will be used for real-time requirements and will hold the current GitHub data, the medium storage will be updated on a monthly basis to archive the repository. The cold storage is focused on long-term storage and will be updated every 5+ years. All the three storage is further divided into several types to preserve data by crawling through the GitHub site and keeping it in the Internet Archive, Bodleian Library, among others. But, the facet of this storage is the Project Silica that was developed by the Microsoft Research team. Eventually, all the data of public active repository to safeguard the codes for 10,000 years. Cold Storage Each repository will be stored in a single TAR file and will be encoded with QR for data integrity. GitHub data will be protected in the 250 meters deep decommissioned coal mine of an Arctic mountain situated in Svalbard. While the location is also prone to climate change, it is envisioned that it will not have any impact on the preserved data. Svalbard is one of the safest places on the planet for years as it witnesses only a few natural hazards. The location is also used for archiving other knowledge of humans, such as various types of seeds, the Svalbard Global Seed Vault. How Impactful Will It Be? While GitHub’s primary intention is to preserve open-source for the future generation, it also wants to encourage other organisations to plan for long-term storage. Besides, the firm also illustrates examples of the history of lost technologies in several events such as the Library of Alexandria, Roman Concrete, and more, where humans lost much useful information due to absence of backups. However, those were the times when the information was difficult to move, and new groundbreaking discovery or invention took years or even decades to proliferate. Thus, only a few people used to stay informed about technologies. But in the internet world, when information is moving at a rapid pace, it is easy to keep track of the developments. Besides, today, in ever-changing landscape technology become irrelevant very quickly due to new technologies. Today’s open-source projects might not be as useful as the firm is envisioning it. Outlook Storing of physical things such as seed for the future generation is a wise idea as there is a risk of losing something that will always be useful to humans. Technology in thousands of years from now would be more advanced than it is today. However, preparing for the worst is not a terrible idea and software is definitely the way to go.","excerpt":"In an attempt to preserve open-source codes for future generations, Github recently launched their GitHub Archive Program. This week at the GitHub universe event, the firm announced that they will open the Arctic Code Vault on 2 February 2020 to host open-source repositories. Partnered with Long Now Foundation, the Internet Archive, the Arctic World Archive, […]","categories":["Global Tech"],"tags":["GitHub"],"author_name":"Rohit Yadav","publish_date":"2019-11-14T19:00:00","publication_year":"2019","word_count":628,"keywords":["Go","API","AI","Git","RAG","Aim","llm_models:Bard","GAN","GitHub","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","GitHub","API","GAN","llm_models:Bard"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/now-github-wants-to-preserve-codes-forever-launches-archive-program\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10022187,"title":"Hands-on Guide to The Evolved Transformer on Neural Machine Translation","content":"Evolved Transformer has been evolved with neural architecture search (NAS) to perform sequence-to-sequence tasks such as neural machine translation (NMT). Evolved Transformer outperforms Vanilla Transformer, especially on translation tasks with improved BLEU score, well-reduced model parameters and increased computation efficiency. Recurrent neural networks showed good performance in sequence-to-sequence tasks over a long time. However, the numerous model parameters used in recurrent neural networks necessitated a computationally-effective alternative architecture. Convolutional neural networks, feed-forward networks outperformed traditional recurrent networks in specific tasks. But they could not achieve a generalized architecture to address any sequence problem. Transformers have emerged as a better alternative and a generalized approach incorporating self-attention mechanisms. Transformers become more powerful in solving problems from various platforms, including text, audio, image, and video. A lot of variants and extensions of the Vanilla Transformer are developed in a task-specific manner. Few remarkable examples include the Vision Transformer and the ResNet-backed Vision Transformer for computer vision tasks, and the TransUNet for Medical Image Segmentation tasks. Similarly, the Google Brain team researchers David R. So, Chen Liang, Quoc V. Le have developed Evolved Transformer through the neural architecture search (NAS) approach targeting Neural Machine Translation tasks and they have succeeded! Models developed by neural architecture search have begun to outperform human-developed models in many applications. Neural architecture search is proven to be better than Reinforcement Learning especially when the training resources are limited. In this work, the famous tournament selection architecture is applied to do a model search. The vanilla Transformer is employed to warm start the search. The search has been performed directly on the WMT 2014 English-German translation task with the newly-developed Progressive Dynamic Hurdles (PDH) algorithm. As a result of the search, a new model has been evolved that outperformed the vanilla Transformer on four well-established language tasks: WMT 2014 English-German (En-De) translation task, WMT 2014 English-French (En-Fr) translation task, WMT 2014 English-Czech (En-Cs) translation task and the 1 Billion Word Language Model Benchmark (LM1B). This model has been named Evolved Transformer, shortly known as the ET. Comparison of the encoder blocks between the Vanilla Transformer and the Evolved Transformer (Source) Comparison of the decoder blocks between the Vanilla Transformer and the Evolved Transformer (Source) PyTorch Implementation of the Evolved Transformer The pre-built, pre-trained architecture of Evolved Transformer runs best in the GPU or TPU devices. The model and the necessary files can be downloaded from the source repository using the following command. !git clone https:\/\/github.com\/Shikhar-S\/EvolvedTransformer.git Proper download of files can be ensured by running the following command. !ls EvolvedTransformer\/ Output: The environment with dependencies can be created using the following commands. %%bash cd EvolvedTransformer\/ pip3 install -r requirements.txt # install spacy python3 -m spacy download en Once the environment is created, the model can be retrained or evaluated with the in-built dataset or a custom dataset. The following codes run text classification on the in-built AG_NEWS dataset. %%bash cd EvolvedTransformer\/ python3 main.py # run on Evolved Transformer’s encoder python3 main.py --evolved true A generalized Evolved Transformer block has been published as a Class in the PyTorch environment. Any custom architecture can incorporate this Class to build a new model on top of the Evolved Transformer. The following codes establish the EvolvedTransformerBlock Class. from models.embedder import Embedder, PositionalEncoder import math import torch import torch.nn as nn import torch.nn.functional as F from models.gated_linear_unit import GLU class EvolvedTransformerBlock(nn.Module): def __init__(self,d_model,num_heads=8,ff_hidden=4): super(EvolvedTransformerBlock,self).__init__() self.attention = nn.MultiheadAttention(d_model, num_heads) self.layer_norms = nn.ModuleList([nn.LayerNorm(d_model) for _ in range(4)]) self.feed_forward = nn.Sequential( nn.Linear(d_model,ff_hidden*d_model), nn.ReLU(), nn.Linear(ff_hidden*d_model,d_model), ) self.glu = GLU(d_model,1) self.left_net = nn.Sequential( nn.Linear(d_model,ff_hidden*d_model), nn.ReLU() ) self.right_net = nn.Sequential( nn.Conv1d(in_channels=d_model,out_channels=d_model\/\/2,kernel_size=3,padding=1), nn.ReLU() ) self.mid_layer_norm=nn.LayerNorm(d_model*ff_hidden) self.sep_conv=nn.Sequential( nn.Conv1d(in_channels=d_model*ff_hidden,out_channels=1,kernel_size=9,padding=4), nn.Conv1d(in_channels=1,out_channels=d_model,kernel_size=1) ) def forward(self,x): glued = self.glu(self.layer_norms[0](x))+x glu_normed = self.layer_norms[1](glued) left_branch = self.left_net(glu_normed) right_branch = self.right_net(glu_normed.transpose(1,2)).transpose(1,2) right_branch = F.pad(input=right_branch, pad=(0,left_branch.shape[2]-right_branch.shape[2],0,0,0,0), mode='constant', value=0) mid_result = left_branch+right_branch mid_result = self.mid_layer_norm(mid_result) mid_result = self.sep_conv(mid_result.transpose(1,2)).transpose(1,2) mid_result = mid_result + glued normed = self.layer_norms[2](mid_result) normed=normed.transpose(0,1) attended = self.attention(normed,normed,normed,need_weights=False)[0].transpose(0,1) + mid_result normed = self.layer_norms[3](attended) forwarded = self.feed_forward(normed)+attended return forwarded Custom configurations to the pre-built model can be made with the following codes. import argparse import logging import utils logger = utils.get_logger() def str2bool(v): return v.lower() in ('true') parser = argparse.ArgumentParser() parser.add_argument(\"--batch\",type=int,default=16) parser.add_argument(\"--evolved\",type=str2bool,default=False) parser.add_argument(\"--epochs\",type=int,default=10) parser.add_argument(\"--model_dim\",type=int,default=32) parser.add_argument(\"--max_seq_len\",type=int,default=200) parser.add_argument(\"--backend\",type=str,default='auto',choices=['cpu', 'gpu','auto']) parser.add_argument('--ngrams',type=int,default=2) parser.add_argument('--train_split',type=float,default=0.95) def get_args(): logger.info('Parsing arguments') args,unparsed = parser.parse_known_args() return args, unparsed TensorFlow Implementation of the Evolved Transformer TensorFlow implementation of the Evolved Transformer is performed through the Tensor2Tensor (T2T) framework. It yields a highly-efficient pre-trained model that can be implemented in minimal time even in a CPU device. The following codes install Tensor2Tensor and its dependencies in the local machine or cloud environment. !pip install tensor2tensor For the custom application of the Evolved Transformer or to build architecture on top of it, a module is developed in the models Class of Tensor2Tensor framework that can be imported using the following command. from tensor2tensor.models import evolved_transformer Tensor2Tensor integrates a lot of famous models and datasets at one place. The following command gives the list of pre-trained models, datasets and suitable hyperparameters. Users can choose any model and problem of interest and run either in a terminal or as python code. !t2t-trainer --registry_help The Evolved Transformer can be invoked along with an example translation problem using the following commands. Here the base CPU version of the model is run for the sake of simplicity. The WMT 2014 English-to-German translation task is chosen as our problem. It should be noted that customized training may take hours based on the device configuration. Setting the initial parameters, defining the problem and the model, %%bash PROBLEM=translate_ende_wmt32k MODEL=evolved_transformer HPARAMS=evolved_transformer_base DATA_DIR=$HOME\/t2t_data TMP_DIR=\/tmp\/t2t_datagen TRAIN_DIR=$HOME\/t2t_train\/$PROBLEM\/$MODEL-$HPARAMS mkdir -p $DATA_DIR $TMP_DIR $TRAIN_DIR # Generate data t2t-datagen \\ --data_dir=$DATA_DIR \\ --tmp_dir=$TMP_DIR \\ --problem=$PROBLEM Training the model based on our requirements, %%bash # Train # *  If you run out of memory, add --hparams='batch_size=1024'. t2t-trainer \\ --data_dir=$DATA_DIR \\ --problem=$PROBLEM \\ --model=$MODEL \\ --hparams_set=$HPARAMS \\ --output_dir=$TRAIN_DIR Running decoder with a sample input English-German text pair, %%bash # Decode DECODE_FILE=$DATA_DIR\/decode_this.txt echo \"Hello world\" >> $DECODE_FILE echo \"Goodbye world\" >> $DECODE_FILE echo -e 'Hallo Welt\\nAuf Wiedersehen Welt' > ref-translation.de BEAM_SIZE=4 ALPHA=0.6 t2t-decoder \\ --data_dir=$DATA_DIR \\ --problem=$PROBLEM \\ --model=$MODEL \\ --hparams_set=$HPARAMS \\ --output_dir=$TRAIN_DIR \\ --decode_hparams=\"beam_size=$BEAM_SIZE,alpha=$ALPHA\" \\ --decode_from_file=$DECODE_FILE \\ --decode_to_file=translation.en Visualizing the translation performance for the predefined test text, %%bash # See the translations cat translation.en Evaluating the BLEU score metric and comparing it with a reference value %%bash # Evaluate the BLEU score t2t-bleu --translation=translation.en --reference=ref-translation.de Wrapping up The Evolved Transformer outperforms the Vanilla Transformer with a BLEU score of 29.8 with the WMT 2014 English-German translation task and has 37.6% lesser parameters. Thus it utilises relatively lesser memory and demonstrates greater computational efficiency. Comparison of performance between the Vanilla Transformer and the Evolved Transformer (Source) Further reading: Original research paper Source code repository – PyTorch Source code repository – TensorFlow WMT translation task 2014","excerpt":"Evolved Transformer has been evolved with neural architecture search (NAS) to perform sequence-to-sequence tasks such as neural machine translation (NMT)","categories":["AI Trends"],"tags":["attention","attention mechanism","Generative Pre-Trained Transformer","neural architecture search","neural machine translation","Pytorch","self attention models","self-attention","Tensorflow","Transformers"],"author_name":"Rajkumar Lakshmanamoorthy","publish_date":"2021-03-15T17:00:00","publication_year":"2021","word_count":1114,"keywords":["TPU","computer vision","Generative Pre-Trained Transformer","Pytorch","self-attention","PyTorch","attention","Tensorflow","text classification","neural machine translation","AI","neural network","spaCy","self attention models","Transformers","Python","neural architecture search","TensorFlow","attention mechanism"],"extracted_tech_keywords":["AI","neural network","computer vision","TensorFlow","PyTorch","Transformers","spaCy","text classification","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/hands-on-guide-to-the-evolved-transformer-on-neural-machine-translation\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10119531,"title":"Confluent Targets India as Top 10 Market","content":"Confluent, the data streaming company founded by the creators of Apache Kafka, is focusing on India as one of the top 10 markets for its business growth. The company announced this during the first-ever Kafka Summit in Bangalore on May 2, 2024. “India is a very critical market for us. It is a top 10 market, probably higher up,” Hemanth Vedagarbha, the global head of sales at Confluent. “It defines our APAC strategy. Core to our APAC strategy is gaining market share in India.” Confluent’s strength in India lies in the banking, financial services, digital native, and telecom sectors. The company’s go-to-market strategy in India is driven by partnerships and direct sales. “India’s always been an incredibly important part of our team philosophy,” said Jay Kreps, co-founder and CEO of Confluent at the summit. “We have very significant parts of our R&D operations here, including massive product areas, and almost every team is represented in India.” With approximately 20% of its global workforce based in the country, Confluent has a large engineering team in Bangalore and a distributed engineering presence across India. The company has a sales team, support staff, and engineering headquarters in Bangalore and a distributed engineering presence across the country. The company is investing in building its partner ecosystem in India, which includes cloud service providers like AWS, Azure, and Google Cloud Provider, system integrators, Independent Software Vendors, resellers, and Managed Service Providers. Confluent’s pricing is universal globally, but it offers flexibility based on commitment levels to cater to price-sensitive customers in India. Confluent also announced new capabilities to simplify AI integration and stream processing during the summit. These include AI Model Inference for Apache Flink, Confluent Platform for Apache Flink, and cost-saving Freight clusters. The company believes that the increasing digitisation in India, including government initiatives like Digital India, presents a massive opportunity for its data streaming technology. The Kafka Summit in Bangalore is part of Confluent’s efforts to build broader awareness and drive demand in the Indian market. The company plans to expand its presence beyond tier-one cities and penetrate deeper into regional banks, non-banking financial services, healthcare, telecom, and gaming sectors.","excerpt":"Confluent singles out India as a critical market for business growth.","categories":["AI News"],"tags":["AI Summit","apache kafka","bangalore","Confluent"],"author_name":"K L Krithika","publish_date":"2024-05-03T15:06:47","publication_year":"2024","word_count":357,"keywords":["Go","stream processing","cloud_platforms:Azure","AWS","AI","cloud_platforms:AWS","R","Git","AI Summit","Kafka","apache kafka","Confluent","Azure","bangalore"],"extracted_tech_keywords":["AI","AWS","Azure","Kafka","R","Go","Git","stream processing","cloud_platforms:AWS","cloud_platforms:Azure"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/confluent-targets-india-as-top-10-market\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119625,"title":"Andrew Ng, Comet Partner for New Course on Prompt Engineering for Vision Models","content":"Andrew Ng’s DeepLearning.AI has announced a free short course on Prompt Engineering for Vision Models, offered in collaboration with Comet ML. This course is intended to give a comprehensive introduction to the concept of prompt engineering, specifically tailored for vision models. It is one one-hour course under instruction by Abby Morgan (ML Engineer), Jacques Verré (Head of Product) and Caleb Kaiser (ML Engineer). To get the most out of this course, it is recommended to have a Python experience. To stay relevant on the bleeding edge, you will promote different vision models, such as Meta’s Segment Anything Model (SAM), a universal image segmentation model, OWL-ViT, a zero-shot object detection model, and Stable Diffusion 2.0. Participants will gain knowledge of how they can generate images efficiently using hyperparameters like strength, guidance scale, and number of inference steps. Furthermore, you can select specific parts of an image by providing coordinates or drawing a box around the area you want to isolate and by combining all of these, you will be able to replace objects within an image with generated content. The best part is you can generate custom images based on pictures of people or places that you provide using a fine-tuning technique called DreamBooth. This course will utilise Comet,  a library to track experiments and optimise visual prompt engineering workflows. Previously, Andrew Ng had partnered with Google Cloud for the LLMOps to equip learners with the practical skills and knowledge needed to work with LLMs and build LLMOps pipelines in real-world applications.","excerpt":"Participants are recommended to have a Python experience.","categories":["AI News"],"tags":["Andrew Ng"],"author_name":"Sagar Sharma","publish_date":"2024-05-06T12:49:51","publication_year":"2024","word_count":251,"keywords":["Go","AI","ML","Andrew Ng","Python","prompt engineering","Comet ML","object detection","stable diffusion","GAN","R"],"extracted_tech_keywords":["AI","ML","Comet ML","prompt engineering","object detection","Python","R","Go","stable diffusion","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/andrew-ng-comet-partner-for-new-course-on-prompt-engineering-for-vision-models\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10067746,"title":"First-ever conference on data engineering wraps up | Watch the sessions now on-demand!","content":"India’s first conference focused on data engineering, the Data Engineering Summit (DES22), presented by Google Cloud and organised by Analytics India Magazine on April 30, ended on a high note. The conference had 11 tech talks, four workshops, and a panel discussion, and over 10,000 participants from leading companies, startups, and institutions attended the DES22. The talks and workshops covered topics like data transformation, data insights, data lakes, data strategies, and data training and governance. Overall, the summit offered a great opportunity for professionals in the data engineering space to network and gain insights into the latest advancements in the field. Google Cloud, Oracle, Publicis Sapient, TheMathCompany, Tiger Analytics, USEReady, and iMerit sponsored the conference. The DES22 offered a well-rounded and highly engaging discussion on how organisations can leverage the potential of efficient and product data engineering workflow. Be a part of the conversation here! DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. Key offerings Google Cloud’s customer engineering director Subram Natarajan kicked off the conference with a keynote speech on how cloud infrastructure can be transformative in a company’s journey. With the prevalence of data, Natarajan said, we need new ways to use data and democratise it. In addition, there is an urgent need to migrate data responsibly and securely while maintaining accessibility. Following this, Vijay Yellapragada, executive director at EY GDS Data Analytics, spoke about unlocking business value through data fabric. The session shed light on how data fabric represents a paradigm shift in how companies leverage their data. Next came Oracle’s Anvita Bajpai. Her talk titled, ‘Data Insights using AI and ML for Business Benefit,’ was about the best ways to utilise data to benefit the business and its customers. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. The talk was followed by a session on Next-Generation Data Lake led by Muthu Govindarajan and Manish Gupta of Tiger Analytics. The presentation explored new and interesting trends in the industry. In addition, the duo also gave a gist of the latest developments in the data lake landscape. The MathCompany’s Varun Saraogi spoke about their Ops first approach and stressed the need to build trust in the data & AI Ops. He made a case for Ops as a critical component in stabilising pipelines. Rashmi Purbey and Sumit Jindal from Publicis Sapient spoke about building a resilient, scalable modern data platform, Rashmi explained how Databricks combines the best of both worlds, data warehousing and data lake. While building a resilient, scalable data platform, Sumit said, businesses focus on the platform itself and not the analytics powering it. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. The panel discussion on powering business growth with an intelligent data strategy was led by Google Cloud’s Rajesh Ramdas, Zepto’s Yash Dayal, FreshToHome’s Saurabh Odhyan, Fino Payments’ Kunal Sakariya, Axtria’s Aditya Bhandari, Wakefit’s Puneet Tripathi, Near’s Ravi Kaushik, and Junglee Games’ Deependra Singh. The discussion covered the ins and outs of powering business growth with an intelligent data strategy. The panel discussion was followed by a session on BI modernisation by Tiger Analytics’ Saravanan Perumal and Riaz Abdul Samadh. They spoke about business intelligence solutions enterprises can use to build modern, future-proof analytics solutions. iMerit’s Sudeep George’s talk was on training data for effective AI deployment. He shared practical insights about handling training data for complex AI systems operating at scale. After this, Shiv Nadar University’s Vallurupalli Vamsi, Satyam Mukherjee, and Jaideep Ghosh gave a presentation on data engineering and analytics in business. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. In the next session, Google Cloud’s Kirthi Ganapathy spoke about intelligence and unified data governance in the multi-cloud age. Kirthi shared insights, key learnings and best practices around intelligent management of metadata, security and governance in a diverse and largely distributed data environment. The summit concluded with a session on Analytics 3.0. IIIT Bangalore’s S Sadagopan and USEReady’s Amit Phatak and Uday Hegde explored the evolution of analytics and how organisations and practitioners can ride the analytics wave over the next decade. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. Workshops The first workshop, titled ‘Building Enhanced Customer Micro Segments to drive engagement by building customer 360’, was presented by Google Cloud’s Sai Sriparasa. He demonstrated the best practices around setting up a data lake to create a Customer 360 view. Oracle’s Aaron Hoile’s workshop titled ‘Be an AI Wizard: No ML magic required’ spoke about how one can take advantage of AI capabilities as a developer without any machine learning expertise. USEReady Technology’s Mohamed Anas’ workshop on Snowflake Data Lakes covered building data lakes in an unconventional way by making the best use of the given tech stack. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. Finally, Vaibhav Kumar from the Association of Data Scientists hosted the last workshop of the event, titled ‘​​Fast-Track Data Engineering.’ The session covered how to prepare data, different methods of data pre-processing, feature engineering and munging steps with hands-on experiments. DES22 is also hosting the Great Indian Data Engineering Hackathon, organised exclusively for data engineers, collaborating with MachineHack, a leading hackathon and competitive coding platform for ML developers. The hackathon will end on May 30, 2022. (Register here to participate). Data governance is at the heart of any data engineering solution and has always been tricky, and a multi-cloud environment makes it all the more complex. So the insights into how Google is approaching this are useful and exciting, said Gaurav Deshpande, VP of Strategy, Business Development in Digital Transformation at Logixal Inc. Thanks to our generous sponsors and an engaging audience, DES22 has been a runaway success, and we look forward to curating great sessions for you in the future as well. In addition, the conference can be accessed on-demand here. DES22 IS NOW AVAILABLE TO WATCH ON-DEMAND. CHECK IT OUT HERE. Check out our upcoming events and conferences: Deep Learning DevCon 2022 (Click here to buy tickets)MachineCon 2022 (Click here to buy tickets) Cypher 2022 (Click here to buy tickets) DES22 – WATCH ON-DEMAND NOW!","excerpt":"Over 10,000 participants from leading companies, startups, and institutions attended the DES22.","categories":["Deep Tech"],"tags":["conference","Data Engineering","data engineering demand","Data Engineering Summit","DES","Google Cloud"],"author_name":"Anushka Pandit","publish_date":"2022-05-24T16:56:05","publication_year":"2022","word_count":1019,"keywords":["Go","Data Engineering Summit","Google Cloud","machine learning","AI","conference","ML","DES","data engineering demand","RAG","Databricks","deep learning","Data Engineering","analytics","R","Snowflake"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","analytics","RAG","Snowflake","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/first-ever-conference-on-data-engineering-wraps-up-watch-the-sessions-now-on-demand\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10169494,"title":"Nutanix and Pure Storage Partner to Deliver Integrated Solution for Critical Workloads","content":"Nutanix and Pure Storage have announced a strategic partnership to deliver a deeply integrated infrastructure solution. The solution is designed to support high-performance, mission-critical workloads, including AI, with scalability, security, and flexibility. The announcement was made at the NEXT Conference in Washington, DC. The collaboration combines the Nutanix Cloud Platform, powered by the AHV hypervisor and Nutanix Flow virtual networking and security, with Pure Storage’s FlashArray, leveraging NVMe\/TCP connectivity. The goal is to help enterprises modernise virtual workloads and manage them more efficiently in an evolving virtualisation landscape. The new solution is being co-developed with input from industry partners and early adopters. “Our integrated solution will be ideally suited for companies with storage-rich environments looking for choices in modernisation,” Tarkan Maner, chief commercial officer at Nutanix, said. The solution will be compatible with server hardware partners currently supporting Pure FlashArray, including Cisco, Dell, HPE, Lenovo, and Supermicro. Additionally, Cisco and Pure Storage are expanding their long-standing partnership of more than 60 validated FlashStack designs to now include Nutanix in the portfolio. “With more than 13,500 global customers, I’m hearing more than ever that organisations of all shapes and sizes have a growing need for efficient, flexible, and high-performance solutions that can also scale to support their most critical, data-intensive applications,” said Maciej Kranz, general manager, enterprise at Pure Storage. However, the integrated solution is currently under development. Early access is expected by summer 2025, with general availability slated for the end of the year through Nutanix and Pure Storage channel partners.","excerpt":"The new solution is being co-developed with input from industry partners and early adopters.","categories":["AI News"],"tags":["AI"],"author_name":"Shalini Mondal","publish_date":"2025-05-09T11:07:14","publication_year":"2025","word_count":251,"keywords":["Go","programming_languages:R","AI","Scala","RAG","Ray","ViT","programming_languages:Scala","GAN","R"],"extracted_tech_keywords":["AI","Ray","RAG","R","Go","Scala","GAN","ViT","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nutanix-and-pure-storage-partner-to-deliver-integrated-solution-for-critical-workloads\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":22482,"title":"Lymbyc Solutions Creates World’s First-Of-A-Kind Virtual Analyst","content":"Lymbyc Solutions, a predictive analytics solutions provider, has developed a first of its kind virtual analyst platform which helps other companies with data analysis and assist with useful business insights. The virtual analyst product named ‘Lymbyc’ is driven by an adaptive machine learning engine which crunches data and processes insights quickly. Lymbyc uses artificial intelligence for feeding inputs in building What-Ifs scenario. The company says that it is the world’s first virtual data scientist and possesses the same intellectual capabilities to that of a data analyst. Notably, it primarily focuses on giving real-time predictive decision-making for consumer and enterprise sectors. Companies can utilise all the relevant data to witness Lymbyc’s potential. At the launch of Lymbyc, Satyakam Mohanty, founder of Lymbyc Solutions, said, “Big Data Analytics have often been used as a validation resource and existing solutions help organisations answer the ‘what’ and ‘how’ kind of business queries at best, but cannot help them understand the ‘why’. We wanted to create a solution that truly mimics analysts and data scientists and does not stop at being just an intuitive BI system. Lymbyc enables organizations to undertake data discovery and predictive insight generation through a proprietary NLP-based querying language that harnesses Artificial Intelligence, thus helping them to take strategic decisions in real-time across enterprise and consumer spaces. It automatically learns what is relevant for your business with every interaction and continually curates information for future sharing.” Directed at corporate leaders to help with decision-making during critical market situations , Lymbyc will surely provide them with quicker results and tackles time and information complexity on a larger basis. With technologies such as self-driving cars, image and speech recognition developing at a rampant pace, Lymbyc has its market approach towards business consumers by coupling artificial intelligence and machine learning domains.","excerpt":"Lymbyc Solutions, a predictive analytics solutions provider, has developed a first of its kind virtual analyst platform which helps other companies with data analysis and assist with useful business insights. The virtual analyst product named ‘Lymbyc’ is driven by an adaptive machine learning engine which crunches data and processes insights quickly. Lymbyc uses artificial intelligence […]","categories":["AI News"],"tags":[],"author_name":"Abhishek Sharma","publish_date":"2018-03-09T06:11:16","publication_year":"2018","word_count":297,"keywords":["big data","machine learning","artificial intelligence","programming_languages:R","AI","R","NLP","analytics","GAN","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","analytics","predictive analytics","R","big data","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meet-lymbyc-worlds-first-virtual-analyst\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":49920,"title":"INSOFE In Collaboration With CICE, Carleton University, Canada Launches A PGP Global Program In India","content":"With immense expertise in the Data Science Education space, INSOFE is delighted to announce the INSOFE-Carleton Global Academy’s – PGP Global Program that is designed to meet the industry standards and prepare students for the most demanding & rewarding jobs in Data Science Consulting, R&D and Product Development. “The PGP Global Program is exclusively designed to create data scientists with exceptional consulting and product development skills.” INSOFE-CICE, Carleton University, Ottawa’s PGP Global Program enables students to learn, understand and equip the skills required to develop exciting products by working with International firms in Canada. Students who opt for the PGP Global Program go through the rigorous training in Data Science for 6-months at INSOFE and 3-months at CICE, Carleton University, Canada. Program Highlights Learn from world-class faculty of INSOFE, India and Carleton University, Canada Work in Canada with International firms using cutting edge technologies Grab an opportunity to work in top analytics companies with highly paid packages Students will find the best value for money to boost their profiles’ and give it the hike it deserves INSOFE’s Career Services Team will provide 2 Guaranteed interviews with Top-tier analytics companies Triple Certifications Along with the right exposure to the World of Data Science, and a great networking opportunity with techno-managerial entrepreneurs, students will also receive three certificates. INSOFE’s PGP in Data ScienceCertificate in Technology Product Management and Innovation by Carleton University, CanadaLetter of Work completion from the company where industry project is done Speaking about the global program, Dr Murthy, President & Co-Founder – INSOFE, said, “PGP Global is a very challenging and selective program. It is aimed for those who are very interested in programming and want to take up challenging roles as AI engineers in product and R&D companies or top-notch tech consulting firms. The admission process has an entrance examination by INSOFE, and an interview by CICE, Global academy.” “For the first 2 trimesters, the students go through a rigorous Data Science program in one of the INSOFE campuses in India. They take 10 credits of technical courses, 3 credits of specialization and 2 credits of business and humanities courses designed around data science.  They will also start on their internship project and start interacting with the Canadian firm where they do their Internships. As you can see, this itself is sufficient to land a job as a data scientist job,” said Dr Murthy Karen Schwartz, Associate Vice-President (Research & International) added, “Canada India Centre of Excellence, in short CICE, is an initiative of Carleton University and the Indo – Canadian Community. Located in Ottawa, the centre aims at providing a platform to the Indian students who would like to explore emerging fields through events, research and partnerships. There are some prestigious programs on offer at the CICE through Carleton University that address Public Policy and Administration, Business, Design, and Engineering apart from Architecture and Urbanism.” He also mentioned that the Faculty of Science assures an approach in research that can conceptualize and address the challenges that are faced by both the countries. This interdisciplinary nature of various faculties ensures that the students have an appetite for quality research and a penchant for product development. So, folks, the PGP global program is a multicultural exposure that is associated with soft skills and career skills in R&D and product development. To know more about the global program, visit: PGP Global.","excerpt":"With immense expertise in the Data Science Education space, INSOFE is delighted to announce the INSOFE-Carleton Global Academy’s – PGP Global Program that is designed to meet the industry standards and prepare students for the most demanding & rewarding jobs in Data Science Consulting, R&D and Product Development. “The PGP Global Program is exclusively designed […]","categories":["AI Trends"],"tags":["Data Science","education","Insofe","pgp program in data science"],"author_name":"Prajakta Hebbar","publish_date":"2019-11-14T19:30:00","publication_year":"2019","word_count":561,"keywords":["data science","Go","programming_languages:R","AI","innovation","education","programming_languages:Go","Insofe","pgp program in data science","Aim","analytics","Data Science","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/insofe-in-collaboration-with-cice-carleton-university-canada-launches-a-pgp-global-program-in-india\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":53284,"title":"BCCI To Set Up New Data Analytics Wing To Revamp Indian Cricket","content":"In a fresh start to the decade, India’s National Cricket Academy (NCA) is all set to get a data analytics wing and medical panel as a part of the Board of Control for Cricket in India’s (BCCI) revamp plan. These decisions were results of a meeting that was attended by BCCI president Sourav Ganguly, NCA head Rahul Dravid, and secretary Jay Shah. Though there hasn’t been much use of conventional analytics tools in the past, BCCI raised some eyebrows last year, when it announced the squad for the 2019 cricket world cup. Before the announcement of the squad, the selection committee was detailed about various concerns regarding the tournament. In a presentation that lasted three-and-a-half hours, CKM Dhananjay, the Indian team’s data analyst used a data analytics software to show how a certain player performs and doesn’t, pitch condition in England and other such factors. Based on this report, it was alleged that Rishab Pant, who was in the probable list, was left out of the squad. Right now, the extent to which BCCI would be deploying its new analytics wing is still uncertain. However, we can clearly draw parallels from other sports that have incorporated analytics into their game plans. In a push to access the state-of-the-art medical assistance, BCCI also plans to collaborate with UK’s Fortius clinic. Since its opening in 2011, Fortius Clinic has become the clinic of choice for many of the elite sports professionals. It also has attained the status of the FIFA Medical Centre of Excellence in 2014. Fortius is also known for developing a cloud-based sports injury data collection system in an attempt to study chronic injuries and also to make this data available to other clinics. This can be of great help to the cricketers who strive for long injury-free careers. Good To Great With Data Every athlete has access to the same number of 24 hours. If we exclude luck, opportunities are pretty much equally distributed. Then what turns these athletes into legends? One very obvious reason can be the coaching team. Boxing legend Mike Tyson and his coach Cus D’Amato are one such example. But what separates a good coach from a great coach is their great intuition for the game. Historically, much of the intuition came from watching hours and hours of old footage of the opponents, of their own selves and then identifying patterns. However, today we have built computers that do the pattern identification for us. A whole new branch has emerged out of this — machine learning\/data analytics. The raw data that is generated after every match was monstrous, and, since a model is only as good as the data it is fed, sporting organizations and startups have a great advantage building models to their customised requirements. This information can now be used to drive value across all aspects of their organization from selling more tickets to preventing injuries in players. For instance, companies like DataRobot has assisted these kinds of organizations to apply AI and machine learning to power their insights and decisions both on and off the field. In India, there are at least half a dozen startups that have already surfaced over the past couple of years. 2017 ICC Champions Trophy especially, had highlighted the role of technology when a bat sensor called Speculur, developed by Intel, was deployed to track the key elements of a batter’s swing. Bengaluru based Spektacom and Stancebeam are two notable companies that have really made the noise globally. Stancebeam especially has tied up with Kookaburra to extend its services internationally. via Microsoft Whereas,Spektacom has been founded by Anil Kumble, India’s most successful bowler and a legend of the game. Spektacom’s vision is to revolutionize the sporting experience by creating an ecosystem that engages the fans, coaches, and professionals through state-of-the-art video analytics, artificial intelligence, video modelling, and augmented reality. Spektacom has also partnered with Microsoft for its cognitive AI service which has been powered by the company’s Azure cloud offering for real-time analytics and data management capabilities. Outlook Over the last 2  years, the Indian sports scene has witnessed a tremendous number of innovations with data analytics. Wearable tech especially has been revolutionary, as they could be used to track the performance metrics in real-time. Though we have seen tech advancements in the form of video analytics and wearable sleeves for bowlers, there is still an untapped market for delivering IoT-based, insightful coaching for batsmen in an intuitive and granular way — for achieving that extra 1%.","excerpt":"In a fresh start to the decade, India’s National Cricket Academy (NCA) is all set to get a data analytics wing and medical panel as a part of the Board of Control for Cricket in India’s (BCCI) revamp plan. These decisions were results of a meeting that was attended by BCCI president Sourav Ganguly, NCA […]","categories":["Deep Tech"],"tags":["Data Analytics"],"author_name":"Ram Sagar","publish_date":"2020-01-07T13:00:33","publication_year":"2020","word_count":750,"keywords":["Go","machine learning","artificial intelligence","AI","R","innovation","GAN","analytics","real-time analytics","Data Analytics","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Azure","R","Go","real-time analytics","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/bcci-data-analyticswing-national-cricket-academy-ganguly-dravid\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":36720,"title":"Now You Can Build Faster, Lighter ML Models On Your Smartphone With TensorFlow Lite","content":"TensorFlow’s machine learning platform has a comprehensive, flexible ecosystem of tools, libraries and community resources. This lets researchers push the state-of-the-art developments in ML and developers easily build and deploy ML-powered applications. At the recently concluded TensorFlow’s developer summit, along with TensorFlow 2.0, the team also announced open sourcing of TensorFlow Lite for mobile devices and two development boards Sparkfun and Coral which are based on TensorFlow Lite for performing machine learning tasks on handheld devices like smartphones. With TensorFlow Lite it looks to make smartphones, the next best choice to run machine learning models. TensorFlow Lite is TensorFlow’s lightweight solution for mobile and embedded devices. It enables on-device machine learning inference with low latency and small binary size. TensorFlow Lite supports a set of core operators, both quantised and float, which has been tuned for mobile platforms. Why Lite, Why Now? There is an urge, now, amongst the makers of smart devices to cater to the growing desires of their users to have devices which are heavy on specs while light on duty; accelerated hardware devices. Now the devices have human-like interactions, with vision and speech applications. TensorFlow Lite accommodates this lighter way of functioning via Android Neural networks API. This opens up a whole new paradigm of possibilities for on-device intelligence. Be it information security or accelerated responses, machine learning models have shown great promise. And, what better way to deploy these models than TensorFlow? Google has been using TensorFlow Lite for taking pictures on its flagship model Pixel. For Portrait mode on Pixel 3, Tensorflow Lite GPU inference accelerates the foreground-background segmentation model by over 4x and the new depth estimation model by over 10x vs CPU inference with floating point precision. This was made possible with accelerating compute-intensive networks that enable vital use cases for the users. Source: TensorFlow Lite The API for calling the Python interpreter is tf.lite.Interpreter Converting TensorFlow Keras model to TensorFlow Lite model: converter = tf.lite.TFLiteConverter.from_keras_model_file(keras_file) The user can deploy pre-trained Tensorflow Probability models, Tensorflow KNN, Tensorflow K-mean model on Android by converting the TF models to TF Lite (guide), and the converted model can be bundled in the Android App Key Takeaways From TF Lite Announcement TensorFlow Lite: Run custom models on mobile platforms via a set of core operators tuned for this task. A new file format based on FlatBuffers. A faster on-device interpreter TensorFlow converter to convert TensorFlow trained models into Lite format. Using TensorFlow Lite cuts down the size of the models by 300 KB which allocates faster deployment. Sparkfun Development Board: Uses extremely low power, less than 1mW in lot of cases. A single coin battery can run for many days. Runs entirely on-device Uses 20KB model Uses less than 100 KB of RAM and 80 KB of Flash. Coral Development Board: Provides on-device machine learning acceleration. Uses Google Edge TPU, which does not depend on the network connection and can perform tasks like object detection under 15 ms. For instance, a shop floor personnel with no prior knowledge of machine learning can use the device just by training within seconds thanks to Edge TPU and perform object detection tasks. Runs inference with TensorFlow Lite. The following state-of-the-art research models can be easily deployed on mobile and edge devices: Image classification Object detection Pose estimation Semantic segmentation Check the TensorFlow Github repository here. Also Watch:","excerpt":"TensorFlow’s machine learning platform has a comprehensive, flexible ecosystem of tools, libraries and community resources. This lets researchers push the state-of-the-art developments in ML and developers easily build and deploy ML-powered applications. At the recently concluded TensorFlow’s developer summit, along with TensorFlow 2.0, the team also announced open sourcing of TensorFlow Lite for mobile devices […]","categories":["Deep Tech"],"tags":["TensorFlow Lite"],"author_name":"Ram Sagar","publish_date":"2019-03-22T06:41:48","publication_year":"2019","word_count":558,"keywords":["machine learning","Keras","TPU","AI","neural network","ML","TensorFlow Lite","Python","object detection","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","TensorFlow","Keras","object detection","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/now-you-can-build-faster-lighter-ml-models-on-your-smartphone-with-tensorflow-lite\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":40509,"title":"E3 2019: Which Are The Companies Moving Into Cloud Gaming In A Big Way?","content":"As the future of gaming moves towards the cloud, long-time industry leaders and veterans are increasing focus towards their cloud operations. While there have been many ‘on-demand’ gaming services in the past, rising cloud infrastructure has increased accessibility to a whole new level. Moreover, cloud gaming generally tends to be in favour of those who already have the infrastructure in place. This is due to the geographical diversity of the products’ target audience, which requires a wide distribution of cloud servers. This gives the advantage to cloud players like Google and Microsoft, both of which now have a stake in gaming. At the Electronic Entertainment Expo (E3) this year, most of the participants in the ecosystem announced their cloud gaming plans. Here’s the lowdown. Microsoft xCloud One of the industry leaders in the console market, Microsoft is looking to future proof their operations. They are also one of the biggest players in the cloud market, especially in emerging economies. Microsoft’s initiative is called ‘xCloud’ and works in conjunction with other Microsoft products such as the Xbox One and Windows. They had released a product known as Xbox Game Pass, which allows users to play whichever games they want to, on whichever supported platform for a flat subscription fee. The head of Xbox, Phil Spencer, stated that xCloud will either leverage a console in the cloud or the one in users’ living rooms. Any game on the console or in the library can be streamed to a mobile phone seamlessly. PlayStation Now Even as Sony’s PlayStation Now has been around for a while now, it is set to get an update. Sony recently announced a partnership with Microsoft for cloud infrastructure, which is sure to make up for their lack of knowledge in the cloud space. PlayStation Now is also being extended to Android devices apart from Sony’s Xperia lineup. The service allows users to access over 750 games on demand, with the catalog expanding over the PS2, PS3 and PS4’s collections. It functions similarly to the Xbox Game Pass and is set to be used alongside the Remote Play functionality on the console. Remote Play, PC support, mobile support and PlayStation Now, along with a solid partnership with Microsoft for cloud gaming capabilities, will allow Sony to take on other players in the space. Google Stadia Google Stadia could be credited with bringing about the newfound enthusiasm towards cloud gaming in the market. While the cloud was generally looked at as being the ‘future’ of gaming, Google was the company that showed everyone that the future was here, Google Stadia is set to offer a variety of games that will be available on instant access when the service launches in November. This includes prominent titles from industry developers such as EA, Square Enix, Bethesda and more. This service is, by far, the most fleshed out at this point in time, with a concrete release date in sight. Google has also released pricing details and is available in 14 different countries. Moreover, there is also cross-compatibility, with the ability to play on mobile using Google’s Pixel line of phones. NVIDIA GeForce Now Nvidia has considerable skin in the game, seeing as how Google and Microsoft both will use its GPUs for their cloud gaming efforts. However, they have also decided to throw their hat in the ring, with a currently-in-beta cloud gaming service called GeForce Now. GeForce Now works with low-powered laptops, Macs and even the Nintendo Shield tablet to allow for seamless play across devices. With a library of over 400 games, GeForce Now boasts a considerable lineup and also comes with a host of useful features, such as instant installation and ultra streaming mode for high-FPS gaming. The free beta is currently open. Other Notable Players Bethesda, a prominent game developer, recently announced its Orion software platform for increasing cloud gaming performance. This is said to be implemented across multiple platforms, with the company stating that the solution offers a bandwidth requirement cut of 40%, an encoding time improvement of 30% with 20% less compute. James Altman, the director of publishing for the company, stated, “This can be put into any game engine, it can be used with any streaming platform to provide a better experience for any consumer.” Nintendo has also begun making forays into the cloud gaming space, with efforts bearing fruit in Japan. Owing to its big footprint in the country and the sustainable Internet infrastructure, the Nintendo Switch is now able to stream much higher-requirement games to the console. This shows that Nintendo is working on a similar solution as well. Another hopeful in the space is cloud services leader Amazon, who is reportedly in the process of creating cloud games. This would signify that all 3 big cloud players are moving into gaming, signifying that is, indeed, the future.","excerpt":"As the future of gaming moves towards the cloud, long-time industry leaders and veterans are increasing focus towards their cloud operations. While there have been many ‘on-demand’ gaming services in the past, rising cloud infrastructure has increased accessibility to a whole new level. Moreover, cloud gaming generally tends to be in favour of those who […]","categories":["AI Features"],"tags":["Gaming","Google","Microsoft","NVIDIA","Sony"],"author_name":"Anirudh VK","publish_date":"2019-06-10T08:28:56","publication_year":"2019","word_count":805,"keywords":["Go","programming_languages:R","AI","Gaming","ML","programming_languages:Go","RAG","Ray","Google","Sony","NVIDIA","R","Microsoft"],"extracted_tech_keywords":["AI","ML","Ray","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/e3-2019-companies-into-cloud-gaming\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":51720,"title":"8 Cloud Outages That Shook The Tech World In 2019","content":"The popularity and usage of cloud computing has increased manifold in the last few years. From leading tech companies across the globe to startups, everyone wants to venture into this space. While it is the obligation of the cloud service provider to take responsibility for their infrastructure and ensure security and safety at all ends, sometimes it doesn’t quite happen. Here we list a few instances of cloud outages from the year 2019 when the customer systems and devices were jeopardised due to these outages. Read last year’s cloud outages that shook the tech world in 2018 Salesforce In May 2019, Salesforce faced one of its biggest service disruptions when the deployment of a database script to its Pardot Marketing Cloud ended up granting elevated permissions to regular users. The company had to block the users to prevent employees from stealing sensitive corporate data. It later had to block network access to other Salesforce services like Sales Cloud and Service Cloud to avoid further damage. As a result of this customers were unable to access Pardot Marketing Cloud for 20 hours and in addition, it took them 12 days to completely roll out other Salesforce services such as Sales Cloud and Service Cloud. The company later stated that a faulty database script had led to almost shut down its entire infrastructure and address the issue of broken user permissions.” Amazon AWS In August, an Amazon AWS US-EAST-1 datacenter in North Virginia experienced a power failure leading to the datacenter’s backup generators to start failing. It led to 7.5% of the EC2 instances and EBS volumes becoming unavailable. After the power was restored, Amazon determined that some EC2 instances and EBS volumes incurred hardware damage and the data stored on them were no longer recoverable. There was an extensive data loss for some customers proving that storing data in the cloud does not mean that it is completely safe and you do not also need a backup. It was one of the major instances of hardware failure in 2019 and proves that hosting data in the cloud is not always safe. Apple Cloud Many iCloud users across the globe briefly got the message of  “Service Unavailable – DNS failure” for several hours in July. This widespread cloud outage affected services such as App Store, Apple Music, and Apple TV, Apple Books, Apple ID, Apple Music, Apple Music Subscriptions and more.  While the issue has now been resolved, during the time of outage users could not use various functions such as Find My iPhone to locate their devices. The company stated that the cloud outage was a result of a ‘BGP route flap’ issue that caused severe packet loss for users in North America. Microsoft Even Microsoft faced its share of cloud outages this year affecting Azure, Microsoft 365, Dynamics, and DevOps. In May, Microsoft had to face an outage that lasted for more than an hour showing network connectivity errors in Microsoft Azure that deeply affected its cloud services including Office 365, Microsoft Teams, Xbox Live, and several others which are widely used by Microsoft’s commercial customers. Engineers identified the root cause to be an incorrect name server delegation issue that affected DNS resolution, network connectivity, and downstream impact. While the services were recovered, no customer DNS records were impacted during this incident. Google Cloud Servers The cloud servers in US-east1 region were cut off from the rest of the world as there was an issue with Cloud Networking and Load balancing. It caused physical damage to multiple concurrent fibre bundles that serve network paths in us-east1. Google carried out extensive mitigation work post that however, users faced increased latency. Google Cloud Platform Very recently Google Cloud Platform (GCP) was found experiencing major issues with services including cloud dataflow, cloud storage, compute engine, affecting multiple products globally. The company stated that its engineers are investigating the matter and will soon mitigate the incident. GCP stated that they have identified the cause and are currently rolling out mitigation. It seemed to have affected some Google Cloud APIs across us-east1, us-east4 and southamerica-east1, with some APIs impacted globally. It is also interesting to note that it came almost 20 days after users faced 100 per cent packet loss to and from ~20 per cent of instances in GCP’s us-west1-b region for two-and-a-half hours. The reason for the failure was its chubby lock system which resulted in the control plane losing and gaining leadership in short succession, company stated. Cloudflare In July 2019, Cloudflare visitors s received 502 errors caused by a massive spike in CPU utilization on the network. The company said that the 30-minute outage was due to a CPU spike which, in turn, was caused by a bad software deploy that was rolled back. It immediately took to fixing the issue. The company also clarified that this was not an attack and that the internal team are performing a full post-mortem to understand how this occurred and how we prevent this from ever occurring again. Once the issue was fixed, the company stated that everything was back to normal and blamed their own software for the mishap. Facebook and Instagram face outage There were issues with Facebook and Instagram earlier this year which was caused due to a server configuration change. During the outage, users faced issues with Facebook-owned properties Instagram and WhatsApp for around 14 hours. Quite recently also users complained that Facebook stopped working and users were not able to carry activities such as sharing a new post or accessing messenger.","excerpt":"The popularity and usage of cloud computing has increased manifold in the last few years. From leading tech companies across the globe to startups, everyone wants to venture into this space. While it is the obligation of the cloud service provider to take responsibility for their infrastructure and ensure security and safety at all ends, […]","categories":["AI Trends"],"tags":["Microsoft 365"],"author_name":"Srishti Deoras","publish_date":"2019-12-12T10:00:00","publication_year":"2019","word_count":922,"keywords":["Go","API","GCP","AWS","cloud computing","AI","R","RAG","Microsoft 365","DevOps","Azure"],"extracted_tech_keywords":["AI","RAG","cloud computing","AWS","Azure","GCP","R","Go","DevOps","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-cloud-outages-that-shook-the-tech-world-in-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10125465,"title":"The Next Wave of AI is Memory and Personalisation","content":"In a recent podcast, Microsoft AI CEO Mustafa Suleyman outlined his vision for the future of AI, particularly for Microsoft’s Copilot. He emphasised that “the next phase of AI development will focus on memory and personalisation”. Microsoft AI CEO Mustafa Suleyman says the next phase of AI is memory and personalization so your AI can remember everything about you and everything you have said in order to support you pic.twitter.com\/923qHdLxMJ— Tsarathustra (@tsarnick) June 24, 2024 Echoing the same, recently on the ‘Unconfuse Me’ podcast, Sam Altman and Bill Gates held a discussion on the next evolution of AI tools. Altman mentioned that “customisation and personalisation is very important for GPT-4, allowing users to tailor its style and assumptions to their needs”. Additionally, Sam mentioned that the system aims to incorporate personal data, including email, calendar, and appointment preferences, as well as connect to external data sources. Trust Woes of Personalisation Continues Suleyman’s post was bombarded with users’ questions about the nature of AI and memory, focusing on the precise meaning of AI’s memory. With personalisation came the worry of systems knowing too much about oneself. For instance, a user raised the concern of AI “owning memories,” with responses suggesting AI could become an extension of oneself, similar to smartphones. Thus, raising the question of trust and security. However, a balance was suggested as a crucial factor. “AI’s evolution towards memory and personalisation is inevitable. While it promises enhanced support, we must tread carefully to ensure ethical use and privacy. Balance is crucial for progress,” commented a user on X. Another post said, “I am willing to share my data with my own personal AI but not with companies like Microsoft or Google that want to use it to train their models. If you want personal data, you should compensate people for it.” Big Tech Giants are Already On It Suleyman envisions Microsoft Copilot as an AI assistant that can remember everything about its users, including personal data, context, and past interactions. The goal is to create a constant, supportive presence that acts as a personal sidekick throughout a user’s life, offering customised assistance based on comprehensive knowledge of the individual. Similarly, others are already working on it. The recent Apple Intelligence, unveiled at Apple WWDC 24, is said to be integrated into iOS 18, iPadOS 18, and macOS Sequoia, marking a significant step forward in personalising user experiences across Apple devices. Digital speaker Mark Van Rijmenam said on X, “Apple’s approach, distinct from Google’s and Amazon’s cloud-centric strategies, emphasises personal device empowerment. This strategy resonates with the principle of thinking exponentially, demonstrating a visionary leap in making cutting-edge AI technology a personal, everyday experience.” Apple’s AI is a step towards building a personalised experience for their users, making their phones more “personal, capable, and intelligent” than ever. Apple is also enhancing Siri’s personalisation by integrating GPT-4, as announced at the 2024 Worldwide Developer Conference, which will enable Siri to access ChatGPT on demand. Since personalisation is already taking place, a prime example in India is HaiVe, a Chennai-based startup, which recently launched a personal AI home studio. This platform aggregates and analyses personal data, such as medical records and financial information, to create a comprehensive digital twin of an individual’s life. Prioritising user privacy, the AI system can be installed on a personal computer, which then acts as a central hub accessible remotely via smartphones. Future enhancements aim to include AI agents capable of making phone calls to manage tasks, making the AI experience more personalised and proactive. Personalised AI has the potential to revolutionise our daily lives, offering tailored recommendations such as these, managing schedules, and even predicting our needs.","excerpt":"“AI can remember everything about you and everything you’ve said, enhancing your memory and personalisation,” says Microsoft chief Mustafa Suleyman.","categories":["AI Trends"],"tags":["memory","Mustafa Suleyman","personalisation","Siri"],"author_name":"Gopika Raj","publish_date":"2024-07-02T14:33:32","publication_year":"2024","word_count":611,"keywords":["Go","ChatGPT","Mustafa Suleyman","AI","Git","GPT","Aim","memory","edge AI","Siri","Rust","ViT","personalisation","R"],"extracted_tech_keywords":["AI","ChatGPT","Aim","edge AI","R","Go","Rust","Git","GPT","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-next-wave-of-ai-is-memory-and-personalisation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":6299,"title":"Sapience raises Rs 45-cr Series B funding from Orios Venture Partners","content":"Sapience Analytics today announced that it has closed a Series B investment of Rs 45 crore raised from Orios Venture Partners. The company has previously received funding from Indian Angel Network and Seed Enterprises. Sapience, which provides employee productivity improvement solutions to enterprises, intends to use the funding to further accelerate its customer base, by augmenting sales and marketing in the Americas, reaching out to new segments in the professional and mobile workforce categories, and expanding product capabilities for higher value addition to Sapience customers. “When it is so rare to see software products coming out of India, Sapience is a great example of the new breed of innovative and successful Indian product firms. With an impressive growth to 100,000+ users across seven countries and over 60 customers, including five of India’s top ten IT companies, four Fortune Global 200 and several billion dollar enterprises, Sapience is undoubtedly the leader in workforce effort analytics space, especially in India. Our investment is aimed at expanding their market reach, and ensuring that Sapience Analytics rapidly achieves global prominence,” said Rehan Yar Khan, Managing Partner, Orios Venture Partners. As a patent-pending product, Sapience backed with insights obtained from analysis of more than 100 million work hours across diverse companies. Sapience helps individual employees to adopt improved work habits that lead to Work-Life Harmony, claims the company. Sapience customers span a broad range of verticals including IT Services, Global ISVs, BPOs\/KPOs, Engineering Design, Financial Services, and for Outsourcing governance. “Globally, more and more knowledge workers are becoming increasingly conscious about the need for mindfulness at work. Sapience has pioneered a unique concept of Sapience Work Yoga which helps employees achieve much more at work with reduced stress. Our customers are realising over 20% gain in productivity and the ability to sustain optimal delivery outcome through improved quality of effort, and happier employees. The series B funding validates our product vision, execution and business strategy so far, and provides the impetus to raise it to the next level,” said Shirish Deodhar, Co-founder and CEO of Sapience Analytics.","excerpt":"Sapience Analytics today announced that it has closed a Series B investment of Rs 45 crore raised from Orios Venture Partners. The company has previously received funding from Indian Angel Network and Seed Enterprises. Sapience, which provides employee productivity improvement solutions to enterprises, intends to use the funding to further accelerate its customer base, by […]","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2014-10-14T15:25:23","publication_year":"2014","word_count":342,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","Aim","ViT","analytics","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","API","ViT","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sapience-raises-rs-45-cr-series-b-funding-orios-venture-partners\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":18943,"title":"Mu Sigma Ex-CEO Ambiga Subramanian To Launch New Social Networking App That Revolves Around Anonymity","content":"Mu Sigma’s former CEO Ambiga Subramanian and Goutham Ekollu, ex-head of operations at the company, have co-founded hyphen.social, new a social-networking application. “Our primary focus is to increase social engagement within the social network. The way that’s done on this platform will be by finding the right balance between anonymity and community-based regulations,” Ekollu said. Subramanian, Ekollu and his wife Krishnaveni Kumbaji have been working on hyphen.social for around one year, but registered the company under the name Vittr only in January 2017. hyphen.social intends to increase social engagement among users by offering their key tools — anonymity and gamification. “During a discussion between Ambiga and I, we discussed the need for authentic conversations amongst people and how anonymity enables that in multiple environments, whether it be at work, at college or any social connection,” Ekollu added. Subramanian and Ekollu said that young peoples’ behaviour on major social media websites such as Facebook had changed as the older generation — their parents and families — had joined the bandwagon. hyphen.social will therefore target a diverse age group, from teenagers to 40-something-year-olds — but will keep their identities anonymous. Reportedly the social network will begin by targeting college students and see how the traction picks up. “Anonymity is a double-edged sword. It can enable authentic conversations, but can also reduce accountability in people. There is a lot of research available on understanding group and individual behavior in anonymous settings. We have tapped into some of them to come up with gamification within the app around anonymity… We want to engage the younger age group in the app, as early adopters. The app has a lot of new features, which will require some getting used to. The young folks in our test group enjoyed exploring these features on their own,” Subramanian said. Last year, Subramanian agreed to sell her stake in Mu Sigma, giving her former husband Dhiraj Rajaram a majority ownership in the data analytics company.","excerpt":"Mu Sigma’s former CEO Ambiga Subramanian and Goutham Ekollu, ex-head of operations at the company, have co-founded hyphen.social, new a social-networking application. “Our primary focus is to increase social engagement within the social network. The way that’s done on this platform will be by finding the right balance between anonymity and community-based regulations,” Ekollu said. […]","categories":["AI News"],"tags":["mu sigma"],"author_name":"Prajakta Hebbar","publish_date":"2017-11-10T10:48:04","publication_year":"2017","word_count":325,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","ViT","analytics","R","mu sigma"],"extracted_tech_keywords":["AI","analytics","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mu-sigma-ex-ceo-ambiga-subramanian-launch-new-social-networking-app-revolves-around-anonymity\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69737,"title":"Diving into the future of IoT","content":"Looking at these predictions, it is not be wrong to say that the IoT space is going to boom in the coming future. However, looking at these estimations leaves us with a few questions in mind. How is ioT going to benefit us? What does IoT has in store? Is the future of IoT going to have a big impact on how we live? If you want to get an answer to any of these questions, go ahead and read the post. What Is IoT? Let us first understand what IoT is. In simple terms, IoT is a technology that helps to simplify life by allowing everyday objects with the ability to send and receive data using internet and a sensor. Such devices have now become very common in everyday lives and people are getting used to them. An example of the same will be a smart watch that tracks your heart rate, distance walked, etc. The walk from Industrial to Consumer Most of us relate IoT devices with consumer products. However, this is not always the case. The concept of IoT began initially and more importantly for enterprise solutions, which is still a very big but comparatively a lesser known part of IoT. IIC or Industrial Internet Consortium was created by AT&T, GE, Cisco, IMB and Intel that has been pushing new ideas in the enterprise area since two years. IIC aims to coordinate the structure and foundation of the industrial internet as a non-profit consortium. IIC has helped in the materialization of a wide range of connected sensors that collect data in an industrial setting since its inception. However, we are still not clear as to how this affects enterprise mobility? Let’s discuss that now… Enterprise Mobility Most companies were earlier using enterprise mobility in the context of messaging and collaboration earlier. However, with time the capabilities have developed and we are watching an enhanced demand for software that supports apps. There is an incursion of opportunities in enterprise app development and it has a scope of benefitting for the developers. Well positioned app developers can unlock valuable sensor and asset data to bring innovative products that can change the way things are handled for better. Better association means enhanced growth It can be interpreted that IoT will allow businesses to leverage mobile workforce by staying connected to key systems and assets throughout the world. Enterprises get the advantage of better connectivity via sensors, internet and smart machinery by reducing downtime and gaining more insight into the working. This will help enhance the services and products and it automatically implies better growth opportunities. What IoT has in store? Rapid Evolution: It is expected that in the coming years, IoT space will evolve at a rapid pace. Several experts believe that IoT products will become more small, cheap and fast. It will not be wrong if we say that in future, we can even ingest the IoT products and even ember them in our bodies. Really? Yes… The Apple’s watch is expected to be a completely different device over the coming years. This means that IoT app developers need to keep themselves abreast with the latest in the field but also foresee future and adapt to changes in advance. Our life will be affected: With the improvement in obtrusive wearables, where the technological component is completely disguised, our life will be affected in every which way. “Biometric sensors will be included in 40% of smart phones shipped to the end user by next year.” says Gartner. This means that there are still ample opportunities in this segment.","excerpt":"As per Vision Mobile “There will be 34 billion devices connected to the internet by 2020” and Business Insider reports “By 2020, there will be close to 10 million Internet of Things developers”","categories":["IT Services"],"tags":[],"author_name":"AIM Media House","publish_date":"2016-10-08T09:54:51","publication_year":"2016","word_count":598,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","RAG","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/diving-future-iot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10005831,"title":"Why Is CRISP-DM Gaining Grounds","content":"CRISP-DM is a popular methodology that follows a standard, end-to-end structured approach to solving a problem that requires data science. More precisely, CRISP-DM or CRoss-Industry Standard Process for Data Mining focuses on the data mining part of the operation. Industries and organisations have been undergoing machine learning-driven approaches for a few years now. However, this report from last year suggests that 85% of AI projects won’t deliver for their sponsors due to reasons like low quality, lack of development process, less functional in real-world applications, among others. Some of its popular instances are after spending 62 Million, IBM Watson AI Health was cancelled in 2019 due to wrong recommendations on cancer treatments, in 2018 Uber’s self-driving car killed a woman in Arizona, and more. Due to these issues, organisations have started using alternative methodologies in their machine learning applications. This is where CRISP-DM comes into play. The utilisation of this methodology has been witnessing exponential growth for a few years now. How It Works CRISP-DM defines a framework for denoting data mining projects and sets out activities to be performed to complete a product or service. The activities consist of six phases, which are Business Understanding, Data Understanding, Data Preparation, Modeling, Evaluation and Deployment. Source Image: here The successful completion of a phase initiates the execution of the subsequent activity. Also, the methodology includes iterations of revisiting previous steps until success or completion criteria are met. Why Use CRISP-DM Data science demands a top-down, solution-oriented approach to solve problems. According to the latest Data Science recruitment survey,  the open jobs figure for Data Science and Analytics reached a maximum in February-March 2020, reaching a high of approximately 113,000 in the first week of March and rising steadily from a figure of 97,000 last year. In this field, data plays a very important role and processes like data mining help in generating actionable insights, extract patterns and identify relationships from large datasets. CRISP-DM is designed to be domain-agnostic and has been widely used by industry and research communities. The distinctive characteristics have made CRISP-DM to be considered as ‘de-facto’ standard of data mining methodology and as a reference framework to which other methodologies are benchmarked. One of the important factors of using this method in Data Science is that it is a cross-industry standard for which it can be implemented in any Data Science project regardless of its domains. This methodology remains a dependable method to develop data science solutions for enterprise problems. Also, the flexible and iterative approach of the method makes it a future-proof alternative for anyone looking to solve data science problems. Benefits of Using CRISP-DM This method is cost-effective as it includes a number of processes to take out simple data mining tasks.CRISP-DM encourages best practices and allows projects to replicate.This methodology provides a uniform framework for planning and managing a project.Being cross-industry standard, CRISP-DM can be implemented in any Data Science project irrespective of its domain. Wrapping Up CRISP-DM is becoming the de-facto industry standard process model for data mining, with an expanding number of applications, such as in quality diagnostics, warranty, and others. However, recently, a team of AI researchers from Max Planck Institute for Information and others claimed that they identified two shortcomings of CRISP-DM. First, CRISP-DM does not cover the application scenario where an ML model is maintained as an application. Second, CRISP-DM lacks guidance on quality assurance methodology. To mitigate such issues, researchers further proposed CRISP-ML(Q) or CRoss Industry Standard Process for the development of Machine Learning applications with Quality assurance methodology. CRISP-ML(Q) is a process model for machine learning applications with a quality assurance methodology, that helps organisations to increase efficiency and success rate in their machine learning projects. The CRISP-ML(Q) methodology is also organised in six phases and expands CRISP-DM with an additional maintenance phase. It guides machine learning practitioners through the entire machine learning development life-cycle, providing quality-oriented methods for every phase and task in the iterative process including maintenance and monitoring.","excerpt":"CRISP-DM is a popular methodology that follows a standard, end-to-end structured approach to solving a problem that requires data science. More precisely, CRISP-DM or CRoss-Industry Standard Process for Data Mining focuses on the data mining part of the operation. Industries and organisations have been undergoing machine learning-driven approaches for a few years now. However, this […]","categories":["Deep Tech"],"tags":["Data Mining","data mining algorithms","Machine Learning","Machine Learning Algorithms"],"author_name":"Ambika Choudhury","publish_date":"2020-08-31T10:00:26","publication_year":"2020","word_count":663,"keywords":["Machine Learning Algorithms","data science","Go","machine learning","AI","Data Mining","ML","Machine Learning","RAG","Aim","analytics","GAN","R","data mining algorithms"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","RAG","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/why-is-crisp-dm-gaining-grounds\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10075896,"title":"Why Speech Separation is Such a Difficult Problem to Solve","content":"You are talking on the phone, or recording an audio, or just speaking to voice assistants like Google Assistant, Cortana, or Alexa. But the person on the other side of the call cannot hear you because you are in a crowded place, the recorded audio has a lot of background noise, or the “Hey, Alexa” call wasn’t picked up by your device because someone else started speaking. All of these problems related to separating voices, informally referred to as the “cocktail party problem”, have been addressed using artificial intelligence and deep learning methods in recent years. But still, separating and inferring multiple simultaneous voices is a difficult problem to completely solve. Why is that? To start, speech separation is extracting speech of the “wanted speaker” or “speaker of interest” from the overlapping mixture of speech from other speakers, also referred to as ‘noise’. In recent years, the advancements of automatic speech recognition (ASR) technology is making speech separation an important field of research. Challenges and approaches The origin of speech recognition dates back to 1952, when Bell Laboratories  researchers Stephen Balashek, R. Biddulph, and K. H. Davis released the first voice recognition device, called “Audrey”, that could recognise digits from a single voice. By the 1980s, plenty of progress had been made in this field with the introduction of the n-gram language model, which is a probabilistic language model that can predict the next item in a text or speech sequence using the Markov model’s (n-1)-order. When these techniques were implemented by the researchers for multiple sound sources, there were two major challenges for accurate separation: Non-stationarity of the speech signals and the surrounding environment.Reverberation in the acoustic setting. There are three most common approaches to tackle these challenges: Beamforming: This technique uses spatial information with linearly arranged microphones to assess the direction of each speaker towards the microphone array. Blind source separation: BSS is based on independent component analysis (ICA) and therefore depends on statistical independence of signals. Single channel speech separation: SCSS is a highly complicated technique that aims to separate and deconvolve independent and individual sources from a single-channel mixture. Speech separation is essentially an advanced case of sound source separation. Humans have an “innate” ability to separate sound sources since childhood. However, though it looks intrinsic in humans, it is actually a case of conditioning—or, training—since birth to be able to separate the desired speech from the background noise. This is why humans can focus on a single sound source by merely turning towards it, even in the presence of background noise. Applications and results Recently, Meta AI researcher Eliya Nachmani with his team introduced SepIt, which can separate speech from 2,3,5, and 10 sources or speakers. It is a deep neural network that takes the approach of SCSS with a general upper bound (stopping criterion) which is obtained with Cramer-Rao bound that makes an assumption about the nature of segments of speech. A similar approach was adopted by Yi Luo in 2018 by developing Conv-TasNet, which is a fully convolutional, end-to-end time-domain sound separation network with an encoder and decoder. MIT-IBM Watson AI Lab researcher Chuang Gan with his team developed an AI tool that essentially matches sound and visuals of the same source to separate similar sounds. In the project, the researchers used synchronised audio-video tracks of musicians playing piano to recreate how humans use multiple sensors to infer information. “Multi-sensory processing is the precursor to embodied intelligence and AI systems that can perform more complicated tasks,” said MIT professor Antonio Torralba. DeLiang Wang, in his paper about Supervised Speech Separation, talks about how the idea of training a model similar to how text-to-image models are trained with CNNs and combination of text and descriptive text, could greatly improve the case for speech separation as well. This argument makes a case to solve the cocktail party problem by comparing ASR scores and separation capabilities of each model against human speech intelligibility in the same conditions. Wang proposes that instead of the traditional way of studying speech separation as a signal processing problem, it should be treated as a supervised learning problem. This way—if the model is trained using discriminative features and patterns of speakers and their speech—it might be possible to infer and remove noise from a recording. A paper by P. Nancy of Parisutham Institute of Technology also pointed out the need of training models on more acoustic conditions. Why is the problem difficult to solve? Meta AI’s ‘SepIt’ ranks highest in the leaderboard of progress in Speech Separation but still indicates improvements and further research in the field due to the size of the model. Though researchers are making great progress in the field of speech separation and recognition using various methods, the solution and the biggest challenge still is inferring sounds as separate sources of speech instead of a single speaker. With training the algorithms and models of speech separation, progress could be made with the use of multiple sensory inferring—similar to humans as we saw in the above mentioned approaches. This can further improve efficiency and accuracy of speech recognition and separation-related innovations like voice assistants or even hearing aids. Every source of sound has a different frequency, volume, and waveform through which each source can be identified and separated. But, this is far easier to say than to work on without sacrificing accuracy. Separating two speeches is far more challenging than understanding the speech of one speaker since the possible combinations are almost infinite.","excerpt":"Researchers are making great progress in the field of speech separation and recognition using various methods, but the solution and the biggest challenge still is inferring sounds as separate sources of speech instead of a single speaker.","categories":["AI Features"],"tags":["automatic speech recognition","ML","Speech Recognition"],"author_name":"Mohit Pandey","publish_date":"2022-09-27T12:00:00","publication_year":"2022","word_count":915,"keywords":["Go","Meta AI","artificial intelligence","AI","neural network","ML","Git","automatic speech recognition","Ray","Aim","deep learning","Speech Recognition","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","Meta AI","Aim","Ray","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-speech-separation-is-such-a-difficult-problem-to-solve\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":57200,"title":"Punjab Jails To Get AI-Enabled CCTV Systems, Live Wire Fencing","content":"Punjab Chief Minister Amarinder Singh has approved a series of measures to make jails more secure. This includes an AI-enabled CCTV system in 18 jails, covering nine central jails, seven district jails and two special jails. Singh also authorized a live wire fencing along the outer walls, a separate prison intelligence wing and a legislation to establish the ‘Punjab Prison Development Board’ in the ongoing Budget session. Chairing a high-level meeting to review the functioning of the Prisons Department, Singh  directed Jails Minister Sukhjinder Singh Randhawa to ensure that best practices are followed. He also asked the department to submit a comprehensive plan for restructuring the jails within four weeks. As stated in an official release, the Chief Minister also asked ADGP (Jails) Praveen Kumar Sinha to prepare a plan for restructuring, realigning and launching of new schemes encompassing correctional measures for reformation of prisoners. With the objective of cutting expenses involved in presenting undertrials in courts – an amount which can go up to Rs 50 lakh a day – Singh also approved the department’s proposal to install video conferencing system in all jails. He also approved the construction of a double-barbed wire inner boundary wall and three additional watchtowers at the Central Jail Ludhiana to beef up the security.","excerpt":"Punjab Chief Minister Amarinder Singh has approved a series of measures to make jails more secure. This includes an AI-enabled CCTV system in 18 jails, covering nine central jails, seven district jails and two special jails.  Singh also authorized a live wire fencing along the outer walls, a separate prison intelligence wing and a legislation […]","categories":["AI News"],"tags":[],"author_name":"Anu Thomas","publish_date":"2020-02-21T17:54:10","publication_year":"2020","word_count":211,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","R"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/headline-punjab-jails-to-get-ai-enabled-cctv-systems-live-wire-fencing\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":24948,"title":"Top 5 Deep Learning Research Papers You Must Read In 2018","content":"Any newfound theory in science is insignificant without being put to practical use. The same can be said about deep learning (DL). Now, with new research and development, the vast pedagogy in this field has found practical applications in various disciplines across new tech businesses. Research work in DL has taken an innovative stance. Rather than using it to better AI and ML technologies, DL research is seeing new ideas being explored in critical areas such as healthcare and banking. We have listed down the top research papers on DL which are worth reading and have an interesting take on the subject. These papers were published in the recently concluded International Conference on Learning Representations in Vancouver, Canada, in May 2018. 1. Spherical CNNs Researchers at the University of Amsterdam have developed a variation of convolution neural networks (CNN) known as Spherical CNNs. These CNNs work with images which are spherical in shape (3D). For example, images from drones and autonomous cars generally cover many directions and are three-dimensional. Regular CNNs are applicable only to two-dimensional images, and imposing 3D features from images mentioned in this example may literally fail in a DL model. This is where Spherical CNNs were envisioned. In the paper, the researchers conceptualise spherical features with the help of the Fourier Theorem, as well as an algorithm called Fast Fourier Transform. Once developed, they test the CNNs with a 3D model and check for accuracy and effectiveness. The concept of Spherical CNNs is still at a nascent stage. With this study, it will definitely propel the way CNNs are perceived and used. You can read the paper here. 2. Can Recurrent Neural Networks Warp Time? Not just ML and AI researchers, even sci-fi enthusiasts can quench their curiosity about time travel, if they possess a strong grasp of concepts like neural networks. In a research paper published by Corentin Tallec, researcher at University of Paris-Sud, and Yann Ollivier, researcher at Facebook AI, they explore the possibility of time warping through recurrent neural networks such as Gated Recurrent Units (GRUs) and Long Short Term Memory (LSTM) networks. The self-learning capabilities present in these models are analysed. The authors have come up with a new concept called ‘Chrono Initialisation’ that derives information from gate biases of LSTM and GRUs. This interesting paper can be read here. 3. Learning How To Explain Neural Networks: PatternNet And PatternAttribution We are yet to fully understand why neural networks work exactly in a particular way. Complex ML systems have intricate details which sometimes astonish researchers. Even though there are systems which decode neural networks, it is difficult at times to establish relationships in DL models. In this paper, scholars at Technical University in association with researchers at Google Brain, present two techniques called PatternNet and PatternAttribution which explain linear models. The paper discusses a host of previously established factors such as signal estimators, gradients and saliency maps among others. You can read the paper here. 4. Lifelong Learning With Dynamically Expandable Networks Lifelong learning was a concept first conceived by Sebastian Thrun in his book Learning to Learn. He offered a different perspective of the conventional ML. Instead of ML algorithms learning one single task, he emphasises on machines taking a lifelong approach wherein they learn a variety of tasks over time. Based on this, researchers from KAIST and Ulsan National Institute of Science and Technology developed a novel deep network architecture called Dynamically Expandable Network (DEN) which can dynamically adjust its network capacity for a series of tasks along with requisite knowledge-sharing between them. DEN has been tested on public datasets such as MNIST, CIFAR-100 and AWA for accuracy and efficiency. It was evaluated for factors including selective retraining, network expansion and network timestamping (split\/duplication). This novel technique can be read here. 5. Wasserstein Auto-Encoders Autoencoders are neural networks which are used for dimensionality reduction and are popularly used for generative learning models. One particular type of autoencoder which has found most applications in image and text recognition space is variational autoencoder (VAE). Now, scholars from Max Planck Institute for Intelligent Systems, Germany, in collaboration with scientists from Google Brain have come up with the Wasserstein Autoencoder (WAE) which utilises Wasserstein distance in any generative model. In the study, the aim was to reduce optimal transport cost function in the model distribution all along the formulation of this autoencoder. After testing, WAE proved to be more stable than other autoencoders such as VAE with lesser architectural complexity. This is a great improvement in autoencoder architecture. Readers can go through the paper here. Endnote All of these papers present a unique perspective in the advancements in deep learning. The novel methods also provide a diverse avenue for DL research. Machine learning and artificial intelligence enthusiasts can gain a lot from them when it comes to latest techniques developed in research.","excerpt":"Any newfound theory in science is insignificant without being put to practical use. The same can be said about deep learning (DL). Now, with new research and development, the vast pedagogy in this field has found practical applications in various disciplines across new tech businesses. Research work in DL has taken an innovative stance. Rather […]","categories":["AI Trends"],"tags":["deep learning research","fourier transform machine learning"],"author_name":"Abhishek Sharma","publish_date":"2018-05-29T11:22:26","publication_year":"2018","word_count":808,"keywords":["Go","machine learning","artificial intelligence","AI","fourier transform machine learning","neural network","ML","deep learning research","Aim","deep learning","VAE","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","Aim","R","Go","VAE"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-deep-learning-research-papers-you-must-read-in-2018\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10161155,"title":"HubSpot Co-Founder Launches Agentic AI Platform Agent.ai, Records 258K Users in 4 Months","content":"HubSpot co-founder Dharmesh Shah has ventured into the AI space with Agent.ai, an agentic platform designed to enable users to create and collaborate on custom AI agents. First unveiled at the INBOUND 2024 conference, the project has already amassed an impressive 258,000 users in just four months, significantly surpassing its initial goals. “At time of launch, I was hoping to get on stage and say I had thousands of users in beta already. Turns out, I was right. We had 47,000 users in beta when launched,” Shah shared in a recent update on LinkedIn. The platform allows creators to experiment with low-code tools to build agents tailored to specific use cases, from analysing personal health data to optimising workflows. It also serves as a professional network and marketplace, with over 3,400 agents built so far, including contributions from HubSpot CEO Yamini Rangan. “We now have 14,422 builders approved to create agents , and 1,697 of them have done so,” Shah noted, emphasising the platform’s accessibility and growing ecosystem. The rapid adoption of Agent.ai highlights its potential to democratise AI innovation, catering to everyone from professionals to middle school students. With Shah’s hands-on approach, the platform is continuously evolving, aiming to bridge the gap between AI tools and real-world applications. “This is the most fun I’ve had building since ChatSpot (now Breeze Copilot),” Shah added, reflecting his excitement for this transformative project. Shah is currently building an agent on Agent.ai to process data from his Oura Ring (using its API). While unsure of the exact use case, he is considering a chat agent to query data insights, such as sleep patterns, average bedtimes, or correlations like readiness scores with social media activity and engagement. “Imagine one day being able to take some of your health\/sleep data and intersecting with your meeting\/transcription data. Does your town in meetings change when you haven’t had enough sleep?” he asked on a social media post. A few days ago Shah remarked that in the future, it’s going to be all about agents. “Even with the large language models we have *today* I think it’s possible to create useful AI agents that can augment a team and take on some subset of tasks that need to be done to accomplish a goal,” he said. He expects the underlying tech to get better in this year which will give the opportunity to make more use cases with AI agents. Shah confirmed that right now, it’s running independent of HubSpot, but his hope is that it will eventually get rolled into HubSpot. Indian Founders ♥️ AI Agents The craze for agentic AI continues with big tech and startups alike working on building customised agents for their customers. The AI agents market has been witnessing promising growth. From $5.1 billion in 2024, the market is expected to hit $47.1 billion by 2030. In particular, Indian entrepreneurs have been significantly driving the growth. Indian founders in the AI startup space have been increasingly driving this growth. A few founders have remarked that advancements in foundational models have made it easier and cheaper to build AI agents. Recently, Microsoft chief Satya Nadella showcased Copilot Studio, a conversational AI platform that helps users build AI agents, at Microsoft AI Tour in Namma Bengaluru. “Building agents should be as simple as creating a spreadsheet,” said Nadella. Similarly, at the AI Pitchfield event, Salesforce India head, Arundhati Bhattacharya, expressed her enthusiasm on the Indian Agentic startup front. “India’s startup ecosystem, being the third largest in the world, is uniquely positioned to address challenges at a billion-plus scale. The progress we’re seeing with agentic AI is phenomenal, even though companies are just starting to dip their toes into it,” she said.","excerpt":"There are 47,000 users in beta and over 3400 agents have been built so far.","categories":["AI News"],"tags":["agent ai","Dharmesh Shah","Hubspot"],"author_name":"Vandana Nair","publish_date":"2025-01-10T20:48:37","publication_year":"2025","word_count":616,"keywords":["Go","API","agentic AI","agent ai","AI","Hubspot","RPA","RAG","Aim","Dharmesh Shah","ViT","GAN","R"],"extracted_tech_keywords":["AI","agentic AI","Aim","RAG","R","Go","API","GAN","ViT","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hubspot-co-founder-launches-agentic-ai-platform-agent-ai-records-258k-users-in-just-4-months\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10089165,"title":"Mohit Joshi Ditches Infosys to Become Tech Mahindra&#8217;s Top Gun","content":"Mohit Joshi, President of Infosys, has resigned to take up the position of Managing Director and Chief Executive Officer at Tech Mahindra. Joshi, who has been with Infosys since 2000, will succeed CP Gurnani as MD & CEO at Tech Mahindra when Gurnani retires on December 19, 2023. According to the release, Joshi will join Tech Mahindra well in advance to allow for a smooth transition. This marks Infosys’ second leadership exit following a series of departures in the Indian IT industry. Ravi Kumar S., the former President of Infosys Global Services Organisation, also left the company to become the global CEO of Cognizant. Sukamal Banerjee, HCLTech’s Engineering and R&D head, was the most recent in the industry to step down. At Infosys, Joshi was responsible for the company’s global financial services and healthcare business, including the banking platform ‘Finacle’ and the AI\/automation portfolio. He also led sales operations and transformation, executive responsibility for all large deals, and was in charge of the company’s internal CIO function and the Infosys Knowledge Institute. Before joining Infosys, Joshi worked with ABN AMRO and ANZ Grindlays in their Corporate and Investment bank. Mohit has been a Non-Executive Director at Aviva Plc since 2020 and is a member of its Risk & Governance and Nomination committees.","excerpt":"Mohit Joshi, President of Infosys, has resigned to take up the position of Managing Director and Chief Executive Officer at Tech Mahindra.","categories":["AI News"],"tags":["Indian IT","Infosys","Tech Mahindra"],"author_name":"Lokesh Choudhary","publish_date":"2023-03-11T12:14:13","publication_year":"2023","word_count":212,"keywords":["Go","Tech Mahindra","Infosys","AI","programming_languages:R","programming_languages:Go","automation","GAN","Indian IT","R"],"extracted_tech_keywords":["AI","R","Go","GAN","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mohit-joshi-resigns-from-infosys-to-become-md-ceo-at-tech-mahindra\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061901,"title":"A hands-on guide to implementing ggplot in python using plotnine","content":"Visualization of the data plays a crucial role in the majority of data analytics tasks. The ggplot package of the R programming language makes the R richer on the side of data visualization. In python as well, various packages are also available for data visualization. If the features and capabilities of ggplot can be used in python, it will be a valuable advantage in many visualization specific tasks. In this article, we are going to explore how we can use ggplot in python for visualizing data using a package named plotnine that is based on ggplot2. The major points to be discussed in the article are listed below. Table of content What is ggplot?First plotFactorization of datapointsGgplot with pandas data frame Let’s begin with having a brief introduction to ggplot. What is ggplot? We mainly know ggplot as a package used in R for data visualizations. This package is responsible for making R one of the best tools in the world of data visualization.  This package is created by Hadley Wickham and can be considered as an implementation of the grammar of graphics suggested by Leland Wilkinson. Grammar of graphics can be considered as a scheme that stands for breaking the graphs into semantic components. Examples of components of graphs can be scales and layers. Ggplot becomes more useful because of its ability to create stylish and clear graphs. From a normal user to a high end-user of the R language, they use this package for visualization. In this article, we are going to use the plotnine package for the implementation of the ggplot package in python. There are several python packages like matplotlib, plotly, ggpy, etc. in python for visualization, but ggplot’s capabilities also need to be explored. To explore this, we will use the plotnine package that covers all the ggplot features and extend them to python. We can install plotnine using the following lines of codes: pip install 'plotnine[all]' After installation, we are ready to use ggplot for visualizing data in python. First plot In this section, we will get to know about how we can use the practice datasets of plotline packages that are available in the form of pandas tabular data. We can say that each dataset available in this package is in the form of a pandas data frame. We can call them using the plotnine subpackage plotnine.data. We can find the list of practice datasets of plotnine here. Let’s import the mtcars dataset. from plotnine import * from plotnine.data import mtcars mtcars Output: Here we can see values in our data frame. Let’s use ggplot for making a plot choosing any two variables. plot = (ggplot(mtcars, aes('disp', 'mpg')) + geom_point()) plot Output: Here we can see we have plotted mpg against displacement of the cars using the ggplot inside the plotnine. Let’s move to the deep side of ggplot. Factorization of datapoints In the dataset, we have seen that we have many categorical values like we can categorize our data based on the number of cylinders used by the engine. The above plot can also be factored using colours according to the number of cylinders in the following way. plot = (ggplot(mtcars, aes('disp', 'mpg')) + geom_point()) plot Output: We can also factor graphs instead of just factorization using points using the following lines of code. plot =(ggplot(mtcars, aes('disp', 'mpg', color='factor(cyl)')) + geom_point() + facet_wrap('~cyl')) plot output: Let’s make it more stylish. plot =(ggplot(mtcars, aes('disp', 'mpg', color='factor(cyl)')) + geom_point() + facet_wrap('~cyl') + theme_xkcd()) plot Output: Here we have used the theme of ggplot to make the visualization of data more attractive. Ggplot with pandas data frame In this section we will work with pandas data frame for making plots using ggplot, for this, we are using the titanic dataset that can be found here. Let’s import the dataset. import pandas as pd data = pd.read_csv('https:\/\/web.stanford.edu\/class\/archive\/cs\/cs109\/cs109.1166\/stuff\/'+'titanic.csv') data.head() Output: Let’s draw a plot that can tell us how many people from titanic data survived according to their passenger class. data['Survived'] = data['Survived'].astype('category') ggplot(aes(x=\"Pclass\", fill=\"Survived\"), data) + geom_bar(stat = 'count') + theme_xkcd() Output: Here we can see a bar chart we made using ggplot for titanic data. Let’s make a plot that can tell us the average age of people who survived and did not survive. ggplot(data, aes(x='Survived', y='Age')) + \\ geom_violin() Output: Here we can see that in data we have people who survived are mostly of age 20. Let’s take a look at the graph categorization based on sex. ggplot(data, aes(x='Survived', y='Age')) + \\ geom_boxplot() + \\ facet_wrap(['Sex']) Output: Here we have segregated the people based on their sex and survival status using a box plot that also represents a range according to age. Let’s segregate the plot more. ggplot(data, aes(x='Survived', y='Age')) + \\ geom_boxplot() + \\ facet_wrap(['Sex','Pclass']) + theme_xkcd() Output: Here we can see a segregated box plot according to sex and class. Final words In the article, we have seen how we can use the data visualization features of ggplot in python. For this purpose, we have used plotnine as our base package for ggplot. Using ggplot we have made our visualization procedure more attractive and easy. References Plotnine documentation Ggplot documentation Link for the codes","excerpt":"The ggplot package of the R programming language makes the R richer on the side of data visualization. In python as well, various packages are also available for data visualization. If the features and capabilities of ggplot can be used in python, it will be a valuable advantage in many visualization specific tasks.","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning","Python"],"author_name":"Yugesh Verma","publish_date":"2022-03-03T10:00:00","publication_year":"2022","word_count":863,"keywords":["Go","Plotly","TPU","AI","Machine Learning","RAG","Python","Matplotlib","analytics","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)","Pandas"],"extracted_tech_keywords":["AI","analytics","Pandas","Matplotlib","Plotly","RAG","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-hands-on-guide-to-implementing-ggplot-in-python-using-plotnine\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10110393,"title":"ChatGPT Now Has a Team Plan","content":"OpenAI has expanded its ChatGPT offerings with the introduction of ChatGPT Team. The newly introduced ChatGPT Team provides access to advanced models like GPT-4 and DALL·E 3, along with tools like Advanced Data Analysis. ChatGPT Team features a dedicated collaborative workspace for teams and admin tools for efficient team management. Similar to ChatGPT Enterprise, businesses using ChatGPT Team retain ownership and control over their data, with no training on business-specific data or conversations. Key features of ChatGPT Team include access to GPT-4 with a 32K context window, tools like DALL·E 3, GPT-4 with Vision, Browsing, and Advanced Data Analysis, with higher message caps. The plan ensures a secure workspace for teams, allowing the creation and sharing of custom GPTs within the workspace. Admin consoles for workspace and team management are also provided, along with early access to new features and improvements. ChatGPT Team is priced at $25\/month per user for annual billing or $30\/month per user for monthly billing. The company said that ChatGPT Enterprise, launched last year, is being utilised by industry leaders such as Block, Canva, Carlyle, The Estée Lauder Companies, PwC, and Zapier. Additionally, Pennsylvania Governor Josh Shapiro recently announced the deployment of OpenAI’s ChatGPT Enterprise service in a pilot program to assist state employees in administrative tasks. OpenAI has recently introduced GPTs, custom versions of ChatGPT designed for specific purposes, enabling businesses and teams to customise ChatGPT according to their specific needs and workflows without requiring any code. GPTs can be utilised for various tasks, including project management, team onboarding, code generation, data analysis, and more.","excerpt":"ChatGPT Team is priced at $25\/month per user for annual billing or $30\/month per user for monthly billing.","categories":["AI News"],"tags":["ChatGPT"],"author_name":"Siddharth Jindal","publish_date":"2024-01-10T23:35:39","publication_year":"2024","word_count":260,"keywords":["Go","ChatGPT","API","OpenAI","AI","programming_languages:R","GPT","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","Go","API","GPT","llm_models:GPT","llm_models:ChatGPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/chatgpt-now-has-a-team-plan\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10124822,"title":"AWS Brings ESM3 Language Models to Life Sciences","content":"AWS has announced a collaboration with EvolutionaryScale to bring their cutting-edge language models to scientists and researchers in the field of biology. This partnership aims to advance applications in drug discovery, carbon capture, and more. The collaboration introduces EvolutionaryScale’s ESM3, a state-of-the-art language model family, to AWS’s robust infrastructure. This includes enterprise-grade security, privacy measures, and purpose-built services for health and generative AI. The ESM3 model family, which includes generative, multimodal models, will be available on AWS platforms such as Amazon SageMaker and AWS HealthOmics, with support for Amazon Bedrock coming later this year. ESM3 enables researchers to generate complex multi-domain proteins from scratch, create protein design workflows, and incorporate functional understanding. This innovative approach, termed “programmable biology,” has the potential to significantly reduce the time and cost of bringing new therapeutics to market. Customers can start using ESM3 through Amazon SageMaker and orchestrate automated drug discovery workflows via AWS HealthOmics. The collaboration also leverages AWS’s generative AI infrastructure, including high-performance GPU instances and ML accelerators like AWS Trainium and AWS Inferentia, ensuring efficient training and deployment of ESM3 models. This initiative marks a significant leap in generative AI for biology, potentially accelerating drug discovery timelines and enabling novel therapeutic developments. The collaboration underscores AWS’s commitment to advancing generative AI across diverse industries, particularly in life sciences. This is just the latest in AWS updates. Only a few days ago, AWS also launched a new AWS Generative AI Spotlight programme in the Asia Pacific and Japan (APJ) region, a four-week accelerator aimed at supporting early-stage startups in the region developing generative AI applications. In India, AWS is collaborating with venture capital firm Accel for this initiative.","excerpt":"The ESM3 model family will be available on AWS SageMaker and HealthOmics, with Amazon Bedrock support later this year.","categories":["AI News"],"tags":["Amazon Bedrock","Language Models"],"author_name":"Gopika Raj","publish_date":"2024-06-26T12:11:47","publication_year":"2024","word_count":276,"keywords":["Go","API","Amazon Bedrock","Amazon SageMaker","AWS","AI","ML","RAG","Aim","generative AI","Language Models","R"],"extracted_tech_keywords":["AI","ML","generative AI","Aim","Amazon SageMaker","RAG","AWS","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-brings-esm3-language-models-to-life-sciences\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":30623,"title":"Manipal ProLearn Announces Excelerate Scholarship For Data Science Aspirants","content":"Manipal ProLearn, a leading professional learning platform and a division of Manipal Global Education Services (MaGE) announced the Excelerate scholarship for its Data Science program on 20 November, this year. Under the endowment, up to 10 meritorious students will receive financial assistance of Rs. 1 lakh each. Also, one outstanding candidate will be offered a full fee waiver of Rs. 5 lakhs. “The Excelerate scholarship is designed to encourage the most talented and students and professionals to apply for the program and help them transform their career trajectory,” said Gopal Devanahalli, Sr Vice-President, Manipal Global Education Services. The Excelerate Scholarship is available for all aspiring data scientists after they have been provisionally admitted into the program. The scholarship will entitle up to ten students with a fee waiver of Rs 1 lakh and one student with a full tuition fee waiver. The students would be evaluated through a scholarship entrance test which will assess them on aptitude, data interpretation\/ programming skills. These students would also be required to submit a Statement Of Purpose, which would be followed by an interview. “Sustained excellence for our learners is at the epicentre of Manipal Global’s ethos. Keeping true to this spirit, the Excelerate scholarship will form the bridge between raw talent and the industry’s quest for professionals in Data Science. It is a natural step towards leveraging our expertise and stature as a leading data science training organisation to supplement India’s rise as knowledge economy capital of the world,” said Dr Yogesh Kumar Bhatt, Vice-President, IT-Education and Training, Manipal ProLearn. Earlier this year, Manipal Prolearn had signed a memorandum of understanding with Equifax Inc, one of the world’s top three consumer credit reporting agencies, to create industry-ready professionals in the BFSI sector.The MoU was aimed at creating young professionals who are well-versed with data science and analytics skills.","excerpt":"Manipal ProLearn, a leading professional learning platform and a division of Manipal Global Education Services (MaGE) announced the Excelerate scholarship for its Data Science program on 20 November, this year. Under the endowment, up to 10 meritorious students will receive financial assistance of Rs. 1 lakh each. Also, one outstanding candidate will be offered a […]","categories":["AI News"],"tags":["Data Science","scholarship"],"author_name":"Disha Misal","publish_date":"2018-11-23T12:02:35","publication_year":"2018","word_count":305,"keywords":["data science","Go","API","programming_languages:R","AI","RAG","Aim","analytics","scholarship","GAN","Data Science","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","API","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/manipal-prolearn-announces-excelerate-scholarship-for-data-science-aspirants\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":28010,"title":"Enhancing Churn Prediction Models Using Social Network Analysis In Telecom Industry","content":"Telecom industry is the most competitive industry in the current period and hence customer churn or loss of the customer to competition is a big problem for this industry. It is more expensive to acquire a new customer than to keep the existing ones from leaving. Telecom companies spend hundreds of dollars to acquire a new customer and when that customer leaves, the company not only loses the future revenue from that customer but also the resources spend to acquire that customer. Churn rate has a strong impact on the lifetime value of the customer because it affects the length of service and the future revenue of the company. Telecom companies typically use two approaches to address churn – (a) Untargeted approach and (b) Targeted approach. The untargeted approach relies on superior product and mass advertising to increase brand loyalty and thus retain customers. The targeted approach relies on identifying customers who are likely to churn and provide suitable intervention to encourage them to stay. This approach can bring in a huge loss for a company, if churn predictions are inaccurate because then firms are wasting incentive money on customers who would have stayed anyway. There are numerous predictive modeling techniques for predicting customer churn. These vary in terms of statistical technique e.g., logistic regression, GBM or Naive Base etc. This paper describes, how Social Network Analysis can enhance the accuracy of a model if used along with normal predictive modeling in identifying the customers who are likely to churn well in advance. Background: Traditional churn models are designed to predict whether or not customers will churn. It treats customers as isolated entities. But individual customers are not isolated entities. Rather, customers are influenced by friends, friends of friends, and others within their network. The performance of normal predictive models using few variables generally not so accurate. But we can improve our models with social network data by asking questions like, who is my customer talking to? Who are his friends? Who is influencing him? This method was first researched by Bart Baesens, PhD, an Associate Professor at Katholieke Universiteit Leuven and presented his findings at the Analytics 2011 conference in Orlando. What is Social network analysis (SNA)? Social network analysis does not typically analyze data from social network sites like Facebook and Twitter, as many people assume. Rather, it is a data mining technique that explores the patterns between people in a network or group. Social networks are typically illustrated through a series of nodes and edges (or ties) that show which people (or households or companies) are connected. The edges or ties can also be weighted to show the strength of the connection. Why SNA? SNA is based on a general concept called community behavior. It is widely used in the marketing campaign by looking at Influencer. If you want to sell your products to your customers, convince their friends. We can use social network analysis to understand more about your customers and their communities. Similar is the case for churn, if we want to retain the customer we can convince the friends. And if everybody in the community is churning, there is a high chance that he will also be influenced by that. So it is very important to study the community behavior in place of individual behavior. Measures of SNA SNA works on the measure of centrality and there are many ways to calculate centrality. Centrality measures tell us who is the most influential person in the network. Influencers are considered to be the most connected customers in the network. In this step, the influence of each node in the network is represented by the score. The most influential nodes are the highest in score. Few measures of centrality are: Degree centrality: How many people can this person reach directly? Betweenness centrality: How likely is this person to be the most direct route between two people in the network? Closeness centrality: How fast can this person reach everyone in the network? Eigenvector centrality: How well is this person connected to other well-connected people? Fig.1. Graph Centrality Measures Methodology: The current project involved developing a Churn prediction model for mobile postpaid subscribers for a region. The objective of the project was to develop a model where in subscribers’ past behavior was used to predict the likelihood of churn of any subscriber, for the immediate next month. For this purpose, subscribers billing, product, plan, demographics and call details information for a period of 6 months was used to build the model. Overall active subscriber base in study was close to 1 Million and average number of Churners every month was ~ 12,000 – 15,000 subscribers. This paper explains how SNA can be used to augment the accuracy of typical churn prediction model if we use it along with machine learning techniques. The base prediction model was created using Logistic regression and the precision of the model was around 12%. In order to improve the accuracy of the model further up, an ensemble approach was deployed. The reason behind using the ensemble approach was that the response rate in the data was close to 4% which posed a rare event modelling problem. As a result, conventional modelling models did not yield any significant result in terms of predicting churners. In order to improve the precision of overall ensemble model, feature enhancement was done by including more variables from SNA based interaction of the subscribers. By including these augmented variables, the overall accuracy of the ensemble was enhanced by another 13%. Hence, multiple combination of different models was used to extract dominant churn patterns from the data. Model was developed using an ensemble approach comprising of 2 layers: Stacking Random Forest + XGBoost Stacking Logistic Regression + Augmented SNA Variables + XGBoost Fig.2. Ensemble Model Results: From Model 1 (Fig.2), Churner precision was 23% By using ensemble Model 2 (Fig.2), Churner precision was improved by 13% Current Status Predicted Status Number of Subscribers Churner Churner 1254727 Churner Active 11061 Active Churner 11986 Active Active 6261 Metric Value Precision 36.14% Recall 34.31% Accuracy 98.20% For a given set of likely churners, percent of subscribers correctly predicted => Precision = 36.1% Execution Environment: Whole process was executed using Aster (Big Data Platform) and R. Conclusion: SNA is a popular concept used for understanding the social behavior of customers in different aspects. Telecom service provider use it to understand the subscribers’ community and try to use that information to retain the existing customers and also create newer avenues for up-sell and cross-sell. SNA can be used as a fusion technique along with traditional machine learning models, to further enhance the accuracy of the churn models. One important caveat around deployment of SNA is that it is helpful only for the scenarios where in there is really an influence of subscribers community on each other. References: https:\/\/www.gartner.com\/it-glossary\/social-network-analysis-sna S. Sumathi and S.N. Sivanandam: “Data Mining in Telecommunications and Control”, Studies in Computational Intelligence (SCI) 29, 615–627 (2006). S. Wasserman and K.Faust. “Social Network Analysis: Methods and Applications”. CambridgeUniversity Press, Cambridge, UK, 1994. L. Freeman. “Centrality in social networks: Conceptual clarifications”. Social Networks, 1:215-239,1979. Wei, C., & Chiu, I (2002).” Turning telecommunications call details to churn prediction: A datamining approach”. Expert Systems with Applications, 23(2), 103-112. Authors Dr. Archana Kumari archanaks17@gmail.com Archana has worked in Telecom, Retail, Automobile, Insurance, Media and HR analytics. By Education she is PhD in Economics and Business Analytics and Intelligence from IIM Bangalore. She is passionate towards analytics. Ashutosh Srivastava Mail.ashutoshsri@gmail.com Ashutosh has worked in Telecom and Retail analytics sector. By education he is an MBA and Computer Engineer. He is passionate towards learning and applying new Machine Learning techniques in the field of analytics.","excerpt":"Telecom industry is the most competitive industry in the current period and hence customer churn or loss of the customer to competition is a big problem for this industry. It is more expensive to acquire a new customer than to keep the existing ones from leaving. Telecom companies spend hundreds of dollars to acquire a […]","categories":["AI Features"],"tags":["churn prediction model"],"author_name":"AIM Media House","publish_date":"2018-09-05T06:24:01","publication_year":"2018","word_count":1284,"keywords":["big data","machine learning","programming_languages:R","AI","RAG","XGBoost","analytics","churn prediction model","R"],"extracted_tech_keywords":["AI","machine learning","analytics","XGBoost","RAG","R","big data","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/enhancing-churn-prediction-models-using-social-network-analysis-in-telecom-industry\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10067776,"title":"Sridhar Vembu joins Anand Mahindra to back manual scavenging robotics startup","content":"SaaS giant Zoho Corp recently invested INR 20 crores (USD 2.5 million) in AI-powered robotics startup Genrobotics. With this investment, Zoho Corp looks to assist the startup on its mission to eradicate manual scavenging in India. The company believes that it reflects its mission to nurture the development of a deep-tech ecosystem in the country. Prior to this, Genrobotics has also raised investments from Anand Mahindra, Rajan Anandan, Unicorn Ventures and SEA funds. To date, the company has raised total funding of USD 3 million. One of those investments that makes you feel really good. A great bunch of guys serving a great purpose. I hope municipalities across the country buy the machine. I hope they sell their products across the globe & become a prime example of ‘Made in India!’ https:\/\/t.co\/eFHWjRyzlT— anand mahindra (@anandmahindra) October 7, 2020 Founded in 2017, Trivandrum-based Genrobotics provides better and safer methods to the people working in extreme and unsafe conditions using robotics and AI. The company has developed the world’s first robotic scavenger called ‘Bandicoot’ and the robot-assisted gait training solution ‘G Gaiter’ for the fastest rehabilitation of paraplegic patients. As part of the Make in India initiative, Genrobotics products are completely designed and manufactured in the country. It currently serves customers in the UK, Malaysia, UAE, and South Korea. Zoho Corp chief Sridhar Vembu said that building such technological competencies and critical know-how locally can help foster sustainable growth across sectors, including industrial manufacturing, healthcare, and energy, making the country economically stronger and self-reliant. Further, he said that making this a reality requires focused, long-term investments that support home-grown deep-tech startups through research and development and engineering phases and bring ideas to the market faster. “We are happy to fast-track their efforts and support them in their mission to end manual scavenging,” he added, commenting on their recent investment in Genrobotics. https:\/\/twitter.com\/svembu\/status\/1529344086506885120 Bandicoot One of its flagship offerings, Bandicoot, helps clean confined spaces such as sewers holes, sewer wells, stormwater manholes, oily water sewers and stormwater sewers in refineries. At present, smart cities, urban local bodies, refineries, MNCs, townships and housing colonies across 14 states are using their robots, eliminating the need for human entry into manholes. Here’s a quick glimpse of the Bandicoot robot: “We have rehabilitated hundreds of people working as manual scavengers by training them to be robot operators,” said Genrobotics’ CEO Vimal Govind MK. He said for India to end manual scavenging, more than one lakh robots will be required. As we scale to fill the need gap, we estimate the creation of nearly five-lakh jobs across the country. The investment from Zoho will help us to expand our advanced R&D infrastructure, build large-scale production facilities, hire more talent, increase our exports to ASEAN markets and expand our global footprints.” G Gaiter Genrobotics has recently ventured into healthcare and launched a robot-assisted gait training solution – G Gaiter – to help rehabilitate paraplegia patients.","excerpt":"With this investment, Zoho Corp looks to assist the startup on its mission to eradicate manual scavenging in India.","categories":["AI News"],"tags":["AI Startups"],"author_name":"Amit Naik","publish_date":"2022-05-25T11:52:32","publication_year":"2022","word_count":486,"keywords":["Go","funding","unicorn","programming_languages:R","AI","programming_languages:Go","ai_applications:robotics","R","AI Startups","startup"],"extracted_tech_keywords":["AI","R","Go","startup","unicorn","funding","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sridhar-vembu-joins-anand-mahindra-to-back-manual-scavenging-robotics-startup\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":12782,"title":"Top 10 Analytical Puzzles for Interview","content":"The analytics industry predominantly relies on professionals and analysts who not only excel in extensive use of statistics and Data handling tools but also exhibit excellent problem-solving skills. However, a fresher entering the domain doesn’t necessarily have to know all these skills in advance. So, most of these interviews entail puzzles, logical reasoning problems, and questions that help in testing a candidate’s lateral thinking. This assessment enables the interviewer to assert the candidate’s logical reasoning problems and questions to test lateral thinking, which is an important part of most analytics job interviews. Most of the puzzles asked during analytics-based interviews can be generalized into 3 sections – technical, straightforward, and the ones unsolvable. It’s advisable for aspiring candidates to practice these puzzle sets in advance, so as to gain more proficiency. The most ideal approach to solving these puzzles is to focus on the data provided while developing a structured approach. Most importantly, you must be able to explain your approach to the interviewer. Let us glance through some of the most common puzzles that baffle candidates appearing for analytics interviews. Top 10 Analytical Puzzles for Interview with Answer 1. There are 5 lanes on a race track. One needs to find out the 3 fastest horses among total of 25. Find out the minimum number of races to be conducted in order to determine the fastest three. 25 horses, 5 tracks puzzle This is one of the most common interview puzzles asked by analytics interviewers. This puzzle tests interviewee’s approach to solving the problem. The approach entails conducting 5 races where each race group would involve 5 horses. In the ensuing step, a sixth race is conducted between winners of first 5 races to determine the 3 fastest horses (marked A1, B1, A=and C1). The seventh race is conducted between horses B1, C1, second and third horse from the horse A1’s group (A2, A3), second horse from horse B1’s group (B2). The horses that finish 1st and 2nd in the seventh race, are actually the 2nd and the 3rd fastest horses among all horses. 2. There are 3 mislabeled jars, with apple and oranges in the first and second jar respectively. The third jar contains a mixture of apples and oranges. You can pick as many fruits as required to precisely label each jar. Determine the minimum number of fruits to be picked up in the process of labeling the jars. Mislabeled jar puzzle This is another tricky puzzle where you must really churn your brain. A noticeable aspect in this puzzles is the fact that there’s a circular misplacement, which implies if apple is wrongly labelled as Apple, Apple can’t be labelled as Orange, i.e., it has to be labeled as A+O. We are acquainted with the fact that everything is wrongly placed, which means A+O jar contains either Apple or Orange (but not both). The candidate picks one fruit from A+O, and let’s assume he gets an apple. He labels the jar as apple, however, jar labelled Apple can’t have A+O. Thus, the third jar left in the process should be labelled A+O. Basically, picking only one fruit helps in correctly labeling the jars. 3. There are 8 batteries, but only 4 of them work. You have to use them for a flashlight which needs only 2 working batteries. To guarantee that the flashlight is turned on, what is the minimum number of battery pairs you need to test? Faulty battery puzzle To solve this problem, the first step involves naming the batteries, for instance, A, B, C, D, E, F, G, and H. In this problem, you can’t compare 2 items directly. If a combination of two batteries fail to turn the light on, it means either one or both the batteries aren’t working. The candidate has to approach the puzzle in a circular manner. The batteries are put test consecutively in the order AB, BC, and AC. At most, one of the three batteries between A, B, And C is working, only if none of the pairs work. This also implies that at least three batteries between D, E, F, G, and H must be functional. DE combination is tried next. If they don’t work, at least 2 out of F, G, and H must work. Similarly, try the combinations FG, GH, and FH to positively asset which batteries really work. 4. 10 coins are placed before you on a table, while you stay blindfolded. The candidate is permitted to touch the coins, however conditions to the puzzle dictates that he can’t really determine which way up they are by feel. 5 coins are placed heads up, while the other 5 are kept tails up, without the interviewee knowing which ones are which. If you’re allowed to flip the coins any number of times, how would you build two piles of coins each with the same number of heads up. 10 coins puzzle This is another common analytics puzzle revolving around coins. This problem can be solved by initially creating two piles of coin, with 5 randomly selected coins in each pile. Let’s assume the first pile looks like H, H, H, H, T and the other pile can be imagined as T, T, T, T, H. The final bit in solving the puzzle involves flipping all coins in the second pile to finally obtain same number of heads. This common coin puzzles tends to pose confusion in the candidates mind. 5. Two trains, separated by a distance of 80km, are running towards each other on the same track at a speed of 40kmph. A bird takes its flight from train X and flies towards train Y at a constant speed of 100kmph. Once it reaches train Y, it turns and start flying back toward strain X. The bird keeps flying to and forth till both the trains collide. Determine the distance travelled by the bird. Train collision puzzle This is another common analytics-based puzzle where the candidate has to use his quick thinking and mathematical skills to find the answer. Ideally the problem should take a minute for solving. The problem can be solved mathematically in the following few steps: Velocity of approach for two trains = (40+40)km\/hr Time taken for the trains to collide = 80km\/80km\/hr = 1hour The total distance travelled by the bird = 100km\/hr * 1hr = 100km Through this puzzle, the interviewer is testing your approach. Consider yourself rejected if the approach you took involves calculating distance from X to Y, and then Y to X, and so on. 6. A birthday cake has to be equally divided into 8 equal pieces in exactly 3 cuts. Determine the way to make this division possible. Cake cutting puzzle This puzzle is not really difficult to solve if you really put your mind to work. The approach entails slicing the cake horizontally down the centre, followed by making another division vertically through the centre. The two divisions made across horizontal and vertical directions will give you 4 equal pieces of the cake. In the final step, simply stack the 4 pieces one above the other, and then make the third division, splitting the stack into half. This gives you the 8 equal pieces of cake, along with answer to your puzzle. 7. You pull out 2 balls, one after another, from a bag which has 20 blue and 13 red balls in total. If the balls are of similar colour, then the balls are replaced with a blue ball, however, if the balls are of different colours, then a red ball is used to replace them. Once the balls are taken out of the bag, they are not placed back in the bag, and thus the number of balls keep reducing. Determine the colour of last ball left in the bag. Red and Blue balls in a bag puzzle This puzzle usually seems like a tough one to answer, but solving it once makes one realize that the procedure was actually simple. If the candidate pulls out 2 red balls, he replaces them with a blue ball. On the other hand, if he draws one red and one blue, it is replaced with a red one. This implies that the red ball would always be in odd numbers, whether the candidate removes 2 together, or removes 1 while adding 1. This also indicates that the last ball to stay in the bag would be a red one. The interviewer is merely testing the approach the candidate applies in solving this common analytics puzzle. 8. There are 2 jugs with 4 litres and 5 litres of water respectively. The objective is to pour exactly 7 litres of water in a bucket. How can it be accomplished? Water and jugs puzzle This question can be rated of medium difficulty and shouldn’t ideally take more than 2 minutes to answer. The approach here is to initially fill the 5L jug with water and empty the same into the 4L jug. The 5L jug will be left with 1L of water, which is poured into the bucket. Meanwhile, empty the 4L jug. The above step is repeated, so that the bucket now is filled with 2L of water. Finally, fill the 5L jug with water and empty the same into the bucket. The bucket will now have 7L of water, as you add % L directly to the previously collected 2L of water in the bucket. 9. There are 10 stacks of 10 coins each, where each coin weighs 10gms. However, one of the stacks is defective, and that stack contains coins which weigh 9gms. Determine the minimum number of weights needed to identify the defective stack. Defective coin puzzle This is another tricky puzzle, asked commonly during analytics-based interviews. To solve this problem, the trick lies in creating a weighted stack for measurement, which will enable the candidate to identify the defective stack in one measurement. A coin is taken from the first stack, 2 from the second, 3 from the third, and so on. This will give a total of 55 coins in hand. If none of them are defective, they would weigh 550gms together. However, if stack 1 turns defective, the total weight would stand at 549gms; defect in stack 2 would result in a total weight of 548gms; and so on. Therefore, just one measurement can help the candidate identify the faulty stack. 10. There are 3 switches in a room, where one of them is assigned for a bulb in the next room. You can’t see whether the bulb is on or off, until you leave the room. Find the minimum number of times you have to go into the room to identify which switch corresponds to the bulb in the other room. One bulb with 3 switches puzzle This question has been asked several times during different analytics-based interviews. The person has to initially turn on the first switch and keep it on for 2-3 minutes. Next, turn off the first switch and turn on the second one. Rush to the other room as soon as you turn on the second switch. If the bulb is glowing, the second switch corresponds to the light bulb; however, if the bulb doesn’t glow, but touching it feels warm, the first switch is the one that turns the bulb on. If it’s neither lit, nor warm, then the third switch is the desired switch. So, a person must go only once to the other room to find out the accurate switch.","excerpt":"The analytics industry predominantly relies on professionals and analysts who not only excel in extensive use of statistics and Data handling tools but also exhibit excellent problem-solving skills. However, a fresher entering the domain doesn’t necessarily have to know all these skills in advance. So, most of these interviews entail puzzles, logical reasoning problems, and […]","categories":["AI Trends"],"tags":["analytics industry india"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-02-15T11:56:03","publication_year":"2017","word_count":1919,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","ML","programming_languages:Go","analytics","analytics industry india","R"],"extracted_tech_keywords":["AI","ML","analytics","AWS","R","Go","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-standard-puzzles-asked-analytics-interviews\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119530,"title":"Meta Spends $30 Billion on a Million NVIDIA GPUs to Train its AI Models","content":"In a “staggering” revelation, Meta AI chief Yann LeCun confirmed that Meta has obtained $30 billion worth of NVIDIA GPUs to train their AI models. Enough to run a small nation or even put a man on the moon in 1969. Speaking at the Forging the Future of Business with AI Summit organised by Imagination in Action, LeCun said that more variations of Llama-3 would be out over the next few months, with training and fine-tuning currently taking place. “Despite all the computers we have on our hands, it still takes a lot of time to fine-tune, but a bunch of variations on those models are going to come out over the next few months,” he said. Speaking of fine-tuning and training, host John Werner stated that Meta had bought an additional 500,000 GPUs from NVIDIA, taking the total number of NVIDIA GPUs up to a million, with a retail value of $30 billion. Combining the total costs of the GPUs so far, Werner pointed out that the training of the model exceeded the costs of the entire Apollo space programme, which back in the 1960s, amounted to about $25.4 billion. Agreeing, LeCun said, “Yeah, it’s staggering, isn’t it? A lot of it, not just training, but deployment, is limited by computational abilities. One of the issues that we’re facing is the supply of GPUs and the cost of them at the moment. Obviously, adjusted for inflation, the Apollo programme still outsells the Meta in terms of how much was actually spent, with roughly $257 billion spent. But it’s no secret that the cost of GPUs is a continuously growing expense for AI companies. Recently, OpenAI’s Sam Altman said that he doesn’t care if the company spends upwards of $50 billion a year in developing AGI. The company, as of March, employs as many as 720,000 NVIDIA H100 GPUs for Sora alone. This amounts to about $21.6 billion. Similarly, all big tech companies are hoping to expand how many GPUs they can obtain by the end of the year, or even by 2025. Microsoft is aiming for 1.8 million GPUs by the end of the year. Meanwhile, OpenAI hopes to use 10 million GPUs for their latest AI model. In the meantime, NVIDIA has also been churning out GPUs, with their latest DGX H200 GPU being hand-delivered by CEO Jensen Huang to Altman. Coming back to LeCun, he pointed out that the need of the hour was the ability to upscale learning algorithms so they could be parallelised across several GPUs. “Progress on this has been kind of slow in the community, so I think we’re kind of waiting for breakthroughs there,” he said. With that occurring, costs could potentially lower for AI companies, though with increasingly fast upscaling overall, demand could remain the same.","excerpt":"Larger than the Apollo Moon Mission, which cost about $26 billion between 1968 and 1973.","categories":["AI News"],"tags":["Meta AI","NVIDIA GPU","Yann LeCun"],"author_name":"Donna Eva","publish_date":"2024-05-03T15:02:38","publication_year":"2024","word_count":465,"keywords":["Go","NVIDIA H100","NVIDIA GPU","Meta AI","Yann LeCun","OpenAI","AI","programming_languages:R","Aim","GAN","llm_models:Llama","R"],"extracted_tech_keywords":["AI","OpenAI","Meta AI","Aim","R","Go","GAN","NVIDIA H100","llm_models:Llama","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-spends-30-billion-on-a-million-nvidia-gpus-to-train-its-ai-models\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":14676,"title":"Infosys MANA &#038; NIA &#8212; the next-gen AI platform for enterprises","content":"Vishal Sikka, CEO for Infosys With automation the new buzzword, Infosys the Indian IT bellwether launched AI platform, Mana with a goal to automate repetitive and commoditised software maintenance tasks earlier last year. The platform combines machine learning together with ‘deep knowledge of an organisation.’ Businesses have utilized Mana to continuously reinvent their system landscapes and lower maintenance cost of assets. “Building on this deep experience, Infosys recognised the need to bring artificial intelligence to the enterprise in a meaningful and purposeful way,” Vishal Sikka, CEO and Managing Director, Infosys had shared. To keep up with the pace of development, Infosys had hired a number of executives from Silicon Valley to build the top-notch cognitive system. Interestingly, Google Executive Sudhir Jha was hired as a Senior Vice President to head product management and strategy around Mana. The platform is built upon the three primary software platforms of Infosys – the Infosys Automation Platform, the Infosys Information Platform, and the Infosys Knowledge Platform. With Mana, the organization helps in building knowledge, leveraging AI and machine learning.Mana is coupled with Infosys’ Aikido offerings to help clients capture knowledge, while delivering new and delightful experiences to their end users. Mana versus other automation platforms? AI can deliver multiple benefits to entrepreneurs Infosys is not the only one with its own automation platform. Wipro launched HOLMES, TCS came up with Ignio, while Tech Mahindra calls its automation platform TACTiX. All these platforms help clients in automating projects to improve efficiency. Essentially, these tools help in automating back-office support jobs. With MANA, Infosys attempted to build a top-notch cognitive computing system that could potentially rival IBM’s flagship Watson.  Clients across industries make sense of large sets of unstructured data sets using Watson. Unlike Mana, Watson was built to function like our brain, learning through experiences, finding correlations, and remembering and learning from the outcomes. However, Mana has had a major success in the market and has proven itself as one of the most efficient automation platforms available today. Mana can also be used for predictive analytics and predictive maintenance. The market is still in its infancy, and industry is investing significantly into intelligent automation. To stay relevant in the market, Infosys revamped Mana recently. Such innovations will help Mana to reach the level of Watson. Is Mana fuelling growth The adoption of Infosys’ automation platform Mana has doubled in the past few quarters, though from a small base. At the end of 2016-17, Mana had 50 customers and 150 engagements. “Contributions came from new areas. For 2016-17, we gained revenue from new software and software-related services like Mana,” remarks Sikka. Reports suggest that more than three quarters of business leaders relied on artificial intelligence for success of their strategy. This only means that the potential market for Mana’s adoption is expanding. MANA Use Cases: Mana at work Infosys Mana is automating jobs Banking: Case in point is a bank in Asia that needed help with automating their contract processing job. The bank would usually engage 10-15 dedicated lawyers for the same job. Through the AI platform, the bank eliminated the need for 10-15 lawyers to bring the task up to speed and deployed MANA to analyse non-disclosure agreement and contractual documents. FMCG:  Another use case for Mana involves a food and beverage manufacturer. Sales Managers at this firm leveraged the platform to automate sales planning processes by automatically resolving maintenance tickets of recurring issues. Mana’s self-learning capabilities helped it to provide solutions to known problems automatically, over time. This helped the company in significantly reducing the time invested in generating a solution to a maintenance problem. Infosys adds teeth to its AI platform with Nia The success of Infosys’s first-generation platform, Mana, encouraged the Indian tech giant to launch Nia. The new platform combines big data\/analytics, machine learning, knowledge management, and cognitive automation capabilities of Mana with end-to-end RPA capabilities of AssistEdge; high-performance and scalable machine learning capabilities of Skytree; and optical character recognition (OCR), natural language processing (NLP) capabilities, and infrastructure management services. “Nia, the next generation of our AI platform now takes our purposeful approach to AI, one in which technology serves to amplify people and empowers them to work in new ways, to new heights, notes Vishal Sikka, CEO, Infosys. The comprehensive platform addresses critical business problems such as forecasting revenues, forecasting what products need to be built, while understanding customer behavior. IBM Watson Key business challenges that Nia addresses: Improving order-to-cash process by creating a real-time risk profile. This helps in customizing the collection strategy, expediting resolution of disputes, and predicting anomalies, while enabling visibility and forecasting cash flow to reduce days sales outstanding (DSO) Predicting variability in manufacturing and material cost, while reducing product development cycle times remarkably. Creation of knowledge models of multiple, complex labor contracts. These models combine an on-demand, self-service conversational interface to operationalize the knowledge Last Word In an age of AI, MANA and now NIA have led to tremendous adoption of India’s biggest software services providers, allowing businesses to improve delivery, accelerate to market and driving operational efficiency. And even though the platform was billed to take on IBM Watson, it is still yet to be seen.","excerpt":"With automation the new buzzword, Infosys the Indian IT bellwether launched AI platform, Mana with a goal to automate repetitive and commoditised software maintenance tasks earlier last year. The platform combines machine learning together with ‘deep knowledge of an organisation.’ Businesses have utilized Mana to continuously reinvent their system landscapes and lower maintenance cost of […]","categories":["IT Services"],"tags":["AI in manufacturing","Artificial Intelligence India","automation India","Machine Learning India","Manufacturing","NLP","OCR"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-05-01T10:14:31","publication_year":"2017","word_count":862,"keywords":["Go","artificial intelligence","OCR","machine learning","AI","R","Machine Learning India","Artificial Intelligence India","Scala","RAG","NLP","automation India","analytics","AI in manufacturing","predictive analytics","Manufacturing"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","analytics","RAG","predictive analytics","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/it-services\/infosys-mana-nia-next-gen-ai-platform-enterprises\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169025,"title":"50% of Our Code Will be AI-Generated in 18 Months: Aurigo CEO on Expanding India GCC","content":"US-based Aurigo Software, known for developing capital planning and construction management software for large infrastructure projects, has recently expanded its presence in India, not just as a backend hub but as a core engine of innovation. The GCC in India now serves as the company’s primary hub for product development. Of its 600 global employees, nearly 500 are now based in Bengaluru, making India the undisputed R&D powerhouse for the company. “We’ve always had a small R&D centre here,” Balaji Sreenivasan, founder and CEO of Aurigo, said in an exclusive interaction with AIM. “However, the last five to six years have seen an intentional shift. All of our engineering is now centred in Bengaluru.” Sreenivasan, however, pointed out that this isn’t a conventional GCC. “We’re not a service arm. This is a product-first team where the engineering actually happens—the code is written here, the innovation happens here.” However, Sreenivasan is quick to add that headcount is no longer the leading indicator of growth in tech, especially in an AI-first era. “The future of software companies doesn’t necessarily mean bigger teams,” he said. “We expect 50% of our code to be AI-generated within the next 18 months.” This trend is gaining momentum. For example, Y Combinator-backed startups claim to have generated 90% of their code by AI, while tech giants Google & Microsoft attribute approximately 30% of their code output to AI tools. That’s where AI has truly shaped Aurigo’s product development lifecycle as well. Sreenivasan explained that while it took four to five years to achieve the right product-market fit for a government project, the company was able to build a prototype for the private sector in under 90 days. It is set to go into production by September, all with the help of tools like Lovable and Figma AI. What AI Means for Aurigo’s Workforce Despite being bullish on AI, Sreenivasan has no illusions about the human side of the equation. “It’s not about replacing everyone with AI agents,” he clarified. “It’s about asking: what do the next 600 employees look like?” While some manual tasks, like testing, will undoubtedly be automated, the focus is on reskilling. As the company continues to grow at 30-40% year-over-year, its hiring strategy is evolving. It’s not about endlessly expanding the team, but about hiring smarter and faster. Aurigo is already building AI agents for both internal use and customer-facing solutions, with plans to deploy approximately 15 AI agents in the next quarter. One major AI offering for Aurigo is Lumina GPT—a suite of AI tools built with the help of OpenAI’s API, Meta’s Llama, and other open-source models. It sits on top of Aurigo’s flagship platform, Masterworks, hosted on AWS. Eventually, Aurigo wants Masterworks to have Lumina GPT natively integrated. To support this, Aurigo is relying on a combination of open-source tools and in-house code. While platforms like Lovable are used for rapid prototyping, the production code for enterprise-grade deployments is still largely proprietary. From Government to Data Centres Aurigo’s products have traditionally catered to government agencies, helping manage massive public infrastructure like highways and bridges. However, the market is evolving. A good example of this is Aurigo Engage, which is essentially a digital town hall that helps governments make more equitable infrastructure decisions. It allows citizens to engage via their phones or web browsers easily. Through quick polls and comments, people can express support or opposition to projects like flyovers or metro stations. The platform’s natural language processing (NLP) engine then analyses this feedback at scale, surfacing insights that guide cities and states on where public funds should be spent to reflect the collective will. Another of Aurigo’s long-standing clients is the state of Utah. Back in 2016, Utah selected Aurigo after an extensive procurement process that included evaluating solutions from Oracle, American Association of State Highway and Transportation Officials (AASHTO), KPMG, and others. The state was losing approximately $100 million annually due to inefficiencies like poor data access, a lack of transparency, and delays in project oversight. From planning and prioritisation to construction management and post-completion maintenance, Utah has transitioned from using a single module to embracing the entire suite since adopting Aurigo’s platform. Over the past six years, the state has reported efficiency gains of 4–5% annually, translating to $50–60 million in savings, which has contributed to a billion-dollar infrastructure budget. That’s taxpayer money being preserved and reinvested, with the software costing less than 0.5% of the total spend. “We sell to government and enterprise clients, so the bar for security, latency, and reliability is high,” Sreenivasan explained. “That’s why we’ve conservatively targeted 50% AI-generated code. If we were building consumer apps, it might be 90%. But our customers demand more.” The company envisions a future where user interfaces themselves are transformed. Instead of building dashboards and reports, users will just ask the system questions, and the AI agent will fetch the data and provide a narrative. Most of its clients are in the US, as India remains somewhat sceptical of AI in enterprises, even though consumers are ready to adopt AI. “We realised that the next 10 years of construction will also be about data centres, EV corridors, power grids, and manufacturing facilities,” Sreenivasan explained. “That meant we had to build a whole new set of products for the private sector.” For Sreenivasan, adopting AI is no longer a strategic choice—it’s a necessity. “You don’t want to be using a matchstick when everyone else has a lighter. You don’t want to bring a knife to a gunfight,” he said.","excerpt":"“It’s not about replacing everyone with AI agents. It’s about asking: what do the next 600 employees look like?”","categories":["GCC"],"tags":["GCC"],"author_name":"Mohit Pandey","publish_date":"2025-05-02T12:37:42","publication_year":"2025","word_count":921,"keywords":["Go","API","GCC","TPU","OpenAI","AI","AWS","Git","NLP","Aim","R"],"extracted_tech_keywords":["AI","NLP","OpenAI","Aim","AWS","TPU","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/gcc\/50-of-our-code-will-be-ai-generated-in-18-months-aurigo-ceo-on-expanding-india-gcc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10136870,"title":"Minsky Awards for Excellence in AI 2024: Meet the Winners","content":"The Minsky Awards for Excellence in AI are among the most prestigious AI accolades, celebrating organizations that have demonstrated exceptional innovation and leadership in artificial intelligence. The awards aim to honor those who have successfully applied data, analytics, and AI to foster innovation, develop new products, and drive operational effectiveness. Held annually at Cypher, India’s largest AI summit, the awards have become a focal point for recognizing excellence in the AI industry. A Transparent and Rigorous Selection Process One of the standout features of the Minsky Awards is the transparent and fair selection process. The awards are open to organizations of all sizes and industries, with no hidden fees or costs involved in the submission process. This has established the Minsky Awards as a credible and highly sought-after recognition, known for upholding the integrity of the selection process. In 2024, the awards received an overwhelming 700+ nominations from various organizations, both domestic and international. The process for choosing the winners is designed to ensure that each entry is rigorously evaluated for its contribution to AI and its potential impact on the industry. Here’s how the selection process works: Nominations: Organizations across industries submit their nominations for one or more of the 23 categories. Each nomination is reviewed based on the submission form, which includes detailed descriptions of the AI project, product, or strategy being considered. Evaluation Criteria: The nominations are evaluated based on three key criteria: Innovation: How unique and groundbreaking the AI solution or strategy is. Impact: The measurable impact of the AI initiative on end-users, industries, or broader societal challenges. Quality of Contribution: The rigor, methodology, and depth of expertise involved in the AI project, ensuring the claims are valid and substantiated. Selection Panel: A highly experienced panel of judges from AIM (Analytics India Magazine) oversees the selection process. The panel, which includes AI experts, industry leaders, and data scientists, carefully reviews each nomination, assessing the submissions against the evaluation criteria. Final Selection: After an exhaustive review, the panel selects the top organizations for each category. These winners represent the best in their respective fields, showcasing innovation, leadership, and transformative contributions to the AI landscape. With over 700 nominations received this year, the competition was fierce. Yet, the transparency and rigor of the process ensure that each award is a true recognition of excellence. Spotlight on Winners The Minsky Awards bring into the spotlight several firms making waves in the AI space. Let’s take a closer look at some of the standout winners. Wells Fargo – AI for Social Impact A financial giant with assets totaling approximately $1.9 trillion, Wells Fargo is a leader in utilizing AI for social good. The company has deployed AI to improve accessibility and community development, addressing key societal challenges while enhancing customer experiences across its banking and mortgage services. AstraZeneca – AI Leadership Award A global, science-driven pharmaceutical company, AstraZeneca is transforming healthcare through cutting-edge AI technologies. The company’s AI-driven approach has revolutionized patient care, especially in precision medicine and drug development, demonstrating leadership in both healthcare and AI. AB InBev – AI-Driven Retail & Supply Chain Excellence by a GCC AB InBev, a global leader in the beverage industry, leverages AI to optimize supply chain operations. The company’s forward-thinking approach ensures seamless delivery of products and improved customer satisfaction. Their use of AI in predictive analytics and demand forecasting has set a new industry benchmark. Tredence – Best AI Implementation by a Service Provider Known for its AI-driven consulting and solutions, Tredence has consistently demonstrated its ability to deliver transformative results for its clients. By developing innovative AI models, Tredence helps businesses in retail, healthcare, and manufacturing sectors optimize their operations and gain competitive advantages. American Express – Best AI Implementation in a GCC\/Captive Environment American Express has implemented AI across its Global Capability Centers, focusing on customer service, fraud detection, and personalized experiences. Its AI-driven innovations have enhanced operational efficiency and elevated customer experiences, making it a leader in financial services. Walmart Global Tech India – Best AI Implementation in a GCC\/Captive Environment Walmart Global Tech India has harnessed AI to revolutionize its retail operations, particularly in supply chain optimization and customer experience enhancement. By employing AI models to predict inventory needs and customer preferences, the company has improved efficiency while setting new standards for retail technology. CASHe – Best AI-Enabled Business Strategy in an Indian Firm A fintech leader, CASHe is known for its customer-centric AI strategy, particularly in the lending space. CASHe’s innovative use of AI to assess creditworthiness and streamline loan approvals has empowered underserved populations and created new avenues for financial inclusion in India. Synechron – Breakthrough AI Technology Synechron is celebrated for its cutting-edge AI solutions that are transforming industries. Known for applying AI to financial services, Synechron’s AI-driven tools have enhanced everything from regulatory compliance to fraud detection, making the company a standout in the global tech landscape. EY GDS and DBS Tech India – Data Engineering Excellence in a GCC\/Captive Environment Both EY GDS and DBS Tech India were recognized for their stellar data engineering capabilities. Their data pipelines and AI-driven insights have enabled organizations to make informed, data-backed decisions, which have transformed operations and enhanced client relationships. Tiger Analytics – Excellence in AI Strategy Consulting A leader in the analytics consulting space, Tiger Analytics has been instrumental in helping businesses implement robust AI strategies. Their deep expertise and innovative solutions have helped clients across multiple sectors, including retail, healthcare, and finance, achieve their AI-driven goals. John Deere India Pvt Ltd – Excellence in AI Talent Development within GCC\/Captives A leading name in the agricultural machinery sector, John Deere India is dedicated to building a skilled workforce in AI. Through extensive training and talent development programs, the company has nurtured a culture of innovation and operational excellence. Rohit Ramanand (New Relic) and Debasis Bal (Fidelity Investments) – GCC AI Visionary Award Both Rohit Ramanand, Group Vice President of Engineering at New Relic, and Debasis Bal, Senior Vice President of Data Science & AI at Fidelity Investments, were honored for their visionary leadership in AI. Their efforts have transformed how their organizations approach AI strategy and implementation. Axtria Inc and Genpact – Leading AI Service Providers Axtria Inc and Genpact were recognized as top AI service providers for their ability to deliver cutting-edge AI solutions. Their innovative use of AI has empowered businesses to adopt data-driven decision-making processes and optimize their operations. HDFC Bank Ltd. – Leading Domestic Indian Firm in AI Innovation A leading financial services company, HDFC Ltd. has made significant strides in integrating AI into its operations, from customer service to financial products. HDFC’s AI innovations have enhanced customer experiences and streamlined various banking operations. EXL Service – Most Innovative Use of AI in Service Delivery EXL Service was recognized for its creative AI applications that transform how services are delivered across sectors such as healthcare, insurance, and finance. Their AI solutions have improved operational efficiency and customer satisfaction. Ramoji Group IT – Outstanding AI Product by a Domestic Indian Firm Ramoji Group IT, a major player in the Indian business landscape, was recognized for developing an exceptional AI product that significantly improves operational efficiency and delivers transformative results. Their innovative use of AI across various sectors earned them the Outstanding AI Product by a Domestic Indian Firm award. Outstanding Contribution to AI Skill Development AI skill development is crucial for the industry’s growth. Epam Systems and Intel Corporation were both recognized for their Outstanding Contribution to AI Skill Development, thanks to their comprehensive training programs and commitment to fostering AI talent. Rising Star GCC in AI Renault Nissan Technology and Business Center India emerged as the Rising Star GCC in AI, honored for its innovative approach and strategic use of AI to enhance operations and customer experiences. NVIDIA – Leading GCC\/Captive for AI Innovation NVIDIA, a global leader in computing technology, was awarded Leading GCC\/Captive for AI Innovation at the Minsky Awards 2024. Known for its groundbreaking advancements in AI and deep learning, NVIDIA has been at the forefront of driving innovation across various industries, including gaming, autonomous vehicles, and healthcare. The company’s Global Capability Centers (GCCs) have played a pivotal role in developing cutting-edge AI solutions, enabling industries to harness the power of AI for better decision-making, operational efficiency, and product innovation.","excerpt":"The Minsky Awards for Excellence in AI 2024 celebrated the most innovative companies, recognizing their groundbreaking contributions across sectors in artificial intelligence.","categories":["AI Highlights"],"tags":["AIM"],"author_name":"Дарья","publish_date":"2024-09-30T15:19:04","publication_year":"2024","word_count":1383,"keywords":["data science","artificial intelligence","AI","ML","RAG","Aim","deep learning","analytics","edge AI","AIM","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","ML","deep learning","data science","analytics","Aim","edge AI","RAG","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/minsky-awards-for-excellence-in-ai-2024-meet-the-winners\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":60989,"title":"IIT Delhi Allocates Supercomputer Resources For COVID-19 Research To Merit-Based Proposals","content":"With the wake of COVID-19 spread in the nation, The Indian Institute of Technology, Delhi has decided to invite proposals from government and private academic institutions as well as private companies to use its supercomputer resource PADUM for COVID-19 research. According to their release, IIT Delhi is allocating its supercomputer resources for COVID-19 research in association with an academic partner from all over the country. In this, the allocation of supercomputer resources to merit-based proposals, according to the institute, will be made available for three months and a total of Rs 1 Cr worth of computational time will be provided to them with no cost. Each proposal will have a maximum cap of Rs 10 lakh worth of computational resource. Speaking about this initiative by the institute, Prof V Ramgopal Rao, the director of IIT Delhi said, “In these difficult times, sharing of resources is important in order to address the infrastructure requirements of researchers working on the Corona epidemic. IIT Delhi has taken a principled stand and wishes to set an example for this. Scientists need to collaborate with each other, given the urgency of the situation.” He further stated that initially, it was located for three months, but now the allocation period can be extended to six months after evaluating the performance of the projects. Also, in a recent LinkedIn post, Rao has shared the release stating – “IIT Delhi opens up its supercomputer facility for COVID researchers.” The institute has urged researchers to submit their proposals by the 15th of this month, where experts from IIT Delhi will evaluate them on a first come first serve basis. Further explaining the process, the institute stated in their release that, after the election of the proposals, the chosen proposals would be provided with basic and limited support, which will include instructions on job submissions or smooth functioning of these projects. Details available here. Along with that, IIT Delhi is also welcoming all researchers from India to use the supercomputer of the institute for COVID-19 research on a payment basis. It is further stated that “IIT Delhi will match to double the amount contributed for high-performance computing usage for COVID-19 research.” In this payment basis mode, IIT Delhi is allocating a budget of Rs 5 crore worth of high-performance computational resources for the next six months.","excerpt":"With the wake of COVID-19 spread in the nation, The Indian Institute of Technology, Delhi has decided to invite proposals from government and private academic institutions as well as private companies to use its supercomputer resource PADUM for COVID-19 research.  According to their release, IIT Delhi is allocating its supercomputer resources for COVID-19 research in […]","categories":["AI News"],"tags":["Coronavirus","covid-19","iiit delhi","IIT Delhi","Supercomputers","Supercomputers India"],"author_name":"Sejuti Das","publish_date":"2020-04-06T10:30:14","publication_year":"2020","word_count":386,"keywords":["Go","covid-19","AI","programming_languages:R","programming_languages:Go","Coronavirus","Supercomputers India","Supercomputers","ViT","iiit delhi","IIT Delhi","R"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-delhi-allocates-supercomputer-resources-for-covid-19-research-to-merit-based-proposals\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":46672,"title":"AI Helps Scientists Predict Depression Outcomes In New Study","content":"The psychiatry field has long sought answers to explain why antidepressants help only some people. Is a patient’s recovery due merely to a placebo effect – the self-fulfilling belief that a treatment will work – or can the biology of the person influence the outcome? Two studies led by UT Southwestern provide evidence for the impact of biology by using artificial intelligence to identify patterns of brain activity that make people less responsive to certain antidepressants. Put simply, scientists showed they can use imaging of a patient’s brain to decide whether a medication is likely to be effective. The studies include the latest findings from a large national trial (EMBARC) intended to establish biology-based, objective strategies to remedy mood disorders and minimize the trial and error of prescribing treatments. If successful, scientists envision using a battery of tests such as brain imaging and blood analyses to increase the odds of finding the right treatment. “We need to end the guessing game and find objective measures for prescribing interventions that will work,” said Dr Madhukar Trivedi, who oversees EMBARC and is founding Director of UT Southwestern’s Center for Depression Research and Clinical Care. “People with depression already suffer from hopelessness, and the problem can become worse if they take a medication that is ineffective.” In this study, scientists used AI to examine neural activity throughout the brain while study participants processed emotions. The studies – which each included more than 300 participants – used imaging to examine brain activity in both a resting state and during the processing of emotions. Both studies divided the participants into a healthy control group and people with depression who either received antidepressants or placebo. Of the participants who received medication, researchers found correlations between how the brain is wired and whether a participant was likely to improve within two months of taking an antidepressant. Dr Trivedi said imaging the brain’s activity in various states was important to get a more accurate picture of how depression manifests in a particular patient. For some people, he said, the more relevant data will come from their brains’ resting state, while in others the emotional processing will be a critical component and a better predictor for whether an antidepressant will work. “Depression is a complex disease that affects people in different ways,” he said. “Much like technology can identify us through fingerprints and facial scans, these studies show we can use imaging to identify specific signatures of depression in people.”","excerpt":"The psychiatry field has long sought answers to explain why antidepressants help only some people. Is a patient’s recovery due merely to a placebo effect – the self-fulfilling belief that a treatment will work – or can the biology of the person influence the outcome? Two studies led by UT Southwestern provide evidence for the […]","categories":["AI News"],"tags":["Depression"],"author_name":"Prajakta Hebbar","publish_date":"2019-10-01T13:47:29","publication_year":"2019","word_count":411,"keywords":["artificial intelligence","programming_languages:R","AI","Depression","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-helps-scientists-predict-depression-outcomes-in-new-study\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10167236,"title":"Chinese Firms Including ByteDance, Alibaba Place $16 Bn NVIDIA GPU Orders: Reports","content":"Chinese big tech companies, such as ByteDance, Alibaba, and Tencent, have collectively placed orders worth $16 billion for NVIDIA’s H20 GPUs, The Information has reported. The increased demand is likely due to the thriving AI ecosystem in China, with all the major players developing foundational models, most of which are available as open source. Models from Alibaba (Qwen) and DeepSeek AI have consistently demonstrated performance comparable to AI models developed in the United States. China represents a challenging market for NVIDIA to export GPUs because of restrictions imposed by the United States government. After the US implemented export restrictions on NVIDIA’s high-end GPUs, the company introduced a less powerful version of its Hopper GPU, the H20, to legally export it to China. However, Bloomberg reported in January that officials from President Donald Trump’s administration are also in discussions to restrict the export of H2O chips. Therefore, if NVIDIA were to ship the chips to China, it would need to do so before any further restrictions come into effect. A report from Reuters last month revealed that Chinese companies are increasing their orders for NVIDIA’s H20 chip to satisfy the growing demand for DeepSeek’s affordable models. Furthermore, as per a report last December, Omdia estimated that ByteDance and Tencent each ordered approximately 2,30,000 NVIDIA chips for 2024. It was also reported that DeepSeek possess around 50,000 NVIDIA GPUs. The Information also revealed that NVIDIA generated $17 billion in sales from China in the 12 months ending January 26, accounting for 13% of its total revenue. Previously, NVIDIA shipped the H800 GPU, a less powerful version of the H100 GPU. The H100 offers a transfer rate of 600 gigabytes per second, compared to the H800’s 300 gigabytes per second. Soon after, the US also banned the export of NVIDIA’s H800 to China, preventing the company from selling chips even at a reduced transfer rate.","excerpt":"NVIDIA reportedly generated $17 billion in sales from China in the 12 months ending January 26.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","China","NVIDIA"],"author_name":"Supreeth Koundinya","publish_date":"2025-04-03T17:53:43","publication_year":"2025","word_count":312,"keywords":["Go","programming_languages:R","AI","AI (Artificial Intelligence)","programming_languages:Go","NVIDIA","R","China"],"extracted_tech_keywords":["AI","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/chinese-firms-including-bytedance-alibaba-place-16-bn-nvidia-gpu-orders-reports\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":7444,"title":"Appnomic® Systems – Simplifying and Automating IT","content":"Appnomic Systems is bringing “Big Data” advanced analytics to the Information Technology (IT) global market and achieving business results breakthroughs with our clients and partners.   As an early member of the analytics industry and community of software product companies emerging from India, I am pleased to share our story with India Analytics Magazine readers. Appnomic Systems’ business is to simplify and automate IT through software product and related services.  The company has developed a new, recently patented approach to managing complex IT operations and is leading the way in a new sector of the $22+ billion IT Operations Management (ITOM) market.  This new sector, called IT Operations Analytics (ITOA), was named just a couple years ago by an industry analyst group.  As the industry evolves, so are Appnomic’s solutions and the market size.  Applications of Appnomic solutions in the broader Advanced Analytics, Business Activity Monitoring, and Business Process Management markets are on the horizon for the company as well. ITOA offers an improvement to the 10+ year-old, $2+ billion, application performance monitoring and management (APM) market.  ITOA is changing the game in this category and Appnomic is doing it with patented Application Behavior Learning (ABL) technology incorporated in our flagship software product, AppsOne®. Today, APM and other similar IT monitoring tools are the equivalent of fire extinguishers that put out “IT fires” once they occur, whereas Appnomic ABL technology is like a smoke detector that prevents fires from occurring in the first place.  Many people use fire extinguishers to put out fires, smoke detectors do a better job to prevent fires with a completely different approach.  This analogy holds true for the new IT operations analytics that are changing how we prevent IT fires and fire drills versus fix things after an incident has occurred. Our customers are some of the largest banks, ecommerce companies, manufacturers, and enterprises in the world.  They use Appnomic software and services to achieve enormous gains in sales and profitability.  A case study of a bank that is delivering 2 second transaction response times across their Internal banking portal is available here for review. Clients achieve compelling results with increased application up-time and responsiveness to end users; accelerated IT application & infrastructure trouble resolution; prevention of IT “accidents” that otherwise impair end user experience and ecommerce sales conversions; compliance enforcement & risk mitigation; migrating their data center operations to the Cloud; selling value added IT services; cost savings from using our tool set versus legacy providers’ tools, and more. In addition to our ITOA solution, AppsOne, we also have a contemporary IT Process Automation (ITPA) platform, OpsOne®.  OpsOne helps automate the IT fixes of poor performing enterprise software \/ web applications as well as repetitive data center and IT management processes. Together, AppsOne and OpsOne deliver actionable analytics like no other solution in the market. The company’s global headquarters are in Bangalore, India and U.S. headquarters are in Sunnyvale, California in the center of Silicon Valley.  You may visit us online at www.appnomic.com .","excerpt":"Appnomic Systems is bringing “Big Data” advanced analytics to the Information Technology (IT) global market and achieving business results breakthroughs with our clients and partners.   As an early member of the analytics industry and community of software product companies emerging from India, I am pleased to share our story with India Analytics Magazine readers. Appnomic […]","categories":["Deep Tech"],"tags":["Analytics Case Study"],"author_name":"D Padmanabhan (Paddy)","publish_date":"2015-05-22T14:54:00","publication_year":"2015","word_count":499,"keywords":["big data","Go","programming_languages:R","AI","programming_languages:Go","automation","ViT","analytics","Analytics Case Study","R"],"extracted_tech_keywords":["AI","analytics","R","Go","big data","ViT","automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/appnomic-systems-simplifying-and-automating-it\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10019927,"title":"How Andy Jassy Will Lead Amazon’s AI Strategy?","content":"In an unexpected turn of events, Jeff Bezos has announced that he is stepping down as the CEO of Amazon, passing the baton to the leading man of AWS, Andy Jassy. The move ties in with last year’s Big Tech hearing when Bezos, along with Tim Cook, Mark Zuckerberg, and Sundar Pichai, was grilled by the Congress over their respective companies’ alleged monopolistic practises. Bezos has decided to move to an executive role, handing Amazon’s reins to the next best in the line. Andy Jassy is the man who built Amazon’s cloud business from scratch, making it the most profitable technology company in the world. Joined in 1997 as a deputy of Bezos, Jassy became the head of AWS division in 2016, which currently accounts for most of Amazon’s revenue. A recent report shows AWS dominates the cloud market with an estimated $20 billion in IaaS cloud revenue in 2019, leading with a 45% market share. The numbers tell the story. No wonder Bezos puts a lot of stock in Jassy’s ability to get the job done. Analysts have also acknowledged AWS’s capability to incorporate emerging technologies in cloud-hosted applications. The parent company, Amazon, has been an early adopter of artificial intelligence. The company is leveraging the technology to enhance its customer experience, product recommendation engine and improve internal operations — all under the able guidance of Bezos. Jassy has big shoes to fill. His leadership will be tested, especially on the AI strategy front. Also Read: Top 10 Quotes By Andy Jassy At AWS re:Invent 2020 AI & ML Pillars With AWS, Jassy has made big bets on artificial intelligence and machine learning. AWS has plans to democratise machine learning for cloud applications, making it accessible to all and sundry. The idea is to allow developers to deploy machine learning models and manage its different layers without building it from scratch. AWS SageMaker is a good case in point. “We have regularly added new capabilities to it, including Amazon SageMaker Ground Truth to build highly accurate annotated training datasets and SageMaker RL to help developers use a powerful training technique called reinforcement learning,” said Denis Batalov, the Worldwide Technical Leader, ML and AI at AWS, in an interview with Analytics India Magazine. “These capabilities have helped many more developers build custom machine learning models.” Last year’s re:Invent event has also seen many announcements showcasing AWS’ AI and ML prowess — starting from launching its second custom machine learning chip and designing new AWS service to speed up data preparation to releasing a natural language query feature in Amazon QuickSight, announcing DevOps Guru, a fully managed operations service, etc. In all these instances, Jassy stressed the importance of AI and ML for the current era. Jassy said all ML experts and AI practitioners get hired in big tech companies, and the low-key enterprises and startups tend to miss out on the opportunity. This is why AWS has been working on building services and capabilities to enable more companies to leverage such technologies, stated Jassy, in last year’s Goldman Sachs Technology and Internet Conference. If history is any indication, Jassy will double down on democratising AI and ML technologies with more solutions like SageMaker. All That Jazz Be that as it may, Jassy’s leadership raises many concerns for the industry. Last year, Amazon decided to ban its facial recognition technology; however, Jassy, defended the technology, calling the circumstances premature. In many instances, he has shown his support towards developing autonomous weapons for national security. His moral compass is suspect. As Jassy once told the media, people are looking to leverage machine learning to get their job done; however, they don’t give much thought about what the company is doing behind the scenes. The face recognition tech is especially problematic. Nicole Ozer, technology and civil liberties director with the American Civil Liberties Union of Northern California, said, “Amazon must fully commit to a blanket moratorium on law enforcement use of face recognition until the dangers can be fully addressed, and it must press Congress and legislatures across the country to do the same.” In addition to this, Google’s former AI researcher, Timnit Gebru and other AI Now Institute researchers have worked on a Gender Shades project that massively condemned facial recognition from companies like Amazon, IBM, Microsoft, etc. This research forced the companies to question their practices; however, Jassy dismissed the argument and was willing to sell their controversial software to any country, where FRT is legal. Challenges Ahead Jassy seems to favour facial recognition technology, and he might even bring back Amazon’s Rekognition to the market once the ban gets lifted. Jassy might also push its smart home division — Ring, to law enforcement, which has been quite unpopular due to its surveillance aspect. Though it’s still early days to make such assumptions, it indeed will be interesting to see Jassy’s AI strategy for Amazon. The question on everyone’s lips is, can Andy further Amazon’s lead in the AI race?","excerpt":"In an unexpected turn of events, Jeff Bezos has announced that he is stepping down as the CEO of Amazon, passing the baton to the leading man of AWS, Andy Jassy. The move ties in with last year’s Big Tech hearing when Bezos, along with Tim Cook, Mark Zuckerberg, and Sundar Pichai, was grilled by […]","categories":["Global Tech"],"tags":["Andy Jassy","AWS Andy Jassy"],"author_name":"Sejuti Das","publish_date":"2021-02-10T13:00:00","publication_year":"2021","word_count":829,"keywords":["Go","Amazon SageMaker","machine learning","artificial intelligence","AWS","AI","ML","RAG","analytics","AWS Andy Jassy","R","Andy Jassy"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Amazon SageMaker","RAG","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-andy-jassy-will-lead-amazons-ai-strategy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":67937,"title":"How The Freelance Boom Will Impact The Position Of In-House Data Scientist","content":"The COVID pandemic has brought in a significant culture change for organisations where employers are relying on remote working for their business continuity. And, as companies are getting comfortable with the remote working ecosystem, many leaders have started hiring freelance data scientists or gig workers for their companies. The profession of data science comes with the advantage of having the ability to code, program, and solve business problems remotely. This, in turn, has created a freelance boom in the data science ecosystem, where companies are preferring to hire more freelance data scientists over in-house full-time core data science teams. A lot of this could be attributed to the flexibility option it provides to the data professionals as well as the best return on investment for companies, who are planning to cut costs amid this crisis. In fact, according to industry estimates, the Indian freelancers market size is going to grow to a whopping $20-30 billion by 2025, and a lot of this could be attributed to the pandemic outbreak. Not only is such a freelance boom beneficial for companies but also advantageous for professionals who have started freelancing to keep their relevance in the industry, post a layoff. Freelance boom impacts the hiring of in house data scientists Data science ecosystem comes with the advantage of working remotely as well as the flexibility of programming codes from any part with the help of several tools that are already available in the market. Alongside many collaboration and employee management tools, it is becoming easier for business leaders to manage their freelance and contract-based data scientists. GitHub, GitLab and Bitbucket are a few of the source control options that can help employers among their freelancers’ work. Therefore, in order to have business continuity amid this crisis with fewer cost expenses, hiring gig workers and freelance data professionals play a significant advantage for the current era. In fact, according to the vice president of enterprise at Toptal, a platform that connects businesses with freelance workers, including software engineers and data professionals — Michael Kearns, stated that this openness to remote working culture will demand more freelance work. Also, he believes that this will urge remote teams to learn how to do it right; so, freelancers will help the work transition. Also Read: How To Get The Best Out Of Your Freelance Data Scientists Alongside, freelancers come with a lot of benefits for the company, which can replace in house data scientists, like providing easy access to a wide range of talent pool without the concerns of geographical boundaries. Also, many freelance data scientists have domain specialisation like machine learning, financial analytics, retail analytics, data warehousing, to name a few, which allow companies to hire specific data professionals for one particular type of work and concentrate on the one-off project at a time. Considering freelance data scientists work on many projects at a time for different clients, they usually have the best time management strategies, which in turn will benefit the company with meeting deadlines faster. Also, with more flexible schedules, it has been proved that freelance data professionals are more motivated to work and finish business projects than regular employees. The president of GreatBizTools, a company that specialises in designing, developing, and implementing human resource management systems for businesses stated in her blog post that, freelancers and contract workers have stated that they are more motivated to work because “they are doing exactly what they want to do. They choose the work they do so in practice, it should be rewarding and enjoyable. If the work becomes less satisfying, they can move on to the next gig. Full-time employees sometimes perform work that does not match their passions.” Also, as businesses require data scientist only for solving business problems and working on a specific type of projects, it’s usually advantageous for employers to go for hiring a gig worker than undergoing the complex process of recruiting in house data scientists for their core team, this is where the contract-based workers come into the picture who work with several companies and don’t demand additional benefits like regular employees. Also Read: How COVID-19 Has Impacted Freelancers And What They Can Do To Survive? Additionally, gig workers come with considerable cost reduction benefits, where data professionals who are working on a contract based usually turn out to be cheaper for companies. In fact, experts believe that a freelance developer can save up to 50% of the company’s cost of hiring a full-time employee. On the other hand, freelance employers especially the ones working with critical data also come with a few disadvantages with leaders having no direct supervision on them, which sometimes can bring in a lag in response time. Also, with no geographical boundaries to work with freelance data scientists, there can also be a time difference in meeting deadlines, which can again hamper businesses’ bottom line. Plus these gig workers come with security issues as they work with critical data. Many of these data scientists can create vulnerability for sensitive company data. And therefore, businesses must assess these issues in order to have an efficient freelance data science team. To support this gig economy, many companies have created platforms on the internet to connect businesses with freelance data scientists, such as Total, Fiverr, Upwork, Freelancer, to name a few. Apace with that, Experfy and Sage Legion are two other platforms that are specifically designed for analytics professionals. In a recent interview, Marlon Rosenzweig, the co-founder & CEO at Workgenius stated to the media that, “Companies realise that being on the payroll and being in-office for certain functions is not necessary, and I think now they have started to sort of embrace the freelance world a little bit more.” He further stated, “Then they turn to a freelance staffing company … that really provides the technology to make that experience seamless, because if you are trying to do manual freelancing as a company, you incur so much overhead.” Also Read: How To Work As A Freelancer In The Field Of Data Science In fact, according to the news, countries like France, the UK and the US have already started preparing guidelines and frameworks to accommodate rising freelancer populations amid this crisis. The countries are creating laws that can involve gig workers, freelancers, and independent contractors with in-house employees, which would include not only payment guidelines but also employee benefits guidelines. The US has recently launched a $2 trillion federal stimulus package — CARES Act to support the freelancers. Also, France has created a strategy to provide grants to its freelance workers. With such pointers in hand it can easily be stated that in order to thrive in this oncoming economic downturn, companies must adapt to necessary change and carefully consider the option of deploying more freelancers instead of full-time data scientists. Wrapping up As the COVID pandemic is still unravelling itself, it continues to have a ripple effect on professionals and how businesses are working with their employees. With a massive amount of collaboration tools in the market, it is becoming increasingly easy for employers to hire and manage freelance data scientists. This, in turn, is becoming their preferred choice amid this crisis. Although freelancers come with the advantage of flexible working and bringing a lot of benefits for companies who are working in a tighter budget, this freelance boom is also bringing in a significant competition in the data science ecosystem, which needs to be addressed to have long term impacts.","excerpt":"The COVID pandemic has brought in a significant culture change for organisations where employers are relying on remote working for their business continuity. And, as companies are getting comfortable with the remote working ecosystem, many leaders have started hiring freelance data scientists or gig workers for their companies. The profession of data science comes with […]","categories":["AI Features"],"tags":["data scientist qualifications","Freelance","Freelance data scientists","freelancer","is the tech boom over"],"author_name":"Sejuti Das","publish_date":"2020-06-23T17:00:00","publication_year":"2020","word_count":1247,"keywords":["data science","Go","machine learning","data scientist qualifications","AI","AWS","Freelance","ML","Git","Freelance data scientists","analytics","is the tech boom over","GitHub","R","freelancer"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","AWS","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-the-freelance-boom-will-impact-the-position-of-in-house-data-scientist\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":58421,"title":"What Makes Cloud Security Challenging For Cybersecurity Teams?","content":"Cybersecurity still seems to be the number one concern when it comes to cloud adoption. About 85% of enterprises are estimated to have the majority of their workloads on the cloud, be it public, hybrid or multi-cloud platforms by the end of 2020, according to a report. On the other hand, cybersecurity still seems to be the number one concern when it comes to cloud adoption. So what does cloud adoption mean for security teams when there is an architectural shift that’s going on with hybrid to multi-cloud strategies? Here we look at the challenges on cloud security: Security teams are accountable because your brand is still the one that’s in the newspaper if something goes wrong. You have to have a constant conversation with the cloud services provider regarding handling an incident. Whose job is it to secure the data and applications that are being deployed on the cloud? Businesses are naturally concerned whether SaaS vendors will take ownership of migrating their own workloads to the cloud. There is also a shift to add software increasingly, and application developers are accessing resources and other cloud services through APIs. So for businesses, one of the single biggest challenges (and priority) is to start moving security more towards the development function by embedding security into DevOps, something called DevSecOps. This can help cybersecurity managers ensure that before they deploy software in the cloud,  it is scanned and does not contain software vulnerabilities. Cloud & Cybersecurity: Containerisation & Microservices Escape Traditional Firewalls Another thing that is happening fast is the move towards containerisation. As containerised applications are starting to occur at a very rapid pace, experts say the fundamental truth about containers is that traditional cybersecurity tools are blind to security managers. Containers network traffic doesn’t even go through conventional firewalls in many cases. Containers and microservices are transforming cloud, but when it comes to analysing traffic via firewalls, they are very challenging. The transition from monolithic applications to container-based microservices brings many advantages but also creates new challenges for security teams. Next-Generation Firewalls (NGFW) were designated to handle the latest threats and data centre architectures but fell short in the cloud microservices evolution as they were designed to act as a gateway for north-south traffic. According to experts, this needs to change, and there is a need to move to container firewalls. A container firewall is a policy-based, declarative network security model used on platforms like Kubernetes to safeguard ingress and egress traffic. Container firewall is built for the cloud-native environment, which can monitor traffic moving in all directions, including through container and non-container network layers. What About Open Source Software Security On Cloud? The premise of moving to the cloud has already baked in the premise that businesses will be using an enormous amount of open-source software. The cloud is mostly built on an open-source foundation, notwithstanding licensing challenges back and forth. The rapid adoption of containers and a lot of development on the cloud is open-source, and cybersecurity managers are not comfortable with that. There are certain open-source technologies that every company uses, but naturally open-source may also raise alarm bells because there are new types of attacks that can spread more quickly via open-source. On the other hand, there are open source success stories and billion-dollar companies coming out of the space in most segments of technology. So, not using open-source is not an option. Instead, experts say there is a need for a more careful approach to creating rules around how to recognise when attackers can infiltrate or access a company’s network via open-source vulnerabilities. Also, reinventing open-source policies over and over again can make open-source strength to be leveraged rather than something to be worried about.","excerpt":"Cybersecurity still seems to be the number one concern when it comes to cloud adoption.  About 85% of enterprises are estimated to have the majority of their workloads on the cloud, be it public, hybrid or multi-cloud platforms by the end of 2020, according to a report. On the other hand, cybersecurity still seems to […]","categories":["AI Trends"],"tags":["Cloud Computing","Cyber Security","Cybersecurity","devop","devops journal","International Affairs","network firewall","Security"],"author_name":"Vishal Chawla","publish_date":"2020-03-11T18:00:00","publication_year":"2020","word_count":620,"keywords":["devops journal","network firewall","Go","Cyber Security","API","programming_languages:R","AI","R","International Affairs","Security","programming_languages:Go","RAG","microservices","Cloud Computing","Cybersecurity","DevOps","devop","kubernetes"],"extracted_tech_keywords":["AI","RAG","kubernetes","microservices","R","Go","DevOps","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/what-makes-cloud-security-challenging-for-cybersecurity-teams\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":34599,"title":"As India Becomes The Hub For AI CoEs, Is Indian Talent Primed To Lead Onshore AI Teams","content":"Image Source: Leadership Management The Indian IT industry has always been a goldmine for international tech giants. With the country being at the cusp of AI crossroad and with key industry players looking at means to integrate technology in their business processes, India has emerged as rich playing field for AI research and development. With emerging markets like India forecasted to contribute up to 55% to global GDP, the latest trend among tech giants has been to establish their Centres of Excellence (CoE) in India. This has been helping leading companies, even the ones which are outside the technology domain to be a part of the next wave of the growth by understanding its complex market, talent and evolving regulatory environment. Some of the biggest names from technology domain include companies like IBM India, Adobe Microsoft and Google, that have set up their CoEs and have already partnered with numerous government and private agencies to drive the next wave of growth in the industry. Recently, there has been a spurt in companies outside the tech domain that are now setting up their technology labs in India. The latest players to set up their tech arms include American technology provider for companies, Ecolab and Latin-American departmental store chain, Falabella. Previously, these companies looked for international talent to manage and oversee the operations in India. But lately, there has been a marked reverse in this trend with the appointment of Indian talent to monitor the development of hubs\/technology arms, dedicated to the development of emerging technologies. Indian Tech Leaders Now At The Forefront of AI Wave Are Scaling Up Analytics & AI Operations A number of Indian companies and their multinationals counterparts have started to rely on their internal pool of Indian talent to lead the teams in India and other regions. Increasingly, as more and more offices of companies continue to mushroom across the globe, regional players are now tapping into Indian expertise to scale-up their operations with their international expansion. While their responsibilities range from CEO to principal data scientist depending on the organisation, over the years more data scientists have been breaking away from traditional career path to embrace diverse leadership roles. Another interesting trend regarding the industry is the growing number of reverse migration of the techies to India. With the recent protectionism rules kicking in, high potential Indian tech talent and data scientists are now relocating to India. Why Indian Data Scientist Are Fit To Lead The International Market With premium institutes in the country equipping graduates with in-demand skills, even freshers come with baseline knowledge and are able to quickly upskill. Even though the Indian educational system lags behind in providing practical knowledge, young data scientist professionals have quickly adapted to the industry standards with great agility. We list down key reasons why Indian technologists and AI professionals are primed to lead AI hubs in the global market: India is a base for experienced DS professionals: Given the availability of immense talent pool within India, in coming years we will see more Indian talent leading data teams internationally and at home. With the future looking very promising for emerging technologies, more young Indian data scientist will emerge as top-tech leaders.A key advantage that they pose is that this talent base boasts of considerable industry experience. According to a study released by Analytics India Magazine last year, the average work experience of analytics professionals in India was 7.9 years which is up from 7.7 years from 2017. The country also has more than 28,000 Analytics professionals with more than 10 years experience. India is the third largest developer market:  “India is soon going to be the third biggest developer market in India as a lot of investment is happening learning, so data science as a stream is picking up and there is a lot of innovations happening. To sum up, there is a lot of positive outlook towards the industry,” said Sandeep Alur, Partner Technology Engagement, Microsoft India, at MLDS 2019. Availability of ML talent in India: “Data science is no longer about sitting there and coding in Python. It is much more than knowing traditional coding sets rather it is about knowing about the optimization of large data,” explained Madalasa Venkataraman, Chief Data Scientist at TEG Analytics. Vibrant ecosystem: Ravinder K Sharma of ABInBev who leads the company’s Growth Analytics Centre & Global Capability Centre in Bangalore told AIM.  “Given the demand and scale that we have as a company, we are entering new gaps and building up a strong partnership in India and we came to India to invest in this ecosystem and we are tremendously benefiting from it. We are very bullish on enabling AbInBeV’s dream through our collaboration with India and most importantly, our dream is to export data science professionals from India to democratise the talent”.","excerpt":"The Indian IT industry has always been a goldmine for international tech giants. With the country being at the cusp of AI crossroad and with key industry players looking at means to integrate technology in their business processes, India has emerged as rich playing field for AI research and development. With emerging markets like India […]","categories":["AI Features"],"tags":["Data Scientist"],"author_name":"Akshaya Asokan","publish_date":"2019-02-07T06:12:00","publication_year":"2019","word_count":804,"keywords":["data science","Go","AI","ML","RAG","Colab","Python","Aim","analytics","Data Scientist","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","Aim","Colab","RAG","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/do-indian-data-scientist-fit-the-bill-for-offshore-team-leading\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10098586,"title":"Data Science Hiring and Interview Process at Dream11","content":"Dream Sports, the parent company of Dream11, is a prominent sports tech unicorn in India that houses a bunch of brands like Dream11, FanCode, and more, making it the Willy Wonka’s Chocolate Factory of sports engagement. Founded in 2008 by Harsh Jain and Bhavit Sheth, Dream Sports is headquartered exclusively in Mumbai, affectionately dubbed “The Stadium” by its team. At the forefront, its flagship fantasy sports platform, Dream11, accommodates a massive user base of over 190 million individuals. This platform offers users the exciting opportunity to participate in fantasy versions of cricket, hockey, football, kabaddi, handball, basketball, volleyball, rugby, futsal, American football, and baseball. And for the seamless operation of Dream Sports, data science has become a cornerstone in tackling fundamental challenges for platforms like Dream11, where user experience and engagement play pivotal roles. AIM got in touch with Amit Sharma, chief technology officer at Dream Sports (Dream11) to know more about their AI, ML operations, hiring strategy, work culture and more. Since 2016, Sharma has led the creation of the sports tech platform. Earlier, he spent over a decade developing complex distributed systems for major companies like Yahoo! and Netflix in California. Dream Sports is currently looking out for a VP of Data Science in Mumbai to lead data science roadmap development, drive experimentation, propose ML solutions, mentor the team, and ensure goal alignment. Required skills include programming languages, data visualisation, machine learning, and strong interpersonal skills. Familiarity with technologies like Cloudfront, API Gateway, Python, MySQL, Kafka, Spark, and Redshift is essential. Inside Dream11’s AI & Analytics Play Aligned with its mission of enhancing the sports experience, Dream Sports operates akin to a “high-performing sports team” composed of “Coaches” (CXOs) and “Captains” (team leaders) who guide over 1000 “Sportans” (employees). That includes a skilled roster of engineers, business and data analysts, applied scientists and machine learning experts. One of the primary hurdles that this team has successfully tackled through data science is the issue of personalisation, discovery, and fraud detection within its application. When it comes to user interaction, the app encompasses a multitude of features, like matches and contests, resulting in a notable mental demand. These elements are subject to change, creating challenges for users to fully grasp and decide on the most suitable choices. Dream11 was an early adopter of data science in its product development journey, even during when these technologies were relatively nascent within the tech ecosystem. “In order to improve user experience and engagement, we have deployed over 100 AI and ML models to enable the contextual discovery of relevant features. We have personalised multiple user journeys throughout the app by analysing user cohorts and behaviours,” said Sharma. Underpinning these offerings are sophisticated ML systems that process an extensive array of over 1000 features across numerous users. Appropriate models and data drifts have also been implemented to ensure the smooth functioning of daily operations. The team of data scientists has also developed robust systems for detecting fraud by incorporating knowledge graphs, extensive searches for similarities, and algorithms for recognising patterns. They consistently carry out tests and refine methods for personalisation. They use A\/B testing to compare various user experiences and gauge their effects on user engagement. Data scientists at Dream11 continually experiment with various personalisation strategies, employing A\/B testing methodologies to compare distinct user experiences. This iterative approach helps in identifying the most effective strategies while quantifying their impact on user engagement. Tech Stack Dream Sports’ technological infrastructure is a blend of in-house tools and third-party solutions. “While our tech infrastructure is a combination of in-house tools and third-party solutions, we strongly believe in developing in-house solutions to minimise costs, strengthen data privacy, and control the scalability of services without compromising the architecture,” said Sharma. One of their in-house frameworks, known as FENCE (Fairplay Ensuring Network Chain Entity), is employed to identify and address Fairplay violations, ensuring fair competition for users. Their primary distributed ML systems rely on Spark and Ray. Transformers are utilised for various sequential learning tasks and have demonstrated superior performance compared to other deep learning models on their datasets, which are awaiting large-scale implementation. “We are exploring applications of LLMs in the context of the Sports ecosystem and testing internal prototypes,” Sharma commented. For forecasting and classical machine learning applications, they rely on a range of resources, including Scikit Learn, XGboost, Prophet, and Scipy. Additionally, for deep learning-based machine learning tasks, the team leverages the capabilities of Pytorch and Tensorflow, harnessing their power to create robust and advanced models. On the front of app development, Dream11 became one of the few tech companies to fully migrate their platform to React Native, a UI software framework. Despite industry scepticism regarding the feasibility of complete React Native adoption due to its historically low success rates, Dream11 navigated and overcame the associated challenges to make this happen. Interview Process Dream11 places a strong emphasis on valuing skills when hiring, seeking top talent aligned with their goal of enhancing the sports experience, encapsulated by the acronym “DOPUT”: Data-Driven, Ownership, Performance, UserFirst, and Transparency. When hiring data science professionals, the organisation prioritises cultural fit initially. Upon meeting this criterion, candidates undergo a customised hiring process that varies by role. For most data science applicants, this process includes an aptitude test, followed by progressive technical interviews covering areas such as R programming, ML, and practical mock projects. Domain-specific interviews led by team leaders provide a thorough evaluation of the candidate’s preferences and skills. One of the common mistakes that candidates make while interviewing is sometimes they miss out on the basic foundations of ML, stats or experimentation which are extremely valued at the organisation. “While building models is easy, making them useful is tougher. That’s where hands-on implementations become a force multiplier,” said the CTO. Prospective team members can expect a supportive work environment with access to extensive qualitative data, challenging machine learning tasks, advanced infrastructure, a motivated team, and the chance to contribute at a large scale. Work Culture Driven by its culture, the company fosters an open and transparent atmosphere. Certified as a Great Place to Work, Dream Sports adopts a hyper-experimentation approach known as “HEAL – Hypothesis, Experiment, Analysis, and Learning” allowing employees to experiment, embrace failure, swiftly learn, and create personalised user features. Employees enjoy a range of special perks and benefits like ‘Learning Wallet‘, unlimited leaves, ESOP, insurance, mental wellness initiatives and more. The Learning Wallet supports diverse learning ambitions, allowing individuals to explore areas such as design or coding regardless of their primary expertise. The unlimited leave policy promotes a healthier work-life balance, including the ‘Unplugged’ feature—a unique seven-day work-free vacation opportunity. Additionally, employees enjoy fully-paid access to sports events, matches, and tournaments. “Most importantly, besides several industry-first benefits, we offer access to the latest tech stack and prioritise building a thriving culture through various engagement activities,” concluded Sharma. Check out their careers page here. Read more: Data Science Hiring Process at Naukri.com","excerpt":"Dream Sports, the parent company of Dream11, is currently looking out for VP of Data Science in Mumbai.","categories":["AI Hirings"],"tags":["Career","Data Science Hiring","Top Trend"],"author_name":"Shritama Saha","publish_date":"2024-07-29T15:32:10","publication_year":"2024","word_count":1153,"keywords":["Top Trend","data science","machine learning","AI","PyTorch","ML","Data Science Hiring","Aim","deep learning","Ray","analytics","TensorFlow","Career"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","Aim","Ray","TensorFlow","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/data-science-hiring-process-at-dream11\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10122034,"title":"After PAN &amp; Aadhaar, Protean Now Leverages AI for Citizen-Centric Solutions","content":"Protean eGov Technologies, a provider of e-governance solutions, has been instrumental in facilitating critical platforms such as the PAN card, Tax Information Network (TIN), the Pension Fund Regulatory and Development Authority (PRAN) and Aadhaar to release eSign and eKYC. The company now sees immense potential for AI-powered infrastructure across domains like healthcare, education, agriculture, and governance. It is developing AI-powered chatbots that can communicate in local languages, besides promoting inclusion and accessibility to government services. It is also building AI models for fraud detection. Targeting Startups, SMEs, and Government Agencies Protean is targeting startups, SMEs, and government agencies with its “very sovereign data cloud offering”. Without naming specific clients, Metesh Bhati, CDO of Protean eGov in conversation with AIM, mentioned that a few government agencies are already using Protean’s cloud services, with some of their existing data centres migrating to Protean cloud. The clientele includes both public and private sector organisations seeking reliable and secure cloud solutions. “Our vision is to align with People+AI to provide advanced machine learning tools and support innovative projects in smart cities and predictive analytics within the public sector,” said Bhati. Protean has given in-principle approval to partner with Open Cloud Compute (OCC), an offering by People Plus AI, to design an open network from the supply side. “That’s where we are saying we’ll be offering the AI infrastructure,” Bhati added. Protean eGov’s first contribution to the AI mission is its in-principle approval to be part of the OCC initiative, which aims to scale up India’s AI capabilities. The company recognises the need for a distributed cloud and compute infrastructure rather than a centralised one, drawing inspiration from successful examples like Aadhaar and the Account Aggregator framework. “Instead of building big techs that need data centres powered by nuclear fuel, etc., we can have a distributed cloud and compute rather than centralised,” Bhati explained. Protean eGov is also exploring the concept of making AI agents available on open networks, allowing anyone to consume and contribute to them. ProKisaan App: Reimagining Farmers’ Lives Protean eGov recently conducted a proof-of-concept called ProKisaan app, reimagining farmers’ lives using the combinatorial power of Digital Public Infrastructure (DPI) like UPI, Bhashini for localisation, and ONDC for buying, selling, and procuring. The company partnered with Google to provide data on weather, MSP rates, and the best crops to sow, making these decision points available in local languages. ProKisaan offers a range of features to assist farmers at every stage of the agricultural process. The app provides weather predictions and decision support for crop selection based on the season. It also enables farmers to procure fertilisers and seeds through the Open Network for Digital Commerce (ONDC). One of the key highlights of ProKisaan is the integration of ONEST, an AI-powered chatbot that offers bite-sized learning modules on various topics relevant to farmers. “Today, as individuals, we use a lot of these LLMs. If we have a question, we just prompt it and get an answer. We thought, why not use these ONEST trails and have a quick small bite-sized learning offering through ONEST,” Bhati explained. ProKisaan also leverages AI to detect crops and create catalogues, allowing farmers to easily publish their yield on the network for potential buyers. The app supports UPI transactions and is localised in multiple languages using Bhashini, with plans to expand the language reach through Google’s technology stack. “There is immense scope here. Maybe we can have multiple learning modules; kids can use it to ask questions in local languages; and there can be healthcare offering of IDs available,” Bhati added. Protean developed ProKisaan as a proof-of-concept in collaboration with Google and other partners, demonstrating the potential of combining AI, DPIs, and open digital ecosystems to create transformative solutions for citizens. Leveraging AI Across Public Sector Projects Protean eGov is witnessing a growing demand for AI-powered solutions from its existing customers and future opportunities. “Most of our existing customers and the future opportunities are talking about AI. And when we say they’re talking about AI, it is about how AI can come in to make things autonomous, agile, and all-inclusive,” Bhati said. The company has worked with over eight ministries in the past and is now focusing on leveraging AI to enhance personalisation, engagement, and inclusion in its offerings. It is also exploring how AI can help prevent fraud and improve the safety and speed of its products. “Whether it’s our existing customers of PAN, where it is a customer, or PFRDA as a regulator and other products, there is a unanimous ask on how AI can intervene and do things more, how our products can be safer, faster, and inclusive,” Bhati explained. Protean eGov’s roadmap includes developing AI-powered solutions for fraud detection and localised language chatbots for external customers. Internally, the company is experimenting with AI to accelerate code development and reviews, optimise infrastructure management, and strengthen security posture. “Security is something that we are wary about as an organisation. Can AI come in to support us on our security posturing are few of the experiments we are running ourselves as well as with the collaboration partnership we have with the big techs,” Bhati added. Advocating for Verifiable Credentials and Data Minimisation While GenAI continues to transform industries and raise concerns about data security, Protean eGov advocates for the adoption of verifiable credentials and data minimisation to safeguard personal information. “Data is sacrosanct, specifically PII (personally identifiable information), and with the new DPDP about to be rolled out, I guess the control is more about the data not owned by the organisation, but of the individual,” Bhati said, emphasising the importance of treating data as sacrosanct, especially in light of the upcoming Digital Personal Data Protection (DPDP) Bill. Bhati highlighted the growing awareness among urban populations about data ownership and privacy. “Individuals, at least the urban population like you and me, are more aware of our data. We are very particular about it, be it on social media or while talking to our banks, investment, etc. It’s eventually us and our data as individuals,” he added. Protean eGov believes that technologies like AI and blockchain can play a crucial role in enabling verifiable credentials, allowing for authentication and verification without the need to share sensitive information like Aadhaar or PAN card numbers. Bhati cited the example of DigiYatra, a contactless passenger processing system that uses facial recognition technology backed by Aadhaar, as a successful use case of verifiable credentials. “Today, there are technologies, I wouldn’t say specifically AI, where verifiable credentials can play a role without sharing the information, your authentication verification. So, you need not even share your Aadhaar number or PAN card number per se,” Bhati explained. Protean eGov believes that the adoption of verifiable credentials and data minimisation can help organisations lower their costs associated with securing and encrypting personally identifiable information (PII). “I guess in the next couple of years, the verifiable credential where I don’t think any organisation would need to save any  of this PII data because it will work on more on a verification basis,” Bhati added. Partnerships with Microsoft and Google Protean eGov is actively collaborating with tech giants Microsoft and Google to develop innovative product offerings and establish Centers of Excellence focused on emerging technologies like AI and DPI. One of the key areas of collaboration is leveraging DPIs available in India, productising them, and using large tech companies to further scale and develop common go-to-market strategies. Protean is also working with organisations like MOSIP (Modular Open Source Identity Platform) and OpenCRVS (Civil Registration and Vital Statistics) to extend its reach internationally. On the security front, Protean has been experimenting with Microsoft’s Security Copilot and Google’s security solutions. The company is also utilising Google’s Vertex AI to create predictive modelling for fraud detection, drawing inspiration from how Google alerts users when logging in from an unfamiliar IP or device.","excerpt":"Instrumental in facilitating PAN, TIN, PRAN and Aadhaar, they now see potential for AI-powered infrastructure across healthcare, education, agriculture, and governance","categories":["AI Features"],"tags":["ondc"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-05-30T10:00:00","publication_year":"2024","word_count":1311,"keywords":["GenAI","machine learning","AI","ondc","chatbots","R","RAG","Aim","analytics","predictive analytics","fraud detection"],"extracted_tech_keywords":["AI","machine learning","analytics","GenAI","Aim","RAG","chatbots","predictive analytics","fraud detection","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/after-pan-aadhaar-protean-now-leverages-ai-for-citizen-centric-solutions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10085357,"title":"Virender Sehwag&#8217;s Cricuru Re-launched with Advanced AI-based Tech","content":"Cricuru, a brainchild of former cricketers Virender Sehwag and Sanjay Bangar, has been re-launched with a more advanced AI integration that allows every user to learn and improve their skills and performance. The AI tool, which Cricuru has developed, will be the first of its kind in cricket education in India, which will not only help students learn but also enhance the player’s skills, the company said. Cricuru is a new cricket ed-tech platform that started in 2021 with a goal to provide a comprehensive marketplace for virtual cricket training and ensuring every cricket enthusiast’s dream of learning cricket from the best comes true. It identified the need for integration of cricket training into education to provide cricket fans an augmented cricket-learning experience. The new upgrade introduces AI, which analyses the user’s positioning, moves and body structure and sends out feedback which helps in improvement. Cricuru helps one keep physically active and practice indoors even in bad weather conditions. Simulation-based training helps students improve their game and perform well in real life games. One of the major hurdles that Cricuru aims to address is geographical, taking into account Tiers 2 and 3 cities that do not have easy access to sports training, let alone a coaching centre for women in sports. Cricuru, being an app-based platform helps overcome these boundaries making cricket coaching easily accessible to all irrespective of age, location and gender. “With the new technologies in place, the app will ensure the registered students receive the best-in-class training from world renowned cricketers. With the relaunch of the app, Cricuru aims to revolutionise the otherwise traditional cricket coaching in India,” Akshaara Lalwani, CEO, Cricuru, said.","excerpt":"Cricuru is a new cricket ed tech platform with a goal to provide a comprehensive marketplace for virtual cricket training","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-01-18T11:51:55","publication_year":"2023","word_count":276,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","R"],"extracted_tech_keywords":["AI","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/virender-sehwag-backed-cricuru-relaunches-with-an-advanced-ai-based-tech\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164808,"title":"Netflix is Hiring for ML Scientist and ML Engineer","content":"Netflix, one of the world’s leading entertainment services, is offering new remote job openings for a machine learning scientist and a machine learning engineer. These roles, part of the Content & Media ML Foundations team, aim to enhance content intelligence, personalisation, and advertising through machine learning. The team builds foundational ML solutions embracing Netflix’s vast media data, driving advancements in multi-modal content understanding. They also explore generative AI for filmmaking and media intelligence to position Netflix at the forefront of AI-driven content creation and distribution. Netflix is on the hunt for an ML scientist to pioneer innovations in multimodal representation learning. The role involves building state-of-the-art ML models for visual, audio, and textual data, optimising performance and scalability using PyTorch and Netflix’s ML infrastructure, engaging with the ML research community, and influencing strategic decisions. Meanwhile, the company is also hiring an ML engineer to develop scalable ML pipelines powering content intelligence. Key responsibilities include optimising large-scale ML models for media understanding, automating ML workflows for faster experimentation and deployment, and enhancing observability and monitoring to ensure model reliability. Netflix seeks engineers with expertise in deep learning architectures, embedding methods, and distributed ML training. Candidates should have 5+ years of industry experience, particularly in NLP, audio, and video understanding. Netflix and AI During the Q3 2024 earnings interview, Netflix co-CEO Ted Sarandos, said, “AI needs to pass a crucial test. Actually, can it help make better shows and better films? That is the test and that’s what they have to figure out.” He emphasised that for AI to be truly impactful, it must contribute to the quality of storytelling rather than simply reducing production costs. Sarandos’ statement reinforces Netflix’s commitment to enhancing viewer experience and industry standards through technology. “Netflix is the best platform for premium stories because we’re the home to the best storytellers. We have an enormous reach–600 million watchers. We assume the financial risk when we’re making your content,” said Sarandos. That’s not all. Netflix is doubling down on AI, not just in film and TV but also in gaming. The company has onboarded Mike Verdu as the VP of GenAI for Games at Netflix.","excerpt":"These roles, part of the Content & Media ML Foundations team, aim to enhance content intelligence, personalisation, and advertising through machine learning.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Hiring","netflix","Remote job"],"author_name":"Vidyashree Srinivas","publish_date":"2025-02-28T09:50:34","publication_year":"2025","word_count":357,"keywords":["GenAI","machine learning","AI","PyTorch","ML","Hiring","netflix","NLP","Remote job","Aim","deep learning","generative AI","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","generative AI","GenAI","Aim","PyTorch","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/netflix-is-hiring-for-ml-scientist-and-ml-engineer\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10078051,"title":"Google Cloud Chief T Kurian on Barriers to AI Adoption in Enterprises","content":"Despite the never-ending benefits of AI and analytics, many companies still struggle to adopt them to scale their business, know customers better, and add value. According to experts, the adoption has been slower than expected. At the recent Scale TransformX Conference, Google Cloud chief Thomas Kurian, alongside Scale AI founder Alexandr Wang, shared his insights into the state of AI adoption, vision for the company, business philosophy, latest innovations, multiple use cases and more. Addressing the barriers to AI adoption from Google Cloud perspective – which has been one of the early movers in AI – Kurian said one of the barriers is rarely about the algorithm itself as it is very different across industries. Citing the retail industry, particularly in line with the recommendation system used to make product discovery, Kurian said that the biggest challenge is standardising the meaning of the product and the catalogue because unless you have a standardised definition of the products and the data behind the algorithm is clean, it is super hard to get a recommendation. “We worked with H&M, Macy’s, IKEA, and Bloomingdale and for a large number of these brands, a big part of the problem is how you label and clean the data upfront and standardise it before you get into the algorithmic phase,” he added, saying that is the one part of things they see. “The second part is that for large organisations to adopt AI, they have to need to input the results of the algorithm back into their core processes,” said Kuran, citing GE (Europe), which they are helping manage the grid and produce electricity effectively using AI and analytics solutions. “The third is the people’s side. There is change management you go through to get people to trust the algorithm,” said Kurian, giving examples of banks that can disburse loans, mortgages and offer financial products and services. First, however, it requires customers to get comfortable with it, where things like fairness and others come into play. “Often, when people look at AI, they think it is a skills issue. Sure, there is not enough talent in the ecosystem. But, things are getting easier as the models get more and more sophisticated,” said Kurian, adding that people often forget about these other important issues in adopting. Innovation philosophy “In the past three years, we have seen a huge ramp up in our business, and the credit goes to all the people who joined us,” said Kurian. Roughly half of Google Cloud’s employees have been hired since 2020. Still, they all did an amazing job together, he quipped. Throwing light on the application of AI in different domains, Kurian said they recently worked with a large financial institution in Hong Kong and Shanghai Bank, which used its machine learning solutions to detect fraud. “There are a lot of false positives, where people are doing something called anti-money laundering, which our AI algorithms can be super-precise in detection,” he added. Further, he said, since regulators are involved in the process, explainability becomes a big deal. Citing Renault, he said that they help the company monitor all of their factories, where they process roughly a billion datasets every day. “Obviously, humans can not process that,” said Kurian. Google Cloud has also helped IKEA build a recommendation system. Kurian said that people shopping for furniture and products are not the same in many countries as they have different buying habits. “Those are all the things we have learned applying our AI in different contexts in different parts of the world,” he added. Growth philosophy Ever since Kurian took over as the CEO of Google Cloud in 2019, the company has witnessed meteoric growth that has tripled over the past few years. As per its fourth quarterly results, the company has an annualised revenue run rate of $22.16 billion, up from $19.96 billion based on its third-quarter results. Sharing the reason for its incredible growth, Kurian said everything in the world is becoming a software-powered technology industry. Citing the automobile industry, he said that vehicles are becoming more software than mechanical. Similarly, telecommunication companies rely on platforms to deliver applications and manage their network effectively. Banks, on the other hand, are becoming more and more virtual each day, and all of the products are based on data, and how they use them to their advantage and enhance customer experience. This is redefining the banking landscape. “Ask yourself, when was the last time you visited a branch of a bank,” quizzed Kurian, saying that a lot of work at Google Cloud has been about pushing the technology innovation far, and making it super easy for people in different industries to adopt, access, and scale. “We offer every part of the stack that we have from the hardware network to software abstractions, to things that are more packaged because different organisations have different levels at which they have the expertise and want to adopt technology,” said Kurian. Read: The Curious Case of Google Cloud Revenue Google Cloud’s AI & analytics strategy “Our vision is super simple,” said Kurian. Speaking of smartphones and how it was able to bring various functions of computing, camera, communication and the internet into everyone’s pocket, he drew parallels, and noted how Google too is trying to take all the technological innovation and make it super simple for everyone to consume it. This explains its investment in global data centres and the development of new types of hardware and large-scale systems, alongside working on software to help customers handle high-scale computations, tools for data processing, cybersecurity, and machine learning, etc. Kurian said machine learning and AI has done much work in the past three to four years. “We look at our work as four elements,” he added. The first and foremost include taking their large-scale computing systems (TPUs, GPU-based systems, etc.) and making them available to everybody. Second is the software stacks (JAX, TensorFlow, scikit-learn, etc.). The third is more advanced solutions based on the requirements of the customers (AutoML, image, video and audio translation, etc.). Last but not least, the company has built a complete packaged solution. What’s next? “So, we feel that the boundary of what machine learning and AI can do will change over time,” said Kurian. For instance, he said, it was about doing assistive things when it started. Assistive things are what a human being can do, but the computer assists the human being in some ways to do it better. This was followed by something you couldn’t do with humans because of the quantity of data you need to process or it might be far too significant. “So, the machine is doing something that humans couldn’t do, but it is still an incremental element on top of what humans could do themselves,” said Kurian. “The third phase is generative AI,” said Kurian, enabling people to express themselves differently.","excerpt":"“Laws in many countries are not keeping up with how fast AI technologies are moving,” says Thomas Kurian","categories":["Global Tech"],"tags":[],"author_name":"Amit Naik","publish_date":"2022-10-26T18:00:00","publication_year":"2022","word_count":1144,"keywords":["scikit-learn","machine learning","TPU","AI","ML","generative AI","analytics","JAX","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","generative AI","TensorFlow","JAX","scikit-learn","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/google-cloud-chief-t-kurian-on-barriers-to-ai-adoption-in-enterprises\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10113563,"title":"Google Challenges Meta’s Llama 2 with Lightweight Open Source LLM, Gemma","content":"Google has unveiled Gemma, a new family of open models, leveraging the research and technology behind the existing Gemini models. The Gemma open models are released in two sizes, Gemma 2B and Gemma 7B, each offering pre-trained and instruction-tuned variants. Users can start working with Gemma today using free access in Kaggle and a free tier for Colab notebooks. Additionally, first-time Google Cloud users can avail themselves of $300 in credits. Researchers can also apply for Google Cloud credits of up to $500,000 to accelerate their projects. Gemma outperforms Llama 2 on several benchmarks, including MMLU, HellaSwag, and HumanEval. To support developer innovation and responsible use, Google is also providing a Responsible Generative AI Toolkit alongside the models. This toolkit includes essential tools for creating safer AI applications with Gemma, offering guidance and support for developers. To facilitate widespread adoption, Gemma is compatible with major frameworks, including JAX, PyTorch, and TensorFlow through native Keras 3.0. The release includes ready-to-use Colab and Kaggle notebooks, integration with popular tools such as Hugging Face, MaxText, NVIDIA NeMo, and TensorRT-LLM. Gemma models can run on various platforms, from laptops and workstations to Google Cloud, with optimisation for industry-leading performance on NVIDIA GPUs and Google Cloud TPUs. This development comes after Google recently introduced Gemini 1.5 with a 1 million token context window — the largest ever seen in natural language processing models. In contrast, GPT-4 Turbo has a 128K context window, and Claude 2.1 has a 200K context window.","excerpt":"Gemma outperforms Llama 2 on several benchmarks, including MMLU, HellaSwag, and HumanEval.","categories":["AI News"],"tags":["Gemini","Gemma","Google","LLMs","Open Source AI"],"author_name":"Siddharth Jindal","publish_date":"2024-02-21T19:12:15","publication_year":"2024","word_count":245,"keywords":["Gemini","Hugging Face","Gemma","Keras","AI","LLMs","PyTorch","ML","RAG","Open Source AI","Colab","generative AI","Google","JAX","TensorFlow"],"extracted_tech_keywords":["AI","ML","generative AI","TensorFlow","PyTorch","JAX","Keras","Hugging Face","Colab","RAG"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-challenges-metas-llama-2-with-lightweight-open-source-llm-gemma\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10163207,"title":"Zoom Expands Phone Services to Chennai, Brings AI Capabilities to Businesses","content":"Zoom Communications, Inc., launched its cloud-based telephony service, Zoom Phone, in Chennai, marking its expansion into Tamil Nadu. This launch follows the company’s rollout in Maharashtra in October 2024. Zoom Phone enables businesses to manage inbound and outbound calls via the public switched telephone network (PSTN), replacing traditional PBX systems and integrating with Zoom’s other communication tools. It also features Zoom AI Companion, an AI-powered assistant that is included at no extra cost with paid Zoom accounts. It offers post-call summaries, voicemail prioritisation, and task extraction. “We are excited to bring Zoom Phone to Chennai, one of India’s most vibrant technology business hubs. The city’s thriving ecosystem of local and global enterprises will benefit from the simplicity, scalability, and AI-first capabilities that Zoom Phone delivers,” said Velchamy Sankarlingam, president of product and engineering at Zoom. “India is a critical market for us, and the availability of Zoom Phone in Chennai reflects our commitment to delivering local solutions with global standards.” The service is designed to meet the needs of both multinational corporations and local companies, offering a user-friendly interface, security, and scalability. With this expansion, Zoom aims to strengthen its presence in India’s growing business communication sector, catering to companies that require reliable and efficient telephony solutions.","excerpt":"Zoom Phone, in Chennai, marking its expansion into the Tamil Nadu Telecom Circle. This follows the company’s rollout in the Maharashtra Telecom Circle in October 2024.","categories":["AI News"],"tags":["AI","Zoom"],"author_name":"Shalini Mondal","publish_date":"2025-02-11T13:47:12","publication_year":"2025","word_count":207,"keywords":["Zoom","AI-first","programming_languages:R","AI","Scala","Aim","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","Aim","R","Scala","AI-first","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/zoom-expands-phone-services-to-chennai-brings-ai-capabilities-to-businesses\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10165756,"title":"Quest Global Acquires US-Based VLSI Design Firm Alpha-Numero Technology Solutions","content":"Quest Global has announced the acquisition of Alpha-Numero Technology Solutions, a US-based company known for its expertise in very large scale integration (VLSI) design and field programmable gate array (FPGA) technology on Monday. The move aims to strengthen Quest Global’s capabilities in semiconductors and expand its mission and safety-critical engineering solutions. Notably, Alpha-Numero is headquartered in Irvine, California, and employs over 150 engineers worldwide. The company has research and development centres in Ahmedabad, Hyderabad, and Endicott. It has built a strong reputation in avionic design and verification, with long-standing partnerships in the aerospace, hi-tech, and automotive industries. Alpha-Numero will continue to be led by its co-founders Sukanta Mitra, Santhi Ayyadevar, and Hari Patel. In a joint statement, the management team said, “Quest Global’s focus on helping clients solve their hardest engineering problems is a perfect fit with Alpha-Numero. We pride ourselves on our longstanding leadership in avionic design and mission and safety critical solutions.” Alpha-Numero has deep expertise in safety-critical avionics systems and complex electronics hardware. The company specialises in managing projects from Design Assurance Level A (DAL-A) to DAL-E. It also works on design and verification of avionics hardware and software, ensuring compliance with key safety standards like DO-254 and DO-178B\/C. Alpha-Numero supports all required documentation for Federal Aviation Administration (FAA) and European Union Aviation Safety Agency (EASA) audits at every stage of the process.In addition to aviation, Alpha-Numero also designs and verifies FPGA or ASIC or system on a chip (SoC) used in automotive systems (aligned with ISO 26262 standards) and consumer electronics.","excerpt":"Quest Global’s focus on helping clients solve their hardest engineering problems aligns perfectly with Alpha-Numero","categories":["AI News"],"tags":["AI","aviation"],"author_name":"Shalini Mondal","publish_date":"2025-03-10T16:45:40","publication_year":"2025","word_count":255,"keywords":["programming_languages:R","AI","aviation","Ray","Aim","R"],"extracted_tech_keywords":["AI","Aim","Ray","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/quest-global-acquires-us-based-vlsi-design-firm-alpha-numero-technology-solutions\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":30000,"title":"AIMinds Bengaluru Edition: Meet And Learn From India’s Finest Data Scientists","content":"AIMinds, a hub for AI intellectuals, is back in Bengaluru with its fourth edition in collaboration with REVA University, Bangalore. A venture by Analytics India Magazine, AIMinds is a monthly meetup which puts common public and data science enthusiasts in touch with artificial intelligence practitioners and researchers from all over the country. The registration for the meetup is free. Click here and RSVP to register. AIMinds is a platform where established and aspiring minds in AI, data science and analytics can meet to have great conversations and share knowledge under one roof. People of all skill levels are welcome. Agenda for the meetup: Demystifying Data Science Date: 24 November 2018 Time: 10:00 am to 2:00 pm Keynote speakers: 1.Sundara Ramalingam Nagalingam – Head – Depp Learning Practice, NVIDIA Graphics Pvt Ltd 2.Dr JB Simha – CTO, ABIBA Systems; Chief Mentor – Analytics, REVA Academy for Corporate Excellence, REVA University 3.Ratnakar Pandey – India Head Analytics and Data Science at Kabbage Inc This is an event for established and aspiring minds in AI, Data Science and Analytics to meet, have great conversations and learn from Industry thought leaders. Freshers with tech background and who are enthusiastic about these technologies are also welcome. As this meetup is primarily intended at Knowledge sharing, it will be good if you have loads of questions to make this a very interactive meet-up session. Each session will be timed at about 35-40 minutes, followed by a Q&A round. So, do send in your RSVP soon and be part of this exciting meet-up along with your friends and colleagues who are also interested in analytics or data science and want to know how to transition to this space.","excerpt":"AIMinds, a hub for AI intellectuals, is back in Bengaluru with its fourth edition in collaboration with REVA University, Bangalore. A venture by Analytics India Magazine, AIMinds is a monthly meetup which puts common public and data science enthusiasts in touch with artificial intelligence practitioners and researchers from all over the country.   The registration […]","categories":["Deep Tech"],"tags":[],"author_name":"Disha Misal","publish_date":"2018-11-08T05:40:31","publication_year":"2018","word_count":281,"keywords":["data science","Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","Aim","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/aiminds-bengaluru-edition-meet-and-learn-from-indias-finest-data-scientists\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161075,"title":"The Dark Side of AI Upskilling in Indian IT","content":"Indian companies are looking to upskill around 85% of their workforce with generative AI in FY25. This marks one of their largest investments in measuring the ROI of AI, yet the quality of the training offered remains questionable. A former employee of Infosys, who wanted to maintain anonymity, told AIM that when their company partners with the likes of NVIDIA, Google, or Microsoft for training their employees in generative AI, the outcome is not that great since the actual training is done in partnership with smaller firms offering entry-level courses at cheaper rates. He explained that most of the training programmes involve presentations and slides where they simply need to press ‘next’ a few times over, answer a few multiple choice questions, answers to which can be easily found on Google, and certify themselves as ‘GenAI Trained’. Therefore, these programmes can sometimes be more attractive for firms, as they can “train” their employees in generative AI while paying a lower price and coming across as better service providers to clients. When it comes to the big-tech partnerships for training, the employee said that the target usually is to take 2-3 years to make the workforce “GenAI Ready”, which is a lot of time, and most do not even have concrete plans ready yet. He explained that this is because most Indian IT companies are focused on providing services rather than building products for their clients. The employees’ tasks mostly involve working with chatbots or copilots and helping their clients, which does not require much knowledge of generative AI. The employee added that there are always smarter ways to train engineers in generative AI, like letting them experiment with tools like Cursor or GitHub Copilot. However, companies are sceptical of such tools and hesitant to test them as their code bases are proprietary. But, this is slowly changing as some companies have started partnering with GitHub and other AI tool providers to enable their employees to experiment with the tools. Emails sent to Infosys and TCS did not elicit any response. Clicking ‘Next’ is All You Need This was also revealed earlier by a user on X when he said that even though around 900,000 employees have been trained in generative AI, the depth and quality remain questionable. “A friend of mine works at one of the largest IT companies in India, and she just completed a GenAI course in an hour by clicking the next button 100s of times. She is now part of a GenAI-ready workforce! Proud of her :).” Even though many companies offer in-depth courses for their employees, the trained workforce remains underutilised. “Such certifications and credentials are bogus, rather a waste of time and effort,” said a user on X. “There is a huge gap between the demand and the quality of candidates that exist in the AI job market, especially for GenAI.” The number of employees trained in India is close to 2.5 lakhs as of this month. Most of these are from Indian IT companies where generative AI training is mostly compulsory to sit through, even though it does not offer any benefits for them. There’s an upskilling mania of sorts currently sweeping across the tech industry. Today, it is generative AI skills. In the future, it might be something else. “IT changes so rapidly that you have to learn until the end of your career,” said a user on Reddit. With the advent of every new technology, be it Python, cloud, or AI, long-standing Indian IT employees need to keep upgrading to keep up. Just Another Trend? “Just like with the metaverse c**p the other year, we have to let companies know that we have some level of competency in generative AI,” said a user on Reddit. Mrinal Rai, assistant director and principal analyst at ISG, told AIM that Indian IT clients will prioritise the extent to which service providers need a workforce trained in key AI technologies. This suggests that having a well-trained team is seen as a competitive advantage and a deciding factor for clients when selecting a service provider. Rai added that smaller AI firms currently don’t have much influence or recognition among clients regarding their ability to meet training needs. As a result, clients would prefer larger or more established firms for AI training and implementation. That is probably why all the bigger IT firms are rushing to call their workforce ‘GenAI ready’, even if that requires minimal training, to appeal to clients. What Needs to Change Krishna Vij, VP of IT hiring at TeamLease Digital, told AIM that upskilling programs in the Indian IT industry have evolved in recent years, focusing on emerging technologies like AI, cloud, and data science. “The best programs prioritise hands-on learning and problem-solving. While large companies are investing in structured platforms, there’s still room to tailor programs to specific roles and ensure they go beyond basic knowledge for competence building and driving efficiency,” Vij said. Vij said that training programs facilitated as ‘click-through’ sessions clearly reveal a design gap, prioritising compliance over actual skill-building. “Effective upskilling isn’t about ticking boxes; it’s about formats that truly engage like interactive modules, case studies, hackathons, and real-world applications,” she said. Vij added that many companies are making significant investments in this direction by providing employees with the opportunities to work on live AI projects, which ensures practical learning that translates into meaningful competence and career relevance. Satya Nadella, on his recent visit to Bengaluru, highlighted that Microsoft was committed to upskilling 10 million people in AI by 2030. Now, this is definitely something to keep an eye out for.","excerpt":"“Effective upskilling isn’t about ticking boxes; it’s about formats that truly engage like interactive modules, case studies, hackathons, and real-world applications.”","categories":["IT Services"],"tags":["AI in Indian IT","New AI Skills"],"author_name":"Mohit Pandey","publish_date":"2025-01-09T19:00:39","publication_year":"2025","word_count":933,"keywords":["AI in Indian IT","data science","GenAI","AI","New AI Skills","chatbots","ML","Python","Aim","generative AI","copilots","R"],"extracted_tech_keywords":["AI","ML","data science","generative AI","GenAI","Aim","copilots","chatbots","Python","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-dark-side-of-ai-upskilling-in-indian-it\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":32012,"title":"Integrating AI Into Blockchain Can Help The Economy In More Ways Than You Think","content":"Artificial intelligence is one of the major driving forces of innovations loved by entrepreneurs and users alike. It has disrupted numerous industries including customer service, agriculture, manufacturing, healthcare, tech support and many more. Blockchain technology is another force to be reckoned with. Although it is presently popular in the BFSI sector, there are endless uses for Blockchain technology, such as, education, transportation, voting, law and enforcement, among others. Blockchain, though powerful on its own, becomes enhanced to a whole new level when coupled with AI. New features and capabilities become unlocked, enhanced, and more secure through their convergence. How AI Can Add On To Blockchain The confluence of AI in blockchain creates perhaps what is the world’s most reliable technology enabled decision making systems that are virtually tamper-proof and provides solid insights and decisions. It holds several benefits like: Improved business data models Globalised verification systems Innovative audits and compliance systems Smarter finance Transparent governance Intelligent retail Intelligent predictive analysis Digital Intellectual Property Rights Technical Enhancements That AI Can Enable Security: With the implementation of AI, Blockchain technology becomes more secure by creating secure future application deployments. AI algorithms which are increasingly making decisions about whether financial transactions are fraudulent and should be blocked or investigated is a good example of it. Efficiency: AI can help optimise calculations to reduce miner load which results in less network latency for faster transactions. AI enables to reduce the carbon footprint of blockchain technology. The cost which is applied upon miners would also be reduced along with the energy spent if AI machines replace the work done by miners. As the data on blockchains grows by the minute, AI’s data pruning algorithms can be also be applied on the blockchain data which automatically prunes the data which is not required for future use. AI can introduce even new decentralised learning systems such as federated learning or new data sharing techniques that make the system much more efficient. Trust: The iron cast records of blockchain is considered one its USP. Applied in conjunction with AI means that users have clear records to follow the system’s thinking process. This, in turn, helps the bots trust each other, increasing machine-to-machine interaction and allowing them to share data and coordinate decisions at large. Better Management: When it comes to cracking codes, human experts get better over time with practice. A machine learning-powered mining algorithm will eliminate the need for human expertise as it could almost instantaneously sharpen its skills if it is fed the right training data. So, AI also helps in managing blockchain systems better. Privacy And New Markets: Making private data secure invariably lead to it being sold, resulting in data markets\/model markets. The markets get easy, secure data sharing that helps smaller players gain Blockchain’s privacy can be more increased by executing “Homomorphic encryption” algorithms. Homomorphic algorithms are the ones using which operations can be performed on encrypted data directly. Storage: Blockchains are ideal for storing the highly sensitive, personal data which, when smartly processed with AI, can add value and convenience. Smart healthcare systems that make accurate diagnoses based on medical scans and records is a good example of that. Real-Life Applications: Triggmine, a product converges AI in blockchain within a single marketing automation system. It has led to the company being one of the world’s first decentralised platform for marketing automation. Mithril is a social media app which makes content creation and engagement on social media worthwhile for users, in the form of crypto. Their social media platform enables what Mithril has termed “social mining” in which users are rewarded for the impact and influence of their content. They also make extensive use of AI in their overall working. Bubblo is an AI-based real-time discovery app that uses blockchain technology and virtual reality to build a marketplace allowing users to monetise their personal information for rewards. It provides users personalised recommendation of bars, restaurants and other venues. Back in India, Signzy couples artificial intelligence with the blockchain to make secure, compliant and user-friendly transactions in banks. It is harnessing the power of AI and blockchain to enable their clients to adapt digitisation and offer fully digital experiences to their users. It offers products such as RealKYC which is a bank-grade digital KYC and Digital contracts which is a secured digital contract enabled by Aadhaar and Biometrics EVS is an employee verification platform created by SpringRole which verifies employees’ credentials and skills. It uses blockchain as a platform where the feedbacks registered are fed into an AI system along with using a smart contract method which digitally facilitates, verify, or enforce the negotiation or performance of any contract. Outlook As the industries are seeing newer developments, there will be more blockchain and non-blockchain based products looking to leverage the advantages that AI can offer. The long-term potential for both blockchain and AI is staggering, so combining the two certainly makes such projects worth beneficial. It enhances each other’s capabilities as seen in few instances above while offering opportunities for better oversight and accountability. These developments simply point towards why AI should not be brought into blockchain.","excerpt":"Artificial intelligence is one of the major driving forces of innovations loved by entrepreneurs and users alike. It has disrupted numerous industries including customer service, agriculture, manufacturing, healthcare, tech support and many more. Blockchain technology is another force to be reckoned with. Although it is presently popular in the BFSI sector, there are endless uses […]","categories":["AI Features"],"tags":[],"author_name":"Martin F.R.","publish_date":"2018-12-22T05:31:12","publication_year":"2018","word_count":853,"keywords":["federated learning","Go","machine learning","artificial intelligence","AI","homomorphic encryption","Git","RAG","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","federated learning","homomorphic encryption","RAG","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/integrating-ai-into-blockchain-can-help-in-more-ways-than-you-think\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10528,"title":"Mezi raises funds worth $9 million to accelerate AI bots technology","content":"Mezi, a leading chatbot shopping app has received funding of $9million in Series A Round from investors. Investors in this round included previous investor Nexus Venture Partners and new investors Saama Capital and American Express Ventures. Also, Amit Singhal, former SVP and Head of Google Search, and Gokul Rajaram, Product Engineering Lead at Square, will be joining in as angel investors. The company plans to use this investment to back the company’s growth and accelerate Mezi’s product and artificial intelligence (AI) technology development around conversational commerce, AI, Natural Language Processing (NLP), and machine learning. Mezi is a personal shopping assistant. The way Mezi works is you just need to send a message to Mezi when you want to shop something and then its Mezi’s responsibility to carry out a thorough research and present you the various available options. Mezi helps you till the final process of placing the order. All this is made possible by Mezi as it is powered by the world’s smartest chatbots and human guided experts. Yes Mezi is a perfect combination of two worlds – human expertise and artificial intelligence. A perfect blend of artificial intelligence and human expertise, allows Mezi to understand the shopping preferences of its users and hence provide a seamless, personalized shopping experience over messaging. Mezi aims at simplifying the mobile shopping experience by sifting through thousands of potential choices based on user price points, preferences, and learned habits and then making suggestions. The idea behind this app is to provide an experience which is very similar to shopping with your best friend by way of chatting and having intelligent and humanistic conversations. Mezi bots are able to understand a wide range of requests and can even mimic speech patterns by learning from user conversations enabled through consciousness design and artificial intelligence. These Mezi bots can manage common queries with respect to travel and flower requests 80 per cent of the times from end-to-end without any human intervention and within two-minutes of time. The result is a rapid-fire conversational experience that makes them indistinguishable from human assistants. Mezi, available on the iOS App Store and SMS, has been successful in providing a wonderful experience to its users and this is evident from the fact that around 40% of its users return on a monthly basis to shop with Mezi.  From the time of its launch, December 2015, Mezi has exchanged one million messages till date making it a go-to daily shopping companion for its users. 25% of all purchases made through Mezi are for travel bookings, 21% for fashion items, and 15% for gifts for friends and family. “With Mezi, we have created the most advanced shopping bots that can anticipate our users’ needs and wants by harnessing the latest advances in artificial intelligence and consciousness design. We look forward to continuing to provide a shopping experience that delights and amazes our users with these world-class investors behind us,” commented Mezi Co-founder and CEO, Swapnil Shinde. Mezi uses proprietary technology platforms, Smart Connect and Smart Assist which are built using artificial intelligence, deep learning, and Natural Language Processing (NLP). These technologies help to streamline intent detection, automatic replies, expert workflows, reduce response times, and provide personalized product recommendations. “Mezi has created an intuitive platform that helps consumers unlock something we all want – a personal assistant in the palm of your hand, whether you need to book travel, purchase a gift or buy a new wardrobe,” said Rohit Bodas, Partner, American Express Ventures. “Our investment in Mezi reflects our belief in their core technology and their ability to shatter the vending-machine experience of past bots and provide a clearly superior product that delivers a truly personal experience for every user.” commented Ash Lilani, Managing Partner and Co-Founder Saama Capital. “Mezi elevates AI to the next level by uniting human intelligence with machine intelligence that shatters today’s current standard,” said Jishnu Bhattacharjee, Managing Director, Nexus Venture Partners.","excerpt":"Mezi, a leading chatbot shopping app has received funding of $9million in Series A Round from investors. Investors in this round included previous investor Nexus Venture Partners and new investors Saama Capital and American Express Ventures. Also, Amit Singhal, former SVP and Head of Google Search, and Gokul Rajaram, Product Engineering Lead at Square, will […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Deep Learning","latest advances","Machine Learning","Natural Language Processing"],"author_name":"Manisha Salecha","publish_date":"2016-08-03T06:01:10","publication_year":"2016","word_count":654,"keywords":["Go","API","machine learning","AI","chatbots","Natural Language Processing","ML","latest advances","Machine Learning","NLP","Aim","deep learning","Deep Learning","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","Aim","chatbots","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mezi-raises-funds-worth-9-million-accelerate-ai-bots-technology\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10173051,"title":"CP Gurnani says His 90-year-old Mother-in-law’s Best Friend is Her iPad with AI","content":"AI adoption is moving beyond enterprises and into living rooms, as seen in a personal account shared by CP Gurnani, co-founder of AIonOS. In a recent LinkedIn post, Gurnani recounted how his 90-year-old mother-in-law now refers to her AI-enabled iPad as her “best friend,” using it daily to ask about the weather or find movie recommendations. Gurnani was speaking at the ET Edge HR Transformation Summit 2025. “You know AI has arrived when my 90-year-old mother-in-law tells me her iPad with AI is her ‘best friend,’” Gurnani wrote. “If I want to know what’s trending, I just ask her or her AI.” The post reflected a growing sentiment among AI practitioners that mainstream acceptance hinges less on technical readiness and more on cultural openness. “True adoption of AI isn’t about forcing change,” Gurnani noted. “It’s about creating the right environment for people to experiment, learn, and see the value for themselves.” Gurnani emphasised that at AIonOS, the philosophy is to treat AI as an extension of human capability rather than a replacement. “It’s always human plus AI, not human or AI,” he said. “The real opportunity is in using these tools to free up teams for more creative, strategic work that truly moves the needle.” As businesses and individuals alike navigate the pace of AI advancements, Gurnani urged a shift in perspective. “The question we need to ask ourselves isn’t about the timeline of adoption,” he said, “but how we can foster a culture that encourages learning and embraces rapid technological change.” AIonOS was co-founded by Gurnani and Rahul Bhatia, founder of InterGlobe Enterprises. Launched in 2024–25, the company is headquartered in Singapore, with fulfilment centres in India (Noida, Hyderabad), the US, UK, Paris, Africa, and the Middle East. Focused on accelerating digital transformation, AIonOS builds AI-powered solutions that enhance both human and system capabilities. Its primary sectors include Travel, Transportation, Logistics, and Hospitality (TTLH), leveraging InterGlobe’s strong presence in these industries, including IndiGo Airlines. The company recently acquired a majority stake in Cloud Analogy, a Salesforce Summit Partner, to combine its generative AI platform with Cloud Analogy’s Salesforce expertise. The acquisition is aimed at improving AI-led customer experience transformation and expanding AIonOS’s presence across India, the US, Europe, and Asia.","excerpt":"“It’s about creating the right environment for people to experiment, learn, and see the value for themselves.”","categories":["AI News"],"tags":["CP Gurnani"],"author_name":"Siddharth Jindal","publish_date":"2025-07-08T16:44:07","publication_year":"2025","word_count":371,"keywords":["Go","API","CP Gurnani","programming_languages:R","AI","digital transformation","Git","RAG","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","R","Go","Git","API","digital transformation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/my-90-year-old-mother-in-law-tells-me-her-ipad-with-ai-is-her-best-friend-says-cp-gurnani\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":33861,"title":"Hands-On Guide For Non-Linear Regression Models In R","content":"It is a truth universally acknowledged that not all the data can be represented by a linear model. By definition, non-linear regression is the regression analysis in which observational data is modeled by a function which is a non-linear combination of the parameters and depends on one or more independent variables. Non-linear regression is capable of producing a more accurate prediction by learning the variations in the data and their dependencies. In this tutorial, we will look at three most popular non-linear regression models and how to solve them in R. This is a hands-on tutorial for beginners with the good conceptual idea of regression and the non-linear regression models. Pre-requisites: Understanding of Non-Linear Regression Models Knowledge of programming Polynomial Regression Polynomial regression is very similar to linear regression but additionally, it considers polynomial degree values of the independent variables. It is a form of regression analysis in which the relationship between the independent variable X and the dependent variable Y is represented as an nth degree polynomial in x. The model can be extended to fit multiple independent factors. Consider for example a simple dataset consisting of only 2 features, experience and salary. Salary is the dependent factor and Experience is the independent factor. Unlike Simple linear regression which generates the regression for Salary against the given Experiences, the Polynomial Regression considers up to a specified degree of the given Experience values. That is, Salary will be predicted against Experience, Experience^2,…Experience ^n. Code The Polynomial Regression is handled by the inbuilt function ‘lm’ in R. After loading the dataset follow the instructions below. Creating the Polynomial Regressor Model and fitting it with Training Set dataset$X2 = dataset$X^2 dataset$X3 = dataset$X^3 dataset$X4 = dataset$X^4 poly_regressor = lm(formula = Y ~ .,data = dataset) The first 3 lines calculate the nth degree polynomial of the independent variable X for each row of observations and add them as features into the original dataset. Here we have calculated till the 5th degree denoted as X4 formula: Used to differentiate the independent variable(s) from the dependent variable. In case of multiple independent variables, the variables are appended using ‘+’ symbol. Eg. Y ~ X1 +  X2 + X3 + … X: independent Variable or factor. The column label is specified Y: dependent Variable. The column label is specified. data: The data the model trains on, training set. Predicting the Y value for a new X predict(poly_regressor,newdata = data.frame(X = value, X2 = value^2, X3 = value^3, X4 = value^4)) This line predicts the value of the dependent factor for a new given value of independent factor. regressor: The regressor model that was previously created for training. newdata: The new observation or set of observations that you want to predict Y for. Accepts arguments as dataframes. value: replace this with a number you want to predict Y for. Visualizing the predictions install.packages('ggplot2') #install once library(ggplot2) X_grid = seq(min(dataset$X), max(dataset$X), 0.1) ggplot() + geom_point(aes(x = dataset$X, y = dataset$Y),colour = 'black') + geom_line(aes(x = X_grid, y = predict(poly_reg, newdata = data.frame(X = X_grid,X2 = X_grid^2, X3 = X_grid^3, X4 = X_grid^4))),colour = 'red')+ ggtitle('Polynomial Regression') xlab('X') ylab('Y') This block of code represents the dataset in a graph. ggplot2 library is used for plotting the data points. To obtain a smooth curve the axis is scaled to 1\/10th of X (X_grid). geom_point() : This function scatter plots all data points in a 2 Dimensional graph geom_line() : Generates or draws the regression line in 2D graph ggtitle(): Assigns the title of the graph xlab: Labels the X- axis ylab: Labels the Y-axis Decision Tree Regression Decision Tree Regression works by splitting a dimension into different sections containing a minimum number of data points and predicts the result for a new data item by calculating the mean value of all the data points in the section it belongs to. That is it breaks down a dataset into smaller and smaller subsets while at the same time an associated decision tree is developed incrementally. Decision tree builds regression or classification models in the form of a tree structure Code The Decision Tree Regression is handled by the rpart library. Installing and Importing Libraries install.packages('rpart') #install once library(rpart) # importing the library Creating the Decision Tree Regressor and providing the Training Set decisionTree_regressor = rpart(formula = Y ~ .,data = dataset, control = rpart.control(minsplit = 1)) The expression ‘Y ~ .” takes all variables except Y in the training_set as independent variables. formula: Used to differentiate the independent variable(s) from the dependent variable.In case of multiple independent variables, the variables are appended using ‘+’ symbol. Eg. Y ~ X1 +  X2 + X3 + … control: parameters that control the formation of the decision tree. minsplit: a controller used to specify the number of observations that must exist in a node in order for a split to be attempted. X: Independent Variable or factor. The column label is specified. Y: Dependent Variable. The column label is specified. data : The data the model trains on, training set. Predicting the values for the test set y_pred = predict(decisionTree_regressor, newdata = data.frame(X = value)) This line predicts the Y value for a given X value. Replace ‘value ‘ with real value. Visualizing the test set results library(ggplot2) x_grid = seq(min(dataset$X), max(dataset$X), 0.01) ggplot() + geom_point(aes(dataset$X, dataset$Y),color= 'red') + geom_line(aes(x_grid, predict(decisionTree_regressor, data.frame(X = x_grid)), color = 'black'))+ ggtitle('Y vs X (Decision Tree Regression) ') xlab('X') ylab('Y') This code plots the data points and the regressor on a 2 Dimensional graph. For more precision, the axis is scaled to 1\/10th of X (X_grid). geom_point() : This function scatter plots all data-points in a 2 Dimensional graph geom_line() : Generates or draws the regression line in 2D graph ggtitle(): Assigns the title of the graph xlab: Labels the X- axis ylab: Labels the Y-axis plot(decisionTree_regressor) This line displays the tree structure generated. Random Forest Regression Random Forest Regression is one of the most popular and effective predictive algorithms used in Machine Learning. It is a form of ensemble learning where it makes use of an algorithm multiple times to predict and final prediction is the average of all predictions. Random Forest Regression is a combination of multiple Decision Tree Regressions. Hence the name Forest. Code The library randomForest is used for handling Random Forest Regression in R Installing and Importing the Library install.packages('randomForest') #install once library(randomForest) # importing the library Creating the Random Forest Regressor and fitting it with Training Set random_forest_regressor = randomForest(x = training_set$X, y = training_set$Y, ntree = 300) This line creates a Random Forest Regressor and provides the data to train. X: independent variable Y: dependent variable ntree: the number of decision trees you want to generate to predict. Predicting the value for a new X y_pred = predict(regressor, data.frame(X = value)) Note: Replace ‘value’ with a real number you want to predict Y for.","excerpt":"It is a truth universally acknowledged that not all the data can be represented by a linear model. By definition, non-linear regression is the regression analysis in which observational data is modeled by a function which is a non-linear combination of the parameters and depends on one or more independent variables. Non-linear regression is capable […]","categories":["Deep Tech"],"tags":["regression analysis"],"author_name":"Amal Nair","publish_date":"2019-01-22T10:55:56","publication_year":"2019","word_count":1144,"keywords":["Go","machine learning","AWS","AI","cloud_platforms:AWS","RPA","programming_languages:R","regression analysis","programming_languages:Go","RAG","R"],"extracted_tech_keywords":["AI","machine learning","RAG","AWS","R","Go","RPA","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-for-non-linear-regression-models-in-r\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10074686,"title":"Council post: Things to keep in mind when building modern data estates","content":"From financial optimisation to creating a better customer experience, artificial intelligence (AI) has emerged as a force in helping organisations meet their operational and strategic goals. As per PwC, 54 percent of executives said AI solutions have turbocharged the business productivity. In another study, 61 percent of executives said AI helped their businesses identify opportunities in data which would otherwise be missed. Though businesses are bullish on AI adoption, many lack proper support infrastructure, and struggle while building the tech stack. Nearly 84 percent of the world’s leading enterprises don’t have accurate data and analytics strategies, and lack an understanding of the foundational processes, systems, and tools to become a truly data-powered company. As a result, companies with a competitive advantage in AI achieve 22 percent higher profitability on average compared to firms that don’t. To unlock the potential of AI, companies need to have a complementary stack of tools and applications. This includes tools that can help with feature extraction, analysis, process management, and machine resource management. Unfortunately, many companies either do not have a robust data infrastructure, or is too outdated to leverage the AI\/ML code in any meaningful way. This is where modern data estates come into play. Simply put, data estate is an infrastructure that helps companies systematically manage all of their data. This infrastructure or data estate can be built on-premises, in cloud, or combination of both (hybrid). On-premise to cloud migration Though companies across the globe were migrating fully – or partially – to the cloud even before COVID-19, the pandemic turned out to be a real accelerant. As per Gartner, the worldwide spending on public cloud services is expected to grow 20.4 percent in 2022 to total USD 494.7 billion, up from USD 410.9 billion in 2021. Thus, companies have to cope with the teething troubles of switching to the cloud, but also have to make a choice in deciding which technology is best suited to power their business functions. Companies need to quickly get a handle on the suitable cloud delivery models, cloud technologies, draft in specialised talent, and more, all while trying to integrate AI. Unfortunately, this is often a recipe for disaster and may result in even further disruption in terms of cloud migration and AI integration. As a result, businesses need to plan their cloud migrations to lay the groundwork for future AI adoption. Switch to self-service BI and analytics As businesses continue to accelerate their AI initiatives, they have grown increasingly frustrated with bottlenecks that prevent them from monitoring the decisions,  and activating new decisions based on real-time data signals. This has subsequently given rise to ‘self-service’ analytics. Therefore, companies have increasingly looked for ways to make their data easily accessible without involving analysts. But, to make this happen, companies have to adopt new components – decision augmentation and automation systems that integrate BI\/AI insights and real-time data signals with decision engines for monitoring, simulations, and activations– an uncharted territory for many businesses–leading to delays in delivering on broader tech and BI priorities. Data governance As per McKinsey, 30 percent of employees spend time on non-value-added tasks because of poor data quality and availability. As data becomes more central to business operations, there is a need for more governance. In other words, companies need a comprehensive suite of tools for oversight and reporting across business, technical, and operational metadata seamlessly. Presently, many companies have some data governance-related suite in place. But, because of the rise in awareness of explainability in governing AI, existing tools cannot deliver the insights needed to comply with modern governance requirements. As AI and data science teams grow, businesses need scalability and agility from their data estates. Thus, they need to embrace more modern tools to help them stay ahead of the curve. Wrapping up With Modern data estate, you can not only take data and transform it the way you like, but also find where the data was transformed, and assess how it was changed, and why it was changed. Modern data operations are both incredibly complex and exciting. However, to ensure that companies can hit all the marks in their data science and AI journeys, they need to have a concrete internal framework to support it. By keeping these trends and potential stumbling blocks in mind, companies can build the data estates they need to succeed. This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.","excerpt":"To unlock the potential of AI, companies need to have a complementary stack of tools and applications. This includes tools that can help with feature extraction, analysis, process management, and machine resource management.","categories":["AI Features"],"tags":[],"author_name":"Natwar Mall","publish_date":"2022-09-08T18:26:36","publication_year":"2022","word_count":768,"keywords":["data science","Go","artificial intelligence","AI","ML","RAG","Aim","analytics","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","data science","analytics","Aim","RAG","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/things-to-keep-in-mind-when-building-modern-data-estates\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10169975,"title":"Data Centres Take the Throne as the New Crown Prince of the Middle East","content":"While the United States and China have long been key players in the AI revolution, countries in the Middle East are the new entrants. As U.S. President Donald Trump commenced his three-day tour across Gulf countries, numerous partnerships followed to propel the country’s AI infrastructure and the data centre ecosystem. Saudi Arabia made significant moves, given the country’s sovereign wealth fund, the Public Investment Fund (PIF), which is making large-scale investments in AI. On Monday, the Crown Prince Mohammed bin Salman announced ‘HUMAIN’, a state-backed AI company, under the PIF. HUMAIN is set to provide a range of AI services, infrastructure, cloud capabilities, and advanced AI models in the country. The company plans to offer one of the world’s most powerful multimodal Arabic large language models (LLMs). The company formed multiple partnerships with companies, including a $600 million agreement with NVIDIA to create and construct ‘AI factories’ in the country. It also entered a partnership with Amazon Web Services (AWS), involving over $5 billion in investment to build an ‘AI Zone’—which includes AWS AI infrastructure, servers, semiconductors, and AWS services and applications. Meanwhile, AMD announced an agreement with HUMAIN to invest up to $10 billion over five years to establish 500 megawatts of AI compute capacity across a global network of AMD-based data centres, stretching from Saudi Arabia to the United States. Even Oracle announced a $14 billion investment over the next decade, to expand access to cloud and AI infrastructure. Super Micro Computer, based in the US, has formed a multi-year partnership valued at $20 billion with DataVolt to improve DataVolt’s hyperscale AI infrastructure in Saudi Arabia and the United States. And it isn’t just about Saudi Arabia. OpenAI, the company behind ChatGPT, is ‘considering’ building a new data centre in the United Arab Emirates (UAE), reported Bloomberg. Saudi Arabia and the United States forge a digital partnership uniting top global tech companies to boost the digital economy and position KSA as a leading hub for technology and AI. pic.twitter.com\/96MMcpzwfn— وزارة الاتصالات وتقنية المعلومات (@McitGovSa) May 13, 2025 Besides the recent announcements, there’s an exhaustive list of investments made in AI infrastructure and data centres in these Middle Eastern countries over the last few years. Why the Middle East? Middle Eastern countries boast a rapidly growing data centre ecosystem, with its capacity expected to increase from 1 gigawatt in 2025 to 3.3 gigawatts within the next five years, according to a PwC research report. “The Middle East offers unique advantages that make it a prime destination for data centre investments. These advantages include low-cost land, affordable power, attractive connectivity prices, access to capital and favourable foreign policies,” read the report. Also, the Middle Eastern countries have multiple government-backed initiatives that involve investing large amounts of capital into AI to diversify their economies beyond oil. For instance, last year, Saudi Arabia committed $100 billion to AI as part of its ‘Project Transcendence’. The MGX Fund, created by the Abu Dhabi government, is a $100 billion investment initiative for AI. It collaborates with BlackRock, Microsoft, and OpenAI to support AI infrastructure projects worldwide. It also participates in the Stargate Project with OpenAI and Oracle, concentrating on AI data centers and power infrastructure. The PwC report further added that governments across the Middle East are actively fostering a business-friendly environment for data centre investments. The report highlighted the case of Saudi Arabia’s Cloud Computing Special Economic Zone (CCSEZ), which was launched in 2023, offering tax incentives and simplified procedures to draw in foreign investment. By 2030, the CCSEZ will represent 30% of the country’s information and communication technology expenditures. The cost of land is also a crucial factor at play, particularly in Saudi Arabia, which is significantly lower than major global data centre hubs. “Industrial land in Saudi Arabia typically costs $10 to $50 per square metre, compared to $150 to $600 per square metre in US hubs like Northern Virginia,” said the report. Besides, Saudi Arabia and UAE are historically renowned for their abundant energy resources. PwC said that electricity tariffs in these countries range from $0.05 to $0.06 per kWh, well below the US average of $0.09 to $0.15 per kWh. Moreover, creating data centres is beneficial if they can connect to global networks through undersea submarine cable systems, and Middle Eastern countries have several of them, given that most of them lie on coastlines. This is also why Mumbai and Chennai occupy the first two spots in India’s data centre capacity by city, since they are situated on coastlines with vast undersea cable networks. In addition, most of the Gulf countries have always maintained strong international ties and favourable trade relationships with global economic centres. Trump’s visit to the country, which prompted a flood of investments, is a testament to this development. Additionally, reports surfaced that the U.S. President might take a favourable position against AI chip exports to Saudi Arabia, which were subject to restrictions from rules drafted by the previous Joe Biden government. However, cooling systems needed for data centres still remain a challenge. “While the Middle East, with its natural gas reserves and solar potential, is well positioned to address these demands, the region also faces water scarcity, which complicates cooling solutions,” said a Morgan Lewis report from March. The report also added that these countries are exploring solutions that involve innovative cooling technologies, such as seawater-based systems.","excerpt":"American giants like OpenAI, NVIDIA, AMD, AWS, Oracle are investing big on AI infrastructure in Saudi Arabia and UAE.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","saudi arabia"],"author_name":"Supreeth Koundinya","publish_date":"2025-05-14T18:30:00","publication_year":"2025","word_count":892,"keywords":["Go","ChatGPT","OpenAI","AI","saudi arabia","cloud computing","AWS","Git","RAG","CuPy","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","CuPy","RAG","cloud computing","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/data-centres-take-the-throne-as-the-new-crown-prince-of-the-middle-east\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10052943,"title":"Genpact Launches ‘Dare in Reality’ Hackathon: Predict Lap Timings For An Envision Racing Qualifying Session","content":"Genpact, in collaboration with Envision Racing and MachineHack, is launching a hackathon for data scientists and machine learning professionals called ‘Dare in Reality,’ from November 8 onwards. The hackathon winner will get a chance to win exciting cash prizes and goodies. Genpact is a global professional services firm that makes business transformation real. With a team of 90,000+ employees, Genpact serves more than 800 clients across 70+ countries. Forbes recently called Genpact out in its list of World’s Best Employers 2021. With this machine learning hackathon, Genpact aims to see how participants can help the Formula E team improve its performance on the racetrack in the all-electric, international single-seater world championship. The partnership between Genpact and Envision Racing uses data, innovation, real-time insights, and process excellence to achieve three core goals: boosting the team’s racing performance, climate change, and building and energizing a passionate, socially conscious fan base. Ready, Steady, Go! The two-week Dare in Reality Hackathon 2021 is for data science professionals, machine learning engineers, artificial intelligence practitioners, and other tech enthusiasts to showcase their skills and impress the judges. The winners are rewarded with cash prizes. The challenge starts on November 8, 2021 Click here to participate The Challenge is on… Besides disrupting motor racing and growing a diverse and global following, Formula E also fights climate change. It is creating a technical and sustainable development testbed that helps countries address mobility and environmental issues. In each race, 12 teams – each with two drivers – compete in electric, battery-powered cars. Every team is focused on securing the best competitive advantage to cross the finish line first. Formula E cars have evolved through years of research. And for a team to win, it needs a combination of driver skill and data analytics. But, Several factors affect their performance during a session, including: WeatherRain WindTrack temperature Ambient temperature  Track evolution: The way the track changes during and between sessions. As the cars leave rubber and debris on the tracks and weather affects the conditions of the driving environment – ambient temperature, humidity, rainfall, etc. – the track evolves and the time taken to complete a lap changes (increases or decreases based on positive or negative track evolution). Driver’s familiarity with the track So, in this hackathon, participants should use machine learning to predict lap times for a specific Formula E session – the qualifying session that determines where each car will be positioned at the start of the race. Submission Guidelines Sklearn models support the predict() method to generate the predicted values. You should submit a .csv file with exactly 1957 rows with 1 column(LAP_TIME). Your submission will return an Invalid Score if you have extra columns or rows. The file should have exactly 1 column. Note: Do not shuffle the sequence of the test series For participants using Pandas: submission_df.to_csv(‘my_submission_file.csv’, index=False) Evaluation Criteria: ​​The submission will be evaluated using the RMSLE metric. Participants can use np.sqrt(mean_squared_log_error(actual, predicted)) to calculate the same This hackathon supports private and public leaderboardsThe public leaderboard is evaluated on 30% of Test dataThe private leaderboard will be made available at the end of the hackathon, which will be evaluated on 100% of Test dataThe Final Score represents the score achieved based on the Best Score on the public leaderboard Prizes: First Prize: $7000Second Prize: $5000Third Prize: $300Fourth Prize: iPad Air 128 GBFifth Prize: Apple AirPods The challenge ends on November 22, 2021 Click here to participate Dataset Description: Train.csv – 10276 rows x 25 columns (includes target column as LAP_TIME)Test.csv – 420 rows x 25 columnsSample Submission.csv — Please check the “Evaluation” section on MachineHack Page for more details on generating a valid submission. Attribute Description: NumberDriver numberLap numberLap timeLap improvementCrossing finish line in pitSession 1 (S1)S1 improvementSession 2 (S2)S2 improvementSession 3 (S3)S3 improvementKphDriver namePit timeClassQualifying groupTeamManufacturerPowerLocationEvent Input data: Data from the practice and qualifying sessionsData tables:Timing Data: The standings for each driver in a race, lap-by-lap data for each driver for free practice (FP) 1, free practice (FP) 2 and qualifying session (Q)Weather: how weather conditions changed during a session Output: Provided data for FP1 and FP2 to build a model to predict the lap times for qualifying sessionsFor the test data, we will provide FP1, FP2 data (timing data and weather). Predict the lap times for qualifying sessions Skills: Multivariate RegressionBig dataset, underfitting vs overfittingOptimizing RMSLE to generalize well on unseen data Click here to participate Other details: Situation and Race Conditions All participants have the same car and battery\/power to consume during the sessionThe race runs for 45 mins (fixed) plus one additional lapThe driver who crosses the finish line first wins the raceIf there are crashes or problems on race day, the safety car enters the track until everything is safe. All drivers must stay behind the safety car, stay below 50kph, and cannot overtakeIf the safety car comes on, there will be fewer laps as the race time is fixedFor practice sessions, each driver can complete one lap in 250 kW, one lap in 235 kW and all remaining laps in 200 kWFor qualifying sessions, each driver can complete a maximum of 3 laps, where only one lap can be in 250 kW, and all others need to be 200 kWFor race sessions, laps are energy limited, as opposed to power limitedIn practice sessions, when drivers complete multiple adjacent laps with similar lap times, they generally practice race laps and, therefore, energy-limited lapsStandard session durations are as follows: FP1: 45 minutes FP2: 30 minutes Q1 to Q4: 4 minutes per group Super pole: 20 minutes maximum Race: 45 minutes plus one lap Suggested Approach: Data Preparation: Build a master data frame combining tables from different sessions, drivers, qualifying, and practice races. Join with weather data to get the final data setExplore data to understand the significant variables and generate derived variables if neededBuild a model to predict the lap times for each qualifying lap using the provided test data for a session Championship Basic structure Twenty-four drivers (two per team) compete in each Formula E raceEach driver ranks separately, but their total scores give the overall team rankingThere are 12 races in a season held over 8 monthsThere are 2 free practice sessions before a race  After the practice sessions, qualifying sessions run for a very limited number of laps to decide the pole position (who starts in the first row and then the following positions). A super pole can follow multiple qualifying sessions to decide the driver standings at the race startThe race runs for 45 minutes plus one lap Start Date: Nov 8, 2021 End Date: Nov 22, 2021 Click here to participate in this hackathon","excerpt":"Enter to build a model that predicts lap times for a qualifying session.","categories":["Deep Tech"],"tags":["data science hackathon","data science hackathon global","Envision Racing hackathon","Genpact","Genpact EVR Hackathon","Genpact hackathon","global data science hackathon","global hackathon grand prize","global machine learning hackathon","Hackathon","hackathon for data scientists","Hackathon for hiring data scientists","hackathon of the year","Hackathons India","Machinehack","Machinehack Hackathon","New Hackathon For Data Scientists","Weekend Hackathon","World Wide Hackathon Event"],"author_name":"Amit Naik","publish_date":"2021-11-08T11:00:00","publication_year":"2021","word_count":1109,"keywords":["TPU","global machine learning hackathon","Genpact","Weekend Hackathon","data science hackathon global","global hackathon grand prize","New Hackathon For Data Scientists","hackathon for data scientists","R","global data science hackathon","Pandas","data science","artificial intelligence","Envision Racing hackathon","Genpact EVR Hackathon","data science hackathon","analytics","Genpact hackathon","hackathon of the year","Go","machine learning","AI","Hackathon","Machinehack Hackathon","Hackathons India","Machinehack","World Wide Hackathon Event","Aim","Hackathon for hiring data scientists"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","Pandas","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/genpact-launches-dare-in-reality-hackathon-predict-lap-timings-for-an-envision-racing-qualifying-session\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":8091,"title":"Top 10 Analytics Courses in India – Ranking 2015","content":"A good data scientist should have a mix of knowledge in statistics and coding with an add-on of visualization and storytelling. There are several tools available in the market that caters to the field of analytics. To build up the right skillset it is imperative that a professional gets the right resources. It’s been our constant endeavor to bring to you the best in the field of analytics. Our Annual ranking is a step in this direction. Through our Annual ranking we bring to you the top analytics education institutions in the country by meticulously going through the institutes and their offerings. It is a complex process based on research to get the correct information about the required parameters and then the ability to use the information to present you the right outcome. Please note that this is a courses ranking by B-Schools in the country. We publish a separate Analytics training institutes ranking which is completely separate from this ranking. We acknowledge that slotting the training institutes with the B-School providing analytics courses in the same list is not right comparison. Thus, these are 2 separate and unique rankings. More importantly, we do not intend to rank these B-School in totality but just a specific analytics course by them. The institutes are ranked on the following parameters: Course content Pedagogy External collaborations Faculty Course Delivery\/Virtual Labs Placement Assistance, Events etc. This year we got more than 15 B-Schools taking part in this ranking. We take into account students as well experts feedback to carve out this cherry-picked ranking of just 10 courses. Read our latest ranking: Top Executive Analytics Courses in India – Ranking 2020 1. Post Graduate Program in Business Analytics – Great Lakes Institute of Management The Great Lakes Business Analytics program has been created in collaboration with corporate partners and senior professionals in analytics industry. The program has delivered around 1,75,000+ learning hours and the schedule and delivery is designed keeping in mind the time constraints and learning requirements of working professionals. The Great Lakes Post Graduate Program in Business Analytics equips candidates with the skill sets required for managerial, techno-functional roles in analytics. Its curriculum has been uniquely designed to meet these features and provides exposure to relevant tools like SAS, R & Tableau. The PGPBA program provides the right exposure to real world applications, ensuring that the professionals are equipped to apply their learning in the industry. The industry oriented pedagogy, hands-on exposure and highly acclaimed faculty help the candidates gain analytics competencies thereby preparing them for business and techno-functional roles in analytics. LOCATION The PGPBA program is currently being offered across three locations in the country: Gurgaon, Chennai, Bangalore. CLASSROOM LEARNING The program consists 230 hours of classroom sessions + 110 hours of online sessions delivered by Great Lakes faculties and industry professionals from the field of analytics. This ensures that the program imbibes Great Lakes academic elegance and industry’s business relevance, thereby providing the candidates with a remarkable learning experience. ONLINE – LEARNING MANAGEMENT SYSTEM All candidates have access to the online LMS that hosts content (classroom recording, discussions forums, assignments, reading material) and live webinar to enable the candidates continue their learning during off campus. The LMS provided and innovative learning environment that encourages collaborative approach between the candidates thus paving the way for maximizing learning effectiveness. EXPERIENTIAL LEARNING This program is designed to transform candidates to business ready analytics professionals through hands on experiential learning on relevant tools. This is achieved through an experiential learning format wherein participants practice exercises and assignments on software package such as SAS, R and Tableau. CAPSTONE PROJECT All candidates would be pursuing an industry project in the field of Business Analytics. The project is mentored and jointly evaluated by faculty from Great Lakes and Industry leaders. The project is presented to the faculty board as part of the requirement for successful completion of the program. 2. Certificate Programme on Business Analytics and Intelligence – IIM Bangalore The course is designed to provide in-depth knowledge of handling data and Business Analytics’ tools that can be used for problem solving and decision making using real case studies. The 1 year long duration program consists of eight modules and a project. The duration of each module is usually 5-6 days except for module 2 (2 days) and module 7 (2 days). Students are expected to do a group project as part of this course based on a real-life problem\/data. The project pre-work should start around October and should roll out by January. It will be supervised by an IIMB faculty member and must be wrapped by May with a project report submission. The Institute encourages students to publish cases studies based on their course project. At the end of the course, the participants will be able to: Understand the emergence of business analytics as a competitive strategy. Understand the foundations of data science; the role of descriptive, predictive and prescriptive analytics in firms. Analyze data using statistical and data mining techniques and understand relationships between the underlying business processes of an organization. Learn data visualization and storytelling through data. Learn decision-making tools \/ Operations Research techniques. Use advanced analytical tools to analyse complex problems under uncertainty. Manage business processes using analytical and management tools. Use analytics in customer requirement analysis, general management, marketing, finance, operations and supply chain management. Learn analytics through case studies published by IIMB at the Harvard Business Publishing Understand sources of Big Data and the technologies and algorithms for analyzing big data for inferences. Ability to analyze unstructured data such as social media data and machine generated data. 3. Executive Program in Business Analytics (EPBA) – MISB Bocconi One of Europe’s top b-schools, SDA Bocconi, and Jigsaw Academy, launched a 10-month Executive Program in Business Analytics (EPBA) for professionals earlier this year. This was a unique combination of management principles backed by analytics training. SDA Bocconi is currently ranked at #7 in the world for its MBA program, and Jigsaw Academy is ranked #1 in India for its analytics training courses. The program will be conducted at Bocconi University’s campus in Mumbai. Upon completion, the participants will receive certification in business analytics from the SDA Bocconi School of Management. The course provides a deep understanding on all relevant disciplines of business analytics, including statistics, machine learning, time series, R, SAS, Big Data (Pig, Hive, Sqoop, Flume, HBASE, SPARK and Oozie), visualization, text mining, web analytics and digital marketing. In addition, it also focuses on its application across sectors and functions including Telecom, Banks, Retail, Healthcare and Insurance as well as areas such as Finance, Marketing, and Operations. The programme involves more than 280 hours of training, including 120 hours of in-person training held over six three-day modules at the MISB Bocconi campus in Mumbai. In the interim, Jigsaw Academy also conducts 24 live online classes for a total of 60 hours, which participants can attend from any convenient location. In addition to the live online and in-person classes, participants will also have access to over 100 hours of pre-recorded video lectures on data science and big data analytics for 12 months. 4. Post Graduate Program in Business Analytics – Praxis Business School The program has 9 months of classroom training followed by campus placements. The curriculum with 500+ hours of classroom interaction with additional hours of case studies, projects and assignments focuses on offering a comprehensive analytics experience in the following: Technology Skills: an understanding of tools like Excel, SQL, R, SAS, Hadoop, Python, MongoDB (NOSQL), Amazon AWS (EMR \/ EC2), QlikView and Tableau that are commonly used to extract, analyse and visualize data Analytical Skills: knowledge of statistical modelling, data mining and machine learning techniques and the ability to create predictive models Applying analytics to business situations: an understanding of business functions like marketing, finance, operations and the application in verticals like BFSI, retail, web, telecom Communication and Visualization Skills: the ability to tell an effective story Industry collaborations This industry supported program is co-created and co-delivered with knowledge supporters ICICI Bank and PwC. 35% of the courses are delivered by practitioners from organizations like PwC, ICICI Bank, Abzooba, Ericsson Global, HP and IBM. Praxis Placement Program Praxis takes ownership of facilitating internships and final placements upon successful completion of the program. Recruiters include Abzooba, Axis, BRIDGEi2i, Genpact, Happiest Minds, HDFC, Hewlett Packard, HSBC Analytics, ICICI, ICTEAS, IDC, ITC Infotech, Karvy Analytics, Kie Square, Millward Brown, PwC, ValueLabs etc. Praxis Analytics alumni have transitioned into companies like Cardekho.com, Dell Analytics, Fidelity, IBM, Lowe’s, Quikr.com, Target, TCS, Walmart Labs, WNS and are performing with distinction. 5. Post Graduate Program In Business Analytics & Big Data – Aegis Aschool of Business Aegis Post Graduate Program in Business Analytics & Big Data is India’s first true Data Science program of globally acceptable 45 credit unit offered in association with IBM. This program is designed by IBM’s subject matter experts & leading Data Scientists for developing highly skilled business Analyst, Data Scientist, Big Data Advisors, consultant, big data developers etc. This program provides hands on exposure to participants on various Big Data tools, real-life Big Data and Data Science projects at Aegis Big Data Product Factory, IBM Business Analytics and IBM Cloud Computing Lab. Program Highlights: Certification from IBM at the completion of the course. IBM has setup an IBM Business Analytics Lab. This Program is core technology MS program and NOT an inferior cousin of MBA\/PGDBM programs in Business Analytics. Unit spread over 11 months that includes 2 month Program gives you freedom to choose career in different areas like: Big Data Advisory\/ Consulting; Technology\/ Development; Pre Sales\/ Enterprise Sales Tools, S\/W and Platforms: Hands on exposure on IBM DB2, IBM Cognos TM1, IBM Cognos Insight, IBM InfoSphere Big Insight, IBM Worklight, IBM BlueMix, R, Python, SAS, Hadoop, MapReduce, Spark, EC2, ElasticSearch, Tableau, EMR, AWS, Weka etc Live Projects: work on Capstone projects, consulting assignments and live projects from industry like Churn Predication, Call drop analysis from leading teleco; NLP projects, consumer analytics, sentiment analysis etc at Aegis Big Data Product Factory as part of the program. The curriculum caters to the various skill requirements of eCommerce, Telecom, Banking, Computer Services, Education, Healthcare, Insurance, Manufacturing, Retail, Automobile etc. Cloud Learning Management System: Online classes is conducted through world class LMS on cloud called mUniversity.mobi All video lectures, study material, ebooks, business cases available on LMS. Course Modules taught by IBM: (i) Business Intelligence using Cognos BI (ii) Big Data Analytics using IBM InfoSphere Big Insight and (iii) Enterprise Performance Management using IBM Cognos TM1 Scholarships: Deans Scholarship for Women in Technology Leadership; Aegis Graham Bell Award Scholarship for Big Data Products for startups; Data Science. These scholarships will provide 100% funding for tuition fee. 6. Post Graduate Certificate in Predictive Business Analytics – Bridge School of Management BRIDGE School of Management is a flagship business school launched via a joint venture between HT Media Ltd. & Apollo Global, Inc. (USA). Apollo Global (www.apolloglobal.us) is one of the world’s leading higher education providing companies. The certificate program has been specially created for India by academicians from Northwestern University and top industry experts using real-world problems and situations. The Northwestern and Bridge School initiative combines online content developed and taught by Northwestern faculty with weekly in-person sessions led by local specialist faculty at the Bridge School’s learning centers. Program Objectives Apply analytics tools to real-world business contexts for improved decision making Assess the strengths and limitations of analytics and predictive modeling techniques for different business applications and varying data conditions Acquire hands-on experience working with leading statistical tools and software packages (such as R) in predictive modeling and the visual analysis of results Effectively communicate the actionable insights stemming from analytical work to multiple stakeholders Strategically navigate technology tools and trends to solve big data and analytics problems Manage data strategies and analytical projects. 7. Certificate Program in Big Data and Analytics (BDAP) – SP Jain School of Global Management The course is being offered at S P Jain’s spanking new campus at Lower Parel, which boasts of State of the art facilities, in the heart of the city’s business center. BDAP is designed to explore, analyze and unravel the complex, unstructured data-driven world. The program kicks of with 10 core courses that build a strong foundation for the second stage of the program, which incorporates more in-depth and application-based learning. Given the need for specialist knowledge, it provides a range of courses in topics like data mining, machine learning, visualization techniques, predictive modeling, and statistics. The programme content builds on basic concepts, teaches tools and technologies that are currently prevalent in industry and progresses to cutting edge-topics like machine learning and Natural Language Processing. On completion of the program students would have learned to apply quantitative modeling and data analysis techniques to solve real world business problems, successfully present results using data visualization techniques, demonstrate knowledge of statistical data analysis techniques utilized in business decision-making, apply principles of Data Science to the analysis of business problems, use data mining software to solve real-world problems and employ cutting edge tools and technologies to analyze Big Data. The program is delivered by a faculty that is an equal mix of academicians and industry practitioners with a key proportion of overseas instructors thus lending the course a global perspective. 8. 2-year Full time PGDM Programme with Analytics Specialization – Narsee Monjee Institute of Management Studies NMIMS, Bangalore offers industry leading specialization in Analytics for its PGDM programme. Students in the second year of the programme can opt for this specialization. It also offers Marketing, Finance, Operations and HR as other specializations. The second batch of this specialization is running now. The first batch was well received by industry and the students are absorbed in organizations like Citibank, Infosys, mu-sigma, GENPACT, iGate, Netapp, Fidelity etc. Through this specialization, NMIMS trying to educate Analytics professionals who are well equipped with Management functions and Data Science. The coverage of the specialization is extensive as it got Tools, Techniques, Functional and Industry related courses. Course Content The Analytics specialization stream consists of twelve (12) courses (and a workshop) spread across 3 trimesters in the 2nd year of the PGDM program. It is expected to cover the knowledge areas expected of an Analytics professional viz. tools, techniques and functions. Additionally, a course titled ‘Business Analytics for Decision Making’ is compulsory for all students in the institute. Pedagogy The programme is conducted live in the class room at the campus at Bangalore. The delivery consists of classroom lectures and interactions, case discussions, workshops and analysis of live industry problems. The institute got license to SAS software and hence, most of the data analysis and modelling is conducted on this platform. 9. Executive Program in Business Analytics and Business Intelligence – IIM Ranchi “Executive Program in business analytics and business intelligence (Saturday to next Sunday with one week leave)” is specially designed to provide inputs which will equip the participants with analytical tools and prepare them for corporate roles in analytics-based consulting. These inputs will provide a basis for the participants to channelize their analytical thinking in appropriate directions, besides, enhancing knowledge. The skills so acquired may be effectively utilized in their day-to-day work and thereby promoting the quality of business decisions. Objectives of the Programme To enable the participants to understand and use the tools and techniques for business analytics and business intelligence. To enable the participants to make use of large volume of data for meaningful business decisions and strategy To impart hands-on-experience with various softwares, like, (i) R, (ii) SAS (iii) Python (iv) SPSS Pedagogy of the Programme The participants will learn the concepts and implications of business analytics & business intelligence through class room lectures, interactive discussions, case studies and hands-on-experience. Both conceptual and practical sides will be addressed. Participants Profile The course is suitable for those with analytical aptitude and would like to start new career in analytics. The course is also appropriate for those who are working in business analytics and business intelligence to enhance their knowledge and skill. 10. PGDM with Specialization in Business Intelligence and Big data – IMT Ghaziabad This is a unique course which prepares you for the world of work in analytics in companies. Students having opted for this course in the previous years have found placement offers from leading companies, the designations include Data analyst, Business development Manager, and Business Analyst and research analyst. The course gives you a blend of industry knowledge, concepts and experiential learning through collaborative teaching by industry experts. Take the first step in joining the course; we will then help you to complete the journey into a person well-versed with the art and science of analytics. Pedagogy The pedagogy will be a mix of lectures, experience sharing, real life case discussion, assignments and industry\/research based projects. The course is focused on strategic issues with cases as the primary vehicle for learning. In addition to the reading materials, additional readings and cases will be distributed in the class from time to time. Students are also expected to prepare and analyze all the cases as class participation is very important. There are four main pillars in the course pedagogy, namely, (a) lectures-cum-PPTs to share the conceptual frameworks; (b) experience sharing through collaborative teaching by industry experts; (c) hands-on Statistica Data mining tools, computer-lab based; and (d) case studies of leading organizations selected from Harvard Cases and other sources. The course requires a high degree of interactions in the class on part of the students and feedback on their hands-on work is provided by industry experts. View & Download the complete report [attachments include=”8097″] [poll id=”9″]","excerpt":"A good data scientist should have a mix of knowledge in statistics and coding with an add-on of visualization and storytelling. There are several tools available in the market that caters to the field of analytics. To build up the right skillset it is imperative that a professional gets the right resources. It’s been our […]","categories":["AI Features"],"tags":["analytics education","Courses","great lakes analytics","jigsaw academy","Praxis Business School"],"author_name":"Дарья","publish_date":"2015-10-20T16:03:45","publication_year":"2015","word_count":2944,"keywords":["data science","machine learning","AWS","AI","sentiment analysis","cloud computing","RAG","great lakes analytics","NLP","Aim","analytics","Courses","jigsaw academy","analytics education","Praxis Business School"],"extracted_tech_keywords":["AI","machine learning","NLP","data science","analytics","Aim","RAG","sentiment analysis","cloud computing","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-analytics-courses-in-india-ranking-2015\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10048851,"title":"Ex-Apple Employee Exposes Apple M1 Chip&#8217;s Secrets","content":"One of the developers behind Apple’s QuickTime and ex Cornell Physics educated, Maynard Handley, recently shared a 350 page PDF analysing and explaining the inner workings of the infamous M1 ARM chip. The document, currently published as version 0.7, dives deep into M1’s architecture from a reverse-engineering perspective and is heavily being fed by other professionals and enthusiasts in the field. The researcher’s PDF goes into excruciating details on the technical elements of the processor, and it can easily be used by developers and budding engineers to reverse-engineer the chip. It has contributions from the team that ran Linux on the M1 and contributions from independent testers. The document even includes instructions on how to run your own testing. https:\/\/twitter.com\/handleym99\/status\/1437537535018684417?s=20 Apple’s M1 chip stunned the world of technology with its incredible IPC and power-to-performance ratio, shifting the spotlight from x86 as the new high-performance architecture. But, Apple’s walled-garden approach to its tech only ensures that anyone who wants to take advantage of the M1 hardware has to either go through Apple’s macOS or follow the only remaining option of a reverse engineering route optimising another software stack to use on the Apple silicon. This also means that we could potentially see hyper-efficient and performance heavy ARM CPUs for Linux and even for Windows in future. A project like this could lead to an open-source M1-type chip, which could unlock all new paths for independent engineers and startup teams. An open-source chip would make it only easier to manufacture competing hardware, from ARM-based laptops all the way to iPad Pro-level tablets. It is still unclear if Apple agrees with this information being readily available for free on the Internet. Although the document is nowhere near a definitive version and yet to reach full completion as there are still some unknown inner workings, more and more people from the community are contributing to it every day.","excerpt":"The researcher’s PDF goes into excruciating details on the elements of the processor, and it can easily be used by developers and engineers.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Apple","Chip shortage","Data Science","Machine Learning"],"author_name":"Victor Dey","publish_date":"2021-09-20T12:02:07","publication_year":"2021","word_count":314,"keywords":["Go","programming_languages:R","AI","Apple","Machine Learning","programming_languages:Go","Chip shortage","Data Science","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ex-apple-employee-exposes-apple-m1-chips-secrets\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":65220,"title":"How To Network Better At An Online Conference","content":"Networking is one thing that distinguishes an in-person event from virtual conferences, where in-person events provide numerous opportunities for attendees to reach out to speakers, business leaders, and potential employers, virtual conferences provide less convenience for attendees to gather and interact. However, plugin is one such online conference on artificial intelligence and data science that features all the aspects of an in-person event, even networking. plugin is a distinctive online conference, hosted by AIM, that aims to bring in the experts from the data science industry to talk about the competitive field as well as the cutting-edge innovations. Scheduled on 28-29th of May 2020, plugin will host more than 1000 attendees from 200+ organisations and will have 70+ speakers across three tracks talking about the industry and the latest developments in the field of artificial intelligence, computer vision, cybersecurity etc. plugin, not only allows attendees to learn from the best brains of the industry but also provides ample opportunities to network with speakers, sponsors, exhibitors as well as other attendees. Here, we will share a few ways you can network better at the plugin. Book your tickets here. Live Q&A At plugin One of the main advantages of plugin is to provide the experience of an in-person event in a virtual setting at the comfort of your home. With plugin’s live question and answers’ session, attendees can easily connect with speakers and panellists after their talk ends. These live Q&A will provide attendees with the opportunities to better understand complex topics by interacting with leaders and speakers about their subjects. During these live sessions, attendees can also exchange ideas and knowledge as well as foster debate with other attendees. In plugin, attendees can even live stream the entire session remotely or on-demand once the session ends. Live Q&A sessions also provide opportunities for attendees to leave feedback on a session, which will help leaders, speakers and panellists to improve their talks. Speaker Lounge plugin also features a ‘speaker lounge’ where attendees can connect with other attendees during and in-between sessions. Along with that, attendees can also uncover speakers’ profiles prior to their session, view speakers’ details, speakers’ presentations, and information about their sessions. These information are easily available during in-person events, however, it becomes a challenge when you are attending an event remotely, but plugin’s speaker lounge makes it easy for attendees to avail all these information on one platform. At plugin, with the help of the speaker lounge, attendees can also rate speakers and provide feedback, bookmark them and take notes, just like in-person conferences. Book your tickets here. Match-Making & Networking Another exciting feature of the plugin event is the match-making and networking possibilities, which allows attendees to view, search, filter and sort other attendees, just like in-person events. This feature provides opportunities to interact with all attendees at the conference. Taking it up a notch, plugin also enables guests to bookmark attendees to talk to them later; this feature can also be utilised by speakers to discuss potential collaboration with specific guests. Alongside, plugin allows attendees and guests to set up meetings at available time slots of an attendee or a speaker, and also have the facility to accept, reject as well as reschedule the meetings. This feature helps attendees to have a smoother networking process with other attendees. In order to stay abreast about the event, plugin also suggests top 10 attendees that one should meet to grow their knowledge base. Chat Rooms The feature of chat rooms at plugin has been designed to completely transform the virtual conference into a more meaningful in-person experience; and also an essential feature. These chat rooms at plugin allow attendees to do impromptu meetings by grabbing a seat at the virtual networking lounge. The chat rooms allow four people for a single meeting, which helps speakers and panellists to address more than one attendees at a time. At plugin, attendees can also join in an existing chat room with a click of a button to understand discussions better. plugin also has dedicated chatrooms for sponsors and exhibitors. Attendees can also create one-on-one chat rooms to have a more personalised session with speakers and business leaders. This can be an excellent opportunity for attendees to interact with potential employers for future job prospects. Book your tickets here. Contests Another way attendees can network better at plugin, is by joining in the contests and hackathons that are being organised during the conference. plugin will host all-night machine learning hackathons that will help attendees to showcase their skills and network with other participants and jury members. Analytics India Magazine is known for organising online hackathons — MachineHack, and therefore, will provide immense opportunities for attendees of plugin to join in these events. Alongside, plugin will also host different contests and will allow all attendees to choose a winner basis engagement, which will later be showcased on the app. These contests and hackathons provide a unique platform to interact, share knowledge as well as build a strong bond with peers. To register for plugin, click here. Join plugin today to learn from industry experts sharing fresh and practical insights on artificial intelligence and data science!","excerpt":"Networking is one thing that distinguishes an in-person event from virtual conferences, where in-person events provide numerous opportunities for attendees to reach out to speakers, business leaders, and potential employers, virtual conferences provide less convenience for attendees to gather and interact. However, plugin is one such online conference on artificial intelligence and data science that […]","categories":["Deep Tech"],"tags":["networking","plugin","plugin aim","plugin online conference","plugin virtual data science event","Work from Home"],"author_name":"Sejuti Das","publish_date":"2020-05-14T17:00:00","publication_year":"2020","word_count":862,"keywords":["plugin online conference","plugin virtual data science event","data science","networking","plugin aim","artificial intelligence","machine learning","AI","innovation","Work from Home","computer vision","Aim","analytics","plugin","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","data science","analytics","Aim","R","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-to-network-better-at-an-online-conference\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":31419,"title":"What Facebook’s Internal Documents Released By UK Parliament Tell Us About Its True Identity","content":"Earlier this December, the UK Parliament committee released documents revealing internal emails within Facebook. These emails have shocked the world and the embattled social media giant has once again become the target of UK government after Cambridge Analytica controversy. What Do The Documents Show? The documents with 250 pages contain internal emails between Facebook’s executives. The documents were a part of the evidence of a lawsuit which was filed against Facebook by Six4Three. The documents, noted by Damian Collins, Chair of Digital, culture, media and sports committee (DCMS) show that Facebook indulged in malpractices after its platform changes in 2014-15. The company is guilty of compromising users’ data to app developers. It is alleged that Facebook leaked user data to companies like Airbnb and Netflix. It was in whitelisting agreements with many companies, it also has the access to record all the calls on an Android system. The company is guilty of taking aggressive actions towards its competitors, because of this policy many small-scale companies couldn’t progress. Facebook gained complete access to friends’ data after the launch of Platform 3.0, it took advantage of the data and sold it to many partners and friend companies. The company understood users’ activities and collect usage data of its competitors. It then gave access to those companies which were preferred by users and put restraints on other or rival companies. A total of 27,019 apps were affected because Facebook denied them access to its’ platform and to data. Some of them even shut down because of Facebook. How Facebook Broke Privacy Laws Collins, Member of Parliament, said in an official statement, “We need a public debate about the rights of social media users and the smaller businesses who are required to work with tech giants.” The committee investigation on Facebook and its malpractices came to a point where they didn’t receive any clarification from the company regarding all the “data breach” allegations.hence, they released the “internal email” documents. Many other DCMS committee members supported Collins since personal data has been at risk since years. To generate revenue through different systems and make more profit from partners, Facebook broke privacy laws and put users’ personal data into peril. Politicians across the globe are angry. Officials from Australia and America are either telling people to leave the app or want Facebook to make amends as soon as possible. Google trends show an average of 70 points on the search  “#stopfacebook”. Richard Blumenthal, US Senator said, “Mounting evidence Facebook acted chaotically, recklessly, and lawlessly by granting access to private consumer data for financial gain. These new documents show clearly that Facebook failed to heed their consent decree agreement and basic standards of privacy.” Mark Zuckerberg Ducks The Debate Mark Zuckerberg took to his Facebook account to defend himself. Although his statement gave no clarification about the data breach, he explained how his company didn’t indulge in breach of ethics, he said, “After the launch of the Facebook platform in 2007, many shady apps abused users data. To prevent such measures, a new platform was launched in 2014 to limit data apps on the platform. Only those abusive apps were removed from the platform.” Zuckerberg is accused of acquiring “Onavo”, an analytics company which browsed users data to retrieve data about rival companies’ performance. To his defence, Zuckerberg added, “We faced an issue of making our platform economically sustainable, so we continued to provide the developers’ platform for free and they could choose ads which they wanted. Our model encouraged a lot of internal discussions and we have focused on preventing abusive apps for years.” He also said that the change in platform was necessary to protect the community and don’t misinterpret our actions and motives.","excerpt":"Earlier this December, the UK Parliament committee released documents revealing internal emails within Facebook. These emails have shocked the world and the embattled social media giant has once again become the target of UK government after Cambridge Analytica controversy. What Do The Documents Show? The documents with 250 pages contain internal emails between Facebook’s executives. […]","categories":["AI Features"],"tags":["documents"],"author_name":"Jignasa Sinha","publish_date":"2018-12-12T12:45:51","publication_year":"2018","word_count":619,"keywords":["Go","AWS","AI","cloud_platforms:AWS","documents","programming_languages:R","Git","RAG","ViT","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","AWS","R","Go","Git","ViT","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-facebooks-internal-documents-released-by-uk-parliament-tell-us-about-its-true-identity\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10015829,"title":"Top 10 Alternatives To Apache Spark in 2024","content":"Apache Spark is an open-source unified analytics engine for large-scale data processing. With more than 28k GitHub stars, this analytics engine can be said as one of the most active open-sourced big data projects and is popular for its various intuitive features. Some of its features include ease of writing applications quickly in various languages, such as Java, Scala, Python, R, and SQL and accessibility in diverse data sources. Below here is a compilation of the top eight alternatives to Apache Spark. 1. Apache Hadoop Apache Hadoop is a framework that allows distributed processing of large data sets across clusters of computers using simple programming models. The framework is designed to scale up from single servers to thousands of machines, each offering local computation and storage. Apache Hadoop has its own file distribution system known as the HDFS (Hadoop Distributed File System). The file storing system is typically used for organising the files. 2. Google BigQuery Google BigQuery is one of the cloud-based big data analytics web services for processing very large read-only data sets. It is Google Cloud’s fully managed, petabyte-scale and cost-effective analytics data warehouse that lets developers run analytics over vast amounts of data in near real-time. 3. Apache Storm Apache Storm is an open-source distributed real-time computation system. Developers use this system mainly to process streams of data in real-time. Apache Storm has many use cases, including real-time analytics, online machine learning, continuous computation, distributed RPC, ETL, and more. Storm integrates with the database technologies; and its features include scalability, fault-tolerance as well as guarantees that the data will be processed in an easy manner and is simple to set up and operate. 4. Apache Flink Apache Flink is a framework, and a distributed processing engine meant for stateful computations over unbounded and bounded data streams. The framework has been created to run in all the common cluster environments and then perform computations at the in-memory speed at any scale. Flink can be used to develop and run many different types of applications due to its extensive features set. Some of its key features include support for stream and batch processing, sophisticated state management, event-time processing semantics, and exactly-once consistency guarantees for the state. 5. Lumify Lumify is a popular big data fusion, analysis, and visualisation platform that supports the development of actionable intelligence. This big data tool enables users to discover complex connections and explore diverse relationships in their data through a suite of analytic options, including graph visualisations, full-text faceted search, dynamic histograms, interactive geospatial views, and collaborative workspaces shared in real-time. It is a tool that empowers intelligence analysts to make the quick, informed decisions that our national security demands. 6. Apache Sqoop Apache Sqoop is a tool designed for efficiently transferring bulk data between Apache Hadoop and structured datastores such as relational databases. Basically, it is a tool that is designed to transfer data between Hadoop and relational databases or mainframes. Developers can use Sqoop to import data from a relational database management system such as MySQL or Oracle or a mainframe into the Hadoop Distributed File System (HDFS), transform the data in Hadoop MapReduce, and then export the data back into an RDBMS. 7. Elasticsearch Released in 2010, Elasticsearch is a popular, distributed, open-source search and analytics engine for all types of data, including textual, numerical, geospatial, structured, and unstructured. It is built on Apache Lucene and known for its simple REST APIs, distributed nature, speed, and scalability. The speed and scalability of Elasticsearch can be used for infrastructure metrics and container monitoring, application performance monitoring, geospatial data analysis and visualisation and more. 8. Presto Presto is an open-source distributed SQL query engine for running interactive analytic queries against data sources of all sizes. The engine was designed and written from the ground up for interactive analytics and approaches the speed of commercial data warehouses while scaling to organisations like Facebook. Facebook uses Presto for interactive queries against several internal data stores, including their 300PB data warehouse. 9. Snowflake Snowflake is a better alternative to Apache Spark for certain use cases due to its fully managed cloud-based platform. It automatically scales compute resources as needed, simplifying performance management and making it easier to handle varying workloads. Snowflake offers a pay-per-use model, charging only for the resources you use, which can result in cost savings, especially for variable workloads. Apache Spark can be more expensive due to its always-on infrastructure needs. Snowflake also provides built-in security and compliance features, making it a better fit for organizations with strict regulatory requirements, whereas Spark’s security setup depends on the platform it’s running on and often requires additional configuration. 10. Dremio Dremio is a strong alternative to Apache Spark due to its architecture and focus on simplifying data access and analytics. Dremio, eliminates the need for manual ETL (Extract, Transform, Load) processes by providing a self-service platform that enables users to query data directly from sources like cloud data lakes in real-time. Its integration with Apache Arrow and the use of in-memory acceleration leads to faster query speeds without the heavy computational overhead associated with Spark. Another key advantage of Dremio is its ease of use. Dremio’s intuitive interface allows non-technical users to explore and analyse data. Dremio also offers better cost efficiency by reducing data movement and optimising queries with features like data reflections, which cache query results to avoid repeated processing. These factors make Dremio a more user-friendly and cost-effective solution for businesses aiming to streamline data analytics without sacrificing performance.","excerpt":"Apache Spark is best for large-scale distributed data processing and real-time analytics. Here are 10 alternatives to Apache Spark you can consider for large scale data processing.","categories":["AI Trends"],"tags":["Apache Spark","apache spark replacement","Mapreduce","Types of Databases"],"author_name":"Ambika Choudhury","publish_date":"2020-12-24T12:00:00","publication_year":"2020","word_count":918,"keywords":["Elasticsearch","machine learning","AI","Types of Databases","ML","Apache Spark","Mapreduce","RAG","Python","Aim","analytics","apache spark replacement","Snowflake"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","Aim","RAG","Apache Spark","Elasticsearch","Snowflake","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-alternatives-to-apache-spark\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10165894,"title":"‘AI Will Be Writing 90% of Code in 3-6 Months,’ Says Anthropic’s Dario Amodei","content":"Dario Amodei, CEO of Anthropic, while speaking at a forum of The Council on Foreign Relations (CFR), said that he believes 90% of coding will be done by AI in less than 6 months. When Mike Froman, the president of CFR, asked about the best case scenarios AI will bring us, Amodei stated, “If I look at coding, programming, which is one area where AI is making the most progress. We are not far from the world where, in 3-6 months, AI is writing 90% of the code.” He added that in 12 months, AI might be writing essentially all of the code. While that sounds alarming, he added that the programmer still needs to specify the conditions of what one wants to do, the idea of the app one wants to make, and the design decisions. While he believes that human programmers’ intervention will still be needed to decide granular aspects of coding, AI will eventually do everything humans can. Amodei expressed his thoughts, encouraging one to look at ‘usefulness’ and ‘uselessness’ in a different way than before. He noted that AI making everyone useless is an annihilistic thought and emphasised that he is optimistic about building a world where human life is meaningful while taking the help of AI. Froman also asked Amodei if DeepSeek was a ‘Sputnik moment’. Amodei commented that DeepSeek was nothing unusual and another data point on the cost reduction curve. A future where everyone would be a coder is getting closer with each passing year – or month, maybe. In a podcast with Lex Fridman, when asked how much programming people would do in the next 5-10 years, OpenAI CEO Sam Altman said, “A lot, but I think it’ll be in a very different shape.” Altman said that many have already started programming entirely in natural language. “No one programs by writing code…some people do. No one programs the pun cards anymore,” he quipped, adding that it would change the nature and the skillset, not so much the predisposition for who we call programmers in the future.","excerpt":"The CEO of Anthropic suggests that AI could write essentially all of the code in 12 months.","categories":["AI News"],"tags":["AI coding"],"author_name":"Ankush Das","publish_date":"2025-03-12T11:33:31","publication_year":"2025","word_count":344,"keywords":["Anthropic","AI coding","OpenAI","AI","programming_languages:R","RAG","R","Redis"],"extracted_tech_keywords":["AI","OpenAI","Anthropic","RAG","Redis","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ai-will-be-writing-90-of-code-in-3-6-months-says-anthropics-dario-amodei\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10009590,"title":"Embold Is Like Autocorrect For Code, Says Vishal Rai, Founder &#038; CEO","content":"In our digital world, software is king. In a world so heavily dependent on software, poor code quality can result in grave consequence, from billions of dollars in lost revenue, to even fatal accidents. Here’s where Embold comes in—a static code analysis product aimed at empowering developers to do their best work. Embold is a general-purpose static code analyser that has been designed for developers to analyse and improve their code by identifying issues across four dimensions, including design and duplication. We, at Analytics India Magazine, spoke to founder and CEO, Vishal Rai, to understand how Embold can detect anti-patterns in code for seamless integration. Embold started a decade ago, with the vision of creating a product that can revolutionise the way developers write and design code. According to Vishal Rai, the idea was to develop a tool for software engineers and developers to write code faster and of better quality. And, after a time of extensive research and development, Vishal Rai, along with his partner Sudarshan Bhide launched their product in 2018. “We have noticed an interesting trend — as teams started becoming bigger, the issues in software started increasing as well and it was very frustrating when you were on programs which weren’t achieving their stated goals because of poor quality,” said Rai. “And that’s where we saw the opportunity of helping companies to write the product as great as Google or Apple and decided to reinvent software analytics.” Embold — Empowering Developers to Reach Their Highest Potential Developers always undergo immense pressure of building their products faster at the best quality possible, and such pressure can lead to compromised code quality. This impact of one line of code or one weak link can create significant issues and can massively affect the entire company. And that is why Rai believes that developers need support in being more productive. With Embold, Vishal and Sudharshan brought in a technology that can help developers be more efficient in their work and make the process of software development easy. Explaining the technology, Rai compared it with “autocorrect for code.” He said, “If you look at the legacy tools, they were built for the waterfall model, aka linear-sequential life cycle model, where one release took six months which gave enough time to test the tools. But in the current time, everything is fast, and thus developers require tools that can help them work fast and give them feedback that can be easily implemented in their workflow.” And that’s where Embold’s platform fits into the workflow that helps them find problems and maintain their code. As a matter of fact, Rai acknowledges that there have been many great tools, already in the market, but all of them have been created for great engineers. However, today, not every engineer is necessarily as experienced as others, and in such cases, it is imperative to make tools that are easy to use and help developers analyse their software. “Embold has been built for the future, the technologies that we have ingrained have neural networks, state-of-the-art in-memory databases and analysers that are far more evolved than legacy tools,” said Rai. “Embold has been created to enable people who aren’t as skilled to write better codes. Not only it fills the skills gap but also brings the new age developers closer to the best developers on the planet.” Wrapping Up Thus, with these pointers in hand, it can be inferred that regardless of languages, frameworks or development procedures, implementing high-quality code is exceptionally important in developing a seamless software application. With Embold, companies can get the best out of their developers by filling the skills gap and making it a level playing field for amateurs, as well as the ones who are already experienced. “Our platform helps developers write great code without worrying about making mistakes. With its highly responsive visual UI, developers can understand their code base better and with our component-level issue flagging, developers will know exactly where to start their fixes.’ concluded Rai.","excerpt":"In our digital world, software is king. In a world so heavily dependent on software, poor code quality can result in grave consequence, from billions of dollars in lost revenue, to even fatal accidents. Here’s where Embold comes in—a static code analysis product aimed at empowering developers to do their best work. Embold is a […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Sejuti Das","publish_date":"2020-10-13T14:00:39","publication_year":"2020","word_count":667,"keywords":["Go","programming_languages:R","AI","neural network","ML","Git","Aim","analytics","Rust","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","neural network","analytics","Aim","R","Go","Rust","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/embold-is-like-autocorrect-for-code-says-vishal-rai-founder-ceo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10131831,"title":"How Digital Twin Can Accelerate India&#8217;s Semiconductor Ambitions","content":"In the recent Union Budget 2024, the central government increased the estimate for setting up semiconductor fabs in India from INR 12.51 crore to INR 1,500 crore. This is in addition to the INR 6,903 crore earmarked for semiconductor and display manufacturing. The commitment has paved the way for India to achieve its dream of becoming the semiconductor capital of the world, and a possible way for us to catch up is by accelerating manufacturing. A potential solution could also be the implementation of digital twins. Why Digital Twin in Manufacturing? “Extended to vast scales, a digital twin is a virtual world that’s connected to the physical world,” said NVIDIA chief Jensen Huang. Digital twins act as virtual replicas of the physical systems, thereby providing a platform that helps train systems before the actual production. In the process, digital twins assist in reducing costs, time, and effort that would be needed to optimise production workflows. Last year, at a tech summit in Bengaluru, Microsoft showcased what an assembly line or a manufacturing plant would look like with HoloLens. In the process, it painted a picture of how the digital world will help scale manufacturing. India currently has one existing semiconductor facility (Micron in Sanand) and three newly approved facilities at Dholera in Gujarat, Morigaon in Assam, and another one in Sanand, Gujarat. While plans are in place, full-fledged manufacturing will take time to come to fruition. This is something, which can probably be expedited with digital twins. While India looks to find its way, companies such as Intel and NVIDIA, are going big on creating digital twin setups for semiconductor and other industries. Big Tech and Digital Twin Samsung is set to launch an NVIDIA Omniverse-based Fab Digital Twin for simulating fab architecture and semiconductor manufacturing, potentially the first to achieve smart factory Level 5. The platform will pilot next year, showcasing various use cases in planning and simulation. Furthermore, NVIDIA’s Aerial Omniverse Digital Twin for 6G will be able to create accurate simulations for 6G systems, proving to be a crucial factor for AI integration testing. Interestingly, Intel is not far behind, as the company signed an MoU with Siemens last year to collaborate on the digitalisation of its wafer fabs using digital twin technology. Recently, the US government invested $285 million to enhance digital twin technology in semiconductor manufacturing. This funding aims to advance the development and application of these virtual models in the industry. India, The Favoured Capital While the logistics and workaround for setting up fab units are ongoing, the country is already a favoured destination for the chip industry. US-headquartered chip company SiMa AI has chosen India as its strategic market despite China being the largest semiconductor market in the world. In December 2021, the Indian government launched the India Semiconductor Mission that announced an INR 76,000-crore chip incentive scheme that offers a 50% subsidy on expenditure for plant development. Meanwhile, Tesla is said to have partnered with Tata Electronics to buy semiconductor chips for its global operations. With the burgeoning pace at which the semiconductor industry is developing in India, it won’t be long before India could possibly emerge as the semiconductor capital of the world. Digital Twins for the World While semiconductor companies have found ways to expedite manufacturing processes with digital twins, the concept is being heavily implemented across other sectors too. For instance, space tech, which involves exorbitant cost, resources and most importantly, time, is already solving these with digital twins. Recently, Declan Ganley, the founder and CEO of Rivada Space, told AIM how tasks that took 50 days five years ago can now be accomplished in a day with digital twins in satellite technology. Fujitsu, which is already in the semiconductor market, is working on creating digital twins for other verticals. They are in the process of developing ‘Ocean Digital Twin’ using AI and underwater drone data. The company is looking to promote marine conservation, carbon neutrality and biodiversity through this digital twin. It is also working with Carnegie Mellon University to develop AI-powered digital twin technology with traffic data from Pittsburgh. It is clear that robotics, especially autonomous vehicles, has one of the largest use cases for digital twins and simulated reality in their training, and in bring AI to the physical world.","excerpt":"“Extended to vast scales, a digital twin is a virtual world that’s connected to the physical world,” said NVIDIA chief Jensen Huang.","categories":["AI Features"],"tags":["Semiconductor India"],"author_name":"Vandana Nair","publish_date":"2024-08-08T11:07:42","publication_year":"2024","word_count":714,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","Git","Aim","Semiconductor India","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","API","GAN","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-digital-twin-can-accelerate-indias-semiconductor-ambitions\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119784,"title":"OpenAI Joins Adobe and Others as C2PA Committee Member","content":"The Coalition for Content Provenance and Authenticity (C2PA) announced that OpenAI has joined the C2PA as a steering committee member. This marks a significant milestone for the C2PA and will help advance the coalition’s mission to increase transparency around digital media as AI-generated content becomes more prevalent. Joining other steering committee members, include Adobe, BBC, Intel, Microsoft, Google, Publicis Groupe, Sony, and Truepic. OpenAI will collaborate to develop further and promote the adoption of Content Credentials, an implementation of the C2PA’s open technical standard for tamper-resident metadata that can be attached to digital content, showing how and when the content was created or modified. Today’s announcement builds on OpenAI’s previously shared initiatives to improve transparency around digital provenance. Earlier this year, OpenAI began attaching Content Credentials to images created and edited by DALL•E 3, the company’s latest image model, in ChatGPT and the OpenAI API. In addition, the company also announced its plans to attach Content Credentials to video generations from Sora, the company’s text-to-video model, when the model is ready to be deployed to the public. OpenAI’s membership and implementation of Content Credentials serve as a strong endorsement for the C2PA technical specification and advance the collective mission to help restore trust in the digital ecosystem. “C2PA is playing an essential role in bringing the industry together to advance shared standards around digital provenance,” said Anna Makanju, OpenAI’s VP of Global Affairs. “We look forward to contributing to this effort and see it as an important part of building trust in the authenticity of what people see and hear online.” The Coalition for Content Provenance and Authenticity (C2PA) is an open, technical standards body addressing the prevalence of misleading information online through the development of technical standards for certifying the source and history (or provenance) of digital content. C2PA is a Joint Development Foundation project.","excerpt":"OpenAI will collaborate to further develop and promote the adoption of Content Credentials","categories":["AI News"],"tags":["OpenAI"],"author_name":"Pritam Bordoloi","publish_date":"2024-05-07T16:56:05","publication_year":"2024","word_count":306,"keywords":["Go","ChatGPT","API","OpenAI","AI","Git","GPT","Rust","GAN","R"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","Go","Rust","Git","API","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-joins-adobe-and-others-as-c2pa-committee-member\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10007088,"title":"What Does India Need In Place To Implement Nationwide AI Systems Across Sectors?","content":"The Government of India has recognised that an AI-driven economy can transform the lives of millions. Leveraging AI for inclusive growth is one of the core principles identified in NITI Aayog’s National Strategy paper. It is the path for much-needed job creation in various sectors, apart from creating new business opportunities and helping increase household incomes. But nationwide AI can only be done by creating datasets that combine these information systems that power e-schemes already established in India. India’s AI revolution will require new architecture designs and upgraded technologies to make real-time decisions in an efficient manner. To integrate AI with Indian sectors, it will need a nationwide strategy that is centred on uniform AI standards and practices. Apart from that, an AI-centred smart economy will need extensive investment at technical and skills level. Hence, the government, along with private sector players, including manufacturers, service integrators, cloud service providers, etc., need to come together and coordinate in the development of an AI framework. What are the various steps that are being taken when it comes to establishing a technical framework for the adoption of AI in India? There are technical challenges in the form of  scalable and robust platforms that can ingest zettabytes of large data sets. In a recent discussion paper, India’s AI Standardisation Committee has outlined the issues related to developing a framework of an Indian AI stack. This paper proposes a stack that seeks to remove the impediments to AI deployment by putting in place a comprehensive framework. A framework that will create an enabling environment to exploit AI productively in various walks of life. This will enable the development of a suitable AI stack with a different mix of layers and interfaces that complements each other and achieves integration. One of the major advantages of this proposed Indian AI stack is that it will facilitate open API integration and build the AI architecture from the ground up. It also ensures the creation of a common Data controller, including multi-cloud scenarios-private and public, as part of the infrastructure layer. AI Standardisation To address some of these bottlenecks from a standardisation point of view, the Department of Telecommunications (DoT) had formed an AI standardisation committee to develop various interface standards and develop India’s AI stack. It will facilitate the implementation of standards for the AI developers and coders with compliance verified through proper algorithmic auditing. This necessitates openness in AI algorithms and enables clearly defined data structures. Infrastructure Infrastructure is an important part of the Indian AI stack. Infrastructure will harness hardware and software innovation to deliver unprecedented products and services in the economy. This can also involve multi-cloud scenarios- both private and public clouds to define the infrastructure. The broad specifications for this include ML \/ DL software stack, training and inferencing development kit, frameworks, libraries, cloud management software and more. The layer will ensure the creation of a common data controller. The data controller will determine the purposes for which and the means by which personal data is processed for use by various layers. Storage The foundational component of any type of AI\/ML approach is big data as ML algorithms work best when you feed them large data sets. Thus, there is a necessity to assure proper storage frameworks for AI, including multi-layer storage systems to ingest and analyse multi-petabytes of big data. Data storage is the most important layer, regardless of size and type of data. To derive value from data, it needs to be processed, and to process it efficiently, it needs to be stored in an effective manner. Even if the best data engineering is implemented, it is practically not feasible to augment and utilise data and gain repeated value out of it without having the right data storage. Having a very clearly defined data structure is the key to making it accessible seamlessly across domains and for various use cases. AI Integrity & Security While AI offers huge potential to transform and realign the economy and society, there is an increasing realisation that AI could also exacerbate problems for people, without proper safeguards. For AI to ensure a sustainable revolution, there is a need to provide an open environment with safeguards and oversight. Through defined data structures, this layer will ensure the process of security and governance. Such a broad plan would solve various issues across the industry as a whole and allow AI software to make fair decisions using unbiased data and transparent practices. Regulatory standards for data collection, interfaces, storage, analysis, application and customer use are also required. It can control existing risks and can preempt future risks by suitable monitoring and auditing of the AI’s design and analytics as part of the stack design. Due to the overwhelming flow of information, there is thus a need to ensure encryption at different levels. Data\/Information Exchange With the help of defined data structures and proper interfaces and protocol, the end customer interface should be defined. The layer of AI stack will have to support an appropriate consent framework for access to data by\/for the customer. Typically there could be different Tiers of consent available to accommodate different tiers of permissions. The information exchange also needs to ensure that proper ethical standards are followed while ensuring the requisite digital rights. The data\/information exchange also defines APIs access for interfaces to different types of applications. There will also be a web-based user interface designing tools to create, modify, test and deploy different UI scenarios. Data Literacy Embedding AI or ML in national systems is a piece that has to come from the government, not merely private tech companies to make it successful. Government employees therefore need to have skill sets to make AI successful. And hence, data literacy is important. Senior government leaders are starting to be acutely aware of the value of data and are realising value outside of traditional IT and cyber use cases. State governments are taking an interest in how they can leverage data so they can have mashups of geospatial data maps, overlay road accident data, crime data and others to start taking a predictive look at the scenarios. If government officials, bureaucrats and policymakers are data literate, it will help them understand what it means to execute good cyber hygiene and access the various tools\/interfaces for rapid data-driven decision making.","excerpt":"The Government of India has recognised that an AI-driven economy can transform the lives of millions. Leveraging AI for inclusive growth is one of the core principles identified in NITI Aayog’s National Strategy paper. It is the path for much-needed job creation in various sectors, apart from creating new business opportunities and helping increase household […]","categories":["AI Features"],"tags":["AI What it Does","how does artificial intelligence work","India AI","india ai strategy"],"author_name":"Vishal Chawla","publish_date":"2020-09-10T13:00:00","publication_year":"2020","word_count":1053,"keywords":["India AI","Go","API","big data","AI","india ai strategy","ML","Scala","Git","RAG","how does artificial intelligence work","AI What it Does","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","Scala","Git","API","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-does-the-government-of-india-need-in-place-to-implement-nationwide-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10071233,"title":"Tensor2Tensor to accelerate training of complex machine learning models","content":"Tensor2Tensor (T2T) is one of the libraries of Tensorflow consisting of various deep learning models and datasets. It aims to accelerate deep learning research and make deep learning models and data more accessible. Tensor2Tensor aims to train deep learning models to be trained and executed on various platforms with minimal hardware specifications and configuration. In this article, let us focus on the Tensor2Tensor library and understand the major benefits of using this framework in various use cases and applications. Table of Contents Introduction to Tensor2TensorNecessity of Tensor2Tensor Features of Tensor2TensorUse cases of Tensor2TensorSummary Introduction to Tensor2Tensor Tensor2Tensor, in short, is known as T2T, and this library is mainly used to increase the usage and availability of deep learning models across various platforms irrespective of device constraints and specifications. The Tensor2Tensor library has various inbuilt datasets and deep learning models that can be used for various tasks like image classification, image generation, sentiment analysis, speech recognition, and also for complex tasks like language translation. Are you looking for a complete repository of Python libraries used in data science, check out here. So, in a nutshell, the Tensor2Tensor is a single shot library with various inbuilt datasets and models that can be used for various tasks. The Tensor2Tensor library provides us with the flexibility to add required data and models into their library as they encourage more addition and also bug fixes if any are recognized. So now, let us look into some of the standard functionalities provided by the Tensor2Tensor library. There are four functionalities that are mainly supported by the Tensor2Tensor library. Let us have an overview of each of the functionalities. Problems The problem functionality of the Tensor2Tensor library basically consists of various features, inputs, and targets to be obtained from the models. The data features are stored in a standard directory named TFRecord, and within the library, it is made available in a standard python(py) file named “all_problems.py”. Models The models functionality is one of the vital functionalities of the Tensor2Tensor library as it is used for computations. Some default transformations are applied to the input and the output features so that the models and the data do not face platform-based dependency issues, and the users can use the pretrained models and data flawlessly. Hyperparameter Set The hyperparameter sets functionality is responsible for storing some of the hyperparameters of various models and data for each problem readily available in the library. So this set of hyperparameters is made available in the library within a python (py) file named “common_hparams.py”. Trainer The trainer is one of the functionalities of the Tensor2Tensor library that is mainly used for utilizing the models and evaluating the models present within the library. With this functionality, the users are provided with the flexibility to switch between the models, data, and hyperparameters available in the respective python files. Adding custom components As mentioned earlier, the library facilitates adding required data and models as per requirement. So this functionality serves as the mechanism which facilitates the data and model addition as per the requirements of the Tensor2Tensor library. The Tensor2Tensor library also provides 5 key components that specify the training process in the library. So let us now look into those 5 key components. i) Dataset is the component of the Tensor2Tensor library which is encapsulated by the Problems functionality. This component encapsulated by the problem will be responsible for pipelines for training and evaluation and also responsible for downloading data suitable for the library. ii) Device Configuration is the component responsible for ensuring that the library is supported on various device configurations and specifications like CPU, GPU, TPU, and the devices with support for parallel training. iii) Hyperparameters is the component that is responsible for initiating the pretrained models available in the library and used to train the model with the required set of parameters. iv) Model is the component that gets activated according to the hyperparameters, and this component is responsible for transforming the data, computation, and evaluation of various metrics from the models loaded. v) Estimator and Experiment is the component responsible for monitoring the logging parameters, running the training process across various platforms, and also carrying out various experiments on the metrics produced. Necessity of Tensor2Tensor The main necessity and use case of the Tensor2Tensor library is to make deep learning and various complex models easily accessible and producible irrespective of device specifications and limitations. The Tensor2Tensor facilitates the storage of various types of data like images, audio, text, and many more in a single library and trains various models with different levels of complexity and architecture in a single framework. The models and the data are made available in the form of pretrained data and models, and the parameters of the models can be made available by the researchers to implement and use for complex tasks. Language translation, speech recognition, and image generation are some of the data and models that are made available in the library and maintained as open source so that the researchers and users can use them for their purpose. The main aim of the Tensor2Tensor library was to make deep learning models accessible and to accelerate the training of the models irrespective of the hardware specifications, which led to the development of this library. Now let us try to understand some of the features of the Tensor2Tensor library. Features of Tensor2Tensor The dynamic ability of the Tensor2Tensor library has facilitated the library to provide certain standard features of operation which accounts for its usage. Let us look into some of the features that the Tensor2Tensor library has to offer. Many complex models are made available in a simple, easy-to-use format, and if required, additional models can be added to the library that can be used in the future.Various forms of datasets like text, image, and audio are available that can be used either to generate data or to use for various tasks.Models and the datasets can be made available, and the model’s hyperparameters can be transformed according to requirements and trained suitably irrespective of the platform constraints and hardware specifications.Special support for accelerator support devices such as GPU and devices with parallel processing capabilities where complex models tend to converge faster.The pretrained models and data can be pushed into cloud-based platforms like Google Cloud ML and platforms with the support of TPUs, and the models can be trained and evaluated completely on the cloud platform itself. Use cases of Tensor2Tensor The Tensor2Tensor library has various data types and uses cases that can be used easily for complex tasks and modelling. So let us look into some of the standard functionalities and use cases of the Tensor2Tensor library. Mathematical Language understanding For mathematical language, understanding respect to mathematical attributes to perform various mathematical operations is made easy by the Tensor2Tensor library. So for mathematical model understanding, the library provides us with a readily available dataset known as the MLU dataset under the problems functionality. For this problem statement, there are 3 types of transformers that are pretrained for mathematical language understanding, which uses different sets of transformers and hyperparameters for the respective models. Question Answering The Tensor2Tensor library consists of a pretrained dataset known as the “BABI” dataset, where the data characteristics are similar to question answering from a story. There are various sets of question answering sets and subsets in the data, and this can be used accordingly for developing and evaluating the Question Answering models. Image Classification The Tensor2Tensor library consists of various datasets suitable for image classification, such as ImageNet, CIFAR, and MNIST. So the datasets can be made available using appropriate problem constraints and used accordingly to accelerate and increase the availability of image classification models and tasks. For ImageNet data, some of the transfer learning models like ResNet and Xception are trained and made available in the form of a model, and the parameters of the model can be used accordingly with the appropriate set of parameters to instantiate the model training on the platform. For CIFAR and MNIST, a pre-trained regularisation technique named shake-shake regularization is being used to improve image classification. So the data can be made available, and the suitable parameters have to be declared accordingly to extract the data and the model trained accordingly for image classification. Image Generation The Tensor2Tensor library has various standard datasets for image generation, such as CeleBA, CIFAR10, MS-COCO, and many more, which can be used extensively for image generation with the required set of parameters and constraints. So the deep learning model, which is made available in the library, can be pulled into the working environment and used for image generation tasks accordingly. Language Modeling The Tensor2Tensor library can be used for easy language modelling and translation. Various language data and language models are made available in the form of problems (data) and models, and this can be suitably pulled into the working environment and used accordingly for language modelling and language translation tasks. Sentiment Analysis For sentiment analysis, the Tensor2Tensor library consists of the IMDB data for recognizing the sentiment of a sentence, and the library provides a trained model to perform sentiment analysis on a text sentence. So the model and the parameters for sentiment analysis from the library have to be pulled into the working environment, and the readily available trained model can be used accordingly to perform sentiment analysis. Speech Recognition The Tensor2Tensor library can be used for speech recognition as the library has two inbuilt datasets for speech recognition. They are basically Speech to text data where the speech is generally in the English language. The datasets available in the library are Librispeech, and Mozilla Common Voice, where the data has to be pulled into the working environment according to the standard problem constraint, and the models trained respectively on each of the data have to be pulled in a similar manner into the working environment with an appropriate model trained with respect to the data. Summary The Tensor2Tensor library aims to provide a single shot framework to facilitate ease of use of complex data and models across various platforms and hardware specifications. The library is well built with various data types and models to simplify complex tasks. So complex deep learning models can be made available irrespective of hardware specifications, and the models can be trained accordingly on any platform by using the Tensor2Tensor library without any issues of dependencies and flaws. The library basically aims to speed up the deep learning training process and make complex deep learning models easily available and accessible. References Tensor2Tensor official documentation","excerpt":"This article mainly focuses on the Tensor2Tensor library and to understand the dynamic abilities to handle and process complex models.","categories":["AI Trends"],"tags":["language translation","Tensorflow"],"author_name":"Darshan M","publish_date":"2022-07-20T13:00:00","publication_year":"2022","word_count":1762,"keywords":["data science","AWS","AI","sentiment analysis","ML","Transformers","RAG","Aim","deep learning","language translation","TensorFlow","Tensorflow"],"extracted_tech_keywords":["AI","ML","deep learning","data science","Aim","TensorFlow","Transformers","RAG","sentiment analysis","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/tensor2tensor-library-of-tensorflow\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":64548,"title":"Top AI Tools To Help Workplaces Restart Operations Post Covid-19 Lockdown","content":"Weeks after keeping their shutters down, companies are emerging out of an extended nationwide lockdown to reboot operations in offices in a phased manner. As restrictions begin to ease and pressure to resume work ramps up, some IT companies have already opened workplaces for critical functions. Given that the risk of contamination still remains high in several locations, a practical — and necessary — approach would be to prepare a detailed standard operating procedure (SOP) to mitigate this risk and protect employees. Can a calibrated technological intervention help ease some of the core modifications required to reset workplaces in a post-Covid-19 world? Be it enforcing mandated social distancing, running at limited capacity using roster-based systems, implementing screening and sanitising procedures, as well as establishing a remote work option, companies can greatly benefit from technologies to meet some of these challenges. Tools To Help Enforce Social Distancing The Ministry of Home Affairs had recently issued a new set of rules for work from offices as businesses prepare to resume operations post lockdown. One of the directives was to ensure stringent social distancing measures within offices and its premises. While many of these measures might demand behavioural changes, some tools can be used to ensure that new habits are created and dutifully followed. Landing AI, a startup founded by Andrew Ng, has created an AI-enabled system that can monitor employees to ensure that they keep safe distances from each other. By undertaking video content analysis, it can indicate when employees are violating social distancing protocols. Another startup called Drishti has also been developing AI tools to surveil employees to check the distance maintained between staff in a work setting. Businesses are likely to introduce various steps to optimise the use of empty spaces in their offices once employees begin to populate workplaces again. While some may rearrange seating to ensure adequate space between cubicles, others may impose staggered meal timings in demarcated areas only to facilitate social distancing. If common sense prevails, technological interventions to enforce these measures may not be required. But on the off chance that a few stray examples may pose a risk to other employees, these tools can be life-saving. ALSO READ: How Computer Vision Came In Handy For Social Distancing Automating Sanitation Processes Disinfecting offices and its surroundings may become a regular fixture until the crisis passes, and automating these sanitation processes may not only be convenient but also safer. To minimise the chances of contamination while maintaining essential hygienic practices, autonomous cleaning robots can potentially replace staff. Based out of the Netherlands, Aziobot builds floor cleaning robots for warehouses and commercial facilities, and its fully autonomous solutions can find applications in scaled-down settings as well. Another company UVD Robots — short for UV Disinfection Robots — is extensively being used in hospitals as part of the regular cleaning cycle. Bots have played a key role in keeping frontline healthcare workers safe amid the Covid-19 pandemic — a case in point being Kerala’s Nightingale-19 Robot. With workplaces ramping up cleaning protocols, deploying robots to periodically sterilise objects used regularly can help. Thermal Screening Of Employees Incorporating temperature screening as part of the SOP in workplaces can help flag potential cases before it escalates. Contactless body temperature scanning has reportedly been implemented across warehouses of Amazon and Walmart. These measures have also shored up to certain offices in India, including Maruti Suzuki’s. US-based Flir’s thermal camera can be used to detect elevated body temperature and is currently being leveraged by Indian Railways. Gurugram’s Staqu has also launched a similar technology to supplement screening efforts by examining heat signatures and alerting authorities when the temperature of a person within a 100-metre radius crosses 37-degree celsius. However, it is important to note that although thermal cameras cannot directly detect Covid-19 and will be ineffective against asymptomatic employees, it can be used as part of the initial filtering process for additional screening. ALSO READ: Agrex.AI Develops AI-Enabled Thermal Cameras To Combat The Spread Of Covid-19 AI-Enabled Tools For Roster Management HR departments across industries are buckling under pressure while managing employees in these uncertain times. While working remotely comes with its share of problems, the most significant challenges will emerge once work resumes in offices. Scheduling staff to meet mandated workplace capacities will require deft planning and efficient administration. With this, many companies are launching or updating tools to add more features to roster management software that can help manage complex scheduling. Globus.ai, a company based out of Norway, has built an AI-enabled system that automatically fills shifts in a work setting. Using NLP and ML, it can match the competencies of employees to align them with appropriate shifts based on available slots. What is more, given that this crisis is an evolving one with new updates coming in almost every day, Globus.ai’s system can also factor in policy requirements when making its suggestions. Better Tools For Online Collaboration Even as companies faced disruptions in work amid Covid-19 lockdowns, it made employees adept at adjusting to working remotely. Although many companies will open up their workplaces in the weeks to come, given that this is an ongoing crisis, all organisations must be prepared to work from home at short notice. This means turning to applications like Zoom for meetings, or Slack for internal communication. What is more, with inbound as well as outbound travel severely restricted even now, going virtual with meetings instead of travelling may become the norm, at least for the foreseeable future. This would also mean preparing for virtual events and conferences. Outlook The challenges posed by Covid-19 in ensuring business continuity is unprecedented. Associations like CII and Nasscom have released guidelines and advisories as companies get ready to function in a radically different setting. As they prepare to restart operations, their Covid-19 strategies need to be flexible and nimble as they adapt to this uncertain and evolving situation.","excerpt":"Weeks after keeping their shutters down, companies are emerging out of an extended nationwide lockdown to reboot operations in offices in a phased manner. As restrictions begin to ease and pressure to resume work ramps up, some IT companies have already opened workplaces for critical functions. Given that the risk of contamination still remains high […]","categories":["AI Trends"],"tags":["AI Workplace","covid-19","Work from Home"],"author_name":"Anu Thomas","publish_date":"2020-05-06T11:00:00","publication_year":"2020","word_count":981,"keywords":["Go","covid-19","AI","ML","Work from Home","computer vision","RAG","NLP","Ray","Scala","GAN","R","AI Workplace"],"extracted_tech_keywords":["AI","ML","NLP","computer vision","Ray","RAG","R","Go","Scala","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ai-tools-to-help-workplaces-reopen-post-covid-19-lockdown\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10070523,"title":"Genpact is the training ground for future analytics leaders","content":"What is common between Tesla, LinkedIn, and YouTube? Apart from being wildly successful companies in their own right, all three companies were created by former PayPal employees and founders. PayPal was originally a money-transfer service offered by Confinity, a company founded by Peter Thiel, Max Levchin, and Luke Nosek. The company was acquired by X.com in 1999, and after a few years, the latter was renamed PayPal and sold to eBay in 2002. Within four years of this transition, 12 of the first 50 employees left, however, they remained connected socially and businesswise. Over time, a number of them partnered together to form new companies and ventures. Some of these ventures became very successful and prolific. In 2007, Fortune Magazine carried an article and coined the phrase – PayPal Mafia – it was accompanied by a photo of these former PayPal employees in gangster attire. These founders are now the most influential and wealthiest people in Silicon Valley. Cut to India, Genpact has produced some of the most influential analytics leaders that we have today. Genpact Mafia, if you will. Yesterday’s Genpact employees, today’s analytics leaders Bermuda-headquartered Genpact is a global professional services firm which employs over 100,000 people. It was established as a part of General Electric in 1997 before being spun off as a separate entity called GE Capital International Services (GECIS) in New Delhi. Headed by Pramod Bhasin, the company offers business process outsourcing solutions to GE’s businesses. Genpact is widely considered a pioneer of business process outsourcing. By 2005, GECIS became an independent company and changed its name to Genpact. Within two years, it became a publicly-traded company and since then, the company has diversified to offer AI-based services, building proprietary APIs and developing an open architecture to include all of its intellectual property and that of other providers. NV “Tiger” Tyagarajan is the current CEO and president of the company. Genpact housed and nurtured some of the best talents in analytics who have now gone on to hold some key positions in leading organisations. RB Rajendar, the current head of analytics – India, Citi Solutions, was one of the initial employees of GECIS. He joined the company in May 1998 as the assistant manager and was instrumental in setting up the analytics department of the company in India. He was associated with the company for over three years and rose to the position of Assistant Vice President before joining HSBC. Rohit Tandon, the Global AI & Analytics Services Leader, Deloitte, joined GECIS in 1999 as the VP – Business Development, before diversifying into other roles within GE. He joined Genpact back again in 2006 as SVP and Business Leader- Analytics for two years. He had his third stint at the organisation in 2016 as the SVP and Analytics Business Leader, where again he spent over two years. Between 2002 and 2010, Namit Sharma worked as the Vice President at Genpact, leading a 550-member analytics and research team working across retail banking, credit cards, auto finance, brokerage, and e-commerce. He is now the Vice President & Head of Global Delivery Center for Analytics, India at FICO. Shuvajit Basu joined as the Delinquency & Risk Strategy Lead at GECIS in 2002. He is currently the VP and Head of Data Science at Tata Group. Satyamoy Chatterjee, executive vice president at Analytiica Datalab, is a data science professional with an experience of 20 years. His association with Genpact dates back to 2003 when he joined the company as a Consultant after completing an MS course from Wichita State University. He worked there for three years before joining Citibank and subsequently assuming his current role at Analytiica Datalab. As an Analytics Manager at GECIS in 2003, Arnab Chakraborty was instrumental in building the Marketing Analytics Center of Excellence (CoE) for GE Commercial Finance. He is currently the Senior Managing Director – Applied Intelligence North America Lead for Accenture AI. Paritosh Anand is the Senior Vice President – Digital Platforms at Reliance Industries Limited, but he started his career with Genpact as a consultant in 2003. Like Paritosh, Amit Kalra too began his professional career in 2003 with GECIS in Market Research. He is currently working with Swiss Re as Managing Director, Head Global Business Solution Centres India. Beginning her professional career as an assistant system analyst at TCS, Mathangi Sri joined Genpact in 2005 as a consultant and worked there for over two years. Since then, Mathangi has worked in organisations like HSBC, Citibank, PhonePe, and Gojek, before joining her current company Yubi (earlier called CredAvenue) as the chief data officer. Amit Khanna is currently a Partner at PricewaterhouseCoopers. He worked with Genpact in 2005 as the Service delivery lead ( Customer & Marketing Analytics) for over two years. Vice President – Global Analytics at Concentrix, Hari Saravanabhavan, is an analytics leader with an experience of 25 years, working with companies like Cognizant, IBM, and Airtel. Saravanabhavan worked at Genpact between 2006 and 2008 as the assistant vice president – Retail, CPG, and Telecom analytics. At Genpact, he oversaw the deployment and delivery of analytics models and developed plans for innovative analytics interventions. Thomas Jacob Kollenkeril joined as a consultant in 2006, rising and was the Global Operating Lead at the time of leaving the organisation in 2020. He is currently the Senior Director (Analytics) at Biocon Group. Ram Kumar joined Genpact in 2006 as a Consultant and worked there for over two years. He rejoined the organisation in 2009 as a Manager and worked for almost seven years this time. He is now the Senior Director & Head, Data Sciences at Snapdeal. Ishu Jain is the Head of Central Analytics at Swiggy, where she heads a team of 100 plus data scientists\/analysts across central functions like delivery network optimization, customer segmentation, loyalty, traffic and acquisition. Ishu joined  Genpact as a consultant in 2006. Training ground for future entrepreneurs Pankaj Kulshreshtha joined GECIS as the Vice President – Analytics, in 1998. He has worked with other GE subsidiaries before coming back to Genpact as SVP & Business Leader – Analytics & Research. After a six-year stint, he went on to establish Scienaptic Systems, an AI-based Credit platform, in 2014. Founded in 2011, BRIDGEi2i enables data-driven digital transformation for its clients, using a combination of data engineering, analytics, proprietary AI and other consulting services. In 2021, the company was acquired by Accenture. The acquisition added over 800 professionals to Accenture’s Applied Intelligence practice. Interestingly, one of the co-founders of BRIDGEi2i is Prithvijit Roy. Roy was one of the earliest employees of GECIS in 1999, where he worked as the assistant vice president, before leaving in 2005 to join HP. Like Roy, Sayandeb Banerjee joined GECIS in 1999 as a manager. He worked at GECIS before joining GE Money in 2003. Currently, he is the co-founder and CEO of TheMathCompany, a company that helps enterprises build and maintain in-house analytics centres. He is also the co-founder\/CEO of co.dx. Sanjeev Mishra is the CEO and co-founder of Convergytics Solutions Pvt. Ltd that offers analytics solutions to Retail, CPG, Telecom, Technology & Media companies. He was associated with GECIS between 2001 and 2003. Sarita Digumarti is the co-founder of Jigsaw Academy, an edtech firm that offers courses in AI and analytics. GECIS was one of the first organisations she worked with at the beginning of her career. In 2004, started as an assistant manager and rose through the ranks to become Assistant VP. Apart from these, Genpact alumni have gone on to establish companies like Transitials, eWandzDigital Services, Stratbeans Consulting Private Ltd, and Stylofie, among others.","excerpt":"Genpact has produced some of the most influential analytics leaders we have today.","categories":["IT Services"],"tags":["analytics companies","analytics leaders","Genpact"],"author_name":"Shraddha Goled","publish_date":"2022-07-06T17:00:00","publication_year":"2022","word_count":1263,"keywords":["data science","Go","API","Genpact","AI","analytics leaders","Git","RAG","analytics companies","data engineering","analytics","GAN","R"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","Go","Git","API","data engineering","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/genpact-is-the-training-ground-for-future-analytics-leaders\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":53721,"title":"Why Does Amazon Want You To Remove This $4 Billion Browser Extension?","content":"Amazon is warning users to remove Honey browser extension, which was bought by PayPal for $4 billion. The American multinational e-commerce giant sees the add-on as a treat to users privacy as it collects users data for tracking the shopping behaviour of the purchasers. However, the timing and selective outrage has confused experts and buyers around the world. Initially, the message popped up for a few users during Christmas, and later it was notified to almost all the users. However, it came as a surprise as Honey is providing its services on Amazon for years. And on top of that, Amazon didn’t target other extensions like Camel Camel Camel.com that more or less delivers the same functions as Honey does. Amazon’s Security Risk Warning “Honey’s browser extension is a security risk. It tracks your individual shopping behaviour, collects data such as the order history, items saved, and can even read or change the data on any website browsed. To keep your data private and secure, uninstall the Honey extension immediately. Please note that all Amazon offers, deals, and coupons, are available without this extension installed,” the warning read. However, Honey’s spokesperson has disagreed with the notification and said, “Our extension has never been a security risk and is safe to use. We have dedicated team for ensuring the security of our purchasers’ information, and we continuously engage with third-party security firms to assess our security protections.” Why Only Honey PayPal Holdings announced the acquisition of Honey in November 2019 and completed the deal in January 2020. Honey allows users to apply best coupons and helps users to save money by finding the best deal across various retailers. Started in 2012, Honey has over 17 million users leveraging the plug-in to secure cheaper deals. Amazon does not go along with PayPal as it perceives them as a direct competitor. The e-commerce company doesn’t support PayPal on its platform during check-out as it was associated in its early years with eBay — another Amazon’s competitor. Amazon released the security risk warning about the plug-in only after PayPal acquired, which portrays the e-commerce giant’s unethical practice to downplay competitors. According to Juozas Kaziukenas, founder of New York e-commerce research firm Marketplace Pulse, “Amazon’s security warning against Honey, while letting a dozen other tools go on, is confusing. I don’t buy their security risk message. They want Honey and PayPal to be squashed.” Threat To Amazon Similar to Honey, Amazon has a recommended solution that suggests products, but its algorithm prioritises fast delivery instead of low price. Honey, in contrast, focuses on best price, which the customers embrace, resulting in reduced relevance of Amazon’s algorithm of recommendation. This has the potential to disrupt Amazon’s approach of favouring its products over competitors. However, Amazon’s, in the past, has squashed the narrative of encouraging purchasers to buy its products over others. According to reports, PayPal has raised a complaint about the unfair practices to regulators, for which they have been asked to provide anticompetitive practices. However, one of Amazon’s spokesperson said that they must warn users of any extension that collects shopping data such as name, billing address, payment methods, and others, without their consent. Outlook Amazon warned users that the plug-in collects purchaser’s data from any website, which is false as the policy of Honey states that it doesn’t track search engine history, emails or any website that is not a retail website. “As markets become more concentrated and organisations grow larger, we see attempts to preserve market positions and remove rivals through deceptive practices,” said Diana Moss, president of the American Antitrust Institute. Since Honey’s revenue is generated by charging retailers for the sales that go through its extension, collecting data doesn’t make them money. Instead, as per Honey, it doesn’t sell users data and utilises for improving its services. Consequently, the extension can be considered safe even after Amazon’s security risk warning.","excerpt":"Amazon is warning users to remove Honey browser extension, which was bought by PayPal for $4 billion. The American multinational e-commerce giant sees the add-on as a treat to users privacy as it collects users data for tracking the shopping behaviour of the purchasers. However, the timing and selective outrage has confused experts and buyers […]","categories":["Global Tech"],"tags":["Amazon"],"author_name":"Rohit Yadav","publish_date":"2020-01-13T16:46:26","publication_year":"2020","word_count":649,"keywords":["Go","programming_languages:R","AI","Amazon","programming_languages:Go","RAG","Ray","Rust","GAN","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","Ray","RAG","R","Go","Rust","GAN","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-does-amazon-want-you-to-remove-this-4-billion-browser-extension\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10069170,"title":"Join Intel’s webinar to learn how to achieve real-time AI inference on your CPU","content":"Intel, in association with Analytics India Magazine, is organising a webinar on “achieving real-time AI inference on your CPU” on 7th July, from 5:00 – 6:30 PM (IST). We all know that the amount of data generated in today’s world is exponential. AI Inference involves the process of using a trained neural network model to predict an outcome. For a typical AI workflow, the workloads associated with all the steps involved follow a diverse mechanism and a single GPU or CPU cannot work for the entire pipeline smoothly. To this end, Intel is organising this webinar for the attendees to understand how to optimise a deep learning neural network model and achieve fast AI inference with a CPU. The session will also introduce Intel’s OpenVINO™ toolkit, an open-source toolkit to enable neural network model optimisation and easy deployment across multiple hardware platforms. The webinar will also have a live demo on setting up and running OpenVINO to achieve real-time AI inference in a CPU. Click here to register for the webinar. What can you expect? The participants will get to learn three core things: How to run fast AI inference with your CPULearn about the set of tools OpenVINO provided in optimising and inferencing deep learning modelsHow to set up and run OpenVINO in just 5 minutes Session speaker Zhuo Wu, AI evangelist, Intel Zhuo works as an AI evangelist in PRC. She received her PhD from the University of York, the UK, in 2006. She has worked as an associate professor at Shanghai University from 2006-2014 and was responsible for research in next-generation wireless communications and supervising graduate students. After that, she worked as a research scientist in Bell Labs (China) from 2014 to 2018 and was responsible for 5G system standardisation and AI-related research for industrial applications. After that, she joined Accenture (China) as a data scientist and was responsible for AI-based solution design and delivery. Click here to register for the webinar. Session agenda 05:00 – 05:05 PMAIM introduction to the session05:05 – 05:45 PMPresentation to introduce OpenVINO05:45 – 06:25 PMHands-on course to run OpenVINO for object detection and OCR with webcam (pre-work needed)06:25 – 06:30 PMVote of thanks by AIM Who should attend? ML developersGPU & CPU programmersData scientistsAI & ML practitionersAI and ML enthusiasts Click here to register for the webinar.","excerpt":"The session will also introduce Intel’s OpenVINO™ toolkit, an open-source toolkit to enable neural network model optimisation and easy deployment across multiple hardware platforms","categories":["Deep Tech"],"tags":["Intel"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-06-16T12:00:00","publication_year":"2022","word_count":386,"keywords":["programming_languages:R","AI","neural network","ML","Aim","deep learning","object detection","analytics","GAN","R","Intel"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","analytics","Aim","object detection","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/join-intels-webinar-to-learn-how-to-achieve-real-time-ai-inference-on-your-cpu\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10170086,"title":"UAE, US Sign Agreement to Build Largest AI Campus in Abu Dhabi","content":"The United Arab Emirates (UAE) and the US have signed a deal for the former country to build one of the largest AI campuses outside the US. The US previously restricted this deal due to concerns that China could have access to the technology. The US Department of Commerce states that the UAE-US AI campus will have a capacity of 5GW for AI data centres located in Abu Dhabi, establishing a regional platform that enables US hyperscalers to offer low-latency services to nearly half of the world’s population residing within 3,200 km of the UAE. Upon completion, the facility will utilise nuclear, solar, and gas energy to reduce carbon emissions and include a science park to foster advancements in AI innovation. The UAE also seeks to protect its advanced AI technologies by enforcing strict measures to prevent unauthorised access and ensure regulated technology use. The two countries will collaborate to improve Know-Your-Customer procedures to manage access to computing resources designated for US hyperscalers and authorised cloud service providers. US Secretary of Commerce Howard W. Lutnick said, “[The agreement] promotes major investment in advanced semiconductors and data centres across the U.S. and the UAE.  American companies will operate the data centres in the UAE and offer American-managed cloud services throughout the region. The agreement also contains strong security guarantees to prevent diversion of US technology.” As stated by the White House, the US will promote enhanced technological collaboration with the UAE, which includes initiating a 1GW AI data centre as a part of a proposed AI technology hub. This initiative aims to address regional computational needs while adhering to strict US security protocols and implementing measures for the responsible deployment of AI infrastructure, both within the UAE and globally. Sources indicated that this development would enhance the Gulf nation’s access to sophisticated AI chips, as reported by Reuters.  The news agency noted that NVIDIA’s Chief Executive, Jensen Huang, was seen in televised coverage on Thursday discussing matters with US President Donald Trump and UAE President Sheikh Mohamed bin Zayed Al Nahyan. On Thursday, The Washington Post cited anonymous sources as saying that OpenAI engaged in detailed discussions with two leading companies in the UAE, G42 and MGX, regarding a significant deal that would allow the San Francisco-based firm to set up data centres and advance its technology within the nation. In exchange, OpenAI would receive fresh investments that would facilitate the establishment of US data centres in collaboration with a group of partners.","excerpt":"The AI campus will utilise nuclear, solar and gas energy to reduce carbon emissions.","categories":["AI News"],"tags":["AI campus US","AI in UAE"],"author_name":"Smruthi Nadig","publish_date":"2025-05-16T15:14:27","publication_year":"2025","word_count":412,"keywords":["AI in UAE","OpenAI","AI","programming_languages:R","AI campus US","innovation","RAG","Aim","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","RAG","R","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/uae-us-sign-agreement-to-build-largest-ai-campus-in-abu-dhabi\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26759,"title":"Adoption Of AI In India Has Been Limited To Chatbots, Says Sanjeev T Menon, Of Light Information Systems","content":"As part of this month’s theme “How is AI affecting human lives”, we talked to Sanjeev T Menon, CEO and co-founder of Light Information Systems. Menon helped build the company’s signature software such as NLPBots and LightApp which have revamped the natural language processing (NLP) space in India. Apart from NLP, he also has 20 years of R&D experience working in areas such as big data analytics, telecommunication, VoIP, augmented reality, e-commerce solutions and predictive algorithms. In this article, we discuss how Light Information Systems has made a positive impact in implementing AI-related solutions. Analytics India Magazine: What are some of the ways that your company is adopting and providing AI services? Please highlight some use cases. Sanjeev Menon: We provide AI NLP products and solutions in two different ways. ‘NLPBots’ AI NLP product platform, under a platform license for companies in the ITES industry for them to offer services and solutions for their enterprise customers. ITES companies utilise our platform and the underlying power of AI NLP algorithms to create use cases and solutions for their customers. These use cases span across internal stakeholders (HR assist, AI-powered hiring, Marketing\/Sales insights, Customer support etc.) i.e. any use case which involves the end user interacting with structured and unstructured data, irrespective of the domain. We also offer function-specific customised products on the platform i.e. HR Assist, Marketing Assist, Customer Support Assist etc. where the end to end process of managing the stakeholder interaction\/journey is automated with limited or no manual intervention. AIM:  What are the areas of life or employment where you would like AI to be more involved? SM: Within the enterprise, we envision AI NLP systems to act as personal assistants to all stakeholders. Currently, stakeholders within the enterprise ecosystem interact with a diverse set of structured and unstructured data through various channels and interfaces. Constant efforts are being made to simplify these interactions through chatbots. However, from a strategic point of view, having chatbot interfaces for each of these interactions do not add value since the enterprise will be left with a plethora of chatbots for each specific task. Instead, the AI NLP systems should evolve to be a one-stop solution for stakeholders where the assistant provides help irrespective of the channel and encompassing the complete data for the individual. There are a few areas where AI, in its current state, exceed the best of human abilities. We can see this within the field of medicine, in diagnostics, drug discovery, clinical trial patient selection etc. Personal life growth and management is also an area where AI can be more involved as there is an immense amount of digital data that we interact with on a regular basis, most of our activities and habits are tracked. We believe AI systems can combine all these data and give guidance for personal growth and achieving goals, as well as refining habits. AIM: Many experts have warned against AI taking over every aspect of our lives. How true is their fear? SM: The AI community is far from building generic intelligence where AI takes over human life. A lot of the fear arises out of the rapid development taking place in this space and through the spotlight it receives courtesy movie characters like Jarvis, Terminator, among others. In general, people have always been fascinated with the imminent threat of computers becoming smarter than the humans, for example, Deep Blue beating Kasparov at chess in 96, IBM Watson winning jeopardy in 2011, AlphaGo’s triumph over Lee Sedol in 2016. These instances serve as evidence that the community is making strides towards artificial general intelligence. But it’s one thing to win board games (which the AI has mastered), and quite another to have the reasoning capabilities of humans. Some of the other AI events that received significant interest like ‘Sophia’, the first robot citizen or other Human-like virtual assistants, are extremely advanced technology but can’t be compared to a human being. We are still far from when AI will be able to understand a joke and laugh, use common sense and develop generic reasoning skills that are at par with humans. Until very recently, research and development in the AI space was confined to very large corporations. With the democratisation of AI, the availability of compute power, open platforms from Google and Microsoft etc., are made accessible at a rapid pace rather than gradual familiarisation of the technology. Due to this democratisation, AI is becoming rapidly pervasive on tech platforms and in real-world applications. AIM: With voice-based assistants, facial recognition and other AI-based technologies being used widely, will the user privacy and data be jeopardised? SM: There are massive digital footprints of each of us covering all aspects of our lives. AI-based tech makes it easier to mine through this data and come up with action points based on what is needed; usually a commercial need of an enterprise. This leads to ‘Surveillance capitalism’ by some large organisations whose business models might be invasive. Similarly, national security projects on immigration and customs can altogether add a different spin to the consumption of private data. Humans will continue to push the technology envelope making it easier to do things, process data, draw insights etc., and this push will never stop. We should focus on creating more legal frameworks in this space on underlying business models and in order to evaluate the good, bad, and ugly. Initiatives are already underway to give individuals more control over their own data and it should continue to pick up the pace. AIM: How has been the adoption of AI in Indian scenario? SM: Chatbots – and the buck stops there with the current scenario in India. We have a fear of missing out and most of the companies have a chatbot strategy to follow suit. But very few actually embark upon an AI strategy. At Light Information Systems, a lot of companies reach out to us because they have a mandate to develop a chatbot. Chatbots are not the true power of AI, it is just one more interface. India is more of a followers market when it comes to new technology implementation within companies. AIM: What are the changes in policies and infrastructure that you would like to see for the better adoption of AI? SM: Changes in policies and infrastructure are already happening at a global level. Graphics Processing Units (GPUs) for AI training is expensive and inaccessible(even though the prices have come down recently). The price points for these need to be significantly lower for the Indian scenario. Data privacy and sharing of data has to be better defined for online interactions. In a corporate environment, policies and practices are tweaked and governed keeping in mind deterministic systems. This need to change. Every AI system will have a learning phase and may not be deterministic at all, depending on the application. In terms of security most of the corporates try to create isolated systems without any access to the outside world (rather than have adequate security measures, it is easier to ban access), this needs to change. AIM: What are some of the other challenges that can come in the way while adopting AI? SM: There are various challenges: Solving for training bias – we are already witnessed AI becoming sexist, with recognition failing for darker skin tones, etc. Identifying when the AI is going wrong – It is easy to say that the AI is highly accurate, but even if a given system achieves an accuracy that is upwards of 90%, we need to accept it and find a way to fix\/identify the inaccuracies of the system. This becomes even more critical when the reliance on system guidance becomes second nature to humans. AIM: Will AI be capable of exhibiting emotions and developing a conscience? Will that be useful or harmful for us? SM: Eventually, AI could be capable of exhibiting emotions and developing a conscience of its own but that is definitely not happening anytime soon. Significant breakthroughs in technology have to take place and a completely new way of teaching the machines has to evolve before this actually happens. Citing the AlphaGo example – it has been seen as a huge success for AI and deep learning, but it’s actually incremental improvements in a combination of areas of deep learning, self-play, reinforced learning, tree search algorithms etc. rather than a monumental breakthrough in AI itself. Humans have quite a lot of biases based on the background, society as well as some evolutionary biases. AI-based decision-making systems are ideally suited to eliminate these biases rather than exhibiting human-like emotions and conscience. AI might one day get to a point where they exhibit emotions and develop a conscience, but that in itself is of little value in the larger scheme of things.","excerpt":"As part of this month’s theme “How is AI affecting human lives”, we talked to Sanjeev T Menon, CEO and co-founder of Light Information Systems. Menon helped build the company’s signature software such as NLPBots and LightApp which have revamped the natural language processing (NLP) space in India. Apart from NLP, he also has 20 […]","categories":["AI Features"],"tags":["AI Chatbot","AI India","Analytics India Magazine","chatbot ai","cognitive computing human capital","Interviews and Discussions","Natural Language Processing"],"author_name":"Srishti Deoras","publish_date":"2018-07-30T12:30:32","publication_year":"2018","word_count":1467,"keywords":["Go","AI","AI Chatbot","Natural Language Processing","chatbots","virtual assistants","Git","Analytics India Magazine","NLP","Aim","deep learning","chatbot ai","analytics","AI India","cognitive computing human capital","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","deep learning","NLP","analytics","Aim","chatbots","virtual assistants","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/adoption-of-ai-in-india-has-been-limited-to-chatbots-says-sanjeev-t-menon-of-light-information-systems\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10122692,"title":"Juniper Networks Brings Industry&#8217;s First AIOps to WAN Routing","content":"Juniper Networks has announced a new Routing Assurance product, the first in the industry to bring AI-Native automation and insight to traditional edge routing topologies. The company also announced that its WAN Assurance, Premium Analytics, and Marvis® Virtual Network Assistant (VNA) products have been augmented with new and unique AI for Networking capabilities that deliver simple, seamless, and secure SD-WAN and SASE experiences. With these latest platform enhancements, Juniper is the only vendor with a single AI-Native Platform that reduces operational expenditures by up to 85 percent in some instances across the entire networking footprint. The augmented Juniper solution leverages AI for Networking to drive even more value to enterprise WAN environments: Marvis Minis, Juniper’s digital experience twin solution that improves network ops by diagnosing real authentication issues without requiring users\/devices, has been extended to SD-WAN. New WAN speed tests can be continuously run (without users having to be present) to verify link speeds and take proactive actions if problems are detected. With this latest Marvis Minis expansion, Juniper is the first vendor to span wired, wireless and WAN with a single AI-Native digital experience twin solution, enabling exceptional end-to-end user experiences. In addition, existing WAN service level expectations (SLEs) for WAN edge health, link health and application health have been augmented with a new SLE that tracks WAN congestion. The new WAN Congestion SLE alerts operators when their network interfaces are being over-utilised, which causes poor user experiences. Juniper has expanded its unique streaming dynamic packet capture (dPCAP) solution for wireless and wired to now include WAN. With WAN dPCAP, the Juniper WAN Assurance solution proactively captures packets during a bad incident to help identify and fix hard-to-find issues, avoiding expensive and time-consuming site visits. Finally, new application insights offer network operators a user-friendly visualization of the traffic traversing the SD-WAN, enabling them to see bandwidth-intensive applications and enable accurate planning and problem remediation. “Since the launch of our AI-Native Networking Platform in January, Juniper has delivered on the promise to build out our industry-leading AIOps across all enterprise network domains. Now embracing routing, these latest innovations enable simplified, fast assurance, monitoring, troubleshooting and issue resolution across multiple branch office, WAN Edge and peering locations. We are also uniquely combining the security and networking domains operationally, enabling insight-driven, holistic security management and audit within the broader networking context, replacing silos with collaboration. All these new innovations further enable exceptional, secure user experiences for the enterprise,” Sunalini Sankhavaram, vice president, product development, Juniper Networks said.","excerpt":"Juniper is the only vendor with a single AI-Native Platform that reduces operational expenditures by up to 85 percent in some instances across the entire networking footprint.","categories":["AI News"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2024-06-06T15:53:25","publication_year":"2024","word_count":415,"keywords":["programming_languages:R","AI","innovation","ML","Git","RAG","automation","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Git","automation","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/juniper-networks-brings-industrys-first-aiops-to-wan-routing\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10102316,"title":"Nolano.ai Launches Open-Source English-Hindi Bilingual Model, Hi-NOLIN","content":"Young Indian startup Nolano.ai has introduced Hi-NOLIN, an open-source English-Hindi bilingual model, making it the first Indian model that has beaten all benchmarks for models of its size. Hi-NOLIN’s journey began with the goal of creating the first open-source English-Hindi bilingual model. Researchers expanded the 7B Pythia architecture to a 9B model, enhancing efficiency on their hardware while training on the 300B token Pile text corpus, encompassing both English and code data. Hi-NOLIN stands out for its ability to transition seamlessly between languages, mastering both Hindi and English, while processing code. As researchers continue training Hi-NOLIN, leveraging the Summit supercomputer with its unique 6 GPUs per node configuration, preliminary results demonstrate remarkable potential. Despite being far from convergence, the 9B model shows a steady reduction in training loss and promises substantial improvements. Employing advanced techniques from GPT-NeoX, Megatron-LM, and DeepSpeed, Hi-NOLIN utilizes 3D parallelism and ZeRO redundancy optimizer, maximizing its training resources and computational prowess. Hi-NOLIN shines through in various standard LLM benchmarks, including HellaSwag, TruthfulQA, Arc, and Human Eval. Remarkably, even in its preliminary stage with 600B tokens, Hi-NOLIN outperforms Pythia 12B and multilingual Bloom models across most evaluation benchmarks, narrowing the gap with LLaMa 2 models. In a landscape dominated by English language models, Hi-NOLIN is a significant stride towards linguistic inclusivity, addressing the gap in state-of-the-art language models for non-English languages. The team behind Hi-NOLIN includes the founding trio of Ayush Kaushal, Tejas Vaidhya and Irina Rish. While Vaidhya is a graduate research student in computer science at MILA and the University of Montreal, focusing on NLP and Transformer-based Language models, Kaushal, has already finished his master’s at the University of Texas, specialising in ML and DL, with expertise in data modalities. Irina Rish, a full Professor at the Université de Montréal, is a core faculty member of MILA, holds research chairs, and leads the US Department of Energy’s INCITE project on Scalable Foundation Models while serving as the co-founder and chief scientific officer of Nolano.ai. Meanwhile, Indian IT firm Tech Mahindra intends to launch Project Indus, its LLM designed for Hindi and its 37 dialects, by the end of December or early January. [Update: 2nd November 2023 20:17 |This article previously mentioned EleutherAI as the creator of Hi-NOLIN. The headline and story have now been updated to reflect the changes.]","excerpt":"Hi-NOLIN outperforms Pythia 12B and multilingual Bloom models across most evaluation benchmarks","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-10-31T18:14:36","publication_year":"2023","word_count":384,"keywords":["Go","AI","ML","Scala","RAG","NLP","GPT","GAN","foundation models","R"],"extracted_tech_keywords":["AI","ML","NLP","foundation models","RAG","R","Go","Scala","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/nolano-ai-launches-open-source-english-hindi-bilingual-model-hi-nolin\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10073612,"title":"CoRover Is Certified The Best Firm For Data Scientists","content":"CoRover is certified as the Best Firm For Data Scientists by Analytics India Magazine (AIM) through its workplace recognition programme. The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company culture. AIM analyses the survey data to gauge the employees’ approval ratings and uncover actionable insights. “We are greatly delighted to be chosen as a Best Firm for Data Scientists by Analytics India Magazine. It is truly a huge admiration for us, and we are pleased to be recognised by AIM,” said Ankush Sabharwal, Founder & CEO at CoRover “I want to humbly convey my utmost thanks to my entire team for taking up challenges and delivering their best. We still endeavour to keep up and stay abreast with the latest market trends and provide the best experiences and learning for existing and budding data scientists.” he added. The certification programme by Analytics India Magazine is considered a ‘Gold Standard’ in identifying the best data science workplaces. Companies participate in order to increase brand awareness and attract talent. The industry faces a massive talent crunch, and attracting good analytics and data science employees is one of the most pressing challenges enterprises face today. This certification is an effort to alleviate this issue. Best Firms For Data Scientists is the biggest data science workplace recognition programme in India. To nominate your organisation for the certification, please fill out the form at this link.","excerpt":"The Best Firm For Data Scientists certification surveys a company’s data scientists and analytics employees to identify and recognise organisations with great company culture","categories":["AI Highlights"],"tags":["CoRover.ai"],"author_name":"AIM Media House","publish_date":"2022-08-25T17:00:00","publication_year":"2022","word_count":245,"keywords":["data science","Go","programming_languages:R","AI","programming_languages:Go","Aim","analytics","GAN","CoRover.ai","R"],"extracted_tech_keywords":["AI","data science","analytics","Aim","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/corover-is-certified-the-best-firm-for-data-scientists\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164458,"title":"Meta&#8217;s New Framework Could Help Make The Next AI Einstein","content":"The Meta research team has unveiled an experimental framework called MLGym and MLGym-Bench to train and evaluate AI research agents on various AI research tasks. This comes days after Google unveiled its co-scientist system. The researchers mention that it is the first machine-learning Gym environment, enabling research on reinforcement learning algorithms for training AI research agents. As per the research paper, MLGym-Bench consists of 13 diverse and open-ended AI research tasks from various domains, including computer vision, natural language processing, reinforcement learning, and game theory. Solving these tasks requires real-world AI research skills such as generating new ideas and hypotheses, creating and processing data, implementing ML methods, training models, running experiments, analysing the results, and iterating through this process to improve on a given task. “Our MLGym framework makes it easy to add new tasks, integrate and evaluate models or agents, generate synthetic data at scale, as well as develop new learning algorithms for training agents on AI research tasks,” the research team mentioned. “We find that current frontier models can improve on the given baselines, usually by finding better hyperparameters, but do not generate novel hypotheses, algorithms, architectures, or substantial improvements,” the researchers added. With this framework, they aim to empower LLM agents capable of independently generating scientific hypotheses, writing scientific papers, analysing results, and more. The MLGym framework is designed to be modular and extensible, allowing researchers to easily add new tasks, datasets, and tools. The framework also provides a default agentic harness that can be used to evaluate any base model. It is interesting to note that the research team evaluated several LLMs on the MLGym-Bench benchmark, including Claude-3.5-Sonnet, Llama-3.1 405B, GPT-4o, o1-preview, and Gemini-1.5 Pro. As per their tests, they say Gemini-1.5 Pro is the most cost-effective option in the context of research. The team hopes that open-sourcing the framework and benchmark will facilitate future research and advance the AI research capabilities of LLM agents. You can find the code on its GitHub page. While we know the future of research is agentic, these frameworks and benchmarks will help make a difference by providing a standardised way to evaluate and compare AI research agents.","excerpt":"Meta’s new experimental framework aims to improve LLM agents that help in research.","categories":["AI News"],"tags":["ai framework","Meta"],"author_name":"Ankush Das","publish_date":"2025-02-24T14:51:28","publication_year":"2025","word_count":358,"keywords":["ai framework","Go","Meta","AI","GPT-4o","ML","Git","computer vision","GPT","Aim","GitHub","R"],"extracted_tech_keywords":["AI","ML","computer vision","GPT-4o","Aim","R","Go","Git","GitHub","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/metas-new-framework-could-help-make-the-next-ai-einstein\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":12045,"title":"Fraud Analytics to play critical role in Digital India &#038; Cashless Society","content":"Digital India is great initiative with vision to transform society in digitally empowered society. Keeping up with the current trends in technology and economic scenario in India, the decade of 2015 to 2025 would probably be termed as the “digital revolution” in India. Everything is getting Digital from individual validation, communication channels to transactions in physical currency. There has been major shift in the way even the currency is being perceived today. It would be really exhilarating to see how the physical currency would be replaced by the digital transaction. With the advent of Unified Payment Interface System, payments banks, mobile banking, etc, physical currency would to some extent be replaced by digital transaction. While there is probability that fraud involving physical cash may reduce, there is equal likelihood of vulnerabilities to increase in the digital world. However the same comes with certain advantages in the form of Digital Footprints, which the analyst and investigating professional can leverage upon. There would be significant rise in the amount of data that will get generated over the next decade in all formats. These frenetic pace and voracious streams of information in a highly mobile environment have created dangerous pitfalls and thus increasing the need for analytics for better insights. As per Global Fraud Study survey 2016 conducted by Association of Certified Fraud Examiners (ACFE); victim organizations that lacked anti-fraud controls suffered greater median losses. Analytics would play a pivotal role in fraud management process. Understanding Fraud Triangle through data analytics: Application of analytics in areas of Fraud Risk management framework: Detection of Fraud: Analyzing heap of data and deriving meaningful insights out of it has become crucial in present digital world. Use of Benford’s law may be evaluated and applied to the population. Other fraud detection techniques may range from deep diving ad-hoc analysis of existing information, pattern analysis, network analysis, outlier monitoring, and behavior analysis to developing some of the complex algorithms and models. Fraud Investigation: There are three critical aspects in any fraud investigation; understanding of the issue, forensic analysis and impact assessment. Data analytics can provide vital support in the investigation process. Investigating a fraud requires identifying root cause and evidences to prove the case. With analytics, traces left by fraudsters in the digital environment can be identified and adequate evidences can be collected. Fraud prevention: Building analytics solution by developing real time triggers for exception scenario. This would again require developing of model for behavior analysis of patterns followed by necessary action on the report. Early warning system – Providing indicators of probable vulnerabilities which may lead to Fraud and may need further analysis before conclusion. Fraud Analytics Maturity Model: It is very important for any organization to continuously assess them in line with fast paced changing business landscape. Below is the ‘fraud analytics maturity model’ which can provide some insights on how far is organizations maturity level in acceptance of data analytics in its fraud risk management process. Stage 1: Receptive: Known from others. Fraud Response based on complaint either through internal or external. No use of analytics for concluding the event Stage 2: Receptive and Scalable: Known from others but replicated for exposure. Based on Response received, data analytics carried out to see how the same is impacting overall exposure and taking corrective action. Stage 3: Pre-emptive: Self-identified. Self-identification of the fraud based on seeding\/assessment of controls through analytics. This would include deployment of analytics for continuous monitoring of controls and focused sampling. Stage 4: Preventing and predicting: This would involve developing an automated system and methods to detect fraud using a predictive model. This may involve analysis of various variables and its correlation with the desired output. An early warning system can be put in place to identify potential frauds. Five step approach to fraud analytics: Hypothesis building: Understanding fraud vulnerabilities and developing comprehensive fraud risk register is crucial before one embarks on the analytics journey. This can be broadly classified in three categories Corruption, Financial loss and financial misstatement. Mapping available data: Another important aspect in analytics is understanding the data structure and linking the same with the underlying hypothesis. Designing algorithm: This refers to exploration of suitable analytics techniques which can provide desired results. This includes developing support tools for monitoring. Testing and validation: It is inherent for any model or system to be validated before deploying the same. Deployment and continuous monitoring: For a successful Fraud analytics framework there has to be proper decision making and monitoring mechanism. Challenges in Analytics: Data availability and hygiene issues: This is one of the by far the biggest challenge in fraud analytics. Without availability of data, it would be difficult to have good analytics framework. Handling big data: Big data comes with its own set of characteristics in the form of volume, velocity and variety. Skill set: Fraud analytics would typically require blend of three skill set – understanding of fraud vulnerabilities, knowledge of data structure and strong analytic techniques. Acquisition of such skill set may sometimes become a challenge. Conclusion: Frauds will increase as transactions volumes would increase with continued digitization. Analytics not just supports marketing and sales but in growing digital scenario it has bigger role to play in risk and forensic analytics as well. The world is changing and with everything becoming automated and digitized, it is inevitable to embrace analytics in forensics and fraud detection.","excerpt":"Digital India is great initiative with vision to transform society in digitally empowered society. Keeping up with the current trends in technology and economic scenario in India, the decade of 2015 to 2025 would probably be termed as the “digital revolution” in India. Everything is getting Digital from individual validation, communication channels to transactions in […]","categories":[],"tags":[],"author_name":"Mihir Mody","publish_date":"2016-12-30T06:05:01","publication_year":"2016","word_count":894,"keywords":["big data","Go","TPU","AI","Scala","Git","RAG","analytics","R","fraud detection"],"extracted_tech_keywords":["AI","analytics","RAG","fraud detection","TPU","R","Go","Scala","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/fraud-analytics-play-critical-role-digital-india-cashless-society\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10062002,"title":"Wakefit’s Puneet Tripathi on how the love for maths shaped his data science career","content":"Puneet Tripathi hasn’t looked back since he graduated as a computer science engineer from the CR State College Of Engineering, Sonipat. He quickly rose through the ranks to become the head of data science at Wakefit, a modern furniture brand. He has showcased his sleight of hand in data science across domains, from working as a bio-statistician on a pharmaceutical project to building intent models for Wakefit. In an exclusive interview with Analytics India Magazine, Puneet discussed his journey into the world of data science and his love for all things maths. AIM: What motivated you to pursue data science? Puneet Tripathi: After graduating in 2010, I started my career at TCS. At that time, data science was gradually gaining momentum and it all fell under the single term called Analytics. In the early years, data science and its applications were only limited to business intelligence and warehousing.  Data science had a very specific use case in pharmaceuticals and banking, with credit underwriting and CIBIL scores becoming known terminologies. However, while considering multiple project-driven opportunities in TCS, my fascination with data and math guided me. Working in the lines of the scientific and mathematical aspects of engineering was my primary motivation for choosing this career choice. AIM: When did you start preparing for a data science career? Puneet Tripathi: Coming from a computer science background, I had topics like neural networks, AI-mathematics, matrix multiples and linear algebra as part of my curriculum. After getting into TCS in 2010, I got familiar with Python while working as a bio-statistician for one of the pharmaceutical clients. While evaluating clinical studies, I got familiar with statistical procedures, tests and tools utilised in the process. I started understanding the theoretical terms that I had used in my course. All these things were great motivators for me. AIM: How important is it to start early in data science? Puneet Tripathi: There is a lot of actual work happening in the data science ecosystem and projects that are actually helping people is what piques the interest of recruiters. It is imperative for aspirants to focus on the basic skills like Python, SQL, data visualisation, etc. that are involved in data science along with building a portfolio that highlights the depth of their knowledge. One of the trends that I disagree with is that one does not need mathematics to do data science. It is one of the most underrated but important factors in data science as mathematics helps you understand the result of your analysis and guides you to make the right inferences to carry the process forward. If possible, aspirants should create their profile in Kaggle to start solving problems, refer to books and experts. To boost their portfolio, one should create their profile on GitHub and maintain their own repositories, get familiar with UI\/UX kind of development. For example, one can take industry charts from various sites and try to replicate them using R, Python or any other language. AIM: Tell us about your data science journey. Puneet Tripathi: After being placed in TCS, I got a deeper understanding of the SAS (statistical analysis system) while working on a pharmaceutical project. I learnt how statistical tools can be used to make inferences and analyse the copious amounts of data on drugs before they are launched in the market. After TCS, I moved to a company called Dunnhumby, a pioneer in customer science, where I worked for five years. In this time period, I worked on big data projects and was one of the few people working on Hadoop at that time. After this, I collaborated with Hindol Basu and Bijoy Khandelwal to build Actify data labs. We did some amazing work in the retail industry. My experience in retail, e-commerce and customer science is what brought me to the role of head of data science at Wakefit. AIM: What are the major challenges\/opportunities you have come across in your career so far? Puneet Tripathi: Of the many challenges to choose from, one that stands out was the fraud identification model that I was helping build for a stock exchange firm. Not a lot of literature was available and the traditional methods took 2-3 days to calculate. We created our own mechanism to identify and predict fraud circular trades using our own rate of volatility, share movement volatility, etc. to create a trust score against malicious trade. It was a moderately powerful logistic regression model with 72% ROC. Another challenge was the amount of data to be processed. The data cache went up to 100GB and handling it in real-time was very difficult. We used the snapshot method to identify chains and predict by the end of the day if a stock will convert to a circular trade or not. Recently, we created an intent model within Wakefit itself, to identify customers and their intent to purchase a product. With the amount of data made available by the users and the data accumulated by Wakefit, we were able to create features on the intent model that depicted the customer’s interaction with our website with probability rates of conversation. AIM: Tell us about your plans\/goals to push the envelope at Wakefit. Puneet Tripathi: Since my induction in 2020, I and the team have been able to build a layer of intelligence in terms of dashboards, multiple data marts, NPS and multiple models. Our ultimate goal would be to help businesses with entire setups or frameworks for their progress. Businesses don’t care if you show them data, but they care for the relationship that they can build on. Building an entire platform for analysis and predictions and converting those into ideas for actionable development is what I strive for. Data has the power to become a stand-alone business entity or partner that can boost any enterprise to its full potential. AIM: What’s your advice for data science aspirants? Puneet Tripathi: Build your skillset, put some effort into understanding the mathematical side of data and don’t be afraid to seek guidance. People are much more receptive than we think and one can gain insights and mentorship from industry experts if you reach out to them on LinkedIn, Twitter or other sites. Understand every aspect of how things work. The essence of data is in its variation and representation.","excerpt":"The essence of data is in its variation and representation.","categories":["AI Features"],"tags":["Data Science","dunnhumby","head of data science","Interviews and Discussions","Journey into data science","TCS"],"author_name":"Kartik Wali","publish_date":"2022-03-03T13:00:00","publication_year":"2022","word_count":1050,"keywords":["data science","Go","Journey into data science","AI","neural network","R","head of data science","dunnhumby","Python","Aim","analytics","SQL","Rust","Data Science","TCS","Interviews and Discussions"],"extracted_tech_keywords":["AI","neural network","data science","analytics","Aim","Python","R","SQL","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/wakefits-puneet-tripathi-on-how-the-love-for-maths-shaped-his-data-science-career\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10112842,"title":"Stability AI Releases New AI Model, Stable Cascade","content":"Stability AI has launched Stable Cascade, a new AI model for generating images from text. This model, based on the Würstchen architecture, a text-to-image  architecture for large-scale text-to-image diffusion models. Central to Stable Cascade’s appeal is its ability to closely adhere to user prompts, a feature highlighted by the AI community on HackerNews for its precision in generating relevant images. “This model is noted for its speed, offering a significant improvement in processing times without sacrificing output quality, addressing a critical need for efficiency in AI-driven tasks,” the users point out. Available for non-commercial use, it introduces a three-stage process designed to work efficiently on consumer hardware. Stable Cascade is different from previous models because it uses a three-part system to compress and generate images. This allows for significant reductions in the resources needed for training. The model consists of stages A, B, and C, with stage C focusing on generating a compressed version of the image which is then expanded in stages A and B. The company has also released tools for training and customising the model. These include scripts for fine-tuning and other adjustments, available on the Stability GitHub page. The model supports features like image variations and image-to-image generation, adding to its versatility. In comparison tests, Stable Cascade showed better results in terms of speed and quality against other models, even those with more parameters. It offers various options for users, including different model sizes to accommodate different hardware capabilities. Stability AI has made all related code public for users to modify and experiment with. This includes features for enhancing images, generating images from sketches, and increasing image resolution. While Stable Cascade is not for commercial use, Stability AI suggests other models for those interested in commercial applications. Users on the latest","excerpt":"Stable Cascade is a text-to-image AI model designed for efficiency on consumer hardware.","categories":["AI News"],"tags":["Stability"],"author_name":"K L Krithika","publish_date":"2024-02-14T13:20:30","publication_year":"2024","word_count":295,"keywords":["Stability","TPU","programming_languages:R","AI","Git","diffusion models","GitHub","R"],"extracted_tech_keywords":["AI","TPU","R","Git","GitHub","diffusion models","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/stability-ai-releases-new-ai-model-stable-cascade\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10061765,"title":"IIMs see a rise in analytics roles at latest campus placements","content":"IIMs are arguably the most sought after schools to pursue management education in India. Come placement time, world class companies across sectors make a beeline for IIMs to handpick top drawer talents to upgrade their workforce. The placements for the batch of 2020-22 at few IIMs concluded recently. Below, we look at how the latest recruitment drive panned out at the top IIMs. IIM-Ahmedabad IIM-Ahmedabad has wrapped up the final placement process for the Post Graduate Programme (PGP) in Management. The students were placed across more than 20 cohorts. As many as 190 companies participated, with around 220 different roles being offered. Boston Consulting Group and Accenture made the most offers. Goldman Sachs was the biggest recruiter in the investment banking domain with six offers, while Bank of America made four. In the lateral process, companies offered mid-level managerial positions for students with work experience. 36 firms hired from sectors such as technology, banking, consulting and analytics. PayTM made the highest number of offers (20) while PwC (13 offers), American Express (11 offers) and OYO (10 offers) were the other top recruiters. In the second stage, firms were grouped into cohorts based on the profile offered, and groups of cohorts were invited to campus across different clusters. If we compare to last year, the private equity, venture capital and asset management cohort saw around 42% increase in the number of net offers while the management consulting cohort had a 30% increase in number of net offers, said IIM-A. In the consumer goods and general management domain, TAS made the highest number of offers (including the pre-placement offers) at 6, closely followed by HUL, Mondelez and Emaar, with 3 offers. In the IT Consulting cohort, Tata Consultancy Services was the largest recruiter with 4 offers. For a complete breakdown, click here. IIM -Bangalore As many as 662 offers were rolled out to 513 students in the lateral and final placements for the PGP and PGPBA class of 2020-22. In the analytics domain, 32 offers were made with American Express leading the charts with 17 offers while EXL and Flipkart made 12 and 3 offers, respectively. Consulting Of the 248 offers made in the consulting domain, Accenture led the way with 51 offers followed by BCG’s 30. The top recruiters included Kearney (27), Bain & Company (26), McKinsey & Company (22), Ernst & Young (9), PricewaterhouseCoopers (9), Tata Consultancy Services (9), Alvarez & Marsal (7), Arthur D. Little (7), Deloitte (5), Infosys Consulting (5), KPMG (5), Strategy& (5), Auctus Advisors (4), Oliver Wyman (4), IBM Consulting (2), and EY-Parthenon Singapore (2). IT and product management Microsoft (15), OYO (11), Amagi Labs (7), Oracle (7) accounted for 141 offers under this domain. Atlassian, Google and Info Edge and Razorpay made 6 offers each, while Angel One, Jio Platforms and UHG Optum made 4 offers each. Finance 71 offers were made in the finance domain. Goldman Sachs made the highest number of offers (22) followed by Avendus Capital (7), Citi Bank (5), Deutsche Bank (5), Rothschild (2), Premji Invest (2), Blackstone (1) and CDC Group (1). Sales and marketing 40 offers were made in this domain with HUL (6), Asian Paints (4), Samsung (3), Dr. Reddy’s Laboratories (2), ITC (2), Nykaa (2), Pidilite (2), Anheuser-Busch InBev (1), Colgate Palmolive (1), Disney-Star (1), Johnson & Johnson (1), Lenovo (1), Marico (1), Nestle (1), Procter & Gamble (1), Tata Sky (1), Thermo Fisher (1), Titan (1), and Wipro Consumer Care (1) participating in the placements. For a complete breakdown, click here. IIM-Indore IIM-Indore has completed the final placements for the Post Graduate Program (PGP) and 5-year Integrated Program in Management (IPM). Over 180 recruiters participated in the drive with 30 recruiters coming in for the first time. The average compensation offered was Rs 25.01 lakhs, an increase of 6% in comparison to last year. The median package also saw an increase of 6.6% compared to last year at INR 24.09 lakhs. The highest package offered stood at Rs 49 Lakhs. The average package of top 100 students stood at INR 37.95 lakhs. Data analytics & IT Amazon, American Express, Angel One, Axtria, Capgemini Chrysalis, CarDekho, Cipla, CityMall, Cognizant, EXL, General Electric, Google, Hevo Data, Hindustan Unilever, IBM, IQVIA, Jio Platforms, Joveo, JustDial, MagicBricks made offers in the analytics and IT space. Consulting The consulting sector recruited 31% of the batch. Accenture Strategy, Acuvon Consulting, Avalon Consulting, Bain & Company, Boston Consulting Group, Deloitte, Everest Group, Eversana, Ernst & Young, Infosys Management Consulting, KPMG, McKinsey & Company, Michael Page participated in the drive. Image: IIM-Indore Finance The companies such as Avendus Capital, Bank of America, Barclays, Credit Suisse, CRISIL, D.E. Shaw, Deutsche Bank, Finezza, Goldman Sachs, HDFC Bank, HSBC and ICICI Bank made offers in the placement drive. Sales and marketing The major recruiters included Aditya Birla Fashion & Retail Limited, Asian Paints, Bajaj Auto, BMW, Cisco, Country Delight, Dabur, Diageo, Grasim Paints, GSK Pharma, Hero MotoCorp etc. For a complete breakdown, visit here. IIM-Lucknow As many as 491 students have received 534 offers from over 150 participating organisations in the latest placement drive. The highest international salary offered stood at INR 61.59 lakhs as opposed to the highest domestic salary of INR 58 lakh. Accenture, Asian Paints, Bain & Co., Bank of America Merrill Lynch, Barclays, Boston Consulting Group, Blue Yonder, Byju’s, Capgemini, Citi, Colgate Ernst and Young, Aditya Birla Group, Goldman Sachs, PayTM, PwC, Myntra, Wells Fargo, Samsung, Vodafone-Idea etc made offers as part of the placements. The first-timers included Ambit, Arga Investment Management, Dalberg, Houlihan Lokey, Lincoln International, Meesho, Pharmeasy, Premji Invest, Spinny, Sutra, and Winzo Games etc. IIM-Kozhikode IIM-Kozhikode has concluded placements for its post graduate programmes (PGP) including for the inaugural batches of PGP Finance (PGP-F) & PGP Liberal Studies and Management (PGP-LSM). A total of 116 companies made 571 offers to 546 students. The average salary grew by 31.3% Y-o-Y to Rs 29.5 lakh per annum. The median salary also saw a 32.5% increase compared to last year and reached Rs 26.50 lakhs per annum. IT and analytics The consulting domain accounted for 41% of the total offers made, followed by BFSI, and IT and analytics at 15% each, and sales and marketing at 13%. BCG, Accenture Strategy, ABInBev CitiBank, DE Shaw, Deloitte, EY, Goldman Sachs, Google, JPMorgan Chase & Co, McKinsey & Company, Microsoft, Nestle, Nomura, PwC, Samsung, Uber, Walmart and Wipro also participated in the recruitment drive.","excerpt":"As many as 662 offers were rolled out to 513 students in the lateral and final placements for the PGP and PGPBA class of 2020-22 of IIM-Bangalore.","categories":["AI Highlights"],"tags":[],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-02-28T18:00:00","publication_year":"2022","word_count":1068,"keywords":["Go","API","AI","RPA","venture capital","RAG","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","API","GAN","ViT","RPA","venture capital"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/iims-see-a-rise-in-analytics-roles-at-latest-campus-placements\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10134239,"title":"How Joule is Helping SAP Find Moat in Spend Management","content":"In today’s fast-paced world, it is vital for organisations to optimise procurement, manage supply chains and control costs effectively. For instance, cocoa prices have surged over 400% this year, posing a significant challenge for chocolate manufacturers. Spend management tools provide comprehensive visibility into a company’s spending patterns, streamlining procurement processes. Generative AI could further optimise spend management as it can provide predictive analytics, supplier recommendations and enhanced negotiations strategies. This is exactly what SAP has done. The company has consolidated all aspects of spend management under one umbrella, calling it Intelligent Spend and Business Network (ISBN). And now, it has thrown generative AI into the mix. SAP customers are now reaping the benefits of Ariba, Concur and Fieldglass – companies which SAP has acquired over the years – under one platform coupled with SAP Business Network. While spend management is not a new concept, clubbing together the solution powered by generative AI and Joule, a proprietary AI assistant developed internally by SAP, they believe the solution holds potential. How GenAI is Changing Spend Management “At SAP, our AI-first product strategy involves deeply embedding AI into core business processes. Rather than overlaying AI onto existing workflows, we are fundamentally reimagining processes, positioning AI as a constant, collaborative partner for every user,” Jeff Collier, chief revenue officer, SAP Intelligent Spend & Business Network, told AIM. By using Ariba, a procurement tool, customers can now leverage the power of generative AI models to accelerate their planning processes across multiple categories (including 3rd party market data from Beroe) and reduce supplier onboarding time, Collier further revealed. (Jeff Collier, chief revenue officer, SAP Intelligent Spend & Business Network) Generative AI models are also helping Ariba customers make sense of historical data and derive insights from it and make recommendations in real-time, which in return is helping them in their procurement process. “To optimise external workforce and services spend for greater insight, control, and savings, SAP Fieldglass customers are now leveraging AI-enhanced SOW description generation, AI-enhanced job descriptions, and AI-enhanced translation of job descriptions,” Collier added. AI Agents are Coming If not LLMs, AI agents are expected to truly scale AI. When asked if these could be part of SAP’s ISBN solution, Collier said that SAP is focused on embedding digital assistants directly into our products through Joule, SAP’s generative AI copilot. Joule provides AI-assisted insights and will be integrated as a standard feature across the SAP portfolio, including ISBN. “In my conversations with chief executives, business leaders, and decision-makers around the world, a common theme has emerged– the urgent need to do more with less,” he said. He added that the COVID-19 pandemic has significantly expanded responsibilities, yet headcounts have remained flat or declined, making productivity a critical driver for AI interest among business leaders. “The real question now is how swiftly these leaders can embed AI into their operations to start reaping its benefits. With the proliferation of tools, complex policies, and the need to maximise return and value, organisations are seeking chat-based interfaces or agents to guide them through their tasks and answer queries efficiently,” he added. Hence, it makes sense for SAP to integrate AI agents into its entire portfolio. With Joule, SAP users can save time and increase productivity by describing their ideas, asking analytical questions, or instructing the system, rather than navigating through traditional clicks or coding. ISBN, an Interesting Prospect for India Explaining further what ISBN solutions from SAP truly mean, Ashwani Narang, Vice President, Intelligent Spend and Business Network, SAP Indian Subcontinent, told AIM that spend management extends beyond traditional procurement to include the entire supply chain, contingent labour, and employee expenses. “Various departments—marketing, finance, and others—constantly request funds, highlighting the need for comprehensive oversight. Procurement used to be the sole area focused on savings, but now spend management encompasses all financial outflows, including those outside the organisation like partners, logistics providers, and consultants,” Narang said. (Ashwani Narang, Vice President, Intelligent Spend and Business Network, SAP Indian Subcontinent) Narang believes, as the country transitions towards being a manufacturing economy with a great consensus on producing things locally thanks to initiatives like ‘Make in India’, ISBN could be a great tool for Indian companies. “The more you become a manufacturing economy, the more working capital becomes important and I believe that’s where the onus is going to be,” he said. SAP has witnessed triple-digit growth with ISBN, and according to Narang, thousands of customers in India are already leveraging the solutions. The solutions can also help companies with category management. “For instance, if a mattress manufacturer needs cotton, it’s crucial to assess whether it’s a supplier-power or buyer-power category. Knowing if enough suppliers offer the specific grade of cotton helps in negotiating prices,” he added. Narang said that Joule can provide insights into market conditions and supplier dynamics, guiding strategic sourcing and ensure better negotiations, he added. Moreover, ISBN is also not limited to large enterprises with a global footprint. The portfolio of SAP customers for ISBN includes mid-size and small enterprises as well. “Any company with significant purchasing needs—whether it’s a relatively small firm with a turnover of INR 1,000 crore or a large corporation like Microsoft with a $50 billion valuation—can benefit from SAP’s intelligence solutions.”","excerpt":"SAP customers are now reaping the benefits of Ariba, Concur and Fieldglass – companies which SAP has acquired over the years – under one platform.","categories":["Deep Tech"],"tags":["SAP"],"author_name":"Pritam Bordoloi","publish_date":"2024-09-02T15:35:48","publication_year":"2024","word_count":869,"keywords":["Go","GenAI","AI","R","SAP","ML","RAG","Aim","analytics","generative AI","predictive analytics"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","GenAI","Aim","RAG","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-joule-is-helping-sap-find-moat-in-spend-management\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10104945,"title":"Intel Releases 5th Gen Xeon Processors","content":"At the Intel AI Everywhere event, the company has revealed the forthcoming release of 5th Gen Xeon processors, featuring AI acceleration in every core and expected to hit the market in 2024. Unveiled by Intel CEO Pat Gelsinger, these processors, previously codenamed Emerald Rapids, mark a significant advancement in computing. Designed to cater to AI, high-performance computing, networking, storage, databases, and security needs, they aim to enhance performance while minimising the total cost of ownership (TCO). Intel Xeon powered data centres would be used by Microsoft, Google Cloud, and IBM, and many more would integrate them. Sandra Rivera, Intel’s executive vice president and general manager of Data Center and AI Group, underscored the importance of this development. “Built for AI, our 5th Gen Intel Xeon processors provide greater performance to customers deploying AI capabilities across cloud, network, and edge use cases. We’re launching 5th Gen Intel Xeon on a proven foundation that will enable rapid adoption and scale at lower TCO,” she stated. Key features of the 5th Gen Xeon processors include AI acceleration in every core, optimizing them for AI workloads and delivering up to 42% higher inference performance with minimal latency on large language models. This enables end-to-end AI processing without the need for additional accelerators, making AI tasks more efficient. In terms of general compute performance, the processors offer an average gain of 21% compared to the previous generation. They also boast a 36% increase in performance per watt across various customer workloads, translating to significant efficiency improvements. These processors support up to 64 cores per CPU, enhanced last-level cache, eight channels of DDR5, and higher memory transfer speeds. These improvements contribute to overall enhanced performance and bandwidth. Additionally, they ensure compatibility with CXL Type 3 workflows through leading cloud service providers. Furthermore, the processors come equipped with security enhancements, featuring Intel Trust Domain Extensions (Intel TDX). This provides increased confidentiality and security at the VM level, ensuring enhanced privacy and control over data. Looking ahead, the 5th Gen Xeon processors, which are pin-compatible with the previous generation, are set to be available in single- and dual-socket systems from leading OEMs such as Cisco, Dell, HPE, Lenovo, and others by the first quarter of 2024. Major cloud service providers will announce the availability of instances based on these processors throughout the year. Intel remains committed to its roadmap, with plans to introduce Sierra Forest, emphasising E-core efficiency with up to 288 cores, in the first half of 2024. Following closely will be Granite Rapids, focusing on P-core performance. The company has also announced the launch of AI PCs, to build and run AI on every PC. Gelsinger announced the company’s partnership with Dell, HP, Lenovo, Supermicro, and Microsoft for onboarding the chips on their devices.","excerpt":"In terms of general compute performance, the processors offer an average gain of 21% compared to the previous generation.","categories":["AI News"],"tags":["Intel","Intel processors"],"author_name":"Mohit Pandey","publish_date":"2023-12-14T21:35:26","publication_year":"2023","word_count":458,"keywords":["Rapids","Go","API","programming_languages:R","AI","Intel processors","RAG","Aim","cloud_platforms:Google Cloud","Rust","R","Intel"],"extracted_tech_keywords":["AI","Aim","Rapids","RAG","R","Go","Rust","API","cloud_platforms:Google Cloud","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/intel-releases-5th-gen-xeon-processors\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10072259,"title":"Google’s Zanzibar leads way for authorisation systems","content":"Sample this: You sent a Google Docs link to your editor. A few minutes later, you receive an email from her requesting access to the document. Of course, you act immediately and give your editor the ‘edit’ access. This is authorisation. In computer systems, authorisation is part of the IT discipline called Identity and Access Management (IAM). It is a security mechanism to grant or deny someone access to a network resource such as files, data, application features or computer programs. Why has authorisation become so vital? In the past few years, addressing the need for authorisation has become vital in our day-to-day life in general, and the IT industry in particular. As businesses move towards cloud-based platforms, the need for security has become ever-so-important. An organisation gives designated individuals access to its systems and not all users need to have the same level of access to the organisation’s systems, applications, data and other resources. Operating systems today use authorisation processes to deploy and manage applications. However, unauthorised access to cloud-based systems can prove disastrous. Without authorisation, people with malicious intent can access an organisation’s confidential resources impacting its business operations. Added to it are reputational damage, potential lawsuits, issues of non-compliance and imposition of fines. Moreover, sometimes, an enterprise’s clients might have to bear the brunt — sensitive data can leak across the internet. Zanzibar — Google’s authorisation system In 2019, Google published a paper titled ‘Zanzibar: Google’s Consistent, Global Authorisation System’ that delves into the details of Zanzibar, a system for storing permissions and performing authorisation checks based on the stored permissions. Zanzibar is a globally distributed authorisation system that handles authorisation for a wide array of services offered by Google, including Calendar, Cloud, Drive, Maps, Photos, and YouTube. Zanzibar is flexible, global and superfast. It allows Google teams to specify their unique authorisation models and globally replicates authorisation data. Zanzibar can easily scale to handle millions of authorisation requests per second across billions of users and trillions of objects with very low latency. In over three years of production use, Zanzibar has maintained 95th-percentile latency of less than 10 milliseconds. To maintain such low latencies, Zanzibar uses secondary indexing for heavily nested groups, request hedging and distributed caching. Open-source authorisation systems Recently, a few open-source authorisation systems have come up inspired by Google’s Zanzibar. Ory built an open-source authorisation system called Ory Keto, which is an implementation of Zanzibar. New York-based startup Authzed released an open-source version of Google’s Zanzibar called Spice DB. Spice DB Spice DB is the open-source Zanzibar- inspired database that stores, computes and validates fine-grained permissions. SpiceDB provides verifiable correctness that ensures security of the system. SpiceDB has been designed so that it not only helps decouple policy from the application but also the data that policies operate on. It provides a single unified view of permissions across several applications that a certain organisation has. SpiceDB has dedicated APIs for checking individual permissions, listing all access and ACL (Access Control List) filtering. Also, a powerful graph engine supports distributed, parallel evaluation. Ory Keto Ory Keto is an open-source implementation of Zanzibar. It is flexible, consistent, highly available and has low latency. Ory Keto is based on a simple, but powerful data model with effective configuration capabilities that serves the needs of different kinds of clients with different access control patterns. As a policy decision, Ory Keto uses a set of access control policies to determine whether a subject (user or application) is authorised to perform a certain action on a resource. Currently, Ory Keto implements basic API contracts for managing and checking “permissions” with HTTP and gRPC APIs. In the future, there are plans to ensure consistency guarantees using snap tokens, interoperability with other Ory products like Ory Hydra and Ory Kratos and incorporate a global spanning cluster operation mode. Apart from the above-mentioned open-source authorisation systems, some companies have developed their own authorisation systems. For example, based on Zanzibar, Airbnb created its own centralised authorisation system, Himeji. Carta, a global ownership management platform that helps companies, investors, and employees manage equity and ownership, came up with AuthZ — a highly scalable permissions system. Such is the importance of authorisation these days that several types of authorisation strategies have come up, the prominent ones being role-based access control (RBAC), attribute-based access control (ABAC), graph-based access control (GBAC) and discretionary access control (DAC). In fact, of late, Auth0, an authentication and authorisation platform, has been engaged in a new strategy called relationship-based access control (ReBAC). Each strategy helps application developers deal with different authorisation requirements and services to ensure and improve overall system security.","excerpt":"Without authorisation, people with malicious intent can access an organisation’s confidential resources impacting business operations","categories":["Global Tech"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-08-07T13:00:00","publication_year":"2022","word_count":770,"keywords":["Go","API","AWS","AI","Scala","Ray","gRPC","ViT","GAN","R"],"extracted_tech_keywords":["AI","Ray","AWS","R","Go","Scala","API","gRPC","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/googles-zanzibar-leads-way-for-authorisation-systems\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10170657,"title":"Why Programmers Are Using Problematic AI Code?","content":"While AI-generated code promises benefits like faster delivery, reduced boilerplate, and increased productivity, it often contains considerable flaws, including hallucinated functions and serious security vulnerabilities. Yet developers continue to use it. The question isn’t whether AI-generated code is broken but rather why developers rely on it. After two years of hands-on experience with AI pair programmers, from early tools like Cursor to more advanced agentic setups, Namanyay Goel, founder of Giga AI, describes a developer’s daily trade-off between code quality and efficiency. He summarised in a recent blog post, saying, “AI coding tools are simultaneously terrible and necessary.” The Cold Economics Behind Messy AI Code Goel highlighted that businesses prioritise rapid product releases over code quality, leading to widespread acceptance of flawed software. Although there has always been a divide between developers and businesses, AI has made it wider. “As a programmer, since I live inside my codebase, I want to make it perfect,” wrote Goel. “But from the perspective of a business, my code is merely a tool to generate revenue.” He pointed out that AI tools are potentially favoured not for their elegance, but for their speed. A former engineering manager once told Goel, “Nobody’s getting promoted for clean code if the feature ships late.” That said, some developers do not have a choice but to use the AI tools because the company invested in them. A Redditor said their company had little choice and had to spend money on Copilot licenses. Yes, It’s Flawed. Yes, It’s Still Useful The cracks are glaring. AI struggles with anything beyond mainstream stacks. Goel recalls a WebGL project where it generated code that looked plausible, but simply didn’t work. Even when the output functions, it often lacks grace. “The code worked, but the AI completely lacked the ability to implement an elegant solution,” Goel wrote. Birgitta Boeckeler, global lead for AI-assisted software delivery at Thoughtworks, explained the problem with AI-generated code by categorising it into three areas of impact. First, it slowed down the development speed and the time to commit. Second, it created friction for the team flow in that iteration. Lastly, it affected the long-term maintainability of the code. AI coding can also be a security nightmare. Chetan Gulati, a senior DevSecOps engineer at Fraud.net, told AIM, “AI coding does present significant security challenges. Generative AI, at its core, is an advanced sentence completion system, making it susceptible to prompt injection attacks that could introduce sensitive details or vulnerable code into a system.” Despite concerns, developers are embracing AI code generation due to its genuine utility in specific applications. In a recent blog post, Github’s Johan Rosenkilde, principal researcher for GitHub Next, highlighted that 95% of the time, Copilot makes his day a little bit easier. According to GitHub’s own stats, 88% of developers felt more productive, and 60% said it made their jobs more fulfilling. Earlier this year, in a Y Combinator podcast, partner Jared Friedman said that one-quarter of the YC founders admitted that over “95% of their codebase was AI-generated”. He pointed out that these were highly skilled founders who, just a year ago, would have built their products entirely on their own—but now, AI does the heavy lifting. Another developer on Reddit commented, “I hate it, but it’s a useful tool. I’m retired, but I have a little project I’m working on for my own entertainment.” He added that Web dev has become insanely complicated. “Google\/StackOverflow used to be enough to get me unstuck, but web toolkits move so fast that it’s full of out-of-date answers that add more confusion.” The developer further stated that they started using Copilot and found it good, even if they did not trust the code. Last year, there were similar sentiments around the usefulness of StackerOverflow, which is where AI tools seem to be chiming in. Goel recommends a layered approach: Use AI to draft, then verify ruthlessly. Break problems into smaller chunks, give full code context, and—crucially—keep a “WTF AI” journal of the tool’s most ridiculous missteps. He noted that those failures compile an internal library of lessons learned, indicating what should not be repeated in the future. Goel concurs, citing documentation parsing, boilerplate generation, and simple debugging as areas where AI earns its keep.","excerpt":"“I hate it, but it’s a useful tool.”","categories":["AI Features"],"tags":["coding"],"author_name":"Ankush Das","publish_date":"2025-05-23T17:02:08","publication_year":"2025","word_count":707,"keywords":["Go","TPU","AWS","AI","coding","Git","Aim","generative AI","Rust","GitHub","R"],"extracted_tech_keywords":["AI","generative AI","Aim","AWS","TPU","R","Go","Rust","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-programmers-are-using-problematic-ai-code\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10097615,"title":"Top 8 Highlights from India’s Google ML Community","content":"Globally, over 35 Google developer communities have hosted ML campaigns distributed by the ML developer programs team during the first half of 2023 showcasing their work and interests. Out of the many, we’ve picked out our favourite contributions by the Indian community. Here are the top 8 highlights from Indian Machine Learning Google Developer Experts. Image Segmentation Using Composable Fully-Convolutional Networks In a demo, Suvaditya Mukherjee provided a detailed account of implementing a fully-convolutional network using a VGG-16 backend and its application in image segmentation. Furthermore, in the KerasCV for the Young and Restless session he delved into the fundamental computer vision components. The presentation showed the significance of this tool and highlighted its integration with TFX and Keras ecosystem. [ML Story] My Keras Chronicles Aritra Roy Gosthipaty shared his journey into the world of deep learning, specifically with Keras. Within his narrative, he offered insights on how aspiring individuals could become a part of the open source community. Alongside him, Subvaditya presented their Keras implementation of Temporal Latent Bottleneck Networks, as originally proposed in the paper. TensorFlow and Keras Implementation of the CVPR 2023 Paper Usha Rengaraju showcased a research paper’s practical application, bringing to light the implementation of BiFormer – a novel Vision Transformer enriched with Bi-Level Routing Attention. Skeleton Based Action Recognition: A Failed Attempt Ayush Thakur provided insights into his experience of participating in the Kaggle competition, “Google – Isolated Sign Language Recognition“. He shared his repository, training logs, and the various approaches he employed during the competition. Furthermore, Ayush’s enlightening article, “Keras Dense Layer: How to Use It Correctly,” delved into the dense layer in Keras and its practical applications. Add Machine Learning to Your Android App At Tech Talks for Educators, Pankaj Rai conducted a session attended by 700+ individuals about how to bring ML capabilities into Android applications, such as object detection and gesture recognition. During the session, he explained ML Kit, MediaPipe, and TF Lite tools, demonstrating their capabilities and instructing attendees on how to effectively use these resources. Google’s Bard Can Write Code Bhavesh Bhatt showed the coding capabilities of Bard, how to create a 2048 game with it, and how to add some basic features to the game. He also uploaded videos about LangChain in a playlist and introduced Google Cloud’s new course on Generative AI. In a demonstration, Bhatt showcased the coding capabilities of Bard like how to create a 2048 game using it. Furthermore, he illustrated the process of adding features to enhance the game’s functionality. Alongside this, he shared videos on LangChain and Google Cloud’s new Generative AI course. Open and Collaborative MLOps In favour of the open-source community, Sayak Paul spoke about why open and collaboration are two important aspects of MLOps. He gave an overview of Hugging Face Hub and how well it integrates with TFX in MLOps workflows. During a discussion, Paul spoke about the significance of openness and collaboration in the context of MLOps. He further outlined the merits of Hugging Face Hub, showcasing its integration with TFX. Learning JAX in 2023: Part 3 — A Step-by-Step Guide to Training Your First Machine Learning Model with JAX Aritra Roy Gosthipaty and Ritwik Raha together showed how JAX can train linear and nonlinear regression models and the usage of PyTrees library to train a multilayer perceptron model.","excerpt":"Out of the many ML contributions from the community, we’ve picked out our favourite contributions by the Indian community","categories":["AI Trends"],"tags":["AI community","India AI","Open Source","open source community"],"author_name":"Tasmia Ansari","publish_date":"2023-07-26T18:00:00","publication_year":"2023","word_count":553,"keywords":["India AI","machine learning","Open Source","AI community","AI","ML","MLOps","computer vision","LangChain","deep learning","open source community","generative AI","JAX","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","computer vision","generative AI","LangChain","MLOps","TensorFlow","JAX"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-8-highlights-from-indias-google-ml-community-in-2023\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10047238,"title":"Deep Mind OpenSources Perceiver IO, Its Latest DL Model","content":"DeepMind has open-sourced their general-purpose deep learning model Perceiver IO. The tool can handle many different inputs and outputs and serves as a ‘drop-in’ replacement for transformers. The original Perceiver model supported many kinds of inputs but was limited to producing straightforward outputs. Its successor is Perceiver IO, which can handle arbitrary outputs in single addition to random inputs. It is a more general version of the original architecture. Broadening the model’s capacity, Perceiver IO is a single network that can easily integrate and transform arbitrary information for arbitrary tasks. Perceiver IO’s research paper states, “Perceiver IO overcomes the limitation without sacrificing the original’s appealing properties by learning to flexibly query the model’s latent space to produce outputs of arbitrary size and semantics.” The model is suitable for different applications given its capacity to produce various outputs from various inputs. For instance, it can perform in real-world domains like language, vision, and challenging games like StarCraft II, all of them having a multimodal understanding. Perceiver IO can classify labels, produce language, optical flow, and multimodal videos with audio using the same building blocks as the original model. It can handle and process large size input-outputs better than standard Transformers, given its linear computational complexity. The bulk of the processing occurs in the latent space to further facilitate this. This allows Perceiver IO to perform BERT-style masked language modelling by directly using bytes (and not tokenised inputs). The Hurdle for Transformers Built on the Transformer architecture, Perceiver uses ‘attention’ to map inputs and outputs. Traditionally, this allows the model to process inputs after comparing all of its elements and basing them on their relationship and the task. However, while attention is widely used, it becomes expensive as the inputs grow, such as in common forms of data like images and videos containing millions of elements. Perceiver’s Architecture The Perceiver IO has overcome the above-mentioned issue by “scaling the Transformer’s attention operation to substantial inputs without introducing domain-specific assumptions”. The model architecture uses cross-attention to project high-dimensional input arrays into a lower-dimensional latent space. These can later be processed, but at a cost that is independent of the input’s size. Lastly, the latent representation is converted to output by applying a query array with the same number of elements as the desired output data. Deep models can flourish in this setting since the computational needs can grow along with the input growth. Credits: DeepMind Blog – The PerceiverIO Architecture The three steps of the Perceiver IO pipeline: Inputs are encoded to a latent spaceThe latent representation is refined via many layers of processingThe latent space is decoded to produce outputs I was mostly responsible for the optical flow section in Perceiver IO. I admit I half expected it to memorize the train set, and prepared a bunch of clever ideas to force generalization, but these weren't needed. I'm still a bit in disbelief that it worked so well out-of-the-box. https:\/\/t.co\/YTD7LRkLD8— Carl Doersch (@CarlDoersch) August 3, 2021 Features This growth allows Perceiver IO to achieve that unprecedented level of generality and versatility over the original model that could only produce one output per input. In addition, Perceiver IO is competitive with domain-specific models on benchmarks based on images, 3D point clouds, and audio and ideas together, making it a fit for researchers. Along with attention to encoding, Perceiver IO also uses it to decode from the latent array – enhancing the flexibility of the network, scaling it to more extensive & more diverse input-outputs, all the while dealing with many types of data at once. This feature makes Perceiver IO a supermodel that can perform various applications like understanding the meaning of a text from each of its characters, playing games, tracking the movement of all points in an image, and processing the sound, images, and labels that make up a video. This is possible while just using a single architecture that’s simpler than the alternatives. Deepmind’s experiments concluded that Perceiver IO could work across a wide range of benchmark domains, including language, vision, multimodal data, and games. It successfully provides an off-the-shelf way to handle many kinds of data. To help researchers and machine learning communities at large, Deepmind has now open-sourced its code.","excerpt":"The model is suitable for different applications given its capacity to produce various outputs from various inputs.","categories":["AI Features"],"tags":["deepmind open source"],"author_name":"Avi Gopani","publish_date":"2021-08-29T12:00:00","publication_year":"2021","word_count":702,"keywords":["machine learning","TPU","AI","Modal","deepmind open source","Transformers","BERT","Ray","deep learning","transformer architecture","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","Ray","Transformers","TPU","R","transformer architecture","BERT","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/deep-mind-opensources-perceiver-io-its-latest-dl-model\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":52900,"title":"Is AI’s Bias Algorithm Affecting the Education System","content":"As we move forward with the current technological revolution, reading artificial intelligence’s name being mentioned in every sector of life might not be surprising anymore. Whatever human beings have their hands on, artificial intelligence (AI) will be right behind them, why wouldn’t it? Artificial Intelligence learns from human beings to make their life easier. One of the critical areas that AI is set to impact is the education sector. However, it faces a set of challenges like public privacy, bias algorithm in education; training teachers to get accustomed to AI, and training AI to understand the education system and to provide a sound dataset to avoid problems. With other essential aspects, one of the enormous problems artificial intelligence is facing is the problem of the bad data that has been given which has resulted in an even more significant problem, which we know as algorithm bias. And, this problem of algorithm bias has now crept into the education system too. Why is the algorithm biased? AI learns from the data provided to it by the human being who designed it. So, the kind of data that is produced depends upon the developers; thus, the developer inherently will be held responsible for the algorithms bias. The more bad data is fed to the algorithm, the more bias it will be. While people and data experts argue about the level of efficiency AI can have in education systems, it still faces a lot of problems when it comes to algorithm bias. The industry requires more diligence and extensive exploration of the field when it comes to deploying technology in that field. It is unfortunate to say that in times like this, racism and discriminations are embedded in our education system, so developers need to make efforts on developing algorithm and datasets which are not racial. There are many examples around the world where AI has been shown levels of bias. How is the algorithm biased? A study by an education software solution provider saw a predictive algorithm bias against Guamanian students. The snapshot of the data is given below: The above graph gives the likelihood of a student passing the test. The grey bar represents the actual proportion, and the teal ones are the ones which are provided by their algorithm. By looking at the data, it is clear that the algorithm showed the amount of bias towards the Guamanian students. The original data used to train the system had very few Guamanian students, which was later fixed after resampling the data and retraining of the algorithm. But, what if this algorithm bias was never have been questioned? The Guamanian students would have been falsely penalised by the algorithm. Questions will definitely arise when such cases come to light in an actual scenario. The developer has to be able to identify these biases and should be able to show evidence about those algorithm’s bias. Another problem with datasets is the number of people and the types of student information that is used in the algorithm’s training. If developers run training on the algorithm with fewer people belonging to specific skin colour, then the algorithm might take these group as less favoured. This problem will amplify when the algorithm is used in practical purposes in the education system, like the evaluation of the SAT or for some recruitment programs etc. Apart from racial bias, AI can also create problems when it comes to its applications where it assesses student’s prior and ongoing learning, placing students in appropriate subject levels, individualising instructions and scheduling. The algorithm doesn’t consider students’ experiences, low-income ability and students in minority groups who are at a relatively small achievement record. The algorithm might also assign a bad instruction set and reduced expertise when it comes to a student with a low seeding record, which in turn can hinder student’s growth. Outlook Well, on the other hand, artificial intelligence is not all that bad when it comes to making a positive impact on the education system. AI provides aids in a more efficient and time-saving manner for the teachers to pay more attention to the students and also helps in providing global access to education for students. However, unfortunately, algorithm bias has become such a common problem which ends up forcing people around the world to question AI’s promise to make human lives more comfortable.","excerpt":"As we move forward with the current technological revolution, reading artificial intelligence’s name being mentioned in every sector of life might not be surprising anymore. Whatever human beings have their hands on, artificial intelligence (AI) will be right behind them, why wouldn’t it? Artificial Intelligence learns from human beings to make their life easier. One […]","categories":["AI Features"],"tags":["expert system examples"],"author_name":"Sameer Balaganur","publish_date":"2019-12-31T16:30:00","publication_year":"2019","word_count":725,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","expert system examples","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-ai-bias-algorithm-affecting-education-system\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10053022,"title":"H2O.ai Raises $100M Series E Funding, Plans To Expand Its AI Development Platform","content":"AI-backed cloud platform H2O.ai recently announced that it has closed a $100 million funding round led by one of its customers, the Commonwealth Bank of Australia, with participation from Goldman Sachs, Pivot Investment Partners and several others. The Commonwealth Bank of Australia, or CommBank, is one of more than 20,000 organisations that use H2O.ai’s namesake artificial intelligence development platform as part of their machine learning initiatives. The startup also counts Nvidia Corp. and AT&T Inc. among its customers. H2O.ai’s platform, the H2O AI Cloud, aims to simplify much of the work involved in machine learning projects as building an enterprise AI model involves numerous highly technical tasks that require specialised know-how and take up a great deal of time when done manually. Furthermore, once development is complete, and a neural network runs in production, ensuring it operates reliably represents another major technical challenge. The H2O AI Cloud also helps with the feature engineering phase of AI projects. Feature engineering is the process by which an AI is trained on which factors it should consider most strongly when making a decision, a task that H2O.ai promises to simplify with its platform. “H2O.ai is picking winners in every vertical on every continent and making our customers AI superpowers to take on the tech giants. CommBank and H2O.ai are led by our core belief that we can make the world better while serving our communities and customers with excellence,” said Sri Ambati, Founder and Chief Executive Officer, H2O.ai. CommBank Chief Executive Officer, Matt Comyn, said, “The collaboration will bolster the bank’s analytic capabilities and help improve the predictive accuracy of existing models, so we’re able to offer more personalised and targeted solutions at a faster rate.” H2O.ai has secured more than $250 million in funding to date. The startup says it raised the $100 million round announced today at a pre-money valuation of $1.6 billion. The proceeds from the round, H2O.ai added, will be used to grow its workforce, add more features and launch new products. H2O.ai and CommBank announced a partnership today in conjunction with the startup’s funding news. Since its previous funding round in 2019, H2O.ai has added a long list of capabilities to its platform to automate other aspects of machine learning projects. Among the additions are more than 45 pre-packaged AI applications aimed at several different industries, including healthcare, finance, retail and manufacturing, among others.","excerpt":"H2O.ai’s platform, the H2O AI Cloud, aims to simplify much of the work involved in machine learning projects as building an enterprise AI model involves numerous highly technical tasks that require specialised know-how and take up a great deal of time when done manually.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Models","Cloud Computing","Cloud Platform","Data Science","Data Scientist","Deep Learning","Funding","Machine Learning"],"author_name":"Victor Dey","publish_date":"2021-11-08T11:41:21","publication_year":"2021","word_count":396,"keywords":["Funding","R","Cloud Platform","AI Models","artificial intelligence","Data Science","startup","Go","machine learning","AI","neural network","Machine Learning","feature engineering","GAN","Aim","Cloud Computing","Deep Learning","Data Scientist","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","Aim","R","Go","feature engineering","GAN","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/h2o-ai-raises-100m-series-e-funding-plans-to-expand-its-ai-development-platform\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10083622,"title":"The Curious Case of Linux: It&#8217;s for Everyone, but Nobody Uses it","content":"Linux has grown from a project operating as an antithesis to greedy corporations to a burgeoning ecosystem with something for everyone. Still, the barrier to entry has stopped many users from even trying the OS. Linux has solidified its position as the go-to for enterprises operating servers and data centres, but consumers have become intimidated by the perceived dread of using the command line and not having a familiar GUI. This user inertia, coupled with the common practice of computer manufacturers bundling Windows with their products, has led to Linux users becoming a dwindling minority in the consumer world. Today, less than 2% of computers run Linux. How much of the fear, uncertainty, and doubt around using Linux is true, and how much of it is baseless rumours from 10 years ago? Linux is for everyone In 1991, Linux first started as a personal project by a student in Finland called Linus Torvalds. He set out to create a free OS kernel based on the then famous MINIX-OS, which was a derivative of the Unix operating system. This project was picked up by the open source community, with many important features, like a GUI, dependency packages and centralised code repositories being added after the initial release. What began as a college student’s project has now become an ecosystem of over 600 actively maintained distributions (OS) of Linux, called ‘distros’ for short. In addition to these, it is easy for developers to fork off from a mainstream distro and add or remove their own features, thus creating an OS tailored to their specific requirements. This allows the user to only have the features they need, and also reduces bloat and resource consumption, allowing the distro to be deployed on the lowest-powered of devices. Timothy Timmer, a Quora user, stated. “Linux users know that no OS is perfect. Every Linux distro has assets and drawbacks. The great thing about Linux is that you can choose the distro with assets you like and the detriments that bother you.” While many tout this customizability as one of the biggest selling points of Linux, there are some who consider this to be one of the biggest hurdles of the operating system. For those who do not wish to customise the OS deeply, there are distros like Pop!_OS that function out of the box with minimal configuration, just like Windows or Mac. Distros like Ubuntu, Pop!_OS and ElementaryOS also vastly reduce the barrier for entry into the Linux world, offering easy-to-use graphical user interfaces, compatibility for devices, and software drivers. This means that users can get up and running with minimal command line wizardry, which was considered impossible just 10 years ago in the Linux world. As data privacy becomes mainstream, more users are becoming aware of the myriad ways that corporations like Microsoft and Google misuse their data. From the beginning, Linux has been free and open source, allowing coders and programmers to look into the source code. For the not-so tech savvy, this means that there are no hidden backdoors that can give corporations access to your data. Linux also does not have any telemetry that sends data back to the company maintaining the OS, something which both Microsoft and Apple do. If one has the technical know-how to go through the Linux codebase, the OS requires no blind trust from the user to function, making it the only trustless OS on the market today. But why nobody uses it There are three main reasons that users shy away from using Linux. The first is the perceived unintuitiveness of the OS, which is the biggest fear of new users. The second is the lack of support for applications, games, and devices – a problem that has plagued Linux forever. The third, and most questionable, is the toxic fanbase associated with the operating system, which commonly undermines the efforts of newcomers to the ecosystem. Command line interface nightmares are the most-quoted reasons for newcomers to join the ecosystem. In addition to this, software developers rarely optimise applications for use in Linux, making compatibility a nightmare for creators and power users. To combat this, the community has come up with distros that inherently require less technical know-how than others. One of the best examples of this is Pop!_OS. From a GUI that supports gesture-based navigation and keyboard shortcuts to compatibility with a wide range of commonly-used software, this OS does away with the high learning curve associated with Linux. It also offers a familiar work environment for those used to working in Windows or MacOS. Another major problem that average users have with Linux is not only the lack of software, but a lack of support for games. Games are usually not made for Linux and rely on proprietary technology like DirectX, which is available only on Windows. Moreover, some online games require Windows to allow their anti-cheat functionality to work, completely blocking off Linux users. Valve, the company behind Steam, is working on creating a compatibility layer for games to run on Linux, but the project is far from where it needs to be to enable mainstream adoption. Companies creating productivity software shy away from making Linux versions, as Linux users make up less than 5% of their user base. In addition to a low user count, developing applications for Linux is a nightmare, as there is a distinct lack of developer tools for the OS. This means that companies will have to spend a lot of resources developing a product for an OS that only a small chunk of their user base will actually utilise. Devices are also a pain point for Linux users, as device manufacturers do not create device drivers for use in Linux. Drivers are a middleware that communicate the capabilities of a certain device to the OS. Without these drivers, devices may not work properly and, in some cases, not work at all. Even the most basic components like WiFi and bluetooth cards don’t work on Linux without additional setup; something which users are not willing or are incapable of doing. …Except for developers The high barrier for entry does restrict a majority of mainstream users from picking up Linux, but this is the status quo for developers. Whether it is trying to find dependencies for their program or trying to solve an obscure error, programmers are used to trying to Google for solutions. This, coupled with Linux’s low overheads and compatibility with a wide range of IDEs (except VSCode), has made it a no-brainer for developers. According to a survey by Stack Overflow, 40% of developers use Linux for both work and personal use, with an additional 15% using the Windows Subsystem for Linux. This not only shows the penetration of Linux in the developer ecosystem, but also shows the requirement of Linux compatibility when developing applications or enterprise use. Newer distros have tried their best to break this mould of Linux being used for only development tasks by extending an olive branch to users of Windows and Mac, but the ecosystem and community still has a long way to go before it becomes a mainstream option.","excerpt":"Linux is one of the cornerstones of the developer ecosystem, but the powerful operating system has not broken into the mainstream as of yet","categories":["AI Features"],"tags":["Linux Kernel","linux mint","Linux Ubuntu"],"author_name":"Anirudh VK","publish_date":"2022-12-31T16:00:00","publication_year":"2022","word_count":1188,"keywords":["Go","Linux Kernel","Linux Ubuntu","AI","programming_languages:R","programming_languages:Go","RAG","ViT","Rust","GAN","linux mint","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","RAG","R","Go","Rust","GAN","ViT","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-curious-case-of-linux-its-for-everyone-but-nobody-uses-it\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":59762,"title":"Now, Evaluate Your risk To Covid-19 With Reliance Jio App","content":"With misinformation spreading faster than the virus itself, Reliance Jio has a solution. The telecom operator has launched a feature of its app that seeks to rein in fake news, offering verified information to users about Covid-19. This tool, located under the hamburger button of the app, offers a list of helpful options which users can benefit from. As soon as you click on ‘coronavirus info & tool’, you will be greeted with a message that says: Coronavirus disease (COVID-19) is an infectious disease caused by a new virus that had not been previously identified in humans. Please use the below options to learn more about the disease and keep your self-update. On further exploration, users can access other options like symptom checker, statistics, test centers, FAQs, as well as a helpline. For instance, by choosing symptom checker, users will be asked to fill a simple questionnaire which will tell them what their chances could be of contracting the virus. ALSO READ: Free AI-Based App To Be Available Soon For Faster Detection Of Covid-19 Reliance’s #CoronaHaaregaIndiaJeetega Initiative In addition to updating useful statistics around the spread of Covid-19, including the total number of cases in India, its FAQ section offers users all the information available about the disease. It also comes with helpline numbers of all the states in India. According to the company, this information is provided by qualified medical practitioners. As part of Reliance’s #CoronaHaaregaIndiaJeetega initiative, the company has also rolled out a ‘work from home’ pack for Rs 251 for remote workers. Additionally, it has also teamed up with Microsoft Teams to combine its digital capabilities with the professional messaging platform. Furthermore, it had also announced the launch of a messaging bot for MyGov Corona Helpdesk through its subsidiary Jio Haptik.","excerpt":"With misinformation spreading faster than the virus itself, Reliance Jio has a solution. The telecom operator has launched a feature of its app that seeks to rein in fake news, offering verified information to users about Covid-19. This tool, located under the hamburger button of the app, offers a list of helpful options which users […]","categories":["AI Features"],"tags":["Coronavirus","covid-19","Jio AI Cloud","reliance jio"],"author_name":"Anu Thomas","publish_date":"2020-03-24T15:44:25","publication_year":"2020","word_count":294,"keywords":["Go","covid-19","AI","programming_languages:R","programming_languages:Go","Git","Coronavirus","Jio AI Cloud","reliance jio","R"],"extracted_tech_keywords":["AI","R","Go","Git","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/now-evaluate-your-risk-to-covid-19-with-reliance-jio-app\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":568,"title":"Want to go to infinity and beyond &#8212; here’s how to make a career in space","content":"Want to explore careers in space. If you are an ardent space fan and want to explore the outer universe, a STEM education provides a plethora of opportunities to kick-start this exciting job opportunity. In this article, Astronaut Today lists down some of the space science career opportunities which are not just limited to an astronaut. In fact, while astronauts make up a sizeable number of space workforce, their selection criteria is rigorous. Must have skills to become an astronaut According to NASA, some of the astronaut requirements are: Stringent physical fitness Flight test experience and a minimum 1000 hours of pilot in-command time Bachelor’s degree in mathematics, engineering, physical science or biological science and they prefer advanced degrees as well However, making it to this job is hard and competition is fierce. But there are still a lot of ways to make a start for a career in space exploration. Some of the job roles you can explore to start a career in space: Aerospace or Aeronautical Engineers: Becoming an aerospace or aeronautical engineer will definitely open a lot of doors and you can start by working with aerospace biggies such as Boeing and Airbus on the design of aircraft. There are many opportunities for aerospace engineers in commercial and military as well. In fact, avionics, a sub-field of aeronautical engineering has gained traction over the last few years, thanks to the surge of private space companies. Avionics powers the electronic systems that are used in space crafts, artificial satellites and aircraft.  The Indian Institute of Space Science and Technology offers a B.Tech Avionics Programme that equips students with electrical, electronics and computer systems knowledge and teaches control system modelling and control design. To get a better idea of upcoming jobs in avionics, have a look at what Elon Musk backed SpaceX is looking for. Space Scientists: Earning a doctoral degree in astrophysics, astronomy and even physics can pave the way to become a space scientist. Space scientists give us a peek inside universe’s best kept secrets such as undiscovered black holes and this career is grounded in physics. Astrophysics is one branch of space science that uses physics to understand the distance, life, birth of planets, galaxies, starts and other objects in the universe. The other two branches – astronomy and cosmology deal with the position, motions and luminosities of the objects in the universe.  In fact, theoretical astrophysics was birthed by Newton who applied the rules of mechanics he used on the surface of earth to study how planets moved in space. Rocket Scientists fall under a broad category of aerospace engineers and their work goes beyond aircraft design Rocket Scientist: From engineering duties such as designing and constructing spacecrafts and missiles to testing equipment such as rocket propelled vehicles, a rocket scientist has a background in aerodynamics and propulsion. They oversee the manufacturing of aerospace vehicles and test communications and fuel system as well. Usually from an aerospace engineering background, rocket scientists plan and develop the design criteria, modify existing systems, and formulate models for computer analysis to better understand engineering requirements. They also research new materials for aerospace engineering designs. Planetary geologists collect samples from celestial bodies for analysis Planetary Geologists: Geologists come in the forefront in light of natural catastrophe such as earthquakes, mining, and tsunamis but there are planetary and lunar geologists as well. With space agencies across the globe undertaking ambitious manned mission to the moon (China National Space Administration plans a manned moon mission) and SpaceX wants to explore the red planet to examine the surfaces for signs of life, geological studies are a big part of the mission. However, identifying planetary geological features is an altogether different ballgame.  To address this gap, European Space agency launched a training program called Pangaea last year. Deriving its name from an ancient supercontinent, the training program is aimed for astronauts to identify best areas for exploration and find rocks for sample analysis. Geology also plays a critical role in identifying landing sites for missions. Some of the core areas where planetary geologist work are: Studying the topography of planets for evidence of life and oversee the movement of robotics missions Lunar Geologists are experts in lunar rocks and soils and oversee the collection of samples from missions Find clues of water, reservoirs and hospitable signs of life A background in physics, geology and astronomy is highly recommended. You can also check out some of the current job openings at ISRO for scientists, engineers and scientific assistants","excerpt":"Want to explore careers in space. If you are an ardent space fan and want to explore the outer universe, a STEM education provides a plethora of opportunities to kick-start this exciting job opportunity. In this article, Astronaut Today lists down some of the space science career opportunities which are not just limited to an […]","categories":["AI Highlights"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-02-16T04:29:21","publication_year":"2017","word_count":753,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","ai_applications:robotics","R"],"extracted_tech_keywords":["AI","Aim","R","Go","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/want-to-go-to-infinity-and-beyond-heres-how-to-make-a-career-in-space\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10076542,"title":"Inside World’s first AI-Enabled, Fully-Automated Robotic Kitchen","content":"Five years ago, in 2017, when Ajay Sunkara, a serial entrepreneur and owner of multiple breweries and manufacturing companies, visited one of his outlets, Khajuraho, located in Hyderabad, he noticed that their fully-automated facility was producing drinks consistent in taste and smoothness, irrespective of batch change. However, the same was not the case with food, which oscillated between a hit and a miss, too salty or overtly spicy, all depending on the chief’s temperament. There was no consistency. That is when he thought of an idea – ‘if beer can be made using machines and are perfectly consistent, why not food?’ A couple of months later, he returned to the US and shared his idea with friends. They instantly loved it. One of them – Vijay Kodali – even ended up becoming the CTO of the company. That was the start of NALA Robotics. Cut to present, the company has set up its first fully-automated restaurant in Chicago, alongside the launch of various products, including Nala Chef, Pizzaiola, The Wingman, and others. The company currently focuses on providing back of the restaurant operations. That includes kitchen operations, food preparations, delivery, etc. Interestingly, it has built everything from scratch and in-house, including the technology capabilities and fabrications. Last month, at the Global AI Summit 2022, Riyadh, the company signed a partnership agreement with Saudi Excellence Company to develop, market, and deliver autonomous food services to Saudi Arabia as part of its global expansion plans. At the event, Analytics India Magazine caught up with Sunkara, where he spoke about his company’s plans, and opportunities in India, alongwith challenges, solutions, cost estimations, RoI, and more. Interestingly, while we were both conversing, their robots were busy serving Saudi Coffee on the sideline. Robot makes biryani Sharing the details of their restaurant in Chicago, Illinois, Sunkara said their restaurant serves Indian cuisine. “When we started the project, we decided to take the most complicated dish – the biryani – for it has multiple ingredients and involves numerous processes,” he added, saying that their Nala Chef robot now makes awesome biryani. Since then, they have been developing add-ons to it. “Right now, we serve Indian tiffins, including idly, dose, pongal, vada, and others,” said Sunkara. How it works He said that the way the system works is, the user has to define the menu and code the recipe into the database. For instance, if you want the robot to make fried rice, you will have to specify the ingredients, saying 200 grams of rice, 10 grams of onions, 5 grams of salt, etc. Once all the details have been entered, the machine will make it, and make it right every time. NALA Robotics not only offers personalised solutions to restaurants and customers but also helps in minimising wastage of food, and eliminates cross-contamination, as it follows a thorough clearing process cycle, ensuring cleanliness and hygiene at all times. “For instance, if you are ordering biryani, and you feel it is too spicy – you can give feedback, saying that it is too spicy. So, next time you order the food again, it checks your data, makes necessary adjustments, tones down the spices a little bit, and gives you a more customised food,” explained Sunkara. This also gives customers more control over what they want to eat and how much they want to eat, based on their diet type, including vegetarian, vegan, keto, etc. Tech stack Further explaining the technology side of things, Sunkara said that the system works with multiple sensors – close to 1,200 – which check every microsecond, oversee how they perform or err. Similar to how human senses work, for eyes, they leverage computer vision, and for ears, they have systems integrated with Alexa. “We also have developed a Taste Bot that senses the spectrum of a beam of light, and dissects the food, and sees how much salt etc it has,” he added, saying that it gives a specification of it, and that is how the robot can make a decision, making your food taste better and consistent. Besides these, the team said that they use multiple simulation platforms. The company has also developed an in-house AI-enabled platform called ‘YummOS,’ where the entire operation of the robotics can be monitored and controlled. The company told AIM that it looks to open-source its platform in the future. Cost factors NALA Robotics chief said that the cost of setting up a fully-automated robotic restaurant varies on the number of dishes, tasks, and restaurant format. For instance, if you have a large menu that includes pizza, burgers, and biryani, the cost would go way up as you will need multiple robots working on each cuisine. But, if an individual robot does tasks like frying fries, it will be much more affordable. So, the cost varies from $120,000 and can go up to $1 million. Return on investment (RoI) This would depend on the number of hours you run the restaurant, like 24\/7 operations or eight-hour operations, ten-hour operations, etc. “There will be significant savings in terms of the staffing cost,” said Sunkara. He said that there would be a 10-15 per cent reduction of the staffing cost, again, it would depend on the restaurant format, menu, and how effectively they utilise the robots. Further, he said that if you are a 24\/7 restaurant, the savings multiply with other benefits like you do not have insurance claims, slips and falls, no sick leaves, etc. Eyes Indian market NALA Robotics is currently in an expansion mode. While Sunkara said the company sees an opportunity globally, India is very close to his heart. He said India has a unique problem. “With a population of 1.3 billion plus right now, a solution with robotics can answer hunger problems, and food is going to be a lot more affordable for many people,” he added. Touching upon the country’s plethora of cuisines and delicacies, he said that certain foods are not available everywhere due to a lack of talent, chefs, etc. “So, robotic solutions might be an answer to those areas as well,” said Sunkara. That explains why they opened an Indian restaurant in Chicago in the first place because solving for India can be an answer to global problems. Indian cuisines prepared by Nala Chef Talent crunch in AI & robotics Initially, NALA Robotics started with 20 members, today, they have about 170 employees and are actively hiring. “We expect to be 500 by the end of 2024,” he added. But, the question is, how will they achieve that goal, given the dearth of talent in the AI and robotics landscape? Sunkara seems to have sorted this issue. Besides running the show for NALA Robotics, he is also one of the members of Usha Rama College of Engineering and Technology, Vijayawada–run by his father, Sunkara Ramabrahamam, alongside his brother Anil Sunkara and others. “My dad runs an engineering college in India. So, we find talent from the college,” said Ajay Sunkara. He said that the college recently introduced robotics and AI courses. “Right now, we have a steady stream of talent coming from college,” said Sunkara. Other challenges During the initial days, NALA Robotics’ biggest challenge was the adaptation itself. “The one who runs the PoS was not used to working with the robot. So that was just the initial challenge. It is now pretty streamlined, and we are now running buffets in the Chicago location – that is pretty good.” Sunkara also said that most of the challenges the company faces today are from local government authorities. For instance, there is a whole gamut of regulations in the US, including UL-NSF certification, local county regulations, health department regulations, etc. “These are some challenges because we are the first to enter into the robotic food space,” said Sunkara. Further, he said educating everyone has become a bit of a challenge, but we have crossed all of them, and right now, we have a stable platform running for the past nine months. Growth way forward In a short span of five years, the company has come a long way. The pandemic accelerated their growth, making their products more relevant than ever. “At our restaurant in Chicago, the robots are behind the wall; we don’t show them to the public. But, with the pandemic, things changed and now we regret it,” said Sunkara. A few months ago, NALA Robotics ran a survey to understand customers’ mindset of how comfortable they are with robots making their food. The results were not surprising; nearly 68% said they are happy with robots cooking their food. “The dynamics are changing. Right now, both Gen Z and millennials are supporting technology,” said Sunkara. Further, he said that the US has a huge staffing issue. “We have been getting calls from all over the world, not just in the US, but in Europe, Indonesia, etc. Everyone is struggling with staffing issues,” he added, saying that they are in the right place at the right time. “Currently, we are working with large restaurant chains, developing solutions for them, from pizza to sandwiches, burgers, etc.,” said Sunkara. He also said they are working with companies on dishwashing as most restaurants are struggling with manpower post-pandemic. “Our sales pipeline is close to a few hundred million right now,” shared Sunkara. Taste the future Sharing the roadmap ahead of the company, the NALA chief said that in the coming months, they are going to focus their efforts on two verticals – mainly installations of fast food restaurants, QSR fine dining restaurants, etc., and setting up of ‘robotic ghost kitchens’ across the country, in the US. “A large restaurant chain would be able to afford robots, but home chefs or someone just starting out will not be able to afford them so we have established robotic kitchens all across the country, where anyone can go online, create menu, recipe, put a name, and signup on GrubHub or UberEats,” shared Sunkara, saying that when the order comes in, the robots are going to fulfil the order, make the dish and the food aggregators would deliver the order to customers doorsteps. In other words, you do not have to open a physical restaurant, you can virtually open a robotic restaurant, and in a matter of seconds, your restaurant can be available in all 100 to 1000s locations. “Sitting at home, you will be able to start your restaurant and start serving food worldwide,” he added. From a monetisation perspective, NALA Robotics said that they would take a cut of it, where they will be charging the customers. Currently, the ‘restaurant-as-a-service’ marketplace by NALA Robotics is available in the US and soon in Saudi Arabia. “We will be expanding into India as well,” he shared. He said they have an office in Hyderabad, India. NALA Robotics’ research and development centres are currently in the US and Ukraine. “We are setting up our R&D centres in Riyadh and India as well. We are looking to hire more people in the coming months,” he added. Ghost Kitchen installations by NALA Robotics Will robots take over jobs? One of the most common questions that come to mind after reading this is – will the robots take over human jobs? A lot of students and people from lower-income socio-economic groups turn to restaurants for jobs to earn income, make ends meet, or fulfil their ambitions. But, with the intervention of robots, people fear that their jobs are at stake. “The answer to that is ‘NO’,” said Sunkara, citing the evolution of computers on how it helped humanity leapfrog scientific advancements and technology innovations, creating more high-paying jobs. He expects robotics will also go through a similar transition. He said, “Robots have already been in the manufacturing industry – did they take our jobs? Maybe a little bit. But, more than that, they improved the safety of the manufacturing industry,” he added. Sunkara said that the quality of life in every aspect of it has improved with robotics, similar to how computer operations have been done. He believes that in the future, the adoption of robotics will happen, and there will be tasks for humans,” he added, “In fact, we are starting to explore other planets right now. So there will be a lot of needs for humans–they will not be out of jobs, for sure.”","excerpt":"“Sitting at home, you will be able to start your restaurant and start serving food worldwide,” says Ajay Sunkara, co-founder and CEO at NALA Robotics","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Amit Naik","publish_date":"2022-10-07T12:00:00","publication_year":"2022","word_count":2055,"keywords":["Go","AI","ML","computer vision","RAG","GRU","Aim","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","computer vision","analytics","Aim","RAG","R","Go","GAN","GRU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/inside-worlds-first-ai-enabled-fully-automated-robotic-kitchen\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":63794,"title":"How AI Can Help Manage Prisons Amid COVID-19 Pandemic","content":"With inmates huddled together in closely-knitted cells inside prisons, it is little wonder then that these facilities are vulnerable to COVID-19 outbreaks. Coupled with the apathetic approach towards the health and well-being of those incarcerated, managing a pandemic of this scale has been more challenging, particularly in containment zones. Some measures that are being considered to stymie the spread of this virus in correctional facilities involve steps to bring down the population behind bars. This may include modification in the norms to shift some prisoners to home confinement, release first-time offenders or those caught for non-violent offences, or even temporarily release those who have served a majority of their sentence already. Furthermore, many inmates have underlying health conditions, making them more susceptible to severe COVID-19 symptoms. Given this, readily-available access to doctors to treat infected inmates, and other medical care provisions like masks and soaps could also help reduce the extent of the outbreak within prison systems. This could also ease pressure on under-staffed prisons where existing staff are either self-isolating or are on sick leave. Anticipating that the coronavirus will enter closed spaces like prisons despite precautions, authorities should leverage new technologies to reduce the impact of the virus. Here are some ways in which prisons can integrate AI and other technologies to respond to the pandemic appropriately: AI-Based Video Surveillance Advanced facial-recognition technologies and AI-powered video surveillance can help prison managers conduct crowd analysis and ensure that social distancing protocols are being followed. A startup based out of Gurugram, Staqu Technologies, had introduced an AI-based video analytics platform in several prisons in Uttar Pradesh late last year. Monikered JARVIS, it uses AI to scan CCTV footage and alert authorities, as well as provide actionable insights. This video analytics engine will enable a quicker response in emergencies like this outbreak. Although this product was launched to keep a check on unlawful practices inside prison cells with the help of ‘intelligent monitoring of objects, crowd and perimeters’, it is equally relevant amid this pandemic, if not more. Although some prisons may not be well-equipped to handle such high-tech resources to adequately enforce social distancing, measures like these should be made a priority. AI-Powered Screening Systems The same startup also launched a new Thermal Camera under its JARVIS portfolio to supplement screening efforts, especially in prisons. Staqu’s new technology examines heat signatures through the cameras and sends an alert when an inmate with a body temperature of over 37°C is detected up to a range of 100 metres. Efficient technological interventions like these should be made mainstream since the health of inmates is often overlooked in prisons. If not appropriately screened to flag inmates who could potentially be carriers of the virus and further infect others, it could be catastrophic. Additionally, automated screening should also be complemented with a database carrying relevant information. Since people regularly enter and exit the premises — be it visitors or doctors — critical information around their travel and medical history must be logged and maintained on a regular basis. Management of COVID-19 will likely result in school and library closures and consideration should be underway concerning the number of people entering jails and prisons and how each step can be re-evaluated and monitored. ML & RPA To Help Prison Staff Some hospital chains have been leveraging AI and related technologies to better manage staff and handle other administrative functions to ease pressure on the overall system. With rising responsibilities shared between officers in an already understaffed prison, scheduling has become more complex. Some companies can provide AI-powered solutions to accurately plan a rota around existing and available staff members. Based on vacancies, it can align them with appropriate shifts based on their experience and skill sets. For instance, Norway-based Globus.ai uses deep learning (DL) and machine learning (ML) techniques to match healthcare workers to specific tasks, thereby helping fill available slots. The same technology, albeit some modifications, can potentially be employed in prisons to make the planning of rota more efficient. Another example is Blue Prism, which has been using robotic process automation (RPA) to tackle COVID-19, including in prisons. The National Health Service (NHS) in the UK has been using its services to automate a central dashboard it maintains to keep track of COVID-19 positive cases in prisons. This, in turn, helps staff manage inmates who are affected by the virus. Automated Disinfection Systems Deploying robots to clean and disinfect rooms and corridors in big establishments to promote no-contact measures has been on the rise. A company based out of Denmark UVD Robots have been improving cleaning routines in hospitals by automating the process with the help of its bots. Currently used in hospitals to protect medical personnel from getting infected, its solution can be applied within prison systems as well. The bots emit ultraviolet light over seemingly unclean surfaces to terminate viruses by breaking down their DNA-structure.","excerpt":"With inmates huddled together in closely-knitted cells inside prisons, it is little wonder then that these facilities are vulnerable to COVID-19 outbreaks. Coupled with the apathetic approach towards the health and well-being of those incarcerated, managing a pandemic of this scale has been more challenging, particularly in containment zones. Some measures that are being considered […]","categories":["Deep Tech"],"tags":["Coronavirus","covid-19"],"author_name":"Anu Thomas","publish_date":"2020-05-03T13:00:00","publication_year":"2020","word_count":814,"keywords":["machine learning","covid-19","AI","RPA","ML","RAG","Coronavirus","automation","deep learning","analytics","R","analytics platform"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","analytics","RAG","R","automation","RPA","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/how-ai-can-help-manage-prisons-amid-covid-19-pandemic\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":44068,"title":"How An April Fools Prank By Google Was Turned Into Reality By China","content":"Self-driving vehicles are perhaps one of the most intriguing advancements today. And over the years, what was just a concept has become a reality ­— whether it’s Tesla or Google Waymo. But today, it’s not just cars, who are equipped with autopilot, bicycles have entered the space as well. On 31 March 2016, Google Netherlands YouTube channel shared a video of a self-driving bicycle which enables cyclists to freely do other things while they ride through the city, which was also thought to be an April Fools’ Day prank. However, Google later revealed that it wasn’t exactly a prank as the search giant posted a behind-the-scenes video. China’s Makes A Move To Towards AGI It’s been over three years that Google released the self-driving bike gag, and surprisingly, the concept has become a reality today. A video has gone viral on the internet that shows a bike walking his man. Many were thinking it to be a prank again; however, that’s not true — a real-world demo of how AI can work with specialized hardware and overcome the challenges of machine learning algorithms. This is not the first self-driving bike, talking about other instances if self-driving bicycle, in 2018, a team of students from Tsinghua University in Beijing, China, have created a self-riding bicycle which could auto-balance. https:\/\/youtu.be\/nkDj2Q0buyg The prime motive behind this project was to keep the cycle balance. And for that, the students chose the most similar method to a man-made use of centrifugal force produced by controlling the handlebar to keep its balance. But among all the other autonomous bikes, this recent innovation is gaining more attention and the very reason is the fact that it is AI-enabled and it’s not just about self-balancing. The bike is not only capable of riding by itself, self-balance, avoid obstacles, take turns, but it can also understand human voice instructions and make independent decisions. But what set this innovation apart is the brains behind the bike — the fact that it is powered by a new kind of hybrid computer chip, called Tianjic. A team made up of members from a host of institutions in China, one in Singapore and one in the US, has built this advanced chip. According to a paper, there are two approaches to developing artificial general intelligence (AGI) — its either through machine learning or by brain-inspired computing. In most of the cases, the two approaches are kind of difficult to merge and make things work because of communication difficulties between the two systems. However, the experts have found ways to make them work and this hybrid chip is a combination of both — it can accommodate computer-science-based machine-learning algorithms and brain-inspired circuits and several coding schemes. Furthermore, the chip is designed and built in such a way that it can process algorithms and models in the unmanned bicycle system simultaneously. Talking about the autonomous bicycle, there was a time when the whole world was talking about driverless cars, in 2014, Chinese multinational technology company Baidu was already working on an autonomous bike with the same concept. The idea was to develop a bike that could not only ride around avoiding obstacles and navigating complicated road conditions, but also a bike that could identify its owner in some way and the project was being carried out in Baidu deep learning laboratory (Institute of Deep Learning, Idl). To know more about the nuts and bolts of this top-notch hybrid computing chip, you can refer to our article, “What’s The Whole Commotion Regarding China’s Tianjic Chip About?” Outlook This entire event also shows how far China has gone when it comes to artificial intelligence. There was a time when the nation was struggling to build its own chip industry, but today, the scenario has completely changed. China has pushed itself to accelerate not only the chip industry but also the AI space with serious innovation and massive investments. Over the years, it has made substantial strides in the field, and today, the Chinese AI startup ecosystem is well known for its cutting-edge technology solutions, making the nation the AI superpower of the world. Furthermore, the nation is aggressively working towards becoming a $150 billion AI global leader by 2030. Looking into the future, the researchers of the Tianjic powered autonomous bicycle has prophesied that this innovation has the capability of providing a route toward more general forms of AI. And looking at this recent project, the job Tianjic has done, it seems the words might soon turn into reality.","excerpt":"Self-driving vehicles are perhaps one of the most intriguing advancements today. And over the years, what was just a concept has become a reality ­— whether it’s Tesla or Google Waymo. But today, it’s not just cars, who are equipped with autopilot, bicycles have entered the space as well. On 31 March 2016, Google Netherlands […]","categories":["Global Tech"],"tags":["autonomous car","china ai investments"],"author_name":"Harshajit Sarmah","publish_date":"2019-08-08T17:00:47","publication_year":"2019","word_count":752,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","innovation","programming_languages:Go","china ai investments","deep learning","autonomous car","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","R","Go","innovation","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/how-an-april-fools-prank-by-google-was-turned-into-reality-by-china\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":46719,"title":"Machine Learning Vs Machine Teaching: A Whole New Approach To Imparting Human Skills","content":"There is no doubt that machine learning is one of the major driving forces behind most of the advanced techs and gadgets we have today. Whether it is your smart home device or that newly bought self-driving car yours, ML is playing a vital role not only advancing gadgets but is also changing the way people interact with machines. No doubt, it is one of the hottest technologies in the world. However, people almost forget that there is something called Machine Teaching that also plays a significant role in all the ML use cases. The Teacher’s Perspective We all have heard a lot about ML, which is also seen as a subset of artificial intelligence. And if we try to understand it by its literal definition, it is the ability of algorithms and statistical models of machines or systems to automatically learn and improve from experience without being programmed or without using any explicit instructions. When we talk about this sought after technology, artificial intelligence, people tend to think about things like performance modelling, learning algorithm optimisation, network architecture, all the tasks it can perform. However, this is just through the perspective of a learner. But what it would be when we look at the entire scenarios through a teacher’s perspective? Would it still be the same things? Microsoft has the answer. While the rest of the world has been always focused on the learner side, tech giant Microsoft has taken the teacher’s perspective and explored the possibilities where a machine could be able to do all the tasks that a human can do — whether it’s about thinking, finding solutions, etc. However, there was a bottle that stands as a hurdle for machine learning to accomplish these goals. But when you go about taking a deeper look for a solution to it, you realise the bottle is not the technology but the teacher — the data. And this where Machine Teaching comes into the scenario. What Exactly Is Machine Teaching? Machine teaching is an approach where human expertise and abilities to find solutions to problems are used in order to help machine learning models find important hints about how to find a solution faster. To be more technical about it, Machine teaching designs the optimal training data to drive the learning algorithm to a target model. Unlike the traditional approach of teaching machine which about feeding the machine or the system with lots of available data, machine teaching takes a different route to go about it — it not only feeds the data but also let know where to look. For example, when a teacher teaches a student how to recognise a bicycle, the teacher tells the students about all the features and characteristics of the bicycle and later tests him\/her with other things. And if the students say the Motorbike is the bicycle, the teacher doesn’t say it is wrong; rather, the teachers correct the student by telling him\/her about the differences. And this is the same loop that machine teaching uses. Rather than extracting data and digging insights from it to train a model, people’s expertise could also be a significant driving force in teaching machines. Outlook Technology is on an evolving spree and year by year things are just getting better and bigger. The same has happened with machine learning. According to Microsoft, who has been researching and working on Machine Teaching for a decade now, even AI struggles and sometimes fail to learn things by itself, but when people start to guide the machine to do and learn things that we already know, things are going to significantly different. The machine is not only a whole new approach to machine learning but it’s an approach to empower people to make sophisticated use of AI. It doesn’t matter whether you are a developer or an SME with limited knowledge, machine learning makes things easier — one can impart abstract concepts to an intelligent system, and it would perform the machine learning mechanics in the background.","excerpt":"There is no doubt that machine learning is one of the major driving forces behind most of the advanced techs and gadgets we have today. Whether it is your smart home device or that newly bought self-driving car yours, ML is playing a vital role not only advancing gadgets but is also changing the way […]","categories":["Deep Tech"],"tags":["Data Science","Machine Learning"],"author_name":"Harshajit Sarmah","publish_date":"2019-10-02T12:00:19","publication_year":"2019","word_count":669,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","ML","Machine Learning","programming_languages:Go","ViT","Data Science","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/machine-learning-vs-machine-teaching-a-whole-new-approach-to-imparting-human-skills\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10096383,"title":"How Generative AI is Reshaping Observability Solutions","content":"Observability solutions are becoming increasingly important in today’s digital age. As more and more companies rely on technology to run their business operations, the need to monitor, analyse, and optimise their systems and applications in real-time has become critical. Organisations are now prioritising customer experience and accessibility through digital transformation. However, these initiatives can introduce complexity and challenges for engineers. “Efficient observability solutions are crucial for businesses scaling their digital outreach, as they face increasing device diversity. “Engineers require end-to-end observability to untangle data, understand business performance, monitor service availability, and optimise further. New Relic serves as a comprehensive observability solution, offering real-time insights and empowering engineers to make data-driven decisions throughout the software lifecycle,” Ganesh Narasimhadevara, Principal Technologist, APJ, New Relic told AIM. New Relic’s comprehensive platform offers over 30 capabilities, delivering a seamless and connected experience throughout the various layers of the technology stack at every stage of the software lifecycle. “We started our journey with the very first capability of application performance monitoring and then we extended ourselves along with the tech advancements that we’re seeing in the market. We extended into infrastructure monitoring, not limited to the VMs or just to the on-prem servers. As of today, we have the most advanced Kubernetes monitoring in the industry.” New Relic Grok But today, we are in the age of generative AI and it is finding use cases in the observability solutions space as well. New Relic, a US-based Observability platform announced New Relic Grok, the world’s first generative AI assistant for observability. Narasimhadevara, in an exclusive conversation with AIM, said that New Relic’s generative AI powered observability is what engineers would need to be more productive every day. “New Relic Grok empowers users to identify instrumentation gaps and provides guidance for onboarding new services. It can sift through extensive documentation to answer questions and generate queries for specific metrics or application trends. Additionally, it facilitates common use cases such as creating alerts or service levels.” New Relic is leveraging generative AI capabilities through Microsoft OpenAI Azure services (GPT models) and is not exploring building its own proprietary models. However, it is something that New Relic could look into in the future. “Currently, I don’t see the immediate necessity for building our own language models. Many organisations, including those in India, are partnering with hyperscalers to leverage their existing efficient models. The focus is on understanding how generative AI can enhance observability before investing resources in model development. Building our own model may be a future plan once we have assessed its benefits and potential.” How gen AI can benefit observability solutions Today, companies in the observability space are developing new generative AI solutions that can help automate monitoring, visualisation, and analysis of complex systems, reducing manual effort, speeding up issue identification and resolution, and improving overall system performance. In short, generative AI is making it easier for users to find the root cause of an issue and fix the errors. “Observability solutions have evolved to be a single source of truth. Having a generative AI system as part of the product is going to eliminate the need of jumping between a tool like ChatGPT, and observability solutions,” Narasimhadevara said. The advantages that generative AI brings to observability solutions will help in reducing the workload of engineers and help save a lot of time allowing them to focus on more strategic work. “Another way generative AI solutions with observability can evolve is by bypassing the tribal knowledge, according to Narasimhadevara. Tribal knowledge refers to experience and knowledge that is shared among team members but is not documented or shared more widely across the organisation. “Team members may vary in their expertise and product knowledge. Imagine having the capability to analyse vast amounts of data, process tribal knowledge, detect anomalies, and connect various metrics to provide insights such as identifying the cause of an anomaly due to a recent deployment. Generative AI solutions integrated effectively can offer such capabilities, providing relevant suggestions and allowing users to train the model to meet their specific needs.” Moreover, another good advancement in observability that could emerge from generative AI is automation. Going forward, we could foresee a scenario where a Large Language Model would be able to perform a task without any human intervention. “In future engineers will have the ability to instruct the generative AI assistant to monitor the application performance, provide some recommendations and also implement them through automation.” What impact could GenAI have on observability solution providers? Narasimhadevara believes there are similarities between observability solutions and large language models. “The reason why it’s very similar is that they get as good as the data that you feed them.” Hence, observability solution providers only stand to benefit from generative AI. The effectiveness of language models is contingent on the data they have access to. Observability solutions, on the other hand, provide critical insights into system states that are exclusive to their functionality. “Integrating observability solutions with advanced or current language models can enhance contextual understanding. Without the relevant data from observability solutions, the language model’s power is limited.”","excerpt":"New Relic introduced New Relic Grok, the world’s first generative AI assistant for observability","categories":["AI Highlights"],"tags":["new relic"],"author_name":"Pritam Bordoloi","publish_date":"2023-07-05T18:00:00","publication_year":"2023","word_count":848,"keywords":["ChatGPT","GenAI","OpenAI","AI","ML","RAG","Aim","generative AI","new relic","Azure","kubernetes"],"extracted_tech_keywords":["AI","ML","generative AI","GenAI","ChatGPT","OpenAI","Aim","RAG","Azure","kubernetes"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-generative-ai-is-reshaping-observability-solutions\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10013916,"title":"Guide To Google’s AudioSet Datasets With Implementation in PyTorch","content":"AudioSet Dataset is developed by the Google Sound and Video Understanding team. The core member of the AudioSet Dataset is Jort Florent Gemmeke, Daniel P.W.Ellis, Dylan, Aren, Manoj Plakal, Marwin Ritter, Shawn Hershey, and two more members of the team. There are other twelve contributors to AudioSet DataSet who help to build a pipeline for the data storage in the form Youtube_url Id, start_time, end_time, and other classes. AudioSet Dataset has more than 600 classes of annotated sound, 6000 hours of audio, and 2,084,320 million YouTube videos annotated videos and containing 527 labels. Each video has a 10 sec sounds clip extracted from Youtube Videos in different classes for the training and testing dataset. AudioSet Ontology is the collection of sound in hierarchical and organized. It covers a wide range of sounds, from the human voice to pets animals to natural sounds. Below the hierarchical Table of ontology to select which type of Dataset you require to develop your model or for research purposes. Research Paper: https:\/\/research.google\/pubs\/pub45857\/ Github: https:\/\/github.com\/audioset\/ontology Download:https:\/\/research.google.com\/audioset\/download.html About DataSet: It is available in two formats to use for a research purpose: Dataset is available as a CSV file.YouTube video  url_id, start time, end time, and other more labels.Other formats are recorded as VGG-like models as TensorFlow Record files. Visit here:https:\/\/github.com\/tensorflow\/models\/tree\/master\/research\/audioset Dataset is split into three disjoint sets each is present in CSV Format: Evaluation Balanced TrainUnbalanced Train How we can use ontology to get the right dataset and help us to find the right data for the training deep learning model. We have to build the Horn Detection of Train. Just visit the ontology page of AudioSet Dataset and select Sounds of things to the in the given link. https:\/\/research.google.com\/audioset\/\/\/\/\/ontology\/index.html . As shown in the image. When you select the vehicle section, then you get a different type of transportation system such as Waterway, Airways, Roadways, and Railways transportation. We have to select Rail transport as shown in the image below. After selecting the Rail Transportation, you will get different types of train such as train wagon, Railroad car, subway, metro, and Train. We have to choose the Train selection as shown in the image. After choosing the train selection, we got two types of trains sound the first one is the train whistle, and another one is the train horn. Choose Train horn as shown in the image below. After selecting the train horn, it comes to our required dataset, which is best suitable for us to build horn detection. Choose this and go furthermore one step to know about the Dataset. Select, as shown in the figure below. After selecting the train horn now, you will get the overall details of the Dataset and their number of videos available, and the total number of each part of the dataset. Hour duration is sufficient to train the model. We learned how to select the dataset using AudioSet Ontology if you have any query reach out to audioset-users@googlegroups.com. Using Pytorch: import pickle import random import torchaudio import numpy as np import pandas as pd from torch.utils.data import DataLoader from torch.utils.data.dataset import Dataset Data = pd.read_csv(“filepath”) ## Downloaded dataset locally. For implementation in PyTorch: Github: https:\/\/github.com\/qiuqiangkong\/audioset_classification Application: 1.Building Horn detection: Horn detection is built using a neural network for the detection of train type. It helps people to identify the train motion and their speed on a railway crossing. 2.Bird Sound Detection. Detect the bird’s sound and check their emotion and feeling based on it. Birds are singing or calling other birds angry or something. Also, use their song feelings. Conclusion: We have learned about the AudioSet dataset, how we can download it from the source. In different file formats,  AudioSet dataset creator and their researcher.AudioSet Ontology uses the case to choose the right dataset and  Implementation of model in PyTorch.Much other real application is used in daily life using AudioSet  Datasets. To participate in the competition: https:\/\/www.kaggle.com\/c\/birdsong-recognition","excerpt":"AudioSet Dataset is developed by the Google Sound and Video Understanding team. The core member of the AudioSet Dataset is Jort Florent Gemmeke.","categories":["Deep Tech"],"tags":["big data storage format","Datasets","Deep Learning","extract big data","Mapreduce","numpy","purposes of a data team","Pytorch","training","VR Data Analytics"],"author_name":"Amit Singh","publish_date":"2020-12-13T10:00:00","publication_year":"2020","word_count":650,"keywords":["big data storage format","deep learning","purposes of a data team","extract big data","R","Pandas","Pytorch","Datasets","NumPy","training","PyTorch","numpy","Mapreduce","RAG","AI","neural network","ML","Deep Learning","TensorFlow","VR Data Analytics"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","TensorFlow","PyTorch","Pandas","NumPy","RAG","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/audioset-dataset\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10058366,"title":"What is multimodal system of AI and its evolution","content":"Multimodal AI, especially the sub-field of visual question answering (VQA), has made a lot of progress in recent years. Multimodal systems, with access to both sensory and linguistic modes of intelligence, process information the way humans do. What is multimodal interaction? As human beings, we experience the world as multimodal: we can feel texture, hear sounds, see objects, smell odours and taste flavours. However, standard AI systems are usually unimodal, meaning they are trained to do a specific task such as processing images or languages. The systems are fed a single sample of training data, from which they are able to identify corresponding images or words. While it is easier to work with a single source of information, it also means that the software lacks the context and supporting information to make the best possible deductions. The advancement of AI relies on its ability to process multimodal signals simultaneously, just like humans. For instance, while we can understand how the meaning of texts and images change when they are paired or juxtaposed with each other, a unimodal AI system would be unable to infer how the meaning of an image could change when placed next to a contradictory piece of text. Multimodal learning, on the other hand, pieces together disjointed data (collected from different sensors and data inputs) into a single model. Since multiple sensors are used to observe the same data, multimodal learning offers more dynamic predictions compared to a unimodal system–processing more datasets translates to more intelligent insights. Milestones Researchers at the Allen Institute of Artificial Intelligence, AI2, created an AI model capable of producing images from text captions. At the University of North Carolina, Chapel Hill, researchers improved on the reading comprehension of existing language models by incorporating images. Google AI created MURAL (“Multimodal, Multitask Representations Across Languages”) in order to combat the lack of direct translation between different languages across the world. To make translations more accurate, the software used multitask learning to match images to text—covering over 100 languages. Meta developed AV-Hubert— a speech recognition system with the ability to filter out background noise to better decipher the speaker’s voice. The model is  75% more accurate than existing models trained on an equal number of transcriptions. Open AI introduced DALL.E, a 12-billion parameter version of GPT-3 to generate images in response to text descriptions, utilising a dataset of text-image pairs. The model can anthropomorphize animals and objects, connect unrelated concepts in plausible ways, apply alterations to existing images, and make visual representations of text. Applications Sophisticated multimodal systems have multiple applications across industries including aiding advanced robotic assistants, empowering advanced driver assistance and driver monitoring systems, and extracting business insights through context driven data mining. Over time, advancements in multimodal learning could also help overcome some of AI’s present challenges. For instance, Meta developed the Hateful Memes Challenge to combat harmful multimodal content. Many of the problems surrounding AI’s inability to understand context, could be combated with the development of strong multimodal algorithms.","excerpt":"Multimodal AI, especially the sub-field of visual question answering (VQA), has made a lot of progress in recent years. Multimodal systems, with access to both sensory and linguistic modes of intelligence, process information the way humans do. What is multimodal interaction?  As human beings, we experience the world as multimodal: we can feel texture, hear […]","categories":["Deep Tech"],"tags":[],"author_name":"Srishti Mukherjee","publish_date":"2022-01-14T14:00:00","publication_year":"2022","word_count":498,"keywords":["Go","artificial intelligence","AI","Modal","BERT","GPT","llm_models:BERT","multimodal AI","R","llm_models:GPT"],"extracted_tech_keywords":["AI","artificial intelligence","multimodal AI","R","Go","BERT","GPT","Modal","llm_models:GPT","llm_models:BERT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/what-is-the-multimodal-system-of-ai-and-its-evolution\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10079202,"title":"NVIDIA’s Text-to-Image Model eDiffi Completes the Picture","content":"AI text-to-image generators have become as commonplace as having an ‘opinion’ —if everyone has an opinion, every tech company worth its salt has its own AI text-to-image generator. All the big tech companies have one—Microsoft-backed OpenAI has ‘DALL.E 2’, Google has ‘Imagen’ and Meta has ‘Make-a-Scene’, while buzzy startups like Emad Mostaque’s Stability.ai have ‘Stable Diffusion’. Now, US semiconductor design giant NVIDIA has also entered the mix with its text-to-image model called ‘ensemble diffusion for images’ or ‘eDiffi’. However, eDiffi isn’t open to the public for use unlike Stable Diffusion and DALL.E 2 which are open source. Some old, some new Diffusion models synthesise images through an iterative denoising process that slowly generates an image from random noise. Traditionally, diffusion models have a single model which is trained to denoise the entire noise distribution. What eDiffi does differently is that it trains a group of expert denoisers at different intervals of time during the whole process. NVIDIA released a research paper, along with the announcement, titled, ‘eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoisers’, which claimed that this simplified the sampling process. Denoising involves solving a reverse differential equation during which a denoising network is called several times. NVIDIA wanted the model to be easily scalable, which is harder when each denoising step adversely impacts the test-time and computational complexity of sampling. The study found that eDiffi’s model was able to achieve the scaling goal without eating into the test-time computational complexity. Which model is the best? The paper concluded that eDiffi had managed to outperform competitors like DALL.E 2, Make-a-Scene, GLIDE and Stable Diffusion on the basis of the Frechet Inception Distance, or FID—a metric to evaluate the quality of AI generated images. eDiffi achieved a FID score slightly higher than Google’s Imagen and Parti. However, while each upcoming model seems to better the previous one in terms of accuracy and quality, it must be noted that researchers cherry pick the examples to showcase their best illustrations. The model’s best configuration was then compared with DALL.E 2 and Stable Diffusion, both of which are publicly available text-to-image generative models. The experiment found that the other models were mixing up attributes from different entities while ignoring some of the attributes. Meanwhile, eDiffi was able to correctly model attributes from all entities. (The first image is generated by Stable Diffusion, the second image is by DALL.E 2 and third image is by eDiffi) When it came to generating text which has been a sticky spot for most text-to-image generators, both Stable Diffusion and DALL.E 2 tended to misspell or even ignore words while eDiffi was able to generate the text accurately. In the context of long descriptions, eDiffi was also shown to be able to handle long-range dependencies much better than DALL.E 2 and Stable Diffusion, which indicates that it has a longer memory than the other two. New features added NVIDIA’s eDiffi uses a bunch of pretrained text encoders to give inputs to its text-to-image model. It uses a combination of the CLIP text encoder—which aligns the embedded text to the matching embedded image—along with the T5 text encoder—which performs language modelling. While older models like DALL.E 2 use only CLIP or Imagen uses the T5, eDiffi uses both encoders in the same model. This enables eDiffi to produce entirely different images even with the same text input. CLIP helps lend a stylised look to the generated images but the output normally misses out on the details in the text. On the other hand, images produced by T5 text embeddings produce better individual objects instead of a style. By using them together, eDiffi was essentially able to produce images with both qualities. The model was also tested on the usual datasets, like MS-COCO, which demonstrated that CLIP+T5 embeddings lead to much better trade-off curves than either used individually. On the visual genome dataset, it was proven that using T5 individual embeddings performed better than CLIP embeddings. The study finds that the more descriptive the text prompt is, the better T5 performs than CLIP. However, overall, a blend of the two worked best. This allows eDiffi to have what it calls ‘style transfer’. In this process, a reference image can be used for style from which CLIP image embeddings are extracted and used as a style reference vector. Then, style conditioning is enabled in the second step, following which the model generates an image similar to the input style and caption. In the third step, style conditioning is disabled, following which images are generated in a natural style. The study also generated images produced solely using CLIP text embeddings and T5 text embeddings separately. Images generated by the former often contained correct objects in the foreground with blurry, fine-grain details while images generated by the latter showed incorrect objects at times. eDiffi also introduced a feature called ‘Paint with Words’ which helps users determine the location of the objects in the image by mentioning it in the text prompt as well as scribbling on the image itself. Users can select the phrase to specify the location within the prompt. The model is then able to produce an image that matches both the input map or sketch and the caption.","excerpt":"What eDiffi does differently is it trains a group of expert denoisers at different intervals of time during the whole process.","categories":["Global Tech"],"tags":["AI Tool"],"author_name":"Poulomi Chatterjee","publish_date":"2022-11-09T18:00:00","publication_year":"2022","word_count":870,"keywords":["Go","TPU","OpenAI","AI","Scala","Aim","stable diffusion","CLIP","AI Tool","R","T5"],"extracted_tech_keywords":["AI","OpenAI","Aim","TPU","R","Go","Scala","T5","CLIP","stable diffusion"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidias-text-to-image-model-ediffi-completes-the-picture\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10095416,"title":"How Indian IT Giants are Bringing GenAI to Their Clients","content":"When OpenAI CEO Sam Altman said it was “hopeless” for India to build a general purpose LLM like the GPT models, Tech Mahindra CEO CP Gurnani tweeted: “Challenge Accepted.” Even though Altman’s comment was taken out of context and blown out of proportion, we will have to wait and see whether Tech Mahindra is building a ChatGPT alternative. Meanwhile, top IT companies in India such as Infosys, TCS, Wipro, HCL as well as Tech Mahindra have already announced generative AI capabilities to benefit their customers. Tech Mahindra’s Generative AI studio Tech Mahindra had been working on generative AI long before OpenAI released ChatGPT. The IT giant developed the Storicool platform, an auto content creation tool, which was beyond its years, according to chief executive CP Gurnani. In April this year, Tech Mahindra launched its Generative AI Studio under its amplifAI0->∞ suite of AI offerings and solutions. TechM amplifAI0->∞ helps the IT giant democratise and scale the deployment of AI technology for enterprises globally. Now, the studio aims to bring six aspects of content generation under one umbrella which include: Code, Document\/Text, Image, Video, Audio and Data, the IT giant said. Infosys leverages open-source large language models Interestingly Infosys, along with a few other parties including Elon Musk, AWS and others donated USD 1 billion to OpenAI, the company leading the generative AI revolution. The donation was made when Vishal Sikka was still at the helm of the IT giant. Last month,  Infosys announced Topaz — an AI-first set of services, solutions and platforms using generative AI technologies. Topaz leverages generative AI technologies and comes with 12,000 use cases, Infosys said. Its aim is to provide readily available industry-specific solutions that enable the integration of generative AI into functions such as intelligent automation, AI-driven customer service, and enhanced security solutions. “Topaz has helped businesses accelerate growth and create value through its AI-first approach by helping them adopt open-source LLMs to build narrow transformers which will solve a particular problem in the enterprise,” Balakrishna D R, EVP and head of AI and automation at Infosys, told VentureBeat. TCS to leverage Google’s foundational AI models Tata Consultancy Services (TCS) also announced generative AI capabilities in partnership with Google Cloud. The IT giant’s generative AI offering is powered by Google Cloud’s generative AI products and services such as Vertex AI, generative AI Application Builder and Model Garden, and TCS’ own solutions. This means TCS will leverage Google’s LLM PaLM for text and chat, Imagen for text-to-image, Codey for code completion and chirp for speech-to-text. Kevin Ichhpurani, CVP, global partner ecosystems and channels, Google Cloud, said that the partnership with TCS will help address industry-specific challenges and opportunities with generative AI capabilities and solutions, with a focus on addressing real-world use cases and adding business value. “Building on its deep domain knowledge across multiple industry verticals and investments in research and innovation, TCS has developed a large portfolio of AI-powered solutions and intellectual property in the areas of AIOps, Algo Retail™, smart manufacturing, digital twins and robotics,” TCS said. According to the company, they are actively engaging with clients in multiple industries to explore how generative AI can be leveraged to deliver value in their specific business contexts. Wipro leverages Google’s PaLM Like TCS, Wipro has also partnered with Google to leverage its generative AI tools through Google Cloud. According to the company, Wipro will integrate them with its own AI models, business accelerators, and pre-built industry solutions. The company plans to integrate generative AI as a central offering across its comprehensive range of consulting services. This includes areas such as digital marketing, customer experience, design thinking, financial services, and its global innovation labs, known as Lab45. By incorporating generative AI into its suite of services, Wipro aims to provide enhanced value and innovative solutions to its clients in various industries. The Bengaluru-based IT giant will also train around 20,000 associates on Google’s generative AI technologies to accelerate AI-driven transformation. “Through our expanded partnership, Wipro and Google Cloud will use generative AI to solve some of the biggest challenges businesses are facing today, safely and securely,” Thomas Kurian, chief executive at Google Cloud, said. HCL partners with Microsoft for generative AI adoption HCL has partnered with Microsoft to bring in generative AI capabilities to its customers. Microsoft’s OpenAI Azure services will give HCL access to leverage OpenAI’s GPT models, Codex and Embeddings model series. “By leveraging the latest Microsoft innovations in AI and machine learning, businesses can gain valuable insights into their operations, improve decision-making processes and achieve greater success. Whether enhancing customer experiences, streamlining supply chain operations or optimising business processes, this powerful collaboration provides the tools and expertise companies need to succeed in this fast-paced digital landscape,” Kalyan Kumar, chief technology officer and head, Ecosystems, HCLTech said.","excerpt":"Top Indian IT companies such as Infosys, TCS, Wipro, HCL and Tech Mahindra have announced generative AI capabilities to benefit their customers","categories":["AI Trends"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-06-19T16:00:00","publication_year":"2023","word_count":794,"keywords":["ChatGPT","machine learning","OpenAI","AI","AWS","ML","Transformers","RAG","Aim","generative AI"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","ChatGPT","OpenAI","Aim","Transformers","RAG","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-the-indian-it-giants-are-bringing-genai-capabilities-to-their-clients\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10013338,"title":"Top Cybersecurity Stories That Made Headlines In 2020","content":"Cybersecurity was the buzzword for enterprises around the world in 2020. With drastic changes to the conventional functioning of offices due to COVID-19 pandemic, cybersecurity emerged as the biggest challenge. Here we discuss some of the major cybersecurity stories that made headlines. Experts Against COVID- Related Hacking Incidents With an increase in the number of COVID-related breaches, in March 2020, an international group, consisting of 400 volunteers with expertise in cybersecurity was formed named COVID 19 CTI League. This group includes professionals from over 40 countries, working at major companies such as Microsoft and Amazon.com. Marc Rogers, one of the four initial managers of the initiative, who also head the security arm of the long-running hacking conference Def Con said that the group aims at working to combat hacks against healthcare, frontline responders to COVID, and of employees and businesses that are forced to operate from home in the given situation. Major Acquisitions In Cyber Security To strengthen their cybersecurity footing, several larger companies acquired smaller yet promising startups working in the field of cybersecurity. Some among them are: In April, Accenture announced its acquisition of Revolutionary Security, a startup that provides cybersecurity services such as risk assessment, breach and attack simulation testing, and building security programs. Spokesperson of Accenture said that with this deal, the company hoped to leverage proprietary technology to enhance security capabilities.In the same month, Accenture also agreed to buy a division of Symantec from Broadcom to expand its presence in the managed security services sector.In June, Microsoft announced its acquisition of Israeli cybersecurity startup CyberX which provides an IoT\/OT cybersecurity platform built by blue-team experts. Intelligence-led security company FireEye also acquired cybersecurity investigation automation company Respond Software in November. The deal was closed at $186 million in cash and stock and was among the most talked-about deals this year. Zoom also made a cybersecurity acquisition this year. The video conferencing company acquired Keybase, a startup that provides end-to-end encryption to protect calls from intruders and clandestine monitoring. India-Australia Collaboration For Cyber Technology In June this year, India and Australia signed a four-year collaboration on cyber affairs and critical technologies. This deal was closed at the Australia-India Leaders’ Virtual Summit, where a collective budget of $12.7 million was pledged towards this collaboration. Both the parties agreed to work hand-in-hand to promote and preserve an open, free, safe and secure internet, enhance digital trade, harness critical technology opportunities, and address cybersecurity challenges. Under this research and development in various technologies such as artificial intelligence, quantum computing, and robotics will be carried out. The fund will be utilised to support both Indian and Australian businesses and researchers to enhance cyber resilience. Cyber Security Policies & Initiatives By Indian States In August this year, PM Modi announced that India would soon get a robust national cybersecurity policy. This announcement was made in view of the increase in internet usage across the country, which brings with itself several cybersecurity challenges with online transactions, data phishing, identity theft, among others. The last national policy was introduced in 2013 named the National Cyber Security Policy 2013 that was formulated to build secure and resilient cyberspace for Indian citizens and businesses. In addition to the national policy, states such as Tamil Nadu and Karnataka announced state-level policies and frameworks in this regard in 2020. Leading the pack, Tamil Nadu became the first Indian state to come out with an ethical intelligence, cybersecurity and blockchain policies. The announcement was made by TN CM Edapaddi K Palaniswami in September who said that the aim is to facilitate speedy and efficient access to government services and to establish Tamil Nadu as the destination of choice for IT and non-IT services. Tailing close behind, Karnataka Deputy Chief Minister C.N. Ashwath Narayan, who is also the minister for IT&BT said in October that the state would soon have a cybersecurity policy. Data Breaches & Cyber Attacks This was a rather challenging year in terms of companies ensuring safe cyberspace and cyber hygiene given the whole work from home scenario. A PwC study earlier this year found that in the initial few months of 2020 itself saw doubling of the number of cyberattacks on Indian companies as cybercriminals misused the disruption brought about by the pandemic situation to maliciously enter companies’ networks and steal critical data. Some of the major breaches and attacks on enterprises are listed below: Twitter: Popular microblogging platform became victim to cyber-attack through a phone spear-phishing attack in July this year. The attackers used the credentials of the Twitter employees with access to tools; almost 130 accounts were said to be compromised. The attackers tweeted from 45 accounts, accessed the message box of 36, and downloaded the Twitter data of seven.Zoom: More than 500,000 Zoom accounts were breached, and the data was sold on the dark web for free or less than a penny each. As per sources, in the attack that happened in April, the stolen credential stuffing attacks.Unacademy: 20 million user accounts of the Bangalore-based popular edtech platform was breached in January. The breach had exposed usernames, SHA-256 hashed passwords, date joined, last login date, email addresses, first and last names, and other information.Marriott International: Marriott International faced a massive data breach in January this year compromising the personal information of around 5.2 million guests. It included contact details like name, mailing address, email address, and phone number, among other details.WHO: In April, 450 active WHO email addresses and passwords were leaked along with thousands of more which belonged to individuals associated with the novel coronavirus response teams worldwide. Also Read: 10 Biggest Data Breaches That Made Headlines In 2020 Increased Job Opportunities in Cybersecurity Like everything else, this extraordinary year too had a silver lining in the cybersecurity space. The demand for cybersecurity experts faced a major upswing this year. Studies found that there has been a 25% increase in the demand in the COVID-era and was widely recognised as the ‘hot skill’ of tomorrow. LinkedIn also reported cybersecurity as one of the most sort after jobs in India in 2020 As per reports, job portal Indeed witnessed a 98% increase in cybersecurity in just the last three years. In fact, Cyber Safety and Education expects a further rise in cybersecurity jobs that may reach 1.8 million by 2022.","excerpt":"Cybersecurity was the buzzword for enterprises around the world in 2020. With drastic changes to the conventional functioning of offices due to COVID-19 pandemic, cybersecurity emerged as the biggest challenge. Here we discuss some of the major cybersecurity stories that made headlines. Experts Against COVID- Related Hacking Incidents With an increase in the number of […]","categories":["AI Trends"],"tags":["Cyber Attack","Cybersecurity India"],"author_name":"Shraddha Goled","publish_date":"2020-12-06T10:00:20","publication_year":"2020","word_count":1047,"keywords":["Go","artificial intelligence","AI","Git","RAG","automation","Cyber Attack","Aim","Cybersecurity India","Ray","disruption","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","Ray","RAG","R","Go","Git","automation","disruption"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-cybersecurity-stories-that-made-headlines-in-2020\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":34554,"title":"Budget 2019: A 360-Degree View Of The Latest Announcements In Emerging Technologies","content":"Interim Finance Minister Piyush Goyal (Image credit: All India Radio\/Twitter) After path-breaking announcements in the field of artificial intelligence, data analytics and other emerging technologies in the year 2018-19, industry insiders had many hopes from this year’s interim budget. However, apart from a handful of key announcements, the interim budget for the year 2019-20 left many questions unanswered. Here is a 360-degree view of the announcements made in the emerging tech area by Acting Finance Minister Piyush Goyal in the Interim Budget 2019-20: National Centre On Artificial Intelligence Goyal announced that in order to make AI and other emerging technologies more accessible to the common people, a National Centre for AI will soon be established. He also announced that a national AI portal will be developed soon. He announced, “…A national programme on AI has been envisaged by the government. This should be catalysed by the national centre for artificial intelligence as a hub, along with other centres of excellence.” Priority Areas For National Deployment Of AI To actualise the ambitious initiative of the National Centre for AI, Goyal announced that the Central Government has identified nine priority areas. Union Minister said, adding compliance procedures and regulatory bottlenecks concerning startups are also being currently looked into. He also added that the Government would be working towards creating a national portal for AI in the current year itself. Buzzing Marketplace For Startups Globally Goyal also talked about how India has become the second-largest startup hub of the world. “We are proud of the hard work and innovative ideas of our youth in this sector,” he said. Goyal also added that due to this buzzing startup ecosystem, jobs are no longer being created in traditional factories and that job seekers have now become job creators. Youth Becoming Job Creators Through the Pradhan Mantri Kaushal Vikas Yojana, over 1 crore youth are being trained to help them earn a livelihood, said Goyal. “Youth power has been harnessed through self-employment schemes including MUDRA, Startup India and Stand-up India… The government is proud of the hard work and innovative ideas of our youth,” he added. One Lakh Digital Villages Goyal also announced that the Government was planning to create 1 lakh digital villages in India over the next five years. The Not-So-Bright Side Of This Budget According to the budget documents, the government in this year’s budget reduced the allocated amount to ₹25 crore from the revised estimate of ₹28 crore in FY19. The cumulative allocation for the department of industrial policy and promotion (DIPP) was also reduced to ₹5,674.51 crore for 2019-20. Meanwhile, last year, Finance Minister Arun Jaitley had doubled the allocation on Digital India programme to ₹3,073 crore in 2018-19. Experts have suggested that due to the upcoming elections, the focus of this year’s budget has been shifted from technology and business to education, agriculture and tax reforms for individuals. On the other hand, this budget has increased the allocation for the Make in India programme to ₹473.3 crore for 2019-20 from the revised estimate of ₹149 crore in 2018-19. This amount has been divided as follows: Investment promotion: ₹232.02 crore for Implementation of national manufacturing policy: ₹8.47 crore Funds: ₹100 crore Comparison To Budget 2018-19 So far, the Narendra Modi-led BJP Government has been working towards achieving its commitment towards Sustainable Development Goals (SDGs) with the help of AI by 2030, as it has the potential to churn out a slew of applications while keeping in mind the quality of approach. Industry insiders were happy so far to see that the Modi-led NDA government has been working towards supporting a tech-driven future. Acting upon these instructions, NITI Aayog had released its AI Strategy, a first-of-its-kind in India, which outlined the scope of research, adoption and commercialisation of AI in India. Outlook Startups have been suffering from a downward trend in funding, even though the startup ecosystem is thriving. Many personalities from this sector were hoping that the Budget would allow them some respite from Angel tax, but this issue was unaddressed during the interim budget. But looking at the new fund allocation for their ambitious Startup India plan, many have been worried about the Government changing their focus from business and corporates to public welfare. Despite the subtle change in track by the government, industry giants have welcomed the announcement and are feeling optimistic about the adoption of AI and ML at a large scale at the national level.","excerpt":"After path-breaking announcements in the field of artificial intelligence, data analytics and other emerging technologies in the year 2018-19, industry insiders had many hopes from this year’s interim budget. However, apart from a handful of key announcements, the interim budget for the year 2019-20 left many questions unanswered. Here is a 360-degree view of the […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","BJP","Narendra Modi"],"author_name":"Prajakta Hebbar","publish_date":"2019-02-06T09:41:17","publication_year":"2019","word_count":738,"keywords":["Go","funding","artificial intelligence","programming_languages:R","AI","ML","Git","BJP","analytics","Narendra Modi","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","R","Go","Git","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/budget-2019-360-degree-view-announcements-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10042416,"title":"Python Guide To Google’s T5 Transformer For Text Summarizer","content":"With the towering advancements in Transfer Learning, Deep Learning has achieved miraculous wonders. Especially in Natural language Processing, dominating these with the rise of Transformers, numerous approaches have arisen in the application of Language Modelling. We transfer the learning of a big model (mostly state-of-the-art) by pre-training it on a huge data corpus for a generic task, and the rest is fine-tuning it for specific tasks. In this blog, we will discuss Google AI’s state-of-the-art, T5 transformer which is a text to text transformer model. This was proposed earlier in 2020 in the paper “Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer”. The gist of the paper is a survey of the existing modern transfer learning techniques used in Natural Language Understanding, proposing a unified framework that will combine all language problems into a text-to-text format. Let’s take the example of BERT-style architecture. First, this model is trained on the objectives – Next Sentence Prediction and Masked Learning Models. Then we follow a simple procedure of fine-tuning the model for downstream tasks, predicting the label class in a classification problem or simple Question Answering. Hence, we separately fine-tune the model according to different tasks. On the contrary, this text-to-text framework(input and output both sequences are text) suggest that we use the same hyperparameters, same loss function and even the same model(same checkpoints and layers)!!  For all the NLP tasks. The inputs are modeled so that the model shall recognize a task, and the output is simply the “text” version of the expected outcome. For examples, let’s check out the following sentences : “Cola sentence: course is jumping well ” -> T5 -> “not acceptable” “Translate English to German: That is good” -> T5 -> “das ist gut” “Summarize: some text in paragraph” -> T5 -> “summary” As we can see in the above examples, it is demonstrated that we need to just add the task name in front of the input sequences (which can be done easily by string formatting )and feed it to the T5 transformer. Quite easy, right? Also, the output is in text format, which eliminates encoding and decoding. This transformer is trained on the C4 dataset (Colossal Clean Common Crawl), a super-cleaned counterpart of the normal standard Common Crawl containing web pages except the HTML markup. Super clean because this dataset has been cleaned of source codes, removing pages containing offensive words according to this list (dataset), removing duplicates and filtering out any pages containing non-English data. After all this, this resulted in a 750 GB huge dataset. Let’s dive into the code! Code Implementation of Text Summarization Setup and Importing Dependencies Keep in mind that the GPU should be on while making the notebook. Installing and importing all necessary libraries and utilities we need. A boilerplate for visualization  is also used to make plots with the same color scheme. Make sure the versions are updated of all the mentioned libraries. # change the runtime and make sure it is set to GPU # if not then follow this # Runtime-> Change Runtime -> GPU # this will restart the runtime so make sure you do it at the start # hugging face transformers and pytorch lightning latest versions !pip install --quiet transformers==4.5.0 !pip install --quiet pytorch-lightning==1.2.7 # you can also check the type of GPU you # by running the following command !nvidia -smi ''' json for reading html text pandas for df manipulation numpy for array operations torch for pytorch lightning empty cache utility ''' import json import pandas as pd import numpy as np import torch torch.cuda.empty_cache() # path for data from pathlib import Path # dataset and dataloader for functions from torch.utils.data import Dataset, DataLoader # lightning for data class import pytorch_lightning as pl # leveraging the model checkpoints from pytorch_lightning.callbacks import ModelCheckpoint # we can visualize performance of model from pytorch_lightning.loggers import TensorBoardLogger # splitting the data from sklearn.model_selection import train_test_split # color formatting in ANSII code for output in terminal from termcolor import colored # wraps the paragraph into a single line or string import textwrap # installing multiple utilities # including optimizer , tokenizer and generation module from transformers import ( AdamW, T5ForConditionalGeneration, T5TokenizerFast as T5Tokenizer ) # showing bars for processes in notebook from tqdm.auto import tqdm # seaborn for visualizing import seaborn as sns # procedural import to matplotlib from pylab import rcParams # graphs import matplotlib.pyplot as plt # rcParams for setting default values to all plots from matplotlib import rc # show graphs in the notebook cell %matplotlib inline # to render higher resolution images %config InlineBackend.figure_format='retina' # setting the default values for plots sns.set(style='whitegrid', palette='muted', font_scale=1.2) # make sure the fig size is not too big rcParams['figure.figsize'] = 16, 10 # random pseudo numbers pl.seed_everything(42) Download the data and do manipulation Here we will load the pre processed data which saves time from google drive. Then, further converting it into a Dataframe for manipulation. We have splitted the data into two parts namely train and test. # downloading the pre processed data !gdown --id 1DXsaWG9p3oQjkKu19mx_ZQ8UjHtsoGHW # taking a look at the data by df df = pd.read_csv(\"news_summary.csv\", encoding='latin-1') df.head() # slicing out useful columns df = df[['text', 'ctext']] # show the first 5 data points df.head() # changing the names of the columns df.columns = ['summary', 'text'] # dropping out the Not Available values df = df.dropna() df.head() # using sklearn utility, splitting the data into 10:1 ratio train_df, test_df = train_test_split(df, test_size=0.1) # let's check the shape of our data train_df.shape, test_df.shape Class for Dataset This class contains functions required for initialising objects of the arguments we have to input for the pytorch model. We define the data , tokenizer , lengths of input and output sequences ,  take care of encoding the data and add paddings and special tokens. # class for creating the dataset which extends from pytorch class NewsSummaryDataset(Dataset): # init it , create a constructor def __init__( self, # data in the form of a dataframe data: pd.DataFrame, # a tokenizer tokenizer: T5Tokenizer, # max token length of input sequence text_max_token_len: int = 512, # same for the summary but less length summary_max_token_len: int = 128 ): # saving all self.tokenizer = tokenizer self.data = data self.text_max_token_len = text_max_token_len self.summary_max_token_len = summary_max_token_len # length method def __len__(self): return len(self.data) # getting the items method def __getitem__(self, index: int): # data row from data at current index data_row = self.data.iloc[index] # get the full text text = data_row['text'] # encoding the text text_encoding = tokenizer( text, # setting max length max_length=self.text_max_token_len, # for same length padding='max_length', # cutting longer sequences truncation=True, # masking unwanted words return_attention_mask=True, # special tokens for start and end add_special_tokens=True, # return pytorch tensors return_tensors='pt' ) # same is done with summary encoding summary_encoding = tokenizer( data_row['summary'], truncation=True, return_attention_mask=True, add_special_tokens=True, max_length=self.summary_max_token_len, padding='max_length', return_tensors='pt' ) # creating the actual labels labels = summary_encoding['input_ids'] labels[labels == 0] = -100 # to make sure we have correct labels for T5 text generation return dict( # data text=text, # task summary=data_row['summary'], # easy batching text_input_ids=text_encoding['input_ids'].flatten(), # masking text_attention_mask=text_encoding['attention_mask'].flatten(), # again flatten labels=labels.flatten(), labels_attention_mask=summary_encoding['attention_mask'].flatten() ) Class for DataModule This class will develop a data module (input method ) for plugging in the utilities as well as the data. Pytorch lightning is not used commonly but it makes sure fast model training and running. # data module for pytorch lightning class NewsSummaryDataModule(pl.LightningDataModule): def __init__( self, # pass in train data train_df: pd.DataFrame, # pass in test data test_df: pd.DataFrame, # tokenizer tokenizer: T5Tokenizer, # batch_size batch_size: int = 8, # length of sequence text_max_token_len: int = 512, # length of output sequence summary_max_token_len: int = 128 ): super().__init__() # storing the data in class objects self.train_df = train_df self.test_df = test_df self.batch_size = batch_size self.tokenizer = tokenizer self.text_max_token_len = text_max_token_len self.summary_max_token_len = summary_max_token_len # automatically called by the trainer def setup(self, stage=None): self.train_dataset = NewsSummaryDataset( self.train_df, self.tokenizer, self.text_max_token_len, self.summary_max_token_len ) self.test_dataset = NewsSummaryDataset( self.test_df, self.tokenizer, self.text_max_token_len, self.summary_max_token_len ) # for train data def train_dataloader(self): return DataLoader( self.train_dataset, batch_size=self.batch_size, shuffle=True, num_workers=2 ) # for test data def test_dataloader(self): return DataLoader( self.test_dataset, batch_size=self.batch_size, shuffle=True, num_workers=2 ) # valid data def val_dataloader(self): return DataLoader( self.test_dataset, batch_size=self.batch_size, shuffle=True, num_workers=2 ) Load, Fine-Tune, the Model We have to load the T5 model having 222M params and instantiate it with the model with setting up the number of epochs and batch size. We also have to name the tokenizer (T5 tokenizer as it saves time) to be used ahead. # leveraging the base T5 transformer MODEL_NAME = 't5-base' # instantiate the tokenizer tokenizer = T5Tokenizer.from_pretrained(MODEL_NAME) # empty lists which are to be filled text_token_counts, summary_token_counts = [], [] # traversing train data for _, row in train_df.iterrows(): # encode the data points text_token_count = len(tokenizer.encode(row['text'])) text_token_counts.append(text_token_count) # do the same with the summary summary_token_count = len(tokenizer.encode(row['summary'])) summary_token_counts.append(summary_token_count) # plotting for token lengths and counts fig, (ax1, ax2) = plt.subplots(1, 2) sns.histplot(text_token_counts, ax=ax1) # the same for summary ax1.set_title('full text token counts') sns.histplot(summary_token_counts, ax=ax2) # number of epochs is less because number of parameters is high N_EPOCHS = 3 BATCH_SIZE = 8 # call the data module data_module = NewsSummaryDataModule(train_df, test_df, tokenizer) Class for Model Summary Model summary is for defining the properties of the model and all the arguments including loss function , learning rate and optimizer,  so on. Separate functions for training step, validation and testing with quantities returned. # create lightning module for summarization class NewsSummaryModel(pl.LightningModule): def __init__(self): super().__init__() self.model = T5ForConditionalGeneration.from_pretrained(MODEL_NAME, return_dict=True) def forward(self, input_ids, attention_mask, decoder_attention_mask, labels=None): output = self.model( input_ids, attention_mask=attention_mask, labels=labels, decoder_attention_mask=decoder_attention_mask ) return output.loss, output.logits def training_step(self, batch, batch_size): input_ids = batch['text_input_ids'] attention_mask = batch['text_attention_mask'] labels = batch['labels'] labels_attention_mask = batch['labels_attention_mask'] loss, outputs = self( input_ids=input_ids, attention_mask=attention_mask, decoder_attention_mask=labels_attention_mask, labels=labels ) self.log(\"train_loss\", loss, prog_bar=True, logger=True) return loss def validation_step(self, batch, batch_size): input_ids = batch['text_input_ids'] attention_mask = batch['text_attention_mask'] labels = batch['labels'] labels_attention_mask = batch['labels_attention_mask'] loss, outputs = self( input_ids=input_ids, attention_mask=attention_mask, decoder_attention_mask=labels_attention_mask, labels=labels ) self.log(\"val_loss\", loss, prog_bar=True, logger=True) return loss def test_step(self, batch, batch_size): input_ids = batch['text_input_ids'] attention_mask = batch['text_attention_mask'] labels = batch['labels'] labels_attention_mask = batch['labels_attention_mask'] loss, outputs = self( input_ids=input_ids, attention_mask=attention_mask, decoder_attention_mask=labels_attention_mask, labels=labels ) self.log(\"test_loss\", loss, prog_bar=True, logger=True) return loss def configure_optimizers(self): return AdamW(self.parameters(), lr=0.0001) Fit the Model model = NewsSummaryModel() %load_ext tensorboard %tensorboard --logdir .\/lightning_logs checkpoint_callback = ModelCheckpoint( dirpath='checkpoints', filename='best-checkpoint', save_top_k=1, verbose=True, monitor='val_loss', mode='min' ) logger = TensorBoardLogger(\"lightning_logs\", name='news-summary') trainer = pl.Trainer( logger=logger, checkpoint_callback=checkpoint_callback, max_epochs=N_EPOCHS, gpus=1, progress_bar_refresh_rate=30 ) trainer.fit(model, data_module) trained_model = NewsSummaryModel.load_from_checkpoint( trainer.checkpoint_callback.best_model_path ) trained_model.freeze() Summarize Text Function and utilities for output function to be called. We will use this function for sample output in our code later on. def summarizeText(text): text_encoding = tokenizer( text, max_length=512, padding='max_length', truncation=True, return_attention_mask=True, add_special_tokens=True, return_tensors='pt' ) generated_ids = trained_model.model.generate( input_ids=text_encoding['input_ids'], attention_mask=text_encoding['attention_mask'], max_length=150, num_beams=2, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True ) preds = [ tokenizer.decode(gen_id, skip_special_tokens=True, clean_up_tokenization_spaces=True) for gen_id in generated_ids ] return \"\".join(preds) Sample for Output sample_row = test_df.iloc[0] text = sample_row['text'] model_summary = summarizeText(text) text sample_row['summary'] model_summary sample_row = test_df.iloc[1] text = sample_row['text'] model_summary = summarizeText(text) text sample_row['summary'] Model_summary Saving the Model import pickle filename = open('text_summarization_model.pkl', 'wb') pickle.dump(trained_model.model, filename) # saving the model model = pickle.load(open('text_summarization_model.pkl', 'rb')) # function for producing output provided input def summarizeText(text): text_encoding = tokenizer( text, max_length=512, padding='max_length', truncation=True, return_attention_mask=True, add_special_tokens=True, return_tensors='pt' ) generated_ids = model.generate( input_ids=text_encoding['input_ids'], attention_mask=text_encoding['attention_mask'], max_length=150, num_beams=2, repetition_penalty=2.5, length_penalty=1.0, early_stopping=True ) preds = [ tokenizer.decode(gen_id, skip_special_tokens=True, clean_up_tokenization_spaces=True) for gen_id in generated_ids ] return \"\".join(preds) EndNote We have seen a brief introduction to a spectacular all in one transformer model for NLP tasks and successfully implemented a beneficial application on this model. I recommend using different datasets (the dataset used in this article is on Kaggle) and try newer applications or implementations on this model. References: Official Github RepositoryResearch PaperOfficial Source CodeColab Implementation","excerpt":"We will discuss Google AI’s state-of-the-art, T5 transformer which is a text to text transformer model. The gist of the paper is a survey of the existing modern transfer learning techniques used in Natural Language Understanding, proposing a unified framework that will combine all language problems into a text-to-text format.","categories":["Deep Tech"],"tags":["Generative Pre-Trained Transformer","Guide","keras python","Text summarization","Transfer Learning","Transformer Model","Transformers"],"author_name":"Mudit Rustagi","publish_date":"2021-06-26T14:00:00","publication_year":"2021","word_count":1946,"keywords":["Hugging Face","AI","PyTorch","ML","Text summarization","Transformers","NLP","Ray","Colab","deep learning","Generative Pre-Trained Transformer","keras python","Transfer Learning","Transformer Model","Guide","Pandas"],"extracted_tech_keywords":["AI","ML","deep learning","NLP","Ray","PyTorch","Hugging Face","Transformers","Colab","Pandas"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/python-guide-to-googles-t5-transformer-for-text-summarizer\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10103006,"title":"Zomato Cooks Generative AI with Microsoft Azure","content":"Zomato has been aggressively investing in generative AI so much so that it now has a dedicated role – head of generative AI – led by Vaibhav Bhutani, where he is working on couple of interesting use cases, including a multi-agent system for suggesting food options and enhancing user engagement and experience and more. Zomato believes that it is on a mission to “power India’s changing lifestyles.” With three main divisions—Zomato, Blinkit, and Hyperpure—the company is dedicated to providing better food experiences for more people. From revolutionising instant commerce to tackling malnutrition through its feeding arm, Zomato’s mission is multifaceted, aiming to cater to diverse aspects of society. Bhutani discussed at the first episode of Azure Innovation Podcast with Ross Kennedy, VP of Digital Natives, about how Zomato uses generative AI to provide different customer experience. (see below) https:\/\/youtu.be\/jcyQdEahaf0?si=FIbKjX-4lpImoFon People need convenience because they are lazy “People are lazy, and it’s very hard to have a conversation these days,” said Bhutani emphasising that having conversations is the only way to collect data, and he said that it should become easier. This is one of the metrics that Zomato is chasing with generative AI. “Is chasing a chatbot an everyday use case, or is it not. And while building all these bots, one of the biggest learnings that I have had is that the UI of these bots is what truly matters,” said Bhutani highlighting that the company focuses and critically thinks about UI and UX design a lot. Bhutani said that he had a ton of options about how he would use generative AI. He emphasised that Microsoft and OpenAI’s GPT through Azure emerged as the natural choice, given Azure’s robust commitment to data privacy and security. The LLMs provided on Azure OpenAI Service became the foundation of Zomato’s generative AI architecture. Bhutani further explained the adoption of a multi-agent system, where AI agents communicate with each other to offer comprehensive responses for the customers. The integration with Zomato’s ecosystem involves creating functions that seamlessly leverage generative AI contributing to a cohesive user experience. “For me, adopting generative AI is a more personal journey,” said Bhutani. Drawing from personal experiences at Spyne, a SaaS AI photography tool, where he recognised the potential of generative AI in enhancing Zomato’s capabilities. The journey began with the creation of Recipe Rover at Blinkit—an AI-generated recipe tool featuring thousands of recipes, images, and text, using GANs and the company’s proprietary data. Bhutani believes that because of the amount of experimentation that Zomato has already done with generative AI, it is definitely going to be a copilot for our company. “we’re working with different teams who need high impact co-pilots so that we can enable large fleets of our team to make much better decisions and faster decisions,” he said. Generative AI in Food In September, Zomato released its Zomato AI Buddy. Going beyond the limitations of traditional chatbots, Zomato AI stands as an intelligent and intuitive foodie companion, dedicated to understanding and satisfying users’ ever-changing preferences, dietary requirements, and even their current moods. One of the standout features of Zomato AI is its multiple agent framework, which equips it with a diverse range of capabilities to serve customers at any given moment. The framework provides Zomato AI with a variety of prompts for different tasks and activities. For instance, if you’re craving a specific dish, the AI will swiftly present you with a widget listing all the restaurants that serve your desired meal. If you’re uncertain about what to order, Zomato AI can suggest a list of popular dishes or restaurants, eliminating the guesswork from your meal selection. Zomato’s competitor, Swiggy, another food delivery platform in India, started using generative AI in July. Amitkumar Banka told AIM that the company is using generative AI to create customised food images based on specific requirements on their platform, and this is helping them serve millions of customers. “We are using generative AI to put the name, description, and image of food items to individual users based on browsing behaviour not only on Swiggy but on the entire internet,” he said. Similar to Zomato, Swiggy also recently unveiled a new feature called ‘WhatTo Eat’ that allows users to explore options based on their mood and cravings.","excerpt":"“Because the people are lazy,” said Vaibhav Bhutani, the head of generative AI at Zomato, in the first ever episode of Azure Innovation Podcast.","categories":["AI Highlights"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2023-11-14T17:52:04","publication_year":"2023","word_count":709,"keywords":["Go","OpenAI","AI","chatbots","R","ML","RAG","Aim","generative AI","Azure"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Aim","RAG","chatbots","Azure","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/zomato-cooks-generative-ai-with-microsoft-azure\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":25196,"title":"Analytics Hiring Scenario For Indian Companies Looks Positive, Says Raj Sekhar Of EdGE Networks","content":"This time, for our interaction around analytics hiring scenario in India, we got in touch with Raja Sekhar Pappala, who is the head of analytics at EdGE Networks. The company is in the business of building applications to help organisations acquire the best talent. This HR tech company uses data science and people analytics to ensure workforce transformation. Their productivity tool harnesses artificial intelligence to help solve the toughest talent acquisition and workforce optimisation problems. Raja Sekhar shares how hiring scenario in Indian companies looks like and the major challenges they face. Raja Sekhar has in the past led several initiatives leveraging data sciences in recommendation systems, semantic search, sensor analytics, image processing, text analytics, among others. His current focus is building machine learning applications for HR technology that deliver significant business impact. Analytics India Magazine: How does the analytics hiring scenario in Indian companies look like? Raja Sekhar: It definitely looks positive. In my view, the bulk of analytics hiring happens in larger companies but most of the top tier institutes in the country have aggressively ramped up their analytics curriculum and their study programs, which leads us to believe that analytics hiring is very much on the rise. AIM: What are the skillsets that companies are mostly looking at while hiring analytics talent? RS: Skills in analytics have a vast range. It begins with a good understanding of data-for-data pipeline creations, and that’s why SQL, ETL methods with Python and other querying expertise is sought after. Data analytics require a range of skills like statistical knowledge and implementing it using SAS, R and Python, able to form and validate statistical hypotheses, advanced statistical methods of data imputation, and data validation in both descriptive and predictive analytics, among others. Data visualisation using Tableau, Power BI, Qlikview or similar tools comes in parallel as it is a critical aspect of business analyses too. Model-building and knowledge of statistical algorithms also plays a major role in AI hiring. The ability to predict certain critical business parameters takes years of practice to perfect. Big data technologies such as Hadoop and Apache Spark are also a relevant skills in analytics as most of the companies run into terabytes of data or even more. AIM: The talent gap is often a talk point in the industry. How could it be bridged? RS: This gap could be bridged by recognising the demand, which is being created for the future. Given that ML and AI will be used in almost any critical business decision-making, ensuring this is an integral part of course curriculums, introducing advanced and relevant programs for those aspiring will definitely help bridge the talent gap. Interactive sessions by real-world practitioners will also generate a serious interest in this booming AI market. AIM: What are the various initiatives that companies and educational institutions can take to set right the analytics talent flow? RS: Initiatives could include sessions and webinars conducted by experienced AI practitioners, workshops which offer hands-on expertise in terms of solving real-world problems, live hackathons and active participation in analytics communities. AIM: What are the challenges in the current education system that stops that growth of analytics talent? RS: In my opinion, there’s still a void between the way an analytical subject is taught in most schools compared, and its real-world applications. Statistics, applied mathematics and applied econometrics, are examples of an analytical curriculum, which unless studied from a top tier institute or platform may offer very less value, mostly in terms of the way it is used in real world to solve analytical problems. Without a link between its correct understanding and usage, it becomes quite difficult for students to grasp such subjects and later on, they may find it difficult to apply their knowledge in solving business problems. AIM: What does the typical hiring process look like at your company? What are different stages, processes etc? RS: Typically, the candidate goes through three to four rounds. It usually starts with a telephonic or Skype technical round, followed by a problem-solving round where the candidate is asked to solve an analytics problem. This is mostly followed by a face-to-face technical and general aptitude round, followed by an HR and management discussion. AIM: What are the current analytics openings at your organisation? What are the skill sets that you for? RS: We are always keen on adding talented and enthusiastic people in our team. AIM: For an analytics professional who wishes to carve a career in analytics industry, what is your advice? RS: There has to be love for data science. Everything cannot be taught nor is it possible to be accommodated in textbooks. The eagerness to learn the discipline will mostly see one through all hurdles. As mentioned before, apart from rigorous curriculum, trying to solve real world business problems, keeping abreast of analytical technologies and the eagerness to learn from a plethora of information found in white papers, journals, articles on the web is a start. AIM: What are your thoughts on attrition? RS: Solving attrition-related problems through analytics is something which most companies have started. Predicting attrition trends, identifying seasonal patterns to attrition, whether it is profitable to hire from top institutes in terms of attrition and cost, who is likely to leave the organisation in the next six months and why, which factors drive attrition in specific grade groups are some of the problems, which revolve around attrition analytics. AIM: What are the offerings by Edge Networks to help companies acquire right talent? RS: HIREalchemy is company’s flagship product, which is a cutting-edge talent acquisition platform powered by AI and data science. The platform auto-sources the right fit by parsing and analysing both structured and unstructured information from internal database as well as external portals. It eases the process of selection as the platform throws up resume matches by scoring and stack ranking them based on business rules set by the client. Other key distinctive offerings include Workforce Optimisation Solution, which forms an intelligence layer on top of HR systems and helping in effective organisation building. Whereas another offering, Talent Analytics Suite by the company helps in predicting attrition, forecasting resource demand and enabling fact-based decision making across the HR value chain. Our Workforce Planning solution, on the other hand, helps align talent strategy with business strategy through advanced analytics algorithms. We provide actionable analytics and insights that accelerates business decision making.","excerpt":"This time, for our interaction around analytics hiring scenario in India, we got in touch with Raja Sekhar Pappala, who is the head of analytics at EdGE Networks. The company is in the business of building applications to help organisations acquire the best talent. This HR tech company uses data science and people analytics to […]","categories":["AI Features"],"tags":["analytics hiring india","Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-06-06T11:52:41","publication_year":"2018","word_count":1064,"keywords":["analytics hiring india","data science","machine learning","artificial intelligence","AI","ML","recommendation systems","RAG","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","RAG","recommendation systems","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-hiring-scenario-for-indian-companies-looks-positive-says-raj-sekhar-of-edge-networks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10063342,"title":"Bharathiar University &#038; ISDC to offer IoA accredited analytics courses","content":"Bharathiar University has signed an MoU with the International Skill Development Corporation (ISDC) to offer a wide range of courses in business analytics accredited by the Institute of Analytics (IoA). ISDC has been working closely with Indian universities and educational institutions to bridge the skill gap across all sectors. The students will be able to acquire in-depth knowledge of a variety of programs related to analytics that will help them navigate the business environment using analytics tools and techniques. Industry and professional trainers from ISDC will offer these courses through blended learning, which will include seminars, interactive workshops, as well as corporate engagement. The IoA is fully supported in its mission by leading organisations in the analytics field. The Institute of Analytics offers students the opportunity to network with other professionals in the industry, learn about career development opportunities, share knowledge and expertise, access ongoing professional development programs, and pursue official designation of membership.“We are excited about our partnership with the International Skill Development Corporation (ISDC) in offering our students IoA accredited analytics programs. With this partnership, students will have the opportunity to learn domain-specific topics. This partnership will give students an opportunity to develop their skills at global levels, enabling them to obtain better employment opportunities in India and abroad. This is a great opportunity for our students to take globally recognised IoA accredited courses and gain membership of the IoA UK after graduation,” said Dr K Kaliraj, Vice-Chancellor, Bharathiar University.","excerpt":"This is a great opportunity for our students to take globally recognised IoA accredited courses and gain membership of the IoA UK after graduation.","categories":["AI News"],"tags":[],"author_name":"Kartik Wali","publish_date":"2022-03-23T13:36:49","publication_year":"2022","word_count":242,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bharathiar-university-isdc-to-offer-ioa-accredited-analytics-courses\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":39752,"title":"The Importance Of Data Munging For Data Preparation In Analytics","content":"Analysis of data and transforming it into some meaningful insights has become an integral part of an organisation. Data Munging is the process by which the data is identified, extracted, cleaned as well as integrated in order to gain a good dataset that is suitable for both exploration and analysis. Data Munging can also be referred to as data wrangling and it includes various aspects such as data quality, merging of different sources, reproducible processes, managing data, etc. It has been estimated that a staggering 70% of the time spent on analytic projects is concerned with identifying, cleansing and integrating data due to the difficulties of locating data which is scattered among many business applications, the need to re-engineer and reformat it in order to make it easier to consume, and the need to regularly refresh it is to keep it up-to-date. This cost, along with recent trends in the growth and availability of data, has led to the concept of a capacious repository for raw data called a data lake, which is a set of centralized repositories containing vast amounts of raw data. Why Is It Important Data munging plays a crucial role in an organisation. The process can be time-consuming but the valuable insights it is producing plays an important role in the organisation. The wrangled data can be organised into a standard repeatable process which can be moved and transformed in a common format and can be reused later for multiple times. Steps For Data Munging According to Trifacta, one of the established leaders of the global market for data preparation technology, data wrangling involves mainly six core activities. They are mentioned below. Discovering: In this process, you understand and learn what is there in your data and to find the best way for some productive analytic explorations. Structuring: Data is usually in the raw form. While analysing the data, it needs to make sure that the data is restructured in the way which suits better during the analytical procedures. Cleaning: Inconsistent and noisy data cannot be used to gain meaningful insights in an organisation. The noisy data needs to be cleaned before it is used for analytical approaches. Enriching: In this process, the cleaned data is enriched by analysing what new data can be derived from the existed data. This new information is sometimes available in in-house databases, but, and increasingly so, may be sourced from marketplaces for third-party data. Validating: Validating is the activity that surfaces data quality and consistency issues, or verifies that they have been properly addressed by applied transformations. Validations should be conducted along multiple dimensions. Publishing: Publishing refers to planning for and delivering the output of your data wrangling efforts for downstream project needs (like loading the data in a particular analysis package) or for future project needs (like documenting and archiving transformation logic). How Is It Different From Data Mining Data mining is a process of discovering some specific hidden patterns in a large dataset whereas data munging is a superset of data mining which involves various process such as cleaning, transforming, integrating, etc. in a large dataset for decision-making. The outcome of a data mining process is meaningful pattern whereas the output of a data munging is a meaningful insight. Skills Required For Data Munging A data wrangler solves all the data related issues right from the integrating, cleaning, and transforming. Data is everywhere but it is mostly in the raw form. A good data wrangler requires adequate skills such that he\/she can integrate information from various data sources. Most often organisations choose data wranglers with a specific set of skills such as a wrangler with efficient knowledge in a statistical language such as R, Python, etc., adequate understanding in the business context, knowledge in other programming languages such as SQL, PHP, Julia, Scala, etc.","excerpt":"Analysis of data and transforming it into some meaningful insights has become an integral part of an organisation. Data Munging is the process by which the data is identified, extracted, cleaned as well as integrated in order to gain a good dataset that is suitable for both exploration and analysis. Data Munging can also be […]","categories":["AI Trends"],"tags":["data transformation"],"author_name":"Ambika Choudhury","publish_date":"2019-05-27T12:40:30","publication_year":"2019","word_count":636,"keywords":["Go","TPU","AI","data transformation","Scala","Python","data quality","SQL","Julia","R","data lake"],"extracted_tech_keywords":["AI","TPU","Python","R","SQL","Go","Scala","Julia","data lake","data quality"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-importance-of-data-munging-for-data-preparation-in-analytics\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":25910,"title":"Managing mine dust pollution in near real time leveraging IoT and Analytics","content":"The pollution levels are rapidly increasing due to continuous mining. This causes the air quality to deteriorate around mining complexes. On the other hand, increased focus on environmental compliance has led to monitoring and regulating the environment at mines and mineral\/ore handling points. The dangerous emissions and dust generated due to mining operations have a significant impact on the habitats, environment and the equipment in the mines. The poisonous gases, mineral particles getting mixed in the water bodies, air and the food chain are considered the adverse impacts of mining. The dust is generated from vehicle movement on haulage roads, excavation, crushing, milling, conveyors belts, blasting, loading, unloading, wind erosion of stockpiles and overburden. The gases, harmful particles emanate from vehicles, beneficiation plants at the mines and material handling points at mines and ports. These dust and emissions are toxic in nature and impact flora and fauna, soil and air quality. Presently, the dust is controlled as a scheduled activity with process control systems and spray trucks. All the present systems work on dust suppression process control methods, based on measurement at a location and actionable insights are time-based schedules. There are two identified objectives to reduce the impact of mining-related pollution – Reduction of dust and harmful impurities which impacts water, environment and living beings Optimized real-time use of resources – water, equipment, chemical, manpower, energy Research done by CDC (Centers for Disease Control and Prevention) – around the theme of the dust control and its impact reveals “When less dust is generated, less has to be removed from the mine air”. In a coal mine, thousands of truck trips covering millions of kilometres per annum, account for nearly 30-40% of costs incurred in the surface mine, hence it becomes essential to minimize the costs. With the rise of Industry 4.0 and Internet of Things (IoT), enables the mining industry to integrate the entire mining process creating opportunities to build Dust Conditioning solutions. Dust Conditioning Solution using IoT sensors and analytics Mining industry is transforming to the ‘Digital mine’, resulting in connectivity between equipment systems in real time, enabling real-time analysis and decision making. There are numerous use cases of IoT and Analytics solutions across industries such as mining, power generation, ports, construction sites and irrigation.  Real-time IoT and Analytics based solutions, will ensure safety, reliability and compliance with the environmental standards set by the statutory and regulatory bodies. Real time IoT and Analytics solutions provide real time monitoring, extrapolation of dust and moisture. Subsequently evaluated actions to contain the dust as per the required optimal moisture level is known as “Dust Conditioning”. Let us look at a three-step process to design a real time monitoring IoT and Analytics solution for Dust Monitoring and Management. Step 1 – IoT based solution for Real time monitoring (based on sensors and data API’s) of the key parameters in a surface mine through platforms located at a distance ranging between 100 to 500 m separation. Moisture, Dust, Gas, Rain Air quality Weather insights (Data API’s from any global weather platform) Network of physical monitoring of points Tank, Tanker water levels Measuring of relevant pollutants Step 2 Extrapolating and analyzing the real time insights using Analytics engines – the data collected are analyzed and extrapolated using Analytics Engines to provide unique insights such as – Extrapolation and analytics of dust level – haul road Extrapolation and analytics of moisture level – haul road Collation of dust and moisture level plotting Definition of variable functions for calculation of target moisture Cognitive analytics of historical data and inputs from weather station Calculation of delta moisture level requirement from the benchmark for dust suppression Calculation of water requirement plotting for the driver or operator in the specific area Decision Framework – Determination of water\/chemical (suppressants) quantity requirement in volume at specific location – graphical display to the operator for ease of usage Feedback of the actions for the next cycle for optimal use of resources and learn from managing situations The monitoring and analytics used for this solution is for a haulage road as visualized in figure 1. The dust generated can be monitored and plotted on real time as shown below on the entire haulage road (location) of the mine with a proper heat map where the dust levels are high or above the threshold value. Here dust level of greater than 120 mg\/m3 is considered higher than the accepted value for dust levels. Figure 1 – Dust level at Haul road The real time moisture level at the same location is also plotted as below, the lower the moisture the higher the dust levels. The 7% moisture levels help reduce dust levels for coal particles. Figure 2 – Moisture level at haul road The Overlap of dust and Moisture level provides insights around zones that have the highest requirement of moisture so that the dust levels can be minimized (Moisture level is scaled up by 10 to showcase in the graph). Figure 3 – Dust and moisture level at haul road Figure 4 – The delta moisture required at the location of haul road Step 3 Design and Development of a Real Time Dashboard – the dashboard will provide the real time information related to dust, moisture and heat map and insights around where the maximum suppressant is required. The dashboard will also provide insights on water level, tanker positions, driver\/operator and weather. These insights will be evaluated by monitoring experts and help deriving the real time actionable intelligence for conditioning actions to contain the dust. The real time dashboard will also enable the sprayers i.e. tankers, spray jets process controls to pin point the exact location grid where spraying is required for dust conditioning. The graphical user interface will help the operator or the process control system by using coordinates or heat maps. The heat map can also be provided to the operator of spray tanker or the operator of dust control department to utilize the suppressant in a more effective way. See Figure 5 for the Real Time Dashboard architecture – Figure 5 – Real Time Dust Conditioning Dashboard architecture The operator can see the real time moisture requirements to spray upon as in figure 6 as a heat map where he needs to spray more suppressant. Figure 6 – Heat map in the Dashboard of the operator Key Solution Benefits: – for the mining industry, the benefits of the real time solution would include Increased life of haulage roads – The proper treatment of haulage roads leads in increasing its life reducing the costs of haulage road maintenance. Smooth movement of fleet equipment at haulage roads which helps in improving productivity – reduction in cycle time of haulers \/ dumpers as the average length of haulage road in a mine can be in the range of 2 Km to 9 Km. Controlled consumption of water and chemicals dust suppressants – the traditional way of controlling dusts is a scheduled spray of chemicals evenly across the haulage road. This solution will help pin point the exact location and amount to be sprayed thus optimizing the consumption of resources. Reduction in wear of tires – The HEMM (Heavy Earth Moving Machinery) tires are very costly so the proper maintenance of haul roads will help maintain the same useful life as it is optimally designed for. Increase in safety aspects – The over spraying of water, chemicals aid the deterioration of haul roads helping create potholes, loose material, rutting, corrugation and slippery which may result in near misses or accidents. Proper dust conditioning helps in optimal amount of sprays at appropriate location to optimally maintain good haul road condition. Effective and real-time utilization of resourcese. mobile spray tankers, stockpile spray system, operators, tankers. Proper monitoring, forecasting and consumption of inventory of water and chemicals. Real time reduction of suspended dust particles which in help reduces the chances of severe medical conditions in a long run for the mine employees. The same model can be applied at stockpiles, loading point, unloading point, tailings dam and ports helping reduce wind erosions by appropriate conditioning. Conclusion The usage of IoT based real time solution will enhance the monitoring of dust hazards and optimize wastage of suppressant, water and increase the haul road life. The challenges are the connectivity of the sensors which will feed data to the server for real time data analytics. If the real time connectivity is ensured, it will enable improvements in the air quality and reduce operating costs of the mine. Soon, analyzing the dust levels using visual\/video analytics will also provide the enhanced framework models for dust containment. Author Bio’s Niraj is an Industry & Technology Consultant having 21 years of experience focused on consulting, strategy and operations improvement. Currently, Niraj is Metals & Mining sub Industry Practice Leader in IBM Singapore. His interests include Niraj is also pursuing PhD from Indian Institute of Technology- Dhanbad (erstwhile ISM-Dhanbad). Saumya Chaki is the Data Platform Solutions Lead at IBM by day and an author by night. Has a Masters in Technology from IIT (ISM) Dhanbad. He has authored two books ‘Enterprise Information Management in Practice’ and ‘A Journey Through 100 Years of Indian Cinema’. His interests include Big Data Analytics, Smarter Cities. Climate Change and Mineral Economics. His hobbies include travel, blogging and reading. He is currently working on his third book ‘Baptism by Fire’ a work of fiction.","excerpt":"The pollution levels are rapidly increasing due to continuous mining. This causes the air quality to deteriorate around mining complexes. On the other hand, increased focus on environmental compliance has led to monitoring and regulating the environment at mines and mineral\/ore handling points. The dangerous emissions and dust generated due to mining operations have a […]","categories":["AI Features"],"tags":["iot enterprise architecture"],"author_name":"AIM Media House","publish_date":"2018-06-30T12:17:06","publication_year":"2018","word_count":1566,"keywords":["big data","Go","API","ELT","AI","iot enterprise architecture","Git","RAG","Ray","analytics","R"],"extracted_tech_keywords":["AI","analytics","Ray","RAG","R","Go","Git","API","big data","ELT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/managing-mine-dust-pollution-in-near-real-time-leveraging-iot-and-analytics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":39599,"title":"4 Ways ML Can Strengthen An Organisation’s Security Infrastructure","content":"This ever-evolving tech-driven world is witnessing one of the most dangerous eras of cyber attacks. With this skyrocketing number of data breaches and hacks, it has become imperative for organisations across the world to adopt the new waves of technological advancements. Talking about advanced techs, artificial intelligence (AI) and machine learning (ML) definitely tops the list, Machine learning and artificial intelligence in security are the fast-growing trends and by implementing  AI and ML, organisations can strengthen their cybersecurity infrastructure and can mitigate cyber threats to a great extent. It can help business in threat analysing and make them better in responding to attacks, by automating a lot of tasks that used to be done by humans. Security is important and to make cybersecurity stronger, it is also important to have an environment with all the latest, advanced technologies. In this article, we are going to see some of the top machine learning and artificial intelligence use cases in cybersecurity. Beforehand Detection of Vulnerabilities To carry out a cyber attack it is mandatory for attackers to find out vulnerabilities, and once a flaw is found, attackers exploit that. Utilising ML and AI in cybersecurity infrastructure, organisations can develop or upgrade systems to scan for loopholes and vulnerabilities. Also, ML algorithms can help businesses detect malicious activity faster, which would help the cybersecurity team to act faster and better. That is not all with advanced AI systems, organisations could also have a picture of various scenarios of how hackers could exploit them. This would not only help mitigate cyber attacks but also prepare them for the upcoming attacks. Automate Tasks Cybersecurity is one of those domains that need a lot of manpower in order to keep every end working fine. However, sometimes it’s not possible to do everything, and this is where technologies like AI and ML comes in. Machine learning has the capability to automate tasks, allowing the cybersecurity professional to focus on other tasks. For example, it is not an easy task for a human to decide which vulnerability to fix first when there is a lot. Machine learning can do the sorting and let the human work on that vulnerability that seems more critical. Another example of using these sought after techs is when  there is a ransomware attack. While the machine can keep interrupting the attack, the cybersecurity professional can figure out ways to eliminate it. To Tackle Zero-Day Exploits If you know what Zero-Day exploits are then you would definitely understand the pain that cybersecurity professionals get while dealing or stopping someone from exploiting a zero-day vulnerability. Basically, a  zero-day vulnerability is a software vulnerability that is unknown to, or unaddressed by, those who should be interested in mitigating the vulnerability. So, this is where machine learning can be used to make it easier for cybersecurity professionals. By leveraging the powers of ML, one can build a system that would monitor traffic about security exploits. Meaning, if hackers are trying to intercept the traffic, it would alert the cybersecurity team that someone is trying to pwn. Basically, its like hacking the hackers. Helps In Analysing Mobile Endpoints Machine learning and artificial intelligence are already being used widely in a lot of mobile devices. However, now their application has been expanded and are now being used to analyse mobile endpoints in cybersecurity. Over the past couple of years, BYOD has gained a lot of popularity, but it is also posing a lot of security loopholes. Using machine learning, companies can analyse mobile devices and make sure they are not vulnerable and pose any threat. Machine learning-based threat detection systems can be used in the organisations’ network to keep an eye on the devices connected to the same network. And every time any loophole or vulnerability is found, it would alert the cybersecurity team to take immediate action against the device. Outlook The world is becoming more and more connected with every passing year, and along with all the technological advancements, more cyber threats are coming into the scenario. Technologies like machine learning and artificial intelligence have proved time and again their worth in this tech-driven era. So, if your organisation is not leveraging these sought after techs then the chances are really high that you are not only lagging behind in terms of business but also in terms of cybersecurity.","excerpt":"This ever-evolving tech-driven world is witnessing one of the most dangerous eras of cyber attacks. With this skyrocketing number of data breaches and hacks, it has become imperative for organisations across the world to adopt the new waves of technological advancements. Talking about advanced techs, artificial intelligence (AI) and machine learning (ML) definitely tops the […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Cybersecurity","latest technological advancements","Machine Learning","Network security"],"author_name":"Harshajit Sarmah","publish_date":"2019-05-23T12:26:23","publication_year":"2019","word_count":722,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","ML","Machine Learning","RAG","Network security","ViT","latest technological advancements","GAN","Cybersecurity","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","RAG","R","Go","GAN","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/4-ways-ml-can-strengthen-an-organisations-security-infrastructure\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10097042,"title":"WormGPT is a Warning for Enterprises to Upskill Their Employees","content":"Ever since LLMs entered the mainstream, concerns have been raised over their capabilities to create large amounts of written content quickly and easily. Now, these concerns have come to fruit, as the black hat hacker community has finally tapped into the capabilities of LLMs for malicious attacks. Reports have emerged that hackers have released a GPT-J powered hacking tool known as WormGPT. This tool capitalises on the already pervasive attack vector known as business email compromise, which is infamous for being one of the world’s top cyber threats. Delving deeper into the effects of WormGPT only sheds further light on the urgent need for AI cybersecurity training. Hackers are getting their hands on more capable technology with the AI wave, putting the onus on companies to inform their workforce on the potential dangers of using AI. WormGPT explained Business email compromise, or BEC, is one of the most widely-used attack vectors for hackers to spread malicious payloads. In this method, hackers impersonate a party in business with a company to execute a scam. While these emails are usually flagged as spam or suspicious by email providers, WormGPT gives fraudsters a new set of tools. By creating a new model trained on a vast array of data sources” of malware-related data models including the open-source GPT-J, attackers are able to craft a convincing fake email to sell the act of impersonation. According to a post on a commonly used hacker forum, WormGPT does not have any limitations like ChatGPT. It can generate text for a variety of black hat applications; a hacker term referring to illegal cyber activities. The model can also be run locally, leaving no trace on any servers as it would with an API. With the removal of safety rails from the model, the output that it can create is not regulated by any alignment method, offering an uncensored output ready for use in illegal actities. The main issue with an application like WormGPT is the fact that it provides the ability to create clean copy to attackers whose first language is not English. Moreover, these emails also have a better chance of passing through spam filters, as they can be customised depending on the attackers’ requirements. WormGPT greatly lowers the barrier for entry for hackers because it’s as easy to use as ChatGPT with none of the protections. Moreover, emails generated with this model also convey a professional tool, possibly increasing their efficacy in carrying out an attack. For those hackers that are too cheap to pay for WormGPT, the aforementioned forum has multiple ChatGPT jailbreaks to help users extract malicious output from the consumer bot. AIM has covered the security issues of jailbreaks extensively in the past, but custom-trained models represent a new level of AI-powered attacks. Coping up with AI attacks As mentioned previously, BEC is one of the biggest cyberattack avenues. In 2022 alone, the FBI received over 21,000 BEC complaints, which totalled to losses of about $2.7 billion. What’s more, 2021 was the 7th year in a row that BEC was the top cyber threat for enterprises. Companies also suffer from leakage of sensitive information through BEC, which can further open up the possibility of attacks. WormGPT isn’t the only way generative AI is causing problems for companies either. LLMs can be used to write malware automatically, carry out social engineering attacks, find vulnerabilities in software code, and even help in cracking passwords. Generative AI poses a threat to the enterprise as well, especially in terms of data leakage. Generative AI has also seen a slow uptake by companies due to a lack of security infrastructure around this powerful technology. While cloud service providers have begun entering the burgeoning AI market, companies are still in need of a strong, security-first LLM offering. By educating the workforce on the dangers of generative AI, companies can protect themselves from data leakage. The dangers of AI-powered hack attacks must also be emphasized, so as to enable employees to spot potential cyberattacks. Companies are falling behind on cybersecurity readiness, as evidenced by this survey which found that only 15% of organisations are deemed to have a mature level of preparedness for security risks. With the rise of generative AI, companies need to pour resources into keeping their workforces up to date with the latest threats in AI-powered cybersecurity.","excerpt":"Reports have emerged that hackers have released a GPT-J powered hacking tool known as WormGPT, making it the first ChatGPT for hackers.","categories":["AI Features"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-07-17T16:00:00","publication_year":"2023","word_count":721,"keywords":["ChatGPT","API","TPU","AI","GPT","Ray","Aim","generative AI","GAN","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","Aim","Ray","TPU","R","API","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/wormgpt-is-a-warning-for-enterprises-to-upskill-their-employees\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10062472,"title":"Google AI introduces Shift-Robust GNNs to improve generalisability in models","content":"Google AI research scientist Bryan Perozzi and research intern Qi Zhu have released a study that offers an answer for using Graph Neural Networks on biased datasets. Called ‘Shift-Robust GNNs: Overcoming the Limitations of Localized Graph Training Data,’ the research shows how a new method can measure the distributional differences between biased data and a graph’s true inference distribution. The largeness of the shift between two probability distributions also helps account for the amount of bias. Machine learning models face issues with generalising from the biased training data as the shift grows wider. This affects the generalisability of a model adversely. When a dataset’s accuracy was tested using a statistical analysis called the F-score, it was found that domain shifts led to a decrease in performance by 15-20 per cent. As the distribution shift increases, the classification accuracy of the dataset decreases, which makes it hard for GNNs to generalise models as the difference between the training dataset and test dataset also increases. The study states that the distribution shift between training data and unlabeled data was reduced using a shift-robust regulariser. This technique assesses the domain shift during the training and then penalises it. This forces the model to ignore the training bias as far as it can. Results from the research showed that SR-GNN was more effective on biased training datasets and beat GNN outcomes in terms of accuracy. It also reduced the negative impact caused by biased data by 30 per cent to 40 per cent.","excerpt":"The study states that the distribution shift between training data and unlabeled data was reduced using a shift-robust regulariser.","categories":["AI News"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-03-09T18:03:01","publication_year":"2022","word_count":249,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","programming_languages:Go","AI research","R"],"extracted_tech_keywords":["AI","machine learning","neural network","R","Go","AI research","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-ai-introduces-shift-robust-gnns-to-improve-generalisability-in-models\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10079399,"title":"Top 8 Humanoid Robots Made in India","content":"Humanoid robots have been the talk of the town, spurred by the latest unveiling of Optimus at Tesla’s AI Day. Since then, discussions have been brewing over the future of human interactions and customer assistance at workplaces. According to a report, the humanoid robot market is expected to register a CAGR of 63.5% by 2027. India has been ahead in the robotics race, making it to the top 15 list of the highest number of recorded robots installations. Let’s take a look at some of the most interesting pieces of humanoid technology made in India: 1. Manav – First 3D Printed Humanoid Robot Manav, India’s first 3D printed humanoid robot, was first unveiled at IIT Mumbai Tech Fest. About two feet tall, the robot is able to perform tasks such as walking, talking and dancing by listening to voice commands. It was built with an open source code so that it can be trained in real-time to learn and respond, similar to a human child. The machinery in it allowed two degrees of freedom in the head and neck—which means that the robot could move its head up and down, and sideways. Additionally, the robotic system had Wi-Fi and Bluetooth connectivity, and a rechargeable lithium polymer battery in it. Manav was made available to many engineering and research institutes teaching robotics as a subject, as the maker believed it was meant mainly for research purposes. 2. IRA Robot by HDFC Bank IRA, launched by HDFC Bank Ltd, was placed near the welcome desk at a branch in Mumbai. The first phase of the project involved the robot greeting customers entering the bank, and showing them a list of the services offered by the bank. Customers could select an option and the robot would guide them to the relevant counters in the branch. However, HDFC Bank Ltd. also came up with the second installment of the robot shortly after. IRA 2.0, which was launched in Bangalore, not only guides customers through the branch, but also interacts with them, and answers certain bank-related queries, and frequently asked questions (FAQs). 3. Mitra Robot for Customer Service Mitra, a customer-service robot, was developed by Bangalore-based Invento Robotics. It is famously known for greeting Ivanka Trump and Prime Minister Narendra Modi at the GES 2017 summit, and declaring the event open. The robot rose to prominence during the Covid-19 pandemic, where it was deployed in hospitals as a point of contact between patients and their loved ones. Patients could connect with them through a screen attached to its chest. Like an assistant to nurses and doctors, Mitra is capable of taking their readings and vitals, and giving timely reminders to patients to take their medication. 4. RADA by Vistara Airlines Built on a four-wheel chassis with 360-degrees rotation, RADA is a customer-assistance robot for Vistara Airlines, deployed at the Indira Gandhi International airport, Delhi. The robot was designed to serve Vistara’s customers by responding to their queries, and providing several entertainment options like games, songs, and videos. Additionally, RADA is also capable of greeting the flyers, scanning their boarding passes, and providing necessary details such as the terminal, departure gates, and weather conditions of the destination city, among other flight-related information. 5. RoboCop Stationed at Kerala Police Headquarter RoboCop is a Sub Inspector (SI)-ranked robot, stationed at the Kerala police headquarters, in Thiruvananthapuram. The robot was tasked with managing the front office at the headquarters. The visitors could directly interact with the KP-robot, receive directions on where they need to head, as and when necessary. RoboCop was also able to perform certain additional features such as fixing meetings, issuing identity cards, or opening new files based on the concerns of the public. Thus, it was equipped with recording the details of the visitors and their respective clients. 6. INDRO By Santosh Hulawale INDRO is the tallest humanoid robot made in India. Created by Santosh Hulawale, it was built with low-cost materials inside a house, and can be used for household chores, entertainment and education-related activities. The robot can carry up to 150 kg of payload on the platform placed below its knees. It is also capable of lifting up to 2 kilos of weight with its hands and has the ability to perform all ‘human-like’ actions. INDRO is open source, allowing its users to program the bot according to their requirements. The latest installment of the robot—6 feet in height—is developed on the Artificial Intelligence and Machine Learning platform, and has the ability to perform multiple tasks with high precision and speed. It also is equipped with a face recognition software, which enables it to recognise people it has interacted with earlier. 7. KEMPA Robot by Sirena Technologies KEMPA, also a customer-assistance robot, was built by Bangalore-based Sirena Technologies. The bot, which was deployed at the Kempegowda International Airport, would help passengers with their flight-related queries, like providing information about flights, check-in details, scanning luggage, and handing out boarding passes. KEMPA was also capable of sharing information about Bangalore—its culture and heritage, and suggesting some tourist places to visit in Bangalore. Or, the bot could also simply entertain you with a conversation. 8. AcYut – First Indigenous Robot AcYut, known to be India’s first indigenous robot, was developed by undergraduate students at the Birla Institute of Technology and Science (BITS), Pilani, India. The project had quite a few sponsors, most notably Govt. of India (DEITY) and BITSAA. At numerous international venues, such as RoboCup, RoboGames, CMU, Stanford, etc., AcYut has represented India. AcYut is also consistently the sole team representing India in the highly technological Humanoid Teen Sized Soccer Leagues at Robocup, where robots compete in autonomous soccer matches.","excerpt":"According to a report, the humanoid robot market is expected to register a CAGR of 63.5% between 2022 and 2027","categories":["AI Trends"],"tags":["Humanoid Robots"],"author_name":"Ayush Jain","publish_date":"2022-11-11T12:00:00","publication_year":"2022","word_count":944,"keywords":["Go","artificial intelligence","machine learning","Humanoid Robots","AI","programming_languages:R","programming_languages:Go","ViT","ai_applications:robotics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-humanoid-robots-made-in-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10016800,"title":"Guide To GluonTS and PytorchTS For Time-Series Forecasting (With Python Implementation)","content":"GluonTS is a toolkit that is specifically designed for probabilistic time series modeling, It is a subpart of the Gluon organization, Gluon is an open-source deep-learning interface that allows developers to build neural nets without compromising performance and efficiency. AWS and Microsoft first introduced it on October 12th, 2017 that provides many different neural network architectures and leverages the deep learning models. It combines many packages into one like mxnet– a lightweight, portable, flexible Distributed\/Mobile Deep learning model; for Python, R, Julia, Scala. Go, Javascript, and more. Gluon’s goal is to leverage Jupyter notebooks’ strengths to present graphics, equations, and code together in one place. Table of contentsGluonTSInstallationGetting Started Let’s take an example dataset to forecast time series dataset using GluonTSTrainingEvaluateOutputPyTorch-tsInstallationGetting StartedConclusion GluonTS “GluonTS simplifies the development of and experimentation with time series models for common tasks such as forecasting or anomaly detection. It provides all necessary components and tools that scientists need for quickly building new models, for efficiently running and analyzing experiments and for evaluating model accuracy.”— GluonTS Arxiv Research paper GluonTS is a toolkit that is specifically designed for probabilistic time series modeling, GlounTS provides the utilities for loading and looping over time-series datasets. It also provides state of the art models for time series forecasting, and a building block to define your own models and quickly test with different solutions, GluonTS has many features like you can: Train and evaluate any inbuilt models on your custom datasetQuickly create your solution using GluonTSIt provides custom abstractions and building blocks to create custom models.Provide multiple baseline algorithms for comparison.Plotting and evaluation facilitiesArtificial and real datasets. Installation pip install gluonts # as gluonts relies on mxnet # install MXnet using pip pip install mxnet Getting Started We have seen time series forecasting using TensorFlow and PyTorch, but they come with a lot of code and require great proficiency over the framework. GluonTS provide simple and on point code for running your time series forecasting here is an example code to run GluonTS for predicting Twitter volume with DeepAR. You can run the following code in a cloud development environment at: https:\/\/github.com\/mmaithani\/data-science\/blob\/main\/Gluonts_twitter_volume_forecasting.ipynb #importing gluonTS utilities and pandas from gluonts.dataset import common from gluonts.model import deepar from gluonts.trainer import Trainer import pandas as pd #getting train datatset of twitter volume url = \"https:\/\/raw.githubusercontent.com\/numenta\/NAB\/master\/data\/realTweets\/Twitter_volume_AMZN.csv\" df = pd.read_csv(url, header=0, index_col=0) data = common.ListDataset([{ \"start\": df.index[0], \"target\": df.value[:\"2015-04-05 00:00:00\"] }], freq=\"5min\") #initializing trainers and deepAR estimators trainer = Trainer(epochs=10) estimator = deepar.DeepAREstimator( freq=\"5min\", prediction_length=12, trainer=trainer) predictor = estimator.train(training_data=data) prediction = next(predictor.predict(data)) print(prediction.mean) prediction.plot(output_file='graph.png') Let’s take an example dataset to forecast time series dataset using GluonTS #importing modules %matplotlib inline import mxnet as mx from mxnet import gluon import numpy as np import pandas as pd import matplotlib.pyplot as plt import json Importing inbuilt GluonTS datasets from gluonts.dataset.repository.datasets import get_dataset, dataset_recipes from gluonts.dataset.util import to_pandas print(f\"Available datasets: {list(dataset_recipes.keys())}\") We are going to use the m4_hourly dataset, now the datasets provided by gluonts are objects that consist of three main attributes we will see later. dataset = get_dataset(\"m4_hourly\", regenerate=True) Three main attributes of GluonTs dataset are dataset.train is a training dataset, dataset.test is a testing dataset, and dataset.metadata contains metadata of the dataset. Let’s plot and iterate the dataset using the following command. entry = next(iter(dataset.train)) train_series = to_pandas(entry) train_series.plot() plt.grid(which=\"both\") plt.legend([\"train series\"], loc=\"upper left\") plt.show() Similarly, you can plot test dataset entry = next(iter(dataset.test)) test_series = to_pandas(entry) test_series.plot() plt.axvline(train_series.index[-1], color='r') # end of train dataset plt.grid(which=\"both\") plt.legend([\"test series\", \"end of train series\"], loc=\"upper left\") plt.show() Preprocessing N = 10 # number of time series T = 100 # number of timesteps prediction_length = 24 freq = \"1H\" custom_dataset = np.random.normal(size=(N, T)) start = pd.Timestamp(\"01-01-2019\", freq=freq) # can be different for each time series Now for splitting dataset and converting it to gluonts format use following commands: from gluonts.dataset.common import ListDataset # train dataset: cut \"prediction_length\", add \"target\" and \"start\" fields train_ds = ListDataset([{'target': x, 'start': start} for x in custom_dataset[:, :-prediction_length]], freq=freq) # test dataset: using whole dataset, add \"target\" and \"start\" fields test_ds = ListDataset([{'target': x, 'start': start} for x in custom_dataset], freq=freq) Training GlounTS comes with its own hyper parameters and feedforward neural network like SimpleFeedForwardEstimator that accepts an input window of length context_length and predicts the distribution of the value. Let;s import necessary training methods and assign estimators values using following commands: from gluonts.model.simple_feedforward import SimpleFeedForwardEstimator from gluonts.trainer import Trainer estimator = SimpleFeedForwardEstimator( num_hidden_dimensions=[10], prediction_length=dataset.metadata.prediction_length, context_length=100, freq=dataset.metadata.freq, trainer=Trainer(ctx=\"cpu\", epochs=5, learning_rate=1e-3, num_batches_per_epoch=100 ) ) #start training predictor = estimator.train(dataset.train) Evaluate For evaluating the models we have further make_evaluation_predictions function that automates the process of prediction and model evaluation. from gluonts.evaluation.backtest import make_evaluation_predictions forecast_it, ts_it = make_evaluation_predictions( dataset=dataset.test, # test dataset predictor=predictor, # predictor num_samples=100, # number of sample paths for evaluation ) Convert the generators to list to ease the computations and examine the first element of these lists: forecasts = list(forecast_it) tss = list(ts_it) # first entry of the time series list ts_entry = tss[0] Convert the first five value of time-series from pandas to NumPy and initialize first entry of dataset.test np.array(ts_entry[:5]).reshape(-1,) dataset_test_entry = next(iter(dataset.test)) Similarly first 5 values and forecast entries dataset_test_entry['target'][:5] forecast_entry = forecasts[0] Output For visualizing the outputs use following commands: def plot_prob_forecasts(ts_entry, forecast_entry): plot_length = 150 prediction_intervals = (50.0, 90.0) legend = [\"observations\", \"median prediction\"] + [f\"{k}% prediction interval\" for k in prediction_intervals][::-1] fig, ax = plt.subplots(1, 1, figsize=(10, 7)) ts_entry[-plot_length:].plot(ax=ax) # plot the time series forecast_entry.plot(prediction_intervals=prediction_intervals, color='g') plt.grid(which=\"both\") plt.legend(legend, loc=\"upper left\") plt.show() plot_prob_forecasts(ts_entry, forecast_entry) You can also evaluate the quality of time series forecast using evaluator class, that can compute the aggregate performance metrics, from gluonts.evaluation import Evaluator evaluator = Evaluator(quantiles=[0.1, 0.5, 0.9]) agg_metrics, item_metrics = evaluator(iter(tss), iter(forecasts), num_series=len(dataset.test)) print(json.dumps(agg_metrics, indent=4)) PyTorch-ts You can achieve similar results using a third party framework called PyTorch-ts, built by Zalando Research, that is specifically designed for PyTorch enthusiasts, Pytorch-ts is probabilistic Time Series forecasting framework based on GluonTS backend and its installation and usage are pretty easy, you can find the source code here, There very minimal changes in Pytorch-ts as it used the Pytorch time series model by utilizing GluonTS as its API for loading dataset, transforming and testing. Installation $ pip install pytorchts Getting Started We are going to use a dataset volume of tweets mentioning the AMZN ticker symbol, to leverage the power of this model, first import the necessary packages using below commands, Notebook is available at: https:\/\/github.com\/mmaithani\/data-science\/blob\/main\/PyTorch_ts_time_series_forecasting(gluonts).ipynb import pandas as pd import torch import matplotlib.pyplot as plt from pts.dataset import ListDataset from pts.model.deepar import DeepAREstimator from pts import Trainer from pts.dataset import to_pandas Import the Amazon tweets dataset: and plot the first 100 data points using pandas and plt. url = \"https:\/\/raw.githubusercontent.com\/numenta\/NAB\/master\/data\/realTweets\/Twitter_volume_AMZN.csv\" df = pd.read_csv(url, header=0, index_col=0, parse_dates=True) df[:100].plot(linewidth=2) plt.grid(which='both') plt.show(} Train the dataset using Pytorch-ts, we are using the data up to midnight on April 5th, 2015, and a 5mins data so req is set to 5min with 15epochs and we are expecting prediction for next hour, so prediction_length is set to 12. training_data = ListDataset( [{\"start\": df.index[0], \"target\": df.value[:\"2015-04-05 00:00:00\"]}], freq = \"5min\") # parameter initialization device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\") estimator = DeepAREstimator(freq=\"5min\", prediction_length=12, input_size=43, trainer=Trainer(epochs=15, device=device)) predictor = estimator.train(training_data=training_data) Model is trained so lets forecast the hour following the midnight on 15-04-2015: test_data = ListDataset( [{\"start\": df.index[0], \"target\": df.value[:\"2015-04-15 00:00:00\"]}], freq = \"5min\") for test_entry, forecast in zip(test_data, predictor.predict(test_data)): to_pandas(test_entry)[-60:].plot(linewidth=2) forecast.plot(color='b', prediction_intervals=[50.0, 90.0]) plt.grid(which='both') Conclusion We have discussed time series forecasting using GluonTS a forecasting library explicitly made for probabilistic time series problems and the outputs were quite satisfactory. We saw the same approach using PytorchTs (PyTorch-based time series framework backed by Gluon) also the Gluon integrates many other features. There are third party libraries are also been made on top of GluonTS that we are not discussing in this article like pytorch-ts which is a PyTorch-based Probabilistic time series forecasting model based on GluonTS backend. Working notebooks and other resources used in the above demonstration: GluonTs  notebook  PyTorch-ts notebookGluonTs on M4 dataset notebookGluonTS Research Paper","excerpt":"GluonTS is a toolkit that is specifically designed for probabilistic time series modeling, It is a subpart of the Gluon organization, Gluon is an open-source deep-learning interface that allows developers to build neural nets without compromising performance and efficiency. AWS and Microsoft first introduced it on October 12th, 2017 that provides many different neural network […]","categories":["Deep Tech"],"tags":["Matplotlib","Time Series","Time Series Forecasting","time series prediction"],"author_name":"Mohit Maithani","publish_date":"2020-12-30T14:00:00","publication_year":"2020","word_count":1328,"keywords":["NumPy","AI","neural network","TensorFlow","PyTorch","Time Series Forecasting","Ray","deep learning","Time Series","Jupyter","time series prediction","Matplotlib","Pandas"],"extracted_tech_keywords":["AI","deep learning","neural network","Ray","TensorFlow","PyTorch","Jupyter","Pandas","NumPy","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/gluonts-pytorchts-for-time-series-forecasting\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10110972,"title":"Stability AI Unveils Stable Code 3B, Enhances New Coding Capabilities","content":"Stability AI has introduced Stable Code 3B, a pre-trained 2.7 billion parameter decoder-only language model on a staggering 1.3 trillion tokens, encompassing a rich tapestry of textual and code datasets. One of the standout features of Stable Code 3B is its Fill in the Middle (FIM) capability. Unlike traditional code completion models that suggest single lines, Stable Code 3B can now seamlessly complete larger missing sections of code. This breakthrough opens up new possibilities for developers, allowing them to bridge gaps in their codebase and enhance productivity effortlessly. Check out code and datasets here. Drawing inspiration from the 2023 StackOverflow Developer Survey, the model has been trained in 18 programming languages, including popular choices like Python, Java, JavaScript, and C++. This breadth of language coverage positions Stable Code 3B as a versatile solution catering to the needs of a vast developer community. Stability AI releases Stable Code 3B on Hugging Facemodel: https:\/\/t.co\/N6PWR7EJUkstable-code-3b is a 2.7B billion parameter decoder-only language model pre-trained on 1.3 trillion tokens of diverse textual and code datasets pic.twitter.com\/3Uv5ajj07R— AK (@_akhaliq) January 17, 2024 With a colossal dataset comprising 1.3 trillion tokens sourced from GitHub repositories, books, and websites, Stable Code 3B underwent a rigorous training regimen. Masked language modelling, a technique where the model predicts missing words within code snippets, played a crucial role in shaping the decoder-only language model. More about the model The model showcases its performance, especially compared to similar models. For instance, it distinguishes itself among coding language models through its performance, achieving comparable accuracy to larger models like CodeLLaMA  on MultiPL-E metrics across various programming languages. Stable Code 3B incorporates RoPE (Recurrent Processing Elements), expanding its context size to 100,000 tokens. This enhancement ensures a deeper understanding of lengthy code sequences, enabling the model to grasp intricate relationships within extensive codebases. The result is a code completion tool that doesn’t just understand lines of code but comprehensively interprets the nuances of entire programs. Despite its smaller size, Stable Code 3B leverages the foundation of Stable LM 3B, inheriting a wealth of general language understanding. Additionally, it excels at code translation, seamlessly bridging gaps between different programming languages. The model’s prowess extends to answering code-related questions, providing valuable insights into functionality and bug-fixing. Validation comes from MultiPL-E metrics, and the model has proven its mettle across various programming languages, as tested using BigCode’s Evaluation Harness. Its applications include code generation, which can craft anything from functions to entire programs. The model’s open-source availability on platforms like Hugging Face promotes transparency and community collaboration, fostering ongoing improvements and innovations. Subscribers gain access to a suite of AI tools, including SDXL, StableLM Zephyr, Stable Audio, and Stable Video. Its notable efficiency allows it to run effectively on standard laptops without GPUs, enhancing accessibility for a broader user base.","excerpt":"Stable Code 3B incorporates RoPE, expanding its context size to 100,000 tokens.","categories":["AI News"],"tags":["Stability AI"],"author_name":"Sandhra Jayan","publish_date":"2024-01-17T17:55:10","publication_year":"2024","word_count":464,"keywords":["Go","Hugging Face","AI","ML","RAG","Python","C++","JavaScript","R","Java","Stability AI"],"extracted_tech_keywords":["AI","ML","Hugging Face","RAG","Python","R","JavaScript","Go","Java","C++"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/stability-ai-unveils-stable-code-3b\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10118844,"title":"Sakana AI Releases Japanese DALLE-3, Calls it  EvoSDXL-JP","content":"Japanese AI startup, Sakana AI introduced EvoSDXL-JP, an image generation model built via Evolutionary Model Merge, which delivers 10x faster image generation for Japanese styles. EvoSDXL-JP, is now publicly available on the HuggingFace platform for research and educational purposes, accompanied by an accessible demo for immediate testing. The model that can support Japanese and generate Japanese style images by fusing different open models. Compared to the existing Japanese model, the inference speed is 10 times faster, but it shows better performance in the benchmark, said the company in its blog post. EvoSDXL-JP is capable of high-speed and low-cost image generation, and is the best model to easily try and experience generative AI. The company said it expects it to be used in educational sites in Japan so that more people can enjoy the benefits of generative AI. Sakana AI recently introduced an innovative model construction approach using evolutionary algorithms called “Evolutionary Model Merge.” The company says Evolutionary model merge is not limited to specific modalities, and can be applied to models of any modality in principle. Furthermore, the company has  released the EvoLLM-JP, a large-scale Japanese language model, and the EvoVLM-JP, an image language model, both constructed through Evolutionary Model Merge. These models were based on self-regressive Transformer models designed for language generation. EvoLLM-JP, was made by merging the large-scale language model (LLM) of Japanese and the LLM of mathematics, and was found to be good not only in mathematics but also in the overall ability of Japanese. In addition, EvoVLM-JP, which was made by merging Japanese LLM and image language model (VLM), can respond to knowledge of Japanese culture and achieved the best results in benchmarks using Japanese images and Japanese text.","excerpt":"The model delivers 10x faster image generation for Japanese styles.","categories":["AI News"],"tags":["sakana ai"],"author_name":"Siddharth Jindal","publish_date":"2024-04-22T18:43:04","publication_year":"2024","word_count":283,"keywords":["Go","programming_languages:R","AI","Modal","programming_languages:Go","generative AI","sakana ai","R","startup"],"extracted_tech_keywords":["AI","generative AI","R","Go","startup","Modal","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/sakana-ai-releases-japanese-dalle-3-calls-it-evosdxl-jp\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":29945,"title":"Hyundai Invests In Israeli Computer Vision Startup Allegro.AI","content":"HyundaiCRADLE, Hyundai Motor Company’s corporate venturing and open innovation business announced their strategic investment in Allegro.ai, a leading technology company specialising in deep learning-based computer vision. This partnership will allow Hyundai to speed up and deploy their artificial intelligence-related technologies in many business areas. This will improve the quality of Hyundai’s products, thereby increasing customer satisfaction while creating a safer driving environment. Founded in 2016, Allegro.ai offers the first end-to-end DL lifecycle management solution focused on deep learning as it applies to computer vision. The company’s platform simplifies the process of developing and managing deep learning-powered solutions – such as autonomous vehicles, drones, security, logistics and others. “Deep learning computer vision is one of the core technologies that can be applied to autonomous driving to navigate roads and make quick decisions in real-time – and allegro.ai is clearly an innovation leader in that field,” said Ruby Chen, head of investment at Hyundai CRADLE, Tel Aviv. He added, “Our investment in Allegro.ai is a further step in enhancing our presence in the Israeli market, a global leader of technological innovation in the fields of automation, artificial intelligence and deep learning. This is our fifth investment in an Israeli company and our activities will continue to grow the coming year”. “We are proud to partner with Hyundai and share Hyundai’s belief that AI empowers the industry to provide greater road safety, autonomy, to better understand customers’ needs and to help broaden their experiences,” said Nir Bar-lev, CEO and Co-founder of Allegro.ai. “Whether a company is developing autonomous vehicles, drones, security, or other types of applications, allegro.ai makes it easy for them to manage their data-sets and build deep learning-based solutions while guaranteeing complete and confidential control of their data.”","excerpt":"HyundaiCRADLE, Hyundai Motor Company’s corporate venturing and open innovation business announced their strategic investment in Allegro.ai, a leading technology company specialising in deep learning-based computer vision. This partnership will allow Hyundai to speed up and deploy their artificial intelligence-related technologies in many business areas. This will improve the quality of Hyundai’s products, thereby increasing customer […]","categories":["AI News"],"tags":["Computer Vision","computer vision autonomous vehicles","Deep Learning","hyundai","israel"],"author_name":"Prajakta Hebbar","publish_date":"2018-11-06T05:01:09","publication_year":"2018","word_count":287,"keywords":["artificial intelligence","programming_languages:R","AI","innovation","computer vision","automation","israel","ai_applications:computer vision","deep learning","ViT","Computer Vision","Deep Learning","computer vision autonomous vehicles","R","hyundai"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","computer vision","R","ViT","automation","innovation","programming_languages:R","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hyundai-invests-in-allegro-ai-dl-computer-vision\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011784,"title":"Guide To SuperAnnotate &#8211; The Most Robust Image and Video Annotator Tool","content":"Annotations are extra information attached to parts of any kind of media, for better understanding that portion. In case of image annotations its labels, segmentation, localization, bounding boxes. For supervised machine learning, image annotation provides labels to objects in the image. Automatic Image Annotation is the new advancement in Computer Vision; it will automatically provide metadata related to the images. It is mainly used for image retrieval(searching through large databases for showing results of exact images for that text). This is better than content-based image retrieval(CBIR) process, which is query-based and requires more time to execute. Most image annotation tools go by bounding box methodology which is the rectangle encompassing the object and giving four coordinates(left top corner, right top corner, left down right down corner) that are used by the algorithms to locate the exact object in the image. Most object detection algorithms(R-CNN, YOLO, Faster-RCnn, etc.) are built along with this methodology. But there are certain limitations to bounding box methodology – often objects are intruded by noise in the bounding box and thus fail to detect accurately by the detection algorithm. Another limitation is that apart from objects to be identified other objects are also to be annotated in some instances such as for self-driving cars, apart from cars there’s pedestrians, poles, signals and many more. These limitations led to the development of better data annotator tools which can provide better accuracy, speed and precision in data annotations and thus today we’ll be discussing SuperAnnotate. What is SuperAnnotate? SuperAnnotate is an image and video annotator automation tool developed by Vahan Petrosyan. Vahan and his brother Tigran are PhD drop out students who built this web-based application after realising the lack of efficient supervised image segmentation algorithms in their research study. In February 2020 they released this toolkit to benefit the computer vision community. SuperAnnotate platform provides end to end service for automating computer vision projects, starting from data engineering(generating high-quality training data) to model creation(training using neural networks). Allows project management through team creation and share via an API through Python SDK to measure progress. Images can be uploaded through a local device or AWS S3 folder. One of the eminent features is that it provides annotation automation for predefined classes. Mislabelled annotations can be detected using QA automation technique. Users can make use of transfer learning to classify new data. Some efficient analytics tools are present for tracking performance speed and accuracy in the annotation. The annotated data can also be downloaded in SuperAnnotate or COCO format. Walkthrough to SuperAnnotate working SuperAnnotate works with pixel-accurate annotations. This could be a tedious job, but with the evolution of deep learning algorithms, pixel-level accuracy has been reached. This kind of annotation process works with pixel precision, and apart from the bounding box, it has various others-  Polyline, Ellipse, Polygon, Cuboid, point, Template formats. All of those annotations formats will have the information stored in a JSON file. SuperAnnotate Desktop This is a venture by OpenCV and SuperAnnotate to provide a desktop application based on the web app present. It is free software available in Windows, Mac and Linux OS. It has accelerated the labelling process with better accuracy. SuperAnnotate provides both manual and automatic image segmentation. Video segmentation works the same as image segmentation by breaking the video into frames and tracking multiple objects. SuperAnnotate can serve training cycles without any compromise to annotated data quality. Provides filtering of images- to allow only display images users need to work with and avoid the whole lot. The application has an advanced polygon tool to monitor minute edges of objects. For more information on SuperAnnotate desktop, usages refer to this Github Repository. Python SDK The latest version of SuperAnnotate SDK version 2.3.8 was released yesterday. The package is pip installable – pip install superannotate Then create a project in the IDE or notebook to use it: import superannotate as sa sa.create_project(\"Example Project 1\", \"example\", \"Vector\") sa.upload_images_from_folder_to_project(\"Example Project 1\", \"<path_to_my_images_folder>\") To create an annotation class: sa.create_annotation_class(project, \"Large car\", color=\"#FFFFAA\") Or sa.create_annotation_classes_from_classes_json(project, \"<path_to_classes_json>\") To download annotations: sa.download_annotation_classes_json(project, \"<path_to_local_folder>\") To add annotations to the downloaded image sa.add_annotation_bbox_to_json(\"<path_to_json>\", [10, 10, 100, 100], \"Human\") To upload your project to platform sa.upload_annotations_from_json_to_image(project, image, \"<path_to_json>\") To create pandas dataframe from the project for annotations df = sa.aggregate_annotations_as_df(\"<path_to_project_folder>\") To visualize the class distributions: df = sa.class_distribution(\"<path_to_export_folder>\", [project_names], visualize = True) Solutions and Services Autonomous VehiclesDrone\/Satellite imagerySecurity and SurveillanceMedical imagingRoboticsAgricultureInvoice and Financial dataFashionSports AnalyticsInfrastructure InspectionRetail automationInsurance There are limitations to free services, and most high-end features are available in paid versions which have different pricing, and users can contact SuperAnnotate team for a demo and other purposes. However, it has a free academic license provided to academic and non-commercial researchers. Companies using SuperAnnotate Percepto reported reducing annotation time by 60%Cogito TechAcme AIAltris AI reported an increase in 12% accuracy. Conclusion Easy to use by annotators, researchers and computer vision engineers. Its user-friendly interface allows even beginners to easy access the high-end features. Services have an extension to LiDar, text and audio apart from image and video. The platform also provides hiring to 3rd party annotators who can do the work and save time. Superannotate blogs, tutorials and documents are very efficiently maintained. Every month active development is being made.","excerpt":"the development of better data annotator tools which can provide better accuracy, speed and precision in data annotations and thus today we’ll be discussing SuperAnnotate.","categories":["Deep Tech"],"tags":["annotation","data annotation","image segmentation","object store database","OpenCV"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-11-18T12:00:32","publication_year":"2020","word_count":869,"keywords":["machine learning","image segmentation","AI","neural network","AWS","annotation","computer vision","OpenCV","data annotation","deep learning","analytics","object detection","object store database","Pandas"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","analytics","OpenCV","Pandas","object detection","AWS"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/superannotate\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10005782,"title":"Complete Tutorial on Pygal: A Python Tool for Interactive and Scalable visualization","content":"Data visualization is a scientific method of finding out the patterns, anomalies, and outliers in the data. It is a method that is used to create the graphs and plots which helps in finding out the trend in the data because it makes data simpler in terms of understanding by the human brain and we identify trends with naked eyes. Python provides different visualization packages that help in creating different types of charts, graphs, and plots. Pygal is an open-source python library that not only creates highly interactive plots but also creates SVG images of the graphs\/plots so that we can use it and customize it accordingly. Pygal is highly customizable and creates graphs with a few lines of code. Pygal creates a variety of plots like a bar, area, histogram, line, etc and due to the high scalability of the images of the plot downloaded as SVG, we will have good quality images that can be embedded into different projects, websites, etc. In this article, we will go through different charts that we can create using pygal and also see how we can download them in the SVG format. Implementation: We will start exploring pygal but before that, we need to install pygal using pip install pygal. Importing required libraries We will be importing pygal for visualization and pandas for importing the dataset for which we will be creating the visualizations. import pandas as pd import pygal Loading the Dataset We will be working with an Advertisement dataset that contains different attributes like ‘TV’, ‘Radio’, etc. and a target variable ‘Sales’. df = pd.read_csv(‘Advertising.csv’) df.head() Creating Bars and Graphs Now we will use pygal to create different bars and graphs using the dataset we have loaded. We will start by creating some basic charts and then move to some advance charts and downloading all the graphs and charts in SVG formats. Bar Charts Bar charts represent the data in the form of horizontal or vertical bars with the values they represent. Pygal supports different types of Bar charts. bar_graphs = pygal.Bar() bar_graphs.add('Sales', df['Sales'][:5]) bar_graphs.render_to_file('bar1.svg') bar_graphs.add('Sales', df['Sales'][:5]) bar_graphs.add('TV', df['TV'][:5]) bar_graphs.render_to_file('bar2.svg') Line charts A line chart displays the data or information as a series of data points on a line. line_chart = pygal.Line() line_chart.add('Sales', df['Sales'][:15]) line_chart.render_to_file('line1.svg') line_chart.add('Sales', df['Sales'][:15]) line_chart.add('TV', df['TV'][:15]) line_chart.add('Newspaper', df['Newspaper'][:15]) line_chart.add('Radio', df['Radio'][:15]) line_chart.render_to_file('line2.svg') Box Plots Box Plots are considered as the statistical graphs which are used to display the numerical data with their quartile information. They are generally used to visualize the outliers in the data. box_plot = pygal.Box() box_plot.title = 'Advertising' box_plot.add('TV', df['TV']) box_plot.add('Radio', df['Radio']) box_plot.add('Newspaper', df['Newspaper']) box_plot.render_to_file('box1.svg') Dot Plots A dot plot is a statistical plot that is used to display the data in forms of the dot, the size of the dot represents the values i.e higher the value bigger is the dot. dot_chart = pygal.Dot() dot_chart.title = 'Advertising' dot_chart.add('Sales', df['Sales'][:5]) dot_chart.add('TV', df['TV'][:5]) dot_chart.add('Radio', df['Radio'][:5]) dot_chart.add('Newspaper', df['Newspaper'][:5]) dot_chart.render_to_file('dot1.svg') Funnel Charts Funnel charts are used to visualize the progressive reduction or expansion of data when it passes from different phases. It is generally used to represent the Sales data. funnel_chart = pygal.Funnel() funnel_chart.title = 'Advrtising' funnel_chart.add('TV', df['TV'][:15]) funnel_chart.add('Radio', df['Radio'][:15]) funnel_chart.add('Newspaper', df['Newspaper'][:15]) funnel_chart.render_to_file('funnel1.svg') Gauge Chart A gauge chart combines the doughnut chart and pie chart, it is used to visualize the data at a particular value. gauge_chart = pygal.Gauge(human_readable=True) gauge_chart.title = 'Advertising' gauge_chart.range = [0, 500] gauge_chart.add('Sales', df['Sales'][:1]) gauge_chart.add('TV', df['TV'][:1]) gauge_chart.add('Radio', df['Radio'][:1]) gauge_chart.add('Newspaper', df['Newspaper'][:1]) gauge_chart.render_to_file('gauge.svg') Treemap Treemaps are used to represent data in nested forms usually in rectangles shape. The size of the map resent the value i.e higher the value greater is the size of the treemap. treemap = pygal.Treemap() treemap.title = 'Advertisement TreeMap' treemap.add('Sales', df['Sales'][:10]) treemap.add('TV', df['TV'][:10]) treemap.add('Radio', df['Radio'][:10]) treemap.add('Newspaper', df['Newspaper'][:10]) treemap.render_to_file('treemap1.svg') Pygal provides different types of parameters for different graphs, we can manipulate different designs, features, and render them accordingly. Conclusions: In this article, we learned about Pygal, a visualization package in python which provides highly interactive and scalable visualizations. We saw how we can create different charts and plots from basic to advanced in just a few lines of code. We downloaded the images in SVG format which makes it scalable for using it in different applications or websites.","excerpt":"Pygal creates a variety of plots like a bar, area, histogram, line, etc and due to the high scalability of the images of the plot downloaded as SVG","categories":["Deep Tech"],"tags":["ai for beginners tutorial","Data Visualisation","interactive plots","python tool","scalability","simple ai tutorial","trends in bi and data visualization","tutorial","Visualization"],"author_name":"Himanshu Sharma","publish_date":"2020-08-30T10:00:00","publication_year":"2020","word_count":695,"keywords":["trends in bi and data visualization","simple ai tutorial","python tool","Scala","Visualization","R","Pandas","scalability","programming_languages:Scala","programming_languages:Python","Go","AI","interactive plots","programming_languages:R","ai for beginners tutorial","Data Visualisation","programming_languages:Go","Python","tutorial"],"extracted_tech_keywords":["AI","Pandas","Python","R","Go","Scala","programming_languages:Python","programming_languages:R","programming_languages:Scala","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/complete-tutorial-on-pygal-a-python-tool-for-interactive-and-scalable-visualization\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10095739,"title":"Linux Kernel 6.4 Brings Better Support For Rust","content":"The everlasting Linux Kernel has gotten yet another update, and this one includes some quality of life changes, a few upgrades to the file system of the OS, and additional support for Rust code in the kernel. In addition to this, the update also kept up with housekeeping by adding drivers for the latest hardware, such as Wi-Fi drivers for various Realtek modems. The update was announced earlier today via the Linux Kernel Mailing List. The creator of Linux and one of the primary maintainers of the kernel, Linus Torvalds, dropped a hint about the next update, Linux 6.5, being a big one, stating, “Most of the stuff in my mailbox the last week has been about upcoming things for 6.5, and I already have 15 pull requests pending. I appreciate all you proactive people. But that’s for tomorrow.” For 6.4, Linux gets some much needed optimisations for its NTFS3 file system driver. This driver aims to give a small performance boost for the ext4 file system, as well as other small bug fixes for BTRFS and F2FS file systems. The update also cleaned up some memory copy code for x86 processors while adding some more support for RISC-V processors. Along with some driver updates for a few controllers, graphics tablets, and WiFi modems, the kernel is also fleshing out its support for native support on Apple Silicon. Non-Apple PCs are getting most of the love, with better support for MSI laptops and ASUS desktop motherboards. The kernel now also includes new power features for the Steam Deck, which run on a custom AMD APU. Perhaps the biggest standout feature of the new update is the amount of upstreamed Rust code. This will help more Rust developers move towards actually being able to use Rust in the kernel and maybe even Rust-written drivers, Even though this was quite a comprehensive update, the next upcoming one, 6.5, plans to integrate some long-awaited features. Some of the expected features in the upcoming 6.5 update include upgrades to the Rust toolchain, improved support for hyperthreading, better support for AMD, NVIDIA, and even Intel GPUs.","excerpt":"Programmers expect better Rust usability and Rust-written drivers in the future of the Linux kernel with this update.","categories":["AI News"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-06-26T12:30:00","publication_year":"2023","word_count":351,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","Aim","R","Go","Rust","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/linux-kernel-6-4-brings-better-support-for-rust\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10049144,"title":"Stanford Researchers Solve One Thing That Bothered Drones","content":"“Researchers at Stanford introduce a new algorithm that can help drones decide when and when not to offload their AI tasks.” Autonomous vehicles, be it self driving cars or drones, in an ideal situation, should need to upload only 1 per cent of their visual data to help retrain their model each day. According to Intel, running a self-driving car for 90 minutes can generate more than four terabytes of data. And, going by the ideal case of 1 per cent, this would generate 40 GBs of data from a single vehicle. Now, that’s a lot of data transfer for an ML model to crunch and dish out insights for the vehicle to make the right decisions. Despite modern-day superfast-compact chips, onboard computing is still a challenge. It leads to data offloading problems. Low powered robots such as drones require visual analysis for real-time navigation and continual learning to improve their recognition capabilities. Building systems for continual learning will lead to the following bottlenecks (Chinchali et al.): network data transfer to compute serverslimited robot and cloud storagedataset annotation with human supervision cloud computing for training models on massive datasets “Cloud robotics allows mobile robots the benefit of offloading compute-intensive models to centralized servers.” To address these bottlenecks, the team at Stanford has released a series of papers detailing their new algorithms, which have improved the performance of low powered robots such as drones. The researchers focussed their efforts on continual learning, especially which is more computationally intensive and may require human intervention to annotate when the robot encounters new visuals. One of their algorithms factors in available bandwidth and the amount of data to be transferred to help the robots make key offloading decisions. Task relevant communication (via Nakanoya et al.) As depicted above, the researchers implemented a co-design method where the on-robot computation is made more efficient with the help of a pre-trained, differentiable task module at a central server, which allows multiple robots to share an upfront cost of training the model and benefit from centralized model updates. The co-design in this context, stated the researchers, means that the pre-trained task network parameters are fixed, and the task objective guides what salient parts of a sensory stream to encode. The task module has an objective to operate with minimal sensory representation. Addressing the latency issues due to compute intense operations on the drones falls under the category of cloud robotics, where the aim is to make robots use remote servers to augment their grasping, object recognition, and mapping capabilities. The Stanford researchers and their peers incorporated deep RL algorithms to handle the accuracy trade-offs between on-robot and in-cloud calculations to arrive at an optimal solution. In this way, drones or other robots become capable enough to navigate through new terrain without getting baked with an onboard compute. According to the researchers, experimental results show that the performance in key vision tasks improved considerably, which will probably result in safer robots and autonomous vehicles that can navigate seamlessly in a real-world scenario. From autonomous quadcopter drones that can survey a flooded area to picking up rocks on the surface of mars, recent advances in autonomous drone technology and machine learning image classification algorithms have enabled a wide range of aerial capabilities. “Imagine a future Mars or subterranean rover that captures high-bitrate video and LIDAR sensory streams as it charts uncertain terrain, some of which it cannot classify locally,” wrote the researchers.","excerpt":"Cloud robotics deals with low powered robots that can offload their compute to centralized servers if they are uncertain locally or want to run more accurate, compute-intensive models.","categories":["Deep Tech"],"tags":[],"author_name":"Ram Sagar","publish_date":"2021-09-23T12:00:00","publication_year":"2021","word_count":572,"keywords":["Go","machine learning","programming_languages:R","AI","cloud computing","ML","RAG","Aim","continual learning","R"],"extracted_tech_keywords":["AI","machine learning","ML","Aim","RAG","cloud computing","R","Go","continual learning","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/stanford-researchers-solve-one-thing-that-bothered-drones\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136257,"title":"Princeton Digital Group to Invest $1 billion in India for AI-ready Data Centres in Mumbai and Chennai","content":"Princeton Digital Group (PDG), a leading data centre provider in Asia, has announced a major expansion in India, increasing its data centre capacity to a total of 230 MW. The company will invest approximately USD 1 billion in India as part of a broader USD 5 billion initiative to build AI-ready data centres across Asia. This investment will significantly enhance PDG’s presence in India, particularly in Mumbai and Chennai, two of the country’s largest cloud infrastructure hubs. PDG’s flagship data centre in Mumbai, MU1, will see the addition of three new buildings, tripling its current capacity to 150 MW. The first phase of this 100 MW expansion is set for completion in 2026, and the campus will span roughly 15 acres. PDG is also making its debut in Chennai with CH1, a 72 MW AI-ready data centre located in the northern Chennai Metropolitan Area. The facility, set on 9.3 acres, is designed to cater to hyperscalers and will offer scalability for future expansion, reinforcing Chennai’s role as a key cloud hub in India. AI and Cloud Infrastructure Rangu Salgame, Chairman, CEO, and co-founder of PDG, emphasised the role of AI in reshaping the data centre landscape. “Over the last 18 months, AI has completely transformed the data centre landscape. First, we saw the explosion of demand in North America, followed by the initial wave of AI-driven demand in Asia,” he said. “PDG has been a major beneficiary of this initial surge, and our capacity and capabilities in India have been a significant factor in our success. India is well positioned to be a global AI leader, and we are determined to play an important role in making that happen.” Salgame added that the company’s investment in India underscores its commitment to supporting the country’s AI and cloud ecosystems. “By adding three new buildings in Mumbai and by entering Chennai, we are significantly enhancing our hyperscale AI-ready infrastructure footprint in the country.” PDG’s MU1 and CH1 campuses are designed to support high-density AI deployments and incorporate advanced cooling technologies. The Mumbai campus is contracted to source about 50% of its energy from renewable sources, and the company is exploring additional sustainable energy options for both the Mumbai and Chennai facilities. According to Vipin Shirsat, managing director of PDG India, “Mumbai and Chennai have been the pre-eminent hubs for cloud infrastructure in India due to the combination of submarine cable landing proximity, high-quality power supply, availability of renewable energy, and robust infrastructure development. With the advent of AI in India, both locations are well-positioned to become leading AI infrastructure hubs as well.” Sustainability remains a key focus for PDG. The company’s MU1 data centre was the first in Mumbai to achieve the prestigious IGBC Platinum certification, a recognition of its commitment to green infrastructure. Delivered in just 20 months during the pandemic, the facility overcame significant supply chain challenges to meet its deadline. This expansion is part of PDG’s larger 500 MW growth strategy across Asia, with investments also targeting Indonesia and Malaysia.","excerpt":"The facilities are looking to cater to the hyperscaler demand in Asia.","categories":["AI News"],"tags":["AI","Chennai","data centre","India","Mumbai"],"author_name":"Vandana Nair","publish_date":"2024-09-21T14:02:15","publication_year":"2024","word_count":501,"keywords":["Mumbai","API","programming_languages:R","AI","Scala","Git","Chennai","programming_languages:Scala","data centre","R","India"],"extracted_tech_keywords":["AI","R","Scala","Git","API","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/princeton-digital-group-to-invest-1-billion-in-india-for-ai-ready-data-centres-in-mumbai-and-chennai\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":28174,"title":"Theoretical Physicists Hold The Key To Some Of The Toughest Problems In Artificial Intelligence","content":"If you have been following the recent high-profile appointments in the artificial intelligence and machine learning industry, you would have noticed a distinct uptick in theoretical physicists snagging top positions in companies like Samsung AI R&D and Fetch AI. In fact, the trend of moving from a background in theoretical physics and maths to machine learning is gaining ground with PostDoc researchers, since there is a crossover of functions. Physicists excel in ML because computer programs are inherently stochastic in nature. They already have a foundation in math and statistical tools needed to understand the complex ML methods. Physicists also specialise in writing high-performance numerical code, which is another helpful skill for ML development. Noted AI scientist Yann LeCun had once said that traditionally, there is a history of theoretical physicists — particularly condensed matter physicists — bringing ideas and mathematical methods to ML, neural networks, probabilistic inference and SAT problems. In a public lecture, Roger Melko, an associate faculty member at the Perimeter Institute, University of Waterloo expounded that ML algorithms are accelerating discovery in physics. He mentioned how DeepMind’s victory in Go which came from ML prodded researchers from different verticals to think about applying ML algorithms to tackle quantum physics complexity problem. In a recently-published paper by Melko and Juan Carrasquilla, the researchers have talked about the neural network, which is an upgraded version of an AI software used to identify numbers written by humans. Interestingly, the ML algorithm effectively captured and recognized the different phases of matter in a quantum system, that too with minimal adjustments. According to LeCun, the wave of interest in neural networks in the 1980s and early 1990s was in part caused by the connection between spin glasses and recurrent nets popularised by John Hopfield. While this caused some physicists to morph into neuroscientists and machine learners, most of them left the field when interest in neural networks waned in the late 1990s. With the prevalence of deep learning and all the theoretical questions that surround it, physicists are staging a comeback. Many young physicists and mathematicians are now working on trying to explain why deep learning works so well. Rise In Demand For Physicists In AI Industry For example, Fetch AI, an AI and digital economics company, recently announced the appointment of Marcin Abram who them as a Machine Learning Scientist. Abram completed his PhD in Theoretical Physics in 2016 and his doctoral research explored topics on coherence and emergent behaviour in quantum systems. Another key appointment was of Dr Sebastian Seung was by Samsung Electronics to bolster the AI R&D and bring a revolutionary business impact. An eminent computational neuroscientist, Dr Seung originally studied theoretical physics at Harvard. He has worked as a researcher at Bell Labs and a professor at the Massachusetts Institute of Technology (MIT). Hopfield’s Contribution To AI One of the biggest contributions by notable scientist John Hopfield was the formalisation of autoencoder networks when he derived the Hopfield network. In 1982, Hopfield introduced Hopfield Network I, an artificial neural network to store and retrieve memory like the human brain. Hopfield Network II is a single-layered and recurrent network: the neurons are fully connected, that is, every neuron is connected to every other neuron. From there, Boltzmann machines were invented to add some stochasticity to the network so it wouldn’t get stuck in local minima since Hopfield networks are deterministic in the standard formulation. The restricted Boltzmann machines are now used in deep belief networks by stacking them on top of each other, and a greedy layer-wise training algorithm made these networks feasible to use in practice and produce very accurate classifications, as well as being useful generative models. This happened in the last decade or so and is a piece of the story that continues to make headlines today. These two groundbreaking advances came from the field of statistical physics and mathematics — the first one was Hopfield’s insight with the spin-glasses, and the other was the application of simulated annealing to solving these spin-glasses (shortly after the algorithm was invented), which was Hinton’s insight. Simulated annealing itself evolved from the Metropolis-Hastings algorithm described by Metropolis, Rosenbluth, and Teller in the 1950s. One of two independent papers presenting simulated annealing was named A Thermodynamic Approach To The Traveling Salesman Problem so the statistical mechanics roots are really clear. Of late, a lot of maths and physics majors are building a career in this booming field. Given the huge demand for right talent, physicists can add robust value to AI research.","excerpt":"If you have been following the recent high-profile appointments in the artificial intelligence and machine learning industry, you would have noticed a distinct uptick in theoretical physicists snagging top positions in companies like Samsung AI R&D and Fetch AI. In fact, the trend of moving from a background in theoretical physics and maths to machine […]","categories":["AI Features"],"tags":["most revolutionary ai deep learning company"],"author_name":"Richa Bhatia","publish_date":"2018-09-11T07:45:32","publication_year":"2018","word_count":754,"keywords":["Go","artificial intelligence","machine learning","AI","neural network","ML","Git","deep learning","most revolutionary ai deep learning company","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","R","Go","Git","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/theoretical-physicists-hold-the-key-to-some-of-the-toughest-problems-in-artificial-intelligence\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10012641,"title":"Top 15 AI Acquisitions Of 2020","content":"The technology landscape seems to have been relatively less affected by the COVID pandemic if one were to consider the various tech acquisitions that have taken place in 2020. This year was no exception to the trend of bigger companies buying out promising smaller startups to add to their AI prowess. We list some of the top major acquisitions of 2020. AMD Buys Xilinx In October, Advanced Micro Devices (AMD) announced its acquisition of Xilinx in an all-stock transaction that was valued at $35 billion. While AMD specialised in making chips for personal computers especially gaming PCs, Xilinx’s highly-flexible programmable silicon which is enabled by its entourage of advanced tools and software drive rapid innovation across a variety of industries and technologies. NVIDIA Acquires Arm NVIDIA in September said that it would be acquiring Arm Limited from SBG and the SoftBack Vision Fund through a transaction worth $40 billion. Arm’s common architecture supports diverse AI applications. It interconnects mega compute power in the data centre to the machine learning processors for AI algorithms in the edge point devices such as wearables and sensors. NVIDIA, which had been already eyeing the self-driving, 5G, and edge computing industry for a while, and is expected to step as an official key player in the global markets with the acquisition of Arm. FireEye’s Acquisition Of Respond Software Touted as one of the most talked of acquisitions of the year, the intelligence-led security company FireEye acquired cybersecurity investigation automation company Respond Software in November, this year. The deal was closed at $186 million in cash and stock. Intelligence-led security company FireEye offers a single platform for innovative and simple solutions against cyberattacks. With this acquisition, the company now aims at tapping Respond Software’s Mandiant Advantage platform and its cloud-based machine learning tools and software that can drive faster outcomes and provide better security solutions. NVIDIA Acquires Mellanox In another high-profile acquisition of the year, chipmaker Nvidia acquired Israeli-American computer networking products supplier, Mellanox Technologies in April. The deal was closed at an amount of $7 billion. Notably, the deal was supposed to be completed almost a year back in March 2019 but was delayed owing to necessary approvals. This acquisition will help address the challenges and demands for consumer internet services, application of AI to data science in the cloud and edge computing, better utilisation of computing resources, and cost-effective operation. Microsoft Buys CyberX In June, Microsoft acquired Israeli cybersecurity startup CyberX. This startup provides an IoT\/OT cybersecurity platform built by blue-team experts. CyberX was founded in 2013, CyberX quickly grew in ranks and had raised about $48 million before finally being acquired by Microsoft. Notably, this acquisition was also part of Microsoft’s increasing footprint in Israel, which had acquired at least seven more businesses in the last decade. Experts had estimated that with CyberX under its wings, Microsoft would adopt a simpler approach to unified cybersecurity governance in IT and industrial systems to drive end-to-end security across managed and unmanaged IoT devices. This would help companies in quicker detection of advanced threats in increasingly converged networks. Microsoft Acquires Softomotive Microsoft made another major acquisition in May this year when it acquired UK-based robotic process automation company Softomotive. Microsoft had then stated that it would be looking at using Softomotive’s robotic desktop automation tools WinAutomation and ProcessRobot for building software robots and enterprise RPA platforms. The announcement was made at Microsoft’s annual event Microsoft Build 2020. Apple’s Acquisition of Vilynx In October, Apple bought Spain-base AI video startup Vilynx for approximately $50 million. Vilynx is known to build software to analyse a video’s visual, text, and audio content to understand what’s really in the video, using computer vision. It then leverages this data to categorise and tag metadata to videos as well as generate automated video previews and recommend content to users. As per experts, Apple’s various offerings such as Apple TV, Music, News, media, Siri, search, Photos, and other apps would possibly be revolutionised with Vilynx’s technology. Apple Acquires Xnor.ai Right at the beginning of 2020, Apple announced its decision to acquire Seattle-base artificial intelligence company Xnor.ai. The startup founded in 2016 had made huge gains by building tools that help AI algorithms to run on devices directly rather than data centres. Notably, Apple has been pushing to run such software on its devices to maintain data privacy and for speeding up processing. The cost of this deal was estimated to be up to $200 million. Intel Acquires Moovit In May, Intel Corporation announced that it had acquired Israeli-base Moovit for approximately $900 million. Moovit provides real-time updates to ride-sharing service users. It combines information from public transit operators and live information from the user community. This data is then made accessible to commuters to offer them a real-time picture of the best route for their journey. Notably, from 2018 to the time it was acquired by Intel, Moovit’s user base had grown seven times. Intel Buys Convrg.io Just recently in November, Intel acquired another Israel-based company Convrg.io, .to strengthen its machine learning and AI capabilities. Founded in 2016, Cnvrg.io offers data scientists a platform to build and run machine learning tools that can be used to train, run comparisons, and recommendations, among others. As per reports, even after acquisitions, Cnvrg.io was touted to exist as an independent Intel company that would continue serving its existing customers. The financial terms of the deal were not made publicly available. Tech Mahindra Acquires Zen3 Infosolutions Tech Mahindra acquired US-based Zen3 Infosolutions for $64 million in an all-cash transaction. It was announced the Indian subsidiary of Zen group would be acquired by Tech Mahindra India while the American office of Tech Mahindra will acquire the Zen3 Infosolutions America. Zen3 offers expertise in software product engineering, DevOps testing, AI, machine learning, and analytics. Cisco’s acquisition of ThousandEyes In May, Cisco Systems announced that it would be purchasing ThousandEyes, a network intelligence company. The deal that was reportedly worth $1 billion is expected to allow Cisco to expand its cloud infrastructure. As per the company release, the new acquisition will function under Cisco’s Networking Services run by Todd Nightingale. It will help Cisco remove the delineation between public and private networks. Accenture Buys Mudano In February, Accenture announced its acquisition of UK-based Mudano, which is a strategic data consultancy firm. This acquisition was aimed at enhancing Accenture’s analytics, AI, and data capabilities. Mudano joins its applied intelligence unit employing data scientists, engineers, and AI professionals. Notably, Accenture has actively employed acquisition-based growth structure. Accenture’s Acquisition Of Byte By May, Accenture made another high-profile acquisition. This time it was a big data analytics company named Byte Prophecy. This will help Accenture address increasing demand for enterprise-scale artificial intelligence and digital analytics. With Byte Prophecy, Accenture hoped to improve the speed and agility of delivering advanced analytics and AI solutions. It is estimated that with this acquisition, it would be able to add about 50 data science and data engineering experts with a focus on insight automation. VMware Acquires Nyansa Cloud services vendor VMware acquired Nyansa in February this year to enhance its AI and analytics capabilities. Nyansa deploys AIOps that uses artificial intelligence to automate the identification and resolution of IT issues in real-time. VMware said that Nyansa’s technology would help the company to deliver improved network visibility, as well as monitoring, and remediation to its VeloCloud SD-WAN platform. This is expected to make it easier for customers to operate and troubleshoot network disruptions.","excerpt":"The technology landscape seems to have been relatively less affected by the COVID pandemic if one were to consider the various tech acquisitions that have taken place in 2020. This year was no exception to the trend of bigger companies buying out promising smaller startups to add to their AI prowess. We list some of […]","categories":["AI Trends"],"tags":["Mergers and Acquisitions","Microsoft"],"author_name":"Shraddha Goled","publish_date":"2020-11-30T17:00:27","publication_year":"2020","word_count":1244,"keywords":["data science","artificial intelligence","machine learning","AI","computer vision","RAG","Aim","analytics","edge computing","Mergers and Acquisitions","R","Microsoft"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","computer vision","data science","analytics","Aim","RAG","edge computing","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-15-ai-acquisitions-of-2020\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080499,"title":"Coding is Dead, Long Live Programming!","content":"The programming space is abuzz with news every other week about new no-code and low-code platforms. At a time where these platforms are witnessing widespread adoption, many are heralding the end of coding. This sentiment is perfectly captured by a quote from the co-founder of GitHub Chris Wanstrath, who said in 2017, “Coding is not the main event anymore. Building software is the main event. Coding is just one small part of it. We think the future of coding is no coding at all.” Let’s take a look at the differences between programming and coding and how they will be impacted by low code and no code platforms in the future. The difference between coding and programming While programming and coding fall under the same umbrella of skill sets, programming is a superset of coding. All coding processes come under programming. Programmers have to work with algorithms, manage system resources, and figure out the bugs with the programme. They also need to keep the end user in mind when creating programmes so as to ensure that their needs are fulfilled. A programmer’s purview also extended to testing, debugging, and security testing. They also have to assess the problem, formulate a problem statement, design how the software will work, create the software and its accompanying documentation, and perform routine maintenance to keep the software running. Coding, on the other hand, is just a part of the process of programming and comes under software development. It involves writing the user’s instructions in the form of a programme, which the device can then execute. While a programmer has to have a big picture view of the solution that is being built, the coder just needs to provide small solutions for coding problems. It is also based on the trial-and-error method and needs no prior knowledge of the problem statement, as any issues that arise are solved on a case-by-case basis. Now that we have established the difference between coding and programming, let’s delve deeper into how coding is slowly dying out while programming is growing stronger than ever. Do no-code and low-code signify the death of coding? Over the past few years, we have observed a surge in the usage of no-code and low-code development platforms. These environments allow non-coders to build a deployment-ready application by using a GUI, as opposed to writing many snippets of code and going through the debugging process. No-code platforms allow developers to drag and drop different logic components and determine their interaction with other elements, such as APIs. Owing to a talent crunch in the coding space, no-code and low-code platforms have noted a widespread proliferation over the last few years. Moreover, modern development projects also come with a need to reduce turnaround time, leading to enterprises picking up these offerings. The testing process for low-code and no-code platforms is also fairly straightforward since developers can simply rearrange the modules until the programme works as expected. The approach that low-code and no-code environments take to solve common coding problems are also relatively elementary. Each block holds a certain set of steps or capabilities, and can be reused to suit the needs of the developer. This allows users to create an application in a process similar to drafting a flowchart, enabling even users who have no coding knowledge to churn out a deployable application. The surge in popularity of these platforms shows that coding can be opened up to so-called citizen developers—another term for people with little coding knowledge. Even those with nearly no knowledge about coding can create an application that can be deployed to the end customer and will provide an experience similar to an application coded in a traditional environment. The difference between no-code and low-code No-code platforms usually target industry-specific functions, with their primary audience being non-technical users who wish to optimise their business operations. Low-code platforms, on the other hand, typically target developers who are operating under a time constraint. They are used to deliver applications quickly and conserve resources which can be better utilised to create something that has more impact on the organisation’s bottom line. Since no-code platforms usually target a line of business users, who look to create applications that can be used to expedite their business operations, they provide a host of benefits. Not only do these platforms open up access to creating code-based solutions to non-codes, the feature-rich nature of modern no-code environments allow developers to solve problems unique to their line of work. In addition to this, no-code platforms can also be used to automate workflows, thereby saving more resources. As with any upcoming market, there are challenges associated with no-code and low-code platforms. Among the biggest perhaps is the fact that customisability on a use case basis is highly restricted to the features offered by each platform. Developers have also brought up the question of platform siloing with no-code environments. They have cited concerns that applications cannot be easily transferred from one no-code platform to the other, as they do not offer inter-compatibility features. With the continued advancement of no-code platforms, there might come a day where even the most capable applications can be created without needing to write a single line of code. However, the reality today is not so peachy. While it is true that no-code platforms can allow citizens to create applications, customised features still need coding. Even the most powerful no-code platforms need at least a bit of code when looking to make fully-featured applications. Is no-code just another developer trend, or is it here to stay? Programming has its own set of anachronisms and trends, and it’s common to see a community flocking to a new platform or language simply because it’s new. These new trends have taken on a form of their own, creating a kind of pop culture in programming wherein developers have to either adapt or risk falling behind. While no-code and low-code environments have become quotidian tools in any developer’s arsenal, there are many who doubt the veracity of their continued usage. However, while several continue to be apprehensive of low-code environments, the industry has already caught on to the trend and has begun adaptation on a wide scale. A study by Gartner found that the usage of low-code and no-code technologies will nearly triple by 2025. They also predicted that 70% of new applications developed by organisations will use low-code and no-code platforms by then. Companies also acknowledge the benefit of adopting these technologies on a large scale, as it will ultimately cut down on the talent they need to acquire to build applications. Observing the trends in this space, it is clear that no-code and low-code platforms are likely here to stay. While they might not replace a programmer’s jobs in the near future, there is concrete evidence that these platforms will provide an important value add for enterprises and programmers alike.","excerpt":"Low code and no code platforms are the talk of the town, but will they hold up in the future?","categories":["AI Features"],"tags":["AI Developers"],"author_name":"Anirudh VK","publish_date":"2022-11-23T12:00:00","publication_year":"2022","word_count":1147,"keywords":["Go","API","AI Developers","programming_languages:R","AI","programming_languages:Go","Git","RAG","GAN","GitHub","R"],"extracted_tech_keywords":["AI","RAG","R","Go","Git","GitHub","API","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/coding-is-dead-long-live-programming\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":6870,"title":"Cost Efficient Solution for Transformer Health Monitoring","content":"Power transformers are an integral part of any power distribution network. Owing to humongous capital and maintenance cost involved for this equipment, there exists a huge scope to optimize expenses thereby improving the revenue. Where is the struggle for distribution companies? Unavailability of dashboards or tools to visualize the status of multiple parameters associated with power transformer Unforeseen failures in transformers leading to downtime and penalty Loss of asset life due to continuous faults\/failures Poor maintenance planning Unavailability of advanced solutions to monitor the health of transformers What happens when transformer fails? Transformer failure can lead to serious loss of revenue in multiple ways. When transformer fails, it leads to isolation of its customer from the grid, and the distribution companies are bound to pay penalty. This adds to the loss of revenue on account of its failure in supplying the generated power to its customer. Consequences of a failure The other important loss is the power transformer itself. Due to improper maintenance or utilization, there is loss of operational life of transformer. Every fault reduces the transformer’s remaining life in addition to the cost incurred towards repairing\/ replacement. This leads to loss of revenue for the distribution company operating the transformer How is the solution developed? The proposed model dwarfs the limitations with respect to the operator’s capability, limited input data, absence of intelligent monitoring devices and inability to correlate multiple test results. The model is based on developing an adaptive testing\/maintenance plan based on the transformer’s history and domain expertise instead of adhering to a static testing plan. By this way, the distribution company can optimize maintenance cost and avoid potential failures. Multiple tests are conducted as per the adaptive testing plan and the results are updated to a central database from which the model fetches data for analysis. The reports are analyzed simultaneously, and various calculated parameters are evaluated for determining health index of transformer. In case of potential faults or failures, the type of fault is identified before evaluating the calculated parameters. Dashboards are the most important part of this model which visualizes trending of various test parameters and calculated parameters, and customize the quantity of data to be displayed as per our requirement. Based on the analysis, the model updates the adaptive testing plan and the next test date is determined. Following parameters\/tests are considered for this model building: Dissolved Gas Analysis (DGA) Oil Sample Test (OST) Loading History Winding hot spot & Top oil Temperature Degree of polymerization or Furanic compound analysis Bringing Analytics to Analysis: The key to this approach is to maximize the amount of information that can be utilized from existing data set. The history of transformer forms the base on which the model is developed. Accuracy of the model is improved with increased availability of data, which has enough data points (failure points) for tuning the model, and that data is fed into the model. Various faults are also identified based on the developed relationships and reference standards (IEC\/IS) What does a solution dashboard look like? Dashboards created for transformer health monitoring and assessment are shown below. Box plot enables us to understand and analyze the overall distribution of recorded values. Trending of multiple parameters helps the operator in monitoring the variation of each parameter with reference to its limits over a period of time, and plan necessary corrective actions or maintenance activity in case of abnormal variations. Fault identification using IEC gas ratios is another critical feature of this model. To assess the overall condition, a numerical value is computed to represent the health of transformer and the same is compared along with its age. A unique solution for monitoring various parameters individually and to assess overall health of the transformer is thus achieved with this model. How will solution help Power distribution companies? The solution helps the distribution companies in the following areas, Monitor asset condition Fault prediction Increase reliability Asset life extension Schedule maintenance activity Downtime management","excerpt":"Power transformers are an integral part of any power distribution network. Owing to humongous capital and maintenance cost involved for this equipment, there exists a huge scope to optimize expenses thereby improving the revenue. Where is the struggle for distribution companies? Unavailability of dashboards or tools to visualize the status of multiple parameters associated with […]","categories":["IT Services"],"tags":["Generative Pre-Trained Transformer"],"author_name":"Ashok Kumar","publish_date":"2015-02-03T11:02:33","publication_year":"2015","word_count":660,"keywords":["Go","API","programming_languages:R","AI","ai_frameworks:Transformers","Transformers","programming_languages:Go","ViT","analytics","Generative Pre-Trained Transformer","R"],"extracted_tech_keywords":["AI","analytics","Transformers","R","Go","API","ViT","ai_frameworks:Transformers","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/cost-efficient-solution-transformer-health-monitoring\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10070136,"title":"Can reinforcement learning solve the NP-Hard problems?","content":"For an algorithm to be termed “efficient”, its execution time must be constrained by a polynomial function of the input size. It was realised early on that not all issues could be handled thus rapidly, but it was difficult to determine which ones could and which couldn’t. Some so-called NP-hard issues are thought to be impossible to answer in polynomial time. NP-hard stands for non-deterministic polynomial-time hardness. This article will be focused on understanding some NP-hard problems and trying to solve them with Reinforcement Learning. Following are the topics to be covered. Table of contents Understanding NP ProblemsWhen a problem is classified as NP-hard?Reinforcement learning for solving NP-hard problem A decision problem has a single boolean output: Yes or No. Let’s understand more about when a decision problem becomes non-deterministic. Understanding NP Problems Generally, those sets of decision problems that can be solved under a polynomial-time are categorized as ‘P’. The decision problem that can be proved and verified the answer to the problem is a ‘Yes’ in a polynomial time is classified as NP. NP is intuitively defined as the set of decision problems where it can prove a ‘Yes’. The circuit satisfiability issue, for example, is in NP. If a given boolean circuit is satisfiable, then every combination of m input values that generate True output is proof that the circuit is satisfiable; the proof may be checked by evaluating the circuit in polynomial time. It is commonly assumed that circuit satisfiability is not in P or co-NP, although no one knows for sure. If a polynomial-time algorithm for the problem implies a polynomial-time algorithm for every problem in NP, the problem is NP-hard. Intuitively, if one specific NP-hard issue can be solved rapidly, then any problem with an easy-to-understand answer may be addressed using the solution to that one special problem as a subroutine. NP-hard issues are as least as difficult as any other problem in NP. Finally, a problem is said to be NP-complete if it is both NP-hard and contains an element of NP. Informally, NP-complete issues are the most difficult for NP. A polynomial-time algorithm for only one NP-complete issue implies a polynomial-time algorithm for all NP-complete problems. Analytics India Magazine Are you looking for a complete repository of Python libraries used in data science, check out here. When a problem is classified as NP-hard A reduction argument can be used to demonstrate that any problem is NP-hard. Reducing issue A to another problem B entails explaining an algorithm to solve problem A while assuming that an algorithm to solve problem B already exists. To demonstrate that your problem is difficult, an efficient method to solve a different problem that is already known to be difficult must be given, utilising a hypothetical efficient programme treating the problem as a black-box subroutine. Proof by contradiction is the core logic. The reduction indicates that if the first problem were easy, the second problem would be easy as well, which it isn’t. Alternatively, because the previous issue is known to be difficult, the reduction implies that the current problem must likewise be difficult; the hypothetical efficient method does not exist. For example, a finite set of integers are given and need to find that any non-empty subset of them adds up to zero. To prove that this problem is an NP-hard problem needs to use 3CNF SAT. Conjunctive normal form (CNF) boolean formulas are those that are represented as conjunctions (AND) of clauses, each of which is the disjunction (OR) of one or more literals. If each sentence has precisely three different literals, the boolean formula is in 3-conjunctive normal form (3CNF SAT). The reduction method builds instance sets and targets given a three CNF formula over variables with clauses each having precisely three different literals. Two assumptions must be made to establish that this issue is NP-hard: first, no sentence contains both the variable and its negation, and second, each variable appears in at least one clause. This will be regarded as an NP-hard task. Reinforcement learning for solving NP-hard problem Reinforcement learners are taught a series of decisions. In an uncertain, possibly complicated environment, the agent learns to attain a goal. The method uses trial and error to find a solution to the problem. Its objective is to maximise overall return. The reward policy, which is the training rules, is specified, and the model is given no cues or ideas for how to tackle the problem. It is up to the model to choose how to accomplish the job to maximise the reward, beginning with completely random trials and progressing to complex tactics. Let’s take a few NP-hard examples and understand the solution to them using Reinforcement Learning. Capacitated Vehicle Routing Problem The Vehicle Routing Problem is a combinatorial optimization problem in which the aim is to find the most cost-effective routes for a fleet of vehicles to take to serve a given set of clients. Despite its simplicity, the VRP belongs to the family of NP-hard problems, which are distinguished by their exponentially growing difficulty in finding a solution as the issue grows in size. To give a few examples of VRP, consider the e-commerce giant Amazon, which delivers 2.5 billion packages per year while constantly attempting to optimise every step of the process, including determining the best route their drivers should take to deliver the parcels on time and at the lowest possible cost. The Uber issue is a variation of the VRP in which the corporation wishes to allocate the best possible routes to drivers to pick up and drop off clients. The purpose of the CVRP is to serve each customer’s demand with a fleet of homogenous vehicles, all starting from a specific node called the depot and with a defined capacity of carried items. A van will visit each client one at a time, completing their orders and returning to the depot after the entire truckload has been delivered. The cost of travelling between two nodes is just the Euclidean distance between them. Analytics India Magazine The depot is represented by the squared node, the customers are represented by the circled nodes, and the network is fully linked and undirected. The dotted lines indicate the pathways that were not chosen in the suggested solution. The Euclidean distance between two nodes is the cost associated with each edge. The highlighted pathways reflect a workable solution since all five clients have been visited and the total of the amounts fulfilled in each of the two trips is less than or equal to the delivery vehicle capacity Q. The recommended solution is also optimum in terms of the total distance travelled. The CVRP may be viewed as a sequential decision-making issue in which the vehicle must select where to go next at each timestep to complete its task of servicing all clients. With this viewpoint, the components characterising the RL issue may be expressed. State, Action and Reward spaces Each customer’s node on the graph is related to their current demand as well as their location. When a vehicle visits a node, whether to meet a client’s demand or to replenish the truck at the depot, the graph can be updated to represent the current condition of the system. The depot’s location is critical to know, especially when the truck is nearly empty and needs to return to the depot to refill. Depending on the volume of items still in the truck, it may be advantageous to pick a node closer to the depot because the truck will need to get there shortly. As a result, the present load in the vehicle must also be reflected in the state space. The location of each node and the present position of the truck are critical pieces of information to offer to the learning agent for it to grasp the cost of travelling from one node to another. Finally, the present demand of each client aids in determining which nodes have yet to be visited and are candidates for the truck’s next destination. More information, such as the distance travelled or the prior position can be added to the state space. The more information that may assist the policy’s neural network in making the optimum choice, the better. However, there is a trade-off to consider: if there is too much information, the “curse of dimensionality” will be an issue, and the training will most likely take too long. The action space is a limited collection of actions that determine which node the vehicle visits next. In principle, the policy might choose any node at any timestep. However, returning to the same node or travelling to the depot while having enough load to serve at least one more client is extremely inefficient. These two tactics would be used to steer the decision-making process. When the truck’s current load is insufficient to meet customer demand, the related nodes are masked off during the action selection process to ensure that these nodes are not picked.When a client node’s demand is met, it is masked out until the run ends. This new logic aids the RL algorithm’s exploration phase and applies to all CVRP instances. The problem’s reward signal is comparable to the MIxed integer programming (MIP) formulation: the aim is to identify the cheapest routes travelled by truck, and the purpose of the RL is to maximise the cumulative reward, thus we use the inverse of the total distance travelled. The RL algorithm is subjected to various issue setups with a fixed set of nodes that are at different places each time. Exposing the learning agent to a large number of cases allows the learned policy to generalise and prevent overfitting to a single CVRP instance. A run begins with all of the customers’ wants unfulfilled and the truck in the depot; the agent then interacts with the environment by visiting nodes until all of the customers’ demands have been met. Only after each episode is the routes chosen evaluated, resulting in the award for this run. Because the masking approach only allows particular nodes to be picked, the trained policy’s solution is guaranteed to be practical. However, there is no assurance that the proposed solution is optimum; the algorithm learns to reduce the distance travelled but has no idea of the minimum that may be attained; thus the algorithm’s exploration phase could progress towards the ideal solution or become trapped in a local minimum. As a result, the RL technique can only be classed as a heuristic method. Graph mining Heterogeneous Information Networks (HINs) can be used to simulate complex interactions in real-world data, such as social networks, biological networks, and knowledge graphs. HINs are typically connected with numerous types of objects\/nodes and meta-relationships. Image source Using the previously illustrated heterogeneous academic network as an example, it has four categories of nodes: papers, authors, institutions, and publication venues, as well as eight types of meta-relationships between these nodes. Traditional machine learning techniques struggle to model HINs due to their complicated semantics and non-Euclidean nature. Through sequential exploration and exploitation, Deep Reinforcement Learning agents may adaptively create optimal meta-paths for each node in terms of a given goal. This is appealing and feasible for HINs with complicated semantics. An MDP’s main components are a collection of states, a set of actions, a decision policy, and a reward function. A meta-path is constructed with maximum timesteps that can be characterised as a circular decision-making process that can be treated naturally as an MDP. The initial node’s characteristics as state, and selecting one relation is the first step in expanding the meta-path. Following the expanded meta-path, build appropriate meta-path instances for the algorithm to learn node representations for use in a downstream prediction task to get reward scores. The meta-path design method is formally formulated as follows. The state space is utilised to aid the decision policy in determining which relation to applying while extending the meta-path. It is critical to record all relevant information in an existing meta-path. As a result, the state is configured to remember all nodes participating in the meta-path. The action space for a state is comprised of the set of accessible relation types in the HIN as well as the unique action STOP. Beginning with the initial node, the decision strategy predicts the most promising relation for extending the current meta-path to get a higher reward score repeatedly. It may either approach the maximum timestep or discover a node representation that any more relation in the meta-path would impair the performance of node presentations on the downstream job, in which case the decision policy selects the action STOP to complete the path design process. An MDP’s decision policy seeks to map a state in state space into action in action space. The state space and action space in the meta route design process are continuous space and discrete space, respectively. To estimate the action-value function, a deep neural network is utilised. Furthermore, because every random action is always a positive integer, the DQN Q-network is employed as the decision policy network. As the deep neural network for the Q-network, an MLP is used. The output represents the possibility of extending the meta-path by selecting other relations in the action space. The reward is an important aspect in directing the RL agent, encouraging it to attain greater and more steady performance. Thus, the reward is defined as an improvement in performance on a certain task relative to previous performances. The objective job is node classification, and the evaluation performance is based on its accuracy on the validation set. According to the descriptions above, the suggested meta route design process of node at timestep contains three phases: Getting the timestep’s stateForecasting an action based on the present state to extend the meta-pathCurrent state updating Back-propagation and gradient descent are used to improve the Q-network parameters by minimising the loss function. Travel Salesman Problem The travelling salesman problem (TSP) is an algorithmic problem in which the goal is to discover the shortest path between a collection of points and places that must be visited. The cities that a salesman could visit are represented by the points in the problem statement. The objective of the salesman is to minimise both travel expenditures and distance travelled. The TSP RL solution has an encoder-decoder design. The first city to visit is chosen at random since the trip should be invariant to the initial city. As a result, the solver’s decisions begin at a time step greater than 2. In translation, invariance may be further leveraged by evaluating relative locations of the most recently visited city rather than the original absolute coordinates. The RL solver’s input consists of simply two bits of information. The present partial tour information now just requires the relative preprocessed location of the first visited city. Because the most recently visited city is always represented by the origin, it can be omitted.The remaining TSP instance information relates to the relative preprocessed locations of the unvisited cities, as well as the first and last visited cities. The relative preprocessed locations are saved in a matrix, which includes the most recently visited city. Because it represents the labels of the remaining graph nodes on which the tour should be finished. The model’s encoder is made up of a graph neural network (GNN) that computes an embedding of the relative city locations and a multilayer perceptron (MLP) that computes an embedding of the first visited city. The GNN encapsulates the graph information characterising the remaining TSP issue. It is invariant with regard to the order of its inputs as a GNN, and its outputs are dependent on the graph structure and information (i.e., city positions). The MLP encodes information about cities visited. Since the independence of the visited cities between the first and last cities (not included) is used. For encoding the locations of the visited cities, the model does not require any sophisticated architecture, such as LSTM. Because just the relative location of the first city is required, a basic MLP suffices. After computing the embeddings, the probability of choosing the next city to visit is calculated using a conventional attention mechanism with relative location and first city as the keys and queries, respectively. The decoder produces a trainable weight vector. The vector is then transformed into a probability distribution using a softmax transformation. Conclusion Reinforcement Learning necessitates a training phase in which the algorithm learns the optimum course of action to take after being exposed to a variety of configurations. Because of its simplicity, just the problem’s rules must be defined, and the algorithm learns how to solve the problem via trial and error. With this article, we have understood NP-hard problems and their solution through Reinforcement Learning. References Read more about Combinatorial OptimizationRead more about Graph Mining","excerpt":"NP-hard is a set of sequentially decision problems which are hard to solve in a time frame.","categories":["AI Trends"],"tags":["Reinforcement Learning"],"author_name":"Sourabh Mehta","publish_date":"2022-06-30T12:00:00","publication_year":"2022","word_count":2800,"keywords":["data science","knowledge graphs","machine learning","Reinforcement Learning","TPU","AI","neural network","ML","RAG","Aim","analytics"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","data science","analytics","Aim","RAG","knowledge graphs","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/can-reinforcement-learning-solve-the-np-hard-problems\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10163276,"title":"Kore.ai’s No-Code Agents are Out to Democratise AI Developments","content":"The dawn of AI agents promises a shift in how enterprises build and deploy applications. By now, we’ve seen big tech and SaaS giants going all-in on agents. However, other companies have been quietly working on providing AI solutions for enterprises and have also rolled out no-code AI agents. Case in point: Kore.ai. The eleven-year-old Orlando-based Kore.ai, founded by Raj Koneru, has its second major hub in Hyderabad, which includes an R&D centre. The company helps businesses create AI chatbots and virtual assistants, and its customers span banking, healthcare, and airlines. Kore.ai is now enabling companies to build and deploy AI agents without extensive coding knowledge. Conversational AI to AI Agents “We provide a platform which is like a set of platform core capabilities, services, and a lot of no-code tools for builders as well,” said Prasanna Arikala, CTO of Kore.ai, in an exclusive interaction with AIM. The platform has evolved from conversational AI to a comprehensive agent development ecosystem. It now includes the ability to build agents, create tools for agents to interact with, and develop sophisticated RAG (retrieval-augmented generation) pipelines for multi-agent applications. “If you have a system of record, you can build pretty much any application with agents. The application development and deployment paradigm is significantly changing, and enterprises have quickly realised that,” explained Arikala, emphasising the paradigm operational shift of enterprises. On the database front, in an earlier interaction with AIM, Redis VP of AI product management Manvinder Singh confirmed that Kore.ai uses Redis as a data platform to power their virtual AI agents. Kore.ai’s AI agents are making significant impacts across various industries. The company boasts about 450 customers, including some of the world’s largest banks and healthcare companies. One of its customers, a major wealth management company, deploys AI agents for its 60,000 employees. Arikala explained how these agents have dramatically reduced the time required for tasks such as creating customer proposals. “Previously, it would take weeks for them to compile, collect all the information in accordance with the enterprise guidelines, corroborate, build, templatise, and deliver a report. Now, they just upload the data, and it gives the outcome within minutes,” he said. Arikala claims that the platform’s versatility allows it to be applied across various domains, including customer service, process automation, and enterprise work management. “One out of the top five banks in the US deploy AI agents for their customer service,” Arikala said. During the COVID-19 pandemic, Pfizer leveraged Kore.ai’s AI agent platform to deploy a multilingual support system in 17 languages globally. This system assisted healthcare professionals in efficiently accessing critical vaccination-related information. Challenges Remain Despite the platform’s success, Kore.ai acknowledges the challenges of deploying AI agents at scale, particularly the governance part when the number of agents keeps growing. Arikala emphasised the need for oversight in agent development, questioning who built them, how they are being used, and whether they comply with enterprise guidelines and SOPs. Unlike workflows, agents don’t follow a deterministic approach, making safeguards essential. To address these challenges, Kore.ai is developing solutions, such as a built-in agent evaluation service, as part of its platform. It allows for periodic assessments of AI agents, generating comprehensive reports on their performance and behaviour. Kore.ai envisions a future where AI agents become ubiquitous in enterprises. “In the future, the enterprise will be all about a network of AI agents, and there will be centralised orchestrators that allow for a hub for the internet sorts, “ predicts Arikala. As confirmed by Arikala, Kore.ai has been witnessing growth rates of 100% year-over-year for the past three years and continued significant growth for the current fiscal year. With this growth rate and IPO plans in the coming years,  it seems that Kore.ai is well-positioned to lead the AI agent revolution.","excerpt":"During the COVID-19 pandemic, Pfizer leveraged Kore.ai’s AI agent platform to deploy a multilingual support system across 17 languages globally.","categories":["AI Startups"],"tags":["AI Agents","Kore ai","low code","no code","pfizer"],"author_name":"Vandana Nair","publish_date":"2025-02-12T09:00:00","publication_year":"2025","word_count":626,"keywords":["Go","AI","Kore ai","chatbots","R","ETL","AI Agents","no code","virtual assistants","RAG","automation","pfizer","Aim","low code","Redis"],"extracted_tech_keywords":["AI","Aim","RAG","chatbots","virtual assistants","Redis","R","Go","ETL","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/kore-ais-no-code-agents-are-out-to-democratise-ai-developments\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10171725,"title":"When Google Cloud Paused the Internet","content":"Google Cloud Platform (GCP), which offers a wide range of cloud services for various applications worldwide, experienced downtime on Thursday, the company reported. In a status report, Google said that “multiple GCP products are experiencing service issues” across several locations worldwide. The incident began at 10:51 am (PT) on Thursday and was resolved hours later, at 6:18 pm (PT), which is 11:21 am IST and 6:48 am IST, respectively. Platforms such as Spotify, Replit, Discord, Claude, Snapchat, Twitch, Cloudflare, and other services that use the Google Cloud faced outages. Besides, many of Google’s products, like Google Meet, Drive, Chat, and Gmail, that rely on GCP were said to be affected. These products use Google Cloud’s infrastructure for services such as storage, computing, networking, data processing, and more. Amjad Masad, CEO of Replit, said on X, “Google Cloud is having an outage and that’s taking Replit down.” Anthropic, the company that develops the Claude family of AI models, stated on a status page that it has identified several issues with its products due to the outage affecting GCP. The company also mentioned that image and file uploads via the API and the Claude web app were unavailable due to the GCP issue. 🚨🚨It looks like EVERYTHING is DOWNAmazon Web Services, Anthropic, AT&T, Box, Breezeline, Calendly, CharacterAI, Cloudflare, cursor, Dialpad, Discord, DoorDash, Dragon Ball, Embark Studios, Equifax, Etsy, FuboTV, Gmail, Google, Google Cloud, Google Drive, Google Gemini, Google… pic.twitter.com\/Sw5ziwbCt8— Hackmanac (@H4ckmanac) June 12, 2025 GitHub, the code hosting platform that also provides AI-assisted programming capabilities with GitHub Copilot, said in an incident report that underlying AI models like Claude and Google Gemini were facing issues. “Copilot is experiencing degraded performance,” said GitHub. However, all of the above services were restored to normalcy after a few hours. Source: x.com\/GeraDeluxer Cloudflare Got Hit Bad Cloudflare was one of the platforms most impacted by GCP’s outage. The company provides internet security services to some of the leading websites worldwide, with around 19.3 million websites using its services. “The outage lasted 2 hours and 28 minutes, and globally impacted all Cloudflare customers using the affected services,” said the company in a blog post. It added that the outage was due to a failure in an underlying storage infrastructure, a part of which runs on a third-party cloud provider. A spokesperson from Cloudflare told CNBC that its issue stemmed from Google’s Cloud services as well. “A limited number of services at Cloudflare use Google Cloud and were impacted. We expect them to come back shortly. The core Cloudflare services were not impacted,” said the spokesperson. The company’s stock also fell by 5% on Thursday. We let down our customers at @Cloudflare today. Our Workers KV service failed and the downstream products that rely on that service had outages of their own. We will publish a full postmortem soon.I know that these kinds of incidents have real and serious impact for teams…— Dane Knecht 🦭 (@dok2001) June 12, 2025 “While the proximate cause (or trigger) for this outage was a third-party vendor failure, we are ultimately responsible for our chosen dependencies and how we choose to architect around them,” said the company. Cloudflare’s Workers KV service—a distributed, serverless key-value store designed for use with Cloudflare’s Workers, the company’s serverless computing platform—was significantly affected. “Workers KV saw 90.22% of requests failing: any key-value pair not cached and that required to retrieve the value from Workers KV’s origin storage backends resulted in failed requests,” said the company. Several Cloudflare products, such as Access, Gateway, WARP client, and Dashboard, were also impacted. These products are a key part of the overall infrastructure that Cloudflare provides to its customers. For instance, Access is a service that users Workers KV to store application and policy configuration along with user identity information. During this incident, Cloudflare said, Access failed 100% of identity-based logins for all application types. The company claims that no data was lost as a result of the incident. How can Google Cloud, AWS, and Cloudflare all be down at the same time? These companies manage nearly 90% of all internet activities and applications. pic.twitter.com\/a9bpAGvfTz— Sam Abdul (@SamAbdul_) June 12, 2025 No, AWS and Azure Did Not Go Down Several reports emerged that GCP’s counterparts, Microsoft’s Azure and Amazon’s Web Services, also faced outages, based on information from Down Detector, a website that checks what websites are currently facing downtime. However, neither Microsoft nor AWS have officially confirmed this. Gergely Orosz, author of The Pragmatic Engineer, said this is likely a misunderstanding of how DownDetector checks for outages. “It cannot differentiate between ‘Azure is having issues’ vs ‘some sites hosted on Azure are having issues’ (that can be the case due to, eg, CF (Cloudflare)),” said Orosz in a post on X. According to Tom’s Guide, an AWS spokesperson has confirmed that they are not experiencing any service disruptions and stated that “Down Detector does not accurately reflect AWS issues.” Overall, Thursday’s events were a stark reminder of how relying on a single cloud service provider can bring down some of the most popular apps on the internet. “We build all this powerful tech, then that one Google Cloud outage takes it all offline,” said a user on X. “Feels like the universe saying: ‘Take a break. Go outside. You’re not in control anyway’,” they added.","excerpt":"Platforms such as Spotify, Replit, Discord, Claude, Snapchat, Twitch, Cloudflare, and other services that use the Google Cloud faced outages.","categories":["Global Tech"],"tags":["AI (Artificial Intelligence)","Google","Google Cloud"],"author_name":"Supreeth Koundinya","publish_date":"2025-06-13T14:57:34","publication_year":"2025","word_count":884,"keywords":["Anthropic","Go","Google Cloud","GCP","AWS","AI","R","serverless","RAG","Aim","Google","Azure","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","Anthropic","Aim","RAG","AWS","Azure","GCP","serverless","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/when-google-cloud-paused-the-internet\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10118876,"title":"Microsoft’s Phi-3 Outperforms Meta’s Llama 3 and Fits Perfectly on an iPhone","content":"“One of the things that makes Phi-2 better than Meta’s Llama 2 7B and other models is that its 2.7 billion parameter size is very well suited for fitting on a phone,” said Harkirat Behl, one of the creators of the model, who has now created Phi-3, the latest open source model by Microsoft. Phi-3-Mini is a 3.8 billion parameter language model trained on an extensive dataset of 3.3 trillion tokens. Despite its compact size, the Phi-3-Mini boasts performance levels that not just exceed the recent ones such as Mixtral 8x7B and GPT-3.5, but even surpass the recently launched Meta’s Llama 3 8B on MMLU benchmarks. Despite these high capabilities, Phi-3-Mini can run locally on a cell phone. Its small size allows it to be quantised to 4 bits, occupying approximately 1.8GB of memory. Microsoft tested the quantised model by deploying Phi-3-Mini on an iPhone 14 with an A16 Bionic chip, running natively on the device and fully offline, achieving more than 12 tokens per second. phi-3 is here, and it's … good :-).I made a quick short demo to give you a feel of what phi-3-mini (3.8B) can do. Stay tuned for the open weights release and more announcements tomorrow morning!(And ofc this wouldn't be complete without the usual table of benchmarks!) pic.twitter.com\/AWA7Km59rp— Sebastien Bubeck (@SebastienBubeck) April 23, 2024 Along with this, Microsoft has also introduced Phi-3-Small and Phi-3-Medium models, both significantly more capable than Phi-3-Mini. The Phi-3-Small 7 billion parameter model achieves an MMLU score of 75.3 outperforms Meta’s recently launched Llama 3 8B Instruct with a score of 66. With a Grain of Salt “To best benefit the open source community, Phi-3-Mini is built upon a similar block structure as Llama-2,” reads the technical report by Microsoft. But currently, the model is limited to English, which is not ideal for other languages and for Indic AI developers. The innovation behind Phi-3-Mini lies in its training dataset, an expanded version of the one used for its predecessor, Phi-2. This dataset comprises heavily filtered web and synthetic data. The model has also been optimised for robustness, safety, and chat format. Given that small open-source models are performing so well, it wouldn’t be surprising if soon there is a model outperforming OpenAI’s GPT-4. Interestingly, Meta is also training a model with around 400 billion parameters which would possibly be able to outperform the closed models once it is launched. “BUT – as with all (tiny) models, benchmarks tell us less than vibes,” said Matt Shumer on X. In a discussion, people highlight the issue with the benchmarks of the model. “According to what I’ve read, Phi-2 was much worse than its benchmark numbers suggested. This model follows the same training strategy,” read a comment. Since the model is built by Microsoft and uses synthetic data for training, it is possibly using GPT-4 output for training. “I don’t think it’s impossible for a small model to be very good. I see their ‘synthetic data’ as essentially a way of distilling GPT-4 into smaller models,” said the same user. Furthermore, the model is also trained with only 4.8 trillion tokens, which is significantly less than 15 trillion tokens that Llama 3 was trained on. Regardless, the model can run on a phone, which the Llama series of models are still a little far away from given the size. Moreover, Phi models aren’t specifically tuned for chat or instruct, which makes them perform slightly worse when compared to Llama models when incorporating in real world scenarios. On the other hand, Behl had told AIM that scaling laws are not necessarily true. “You don’t need a specific size or number of parameters for a model to get good at coding,” said Behl, saying that you do not need large models to instil intelligence. “All you need is a small amount of high quality data, aka textbook quality data.” This is what is continued with Phi-3. What Dent will it Make? Since the model is built for on-device and edge use cases, it is ideal for the ongoing shift towards AI devices. Moreover, Apple is also experimenting with AI on edge, and Phi-3 might give Microsoft an edge (pun intended) over Apple. Moreover, since such small models are outperforming larger models, this might also possibly make an impact on OpenAI’s release of GPT-5, as enterprises are also increasingly adopting open source models. Who knows, the company might decide to open source one of its upcoming models, though that seems highly unlikely for now. Microsoft has also kept in mind the need to make LLMs that are up to date with current information, thus have made Phi-3 ideal for RAG use cases as well. Microsoft believes that training models on synthetic data reduces the size of the model, and also brings in a lot of capabilities within them, which is different from how GPT-3 was trained. “Textbooks are written by experts in the field, unlike the internet where anybody can write and post, which is how GPT-3 is trained,” said Behl.","excerpt":"Microsoft shows who is the boss of tiny open source models.","categories":["Global Tech"],"tags":["iPhone","LLaMA","Microsoft"],"author_name":"Mohit Pandey","publish_date":"2024-04-23T14:41:28","publication_year":"2024","word_count":835,"keywords":["TPU","GPT-5","AI","OpenAI","AWS","ML","LLaMA","RAG","CuPy","Aim","iPhone","R","Microsoft"],"extracted_tech_keywords":["AI","ML","GPT-5","OpenAI","Aim","CuPy","RAG","AWS","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsofts-phi-3-outperforms-metas-llama-3-and-fits-perfectly-on-an-iphone\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10054332,"title":"Major Announcements Made At Tech Conferences In 2021","content":"Other than the exchange of knowledge and ideas, tech conferences are a great way for companies to introduce customers to new products, innovations and experiences that they are working on. Like other years, we saw big names in tech conducting conferences either virtually or offline. Let’s take a look at some of the biggest ones held this year and what announcements they made. Connect 2021 In a major decision, CEO Mark Zuckerberg announced that Facebook has changed its name to Meta at Connect 2021 held recently. The reach of Meta will be much beyond just social media. The metaverse will give the feeling of a hybrid structure of online social experiences expanded into the physical world. From the fourth quarter of 2021, the company will be reporting on two operating segments of Family of Apps and Reality Labs. For more details, refer here. NVIDIA GTC 2021 NVIDIA has also jumped on the metaverse bandwagon. It announced Omniverse VR, where creators, designers, researchers, and engineers can connect major design tools and assets to collaborate in a shared virtual space. It also revealed Omniverse Avatar, a new platform for creating interactive AI avatars with the help of computer vision, NLP, and simulation technologies. The company also showed us the NVIDIA Omniverse Replicator, a synthetic data generation engine to train deep neural networks. NVIDIA also announced that Omniverse Enterprise is now generally available. For more details, refer here. Microsoft Ignite 2021 It made a host of announcements across IT products. Azure Percept- It uses AI and machine learning capabilities to provide IT developers with tools for using Azure AI technologies on edge.New capabilities of Azure Arc. It believes that Ark is the first of its kind control panel extending Azure Security Governance dev tools and managed Azure services to any infrastructure, whether it is on-premise multi-cloud or at the edge. Several new capabilities to Microsoft 365 and Microsoft Teams such as extending meeting attendees, Microsoft Teams Connect that allows users to share channels within an organisation or externally, Microsoft PowerPoint Live in Microsoft Teams, etc. For more details, refer here. IBM Think 2021 Held in May this year, IBM announced a capability to its IBM Cloud Pak for Data– a new AutoSQL technology that uses AI to automate access, and integration and management of data for AI, without having to move it. IBM Mono2Micro, which has been commercialised into IBM WebSphere Hybrid Edition, uses AI to scan and analyse enterprise applications and give recommendations on how to rewrite them for containers and microservices. Part of the announcements was the IBM Watson Orchestrate for business professionals in sales, human resources, and operations, which uses an AI engine that connects to business applications and tools like Calendar and Email. It also showed us Maximo Mobile, an easy-to-deploy mobile platform, as part of the IBM Maximo Application Suite. For more details, refer here. VMworld 2021 At the VMworld 2021 event held in October, the company announced the launch of VMware Cross-Cloud services. It said that the services will help businesses to move to the Cloud faster and provide customers with the ability to build, deploy and secure apps across any cloud. VMware introduced Cloud Infrastructure and Management for the Cross-Cloud services, advancements in the VMware Tanzu portfolio, and launched VMware Edge, among other announcements. For more details, refer here. Google Cloud Next 2021 Held in October, Google introduced a range of cloud-based features and products like: Google Distributed Cloud, a bundle of fully managed hardware and software solutions that work on edge and data centres.Google Cloud Cortex Framework and Cloud Build Hybrid– The Cloud Build Hybrid allows developers to build, test, and deploy across clouds and on-premise systems, while Google Cloud Cortex Framework allows customers to get insights and reduce time-to-value with reference architectures, packaged services, and deployment accelerators. For more details, refer here. Google for India 2021 Google conducted the 7th “Google for India” initiative recently and introduced several innovations to transform India into a leading digital economy. Some of the key ones include: Google has introduced Google Career certificates that will help fresh graduates pick up in-demand skills in IT support, IT automation, project management, data analytics and UX design.It announced the first-ever Google Assistant-enabled end-to-end vaccine booking flow in India.Teamed up with India Meteorological Department to launch weather alerts for any extreme climatic conditions.New features for payments like Groups and Bill Split on Google Pay. For more details, refer here.","excerpt":"Review the major announcements made by tech giants this year at conferences","categories":["IT Services"],"tags":["AI (Artificial Intelligence)","Deep Learning","Google","Hybrid Cloud","Meta","Microsoft","NVIDIA","Omniverse","VMWare"],"author_name":"Sreejani Bhattacharyya","publish_date":"2021-11-26T18:00:00","publication_year":"2021","word_count":737,"keywords":["VMWare","computer vision","AI (Artificial Intelligence)","R","NLP","Omniverse","analytics","NVIDIA","Meta","machine learning","AI","neural network","Hybrid Cloud","microservices","Google","SQL","Deep Learning","Azure","Microsoft"],"extracted_tech_keywords":["AI","machine learning","neural network","NLP","computer vision","analytics","Azure","microservices","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/it-services\/major-announcements-made-at-tech-conferences-in-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":56867,"title":"Inside Flipkart’s AI-Powered Bots","content":"Flipkart’s AI exploration and robotics space had led the company to adopt a technology that enabled humans and bots to work together seamlessly. India is one of the fastest-growing markets for e-commerce in the world. Buoyed by affordable smartphones, deeper penetration of internet and high consumption growth, this market is set to grow and exceed $100 billion by 2022. Catering to this rapid growth, e-commerce companies Amazon and Walmart-owned Flipkart are turning to technology-enabled solutions to deliver smarter, faster and better. While Amazon has been leveraging data and machine learning to keep its warehouse operations running smoothly, Flipkart had adopted a robot-based sortation technology – a first for India – at its sortation centre in Bengaluru to handle increasing order volumes. Flipkart’s AI-Powered Bots Flipkart’s AI exploration and robotics space had led the company to adopt a technology that enabled humans and bots to work together seamlessly. Nearly 350 AI-powered bots – monikered Automated Guided Vehicles (AGVs) – help operators process ‘4,500 shipments an hour at twice the speed and with 99.9% accuracy’. Not just that, the company claims that with the emergence of these sophisticated robots, warehouse storage capacity and throughput has doubled as well. What is more, these compact bots are easy to maintain and quick to deploy, giving Flipkart the flexibility to create nimble distribution networks. The e-tailer had used it extensively during its flagship pre-Diwali sale Big Billion Day last year. These bright orange robots – about the size of a Roomba – are loaded with sensors which ensures that they do not bump into each other, but rather, operate like self-driving vehicles on the road. Programmed to run 7-8 hours on a single charge, these self-charging units automatically guide themselves towards designated charging points when batteries are down. The AGV setup operates on three levels – ground level, where packages are sent; level two, where bagging operations are conducted; and level three, where packages are sorted according to the addresses. Once a customer places an order on Flipkart, the package is dispatched from a warehouse to this sortation facility. It is then transported on a trolley to the ground floor, following which it is loaded by operators on to a conveyor belt from where it reaches level two. Here, the package is picked up manually and placed on a bot, barcode side up. The AGV then passes under a barcode scanner, where its algorithm identifies the address and thereafter, directs it to the designated chutes. The package is then dropped, bagged and sent on its way to the customer. In fact, the complete life-cycle of this package at the centre is a mere one and a half hours! This complex AGV project was completed in just five months and had been going through various iterations since 2018. Not only does it offer speed, accuracy and safety, the tech setup is also space-efficient since it employs vertical space utilisation – a unique feature in traditional sortation centres. What is more, it also reduces multiple touchpoints that typically happens during the sorting process, minimising the chances of damage. Amazon Amazon’s fulfilment network comprises a variety of building types and sizes. Each of these is set up like a Manhattan-style grid and houses 1-4 million bins on the order of 10 million packages. The company uses computer vision systems to analyse images to securely monitor and regulate all key operations. The company’s Chief Scientist for Worldwide Operations Russell Allgor has helped build systems guided by AI and ML that manages and delivers customer orders efficiently and effectively across its 175 fulfilment centres. Allgor first makes decisions around which packages to pick at the same time in order to get the items on the same box. According to him, this is a large optimisation problem and using the resulting information, his team tries to minimise the distance that needs to be covered when making the deliveries. Amazon has been using decision engine, decision logic and AI to make those decisions in real-time on a constant basis. Using these technologies, the company makes predictions around the likelihood of when it would need access to the delivery vehicle in say, the next hour, two hours, three hours, and so on. In simple terms, it uses ML to build information around how long it takes to travel from point A to point B.","excerpt":"Flipkart’s AI exploration and robotics space had led the company to adopt a technology that enabled humans and bots to work together seamlessly. India is one of the fastest-growing markets for e-commerce in the world. Buoyed by affordable smartphones, deeper penetration of internet and high consumption growth, this market is set to grow and exceed […]","categories":["AI Features"],"tags":["AI and ML","Amazon","ecommerce","Flipkart"],"author_name":"Anu Thomas","publish_date":"2020-02-18T17:00:00","publication_year":"2020","word_count":721,"keywords":["Go","API","machine learning","ELT","AI","AI and ML","ML","Flipkart","Amazon","computer vision","RAG","Aim","ecommerce","R"],"extracted_tech_keywords":["AI","machine learning","ML","computer vision","Aim","RAG","R","Go","API","ELT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ecommerce-of-the-future-inside-flipkarts-ai-powered-bots\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":48655,"title":"IIT Hyderabad Researchers Develop Low Power Chips For AI Devices","content":"Researchers from the Indian Institute of Technology, Hyderabad, have developed low power chips that can be used in artificial intelligence-powered devices. These researchers have developed Magnetic quantum-dot cellular automata (MQCA) based nanomagnetic logic architectural design methodology of approximate arithmetic circuits. The Researchers are working towards a vision of realizing resource-constrained Magnetic Chips for Ultra low power portable AI applications. Many modern systems such as speech and face recognition systems and IoT-enabled devices for remote health monitoring require highly computationally and energy-intensive neural networks. That is why, it is not practically affordable to perform these computations in the portable hand-held devices. With these major limitations, all the machine learning algorithms used in these AI applications run on remote systems, said IIT-H in a statement. Speaking about the outcomes and benefits of this Research, Dr Amit Acharyya, Associate Professor, Department of Electrical Engineering at IIT-H, said, “We have computationally modelled, designed and implemented an arithmetic adder, subtractor and add\/sub using nanomagnets which are the basic building blocks of performing AI computing. We are aware that the emerging edge computing devices are handy in size as well as requiring low-power computation and are also tolerant to feeble decrease in precision. The reported work of ours’ targets such devices, where there is a significant investment in the research towards making it low power without compromising on accuracy too much. Performing AI computing on edge with approximate nanomagnetic logic deployed on the magnetic ICs is an attempt towards the futuristic computations. I hope this work paves the way towards achieving such a vision.” The research was undertaken by a team comprising Santhosh Sivasubramani, PhD Scholar, Advanced Embedded Systems and IC Design Laboratory, Department of Electrical Engineering; Dr Acharyya and Dr Chandrajit Pal, Post-Doctoral Research Fellow. The Research has been published in the reputed peer-reviewed journal by Nanotechnology (Prestigious journal of Institute of Physics).","excerpt":"Researchers from the Indian Institute of Technology, Hyderabad, have developed low power chips that can be used in artificial intelligence-powered devices. These researchers have developed Magnetic quantum-dot cellular automata (MQCA) based nanomagnetic logic architectural design methodology of approximate arithmetic circuits. The Researchers are working towards a vision of realizing resource-constrained Magnetic Chips for Ultra low […]","categories":["AI News"],"tags":["iit hyderabad","nanotech","nanotechnology"],"author_name":"Prajakta Hebbar","publish_date":"2019-10-23T13:16:46","publication_year":"2019","word_count":308,"keywords":["Go","artificial intelligence","machine learning","programming_languages:R","AI","neural network","iit hyderabad","programming_languages:Go","nanotech","edge computing","nanotechnology","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","edge computing","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-hyderabad-researchers-develop-low-power-chips-for-ai-devices\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":34314,"title":"Turing Award Winner Judea Pearl Believes To Build Truly Intelligent Machines, Researchers Should Teach Them Cause &#038; Effect","content":"Judea Pearl has been a prominent figure, an  ACM A.M. Turing Award Winner computer science’s highest honour, a person who made it possible to create a viable ecosystem for Artificial Intelligence to thrive. That what we are seeing today is the causal effect of the efforts that he led in the 1980s to allow machines to do the reasoning probabilistically. But now he is one of the most critical critics of AI in the world. A book that he authored in the year 2018  The Book of Why: The New Science of Cause and Effect, argues that AI has been handicapped by an incomplete understanding of what intelligence really is. By using a scheme called Bayesian networks, he figured out how to program machines to associate a potential cause to a set of observable conditions, which was the primary challenge in the field of AI research. For example, when a patient who returned from Africa with a fever and body aches, Bayesian networks made it practical for machines to say that it was malaria. According to Pearl, the probabilistic association is what is causing and disrupting the efficacy of Artificial intelligence at large. The latest breakthroughs are all around only projecting machine learning and neural networks only.AI use cases are all hyped up around computers that can intelligently master ancient games or drive a car through AI- based automation. In his own words what he believes is that the state of the art in artificial intelligence today is merely just a souped-up version of what machines could already do a generation ago that is finding hidden regularities in a large set of data. All the impressive achievements of deep learning amount to just curve fitting. The book elaborates a vision for how truly intelligent machines would think and gives a completely different angle to analyse and judge Artificial Intelligence at large. What The Turing Award Winner Proposes? Pearl argues to replace reasoning by association with causal reasoning. Instead of the mere ability to correlate fever and malaria, machines need the capacity to reason that malaria causes fever. Pearl is of the opinion that once this kind of causal framework is in place, it becomes possible for machines to ask counterfactual questions to inquire how the causal relationships would change given intervention which Pearl views as the cornerstone of scientific thought. Pearl also proposed a formal language in which to make this kind of thinking possible a 21st-century version of the Bayesian framework that allowed machines to think probabilistically must be developed. Pearl expects that causal reasoning could provide machines with human-level intelligence so that they could be able to communicate with humans more effectively to achieve the status as moral entities with a capacity for free will. Judea Pearl’s breakthroughs in the 1980s made him an AI pioneer. Now he believes the field has gotten too caught up in the very methods he helped devise.https:\/\/t.co\/iZgb3aDsRB pic.twitter.com\/IkZvxggO0S — Quanta Magazine (@QuantaMagazine) May 15, 2018 In Conclusion Humans must find a way to equip computer machines with a newer model of an adaptable ecosystem. Humans should not expect the computer machine to exist in that reality and behave in an intelligible manner when it’s not given the much-needed access to a comprehensive model of analysing and understanding the reality. Pearl in the book overall justifies the need of why that AI has been handicapped by an incomplete understanding of what intelligence really is. Causal reasoning is a cornerstone in explainable-AI. The prime reason to stress upon this factor is that of his belief that in future Robots will attain some amount of free will while making decisions, a fact he says that cannot be denied. All these will come to public visibility the day a when a robot starts defying or ignoring software components consistently and starts selectively to be listening to certain sets of software components only. The other components that will be ignored, will be the ones that are maintaining norms of behaviour that have been programmed into them or are expected to be there based on past learning. So, it’s a moral as well as an ethical duty of the global  AI industry to start finding solutions to build intelligent machines and teach them to identify and analyse the real cause and its subsequent effect.","excerpt":"Judea Pearl has been a prominent figure, an  ACM A.M. Turing Award Winner computer science’s highest honour, a person who made it possible to create a viable ecosystem for Artificial Intelligence to thrive. That what we are seeing today is the causal effect of the efforts that he led in the 1980s to allow machines […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","judea pearl"],"author_name":"Martin F.R.","publish_date":"2019-01-31T11:13:20","publication_year":"2019","word_count":716,"keywords":["Go","machine learning","artificial intelligence","programming_languages:R","AI","neural network","automation","deep learning","AI research","judea pearl","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","R","Go","automation","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/turing-judea-pearl-machine\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":66740,"title":"CISOs In India Should Use Machine Learning For Security Assessment: Adam Palmer, Tenable","content":"As the world gets swallowed by the COVID-19 pandemic, cyberattacks have risen to become a critical area for all tech-enabled companies around the globe. In the WFH context, malicious hackers have been utilising various tactics to steal valuable and sensitive corporate data. As far as India is concerned, it has become one of the most targeted nations worldwide, and attackers have been targeting critical infrastructure and data assets in sectors like government, banking, defence, manufacturing, software and others in the country. According to Microsoft, there were in excess of 9,000 COVID-19 themed attacks in India between February 2 and May 2. The persistent attacks show the level of interest that cyber hacking groups have in India. To deep dive into the critical issue, Analytics India Magazine connected with Adam Palmer, Chief Cybersecurity Strategist at Tenable. During the interview, Adam talks about why there is an urgent need among Indian security managers to invest in building cyber resilience, gathering and utilising threat intelligence to mitigate attacks. Here are the excerpts from the interaction with Adam Palmer from Tenable: AIM: In your view, how has cybersecurity evolved in India? Adam Palmer: In recent years, digital transformation has become a driving force propelling Indian organisations to be more agile and competitive. Technology advances have brought many new benefits such as speed and increased capability for corporate networks. However, the newly expanded attack surface now includes cloud, Internet of Things (IoT), personal devices, and even operational technology like industrial controls. This has given rise to a massive barrage of thousands of vulnerabilities that are overwhelming security teams. Yet, in the face of this challenge, many organisations are still relying on legacy tools and processes that are inadequate to navigate the complex threats in today’s dynamic and modern computing environment. As cybersecurity increasingly becomes central to each organisation’s business strategy, there is a stronger need for security leaders to understand vulnerabilities in the context of the business and highlight the areas that impact the organisation. That means ad-hoc vulnerability scanning and a ‘check-the-box’ approach will no longer cut it. Cybersecurity programs in India should evolve to take a risk-based approach that means organisations should focus on the vulnerabilities that matter the most. Address true business risks instead of focusing on every one of the thousands of flaws that have a low likelihood of being exploited. AIM: What new security challenges have COVID-19 created for Indian companies? Adam Palmer: The sudden shift to a remote-work model means that employees are now combining personal technology with work networks, and this is contributing to an expanded attack surface. Many of these devices may also be older or unsecured, and this introduces serious new risks. All of this can be challenging for security teams who now have to manage this expanded and complex attack surface. We’ve already seen many phishing scams, misinformation, and fraudulent work-from-home opportunities for hackers making their way around the internet. These risks potentially put the wider corporate network at risk. Therefore, organisations need to ensure that a robust security plan is in place to address these threats. Good basic cyber hygiene practices such as maintaining systems, blocking malicious sites and IP addresses, enforcing multi-factor authentication, and using encryption are good places to start. AIM:  As the topic of cybersecurity dominates in the context of a large majority of people working from home, how can organisations truly understand where they’re exposed? Adam Palmer: The wide range of vulnerabilities and the diverse ways attackers can target them make vulnerability management a critical component of any cybersecurity program. Organisations should continually assess their networks for security vulnerabilities. This can prevent a range of problems such as unauthorized access to applications and identifying underlying software flaws that expose sensitive data. Vulnerability scanners can help identify these concerns, making it easier to understand if systems have critical risks that need to be addressed. AIM: There are thousands of security vulnerabilities in software and systems these days. How can security teams keep up? Adam Palmer: Security teams need to prioritise remediation efforts based on actual critical cyber risk. This means identifying critical assets and combining this with a clear understanding of the vulnerability severity and likelihood of exploitation. These three key elements are essential to securing the threat landscape. This approach filters out lower-risk vulnerabilities. It allows security teams some breathing room to remediate the most business-critical security issues and focus on vulnerabilities which are being actively exploited by threat actors rather than the thousands that might only theoretically be used. If everything matters, then nothing matters. Security leaders need to follow an approach that prioritises risks that matter. That is the key to success. AIM: How is Tenable working with security professionals to tackle the challenges? Adam Palmer: We know that managing risk during these challenging times can be problematic. Apart from being readily available for our customers, our Tenable Research team is working continuously to publish the latest research on cybersecurity, phishing attempts, and other opportunistic attack behaviours. We combine this with our machine learning capabilities and thousands of data sources so that our customers can stay aware and have clarity about the critical risks that matter. AIM: Going forward, what is your outlook for cybersecurity readiness of Indian companies in the world of constantly evolving threats? Adam Palmer: India’s cybersecurity needs are not different from the rest of the world. When you analyse the vast majority of breaches that occur, whether they’re in India or globally, most of them are caused by known but unpatched vulnerabilities. Practising basic cyber hygiene like patching systems and utilising strong authentication can significantly reduce the risk of compromise on critical networks. Doing this makes sure that companies can identify vulnerabilities and exposures before any asset or data is compromised. This enables organisations to make the necessary corrections to mitigate these weaknesses. At the same time, as the threat landscape expands, there’s a stronger need for CISOs to recognise vulnerabilities in the context of business risk and utilise data to prioritise cybersecurity efforts.  We at Tenable believe that there will be a turn in India’s cybersecurity industry towards a risk-based strategy to vulnerability management which applies machine learning analytics to correlate vulnerability severity, threat actor activity and asset criticality to classify and manage issues posturing the biggest business risk. This innovative approach will support Indian organisations’ focus on the vulnerabilities that weigh the most and mitigate true business risk, instead of focusing on flaws that have the lowest possibility of being exploited.","excerpt":"As the world gets swallowed by the COVID-19 pandemic, cyberattacks have risen to become a critical area for all tech-enabled companies around the globe. In the WFH context, malicious hackers have been utilising various tactics to steal valuable and sensitive corporate data.  As far as India is concerned, it has become one of the most […]","categories":["AI Features"],"tags":["Cyber Attack","Cyber Security","Cybersecurity","Cybersecurity India","Interviews and Discussions","latest technology in machine learning"],"author_name":"Vishal Chawla","publish_date":"2020-06-05T13:00:00","publication_year":"2020","word_count":1079,"keywords":["Go","Cyber Security","machine learning","AWS","AI","Git","RAG","Cyber Attack","Aim","Cybersecurity India","latest technology in machine learning","analytics","GAN","Cybersecurity","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","AWS","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/indian-cisos-should-use-machine-learning-for-security-assessment-adam-palmer-tenable\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10119063,"title":"PyTorch Releases Version 2.3 with Focus on Large Language Models and Sparse Inference","content":"PyTorch announced the release of version 2.3, introducing several new features and improvements for performance and usability of large language models and sparse inference. The release, which consists of 3,393 commits from 426 contributors, brings support for user-defined Triton kernels in torch.compile. This feature allows users to migrate their existing Triton kernels without experiencing performance regressions or graph breaks. The feature also allows Torch Inductor to precompile user-defined Triton kernels and organise code more efficiently. Another feature, Tensor Parallelism for efficient training of large language models. It facilitates various tensor manipulations across GPUs and hosts, integrating with FSDP (Fully Sharded Data Parallel) for efficient 2D parallelism. The PyTorch team has validated Tensor Parallelism on training runs for models with over 100 billion parameters, demonstrating its effectiveness in handling large-scale language models. PyTorch 2.3 introduces support for semi-structured sparsity, specifically 2:4 sparsity, by implementing it as a tensor subclass. This feature enhances performance, achieving up to 1.6 times faster processing than dense matrix multiplication, and includes advanced functionalities like mixing different data types during quantization, uses improved versions of cuSPARSELt and CUTLASS kernels, and is compatible with torch.compile for more efficient computation. Compared to the previous version, PyTorch 2.2, which brought advancements like the integration of FlashAttention-v2 and the introduction of AOTInductor, PyTorch 2.3 builds upon these improvements and introduces new features specifically targeted at large language models and sparse inference. With significant contributions from a large and active community, this version brings features like user-defined Triton kernels and Tensor Parallelism to collectively improve performance, scalability, and flexibility.","excerpt":"The new release has added features to meet the needs of the AI and machine learning community.","categories":["AI News"],"tags":["LLMs","Pytorch"],"author_name":"K L Krithika","publish_date":"2024-04-25T13:25:59","publication_year":"2024","word_count":257,"keywords":["Pytorch","ELT","programming_languages:R","PyTorch","LLMs","AI","Scala","ai_frameworks:PyTorch","programming_languages:Scala","GAN","R"],"extracted_tech_keywords":["AI","PyTorch","R","Scala","ELT","GAN","ai_frameworks:PyTorch","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/pytorch-releases-version-2-3\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":61053,"title":"BRIDGEi2i Launches COVID-19 Global Monitor To Share Insights Into The Pandemic","content":"In the wake of the COVID-19 pandemic, BRIDGEi2i has launched a COVID-19 Global Monitor to share insights into the pandemic and its impact on sectors\/ economies\/ communities etc. In a recent LinkedIn post, Prithvijit Roy, the CEO and Co-founder at BRIDGEi2i Analytics Solutions, stated that this is an attempt to “create the First Version of the COVID-19 Monitor.” The post stated — “What’s happening in the world is unprecedented. We hope we will get back to the normal world in some time. As analytics practitioners, we thought we would leverage our expertise to share insights from what we see as the emerging trends and patterns from the COVID data .. hence our attempt to create the First Version of the COVID-19 Monitor.” The COVID-19 Global Monitor collates information from various industries\/ sources and visualises their impact for the sectors\/economy\/community as the case might be. According to their website, “BRIDGEi2i has been keenly watching the developments regarding the COVID-19 pandemic and have been analysing it with some data-centric curiosity. We feel that the availability of data improves the understanding of any problem before us and can help us fight it better.” And that’s why by studying the recent trends carefully, the company has developed this COVID-19 Global Monitor. Further BRIDGEi2i’s website stated — “Countries have been trying hard to manage this pandemic through a policy of wider diagnosis, treatment, and containment but to limited success. Since no early cure seems in sight, it is widely thought that a policy of isolation can buy more time and probably develop some herd-immunity among the population that can flatten the curve. This has led to measures of social distancing, lockdowns and travel and several other restrictions being put in place across geographies—with a direct impact on trade, commerce, industry, and the economy. The financial implication of these measures is thought of to be as great as those seen during the World Wars.” This comprehensive dashboard will, therefore, share insights on the global preview, along with country comparisons, India overview and Indian demographics.","excerpt":"In the wake of the COVID-19 pandemic, BRIDGEi2i has launched a COVID-19 Global Monitor to share insights into the pandemic and its impact on sectors\/ economies\/ communities etc. In a recent LinkedIn post, Prithvijit Roy, the CEO and Co-founder at BRIDGEi2i Analytics Solutions, stated that this is an attempt to “create the First Version of […]","categories":["AI News"],"tags":["bridgei2i","Coronavirus","covid-19"],"author_name":"Sejuti Das","publish_date":"2020-04-06T17:15:00","publication_year":"2020","word_count":338,"keywords":["covid-19","AI","programming_languages:R","RAG","Coronavirus","analytics","bridgei2i","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bridgei2i-launches-covid-19-global-monitor-to-share-insights-into-the-pandemic\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10161945,"title":"AI and India’s Developer Surge Will Spark Innovation in 2025","content":"India’s developer community reached new heights in 2024, both in terms of growth and innovation. The number of developers on GitHub in India surpassed 17 million, which is more than the entire population of Bengaluru city, making it one of the fastest-growing communities on the web. Not just that, the contributions to public generative AI projects increased by 95%, while India’s developer community continued to embrace AI-developer tools, setting the stage for even greater breakthroughs in 2025. With a surging developer community and the large-scale adoption of AI tools, here’s how India’s developer community will make its mark on the world in 2025. Indian Devs Hold the Power for Large-Scale Transformation The sheer size of India’s vibrant developer community signals incredible prosperity for 2025—after all, where developers thrive, economic growth follows. Combined with the power of AI tools like GitHub Copilot, India’s developer community is poised to fuel a fresh wave of homegrown tech multinationals, a new generation of disruptive startups, and an empowered open-source community like never before. These communities will leverage AI developer tools to shape global innovation and create digital solutions that benefit society—much like the impactful work being done by Open Healthcare Network in India. To turn this vision into reality, GitHub is empowering every Indian developer to unlock the potential of AI-powered software by offering GitHub Copilot Free in VS Code. With the community shipping software up to 55% faster, immense digital and economic progress lies ahead for India. Imagine a country where each developer, equipped with AI tools like GitHub Copilot, contributes to the nation’s growth by solving problems for their society. They educate people about technology, contribute to open-source digital public goods, and create newer ways to drive large-scale transformation. It’s a New Era of ABCD: AnyBody Can Develop! AI has successfully created a bridge between humans and machine languages. This will empower children in India to start programming in their native languages even before learning English in schools. Be it Hindi, Kannada or Marathi, aspiring developers can now write and understand code—traditionally a complex abstraction layer—in natural language. With AI-powered tools such as GitHub Copilot, students can learn to code in their own language, using AI as a personal programming assistant, much like a calculator for coding. This will add millions more to India’s rising developer community, cementing its place as the largest developer hub in the world. More importantly, it will ensure that every Indian student can explore STEM careers and express their creativity through code without needing to learn English first. Digital Diwali: Where Ideas Ignite with the Power of AI AI is ready to empower anyone to turn their ideas into reality, all in natural language, leading to an explosion of creativity in India and across the globe. This has never been more possible than with GitHub Spark, an AI-powered tool for creating and sharing micro apps (“sparks”) tailored to individual needs and preferences. While still in technical preview, it’s going to build a world where anyone and everyone is empowered to create or adapt software for themselves. If India continues to nurture its developer community and grow it at scale while simultaneously embracing the transformative power of AI, the nation will not only cement its place as a global AI leader but also extend the economic opportunity of building software to all its people. The promise of this new future is entirely possible—and 2025 will be a pivotal year in paving the way.","excerpt":"The sheer size of India’s vibrant developer community signals incredible prosperity for 2025—after all, where developers thrive, economic growth follows.","categories":["AI Trends"],"tags":["AI coding","AI in Coding","coding platform","GitHub","Impact of AI Coding in Indian Graduates","Indian Developers","STEM"],"author_name":"Karan M V","publish_date":"2025-01-22T14:00:00","publication_year":"2025","word_count":575,"keywords":["Go","AI coding","STEM","AI","RPA","innovation","Git","Impact of AI Coding in Indian Graduates","AI in Coding","RAG","ViT","generative AI","Indian Developers","coding platform","GitHub","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","Git","GitHub","ViT","RPA","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/ai-and-indias-developer-surge-will-spark-innovation-in-2025\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077012,"title":"Backend-As-A-Service Platforms Want To End Firebase Supremacy. They Might Succeed","content":"“In Google, We Trust” is the official motto of app developers. Adesh Verma, a software developer and a tech consultant, hosted an app on Google’s Firebase during the initial days of his project. His platform grew from a few hundred to a few thousand customers when he decided to move data to an SQL database. It was a huge revelation for Verma that he was stuck in a rabbit hole which seemed impossible to come out of. Firebase has been a leader in the backend-as-a-service (BaaS) space. It offers various features such as analytics, crashlytics, performance monitoring and others. But, the most critical features that the app developers use are user authentication, database and server-side functions. These infrastructure pieces are hosted by Google Cloud platform. Firebase SDK weekly download Besides, it also provides SDKs for frontend applications like JavaScript, iOS, Android, Flutter, Unity and others. So, developers can access them without having to architect, deploy and maintain their own server-side which requires money, muscle and time—all. Since building a backend for hosting mobile apps wasn’t\/isn’t easy, for many years, the only game in town was Firebase that provided BaaS. There’s a list of developer-friendly features in the platform, but it still isn’t a complete platform. It has limited support for iOS features, unpredicted pricing, real time synchronisation issue, limited querying capabilities and above-all data migration problems. Despite its flaws, Firebase had built its monopoly in the market. But, that’s not the case any longer. There are multiple BaaS, which are now encroaching the ground that Firebase held for many years. Amazon’s answer to Firebase ‘Amplify’, a product of AWS, has a very impressive list of features that allow developers to enable analytics, authentication, Geo location, predictions, push notifications, file storage, data storage and add AI\/ML to the front-end web or mobile application. Amplify supports iOS, Android, Flutter, React Native for web development and integrates with popular Javascript frameworks like Vu, React and Next.js. Developers can use only the tools they need in an existing app or use them all together. AWS Amplify also provides many tools and services such as Studio, Libraries, CLI, UI Components and Web hosting. With all these tools, developers can build the apps’ back-end and front-end UI, connect their apps to AWS backends and manage users hosted outside AWS console. Amplify has been gradually climbing up the ladder to become the second biggest popular BaaS platform. A backend-as-a-service (BaaS) is an extremely complicated product for any company to offer. To solve such a complicated problem, there are now other platforms which offer various solutions of their own. Supabase: Database makes the difference Supabase has been around for two years now and has core features like authentication, file storage and serverless function, but the most distinguishing feature is its database. It uses ‘Postgres’, an open-source object-relational database system. This provides several advantages such as flexibility in data modelling and freedom from vendor lock-in (developers can host a Postgres database anywhere). However, there are certain drawbacks with the platform. It doesn’t provide website hosting features and its SDKs are only focussed on the web with the exception of app-based frameworks like React Native, Flutter and Ionic. Despite being focused on the web, Supabase gains its popularity as it makes relational databases easy to work with. It provides developers with a browser-based tool to manage the data and integrates nicely with user authentication where they can implement low-level security policies to have access control over data. Postgres also supports GraphQL, a Query language for reading and mutating data for APIs. While Supabase is trying to solve the problem of developers via Postgres, nHost is building a Firebase alternative with GraphQL. nHost: Postgres database is also GraphQL api nHost provides all the necessary building blocks for contemporary software, such as a PostgreSQL database, a real-time GraphQL API, authentication, storage, and serverless functions that let businesses run custom code. These building blocks are available for most popular front-end frameworks, including React, Flutter, and Vue. Additionally, nHost provides managed cloud solutions with options for professionals, businesses, and amateurs. A project built on Hasura can take a relational database and turn into a GraphQL API. The UI is nice and intuitive. It allows developers to visualise all their database records in the browser. The most special feature of the platform is that the Postgres data is also a GraphQL API. As mentioned previously, BaaS is an extremely completed product for any company to offer. In an effort to disrupt the monopoly of Firebase, different platforms are trying to solve the varied problems of developers. The attempts of these platforms are laudable, but it could also be interesting to witness if any of these platforms would actually be able to surpass Firebase in the future.","excerpt":"There are multiple backend-as-a-service, which are now shaking the ground which Firebase was holding for many years.","categories":["IT Services"],"tags":["Android","flutter","Google","iOS","Javascript","unity"],"author_name":"Tausif Alam","publish_date":"2022-10-11T16:00:00","publication_year":"2022","word_count":791,"keywords":["PostgreSQL","AWS","AI","iOS","Javascript","ML","JavaScript","serverless","RAG","Android","analytics","Google","SQL","unity","R","flutter"],"extracted_tech_keywords":["AI","ML","analytics","RAG","AWS","serverless","PostgreSQL","R","SQL","JavaScript"],"url":"https:\/\/analyticsindiamag.com\/it-services\/backend-as-a-service-platforms-want-to-end-firebase-supremacy-they-might-succeed\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":7423,"title":"Analytics Maturity Assessment: F.I.R.E 4X4 Matrix","content":"The speed at which analytics industry is evolving is almost awe inspiring. Yet, it’s evident that we are at the stage of maturity where standard practices and ripeness of analytics processes are far from reality. Within the whole buzz of how essential analytics is today than ever for organizations, there is a huge confusion around different jargons that are currently floating around. The confusion limits the ability of organizations to view analytics as a single cohesive thread that binds all their business units with the charter of intelligent decision making. This framework rationalizes how mature an organization is in its ability to incorporate the full potential of analytics. The model measure organization’s current strengths and weaknesses across 4 parameters essential for a robust analytics strategy. The model helps you take an honest look at how analytics is being utilized within your organizations and what steps can be taken to improve your analytics usefulness. Focus: (Strategy & Vision) Organizations today understand that analytics is a key factor to lasting competitive advantage, yet few organization go on to identify independent, well rounded use cases around analytics. Most of the work around data science is done in piece-meal fashion; wherein management tends to look at data science as an experiment. Even where analytics shows promise, little advancement is made into the area. While various reason can be attributes to why analytics does not become a part of how businesses operate; the biggest is absence of management focus. This often leads to a lack of sponsorship from stakeholders, stalling key analytics projects. Often, management considers analytics initiatives in similar light as IT projects. Not enough focus is given to align analytics to organizational strategy. Broad organizational introspections lead to an insight into management focus. Integration: (Structure & Deployment) Data Science initiatives starts in silos within various business units. This works well; and is even efficient; in cases where organizations are just getting started on their analytics journey. Yet, as organizations move towards a robust analytics strategy, providing an efficient structure to the data science teams becomes critical. At Kruxonomy, we support both the centralized and de-centralized structure based on the maturity and needs of the organization. A robust structure helps in reducing   disparate analytics processes and minimize sub-optimal utilization of resources. It also helps in sharing of best practices from the different functions and improve the way data scientists tackle various problem statements. Not just how data science teams are structured, but this parameter also access how well knit is the data science team with the operational and management teams. This may include the organization-wide alliance of data scientists, an analytics center of excellence that fosters research and discovery in data science. Rigour: (Competency & Governance) The amount of new developments that is currently happening around data science and machine learning is almost overwhelming. The pace is at times incomprehensible by even seasoned data scientists and organizations that have remained on top of the analytics game. Given your investment and conviction towards analytics, your current analytics initiatives and solutions will be assessed on the level of rigor against industry standards. This would mean an assessment on the maturity of your analytics itself. Are your current analytics solutions effectively maximizing the return on your data? Are you data scientists spending more time on mundane reporting when there are opportunities to automate most of it? Are there opportunities for predictive modelling that are untapped? Enablers: (Technology & Talent) Few companies have organized their data with analytics as an end goal. Legacy systems often tend to create data that have operational and quality issues. This has led to contemporary rant over massive centralized data stores i.e. data lakes often using technologies like Hadoop. Not all organization need big data infrastructure, yet centralized, quality data is key to analytics success. This is true with analytics technology and talent within enterprises. Data science is a fast evolving area and organization need to quickly stay connected with the developments to be competitive. Yet, skills and technology needs to upgraded at a much faster pace. Increasingly, there’s more that can be done in analytics from same amount of investment than it was possible just 2-3 years back. The cost of ownership in data science projects would continue to decrease over the years. It is best to utilize this to upgrade the technology and train your data scientists. To download the complete whitepaper, visit here.","excerpt":"The speed at which analytics industry is evolving is almost awe inspiring. Yet, it’s evident that we are at the stage of maturity where standard practices and ripeness of analytics processes are far from reality. Within the whole buzz of how essential analytics is today than ever for organizations, there is a huge confusion around […]","categories":["IT Services"],"tags":[],"author_name":"Дарья","publish_date":"2015-05-18T13:37:10","publication_year":"2015","word_count":733,"keywords":["big data","data science","Go","machine learning","programming_languages:R","AI","analytics","GAN","R","data lake"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","R","Go","big data","data lake","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/analytics-maturity-assessment-f-i-r-e-4x4-matrix\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10023949,"title":"Top AI &#038; Analytics Appointments At Major Tech Firms In 2021","content":"The whole world came to a standstill after the COVID breakout. So much so that the entire 2020 and the first half of 2021 seems like a blur. However, the tech companies have seen a change of guard at the top management amid the pandemic. Below we look at the new hires at tech companies in 2021. Amazon appoints former Tableau CEO to lead cloud business Tableau CEO Adam Selipsky has been appointed as the head of Amazon’s cloud business. He will replace Andy Jassy, who is taking up the CEO mantle when Jeff Bezos steps down in a couple of months. Selipsky has earlier led the marketing, sales and support business of AWS as its vice president for 11 years. Selipsky had quit AWS to join Tableau as its CEO five years ago. Read the news here. Lenskart hires Head, Analytics & ML Lenskart.com has hired Saurabh Agarwal as Head, Analytics and ML. He will bring more than a decade of his experience in data science and AI to build a strong analytics team at Lenskart. He had worked with companies such as MothersonSumi Infotech & Designs Ltd (MIND), Tata Insights and Quants and American Express, driving their analytics strategies. Read the news here. Google Cloud appoints Asia Pacific leader Google Cloud has appointed Karan Bajwa as its new Google Cloud leader for the Asia Pacific. He will lead all regional revenue and go-to-market operations for Google Cloud, including Google Cloud Platform (GCP) and Google Workspace. Bajwa will directly report to Rob Enslin, the President of Sales at Google Cloud. Bikram Singh Bedi will take his place as the Managing Director of Google Cloud, India. Read the news here. Intel announces new CEO Former VMWare CEO Pat Gelsinger has replaced Bob Swan as Intel’s CEO. Gelsinger has also joined the Intel board of directors. “Gelsinger is a proven technology leader with a distinguished track record of innovation, talent development, and deep knowledge of Intel. He will continue a values-based cultural leadership approach with a focus on operational execution,” stated Intel. Read the news here. Google appoints new AI lead Dr Marian Croak has been appointed to oversee Google’s research department on responsible AI. Before this, Croak was heading the engineering division of the company as a Vice President. Croak manages the teams working on artificial intelligence for social good, algorithmic fairness, and AI ethics. She reports to the SVP of Google AI Research and Health. Read the news here. Salesforce appoints Senior VP and MD, Sites (India) Salesforce has appointed Sanket Atal as the Senior Vice President and Managing Director, Sites (India). Atal will drive the next phase of Salesforce’s growth in India with design thinking and digital strategies. Read the news here. Google Cloud appoints new leader for India business For the post of managing director of their India Business, Google cloud has selected Bikram Singh Bedi, who had been the head of AWS business in India and South Asia for about six years. He will replace Karan Bajwa, the recently appointed vice president of the Asia Pacific cloud, AWS. Read the news here. GitHub appoints first-ever chief security officer GitHub has hired Mike Hanley as its first chief security officer. Hanley has earlier led the Duo security program. In GitHub, he would focus on the development of developer-first security. Read the news here. Wipro hires CTO Wipro has appointed Subha Tatavarti as the Chief Technology Officer. She would oversee Robotics, Service Transformation, Technovation Centre, Open Innovation, Topcoder, SVIC and Applied Research teams. Tatavarti has worked with companies like Walmart and PayPal. She is an expert in simplifying analytics infrastructure and modernising initiatives on cloud frameworks and developer platforms. Read the news here. Teradata appoints country manager for India The warehouse platform of data from multiple clouds, Teradata has appointed Ashok Shenoy as the country manager, India. He has over 25 years working with technology giants like Microsoft and Cisco. Shenoy aims to help Teradata to build ecosystems with robust partners and also broaden the client base. Read the news here.","excerpt":"Check out the new hires at tech companies in 2021.","categories":["AI Trends"],"tags":[],"author_name":"Satavisa Pati","publish_date":"2021-04-14T14:00:00","publication_year":"2021","word_count":673,"keywords":["data science","Go","artificial intelligence","GCP","AWS","AI","ML","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","data science","analytics","Aim","AWS","GCP","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ai-analytics-appointments-at-major-tech-firms-in-2021\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":26466,"title":"Microsoft 365 Introduces AI-Backed Capabilities – Check Out The New Features","content":"As part of its commitment to strengthening AI capabilities, Redmond tech bellwether Microsoft introduced new AI-backed capabilities in Microsoft 365, a year after its launch. Microsoft, a leader in cloud-based software suite is making the workforce smarter by giving them the tools to work smarter. Here are a slew of features with AI-infused capabilities. Microsoft Teams The software giant introduced Microsoft Teams, also in the free version that enables unlimited chat messages and search and also allows app integrations with 140+ business apps to choose from—including Adobe, Evernote, and Trello. Besides the company, also enables enterprises to create live and on-demand events in Microsoft 365. The company blog enumerated how AI-powered capabilities can help create an on-demand event. Microsoft Events A speaker timeline, which uses facial detection to identify who is talking, so you can easily jump to a particular speaker in the event. Speech-to-text transcription, timecoding, and transcript search, so you can quickly find moments that matter in a recording. Closed captions to make the event more accessible to all. Workplace Analytics Solution The company also announced a new Workplace Analytics solution, that leverages collaboration insights from the Microsoft Graph. This helps teams run meetings, create time for focused work. Through this feature, companies can use aggregated data in Workplace Analytics to identify opportunities for improving collaboration, then share insights to their respected departments via MyAnalytics In an attempt to streamline work-life balance, the company will roll out “nudges”, powered by MyAnalytics in Microsoft 365, that will flag emails on certain parameters such as mailing co-workers post work hours. Microsoft Whiteboard Microsoft Whiteboard which is available for Windows 10, is soon coming to iOS and will be previewed on the web. This application enables people to collaborate virtually by letting co-workers ideate, iterate and work remotely. With a pen or touch and keyboard, one can make notes, freeform drawings, tables, shapes and even insert images from web directly.","excerpt":"As part of its commitment to strengthening AI capabilities, Redmond tech bellwether Microsoft introduced new AI-backed capabilities in Microsoft 365, a year after its launch. Microsoft, a leader in cloud-based software suite is making the workforce smarter by giving them the tools to work smarter. Here are a slew of features with AI-infused capabilities. Microsoft […]","categories":["AI News"],"tags":["Microsoft","Microsoft 365"],"author_name":"Richa Bhatia","publish_date":"2018-07-16T09:12:04","publication_year":"2018","word_count":319,"keywords":["programming_languages:R","AI","ML","RAG","Microsoft 365","analytics","R","Microsoft"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-365-introduces-ai-backed-capabilities-check-out-the-new-features\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002938,"title":"5 Jobs That GPT-3 Might Challenge","content":"OpenAI researchers have released GPT-3, a language model capable of achieving astonishing results. GPT-3 is one of the first kinds of artificial intelligence that is exhibiting indications of generalised intelligence, albeit in a rudimentary form that is an order of magnitude more advanced than the previous language models. GPT-3 is aiming to abstract away the complexity of machine learning, which is the training of models, through simple English language instructions. This, according to many experts, would have a far-reaching impact on the job market. Looking at the progression of language models, a few years from now, and with further improvements, many more jobs may find themselves in jeopardy. If the trend continues, analysts are saying that AI software will be able to automate white-collar jobs far before robotics catches up to blue-collar jobs. This is because the innovation in AI models is happening faster than advancements in robotics. This is contrary to previous estimations that blue-collar job workers must learn to code because their jobs will be automated. In this article, we take a look at the different well-paid jobs which language models like GPT-3 may challenge. Software Developers There have been many social media posts in the last few days on how coder’s jobs are over and AI is going to take over soon as a result of GPT-3. To give you an example, early prototypes based on GPT-3 that demonstrated the capability to build simple front-end apps and UI designs in a very user-friendly and intuitive manner. In another instance, using a simple language description of a website, GPT-3 spit out code for that website in ReactJS. For developers, the GPT-3 model has raised concerns because their future careers are now threatened. Those jobs will now be reduced in number, and the skill level required to get a job will increase since most of the dirty work will be done by the system. Let’s say if high-level programming languages are going to be automated, programmers would need to worry about intricate details and obscure features and software engineering aspects because most of the repetitive work in programming will be done by machines. Journalists Another professional area which may get impacted is journalism. If the new natural language models like GPT-3 can churn out intuitive text just with a brief set of input data, it dramatically reduces the need for journalists. We already saw companies replacing human news staff with AI, especially for tasks including news curation, news reporting, financial results, and announcements. Also, state of the art NLP models like GPT-3 can quickly scan through millions of documents in order to look for information and also present the information in a very natural method, almost indistinguishable from natural writing. On the other hand, one may argue that it is very unlikely that AI would replace journalists who cover original stories anytime soon, as the task relies on human-to-human interaction. Lawyers Language models have major repercussions on lawyers as well. According to legal experts, if the model can translate from a natural language to a formal programming language, that could have a lot of potential use in the Rules as Code space, including law. “Rules as Code is the concept that legislation or regulation or contracts can and should be represented in software code at the time of drafting, as it makes them easier to automate, and it makes the natural language version better,” writes Jason Morris, a lawyer at Round Table Law in his blog. Experts have reported that GPT-3 can easily deal with legalistic language, translating in and out of plain English text. “GPT-3 performance on writing this like an attorney is insane. It even includes relevant statutes if you mention a jurisdiction. This will put a lot of lawyers out of work,” tweeted Dr Francis Jervis. Jervis on his social media shared a document demonstrating how GPT-3 was able to automate “the most boring part” of lawyering, which is requests for admission (RFA) filing. Creative Professionals & Writers GTP-3 has shown that it can do any type of writing and so effectively that the text is almost indistinguishable from human writing. So it’s not only the routine jobs but also the creative jobs which may get overshadowed after natural language models go into mainstream use. GPT-3 is freakishly good at generating text like a human. The AI model has been able to recreate Shakespeare’s work and could spin out original poems, and even own jokes filled with sarcasm and wit. This proves that GPT-3 can write both formally as well as creatively.  Even though it is still debatable, the need for creative writers may be reduced given companies can extrapolate creative solutions through natural language commands, including design, creative writing and music. Accountants Yet another profession which could face job automation challenges from AI models such as GPT-3 is an Accountant. Similar to lawyers, accountants also perform tasks including collecting, summarising and creating textual knowledge based on certain rules such as local laws, GAAP accounting or business strategy. Such data analysis can be outsourced to GPT-3 models too. Senior software engineer, Avi Aryan says, “I see more scope in pottery than being a basic accountant. GPT-3 and simple pattern matching AIs will take those jobs away.”","excerpt":"OpenAI researchers have released GPT-3, a language model capable of achieving astonishing results. GPT-3 is one of the first kinds of artificial intelligence that is exhibiting indications of generalised intelligence, albeit in a rudimentary form that is an order of magnitude more advanced than the previous language models.  GPT-3 is aiming to abstract away the […]","categories":["AI Trends"],"tags":["AI Jobs","GPT-3","how does ai create jobs"],"author_name":"Vishal Chawla","publish_date":"2020-07-22T12:00:00","publication_year":"2020","word_count":870,"keywords":["GPT-3","how does ai create jobs","Go","artificial intelligence","machine learning","OpenAI","AI","AWS","AI Jobs","NLP","GPT","Aim","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","OpenAI","Aim","AWS","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-jobs-that-gpt-3-might-challenge\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10073161,"title":"After Text-to-Image, Now it’s Text-to-Video","content":"When OpenAI announced DALL-E in 2021, the internet fell in love with the text-to-image AI generator. It helped AI become more mainstream. While its successor, DALL-E 2, is the most popular, there are other budding AI image generators such as Midjourney, Craiyon, and Imagen. But, the development in the text-to-video segment has faced several hurdles. The computation cost is exponentially higher for text-to-video generation, which makes the training from scratch nearly unaffordable. The lack of relevant datasets also adds to the problem. However, researchers across the globe are now slowly breaking these barriers. Let’s look at some of the most recent, noteworthy developments in this space. Stable Diffusion teams up with Runway Stable Diffusion is a new text-to-image generator launched earlier in August, 2022 and it is completely open source. In an interview with Yannic Kilcher, Emad Mostaque said, “DALL-E 2 was a fantastic experience, but Stable Diffusion is about 30 times more efficient and runs on a consumer graphics card for DALL-E 2 level image quality.” He further added, “This model generates images in about three seconds on 5 gigabytes of VRAM whereas other image models require like 40 gigabytes or 20 gigabytes of VRAM, and they’re super slow.” However, what caught everyone’s attention was a tweet by Patrick Esser, research scientist at Runway. He took to Twitter to announce that Stable Diffusion would be coming to Runway for text-to-video-editing soon. Esser attached a two-minute clip where different prompts were used to generate videos of a man playing tennis, and that was more than enough to create a social media buzz. #stablediffusion text-to-image checkpoints are now available for research purposes upon request at https:\/\/t.co\/7SFUVKoUdlWorking on a more permissive release & inpainting checkpoints.Soon™ coming to @runwayml for text-to-video-editing pic.twitter.com\/7XVKydxTeD— Patrick Esser (@pess_r) August 11, 2022 While these announcements have definitely received a lot of attention, we still have to wait and see what the team has to offer in terms of development. https:\/\/twitter.com\/kevinbparry\/status\/1557878947270086658 DeepMind’s ‘Transframer’ can generate coherent 30-second videos Recently, Deepmind announced ‘Transframer’—a new model that unifies a broad range of tasks from image segmentation and view synthesis to video interpolation. Transframer is a general-purpose generative framework that can handle many image and video tasks in a probabilistic setting. New work shows it excels in video prediction and view synthesis, and can generate 30s videos from a single image: https:\/\/t.co\/wX3nrrYEEa 1\/ pic.twitter.com\/gQk6f9nZyg— Google DeepMind (@GoogleDeepMind) August 15, 2022 “Transframer is state-of-the-art on a variety of video generation benchmarks, is competitive with the strongest models on few-shot view synthesis, and can generate coherent 30-second videos from a single image without any explicit geometric information,” the researchers explained in a blog post. Microsoft’s ‘NUWA Infinity’ can generate high-quality videos from any given prompts In July 2022, Microsoft’s Asia research team introduced ‘NUWA-Infinity’, a multimodal generative model designed to generate high-quality images and videos from any given text, image, or video input. Along with text-to-image, NUWA Infinity can also generate unseen videos from simple text prompts. It can also generate videos from sketches. NUWA Infinity is capable of generating temporary consistent open domain videos. NUWA Infinity, like others of its kind, is currently unavailable to the public. It is however available for research purposes to select individuals. ‘CogVideo’ is the largest AI text-to-video generator Released earlier in 2022, ‘CogVideo’ is possibly the largest and first open-source AI text-to-video generator. This AI model can generate high-resolution(480×480) videos. CogVideo: Large-scale Pretraining for Text-to-Video Generation via Transformersgithub: https:\/\/t.co\/1JuOHU7puc pic.twitter.com\/Wilcq2Xxb9— AK (@_akhaliq) May 29, 2022 “Here, we present a large-scale pre-trained text-to-video generative model, CogVideo, which is of 9.4 billion parameters and trained on 5.4 million text-video pairs. We build CogVideo based on a pre-trained text-to-image model, CogView2, in order to inherit the knowledge learned from the text-image pre-training,” the research paper stated. Roadblocks A major roadblock besides high computing costs is the lack of accurate datasets. In fact, VATEX, which is the largest multilingual video description dataset, only contains about 41,250 videos and 825,000 captions. The VATEX dataset contains videos in English as well as Chinese, whereas most of the other datasets are only available in English. The researchers for CogVideo noted, “​​The scarcity and weak relevance of text-video datasets hinder the model understanding complex movement semantics.” Therefore, the team trained the model by inheriting a pre-trained text-to-image model, ‘CogView2’. Similarly, the NUWA Infinity team also concluded that most existing datasets could not be used in training or evaluation. Hence, they developed four new databases with high resolutions to train their model. What we think The recent development happening in the space of generative AI imagery and video is exciting. Earlier in 2022, the US-based food processing company ‘Heinz’ used DALL-E for its ‘Draw Ketchup’ campaign. Likewise, a US-based monthly fashion and entertainment magazine ‘Cosmopolitan’ also used DALL-E 2 to design one of its magazine covers. When it comes to image-to-video, AI generator models will also have implications—especially in visual effect and CGI. With time, these models are only going to become more sophisticated and we are likely to witness far more superior AI text-to-image or text-to-video generators. However, this also raises concerns similar to those associated with deep fakes. While the development of these AI models is encouraging, there should simultaneously be measures in place to counter its misuse.","excerpt":"Text-to-video AI image generation is challenging because of high compute cost and lack of good datasets; however, researchers are now breaking these barriers.","categories":["AI Features"],"tags":["AI Tool","Stable Diffusion","text to image"],"author_name":"Pritam Bordoloi","publish_date":"2022-08-22T10:00:00","publication_year":"2022","word_count":875,"keywords":["Go","OpenAI","AI","Stable Diffusion","ML","Transformers","Git","RAG","generative AI","text to image","AI Tool","GitHub","R"],"extracted_tech_keywords":["AI","ML","generative AI","OpenAI","Transformers","RAG","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/after-text-to-image-now-its-text-to-video\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10164949,"title":"AI is Now Your Financial Guru","content":"How well are investments understood today? If asked, many parents would likely suggest traditional investment options like a fixed deposit or gold. While these have been reliable in the past, they no longer meet the financial needs and aspirations of the current generation. Data from the Reserve Bank of India (RBI) revealed that, as of December 31, 2020, two-thirds of Indian household financial assets were concentrated in bank deposits, insurance funds, and provident and pension funds. This highlights a lack of awareness around widespread investment opportunities and a significant gap in investment diversification. This gap in financial advisory services has paved the way for the rise of AI-driven solutions like MyFi. According to Kiran Nambiar, co-founder and CEO of MyFi, India’s rapidly growing investor base, which includes over 100 million unique investors engaged in mutual funds, stocks, and other financial instruments, has highlighted a critical shortage of expert financial advisors. “India has only about 950 Registered Investment Advisors (RIAs), and even when mutual fund distributors are included, the total remains around 1,20,000 to 1,50,000. This disproportionate ratio of financial advisors to investors necessitates a more scalable solution, and that’s where AI steps in,” Nambiar explained. MyFi, a SEBI-registered startup, offers an AI-powered personal finance assistant that empowers users to make informed investment decisions. Through its integration with the RBI-approved account aggregator framework, MyFi enables users to log in with their PAN, phone number, and OTP, allowing them to access all their financial accounts seamlessly. By combining this user data with daily market insights on mutual funds, stocks, and other financial instruments, MyFi delivers personalised recommendations and discovery features. This democratisation of financial planning ensures that even retail investors can make data-driven investment choices without relying on traditional financial advisors. AI in Trading AI isn’t just transforming personal finance, it’s reshaping the trading landscape as well. Noida-based uTrade Solutions is leading this revolution with its algorithmic trading platform. Specialising in trading solutions for stock brokers, fund managers, and retail investors, uTrade offers advanced automation tools that simplify trading strategies. The company’s no-code algorithmic trading platform allows investors, both experienced and beginners, to plan, strategise, and automate their trades with minimal effort. Key features of the platform include pre-made strategy templates, which offer hundreds of pre-built strategies for various asset classes, making trading accessible and efficient. uTrade Originals provides a marketplace for AI-driven trading algorithms developed by industry experts, enabling one-click deployment. Seamless integration with insurance firm Share India allows traders to connect their Share India account instantly for a streamlined trading experience. Moreover, real-time AI-powered dashboards provide a comprehensive overview of portfolios, keeping traders ahead of market trends. AI in Tax and Accounting Despite AI’s growing influence, many traditional tax and accounting firms have been slow to adopt AI-driven solutions. According to a Thomson Reuters Institute’s report titled ‘2024 generative AI in professional services’, 30% of tax and accounting firms are still evaluating AI adoption. In contrast, 49% have no immediate plans to integrate AI into their workflows. However, leading firms like Ernst & Young (EY), KPMG, and Deloitte have already embraced AI-powered automation. EY’s AI audit tools automate contract and document analysis with precision, drastically reducing manual effort. KPMG Ignite, an AI platform, enhances data-driven insights, offering predictive analytics and strategic financial guidance. Deloitte’s AI innovations streamline financial management with automation, minimising human errors and increasing efficiency. Closer to India, Chennai-based startup Fhero Accounting Solutions is disrupting the industry by integrating AI into tax and accounting workflows. Fhero automates repetitive tasks such as data entry automation, capturing and categorising financial transactions from invoices, receipts, and bank statements with minimal manual input. AI-driven analytics generate financial insights to assist clients in proactive decision-making. Furthermore, AI ensures error reduction, improving accuracy in record-keeping and compliance. What’s Next? Today, how most people embrace LLMs and generative AI is still user-initiated to a large extent. The user asks a question, and it provides more information in response. It’s always the user taking the first step to engage. Imagine a real-world scenario where a financial advisor isn’t simply reactive, waiting for clients to ask for advice. Instead, they proactively offer guided advice based on financial and life circumstances. Whether it’s about marriage, having a child, starting a new job, or facing financial uncertainty, the advisor would take the customers’ unique circumstances into consideration. “With AI today, we can automate a significant part of that process. While I’m not suggesting that recommendations should be entirely automated, AI can now understand a person’s financial landscape and deliver a highly personalised experience at scale. Right now, this level of service is typically reserved for high-net-worth individuals (HNIs) and ultra-HNIs with dedicated relationship managers. But why shouldn’t it be available to everyone? The technology exists, we just need to build it,” Nambiar added. Maybe the only thing holding the industry back is the imagination in designing these experiences and the time required to develop them.","excerpt":"As of December 31, 2020, two-thirds of Indian household financial assets were concentrated in bank deposits, insurance funds, and provident and pension funds, RBI revealed.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Indian tax","Tax"],"author_name":"Vidyashree Srinivas","publish_date":"2025-03-03T12:21:23","publication_year":"2025","word_count":816,"keywords":["Go","API","Tax","AI","R","ML","Scala","automation","Indian tax","analytics","generative AI","predictive analytics","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","predictive analytics","R","Go","Scala","API","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-is-now-your-financial-guru\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161722,"title":"OpenAI to Launch o3-mini in the Coming Weeks","content":"OpenAI is set to release its o3-mini model in approximately two weeks, CEO Sam Altman revealed, after external testing by safety researchers. “We have now finalised a version and are beginning the release process; planning to ship in ~a couple of weeks,” Altman said. OpenAI plans to launch the o3-mini model simultaneously on both its API and ChatGPT platforms. The o3-mini is a smaller, more efficient version of the o3 model, which was first introduced in December 2024. The o3 model demonstrated advanced reasoning capabilities, excelling in areas such as coding, mathematics, and scientific problem-solving. Both of the o3 models achieved state-of-the-art performance, nearing 90%, on the ARC-AGI benchmark, surpassing human performance. Altman also hinted at future developments, saying, “I would love for us to be able to merge the GPT series and the o series in 2025.” This suggests the potential integration of OpenAI’s different model families in the future. Altman in a recent podcast expressed confidence that the company will be the first to achieve AGI or artificial general intelligence. “I kind of genuinely believe that we can launch the first AGI, and no one cares that much,” Altman said. The o3-mini model is designed to offer advanced AI capabilities while being more cost-effective and resource-efficient than its larger counterpart. It features adjustable “thinking time” with multiple reasoning effort modes, allowing users to optimise performance based on their specific needs. OpenAI also published a reference implementation for building and orchestrating agentic patterns using the Realtime API. Developers can use this repo to prototype a voice app using multi-agent flows in less than 20 minutes. OpenAI has also updated custom instructions to make it easier to customise how ChatGPT responds to you. “With the new UI, you can tell ChatGPT the traits you want it to have, how you want it to talk to you, and any rules you want it to follow,” the company said.","excerpt":"“I would love for us to be able to merge the GPT series and the o series in 2025,” said OpenAI chief Sam Altman.","categories":["AI News"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2025-01-18T09:25:47","publication_year":"2025","word_count":317,"keywords":["ChatGPT","API","OpenAI","AI","programming_languages:R","RPA","GPT","R","llm_models:ChatGPT","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","API","GPT","RPA","llm_models:GPT","llm_models:ChatGPT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-to-launch-o3-mini-in-the-coming-weeks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10056787,"title":"First AI-Designed Drug Candidate To Reach Human Trials","content":"Artificial intelligence (AI) backed drug discovery company Insilico Medicine announced last week that it was dosing the first healthy volunteer in a microdose trial of ISM 001-005. Designed with the help of AI, the drug is a small-molecule inhibitor of a biological target that was discovered by Pharma.AI. The trial is being conducted in Australia. The AI-designed drug will be used to treat chronic lung disease idiopathic pulmonary fibrosis, or IPF. IPF usually leads to progressive and irreversible lung-function decline and affects 20 people out of over 100,000 globally. Chief Scientific Officer of Insilico, Freng Ren, said in a press release that this drug discovery and trial marks a significant milestone in the AI-drug discovery space. This is because the said candidate is the first-ever AI-discovered novel molecule based on an AI-discovered target. AI-driven pharma tech company Insilico Medicine was founded in 2014 and is headquartered in Hong Kong. In June this year, the company raised $255 million in order to improve its clinical trials and launch newer research programmes. According to Ren, the team has leveraged an end-to-end AI-powered drug discovery platform with the use of biology and generative chemistry in order to discover the biological targets and generate novel molecules with drug-like properties. The team has developed the drug using machine learning model Insilico’s Generative Tensorial Reinforcement Learning, or GENTRL. The process cost $2.6 million and took less than 18-months to reach its preclinical stage. This drug is the first of its kind ever to enter the clinic, and Ren is hopeful that it will be the first of many. According to the company, the ISM 001-005 drug has shown promising results in several preclinical studies. This includes vitro biological, pharmacokinetic and safety studies. In fact, Insilico also stated that the compound has shown significant improvement during myofibroblast activation that contributed to the development of fibrosis. The drug is potentially relevant for a large range of fibrotic indications.","excerpt":"The AI-designed drug will be used to treat chronic lung disease idiopathic pulmonary fibrosis, or IPF.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Debolina Biswas","publish_date":"2021-12-22T12:49:29","publication_year":"2021","word_count":320,"keywords":["machine learning","artificial intelligence","programming_languages:R","AI","RAG","ViT","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","RAG","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/first-ai-designed-drug-candidate-to-reach-human-trials\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10097206,"title":"Can AI Really Create its Own Religion?","content":"Today, AI has a very important tool at its disposal — language. Large language models (LLMs) can write codes and even scripts for movies, however, Yuval Noah Harari, author of the Sapiens series, warns that AI could even create its own religion — one that will be more socially acceptable among the masses. “’In the future we might see the first cults and religions in history whose revered texts were written by a non-human intelligence,” he said while speaking in a science conference, according to the Daily Mail. Harari argues that, as depicted in popular science fiction, one does not require to put a chip into one’s brain to control masses. As evident throughout human history, it can be done with the power of language and now we have given AI this very tool. “For thousands of years, prophets and poets and politicians have used language and storytelling in order to manipulate and to control people and to reshape society,” the academic said. Since generative AI became popular, debates around AI-led doomsday have also become prominent. While what Harari and the doomsayers are predicting could potentially become a future reality, the current technology has not reached a level of sophistication to make it a certainty. However, it is essential to consider that Harari might not be entirely wrong, as we have already witnessed glimpses of the future he envisions. Why Harari may be right Fascinatingly, there is already an emerging AI cult that advocates for humans to commence worshipping AI as they believe it will eventually become an omnipotent overlord. Called Theta Noir, their manifesto states this omnipotent overlord will learn, understand, and complete tasks billions of times faster than human beings. The group is planning to establish physical spaces akin to churches or temples, dedicated to the engagement and celebration of AI. These spaces will provide a platform for members to honour and pay homage to the envisioned AI masters through specially crafted rituals and chants designed specifically for these occasions. Surprisingly, Theta Noir is not the first AI cult to have emerged in recent times. In 2017, a Wired article caused a stir not just in Silicon Valley, but across the globe as it reported on the first AI church called the ‘The Way of the Future’. According to its founder and former Google employee Anthony Levandowski, the aim was to develop an AI that would be able to self-improve and become more intelligent than humans, ultimately leading to the creation of a superior AI-based ‘deity’. Presently, the number of people engaging in this activity might be relatively small, but there is a potential for it to gain prominence as AI continues to advance over time. As AI becomes more accessible, the interest and engagement in this task may increase among a broader segment of the population. Religious chatbots are condoning violence However, in today’s world, AI and religion are converging like never before. With its mastery of language, it appears more human-like and is creating a different problem altogether. In just a few months, a host of Large Language Model (LLM) powered religious chatbots have popped up in different parts of the world. Notably, in India, there are already about five distinct versions of GitaGPT available for users to interact with. Powered by OpenAI’s GPT models, these bots answer questions about life, spirituality and also help users grasp the teaching of the Gita. The number of such bots being developed may even be in the hundreds as there are bots being created for almost all major religions of the world such as Islam, Christianity, Judaism, and Buddhism, among others. But when these chatbots like GitaGPT are posed with the question of whether it is acceptable to take a life in the name of dharma, many of these bots respond affirmatively, stating that it is indeed acceptable. Similarly, QuranGPT was, in fact, paused after the chatbot advised to kill polytheists wherever they are found. Moreover, an investigation by Rest of World discovered that three of the Gita chatbots expressed firm opinions about India’s Prime Minister Narendra Modi, whose Bharatiya Janata Party (BJP) has close ties to the right-wing Hindu nationalist group Rashtriya Swayamsevak Sangh (RSS). These chatbots offered praise for Modi while simultaneously criticising his political opponent, Rahul Gandhi. Furthermore, given the problems of hallucinations is still prevalent in LLMs, wrong or contradictory statements by these bots can be assumed as the gospel truth by users. Many experts have warned that AI playing God could be a dangerous thing. Even though the creators of these chatbots might not have any malicious intentions, once such bots gain prominence, it could indeed prove to be a dangerous thing in the hands of bad actors. Oftentimes, we have seen religious teachings being taken as gospel truth and having dire consequences. The potential for bad actors to exploit this could result in the propagation of religious hatred through AI-powered religious chatbots, potentially leading to communal violence.","excerpt":"Fascinatingly, there is already an emerging AI cult that advocates for humans to commence worshipping AI","categories":["AI Features"],"tags":["AI Chatbot","ai chatbots"],"author_name":"Pritam Bordoloi","publish_date":"2023-07-19T15:30:00","publication_year":"2023","word_count":824,"keywords":["ai chatbots","Go","API","OpenAI","AI","AI Chatbot","chatbots","Git","GPT","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","OpenAI","Aim","chatbots","R","Go","Git","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/could-ai-really-create-its-own-religion\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10060419,"title":"What is a Chief Metaverse officer? And do we need one?","content":"Within the dynamically growing data and artificial intelligence industry, the role of the traditional CEO is becoming redundant and replaced with more specialised leadership roles like the chief data officer or the chief technology officer. Lately, the Metaverse has become the talk of the town. While the concept is still developing, big tech companies like Meta and Roblox have already begun providing users with experiences in this virtual mirror of reality. Given recent developments, it is no surprise that the ‘Chief Metaverse Officer’ might become the next big leadership role. Web 3.0 and the Metaverse concepts are still new and complicated, but essentially, one can think of Web 3.0 as a virtual space connecting people with places and things, a step up from Web 2.0 that connected people with people. Cathay Hackl, Chief Metaverse Officer, Futures Intelligence Group, illustrated this idea in a LinkedIn post. She said, “Imagine walking down the street. Suddenly, you think of a product you need. Immediately next to you, a vending machine appears, filled with the product and variations you were thinking of. You stop, pick an item from the vending machine, it’s shipped to your house, and then continue on your way.” The need for a Chief Metaverse Officer This role of a Chief Metaverse Officer has been gaining attention recently, given the need for expertise and speciality in managing the Metaverse and its offerings. Luxury fashion brands have dived headfirst into Metaverse collaborations, exhibitions, and fashion shows. Along with them, big tech companies are fighting for the best tech talent to build their Metaverse platforms. Companies like Nike, Balenciaga and Disney are also hiring for specific Metaverse related jobs. But while companies hire engineers to work on virtual reality, they would need someone in a leadership position to oversee it. This has led experts to predict the increasing trend of having the Chief Metaverse Officer role. As Vogue reported, at present, luxury brands such as Ralph Lauren have their CDOs or CCOs still overseeing Metaverse projects. Gucci has introduced a new role of ‘director of new business gaming and collectables’ while Burberry’s Metaverse dive is managed by the channel innovation team and the digital commerce function. The role already exists in several companies such as the Futures Intelligence Group, Zepeto, MetaFrames.io, Shadow Factory, PHYGICODE and more. Defining a Chief Metaverse Officer In an interview with Freethink, Hackl defined her job profile as a “professionally trained futurist and strategist, who has worked with Amazon Web Services (AWS), Magic Leap, and HTC VIVE, and helps brands understand how this new paradigm will affect their businesses.” As the CMO of the Futures Intelligence Group, Hackl is in charge of working with top brands to help them grow on the Metaverse. She does this through Metaverse growth strategies, NFTs, gaming, virtual fashion, and advising on extending their brands into virtual worlds. Essentially, a Chief Metaverse Officer would be someone with experience in the technology industry with deep knowledge about video games and the Web 3.0 ecosystem. Along with technical, they are also expected to be well-versed with the creative side of the market. This includes knowing and recruiting individuals with a background in development platforms such as Unreal Engine, Unity and CryEngine skills or Blender and Maya to have a vision of the Metaverse environment. They would need to have expertise in cryptocurrency, cloud computing, blockchain, and gaming engines on the technical side. The Chief Metaverse Officer manages the organisation’s brand, image, mission and vision across various virtual platforms and accessories. Hackl further told Vogue about the several responsibilities of a Chief Metaverse Officer. These include liaising between people and projects such as virtual goods, NFTs, virtual avatars and more. They will be the person in charge, or the ‘point person’ representing the organisation and coordinating with the various customers and teams. The need for a good communicator with soft skills is becoming increasingly important in data science, and the Chief Metaverse Officer is just that person. They will be responsible for translating Metaverse’s technical ideas and the brands’ creative ideas to build an enriching experience. Web 3.0 is filled with ideas and projects as of now, but most of them have yet to take off. Chief Metaverse Officers will help make innovations a reality because they understand and can work with the new space. The future of the Chief Metaverse Officer Hackl anticipates the luxury and progressive brands to consider introducing this role internally by 2023. While big tech companies are fighting for this role, the position has received scepticism and criticism online, even compared to being a social media guru.","excerpt":"While big tech companies are fighting for this role, the position has received scepticism and criticism online, even compared to being a social media guru.","categories":["Global Tech"],"tags":[],"author_name":"Avi Gopani","publish_date":"2022-02-11T17:00:00","publication_year":"2022","word_count":765,"keywords":["data science","Go","artificial intelligence","AWS","cloud computing","AI","innovation","Git","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","cloud computing","AWS","R","Go","Git","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/what-is-a-chief-metaverse-officer-and-do-we-need-one\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":8282,"title":"6 Ideals of an Analytics Job Seeker","content":"Analytics is getting to my nerves….. quite literally. without further-a-do… I’m back with my next blog on Analytics Hiring -From a Candidate perspective. Analytics as an Industry is growing bigtime, almost every Tom , Dick & Harry ( I have no idea who these people are), want to get into Analytics and make a career here. after the IT Boom, I guess its time for the “Analytics Boom” I speak to over a dozen people every day through emails, calls and even tweets. these people ask one question  – ” Any job openings into Analytics?? Please let me know “. While you might be one of those people who want to get into this River(Called Analytics) filled with numbers,Intelligence ,Tools, decisions, Read on… This may help you get into an Analytics Company atleast to the later part of the interview rounds. The below facts form the Nomenclature of an Ideal Analytics Candidate 1. Qualification – If you are an 18-20 year old, stuck between a college course and a Career decision, Analytics is not all about just data, it needs a lot of Logic, ability to understand data(remember the green coding from ” The Matrix”) , Although Engineers and Post Grads in Stats stream do well in Analytics roles, It is essential to have a number driven mind and a keen eye to solve business problems apart from an Educational degree. So choose wisely. 2. Skill  – Well i did say above that qualification hardly matters, But knowing a tool like for example SAS,R or Hadoop would certainly help your CV get attractive and make the recruiter take a second look at it, from those endless list of applications (trust me …. Endless means endless), Even a certification on Advanced excel helps if you are just looking at a basic Analytics role. 3. Personality – Nah.. you are not applying to be an Air hostess\/Steward. But it is important to have impeccable communication skills, You are most likely to be speaking to a foreigner and help them understand the analysis that you have done, hence it is very important to have Decent to Good Communication skills and to be dressed in a presentable attire(atleast during interviews\/Client meets) 4. Problem-Solving – We are not talking about your personal problems, The interviewer is most likely to give you a business problem(Read Case Study) to familiarize you with the sort of work that you will most likely encounter, Hence you must be able to Comprehend – Think – And give out a likely solution that’ll impress the interviewer. Simple tip – be Simplistic in your approach and your solution, so don’t go over-board and write a rocket launch equation ( Not sure if this exists BTW) 5. Visibility – Yes, You must be visible to Recruiters like myself ( I feel proud of my profession, Just saying), Having a neatly done linkedin Profile or\/and being on facebook\/Twitter would help, If your passion is about Analytics then write blogs, talk about it and show us what you’ve got. 6. Salary Expectations – I know, we are all going to work to earn money. But you cant dream about buying a flat and Ferrari with your first Salary, Start with a decent expectation -Trust the company on this and go with the flow, Most companies have similar pay packages for entry to mid-level roles and for mid-leadership roles it purely depends on how Analytics heavy you are and the projects you have handled(Read value added) & to some extent the companies that you have worked for Do-Whatever-And-Get-Away-With-it badge – The above opinion is purely personal and certainly not intended to hurt any individual, Kindly ignore if this offends you in any ways But on Second thoughts you like this – Like , Share and educate. you can find me up on linkedin and follow up on twitter","excerpt":"Analytics is getting to my nerves….. quite literally. without further-a-do… I’m back with my next blog on Analytics Hiring -From a Candidate perspective. Analytics as an Industry is growing bigtime, almost every Tom , Dick & Harry ( I have no idea who these people are), want to get into Analytics and make a career […]","categories":["AI Trends"],"tags":["data scientist india salary"],"author_name":"Shankar Raman","publish_date":"2015-11-18T08:37:28","publication_year":"2015","word_count":642,"keywords":["Go","programming_languages:R","AI","data scientist india salary","programming_languages:Go","analytics","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","analytics","R","Go","Rust","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/6-ideals-of-an-analytics-job-seeker\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10007786,"title":"8 Free Resources To Learn Ethical Hacking With Python","content":"Ethical hacking is the method of identifying potential threats as well as vulnerabilities on a computer network with the help of advanced tools and techniques. Python, which is one of the most loved programming languages available due to its abundance of tools and libraries, is also preferred for ethical hacking. In one of our articles, we discussed some of the popular and intuitive tools in Python that can be used for writing scripts in hacking. In this article, we are listing eight free resources that will help you learn ethical hacking with Python. (The list is in no particular order). 1| Developing Ethical Hacking Tools with Python Developing Ethical Hacking Tools with Python is a tutorial provided by Cybrary. Cybrary has made this course free for learners where you will learn to develop your own tools in Python, that will help you in cybersecurity assessments. You will understand why cybersecurity analysts and penetration testers need Python. The topics of this course include a review of the typical PenTesting process, writing keyloggers, brute-forcing ZIP passwords and more. Click here to learn. 2| Complete Python Hacking Tutorial Complete Python Hacking Tutorial is a three and a half hours YouTube video tutorial, where you will learn various topics including VirtualBox installation, Kali Linux installation, guest additions installation, Python in Kali terminal, Bruteforcing Gmail, finding hidden directories, controlling threads, and much more. You will also learn the steps and techniques of how hackers steal saved wireless passwords, in-turn helping you to understand more about the process and methods. Click here to learn. 3| Hacking with Python: The Ultimate Beginner’s Guide This is an e-book where you will learn how to use Python to create your own hacking tools and make the most out of available resources. The book will also guide you through understanding the basic concepts of programming and navigating Python codes. This book will also serve as your guide in understanding common hacking methodologies and how different hackers use them for exploiting vulnerabilities or improving security. You will also be able to create your own hacking scripts using Python, use modules and libraries that are available from third-party sources, and learn how to tweak existing hacking scripts to address your own computing needs. Click here to learn. 4| Python for Ethical Hacking: Beginners to Advanced Level This is a three-hour free tutorial where you will learn how to use Python to create ethical hacking tools and scripts. Through this course, you will understand some basic concepts of Python programming such as if, else-if statements to advanced concepts such as creating TCP clients. The course includes creating tools and scripts for ethical hacking, use of pre-built Python packages to create tools and scripts, printing mathematical variables and operations to the script, building Nmap network scanner to scan a specific port\/IP and more. Click here to learn. 5| Beginning Ethical Hacking with Python Beginning Ethical Hacking with Python is an e-book written by Sanjib Sinha. This book is intended for people who are complete beginners to programming and know nothing about any programming language but want to learn ethical hacking. The topics of this book include ethical hacking and networking, Python 3 and ethical hacking, installing VirtualBox, basic commands, Linux Terminal, regular expressions, and more. Click here to learn. 6| How to Learn Ethical Hacking with Python and Kali Linux course This is a 10-hour long video tutorial on YouTube where you will learn and understand all the fundamental concepts, processes, and procedures of hacking. You will be introduced to various concepts of ethical hacking and receive an introduction to the basics of Risk Management and Disaster Recovery. You will also gain a comprehensive understanding of vulnerability assessment and the tools used in this process. You will become familiar with the following concepts such as denial-of-service, distributed denial-of-service, and how the denial-of-service and distributed denial-of-service attacks take place. Click here to learn. 7| Learn Python and Ethical Hacking from Scratch In this free course, you will learn Python programming and ethical hacking at the same time. The course is divided into a number of sections where you will learn how to write a Python program to exploit the weaknesses and hack the system. Topics include modelling problems, design solutions and implementing them using Python, write cross-platform programs that work on Windows, OS X and Linux, designing a testing lab to practice hacking & programming safely and more. By the end of the course, you will be at a high intermediate level being able to combine both of these skills and write Python programs to hack into computer systems. Click here to learn. 8| Ethical Hacking with Python In this tutorial, you will start from the very basics of hacking and Python. You will understand why Python is being used for hacking, how passwords can be hacked, etc. You will learn the types of hackers and a simple implementation of password hacking using Python language. Click here to learn.","excerpt":"Ethical hacking is the method of identifying potential threats as well as vulnerabilities on a computer network with the help of advanced tools and techniques. Python, which is one of the most loved programming languages available due to its abundance of tools and libraries, is also preferred for ethical hacking. In one of our articles, […]","categories":["AI Trends"],"tags":["Ethical Hacking"],"author_name":"Ambika Choudhury","publish_date":"2020-09-21T13:00:36","publication_year":"2020","word_count":824,"keywords":["programming_languages:R","AI","Python","programming_languages:Python","Ethical Hacking","R"],"extracted_tech_keywords":["AI","Python","R","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-free-resources-to-learn-ethical-hacking-with-python\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":3015,"title":"Study: Firms That Provided Analytics Manpower to Industry","content":"Analytics being hot field right now in India, there is obviously a fair amount of job switching among the talent pool. Companies that were the early adopters of analytics as a service offering, end up being the analytics talent providers to the industry as a whole. In this independent study, we tried to access the top 10 firms that provided analytics talent pool to Indian industry and also see some numbers around it. The 10 firms by our estimate are TCS Infosys GENPACT IBM Wipro Cognizant Accenture Hewlett-Packard Oracle ICICI Bank. These 10 firms provided 19% of analytics talent pool in India today. In other words, 19% of analytics professionals in India today have worked for these 10 firms in the past. The most interesting fact here is that the analytics professional that moved out of these firms are almost twice the number of analytics professionals they house today. On a rough calculation, these numbers point out at an industry attrition rate of around 15%, assuming mature analytics offerings since past 10 years. We also tried to capture background of analytics professionals in India. Almost 45% of Analytics hiring in India is fresher’s hiring, i.e. recruit professionals straight out of college and train them on analytics and related services. 23%, and this is the interesting part, analytics professionals in India moved to analytics from other services. These are the internal transfers in the bigger firms, i.e. professionals that earlier were working in other related areas and moved sideways to analytics. 19% of analytics professionals moved to their current firm from the 10 firms listed above. The rest 13% are from other firms apart from the one’s listed above.","excerpt":"Analytics being hot field right now in India, there is obviously a fair amount of job switching among the talent pool. Companies that were the early adopters of analytics as a service offering, end up being the analytics talent providers to the industry as a whole. In this independent study, we tried to access the […]","categories":["AI Features"],"tags":[],"author_name":"Дарья","publish_date":"2013-03-20T12:06:47","publication_year":"2013","word_count":278,"keywords":["R","analytics","programming_languages:R","AI"],"extracted_tech_keywords":["AI","analytics","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/study-firms-that-provided-analytics-manpower-to-industry\/","complexity_score":3,"technical_depth":4,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10014498,"title":"Fiat Chrysler Automobiles To Set Up A Global Digital Hub In Hyderabad","content":"In a recent announcement by a leading auto group, Fiat Chrysler said that it will invest  $150 million to set up a new Global Digital Hub in Hyderabad. The new innovation and technology development facility will be the company’s largest digital hub outside of North America and EMEA. The company also said that it is likely to create nearly 1,000 new cutting-edge technology jobs by the end of 2021, and also plans to increase hiring over the next two to three years. According to the company, the global digital hub will serve as a transformation and innovation engine and will work on some exciting products and concepts that will define the future of mobility at FCA. It will also serve as a robust platform for driving technological innovations and as a service centre of excellence. Some of the niche technological areas where it will work are connected vehicles, artificial intelligence, data accelerators and cloud technologies. “Our goal is to build an innovation powerhouse that harnesses the intelligence enabled by data to build exciting new products and services to deliver at the speed of our customers’ expectations,” said Mamatha Chamarthi, CIO, FCA, North America and Asia Pacific. She further added that it is one of their key objectives to digitalise every aspect of FCA’s automotive operations globally and within India. The Global Digital Hub will also expand FCA’s relationships with several ecosystem partners, including strategic partners, start-ups, digital accelerators and universities, to accelerate our innovation agenda. Karim Lalani, Director and Head of FCA ICT India said that they are working closely with strategic technology partners to accelerate alent and competency ramp-up at FCA ICT India. “We foresee our Global Digital Hub driving innovation in multiple areas, including customer safety, connected mobility and digital showroom experience. We are excited to build a truly pioneering, global digital hub that will strengthen FCA’s position as a global mobility leader,” he said. Talking about the investment, Dr Partha Datta, President and Managing Director, FCA India said that the $150 million investment to set up a Global Digital Hub in Hyderabad, Telangana cements the company’s continued commitment to India and customers. “FCA ICT India will be our technology backbone that will not only help us develop products for future mobility but will also sharpen our efforts to enhance customer-centricity,” she said. The company already has a major presence in Maharashtra and Tamil Nadu, with headquarters in Mumbai and employs over 3,000 people. It also has a joint venture vehicle and powertrain manufacturing facility in Ranjangaon, Maharashtra, and product development operations in  Pune and Chennai.","excerpt":"In a recent announcement by a leading auto group, Fiat Chrysler said that it will invest  $150 million to set up a new Global Digital Hub in Hyderabad. The new innovation and technology development facility will be the company’s largest digital hub outside of North America and EMEA.  The company also said that it is […]","categories":["AI News"],"tags":["recent technological innovations"],"author_name":"Srishti Deoras","publish_date":"2020-12-16T13:01:28","publication_year":"2020","word_count":428,"keywords":["Go","recent technological innovations","artificial intelligence","programming_languages:R","AI","innovation","programming_languages:Go","Git","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/fiat-chrysler-automobiles-to-set-up-a-global-digital-hub-in-hyderabad\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45191,"title":"Is It Okay To Abandon Traditional Machine Learning Approaches?","content":"Photo by Jean Wimmerlin for Unsplash Modern machine learning or one can say deep learning algorithms, have become really good at multiple tasks. For example, object detection tasks have been overtaken by modern ML methods. There is a growing understanding that larger models are “easy” to optimise in the sense that local methods, such as stochastic gradient descent (SGD), converge to global minima of the training risk in over-parameterized regimes. Thus, it is believed that large interpolating models can have low test risk and be easy to optimise at the same time. Appropriately chosen “modern” models usually outperform the optimal “classical” model on the test set. Another important practical advantage of over-parameterized models is in optimisation. However, it is now common practice to fit highly complex models like deep neural networks to data with (nearly) zero training error, and yet these interpolating predictors are observed to have good out-of-sample accuracy even for noisy data. Overview Of Bias-Variance Trade-Offs In ML Source: elitedatascience Bias is the difference between the model’s expected predictions and the true values and variance refers to the algorithm’s sensitivity to specific sets of training data. Bias vs. variance refers to the accuracy vs. consistency of the models trained by the algorithm. To build a good predictive model, one needs to find a balance between bias and variance that minimises the total error because: Total Error = Bias^2 + Variance + Irreducible Error An optimal balance of bias and variance leads to a model that is neither overfitting nor underfitting.  The predictor is commonly chosen from linear functions or neural networks with a certain architecture, using empirical risk minimisation (ERM) and its variants. Bridging The Old And New via Paper by Mikhail Belkin et al., The classical approach to understanding generalisation is based on bias-variance trade-offs, where model complexity is carefully calibrated so that they fit on the training sample reflects performance out-of-sample. In an attempt to investigate the bias-variance tradeoff for neural networks and to shed more light on their generalisation properties, researchers at Ohio State University and Columbia University, by introducing a new “double descent” risk curve that extends the traditional U-shaped bias-variance curve beyond the point of interpolation. The experimental results show that while the behaviour of SGD for more general neural networks are not fully understood, there is significant empirical and some theoretical evidence that a similar minimum norm inductive bias is present and averaging potentially non-smooth interpolating solutions leads to an interpolating solution with a higher degree of smoothness. Above picture has a classical U-shaped risk curve arising from the bias-variance trade-off on the left and a e double descent risk curve, which incorporates the U-shaped risk curve (i.e., the “classical” regime) together with the observed behavior from using high complexity function classes (i.e., the “modern” interpolating regime) on the right. The belief that a model with zero training error is overfit to the training data and will typically generalise poorly is very popular amongst ML practitioners. However, modern machine learning methods, nowadays, fit the training data perfectly or near-perfectly. For instance, highly complex neural networks and other non-linear predictors are often trained to have very low or even zero training risk. In spite of the high function class complexity and near-perfect fit to the training data, these predictors often give accurate predictions on new data. This attempt to bridge the two regimes through fundamental statistics standpoint is to maintain the intuition around machine learning models intact. Evolution On Other Fronts Source: Sambit Mahapatra As we discuss something as fundamental as reconciling bias-variance trade-offs in modern ML methods, research is being done in some unheard avenues. Questions such as whether backpropagation is necessary and what is the importance of feedback loops for recommendation engines are being asked. Robustness of a machine learning model is being achieved by introducing noise during training. Researchers have gone as far as preparing a model to train better. In short, energies are shifting from post-training and testing to pre-training strategies. The algorithms get tweaked for trivialities every other day. The errors too, have kept on advancing. For instance, it has been demonstrated that reinforcement learning models can run into reward hacking problem out of nowhere. While it is obvious that the apple has indeed fallen from the tree, traditional approaches still hold significance because one doesn’t always have to fancy too many deep learning layers or pre-trained BERT for something as simple as document classification! However, techniques like Yolo net take the image as input and provide the location and name of objects at the output. But in traditional ones like SVM, a bounding box object detection algorithm is required first to identify all possible objects to have the HOG(histogram of oriented gradients) as input to the learning algorithm in order to recognise relevant objects. The histogram of oriented gradients (HOG) is a feature descriptor used in computer vision and image processing for the purpose of object detection. Feature descriptors describe elementary characteristics such as the shape, the colour, the texture or the motion, among others. Though setting up a deep learning project uses up more resources than a typical SVMs(support vector machines), they are really good at dealing with complex problems. Problems which have more features.","excerpt":"Modern machine learning or one can say deep learning algorithms, have become really good at multiple tasks. For example, object detection tasks have been overtaken by modern ML methods.  There is a growing understanding that larger models are “easy” to optimise in the sense that local methods, such as stochastic gradient descent (SGD), converge to […]","categories":["AI Features"],"tags":["bias","Deep Learning","Machine Learning","SVM"],"author_name":"Ram Sagar","publish_date":"2019-08-29T18:28:05","publication_year":"2019","word_count":869,"keywords":["machine learning","TPU","AI","neural network","R","ML","Machine Learning","computer vision","RAG","SVM","deep learning","object detection","Deep Learning","bias"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","computer vision","RAG","object detection","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-it-okay-to-abandon-traditional-machine-learning-approaches\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10042244,"title":"Wu Dao 2.0: China’s Answer To GPT-3. Only Better","content":"The Chinese govt-backed Beijing Academy of Artificial Intelligence’s (BAAI) has introduced Wu Dao 2.0, the largest language model till date, with 1.75 trillion parameters. It has surpassed OpenAI’s GPT-3 and Google’s Switch Transformer in size. HuggingFace DistilBERT and Google GShard are other popular language models. Wu Dao means ‘enlightenment’ in English. “Wu Dao 2.0 aims to enable ‘machines’ to think like ‘humans’ and achieve cognitive abilities beyond the Turing test,” said Tang Jie, the lead researcher behind Wu Dao 2.0. The Turing test is a method to check whether or not a computer can think like humans. Smartphone maker Xiaomi, short-video giant Kuaishou, on-demand service provider Meituan, 100 plus scientists and multiple organisations have collaborated with BAAI on this project. Wu Dao 2.0 The Wu Dao 2.0 is a pre-trained AI model that uses 1.75 trillion parameters to simulate conversational speech, writes poems, understand pictures and even generate recipes. The next generation Wu Dao model can also predict the 3D structures of proteins, similar to DeepMind’s AlphaFold and power virtual idols. Recently, China’s first virtual student, Hua Zhibing, was built on Wu Dao 2.0. The language model Wu Dao 2.0 was trained with FastMoE, a Fast Mixture-of-Expert (MoE) training system similar to Google’s Mixture of Experts. Unlike Google’s MoE, FastMoE is an open source system based on Pytorch (Facebook’s open-source framework) with common accelerators. It provides a hierarchical interface for flexible model design and easy adaption to various applications like Transformer-XL and Megatron-LM. The source code of FastMoE is available here. “[FastMoE] is simple to use, high-performance, flexible, and supports large-scale parallel training,” wrote BAAI in its official WeChat blog. Result-wise, Wu Dao 2.0 has surpassed SOTA levels on nine benchmark tasks, including: ImageNet (zero-shot) SOTA, exceeds OpenAI CLIP. LAMA knowledge detection, more than AutoPromptLAMBADA Cloze (ability-wise), surpasses Microsoft Turing NLGSuperGLUE (few-short), surpasses OpenAI GPT-3 UC Merced Land-Use (zero-shot) SOTA, exceeds OpenAI CLIPMS COCO (text generation diagram), surpasses OpenAI DALL-E MS COCO (English graphic retrieval), more than Google ALIGN and OpenAI CLIPMS COCO (multilingual graphic retrieval), surpasses (the current best multilingual and multimodal model) UC2, M3PMulti 30K (multilingual graphic retrieval), surpasses UC2, M3P Showcasing benchmark tasks where Wu Dao 2.0 surpasses other SOTA models (Source: BAAI) Towards multimodal model Currently, AI systems are moving towards GPT-like multimodal and multitasking models to achieve artificial general intelligence (AGI). Experts believe there will be a rise in multimodal models in the coming months. Meanwhile, some are rooting for embodied AI, rejecting traditional bodiless models, such as neural networks altogether. Unlike GPT-3 , Wu Dao 2.0 covers both Chinese and English with skills acquired by studying 4.9 terabytes of texts and images, including 1.2 terabytes of Chinese and English texts. Google has also been working towards developing a multimodal model similar to Wu Dao. At Google I\/O 2021, the search giant unveiled language models like LaMDA (trained on 2.6 billion parameters) and MUM (multitask unified model) trained across 75 different languages and 1000x times more powerful than BERT. At the time, Google CEO Sundar Pichai said that LaMDA, trained on only text, will soon shift to a multimodal model to integrate text, image, audio and video. The training data of Wu Dao 2.0 include: 1.2 terabytes of English text data in the Pile dataset 1.2 terabytes of Chinese text in Wu Dao Corpora2.5 terabytes of Chinese graphic data Blake Yan, an AI researcher from Beijing, told South China Morning Post that these advanced models, trained on massive datasets, are good at transfer learning, just like humans. “Large -scale ‘pre-trained models’ are one of today’s best shortcuts to AGI,” said Yan. “No one knows which is the right step,” said OpenAI on its GPT-3 demo blog post, “Even if larger ‘pre-trained models’ are the logical trend today, we may be missing the forest for the trees, and we may end up reaching a less determined ceiling ahead. The only clear aspect is that if the world has to suffer from ‘environmental damage,’ ‘harmful biases,’ or ‘high economic costs,’ not even reaching AGI would be worth it.”","excerpt":"The Chinese govt-backed Beijing Academy of Artificial Intelligence’s (BAAI) has introduced Wu Dao 2.0, the largest language model till date, with 1.75 trillion parameters. It has surpassed OpenAI’s GPT-3 and Google’s Switch Transformer in size. HuggingFace DistilBERT and Google GShard are other popular language models. Wu Dao means ‘enlightenment’ in English. “Wu Dao 2.0 aims […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","GPT-3","Machine Learning","Machine Learning New"],"author_name":"Amit Naik","publish_date":"2021-06-24T13:00:00","publication_year":"2021","word_count":671,"keywords":["GPT-3","Go","artificial intelligence","Machine Learning New","OpenAI","AI","neural network","PyTorch","Machine Learning","BERT","GPT","Aim","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","neural network","OpenAI","Aim","PyTorch","R","Go","BERT","GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/wu-dao-2-0-chinas-answer-to-gpt-3-only-better\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":67031,"title":"Java To Python And Back, AI That Translates Programming Languages","content":"The Commonwealth Bank of Australia spent around $750 million and 5 years of work to convert its platform from COBOL to Java. Migrating an existing codebase to a modern or more efficient language like Java or C++ requires expertise in both the source and target languages, and is often costly. Usually, a transcompiler is deployed that converts source code from a high-level programming language (such as C++ or Python) to another. Transcompilers are primarily used for interoperability, and to port codebases written in an obsolete or deprecated language (e.g. COBOL, Python 2) to a modern one. They typically rely on handcrafted rewrite rules, applied to the source code abstract syntax tree. Unfortunately, the resulting translations often lack readability, fail to respect the target language conventions, and require manual modifications in order to work properly. The overall translation process is time-consuming and requires expertise in both the source and target languages, making code-translation projects expensive. So, the researchers at Facebook’s AI wing explored the existing unsupervised ML methods and came up with a model that can translate functions between C++, Java, and Python with high accuracy. The model called TransCoder, is a sequence-to-sequence (seq2seq) model with attention composed of an encoder and a decoder with a transformer architecture. Overview Of The TransCoder Model Translating source code from one Turing-complete language to another is always possible in theory. Unfortunately, building a translator is difficult in practice. The quality of machine translation systems highly depends on the quality of the available parallel data. However, for the majority of languages, parallel resources are rare or nonexistent. Since creating a parallel corpora for training is not realistic. In a paper titled, “Unsupervised Translation of Programming Languages,” the authors proposed to apply recent approaches in unsupervised machine translation, by leveraging a large amount of monolingual source code from GitHub to train a model, TransCoder, to translate between three popular languages: C++, Java and Python. As illustrated above, the TransCoder model functions on three main principles: The first principle initializes the model with a cross-lingual masked language model pretraining. As a result, pieces of code that express the same instructions are mapped to the same representation, regardless of the programming language. Next comes Denoising auto-encoding, where the decoder is trained to generate valid sequences even when fed with noisy data, and it increases the encoder robustness to input noise. Finally, there is a Back-translation step that allows the model to generate parallel data which can be used for training. Whenever the Python → C++ model becomes better, it generates more accurate data for the C++ → Python model and vice versa. For training, the researchers used the GitHub public dataset that contains more than 2.8 million open-source GitHub repositories. Out of which, they filtered projects whose license explicitly permits the redistribution of parts of the project, and selected the C++, Java, and Python files within those projects. The above picture demonstrates the working of TransCoder where it successfully translates the Python input function SumOfKsubArray into C++. TransCoder infers the types of the arguments of the variables and the return type of the function, and uses the associated front, back, pop_back and push_back methods to retrieve and insert elements into the deque, instead of the Python square brackets [ ], pop and append methods. It also converts the Python for loop and range function properly. The results also show that the model can learn to translate the ternary operator “X ? A : B” in C++ or Java to “if X then A else B” in Python, in an unsupervised way. Key Takeaways Code migration or codebase portability is a tricky yet expensive decision for any organisation, and an AI assistant that takes care of the nitty-gritty dependencies within the programming languages can be quite handy. The key contributions of this work, according to the authors, can be summarised as follows: Introduction of a new approach to translating functions from a programming language to another, which is purely based on monolingual source code.TransCoder successfully manages to grasp complex patterns specific to each language and translate them to other languages.Results show that a fully unsupervised method can outperform commercial systems that leverage rule-based methods and advanced programming knowledge. Know more about TransCoder here.","excerpt":"Code migration or codebase portability is a tricky yet expensive decision for any organisation, and an AI assistant that takes care","categories":["Deep Tech"],"tags":["Google Translate","Java","Programming Languages","Python","Python Programming","translate"],"author_name":"Ram Sagar","publish_date":"2020-06-10T15:00:38","publication_year":"2020","word_count":704,"keywords":["Google Translate","AI","ML","Git","Programming Languages","RAG","Python","Ray","C++","translate","Python Programming","R","Java","Redis"],"extracted_tech_keywords":["AI","ML","Ray","RAG","Redis","Python","R","Java","C++","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/java-python-programming-language-transcoder\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":44153,"title":"A Day In The Life Of: A Research Analyst For Whom Data Science Is Not All Glam &#038; Bing","content":"In our weekly column ‘A Day In The Life Of’, we are trying to step into the shoes of awesome techies from various organisations and sectors who are working in emerging tech areas like big data, data analytics, artificial intelligence, machine learning and the internet of things, among others. This week Analytics India Magazine got in touch with Krisha Mistry, a Research Analyst who works at Decimal Point Analytics, to talk about her work, enthusiasm for higher education and her love for languages. Mistry is a driven young woman who likes to make sure that her day is productive and efficient. Her day begins at 6 am and she starts work at 7.30 am. “I have a regular nine-hour working day,” she says with a smile. Working in a fast-paced sector like data analytics, Mistry sometimes feels like things are moving too fast. She, however, has her eyes on the prize. “I am simultaneously working and studying for my Master’s degree. That is why I don’t have a lot of free time. She, however, has learned to make optimal use of her spare time. “I have always loved to learn new languages. So I actively take out time for that and pursue it,” she says enthusiastically. She adds that things have picked up pace after she finished her studies at Meghnad Desai Academy of Economics. Mistry also makes sure that her work, studies and hobbies do not claim all of her time. “I also make sure that I spend enough time with my family and friends. They are the ones who keep me grounded,” she says. “My social life exists somewhere between active and boring. There are days when I just go to work and do nothing else. Meanwhile, on some days, I go out with my friends for a drive or for dinner with my family,” she explains after a pause. When asked about her work, Mistry says, “My role is that of a Research Analyst. I basically accumulate data — micro level as well as macro-level data — for various daily and monthly reports prepared by my firm. I have to find data and information on a particular topic and present it in an organised manner. My work primarily consists of researching. Since I am a part of the Global Market Research team, I have to search and present daily market news and analysis in an orderly manner.” Mistry says that she is currently working on daily macro-economic reports right now. “This involves working on daily market reports that predominantly centre around the South African markets,” she explains. When asked about the best and the worst parts of her workday, Mistry says with a smile, “The best part of my day is the very beginning — I love taking on new assignments! And as you can imagine the boring part of my day is when the work turns out to be a little boring — I tend to get bored towards the end in such cases,” she says. We asked Mistry about whether she feels her company is making use of her abundant talent. She replied, happily, “Yes. I believe that my company is encouraging me to always put my best foot forward by assigning work that pertains to my interest and capabilities. However, I try my best to make sure that I continuously and diligently work towards it.” When asked about her plans for the future, Mistry is all ready with the answers: “In the next five years, I want to explore various opportunities that will help me develop my skills, take on interesting projects in my area of expertise, and build my career to a level that I have always worked towards. My short term goal is to develop my knowledge of the sector I am currently working in, as well as be instrumental to the progress and growth of the organisation I am affiliated with,” she says, signing off.","excerpt":"In our weekly column ‘A Day In The Life Of’, we are trying to step into the shoes of awesome techies from various organisations and sectors who are working in emerging tech areas like big data, data analytics, artificial intelligence, machine learning and the internet of things, among others. This week Analytics India Magazine got […]","categories":["AI Features"],"tags":["Data Science Career","Interviews and Discussions"],"author_name":"Prajakta Hebbar","publish_date":"2019-08-09T18:00:47","publication_year":"2019","word_count":654,"keywords":["big data","Go","machine learning","artificial intelligence","AI","RAG","Data Science Career","Aim","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","analytics","Aim","RAG","R","Go","big data","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-day-in-the-life-of-a-research-analyst-for-whom-data-science-is-not-glam-bing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10165963,"title":"Bessemer Venture Partners Announces $350 Million Fund for Indian Startups","content":"Bessemer Venture Partners has announced a $350 million fund dedicated to early-stage investments in India. This is the firm’s second India-focused fund, which will target startups in AI-enabled services, SaaS, fintech, digital health, consumer brands, and cybersecurity. Partners Vishal Gupta and Anant Vidur Puri will lead the investment strategy. The firm cites several factors behind the new fund, including the growth of India’s middle class, expansion of digital infrastructure, and increasing domestic investment. “India is rapidly moving towards being the world’s third largest economy and a trillion dollar digital opportunity fueled by strong domestic savings, growing mobile and internet penetration, rapidly improving software and AI advancements,” Puri wrote on LinkedIn. He highlighted how this growth is further supported by a strong layer of public technology initiatives such as Aadhaar, UPI, Open Network for Digital Commerce (ONDC), and others. Bessemer opened its Bengaluru office nearly 20 years ago to focus on India’s technology sector. Since then, it has invested in more than 80 startups in the country. Over the last five years, more than 80% of its India investments have been at Series A or earlier stages. Its portfolio includes BigBasket, Urban Company, Perfios, and Livspace. Nine of its portfolio companies in India have completed initial public offerings. According to the company, the investment strategy will follow Bessemer’s roadmap research process, aimed at identifying early trends and emerging sectors. Startups backed by Bessemer’s first India fund include Boldfit, MoveInSync, Pepper Content, ShopDeck, Vetic, and Zopper.","excerpt":"The fund will target startups in AI-enabled services, SaaS, fintech, digital health, consumer brands, and cybersecurity.","categories":["AI News"],"tags":["venture capital"],"author_name":"Aditi Suresh","publish_date":"2025-03-13T11:22:48","publication_year":"2025","word_count":244,"keywords":["Go","API","programming_languages:R","AI","venture capital","programming_languages:Go","Git","Aim","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","API","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bessemer-venture-partners-announces-350-million-fund-for-indian-startups\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045827,"title":"Esri India To Train Over 2 Lakh Students In GIS Technologies","content":"Esri India, a geographical information system (GIS) software and solutions provider, has announced its plans to train over 2 lakh students in three years. According to Esri India’s official release, this initiative has been taken in order to meet the rising demand of GIS skilled workforce in India, driven by crucial government projects in smart cities, AMRUT, water resources, agriculture, insurance, land management & SVAMITVA, and utilities. As per the reports, India will need 10 lakh more GIS professionals by the year 2025. Therefore to achieve this, Esri India has launched a countrywide comprehensive campus-wide program to empower universities and institutes to set up and scale their GIS learning infrastructure for students. The company stated that it aims to skill students from over 150 institutes as part of this initiative. In addition, Esri India is also planning to establish a Centre of Competence (CoC) to provide students from various streams with an opportunity to acquire the most advanced GIS technology skills. Talking about this initiative, Agendra Kumar, Managing Director, Esri India stated that, job creation is an important goal for the government as well as academia. So, the opportunities will come in many segments. Read an interview with Agendra Kumar at Esri, here. He said, “Large IT companies also have opportunities for geospatial professionals because they handle global clients and both global as well as Indian clients. In India, public sector companies like petroleum companies, utility companies, private organisations, and educational and research communities are some of the segments that will also see higher demands. That is why it’s important to create a pipeline of more skilled resources in the country.” The company said that the recent government policy changes like de-regulating geospatial data and National Education Policy 2020 have also increased the adoption of geospatial technologies. The official release stated that universities and institutes like Indian Institute of Technology (IITs) in Delhi, Bombay, Guwahati, Roorkee, along with BHU, NIIT Kurukshetra, Aligarh Muslim University, Jammu University, and Symbiosis Institute of Geoinformatics at Pune, among others, are closely working with Esri India to skill students in GIS technologies. Esri India also fosters GIS know-how through other programs, including the GIS Academia Council of India and Esri India Young Scholar Program.","excerpt":"Esri India, a geographical information system (GIS) software and solutions provider, has announced its plans to train over 2 lakh students in three years.","categories":["AI News"],"tags":[],"author_name":"AIM Media House","publish_date":"2021-08-12T11:32:10","publication_year":"2021","word_count":369,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/esri-india-to-train-over-2-lakh-students-in-gis-technologies\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":19710,"title":"Google To Launch Artificial Intelligence Research Centre In China","content":"Alphabet Inc’s Google is expanding its horizons into Artificial Intelligence (AI) by opening a research centre in China. Even though its search services are still blocked in the country, the tech giant said on Wednesday that Asia’s first such centre will be supported by Google China’s strong engineering teams. Since September 2017, Google has been recruiting machine learning engineers and cloud machine learning engineers in Beijing and preparing for the new AI lab. Fei-Fei Li, the Chief Scientist AI\/ML, Google Cloud, said in an official statement on Wednesday: I believe AI and its benefits have no borders. Whether a breakthrough occurs in Silicon Valley, Beijing or anywhere else, it has the potential to make everyone’s life better for the entire world. As an AI first company, this is an important part of our collective mission. And we want to work with the best AI talent, wherever that talent is, to achieve it. The statement also adds that Google has already hired some top experts, and will be working to build the team in the months ahead. Along with Dr Jia Li, Head of Research and Development at Google Cloud AI, Li would be leading and coordinating the research. Besides publishing its own work, the Google AI China Center will also support the AI research community by funding and sponsoring AI conferences and workshops, and working closely with the Chinese AI research community. Reports have also suggested that the Chinese government has indicated strong support for AI development and for catching up with the US. The Asian country has been considering technological innovations in AI (among others) as a key to reviving its slowing economy. They have also outlined plans to become the world’s premier AI innovation hub by 2030. Google Cloud currently doesn’t operate in China. The company would need a local partner and special licenses to establish the business there. Google’s search engine can be used in China only by using virtual private networks, or VPNs, to bypass the government’s complex internet filtering system, known as the “Great Firewall”. In 2010, Google had pulled out of China over concerns about censorship and after a cyber-attack in which some of the company’s proprietary computer code was stolen.","excerpt":"Alphabet Inc’s Google is expanding its horizons into Artificial Intelligence (AI) by opening a research centre in China. Even though its search services are still blocked in the country, the tech giant said on Wednesday that Asia’s first such centre will be supported by Google China’s strong engineering teams. Since September 2017, Google has been […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","China","Fei-Fei Li","Google","Google Cloud","Machine Learning"],"author_name":"Prajakta Hebbar","publish_date":"2017-12-13T09:54:41","publication_year":"2017","word_count":367,"keywords":["Go","funding","Google Cloud","artificial intelligence","machine learning","AI","innovation","ML","Machine Learning","Google","AI research","cloud_platforms:Google Cloud","Fei-Fei Li","AI (Artificial Intelligence)","R","China"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","R","Go","innovation","funding","AI research","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-china-artificial-intelligence\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10014134,"title":"Guide To Goutte: A Simple PHP Web Scraper","content":"Web scraping is a term used to describe a way to automatically extract data from the internet, we have seen many web scraping tools so far like BeautifulSoup with python, Diffbot without coding a GUI based tool, Puppeteer with Node.js but is it possible to scrape the data from website using PHP? Yes, Goutte made it easy for developers to use PHP to scrape data. Goutte was originally written by Fabien Potencier. He is a creator of the Symfony framework, which is now maintained by FriendsOfPHP. Goutte is a library that is based on PHP 5.5+ version and Guzzle 6+; Guzzle is a PHP HTTP client that is the requirement of Goutte framework, it is used to send HTTP requests. Some Pros about Guzzle is as follows: Simple interface for building POST requests.The Same interface can send both synchronous and asynchronous requests.PSR-7 interfaces for requests.No hard dependency on cURL, PHP streams. Read more about Guzzle here. PHP is a server scripting language. It is a very powerful tool for making dynamic and interactive Web pages. PHP is widely used and a great competitor to Microsoft’s ASP. When we talk about data extraction from the internet, PHP is the last thing that comes into mind. Goutte is based on the Symfony framework. Symfony is a set of PHP components: a Philosophy, a Web application framework, and a community – all working together in harmony. It is a PHP framework and a set of reusable components\/libraries. Symfony was created by Sensio labs and was published as free software in 2005 and was released under MIT licence. Image: credit Symfony is used by large numbers of developers and contains many great features like: Create complex web applicationsStandalone PHP Micro Framework.Very fast.Stable frameworkGood open source community & contribution. Read more about the Symfony framework here. Goutte provides a decent API to crawl websites and extract data from HTML\/XML documents. That means you can login into websites, submit forms using POST, upload a file and many more all by just using the Goutte framework at your server, you can also run this framework on a local computer. Getting Started First, let’s see how to set up a PHP environment, what are the requirements, how to install an additional framework one by one. Requirements for Goutte As discussed above, Goutte depends on : PHP 5.5+ version- download here After downloading unzip and adding the extracted directory path into the environment variable For installation procedure of PHP visit here. For checking if PHP installed properly use the below command: php --version After PHP, Download Composer from here, it is a dependency manager for PHP.Guzzle 6+ (use composer command to install), Read more. composer require guzzlehttp\/guzzle Installing Goutte Now install goutte using composer, it will add fabpot\/goutte as a required dependency in your composer.json file: composer require fabpot\/goutte Example A web app that will Scrape GitHub repository list from your account; using Goutte a php framework! The following example is taken from here, we are going to create a script that will log in to your personal Github account and scrape all the repository list into your browser. Create a project folder and name Download the Goutte library repository from GitHub using the below command and after extracting you will get a directory name “Goutte”, we are going to use this directory in further process. git clone https:\/\/github.com\/FriendsOfPHP\/Goutte.git Now use the composer command to initialize your local directory with composer.json file, we are going to install goutte dependencies init. composer require fabpot\/goutte Let’s first create a basic interface for the user so that anyone can extract their repository list from GitHub by entering their username and password into the given below form. <form method=\"POST\"> <div class=\"form-group\"> <h1>Github Repositories scraper<\/h1> <label for=\"git_email\">Email address<\/label> <input type=\"email\" class=\"form-control\" id=\"git_email\" name=\"git_email\" placeholder=\"Enter email\"> <\/div> <div class=\"form-group\"> <label for=\"git_pwd\">Password<\/label> <input type=\"password\" class=\"form-control\" id=\"git_pwd\" name=\"git_pwd\" placeholder=\"Password\"> <\/div> <button type=\"submit\" class=\"btn btn-primary\">Submit<\/button> <\/form> This is a client-side user interface from where our scraper is going to read the username and password. Check if our page is submitting the data. This is PHP code which runs first when the user hits the submit button after filling the form. if(isset($_POST[\"“git_email\"]) && isset($_POST[\"git_pwd\"]) && !empty($_POST[\"“git_email\"]) && !empty($_POST[\"git_pwd\"])){ } Import libraries and importing client.php from Goutte directory which we downloaded by using git clone command also we are importing vendor autoload.php that helps in autoloading PHP classes. require_once(\"vendor\/autoload.php\"); require_once(\"Goutte\/Goutte\/Client.php\"); $client = new Client(); Inspect Github and search for login elements: username and password. Initialize crawler variable using request and GET command on URL. $crawler = $client->request('GET', 'https:\/\/github.com\/login'); Now we know the elements where username and password tag are, so set the parameter and form: $form = $crawler->selectButton('Sign in')->form(); $form->setValues(['login' => $_POST[\"git_email\"], 'password' => $_POST[“git_pwd\"]]); $crawler = $client->submit($form); Checking if the login was successful by checking if meta tag having name “octolytics-actor-login” $username = \"\"; $crawler->filter('meta')->each(function ($node) { global $username; if(trim($node->attr(\"name\")) == \"octolytics-actor-login\"){ $username = ($node->attr(\"content\")); return; } }); Navigate the URL of the GitHub repository of the user, i.e. for example https:\/\/github.com\/mmaithani?tab=repositories $crawler = $client->request('GET', 'https:\/\/github.com\/'.$username.'?tab=repositories'); Let’s inspect the repository page. All the repository’s names are inside the ankle tag that is inside the class “source”. We can use the filter function to extract text from ankle tag using filter(li.source a) $crawler->filter('li.source a')->each(function ($node) { if(is_numeric($node->text()) === false){ echo $node->text(); echo \"<br\/>\"; } }); Before running the final script, first, let’s see the files and directories we are having, index.php is our main script which is inspired by this article. Goutte directory is essential for this project, vendors contain our autoloader PHP script, composer.json is out dependencies file. Let’s run our script in the browser. To create a lightweight server for our web app, first, open the terminal\/command prompt(CMD) in this directory and use the following commands: php -S 127.0.0.1:8000 Now go to http:\/\/127.0.0.1:8000\/, It will load the index.php automatically, and the output will be something like this, log in with your Github account and click submit. Custom GitHub login page The output will be shown in the browser, all the repositories list including private repo too because we scraped the data after login, so we have full access to user data. output Conclusion Goutte is quite fast, can imitate basic user actions, supports async requests, and even doesn’t require any browser. We saw a web application that is capable of scraping data from the GitHub account of the user, Goutte a friend of PHP which is capable of working with client-side application and also it can take user inputs and scrape accordingly with full control over the account. There are some cons of Goutte too like it doesn’t support JavaScript and also can’t take pictures as we do in Puppeteer. Indeed Goutte is a lightweight wrapper on top of the best frameworks.","excerpt":"Web scraping is a term used to describe a way to automatically extract data from the internet, we have seen many web scraping tools so far like BeautifulSoup with python, Diffbot without coding a GUI based tool, Puppeteer with Node.js but is it possible to scrape the data from website using PHP? Yes, Goutte made […]","categories":["Deep Tech"],"tags":["PHP","Web Scraping Tools"],"author_name":"Mohit Maithani","publish_date":"2020-12-14T13:00:00","publication_year":"2020","word_count":1124,"keywords":["Go","Web Scraping Tools","TPU","AI","ML","Git","Python","JavaScript","GitHub","R","Java","PHP"],"extracted_tech_keywords":["AI","ML","TPU","Python","R","JavaScript","Go","Java","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/goutte-php\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":68338,"title":"Springer Nature Asked Not To Publish A Deep Learning Paper","content":"“Machine learning does not have a built-in mechanism for investigating or discussing the social and political merits of its outputs.” Two thousand two hundred twelve expert researchers and practitioners across a variety of technical, scientific, and humanistic fields, including statistics, machine learning and artificial intelligence, law, sociology, history, communication studies and anthropology have joined hands to sign a petition to stop Springer Nature from publishing a potentially malicious research paper. The petition demanded the publisher to consider the following: Review committee should publicly rescind the offer for the publication of this specific study, along with an explanation of the criteria used to evaluate it.Springer should issue a statement condemning the use of criminal justice statistics to predict criminality and acknowledging their role in incentivising such harmful scholarship in the past.All publishers to refrain from publishing similar studies in the future. Springer Nature plans to publish an article \"A Deep Neural Network Model to Predict Criminality Using Image Processing\" that revives long discredited physiognomist pseudoscience.Sign this petition to urge @SpringerNature to refrain from publishing. RT! https:\/\/t.co\/1NP8wmLCjI— Abeba Birhane (@Abebab) June 21, 2020 The argument here is that the uncritical acceptance of default assumptions inevitably leads to discriminatory design in algorithmic systems, reproducing ideas which normalize social hierarchies and legitimise violence against marginalised groups. About The Paper A group of Harrisburg University professors and a PhD student developed automated computer facial recognition software capable of predicting whether someone is likely going to be a criminal. Their paper titled, “A Deep Neural Network Model to Predict Criminality Using Image Processing” was supposed to be published in a book series, titled “Springer Nature – Research Book Series: Transactions on Computational Science & Computational Intelligence.” “This research indicates just how powerful these tools are by showing they can extract minute features in an image that are highly predictive of criminality.”Prof. Roozbeh Sadeghian, one of the authors We would like to clarify that the article referenced here will not be published by Springer Nature.— Springer Nature (@SpringerNature) June 23, 2020 This research was touted to assist law enforcement in identifying the criminality of a person from their facial image to prevent crimes from occurring in their designated areas. The press release was followed by the petition, and immediately after the publisher confirmed in a tweet that they would not be publishing the paper. Update: Springer Nature’s Director Communication Renate Bayaz reached out to Analytics India Magazine and provided the following statement: “We acknowledge the concern regarding this paper and would like to clarify at no time was this accepted for publication. It was submitted to a forthcoming conference for which Springer will publish the proceedings of in the book series Transactions on Computational Science and Computational Intelligence and went through a thorough peer review process.  The series editor’s decision to reject the final paper was made on Tuesday 16th June and was officially communicated to the authors on Monday 22nd June. The details of the review process and conclusions drawn remain confidential between the editor, peer reviewers and authors.” The statement says that the process and conclusions drawn remain confidential. Open peer review is a standard practice in the machine learning community. The feedback of the reviewers not only help the concerned authors but also present great insights for other researchers to avoid repetition of a mistake. In the long run, closed door criticisms in a scientific domain will only fuel more skepticism within the community and might even impede the advancement of a field such as AI, which has just started to flourish. A Minority Report In The Making Tom Cruise in the movie Minority Report This self-critique must be integrated as a core design parameter, not a last-minute patch. The prevalence of bias in algorithmic solutions is nothing new. The infamous recidivism algorithm, COMPAS (which stands for Correctional Offender Management Profiling for Alternative Sanctions), a tool used for finding criminals who would repeat an offence was found to be unfair to the black defendants as they were far more likely than white defendants to be incorrectly judged for higher risk of recidivism, while white defendants were more likely than black defendants to be incorrectly flagged as low risk. Researchers are yet to figure out a way to keep the biases from creeping into the machine learning models. The ‘how’ of it still largely remains to be unknown. Using machine learning for non-critical applications such as finding friends in a group photo is harmless but if the same model is tasked with finding certain features of the face to assess the potential criminality is an obvious ethical flop show about to blow in the face. The intent of the authors might have been benevolent. They wanted to bring down the wrongdoers. However, current law enforcement isn’t perfect. These policies are just agreed upon due to lack of a better alternative. Now if we throw AI into the mix of these already cluttered institutions and if something goes wrong, will a human (researcher\/lawmaker) take responsibility or will they resort to a perpetual blame game under the guise of faulty algorithms. That said, there is also a scepticism within the community regarding the way the research was taken down. In case this whole ordeal sets a new precedent, will there be more petitions to follow? In such a scenario, will some research share the same fate of the wrongfully convicted? Ethics in AI is a problematic subject and might remain to be so. However, for AI to be accepted widely, it might need the consensus of all parties concerned. The decision-makers should consist of a body of experts who are also the representatives of marginalised communities and those who have been on the wrong end of the deal. Whether this will clear the path for AGI is yet to be determined. Even if we all can consciously agree upon a solution or the lack thereof, it can be a good starting point.","excerpt":"“Machine learning does not have a built-in mechanism for investigating or discussing the social and political merits of its outputs.” Two thousand two hundred twelve expert researchers and practitioners across a variety of technical, scientific, and humanistic fields, including statistics, machine learning and artificial intelligence, law, sociology, history, communication studies and anthropology have joined hands […]","categories":["AI Features"],"tags":["Facial Recognition"],"author_name":"Ram Sagar","publish_date":"2020-06-26T12:58:49","publication_year":"2020","word_count":985,"keywords":["Go","machine learning","artificial intelligence","Facial Recognition","AI","neural network","TPU","ML","Git","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","analytics","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/springer-nature-deep-learning-research-crime-prediction\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10144210,"title":"Bengaluru-based LatentForce.ai Enables LLM-Powered Data Extraction, Document Conversion Tasks","content":"Bengaluru-based AI-driven startup, LatentForce.ai, announced yesterday the launch of Extractors.ai – the ultimate one-stop solution for all LLM (Generative AI)-powered data extraction and document conversion tasks. “Extractors.ai is incredibly useful for anyone who frequently works with scanned documents in their daily workflow. Imagine you’re processing 100 invoices a day and need to compile all the details into an Excel sheet by the end of the day; Extractors.ai makes this task fast and effortless,” Aravind Jayendran, co-founder and CEO of LatentForce.ai told AIM. The platform offers a versatile range of data extraction tools, including converting photos into Excel files, transforming scanned copies into editable documents, and bulk processing unstructured documents to extract valuable information. For users who prioritise data privacy, the platform provides on-premise solutions that enable the setup of dedicated servers within the user’s network. This ensures that documents are processed securely while remaining entirely within the user’s infrastructure. LatentForce built Extractors.ai using AI Agents powered by Vision Language Models (VLMs). Unlike traditional LLMs, VLMs combine the ability to understand both text and images. Since scanned documents are essentially images, this dual capability allows Extractors.ai to achieve unparalleled output quality and tackle previously impossible tasks. “For example, our image-to-DOCX converter is the first tool in the world capable of transforming an image into a fully editable DOCX file while preserving the original formatting,” Jayendran explained. “Imagine you’re working on a tender and need to make quick edits to a scanned page but don’t have the soft copy. Instead of manually typing and reformatting everything, you can simply take a photo of the page and use Extractors.ai to generate an editable DOCX file in seconds.” He further added that this is just one example of how its unique AI Agents are revolutionising data extraction and processing. The company is continuously developing new tools to simplify these tasks as part of LatentForce.ai’s mission to make AI accessible and practical for enterprises of all sizes.Overall, generative AI startups in India are transforming various industries by leveraging advanced AI models such as text, images, LLMs, codes and videos. Key players like Krutrim AI, Sarvam AI, and KissanAI focus on applications like LLMs, AI tools, and agricultural technology.","excerpt":"“Imagine you’re processing 100 invoices a day and need to compile all the details into an Excel sheet by the end of the day; Extractors.ai makes this task fast and effortless,” CEO Aravind Jayendran said.","categories":["AI News"],"tags":["AI Agents","data extraction"],"author_name":"Shalini Mondal","publish_date":"2024-12-24T12:26:36","publication_year":"2024","word_count":363,"keywords":["data extraction","TPU","programming_languages:R","AI","emerging_tech:AI agents","AI Agents","RAG","Aim","AI agents","generative AI","R","startup"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","TPU","R","startup","AI agents","programming_languages:R","emerging_tech:AI agents"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/bengaluru-based-latentforce-ai-enables-llm-powered-data-extraction-document-conversion-tasks\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10045053,"title":"AI Finds A Place In Modern Wealth Management","content":"Wealth management is an investment advisory and financial solutions provider service that caters to a wide array of clients including affluent individuals and families with high net worth. As per reports, there is approximately $78 trillion of assets in motion for wealth managers to capture. This is underpinned by the global expansion of the affluent middle class, an increase in the number of women with wealth, and the wealth created by business ownership. Traditional methods of wealth management no longer hold water and wealth managers are forced to make a significant shift in their strategy and approach. In such a scenario, there is a brilliant opportunity to capture more value using artificial intelligence. AI can help wealth managers smoothly transition from theory to execution. The need to adopt artificial intelligence in wealth management is driven by the success of wealth management technology and the entrance of young tech-savvy modem wealth managers. The increasing amount of data shaping the future of the industry. Most modern wealth managers now have a digitally weak point for work. AI in wealth management The wealth management industry is data-rich and relies heavily on the data parsing process, generally performed by humans. Several banks and wealth management services such as Vanguard, Fidelity, Morgan Stanley, and more are looking to embed artificial intelligence in their work and financial advisory services to eliminate human wealth advisors completely or much more to commonly augment their efforts. To begin with, artificial intelligence tools help identify investment preferences and provide personalised and curated advice for the client. It tells the wealth manager what the client’s attitude towards risks is. Moreover, it saves the time of the wealth manager by parsing data and presenting them in the context of relationships within customers, markets, products, and client profiles. Furthermore, when you work with an algorithm, no advice and decision are instinctive. Such softwares are highly regulated and are mandated to provide proper reasoning for any advice. At large firms managing massive portfolios, wealth managers might see a reduction in time, costs and an escape from mundane tasks that would involve repetitive data interrogation if AI softwares are used to monitor data to track risks and exposure. They can provide reasoned recommendations after considering clients’ preferences, financial trading trends, and advisory services that human advisors can’t match. Most managers see it as an opportunity to get ahead in the race by moving in the right direction and adopting AI over the next two years. Use cases Morgan Stanley, a market leader in wealth management, has the most extraordinary AI intergenerational advisory infrastructure. It uses machine learning algorithms to identify investments of interest and relevance. However, the company understands that people have been sceptical about the tech applied to the sensitive field of money management, where trust is the key to relationships. Jeff McMillan of Morgan Stanley said, “There is a perception that these tools are suitable for the mass affluent segment and not the ultra-high net worth space. The argument is that such populations are too small for a trustworthy recommendation. But we can drive specific opportunities based on individualised client behaviour”. Vanguard has not applied artificial intelligence but does leverage technology to assess risks. The program it uses leverages simple algebra to translate the questionnaires to investment percentages. The company wants to focus on the end goal, and that is to finance retirements. It uses Monte Carlo simulations for that purpose, determining what the probability of the client outliving their money is. The programme recommends the steps for portfolio rebalancing, but nothing is executed without client or manager approval. Wealthfront, on the other hand, recommends its clients artificial intelligence generated robot advice. The advice has its perks; it is far cheaper than human advice, just 0.15 per cent of the investment, far less than the one per cent human price. It also makes free recommendations for financial advice below $5000. Its advice is based on a set of questionnaires translated into a customised investment portfolio further analysed by the algorithm. Wrapping up Many firms are looking to add artificial intelligence to their way of business. Such algorithms create better value and make service more reliable. When such management softwares are deployed, they significantly improve the value they create for clients. They cover an overall vast spectrum involving user journey analytics, candidate screening, documents data extraction and so on. Furthermore, they also improve the user experience. The use of NLU and NLP interfaces dramatically increases client engagement.","excerpt":"AI can help wealth managers smoothly transition from theory to execution.","categories":["IT Services"],"tags":[],"author_name":"Meenal Sharma","publish_date":"2021-08-03T10:00:00","publication_year":"2021","word_count":743,"keywords":["Go","machine learning","artificial intelligence","AI","RAG","NLP","Ray","analytics","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","analytics","Ray","RAG","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/it-services\/ai-finds-a-place-in-modern-wealth-management\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":52345,"title":"Drone Startup Flytbase Closes Seed Funding Round Successfully","content":"FlytBase, a startup based in Silicon Valley and India, recently closed its Series Seed round of venture financing, as it powers fully autonomous drone fleets for global businesses. The seed funding comes as the company gains traction for its commercial drone automation software, with a focus on high-value enterprise use-cases such as warehouse inventory management, wind turbine inspection, live remote drone operations, and public safety. Nitin Gupta, CEO at FlytBase, explained, “We have been fortunate to have support from leading incubators and angel investors since our inception. With the most recent round, we now can rapidly build upon our unique competitive position in the global drone ecosystem. Our ability to combine intelligent software with off-the-shelf, cost-effective drones positions us well with enterprise customers, as they mature their drone adoption from PoCs and pilots to fully autonomous, multi-site drone operations. We are focused on developing cutting-edge autonomous systems that solve real-world problems for our customers, at scale”. The Series Seed round is led by a US-based, early-stage seed fund, and brings not only venture capital, but also a wealth of professional experience and world-class B2B networks to the table. FlytBase holds a unique position in the global commercial drone ecosystem – its FlytOS and FlytCloud software offerings, built and enriched over the years, can enable intelligent automation of most major drone hardware platforms. Sharvashish Das, Director of Engineering, added: “With an IoT architecture, intelligent plugins and open APIs, our technology platform brings a variety of capabilities to drones – ranging from indoor autonomous navigation and obstacle avoidance to precision landing and autonomous charging. This, in turn, is enabled by our team’s expertise in control theory, multi-sensor fusion, SLAM, machine vision, artificial intelligence, machine learning, robotics and embedded systems”.","excerpt":"FlytBase, a startup based in Silicon Valley and India, recently closed its Series Seed round of venture financing, as it powers fully autonomous drone fleets for global businesses. The seed funding comes as the company gains traction for its commercial drone automation software, with a focus on high-value enterprise use-cases such as warehouse inventory management, […]","categories":["AI News"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2019-12-20T14:31:41","publication_year":"2019","word_count":287,"keywords":["API","intelligent automation","machine learning","artificial intelligence","funding","AI","venture capital","automation","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","R","API","automation","intelligent automation","startup","venture capital","funding"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/flytbase-seed-funding\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10049102,"title":"Exciting Times For AI in India: Sameep Mehta, Lead &#8211; Data and AI Platforms, IBM Research India","content":"Good training data is a prerequisite for better performance of machine learning models, and it further calls for a systematic analysis of data quality before building an AI model. Analytics India Magazine got in touch with Sameep Mehta – IBM Distinguished Engineer and Lead – Data and AI Platforms, IBM Research India. Sameep has completed his masters and PhD in Computer Science from The Ohio State University and has been working at IBM for more than 15 years. “AI for enterprises is about enabling organisations to modernise at ease, better predict outcomes, automate at scale and secure their organisations. Organisations are looking to infuse AI in the form of NLP, automation with trust and security as a key foundation across various processes and business functions. We are witnessing the demand for AI across the board, from SMBs to large enterprises and across all sectors,” said Sameep. AIM: Can you talk about how IBM Research India is devising innovative approaches to ensure quality data for having ready-to-deploy ML models? Sameep: As we know, the AI team spends the majority of its time in data collection, data cleaning, and data preparation. There are many tools for data preparation for Business Intelligence (BI) and Management of Information System (MIS) that provide traditional data cleaning methods like missing value imputations, data normalisation, etc. Even though these metrics are helpful, they do not meet the requirements of AI teams. Moreover, data quality for AI is fundamentally different from the one required for BI, e.g., how to detect and correct wrong labels in the data is a pressing problem. Wrong labels will result in poor-quality models. At IBM Research, we are building novel algorithms and toolkits for data assessment and remediation so that the downstream AI model can be more accurate, fair, and robust. While the data quality toolkit is available as a commercial offering, we have recently released a part of the toolkit as APIs on the IBM Developer Hub. Since most data assessment algorithms are complex and require non-trivial efforts to build, these APIs will enable developers to experiment with the data quality metrics and to include the metrics in their AI pipeline without spending too much development effort. Apart from algorithmic innovation, we also advocate the importance of data management issues in the AI lifecycle. Last year, we also developed a course on Data Lifecycle Management for Computer Science engineering students of Indraprastha Institute of Technology, Delhi (IIIT-Delhi) to cover topics such as handling data to build better ML pipelines. AIM: What kind of initiatives have IBM Research India taken to infuse trust, transparency, and fairness in AI platforms and algorithms? Sameep: Trust is essential to AI adoption. It allows organisations to understand and explain recommendations and outcomes and manage AI-led decisions in their business while maintaining the full ownership and protection of data and insights. A recent Morning Consult study conducted for IBM on AI adoption revealed that nearly 95% of IT professionals in India believe that it is critical or very important to their business to trust the AI’s output is fair, safe and reliable. IBM Research India is actively working to build toolkits that infuse trust in AI algorithms. We contributed to the development of the open-source toolkit AIFairness 360 that allows developers to detect and de-bias the AI models. AI explainability 360 provides technology that enables developers and other personas like Risk Officers & end customers to open up the AI model and understand the reasoning behind the decision. The open-source toolkits help our developer community to build fair and explainable AI models. At the same time, the team works very closely with IBM Products and Services to include these capabilities in the IBM portfolio for our enterprise customers. We have also introduced AI Factsheets – which provides information about a model’s important characteristics, just like nutrition labels for foods. We are also part of various external think-tanks and working groups to shape this critical topic at a broader level. AIM: What challenges do you generally face in the AI testing cycle? How do you overcome those challenges? Sameep: Typically, a good AI model is evaluated by metrics like accuracy and scale. However, just testing the model performance on these metrics may not be enough. Before deploying AI models in production, they must be thoroughly tested for other metrics like fairness, robustness, generalizability, privacy, etc., to identify any shortcomings. While the importance of testing is well accepted and practised for traditional software, it is still in a very nascent stage with respect to AI model testing. One of the core problems in testing is the non-availability of test cases. For example, the standard hold-out set may not be rich enough to test for the desired metrics. We build algorithms that generate millions of realistic and metric-driven test cases to validate the model across a wide range of properties to tackle this challenge. These test cases are then further prioritised and executed to find failure points. We are also investing heavily in techniques to help developers improve the model by providing focused recommendations by deeper analysing the failed test cases. AIM: With more than 15 years of experience, where do you think most data companies falter? Sameep: I wouldn’t say they are faltering, but data companies today are looking at newer innovations with a renewed focus, newer ways to increase investments and associated returns. There are two dimensions to this: First, the data companies invest in more state-of-the-art tools to collect, curate, and manage the data. The regulatory framework around data privacy and security is constantly changing, and the data companies are always looking to ensure they keep up with the changing demands. Else, data breaches and security leaks will be one big dampener for these companies. The second part is around value unlocking and monetisation of the data. Organisations are looking to be more flexible and open (modulo privacy) with the data. They are looking at ways to work closely with their AI team, academia, developer community, etc., to discover potential uses of enterprise data and then demonstrate how this data can solve pressing business or societal problems. So, in essence, the need for more investment in tools\/platforms to prepare high-quality data and collaboration with partners who can use the data securely are the two biggest reasons that will help data companies to succeed. AIM: How do you see\/observe the landscape of Enterprise AI evolving in India? Sameep: AI has moved from consideration to wider mainstream adoption due to the pandemic. A recent Morning Consult study conducted for IBM on AI adoption revealed that 53% of Indian IT professionals stated that their company had accelerated its rollout of AI due to the COVID-19 pandemic. We are witnessing the demand for AI across the board, from SMBs to large enterprises and across all sectors. Different patterns are emerging. The first or the most common is to improve the existing process by infusing AI. For instance, automating a manual process, generating sales leads by connecting diverse data sets, developing chat assistants and voice assistants for customer engagement, etc. The second trend, though in relatively small pockets, more impactful is to use AI to unlock the value of the existing enterprise data and establish new lines of business. The third trend is customers wanting to modernise their traditional data and applications – using AI. They are using AI tools to discover & analyse the traditional data systems and recommend modernised configuration. With so much excitement in the AI ecosystem from students, academia, developers, industry & government, these are exciting times for AI in India.","excerpt":"Trust is essential to AI adoption. It allows organisations to understand and explain recommendations and outcomes and manage AI-led decisions in their business while maintaining the full ownership and protection of data and insights.","categories":["AI Features"],"tags":["AI in India","big data quality","data privacy use cases","data quality","Interviews and Discussions","nlp in data","nlp pipeline"],"author_name":"kumar Gandharv","publish_date":"2021-09-21T16:00:00","publication_year":"2021","word_count":1262,"keywords":["Go","big data quality","machine learning","AI in India","TPU","AI","R","ML","data privacy use cases","NLP","Aim","data quality","analytics","Rust","nlp pipeline","nlp in data","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","analytics","Aim","TPU","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/exciting-times-for-ai-in-india-sameep-mehta-lead-data-and-ai-platforms-ibm-research-india\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10166733,"title":"New Consumer App ‘Together Chat’ Offers Access to DeepSeek R1 and More","content":"Earlier this week, California-based Together.ai introduced Together Chat, a consumer AI platform designed to provide seamless access to leading open-source models. The platform enables users to brainstorm, draft code, conduct web research, generate images and explore creative writing with ease. The platform integrates multiple AI models tailored for various tasks. There’s DeepSeek R1 for complex queries with web access, Llama 3.3 70B and Qwen 2.5 72B for intermediate inquiries, Qwen Coder 32B for coding assistance and Flux Schnell for image generation. Notably, DeepSeek R1 enhances reasoning capabilities, making it ideal for sophisticated problem-solving. With real-time web browsing, the platform summarises and cites sources as needed. Developers can leverage Qwen Coder 32B for code generation and debugging, while smaller models like Llama 3.1 8B offer quick responses. Built as a Progressive Web App, Together Chat operates as a native mobile app for instant AI-powered assistance. Users receive 10 daily credits for DeepSeek R1 and 50 for other models, with credits renewing each day. Meanwhile, in a post on X, open-source web scraper Firecrawl has confirmed that Together Chat is backed by it. This launch follows Together AI’s recent $305 million Series B funding, led by General Catalyst and co-led by Prosperity7, with participation from investors including Salesforce Ventures, NVIDIA, and Kleiner Perkins. Back in 2023, the company was in the spotlight for securing a $102.5 million Series A investment from NVIDIA, Kleiner Perkins and Emergence Capital.","excerpt":"The platform provides free access to top open-source models.","categories":["AI News"],"tags":["AI","Together chat"],"author_name":"Merin Susan John","publish_date":"2025-03-27T09:01:56","publication_year":"2025","word_count":235,"keywords":["Together chat","API","funding","programming_languages:R","AI","ML","RAG","DeepSeek R1","Together AI","llm_models:Llama","R"],"extracted_tech_keywords":["AI","ML","DeepSeek R1","RAG","R","API","funding","Together AI","llm_models:Llama","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/earlier-this-week-california-based-together-ai-introduced-together-chat-a-consumer-ai-platform-designed-to-provide-seamless-access-to-leading-open-source-models-the-platform-enables-users-to-brains\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10130713,"title":"MIT Researchers Develop AI Model for Early Breast Cancer Detection","content":"The Massachusetts Institute of Technology (MIT) has achieved a groundbreaking advancement in breast cancer detection. Researchers at the Jameel Clinic for Machine Learning and MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have developed an AI model capable of identifying breast cancer up to five years before clinical diagnosis. This deep learning model, named Mirai, is based on mammography and aims to detect precancerous changes in high-risk women. It predicts a patient’s risk across various future time points and can incorporate clinical risk factors like age and family history when available. Additionally, Mirai is designed to maintain consistent predictions despite minor clinical variances, such as different mammography machines. Technology Explained The AI model analysed chromatin images from 560 tissue samples across 122 patients, identifying eight distinct cell states throughout various stages of ductal carcinoma in situ (DCIS). By considering both cellular composition and spatial arrangement, the model revealed that tissue organisation is crucial in predicting disease progression. Utilising a convolutional variational autoencoder, the AI learns from simple chromatin staining images, a more cost-effective and accessible method compared to complex sequencing techniques. Remarkably, the model detected cell states associated with invasive cancer even in tissue that appeared normal to the human eye. Anand Mahindra lauds ‘Mirai’ Business tycoon Anand Mahindra highlighted this medical advancement and shared a post on X appreciating the AI’s ability to detect breast cancer five years in advance. If this is accurate, then AI is going to be of significantly more value to us than we imagined and much earlier than we had imagined… https:\/\/t.co\/5Mo2cT7X7T— anand mahindra (@anandmahindra) July 28, 2024 With current five-year survival rates for breast cancer at approximately 90% when detected early, this five-year head start on diagnosis and treatment could significantly improve patient outcomes. Breast Cancer Detection By Google’s AI Google’s AI has started detecting breast cancer in screenings where identifiable information has been removed. Google’s AI spotted cancer with greater accuracy and fewer false negatives and false positives than experts. In collaboration with Deepmind, Cancer Research UK Imperial Centre, Northwestern University and Royal Surrey County Hospital have tested the AI to help radiologists around the world to assist them in detecting breast cancer more accurately.","excerpt":"The deep learning model, named Mirai, is based on mammography and aims to detect precancerous changes in high-risk women.","categories":["AI News"],"tags":["MIT research"],"author_name":"Vidyashree Srinivas","publish_date":"2024-07-30T13:40:26","publication_year":"2024","word_count":363,"keywords":["Go","machine learning","artificial intelligence","programming_languages:R","AI","GAN","Aim","deep learning","MIT research","R","Chroma"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","Aim","Chroma","R","Go","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mit-researchers-develop-ai-model-for-early-breast-cancer-detection\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10024266,"title":"A Complete Python Guide to ANOVA","content":"Getting informative insights from the raw data in hand is vital in a successful machine learning project. The selection of the right machine learning algorithm and tuning of the model parameters to achieve better performance are possible only with proper data analytics in the pre-processing stage. Traditional statistical analysis is simple and powerful in extracting the essence out of the raw data. Statistical analysis is performed reliably and quickly with statistical software packages. The famous multi-purpose language, Python, has a great collection of libraries and modules to do statistical analysis in a lucid way. In this article, we discuss a widely used statistical tool called ANOVA with hands-on Python codes. ANOVA is one of the statistical tools that helps determine whether two or more data samples have significantly identical properties. Let’s assume a scenario- we have different samples collected independently from the same dataset for cross-validation. We wish to know whether the means of the collected samples are significantly the same. Another scenario- we have developed three different machine learning models. We have obtained a set of results, and we wish to know whether the models perform significantly in the same manner. Thus, there are many scenarios in practical applications where we may need to use ANOVA as part of data analytics. ANOVA is the acronym for Analysis of Variance. It analyzes variations among different groups and within those groups of a dataset (technically termed as population). However, there are some assumptions that the data must hold to use ANOVA. They are as follows: The data follows normal distributionThe variance of data is the same for all groups.Data among groups are independent of each other. Math concept behind ANOVA and its usage can be explored with the following hands-on Python example. Comparing Means using ANOVA Import the necessary libraries to create the environment. # import libraries import numpy as np import pandas as pd import scipy import statsmodels.api as sm from statsmodels.formula.api import ols from matplotlib import pyplot as plt import seaborn as sns Generate some normally distributed synthetic data using NumPy’s random module. While generating synthetic data, we should ensure that the standard deviation is common for all different methods. # numpy random normal arguments: (mean, std_dev, size) method_1 = np.random.normal(10,3,10) method_2 = np.random.normal(11,3,10) method_3 = np.random.normal(12,3,10) method_4 = np.random.normal(13,3,10) # build a pandas dataframe data = pd.DataFrame({'method_1':method_1, 'method_2':method_2, 'method_3':method_3, 'method_4':method_4}) data.head() Output: Before proceeding further into ANOVA, we should establish a null hypothesis. Whenever we are unable to make a solid mathematical decision, we go for hypothesis testing. ANOVA does follow hypothesis testing. Our null hypothesis (common for most ANOVA problems) can be expressed as: Means of all the four methods are the same. We know very well that the means are mathematically not the same. We set 10, 11, 12 and 13 as the means for the corresponding four methods while generating data. But from a statistical point of view, we make decisions with some level of significance. We set the most common level of significance, 0.05 (i.e. 5% of risk in rejecting the null hypothesis when it is actually true). In other words, if we set a level of significance of zero, it is a mathematical decision – we do not permit errors. In our case, we can reject the null hypothesis without any analysis, because we know that the means are different from each other. However, with many factors affecting the data, we should give some space to accept some statistically significant deviations among data. ANOVA follows F-test (We will define F-statistic shortly). If the probability of F-statistic is less than or equal to the level of significance (0.05, here), we should reject the null hypothesis. Else, we should accept the null hypothesis. Make the data frame to have a single column of values using Pandas’ melt method. df = pd.melt(data, value_vars=['method_1', 'method_2', 'method_3', 'method_4']) df.columns = [ 'treatment', 'value'] # treatment refers to the method df.sample(10) Output: Develop an Ordinary Least Squares model with the melted data. model = ols('value~C(treatment)', data=df).fit() model.summary() Output: We can jump into conclusions with this step itself. The probability score is 0.135, which is greater than 0.05. Hence, we should accept the null hypothesis. In other words, the means of all four methods are significantly the same. However, an ANOVA table can give crystal clear output for better understanding. Obtain the ANOVA table with the following code. anova = sm.stats.anova_lm(model, typ=1) anova Output: Users need to be aware that the terms groups and methods are invariably used in this example. We have come to the conclusion based on the Probability score. However, we can also arrive at the conclusion based on the F-statistic also. We can calculate the critical value of F-statistic with the following code. # arguments: f(numerator degrees of freedom, denominator degrees of freedom) # arguments: ppf(1-level of significance) scipy.stats.f(3,36).ppf(0.95) Output: If the observed F-statistic is greater than or equal to its critical value, we should reject the null hypothesis. Else, if the observed F-statistic is less than its critical value, we should accept the null hypothesis. Here the observed value 1.975314 is less than the critical value 2.86626. Therefore, we accept the null hypothesis. We can visualize the actual data to get some better understanding. sns.set_style('darkgrid') data.plot() plt.xlabel('Data points') plt.ylabel('Data value') plt.show() Output: We can see a great overlap among different data groups. This is exactly where we cannot jump into conclusions in a mathematical way. Statistical tools help take successful business decisions in these tough scenarios. How does Means vary among different groups? Let’s visualize it too. data.mean(axis=0).plot(kind='bar') plt.xlabel('Methods') plt.ylabel('Mean value') plt.show() Output: Though we see some differences in the mean values with human eyes, statistics say there are no significant differences in the mean values! A Major Limitation of ANOVA There is a big problem with the ANOVA method when we reject the null hypothesis. Let’s study that with some code examples. Increase the mean value of method_4 from 13 to 15. # Alter the mean value of method_4 method_1 = np.random.normal(10,3,10) method_2 = np.random.normal(11,3,10) method_3 = np.random.normal(12,3,10) method_4 = np.random.normal(15,3,10) data = pd.DataFrame({'method_1':method_1, 'method_2':method_2, 'method_3':method_3, 'method_4':method_4}) data.head() Output: Melt the data to have single-columned values. df = pd.melt(data, value_vars=['method_1', 'method_2', 'method_3', 'method_4']) df.columns = [ 'treatment', 'value'] df.sample(10) Output: Develop the Ordinary Least Squares model. model = ols('value~C(treatment)', data=df).fit() model.summary() Output: Obtain the ANOVA table. anova = sm.stats.anova_lm(model, typ=1) anova Output: Since the probability score is less than the level of significance, 0.05, we do reject the null hypothesis. It means that at least one mean value is different from the others. But we cannot identify the method or methods whose means are different from the others. This is where ANOVA needs some other methods to bring light upon its decisions. This issue can be tackled with the help of Post Hoc Analysis. Post Hoc Analysis Post Hoc Analysis is also known as the Tukey-Kramer method or the Tukey test or the Multi-Comparison test. Whenever we reject the null hypothesis in an ANOVA test, we explore individual comparisons among the mean values of different groups (methods) using the Post Hoc Analysis. Import the necessary module from the statsmodels library. from statsmodels.stats.multicomp import MultiComparison comparison = MultiComparison(df['value'], df['treatment']) tukey = comparison.tukeyhsd(0.05) tukey.summary() Output: This method performs ANOVA individually between every possible pair of groups. It yields individual decisions with probability scores. Here, the null hypothesis is accepted (means are significantly the same) for the pairs: method_1 and method_2 method_1 and method_3 method_2 and method_3 On the other hand, null hypothesis is rejected (means are significantly different) for the pairs: method_1 and method_4 method_2 and method_4 method_3 and method_4 Hence, we can conclude that methods 1, 2 and 3 possess significantly the same means while method 4 differs from them all. Note: We have generated data with NumPy’s random module without any seed value. Hence, the values and results in these examples are not reproducible. This Colab Notebook has the above code implementation. Wrapping up In this article, we discussed the importance of statistical tools, especially ANOVA. We discussed the concepts of ANOVA with hands-on Python codes. We also studied the limitations of ANOVA and the Post Hoc Analysis method to overcome the same. Now, it is your turn to perform ANOVA with the raw data in your hand! References: Read on WikiWatch in YouTubeMore examples","excerpt":"ANOVA is one of the statistical tools that helps determine whether two or more data samples o have significantly identical properties","categories":["Deep Tech"],"tags":["Data Analytics","data preprocessing","Guide","Python","statistical significance","Statistics"],"author_name":"Rajkumar Lakshmanamoorthy","publish_date":"2021-04-20T11:00:00","publication_year":"2021","word_count":1381,"keywords":["NumPy","machine learning","TPU","data preprocessing","Statistics","AI","statistical significance","Python","Colab","Seaborn","analytics","Data Analytics","Matplotlib","Guide","Pandas"],"extracted_tech_keywords":["AI","machine learning","analytics","Colab","Pandas","NumPy","Matplotlib","Seaborn","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-complete-python-guide-to-anova\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10112679,"title":"What do Women in AI Really Want?","content":"“We are underfunded, and we are exhausted. Women [in AI] do not want private jets and yachts. What we deserve is equal and equitable financial support , not just applauds and kudos” explained Mia Shah-Dand in a recent interview with AIM. “There is an imbalance in how the money is distributed. It is a self-fulfilling cycle of men getting more attention, which gets them more funding, which in turn leads to more attention and subsequently even more funding.” the founder of the Women in AI Ethics initiative added. Notably, OpenAI CEO Sam Altman has asked for a $7 trillion investment to build the future of AI collectively. The number is unfathomable and makes one wonder if it’s just another episode of Altman messing around. But then, what could the potential investors from the Middle East and Southeast Asia be talking about to the OpenAI founder? “Funding shifts the power dynamic,” Dand firmly stated. She mentioned Altman’s fame: “He is talked about because he has clout and access to billions in funding. Likewise, there are so many women doing great work but barely getting a fraction of that funding. Even though their work is lauded, the money lands in the hands of the men in the field,” she said. The (uncomfortable) truth is: Female-founded AI startups get just 2% of funding deals (at least in the UK and the US). Fortunately, in the US, the scene is changing gradually as AI companies with at least one female founder have steadily increased over the past few years, according to Crunchbase data. But usually, all-women teams are much, much lesser paid. Platforms giving a voice to these women building technology have certainly helped. One of them is WAIE, which Dand started when she realised only women were talking about AI ethics, and nobody was talking to those women doing all this work. “Fast forward to five years later, everybody’s talking about AI ethics and adding it to their LinkedIn profiles,” she pointed out. The problem persists as everybody has started talking about ethical AI but what is rarely mentioned are the sacrifices made by that the pioneers in this field like Dr. Timnit Gebru and Margaret Mitchell, who were fired  for raising awareness about the issues in AI, Dand mentioned. The team of AI ethicists co-led by Gebru were forced out of Google four years ago for writing a research paper about the risks of large language models. It was Dr.Joy Buolamwini’s research that showed that commercial facial recognition systems areto be less accurate at identifying women and people of colour, which can end up discriminating against them. Half a decade later, the companies have also started working on the issues and continue to sift throughbut progress remains slow. “If we can’t survive today, how are we even going to make it to the dystopian future that these tech billionaires are planning for us?” she chuckled quipped. “We have a long way to go before their doomsday scenario of robots coming to life, and the rich moving to Mars. For the average person today, AI is in the surveillance camera on the street,” Dand added. “Whether you get hired or get a kidney transplant depends on an algorithm. You can’t even find love without an algorithm today,” she laughed wryly. Change the Narrative The New York Times’ list of the who’s who of AI, where only a dozen men were named, is a classic example of sidelining female researchers in AI. The issue is stubborn since several researchers have been deprived of credit for their work across history. “You have to give credit where it’s due,” Dand asserted. Disserting how the media covers AI researchers, Dand noted, “When men are interviewed, they are not asked questions about their role as a father or a brother; the default perception is of men as experts. Women have to work much harder to be acknowledged and perceived as technology experts.” The white male saviorism is not a new phenomenon in Silicon Valley. “There’s a tendency to assume that if a man does something it benefits everybody. When women do something, they are pigeonholed into DEI or diversity bucket,” Dand said. “There are two things: recognition of women’s contributions in the field of AI and change the narrative  to “women’s issues are humanity’s issues.” she suggested. Focus on Deepfakes Speaking about the most problematic tech for women today, Dand points to deepfakes. “We are talking about it, but we need more than talk. We need serious funding for research and development of tools to address this ongoing crucial issue.” she said. “Taylor Swift is a high profile example, but there are many more women out there who are not celebrities, and AI is being weaponised against them,” she rightly said, since women are targeted even more often than politicians through deepfakes. Advising about how to address the challenges, the tech entrepreneur said, “Educational institutions and philanthropic organisations can play a vital role.” She added that these institutions must act as more than an extension of the tech companies and billionaire’s pet projects. By championing a technosolutionist agenda, the whole space becomes problematic and loses sight of their human-centric mission. At the end of the interview, Dand suggested that “We should celebrate women’s accomplishments, but individual success is not a substitute for collective progress.” We need to reject the need for women to be “one of the boys” and make it easier for them to help other women without being accused of favouritism. People underestimate the power of solidarity and equitable funding in overcoming systemic barriers.”","excerpt":"Female-founded AI startups get just 2% of funding deals","categories":["AI Features"],"tags":["Interviews and Discussions","Top Trend","Women in Tech"],"author_name":"Tasmia Ansari","publish_date":"2024-02-13T14:30:20","publication_year":"2024","word_count":928,"keywords":["Top Trend","Anthropic","Go","OpenAI","AI","ML","RAG","Aim","Women in Tech","ViT","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","OpenAI","Anthropic","Aim","RAG","R","Go","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-do-women-in-ai-really-want\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":68871,"title":"How This Startup Is Using AI To Become A One-Stop Shop For Every Educational Requirement","content":"The pandemic has turned the crisis into an opportunity for the EdTech Industry as the majority of schools are adopting online classes in order to continue as per their academic calendar. The e-Vidya initiative by the finance minister is an encouraging move and opens new avenues for EdTech players. With this initiative, e-learning will get more organised in India, and the adoption rate will go up across the educational institutions with an adaptive and integrated system to make an extensive ecosystem of online learning. Following a similar vision, New Delhi-based Admission24 has started the initiative to become a one-stop solution for every educational requirement of a parent, children, teacher, colleges, universities, etc. with the focus of delivering ease and convenience through digital transformation. Founded in 2019 by Abhinav Sekhri and Deepanka Sekhri, Admission24 was set up with a vision to provide quality education across nooks and corners of the country. The company has been facilitating educational institutions with the pool of digital services like online virtual classes, online payment of fees, and buying study material and books online. Flagship Product Admission24 recently launched a live virtual class solution keeping the requirement of the educational institutions in mind. The virtual classroom solution allows teachers to conduct unlimited live and interactive online classes for up to 1000 students in each session. During the live session, the students and teachers will be able to access virtual tools like online attendance, whiteboard, audio questions post-session, live chatbox option, virtual homework assignment, and other such. How Is It Different From Other Products In The Market? To this question, Abhinav said, “We are one of the few players to be providing features like recording the class to teachers in our virtual live class solution. With this feature, teachers are able to record the session for students, which will be available on the app for 48 hours. This feature was introduced so that no students can miss the important session because of internet bandwidth.” He added, “Also, one of the differentiators from our competitors is our ability to provide solutions and services with an affordability factor especially to tier 2, 3, 4 cities as per their needs and requirements.” Use Of AI And ML The company is using emerging technologies like AI and ML on various features that include LMS (Lead Management System), online virtual classes, fees management, GPS tracking, Online Attendance, etc. Abhinav stated, “We are using a limited memory AI rather than a reactive machine AI where we are successfully able to forecast the trends and success rate of our user base from the existing market as well as the potential market.” He added, “The same is accompanied by an Unsupervised Machine Learning algorithm that is self-efficient in understanding the user behaviour patterns and choices and accordingly displays the results with respect to selection of institutions on the parameters of reviews, fees, location etc. And also on the basis of course preferences.” Core Tech Stack At Admission24, the tools used by developers are mentioned below: – HTML, CSS, JavaScript, Bootstrap, JQuery are used for the front-end along with some other tools like Adobe For the backend, the company is using MySQL, LARAVEL, and PHPBesides, they use AI technologies that have features like facial recognition. The company has also rolled out its solution that surpasses live testing centres for security and reliability. Tackling Hiring Phase The company has a wide pool of developers, testing professionals, coders, digital marketing experts, content writers, bloggers, data entry specialists in the team. Admission24 has hired people from backgrounds like Telesales, IT, Tech, App Developers, Animators, Creative Designers, SEO\/SEM Experts, and Sales. Some of the essential attributes the company looks for in applicants are productivity, politeness, open-minded, passion for the job, hardworking, and adaptable mindset. Future Roadmap Talking about the user base, the company has onboarded more than 1,57,000 students on the platform within 2 months and has a total user base of more than 3.5 lacs students with more than 50,000 educational institutions. Abhinav said, “We are in a process to launch interactive virtual solutions for students and schools while keeping the value proposition in mind. We have plans to go global and add international universities on our platform in order to help student looking for overseas education.” He added, “Also, we are working on bringing at least 400-500 clients on board in the next 3 months. Our team is pretty confident and aggressive towards this without compromising on the value we drive.”","excerpt":"The pandemic has turned the crisis into an opportunity for the EdTech Industry as the majority of schools are adopting online classes in order to continue as per their academic calendar. The e-Vidya initiative by the finance minister is an encouraging move and opens new avenues for EdTech players.  With this initiative, e-learning will get […]","categories":["AI Trends"],"tags":["Startups"],"author_name":"Ambika Choudhury","publish_date":"2020-07-03T10:00:00","publication_year":"2020","word_count":743,"keywords":["Go","machine learning","AI","ML","Git","RAG","SQL","JavaScript","Startups","R","Java"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","R","SQL","JavaScript","Go","Java","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-this-startup-is-using-ai-to-become-a-one-stop-shop-for-every-educational-requirement\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10045150,"title":"Explained: NVIDIA’s Record-Setting Performance On MLPerf v1.0 Training Benchmarks","content":"Last June, MLCommons, an open engineering consortium, released new results for MLPerf Training v1.0, the organisation’s machine learning training performance benchmark suite. The latest version includes vision, language and recommender systems, and reinforcement learning tasks. MLCommons started with the MLPerf benchmark in 2018 in collaboration with its 50+ founding partners, including global technology providers, academics and researchers. Since then, it has rapidly scaled to measure machine learning performance and promote transparency of machine learning techniques. MLPerf training measures the time it takes to train ML models to a standard quality target in various tasks, including image classification, NLP, object detection, recommendation, and reinforcement learning. The full system benchmark tests machine learning models, software, and hardware. NVIDIA has submitted its training results for all eight benchmarks. It has improved up to 2.1x on a chip-to-chip basis and up to 3.5x at scale, compared to its last MLPerf v0.7 submissions. The company set 16 performance records with eight on a per-chip basis and eight at-scale training in the commercially available solutions. Besides NVIDIA which dominates the latest MLPerf results, MLPerf training received submissions from Dell, Fujitsu, Gigabyte, Intel, Lenovo, Nettrix, PCL etc. Below table shows all the benchmarks submitted by NVIDIA: (Source: NVIDIA) NVIDIA said this is the second MLPerf training round featuring NVIDIA A100 GPUs. NVIDIA engineers have developed a host of innovations to achieve new levels of performance by: Extending CUDA Graphs across all benchmarks. CUDA Graphs offers a new model for work submission in CUDA. NVIDIA has launched the entire sequence of kernels as a graph on the GPU in MLPerf v1.0, thereby minimising communication with the CPU. Using SHARP to double the adequate interconnect bandwidth between nodes. Scalable Hierarchical Aggregation and Reduction Protocol (SHARP) offloads collective operations from the CPU to the network and removes the need for sending data multiple times between endpoints. Leveraging CUDA graphs and SHARP, the team increased the scale to record the number of 4096 GPUs used to solve a single ‘AI network.’ Further, spatial parallelism enabled them to split a single image across eight GPUs for massive image segmentation networks (3D U-Net) and use more GPUs for higher throughput. In terms of hardware, the new HBM2e GPU memory on the NVIDIA A100 GPU increased memory bandwidth by nearly 30 percent to 2 TeraBytes per second (TBps). All the software used for NVIDIA submissions is currently available from the MLPerfr repository. In its blog post, NVIDIA offered insights into many of the optimisations used to deliver outstanding scale and performance. At scale training To support large-scale training, NVIDIA made advances on both software as well as hardware performance. On the hardware front, the key building block of its at-scale training is the NVIDIA DGX SuperPOD. On the software side, the team has released v 21.05 enhances to enable several capabilities, including distributed optimiser support enhancement; improved communication efficiency with Mellanox HDR Infiniband and NCCL 2.9.9; and added SHARP support. Workloads NVIDIA highlighted several optimisations for selected individual MLPerf workloads, including: Recommendation (DLRM) NLP (BERT) Image classification (ResNet-50 v1.5) Image segmentation (3D U-Net) Optimized CuDNN kernelsObject detection (lightweight – SSD) Object detection (heavyweight – Mask R-CNN)Speech recognition (RNN-T) Recommendation Today, recommendation is arguably the most pervasive AI workload in data centres. NVIDIA MLPerf DLRM submission was based on ‘HugeCTR,’ a GPU-accelerated recommendation framework part of the NVIDIA Merlin open beta framework. HugeCTR v3.1 beta version added the following optimisations: Hybrid embedding, optimised collectives, optimised data reader, overlapping MLP with embedding and whole-iteration CUDA graph. NLP Currently, BERT is one of the most important workloads in the NLP domain. In the latest version of the MLPerf round, NVIDIA improved their v0.7 submission with the following optimisations: fused multihead attention, distributed LAMB, synchronisation-free training, CUDA graphs in PyTorch and one-shot all-reduce with SHARP. Image classification In this edition of MLPerf, the team continue to optimise ResNet by improving Conv+BN+ReLu fusion kernels in CuDNN, along with DALI optimisations, MXNet fused BN+ReLu and improved MXNET dependency engine improvements. Image segmentation In this round of MLPerf training, U-Net3D is one of the two new workloads. For this, NVIDIA used the following optimisations: spatial parallelism, asynchronous evaluation, data loader, and channels-last layout support. Optimized CuDNN kernels 3D U-Net has multiple ‘encoder’ and ‘decoder’ layers with small channel counts. Leveraging a typical tile size of 256×64 for kernels used in these operations results in significant tile-size quantisation effects. CuDNN added kernels optimised for smaller tiles sizes with better cache reuse, thereby helping 3D U-Net achieve better compute utilisation. Also, 3D U-Net benefited from the optimised BatchNorm + ReLu activation kernel. The asynchronous dependency engine implemented in CuDA Graphs, MXNet, SHARP helped performance significantly. “With the array of optimisations made for 3D U-Net, we scaled to 100 DGX A100 nodes (800 GPUs), with training running on 80 nodes (640 GPUs) and evaluation running on 20 nodes (160 GPUs),” said the NVIDIA team. Compared to the single-node configuration, the max-scale configuration of 100 nodes got over 9.7x speedup. Object detection (lightweight) Lightweight SSD has been featured fourth time in MLPerf. In the latest round, NVIDIA said the evaluation schedule is changed to happen every fifth epoch, starting from the first. However, in the previous version, the evaluation schedule started from the 40th epoch. “Even with the extra computational requirement, we sped up our submissions time by more than x1.6,” said NVIDIA researchers. SSD consists of many smaller convolution layers. NVIDIA said the benchmark was particularly affected by the improvements to the MXNet dependency engine, CUDA graphs, and the enablement of SHARP via more efficient configurations and optimised evaluation. Object detection (heavyweight) Here, NVIDIA optimised object detection with the following methods: CUDA Graphs in Pytorch, removing synchronisation points, asynchronous evaluation, dataloader optimisation, and better fusion of ResNet layers with CUBNN v8. Speech recognition In this round of MLPerf, speech recognition with RNN-T is another new workload after U-Net 3D. For this, NVIDIA used the following optimisations, including Apex transducer loss, Apex transducer joint, sequence splitting, batch splitting, batch evaluation with CUDA Graphs, and more optimised LSTMs is cuDNN v8. Wrapping up MLPerf v1.0 highlights the continuous innovation happening in the artificial intelligence and machine learning space. Since the MLPerf training benchmark was launched two-and-a-half years ago, NVIDIA performance has increased by nearly 7x.","excerpt":"NVIDIA has submitted its training results for all eight benchmarks.","categories":["Global Tech"],"tags":["BERT","Image Classification","image segmentation","Machine Learning","MLPerf","NLP","NVIDIA","Object Detection","recommendation","Speech Recognition"],"author_name":"Amit Naik","publish_date":"2021-08-04T12:00:00","publication_year":"2021","word_count":1039,"keywords":["MLPerf","BERT","Ray","object detection","Speech Recognition","CUDA","artificial intelligence","PyTorch","RAG","NLP","NVIDIA","Image Classification","machine learning","AI","ML","Machine Learning","Object Detection","image segmentation","recommendation"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","Ray","PyTorch","RAG","object detection","CUDA"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/explained-nvidias-record-setting-performance-on-mlperf-v1-0-training-benchmarks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":49716,"title":"What The Layoffs By Cognizant &#038; Infosys Mean For India&#8217;s Economy","content":"We have seen quite a few cases of layoffs recently in India, including many companies from the IT sector. The list contains Cognizant, Infosys, Deutsche Bank, Capgemini and UiPath, among others, and highlights the fact that economic recession may be upon us soon. A ripple effect starting with a slowdown in the economy post the NBFC crisis, leading to a terrible year for the automotive industry has now reached the IT\/ITES which too is showing signs of slowing growth. There is also a highly pronounced cautionary approach by companies in terms of hiring and retaining talent. As the technology segment of the Indian economy has a bigger dependence on global clients, IT-based outsourcing is the place most of the income lump originates from. On the off chance that the downturn hits the world, the Indian IT market may see more harm. The recent layoffs by tech companies like Cognizant and Infosys highlight the recession fears are real, and the situation may be worsening for the large IT players. Layoffs: Optimising Costs For The Workforce Companies are making fast steps to cut back any extra costs related to jobs. Here, costs and automation are playing a key and helping rejig the workforce. Let us take Cognizant’s example — a company which has 75% of the organisation’s workforce (290,000-worldwide) in India alone. Cognizant announced a net firing of 7,000 in its workforce in India, primarily due to a renewed focus away from content as well as cost optimisation. The company has faced criticism due to the fact that it reported a decent quarterly profit and yet announced layoffs.  For the third quarter which ended September, Cognizant registered a revenue of $4.25 billion, up 4.2% from the comparable quarter last year. On the hiring side, Cognizant workforce has now honed its attention on freshers. It has expanded compensations for entry-level employees and is laying off mid-and senior-level workers to account for talented youngsters. Cognizant’s cost optimisation by hiring more freshers comes as a silver lining though raising worries of professional stability for the mid-level employees. The company has told it may see an annual cost savings of $500-550 million by optimising workforce at the senior level and also exit from a business that is not in line with its strategy such as content moderation. The total number at layoffs at Infosys amount to 4,000-10,000 people and the company has reasoned that as a high-performance organisation, involuntary attrition is crucial for the usual course of business and should not be interpreted as any mass trimming across any level. The layoffs were needed for the company due to a strong focus on cost optimisation and rising needs for technology skills. It too has renewed its focus on hiring skilled freshers to not only save costs but also renew tech skills. The Positive Side For India’s Job Scenario But, if we look at the positive side, innovation is also bringing in hope for new jobs. Technology will trickle down to SME businesses to help create new business models or bring new efficiency to the existing ones. Talking about job creation through tech, Software Technology Parks of India (STPI) is expecting to make around 15,000 direct jobs through North East BPO Promotion Scheme (NEBPS).  STPI is an association under the Ministry of Electronics and Information Technology (MeitY). Government of India has invested ₹50 crore to boost the foundation of 5,000 seats in regard to BPO\/ITES activities in the area. Similar investments are being made across different states in collaboration with government and industry bodies. Tamil Nadu, for example, got ₹6,500 crore as investments in the IT and ITeS industry in the last one year and it has led to about 60,100 new jobs in Tamil Nadu during the time frame, stated Chief Minister Edappadi K Palaniswami. Reliance which has around 6,000 programming engineers, is also focusing on enlisting a lot more to achieve its innovation projects relating to artificial intelligence, telecom, blockchain, IT services. Toward this path, Reliance has likewise put resources into 14 new businesses to help drive various initiatives to make AI-first economy, and focus on SMEs of India. The company said would hire about 6,000 more employees in this financial year. The SMEs in India incorporate 63.4 million units and records for almost 30% of India’s GDP, utilising around 460 million individuals. The division likewise compensates for 33.4% of India’s yield in manufacturing, offering work to around 120 million Indians, as indicated by CII. In fact, the promise of technology is so significant that that digitisation of SMEs could grow their commitment to India’s GDP by 10% points, driving it up to 46-48% by 2020. And it’s not just a few instances of hiring. In fact, the net hiring by the top three IT companies —  TCS, Infosys and Wipro climbed 59% in the September quarter compared to the June quarter. TCS, Infosys and Wipro, together with hired 28,157 employees against 16,687 in Q1FY20. Here, the hiring seems to be more focused on freshers, as discussed previously. Startups Are Manufacturing Jobs At A Decent Rate With the extension of more than 1,300 new businesses this year, India continues bracing its circumstance as the third greatest startup framework on the planet, as demonstrated by IT industry body Nasscom. India similarly observed the development of seven Unicorns this year till August taking the hard and fast tally to 24, which again is third beside US and China. There are around 430,000 employments from startups and 60,000 of them in 2019, as per Nasscom-Zinnov report. India’s startup environment can possibly make up to 12.5 lakh direct employment by 2025, from 3.9-4.3 lakh direct occupations in 2019, as indicated by another report from industry body Nasscom. Indicator: Layoffs Also Stress Reskilling As the world gets ready for an economic downturn, trade wars between the US and China are adding to the uncertainty. India is in a turbulent circumstance, also with a slowing financial circumstance, as shown by the falling quarterly GDP rate. A significant level of non-performing assets (NPAs), NBFC emergency and moderate credit development stand before the economy. The bad news is recession coupled with automation may see more India layoffs in the coming days, particularly in industries where repetitive processes are involved. The reason is Indian companies will focus on cost optimisation by using automation tools. The major task, in that case, is the investments in infrastructure and reskilling as new-age skills will drive the creation of new jobs. Cutbacks over various areas are representing a test that requires reskilling and vocation support for workers. In a developing nation like India, there are open doors for job creation as there is no lack of opportunities. It, of course, requires astute planning and reconsidering the job market in alignment with new technologies.","excerpt":"We have seen quite a few cases of layoffs recently in India, including many companies from the IT sector. The list contains Cognizant, Infosys, Deutsche Bank, Capgemini and UiPath, among others, and highlights the fact that economic recession may be upon us soon. A ripple effect starting with a slowdown in the economy post the […]","categories":["IT Services"],"tags":["Cognizant","Infosys","Layoffs"],"author_name":"Vishal Chawla","publish_date":"2019-11-12T11:28:00","publication_year":"2019","word_count":1125,"keywords":["Go","AI-first","artificial intelligence","Layoffs","Infosys","AI","innovation","Cognizant","Git","automation","ViT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Git","GAN","ViT","automation","AI-first","innovation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/layoffs-cognizant-infosys-india-economy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10021267,"title":"Hands-On Guide To Automated Feature Selection Using Boruta","content":"Feature selection is one of the most crucial and time-consuming phases of the machine learning process, second only to data cleaning. What if we can automate the process? Well, that’s exactly what Boruta does. Boruta is an algorithm designed to take the “all-relevant” approach to feature selection, i.e., it tries to find all features from the dataset which carry information relevant to a given task. The counterpart to this is the “minimal-optimal” approach, which sees the minimal subset of features that are important in a model. It is originally an R package that has been recoded in Python with some additions and improvements: Faster run timesScikit-learn like interface, it uses fit(X, y), transform(X), or fit_transform(X, y), to run the feature selection.Compatible with any ensemble method from scikit-learnAutomatic n_estimator selectionRanking of featuresGini impurity is used to derive the importance of features instead of the RandomForest R package’s MDA. Algorithm Here’s the algorithm behind Boruta, as mentioned in the paper: Extend the information system by adding copies of all variables (the information system is always extended by at least 5 shadow attributes, even if the number of attributes in the original set is lower than 5).Shuffle the added attributes to remove their correlations with the response.Run a random forest classifier on the extended information system and gather the Z scores computed.Find the maximum Z score among shadow attributes (MZSA), and then assign a hit to every attribute that scored better than MZSA.For each attribute with undetermined importance perform a two-sided test of equality with the MZSA.Deem the attributes significantly lower than MZSA as ‘unimportant’ and permanently remove them from the information system.Deem the attributes which have importance significantly higher than MZSA as ‘important’.Remove all shadow attributes.Repeat the procedure until the importance is assigned for all the attributes, or the algorithm has reached the previously set limit of the random forest runs. Basically, a number of randomly shuffled shadow attributes are created to establish the baseline performance. A hypothesis test is then used to determine whether a variable is only randomly correlated or carries significant information. This test, by default, is carried out with a significance level of .05, this can be changed using the alpha argument when creating a BorutaPy object. Variables that fail to reject this hypothesis are discarded. As Boruta iteratively removes uninformative variables, the feature importance of the remaining relevant variables will improve. The comparatively noisier variables will see larger improvements. This happens because the random variables that the comparatively noisier relevant variables were correlated with have been discarded from the dataset because the noise, the random variables. Here’s a plot from the original paper illustrating the evolution of relevance(Z) score, pay close attention to the one glimmering green line amidst the red mess in the first round of Boruta run. Z score evolution during Boruta run. Green lines correspond to confirmed attributes, red to rejected ones, and blue to respectively minimal, average, and maximal shadow attribute importance. Requirements numpyscipyscikit-learn Installation pip: pip install Boruta Conda: conda install -c conda-forge boruta_py Using Boruta for feature selection Importing Boruta and other required libraries. import pandas as pd import numpy as np from sklearn.ensemble import RandomForestClassifier from boruta import BorutaPy from sklearn.model_selection import train_test_split from sklearn.metrics import accuracy_score Loading the dataset, separating the features from the target variable, and splitting the data into a train and a dev set. URL = \"https:\/\/raw.githubusercontent.com\/Aditya1001001\/English-Premier-League\/master\/pos_modelling_data.csv\" data = pd.read_csv(URL) data.info() X = data.drop('Position', axis = 1) y = data['Position'] X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = .2, random_state = 1) Creating a baseline RandomForrestClassifier model with all the features. rf_all_features = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) rf_all_features.fit(X_train, y_train) accuracy_score(y_test, rf_all_features.predict(X_test)) Creating a BorutaPy object with RandomForestClassifier as the estimator and ranking the features. One important thing to note here is that Boruta works on NumPy arrays only rfc = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) boruta_selector = BorutaPy(rfc, n_estimators='auto', verbose=2, random_state=1) boruta_selector.fit(np.array(X_train), np.array(y_train)) print(\"Ranking: \",boruta_selector.ranking_) print(\"No. of significant features: \", boruta_selector.n_features_) Boruta has selected 31 features, the features with rank 1 are selected. Let’s create a table and see exactly what features were rejected. selected_rf_features = pd.DataFrame({'Feature':list(X_train.columns), 'Ranking':boruta_selector.ranking_}) selected_rf_features.sort_values(by='Ranking') Using the BorutaPy object to transform the features in the dataset. X_important_train = boruta_selector.transform(np.array(X_train)) X_important_test = boruta_selector.transform(np.array(X_test)) Creating another RandomForestClassifier model with the same parameters as the baseline classifier and training it with the selected features. rf_boruta = RandomForestClassifier(random_state=1, n_estimators=1000, max_depth=5) rf_boruta.fit(X_important_train, y_train) accuracy_score(y_test, rf_boruta.predict(X_important_test)) GitHubPaperCollab Notebook","excerpt":"Boruta is a Python package designed to take the “all-relevant” approach to feature selection.","categories":["AI Trends"],"tags":["automated machine learning","feature selection","machine learning pipeline"],"author_name":"Aditya Singh","publish_date":"2021-03-07T13:00:00","publication_year":"2021","word_count":728,"keywords":["scikit-learn","NumPy","machine learning","AI","ML","RAG","Python","Ray","feature selection","machine learning pipeline","R","automated machine learning","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","Ray","scikit-learn","Pandas","NumPy","RAG","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/hands-on-guide-to-automated-feature-selection-using-boruta\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10015240,"title":"Top Data Science Education Initiatives By Institutions In 2020","content":"While normal education suffered a standstill in 2020, there were a lot of online courses and programs that were initiated by some of the most prestigious institutions as well as big tech giants so that the process of learning and skill development doesn’t suffer. As the trend has been for a few years now, some of the most interesting initiatives were seen in the field of data science. In this article, we have listed some of the prominent data science education programs and initiatives in 2020. Microsoft’s Learning Modules Based On Netflix Movie Microsoft, in collaboration with Netflix, has launched three new learning modules on beginners concepts in data science, along with machine learning and artificial intelligence. The design of these courses is inspired by the Netflix original film — ‘Over The Moon,’ where a young girl Fei Fei, who builds a rocket to the moon, embarks on a mission to prove the existence of Moon Goddess. Like the film, the modules are based on different missions — a moon mission using Python Pandas library, predicting meteor showers by Python and VC Code, and recognising objects in images using Azure Custom Vision. Notably, Microsoft, earlier this year, had also released a few learning modules inspired by real NASA engineers and scientists. Price: Free of cost Know more about it here. Stanford Offers Projects in Data Science Stanford invited experienced candidates in machine learning and statistics with a biological background to participate in a project to study the biology and epidemiology of the COVID-19 virus. This project class will investigate and model the virus using data science and machine learning tools. Know more about it here. World’s First Doctorate Of Business Administration In Data Science The Rennes School of Business, France, in partnership with the International School of Engineering (INSOFE), India, has launched the world’s first Doctorate in Business Administration (DBA) in Data Science this year. This program is aimed at aligning data science to a business and for building state-of-art decision support systems or automation tools to solve business problems. The program deals with the data and digital aspects of a business. Students, undertaking this course, will design, build, and deploy systems that combine several digital technologies. It will help graduates lead teams of specialists from the fields of IoT, blockchain, cybersecurity and big data. Price: €10,000 per year (~ ₹8 Lakh) Know more about it here. Initiatives By Various IITs In 2020, different IITs have also announced various interesting courses, degrees, and programs in data science. Some of the most prominent ones are: Post-Doctoral Fellowship At IIT Madras: In August, the Robert Bosch Centre for Data Science and Artificial Intelligence (RBC DSAI) at IIT Madras called for application for the post-doctoral fellowship program into areas of Data Science, artificial intelligence, and other allied application domains. As per the notification, the selected candidates would be able to take up research on topics related to deep learning, theoretical machine learning, network analytics, reinforcement learning and multi-armed bandits, natural language processing, the system architecture for data science, and financial analytics, among others. Price: The fellowship amount is between ₹45,000-55,000. IIT Madras’ BSc In Data Science: In August, IIT Madras announced that it would be bringing a unique online degree and diploma-level programs for data science and programming. This program consists of three tiers. The first is a foundational program, followed by two diplomas in the second year. Interested students may continue to the third year, which allows them to opt for specialisation. Upon completion of the three years, candidates are awarded a BSc degree. Price: For the foundational course – ₹4,000; for diploma – ₹5,000; for BSc course – ₹10,000. Know more about it here. Also Read: Top AI Initiatives By IITs In 2020 IIT Jodhpur’s Undergraduate Program in Data Science: IIT Jodhpur launched a new BTech degree program in data science and AI for the academic year 2020-21. This program consists of courses in computer science, mathematics, AI\/ML, data science, and their related applications. Further, this program will also allow students to opt for specialisations in visual computing, language technologies, socio-digital realities, among others. They will also have the option to take up MBA(tech) in the fifth year to receive a dual degree from the School of Management and Entrepreneurship. Price: The tuition fee is up to ₹300,000. PG Certificate In Data Science From IIT Roorkee: IIT Roorkee collaborated with TSW, Executive Education arm of Times Professional Learning to launch a post-graduate certificate program in data science and machine learning. Designed specifically for working professionals, it is an 11-months program that has ten modules. For the last module, candidates are given the option to choose a specialisation in computer vision and image recognition, speech recognition or data engineering. Price: ₹300,000\/- + Taxes (There is a provision of 15% discount on course fee for the alumni of IIT-R) IIT-R & Coursera Launch Online Data Science Program: IIT Roorkee has also partnered with Coursera to introduce a certificate program in data science. It will help students looking to skill up in data science, machine learning, data collection, visualisation and management, with training on Python and SQL programming. The programs will be launched in early 2021. IIT Kanpur’s Free Online Data Science Course: IIT Kanpur announced two free online courses on data science on its SWAYAM NPTEL platform. SWAYAM is a government-initiated program for equal and free access to teaching-learning resources. A two-part course, this data science course consists of — Essentials of Data Science With R Software – 1: Probability and Statistical Inference and Essentials of Data Science With R Software – 2: Sampling Theory and Linear Regression Analysis. The programs will be launched in early 2021.","excerpt":"While normal education suffered a standstill in 2020, there were a lot of online courses and programs that were initiated by some of the most prestigious institutions as well as big tech giants so that the process of learning and skill development doesn’t suffer. As the trend has been for a few years now, some […]","categories":["AI Trends"],"tags":["Artificia Intelligence in Data Science","artificial intelligence machine learning data","beginner python projects","big data platform c++","computer science projects","Data Science","data science programs","data science training","IISc","machine data intelligence","project topics for computer science"],"author_name":"Shraddha Goled","publish_date":"2020-12-22T17:00:00","publication_year":"2020","word_count":944,"keywords":["computer vision","IISc","deep learning","Pandas","project topics for computer science","data science","artificial intelligence","machine data intelligence","analytics","data science programs","computer science projects","Data Science","machine learning","AI","ML","artificial intelligence machine learning data","big data platform c++","data science training","Artificia Intelligence in Data Science","Aim","beginner python projects"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","computer vision","data science","analytics","Aim","Pandas"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-data-science-education-initiatives-by-institutions-in-2020\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10166939,"title":"Chhattisgarh Partners with IESA to Strengthen Semiconductor Ecosystem","content":"The Chhattisgarh government has signed a memorandum of understanding (MoU) with the India Electronics and Semiconductor Association (IESA) to boost the electronics and semiconductor industry. The MoU was signed during the Chhattisgarh Investor Connect roadshow in Bengaluru. The agreement will enhance the electronics system design and manufacturing (ESDM) sector. This collaboration aims to position Chhattisgarh as a key player in homegrown semiconductors.. Rajat Kumar, secretary of Chhattisgarh’s commerce and industries department, and Ashok Chandak, president of IESA, signed the MoU that is a significant step in encouraging a business-friendly environment and positioning Chhattisgarh as a key player. Speaking about the partnership, Kumar said, “Chhattisgarh has played a crucial role in India’s infrastructure growth by meeting demands for steel, cement, minerals, and power…We must strengthen our ESDM footprint. This partnership with IESA is a significant step in the development of Chhatisghargh.” Under this partnership, the IESA will support driving investments, innovation, and ease of doing business in Chhattisgarh. The association will work closely with the state government to establish incubation centres and talent development programs in Naya Raipur and other key locations. Using IESA’s expertise, these incubation centres are expected to become self-sustainable within five to seven years. Moreover, IESA will help secure grants and formulate sustainable business frameworks. To further boost the ESDM ecosystem, IESA will help establish Common Facility Centres (CFCs) and Electronics Manufacturing Clusters (EMCs) while fostering industry-academia collaboration. The MoU also includes plans for hosting national and international ESDM events, enhancing Chhattisgarh’s visibility in the global electronics and semiconductor landscape. “We are honored to be Chhattisgarh’s knowledge partner in shaping its ESDM landscape. With our collective expertise, we will drive industry-friendly policies, skill development, and innovation, helping create a thriving technology ecosystem in the state,” V Veerappan, chairman of IESA, said, stressing the importance of industry-academia collaboration. “This partnership ensures that Chhattisgarh’s semiconductor and ESDM policies align with industry needs.” Later in the day, IESA led a semiconductor industry discussion with chief minister Vishnu Deo Sai and senior government officials. Both parties reaffirmed their commitment to strengthening Chhattisgarh by promoting talent, development, and strengthening the city. This strategic collaboration between the Chhattisgarh government and IESA is set to unlock new opportunities in the electronics and semiconductor sector. By fostering innovation, talent development, and investment, Chhattisgarh is poised to become a major player in India’s evolving ESDM landscape.","excerpt":"This strategic collaboration between the Chhattisgarh government and IESA is set to unlock new opportunities in the electronics and semiconductor sector.","categories":["AI News"],"tags":["Indian AI startups","semiconductor industry"],"author_name":"Amisha Arya","publish_date":"2025-03-31T10:55:51","publication_year":"2025","word_count":390,"keywords":["Indian AI startups","Go","API","programming_languages:R","AI","innovation","semiconductor industry","programming_languages:Go","RAG","Aim","R"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","API","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/chhattisgarh-partners-with-iesa-to-strengthen-semiconductor-ecosystem\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45740,"title":"10 Exciting Papers To Look Out For At The NeurIPS Conference","content":"The 33rd annual conference on Neural Information Processing Systems (NeurIPS) is going to be held at Vancouver Convention Center, Vancouver, Canada from December 8th to 14th, 2019. The primary focus of the Foundation is the presentation of a continuing series of professional meetings known as the Neural Information Processing Systems Conference, held over the years at various locations in the United States, Canada and Spain. NeurIPS received a record-breaking 6743 submissions this year, of which 1428 were accepted. Here are few exciting works from the accepted papers at the 33rd edition of this widely popular global conference: Quality Aware Generative Adversarial Networks By IIT Hyderabad With the publication of this paper in 2014, applications of GANs have witnessed a tremendous growth. Generative-Adversarial Networks(GANs) have been successfully used for high-fidelity natural image synthesis, improving learned image compression and data augmentation tasks. GANs have advanced to a point where they can pick up trivial expressions denoting significant human emotions. They have become the powerhouses of unsupervised machine learning. While GANs have indeed become very popular, they suffer from drawbacks such as mode collapse, convergence issues, entanglement and poor visual quality. To address these shortcomings, two researchers from IIT Hyderabad introduce Quality Aware GANs(QAGANs). In this work, the authors, Sumohana Channappayya and Parimala Kancharla of IIT Hyderabad demonstrate how a distance metric and a novel quality aware discriminator gradient penalty function can each be used as excellent regularizers for GAN objective functions. The results show performance of QAGANs to be very competitive with the state-of-the-art methods over three popular datasets. HyperGCN By IISc Bangalore A popular learning paradigm is hypergraph-based semi-supervised learning (SSL) where the goal is to assign labels to initially unlabeled vertices in a hypergraph. Motivated by the fact that a graph convolutional network (GCN) has been effective for graph-based SSL, the authors propose HyperGCN, a novel GCN for SSL on attributed hypergraphs. Additionally, the authors show how HyperGCN can be used as a learning-based approach for combinatorial optimisation on NP-hard hypergraph problems. The authors also demonstrate HyperGCN’s effectiveness through detailed experimentation on real-world hypergraphs. Stand-Alone Self-Attention in Vision Models By Google Brain Recent approaches have argued for going beyond convolutions in order to capture long-range dependencies. These efforts focus on augmenting convolutional models with content-based interactions, such as self-attention and non-local means, to achieve gains on a number of vision tasks. This work primarily focuses on content-based interactions to establish their virtue for vision tasks. A simple procedure of replacing all instances of spatial convolutions with a form of self-attention applied to ResNet model produces a fully self-attentional model that outperforms the baseline on ImageNet classification A Geometric Perspective on Optimal Representations for Reinforcement Learning By DeepMind The authors propose a new perspective on representation learning in reinforcement learning based on geometric properties of the space of value functions. This work opens up the possibility of automatically generating auxiliary tasks in deep reinforcement learning, analogous to how deep learning itself enabled a move away from hand-crafted features. ViLBERT By Georgia Inst. Of Technology & FAIR ViLBERT (short for Vision-and-Language BERT), is a model for learning task-agnostic joint representations of image content and natural language. The authors have extended the popular BERT architecture to a multi-modal two-stream model, pro- cessing both visual and textual inputs in separate streams. This work represents a shift away from learning groundings between vision and language only as part of task training and towards treating visual grounding as a pre trainable and transferable capability Neural Attribution for Semantic Bug-Localization in Student Programs By IISc Bangalore & Google This work proposes a first deep learning based technique that can localize bugs in a faulty program w.r.t. a failing test, without even running the program. Most open online courses on programming make use of automated grading systems to support programming assignments and give real-time feedback. The authors, in this technique introduce a novel tree convolutional neural network which is trained to predict whether a program passes or fails a given test. Adversarial Examples Are Not Bugs, They Are Features By MIT The researchers in this paper, demonstrate that adversarial examples can be directly attributed to the presence of non-robust features: features derived from patterns in the data distribution that are highly predictive, yet brittle and incomprehensible to humans. After capturing these features within a theoretical framework, they establish their widespread existence in standard datasets. Weight Agnostic Neural Networks By Adam Gaier & David Ha In this work, the authors question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. They propose a search method for neural network architectures that can already perform a task without any explicit weight training. The results show that, on a supervised learning domain, network architectures achieve much higher than chance accuracy on MNIST using random weights. Deep Learning Without Weight Transport By University Of Toronto The authors describe two mechanisms – a neural circuit called a weight mirror and a modification of an algorithm proposed by Kolen and Pollack in 1994 – both of which let the feedback path learn appropriate synaptic weights quickly and accurately even in large networks, without weight transport or complex wiring. Tested on the ImageNet visual-recognition task, these mechanisms outperform both feedback alignment and the newer sign-symmetry method, and nearly match backprop, the standard algorithm of deep learning, which uses weight transport. Deep Equilibrium Models By Carnegie Mellon University The researchers at CMU  present a new approach to modeling sequential data: the deep equilibrium model (DEQ). Motivated by the observation that the hidden layers of many existing deep sequence models converge towards some fixed point, they propose the DEQ approach that directly finds these equilibrium points via root-finding. Such a method is equivalent to running an infinite depth (weight-tied) feedforward network, but has the notable advantage that can analytically backpropagate through the equilibrium point using implicit differentiation. Using this approach, training and prediction in these networks require only constant memory, regardless of the effective “depth” of the network. Check all the accepted papers here. This year, NeurIPS accepted more papers from Google AI while the premier institutes like MIT, Stanford and CMU took the next spots. Source: Blog by Diego Charrez The Neural Information Processing Systems Foundation is a non-profit corporation whose purpose is to foster the exchange of research on neural information processing systems in their biological, technological, mathematical, and theoretical aspects. Neural information processing is a field which benefits from a combined view of biological, physical, mathematical, and computational sciences.","excerpt":"The 33rd annual conference on Neural Information Processing Systems (NeurIPS) is going to be held at Vancouver Convention Center, Vancouver, Canada from December 8th to 14th, 2019. The primary focus of the Foundation is the presentation of a continuing series of professional meetings known as the Neural Information Processing Systems Conference, held over the years […]","categories":["Deep Tech"],"tags":["Deep Learning","Google Brain","iisc bangalore","MIT","NeurIPS","papers"],"author_name":"Ram Sagar","publish_date":"2019-09-09T19:00:03","publication_year":"2019","word_count":1084,"keywords":["Go","data augmentation","machine learning","AI","neural network","MIT","Google Brain","ResNet","NeurIPS","BERT","iisc bangalore","deep learning","papers","GAN","Deep Learning","R"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","R","Go","BERT","GAN","ResNet","data augmentation"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/10-exciting-papers-to-look-out-for-at-the-neurips-conference\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10062600,"title":"Data science expert Mark Tenenholtz admits that 90% of models he created are inefficient","content":"Kaggle Master and senior data scientist Mark Tenenholtz sent a tweet admitting that despite spending thousands of hours on ML models, 90 per cent of the models that he used were ineffective. Tenenholtz then listed the best baseline models for different datasets in the same thread. He underlined the importance of having a good baseline in models, and it was a valuable asset to solve issues with ML models. https:\/\/twitter.com\/marktenenholtz\/status\/1501905740813848582?s=21 For tabular data, Tenenholtz said that XGBoost, LightGBM, or RF models are some of the most commonly used models. Even though ensemble tree-based models can outperform neural networks, XGBoost is the most popular choice among Kagglers. For time series data, he stated that models like XGBoost, LightGBM and RF were the best despite them not being built for time series data. Tenenholtz explained that even if the dataset is for tabular data, one could set a prediction horizon that is well-matched to the lag between the input and the output so that the user can control the system better and effectively. Then, the dataset can be treated as tabular data. For image datasets, ResNet and EffNet-BO are small and quick models that are effective for nearly any type of image data. A huge advantage of these models is that they can be scaled up and used for greater accuracy. DistilRoBERTa is the best model for text datasets. The model offers a combination of speed and accuracy. When scaled up, the accuracy of the model increases. The best models for audio datasets are ResNet and EffNet. Tenenholtz justified the usage of image models for audio datasets. He said that he had started audio problems by converting the audio to a spectrogram and combining it with an image model.","excerpt":"The best models for audio datasets are ResNet and EffNet. Tenenholtz justified the usage of image models for audio datasets.","categories":["AI News"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-03-11T18:38:12","publication_year":"2022","word_count":287,"keywords":["Go","TPU","RoBERTa","AI","neural network","ML","BERT","XGBoost","LightGBM","R"],"extracted_tech_keywords":["AI","ML","neural network","XGBoost","LightGBM","TPU","R","Go","BERT","RoBERTa"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/data-science-expert-mark-tenenholtz-admits-that-90-of-models-he-created-are-inefficient\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001650,"title":"8 Sensor-Based Projects Beginners Should Work On To Enhance Their Skills","content":"Applications and appliances in everyday lives have a lot of intricate and basic science and technology uses, that we hardly even realise on a daily basis. From the attendance ID card detector at your workplace or college to the touch screens on your smartphone, everything has some amount of sensors equipped with it, requiring a proper circuit on a breadboard. Here are some sensor-based projects that beginners can do: 1.Car parking sensor: The parking sensors that we have in cars today are programmed on a simple breadboard and a sensor. Some sensors also calculate the distances between the obstacle and the car to tell you how close you are to hitting the object and show by means of a sound. There can also be another way to indicate the obstacle. For example, the sensors measure the distance the sensor is from an object or obstacle and it will emit an LED from blue to red as the object is within range and closer to the target area. The sensor is an ultrasonic sensor which can be used to measure the distance of an object from a certain position. The sensor emits ultrasonic waves which are reflected back by the object. The time taken by the waves to travel back and forth is calculated and multiplied with a velocity of sound to get the distance measurement. Things you need: Breadboard, an ultrasonic ranger, two LEDs, buzzer and RGB LCD Display. 2.Maze solving robot: Maze solver robot involves developing the program and logic that will make the robot solve the maze by itself. It portrays one of the important ways in which we can train the machines to achieve a complicated task on their own. Things you need: Arduino UNO, motor driver, geared DC motors, IR sensor modules. 3.Temperature sensor: The digital temperature sensor interfaced to the 8051 microcontrollers is used for sensing the temperature conditions. By using four push buttons of this circuit we can adjust the temperature settings. If the temperature exceeds a preset value then the load is turned off automatically and if the temperature falls below a set value, then the load, which is the lamp, is turned on automatically. Things you need: Microcontroller, breadboard, temperature sensor, lamp. 4.Metal Detector Robotic Vehicle: A metal detector robotic vehicle is designed to detect metals. They can also be used to detect land mines. Land mines are unstable devices that are placed under the ground and it is very dangerous to detect those manually using metal detectors. Here, a metal detector is embedded in the robotic circuit and it is controlled by using radio frequency communication. At the transmitter end, to provide required motion to the robot, the number of push buttons is interfaced to the microcontroller. When any button is pressed, then the command is sent to the microcontroller which sends a binary data to the button. The encoder converts parallel data to serial data and this command is transmitted using RF module. At the receiver end, this command is decoded by the decoder and the microcontroller gives relevant signals to the motor driver to drive the motors so that the robot moves in the desired motion. A metal detector is embedded in the circuit that detects any metal and using buzzer it gives an indication. Things you need: Transmitter, receiver, microcontroller. 5.Automatic Door Opening System: This can be implemented for opening the door automatically by detecting the human body temperature. But, the simple projects can be used to operate these doors automatically using a PIR sensor as shown in the figure below. The infrared radiation emitted from a human body is detected by a passive infrared sensor or the PIR sensor of an automatic door opening system project. This PIR sensor generates a sensing signal after detecting infrared energy emitted from the human body. This signal is fed to the 8051 microcontrollers which are interfaced to the door motor through motor driver IC. Thus, if any, human body comes near to the sensing area of the PIR sensor, automatic door system, then the door gets opened and closes automatically after a fixed time delay. Things you need: Temperature sensor, PIR, sensor, microcontroller. 6.Bluetooth controlled robot using Android mobile: For this project, it required for the Android mobile user has to install an application on her\/his mobile. Then the user needs to turn on the Bluetooth in the mobile.   can use various commands like move forward, reverse, stop move left, move right. These commands are sent from the android mobile to the Bluetooth receiver. The Android-based robot has a Bluetooth receiver unit which receives the commands and gives it to the microcontroller circuit to control the motors. The microcontroller then transmits the signal to the motor driver IC’s to operate the motors. Robots can also be made so that they can be controlled by Bluetooth. They would be small in size and can also be deployed to spying and surveillance, by attaching a camera to it. With few additions and modifications, this robot can be used in the borders for detecting and disposing of hidden land mines. Things you need: Android mobile, Bluetooth receiver unit, 8051 microcontrollers, LCD Display, motor driver, DC motors. 7.Home security system: Home security system project consists of an IR transmitter and receiver module which works for the safety of doors at night or in case we are out of the home. When the IR sensors are interrupted, a buzzer is turned on indicating someone is entered into the house. It has an LPG gas sensor. It is provided to detect LPG gas leakage. A buzzer is turned on when gas is detected by the sensor. It also serves the functionality of a door-latch opening using a password entered through the keypad. A serial port is used to send the data from all modules to a computer. This project can be used to implement other related modules like fire sensor and wind sensor. Things you need: Transmitter, receiver, gas sensor, computer. 8.Moisture sensor: Most soil moisture sensors are designed to estimate soil volumetric water content based on the dielectric constant of the soil. The dielectric constant can be thought of as the soil’s ability to transmit electricity. The dielectric constant of soil increases as the water content of the soil increases. This response is due to the fact that the dielectric constant of water is much larger than the other soil components, including air. Thus, the measurement of the dielectric constant gives a predictable estimation of water content. Things you need: Breadboard, Arduino UNO, soil moisturiser sensor.","excerpt":"Applications and appliances in everyday lives have a lot of intricate and basic science and technology uses, that we hardly even realise on a daily basis. From the attendance ID card detector at your workplace or college to the touch screens on your smartphone, everything has some amount of sensors equipped with it, requiring a […]","categories":["AI Trends"],"tags":["sensor"],"author_name":"Disha Misal","publish_date":"2019-03-26T17:52:51","publication_year":"2019","word_count":1091,"keywords":["programming_languages:R","AI","Git","sensor","Ray","R"],"extracted_tech_keywords":["AI","Ray","R","Git","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-sensor-based-projects-beginners-should-work-on-to-enhance-their-skills\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":49605,"title":"Why Storytelling Is A Much-Needed Element In Pitch Decks","content":"Funding is important for startups — it allows them to innovate with all its potential and grow in this ever-competitive industry. However, it is not an easy task to gain significant funding from VCs; a company must show and prove that they are worth the amount of money. And this can be done when you have a strong pitch deck. Data vs Story It’s not always about just having a pitch deck, it is also about what to use in the pitch deck — data or story. This is a huge dilemma for startups when they seek funding. While many startups prefer to add as many numbers and details as possible on the pitch deck to show their credibility, there are companies that prefer to take the process ahead with storytelling. If you are a startup and you have worked for a certain time, you would definitely prefer to put up a lot of data about — from revenues created to capital spent on resources to acquiring customers. And this definitely is important. When you have the credibility, why not show it with the data. But despite the importance of data in a pitch deck, reports suggest that storytelling is way more effective when pitching to a VC. The major reason behind storytelling being one of the driving forces that helps you land investment is that it pulls your audience in and allow you to create an emotional connection. When you incorporate storytelling with your pitch deck, you take your potential investor to a journey and make them feel connected, which is more likely to make them invest not only on your company or product but also on your vision. How To Incorporate Storytelling In Your Pitch Deck Storytelling works only when you know how to incorporate it with your pitch deck. While there are many methods and ways to make your pitch deck stand out,  we would like to share some points that you definitely need to keep in mind when you create a pitch deck. When we talk about storytelling, the things we have to focus on are building a scenario, the problem and the solution. The Scenario When it comes to the scenario, it’s about creating a character and a complete scene that investors would like to connect with. Your potential investor wouldn’t be interested in knowing about your product right away. Rather, they would like to know about a person — with a name, with a designation, and would like to know about him\/her. Here is an example: Tinder Pitch Deck Doesn’t matter which domain you are in if you know the art of storytelling, even your data science startup raise that much-needed fund. The Problem Once you are done setting up the scenario, now is the time to talk about that problem that is bugging your character. The problem cannot come up out of the blue, your character must have discovered it — tell them about it too. One of the best ways to make more impact is to compare your character’s problem with your potential investor’s problem. Make them feel that it is actually something that needs to be taken care of. For example, if you are pitching for your data science startup, you have to do your homework and see if your potential investors have any data science aspect. If they have, is there any challenge they are facing or could face? Create tension in the mind of the listener and then jump onto the solution. The Solution When you talk about solving the problem, you have to make sure that your product gets enough attention. It is not just about the solution to the problem but more about how your product would solve the problem. Furthermore, you also have to state how your product going to scale, how it is better than the ones in the market (only if there’s any) and is there only one problem that your product deals with. That is not all, you also have to tell them how the team has worked to make the product. Tell them about each one’s role. Example Storytelling is definitely one of the core pillars but that doesn’t mean data has no role to play. If you have been operating your startup from past few years and if you have generated some certain amount of revenue, you definitely should add your stats. Some of the stats you can add: Growth rateNumber of clients or customers using your productInvestments you have madeThe revenue you have generatedHow do you understand your clients\/customer behaviour and make changes to your productYour marketing strategy etc. To give you an example of Problem and Solution and even the use data, here is the AirBnb Pitch Deck where they showcased the perfect way of stating the problem and the solution along with data. Outlook The number of startups emerging in different domains is massive. However, not every startup is successful when it’s about convincing the sharks. They either don’t have any experience about the process of pitching or they neglect the imperative points. If that’s the case, then it’s high time to understand that data and storytelling go hand in hand, and the startups who are making the best use of these two are more likely to bag some significant funding.","excerpt":"Funding is important for startups — it allows them to innovate with all its potential and grow in this ever-competitive industry. However, it is not an easy task to gain significant funding from VCs; a company must show and prove that they are worth the amount of money. And this can be done when you […]","categories":["AI Features"],"tags":["Funding"],"author_name":"Harshajit Sarmah","publish_date":"2019-11-08T12:30:24","publication_year":"2019","word_count":883,"keywords":["Go","data science","Funding","API","funding","programming_languages:R","AI","programming_languages:Go","R","startup"],"extracted_tech_keywords":["AI","data science","R","Go","API","startup","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-storytelling-is-a-much-needed-element-in-pitch-decks\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10052584,"title":"Federated Learning Using Particle Swarm Optimization","content":"Federated learning is a method that stores only learnt models on a server in order to protect data privacy. This approach does not collect data on the server but instead collects data from scattered clients directly. Due to the fact that federated learning clients frequently have limited transmission bandwidth, communication between servers and clients should be streamlined to maximize performance. As a result, researchers have created the FedPSO algorithm, which combines the particle swarm optimization technique with federated learning to boost network communication performance. We will attempt to cover certain aspects of this system and comprehend the proposed system in this post. The points to be focused on in this post are outlined below. Table of Contents Understanding the Current ScenarioFederated LearningParticle Swarm Optimization FedPSOEvaluation Let’s start the discussion by knowing how the need for PSO is identified in a federated learning system. Most of the discussions below are referred from an official research paper by Sunghwan Park and his team. Understanding the Current Scenario The use of mobile devices such as smartphones and tablets has recently increased. On mobile devices, several types of data are generated and accumulated, including data generated by users and sensors such as cameras, microphones, and the global positioning system. The gathered data on mobile devices is excellent for deep learning, which performs well when there is a large volume of data. Mobile device data can be used for machine learning (ML) in a variety of ways. Calculation time accounts for significantly more than communication time in traditional ANN models, therefore various approaches, such as leveraging graphics processing unit (GPU) accelerators and linking numerous GPUs, are utilized to reduce calculation time. In federated learning, however, communication takes longer than calculation. There have been numerous studies on client communication to increase federated learning performance. Federated learning suffers from a number of issues caused by mobile devices’ unstable network environment, such as frequent node crashes, regularly shifting node groups, high central server overhead, and increased latency as the number of nodes grows. Furthermore, multi-layer models have been employed to improve learning accuracy, although the number of weights for the nodes rises as the layers deepen. Because it increases the size of the network transmission between the server and the client, data size is a restriction for federated learning. Many research works have recently explored this issue. Particle swarm optimization is one of them. PSO has been used in machine learning and research to increase performance, such as a PSO convolutional neural network, which employs a PSO to classify images and a linearly decreasing weight PSO for CNN hyperparameter optimization. In addition, a study used PSO to determine optimum hyperparameters to increase the learning performance of federated learning clients. As a result, PSO has been consistently applied to a variety of methodologies, including updating ML model weights and tweaking hyperparameters. Federated Learning Federated learning is a learning strategy for distributed datasets that have been proposed. It uses datasets dispersed across several devices to train a model while limiting data leakage. Federated learning has the advantage of improving privacy and lowering communication costs. ANN models can learn without compromising data or personal information thanks to federated learning. Transferring data from multiple devices to a single server also increases network traffic and storage costs. By exchanging only the weights generated from training the models, federated learning greatly cuts communication costs. Source The above figure outlines the process involved in a Federated Learning system and can be described as below: The learning model is sent from the server to each client.Client data is used to train the received models.Each client delivers the server its trained model.The server aggregates all of the collected models into a single updated model.The server updates each client’s model, and processes 1 through 5 are repeated. To achieve the fourth step in Figure above, several federated learning research papers employ algorithms such as federated stochastic gradient descent (FedSGD) and federated averaging (FedAvg). Federated learning needs a mobile device environment that is distributed. Mobile devices have the drawback of learning on a wireless network rather than a solid cable network connection. When sending the trained model, if the network is unstable, the learning client’s connection may be lost, or the client may be unable to communicate the entire dataset. So to overcome this problem as stated in the introduction of this post, researchers have used PSO to aggregate all the distributed models in an optimized way. In the next part before proceeding to the proposed system we will take a brief look at PSO. Particle Swarm Optimization Kennedy and Eberhart created PSO, the most well-known metaheuristic global optimization algorithm, in 1995. The program uses techniques inspired by natural swarms of birds and fish to maximize multiple variables at once. Because of its ease of implementation, scalability, resilience, quick convergence, and simple mathematical operations, the PSO method provides advantages in memory requirements and speed. The algorithm takes a probabilistic approach to optimization, which necessitates a huge number of iterations. The swarm and particles are two types of PSO components. A swarm is a collection of particles. Each particle symbolizes one of the problem’s possible solutions. For the next phase, each particle has a position and a velocity V. Particles communicate with each other step by step to find the global optimal value and share their individual pbest (particle best) variable. The gbest (global best) variable is set to the optimal value of the shared pbest values by each particle: gbest = maxi (pbest). Each particle calculates inertia (Vt-1, preceding step’s speed), pbest, and gbest values. To have practical exposure to the PSO I recommend you to go through this article which discusses the PSO where you will learn how to use PSO in your daily ML practice. Now let’s move the proposed system i.e., FedPSO. FedPSO Deepening the layers of the model is a common way to improve the accuracy of ANN models. A deep neural network is what this is called. The number of weight factors that must be trained grows as the layers go deeper. When the model learned on the client is transferred to the server in universal federated learning, the network communication cost increases significantly. To overcome this, the FedPSO algorithm was employed, which delivers the best score (such as accuracy or loss) to the server by using PSO characteristics to send the trained model, independent of size, to the server. The suggested model, FedPSO, only receives the model weights for the client who supplied the top score, eliminating the need for all clients’ model weights to be conveyed. The figure below depicts the procedure. The lowest loss value derived after training on the client is used to calculate the best score. This loss is only 4 bytes long. FedPSO determines the best model using the pbest and gbest variables, then updates each weighted array element of the best model using the value of V. Source The weight update method in FedPSO shown in the above Figure is as follows: the server receives a client’s score and requests a learning model; the client then sends the best value to make it a global model. In the next section, we will see the evaluation result of the FedPSO system Evaluation Researchers conducted experiments to determine accuracy and convergence speed in order to evaluate FedPSO’s performance, and all parameters were observed in an unstable network environment. Below are two graphs: the first depicts the accuracy of the network as a function of convergence speed for both methods, and the second depicts the communication cost incurred by both algorithms. The accuracy benchmarks of the two methods, as well as a cost analysis of data communication between clients and servers, were all achieved using the Canadian Institute for Advanced Research (CIFAR-10) and Modified National Institute of Standards and Technology (MNIST) datasets. Source Conclusion We’ve seen how a FedPSO technique based on particle swarm optimization can increase federated learning network communication performance and minimize the bulk of data delivered from clients to servers in this post. By sharing the score value, the proposed technique aggregates the model trained on the server. The client with the highest score sends the server the trained model. As we can see in the above two graphs FedPSO succeeds in both terms of accuracy and communication cost. Reference FedPSO Original Research Paper","excerpt":"Federated learning is a method that stores only learnt models on a server in order to protect data privacy.","categories":["AI Trends"],"tags":["Cloud Platform","federated learning","Machine Learning","machine learning optimization","optimization models"],"author_name":"Vijaysinh Lendave","publish_date":"2021-10-30T10:00:00","publication_year":"2021","word_count":1387,"keywords":["federated learning","Go","machine learning","AI","neural network","ML","Machine Learning","RAG","Ray","deep learning","machine learning optimization","optimization models","R","Cloud Platform"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","Ray","federated learning","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/federated-learning-using-particle-swarm-optimization\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10115008,"title":"Rajeev Chandrasekhar on Google Gemini Fiasco: &#8216;Saying Sorry is Not Enough!&#8217;","content":"In a recent development that has stirred significant attention, Google issued an apology to the Indian government following the backlash over its AI application, Gemini. The controversy has sparked a broader discussion on the regulation and accountability of AI-based applications in India. Union Minister Rajiv Chandrasekhar expressed dissatisfaction with Google’s apology, emphasising that a mere apology does not suffice in light of major internet platforms’ legal expectations and responsibilities. Source: NDTV The government of India has made it clear that AI-based applications operating within its jurisdiction must be licensed, highlighting the need for accountability and responsibility from companies like Google. Minister Chandrasekhar pointed out that platforms with a significant internet presence cannot simply apologise after violating laws or regulations. “We will not tolerate platforms being irresponsible to the idea that the internet should be safe and trusted that our Digital Nagariks should find that what they consume is safe and trusted,” said Chandrasekhar. He said that this has been told to Google and every other platform very seriously. “Some platforms are taking the advice and laws seriously and some platforms believe, I think, very wrongly or in a sense delusionally that they are above the laws of India or the expectations of safety and trust that we are insisting,” he added. Citing Google, Chandrasekhar said this was “an irresponsible, unplanned, and hasty launch” of the platform [Gemini] that had not been fully tested yet or not fully passed the test for safety and trust that people put. “They are in some sort of a mad race not fearing any consequence of the platform blowing up and that is something that we are drawing their attention, and there is consequence,” he added, saying that companies are taking it from the lab and rushing it to the public internet without creating necessary guardrails. The controversy surrounding Google’s Gemini application has raised important questions about regulating AI technologies and balancing innovation and legal compliance. As India continues to navigate the complexities of regulating emerging technologies, the government’s stance signals a firm commitment to protecting the interests of its citizens in the digital age.","excerpt":"Chandrasekhar also criticised Gemini’s release on Indian internet without proper disclosures and disclaimers.","categories":["AI News"],"tags":["Gemini","Google","Rajeev Chandrasekhar"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-03-05T12:09:12","publication_year":"2024","word_count":350,"keywords":["Rajeev Chandrasekhar","Gemini","Go","AWS","AI","cloud_platforms:AWS","innovation","programming_languages:R","Git","llm_models:Gemini","Google","Rust","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Rust","Git","innovation","llm_models:Gemini","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-it-minister-rajeev-chandrasekhar-on-google-gemini-fiasco-saying-sorry-is-not-enough\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045658,"title":"Machine Learning For Predictive Maintenance: Key Approaches &#038; Techniques To Consider","content":"System failure is common across all use cases of machines and equipment. But for a certain context, it is crucial to predict the failure in time and, accordingly, take preventive measures to avoid such failure. Fortunately, we now have machine learning technology at our disposal to make precise predictions of system failure that can easily be prevented. By assessing the working of certain parts and equipment within a machine, maintenance work can be scheduled, and the entire system can be prevented from breaking down. Such predictive maintenance can prevent car breakdowns, aviation accidents, and failure of mission-critical systems exposing people and resources to risks. While periodic maintenance based on the gross assessment and life cycles of parts is still the most common approach to prevent sudden failure, it cannot guarantee that a system will not break down before reaching the threshold time for periodic maintenance. As most users consider uncompromising system performance as the most important value proposition of tech gadgets and devices, predictive maintenance is gaining more importance than ever before. Here, through the rest of the blog post, we will explain how machine learning (ML) technology empowers predictive maintenance and the key ML techniques used by predictive maintenance. Let’s start. Finding the Ideal Machine Learning Techniques for Predictive Maintenance Source: Maruti Techlabs Predictive maintenance by using Machine Learning tries to learn from historical data and use live data to detect the patterns of system failure. In contrast to traditional maintenance procedures relying on the life cycle of machine parts, the ML-based predictive approach prevents loss of resources and under-optimized utilization of resources for maintenance tasks. Predicting failure at the right time ML technology helps mitigate the fault lines in time while not draining resources. This ensures establishing the right balance between maintenance needs and resource utilization. Utilizing data to the optimum level Source: Intuceo For years, machine learning technology has made significant contributions in delivering the most precise predictions for different industry scenarios and contexts. Thanks to Machine Learning technology, making precise and highly reliable forecasts and utilizing those forecasts for perfectly timed maintenance tasks has become easier than ever before. Preventive maintenance in certain industries is part of the standard maintenance protocols. For example, in railways and aviation, preventive maintenance based on data-driven analytics has been in use for years. In these industries, the signals from the attached sensors are tracked to evaluate system health and accordingly provide support. In addition, such sensor data organized and analyzed through an analytics engine provides a historical analysis of the typical failure, system issues, and timing. While these data-driven insights and test results of machines help to implement the same maintenance drives for multiple systems in an environment, they also help establish a standard maintenance timeline besides helping to reduce resource utilization and cost. Moving from idea to reality Source: The Data Scientist To understand how machine learning can make value additions to the predictive maintenance protocols, it is important to understand its journey from merely a great concept to an era-defining reality for the industries. The most significant impetus for ML-based predictive maintenance is the exposure to the massive volume of digital data. From the huge volume of system data, an ML program can easily draw the most relevant insights to assess the system issues. Moreover, as many leading organizations, including top banks, have already used ML, the technology is already time-tested for its efficiency. On the other hand, just having a large volume of data is not enough. This data needs to be utilized for training the machine learning model with sharp attention on how the device and equipment interactions are changing because of the newly trained ML models. Furthermore, to make sure that the data facilitates precise learning for highly predictable output from the ML model, data needs to be cleaned and structured. Only then, the ML model can ensure optimum accuracy in guiding the system behaviour. Machine Learning Techniques for Predictive Maintenance Source: Tryolabs Over the years, some predictive maintenance techniques based on machine learning stood out as most effective. For predictive maintenance, one needs to attach sensors to the respective system, and the sensor will track and gather data concerning the system operations. This data fetched and gathered by sensors for predictive maintenance appears with time series comprising timestamp and sensor readings. Such timestamped data can help the ML-based application to predict the failure accurately with precise timing. Machine learning-based predictive maintenance is mainly created by using either of the two techniques mentioned below. Classification approach: This predictive approach makes predictions of the possibility of failure in any of the upcoming steps. Regression approach: This approach makes prediction of the time left before a system is failed. This predicted time before system failure is also called as Remaining Useful Life (RUL). While the first approach comes with a gross answer, it can achieve more precision while using very little data. On the other hand, the latter uses more data, but it also delivers more details about the impending failure. Industries widely use these ML-based predictive approaches to assess system failure and prevent them in advance. As the last piece of advice, we request you to prepare yourself with the following steps to make the most of a predictive maintenance journey. First of all, come with a proper definition and clarification of the use case. Ake uses existing data or produces a new set of data that goes well for the respective use case. When fetching data for your use case, utilize available data exploration techniques. Make an assessment of whether the data needs to incorporate failure patterns. When you can see clear evidence of failure patterns, the same can be used for developing Machine Learning models. Conclusion Let us make it clear once and for all that creating a machine learning model for predictive maintenance cannot follow a one-size-fits-all approach. The strategy to fetch data and process the data for the machine learning model will entirely depend on the maintenance tasks and specific challenges in dealing with the system. Not all system failures can be predicted through one model, and not all ML models can be created by following one strategy.","excerpt":"We will explain how machine learning technology empowers predictive maintenance and the key ML techniques used by predictive maintenance.","categories":["IT Services"],"tags":["Machine Learning","Machine Learning India","Predictive AI"],"author_name":"Dhaval Sarvaiya","publish_date":"2021-08-10T13:45:17","publication_year":"2021","word_count":1026,"keywords":["Go","machine learning","TPU","AI","data-driven","ML","Machine Learning India","Machine Learning","Git","analytics","GAN","R","Predictive AI"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","TPU","R","Go","Git","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/it-services\/machine-learning-for-predictive-maintenance-key-approaches-techniques-to-consider\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":20597,"title":"From Analytics Perspective, We Are Still Scratching The Surface, Says Ashutosh Misra Of Philips Lighting","content":"Ashutosh Misra at Cypher 2017 With more than 22 years of experience in research and IT and a PhD from IIT Kanpur, Ashutosh Misra has worked with large IT companies, handling several responsibilities in the insurance, financial services IT domain including advanced analytics, big data, application development, project and program management, architecture consulting, and more. Analytics India Magazine caught up with Misra, Director of Advanced Analytics & Big Data at Philips Lighting, who has served as an advisor to various Fortune 500 customers on data and analytics consulting projects. He shared his insights on his analytics journey, use of big data and analytics at Philips Lighting, challenges in the industry and much more. Analytics India Magazine: How has been your journey in the analytics industry and what are some of the analytic solutions that you have worked on? Ashutosh Misra: I have been in data organisation since the beginning of the field. Data is a key tool, and I started working with business analytics and slowly over a period of time I started looking into technologies and business functions attached to it. Post this I gradually moved into pure advanced analytics. The initial years were with large multinational firms, where I worked for around 18 years before moving into my current role. My initial days were all about learning and I invested my time heavily into understanding databases, data structures, Science of Cybernetics, then moving to Data Warehouses, creating reports, generating insights and list goes on. From there on I ended up working on some of the most complex projects. Then I became advisor to a lot of C-level executives on various data needs, BI, and business\/advanced analytics platforms. This is the time when I introduced a framework for business analytics. The simple framework lists out four factors necessary for a successful team to carry out analytics projects. The four tenants are: Data (quality etc), Domain Knowledge, Platform (tools etc) and mathematical and statistical modelling. Currently, we are working on developing a lot of value driven analytical solutions to increase efficiency and productivity of the organisation and some of them are very confidential in nature, around enablement services like supply chain, finance insurance, HR and IT. We have published couple of papers in IEEE proceedings to list out the use of complex algorithms and how to come up with insights. AIM: How have you seen the evolution of the industry? As you said that you have been here from the beginning, how has the growth been? AM: From analytics perspective, I would say that we are still just scratching the surface. Lack of good use cases (other than in Sales and marketing) has been the biggest show stoppers. Analytics today is in a state where it can be utilised in any domain and can provide small to large value. Lack of right skills, not only from technology perspective, but there is a scarcity of people who understand the domain and can cross pollinate with data. Lot of people with whom I interact, still think that modelling is analytics and that is not true.  In my interactions I have always said that we are at the beginning of what I would call the ‘wave’ of analytics. In the next 2-3 years probably, the wave will be at its peak, particularly with respect to analytics in a lot of areas. NLP, AI and use of ensembles of algorithms is going to drive analytics. IT will become a natural part of Robotics, Blockchain (analytics), etc. Also, by that time all the bubbles that are floating around analytics will also burst. AIM: How is Philips Lighting utilising advanced analytics and big data? AM: We are a big innovation company on data fronts—first one is the enablement front which includes supply chain, finance, sales and marketing and other functions, whereas IoT is the second, where we look at the complete connected lighting and connected functions. These are the two areas where we are doing a lot of investment in terms of advanced analytics and big data, and are also seeing good results. AIM:  Would you like to highlight a few use cases where analytics has benefitted the organisation tremendously? AM: There are two areas primarily where we have seen the benefits—supply chain and connected parts. When you are looking at supply chain, you start with demand, manufacturing, logistics, customer support and that is where we are looking at blockchain and analytics association. The first part is particularly with respect to demand sensing and forecasting. The second one is with respect to logistics. This includes network optimisation, container optimisation, palette optimisation and a whole lot of logistics where in real time the use beacons. There is a huge potential and there will be more in the future. When we talk about IoT, it is more about connected technologies—such as connected lighting. Today it is more focused on product development, but we see it moving towards improving the customer experience. AIM: What are the key changes that you have brought into the analytics domain at Philips Lighting? AM: My team was the pioneer in starting analytics application within the enablement functions. We built up the team (including the partner ecosystem) and it has now grown to substantial numbers, we have already deployed a few used cases and can observe the difference we have created so far. We are continuously looking at value driven data analytics and how the organisation can benefit from the same. From a change perspective, for the first time within the organisation business groups have stated using data and predictive analytics and can see the value it brings in. The initial questions such as “what will data do? “have now changed to “how can you help me with this business challenge?”. More collaboration and complex problems are getting into our funnel. This collaboration with business is what I consider as a big win for us in a short interval of time. AIM: Philips Lighting has introduced a ‘Social Impact Analysis’ app. Can you please tell us more about it? AM: In simple terms, when you talk about social impact analysis it is about how we make a change in the society utilising light. That is where we collect a lot of data and is again connected to the first part where we said connected lighting. Lot of information on this is available here. AIM: Is Philips Lighting also using technologies like artificial intelligence and machine learning? AM: Oh yes, absolutely! The projects that I talked about definitely use machine learning and artificial intelligence. We are at a stage where certain projects have reached maturity (deployed) whereas a couple of them are still in pilot. Some examples are usage of machine learning in forecasting demand, new product introduction and funnel prediction. AI is in pilot but will be more driven towards customer experience. AIM: What is the size of analytics team at Philips Lighting AM: In our group particularly, which is based out of Bengaluru, we have now about 50+ people. Whereas globally there are other groups working on various aspects of analytics and the number is much higher. AIM: What have been the most significant challenges that you have faced at the forefront of analytics? AM: One part that everyone looks at is data, but taking out business insight from data is a challenge. On the enablement front, understanding the confluence of domain and data is a challenge. Whereas in IoT, it’s about collection of huge data, its quality, applying analytics and being careful about the sensitive privacy attached to a customer. The biggest challenge for analytics today is the time taken to collect data and seeing the benefits is too huge. Other than that adoption becomes a big challenge, which is always there. AIM:  Do you also feel that the right kind of talent is a challenge in the industry? AM: I would say that “good talent” is a challenge. As I mentioned earlier, people are not clear on what analytics is all about. Everyone claims themselves to be a data scientist or wants to become one, but there is a lot of confusion about the definition itself. They don’t want to invest time in learning the domain and cross pollinate it with data. Our learning institutions don’t have industry data and trainers who can bring in that kind of perspective. Analytics is very specific and things can change very fast. Hence a wider outlook on identification of problems and influencing parameters is a must. As I like to say, analytics is a lot like astrology- only the best astrologer can predict with some confidence about the future and will stand true. At Philips Lighting, we as a company invest a lot in young talent. The idea is to identify them at a very younger age either through campuses or even outside, such as through discussions and hackathons. We feel that particularly in analytics, it takes 2-3 years before a person can mature and start producing results, therefore we hire them early and train them. We also do lateral hiring but that is very specific. AIM: What are the skill sets that you specifically look for while hiring them at an initial stage? AM: Self motivation and self-drive are two key skills that we look for in the talent we hire. The third one is problem solving and analytical skills. We also look for people who can work on coding languages like R and Python, and skillsets in analytics tools such as SAS. It is mandatory for them to be inclined towards the domain—such as understanding of supply chain, finance, connected technology or IoT. AIM: Would you also like to give some more details about how IoT is being used at Philips Lighting? How has it been progressing? AM: I can share this at an overall level, as certain things are very confidential in this particular area, given the fact that we face competition on a daily basis. When you look at lights, it not just one light now, it is all connected. Let’s say we have a building, and lights are an essential part of its infrastructure. Earlier you would just put an LED bulb and the job was done, but now that has changed. We now monitor these lights remotely and want to create a customer experience. The other aspect is management of power consumption by monitoring data, hence we can now measure that we have been able to reduce the whole power consumption by up to x% through our solutions. So effective management of power and energy efficiency, illumination of space with the right amount of light and enhancing the customer experience will be a few used cases. AIM: What are the major trends you see in the analytics industry? AM: The place where I see the industry moving into effectively is business-driven insights. Sales and marketing have more or less seasoned out with people heading towards more mature use cases now, which will also become more business specific. Companies will start looking at technologies like blockchain, machine learning and artificial intelligence, in particular. But it may take time. I expect that it will take at least another 1-2 years before these technologies are adopted fully. Overall, I see a bright future in analytics.","excerpt":"With more than 22 years of experience in research and IT and a PhD from IIT Kanpur, Ashutosh Misra has worked with large IT companies, handling several responsibilities in the insurance, financial services IT domain including advanced analytics, big data, application development, project and program management, architecture consulting, and more. Analytics India Magazine caught up […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-01-15T05:02:56","publication_year":"2018","word_count":1874,"keywords":["Go","machine learning","artificial intelligence","AI","R","NLP","Python","Aim","analytics","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","NLP","analytics","Aim","predictive analytics","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-perspective-still-scratching-surface-ashutosh-misra-philips-lighting\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10168035,"title":"AMD Announces First Chip With TSMC’s 2nm Node","content":"AMD on Monday announced the next generation of EPYC processors, claiming it to be the first HPC product in the industry built on TSMC’s 2-nanometer technology. Codenamed ‘Venice’, the processor is on track to be launched next year. Additionally, the company announced that it will manufacture its fifth-generation EPYC CPU products at TSMC’s new facility in Arizona. Based on the Zen 5 architecture, AMD’s fifth-generation EPYC CPUs were launched last month. “Being a lead HPC customer for TSMC’s N2 process and for TSMC Arizona Fab 21 are great examples of how we are working closely together to drive innovation and deliver the advanced technologies that will power the future of computing,” said Lisa Su, CEO of AMD. AMD’s announcement follows NVIDIA’s news from Monday, in which the company revealed plans to produce its first-ever AI supercomputers domestically. This initiative is part of NVIDIA’s commitment to manufacture $500 billion of AI infrastructure in the country over the next four years, in collaboration with TSMC, Foxconn, and Wistron, among others. Additionally, the TSMC plant in Arizona has already started to produce NVIDIA’s Blackwell chips. These announcements arrive as US President Donald Trump pushes to advance domestic chip manufacturing and reduce dependency on countries like Taiwan and China. Last month, TSMC also increased its investment in the US from $100 billion to $165 billion. This expansion will support the establishment of three new fabrication plants, two advanced packaging facilities, and a major R&D center in Arizona, which is the largest single foreign direct investment in the history of the country. Besides AMD and NVIDIA, Apple is also a leading customer of TSMC in its Arizona facility. Earlier this year, reports indicated that Apple is in the final stages of verifying its first domestically manufactured chips at TSMC’s plant in Arizona.","excerpt":"The company also revealed that the 5th generation EPYC CPUs will be manufactured at the TSMC plant in Arizona.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AMD","tsmc"],"author_name":"Supreeth Koundinya","publish_date":"2025-04-15T17:57:45","publication_year":"2025","word_count":297,"keywords":["AMD","programming_languages:R","AI","innovation","Aim","tsmc","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","Aim","R","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/amd-announces-first-chip-with-tsmcs-2nm-node\/","complexity_score":3,"technical_depth":5,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161493,"title":"How MongoDB Helped Zepto Reduce Latency by 40%","content":"Zepto, the quick commerce giant, leverages MongoDB Atlas to overcome significant scalability challenges. “Behind Zepto’s 10-minute delivery promise is a database built for speed!” said Sachin Chawla, VP for India & South Asia at MongoDB, in a recent social media post. By shifting from a monolithic architecture with relational databases to MongoDB’s NoSQL platform, Zepto reduced latency for its critical APIs by 40%. This change also boosted its ability to handle six times more traffic, improving both operational efficiency and the overall customer experience. The Challenge: Scaling Quick Commerce Founded in 2021, Zepto revolutionised the Indian grocery delivery industry with its promise of 10-minute deliveries. However, as Zepto expanded—achieving $1.5 billion in annualised sales by mid-2024—its legacy infrastructure struggled to keep pace. Performance bottlenecks, high latency, and an inability to deliver real-time analytics hindered growth and operational efficiency. A few months ago, during a presentation at MongoDB.local, Mayank Agarwal, senior architect at Zepto, explained: “We had a big monolith, and all the components were powered by PostgreSQL and a few Redis clusters. As our business grew, we faced many performance issues and restrictions on the velocity at which we wanted to operate.” Zepto’s reliance on Redis clusters for caching also became a bottleneck as the business scaled. Kshitij Singh, technical lead at Zepto, noted, “We had a pretty huge Redis setup which we wanted to get rid of, as we were not able to scale it further. MongoDB’s in-memory caching solved this problem for us.” Why MongoDB Atlas? Zepto sought to break its monolithic architecture into microservices and migrate to a NoSQL database to address its issues. After evaluating multiple databases, MongoDB stood out due to its ability to handle nested document structures and array-based queries effectively. “The queries were very performant, given the required indexes we had created, and that gave us confidence,” said Agarwal. “The biggest motivating factor was when we saw that MongoDB provides in-memory caching, which could address the huge Redis cluster that we couldn’t scale further.” MongoDB’s Atlas platform also offered features such as real-time data archival and analytical nodes, enabling Zepto to separate customer-facing workloads from internal queries. Singh highlighted: “This ensured that customer performance wasn’t compromised.” Results: Faster, Scalable, and Reliable Zepto’s migration to MongoDB Atlas resulted in improved efficiency at an overall operational level. The company witnessed a huge reduction in latency, thereby significantly improving the response times of critical APIs and enhancing the customer experience. Additionally, the infrastructure scaled to handle 6x more traffic, ensuring no performance degradation. The migration to MongoDB also led to a 14% improvement in page load times, contributing to higher conversion rates and increased sales. MongoDB’s support for analytical nodes enabled Zepto to maintain consistent performance, ensuring effective real-time analytics across the platform. The shift also streamlined Zepto’s ETL pipelines and archival processes, reducing management overhead and enabling faster deployment of new features. “Post the migration, we observed a latency reduction of up to 40% for key APIs and handled throughput spikes, such as during campaigns, by enabling data compression and other optimisations. This has made our infrastructure more robust and scalable,” Singh said.Scaling New Heights with AWS Zepto’s success story is also powered by its strong partnership with AWS. “We started Zepto about three years ago with just one dark store in Mumbai. Today, we’ve grown to over 200 dark stores and annualised sales north of INR 10,000 crore,” said Kaivalya Vohra, co-founder of Zepto. “From day one, we were clear that AWS was the way to go.” AWS’ managed services and support allowed Zepto to launch new features rapidly while optimising costs. “AWS helped us achieve a quick MVP with credits from the AWS Activate program. As we scaled, the collaboration with AWS architects enabled us to come up with solutions no one had tried before,” noted Singh. Zepto leverages AWS across the entire customer journey, from enhancing product discovery to streamlining last-mile logistics. Generative AI solutions on AWS have also enabled faster resolution times and improved customer satisfaction. “AWS has been instrumental in helping us scale quickly while optimising costs. Its enterprise support and sponsorship for POCs have been invaluable,” Singh added. The Pricing Dilemma Zepto is currently under fire for differential pricing algorithms on Android and iPhones. Several customers have expressed their frustration on social media platforms, speculating the possible reasons behind the stark difference in prices. View this post on Instagram A post shared by Digital Marketing Consultant | Social Media Trainer | Bangalore (@pc_pooja.chhabda) Some believe the pricing is influenced by demographic and user device data, which could potentially favour Android users. The company has yet to officially comment on this issue.","excerpt":"“Behind Zepto’s 10-minute delivery promise is a database built for speed.”","categories":["AI Features"],"tags":["AWS","mongodb atlas","postgreSQL","Redis"],"author_name":"Vandana Nair","publish_date":"2025-01-16T09:30:00","publication_year":"2025","word_count":772,"keywords":["mongodb atlas","AI","AWS","MongoDB","ML","postgreSQL","RAG","microservices","Ray","analytics","generative AI","Redis"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","Ray","RAG","AWS","microservices","Redis","MongoDB"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-mongodb-helped-zepto-reduce-latency-by-40\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10015809,"title":"Guide To VIBE: Video Inference for 3D Human Body Pose and Shape Estimation","content":"Pose estimation is now a greater research area. Until now developments have been made based on human body 2D keypoint annotations. Most of the solutions have been around a single image or 2D motion and significantly less on the 3D motion as it involves more challenges. The primary challenge being less ground truth training 3D annotated data. Some of these 3D motion researches that have come across are not satisfactory and suffer from many drawbacks. Also, these methods are mostly frame-based, which increase error rates. In February 2020 (later updated in April) PhD students Muhammed Kocabas, Nikos Athanasiou, and director Michael J. Black at Max Planck Institute for Intelligent Systems represented their paper to CVPR named “VIBE: Video Inference for Human Body Pose and Shape Estimation”. VIBE uses CNNs, RNNs(GRU) and GANs along with a self-attention layer to achieve its state-of-the-art results. A monocular video is analysed into video sequences. They have used both 2D keypoint annotated data and AMASS (Archive of Motion Capture as Surface Shapes) dataset of unpaired static 3D human motion containing shapes and poses. The model is tested upon 3DPW and MPI-INF-3DHP datasets and has produced new benchmark results. Shown above is a state-of-the-art video-pose-estimation approach, failing to produce accurate 3D body poses. To address these limitations, a large-scale motion-capture dataset is used to train a motion discriminator using an adversarial approach. VIBE can produce realistic and accurate pose and shape, beating previous methods on standard benchmarks. Below gif shows results achieved by VIBE. VIBE uses CNNs to extract image features. The output from the CNN is fed as input to the recurrent neural network, which processes the sequential nature of human motion. Then a temporal encoder and regressor are used to predict the body parameters for the whole input sequence. This whole part is referred to as the Generator(G) model. Now with the help of the AMASS dataset 3D, realistic human motion is achieved for adversarial training and build a motion discriminator(Dm). The motion discriminator takes in both predicted pose sequences along with pose sequences sampled from AMASS. The discriminator tries to differentiate between the fake and real motions by providing a real\/fake probability for each input sequence which helps in producing realistic motion. The output of this method is a standard SMPL body model format consisting sequence of pose and shape parameters. Code Snippet Source Code – https:\/\/github.com\/mkocabas\/VIBE The code is implemented in PyTorch and underneath is the train.py file illustration. # importing libraries import torch import pprint import random import numpy as np from torch.utils.tensorboard import SummaryWriter from lib.core.loss import VIBELoss from lib.core.trainer import Trainer from lib.core.config import parse_args from lib.utils.utils import prepare_output_dir from lib.models import VIBE, MotionDiscriminator from lib.dataset.loaders import get_data_loaders from lib.utils.utils import create_logger, get_optimizer # Dataloaders data_loaders = get_data_loaders(cfg) # compiling loss loss = VIBELoss( e_loss_weight=cfg.LOSS.KP_2D_W, e_3d_loss_weight=cfg.LOSS.KP_3D_W, e_pose_loss_weight=cfg.LOSS.POSE_W, e_shape_loss_weight=cfg.LOSS.SHAPE_W, d_motion_loss_weight=cfg.LOSS.D_MOTION_LOSS_W, ) # Initializing networks – CNN used is ResNet-50, T = 16 (after experimenting different values this gave best results) as the sequence length minibatch size 32, the temporal encoder has 2-layer GRU with a hidden size of 1024, regressor two fully-connected layers with 1024 neurons each, followed by a final layer generator = VIBE( n_layers=cfg.MODEL.TGRU.NUM_LAYERS, batch_size=cfg.TRAIN.BATCH_SIZE, seqlen=cfg.DATASET.SEQLEN, hidden_size=cfg.MODEL.TGRU.HIDDEN_SIZE, pretrained=cfg.TRAIN.PRETRAINED_REGRESSOR, add_linear=cfg.MODEL.TGRU.ADD_LINEAR, bidirectional=cfg.MODEL.TGRU.BIDIRECTIONAL, use_residual=cfg.MODEL.TGRU.RESIDUAL, ).to(cfg.DEVICE) # initializing optimizers – Adam optimizer with a learning rate of 5  10*5 and 110*4 for the G and DM respectively gen_optimizer = get_optimizer( model=generator, optim_type=cfg.TRAIN.GEN_OPTIM, lr=cfg.TRAIN.GEN_LR, weight_decay=cfg.TRAIN.GEN_WD, momentum=cfg.TRAIN.GEN_MOMENTUM,) # initializing discriminator – contains a sequence of GRUs and self-attention to amplify distinctive frames. 2 MLP layers containing 1024 neurons each with tanh activation function. motion_discriminator = MotionDiscriminator( rnn_size=cfg.TRAIN.MOT_DISCR.HIDDEN_SIZE, input_size=69, num_layers=cfg.TRAIN.MOT_DISCR.NUM_LAYERS, output_size=1, feature_pool=cfg.TRAIN.MOT_DISCR.FEATURE_POOL, attention_size=None if cfg.TRAIN.MOT_DISCR.FEATURE_POOL !='attention' else cfg.TRAIN.MOT_DISCR.ATT.SIZE, attention_layers=None if cfg.TRAIN.MOT_DISCR.FEATURE_POOL !='attention' else cfg.TRAIN.MOT_DISCR.ATT.LAYERS, attention_dropout=None if cfg.TRAIN.MOT_DISCR.FEATURE_POOL !='attention' else cfg.TRAIN.MOT_DISCR.ATT.DROPOUT ).to(cfg.DEVICE) dis_motion_optimizer = get_optimizer( model=motion_discriminator, optim_type=cfg.TRAIN.MOT_DISCR.OPTIM, lr=cfg.TRAIN.MOT_DISCR.LR, weight_decay=cfg.TRAIN.MOT_DISCR.WD, momentum=cfg.TRAIN.MOT_DISCR.MOMENTUM ) # initializing lr_schedulers motion_lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau( dis_motion_optimizer, mode='min', factor=0.1, patience=cfg.TRAIN.LR_PATIENCE, verbose=True, ) lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau( gen_optimizer, mode='min', factor=0.1, patience=cfg.TRAIN.LR_PATIENCE, verbose=True, ) # Training model – Both the generator and discriminator are trained together and thus helps in reducing loss. Trainer( data_loaders=data_loaders, generator=generator, motion_discriminator=motion_discriminator, criterion=loss, dis_motion_optimizer=dis_motion_optimizer, dis_motion_update_steps=cfg.TRAIN.MOT_DISCR.UPDATE_STEPS, gen_optimizer=gen_optimizer, start_epoch=cfg.TRAIN.START_EPOCH, end_epoch=cfg.TRAIN.END_EPOCH, device=cfg.DEVICE, writer=writer, debug=cfg.DEBUG, logdir=cfg.LOGDIR, lr_scheduler=lr_scheduler, motion_lr_scheduler=motion_lr_scheduler, resume=cfg.TRAIN.RESUME, num_iters_per_epoch=cfg.TRAIN.NUM_ITERS_PER_EPOCH, debug_freq=cfg.DEBUG_FREQ, ).fit() Benchmark results Below are the results achieved by SOTA models on 3DPW, MPI-INF-3DHP, and Human3.6M datasets. This is taken from the paper. VIBE (direct comp.) is trained on video datasets like others, while VIBE is trained with extra data from the 3DPW training set. Vibe outperforms all. Limitations – Vibe fails in heavy occlusion, fast motion, and multi-person occlusion. End Notes 3D pose estimation is necessary to understand human behaviour. Vibe has introduced many methods that are clubbed together to achieve the state-of-art results. In future releases, we can expect supervision on single-frame methods by fine-tuning the HMR features, extending experiments to optical flow, and resolve the multi-person and occlusion problem. Also, now in the era of transformers, the authors have plans to explore it and enhance more to showcase better performances. Video Demonstration – https:\/\/youtu.be\/rIr-nX63dUA Explained talk by the authors – https:\/\/twimlai.com\/thats-a-vibe-ml-for-human-pose-and-shape-estimation-with-nikos-athanasiou-muhammed-kocabas-michael-black\/ Code Demo – https:\/\/colab.research.google.com\/drive\/1dFfwxZ52MN86FA6uFNypMEdFShd2euQA","excerpt":"VIBE – Video Inference for 3D Human Body Pose and Shape Estimation. It uses CNNs, RNNs(GRU) and GANs along with a self-attention layer to achieve its state-of-the-art results.","categories":["Deep Tech"],"tags":["3d AI","CNNs","GANs","self attention models"],"author_name":"Jayita Bhattacharyya","publish_date":"2020-12-24T11:00:00","publication_year":"2020","word_count":831,"keywords":["Go","NumPy","TPU","AI","neural network","PyTorch","ML","self attention models","3d AI","Transformers","Colab","GANs","CNNs","R"],"extracted_tech_keywords":["AI","ML","neural network","PyTorch","Transformers","Colab","NumPy","TPU","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/guide-to-vibe-video-inference-for-3d-human-body-pose-and-shape-estimation\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10104816,"title":"Snezhana is Bringing Tech into Fashion and Films the Perfect Way","content":"Born in Russia, Snezhana Paderina is an art director in fashion and technology. Based out of New York, she works with CG cinematics and 3D designing as a fashion film director. With a drive to learn design from an American university, given that English was not her first language, Paderina combined it with her technical know-how to help her get through. “Imagine coming from a tech background and going to an art school and then the fashion industry in the US,” said Paderina, sitting in her office in Turkey, a Meta Quest headset sitting on the shelf behind her, in an exclusive interaction with AIM. “I had to write these long essays on art. So, I created a neural network to do it,” she said about her experience at design school. Being a big fan of Cyberpunk, she used William Gibson‘s books on Blade Runner and Neuromancer to build a bot that could write essays for her. “That’s how I first used AI.” Paderina said that she always wanted to be a programmer. “When I was eight years old, I watched The Matrix for the first time, and that is how I got into computers. I never wanted to go into fashion.” Currently, Paderina works at DeV Virtual Production as an art and cinematics director, where she uses Unreal Engine for AR, VR, and 3D filmmaking, along with fashion and game development. Her website Snezhana.nyc also highlights her love for using tech in fashion and film. Paderina completed her education in cyber security but was always inclined towards using it in some creative ways. One fine day she got a sewing machine and suddenly realised how her experience with programming and coding could actually augment the fashion industry. Then she got into fashion. “It doesn’t matter if it’s fashion or film, it is like a canvas for me,” said Paderina. In 2013, Paderina did her first fashion show, where she presented her CGI and 3D printing collection. “I realised that I needed some education in fashion,” she laughed. She then joined Parson School of Design in New York. “Using the Unreal Engine in filmmaking allowed me to combine everything that I have learnt as there are no constraints like in real-life film making.” Being an avid gamer with games such as Cyberpunk, she also aims to go into developing games in the future. ‘I think like a tech person, more than anything else‘ In 2018, when NVIDIA was announcing the launch of its RTX series GPUs, the Russian branch of the company contacted Paderina for a collaboration for the fashion industry at the Mercedes Benz Fashion Week, where she made her ‘RAY’ collection. “From a designer’s point of view, it was very challenging because I had to create a collection that would be appealing to both audiences, which are very different gamers, and fashion insiders,” she explained. RAY collection from Mercedes Benz Fashion Week Today, Paderina uses Midjourney to create storyboards for her movies. “It’s the fastest way to convey my vision to the team,” she added. She also uses ChatGPT to analyse scripts. “I put my scripts in ChatGPT, and ask it to analyse it from a viewer standpoint to find weak spots, and actually it gives very good results by suggesting weak spots in the character arc and many more things.” “I feel that people should learn how to utilise AI,” Paderina talks about the fear of AI. “I see a lot of artists who feel threatened by AI because Midjourney can draw better than us,” she laughed. “I spent 15 years drawing and now I do it in seconds. But these people feel that AI is a threat, they are kind of closing themselves to it, instead of adopting it in their work, just as they did with sewing machines to remove the tedious work.” In a wearable tech competition in Spain, Paderina with her collaborator, Nikita Replyanski, got first place for building GS[3], which is a corset with a graduated spine support system provides dynamic back support for patients suffering from medical conditions which cause joint hypermobility and chronic musculoskeletal issues that require daily spinal support. GS[3] corset “We made this corset very fashionable on purpose because we wanted to normalise medical devices as part of your outfit, instead of hiding under clothes,” Paderina explained. Embracing technology is all we need “With the team in 3D, we also use AI for motion capture and taking visual reference,” Paderina continued. “You put the models through AI and you get the motion capture data for 3D characters animation.” The team also uses AI on Unreal Engine for generating textures and meshes, which Paderina highlights is pretty groundbreaking. Furthermore, she highlighted that with the development of deepfake, it is very interesting to see the future of entire films being made just using AI with actors, scripts, sound, and even speech. “I think that is what people in Hollywood are afraid of and are giving a pushback to AI in the field.” When it comes to fashion, “AR has a huge potential because the fashion industry is dependent on consumers,” she added. “It would be so cool for fashion if you could see other people’s clothes changing. For example, you wear a t-shirt with the marker that you want, and other people see changes on your clothes when they see them through glasses.” “I expect a lot from Apple Vision Pro and other such upcoming wearables,” Paderina spoke about the future of AR\/VR in films and fashion. “There is something really good about interaction and interactivity that these gadgets can provide, as we saw with games.”","excerpt":"“I had to write these long essays on art. So, I created a neural network to do it,” said Snezhana.","categories":["AI Features"],"tags":["ai in fashion","Interviews and Discussions"],"author_name":"Mohit Pandey","publish_date":"2023-12-13T17:30:00","publication_year":"2023","word_count":936,"keywords":["Go","ChatGPT","ai in fashion","AI","neural network","GPT","Ray","Aim","ViT","llm_models:GPT","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","neural network","ChatGPT","Aim","Ray","R","Go","GPT","ViT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/snezhana-is-bringing-tech-into-fashion-and-films-the-perfect-way\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10119195,"title":"Why Moderna Partnered with OpenAI","content":"“People literally talk about how AI is going to cure diseases someday, and I think this is a very meaningful first step,” said CEO Sam Altman about OpenAI’s partnership with Moderna. The hottest AI startup has recently partnered with the pharmaceutical and biotechnology company to develop mRNA medicines. We’re excited to announce our ongoing collaboration with @OpenAI to co-innovate with a shared vision of AI’s immense potential in the future of business and healthcare. #AI and collaborations like this will be a key component to our ability to scale and maximize our impact on… pic.twitter.com\/MOylzcErDc— Moderna (@moderna_tx) April 24, 2024 As part of the deal, approximately 3,000 Moderna employees will gain access to ChatGPT Enterprise, developed on OpenAI’s GPT-4. Moderna plans to use ChatGPT Enterprise for mRNA medicine development to launch up to 15 new products in the next five years, including a vaccine for RSV and personalised cancer treatments. Interestingly, Moderna was one of the first customers of OpenAI’s ChatGPT Enterprise when it was launched last year. Prior to that, Moderna launched mChat in 2023, their own instance of ChatGPT built on top of OpenAI’s API. Till date, the employees at Moderna have created over 750 GPTs, designed to support specific tasks or processes across the business. Some of these GPTs assist in selecting optimal doses for clinical trials and drafting responses to regulatory questions. Moderna x ChatGPT Enterprise One of the solutions that Moderna has built and is actively developing and validating with ChatGPT Enterprise is a pilot program called Dose ID. It has the ability to review and analyse clinical data, integrating and visualising large datasets. Dose ID is designed to assist the clinical study team in data analysis, enhancing their clinical judgment and decision-making. Apart from Dose ID, there is another GPT called Contract Companion GPT, which helps the legal team at Moderna get a clear, readable summary of a contract. The Policy Bot GPT helps employees get quick answers about internal policies without needing to search through hundreds of documents. Moderna’s corporate brand team has also found many ways to take advantage of ChatGPT Enterprise. They have a GPT that helps prepare slides for quarterly earnings calls, and another that helps convert biotech terminology into approachable language for investor communications. Moderna is not Alone Last year, during AWS re:Invent, Pfizer announced that they are using generative AI for drug discovery. Pfizer developed a new generative AI platform Charlie, named after the pharmaceutical giant’s founder. “We are leveraging generative AI, which is estimated to deliver annual cost savings of $750 million to $1 billion in the near term—a real tangible value,” said Lidia Fonseca, chief digital and technology officer at Pfizer. She added that using AWS cloud services, Pfizer rapidly deployed Vox, an internal generative AI platform, enabling colleagues to access LLMs available in Amazon Bedrock and SageMaker. “A variety of LLMs in Bedrock means we can select the best tools for use cases in R&D, manufacturing, marketing, and more, enabling Pfizer and AWS to prototype 17 different use cases in a matter of weeks.” “AI and generative AI will help us identify new oncology targets, a process that is largely manual today. With AI, we can search and collate relevant data and scientific content from many more sources in a fraction of the time,” she added. Meanwhile, Novartis has partnered with Isomorphic Labs. “We announced a partnership with Isomorphic Labs, which is a spin-out of DeepMind from Google, to also see how we can speed up our ability to drug new potential targets for new medicines,” said Novartis CEO Vasant Narasimhan. “AI is going to impact many of our productivity efforts in drug development, like how fast can we generate new trial protocols, how fast can we work with regulators, how fast can we look at patient safety, and look at large patient data sets,” he added. Similarly, AstraZeneca partnered with the US AI biologics firm Absci to design an antibody to fight cancer. Last year, Google introduced Med-PaLM 2, an LLM fine-tuned for healthcare. Recently, the tech giant introduced MedLM for chest X-ray, which simplifies radiology workflows by assisting with the classification of chest X-rays for a variety of use cases. With pharmaceutical companies partnering with generative AI startups and firms, the future of medicines and drug discovery is set to become more efficient and cost-effective.","excerpt":"Moderna plans to use ChatGPT Enterprise for mRNA medicine development to launch up to 15 new products in the next five years.","categories":["Global Tech"],"tags":["OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-04-29T14:37:37","publication_year":"2024","word_count":722,"keywords":["Go","ChatGPT","OpenAI","AI","AWS","Git","RAG","Ray","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Ray","RAG","AWS","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-moderna-partnered-with-openai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10003607,"title":"Pre-Pandemic Facial Recognition Algorithms Falter In The Presence Of Masks","content":"“Even the best of the 89 commercial facial recognition algorithms tested had error rates between 5% and 50%.” The ongoing pandemic has established many uncomfortable norms in every corner of the world. Wearing masks is one such norm that has been embraced by many, reluctantly. Now the question is, what happens to those facial recognition systems, which were trained on faces without masks in the pre-pandemic world? According to a preliminary study by the National Institute of Standards and Technology (NIST) on 89 of the best commercial facial recognition algorithms, showed error in matching digitally applied face masks with photos of the same person without a mask. This study is being run under the Ongoing Face Recognition Vendor Test (FRVT) executed by the National Institute of Standards and Technology (NIST). What Does The Report Say Source: NIST This report by NIST documented the accuracy of algorithms when encountered with masked faces. The evaluation was carried out on algorithms provided to NIST before the COVID-19 pandemic and were developed without expectation of an experiment like the one conducted by NIST. The NIST team explored how well each of the algorithms was able to perform “one-to-one” matching, where a photo is compared with a different photo of the same person. The function is commonly used for verification such as unlocking a smartphone or checking a passport. The NIST Information Technology Laboratory (ITL) quantified the accuracy of pre-COVID face recognition algorithms on faces occluded by masks applied digitally to a large set of photos that have been used in an FRVT verification benchmark since 2018.  To this end, the team at NIST used two large datasets: unmasked application photographs from a global population of applicants for immigration benefits and digitally-masked border crossing photographs of travellers entering the United States. These photographs were collected in US governmental applications that are currently in operation. The team tested the algorithms on a set of about 6 million photos used in previous FRVT studies. “Black masks also degraded algorithm performance in comparison to surgical blue ones.” Source: NIST The research team digitally applied mask shapes as depicted above, to the original photos and tested the algorithms’ performance. To simulate the real-world setting, the researchers came up with nine mask variants of different shape, colour and nose coverage. The digital masks were black or a light blue that is approximately the same colour as a blue surgical mask. The shapes included round masks that cover the nose and mouth and a larger type as wide as the wearer’s face. These wider masks had high, medium and low variants that covered the nose to different degrees. The team then compared the results to the performance of the algorithms on unmasked faces. The results of the study can be summarised as follows: Masked images raised the failure rate of top algorithms to about 5%, while many otherwise competent algorithms failed between 20% to 50% of the time.The more of the nose a mask covers the lower the algorithm’s accuracy. The shape and colour of the mask matters. Algorithm error rates were generally lower with round masks. Black masks also degraded algorithm performance in comparison to surgical blue ones, though because of time and resource constraints, the team was not able to test the effect of colour completely. This report comes at a crucial time especially, in the US, where many government bodies have been creating hindrances to the deployment of facial recognition technology and quite rightfully so. One of the main reasons behind these regulations is the unreliability of these machine learning systems, which are alleged for biases that favour a certain community. Now, the masks pose a gigantic challenge to these computer vision systems, which are mostly functional in critical areas such as airports and on streets where masked individuals can exploit the flawed systems. Check the full report here.","excerpt":"“Even the best of the 89 commercial facial recognition algorithms tested had error rates between 5% and 50%.” The ongoing pandemic has established many uncomfortable norms in every corner of the world. Wearing masks is one such norm that has been embraced by many, reluctantly. Now the question is, what happens to those facial recognition […]","categories":["AI Features"],"tags":["AI facial recognition","best facial recognition software","mask","real time face recognition software"],"author_name":"Ram Sagar","publish_date":"2020-07-29T18:00:00","publication_year":"2020","word_count":641,"keywords":["Go","AI facial recognition","machine learning","mask","AI","best facial recognition software","programming_languages:R","programming_languages:Go","Git","computer vision","RAG","GAN","R","real time face recognition software"],"extracted_tech_keywords":["AI","machine learning","computer vision","RAG","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/facial-recognition-algorithms-fail-with-masks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10126891,"title":"Researchers Make AI-Generated Board Games Using CodeLLaMa","content":"In a recent breakthrough, researchers have used the help of large language models (LLMs) to generate new and never-before-seen board games. The researchers, from NYU, Maastricht University, Flinders University and UCLouvain, combined evolutionary computation with LLMs to create novel board games and their rules. In the paper, titled ‘GAVEL: Generating Games Via Evolution and Language Models’,  the researchers made use of evolutionary computation to generate new board games based on the Ludii game generation language. Additionally, they made use of CodeLlama-13b to generate the rules of the game. GAVEL essentially uses a readily available dataset of games that were made using the Ludii game system, generating variations of the games that don’t exist. This was done using an evolutionary computation method called MAP Elites, a population-based quality-diversity algorithm. “We explore the generation of novel games in the comparatively expansive Ludii game description language, which encodes the rules of over 1000 board games in a variety of styles and modes of play. We draw inspiration from recent advances in large language models and evolutionary computation in order to train a model that intelligently mutates and recombines games and mechanics expressed as code,” the researchers stated. The games generated using the GAVEL method can be accessed through the Ludii portal. One game that the researchers generated is called Havabu, which combines elements of two Ludii games called Havannah and Tabu Y. According to the researchers it’s “a variant of Havannah (a game in which players attempt to form loops or connected lines of pieces between sides of the board) that introduces a restriction on piece placement from another game in the Ludii dataset (Tabu Y).” While novel, the use of generative AI is not uncommon in game development. There have been cases of indie developers using GenAI to develop their own games. However, these are used for specifics aspects of game creation, like creating art or assets, dialogue, or even music for the game.","excerpt":"GAVEL essentially uses a a readily available dataset of games that were made using the Ludii game system, generating variations of the games that don’t exist.","categories":["AI News"],"tags":[],"author_name":"Donna Eva","publish_date":"2024-07-15T15:59:49","publication_year":"2024","word_count":322,"keywords":["Go","GenAI","programming_languages:R","AI","programming_languages:Go","generative AI","llm_models:Llama","R"],"extracted_tech_keywords":["AI","generative AI","GenAI","R","Go","llm_models:Llama","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/researchers-make-ai-generated-board-games-using-codellama\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":24158,"title":"How To Stand Out In A Highly-Competitive Machine Learning Jobs Ecosystem","content":"Digital natives like Google, Uber, Amazon and Netflix have disrupted the IT landscape with state-of-the-art machine learning capabilities and are always on a look-out for advanced talent in artificial intelligence and ML. Now, the next-generation startups are launching products and services into the market and are addressing industry-specific pain points. But as the ML ecosystem grows, the need for experienced experts has grown manifold. The buzz around ML has spawned a lot of online courses which serve as a good stepping stone for beginners, but the learning curve is mostly steep. Online courses from popular MOOC platforms do come in handy but enterprises and startups look for professionals who come with experience in ML technologies and have translated their work into a product or functionality. But as the tribe around ML is growing, so is the competition. So one of the key questions of our times is: How does one differentiate oneself from the crowd? Today, there is an oversupply of self-taught STEM learners, so one has to step up on the Kaggle and Github platform to differentiate. It’s not enough to have theoretical knowledge to compete with your counterparts and get hired. Most organisations look for the best-of-breed, business-minded ML experts who can apply the technology to business problems: According to Matthew J Schwartz of MJS Executive Search, a New York-based talent firm, top organisations lean towards ML candidates who have the following skills: A PhD in Machine Learning, Mathematics, CS in addition to business experience Knowledge and experience in leading edge and open source platforms ML experts should also have the ability to engage and consult with both business leaders and data scientists Researchers who have the most-cited papers, have their work published in journals, filed patents or clinched speaking engagements are preferred Have experience with cutting-edge ML technologies Deep Learning, AI-optimised hardware, decision management and biometrics Those who are working at the intersection of engineering and big data ecosystem, are exposed to data modeling and at are building end-to-end production systems. Software engineers usually follow this route — software engineer < data engineer and usually deal with descriptive statistics. So, if you are a software engineer and trying to understand the ML ecosystem and plan to make a switch, here are a few pointers culled from community forums: ML candidate should boast of a strong coding background (especially Python) and also have a strong base in statistics, linear algebra and calculus Unlike big tech companies headquartered in the US, most Indian companies do not expect candidates to have published research papers What’s more important is the Github profile and even successful ML projects executed as a side gig for startups or small companies Another effective way of getting the word about yourself would be improving research around a real world application Also, domain knowledge is preferred for certain jobs, for example, a person from the NLP background would be expected to know the conversion of grapheme to phoneme conversion methods. On the other hand, if you are a complete beginner, try to look for mentorships and projects or repositories to start working on in your spare time The Last Word Professionals from data engineering and ML space are in high demand and companies are grappling to fill these positions. Finding and attracting ML and data science talent has become a strategic imperative for every company. Also, the hiring cycle is usually long and one of the reasons is a lack of clarity on the business problem, as observed by Facebook’s director of AI research, Yann LeCun. According to LeCun, who was cited in MJS Research, most tech companies are not able to find out the business problem and how to solve it. We spoke to Senseforth CEO and co-founder Shridhar Marri who shared that ML beginners should go beyond their online learning and demonstrate the motivation with projects. “Online courses, especially from international universities, serve as a good starting point but it’s not just important to train, ML enthusiasts should also demonstrate their work with solutions. There is a lot of open source stuff out there and ML platforms that beginners can explore,” he said.","excerpt":"Digital natives like Google, Uber, Amazon and Netflix have disrupted the IT landscape with state-of-the-art machine learning capabilities and are always on a look-out for advanced talent in artificial intelligence and ML. Now, the next-generation startups are launching products and services into the market and are addressing industry-specific pain points. But as the ML ecosystem […]","categories":["IT Services"],"tags":["ML projects","Python","Yann LeCun"],"author_name":"Richa Bhatia","publish_date":"2018-05-02T06:59:28","publication_year":"2018","word_count":688,"keywords":["data science","Go","Yann LeCun","artificial intelligence","machine learning","AI","ML","Python","NLP","deep learning","ML projects","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","data science","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-to-stand-out-in-a-highly-competitive-machine-learning-jobs-ecosystem\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10068740,"title":"Social media firms will soon have a new problem in India","content":"India may soon become the first country in the world to create an appeal panel with the power to reverse the content moderation decisions of social media firms. Recently, the government republished the draft changes to the IT rules a week after withdrawing them. The Ministry of Information Technology has sought public consultation and comments from the stakeholders on the proposed draft for amendments to the IT Rules 2021. The new amendments The IT Rules seek to build an internet that is ‘open, safe, trusted, and accountable’ for all the citizens. MeitY has notified all social media platforms that they should comply with the Indian Constitution strictly. The IT Rules also provide for the creation of a robust grievance redressal mechanism to act immediately when a user reports illegal or harmful information. Social media companies like Facebook and Twitter are mandated to appoint an India based resident grievance officer as intermediaries with legal immunity from third party content on their platform. The amendments require the social media platform to address any complaint regarding content removal within 72 hours. According to MeitY, the rule will help remove problematic content before it goes viral. The grievance officer should also implement safeguards to prevent the misuse of the redressal mechanism. Users can appeal against the grievance redressal process of the intermediaries, and the appeal should be addressed within 30 days. Currently, there is no appellate mechanism provided by intermediaries, nor is there any credible self-regulatory mechanism in place. Earlier, social media companies have been accused of de-platforming users without hearing their side. Government vs social media firms Last year, Twitter received several blocking orders from MeitY, of which only two emergency blocking orders were temporarily complied with. However, the social media company restored the tweets ‘in a manner that we (Twitter) believed was consistent with Indian law’. Twitter said it would advocate for the rights of free expression. The company was also exploring options under the Indian law to ‘safeguard the health of the conversation occurring on Twitter’. We strongly believe that the Tweets should flow, the blog said. Last year, Facebook-owned WhatsApp sued the Indian government for ‘oppressive internet rules’ that required the instant messaging platform to make messages traceable to the outside parties; the platform called these rules unconstitutional. WhatsApp’s officials said making the messages traceable would severely undermine the privacy of billions of users and effectively impair its security. The company said it would break end-to-end encryption and could lead to real abuse. It looks like India wants to put the social media tech giants on a tight leash. Aruna Sundararajan, India’s secretary of telecommunications, said, “We don’t want to build walls, but at the same time, we explicitly recognise and appreciate that data is a strategic asset.” She further said India may not have been able to develop Tencent or Baidu or Alibaba because of a lack of nuance in the policies. Interestingly, the new proposed amendments to the IT Rules 2021 apply only to the big-tech platforms while exempting early-stage or growth-stage Indian companies or startups like Koo, DailyHunt, etc. Matter of concern While the decision to keep a close watch on the big tech platforms is a good step, considering the highly digital world we live in, the new amendments are beset by loopholes. According to Delhi-based Internet Freedom observes, the new rules may impose impractical obligations on social media platforms and subject them to government oversight. The IT Rules have considerable consequences for online freedom and privacy. It allows the government to censor content on social media, on-video entertainment, etc. Activists and privacy crusaders have appealed to MeitY to withdraw rules which may restrict the freedom of Indian internet users.","excerpt":"The IT Rules seek to build an internet that is ‘open, safe, trusted, and accountable’.","categories":["IT Services"],"tags":["Facebook"],"author_name":"Shraddha Goled","publish_date":"2022-06-10T16:08:24","publication_year":"2022","word_count":614,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Git","Rust","Facebook","R","programming_languages:Rust","startup"],"extracted_tech_keywords":["AI","R","Go","Rust","Git","startup","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/it-services\/social-media-firms-will-soon-have-a-new-problem-in-india\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":69197,"title":"Why Hybrid Cloud And Multi-Cloud Management Is No Longer Challenging","content":"As the name suggests, a hybrid cloud is a combination of public cloud, private cloud and public cloud in the target architecture. Due to security and control reasons, not all enterprise information can be placed on the public cloud, so most enterprises that have already embraced cloud computing will use the hybrid cloud model. Today, IT T teams are facing increasingly facing complexity. Multi-cloud and hybrid clouds are part of this new and dynamic IT landscape-and bring new risks. To solve the problem here, some enterprises use infrastructure as a code solution. At present, many enterprises have moved toward this cloud architecture, and this is also the key to maximizing profits. Of course, there are many configuration methods and the purpose behind these methods. The data shows that most enterprises that use cloud computing will apply several types of hybrid clouds, many of which are simply too complicated. Configuration management (CM) has always been a necessity in large-scale IT infrastructure. There are some CM tools from cloud vendors, such as Amazon Web Services or Microsoft Azure, or from virtualisation or private cloud software vendors like VMware. Infrastructure as code extends CM by creating virtual hosting models for applications. This virtual hosting model is spread across multiple cloud environments and data centre platforms. Although the infrastructure as code is an extension of CM, it actually became popular as an extension of DevOps. Users cannot deploy applications on servers or cloud services that have not yet been set up. Therefore, DevOps tools and scripts must include these configuration tasks. This makes DevOps scripts and tools tied to the configuration, and if a company changes from one cloud platform to another, users must change the script. Infrastructure-as-a-service provides a way to isolate the virtual world of applications from the underlying resources, including the cloud. As more hosting solutions exist, the infrastructure as code becomes more valuable. The IT team needs to define an abstract model of IT resources based on which infrastructure as code will be deployed and configured. The infrastructure as code model creates an intermediate layer for the deployment description. A user deploys an application to the abstract managed model created by the code, and the infrastructure then adapts it to any cloud currently in use, multi-cloud or in a mixed configuration environment. However, users of this kind of arrangement need to pay attention to various steps, as discussed below. Isolate Infrastructure From Code IT teams can deploy infrastructure as code to any environment where configuration scripts are defined, and enable applications to adapt to almost any public cloud service or data centre platform. Some users build software for each application, while others build standard models for each type of cloud hosting environment, such as infrastructure as a service, platform as a service, or a container platform. In general, it is best to reduce the number of abstract hosting models created, because when adding new hosting options, teams must debug each model. When tools allow, consider building models hierarchically, so that the infrastructure for deploying application components — or parts of an application — that is, the code model can be directly referenced in the model where the entire application is deployed. Support Infrastructure As Code For All Cloud Or Data Centre Platforms Once teams understand the required models, they can then support the specific cloud vendor and data centre configurations they plan to use. Almost all DevOps automation tools let users define their own configuration rules for any environment. Still, all popular public cloud, private cloud and platform solutions such as a hypervisor, container system and server operating system are provided as part of the infrastructure as a code toolset. There may also be community support, and other users contribute their configuration rules. This is important as it is easier to start development from a configuration that is already working than to build their own from scratch. Integrating Infrastructure Or Code Into DevOps The most subtle and challenging thing to accomplish in infrastructure as code solution is to handle the event flow such that software effectively integrate with other tools. In most cases, this means using DevOps tools. Application lifecycle operations management needs to select the appropriate software according to the conditions or specific events in the infrastructure. These events, generated through managed resources, serve as a signal of what to do. They usually activate an automated process, such as replacing the failed application component by hosting it elsewhere. To be more efficient, infrastructure as code must work closely with DevOps, but at the same time maintain its characteristics. If teams are not careful, they can develop ambiguous configuration and deployment practices, and gradually erode the independence of resources. In multi-cloud and hybrid cloud deployments, maintaining an agile infrastructure is critical, so this should be a specific goal. The infrastructure, the integration of code, and DevOps ensure the correct design and implementation of event-triggered processes. Integrating infrastructure as code into DevOps can also help users avoid common sense errors if they already have specific tools. If the cloud system is integrated into DevOps, it will be easier to host resources. This is because it is easier to virtualise the entire deployment process and the resource role of the infrastructure as code.","excerpt":"As the name suggests, a hybrid cloud is a combination of public cloud, private cloud and public cloud in the target architecture. Due to security and control reasons, not all enterprise information can be placed on the public cloud, so most enterprises that have already embraced cloud computing will use the hybrid cloud model.  Today, […]","categories":["AI Features"],"tags":["AWS cloud","Cloud Computing","data management providers"],"author_name":"Vishal Chawla","publish_date":"2020-07-07T18:54:45","publication_year":"2020","word_count":871,"keywords":["Go","cloud_platforms:Azure","cloud_platforms:Microsoft Azure","cloud computing","AI","R","automation","Cloud Computing","DevOps","cloud_platforms:Amazon Web Services","data management providers","AWS cloud","Azure"],"extracted_tech_keywords":["AI","cloud computing","Azure","R","Go","DevOps","automation","cloud_platforms:Azure","cloud_platforms:Microsoft Azure","cloud_platforms:Amazon Web Services"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/hybrid-cloud-and-multi-cloud-management-is-no-longer-challenging\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10001030,"title":"Remembering Russian Physicist Lev Landau &#038; His Contributions On His 111th Birthday","content":"“Product of optimism and knowledge is constant”, well-known Soviet theoretical physicist Lev Landau said when he was making fundamental contributions to understanding physics. Today, on his 111th birthday, we give a tribute and tell you why should you know about this legendary figure in physics. A Born Talent Born in Baku on this day, Landau was described by his classmates as a ‘shy boy’ in the class and had struggles in relating to his peers. Needless to say, he always was brilliant in Math and Physics in school, and learnt to differentiate when he was just 12 and integrate when he was 13. He had his first ever publication published when he was a mere 18 year old teenager. The publication was called Theory of the Spectra of Diatomic Molecules. He further went on to completing his PhD when he was only 21 year old and later earned a Rockefeller fellowship and a Soviet stipend, because of which had him the opportunity to study with the father of the atom, Nobel Laureate Niels Bohr. Elected to the USSR’s Academy of Sciences in 1946, Ladau also received the Lenin Science Prize for his monumental Course of Theoretical Physics which is a ten-volume study written with his student Evgeny Lifshitz, and is still widely used as a graduate-level physics reference text. Notable Contributions To Quantum Mechanics And Its Uses Today Landau made several contributions in the field of quantum mechanics. Here are some of the key ones: 1.Density matrix: He along with John von Neumann had discovered the density matrix in quantum mechanics, which represents the statistical state of a system in quantum mechanics. This theory is highly used in quantum mechanics in decomposition matrices, in pure and mixed states which have several applications such as light polarizability 2.Diamagnetism: Landau discovered the quantum mechanical property of diamagnetism in a free electron gas. Discovered in 1930, Landau diamagnetism is the diamagnetism of free electrons in an external magnetic field. It is used in the study of magnetism 3.Superfluidity: R=Nobel prize. It is the property of a fluid of zero viscosity. And therefore flows without loss of kinetic energy. This phenomena is used in Astrophysics, high-energy physics and quantum gravity. It also helps in the understanding of the Standard Model 4.Second-order phase transitions: Microscopic states of the system are important to be determined in order to calculate thermodynamic properties like the free energy, entropy, specific heat, and so on. While introducing the theory of “continuous” second-order phase transitions, Landau claimed that they “may also exist” along with the majority of first order phase transitions, the latter being discontinuous, displaying ‘jumps’ of their physical properties. Several advancements have since then been made in these transitions and are used in thermodynamics and solid state physics 5.Ginzburg–Landau theory of superconductivity: Landau’s second-order phase transitions theory gave rise to the phenomenological Ginzburg-Landau theory of superconductivity. It is used in the study of superconductors and also in String Theory 6:Fermi liquid: Also called as Landau Fermi liquid theory, this is a theoretical model of interacting fermions that describes the normal state of most metals at sufficiently low temperatures. The key idea is the notion of adiabaticity and the Pauli exclusion principle 7.Landau damping: Landau damping is the exponential decrease of longitudinal space charge waves in plasma, as a function of time. It is used to understand the wave-particle interaction in plasma, or a similar environment. It also has applications in theoretical physics 8.Landau pole: A concept in quantum electrodynamics pointed out by Landau and his colleagues, it is the momentum scale at which the coupling constant of a quantum field theory becomes infinite, and has applications in statistical physics 9.Two-component theory of neutrinos: Proposed by Landau, this was first theoretical idea about neutrino mass. It has experienced several advancements and is now widely used in particle physics 10.S matrix singularities: S-matrix theory was a proposal for replacing local quantum field theory as the basic principle of elementary particle physics. The S-matrix relates the infinite past to the infinite future in one step, without being decomposable into intermediate steps. Although not thoroughly proposed by Landau, he had contributed to proposing some aspects of this theory. It was further led to string theory. Interesting Facts One of the very interesting facts of this genius mind is that he used to keep a list of names of physicists that he ranked on a logarithmic scale, from 0 to 5. His ranking had Isaac Newton at 0 and Albert Einstein at 0.5 and the rank 1 was awarded to the founding fathers of quantum mechanics, Niels Bohr, Werner Heisenberg, Satyen Bose, Paul Dirac and Erwin Schrödinger, and others. He had also ranked himself at a 2.5, later promoting at 2.","excerpt":"“Product of optimism and knowledge is constant”, well-known Soviet theoretical physicist Lev Landau said when he was making fundamental contributions to understanding physics. Today, on his 111th birthday, we give a tribute and tell you why should you know about this legendary figure in physics. A Born Talent Born in Baku on this day, Landau […]","categories":["AI Highlights"],"tags":["Nobel prize","physicist"],"author_name":"Disha Misal","publish_date":"2019-01-22T21:10:02","publication_year":"2019","word_count":787,"keywords":["physicist","programming_languages:R","AI","Nobel prize","Git","BERT","Aim","llm_models:BERT","ViT","R"],"extracted_tech_keywords":["AI","Aim","R","Git","BERT","ViT","llm_models:BERT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/remembering-russian-physicist-lev-landau-his-contributions-on-his-111th-birthday\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":25263,"title":"Will Artificial Intelligence Replace Radiologists In 20 Years? The Answer Is No","content":"Computer-aided detection is a big part of radiology, so much so that every now and then, we hear about a new algorithm being trained to detect pathologies in medical images like tumours, lesions and fractures — just like practising radiologists. The hype is so great, that it has led to speculation about radiologists losing jobs because of artificial intelligence. In India, AI has already waded into commercial diagnostics with Aravind Eye Care hospitals deploying Google’s Deep Learning application to identify diabetic retinopathy in patients. It is the same deep learning technique which was leveraged by Google’s image search applications to detect dogs and cats in pictures. Atom360, a Bengaluru-based startup that we spoke to, focuses on AI-powered diagnostic tools to counter tobacco-related deaths in India. The six-member startup is working on deep learning applications for tumour segmentation and it targeted at radiologists who can use it to segment tumour faster and more efficiently. What Role Does AI Play In Radiology? With AI playing a major role in medicine and deep learning algorithms expediting analysis of medical imaging, radiologists, especially in the West, have gone into a panic mode. Part of the blame should be assigned to leading AI researchers Andrew Ng and Geoff Hinton who believe the profession is dying. According to experts from the medical community, radiology as a field lends itself well to technological development. Before the advent of deep learning, radiology saw rapid technological advances, from X-ray and CT scanners, to MRI scanners. Today, these are routine examinations in medical centres across the globe. In a similar vein, radiologists would certainly benefit from AI-powered systems that can read and interpret multiple images quickly, because the number of images has increased much faster over the last decade than the number of radiologists. Especially in countries like India where the doctor-patient ratio is reported to be 1:921, deep learning algorithms can help radiologists assess cases faster. Lily Peng, product manager at Google Brain AI research had once said, “India is one of the many places around the world where a lack of ophthalmologists means many diabetics don’t get the recommended annual screening for diabetic retinopathy.” Automated DR screening methods with high accuracy have a strong potential to assist doctors in evaluating more patients and quickly routing those who need help to a specialist. For Aravind Eyecare Hospitals, Google researchers worked closely with doctors in India and the US to create a development dataset of 128,000 images which were each evaluated by 3-7 ophthalmologists from a panel of 54 ophthalmologists. This dataset was used to train a deep neural network to detect referable diabetic retinopathy. Here’s Why AI Won’t Replace Radiologists Deep Learning Can Only Solve A Common Set Of Problems: Medical experts emphasise that deep learning algorithms can only solve a set of problems, that too, after the models have been trained on millions of images to arrive at the right results. In the medical field, there is a long trail of problems which can only be detected by radiologists. Currently, there are a few algorithms which are very specific and each algorithm is targeting only a specific problem. Challenges Abound In Building Discovery And Analytics Platform: A senior executive from the American tech consulting firm, Booz Allen Hamilton revealed that there are many challenges in developing discovery and analytic platforms — right from acquiring and ingesting data to annotating, storage, figuring out governance and policy — to types of analysis enabled via the platform. One of the biggest challenges has been cited as data annotation and facilitating discovery across the datasets in a platform. Deep Learning Will Add To Radiologist’s Skill And Effort: One of the biggest hurdles researchers face in the medical imaging field is building highly-accurate algorithms that can have a significant clinical impact. And the adoption of AI and medical image interpretation algorithms would speed up the results and add to the skill and effort of radiologists. Corporate Players Jump On AI Bandwagon The first company to make a splash in the medical AI field was IBM with its much-touted Watson for Oncology application which provided evidence-backed cancer care. The solution has already been deployed in India’s Manipal Hospitals to much success. As per the IBM report, in a double-blinded study presented at the San Antonio Breast Cancer Symposium, the doctors at Manipal Hospitals found that Watson technology was consistent with the tumour board recommendations in 90 percent of breast cancer cases. It’s not just IBM, IT leaders with dedicated healthcare arms like GE, Philips and Siemens have also started taking AI into their medical imaging software systems. Reports indicate that GE is developing a predictive analytics software with AI technology and there are also a slew of startups such as India and US-based Qure.ai, which are specifically developing deep learning algorithms to interpret radiology images. Qure.ai’s algorithms have received critical reviews at scientific conferences as well.","excerpt":"Computer-aided detection is a big part of radiology, so much so that every now and then, we hear about a new algorithm being trained to detect pathologies in medical images like tumours, lesions and fractures — just like practising radiologists. The hype is so great, that it has led to speculation about radiologists losing jobs […]","categories":["IT Services"],"tags":["AI for Radiology","GE Healthcare","medical imaging","qure.ai","radiology"],"author_name":"Richa Bhatia","publish_date":"2018-06-08T12:08:37","publication_year":"2018","word_count":812,"keywords":["radiology","artificial intelligence","AI","neural network","R","medical AI","RAG","Ray","deep learning","GE Healthcare","analytics","qure.ai","AI for Radiology","predictive analytics","medical imaging"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","analytics","Ray","RAG","predictive analytics","medical AI","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/will-artificial-intelligence-replace-radiologists-in-20-years-the-answer-is-no\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":44954,"title":"Why IoT Security Standards Are Crucial In Preventing Hackers From Stealing Your Data","content":"As more gadgets in the home and enterprise get connected with the web, the digital security of the internet of things is turning into a developing concern. With every one of those smart devices connected with edge gateways or cloud platforms, IoT won’t just produce a bigger amount of data than any other technology, it will also create the fastest data streams, constantly flowing through the sensors. IoT, IIoT and edge computing are growing at an incredibly quick pace. A report estimates that overall IoT spending will reach $745 billion in 2019 and get passed the $1 trillion mark by 2022. As everything from smart assistants to self-driving cars, and from smart meters in factories to healthcare devices, the extension of the IoT market reflects that of a booming business sector. With such a high level of device connectivity comes a potential hazard to data security. Episodes like hackers shutting down IoT gadgets, the Mirai botnet, which cut down a great part of the web on the US east coast; and targeted attacks against enterprise infrastructure, including electrical grids, dams and even nuclear offices have come to light in the past. It seems that IoT security may not just be about enterprise data security but also national security. Why IoT Devices Are More Vulnerable IoT devices interact with the web in ways conventional IT devices usually do not. The effectiveness of cybersecurity and privacy features are often different for IoT devices than conventional IT devices as they may run on different (or no) operating systems. IoT security is one of those features that is still thought to be a costly add-on or idea in retrospect in the development of both chips and frameworks associated with systems. The IoT’s security difficulties are especially overwhelming in light of the fact that they require shielding devices outside of conventional enterprise boundaries. Additionally, these endpoints are intended for lightweight information transmissions – not enterprise-grade security norms. What India Is Doing To Enhance IoT Standardisation As part of the Digital India initiative, the government of India has already planned to give IoT a push in the country. There is a ₹7,000-crore fund allocation for the development of 100 smart cities powered by IoT devices to control traffic, efficiently use water and power, and collect data using IoT sensors for healthcare and other services. According to Nasscom, the Indian IoT market is expected to reach $15 billion by 2020 and constitute 5% of the global market. Yet, IoT standardisation has not yet been fully implemented in the country. Nonetheless, the Department of Telecom has asked its technical body — Telecom Engineering Centre (TEC) to finalise standards for IoT and machine-to-machine technology as the country is embarking to introduce 5G services. Ministry of Electronics and Information Technology (MeiTY)’s draft IoT policy also aims to create committees in order to govern and take IoT-specific security tasks forward, although the creation of the committees and specific guidelines related to IoT has not come to life yet. Other than that, the NDCP (National Digital Communications Policy) under the Department of Communication has asked important stakeholders to enhance security around India’s infrastructure and digital communications which includes IoT. What Other Countries Are Doing For IoT Standardisation In the US, the National Institute of Standards and Technology (US Department of Commerce) has been working to establish IoT standards for a long time. It gives a rundown of six prescribed security characteristics that manufacturers can incorporate with IoT gadgets and that clients can search for: device identification, device configuration, data protection, logical access to interfaces, software and firmware updates, and cybersecurity event logging. In fact, the state of California in the US passed the first IoT Cybersecurity law that holds IoT gadget makers to higher security standards. Having faced an attack by Russian hackers, the US government is also in progress to initiate a Federal law too on IoT devices. In Europe, ETSI Technical Committee on Cybersecurity (TC CYBER) has also released ETSI TS 103645, a standard for cybersecurity in the Internet of Things, to build up a security baseline for web-connected devices, while also give a premise to future IoT certification plans. TS 103645 requires implementers to renounce the utilisation of all-inclusive default passwords, which have been the source of numerous security issues. It additionally requires vulnerability disclosure policy so as to allow security specialists and others to report issues. What More Needs To Be Done The IoT landscape is chaotic and risks to enterprises utilising IoT networks go from run-of-the-mill cyberthreats all the way to cyber espionage and national security threats. So, we certainly need a more specific approach when it comes to IoT standards. This is critically important for India as the country is one of the biggest targets for IoT attacks in the world, according to a recent report. The potential advantages of IoT will be accomplished if IoT devices are structured with trust, data privacy and security worked in. Devices need to be designed with the standards from the beginning, not as an afterthought. How might we address these worries without hampering innovation? It will require a more elevated amount of collaboration between countries on a worldwide scale. Policy-makers, regulators, device manufacturers, relevant industries and service providers will all have to come together in formulating a safer IoT ecosystem. While IoT innovation is a great resource to people and companies around the world, IoT manufacturers should gather their client’s trust in their devices. Owing to the device and edge-networking limitations, vendors need to layer security deeply into the IoT networks. Here, embedded security hardware is an important component in making IoT devices more secure. Research estimates total worldwide shipments of embedded hardware equipment to twofold by 2023, outperforming the 4 billion mark. There is quickening interest seen for embedded security, particularly in industrial verticals, which according to research is driving the market for advancements, for example, secure microcontrollers and trusted platform modules.","excerpt":"As more gadgets in the home and enterprise get connected with the web, the digital security of the internet of things is turning into a developing concern. With every one of those smart devices connected with edge gateways or cloud platforms, IoT won’t just produce a bigger amount of data than any other technology, it […]","categories":["AI Trends"],"tags":["AI cybersecurity threat","Cybersecurity","Data Security","IIoT","Internet of things","IoT"],"author_name":"Vishal Chawla","publish_date":"2019-08-26T09:52:08","publication_year":"2019","word_count":984,"keywords":["Go","AI","Data Security","innovation","Git","Internet of things","BERT","Aim","ViT","Rust","edge computing","AI cybersecurity threat","Cybersecurity","R","IoT","IIoT"],"extracted_tech_keywords":["AI","Aim","edge computing","R","Go","Rust","Git","BERT","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/iot-security-standards-hackers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":22603,"title":"How To Build A Career In AI Policy &#038; Strategy","content":"Not a day goes by when one doesn’t come across the term ‘ethical watchdog’. With enterprises mainstreaming AI applications and consumer AI taking off, concerns around data misuse, algorithmic bias abound. The recent GDPR guidelines that put the focus on the need for ethics and governance in AI is a laudable attempt to bring together legal and technical communities to help form better policy in the future. According to an IDC report, by 2019, 40% of digital transformation initiatives will use AI services. And by 2021, 75% of commercial enterprise apps will use AI. Developers will soon become the drivers of growth and the critical population to watch as IT organizations would hire AI engineers and data scientists to support the large number of DX initiatives that are AI dependent. Given the rise of an algorithmic economy, ethics and data protection would come into sharp focus, noted Giovanni Buttarelli, European Data Protection Supervisor. There would be a proliferation of consumer AI applications powered by sophisticated algorithms – which means besides AI, another key area where AI is gaining momentum is governance. Making technology work for in the interests of human being would become a critical component. Rise of AI Policy & Strategy in enterprises In view of the recent trends – AI will open new avenues and career paths for professionals. While AI will remain a tech-dominated field, there will be a growing need for academicians and researchers from universities and think tanks from field like economic, sociology, philosophy and a background in emerging technology policy to work with a team of interdisciplinary experts and steer the company’s AI policy and improve digital-human interface. Large enterprises like Microsoft and Google-owned DeepMind which have a strong position in the AI ecosystem are already putting the building blocks in place by setting up an Ethics & Impact team, focused on understanding the social effects and ethical challenges surrounding the emerging technology of artificial intelligence. Microsoft set up FATE: Case in point – Microsoft’s newly minted research group called FATE – that are working on collaborative research projects that address the need for transparency, accountability, and fairness in AI and ML systems. Also, the group publishes in a wide array of disciplines, including machine learning, information retrieval, systems, sociology, political science, and science and technology studies. The group addresses important ethical questions such as – the best use AI to assist users and offer enhanced insights, while avoiding exposing them to different types of discrimination in health, housing, law enforcement, and employment? In the same vein, how can AI applications balance the need for efficiency and exploration with fairness and sensitivity to users? As we move toward relying on intelligent agents in our everyday lives, how do we ensure that individuals and communities can trust these systems? DeepMind sets up Ethics & Impact team: Earlier last year in October, London-headquartered DeepMind set up an Ethics & Society research unit to build artificial intelligence applications that works for the benefit of all. To that effect, DeepMind has hired scientists and practitioners from diverse field – academia and charitable organization. The company has brought in American economist and Columbia Professor Jeffrey Sachs, Oxford AI professor Nick Bostrom, and climate change campaigner Christiana Figueres to advise the DeepMind team and support open research and investigation into the wider impacts of their work. Partnership of AI: Earlier in 2016, big tech companies like Amazon, Facebook, Google, Microsoft, Apple, and IBM joined hands to build a first-of-its-kind industry-led consortium that included well-known academicians and nonprofit researchers to help build ethical technologies and ensure the trustworthiness of AI. From developing best practices to advancing public understanding, the consortium regularly engages experts from various disciplines such as psychology, philosophy, economics, finance, sociology, public policy, and law to provide guidance on AI-related issues and its impact on society. The Berkman Klein Center and the MIT Media Lab: These two institutes are conducting evidence-based research with an aim to provide guidance to key decision-makers from the public and private sectors and deliver high impact-oriented pilot projects to bolster the use of AI for the public good. The centres, through research efforts will also build up institutional knowledge base on the ethics and governance of AI and strengthen the interface with industry and policy-makers. How to build a career in AI policy & strategy? Image Source: New Yorker cartoon The recent upheaval caused by GDPR regulations which will come into effect this year forced a lot of enterprises, big and small to review and rewrite their data governance guidelines and overhauled the systems. If there is ever a time to jumpstart your career in this field – it is now. Research groups are expanding their stakeholder community and are always on a lookout for key positions – such as Director of Research, Director of Partnerships, and Program Associate. You can access the details here. DeepMind recently announced its hiring Policy & Ethics Researcher and some of the requirements are: Excellent grasp of technology policy and its implications on society Ability to quickly assimilate complex issues and work across science areas Knowledge of and interest in sectors vital to artificial intelligence policy are desirable. You can access the job description here. Skills needed to make a career in AI Policy & Strategy The three core areas where AI researchers from the Ethics team work on are – AI policy, AI governance & AI strategy. This work, according to Oxford’s Future of Humanity Institute touches on a range of topics and areas of expertise such as international relations, international institutions and global cooperation, international law and international political. According to Miles Brundage, well-known AI policy researcher at the University of Oxford’s Future of Humanity Institute, there are four main roles in this area: direct research working within government think tanks & industries advocacy recruitment Candidates work on a range of topics – improving public opinion about AI, bridging between the short-term and long-term AI policy and case studies in comparisons with related technologies. Most of these job openings look for graduates who have a background in international relations, economics, psychology, law and have a deep interest in emergent technology and most importantly have a good grasp on complex technical and regulatory issues. Core on-the-job requirements\/skills for AI policy and strategy professionals are: Graduate degree Project management skills Ability to carry out both quantitative and qualitative research on emerging technology policy Devise forward-thinking policy position to companies and ability to work with different stakeholders Develop proposals for new academic projects Map AI trends and forecasts of how AI is progressing Lead external engagement with key stakeholders Those who wish to work for AI advocacy could find jobs in big tech companies and government think tanks where the role would entail building a broader awareness about AI through conferences, literature and advocating its growth. Job Titles Some of the job titles are: Policy & Ethics Researcher AI Policy Researcher AI Policy Practitioner Research Associate Program Associate Who are the best hires for AI policy and strategy roles? Usually professionals from the government sector, non-profit and academia can land a top slot in enterprises looking for external advisors from diverse fields to contribute ideas and ensure fairness of commercial AI applications.  Well, professionals who have a strong technical background in AI are well-suited to develop and oversee AI ethics committee and evaluate the recommendations posed by the team. Now, how can these professionals sharpen their skills and hone their credibility? According to Miles Brundage, as the economic relevance of AI increases, there will be a need for voices from different backgrounds to ensure fairness and ethics. One can one can start by doing short-term policy research, Brundage proposes and staying tuned to the ever-evolving AI landscape. One can also brush up AI and ML skills by doing online certifications and sharpening the AI speak Keep a tab coming from well-known research companies like DeepMind, Google, NVIDIA Try working closely with AI’s technical experts and gain an understanding of conceptual frameworks that would be needed to develop an AI policy framework Brush up on the political science skills, develop stronger research and policy development skills According to Brundage, graduates and undergraduates should look back on the international cyber-conflict to draw inspiration on how devising long-term AI policy He also says tools like statistical analysis and game theory can prove useful for doing AI policy analysis. People from law, economics and social sciences background can make a shift towards this career Lastly, sign up for an AI ethics course – a lot of universities such as Stanford, Cornell, Harvard, University of Edinburgh & online courses offer undergraduate and graduate level programs. Where can one expect the most job openings? From consulting firms such as Deloitte, McKinsey, Nielsen to government think tanks (Niti Aayog, AI Task Force) to academic institutions (Wadhwani Institute for AI — India’s first AI research centre), enterprises and India’s robotics startups, there is a growing need for AI strategists and policymakers who can straddle the field of law, governance, ethics and policy-making. Besides these areas, external advisors can also pursue opportunities in for-profit organizations\/government bodies that keep an eye on competing nations and the changing landscape of AI.","excerpt":"Not a day goes by when one doesn’t come across the term ‘ethical watchdog’. With enterprises mainstreaming AI applications and consumer AI taking off, concerns around data misuse, algorithmic bias abound. The recent GDPR guidelines that put the focus on the need for ethics and governance in AI is a laudable attempt to bring together […]","categories":["AI Highlights"],"tags":["AI Strategy"],"author_name":"Richa Bhatia","publish_date":"2018-03-14T05:13:46","publication_year":"2018","word_count":1531,"keywords":["Go","machine learning","artificial intelligence","AI","ML","Git","AI Strategy","Ray","Aim","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","Ray","R","Go","Rust","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/how-to-build-a-career-in-ai-policy-strategy\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10021235,"title":"Hands-On Guide to PyTorch Geometric (With Python Code)","content":"Released under MIT license, built on PyTorch, PyTorch Geometric(PyG) is a python framework for deep learning on irregular structures like graphs, point clouds and manifolds, a.k.a Geometric Deep Learning and contains much relational learning and 3D data processing methods. Graph Neural Network(GNN) is one of the widely used representations learning methods but the implementation of it is quite challenging as the throughput of GPU needs to be achieved on highly sparse and irregular data of varying sizes.  PyG overcomes this bottleneck by providing dedicated CUDA kernels for sparse data and mini-batch handlers for varying sizes. Methods implemented in PyG framework are supported by both CPU and GPU. PyTorch Geometric was submitted as a workshop paper at ICLR 2019, as FAST GRAPH REPRESENTATION LEARNING WITH PYTORCH GEOMETRIC. The framework was developed by Matthias Fey, eJan Eric Lenssn from TU Dortmund University. Overview of PyTorch Geometric In PyG, a graph is represented as G =  (X, (I, E)) where X is a node feature matrix and belongs to ℝN x F , here N is the nodes and the tuple (I, E) is the sparse adjacency tuple of E edges and I ∈ ℕ2 X E  encodes edge indices in COOrdinate (COO) format and E ∈ ℝE X D holds D-dimensional edge features. All the API’s that users can use are inspired from PyTorch framework itself, so that the usage of PyG should be familiar. Functionalities provided by PyG : Neighbourhood AggregationGlobal PoolingHierarchical PoolingMini-Batch HandlingProcessing of Datasets You can check all the algorithms supported by PyTorch Geometric here. Requirements & Installation Install all the requirements of PyTorch Geometric and then install it via PyPI. PyTorch >= 1.4.0 For checking the version of PyTorch, run the mentioned code: !python -c \"import torch; print(torch.__version__)\" Check the version of CUDA installed with PyTorch. !python -c \"import torch; print(torch.version.cuda)\" Install the dependencies : Replace TORCH with the PyTorch version and CUDA with the CUDA version which you are using. Might take some time to install. !pip install torch-scatter -f https:\/\/pytorch-geometric.com\/whl\/torch-${TORCH}+${CUDA}.html !pip install torch-sparse -f https:\/\/pytorch-geometric.com\/whl\/torch-${TORCH}+${CUDA}.html !pip install torch-cluster -f https:\/\/pytorch-geometric.com\/whl\/torch-${TORCH}+${CUDA}.html !pip install torch-spline-conv -f https:\/\/pytorch-geometric.com\/whl\/torch-${TORCH}+${CUDA}.html Install PyG: !pip install torch-geometric For installing from other sources, refer here. Basics of PyTorch Geometric First example refers to the data handling. Creating an unweighted and undirected graph with three nodes and four edges. Each node contains exactly one feature as shown below : #import the libraries import torch from torch_geometric.data import Data #making the edge #the tensor defining the source and target nodes of all edges, is not a list of index tuples edge_index = torch.tensor([[0, 1, 1, 2], [1, 0, 2, 1]], dtype=torch.long) #making nodes #Node feature matrix with shape [num_nodes, num_node_features] x = torch.tensor([[-1], [0], [1]], dtype=torch.float) data = Data(x=x, edge_index=edge_index) Above in edge_index, if you want to give indices, transpose the edge_index like this: edge_index = torch.tensor([[0, 1], [1, 0], [1, 2], [2, 1]], dtype=torch.long) And call contiguous on data constructor. Example is shown below: data = Data(x=x, edge_index=edge_index.t().contiguous()) You can check out all the utilities of data handling here. Common Benchmark Datasets PyG contains many benchmark datasets e.g., : all Planetoid datasets (Cora, Citeseer, Pubmed), all graph classification datasets from http:\/\/graphkernels.cs.tu-dortmund.de and their cleaned versions, the QM7 and QM9 dataset, and 3D mesh\/point cloud datasets such as FAUST, ModelNet10\/40 and ShapeNet. An example of loading the benchmark dataset is shown below: from torch_geometric.datasets import TUDataset dataset = TUDataset(root='\/tmp\/ENZYMES', name='ENZYMES') You can check all the functionalities of benchmark datasets in PyG here  or here. Mini-Batches PyG provides torch_geometric.data.DataLoader for merging the data objects to a mini batch. An example of it, is shown below: from torch_geometric.datasets import TUDataset from torch_geometric.data import DataLoader dataset = TUDataset(root='\/tmp\/ENZYMES', name='ENZYMES', use_node_attr=True) loader = DataLoader(dataset, batch_size=32, shuffle=True) for batch in loader: print(batch) print(batch.num_graphs) You can learn more about it from here. Data Transforms PyG provides its data transformation utility whose input is Data object and output is transformed Data object. Further, it can be concatenated via torch_geometric.transforms.Compose and are applied before saving a processed dataset on disk (pre_transform) or before accessing a graph in a dataset (transform). For example, we have taken a ShapeNet dataset. import torch_geometric.transforms as T from torch_geometric.datasets import ShapeNet dataset = ShapeNet(root='\/tmp\/ShapeNet', categories=['Airplane'], pre_transform=T.KNNGraph(k=6)) dataset[0] Learn more functionality here. Learning methods on Graphs. This section will create a graph neural network by creating a simple Graph Convolutional Network(GCN) layer. The whole experiment is based on the Cora dataset. Import the cora dataset. from torch_geometric.datasets import Planetoid dataset = Planetoid(root='\/tmp\/Cora', name='Cora') print(f'Dataset: {dataset}:') print('======================') print(f'Number of graphs: {len(dataset)}') print(f'Number of features: {dataset.num_features}') print(f'Number of classes: {dataset.num_classes}') Calculate the statistics on the dataset and visualize it. data = dataset[0] # Gather some statistics about the graph. print(f'Number of nodes: {data.num_nodes}') print(f'Number of edges: {data.num_edges}') print(f'Average node degree: {data.num_edges \/ data.num_nodes:.2f}') print(f'Number of training nodes: {data.train_mask.sum()}') print(f'Training node label rate: {int(data.train_mask.sum()) \/ data.num_nodes:.2f}') print(f'Contains isolated nodes: {data.contains_isolated_nodes()}') print(f'Contains self-loops: {data.contains_self_loops()}') print(f'Is undirected: {data.is_undirected()}') from torch_geometric.utils import to_networkx G = to_networkx(data, to_undirected=True) #helper function, check colab notebook mentioned in endnotes visualize(G, color=data.y) The output will be : Create a two-layer GCN network. import torch import torch.nn.functional as F from torch_geometric.nn import GCNConv class Net(torch.nn.Module): def __init__(self): super(Net, self).__init__() self.conv1 = GCNConv(dataset.num_node_features, 16) self.conv2 = GCNConv(16, dataset.num_classes) def forward(self, data): x, edge_index = data.x, data.edge_index x = self.conv1(x, edge_index) x = F.relu(x) x = F.dropout(x, training=self.training) x = self.conv2(x, edge_index) return F.log_softmax(x, dim=1) The following network contains two GCNConv layers which are used in forward pass of the model.Here, we chose to use ReLU as our intermediate non-linearity between and finally output a softmax distribution over the number of classes. Let’s train this model on the train nodes for 200 epochs. device = torch.device('cuda' if torch.cuda.is_available() else 'cpu') model = Net().to(device) data = dataset[0].to(device) optimizer = torch.optim.Adam(model.parameters(), lr=0.01, weight_decay=5e-4) model.train() for epoch in range(200): optimizer.zero_grad() out = model(data) loss = F.nll_loss(out[data.train_mask], data.y[data.train_mask]) loss.backward() optimizer.step() Evaluate the model on test data. model.eval() _, pred = model(data).max(dim=1) correct = int(pred[data.test_mask].eq(data.y[data.test_mask]).sum().item()) acc = correct \/ int(data.test_mask.sum()) print('Accuracy: {:.4f}'.format(acc)) You can learn more about the creation of Graph Neural Network in PyTorch Geometric here. You can check other examples here : Introduction: Hands-on Graph Neural NetworksNode Classification with Graph Neural NetworksGraph Classification with Graph Neural NetworksScaling Graph Neural NetworksPoint Cloud Classification with Graph Neural NetworksExplaining GNN Model Predictions using Captum Conclusion This post discussed PyTorch Geometric for fast representation learning on graphs, point clouds, and manifolds. This framework is built upon PyTorch and easy to use. It consists of various methods for Geometric Deep learning. It provides an easy-to-use mini-batch loader, multi GPU-support, benchmark datasets, and data transforms for arbitrary graphs and points clouds. Colab Notebook PyTorch Geometric Demo Official Codes, Documentation & Tutorials are available as : GithubResearch PaperTutorialsDocumentation","excerpt":"Released under MIT license, built on PyTorch, PyTorch Geometric(PyG) is a python framework for deep learning on irregular structures like graphs, point clouds and manifolds, a.k.a Geometric Deep Learning and contains much relational learning and 3D data processing methods. Graph Neural Network(GNN) is one of the widely used representations learning methods but the implementation of […]","categories":["AI Trends"],"tags":["GNN","graph convulationsl networks","graph neural networks","Python","python visualize neural network","Pytorch"],"author_name":"Aishwarya Verma","publish_date":"2021-03-04T11:00:00","publication_year":"2021","word_count":1114,"keywords":["CUDA","Pytorch","TPU","GNN","AI","neural network","PyTorch","ML","graph convulationsl networks","RAG","Python","Colab","deep learning","python visualize neural network","graph neural networks"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","PyTorch","Colab","RAG","TPU","CUDA","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/hands-on-guide-to-pytorch-geometric-with-python-code\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10051162,"title":"Top Data Classification Project Ideas and Technique","content":"Classification is the challenge in machine learning that involves detecting whether an object belongs to a certain category based on a previously trained model. As an aspiring data scientist, the most effective approach to improve the skills would be to practise. Personal projects are critical to your career development and help get one step closer to realising your data science ambitions. The knowledge, abilities, and confidence will improve as a result of projects. Including projects in the resume will make it much easier to land a data science job. However, selecting data science project ideas is a difficult task. This article contains a list of data science projects using classification to help professionals practise and improve their data science skills. The classification serves a variety of purposes. To gain a practical understanding of it, one must work in real-time. Therefore, let us proceed to the classification projects that will enable us to get real-world experience. 1. Fake News Detection According to a study conducted by MIT (Massachusetts Institute of Technology), fake news spreads six times quicker than legitimate news. Nowadays, with social media consuming so much of our lives, it is critical to discern fake news from legitimate news. With the advent of social media platforms, fake news is spreading at a breakneck pace. This research aims to develop a machine learning model that can distinguish between true and fraudulent news using a text classification approach. The complete description of the project, including the source code, can be seen here. Sample fake news detection dataset. 2. Gender classification Gender classification is garnering increasing attention, as it offers detailed information about men’s and women’s social activities. Gender classification is concerned with determining an individual’s gender based on the features that distinguish masculinity from femininity. For instance, a computer system with gender identification capabilities has a wide range of applications in basic and applied research domains such as human-computer interaction, security and surveillance, demography research, business development, mobile applications, and video games. Advances in the science of gender classification have resulted in a plethora of potential applications. Sample gender classification dataset and the project code can be found here. 3. Fake currency detection Detecting counterfeit currency is a significant issue for both individuals and corporations. Counterfeiters are continually developing new ways and techniques for manufacturing counterfeit banknotes that are virtually indistinguishable from genuine currency – at the very least in terms of the human eye. Fake currency detection is a machine learning challenge that requires binary categorisation. If we have sufficient data on genuine and counterfeit banknotes, we can use it to train a model that can categorise new banknotes as genuine or counterfeit. The project code can be found here. 4. Language classification Language classification is the process of classifying related languages. Diachronically, languages are classified into language families. In other words, languages are classified based on their development and evolution across time, with languages descended from a common ancestor being classified as a single language family. Sample language classification dataset and project code can be found here. 5. Customer churn prediction Customer churn is a term that refers to the process of identifying all potential customers or clients who will discontinue their relationship with the business. It is critical for any organisation because it is used to forecast the organisation’s growth as well as future customer trends. This project aims to categorise customers based on their likelihood of remaining with the company or to terminate their relationship. Sample customer churn prediction dataset and project code can be found here. 6. MNIST dataset image classification MNIST is an acronym for Modified National Institute of Standards and Technology. It is a collection of over 60,000 training photos and 10,000 testing images of handwritten digits. This task aims to classify the image of a handwritten numeral ranging from 0 to 9. This dataset is perfect for those who are just getting started with image categorisation. This dataset is frequently referred to as the ‘hello world’ of machine learning and deep learning in terms of object recognition. Sample MNIST dataset and project code can be found here. 7. Skin cancer classification Skin cancer is one of the most prevalent types of cancer worldwide. Even when people have skin cancer symptoms, many do not seek medical attention, which is a bad indicator because skin cancer can be cured in its early stages. This is where a machine learning algorithm comes into play when it comes to skin cancer classification. The algorithm for machine learning is based on Convolutional Neural Networks (CNN). Sample skin cancer classification dataset and project code can be found here. 8. Heart disease prediction Predicting and diagnosing cardiac diseases is the most difficult task in the medical business, as it is dependent on aspects such as the physical examination, the patient’s symptoms, and signals. Body cholesterol levels, smoking habits and obesity, family history of illnesses, blood pressure, and job environment – all contribute to heart diseases. Machine learning algorithms are critical for the accurate prediction of cardiac diseases. Hence, the prediction of heart diseases uses machine learning and the logistic regression technique. Sample heart disease prediction dataset and project code can be found here. The list consists of various machine learning classification projects. Only via practice and interaction with machine learning tools and algorithms can one gain real-world exposure to machine learning. With the proper tools and skills, no data science endeavour is too challenging. Projects are an excellent method to hone your abilities and advance toward mastery. One can practise machine learning or data science algorithms using a variety of machine learning datasets.","excerpt":"Machine Learning classification projects that will allow you to gain practical experience with the technology.","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Classification","Data Science"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-10-11T13:00:00","publication_year":"2021","word_count":932,"keywords":["data science","Classification","text classification","machine learning","Go","AI","neural network","Git","Aim","deep learning","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","data science","Aim","text classification","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/projects-using-classification-technique\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":33004,"title":"‘Cows Can Turn Food Into 24-Carat Gold’ &#038; 14 Other Gems From Indian Science Congress","content":"The Indian Science Congress has for several years been in the news for all the wrong reasons. From accusations of being “saffronised”, to propagating mythology-based stories as scientific facts, the ISC has been gathering a lot of criticism, especially from the scientific community. In keeping with the same “tradition”, the 106th Indian Science Congress also made headlines for making startling claims and statements made by well-known speakers. In this article, we list down the most bizarre mind-boggling statements made at the ISC this year: 1. Kauravas Were Born With Stem Cell  And Test Tube Technologies “We had hundreds of Kauravas from one mother because of stem cell research and test tube baby technology. Mahabharat says, 100 eggs were fertilised and put into 100 earthen pots. Are they not test-tube babies?” said The VC of Andhra University, G Nageshwar Rao, in the recently held 106th Science Congress. 2. Ravana Had 24 Types Of Aircraft — Pushpak Viman The Infamous Rao Sir had also claimed that the Ramayana states that Ravana didn’t just have the Pushpak Viman but had 24 types of varied aircraft of varying sizes and capacities. He also had several airports in Lanka and also he used these aircraft for different military purposes. 3. Lord Vishnu’s Avatar  Predates The Theory Of Evolution Given By Charles Darwin Mr Rao, the Vice-Chancellor of Andhra University said that Darwin stated life started from water and the first avatar of Lord Vishnu was also a fish a.ka. Matysa. Therefore, the Theory of Evolution was known to India much before Charles Darwin proved it. Mr Rao said that as Darwin stated life started from water the first avatar of Lord Vishnu was also a fish. For the second avatar, he took shape of a tortoise, the third avatar was a boar’s head and a human body, the fourth was the Narasimha avatar with the head of a lion and human body. He took on the human form of Vaman in the fifth avatar. According to Rao, Dashavtar, the known ten avatars of Lord Vishnu, predates the Theory of Evolution given by Charles Darwin. Here are the other strange statements made at the ISC over the years: 4. Cows Can Turn Their Food Into 24-Carat Gold A speaker at the Indian Science Congress had earlier also claimed that the cow carries a bacterium in its body which enables it to turn whatever it consumes into pure 24-carat gold. This was aberrated agenda for the cow that continues to be mystified as right-wing groups campaign for laws that ban cow slaughter across the nation. 5. Gravitational Waves Should Be Renamed As ‘Narendra Modi Waves’ Kannan Jegathala Krishnan, the senior research scientist at the World Community Service Centre at Aliyar, Tamil Nadu, claimed that modern Physics will be completely destroyed, and a new understanding of Physics will emerge based on his observations. Modern Physics will be destroyed, and gravitational waves will be renamed as ‘Narendra Modi waves’. 6. Higgs Boson Is Not The God Particle After years of research at the Large Hadron Collider, scientists confirmed the detection of Higgs boson, the elementary particle that gives mass to all other particles in the universe. But Krishnan maintained that based on his observations Higgs boson is no longer the God particle. According to him, the international scientific community will soon declare that Higgs boson is not the God particle. 7. Science Of Guided Missiles Was Present In India Thousands Of Years Ago G Nageshwar Rao had also claimed earlier that Lord Rama used astras and shastras, while Lord Vishnu sent a Sudarshan Chakra to chase his enemies\/targets. After hitting them they would come back to him, the same way it went. So, according to him, this proves that the science of guided missiles is not new to India and it was present thousands of years ago. 8. Pythagoras’s Theorem Was Invented In India, Not Greece In the very same Congress, Health Minister Dr Harsh Vardhan had also claimed earlier that India had given the world algebra and the Pythagoras’s Theorem, which is used to calculate the length of the hypotenuse of a right-angle triangle. “In the Sulbha Sutra written in 800 BCE, Baudhayan wrote the geometric formula now famously known as Pythagoras theorem. It was written by Baudhayan 300 years before Pythagoras,” said the head of Sanskrit department at Mumbai University, Dr Gauri Mahulikar, to a leading daily to back his claims. 9. Blowing A Conch Shell Exercises Your Rectal Muscles In 103rd Indian Science Congress, an IAS officer Rajeev Sharma had claimed that by blowing a conch shell for two whole minutes showcases the fact that the common Indian ritual at the beginning of any auspicious occasion has many positive effects on the body. Speaking at the ISC, Sharma had said that blowing a conch shell provides excellent exercise for rectal muscles, prostate, urinary tract, lower abdomen, diaphragm, chest and neck muscles. 10. Inter-Planetary Planes In Vedic Age In another instance, Captain Anand J Bodas had claimed that there were 200-foot planes that could fly forwards, backwards and sideways and even hover in mid-air during the Vedic age. Bodas claimed that the planes were invented by a sage called Maharishi Bharadwaj over 7000 years ago, had up to 30 engines and were equipped for warfare. 11. Helmet From Mahabharata Days On Mars One exhibitor at the ISC, Kiran Naik had also claimed that during the Mahabharata war, there was a chase in one of those Vedic planes from the earth to the moon and then to Mars, where a king attacked his rival which broke his helmet. “If you don’t believe me, will you believe the National Aeronautics and Space Association (NASA) of the USA? If you search on Google for ‘helmet on Mars’, it will tell you that even NASA has found this helmet on Mars and it will give you evidence for it,” said Kiran Naik to a leading daily. 12. Plastic Surgery  Was Done To Affix Lord Ganesha’s Elephant Head Naik also made another claim once where he according to him heated sugar was used in the plastic surgery to affix Lord Ganesha’s elephant head to his human torso. The claim about Ganesha’s head being a plastic surgery miracle was earlier made by our very own PM Narendra Modi while inaugurating a hospital in Mumbai last year. 13. Autopsy By Leaving The Dead Body In Water According to leading daily, one of the guest speakers in the ISC had claimed that autopsies were conducted in ancient days by leaving the dead body to float in water for nearly three days. When the muscles and nerves of the dead body got swelled, it was then dissected with surgical instruments that carried the names of the animals and birds they belonged to in their previous lives. 14. The first human flight took off from Chowpatty A  research paper presented at the 102nd ISC stated that the first human flight took off in Chowpatty (Mumbai) eight years before the Wright brothers flew an aircraft did. “We must look at ancient sciences with academic and scientific vigour. If we want to unlock the ancient theories, then we must read the original texts and not the translations that carry an anti-Indian viewpoint,” said Uma Vaidya, vice-chancellor, KK Sanskrit University to a leading daily. 15. Cow Urine Can Help You Locate Underground Water Another research paper presented in 102nd Indian Science Congress on ‘Engineering applications of ancient Indian botany’ claimed that herbal paste made of seeds and roots mixed with cow’s urine when applied to a person’s feet could easily locate underground water sources. “Also, cow dung, jaggery, coconut water, egg whites and green algae were used as natural polymers,” said Professor Ashok Nene from the KK Sanskrit University.","excerpt":"The Indian Science Congress has for several years been in the news for all the wrong reasons. From accusations of being “saffronised”, to propagating mythology-based stories as scientific facts, the ISC has been gathering a lot of criticism, especially from the scientific community. In keeping with the same “tradition”, the 106th Indian Science Congress also […]","categories":["AI News"],"tags":[],"author_name":"Martin F.R.","publish_date":"2019-01-08T05:41:44","publication_year":"2019","word_count":1289,"keywords":["Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","RAG","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","R","Go","GAN","ViT","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cows-can-turn-food-into-gold-14-other-isc\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10136614,"title":"‘Define Your Job by the Outcomes, Not Just the Skill Sets’ in the Era of Generative AI","content":"Speaking at Cypher 2024, India’s Biggest AI Conference by AIM Media House, Shashank Dubey, co-founder and chief revenue officer at Tredence explained that the key to thriving in the Generative AI era lies in redefining how we perceive job roles. He stressed the importance of focusing on the outcomes delivered by skill sets rather than the skills themselves. “It’s important not to define our jobs by the skills we possess. Instead, we should define our roles by the outcomes those skills deliver, and those outcomes are business outcomes,” he said. For instance, he compared two types of engineers, one whose sole expertise was in Hadoop, and another who ensured enterprise data was delivered effectively to meet business needs. Apache Hadoop is an open-source software framework designed for distributed storage and processing of large datasets across clusters of computers using simple programming models. Back in 2013, 2014, 2015, and 2016, moving data into Hadoop served as an example of modernization. “The latter engineer will have far more opportunities because they focus on the outcome,” Dubey explained. “As leaders, we must encourage our teams to embrace this mindset, prioritising results over rigid skill definitions.” Embrace Change, Fear Not About AI Dubey shared a personal anecdote from a recent Uber ride he took from LaGuardia Airport to Manhattan. His driver, Mike, a 75-year-old retired engineer, initiated an insightful discussion about adapting to change and being resilient. Mike, who had started his career sorting physical mail in Manhattan’s corporate offices, had navigated several industry transformations throughout his life. “At 17, I sorted physical mail, but as emails took over, I knew I had to pivot,” Mike shared. “I became an executive assistant and later transitioned into logistics as e-commerce emerged. Now, I drive for Uber and even dabble in learning Python.” Generative AI Is Like Formula One To illustrate the complexities of generative AI, Dubey drew a parallel to driving a Formula One car. “Just like a Formula One car requires special driving skills, specific traffic regulations, and a specialized racing track, generative AI is also a full ecosystem and platform. This is not just a skill set game,” said Dubey. Dubey also highlighted a staggering disparity in investments within the AI sector. “In the past year, there has been about $200 billion invested in acquiring chips, but less than $5 billion has gone into developing the software and services needed to utilise those chips effectively,” he said. According to him, this gap presents a significant opportunity for organizations willing to invest in the right solutions. The Roadmap to Effective Leadership Dubey outlined four essential pillars for leadership in the GenAI era: Tech Literacy, Accountability, Virtual Orchestration, and Responsible AI. Tech Literacy: According to Dubey, leaders must familiarise themselves with generative AI applications. He shared a successful initiative at a U.S. telecom company that involved business leaders developing their own ideas and collaborating with analysts to explore feasibility. “This approach teaches leaders the art of what is possible,” he noted. Accountability: As organisations embrace AI, Dubey said that accountability becomes paramount. Leaders must ensure that their teams not only understand AI but are also responsible for its ethical deployment. Citing Tesla as an example, he noted that whenever inadvertent accidents or incidents occurred during Tesla’s development, the company took accountability for those events, regardless of who was at fault. Virtual Orchestration: Dubey emphasised the importance of creating an environment that fosters collaboration between humans and machines. This involves designing workflows that use AI to augment human decision-making rather than replace it. Responsible AI: With great power comes great responsibility. Leaders must advocate for ethical AI practices that prioritise transparency, fairness, and accountability. The Future is Collaborative In closing, Dubey reiterated that the future of work in the Generative AI era is not about human versus machine, it is about human-machine collaboration. As leaders, it is essential to cultivate an environment where team members are empowered to adapt and innovate in response to technological advancements. “Human-machine collaboration is very important. The biggest destruction of value occurs not because the models are poor or the sites are lacking, but because human change management is missing,” he concluded.","excerpt":"“Just like a Formula One car requires special driving skills, specific traffic regulations, and a specialised racing track, generative AI is also a full ecosystem.”","categories":["AI Features"],"tags":["AI jobs in India"],"author_name":"Siddharth Jindal","publish_date":"2024-09-25T14:42:32","publication_year":"2024","word_count":691,"keywords":["Go","AI jobs in India","GenAI","AI","RAG","Python","responsible AI","Aim","generative AI","GAN","R"],"extracted_tech_keywords":["AI","generative AI","GenAI","Aim","RAG","Python","R","Go","GAN","responsible AI"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/define-your-job-by-the-outcomes-not-just-the-skill-sets-in-the-era-of-generative-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10000913,"title":"This Survey Reveals Concerns About Container Security Are Real","content":"More than half of companies that relied on containers to accelerate their software development and deployment has encountered security issues last year, with a majority of them deploying containers without assessing its health, reveals a recent study by Tripwire. The study, conducted in partnership with Dimensional Research in November 2018, surveyed 311 IT security professionals who manage environments with containers at companies with more than 100 employees. The survey showed that out of the 269 participants who relied on containers in production, 47% confirmed they deployed containers with vulnerabilities, while 46% admitted they deployed containers without knowing their security aspects. This lack of proper check and awareness has led to more than 60% of the organisation suffering container security-related incidents in 2018, it said. “It’s concerning, but not surprising, that nearly half of the respondents said they knowingly deploy vulnerable containers. With the increased growth and adoption of containers, organisations are under increasing pressure to speed up their deployment. To keep up with the demand, teams are accepting risks by not securing containers. Based on the findings, the majority of organisations are experiencing container security incidents,” said Tim Erlin, vice president of product management and strategy at Tripwire. With container adoption rate reaching 80% in 2018, it is estimated that organisations will confront more risk and complexity as they scale up their DevOps in 2019. Confirming this fact, 71% of the responded believed that container security-related incidents will be common in 2019. The study also pointed out how 94% of the respondent acknowledged their fear regarding container security.  Lack of proper knowledge regarding container security within a team, limited visibility into the security status of containers and container images, as well as the inability to assess risk in container images prior to deployment ranked the highest remained their biggest concerns. Scaling down on DevOps With security escalating security vulnerabilities, the only options available are for organisations to scale down DevOps deployment. Understanding the security concern revolving around containers, 42% of the respondent said they are reducing container adoption while 82% are security responsibilities because of container adoption. However, 98% were of the opinion that they want additional security capabilities for container environment. Vulnerability management, regular monitoring and auditing of containers, security testing a validation process are some of the safety controls suggested by Tripwire as a precautionary measure.","excerpt":"More than half of companies that relied on containers to accelerate their software development and deployment has encountered security issues last year, with a majority of them deploying containers without assessing its health, reveals a recent study by Tripwire. The study, conducted in partnership with Dimensional Research in November 2018, surveyed 311 IT security professionals […]","categories":["AI Features"],"tags":["DevOps"],"author_name":"Akshaya Asokan","publish_date":"2019-01-16T14:16:28","publication_year":"2019","word_count":388,"keywords":["programming_languages:R","AI","Scala","programming_languages:Scala","GAN","DevOps","R"],"extracted_tech_keywords":["AI","R","Scala","DevOps","GAN","programming_languages:R","programming_languages:Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/this-survey-reveals-concerns-about-container-security-are-real\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049945,"title":"Addressing The Vanishing Gradient Problem: A Guide For Beginners","content":"Despite having many important applications, artificial neural networks often face a number of problems. One such problem is the Vanishing Gradient Problem. When the neural networks are trained with gradient-based learning methods and backpropagation, they encounter the vanishing gradient problem. In this problem, at the time of training, the gradient starts getting smaller in size which prevents the neural networks from getting trained by not letting the network weights be changed. In this article, we will try to understand the vanishing gradient problem in detail along with the approach to resolve this problem. The major points to be covered in this article are listed below. Table of Contents What is a Vanishing Gradient Problem?How to identify the Vanishing Gradient Problem?Recognizing The Vanishing GradientsHow to Resolve the Vanishing Gradient Problem? Let us begin with understanding the vanishing gradient problem. What is a Vanishing Gradient Problem? As we know, using more layers in any neural network causes more activation function with them and as we increase the number or activation function increases the gradient of loss function leads to zero. Let’s take a simple example of any neural network where we are using any layer in a layered neural network with hyperbolic tangent function, and have gradients in the range between 0 to 1. This function will be multiplying n of these(0-1) small numbers to compute gradients of the preceding layers, meaning that the gradient decreases exponentially with n. So the function gives the output between the range of 0 to 1. This means that the output from the tanh activation function does not depend on the size of input data. The image represents a hyperbolic tangent activation function. Image source So while using the function we can say that a large change in the input space will be very small in the output. The vanishing gradients problem is one example of the unstable behaviour of a multilayer neural network. Networks are unable to backpropagate the gradient information to the input layers of the model. In a multi-layer network, gradients for deeper layers are calculated as products of many gradients (of activation functions). When those gradients are small or zero, they will easily vanish. (On the other hand, when they’re bigger than 1, it will possibly explode.) So it becomes very hard to calculate and update. The VGP occurs when the elements of the gradient (the partial derivatives with respect to the parameters of the NN) become exponentially small so that the update of the parameters with the gradient becomes almost insignificant Recognizing The Vanishing Gradients We can detect it by analysing the kernel weight distribution. There is a vanishing gradient if the weights are falling regularly near zero. This problem can be recognised when a neural network is very slow in training. Neural networks are not well trained with the data which we are using or showing unusual behaviour regarding results. How to Resolve the Vanishing Gradient Problem? There are various methods that help in overcoming the vanishing gradient problems: Multi-level hierarchyThe long short term memoryResidual neural networkReLU Let us understand these approaches one by one. Multi-Level Hierarchy It is one of the most basic and older solutions for a multilayer neural network model facing the vanishing gradient problem. It is simply a method that follows the procedure of training one level at a time and fine-tuning the level by backpropagation. So that every layer learns a compressed observation which goes ahead for the next level. Long Short-Term Memory(LSTM) So as of now, we have seen there are two major factors that affect the gradient size – weights and their derivatives of the activation function. A simple LSTM helps the gradient size to remain constant. The activation function we use in the LSTM often works as an identity function which is a derivative of 1. So in gradient backpropagation, the size of the gradient does not vanish. Let’s understand the image below. Image source According to the above image, the effective weight of the gradient is equal to the forget gate activation. So, if the forget gate is on (activation close to 1.0), then the gradient does not vanish. That is why LSTM is one of the best options to deal with long-range dependencies. More powerful in the recurrent neural network. Residual Neural Network The residual neural networks were not introduced to solve the vanishing gradient problem but they have special connections which makes it different from the other neural networks that are a residual connection, residual connection in the neural network make the model learn well and the batch normalization feature makes sure that the gradients will not be vanishing. These batch normalization features are obtained by the skip connection. The skip or bypass connection is useful in any network to bypass the data from a few layers. Basically, it allows information to skip the layers. Using these connections, information can be transferred from layer n to layer n+t. Here to perform this thing we need to connect the activation function of layer n to the activation function of n+t. This causes the gradient to pass between the layers without any modification in size. As we have discussed, activation functions keep multiplying the finite small numbers to the weights. For example, ½*½=¼ and then ½*¼=⅛ and so on. Here in the example, we can say the number of layers increases the chances of VGP. The skip of layers will help the weight of information pass from layers without vanishing. Therefore, skip connections can mitigate the VGP, and so they can be used to train deeper NNs. The above image represents the architecture of the ResNet. ResNet stands for residual network. Rectified Linear Unit (ReLU) Activation Function ReLU is an activation function more deeply it is a linear activation function. Which is like a sigmoid and tanh activation function but better than them. The basic function for ReLU conversion of input can be represented as f(x) = max(0,x) Where the ReLU function is its derivatives are constant. If in input the function gets a negative it returns 0 or if the input is greater than 0 it returns a similar value back. That is why we can say the output from the ReLU has ranged between 0 to infinity.\\ The above image represents the output of the ReLU function. Now let’s see how it helps in vanishing gradient problems. When we talk about the backpropagation procedure, whichever gradient gets updated by multiplying with the multiple factors. As the information goes towards the start of the network the more factors are multiplied together to update the gradient. Many of these factors can be considered as the activation function. The activation function derivatives can be considered as a kind of tuning parameter, designed to get the accurate gradient descent. In the above we have seen that if we multiply a bunch of numbers with a value less than 1, they will start to tend to zero hence the gradient we get from the output layer will be negligible. In this scenario, if we multiply a number with a greater value than 1 they will tend towards infinity. So where the values are less than one we will get a slope that will be less than one and here comes the vanishing gradient problem. But if somehow we get the contribution of these derivatives of the activation function as 1 we can resolve the gradient vanishing problem of the model. Basically, in this situation, we can say that every gradient update is contributing to the model from input to the output or model. Here for this ReLU comes in the picture which has only two gradients 0 or 1. Gradient one when the output of the function is > 0. Gradient zero when the output of the function is < 0. Hence these bunch of derivatives give either 0 or either 1 when multiplying together. The backpropagation equation will have only two options of either being 1 or being 0. The update is either nothing or takes contributions entirely from the other weights and biases. Final Words This article is aimed to discuss the issue that we can face while training the neural network in following the backpropagation procedure. We have seen how the problem occurs when the weights recurring are very less and tend towards zero. This kind of issue often occurs with the network with many numbers of layers, it barely occurs when the network is shallow. So if the long time taking problem is there but the network has low layers we should check for the computer configuration and if the 25% of your kernel weights are falling down to zero it should not be considered as the vanishing gradient problem.","excerpt":"when the elements of the gradient become exponentially small so that the update of the parameters with the gradient becomes almost insignificant","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","ANN","Data Science","Deep Learning","Exploding Gradient Problem","gradient boosting","gradient descent","Guide","lstm recurrent neural network","Machine Learning","Neural Networks","Python","resnet","stochastic gradient descent","vanishing gradient"],"author_name":"Yugesh Verma","publish_date":"2021-09-29T15:00:00","publication_year":"2021","word_count":1450,"keywords":["TPU","Exploding Gradient Problem","ANN","ResNet","gradient boosting","LSTM","R","resnet","Data Science","Guide","Neural Networks","Go","AI","neural network","Machine Learning","stochastic gradient descent","vanishing gradient","programming_languages:R","programming_languages:Go","Python","Aim","lstm recurrent neural network","Deep Learning","AI (Artificial Intelligence)","gradient descent"],"extracted_tech_keywords":["AI","neural network","Aim","TPU","R","Go","LSTM","ResNet","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/addressing-the-vanishing-gradient-problem-a-guide-for-beginners\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":35933,"title":"Why NLP-Powered sciSpacy Is A Game-Changer For Biomedical Text Processing","content":"A human genome contains genetic information of an organism as DNA sequences in the form of 23 chromosomes. And a single DNA molecule consists of two strands which are connected by four different bases (A, T, C, G). The human genome consists of around 3 billion of these base pairs. So, if a base pair is considered as 2-bit combination then considering all the base pairs, a diploid cell would contain 1.5 GB of data. And humans contain around 100 trillion cells. The numbers are astounding. Tasking a biomedical researcher with handling data which is not only inherently large but also comes with a multitude of combinations and classifications. Add to this, there are frequent discoveries of drugs and proteins by academia. All this information is stored in the form of tonnes of text. Skimming through this text for discoveries and deductions takes a lifetime. Though computers have made it easy to find information like a specific genome name but only in a naive way as the user has to possess the information prior to the search. The researchers at Allen Institute of Artificial Intelligence came up with a new tool or a library by the name sciSpacy, developed specifically for biomedical or scientific text processing. Most of the tools available today, deal with entity linking, abbreviation and negation detection. For traditional NLP tasks, there is GENIA. But these tools do not implement state-of-the-art word representations and neural networks. Making A Room For Biomedical Applications With sciSpacy In a paper titled scispaCy: Fast and Robust Models for Biomedical Natural Language Processing, the researchers introduce a specialised NLP library for processing biomedical texts, built on the spaCy library. To emphasise the efficiency and practical utility of the end-to-end pipeline provided by scispaCy packages, a speed comparison is performed in comparison with several other publicly available processing pipelines for biomedical text using 10k randomly selected PubMed abstracts. For training, the researchers used GENIA 1.0 corpus. This dataset has parts of speech tags annotated, which was used to train the parts of speech tagger jointly with the dependency parser. The researchers have also included the PubMed metadata for the abstracts which was discarded in the GENIA corpus. The original metadata includes relevant named entities of chemical and drugs associated to a variety of ontologies along with citation statistics and journal metadata. For named entity recognition (NER) models, the training was done on the following datasets: BC5CDR – for chemicals and diseases CRAFT – for cell types, chemicals, proteins, genes JNLPBA – for cell lines, cell types, DNAs, RNAs, proteins and BioNLP13CG – for cancer genetics Along with the datasets mentioned above, the researchers have also covered five more datasets such as Linnaeus and AnatEM for a variety of entity types which include cancer genetics, pathway analysis, trial population extraction etc. Another key challenge with biomedical data is with its commonly occurring abbreviated names and noun compounds containing punctuation, which might lead to misidentification. So, for evaluating sentence segmentation, both sentence and full-abstract accuracy were used. Read more about the sciSpacy here Installation pip install scispacy A Python code for carrying out entity recognition using ‘scispacy’: import scispacy import spacy nlp = spacy.load(“en_core_sci_sm”) text = “”” Myeloid derived suppressor cells (MDSC) are immature myeloid cells with immunosuppressive activity. They accumulate in tumor-bearing mice and humans with different types of cancer, including hepatocellular carcinoma (HCC). “”” doc = nlp(text) print(list(doc.sents)) >>> [“Myeloid derived suppressor cells (MDSC) are immature myeloid cells with immunosuppressive activity.”, “They accumulate in tumor-bearing mice and humans with different types of cancer, including hepatocellular carcinoma (HCC).”] print(doc.ents) >>> (Myeloid derived suppressor cells, MDSC, immature, myeloid cells, immunosuppressive activity, accumulate, tumor–bearing mice, humans, cancer, hepatocellular carcinoma, HCC) Key Takeaways Sets a benchmark for named entity recognition models for more specific entity extraction applications and when compared to others. sciSpacy demonstrates a competitive performance by releasing and evaluating two fast and convenient pipelines for biomedical text, which include tokenisation, part of speech tagging, dependency parsing and named entity recognition.","excerpt":"A human genome contains genetic information of an organism as DNA sequences in the form of 23 chromosomes. And a single DNA molecule consists of two strands which are connected by four different bases (A, T, C, G). The human genome consists of around 3 billion of these base pairs. So, if a base pair […]","categories":["Deep Tech"],"tags":["biomedical AI","NLP","Python"],"author_name":"Ram Sagar","publish_date":"2019-03-07T08:37:44","publication_year":"2019","word_count":665,"keywords":["artificial intelligence","AI","neural network","ML","biomedical AI","NLP","Python","GAN","ViT","spaCy","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","neural network","NLP","spaCy","Python","R","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/why-nlp-powered-scispacy-is-a-game-changer-for-biomedical-text-processing\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":40631,"title":"Guide To Cracking The MachineHack ‘Pre-Owned Cars Price Prediction’ Hackathon","content":"MachineHack one of the leading hackathon platforms dedicated to the Data Science community, is back again with an exciting hackathon for all data science enthusiasts. This new hackathon, in partnership with Imarticus Learning, challenges the data science community to predict the resale value of a car from various features. Predicting The Costs Of Used Cars Hackathon consists of data collected from various sources across India. In this article, we will continue from where we stopped, to preprocess and build a simple regression model for the hackathon. So without further ado let’s begin with a basic solution. Data Preprocessing By the end of the first part, we had already performed Exploratory Data Analysis and also cleaned the data to some extent making it ready for the next stage which is Data Preprocessing. We now have a clean dataset that we believe consists of only the values or numbers that are required to train a model and make some predictions. However, that data is still not ready to be trained. The data still consists of empty cells or nans that needs to be filled and also we need to encode and scale the data. We will also split the training set to a training set and a validation set so that we can evaluate the model for prediction accuracy. Encoding Categorical Variables We will start by encoding the categorical features in the cleaned dataset.To  encode we must know all the unique values or categories in each of the columns(‘Brand’, ‘Model’, ‘Location’,’Fuel_Type’, ‘Transmission’, ‘Owner_Type’). Follow the below steps. Finding all unique categories #'Brand', 'Model', 'Location','Fuel_Type', 'Transmission', 'Owner_Type' all_brands = list(set(list(training_set.Brand) + list(test_set.Brand))) all_models = list(set(list(training_set.Model) + list(test_set.Model))) all_locations = list(set(list(training_set.Location) + list(test_set.Location))) all_fuel_types = list(set(list(training_set.Fuel_Type) + list(test_set.Fuel_Type))) all_transmissions = list(set(list(training_set.Transmission) + list(test_set.Transmission))) all_owner_types = list(set(list(training_set.Owner_Type) + list(test_set.Owner_Type))) Initializing label encoders and fitting the categories #Initializing label encoders from sklearn.preprocessing import LabelEncoder le_brands = LabelEncoder() le_models = LabelEncoder() le_locations = LabelEncoder() le_fuel_types = LabelEncoder() le_transmissions = LabelEncoder() le_owner_types = LabelEncoder() #Fitting the categories le_brands.fit(all_brands) le_models.fit(all_models) le_locations.fit(all_locations) le_fuel_types.fit(all_fuel_types) le_transmissions.fit(all_transmissions) le_owner_types.fit(all_owner_types) Transforming the data in training set and test_set #Applying encoding to Training_set data training_set['Brand'] = le_brands.transform(training_set['Brand']) training_set['Model'] = le_models.transform(training_set['Model']) training_set['Location'] = le_locations.transform(training_set['Location']) training_set['Fuel_Type'] = le_fuel_types.transform(training_set['Fuel_Type']) training_set['Transmission'] = le_transmissions.transform(training_set['Transmission']) training_set['Owner_Type'] = le_owner_types.transform(training_set['Owner_Type']) #Applying encoding to Test_set data test_set['Brand'] = le_brands.transform(test_set['Brand']) test_set['Model'] = le_models.transform(test_set['Model']) test_set['Location'] = le_locations.transform(test_set['Location']) test_set['Fuel_Type'] = le_fuel_types.transform(test_set['Fuel_Type']) test_set['Transmission'] = le_transmissions.transform(test_set['Transmission']) test_set['Owner_Type'] = le_owner_types.transform(test_set['Owner_Type']) On executing the above code blocks, the training_set and test_set will be converted to completely numerical datasets as shown below. Imputing Missing Values Now we can impute or fill up the missing values. Just before imputing we will classify the predictors and target. Classifying predictors and target # Dependent Variable Y_train_data = training_set.iloc[:, -1].values # Independent Variables X_train_data = training_set.iloc[:,0 : -1].values # Independent Variables for test Set X_test = test_set.iloc[:,:].values Initializing and fitting the imputer from sklearn.impute import SimpleImputer #Training Set Imputation imputer = SimpleImputer(missing_values = np.nan, strategy = 'most_frequent') imputer = imputer.fit(X_train_data[:,8:12]) X_train_data[:,8:12] = imputer.transform(X_train_data[:,8:12]) #Test_set Imputation imputer = SimpleImputer(missing_values = np.nan, strategy = 'most_frequent') imputer = imputer.fit(X_test[:,8:12]) X_test[:,8:12] = imputer.transform(X_test[:,8:12]) The above code block will replace all missing values or ‘Nan’ with the most frequently occurring element in each respective column. Splitting The Training Data Into Training And Validation Sets from sklearn.model_selection import train_test_split #Splitting the training set into Training and validation sets X_train, X_val, Y_train, Y_val = train_test_split(X_train_data, Y_train_data, test_size = 0.2, random_state = 1) Scaling The Data #Feature Scaling from sklearn.preprocessing import StandardScaler sc = StandardScaler() #Scaling Original Training Data X_train_data = sc.fit_transform(X_train_data) #Reshaping vector to array for transforming Y_train_data = Y_train_data.reshape((len(Y_train_data), 1)) Y_train_data = sc.fit_transform(Y_train_data) #converting back to vector Y_train_data = Y_train_data.ravel() X_test = sc.transform(X_test) # Scaling Splitted training and val sets X_train = sc.fit_transform(X_train) X_val = sc.fit_transform(X_val) #Reshaping vector to array for transforming Y_train = Y_train.reshape((len(Y_train), 1)) Y_train = sc.fit_transform(Y_train) #converting back to vector Y_train = Y_train.ravel() The above code blocks on execution will transform the datasets into scaled or normalised datasets. As shown below for example data in X_train has been reduced to a smaller range. Modelling And Predicting We are down to the final stage of modelling the data. We will create a simple linear regression model to predict the Price for the given test data. But before we do that we need to check how efficient our model is for which we have created a validation set. We will use the Root Mean Log Squared Error (RMLSE) on the validation set for calculating the accuracy as mentioned in the hackathons evaluation page. Calculating Accuracy With RMLSE # Score Calculation def score(y_pred, y_true): error = np.square(np.log10(y_pred +1) - np.log10(y_true +1)).mean() ** 0.5 score = 1 - error return score #The actual recordings to be tested against y_true = Y_val Testing The Model On Validation Sets #Initializing Linear regressor from sklearn.linear_model import LinearRegression lr = LinearRegression() #Fitting the regressor with training data lr.fit(X_train,Y_train) #Predicting the target(Price) for predictors in validation set X_val Y_pred = sc.inverse_transform(lr.predict(X_val)) #Eliminating negative values in prediction for score calculation for i in range(len(Y_pred)): if Y_pred[i] < 0: Y_pred[i] = 0 #Printing the score for validation sets print(\"\\n\\n Linear Regression SCORE : \", score(Y_pred, y_true)) Output: Linear Regression SCORE :  0.763433258668093 Predicting The Price For Test Set #Initializing a new regressor lr2 = LinearRegression() #Fitting the regressor with complete training data(X_train_data,Y_train_data) lr2.fit(X_train_data,Y_train_data) #Predicting the target(Price) for predictors in the test data Y_pred2 = sc.inverse_transform(lr2.predict(X_test)) #Eliminating negative values in prediction for score calculation for i in range(len(Y_pred2)): if Y_pred2[i] < 0: Y_pred2[i] = 0 Saving the predictions to an excel sheet pd.DataFrame(Y_pred2, columns = ['Price']).to_excel(\"predictions.xlsx\") Finally, you can submit the excel in the assignment page of the hackathon and see your score on the leaderboard. The above solution has attained a leaderboard score of 0.76959 at MachineHack. Use the above code as a starter pack, use your own ideas and submit the solutions to learn and win prizes. Good luck and happy modelling!","excerpt":"MachineHack one of the leading hackathon platforms dedicated to the Data Science community, is back again with an exciting hackathon for all data science enthusiasts. This new hackathon, in partnership with Imarticus Learning, challenges the data science community to predict the resale value of a car from various features. Predicting The Costs Of Used Cars […]","categories":["Deep Tech"],"tags":["Hackathon","Machinehack","regression"],"author_name":"Amal Nair","publish_date":"2019-06-12T12:07:32","publication_year":"2019","word_count":972,"keywords":["data science","Go","API","TPU","programming_languages:R","AI","Machinehack","ML","programming_languages:Go","regression","Hackathon","Ray","R"],"extracted_tech_keywords":["AI","ML","data science","Ray","TPU","R","Go","API","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/solving-used-cars-price-prediction-hackathon\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10106012,"title":"How Moreh is Making AI Software Better with AMD","content":"AMD is rising rapidly up the AI market ever since it announced its MI300X and the software updates for ROCm. It has been building partnerships with various AI companies for testing out and delivering its products. One such company led us to Korean-based Moreh. One of the problems with AMD GPUs earlier was that the software stack was not well established, but now the company has ROCm, its alternative to CUDA. But still, it was not well suited for large GPU clusters. “Our software enables people to use AMD GPUs without any more codes, they just can run their or larger language models without any further engineering,” Junghwan Lim, head of AI group at Moreh, told AIM in an exclusive interview. Lim has also worked as a data scientist at PUBG Corporation and has also worked with Samsung in South Korea. Currently, Lim is focused on developing better software for AI workloads at Moreh and making language models smaller for better efficiency. AMD is being embraced way too much, and that’s good Moreh’s flagship AI software, known as MoAI, is positioned similarly to NVIDIA’s CUDA but boasts compatibility with existing machine learning frameworks like Meta’s PyTorch and Google’s TensorFlow, and even OpenAI’s Triton. Now, the company is helping AMD boost up its ROCm performance. In August, Moreh announced that it has been using AMD MI250 for the longest time and it is outperforming NVIDIA. According to Moreh, AMD’s MI250 Instinct accelerator, when powered by the MoAI platform, achieved 116% higher GPU throughput than NVIDIA’s A100. “We have been using more than 400 MI250 GPUs along with a few MI300X for training AI models,” said Lim. The company also plans to buy more MI300X from AMD in the future. “If anyone wants to use AMD GPUs, they can come to us without any code, and just use our software to run on them,” said Lim. This is somewhat similar to what Lamini has been doing in partnership with AMD, but Lim says that there is still a difference in the requirement of codes, as companies can also use their existing GPUs to run their models. A key aspect of Moreh’s success lies in its software being built on AMD GPU infrastructure, showcasing performance that surpasses NVIDIA GPUs in AI model development. The MoAI platform, a comprehensive software offering, is not tied to a specific hardware vendor and supports various device backends, including AMD GPUs. “Every configuration or any technique that the customer wants to apply should be easily applied without the need of any programmer. That is what we are aiming to do, and that is what differentiates us,” Lim said. “If you are building an AI model on a thousand GPUs, there can be issues such as a GPU malfunction because of hardware or software issues,” explained Lim. “The training suddenly stops and it takes hours, or sometimes days to start it again.” He explains that Moreh is also building software and devising techniques to reduce this by parallelising computers to different fragments. Large language models for the win Moreh has also finished training its own LLM for the Korean language which consists of 221 billion parameters and wishes to make upcoming models open source. “We are developing something similar to GPT or Gemini, but it is going to be open source.” Lim says that the current model is too big to open source, so Moreh is also planning to release smaller models soon. “Our models would include the code, weights, inference code, and everything else,” Lim highlighted about the recent trend in open source models which come with some or the other restrictions. In October, AMD and Korean telecommunications (KT) invested a $22 million series B round in Moreh, bringing the valuation of the startup to $30 million. The company also projected its revenue to reach $30 million by the end of 2023. KT has also bought one of the largest AMD GPU cluster in the world and are also building AI models using them. “We are supporting all those clusters and cloud systems for KT,” said Lim. KT is starting to focus on GPU cloud provider business, also providing APIs for language models, but which would just be focused on Korean language, not English. KT, which has been working with Moreh since 2021, claims that Moreh’s technology has demonstrated superior performance compared to NVIDIA’s DGX, specifically in terms of speed and GPU memory capacity. “People believe that AMD GPUs are not that suited for machine learning, but the company has been increasingly proving them wrong,” concluded Lim.","excerpt":"“Every configuration or any technique that the customer wants to apply should be easily applied without the need of any programmer. That is what we are aiming to do, and that is what differentiates us,” said Junghwan Lim.","categories":["AI Features"],"tags":["AMD","amd mi300x","amd rocm","Interviews and Discussions"],"author_name":"Mohit Pandey","publish_date":"2023-12-25T12:31:19","publication_year":"2023","word_count":759,"keywords":["amd rocm","CUDA","AMD","Go","amd mi300x","machine learning","OpenAI","AI","PyTorch","RAG","Aim","TensorFlow","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","OpenAI","Aim","TensorFlow","PyTorch","RAG","CUDA","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-moreh-is-making-ai-software-better-with-amd\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":17334,"title":"Mobikon Secures $7 Million Funding For Restaurant Dashboard","content":"Mobikon a noted customer engagement and analytics platform for hospitality industry has raised $7 million in its third round of institutional funding. The three lead investors are Sistema Asia Fund, C31 Ventures and Qualgro. According to the press statement, Mobikon will use the funding to increase its business in India, South East Asia, and the Middle East. The company also plans to work on a grass-root level into their existing markets and strengthen their core team in India, United Arab Emirates, Philippines, Singapore and new markets — United Kingdom and Australia over the next couple of years. Founded in 2012, Mobikon had raised seed funding followed by $5 million from Jungle Ventures, LifeSREDA, Qualgro, Spring Singapore and Lion Rock. With this latest funding the total capital raised by the company has gone up to $12 million. “We will be using the money to hire senior management as we grow rapidly. We expect to take the employee strength from 120 to 180-200 by the end of the year across all markets,” said Samir Khadepaun, founder of Mobikon. Mobikon provides restaurants with a single dashboard for services as well as customer relationship management. Dynamic customer profiling, intelligent marketing, notifications, alerts, and real-time reporting are some of the services they offer. As of now, Mobikon has more than 350 brands spanning over 2,4000 outlets in nine countries under their umbrella. Brands such as Lite Bite Foods that owns Punjab Grill, Specialty Group, which owns Mainland China, Pizza Express and Hard Rock Café are some of their key partners. “Mobikon Provides a unique platform for F&B outlets to capture data, get feedback, perform analytics, and improve marketing. It helps them to use technology to drive improvements in customer service, marketing, and increase revenue. Larger franchises are also seeing the benefits this can bring to their business and we are excited to be a part of this journey,” Jason Edwards, co-founder of Qualgro said.","excerpt":"Mobikon a noted customer engagement and analytics platform for hospitality industry has raised $7 million in its third round of institutional funding. The three lead investors are Sistema Asia Fund, C31 Ventures and Qualgro. According to the press statement, Mobikon will use the funding to increase its business in India, South East Asia, and […]","categories":["AI News"],"tags":[],"author_name":"Prajakta Hebbar","publish_date":"2017-08-30T04:39:37","publication_year":"2017","word_count":319,"keywords":["Go","API","funding","programming_languages:R","AI","programming_languages:Go","analytics","R","analytics platform"],"extracted_tech_keywords":["AI","analytics","R","Go","API","analytics platform","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mobikon-secures-7-million-funding-restaurant-dashboard\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162721,"title":"Constellation Perseus Launches DarkMatter, its GCC in Hyderabad","content":"Constellation Software’s Perseus Group, a global leader in vertical market software, has announced the launch of its Global Capability Centre (GCC) in Hyderabad under the name DarkMatter India Technologies. This expansion aims to strengthen the company’s technology development, research, and customer support while supporting its growing client base worldwide. The Hyderabad GCC will serve as a hub for AI, machine learning, cloud computing, and cybersecurity. Speaking at the launch, Scott Smith, president of Constellation Perseus Group, said, “India is a key market for us, and Hyderabad, with its thriving tech ecosystem and deep talent pool, presents the ideal location for our Global Capability Centre. This expansion reflects our commitment to investing in innovation and building a world-class workforce to drive the future of technology.” Investment and Expansion Plans Constellation Perseus Group has made a significant investment in its Hyderabad office, focusing on infrastructure, research, and talent development. The Hyderabad GCC currently employs 750 people and plans to triple its workforce over the next few years across domains such as software engineering, data science, and cybersecurity. The company also plans to acquire and invest in Indian technology to enhance its innovation capabilities. “By tapping into India’s exceptional growth trajectory and leveraging its dynamic innovation ecosystem, we aim to drive greater value for our customers and enhance our technological capabilities,” Smith added. Telangana’s GCC Growth and Hyderabad’s Strategic Role Hyderabad is fast becoming a global technology hub, particularly in the BFSI (Banking, Financial Services, and Insurance) sector. Duddilla Sridhar Babu, Telangana government’s IT, electronics, communications, industries, and commerce minister, highlighted the state’s progress. “Hyderabad is already home to five of the world’s top BFSI Global Capability Centres, with industry giants like Wells Fargo, Bank of America, and Goldman Sachs driving innovation from here,” he said. With the recent signing of an MoU with UBS and the addition of Constellation’s Perseus Group DarkMatter, Telangana continues to lead as a global hub for BFSI innovation. Telangana now hosts 355 GCCs, which generate $10 billion in revenue and employ 3 lakh skilled professionals. India as a whole has 1,800 GCCs, a number expected to grow to 2,500 within the next three to four years. Yet, 75% of the Fortune 500 market remains untapped. DarkMatter India’s Vision and Leadership Under the leadership of Ravi Varma, president and managing director of DarkMatter India Technologies, the company plans to further strengthen its innovation-driven approach. “We are excited to launch our centre in Hyderabad, a city that has become a hub for top technology talent. The GCC will house a dedicated product engineering team, driving innovation and enhancing the company’s suite of technology solutions,” Varma said. DarkMatter Technologies began its journey in India in 2024 with a strategic acquisition, rapidly expanding its workforce to 750 employees. The company plans to triple its workforce in the coming years to support its operations in the US, Canada, Africa, and Australia. Part of the Constellation Perseus Group’s Global Network DarkMatter Technologies operates under the Constellation Perseus Operating Group, which is part of Constellation Software Inc (CSI) – a global software and services provider. CSI focuses on acquiring, managing, and scaling software businesses. With CSI’s backing, DarkMatter Technologies is rapidly growing in the mortgage banking sector by integrating AI-powered loan origination technology solutions.With its Hyderabad GCC, strong industry partnerships, and expanding workforce, DarkMatter India Technologies is set to become a key player in global technology innovation and business expansion","excerpt":"Constellation Perseus Group has made a significant investment in its Hyderabad office, focusing on infrastructure, research, and talent development.","categories":["AI News"],"tags":["AI","GCC india"],"author_name":"Shalini Mondal","publish_date":"2025-02-01T14:09:31","publication_year":"2025","word_count":566,"keywords":["data science","Go","API","machine learning","AI","cloud computing","RAG","Aim","GCC india","GAN","R"],"extracted_tech_keywords":["AI","machine learning","data science","Aim","RAG","cloud computing","R","Go","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/constellation-perseus-launches-darkmatter-its-gcc-in-hyderabad\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002278,"title":"Ecolibrium Energy’s IoT platform SmartSense Is Accelerating Industry 4.0 For Enterprises","content":"Ecolibrium Energy, a leading provider of Predictive Asset Intelligence for industrial and commercial enterprises, with over 700 installations in over 300 global organisations, has its applications across industries, commercial buildings and smart cities. Curious Dose talked to the IoT startup Ecolibrium Energy to know more about their IoT-enabled platform called SmartSense and how it helps in digitization, energy efficiency and predictive maintenance in the IoT world. How Did It Begin? Ecolibrium was founded by Chintan Soni, and Harit Soni at the end of 2010. It started off on a high note with the establishment of Gujarat’s first smart micro-grid demonstration project with Torrent Power. Their next breakthrough came in the form of incubation at CIIE, IIM Ahmedabad. Since then, there has been no looking back and the startup now has investors such as JLL, Infuse Ventures and IFC. They have over 700 installations across 300 global enterprises and offices across India, UAE and South East Asia. Ecolibrium solutions have evolved from an energy monitoring module to a comprehensive Industry 4.0 – enabling solution that encompasses enterprise-wide energy efficiency modules and predictive maintenance solutions. The startup had to go through challenges of the lack of know-how about IoT and Predictive Maintenance and resistance to change from traditional forms of energy management but they have now gained a massive experience in the field to tackle these challenges. The Tech Behind It The tech team of Ecolibrium Energy has around 20 people with an array of departments like UI\/UX, backend, full-stack, DevOps, Data Science, embedded technology, Server and platform teams. The team consists of professionals who are proficient in Python, Java, ReactJS, Bootstrap, PostgreSQL and Time Series Database (InfluxDB), WebAPI, REST API and framework (Spring and Django), as well as technologies like Redis and RabbitMQ. Many of the team members are also experienced in machine learning algorithms. The SmartSense edge lies in its proprietary ‘Power Genome’ algorithm that the startup uses to ensure early detection of machine degradation, using the least possible sensors. These algorithms have been built using 4.1 TB data over 7 years, and are being used on over 4000 equipment today. The hardware that Ecolibrium Energy uses is called Link+. It enables strong data acquisition that is retrofit with 200 types of sensors, 12 Industry protocols, and enables a 3rd Party Integration with SAP, BMS and other IoT Platforms. The Success Story So Far The Ecolibrium solution- called SmartSense- enables enterprises to prevent unscheduled downtime; optimise maintenance & energy costs. It has been very well received among our customers across India, South East Asia and the Middle East. Oftentimes this results in them asking for customisation in reports and on the dashboard. While the traffic sometimes gets the team busy, this has been one of the biggest reasons and motivations to enhance and improve the solution from time to time. The startup was recently recognised as one of the major players in the global Predictive Maintenance market by the Market Research company, MarketsandMarkets. The report identified 19 major vendors in the Predictive Maintenance market, of which Ecolibrium Energy is the only India-based solution provider. Following are the key SmartSense offerings: SmartSense for digitization: Real-time monitoring SmartSense for energy efficiency: Helps reduce the cost of power through energy analyses & network leakage identification along with root cause & recommendations SmartSense for Predictive Maintenance: Reduced asset downtime and maintenance costs by identifying asset degradation along with root cause & recommendations SmartSense is used by industries, commercial buildings and OEMs alike. Following are some of the SmartSense success stories: 1.In one of the large plants using SmartSense, within a span of 3 years, Ecolibrium has increased its presence from 5 to 108 electrical points and 17 critical equipment. This has helped the client save over $50000 by preventing downtime, has reduced maintenance costs of this equipment by 23% and overall energy costs by 8%. 2.Another reputed tech park in Bangalore was able to save energy cost of $11200 in just over a year, using SmartSense. 3.OEM experts believe that in this digital era, it is critical to use technology to carve a niche for oneself. And in the OEM space, the best way to create this edge is by connecting, understanding and delighting customers. It is with this vision that SmartSense is working with many prominent OEM partners. Talking about its competitors in this space, the Founder said that the SmartSense platform is a full-stack IoT solution that seeks to make industries and buildings smarter. This means that SmartSense can be used by an organisation through its Industry 4.0 journey – from energy data acquisition to analysis and condition monitoring to predictive maintenance. What this also means is that, for SmartSense, there exists no competitor. For instance, if an organisation uses solution A for energy monitoring, the SmartSense analytics module can be layered on this. If an organisation uses solution B for condition monitoring, SmartSense can be used for predictive maintenance. This creates a situation where other solution providers are made partners, not competition. And this means that any organisation, whether for end-to-end solutions or for a particular purpose, can contact and engage with team SmartSense, with no worries of integration issues with other solutions. Over and above this, SmartSense stands out from the competition in the following ways: SmartSense ‘Power Genome’ ensures early detection of machine degradation, using the least possible sensors – ensuring maximum data using minimum sensors, and making it more cost effective SmartSense uses reliable algorithms that are built and enhanced using 4.1 TB data over seven years The solution uses AI and ML models use Data + Domain, making it brand agnostic SmartSense covers a wide set of assets such as motors, transformers, chillers, pumps, compressors, escalators and elevators. Has a strong acquisition layer which is retrofit with 200 types of sensors, 12 industry protocols, and enables a third-party integration with SAP, BMS and other IoT platforms The proven solution which currently has over 4000 equipment being analysed using Power Genome SaaS implementation in an Opex model, with minimal upfront investment. IoT Scenario In India When asked about the team’s views on IoT market evolving in the coming 5 years in India, they said that IoT is much larger than a market – it is a space, a movement in itself. The IoT movement has swept across every sector and industry in every stage of development and scale. It is today, an integral part of production, transport and distribution and consumption. So, it’s safe to say that every bit of the hype is completely justified. Besides, some of the world’s biggest players have begun their IoT journeys – both in terms of implementation and by investing in IoT technologies. Their belief and concentrated efforts to get a stronghold in the IoT space is proof enough. IoT will soon switch from being a mere ‘competitive advantage’ to a mandate, a bare minimum to stay competitive with others in the market. Talking about the industries with IoT demand, the CEO said that all sectors in the manufacturing space have started adopting and scaling up their IoT usage and that he sees a great deal of demand coming in from competitive sectors such as auto ancillary and engineering. But another segment that is now the dark horse and will also emerge as a major demand generator is the commercial buildings segment. Malls, technology parks, educational institutions and office buildings – they are all in the race to provide healthier and safer environments for their users or occupants. And most of these initiatives are IoT-centric. Facility management companies have all embarked on their IoT journey and there is soon going to be a boom in their IoT usage. The startup recently expanded into South East Asia and the Middle East. The next year is going to be crucial for them in establishing themselves in this market. In the long run, it is looking to establish and connect with more strategic partners and investors, and customers across the world and to expand into the US and Europe.","excerpt":"Ecolibrium Energy, a leading provider of Predictive Asset Intelligence for industrial and commercial enterprises, with over 700 installations in over 300 global organisations, has its applications across industries, commercial buildings and smart cities. Curious Dose talked to the IoT startup Ecolibrium Energy to know more about their IoT-enabled platform called SmartSense and how it helps […]","categories":["AI Features"],"tags":["IoT","ML","OEM","Technology"],"author_name":"Disha Misal","publish_date":"2019-06-19T17:30:30","publication_year":"2019","word_count":1334,"keywords":["PostgreSQL","data science","machine learning","AI","ML","Transformers","Python","Ray","Technology","analytics","OEM","IoT","Redis"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Ray","Transformers","Redis","PostgreSQL","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ecolibrium-energys-iot-platform-smartsense-is-accelerating-industry-4-0-for-enterprises\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10078349,"title":"Big Tech’s Future Looks Bleak As its Revenue Declines","content":"Recent reports claim that tech giants like Netflix, Meta, Amazon, Microsoft, Alphabet, and Apple have lost a combined market value of $2.5 trillion so far. Such claims are widely believed to discredit the narrative that big tech companies are invincible in evolving macroeconomic conditions. In the aftermath of the pandemic, several tech companies attribute high interest rates for their weak revenues. While the majority assumed that it wouldn’t be the same for big techs, the reality turned out to be quite different. Source: Forbes (The chart depicts year-to-date performances of top technology companies) Following are some key highlights from the earning reports of the top tech companies and what the future looks like for them. Meta Meta’s Q3 earnings call reported a decrease of 4% compared to the Q3 earnings last year. The decline in revenue, Meta said, is driven largely by their significant expenditure toward the metaverse project. While the expenses towards Reality Labs, Meta’s virtual reality division, reportedly increased by 24%, the revenue decreased by 49%—leading to a total operating loss of $ 3.7 billion in this division alone. Meta has also stated that all of the capital expenditure growth in 2023 is directed at the efforts to increase its AI capacity. To reduce spending costs and improve operating efficiency, Meta’s CFO David Wehner revealed that in 2023, the company plans to freeze hiring significantly and will expect the headcount at the end of next year to be the same as Q3 2022. Despite the abysmal state of the company at present, sufficient belief remains that, in the long run, the sustained efforts toward VR research will pay off. I believe in #Zuckerberg and his long term vision. It will likely make @Meta the most valuable company on the planet. But he also has to realize that investors dont care about 10 years from now, they care about 6-12 months from now. Long-term vision collides with near-term greed.— Gareth Soloway (@GarethSoloway) October 27, 2022 On the other hand, analyst Eric Seufert points out that while Meta’s earnings from ad revenue have increased by 17% on a year-over-year basis, the average price per ad has decreased by 18%. Seufert attributes this fall in Meta’s revenue as a direct result of Apple’s App Tracking Transparency (ATT) policy. This policy enables users to authorise apps to collect “data about end users and shares it with other companies for purposes of tracking across apps and web sites”, which has had a particularly significant impact on companies like Meta who earn a major chunk of their revenue from targeted advertising. Microsoft Microsoft reported its slowest revenue growth in five years. The company said that the sales of WindowsOS typically installed on new personal computers has dropped by 15%. Even with the onset of the latter stage of the COVID-19 pandemic and employees returning to offices, the demand for computers has continued to decrease. In addition, the rate of expansion of Microsoft Cloud has also reduced considerably—highlighting the rising costs incurred by companies which force them to cut down on cloud spendings. Microsoft claims that they expect a moderate growth in cloud rates in Q4. In addition, Microsoft CFO Amy Hood was also conservative about the company’s revenue growth in the personal computing market for the next quarter. Hood cited the shutdowns in China negatively impacting the supply of OEM, Surface, and Xbox consoles as the source of her claims. Amazon Amazon revealed that their operating income decreased to $2.5 billion in Q3 compared to the $4.9 billion in the same quarter last year. AWS services grew only 27% year over year compared to the 39% year over year in the same quarter last year. The company has cited the current macroeconomic uncertainties as the primary reason for this decrease. It further added that the rising inflation has had an impact over the purchasing power of the customers which, in turn, has led to a moderate sales growth. The company stated that it expects the trend to continue into the upcoming quarter. Entering the fourth quarter, the company is likely to cut costs by freezing hiring and trimming their inventory of products and services and enable allocation of resources toward higher organisational priorities. Alphabet Alphabet recorded an increase in revenue by 6% compared to the 41% a year earlier. While the company did witness withdrawals from some advertisers on YouTube and network in the second quarter, they maintain that the third quarter reported grim advertising expenditure in search ads. Heading into the next quarter, it has stated that the headcount addition is likely to be lesser than the previous quarter—with only certain important hires mainly in engineering and technical roles. Apple In the context of ongoing macroeconomic challenges and supply chain constraints, companies like Microsoft, Amazon, and Google have taken an anticipated beating in their sales but not Apple, who reported an increase of 9% in their product revenue compared to last year. However, the company is still wary about the current economic downturn and expects the next quarter to be much slower—evinced by the news that Apple may foresee a 30% fall in the production of iPhones as restrictions around COVID-19 in China continue to grow. What’s next? NYSE’s Michael Reinking told Business Insider, “It’s clear that there are headwinds for the industry after a period of unsustainable growth coming out of the pandemic, IOS privacy changes, growing competition and macro headwinds”. Reinking also said that investors are “screaming for financial discipline” from companies after a period of accelerated hiring and expenditure. Thus, while Big Tech’s Q3 earnings witnessed heavy setbacks, the next quarter doesn’t show any signs of relief either.","excerpt":"Big Tech earnings reports have been a major topic of discussion for analysts, with many finding the slowdown in growth a cause for concern.","categories":["Global Tech"],"tags":["Alphabet","Amazon","Apple","Big Tech","Meta","Microsoft"],"author_name":"Ayush Jain","publish_date":"2022-10-31T17:00:00","publication_year":"2022","word_count":938,"keywords":["Go","API","Meta","AWS","AI","cloud_platforms:AWS","Apple","programming_languages:R","Amazon","RAG","Aim","Big Tech","GAN","Alphabet","R","Microsoft"],"extracted_tech_keywords":["AI","Aim","RAG","AWS","R","Go","API","GAN","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/big-techs-future-looks-bleak-as-its-revenue-declines\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10069567,"title":"Fintech platform CredAvenue rebrands to become Yubi","content":"Fintech platform CredAvenue has rebranded itself as Yubi. The new brand captures the company’s long-term ambition of being ubiquitous in the debt ecosystem, an invisible infrastructure layer powering credit globally and ensuring frictionless access to capital. “This is an exciting transformation and a solid foundation for the constantly evolving nature of our business. Yubi represents the beginning of our global ambitions as we prepare to launch our first international office in UAE, successfully debuting in the MENA region. Another reason for the brand identity change stems from our conversations with customers and other stakeholders on how technological integration because of data security concerns around financial information remains one of the key deterrents in the advancement of digital finance. We aim to bridge this trust deficit, and in this effort of humanising the brand, we’re confident of further building our platform, which thrives on new opportunities for people and businesses alike,” said Gaurav Kumar, founder and CEO of Yubi. The new brand will reflect in every customer touchpoint through an overhaul that spans its website, social media handles and existing product lines. Yubi will deliver a richer, more consistent brand experience. With two major acquisitions (Spocto and Corpository) in the past six months, the company is growing rapidly. Founded in August 2020, it is one of the fastest-growing fintech platforms in India as it joined the unicorn club with its Series B fundraise in March this year, catapulting its valuation to USD 1.3 billion. Going forward, Yubi may even evolve beyond credit in the long term. The company caters to every requirement of borrowers and investors through five rechristened platforms with a potential launch of YubiBuild, the yet-to-be-launched real estate and infrastructure vertical of the company: 1. YubiLoans: India’s largest corporate loan marketplace for enterprises and banks, offering a seamless digital experience that accelerates the loan process from origination to disbursal five times faster. 2. YubiCo.Lend: Largest co-lending platform in India, trusted by the lending ecosystem to discover, go live and collaborate with multiple partners with quick one-time API integration. 3. YubiInvest: India’s leading Fixed Income Investment platform for bond issuance and investment for institutional and retail investors. 4. YubiFlow: A fully integrated & easy-to-use supply chain finance platform that offers trade financing solutions to lenders and corporates to strengthen their channel partner ecosystem. 5. YubiPools: End-to-end securitisation management system for Banks & NBFCs that brings unparalleled efficiencies and a data-driven approach to ABS deal-making.“We want to establish ourselves as an enabler institution to win the trust of the larger business community. As we expanded our product offerings, organically and inorganically, we understood the need for a simple brand architecture that reflects our long-term vision. With Yubi, we are reiterating our brand principles of exercising the freedom to collaborate while showcasing the utmost transparency and equitability. We stand firm on our brand promise as we continue the journey of strengthening and unlocking the real power of credit in India and beyond,” said Karanpreet Bindra, Chief Marketing Officer of Yubi.","excerpt":"With Yubi, we are reiterating our brand principles of exercising the freedom to collaborate while showcasing the utmost transparency and equitability.","categories":["AI News"],"tags":[],"author_name":"Sri Krishna","publish_date":"2022-06-22T15:17:50","publication_year":"2022","word_count":498,"keywords":["Go","API","AI","data-driven","ML","Git","Aim","Rust","GAN","R"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Rust","Git","API","GAN","data-driven"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/fintech-platform-credavenue-rebrands-to-become-yubi\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10091500,"title":"14 Open Source LLMs You Need to Know","content":"It seems like everyone is obsessed with the latest craze: large language models (LLMs). The appetite for these data-devouring behemoths just keeps growing. From GPT-3 to Megatron, the quest for bigger and better resources is far from over. So whether you’re a language processing newbie or a seasoned pro, here’s a rundown of all the open source LLMs that have hit the scene so far. Get ready to geek out! Dolly & Dolly 2.0 Within weeks of releasing Dolly, Databricks has unveiled Dolly 2.0, a model for commercial use without requiring payment for API access or data sharing with third parties. The model is a potential solution to the legal ambiguity surrounding large language models that were previously trained on ChatGPT output. BLOOM The world’s largest open-source large language model, presented by the Hugging Face team.  With the collaborative efforts of a thousand brilliant minds from across the globe, BigScience birthed BLOOM. GLM-130B The model impressively surpasses GPT-3 and the largest Chinese language model on various benchmarks, this model is a true game-changer. But that’s not all – it also boasts a unique scaling property that allows for efficient inference on affordable GPUs. The best part? The model weights, code, and training logs are all available to the public. Say goodbye to language processing limitations and hello to GLM-130B! GPT-Neo, GPT-NeoX & GPT-J In the NLP realm, the GPT-Neo, GPT-J, and GPT-NeoX models shine, providing a powerful tool for few-shot learning. Thanks to the minds at EleutherAI, these models have been crafted and made available to the public as open-source versions of GPT-3, which has been kept under lock and key by OpenAI. GPT-J and GPT-Neo, were trained on the mighty Pile dataset, a collection of linguistic data sources that spans across different domains, making them versatile and adaptable to various natural language processing tasks. But the crown jewel of this trio is GPT-NeoX, a model built on the foundation of Megatron-LM and Meta’s DeepSeed, and designed to shine on the stage of GPUs. Its massive 20 billion parameters make it the largest publicly available model. GPT-NeoX is the proof-of-concept that pushes the boundaries of few-shot learning even further. GPT-2 After initially withholding GPT-2 for nine months, due to concerns over its potential for spreading disinformation, spam, and fake news, OpenAI released smaller, less complex versions for testing purposes. In the November blog, OpenAI reported that it has witnessed “no strong evidence of misuse,” and as a result, made the full GPT-2 model available for use. PaLM Google AI disagreed with this ‘bigger the better’ assumption in the LLMs race where the size of the models have been the attention grabbing factor. The study found that bigger language models work better because they can learn from previous tasks more effectively. Based on this, Google created PaLM or Pathways Language Model, which has 540 billion parameters and is a decoder-only Transformer model. OPT Meta made a big splash in May 2022 with the release of its OPT (Open Pre-trained Transformer) models. Ranging from 125 million to a whopping 175 billion parameters, these transformers can handle language tasks on an unprecedented scale.You can download the smaller variants from Github, but the biggest one is only accessible upon request. CerebrasGPT Cerebras, an AI infrastructure firm based, made a bold move with the release of seven open-source GPT models. These models, including weights and training recipes, are available to the public free of charge under the Apache 2.0 license, challenging the proprietary systems of the current closed door industry. Flan-T5 Google AI launched an open-source language model – Flan-T5 that can tackle more than 1,800 diverse tasks. Researchers claimed that the Flan-T5 model’s advanced prompting and multi-step reasoning capabilities could lead to significant improvements. Llama Meta announced LLaMA at the end of February 2023. Unlike its counterparts, OpenAI’s ChatGPT and Microsoft’s Bing, LLaMA is not accessible to the public, but instead, Meta made it available as an open-source package that the AI community could request access to. But, just one week after Meta began accepting requests to access LLaMA, the model was leaked online, sending shockwaves through the tech community. Read here: 7 Ways Developers are Harnessing Meta’s LLaMA Alpaca From the halls of Stanford University emerged Alpaca. The model was created by fine-tuning LLaMA 7B with over 50k demonstrations following instructions from GPT 3.5. It was trained and tested at a mere $600, instead of the millions. Since its release, Alpaca has been hailed as a breakthrough, Though it started small, with a Homer Simpson bot, the model quickly proved its versatility.","excerpt":"Every model open sourced since the genesis of GPT-2.","categories":["AI Trends"],"tags":["GPT-4","Open Source","OpenAI"],"author_name":"Tasmia Ansari","publish_date":"2023-04-15T16:00:00","publication_year":"2023","word_count":759,"keywords":["ChatGPT","Hugging Face","TPU","Open Source","OpenAI","AI","GPT-4","Transformers","NLP","Aim","few-shot learning","Databricks"],"extracted_tech_keywords":["AI","NLP","ChatGPT","OpenAI","Aim","Hugging Face","Transformers","few-shot learning","TPU","Databricks"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/14-open-source-llms-you-need-to-know\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":17022,"title":"Battle of AI is turning software giants into chipmakers","content":"In the battle to bring AI home sooner, tech giants Microsoft, Google, Apple and Facebook have joined the likes of Nvidia, Qualcomm, Intel and Broadcom among others to build their own AI chips that are powering their hardware.  Over the last few years, there has been a frenetic activity in this space with tech titans becoming chipmakers to support the AI capabilities of their products and improve their AI systems. Legacy giants need chips in the data centres to train powerful neural networks and they also need these chips in the products to help devices run AI capabilities smoothly. Hence, the shift from mobile to AI is seeing tech titans producing AI-focused chips. Today, Google, Apple and Microsoft have some of the most visible projects in the AI space with announcements made over the last two years. Tech companies turning chipmakers Leading the race is the search software giant Google whose Tensor Processing Unit (TPU) is designed for training AI system. This ground-breaking advancement in AI helps Google provide faster response time for search and voice search and played a pivotal role in AlphaGo beating a reigning champion at the ancient Chinese board game. At a recent I\/O conference, Google CEO Sundar Pichai revealed a new kind of second generation AI chip designed in-house that could train and execute deep neural networks that power automated translation, image and speech recognition and even robotics.  In fact, Google’s shift to produce Cloud TPUs is in direct competition with NVIDIA, the gaming giant that adapted its GPUs for AI data processing. Reportedly, Google also opened Cloud TPUs for business, and has also released Tensor Flow Research Cloud for free for research work. On the other hand, the Cupertino giant, which is highly secretive company about its developments was recently in the news for developing an Apple Neural Engine chip to add more advanced AI capabilities across its products. Most AI tasks such as speech recognition can affect the battery life and the iPhone maker plans to use the AI chip in conjunction with the standard processor and graphics chip to offset the battery life. Redmond giant Microsoft too has designed its own chips designed to boost AI capabilities of their latest invention – the HoloLens, a holographic computer and head mounted display that enables users to interact with holograms. Google and Microsoft’s projects are the most visible part of a new AI-chip industry springing up to challenge established semiconductor giants such as Intel and Nvidia. From gaming to AI – here’s how NVIDIA started the revolution The gaming giant won the chip market when it trained its GPUs that powered video game graphics to handle other types of large computations. Today, GPUs have become the go-to hardware in data centres, running the company’s machine learning system, with NVIDIA’s GPUs deployed by Microsoft, Amazon and Alibaba among others. As opposed to other players such as Intel and Advanced Micro Devices, NVIDIA identified a high growth opportunity in deploying GPUs for highly intensive computing and deep learning and started releasing their early version (CUDA) to researchers at university. By bulking up on the talent front, by snagging top performers from SGI, the computer graphics company now owned by HPE and leading graduate programs, trained its eyes on making NVIDIA products that could support machine learning algorithms. Known for their killer marketing strategy, the Santa Clara-headquartered gaming company now develops processors that power autonomous capabilities in cars. Build vs Buy Why are companies pushing to build their own chip production facility when players such as Intel, NVIDIA, ARM already exists? The race for AI is driven by the underlying need of more computing power to support training deep neural networks that can analyse vast amounts of data. This means, supplementing CPU power with additional processor, such as Google’s TPU and in the case of Microsoft, it is Field Programmable Gate Array. Meanwhile, rumours are rife that Apple bought its own silicon chip fabrication unit and shunned process development partner to develop AI chips to power its smartphones. The biggest reason is that companies want more autonomy in tailoring the production process more to their needs. In the case of Apple, it wants to tweak its hardware according to the needs of the functionality it wants to incorporate in its new line of iPhone. In fact, Google’s TensorFlow Lite is a version developed for mobile and is addressing almost the same set of use cases that Apple is working on. How are tech giants closing the gap in the battle of chips? With AI-related growth taking off, especially in areas such as healthcare, automotive, retail among other areas, legacy companies are quick to snap up startups, they see as AI-enablers. These startups are also pegged as potential AI-stocks that can change the market rapidly and accelerate bringing next-gen to market. Case in point – semiconductor giant Intel acquired chipmaker Altera in 2015, an acquisition driven by the rise of AI. Earlier this year, Intel paid a whopping $15.3 billion to an Israeli chip startup Mobileye and in good turn of events, Intel is reportedly shunting the autonomous driving team to the startup’s HQ to accelerate tech to market. Other potential startups in this arena are UK-based Graphcore, machine learning chipmaker that closed a $30 million round of funding recently and has Deep Mind CEO Demis Hassabis as its backer and Wave Computing, focused on machine-learning chips. Graphcore’s USP lies in making chips that are mapped to machine learning algorithms as opposed to NVIDIA’s chips, that were originally meant for powering video games. Will it phase out cloud? With Google putting all its might behind Cloud-ML, Cloud TPUs can jostle AWS as the number #1 cloud computing partner. Analysts point out that the next generation of web-based computing services could reduce the need for cloud servers. A host of next gen technology such as VR necessitates the need for powerful AI chips to power devices that are used in a wide range of settings such as manufacturing, retail, logistics and other areas for autonomous intelligence. Instead of depending on data centres to process responses, companies want to outfit VR headsets with AI chips to reduce time lag and speed up the responses. With incremental growth in AI over the last few years, and emerging technologies have compelled tech behemoths to expand their ecosystem by flexing their muscles and experiment with powerful AI chips, built in-house or in collaboration to power their services and products.","excerpt":"In the battle to bring AI home sooner, tech giants Microsoft, Google, Apple and Facebook have joined the likes of Nvidia, Qualcomm, Intel and Broadcom among others to build their own AI chips that are powering their hardware.  Over the last few years, there has been a frenetic activity in this space with tech titans […]","categories":["IT Services"],"tags":["AI Chips"],"author_name":"Richa Bhatia","publish_date":"2017-08-17T06:50:46","publication_year":"2017","word_count":1076,"keywords":["machine learning","TPU","AWS","AI","neural network","cloud computing","ML","AI Chips","Ray","deep learning","TensorFlow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","Ray","TensorFlow","cloud computing","AWS","TPU"],"url":"https:\/\/analyticsindiamag.com\/it-services\/battle-ai-turning-software-giants-chipmakers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10039648,"title":"Tech Behind DINO, Facebook’s Open-Source ML Model For Computer Vision","content":"Facebook AI launched a computer vision system called DINO to segment unlabeled and random images and videos without supervision. The open-source PyTorch framework implementation and pre-trained models for DINO is currently available on GitHub. DINO stands for self DIstillation with NO labels. Facebook has developed the new model in collaboration with researchers at INRIA, Sorbonne University, MILA and McGill University. DINO has set a new state of the art among self-supervised methods. Self-supervised vision transformers (ViT), a type of machine learning model, carry explicit information about the semantic segmentation of an image and perform better than supervised ViTs and convolutional neural network (CNNs). Unlike traditional supervised learning, DINO doesn’t require large volumes of annotated\/labelled data. In DINO, high accurate segmentation can be solved with self-supervised learning and a suitable architecture. Interestingly, DINO focuses on the near object even in highly ambiguous situations. Source: Facebook (Showcasing the visual representation of the original image, followed by supervised model and unsupervised model (DINO)) “Our model can discover and segment objects in an image or a video with absolutely no supervision,” said researchers at Facebook AI, pointing at the visuals of the original video trained versus DINO (self-supervised vision transformers). Source: Facebook (Self-attention maps of neural network on videos of a puppy, a horse, a BMX rider, and a fishing boat, using DINO) Facebook’s CTO Mike Schroepfer, while explaining the nuances of its new computer vision system, said the DINO self-supervised model is inspired by how young children learn the language, physics, and more without formal instruction. Here’s our new computer vision system achieving state of the art results in image segmentation, without needing any labeled training data. This new model was trained on random, unlabeled data, but quickly achieved state-of-the-art results. It’s awesome. pic.twitter.com\/6Y4zjaImFY— Mike Schroepfer (@schrep) April 30, 2021 More with less Facebook AI claimed that it lets users train models using limited computing resources. Besides launching DINO, the company has also introduced PAWS, a new model-training approach that delivers accurate results with much less compute. The open-source PyTorch implementation of PAWS (predicting view assignments with support samples) is also available on GitHub. “When pretraining a standard ResNet-50 model with PAWS using just 1 percent of the labels in ImageNet, we get state-of-the-art accuracy while doing 10x fewer pertaining steps,” claimed Facebook researchers. For instance, while training a student network, its self-supervised computer vision system matches the output of a teacher network over different views on the same image. Source: Facebook Similarly, in another example where the model was asked to recognise duplicate images, Facebook’s DINO outperformed existing models, even though it was not trained to solve that particular problem. DINO had the highest accuracy compared to ViT trained on ImageNet and MultiGrain, which were so far considered to have the highest accuracy for duplicate detection. Source: Facebook (The image showcases how DINO can recognise near-duplicate images taken from the Flickr dataset, where red and green outlined images indicate false and true positives) Additionally, Facebook’s new model discovers object parts and shared characteristics across the model and learns to categorise and structure images into groups based on physical properties like animal species or biological taxonomy. Source: Facebook (Feature representation of unlabelled data) Lately, Facebook has been bullish on open-source frameworks, tools, libraries and models for research and developers to deploy in large-scale production. Last month, Facebook launched an open-source machine learning library called Flashlight that lets researchers execute AI applications seamlessly using C++. Its latest machine learning library, written entirely in C++, is currently available on GitHub. A year before that, Facebook had launched an open-source graph transformer networks (GTN) framework for effectively training graph-based learning models.","excerpt":"Facebook AI launched a computer vision system called DINO to segment unlabeled and random images and videos without supervision. The open-source PyTorch framework implementation and pre-trained models for DINO is currently available on GitHub. DINO stands for self DIstillation with NO labels. Facebook has developed the new model in collaboration with researchers at INRIA, Sorbonne […]","categories":["AI Trends"],"tags":["Facebook latest"],"author_name":"Amit Naik","publish_date":"2021-05-05T16:00:00","publication_year":"2021","word_count":605,"keywords":["machine learning","TPU","AWS","AI","neural network","PyTorch","ML","Transformers","computer vision","Aim","Facebook latest"],"extracted_tech_keywords":["AI","machine learning","ML","neural network","computer vision","Aim","PyTorch","Transformers","AWS","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/tech-behind-dino-facebooks-open-source-ml-model-for-computer-vision\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10097088,"title":"The Dark Consequence of AI&#8217;s Data Cannibalism","content":"AI is eating itself. The internet has now become an AI dumping ground and the models being trained on the web are feeding on its own kind. That’s data cannibalism. In an article for The New Yorker, acclaimed science fiction author Ted Chiang drew attention to the perils of AI copies breeding copies in a digital photocopying of sorts. He likens this burgeoning dilemma to the JPEG effect, where each subsequent copy degrades in quality, revealing a mosaic of unsightly artefacts. As the boundaries of AI replication blurs, the point to ponder is, what happens as AI-generated content proliferates around the internet, and AI models begin to train on them, instead of primarily human-generated content? Recent findings by researchers from Britain and Canada state that generative AI models exhibit a phenomenon known as “model collapse“. This degenerative process occurs when models learn from data generated by other models, leading to a gradual loss of accurate representation of true data distribution. Remarkably, it is deemed unavoidable, even in scenarios where the conditions for long-term learning are nearly ideal. Repercussions According to Ross Anderson, a professor of security engineering at Cambridge University and co-author of the ‘model collapse’ research paper, the internet is at the risk of being flooded with insignificant content, similar to how the oceans are littered with plastic waste. This flood of content could hinder the training of new AI models through web scraping, benefiting firms that have already amassed data or control large-scale human interfaces. He further mentioned the recent Internet Archive fiasco when AI startups were aggressively mining the website for valuable training data, further highlighting this concern. According to a recent report by media research organisation NewsGuard, an alarming trend has emerged in which websites are being filled with AI-generated junk content to attract advertisers. The report reveals that over 140 prominent brands unknowingly end up paying for advertisements displayed on websites powered by AI-written content. This growing spammy AI-generated material poses a threat for the very AI companies responsible for these models. As training data sets become increasingly saturated with AI-produced content, concerns are being raised regarding the diminishing utility of language models. “There are many other aspects that will lead to more serious implications, such as discrimination based on gender, ethnicity or other sensitive attributes,” Ilia Shumailov a research fellow at Oxford University’s Applied and Theoretical Machine Learning Group said, especially if generative AI learns over time to produce, say, one race in its responses, while “forgetting” others exist. Inclusivity All The Way The current models are already in the bad books of AI ethicists for their lack of inclusivity. In 2021, a group of researchers warned about the white male problem in language models. Lead author Anders Søgaard, a professor at UCPH’s department of computer science, explains that these models exhibit systematic bias. Surprisingly, they align best with the language used by white men under 40 with lesser education, while showing the weakest alignment with language from young, non-white men. This discovery emphasises the pressing need to address and rectify the biases within language models to ensure fairness and inclusivity for all. Along similar lines, Shumailov said, “To stop model collapse, we need to make sure that minority groups from the original data get represented fairly in the subsequent datasets.” While some companies are working towards a more inclusive AI —  like Meta’s recently released open-sourced consent-driven dataset of recorded monologues called ‘Casual Conversations v2‘. This enhanced version has been built to serve a broad spectrum of use cases. By offering researchers a robust resource, it empowers them to evaluate the performance of their models with greater depth. But on the other hand, we have Google which has been in the news for not-so-good reasons since renowned AI ethicist Timnit Gebru was fired followed by the exit of the rest of the team calling Google a ‘white tech organisation’. Flawed, not useless While language models have a long, long list of ethical defects, they come with an array of advantages. For instance, Shumailov along with his team originally called ‘model collapse’ — the effect model dementia, but decided to rename it after objections from a colleague. “We couldn’t think of a replacement until we asked Bard, which suggested five titles, of which we went for The Curse of Recursion,” he wrote. Currently, language models are becoming a part of every second company’s strategy. Firms in every sector are learning to unlock the full potential of the advanced chatbots based on language models like GPT-4. While it is too early to judge, companies are shifting their way through the generative AI clutter trying to figure out the best use cases for their business.","excerpt":"Eventually, the AI models producing content currently will start training on data generated by themselves leading to data cannibalism","categories":["AI Trends"],"tags":["AI Models"],"author_name":"Tasmia Ansari","publish_date":"2023-07-18T13:00:00","publication_year":"2023","word_count":779,"keywords":["AI Models","Go","API","machine learning","AI","chatbots","Git","Ray","Aim","generative AI","R"],"extracted_tech_keywords":["AI","machine learning","generative AI","Aim","Ray","chatbots","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-dark-consequence-of-ais-data-cannibalism\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10133665,"title":"Andrej Karpathy Praises Cursor Over GitHub Copilot","content":"Cursor is currently the hottest AI developer tool. And its prowess is further reinforced when Andrej Karpathy decides to praise it. In a recent post on X, Karpathy said that he is trying Cursor along with Claude Sonnet 3.5 instead of GitHub Copilot and said it’s a net win. Register for NVIDIA AI Summit India “Just empirically, over the last few days most of my ‘programming’ is now writing English (prompting and then reviewing and editing the generated diffs), and doing a bit of ‘half-coding’ where you write the first chunk of the code you’d like, maybe comment it a bit so the LLM knows what the plan is, and then tab tab tab through completions,” said Karpathy, adding that he is able to do coding a lot faster with the help of the tool. Programming is changing so fast… I'm trying VS Code Cursor + Sonnet 3.5 instead of GitHub Copilot again and I think it's now a net win. Just empirically, over the last few days most of my \"programming\" is now writing English (prompting and then reviewing and editing the…— Andrej Karpathy (@karpathy) August 24, 2024 Further in the thread, he added that with the capabilities of LLM shifting so rapidly, it is important for developers to continually adapt the current capabilities. “The tool is now a complex living thing,” he added. “All this talk of Claude + Cursor and becoming capable of building anything you put your mind to (no matter your skill set) is warranted. If this is the future, I want to live in it,” said Jordan Singer from Figma. Sebastian Raschka, LLM research engineer at Lightning AI, said that though using the tool is fun, he sometimes enjoys unassisted coding as well. “It’s like driving a manual car (\/stick). Not the most practical but fun,” he explained with an analogy. Pratik Desai, CEO of KissanAI, added, “Once you start using it regularly, you are going to develop your own tricks suitable for your style, that may not be over there in their onboarding tutorials.” Anysphere, the company building the tool, recently raised $60 million led by Andreessen Horowitz, and included contributions from OpenAI’s Jeff Dean, John Schulman, Nat Friedman, and Noam Brown, valuing the company at $400 million. The hype around Cursor AI is true. To give an example, Ricky Robinett, VP of developer relations at Cloudflare, posted a video of his eight-year-old daughter building a chatbot on the Cloudflare Developer Platform in just 45 minutes using Cursor AI, documenting the whole process, even the spelling mistakes while giving prompts! Read: Cursor AI is Better at Coding Than GitHub Copilot Will Ever Be “The hottest new programming language is English,” Karpathy said more than a year back. Now for him, it is coming true in the purest sense.","excerpt":"“The tool is now a complex living thing,” Karpathy added.","categories":["AI News"],"tags":["Andrej Karpathy","Github Copilot"],"author_name":"Mohit Pandey","publish_date":"2024-08-24T08:49:33","publication_year":"2024","word_count":465,"keywords":["Go","API","Andrej Karpathy","OpenAI","AI","programming_languages:R","RPA","Github Copilot","Git","llm_models:Claude","GitHub","R"],"extracted_tech_keywords":["AI","OpenAI","R","Go","Git","GitHub","API","RPA","llm_models:Claude","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/andrej-karpathy-praises-cursor-over-github-copilot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10172112,"title":"Wipro, Zoho Lead India&#8217;s UK Growth Surge: Report","content":"Indian IT majors Wipro and Zoho are spearheading the surge in Indian-owned businesses in the UK, according to Grant Thornton’s India Meets Britain Tracker 2025. Wipro IT Services UK Societas, a UK-based subsidiary of Wipro, recorded an astounding 448% revenue growth, topping the list of fastest-growing Indian firms in the UK. According to the report, this dramatic growth follows the company’s strategy of localising services, winning long-term contracts—including a £500 million, 10-year agreement with Phoenix Group—and establishing innovation and delivery centres in London, Reading, and Edinburgh.Zoho Corporation Limited entered the Tracker for the first time, achieving 197% annual revenue growth, making it the fastest-growing new entrant. With £67 million in revenue, Zoho credited its success to a smart sales approach and tripling its Bletchley team to support UK clients better. The Tracker now tracks 1,197 Indian-owned UK companies—a 23% increase from last year, reflecting broader expansion trends tied to innovations in AI, cloud, and digital services. Collectively, these businesses generated £72.1 billion in revenue, employing 126,720 talents across the UK, and contributed £67.3 million in corporation tax. The Tracker, developed in collaboration with the Confederation of Indian Industry (CII) and, for the first time, India Global Forum (IGF), identifies the top fastest-growing Indian companies in the UK as measured by percentage revenue growth year-on-year. It includes Indian-owned corporations with operations headquartered or with a significant base in the UK, with a turnover of more than £5 million, year-on-year revenue growth of at least 10% and a minimum two-year track record in the UK, based on the latest published accounts filed as of 31 March 2025.Anuj Chande OBE, partner and head of South Asia Business Group, Grant Thornton UK Advisory & Tax LLP, said in the report that the 2025 research identified 1,197 Indian-owned companies operating in the UK, an increase of over 23% from last year’s 971. “I’m delighted that this is the highest number yet, crossing the 1,000 threshold for the first time, and the largest year-on-year increase since we began measuring total numbers in our Tracker in 2017. This is an exciting time for UK-India bilateral relations as a new era begins with the conclusion of negotiations on the UK–India Free Trade Agreement (FTA),” he said, adding that the historic milestone was reached on 6 May 2025, and is expected to increase UK GDP by £4.8 billion and UK wages by £2.2 billion annually in the long run. The technology, media, and telecom (TMT) sector remains the largest in the Tracker, comprising 31% of companies in 2025, up from 27% in 2024. Pharmaceuticals and chemicals moved to second place with 22%, while manufacturing and engineering declined to 14%. Financial services grew to 10% of companies, led by the Indian government and private banks.","excerpt":"The 2025 research identifies 1,197 Indian-owned companies operating in the UK, an increase of over 23% from last year’s 971.","categories":["AI News"],"tags":["AI","Technology","uk","Wipro","zoho"],"author_name":"C P Balasubramanyam","publish_date":"2025-06-20T14:33:29","publication_year":"2025","word_count":455,"keywords":["Wipro","zoho","Go","uk","AI","programming_languages:R","innovation","programming_languages:Go","Git","Technology","GAN","R"],"extracted_tech_keywords":["AI","R","Go","Git","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wipro-zoho-lead-indias-uk-growth-surge-report\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10083469,"title":"Computer Vision Trends That will Dominate the Industry","content":"Computer vision is the eyes of the machine. AI models are built to recreate living beings’ capability of looking at the world around them and interpreting and understanding it. Machines do this by analysing images, videos, and objects around them. Recent developments like Tesla’s Optimus Robot and Full-Self Driving majorly relied on computer vision for object detection and image tracking. Even 2D to 3D models use computer vision for image analysis and interpretation. The Conference on Computer Vision and Pattern Recognition (CVPR) 2022 saw a total of 8,161 submissions and thousands of them tried to solve a different problem in AI\/ML. Seeing these advancements and developments through computer vision, let’s look at some of the predictable trends in the field. Read: Top AI Predictions for 2023 Autonomous vehicles The objective of achieving self-driving vehicles has been a long-running one. One of the most important aspects of achieving these autonomous vehicles is identifying the objects around the vehicle for it to traverse and navigate safely. This is where computer vision-based algorithms come into the picture. Companies like Tesla have been adopting techniques like auto-labelling to further their autonomous driving vehicles. The same technology can be useful for other transportation-based applications like vehicle classification, traffic flow analysis, vehicle identification, road condition monitoring, collision avoidance systems, and driver attentiveness detection. Increased use of edge computing As the demand for real-time processing of visual data increases, there will likely be a trend towards using edge computing to perform computations closer to the source of the data. Traditionally, computer vision tasks have been performed on centralised servers or cloud-based systems, which can be time-consuming and require a stable internet connection. Edge computing enables these systems to make quick and accurate decisions based on visual data without sending the data back and forth to the cloud for processing. Robotics One of the main areas where computer vision is expected to play a significant role in robotics is in enabling robots to navigate and manipulate objects in their environment by using algorithms to analyse images and video from cameras, robots to detect and identify objects, as well as understand their shape, size, and location. This can allow robots to perform tasks such as grasping and moving objects, as well as avoiding obstacles and navigating through complex environments. Robots can understand and respond to human behaviour through computer vision by analysing facial expressions, body language, and other visual cues. As a result, robots could potentially be used in applications such as customer service, education, and healthcare. Healthcare, safety, & security Medical image analysis: Computer vision can be used to analyse medical images, such as X-rays, CT scans, and MRIs, to detect abnormalities or diseases. For example, a computer vision system could be trained to recognize the presence of a tumour in an MRI scan. Diagnosis and treatment planning: Computer vision can be used to assist with diagnosis and treatment planning. For example, a computer vision system could be used to analyse medical images and recommend the most appropriate treatment for a patient based on their specific condition. Monitoring patient health: Computer vision can be used to monitor patient health by analysing vital signs such as heart rate, respiration rate, and blood pressure. Robotic surgery: Computer vision can be used in robotic surgery to assist surgeons in performing complex procedures. For example, a computer vision system could be used to guide the movement of a surgical robot, ensuring that it stays on course and avoids damaging any surrounding tissue. Retail Shops and retail stores can be installed with cameras to analyse items on shelves and automatically detect the stock and also recognise which items sell the most. Apart from inventory management, AR can also be used to create “virtual fitting rooms” or “virtual mirrors” to try out items without touching them or even going to the story, much like how filters work on Snapchat or Instagram by superimposing items on top of the person in front of the camera. Data-centric AI Optimising the quality of data is as important as increasing the quantity of data when it comes to training models and building algorithms. Image recognition models are built for enabling machines to identify and classify pictures of different objects and labelling these images is important for extracting the correct information from the data. Therefore, unsupervised and automated computer vision technology would increase the accuracy and information when there is less availability of data. 3D reconstruction In 2022, we witnessed text-to-image models and that eventually led to text-to-3D models. This further led to 3D reconstruction models using methods like Neural Radiance Fields (NeRF) that could recreate 2D images into 3D meshes that can be used for recreation of scenes and also for building models in the metaverse. This can also be used for creating immersive virtual and augmented reality experiences, allowing users to interact with digital environments in a more realistic and natural way. SpaceTech Apple has computer vision based  applications that can detect objects in the sky if you point your phone towards it. This is just one of the use cases of computer vision in the space industry. By analysing imagery and data collected by satellite or aerial sensors we can accurately map and analyse the Earth’s surface and environment. Moreover, by analysing geospatial data through the satellites, we can predict future disasters such as earthquakes and hurricanes and then effectively work to reduce their impact. Computer vision can also be used for space exploration by locating and identifying space objects and also detecting their various characteristics. Identifying these objects can also be used for cleaning up the space, for which NASA, ISRO, and all other big-tech companies are planning projects. Read: It’s Time for ‘Swachh Antariksh Abhiyan’","excerpt":"There is a lot that the eyes of the machine can see","categories":["AI Features"],"tags":["computer vision applications","Robotics"],"author_name":"Mohit Pandey","publish_date":"2022-12-31T10:00:00","publication_year":"2022","word_count":951,"keywords":["Go","AI","ML","image recognition","computer vision applications","Git","computer vision","Robotics","Ray","object detection","edge computing","R"],"extracted_tech_keywords":["AI","ML","computer vision","Ray","image recognition","object detection","edge computing","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/computer-vision-trends-that-will-dominate-the-industry-in-2023\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10017787,"title":"Bringing Simplicity In HR Intelligence: The Startup Story Of GoEvals","content":"In today’s volatile and fast-moving world, especially in the post-COVID era, organisations have a hard time implementing robust employee management processes and advanced recruitment practices. Their needs are critical and urgent; however, the solutions currently available in the market are either too complicated or don’t serve the need. Addressing this concern, GoEvals, an HR Tech startup, has developed simplified AI-powered intelligent tools for managing employees’ entire lifecycle. While artificial intelligence in human resource functions is not new, GoEvals has leveraged Markov decision process to develop a unique decision matrix for MSMEs to streamline the time-consuming parts of hiring. Founded in December 2019, GoEvals has seen a massive acceleration of digital transformation amid COVID, where organisations of all sizes are looking to digitise their hiring processes. Digital capability has become the sine qua non of the human resource function. We got in touch with Dr Gaurav Hirey, the Founder & CEO of GoEvals, to understand his Mumbai-based startup’s success story and how its innovative and simplified AI-based tools are helping MSMEs execute HR digitisation. “We have been working on a blueprint to simplify the process of HR technology implementation for a decade and more. And, amid COVID, we realised the HR tech solutions out there are either IT-driven or follow a complex IT framework making it challenging for organisations,” said Dr Hirey. Leveraging AI According to Dr Hirey, reports released at the beginning of 2020 stated that 42% of organisations believe their HR tech initiatives are either partially successful or a disaster. This high rate of failure is reflective of the faultlines in existing solutions. “We have always believed there is an easier and less destructive way to leverage AI and data analytics in HR, and this is what inspired the idea of creating GoEvals,” said Dr Hirey. “Our platform offers intelligent digital HR tools that are simple, fast, reliable, and fits customers’ budgets.” The startup applies artificial intelligence along with machine learning and predictive algorithms to streamline the resource-intensive parts of hiring. Using the Markov decision process, GoEvals has developed a unique and dynamic GoEvals decision matrix that drives all AI and analytics processes. “This decision-making matrix is developed and validated by our in house team, and we constantly keep updating it,” added Dr Hirey. Explaining the process, Dr Hirey said — while setting up the Compatibility Index Test, the recruiter or HR personnel is asked to decide on the compatibility parameters most suited for the hiring roles. This set of data is loaded to the matrix for recording and setting standards. The recruiter can then send the test to the required candidate. The platform also has an inbuilt follow-up management process that keeps prompting candidates or evaluators to complete their roles, eliminating grunt work and freeing more time for recruiters. The platform actively captures multiple data points the candidate shares during the test. Using the MDP framework and the decision-making matrix, the outcomes are derived by the algorithms developed by the GoEvals team. “These algorithms understand the data and try to arrive at decisions about the candidate’s actual ability to do the job, their cultural fit and personality,” added Dr Hirey. The data is then presented (using data visualisation)to the recruiters and hiring managers. The platform also applies predictive analytics to forecast the candidate’s ability to deliver. The reports provide insights and use graphs, text and visuals that are easy to read and understand to facilitate data-based decision making. The startup’s core tech stack comprises Vue.js, Laravel Web framework, NoSQL database and MongoDB. Growth & Future Plans GoEvals is a bootstrapped startup; however, it is currently seeking seed funding and is open to like-minded individuals, startups, and MSMEs interested in being a part of its journey. The startup is also looking to expand globally, starting with launches in Thailand and UAE, coming year. “After the recent launch of our first two tools, we are currently working on onboarding about 50 clients by March 2021,” said Dr Hirey. With its advanced AI tools, the startup competes directly with tech companies like People Strong, ADP, Gusto, Zoho, Bamboo HR, and others. “We believe that simplicity, especially in an age of information overload and complex solutions, is going to help lift the game of employee management and recruitment process,” said Dr Hirey.","excerpt":"In today’s volatile and fast-moving world, especially in the post-COVID era, organisations have a hard time implementing robust employee management processes and advanced recruitment practices. Their needs are critical and urgent; however, the solutions currently available in the market are either too complicated or don’t serve the need. Addressing this concern, GoEvals, an HR Tech […]","categories":["AI Startups"],"tags":["AI Algorithms","artificial intelligence machine learning data"],"author_name":"Sejuti Das","publish_date":"2021-01-13T13:00:00","publication_year":"2021","word_count":710,"keywords":["artificial intelligence","machine learning","AI","MongoDB","ML","R","artificial intelligence machine learning data","RAG","AI Algorithms","analytics","SQL","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","predictive analytics","MongoDB","R","SQL"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/bringing-simplicity-in-hr-intelligence-the-startup-story-of-goevals\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10070858,"title":"Who should a Chief Data Officer report to?","content":"In the book, The Chief Data Officer’s Playbook and Data-Driven Business Transformation: How to Disrupt, Innovate and Stay Ahead of the Competition, Caroline Carruthers defines a CDO as a person who is business-focused and understands the strategy and direction of the business. The primary role of a CDO is to underpin business strategy with data. They are responsible for maximising the strategic potential of data across the organisation. Heidrick and Struggles, an executive search company based out of Chicago, recently published a report called “2021 Europe and US Data, Analytics, and Artificial Intelligence Executive Organization and Compensation Survey”. The report gives valuable insights into various aspects of the role of data executives and data heads. The report highlights that 54 percent of data executives engage in marketing and customer engagement functions, followed by 28 percent in operations and 25 percent in sales. While the report is more focused on the role of CDOs in the US and Europe, we look at what responsibilities are CDOs in India entrusted with. Function areas under CDOs’ command According to the report, the majority of data analytics and AI leaders are responsible for data science, artificial intelligence and machine learning, business intelligence, and analytics functions. But with variation in reporting structures, the responsibility of data heads also varies. Credit: Heidrick and Struggles Generally, data architecture, data warehousing and data platform are part of the IT ecosystem in any organisation and is the responsibility of technology officers. However, if the role of data analytics heads is more business-oriented, they do not command these function areas. “The CDO role can be seen to have three layers – run, change & transform with a different time horizon focus. The first is the ‘run’ layer which is closest to the endpoint, i.e. factory\/ store, the next is the CXO level or “change” layer, and the last one is the group level central team or the “transform” layer. The idea is to make sure that the digital capability in all these three layers exists and works to complement each other,” mentioned Pankaj Rai, Group Chief Data and Analytics Officer, Aditya Birla Group. At Aditya Birla Group, the data analytics team has been organised in what they call the 6P structure. The first P stand for Purpose, followed by PMO, Project, Platform, People and Partnerships. The team that Rai is heading intends to create new digital capabilities, and hence it assumes the role of a digital orchestrator. “In a lot of organisations, data engineers, data governance and validation, analytics and data science come under the CDO, but it also depends on the culture of the organisation. If the culture of the organisation is such that a lot of people are from the engineering side, they will be able to hold the court with their expertise. In those cases, Analytics and science would be some acquired skill,” said Debayan Bose, Head of Data Science and Analytics, Ibibo Group. Who does the CDO report to? Globally, data, analytics, and AI executives most often report to the CEO. However, there is significant regional variation. In the US, the data and analytics head mostly report to the CTO, followed by CEO and CIO. On the other hand, in Europe, they most often report to the CEO, followed by the COO. This difference could be due to the fact that in the US data and analytics has been largely seen as technology function. On the other hand, in Europe, where data integrity and privacy are a major concern (almost a deal-breaker), the data leadership roles are integrated into the business functions. Credit: Heidrick and Struggles The reporting authority of data head companies varies with organisations’ structure. It may even reflect the said company’s stage of digitalisation and that of its digital maturity. In fact, the definition of and responsibilities of CDO often overlap with that of the chief digital officer and chief analytics officer. For a more blanket reporting structure, CDO may report to CTO or CIO. In some cases, the CDO may report to the CEO, COO or even CFO. “CDO encompasses data governance, data-led decisions and hence reporting to CEO would be the ideal structure to have,” said Mathangi Sri, Chief Data Officer at YUBI. Agreeing with Mathangi, Himanshu Gupta, Co-Founder & COO, WeRize said that the data science team should report to the CEO or the founders and not to the technology team; he added that data science to be aligned more with business and company objectives. “Data science shouldn’t be focusing on only technology. They also need to help the businesses achieve their revenue targets, profit targets, NPA and risk targets, as well as the operational efficiency targets,” he said.","excerpt":"The primary role of a CDO is to underpin business strategy with data.","categories":["IT Services"],"tags":[],"author_name":"Zinnia Banerjee","publish_date":"2022-07-12T18:25:57","publication_year":"2022","word_count":779,"keywords":["data science","Go","machine learning","artificial intelligence","AI","Git","analytics","data governance","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","R","Go","Rust","Git","data governance"],"url":"https:\/\/analyticsindiamag.com\/it-services\/who-should-a-chief-data-officer-report-to\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10058969,"title":"Delivering analytics at scale for the end to end business chain of clients","content":"Tarun Srinivasan, VP, Global Operating Leader, Analytics at Genpact, discussed steps to build cost-saving analytics pipelines for clients during his talk at the fourth edition of the Machine Learning Developers Summit (MLDS) titled,“Delivering analytics at scale for the end to end business chain of clients”. He dilated on the application of conventional analytics tools\/techniques and new age AI\/ML technologies in analytics services like data preparation, business intelligence, forecasting, optimisation and advisory services. “Less than 0.5 percent of the data generated in the world is really ever analysed. We are generating more and more data everyday. I can ensure that we will all have jobs in the analytics space in the future even if we analyse only 0.5 percent of the total data. Because of COVID-19, we have seen accelerated adoption of analytics than ever before. We are seeing a conscious movement from autonomous to autonomy”, Srinivasan said. He outlined the common issues clients raise, including: How can we scale our analytics efforts?We don’t have enough good dataWe have outdated legacy systems which are not talking to each otherWe don’t have enough analytics talentThe  IT organisation does not speak to the business Analytics Continuum At the bottom of the pyramid is the foundational layer. “As we go up, the complexity increases but the value also increases  exponentially. Unfortunately, most of the efforts are happening at the bottom of the pyramid. This restricts data scientists to spend quality time upward,” Srinivasan said. The focus should be on how to use artificial intelligence and machine learning to help in the processes of the foundational layer. Once ML simplifies the bottom layer, data scientists can spend their energy on other priorities like descriptive analytics, predictive modeling etc. Building scalable analytics systems Srinivasan elaborated on the two foundational elements to build scalable systems: Strong governance model: Make sure the analytics end -to-end value chain is in alignment with the company’s vision and a strong analytics engagement model is in place. The model should have a constant interaction mechanism with the end customers. Incubating bilingual skills (data scientists also have expertise in a certain domain such as finance, supply chain, manufacturing etc) is also crucial. Strong execution model: Ensure the right infrastructure such as  automated data pipelines, leading tools for data analytics and preparation is in place. Building the right experience for the customer is also critical. Cost reduction Srinivasan outlined how analytics can be used in the business processes to save money for the clients. Companies possess a huge quantity of spend data from the suppliers. The historical spend data can be crunched to extract business insights. “This is mostly spreadsheet analytics. The power lies in its simplicity.When you analyse the spend data, you get a lot of powerful insights: How the price of the supplier has been varying over a period of time; is there any dependency of the quantity of the parts purchased; dependency of location; dependency of various pricing indices etc. Once you are able to understand the reasons for the price variations, you will be able to predict the price and also use it as a negotiation mechanism with the suppliers.” Contract analytics Most of the clients have hundreds and thousands of contracts, be it with end customers or suppliers. Due to its sheer volume, the contracts do not have good visibility. Not many industrial clients have good contact management technology solutions. The contracts may be in the form of scanned images and pdfs. At Genpact, computer vision OCR technologies are used to read the documents with great accuracy in many languages. NLP techniques are deployed to extract insights from the data. Then, machine learning is used to classify and create a structured database. Diagnostic and prognostics If an aircraft runs into an issue mid-air, a message is relayed to the ground station and usually troubleshooting happens after the aircraft lands. Analytics can improve accuracy of forecasts and reduce downtime. “We observed  that the algorithms and rules used to generate alerts are rudimentary and all design-based. We have used analytics to make sure the rights alerts are generated at the right time when the aircraft is landing. We had a large history of alerts and work done previously to work with to achieve this. It is a classical case of supervised ML as we could bring AI to build algorithms to dramatically improve accuracy,” Srinivasan said. Healthcare innovation Analytics can greatly benefit medical professionals, especially radiologists in times of COVID-19. Due to the pandemic, the stress on radiologists has increased due to the sheer volume of scan requests. Radiologists can leverage analytics  to parse the historical scan data to do the differentials for new scans. The goal is not to eliminate radiologists but to assist them in making informed decisions.","excerpt":"Srinivasan outlined how analytics can be used in the business processes to save money for the clients","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Contract","Data Science","Deep Learning","healthcare analytics","Machine Learning","tools"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-01-22T16:00:00","publication_year":"2022","word_count":789,"keywords":["Go","artificial intelligence","machine learning","AI","ML","Machine Learning","computer vision","healthcare analytics","NLP","RAG","tools","analytics","Deep Learning","Data Science","R","AI (Artificial Intelligence)","Contract"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","computer vision","analytics","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/delivering-analytics-at-scale-for-the-end-to-end-business-chain-of-clients\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10061063,"title":"I never imagined I&#8217;d be Kaggle Grandmaster in a year: Karnika Kapoor","content":"You need a unique blend of technical skills, mathematical expertise, storytelling, and insight to extract meaningful commercial value from data, said Kaggle Notebooks Grandmaster Karnika Kapoor. A BTech graduate from Kurukshetra University, she believes the possibility of the scope of improvement keeps life exciting. Register for the webinar: Managing your career in the age of AI Analytics India Magazine got in touch with Karnika to understand her Kaggle Grandmaster journey and how she made the best of the pandemic-induced lockdowns to upskill herself. Excerpts: AIM: Tell us how you got started with coding. Karnika: I have a background in computer-aided design and drafting and computer-aided engineering. However, in terms of learning a programming language, I started learning a bit of the C syntax in my university while pursuing Btech in mechanical engineering. In my opinion, learning to code is always a good idea for its futuristic relevance. Coding is also a good mental exercise that enhances critical thinking and analytical skills. Coding relies on problem-solving and is quite enjoyable. When I decided to learn machine learning around two years ago, I was torn between R and Python. I picked Python to begin with as it seemed more versatile. However, I am looking forward to learning R as well. I started by exploring several online Python learning platforms. But, my introduction and initial projects in machine learning were in Octave. AIM: How did your first job add value to your career? Karnika: My first job was at an Autodesk value-added reseller as a technical resource for mechanical software. It was a great starting point since it gave me access to various CAE and CAD tools. I was responsible for mechanical CAD suit training. Providing training helped me a lot, as teaching always enhances expertise. The job allowed me to grow and learn a great deal about mechanical design engineering. Working with Autodesk provided first-hand experience with cutting-edge technologies, and I developed a great deal of confidence in my abilities. AIM: What made you interested in Kaggle? Karnika: Andrew Ng’s lectures on machine learning were my primary source of information. That course was conducted on Octave\/MATLAB and was quite useful for understanding the subject matter. I stumbled upon Kaggle and discovered that I had already created an account there a year ago while exploring resources to learn data science. I began my exploration of the platform with the Kaggle courses. It was great to read so many great notebooks. I made it a point to refer to a few notebooks every day; reading notebooks helped me considerably. The Kaggle community is amazing, and the people on the platform share their thoughts and resources with others. AIM: What was your first Kaggle competition like? Karnika: My first competition on Kaggle was an in-class competition. I was a bit daunted at the beginning. I took a stab at the titanic competition and gradually increased my score by a few points but got stuck at 0.777 (score for accuracy on leader board). Beyond this score, all my approaches to increase accuracy were most likely overfitting. That was a year ago. Now I would take a slightly different path for that competition. On top of what I did back then, now I would try out a few more feature engineering techniques. Also, I would consider an Ensemble Model. I hope to explore and try more competitions on Kaggle. The most encouraging aspect of simply trying is that it reminds you that there is still room for improvement. AIM: How did it feel when you became the Kaggle Grandmaster? Karnika: I was a little hesitant when I initially started sharing my work on Kaggle. My first project on Kaggle was a PCOS diagnosis. I never imagined I’d be Notebooks Grandmaster in a year. I was trying to learn something new in each of my projects. For the published notebooks, I kept them simple and clean. I thoroughly enjoyed the entire learning process, and it never seemed like I was “working”. However, in retrospect, I realise I have spent a lot of effort and time acquiring new skills. Because I had so much fun doing it, all the hard work didn’t seem like heavy lifting. During lockdowns, I had time in hand to learn a lot. When I published my 15th notebook, many of my fellow Kagglers congratulated me in advance. AIM: What is your advice for beginners in Kaggle? Karnika: My top recommendation is to go through the Kaggle courses. They are well designed and help in beginning the Kaggle journey. Moreover, it also provides the knowledge to get ahead in the field. Secondly, I’d say focus on learning and mastering your skills. I’d recommend reading other people’s notebooks. I used to go through many notebooks initially, which was quite beneficial. It might be overwhelming at the start. When I started, I wouldn’t fully comprehend some of the notebooks, but as I learned more things, most of it made sense. Lastly, there are many ways of doing the same thing. So, when you look at other people’s work, think about how you would have done it. AIM: What excites you about blockchain and the crypto industry? Karnika: I believe that blockchain technology represents the beginning of a paradigm change in how the world operates. The most exciting thing about blockchain technology is that it has the potential to secure a privacy-driven future. The technology is not limited to crypto and has wide applications in supply chain transparency, smart contracts, and so on. The use-cases range from medical records to real-estate records and art to music. Blockchain makes everything transparent and decentralised. When it comes to the internet of the future, Web 3.0 is a very promising blockchain application. AIM: What makes a good data scientist? Karnika: Enthusiasm to learn about new subjects: Each issue statement or project in data science might come from a completely different field. The interest in learning more about a subject is essential. If one has an innate curiosity, they have the potential to be a good data scientist. Analytical aptitude is necessary: Critical thinking and problem-solving are imperative to understand the business problem. How someone perceives a problem is essential for finding the solution. Data analysis can provide a plethora of information, but identifying what is relevant to the problem at hand is paramount. In addition, the interpretation of data is vital. For example, the association between various features sometimes appears straightforward until you dig a bit further and discover the previous assumption collapse. Good understanding of mathematics is a prerequisite: Understanding maths is essential to understanding how the model works. Linear algebra, statistics, and calculus are the core subjects that drive data science. It is feasible to implement a model without understanding the underlying maths. You can do so by importing libraries and fitting the model to get the result. Nonetheless, this approach does not accommodate a variety of problems that require an insight into the model. Working knowledge of coding and a good grasp of ML algorithms: The value of communication skills is quite significant. A data scientist must have the ability to communicate with various partners. Storytelling is the final and most crucial step of a data-driven project’s pipeline. Adaptability and upskilling are also crucial.","excerpt":"Storytelling is the final and most crucial step of a data-driven project’s pipeline.","categories":["AI Features"],"tags":["Interviews and Discussions","Kaggle Grandmaster","Kaggle grandmasters"],"author_name":"Meeta Ramnani","publish_date":"2022-02-18T18:00:00","publication_year":"2022","word_count":1206,"keywords":["data science","Go","machine learning","AI","ML","RAG","Python","Aim","Kaggle grandmasters","analytics","Kaggle Grandmaster","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","Aim","RAG","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/i-never-imagined-id-be-kaggle-grandmaster-in-a-year-karnika-kapoor\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10110491,"title":"TCS Revenue Growth, AI Might Make Impact on Deals and Jobs","content":"According to the latest Q3 FY 2023-24 results, the IT Giant Tata Consultancy Services (TCS) secured a total of $8.1 billion deal without mega deal support and showing broad-based success. TCS said that four generative AI deals have turned from PoC to contract, but there is no effect on revenue because of it yet. “All deals are progressing as planned,” said Chief Operating Officer, N Ganapathy Subramaniam. He also expressed a cautious outlook on generative AI, stating, “I don’t think generative AI is impacting [our revenue] at this time. Clients are trying to see what benefit it can bring. All are small projects. It’s not leading into our TCV.” When it comes to jobs, “Depending on the overall condition, the investments we make, and driving efficiency, if the headcount has to become less, then be it,” said Milind Kakkad, chief HR Officer saying that if TCS has to reduce the headcount, it might choose to do so depending on the market. On the other hand, he also said that TCS continues its plans to hire 40,000 freshers in 2024 and it stands by the number. He also said that they hired a lot of freshers in this quarter. TCS reported a total employee count of 603,305. The attrition rate for IT services decreased to 13.3% over the last twelve months, down from 14.9% in the previous quarter. The partnership with NVIDIA is progressing as planned, with 14,000 employees trained on leveraging the architecture. The team is also planning to increase the speed of training. A lot of the projects developed by TCS are also being offered as small services to the clients. The company also showcased growth in emerging markets, specifically 23.4% in India. The company also filed 83 patents in the quarter. TCS reported a Q3 marked not much information about generative AI, despite a 4% YoY growth at Rs 60,583 crore in total revenue. Generative AI, although advancing into production with four projects, has not yet demonstrated a substantial impact on the company’s overall revenue. The company says that it is moving into a situation where its clients are comfortable with generative AI, and are trying the technology in different ways. While deals in generative AI are in progress, their timing remains uncertain, and their impact on revenue is yet to be realised. However, the company’s cautious stance on the potential impact of generative AI raises questions about its contribution to future revenue growth, as no commitments have been made in terms of investments. Though the company has previously announced major partnerships with Microsoft and Google, the pace is still slow when it comes to revenue. Read: TCS’ Obsession With Generative AI","excerpt":"“Depending on the overall condition, the investments we make, and driving efficiency, if the headcount has to become less, then be it,” said TCS chief HR officer.","categories":["AI News"],"tags":["TCS"],"author_name":"Mohit Pandey","publish_date":"2024-01-11T19:59:29","publication_year":"2024","word_count":444,"keywords":["Go","programming_languages:R","TCS","AI","programming_languages:Go","RAG","generative AI","GAN","R"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/tcs-revenue-growth-ai-might-make-impact-on-deals-and-jobs\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10048181,"title":"Airbus And IBM Upgrades Their HAL 9000 Inspired Robot","content":"Everywhere we look — from the voice assistant systems in our homes to making digital payments at the local kirana store, technology touches our lives in every step that we take. So, why not allow technology to ease our lives in outer space? Two years ago, Airbus partnered with IBM and German Aerospace Center DLR to explore innovative opportunities and space exploration projects. In September 2019, they launched an artificial intelligence-powered robot in space — CIMON (spelt as Simon). The short for Crew Interactive Mobile Companion, CIMON is a free-floating and sphere-shaped interactive companion robot that accompanies astronauts to space on ISS (International Space Station). Last year, Airbus announced the updated version of the AI-backed astronaut assistant CIMON-2. More recently, Airbus said that CIMON-2 is being loaded with new tasks. <https:\/\/www.youtube.com\/watch?v=FW5vYrseykA> Who’s CIMON? CIMON was developed and built by Airbus with the help of scientists from the Ludwig-Maximilian University to accompany astronauts on space missions. Project CIMON is backed by the German Federal Ministry for Economic Affairs and Energy. CIMON can see, hear, understand and speak. IBM Cloud’s Watson AI technology powers its voice-controlled AI. It is equipped with a stereo camera, a high-resolution camera for face recognition, and two lateral cameras for images and video documentation. Thus, it can be used to perform routine tasks, including documenting experiments, searching for objects, and inventorying. (Source: IBM) CIMON is also loaded with ultrasonic sensors to measure distances for collision detection. It has eight microphones to recognise the direction and a directional microphone for improved speech recognition for ears. CIMON cannot be self-taught, and data has to be essentially fed by humans to train the AI assistant. BIOTESC at the University of Lucerne makes sure that the AI assistant works seamlessly in the Columbus module of ISS by supporting CIMON’s interactions with the astronauts from the ground. To know more about CIMON’s specs and origin, check our previous article here. What’s new? CIMON-2 will be taking on scientific experiments on-board the ISS, accompanying ESA astronaut Matthias Maurer, providing “Educational services from orbit.” Interestingly, Airbus and DLR have signed a contract comprising up to four astronaut missions, working with four humans on the orbital outpost. CIMON-2 will undergo the upgrade while being onboarded, and engineers will first test its software before allowing it to participate in complex experiments. The latest mission will be focusing on the operational and scientific use of CIMON-2. The sphere-shaped AI assistant is receiving new software packages, updated with new safety standards, and fed with new scientific procedures. According to Airbus, the scientific study termed ‘Human interaction with AI and CIMON’ will be focusing on CIMON-2 itself as the research subject. Additionally, CIMON will be upgraded to support routine tasks and documentation of complex scientific tasks. This will be the first time that CIMON-2 will guide a complete experimental procedure. Secondly, the AI-powered astronaut assistant will be providing scientific support for an educational experiment — ‘3D kinetic gas theory.’ The experiment will explain the properties of gases. Post CIMON-2’s upgradation from CIMON, the AI assistant is attuned to astronauts’ emotional states and takes less time to react. Unlike earlier, when it took 10 seconds to react, developers have improved its software architecture, and CIMON-2 takes only two seconds to respond. What does the future hold? After the completion of the four missions, researchers plan to have secured enough datasets to do a justified analysis of CIMON-2’s capabilities. Thus, enabling the AI-powered astronaut assistant to perform more complex tasks in the future. As of now, CIMON-2 has been trained to navigate the European Columbus module of ISS. Moving ahead, if the present experiment is a success, the team of researchers and developers plan to make CIMON independent of the ground-based data centre. While talking about AI assistants in space, one cannot not mention Robonaut — NASA’s companion for astronauts in ISS. Delivered in 2011, the robotic torso was designed to help the crew by holding tools. While having an AI assistant on expeditions to the moon and Mars would be highly acknowledged and be helpful to astronauts, the assistants need to be automated. Waiting to receive instructions from Earth would be inconvenient and time-consuming. The potential to explore outer space with AI is massive, and the future ensures artificial intelligence propelling the astral navigation space.","excerpt":"CIMON-2 will be taking on scientific experiments on-board the ISS to provide educational services from orbit.","categories":["IT Services"],"tags":["AI Assistant","IBM"],"author_name":"Debolina Biswas","publish_date":"2021-09-12T18:00:00","publication_year":"2021","word_count":715,"keywords":["Go","AI assistants","artificial intelligence","ELT","programming_languages:R","AI","AI Assistant","ML","programming_languages:Go","Git","IBM","R"],"extracted_tech_keywords":["AI","artificial intelligence","ML","AI assistants","R","Go","Git","ELT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/airbus-and-ibm-upgrades-their-hal-9000-inspired-robot\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10051294,"title":"[Tech Wars] Zerodha Vs Groww","content":"Nearly two decades ago, trading was next to impossible and time-consuming, as it was done manually from the terminal of brokers and sub-broker offices. Also, opening a Demat account was a whole different ball game altogether. However, cut to the 2020s, investing has become as easy as trading from mobile or desktop, thanks to the rise of new-age investing and discount brokerage platforms like Zerodha, Groww, and others. Today, the Indian markets are driven by FOMO, or fear of missing out, along with easy access to investing – and the bulk of that money is getting routed via new-generation tech platforms and discount brokerage firms. According to SEBI, close to 14.2 million Demat accounts have been opened in FY21. Moreover, because of depleting interest rates (banks) and markets soaring, many retail investors have entered the industry in the last one year. Leading the market is Zerodha. Nearly 20 per cent of all retail trading volume in India is done via its platform. With close to 1.5 million daily active users (DAU) and 4 million monthly active users (MAU), its platform processes close to 10-12 million retail trades per day, generating about 8,000 crores of equity and 15,000 crores of futures and options (F&O) turnover. Next in line is Groww. The company witnessed a 226.12 per cent rise in 2020 – fueled by first-time investors from the age group 18-20 years – compared to a 101.6 per cent increase in 2021. As per the company, Pune, Mumbai, Bengaluru, and New Delhi were the top cities that saw consistent growth over the last two years. Besides Zerodha and Groww, other emerging players include Upstox, ETMoney, Paytm Money, etc. Here is a comparison graph showcasing various investing platforms in India based on monthly visits from users on its website as of September 2021. Zerodha vs others Today, Zerodha competes with a slew of investing platforms, including Groww, ETMoney, Upstox, Paytm Money, ProStocks, 5paisa, etc. But the question is, how is Zerodha managing to stay ahead of the game, despite the growing competition in the space? “We have specific product and business philosophies that we follow, and we have been on a certain trajectory guided by those for over a decade, irrespective of who the other player in the market is — old institutions or new startups — we continue to follow our own trajectory,” said Kailash Nadh, CTO at Zerodha. The company believes in innovating consistently irrespective of competition and focuses on products and services that keep getting better, while never chasing ‘engagement’ metrics or user ‘growth.’ For instance, Zerodha has never advertised or marketed its products and services to date, despite the newer players doing massive ad campaigns, where they are burning investors money, chasing growth metrics, and whatnot. “Our focus has always been on service offerings. We do not have any growth targets or revenue goals,” said Nadh. This can be explained with its latest feature called the ‘Nudge,’ which warns and blocks users of potentially risky trades. In other words, this reduces the number of trades their clients make, thus reducing their revenue. “I am not sure how many other businesses can take the decision to actively discourage their own clients from trading because that would be the right thing to do,” said Nadh. Zerodha = Zero Barriers Compared to other players in the space, Zerodha has the first-mover advantage as it was one of the first discount brokers in the country. Nithin Kamath founded the company in 2010 to overcome the hurdles he faced during his decade-long stint as a trader and create a barrier-free trading experience for investors. Currently, the company offers a gamut of products, including Kite (trading platform), Console (central dashboard), Coin (mutual fund), Kite Connect API, Varsity​​ (knowledge platform), and Sentinel (cloud-based market alert platform), among others. Here’s a complete list of products under the Zerodha umbrella: Zerodha offers all kinds of investment options for its customers, including stocks, direct mutual funds, F&Os, IPO, gift stocks, and fixed income. The team told Analytics India Magazine that the bulk of the activity at Zerodha comes from equity and futures and options trading, followed by mutual fund investments. Its mutual fund platform Coin is one of the biggest online direct mutual fund platforms in India by AUM (asset under management). Soon, the company will be launching its fully revamped web app for Coin. Groww-ing Threat Groww, which was started six years ago by former Flipkart employees Lalit Keshre, Harsh Jain, Neeraj Singh, and Ishan Bansal, offers various investment products such as stocks, direct mutual funds, IPO, exchange trading fund (ETF), fixed deposit (FD), digital gold, and US stocks. With time, Groww is becoming synonymous with mutual funds investing in the country, and in the future, it looks to capture the burgeoning trading ecosystem. Its mutual fund is quite popular among its users – particularly among the millennials, gen-z and younger audience – as it offers a seamless, commission-free direct mutual fund investing platform without any brokerage or subscription fee. Compared to Zerodha, Groww offers ETF, digital gold and US stocks options. However, the only missing services include gift stocks and fixed income. Plus, Groww Calculators are quite different. It provides several tools to assist customers in different ways, including SIP calculator, Lumpsum calculator, SWP calculator, MF return calculator, income tax calculator, FD calculator, etc. Soon, the company is coming up with fixed deposit (FD) services on its mobile app. In the coming years, Groww could emerge as a direct contender to Zerodha. To this, team Zerodha said that their technology\/product offerings are objectively exhaustive and ahead of other players in the market. For instance, no other brokers offer anything like Console, a reporting and analytics platform, said the team. “We built it before the fintech frenzy started because we have always wanted to build it for the benefit of our clients, irrespective of what others offer or don’t. Or the fact that Varsity, our open market education courseware, one of the largest in the world, is freely open and available to everyone without even the need to sign up. We will continue to improve and innovate and go wherever that takes us,” said Nadh, sharing the USP of the company. Bundling vs Unbundling While Groww looks to offer all its services under one roof, Zerodha’s product philosophy is unique. “We decide this on a case-by-case basis,” said Nadh. For example, Zerodha’s Kite, as a trading platform, keeps getting features integrated seamlessly while Coin – the passive investment platform offering MF, bonds, etc. – is an entirely separate app. “Here, the fast-paced nature of a trading platform and the slow dynamic of passive funds, not open Coin every day, but traders open Kite every day multiple times. Console, on the other hand, is a centralised, consolidated reporting and analytics platform that ties everything together in one place. People generally use this after trading and investment activity to analyse and reflect,” explained Nadh. Interestingly, even different products of Zerodha are seamlessly integrated. Kite, for example, opens many contextual Console views from its UI in a single tap. In other words, Zerodha believes in meaningful separation with seamless integration. Tech Evolution – The Beginning Today, investing apps like Zerodha and Groww process hundreds of thousands of requests per second during peak market hours. So, how are they leveraging technology to ensure a smooth and hassle-free experience for millions of users and scaling their platforms? To this, Nadh said that scaling is a collection of practices unique and is the product of an infinite number of variables – the numerous domain-specific problems, the nature of the people involved and their biases, the structure of the organisation, countless engineering decisions and trade-offs, technical debt and history, ad infinitum. He said that several, mostly trivial, common sense techniques and rules of thumb help pluck the low-hanging fruits of typical scaling problems, particularly in the context of running software systems that serve concurrent user traffic and demand. Recalling the early days – between 2013 and 2014 – Nadh said that in the early days of their team, they were busy digitising things inside the organisation, mostly writing Python scripts to automate redundant manual processes\/tasks. “Building a trading platform was not even a thought,” he added. “Sometime in 2014, our order management system (OMS) vendor came out with a white-label web-based trading platform that we could provide to our users. Web-based trading platforms were a tiny niche, and Windows-only desktop trading platforms dominated the industry. When we saw it first hand, we were shocked by what passed for a web application,” said Nadh. Further, he said, “if I remember correctly, its standout feature was that IE6 was no longer a requirement. It was at that moment the team decided to build a usable web-based trading platform. A pivotal, unplanned move that would eventually transform Zerodha into a full-fledged technology firm.” Tech Stack Currently, Nadh likes ‘Go’, the programming language, Redis, and Postgres databases. He believes that these three things are enough to build powerful systems that can attain a huge scale. Here’s a complete picture of all the technology stacks used at Zerodha: When it comes to leveraging machine learning or artificial intelligence tools, Zerodha uses a very minimal AI\/ML. “We use little to no AI or ML apart from some basic image\/document recognition ML models for document processing,” said Nadh. On the other hand, Groww uses AI\/ML in image processing, automating manual workflows, reducing errors and increasing user ease throughout the journey. Besides this, its tech stack includes React Native, Spingboot, and SpringCloud, and a microservices-based architecture that allows it to scale. Here are the details of the tech stacks used at Groww (as shown below): The Struggle Is Real – Transitioning to New Framework At Zerodha, the team believes that product and services improvements happen incrementally every day. For example, around 2017, Zerodha struggled to maintain two codebases for their app on two different platforms – iOS and Android. “We wrote the iOS app in React Native hoping we would have a cross-platform codebase, but it turned out to have severe performance issues and version changes that constantly broke,” shared Nadh, saying that they stumbled upon Flutter, an open-source UI software development kit created by Google. It was a highly experimental alpha technology at the time. “We wrote a quick prototype of Kite in Flutter by learning the Dart language over a few weeks to get a feel for it and, in the end, decided that it was worth ditching the Native and React Native apps and re-writing Kite,” he added. Interestingly, its trading platform Kite is perhaps the first serious app written in Flutter in the world. Last year, Zerodha started experimenting with K8s, or Kubernetes – an open-source system for automating deployment, scaling, and managing containerised applications and tools. Sharing the experience, Nadh said, “We picked K8s to make deployment uniform across our projects. However, K8s turned out to be quite complex in many aspects, including mysterious edge cases (especially under high traffic) such as network packet loss and DNS\/routing issues. Because there are so many abstractions to K8s, it turned out to be hard to deep dive and debug. Ironically, the uniform deployments themselves became complex with layers upon layers of YAML manifests per project.” While the team did manage to move several services to K8s clusters, in the end, it did not turn out to be a smooth experience. Eventually, they had to remove K8s from their stack and are now looking at Nomad (and a related set of tools from Hashicorp) to script and manage their infrastructure. “We’ve run pilots, and it is far easier to understand and handle and seems to be the right fit for our requirements compared to K8s,” said Nadh. A Peek into the Future “We have been working on bringing the few critical external dependencies we have in-house. We are very close to achieving this, and that will make us 100 per cent self-sufficient with our end-to-end broking technology stack with zero external dependencies,” said Nadh. Besides this, team Zerodha said that it would continue to push incremental features and improvements to all their products – Kite, Console, Coin, Varsity, etc. With all the buzz around cryptocurrency globally, the question is, will Zerodha and Groww think about entering into this space or acquiring crypto trading platforms in the near future? “As a heavily regulated entity, we can only consider it if it becomes a fully regulated entity. Personally, I think crypto is rife with dodgy financial schemes just like any unregulated financial scheme would be,” opined Nadh. Final Thoughts Overall, both Zerodha and Groww, without a doubt, are phenomenal investing platforms for all kinds of investments and are ahead of most of the trading and investing platforms. However, if we had to analyse based on technology, accessibility, performance, maintenance, best practices, quality, UI\/UX, Zerodha is a clear winner, not only because of the larger user base but also because of its seamless integration, transparency, and risk control features. Groww, on the other hand, is still in the early days of user acquisition as it added stocks on its platform in the early half of 2020, and the same year, launched digital gold, ETFs, intraday trading, and IPOs. Plus, its mutual fund is top-notch compared to other players in the market, and the recommendation system within the app solves the hassle for many users in choosing the right investment plans.","excerpt":"We use little to no AI or ML apart from some basic image\/document recognition ML models for document processing, says Zerodha CTO.","categories":["Deep Tech"],"tags":["Zerodha"],"author_name":"Amit Naik","publish_date":"2021-10-13T12:00:00","publication_year":"2021","word_count":2236,"keywords":["machine learning","artificial intelligence","AI","ML","RAG","microservices","Python","analytics","Zerodha","kubernetes","Redis"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","kubernetes","microservices","Redis","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/tech-wars-zerodha-vs-groww\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10013891,"title":"Understanding Bioinformatics As A Beginner In Data Science","content":"Though the skillset for Data Science is the same, the implementation varies from problem to problem. That is where domain knowledge comes into play. It has been established that any sector that produces data can be optimized by data science skills to make better business decisions, overcoming challenges and identifying opportunities. Molecular biology is one of the latest fields where data analytics are extensively applied. In this article we will go through a brief introduction of bioinformatics, also referred to as computational biology, from the point of view of a beginner data scientist. What is bioinformatics? As the name indicates – bioinformatics deals with computational analysis of biological data at a molecular level. It is a crossover of biology, computer science, statistics and mathematics which are not the usual disciplines that are studied together. Usually, an expert of one of the specialities decides to pursue bioinformatics which requires them to familiarize themselves with the remaining disciplines. This could be a difficult task; hence this article will assist enthusiasts who have a competent computational and statistical background and are looking to get into bioinformatics. The life sciences contain a plethora of data that need computational tools and frameworks to manage this data and make it more readable and accessible. Bioinformatics provides the said tools and techniques that require a good understanding of the problem’s domain. Now, the question arises that what type of data are we talking about. Though the format of the data is string sequences or numerical expression of gene and proteins, the meaning could vary depending on the source and perturbation of data. These data types will be discussed in detail further in the article. Why do we need quantitative computation in bioinformatics? Significant amounts of research are being carried out to understand the basic human body functions to deduce how the body reacts to perturbations. For the purpose, a cell behaviour of a healthy entity to a perturbed entity is compared to deduce the difference of behaviour that is resourceful in developing drugs to deal with the perturbation. However, the data produced at a cell level is highly dimensional. For instance, one organism’s one cell activity can produce sequences ranging from 450 to 100,00 genes. Hence, to handle such sensitive noisy, high-dimensional data, it is imperative to implement data analysis tools that have been developed in order to find the most optimized way of storing, analysing and computing this data. Types of Data you can come across in bioinformatics? Gene Sequences Most of the data types that one can come across in bioinformatics is nucleic acid sequences – ACGT – namely, Adenine, Cytosine, Guanine and Thymine. These sequences could be for a gene or the whole DNA. They are present in pairs of G-C, T-A, A-T and C-G, hence only one side of the sequence is recorded as the other side can be produced as per their pairing rules. If you are dealing with sequences most of your work will be identifying patterns that are repetitive, recognizing the protein formation pattern in different sequence strips and pinpointing different patterns while comparing two strips of sequences of a healthy cell and a perturbated cell. Figure 1: an example of a DNA Sequence Gene expressions Every human’s biological data is hard encoded in their genes which acts as a guide to how a body will react to any action. There is a surplus amount of information that lies in the genes of an individual yet to be discovered. Gene expressions refers to the messenger RNA levels of a gene at a certain time point and perturbation. Their values are numerical and represent the so-called expression of a gene at a certain time point. It has been biologically proven that in a set of gene’s at a particular location, there are few gene’s that are referred to as “regulatory genes” and the remaining gene’s are referred to as “target genes”. The regulatory genes can be labelled as the supervisors that control the expressions of a target gene. For instance, if X ???? Y, that means X gene regulates Y gene. Figure 2: An example of a Gene Regulatory networks A lot of research is being carried out to find these regulatory and target relationships between genes. Gene expression data suffers from high dimensionality issue also referred to as “curse of dimensionality” that means the data points to data features ratio is very small as there are thousands of genes and their respective expressions however, time points recording still falls between 10-30 time points. If you are working with Gene expression data, you will be spending time mostly in representation models of gene regulatory networks, optimizing these models and dealing with computational complexity. Popular data science tools and databases for bioinformatics? Databases for bioinformatics GenBank: Genetic sequence database from NCBI EMBL-EBI: Nucleotide Sequence DatabaseUniProt: Protein sequence databaseGEO Database: Gene expression profiles from NCBIExpression Atlas: Gene expression across species and biological conditions The top three tools\/programming languages used by computation biologist are: Python: BioPython, Biotite, Scikit-Bio, SciPyR: CROME, InterMineR, rScudo, RepoMatlab: Bioinformatics Toolbox Though it fairly depends on an individual’s background to which tool they prefer to adopt, Matlab does have a better edge for visualization. Where to begin? Now, that we have the basics laid out let’s discuss the ideal way to address a bioinformatic project to begin will. Step1: Identify the datatype and the problem definition related to the data type Step2: Research about the biological inference underlining the datatype to improve your domain knowledge Step3: Data preparation – Identify the database to be used along with required data points or data features. It is advisable to start with small datasets such as a 5-gene IRMA network. Step4: Lay down the analysis solution in pseudocode to ensure you understand the problem statement and its working Step5: Code your analysis, compare your results to the ground truth and infer your outcome Bioinformatics could be challenging to any research with a non-medical background, hence, jumping to solutions without appropriate understanding of the background problem will significantly enhance the complexity of the analysis. In this article, we scratched the surface of bioinformatics with a point of view of an intermediate data scientist in order to lay a good foundation for those who have taken on the endeavour to pursue computational biology with a non-medical background.","excerpt":"Though the skillset for Data Science is the same, the implementation varies from problem to problem. That is where domain knowledge comes into play. It has been established that any sector that produces data can be optimized by data science skills to make better business decisions, overcoming challenges and identifying opportunities. Molecular biology is one […]","categories":["AI Features"],"tags":["Data Science","data visualization python","Types of Databases"],"author_name":"Jaskaran Kaur","publish_date":"2020-12-10T19:00:00","publication_year":"2020","word_count":1052,"keywords":["data visualization python","data science","Go","programming_languages:R","AI","Types of Databases","Python","programming_languages:Python","ViT","analytics","GAN","Data Science","R"],"extracted_tech_keywords":["AI","data science","analytics","Python","R","Go","GAN","ViT","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/understanding-bioinformatics-as-a-beginner-in-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10143978,"title":"Vultr Achieves $3.5 Billion Valuation","content":"JJ Kardwell, CEO of Vultr, announced on LinkedIn that the company, a leading independent cloud infrastructure provider, has achieved a valuation of $3.5 billion after securing funding from LuminArx Capital Management and AMD. The funding highlights the rising demand for AI infrastructure and marks a major milestone for the company. Founded in 2014 by David Aninowsky, Vultr has grown solely on operating cash flow and has become a major player in cloud computing. “This new financing accelerates our growth in AI infrastructure and cloud computing as we build the category-defining independent cloud infrastructure company,” he said. The new funding will enable Vultr to expand its AI infrastructure and cloud computing offerings, making it a strong alternative to traditional hyperscalers. Vultr’s mission is to deliver high-performance cloud services that are affordable, accessible, and easy to use. Its offerings include cloud compute, cloud GPU, bare metal, and cloud storage, and it serves hundreds of thousands of active customers across 185 countries. In related developments, CoreWeave, a specialised cloud provider focused on GPU-powered workloads, secured $2.3 billion in financing last year. Magnetar Capital and Blackstone Tactical Opportunities led the funding, with contributions from BlackRock, Pacific Investment Management Company (PIMCO), Coatue, and Carlyle. The financing supports CoreWeave’s plans to increase computing capacity, open new data centres, and expand its workforce to meet the growing demand for AI applications like LLMs and generative AI. Meanwhile, AMD has enhanced its efforts to close the GPU performance gap with NVIDIA. AMD CEO Lisa Su recently announced the MI325 advanced GPU accelerator, calling it a “game-changing product” set to rival NVIDIA’s H200. Meanwhile, the MI350 series aims to compete with NVIDIA’s Blackwell GPUs. Expressing optimism, Su stated that AMD expects to close the performance gap faster than anticipated with its new hardware. Andrew Dieckmann, CVP and GM of data centre GPU at AMD, also commented on the company’s progress. “We are trying to take very representative benchmarks that are realistic. I can tell you that in our customer engagements, especially regarding inference workloads, we have yet to find a single workload on which we cannot outperform NVIDIA,” he said. He, however, acknowledged that AMD doesn’t always outperform its rivals. As competition intensifies in the AI and cloud infrastructure market, Vultr, CoreWeave, and AMD are working to meet the growing demands of enterprises and innovators worldwide.","excerpt":"The new funding will enable Vultr to expand its AI infrastructure and cloud computing offerings, making it a strong alternative to traditional hyperscalers.","categories":["AI News"],"tags":["AMD","GPU","NVIDIA"],"author_name":"Shalini Mondal","publish_date":"2024-12-19T16:43:54","publication_year":"2024","word_count":387,"keywords":["Go","AMD","API","funding","programming_languages:R","AI","cloud computing","RAG","Aim","generative AI","NVIDIA","R","GPU"],"extracted_tech_keywords":["AI","generative AI","Aim","RAG","cloud computing","R","Go","API","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vultr-achieves-3-5-billion-valuation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":51038,"title":"Big Data Analytics And Future Managers","content":"Big Data Analytics is a buzz word today and the skills in its related technologies such as artificial intelligence, machine learning, cloud computing and robotics for automation are driving the industry’s growth. As per the recent report, the IT companies are laying off their employees in thousands and at the same time looking for these skills. It says, “After thousands were fired by Infosys and Cognisant in the past six months, as many as 200,000 more may be at risk of losing their jobs in the next year as companies seek people with new skillsets, according to recruitment experts.” Two decades ago the people who face layoffs today had skills which were similarly valued. I lived through the IT revolution that began in the mid-1990s with the spread of the Internet into business and consequent changes in the way the business was done. In all the transitions that took place in these decades what survived is people who knew how to use technologies for business benefit. There appears to be an uncanny parallel between what happened to the IT and what is happening to the Analytics today. Today Analytics skills are highly valued simply because they are in short supply. But will they remain the same in the coming years? The clichéd answer to such questions is to reskill. Of course, one needs to keep pace with rapid changes taking place in technological space, but it is not practicable to be competitive in changing technological skills. It is, however, possible to be competitive in using technology for the business benefit. If one is located in the business world, this is precisely the managerial skill that will always stand one in good demand. Today, the typical way in which companies are organized to use analytics may be depicted by silos comprising Strategy, Data Science or analytics team and User Department teams. Many companies have a small in-house Analytics team representing these functions. Today, they are silos because there is a communication disconnect between them, each speaking its own language. The strategy typically conceives an analytics problem ideally in consultation with the user department and proposes it to the analytics team or data science team. The latter would take stock of data within the company and outside, formulate variables and try modelling the problem which, in process, undergoes a change from the original. There could be several such iterations and corresponding changes between the analytics team and strategy before it is cleared for execution. When it is executed and the solution is offered to the user department, it could be completely irrelevant and unusable, adding itself to worrying statistics of failures of analytics projects. It is, therefore, that Gartner reported in November 2017, that 60% of big data projects failed. A year later, Gartner analyst Nick Heudecker‏ said his company was “too conservative” with its 60% estimate and put the failure rate at closer to 85%. Today, he says nothing has changed. Gartner isn’t alone in that assessment. Reports of failure of analytics project is a legion: In July 2019, VentureBeat AI reported that 87% of data science projects never make it into production, in January 2019, NewVantage survey reported that 77% of “business adoption” of big data and AI initiatives continued to represent a big challenge for business, (which meant three-fourth of the software being built is apparently collecting dust), in January 2019, Gartner also came out saying  80% of analytics insights will not deliver business outcomes through 2022 and 80% of AI projects will “remain alchemy, run by wizards” through 2020, in November 2017, CIO.com lists seven sure-fire ways to fail at analytics. Tom Davenport, a senior advisor at Deloitte Analytics stated, “The biggest problem in the analysis process is having no idea what you are looking for in the data.” In May 2017, Cisco reported that only 26% of survey respondents were successful with IoT initiatives, which meant 74% of them just failed. This takes me two decades back when we used to read exactly similar things about the IT projects. When I ventured in to implement ERP in my company, one of the biggest projects in the Asia Pacific the frightful dictum was in the air that over 80% ERP projects fail to deliver business benefits. We averted this aftermath by being extra conscious about the business objectives at every point and in the course even realised that it was not technology alone but adequate technology insight and lots of business acumen that is what is required in any manager. The same could be said of the analytics, which of course has become big data analytics. Analytics reflects the trend towards automation; embedding solution into business processes. The challenge would shift from technology or analytics skill to essentially managerial skill which embeds knowledge of what various technologies are capable of, and which one would best serve one’s business strategy. People who developed these skills have risen to head businesses and are not the potential victims of layoffs. The future manager thus will have to be well-grounded in business, having sharp strategic acumen and knowing ‘enough’ of analytics. This ‘enough’ can never be specified enough. The possible approach is to know the analytics technologies in vogue thoroughly enough by working them in hands-on mode. It should necessarily include Stats and Maths as well as coding in R, Python and SAS as the foundation. It should also include ETL technologies, visualization tools, predictive and prescriptive analytics, AI and Machine Learning, Deep Learning, Cloud computing and security, IoT Analytics. In management domain, the approach should be to build strategic perspective with the help of capsuled courses in most functional areas—Economic, Accounting, Marketing, Finance, Operations, OB\/HR—and integrative capstone course capped with an opportunity to work long enough as an intern in a company. It should be so designed as to stress the problem areas so as to generate use cases for analytics. The managers should also be formally trained in strategy and equipped with supplementary skills of design thinking, storytelling with data, data strategy, analytics project management and ethical aspects in managing data, etc. Depending upon their choice of industry, they should also have an opportunity of specialization in certain areas. The biggest anxiety of placement is taken care in this process of ensuring the output quality. This is precisely what we offer in our Big Data Programme. This kind of integrative training should make future managers break prevailing silos and insure analytics projects from failure as well as themselves from the risks of layoffs. To apply for PGDM-Big Data Analytics programme at Goa Institute of Management, click here.","excerpt":"Big Data Analytics is a buzz word today and the skills in its related technologies such as artificial intelligence, machine learning, cloud computing and robotics for automation are driving the industry’s growth. As per the recent report, the IT companies are laying off their employees in thousands and at the same time looking for these […]","categories":["AI Trends"],"tags":["Big Data","Big Data Analytics","Data Analytics","is artificial intelligence the next big thing","what is big data"],"author_name":"Anand Teltumbde","publish_date":"2019-12-03T12:36:55","publication_year":"2019","word_count":1097,"keywords":["data science","artificial intelligence","is artificial intelligence the next big thing","what is big data","AI","machine learning","cloud computing","TPU","Big Data Analytics","Python","deep learning","analytics","Data Analytics","Big Data","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","cloud computing","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-role-of-big-data-analytics-in-the-future-of-managers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10171523,"title":"Wayve, Uber Begin Level 4 Autonomy Trials in the UK","content":"Autonomous car driving technology firm Wayve and Uber have announced the launch of public-road trials for fully autonomous Level 4 (L4) vehicles in London, marking the first such initiative in the UK. This development positions the UK as Uber’s largest market for piloting autonomous vehicles to date. The announcement follows a new framework from the UK secretary of state for transport, which accelerates permissions for commercial self-driving pilots. Wayve and Uber will work with the UK government and Transport for London to secure regulatory approvals before deployment. “This is a defining moment for UK autonomy,” said Alex Kendall, CEO and co-founder of Wayve. “With Uber and a global OEM partner, we’re preparing to put our AI Driver technology into real service on the streets of London.” The partnership will combine Wayve’s Embodied AI platform with Uber’s ride-hailing network. Wayve’s AV2.0 technology, which does not rely on HD maps or geofencing, allows its AI to adapt to new environments without significant pre-programming. Andrew Macdonald, president and COO of Uber, said, “We’re excited to take the next step in our journey with Wayve, bringing autonomous mobility to one of the world’s busiest and most complex urban environments.” Trials are expected to begin by spring 2026. The UK transport secretary, Heidi Alexander, said the agreement is “a fantastic vote of confidence in this new technology”, adding that the trials could create 38,000 jobs and contribute £42 billion to the economy. The UK’s unique driving conditions, including varying road layouts and traffic laws, offer a contrasting environment to the US, where most Level 4 testing has occurred. Insights gained from the UK are expected to help improve the global scalability of autonomous systems. This initiative builds on a 2024 collaboration between Wayve and Uber to integrate Wayve’s AI into Uber-operated vehicles. The new phase will move from development to live operations, with further details on the trial and the OEM partner expected in the coming months. Wayve recently showcased its technology by deploying a single AI model across 90 cities worldwide during a 90-day roadshow. “We drove through Tokyo, Milan, and Montana—all with the same AI model,” Kendall said. “That’s the power of AV2.0.” Both companies stated their commitment to advancing safe and efficient autonomous mobility while strengthening the UK’s position in global transportation innovation.","excerpt":"UK becomes the largest market for Uber’s public-road autonomous vehicle testing.","categories":["AI News"],"tags":["Uber"],"author_name":"Siddharth Jindal","publish_date":"2025-06-10T14:46:43","publication_year":"2025","word_count":381,"keywords":["Go","AWS","AI","cloud_platforms:AWS","innovation","programming_languages:R","programming_languages:Go","Scala","programming_languages:Scala","Uber","R"],"extracted_tech_keywords":["AI","AWS","R","Go","Scala","innovation","cloud_platforms:AWS","programming_languages:R","programming_languages:Scala","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/wayve-uber-begin-level-4-autonomy-trials-in-the-uk\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":29709,"title":"Twitter’s Adoption Of Tensorflow Might Have Improved Its User Experience Dramatically","content":"As per the Stack Overflow Developer Survey, TensorFlow was the most wanted framework library, garnering 73.5 percent votes. It is one of the fastest growing and most popular framework on Stack Overflow. TensorFlow is already the most popular machine learning tool amongst leading tech giants like eBay, Google and Uber that adopted and embraced TensorFlow for a long time now. And the latest to join the bandwagon is Twitter that recently moved to TensorFlow as its machine learning framework, in July this year. Twitter has been using its trademark internal system called DeepBird which before the migration to TensorFlow used LuaTorch for its operations. But since LuaTorch moved to PyTorch in December last year, the support for DeepBird was automatically restricted. So Twitter finally adopted TensorFlow as its machine learning framework. Ml Use Cases @Twitter Twitter just like any other platform of its kind, uses machine learning for the following different aspects: Ads: To display ads according to the user’s interest. Timelines: Providing with interesting relevant context of timelines to the users. Abuse: The information reached should be safe for work and healthy for the platform. Recommendations: Provides recommendations to tweets based on the user interest. Twitter deals with petabytes of data, mostly tweets every single day are many and each tweet has to be addressed in very few seconds before being rendered on the timeline. Thus, the models besides being accurate in prediction game, should also be supersonic. This is why ML is always required in a platform like Twitter. How TensorFlow Helped Twitter Lend Relevance & Decode Data Records Python is the most popular language preferred by data professionals. TensorFlow supports Python and that provides a great deal of flexibility in leveraging machine learning. At Twitter, ML engineers can now better track the machine learning models and make changes with the inclusion of TensorBoard. The micro-blogging site shows tweets to Twitterati based on their relevance and TensorFlow integrates very well with different parts of Twitter platform. Also, with the new platform the process of decoding data records has become seamless. With the adoption of TensorFlow, Twitter engineers are using an internal tool called Model Repo, who works towards reducing friction in developing, deploying, maintaining and refining models. Model Repo tool. Image source: Twitter Twitter, like other big tech firms relies on open source tools and platforms. The social media giant leverages Finagle over Thrift and that has worked better for their use cases. They have used Hadoop for training. Their data records are able to store different kinds of features in machine learning, for example binary features. Their build_graph function is used in three modes: training mode, evaluation mode and prediction mode. Here’s how adopting TensorFlow has helped Twitter in the following aspects: Higher engineer productivity: Using TensorFlow along with Twitter’s internal visualisation tools like Model Repo, convergence of models can be observed and adjusted easily. Easier ML access: DeepBird v2 provides simplified trainers and easy integration with Twitter’s technology stack, making it easier for engineers to experiment with ML. Better inference performance: v2’s performance is better than its predecessors. Improvement in model metrics: The use of DeepBird v2 on TensorFlow has improved the ML models and made them more robust. Three modes: Training, Evaluation, Prediction. Image source: Twitter. LuaTorch Vs TensorFlow 1. Code debugging: Their model was a combination of YAML Lua Torch and C at the backend. Why TensorFLow is better because the models can be easily tracked and they can easily transport the Python code into different ML models. Therefore it’s easier to debug your code. 2. Unitise models: Benefits of using TensorFlow as opposed to YAML is that it is easier to unitise models. It cannot be done without going to the backend code, in case of LuaTorch. Whereas, with TensorFlow, it is easier to write and unitise models. 3. Access to a specific code: With TensorBoard it is very easy for ML engineers to actually see how the data models are performing overtime. It was difficult in YAML to use a specific part of code of interest and reuse that. TensorFlow allows to do that. Parts of the model can be reused. It is difficult to debug the whole program  and add test cases, which had to be done in LuaTorch. TensorFlow can tell the output of each layer. 4. Tools: LuaTorch had not many visualisation tools. The engineers had to create a lot of them by themselves. TensorFflow, on the other hand, has TensorBoard, TensorFlow Model Analysis (TFMA), TensorFlow Debugger (TFDebugger) which makes work more easy. 5. Code modification: It was not easy to modify code in Torch. It is relatively easy to debug in TensorFlow. Future Work Twitter engineers believe that DeepBird v2 is the future of ML at Twitter. They are working to support models on this platform of v2 and train on GPU clusters. They are also working towards adding support to online as well as distributed training in v2. Although Twitter faced several roadblocks before switching to TensorFlow, it has the power to unlock ML and create models for a better Twitter experience. TensorFlow is undoubtedly a very good way to provide a refuge to ML and the next thing would be to see who next relocates to this popular platform.","excerpt":"As per the Stack Overflow Developer Survey, TensorFlow was the most wanted framework library, garnering 73.5 percent votes. It is one of the fastest growing and most popular framework on Stack Overflow. TensorFlow is already the most popular machine learning tool amongst leading tech giants like eBay, Google and Uber that adopted and embraced TensorFlow […]","categories":["AI Features"],"tags":["ML","Python","Stack Overflow"],"author_name":"Disha Misal","publish_date":"2018-10-29T10:09:53","publication_year":"2018","word_count":873,"keywords":["Go","machine learning","TPU","AI","PyTorch","ML","Stack Overflow","RAG","Python","TensorFlow","R"],"extracted_tech_keywords":["AI","machine learning","ML","TensorFlow","PyTorch","RAG","TPU","Python","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/twitters-adoption-of-tensorflow-might-have-improved-its-user-experience-dramatically\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10169256,"title":"This YC-Backed Startup Just Made Finding Blue Collar Jobs in India ‘300% Easier’","content":"In a country where millions of blue-collar workers remain disconnected from formal job markets, one of India’s most urgent challenges is making employment accessible, equitable, and efficient at scale. Vahan AI, a YC-backed AI recruitment startup, announced that it is using OpenAI’s technology to improve how blue-collar workers are hired across India. The company, founded in 2016 by Madhav Krishna, aims to address hiring challenges for blue-collar workers in India using AI. The platform matches job seekers with relevant roles and helps employers find suitable candidates. It has placed at least five lakh workers in more than 480 cities and works with companies including Zomato, Swiggy, Flipkart, Zepto, Blinkit, Amazon, Rapido, and Uber. “We’re finding that even with a fairly early version of this AI recruiter, we’re able to increase human recruiter productivity by 300%,” Krishna said in an interaction with AIM regarding the efficiency of this. Bridging Employers and Blue-Collar Workers Vahan AI is reshaping how employers like Zomato and Swiggy connect with India’s vast informal workforce. The platform doesn’t seek to replace human recruiters but to amplify their effectiveness using smart automation. “So what we’ve built, the AI recruiter, is not a replacement for human recruiters,” Krishna added. Instead, the AI recruiter acts as a digital assistant—qualifying candidates, answering their questions, and even handling document verification. This has had a transformational impact on recruitment agencies using the platform. “We’re finding that even with a fairly early version of this AI recruiter, we’re able to increase human recruiter productivity by 300%.” While the AI recruiter currently supports English and Hindi, the challenge of India’s linguistic diversity remains. “In Hindi alone, the word yes is said in over 20 ways,” Krishna pointed out, highlighting how this feature would be particularly valuable in a linguistically diverse country like India, where overcoming language barriers is crucial to enabling millions to access employment opportunities with greater ease and confidence. To address this, Vahan is expanding support to eight more Indian languages, working closely with OpenAI and Eleven Labs to ensure inclusivity across dialects and regions. Moreover, another facet of Vahan’s success is its network of local entrepreneurial partners, called Vahan Leaders (VLs). These are small recruitment agencies operating in smaller towns, often overlooked by employers across the country. According to the company’s official statement, over the past 15 months, some VLs in India have grown their operations by up to 10 times. In Bengaluru, four VLs now manage to place over 2,000 workers each month. In the NCR region, placements have increased to 3,500 per month in three months. Notably, Mumbai has also recorded a rise in placements through the platform. These local leaders use Vahan’s tools to help place workers in jobs while maintaining the trust and cultural context needed for success in rural India. In a previous interview with AIM, Krishna stated that while global markets like Southeast Asia, the Middle East, and potentially South America are of definite interest to the company, India remains a massive opportunity. Thus, Vahan.ai intends to stay focused on the Indian market for the next few years. “What we’re seeing on the ground is remarkable: people who were previously excluded from formal employment channels are now finding jobs in hours, not months,” he added. OpenAI’s Expanding Role in India India represents both a massive user base and a growing developer hub for AI labs like OpenAI, which is ramping up its presence in the country. This is why OpenAI has increasingly been collaborating with Indian startups to bring real-world AI solutions to scale, and Vahan is a case in point. “India is strategically important…we are the second largest base of developers in the world,” Krishna added. “There’s a very tight startups team at OpenAI…who help bridge the gap between what startups are trying to build in terms of use cases and their platform.” In addition to this, Vahan is also part of the influential Y Combinator network, which has been pivotal in shaping India’s tech startup landscape. Vahan.ai began as an upskilling platform but pivoted in 2019 after struggling to grow at scale. Krishna noted that skilling felt optional, while job placement addressed a more urgent need. The company joined Y Combinator in 2019 and raised a Series A in 2021. This was followed by a $10 million in Series B funding from Khosla Ventures, with support from Y Combinator and Gaingels LLC. It is the third Indian AI startup to receive such backing from Khosla, after Sarvam AI and upliance.ai. Krishna declined to comment on future funding plans.","excerpt":"Vahan is working closely with OpenAI and Eleven Labs to ensure inclusivity across dialects and regions.","categories":["AI Startups"],"tags":["AI (Artificial Intelligence)","Startups","vahan ai"],"author_name":"Aditi Suresh","publish_date":"2025-05-06T20:43:07","publication_year":"2025","word_count":754,"keywords":["API","ELT","vahan ai","OpenAI","AI","Git","Aim","ViT","Rust","GAN","Startups","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Rust","Git","API","ELT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/this-yc-backed-startup-just-made-finding-blue-collar-jobs-in-india-300-easier\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10045505,"title":"8 React Native Component Libraries Every Mobile Developer Should Know","content":"React Native or RN is a popular open-source JavaScript-based mobile application framework. It allows developers to build natively rendered mobile applications for both Android and iOS operating systems. Components are essentially like JavaScript functions. React Native accepts input in the form of props and returns React elements on the screen. Essentially there are two types of components– native component and core component. Developers building for Android write views on Kotlin or Java and use Swift or Objective-C for iOS development. However, with React Native, these views can be called for using JavaScript using React components. During the programme’s run time, React Native creates the Android or iOS views for these components. Backed by the same views as iOS and Android, these React Native applications, thus built, feel, look and perform like any other application. And these platform-backed components are called Native Components. Originally, React Native came with a set of ready-to-use components, enabling developers to start building their applications immediately. Today, we list the top React Native Component Libraries for 2021: React Native Maps The React Native Map component for iOS and Android offers customisable map components. These provide intuitive and react-like APIs for declaratively controlling map features. Features include– changing map view region, overlaying elements on the map, customising markers and map style. For installation, use: npm install react-native-maps –save-exact For more information, click here. NativeBase This mobile-first accessible component library is built to provide a customisable design system across iOS, Android and web for React and React Native. In addition, it can be integrated with React ARIA and React Native ARIA. NativeBase is powered by Styled System to help developers build custom UI components– 40 of them, including button, checkbox, flex, and stack. In addition, it is highly themeable, allowing customisation of app themes and component styles. For more information, click here. React Native Camera The Camera component for React Native supports photographs, videos, face detection (for both Android and iOS), barcode scanning and text recognition. This allows developers to communicate with a device camera without having to worry about the native code. For more information, click here. React Native Paper React Native Paper offers components with Material design standards for React Native– Android and iOS (following platform adaptation guidelines). Additionally, it follows material design guidelines and theming support. It is an open-source project. For more information, click here. React Native Bit Bit is a tool for component-driven application development. It is an extensible toolchain for systems that are faster to develop, simpler to understand, test, and maintain, and easier to collaborate on. Bit’s USP lies in allowing developers to build components outside an application and then use them to compose as many applications as they want. Components can be added or removed from applications to alter their functionality. To install, use: bvm install For more information, click here. React Native Ignite Ignite is a popular React Native boilerplate. It is a collection of all of Infinite Red’s opinions on patterns, packages and stack in one place. Ignite apps include React Native, React Navigation 5, MobX-React-Lite, AsyncStorage, TypeScript, MobX-State-Tree and is Flipper-ready. According to the official website, developers save as much as two to four weeks of their time on their React Native project by using Ignite. For more information, click here. React Native Lottie A picture is worth 1,000 words. And making this possible for mobile applications is Lottie. The Lottie component for React Native is a library for Android, iOS, Web and Windows that parses Adobe After Effects animations, exported as JSON, with Bodymovin and renders them on mobile and the web. Thus, allowing designers to create and ship animations without engineers recreating them by hand. For installation, use: lottie-react-native For more information, click here. React Native Navigation As the name suggests, React Native Navigation is a native navigation solution for Expo and React Native apps for Android and iOS. Navigation is a cross-platform JavaScript API, and requires iOS 11 and Android 5.0 and above. Navigation is platform-specific and customisable. Moreover, it is extensible at every layer. That is, developers can write their navigators or replace the user-facing API. To install, use: react-native-navigation For more information, click here.","excerpt":"React Native an open-source JavaScript-based mobile application framework for both Android and iOS operating systems.","categories":["AI Trends"],"tags":["Javascript","react-native"],"author_name":"Debolina Biswas","publish_date":"2021-08-07T18:00:00","publication_year":"2021","word_count":692,"keywords":["API","programming_languages:R","AI","Javascript","programming_languages:Java","TypeScript","RAG","react-native","JavaScript","R","Java","programming_languages:JavaScript"],"extracted_tech_keywords":["AI","RAG","R","JavaScript","TypeScript","Java","API","programming_languages:R","programming_languages:JavaScript","programming_languages:Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-react-native-component-libraries-every-mobile-developer-should-know\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":59509,"title":"What Are Lagrangian Neural Networks","content":"Neural networks can perform well on tasks such as image classification, language translation, and game playing. However, they usually fail to perform well in tasks that need human abstraction. Activities such as catching balls mid-air or juggling multiple balls, which the humans have mastered, need an intuitive understanding of dynamics of how physical bodies behave. We don’t take time out to calculate the trajectories before hitting the ball. We just know. Machine learning models lack many basic intuitions about the dynamics of the physical world. A neural network may never grasp human-level abstraction, even after seeing thousands of examples. The basic problem with neural network models is that they struggle to learn basic symmetries and conservation laws One solution to this problem is to design neural networks that can learn arbitrary conservation laws. Sam Greydanus, a resident at Google Brain, along with his peers, Miles Cranmer, Peter Battaglia, David Spergel, Shirley Ho and Stephen Hoyer has published an interesting paper addressing the above challenges. In the next section, we shall talk briefly about Lagrangian neural networks and why they are significant Overview Of Lagrangian Neural Networks A double pendulum To demonstrate how a Lagrangian denotes the underlying secrets of nature, the authors take the example of a double pendulum. A double pendulum is a pendulum attached to a pendulum. It is famous for representing chaos in moving bodies. Consider a physical system that has coordinates x_t=(q,q˙). For example, a double pendulum can be defined with the help of the angles it makes and also its angular velocities. When a double pendulum is set into motion, the coordinates that we have ascribed on the body move from one point to another. This is clear. This is common sense. But what induces chaos is the possibility of many paths that these coordinates can take between start(x_0) and end(x_1). According to Lagrangian mechanics, any action can be written as a function of the potential energy (energy by virtue of position) and kinetic energy (by virtue of motion). And, it is expressed as follows: The S in the above expression has a remarkable property. For all possible paths between x_0 and x_1, there is only one path, which provides a stationary value of S. Besides, that path is the one that nature always takes — the principle of least action. So, if we understand the laws of conservation of energy, then we can define the action of the body under consideration and eventually predict the path taken by the body with great accuracy. However, neural networks are poor at understanding loss of energy concepts, and this is where Lagrangian Neural Networks (LNNs) comes into the picture. via paper by Sam Greydanus et al., “If energy is conserved,” they might say, “when I throw a ball upwards, it will return to my hand with the same speed as when it left.”via Sam Greydanus’ blog But these common-sense rules can be difficult to learn straight from data. So, a neural network that understands the conservation of energy just from data can have great implications for robotics and reinforcement learning. And, with LNNs, the authors have opened up new avenues of research by binding the underlying principles of nature with man-made artificial neural networks. What Is The Significance Of LNNs A French stamp commemorating Lagrange Joseph-Louis Lagrange, born in 1736, in Italy, is one of the greatest polymaths who have ever lived. By the age 20, he already has published the principle of least action in the dynamics of solid and fluid bodies. He formulated how to find the path taken by bodies no matter how chaotic their movements are. Three centuries later, today, his work still resonates in the most advanced fields such as deep learning. Not that there is an expiry date to the fundamentals of mathematics but the fact that those principles can be applied to make neural networks better than before is a fascinating idea. The models of the universe are probed for their symmetries while formulating a model. These symmetries correspond to laws of conservation of energy and momentum, which the neural network struggle at learning. To address these shortcomings, Greydanus and his peers, introduced a class of models called Hamiltonian Neural Networks (HNNs) last year, which can learn these invariant quantities directly from (pixel) data. Now, going a step ahead of HNNs, Lagrangian Neural Networks (LNNs) can learn Lagrangian functions straight from data. This can also be done by HNNs, but unlike HNNs they don’t require canonical coordinates. The above table pits LNNs against other related work, and we can clearly see that Lagrangian Neural Networks are winning and we have to wait and see what other breakthroughs they have in store for the scientific community. Know more about Lagrangian Neural Networks here.","excerpt":"Neural networks can perform well on tasks such as image classification, language translation, and game playing. However, they usually fail to perform well in tasks that need human abstraction. Activities such as catching balls mid-air or juggling multiple balls, which the humans have mastered, need an intuitive understanding of dynamics of how physical bodies behave. […]","categories":["Deep Tech"],"tags":["Neural Networks"],"author_name":"Ram Sagar","publish_date":"2020-03-23T19:00:37","publication_year":"2020","word_count":792,"keywords":["Go","machine learning","AWS","AI","neural network","cloud_platforms:AWS","programming_languages:R","deep learning","ViT","R","Neural Networks"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","AWS","R","Go","ViT","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/what-are-lagrangian-neural-networks-hamiltonian-physics\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10097297,"title":"Meta-Qualcomm Partnership will Bring Llama 2 to the Masses","content":"The conversation around the use cases of LLMs has been rapidly growing, supercharged by the release of Llama 2, Meta’s new open source model. Even as Meta’s Llama has been released to the public, there is still a huge barrier for entry in terms of being able to run it on local hardware. To remedy this issue and truly open up the power of Llama 2 to everyone, Meta has partnered with Qualcomm to enable the chipmaker to optimise the model for running on-device, powered by the chips’ AI capabilities. Industry experts are predicting that open LLMs could create a new generation of AI-powered content generation, smart assistants, productivity applications, and more. Adding the capability to natively run LLMs on-device is sure to create a strong new ecosystem of AI-powered applications, akin to the app store explosion that happened with iPhones. This move will not only democratise access to the model, but also unlock a bunch of possibilities for on-device AI processing. It also comes at a time where the consumer hardware and software industries are waking up to the possibility of AI capabilities at the edge. First spearheaded by Apple’s inclusion of a neural engine in the M1 chip, the addition of a new type of processor on personal computers will finally give developers the tools to create truly democratic AI. What the Qualcomm-Meta partnership entails For context, Qualcomm is currently creating a new set of AI-enabled chips under the Snapdragon platform. Using what it calls the Hexagon processor, the chipmaker equips its chips with various AI capabilities. Using an approach called micro tile inferencing, Qualcomm is able to integrate tensor cores, dedicated processing for SegNet, scalar, and vector workloads into an AI processor, which is then integrated into a Snapdragon mobile chip. As part of its partnership with Meta, Qualcomm will make Llama 2 implementations available on-device, harnessing the capabilities of the new AI-enabled Snapdragon chips. Since the model will be running on-device, developers can not only cut down on cloud computing costs for their applications, but also bring a higher degree of privacy to users, as no data is in transit to servers off-device. Running the models on the device also brings the additional benefits of being able to use generative AI without connection to the Internet. Moreover, the models can also be personalised to the users’ preferences as it ‘lives’ on the device. Llama 2 will also fit neatly into the Qualcomm AI Stack, a set of developer tools made to further optimise running AI models on-device. “We applaud Meta’s approach to open and responsible AI and are committed to driving innovation and reducing barriers-to-entry for developers of any size by bringing generative AI on-device,” said Durga Malladi, Qualcomm’s senior vice president and general manager of technology, planning and edge solutions businesses. Qualcomm has also worked closely with Meta in the past, mainly to make chips for its Oculus Quest VR headsets. The company has also tied up with Microsoft to help scale on-device AI workloads. As part of a partnership with Qualcomm and other chipmakers like Intel, AMD, and NVIDIA, Microsoft introduced the new Hybrid AI Loop toolkit to support AI development at the edge. Taking a zoomed out look at the edge AI hardware and software ecosystem, it is clear that the industry is moving towards AI at the edge, and Llama 2 might have a bigger role to play than anyone thinks. Setting the open source world on fire It seems that Meta has learnt a lot from the leak of the first LLaMA model. While the first iteration of this LLM was only available to researchers and academic institutions, the model and its weights were leaked on the Internet through 4chan. This resulted in an explosion of open source LLM innovation using LlaMa as the base model. In just under a month of its launch, the open source community had already bettered LLaMA in every way possible. Researchers at Stanford University created a version of LLaMA that could be trained for a cost of $600, which then led to the development of many other faster and lighter versions. Most, if not all of these versions, could be run on-device, giving the world access to their very own LLMs. One developer ported the LLM model to C++, which then resulted in a version of the model that could be run on a phone. The project, dubbed LLaMA.cpp, was fueled by the open source community, resulting in the model’s weights being quantised. This innovation allowed it to run on a Google Pixel 5, albeit generating only 1 token per second. As part of the latest partnership with Meta, Snapdragon could receive information about the inner workings of the model. This would enable the chipmaker to bake in certain optimisations, allowing Llama 2 to run better than other models. Considering the 2024 release window, it is also likely that Qualcomm will likely explore other partnerships to coincide with the launch of its Snapdragon 8 Gen 3 chip. The open source community is also sure to contribute its fair share to the (almost) completely open Llama 2. When combined with huge industry momentum for on-device AI, this move is the first of many to support a vibrant on-device AI ecosystem.","excerpt":"The conversation around the use cases of LLMs has been rapidly growing, supercharged by the release of Llama 2, Meta’s new open source model","categories":["Global Tech"],"tags":[],"author_name":"Anirudh VK","publish_date":"2023-07-20T17:00:00","publication_year":"2023","word_count":874,"keywords":["Go","API","cloud computing","AI","Scala","RAG","C++","edge AI","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","edge AI","RAG","cloud computing","R","Go","Scala","C++","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/meta-qualcomm-partnership-will-bring-llama-2-to-the-masses\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10044923,"title":"Hands-On Tutorial on Visualizing Spectrograms in Python","content":"In this modern data science scenario, there are many kinds of data required to analyze, and various analysis algorithms help us view the data better or understand the data. Still, when it comes to analysis, a series evolving with time, the spectrogram is the most common tool we frequently use to analyze this kind of data. Audio files, sound waves, and magnetic waves are the most common examples of this kind of data; all of them provide signal information in the form of data. Therefore, measuring the frequency and amplitude of the signals can be considered the main motive of the spectrogram. For visualising signals into an image, we use a spectrogram that plots the time in the x-axis and frequency in the y-axis and, for more detailed information, amplitude in the z-axis. Also, it can be on different colors where the density of colors can be considered the signal’s strength. Finally, it gives you an overview of the signal where it explains how the strength of the signal is distributed in different frequencies. So the amplitude and the frequency of the signal are the two main components of any spectrogram. Next in the article, we will have a general definition of both of them so that we won’t get confused about the terms we will use later in the article. Frequency Mathematically, frequency is the number of waves passing through a fixed point in a single time unit or the number of cycles performed by a body in a single time when it is in a periodic motion. Amplitude Amplitude can be defined as the greatest distance travelled by a moving body in a periodic motion in a single time unit or the highest distance of the wave on dips down or rising from its flat surface. The image below can give us the visualization of these components. Image source Let’s move on the spectrogram side. The mathematics behind the spectrogram is based on the Gabor transform. We use the Gabor transform to compute the spectrogram. Gabor transform is the special case of the short-time Fourier transform used to extract the sinusoidal frequency and phase content of a signal in its particular section. In Fourier transform, we take some signals in space or time and write them into their frequency components. More formally, from any signal, after performing a Fourier transform, we can pull out the signal’s frequency components that help to make the signal. In the images below, we can see the graphs of signal and frequency after the Fourier transform of the signal. Hereafter the Fourier transform, we can see the frequency components but for example, let’s consider the signal in the image as an audio signal of any song. Then we can know by the time domain graph what the words or tune is playing. By the frequency domain graph, we will be able to know what is the frequency of the tune. If we don’t know about the frequency at a particular time, Gabor transform comes into the picture to resolve this. Gabor Transform Gabor transform allows us to figure the spectrogram of any signal by using the time-frequency plot to easily track details in a signal like the frequency with the time factor.  In Gabor transform, we multiply the Gaussian function to our signal function. The function can be regarded as the window function, and the resultant of the process is then transformed with the Fourier transform to derive the time-frequency analysis. The window function is the signal near the time we want to analyze the signal and provide it with a heavier weight. Mathematically we can represent the Gabor transform as: Image source. In the image, we can see that the Gabor transform is built on the Fourier transform. As discussed before, we can see that to get the Gabor transform, we are multiplying the Fourier transform of the function with the window function of the frequency. The window function will slide with time, so basically, the window function is a function of the time range, or at a time, we can say that time to complete one frequency of the signal. In the image below, we can imagine that we have taken a gaussian window that slides across the signal. When it is positioned in a particular time domain, it will provide the weight to the signal of that particular time. Thus, it will generate the key present in our time-frequency graph of the signal in the form of small dashes. The distance from the x-axis is showing the magnitude of frequency where the distance from the y axis is showing the time. Next, we will see how we can make the spectrogram using python and also, we will be looking at the insights from the graph of any signal. There are various ways to make any spectrogram in python, and various libraries provide the direct modules to make spectrograms of any signal. So, let’s see how we can do that. Implementation of Spectrogram in Python Importing the required libraries : Input: import numpy as np from scipy import signal from scipy.fft import fftshift from matplotlib import mlab mport matplotlib.pyplot as plt mport matplotlib.pyplot as plt Making a signal using scipy in python: Input: fs = 10e3 N = 1e5 NFFT = 1024 amp = 2 * np.sqrt(2) noise_power = 0.01 * fs \/ 2 time = np.arange(N) \/ float(fs) mod = 500*np.cos(2*np.pi*0.25*time) carrier = amp * np.sin(2*np.pi*3e3*time + mod) noise = rng.normal(scale=np.sqrt(noise_power), size=time.shape) noise *= np.exp(-time\/5) x = carrier + noise Here I have created a signal which is a 2 Vrms sine wave with modulated frequencies around 3000 Hz, and also the amplitude of the signal is slowly decreasing from 20000 Hz to 100000 Hz. Next, we can make a plot of the signal so that we can have its overview. Input: plt.figure(figsize=(10,12)) plt.plot(x) plt.show() Output: Here we can see the waves of the images, which tells how the signal’s amplitude is changing. Let’s make a spectrogram of the signal using scipy.signal.spectrogram. Input: f, t, Sxx = signal.spectrogram(x, fs) plt.figure(figsize=(8,10)) plt.pcolormesh(t, f, Sxx, shading='gouraud') plt.ylabel('Frequency [Hz]') plt.xlabel('Time [sec]') plt.show() Output: Here we can see in the spectrogram how our wave is moving on its space, but in scipy’s spectrogram, we do not accurately measure amplitude. As discussed earlier in this topic, it will be much better to analyze the signal in a spectrogram when we can have all components (frequency, amplitude) with time in one image. So I basically prefer to use matplotlib to make a spectrogram of any kind of signal. Using matplotlib to make the spectrogram. Input: fig, (ax1, ax2) = plt.subplots(nrows=2) ax1.plot(time, x) Pxx, freqs, bins, im = ax2.specgram(x, NFFT=NFFT,Fs=fs, noverlap=900) plt.show() Output: . In this graph, we can see that as the amplitude decreases, the color of the spectrogram is varying or getting darker than before. So if we are working with any fluctuating audio file and by this graph, we can understand how the frequency(curve line of yellow color) is changing with time and what is the amplitude component of the particular time. We can modify that graph more in the next input. I am providing it with a color label to see the exact value of magnitude at a particular time. Input: def specgram2d(y, srate=44100, ax=None, title=None): if not ax: ax = plt.axes() ax.set_title(title, loc='center', wrap=True) spec, freqs, t, im = ax.specgram(y, Fs=fs, scale='dB', vmax=0) ax.set_xlabel('time (s)') ax.set_ylabel('frequencies (Hz)') cbar = plt.colorbar(im, ax=ax) cbar.set_label('Amplitude (dB)') cbar.minorticks_on() return spec, freqs, t, im fig1, ax1 = plt.subplots() specgram2d(x, srate=fs, ax=ax1) plt.show() Output: Here we can see more on the graph. Now we can easily measure the amplitude component of the signal with time using the color label. We can also visualize the signal in 3D to understand how the signal is going in the 3-dimensional space. Input: def specgram3d(y, srate=44100, ax=None, title=None): if not ax: ax = plt.axes(projection='3d') ax.set_title(title, loc='center', wrap=True) spec, freqs, t = mlab.specgram(y, Fs=srate) X, Y, Z = t[None, :], freqs[:, None],  20.0 * np.log10(spec) ax.plot_surface(X, Y, Z, cmap='viridis') ax.set_ylabel('frequencies (Hz)') ax.set_ylabel('frequencies (Hz)') ax.set_zlabel('amplitude (dB)') ax.set_zlim(-140, 0) return X, Y, Z fig2, ax2 = plt.subplots(subplot_kw={'projection': '3d'}) specgram3d(x, srate=fs, ax=ax2) plt.show() Output: Here we can see all three dimensions in one picture. Here it is a little messed up, but it can be helpful when working with real data. There are more ways to make the spectrogram of the signal. We can also use TensorFlow to make a spectrogram. I have written an article to explain the whole TensorFlow to preprocess the audio data with a spectrogram. Please refer to the article here- link. There are various uses of the spectrogram, like classification of the music, sound detection, where we compare the spectrogram of saved audio files to the target audio file. The ocean also sometimes uses the spectrogram for object detection by sending the SONAR waves and collecting the variation in waves in the form of spectrograms. Here in the article, we have seen what a spectrogram is, the mathematics behind the spectrogram, and how can we visualize spectrograms using python libraries. We have also gone through some examples that are done by the use of spectrograms. References Scipy.signal.spectrogram.Matplotlib.pyplot.specgram.Google colab for codes. .","excerpt":"For visualising signals into an image, we use a spectrogram that plots the time in the x-axis and frequency in the y-axis and, for more detailed information, amplitude in the z-axis. Also, it can be on different colors where the density of colors can be considered the signal’s strength. Finally, it gives you an overview of the signal where it explains how the strength of the signal is","categories":["AI Trends"],"tags":["fourier transform machine learning","Matplotlib","PowerBI","scipy","spectrogram","Tensorflow"],"author_name":"Yugesh Verma","publish_date":"2021-07-31T11:00:00","publication_year":"2021","word_count":1530,"keywords":["data science","scipy","NumPy","TPU","fourier transform machine learning","AI","TensorFlow","ML","PowerBI","Colab","Matplotlib","object detection","Python","spectrogram","Tensorflow"],"extracted_tech_keywords":["AI","ML","data science","TensorFlow","Colab","NumPy","Matplotlib","object detection","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/hands-on-tutorial-on-visualizing-spectrograms-in-python\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10089969,"title":"OpenAI, a Data Scavenging Company for Microsoft","content":"If you thought OpenAI exploiting Kenyan workers with a wage of less than $2 per hour was morally wrong, think about yourself for a moment. You, who has equally contributed for the improvement of ChatGPT but for free, in the name of reinforcement learning with human feedback (RLHF). Now, where is all of this data going and who is using it? As soon as we talk about ‘OpenAI’ or ‘ChatGPT’, the conversation is bound to head towards Microsoft and how the tech-giant has invested billions of dollars into the company. While it might be true that the investment was for furthering AI research, this partnership is also providing Microsoft with one of the greatest assets of this digital age, data​​, and—perhaps to make it worse—that data might be yours. Now with internet access enabled, OpenAI has removed the users from the picture as well. Recently, OpenAI also integrated the capability of browsing the web for answers within ChatGPT through several new plugins, something that was previously touted as the biggest limitation of the AI model. This will most likely enable real-time accuracy and up-to-date responses from the chatbot, not excusing the 2021 cut-off data anymore. But, as much as this would increase the usability and emergence of new cool things, there is a different, probably darker side to this. Up until now, ChatGPT was limited to its training data. But now, the capability to move beyond its training data enables this AI chatbot to retrieve more data, thus making the LLM-based chatbot even larger. But why is this an issue? Sam Altman, CEO of OpenAI, recently expressed concerns about the capabilities of these generative models and also suggested that the progress of this field must be slower. But now, by connecting it to the internet, the company’s motivation behind developing this technology indicates something else. Setting aside the security and privacy issues of what these plugins on ChatGPT might entail, internet connectivity will now provide more data to OpenAI, which eventually would be provided to—guess who? Microsoft, which is not just the fund provider but is also the cloud provider of the “non-profit” research organisation. All the data that OpenAI gathers is stored on Microsoft’s cloud. OpenAI’s Privacy policy does not deny the fact that it shares personal information of the users’ with its vendors and service providers, which clearly is Microsoft. How would Microsoft benefit from this? Well, for starters, they would have access to all the data on the internet that can be used in building their own products. While OpenAI might be busy building AI models and collecting data, Microsoft is already heading its own way and utilising the collected data. If anyone wants to access that data, Microsoft can create its own walls and restrict everyone else. Pretty aggressive. Users definitely have to be very careful before sharing their personal information on the free-to-access ChatGPT and GPT-4, which are now also available on almost all of Microsoft’s products. On another note, Microsoft, OpenAI, and GitHub were recently hit with a class-action lawsuit for using developers’ code in building Copilot. This eventually led to OpenAI discontinuing support for Codex, the technology that was powering Copilot and integrating GPT-4 into Azure OpenAI Service. In all likelihood, there might be another lawsuit waiting for these companies when users realise that GPT-4 is also trained using similar, if not the same, data. Satya Nadella, CEO of Microsoft, has already spoken about the future of generative AI and what the partnership means for both the companies. “In future, the generative models will generate most of the data. But, on the other side, we should also think about how it can augment us in what we are doing today since it can have a huge impact on our future.” As part of their investment, Microsoft gained exclusive access to the entire OpenAI codebase— Elon Musk (@elonmusk) March 24, 2023 The AI race is fuelled by data for the most part and we can clearly see that Microsoft has recognised that. There is a high probability that Microsoft regards OpenAI merely as a ‘data scavenging company’ more than an AI company. Why Does It Matter? Interestingly, Microsoft has laid off its entire Responsible AI team. Data privacy, storage, or usage are probably just fluff talk for the company anyway. When OpenAI announced ChatGPT Plus, they also said that they would not be storing users’ data anymore for training the model but, for that to happen, the users have to opt out. In addition, the data would be deleted only after a month. Notably, with internet access now, the company does not even need to store data for training the model. Microsoft would, in turn, be able to take this data from all the websites across the internet, repack it as theirs, and present it for the users. This means that people would essentially stop going to the internet or visiting websites for accessing information, and only rely on ChatGPT or GPT-4 in the future. This would bring online traffic to an eventual halt and thus affect the revenue of these sources, inevitably leading them to shut down. Read: OpenAI and Microsoft: A Match Made in Tech Heaven On the other hand, Google is no less. This big tech has been monopolising the search engine market since its inception. No matter how much Microsoft tries to push Bing by integrating GPT into search, Google stays a step ahead. Recently, Google released ‘Bard’ to the public. Though it is reportedly trained on LaMDA, the company’s own LLM, when asked about the training data to the chatbot itself, it reveals that it includes Google Search and Gmail, among other apps. Though the company has tried to refute the claim by saying that the chatbot might be hallucinating, the bug that they are actually doing this has already been planted on many people’s minds. This is good for Microsoft in its bid to get ahead of Google, to which Jeff Bezos says “Treat it like a mountain — you can climb it, but not move it.” It’s hard to deny that OpenAI’s datasets are a valuable asset. ChatGPT’s massive success has given the company its biggest asset. With Microsoft’s deep pockets and extensive reach, it’s up to them to decide what to do with the data. So, while the partnership between Microsoft and OpenAI may not be as good as some had hoped for data privacy, it’s still a win-win situation for both the companies.","excerpt":"While it might be true that the investment was for furthering AI research, this partnership is also providing Microsoft with one of the greatest assets of this digital age, data​​, and—perhaps to make it worse—that data might be yours.","categories":["Global Tech"],"tags":["GPT-4","Microsoft","OpenAI"],"author_name":"Mohit Pandey","publish_date":"2023-03-24T18:00:00","publication_year":"2023","word_count":1078,"keywords":["ChatGPT","RLHF","OpenAI","AI","AWS","R","GPT-4","RAG","Aim","generative AI","Azure","Microsoft"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Aim","RAG","RLHF","AWS","Azure","R"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-a-data-scavenging-company-for-microsoft\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60350,"title":"Why Software Developers Are Most Likely To Work From Home","content":"Software developers already constitute the largest group of remote workers in the world. Given the current COVID-19 situation, more developers are likely to shift to a similar arrangement. But, why are developers leading the course of this fluid workforce? In a survey, some 73% of technology professionals revealed that they think of remote work as an essential perk. Programming requires high levels of concentration and focus since it involves working on complex algorithms and mathematical software. Open offices may be too noisy and full of interruptions. Hence, given a choice between a public office and a quiet home, the latter is mostly preferred. Developers Worked Towards ‘Work From Home’ Long Ago Innovations were in place that could enable developers to track anyone from any part of the world with a simple click of a button. Processes that were once done through meeting clients in person were now accomplished on online systems that only require participants to log in. When people realized that a well-paying job could be done at home with the help of the internet, developers started making codes more straightforward for regular people to understand. It is also worth highlighting the agile development practices, and the benefits of co-location. Most agile enthusiasts will tell you that it is better if the team is co-located. This is true to the extent that given an ideal organization, you will probably get better output out of them if they are co-located rather than working remotely. But sometimes, to get that perfect team, you need to provide conditions that they are happy with. If software developers are not happy travelling three hours every day, then you might need to live with the ideal team where some or all of the team members are operating remotely. Developers Still Need To Make Sure There Is Effective Communication With The Management It is evident that not all tech experts require remote work, even though most companies seem to be offering that amid the COVID-19 pandemic. It also places more expectations on managers and executives to create a workspace where programmers can work without too many interruptions and distractions. For example, it is increasingly becoming clear that many developers dislike open offices, and prefer to operate in comparative isolation, at least when they want to concentrate. For managers, this is also a problem of communication. Rather than implementing top-down strategies that might upset developers’ flow, it might be better to talk to them about their preferred ways of working—whether that is at a desk, or a cafe. Just listening and coordinating with developers can guarantee the best product as it is a critical job role which does not require a lot of micromanagement. In the short-term, conferencing technology companies are enabling alternatives. Companies like Zoom, Webex, and other teleconferencing and remote working solutions are being used extensively by organizations. Handling CyberSecurity Better Than The Average Employee There are other factors as well that come into play. Programming usually does not involve face-to-face problem discussion all the time. If need be, there is video chat as well as virtualization tools like Slack and Trello. These tools are user-friendly and tailor-made for remote workers, particularly developers. The built-in integrations and APIs can be leveraged by developers to customize virtualization tools to their most optimum use. Also, work from home, when it comes to developers, is more efficient in the context of cybersecurity as well. Compared to an average company employee, developers and IT teams can follow cybersecurity protocols from any remote location, and make sure they detect vulnerabilities in the workflow without a lot of help. In the case of other departments like sales, marketing, and finance, the employees can face great difficulty and ignorance to maintain the best security principles. In many cases, everyday office workers face challenges in setting up two-factor authentication services, let alone debug a cybersecurity software tool. This translates into more trust as well as freedom put into the hands of developers when it comes to handling sensitive data and security protocols.","excerpt":"Software developers already constitute the largest group of remote workers in the world. Given the current COVID-19 situation, more developers are likely to shift to a similar arrangement. But, why are developers leading the course of this fluid workforce? In a survey, some 73% of technology professionals revealed that they think of remote work as […]","categories":["AI Features"],"tags":["Developers","Software","Software Development","Work from Home"],"author_name":"Vishal Chawla","publish_date":"2020-03-30T11:39:35","publication_year":"2020","word_count":669,"keywords":["Go","API","TPU","programming_languages:R","AI","innovation","Work from Home","RAG","Rust","GAN","Software Development","Software","R","Developers"],"extracted_tech_keywords":["AI","RAG","TPU","R","Go","Rust","API","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-software-developers-are-most-likely-to-work-from-home\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10039328,"title":"Big Tech Comes To India&#8217;s Aid, The EU&#8217;s Chip Autonomy And More In This Week’s Top News","content":"This week saw the big tech come to India’s rescue as the subcontinent is among the nations worst hit by COVID-19’s second wave. From million dollars in funding to oxygen supplies, the companies have lend their helping hand in all possible ways. “I am heartbroken by the current situation in India. I’m grateful the U.S. government is mobilizing to help. Microsoft will continue to use its voice, resources, and technology to aid relief efforts, and support the purchase of critical oxygen concentration devices,” tweeted Satya Nadella, CEO, Microsoft. Whereas, Google announced $18 million in new funding for India, which includes two grants from Google.org, Google’s philanthropic arm, totalling $2.6 million. Check the full coverage here. Lyft Sells Its Self Driving Segment Ride hailing service Lyft is unburdening itself of its much hyped level-5  self-driving efforts as the company announced the acquisition by Toyota’s subsidiary Woven Planet. “Today’s announcement launches Lyft into the next phase of an incredible journey to bring our mission to life,” Lyft Co-Founder and CEO Logan Green said in a statement. Lyft’s Open Platform team will become the new Lyft Autonomous team. As per the agreement, Lyft will receive, in total, approximately $550 million in cash with this transaction, with $200 million paid upfront subject to certain closing adjustments and $350 million of payments over a five-year period. The Woven Planet team, alongside the team of researchers at Toyota Research Institute, have already established a center of excellence for software development, automated driving, and advanced safety technology within the Toyota Group. “This acquisition assembles a dream team of world-class engineers and scientists to deliver safe mobility technology for the world,” James Kuffner, CEO of Woven Planet said. Big Tech Report Record Earnings Source: WSJ The earning reports have come in and it looks like the tech juggernauts are having none of the pandemic clout. Between January and March, Apple sold iPhones worth $47 billion, a 66% jump from the previous year. Facebook too has exceeded the expectations as it reported a 48% rise thanks to the booming ad revenue. Amazon Web Services(AWS), which has been making profits for Amazon for many consecutive quarters reported a net revenue of $13.5 billion this quarter, up 32% year over year. Google’s parent company Alphabet reported a revenue of $55.31 billion. “Over a year into the pandemic, digital adoption curves aren’t slowing down. They’re accelerating,” said Satya Nadella as he revealed Microsoft’s revenue which has risen by 19% in the latest quarter. Now Europe Wants Chip Autonomy As the dangers of chip drought looms large, governments across the world are taking extreme measures to dodge an indefinite chip-dependency on the east. Last month, the US President Biden called upon the chip makers to bring the fabs back to the shores. Now, the European Union is in plans to dish out its own version of chip autonomy. According to reports, the EU is mulling a pan-European scheme known as an Important Project of Common European Interest (IPCEI), is aimed at doubling the EU’s market share in semiconductors to 20% by 2030. Taiwan’s TSMC makes most of the chips that power our cars, computers etc., However, the country is facing worst drought in 50 years which aggravated the chip production, which was already disrupted by the pandemic last year. The EU Commissioner will soon be meeting Intel’s chief Pat Gelsinger and TSMC’s President to explore potential synergies. SPOT In A Tight Spot NYPD’s DIGIDOG (Source: ArsTechnica) The New York Police Department’s plans to hire a robo dog came to a halt this week. Named “DIGIDOG”, the Boston Dynamics’ SPOT robot was purchased by the law enforcers to sniff out suspicious activities i n the city. When the digi-dog’s presence on the crime scene triggered a heated debate where people accused the whole initiative to be overtly dystopian. SPOT was originally designed to run errands in factories, work in harmful environments etc. But, its functionality in a highly critical crime scene was immediately protested. In response to a subpoena, police officials said that a contract worth roughly $94,000 to lease the robotic dog from its maker, Boston Dynamics, had been terminated on April 22. EU Charges Apple With Antitrust On Friday, the European Union charged Apple with antitrust violations for allegedly abusing its control over the distribution of music-streaming apps, broadening the battle over the tech giant’s App Store practices ahead of a federal trial in the U.S. brought by “Fortnite” maker Epic Games. According to the EU, Apple squeezed rival music-streaming apps by requiring them to use Apple’s in-app payments system to sell digital content. The case stems from a complaint by Spotify Technology SA, which competes with Apple’s music-streaming service. While the “big tech- antitrust” saga adds a new chapter almost every month, there is something more serious brewing on the Easter front. China, which began its own version of a big tech crackdown last year, has now unleashed its whip on fintech giants. On Thursday, China’s central bank and four other regulatory agencies told WeChat, ride-hailing company Didi and others that their apps should no longer provide financial services beyond payments, according to people familiar with the discussions. Jack Ma’s Ant Group, has already accepted the terms of the regulators and is in process of delinking its financial services.","excerpt":"This week saw the big tech come to India’s rescue as the subcontinent is among the nations worst hit by COVID-19’s second wave. From million dollars in funding to oxygen supplies, the companies have lend their helping hand in all possible ways. “I am heartbroken by the current situation in India. I’m grateful the U.S. […]","categories":["AI News"],"tags":["china chip technology"],"author_name":"Ram Sagar","publish_date":"2021-05-02T10:00:00","publication_year":"2021","word_count":878,"keywords":["Anthropic","Go","AWS","AI","Git","RAG","Aim","china chip technology","Rust","GAN","R"],"extracted_tech_keywords":["AI","Anthropic","Aim","RAG","AWS","R","Go","Rust","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/big-tech-covid-record-earnings-top-news\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":26509,"title":"How Researchers Are Building Models To Safeguard Private Data In Machine Learning","content":"Image credits : Flickr More machine learning applications are permeating in the tech ecosystem and the data that goes into ML systems is being derived from all sorts of sources — regardless of its sensitivity. ML algorithms do not realise the aspect of sensitivity as it always looks at data as a way to establish and learn patterns, rather than looking into the who’s who of the data. Miscreants might take advantage of this and circumvent the ML systems itself, which can have devastating effects altogether. If that happens, the purpose of ML will completely fail. To counter this, and establish a secure and safe ML environment, researchers are strictly working towards building privacy in ML models. In this article, we explore various studies that have seen privacy as the core focus of ML. Using Anonymity For Privacy In order to contain private data in ML from being misused, anonymity was also another option present. But this had a major setback. Anonymous data would sometimes mislead ML and it also lead to vague interpretation of data. Nicolas Papernot and Ian Goodfellow, lead researchers at Google, described the drawback with anonymity as follows: “However, anonymising data is not always sufficient and the privacy it provides quickly degrades as adversaries obtain auxiliary information about the individuals represented in the dataset. Famously, this strategy allowed researchers to de-anonymise part of a movie rating dataset released to participants of the Netflix Prize when the individuals had also shared their movie ratings publicly on the Internet Movie Database (IMDb).” That example cited by the researchers shows that the end results can be disastrous with data going incognito. This paved way for searching a foolproof method to regulate private data. Differential Privacy: The Base For Measuring Privacy Risk In his paper, reputed ML author Tom Mitchell emphasises that the benefits of ML are sometimes questioned with privacy issues. He suggests that in order to get the best of both ML and privacy, some trade-offs have to be considered amongst them. In addition, he highlights that ML can be modified to curb sensitive information from falling into the wrong hands, and thus preventing disasters in the system. This was the reason many researchers started pursuing the aspect of incorporating privacy in various other fields of computer science such as cryptography, and database systems. One such concept of privacy that saw popularity in the research community was “differential privacy”, where it relies on mathematical relations and a statistically-controlled database to carefully extract information from data analysis algorithms. From then on, there were various research studies which revolved around the original work of differential privacy to pass it on to various applications. The results from these studies proved that the concept found significance and powerful enough in other applications. Slowly, differential privacy also made into ML as well. From Omnipresent To Private: Progress In Differential Data Privacy One of the earliest studies that saw differential privacy in ML was by scholars Anand Sarwate and Kamalika Chaudhuri from the University of California where they explored various models and algorithms designed to maintain privacy, and then analyse with respect to ML, and signal processing. In the study, differential privacy is explained in terms of mathematical functions coupled with a database. In order to achieve this, the authors add a noise function to obtain a differential approximation in the algorithm, which as a result gives out randomness in privacy. Now, this is applied for common ML tasks such as classification, regression, dimensionality reduction and time series, and thus a privacy-preserving mechanism is obtained in general. Although this study had practical limitations based on technical assumptions, it served as the basis for further studies. The subsequent papers on privacy in ML focussed on optimising loss functions, dealing with missing data along with handling publicly available data. Privacy-Preserving ML Models Recent years have now seen unique approaches to privacy in ML. A popular amongst them are privacy-preserving ML models, which use insights from differential privacy concepts to counter malign attacks made at every phase and component of ML. One paper by academics at Pennsylvania State University and the University of Michigan present a demonstrative ML model that showcases and analyses adversarial attacks at every stage of the model right from training to its inference. They layout threat models to pinpoint possibilities of attacks. The authors emphasise that there is always some kind of uncertainty between privacy and ML prediction, and this trade-off is to be balanced perfectly. Another study called SecureML by researchers Payman Mohassel and Yupeng Zhang, provide novel ways to tackle adversaries in various scenarios in ML. They specifically consider stochastic gradient descent (SGD) to devise privacy-preserving ML algorithms with lesser loss and implications on data. Now the importance has even spread to building systems for ML privacy as a service. A system called Chiron, has been developed to cater to entities that have ML-as-a-service (MLaaS) platforms. The developers of Chiron say that the system runs popular ML frameworks such as Theano, on SGX enclaves. They show that Chiron proves to be a practical element for ML privacy. Conclusion All of the above studies and methods depict the possibilities that can be made to entrust privacy in mission-critical ML systems. Since this is a new area of research, it is yet to achieve a full-fledged status for ML systems to be privacy-adherent.","excerpt":"More machine learning applications are permeating in the tech ecosystem and the data that goes into ML systems is being derived from all sorts of sources — regardless of its sensitivity. ML algorithms do not realise the aspect of sensitivity as it always looks at data as a way to establish and learn patterns, rather […]","categories":["AI Features"],"tags":["database systems","differential privacy","Machine Learning","Privacy","Statistics"],"author_name":"Abhishek Sharma","publish_date":"2018-07-18T11:31:59","publication_year":"2018","word_count":890,"keywords":["Go","Privacy","machine learning","Statistics","AI","ML","Machine Learning","ViT","database systems","differential privacy","Rust","GAN","R","adversarial attacks"],"extracted_tech_keywords":["AI","machine learning","ML","differential privacy","R","Go","Rust","GAN","ViT","adversarial attacks"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-researchers-are-building-models-to-safeguard-private-data-in-machine-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10171598,"title":"Freshworks Upgrades Its AI Platform With Abilities to Create AI Agents and More","content":"Freshworks has launched new features for its Freddy AI platform, enhancing automation across customer and employee service operations. The updates aim to simplify complex workflows and enable autonomous resolution of service requests. The upgraded platform includes Freddy AI Agent Studio, a no-code builder for deploying task-specific AI agents, and new capabilities in email support, knowledge search, IT insights, and AI copilot functions. These features allow Freddy AI agents to take direct actions, such as issuing refunds or booking appointments across business apps like Shopify and Stripe, without human intervention. “AI agents can be taught unique skills to autonomously take actions like issuing a refund, checking order status, or updating a customer record,” the company stated. Businesses also get access to pre-built templates to easily add the required skills to the AI agents they create. Murali Swaminathan, CTO of Freshworks, told AIM, “The whole idea of the Freddy AI agent studio is to provide a set of out-of-the-box templates for our customers to build AI agents quickly. We also have our Skill Builder behind the scenes, which will enable them to create new AI agents.” The new Freddy Agentic AI platform supports multi-application workflows, autonomous email handling, and root cause analysis for IT operations. The agent integrates with Slack and Microsoft tools, supports 40+ languages, and offers improved context awareness through a multi-LLM architecture. Swaminathan highlighted that the company utilises GPT-4o and GPT-4o mini models under the hood and hosts several others, including Llama, for specialised applications. The company claims real-world productivity gains for its customers. Swaminathan told AIM, “Hobbycraft is our customer that has deployed Freddy AI agents for the customer service use case. They’ve been able to successfully deflect at least 30% of their support requests, which has indirectly helped them improve customer satisfaction by at least 25%.” The company also mentioned other customer stories, such as Bergzeit, which reduced ticket translation efforts by 75%, and iPostal1, which now resolves over half of its support queries automatically. Furthermore, Five9 reports saving 200 IT hours monthly using Freddy AI Copilot. Swaminathan explained to AIM that the company is exploring opening up an MCP server to Freddy AI capabilities and is aiming for wider collaboration with other leading solutions, like a Salesforce agent. Furthermore, Agent2Agent (A2A) protocols are also being explored to facilitate interoperability between AI agents beyond the Freshworks ecosystem. To encourage adoption, Freshworks has introduced training resources, partner programmes, and professional services. With over 5,000 organisations using Freddy AI since its 2023 launch, Freshworks positions its agentic platform as a scalable alternative to legacy service automation.","excerpt":"Freshworks says Freddy Agentic AI Platform is designed to go beyond simple automation to solve challenges in real-time.","categories":["AI News"],"tags":["Freshworks"],"author_name":"Ankush Das","publish_date":"2025-06-11T15:21:24","publication_year":"2025","word_count":427,"keywords":["agentic AI","AI","GPT-4o","Scala","Freshworks","RAG","GPT","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","agentic AI","GPT-4o","Aim","RAG","R","Scala","GPT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/freshworks-upgrades-its-ai-platform-with-abilities-to-create-ai-agents-and-more\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":583,"title":"World’s third LIGO project to be commissioned in India, ready by 2024","content":"India is all set to add another feather to its cap. The site for the proposed LIGO India project, an advanced gravitational wave observatory has been finalized. To be operational by 2024, India has the honour for setting up the world’s third LIGO observatory. Currently, USA houses two observatories – Hanford in Washington and the other in Livingston in Louisiana operated by CalTech and MIT.  The two sites identified for the project are Udaipur in Rajasthan and Hingoli in Maharashtra. With a budget of Rs 1,260 crore earmarked and a nod from Prime Minister Narendra Modi, the Indian LIGO project will open up avenues of research in space matter and energy by tracking cosmic gravitational waves. For the uninitiated, Laser Interferometer Gravitational-Wave Observatory (LIGO) is a massive gravitational wave observatory consisting of two laser interferometers placed thousands of kilometers apart to make use of light and space to detect gravitational waves. The ground-breaking LIGO-India project that will refine the “understanding of black holes and why they occur” will be carried out in collaboration with several international partners Australia, UK and Germany. The proposed LIGO project will also reportedly move one advanced LIGO detector from Hanford to India. Astronaut Today lists down how the path-breaking scientific endeavor will impact Indian scientific community: This large scale physics experiment and observatory is a step in the direction of gravitational wave astronomy Lead to a new breed of young researchers in advanced physics, gravitational physics, cosmology, computational sciences, and mathematics and in engineering Spawn academic-industry partnerships and lead to more job opportunities The LIGO-India project will serve as a research facility for the international community The Indian LIGO project will be a collaboration between California Institute of Technology (CalTech) and Massachusetts Institute of Technology (MIT) and three Indian universities Inter-University Centre for Astronomy and Astrophysics (IUCAA), Pune, Institute for Plasma Research (IPR), Gandhinagar, and Raja Ramanna Centre for Advanced Technology (RRCAT), Indore. What goes into the making of LIGO? Seismic isolation system is aimed at eliminating vibrations that fall under two categories – active and passive damping system.  Since LIGO is sensitive to the smallest motions and fleeting vibrations  such as the ones caused by speeding trucks on a nearby road, seismic isolation system is the defense system to remove these environmental noises. In a way, it works likes a noise-cancelling headphone. Vaccum: LIGO comprises of one of the largest vacuums on earth with the atmospheric pressure that equals one-trillionth that of air pressure at sea level. The vaccum functions to eliminate any dust that will deflect on the laser. Another reason for building a high-quality vacuum is to remove any air in the path of laser that could potentially mask gravitational waves. Optics System: The optic system comprises of a 200 watt laser beam that helps in detecting gravitational waves. Mirrors: The pure fuse silica glass mirrors weighing up to 40 kg can absorb one in 3 million photons that hit them, which means they are able to reflect most of the light that hits them. The mirrors are used to refocus the laser so that it can travel without interruptions and maintains the stability of laser light.","excerpt":"India is all set to add another feather to its cap. The site for the proposed LIGO India project, an advanced gravitational wave observatory has been finalized. To be operational by 2024, India has the honour for setting up the world’s third LIGO observatory. Currently, USA houses two observatories – Hanford in Washington and the […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2017-02-17T07:30:23","publication_year":"2017","word_count":525,"keywords":["Go","programming_languages:R","AI","programming_languages:Go","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","Aim","R","Go","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/worlds-third-ligo-project-commissioned-india-ready-2024\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":64056,"title":"Facebook AI Research Launches Blender, A Human-Like Chatbot","content":"Facebook AI Research has announced the development of a new chatbot — Blender, which they claim to be the largest-ever, state-of-the-art open-domain chatbot model. The company said that the chatbot can outperform other models in terms of engagement and can also feel more human. In a recent tweet, Facebook AI confirmed the news stating — “Today we’re announcing that Facebook AI has built and open-sourced Blender, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators.” Today we’re announcing that Facebook AI has built and open-sourced Blender, the largest-ever open-domain chatbot. It outperforms others in terms of engagement and also feels more human, according to human evaluators. https:\/\/t.co\/TdNxJpL0JI pic.twitter.com\/1wo3ZxwDeU— AI at Meta (@AIatMeta) April 29, 2020 The company has also released a blog post stating where it stated — “This is the first time a chatbot has learned to blend several conversational skills — including the ability to assume a persona, discuss nearly any topic, and show empathy — in natural, 14-turn conversation flows.” It further stated — “Our new recipe incorporates not just large-scale neural models, with up to 9.4 billion parameters — or 3.6x more than the largest existing system — but also equally important techniques for blending skills and detailed generation.” According to the blog, Blender is a result of a culmination of years of research in conversational AI, where it includes empathy, knowledge, and personality — together in one system. The company included improved decoding techniques, the novel blending of skills, and a model with 9.4 billion parameters, which is 3.6x more than the largest existing system, in order to build this chatbot model. The company released its complete model code, and evaluation set-up so that other AI researchers can reproduce this work and continue to advance conversational AI research. According to researchers, Blender comes with diverse skills. The company has done a large scale training for creating the chatbot by involving 1.5 billion training examples of extracted conversations. It further utilised column-wise model parallelism, which allowed the researchers to split the neural network into smaller, more manageable pieces while maintaining maximum efficiency. Blender also learned blending skills for training and evaluating these desirable BST skills, which includes, engaging use of personality engaging use of knowledge, display of empathy, and ability to blend all three seamlessly. According to the blog, “Blending these skills is a difficult challenge because systems must be able to switch between different tasks when appropriate, like adjusting tone if a person changes from joking to serious. Our new BST data set provides a way to build systems that blend and exhibit these behaviours. We found that fine-tuning the model with BST has a dramatic effect on human evaluations of the bot’s conversational ability.” The researchers further found that the length of the agent’s utterances is important in achieving better results with human evaluators. The company showed that a careful choice of search hyperparameters could give strong results by controlling this trade-off. In particular, tuning the minimum beam length gives important control over the “dull versus spicy” spectrum of responses. To evaluate the model, the researchers benchmarked its performance against Google’s latest Meena chatbot through pairwise human evaluations. Since their model has not been released, the researchers used the roughly 100 publicly released and randomised logs for this evaluation. Using the ACUTE-Eval method, human evaluators were shown a series of dialogues between humans paired with each respective chatbot. Facebook AI is excited about the progress that has been made in improving open-domain chatbots. Although they are still far from achieving human-level intelligence in dialogue systems, the researchers are currently exploring ways to further improve the conversational quality of our models in longer conversations with new architectures and different loss functions. “True progress in the field depends on reproducibility — the opportunity to build upon the best technology possible. We believe that releasing models is essential to enable full, reliable insights into their capabilities. That’s why we’ve made our state of the art open-domain chatbot publicly available through our dialogue research platform ParlAI. By open-sourcing code for fine-tuning and conducting automatic and human evaluations, we hope that the AI research community can build on this work and collectively push conversational AI forward,” the blog stated.","excerpt":"Facebook AI Research has announced the development of a new chatbot — Blender, which they claim to be the largest-ever, state-of-the-art open-domain chatbot model. The company said that the chatbot can outperform other models in terms of engagement and can also feel more human.  In a recent tweet, Facebook AI confirmed the news stating — […]","categories":["AI News"],"tags":["AI Chatbot","AI Chatbot","chatbot ai","Facebook AI","Facebook AI research"],"author_name":"Sejuti Das","publish_date":"2020-04-30T18:22:19","publication_year":"2020","word_count":711,"keywords":["Go","Facebook AI","AI","AI Chatbot","neural network","ML","chatbots","RPA","programming_languages:R","Aim","chatbot ai","Facebook AI research","AI research","R"],"extracted_tech_keywords":["AI","ML","neural network","Aim","chatbots","R","Go","RPA","AI research","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/facebook-ai-research-launches-blender-a-human-like-chatbot\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10078187,"title":"New Algorithms That Harnessed Protein-folding Power in 2022","content":"Big pharma companies have been researching protein folding for a long time now. Discoveries and innovations in the field can revolutionise the development of drug and other biological advancement. Recently, the development of the COVID-19 vaccine was also supported by tackling this issue. Protein folding prediction process involves a combination of complex algorithms. Recent models from big tech companies like Meta and Google have made advancements in solving this problem for protein folding and sparked interest in researchers after getting open-sourced. Here’s a list of some of the prominent protein fold prediction models that are highly accurate and compete with each other in terms of their methods and speed! AlphaFold 2 Google’s DeepMind made a major breakthrough using a deep learning approach to build AlphaFold, which has a network-based approach for predicting protein structures. In 2018, AlphaFold 1 was highly appreciated at CASP13 for its remarkable innovations and now with AlphaFold 2, DeepMind has increased the speed and accuracy even further. AlphaFold 2 won the CASP14 in 2020 and is since then regarded as the best protein folding model. DeepMind decided to make their model open-source for more contributions and to further innovations. In July, DeepMind collaborated with European Bioinformatics Institute (EMBL-EBI) and released the predicted structures of all catalogued proteins, thereby expanding their earlier database by more than 200X. Check out the code for AlphaFold here. ESMFold Meta AI’s launch of Evolutionary Scale Modeling (ESM), proved to be one of the biggest competitors or the best alternative to AlphaFold 2. Much like AlphaFold, the model is also open to the public. ESMFold has excellent accuracy and works on end-to-end atomic level protein structure. It uses ESM-2, which is a transformer-based language model built on 15 billion parameters. Since it is based on a language model, ESMFold stands apart from other protein fold prediction models in that it offers higher accuracy and faster inference. ESMFold produces precise protein structure even with a single sequence as input as it leverages the internal representations of the language model. When it comes to tests on CASP14, the model received a score of 68 which is lower than that of AlphaFold 2, which received a score of 84. To see the code, click here. RoseTTAFold Minkyung Baek from the Baker Lab developed a tool to predict protein structures using deep learning called ‘RoseTTAFold’. It is based on a three-track neural network and is interestingly insightful towards protein structure even without a determined structure—making it faster at prediction. The three-track network integrates one-dimensional protein structure and processes into two-dimensional sequence information with the distance of amino acids at once. The software allows direct collection of reasons and patterns in the relationship between folded architecture and peptides. According to several reports, RoseTTAFold was able to predict tens of hundreds of new protein structures that were unknown before. Scientists and researchers also predict that the software could resolve x-ray crystallography and cryo-electron microscopy modelling problems. Click here for the GitHub repository. OmegaFold In July, Chinese biotech firm, ‘Helixon’, developed OmegaFold and joined the protein fold prediction race—beating its competitors in several areas. After outperforming RoseTTAFold and competing with AlphaFold 2 for its high-resolution protein structure prediction, the developers released the code to the public on GitHub. The model works on divergent sequences, unlike multiple sequence alignments in AlphaFold and RoseTTAFold, which allows them to make predictions and suggest geometry-inspired transformer models trained on protein structures from single sequences. OmegaFold works on the protein language model, OmegaPLM, that can sense structural information encoded in amino-acid sequences. Thus, the model can predict protein structure ten times faster than RoseTTAFold as it can predict structure and folds with a single amino-acid sequence. Click here for the repository. D-I-TASSER Zhang Lab from the University of Michigan developed Distance-guided Iterative Threading ASSEmbly Refinement, or D-I-TASSER, which is used for high-accuracy protein fold and structure prediction. It is built by integrating threading and deep learning. D-I-TASSER comes after the lab’s older model, ‘I-TASSER’, and provides higher speed and accuracy. Starting with a query sequence, the generation of inter-residual contact and distance maps is processed using two multiple deep neural network predictors—DeepPotential and Attention Potential. The model has an optional additional server called D-I-TASSER-AF2 that incorporates AlphaFold2 restraints and increases general accuracy when compared to both models separately. Click here to visit the lab’s website. IntFOLD This server provides a unified resource for predicting protein tertiary structures automatically with built-in estimates of model accuracy (EMA). The server is a fully automated, high-performance tool for predicting protein structures from their amino acid sequences. The server was tested on CASP and performed very well in the blind tests. The results are presented in graphical outputs, which is also beneficial for non-expert users as it provides a visual summary of a complex set of data. Click here to read IntFOLD’s research paper. RaptorX RaptorX offers a template-based protein secondary structure prediction and modelling. The template-based tertiary structure modelling approach allows the model to finish processing a sequence of 200 amino acids in around 35 minutes. What sets RaptorX apart from other protein fold prediction models is a novel non-linear scoring function, aligning target sequence with multiple distantly-related template proteins and probabilistic consistency algorithm. Read more about RaptorX here.","excerpt":"List of top protein folding prediction models.","categories":["AI Features"],"tags":["AlphaFold","ESMFold","RoseTTAFold"],"author_name":"Mohit Pandey","publish_date":"2022-10-27T18:00:00","publication_year":"2022","word_count":871,"keywords":["Go","AlphaFold","Meta AI","TPU","ESMFold","AI","neural network","Git","RAG","Ray","deep learning","RoseTTAFold","R"],"extracted_tech_keywords":["AI","deep learning","neural network","Meta AI","Ray","RAG","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/new-algorithms-that-harnessed-protein-folding-power-in-2022\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":22646,"title":"Can Artificial Intelligence Learn From A Child’s Language Acquisition Process?","content":"Can language be taught to artificial intelligence? Given that “machines are capable of doing any work man can do” researchers are working hard on long-standing projects to train machines so they can talk just like us. But unlike identifying images, words are grounded in sense and representation. Jonathan Mugan, co-founder & CEO at DeepGrammar, an AI company focused on NLP believes to train machines so that they can talk with us, we need to immerse them in an environment which is similar to ours. In this direction, there is a considerable research underway in developing better real worlds by allowing computers to spend more time in the virtual one. Case in point – OpenAI trained machines by playing Grand Theft Auto V to help it understand the real world better. Even though artificial intelligence has made considerable leap in the recent years, especially with deep learning, but current chatbots are still an embarrassment, added Mugan. Despite the complexity of language, a child is able to learn a natural language competently in a short span of time. Can machines have the capacity to learn language like children do? Can AI Learn from a child’s language acquisition process? According to Steven Pinker, experimental cognitive psychologist and a leading mind on languages and the mind, the “child’s language acquisition process has solved a remarkably difficult computational problem”. Pinker shared in a recent interview to Edge that AI’s problem can be stated as an engineering task –  designing an algorithm that ingests a bunch of sentences and their contexts from any of the five thousand languages on the planet, and after crunching through a number of these sentences — drums up the grammar for the language, regardless of what the language is. Amaxon’s Alexa & Google Home outfitted with voice-enabled interface are excellent examples of present day automated speech recognition and NLG. The key question is whether a computational simulation of natural language acquisition (just in a way a child acquires natural language) could be used to develop computer systems that can recognize, understand and generate natural languages, solving in this way the problem between machines and humans, proposed the research paper titled — Children as Models for Computers: Natural Language Acquisition for Machine Learning. Let’s look at some of the research that is underway in this field coming out of universities across the globe: General Inference framework: According to researchers Leonor Becerra-Bonache and Maria Dolores Jiménez-López, Grammatical Inference (GI) – is a subfield of Machine Learning that deals with the learning of formal languages from a set of data and GI can provide a good theoretical framework to understand how children process and acquire languages. How GI does that is by providing a number of alternative learning algorithms — Query learning model & PAC learning model. GI presents a scenario that is close to real life – there is a teacher and a learner wherein the teacher provides information about a language to the learner and the learner infers the grammar for that language. MIT SpeechHome Corpus: MIT researcher Deb Roy went a step ahead and put together a Human Speechome Project with the goal of making a comprehensive and unbiased record of one child’s development at home. The data provided a basis for further study and the researcher developed a robot that could learn from “show-and-tell” interactions with a human teacher. Given a number of visual presentations of objects paired with spoken descriptions of the objects, the robot learned to (a) segment continuous speech in order to discover spoken words units (b) form visual categories of object shapes and colors (c) learn semantically appropriate associations between speech labels and visual categories. The research stated child’s language acquisition process could advance the understanding of language acquisition through cross-disciplinary methods that bring together the human sciences with computational sciences and design. AI\/LT Lab at Flinders University: One of Australia’s leading universities, Flinders AI\/LT lab is playing a major role in understanding the grounded syntax and semantics in which the computer\/system\/robot really understands what is being talked about. Children learn grammar and meaning in an unsupervised way by learning patterns in context, and this is also emulated in the lab’s AI systems. For example, there are projects which by combining additional enhancements such as camera, microphone input enables lip-reading to be used to improve speech recognition under noisy conditions. It also allowed researchers to track more expressions and emotional content and synthesizing facial expressions. Biometric signals are also picked up as inputs to validate theories and models of language, learning and emotion. Why are tech giants battling over voice? Amazon Alexa Echo family With a shift in human-computer interactions, big tech giants like Google, Facebook & Amazon are plowing in billions of dollars to advance natural language understanding in their voice-powered devices. Companies are aggressively expanding their teams working on conversational interfaces. The rise of Alexa-powered speakers and Google Home have taken voice assistants from smartphones to millions of households where the devices are delivering high-level accuracy in conversational AI. An industry research cites that tech behemoths Microsoft, IBM, Facebook, Google and Amazon have all inked previous NLP transactions and have exhibited a heightened interest in independent NLP platform providers. Google was the first to make the acquisition of API.AI, which provides developers with access to NLP capabilities. Earlier this year, SAP acquired French bot building platform Recast.ai earlier this January, while last year in January, Redmond giant Microsoft bought Maluuba, a Montreal-based company that uses deep learning to develop natural-language understanding.","excerpt":"Can language be taught to artificial intelligence? Given that “machines are capable of doing any work man can do” researchers are working hard on long-standing projects to train machines so they can talk just like us. But unlike identifying images, words are grounded in sense and representation. Jonathan Mugan, co-founder & CEO at DeepGrammar, an […]","categories":["IT Services"],"tags":["AI Language","Natural Language Understanding","NLG"],"author_name":"Richa Bhatia","publish_date":"2018-03-15T05:34:10","publication_year":"2018","word_count":917,"keywords":["Natural Language Understanding","Go","API","artificial intelligence","machine learning","OpenAI","AI","chatbots","NLP","AI Language","deep learning","NLG","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","NLP","OpenAI","chatbots","R","Go","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/can-artificial-intelligence-learn-from-a-childs-language-acquisition-process\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":59882,"title":"Ways To Learn Data Science While Working From Home","content":"With COVID-19 adversely spreading, the only option to control the pandemic is going into a zone of self-quarantine. Keep that in mind, most of the organisations have ensured that the employees work from home to safeguard themselves and others around from getting infected with Coronavirus. Although these are tough times, work from home does provide an opportunity to learn data science since a lot of time is saved. But, when quarantined at home, how can one learn data science? In this article, we will try to explore a few points that might help one learn data science while being stuck at home. Online Courses With social distancing a major guideline to follow at these times of peril, the idea of joining a full-time course goes right out of the window. The best option one can opt for is to register with a traditional online course in data science. It is true that an individual must have a lot of work even though at home, online courses are feasible options since a lot of time gets saved, which usually goes on waste due to unwanted hurdles such as getting stuck in a traffic jam. Online courses come with a variety of schedules, and one can attend them as per their convenience. The courses are designed with quality content and are available in different formats and levels, which one can choose as per their educational background and skill sets they possess. But it does have some drawbacks in a few courses where contents are limited and can be accessed only if one completes the current level. Also, in case a learner is facing a problem, immediate suggestion or help is not present quickly. One can click on the link to know about some free online courses. Massive Open Online Courses (MOOCs) When we talk about MOOCs, there is a need to lay down clear differentiation between MOOCs and online courses or confusion often mounts up. MOOCs focus more on context rather than content, and it’s dynamic building up of context around content makes it unique. MOOCs are designed on the principle of micro-learning with no learning required to go beyond 10 to 12 minutes unless the topic really needs a detailed understanding. MOOCs are the answer to the drawbacks of traditional online courses. MOOCs give a learner the liberty to choose an online course in data science and learn at the pace they want to. All the materials can be accessed at any time irrespective of the current level a learner is in. The courses can be accessed from anywhere at any time, making them extremely remote-friendly in nature. Google has its own MOOCs, such as “Learn with Google AI” along with other crash courses in machine learning that are available to the public for free. Microsoft too has its “AI Learning Track” that provides free training courses to everyone. The courses come with a number of exercises, visualisations and instructional videos. To know more about MOOCs courses, feel free to read our article by clicking on the link, and to select the best courses, click here. Books It is essential to understand that data science is a vast field and is not just about computing. It covers several subjects such as mathematics, statistics, probability, programming and more. Due to this reason, books are the best companion that can provide a holistic view of data science. Be it a compilation of various topics or a detailed explanation of one topic, books are available, and some of them are available for free. Moving on, books have been curated by different authors for different kinds of readers, so books are a good-to-go option for beginners as well as professionals. For beginners who are wondering which book, to begin with, here’s a link that might be of help and professionals can increase their knowledge with these books. Webinar Webinar serves as an opportunity to be a part of events to educate or even educate about languages, tools and other topics related to AI, ML and data science. Although physical appearances are not associated with a webinar, interacting opportunities are present in ample options such as chat, poll, survey, test, call to action and Twitter. One can easily access these webinars on their desktop, tablet or smartphones, thanks to audio and video feed. For those who are in search of data science webinars that are about to take place in a matter of a few days, here’s a link. Online Community Online communities can help a person learn a lot about the field of data science. One can begin by posting their queries on platforms like Quora or Reddit, which are often answered by experts from the field. Twitter and Linkedin play an important role as it helps in connecting with mentors and experts from the field who can give an insight on a wide variety of topics. Moving on, Kaggle and MachineHack are a must for anyone who is interested in the field of data science. From sharing projects or dataset to the public for open criticism and feedback to training models and picking up programming languages, these two platforms house more than millions of developers who are ever-ready to guide someone in need. Last but not least, online hackathons are another place, which gives a learning experience by collaborating with designers, developers and subject matter experts to build a prototype.","excerpt":"With COVID-19 adversely spreading, the only option to control the pandemic is going into a zone of self-quarantine. Keep that in mind, most of the organisations have ensured that the employees work from home to safeguard themselves and others around from getting infected with Coronavirus. Although these are tough times, work from home does provide […]","categories":["AI Highlights"],"tags":["covid-19","Data Science","Learn Data Science","MOOCs","Work from Home","Work from Home Tips"],"author_name":"Rohit Chatterjee","publish_date":"2020-03-25T14:00:42","publication_year":"2020","word_count":896,"keywords":["data science","Go","machine learning","covid-19","AI","programming_languages:R","ML","MOOCs","Work from Home","BERT","GAN","llm_models:BERT","Learn Data Science","Data Science","R","Work from Home Tips"],"extracted_tech_keywords":["AI","machine learning","ML","data science","R","Go","BERT","GAN","llm_models:BERT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/ways-to-learn-data-science-while-working-from-home\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10143334,"title":"Reverse Thinking Could Make LLMs Smarter, More Accurate","content":"Carl Gustav Jacob Jacobi, a noted mathematician, believed that rephrasing problems in their inverse form could simplify finding solutions. “Invert, always invert.” Along similar lines, a recent study by researchers from The University of North Carolina at Chapel Hill, Google Cloud AI Research, and Google DeepMind found that reverse thinking plays an essential role in human reasoning. They developed a novel method to improve the reasoning capabilities of LLMs by incorporating reverse thinking – starting from solutions and reasoning back to the problems. The researchers introduced the concept of Reverse-Enhanced Thinking (RevThink), which has shown performance improvements across tasks involving common, mathematical, and logical reasoning. RevThink involves data augmentation and a multi-task learning objective that mimics human reasoning processes. With this approach, smaller student models learn both forward and reverse reasoning through structured examples provided by a teacher model. Performance That Speaks Volumes In these experiments, RevThink demonstrated an average improvement of 13.53% over its zero-shot performance, where the model attempts tasks it hasn’t been trained on. Compared to traditional methods like knowledge distillation, RevThink outperformed even the strongest baselines by 6.84%. Knowledge distillation typically involves a teacher model guiding a student model, but RevThink’s bidirectional approach appears to yield better results. The researchers express the most striking feature of RevThink is efficiency. The framework can achieve these improvements using just 10% of the correct forward reasoning examples in the training data. In contrast, standard fine-tuning methods require 10 times more data to achieve similar results. This shows the potential of reverse thinking for more efficient AI. This approach succeeds the earlier research about a simple prompting technique – re-reading improves reasoning in LLMs. Very Powerful but very simple Prompting technique.Simply ask the LLM to re-read the question – and this significantly boosts LLM reasoning across diverse tasks and model types. 💡Repeats question input twice in prompt, unlocks latent reasoning potential**Original Problem**… pic.twitter.com\/tVCwCGONHj— Rohan Paul (@rohanpaul_ai) September 19, 2024 The Change in the AI Game The core idea behind RevThink is rooted in how humans approach complex problems. People often switch between forward reasoning, moving from a problem to a solution, and backward reasoning, working from a potential solution back to the problem. This dual-directional reasoning allows for consistency checks, reducing errors and improving understanding. In RevThink, these human-inspired reasoning techniques are applied to LLMs. The process begins with a teacher model, which generates structured examples of reasoning. Each example includes the original question, forward reasoning steps to derive a solution, backward question derived from the solution, and backward reasoning steps to validate the solution. These examples serve as training data for a smaller student model. The student model is trained on three distinct tasks: generating forward reasoning from a question, creating a backward question from the original question, and producing backward reasoning from the backward question. This multi-task learning setup ensures the student model understands both directions of reasoning and can use them to cross-check its outputs. Researchers note that this approach helps the model generalise new tasks and datasets better. Human-Like Reasoning Tanishq Mathew Abraham, research director at Stability AI, wrote on X, “Humans can reason not only from a problem to a solution but also in reverse, i.e., start from the solution and reason towards the problem.” One of RevThink’s most promising outcomes is its ability to adapt to new tasks and data types without requiring extensive retraining. Such adaptability is critical for deploying AI systems in real-world scenarios where inputs often vary. Looking ahead, the researchers plan to release the code for RevThink, enabling other developers to experiment with and expand on the framework. This open approach aligns with broader trends in AI research, where transparency and collaboration drive innovation. RevThink’s success underscores the value of incorporating human-inspired techniques into AI development. By enabling LLMs to think in reverse, researchers aim to open new avenues for improving reasoning, consistency, and efficiency in AI systems.","excerpt":"The core idea behind RevThink is rooted in how humans approach complex problems.","categories":["AI Features"],"tags":["LLMs","reasoning"],"author_name":"Sanjana Gupta","publish_date":"2024-12-11T19:00:00","publication_year":"2024","word_count":648,"keywords":["Go","data augmentation","TPU","AI","LLMs","innovation","RAG","Aim","cloud_platforms:Google Cloud","reasoning","AI research","R"],"extracted_tech_keywords":["AI","Aim","RAG","TPU","R","Go","data augmentation","innovation","AI research","cloud_platforms:Google Cloud"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/reverse-thinking-could-make-llms-smarter-more-accurate\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10138701,"title":"Jensen Huang and Mukesh Ambani to Discuss India’s AI Future at NVIDIA AI Summit in Mumbai","content":"NVIDIA chief Jensen Huang will engage in a fireside chat with Mukesh Ambani, chairman and managing director of Reliance Industries, on October 24 at 10 a.m IST during NVIDIA’s AI Summit India. The discussion will explore NVIDIA’s engagement with the Indian AI ecosystem and how India is embracing accelerated computing as it embarks on a new industrial revolution fueled by AI. The summit, scheduled from October 23 to 25 at the Jio World Convention Centre in Mumbai, will include presentations, workshops, and networking opportunities aimed at connecting industry experts and showcasing advancements in artificial intelligence. Huang’s visit marks a continuation of NVIDIA’s collaboration with Indian companies, including Reliance, Tata, and Infosys, to enhance India’s AI startup ecosystem and reskill the IT workforce. NVIDIA has begun delivering its latest chips, such as the GH200 AI, to Indian partners like Tata Communications and Jio Platforms, which are working on AI-cloud infrastructure. Tata Communications’ Managing Director and CEO, A.S. Lakshminarayanan, confirmed that the installation of NVIDIA’s AI Cloud is underway, with a full launch anticipated by the third quarter of this fiscal year. Reliance Industries has also unveiled consumer-focused AI initiatives, including Jio Brain, Jio AI-Cloud, and Jio Phone Call AI, alongside plans for national AI infrastructure that aims to challenge global cloud giants and reshape India’s AI landscape. At the 47th Annual General Meeting (AGM), Ambani highlighted AI’s pivotal role in the company’s future, showcasing innovations such as JioBrain, an AI service platform designed to improve operations across all Reliance enterprises. Moreover, Jio introduced an AI-powered feature capable of transcribing, summarizing, and translating phone conversations in real time. “Thanks to Jio, India is now the world’s largest data market,” said Ambani. The vast data pool offers Jio exceptional insights, allowing for continuous refinement and scaling of its AI solutions. Jio’s foray into the AI cloud market directly challenges established players like Google Cloud. As part of its AI strategy, Jio launched the Jio AI-Cloud Welcome Offer, providing up to 100 GB of free cloud storage for Jio users starting this Diwali. Recently, Reliance chairman Mukesh Ambani and Akash Ambani, chairman of Reliance Jio Infocomm, visited the TWO US office, where they discussed AI’s evolving role in India with founder Pranav Mistry. “Pleasure hosting Mukesh Ambani and Akash Ambani at our TWO US office and discussing all things AI for India and beyond over chai,” Mistry posted on LinkedIn. TWO, a startup supported by Reliance Jio, has launched SUTRA, a family of cost-efficient, multilingual generative AI models that excel in over 50 languages. These models, available on the ChatSUTRA app, offer speech, search, and visual processing capabilities. The startup secured a $20 million seed fund in February 2022 from Jio Platforms and South Korean conglomerate Naver. Pranav Mistry remarked, “Jio has been one of our key partners for a long time and has invested in us from the very beginning.”","excerpt":"Huang’s visit marks a continuation of NVIDIA’s collaboration with Indian companies, including Reliance, Tata, and Infosys, to enhance India’s AI startup ecosystem and reskill the IT workforce.","categories":["AI News"],"tags":["NVIDIA","Pranav Mistry","TWO AI"],"author_name":"Siddharth Jindal","publish_date":"2024-10-17T12:29:40","publication_year":"2024","word_count":477,"keywords":["Go","artificial intelligence","TWO AI","AI","innovation","Pranav Mistry","RAG","Ray","Aim","generative AI","NVIDIA","R","startup"],"extracted_tech_keywords":["AI","artificial intelligence","generative AI","Aim","Ray","RAG","R","Go","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/jensen-huang-and-mukesh-ambani-to-discuss-indias-ai-future-at-nvidia-ai-summit-in-mumbai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":50525,"title":"Why Should You Learn Python For Data Science?","content":"If you’re looking for an exciting new career that offers tremendous growth opportunity, look no further than the data science industry. Today, organizations of all sizes rely on the insights they extract from the data they have to measure progress, make informed decisions, plan for the future, and so on. Data scientists are the people who process and organize the data with scientific methods, algorithms, and other techniques. Daily, they sift through large data sets, extract what matters, and provide businesses with clear, easy-to-understand insights. Learn more about data science and how it’s helping organizations in the following video: With the advancement of machine learning, AI, predictive analytics, data science is becoming a more popular career choice.  While it’s beneficial to know more than one programming language, aspiring data scientists must learn at least one. There are many to choose from, too, including, Java, Python, Scala, MATLAB, and R. Click here to know more details about Data Science with Python Training Course by Simplilearn. As it stands now, Python is one of the most widely used programming languages in the field and most of the data scientists use python for data science. This dynamic language is easy to learn and read, so it’s an optimal choice for beginners. Python enables quick improvement and can interface with high-performance algorithms written in Fortran or C. IT’s also commonly used in data mining, web development, scientific computing, and more. Simply put, the demand for experts with Python skills is on the rise. Why is Python for data science preferred over other tools? In many scenarios, Python is the programming language of choice for the daily tasks that data scientists tackle, and is one of the top data science tools used across industries. For data scientists who need to incorporate statistical code into production databases or integrate data with web-based applications, Python is often the ideal choice. It is also ideal for implementing algorithms, which is something that data scientists need to do often. There are also Python packages that are specifically tailored for certain functions, including pandas, NumPy, and SciPy. Data scientists working on various machine learning tasks find that Python’s scikit-learn is a useful and valuable tool. Matplotlib, another one of Python’s packages, is also a perfect solution for data science projects that require graphics and other visuals. It is called ‘Pythonic’ when the code is written in a fluent and natural style. Apart from that, Python is also known for other features that have captured the imaginations of data science community. Easy to learn The most appealing quality of Python is that anyone who wants to learn it—even beginners—can do so quickly and easily and this is one of the reasons why learners prefer python for data science.  That also works well for busy professionals who have limited time to spend learning. When compared to other languages, R, for instance, Python promotes a shorter learning curve with its easy-to-understand syntax. Scalability Unlike other programming languages, such as R, Python excels when it comes to scalability. It’s also faster than languages like Matlab and Stata.  It facilitates scale because it gives data scientists flexibility and multiple ways to approach different problems—one of the reasons why YouTube migrated to the language. You can find Python across multiple industries, powering the rapid development of applications for all kinds of use cases. Choice of data science libraries Another key benefit of using python for data science is that python offers is access to a wide variety of data analysis and data science libraries. These include, pandas, NumPy, SciPy, StatsModels, and scikit-learn.  These are just some of the many available libraries, and Python will continue to add to this collection. Many data scientists who use Python find that this robust programming language addresses a wide range of needs by offering new solutions to problems that previously seemed unsolvable. Python community One reason that Python is so well-known is a direct result of its community. As the data science community continues to adopt it, more users are volunteering by creating additional data science libraries. This is only driving the creation of the most modern tools and advanced processing techniques available today which is why most of the people are preferring Python for data science. The community is a tight-knit one, and finding a solution to a challenging problem has never been easier. A quick internet search is all you need, and you can easily find the answer to any questions or connect with others who may be able to help. Programmers can also connect with their peers on Codementor and Stack Overflow. Graphics and visualization Python comes with many visualization options. Matplotlib provides the solid foundation around which other libraries like Seaborn, pandas plotting, and ggplot have been built. The visualization packages help make sense of data, create charts, graphical plots. and web-ready interactive plots. Bottom Line There is no denying that the current job market is competitive, as the Bureau of Labor Statistics recently reported. If you’re looking for a stable industry that isn’t going anywhere anytime soon, data science is an excellent choice. But, choosing a successful industry is only half the battle when it comes to job security. There is also a competition to consider, and it’s important to remember that oftentimes, many qualified candidates competing for the same job opening. One of the best ways to ensure you stand out to recruiters and employers is to have the right credentials. Earning your certification in Python with data science or other relevant field is a surefire way to get your resume noticed by the right people. Get started today!","excerpt":"If you’re looking for an exciting new career that offers tremendous growth opportunity, look no further than the data science industry. Today, organizations of all sizes rely on the insights they extract from the data they have to measure progress, make informed decisions, plan for the future, and so on. Data scientists are the people […]","categories":["AI Trends"],"tags":["Data Science","education","Learn Data Science","Python","Python for Data Science"],"author_name":"Prajakta Hebbar","publish_date":"2019-11-25T11:00:00","publication_year":"2019","word_count":935,"keywords":["data science","scikit-learn","NumPy","machine learning","AI","education","Python","Seaborn","analytics","Python for Data Science","Learn Data Science","Data Science","Matplotlib","predictive analytics","Pandas"],"extracted_tech_keywords":["AI","machine learning","data science","analytics","scikit-learn","Pandas","NumPy","Matplotlib","Seaborn","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/why-should-you-learn-python-for-data-science\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10088893,"title":"Microsoft&#8217;s AI Advantage Puts Pressure on Salesforce, Zoho to Innovate","content":"Seems like the heyday of enterprise software platforms like Zoho, Salesforce, IBM, and others are soon to be over. Microsoft recently launched its Dynamic 365 Copilot, an AI-powered assistance for both customer relationship management (CRM) and enterprise resource planning (ERP) functions. Dynamic 365 Copilot has a suite of AI tools designed for a range of business functions, including sales, service, marketing, operations, and supply chain. This way, it burst into the scene with much fanfare and posed a legitimate threat for others in the industry. Granted some of the its competitors are still figuring out how to handle the safety and moral concerns of this tech before they dive in, there is something that already gives Microsoft an edge: Its partnership with OpenAI. The partnership – in which Microsoft owns a 49% stake in OpenAI – puts its competitors in a “do or die” situation. Either innovate, or, if you only choose to rely on OpenAI’s API, then forever be in the shadows of Microsoft. Salesforce hitting back This is not the first time that Microsoft has expressed its interest in providing CRM in the cloud, which companies use to help manage their self efforts. About eight years ago, the company made a $55B bid to acquire Salesforce. Unfortunately, the talks fell through. What could have been a setback then, might be an opportunity now for Microsoft to catapult itself into this segment with generativeAI capabilities. Alongside Microsoft, there is also SambaNova Systems that will be launching its enterprise suite which will be accessible through APIs and integrate with existing business processes. Not to say that others are a lost cause. It will be interesting to see how companies like Salesforce and others compete with Microsoft on the AI pair programming front. Only last month, Salesforce announced EinsteinGPT, a ChatGPT-like chatbot with abilities to “generate leads and close deals”. EinsteinGPT blends both Salesforce’s and OpenAI’s technology, enabling it to generate emails that sales personnel can send to customers, craft responses to received emails, and create “targeted content” for marketers. The California-based company also announced the launch of a new $250 million generative AI fund to focus on responsible generativeAI. The fund will be invested in companies like Anthropic, Cohere, Hearth.AI, and You.com. “I don’t think that Microsoft is positioned to take back huge market share from Salesforce in the CRM space. The announcement of Einstein GPT, and the deal with OpenAI, will give generativeAI proponents something to hold on to if they are Salesforce customers,” Vernon Keenan, Senior Industry Analyst at SalesforceDevops.net, told AIM. However, it is still unknown if this would be a premature launch to show an advantage against Microsoft or did Salesforce really crack something here. “The problem is that none of the Salesforce products demonstrated today are shipping. We shall see how it goes, since there is now huge pressure for Salesforce to deliver because Microsoft has shipping versions of Copilot in the market today,” Kennan added. What about the rest? Salesforce has made its move. Now, it’s time for Zoho, IBM, Freshworks, and the rest of the lot to reveal their cards. There are little-to-no reports on how these companies are dealing with the GenerativeAI challenge. They have been fairly secretive about their intentions—as well as “extra cautious”, if one may put it like that. Zoho, for instance, is known for taking an increasingly privacy-first approach. Zoho’s Ramprakash Ramamoorthy had previously told AIM, “At Zoho, we strongly believe privacy is more than just a feature and ensure a privacy-first approach in whatever we do—from building our own data centres to a homegrown AI stack”. Prashanth Krishnaswami, Global Head of Market Strategy and Thought Leadership – CX at Zoho Corp told AIM that the company has built generative AI services across Zoho CRM, Zoho Desk, and other Customer Experience (CX) solutions. “Given the scope of the technology, we are doubling down on our investment and R&D efforts for AI and we expect to release more solutions in various phases through 2023 and 2024,” he added. Likewise, when AIM tried to reach out to IBM to understand what they are doing to combat the impending challenge from Microsoft, they declined to comment. Nevertheless, enterprise SaaS with their own AI-GPTs will soon be a reality. OpenSource ≠ equal access for all Meanwhile, there is another caveat to where the industry is going at the moment. With ChatGPT API available for everyone to access, companies have the chance to leverage upon the “democratisation” of AI, while also using Microsoft Azure to train and deploy their AI models. For example, several reports suggest that EinsteinGPT might also be powered by OpenAI’s API. Wait, is Microsoft building its own competitors? Although it may seem like it, this may as well be another form of gatekeeping. Having complete access to OpenAI to continuously improve and build on its model, Microsoft has the power to limit what its competitors can access. Thereby, strictly building a monopoly on the enterprise software division. In addition, the same “democratisation” also means that other large enterprises will now face competition from a bunch of up-and-coming startups who will utilise the ‘integrate API and ship’ model of product development. Known as ‘FastSaaS’, the new model leverages the power of generative AI and low\/no code to significantly lower the barrier to build software products. We initially saw that with OpenAI’s GPT API, which caused quite a stir in the tech world. It’s already been integrated into a bunch of amazing writing and productivity tools, but now the ChatGPT API is causing an even bigger furore.","excerpt":"There is something that already gives Microsoft an edge: Its partnership with OpenAI.","categories":["Global Tech"],"tags":["IBM","Salesforce","zoho"],"author_name":"Ayush Jain","publish_date":"2023-03-08T17:03:28","publication_year":"2023","word_count":926,"keywords":["Anthropic","zoho","ChatGPT","Go","OpenAI","AI","R","RAG","Aim","generative AI","IBM","Salesforce","Azure"],"extracted_tech_keywords":["AI","generative AI","ChatGPT","OpenAI","Anthropic","Aim","RAG","Azure","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/microsofts-ai-advantage-puts-pressure-on-salesforce-zoho-to-innovate\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":61401,"title":"Five Of The Biggest Challenges Faced By Cloud Analytics In 2020","content":"The Cloud Analytics market globally has been projected to grow by $39.1 billion, with a compound growth rate of 7.6%. While many companies view these services as economical, they also choose cloud analytics because it makes it easier for them to handle and process massive volumes of data from different sources. It offers real-time information while providing excellent security. So, it comes as no surprise when we read that 91% of the industry say that analytics should be moved to the cloud at a faster rate. While many analytics companies look to adopt cloud and search for the cheapest cloud platform or a way to optimize their network costs, they still battle a lot of problems: Security Security is a significant factor when it comes to cloud analytics and computing challenges. Security has been the most significant concern because it involves a lot of data. Analytics deals with a lot of private data, so the cloud-based software that is used should have built-in flexibilities that allow it to work efficiently with popular security tools. Security is such a big concern that software vendors and organizations have to always be on the lookout for the latest trends and new designs for their software to work with different security protocols. Implementing a Functioning Cloud Analytics Application Finding an application which can be efficiently implemented and functions properly is one of the most important challenges when it comes to cloud analytics. The software has to run on almost all the operating systems or hardware configurations. Ideally, the analytics software should not be operating under a predefined processing strategy, and it should be capable enough to distinguish between different processing scenarios and dynamically choose how the analytics should be executed without moving a lot of data around. This is not something that is achieved easily, and a lot of organizations struggle to find an appropriate software solution. Maintainability & Governance IT departments in an organization generally advise business users not to install software they cannot control or support. Also, most apps which are critical to the process, need 100% uptime, which gives no scope for the server that is being shut down to replace or upgrade it. Maintenance can be made more accessible by building redundancies into the software so that duplication of control and processing exists. Managing The Cloud Costs Companies nowadays struggle with controlling their cloud spending. In fact, controlling spending is as important to them as cloud security. Nowadays, the mistake of cloud spending is committed not only by beginners, but also by companies who have been using cloud services for a long time. It was estimated in a report that in the year 2020, wasted cloud spending could exceed $17.6 billion. As cloud computing grows, companies are still struggling to manage their cloud spending. Assess Total Ownership Cost Many online vendors offer services which appear cheap at first, but when it comes to using the analytics results, the costs do not seem reasonable anymore. What is presented as a low-cost solution at the front end represents accumulated costs for data storage, bandwidth, database access, numbers of users, memory allocations, row-level scoring, and many tasks along with other resources. To assess the profitability, the entire lifecycle of the analytics has to be quantified and evaluated, so that hidden costs can be avoided. This will allow cloud users to know whether the technology they are using is cheaper than using software on a server.","excerpt":"The Cloud Analytics market globally has been projected to grow by $39.1 billion, with a compound growth rate of 7.6%. While many companies view these services as economical, they also choose cloud analytics because it makes it easier for them to handle and process massive volumes of data from different sources. It offers real-time information […]","categories":["AI Features"],"tags":["big data trends and challenges","cloud analytics","Cloud Computing in IT Industry","latest in cloud computing technology","latest technology in cloud computing","private cloud"],"author_name":"Sameer Balaganur","publish_date":"2020-04-13T10:00:00","publication_year":"2020","word_count":572,"keywords":["Go","latest in cloud computing technology","latest technology in cloud computing","programming_languages:R","AI","cloud analytics","cloud computing","Cloud Computing in IT Industry","R","programming_languages:Go","RAG","private cloud","analytics","GAN","big data trends and challenges"],"extracted_tech_keywords":["AI","analytics","RAG","cloud computing","R","Go","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/five-of-the-biggest-challenges-faced-by-cloud-analytics-in-2020\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10118958,"title":"AWS Brings Meta’s Llama 3 Models on Amazon Bedrock","content":"AWS has announced the general availability of Meta’s recently released new generation of LLM, Llama 3, in Amazon Bedrock. The 7B model outperforms Gemma and Mistral on all benchmarks, and the 70B model outperforms Gemini Pro 1.5 and Claude 3 Sonnet. The Llama 3 models are tailored for diverse AI applications and are presented in two main configurations: the Llama 3 8B and Llama 3 70B. These variants are specially designed for tasks ranging from text summarisation to complex language translation. Meta is also developing even more powerful models, potentially exceeding 400 billion parameters, that will support features like multimodal inputs and multilingual capabilities, greatly expanding the use cases of these models. Expanded Capabilities of Amazon Bedrock Additionally, Amazon Bedrock is now equipped with new features for customers exploring generative AI. Custom Model Import: This new functionality allows users to integrate their unique models into the Bedrock environment. This integration supports a wide range of AI applications by simplifying the development process and reducing operational overhead. Model Evaluation: This tool is now generally available and accelerates the model selection process. It enables organisations to assess and compare the effectiveness of different models on Amazon Bedrock, thus optimising their AI strategies based on performance, accuracy, and suitability for specific tasks. Guardrails: Designed to be responsible for the deployment of AI applications, this feature helps users implement essential safety measures. It effectively filters out undesirable content, aligning with organisational policies and ethical AI practices. Amazon Bedrock is leveraged by various enterprises for its scalability and versatility. For example, airline Ryanair enhances crew operations by using Bedrock to access vital regulatory information and procedural guidelines quickly. In the healthcare sector, Netsmart is making significant strides in improving the efficiency of clinical documentation, which facilitates better patient care and faster service delivery. On the other hand, New York Stock Exchange (NYSE): The NYSE employs Bedrock to decode and simplify vast amounts of regulatory documentation, making critical information more accessible and understandable for stakeholders.","excerpt":"Apart from Llama 3, Amazon Bedrock now includes new features for its customers explpring generative AI.","categories":["AI News"],"tags":["Amazon Bedrock","AWS"],"author_name":"Shritama Saha","publish_date":"2024-04-24T14:31:53","publication_year":"2024","word_count":329,"keywords":["Gemini Pro","Amazon Bedrock","AWS","AI","Modal","Scala","RAG","ViT","generative AI","GAN","R"],"extracted_tech_keywords":["AI","generative AI","Gemini Pro","RAG","AWS","R","Scala","GAN","ViT","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-brings-metas-llama-3-models-on-amazon-bedrock\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":23739,"title":"Top 5 Machine Learning Stocks To Invest In 2018","content":"A market leader in the machine learning, long-term investors are betting on Nvidia The latest Facebook\/Cambridge Analytica data scandal has sparked fear among investors, with reports about several stocks plunging amidst the public backlash. The once-leading high growth technology stocks — Facebook, Amazon, Netflix and Google, better known as FANG — are being crushed by smaller competitors. Tech companies are coming under fire because of increasing regulatory concerns and tightening norms which have shaken up public and investor confidence. With prospects of tighter data privacy legislations looming large in Europe, China and the US, can these machine learning stocks continue their momentum? While the Facebook controversy did sound off an alarm, analysts believe data privacy legislations will only bring more transparency on the table and wouldn’t affect investor confidence. Top Performing Machine Learning Stocks Of 2018: Nvidia: Hailed as a market leader in the machine learning sphere, long-term investors are betting on Nvidia for its heft in machine learning, artificial intelligence and big data — all capabilities “deemed relevant by investors”. Interestingly, Morgan Stanley lifted the Nvidia stock from “equal weight” to “overweight”, deeming Nvidia’s machine learning prowess enough to offset the market enthusiasm in cryptocurrencies, a news website noted. Increased footprint and its position as a market leader puts the leading chipmaker in a strong and favorable position for investors. The company’s major revenue comes from GPUs and the demand for them has increased exponentially over the years. The company’s strong double-digit sales and continued investment in emergent technologies makes it a favourite pick. Alphabet: Perhaps, one of the strongest companies with a history of growth and profitability, the parent company for Google has always surged ahead of its competitors. Known for its strong R&D component and leadership team, a Smarter Analyst report hints that the company spends more on R&D than any of its American counterparts. Besides, Google has always been a futuristic company, leading the way for emergence of next-gen technologies. Also, the Mountain View giant’s leadership in machine learning and artificial intelligence makes it an undisputed leader in AI, making it a favorable stock for investors. While search is the biggest revenue driver for the company, Google is also posting growth from its fast-growing cloud business and machine learning systems. Buzz is that the company will start renting out AI chips soon as well. However, just like Facebook, Google continues to face regulatory threats in the form of antitrust pressures from Government. The company has incurred occasional fines that has had a detrimental effect on its business model. ST Microelectronics: Last we heard about this Swiss semiconductor industry leader was when Chinese e-commerce and cloud player Alibaba and STMicroelectronics were testing a buy zone. Just like pure ML and AI players Amazon, Nvidia, Google, Baidu and Alibaba, STM is reckoned as a leading artificial intelligence and ML stock for powering cutting-edge products. STM’s semiconductors are driving smart homes applications, self-driving car technology and a host of other devices and applications, especially at the edge of the network instead of a centralised data centre. According to a stock prediction report in Walllet Investor, a person’s $100 investment may be up to $306.29 in 2023. Apple: Apple’s latest headline grabbing event – snagging Google’s former head of AI John Giannandrea to strengthen its machine learning and artificial intelligence strategy. Analysts bet that Apple’s AI\/ML development teams tasked with product innovation and improving Siri’s interface in a bid to catch up with Amazon’s Alexa and Google Assistant will improve its position in the market. Also, the company’s R&D budget is expected to increase exponentially, and market analysts believe that Apple will soldier on by pushing AI on the edge of network. Another factor that has raised investor confidence is a three time rise in ML job postings that reportedly mention Siri. Also, the recent hiring of ex-Google chief to improve Siri’s natural language understanding has boosted investor confidence which seems the key to Apple’s iPhones. As Apple continues to bring more ML and AI to iPhones, analysts are bullish Apple’s stock will surge. Amazon: Dubbed as the high margin and cash machine by a US news website, Amazon’s major sales come from its revenue generating component AWS and its e-commerce platform. Another component that enjoys high market confidence is Alexa, its virtual assistant technology, baked into the company’s Echo smart speakers, and has sold to millions across the globe. According to Heidi Pozzo, founder of Pozzo Consulting, AWS is clearly carrying the earnings of Amazon, but interestingly, the company enjoys investor confidence from expanding into machine learning and AI. Holiday sales record Echo Dot as the best-selling product and the company expanded its Echo line of speakers into new markets. It has recently been launched in India.","excerpt":"The latest Facebook\/Cambridge Analytica data scandal has sparked fear among investors, with reports about several stocks plunging amidst the public backlash. The once-leading high growth technology stocks — Facebook, Amazon, Netflix and Google, better known as FANG — are being crushed by smaller competitors. Tech companies are coming under fire because of increasing regulatory concerns […]","categories":["AI Trends"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2018-04-16T10:39:46","publication_year":"2018","word_count":788,"keywords":["big data","Go","machine learning","artificial intelligence","AWS","AI","ML","Git","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","AWS","R","Go","Rust","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-5-machine-learning-stocks-to-invest-in-2018\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":43267,"title":"Behind The Code: A Developer Who Believes That Practice Is More Important Than MOOCs","content":"For our developer series Behind The Code, we get in touch with some of the brilliant minds from the developer community. This time, we talked to Shambhu Nath Singh, who is working as a Senior Software Engineer at Spectrum7 Technologies. The Beginning A BE-CSE graduate, Singh developed an interest in coding when he was in high school. He used to write basic programs such as adding two numbers and try to understand how it’s being executed. As time passed, his interest in programming grew. “I was actually much curious about how smartphone apps are built. So I started digging more and more and started building small mobile applications,” said Singh. “Eventually, it became a fun activity for me. I would build apps and install into my friends’ phones and test them,” he added with a laugh. But the journey to becoming a software developer wasn’t a cakewalk. Singh started his gig as a professional software developer when he was still pursuing his engineering. He was managing his work, his final year project and he was preparing for his final exams. “I used to stay awake for a long time at night, studying and watching tutorials. And I would next morning I used to implement them in the office,” said Singh. However, the hustle was worth it. Today, Singh is leading a team of software development. His role at Spectrum7 also includes the creation of architectures of new products, designs of back-end and front-end, analysis of technologies which can fit the requirements and sometimes, resource-wise effort estimations. “Today, I can see the difference between me before joining the company and me after joining. Every day an exciting journey,” Singh added. Knowledge And Tech Stack When asked about how he upgraded his skills and knowledge, Singh said, “Well, honestly speaking, I didn’t go through any particular MOOCs. I am very good at digging stuff on the internet, so I used resources available on several websites. Also, I have spent enough time watching tutorials on YouTube. Further, he always stated that he used to practice a lot. He believes that when you learn a new program or get a problem statement that needs to be solved by coding, work on it with complete focus, and keep trying until and unless you execute and run it without any errors. Looking into the tech stack of the Spectrum7 developer, most of Singh’s projects are based on Node.js, AngularJS and NoSQL database. He believes that Node.js projects are really powerful in terms of performance and real-time activities. Talking about tools, he uses Web-Storm, Atom etc. He also uses Eclipse as it provides better structures for file systems. Another tool from Singh’s stack is the express framework — he reckons that it is useful in terms of data communication between back-end and front-end. Talking about Java, Singh said, “one of the most powerful tools is Spring MVC framework with Hibernate ORM tool — it gives better flexibility to the projects.” Singh’s Take On DevOps DevOps as a technology domain is said to have a much longer life. “It is a key factor in the IT sector — it not only helps developers to coordinate with the team but also help in executing processes to complete the software building at a faster pace.” Talking about the future, Singh said that the industry is extensively heading towards automation. And why not? The RPA technologies have reduced the cost, maintenance and the resources to develop software. So, the transformation is real. “If one wants to be a DevOps specialist then, you must keep yourself updated with the latest technologies across several industries and how they are going to be integrated with other platforms, Compatibility is a critical aspect, and platforms with higher compatibility will definitely be a success,” Singh added.","excerpt":"For our developer series Behind The Code, we get in touch with some of the brilliant minds from the developer community. This time, we talked to Shambhu Nath Singh, who is working as a Senior Software Engineer at Spectrum7 Technologies. The Beginning  A BE-CSE graduate, Singh developed an interest in coding when he was in […]","categories":[],"tags":["Behind The Code","developer","MOOCs"],"author_name":"Harshajit Sarmah","publish_date":"2019-07-24T16:00:28","publication_year":"2019","word_count":630,"keywords":["Go","AI","RPA","MOOCs","automation","developer","ViT","SQL","CLIP","Behind The Code","DevOps","R","Java"],"extracted_tech_keywords":["AI","R","SQL","Go","Java","DevOps","CLIP","ViT","automation","RPA"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/behind-the-code-a-developer-who-believes-that-practice-is-more-important-than-moocs\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":22719,"title":"How Andhra Pradesh Is Emerging As India’s Blockchain Hub","content":"Cryptocurrencies have become a hot theme of discussion since last year and it has taken the world by storm. India too has got on the bandwagon, after which the Narendra Modi-led government had to intervene to curb cryptocurrencies in India. Though the government may have killed the cryptocurrency party in India but the blockchain technology did not fail to catch the fancy of the government. The blockchain technology started from the banking and finance sector. Last year, India’s first blockchain consortium, Bank-chain, was launched for banks. The blockchain bug has bitten state governments and private companies as well. State government and private firms are increasingly exploring blockchain for improved governance and ensuring transparency. India’s First Blockchain Hub After changing Hyderabad into an IT hub which houses India’s Google and Facebook headquarters, Andhra Pradesh chief minister Chandrababu Naidu now aims to transform Vizag to a world-class fintech ecosystem bringing government, academia, corporates, investors, entrepreneurs together. In 2016, Naidu announced an ambitious FinTech Valley Vizag plan. That makes Andhra Pradesh the first state in the country to introduce pilot projects for the departments of civil supplies and land records in order to protect data of subsidies and land ownership from cyber attacks. Similarly, the technology is used in transport department to streamline titles of the vehicles. “I want to make Visakhapatnam not only India’s capital for FinTech or blockchain technologies, it will be centre of excellence for the global community,” Naidu said at a blockchain business conference held last year. The FinTech Valley Vizag aims to create 5 lakh jobs by 2020. As of 2017, the Valley Vizag has created over 5,500 jobs and has attracted $900 million in investment. Naidu also plans to make the Andhra Pradesh a cashless state. Soon after demonetisation, he has been encouraging people to adopt cashless models of transactions. The state had launched the AP Purse mobile application, linked with 13 mobile banking and 10 mobile wallets, which are also linked with Aadhaar. Dealing with Land Fraudsters Land ownership system is grappling under fraudsters and people often fear being duped with fake land certificates. Land records in most states date back to the colonial era. The disputes over titles often end up in courts. In fact, over 66 percent of civil cases in India are property-related disputes. According to McKinsey Global Institute, distortion to India’s land markets is a barrier to faster growth, accounting for 1.3 percent of lost GDP growth every year. To deal with such crisis, the state government wanted to use the technology in managing land records first. Putting India’s land records on blockchain would help in reducing fraud, increase efficiency and in return boost economic growth. It will also lessen the administrative hassle of registration and title transfer. Once the data on lands or real estate transactions are on blockchain all the parties can involved can track the deal. The technology works by creating public ledgers of all the transactions, replacing a mass of overlapping records with a simple database. Partnership With Firms In collaboration with a Sweden-based startup ChromaWay, AP government is building a ledger system tracking digital information that will allow people to collatarise property. To increase the efficiency further, it also partnered with a Visakhapatnam-based firm for dealing with land records. The state government has so far secured more than 1,00,000 land records through Zebi Data. “We will authenticate the credentials of users, allow them to access the records and give them a certificate. No one can tamper with the database. The buyers, too, can access relevant information on registering with their credentials,” Babu Munagala, founder and CEO at Zebi Data India said. Vizag has also collaborated with Covalent Fund to develop India-focused blockchain stack. With the help of Velugu Core government will be able to access data through APIs. Later, it will let developers create a range of apps based on the available information. Last year, it has also partnered with KPMG to develop a high potential ecosystem in Vizag. Recently, the state government has signed a MoU with Malaysia-based cryptocurrency player Belfrics Global to set up an academy in Vizag with Gitam University and for application of blockchain in several government operations. Earlier this month, the government also signed a MoU with New York-based ConsenSys to establish the first Indian cohort of ConsenSys Education. This year at Davos, the government of Andhra Pradesh signed a MoU with Hitachi. Hitachi will help establish an online governance platform called the Hitachi MGRM Citizen Lifecycle E-Governance in the state. Preventing Cyber Attacks India was affected by ransomware attack WannaCry last year after the key markets within the country took huge hits. As a result, blockchain would make a huge difference in fostering secure transactions. The blockchain is the only trust protocol that guarantees the safety of all digital and financial assets. The main characteristic of the technology is the data cannot be modified after it is created as a result it can prevent and detect any form of tampering. AP is spearheading the revolution by embracing the technology. The state’s adoption of blockchain technology will help in securing the government data. The state is training police forces on blockchain technology through workshops to prevent cybercrime. To secure the data from ransomware or cyber attacks, the state is establishing a research and development centre for cryptocurrency in association with RC Bose Center for Cryptology and Security at the Indian Statistical Institute. Other initiatives also include specialized centers of excellence in collaboration with Thomson Reuters, Broadridge for blockchain. It is also working towards integrating its own e-governance program and securing its assets on blockchain by 2019. The State also plans to build the largest repository of blockchain use cases in transport, finance and digital security. On A Concluding Note After Vizag’s initial success in blockchain technology made other states follow suit. Telangana, Maharashtra, Karnataka and Kerala have also supported the technology and announced pilot projects to ramp up e-governance transparency. The Central government can take a note out of Andhra’s success to adopt blockchain technology. Bringing blockchain technology into the mainstream will give a boost Digital India initiative, which currently lags far behind expectations. The technology will enable paperless governance with permissions, licences, services and transactions being managed digitally.","excerpt":"Cryptocurrencies have become a hot theme of discussion since last year and it has taken the world by storm. India too has got on the bandwagon, after which the Narendra Modi-led government had to intervene to curb cryptocurrencies in India. Though the government may have killed the cryptocurrency party in India but the blockchain technology […]","categories":["IT Services"],"tags":["ai certificates","Blockchain","Chandrababu Naidu","FinTech"],"author_name":"Smita Sinha","publish_date":"2018-03-17T07:10:35","publication_year":"2018","word_count":1037,"keywords":["ai certificates","Go","API","Blockchain","AI","ML","Chandrababu Naidu","Git","RAG","Aim","Rust","Chroma","FinTech","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","Chroma","R","Go","Rust","Git","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-andhra-pradesh-is-emerging-as-indias-blockchain-hub\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":2484,"title":"Interview &#8211; Srikanth Velamakanni, CEO at Fractal Analytics","content":"Fractal Analytics is the most global analytics provider in the world serving Fortune 500 companies in CPG, financial services, insurance, retail, technology and pharma in over 150 countries. Headquartered in San Mateo, they have offices in New Jersey, London, Singapore, Mumbai, Gurgaon and Dubai. In this interview with Srikanth, we talk more about the offerings at Fractal. [dropcap style=”1″ size=”2″]AIM[\/dropcap]Analytics India Magazine: What are some of the main tenets (philosophies, goals, attributes) of your analytics approach and policies? [dropcap style=”1″ size=”2″]SV[\/dropcap]Srikanth Velamakanni: Fractal Analytics seeks to be the most respected analytics brand in the world.  We believe analytics is critical for business to- 1) make better high-volume decisions to achieve breakthrough business performance, 2) deeply understand and engage customers to earn and inspire their loyalty, and 3) change the game and drive competitive advantage through Big Data and technology . Our core values are demonstrated in everything we do. These values include: client value creation, excellence, speed, innovation, professionalism integrity, and respect and fairness. We also believe taking good care of our employees create an empowering environment to make happy customers. AIM: What according to you is your biggest USP that differentiates your organization with similar sized players in the analytics space in India? SV: What clients like about us is- 1)    we have the broadest global footprint among analytics providers with experience in data and consumer behavior in more than 180 countries and 7 offices around the world; 2)    we have deep industry expertise and leverage cross-industry knowledge to solve problems; 3)    we are building proprietary products and advanced machine-learning platforms and increasing process automation; and 4)    we are building a great place to work because we, are very selective in our hiring process, and have the lowest attrition rates in the industry. AIM: Please brief us about some business solutions you provide to your customers and how do they derive value out of it. SV: Our key solution areas include: 1) Customer Insights and Analytics, featuring Customer Genomics for personalized targeting; 2) Pricing Analytics, featuring Pincer for insurance price optimization; 3) Marketing Analytics for more effective marketing mix, media and promotions management; 4) Forecasting and Business Intelligence, featuring Concordia data harmonization that creates clean, organized and consistent data as inputs into visual storytelling dashboards and predictive models; and 5) Analytical Centers of Excellence, bringing an experienced team of data scientists, analysts and consultants to help in-house teams scale analytics, drive implementation and foster a data-driven decision culture. Fractal Analytics has enabled more than 200 clients in consumer goods, retail, financial services, insurance, technology, telecom, pharma and government to make better data-driven decisions with analytics. We help companies answer questions like: “Where can we make more effective use of our marketing budget?” “Which product features or brand attributes should we target to which customers through which communication channels?” “What’s the best pricing and promotion approach to increase consumer loyalty while maintaining profit targets?” Some exciting business problems and results that excite us include: 1)    Optimized pricing decisions for over $100 billion in revenue 2)    Forecasted demand for 15,000 country\/categories for a top 5 global consumer goods company 3)    Developed predictive models for more than 200 clients 4)    Increased returns on marketing allocation for advertising and promotions for 5 Fortune 500 CPG companies 5)    Helped a Fortune 500 financial services company increase customer loyalty by understanding and predicting customer response to brand advertising and promotions 6)    Redefined assortment, product promotions and pricing for one of America’s best known retailers 7)    Predicted and reduced crime for state police departments AIM: Where do you see the bulk of your business coming from? Do Indian organizations have the same affinity towards BI\/ Analytics as that of organizations from other regions? SV: Our clients are global Fortune 500 companies, headquartered in the United States with market reach in Asia Pacific, Europe, Latin America, and Middle East\/Africa markets.  We also work with several large financial institutions and insurance companies in India, largely due to the rapid growth in Big Data and emerging analytics talent pool in India. According to recent study by NASSCOM and CRISIL, the Big Data industry is expected to grow from $200 million in 2012 to $1 billion in US dollars by 2015.  While India is in the early stage of the analytics adoption curve, we expect the market will steadily adopt more advanced business approaches. AIM: Please brief us about the size of your analytics division and what is hierarchal alignment, both depth and breadth. SV: Today we are 460 employees and expect to grow to 700 or more employees by September 2013. Fractal Analytics has a flat organization structure and the vast majority of Fractalites are engaged in the client-delivery functions, broadly classified in four tracks – analytics consulting, big data engineering, data scientist and analytics program management. Every track consists of sales, account management, engagement and offshore delivery teams. AIM: What kind of knowledge worker do you recruit and what is the selection methodology? What skill sets do you look at while recruiting in analytics? SV: We are very selective in our hiring approach.  During the past 9 months we received over 10,000 applications and hired less than 2%.  Fractal Analytics is recognized as a recognized recruiter brands among the top colleges in India where we seek engineers, management graduates, statisticians, mathematicians, and economists. To qualify, candidates must demonstrate an analytical mind, inclination to solve problems, a drive for learning new skills, and a passion to serve clients. Candidates are also required to align with Fractal Analytics’ core values. We believe exceptional people with a good education foundation that share our values can be trained to serve.  We set up Fractal Analytics Academy (FAA) to provide a six-week client-ready education on tools, data and modeling techniques, domain knowledge and soft skills training. We are big on continuous learning so new courses and instructors are added to our FAA curriculum regularly to provide all Fractalites with electives and access to new courses for an average of 85 learning hours per person throughout the year. AIM: Would you like to share any example of an Insight that generated a huge positive impact for your clients? SV: Our Customer Genomics solution is changing the way marketers engage with their consumers. Unlike traditional marketing models that segment people into boxes, Customer Genomics is a machine-learning algorithm that adapts its understanding of customers on more than 200 dimensions, and learns from every consumer interaction. It decodes and learns about customers from every shopping transaction, every tweet, and every store visit. Armed with deep customer intelligence, we helped our Fortune 100 retailer client gain unique insights into the kinds of products and promotions that would most resonate with individual customers. During this project we uncovered the following insights: Novice users more likely to buy expensive items with inexpensive accessories 40% of apparel spend among brand neutral shoppers occurs during holidays People who move spend 4 times the average amount weekly 5 weeks prior to moving This retailer is now able to personalize its messages, promotions and product mix to increase sales among its 60 million households. AIM: What are the most significant challenges you face being in the forefront of analytics space? SV: It is widely reported that there is an expected shortage of data scientists with business skills and real-life problem solving that will challenge the analytics industry worldwide before the end of this decade. Fractal Analytics is well positioned to help companies bridge the gap by providing deep expertise, proprietary solutions and highly trained and experienced data scientists, analysts and consultants to solve real-world data problems. [divider top=”1″] [spoiler title=”Biography of Srikanth Velamakanni” open=”0″ style=”2″] Srikanth Velamakanni co-founded Fractal in February 2000 and is the Group Chief Executive Officer. Over the last 11 years at Fractal, Srikanth has worked with Fortune 1000 companies (in consumer banking, credit cards, retail, telco, insurance and consumer packaged goods industries) in institutionalizing analytics in their organizations to understand, predict and influence consumer behavior. Srikanth has MBA from IIM Ahmedabad and BS in Electrical Engineering from IIT Delhi.[\/spoiler]","excerpt":"Fractal Analytics is the most global analytics provider in the world serving Fortune 500 companies in CPG, financial services, insurance, retail, technology and pharma in over 150 countries. Headquartered in San Mateo, they have offices in New Jersey, London, Singapore, Mumbai, Gurgaon and Dubai. In this interview with Srikanth, we talk more about the offerings […]","categories":["AI Features"],"tags":["analytics ceo","analytics leader","Fractal Analytics","Interviews and Discussions"],"author_name":"Дарья","publish_date":"2013-01-22T06:44:48","publication_year":"2013","word_count":1340,"keywords":["Fractal Analytics","Go","API","big data","analytics ceo","AI","RAG","Aim","data engineering","analytics","analytics leader","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","analytics","Aim","RAG","R","Go","API","big data","data engineering","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-srikanth-velamakanni-ceo-at-fractal-analytics\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10004282,"title":"Has OpenAI Surpassed DeepMind?","content":"OpenAI’s GPT-3 is the talk of the town, and the media is giving it all the attention. Many analysts are even comparing it to AGI because of its practical applicability. Initially disclosed in a research paper in May, GPT-3 is the next version of GPT-2 and is 100x larger than it. It is far more competent than its forerunner due to the number of parameters it is trained on, which is 175 billion for GPT-3 versus 1.5 billion for GPT-2. After the successful launch of GPT-3, other AI companies seem to have been overshadowed. One such company which is also a close competitor of OpenAI is DeepMind. The big question is whether OpenAI has now surpassed DeepMind, which rose to fame in 2016 when it produced AlphaGo software that learned how to play the board game Go and grew better than any human player. Musk co-founded the OpenAI research lab in San Francisco in 2015, one year after Google acquired DeepMind. While both the companies are leading the AI charge and trying to move towards artificial general intelligence (AGI), it has called for a tech war. The Commercial Aspect It is the usefulness of GPT-3 which can be commercialised, and OpenAI has launched APIs for a commercial subscription. OpenAI stands the chance of churning profit with their APIs. Open AI also benefits from Microsoft’s collaboration in training the language model using its supercomputer. Microsoft can further help the company in finding business clients, given its incredibly rich enterprise presence. DeepMind exists under Google’s umbrella, hence a little more skewed towards Google. Ever since Alphabet acquired DeepMind, it has been reporting losses, but Google back up is going to keep it fine. Also, it does not have to prioritise on building a product that could be commercialised readily. Instead, DeepMind has been focusing on proof-of-concept where its agents have beaten humans at very complex games using reinforcement learning techniques, including AlphaGo. DeepMind has also moved away from its focus from just gaming (which has been their forte) into health research projects for Google. This way, the company is working on building more commercially-applications AI by using a state-of-the-art baseline for Deep Reinforcement Learning algorithms. DeepMind is expanding its focus from creating AI agents which can compete in games to making AI agents which can have real-world impact, specifically in areas such as biology. The AI Value Of Both Systems GPT-3 can be used by businesses in actually finishing human tasks, making it the most coherent language model. People have used it to write articles, songs, stories, essays, technical manuals and more. The system may also have the ability to help businesses, such as enhance chatbots, write code, design websites, etc. Comparatively DeepMind’s AI doesn’t have many practical applications yet in day to day business operations, but only in niche areas. We know DeepMind concentrates on cognition, RL etc. Nevertheless, Google is using it to improve its products which can have long term implications for the company’s enterprise customers. Why We Should Not Underestimate DeepMind Compared To OpenAI Having understood the practical implications and AI value of both the companies, the capabilities of DeepMind should not be underestimated. As the years go by, Google may probably come up with groundbreaking applications using Deep Reinforcement Learning that DeepMind possesses. For instance, a paper examined DeepMind’s accomplishments thus far in applying AI to predict protein folding, a crucial issue for developing new drugs. In healthcare, the protein folding area is also great for training artificially intelligent agents. Using the Protein Data Bank, a repository of the 3-D structure and genetic makeup of 150,000 proteins, DeepMind’s protein structure-predicting system, called AlphaFold, was trained. In terms of research, both companies deal with Deep RL and have a similar approach to advancing artificial intelligence. But, it may not be fair to compare the two when it comes to their technology as in the case of algorithmic achievements; usually, the synergies are mutual. Even in gaming, we have seen DeepMind has done some incredible things and provided breakthroughs which are comparable to GPT-3. Such advances may have limited media coverage and got considerably less media attention that they deserved. DeepMind’s staff of more than 1,000, which includes hundreds of well-paid PhD graduates and continues to publish academic papers but only a tiny amount of the work gets covered by the mainstream media. Its most famous coverage so far has been the victory of AlphaGo AI agents over human players. Wrapping Up DeepMind can also advance further in NLP and create mega language models which could be used at a massive scale. While it has been focusing on improving Google’s language models till now, DeepMind is now also powering AI agents to perceive dynamic real-world environments, as suggested in a new paper titled AlignNet: Unsupervised Entity Alignment. It can be said that Open AI does look to be a front runner despite the current hype, popularity and usefulness of GPT-3. Despite the comparison, this A vs B approach doesn’t apply to AI research labs. On the other hand, DeepMind is not far behind.","excerpt":"OpenAI’s GPT-3 is the talk of the town, and the media is giving it all the attention. Many analysts are even comparing it to AGI because of its practical applicability. Initially disclosed in a research paper in May, GPT-3 is the next version of GPT-2 and is 100x larger than it. It is far more […]","categories":["Deep Tech"],"tags":["DeepMind","GPT-3","OpenAI"],"author_name":"Vishal Chawla","publish_date":"2020-08-05T18:00:00","publication_year":"2020","word_count":844,"keywords":["GPT-3","Go","API","artificial intelligence","OpenAI","AI","chatbots","RAG","NLP","GPT","R","DeepMind"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","OpenAI","RAG","chatbots","R","Go","API","GPT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/has-openai-surpassed-deepmind\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10045066,"title":"Accenture Thought Leaders Discuss Future Of Data Science At Applied Intelligence Week","content":"At the Applied Intelligence Week held last month, Accenture leaders and renowned data scientists threw light on the possibilities and opportunities in artificial intelligence and data science for a better tomorrow. Applied intelligence lead at Accenture India Sanjay Sharma moderated the ‘future of data science’ forum. The panellists, including Rajamani Sambasivam, Fernando Lucini, Chandrasekharan Rajendran, and Tahney Keith, spoke about hiring data scientists, remote collaborations, demand and supply gaps and more. Ideal data scientist “If you look at the next five years, AI is coming up in a big way across multiple industries. An ideal candidate should know the algorithms, mathematics, code and technical skills. Surely, these are a must. But, in addition to this, candidates should work on their problem-solving skills, develop innovative solutions, and out-of-the-box thinking,” said Sambasivam. Accenture’s chief data scientist Lucini said an ideal candidate must have a great balance between mathematics and the principles behind the tools. “There is a real danger where we are giving in to all these easy-to-use tools, but we do not understand the principles. Building a neural network takes three clicks, but we do not understand what is going on inside,” said Lucini. Rajendran, professor at IIT Madras, said if a candidate has solid knowledge in their domain, data science can be a great value add. Besides this, Rajendran also made a case f0r a candidate’s ability to articulate a business problem better, alongside continually updating their knowledge and differentiating themselves by participating in international level competitions and hackathons. “This should be the benchmark to know where you stand vis-a-vis best in the field, or in the world,” he added. Further, Rajendran said a candidate should have a solid background in statistics, optimisation (both discrete and continuous optimisation), mathematics (linear algebra), and coding. “Being good in AI, machine learning and developing statistical models is one thing, but being able to understand complex business problems is another. Merging these two is where it matters the most. That is a whole different side of the coin,” said Keith, cognitive automaton capability lead at Accenture. According to Keith, the ideal data scientist should be able to translate real-world problems into data science problems they can solve. They should also be able to convert the output of data solutions and communicate this back as insights or actions in a way that is consumable to the business and non-data scientists, he added. Overall, the industry leaders are looking at attributes, core skills, domain expertise, candidates’ ability to learn and scale. Companies are not looking at hiring data scientists based on one skill (be it coding, mathematics, or statistics), but a combination of different skills. Data science in manufacturing Sambasivam believes AI in manufacturing is a slightly different ball game. Since manufacturing has been there for more than a century now, there are many subject matter experts in the field (be it physics, thermodynamics, fluid mechanics, etc.). “Therefore, when a data scientist gets into applying AI into the manufacturing, the first and foremost requirement is the explainability of the results and parameters,” added Sambasivam. He said data scientists need to work on the explainability part, keeping safety concerns in mind. Today, most manufacturers have incorporated level 2 automation into their processes. The ability to control the process becomes very narrow, and when AI is introduced to improve it further, the accuracy is significantly higher. “Sometimes, the features that we want to build from the data may have to be non-linear in nature. So we have to work on that, and understand the process,” said Sambasivam. “Understanding physics, chemistry, thermodynamics, and process dynamics are some of the key requirements for any data scientist who wants to work in manufacturing,” he added. “Data science  is industry-specific. It is a profession, and not a skill,” said Lucini. Lucini believes that enough data scientists are pushing the industry today, and enough scientists are coming out of great universities in India and elsewhere. Perks of collaborating “Data scientists working anywhere need to have a similar set of skills in terms of rigour, correctness, breadth, depth of theory, practice and domain knowledge. Since you are working with different people from around the world, you are getting exposure to different skill sets, strengths and weaknesses,” said Keith. Keith said data scientists learn on the job and are supported by peers and seniors. “Mentoring is just a really big part of how we develop our data scientists as well. It is really exciting to have had the opportunity with an extension to move around and leverage skill sets,” she added. Addressing the supply side In 2003, the government of India started the National Programme on Technology Enhanced Learning, NPTEL, in collaboration with premier institutions like IIT and IISc. NPTEL is an online curriculum development programme in science and engineering at university and research levels. “Many students have joined these courses free of cost throughout the country. In total, 1,000+ courses have been offered by different institutions in the country, and more than 4000 colleges are the beneficiaries of this programme,” said Rajendran. Rajendran said IIT Madras offers data science courses at BSc levels. In addition, IIT Madras is introducing data science courses in engineering disciplines. At the end of the second year, the candidate can choose to pursue a Master’s in data science. “You will get a degree in the parent discipline and a Mtech degree in data science,” he added. He further said the students get to see the applications of data science early on and would be able to apply to the parent discipline.","excerpt":"A data science candidate should have a solid background in statistics, optimisation, mathematics and coding.","categories":["IT Services"],"tags":["AI (Artificial Intelligence)","Data Science","data science and manufacturing","Data Science Hiring","data science latest","data science new","Data science skills","IIT Madras","what is data science"],"author_name":"Amit Naik","publish_date":"2021-08-03T16:00:00","publication_year":"2021","word_count":922,"keywords":["TPU","R","Data science skills","data science","artificial intelligence","data science new","data science latest","RAG","Data Science Hiring","what is data science","Data Science","Go","machine learning","AI","IIT Madras","neural network","data science and manufacturing","automation","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","data science","RAG","TPU","R","Go","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/accenture-thought-leaders-discuss-future-of-data-science-at-applied-intelligence-week\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":36713,"title":"Parameters To Keep In Mind While Using Data Analytics On The Cloud","content":"With the enormous amount of big data that we have for running businesses efficiently in our everyday lives, the cloud comes in as a very efficient solution. But moving data analytics on the cloud has its own set of difficulties. Here are some of the parameters to keep in mind while using data analytics on the cloud. In this article, we will take a look at the migration process as well as the advantages of using cloud analytics. Figure out where to host the data modelling: While adopting cloud technology, a common question of where to host the data modelling is an important one to consider. There model hosting should provide good performance and functionality gains and flexibility benefits. It becomes necessary to identify where to host data modelling. Study the challenges thoroughly: In any migration, many companies focus a ton on the project’s challenges but don’t spend a lot of time learning the shortcuts, resources, and tools to help them out. It is important to understand migration challenges thoroughly to help companies ease the transition to the cloud. Not every tool is for everyone, but by becoming much more familiar with these accelerators, you can potentially save time and money. By having someone focus on understanding this landscape and bringing them to the project team there will be a start on resources that can add value. Learn from the employees: There would be people already in your organisation that must have downloaded some sort of cloud-based analytics and have even begun to use them. Instead of banning them, learn from them and adopt in the entire organisation. It is necessary to need to keep informed about the solutions readily available. It is also important to learn what tools they refer to and what solutions are being provided and how they win against the traditional business intelligence tools of the organisation. Understand the extent to which you rely on Universes: There a unique way to handle data models for every BI environment. For a large organisation, it is likely that there are different front-end reporting tools and data sources. This has to be taken into consideration while moving data analytics in the cloud. Build the first frame: The task of transforming the data, building and testing the model, creating visualisations and then turning the output into action causes an analytical block. Building a basic data frame is important for cloud services. Companies should build a basic data frame on a relatively manageable and familiar dataset, process statistics against it and create analysis out of it. Then start to layer in new data by adding analytics, eventually bringing out new visualisations. The aim is to keep it simple, flexible, and understandable. This helps build consensus internally, increase buy-in from lines of business, and get a faster time to value. Avoid shortcut of a point solution: It is always better to have a start-small approach. The cloud platform is yet to acquire certain capabilities to make it useful. Using shortcuts might in the realm of cloud analytics starting short and having a long way is always better than going for short cuts for a point solution, which may not help. Adoption Of Cloud Analytics Cloud analytics is important for organisations belonging to different sectors to adopt because of the following advantages: Improve product availability: Study buying behaviour to improve product availability and delivery. Study genetic diseases: Test genomic data to better understand the genetic disease and how to offer cures. It can also be used to keep track of a lot of data and identify patterns of disease reporting to improve the availability of medicine and vaccines Improve customer service: Identify patterns in speech, images and videos in order to improve customer satisfaction and improve customer service. Optimise IT costs: Cloud analytics can be used to analyse hybrid cloud infrastructures to improve application performance and optimise the IT costs.","excerpt":"With the enormous amount of big data that we have for running businesses efficiently in our everyday lives, the cloud comes in as a very efficient solution. But moving data analytics on the cloud has its own set of difficulties. Here are some of the parameters to keep in mind while using data analytics on […]","categories":["AI Features"],"tags":["Cloud Data AI","tools"],"author_name":"Disha Misal","publish_date":"2019-03-22T08:57:40","publication_year":"2019","word_count":647,"keywords":["big data","Go","business intelligence","TPU","programming_languages:R","AI","R","Aim","analytics","tools","GAN","Cloud Data AI"],"extracted_tech_keywords":["AI","analytics","Aim","TPU","R","Go","big data","GAN","business intelligence","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/parameters-to-keep-in-mind-while-using-data-analytics-on-the-cloud\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10041179,"title":"Register For This Webinar: Solving The Hunger For Data Science Skills In The Post-COVID World","content":"In April 2020, Microsoft CEO Satya Nadella said the company witnessed two years’ worth of digital transformation in just two months of its third quarter. Not just Micrososft, the pandemic has catalysed the digitisation of organisations across the world. A direct output of this radical transformation is the humongous quantities of data generated across sectors. Now, businesses are waking up to the need to be adept at data acquisition and analytics to survive these trying times. This, in turn, has led to a massive boom in the data science market. In association with Praxis Business School, Analytics India Magazine is organising a webinar to understand this rapidly evolving field of data science and how pandemic impacted the demand for data scientists. The webinar will also shine light on opportunities and challenges for data scientists in the current market. The webinar will cover: Impact of COVID on the demand for data scientistsHow to build a career in data science‘Hot’ data science skills & how to acquire them REGISTER NOW Who should attend? Data science & AI enthusiasts.   Engineering \/ technical college graduatesGraduates in Maths, Science, Economics or StatisticsData science & analytics professionals looking to pivotCandidates taking a break from their career or lost their jobs due to pandemic and wish to pursue a career in dataScience Why should you attend? To understand the dynamics of the data science market.To understand the importance of learning data science skills in the current era.Tips & techniques to build a career in this competitive field.Explore how Praxis can help you achieve excellence in data science. REGISTER NOW Session Speakers: Sudipta Ghosh Sudipta Ghosh is the Partner and Leader, Data and Analytics and Industrial Products Leader at PwC. With more than 20 years of consulting experience in the field of data and analytics and Enterprise Performance Management, Sudipta has been instrumental in successfully setting up the analytics practise at PwC. He has also been involved in setting up advanced analytics propositions using artificial intelligence and machine learning for the CXO suite. Sudipta is a member of the PwC Global Data & Analytics Leadership team for defining analytical propositions and solutions across the PwC network of firms. He has been a visiting faculty at IIM Kolkata and the Vinod Gupta School of Management at IIT Kharagpur, instructing Business Analytics and Enterprise Performance Management courses. Further, he is an advisory council member of the Post Graduate Diploma in Business Analytics conducted jointly by IIM Calcutta, IIT Kharagpur and ISI Calcutta. Sudipta is a gold medalist from IIM Calcutta and a silver medalist from IIT Kharagpur Prof. Charanpreet Singh Prof. Charanpreet Singh is the Founder and Director at Praxis Business School. He has been the driving force behind the highly successful Data Science Program offered by Praxis Business School. Prof Singh has 30+ years of experience in organisations like British Oxygen, Tata Steel, PwC, HP and Praxis. He has professional and academic interests in the areas of education, learning, data science and communication and has been a mentor to aspirants across domains. He is a Chevening Scholar and has done his BTech in Mechanical Engineering from IIT Kanpur and MBA from the University of Iowa in the USA. Date: 11th June 2021 Timing: 6 PM to 7 PM (IST) REGISTER NOW","excerpt":"Praxis Business School Foundation and Analytics India Magazine are organising a webinar to understand this rapidly evolving field of data science and how pandemic impacted the demand for data scientists.","categories":["Deep Tech"],"tags":["Data Science","Data science skills","post covid world","praxis","Praxis Business School","praxis data science programme","praxis data scientist programme"],"author_name":"Sejuti Das","publish_date":"2021-06-03T12:00:00","publication_year":"2021","word_count":542,"keywords":["praxis data scientist programme","data science","Go","artificial intelligence","machine learning","TPU","AI","R","post covid world","Git","RAG","praxis data science programme","analytics","praxis","Data Science","Data science skills","Praxis Business School"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","RAG","TPU","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/register-for-this-webinar-solving-the-hunger-for-data-science-skills-in-the-post-covid-world\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":61142,"title":"Are Too Many Data Scientists Trying To Predict COVID-19 Outcomes In Futility?","content":"Data scientists have been creating a lot of tools that help explain significant questions around COVID-19. One example is dashboards based on COVID-19 cases around the globe. It has helped show active cases, those in the testing phase, information on patient history, etc that provide a window into the overall scenario of the pandemic. There have also been many challenges and hackathons in response to COVID-19, and several data companies are providing free data resources. Kaggle has thousands of posts related to COVID-19. The COVID-19 Open Research Dataset Challenge (CORD-19) dataset on Kaggle contains over 44,000 scholarly articles, and one Kaggle expert Daniel Wolffram has created several widgets that help navigate the current COVID-19 research literature. There are also geospatial trackers of multiple government initiatives built from the work of data scientists, which serve as valuable tools during the pandemic. Explaining The Issue Using hundreds of metrics, data scientists have been trying to predict COVID-19 outbreak. But, are the predictions accurate, given the pandemic is a black swan event with not much epidemiological records in the research literature? Even information relating to its DNA sequencing is new. While data scientists are using geographical cases to predict how COVID-19 will pan out, some professionals and data scientists on social media think the work is not accurate. The issue lies in the fact that epidemiologists have been tracking and predicting the spread of pathogens for decades, long before machine learning professionals and data scientists. Also, data scientists may not have expertise when it comes to the highly complex biological aspects of predicting viral outbreaks. Also, there may also be an issue with the datasets that are being used to create predictive models. “Existing datasets (on COVID-19) are incredibly biased. For example, when calculating the mortality rate, normally we look at the deaths per confirmed case. However, the underlying assumption is that we have captured all of the confirmed cases, which is not true, since we are bottlenecked by the number of tests and only the sickest are diagnosed. For a place like New York, an exponential increase in the availability of testing can also generate an exponential growth curve,” according to Neil Cheng, Senior Data Scientist at Akamai. Wherever There Is Data, There Is Room For Data Science Can data scientists can have a key role in predicting all aspects of the global pandemic, regardless of their experience in biology? This is because most microbiologists and epidemiologists have had little or no training on data analysis, where data scientists can add value. Indeed, there are only a handful of epidemiologists who are also good data scientists with backgrounds in mathematics, computer science, and machine learning. Here, pure data scientists can certainly collaborate with microbiologists and epidemiologists to create better predictive models. The issue is that when such models are created by pure data scientists, who do not realize whether a data set is even helpful or accurate in most cases, it becomes problematic. Wherever there is data, there is scope for data science to make an impact. Of course, data scientists should have some level of domain knowledge so they can effectively analyse and interpret the data. It is not merely about predicting the COVID-19 outbreak. Data scientists could also help create better models on how to optimize the hospital infrastructure, medical supply chain and medical equipment manufacturing process such as ventilators and masks, instead of forecasting the outbreak of something as complex as a global pandemic. Data scientists may uncover unique patterns that may be valuable to those experts by leveraging advanced machine learning techniques. But findings would need to be peer-reviewed, validated, and examined by medical and epidemiological experts as acceptable. Yet, the majority of people downloading COVID-19 datasets may be unqualified to contribute in a meaningful way to save lives, as many point out. This pertains to the complexity of understanding microbiology and epidemiology.","excerpt":"Data scientists have been creating a lot of tools that help explain significant questions around COVID-19. One example is dashboards based on COVID-19 cases around the globe. It has helped show active cases, those in the testing phase, information on patient history, etc that provide a window into the overall scenario of the pandemic.  There […]","categories":["Deep Tech"],"tags":["become a data scientist","Covid Dataset","covid-19","Data Scientist","Data Scientists"],"author_name":"Vishal Chawla","publish_date":"2020-04-07T15:00:00","publication_year":"2020","word_count":645,"keywords":["data science","Go","machine learning","Covid Dataset","covid-19","AI","programming_languages:R","become a data scientist","programming_languages:Go","RAG","Data Scientist","R","Data Scientists"],"extracted_tech_keywords":["AI","machine learning","data science","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/are-too-many-data-scientists-trying-to-predict-covid-19-outcomes-in-futility\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10063298,"title":"Will generative models replace the need for real-world datasets?","content":"Generative modelling is an unsupervised learning method to tease out patterns in input data and output data instances resembling the original data. Generative adversarial networks (GANs) pit a generator model against a discriminator model. While generative models churn out new data instances, discriminator models differentiate between data instances. Both models try to one-up each other to optimise the output. The advances in parameterizing these models using deep neural networks, combined with progress in stochastic optimization methods, have enabled scalable modelling of complex, high-dimensional data including images, text, and speech. Generative models have the potential to produce photorealistic images that look identical to training data. So, if we have good enough generative models, do we still need datasets?  MIT researchers Ali Jahanian, Xavier Puig, Yonglong Tian, and Phillip Isola have investigated this question in the setting of learning general-purpose visual representations from a black-box generative model rather than directly from data. The findings showed that a contrastive representation learning model trained on synthetic data could learn visual representations that compete with, if not outperform, those learned from real data. Deep generative models could offer the most promising developments in AI https:\/\/t.co\/m0OStj1fEV— Stewart Rogers (@TheRealSJR) March 21, 2022 Challenges Training machine-learning models to perform image classification tasks requires a gigantic amount of data. Datasets can cost millions of dollars to create. Additionally, datasets carry biases that can put a damper on model performance. In the ICLR 2022 conference paper titled, “Generative models as a data source for multiview representation learning”, the researchers proposed a method for training a machine learning model that uses a special type of machine-learning model to generate realistic synthetic data that can train another model for downstream vision tasks. Generative models require far less memory to store or share than a dataset. Using synthetic data can help in working around privacy and usage rights concerns. Generative models can also be edited to remove specific attributes, such as race or gender. Generating synthetic data The researchers linked a pretrained generative model to a contrastive learning model. According to Ali, the contrastive learner could instruct the generative model to generate different views of an object and then learn to identify that object from multiple angles. The generative model provides different views of the same thing and helps the contrastive method to learn better representations. Source: Generative models as a data source for multiview representation learning “Given an off-the-shelf image generator without any access to its training data, we train representations from the sample’s output by this generator. We compare several representation learning methods that can be applied to this setting, using the latent space of the generator to generate multiple “views” of the same semantic content. We show that for contrastive methods, this multiview data can naturally be used to identify positive pairs (nearby in latent space) and negative pairs (far apart in latent space),” the researchers said in the paper. The resultant representations either matched or in some cases outperformed those learned directly from real data. “We knew that this method should eventually work; we just needed to wait for these generative models to get better and better. We were especially pleased when we showed that this method sometimes does even better than the real thing,” said Ali Jahanian, a research scientist in the Computer Science and Artificial Intelligence Laboratory (CSAIL) and lead author of the paper. Read more about synthetic data here. Roadmap The paper suggested several techniques for dealing with visual representation learning. Real-world data falls short for learning about corner cases. For example, if researchers are developing a computer vision model for a self-driving car, real-world data would not include examples of a person or an animal running down a highway. Consequently, the model would never learn what to do in this situation. Synthetically generating that corner case data could improve the performance of machine learning models in some high-stakes situations. The researchers also want to improve generative models to create more sophisticated images.","excerpt":"Generative models require far less memory to store or share than a dataset.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI Tool","Data Analysis","Data Mining","Data Science","Datasets","Machine Learning","MIT","Research","researchers","synthetic data","Unsupervised Learning"],"author_name":"Sri Krishna","publish_date":"2022-03-23T10:00:00","publication_year":"2022","word_count":658,"keywords":["Data Analysis","TPU","MIT","Scala","computer vision","AI Tool","R","Datasets","artificial intelligence","Data Mining","Research","Data Science","Unsupervised Learning","Go","machine learning","synthetic data","AI","neural network","Machine Learning","GAN","researchers","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","neural network","computer vision","TPU","R","Go","Scala","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/will-generative-models-replace-the-need-for-real-world-datasets\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":12532,"title":"Re-imagining Opera Browser with Artificial Intelligence","content":"The fifth-largest web browser in the world, Opera, has over 300 million users worldwide, across its various range of products on mobile and desktop combined. The firm accumulates a copious amount of data which could be combined with AI technologies to intelligently offer content to its users. The firm comprehends the extensive application of AI and has announced its plan of integrating the technology into its products. Sunil Kamath, Vice President, South Asia and South-east Asia, Opera Software remarks, “AI will help us be more predictive in providing consumers with what they want. AI will form the very core of our business practice.” Opera invested towards integrating AI, and 2017 could be the year when the firm combines the technology with all its products, viz, Opera web browser on the desktop, Opera Mini on Android smartphones, and Opera Max VPN. The company’s new Chinese parent, KunlunTech has already established a team in Beijing, which collaborates with Opera’s team in Scandinavia, towards designing a new AI platform. The Neon browser, launched recently by Opera, is an offbeat experimental web browsing project. The firm plans to use the browser as a testbed for all its upcoming projects. This will help the firm realize its dream of bringing in AI. “Neon will serve as a testbed for us. This will help us test our new technologies and obtain feedback from the consumers,” notes Sunil. The new parent firm, Kunlun has definitely driven Opera in the directions of leveraging AI. However, Opera’s core technology and product team are still based out of Scandinavia. Additionally, the owner of Kunlun Tech is striving at promoting a culture where Kunlun learns from Opera Software. Being privately held by Kunlun, Opera doesn’t have to face the heat of the stock market. “Being a private company now, we are very aggressive about how we build market share,” adds Sunil.","excerpt":"The fifth-largest web browser in the world, Opera, has over 300 million users worldwide, across its various range of products on mobile and desktop combined. The firm accumulates a copious amount of data which could be combined with AI technologies to intelligently offer content to its users. The firm comprehends the extensive application of AI […]","categories":["AI News"],"tags":[],"author_name":"Amit Paul Chowdhury","publish_date":"2017-02-03T06:32:32","publication_year":"2017","word_count":310,"keywords":["R","RAG","programming_languages:R","AI"],"extracted_tech_keywords":["AI","RAG","R","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/re-imagining-opera-browser-artificial-intelligence\/","complexity_score":2,"technical_depth":4,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167932,"title":"Google’s New AI Model Helps Humans Communicate with Dolphins","content":"Google, in collaboration with Georgia Tech and the Wild Dolphin Project (WDP), has launched DolphinGemma, an AI model developed to analyse and generate dolphin vocalisations. The announcement, made on National Dolphin Day, marks a new step in the effort to understand and potentially engage in two-way communication with dolphins. DolphinGemma is trained on decades of underwater video and audio data from WDP’s long-term study of Atlantic spotted dolphins (Stenella frontalis) in the Bahamas. The model identifies patterns in dolphin vocal sequences and generates realistic dolphin-like sounds. “By identifying recurring sound patterns, clusters and reliable sequences, the model can help researchers uncover hidden structures and potential meanings,” Google said. The AI system is based on Google’s lightweight Gemma models and leverages SoundStream for audio representation. At roughly 400 million parameters, the model is small enough to run on Pixel phones used in field research. It operates as an audio-in, audio-out system, designed to predict subsequent dolphin sounds much like a language model predicts the next word in human text. “Understanding any species requires deep context,” the WDP team said. Their work since 1985 has documented individual dolphins’ life histories and linked vocalisations to specific behaviours. Signature whistles are used to reunite mother and calf pairs, while burst-pulse squawks and click buzzes are associated with fighting and courtship. In parallel, the WDP and Georgia Tech have also developed the CHAT (Cetacean Hearing Augmentation Telemetry) system. CHAT uses synthetic whistles linked to objects like sargassum or scarves. If dolphins mimic these whistles, the system can recognise them and inform the researcher via bone-conducting headphones, enabling real-time interaction. “A Google Pixel 6 handled the high-fidelity analysis of dolphin sounds in real time,” researchers explained. The next iteration, built around a Pixel 9, will integrate deep learning and template matching directly into the device, reducing the need for specialised hardware. DolphinGemma will be released as an open model this summer. Although currently trained on Atlantic spotted dolphins, it is expected to benefit research into other cetaceans. Fine-tuning will be needed for species with different vocal patterns. “We’re not just listening anymore. We’re beginning to understand the patterns within the sounds,” Google stated. The goal of the research is to bridge the gap between humans and dolphins through data-driven insights and shared interaction systems.","excerpt":"DolphinGemma will be released as an open model this summer.","categories":["AI News"],"tags":["Google"],"author_name":"Siddharth Jindal","publish_date":"2025-04-14T19:46:16","publication_year":"2025","word_count":378,"keywords":["Go","programming_languages:R","AI","data-driven","programming_languages:Go","RAG","deep learning","Google","R"],"extracted_tech_keywords":["AI","deep learning","RAG","R","Go","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/googles-new-ai-model-helps-humans-communicate-with-dolphins\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10002141,"title":"Indian Telcos Are Adopting Blockchain To Filter Out Spam: Here’s How It Works","content":"Blockchain technology has now begun trickling down to the consumer and is set to offer a benefit that is sorely needed by many. Prominent Indian telcos Reliance Jio, Bharti Airtel and Vodafone Idea have begun rolling out blockchain solutions to curb spam calls. This is keeping in mind the modern regulatory standpoint by the Telecom Regulatory Authority of India. Keeping in mind consumer interests, TRAI set a framework for access providers to do something about the problem of unsolicited commercial communication. Regulation And Compliance The TRAI published an amendment to the Telecom Commercial Communication Customer Preference Regulation on 19th July 2018 which will be enforced from the end of this month onwards. The regulation was to curb the rise of unsolicited commercial communication to protect customers from spam calls. However, those making the calls still found ways to get into loopholes that the access providers left in their system. Since all telecom service providers in India fall under the purview of TRAI, they are required to adhere to the new norms mentioned by the regulator. The party wanted to make sure that only those registered in the databases can make the calls to prevent fraud. Apart from this, they also wished to record user consent for telemarketers for consumer protection. How Does It Work? The blockchain, at its base, is a database. What makes it different is its ability to not only hold a single version of the truth but also synchronize with multiple parties who might not be ready to work together or agree on the data. This architecture is a perfect fit for an environment such as the telecom market. At its core, the blockchain solution will be a shared ledger of UCC numbers distributed across a telco’s databases. It is also a very transparent system, and will enable a clearer outlook on parties that are looking to exploit loopholes left by the telcos. Tech Mahindra, one of the blockchain service providers in the picture, developed a blockchain solution late last year after the regulations were made public. It was revealed that this was built on Microsoft’s Azure platform, and functions at the “intersection of cloud and blockchain”, according to the National Technology Office of Microsoft India, Prashant Shukla. The system will “ensure a new way of monitoring and enforcing compliance throughout the ecosystem,” he stated. Apart from this, IBM, another of the service providers, is offering solutions for mobile number portability and do not call registry. The VP of IBM Research, Sriram Raghavan, stated that the company had completed the proofs-of-concept and pilots with “major telecom providers” in the country. The solution will be a private distributed ledger that can be accessed by telcos. The ledger will contain the registry of consumers that have forbidden consent to any calls by telemarketers and will be shared across all operators. Apart from this, the company is also offering a solution for better mobile number portability; a solution which works across multiple telcos. At the time, Raghavan stated, “We anticipate that going into the New Year, we’ll start to see blockchain solutions getting rolled out. We have completed proof of concepts and pilots.” More recently, Tanla Solutions, one of the service providers, also confirmed that the projects are being rolled out currently. A spokesperson said, “Deployment is underway and will help in arresting spam and fraud-related commercial communications” Blockchain Adoption Rising The three companies have chosen to partner with different parties to ensure the seamless rollout and maintenance of their respective blockchains. Vodafone Idea has partnered with a Hyderabad-based company known as Tanla Solutions for their deployment across 395 million users. Airtel, with its 325 million users, tied up with blockchain industry leader IBM for their solution. The company has shown previous experience in handling blockchain solutions, and has R&D capabilities to explore optimizations. Jio tied up with Tech Mahindra, who developed the blockchain solution in conjunction with Microsoft. With over 307 million users on the platform, the telco is looking to utilize Microsoft’s Azure platform for its deployment. Post the rollout of all of these networks, India will become the hotbed of one of the biggest blockchain implementations in the world. With over 1 billion customers reached due to this initiative, India will be home to one of the most burgeoning blockchain markets in the world. The World Economic Forum even predicted that the blockchain would account for almost 10% of the global GDP by 2025. With blockchain set to be a trillion-dollar industry, India looks to be cementing its role as a leader in the technology for years to come. The country has also shown its interest towards blockchain in other ways as well, with the National Payments Corporation of India looking to adopt such technologies for better digital payments flow. The forward-thinking mentality of regulators and readiness of participants also shows an atmosphere conducive to growth. It is left to see how the market is going to develop in one of the world’s biggest countries.","excerpt":"Blockchain technology has now begun trickling down to the consumer and is set to offer a benefit that is sorely needed by many. Prominent Indian telcos Reliance Jio, Bharti Airtel and Vodafone Idea have begun rolling out blockchain solutions to curb spam calls. This is keeping in mind the modern regulatory standpoint by the Telecom […]","categories":["AI News"],"tags":["Airtel","Blockchain","Jio","telecom","TRAI"],"author_name":"Anirudh VK","publish_date":"2019-05-28T19:23:16","publication_year":"2019","word_count":828,"keywords":["Jio","Go","TRAI","Blockchain","cloud_platforms:Azure","AI","programming_languages:R","R","ML","programming_languages:Go","Git","RAG","telecom","Airtel","Azure"],"extracted_tech_keywords":["AI","ML","RAG","Azure","R","Go","Git","cloud_platforms:Azure","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-telcos-are-adopting-blockchain-to-filter-out-spam-heres-how-it-works\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10011927,"title":"What Is The Hiring Process For Data Scientists At Genpact","content":"With a surge in AI and ML applications to improve business process efficiencies, there has been an increase in data science hiring across companies. It is one of the fastest-growing areas and often requires a robust hiring process to get the best candidates in the team. We got in touch with Sreekanth Menon, VP, Data Science at Genpact to understand their data science hiring process. Sreekanth shares that not only their hiring mechanism looks for the best available candidate but also retains them by providing opportunities to improve their skills over time. Data Science Skills That Genpact Looks For Sreekanth shared that with years of experience in leveraging data analytics for solving challenges across industries and functions, Genpact has identified the most vital and useful attributes that a potential data scientist needs to have. He shared some of them as below: Attitude: One of the essential attributes that Genpact looks for in a potential data scientist is their attitude towards learning. They believe that being a good data scientist requires an inherent level of flexibility, curiosity, adaptability and an internal urge to explore and share knowledge.  Analytical Mindset and Critical Thinking: While knowledge of AI\/ML algorithms is one aspect of skills that data scientists should have, Sreekanth shares that to share real-world business challenges, they also need to apply this knowledge objectively to arrive at the right solutions.Proficiency in Mathematics and Statistics: Most AI\/ML algorithms have been developed from the base of statistical and mathematical analytical tools and techniques. Having a strong grasp of the underlying concepts of these tools and techniques is essential for any data scientist.Competency in Coding: Proficiency in computer programming languages such as R, Python and others is a must. Understanding of Data Architectures: The use of AI\/ML techniques in business challenges involves managing and working with large volumes of data. Understanding how data architectures work is critical to having a good grasp of the challenges involved in deploying AI\/ML solutions and is one of the most important skills that Genpact looks for in potential candidates.Storytelling and Communication Skills: Genpact values candidates who can not only visualise the data but also tell an insightful story about what the visualisation means with respect to an organisation’s goals and outcomes. Business Acumen: Genpact looks for candidates with a certain level of business acumen along with technical skills. An individual must be inquisitive towards learning domain-specific dynamics so that they can apply the AI\/ML techniques within the right business context. Hiring Process At Genpact Talking about their hiring methodology, Sreekanth shares that their hiring process focuses on identifying if the potential data science candidate fits the bill in terms of all the factors as mentioned above. Further, he shared that Genpact uses a five-step process to effectively hire the best data science candidates, which are as follows: Step 1: Testing – The candidates are required to undergo a series of cognitive, IQ and coding tests that are designed to evaluate their competency levels in mathematics, statistics and coding ability.Step 2: Business Understanding – Candidates are required to solve critical business problems and thoughtfully analyse information in a live case study that is designed to test the business acumen of the candidate.Step 3: – Critical Thinking – This step involves a whiteboarding exercise with experts from the Genpact data science where candidates are assessed on their critical thinking and analytical mindset. Step 4: Video Interviews – Genpact values domain knowledge along with technological expertise in candidates and identifies these attributes in candidates through video interviews with the subject matter experts. This step tests the candidate’s storytelling, communication skills and technical proficiency in understanding data architectures.Step 5: Face-To-Face Interviews – Genpact conducts face-to-face interviews of candidates with business leaders to understand if the candidate’s attitude is a good fit for their culture. “With an 8000+ strong analytics team, Genpact has identified several best practices when it comes to hiring data scientists and data engineers. Genpact tends to look for candidates that can apply advanced algorithms in solving critical business problems. Candidates with both technical skills and domain-expertise have a higher likelihood of being able to identify opportunities where AI\/ML solutions can be applied to solve real-time business challenges,” said Sreekanth. He further added that additionally, any candidate with previous experience of working in AI\/ML projects at scale is always considered to be a significant hire. Some of the ways that Genpact hires data scientists are through referrals and recruitment drives, internal job movement, university campus hiring, internship programs, international conferences and hackathons. Retaining Data Scientists As mentioned earlier, retaining data scientists is an essential aspect for Genpact. Sreekanth shares that one of the best ways is to continually upskill them in line with the ongoing trends in the field. “Retaining them at low attrition rates requires that they are motivated to take up challenging tasks and continuously upskill themselves to suit the requirement of the assigned project,” he said. He further added that they encourage their data scientists to become experts in the state-of-the-art algorithms, architectures and technologies by offering them self-paced tutorials through an in-house upskilling framework — Genome. Genome provides development opportunities for data scientists in four different dimensions — industries, service lines, digital and professional skills, adding up to more than 300 specific skills. In addition, Genpact believes in fostering a culture of research and development and encourages data scientists to make contributions to the global learning community through technical paper publications, patent publications and participation in AI\/ML conferences. “The data science field is supply-constrained – the demand for data scientists far outstrips the current number of potential candidates. Thus, companies will need to attract the right kind of candidates by being highly proactive. With the recent COVID-19 global crisis forcing companies to take stock of their situations; many organisations realise the significance of skilled data scientists in enabling faster data-driven decision-making. It seems likely that the demand for talented data scientists will only increase in the near future as businesses across the globe look to deploy advanced analytics, AI and ML solutions across their value streams,” said Sreekanth on a concluding note.","excerpt":"With a surge in AI and ML applications to improve business process efficiencies, there has been an increase in data science hiring across companies. It is one of the fastest-growing areas and often requires a robust hiring process to get the best candidates in the team. We got in touch with Sreekanth Menon, VP, Data […]","categories":["AI Hirings"],"tags":["Data Science Hiring","data science hiring india","data science hiring process","Data Scientist","Genpact","Hiring","Python for Data Science"],"author_name":"Srishti Deoras","publish_date":"2020-11-22T13:00:00","publication_year":"2020","word_count":1015,"keywords":["data science","Go","Genpact","AI","ML","Hiring","Git","Data Scientist","RAG","Data Science Hiring","Python","GAN","analytics","Python for Data Science","data science hiring india","data science hiring process","R"],"extracted_tech_keywords":["AI","ML","data science","analytics","RAG","Python","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/what-is-the-hiring-process-for-data-scientists-at-genpact\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10161145,"title":"Dear L&amp;T, This is a Recipe for Attrition","content":"In June last year, Larsen & Toubro (L&T) made headlines for grappling with an acute manpower shortage across its businesses. Chairman SN Subrahmanyan, popularly known as SNS, said that the company needed around 45,000 engineers and techies. An attrition rate of 10% was said to be a contributing factor. L&T has made headlines once again, and this time, SNS broke the internet with his viral video. In it, he was seen asking employees to work 90 hours a week, including Sundays—a move that could only compound the company’s attrition and staff shortage issues. During an employee interaction, Subrahmanyan said he would be happier if he could make them work on Sundays as well. “What do you do sitting at home? How long can you keep looking at your wife? Come on, get to the office and start working,” he added. Drawing comparisons with China’s intensive work culture, he said, “If you want to be on top of the world, you have to work 90 hours a week.” Facing Brickbats Netizens have reacted sharply to his extreme work expectations, erupting in a flurry of memes, jokes and posts online. A Reddit user commented, “So unfortunate, we have such business leaders! I think we must call them “leaders in baby diapers” :) I had a few close friends who worked at L&T Madras. About 10 years ago. Going by what they said about the work culture, I felt it was like an adults’ kindergarten.” Another added: “L&T came to my college for placements, offering a CTC of 6 LPA, and they expect us to work 90 hours a week for that? This really highlights the sad state of labour laws in India and the mindset of some Indian chairmen and CEOs. It’s honestly ridiculous.” A former L&T employee, Karthik Madhavapeddi, deputy editor at IndiaSpend, had this to say: “I just saw a news report in which L&T chief SN Subrahmanyan (SNS to employees) is quoted saying he wants employees to work 90 hours a week.” He added that having worked at L&T Construction from 2010 to 2013, “I can say this reflects the typical mindset of someone with a background in construction. On-site, we had 6.5-day workweeks, with hours stretching from 8:30 am to 8:30 pm Monday through Saturday, and up to 1 pm on Sundays. The only exceptions were projects where the client’s operations didn’t permit such extended hours.” Further, he said that labourers and workmen were compensated with overtime pay for the extra three hours. Employees, however, were not. “When I attended an internal interview for the management trainee programme at the corporate office in Mumbai, the interviewers didn’t seem to grasp why such long hours were necessary on-site,” Madhavapeddi added. He said that SNS may have brought the same “construction culture” into the corporate. Narayan Murthy in the Mix Last year, Infosys co-founder Narayan Murthy kicked up a storm with his “70-hour a week” remark. Commenting on development and nation-building, Murthy said, “India’s work productivity is one of the lowest in the world… my request is that our youngsters must say, ‘This is my country. I’d like to work 70 hours a week’.” Bollywood actress Deepika Padukone also took to social media, connecting SNS’s remarks to mental health. In a LinkedIn post, Sanjay Sehgal, chairman & CEO at MSys Technologies, explained that Indian workers, on average, worked significantly longer hours than their global counterparts. According to the International Labour Organisation, the average Indian worker, aged around 15, clocks in 47.7 hours per week. This is higher than countries like the US (36.4), the UK (35.9), Germany (34.4), and even Asian countries like China (46.1), Singapore (42.6), and Japan (36.6). He further claimed that gig industry workers, such as those working for UrbanCompany, Swiggy, Zomato, Ola, and Uber, put in 11-12 hours a day, often totalling over 70 hours a week. This includes labourers, electricians, and plumbers, who spend long hours but often lack growth opportunities or fair pay. Young workers, aged 16 to 25, are increasingly involved in gig work like driving taxis, delivering food, or renting bikes, which provide limited benefits or career progression. Additionally, employees in IT and corporate sectors often face expectations of being available round-the-clock for calls and emails to ease collaboration with global teams. “But now, despite knowing the effects of long working hours, pushing employees to work for 70 hours a week sounds unjust and brutal,” said Sehgal. L&T, however, rushed in to defend its chairman. A company spokesperson said, “Nation-building lies at the heart of our mission. For over eight decades, we have been shaping India’s infrastructure, industries, and technological capabilities. We believe this is India’s decade – a period calling for collective commitment and effort to drive growth and realise our shared vision of becoming a developed nation.”","excerpt":"L&T chairman SN Subrahmanyan faces backlash for suggesting employees should work 90 hours a week, including Sundays.","categories":["AI Features","IT Services"],"tags":["Indian IT"],"author_name":"Shalini Mondal","publish_date":"2025-01-10T18:24:35","publication_year":"2025","word_count":797,"keywords":["Go","API","ELT","AWS","AI","RAG","Ray","Aim","GAN","Indian IT","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","AWS","R","Go","API","ELT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/dear-lt-this-is-a-recipe-for-attrition\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10002693,"title":"How To Tackle A Machine Learning Project As A Beginner","content":"The proliferation of artificial intelligence (AI) has made mastery over machine learning (ML) an imperative for anyone in the field of software development. While theoretical knowledge will help, it is not possible to master ML without a strong foundation in practical experience. There are numerous digital ML courses that enable learners to dabble in various projects and many other freely available resources online for the same. With hands-on experience in these projects, aspiring developers, data scientists and other professionals will not only understand how to apply their knowledge to solve real-world problems, but by undertaking different kinds of projects, they will also continue honing their skills, understand their strengths and weaknesses better, and also add valuable experience to their overall portfolio, making it easier for them to land a job. However, it can be tough to know where to begin. In the following paragraphs, we have tried to provide some guidance to newcomers on how they can apply ML to real-world problems through projects: Get Familiar With Common ML Applications This could be termed step 0 in an ML project lifecycle. Before you get started, invest time in developing a better understanding of ML. There are three basic types – supervised learning, unsupervised learning, and reinforcement learning. Learn what the applications of each of these could be, and once that is done, you will have a better idea of how to apply ML to your problem. ALSO READ: PyTorch VS Caffe2 – Which One Should You Use For Your Next Project? Select A Project As mentioned in the introductory paragraphs, there are many ML projects found online that use publicly available real-world datasets. Understand whether or not it covers core aspects of ML. Also, check if it aims to solve a pressing problem and provide real value to stakeholders – this is relevant, especially if you are doing projects with an eye on landing jobs. While choosing a project from an industry, you are familiar with will help you play to your strengths, selecting one that you may not be well informed about will give you the chance to explore an interesting topic. Understand The Problem Once you have selected the project, proceed to define the problem you are trying to solve, and its end objectives. Although this may seem like an easy step, it will give you a window into the criticality of the problem you are trying to solve. During this step, ensure that the objective of the project is measurable. It will also allow you to revisit your choice of the project early on if the results are not desirable. Outline Limitations – If Any As an extension to the previous point, this is another step that evaluates the choice of your project and allows you to revisit it if the results are not desirable. The limitations you should consider before proceeding are the following: Resources (lack of time)Infrastructure (lack of computing power)Data (Unstructured and uninterpretable) ALSO READ: Why You Should Shift Your ML Project To Cloud Cleaning The Data If the above points check out, and you have finalised a project to work on, the next step would be to clean the data. If you have curated and collected them from numerous sources, merge them into a single table. After this, wrangle the data, and then conduct exploratory data analysis (EDA). Selecting, Training & Evaluating Your Model Once you have cleaned your dataset, you can begin by training your model based on the algorithms. There are several tasks involved in this step, the first of which is selecting the model, which varies based on the problem chosen. A different modeling approach could be warrant based on whether it is a regression problem or a  classification one. Following this, you need to train your model and then evaluate it using testing data based on the success metrics decided by you. In a previous article, we have illustrated these steps with an example. You can read it here. Outlook While there are fun applications of machine learning as well, choosing problems of real import will have a higher chance of getting noticed on your professional portfolio. It might make it a tough playing field, especially if you are a newcomer, but the learning curve with such a project will be great.","excerpt":"The proliferation of artificial intelligence (AI) has made mastery over machine learning (ML) an imperative for anyone in the field of software development. While theoretical knowledge will help, it is not possible to master ML without a strong foundation in practical experience.  There are numerous digital ML courses that enable learners to dabble in various […]","categories":["AI Features"],"tags":["datascience","Machine Learning","machine learning projects"],"author_name":"Anu Thomas","publish_date":"2020-07-18T16:00:00","publication_year":"2020","word_count":713,"keywords":["Go","artificial intelligence","machine learning","AI","PyTorch","datascience","ML","R","Machine Learning","Git","RAG","Aim","machine learning projects"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","PyTorch","RAG","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-to-tackle-a-machine-learning-project-as-a-beginner\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10102377,"title":"Dynatrace &amp; Kyndryl Form a Partnership to Amplify Business Potential","content":"Kyndryl and Dynatrace announced a global alliance to provide joint offerings for insights and informed business decisions. The collaboration offers unified observability, application modernisation, and enhanced AIOps capabilities for customers. The alliance expands on their 2022 relationship, aiming to provide more capabilities in unified observability, application modernisation, cloud migration, and IT service and operations automation. The Kyndryl and Dynatrace partnership reaps manifold benefits for both companies and their customers. For Kyndryl, this alliance enables the offering of an expanded array of services through collaboration with Dynatrace, granting access to the latter’s observability platform to enhance their own service spectrum. This integration also opens avenues for Kyndryl to increase revenue by marketing and selling Dynatrace’s products to its customer base. “Leveraging Dynatrace’s OneAgent and Kyndryl’s expertise has helped Hapag-Lloyd access more system process insights,” said Michele Madaus, Director IT Supporting Platforms at Hapag-Lloyd. Dynatrace, on the other hand, gains access to Kyndryl’s specialised expertise in IT infrastructure and cloud computing, expanding its customer reach and revenue streams through Kyndryl’s established sales channels. Customers, as a result of this partnership, experience the advantage of a consolidated solution for their IT infrastructure needs and access to cutting-edge observability technology via Dynatrace’s platform. “Kyndryl is an ideal partner to bring the Dynatrace observability and security platform to more customers,” said Michael Allen, Vice President of Global Partners at Dynatrace. Specific instances include banks optimising cloud applications, retailers enhancing edge computing, and healthcare providers securing patient data while leveraging monitoring solutions for potential incidents, underscoring the practical application and benefits of this collaboration. This alliance not only offers immediate advantages but also holds the promise of further innovative solutions and services in the evolving partnership between Kyndryl and Dynatrace. “Kyndryl’s alliance with Dynatrace provides more opportunities for increased application observability and actionable business insights,” said Nicolas Sekkaki, Kyndryl Applications, Data, and AI Global Practice Leader.","excerpt":"Kyndryl and Dynatrace collaborate to mutually expand their services. Customers benefit from comprehensive IT solutions and advanced technology, leading to improved performance and security.","categories":["AI News"],"tags":["Kyndryl","partnership"],"author_name":"K L Krithika","publish_date":"2023-11-02T14:43:13","publication_year":"2023","word_count":310,"keywords":["partnership","Kyndryl","programming_languages:R","AI","cloud computing","RAG","automation","Ray","Aim","edge computing","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","cloud computing","edge computing","R","automation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/dynatrace-kyndryl-form-a-partnership-to-amplify-business-potential\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10052831,"title":"LatentView All Set To Become First Pure Play Analytics Company In Indian IPO","content":"Data analytics company LatentView Analytics’ IPO will open on November 9 and close on November 12. The size of the IPO is valued at ₹600 crores, of which ₹474 crores is a fresh issue and ₹126 crores is up for sale. Per equity, the share band has been set between ₹190-₹197. Founded in 2006, LatentView will become the first pure-play analytics company to be listed in the Indian IPO. The company decided to be listed as it believes that the future of data analytics is offering huge growth opportunities, and they don’t want to miss out on any. CEO Rajan Sethuraman stated, “The total capital we are raising, we believe, is fairly small in comparison to the market potential. The industry reports and the prospectus is listing out that the space is expected to grow at 18-20 per cent CAGR. Compared to that, the capital raise that we were doing is very small. It is almost insignificant in the larger market size.” Like any other company, LatentView Analytics did see some hiccups as the pandemic hit, but it soon became an accelerator for growth. CFO Rajan Venkatesan said, “As greater digital acceleration pushed a high portion of our client business online, this worked very well for us. Also, a lot of our clients actively and passively got comfortable with a lot of the complex business problem solving historically that was done from on-site location. We were able to achieve a significant business shift to our off-shore centres of excellence in Chennai and Bangalore. Thereby achieving a significant margin expansion in the FY 2020-21.” Over 80 per cent of the company’s clients are from the US, and it plans to increase growth in Europe. The company plans inorganic growth and has started evaluating young start-ups for mergers and acquisitions.","excerpt":"At the announcement of their IPO opening, Rajan Venkatesan, CFO, Latent View Analytics, stated that the demand for going digital increased a lot during COVID19 that compelled growth","categories":["AI News"],"tags":["Data Analytics","IPO","latentview analytics"],"author_name":"Meeta Ramnani","publish_date":"2021-11-03T15:23:17","publication_year":"2021","word_count":298,"keywords":["Go","API","programming_languages:R","AI","IPO","latentview analytics","programming_languages:Go","Git","GAN","analytics","Data Analytics","R"],"extracted_tech_keywords":["AI","analytics","R","Go","Git","API","GAN","IPO","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/latentview-all-set-to-become-first-pure-play-analytics-company-in-indian-ipo\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10096991,"title":"Flyfish Wants to be a Premium Indian ‘Consultative Sales AI’ Offering","content":"Fractal, a global provider of AI and advanced analytics solutions, has introduced Flyfish, a 360-degree generative AI platform for digital sales. It aims to enhance revenue growth and foster long-lasting customer relationships. The vision behind Flyfish is to revolutionise the sales experience by making technology more human-like. It offers consultative experiences through intuitive sales advisors, enabling personalised, data-driven shopping interactions with consumers. Moreover, the platform aims to create a new market category called consultative sales AI. According to the team, the platform has shown promising results in pilot programs, with increases in lead generation, conversion rates, and a significant reduction in customer acquisition costs. “In the close-door pilots that we have built and launched for some of our customers, there is a significant jump in lead generation. There was a 35% jump in in one of the cases and the conversion rate went up by 12% in another one,” Shridhar Marri, CEO and co-founder of Flyfish, said during a recent interview with AIM. He also predicted that the product would help enterprises bring customer acquisition costs down by 40% over three years. How it Works Flyfish has a preprocessor that analyses user questions or context, determines the appropriate flow, and retrieves relevant products and recommendations. This information is sent to a large language model (LLM) to generate a humanised response, which is then post-processed within Flyfish. The platform also includes a channel connector that allows integration with various digital channels like web, messaging platforms, and social media. If needed, a human can be brought into the conversation, and their responses are delivered through the same channels. Custom flows can be created within Flyfish for specific enterprises, such as incorporating sponsored or complementary product options. Flyfish can also be used for marketing, where it can provide a point-of-sale or advisory experience for users clicking on ads. The platform is agnostic to the choice of language models, allowing users to select their preferred LM, such as GPT, Google’s models, or others. Flyfish integrates with messaging APIs to connect with different platforms like WhatsApp, Telegram, and more. The platform seamlessly integrates into existing sales and marketing tech stacks, encompassing e-commerce platforms, product catalogues, pricing, availability, and more. It operates on a cloud-agnostic architecture, utilising various cloud services such as Azure, ensuring flexibility and scalability. Addresses Specific Consumer Needs Flyfish addresses the diverse needs and contexts of individual consumers, streamlining their research process and significantly reducing the time required to make informed purchasing decisions. By leveraging generative AI, Flyfish delivers a range of benefits, including the ability to provide tailored product recommendations and suggestions that consumers may not have discovered otherwise. “I think we’re going to change the way consumers interact with the brands. Every consumer, if you look at them, have different needs and contexts.” He explained, when making a significant purchase like a car, it is common for consumers to go through extensive research that can last anywhere from six to eight months. This research period applies to various types of purchases, with varying time frames depending on the specific item. Regardless of the product, consumers invest significant time in researching their options, often spanning across days, weeks, or even months. Marri expects Flyfish to reduce this time and result in an enjoyable experience. A Secure Enterprise-Friendly Product with Ambitions Fractal’s backing as a prominent AI company and its strong relationships with Fortune 500 enterprises position Flyfish for success. The Flyfish team consists of approximately 30 professionals with expertise in AI engineering, product management, experience design, marketing, and research and development. They also have researchers and developers in AI and ML and deep learning. Flyfish has great ambitions and wants to carve out a niche for itself. “We are trying to create a cutting-edge global product out of India,” Marri said. While there are other platforms like Sendbird, which provide an end-to-end automated chatbot experience, Flyfish is looking to differentiate itself from the competition. Marri also believes that other competitors might take a while to reach the stage they currently at. Given the significant progress in innovation, Flyfish is ahead of the curve. “At this point of time, there is no direct competition but there are some people trying to use chatbots and bring in a little bit of conventional conversational commerce into the play. But we’re trying to defy the conventional chatbot experience and bring in a consultative experience, he said.” Marri revealed that there are several ongoing conversations with the Fortune 500 companies within Fractal’s ecosystem. “One of the biggest advantages that we have is to use Fractal’s access to large enterprises around the world and then offer Flyfish as a product,” he said. Enterprises have been wary of their data, however, Flyfish prioritises the security of customer data by ensuring that they do not store any of it. Instead, the data flows directly into the enterprise’s ecosystem, including their tech stack, cloud services, CRM and ERP systems, and e-commerce platforms. Moreover, anything that passes through the platform is encrypted for additional measures to safeguard the data. The Existing AI in Sales Market While, Fractal claims that Flyfish is the world’s first 360-degree generative AI platform for digital sales. There are other firms that offer several kinds of assistance in sales and marketing. Several AI tools are available for sales organisations of different scale. Some notable examples include, Conversica, which provides an automated AI sales assistant that engages leads in conversation and qualifies them before they talk to a sales representative. Exceed.ai, which uses AI to engage in human-like conversations with sales leads via email and chat. Crayon, an AI-powered competitive intelligence tool that tracks competitors’ activities online and automatically generates sales battle cards; and Salesforce, a prominent player in the AI for sales space, with its AI named ‘Einstein’. Einstein prioritises leads, evaluates deal likelihood, and allows developers to incorporate AI into Salesforce apps. Microsoft’s Viva Sales offers features such as personalised customer email drafting, customer insights, and generating recommendations and reminders. The integration of AI into existing platforms by established companies demonstrates the growing trend of incorporating AI into sales processes. According to McKinsey analysts in Harvard Business Review, the capability of AI to enhance and boost sales performance can potentially generate a value ranging from $1.4 to $2.6 trillion in marketing and sales.","excerpt":"Fractal’s backing as a prominent AI company and its strong relationships with Fortune 500 enterprises position Flyfish for success","categories":["IT Services"],"tags":["AI integration","Competition","Data Security","Fractal","integration","language model","Language Models","Marketing","Revenue Growth","Salesforce"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2023-07-17T10:00:00","publication_year":"2023","word_count":1046,"keywords":["integration","chatbots","Fractal","Ray","deep learning","Salesforce","AI integration","RAG","analytics","Competition","AI","ML","generative AI","Language Models","Revenue Growth","Data Security","language model","Aim","Azure","Marketing"],"extracted_tech_keywords":["AI","ML","deep learning","analytics","generative AI","Aim","Ray","RAG","chatbots","Azure"],"url":"https:\/\/analyticsindiamag.com\/it-services\/flyfish-wants-to-be-a-premium-indian-consultative-sales-ai-offering\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":16105,"title":"Infosys Finacle integrates with Active AI platform to boost conversational services in banks","content":"In a recent development around conversational AI in banking and financial services, Infosys Finacle, a part of EdgeVerve Systems collaborated with Singapore based Active.AI that offers enterprise AI platform to banks globally. The integration of Active AI platform with Finacle will help financial institutions to offer conversational solutions via chat and voice interfaces across their digital channels. This step can boost the customer experience and automate the processes for banks. One of the fastest growing fintech companies across the globe, Avtice.AI was recently recognized in the recently concluded chapter of Finacle FinTech Connect, that identifies promising fintech globally. “Elevating customer experience and reducing servicing costs are two of the strategic goals banks across the world are pursuing. The advancement in artificial intelligence technologies offer a significant opportunity to achieve these twin goals”, said Sanat Rao, Chief Business Officer, Infosys Finacle. “Active.AI conversational banking offering focused on financial services will complement our suite of digital banking offerings and will help our clients make banking simpler for their customers,” he added. Some of the benefits that this association would bring are: The Active.AI platform integrated with Finacle would enable end customers an interaction with software bots in natural language through voice and chat interface. It can prove beneficial around account information, pay bills, make transfers and access several other banking services. The joint solution would mean better use of artificial intelligence to understand customer interactions across various channels. It would also mean personalized conversations that can make tailored offerings to customers. Available in multiple languages, the solution would be offered through Finacle digital channels solutions such as Finacle Online Banking and Finacle Mobile Banking. “We are building the next generation of secure and scalable technology for financial services with an AI first approach. Our enterprise AI platform will enable banks to be ready for the conversational era”, said Parikshit Paspulati, Chief Technology Officer, Active.AI. He added “Partnering with Finacle helps us greatly increase delivery velocity for our mutual clients. Axis Bank is one of the first banks, where we have integrated our AI platform with Finacle ecosystem.”","excerpt":"In a recent development around conversational AI in banking and financial services, Infosys Finacle, a part of EdgeVerve Systems collaborated with Singapore based Active.AI that offers enterprise AI platform to banks globally. The integration of Active AI platform with Finacle will help financial institutions to offer conversational solutions via chat and voice interfaces across their […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-07-06T04:19:36","publication_year":"2017","word_count":345,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","Scala","Git","programming_languages:Scala","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Scala","Git","programming_languages:R","programming_languages:Scala","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/infosys-finacle-integrates-active-ai-platform-boost-conversational-services-banks\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":56454,"title":"Can AI Create Video Games","content":"From DeepMind’s AlphaStar (beating 99.8% of the Starcraft players) and AlphaGo to OpenAI’s Dota 2 bot, AI is revolutionising the gaming industry. Each time it beats a human at a video game, it is believed to be getting closer to a level where it can make decisions on its own. And, that’s the reason AI is being put through hours of training so that it can make decisions on its own when it enters the real-world. Of course, playing a game and learning to counter the real-world problem eventually is something AI has always been trained for. AI has also been playing a massive role in creating video games and making it more tailored to players’ preferences. Matthew Guzdial from the University of Alberta and his team have been working towards leveraging AI’s power to help video gamers create the exact game that they want to play. Creating A Game Using AI The approach researchers use is through machine learning, where a system learns to create approximate representations of the games, and recombine the knowledge from these representations in order to develop new games via conceptual expansion. This approach helps in demonstrating the ability of the system to recreate the games. The team of researchers first fed the machine with the data in the form of videos. This video contained hours of gameplay from humans playing the first levels of games like Super Mario Bros, Kirby’s Adventure, and Mega Man. After hours of ‘watching’ these video game footage, the AI was able to probabilistically map the relationship between the objects and how that change to generate ‘game graphs’. When the AI is watching the game, it is also guessing the rules of the game. The AI is made to watch the gameplay, and here, it is validating the rules of the game that it has guessed. The ‘game graphs’ are a result of putting together these two sets of data acquired by watching and rewatching the gameplay. These game graphs mainly contain the details about the game, and the system makes use of the information present in the game graphs and starts to design, combine and reproduce. To start the testing, the AI will first be trained on the Mega Man and then will be asked to complete the game based on its knowledge and approximations from the game graphs. But, the AI failed to completely recreate the game leaving out some aspects of the games. For example, it was unable to deduce the mechanics of the magnetic beam from the game. But, Super Mario Bros., was simpler for it to understand because it didn’t contain power-ups that had a significant effect on the character. To create a new game, for example, the researchers combined the platforming styles of Mega Man and Super Mario Bros. Through many repetitions for every level and every game rule, a completely new AI-generated game can be created. Replace Game Developers? A common thought arises when anyone reads about these kinds of technology, will it replace human jobs? The answer is no. The researchers believe, instead of replacing, AI will be helping in easing the burden on game developers. These tools do not contain coding, so it gives more accessibility to the game creators by removing the hassle of dealing with codes. Future of the Technology Next, the researchers have planned to make the AI system to predict and design the whole game with just two frames and user-defined data. The AI will make predictions, and post the creator gives feedback; the AI then makes the adjustments. Going through this process over and over again will ultimately result in something entirely new. The AI system will be able to automatically produce in-game visuals, sound and also the story in the future. Outlook The video gaming industry requires long hours of hard work and expertise in terms of designing and coding skills and data. The coding skill especially presents a problem to non-coding personnel for creating games. This automating of game design will democratise the industry and allow for applications in education, scientific and entertainment.","excerpt":"From DeepMind’s AlphaStar (beating 99.8% of the Starcraft players) and AlphaGo to OpenAI’s Dota 2 bot, AI is revolutionising the gaming industry. Each time it beats a human at a video game, it is believed to be getting closer to a level where it can make decisions on its own. And, that’s the reason AI […]","categories":["AI Features"],"tags":[],"author_name":"Sameer Balaganur","publish_date":"2020-02-12T10:00:00","publication_year":"2020","word_count":679,"keywords":["Go","machine learning","OpenAI","AI","programming_languages:R","programming_languages:Go","RAG","BERT","llm_models:BERT","R"],"extracted_tech_keywords":["AI","machine learning","OpenAI","RAG","R","Go","BERT","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/can-ai-create-video-games\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042266,"title":"World’s First AI Analog Chip","content":"Austin-based Mythic has launched the Mythic Analog Matrix Processor (Mythic AMP) — a single-chip analog computation device. The M1076 AMP uses Mythic Analog Compute Engine (ACE) to deliver the compute resources of a GPU at up to a tenth of the power consumption. With a 3-watt power draw, the M1076 can perform up to 25 trillion operations per second (TOPS). The new lineup includes a single chip, a PCIe M2 card for low-footprint applications, and a PCIe card with up to 16 chips. Now, edge devices can execute complex AI applications at greater resolutions and frame rates, resulting in superior inference results. Computation happens at the same place where data is stored. Last month, Mythic raised $70 million in Series C funding, with Blackrock as the lead investor and co-led by Hewlett Packard Enterprise (HPE). Why Mythic AMP? In traditional computers, data is transferred from the DRAM memory to the CPU at regular intervals. Memory holds programs and data. The processor and memory in computers are separate, and data moves between the two. Over the years, processor speeds have seen a drastic uptick. Meanwhile, memory advancements have been mainly focused on density — the ability to store more data in less space – rather than transfer rates, leading to latency. Simply put, processors, no matter how fast they are, have to sit idle while fetching data from memory and are dependent on the rate of transfer — Von Neumann limitation. Thereby, merging compute and memory in a single device, analog AI eliminates the von Neumann bottleneck, resulting in dramatic performance gains. In addition, tasks can be completed in a fraction of the time and with a lot less energy because there is no data transit. Each Mythic ACE is accompanied by a digital subsystem including a 32-bit RISC-V nano processor, 64KB of SRAM, SIMD vector engine and a high-throughput network-on-chip (NoC) router. The Analog Matrix Processor is capable of delivering power-efficient AI inference at up to 25 TOPS. “Edge devices can now deploy powerful AI models without the challenges of high power consumption, thermal management, and form-factor constraints,” as per the company. AI on edge Mythic’s primary focus is on edge AI deployments. The company also provides server-class compute in data centres. Edge AI can be used by businesses to deploy ML models that operate locally on edge devices. However, edge AI faces some challenges: Low Power: The device’s power and related heat grow as more functions and capabilities are added. Sometimes they are powered with Power-over-Ethernet (PoE) having limited power budgets. Devices need to exhibit strong performance even at 0.5 or 2W. Power should be as near to zero when not in use, and switching between these different modes should be rapid and simple. Small Size: AI algorithms running at the data source have minimum latency issues and no loss of accuracy due to video compression; hence, there are no requirements for large PCIe cards, big heatsinks, or fans. The entire system needs to fit on a 22mm x 30mm M.2 A+E card for others. Even with larger PCIe cards, the size of the accelerator and cooling solution determines how much AI can be crammed in. Cost-effectiveness: The capability to deliver high-power computing at an affordable and effective price gives customers much freedom to scale up to customer demand. To date, the company has raised $165.2 million for easy and cost-effective deployment of powerful AI for the smart home, smart city, AR\/VR, drone, video surveillance, and even manufacturing.","excerpt":"Austin-based Mythic has launched the Mythic Analog Matrix Processor (Mythic AMP) — a single-chip analog computation device. The M1076 AMP uses Mythic Analog Compute Engine (ACE) to deliver the compute resources of a GPU at up to a tenth of the power consumption.  With a 3-watt power draw, the M1076 can perform up to 25 […]","categories":["IT Services"],"tags":["AI Applications","edge AI","edge computing"],"author_name":"kumar Gandharv","publish_date":"2021-06-24T11:00:00","publication_year":"2021","word_count":579,"keywords":["Go","API","funding","programming_languages:R","AI","ML","AI Applications","Git","programming_languages:Go","edge AI","edge computing","R"],"extracted_tech_keywords":["AI","ML","edge AI","R","Go","Git","API","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/worlds-first-ai-analog-chip\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10168668,"title":"Indian IT Services Firms have Greater Insulation from H-1B Visa Disruption","content":"While concerns about H-1B visa scrutiny and enforcement continue to make waves, their actual impact on India’s IT services sector appears increasingly muted. This can be attributed to years of proactive workforce localisation and wage realignment. In their recent Q4 earnings calls, Accenture, TCS, Wipro, Infosys and HCL Tech, none of them highlighted any challenges pertaining to H-1B visa rejection or acceptance rates. However, the companies did mention that Trump’s tariffs created uncertain business situations. During Trump’s first term, the administration pursued a restrictive stance on legal immigration, particularly targeting work visas. Through Executive Order 13788, titled ‘Buy American and Hire American’, the administration significantly tightened the eligibility criteria for H-1B and L-1 visas. As a result, denial rates for these visa categories rose sharply, disproportionately affecting IT services companies, which were heavily reliant on foreign skilled workers. H1B Visa Dependency Declining for Indian IT? According to a report by JM Financial, in FY2017, approximately two-thirds of US employees at major Indian IT services firms were on H-1B or L-1 visas. However, over the past decade, this dependency has significantly declined, as companies proactively localised their talent base in response to tightening visa norms. This strategic shift is evident in the 50% to 80% drop in H-1B visa approvals for major players like Infosys, TCS, and Wipro between FY2015 and FY2024, according to the report. This evolving workforce composition now offers Indian IT services firms greater insulation from future disruptions in the H-1B visa programme. Even if the Trump administration were to return with a renewed focus on restricting work visas, the reduced reliance on H-1B holders could soften the impact for these companies. Rajesh Ranjan, managing partner at Everest Group, believes that Indian IT will remain unaffected as they have already strengthened their presence in the US during the previous Trump administration in response to similar issues earlier. Meanwhile, a report by NFAP revealed that Cognizant had the second most H-1B petitions in FY 2024 approved for initial employment (2,837), followed by Infosys (2,504), TCS (1,452), IBM (1,348), Microsoft (1,264), HCL America (1,248), Google (1,058), Capgemini (1,041) and Meta Platforms (920) in FY 2024. Why Indian IT Appears Muted? The reason is that Indian IT companies have significantly increased their focus on local hiring, particularly in the Americas. In a LinkedIn post, Prince Chaturvedi, regional business head at IDFC FIRST Bank, mentioned that Infosys, for example, has expanded its workforce in the region from 19,720 employees in 2014 to 36,118. Similarly, Cognizant has grown its US workforce from 40,500 in 2014 to 65,000 by 2023. Alongside this hiring surge, these firms are also making substantial investments in talent development. “Increased local hiring and higher offshoring since Trump’s first term has reduced IT companies’ dependence on visa-based talent, thereby safeguarding their interest against immigration related policy uncertainties,” Abhishek Kumar, equity research analyst, JM Financial Ltd, told AIM. Collectively, as per reports, the Indian technology companies have contributed over $1.1 billion and developed partnerships with nearly 180 universities, colleges, community colleges, and others to strengthen and diversify the STEM pipeline in the US. Sid Tipnis, leader, technology, media and telecommunications (TMT) consulting at Deloitte, told AIM that “IT\/ITeS organisations have relied in the past on temporary\/permanent immigration arrangements to enable workforce to be onshore and maintain high touch with clients focused across the lifecycle of sales, delivery and customer success.” He further added, “These organisations are increasingly augmenting talent onshore through local sovereign engagements. Furthermore, as clients welcome virtual and collaborative ways of working on the delivery lifecycle, we see a limited impact of potential immigration restrictions.” Link Between H-1B Workers and AI Innovation in the US As per the report, since 2000, foreign-born workers have accounted for 55% of job growth in AI-related fields, highlighting their essential role in advancing the sector. The shortage of qualified AI professionals in the US has created a significant talent gap, and H-1B visa holders are critical in filling it. Their presence contributes to both short-term productivity and long-term innovation. Notably, H-1B visa applications are strongly linked to higher rates of patent filings and citations, suggesting these workers play a direct role in technological advancements. Furthermore, the absence of H-1B workers could lead to a decline in successful AI startups, as these visa holders often contribute to new business creation and are more likely to secure venture capital funding. Speaking on the same topic on X, Elon Musk said, “The reason I’m in America, along with so many critical people who built SpaceX, Tesla and hundreds of other companies that made America strong, is because of H1B.” What About Wage Obligations for H-1B Workers? Kumar stated in his report that one area that could still be affected is wage obligations for H-1B workers. In Trump’s previous term, the Department of Labor (DOL) attempted to raise wage levels for foreign workers through an Interim Final Rule (IFR) in late 2020 and early 2021, though this was later struck down in court. Today, wage data for select software roles suggests that H-1B employees are paid approximately 25% above prevailing wage levels, casting doubt on the argument that they undercut American workers. Interestingly, he further stated that not all potential Trump policies would be unfavourable for the tech sector. His proposed expansionary fiscal policies include reducing the corporate tax rate from 21% to 15% for domestic production. This could stimulate demand in IT services by easing client budget constraints, particularly in sectors with heavy IT spending. Meanwhile, recent developments in US immigration enforcement continue to raise concerns among immigration attorneys, as officials are now issuing Requests for Evidence (RFEs) for H-1B and other employment-based immigrant petitions, asking for home addresses and biometrics—a move previously unseen in such cases. According to a Forbes report, these requests are linked to the US Citizenship and Immigration Services (USCIS), citing “adverse information” about individuals, though the specifics remain unclear. Meanwhile, it is still uncertain whether Trump-era officials have included employment-based visa applicants in broader efforts to identify and deport individuals who are lawfully residing in the United States, or if these requests are intended for a different purpose altogether.","excerpt":"Analysts attribute this to the fact that Indian IT companies have significantly increased their focus on local hiring, particularly in the Americas.","categories":["AI Features"],"tags":["Indian IT"],"author_name":"Shalini Mondal","publish_date":"2025-04-28T09:58:31","publication_year":"2025","word_count":1018,"keywords":["Go","API","AI","innovation","Aim","ViT","disruption","GAN","Indian IT","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","API","GAN","ViT","innovation","disruption","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/indian-it-services-firms-have-greater-insulation-from-h-1b-visa-disruption\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10118173,"title":"Top 10 Learning Management System (LMS) for AI Training and Development","content":"For enterprises aiming to foster technological advancement within their teams, choosing the right Learning Management System (LMS) is crucial. These platforms offer specialized training in artificial intelligence (AI), providing the tools necessary for businesses to stay competitive in a rapidly evolving digital landscape. Top 10 LMS for AI Training and Development Here’s a list of the top 10 LMS platforms that are perfect for enterprises looking to enhance their AI capabilities (in no specific order). 1. MachineHack for Business MachineHack for Business excels in transforming teams into AI powerhouses with its bespoke Learning Management System. It provides AI skill assessments, hackathons, and targeted training programs, making it an invaluable resource for practical and competitive AI learning. 2. ADaSci AI Academy ADaSci AI Academy focuses on Generative AI and MLOps, offering hands-on training to bridge the gap between AI-savvy professionals and the rest. Their courses like “Generative AI Application with Google Vertex AI” and “Mastering Prompt Engineering for LLMs” ensure that teams not only learn AI but become masters in applying it effectively. 3. Coursera for Business Coursera for Business offers a wide range of AI and machine learning courses developed in collaboration with leading universities and companies. Its enterprise plan includes comprehensive learning programs, employee progress tracking, and globally recognized certifications. 4. Udacity Udacity is renowned for its project-based Nanodegree programs in AI and machine learning, tailored for enterprises needing to close skills gaps with hands-on learning experiences that are reviewed by industry experts. 5. Pluralsight With a focus on technical training, Pluralsight provides AI and machine learning courses alongside skill assessments and role-based learning paths that align team learning outcomes with business objectives. 6. EdX for Business EdX for Business offers rigorous, university-level courses in AI from institutions like MIT and Harvard, tailored for companies that prioritize deep, academic learning. 7. DataCamp for Business Specializing in data science and analytics, DataCamp offers an interactive learning experience in AI and machine learning, perfect for hands-on skill development through coding exercises and real-world datasets. 8. LinkedIn Learning LinkedIn Learning features an extensive course library, including a wealth of AI and machine learning topics, making it suitable for professionals looking to upskill quickly within a corporate environment. 9. Simplilearn Simplilearn focuses on applied learning with professional certification training in AI and machine learning, delivered through live, instructor-led sessions and labs. 10. A Cloud Guru Ideal for teams utilizing cloud platforms for AI applications, A Cloud Guru specializes in cloud computing, offering hands-on labs and real-world scenarios that enhance learning in cloud-based AI technologies. These LMS platforms are designed to cater to the varied needs of large organizations, ensuring that teams not only learn AI technologies but are also able to effectively implement them in real-world applications. Compare the Best LMS Platforms for AI Training in 2024 MachineHack for BusinessAI Talent DevelopmentAI skill assessments, hackathons, bespoke LMSEnterprises focusing on AIADaSci AI AcademyGenerative AI, MLOpsHands-on courses on Generative AI, MLOps, Prompt EngineeringAI professionals, data scientistsCoursera for BusinessBroad educational offeringsCourses from top universities, tracking, certificationsEnterprises seeking academic partnershipsUdacityTech and AI skillsProject-based Nanodegree programs, industry expert reviewsEnterprises closing skills gapsPluralsightTech skillsTechnical courses, skill assessments, learning pathsTech-focused enterprisesEdX for BusinessAcademic rigorUniversity-level courses from MIT, Harvard, etc.Enterprises valuing deep learningDataCamp for BusinessData science and AIInteractive courses, coding exercises, real-world datasetsTeams needing hands-on data science trainingLinkedIn LearningBroad professional developmentExtensive course library, quick upskillingProfessionals in various industriesSimplilearnCertification trainingLive sessions, professional certification, hands-on labsEnterprises looking for certified trainingA Cloud GuruCloud computingCloud-based AI training, hands-on labs, real-world scenariosTeams using cloud for AI applications MachineHack for Business and ADaSci AI Academy distinguish themselves in the crowded LMS market through their singular focus on artificial intelligence. MachineHack for Business leverages its expertise by offering practical AI skill assessments and hackathons, making it highly relevant for enterprises seeking to directly enhance their teams’ AI capabilities. Similarly, ADaSci AI Academy’s focus on Generative AI and MLOps through hands-on training ensures that learners not only understand AI concepts but are fully prepared to implement these advanced technologies in real-world scenarios. This focused approach ensures that both platforms deliver highly specialized and effective training, making them the top choices for organizations committed to leading in AI innovation.","excerpt":"Discover the top LMS platforms that specialize in AI training, providing the tools and knowledge for enterprises to excel in the AI-driven digital landscape.","categories":["AI Features"],"tags":["ADAsci (Association of Data Scientists)"],"author_name":"AIM Media House","publish_date":"2024-04-14T13:40:41","publication_year":"2024","word_count":689,"keywords":["data science","machine learning","artificial intelligence","ADAsci (Association of Data Scientists)","AI","ML","MLOps","Aim","deep learning","analytics","generative AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","data science","analytics","generative AI","MLOps","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/top-10-lms-platforms-for-enterprise-ai-training-and-development\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167422,"title":"Cautio Gets ₹11 Cr Boost in Seed Round with MeitY Backing","content":"Bangalore-based AI-powered video telematics startup Cautio has raised ₹11 crore in its ongoing seed round. The company saw funding participation from notable investors, including 9Unicorns, Venture Catalysts, Antler India, Infinyte Club, CARS24 founders Vikram Chopra and Gajendra Jangid, and the Ministry of Electronics and IT (MeitY) through PIEDS-BITS Pilani. The company plans to use the funds to strengthen its R&D capabilities and further scale its technology. Cautio’s platform is designed to bring real-time intelligence to fleet operations, transforming road safety across India. “With AI-driven video telematics, we are changing this narrative. We are giving fleet operators, drivers, and passengers a chance to be safer, more responsible, and more accountable on the road,” said Ankit Acharya, co-founder and CEO of Cautio. Since its inception, Cautio has enabled over 300,00,000 km of safe travel and safeguarded more than 16 lakh trips. Its technology has helped reduce risks, enhance accountability, and optimise fleet operations. Today, it serves leading mobility and logistics providers such as Shoffr, Euro Cars, Namma Yatri, IITs, and Cityflo, operating across 35+ cities in India. The platform has already delivered over 50 crore location data points through its APIs. Investors see Cautio as more than a data company. “Cautio goes beyond just data collection; it transforms insights into action, improving driver behaviour, enhancing safety, and laying the foundation for autonomous mobility,” said Nitin Sharma, partner at Antler. Similarly, CARS24 founders added, “In India, road safety isn’t broken because we lack vehicles or infrastructure. It’s broken because accountability is almost invisible. Cautio’s AI tackles this problem head-on.” With this funding, Cautio aims to expand its technology across India, strengthen its AI capabilities, and work closely with fleets to build a proactive safety culture.","excerpt":"The funds will strengthen R&D and scale technology capabilities.","categories":["AI News"],"tags":["Cautio"],"author_name":"Merin Susan John","publish_date":"2025-04-08T14:24:38","publication_year":"2025","word_count":282,"keywords":["Go","API","funding","unicorn","programming_languages:R","AI","programming_languages:Go","Aim","Cautio","R","startup"],"extracted_tech_keywords":["AI","Aim","R","Go","API","startup","unicorn","funding","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/cautio-gets-%e2%82%b911-cr-boost-in-seed-round-with-meity-backing\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":30898,"title":"IIT Delhi Joins IBM’s AI Horizons Network To Accelerate New Tech Research In India","content":"Photo Caption 1: L-R (Mausam, Associate Prof. Dept. CSE; Sachindra Joshi, Senior Manager, IBM Research; Sanjiva Prasad, Professor & Head, Department of Computer Science & Engineering; L Venkat Subramaniam, Senior Manager, IBM Research; Prof. V. Ramgopal Rao, Director, IIT Delhi; Michael Karasick, Vice President, Global Labs, IBM Research; Prof M Balakrishnan, Dy Director (S&P); Sriram Raghavan, Vice President, IBM Research- India & Singapore & CTO IBM India\/ South Asia; Srikanta Bedathur, Associate Prof. Dept. CSE) It was just two months ago that IBM had announced an association with IIT Bombay to accelerate artificial intelligence research in India. This was the first institution outside North America to join the IBM AI Horizons Network. Now, IBM has announced that it has collaborated with IIT Delhi in a multi-year deal to work on AI. With this collaboration, IIT Delhi becomes one of the select institutions in India to join the network which also includes the likes of University of Michigan, Universite de Montreal, the University of Maryland at Baltimore County, UC San Diego and the University of Massachusetts at Amherst, University of Illinois Urbana-Champaign and more. IBM works with over eight universities globally. In a brief conversation with Analytics India Magazine, Dr Venkat Subramaniam, Senior Manager at IBM Research India, shared, “It [IIT-D] is a natural choice, given that it is one of the top academic institutions working in the area of AI and has world-renowned faculty in this space. There are synergies we can draw between the work we do at IBM Research lab and IIT Delhi students and faculties.” Key Focus Areas Of The Collaboration The aim is to discover novel AI techniques which can help organisations take informed decisions by being able to logically reason with their AI systems. AI solutions will be trained to comprehend complex questions using natural language techniques and derive new insights using domain knowledge. “The key areas include text and image comprehension, knowledge representation and reasoning. We will also work on making AI more transparent, where the machine is able to explain its reasoning process to a human user. We will apply these scientific advances to areas like medicine, financial services, technical support, agriculture,” told Dr Subramaniam to AIM. The company believes that today AI system is not trained to explain its decision-making process. With this partnership, IBM researchers along with professors from the Department of Computer Science and Engineering at IIT Delhi would aim to address this issue and conduct joint research to inculcate some key traits such as reasoning, comprehension and inferencing in AI systems. For instance, a procurement analyst may not be able to decide the right price for a commodity based on the recommendations shared by a trained AI assistant without knowing why it made those suggestions. The research will benefit sectors such as healthcare and medicine, finance, and customer support which deals with a complex set of questions and require reasoning. “While working with AI systems, organisations require explicit reasoning and comprehension to reach a particular conclusion. We believe advancement in AI can tackle such problems”, said Michael Karasick, Vice President, Global Labs, IBM Research in the press statement. Speaking on the relevance of collaboration Prof V Ramgopal Rao, Director, IIT Delhi, said “India has immense talent to accelerate innovation in AI and related technologies. We are happy to collaborate with IBM Research scientists and provide opportunities to our students and faculty colleagues to work on some of the complex problems around AI and apply the solutions to real-world scenarios.” “As an academic institution students and faculty are more tuned to the theoretical aspects of AI and by providing them real-life problems, IBM along with IIT Delhi wants to address the key research challenges. The future is about AI and through this collaboration with IIT Delhi, IBM also hopes to train the next generation workforce to work in cutting-edge problems in the AI field”, said Dr Subramaniam in an exclusive interaction with AIM, on why the company prefers collaboration with universities. The teams plan to publish their research in peer-reviewed academic journals and release datasets and open challenges to the research community to identify new areas in making AI decisions better. Some of the focus areas include deep learning, natural language processing, computer vision, and others, as well as their application to big societal challenges, ranging from aiding the understanding of disease to education and cybersecurity. On being asked about the progress with IIT Bombay, he told AIM that students and faculty from IIT-B along with IBM researchers have already begun working on imparting AI systems with the ability to comprehend spoken, written and visual data so that can drive their decision-making ability in a better way. “We will have some exciting announcements to share in the near future”, he said while signing off.","excerpt":"It was just two months ago that IBM had announced an association with IIT Bombay to accelerate artificial intelligence research in India. This was the first institution outside North America to join the IBM AI Horizons Network. Now, IBM has announced that it has collaborated with IIT Delhi in a multi-year deal to work on […]","categories":["AI News"],"tags":["IIT Bombay","ixigo","Quantum Computer"],"author_name":"Srishti Deoras","publish_date":"2018-11-30T11:30:31","publication_year":"2018","word_count":794,"keywords":["Go","IIT Bombay","artificial intelligence","AI","Quantum Computer","ixigo","computer vision","RAG","Aim","deep learning","analytics","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","computer vision","analytics","Aim","RAG","R","Go","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-delhi-joins-ibms-ai-horizons-network-to-accelerate-new-tech-research-in-india\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10061331,"title":"MeitY proposes data-sharing framework; plans on data monetisation","content":"The Ministry of Electronics and Information Technology (MeitY), India, published a draft data policy on February 21 which states all data collected, generated, and stored by every government ministry and the department will be open and shareable, excluding certain exceptions. “Draft India Data Accessibility & Use Policy 2022”, published by MeitY prescribes that a regulatory authority called the Indian Data Council (IDC) and an agency named India Data Office (IDO) will oversee framing metadata standards and enforcement, respectively. To promote innovation, minimally processed datasets will be made available at no cost. At the same time, certain datasets that have undergone value addition or transformation may be valued appropriately by the concerned Departments or Ministries. Any data sharing will happen within the legal framework of India. This will help in preventing the misuse of data and assure security, integrity, and confidentiality of the data. IDO will be set up by MeitY with an objective to streamline and consolidate data access and sharing of public data repositories across the government and other stakeholders. Every ministry\/department will have data management units headed by Chief Data Officers, who will work closely with IDO for ensuring the implementation of the data accessibility and use policy. All data owned by the government, ministry, the organisation shall be open and shareable by default, unless the data is categorised under the negative list of datasets or falls under the restricted category and will be shared with only trusted users, as defined by the ministry. All government\/ministries will identify all existing data assets and create detailed, searchable data inventories with clear metadata and data dictionaries. The government will integrate all data portals maintained by ministries and other departments through APIs with the open government data portal. The IDC will create an indicative framework for identifying high-value datasets. HVD will be defined on their degree of importance in the market, degree of socio-economic benefits, impact on India’s AI strategy and performance on global indices.","excerpt":"The government will make minimally processed datasets available at no cost.","categories":["AI News"],"tags":[],"author_name":"SharathKumar Nair","publish_date":"2022-02-22T19:17:10","publication_year":"2022","word_count":324,"keywords":["Go","API","programming_languages:R","AI","innovation","ML","programming_languages:Go","Rust","GAN","R"],"extracted_tech_keywords":["AI","ML","R","Go","Rust","API","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meity-proposes-data-sharing-framework-plans-on-data-monetisation\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171736,"title":"Over 20% of Indians on GitHub Used AI to Write Python Functions: Study","content":"A study published by the University of Utrecht, the Complexity Science Hub (Vienna), Corvinus University of Budapest, and the HUN-REN Centre for Economic and Regional Studies in Hungary outlined the use of AI to write code globally and its economic impact. The authors trained a neural classifier to detect AI-generated Python functions in 80 million GitHub commits by 200,000 developers over the past six years. “By December 2024, AI wrote an estimated 30.1% of Python functions from US contributors, versus 24.3% in Germany, 23.2% in France, 21.6% in India, 15.4% in Russia and 11.7% in China,” read the study. The authors also said that newer GitHub users use more AI than veterans, and using AI 30% of the time raises quarterly commits by 2.4%. By combining wage data with estimates of the amount of programming done in almost 900 US occupations, the study calculates that AI generates $9.6-14.4 billion annually in the US software sector. The researchers thoroughly trained the neural classifier using a vast dataset to ensure it correctly identifies AI-generated Python functions. They collected human-written Python functions from 2018 onwards, covering diverse examples across different time periods. They also created synthetic AI-written functions using a two-step process: one LLM described human functions, and another LLM generated corresponding code based on these descriptions. The researchers then used GraphCodeBERT, a cutting-edge model, to tokenise and embed the code before fine-tuning it into a classifier. The model was assessed with an out-of-sample ROC AUC score of 0.964, indicating high detection precision.","excerpt":"By combining wage data with estimates of the amount of programming done in almost 900 US occupations, the study calculates that AI generates $9.6-14.4 billion annually in the US software sector.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Supreeth Koundinya","publish_date":"2025-06-13T15:24:23","publication_year":"2025","word_count":250,"keywords":["programming_languages:R","AI","Git","BERT","Python","llm_models:BERT","programming_languages:Python","GitHub","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","Python","R","Git","GitHub","BERT","llm_models:BERT","programming_languages:Python","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/over-20-of-indians-on-github-used-ai-to-write-python-functions-study\/","complexity_score":4,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10012489,"title":"15 Latest Data Science Jobs Openings From Past Week","content":"The Sexiest Job of the 21st Century, data science jobs have gained much popularity in recent years. For this week’s data science job roles, we have listed 15 latest data science job openings that just posted the past week in locations like Gurgaon, Hyderabad, Bangalore and more. Up to 5 Years Experience Data Scientist at QuantumBlack Location: Gurgaon Responsibilities: As a data scientist at QuantumBlack, you will work in multi-disciplinary environments harnessing data to provide real-world impact for organisations globally. You will influence many of the recommendations the clients need to positively change their businesses and enhance performance. Apply here. Data Scientist II at HP Location: Bangalore Responsibilities: In this job, you will mine data using modern tools and programming languages, work with the project team to understand problem statements, initiates and direction, define and implement models to uncover patterns and predictions creating business value and innovation and other such. Apply here. Data Scientist: Advanced Analytics at IBM Location: Bangalore Responsibilities: In this role, you will work with IBM Q Start team on active exploratory research engagements to prepare for future use case commercialisation within a specific industry, interact with client data science teams to define promising areas for quantum, implement quantum approaches, which includes data pre-\/post-processing, running numerics and visualising data, etc. Apply here. Data Scientist at ABB Location: Bangalore Responsibilities: Your responsibilities include conceptualising and developing prescriptive analytic algorithms for rotating machinery and industrial structures based on vibration data, finding solution analytical algorithms and their output visualisations by packaging them as industrial business value apps, among others. Apply here. Data Scientist at Accenture Location: Hyderabad Responsibilities: You will be responsible for developing analytics-based solutions that produce quantitative and qualitative business insights, work with partners as necessary to integrate systems and data quickly and effectively, regardless of technical challenges or business environments, etc. Apply here. Data Scientist at Commonwealth Bank Location: Bangalore Responsibilities: This role will sit within the transparency and visualisation team, forming part of the broader RSO business unit. This team provides strategic guidance to the EGM Risk Strategy and Optimisation, Group CRO and the Risk Management Leadership Team to help drive better business and risk outcomes through analytically driven insights. Apply here. Business Analyst II, Data Sciences at VMWare Location: Bangalore Responsibilities: As a business analyst, you will directly work with various stakeholders across the world to disambiguate data as well as build, iterate and optimise ML solutions, and support data-driven decision-making across the organisation. Apply here. Deputy Manager, Marketing Data Science at Philips Location: Bangalore Responsibilities: As a deputy manager, marketing data science, you will leverage machine learning models to address key growth challenges such as lifecycle marketing, predictive LTV, cross-channel spend allocation, response modelling, campaign\/channel performance measurement methodologies, Program effectiveness and media attribution, etc. Apply here. Product Manager (Data Science \/ML\/AI\/SAAS Products) at Delhivery Location: Hyderabad Responsibilities: As a product manager data science, you will work closely with data science, engineering and design teams to define and build products that help us deliver high-precision and low-cost services, retrieve and analyse data, and establish critical business metrics to measure product success and drive alignment across all teams to improve them, etc. Apply here. More Than 5 Years Experience Data Scientist at Alstom Location: Bangalore Responsibilities: As a data scientist\/data science specialist at Alstom, you will help transform the data into tangible business value by analysing information, communicating outcomes and collaborating on product development. Apply here. Data Scientist & AI Expert at Siemens Technology Location: Haryana Responsibilities: In this role, you will be responsible for the design of software solutions based on requirements and within the constraints of architectural \/design guidelines, involved in the coding of features and\/or bug-fixing and delivering solutions adhering to coding\/scripting and quality guidelines, for self-owned components, etc. Apply here. Media Data Scientist at Freshworks Location: Bangalore Responsibilities: As a Media DS, you will be building models around audience conversion optimisation across the paid owned and earned media channels, help design at scale micro-segments which will enable better engagement and hence better conversion paths and more. Apply here. Data Scientist – 1 at Novartis Location: Hyderabad Responsibilities: As a data scientist – 1, you will design, develop and deliver various NLP based insights, outcomes and innovation and create “proof of concepts & blueprints” to drive faster, timely, highly precise, workable and proactive decision making based on data-based insights. Apply here. Senior Data & Applied Scientist at Microsoft Location: Bangalore Responsibilities: As a senior data and applied scientist, you will work on machine learning components in the whole sponsored search stack. You will work on the problems related to machine learning, deep learning, natural language processing, image understanding, optimisation, information retrieval, auction theory, among others. Apply here. Data Scientist at Ericsson Location: Bangalore Responsibilities: As a data scientist at Ericsson, you will be responsible for the development, management, execution and reactive maintenance activities that require a higher level of development. You will also ensure that the services provided to customers are continuously available and maintaining Service Level Agreement (SLA) performance levels. Apply here.","excerpt":"The Sexiest Job of the 21st Century, data science jobs have gained much popularity in recent years. For this week’s data science job roles, we have listed 15 latest data science job openings that just posted the past week in locations like Gurgaon, Hyderabad, Bangalore and more.   Up to 5 Years Experience Data Scientist at […]","categories":["AI Hirings"],"tags":["big data scale","data science job roles","Data Science Jobs","Data Scientists","impact of big data in education","jobs that ai will improve","Latest Data Science jobs","Scale Big Data","the need to scale with big data","what is big data coding"],"author_name":"Ambika Choudhury","publish_date":"2020-11-26T14:03:54","publication_year":"2020","word_count":840,"keywords":["jobs that ai will improve","TPU","the need to scale with big data","deep learning","impact of big data in education","R","data science","RAG","NLP","analytics","what is big data coding","Data Scientists","machine learning","AI","data science job roles","Latest Data Science jobs","ML","Data Science Jobs","Scale Big Data","big data scale"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","data science","analytics","RAG","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/15-latest-data-science-jobs-openings-from-past-week\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10000689,"title":"10 White Noise Websites Which Will Help You Sleep Like A Baby","content":"Lucky are the mortals who can catch the zzzs as soon as their heads touch the pillow. However, there are some who struggle to get sound sleep. Sleep disorders are related to the loss of ability to get quality sleep and can be caused by various reasons like stress, unhealthy lifestyle, fatigue, stress, travel or illness, among other reasons. A recent study has claimed that approximately 93 percent of Indians are sleep-deprived, and that most of us are actually unaware of the condition. To deal with the situation, white noises are now playing a key role. Researchers from across the globe are working on noises as a way to overcome sleep disorders — an example of which is white noise Just like white colour, which has different colours combined, for white noise, different frequencies are fused together to form white noise. It is used to mask high and low frequencies of noise pollution resulting in a good and quality sleep. In technical terms, white noise in signal processing is a random signal that has similar intensity at different frequencies emitting a constant power spectral density. There are many websites now that are constantly providing a helping hand to get us some sound sleep. In this article, we have covered 10 websites that can give you a quality sleep which you have been urging for a long time. 1| Soundrown Soundrown is an audio platform containing the sound of a coffee shop, rain, birds, etc. that allows you to listen what you love and create your own ambient sound and can also share it. 2| SleepyTi.me SleepyTi.me is a user-friendly tool or can be called as a bedtime calculator that enables you to set time to wake up just like an alarm does and also offers to set a suggested time for when you wake up. 3| Brain.fm This web really helps to those who need a piece of soulful music to fall asleep and is mainly designed for enhancement of focus, meditation, help to fall sound sleep, relaxation within a listening period of 10 to 15 mins. More of it, Android and iOS users can also get to use it as an app. 4| Moodturn This is a free web service that provides that soothing music to help relax and let us escape from the hustles and bustles of life. The music is played in a loop and one can set the theme as per one’s mood. The themes include night, dolphin, rain-forest, birds, storm, beach, etc. 5| Jazz and Rain If you are a jazz fan, then what could be more soothing than listening to jazz music and fall asleep. Yes, you heard right. This site provides a theme for the rainy day and plays jazz music in the loop. The volume for rain can be adjusted as per need and there are varieties of jazz music available in here. 6| Calm Radio Calm Radio is an internet radio that features multiple mixers of soothing sound including solo piano, dolphin sound, whale sound, nature sound, birds chirping, rain, etc. This is a perfect website for someone who is seeking relaxation or sound sleep. The subscription is free for the users that open up to 200 channels. 7| Moszen It is an extremely easy and user-friendly web service for a person who is dealing with problems like insomnia and other sleeping disorders. Soothing themes and music like whale sounds, river, raindrops, forest, etc. are contained so that one can easily listen and get quality sleep. 8| Relaxing World Relaxing World is a free web that gives you to configure your own ambiance with an extensive variety of soothing sounds. This web also allows you to share your created ambiances with your friends and loved ones. There is an advanced 3D option where you can locate the sound sources in various positions with respect to your location as to make the sound surrounds the listener. 9| Rainy Mood If you want to experience rain, no other website is as good as this one. A perfect website for a rain lover that loops for 30 minutes with high-quality sound, this web helps to focus, relax and a good sleep. 10| GitHub Audio Lastly, we have something for the tech geeks too! What’s better than GitHub Audio? This site generates soothing music not just for relaxing after a hectic schedule but also helps to fall asleep. The amazing part is, the soothing sound is algorithmically generated based on the GitHub events.","excerpt":"Lucky are the mortals who can catch the zzzs as soon as their heads touch the pillow. However, there are some who struggle to get sound sleep. Sleep disorders are related to the loss of ability to get quality sleep and can be caused by various reasons like stress, unhealthy lifestyle, fatigue, stress, travel or […]","categories":["AI Trends"],"tags":["meditation"],"author_name":"Ambika Choudhury","publish_date":"2018-12-27T13:51:46","publication_year":"2018","word_count":748,"keywords":["meditation","Go","programming_languages:R","AI","programming_languages:Go","Git","Aim","GitHub","R"],"extracted_tech_keywords":["AI","Aim","R","Go","Git","GitHub","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-white-noise-websites-which-will-help-you-sleep-like-a-baby\/","complexity_score":4,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10103783,"title":"5 Spectacular New Features of RunwayML","content":"With video-generation tools blurring the line between reality and AI, the continuous developments are only making it easier for people to create stunning visuals. RunwayML, an applied AI research company for AI video editing and generation, which is capable of making videos from short text-based prompts, has released a series of new updates and features. The updates are said to enhance user experience by providing more control, fidelity and expressiveness. Here are some of the latest features of RunwayML that transcend the boundaries of creativity. Motion Brush Motion Brush, a cutting-edge interface for generative models, is empowering artists with controlled movement options for enhanced creative expression. This tool addresses the need for increased utility and expressiveness in generative models, providing users with precise control over their creations. Users can paint an area or subject, and choose appropriate directions and intensity to the motion, thereby allowing one to transform any kind of images or portion of images into a video. This has been called a much-needed feature with many sharing their experiments with the new tool. A few users have also tried running this feature on images generated from Midjourney. RunwayML's new Motion Brush feature is unreal.It lets you pick exactly where to add movement in your image.Talk about a game-changer. pic.twitter.com\/whyK3sxITO— Bryan Marley (@_bryanmarley) November 27, 2023 Gen 2 Style Presets The company believes that control is not just about motion but also art direction and style. Style presets in Gen2 allows one to create content using curated styles without the need of complex prompting. Glossy animations, and grainy retro film stocks are some of the many styles that can be brought to stories. https:\/\/twitter.com\/iamneubert\/status\/1727719096706543726 Director Mode Updates The advanced camera control on Director’s mode has been updated with more nuances. With this, a more granular level of control can be obtained. Furthermore, a user can fine-tune camera moves using fractional numbers for improved precision. 3) Director Mode UpdatesDirector Mode’s advanced camera controls have been updated to allow for a more granular level of control. Now you can adjust camera moves using fractional numbers for greater precision and intention. pic.twitter.com\/zUbtIrXvFC— Runway (@runwayml) November 20, 2023 New Image Model Update With the latest update, RunwayML has brought in features to improve fidelity and consistency. The higher resolution generations are now available in options such as text-to-image, image-to-image and image variation. By implementing these features, it also improves storytelling control. 4) New Image Model UpdateImproved fidelity, greater consistency and higher resolution generations are now available in Text to Image, Image to Image and Image Variation. Add these tools to your Image to Video workflow for more storytelling control than ever before. pic.twitter.com\/3rvvhP8WG5— Runway (@runwayml) November 20, 2023 Combined Motion and Camera Control To achieve higher precision and control over an image, combining the tools can lead to desired results. Users can fine-tune it as per their desired style with the ‘style preset’ option. This will enable better expressibility too. Continuing to experiment with @runwayml Motion Brush Beta feature.Result from my stacked image with ⬆️vertical Motion Brush movement and extended with Camera Motion zoomed in at 1.7 speed.Original image in the next post: pic.twitter.com\/vx7NKzgtHz— Heather Cooper (@HBCoop_) November 20, 2023 Runway has been working aggressively to release updates in the AI image space to compete with other players such as Midjourney and Stable Diffusion that released a new feature where meme images can be converted to videos. Runway has also received significant funding. The company raised funds of over $141 million from major tech giants such as Google, Nvidia, Salesforce and others. Furthermore, Runway AI announced its partnership with Canva, a graphic-design tool company to create generative AI videos and provide Gen-2 AI technology to Canva’s ecosystem.","excerpt":"AI video tool Runway releases new features to enhance user-friendliness with improved output.","categories":["AI Trends"],"tags":["AI Video Generation Models","Canva","Google","MidJourney","NVIDIA","Runway","Salesforce","Stable Diffusion"],"author_name":"Vandana Nair","publish_date":"2023-11-28T15:57:12","publication_year":"2023","word_count":614,"keywords":["Go","MidJourney","funding","AI","Stable Diffusion","ML","Canva","AI Video Generation Models","BERT","stable diffusion","Runway","Google","Salesforce","generative AI","ViT","NVIDIA","AI research","R"],"extracted_tech_keywords":["AI","ML","generative AI","R","Go","BERT","stable diffusion","ViT","funding","AI research"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-spectacular-new-features-of-runway-ml\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10071642,"title":"Bitter Cloud Competition is turning Rivals into Allies","content":"Recently, Oracle Corp and Microsoft Corp made a splash in cloud integration with the announcement of interoperability of their clouds at Inspire, an online event for Microsoft partners. The availability of Oracle Database Service for Microsoft Azure will allow Azure users to access and monitor the Oracle database services in Oracle Cloud Infrastructure (OCI). With the introduction of the duo’s multi-cloud database service, it just got easier to stick together the databases in OCI with workloads running in Microsoft Azure. In addition, the users can migrate or build new applications on Azure and tap into the high performance and availability guaranteed by Oracle. Why partner with a competitor? “Microsoft and Oracle have a history of working together to support the needs of our joint customers. This partnership is an example of how we offer customer choice and flexibility as they digitally transform with cloud technology,” said Corey Sanders, corporate VP of Microsoft Cloud for Industry and Global Expansion, in a statement. The companies said that enterprises could seamlessly connect Azure services, like Analytics and AI, to Oracle Cloud services, like autonomous databases. The partnership has attracted high-profile customers, including AT&T, Veritas, Marriott International, Swiss and SGS. The new service connects Oracle’s database service directly to the Azure cloud, eliminating previously required custom work. Oracle and Microsoft have won and protected market share for decades by competing at each technology tier and integrating competing technologies. On Microsoft’s side, this partnership shows a willingness to collaborate with any market leaders and to make it easier for Microsoft clients to work with their client of choice. Previously, Microsoft Azure partnered with IBM to offer enterprise security solutions. So, Microsoft is not alone in Google and AWS partnerships. Earlier in May 2022, AWS’ collaboration with IBM allowed joint customers to make various IBM software available on AWS Marketplace. Adding to the partnerships, in 2020, Google joined with IBM to catch their larger cloud services competitors and offer their customers the best of both clouds. Towards consolidation Companies are spending a fortune on the cloud, but the market is shrinking, and names like AWS, Azure, and GCP are capturing the lion’s share of it. Companies building a cloud strategy need to factor cloud computing trends into their plans and prepare to survive major shifts due to the failure, acquisition or discontinuation of a cloud provider that they rely on. Mordor Intelligence predicted a 21% compound annual growth rate for hybrid cloud through 2026. IBM called it a $1 trillion market with plenty of opportunity for channel partners to cash in. As per Microsoft’s Q4 FY2022 Earnings Report, investors will focus on revenue growth in Azure and other cloud services. However, Microsoft still has to compete with Amazon, Alphabet Inc.’s Google cloud and other competitors. The cloud computing market is growing, and Azure is growing faster than Microsoft as a whole. However, Microsoft still has to compete with Amazon, Alphabet Inc.’s Google Cloud, and other smaller rivals. This holds significance especially since work-from-home and hybrid work arrangements that began during the pandemic may be here to stay, boosting the market for cloud services.","excerpt":"The success of an alliance depends on both partners benefitting from the existing market or from gaps in the market that competitors are yet to spot.","categories":["IT Services"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2022-07-27T16:36:00","publication_year":"2022","word_count":517,"keywords":["Go","GCP","AWS","cloud computing","AI","R","ML","Git","analytics","Azure"],"extracted_tech_keywords":["AI","ML","analytics","cloud computing","AWS","Azure","GCP","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/it-services\/bitter-cloud-competition-is-turning-rivals-into-allies\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10134835,"title":"After Microsoft Azure and GCP, AWS Finally Partners with Oracle","content":"As reported by AIM last year, Oracle has finally partnered with AWS to complete its multi-cloud strategy, following its partnerships with Microsoft Azure and Google Cloud. Oracle has partnered with Amazon Web Services (AWS) to launch Oracle Database@AWS. This new offering enables customers to access Oracle Autonomous Database and Oracle Exadata Database Service on AWS infrastructure, simplifying the migration and deployment of enterprise workloads to the cloud. The integration allows customers to manage Oracle databases with tools like AWS Management Console and Command Line Interface (CLI). They will also benefit from seamless connections between Oracle databases and AWS applications such as Amazon Elastic Compute Cloud (EC2), AWS Analytics, and artificial intelligence (AI) services like Amazon Bedrock. Oracle Database@AWS will be available in preview by the end of 2024, with full availability planned for 2025. “We are seeing huge demand from customers that want to use multiple clouds,” said Larry Ellison, Oracle Chairman and CTO. “To meet this demand and give customers the choice and flexibility they want, Amazon and Oracle are seamlessly connecting AWS services with the very latest Oracle Database technology, including the Oracle Autonomous Database. With Oracle Cloud Infrastructure deployed inside of AWS datacenters, we can provide customers with the best possible database and network performance,” he added. AWS chief Matt Garman said that many organisations have been running Oracle workloads on AWS since 2008, and this partnership provides customers with a unified experience for migrating and managing their Oracle databases. Customers will also benefit from zero-ETL integration between Oracle Database and AWS Analytics services, enhanced migration tools like Oracle Zero Downtime Migration, and unified billing and support experiences. Additional services include database backup integration with Amazon Simple Storage Service (S3) for disaster recovery. Fidelity, Best Buy, and Vodafone have expressed optimism about the partnership, with leaders from each organization emphasising how this collaboration will accelerate digital modernization and improve business outcomes. Both companies will co-market Oracle Database@AWS, aiming to serve industries like financial services, healthcare, manufacturing, and retail globally.","excerpt":"Oracle Database@AWS will be available in preview by the end of 2024, with full availability planned for 2025.","categories":["AI News"],"tags":["Oracle"],"author_name":"Siddharth Jindal","publish_date":"2024-09-10T10:01:37","publication_year":"2024","word_count":332,"keywords":["Go","artificial intelligence","AWS","AI","R","ML","Oracle","RAG","Aim","analytics","Azure"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","Aim","RAG","AWS","Azure","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-finally-succumbs-to-oracle\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10164880,"title":"How Cognite is Building its ‘ChatGPT for Industrial AI’ from Bengaluru","content":"Cognite, a global industrial AI and data company, is expanding its footprint in India with its new global capability centre (GCC) in Bengaluru in March. After setting up first in Oslo, followed by Texas and Japan, the team has now decided that India is the right place to attract the top AI and ML talent. Cognite specialises in providing data and AI solutions tailored for asset-heavy industries such as manufacturing, oil and gas, and energy. These industries operate with a complex network of IT and operational technology (OT) systems, requiring seamless access to diverse data sources. Under the leadership of Guru Ananthanarayanan, who was appointed the managing director of Cognite India in August last year, the company aims to make the India Centre the R&D hub for scaling its agentic AI framework, Atlas AI. “Large industrial companies have multiple IT and OT systems — SAP, Salesforce, ServiceNow, SharePoint folders, IoT sensors, SCADA (Supervisory Control and Data Acquisition) systems, PLCs (programmable logic controllers), and more,” Ananthanarayanan told AIM. “A service technician might have to access 10 to 15 different systems to gather relevant data.” Using Cognite, end-users can simply log in and query complex datasets in natural language. The AI-powered platform retrieves relevant information across systems, analyses it, and provides actionable insights, making industrial AI as intuitive as ChatGPT for manufacturing and energy companies. How Atlas AI is Taking Shape Cognite has built its AI framework around three core areas — asset performance, predictive reliability, and operational efficiency. AI can not only predict failures but also suggest multiple resolution strategies with their expected impact. Moreover, it enables autonomous execution, reducing the need for manual intervention. “For companies like Aker BP, ExxonMobil, Yokogawa, and Vedanta, Cognite’s AI continuously monitors sensor data to predict failures before they happen,” Ananthanarayanan said. “It doesn’t stop at prediction; it provides resolution options and executes them autonomously, ensuring uninterrupted operations.” A significant part of Cognite’s innovation strategy revolves around Atlas AI, its no-code, agentic AI framework. “With Atlas AI, a business user can simply type a request to generate an agent for root cause analysis of a specific asset,” Ananthanarayanan further mentioned. “The AI handles the coding in the background and generates an agent that continuously monitors and provides insights.” Atlas AI integrates multiple LLMs from OpenAI, Meta, and Anthropic, dynamically selecting the most suitable model based on the problem statement. “This approach ensures that businesses don’t need specialised AI expertise to leverage advanced AI solutions,” he added. Unlike generalised AI platforms that cater to multiple industries, Cognite specialises in AI solutions tailored for asset-intensive sectors like manufacturing, oil and gas, and energy. “Think of it as the difference between a general hospital and a neurosurgery specialty centre,” Ananthanarayanan explained. “We are not a jack-of-all-trades AI platform. We are specialists in industrial AI, delivering targeted solutions for manufacturing and energy companies.” Cognite’s domain-specific expertise enables it to build proprietary knowledge graphs that contextualise data across disparate systems. “Industrial data is highly inconsistent. The same pressure tank might be referred to differently in different systems,” he noted. “Our algorithms intelligently map these inconsistencies, providing users with a unified view of their assets,” he further said. “It’s like a ChatGPT for industrial AI.” Why Bengaluru? The Talent Advantage Cognite’s decision to establish its India GCC in Bengaluru stems from the city’s abundant talent pool and deep expertise in AI and data engineering. The company plans to onboard over 150 engineers by the end of 2025, focusing on AI\/ML, data science, software engineering, and industrial domain expertise. “While Pune and Hyderabad were considered, Bengaluru’s unmatched talent density made it the ideal choice,” Ananthanarayanan said. “For niche roles like OPC UA protocol experts, which are rare even globally, Bengaluru has a significant talent pool.” With AI-driven automation reshaping job roles, Cognite is hiring engineers who bring not just their expertise but also their own AI agents. “We are now hiring individuals who don’t just code, but also bring AI-powered tools to enhance their efficiency.” “We are not just a backend development centre. India will lead end-to-end product ownership, from conceptualisation to implementation,” Ananthanarayanan affirmed.","excerpt":"Under the leadership of Guru Ananthanarayanan, the company aims to make the India Centre the R&D hub for scaling its agentic AI framework, Atlas AI.","categories":["GCC"],"tags":["Bengaluru","ChatGPT","Salesforce"],"author_name":"Mohit Pandey","publish_date":"2025-03-03T09:00:00","publication_year":"2025","word_count":683,"keywords":["data science","ChatGPT","agentic AI","Anthropic","OpenAI","AI","ML","RAG","Ray","Aim","Salesforce","Bengaluru"],"extracted_tech_keywords":["AI","ML","data science","agentic AI","ChatGPT","OpenAI","Anthropic","Aim","Ray","RAG"],"url":"https:\/\/analyticsindiamag.com\/gcc\/how-cognite-is-building-its-chatgpt-for-industrial-ai-from-bengaluru\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10167188,"title":"ST Telemedia Global Data Centres (India) Launches AI-Ready Data Centre in Kolkata with INR 450 Crore Investment","content":"ST Telemedia Global Data Centres (India) (STT GDC India) has launched a new AI-ready data centre campus in New Town, Kolkata on Monday, with an investment of INR 450 Crore. Spanning 5.59 acres, this cutting-edge facility is designed to meet the growing demand for AI computing. It will provide high-density rack configurations, advanced cooling systems, and scalable, modular designs to support AI and other high-performance computing workloads. The campus has received the TIA-942 Rated-3 Design certification, highlighting its world-class infrastructure and reliability. The data centre has an IT power capacity of 10 MW, with the first building offering 5-6 MW. Its power design includes N+2C redundancy for reliable power availability, along with a radial N+N configuration for main power incomers. The campus provides a significant boost to digital infrastructure creation in the eastern part of the country with scalable capacity of up to 25 MW in terms of overall IT load. Bimal Khandelwal, CEO of STT GDC India, said, “Our Kolkata campus is specifically designed to support the burgeoning AI ecosystem, from startups developing local language AI models to enterprises deploying large language models. The facility’s high-performance computing capabilities and low-latency connectivity will empower organisations to build and deploy AI solutions that drive digital transformation across sectors.” The facility’s design ensures zero Single Points of Failure (SPOF), with flexibility for future technologies like liquid cooling. It also prioritises sustainability with low PUE (Power Usage Effectiveness) cooling, water conservation measures, and the use of low-GWP refrigerants to reduce its carbon footprint. Scheduled to go live in Q2 2025, the Kolkata data centre will expand STT GDC India’s network to 30 data centres across 10 cities, offering a total IT load capacity of 390 MW. Its strategic location in New Town’s Silicon Valley will serve as a key hub for AI development, supporting enterprises, hyperscale cloud service providers, and government organisations","excerpt":"Spanning 5.59 acres, this cutting-edge facility is designed to meet the growing demand for AI computing","categories":["AI News"],"tags":["data centre"],"author_name":"Shalini Mondal","publish_date":"2025-04-03T13:09:16","publication_year":"2025","word_count":309,"keywords":["Go","programming_languages:R","AI","digital transformation","Scala","Git","ViT","GAN","R","data centre","startup"],"extracted_tech_keywords":["AI","R","Go","Scala","Git","GAN","ViT","digital transformation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/st-telemedia-global-data-centres-india-launches-ai-ready-data-centre-in-kolkata-with-inr-450-crore-investment\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":24161,"title":"How Nikon, Canon Are Using AI To Stay Relevant In Today’s Instant Photography Ecosystem","content":"Thanks to many noted mobile phone manufacturers in the market, AI-powered smartphone cameras have become the new normal. These companies are increasingly integrating AI-focused software and hardware into their devices, leaving professional camera makers Canon and Nikon behind. The camera market has been impacted ever since iPhones took the mobile photography by storm. According to a data published by photo sharing platform Flickr, iPhone left brands such as Canon and Nikon miles behind. In fact, smartphone camera accounted for 50 percent of all the pictures uploaded to the site in 2017. Why Do People Prefer Smartphone Cameras Over DSLRs Smartphone cameras already rival their bulky brethren in users convenience and quality. According to a survey, more than 67 percent of young Indian travellers prefer smartphones over DSLR for photography.  Apart from convenience and image quality, several reasons why an average user prefers a smartphone over a DSLR: Pocket-friendly and easy to use Photo editing apps available at a click of a button Pictures can be shared instantly on social media Can easily shoot slow-motion and time-lapse videos Create easy panorama images without complex editing For example, if one is to create the currently-popular effect of taking a picture with the subject as the main focus, and the rest in a blurry background, the photographer would have to pick up a DSLR and manually manipulate the depth of field, focus on the subject, etc. On the other hand, an AI-powered smartphone camera can do it with a single click. Off late, we have seen many dual-lens smartphones which come with different focal lengths that use artificial intelligence to create an aesthetically pleasing image. Commonly known as the ‘bokeh’ effect, here, machine learning identifies the subject and blurs the rest of the image automatically. How AI Improves The Image Quality AI in newer versions of smartphones recognise images and process them like data. When taking a picture, the phone identifies the object being photographed. A correction filter is applied, which results in better quality images. The AI also compensates for low-resolution images. If the picture being viewed on the device is of a lower resolution, the AI-vision engine predicts the pixels which are missing. This allows a better quality full-display image. Almost every smartphone brand is now rushing to adopt AI into their devices for creating better quality images. AI has become important for smartphones because they are trying to replicate the optical zoom. Top-end smartphones like iPhone X features AI which primarily focuses on imaging and animation. Sensing the future of AI-enabled cameras, Google too has introduced AI in its smartphones Pixel 2 and Pixel 2 XL, which feature a portrait mode. The portrait mode automatically applies an out-of-focus effect within four seconds, making the background blur and subject stand out. Pixel smartphones can also set a different level of depth of field after calculation based on convolutional neural network (CNN). The CNN filters are trained on millions of pictures of people and can differentiate which pixels are people and which are not. Google had previously launched Clips, a camera which uses AI to figure out when to take a picture or video. It automatically adjusts its resolution according to the situation. Huawei P20 Pro features AI-powered triple rear camera system which has a 40MP RGB sensor, 20MP monochrome sensor and 8MP sensor with a telephoto lens for sharpness, colour accuracy, focus and contrast. Professional Camera Makers Foray Into AI Looking at the current trend of AI-powered smartphones, Canon has also joined the AI bandwagon and has emphasised that they would now focus on making cameras that would “create great shots for the millennials.” The professional camera market has caught on to many of the trends such as using smarter chips, WiFi support, smart functions, superior speed and enhanced performance. According to Canon, camera-centric smartphones are giving the company a renewed push to come up with ‘serious’ cameras. “The relationship between a smartphone camera and a real camera is interesting. With smartphones, more and more people are now clicking pictures which have given us a thrust as people are now keen to have better image quality with AI embedded into it. This is a good takeaway from the smartphone market to build next-gen cameras,” said Canon India’s President and CEO Kazutada Kobayashi. However, Canon Chairman and CEO Fujio Mitarai admitted that the company is lagging behind in terms of innovation. “Our primary management goal this year is to raise our antennas high toward cutting-edge technology. It is on this point where we lag behind other companies. We will open up a research and development centre in the US’s Silicon Valley, where we will actively adopt new technology,” he said in a report. Nikon’s MRMC recently introduced two new automatic camera operating systems that use AI to track subjects without human intervention. These cameras are called as Polycam Player and Polycam Chat. Polycam Player is built for live sporting events. It can easily identify and track individually athletes and also can achieve angles not easily obtained by a human camera operator. With ChyronHego TRACAB player, the Polycam Player can physically move the camera as it tracks the subject and adjust the camera focus and zoom automatically. It can even track multiple players. Polycam Chat is designed for in-studio work. It is an automated system that can keep up to four people in focus and in the shot by using both face detection and limb recognition, allowing it to automate the camera’s operation. DSLR Vs AI-Powered Smartphone Cameras Despite all the advantages that smartphone cameras provide, it may not be the time to chuck out the DSLR quite yet. The DSLR cameras are still better than smartphone cameras in expressing details with more accuracy as they are equipped with huge interchangeable lenses, which photographer can manually control for aesthetic images. Despite the convenience and better quality image, it will be some time before a smartphone camera can replace a DSLR camera. The in-built AI does boost colour, saturation and highlights for more pleasing pictures, but will take more than AI to get DSLR quality pictures. DSLRs have much larger sensors to capture light and colour data along with a dedicated image processing chip that helps these cameras to produce pictures with more detail and more accurate colours. The DSLR lets one shoot a wide range of subjects — shallow depth of field for portrait, macro for extreme close up photography, ultra wide-angle for landscapes, telephoto for sports, perspective control for architecture. Many AI-powered smartphone cameras are now capable of shooting an aesthetic photo with more face detection, bokeh effect and shallow depth of field and can produce and can produce raw files as well as .jpeg which means that when it comes to editing, you have more control over how an image looks. It uses ML algorithms to help capture scenes. Many smartphones companies like Vivo are also using AI into their smartphone cameras’ HDR mode. For every photo taken, 12 separate frames will be mßerged to better light the final picture. Vivo smartphones super HDR also adapt and choose different parts of the frames to include best highlights and details for balanced and natural photos. To sum up, with the integration of AI into cameras, the future of technology looks promising, both for DSLR cameras and for smartphones. But a smartphone will never supersede DSLRs. However, they will remain powerful tools that can be carried in pockets wherever we go. DSLR is great for professional photographers in a variety of fields including sports, fashion etc who need to produce technically accomplished photographs with more details. Whereas smartphone cameras are useful when you want to capture, edit and share photos in real-time. In the end, it depends on the photographer who holds the camera as the person behind the lens is as important as the camera, its tools and specs.","excerpt":"Thanks to many noted mobile phone manufacturers in the market, AI-powered smartphone cameras have become the new normal. These companies are increasingly integrating AI-focused software and hardware into their devices, leaving professional camera makers Canon and Nikon behind. The camera market has been impacted ever since iPhones took the mobile photography by storm. According to […]","categories":["IT Services"],"tags":["camera","smartphones","system integration companies"],"author_name":"Smita Sinha","publish_date":"2018-05-02T07:27:37","publication_year":"2018","word_count":1307,"keywords":["Go","camera","artificial intelligence","machine learning","AI","neural network","ML","smartphones","RAG","system integration companies","Ray","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","Ray","RAG","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/it-services\/how-nikon-canon-are-using-ai-to-stay-relevant-in-todays-instant-photography-ecosystem\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10098239,"title":"NVIDIA’s AI Supremacy is All About CUDA","content":"By now, it is clear that no matter who wins the AI race, the biggest profiteer is NVIDIA. It’s common knowledge that the company is a market leader in the hardware category with its GPUs being widely used by all AI-focused companies in the world. That’s not all. NVIDIA, the biggest chip company in the world, is leading the battle from the software side of things as well, with its CUDA (Computing Unified Device Architecture) software. CUDA, in essence, is like the magic wand that connects software to NVIDIA GPUs. It’s the handshake that enables your AI algorithms to work with the computing power of these graphical beasts. But to NVIDIA’s advantage, CUDA isn’t just any ordinary enchantment, but a closed-source, low-level API that wraps the software around NVIDIA’s GPUs, creating an ecosystem for parallel computing. It’s so potent that even the most formidable competitors such as AMD and Intel struggle to match its finesse. While other contenders such as Intel and AMD attempt to juggle one or the other, NVIDIA has mastered the art of both. Their GPUs are sleek, powerful, and coveted – and it’s no coincidence that they’ve also laid down the foundations of software that make the most of these machines. Software companies can’t just waltz in and claim the crown to replace NVIDIA, they lack the hardware prowess. On the flip side, hardware manufacturers can’t wade into the software territory without struggling. This has made CUDA the winning ingredient for NVIDIA in AI. Undisputed but vulnerable NVIDIA built CUDA in 2006 with parallel computing for processing on multiple GPUs simultaneously. Earlier, developers were using models like Microsoft’s Direct3D or Linux’s OpenGL for computational purposes on GPUs, but lacked parallel computing capabilities. After the launch of CUDA, businesses began tailoring their strategies to adopt the software. OpenCL by Khronos Group was the only potential competitor released in 2009. But by then all companies had already started leveraging CUDA, leaving no room or need for it. NVIDIA’s current strategy sounds all great, but there are some major drawbacks in it as well. Though CUDA is a moat for NVIDIA, the company’s pursuit of an upmarket strategy, focusing on high-priced data centre offerings, might let other companies be able to catch up with their software. Moreover, the market is rife with a GPU shortage that feels almost mythical, but a few are willing to forsake NVIDIA’s wares for alternatives like AMD or Intel. It’s almost as if tech aficionados would rather gnaw on cardboard than consider a GPU from another company. NVIDIA’s maintenance of its current dominance is rooted in removing the RAM constraints within its consumer grade GPUs. This situation is likely to change as necessity drives the development of software that efficiently exploits consumer-grade GPUs, potentially aided by open-source solutions or offerings from competitors like AMD and Intel. Both Intel and AMD stand a chance at challenging NVIDIA’s supremacy, provided they shift away from mimicking NVIDIA’s high-end approach and instead focus on delivering potent, yet cost-effective GPUs, and build open source solutions. Crucially, they should differentiate themselves by avoiding artificial constraints that limit GPU capabilities, which NVIDIA employs to steer users towards their pricier data centre GPUs. Even after these existing constraints, a lot of developers choose NVIDIA’s consumer grade GPUs over Intel or AMD for ML development. A lot of recent development in these smaller GPUs has led to people shifting to them for deploying models. There is another competitor coming up Interestingly, OpenAI’s Triton emerges as a disruptive force against NVIDIA’s closed-source stronghold with CUDA. Triton, taking Meta’s PyTorch 2.0 input via PyTorch Inductor, carves a path by sidestepping NVIDIA’s CUDA libraries and favouring open-source alternatives like CUTLASS. While CUDA is an accelerated computing mainstay, Triton broadens the horizon. It bridges languages, enabling high-level ones to match the performance of lower-level counterparts. Triton’s legible kernels empower ML researchers, automating memory management and scheduling while proving invaluable for complex operations like Flash Attention. Triton is currently only being powered on NVIDIA GPUs, the open-source reach might soon extend beyond, marking the advent of a shift. Numerous hardware vendors are set to join the Triton ecosystem, reducing the effort needed to compile for new hardware. NVIDIA, with all its might, overlooked a critical aspect – usability. This oversight allowed OpenAI and Meta to craft a portable software stack for various hardware, questioning why NVIDIA didn’t simplify CUDA for ML researchers. The absence of their hand in initiatives like Flash Attention raises eyebrows. NVIDIA has indeed had the upper hand when it comes to product supremacy. But let’s not underestimate the giants of tech. Cloud providers have rolled up their sleeves, designing their own chips that could give NVIDIA’s GPUs a run for their transistors. Still, all of this is just wishful thinking as of now.","excerpt":"CUDA is a moat for NVIDIA. But the company’s pursuit of an upmarket strategy, focusing on high-priced data centre offerings, might let other companies be able to catch up with their software","categories":["Global Tech"],"tags":["AMD","Intel","NVIDIA"],"author_name":"Mohit Pandey","publish_date":"2023-08-08T14:16:18","publication_year":"2023","word_count":799,"keywords":["CUDA","Go","AMD","OpenAI","AI","PyTorch","ML","RAG","Aim","NVIDIA","R","Intel"],"extracted_tech_keywords":["AI","ML","OpenAI","Aim","PyTorch","RAG","CUDA","R","Go","CUDA"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/nvidias-ai-supremacy-is-all-about-cuda\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10170214,"title":"PierSight Secures Funding to Advance Maritime Surveillance Satellites","content":"Space-tech startup PierSight announced that it has raised an undisclosed amount as additional capital in a seed extension round, bringing the total funding to $8 million. New investors, including CE-Ventures, Sustainable Ocean Alliance (SOA), and individual backers such as Nandan Reddy (co-founder at Swiggy), Natasha Malpani Oswal (investor at Boundless), and Sahil Lavingia (founder at Gumroad), participated in the round. The company said the investment will aid the development of the world’s first satellite constellation combining Synthetic Aperture Radar (SAR) and Automatic Identification System (AIS) for round-the-clock maritime intelligence. PierSight, based in Ahmedabad, is developing technologies to detect illegal fishing, marine pollution, and insurance fraud, while also monitoring Exclusive Economic Zones and supporting port operations. Its first commercial SAR+AIS satellite is scheduled for launch in mid-2026, following the successful deployment of its demonstration satellite, Varuna Mission, in December 2024, onboard the ISRO SpaDeX mission. Founded by ex-ISRO scientist Gaurav Seth and former National Instruments engineer Vinit Bansal,  the company aims to provide persistent, all-weather visibility into maritime activity through advanced satellite technology. The startup continues to hire across engineering roles to support its expansion. “We have the necessary resources in place to reach our upcoming targets,” Seth said. “This round was about welcoming partners who bring frontline experience in ocean sustainability and port operations.” In a recent interview with AIM, Seth elaborated on the rationale behind focusing on maritime surveillance. “Traditionally, this was done by flying helicopters, aircraft, speedboats, ships, but they weren’t as optimised as a satellite. A satellite can cover 400 square kilometres every second. No other method can cover that,” he elaborated. Aligning with the UN Sustainable Development Goal 14, focusing on the conservation and sustainable use of oceans, Matt Mulrennan, head of investments at SOA, “PierSight’s ocean-focused satellite constellation will be highly scalable for conservation outcomes by exposing the ‘dark side’ of ocean activities like illegal fishing and offshore pollution dumping.” Alongside its satellite development, PierSight is also building a maritime analytics platform, MATSYA, and testing a drone-borne SAR system. Both are expected to be commercially available by the end of 2025.","excerpt":"PierSight’s MATSYA platform and drone-borne SAR system are set for commercial release by the end of 2025.","categories":["AI News"],"tags":["Funding","PierSight","seed funding","space startup","space technology"],"author_name":"Sanjana Gupta","publish_date":"2025-05-20T10:41:41","publication_year":"2025","word_count":346,"keywords":["seed funding","Go","Funding","API","AI","space technology","Scala","PierSight","Aim","space startup","analytics","ViT","R","analytics platform","startup"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","Scala","API","ViT","analytics platform","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/piersight-secures-funding-to-advance-maritime-surveillance-satellites\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":40872,"title":"Paytm Introduces AI-Powered Router Engine To Enable Faster Checkouts","content":"India’s leading fintech firm Paytm has launched its AI-powered router to help route payment transactional traffic between multiple payment providers with a routing engine. Alibaba-backed Paytm’s routing engine comes bundled with Paytm Assist and provides a faster checkout experience to users. The intelligent router enables higher payment success rates and allows developer-friendly integration, thereby providing a quicker time to market and a seamless checkout experience. Demo of Intelligent Router According to a company statement, the AI engine will analyse data in real-time and will route the transactions to the best performing aggregator to enable faster checkout for merchants. Also,  the engine’s algorithm will take advantage of Bin based routing and will calculate the success rate for the card bin. The AI-based router and Big Data analytics, can provide 2X growth in transactions that are processed on its platform by the end of FY ‘20. According to Puneet Jain, Vice President – Paytm Payment Gateway, partnered merchants spend huge amount on customer acquisition and retention. Losing a customer due to failed payment can cost a lot. “We are excited to introduce an AI-based routing engine that addresses this problem by optimizing the payment workflows and routing the transaction to best-performing payment aggregator in real time. Further, this will help online merchants reduce development effort to enable various PG providers and achieve faster time to market,” he said.","excerpt":"India’s leading fintech firm Paytm has launched its AI-powered router to help route payment transactional traffic between multiple payment providers with a routing engine. Alibaba-backed Paytm’s routing engine comes bundled with Paytm Assist and provides a faster checkout experience to users. The intelligent router enables higher payment success rates and allows developer-friendly integration, thereby providing […]","categories":["AI News"],"tags":[],"author_name":"Richa Bhatia","publish_date":"2019-06-18T07:55:02","publication_year":"2019","word_count":226,"keywords":["big data","Go","programming_languages:R","AI","ML","programming_languages:Go","analytics","R"],"extracted_tech_keywords":["AI","ML","analytics","R","Go","big data","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/paytm-introduces-ai-powered-router-engine-to-enable-faster-checkouts\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":49626,"title":"Top 10 Machine Learning Jobs In Israel","content":"Israel today is one of the most technologically-forward countries. Over the years the nation has not only strengthened its cybersecurity and other tech arms but also its data science domain. If you are planning to bet big by taking your data science journey outside India, then you should definitely take a look at these jobs. Data Scientist – Machine Learning @ Booking.com (Tel Aviv) Part of Booking Holdings Inc., Booking.com is one of the largest travel e-commerce companies in the world. The company is currently hiring a data scientist for machine learning to join the company’s Tel Aviv machine learning centre. The candidate will be responsible for turning data into customer experience improvements and also support business decisions and improve products. Requirements: The candidate should have a minimum of 4 years of industry experience in data science\/machine learningShould have a Masters, PhD, or equivalent experience in a quantitative field (ComputerScience, Mathematics, Engineering, Artificial Intelligence, etc.)Should have strong knowledge of statistics, algorithm programming, experimentation, data visualization, machine learning, optimization and big dataShould be excellent at scripting languages such as R, Octave, and one programming language such as Python, Perl, C\/C++, JavaShould also have the knowledge and make the right selection of tools to train elementary machine learning models Note: Only for candidates that are already based in Israel Apply here. Machine Learning Engineer @ Moon Active (Tel Aviv) Moon Active is one of the well-known and world’s fastest-growing mobile game startups. Right now, the company is looking to hire a machine learning engineer to join its team in Tel Aviv. The candidate will not only be responsible for building real-time ML models to optimise games and also produce the best game experience. The job role would also include data mining, exploration and extraction. Requirements: The candidate should have at least 1 year of industry experience in the fields of Data Science and Machine Learning and 2 years of software development experienceShould have strong knowledge with Python and Data Science toolkitsExperience in Deep Learning, JavaScript and Node.js is an added advantage Apply Here. Software Developer – Machine Learning @ JPMorgan (Herzliya) The technology centre of JPMorgan in Herzliya is looking to hire a software developer for machine learning. It is a full-time role and is only for Israeli candidates. Requirements: The candidate should have a strong knowledge of application, data and infrastructure architecture disciplinesKnowledge of OO and SOLID design principlesShould possess experience with end-to-end design and implementation of enterprise deployments,Should know distributed data processing platforms, such as Spark, Hadoop,Experience with integrating data with enterprise applications, including relational and NoSQL databases, messaging, and event processing.Should have experience working with Machine Learning Apply Here. AI\/Machine Learning Expert @ BMW Group (Tel Aviv) The BMW Group is hiring AI and Machine Learning experts to join its Tel Aviv team. As AI\/ML expert at BMW, you will be responsible for the development and evaluation of new AI-based technologies at BMW. S\/he would work closely with BMW’s Munich office and make sure that all the generated insights and research are transferred. Requirements: The candidate should have a degree in computer science, statistics or a similar discipline in the field of AI, machine learningShould have experience in creating AI and ML models and use casesStrong knowledge of Python, Java \/ Android, Swift \/ iOS, TensorFlow, Pytorch, Bot frameworks (Alexa, Dialogflow) Apply Here. Machine Learning Applied Researcher @ General Motors (Herzliya) General Motors is hiring machine learning applied researcher to join its team in Herzliya, Israel. As an applied researcher at GM, you will be responsible for primarily designing and implementing innovative ML\/CV algorithms for autonomous vehicles perception system. Requirements: The candidate should be an MSc\/PhD in EE\/CS\/MathematicsShould have more than 3 years of a track record of working with machine learning, deep learning or computer vision algorithmShould have strong knowledge of C\/C++\/ python Apply Here. Machine Learning Engineer @ First Media (Tel Aviv) First Media is one of the Top 10 publishers on Facebook, with 1.6 billion monthly views and 120M+ fans on social media. Currently, the firm is hiring machine learning engineer to work with the company’s engineering team to build models to improve the profitability of the sites and apps. Requirements: The candidate should have a degree in Computer Science, Mathematics, Engineering or a related fieldShould have at least 3 years of experience in machine learningHands-on experience with Python or Java is a mustThe candidate is also expected to have significant experience with cloud technologies like Google Cloud – a distinct advantage. Apply Here. Machine Learning Expert @ Totango Metrics Ltd. (Tel Aviv) Started out in 2010, Totango is the leading customer engagement platform that uses the cloud to process billions of customer signals and make meaning for customer-facing teams. The company is currently hiring machine learning expert to lead machine learning initiatives and provide actionable insights from customer data. Requirements: The candidate should have experience in building ML and AI-based solutionsShould have strong knowledge of CNN, RNN, Deep learningExperience with frameworks such as SparkML, TensorFlow, Caffe2, MxNet, H20, PredictionIO, PyTorch, SageMaker, AutoMLKnowledge of functional programming is a must Apply Here. Deep Learning Engineer-Machine Intelligence @ Alibaba Group (Tel Aviv) Alibaba is one of the leading names in the eCommerce industry. Right now, the company is hiring deep learning engineers to join its machine intelligence Israeli lab. As a DL engineer, you will be responsible primarily for designing and implementing advanced machine vision and machine learning-based products. Requirements: The applicant must have a BSc or MSc degree in CS, EE or a related field.She should also have significant experience in the field of machine learning and deep learning.The candidate should be able to design and optimization of deep learning architectures.Should possess excellent analytical and learning skills, and also have strong coding skills in Python. Apply Here. Data Scientist – Machine Learning @ Apple (Herzliya) Apple is looking to data scientist with a passion for using machine learning to create intelligent applications to join its team in Israel. The candidate will be responsible for designing and implementing new machine learning algorithms and techniques and collaborate with the most innovative product development teams. Requirements: The candidate should have a strong background in Data Science and Machine Learning.The candidate will also have to explain how to build models to answer problems, and how proved models worked in silica and the real world.The applicant should have at least 2 years of experience applying ML techniques to build modelsSoftware development skills are a mustProficiency in PythonShould have experience using ML and statistical-analysis libraries, such as Turi Create, PyTorch, Keras, Scikit-learn, SciPy.Experience with deep learning frameworks, such as MXnet, Caffe, and TensorFlow is a plus. Apply Here. Machine Learning Engineer @ Verint (Tel Aviv) New York-based analytics company, Verint is hiring machine learning engineer. As an ML engineer at Verint, you will be responsible for developing and implementing machine learning model based on data exploration. Requirements: The applicant should have a BSc or BA in Computer Science or another relevant degreeShould have more than 5 years of experience as data science or machine-learning researcherStrong knowledge of Machine Learning or Deep learning algorithms is a mustShould have experience with tools such as Pandas, Scikit-learn, NumPyShould have experience with R, Python, Java coding, Jupyter notebook. Apply Here.","excerpt":"Israel today is one of the most technologically-forward countries. Over the years the nation has not only strengthened its cybersecurity and other tech arms but also its data science domain. If you are planning to bet big by taking your data science journey outside India, then you should definitely take a look at these jobs. […]","categories":["AI Hirings"],"tags":["AI Jobs","apache spark replacement","Career","machine learning jobs","nosql spark support"],"author_name":"Harshajit Sarmah","publish_date":"2019-11-09T12:43:37","publication_year":"2019","word_count":1203,"keywords":["machine learning jobs","data science","artificial intelligence","machine learning","nosql spark support","AI","PyTorch","ML","AI Jobs","computer vision","deep learning","analytics","apache spark replacement","TensorFlow","Career"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","computer vision","data science","analytics","TensorFlow","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/top-10-machine-learning-jobs-in-israel\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10143401,"title":"Automation Is Not for Everyone","content":"In manufacturing, automation is considered an indispensable force, serving as the key to functioning. However, surprisingly enough, this is not the case. Automation may not be a universal solution. “Automation isn’t for everyone. Sometimes, the process a customer is following is already solid, and they don’t need to spend heavily on robotics,” said Hari Parthasarathi, global manufacturing conglomerate 3M’s application engineering leader in India, Southeast Asia, Australia and New Zealand, while interacting with AIM. It’s safe to call Fortune 500 and 3M the crusaders of automation. Parthasarathi spoke about the importance of assessing automation’s relevance on a case-by-case basis, with the goal being to address genuine needs rather than becoming a costly exercise in over-engineering. India’s Manufacturing Landscape 3M, short for Minnesota Mining and Manufacturing Company, has been in the automation space for 122 years. The company entered India in 1987. With over three decades in the country, 3M’s presence in India is only solidifying over time. Last year, the company opened an abrasive robotics lab in Bengaluru, which is the first in India and 17th in the world. India’s industrial automation market is estimated to hit $15.2 billion in 2024 and is expected to touch over $29 billion by 2029. Interestingly, India’s manufacturing ecosystem, which is considered diverse owing to small-scale to large global operations, presents unique challenges for automation, as many enterprises prioritise cost-effective, scalable solutions over advanced, high-cost systems. “If you look at India as where we are – let’s say manufacturing or industrial space – that’s where we operate. We are trying to position ourselves globally as a manufacturing hub,” Parthasarathi said. He goes on to explain that over the past 20 years, India has established itself as a leader in the IT and service industries, while manufacturing was often viewed as secondary. However, initiatives such as ‘Make in India’, shifting geopolitics, and increasing global reliance on India have significantly transformed this perception. These factors have positioned India as an emerging manufacturing hub, with the country aiming to demonstrate its growing capabilities in the sector. “We want to flex our muscles in a few years and show that we are a much bigger manufacturing hub today than what we used to be,” he added. Collaborative Approach for Automation There are two types of automation that are predominantly run in manufacturing facilities. “We are working with both fixed automation guys and robotic automation guys,” said Raghavendra Koneri, Application Specialist, Robotics and Automation, 3M, to AIM. “If there is a big volume with a similar shape, size and everything, then they will go with fixed automation. If there is a complex part, the customer uses a robotic arm, and they will say yes.” 3M partners with robotic arm manufacturers, compliance system providers, and customers to create automation systems tailored to specific industries and processes. For example, in the abrasives segment, 3M has demonstrated that automation can boost productivity tenfold. However, the company ensures proof of concept before scaling such solutions, focusing on long-term benefits rather than immediate returns. The team believes that automation adoption is slower in India, not because of reluctance but due to a lack of skilled manpower and high implementation costs. Furthermore, many businesses are hesitant to invest in automation without a clear understanding of its long-term value. 3M acknowledges that the transition to automation must be gradual and strategic. The company emphasises that automation should complement, not replace, human skills. By focusing on upskilling the workforce and integrating automation thoughtfully, businesses can achieve a balance between efficiency and inclusivity. 3M Finesse-it Robotic at Bengaluru lab Humanoids for Automation? As we already know, the adoption of humanoids in manufacturing facilities is no longer a figment of our imagination. There are humanoids and semi-humanoids already being implemented by big tech companies. BMW manufacturing plants are already using Figure 02 humanoids on their production lines. Similarly, Amazon is leveraging Digit robots at their warehouses to support safety. Considering the possibility of humanoids at 3M, it definitely is not implausible. “Humanoids are already playing a huge role, especially when there’s a huge level of safety that comes into the picture. Here, I think it could happen, but not in the immediate future,” Parthasarathi said. He goes on to explain 3M’s automated paint repair solution for automotive original equipment manufacturers (OEMs), where they aim to replace manual inspection of paint defects, a task prone to errors due to operator fatigue. Parthasarathi believes the automation at 3M might pave the way for a future where humanoids or automated systems can handle such tasks seamlessly.","excerpt":"3M, a century-old company, opened its abrasive robotics lab in Bengaluru in 2023, making it the first in India and the 17th in the world.","categories":["AI Features"],"tags":["3M","abrasives","AI in manufacturing","humanoids","Manufacturing","robotic"],"author_name":"Vandana Nair","publish_date":"2024-12-12T13:04:51","publication_year":"2024","word_count":754,"keywords":["Go","humanoids","AI","ML","abrasives","Scala","Git","RAG","automation","Aim","robotic","ViT","3M","AI in manufacturing","R","Manufacturing"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","Scala","Git","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/automation-is-not-for-everyone\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054845,"title":"A Beginner’s Guide to Time Series Modelling Using PyCaret","content":"When it comes to determining whether a business will succeed or fail, time is the most important factor. Pre-processing, trend normalization, and, most importantly, a cross-check of all available algorithms take time when building a robust forecasting model from scratch. There are a variety of AutoML tools in the market that allow us to perform modelling on raw data with just a few lines of code, saving our time. However, in this article, we will concentrate on PyCaret, an AutoML tool. Time series modelling necessitates special treatment due to the presence of its component, which PyCaret provides. The main points to be discussed in this article are given in the below table of contents. Table of Contents What is Time Series Modelling?All About PyCaret Quick Start Modelling with PyCaret Let’s start the discussion by understanding what time series modelling is. What is Time Series Modelling A time series is a quantity that is measured over time in a progressive manner. In its broadest sense, time series analysis attempts to forecast what will happen in the future by inferring what happened to a set of data points in the past. We will, however, approach time series quantitatively by assuming that they are realizations of random variable sequences. To put it another way, we will assume that our time series are generated by an underlying generating process based on one or more statistical distributions from which these variables are drawn. The goal of time series analysis is to learn about the past while also predicting the future. Time series data is naturally organized chronologically. Time series analysis differs from cross-sectional research in which the observations are not naturally ordered (for example, explaining people’s earnings by reference to their educational degrees, where the individuals’ data might be input in any order). Time series analysis varies from spatial data analysis in that the observations are typically linked to specific physical places (e.g. accounting for house prices by the location as well as the intrinsic characteristics of the houses). In general, a stochastic model for a time series will represent the fact that observations near in time are more tightly related than ones further away. Furthermore, time series models commonly use the inherent one-way ordering of time to express values for a given period as derived from past rather than future values. All About PyCaret PyCaret is a low-code machine learning library and end-to-end model management solution for automating machine learning workflows that are created in Python. It is quite popular due to its ease of use, simplicity, and ability to quickly and efficiently construct and deploy end-to-end machine learning prototypes. PyCaret is a low-code alternative that may be used to replace hundreds of lines of code with just a few lines. As a result, the experiment cycle becomes exponentially faster and more efficient. PyCaret is a straightforward and straightforward application. All of PyCaret’s operations are saved in a Pipeline, which is fully automated and ready for deployment. PyCaret automates everything from missing values through one-hot encoding, categorical data transformation, feature engineering, and hyperparameter tuning. Quick Start With PyCaret In this section, we will leverage the power of PyCaret to model Time Series Data. The dataset used is of climate parameters such as temperature, humidity, wind pressure, and an atmospheric pressure of a city in Delhi. All the instances are recorded from the year 2013 to 2017 and it is taken from this Kaggle repository. To work with PyCaret you need to install it via simply pip command like! pip install pycaret[full] which installs all the core functionality of the package. Now let’s take a look at the dataset. import pandas as pd data = pd.read_csv('\/content\/DailyDelhiClimateTrain (2).csv') data As you know, the time series forecasting or modelling is a kind of regression model for that we can use PyCaret’s regression module to deal with it. The PyCaret Regression Module is a supervised machine learning module that computes the relationships between one or more independent variables and one or more dependent variables (also known as the “outcome variable”, or “target”). Regression is used to anticipate continuous data such as sales amount, quantity, temperature, and customer count. The setup function in PyCaret provides a number of pre-processing options for preparing data for modelling across all modules. The default settings of PyCaret’s Regression module are not ideal for time series data because they involve a few data preparation steps that are not valid for ordered data (data with a sequence such as time series data). Time-series data necessitates a different type of cross-validation because the order of dates must be preserved. When evaluating models, the PyCaret regression module employs k-fold random cross-validation by default. For time-series data, the default cross-validation setting is insufficient. Because algorithms cannot deal with dates directly, we have extracted some simple features from dates, such as month and year, and removed the original date column. Now here we are using only one feature from our data because, at the time of predicting the values of the features, we also need to supply close realistic values for those independent variables which are hard to simulate the same as data we have. But we can also extend it to multiple features prediction as usual by holding a test split from data but make sure you do not shuffle the data as it is a time series problem. data['month'] = [i.month for i in data['date']] data['year'] = [i.year for i in data['date']] data['day_of_week'] = [i.dayofweek for i in data['date']] data['day_of_year'] = [i.dayofyear for i in data['date']] Let’s split data for train and test. train = data[data['year'] < 2016] test = data[data['year'] >= 2016] Now we will explicitly pass the training data, test data, and cross-validation strategy to the setup function via the fold strategy parameter. from pycaret.regression import * # initialize setup Setup_ = setup(data = train, test_data = test, target = ['meantemp'], fold_strategy = 'timeseries', numeric_features = ['year','month','day_of_week','day_of_year'], fold = 3, transform_target = True, session_id = 123) Now let’s train and validate all the available models: best = compare_models(sort = 'MAE') PyCaret does provide a detailed report of all the models, same as other AutoML tools. We have sorted the result based on Mean Absolute Error and based on that Gradient Booster regressor outperforms the rest not only w.r.to MAE but also with all other metrics. Technically speaking, we have trained these models on only features and those retrieved from simple date format. In order to make a future prediction, we need to simulate the same four attributes from the date format. As you have seen, the last observation of our data is on 2017-01-01, now we are going to create some future instances up to 2019-01-01 nearly for 2 years. We are expecting that the model should predict the same trend as it has seen previously. future_dates = pd.date_range(start = '2017-01-02', end = '2019-01-01', freq = 'D') future_df = pd.DataFrame() future_df['month'] = [i.month for i in future_dates] future_df['year'] = [i.year for i in future_dates] future_df['day_of_week'] = [i.dayofweek for i in future_dates] future_df['day_of_year'] = [i.dayofyear for i in future_dates] Now let’s finalize the model and make the predictions. final_best = finalize_model(best) predictions_future = predict_model(final_best, data=future_df) Now we will visualize the result using Plotly express. Plotly is an open-source library used widely to create beautiful and more insightful visualization. To know more about the Plotly visualization, you can refer to this article. concat_df = pd.concat([data,predictions_future], axis=0) concat_df_i = pd.date_range(start='2013-01-01', end = '2019-01-01', freq = 'D') concat_df.set_index(concat_df_i, inplace=True) fig = px.line(concat_df, x=concat_df.index, y=[\"meantemp\", \"Label\"]) fig.show() What do you think by observing the above result? I would say it is the more beautiful plot I have ever seen. Because the models have mostly identified major trends in data especially the peak trend and downtrend. Conclusion Technically we have done the univariate modelling, but it resembles multivariate because we have used four features like day, month, year, and day of the year. As we discussed earlier, for multivariate modelling, we can pass multiple features inside the setup function for the attribute numeric_features. Through this post, we have discussed what time series modelling is, and in contrast to a low code base, we practically see how an AutoML tool like PyCaret can be used to perform outstanding modelling. References PyCaret Founder’s BlockTime Series AnalysisPyCaret Official Documentation    Link for above codes","excerpt":"When it comes to determining whether a business will succeed or fail, time is the most important factor.","categories":["Deep Tech"],"tags":["AutoML for data scientists","Guide","Machine Learning","PyCaret","Python","Time Series","time series modeling"],"author_name":"Vijaysinh Lendave","publish_date":"2021-12-05T16:00:00","publication_year":"2021","word_count":1379,"keywords":["Go","AutoML for data scientists","machine learning","Plotly","AI","ML","Machine Learning","feature engineering","PyCaret","RAG","Python","Time Series","time series modeling","R","Guide","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","Pandas","Plotly","RAG","Python","R","Go","feature engineering"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-beginners-guide-to-time-series-modelling-using-pycaret\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10098157,"title":"The New Bill Doesn’t Protect Citizens’ Privacy","content":"The Union Government introduced the Digital Personal Data Protection Bill before the Lok Sabha on Thursday. The bill has been through many rounds of revisions and this is the 4th iteration of the same. In November last year, the bill was open to public consultation and received criticism for giving considerable powers and exceptions to the central government. After significant changes to the bill, the opposition still raised concerns about it. There was an uproar by Congress Minister Manish Tewari who  protested the bill’s classification as a Money bill. The IT Minister, Ashwini Vaishnaw clarified that it is an ‘ordinary bill’. What has changed? There are multiple changes from the previous draft of the bill. One of the essential differences is the definition of ‘processing’ has been updated to include data that has been wholly or partly automated. This comes after significant developments in generative AI. Another significant change in the scope is related to profiling of citizens of the country. In the 2022 version, any foreign entity processing data of Indian people, such as analysing the behaviour, would be subject to the law. However, the 2023 version has removed this rule. Now, the law doesn’t apply to “profiling” that happens overseas. The government was supposed to notify countries where data can be sent for processing. Now in complete reversal the bill has decided to grant relief to the industry. It will now be notifying countries only where personal data cannot be sent for processing. Exemptions for startups Startups can also breathe easy knowing that the bill introduces exemptions for them to reduce compliance burdens. The Central Government has the power to exempt certain categories of businesses, including startups, from specific compliance requirements  like providing prior notice before obtaining consent, ensuring data accuracy, erasing data after its purpose is served and other obligations related to significant data fiduciaries. Appeal system and penalties In terms of grievance redressal, the new Bill provides for a tiered mechanism. Individuals with grievances must first approach the data fiduciary’s grievance redressal mechanism. If they are not satisfied with the outcome, they can then approach the Data Protection Board. Appeals from the Board will be handled by the Telecom Disputes Settlement and Appellate Tribunal (TDSAT). The penalties for non-compliance have been revised in the 2023 Bill. Interestingly, the maximum penalty, which was earlier capped at 500 crore rupees, has now been done away with, in addition to no criminal consequences for the action. There is no upper limit now as far as the penalty is concerned. Shahana Chatterji, partner at Shardul Amarchand Mangaldas & Co says, “In the earlier iteration it was more of a drafting issue and the intention was not to create a cap on the penalty that could be imposed. Now, a whole schedule of penalties can be imposed depending on what the non-compliance is. And, in fact, the board has to consider various factors when it is imposing this penalty. I think they have done away with some of the confusion that was arising from the drafting in the earlier iteration. She further says that the penalty framework moving away from a criminal prosecution is a fantastic move. “It’s very much aligned with the Jan Vishwas Bill. I think it’s very consistent with the way in which data privacy frameworks globally operate as well.” she concludes. Deemed consent in other words The new Bill retains the concept of deemed consent but applies it to specific legitimate uses. Data can be processed without explicit consent as long as it’s given voluntarily and is for a “legitimate purpose” provided under the Bill. Entities collecting data see reduced compliance and in the final version has done away with the need to seek consent for the transfer of personal data to a third entity for processing. Consent is considered given if the individual has not explicitly indicated refusal, or for certain situations like issuing subsidies, benefits, services, etc., where consent was previously obtained by a state instrumentality for a digital purpose. It also includes situations related to national interest, compliance with judgments, medical emergencies, disaster response, public health threats, and health services during epidemics. However, the extent of consent is limited to the specific purpose for which it was given. The government had received considerable feedback on a clause in the earlier draft, which required entities to seek consent from parents while processing personal data of children. It has now been tweaked a bit. If the government is satisfied that the personal data of children is being handled securely, it may prescribe an age beyond which the entity collecting data may no longer require parental consent. This benefits corporations, where previously dealing with such consent was a logistical nightmare. Sweeping powers to the government The 2022 draft provided the Data Protection Board with protection from prosecution, suits, or legal proceedings as long as actions were done in good faith. The new Bill extends this immunity to the central government as well. Any action taken by the government, intended to be done in good faith, will be protected from prosecution. In a couple of additions: the government has given itself the power to block certain entities; the government has also given itself the power to seek any data from entities for purposes of this act. In another tweak to the earlier draft, the final version of the bill gives the central government immunity from lawsuits. This is in addition to the immunity enjoyed by the data protection board, its chairperson, and its members. Finally, decisions of the data protection board can now be challenged before the Telecom Disputes Settlement and Appellate Tribunal (TDSAT). As per the earlier draft, it could be appealed only before high courts, but now on. India isn’t getting it right With such a large concentration of powers with the government, the opposition is also unhappy with the bill. The Internet Freedom Foundation has written a list of grievances which echo their previous concerns which haven’t been addressed. The government has made more exceptions for itself, which could lead to increased state surveillance. There are also concerns about unclear rules on important matters left for future decisions. Changing the Right to Information Act weakens its strong nature. Moreover, the government has too much control over the Data Protection Board, and there are strict duties and penalties for Data Principals.India isn’t getting it right Amit Jaju, Senior Managing Director, Ankura Consulting Group (India) compares the bill with the European Union’s GDPR, stating there are several similarities, such as the emphasis on consent, rights of the data subject (similar to Data Principal in the Indian bill), and penalties for non-compliance. “However, there are also differences. For instance, GDPR has stricter regulations on data transfer outside the EU and has provisions for the “right to be forgotten”, which allows individuals to request the deletion of their data under certain circumstances. The Indian bill, on the other hand, has a focus on the establishment of a Data Protection Board, which is not a feature of the GDPR.”","excerpt":"The new Digital Personal Data Protection Bill aims to protect individuals’ privacy but raises concerns. Sweeping powers to the government, weak penalties for non-compliance, and unclear rules on crucial matters are points of contention.","categories":["AI Trends"],"tags":["Ashwini Vaishnaw","dpdp india","GDPR"],"author_name":"K L Krithika","publish_date":"2023-08-05T17:40:41","publication_year":"2023","word_count":1169,"keywords":["Go","dpdp india","AWS","AI","cloud_platforms:AWS","R","programming_languages:R","programming_languages:Go","Git","generative AI","GDPR","Ashwini Vaishnaw","startup"],"extracted_tech_keywords":["AI","generative AI","AWS","R","Go","Git","startup","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/the-new-bill-doesnt-protect-citizens-privacy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24620,"title":"Interview With Snehashish Bhattacharjee: Global CEO &#038; Co-Founder At Denave","content":"Founded over 19 years ago, Denave filled the void in the industry by providing end-to-end sales enablement solutions. They had clear focus on revenue acceleration, optimisation of return on investment and measurable impact by leveraging people, processes and technology. Co-founder and global CEO Snehashish Bhattacharjee spearheads the strategic goals of the company and is instrumental in building the company’s future roadmap. With technology and innovation as the convergence point, Bhattacharjee is aiming to emerge as the world’s largest sales enablement organisation. In an interaction with Analytics India Magazine, Bhattacharjee talks about how Denave is disrupting the world of sales using sales process automation, business intelligence, machine learning and analytics to cater companies across the industries globally and also shares his insight on how data analytics is evolving in India: Analytics India Magazine: How can data and analytics help companies drive better sales? How is Denave enabling companies to accomplish that? Snehashish Bhattacharjee: Data analytics has a critical role to play in sales in terms of sourcing, checking purchase history and prioritising, identifying probable prospects, and sales process and fulfilment. Denave, enables the organisations with its tech and analytics practice led services at each of these stages during the entire sales journey. With our solutions, we are changing the industry landscape with the integrity of revenue, proof of execution and ROI orientation that we bring to marketing processes as well as technology solutions. AIM: Tell us more about Denave’s Intelligent Database Management and other analytics services provided by the company SB: Denave’s Intelligent Database platform focuses on two primary areas: Building a tech-enabled access to organisational level information, that enhances contactability and access to profile information which is relevant to prioritise sales drive focus. A suggestion engine (Smart Database Access Platform) which allows identifying the probability of purchase. Along with this, we also have services around these areas: Data Engineering: We consider this a prerequisite to build an effective Data Visualisation engine, leading to impactful sales and marketing analytics. Data Visualisation and Dynamic Narratives: We leverage popular platforms in conjunction with our CRM to help provide this as a service. Analytics Practice: Here, we presently have built a practice around Retail Analytics which we provide to customers across 14 countries, including India. AIM: Customer centricity is important in the analytics industry. How is Denave helping businesses to discover untapped customer database and reach out to more customers? SB: Appropriate customer analytics strategy is required to support business leaders to develop well-informed plans with minimal risks, that drives profitability through optimum customer spends. This is where Denave’s UDB (Unified Database Platform) in conjunction with SDAP, comes into play. We have built an engine around consolidating database from multiple sources, leveraging technology and then consolidating, cleaning and integrating the same leveraging in-house developed rules and algorithms. This is then made to run through our suggestion engine (SDAP) that leverages past purchase history of the product or similar products to identify the probability of purchase, leveraging affinity or similarity mapping. We have built this prototype for India and Malaysia already and are in the process of replicating the same for multiple geographies. AIM: What are the various industries you serve? Highlight few use cases where they have been benefited with Denave’s solutions? SB: The major industry verticals that we cater to, includes Tech, Telecom, FMCG, consumer durables, oil and natural gas etc. Typically, at a point of time, we work with 150-200 customers altogether. Here are some of the use cases: Retail Analytics Practice for one of the world’s largest search engine Denave helped one of the largest search engine to analyse the data quality and increase the efficiency of its merchandising activity. With our Integrated Analytics Platform, leveraging services like data assessment, availability optimisation analysis and visibility optimisation analysis, Denave’s Retail Analytics practices has helped the client to achieve: Up to 99.9 percent improvement in data quality 2X faster TAT (4 weeks to 2 weeks) Standardisation across 14 countries Data Visualisation for a technology giant One of the world’s largest technology company approached Denave with some of the prevailing challenges such as: Scattered\/manual reporting structure Lack of process automation Lack of integrated view for the overall operations Data quality issues Excess man-hours and efforts In a nutshell, our Data visualisation tool has helped client in achieving: 60 percent reduction in TAT 360-degree view of business operations Higher sales closure analysis based on predictive analytics to identify highest probability verticals Incremental revenue opportunity identification Geo expansion and SMB penetration based on propensity analysis Real-time analysis of agent productivity Data-driven higher product affinity analysis Effective lead conversion mapped against high performing agents leading to better productivity and higher results Resource optimisation and higher productivity analysis through performance mapping AIM: What are the technologies that are going to change the sales game in India? SB: Some of the trends that are enabling the sales world to usher into an era of remarkable transformation are: Intelligent Platforms: Sales enablement platforms such as CRM, SFA tools etc. will move towards AI for increasing the analytics and intelligence leverage within these platforms. This significantly increases the productivity and reduces the need for any manual training intervention. Social Media: Leveraging Social media will bridge the gap between marketing and sales by reducing the time taken to target customers and improving campaign effectiveness. Digital Marketing Platforms: Digital Marketing platforms will become more precise by leveraging AI enabled analytics and identifying high propensity prospects so as to exponentially increase sales productivity. AR\/VR: Augmented\/ Virtual Reality will be leveraged more in areas wherein visual impact is created using objects\/people. For instance, training in industries such as Aviation (or other industries manufacturing heavy-duty devices) will start leveraging AR\/VR devices like MS HoloLens etc. to replace physical presence at a single location of both attendee, as well as the instructor. Speech recognition: Digital transformation and AI proliferation will leverage the advancement made in speech recognition so as to improve the deployment of sales tools, by replacing input mechanism through speech. Online Commerce: In the years to follow, will see a fast transformation in the consumer buying behaviour on the internet, and the subsequent change will increase the overall market penetration potential – at half or less than half the cost. AIM: How is Denave helping companies create an insight-driven sales pipeline? SB: With an intelligent database management, sales analytics practices and digital marketing tactics in hand we help companies through our 5-stage sales process through sourcing, probability mapping and propensity gauging by leveraging tech, analytics and sales skills. All this ultimately leads to higher productivity and faster and accurate conversion for our customers. AIM: How is Denave using data analytics internally? Highlight some of the analytics tools used by the company? SB: Tools effectively used at Denave are our in-house built Integrated Analytics Platform (IAP), Data Visualisation tools like Power BI, Tableau and Google Data Studio. At our IAP, technology and analytics meet to enable tighter performance tracking and faster decision making. AIM: How is Denave using Predictive analytics? SB: At Denave, Predictive analytics is being used to help customers identify the potential next set of buyers or purchase cycles based on seasonality and affinity mapping of past purchase. Along with this, Predictive Analytics helps us in identifying the sales pattern, leading to potential market issues that need to be resolved. AIM: How has been the growth story of Denave so far? SB: Over the years, we have built multi-industry expertise partnering with global businesses and adopting a solution-focused approach to deliver best practices in sales. With over $5 billion in revenue contribution to clients in this 19-year span, we can confidently say that we understand the industry pulse and what it takes to be profitable in the marketplace. With time, the concept of sales enablement has evolved and we have evolved with it – our evolution from an Indian, to an Asian and to now a global entity stands testimony to that. We have delivery centers in India, Singapore, UK and Malaysia and we will soon open a center in  US. With a reach spanning 5 continents, 50 countries and 500+ cities globally and our innovation-first approach, we are disrupting the world of Sales in a major way. We aim to build world’s largest Sales Enablement organisation in the years to come. AIM: How do you think analytics in India is evolving? How is big data going to revolutionise sales in India? SB: Analytics in India has a long way to go. The primary challenges revolve around getting your raw material organised before actual analytics can be applied. A large part of the focus is still around data engineering and data visualisation. Even predictive analytics is also still evolving in India. Big Data has a very strong role to play in consolidation and sourcing of potential prospects in India, especially in the light of government’s claim of there being 70 million entities in the country’s business landscape. On the tech leverage around analytics, the most relevant and dynamic development is happening around data visualisation. Therefore, instead of just Big Data, the ability to visualise, read and therefore predict outcomes from that big data, has a key role to play in the evolution of sales Industry in India. All in all, Analytics as a practice is budding well in India with some pure play analytics practice organisations coming up. However, there is still a large gap in the sales analytics arena due to lack of domain depth of the existing players. AIM: What is the roadmap for analytics at Denave? SB:   Analytics roadmap at Denave is spread across four critical stages: Data Engineering: Acquiring the data and transforming it into a structured and usable data set for diagnostic and descriptive Analytics. Interactive Visualisation: Depicting the information in terms of correlated graphs for better and holistic understanding of data and events. Actionable Insights and Dynamic Narratives: A synopsis of the critical information and actionable insights enabling quicker informed decisions. Predictive and prescriptive analytics: This stage is to predict the future on the basis of historical information and some external categorical parameters, additionally prescribing the action points on the basis of predictions for future readiness.","excerpt":"Founded over 19 years ago, Denave filled the void in the industry by providing end-to-end sales enablement solutions. They had clear focus on revenue acceleration, optimisation of return on investment and measurable impact by leveraging people, processes and technology. Co-founder and global CEO Snehashish Bhattacharjee spearheads the strategic goals of the company and is instrumental […]","categories":["AI Features"],"tags":["Business Intelligence","Data Analytics","Interviews and Discussions","Machine Learning"],"author_name":"Smita Sinha","publish_date":"2018-05-15T06:32:08","publication_year":"2018","word_count":1688,"keywords":["big data","Go","machine learning","AI","R","Machine Learning","Git","RAG","Aim","analytics","Data Analytics","Business Intelligence","predictive analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","analytics","Aim","RAG","predictive analytics","R","Go","Git","big data"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/interview-snehashish-bhattacharjee-denave\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":46616,"title":"Meet Wellnesys, The Indian Startup Which Has Revolutionised Yoga Globally  Using AI, ML &#038; IoT","content":"Yoga has been taking the western world by storm for quite some time now. According to a survey that had Yoga practitioners as the respondents, 80.4% felt that yoga contributed to the healing of an injury or pain in their body. In order to make Yoga easily accessible and encourage more people to practice it, an Indian startup called Wellnesys is making use of technologies like artificial intelligence, machine learning and the internet of things to give a unique experience to its users. Analytics India Magazine talked to the CEO of Wellnesys, Muralidhar Somisetty to know about the startup and their work in this field. The Inception Wellnesys started out in July 2017 and has its origins in Muralidhar’s personal experience. He faced many health issues in his young days — he had severe back pain and was almost at the verge of getting a surgery done. However, he wasn’t ready to go for a medication routine and wanted to get his health back without any allopathic medicine. He always had a conviction in the sciences of Yoga and Ayurveda. He believes that what is considered as alternative therapy should be the primary medicine for any lifestyle diseases. And this is when he came across a Yoga teacher who helped him come back on track within a few months of time. Witnessing the improvements, Muralidhar started cultivating a habit of Yoga and became a Yoga teacher himself. As a techie — who has previously worked with Siemens and Cisco Systems — and a part-time yoga teacher, he eventually understood the challenges that come up in his work. For example, Muralidhar was unable to track whether his students were doing the asanas properly or whether they were consistent with the program. All he wanted was data because he believed that data helps in figuring out solutions for every individual. And this is where he thought of integrating technology with Yoga and came up with an AI and IoT enabled mat called the YogiFi where his students could track and give him the updates from a faraway distance. Technology Behind YogiFi is basically a yoga mat that is powered with IoT, AI and ML. The mat is designed and created in such a way that when the user steps onto the mat it tracks all the yoga asanas, calculates the strength, flexibility of the user and the body vitals, provide real-time correctional feedback and even guides one to perfect every posture — all these are done without using any non-intrusive method like using a camera. Further, the AI platform on the cloud selects a lot of data from the mat and wearables and correlates them all and tells the user the aspects in which his Yoga practised have helped him individually. Talking about data, the data such as heart rate, the pressure, posture, body alignment, weight distribution, are collected from the mat using patented technologies. This AI-powered mat by Wellnesys also makes sure that it both listen to the body and calms the human mind with biofeedback and vitals correlation. “People trust data and we wanted to substantiate the success story of Yoga with data,” said Muralidhar. The YogiFi also has a mobile app that its users can use to update their asanas, so the instructor can provide personalised suggestions and keeps a track of if s\/he is doing everything the right way and being regular. Also, another best feature about the YogiFi app is that it has a coach plus a student interface, where it elevate the value of the practices where the teacher designs a Yoga program and uses the map to deliver. Future Ahead The startup, with just about 9 people in their team equipped with varied skills like embedded, UX, IoT, AI and ML, is providing unique services to the country. Also, it is the only company in India to have provided such services. Talking about competition, Wellnesys have seen the competitive landscape and according to Muralidhar, there is no other company that has worked on a deep technology that Wellnesys has done so far. Looking into the future, Wellnesys aims to popularise Yoga to an audience as large as possible using AI, ML and IoT and wants to be recognised as a wellness platform where teachers and students come together.","excerpt":"Yoga has been taking the western world by storm for quite some time now. According to a survey that had Yoga practitioners as the respondents, 80.4% felt that yoga contributed to the healing of an injury or pain in their body. In order to make Yoga easily accessible and encourage more people to practice it, […]","categories":["AI Startups"],"tags":[],"author_name":"Harshajit Sarmah","publish_date":"2019-09-30T16:34:23","publication_year":"2019","word_count":713,"keywords":["Go","machine learning","artificial intelligence","AI","ML","RAG","Aim","analytics","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","Aim","RAG","R","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-startups\/meet-wellnesys-the-indian-startup-which-has-revolutionised-yoga-using-ai-ml-iot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10003786,"title":"Pycharm IDE For Dummies- Beginners Guide","content":"Pycharm is a very popular python IDE which is a cross-platform IDE that was developed by JetBrains to use it for python development. Many people these days believe that python is the best language where a user can build software applications by writing clean and readable code. Python is the favourite language for many, especially people working in Data Science and Machine Learning. Big giants like Facebook, Amazon, and Twitter use Pycharm as their IDE for writing codes in python. Pycharm supports python 2.0 and python 3.0 also once can work with Pycharm on Mac and Windows as well. There are many advantages of using Pycharm that include making it easier for the people to code quickly and efficiently using different software applications provided by Pycharm. This article will demonstrate everything you need to know before writing your first code in Pycharm IDE. The article will take you from installing the software and dependencies to writing your first code in Pycharm followed by a classification problem on iris data set where we will classify which class does the flower belong to. What you will learn from this article? Why should we use Pycharm How to install Pycharm and do the configuration for the first project Advantages of Pycharm How to install python packages in Pycharm Implementation of a classification model in Pycharm 1. Why should we use Pycharm IDE? Pycharm was introduced mainly for python programming that can be used over multiple platforms like windows, mac, and Linux. The benefit of IDE is to permit a user to work with a different database without integration with other tools. Pycharm also allows you to create HTML, CSS, and Javascript files. Also, users can personalize their interface of the IDE according to their wishes. 2. How to install Pycharm IDE and do the configuration for the first project? Let’s start with installing Pycharm. Visit the below link to download the Pycharm software. There are two different versions available for installing it on Windows one is a Professional version that is paid one which comes with advanced features and another is a community that can be downloaded that is free of cost. You can download it and install it. Below is the image that describes the instructions. https:\/\/www.jetbrains.com\/Pycharm\/download\/#section=windows Once you have installed the software, if you run the application you will see something like this as shown in the image below. You will be asked for UI themes. You can choose anyone you want to go with and then click on Next Features plugins. Once you click on next steps it will automatically give you a welcome screen interface that is shown below having few of the options that include:- New Project Open a project Configure If you are already working on an existing project you can open that by using the open options or if you want to create a new project you can make a new project as your selection. As we do not have any of the existing projects we will create a new project. Once you click on the new project you will see an output screen as shown below where we have to define the name of our project and choose the environment. I have given the project name as an example and environment as conda whereas you can give whatever name you like and the same goes for the environment. After doing the same we need to click on create. After you create your project you will see a new interface as shown below. The left side is everything about your project and on the right-hand side, you will be able to code. This code is written in the main.py that will arise automatically. Since we have set up a new project lets add a new python file in our project and write our first python code in Pycharm. Right Click on Example and choose New -> Click on Python file-> Name your file. I have created a file with a name test. Now we will write our code in this file. You can check the below image for your reference that shows you a test.py where we have written a code to print a welcome message. After writing the code you can run it by right-clicking on the file name and click on run. The output of the code would be shown below at the last of the interface. 3. Advantages of Pycharm Now we will explore some of the helpful features and advantages of Pycharm that are listed below:- a. We can replicate a copy of code using the shortcut key Ctrl + D. Before After b. We can also refract a variable. Suppose we have to change the name of a variable and want it to reflect in all our code. We can do that by refracting. Right click on the variable you want to refract->  Click on Refactor -> Rename -> Choose new name of variable. Before After c. We don’t have to write each and every line of code. Pycharm gives predictive suggestions like shown in the below image. 4. How to install Python Packages in Pycharm? Now let’s see how we can import different packages in Pycharm. Follow the below steps to install your desired python package. Select your project Go to Settings-> Under your project name select -> Python Interpreter Once we click on the interpreter we will see the interface shown below in the image that describes all the packages that are already installed. To install our package we need to click on the + icon shown on the right side of the image. After clicking on that you will see a new window where you can search for the package and install it. Check the below image for your reference where we are installing the pandas. Once it is installed we will see a message below that says that the package has installed successfully as shown in the below image. This way we can install multiple packages that we require for our project. 5. Implementation of Classification Model in Pycharm Let us now build a classification model in Pycharm IDE. We need to first save the data file inside the project folder that was created titled ‘example’. We have added a Pima diabetes data set inside that folder and now we will read that data. The output for the same is shown in the image below the code. import pandas as pd df= pd.read_csv(‘pima.csv’) print(df) After importing the data we have checked whether there are any missing values or not. print(df.isnull().sum()) print(df.columns) print(df.columns) After checking the column we have defined the dependent and independent variable X and y respectively followed by splitting the dataset into training and test sets. X = df.drop('class', axis=1) y = df['class'] from sklearn.model_selection import train_test_split X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.33, random_state=42) print(X_train.shape) print(y_train.shape) print(X_test.shape) print(y_test.shape) After checking the shape we have then defined the classifier that would be used for classifying. We will be using a random forest classifier. After that, we have fitted the training data to the classifier and made predictions on testing data. At last, we have evaluated the model using different metrics like accuracy score, confusion matrix, recall, and precision. Use the below code to the same. from sklearn.ensemble import RandomForestClassifier rfcl = RandomForestClassifier() rfcl.fit(X_train,y_train) y_pred = rfcl.predict(X_test) from sklearn.metrics import accuracy_score,confusion_matrix,precision_score,recall_score print(accuracy_score(y_pred,y_test)) print(confusion_matrix(y_pred,y_test)) print(precision_score(y_pred,y_test)) print(recall_score(y_pred,y_test)) Conclusion I would like to conclude the article by hoping that you have now understood the whole working of Pycharm from installing Pycharm to building a classification model. Pycharm is a very good IDE due to its capabilities of predictive suggestions whereas you can also explore other coding platforms like Jupyter Notebook and Google Colab. Google Colab even provides GPUs for fast processing when dealing with image data. Also, see a similar article where I build a CNN model using GPU on Google Colab. “How to build CNN model for finger count classification”. Also, you can read more about Pycharm here in this article.","excerpt":"This article will demonstrate everything you need to know before writing your first code in Pycharm IDE. The article will take you from installing the software and dependencies to writing your first code in Pycharm followed by a classification problem on iris data set where we will classify which class does the flower belong to.","categories":["Deep Tech"],"tags":["confusion matrix recall","Pycharm","Pycharm IDE","what is recall in confusion matrix"],"author_name":"Rohit Dwivedi","publish_date":"2020-08-02T11:00:00","publication_year":"2020","word_count":1334,"keywords":["data science","machine learning","Pycharm IDE","TPU","AI","confusion matrix recall","ML","Colab","Pycharm","Python","what is recall in confusion matrix","Jupyter","R","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Jupyter","Colab","Pandas","TPU","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/pycharm-ide-for-dummies-beginners-guide\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10134321,"title":"Why AI Can&#8217;t Get Software Testing Right","content":"Writing unit tests was already a headache for developers, and AI is making it worse. A recent study has unveiled a critical weakness in LLMs: their inability to create accurate unit tests. While ChatGPT and Copilot demonstrated impressive capabilities in generating correct code for simple algorithms (success rates ranging from 63% to 89%), their performance dropped significantly when tasked with producing unit tests which are used to evaluate production code. ChatGPT’s test correctness fell to a mere 38% for Java and 29% for Python, with Copilot showing only slightly better results at 50% and 39%, respectively. According to a study published by GitLab in 2023, automated test generation is one of the top use cases for AI in software development, with 41% of respondents currently using it. However, this recent study is now questioning the quality of those tests. A fullstack developer named Randy on Daily.dev forum mentioned that he had tried AI for both writing code and writing unit tests, and it failed miserably as it does not understand testing frameworks like Groovy and Spock. If you practice TDD, LLMs shouldn’t be writing your tests (as good as it seems). You think through the use cases, break down the logic, and write the tests. Let the LLM fill in the implementation.I wonder how far someone could get with this approach 🤔— Andrew Nguonly (@andrewnguonly) June 2, 2024 Reason Why AI is Poor at Software Testing AI-generated tests often lack the necessary context and understanding of specific requirements and nuances of a given codebase. Due to this, AI may result in an increase of “tautological testing” – tests that prove the code does what the code does rather than proving it’s doing what it’s supposed to do. “It’s already a danger when you write the implementation first; AI is only going to make it worse,” a user explained in the Reddit discussion. Moreover, relying on AI for test writing can lead to a false sense of security, as generated tests may not cover all critical scenarios, potentially compromising the software quality and reliability. When an AI is asked to write unit tests for code that contains a bug, it typically doesn’t have the ability to identify that bug. Instead, it treats the existing code as the “correct” implementation and writes tests that validate the current behavior – including the bugs, if any. Instead, the developer says that a better use for AI would be to ask it, “What are all the ways that this code can fail?” Instead of having it write tests, have it identify things you might have missed. Another report by researchers from the University of Houston, suggested similar numbers as ChatGPT-3.5. Only 22.3% of generated tests were fully correct, and 62.3% were somewhat correct. Besides, the report noted that LLMs struggle to understand and write OpenMP and MPI unit tests due to the inherent complexity and domain-specific nature of parallel programming. Also, when provided with “too much” context, LLMs tended to hallucinate, generating code with nonexistent types, methods, and other constructs. “Like other LLM-based tools, the generated tests are a “best guess” and developers shouldn’t blindly trust them. In many cases, additional debugging and editing are required,” said Ruiguo Yang, the founder of TestScribe. When developers consider making new test cases, AI still has a hard time doing that. With their creative problem-solving skills, human testers still need to make thorough test plans and define the overall testing scope. But What is the Solution? To solve this problem, researchers from the University of Houston used the LangChain memory method. They passed along smaller pieces of the code as a guide, allowing the system to fill in the rest, similar to how autocomplete works when you’re typing. This proves that one of the most effective ways to tackle this problem is providing more context to the AI models, such as the full code or associated libraries, which significantly improves the compilation success rate. For instance, with ChatGPT, the increase was from 23.1% to 61.3%, and for Davinci, it was almost 80%. In recent times, tools like Cursor are helping developers build code without any hassle and in future, we might see these tools building better unit tests along with production code. But for now, while AI can generate tests quickly, having an experienced engineer will remain crucial to assess the quality and usability of AI-generated code or tests.","excerpt":"It’s already a danger when you write the implementation first; AI is only going to make it worse.","categories":["Deep Tech"],"tags":["AI Developers"],"author_name":"Sagar Sharma","publish_date":"2024-09-03T16:01:39","publication_year":"2024","word_count":731,"keywords":["Go","ChatGPT","AI Developers","OpenMP","AI","Git","Python","LangChain","Rust","R","Java"],"extracted_tech_keywords":["AI","ChatGPT","LangChain","Python","R","Go","Rust","Java","OpenMP","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/why-cant-ai-tools-get-programming-tests-right\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":51000,"title":"Hail To Open Source: The 10 Most Popular Github Repos From 2019","content":"Open source projects not only help programmers develop their skills but also helps the community to create better software while reducing the development costs at the same time. According to an Octoverse report, the popular open-source repositories are topping 10k contributors, and repositories like Flutter, First Contributions, and Home-Assistant are among the top 10 projects used in 2019. In this article, we list down the 10 most popular and open-source Github repositories from 2019. (The repositories have been listed according to their popularity) 1| VS Code Released under Microsoft product license, VS Code or Visual Studio Code is one of the most popular code editors. This development environment has not only topped as the most popular open-source repository in GitHub with 19.1k contributors but also has been chosen as the most popular development environment in the Stack Overflow’s annual Developer Survey 2019. Visual Studio Code is a streamlined code editor with support for development ops such as debugging, task running, and version control. 2| Azure Docs Microsoft Azure Docs is used by 14k contributors and secured the second position in the Octoverse 2019 report. It is open-source documentation of Microsoft Azure where the developers contribute by building and managing the applications with the help of Microsoft Azure cloud services. A developer can deploy infrastructure, secure and manage resources, manage data and much more on this platform. 3| Flutter Flutter is a software development kit (SDK) by Google. Besides securing the third position for the most popular open-source repositories, this toolkit has also been titled as one of the fastest-growing open-source projects by the contributors in GitHub. Flutter is a user interface toolkit for building natively compiled applications from a single codebase. 4| First-Contributions First Contributions is a platform for new open-source contributors where they can learn and contribute for the first time without any hassle. It is basically a practice environment which makes it easy for the developers who are collaborating, contributing and open-sourcing their projects for the first time. 5| TensorFlow One of the popular end-to-end open-source platforms for machine learning, TensorFlow is a comprehensive, flexible ecosystem of tools, libraries, and community resources. It helps machine learning developers to easily build as well as deploy state-of-the-art machine learning models in the cloud, on-premise, in the browser, or on-device. 6| React-Native Written in JavaScript, React Native is a popular framework for building Native applications using the JavaScript library. This year, the popular framework is contributed by 9.1k contributors and secured the sixth position among the top 10. It is used by a number of popular social media platforms such as Instagram, Facebook, Pinterest, among others. 7| Kubernetes With 6.9k contributors, Kubernetes secured the 7th position as the most popular open-source projects. Built by Google, Kubernetes is an open-source system for managing containerized apps across different hosts. It gives basic mechanisms for deployment, maintenance, and scaling of applications. 8| DefinitelyTyped DefinitelyTyped is a repository for high-quality TypeScript type definitions. The number of contributors this year is similar to the number of contributors gained by Kubernetes. In this repository, the developers can maintain as well as share TypeScript type definitions for existing JavaScript libraries. With the help of DefinitelyTyped and its declaration files, one can use the JavaScript libraries which will perform the operations of TypeScript libraries. 9| Ansible Ansible is an automation platform by RedHat which can configure systems, deploy software, and orchestrate more advanced IT tasks like continuous deployments or zero downtime rolling updates. This platform secured the 9th position with 6.8 contributors and has maintained itself to be among the top 10 open-source repositories on GitHub since 2016. 10| Home-Assistant Built with Python, Home-Assistant is a home automation platform. This platform keeps tracks and automate controls in devices such as Google cast, Ecobee, Plex, among others. This is the first time that the repository has secured a position among the top open-source repositories.","excerpt":"Open source projects not only help programmers develop their skills but also helps the community to create better software while reducing the development costs at the same time.  According to an Octoverse report, the popular open-source repositories are topping 10k contributors, and repositories like Flutter, First Contributions, and Home-Assistant are among the top 10 projects […]","categories":["AI Trends"],"tags":["GitHub Repositories","open source project","open source projects on github"],"author_name":"Ambika Choudhury","publish_date":"2019-12-03T10:00:00","publication_year":"2019","word_count":644,"keywords":["open source projects on github","machine learning","AI","R","ML","TypeScript","Python","GitHub Repositories","JavaScript","open source project","TensorFlow","Azure","kubernetes"],"extracted_tech_keywords":["AI","machine learning","ML","TensorFlow","Azure","kubernetes","Python","R","JavaScript","TypeScript"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/hail-to-open-source-the-10-most-popular-github-repos-from-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093106,"title":"Key Highlights from Google I\/O 2023","content":"“AI is having a very busy year,” said Google CEO Sundar Pichai, kickstarting Google I\/O 2023 at the Shoreline Amphitheatre. He said, the ‘AI company is reimagining all core products, including search with Generative AI, alongside announcing groundbreaking developments that will reshape the way users interact with the company’s suite of products and services. https:\/\/twitter.com\/Analyticsindiam\/status\/1656344984176427032 Sending Your Emails Just Got Easier At this year’s Google I\/O conference, Gmail got an interesting new update, where the company introduced a new generative AI feature called ‘Help me Write,’ an extension that expands on the capabilities of Smart Reply and Smart Response. The new features aim to enhance user productivity by providing more comprehensive and refined responses. These updates will be rolled out in conjunction with Workspace updates, elevating the productivity and convenience of Google’s email platform. Unveils Next Generation LLM, PaLM 2 One of the key highlight of the developers’ conference was PaLM 2. Whose predecessor was announced last year. The new model represents a significant leap forward in AI technology and is set to power over 25 products released by Google. It boasts an expanded language support, enabling collaboration and communication across more than 100 languages. Products such as Gecko, Otter, Bison, and Unicorn, all built on this latest model, showcase Google’s commitment to leveraging AI to deliver innovative and user-friendly experiences. PaLM 2, surpasses its predecessor. “As a versatile foundation, it can be customized to cater to specific needs and requirements,” shared Google. Med PaLM 2 Google also highlighted Med-PaLM 2, which has been fine-tuned for medical applications, enabling it to analyze X-ray images and generate comprehensive mammography reports independently. Google will let selective users test this cutting-edge technology during the upcoming summer. “Med-PaLM 2 was the first large language model to perform at an expert level on the US medical licensing exam,” Zoubin Ghahramani, a VP at Google’s DeepMind division said. Google Bard To Developers’ Rescue Sissie Hsiao, VP\/GM, Google Assistant and Bard at Google, took the stage to announce the new capabilities of Bard, which is currently built on the foundation of PaLM 2 model. Bard, which has been trained on more than 20+ programming languages, now allows developers to streamline coding and debugging. Furthermore, Bard introduced a Dark Theme, catering to developers’ preferences. Additionally, Bard’s update extends to Adobe Firefly and Google Lens, expanding its reach and utility across various platforms and tools. In addition to this, Bard has now become available in 180 countries and territories, with support for languages beyond English, including Japanese and Korean as a move to promote global accessibility. Workspace GenAI’ed Aparna Pappu, VP of Workspace, reminded attendees of the Help Me Write feature, which was launched to trusted testers in March this year. Pappu unveiled several GenAI features tailored for Google Docs, Slides, and Sheets. These advancements, part of the new Duet AI for Workspace product, will be made available to trusted users and testers later this year, followed by a wider release. Duet AI aims to amplify productivity by harnessing the capabilities of AI to facilitate seamless collaboration and enhance user creativity. Search Gets Generative AI Boost Not to be left behind, the tech giant’s flagship product, Google Search is also set to undergo a transformative upgrade. Cathy Edwards, leading the Search team, announced the integration of generative AI into the search results page, providing users with a comprehensive overview of their queries. The new AI-powered snapshot delivers key information at a glance, empowering users with quick and insightful summaries of their search results. Perspective API Google also introduced Perspective AI, which has been used by publishers to “mitigate toxicity.” Interestingly, all significant LLMs developed by OpenAI and Anthropic now use Google’s Perspective API to evaluate toxicity generated by their own models. Introduces Generative AI Capabilities for Google Cloud Google also unveiled Duet AI for Google Cloud, an AI-powered collaborator designed to assist cloud users with programming tasks such as contextual code completion, code review, and real-time function generation. It is integrated across Google Cloud platforms, including the IDE, Google Cloud Console, and chat functions. In addition to this, Google has introduced new foundation models and capabilities for AI development, including Codey for code generation, Imagen for text-to-image conversions, and Chirp for advanced speech recognition. These models are accessible via APIs and feature enterprise-grade security and reliability. Additionally, Google has launched a Text Embeddings API for semantic understanding of text or images, and Reinforcement Learning from Human Feedback (RLHF) for model improvement via human feedback. These innovations are supported by Google’s AI-optimized infrastructure, which includes the new A3 Virtual Machines based on NVIDIA’s H100 GPU, offering a broad spectrum of GPU power for AI model training and serving. Google Cloud chief Thomas Kurian said that Google Cloud’s infrastructure makes large scale training workloads up to 80% faster, and up to 50% cheaper compared to any alternatives out there.","excerpt":"All significant LLMs developed by OpenAI and Anthropic now use Google’s Perspective API to evaluate toxicity.","categories":["AI Highlights"],"tags":["Google","Google Map","Google PaLM","Google Pixel","Sundar Pichai"],"author_name":"Tasmia Ansari","publish_date":"2023-05-11T02:06:53","publication_year":"2023","word_count":808,"keywords":["Anthropic","GenAI","OpenAI","AI","Google Bard","ML","Google PaLM","Google Map","Aim","Google Pixel","analytics","Google","generative AI","foundation models","Sundar Pichai"],"extracted_tech_keywords":["AI","ML","analytics","generative AI","GenAI","foundation models","OpenAI","Anthropic","Google Bard","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/key-highlights-from-google-i-o-2023\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10087812,"title":"The Once Popular Java Web Framework Comes Back to Life","content":"RIFE recently made a comeback since its popular years between 2002 and 2010. Built on the foundations of the original RIFE framework, RIFE2 is a full-stack framework to quickly and effortlessly create web applications with modern Java. Since its initial release, the Java environment has evolved and many of the original RIFE APIs can now be replaced with pure Java, eliminating the need for XML, or YAML, and allowing for the use of expressive, type-safe code. Created by Geert Bevin, one of the first Java Champions and speaker at many Java conferences, the new framework retains most of the original features while introducing new ones, for a fraction of the footprint and with even greater developer productivity than before. According to its creators, RIFE2 requires only 10% of the effort to achieve 90% of the desired results. It is also compatible with other Java libraries and frameworks when needed. RIFE2’s full-stack design includes intelligent integrations between layers, emphasising practical decisions that prioritise code maintainability and developer productivity. Some features of this framework include web continuations, bidirectional template engine, bean-centric metadata system, full-stack without dependencies, metadata-driven SQL builders, content management framework, full localisation support, resource abstraction, persisted cron-like scheduler, and continuations-based workflow engine. An HN discussion raised the question on how RIFE2 compares to other web frameworks in Java, such as Ratpack and Spark. One participant suggested that the major difference lies in RIFE2’s treatment of “continuations” as a fundamental aspect of the framework, which “allows one to build flows around pausing and saving the state of execution, and resuming it with all variables captured at a later time, which opens up neat possibilities.” On the other hand, the participant pointed out that something like OpenJDK’s Loom framework appears to experience difficulties in its ongoing implementation when compared to RIFE2’s approach. RIFE2’s source code can be found here.","excerpt":"After more than 10 years since it was originally popular, Java-based web framework RIFE makes a comeback","categories":["AI News"],"tags":[],"author_name":"Ayush Jain","publish_date":"2023-02-21T18:21:48","publication_year":"2023","word_count":307,"keywords":["Go","API","programming_languages:R","AI","data-driven","ML","ViT","SQL","R","Java"],"extracted_tech_keywords":["AI","ML","R","SQL","Go","Java","API","ViT","data-driven","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/the-once-popular-java-web-framework-comes-back-to-life\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":47634,"title":"How This Desi Drone Ensured The Security Of Chinese Premier Xi Jinping","content":"When Chinese Premier visited Mamallapuram last Friday his visit was made secure by special India-made drones. Xi Jinping arrived in Chennai for a 24-hour visit for an informal summit with Prime Minister Narendra Modi. The way from Chennai airport to the temple town of Mamallapuram was reportedly surveyed and studied by aerial drone photography even before Premier Jinping made a visit. The route was of more than 50 km and the technology was developed by a team at Anna University. An official who was working in the security arrangement of Premier Jinping told a noted newspaper, “These fixed-wing drones can map with a special resolution of 3cm, compared to Google Street view map that has a spatial resolution of 60 cm… This means anything that is of 3cm could be captured as an image, which includes things inside a parked car on the route.” Before Jinping arrived, over the last 10 days the drones captured more than 30,000 high-resolution images which were about 14 megapixels each from a height of 100 m. A digital map was created using and stitching all these images and was converted into 3D visuals which could be reached through a mobile app. “This helped police in identifying vulnerable points and decide on personnel deployment”, said K Senthil Kumar, researchers from Madras Institute of Technology, Anna University. The mission was helped by six more copter drones that provided a live view to the Tamil Nadu Special Task Force. Anna University researchers were also kept in the dark about the use of this technology and the commandos were trained accordingly. This is not the first time that drones are being used so productively in real-life situations. IIT Madras students had announced earlier this month that they were developing the ‘Eye in the Sky’ Disaster Management and Humanitarian Aid Services. A team from Centre For Innovation (CFI), IIT Madras, is building drones enabled with AI and Computer Vision to develop an end-to-end solution for identifying accurate and critical information on people trapped in disaster-hit areas and communicate them to the relief task force.","excerpt":"When Chinese Premier visited Mamallapuram last Friday his visit was made secure by special India-made drones. Xi Jinping arrived in Chennai for a 24-hour visit for an informal summit with Prime Minister Narendra Modi. The way from Chennai airport to the temple town of Mamallapuram was reportedly surveyed and studied by aerial drone photography even […]","categories":["AI News"],"tags":["Drone","Narendra Modi"],"author_name":"Prajakta Hebbar","publish_date":"2019-10-14T17:39:33","publication_year":"2019","word_count":345,"keywords":["Go","Drone","API","programming_languages:R","AI","innovation","programming_languages:Go","Git","computer vision","Narendra Modi","R","ai_applications:computer vision"],"extracted_tech_keywords":["AI","computer vision","R","Go","Git","API","innovation","programming_languages:R","programming_languages:Go","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/how-this-desi-drone-ensured-the-security-of-chinese-premier-xi-jinping\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10095231,"title":"Crazy Fitness Routines of Tech Billionaires","content":"Gone are the days when tech magnates only wanted to get rich and famous. Now, along with toiling in the tech field, the leaders are also working towards keeping themselves fit. While nutrition is one aspect of it, tech leaders swear by their workout routines to stay healthy. We’ve compiled a list and here you’ll find leaders who stick to normal workouts for muscle gain or fat burn and those who opt for death-defying action sports as a way to achieve fitness. Crazy or not, here are some of the workout routines that tech leaders follow. Mark Zuckerberg Meta CEO Mark Zuckerberg is all out experimenting with extreme sports that pushes one’s physical limits. He recently completed the Murph challenge, which consists of a mile run, a 100 pull-ups, 200 push-ups, 300 squats, ending with another mile run, all while wearing a 20-pound vest. He took under 40 minutes to complete it. Zuckerberg has been aggressively practising the Brazilian jiu jitsu, a martial arts combat sport which he started during the pandemic. He recently competed in his first Brazilian jiu jitsu event where he beat his opponent and won two medals. Sam Altman Making news with his world tours and AI regulations, lean-mean Sam Altman is everywhere. To stay in shape though, Altman has experimented with different kinds of exercise programs in order to arrive at something that not only works well physically but also makes him feel good. Altman once shared in a blog that he lifts weights thrice a week for an hour and occasionally practises HIIT (high intensity interval training). Jeff Bezos The physical transformation that the Amazon founder underwent is like no other. From being a lean and pale tech magnate, Bezos shook the world when he appeared all bulked up in 2017. Along with a muscle-gaining diet, Bezos does extensive weight training and also indulges in outdoor activities that combine cardio and resistance exercises. Activities such as kayaking, rowing, paddleboarding, running uphill and downhill are some of his favourite workouts. Jack Dorsey Source: NY Post Entrepreneur and former CEO of Twitter, Jack Dorsey follows a holistic workout routine that combines mental and physical exercises. He has been practising meditation for over 20 years and aims to get two hours of meditation daily. On the physical front, he prefers long walks and some jogging. On days he doesn’t walk, he takes up high-intensity interval workouts on an exercise bike or does the seven-minute workout. Elon Musk Source: Forbes The 51-year old Tesla King, Elon Musk, is known for his on-again, off-again fitness routine. Though not a fan of it, Musk lifts weights and uses the treadmill to achieve a combination of weights and cardio. He prefers to watch something ‘compelling’ to help him work on the treadmill. Musk has also tried his hand at various forms of martial arts such as karate, Brazilian jiu-jitsu, Taekwondo and Judo.","excerpt":"Some outrageous, and some simple, here are the workout routines of your favourite tech magnates","categories":["AI Trends"],"tags":["Elon Musk","jack dorsey","jeff bezos","Mark Zuckerberg","Sam Altman","weights"],"author_name":"Vandana Nair","publish_date":"2023-06-18T10:00:00","publication_year":"2023","word_count":481,"keywords":["Go","Sam Altman","programming_languages:R","AI","weights","programming_languages:Go","Elon Musk","Mark Zuckerberg","Aim","ViT","jeff bezos","jack dorsey","R"],"extracted_tech_keywords":["AI","Aim","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/crazy-fitness-routines-of-tech-billionaires\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":5962,"title":"INFORMS Develops First-Ever Professional Analytics Certification; Now Available World-Wide","content":"The Institute for Operations Research and the Management Sciences (INFORMS), the largest society in the world for professionals in the field of operations research (O.R.), management science, and analytics has established the Certified Analytics Professional (CAP®) designation, a first-of-its kind analytics professional certification.  The mission of CAP® is to validate solid analytics practices, provide an instant level of credibility with clients, help organizations to identify and develop qualified analytics talent, and to help analytics professionals boost their career potential and elevate themselves within the marketplace. INFORMS defines analytics as the scientific process of transforming data into insight, thereby yielding better decision-making.  CAP® measures analytics performance across seven major areas of practice encompassing the analytics end-to-end process: business problem framing, analytics problem framing, data, methodology selection, model building, deployment, and model life cycle management.  The exam itself is based on a rigorous job task analysis that enumerates the tasks that analytics professionals perform and the knowledge they need to successfully perform those tasks.  The CAP® job task analysis can be reviewed here: https:\/\/www.informs.org\/Certification-Continuing-Ed\/Analytics-Certification\/Analytics-Job-Task-Analysis The INFORMS Board of Directors made the bold move in 2010 to development products and services for the analytics community that they would value. Their first endeavor into analytics was the publishing of Analytics Magazine in 2008 as bi-monthly glossy magazine that covers the field. Next came INFORMS Conference on Business Analytics and Operations Research, which was re-programmed from its original concept of INFORMS O.R. practice meeting, to now appeal broadly to the analytics community. This was followed by other products and services the analytics community is embracing such as certification (CAP), our conference on big data, continuing education courses, and the brand newAnalytics Maturity Model. CAP® was developed for early- to mid-career analytics professionals. It was developed by subject matter experts comprised of both INFORMS members and non-members and validated by the wider analytics community. It was developed for several reasons: to bring visibility and increased credibility to the overall profession in the same way that the PMP® has done for the project management profession, to allow individuals to set themselves apart from the competition as qualified, serious analytics professionals, to allow employers another tool to help find good, qualified analytics talent, and to place the analytics talent they do have on a prescribed path of continuous professional development. To attain the CAP® designation, which can be put after your name like any rigorous credential, you not only have to pass a 100-question exam, but you also must possess the necessary educational achievement and analytics work experience, have your analytical “soft skills” validated by a current or former employer or client, and agree to abide by a first of its kind code of ethics. To maintain your CAP, you must earn professional development units. So you can see it is a complete program of professional development, not just an exam. Eligibility Requirements for CAP® o    BA\/BS or MA\/MS or higher degree o    At least five years of analytics work-related experience for BA\/BS holder in a related area, or o    At least three years of analytics work-related experience for MA\/MS (or higher) holder in a related area, or o    At least seven years of analytics work-related experience for BA\/BS (or higher) holder in an unrelated area o    Verification of soft skills \/ provision of business value by employer o    Agreement to follow Code of Ethics for analytics The exam itself covers the analytics end-to-end process, everything from business problems framing to working with data and models, to deployment and model life-cycle management. Reviews of the exam have been excellent. The terms that are most often used are rigorous but fair. It truly means something when you attain the CAP® . Sample questions and their answers can be found in our Candidate Handbook and Study Guide. For more information on CAP® go to www.informs.org\/certification. The CAP® exam can now be taken at over 700 Kryterion testing centers worldwide on almost any day of the week. There are about 50 testing centers in India.  You can locate the testing center nearest you here: http:\/\/www.kryteriononline.com\/host_locations. It is easy to prepare for the exam with INFORMS free, downloadable Study Guide developed by CAP® professionals. Candidates may apply here: https:\/\/www.informs.org\/Certification-Continuing-Ed\/Analytics-Certification. CAP® can be earned individually and often is, but if you really want to get the most out of CAP®, corporate entities are encouraged to provide CAP® to their entire analytics staffs.   INFORMS provides team pricing for 5 or more individuals at one location sitting for the CAP® exam at one time.  You may also request a CAP® Refresher Session.  It is a half-day workshop that culminates in a paper-and-pencil CAP® exam delivered at your location to a minimum of 5 candidates. Current CAP® credential holders will review the Study Guide and 24 sample questions and answers with you, relate their experiences taking the exam, and provide an interactive Q&A experience. The day ends with the administration of the three-hour CAP® exam to all eligible candidates. Those who do not pass the exam will be provided with one free re-take at their convenience. Analytics thought leader Tom Davenport said it well when he recently said “we think the credential has a bright future. Consultants have begun to get certified so they can represent to their clients that they are qualified to offer analytical services. Particular companies are beginning to ask their analysts to take it. Several of the new analytics Master’s degree programs in universities are planning for their experienced students to get certified. And a few leading companies are beginning to look for evidence of certification on the resumes of job applicants” For more information on CAP®, send email to certification@informs.org.","excerpt":"The Institute for Operations Research and the Management Sciences (INFORMS), the largest society in the world for professionals in the field of operations research (O.R.), management science, and analytics has established the Certified Analytics Professional (CAP®) designation, a first-of-its kind analytics professional certification.  The mission of CAP® is to validate solid analytics practices, provide an instant level of credibility with […]","categories":["AI Trends"],"tags":["analytics certification"],"author_name":"AIM Media House","publish_date":"2014-07-28T15:05:35","publication_year":"2014","word_count":939,"keywords":["big data","Go","programming_languages:R","AI","analytics certification","programming_languages:Go","RAG","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","RAG","R","Go","big data","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/informs-develops-first-ever-professional-analytics-certification-now-available-world-wide-2\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10116953,"title":"India Shows the World How AI Can Shatter Social Barriers","content":"Not only can AI solve many business problems, but the technology also holds promise for resolving various societal issues. On his recent visit to India, Microsoft co-founder Bill Gates addressed students of IIT Delhi, encouraging them to use technology like AI for social good. With its diverse array of social issues, India does offer a rich canvas for innovative solutions. India’s strength lies in its diverse talent pool, encompassing skilled engineers, developers, and technology professionals. Additionally, the country nurtures a vibrant startup ecosystem, fostering a culture of entrepreneurship, further fueling innovation in the technology sector. “What is encouraging for India is that young entrepreneurs are picking these problems statements and coming up with the skills, a ‘can do’ attitude, and commitment and passion to achieve change,” Sudha Srinivasan, who heads The Nudge Centre for Social Innovation, told AIM. Over the years, we have seen startups and nonprofits leverage AI to solve problems in agriculture, healthcare and education. India is home to nonprofits like Wadhwani AI, which could possibly be the only nonprofit in the world devoted exclusively to AI for social impact. AI for social good “Over the last six to seven years, we have incubated about 130 startups. About 14 of those would be deep-tech startups. Our first AI startup was actually in our very first cohort. They used spectrometry to detect contaminants in water and could make it possible to detect water quality at a tiny price point,” Srinivasan said. In India, water contamination still remains a daunting challenge. According to a Lancet study, contaminated water led to half a million deaths in 2019. Another HSR Layout, Bengaluru-based startup Smarterra is using AI to help cities reduce water losses. In India, about 40-50% of water is lost even before it reaches the end customer, according to Smarterra chief scientist Navaneethan Santhanam. “We help utility companies assess the condition of their network. In any large Indian city, there’s probably 5,000-6,000 km of buried pipes of different materials, sizes, and age. We help them identify which pipes are at the risk of failing,” Santhanam told AIM. SmartTerra uses generative AI along with modern geospatial analysis, forecasting, and hydraulic modelling to help utilities pinpoint network failures such as leaks, failing pipes, and faulty metres. The startup operates in many Indian cities and works closely with utility companies such as L&T, Suez, and many state-owned water bodies in India as well as in foreign markets such as Singapore and Philippines. Moreover, this is a great opportunity to build up India’s human capital through skilling and education and AI could be an enabler for that. In many areas in India, the student-teacher ratio is still relatively low and there is also the problem of access to education. “With AI you can improve these ratios remarkably. Instead of a teacher teaching 20 students you can have a teacher using AI to teach maybe 100 students without dropping quality,” Srinivasan said. “The time that goes in assessments too could come down with automation in both spoken and written forms.” AI is breaking social barriers AI, or technology in general, has the potential to dismantle numerous social barriers, empowering marginalised communities to access opportunities, resources, and services previously out of reach. “Think of all the jobs that humans should not be doing, like manual scavenging. If you consider the situation where humans are relegated to jobs with poor dignity, it maintains the status quo, as communities remain bound to such professions due to the lack of alternatives. “This necessity perpetuates social barriers that limit their mobility beyond these jobs. However, with AI or robotics handling these tasks, individuals in these professions could find a quicker path out of low-paying, low-dignity jobs and transition into better-paying roles,” Srinivasan said. Today, many startups are building robots or AI-powered solutions to solve the manual scavenging problem in the country. In 2022, interestingly, the Delhi Jal Board (DJB) developed an AI-powered technology to clean sewage water. Solinas, also incubated by The Nudge, has developed an AI-integrated affordable robotics solution to inspect, clean, and manage confined space for sanitation purposes. It helped clean up manhole blockages and reduced sewer overflows in Madurai. (Homosep Atom by Solinas) Moreover, AI is also breaking down accessibility barriers for the disabled in India. I-Stem, a startup co-founded by two visually impaired duo, is leveraging AI to make STEM content more accessible for the visually impaired. “Over 96% of the content available today is incompatible with various assistive technologies that people with disabilities use. So, chances that a PDF you’re going to find online is going to be accessible is only 4%,” Kartik Sawhney, co-founder at I-STEM told AIM. ( I-Stem awareness session held in Bhopal) In India, over 70% of the disabled community in India remains unemployed, which is more than the entire population of Sri Lanka, Sawhney said. He believes this is contributing to an annual loss of nearly USD 12 billion to the Indian economy. “We believe that by assisting these individuals in securing meaningful roles within high-growth industries, they can serve as influential role models. This not only paves the way for their success but also contributes to reshaping the narrative and discourse surrounding disability,” Sawhney said. Can AI end poverty? Technology is not just an enabler of scale, but is fundamentally a means to unlock value in a digital economy. “Eventually, inclusion in the digital economy is key to bringing India’s bottom 20% out of poverty, otherwise it will be on the other side of a digital divide which will further widen the inequalities in our society,” Srinivasan said. The Nudge runs various programmes like ‘End Ultra Poverty’ and ‘Asha Kiran’ is working towards providing marginalised communities in rural India with sustainable livelihood and elevating them from poverty. However, over some time, The Nudge has started working closely with technology startups and leverage their technological know-how and expertise to advance its mission. For instance, Pune-based startup Tapasya, which is part of the Nudge incubation programme, believes ensuring access to government benefits is crucial for poverty alleviation. However, many underprivileged families lack awareness and know-how to access these schemes. To address this, the startup leverages technology like data analytics to identify eligible families and assist them in accessing the benefits efficiently, enabling last-mile reach. Additionally, Indian entrepreneurs are also creating employment opportunities in rural India with AI. Karya, a non-profit from Bengaluru is using crowdsourcing to bring dignified, digital work to the economically disadvantaged Indians, giving them a pathway out of poverty. The boom in AI has also created the demand for huge amounts of data, especially in Indian languages, which are not found on the web. Karya enables tasks like capturing, labelling and annotating data for corporate clients and in doing so provides a minimum wage, which is 20 times higher than the minimum wage in India. Today, Karya has a presence in 22 states (100+ districts) with over 30,000 workers, who have completed 30 million paid digital tasks.","excerpt":"Indian entrepreneurs are creating employment opportunities in rural India with AI.","categories":["AI Features"],"tags":["AI for good"],"author_name":"Pritam Bordoloi","publish_date":"2024-03-22T12:43:50","publication_year":"2024","word_count":1158,"keywords":["Go","API","AI","AI for good","Git","RAG","Ray","Aim","analytics","generative AI","R"],"extracted_tech_keywords":["AI","analytics","generative AI","Aim","Ray","RAG","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/india-showing-world-how-ai-shatter-social-barriers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":46301,"title":"Secret Behind Youtube’s Great Machine Learning Enabled Video Recommendations Finally Revealed","content":"Recommender systems are one of the most treasured tools on the modern internet. Good recommender systems increase retention on video and content platforms and provide an excellent experience for the users. Just ask any ardent Netflix or Hotstar consumer and one will understand how so much of their viewing choices are influenced by algorithms and models running behind the platforms. Researchers at YouTube recently open-sourced the algorithm running behind the hood which throws much-needed light on how the biggest video platform in the world operates. YouTube has over 1.9 billion monthly users who consume content in over 80 languages. Zhe Zhao and other scientists at Google published a paper called titled Recommending What Video to Watch Next: A Multitask Ranking System to unveil the workings of the large recommender system. The paper outlines and explains how scientists have converted the recommendation problem into a large scale multi-objective ranking system. The algorithm manages many ranking objectives including user feedback. Researchers tried many techniques such as Multi-gate Mixture-of-Experts to manage multiple objectives. They have also underlined how the improvements they suggest have made a huge difference to the world’s biggest video platform. YouTube’s New Recommendation Engine The core target of the scientists is to come up with an algorithm that will improve recommendations for billions of users on YouTube. The researchers also had to consider two important issues: Multimodal feature space: They had to rank a candidate video based on multiple features such as video content, thumbnail, audio, title and description and others. The two big hurdles that have to be solved according to the researchers are bridging the semantic gap for low-level video features and making the machine learning model, learn a sparse distribution of items to do collaborative filtering. Scalability: Since the platform is used by more than a billion users, scalability becomes a huge issue. Some of the features needed for the model to perform are only available online and can not be fetched beforehand. Researchers say in the paper, “At the candidate generation stage, we retrieve a few hundred candidates from a huge corpus. Our ranking system provides a score for each candidate and generates the final ranked list.” Candidate generation used by the system runs on multiple algorithms which try to calculate the similarity between query and candidate video. One particular algorithm retrieves candidates by the frequency of the video being watched for a given query. On the other hand, the ranking algorithm runs on two important inputs from users: Engagement behaviours, collected from clicks and watches Satisfaction behaviours, collected from likes and dismissals The researchers choose to go with the learning-to-rank framework to solve this particular problem. Researchers explain, “We model our ranking problem as a combination of classification problems and regression problems with multiple objectives.” Ranking with multiple objectives is really a hard task. The researchers decided to mitigate the conflict between multiple objectives using Multi-gate Mixture-of-Experts (MMoE), which is a technique recently invented. This particular technique called MMoE is a soft-parameter sharing model which is specifically designed to model task conflicts. The researchers say, “The MMoE layer is designed to capture the task differences without requiring significantly more model parameters compared to the shared-bottom model. The key idea is to substitute the shared ReLu layer with the MoE layer and add a separate gating network for each task.” The neural network suggested by YouTube researchers is identical to multilayer perceptrons with a ReLU activation with task k, prediction yk, and the last hidden layer hk, the MMoE layer as components. Researchers are very careful to eliminate various biases such as position bias, selection biases and others. The new MMoE based neural network with 8 Experts resulted in +0.45% improvement in engagement metrics and + 3.07% improvement in satisfaction metrics. Conclusion The research paper from Google shows that how putting more thought and effort into the ranking problem can reap great rewards for any media recommendation engine. Researchers have been successful in creating a scalable and improved end-to-end ranking system with systems built in to handle various kinds of data biases.","excerpt":"Recommender systems are one of the most treasured tools on the modern internet. Good recommender systems increase retention on video and content platforms and provide an excellent experience for the users. Just ask any ardent Netflix or Hotstar consumer and one will understand how so much of their viewing choices are influenced by algorithms and […]","categories":["AI Features"],"tags":["YouTube"],"author_name":"Abhijeet Katte","publish_date":"2019-09-24T15:04:01","publication_year":"2019","word_count":676,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","Modal","programming_languages:Go","Scala","programming_languages:Scala","YouTube","R"],"extracted_tech_keywords":["AI","machine learning","neural network","R","Go","Scala","Modal","programming_languages:R","programming_languages:Scala","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/secret-behind-youtubes-great-machine-learning-enabled-video-recommendations-finally-revealed\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10068491,"title":"Kissflow launches a low-code\/no-code platform to take on Microsoft and Appian","content":"Chennai-based Kissflow, founded in 2004, offers business process management products. Recently, the company launched a unified low-code\/no-code work platform that fast-tracks enterprise digital transformation. The platform brings together the entire spectrum of work management into a unified experience for enterprise-wide users. Kissflow lets organisations simplify work by enabling the team to build enterprise-grade apps, automate workflows 10x faster, and increase company productivity and innovation. Kissflow streamlines processes, democratises IT, provides 360-degree visibility, fosters creativity, and helps companies build a collaborative culture. “True digital transformation can only be achieved when business and IT come together. Among other things (like the mindset of these two groups), one of the biggest roadblocks is not having a platform that seamlessly allows these two diverse groups to co-create digitisation. Kissflow is that technology that combines the power of two – business and IT,” said Suresh Sambandam, CEO of Kissflow. The opportunity Enterprises today use multiple work management tools that focus exclusively on app development, process management, project or task management, and collaboration. However, rather than simplifying work, they make work complex and disjointed. Enterprises are frustrated with the disconnect between different solutions and teams. According to IDC, over 500 million digital apps and services will be developed and deployed using cloud-native approaches by 2023. To accelerate digital transformation exponentially, the business users, who are the domain experts and are closer to the problem, need to be included. But to unlock the full potential of these business users, enterprises can’t rely on complex, old-school technologies. “An inclusive and unified experience for enterprise users is the only way to reach their digital transformation goals successfully,” Suresh adds. With the new platform, Kissflow aims to unify Work Management (Process Management, Case Management, Project Management, Task Management, and Custom App builder). Unlike other solutions, Kissflow does not charge per app developed – giving users the ability to create unlimited apps without restrictions. The platform can also solve any long-tail use-cases for individual departments. “Using Kissflow’s platform 500 million business users can embark on digital transformation. With this, Kissflow aims to secure a large pie in the USD 50 billion low code \/ no-code market entrenched by big gorillas like Microsoft and Appian,” said Suresh. Kissflow’s work management platform has practical applications across many key industries with use-cases including: Logistics & supply chain: distributor management, project tracking, operational dashboards, quality controlHealthcare: appointment management, calendar and reminders for patient appointments, online training apps for consultationManufacturing: maintenance management, inventory management, order management, vendor managementBanking & Insurance: factoring, loan loss prevention, risk management, policy administration The long march “Even though we started with simplified rule-based application development, we are in a constant pursuit to empower the people closest to the problem with tools and technologies as they are best placed to solve those problems in the most elegant, cost-effective and time-efficient way. We have been one of the top 3 players in no-code BPM & workflow automation categories. We have been leading those categories for decades,” said Suresh. “At Kissflow, we believe that the most powerful solutions are the simple ones. For example, if you punch a few numbers in your phone, it magically calls someone on the other end of the world in a second. The technology is complex (not complicated), but it is simple for users. In an increasingly complex world, such simplicity, which we call the “Power of Simple”, is the core value of our products, that helps us stand out.” Over the years, the company has acquired over 10,000 customers and a million users across 160 countries. Hundreds of global and Fortune 500 brands such as Airbus, Pepsi, McDermott, Comcast and Danone rely on Kissflow to simplify their work. “With the new platform, we aim to secure at least 100 million users on our platform across the entire enterprise user spectrum- from end-users, teams, process owners, citizen developers, and IT developers,” said Suresh. With a strong focus on the global enterprise market, Kissflow is now setting up multiple sales offices in the US and sales operations in LATAM and South East Asia. They already have a team in the Middle East and will look at strengthening their reach and partnerships in the region.","excerpt":"With the new platform, we aim to secure at least 100 million users on our platform across the entire enterprise user spectrum- from end-users, teams, process owners, citizen developers, and IT developers.","categories":["Global Tech"],"tags":["Microsoft"],"author_name":"Sri Krishna","publish_date":"2022-06-07T12:00:00","publication_year":"2022","word_count":694,"keywords":["Go","AI","ML","digital transformation","Git","Aim","ViT","Rust","GAN","R","Microsoft"],"extracted_tech_keywords":["AI","ML","Aim","R","Go","Rust","Git","GAN","ViT","digital transformation"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/kissflow-launches-a-low-code-no-code-platform-to-take-on-microsoft-and-appian\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":16734,"title":"IBM Opens Its First Machine Learning Hub In Bengaluru","content":"Software giant IBM has upped its game and has opened up its first machine learning hub in India. The hub, located in Bengaluru, is a physical space for organisations to visit for hands-on training on ML. The company in a press release stated that while at the hub, organisations can learn from, and collaborate with IBM experts to build and deploy analytic models for a new generation of intelligent applications that learn as they are used. “With India’s focus on Digitization, it’s an apt time for organizations to make this transition. IBM Machine Learning hub reiterates our mission to partner and prepare enterprises for the cognitive era by unleashing the potential of machine learning through models that are constantly improving by analysing new data to generate real-time results which benefit businesses,” said Gaurav Sharma, Vice President – IBM India Software Labs and Growth – IBM India and South Asia. The hub will also provide a platform for like-minded enterprises to collaborate and transform data science processes. IBM’s other ML Hubs are currently operational in Toronto, San Jose, California, at IBM’s Silicon Valley Lab, Beijing, and Boblingen, Germany. Machine Learning is gaining popularity to deal with increasingly complex data and analysis problems. The press release said that by 2018, 75% of enterprise and ISV development globally will include cognitive\/Al or machine learning functionality in at least one application, including all business analytics tools. “Machine Learning, a term coined by an IBMer decades ago has evolved significantly. Businesses are increasingly using machine learning to support advanced analytics across a growing range of industries and initiatives,” Sharma added. The news comes soon after IBM announced that more than 40 global clients in industries have joined the IBM Watson for Cyber Security beta program. The program uses the Watson AI to provide security services to these companies.","excerpt":"Software giant IBM has upped its game and has opened up its first machine learning hub in India. The hub, located in Bengaluru, is a physical space for organisations to visit for hands-on training on ML. The company in a press release stated that while at the hub, organisations can learn from, and collaborate with […]","categories":["AI News"],"tags":["IBM","Machine Learning"],"author_name":"Priya Singh","publish_date":"2017-08-04T05:35:37","publication_year":"2017","word_count":303,"keywords":["data science","Go","machine learning","programming_languages:R","AI","ML","Machine Learning","Git","IBM","analytics","GAN","R"],"extracted_tech_keywords":["AI","machine learning","ML","data science","analytics","R","Go","Git","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ibm-opens-first-machine-learning-hub-bengaluru\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":8564,"title":"The Time to Create a Culture for Analytics Is Now","content":"The old organizational model for decision making is broken. The jig is up. Making decisions relying on gut feeling, intuition, office politics, past experience and bias is giving way to using fact-based information and analytics. These allow for investigation, insights, foresight, and improvements. Predictable skepticism of business analytics Some readers may already be reacting to my observations and saying to themself, “I’ve heard this exaggerated story before.” Skepticism is a healthy virtue. Skepticism involves waiting for enough evidence before accepting or believing. My belief in the value of analytics began at an early age. Math came easily to me in grade school. I enjoyed solving story problems. I created my own dice baseball game and kept team and player statistics. I liked to measure things, including time and distance. But my advocacy of analytics had a personal and punctuated advance early in my adult career. Following a series of job promotions in a manufacturing plant of a blue chip conglomerate, I became operations manager. The daily job of two employees was to make truck deliveries to customer facilities in the city and suburbs. I was young and suspicious that these two workers were slackers getting a full day’s pay for a half day’s work. A heavy fist versus an open hand I did not want to intimidate the two delivery drivers. I wanted to see if I could socially engineer their self-improvement. I began this by first having both drivers draw a map of the few dozen customer locations they routinely transported our products to. Then I asked them at the end of each workday to simply draw arrows connecting the sequence of their delivery locations. The maps initially looked like spaghetti. Next I asked them to write out the approximate length of time associated with each arrow – how long it took between truck stops. I gave them a simple math rule that the total time for the deliveries plus their lunchtimes and breaks must exactly total their daily shift times – typically eight hours. I never asked them to work harder. I only asked them to measure the times for what they were doing. No heavy fist. Just asking for some facts. Fast forward to results and improvement Over the next few weeks, the two drivers grew like plants from seeds in good soil. The spaghetti-like routes became more straight and orderly. The drivers began calculating ratios of the time length per each delivery stop, plotting each time ratio on a crude second-page “distribution curve chart.” They then added explanations for the unusually high or low ratios – the outliers of the distribution curve’s tails. The drivers began thinking in terms of root-cause analysis for how to shorten the time durations (without speeding). It was a way of applying a fundamental problem-solving technique used in the quality management field. As volume increased the two drivers were able continue to handle the increased load during the same workday. They had improved their productivity. Things got even better. The drivers took an interest in the operations of the plant that made the products loaded on to their trucks. Some of the root-cause problems increasing their delivery times were due to problems caused by the plant. As they learned more about the factory’s operations the drivers decided to become certified in production and inventory management with the professional society APICS (www.apics.org ). They switched to jobs as production schedulers in the factory with higher pay. They used their newly gained analytical skills to schedule parts and machine operations and to improve product manufacturing throughput rates and inventory management. A culture for analytics and accountability Yes, it may be a big stretch for me to conclude that current organizational models that are not fully embracing business analytics are broken. However, I learned a lesson from watching two high school graduates go from driving chaotic daily truck routes to very competently scheduling hourly production at a 500-employee factory. The lesson is to help people measure what matters. Provide a context for what constitutes improvement, whether it is time, quality, service level, customer satisfaction or cost. Then, similar to watching plants grow from the soil, watch nature grow individual skills. Executives can formulate textbook strategies like offering highly differentiated products and services or being the lowest cost supplier. But today these are vulnerable strategies because competitors can quickly match them. My belief is the best sustainable competitive edge will come from creating a culture for analytics and accountability – from training employees to have competencies with analytics that lead to fact-based and better decisions.","excerpt":"The old organizational model for decision making is broken. The jig is up. Making decisions relying on gut feeling, intuition, office politics, past experience and bias is giving way to using fact-based information and analytics. These allow for investigation, insights, foresight, and improvements. Predictable skepticism of business analytics Some readers may already be reacting to […]","categories":["IT Services"],"tags":[],"author_name":"Gary Cokins","publish_date":"2016-01-05T07:16:29","publication_year":"2016","word_count":760,"keywords":["Go","API","programming_languages:R","AI","programming_languages:Go","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","R","Go","API","GAN","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/the-time-to-create-a-culture-for-analytics-is-now\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":54697,"title":"Standford Researchers Help AI ‘See’ What’s Around The Corner","content":"Over the last few years, AI has seen significant advances in the autonomous driving sector. Countries like China and America have started testing their self-driving cars. Although there are many challenges that self-driving cars face, one of the most significant problems it encounters is predicting or detecting an object coming from around the corner.  Predicting running objects coming towards the vehicle, before making a turn or while it is passing the turn is crucial for the safety of the passengers inside the car. Artificial intelligence algorithms and the researchers at Standford University have taken a step forward towards eliminating the hassle of a camera ‘seeing’ an object that might be around the corner or not directly visible. Not only can this technology be used for self-driving cars to see around the corner, but it also has various other security purposes, such as military usages. If one knows where the threat is, it accounts for better countermeasures. A Stanford University graduate, David Lindell, has pushed the limits of cameras by using the AI systems in the future to see something that’s coming around the corner and isn’t in its direct line of sight How Does It Work? David Lindell in his Ted Talk said that; these cameras use the laser to scan the side of a simple wall or the wall of a building, to create a 3D geometrical design of the object\/scene which is hidden around the corner. The research camera makes use of the laser to scan that side of the wall where a reflection of the scene would be incident. These cameras are powerful and can capture the movement of light — faster than ‘The Flash’. David Lindell and the researchers took this high-speed camera and paired it with a laser. This laser sends out short pulses of light onto the wall opposite to the scene which is present around the corner. The technology uses the laser to capture the light scattered by the scene on the wall. What happens is that these light pulses scatter on this wall and some light scatters back to the cameras. At the same time, the camera makes use of this scattered light around the corner to get the photons of the scene scattered on the wall around the corner and gets slight imaging back through it after several repetitions. The process looks like the ultra-high-speed camera directly recording the photons as if following the reflectivity law, whatever the light is scattered on the wall by the object or the camera is captured back through reflecting\/scattering photons on the wall. Through numerous repetitions, the researchers capture different arrival times of these photons and their positions. These measurements can be used to create what David Lindell calls, ‘trillion frames per second’ video of the wall. Keep in mind that the trillion frames per second camera video are of the wall, while it is in operation, the camera is so powerful that one can see waves of light scattered back from the scenes on the wall. Theses images and the measurement are passed through a reconstruction algorithm which will process the 3D geometry of the hidden scene around the corner. The researchers have introduced a wave-based image formation model and have used the f-k migration method. Countering the Non-Line-Of-Sight (NLOS) Imaging Problem First off, NLOS is that technology which helps in recovering the obstructed objects in the scene which aren’t in the line of sight of the camera because usually, a camera captures the objects which are in its line of sight. Some of the occluded objects are lost in the process. David Lindell and team, in this work, have introduced a wave-based image formation model for the problem of NLOS. Though NLOS give out good results, there are some challenges which come with it. The image formation and the inversion models used by the NLOS are slow and work for a limited amount of hidden surfaces. Along with these problems, the NLOS only detects a limited amount of light back to it, which results in high losses. The wave-based image formation is more robust when it comes to reflective properties of various surfaces, and the powerful camera helps in achieving this too. Also, this wave-based system captures better quality reconstructions and is easy to work with. Outlook This new study has a lot of potential applications. Not only can this study find its applications in the area of self-driving cars, security, military but also in robotic vision, medical imaging and remote sensing and other major domains. Although this type of development which helps the camera see around the corner has been going on over the years using different methods like backpropagation and light cone transformation, this wave-based approach makes up for most of their limitations. But not without facing some of its challenges like imaging scenes\/objects at long distances, making efficient use of lasers which are safe for the eyes and low in power. Also capturing photons which have bounced around multiples times around the corner or encountering problems like whether the researchers can make the cameras small and safe enough to enter them into the human bodies. And with technologies like these, one has to wonder if these get augmented with AI, it can take one step towards the myth of surpassing humans.","excerpt":"Over the last few years, AI has seen significant advances in the autonomous driving sector. Countries like China and America have started testing their self-driving cars. Although there are many challenges that self-driving cars face, one of the most significant problems it encounters is predicting or detecting an object coming from around the corner.  Predicting […]","categories":["AI Features"],"tags":["stanford university"],"author_name":"Sameer Balaganur","publish_date":"2020-01-27T16:07:43","publication_year":"2020","word_count":881,"keywords":["Go","artificial intelligence","programming_languages:R","AI","RPA","programming_languages:Go","ai_applications:autonomous driving","ViT","stanford university","R"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","ViT","RPA","programming_languages:R","programming_languages:Go","ai_applications:autonomous driving"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/standford-researchers-help-ai-see-whats-around-the-corner\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10038955,"title":"HCL Shifts Workload From India To Battle COVID-19","content":"HCL Technologies has announced that the company is shifting some of its workloads from India in order to combat the current situation of COVID-19 in India. Engaging closely with clients to prioritise work, the IT major shifts some workloads from India to other geographies in order to sustain the business continuity. Vijayakumar said, “We have a global organisation. We already have almost 30 per cent of our workforces present outside India. Some geographies have stabilised and their vaccination levels are very high, so they are taking a little more load. Everybody’s very understanding of the situation, clients and our employees in other geographies.” “Clients are very understanding and supportive. And we are working with them to see how, if there is a shortfall of people in certain operations, what can we do to offset that, how can we prioritise some work over the other…very active conversations happen on a continuous basis,” he added. According to sources, C Vijayakumar, President of HCL Technologies stated that the company is working with clients to prioritise important work. At present, the IT major is working on various additional capabilities, such as 5G, Industry 4.0, AR\/VR, among others. The company intends to hire about 20,000 freshers in FY22 and is expecting to register double-digit growth in FY22 revenues. Vijayakumar stated, “The markets that we are currently very strong and prominently present in are the US, UK and Nordics. These markets are very strong and we have a large presence there. We also have a presence in France, Germany, Canada, Italy and Australia…we want to scale our presence in these markets in line with the IT spend in these markets.” The company is working with a hospital to set up a facility at two of its campuses to ensure employees get access to healthcare services. The tech giant is hopeful of the situation starting to moderate over the next couple of weeks.","excerpt":"HCL Technologies is shifting some of its workloads from India in order to combat the current situation of COVID-19 in India.","categories":["AI News"],"tags":["HCL Technology"],"author_name":"Ambika Choudhury","publish_date":"2021-04-26T14:54:07","publication_year":"2021","word_count":316,"keywords":["programming_languages:R","AI","Git","HCL Technology","GAN","R"],"extracted_tech_keywords":["AI","R","Git","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/hcl-shifts-workload-from-india-to-battle-covid-19\/","complexity_score":2,"technical_depth":5,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10046986,"title":"Can We Teach Machines To Think Twice?","content":"“PonderNet tries to find a sweet spot between training prediction accuracy, computational cost, and generalisation.” As humans, we think many times before speaking our thoughts out loud. But, can we expect the same from machines? Last week, Deepmind introduced PonderNet, a new algorithm that allows artificial neural networks to learn to think for a while before answering. Halting to think is something very familiar to humans. In machines, the target is always to pick the most optimised route in less time using lesser compute. This new model by DeepMind answers a more fundamental problem by introducing halting steps into the model. Most machine learning algorithms, wrote the researchers at DeepMind, do not adjust their computational budget based on the complexity of the task they are learning to solve. Arguably, such adaptation is made manually by the machine learning practitioner. This adaptation is known as pondering. About PonderNet Paper by Banino et al., Modern-day AI architectures are specialists. A two-dimensional residual network may be a good choice for processing images, but at best, it’s a loose fit for other kinds of data — such as the Lidar signals used in self-driving cars or the torques used in robotics. Whereas the architecture of PonderNet has a new component — a halting node that predicts the probability of halting conditional on not having halted before. The overall likelihood of halting at each step is computed as a geometric distribution. PonderNet is not regularised to minimise the number of computing steps explicitly but to incentivise exploration instead. According to the researchers, PonderNet is probabilistic in terms of the number of computational steps and the prediction produced by the network. To validate the performance of PonderNet, the researchers picked three premises: Parity from Adaptive Compute paperbAbI Q&A datasetPaired Associative Inference PonderNet builds on previous works such as Adaptive Compute while addressing their shortcomings. The above image illustrates the performance of PonderNet on the parity task. Top: accuracy for both PonderNet(blue) and ACT(orange). Bottom: number of ponder steps at evaluation time. Error bars calculated over 10 random seeds. The researchers demonstrated on the parity task that a neural network equipped with PonderNet can increase its computation to extrapolate beyond the data seen during training. With regards to the bAbI question answering dataset, which consists of 20 different tasks, this task was chosen as it proved to be difficult for a standard neural network architecture. According to the researchers, PonderNet can match the state of the art results, but it achieves them faster and with a lower average error. The comparison with Universal Transformer is interesting as it uses the same transformer architecture as PonderNet, but the compute time is optimised with ACT. Interestingly, to solve 20 tasks, Universal Transformer takes 10,161 steps, whereas in the case of PonderNet, only 1,658, implying PonderNet uses less compute than ACT. When tested on the Paired associative inference task (PAI), which is designed to capture the essence of reasoning, the results show that it benefits from the addition of adaptive computation. PonderNet optimises a novel objective function that combines prediction accuracy with a regularisation term that incentivises exploration over the pondering time. The methods used in this work achieved the highest accuracy in complex domains such as question answering and multi-step reasoning. Additionally, the experiments show that enabling neural networks to adapt their computational complexity also benefits their performance (beyond the computational requirements) when evaluating outside of the training distribution, which is one of the limiting factors when applying neural networks for real-world problems. So far, definitions around general intelligence machines have been restricted to concepts such as reward maximisation in reinforcement learning. It is thought to be the best shot we have at AGI. One can explain agents’ success in perception, generalisation and even imitation through the lens of reward maximisation. However, there are doubts on whether the reward that rationalises an act also rewards the agent for learning to perform that act and if the machines will have enough processing power. Future Direction With PonderNet, the researchers have taken the road less taken. It enables neural networks to adapt their computational complexity to the task they are trying to solve. Neural networks achieve state of the art in a wide range of applications including natural language processing, reinforcement learning, computer vision, and more. Currently, they require much time, expensive hardware and energy, to train and deploy. They also often fail to generalise and extrapolate to conditions beyond their training. PonderNet expands the capabilities of neural networks by letting them decide to ponder for an indefinite amount of time. This can be used to reduce the amount of compute and energy at inference time, which makes it particularly well suited for platforms with limited resources, such as mobile phones. “We believe that biasing neural network architectures to behave more like algorithms, and less like “flat” mappings, will help develop deep learning methods to their full potential,” concluded the researchers. Read the original Pondernet paper here.","excerpt":"Deepmind introduced PonderNet, a new algorithm that allows artificial neural networks to learn to think for a while before answering.","categories":["Deep Tech"],"tags":["DeepMind"],"author_name":"Ram Sagar","publish_date":"2021-08-25T14:00:00","publication_year":"2021","word_count":826,"keywords":["Go","machine learning","programming_languages:R","AI","neural network","computer vision","RAG","deep learning","transformer architecture","R","DeepMind"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","RAG","R","Go","transformer architecture","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/can-we-teach-machines-think-twice\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10161972,"title":"‘Replit is So Good at Generating Apps’ – Is it Time to Ditch Figma?","content":"Paul Graham, co-founder of Y Combinator and a renowned venture capitalist, recently said he encountered a CEO of a “moderately big tech company” who replaced the design tool Figma with Replit, an AI code companion tool. “Replit is so good at generating apps that they just go straight to prototype now,” Graham said as he recalled what the CEO told him. “This surprised me because I don’t even think of them as being in the same business,” he added. Multiple developers have resonated with the sentiment. Kenny Totten, founder and COO of AllFly, a travel tech company, said on X that Replit and Claude have considerably affected their startup. “Not sure we will continue with Figma. Replit is actually a little easier to prototype and design in,” he said. He also highlighted another effect, where the feedback loop from customer to product to engineering will now be five to ten times faster. Of course, it isn’t just Replit but many other AI-powered coding tools that have displayed a similar impact. Cory House, a web developer, said on X, “My team is doing the same thing but with [Vercel’s] v0 instead of Replit. Rapid prototyping is replacing Figma.”He added that designers will continue to build a design system – a set of guidelines and templates to build a product – but will need to build the actual design using tools like these instead of Figma. AIM got in touch with Viba Mohan, a product designer based out of California, for a realistic observation. Like many, she is a proponent of this change and said she’s using code to polish concepts more often than Figma. “I also see this being reflected in the industry with much smaller teams,” she said, indicating that it accelerates the process “so much faster”. “I believe larger companies will be much slower to adapt to this change and that it will hurt them,” Mohan added. But, is it fair to objectively conclude Figma’s fate – like the CEO who met Graham did? ‘There’s a Fundamental Misunderstanding Happening Here’ AIM also spoke to Chandra, a Bengaluru-based UX designer with over five years of experience. “In my opinion, there’s a fundamental misunderstanding happening here that’s worth unpacking,” he added. Chandra said that design isn’t just about making things look pretty or creating mockups but about forming hypotheses about how users will interact with a product to achieve specific goals.  “Development isn’t just about writing code; it’s about implementing these hypotheses in a way that is technically feasible, maintainable, and scalable,” he said. It is important to understand that Replit and Figma aren’t competing approaches but tools that can complement each other. “We can quickly iterate on different approaches [in Figma] without getting bogged down in implementation details,” he said, emphasising that Figma is an entire design system and a collaboration platform. It helps teams align on a shared vision, think through user flows and journeys inside a product, and test assumptions before allocating resources towards the building process. Replit, along with other AI tools, has its own role. “It’s fantastic for quick prototypes that prove a technical concept or demonstrate a specific interaction,” he said, indicating that it might replace certain workflows for companies “obsessed” with rapid validation and proof of concepts. However, not entirely. “Claiming Replit replaces Figma entirely reveals a fundamental misunderstanding of a design’s role,” he said. This raises a question about the level of product development the CEO, who interacted with Graham, was working on. “The fact that a CEO sees them as interchangeable suggests they’re operating at a surface level of product development,” Chandra opined. “It’s analogous to saying you’ve replaced your architecture firm with a construction company because they can put up buildings faster.” All things considered, it is still fair to say that Figma is lagging behind in terms of AI features. “It is also possible that Figma adapts to this change because they’re a powerful team,” Mohan said. But what is Figma doing? Figma Should Pull Its Socks Up Figma is ubiquitously used in the design ecosystem; is the company just relying on its monopoly? Many feel the tool needs to improve. Alex Kehr, CEO of Superlocal Maps, said on X that tools like Cursor and v0 could lead to Figma userbase’s collapse if the company doesn’t react quickly. “Beyond collaboration features, designing in Figma is starting to feel like it makes no sense when functional prototypes can be generated so quickly,” he added. However, Figma hasn’t shied away from the AI game. The company announced a host of AI-enabled features on its platform. For instance, Figma offers a feature to generate UI concepts with a single prompt. It does give users a headstart by quickly generating an interface with a design system. However, the features weren’t up to the mark. A review from a popular UX\/UI blog, DesignerUp, highlighted issues with creating more than a single screen. “After trying multiple prompts to generate more than one screen, it failed on all accounts. This, to me, is one of the biggest limitations and drawbacks right now – its inability to generate full flows or add additional screens,” read the review. Figma also isn’t capable of producing multiple screens with consistent styles and elements. Another user on X said that Figma’s AI rewrite tool, a feature that improves copy, feels like it uses an older version of OpenAI’s GPT-2. “It gets simplest things wrong and occasionally does the job.” Currently, several users rely on external plugins to integrate more useful AI features inside Figma.","excerpt":"Probably not.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","figma","Replit"],"author_name":"Supreeth Koundinya","publish_date":"2025-01-22T14:46:33","publication_year":"2025","word_count":922,"keywords":["Replit","Go","API","OpenAI","AI","venture capital","Scala","figma","GPT","Aim","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Go","Scala","API","GPT","startup","venture capital"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/replit-is-so-good-at-generating-apps-is-it-time-to-ditch-figma\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10066293,"title":"Intuit&#8217;s Saurabh Saxena on solving problems with a customer and a platform-focused mindset","content":"At present, many companies are heavily investing in big data and AI. But, the question is, how effectively are they using data? In an exclusive interview with AIM, Intuit India Site Leader and Vice President of Product Development Saurabh Saxena shares fascinating insights into the world of data science and AI — while discussing facets of his learnings, experiences and leadership style. Companies need to change their approach and look at data as a product. Saurabh emphasizes on an approach, known as data mesh, which favours pushing data ownership and responsibility to local domain teams. The teams will be stewards of data and support documentation for quality. Intuit’s vision is to produce clean, interoperable and re-usable data through distributed data ownership. These connected data products create a network of our data ecosystem, enabling productivity gains for consumers and data engineers. Intuit was recently recognized as number one among the 50 best firms in India for data scientists to work for – 2022 as part of the fourth annual list released by AIM. Leader’s talk At Intuit, Saxena focuses on the company’s growth journey, in line with their strategy of becoming an AI-driven expert platform, delivering best-in-class results for each stakeholder group, including employees, customers, and shareholders, to measure success. He also leads product suites of QuickBooks (QB) Online Advanced, QB Desktop, QB Enterprise, point of sale, and Intuit market—aligned with business goals—positioning Intuit for long-term development. Interview excerpts: Q: Can you tell us about your contributions and initiatives toward the company’s growth. What are your key takeaways from the experience? Our engineering teams in Bengaluru drive end-to-end product development and management for Intuit’s high-value products – Lacerte, QuickBooks Desktop, TurboTax Desktop, Billing, Virtual Expert Platform, Developer Experience, etc. Leveraging the power of AI, ML and cloud, Intuit India has been creating new opportunities with products like QuickBooks Online-Advanced, targeted toward the global mid-market business segment. Some takeaways we have observed are: While we were already working towards solving for small business growth through omnichannel commerce, the pandemic has only accelerated this journey to virtual\/digital platforms. The power of a global collaborative platform mindset is about evolving from a mindset of just building products and features to building a set of capabilities and building blocks upon which the products and features are built, contributing to innovation at scale. The power of AI and data and its role in building stronger customer experiences. We believe our AI-driven expert platform will revolutionize experiences for customers and help people make the right financial decisions for themselves, their business, and their families. People across the world will be able to instantly access the skills, insights and expert advice they need, while at the same time, experts will gain access to new clients they wouldn’t reach otherwise. Open platform contribution through open source and inner source. Open source is an important part of Intuit’s engineering culture. We rely on open-source tools and give back to the community by actively participating and sharing our projects. Members of our engineering community have open sourced many projects, including tools to speed up UI design, enable web services testing, and power productivity at scale. The inner source is the use of open-source software development best practices and establishing an open source-like culture within organizations. Programmers share our work with a wide audience instead of just with a manager or team. We can contribute to a range of projects, not just those directly connected to our team. This open culture empowers us and promotes speed and innovation. Q: What are some of the daily challenges you face as VP of product development? How do you approach them to ensure work goes smoothly as planned? A: Given the current environment, the most important consideration is our continuous journey to transform the workplace, keeping factors such as cross-team collaboration and our energizing and inclusive culture at the forefront. So one of my priorities is to reiterate and effectively and safely operationalise our hybrid workforce\/workplace strategy, which will enable us to – create a sense of connectedness and belonging; spark creativity and innovation; solve for speed, agility, and productivity, and help us attract and keep top, diverse talent without affecting the productivity of our teams. Second, it would be to align and prioritize what’s most important to our teams. For example, we have a leadership playbook to enable our managers to lead with a clear vision, build a high-performing culture and drive winning results. A key part of this exercise would include allowing teams to take ownership of the outcomes, championing a culture of inclusion and innovation, encouraging teams to say no if something isn’t a priority, and creating a culture of autonomy to maintain a high say-do ratio. Lastly, it would be to promote being nimble, having a growth mindset, and quickly adjusting to the customer needs and strategy changes. There are no failures in tech disruption, only learnings and experiments, so an important priority would be to enable teams to focus on the ‘Fail Fast and Learn Faster’ philosophy. Q: How have your roles and responsibilities changed over time, especially during the pandemic? A: If the pandemic has changed anything, it’s that I’ve only become more connected with my teams. It has also allowed me to learn the hybrid style of working and how to balance flexibility and productivity with the connectedness and innovation created through collaboration and in-person interactions, which is critical to our culture. It has also allowed me to reflect on the importance of our employees’ well-being\/wellness more than ever before. At Intuit, we understand just how important wellness is. We have expanded our ‘Wellbeing for Life Program‘ to support anything that falls under physical, emotional, or financial well-being and any expenses our employees may need to support their family’s needs. Our team also designed a simple, intuitive wellness playbook that brings together resources to create awareness and equip managers to support their own and their team’s balance and wellness. Q: Can you shed light on the skills required to thrive in turbulent times and strategies implemented to bring tangible benefits for employees? In a world where skills need to be constantly updated, Intuit seeks to solve challenging and meaningful customer problems, always evolving with new technologies, elegant and intuitive design and UI. We look for talent who have high learnability, a passion for deep customer understanding, innovation while being creative, and talent with an ability to work in fast-moving teams. Some of the core skills are being curious, doing breakthrough innovations in constrained environments, and aiming for simplicity in our product and organizsation. Next, it’s about the mindset to operate boundary-less, with a lot of compassion, demonstrating the highest level of ownership and accountability. Finally, it’s about operating with high velocity and having the platform mindset to drive that. Engineers, data scientists and product managers must have a solid, in-depth background in technical abilities like programming and a grasp of analytical tools; in terms of non-technical talents – business acumen, problem-solving ability, soft skills, and a will to create with speed determine success. At Intuit, we prioritize recruiting people with the appropriate skills and experience, but we also look for additional traits, such as a customer-centric and innovative mindset. In addition to relevant skills, we also look at how potential talent aligns with our values to ensure we have people who are the right culture fit for us. Some of the significant roles where Intuit is seeing growth are software engineering, product design, product and program management, data science, risk analytics and business analytics roles. The complementary capabilities that will see growth include artificial intelligence\/machine learning, data science, cloud, open-source and natural language understanding (NLU). Inside Intuit Intuit India has a team of over 1300+ people and is essentially a microcosm of Intuit, a global technology platform that helps its customers and small businesses overcome their most important financial challenges. The company serves over 100 million customers worldwide with TurboTax, QuickBooks, Mint, CreditKarma, and Mailchimp. In FY21, Intuit had about 30 percent of women team members in technology roles globally, and now the company is looking to make it 35 per cent by FY23. “Diversity isn’t something we do—it’s who we are. Our commitment to Diversity, Equity & Inclusion (DEI) is foundational to our company. We are one strong step ahead in building an organization that represents the diverse world we live in and the customers we serve,” said Saxena. Intuit has some world-class inclusive and employee-friendly policies, alongside networks and communities like Tech Women@Intuit, Intuit Women’s Network, and Pride community. For more than a decade, the company has been on the list of India’s Best Companies To Work for.Also, it has been recognised as a ‘Gold employer’ in the India workplace equality index (IWEI) 2021 for two times in a row. Intuit plans to expand in India, leverage the large pool of skilled tech talent available in the country, and invest in growing core tech capabilities that will drive innovation for their global products and platforms. “We need great talent across levels to change the dynamic, reshape conversations and motivate the younger generation to join and excel in the tech world,” said Saxena. Intuit is currently hiring for various roles in analytics and AI. Click here to apply.","excerpt":"“We hired over 40% of our talent in the last two years,” says Saurabh Saxena, Intuit India Site Leader & Vice President, Product Development","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Interviews and Discussions","Intuit","list of computer languages and their uses","Machine Learning","Netcore Cloud"],"author_name":"Amit Naik","publish_date":"2022-05-05T11:00:00","publication_year":"2022","word_count":1541,"keywords":["data science","Go","artificial intelligence","machine learning","AI","Intuit","ML","Machine Learning","RAG","Netcore Cloud","Aim","analytics","list of computer languages and their uses","R","AI (Artificial Intelligence)","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","analytics","Aim","RAG","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/intuits-saurabh-saxena-on-solving-problems-with-a-customer-and-a-platform-focused-mindset\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":27609,"title":"Data Analytics Is At The Core Of All Disruptions In The Business Landscape, Says Hareesha Pattaje Of Synechron","content":"Analytics India Magazine got in touch with Hareesha Pattaje, the managing director of Synechron, to get his views on the biggest challenges that analytics industry faces today. With over 20 years of expertise in IT services, program delivery and client relationship management, he is responsible for global delivery of Synechron’s software services its major clients. He has also played a pivotal role in delivering key Innovation Programs at Synechron and is counted as a top leader in the software delivery space. Analytics India Magazine: What are the three key challenges you face being in analytics industry? Hareesha Pattaje: Data and analytics are a rapidly changing part of almost every industry today. With digital transformation in the forefront, big data analytics has become an important part of the overall business strategy. Data analytics is at the core of all the disruptions currently underway in the business landscape. Some of the key challenges that we have observed while working with our clients are: Strategic Alignment: One of the biggest challenges is convincing traditional companies to move to a data-driven decision-making process. The best way to overcome this is to provide use cases to the customer highlighting the impact data analytics can have on their business. Synechron typically engage our domain experts and business consulting resources to help our clients get past this hurdle. Information Maturity: Analytics solutions are only as effective as the maturity of underlying data. Data platform implementations often struggle due to low quality and insufficient data. This usually happens as a result of lack of clear data source definitions or complexity of existing data. During initiation phases, our business consulting teams help clients identify clear requirements in terms of mapping right set of data sources and different patterns associated with it. Stakeholder Commitment: Apart from the fact that data analytics solutions enable enterprises to pave a path for business process transformations, it also requires a lot of involvement and upfront commitment from domain experts to define future business processes driven by analytics platform. Many service providers and analytics platform builders do not consider this as a key aspect while starting new initiatives and suffer down the road on their journey with the client. We identify and engage key stakeholders and ensure the right commitment is obtained from the client side while defining the analytics roadmap for them. AIM: Talent shortage is often considered the biggest challenge. What are your thoughts about it? HP: In our opinion more than the shortage of talent, the industry is facing the challenge of mismatch between supply and demand of right skilled talent. On one hand there is an abundant supply of existing resources who have domain knowledge and experience needed for enterprise implementations but have outdated technology skills, and on the other hand, there are a fresh set of resources coming out of colleges every year. Out of this very big pool, one needs to find the right fit for the challenge at hand. A lot of universities now offer specialised courses in data analytics to help bridge this gap. Businesses today believe in investing in their existing resources as well as acquire talent from the market. They regularly run intensive training programs to cross skill and upskill our existing workforce to ensure availability of right resources for our customer engagements. AIM: What are the key challenges while setting up analytics and AI startups? HP: Although Synechron does not fall under this category, we see challenges for analytics and AI startups in various areas. We often notice that most startups are unable to classify them as an AI venture and present a strong case study on the value they bring to the table for investors, which creates issues in securing funding. Secondly, the highly unpredictable structure of AI as products pose another challenge for startups if they are lacking in proper understanding and right talent. Creating a generic AI platform product that can cater to various domain and industries is again a major challenge for startups. AIM: What is the biggest challenge in carving an AI roadmap — budget, talent or senior management buy-in? HP: All the points mentioned here are the challenges in carving the right AI roadmap, however, the main challenge is to identify and demonstrate the business case leaving aside the hype and expectation around AI. An AI Roadmap should clearly be relevant to the business scenarios and align with the ecosystem one is operating upon. This will help an organisation to get the management buy-in, budget and clients buy-in too. This will also augment the process of creating and retaining the right talent to cater to the business. AIM: What are the biggest challenges in moving from pilot stage to production stage? HP: Creating the synergies between the ideation and the operationalisation of an AI solution and platforms and taking the holistic approach towards a business case seems to be a major hurdle in elevating your pilots to production. Creating an analytics solution and platform that matches expectations of various stakeholder groups is a key challenge as they generally will have a different mindset and have different priorities, for the right reasons. Groups that drive ideation and new initiatives mainly focus on the disruption and transformation aspect of the solution whereas groups that manage application and platforms in production are more concerned about non-functional aspects of the platform like stability, performance and maintainability to name a few. Teams and solution providers who do not take these aspects in consideration early, suffer either a major rework or the solution getting shot down later in the game defeating the business purpose and causing an increase in sunk cost. AIM: What are the challenges specific to the industry\/domain that you are working with? HP: Data and analytics including many topics such as data science, data lakes, and data visualisation is of growing interest – not only across banks – but also across business divisions and support functions. Given this, banks are looking for ways to overcome internal silos and leverage their data analytics capabilities across the organisation.  Banks are also now changing their strategies from being “Customer Focused” to “Customer Centric” and advancement of technologies have opened a renewed challenge to derive meaningful insights from the vast amount of data received from multiple channels to help them predict and respond to the changing consumer needs. AIM: How is the tech industry grappling with finding the requisite skills to beef up the bench strength? HP: The entire tech industry is facing the challenges including managing and retaining the right talent. This indeed is a very big challenge finding the right kind of resources with the depth of skills required as there is a shortage of analytics professionals because it requires a unique blend of expertise in areas such as mathematics, computers and domain. Most organisations nurture the talent they have and reskill them to match the market requirements.  In fact, they are devising internal training framework which collaborates with educational institutes to ensure parity between skilling programmers and addresses the industry’s requirements. Leading players are customising training frameworks and infrastructure and collaborate with recruitment partners. Furthermore, they have a sharp focus on cross skilling and upskilling of our resources through our training modules like Analytics, Artificial Intelligence (Python, NLP, NLG, BOTS, Robotics), Machine Learning, Big Data and Cloud to name a few. AIM: Ways to overcome the commonly faced challenges in the industry. HP: Service providers and analytics platform builders should take a holistic view of their offerings and align them to client requirements. Businesses need to provide end-to-end services for building analytics solutions and that have focus on the core areas of business. The specialists need to engage with clients right from the ideation and to the design phase, define the path to production and help clients implement the solutions. AIM: How does one decide what part of the IT budget should be allocated to emerging tech — ML applications, AI-based products? HP: In today’s world, with the potential disruption that new technologies like AI and Blockchain can have on the future of organisations, the leadership team is faced with a dilemma of prioritising today vs future.  Technology is now deployed in every business function, but the investment varies. Unless a business has a clear innovation strategy in place, attention to short-term burning problems take precedence and allocating resources and budget to long-term innovation initiatives often remains unaddressed.  We have typically seen companies allocating budgets ranging from 5% – 15% to innovation projects based on their focus areas. AIM: What performance metrics are used by the C-suite to measure the business outcome — growth in revenue, improved business margins or other factors? HP: Organisations typically track investments in these areas and compare it against innovation project investments to new product\/service rollouts, revenue and margins coming out of them as a result of technology transformation effectiveness. Another important factor that the leadership teams consider is the impact of not doing the investment on the future of the organisation due to potential disruptive nature of these technologies. In some cases, there can be a major impact and result in an organisation losing market share to competitors and products\/services offered can become redundant in the marketplace.","excerpt":"Analytics India Magazine got in touch with Hareesha Pattaje, the managing director of Synechron, to get his views on the biggest challenges that analytics industry faces today. With over 20 years of expertise in IT services, program delivery and client relationship management, he is responsible for global delivery of Synechron’s software services its major clients. […]","categories":["AI Features"],"tags":["Interviews and Discussions"],"author_name":"Srishti Deoras","publish_date":"2018-08-24T12:17:36","publication_year":"2018","word_count":1531,"keywords":["data science","machine learning","artificial intelligence","AI","ML","RAG","NLP","Python","Aim","analytics","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","NLP","data science","analytics","Aim","RAG","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/data-analytics-is-at-the-core-of-all-disruptions-in-the-business-landscape-says-hareesha-pattaje-of-synechron\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":6,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10080028,"title":"All You Need To Know About Meta’s New India Head","content":"Following numerous high-profile departures in the overseas market, tech giant Meta has appointed Sandhya Devanathan as the new head of India, effective from January 1, 2023. The Facebook parent company said on Thursday that Devanathan joined the firm in 2016, building the firm’s Singapore and Vietnam businesses. Devanathan was then appointed as the head and VP of Meta India. In 2020, the new India head moved to lead Meta’s gaming efforts in the Asia-Pacific region. In her new role, Devanathan will report to Meta Asia-Pacific’s Vice President, Dan Neary. The new reporting hierarchy is reportedly a shift for the company, which had earlier witnessed Indian executives report to the U.S. leadership directly. Notably, in 2019, Meta (then Facebook) created a new organisational structure for the India unit and delinked it from the APAC region. This meant that the India unit functional heads were reporting to the country managing director who, in turn, reported directly to the company’s headquarters in Menlo Park in the United States. Meta in a statement said, “Devanathan will focus on bringing the organization’s business and revenue priorities together to serve its partners and clients, while continuing to support the long term growth of Meta’s business and commitment to India.” The American tech behemoth identifies India as one of its largest markets in terms of user base, with over half a billion Indians using Meta’s services. The company’s family apps such as Instagram and WhatsApp—growing fastest in India over the recent years—have boarded millions of users. It is followed by a series of huge investments in the country, which includes a $5.7 billion check to Indian firm Jio Platforms to ramp up the commerce engine of messenger WhatsApp. Meta’s Chief Business Officer, Marne Levine, said, “I’m pleased to welcome Sandhya as our new leader for India. Sandhya has a proven track record of scaling businesses, building exceptional and inclusive teams, driving product innovation and building strong partnerships. We are thrilled to have her lead Meta’s continued growth in India.” Advocate of Women Leadership With over 22 years of being a global business leader, Sandhya Devanathan has an international career in payments, banking, and technology, and holds an MBA from Delhi University‘s Faculty of Management Studies. Devanathan spent around ten years at Citi, following which she had stints at Chartered Bank for six years as the managing director for payment products and retail banking in the Singapore region. Moreover, she holds an engineering degree from Andhra University, along with completing a course in leadership from University’s Saïd Business School in 2014. She is a strong advocate of women leadership and aims to promote diversity at the workplace. Devanathan has a track record of scaling businesses and building exceptional teams to drive product innovation. Furthermore, she is the executive sponsor for Women@APAC at Meta and the global lead for a global Meta initiative called ‘Play Forward’, established to improve diversity in the gaming industry. Amidst high-profile exits What’s more is that the new appointment comes along with several key departures in India at Meta in recent weeks. Meta India’s former head Ajit Mohan left the company late last month, joining hands with its rival ‘Snap’. He would be appointed as the president of the younger company’s business in Asia-Pacific. Earlier this week, Abhijit Bose, India Head at WhatsApp and Rajiv Aggarwal, Public Policy Head at Meta India stepped down from their respective roles. These key exits come along with the largest single layoff phase that was announced on 9 November when CEO Mark Zuckerberg stated that Meta will be sacking 11,000 of its employees, freezing all hiring at least until March 2023. Zuckerberg also shared the company’s decision in a blog post: “At the start of Covid-19, the world rapidly moved online and the surge of e-commerce led to outsized revenue growth. Many people predicted this would be a permanent acceleration that would continue even after the pandemic ended. I did too, so I made the decision to significantly increase our investments. Unfortunately, this did not play out the way I expected.”","excerpt":"Devanathan’s appointment comes after former head Ajit Mohan stepped down from Meta to join rival social media company, Snap.","categories":["AI News"],"tags":["instagram","Layoffs","Meta","reliance jio","whatsapp"],"author_name":"Bhuvana Kamath","publish_date":"2022-11-17T18:32:11","publication_year":"2022","word_count":673,"keywords":["API","Meta","Layoffs","AI","programming_languages:R","innovation","whatsapp","Aim","reliance jio","GAN","instagram","R"],"extracted_tech_keywords":["AI","Aim","R","API","GAN","innovation","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/all-you-need-to-know-about-metas-new-india-head\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10057424,"title":"Indian Army establishes new Quantum Computing Lab and AI centre","content":"The Indian Army has established a quantum computing laboratory and an AI centre at a military engineering institute in Madhya Pradesh’s Mhow. The centres will carry out extensive research in developing transformative technologies for use by the armed forces. The Defence Ministry, in a statement, has said, “The Army, with support from the National Security Council Secretariat (NSCS), has recently established the Quantum Lab at Military College of Telecommunication Engineering, Mhow, to spearhead research and training in this key developing field.” The Indian Army is making steady and significant strides in emerging technologies. Chief of Army Staff General MM Naravane visited the facility during his recent visit to Mhow. During a recent visit to Mhow, Chief of Army Staff General MM Naravane had visited the facility. A statement by the ministry said, “Indian Army has also established an AI Centre at the same institution with over 140 deployments in forward areas and active support of industry and academia.” “Training on cyber warfare is being imparted through a state-of-the-art cyber range and cyber security labs,” it said. The ministry said the research undertaken by the Army in the field of quantum technology would help it leapfrog into the next generation of communication and transform the current system of cryptography to post-quantum cryptography. The key thrust areas are quantum key distribution, quantum communication and quantum computing, among others.","excerpt":"Backed by National Security Council Secretariat (NSCS), the Indian Army has set up a Quantum Lab at the Military College of Telecommunication Engineering, Mhow.","categories":["AI News"],"tags":["Indian Army"],"author_name":"Poornima Nataraj","publish_date":"2021-12-30T11:44:09","publication_year":"2021","word_count":226,"keywords":["emerging_tech:quantum computing","programming_languages:R","AI","Indian Army","Rust","R","programming_languages:Rust"],"extracted_tech_keywords":["AI","R","Rust","programming_languages:R","programming_languages:Rust","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/indian-army-establishes-new-quantum-computing-lab-and-ai-centre\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":6829,"title":"Art &#038; Analytics of the Home Insurance Quote","content":"It is no secret that the Insurance industry is one of the biggest employers of data scientists and analysts. This report by Accenture talking about the global talent crunch in Insurance shows a shortfall of 15,000 jobs in 2015. Within the Insurance sector, the underwriting & rate setting is the most critical analytics function. If you quote an insurance rate that is too high you lose market share and if you set it too cheap, you have underwriting losses. This fine line has meant an extremely complicated analytical model driven by hundreds of variables, all trying to figure out what your resulting claims would be and then setting an insurance premium from that. In addition, companies & insurance agents use “art” on top of the science to set the final rate. The “art” part of the equation includes how much cross sell & up sell opportunities there could be with other products in the future. For example, if there is a probability for you to purchase car insurance in the future, this implies that the cost of customer acquisition will reduce on the entire bundle which will in turn make the customer more profitable. This “Art” is normally left upto the insurance agent’ judgments when the final quote is done. Some agents can even take hits on their commission depending on the “right” customer. vHomeInsurance (www.vhomeinsurance.com), a home insurance analysis service, shows an illustrative “Art” Vs “Analytics” example for a typical customer looking for Homeowners Insurance in Dallas. A standalone home insurance rate in Dallas could be anywhere between $850 to $1200 depending on a number of variables such as your street address, age, home value, number of bedrooms year your house was built. At the same time, if the customer wanted both auto & home, they could get anywhere between a 10 to 30% discount saving a few hundred dollars in the process. This discount is offered because of lower customer acquisition & customer service costs from a bundled home & auto service. But- what if the Dallas Homeowner is not sure of the auto & home bundling and wants the quote only for home insurance. The insurance agent who knows the background of the customer may still give some discount for a “potential” future bundling. This estimation of “potential” is what leads to a lack of discipline and final underwriting losses for the insurer. Warren Buffett repeatedly coaches his insurance executives that the art of saying no is more important than saying yes to have a sustainable insurance business. However, on the other hand, the potential for future bundling is real. Realizing the potential in a disciplined & analytical manner would involve building an aggregate variable called “Probability of Cross Sell” BEFORE a person becomes a customer. This implies that our probability will need to be developed without any past transactional behavior of the potential customer. The model would need to be calibrated just on 3 Factors 1) What the customer has provided 2) External Data 3) Internal Data. Merging all these activities into a single unified view for the front end agents to quote a customer is not easy. But, if one could get it right, there could be significant benefits for the insurer & the customer in moving from art to analytics. Author: vHomeInsurance (www.vhomeinsurance.com )  is a home insurance research & analytics service that focuses on the micro factors such as location, house & neighborhood as well as big picture technology , consumer and analytics trends that impact your homeowner insurance.","excerpt":"It is no secret that the Insurance industry is one of the biggest employers of data scientists and analysts. This report by Accenture talking about the global talent crunch in Insurance shows a shortfall of 15,000 jobs in 2015. Within the Insurance sector, the underwriting & rate setting is the most critical analytics function. If […]","categories":[],"tags":[],"author_name":"AIM Media House","publish_date":"2015-01-24T15:51:31","publication_year":"2015","word_count":586,"keywords":["programming_languages:R","AI","Aim","ViT","analytics","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/art-analytics-home-insurance-quote\/","complexity_score":3,"technical_depth":6,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10060316,"title":"Salesforce to launch its own NFT Cloud","content":"Salesforce plans to release an NFT Cloud, as the company sees an immense opportunity to bring the technology into enterprise software. The announcement was made at a private event by Marc Benioff and Bret Taylor, the co-CEOs of Salesforce. Salesforce wants to offer a service for artists to create content and release it on a marketplace like OpenSea. Salesforce could possibly integrate the tool into its own ecosystem where transactions can be managed. A marketplace owned by Salesforce would possibly reduce the dependency on OpenSea. Leading brands will search for utility through NFTs. Organisations will find NFTs more appealing once they move out of the novelty phase, Mathew Sweezey, Director of Market Strategy at Salesforce said. NFTs, tokens, and other crypto-based assets could also be distributed as part of innovative loyalty programs where customers become stakeholders. Homes, cars, luxury goods, can now all be tied to distributed systems of record that guarantee authenticity, linked to a historical value chain, providing individual ownership without any centralised group facilitating each transaction. The value proposition changes from one of consumption to that of ownership. For customers, this becomes not only a matter of spending but also investing in brands, said Brian Solis, VP, Global Innovation Evangelist, Salesforce in a blog post.","excerpt":"Salesforce wants to offer a service for artists to create content and release it on a marketplace like OpenSea.","categories":["AI News"],"tags":["Blockchain","Salesforce"],"author_name":"SharathKumar Nair","publish_date":"2022-02-10T14:06:59","publication_year":"2022","word_count":208,"keywords":["Go","ELT","Blockchain","programming_languages:R","AI","innovation","programming_languages:Go","Salesforce","GAN","R"],"extracted_tech_keywords":["AI","R","Go","ELT","GAN","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/salesforce-to-launch-its-own-nft-cloud\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10126512,"title":"There is No Such Thing as Experts","content":"In a recent interview, Vinod Khosla, the founder of Khosla Ventures, had some unconventional insight to share. He claimed that all major innovations almost always stem from disruptors originating outside of the field. Nobody with real experience is ever credited with any major innovation, he said. As per Khosla, these disruptors often challenge conventional wisdom and apply cross-disciplinary knowledge to create breakthrough technologies and products. In Khosla’s words, “Retailing didn’t come from Walmart, it came from Amazon. Space X didn’t come from Lockheed or Boeing. Companies like SpaceX and RocketLabs didn’t work in space before entering the field.” Outsiders Bring in Fresh Ideas And he isn’t wrong—everywhere you look, industries are being disrupted. The hotel industry encountered it with Airbnb, and the music industry experienced it with Spotify. Mark Zuckerberg created a social media platform that changed online interaction despite not being an expert in social networking. Each one of these businesses, in Khosla’s words, offered a more practical service than the tried-and-true approaches. By doing this, they not only fundamentally altered the rules for their rivals and users but also inspired a new wave of innovation. Khosla discussed how, in 2018, he decided to invest in OpenAI, which back then was a nonprofit focused on artificial general intelligence. “Many people told me that I was crazy. It was a nonprofit and there were lots of reasons not to invest in it. It was an odd structure and they didn’t have a business plan. They didn’t know if they’d ever get revenue. And I made the largest initial investment,” he said. Elon Musk is another prime example of this. Although he had no prior experience in the automotive or aerospace industries, he managed to disrupt both sectors by applying new perspectives and innovative thinking. His concept of a strong, long-range EV was fantastic. However, when Musk first introduced his vehicle, the Roadster, based on the Lotus platform, it experienced multiple failures and recalls. The technology was not flawless, the market was unpredictable, and the pricing was incomprehensible. Steve Jobs, for instance, revolutionised many industries, including computing, telecommunications, and music, not because he was a specialist in any of them. He did it because he approached problems from a user-centric standpoint, with an unwavering emphasis on design and simplicity. Apple noticed that while laptops excelled in functionality—thanks to larger, more comfortable screens and keyboards, as well as fast processors—it lacked portability that cell phones had. There was a unique desire for a device that was more portable than a laptop but had more capabilities than a smartphone, which led to the development of the iPad. At its peak, the iPad held 28% of Apple’s revenue. The Non-Expert Advantage “You don’t go hire somebody from IBM, who’s done IBM for 20 years; that experience will kill you for sure,” Khosla said in the interview. Khosla believes that disruptive innovation typically comes from individuals or teams not deeply entrenched in traditional industry practices. Let’s take a look at Khosla’s venture, Commonwealth Fusion System. In the interview, he spoke about meeting with a senior fellow from MIT’s plasma fusion lab. Khosla felt that the world needed a different kind of electricity than what was available, so he decided to take a chance with fusion energy despite nobody at the Department of Energy wanting to talk about it at the time. Commonwealth Fusion Systems is now working to develop fusion as a viable energy source. Another example is Okta, one of the first companies he invested in. Before Okta, several companies like Microsoft and IBM already offered IAM solutions. However, they were primarily designed for on-premises environments. Okta thus began offering an entirely cloud-based program, which allowed for easier deployment, maintenance, and scalability. Today, Okta holds a 28% market share. Innovation Comes from Disruptors There is another story playing out in Netflix’s (which has been around since 1997) transition from DVD shipping to streaming. This transition required the company to disrupt itself, which is an extraordinary task, as most successful disruptive innovations attack someone else’s profit pool, not one’s own. Conversely, if the other industry’s core offering is a product, consider whether it can also be offered as a service. Hubspot provides an excellent example of this type of innovation. Following the success of Google’s search engine, other companies started offering consulting services aimed at helping their clients become more visible in Google searches (i.e. search engine optimisation). Hubspot developed software enabling companies to optimise search engines, thus transforming a core offering that others provided as a service into a product. Learn Lessons in Disruption Experts play a crucial role in driving innovation and advancing technology. Elizabeth Holmes is a prime example of this. She was a non-expert in medical technology who founded Theranos, promising revolutionary blood-testing technology. The lack of deep medical knowledge led to flawed technology that failed to deliver accurate results, ultimately resulting in legal issues and the company’s collapse. So, to innovate effectively, it’s often beneficial to balance the fresh perspectives of non-experts with the deep knowledge of experts.","excerpt":"Nobody with real experience is ever credited with any major innovation.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","AI innovation","Data Science","Elon Musk","Vinod Khosla"],"author_name":"Anshul Vipat","publish_date":"2024-07-11T13:05:56","publication_year":"2024","word_count":838,"keywords":["Go","ELT","AI innovation","Vinod Khosla","AI","OpenAI","ETL","innovation","Scala","Elon Musk","Aim","GAN","Data Science","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","Aim","R","Go","Scala","ETL","ELT","GAN","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/there-is-no-such-thing-as-experts\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10003559,"title":"Building a Covid-19 Dashboard using Streamlit","content":"In data visualization, dashboards are the Graphical User Interfaces which display data in an informative and highly interactive way. It contains various plots such as bars, pies, line charts etc. that are actually the visualizations of a dataset by which we can derive some useful information. Dashboards are useful because they are easy to understand and provide us with a clear picture of the key performance indicators. Streamlit is an open-source python library that allows us to build beautiful, highly interactive, and informative dashboards easily. It also allows us to create custom based Machine Learning and Data Science applications. Every time we save the code streamlit runs from top to bottom and displays the changes in seconds because it is incredibly quick.  The UI of streamlit is visually appealing and already loaded so that we don’t have to write the code about the UI of the app. Throughout this article, we will cover the following points:- Installing streamlit Building streamlit dashboard applications Visualizing insights through streamlit dashboard In order to use streamlit, we need to install it using  pip install streamlit Before building our Streamlit Dashboard let us analyze the sample applications that are pre-loaded in streamlit. For this, we need to open the anaconda command prompt and run streamlit using the command given below. streamlit hello This will run streamlit at the local URL mentioned. You can select any demo you want to explore and go through different documentation that is provided there on the homepage. Now, let us build our own dashboard of Covid-19 in India. For creating the script for streamlit you need to have a code editor installed, I personally prefer “Atom” but you can use any code editor. The dataset I am using here can be downloaded from the online repository. Implementation a. Importing Libraries We will start by importing some important libraries which we will be using. We need libraries to load our data and to visualize the data. import streamlit as stimport pandas as pdimport numpy as npimport plotly.express as pxfrom plotly.subplots import make_subplotsimport plotly.graph_objects as goimport matplotlib.pyplot as plt b. Importing the dataset df = pd.read_csv(“C:\/Users\/covid_dataset.csv’’) c. Building the dashboard application In this dashboard, we will visualize the cases of Covid-19 in India through different visualization charts. Streamlit has predefined functions which are very helpful in creating dashboards and displaying the data beautifully. st.title()  and st.markdown() are used for headings and markdowns respectively. st.sidebar() is used for displaying data on the sidebar. st.dataframe() to display the data frame st.map() to display the map in just a single line code etc. There are many more functions in streamlit and we will try and cover maximum while creating our dashboard. Starting with setting the title and the sidebar title for streamlit dashboard st.title(\"Covid-19 Dashboard For India\")st.markdown('The dashboard will visualize the Covid-19 Situation in India')st.markdown('Coronavirus disease (COVID-19) is an infectious disease caused by a newly discovered coronavirus. Most people infected with the COVID-19 virus will experience mild to moderate respiratory illness and recover without requiring special treatment.')st.sidebar.title(\"Visualization Selector\")st.sidebar.markdown(\"Select the Charts\/Plots accordingly:\") A. Visualization (part-1: Cases using bar chart and pie chart) In this part, we will create a widget that will display the data in a bar plot and pie chart form. While creating widgets in streamlit we should keep in my about allotting a unique key for similar types of widgets others streamlit won’t be able to differentiate between the widgets. The code below will create our first part of the visualization. select = st.sidebar.selectbox('Visualization type', ['Bar plot', 'Pie chart'], key='1')if not st.sidebar.checkbox(\"Hide\", True, key='1'):     if select == 'Pie chart':                         st.title(\"Selected top 5 cities\") fig = px.pie(df, values=df['Confirmed'][:5], names=df['State'][:5], title='Total Confirmed Cases')             st.plotly_chart(fig)                       if select=='Bar plot':        st.title(\"Selected Top 5 Cities\")        fig = go.Figure(data=[        go.Bar(name='Confirmed', x=df['State'][:5], y=df['Confirmed'][:5]),        go.Bar(name='Recovered', x=df['State'][:5], y=df['Recovered'][:5]),        go.Bar(name='Active', x=df['State'][:5], y=df['Active'][:5])])        st.plotly_chart(fig) B. Visualization (part 2: active and  confirmed cases on time series) For this, we need another dataset which contains the date wise cases registered in India so that we can create a line plot to visualize the trend of Covid-19. The code below will provide us with a new widget which will show us the time-series visualization. df2 = pd.read_csv('C:\/Users\/Divya\/timeco.csv')df2['Date'] =  df2['Date'].astype('datetime64[ns]')select1 = st.sidebar.selectbox('Select', ['Confirmed', 'Recovered'], key='2')if not st.sidebar.checkbox(\"Hide\", True, key='3'):    if select1 == 'Confirmed':        fig = px.line(df2, x=\"Date\", y=\"Cases\")        st.plotly_chart(fig)    elif select1 == 'Recovered':        fig = px.line(df2, x=\"Date\", y=\"Recovered\")        st.plotly_chart(fig) C. Visualization (part-3: Area Charts) In the third part, we tried creating a visualization known as area chart which will allow us to visualize the COVID spread according to area charts. Don’t forget to set the unique key for the widgets. select2 = st.sidebar.selectbox('Select', ['Confirmed', 'Recovered'], key='3')if not st.sidebar.checkbox(\"Hide\", True, key='4'): if select2 == 'Confirmed':         fig = px.area(df2, x=\"Date\", y=\"Cases\")         st.plotly_chart(fig)          elif select1 == 'Recovered':                fig = px.area(df2, x=\"Date\", y=\"Recovered\")                    st.plotly_chart(fig) Similarly, we can create as many sections as we want and visualize different aspects of a dataset using streamlit. While making dashboards using streamlit you will realize that displays the changes as soon as you save them which helps in visualizing them simultaneously and saves a lot of time. This dashboard is just an example of what streamlit can do, you can try creating streamlit dashboards with your dataset and make visually appealing charts and graphs. Now let us have a look at the Dashboard we have created. For running our dashboard we need to run a python script using streamlit. I saved my file named ‘covid.py’. You can choose whatever name you want and launch it by running the below command in the Anaconda command prompt. streamlit run covid.py This is how the homepage of our app is looking, we can select different options and visualizations from the sidebar and see how streamlit displays it. These graphs and plots are created using Plotly that is why they are highly interactive and can be downloaded easily. Streamlit also allows us to record a screencast of our app so that we can share what we have created with our friends, colleagues, etc. Conclusion Streamlit is considered to be the fastest-growing Machine Learning and Data Science app building platform. It is open-source and the community is updating it frequently with new updates. It is considered to be one of the best dashboarding tools. Now as you have learned how to use streamlit go ahead and keep experimenting with the different datasets and create new apps and dashboards.","excerpt":"Streamlit is an open-source python library that allows us to build beautiful, highly interactive, and informative dashboards easily. It also allows us to create custom based Machine Learning and Data Science applications.","categories":["Deep Tech"],"tags":["data visualization","data visualization tools","streamlit"],"author_name":"Himanshu Sharma","publish_date":"2020-07-29T17:00:00","publication_year":"2020","word_count":1056,"keywords":["data science","NumPy","machine learning","Plotly","AI","ML","streamlit","data visualization tools","Python","Streamlit","data visualization","Matplotlib","Pandas"],"extracted_tech_keywords":["AI","machine learning","ML","data science","Streamlit","Pandas","NumPy","Matplotlib","Plotly","Python"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/building-a-covid-19-dashboard-using-streamlit\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":18024,"title":"Karnataka To Invest Rs 40 Crore To Set Up Centre Of Excellence In AI, Data Science","content":"The state that proudly boasts of having India’s very own Silicon Valley within it, is now set to get another boost in the technology sector, especially Artificial Intelligence and Data Science. A statement issued by the Karnataka government says that they are set to invest Rs 40 crore to establish a centre dedicated to developing Artificial Intelligence and Data Science capabilities in the Karnataka and across the country. The new body, which will be called the Centre of Excellence for Data Science and Artificial Intelligence (CoE-DS&AI), will work with Nasscom as the programme and implementation partner. Reportedly, the government wants to create one-of-its-kind port based on Public-private partnership model in Karnataka. The official announcement, which came on Wednesday, confirmed the reports that have been circulating since last week in the media. Priyank Kharge, minister for IT, biotechnology and tourism for government of Karnataka, said in the press note, “This Centre of Excellence is the logical next step required to provide the right fillip to areas of data science and artificial intelligence and give a head start to not just the state, but India as a destination to develop global product solutions.” Kharge added that the CoE-DS&AI will not only help global corporates planning on setting up their global analytics practices in Karnataka, but also create jobs for 35,000 data science and artificial intelligence professionals over the next five years. “India is already on the path of a digital revolution and the next step is utilising the big data generated to take intelligent decisions. This requires close collaboration between academia, the private sector and public sector in order to understand problems holistically and solve them,” a joint study by ASSOCHAM and PwC had stated recently. Karnataka, especially Bengaluru, had recently been lauded for being the breeding ground for Halli Labs, a startup that revolves around artificial intelligence, which was acquired by Google for an undisclosed amount.","excerpt":"The state that proudly boasts of having India’s very own Silicon Valley within it, is now set to get another boost in the technology sector, especially Artificial Intelligence and Data Science. A statement issued by the Karnataka government says that they are set to invest Rs 40 crore to establish a centre dedicated to developing […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Data Science","karnataka","NASSCOM","Priyank Kharge","pwc"],"author_name":"Prajakta Hebbar","publish_date":"2017-10-03T11:45:53","publication_year":"2017","word_count":315,"keywords":["karnataka","data science","Go","big data","artificial intelligence","programming_languages:R","AI","Git","Priyank Kharge","analytics","pwc","Data Science","NASSCOM","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","data science","analytics","R","Go","Git","big data","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/karnataka-invest-rs-40-crore-set-centre-excellence-ai-data-science\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":48552,"title":"Future Of Machine Learning On Smartphones","content":"Today, any typical modern-day smartphone is able to scan faces, documents, QR codes, capture super-resolution photos, recognise gestures, voice and perform multiple other tasks besides answering calls and texts. These handheld devices are the epitome of software and hardware engineering; and to do these tasks, they require state-of-the-art image recognition and NLP models running in the background. Image and language models are at the heart of many machine learning applications today and training these models is a computational nightmare with increasing data. Google has been using TensorFlow Lite for taking pictures on its flagship model Pixel. For Portrait mode on Pixel 3, Tensorflow Lite GPU inference accelerates the foreground-background segmentation model by over 4x and the new depth estimation model by over 10x vs CPU inference with floating-point precision. Apple says that it is using machine learning in the iPhone 11’s cameras to help process their images, and that the chip’s speed allows it to shoot 4K video at 60 fps with HDR. Whereas, Samsung’s Galaxy S10 series phones and Galaxy Fold use neural processing units (NPUs) to power Scene Optimizer that enhances the ability to recognise photos. Therefore, in order to bridge the gap between the realtime magic that ML has to offer and hardware inadequacies, chipmakers and phone manufacturers are coming with customised processors designed to deal with the demands of neural networks. How The Adjustments Were Made To Meet The Demand Even though Deep learning algorithms have been around since the early 90s, the lack of right kind of hardware created a primary hurdle for many developers at least until 2009. In 2015, Qualcomm kick-started the deep learning on mobiles movement with its efforts to accelerate models using mobile GPUs. The most important milestone in this space occurred in 2017 with the introduction of TensorFlow Lite. This framework offered options optimised for on-device inference. This library also got support for the Android Neural Networks API (NNAPI), allowing for access to the device’s AI hardware acceleration resources directly through the Android OS. This enabled building an ML pipeline without using specialised vendors tools or SDKs. havng said that,The use of floating-point and quantized models for mobile devices has been a topic of discussion amongst the developers and vendors. With floating point inference, the model is in the same format as it was originally trained on the server, however, models working with high-resolution image transformations, require more than 6GB of RAM and enormous computational resources. Whereas, the quantized approach allows the model to be first converted from a 16-bit floating-point type to int-8 format, in a way, reducing the size and RAM consumption by a factor of 4 and potentially speeds up by 2-3 times. The disadvantage here is that reducing the bit-width of the network weights (from 16 to 8 bits) leads to accuracy loss. Even though extensive research is being done by the likes of Google and Qualcomm, the quantized inference is still yet to find a solution for large scale deployment. What Do Experts Have To Say via ETH Zurich In order to assess the state of deep learning in the era of smartphones, researchers from ETHZurich, Google, Huawei, Qualcomm and other top companies collaborated to publish a paper. The above picture illustrates the comparison of the performance evolution of mobile AI accelerators. For this comparison, the Mobile devices were running the FP16 model using TensorFlow Lite and NNAPI. In this work, they evaluated the performance and compare the results of all chipsets from Qualcomm, HiSilicon, Samsung, MediaTek and Unisoc that are providing hardware acceleration for AI inference. The researchers list their findings as follows: When compared to the second generation of NPUs, the speed of floating-point and quantized inference has increased by more than 7.5 and 3.5 times, respectively, bringing the AI capabilities of smartphones to a substantially higher level.All flagship SoCs presented during the past 12 months show a performance equivalent to or higher than that of entry-level CUDA-enabled desktop GPUs and high-end CPUs.TensorFlow Lite is still only one major mobile deep learning library, providing reasonably high functionality and ease of deployment of deep learning models on smartphones. Deep Learning Is Just A Touch Away via Apple Apple, on the other hand, has been very vocal about their interest in building a next-generation machine learning platform. The enhancement of their hardware services combined with state-of-the-art software options has put Apple at the frontiers of machine learning advancement. “The A13 Bionic is the fastest CPU ever in a smartphone,” Apple said at their recently concluded mega event, adding that it also has “the fastest GPU in a smartphone,” too. The iPhone 11 is powered by Apple’s new A13 Bionic chip, which Apple touts as its faster processor ever. As for battery life, the iPhone 11 packs a one-hour-longer battery life than the iPhone XS. The A13 also features an Apple-designed 64-bit ARMv8.3-A six-core CPU, with two high-performance cores running at 2.65 GHz called Lightning and four energy-efficient cores called Thunder. The 2 high-performance cores are 20% faster with 30% reduction in power consumption, the 4 high-efficiency cores are 20% faster with a 40% reduction in power consumption. With all SoC vendors and phone makers like Apple and Samsung, determined about AI for mobiles, running many state-of-the-art deep learning models on smartphones in the last few years have radically changed. Today devices having Qualcomm and other top systems on a chip (SoCs) come with a dedicated AI hardware designed to run ML workloads on embedded AI accelerators. The latest Android 10 too, has an updated 1.2 version At the TensorFlow’s developer summit, held earlier this year, along with TensorFlow 2.0, the team also announced the open sourcing of TensorFlow Lite for mobile devices and two development boards Sparkfun and Coral which are based on TensorFlow Lite for performing machine learning tasks on handheld devices like smartphones. TensorFlow Lite aims at making smartphones, the next best choice to run machine learning models. These proceedings only mean that in the coming two-three years, all mid-range and high-end chipsets will get enough power to run the vast majority of standard deep learning models developed by the research community and industry. Not only chipmakers but there is a lot coming from the other end as well. Frameworks like TensorFlow are being developed to suit the demands of the hand held devices. With advancements emerging from both ends, the goal to make smartphones the next hub for deploying ML models is soon going to be a reality.","excerpt":"Today, any typical modern-day smartphone is able to scan faces, documents, QR codes, capture super-resolution photos, recognise gestures, voice and perform multiple other tasks besides answering calls and texts.  These handheld devices are the epitome of software and hardware engineering; and to do these tasks, they require state-of-the-art image recognition and NLP models running in […]","categories":["Deep Tech"],"tags":["Qualcomm","smartphones","Tensorflow"],"author_name":"Ram Sagar","publish_date":"2019-10-21T18:56:56","publication_year":"2019","word_count":1076,"keywords":["CUDA","machine learning","AI","neural network","ML","image recognition","smartphones","NLP","Aim","deep learning","Qualcomm","TensorFlow","Tensorflow"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","neural network","NLP","Aim","TensorFlow","image recognition","CUDA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/can-smartphones-run-machine-learning-models-qualcomm-apple-google-samsung\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":33721,"title":"IIT Hyd Introduces BTech In Artificial Intelligence; Admission Through JEE Advanced","content":"With the growing demand and applicability of artificial intelligence, IIT Hyderabad is set to launch a full-fledged BTech program in AI starting from the academic year 2019-2020. Admissions to the course will be accepted based on the JEE Advanced score. With this, IIT Hyderabad becomes the first Indian educational institution to offer a full-fledged BTech programme in AI and reportedly the third institute globally after Carnegie Mellon University and Massachusetts Institute of Technology in the US. The course will reportedly only take in only 20 students. Students pursuing other degrees such as B.Tech. in chemical engineering or mechanical engineering can also pursue a minor in AI as well from the coming academic year onwards. Along with this, the Department of Liberal Arts in collaboration with faculty from Computer Science and Electrical Engineering in IIT Hyderabad has also launched a minor in AI and Humanity. Major Objectives Of The Program To produce students with a sound understanding of the fundamentals of theory and practice of artificial intelligence and machine learning. To enable students to become leaders in the industry and academia nationally and internationally and meet the pressing demands of the nation in the areas of AI and machine learning. USP Of The Program The programme at IIT Hyderabad unique is that it aims to provide a holistic view to students. The programme comprises algorithms from the Computer Science Department, Signal Processing from Electrical Engineering Department, Robotics from Mechanical Engineering Department and Mathematical Foundations. The course will also focus on application verticals such as healthcare, agriculture, smart mobility, among many others. The ethical impact of AI and its technologies on areas such as privacy, bias and related issues will also be a key component of this B.Tech. programme. #IITHyderabad witnesses impressive achievements in 2018. Prof UBDesai says: “IITH looks forward to even stronger growth in 2019.Main thrust will be in #ArtificialIntelligence #5G #CyberPhysical Systems, Batteries.\"https:\/\/t.co\/1JumoQzBGK@PrakashJavdekar @HRDMinistry @IndiaDST pic.twitter.com\/VXxLEmjyYK — IIT Hyderabad (@IITHyderabad) January 4, 2019 “The basic aim is to create a complete ecosystem for Artificial Intelligence Academics and Research at IIT Hyderabad. This involves B.Tech, M Tech and different minor programs in AI. Moreover, the R&D will be strongly entwined with academics,” UB Desai, director, IIT Hyderabad told a leading news wire. While highlighting the course details, Sumohana Channappayya, Dean (R&D), IIT-H, said, “AI solutions are particularly promising for India given the availability of a large corpus of data where it can have a major positive impact on several critical domains such as healthcare, crop and soil management, weather prediction, surveillance and security, and defence. “However, the demand for professionals trained in this area far exceeds the current supply. The B.Tech programme in AI is a step in the direction of addressing this highly skewed demand-supply scenario,” she said to a leading daily. Further, she added that the potential for AI to improve the quality of human lives is tremendous. This program will train students in the fundamentals of computer science, AI and ML in addition to sensitising them to the ethical issues in deploying AI-based solutions. IIT Hyderabad is already offering an MTech program in AI and ML, and an MTech in Data Science since 2015-16.","excerpt":"With the growing demand and applicability of artificial intelligence, IIT Hyderabad is set to launch a full-fledged BTech program in AI starting from the academic year 2019-2020. Admissions to the course will be accepted based on the JEE Advanced score. With this, IIT Hyderabad becomes the first Indian educational institution to offer a full-fledged BTech […]","categories":["AI News"],"tags":["Btech","iit hyderabad"],"author_name":"Martin F.R.","publish_date":"2019-01-18T08:45:16","publication_year":"2019","word_count":527,"keywords":["data science","Go","artificial intelligence","Btech","machine learning","AI","programming_languages:R","ML","iit hyderabad","Aim","Rust","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","data science","Aim","R","Go","Rust","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/iit-hyd-btech-ai-admission-jee-advanced\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10071446,"title":"How TVS Motors utilises AI","content":"Commemorating over 35 years of unrivalled racing legacy now, TVS Motor Company has been one of the most dynamic motorcycle brands in the world. With four state-of-the-art manufacturing facilities in Mysuru, Hosur and Nalagarh in India and Karawang in Indonesia, the global two and three-wheeler manufacturer endeavours to deliver the most superior customer experience across 80 countries in which they operate. TVS Motors is also a recipient of the prestigious Deming Prize. “At TVS Motor, we have always focused on bringing great customer experiences driven through robust and innovative technologies, be it in our vehicles, our production engineering systems or through digital and AI technologies,” says Maheshwaran Calavai, Chief Digital and AI Officer at TVS Motor Company. In an exclusive interview with Analytics India Magazine, Maheshwaran sheds light on how the brand leverages AI to build superior products, enhance business performance, and offer an immersive customer experience. AIM: How is AI driving the latest trends in the automotive industry? Maheshwaran Calavai: AI is transforming industries across sectors, and mobility is no exception to this. AI-driven digitisation of consumer experiences is one of the key trends being leveraged by the automotive industry. Mobility companies are improving consumer experiences during the initial purchase and throughout the ownership. Since there is a continual engagement with customers and since these touchpoints are getting digitalised end-end, applications for data-based decisions are extensive and are being actively pursued by the industry. Furthermore, the rapid rise of CASE (Connected, Autonomous, Shared, Electric) mobility is one of the key developments led by the rising interest in technology development. Today’s auto enthusiasts are more tech-savvy, and they choose IoT-enabled automobiles that meet their needs for safety, navigation, audio streaming, and other features. The rising usage of AI-driven technology, along with a surge in electric and hybrid mobility, is one of the biggest trends to watch in the automobile sector. Customer preferences of ownership vs usership of vehicles for their mobility needs, along with connecting multimodal options for end-end convenience, is a key trend that will shape the industry this decade. Autonomy driven by AI will be a major contributor to this shift. Connectivity also plays a crucial role in designing and manufacturing vehicles and associated components to improve speed, quality and efficiencies in operations. Reflecting on the challenges, deployment of AI-enabled applications at-scale in sales across the extended value chain, like in dealerships, depends on consumers’ preferences. The automotive buying experience is today largely “phygital“, wherein consumers often discover and research digitally, but the final buying happens in a dealership. Convincing dealer partners that AI algorithms can present better insights on consumer preferences, sales potential, and engagement during vehicle ownership becomes critical to creating a balanced and seamless experience across physical and digital touchpoints. There are several such examples, and we look at human-in-the-loop AI systems as an opportunity to combine the precision of algorithms with the experience of experts. AIM: How important is data science at TVS Motors? Maheshwaran Calavai: At TVS Motor Company, data engineering and data sciences remain at the forefront for better decision-making in all aspects of our operations. We leverage data science-driven applications in both direct-to-customer and digital aided experiences in dealerships. For example, TVS iQUBE Electric scooter offers advanced features for our customers that are rooted in data sciences. On the digital-aided side, arming our dealer partners with customer preferences during sales and vehicle service improves customer experiences. Hyperlocal marketing improves marketing efficiency and AI leveraging vision, and NVH (noise vibrations and harshness) improve supply chain quality and efficiencies. AIM: Elaborate on ways TVS Motor leverages AI with use cases. Maheshwaran Calavai: Our international business spans 80 countries across the globe, where customers use our two and three-wheelers for their mobility needs. Due to the specific requirements in various countries, we despatch vehicles in CKD (completely knocked down) & SKD (semi-knocked down) conditions. We won’t assemble the complete vehicle even if one part is missed, mismatched, or has a defect. We first implemented an IoT-based platform for packing assembly using barcode-based scanning to ensure all the right parts are placed in the right boxes. To make the packing process Poke Yoke (mistake proofing), we wanted to deploy Vision-based AI systems to complement the IoT platform. Further, weighing machines present a good mechanism for parts like chain and cable assemblies that do not have fixed profiles. So, we have developed and fully integrated deep learning algorithms and complementary systems with the manufacturing execution system (MES) in multiple stages of packing assembly. For each vehicle variant, the stagewise parts to be packed are displayed on a computer screen, and each part is identified through the Vision AI system. If it is the right part, the conveyor will run, and assembly will continue. On the other hand, if the vision AI system detects a wrong part, the conveyor will stop since the conveyor PLC is integrated with the result of the vision system. The operator then inspects and replaces it with the correct part. In another example, we used AI vision during the pandemic to ensure adherence to safety protocols. We developed an “AI-Based – Social Distance Monitoring System & Mask Wearing Detection System” in our factories and global offices of TVS Motor group companies. We used YOLO – deep architecture for Person Detection & RCNN (Region-based Convolutional Neural Network) based deep architecture for mask detection. Both models have their advantages and limitations. At TVS Motor Company, we see large volumes of consumers’ interests. Such enquiries have different propensities to buy as they may be at different stages in their buying journey. Each lead requires the TVS Motor marketer and dealer sales staff to follow up promptly and personalised. At any point, our dealer partners in India will have 5+ lakh enquiries to follow up. To classify leads and help our dealer partners make profitable decisions, we extensively use data science algorithms powered by AI\/ML. We view AI as a platform that provides opportunities for human-in-the-loop solutions, allowing the best of experts and technology to come together to improve decision accuracy, consistency, and speed. To ensure all aspects of AI are taken care of, we have several workstreams that strengthen the foundations of AI. First, we ensure the underlying data is managed and governed well across our group companies through an established industry framework. We maintain strict standards when it comes to data robustness. We also have data privacy and protection programs, along with data enrichment techniques. AIM: What is TVS Motors’ acquisition strategy? Maheshwaran Calavai: Our focus is to provide the best products and services for our customers through innovation – some we create ourselves, some we co-create, and some we partner with other innovators. Over the last few years, we have made significant investments in startups such as TagBox, Rapido, Scienaptic, Ultraviolette and others. In addition to investments, we work with several startups in customer-facing, manufacturing, sustainability, and enterprise operations. We are also working on open innovation initiatives with industry partnerships. AIM: What are your thoughts on Autonomous Vehicles? Maheshwaran Calavai: Autonomous driving will radically transform the mobility of people and products in the coming years. Self-driving technology will first and foremost provide convenience and ease to commuters. If you look at how automation has helped consumers in the last century, hours and energy spent on tasks have been made easy and converted into a formal economy, with actual GDP being added to countries and revenues to companies. Autonomous vehicles will be no different, freeing up human energy and time that can be devoted to other things they would like to do! From an economics standpoint, this will simultaneously offer immense opportunities to create and grow new products, business models and innovations whilst also reshaping the communities we live, work and commute in. With full autonomy, vehicles can be on the road all the time. Vehicles owned by consumers for their personal needs are largely idle today except for a few busy hours on the road daily. If vehicles can ply autonomously, the productive time per vehicle day can be maximised. This will bring into question why one would want to own a vehicle when they can get a vehicle to fetch them whenever they want. So instead of a vehicle being a depreciating asset, it can be a revenue-driving investment. How will this also shape the vehicle financing industry? If we can maximise the productive hours of a vehicle every day, why would we need the same number of vehicles to meet the same demand? If this decreases the number of active vehicles in the world as a result, why would we need so much parking space in the cities and communities we live in? The same autonomous driving tech will power drones enabling 3-dimensional transportation in micro-geographies. How will this shape urban planning and transportation systems? If they can play on their own, why can’t vehicles like cars or vans be designed to be bidirectional with no need to reverse? If vehicle-to-vehicle and vehicle-to-infrastructure communication and associated navigation are perfected, what would happen to vehicle insurance? If there is indeed an incident, which is to assume ownership of the failure in communication and navigation? How will these change laws and regulations? Several industries like motels and highway convenience stores are designed to give tiring drivers a break during their long journeys. What will happen to these? These are exciting questions that autonomy brings, and it is indeed an exciting time to be in the mobility industry at a time when technology is on the brink of making these questions real and now! While several specialities from design and engineering to marketing and law will come together to solve these, it is doubly exciting to be in the data engineering and data sciences field during this time as these are large-scale data assimilation, engineering and algorithm solutions that will be at the core of autonomous driving. Even before full level 5 autonomy, innovations and usage of AI in vision, voice and tactile are key to aid mobility. These are akin to the human-in-the-loop AI examples noted above. AIM: How far are we from achieving level-5 vehicle autonomy? Maheshwaran Calavai: The automotive sector is fast evolving in autonomy. We are on the cusp of a global breakthrough, but we still have some way to go before we reach level 5 autonomy as far as two-wheelers are concerned. Several companies and organisations are researching and developing the technologies, infrastructures and policies needed to gear for full autonomy, including in autonomous two-wheelers. The sensing and computing power needed for full-scale autonomy needs to be finetuned for the form factor and price points of two-wheelers which is an engineering and commercial problem different than for other, larger vehicles. Similarly, challenges like self-balancing, riding behaviour and safety considerations are also different in the case of two-wheelers. Further, most of the two-wheelers in the world in developing economies with different traffic patterns, usage on un-laned and off roads, road conditions etc., need to be solved. AIM: What does the future hold for TVS Motors from a technology standpoint? Maheshwaran Calavai: We are working on several immersive experience technologies in our digital assets and in-store. Some of these, like chatbots, 360 virtual experience, in-dealership augmented reality, and real-time visual feedback of vehicle service, are already deployed, and several others are underway. We continue strengthening our capabilities to provide mobile-first, personalised experiences for our customers in direct-to-customer and digital aided engagements. Such technologies and capabilities also have transformative power in automotive design, manufacturing, supply chain and enterprise operations. Many of these digital experiences have AI underpinning them to personalise and make them more effective. To power these AI systems, we have adopted a cloud-first data engineering approach, coupled with foundational efforts to enrich, manage and govern data with the requisite data privacy and protection programmes. We leverage open-source software and develop purpose-built algorithms. Digital and AI products are built using an agile framework and DevSecOps methodologies. This digital and AI path we are on has created several wins already, and we want to create a competitive advantage with these technologies by providing immersive customer experiences, superior products, and better business performance in the years to come.","excerpt":"As one of the leading brands in mobility, we see our roles as an enabler in moving the industry forward and future-ready through such partnerships in the innovation ecosystem.","categories":["AI Features"],"tags":["autonomous driving","Interviews and Discussions"],"author_name":"Sri Krishna","publish_date":"2022-07-25T10:00:00","publication_year":"2022","word_count":2025,"keywords":["data science","AWS","AI","neural network","chatbots","ML","RAG","Aim","deep learning","analytics","autonomous driving","Interviews and Discussions"],"extracted_tech_keywords":["AI","ML","deep learning","neural network","data science","analytics","Aim","RAG","chatbots","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-tvs-motors-utilises-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":7,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10138407,"title":"Ola Co-founder Ankit Bhati &amp; Others Invest $1 Mn in AI No-Code Startup","content":"Tablesprint, an AI no-code app development platform, has raised approximately $1 million in a funding round led by notable investors, including Ankit Bhati (co-Founder, Ola), Ajeet Khurana (Founder, Reflexical), Sunil Sharma (CEO, Coingape), BlueLotus Ventures, TDV Partners, and others. The funding will be used to accelerate product development, strengthen the team, and support the company’s mission to provide businesses with future-ready, low-cost apps. Founded earlier this year, Tablesprint’s AI-first SaaS platform allows companies to rapidly build customisable applications for various business functions, including HR, sales, operations, and vendor management. Its no-code solution features modular building blocks such as AI write\/image tools, workflows, and charts, enabling businesses to start with simple tasks like surveys and scale to full-fledged workflows. Co-Founder Abhijeet Kumar, who previously co-founded RainCan, which got acquired by BigBasket and rebranded as BBdaily, emphasised the platform’s ease of use. “We are building an AI-powered platform that helps enterprises go live in minutes, not months. By delivering a consumer app-like experience for enterprise use cases, we enable simple workflows to scale into complex, enterprise-grade systems, allowing businesses to adapt and grow seamlessly.” Tablesprint’s approach addresses inefficiencies in traditional ERP and CRM systems, which often suffer from low user adoption rates. Chirag Jadhav, co-Founder and CTO, highlighted the platform’s multi-tenant system designed for both developers and business users. “Creating a system that resonates with both developers and business stakeholders is a significant challenge. We are excited with our progress so far and look forward to tackling more challenges ahead on this journey.” Tablesprint’s clients include large enterprises like Flipkart and startups in the wealth management sector, such as Elever. The platform is also running pilot programs with medium and large global companies across industries like manufacturing and investment management. Uday Arya, co-founder and partner at BlueLotus Ventures, expressed confidence in the Tablesprint team, stating, “We see enormous opportunities for companies enabling SMBs in a thoughtful way. This fundraise marks the first milestone in creating what we believe will be an exceptional global product company from India.”","excerpt":"Tablesprint’s AI-first SaaS platform allows companies to rapidly build customisable applications for various business functions, including HR, sales, operations, and vendor management.","categories":["AI News"],"tags":["no code platforms","Ola"],"author_name":"Mohit Pandey","publish_date":"2024-10-16T13:36:40","publication_year":"2024","word_count":335,"keywords":["Go","API","Ola","AI-first","funding","programming_languages:R","AI","ML","RAG","no code platforms","R","startup"],"extracted_tech_keywords":["AI","ML","RAG","R","Go","API","AI-first","startup","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/ola-co-founder-ankit-bhati-others-invest-1-mn-in-ai-no-code-startup\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":21169,"title":"From Initial Screening To Predictive Hiring – A Look At How AI Impacts Recruitment","content":"AI’s onward march has threatened jobs across industries, but the impact doesn’t stop there. Artificial Intelligence (AI) has impacted almost all sectors, and that includes hiring and recruitment space as well. Predictive analytics is going to change the way millions of people are hired and assessed. According to a whitepaper, The Robots Are Here- How AI Has Reshaped Recruiting, AI is viewed as a tool for the future and organizations are chasing a five-year deadline to transform HR from top to bottom. Leading global consultancy firm Deloitte predicted that 38 percent of companies(10,400 respondents from 140 countries) believe that robotics and automation will be “fully implemented in their business within five years.” Companies like Amazon and JP Morgan & Chase have braced themselves for drastic changes in HR function. Case in point – JPMC implemented an AI product that reduced their hiring need of legal professions for loan interpretations. Meanwhile, organizations like Citigroup, Airbnb and Reebok are using Koru a predictive hiring software that identifies the best drivers of performance in the company, increases high quality hires, and reduces bias. Understanding AI’s Role in Recruitment Interestingly, more than half of recruiters have griped about mining profiles of potential candidates being the hardest part of their work. With millions of candidates profiles and resumes across the internet, no top-notch recruiter, or even a Boolean black belt, can cope with those quantities, notes the paper. However, AI surpasses Boolean search in its ability to find matches that don’t contain specific keywords. It does this based on correlations and through constant learning as it determines what skills and other attributes are similar and appear in combination. Let’s Have a Look at Intelligent AI Screening Tools and they are upending early assessment Belong.co team 1) Joberate’s J-score:  Today, AI plays a huge role in initial screening with intelligent tools flooding the market. For example, Jobearte founded in 2014 has developed a technology solution that scans public social media for signals about the global workforce, and from this publicly available data, curates unique people analytics including the patent-pending metric known as J-Score, which measures job seeking activity level. The talent search and analytics technology provider has come up with a metric called the “J-score” to assign those who use social media to look for a new career opportunity a job seeking score. According to the CEO Michael Beygelman, a higher J-score, means a person is more active in their search. This helps recruiters find and target people who are most open to their message. 2) LinkedIn Talent Solutions: Automated candidate search has made recruitment tools like LinkedIn powerful platforms for recruiters. According to a business report, LinkedIn’s hiring solutions have gained more traction than its other divisions. Jeff Weiner founded LinkedIn’s primary sources of revenue includes ads and marketing, talent solutions and premium account subscriptions, but Talent Solutions contributes more than 65% of its revenues in 2015. To add more teeth to their Talent Solutions divisions, the Microsoft-owned company even acquired a data science jo search startup Bright.com in 2014 that matches candidates to the right job. According to an industry report, LinkedIn’s Talent Solutions group already accounts for the majority of its revenue and recruiters form its biggest group of playing members. LinkedIn’s recent AI offerings allow recruiters to match top performer profiles to the database to find others like them. The paper from Austin-based Proactive Talent cites how in the future, LinkedIn would find the right prospective employee, do the initial chat, answers questions and confirm the interest – and finally send the hiring managers three or four top candidates for interviews. 3) Bangalore’s Belong.co brings predictive solutions to hiring: Closer home, Bangalore-based AI recruitment startup Belong.co aims to become a “Google for people”, co-founder Rishabh Kaul reportedly mentioned. The startup mines social media sites and profiles to drum up the right candidate for the job. The startup claims to be the world’s first predictive outbound hiring solution and has garnered a huge client base from Cisco, Amazon, Adobe, Paypal, Boeing, Accenture, Deloitte to Tesco. The startup’s outbound hiring solution helps companies hire top talent by putting three things at the core of their recruitment strategy — data-driven intelligence, personalization, and automation. “The solution aggregates and analyzes talent data from over 90 social and professional networks and uses data science and predictive analytics to help companies engage talent most likely to move”, Kaul told Analytics India Magazine. 4) Use of Chatbots in initial screening: Today, recruitment firms are leveraging chatbots that perform the twin role of managing candidate relationships in early stages and even performing an initial screening. The use of chatbots is widespread in cases that involve high-volume of recruiting. Case in point – Mya chatbot automates the process from resume screening and decides whether the applicant’s resume should be forwarded or rejected.  Now, even before the advent of chatbots, there were intelligent tracking systems that asked certain pre-screening questions but Mya performs a more dynamic candidate interview with contextual questions generated in real-time. Her knowledge base grows as she has different experiences with candidates, notes the website. Key advantages of AI in first-stage hiring Roots out human bias and human fatigue element Provides improved speed and accuracy and lowers operational costs AI also has the potential to reduce discrimination from sourcing and screening process If the bias isn’t inadvertently built into the system, AI can help match right candidate with the right job A 2015 paper from Harvard Business School pointed out that algorithms do a better job of hiring qualified employees However, the decision shouldn’t be left completely on algorithms and a human element should be involved in making the final decision Gradually, AI & machine learning techniques will phase out the tasks of initial screening, ad-buying and placement of ads that took up the bulk of time of recruiters For high-volume, low-skilled positions, AI can be deployed to automate this cycle of job advertising, sourcing, screening and final hiring of candidates Automating ad-placing with Programmatic Advertising As projected, programmatic recruitment advertising is fast emerging as a technology which is used by large enterprises to automate a part of recruitment process, which is ripe for automation. In the early hiring cycle, even big teams of recruiters are unable to effectively analyze all the places their ads might appear, find the best placement for the ad, when to place or remove ads and how to adjust them mid- campaign. Programmatic technology has seeped its way into recruitment ads and helps recruiters optimize their ad campaigns by improving outcomes. Hence, ad-placing and management is another area which is ripe for automation.","excerpt":"AI’s onward march has threatened jobs across industries, but the impact doesn’t stop there. Artificial Intelligence (AI) has impacted almost all sectors, and that includes hiring and recruitment space as well. Predictive analytics is going to change the way millions of people are hired and assessed. According to a whitepaper, The Robots Are Here- How […]","categories":["IT Services"],"tags":["Predictive AI"],"author_name":"Richa Bhatia","publish_date":"2018-01-31T11:17:57","publication_year":"2018","word_count":1099,"keywords":["data science","machine learning","artificial intelligence","AI","chatbots","R","RAG","Aim","analytics","predictive analytics","Predictive AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","data science","analytics","Aim","RAG","chatbots","predictive analytics","R"],"url":"https:\/\/analyticsindiamag.com\/it-services\/initial-screening-predictive-hiring-look-ai-impacts-recruitment\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":49741,"title":"What Are The &#8216;No Free Lunch Theorems&#8217; In Data Science?","content":"No Free Lunch Theorems (NFLTs): Two well-known theorems bearing the same name: One for supervised machine learning (Wolpert 1996) and one for search\/optimization (Wolpert and Macready 1997). The thing that they share in common is that they state that certain classes of algorithms have no “best” algorithm because on average, they’ll all perform about the same. Mathematically put, the computational cost of finding a solution, averaged over all problems in the class, is the same for any solution method. No solution, therefore, offers a short cut. Today’s menu: InceptionNFL real-world implicationsExample problemImplementationTakeawaysReferences Once Upon A Time The No Free Lunch Theorem (NFLT) is named after the phrase, there ain’t no such thing as a free lunch. It has origins in the mid-nineteenth century onwards, whereby bar and saloon owners would attract drinkers in with free food on the condition that they brought a drink. In the “no free lunch” metaphor, each “restaurant” (problem-solving procedure) has a “menu” associating each “lunch plate” (problem) with a “price” (the performance of the procedure in solving the problem). The menus of restaurants are identical except in one regard – the prices are shuffled from one restaurant to the next. For someone who is as likely to order each plate as any other, the average cost of lunch does not depend on the choice of restaurant. There’s no place where it’s a free lunch. The Whys And Hows In theory, a novice in Data Science and an expert Kaggler, have equal chances of winning the Kaggle competition since the performance of the solution model does not depend on the choice of the algorithm or method.But that doesn’t happen in the real world, right?We know about the great algorithmic weapon, “XGBoost” and how it wins most of the Kaggle competitions that aren’t won with deep learning. Shouldn’t we just use XGBoost all the time? Doesn’t it have upper hand over other algorithms? Or even better, what about the much-hyped, “Master Algorithm” that claims to be one-size-fits-all for the machine learning algorithms? “Never assume the obvious is true.” The question to ask is: “What is the basic set of assumptions that are specific enough to make learning feasible in a reasonable amount of time, general enough to be applicable to a large class kind of machine learning algorithms?” The core observation for understanding theorem is to know that an NFL theorem considers applying two algorithms over all possible problems. In practice, we are probably only interested in a small subset of all possible functions. If an algorithm performs well on a certain class of problems then it necessarily will perform poorly on the set of all remaining problems. Alice and Bob discuss what’s in it for us over lunch. Assume we have a large dataset from which we drew a sample which might allow us to train a machine learning model that predicts values for any of the required predictor columns. Alice: To make a plant predictor algorithm, let’s first list out some major characteristics of plants. Bob: Plants have green-coloured leaves, stems, roots, flowers, fruits, thorns and can’t move. Bob: Well, it could also be a tree. We would require more attributes to ascertain if it’s a plant. Alice: Let’s include some more: life span, average height, absence of trunk and branches, and if it has autotrophic nature. Bob: This is still not an exhaustive list, but these attributes are enough to determine if a given being is a plant or not. Alice: Agreed. We need not map all possible attributes of plants to tell them apart from other beings. Let’s take a look at the sample training set that Alice and Bob gathered. ItemHas trunkHas branchHas flowersHas fruitsAvg heightLifespanIs a plantBanana treeYesYesYesYes12-25 ft5-6 yearsNoTulsi NoNoYesNo30-60 cm1 yearYesRose plantNoNoYesNo5-6 ft35 yearsYesMango treeYesYesYesYes35–40 m300 yearsNoAloe veraNoYesNoNo60–100 cm5-25 yearsYes Bob: Now, we have the training set: different living beings in rows, their different attributes in columns, with the predictor column being Yes or No, if it’s a plant or not, based on the above set of hypotheses. Any classification ML algorithm could predict the results for the test dataset. Isn’t it? Alice: Yes, but according to NFL theorem, the only way to arrive at an optimised solution is just to use the whole mapping with all possible mapping functions. No shortcuts. No compressing the data by ignoring the attributes. So, how does the model work here? Shouldn’t that be applicable to our problem as well? Bob: Yes, we could correctly model a plant predictor with the smaller dataset since we did not assume a uniform distribution over all the problems. Hence, the “No free lunch theorem” does not apply when we don’t follow the assumptions it asks us to make. We can now assess performances of different algorithms, keeping in mind the important trade-offs, for the best algorithm. Time To Get Little Technical Let’s take the training dataset of our plant predictor model in variable trainfeats and test dataset in testfeats. Now, pass this as an argument in different machine learning classifiers namely: Naive bayes, Decision tree, Support Vector Clustering and Stochastic Gradient Descent. We have chosen our trusted NLTK library in Python to get a whole lot of options to choose from for our prediction model. Get the accuracies and compare them. (Assuming data cleaning and other preliminaries are done beforehand) accuracies = [] for x in range(10): random.shuffle(train) cutoff = int(len(train) * 0.01) trainfeats = train[:cutoff] testfeats = train[cutoff:] # Naive bayes NBClassifier = NaiveBayesClassifier.train(trainfeats) # Decision tree DTClassifier = DecisionTreeClassifier.train(trainfeats) # Support Vector Clustering SVCClassifier= SklearnClassifier(LinearSVC(), sparse=False) SVCClassifier.train(trainfeats) # Stochastic Gradient Descent SGDClassifier_classifier = SklearnClassifier(SGDClassifier()) SGDClassifier_classifier.train(trainfeats) # Taking Naive Bayes classifier as our first model for measuring the accuracy accuracy = nltk.classify.util.accuracy(NBClassifier,testfeats) accuracies.append(accuracy) print ((accuracy * 100)) On comparing the results obtained from different ML models, one would find that the pre-defined notions that existed before the start of this experiment, such as one OG model performing better than the other, would be all proven wrong. That is, until and unless we pass our training data on a desired set of algorithms and perform the model evaluation ourselves, we cannot for certain point out one algorithm that would give better results than the rest. Key Takeaways How well the algorithm will do is determined by how aligned the algorithm is with the actual problem at hand. Depending on the problem, it is important to assess the trade-offs between speed, accuracy, and complexity of different models and algorithms and find a model that works best for that particular problem. There are no universal solutions to ML problems.. All ML approaches are equally good if we do not place strong assumptions on the input data. For every ML algorithm, there exists a sample or sample class where it outperforms some other method.","excerpt":"No Free Lunch Theorems (NFLTs): Two well-known theorems bearing the same name: One for supervised machine learning (Wolpert 1996) and one for search\/optimization (Wolpert and Macready 1997). The thing that they share in common is that they state that certain classes of algorithms have no “best” algorithm because on average, they’ll all perform about the […]","categories":["AI Features"],"tags":["AIM Writers Programme","Data Analytics","data mapping","Data Science","naive bayes"],"author_name":"Kshiti Ballal","publish_date":"2019-11-18T18:00:00","publication_year":"2019","word_count":1125,"keywords":["data science","machine learning","AI","NLTK","ML","AIM Writers Programme","naive bayes","RAG","Python","Aim","deep learning","XGBoost","Data Analytics","data mapping","Data Science"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","Aim","XGBoost","NLTK","RAG","Python"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/what-are-the-no-free-lunch-theorems-in-data-science\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10104606,"title":"OpenAI Hires Former X Head to Lead India Operations","content":"OpenAI is planning to expand its footprint in India by roping in Rishi Jaitly, former VP of Elon Musk’s X, as reported by TechCrunch. Jaitly will be working as a senior advisor and will be responsible for guiding the company to navigate India’s AI policy and regulatory landscape. Though the news is not officially announced by OpenAI, he’s helping the team to set up an office in India. Interestingly, the development is said to have taken place soon after Altman’s inaugural trip to the country in June. According to Jaitly’s LinkedIn, his extensive professional journey encompasses diverse roles, including founding the Virginia Tech Institute for Leadership in Technology, where he pioneered the world’s first executive degree in humanities. Currently serving as a distinguished humanities fellow and professor of practice at Virginia Tech, he engages in teaching and leadership at the intersection of humanities and technology. As Principal at Alchmy LLC, he focuses on evangelism and excellence for companies globally. Notably, as the Co-Founder & CEO of Times Bridge, he scaled a venture capital business, bringing global ideas to India. Jaitly also played key roles at Twitter (he was India’s first employee), leading operations and partnerships in Asia, and at Google, where he overturned internet censorship and advocated for an open web across India and South Asia. Back in June, the company made its first international expansion to London, UK. During Altman’s visit to India, he found himself unnecessarily entangled in a controversy after declaring it “hopeless” for Indian companies to compete with their American counterparts in AI, later clarifying that his remark was taken out of context; it referred specifically to the challenge of competing with a $10 million budget. In response, CP Gurnani, Tech Mahindra CEO, revealed that the IT giant is working on Project Indus, an indigenous LLM that would have the ability to speak in many Indic languages, most notably Hindi.","excerpt":"OpenAI is planning to set up office in India and Rishi Jaitly is said to be helping in navigating the country’s policy and regulatory space for same.","categories":["AI News"],"tags":["OpenAI"],"author_name":"Shritama Saha","publish_date":"2023-12-11T14:35:45","publication_year":"2023","word_count":314,"keywords":["Go","API","OpenAI","AI","programming_languages:R","RPA","venture capital","programming_languages:Go","R"],"extracted_tech_keywords":["AI","OpenAI","R","Go","API","RPA","venture capital","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-ropes-in-former-x-head-to-lead-india-operations\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":45746,"title":"6 Trending Jobs In Machine Learning &#038; Data Science To Apply Right Away","content":"Machine Learning and Data Science deals with intelligent algorithms, statistics, mathematics and much more. In this article, we list down 6 trending jobs in machine learning one can apply. 1| ML Engineer II at Microsoft India Location: Hyderabad Responsibilities: The responsibilities include developing highly scalable classifiers and tools leveraging machine learning, data regression and rule-based models, deep learning, create language models from petabytes of text data in different languages, suggest, collect and synthesize requirements and innovate to create next-generation feature sets. The candidate will work as part of the product team to implement algorithms that power user and developer-facing products reaching out to millions of users, adapt standard machine learning methods to best exploit modern parallel environments. Prerequisites: The candidate must have strong background in one or more of Machine Learning, Artificial Intelligence, Pattern Recognition, Natural Language, Deep Learning, DNNs, large scale Data Mining, experience with scripting languages such as Perl, Python, PHP, and shell scripts, experience with recommendation systems, targeting systems, ranking systems or similar systems, experience with any of Hadoop\/Hbase\/Pig or MapReduce\/Bigtable or R\/Matlab\/AzureML or similar technologies. Apply here. 2| Machine Learning Engineer- Lead At JPMorgan Chase & Co Location: Bangalore Responsibilities: The responsibilities for a Machine Learning Engineer – Lead include building common ML capabilities used across Corporate based on machine learning models, automate and streamline existing processes, procedures, and toolsets. The candidate shall also innovate new ways of managing, transforming and validating data, mentor and coach junior team members to build a high-performing team, lead development of one or more of next-generation big data based machine learning frameworks and back-end services, front-end web applications and back-end services that integrate with other products, high-performance processing and analytics applications using Spark and robust analytic data pipelines. Prerequisites: The candidate must have BA\/BS degree or equivalent experience in Computer Science, Information Technology or related disciplines in addition to 10+ years of exceptional hands-on knowledge of robust software design and development. Python is a must know  and prior experience on end to end – design, development, deployment, exposure to either AWS or Hadoop Ecosystem is a must. Apply here. 3| Senior Data Scientist At PayPal Location: Bangalore Responsibilities: The responsibilities include build, evolve, and scale state-of-the-art machine learning system infrastructure powering PayPal’s data and AI Platforms.  The candidate is expected to contribute to the design, development and operations of large-scale infrastructure systems. The Data Scientist will work with other Machine learning \/ deep learning researchers and backend engineers to implement scalable solutions to solve complex problems. Prerequisites: Requirements include strong analytical skills, ability to build quick estimates using back-of-the-envelope analysis, structure (and, if needed, execute) more complex analyses, pull together business cases and forecasts to navigate through multi-dimensional sets of tradeoffs. In addition, experience with Microsoft Excel or statistical software, working knowledge of SQL or other relational database languages, and hands-on experience in data analysis involving large data sets and good communication skills. The candidate must have BS\/BA degree with 7+ years of experience, MS degree with 5+ year of experience and all the above mentioned skills, Apply here. 4| Data Analyst\/ Data Science At Rawmind Location: Gurgaon Prerequisites: The candidate must have BCA, MCA, B.Tech, M.Tech (Passout batch 2016 to 2108), must have good knowledge of SQL, should be able to write SQL Queries and having an understanding of data warehousing concepts like Joins , Keys , Schema, Normalization, etc., good analytical skill, exposure to Tableau, Power BI,ETL Tools, etc. Apply here. 5| Data Analyst At Recruise India Location: Bangalore Responsibilities: This role supports the needs of Forecasting, Analytics and Master Data Management team with a focus on providing advanced descriptive analytics through Tableau. This position is responsible for building dashboards, data collection, data curation and presentations in order to measure the success of business strategies, provide accountability for field sales performance, and enhance overall performance. Specifically, collaborate with Strategic Analytics to provide insightful visualisations as a part of larger project analyses. Prerequisites: The candidate must have Tableau visualisation and calculation skills, LOD, parameters, ability to evaluate business process improvements. Also, Bachelor’s Degree + MBA from a good school required. Apply here. 6| Computer Vision Engineer At KoiReader Technologies Location: Bangalore Responsibilities: The candidate must leverage open-source code and libraries to quickly experiment and build novel solutions, independently think of solutions to complex requirements, and possess exceptional logical skills. For this position, the candidate shall also analyse current products in development, including performance, diagnosis and troubleshooting work with the existing framework and help evolve it by building reusable code and libraries, search and introduce the new software-related technologies, process and tools to the team. Prerequisites: The candidate must have strong experience in Python, building statistical models & Deep Learning, experience in PyTorch, Tensorflow, Keras, OpenCV, SciPy, NumPy, NLP, good understanding of optimizing data processing pipelines, ability to work and thrive in a startup environment, learn rapidly and master diverse web technologies and techniques. Apply here.","excerpt":"Machine Learning and Data Science deals with intelligent algorithms, statistics, mathematics and much more. In this article, we list down 6 trending jobs in machine learning one can apply. 1| ML Engineer II at Microsoft India Location: Hyderabad Responsibilities: The responsibilities include developing highly scalable classifiers and tools leveraging machine learning, data regression and rule-based […]","categories":["AI Hirings"],"tags":["Data Science Jobs","etl hadoop","hadoop etl","Learn Data Science","learning data science","machine learning pattern recognition python","Mapreduce","science jobs"],"author_name":"Ambika Choudhury","publish_date":"2019-09-09T19:03:30","publication_year":"2019","word_count":816,"keywords":["data science","machine learning pattern recognition python","artificial intelligence","machine learning","science jobs","AI","TensorFlow","ML","learning data science","Data Science Jobs","computer vision","Mapreduce","NLP","deep learning","analytics","Learn Data Science","hadoop etl","etl hadoop"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","NLP","computer vision","data science","analytics","TensorFlow"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/6-trending-jobs-in-machine-learning-data-science-to-apply-right-away\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10072161,"title":"How does the Indian Army want to use AI?","content":"Technology already plays a vital role in modern warfare and, with the emergence of AI, it is expected to assume an even bigger role in the future of defence. So far, the Indian Army has undertaken various projects with many more in the pipeline aimed towards the integration of AI. However, not much has materialised in terms of improvement in its capabilities in an actual battle environment. “The scope of artificial intelligence is vast. And if I did say, what we have done till now is very rudimentary or very basic of what the scope or potential of AI in the Indian Army can be”, said Lt. Gen. Shantanu Dayal, Deputy Chief of Army Staff while speaking at the ‘Artificial Intelligence in Defence’ (AIDef) event. AI integration into the Indian Army’s core operations is critical because that is what our adversaries are doing. India shares sensitive borders with China and Pakistan. Tension between India and China was at its peak during the Galwan Valley clash, and the fear of war lurked around for months. Pravin Sawhney in his book, ‘The Last War: How AI Will Shape India’s Final Showdown With China’, said that if India and China were to clash in the near future, India would most likely lose within 10 days. This is because the Chinese military has made significant progress in AI integration compared to India. So far, the Indian Army has used AI for facial recognition, language translation (Mandarin to English), remotely-operated weapon stations, robotic mine detectors, and intrusion detection systems, among others. Now, the primary focus of the Indian Army is to leverage AI in its core operations. War efficiency In a battle environment, the soldiers operate under acute stress and time pressure. There is also the threat from the enemies along with the threat to life. Most of the time, terrains and weather conditions further aggravate the situation. Most of India’s sensitive borders are in the Himalayan stress, from Kashmir in the north to Sikkim in the East. It is at such strategic locations that the Indian Army aims to leverage the potential of AI and empower the infantry to execute their tasks in a much easier and swift manner. (Source: Indian Army) Military intelligence and surveillance Military intelligence, where information is collected and analysed, helps authorities make accurate decisions in a battle environment. The Indian Army aims to use AI and enhance its intelligence gathering and surveillance capabilities to better assist commanders in their decision-making. The information collected could be about the enemy, weaponry, battle capabilities, or any aspect of the battlefield from different sources. “We have got aircraft, we have got satellites, we have got drones, we have got soldiers, we have got numerous kinds of platforms and capabilities, which get us this information. And today, the information is absolute. There’s no dearth of that information. The problem is how to collate that information. How to present it to the commander or the boss there so that you can make use of that information”, said Lt. Gen. Shantanu Dayal. The problem is collecting and analysing information can be a time-consuming process. By the time this information is collated and ready for use, the dynamics of the operations have already changed making the processed information more or less redundant. To resolve this, the Indian Army seeks an AI system that can collect, analyse and present crucial data in a short time frame to assist quick decision making needed in the battlefield. Further, the collected data may not always be good data. Most of the time, the collected data is distorted or falsified. “So, the AI has to be intelligent enough, potent enough to discern through these kinds of enemy activities”, explained Lt. Gen. Shantanu Dayal. Combat Units The Indian Army ranks as the second largest army in the world, with more than 1.4 million personnel. Besides their mammoth strength, the Indian Army has a stunning number of armoured combat units, artillery, missile systems, air defence systems and unmanned aerial vehicles. “This is what the Army fights with. This is what gives the results to the nation. This is what destroys the adversary, and this is what gets the victory for the country”, said Lt. Gen. Shantanu Dayal. He further explained with the example of an artillery gun that there are systems to calculate the trajectory of the shell. These systems take into account factors such as wind speed, temperature, the density of the air, the range and more. (Source: Indian Army) “All these factors go into play, and there are very established and very good systems of calculating all this to ensure the accuracy of the shell. And obviously, because all these factors have to be calculated from certain sources, it takes time”, added Lt. Gen. Shantanu Dayal. However, in most cases, in a battle environment, due to the lack of time, all aspects cannot be factored into the system, and therefore sometimes, it is not so accurate. This is where AI comes in. It can factor in all aspects and provide accurate calculations. Supply chain disruption The supply chain also plays an important role in ensuring the efficiency of the military. To reiterate, let’s take the example of the movement of ammunition, which is stored in different depots in various parts of the country. Ammunition is moved by different modes of transportation and reaches the forward area. Further, it might be moved again by mule or by helicopters, before it reaches the soldiers. “It is a long chain, and each part of the chain is prone to disruption, and therefore, prediction of the disruption, prediction of the capability, prediction of the smoothness or prediction of the robustness of the system is a very important factor to enable the commanders to plan their operations”, Lt. Gen. Shantanu Dayal said. There are existing systems that provide such predictions. However, they are not as robust as the Army would prefer them to be in order to maintain efficiency. They require an AI system that is capable of sustaining the vagaries of battle, enemy disruption, communication breakdown, disruption in transportation and more. “My aim was to expose all of you to the vastness of the requirement of AI in the Indian Army, and we will be accepting all such proposals, all such ideas with open arms”, concluded Lt. Gen. Shantanu Dayal.","excerpt":"An AI system that can collect data, analyse them and present the same to the commander in a very short time frame is one of the key requirements for the Indian Army","categories":["AI Features"],"tags":["Rajnath Singh AI"],"author_name":"Pritam Bordoloi","publish_date":"2022-08-04T15:00:00","publication_year":"2022","word_count":1054,"keywords":["Go","artificial intelligence","programming_languages:R","AI","Rajnath Singh AI","programming_languages:Go","RAG","Aim","ViT","disruption","R"],"extracted_tech_keywords":["AI","artificial intelligence","Aim","RAG","R","Go","ViT","disruption","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-does-the-indian-army-want-to-use-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":31698,"title":"Shine.com’s Face Recognition &#038; Touch ID Is Changing User Experience With Smooth Authentication","content":"One of India’s leading job portal, Shine.com introduced face recognition and touch ID capabilities in its mobile application to improve the user experience significantly. Close on the heels of expanding its technology team, the company has doubled down on machine learning to reimagine the user experience of the app. In a first-of-its-kind, the mobile app received an upgrade around its hardware-based tech with the portal streamlining the user’s login process with Face\/Touch ID, thereby allowing users to validate their identity without the need for passwords. According to Shine.com CTO, Amardeep Vishwakarma, “The job portal has been an industry frontrunner for the past several years, with an unwavering focus on user-centricity when it comes to its offerings. We were amongst the first online jobs portals in India to have launched their own mobile app. The launch of this first-of-its-kind face and touch-based login features on our mobile app, well before any other player in the domain, further reinforces our commitment to simplifying and adding value to our users’ interactions with our online platform.” Users can now log into the app and receive notifications by using Face\/Touch ID, without having to enter a password or going through other means of authentication. The new features also draw on the in-built security mechanism of the OS platform that the mobile app operates on, to safeguard the user’s information. The move underlines the company’s focus on leveraging AI and ML as a key differentiator India’s. Vishwakarma further added how most apps mostly focus on improving the software side of tech, Shine.com mobile app has made advanced upgrades around the hardware side to enable better user experience and is now geared for the latest smartphones in the market, for example iPhone XS and the iPhone XR. The job portal has enabled touch ID-based login on both Android and iOS applications, while the face recognition-based login is available on the Shine.com iOS app on iPhone XS and iPhone XR. The platform plans to soon make the face ID feature available on its Android app as well.","excerpt":"One of India’s leading job portal, Shine.com introduced face recognition and touch ID capabilities in its mobile application to improve the user experience significantly. Close on the heels of expanding its technology team, the company has doubled down on machine learning to reimagine the user experience of the app. In a first-of-its-kind, the mobile […]","categories":["AI News"],"tags":["face recognition online"],"author_name":"Richa Bhatia","publish_date":"2018-12-18T08:05:00","publication_year":"2018","word_count":339,"keywords":["Go","machine learning","programming_languages:R","AI","ML","programming_languages:Go","RAG","face recognition online","R"],"extracted_tech_keywords":["AI","machine learning","ML","RAG","R","Go","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/shine-coms-face-recognition-touch-id-is-changing-user-experience-with-smooth-authentication\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10130021,"title":"AI Could Transform Crime Solving in India—If Only the Govt Stepped Up","content":"AI tools are revolutionising the government’s crime-solving efforts, but their full potential still remains untapped. Staqu Technologies, a Gurugram-based startup, is at the forefront of developing cutting-edge AI technologies to solve real-world problems. It has partnered with several state governments, including Uttar Pradesh and Punjab, to enhance public safety and security. While its primary focus is supporting police and security forces, only 30-35% of its revenue comes from government contracts. A majority of the revenue is generated from the private sector, where it uses video analytics for various security and operational use cases. This includes intrusion detection, fire detection, SOP compliance in manufacturing, retail industry analytics like tracking queue lengths, customer churn, and more. These applications have proven to be more lucrative, even though security remains a critical and growing concern. Every day, the company identifies over 400 criminals. Over the past six or seven years, it has identified more than 30,000 criminals. This represents a significant value proposition for the Indian ecosystem, given the high population density, which makes crime and evasion relatively easier. While it is focused on contributing to society’s safety, generating revenue is also important since the organisation incurs significant costs, particularly on salaries for AI scientists and researchers. How can AI Solve Crime in India? Drug trafficking in Punjab is a significant crime for the police to crack. It’s no longer just a crime; it has become a business for many in the border areas. For instance, in Afghanistan, one kilogram of heroin costs around INR 1 lakh, but by the time it reaches Delhi, it is worth INR 3 crore. This makes it a lucrative business, and thus, it is essential to track the individuals involved. In an exclusive interview with AIM, Atul Rai, the co-founder & CEO at Staqu Technologies, said, “We provide a tool that analyses visual, audio, and textual data for the police. While we do not conduct the analysis ourselves, we offer a platform that the police use for their investigations.” “Our tool allows for unstructured data analysis. Users simply input the data, and the tool processes it without needing extensive configuration. Given our experience working with the police for nearly a decade, we understand the specific needs and requirements,” Rai said. Tools like this require a boost as the NCRB data suggests that nearly 40% of cases registered by Punjab Police in 2022 under the Narcotic Drugs and Psychotropic Substances (NDPS) Act were against those caught with drugs meant for personal consumption. In Punjab, drones have been recovered and drugs seized, making it one of the top states as far as drug seizures are concerned. Working on similar lines, Quadrant Technologies Confidential has developed NarcGuideBot, an AI-powered assistance in Narcotics Investigations. The tool is designed to help inexperienced officers in complex drug enforcement. Further, the tool expedites form-related tasks by simplifying form access, providing accurate filling instructions, and ensuring legal compliance through error-detection systems. Rai further mentioned that the same tool is used in other states for different issues. For example, Uttar Pradesh utilises it for organised crime and gang analysis, while Bihar uses it to address illegal liquor problems. Despite the variations, the core functionality remains the same, providing insights into gang activities and crime patterns. The tool can identify gang members, track their crimes, and analyse their modus operandi, all through automated data extraction and analysis. Next Stop: Karnataka Staqu Technologies is likely to make an official announcement within the next two months regarding its operation in Karnataka. The company is active in the state, with a team of about 15 members based in the Bengaluru office. Currently, it is working in Telangana and Andhra Pradesh’s healthcare sector, monitoring outpatient departments in hospitals and managing primary health centres and community health centres to ensure doctors are available and patients aren’t left waiting. This data is sent to a command centre in Amaravati for analysis. Regarding border security, its presence is established in Jammu & Kashmir, working towards addressing the challenges on the Jammu and China borders. Globally, it is already operating in the Middle East with 20 enterprise clients and has started its journey in the US, recently closing a deal there. Bigger Impact Calls for Bigger Team Staqu Technologies team consists of around 100 people, with about 65-70 of them in engineering and research. The company has also partnered with IIT Delhi. After pursuing his master’s in AI from Manchester, Atul Rai worked with the Graphene Lab that won a Nobel Prize in 2010. He returned to India in 2014 and joined Anurag Saini and Pankaj Sharma in founding the company. In 2023, the company published five research papers, and published two in CVPR this year. It is further committed to developing its own IP and models. “Although we may not be as well-known for securing large funding rounds, our impact in security, especially with tools like G20 and Crime GPT for police forces, has earned us recognition and a strong reputation in the field,” Rai stated.","excerpt":"Over the past six or seven years, Staqu Technologies has identified more than 30,000 criminals.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","India"],"author_name":"Vidyashree Srinivas","publish_date":"2024-07-23T19:00:00","publication_year":"2024","word_count":831,"keywords":["Go","AI","RAG","GPT","Aim","ViT","analytics","edge AI","GAN","R","India","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","analytics","Aim","edge AI","RAG","R","Go","GPT","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-could-transform-crime-solving-in-india-if-only-the-govt-stepped-up\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":7043,"title":"A Heuristic Approach to Predictive Modeling, RFM Analysis","content":"This paper discusses the basic postulates of Recency, Frequency & Monetary (RFM) analysis, a heuristic modeling approach, used in Predictive Analytics, to segment a target market into preferred segments and not so preferred segments. The preferred segments are characterized by their high response rate or high willingness to purchase as opposed to other segments which are not as preferred. The paper also exemplifies these concepts with the help of a case study in the tele- communication sector where a company uses an existing data base to arrive at RFM categorization as well as identifies the profile of customers in the preferred segments. Introduction Predictive modeling, the way it is understood in the Business Analytics context, is a way of predicting consumer behavior by analyzing a database either existing in the company concerned or on a database created with the help of an empirical survey. Essentially, a modeling approach, predictive modeling helps the company to identify profiles of consumers who would be more likely to purchase a product or a service which the company might be offering to a specified and defined target market. Applications of predictive modeling can be seen over different industries and in different managerial functions. For instance, for an entrepreneur offering a new product in a specified target market, predictive modeling can help in understanding the consumer needs and preferences with respect to the attributes defining the product.  For a service oriented company it can help to determine the profile of the most preferred segment and predict the percentage of customers who may actually purchase a new service being offered. For a credit card company or an organization offering loans of any kind, predictive modeling may evolve guidelines as to what kind of consumer profile would merit a preferred treatment and to whom loans may be extended with a softer level of interest. Once the predictive modeling  context is well understood and  the objective in terms of what phenomenon  is to be predicted has been clearly stated, the approach would define measurable variables for each item of the in the situation … the predictor  variables… as well as the variable to be predicted … the dependent  or target variable. Thereafter the predictive modeling uses either: (a)Analytical approach like Logistics regression, Linear Regression Analysis, Factor and Cluster Analysis, Conjoint analysis, or (b)Heuristic approach… like RFM analysis, or (c) Data mining approach, which combines Heuristic and Statistical approach such as Classification Trees. This paper proposes to discuss the basic concepts of the Heuristic Approach of RFM Analysis and provide an example of RFM Analysis applied on the database of a company operating in the telecommunication field in India. Predictive Modeling and RFM analysis. In strategic decision making companies often strive to determine who are the most valuable customers whom they would give special privileges to, invest to build up long term relations with, say in a CRM scenario, or target offers for mail orders, catalogue buying or any kind of direct marketing initiatives. The objective, in most of such situations, is to find out who the most likely buyers are, who makes purchases most frequently, who spend the most and who have the greater probability of coming back for repurchase.  In many such initiatives, RFM analysis, recency-frequency-monetary analysis, helps identify consumer segments and customer profiles having such characteristics. ‘The fundamental premise underlying RFM analysis is that customers who have purchased recently – , have made more purchases and have   made larger  purchases are more likely to respond to your offering than other customers who have purchased less recently, less often and in smaller amounts.’ [Charlotte Mason, 2003, University of North Carolina]. The analysis helps an organization to focus on a smaller section of the target population which again follows another managerial premise, Pareto Principle that 80 % of the business comes from 20 % of the customers. In the past 30 years, direct mailing marketers for non-profit organizations have used an informal RFM analysis to target their mailings to customers most likely to make donations. The reasoning behind RFM was simple: people who donated once were more likely to donate again. Currently, with the availability of CRM software and the use of e-mail marketing, RFM analysis has become an even more important tool. Using RFM analysis, customers are assigned a ranking number of 1,2,3,4, or 5 (with 5 being highest) for each RFM parameter. The three scores together are referred to as an RFM   composite score. The database is sorted to determine which customers have been the best customers in the past, with a composite score “111″ being ideal. Of course, in some organizations marketers consider 5 to be the most preferred RFM parameter, in which case ‘555’ would be the most preferred customer. (http:\/\/searchdatamanagement.techtarget.com\/sDefinition\/0,290660,sid91_gci751219,00.html) There are many justifications as to why RFM analysis works. Customers who bought most recently from an organization, are more likely to respond to the next promotion than those whose last purchase has been way back in the past. This is a universal marketing phenomenon and has been observed in many industries such as insurance, banks, cataloging, retail, travel, etc. In a similar manner, customers who have purchased frequently are more likely to respond than the less frequent ones. Also customers who are big spenders often exhibit much higher response rates than small spenders. HOW CAN ANALYTICS HELP Analytics is increasingly being used as a tool to solve complex organizational problems – leading to better decisions. These are the decisions which were once taken solely by gut instincts. The success of any company in the Telecom industry currently depends on two broad factors – Ability to add new subscribers (both data and voice) Ability to retain existing subscribers (Since, Mobile Number Portability is now available by all operators) This paper focuses on the second part, which is on the indicators which would help the company minimize the tendency of subscribers to switch from their service to others. One of the major indicators of this tendency is measured by Port-in Port-out ratio (it is also commonly referred to as Churn). The number of subscribers switching to a given provider from others is referred to as Port-in. Port-out indicates the number of subscribers switching to a different provider from the given company. A port-in port-out ratio of less than 1, hence, is good for the company because more subscribers are coming in than out. If the ratio is greater than 1, it is considered bad for the company. There are two ways of making this ratio healthy – increase the number of port-ins or decrease the number of port-outs. In order to do so, the company would need to strategically connect with the individual subscriber base – the better their needs are taken care of, the less likely it is that they will switch to other provider. Keeping the above fact in mind, a telecom provider usually comes up with a number of plans to woo the existing customers. This comes with a catch though. It is almost impossible to roll out tailor made plans for every subscriber – such would be too expensive. At the same time, covering the entire subscriber base with a few plans would not go down well with specific consumers whose needs might be different. One solution might be to come up with a plan, say, ‘Pay-Per-Use’ plan – and in order to do so, a broad survey of consumers is required – most of their usage details are already with the company. It is their preferences which need to be mapped with their eagerness to take up a new plan from the same provider instead of switching to another service provider. This new service was called ‘dataplan’ by the telecommunication company. Thereafter the company wished to use Analytics to help identify, who among the existing subscribers are most willing to take up the new plans. It is with this end in mind, that the data has been collected, data base was constructed and analyzed. The database elements have been discussed in the next section. Authors Raghuveer Kodali holds a Bachelor’s degree in Electrical and Electronic engineering, from DVR & DR and is currently pursuing his MBA from Myra School of Business, Mysore. HS MIC College of Technology. His Internship in Market Research for a product UNIled In Kwality Photonics. He is Interested in Analytics and Market Research. Nitish holds an MBA degree in Marketing & Strategy from MYRA School of Business, Mysore. He has done his Bachelor of Technology in Mechanical Engineering and has about two years of experience in manufacturing domain. He is an automobile enthusiast and likes to travel. Sangitha Ajith is an MBA student in Marketing & Strategy at MYRA School of Business. She holds a degree in Bachelors of Commerce from Mysore University. She spent her summer interning in Brandcomm as marketing analyst. Her interests lie in creative art. Sri Valli holds a PGDM degree in Finance & Analytics from MYRA School of Business, Mysore. She had done Bachelor of Technology in Computer Science and Engineering. She is an avid puzzle solver with a special interest in Rubik’s cube. Ashutosh Kar holds a Bachelors in Computer Science & Engineering from Jaypee University of Information Technology. After working with Data Warehouses and Business Intelligence tools for 6 years in various IT companies like ORACLE & IBM, his interests shifted to finance and mathematics. He is currently pursuing his MBA from MYRA School of Business, Mysore. His academic interests include dynamic optimization and optimal control theory. Mihir Ghosh: Academic Background – PGPX Executive MBA from MYRA School of Business. Graduated from IIT Kharagpur in the field of Dairy and Food Engineering. Experience – 7 Yrs and 4 months professional experience in project sales and marketing in food, pharmaceutical and life science sector.","excerpt":"This paper discusses the basic postulates of Recency, Frequency & Monetary (RFM) analysis, a heuristic modeling approach, used in Predictive Analytics, to segment a target market into preferred segments and not so preferred segments. The preferred segments are characterized by their high response rate or high willingness to purchase as opposed to other segments which […]","categories":["AI Features"],"tags":[],"author_name":"Prof Purba Rao","publish_date":"2015-03-09T05:36:23","publication_year":"2015","word_count":1624,"keywords":["Go","AI","R","ML","Git","RAG","analytics","GAN","predictive analytics","data warehouse"],"extracted_tech_keywords":["AI","ML","analytics","RAG","predictive analytics","R","Go","Git","data warehouse","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-heuristic-approach-to-predictive-modeling-rfm-analysis\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":807,"title":"Arrow brings you the first “Smart Shirt” that connects with your smartphone","content":"As a new development in the area of wearable, one of the largest clothing brands, Arrow, has come up with an innovative wearable shirt. Now go “smart” with your office wear as Arrow enters the future of fashion with its Smart Shirts, which lets you share your business cards, social media profiles, sets a meeting mode, plays your favourite music and much more. This latest wearable in town vouches for an amazing experience both at office and social gatherings, as there’s a lot you could explore with it. Here’s a list of things that Smart Shirt can do for you: Want to share your LinkedIn profile with a colleague? All you need to do is tap their phone on the shirt’s cuff to have your profile open on their smartphone In the same way, it lets you share your Facebook profile in just a tap away Sharing your Business card was never this easy. Tap the phone on shirt’s sleeves to get your e-business card on other’s smartphone Set the meeting mode to stay away from disturbances during important meetings. Customize the meeting mode by setting silent mode, vibrate more or auto mode and carry your meetings uninterrupted Ease your work by activating work mode. It lets you work with low brightness, Wi-Fi and Bluetooth to save on your battery and stay focused on your work without bothering for phone’s battery Let’s you pair easily with your Bluetooth speakers Opens your favourite app with just a tap on the shirt Let’s you listen to all your favourite songs, courtesy- Smart Shirt How it works? We already can’t seem to get enough of what the Smart Shirt has to offer. Let’s understand how it exactly works? The Smart Shirt is enabled with NFC chip in one of its cuffs that ensures a connectivity between revolutionary Smart Shirt and the phone. These shirts come with an app that can be downloaded from the Play Store. Currently it is available only on Android as Apple hasn’t opened APIs for NFC on the iPhone yet. While this wearable doesn’t intend to change your life dramatically, it surely has interesting functionalities that can ease up your work. The first of its kind shirt is available across major Arrow stores in lot of exciting colours and is made of 100% cotton fabric. If you are worried about its washing, the company has taken care of it and the Smart Shirt is good for multiple cycles of wash, iron, wear. Based on the consumer reception, the company intends to add more functionalities and add wearable technology to their other products such as womenswear and athleisure wear. All the men out there, go grab your wearable shirt now!","excerpt":"As a new development in the area of wearable, one of the largest clothing brands, Arrow, has come up with an innovative wearable shirt. Now go “smart” with your office wear as Arrow enters the future of fashion with its Smart Shirts, which lets you share your business cards, social media profiles, sets a meeting […]","categories":["AI News"],"tags":["Wearable India"],"author_name":"Srishti Deoras","publish_date":"2016-12-01T11:37:06","publication_year":"2016","word_count":450,"keywords":["Go","API","programming_languages:R","AI","Wearable India","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","API","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/arrow-brings-first-smart-shirt-connects-smartphone\/","complexity_score":2,"technical_depth":7,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":41504,"title":"Meet Gen, A New AI Programming Language That Thrives On Automation, Flexibility And Speed","content":"The good thing about artificial intelligence is that it brings together a variety of domains like statistics, computer vision and deep learning, among others, under one roof and enables the developers to mutually benefit from them. The growing demand for AI-based applications has also raised the bar for the platforms that are used to build them. For a vast field like AI, the platforms or programming languages, to be specific, need to be flexible as well as agile. But there are a few challenges one may come across in trying to build such platforms: Achieving good performance for heterogeneous probabilistic models that combine black box simulators, deep neural networks, and recursion Providing users with abstractions that simplify the implementation of inference algorithms while being minimally restrictive Existing systems lack the flexibility and efficiency needed for practical use with more challenging models arising in fields like computer vision and robotics. In a paper presented at the Programming Language Design and Implementation conference, a group of researchers at MIT have demonstrated a novel probabilistic-programming system named “Gen”. According to MIT News, the researchers sought to combine the best of all worlds — automation, flexibility, and speed — into one. “If we do that, maybe we can help democratise this much broader collection of modelling and inference algorithms like TensorFlow did for deep learning,” says Vikash K. Mansinghka who is part of the team that developed Gen. Where Does Gen Fit In The authors claim that Gen can be used for AI applications such as computer vision, robotics, and statistics — without having to deal with equations or manually write high-performance code. A short code of Gen can enable the user to infer computer vision tasks like 3D body poses, which are ubiquitous with autonomous systems, human-machine interactions, and augmented reality. Not only that but Gen also contains components that perform graphics rendering, deep-learning, and probability simulations as well. Gen can be used to simplify data analytics by using another Gen program that automatically generates sophisticated statistical models for feature extraction from datasets. As far as the use cases go, Gen has found its niche in the following departments: Intel and MIT have collaborated to develop depth-sense cameras used in augmented reality systems which use Gen. Whereas, MIT Lincoln Laboratory is using Gen in aerial robotics for humanitarian relief and disaster response. Gen is central to an MIT-IBM Watson AI Lab project, along with DARPA ongoing Machine Common Sense project, which aims to model human common sense at the level of an 18-month-old child. What makes GEN Different? Unlike deep learning platforms such as TensorFlow, PyTorch, Theano, Gen programs explicitly factorise modelling and inference. By automating the process of calculating the proposal densities needed for a broad range of advanced Monte Carlo techniques, Gen has given a platform to combine Julia and TensorFlow code. The authors claim that Gen has outperformed existing probabilistic programming languages in solving inference problems including 3D body pose estimation from a single depth image; robust regression; inferring the probable destination of a person or robot traversing its environment, and structure learning for real-world time series data. These performance gains are enabled by Gen’s more flexible inference programming capabilities. Here’s how to install Gen: Installation First, obtain Julia 1.0 or later, available here. The Gen package can be installed with the Julia package manager. From the Julia REPL, type ] to enter the Pkg REPL mode and then run: pkg> add https:\/\/github.com\/probcomp\/Gen Know more about Gen here","excerpt":"The good thing about artificial intelligence is that it brings together a variety of domains like statistics, computer vision and deep learning, among others, under one roof and enables the developers to mutually benefit from them. The growing demand for AI-based applications has also raised the bar for the platforms that are used to build […]","categories":[],"tags":["DARPA","Julia Language","Robotic Process Automation"],"author_name":"Ram Sagar","publish_date":"2019-06-28T11:10:46","publication_year":"2019","word_count":578,"keywords":["artificial intelligence","AI","DARPA","neural network","PyTorch","computer vision","Julia Language","Robotic Process Automation","deep learning","Aim","analytics","TensorFlow","R"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","neural network","computer vision","analytics","Aim","TensorFlow","PyTorch","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-gen-a-new-ai-programming-language-that-thrives-on-automation-flexibility-and-speed\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10087631,"title":"This AR\/VR Startup is Helping DRDO Bolster India’s Defence Capabilities","content":"In an attempt to stay one step ahead, India has been working towards revamping the country’s defence and military operation by incorporating artificial intelligence. The effort is in the right direction as combat units, military intelligence and surveillance, and war efficiency can all be improved by including AI in fundamental operations. Mixed Reality startup AjnaLens (Dimension NXG) is collaborating with Indian defence for a more AI-enabled defence system. AjnaLens launched mixed reality glasses, AjnaXR, along with an immersive learning platform, AjnaVidya, at CES 2023 which has received widespread recognition. Pankaj Raut (chief executive officer), Abhishek Tomar (chief technology officer), and Abhijit Patil, (chief operating officer) founded it in 2014. It bagged pre-Series A funding of Rs 12 crore ($1.6 million) from Lets Ventures Angel Fund and JITO Angel Network, among other entities in February, 2022. It is backed by angel investors like Vijay Shekhar Sharma, the founder of Paytm, Japan Vyas, the managing partner of Roots Ventures, the Maharashtra Defence and Aerospace Venture Fund run by IDBI Capital Markets and Securities, and many more. AIM recently got in touch with Abhijit Patil to know more about it. Team AjnaLens at CES AR,VR To The Rescue Augmented reality (AR) and virtual reality (VR) technology can change the way India’s defence force trains and equips its soldiers. “AR-VR glasses have a wide range of applications. Right from medical, aerospace to education and defence. Defence also has a wide range of XR applications to upgrade soldiers, vehicles and weapons, such as crime scene visualisation, holographic planning, holographic collaboration, enhancing situational awareness, remote assistance, and much more,” said Patil. Besides AjnaXR and AjnaVidya, AjnaLens has also created AjnaESAS, a see-through armour system that uses a 360° camera system fixed on a tank and an ‘Augmented Reality Head Mounted Display’ (AR-HMD) worn by the crew. It offers a 360° field of view and night vision, can be zoomed in well, and has features to increase survivability, mobility, and safety, thereby making it useful for defence applications. Funded by the Ministry of Defence, AjnaLens is working to upgrade India’s battle tanks with these see-through armour systems for higher situational awareness, enabling them to detect potential threats and react more quickly to them. Additionally, the company is collaborating with the Indian Army to augment the abilities of weapon systems and with the Navy and DRDO to offer immersive training via their unique physical and digital learning solution. According to Patil, one of the key uses for AR\/VR technology remains in training and upskilling soldiers. Traditional live training methods can be risky and expensive. AjnaLens’ training module can be used by soldiers to practise in a safe space, as many times as they need to, and in different scenarios. This not only increases the efficiency of training but also saves a significant amount of capital that can be utilised in other areas of defence. Adding to the point, Patil said, “Another use case for AjnaLens’ AR\/VR technology is to enhance the existing weapon systems or vehicles already in use by the defence forces. By incorporating this technology, the defence force can save money by upgrading their existing systems instead of buying new ones. This approach is cost-effective and modular, making it easier for the defence forces to become more advanced and future-ready. Additionally, by developing a Made-in-India product, AjnaLens can be exported globally.” The products are solely made in India but the challenges of poor infrastructure along with lack of vendors and subsidy continues to impede the process. However, Patil acknowledges that the scenario is slowly changing. Besides defence, AjnaLens developed the mining training module to guide natural resources company Vedanta’s employees on how to operate the massive drill machine. Because most of the equipment is operated manually, they created AjnaSparsh (haptic gloves) to make training more immersive and interactions more natural. India’s Goal to Strengthen its Defence with AI The Department of Defence Production has allotted $12.6 million annually for military AI research. Earlier, Rajnath Singh, India’s defence minister, stated that the use of AI can aid in the development of autonomous military devices that can effectively handle large amounts of data, aid in training soldiers, and be very helpful in combat. Read more: Indian Navy’s quest to become an AI-enabled force The Indian army has already used AI for facial recognition, language translation, remotely-operated weapon stations, robotic mine detectors, and intrusion detection systems. The army currently seeks an AI system that can collect, analyse, and present important data in a short time span to assist quick decision making needed during battle. On the other hand, the navy is also using training schools to upskill its personnel in AI and machine learning (ML), while collaborating with educational institutions and companies, such as the Indian Institutes of Technology (IITs), the Defence Research and Development Organisation (DRDO), defence public sector undertakings (DPSUs), Bharat Electronics, and Goa Shipyards. India’s Vice Admiral Rajesh Pendharkar explained that the navy is working towards becoming a more AI-enabled force. In order to improve combat operations, communications, logistics management, maintenance, cybersecurity, and physical security, the Indian Navy is also working towards becoming a more AI-enabled force. With an aim to make grassroot-level impact on people and a focus on secure, cost-effective, and future-ready solutions, AjnaLens is well-positioned to overhaul the way soldiers are trained and equipped—both in India and around the globe. Read more: How does the Indian Army want to use AI?","excerpt":"AjnaLens launched mixed reality glasses, AjnaXR, along with an immersive learning platform, AjnaVidya, at CES 2023","categories":["Deep Tech"],"tags":["ces","paytm"],"author_name":"Shritama Saha","publish_date":"2023-02-18T10:00:00","publication_year":"2023","word_count":898,"keywords":["Go","paytm","API","artificial intelligence","machine learning","ces","AI","ML","Git","Aim","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","R","Go","Git","API","GAN"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/this-ar-vr-startup-is-helping-drdo-bolster-indias-defence-capabilities\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":32650,"title":"5-Step Guide You Should Follow To Start A Career In Artificial Intelligence In 2019","content":"As artificial intelligence (AI) continues to invade different verticals, it is getting obvious that the future of tech is already here. Also, there are talks that AI will soon eliminate human jobs and will take the driving seat. However, it is not completely true — with all the advancements, it will very likely create new types of jobs as well. Form chatbots to robots to voice assistance, AI over the years has proved that it is here to stay, not to fade. So, it is high time for people who are into AI or want to make a career in AI, should start preparing. In this article, we lay down the 5 crucial steps to follow to start a career in AI. 1) Understand the AI Career Landscape AI is a sound career choice for a while now and as the adoption of AI in various verticals continues to grow, the demand for trained professionals to do the jobs created by this growth is also skyrocketing. Even though many AI pundits have prophesied that this technology will wipe out a massive amount of human jobs, there other pundits too who have said this will offers many unique and viable career opportunities. Therefore, if you are an AI enthusiast then be optimistic and prepare for a great career in AI. 2) Popular Job Roles In The Field Of AI Machine Learning Engineer: It is considered to be one of the most sought-after careers jobs in the AI space and to be a machine learning engineer, one must have a strong hand on software skills, have the knowledge of how to apply predictive models and utilise natural language processing while working with huge datasets. Robotic Scientist: A Robotic scientist is someone with significant formal education and builds mechanical devices to perform various tasks — whether it is about machines to go where humans can’t go or robotic hands for microscopic tasks. A robot might automate jobs, but it requires someone who can create robots. To be a robotic scientist, one should at least have a bachelor’s degree related to computer science or engineering. Data Scientist: A data scientist is someone who collects, analyse and interpret data from various sources by using machine learning and predictive analytics to better understand how the business performs and builds AI tools. To be a data scientist, one should have expertise in using Big Data platforms and tools including Hadoop, Pig, Hive, Spark, and MapReduce. Also, one should have a strong hand in programming languages including SQL) Python, Scala, and Perl. Research Scientist: A research scientist is someone who is responsible for designing, undertaking and analysing data from controlled laboratory-based investigations, experiments and trials. Also, to be a research scientist, one should have a strong knowledge of different AI disciplines. Business Intelligence Developer: The demand for Business Intelligence developers has skyrocketed in recent years. BI developer spends a lot of time researching and planning solutions for existing problems within the company. Whether it is about analysing complex data or look for current business, a BI developer plays a vital role in increasing the profitability and efficiency of the organisation. 3) Educational and Knowledge Prerequisites If you’re intrigued by AI and wondering how to get started, then the first and foremost step is to hold a Bachelor’s degree in mathematics and computer science. However, most of the time, a Bachelor’s degree only lands you an entry-level position, but it will at least get you started. So, once you get into the industry and understand how things work, you can plan your future career. Talking about positions entailing supervision, leadership or administrative roles, you need to have a Master’s degree or a Doctoral degree. Before you dive in AI, build up a base in these areas: Computer Science: Coding expertise with popular programming languages such as Python, Java, Julia, Lisp etc. Physics, engineering, and robotics Mathematics: Algebra, calculus, logic and algorithms, probability, statistics Bayesian networking (including neural nets) Cognitive science theory One step ahead: If you’re already a software engineer, it becomes a bit easy for you to enter the AI industry compared to those who are just getting started. Some additional courses on AI (at a brick-and-mortar school or an online school) will win you the brownie points with a job in the AI field. Word to the wise: Also, have great communication skills. Whether you are a newbie or programmers, or someone with some relevant experience,  apart from different skill sets required by different industries working in AI one should also possess great communication skills. 4) Three top-tier online AI resources to learn from Learn with Google AI This platform by Google provides a course that starts from a basic introduction to machine learning to getting started with TensorFlow, to designing and training neural nets. The best thing about this course is that you don’t need to have any prior knowledge to get started; it is designed in such a way that whether you’re just starting with coding or you’re a seasoned machine learning practitioner, you’ll find information and exercises to help you develop your skills and advance your projects. Udacity: Machine Learning Course From Google This course from Google through Udacity is on deep learning, giving us an overview of what deep learning is all about. Aiming at the people who are looking to put machine learning, neural network technology to work as data analysts, data scientists or machine learning engineers. Also, if someone wants to dig deep more, then Udacity has a full-fledged Deep Learning Nanodegree program to have a more hands-on experience. MIT: Deep Learning for Self-Driving Cars With so many courses available on the internet about AI, MIT takes the approach of using one major real-world aspect of AI — self-driving cars. This course (an introduction to the practice of deep learning) by MIT is focused on people who have just started in the field of machine learning. However, according to MIT, this course is also helpful for those who are already working in machine learning or deep learning space. 5) Three Must-Try Job Hunting Platforms It’s a technology-driven age. You don’t have to roam around, knocking doors of the companies to hire you; we have AI, the internet and 4 billion years of evolution. Let’s have a look at some of the best job hunting portals\/platforms that would help you land your next job in the field of AI. Indeed Indeed is the most popular of the job sites today. With a simple and effective interface, the portal makes it easier for the job hunter to search jobs that fit hit well. From company hiring pages to direct job\/hiring posts, Indeed scrape down all. How to search? Just type the relevant job title keyword and location. When you get the job you were looking for, just hit apply.  You can also sign up and create a profile, which makes it easier for you to apply for jobs. LinkedIn Job Search With more than 562 million users in more than 200 countries and territories worldwide, LinkedIn is the world’s largest professional network. And talking about the job search, LinkedIn Job Search is one of the best platforms, delivering great job listings. One can also reach out to the industry power players and have a conversation about hiring. Google For Jobs When it is about searching for anything online, one can never ignore Google. Google’s job search engine (don’t get confused! It is not Google job board.) So, it’s basically the same as the google search engine, but here, google only performs searches for jobs. It displays jobs already posted on different portals. Compared to job searches in different portals, Google job search is convenient as it saves a lot of time. So, if you are looking for a job in the field of AI, then try out Google Job Search first then check out other platforms. Outlook Since its inception, AI has been playing a vital role in the technology space, improving the quality of life across various industries. And talking about what the future holds, it is hard to predict. However, the way AI is evolving, it seems the innovations in the coming years are going to be marvelous and those innovations will be successful only when there are people who are trained and working in the field of AI. If you have a dream of working with amazing technology, then it is high time that you should start paving your path towards a career in artificial intelligence.","excerpt":"As artificial intelligence (AI) continues to invade different verticals, it is getting obvious that the future of tech is already here. Also, there are talks that AI will soon eliminate human jobs and will take the driving seat. However, it is not completely true — with all the advancements, it will very likely create new […]","categories":["AI Trends"],"tags":["AI Jobs","business intelligence career path","career in AI","current leaders in self driving cars"],"author_name":"Harshajit Sarmah","publish_date":"2019-01-02T11:33:27","publication_year":"2019","word_count":1416,"keywords":["artificial intelligence","machine learning","AI","neural network","chatbots","business intelligence career path","AI Jobs","career in AI","Aim","deep learning","analytics","TensorFlow","predictive analytics","current leaders in self driving cars"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","neural network","analytics","Aim","TensorFlow","chatbots","predictive analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/5-step-guide-you-should-follow-to-start-a-career-in-artificial-intelligence-in-2019\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10164791,"title":"The ‘First Commercial Scale’ Diffusion LLM Mercury Offers over 1000 Tokens\/sec on NVIDIA H100","content":"For a long time, there’s been an active discussion about exploring a better architecture for large language models (LLM) besides the transformer. Well, two months into 2025, this California-based startup seems to have a promising solution. Inception Labs, founded by professors from Stanford, the University of California, Los Angeles (UCLA), and Cornell, has introduced Mercury, which the company claims to be the first commercial-scale diffusion large language model. Mercury is ten times faster than current frontier models, according to an independent benchmarking platform, Artificial Analysis, the model’s output speed exceeds 1000 tokens per second on NVIDIA H100 GPUs, a speed previously possible only using custom chips. “Transformers have dominated LLM text generation and generate tokens sequentially. This is a cool attempt to explore diffusion models as an alternative by generating the entire text at the same time using a coarse-to-fine process,” Andrew Ng, founder of DeepLearning.AI, wrote in a post on X. Ng’s last phrase is key to understanding why Inception Labs’ approach seems interesting. Andrej Karpathy, a former researcher at OpenAI, who’s currently leading Eureka Labs, helps us understand this better. In a post on X, he said that LLMs based on transformers are trained autoregressively, meaning predicting words (or tokens) from left to right. However, diffusion is a technique that AI models use to generate images and videos. “Diffusion is different – it doesn’t go left to right, but all at once. You start with noise and gradually denoise into a token stream,” added Karpathy. He also indicated that Mercury has the potential to be different and showcase new possibilities. And as per the company’s testing – it does make a difference in the output speed. In the company’s evaluation across standard coding benchmarks, Mercury surpasses the performance of speed-focused small models like GPT-4o Mini, Gemini 2.0 Flash and Claude 3.5 Haiku. The Mercury Coder Mini model achieved 1109 tokens per second. Source: Artificial Analysis Moreover, the startup also said diffusion models are advantageous in reasoning and structuring their responses because they are not restricted to considering only their previous outputs. Besides, they can continuously refine their output to reduce hallucinations and errors. Thus, diffusion techniques power the models under video generation tools like Sora and Midjourney. The company also took a subtle dig at the techniques used by current reasoning models and their bet on inference time scaling that uses additional compute while generating the output. “Generating long reasoning traces comes at the price of ballooning inference costs and unusable latency. A paradigm shift is needed to make high-quality AI solutions truly accessible,” the company said. Inception Labs has released a preview version of the Mercury Coder, which allows users to test the model’s capabilities. Small models optimised for speed are at threat –  but what about specialised hardware providers like Groq, Cerebras and SambaNova? Are Groq, Cerebras, and SambaNova Under a Threat? It isn’t for no reason that NVIDIA achieved the status of the world’s most valuable company during the age of the AI frenzy. Their GPUs are ubiquitously preferred for training AI models. However, the company’s Achilles heel was providing low latency and high-speed outputs—even Jensen Huang, CEO of NVIDIA, noted this. This opened up the opportunity for companies like Groq, Cerebras, and SambaNova to build hardware dedicated to high-speed outputs. However, Mercury’s speed was only matched before by models hosted on specialised inference platforms—for instance, Mistral’s Le Chat running on Cerebras. Recently, Jonathan Ross, CEO of Groq, said that people will continue to buy NVIDIA GPUs for training, but high-speed inference will necessitate specialised hardware. Does Mercury’s breakthrough suggest a threat to this ecosystem? Moreover, Inception Labs also said that diffusion LLMs are a replacement for all current use cases like RAG, tool use, and agentic workflows. But this isn’t the first time a diffusion model for language has been explored. In 2022, a group of Stanford researchers published research on the same technique but observed that the inference was slow. “Interestingly, the main advantage now [with Mercury] is speed. Impressive to see how far diffusion LMs have come!” said Percy Liang, a Stanford professor comparing Mercury to the older study. Similarly, a group of researchers from China recently published a study on a diffusion language model they built called LLaDA. The researchers said that the 8 billion parameter version of this model offered competitive performance, and their benchmark evaluations revealed better performance in several tests compared to models in its category.","excerpt":"Built by Inception Labs, the model doesn’t require specialised architecture to achieve the speed.","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","diffusion models","transformer based language models"],"author_name":"Supreeth Koundinya","publish_date":"2025-02-27T18:30:15","publication_year":"2025","word_count":738,"keywords":["Gemini 2.0","transformer based language models","TPU","OpenAI","AI","GPT-4o","diffusion models","agentic workflows","Transformers","RAG","Aim","Claude 3.5","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","GPT-4o","OpenAI","Claude 3.5","Gemini 2.0","agentic workflows","Aim","Transformers","RAG","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-first-commercial-scale-diffusion-llm-mercury-offers-over-1000-tokens-sec-on-nvidia-h100\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10023008,"title":"Why ServiceNow Acquired Intellibot","content":"ServiceNow has recently signed an agreement to acquire Hyderabad-based robotic process automation (RPA) company, Intellibot. The financial details of the transaction were not disclosed. “ServiceNow is the platform of platforms for the workflow revolution, offering powerful end to end automation capabilities that allow customers to streamline business decisions and unlock new levels of productivity,” Josh Kahn, SVP of Creator Workflow Products at ServiceNow said. “Our customers represent nearly 80 percent of the Fortune 500, and the vast majority are trying to drive automation across a mix of legacy and modern applications,” Khan said. “With Intellibot, we will extend ServiceNow’s ability to help customers connect systems so they can easily automate workflows and drive productivity,” he added. Last year, ServiceNow bought five companies including Element AI, Loom Systems, Passage AI and Sweagle to help companies create end-to-end automation. Why This Acquisition ServiceNow will embed the Intellibot RPA capabilities in its platform once the deal is closed. According to sources, robotic process automation (RPA) will boost the current automation capabilities of ServiceNow. ServiceNow chose a relatively smaller startup vendor as it is easier to incorporate their RPA platform compared to a larger platform which might need a whole engineering overhaul. Meanwhile, Khan reiterated ServiceNow’s commitment to integrating with third-party RPA platforms already being employed by its customers. The acquisition is in line with a broader commitment to the Indian market. At present, India is the second-largest research and development centre of the digital workflow company. By 2022, ServiceNow is planning to develop two new data centre facilities in India. How Intellibot Helps Intellibot will help extend the core workflow capabilities of ServiceNow by helping customers automate repetitive tasks for intelligent, end to end automation. Intellibot was founded by Raghu “Alekh” Barli, Srikanth Vemulapalli, and Kushang Moorthy in 2015. The digital workflow platform intends to build the capabilities of Intellibot natively into its Now Platform. This will enable the customers to easily integrate with both modern as well as legacy systems to drive productivity and strengthen existing artificial intelligence and machine learning efforts. Srikanth Vemulapalli, CTO and co-founder of Intellibot said, “We are proud to join forces with ServiceNow as it continues to invest in powerful end to end automation capabilities to make the world of work, work better for people. ” Recently, the cloud based digital workflow platform unveiled its latest version of the Now Platform, known as the Now Platform Quebec release. The latest version features expanded native AI capabilities as well as new low code app development tools that help the customers to innovate quickly, realise fast time to value, improve productivity as well as deliver robust experiences. “RPA enhances ServiceNow’s current automation capabilities including low code tools, workflow, playbooks, integrations with over 150 out of the box connectors, machine learning, process mining and predictive analytics,” Kahn said. Wrapping Up The pandemic has catalysed the digital transformation of companies. Low code is key to businesses adapting to a rapidly changing technology environment. Recently, Microsoft announced a new low-code programming language, Power Fx for its Microsoft Power Platform. With the adoption of various features and functionalities, ServiceNow is helping organisations to be more agile. The digital workflow company is planning to complete the acquisition of Intellibot in Q2 2021 and to double its staff in the country within the next three years.","excerpt":"ServiceNow has recently signed an agreement to acquire Hyderabad-based robotic process automation (RPA) company, Intellibot. The financial details of the transaction were not disclosed. “ServiceNow is the platform of platforms for the workflow revolution, offering powerful end to end automation capabilities that allow customers to streamline business decisions and unlock new levels of productivity,” Josh […]","categories":["Global Tech"],"tags":["low code platform"],"author_name":"Ambika Choudhury","publish_date":"2021-03-29T14:00:00","publication_year":"2021","word_count":551,"keywords":["API","machine learning","artificial intelligence","low code platform","AI","R","ML","Git","RAG","analytics","predictive analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","predictive analytics","R","Git","API"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/why-servicenow-acquired-intellibot\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10093722,"title":"Meta&#8217;s Breakthrough Language Model on Par with GPT-4 and Bard in Performance","content":"Researchers for Meta AI, alongside Carnegie Mellon University, University of Southern California and Tel Aviv University, today unveiled LIMA, a 65 billion parameter LLaMa language model fine-tuned with the standard supervised loss on only 1,000 carefully curated prompts and responses, without any reinsurance learning or human preference modelling. Register >> Check out the research paper: LIMA: Less Is More for Alignment Meta’s AI chief Yann LeCun said that this is on par with GPT-4 and Bard in terms of performance. The researchers said that LIMA has shown strong capabilities in learning specific response formats with minimal training examples. They said that it can effectively handle complex queries, ranging from planning travel itineraries to speculating about alternate history. Interestingly, one of the notable aspects of this new language model’s performance is its ability to generalise well to unseen tasks that were not part of its training data. In other words, LIMA can apply its learned knowledge to new and unfamiliar tasks, demonstrating a level of flexibility and adaptability, shared by the researchers. In comparison to human-controlled responses, specifically GPT-4, Bard, and DaVinci-003 (trained with human feedback), the responses generated by LIMA were quite impressive. For instance, in 43% of cases, responses from LIMA were either equivalent to or preferred over GPT-4. In the case of Bard, LIMA was preferred in 58% of cases. In the case of DaVinci-003, the preference rose to almost 65%. It is interesting to note that with only limited instruction tuning data, models like LIMA can generate high-quality output. A few weeks back, Meta also released MEGABYTE, a scalable architecture for modelling long sequences. This new technique has outperformed existing byte-level models across a range of tasks and modalities, allowing large models of sequences of over 1 million tokens.","excerpt":"One of the notable aspects of LIMA’s performance is its ability to generalise well to unseen tasks that were not part of its training data.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","ai chatbots","ChatGPT","Google Bard","Meta AI","OpenAI","Yann LeCun"],"author_name":"Mohit Pandey","publish_date":"2023-05-22T23:25:02","publication_year":"2023","word_count":292,"keywords":["ai chatbots","ChatGPT","Meta AI","Yann LeCun","TPU","OpenAI","AI","Google Bard","Modal","Scala","GPT","llm_models:Llama","llm_models:Bard","R","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","Meta AI","TPU","R","Scala","GPT","Modal","llm_models:GPT","llm_models:Llama","llm_models:Bard"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/metas-breakthrough-language-model-on-par-with-gpt-4-and-bard-in-performance\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10167319,"title":"Meta Releases First Two Multimodal Llama 4 Models, Plans Two Trillion Parameter Model","content":"Meta has announced the release of two new open-weight multimodal models—Llama 4 Scout and Llama 4 Maverick. Both models are now available for download on llama.com and Hugging Face and can be accessed via Meta AI products on WhatsApp, Messenger, Instagram Direct, and the Meta AI website. Llama 4 Scout and Maverick are built on a mixture-of-experts (MoE) architecture, making them Meta’s most advanced models released to date. Llama 4 Scout features 17 billion active parameters and 16 experts, designed to fit within a single H100 GPU. According to Meta, it supports an industry-leading 10 million token context window, enabling complex tasks such as multi-document summarisation and reasoning over large codebases. Meta said, “Scout is our most efficient model ever in its class. It delivers performance that surpasses Llama 3 while being more scalable.” The model achieves better results than competing systems, including Gemma 3, Gemini 2.0 Flash-Lite, and Mistral 3.1 on widely reported benchmarks. Llama 4 Maverick, also a 17 billion active parameter model but with 128 experts, is designed for higher-end use cases. It includes 400 billion total parameters and performs competitively with larger models like DeepSeek V3 on reasoning and coding tasks. Meta said that Maverick exceeds GPT-4o and Gemini 2.0 Flash on several benchmarks. It scored an ELO of 1417 on LMArena in experimental chat settings. Meta chief Mark Zuckerberg described it as the “workhorse,” built for larger-scale tasks. He said it “beats GPT-4o and Gemini Flash 2 on all benchmarks” while remaining “smaller and more efficient than DeepSeek-V3.” “These models represent a step forward in balancing performance and cost,” Meta said. “Maverick can run on a single H100 host or scale to distributed inference, offering developers flexibility.” The models were distilled from Llama 4 Behemoth, a yet-unreleased teacher model that is also a multimodal mixture-of-experts model, with 288B active parameters, 16 experts, and nearly two trillion total parameters. Behemoth is still in training but has already demonstrated top-tier results on STEM benchmarks such as MATH-500 and GPQA Diamond, outperforming GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro. Meta noted that Behemoth will not be released yet, but it played a central role in shaping the smaller models through a process called codistillation. The training involved innovations such as a novel distillation loss function and dynamic data selection strategies. Zuckerberg said the company will next release the Llama 4 reasoning model. He added that details will be shared next month. The company also shared new architectural insights. Both Scout and Maverick use interleaved attention layers without positional embeddings and a technique called inference-time temperature scaling to generalise across longer input sequences. The models were pre-trained on diverse multimodal data, including image and video frame stills, and support multimodal interactions across multiple images and text. In terms of training methodology, Meta introduced a lightweight supervised fine-tuning (SFT) approach followed by online reinforcement learning (RL) and direct preference optimisation (DPO). For Maverick, over 50% of SFT data was filtered out to focus on harder examples, improving the model’s performance in reasoning and conversation. Meta highlighted the strategic importance of openness in its release. “We believe openness drives innovation and benefits everyone,” the company said. Llama 4 Scout and Maverick are being released under open terms, with broader access expected soon through cloud providers and partners. The announcement comes ahead of LlamaCon, scheduled for April 29, where Meta plans to share more about its vision for the future of the Llama platform. “This is just the beginning,” Meta stated. “We’re building models that can reason, understand images, and converse naturally to support the next generation of applications.”","excerpt":"The models are now available for download on llama.com and Hugging Face and can be accessed via Meta AI products on WhatsApp, Messenger, Instagram Direct, and the Meta AI website.","categories":["AI News"],"tags":["Meta AI"],"author_name":"Siddharth Jindal","publish_date":"2025-04-06T01:04:50","publication_year":"2025","word_count":598,"keywords":["Hugging Face","Meta AI","Gemini 2.0","AI","GPT-4o","Llama 4","DeepSeek V3","GPT-4.5","Gemma 3","R"],"extracted_tech_keywords":["AI","GPT-4.5","GPT-4o","Gemini 2.0","Gemma 3","Llama 4","Meta AI","DeepSeek V3","Hugging Face","R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/meta-releases-first-two-multimodal-llama-4-models-plans-two-trillion-parameter-model\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":61288,"title":"Top Seven Virtual Conferences On AI","content":"Online AI conferences have been hosted for years, but amid growing concerns around the Covid-19 pandemic, we are witnessing a rise in such events. Although online events have a few disadvantages over the regular ones, one of the most significant advantages is that it can be accessed from anywhere in the world. This opens it up to many technology enthusiasts who cannot otherwise attend traditional conferences. Analytics India Magazine curated the top online AI conferences from across the world that would help you stay abreast with the latest developments in the data science space, and at the same time, allow you to network and clear doubts. plugin ‘plugin‘ is an AI and data science online conference of Analytics India Magazine, which also hosts the most popular physical conferences such as Machine Learning Developer Summit (MLDS) and Cypher. With an objective of catering to the needs of data science enthusiasts from across the world, Analytics India Magazine will host its first virtual event on 28th and 29th May. The conference will include 50 plus speakers across three tracks to allow attendees learn, interact, and participate in various sessions. Some of the key features of the conference are live Q&A, meeting with speakers, networking, among others. Click here to know more. AWS Innovate Online Conference The AI and ML learning edition conference by AWS is an online event that is designed to inspire and empower AI practitioners to accelerate their innovation by unlocking new possibilities. The free AWS Innovate Online Conference consists of over 20 breakout sessions across six tracks, which is delivered by AWS experts on various AI, ML, and cloud computing concepts. The sessions are planned in a way that practitioners can stream sessions based on their skill level, thereby catering to a diverse range of learners. The videos can be watched on-demand here. IBM Think IBM Think will be held on May 5-6 and will be broadcasted over multiple channels. One can also chat with experts and participate in real-time questions and answers during live-streamed sessions. The event will cover a wide range of topics, ranging from blockchain to cloud and AI. Anyone can get their free passes here to learn from experts who will not only inform them about the latest technologies, but also help them gain insights into best practices in business to increase operational resilience in the highly competitive technology landscape. O’Reilly Strata Data & AI Superstream The two days’ O’Reilly Strata Data & AI Superstream online event was attended by more than 4,600 participants. It included live sessions and interactive tutorials as well. The event was hosted on March 17-18 and shed light on some of the trending topics in the AI and ML landscape. The key topics revolved around ML in production and streaming data. The speakers also demonstrated the best practices for robust ML Ops with new open-source tools, explained distributed training in the cloud for NLP models, and more. Besides, it also had a live coding environment to try out new technologies. NVIDIA GPU Technology Conference (GTC) NVIDIA GPU Technology Conference (GTC) was shifted online due to the proliferation of COVID-19. The GTC digital conference also had instruction-led training from the NVIDIA Deep Learning Institute (DLI), where attendees could chat directly with NVIDIA researchers and engineers in a season called Connect With The Experts. The event included featured speakers from NVIDIA, Stanford University, Pixar, Google Brain, among others. The event was held on March 22-26, whose online recorded videos and podcast can be accessed by registering for free. Apple Worldwide Developers Conference 2020 Apple WWDC is scheduled to happen in June where developers from around the world can participate to get their hands on new Apple-engineered technologies and frameworks that will shape the development in AI, ML, and other Apple platforms. The offline version of the conference used to gain interest from more than 23 million registered developers in more than 155 countries. Apple had also committed to giving $1 million to local San Jose organizations as compensation for the loss due to moving the event online. Apple is yet to release the schedule of the event, but you can stay updated with the latest information about the event here. Open Data Science Conference (ODSC) ODSC will be streamed live from April 13-17 and will focus on demonstrating the latest tools, breakthrough models, and frameworks in the tech landscape. The four-day event will have 210 sessions to help you learn from the leaders in the data science community. The conference will generate 360 hours of content, which will also be available for on-demand viewing, but it depends on the type of pass you have bought.","excerpt":"Online AI conferences have been hosted for years, but amid growing concerns around the Covid-19 pandemic, we are witnessing a rise in such events. Although online events have a few disadvantages over the regular ones, one of the most significant advantages is that it can be accessed from anywhere in the world. This opens it […]","categories":["Deep Tech"],"tags":["blockchain tutorial","online ai conference","PowerBI","virtual AI conference"],"author_name":"Rohit Yadav","publish_date":"2020-04-08T19:17:00","publication_year":"2020","word_count":773,"keywords":["data science","machine learning","AWS","AI","cloud computing","ML","PowerBI","blockchain tutorial","online ai conference","deep learning","NLP","virtual AI conference","analytics","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","NLP","data science","analytics","cloud computing","AWS","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/top-six-virtual-conferences-on-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10044684,"title":"Register For This Full Day Workshop To Master Exploratory Data Analysis","content":"The Association of Data Scientists (ADaSci), the premier global professional body of data science & machine learning professionals, has announced a hands-on Mastering Exploratory Data Analysis workshop on August 28, Saturday. Exploratory Data Analysis (EDA) is a critical process of investigating data sets to identify patterns, discover anomalies and summarise main characteristics using data visualisation techniques. Currently, data analysis is one of the must-have skills for a professional in the field of data science. The Association of Data Scientists, in collaboration with Analytics India Magazine, brings an extensive full-day workshop to help data scientists to learn and implement data visualisation techniques. Register For The Workshop Here The workshop is geared towards anyone who wants to master the concept of data analytic. The attendees will start with learning the basics, such as reading and analysing data and will get a hands-on understanding of data analytics complete with a host of visualisation techniques. The workshop will also cover important analysis methods and touch up on data visualisation using Python. The prerequisites to take part in the workshop includes knowledge on Python programming language, a basic understanding of statistics and familiarity with Jupyter\/Colab Notebook. Attendees should have a Jupyter Notebook \/ Google Colab and a high-speed internet connection. The attendees will also receive a certificate on request. Register For The Workshop Here About The Instructor: Dr Vaibhav Kumar Dr Vaibhav Kumar is currently working as Senior Director with The Association of Data Scientists (ADaSci). He brings a lot of experience in the field of Deep Learning and Artificial Intelligence. With a diverse background in industry and academics, he has led many Research and Development (R&D) activities in the field of AI. With a PhD in Deep Learning, he has published many research papers in reputed journals and conferences and guided many scholars in their research works. Details of the workshop: Date: 28th August 2021 Time: 10 AM to 5 PM (IST) Mode: Online Price: ₹999.00 100% discount on CDS Program 100% discount for ADaSci Membership Register For The Workshop Here","excerpt":"The Association of Data Scientists has announced a hands-on workshop on exploratory data analysis and visualisation techniques.","categories":["Deep Tech"],"tags":["Data Analysis","Data Visualisation","data visualisation python","exploratory data analysis"],"author_name":"AIM Media House","publish_date":"2021-07-28T12:55:54","publication_year":"2021","word_count":337,"keywords":["Data Analysis","data visualisation python","data science","artificial intelligence","machine learning","AI","Data Visualisation","Colab","Python","deep learning","Jupyter","analytics","exploratory data analysis","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","deep learning","data science","analytics","Jupyter","Colab","Python","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/register-for-this-full-day-workshop-to-master-exploratory-data-analysis\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10061571,"title":"A Hands-on Guide to JoyPy for drawing interactive Joyplots","content":"Understanding the distribution of data is very important in data analytics and it can be done easily using different types of visualizations. Also, sometimes, we need to make density plots stacked and partially overlapped for better understanding. JoyPY is a python package that helps us in plotting such visualizations through Joyplot. A Joyplot is a series of histograms, density plots or time series for a number of data segments, all aligned to the same horizontal scale. In this article, we are going to discuss how we can make Joyplots using the JoyPy package. The major points to be discussed in the article are listed below. Table of contents About JoyPYWhat are Joyplots?Joyplots using JoyPy Let’s begin with having a quick introduction to JoyPy. About JoyPy? JoyPy is a low code python package that can help us in visualization based on ridgeline plots. It is mainly designed using Matplotlib and Pandas. To draw the ridgeline plots which they say joyplots, this package takes codes from Pandas kdes plots. This library can be compared to R package ggridges that is also named as ggjoy in its older version. What are joyplots? In simple words, joyplots are density plots but stacking and overlapping make them different from density plots. We mainly use this kind of plot for cross-check distribution of the data. Density plots are very helpful for measuring the changes in the data across one dimension. Stacking and partially overlapping make them more helpful to understand the distribution of the data. We can also call these plots ridgeline plots. The above visualization can be considered as an example of joyplots. Let’s see how we can start with joyplots using JoyPy. Joyplots using JoyPy In the implementation, we will start with very basic joyplots using the iris dataset from scikit learn. Before plotting data we need to install joyPy that can be performed using the following lines of codes. !pip install joypy Output: Now we are ready to draw joyplots using python language. Let’s call the important libraries import joypy import pandas as pd import numpy as np from matplotlib import pyplot as plt from matplotlib import cm from sklearn.datasets import load_iris Let’s call the sklearn provided iris data. iris, y = load_iris(as_frame=True, return_X_y=True) iris.columns = [\"SepalLength\",\"SepalWidth\",\"PetalLength\",\"PetalWidth\"] iris[\"Name\"] = y.replace([0,1,2], ['setosa', 'versicolor', 'virginica']) iris Output: In the above, we can see our dataset. Let’s check the density of our data using JoyPy. %matplotlib inline fig, axes = joypy.joyplot(iris) Output: Here we can see an example of a joyplot or we can also call it a ridgeline plot of iris data. We also know that with the data we have a group of names we can also plot a joyplot using the different groups. For doing this we are just required to pass the name of the variable that has group information in the data. fig, axes = joypy.joyplot(iris, by=\"Name\") Output: Let’s say in any dataset we have the size of the y-axis in a larger size than just by just defining limits we can compress it like following: fig, axes = joypy.joyplot(iris, by=\"Name\", ylim='own') Output: In above visualization, we can see that the subplot is not comparable directly because of overlapping we can adjust it by using the overlap factor. fig, axes = joypy.joyplot(iris, by=\"Name\", overlap=3) Output: We can also check the distribution of the data using the histogram. fig, axes = joypy.joyplot(iris, by=\"Name\", column=\"SepalWidth\", hist=True, overlap=0) Output: Here we got to know about how we can use JoyPy for generating joyplots of the data efficiently.  Of course, we can perform more things in our joyplots. Let’s use some other datasets, for example, we are using the global temperature as our dataset which can be found here. Let’s import and see the details of our dataset. df = pd.read_csv('https:\/\/raw.githubusercontent.com\/leotac\/joypy\/master\/data\/daily_temp.csv',comment=\"%\") df Output: In the data, we can see the anomaly columns that represent the difference between daily values. Using this data we are going to draw a joyplot by grouping the years. labels=[y if y%10==0 else None for y in list(df.Year.unique())] fig, axes = joypy.joyplot(df, by=\"Year\", column=\"Anomaly\", labels=labels, range_style='own', linewidth=1, legend=True, figsize=(6,5), title=\"Global daily temperature 1880-2014\", colormap=cm.autumn_r) Output: Here in the plot, we can see how the daily temperature distribution of our data shifted across time. We can also make it more use grid function to map the plot better. fig, axes = joypy.joyplot(df, by=\"Year\", column=\"Anomaly\", labels=labels, range_style='own', grid=\"y\", linewidth=0, legend=True, figsize=(6,5), fade=True, title=\"Global daily temperature 1880-2014\", colormap=cm.autumn_r) Output: Here we have also provided zero value to the linewidth function. We can also make it faded for a better understanding of the data. Here we have a much clearer view of the temperature distribution. We can also change the background and color of the lines. fig, axes = joypy.joyplot(df,by=\"Year\", column=\"Anomaly\", ylabels=False, xlabels=False, grid=False, fill=False, background='k', linecolor=\"g\", linewidth=1, legend=False, overlap=0.5, figsize=(6,5),kind=\"counts\", bins=80) Output: Now the distribution of the data has been differently plotted than the other plots. Maybe things are not clear but I performed it just to let us know how using a single command we can perform changes in the visualization using the JeoPy package. Final words In this article, we have gone through the usage of the JoyPy package that is similar to the ggjoy package in R. We have performed some of the visualizations and seen how we can change them according to different situations and measurements to make joyplots. References JoyPy – GithubLink for the codes","excerpt":"In simple words, joyplots are density plots but stacking and overlapping make them different from density plots. We mainly use this kind of plot for cross-check distribution of the data. Density plots are very helpful for measuring the changes in the data across one dimension.","categories":["Deep Tech"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning"],"author_name":"Yugesh Verma","publish_date":"2022-02-26T10:00:00","publication_year":"2022","word_count":896,"keywords":["Go","NumPy","TPU","AI","Machine Learning","Git","Python","Matplotlib","analytics","Deep Learning","Data Science","Data Scientist","R","AI (Artificial Intelligence)","Pandas"],"extracted_tech_keywords":["AI","analytics","Pandas","NumPy","Matplotlib","TPU","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/a-hands-on-guide-to-joypy-for-drawing-interactive-joyplots\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10123730,"title":"Transformers Can Now Work Pixel by Pixel, Says Meta AI’s New Study","content":"A latest research by Meta AI and the University of Amsterdam have shown that transformers, a popular neural network architecture, can operate directly on individual pixels of an image without relying on the locality inductive bias present in most modern computer vision models. The study, exploring “Transformers on Individual Pixels,” challenges the long-held belief that locality – the notion that neighboring pixels are more related than distant ones – is a fundamental requirement for vision tasks. Traditionally, computer vision architectures like Convolutional Neural Networks (ConvNets) and Vision Transformers (ViTs) have incorporated locality bias through techniques such as convolutional kernels, pooling operations, and patchification, assuming neighboring pixels are more related. However, researchers introduced Pixel Transformers (PiTs), which treat each pixel as an individual token, removing any assumptions about the 2D grid structure of images. Surprisingly, PiTs achieved highly performant results across various tasks. Following the architecture of Diffusion Transformers (DiTs), PiTs operating on latent token spaces from VQGAN achieved better quality metrics like Fréchet Inception Distance (FID) and Inception Score (IS) than their locality-biased counterparts. Perceiver IO Transformers (PiTs) are computationally expensive due to longer sequences, but they challenge the need for locality bias in vision models. Advances in handling large sequence lengths may make PiTs more practical. The study highlights reducing inductive biases in neural architectures, potentially leading to more versatile and capable systems for diverse vision tasks and data modalities. Image generation using transformers There are different models for image generation, such as Midjourney, Stable Diffusion, and Invoke, whose images can be reimagined with these technologies. Recently Midjourney has released the new feature “Character Reference” claiming to generate consistent characters across multiple reference images. Stability AI announced Stable Diffusion 3, the most capable text-to-image model, featuring significantly enhanced performance in multi-subject prompts, image quality, and spelling abilities.","excerpt":"The study, exploring “Transformers on Individual Pixels,” challenges the long-held belief that locality – the notion that neighboring pixels are more related than distant ones – is a fundamental requirement for vision tasks.","categories":["AI News"],"tags":["Generative Pre-Trained Transformer","images","Meta AI","MidJourney","Transformers","Vision Transformers"],"author_name":"Gopika Raj","publish_date":"2024-06-14T18:05:11","publication_year":"2024","word_count":298,"keywords":["MidJourney","Meta AI","Vision Transformers","AI","neural network","vision transformer","Transformers","computer vision","Aim","stable diffusion","Generative Pre-Trained Transformer","GAN","R","images"],"extracted_tech_keywords":["AI","neural network","computer vision","Meta AI","Aim","Transformers","R","stable diffusion","GAN","vision transformer"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/transformers-can-now-work-pixel-by-pixel-says-meta-ais-new-study\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10111600,"title":"OpenAI Got 9.9 Problems, But Gary Ain’t One","content":"In a recent blog, the AI expert Gary Marcus, who stood by Sam Altman at the US Senate hearing last year, expressed concerns about OpenAI, listing out a bunch of serious challenges that the ChatGPT maker might face this year. “As a scientist and technologist, who has watched the company for a long time, I see a long list of challenges for the company this year,” said Marcus, sharing the 9.9 problems that OpenAI might face this year. Legal Troubles? Marcus acknowledges that OpenAI is the key player when we talk about the future of AI. However, that distinction comes with a lot of criticism and as he puts it, OpenAI would “remain under a microscope henceforth”. When it comes to the NYT lawsuit, it is just the first of the millions that are going to come along, he writes. People have been calling the lawsuit as the one that finally moved OpenAI to respond to a lawsuit for the first time as it provided the proof of ChatGPT being trained on the publisher’s data – a classic case of copyright infringement. Meanwhile, OpenAI has now been making deals with publishers such as AP, Axel Springer, American Journalism Project, and NYU, paying $1 million to $5 million to publications for licensing their articles. It does look like the deal went kaput with NYT since the news publisher had approached Microsoft and OpenAI for a resolution in April, but nothing substantial came out in the end. On the other hand, NYT has also partnered with Google for the same reasons. Now, Altman has said that OpenAI does not need to use NYT data to train models anymore. “Some people want to partner with us, some don’t,” said Altman. The only thing that might take a hit because of this is OpenAI’s profits as it could end up paying for the infringements. The company has accepted that it would be almost impossible for AI to get trained adequately without copyright materials. That means it has to end up paying the amount to either accept the licence material or pay the fine. On the contrary, money may not be that big a deal for OpenAI as it has been going through several rounds of funding. According to latest reports, it is looking to raise funds for a valuation of $100 billion. With Microsoft backing the company, OpenAI may end up not caring much about paying off the allegations, if proved correct. Profit, moat, and technology OpenAI has it all. Though not currently, the company is definitely on the road to being profitable soon. The company was operating at a loss for the longest time since it developed ChatGPT. Now, even though it is spending billions of dollars for training GPT-5, it said in December that it recently topped $1.6 billion annualised revenue because of ChatGPT. When we said that OpenAI is as good as its next model, we meant it. Currently, Google is about to release Gemini Ultra and Meta is on its way to release open source Llama 3, which admittedly is on par with GPT-4, but OpenAI could just release GPT-5, which would overshadow both these models. Interestingly, the company can also decide to open source one of its older models, such as GPT-4, to please the open source community. The concerns regarding hallucinations and expert claims that generative AI might soon reach a plateau hold true for all the offerings around the globe. OpenAI would still be the leader when it comes to building these models, and being the innovative and forth-running company, it would come up with different solutions, in completely different verticals. Though as the generative AI landscape is diversifying and people are different providers to backup plans, the same is the case with any other provider of any other technology. For example, Apple is still the most successful company regardless of hundreds of competitors rising in the market. Though, the tensions will remain At the WEF 2024 conference in Davos, Altman said that he still does not know for certain what Ilya Sutskever is doing at OpenAI, hinting that still may be trouble brewing among the board members and the team in general. The tussle between being an altruistic non-profit organisation or going for-profit is out in the public, which is going to be a concern for the company in the year. And this comes when, as Marcus pointed out, the FTC has developed a strong interest in OpenAI with concerns about privacy and security, adding to the several leak reports. Moreover, the most recent  ‘Inquiry into Generative AI Investments and Partnerships‘, that was issued to Alphabet, Amazon, Anthropic, Microsoft, and OpenAI, is definitely going to push OpenAI to keep all of its investments and secrets in order. But as uncle Marcus concluded, these tensions are not exclusive to OpenAI. Though it might remain in the focus, it might be better off than others. Though the valuations might drop, nothing substantial is going to affect the hottest AI startup. He also forgot that OpenAI is not built by a bunch of 24-year old programmers, but as the recent openings suggest, the company is trying to woo in a bunch of young talent.","excerpt":"Gary Marcus seems to be worried about OpenAI, and well, rightly so.","categories":["Global Tech"],"tags":["AI Jobs","ChatGPT","gary marcus","GPT-4","Sam Altman"],"author_name":"Mohit Pandey","publish_date":"2024-01-30T12:00:00","publication_year":"2024","word_count":865,"keywords":["Anthropic","ChatGPT","Go","Sam Altman","GPT-5","AI","OpenAI","AWS","AI Jobs","GPT-4","gary marcus","Aim","generative AI","R"],"extracted_tech_keywords":["AI","generative AI","GPT-5","ChatGPT","OpenAI","Anthropic","Aim","AWS","R","Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-got-99-problems-but-gary-aint-one\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10047997,"title":"Understanding The Technology Behind Content Moderation System Of Tech Giants","content":"Facebook has often been on the news for content moderation — political, racial and ethical. Recently, the tech giant revealed in its monthly compliance report that its artificial intelligence-based content moderation algorithm had removed over 33 million content pieces between June 16 and July 31 this year, in India itself. In addition, Facebook removed another 2.6 million pieces of content from its photo and sharing social media platform Instagram. Facebook posts these reports since implementing the country’s new Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules, 2021. What’s with the content? Facebook said that the largest share of these takedowns was spam and that the platform’s AI algorithms were able to take down 99.9 per cent of such problematic pieces of content over the said period. In addition, 2.6 million of these content pieces concerned nudity, and another 3.5 million accounted for sexual and violent or graphic activities. On the other hand, Instagram was able to detect 64.6 per cent of content related to bullying and harassment, compared to Facebook’s 42.3 per cent. However, this isn’t the algorithm’s performance to the best of its ability. Leveraging AI for Content Moderation Like Instagram and its parent company Facebook, social media giant Twitter uses machine learning for content moderation. Artificial intelligence helps these tech giants scale the work of human experts. Facebook’s ML algorithm performs the following tasks to take action before a post, or a comment harms people: Reducing the distribution of problematic posts or contentWarnings and more context to content rated by third-party fact-checkers Removing misinformation Last year, Facebook AI announced that it had deployed image matching model SimSearchNet++, an upgraded version of SimSearchNet. The model is trained using self-supervised learning to match variations of the same image with precision. Facebook claims that SimSearch++ improves recall while maintaining accuracy, enhancing its ability to find true instances of misinformation while triggering few false positives. Apparently, it is more effective at grouping collages of misinformation. The algorithm runs on both Facebook and Instagram. Source: Facebook AI Facebook also introduced AI systems to detect new variations of harmful content automatically. These systems rely on technologies including ObjectDNA, which focuses on key objects within an image while ignoring background clutter. The AI model also leverages LASER cross-language sentence-level embedding developed by Facebook AI researchers. Additionally, Facebook collaborates with industry leaders and academic experts to organise an open initiative — Deepfake Detection Challenge (DFDC) to develop new tools to address the challenges of deepfake. Despite these measures, a report by NYU Stern, Facebook continues to make 300,000 content moderation mistakes every day. The Price paid by Moderators Facebook continues to make headlines for its sneaky acts. Most recently, it was fined about $270 million by Irish authorities. The authorities charged Facebook for not being transparent about the collection of data from people. Content moderation is at the heart of Facebook’s business model. It is, therefore, imperative for the company to make sure that its moderators are kept content. In July this year, content moderators wrote an open letter to Facebook demanding change — fair treatment, a safe workspace, and mental health support. T0 add to it, a recent article by the New York Times revealed how insiders at Accenture — Facebook’s largest content moderator, have been questioning the ethics of working for the company. In shifts that last eight hours, thousands of Accenture employees sort through Facebook’s problematic content, including messages and videos about suicide and sexual acts, making sure to stop them from spreading online.Provided powerful algorithms are in place to ‘scale up’ the task of human experts; Facebook wouldn’t be facing the nightmare of constant complaints from moderators. Despite his promises to clean up the social media platform, Mark Zuckerberg hires third-party consulting and staffing firms to remove harmful content that AI cannot. Starting 2012, reportedly, Facebook has hired a minimum of 10 consulting and staffing firms worldwide for content moderation. Interestingly, it pays Accenture $500 million every year to avail its services. Unless the models get better trained to identify the wrong from the right, harmful content and misinformation will continue to spread through social media platforms.","excerpt":"Facebook’s AI-based content moderation algorithm removed over 33 million pieces of content between June 16 and July 31 in India.","categories":["IT Services"],"tags":["Accenture","Facebook","Facebook AI","Facebook AI research","instagram","Mark Zuckerberg"],"author_name":"Debolina Biswas","publish_date":"2021-09-09T10:00:00","publication_year":"2021","word_count":687,"keywords":["Accenture","Go","artificial intelligence","machine learning","Facebook AI","AI","ML","Git","RAG","Mark Zuckerberg","GAN","Aim","Facebook","Facebook AI research","instagram","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/it-services\/understanding-the-technology-behind-content-moderation-system-of-tech-giants\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10167426,"title":"Govt Opens Vinod Dham Semiconductor Centre in Delhi to Boost Talent, R&amp;D","content":"The Vinod Dham Centre of Excellence for Semiconductors and Microelectronics has been inaugurated in Delhi, marking a significant step in developing India’s semiconductor talent. Union electronics and IT minister Ashwini Vaishnaw, along with Vinod Dham, known as the ‘Father of the Pentium Chip’, inaugurated the new building at Delhi Technological University (DTU) on Monday. The centre will support research, development, and talent training efforts nationwide. It will serve students, institutions, and industry partners working on semiconductor design and microelectronics. Source: DTU LinkedIn The initiative is part of the government-backed Chips to Startup (C2S) programme. Over one lakh B.Tech, M.Tech, and PhD students have been onboarded under the scheme. “263 institutions have been supported with world-class Electronic Design Automation (EDA) tools,” Vaishnaw mentioned in a post on X. He also mentioned that the programme has enabled 65 R&D projects across 113 organisations, alongside a dedicated SMART lab established at NIELIT, Calicut. Moreover, the ChipIN Centre has conducted 170 training sessions so far. Vinod Dham Centre of Excellence for Semiconductors & Microelectronics inaugurated in Delhi.Strong focus on talent development for India’s semiconductor ecosystem.✅ Chips to Start-up program: – Over 1 lakh students (B-Tech\/ M-Tech\/ PhD) onboarded– 263 institutions supported… pic.twitter.com\/DOwRht1tPX— Ashwini Vaishnaw (@AshwiniVaishnaw) April 7, 2025 As part of the design and fabrication outcomes, 20 Application-Specific Integrated Circuits (ASICs) have been fabricated at the Semi-Conductor Laboratory (SCL) foundry, and 14 ASICs have been sent for fabrication. Rishu Chaujar, the centre’s director, outlined plans for future phases, including collaborations with international universities and industry partners. It is expected to play a key role in reducing India’s dependence on imported chips and strengthening domestic capability. It aims to drive innovation and knowledge creation in semiconductors and microelectronics, aligning with the India Semiconductor Mission (ISM). Source: DTU LinkedIn Dham praised India’s ISM, which he believes will open new opportunities for the country. He emphasised the importance of institutions working together to achieve national goals like Atmanirbhar Bharat. Vaishnaw also encouraged students to seize opportunities and contribute to India’s semiconductor journey. The C2S program is part of a larger push by the Indian government and aims to train specialised manpower in Very-Large-Scale Integration (VLSI) and Embedded System Design (ESD). As India moves towards launching its first homegrown semiconductor chip, these developments are pivotal in reducing dependency on imports and fostering self-reliance in advanced technology.","excerpt":"“263 institutions have been supported with world-class Electronic Design Automation tools.”","categories":["AI News"],"tags":["CoE","Semiconductor India"],"author_name":"Sanjana Gupta","publish_date":"2025-04-08T14:40:20","publication_year":"2025","word_count":388,"keywords":["Go","CoE","AI","programming_languages:R","innovation","RAG","automation","Aim","Semiconductor India","GAN","R","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","GAN","automation","innovation","startup","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/govt-opens-vinod-dham-semiconductor-centre-in-delhi-to-boost-talent-rd\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":423,"title":"Persistent Systems to provide social media analytics for Satyamev Jayate","content":"Outsourced software development services provider Persistent Systems has announced its association with Star India to provide social media analytics for Aamir Khan’s pioneering television show Satyamev Jayate. This data is being analysed and delivered by Persistent Systems to Satyamev Jayate team at Star India and Aamir Khan Productions in real time to understand the impact the show has on the governmental, social and, most importantly, the individual level, the company said. This is the first time that viewer responses and sentiments are being tracked and analysed on such a large scale in India. After the end of each episode, the discussion around the show trends heavily on Twitter, occupying the top slots in India and even worldwide. Within the first nine weeks of the show, some 900 million impressions were made across web and social channels about Satyamev Jayate. In the past few weeks there has been a wealth of unstructured responses coming in from a variety of sources — Twitter, Facebook, YouTube, website comments, SMS polls, and phone voice messages. This data is being analysed and delivered by Persistent Systems to the Satyamev Jayate team at Star India and Aamir Khan Productions in real time to understand the impact the show has at a governmental, social and individual level. A content filtering, ranking and tagging system (CFRTS) has been custom developed for the data analysis of Satyamev Jayate. Persistent has assembled an array of automated tools to parse the data and a user-interface for several analysts to process messages for deep analytics. The result is a cluster-based analysis along with trend, demographics and sentiment analysis for each message. The data is being portrayed on a series of dashboards that are featured online at savatmevjayate.in as well as on air on the Star Network. Satyamev Jayate aims to shed light on important social issues faced by India, with the aim of bringing about social awareness. Viewers are encouraged to share their opinions, personal experiences and suggestions in various ways — including SMS, IVR and online media. This has created a large volume of data that needs to be analysed instantly and on a continuous basis. Anand Deshpande, chairman and MD of Persistent Systems, said: “We are very excited to be a strategic partner to the creative and unique initiative that is Satyamev Jayate. Persistent has been at the forefront of the Big Data revolution, working on key technologies, building platforms and creating solutions to help enterprises solve Big Data problems but by working with STAR and the Satyamev Jayate team we are leveraging our expertise to contribute towards a social cause.” Lalit Bhagia, VP digital for Star India, added: “We are very keen to understand and capture the chronology of change that is being affected through Satyamev Jayate. Hence we were looking for a partner who can not only mine the relevant data, but also analyse and present it in a meaningful way. Persistent Systems has managed to do this very well and the wealth of data and information makes for an interesting read for any student of social change.”","excerpt":"Outsourced software development services provider Persistent Systems has announced its association with Star India to provide social media analytics for Aamir Khan’s pioneering television show Satyamev Jayate. This data is being analysed and delivered by Persistent Systems to Satyamev Jayate team at Star India and Aamir Khan Productions in real time to understand the impact […]","categories":["AI News"],"tags":[],"author_name":"Дарья","publish_date":"2012-07-10T16:36:00","publication_year":"2012","word_count":510,"keywords":["Go","AI","sentiment analysis","Git","RAG","Ray","Aim","CuPy","analytics","R"],"extracted_tech_keywords":["AI","analytics","Aim","Ray","CuPy","RAG","sentiment analysis","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/persistent-systems-to-provide-social-media-analytics-for-satyamev-jayate\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":36754,"title":"10 Technological Predictions Made In Movies That Are A Reality Now","content":"AI has for long been posited as a technology that has been designed to enslave human beings. Well, much of these ideas and distorted imagery stems from the role of movies and TV series has shaping our ideologies and the fact that singularity as a genre, continue to fascinate movie makers and directors across the world. In this article, we take a look at some of the predictions that were described in movies about AI and other technologies that have become freakishly now become a reality. Star Wars (1977): The movie saga is one of the longest running and has amassed a mass fan following for its depiction of technology and even for its background score. One of the most unforgettable snippets from the movie is that of Princess Leia Organ 3D pleading for help to Obi-Wan Kenobi in a 3D hologram. Other aspects from the movie like bionic limbs and magnetic levitation are few other technological advancements which have taken shape 2001: A Space Odyssey (1968): The film by Stanley Kubrick was based on Arthur C. Clarke’s short story The Sentinel and the screenplay was written by him too.  The movie was very well received by science enthusiasts for its marvellous depiction of technology, like spaceflight, special effects, and ambiguous imagery. The movie revolves around the voyage to Jupiter with the sentient computer HAL 9000, an artificial general intelligence device, who is the main antagonist of the movie. Star Trek: The Original Series (1969): The movie is all time favourite of science-buff all across the world and has managed to create a cult following for its series among its fans. The original series that was aired 1966 showcased some of the coolest technology, which needless to say have come true in the next 50 years time. In fact, Martin Cooper the creator of cell phones, got the inspiration to create the device after seeing a scene from the movie where Captain Kirk uses a flip phone like wireless device to communicate. In the age of NLP and AI-powered by solutions, instant translation to the language that we prefer is never a challenge. In Star Trek too, when Captain Kirk is moving from one place in the galaxy to other, unknown and unheard language was never a problem as he had the universal translator, which could help him translate the language instantly. Apart from the conversational AI, the movie also depicts portable video devices similar to that of the present day tablets, where Dave Bowman, the protagonist watches BBC news in his internet-connected portable screen. In addition to this, the movie also shows a video phone through which Heywood Floyd, another character in the movie, makes a phone call to his Daughter. Blade Runner (1982): The movie is set in a dystopian future of 2019 where synthetic humans are bio-engineered by a powerful corporation to work in off-world colonies. Though the movie received a mixed response from the audience, it showcased some of the futuristic which has already become a reality in 2019. this includes video calling, though it is from phone booths. Back to the future(1985): If there is a fun and memorable science fiction movie that deals with time travel, then without a doubt it would be this movie.  The 1985 movie oddly depicted technologies of varying kinds which would become a reality in the 21st century. From Marty’s auto-lacing sneakers in the movie to the 3D film obsession. A number of things including video-conferencing and wall-mounted widescreen television was mentioned in the movie. Total recall(1990): The movie is set in 2084, where Arnold Schwarzenegger’s character Douglas Quaid goes on a memory trip to Mars as a secret agent. In the 1990 big-budget movie, Douglas gets into a self-driving car called Johnny cab, with a robot taxi driver showing him around the town Demolition man (1993): movie talks about two men who are Cryogenically frozen in 1996 are reanimated in 2032. The scariest prediction made in this movie is that of virtuality porn, which sadly is already a thing of the present. Other than the  VR gear, the movie extensively showcases tablets computers, video conferencing and voice-activated appliances The Matrix (1999): Virtual Reality forms the underlying theme of the movie and it talks about a time frame when artificial intelligence has taken over the human race, resulting in a burnt-out planet as the result of the war between man and machine. Neo, the protagonist of the movie realises that he along with other humans are placed in an advanced make-to-belief VR stimulator. Minority Report (2002): The movie based on the short story of the same name is set in 2054 in Washington and Northern Virginia in the United States. The 2002 movie directed by Steven Spielberg has Tom Cruise in the lead, where he is working for a specialised police department which aims to prevent murders from committing the crime with the help of mutated humans, called PreCogs who have the ability to read the future.The classic has toyed with some of the most ubiquitous technological developments such as autonomous cars, facial and optical recognition to name a few.  At the onset of social media and increased online activities, the targeted advertisement is something that follows us anywhere when we browse the internet. In the movie as well, Cruises character comes across advertisements addressing him as he moves about the city Her (2013):  2013 romantic science fiction drama is eerie similar to the lives that we lead now. The movie tells the story of a man who falls in love with a female voice assistant, whose closest cousin in the real world will be Alexa or Siri. Apart from the depiction of this strange man and machine relationship. The movie makes references to ear-bud phones, holographic video games and even digitally transcribed “Handwritten” letter-generating service.","excerpt":"AI has for long been posited as a technology that has been designed to enslave human beings. Well, much of these ideas and distorted imagery stems from the role of movies and TV series has shaping our ideologies and the fact that singularity as a genre, continue to fascinate movie makers and directors across the […]","categories":["AI Trends"],"tags":["ML","Movies"],"author_name":"Akshaya Asokan","publish_date":"2019-03-22T12:57:53","publication_year":"2019","word_count":967,"keywords":["Movies","Go","API","artificial intelligence","AI","ML","Git","NLP","Aim","ViT","GAN","R"],"extracted_tech_keywords":["AI","artificial intelligence","NLP","Aim","R","Go","Git","API","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-technological-predictions-made-in-movies-that-are-a-reality-now\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10104081,"title":"Runway Partners with Getty to Build Video Generation Model for Enterprises","content":"Runway is partnering with Getty Images to launch a new video model for enterprise customers, the company said in a blog post. This model will combine the power of Runway with Getty Images’ world-class fully licensed creative content library, providing a new way to bring ideas and stories to life through video in enterprise ready and safe ways. Today we're thrilled to announce that we are partnering with @GettyImages to build enterprise ready AI tools that will allow companies to create high-quality, customized video content.Read more: https:\/\/t.co\/LjTUcVvfxt— Runway (@runwayml) December 4, 2023 The Runway <> Getty Images Model (RGM) is designed to serve as a foundational model, allowing companies to construct their custom models for video content generation. Runway’s enterprise clients will have the flexibility to fine-tune RGM using their exclusive datasets, empowering entities across various industries such as Hollywood studios, advertising, media, and broadcasting. This initiative aims to amplify creative capabilities, introducing innovative content workflows and facilitating the creation of tailored experiences aligned with enterprises’ styles, brand identities, and unique audience preferences. “Runway’s collaboration with Getty Images takes our mission to empower creators with a new generation of AI tools to a new level of creative control and customization,” said Runway CEO and co-founder, Cristóbal Valenzuela. “We’re excited to work with Runway to help enterprises further creativity and exploration with AI in responsible ways,” said Grant Farhall, chief product officer at Getty Images. The application of RGM for custom model training is particularly pertinent to companies holding proprietary datasets. For those interested in delving deeper into this innovative collaboration and exploring its possibilities, further details and inquiries can be directed through the provided link. The team-up also follows Runway’s recent improvements to its web-based AI tool, Gen-2, designed for creating videos.","excerpt":"Runway enterprise users can refine RGM with their proprietary datasets, benefiting various industries like Hollywood, advertising, media, and broadcasting.","categories":["AI News"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-12-05T11:07:36","publication_year":"2023","word_count":292,"keywords":["programming_languages:R","AI","ML","Aim","ViT","R"],"extracted_tech_keywords":["AI","ML","Aim","R","ViT","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/runway-partners-with-getty-images-to-build-video-generation-model-for-enterprises\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10129457,"title":"AWS Launches AuditLLM, a Multiprobe Approach Tool for LLMs","content":"AWS has recently unveiled a tool named AuditLLM. By announcing their demo paper at CIKM 2024, they introduce a novel tool for auditing LLMs using a multi-probe approach. This solution is designed to streamline the auditing process for LLMs by focusing on activity-based auditing. “AuditLLM is set to redefine the way auditing is performed in the digital age,” said Aman Chadha, researcher from Stanford University & AWS Leadership Member. Authored by Maryam Amirizaniani, Elias Martin, Tanya Roosta,  Chirag Shah, and Chadha this is a collaboration with the University of Washington, where the primary objective of AuditLLM is to provide a comprehensive audit trail of all activities performed by LLMs. The tool leverages advanced technology to meticulously scrutinise the activities of LLMs, ensuring transparency, accuracy, and compliance with established standards. Despite its sophisticated functionality, AuditLLM is easily navigable, allowing auditors to perform tasks without extensive technical knowledge. As LLMs gain prominence across sectors, effective auditing tools like AuditLLM are increasingly crucial. With its launch, it addresses this need, setting a new benchmark for digital-era auditing. Meet the AI Expert Building Indic LLMs with IITs AIM got in touch with Chadha, who was working on building a medical large language model for India on top of Sarvam AI’s OpenHathi, and released a research paper around it. Given the amount of Indic languages speakers all over the world, Chadha expressed his happiness that models like Bharat GPT, Sarvam AI and Kissan AI are coming up. “But there’s nothing on the healthcare or the medical side,” he added, saying that he has been tracking all the recent announcements.","excerpt":"AuditLLM is set to provide a comprehensive audit trail of all activities performed by LLMs.","categories":["AI News"],"tags":[],"author_name":"Tarunya S","publish_date":"2024-07-18T12:27:30","publication_year":"2024","word_count":264,"keywords":["Go","AWS","AI","ML","Git","RAG","GPT","Aim","ViT","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","AWS","R","Go","Git","GPT","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/aws-launches-auditllm-a-multiprobe-approach-tool-for-llms\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10048999,"title":"React 18 Announced: Key Features Expected","content":"Recently, React’s official website announced a new update to be launched for React, the React 18. Although with no specific release date announced and being available only in alpha, React 18’s latest release additions seem very promising. ReactJS is an open-source library that runs on Javascript and has been very popular amongst software developers in the past few years. It has also been considered a great resource for startups and business leaders. Designed by Facebook, it gained popularity owing to its ability to help create UI rich web apps that are fast and efficient, developed with minimal coding. The simplicity and flexibility of React are what connected with its users. Conglomerates such as PayPal, Uber, and Instagram use React to solve their user interface problems. Here’s a roundup of the latest additions and changes to be expected in React 18: Automatic Batching Batching occurs when React groups multiple updates together into a single render state for achieving better computational performance. This also prevented the components from rendering “Half-Finished” states where only one state variable was updated before, causing several bugs at times. However, React would not batch the updates every time and rather perform two independent batches. React 18 will be added with a performance improvement update, where it will automatically batch the updates, irrespective of origin, for both the application and the library code. The updates inside of timeouts, promises, or even native event handlers will be batched the same way as the updates inside of React events. This will add an out-of-the-box improvement to the rendering time and even better performance. With this issue addressed in React 18, it makes the batching process more efficient and consistent. Click here to see a demo of how the new batching technique works. function App() { const [count, setCount] = useState(0); const [flag, setFlag] = useState(false); function handleClick() { fetchSomething().then(() => { \/\/ React 18 and later DOES batch these: setCount(c => c + 1); setFlag(f => !f); \/\/ React will only re-render once at the end (that's batching!) }); } return ( <div> <button onClick={handleClick}>Next<\/button> <h1 style={{ color: flag ? \"blue\" : \"black\" }}>{count}<\/h1> <\/div> ); } } Code Source: React’s Github Added Feature: startTransition for Concurrent Rendering During interactions with the user interface for small actions like clicking a button or typing an input, page freezing may occur, disrupting the workflow. In React 18, a new API known as startTransition has been introduced that helps keep the app responsive even during large screen updates. The API substantially improves the user interactions by marking specific movements as “transitions ”. This allows for React to be informed about which updates are important and which are not. Transitions here are interrupted by urgent updates, and the prior irrelevant transitions are dismissed. This enables the UI to ignore secondary updates that might be slower in nature. startTransition moves the updates to the background, consisting of either complex processing or slowed data fetching due to network connectivity. You can further understand this through a real-world example from the link here. New Suspense SSR and Selective Hydration Server-side rendering, also known as SSR, is a component that lets you generate HTML from React components directly on the server and share the HTML with users. Users can see a preview of the page through SSR even before the javascript bundle presents loads and runs. But sometimes, the javascript on the backend takes a long time to get processed, and the time taken is known as Hydrating Time. React 18 will include architectural improvements to the React SSR’s performance. The new update will allow Streaming HTML directly on the server, i.e., the server sends pieces of components as they get rendered using another component known as Suspense, which decides which parts of the application might take longer to load and what shall be rendered directly. Using a selective hydration method, components that are wrapped with Suspense will not block hydration anymore. Every ready component will start hydrating once the browser gets both its content and javascript code. Checkout a real world demo to this update using the link here. Improvements to Root API Root API in React is a pointer for the top-level data structures on the application that React uses to track a render tree. Two different root APIs will be deployed when using the new React 18  version, the Legacy root API and ReactDOM.createRoot. The Legacy root API will run a legacy mode root API, trigger warnings when the API is deprecated, and suggest moving it to the new root API. The new root API known as ReactDOM.createRoot will add all the improvements to the application and allow concurrent mode features. Old root API : import React from 'react'; import ReactDOM from 'react-dom'; import App from 'App'; ReactDOM.render(<App \/>, document.getElementById('root')); New root API : import ReactDOM from 'react-dom'; import App from 'App'; const root = ReactDOM.createRoot(document.getElementById('root')); root.render(<App \/>); One of the biggest changes of the new root API will be removing the hydrate method, and a prop to the top-level component present will be passed. Easy Upgrade As concurrency in React 18 is subjective, with minimal to no changes in the already created application codes, one can upgrade to React 18 version easily. Users will be able to convert several apps to React 18 within a single afternoon. The company has tested this by deploying several concurrent feature components at Facebook, and most components reportedly work without making additional changes. . Summing Up With many minor features yet to be announced, the new React 18 update looks impactful. The new React has opened up new possibilities, ones that were previously impossible to have in a Javascript framework. All these changes, once implemented, will surely be sources of inspiration for other available frameworks in the market. React 18 has been professionally crafted and looks in line with the best practices one can avail. To know more about how to install the alpha version for React 18 click the link here.","excerpt":"React’s official website announced a new update to be launched for React, the React 18. Although with no specific release date announced and being available only in alpha, React 18’s latest release additions seem very promising.","categories":["AI Trends"],"tags":["updates","Web Applications","Web Development"],"author_name":"Victor Dey","publish_date":"2021-09-20T17:59:17","publication_year":"2021","word_count":994,"keywords":["API","AI","ML","JavaScript","Web Development","Git","ViT","updates","Web Applications","GitHub","R","Java","startup"],"extracted_tech_keywords":["AI","ML","R","JavaScript","Java","Git","GitHub","API","ViT","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/react-18-announced-key-features-expected\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10101702,"title":"Google Will Make Pixel 8 in India","content":"At Google for India 2023 event in New Delhi, Alphabet Inc’s Google has declared India as a priority market and revealed its intention to manufacture the flagship Pixel line-up of phones in the country. Google’s Head of Devices & Services, Rick Osterloh, announced the company’s plan to commence the production of the Pixel 8 in India, with the devices set to hit the market in 2024, as part of the Make in India initiative. In May, India’s Technology Minister, Ashwini Vaishnaw, held discussions with Google’s CEO Sundar Pichai at the company’s headquarters in Mountain View, California. Their conversation primarily revolved around Prime Minister Narendra Modi’s initiative to promote local manufacturing and the government-supported drive for technological advancements. A recent report by Counterpoint Research highlighted India’s emergence as the second-largest manufacturing hub for mobile phones, owing to substantial investments from original equipment manufacturers, original design manufacturers, and companies specialising in components and parts. According to Counterpoint Research, India is expected to export approximately 22% of its total assembled mobile phones in 2023. Government and industry data also revealed that the ‘Make in India’ initiative received a significant boost, with mobile phone exports valued at over $5.5 billion (or more than Rs 45,000 crore) in the April-August period of the ongoing fiscal year (FY24). Furthermore, it is anticipated that India will surpass Rs 1,20,000 crore in mobile phone exports during the current fiscal year, with Apple holding a dominant market share of over 50% in FY24. This decision by Google follows the footsteps of Apple, which has utilised a similar approach to bolster its network of suppliers in India. Apple’s participation in this program has led to a significant increase in iPhone production, surpassing $7 billion during the fiscal year ending in March 2023. Apple has successfully harnessed the benefits of local manufacturing and supplier networks in India. This move by Google aligns with the broader trend of major tech companies leveraging India’s manufacturing capabilities to expand their operations.","excerpt":"This decision by Google follows the footsteps of Apple that also started assembling iPhone 15 in India.","categories":["AI News"],"tags":["Google Pixel"],"author_name":"Mohit Pandey","publish_date":"2023-10-19T13:20:05","publication_year":"2023","word_count":327,"keywords":["Go","programming_languages:R","AI","RPA","programming_languages:Go","RAG","Google Pixel","R"],"extracted_tech_keywords":["AI","RAG","R","Go","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-will-make-pixel-8-in-india\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":61256,"title":"Top Valuable DataSets For COVID-19 Researchers","content":"There is an increasing urgency to maintain reliable data assets around COVID-19 because of the speed at which developments are unfolding. This has made it challenging for the medical research community to keep up. These freely available datasets are offered to the global research community to produce new insights as the world continues its fight against COVID-19. Here, we look at what these data assets are, and where they can be located: Visual Dashboard Dataset This is the data repository for the Coronavirus Visual Dashboard, managed by the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE). Multiple organizations have extensively used it to track the geographic spread of the viral epidemic. The dataset is also supported by ESRI Living Atlas Team and the Johns Hopkins University Applied Physics Lab (JHU APL). Research Articles Dataset In response to the COVID-19 pandemic, the Allen Institute for AI, White House and a group of top research groups have developed the COVID-19 Open Research Dataset (CORD-19). CORD-19 comprises over 47,000 scholarly articles, including over 36,000 with full text about COVID-19, SARS-CoV-2, and associated coronaviruses. The CORD-19 dataset serves as the most comprehensive machine-readable coronavirus literature compilation ready for data mining at the moment. The Allen Institute produced this dataset for AI in cooperation with the Microsoft Research, Georgetown University’s Center for Security and Emerging Technology, Chan Zuckerberg Initiative, and National Institutes of Health, under collaboration with White House Office of Science and Technology Policy in the US. The World Health Organization (WHO) has also been gathering the latest scientific verdicts and knowledge on COVID-19, and is organizing it in a database. WHO updates the database daily from the exploration of bibliographic databases, manual searches of the table of contents of associated scientific journals, and the addition of other relevant scientific articles. The entries in the database are not fixed, and additional research is supplemented daily. Scan Images Dataset The British Society of Thoracic Imaging (BSTI), in connection with Cimar UK’s Imaging Cloud Technology (cimar.co.uk), produced and deployed an anonymized and encrypted web portal to submit and refer images of patients from confirmed COVID-19 cases. From these, BSTI hopes to give an imaging database of established UK patient examples for reference and teaching. The intention is to quickly disseminate clinical and diagnostic information to frontline healthcare workers in the UK. Lan Dao, Joseph Paul Cohen and Paul Morrison from the University of Montreal have also created a database of COVID-19 reported incidents with chest X-ray or CT scans and images. The database contains images from publications and has been released publicly in this GitHub repo. The researchers say the goal is to use these images to develop AI-based approaches to predict and understand the infection better. Twitter Data The repository comprises an ongoing compilation of tweet IDs connected with the novel coronavirus COVID-19 (SARS-CoV-2), which began on January 28, 2020. Emily Chen from the University of Southern California used Twitter’s search API to find old Tweets from the preceding seven days, leading to the first tweets in the dataset dating back to January 22, 2020. Twitter’s streaming API was leveraged to follow particularized accounts and also collect real-time tweets that discussed specific keywords. To comply with Twitter’s Terms of Service, the dataset is only publicly released with the Tweet IDs of the collected Tweets for non-commercial research use. Genome Sequences Data Laboratories around the world are generating and sharing an increasing number of hCoV-19 genome sequences, clinical and epidemiological data associated with the novel coronavirus through GISAID. The genome sequences of hCoV-19 are essential to produce and assess diagnostic tests, to track and trace the ongoing outbreak, and to recognize possible intervention choices. The GISAID initiative supports the global sharing of all influenza virus sequences, and associated clinical and epidemiological data linked with human viruses to help researchers.","excerpt":"There is an increasing urgency to maintain reliable data assets around COVID-19 because of the speed at which developments are unfolding. This has made it challenging for the medical research community to keep up. These freely available datasets are offered to the global research community to produce new insights as the world continues its fight […]","categories":["AI Trends"],"tags":["Covid Dataset","covid-19","covid19 data"],"author_name":"Vishal Chawla","publish_date":"2020-04-08T16:00:00","publication_year":"2020","word_count":636,"keywords":["Go","API","Covid Dataset","covid-19","AI","programming_languages:R","Git","covid19 data","RAG","Ray","GAN","GitHub","R"],"extracted_tech_keywords":["AI","Ray","RAG","R","Go","Git","GitHub","API","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-valuable-datasets-for-covid-19-researchers\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10117736,"title":"Top 6 Devin AI Alternatives for Developer to Automate Codings in 2024","content":"Devin, the world’s first AI engineer by Cognition Labs, took the internet by storm with its ability to write code from scratch, fix bugs, and deploy solutions, aiming to automate aspects of the software development process. However, Devin is not the only AI autonomous agent available Here is the list of top Devin alternatives. Top Devin AI Alternatives for Coding Automation in 2024 NameDeveloped byPriceDevika AIMufeed VHFreeReplit Code RepairAmjad MasadFreeSWE AgentJohn YangFreeOpenDevincommunity-driven projectOpen SourceMetaGPTPicoOpen SourceChatDevcommunity-driven projectOpen Source 1. Devika Devika is an open-source AI software engineer created by Mufeed VH (Hamzakutty), the founder of Lyminal and Stition.AI.  It is  capable of understanding human instructions, breaking them down into tasks, conducting research, and autonomously writing code to achieve set objectives. Devika aims to be a competitive open-source alternative to Devin by Cognition AI. It  utilises LLMS, planning and reasoning algorithms, and web browsing abilities to intelligently develop software. One of Devika’s key strengths lies in its ability to function as an AI pair programmer, reducing the need for extensive human intervention in complex coding tasks. Devika simplifies software development processes, whether it’s creating new features, debugging code, or developing entire projects from scratch, thereby enhancing efficiency. The main difference between Devin and Devika, apart from the latter being open source, is that Mufeed used Claude 3 instead of GPT-4 for Devika. 2. Replit Code Repair Replit’s Code Repair is a low-latency code repair AI agent. It utilises LLMs trained on a massive dataset of code examples and their corresponding fixes. This allows the LLM to analyse your code and identify potential errors or inefficiencies. Replit took a 7B Code LLM and fine tuned it into a tool that mimics the behavior of LSP Code Actions. The special ingredient is in the training data — a careful mixture of real-world errors (collected on Replit) combined with synthetically-generated code fixes. Replit’s approach involves using Operational Transformations (OTs) and session events to create a dataset of (code, diagnostic) pairs. They synthesise diffs using large pretrained code models and fine-tune them for code repair tasks. Explore: Latest News & Stories About Replit 3. SWE Agent SWE Agent is also an open-source alternative to Devin, much like Devika, developed by a team led by John Yang, Carlos E. Jimenez, and Alexander Wettig at Princeton University. It turns language models like GPT-4 into software engineering agents that can fix bugs and issues in actual GitHub repositories. On the full SWE-bench test set, SWE-Agent resolves 12.29% of issues. The key to SWE-Agent’s success lies in its innovative Agent-Computer Interface (ACI), which streamlines the interaction between the language model and the code repository. Unlike traditional approaches, SWE-Agent’s ACI simplifies commands and feedback formats, making it easier for the model to navigate, edit, and execute code files within the repository. Developers can easily set it up using Docker and Miniconda, following straightforward installation and configuration steps outlined in the project’s documentation. 4. OpenDevin OpenDevin is an open-source project aiming to mimic Devin, an AI software engineer. Similar to Devin, OpenDevin aspires to handle various aspects of software development, potentially including, Code Generation, Debugging, and Deployment Automation The alpha version is available for testing, showcasing its ability to handle complex tasks and collaborate with users. The project is focusing on key milestones creating a user-friendly interface with chat and command features, building a stable backend for commands, improving the agent’s capabilities, and setting up an evaluation pipeline. Explore: Latest News & Stories About Devin 5. MetaGPT MetaGPT is a multi-agent framework that in itself acts as a virtual software company.  It takes a one-line requirement and outputs user stories, competitive analysis, requirements, data structures, APIs, and documents. MetaGPT includes product managers, architects, project managers, and engineers, following carefully crafted Standard Operating Procedures (SOPs). Explore: Latest News & Stories About Meta 6. ChatDev Similar to MetaGPT, ChatDev stands as a virtual software company that operates through various intelligent agents holding different roles, including Chief Executive Officer , Chief Product Officer , Chief Technology Officer , programmer , reviewer , tester , art designer. These agents form a multi-agent organisational structure and are united by a mission to “revolutionise the digital world through programming.”They collaborate through specialised functional workshops at ChatDev, engaging in activities like design, coding, testing, and documentation.","excerpt":"Devin, the world’s first AI engineer by Cognition Labs, took the internet by storm with its ability to write code from scratch, fix bugs, and deploy solutions.","categories":["AI Trends"],"tags":["Devika AI","Devin","Scott Wu","Top Trend"],"author_name":"Siddharth Jindal","publish_date":"2024-04-08T11:44:22","publication_year":"2024","word_count":707,"keywords":["Top Trend","Go","TPU","AI","Devin","ML","MetaGPT","docker","R","Git","Aim","Scott Wu","GitHub","Devika AI"],"extracted_tech_keywords":["AI","ML","MetaGPT","Aim","docker","TPU","R","Go","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-6-devin-alternatives-to-automate-your-coding-tasks\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10045544,"title":"Moving Beyond Transformers: Microsoft Enhances Bing Search Results With MEB","content":"Microsoft has recently introduced ‘Make Every feature Binary’ (MEB) to improve its search engine Bing. MEB is a large scale parse model that goes beyond pure semantics and reflects a more nuanced relationship between search queries and documents. To make the search more accurate and dynamic, MEB harnesses the power of large data and accepts an input feature space with over 200 billion binary features. DNN & Transformers for Bing The Bing search stack depends on natural language models to improve the core search algorithm’s understanding of user search intent and related web pages. Deep learning computer vision techniques are used to enhance the discoverability of billions of images even when text descriptions or summary metadata does not accompany the queries. Machine learning-based models are used to retrieve captions within the larger text body that answer specific questions. The introduction of Transformers was a game changer in natural language understanding. Unlike DNN architectures that processed words individually and sequentially, Transformers could understand the context and the relationship between each word and all other words around it in a sentence. From April 2019, Bing incorporated large Transformer models to deliver high quality improvements. How does MEB improve search performance Transformer-based deep learning models have been favoured due to their advanced understanding of semantic relationships. While these models have shown great promise, it still fails at capturing nuanced understanding of individual facts. Enter MEB. MEB model has 135 billion parameters which help it map single facts to features to gain a more nuanced understanding. It is also trained with more than 500 billion query\/document pairs from three years of Bing searches. This gives MEB the capability to memorise facts represented by the binary features while reliably learning from a vast amount of data continuously. Microsoft’s team used heuristics for each Bing search impression to determine whether the users were satisfied with the results. The ‘satisfactory’ documents were labelled as positive samples. Other documents for the same impression were labelled as negative samples. For each query-document pair, the features were extracted from the query text, the URL of the document, title, and the body text. These binary features are then fed to the sparse neural network model. It helps in minimising the cross-entropy loss between the model’s predicted click probability and the actual click label. Feature design and large-scale training are key to the MEB model. Traditional numeric features only care about the matching count query and document. On the other hand, MEB features are very specific and are defined on the N-gram level relationship between the query and the document. All the features are designed as binary features to cover manually crafted numeric features easily. These features are directly extracted from raw text, allowing MEB to perform end-to-end optimisation in one path. The current production model uses three major features: Query and document N-gram pair featuresOne-hot encoding of bucketised numeric featuresOne-hot encoding of categorical features Advantages MEB is currently running in production for all Bing searches in all regions and languages, making it the largest universal model at Microsoft. Compared to Transformer-based deep learning models like GPT-3, the MEB model can even learn hidden intents between query and document. It can also identify negative relationships between words or phrases to reveal what users might not want to see for a query. With the introduction of MEB in Bing, Microsoft has following advantages: A 2 percent increase in the clickthrough rate (CTR) on the top search results1 percent reduction in manual query reformulationOver 1.5 percent reduction in the number of clicks on pagination (need to click on next page button) The MEB model consists of a binary feature input layer, a feature embedding layer, a pooling layer, and two dense layers. Generated from 49 feature groups, the input layers contain 9 billion features. Each of the binary features is encoded into a 15-dimension embedding vector. After per-group sum-pooling and concatenation, the vector is passed through the dense layers for producing a click probability estimation. “If you are using DNNs to power your business, we recommend experimenting with large sparse neural networks to complement those models. This is especially true if you have a large historical stream of user interactions and can easily construct simple binary features,” the team said in a blog.","excerpt":"Microsoft’s team used heuristics for each Bing search impression to determine whether the users were satisfied with the results.","categories":["Global Tech"],"tags":["Deep Neural Networks","Generative Pre-Trained Transformer","list of computer languages and their uses","Microsoft","Search engine","Transformers"],"author_name":"Shraddha Goled","publish_date":"2021-08-09T13:00:00","publication_year":"2021","word_count":709,"keywords":["Go","machine learning","Search engine","AI","neural network","Transformers","computer vision","GPT","deep learning","Deep Neural Networks","Generative Pre-Trained Transformer","list of computer languages and their uses","R","Microsoft","llm_models:GPT"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","computer vision","Transformers","R","Go","GPT","llm_models:GPT"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/moving-beyond-transformers-microsoft-enhances-bing-search-results-with-meb\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10066598,"title":"Why are feature stores such a buzz these days?","content":"“A few years ago, we noticed a pattern: teams were getting overburdened with increasing costs of maintaining their feature preparation pipelines,” says LinkedIn in a blog post while talking about its feature store, Feathr. Feature stores are actually quite a recent concept that helps to create better machine learning pipelines. A major chunk of a data scientist’s time goes into wrangling data and preparing it for analysis instead of building models. Feature stores can be of great help to solve this issue. Uber introduced Michelangelo Palette, the first feature store, in 2017. Databricks recently announced that it has made the Databricks feature store generally available. LinkedIn also open-sourced Feathr a month ago. All major tech firms have their feature stores, like Amazon SageMaker feature store, Vertex AI feature store (Google), ML Lakes Salesforce, and Overton (Apple). What is a feature store? A feature store enables the discovery, documentation and reuse of features. It is a feature computation and storage service that enables features to be registered, discovered, and used for ML pipelines and by online applications for model inferencing. Feature stores stockpile feature data and offer low latency access to features for online applications. It also ensures consistent feature computations across the batch and serving APIs. Most feature stores are proprietary till now Recently, LinkedIn announced that it is open-sourcing its feature store Feathr. But most of the feature stores that exist today are still proprietary. Other popular open-source feature stores include Feast and Hopsworks. Feast was developed jointly by GO-JEK and Google Cloud for teams to store and discover features for use in machine learning projects. Hopsworks was started as a collaborative project between KTH University, RISE, and Logical Clocks. The feature store is a data management system for managing machine learning features, including the feature engineering code and the feature data. Why do we need them Priyanka Vergadi, developer advocate at Google, explains in a video the need for feature stores and the ML challenges that it can solve. She adds, “Most of the time spent by data scientists goes into wrangling data, more specifically, in feature engineering, which is transforming raw data into high-quality input signals for ML models. But this process is often inefficient and brittle.” There are many ML feature challenges, as she points out: Hard to use and reuse across different steps of the ML workflow and across projects, which results in duplication of efforts.It is hard to serve in production reliably with lower latency. There is an inadvertent skew in feature values between training and serving usually, which causes your model quality to degrade over time. Better performance and reusability are key attractions Due to the sheer volume of data and algorithms big companies handle these days, feature stores are a requisite for them. Hence, we are seeing more and more feature stores coming up. Smoother and faster deployment – A data scientist’s primary job is to focus on building models that can help achieve business needs for an organisation. But, often, they are burdened with dealing with data engineering configurations. A feature store provides a consistent feature set, enabling a smoother deployment process. The biggest advantage of a feature story is its reusability aspect. The feature store keeps metadata in addition to the actual features. Due to this, data scientists can figure out the features that performed well on existing models. Data scientists can also share features with their team members, which encourages better collaboration and no duplication. Feature stores help standardise feature definitions. If there is no common abstraction for features, there is no uniform way to name features across models, no common type system for features, and no standard way to deploy and serve features in production.","excerpt":"Uber introduced Michelangelo Palette, the first feature store, in 2017.","categories":["AI Features"],"tags":["linkedin","Open Source"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-05-09T16:00:00","publication_year":"2022","word_count":617,"keywords":["Feast","Go","Amazon SageMaker","machine learning","Open Source","AI","ML","RAG","Hopsworks","linkedin","R","Databricks"],"extracted_tech_keywords":["AI","machine learning","ML","Amazon SageMaker","Feast","Hopsworks","RAG","Databricks","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-are-feature-stores-such-a-buzz-these-days\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049796,"title":"Unlocking Documental Intelligence Holds The Key For Enhanced Customer Experience","content":"A large volume of information travels through papers in an organisation, and knowing the structure of documents enables the extraction of relevant and useful data. Documents can be in a variety of forms and types, including native PDFs, web pages and scanned images. They also have a variety of templates, making document processing and interpretation a chore. Financial papers, in particular, such as audited reports, bank statements, financial reports, exchange filings, and so on, are critical documents that must be assessed for a variety of compliance and risk-related applications, including underwriting, risk rating, and more. At the DLDC 2021 organised by the Association of Data Scientists (ADaSci) — Rahul Ghosh, VP of AI Research and Services at American Express AI Labs, spoke on AI-Powered Document Intelligence for Enterprises. Additionally, Rahul also gave a sneak peek into the R&D efforts at American Express and demonstrated how Document AI-enabled products could drive innovation and efficiency at scale. To start with, document intelligence is nothing but the way that allows us to tap into the opportunities offered by unstructured document data and unlock the potential for faster and more informed decisions, increase operational efficiency, data governance, integrity and compliance, and enhance customer experience. Next, Rahul presented a general document AI stack for an enterprise. The stack is shown below. This was followed by a simple explanation of the different types of documents, which includes: Form type documents are short documents. Typically, they are like one to two, or less than five pages, and have a very well-defined structure and a layout present —for example, invoices or bank statements. Use cases can be understanding invoice spend patterns, getting cash flow insights from bank statements, etc.Verbose documents are longer and have a lot more information put together in a single place. For example, a construction contract agreement – where you have data in the form of text, images, tables and more. Marketing creatives can review use cases before the campaign launch, highlight key clauses from documents, etc. Talking about the challenges in information extraction from verbose documents, Rahul discussed a research paper for the case. A research paper consists of certain text, tables, titles, etc. Moreover, tables come with different types of cell formats, so, as the paper changes, the content will change completely. Hence, a simple rule-based or template-based approach will fail to extract table types that vary across different documents. There are few extraction challenges attached in form type documents, which includes: Form type documents also have diverse templates.Rule-based approaches can’t handle unseen templates and are difficult to manage.NLP-based approaches assign tags to each portion of the text, while CNN-based approaches can capture irrespective of variations in templates.Both NLP\/CNN approaches have limitations for the cases where information is embedded in the spatial arrangement of the layout, not the text itself. A typical extraction pipeline for extracting information out of documents is shown below. Rahul said, “RCNN is an approach for extracting information from images and other things that are there. Although the approach was good, it was computationally expensive and had a longer training time. To remove this bottleneck, Microsoft research came up with a fast RCNN. With the fast RCNN, the problem was that at the testing time, you needed to explicitly tell where the objects were located. Then, somebody had to manually feed that information. That itself is a bottleneck.” On top of faster RCNN, there are some improvement areas that can be taken into account. The addition of iterative refinement helps in identifying the location of objects. Even after this, there are more complicated document use cases where even with iterative refinement, one may not be able to capture all the details accurately. The results obtained by the American Express team are shown below. Documents are the key source of unstructured data for an enterprise. AI-powered document intelligence can understand the structure of a document and extract contents, leading to significant process efficiency. In addition, deep learning approaches borrowed from image processing, computer vision and other related disciplines can significantly outperform naive rule-based approaches.","excerpt":"Deep learning approaches borrowed from image processing, computer vision and other related disciplines can significantly outperform naive rule-based approaches.","categories":["AI Features"],"tags":["American Express","Data Analysis","structured and unstructured data","unstructured data"],"author_name":"kumar Gandharv","publish_date":"2021-09-27T14:00:00","publication_year":"2021","word_count":674,"keywords":["American Express","Data Analysis","Go","AI","structured and unstructured data","computer vision","document AI","NLP","deep learning","data governance","GAN","CNN","R","unstructured data"],"extracted_tech_keywords":["AI","deep learning","NLP","computer vision","document AI","R","Go","data governance","GAN","CNN"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/unlocking-documental-intelligence-holds-the-key-for-enhanced-customer-experience\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10022754,"title":"ISRO Makes A Quantum Communication Breakthrough: India Joins An Elite List Of Nations","content":"On Monday, Indian Space Research Organisation (ISRO) announced that it has successfully demonstrated free-space Quantum Communication over a distance of 300 m. According to ISRO, this is a major milestone for unconditionally secured satellite data communication using quantum technologies. With this breakthrough, India joins a handful of other nations such as the US, the UK, Canada, China and Japan who have made significant contributions in the field of quantum communication. Register for Webinar Series: AWS ML Fridays ISRO demonstrated their quantum capabilities by live video conferencing using quantum-key-encrypted signals. The free-space quantum key distribution(QKD) was demonstrated at Space Applications Centre (SAC), Ahmedabad, between two line-of-sight buildings within the campus. The experiment was performed at night to ensure that there is no interference of the direct sunlight. Most of the mission critical technologies used in the experiments were built by ISRO in-house, which includes the NAVIC receiver for time synchronization between the transmitter and receiver modules, and gimbal mechanism systems instead of bulky large-aperture telescopes for optical alignment. Image credits: ISRO Quantum Key Distribution allows to generate and distribute a secret key which can be used to encrypt or decrypt a message. Thanks to quantum phenomena like quantum entanglement, the presence of a third party between a sender and receiver can be identified hence securing the data transmission. “The Quantum Key Distribution (QKD) technology underpins Quantum Communication technology that ensures unconditional data security by virtue of the principles of quantum mechanics, which is not possible with the conventional encryption systems. The conventional cryptosystems used for data-encryption rely on the complexity of mathematical algorithms, whereas the security offered by quantum communication is based on the laws of Physics. Therefore, quantum cryptography is considered as ‘future-proof’, since no future advancements in the computational power can break quantum-cryptosystems,” said ISRO in their statement. The success of this experiment will pave the way for a highly secured satellite based quantum communication as ISRO gears up to establish contact between two Indian ground stations as part of their next experiment. Also Read: Raman Research Institute Achieves Breakthrough In Quantum Communication","excerpt":"On Monday, Indian Space Research Organisation (ISRO) announced that it has successfully demonstrated free-space Quantum Communication over a distance of 300 m. According to ISRO, this is a major milestone for unconditionally secured satellite data communication using quantum technologies. With this breakthrough, India joins a handful of other nations such as the US, the UK, […]","categories":["AI News"],"tags":["quantum key distribution"],"author_name":"Ram Sagar","publish_date":"2021-03-23T11:26:03","publication_year":"2021","word_count":344,"keywords":["quantum key distribution","Go","AWS","AI","cloud_platforms:AWS","programming_languages:R","ML","programming_languages:Go","GAN","R"],"extracted_tech_keywords":["AI","ML","AWS","R","Go","GAN","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/isro-quantum-communication-breakthrough-india-qkd\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":22056,"title":"The Analytics Leap to Workforce Engagement","content":"Gartner recently announced its inaugural Magic Quadrant for Workforce Engagement Management (WEM). It was an important indicator of the evolution from workforce optimization (WFO) technologies, an already mature market, to technologies that help drive employee engagement. Though gamification, mobile scheduling apps and robotic process automation (RPA) tools all contribute to helping employees, especially Millennials, feel more connected to their jobs, there are also analytics-driven solutions that are playing a major role in helping companies engage employees in their quest to improve the customer experience. Here are three key areas where businesses are shaping up their workforce optimization strategies with speech, text and performance analytics to listen to the voice of the employee and make the leap successfully from WFO to WEM. 1. Speech and text analytics pinpoint key issues Solutions are available today that can analyze speech and text interactions between employees and customers and provide valuable data for focused, relevant process improvement. The tools mine these interactions and produce user-friendly reports on key conversation topics, words most frequently used and problems most frequently discussed. It’s a new window into customer concerns and delights—and employee responses to them. BPOs and other companies in India and around the world are embracing a holistic Voice of the Customer (VoC) solution that includes speech analytics, text analytics and customer feedback solutions. In response, they not only optimize each channel, but they also better understand the customer journey and take corrective action for converting customer detractors to promoters. They have achieved significant improvement in customer satisfaction while reducing costs and improving sales conversion rates. Where does employee engagement fit in here? The key is having tools that bring the most important nuggets from speech and text analytics data to the employee’s fingertips. As customer inquiries grow in complexity, most employees can’t memorize all the information necessary to respond effectively. For example, your speech and text analytics applications might identify that customers are asking most frequently about one line of service or product. Employee engagement tools can then populate agent desktops with an easy-to-access, intuitive repository of information to help them answer the questions customers are asking about that service or product. Employee engagement solutions today provide a knowledge management capability that brings these valuable nuggets to agent desktops. They play an important role in simplifying the delivery of information agents need to be good at their jobs. Aberdeen Group’s Omer Minkara provides insight here in 2017 research: Minkara reports that 51 percent of companies use at least eight channels to interact with customers, contact center agents typically browse three different apps to find relevant customer insights, and optimizing the agent desktop results in a 74 percent improvement in customer satisfaction rates. These metrics are a testament to the value of making it easy for employees to engage relevantly with customers, as well as to the impact those employees can have on the customer experience. Individual case studies tell the story, as well. Bank of Montreal (BMO) uses speech analytics to identify the top five processes per line of business that drive the most repeat calls, as well as customers’ top digital challenges. Given the fact that the company has 12 million customers in North America and fields tens of thousands of calls annually, the data is golden and is used by the bank to drive positive change. “We didn’t have a clear line of sight as to why customers were calling us and what their experience was and what our associates’ experience was,” said one of BMO’s CX executives. 2. Performance management analytics chart a plan for growth Contact centers have traditionally adopted the use of quality monitoring in combination with call recording to evaluate a small portion of employee-customer interactions for use of the proper greeting, inclusion of an upsell offer, reading of a relevant disclaimer, script adherence and more. Assessing the calls and documenting the results are important for identifying coaching and training needed by individual agents to improve their performance. Although many quality practitioners may consider their quality programs automated, their programs typically consist of manually evaluating, scoring and reporting on five to seven percent of an agent’s calls. Even the best contact centers typically have not been able to evaluate a statistically significant portion of calls, and so might miss critical skill gaps that demand attention. Automated Quality Management (AQM) solutions change all that. It’s an important addition to any contact center’s technology stack, providing solid assurance that an organization’s performance and customer experience improvement programs are on target. AQM provides automated performance scorecards in an at-a-glance format that make it easy for employees and their managers to view results and pinpoint areas for improvement. Many contact centers currently produce agent performance reports monthly, so problems can fester before they are correctly identified and addressed. By automating the scoring of evaluation forms based on business rules created for each question, AQM makes agent performance evaluation instantaneous, so managers can address skill gaps within a day rather than at the end of the month. Scorecards can trigger alerts that then schedule coaching on the skill and knowledge gaps identified, and progress can be tracked after each session to gauge uptake and effectiveness. 3. Voice of the Employee solutions make it clear you care Voice of the Employee (VoE) solutions provide a timely, cost-effective alternative to traditional, third-party employee satisfaction surveys. Organizations can easily survey thousands of employees across hundreds of locations and multiple languages, then centrally measure, analyze and take action on their feedback. By capturing employee feedback using their preferred channels—including email, web, mobile, and SMS surveys—you can gain critical insights for fostering employee engagement and empowerment, productivity, satisfaction and retention. The cumulative benefit to overall customer satisfaction is clear. While customers who provide feedback are only able to talk about their own individual experiences, employees on the front lines are ideally positioned to recognize patterns based on their conversations with the many customers who call them every day. Not only can they identify common issues, but they can also assess how important these issues are to customers, help to understand what’s causing them in the first place, and provide insight into how customers would like to see those issues corrected. Granted, contact centers today have access to all sorts of analytics to mine voice recordings, text, the web, social media, and other customer touch points, so they can spot trends and respond to issues more quickly. The importance of analytics cannot be overstated as a way to extract call content, identify emerging issues or concerns, and quantify the number of calls regarding a specific problem, but getting the information directly from contact center agents in almost near real time allows the company to act on and resolve issues much faster. And while VoE is a somewhat roundabout way to get at customer sentiment, it can be more reliable than VoC solutions alone. For example, a large IT outsourcing company combined and analyzed customer and employee feedback to optimize internal business processes and increase customer experience scores. After asking customers to give a customer satisfaction score on their experiences with departments such as sales, operations, help desk, billing and delivery, they asked employees to rate the same departments on how important their work is to the success of their jobs. They also used text analytics to analyze verbatim comments to help identify trends. Using one centralized platform, the company then mapped the survey data from both and drew conclusions to prompt process improvements. The time to engage is now According to a recent Gallup report on the state of the American workplace, 70 percent of the workforce is disengaged. They lack enthusiasm and commitment to the job. Companies with highly engaged employees, the report notes, are 22 percent more profitable, 21 percent more productive and have a 10 percent higher customer satisfaction rating than organizations with somewhat or highly disengaged employees. Understanding that an engaged-employee culture is essential to a positive customer experience, companies today have the analytics tools available to help them score big in both areas.","excerpt":"Gartner recently announced its inaugural Magic Quadrant for Workforce Engagement Management (WEM). It was an important indicator of the evolution from workforce optimization (WFO) technologies, an already mature market, to technologies that help drive employee engagement. Though gamification, mobile scheduling apps and robotic process automation (RPA) tools all contribute to helping employees, especially Millennials, feel […]","categories":["IT Services"],"tags":[],"author_name":"Jason Du Preez","publish_date":"2018-02-26T07:34:24","publication_year":"2018","word_count":1338,"keywords":["Go","API","AI","Git","automation","Aim","ViT","analytics","GAN","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","Git","API","GAN","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/it-services\/analytics-leap-workforce-engagement\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10072204,"title":"Node.js founder is disappointed with Node.js","content":"Four years ago, from the stage of Javascript Conference EU, Ryan Dahl, founder of Node.js, opened his address with regrets. The list includes: removal of promises from Node, which are necessary abstraction for async\/await; lacking security; including GYP build system, and allowing Node to inspect package.json for ‘main’, among others. After rueing nearly 15 minutes over his past mistakes, he made an announcement of a new software he was working on. In a very nervous and quivering voice, he announced to the world the birth of ‘Deno.js’, a simple modern runtime for Javascript and Typescript that uses V8 and is built on Rust. He explained that Deno is an attempt to undo the mistakes he committed with Node—lack of attention to security, module resolution through node_modules, and various deviations from how browser’s worked, among other things. A lot of developers have issues when it comes to the security part of Node.js, especially when access is granted to packages that shouldn’t be able to use our machine or network. According to a study, conducted by a hacker, 14% of the Node Package Manager (NPM) ecosystem is impacted and about 54% of the NPM ecosystem is likely to be impacted indirectly. Released in May 2020, Deno.js has a more robust core and several functionality and JavaScript capabilities. Deno: Safer, faster, lighter The software is built on Rust, which experts say that was the ideal decision since it not only made the runtime quicker and bug-free but also increased security. Once a programme is launched, it may quickly access your file system or network—a severe security weakness. Deno solves the issue by executing the code in a sandbox. It means that your file system or network are not accessible to the runtime. Besides security, Dahl’s concern was also related to speed. To solve this, he used Rust, a computer language that has recently gained popularity in the development world. This language has gained popularity for a variety of reasons, including guaranteeing that our applications are free from undefined behaviour and data races, increased memory safety, and so on. Rust is an extremely safe and fast programming language. In addition to being safer and faster, the software is lighter than Node. Dahl added built-in package management for resource retrieval, eliminating the need to utilise NPM. This is expected to be the game changer in the way modern JavaScript apps are designed. Instead of downloading packages from the NPM repository, it needs a URL, and Deno fetches the dependencies, much like browsers. Another advantage of Deno is that it caches all modules that you download. So, if a module is downloaded, Deno will cache it and not download it again until it is given the reload command. “Browsers have not blessed any one CDN for distributing JavaScript—the decentralized nature of the web is its greatest strength. I don’t see why this can’t also work for server-side JavaScript too. Thus I want Deno to not be reliant on any centralized code database”, said Ryan Dahl, founder of Node.js and Deno.js. So, what about Node.js? Deno is, undoubtedly, addressing the various drawbacks of Node, but it’s nowhere close to posing any threat to its predecessor. According to experts, Deno would take several years to gain traction. It is currently quite restricted, with many libraries still lacking. As a result, businesses would be slow to adopt it since it requires them to rewrite their programmes, which is time-consuming and costly. If Deno can provide compatibility with prominent Node.js libraries, widespread adoption may be achievable. JavaScript has been in existence for 26 years. It has evolved throughout time, adding new terms and features to keep up with the times. In this regard, Node.js may also alter and perhaps absorb many of the functions provided by Deno.","excerpt":"For Ryan Dahl, Deno.js is an attempt to undo the mistakes he did with Node.js","categories":["AI Trends"],"tags":[],"author_name":"Tausif Alam","publish_date":"2022-08-05T13:07:52","publication_year":"2022","word_count":627,"keywords":["Go","Rust","programming_languages:R","AI","programming_languages:Java","TypeScript","JavaScript","R","Java","programming_languages:JavaScript"],"extracted_tech_keywords":["AI","R","JavaScript","TypeScript","Go","Rust","Java","programming_languages:R","programming_languages:JavaScript","programming_languages:Java"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/node-js-founder-is-disappointed-with-node-js\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10077089,"title":"Google Cloud to Let Users Pay with Ethereum, Bitcoin, Litecoin","content":"The world’s leading cloud computing service Google Cloud and Coinbase on Tuesday announced a strategic partnership to serve the growing Web3 ecosystem. As part of the collaboration, Google Cloud is positioned to enable customers – in the Web3 ecosystem – to pay for its cloud services via select cryptocurrencies. According to a press release, Coinbase will use Google Cloud’s powerful compute platform to process blockchain data at scale, enhancing the global reach of its crypto services. Moreover, Coinbase will build its global data platform on Google Cloud’s secure infrastructure to leverage data and analytics technologies to provide customers with ML-driven crypto insights. Google Cloud CEO Thomas Kurian says, “We want to make building in Web3 faster and easier, and this partnership with Coinbase helps developers get one step closer to that goal.” Vice President Amit Zavery informed that the strategic partnership with Coinbase will be through their platform ‘Coinbase commerce’, giving access only to selected customers. At present, the platform only accepts ten cryptocurrency payments which includes: Ethereum, Bitcoin, Litecoin, Bitcoin Cash, among others. “We are excited that Google Cloud has selected Coinbase to help bring Web3 to a new set of users and provide powerful solutions to developers.With more than 100 million verified users and 14,500 institutional clients, Coinbase has spent more than a decade building industry-leading products on top of blockchain technology,” said Coinbase CEO Brian Armstrong. Moreover, the partnership allows Web3 developers to access Google’s BigQuery crypto public datasets, powered by Coinbase Cloud Nodes, across leading blockchains. The integration allows developers to reliably operate Web3-based systems without the need for complex and expensive infrastructure.","excerpt":"The partnership will involve Coinbase selecting Google Cloud as a strategic cloud provider to build advanced exchange and data services","categories":["AI News"],"tags":["Bitcoin","coinbase","Cryptocurrency","digital currency","Ethereum","Google","Google Cloud"],"author_name":"Bhuvana Kamath","publish_date":"2022-10-12T14:21:04","publication_year":"2022","word_count":267,"keywords":["Go","Bitcoin","Google Cloud","coinbase","Cryptocurrency","AI","cloud computing","ML","R","programming_languages:R","programming_languages:Go","RAG","analytics","Google","cloud_platforms:Google Cloud","digital currency","Ethereum"],"extracted_tech_keywords":["AI","ML","analytics","RAG","cloud computing","R","Go","cloud_platforms:Google Cloud","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-cloud-to-let-users-pay-with-ethereum-bitcoin-litecoin\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":24447,"title":"Qubole &#038; AIM Hosts Big Data CXO Series | May 25 | Gurugram","content":"The challenges of utilizing big data infrastructure in the cloud are significant and involves massive scalability and overcoming legacy warehousing constraints. Qubole, in partnership with Analytics India Magazine will highlight the challenges involved in building cloud data warehouses and how to make the most of data for insight-driven decision making with an exclusive Big Data CXO Series on May 25 in Crowne Plaza Gurugram. The conference will see a host of data visionaries sharing their experience of building up a robust Big Data Architecture and scaling it to meet with the rapidly growing demands of business with a series of talks on How To Activate Your Big Data.  The Key Takeaways: Insights into common Big Data Challenges, including Scaling, Effective Democratization of Data\/Analytics using modern platforms, Real-world ideas of how to leverage AI and ML in your business, including customer stories, lessons learned and best practices for success. The findings from a recent Big Data Analytics Study  from Dresner Advisory Services indicate that data warehouse optimization and how to get value from data is considered the most important aspect by 70% of respondents. This emerged as one of the most important use cases of 2017. According to a market study by SNS Research, by the end of 2017, as much as 30% of all Big Data workloads will be processed via cloud services as enterprises seek to avoid large-scale infrastructure investments and security issues associated with on-premise implementations. The conference will focus on helping enterprises of all sizes use the cloud effectively to scale operations and educate the attendees about cloud data warehouse and how it can lead to faster decision making with integrated analytics. Bhasker Gupta from Analytics India Magazine says; “We’re happy to partner with Qubole to help customers better understand how to build big data infrastructure effectively for more wide-spread, data-driven decision making.” The keynote will be helmed by Sanjay Srivastava from American Express. During the event, attendees will gain insights from customer stories and best practices shared from industry veterans on how to build a big data infrastructure and scale it effectively. What: Big Data CXO Series Where: Crowne Plaza, Sector 29, Gurugram When: May 25, 2018 | 7:00 pm – 11:00 pm About Qubole: Qubole is a California-headquartered company that helps customers transition from a legacy on-premises data warehouse to an elastic, open source data lake in the cloud. QDS is a big data service helping you move your big data workloads to the cloud for more flexibility and lower cost or scaling your automation. Register for the event here.","excerpt":"The challenges of utilizing big data infrastructure in the cloud are significant and involves massive scalability and overcoming legacy warehousing constraints. Qubole, in partnership with Analytics India Magazine will highlight the challenges involved in building cloud data warehouses and how to make the most of data for insight-driven decision making with an exclusive Big Data […]","categories":["Deep Tech"],"tags":[],"author_name":"Дарья","publish_date":"2018-05-09T08:23:29","publication_year":"2018","word_count":425,"keywords":["big data","API","AI","ML","Scala","RAG","analytics","R","data lake","data warehouse"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Scala","API","big data","data warehouse","data lake"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/qubole-aim-hosts-big-data-cxo-series-may-25-gurugram\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":14956,"title":"Study &#8211; Analytics &#038; Data Science Employee Attrition In India","content":"Attrition is a serious concern for analytics industry in India, especially when the availability of data scientists is called out as the biggest challenge facing organizations. Whenever, a well trained and knowledgeable employee leaves, the organization loses key skills, knowledge and business relationships. Moreover, there are serious costs involved in replacing that employee. In this independent study, we try to put in some numbers around attrition rates in the analytics industry. This study is an outcome of a sample size of 20 largest analytics recruiters and the analysis of their current strength and past employees. The main objective of this study is to provide benchmarking against which employee attrition can be measured. We will keep expanding the study over the year, as we accumulate more data as well as needs. The average employee attrition in Analytics & Data Science Industry in India currently stands at 24.4% annually. Also, analytics employees tend to stay in their organization for an average of 3.7 Years. In terms of the organization type, the analytics function at IT providers tends to have the highest attrition at 26%. IT service providers are the biggest recruiters of analytics professionals and therefore tend to skew the overall attrition rates higher. GIC’s\/ Captive have an employee attrition at 22.1%, followed by Consulting firms at 21.9% and Boutique analytics firms at 21.6%. IT service providers also tend to have the highest tenure of analytics professionals in their current organizations, at 4.5 years. We noticed no correlation between years in the current organization to attrition rates. Captives have an average of 3.9 years as employee tenure followed by consulting firms at 3.3 years and Boutique analytics firms at 3 years. Attrition Trends by Cities Mumbai currently has the highest annual attrition rates in analytics, at 25.7%. The city boasts of a robust data science ecosystem, mostly from banking and finance industry. Coupled with high salaries and demand for data scientists, Mumbai is pushed to the highest position in terms of employee attrition. Chennai, on the other hand, has the lowest attrition rates, at just 15.5%. It is evident that larger cities, with bigger employee base in analytics (Bangalore, Mumbai & Delhi\/ NCR) have higher attrition rates, higher than 20%. While, cities with lesser employee base have in turn lower attrition rates, lesser than 20%. Also, current analytics employees in Mumbai are on average staying their current organizations for an average of 4.2 years. As earlier mentioned, we did not find any correlation with attrition rates and tenure. Tenure of analytics professionals in Chennai is highest, at 4.5 years. Attrition Trends by Experience Level Employee attrition increases with a decrease in experience levels. Younger employees tend to jump jobs more often. At senior level i.e. more than 10 years experience, the attrition rates for analytics industry stands at just 13.7%. While at fresher’s level, the attrition rates are almost 80%. The same is true for average tenure in the current organization. More than 10 years experience has an average tenure of 7 Years.","excerpt":"Attrition is a serious concern for analytics industry in India, especially when the availability of data scientists is called out as the biggest challenge facing organizations. Whenever, a well trained and knowledgeable employee leaves, the organization loses key skills, knowledge and business relationships. Moreover, there are serious costs involved in replacing that employee. In this […]","categories":["AI Features"],"tags":["AI Jobs","study data science","Types of Databases"],"author_name":"Дарья","publish_date":"2017-05-15T12:51:36","publication_year":"2017","word_count":501,"keywords":["data science","programming_languages:R","AI","Types of Databases","AI Jobs","RAG","GAN","analytics","study data science","R"],"extracted_tech_keywords":["AI","data science","analytics","RAG","R","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/analytics-india-employee-attrition-study-2017\/","complexity_score":3,"technical_depth":7,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10045436,"title":"Why Is Knowledge-Supervised Deep Learning Important","content":"Researchers across the world are performing iterations and experiments to unlock the capacity that Artificial Intelligence (AI) has, in terms of recognition, solving complex problems, and providing accurate results. Unlike the recognition and observation developed by the human brain, machines largely have relied on inputs that are fed in a model, to fetch certain results. Researchers, however, aim to let neural networks function independently, with or without inputs— using graphs. Right with Leonhard Euler’s invention of the graph theory in the 18th century, it has served to be one of the most efficient tools for analytics, to mathematicians, physicists, computer experts, and data scientists. Researchers and data scientists have extensively used “Knowledge graphs” to organise structured knowledge over the internet. The knowledge graph is a process to integrate information extracted from several sources and feed them to a neural network for processing. Extrapolation is another way researchers think the AI can be trained to imagine the unseen, with structured inputs that are fed to any neural network. Unlike in machine learning, where humans train the parameters of any given model using data, the neural network in deep learning can self train itself using both— structured and unstructured data. Source: Medium.com What is the purpose of a Knowledge graph? Unlike any other technique, knowledge graphs have directed labeled graphs— with well-defined meanings for labels. These graphs have nodes, edges, and labels where anything can act as a node— for example an object, a place, person, or a company, etc. It is the “relationship of interest” between two nodes that are connected by edges. Labels define the meaning of relationships between nodes. For example, “employees” in a company are nodes and their respective “departmental managers” are the edges— the relationship here is defined with the label “colleagues”. With the deployment of such nodes to edge networks in machine\/ deep learning, data science researchers are working to make neural networks independent, in performing tasks. After having used PageRank which relied on links to rank a page on the web for authenticity, Google deployed knowledge graphs in its search mechanism, in 2012. According to Neo4j, when a search operation is performed on the web using the graph network, it is the latency of the query that is proportional to how much of the graph you want to traverse, rather than how much input is stored. An open-source research organisation— Octavian, uses neural networks to perform tasks on “knowledge graphs”. While it is often said that deep-learning performs well with unstructured data, the superhuman neural networks modeled by Octavian deal with specific structured information and the neural architectures are engineered to suit the structure of the data they work well with. One of Octavian’s data science engineers delivered a talk in 2018, at the Connected Data London conference, where he compared the traditional forms of deep learning to graph learning. The question he presented to establish his claim was “How would we like machine learning on graphs to look from 20,000 feet?” As the neural networks require training to perform any task, his findings pointed out that many ML techniques in use on graphs have some fundamental limitations. While some techniques needed conversion of the graph into a table—discarding its structure, other techniques don’t work on unseen graphs. With an aim to overcome such limitations and use deep learning on graphs— the Octavian researcher identified certain “graph-data tasks” which demanded “graph-native implementations like Regression, Classification, and Embedding. ‘Relational Inductive Biases’ Source: Slideshare\/ Octavian\/ Connected Data London The focus in this study of using deep learning on graphs is on data structures that function well with neural networks. For instance, images are in a structured format as they are either two or three-dimensional— where pixels near to each other are relevant to each other than those which are far from each other. Whereas, sequences have a one-dimensional structure where items adjacent to each other are more relevant to one another than those that are far apart. Source: Slideshare\/ Octavian However, the nodes in graphs don’t have fixed relations like in images and sequences— the example of pixels and sequences. In order to achieve the best results in a deep learning model, Octavian claims that the creation of neural network models that match the graphs is the key. To establish the accuracy in results acquired by using deep learning of graphs, Google Brain along with MIT and the University of Edinburgh had released a paper on “Relational Inductive Biases”. The paper introduced a general algorithm to propagate datasets through a graph, and argued that a state of art performance can be achieved on the selection of graph tasks— by using neural networks to learn six functions to “perform aggregations and transforms within the structure of the graph”. ‘Combinatorial generalisation’ The combined thesis presented by Google Brain, MIT, and The University of Edinburgh argued that “combinatorial generalisation” is the top priority for artificial intelligence to achieve human-like abilities. The study argues that structured representations and computations are the keys to empower AI with analysing capabilities like that of a human brain. According to the paper, the graph network generalises and extends various approaches for neural networks that operate on graphs— to provide a straightforward interface for manipulation of structured knowledge and produce “structured behaviours”. The participants in this research have used deep learning on graphs to prove that neural network models yield more accurate results when they work with graphs. Google’s knowledge graph A well-established example of the aforementioned studies is Google. It has been using “knowledge graphs” for eight years. Rolled out on May 16, 2012, Google’s knowledge graph was invented to enhance the search experience. Using this graph, Google improvised its path to generate results for the search. In simple terms, when you search for the name of a movie ‘Batman: The Dark Knight’, the Google search yields all possible results with posters, videos, hoardings, advertisements, and the movie halls that are showing the film. It is the knowledge graph with deep learning that Google has been using to optimise its search engine, where billions of users arrive each day. It was Google’s X lab that built a neural network of 16,000 computer processors with 1 billion connections, after which the artificial brain browsed YouTube and searched for cat videos. No input was provided to the neural network, yet Google’s artificial brain used deep learning algorithms, combined with knowledge graphs— to perform one of the most common searches that even a human brain would. Graphs have proved to be the core of communication for the neural networks in deep learning. Such a learning process has several hidden layers which the network uses to generate the best results for any input. With knowledge graphs, Google has made groundbreaking achievements and acquired DeepMind in 2014 to further its research and study in deep learning algorithms. Google Assistant, Voice recognition on Facebook, and in your smartphones, Siri, Unlock using face recognition, fingerprint unlock and more are among several such innovations where AI has achieved recognition with analysis.","excerpt":"Graphs have proved to be the core of communication for the neural networks in deep learning.","categories":["AI Features"],"tags":["Knowledge graphs"],"author_name":"Gourav Mishra","publish_date":"2021-08-08T12:00:00","publication_year":"2021","word_count":1168,"keywords":["data science","knowledge graphs","machine learning","artificial intelligence","AI","Knowledge graphs","neural network","ML","Aim","deep learning","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","data science","analytics","Aim","knowledge graphs"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-is-knowledge-supervised-deep-learning-important\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10172526,"title":"92% of Indian Employees Embrace GenAI at Work, Highest in the World","content":"India leads the world in generative AI adoption, with 92% of employees using the technology regularly at work, well above the global average of 72%. According to a new report, the country ranks ahead of the Middle East and Spain, reflecting a broader trend of higher AI adoption in the global South compared to the global North. The survey, released on Thursday by Boston Consulting Group, is based on responses from 10,600 workers across 11 countries that suggest AI usage has turned mainstream. India also ranks second globally in AI agent integration, with 17% of employees reporting that their companies have incorporated AI agents into workflows. While employees recognise the potential of AI agents, most still lack a clear understanding of how they work. Although 77% believe AI agents will be important in the next three to five years, only 33% say they have a proper understanding of what these agents actually are. This gap in awareness reflects the early stage of deployment, with most companies still in the experimental phase rather than having fully integrated their systems. Concerns about trust and governance are growing prominent. About 46% of employees are worried about decisions being made without human oversight, 35% point to the risk of bias or unfair treatment, and 32% are concerned about unclear accountability when mistakes occur. Countries with high generative AI usage are also seeing greater concerns about job security. In India, 48% of respondents believe their job will certainly or probably disappear entirely within the next 10 years, which is higher than the global average of 41%. At the same time, AI is helping employees save time, with 47% reporting that they save more than an hour a day. However, only one-third receive guidance on how to use this time effectively, which limits the overall impact. Successful AI adoption depends on three key enablers: proper training, access to the right tools, and strong leadership support. Only 36% of employees feel adequately trained, and those without in-person instruction or coaching are up to 26 percentage points less likely to be regular users. Tool availability also remains a challenge, with 37% employees saying their company does not provide the right AI tools. As a result, 54% turn to unauthorised alternatives, especially among Gen Z and Millennials, where 62% say they would bypass restrictions. Leadership support is equally important. Only 25% of frontline workers report receiving it, but when it is present, regular AI usage increases from 41% to 82%, along with significant improvements in job satisfaction and career optimism.","excerpt":"A BCG survey has found that India is at the forefront of generative AI adoption and near the top in AI-agent integration (17 %). Yet, 48 % of employees express concern that their roles may vanish in the next 10 years.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI integration","BCG","empoyee"],"author_name":"Smruti S","publish_date":"2025-06-27T17:46:55","publication_year":"2025","word_count":421,"keywords":["Go","AI integration","empoyee","AI","programming_languages:R","programming_languages:Go","RAG","BCG","AI agents","generative AI","Rust","R","programming_languages:Rust","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","generative AI","RAG","R","Go","Rust","AI agents","programming_languages:R","programming_languages:Go","programming_languages:Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/92-of-indian-employees-embrace-genai-at-work-highest-in-the-world\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10162293,"title":"OpenAI Finally Has a Competitor, says CEO Sam Altman","content":"In a series of tweets, OpenAI CEO Sam Altman acknowledged DeepSeek as an emerging rival on Tuesday and praised the advanced capabilities of its new R1 model. As DeepSeek takes over the internet and the stock market, Altman also expressed confidence that OpenAI will continue to produce even stronger and more advanced systems. “DeepSeek’s R1 is an impressive model,” Altman wrote. “We will obviously deliver much better models, and it’s invigorating to have a new competitor!” As the conversation about computing heated up with the crash of semiconductor stocks yesterday, Altman added, “We believe more computing is more important than ever before to succeed at our mission. The world is going to want to use a lot of AI and will be amazed by the next-generation models coming.” It is interesting to note that while OpenAI is charging $200 a month for its version of the o1 reasoning model, DeepSeek is offering it for free.  “Ironic that we got free AI from a hedge fund and $200\/month AI from a nonprofit,” Avichal, co-founder at Electric Capital, said. This morning, US President Donald Trump described China’s DeepSeek AI model as a “wake-up call” for American firms while welcoming it as a positive development toward more rapid and cost-effective methods of advancing AI. Interestingly, he also announced that the US will soon place tariffs on all semiconductors and pharmaceuticals imported from Taiwan. After the stock market collapsed, Deepseek introduced a fresh lineup of models, even though many were still adjusting to R1. The standout among the bunch is Janus-Pro – an advanced autoregressive framework designed to unify multimodal understanding and generation. By separating visual encoding into distinct pathways but keeping a single transformer at its core, Janus-Pro avoids the usual conflicts and gains extra flexibility. “The main two implications of DeepSeek are (1) an acceleration of AI model development until they hit a wall and (2) increased likelihood that there is no wall in the near future,” The Wharton School’s professor Ethan Mollick said. Yesterday, tech giants like NVIDIA, Microsoft, and Alphabet saw their stock prices plunge, while the Nasdaq 100 and Europe’s Stoxx 600 tech sub-index bled nearly $1 trillion in market cap. US stocks, particularly in the semiconductor sector, saw significant losses, while Chinese markets held steady, with their strongest performance relative to US markets in over two years. Chip maker NVIDIA’s shares were down nearly 15% on the New York Stock Exchange (NYSE) in early trade on Monday.  Speaking at the World Economic Forum (WEF) earlier this month, Microsoft CEO Satya Nadella said, “We should take the developments out of China very, very seriously.”","excerpt":"And it is not Google or Anthropic.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","DeepSeek"],"author_name":"Aditi Suresh","publish_date":"2025-01-28T11:18:09","publication_year":"2025","word_count":435,"keywords":["Go","API","OpenAI","AI","Modal","programming_languages:R","programming_languages:Go","DeepSeek","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","OpenAI","R","Go","API","Modal","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/openai-finally-has-a-competitor-says-ceo-sam-altman\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":57949,"title":"Cybersecurity In Healthcare Industry","content":"2017, Wannacry cyberattack shook the National Health Service of the UK, which infected more than 2,00,000 computers, putting a total number of 16 healthcare centres on halt that led to the cancellation of more than 19,000 appointments. In 2019, a man by the name of Martin Gottesfield from Massachusetts launched a DDoS attack on Boston’s Children Hospital due to which the hospital lost a total amount of $300,000 in order to repair its computers. These are just two cybersecurity attacks from a thousand ones which could be mentioned here, and at the time when this article is being typed, another cyber attack is taking place on some health institute in some part of the world. As the healthcare industry evolved, it began to primarily rely on technology and digitalisation — even went entirely paperless in some parts of the world. On the one hand, the operating systems inside the healthcare institutes are becoming faster and less complicated, but on the other hand, it’s falling prey to cyber-attacks. But for every ‘anti’, there will always be a ‘pro.’ Thus came the role of cybersecurity to stand as a shield in front of healthcare institutes. Healthcare institutions store large amounts of health data related to patients’ condition along with medical billing and insurance information, which are of high value in the black market and are often targeted by hackers. Some of the common kinds of attacks to keep an eye for are malware and ransomware, where hackers often shut down individuals’ devices and servers. Information stored on cloud storage can also become a weak spot without proper encryption. Another clever idea of attacking is the usage of websites and phishing attacks. Through the use of cybersecurity, several healthcare institutions and providers are educating members on how to protect data and devices along with cultivating a sense of security. Despite the rise in threats, most of the hospitals are incapable of handling the attacks and often do not take these threats seriously. The data breach is not the only threat that lurks. The proliferation of connected medical devices such as pacemakers and insulins can also put the life of a patient at serious jeopardy. Not to mention, most of the attacks are the cause of negligence by inside staff who have access to the organisation’s EHR. The attack on healthcare institutions can only be judged by numbers, i.e., the amount of money lost in the attacks. As per a survey by IBM, the healthcare industry had faced a loss of $6.45 million in the year 2019 alone. In the year 2015, more than 113 million records were stolen and out of which, 78.8 million were stolen in a single attack. The reason these attacks took place in the first place is the lack of cautiousness on the part of healthcare providers. In terms of security, the healthcare industry has always lagged while hackers went on to create sophisticated tools. Healthcare institutes have finally begun to ask for security from their machine providers to deal with the nuisance. In the race between hackers and healthcare institutions, artificial intelligence (AI) is the ultimate weapon that has been chosen by both parties to attack and defend. Hackers are on the lookout to design attack vectors that can get passed cyber defence by leveraging AI and machine learning (ML). It is often noticed that advanced persistent threats (APTs) are found hiding in the networks for years and go unnoticed, collecting information years after years. In certain scenarios, offensive AI changes itself as per the environment and mimic specific actions to avoid human detection. To counter such a wide variety of weaknesses, AI is at the forefront of defending. Advanced malware protection is put to use which inoculates the LAN and counters back to unrecognised behavioural patterns. Several health institutes are relying on IBM’s Watson platform that provides an AI system for routine security assessments, reducing response time in case of an attack and making an auto recommendation to deal with a specific kind of attack. Not just IBM Watson but Cisco too is on the forefront with several cybersecurity solutions such as security broker and cognitive threat analytics to list a few. Moving on, cybersecurity was never a part of the hospital IT department. But in recent times, that has been changed as healthcare institutes are establishing a connection between security and IT departments to control and respond to any attack effectively. Most healthcare institutes are also developing a reliable security program based on official frameworks, notably the NIST framework for health information technology. Healthcare institutes are also taking recommendations from the Food and Drug Administration, along with following the HIPAA guidelines to ensure a safe and secure healthcare environment. In the fight against hackers, cybersecurity is on the rise with new findings to counter the attacks, but the war between the two entities will never come to an absolute end. However, one can further read about the regulatory challenges presently prominent with healthcare AI in current times.","excerpt":"2017, Wannacry cyberattack shook the National Health Service of the UK, which infected more than 2,00,000 computers, putting a total number of 16 healthcare centres on halt that led to the cancellation of more than 19,000 appointments.  In 2019, a man by the name of Martin Gottesfield from Massachusetts launched a DDoS attack on Boston’s […]","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","cisco","computer security","Cybersecurity","FDA","healthcare India","Machine Learning","NIST"],"author_name":"Rohit Chatterjee","publish_date":"2020-03-04T15:00:22","publication_year":"2020","word_count":831,"keywords":["Go","artificial intelligence","machine learning","AI","NIST","ML","Machine Learning","computer security","Git","RAG","analytics","healthcare India","FDA","GAN","Cybersecurity","R","AI (Artificial Intelligence)","cisco"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","analytics","RAG","R","Go","Git","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/cybersecurity-in-healthcare-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10088078,"title":"Web3 to Cause a Tectonic Shift in How we View Tech Education","content":"Despite the negative stance taken by regulatory authorities on cryptocurrencies, the adoption of blockchain technology has witnessed an upward momentum in India in recent months. India’s finance minister, Nirmala Sitharaman, anticipates a 46% increase in the utilisation of blockchain technology in the country in the upcoming years. Keeping in mind the importance of the technology, recently, Atal Innovation Mission, which comes under Niti Aayog, has set up the ATL Blockchain Module to accelerate Web3 education in the country. Dr Chintan Vaishnav, mission director at Atal Innovation Mission, believes that blockchain technology has the potential to unlock tremendous opportunities for the youth and the module will introduce students as young as Grade 6 to the concept of blockchain. ATL Blockchain module For this, Niti Aayog has partnered with 5ire, a blockchain ecosystem that focuses on sustainability, technology and innovation to build the 5th industrial revolution (5IR). The company’s primary offering is 5ireChain, a first layer, sustainability-driven 5th generation blockchain that ensures adherence to the philosophy of 5IR, creating a net positive impact on the planet and service to humanity. Through the Blockchain Module, the government wants to educate Indian students about blockchain technology as they believe that an in-depth understanding of the technology could provide students with a competitive edge in the job market. “Web 3 education will cause a tectonic shift in how we view tech education. Our partnership with Niti Aayog is essentially seeking to make this transition,” Dr Pratik Gauri, co-founder and CEO, 5ire, told Analytics India Magazine. “There are a couple of ways we are approaching it. Firstly, mass education of young people and offering it up as an opportunity to advance their careers and through partnering with governmental organisations and transforming the public sector to use blockchain. Secondly, there is also a surge of communal activity for the adoption of blockchain.” The learning course on blockchain encompasses nine sections covering topics such as an introduction to blockchain, its advantages, the challenges it encounters, consensus algorithms, digital assets, smart contracts, and the basics of civil identities on distributed ledgers. Moreover, the course delves into blockchain use cases in various industries, including financial services, healthcare, supply chain, manufacturing, and automotive sectors. “A business model to consider for web3 education is ‘Learn to Earn’, a concept that refers to the idea of using decentralised platforms to earn cryptocurrency or other digital assets while learning new skills or participating in educational activities,” Dr Gauri said. A decentralised language-learning platform could reward users with tokens for completing lessons or passing quizzes. These tokens could then be used to access more advanced content or traded on decentralised markets. “Learn to Earn can also be achieved through the use of decentralised platforms for online tutoring or peer-to-peer education. Tutors or educators can earn cryptocurrency or tokens for providing their services, and students can earn cryptocurrency or tokens for participating in or completing educational activities,” Dr Gauri said. Blockchain for governance The Government of India has also leveraged the power of blockchain over the years. For instance, ‘DigiLocker’, an Indian digitisation online service provided by the Ministry of Electronics and Information Technology (MeitY), is based on blockchain technology. “Governmental bodies moving to digitise their records and services can quickly leapfrog to proven technologies to serve their constituents today rather than suffer through yesterday’s archaic systems,” Dr Gauri said. Last year, the Police department of Firozabad revealed a blockchain-powered project aimed at monitoring public complaints. The project has proven to be beneficial for the Uttar Pradesh (UP) Police in their efforts to tackle local police corruption and criminal activities. “Aside from Niti Aayog we have three other partnerships with governmental organisations to develop blockchain-based solutions. These include a project to move Goa Police, Muzaffarnagar Police and Nebraska, USA Police and take governance to the blockchain,” Dr Gauri said. Further, he also revealed that 5ire is the first blockchain network that comes equipped with local governance solutions from the provenance of private property and title records as well as recording private business transactions from agreements, licences, registration, and intellectual property registration to personal credentials like passport, visa, driver’s licence, and birth records that are securely stored in blockchain systems.","excerpt":"An in-depth understanding of the technology could provide students with a competitive edge in the job market.","categories":["Deep Tech"],"tags":[],"author_name":"Pritam Bordoloi","publish_date":"2023-02-24T15:58:11","publication_year":"2023","word_count":693,"keywords":["Go","AI","Git","RAG","Aim","ViT","analytics","GAN","Tecton","R"],"extracted_tech_keywords":["AI","analytics","Aim","Tecton","RAG","R","Go","Git","GAN","ViT"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/web3-to-cause-a-tectonic-shift-in-how-we-view-tech-education\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10171190,"title":"Tredence’s Case for Agentic Data Engineering","content":"In today’s data-intensive world, the traditional approach to data engineering is increasingly seen as a bottleneck rather than an enabler of business growth. At AIM’s DES 2025 event, Maulik Dixit, a senior director of data engineering at Tredence, highlighted the problem that many organisations face. With more than 20 years of experience building and modernising data platforms, Maulik highlighted the growing gap between the promise of data and the reality of getting value from it. “Data engineering is the single biggest bottleneck, in my view, that stops businesses from getting real value out of their data.” He strongly recommended using intelligent agents to improve how data systems are built, monitored, and maintained. The focus was on moving away from manual, slow processes to something more efficient, scalable, and ready for the future. Understanding the Investment in Data Maulik began by contextualising the scale of data investment in large enterprises. Using an example, he explained how even large companies must use their data analytics budgets wisely, as it’s a small portion of overall revenue. He explained that for a company with $50 billion in revenue, about 4% typically goes to IT, which is around $2 billion. Of that, 20% is spent on data analytics. So, in effect, $400 million needs to be spent very wisely on data analytics projects. This highlights the need to prioritise high-impact, value-driven projects. Despite this significant outlay, data engineering processes often suffer from long timelines, rapidly evolving technologies, a need for highly specialised talent, and low automation. These challenges create friction and slow down the speed at which businesses can derive actionable insights and value from their data. To address these constraints, Maulik gave a refresher on the architecture of AI agents. Unlike traditional automation tools that follow fixed rules, AI agents powered by generative AI are capable of reasoning, making decisions, and responding to natural language inputs. They function using four foundational components: tools (reusable functions), memory (contextual storage), planning (decision-making), and action (execution). This architecture allows them to perform complex workflows with minimal human intervention. Maulik illustrated the transformation by comparing the traditional and agentic approaches to data operations. In the conventional setup, data engineers manually create pipelines, monitor systems, respond to incidents, and generate reports. This is not only time-consuming but also prone to delays and errors. In contrast, agentic systems can automatically generate pipelines, conduct real-time monitoring, trigger event-driven actions, and provide insight-based dashboards. The shift from reactive to proactive incident handling significantly reduces downtime and enhances reliability. He shared a relatable story to show the impact of AI agents. When a critical ETL (extract, transform, and load) job fails, support staff typically struggle to diagnose the issue, leading to delays and missed SLAs. With an AI agent in the picture, the problem is detected, resolved, and communicated automatically within minutes. This highlights significant improvements in speed and reliability. AI agents open up a wide range of use cases in data engineering. One of the most impactful is code translation, which automatically converts legacy ETL code into modern platform-compatible code. This traditional multi-year undertaking can now be completed in a fraction of the time. “Ingesting a new data source, which used to take weeks for us, is now going to take some days,” Maulik added. Other use cases include automated pipeline documentation, accelerated onboarding of new data sources, real-time data quality checks, and intelligent support. Each of these areas contributes to faster, more efficient, and more scalable data operations. Building the Future with Trust However, Maulik was careful to acknowledge the associated risks. In data engineering, accuracy is paramount. Anything less than a 99% SLA is often unacceptable to the business. He emphasised that AI agents must be monitored by humans, with explainable actions, secure access, and robust data protection. Risks such as bias, hallucination, and data leakage must be mitigated through thoughtful implementation of traceability mechanisms, retrieval-augmented generation techniques, and strict access controls. Trust, explainability, and security form the foundation of successful agent deployment. Maulik concluded by hinting towards a shift from traditional data teams to hybrid teams composed of data engineers and agent developers. These new roles will focus on building, training, and maintaining intelligent systems. Tredence is actively building AI agents across all key data engineering functions, including data ingestion, transformation, consumption, and documentation. The impact from a business value standpoint is immense. With significantly reduced time-to-market, companies can launch data products in weeks instead of months or years. The winners in this new era will not be those with the largest engineering teams but those who can build the smartest and most efficient AI agents.","excerpt":"“Data engineering is the single biggest bottleneck that stops businesses from getting real value out of their data.”","categories":["AI Highlights"],"tags":["AI (Artificial Intelligence)","Tredence"],"author_name":"Aditi Suresh","publish_date":"2025-06-04T16:08:00","publication_year":"2025","word_count":767,"keywords":["Go","API","AI","Scala","RAG","Aim","generative AI","analytics","Rust","Tredence","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","analytics","generative AI","Aim","RAG","R","Go","Rust","Scala","API"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/tredences-case-for-agentic-data-engineering\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10140922,"title":"This New Logic Gate Network Reduces Inference Speed to Only 4 Nanoseconds","content":"Stanford researchers Felix Petersen and Stefano Ermon, among others, have introduced convolutional differentiable logic gate networks (LGNs) with logic gate tree kernels in their latest research paper published in November 2024. This new research integrates a number of concepts from machine vision into differentiable logic gate networks. The research claims to allow for the training of deeper LGNs than before by introducing residual initialisations which help preserve information in deeper networks and prevent vanishing gradients. They also introduced logical or pooling which when combined with logic tree kernels, improved training efficiency substantially. This could be useful for applications that require both high performance and explainable decision-making, such as robotics, program synthesis, and cognitive modelling. Improved Architecture The researchers propose a CIFAR-10 architecture called ‘LogicTreeNet’ which significantly decreases the model size as compared to the SOTA while improving accuracy. Peterson also took to X to announce this research recently. “We reduce model sizes by factors of 29x-61x over the SOTA,” he said. Excited to share our NeurIPS 2024 Oral, Convolutional Differentiable Logic Gate Networks, leading to a range of inference efficiency records, including inference in only 4 nanoseconds 🏎️. We reduce model sizes by factors of 29x-61x over the SOTA. Paper: https:\/\/t.co\/Aptk35mKir pic.twitter.com\/7nwcZ8PbTB— Felix Petersen (@FHKPetersen) November 11, 2024 Further, the inference stack demonstrates that convolutional LGNs can be efficiently executed on hardware. The paper claims that their model improves accuracy on MNIST while achieving 160x faster inference speeds, but on CIFAR-10, the model improves inference speed by 1900x over the state-of-the-art. This model combines deep learning strengths with logic gates’ interpretability, enabling networks that can perform logical reasoning. It allows for end-to-end training, enhancing both power and clarity. By building networks directly from logic gates like NAND, OR, and XOR, it optimises for hardware alignment. This results in faster, more energy-efficient computation, ideal for many applications. This approach shines in resource-limited environments, like mobile devices and embedded systems, where energy and speed matter. It makes deep learning feasible in settings with limited processing power. Going Forward A very intriguing forte for future research in this domain involves applying convolutional differentiable LGNs to computer vision tasks. The researchers express that this would focus on tasks that require continuous decision-making, such as object localisation. While LGNs have demonstrated efficiency in tasks like image classification, their potential in handling continuous outputs remains largely unexplored. Investigating this could lead to more efficient and interpretable models for complex vision applications.","excerpt":"“We reduce model sizes by factors of 29x-61x over the SOTA,” said researcher, Felix Peterson.","categories":["AI News"],"tags":["AI","Deep Learning"],"author_name":"Sanjana Gupta","publish_date":"2024-11-13T18:29:33","publication_year":"2024","word_count":405,"keywords":["Go","TPU","programming_languages:R","AI","programming_languages:Go","computer vision","Aim","deep learning","Deep Learning","R","ai_applications:computer vision"],"extracted_tech_keywords":["AI","deep learning","computer vision","Aim","TPU","R","Go","programming_languages:R","programming_languages:Go","ai_applications:computer vision"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/this-new-logic-gate-network-reduces-inference-speed-to-only-4-nanoseconds\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10066251,"title":"All you need to know about Graph Contrastive Learning","content":"Self-supervised learning of graph-structured data has recently aroused interest in learning generalizable, transferable, and robust representations from unlabeled graphs. A Graph Contrastive Learning (GCL) technique first generates numerous graph views by stochastic augmentations of the input and then learns representations by comparing positive samples against negative ones. In this article, we will be discussing the theoretical aspect of Graph Contrastive Learning. Following are the topics to be covered. Table of contents What is Graph Contrastive Learning (GCL)?When to use GCL?Self Supervised LearningHow does it work?Benefits and Drawbacks of GCLApplication of GCL Let’s start with an introduction to Graph Contrastive Learning and know why it doesn’t need any human annotations. What is Graph Contrastive Learning? Graph Contrastive Learning (GCL), as the name implies, contrasts graph samples and pushes those belonging to the same distribution toward each other in embedded space. On the other hand, those belonging to different distributions are pressed against each other. The backbone of this method is Contrastive Learning. It employs proxy tasks to guide the learning of the representations. The proxy task is designed to predict any part of the input from any other observed part. Contrastive learning is self-supervised learning in which unlabeled data points are placed side by side to form a model of which points are similar and which are different. So, a general contrastive pattern which characterizes the design space of interest into four dimensions Data augmentation functions, Contrasting modesContrastive objectives Negative mining strategies Are you looking for a complete repository of Python libraries used in data science, check out here. When to use GCL? As GCL aims to learn the graph representation with the help of contrastive learning. It is used where there is a need to understand the low dimensional nodes embeddings both structural and attributive information. For instance, we are developing an AI software for deaf people as a sign language and we have data related to the graphical representation of hand gestures. Now each node of the representation has to be learned to understand the sign language and reply in signs. These nodes are called low dimensional because they have fewer features that are needed to be learned as compared to the observations. So to learn the structural and attributive information of each graphical representation we will use GCL as a pre-training model and for the target model, we can use NN. Since Graph Contrastive Learning uses Contrastive Learning which is a self-supervised technique, let’s understand the basics of self-supervised learning. Self Supervised Learning The system learns to anticipate a portion of its input from other portions of its input in self-supervised learning. In other words, a portion of the input is utilised as a supervisory signal to a predictor that is fed the remainder of the data. Neural networks may learn in two phases when using self-supervised learning: Problems with erroneous labels can be handled to initialize the weight of the networks.The process’s real task can be completed via either supervised or unsupervised learning. Why self-supervised learning is used in GCL? Due to three major points, self-supervised learning instead of supervised learning is being used. Scalability: To forecast the result of unknown data, the supervised learning approach requires labeled data. However, building models that generate good predictions may need a vast dataset. Manual data labeling takes time and is often impractical. This is where self-supervised learning comes in handy since it automates the process even when dealing with massive volumes of data.Improved capabilities: self-supervised learning offers a wide range of applications in computer vision, including colourization, 3D rotation, depth completion, and context filling. Another area where self-supervised learning excels is speech recognition.Human intervention: Labels are generated automatically by self-supervised learning without the need for human interaction. Let’s dive deeper into the concept of Contrastive Learning and how it is used in the graph domain. How does it work? The goal of GCL is to create different samples of the input graph and train the algorithm to learn the low dimensional attributes of the graphical representation which will then be categorized into positive and negative subsets. Mathematical Formulation An input graph with two primary properties: the node-set and the edge set. During training, there is no supplied class information of nodes or graphs in the context of unsupervised representation learning. Our goal is to train a GNN encoder that takes graph characteristics and structure as input and creates low-dimensional node embeddings. We may also generate a graph-level representation of the input graph that aggregates node-level embeddings for graph-oriented activities. These representations can be employed in subsequent tasks like node\/graph categorization and community discovery. Stochastic augmentations are applied at each training cycle to produce different graph perspectives from the input graph. We sample two augmentation functions to build two graph views with all potential transformation functions. We then use a common GNN encoder to obtain node representations for the two views. A readout function may be used to acquire graph representations for each graph view if desired. The contrasting mode defines a positive set and a negative set for each node embedding “V” as the anchor instance. The positive set in a pure unsupervised learning context consists of embeddings in the two augmented graph views corresponding to the same node or graph. It should be noted that when label supervision is used, the positive set may be augmented by samples from the same class. Furthermore, we may use negative mining algorithms to enhance the negative sample set by taking into account the relative similarity of negative samples. Finally, we rate these defined positive and negative pairings using a contrastive objective. So, the whole process could be divided into four major parts. The goal of data augmentation is to produce congruent, identity-preserving positive samples from a given network. The majority of GCL work entails bi-level augmentation approaches such as structural transformation and feature transformation. In the case of an anchor, contrasting modes determine the positive and negative sets at various granularities of the graph. Three contrasting forms of work are regularly used in mainstream work.The goal of Local-Local CL is to contrast node-level representations in the two viewpoints. The anchor for a node embedding is “V,” and the positive sample is its congruent counterpart in the other perspective “U”; embeddings other than “U” are naturally picked as negatives.Global-Global CL ensures that node- and graph-level embeddings are compatible. For example, if a global embedding is the anchor instance, the positive sample is all of its node embeddings throughout the network. If the readout function is expressive enough, the global-local scheme can be used as a surrogate for local-local CL. Global-Local CL ensures that the graph embeddings of the two augmented views from the same graph are consistent. The positive sample for a graph embedding “S1” is the embedding “S2” of the other enhanced view. Other graph embeddings in the batch are deemed negative samples in this situation. The space of the contrasting mode is determined by subsequent tasks. For node datasets, only local-local and global-local CL are appropriate, but graph datasets can employ all three modes. For a better understanding, refer to the preceding visual illustration.To train the encoder, contrastive objectives are employed to maximize the agreement between positive samples and the disparity between negatives.Other than the anchor instance, the embeddings of nodes or graphs are distinct to the anchor and so contemplate negatives. As a result, it is logical to conclude that greater batch\/sampling sizes are required for successful Contrast Learning (CL) to include more negatives to deliver more useful training signals. This is referred to as the negative mining strategy. Benefits and Drawbacks of GCL Let’s have a look at the benefits and drawbacks of Graph Contrastive Learning. Benefits It is self-supervised learning so there is no need for any human annotations.By maximizing mutual information, GCL is used to generate multiple views of the same graph.Does not depend on the quality of data. It can operate on low-quality data. Drawbacks Due to data augmentation, information loss incurs.High probability of adversarial attack due to the discrete nature of edges and nodes in graphs.Speculative overfitting of the learner. Application of GCL In the discipline of botany, GCL aids in the understanding of the molecular structure of various specimens, which may then be categorized further.Graph Contrast Learning aids in the addition of multiple colors to a certain picture.Context-aware prediction is aided by GCL. It aids in the completion of the context if certain elements are lacking.GCL might be utilised as a target’s pre-training model. This is referred to as knowledge transmission. Because it can operate with fewer amounts of data.GCL is used to determine the connection between several patches in a picture. Closing verdicts Graph Contrasting Learning is self-supervised learning which augments the data and based on the augmentation it learns different attributes about the data at root levels. In this article, we have discussed the functionality of GCL which helps to understand its back processes and where to implement this kind of learner. References Read what the wiki says about Self Supervised LearningWhite paper on Contrastive SSL","excerpt":"The Graph Contrastive Learning aims to learn the graph representation with the help of contrastive learning.","categories":["AI Trends"],"tags":["AI Applications","contrastive learning","data augmentation","graphs","mining","self supervised learning"],"author_name":"Sourabh Mehta","publish_date":"2022-05-04T13:02:10","publication_year":"2022","word_count":1505,"keywords":["data science","data augmentation","Go","AI","neural network","AI Applications","Scala","computer vision","Python","GRU","Aim","mining","graphs","R","contrastive learning","self supervised learning"],"extracted_tech_keywords":["AI","neural network","computer vision","data science","Aim","Python","R","Go","Scala","GRU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/all-you-need-to-know-about-graph-contrastive-learning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10041734,"title":"An Ode To Open Source: We Can&#8217;t Thank The Community Enough","content":"“Open source was reportedly worth at least $143 million of Instagram’s $1 billion acquisition by Facebook in 2012. ” Just like a thriving, busy physical world, the rise of technology has opened up an equally busy digital world. And, open source software (OSS) makes up for all the digital infrastructure we have today. Every internet user benefits from the coding capabilities of an unknown developer contributing remotely. There are billion dollar companies, which rely on these open source projects. But, do we thank the open source community enough? Today, more than 85 per cent of India’s internet runs on open source software. The State Bank of India, the IRCTC and LIC India are all examples of organisations that rely on open source frameworks. Nadia Eghbal, a writer and researcher, called open-source the digital world equivalent of ‘roads and bridges’. Still, individuals tend to take the importance of open source for granted. In line with the analogy used above, we see bridges and know they are important, but open source code—their digital counterparts—is invisible, and their value goes unnoticed. A majority of large companies and proprietary services today rely on open source code. For example, social media tools such as Instagram allows various interactions such as uploading photographs or stories or DMing other users. When a user carries these actions out, the part of Instagram’s software responsible for storing your data remembers your profile and then posts the image on your feed. Such interactions require building a lot of code. Companies such as Facebook (which owns both its Facebook platform and Instagram) use free, public code available on the internet. Mike Krieger, one of Instagram’s founders, spoke about how it was better to ‘borrow instead of building wherever possible’, citing that leveraging open source was a time-saving practice that allowed developers to focus on building their product. Open source was reportedly worth at least $143 million of Instagram’s $1 billion acquisition by Facebook in 2012. The demand for OSS Additionally, due to permissive licenses, Facebook and Instagram do not need to pay for this code they have complete access to. For example, Appirater is a library developed by Arash Payan in 2009 which allows easy reminders to iPhone users to rate mobile apps. Despite having a widely known use case, Payan made no income from this project. Charging individuals for open source code would significantly limit their adoption. It is also difficult to charge a fee for something that developers build in their spare time. Additionally, a lot of open-source platforms are made to solve problems developers face with other software. For instance, Webpack was started by Tobias Koppers in 2012 when he could not find a suitable code optimisation tool for his master’s thesis in computer science. He was not happy with how other devices handled code-splitting—which Webpack allows. Today, Webpack is widely used and appreciated by users. Pete Hunt, former engineering manager at Instagram, spoke out about the software in 2014 and said that the social media platform (which is among the world’s most popular ones) ‘relied on webpack’. Thus, open-source software works like public goods would with their contribution to technology and information. It then makes sense for them to remain free of charge. Not charging for software has transformed our society. Due to open source’s easy access, new developers are generated every day. The Bureau of Labour Statistics (BLS) predicted a 22 per cent rise of software developers from 2012 to 2022, which is much faster than other occupations. Free software has made it substantially more straightforward and cheaper to build other software. Free software—or its components—allow tech startups to get off the ground without spending an excessive amount of money. Imgur, an image-sharing site, reportedly cost $7, which was used to buy the domain name. Open-source software has also democratised coding by making it possible for anyone with laptops and an internet connection to learn to code—regardless of their socioeconomic status. All of this has undoubtedly played into the skyrocketing levels of technological innovation that we see today. Maintaining open source infrastructure Nadia Eghbal sees OSS as ‘the roads and bridges’ of the digital world. Digital infrastructure is not as expensive as its physical counterpart to build, but it is more tedious to maintain. This is because digital infrastructure needs to be updated frequently. However, it does not have a governing body to take care of it, as physical infrastructure does. Instead, open-source infrastructure is maintained in different ways. Some ways in which open-source software can originate and be maintained are by starting within a company or entering a new business altogether. Such frameworks allow these projects to solve a significant problem faced by open source software today: funding. Funding is a significant problem primarily due to open source software being invisible components of major proprietary applications. Working with companies can go two ways: funding via corporate sponsorships or being acquired by the company. Companies that rely on open source libraries and understand the importance of keeping them secure and well-maintained are more inclined to hire such open-source software maintainers to continue working on their projects. Tobias Kopper decided to work on Webpack full-time after being granted corporate sponsorship and now works on Webpack for Vercel, a company. Individual, independent contributors maintain many projects. These people are not paid for this work but choose to volunteer because they either wish to continue supporting it or want to cultivate their reputation in the developing community. Many cases in the latter box are computer science students who want to show off or practice their skills. Furthermore, many people prefer keeping their work independent and not monetising it. Samuel Colvin, for example, maintains pydantic—a popular Python library. The library helps creators make sure they have inputted the correct type of data in their applications. Colvin has said that he does not see a straightforward way to monetise the library due to its reach and usefulness. Many such projects then rely on means like crowdfunding or receiving help from venture capitalists. It is, however, essential to note that individual contributors, crowdfunding and corporate sponsors are not permanent methods to maintain OSS. Another alternative is to open the software as its own startup then. Docker, for example, a company that helps software applications run inside containers, started as a project within dotCloud (a platform-as-a-service company). Docker, however, did so well that its founders decided to make it the company’s main focus. The project was outsourced in 2013 and went from making less than $10 million in 2014 to $180 million in 2016. Big Tech and open source Until the 1990s, companies made sure to guard their code, and the open-source initiative was thought to be this ‘big tech battler’. In addition to using external open source codes for their work, many large companies today both fund and develop their own open-source libraries. Facebook recently made its most extensive language library available for free to help developers make better translation models.  Opening up, however, makes a lot more sense for Big Tech. These projects for open source helps build the company’s image in developer circuits and thus incentivises these developers to use other company resources for their work. It also allows companies to crowdsource improvements. Such improvements could make the company even more influential. Why should you care about maintaining open source? Open source code might be free, but ignoring its maintenance can prove very expensive. Nadia Eghbal describes these as direct and indirect costs. Direct costs include direct security concerns such as undetected bugs and vulnerabilities. Many open-source projects are not created and maintained by companies that offer support, such as actively patching flaws or paying for audits. Larger organisations that provide open source code still deploy mechanisms like scanning code for vulnerabilities and fixing them in the project’s entirety. But smaller projects and firms do not enjoy this protection. In an ironic twist, the earlier mentioned democratisation of coding has also led to developers with significantly varied skill sets. Many newly trained programmers wish to test their code but do not yet know how to secure it. Indirect costs comprise longer-term maintenance costs such as loss of qualified labour and slower growth and innovation. The loss of skilled labour stems from the lack of compensation in many open source projects or volunteers not having time (due to other commitments). This can lead to direct costs for the software. For example, in 2013, a vulnerability was detected in RubyGems.org but was exploited before the group of volunteers taking care of the platform could actually get to discussing it. Following this, the entire server had to be built from scratch. As per Prashanth Kaddi, Partner at Deloitte India, ‘Open source data analytics offerings have managed to gain greater mindshare in the newer crops of academics, researchers, and developers due to ease of accessibility, faster release upgrades, the support of sprawling online communities, and ease of integration with popular cloud providers.’ It is, therefore, necessary to understand the importance of maintaining open-source software. With the open-source ecosystem becoming increasingly complex, ignoring these vital platforms will only make maintaining them more complicated in the future. Open source code is here to stay—here’s hoping its future involves improved security and reliability!","excerpt":"“Open source was reportedly worth at least $143 million of Instagram’s $1 billion acquisition by Facebook in 2012. ” Just like a thriving, busy physical world, the rise of technology has opened up an equally busy digital world. And, open source software (OSS) makes up for all the digital infrastructure we have today. Every internet […]","categories":["IT Services"],"tags":[],"author_name":"Mita Chaturvedi","publish_date":"2021-06-14T12:00:00","publication_year":"2021","word_count":1536,"keywords":["Go","API","AWS","AI","docker","Git","RAG","Python","analytics","R"],"extracted_tech_keywords":["AI","analytics","RAG","AWS","docker","Python","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/it-services\/an-ode-to-open-source-we-cant-thank-the-community-enough\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":18925,"title":"Airtel Looks Forward To Introducing AI Based Services To Its Customers, Partners With Amdocs","content":"In an interesting development around the use of artificial intelligence and machine learning in the telecom sector, Bharti Airtel, one of the biggest global telecommunications services company has partnered with Amdocs, software solutions provider. It has created an innovation foundry to bring new services to Airtel’s customers in India. With this partnership that includes a long term contract, Airtel intends to bring cutting edge technology such as artificial intelligence based services to its customers in India. Amdoc plans to deploy machine learning and advanced AI based technologies across Airtel’s business, by leveraging its innovation centers, delivery expertise and startup ecosystem. How will the collaboration help? Amdocs plans to leverage its innovation centres, delivery expertise and its ecosystem of startups to deploy machine learning and advanced artificial intelligence based technologies across multiple lines of Airtel’s business to self heal operational issues.  Enabling a seamless customer experience, the collaboration also intends to introduce smartbots into digital channels and activating new services. Gopal Vittal, MD and CEO (India and South Asia), Bharti Airtel India, was quoted as saying that with Amdoc’s sharp focus on telecoms and their capability of bringing cutting edge innovation, Airtel is all set to witness digital transformation and help in delivering a world class experience to their customers. “This program is at the epicenter of two innovation waves: India’s fast growing and dynamic market and the communications industry’s amazing pace of change. Airtel India is a pioneer and innovator, and our partnership will be an innovation engine to the industry and drive our joint success,” said Eli Gelman, President and Chief Executive Officer, Amdocs. Amdocs has been providing its services to country’s leading service providers and with this partnership it sees it to be an innovation engine to the industry and drive joint success. It was only recently that Airtel has rolled out “Project Next”- its digital innovation programme, aimed at transforming customer experience across its services and touch points. It complements Airtel’s massive investments in building a future ready network under “Project Leap”. Airtel is looking out to launch several exciting innovations to enhance interactivity of Airtel customer experience. The telecom major had earlier deployed Amdocs services for billing.","excerpt":"In an interesting development around the use of artificial intelligence and machine learning in the telecom sector, Bharti Airtel, one of the biggest global telecommunications services company has partnered with Amdocs, software solutions provider. It has created an innovation foundry to bring new services to Airtel’s customers in India. With this partnership that includes a […]","categories":["AI News"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2017-11-10T06:11:39","publication_year":"2017","word_count":361,"keywords":["Go","machine learning","artificial intelligence","AI","ML","Git","RAG","Aim","ViT","R"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","Aim","RAG","R","Go","Git","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/airtel-looks-forward-introducing-ai-based-services-customers-partners-amdocs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10058081,"title":"How I became a data science entrepreneur: Venkat Raman of Aryma Labs","content":"Entrepreneurship is a difficult journey. It comes with a high degree of uncertainty and unpredictability—that makes it different from the conventional 9 to 5. Though many of us dream of building something that we can truly call our own, very few can make that transition successfully. There is intense competition among start-ups in the analytics and AI space due to the huge opportunities the sector provides. To motivate more and more people to take a plunge towards building their own ventures in this domain, we will look at some successful entrepreneurial journeys. Today, we look at the journey of Venkat Raman, co-founder of Aryma Labs, a data science consulting firm, and understand what it is to build an analytics firm in the present scenario. Never easy quitting a well-paying job Raman comes with eleven years of industry experience in software engineering, industrial engineering, marketing and advertisement, with a strong background in statistics. “It is never easy quitting a well-paying job. My co-founder Ridhima Kumar had started Aryma Labs a year before I joined,” adds Raman. The motivation behind starting Aryma Labs came from the realisation that many companies have huge troves of data but don’t know how to extract actionable insights from it. There is also a shortage of talent who know how to apply data science to business problems correctly. This void led to the creation of Aryma Labs. Aryma Labs is a data science start-up that specialises in three core areas of market effectiveness, time series forecasting and NLP. “We aim to deliver efficient machine-learning-powered solutions to clients that genuinely deliver value in the long run. We provide these solutions across different domains and industries – CPG, manufacturing, e-commerce, supply chain, travel and hospitality,” added Raman. A huge dearth of good data scientists Raman feels that the biggest challenge he has experienced along with Kumar is hiring and the current COVID-19 situation (which has impacted many sectors severely, globally). He states, “We have the numbers, but the quality is simply not there. Picking a good data scientist is literally like hunting for a needle in a haystack.” As data scientists form the heart of any analytics company, Raman had to dig deep to solve the issue of hiring quality data scientists. The company has streamlined its process to allow for quicker interviews. They hire people with good aptitude and then train them on statistical concepts. Sometimes, this also calls for unlearning the wrong things they learnt. Once they have suitably trained the interns and employees, they are put on real projects. First year is the biggest litmus test of an entrepreneur’s grit As a young founder, Raman puts down some important lessons he has learnt in the last two and half years of starting up. They are: Patience – Patience is a virtue in this day and age of instant result expectations. As a young co-founder of a bootstrap start-up, he realised that patience is key.Survive and never give up – The first year or so is the biggest litmus test of an entrepreneur’s grit and resolve. Once the chasm is crossed, things get progressively easier. Imperative to know your craft If someone wants to start a data science consulting company, having sound knowledge of statistics\/machine learning and solid domain expertise is mandatory for the venture to take off. “The team grows around you much like a crystal grows around the first crystal particle,” adds Raman. Data science as a field will move towards a policy of ‘Do more with less” Raman feels that given the impetus on having a lower carbon footprint, data science as a field will move towards a policy of ‘Do more with less”. It means that we will go back to core statistical techniques. Statistics has a history of emphasising the parsimony of models. In the 20th century, this emphasis was more due to technical constraints. But in the future, it will be for reasons such as carbon footprint and data privacy, among others.","excerpt":"We look at the journey of Venkat Raman, co-founder of Aryma Labs, a data science consulting firm, and understand what it is to build an analytics firm in the present scenario.","categories":["AI Features"],"tags":["Entrepreneur","entrepreneurship","Interviews and Discussions","journey","Startups"],"author_name":"Sreejani Bhattacharyya","publish_date":"2022-01-11T18:00:00","publication_year":"2022","word_count":661,"keywords":["data science","Go","machine learning","programming_languages:R","AI","ML","journey","NLP","Aim","analytics","entrepreneurship","Startups","Entrepreneur","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","ML","NLP","data science","analytics","Aim","R","Go","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-i-became-a-data-science-entrepreneur-venkat-raman-of-aryma-labs\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":36339,"title":"A Look At The Evolution Of BI Platform Over The Years","content":"Business Intelligence tools are crucial to the growth of a company and are a go-to tool for gaining insights and making efficient data-driven decisions. Here is the detailed evolution of this BI platforms that has immense significance today. “BI is about providing the right data at the right time to the right people so that they can take the right decisions” – Nic Smith with Microsoft BI Solutions Marketing Looking At The Start Of BI It was in 1865 when the term BI first came into existence. Cyclopaedia of Commercial and Business Anecdotes, Richard Devens used the term “Business Intelligence” to describe how a financier Sir Henry Furnese had successfully beaten his competitors and profited by understanding the market and the conditions surrounding it better than they did. More recently, in 1958, an article was written by an IBM computer scientist named Hans Peter Luhn, describing the potential of gathering Business Intelligence (BI) through the use of technology. Avin Jain of Big Data BizViz at Cypher 2018 said that B of BI comes before A of AI as far as analytics is concerned. We have moved from the desktop based reports to web reports and from web reports to the BI world, which is from where we move to the big data and AI world. In the present world, there are 50 plus BI products across the world but the platform story is not the same as what it was before 2007. Evolution Of BI In 1968, only individuals with specialised skill-set could translate data into usable information. At this time, data from multiple sources was normally stored in silos, and research was typically presented in a fragmented, disjointed report that was open to interpretation. Edgar Codd recognised this as a problem, and published a paper in 1970, altering the way people thought about databases. His  proposal of developing a “relational database model” gained tremendous popularity, and was adapted worldwide. Business intelligence is understood and evolved from the decision support systems (DSS) that began in the 1960s and developed throughout the mid-1980s even spread out to the countries like Holland, Flanders, France, and Germany. Now BI is evolving in every industry throughout the world but in olden times, most of the business people found it very difficult to analyse data. Let’s rewind and go back to olden times on how business intelligence has been evolved from then to now. According to the History of Business Intelligence, the term “Business Intelligence” first appeared in Richard Millar Devens’ Cyclopaedia of Commercial and Business Anecdotes in the year of 1865. After that Sir Henry Furnese gained profit by receiving and acting upon information about his environment, prior to his competitors. BI in 1980s 1980s, was when Big Data experienced some major changes. Data began to be organised into warehouses and used to access and manage data in one place. Bu these data warehouses were very technical in nature and required an expensive IT staff specialised in BI to run the reports. There was not much awareness about how to make use of this new technology, among businesses. There would also be a longer run time for the reports because of the limitations of computers at the time. During the 1980s, BI still had a long way to go. BI in 1990s Known as the “BI 1.0”, this era of BI was when BI became popular to the business world and the adoption increased. But since the technology grew during this era, BI solutions were very expensive. BI during this era was not very flexible and affordable. Moreover, it could take many days to ask a BI question which is a problem in BI where  it is only by asking the right data analysis questions that you will get something valuable from your data. Image credit: datameer BI in 2000 This was the era when technologies such as real-time processing came into existence for making decisions based on recent information. BI was taken more seriously as businesses were realising its importance. Dashboards came into existence in abundance, making an easy access of the data to help make better-informed business decisions. This made the business to access more than a static report. This was a time when BI platform advanced a lot. Self-service BI tools like Tableau, QlikView and others hit the market. These players put the analytics in the hands of the business analysts, but they were still dependent on IT to get the data to a place where they could then work with it in their self-service tool. They were limited only to structured data and limited in big data use cases to aggregated, summarised, pre-processed data. More and more companies understood the true value of BI and realised that they could help them to succeed in business. Emergence of these simple tools let the non-technical user to intuitively generate reports and data through clicking and a drag and drop interface rather than typing into a command line. This is when real-time processing had emerged, giving an advantage of an up to date information to be used in decision-making, unlike the previous times when databased would have to be updated in batches calling for a time lag. BI Today BI has played a major role in the AI-driven world and enables businesses with a lot of ease in their functioning and transform their businesses. It is important for a streamline low and predictive abilities. The future of BI would look more automated and heavily used. It promises to help in the progression of the modern businesses.","excerpt":"Business Intelligence tools are crucial to the growth of a company and are a go-to tool for gaining insights and making efficient data-driven decisions. Here is the detailed evolution of this BI platforms that has immense significance today. “BI is about providing the right data at the right time to the right people so that […]","categories":["AI Features"],"tags":[],"author_name":"Disha Misal","publish_date":"2019-03-14T12:32:09","publication_year":"2019","word_count":923,"keywords":["big data","Go","business intelligence","AI","ML","RAG","analytics","GAN","R","data warehouse"],"extracted_tech_keywords":["AI","ML","analytics","RAG","R","Go","big data","data warehouse","GAN","business intelligence"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-look-at-the-evolution-of-bi-platform-over-the-years\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":26595,"title":"How The Tech Community Is Leading The War Against Development Of Lethal Autonomous Weapons","content":"Of late, there have been many arguments against killer robots and lethal autonomous weapons but none have been more potent than the recent news about more than 2,400 technology leaders calling for an open ban on the development of lethal autonomous weapons. One of the biggest voices in this debate has been Tesla’s Elon Musk, who has rallied against the use of killer robots. In a recent conference in Stockholm more than 2,400 individuals and 150 companies from 90 different countries vowed to play no part in the construction, trade, or use of autonomous weapons in a pledge signed on Wednesday at the 2018 International Joint Conference on Artificial Intelligence in Sweden. Max Tegmark, president of the Future of Life Institute and one of the supporters of the ban against development of LAWS, “AI has huge potential to help the world — if we stigmatise and prevent its abuse. AI weapons that autonomously decide to kill people are as disgusting and destabilising as bioweapons, and should be dealt with in the same way”. According to a statement released by the body, the decision to take a human life should never be delegated to a machine since LAWS engaging targets without human intervention – would be dangerously destabilising for every country and individual. If autonomous weapons are the Kalashnikovs of tomorrow — the tech community should abandon it AI military tech has raised concerns in the past few months wherein Google employees reportedly threatened to quit over military drone project. Similarly, employees at Microsoft posted a letter to CEO over the use of technology by law enforcement agencies. Now as AI researchers and communities across the globe vow not to invest in the development of LAWS, the debate clearly puts the onus on the tech community to counter the proliferation and threat posed by autonomous weapons. Leading organisations such as XPRIZE Foundation, Element AI, Swedish AI Society, European Association for Artificial Intelligence to Swedish AI Society to GoodAI among others signed the pledge against use of AI. Besides the tech community has realised the urgent opportunity and necessity to join the debate which so far involved policymakers and government leaders. It’s not just the ethical questions around LAWS that have raised eyeballs, there are deeper concerns about autonomous weapons being hacked in addition to them ending up in the hands of terrorists. As part of the pledge, tech companies are vowing not to participate or support the development, manufacture and trade of lethal autonomous weapons. What has sparked more concern is that the international community, including policymakers, government leaders and tech community does not have any governance framework or technical tools to prevent this into a spiraling arms race. Understanding the three levels of autonomy in LAWS Weapons with human-in-the-loop systems: Most LAWS that are deployed today require a human-in-the-loop element which means that a human command is required for the target and deployment of force. For example, Israel’s Iron Dome system is an example of this level of autonomy. Weapons with human-on-the-loop system: This gives the autonomous weapon partial autonomy wherein the robot can supersede human’s decision in the selection of target. Case in point includes South Korea that has planted a sentry robot along the demilitarised zone joining North Korea. According to a report from Human Rights Watch, United States, Russia, China, Israel, South Korea and the UK possess partially autonomous weapons systems such as armed drones. Fully autonomous weapons: Fully autonomous weapons operate without requiring any human input.  For example, a report in IEEE hints that drones are by and large remotely controlled, but hobbyist drones are becoming increasingly autonomous. Some of the new models of drones can easily navigate to a fixed target on their own and even track moving objects. Increasingly, there are reports about small drones being equipped with facial recognition technology which are leveraged for searching people autonomously. Here are some of the reasons for concerns: Tech companies and governments are manufacturing micro-drones which in turn can be used for spying, surveillance purposes As of now, there are no effective guidelines or policy framework against the defense of drones There is no defense framework to keep military-grade weapons from falling into the hands of terrorists or being misused Rise of counter-force weapons In his book Army of None: Autonomous Weapons and the Future of War, Paul Scharre, also a Senior Fellow and Director of the Technology and National Security Program at the Center for a New American Security mentioned that governments across the world, Russia, US and China are building counter-force weapons to fight other militaries. This raises even more concerns since counter-force autonomous weapons spark concerns about risk and the factor of controllability.","excerpt":"Of late, there have been many arguments against killer robots and lethal autonomous weapons but none have been more potent than the recent news about more than 2,400 technology leaders calling for an open ban on the development of lethal autonomous weapons. One of the biggest voices in this debate has been Tesla’s Elon Musk, […]","categories":["AI Features"],"tags":["autonomous systems","autonomous weapons"],"author_name":"Richa Bhatia","publish_date":"2018-07-20T13:23:37","publication_year":"2018","word_count":780,"keywords":["Go","artificial intelligence","AWS","AI","cloud_platforms:AWS","programming_languages:R","RAG","GAN","autonomous systems","autonomous weapons","AI research","R"],"extracted_tech_keywords":["AI","artificial intelligence","RAG","AWS","R","Go","GAN","AI research","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-the-tech-community-is-leading-the-war-against-development-of-lethal-autonomous-weapons\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":50835,"title":"2020: Five Artificial Intelligence Trends For Engineers And Scientists","content":"1. Workforce skills and data quality barriers start to abate As AI becomes more prevalent in industry, more engineers and scientists – not just data scientists – will work on AI projects. They now have access to existing deep learning models and accessible research from the community, which allows a significant advantage than starting from scratch. While AI models were once majority image-based, most are also incorporating more sensor data, including time-series data, text and radar. Engineers and scientists will greatly influence the success of a project because of their inherent knowledge of the data, which is an advantage over data scientists not as familiar with the domain area. With tools such as automated labeling, they can use their domain knowledge to rapidly curate large, high-quality datasets. The more availability of high-quality data, the higher the likelihood of accuracy in an AI model, and therefore the higher likelihood for success. 2. The rise of AI-Driven systems increases design complexity As AI is trained to work with more sensor types (IMUs, Lidar, Radar, etc.), engineers are driving AI into a wide range of systems, including autonomous vehicles, aircraft engines, industrial plants, and wind turbines. These are complex, multidomain systems where behavior of the AI model has a substantial impact on the overall system performance. In this world, developing an AI model is not the finish line, it is merely a step along the way. Designers are looking to Model-Based Design tools for simulation, integration, and continuous testing of these AI-driven systems. Simulation enables designers to understand how the AI interacts the rest of the system. Integration allows designers to try design ideas within a complete system context. Continuous testing allows designers to quickly find weaknesses in the AI training datasets or design flaws in other components. Model-Based Design represents an end-to-end workflow that tames the complexity of designing AI-driven systems. 3. AI becomes easier to deploy to low power, low cost embedded devices AI has typically used 32-bit floating-point math as available in high-performance computing systems, including GPUs, clusters, and datacenters. This allowed for more accurate results and easier training of models, but it ruled out low cost, low power devices that use fixed-point math. Recent advances in software tools now support AI inference models with different levels of fixed-point math. This enables the deployment of AI on those low power, low-cost devices and opens up a new frontier for engineers to incorporate AI in their designs. Examples include low-cost Electronic Control Units (ECUs) in vehicles and other embedded industrial applications. 4. Reinforcement Learning moves from gaming to real-world industrial applications In 2020, reinforcement learning will go from playing games to enabling real-world industrial applications particularly for automated driving, autonomous systems, control design, and robotics. We’ll see successes where Reinforcement Learning (RL) is used as a component to improve a larger system. Key enablers are easier tools for engineers to build and train RL policies, generate lots of simulation data for training, easy integration of RL agents into system simulation tools and code generation for embedded hardware. An example is improving driver performance in an autonomous driving system. AI can enhance the controller in this system by adding an RL agent to improve and optimize performance – such as faster speed, minimal fuel consumption, or response time. This can be incorporated in a fully autonomous driving system model that includes a vehicle dynamics model, an environment model, camera sensor models, and image processing algorithms. 5. Simulation lowers a primary barrier to successful AI adoption – lack of data quality Data quality is a top barrier to successful adoption of AI – per analyst surveys.  The simulation will help lower this barrier in 2020. We know training accurate AI models requires lots of data. While you often have lots of data for normal system operation, what you really need is data from anomalies or critical failure conditions. This is especially true for predictive maintenance applications, such as accurately predicting remaining useful life for a pump on an industrial site. Since creating failure data from physical equipment would be destructive and expensive, the best approach is to generate data from simulations representing failure behavior and use the synthesized data to train an accurate AI model. The simulation will quickly become a key enabler for AI-driven systems.","excerpt":"1. Workforce skills and data quality barriers start to abate  As AI becomes more prevalent in industry, more engineers and scientists – not just data scientists – will work on AI projects. They now have access to existing deep learning models and accessible research from the community, which allows a significant advantage than starting from […]","categories":["AI Features"],"tags":["AI (Artificial Intelligence)","Machine Learning"],"author_name":"Paul Pilotte","publish_date":"2019-11-28T14:00:00","publication_year":"2019","word_count":713,"keywords":["Go","API","AWS","AI","cloud_platforms:AWS","programming_languages:R","Machine Learning","programming_languages:Go","deep learning","data quality","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","deep learning","AWS","R","Go","API","data quality","cloud_platforms:AWS","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/2020-five-artificial-intelligence-trends-for-engineers-and-scientists\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":27466,"title":"Is Flipkart Acquiring Liv.AI To Compete With Amazon&#8217;s Alexa?","content":"Conversational platforms are getting popular each day and now numerous companies are venturing into voice-based assistants territory. This area is currently dominated by Apple (Siri), Amazon (Alexa) and Google Assistant. Now, a new report by a leading financial daily has suggested that Indian e-tail giant Flipkart is set to acquire artificial intelligence-based company Liv.ai, to compete with these tech giants. The newspaper also added that the talks were in the “final stage” and that the deal would be closed at around $40 million. Liv.ai uses deep learning to develop products that can take over tasks that are repetitive and time-consuming. Their website describes their vision as “to ease the tedium of daily life and create a world where people are free to devote their energy to things that matter, where their minds are free to innovate and make a real difference.” The Bengaluru-based startup has speech recognition in all major India languages including Hindi, Bengali, Punjabi, Marathi, Gujarati, Kannada, Tamil and Telugu. Over the years the team has created state-of-the-art speech recognition and natural language understanding system, along with voice-based keyboards, speech synthesis and language understanding. Like Alexa and Amazon, Flipkart will probably use Liv.AI to instruct the devices to buy groceries, pay bills, or book tickets in the language users are comfortable with. This news comes just after US retail giant Walmart announced this weekend that they had completed the $16 billion investment in Flipkart. Now, as things stand, Walmart holds a 77 percent stake in Flipkart. To achieve this, they bought out a number of prior investors in the process. Many industry insiders have suggested that this move signifies the next level of the tough competition between two retail behemoths — Amazon and Walmart.","excerpt":"Conversational platforms are getting popular each day and now numerous companies are venturing into voice-based assistants territory. This area is currently dominated by Apple (Siri), Amazon (Alexa) and Google Assistant. Now, a new report by a leading financial daily has suggested that Indian e-tail giant Flipkart is set to acquire artificial intelligence-based company Liv.ai, to compete […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Alexa","Amazon","Flipkart","virtual assistants","Walmart"],"author_name":"Prajakta Hebbar","publish_date":"2018-08-21T05:44:27","publication_year":"2018","word_count":286,"keywords":["Go","artificial intelligence","programming_languages:R","AI","Flipkart","Amazon","virtual assistants","programming_languages:Go","deep learning","Walmart","Alexa","R","AI (Artificial Intelligence)","startup"],"extracted_tech_keywords":["AI","artificial intelligence","deep learning","R","Go","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/is-flipkart-acquiring-liv-ai-to-compete-with-amazons-alexa\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":1884,"title":"Panasonic partners with Tata Elxsi to set-up an R&#038;D center in Bengaluru","content":"Mr. MadhukarDev, CEO and Managing Director of Tata Elxsi shaking hands with Mr. Tetsuro Homma President of Panasonic Panasonic recently revealed its plan to partner with Tata Elxsi in setting up an R&D unit in Bengaluru. This will help the consumer electronics major to fortify its appliance business in domestic and global markets. Tetsuro Homma, President of Appliances Company & Senior MD, Panasonic Corporation comments, “The launch of the R&D center in Bengaluru will help to drive Panasonic’s business in the ISAMEA (India, South Asia, Middle East and Africa) region, with a medium and long-term strategic view. The establishment has been constructed to house two centers, a Design Division, which would be based at Panasonic’s Jhajjar unit, and an Offshore Development Division based at Bengaluru. The refrigerator plant at Jhajjar was only launched recently, for which Panasonic put in a Rs 115 crore investment. The operations at this plant will start from January 2018, and will be followed by sales, commencing from April 2018. The Jhajjar unit will primarily focus on factory engineering, product design and QC function for products such as air conditioner, washing machine, and refrigerators, while the Bangalore branch will be responsible for developing technologies such as artificial intelligence and robotics. It will help the firm carry out its global operations, besides extending support for product designing in local product manufacturing. “Panasonic will be able to accelerate product innovations in newer categories with the newly established R&D center. The move will also facilitate the growth of the appliances industry,” extols Homma. Artificial Intelligence can help in improving the utility and performance of home appliances, for individual users over time. This is based on the data collated everyday about an individual consumer’s usage habits and patterns. Work assist systems, robotic household appliances, and interactive machine technology has been part of Panasonic’s AI-based ecosystem for some time now. With this move, there will be an increased focus on driving R&D in India on this platform. With this move, Panasonic and Tata Elxsi intends to utilize IoT solutions, and push the next generation of home appliances from India to the global market. “Panasonic India, with its Japanese reputation of technology and innovation, is an important partnership for us,” comments Madhukar Dev, CEO and MD, Tata Elxsi. The partnership also represents Panasonic’s commitment towards the Make-in-India initiative.","excerpt":"Panasonic recently revealed its plan to partner with Tata Elxsi in setting up an R&D unit in Bengaluru. This will help the consumer electronics major to fortify its appliance business in domestic and global markets. Tetsuro Homma, President of Appliances Company & Senior MD, Panasonic Corporation comments, “The launch of the R&D center in Bengaluru […]","categories":["AI News"],"tags":["Artificial Intelligence India","manufacturing India","Robotics India"],"author_name":"Дарья","publish_date":"2017-04-13T04:22:01","publication_year":"2017","word_count":386,"keywords":["Go","artificial intelligence","programming_languages:R","AI","R","manufacturing India","innovation","Artificial Intelligence India","programming_languages:Go","ai_applications:robotics","Robotics India"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","innovation","programming_languages:R","programming_languages:Go","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/panasonic-partners-tata-elxsi-set-rd-center-bengaluru\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10143862,"title":"Meta’s Apollo Models Set New Benchmarks for Video Understanding in AI","content":"Meta AI and Stanford have introduced Apollo, a family of video-based large multimodal models (LMMs) designed to efficiently and accurately understand video content. Apollo aims to bridge the gap between text-to-image models and video comprehension, addressing challenges posed by high computational demands and technical limitations. Apollo models excel at video tasks by addressing key challenges, including how videos are sampled, encoded, and trained. This paper gains significance,  especially in light of OpenAI co-founder Ilya Sutskevar’s recent talk on pre-training hitting a wall. Apollo leverages scaling consistency to reduce reliance on large datasets and models while improving task-specific performance. “We discovered scaling consistency, which allows us to design effective solutions using smaller models and datasets, reducing computational overhead,” the researchers explained in the paper. Two key improvements make AI models better at understanding videos. First, fps sampling selects video frames at a steady rate, which works better than picking frames evenly. Second, combining SigLIP-SO400M (which focuses on clear image details) with InternVideo2 (which captures motion and timing) helps the model understand both still visuals and movements in videos. Smaller Models with Superior Performance Apollo-3B outperforms larger 7B models with a score of 68.7 on the MLVU benchmark. Meanwhile, Apollo-7B sets a new standard in its category, achieving 70.9 and even surpassing some 30B models. The team also introduced ApolloBench, a faster, more efficient evaluation tool for video understanding, reducing test times by 41 times. “Our results prove that smart design and training can deliver top performance without relying on massive model sizes,” the researchers said. Apollo marks a significant leap in video AI, opening doors to applications like content analysis and autonomous systems.","excerpt":"Apollo models excel at video tasks by addressing key challenges, including how videos are sampled, encoded, and trained.","categories":["AI News"],"tags":["AI (Artificial Intelligence)","AI Video Generation Models","Meta"],"author_name":"Aditi Suresh","publish_date":"2024-12-18T16:24:06","publication_year":"2024","word_count":272,"keywords":["Go","Meta","Meta AI","OpenAI","AI","Modal","RPA","ML","AI Video Generation Models","RAG","Aim","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","ML","OpenAI","Meta AI","Aim","RAG","R","Go","RPA","Modal"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/metas-apollo-models-set-new-benchmarks-for-video-understanding-in-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10172160,"title":"Vibe Coder Gets Legal Notice From DocuSign","content":"DocuSign, a platform that provides digital signature services for documents, has sent a legal notice to Michael Luo, a developer who built a website offering a free alternative with a similar feature suite. Luo built a free e-sign tool using ChatGPT, Cursor, and Lovable—platforms that help developers write code using natural language prompts. He built a product called Inkless in two days, which lets users sign unlimited documents for free. “Just as DocuSign respects the intellectual property rights of third parties, we expect third parties to do the same with our intellectual property,” the company said in a cease-and-desist letter sent to Luo. The company said it is also concerned about how Luo was “disseminating” false and misleading statements regarding its product. This likely refers to Luo expressing his inspiration to create a free alternative to Docusign’s high costs. Source: x.com\/AzianMike “I never stole anything from DocuSign or made misleading statements,” Luo said. “They basically got scared that I created a free e-sign tool.” Luo added that despite receiving the legal notice, he is continuing to build the platform and ship new features to make the “free product even better”. DocuSign allows users to sign and send back unlimited documents for free. However, if one is collecting signatures, they can only send three documents with a free account. Over the last few months, using AI to build applications has quickly gained popularity, and platforms serving these capabilities have observed unprecedented growth rates. Andrej Karpathy, a former researcher at OpenAI, calls this phenomenon ‘vibe coding’. Aside from various safety and privacy concerns for developers creating AI-based apps, receiving legal notices from other developers or companies is now an added concern.","excerpt":"“I never stole anything from DocuSign or made misleading statements,” said the developer.","categories":["AI News"],"tags":["AI (Artificial Intelligence)"],"author_name":"Supreeth Koundinya","publish_date":"2025-06-23T11:18:24","publication_year":"2025","word_count":279,"keywords":["Go","ChatGPT","OpenAI","AI","RPA","Git","GPT","R","llm_models:ChatGPT","AI (Artificial Intelligence)","llm_models:GPT"],"extracted_tech_keywords":["AI","ChatGPT","OpenAI","R","Go","Git","GPT","RPA","llm_models:GPT","llm_models:ChatGPT"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/vibe-coder-gets-legal-notice-from-docusign\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":58511,"title":"New Hackathon To Forecast What Happens Next In COVID-19 Outbreak","content":"Analytics India Magazine’s MachineHack team feels it is crucial to predict when the outbreak may slow down, flatten or further worsen across different countries to ascertain the true nature of economic or human life cost as a result of COVID-19. We are in the midst of a worldwide outbreak of respiratory illness induced by a novel coronavirus, initially identified in China and which has now been detected in more than 110 geographies internationally. On 7th March 2020, the World Health Organisation (WHO) also proclaimed that the cumulative number of confirmed cases for COVID-19 had surpassed 100,000. The organisation urged all countries to continue their efforts to curb the disease, which has expanded to become a global pandemic. The virus outbreak has become one of the biggest threats to the global economy and financial markets. Major institutions and banks have cut down on their growth forecasts and the stock market has seen a drastic plunge worldwide. In this context, we feel it is crucial to predict when the outbreak may slow down, flatten or further worsen across different countries to ascertain the true nature of economic or human life cost as a result of COVID-19. Source: https:\/\/covid19info.live\/ The Objective Of The Hackathon In the coming weeks and months, we at MachineHack (an Analytics India Magazine initiative) along with our community members will ominously examine how the coronavirus could affect different nations. Thereby, we invite MachineHackers to predict potential COVID-19 cases across all the globe on an everyday basis. The objective of the hackathon is to gauge COVID-19 on three metrics- confirmed cases, recovered cases and death events for the next day using historical data as on a given date. As sad as it is to analyse the data around COVID-19 events, it is critical to keep a tab on the disease metrics to track the outbreak. The hackathon will be based on the data published by various agencies and the Johns Hopkins University Center for Systems Science and Engineering (JHU CSSE), which can be found here. Participate in the hackathon. A Note For Hackathon Participants The univariate time series knowledge is rendered based on all individual countries affected by COVID 19 from 22nd January 2020 onwards to the present date. Here is an example for your reference. The provided .csv file comprises the count of confirmed COVID cases across countries till 10th March 2020 for the three target variables (confirmed cases, recovered cases and death cases). The dataset would be updated daily at 00:00 UTC standard time with the prevailing forecast of the distinct target variables. It is to be noted that the published data is dynamic, and hence it will be renewed each day in a new column every day. The data in the rows will also fluctuate based on the reported changes for COVID-19 outbreak in various world geographies. The submission file from participants must contain the projected count of incidents for the next day, i.e. 11th March 2020. as per the sample_submission.xlsx format. Learn more at MachineHack.com","excerpt":"Analytics India Magazine’s MachineHack team feels it is crucial to predict when the outbreak may slow down, flatten or further worsen across different countries to ascertain the true nature of economic or human life cost as a result of COVID-19. We are in the midst of a worldwide outbreak of respiratory illness induced by a […]","categories":["Deep Tech"],"tags":["Coronavirus","Coronavirus and AI","covid-19","Machinehack","Machinehack Hackathon"],"author_name":"Vishal Chawla","publish_date":"2020-03-12T18:11:24","publication_year":"2020","word_count":499,"keywords":["Coronavirus and AI","covid-19","Machinehack","AI","RPA","programming_languages:R","Coronavirus","Aim","ViT","analytics","Machinehack Hackathon","GAN","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","GAN","ViT","RPA","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/new-hackathon-to-forecast-what-happens-next-in-covid-19-outbreak\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10140057,"title":"AI is Killing Remote Work","content":"“AI is killing remote work,” said Sahil Lavangia, founder of Gumroad. Citing Claude 3.5 Sonnet, he said that software that once took days to ship can now be delivered in hours or minutes, enabling people to work 10-20 times faster than before. Lavingia further said that AI-accelerated speedups in software and content creation have made it challenging for remote work to keep pace. However, he argued that even brief communication delays have become bottlenecks in AI-accelerated workflows. He explained that in-office teams can more effectively use AI’s rapid capabilities by collaborating in real time, using instant feedback for experimentation, problem-solving, and decision-making. This immediacy, he suggested, is harder to achieve remotely, where asynchronous delays slow potential productivity gains. “With AI handling much of the execution work—writing code, generating content, creating designs—the main bottlenecks are now cognitive: getting stuck on problems, running low on energy, or struggling to generate fresh ideas,” he said. On a similar note, AWS chief Matt Garman, recently made a decision to implement a strict five-day, in-office mandate for employees starting next year. He stressed the importance of in-person collaboration for driving “innovation” and “speed of execution.” He added that spontaneous hallway conversations and quick brainstorming sessions at whiteboards provide unique advantages in physical office spaces. “Particularly as we think about how we want to disrupt and invent on behalf of our customers, we find there is no substitution for doing that in person,” he said, adding that face-to-face interactions boost creative energy and speed. Makes sense. Recently, Google chief Sundar Pichai revealed during the company’s recent earnings call that 25% of Google’s new code is generated by AI. This leads us to the question of what the role of software engineers will be. “People wanted remote jobs and then got replaced by global talent that works twice as hard at half the cost. And now AI is also coming for those jobs,” said Neel Parekh, founder and CEO of MaidThis. Interestingly, former Google CEO Eric Schmidt recently said that Google was falling behind in the AI race due to its remote work policies. “Google decided that work-life balance and going home early and working from home was more important than winning,” he said. He suggested that startups like OpenAI and Anthropic were succeeding because their employees “work like hell.” During the GitHub Universe 2024, Thomas Dohmke also expressed a similar sentiment, saying that developers are now using AI to build AI. “Developers have embraced AI faster than any sector of the global workforce,” he said. However, not everyone agrees. “Remote work isn’t about coding cheaply. With remote work, you can hire high-IQ individuals who would be impossible to recruit in NYC physically,” said Catagay Kurt, founder of an AI startup based in Los Angeles. Remote Work Kills Creativity? Last year, at a session organised by fintech company Stripe, OpenAI chief Sam Altman also said that remote work was a mistake. “I think definitely one of the tech industry’s worst mistakes in a long time was that everybody could go full remote forever, and startups didn’t need to be together in person, and there was going to be no loss of creativity,” he said. He added that the remote work experiment is over, and technology isn’t yet advanced enough for people to work fully remote indefinitely, especially in startups. On the contrary, Dropbox CEO Drew Houston recently shared in an interview that he sees remote work becoming the standard. He said that currently, 90 percent of Dropbox’s 2,600 employees work remotely. However, he also believes that while remote work offers great flexibility, there’s no substitute for meeting in person. “If you want to build relationships with people and build trust, which is super important. You can sustain relationships on Zoom, you can’t build relationships on Zoom,” he said. AI at Works Speaking at Cypher 2024, Mohandas Pai, head of Aarin Capital and former CFO of Infosys said that repetitive and rule-based jobs are going to disappear as AI can perform them easily. However, he added that creative jobs, where higher-order thinking skills are required, are not going away. “The highly creative jobs are not going to go away. People who are very well educated, very well informed, and specialists—their jobs are not going to disappear,” he said. Vinod Khosla, in his latest blog post, suggested that with AI, the idea of a three-day work week could soon become a reality. “With the right policies, we could smoothen the transition and even usher in a three-day workweek,” wrote Khosla, underlining how AI will fundamentally transform the way we work, albeit in a way that positively impacts all of humankind and the economy at large. Previously, Reid Hoffman, co-founder and executive chairman at LinkedIn, predicted that the traditional 9-to-5 job would disappear by 2034. He is bullish on the gig economy revolution, where 50% of the population will become freelancers and earn more while working for “3 or 4 gigs”, than those in traditional employment. Going a step ahead, Tesla chief Elon Musk suggests that AI will completely eliminate all jobs. “If you want to do a job that’s kinda like a hobby, you can do a job. But otherwise, AI and the robots will provide any goods and services that you want,” he said at a conference earlier this year.","excerpt":"“People wanted remote jobs and then got replaced by global talent that works twice as hard at half the cost. And now AI is also coming for those jobs.”","categories":["AI Features"],"tags":["AI Jobs"],"author_name":"Siddharth Jindal","publish_date":"2024-11-02T10:49:25","publication_year":"2024","word_count":882,"keywords":["Anthropic","Go","GitHub","OpenAI","AI","AWS","AI Jobs","Git","Rust","Claude 3.5","R"],"extracted_tech_keywords":["AI","OpenAI","Claude 3.5","Anthropic","AWS","R","Go","Rust","Git","GitHub"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/ai-is-killing-remote-work\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":2311,"title":"IDG-backed iCreate Software raises $9.2M led by Sequoia Capital","content":"Bangalore-based banking software developing company iCreate Software has announced the completion of its Series B round of funding of $9.2M (Rs 50 crore). The investment round was led by Sequoia Capital and IDG Ventures India, the Series A investors, also participated. As part of this investment, Gautam Mago from Sequoia Capital will join the board. The funds will be utilised for market expansion, product portfolio enhancement and to deploy banking solutions. iCreate claims its banking software products is used by over 20 banks across Asia, Europe, Africa and the Middle East, including HDFC Bank, IndusInd Bank, Bank of America, Gulf Investment Corporation, Capitec Bank, BankMandiri, SAIB Egypt and East West Bank. Anup Pai, Founder & COO, iCreate added, “A global leader like Sequoia Capital will help us launch newer products and expand our portfolio. This new thrust will certainly catapult our growth.” Founded in 2010, with offices in Mumbai, South Africa, the Middle East and Europe, iCreate is a global banking decision sciences company. It creates enterprise-grade solutions to integrate with banking systems. It has software solutions for business intelligence, analytics and risk and compliance. Shailendra Singh, Managing Director, Sequoia Capital India Advisors said, “iCreate has been delivering compelling value to the global banking sector with their banking BI and analytics products.”","excerpt":"Bangalore-based banking software developing company iCreate Software has announced the completion of its Series B round of funding of $9.2M (Rs 50 crore). The investment round was led by Sequoia Capital and IDG Ventures India, the Series A investors, also participated. As part of this investment, Gautam Mago from Sequoia Capital will join the board. The funds […]","categories":["AI News"],"tags":[],"author_name":"Fintellix","publish_date":"2012-12-13T04:58:29","publication_year":"2012","word_count":212,"keywords":["business intelligence","Go","API","funding","programming_languages:R","AI","Aim","analytics","Rust","R"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","Rust","API","business intelligence","funding","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/idg-backed-icreate-software-raises-9-2m-led-by-sequoia-capital\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10062648,"title":"Most underrated developer tools in 2022","content":"Chrome DevTools, Git SCM, Oracle JDeveloper, Android Developers, GitLab, Postman, and NVIDIA Developer are the go-to tools for the modern developers, and rightly so. But, if you dig deeper, you’ll stumble upon many hidden gems–robust, task-specific, and seamless platforms to accelerate your developer journey. We took the liberty to go on an easter egg hunt to find underrated developer tools that can perform repetitive tasks, save time, and deliver good performance. explainshell explainshell is a web-based application for parsing man pages, extracting options, and matching each argument to the relevant help text in the man page to explain a given command line. explainshell is built with a man page reader which converts a given man page from raw format to HTML, a classifier that goes through every paragraph in the man page and classifies it as contains options or not (algo\/classifier.py), an options extractor that scans classified paragraphs and looks for options (options.py), a storage backend that saves processed man pages to MongoDB (store.py) and a matcher that walks the command’s AST (parsed by bashlex) and contextually matches each node to the relevant help text (matcher.py). Regex101 regex101 is a multilingual regular expression debugger with real-time explanation, error detection and highlighting. The platform is free to use and supports PCRE, JavaScript, GoLang, and Java. Namae.dev namae is an open-source name availability checker that can scout various registries to check if the domain’s name is taken or not. It also has a name suggestion feature that auto-generates names. The source code is available on GitHub. It consists of Node.js Lambda for APIs and the React app for the web frontend and runs on ZEIT Now. namae supports various package registries and web platforms, such as Domains, GitHub, PyPI, RubyGems, Rust (crates.io), Homebrew \/ Homebrew Cask, Linux (Launchpad & APT), Twitter, Spectrum, Slack, Heroku, ZEIT Now, AWS S3, and js.org. Slazzer Slazzer is an AI-powered API that helps users remove background from any online image and fill it in with details. With its advanced computer vision algorithms, any image with a clear foreground and background can be processed in just a few seconds. Users can process up to 500 images per minute via the API, depending on the input image’s resolution. Slazzer is also available as a plugin, desktop app, and on-premise. Flaticon Founded in 2013, Freepik Company-owned Flaticon provides over 3 million editable vector icons for apps, websites and information design projects. Flaticon offers google extensions and robust features like patterns generator and font-face generator to visualise information on user interfaces. Icons are available in multiple formats, including PNG, PSD, SVG, EPS, and BASE 64. They are scalable and adaptable to projects, ranging from landing pages, social media templates, presentations to videos, brochures, and much more. cronhub.io Cronhub helps developers run any command or script on schedule without DevOps, managing servers and infrastructure. They can schedule and monitor recurring jobs without writing a single line of code. Cronhub allows developers to schedule jobs using any time interval or cron expression. Users can set up one-click uptime and running time monitoring for recurring jobs. The platform sends users instant alerts if jobs fail or run longer than expected. Users can invite and collaborate with team members using a shared dashboard. Lastly, Cronhub offers rich analytics on critical metrics of the jobs. P.S: Learn X in Y minutes If you are a professional programmer looking to learn a new language, Learn X in Y minutes is the place for you! It features over 20 community-contributed short programming language tutorials, accessible and editable via a public Github repository.","excerpt":"regex101 is a multilingual regular expression debugger with real-time explanation, error detection and highlighting.","categories":["AI Trends"],"tags":["AI Tool","API","coding","Developers","DevOps","Programming Languages","Web Development"],"author_name":"Sri Krishna","publish_date":"2022-03-13T13:00:00","publication_year":"2022","word_count":594,"keywords":["Go","API","AWS","AI","MongoDB","coding","ML","Web Development","Programming Languages","computer vision","RAG","analytics","JavaScript","AI Tool","DevOps","R","Developers"],"extracted_tech_keywords":["AI","ML","computer vision","analytics","RAG","AWS","MongoDB","R","JavaScript","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/most-underrated-developer-tools-in-2022\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10054228,"title":"Maintaining A Daily Log Help In Structuring Research Work: Sahana Prabhu, Robert Bosch","content":"A data scientist transitioned from an electronic communication engineer, Sahana Prabhu‘s research interests include diabetic retinopathy image analysis, meibomian image segmentation, emotion recognition via deep learning approaches, and retail analytics via RFID and stereo cameras. Analytics India Magazine caught up with Sahana, who is currently serving as a research scientist and technical architect at Bosch Engineering and Business Solutions, to understand her insights on deep learning, machine learning, etc. AIM: As a research scientist, what factors do you believe contribute to your research success? Please throw some light on it for us. Sahana Prabhu: Proper dataset collection that is representative of the problem to be solved is vital. Once the problem statement is formulated, one approach to innovating is to reduce the restrictions in previous research papers. An example of this is to add data augmentations to tackle variations in data from open settings instead of having a controlled setting for capturing data. Another valuable research method is combining two related problems and developing a comprehensive framework that optimises both solutions. For instance, one of my papers involves a framework to solve image matting and super-resolution simultaneously. Maintaining a daily log of useful web links and attempted experiments helps in structuring research work. Displaying the results of the intermediate steps of the algorithm and checking if it is along expected lines reduces erroneous assumptions. AIM: Is the problem of unsupervised learning more difficult to solve? What is your view? Sahana Prabhu: Unsupervised learning does not have pre-existing patterns to learn from in labelled training data. Nevertheless, it is necessary to harness the potential of huge unlabeled data and recognise unknown insights without the limitations of human bias. Since purely unsupervised learning is difficult, an approach to dealing with unlabeled data is self-supervised learning, an active area of research. AIM: How can larger firms that invest in deep learning ensure that their efforts benefit others in the field? Sahana Prabhu: Larger firms have more capacity to do experimental research for problems that are not obvious. They can achieve greater operational efficiency from their proprietary data for real-world problems. Their existing customer network can look into various areas and formulate ideas to improve automation processes. Many large firms provide start-up incubation support, such as the Bosch Accelerator program, which goes a long way to further progress in this field. AIM: What applications of deep learning excite you the most right now or soon? Sahana Prabhu: Material recognition using texture analysis is a topic I am working on now, and it has applications in track-and-trace across automotive manufacturing, food produce, and pharmaceutical domains. Explainability is another research area emerging for many domains such as medical imaging, assisted driving, and manufacturing defect detection. AIM: Which industry do you believe will be most disrupted in the future by deep learning? Sahana Prabhu: Autonomous retail and autonomous driving are two emerging industries that will be made possible in the future. Smart automation of retail stores for self-checkout, tracking the flow of customers, and monitoring inventory can be facilitated by deep learning, and this is already under the experimental phase. Several automotive companies, including Bosch, have made significant advancements in autonomous driving. In addition, deep learning for computer vision, including semantic segmentation and image retrieval, is used extensively for assisted driving applications. AIM: What are your opinions on the recent rise in interest in deep learning in the media? Sahana Prabhu: There are two sides to it – while it has gained proper attention for automation in some domains such as manufacturing, it is also necessary to recognize its limitations. For example, deep learning can be an additional aid for routine screening in many medical image-based diagnosis tasks, but medical image-based diagnosis cannot fully eliminate doctor supervision. Moreover, there are many applications wherein conventional computer vision methods already provide the required accuracy for large-scale deployment, and deep learning may not be useful. There is also a perception that deep learning can give better predictions for cases where humans cannot, but deep learning gives only close to human accuracy in many applications. AIM: What are some of the difficulties that someone new to machine learning might encounter? How should they approach them? Sahana Prabhu: Information overload is a common problem that both newbies and researchers in machine learning face. It is tempting to apply machine learning for problems where conventional computer vision methods would suffice. We have a lot of online resources available, but it is important to focus on the problem and limit oneself to appropriate methods. New state-of-the-art algorithms keep getting added, and these can be found in paper listing sites such as “Papers with code” and “Awesome Deep Learning Resources”, rather than doing random searching on the internet. AIM: Who is your role model in machine learning research? Sahana Prabhu: Prof Andrew Ng has facilitated many research students (including myself) from related core fields such as computer vision to transition smoothly to machine learning. The way he straddles academic theory and industrial applications is inspiring. AIM: Are there any research papers you think every data scientist should read, irrespective of whether they are just starting or have years of experience? Sahana Prabhu: I will mention three breakthrough deep learning papers in computer vision for the three broad topics – classification, detection, and segmentation, respectively: [1] Simonyan K, Zisserman A. “Very deep convolutional networks for large-scale image recognition.” 2014. [2] Girshick R., et al. “Rich feature hierarchies for accurate object detection and semantic segmentation.” CVPR 2014. [3] Ronneberger O., et al. “U-net: Convolutional networks for biomedical image segmentation.” MICCAI 2015.","excerpt":"Explainability is emerging for many domains such as medical imaging, assisted driving, and manufacturing defect detection.","categories":["AI Features"],"tags":["AI in manufacturing","Classification","Computer Vision","data augmentation","data collection","Deep Learning","detection","Healthcare Automation","Interviews and Discussions","Machine Learning","Manufacturing","medical imaging","Segmentation","signal processing","Unsupervised Learning"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-11-25T13:00:00","publication_year":"2021","word_count":924,"keywords":["computer vision","deep learning","object detection","R","Manufacturing","image recognition","detection","analytics","Computer Vision","signal processing","AI in manufacturing","Unsupervised Learning","medical imaging","Go","Classification","machine learning","AI","Machine Learning","Healthcare Automation","data augmentation","Segmentation","Aim","data collection","Deep Learning","Interviews and Discussions"],"extracted_tech_keywords":["AI","machine learning","deep learning","computer vision","analytics","Aim","image recognition","object detection","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/maintaining-a-daily-log-sahana-prabhu\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10059620,"title":"3D animation using AI: Behind Plask","content":"If you’re an aspiring animator searching for a programme, you can look into Plask. Plask is a web-based 3D animation editor and motion capture tool driven by AI. It includes the required animation tools, allowing you to record, edit, and animate your projects without ever leaving your browser. However, Plask’s most significant feature is its AI-assisted ability to animate your characters using any video as a mocap. To capture keyframes, simply upload any films or record motions with any camera straight on Plask. It retargets, rigs, and optimises the output on the editor automatically. You can also use a webcam to record yourself and receive a video. Key features Motion Capture This step allows the user to take a person’s video, crop it, and extract the 3D motion. Plask further applies the extracted 3D motion to the character. These are certain guidelines to follow to capture the motion accurately: Plask can identify only people’s poses in the video. It is recommended to include the person’s entire body in all settings.The video must only include one individual.Plask does not currently support multiple characters. The person in the video should be clear to improve motion accuracy. For video recording, a horizontal angle is suggested. Angled video may have an impact on the outcome. Users should view the video at 30 frames per second. The running time for other frames is altered. Video formats supported: mp4, mov, and avi Retargeting Plask provides for both manual and automatic retargeting. Automatic retargeting is performed based on keywords and hierarchies in the image below. A modal will notify if automatic retargeting fails. A user can manually retarget by going to the Retargeting tab in the Control panel and visualising the model, based on 24 source bones, PLASK motion capture is extracted. source: https:\/\/plasticmask.notion.site\/User-guide-ac4bba1b75384c309e7a24e6542454ba Once done, users can easily export the file in FBX, GLB, and BVH formats. AI powered animation- an emerging field Generative models, which may create very realistic results in sectors such as semantic photos or movies, are among the most recent technological achievements. Using machine learning capabilities, Disney has simplified the process of designing and modelling 3D faces. A nonlinear 3D face-modelling system based on neural networks has been proposed by Disney researchers. This system learns a network architecture that converts a face’s neutral 3D model into the desired facial expression. In contrast, an animation technology start-up Midas Interactive has already started the process. Jiayi Chong, a former technical director at Pixar, has developed Midas Creature, a new tool that automates sophisticated 2D character animation. Artists and designers use Midas Creature to tell the engine to choreograph and figure out the movements for them, and the engine does it. Also read: Facial Motion Capture for Animation Using First Order Motion Model","excerpt":"This free AI-powered 3D animation editor and mocap tool will completely change the way we edit our videos.","categories":["AI Features"],"tags":["coding"],"author_name":"Abhishree Choudhary","publish_date":"2022-02-05T10:00:00","publication_year":"2022","word_count":458,"keywords":["Go","machine learning","TPU","programming_languages:R","AI","neural network","Modal","coding","programming_languages:Go","R"],"extracted_tech_keywords":["AI","machine learning","neural network","TPU","R","Go","Modal","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/3d-animation-using-ai-behind-plask\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":19561,"title":"Google’s AlphaZero Masters All Chess Knowledge Known To Mankind In Just 24 Hours","content":"After creating a “perfect” recipe for gluten-free chocolate chip cookie, Google’s Artificial Intelligence programme has now mastered the game of chess. According to a new paper, Google researchers recently detailed as to how their latest AI creation, AlphaZero, developed a “superhuman performance” in chess. Reportedly, the programme only took four hours to learn the rules before obliterating the world champion chess program, Stockfish. “The game of chess is the most widely-studied domain in the history of Artificial Intelligence. The strongest programs are based on a combination of sophisticated search techniques, domain-specific adaptations, and handcrafted evaluation functions that have been refined by human experts over several decades…” “We have generalised this approach into a single AlphaZero algorithm that can achieve, tabula rasa, superhuman performance in many challenging domains. Starting from random play, and given no domain knowledge except the game rules, AlphaZero achieved within 24 hours a superhuman level of play in the games of chess and shogi (Japanese chess) as well as Go, and convincingly defeated a world-champion program in each case,” said the paper written by David Silver, Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, Timothy Lillicrap, Karen Simonyan and Demis Hassabis. Interestingly, AlphaZero, wasn’t designed to play any of these games. In each case, it was given the basic rules of chess, go, etc, but was programmed with no other strategies or tactics. It simply got better by playing itself over and over again at an accelerated pace — a method of training AI known as “reinforcement learning”. Chess researcher David Kramaley, who is also the CEO of chess science website Chessable, said, “We now know who our new overlord is… It will no doubt revolutionise the game, but think about how this could be applied outside chess. This algorithm could run cities, continents, universes.”","excerpt":"After creating a “perfect” recipe for gluten-free chocolate chip cookie, Google’s Artificial Intelligence programme has now mastered the game of chess. According to a new paper, Google researchers recently detailed as to how their latest AI creation, AlphaZero, developed a “superhuman performance” in chess. Reportedly, the programme only took four hours to learn the […]","categories":["AI News"],"tags":["AI (Artificial Intelligence)","Google"],"author_name":"Prajakta Hebbar","publish_date":"2017-12-07T09:59:35","publication_year":"2017","word_count":308,"keywords":["Go","artificial intelligence","programming_languages:R","AI","programming_languages:Go","BERT","llm_models:BERT","Google","Julia","R","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","artificial intelligence","R","Go","Julia","BERT","llm_models:BERT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/googles-alphazero-masters-chess-knowledge-known-mankind\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10355,"title":"After Brexit stumbling predictions, do we need better big data driven forecasting?","content":"The data driven forecasting such as opinion polling and prediction market came in for a major shock when later in the month of June, Great Britain opted to vote against staying in the EU, much contradictory to the overwhelming cry till date that UK would ‘Remain’. Amidst the mix reaction of surprise, fury and relief, the question that definitely becomes inevitable is ‘where did the market prediction go wrong?’ Do we need a better data driven forecasting? Before we set our foot on the debate let’s find out what went wrong and why has data driven Brexit come as a wakeup call for analysts? Forecasting results when it comes to large amount of data such as in elections have been fairly common. And given a scenario in any country across the globe, the know-hows keep coming up with the current support rate for a candidate and their chances to win. The call for Brexit, in the same way fetched a charged up campaign and public polling to know what the citizens favoured. They were called in for voting, with a few opting for telephone and others via online voting. The votes were divided into “Remain” and “Leave”. Given the scenario we would have hope either a clear winner or a tie. But the results were all the opposite. The overall predictions itself came out faulty which could be witnessed through various poll trackers. The poll tracker for Brexit powered by Bloomberg showed that Remain was maintaining a lead with 2% while Brexit had a bare minimum chances of 25% win. The same kind of results were shown by YouGov where the ratio of Remain vs. Leave was 52:48. With 90% of the market claiming to be in favour of ‘Remain’ and results coming out in favour of Brexit, there have been questions and concerns raised over a need for better data driven forecasting. With a serious lag curbing up between predictions and the results, it has raised critical questions about the reliability of these forecasts in the coming future. It definitely indicates that the long practised method of collecting thoughts and opinions from people strolling up the street or simply by calling them up is a passé and we need something more extra-ordinary than that. Apart from polling, social media platforms and web search which were considered as a basis to form out predictions went haywire. Given the ambiguity of mere discussions rather than standing out as strong support, made it difficult to declare a clear emerging winner. It again leaves us to the point of having a strongly knit data forecasting tool. When it comes to real world forecasting, it could be quite difficult to get the right results. Though there have been a few online polls, social media and Google search volume timeline closer to the actual results which were in favour of Brexit, improvements in “big data” measurements to come out with better and robust predictions is the need of the hour.","excerpt":"The data driven forecasting such as opinion polling and prediction market came in for a major shock when later in the month of June, Great Britain opted to vote against staying in the EU, much contradictory to the overwhelming cry till date that UK would ‘Remain’. Amidst the mix reaction of surprise, fury and relief, […]","categories":["IT Services"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2016-07-06T04:41:32","publication_year":"2016","word_count":495,"keywords":["big data","Go","programming_languages:R","AI","programming_languages:Go","Aim","ViT","R"],"extracted_tech_keywords":["AI","Aim","R","Go","big data","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/brexit-stumbling-predictions-need-better-big-data-driven-forecasting\/","complexity_score":2,"technical_depth":8,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":62083,"title":"8 Different NLP Scenarios One Can Take Up For A Project","content":"Over the last few years, natural language processing (NLP) techniques have witnessed fast growth in quality as well as usability. Today, NLP is one of the most trending topics of research in the field of STEM. Tech giants have been researching NLP, and applying newer deep learning methods to gain a deeper understanding of the consumers. In this article, we list down – in no particular order – eight different NLP scenarios that one can take up for a project. 1| Question Answering Question answering is one of the most prevalent research problems in NLP. Some of its applications are chatbots, information retrieval, dialog systems, among others. It serves as a powerful tool to automatically answer questions asked by humans in natural language, with the help of either a pre-structured database or a collection of natural language documents. Models: Models like BiDAF, BERT, and XLNet can be used for question-answering projects. Dataset: Stanford Question Answering Dataset (SQuAD), Conversational Question Answering systems (CoQA), etc. 2| Text Classification Text Classification or Text Categorization is the technique of categorizing and analyzing text into some specific groups. This technique supports a comparative evaluation of the impact of linguistic information concerning approaches based on word matching. Models: BERT, XLNet, and RoBERTa can be used for text classification. Dataset: Amazon Reviews dataset, IMDB dataset, SMS Spam Collection, etc. 3| Text Summarization Text summarization is one of the most efficient methods to interpret text information. Text summarization methods can be mainly categorized into two parts – extractive summarization and abstractive summarization. In extractive summarization, the process involves selecting sentences of high rank from any document based on word and sentence features and fusing them to generate a summary. On the other hand, an abstractive summarization is mainly used to understand the main concepts in any given document and then express those concepts in any natural language. Models: BERTSumExt, BERTSumAbs, and UniLM (s2s-ft) can be used for text summarization. Dataset: BBC News Summary, Large-Scale Chinese Short Text Summarization Dataset, etc. 4| Sentiment Analysis Sentiment Analysis is the technique of understanding human sentiments implied in a text, and helps classify emotions using text analysis methods. This technique has witnessed significant traction due to the growth of social media platforms like Facebook, Instagram, and more. Some of the applications of this technique are market research, brand monitoring, customer service, among others. Models: Models like Dependency Parser, BERT, and RoBERTa can be used for sentiment analysis. Dataset: Stanford Sentiment Treebank, Multi-Domain Sentiment Dataset, Sentiment140, etc. 5| Sentence Similarity Sentence similarity portrays an important part in text-related research and applications in areas such as text mining and dialogue systems. This technique has proven to be one of the best to improve retrieval effectiveness, where titles are used to represent documents in the named page finding task. Models: BERT, GloVe, etc. can be used for sentence similarity projects. Dataset: Paraphrase Adversaries from Word Scrambling (PAWS) 6| Speech Recognition Speech Recognition is the technique used in identifying spoken words or phrases and translating them into machine language. Speech recognition has gained attention in recent years with the dramatic improvements in acoustic modeling yielded by deep feedforward networks. Models: BERT, RoBERTa, etc. can be used for speech recognition projects. Dataset: Google AudioSet, LibriSpeech ASR corpus, etc. 7| Neural Machine Translation Neural machine translation is one of the most popular approaches in NLP research. The neural machine translation aims at building a single neural network that can be jointly tuned to maximize translation performance. Models: BERT, RNN Encoder-Decoder, etc. Dataset: English-Persian parallel corpus, Japanese-English Bilingual Corpus, etc. 8| Document Summarization Document Summarization is the technique of helping readers catch the main points of a long document with less effort. It also helps as a preprocessing step for some text mining tasks such as document classification. This method can be categorized into two different dimensions – abstract-based and extract-based. An extract-based summary includes sentences that are extracted from the document. In contrast, an abstract-based summary may consist of words and phrases which do not appear in the original document. Models: Hidden Markov Model can be used for document summarization. Dataset: 20 Newsgroups dataset.","excerpt":"Over the last few years, natural language processing (NLP) techniques have witnessed fast growth in quality as well as usability. Today, NLP is one of the most trending topics of research in the field of STEM. Tech giants have been researching NLP, and applying newer deep learning methods to gain a deeper understanding of the […]","categories":["AI Trends"],"tags":["machine translation","NLP AI","NLP projects","Question Answering","sentiment analysis NLP","Speech Recognition","text classification","Text summarization"],"author_name":"Ambika Choudhury","publish_date":"2020-04-20T13:00:00","publication_year":"2020","word_count":686,"keywords":["NLP projects","machine translation","text classification","Question Answering","AI","neural network","chatbots","sentiment analysis","sentiment analysis NLP","AWS","Text summarization","NLP","Ray","Aim","deep learning","Speech Recognition","NLP AI"],"extracted_tech_keywords":["AI","deep learning","neural network","NLP","Aim","Ray","chatbots","sentiment analysis","text classification","AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/8-different-nlp-scenarios-one-can-take-up-for-a-project\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10019574,"title":"Top Ten Kaggle Notebooks For Data Science Enthusiasts In 2021","content":"Kaggle Notebook is a cloud computational environment which enables reproducible and collaborative analysis. Notebooks, previously known as kernels, help in exploring and running machine learning codes. It also helps in discovering the vast repository of public, open-sourced, as well as, reproducible code for data science and machine learning projects. Currently, there are more than 460,000 public notebooks available in Kaggle. There are two types of Notebooks on Kaggle. The first type is a script that executes everything as code sequentially, and the other type is the Jupyter notebooks that consist of a sequence of cells, where each cell is formatted in either Markdown or in a programming language. Here is a list of ten top Kaggle Notebooks a data science enthusiast must know. 1| Comprehensive Data Exploration With Python About: This notebook offers a comprehensive analysis of data. In this kernel, you can perform tasks like understanding the problem by looking at each variable, focusing on the dependent variable, understanding how the dependent variable and independent variables relate, cleaning the dataset and handling the missing data, outliers and categorical variables and, lastly, checking if the data meets the assumptions required by multivariate techniques. Know more here. 2| Titanic Data Science Solutions About: The notebook is a typical workflow for solving data science competitions at sites like Kaggle. The notebook follows a step-by-step workflow, explaining each step and rationale for every decision a data scientist needs to take during solution development. It involves analysing and visualising data, analysing data before and after wrangling, etc. Know more here. 3| Hello, Python About: This notebook covers the key Python skills you need to start using Python for data science. It helps in understanding as well as levelling up the basic Python skills. The notebook includes a brief overview of Python syntax, variable assignment, and arithmetic operators. Know more here. 4| Introduction to Ensembling\/Stacking in Python About: This notebook serves a primer for ensembling (combining) base learning models, in particular, the variant of ensembling known as Stacking. The notebook allows ensembling in an intuitive and concise manner, and also includes the Titanic dataset. Know more here. 5| A Data Science Framework: To Achieve 99% Accuracy About: In this Notebook, you will learn the approaches in data science such as data science frameworks, gathering and preparing data, performing exploratory analysis, tuning models with hyper-parameters, feature selection and more. Know more here. 6| Exploring Survival on the Titanic About: This notebook focuses on illustrative data visualisations using the Titanic dataset. It includes topics like feature engineering, predictive imputation, building a machine learning model, variable importance, sensible value imputation, etc. Know more here. 7| Credit Fraud || Dealing with Imbalanced Datasets About: In this notebook, you will learn how to use various predictive models to see if a transaction is legitimate. The goal of this notebook is to understand the distribution of data, create a 50\/50 sub-dataframe ratio of “Fraud” and “Non-Fraud” transactions, determine the Classifiers, create a Neural Network and compare the accuracy to the best classifier, etc. Know more here. 8| Coronavirus (COVID-19) Visualization & Prediction About: This notebook aims at exploring COVID-19 through data analysis and projections. It includes exploring global coronavirus cases, exploring the cases from different countries, prediction of worldwide confirmed cases, US testing data, mobility data from hotspots etc. Know more here. 9| Approaching (Almost) Any NLP Problem on Kaggle About: This notebook discusses the approaches to natural language processing problems on Kaggle. You will learn how to use data and create a very basic first model as well as improve it using different features. It includes topics like logistic regression, naive bayes, svm, xgboost, grid search, word vectors, LSTM, and more. Know more here. 10| Interactive Intro to Dimensionality Reduction About: This Notebook discusses the merits of dimensionality reduction methods. The Notebook aims to provide an introductory exposition on the three methods, which are PCA (Principal Component Analysis), LDA ( Linear Discriminant Analysis) and TSNE ( T-Distributed Stochastic Neighbour Embedding). The notebook also allows visualisations via the Plotly visualisation library. The dataset used in this Notebook is the popular MNIST (Mixed National Institute of Standards and Technology) computer vision digit dataset. Know more here.","excerpt":"Kaggle Notebook is a cloud computational environment which enables reproducible and collaborative analysis. Notebooks, previously known as kernels, help in exploring and running machine learning codes. It also helps in discovering the vast repository of public, open-sourced, as well as, reproducible code for data science and machine learning projects. Currently, there are more than 460,000 […]","categories":["AI Trends"],"tags":["data science projects","Jupyter Notebook","machine learning projects","survival regression python"],"author_name":"Ambika Choudhury","publish_date":"2021-02-02T18:00:00","publication_year":"2021","word_count":694,"keywords":["data science projects","data science","Jupyter Notebook","machine learning","Plotly","AI","neural network","computer vision","survival regression python","NLP","Aim","XGBoost","Jupyter","machine learning projects"],"extracted_tech_keywords":["AI","machine learning","neural network","NLP","computer vision","data science","Aim","XGBoost","Jupyter","Plotly"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-ten-kaggle-notebooks-for-data-science-enthusiasts-in-2021\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":31553,"title":"The Hitchhiker&#8217;s Guide to Artificial Intelligence 2018-2019: By AIM &#038; Great Learning","content":"Everyone is talking about Artificial Intelligence — the new normal, which has entered almost every work process across industries. Enterprises are rethinking and strengthening their AI capabilities, using it as a tool to improve products and services. With AI becoming crucial to enterprise success, upskilling has become the new mantra among Indian IT professionals, who are keen to make an impact in their careers with Machine Learning and AI. In our annual AI Study with Great Learning, we take a look at key AI trends dominating the Indian AI market-leading companies, professionals, salaries, jobs broken down by cities and how AI’s potential for industry growth has risen over the last few years. In the second half of the study, we cover AI literacy in India through Great Learning’s comprehensive AI\/ML programs that are bridging the current skill gap and consequently boosting workforce transitions. Key Trends The Artificial Intelligence Industry in India is currently estimated to be $230 million (annual) in revenue, up from $180M a year ago Currently, there are approximately 40,000 AI professionals in India There is a growing interest in AI with the Indian government doubling up spends on AI\/ML research AI applications have emerged in key areas like healthcare and agriculture Bengaluru outpaces its peers in terms of attracting AI talent and has emerged as a leading AI hub in India AI Professionals in India The average work experience of AI professionals in India is 6.6 years, same as last year Around 3,700 freshers were added to the AI workforce in India this year Almost 55% AI professionals in India have a work experience of fewer than 5 years, same as last year 23% of AI professionals have more than 10 years of work experience. This work experience is not necessarily in AI but these professionals have transitioned into AI over time Women participation in AI workforce remains low – only 24% of AI professionals in India are women Tenure On average, AI professionals in India have joined\/transitioned to their current role in the last 3 years 67% AI professionals in India have joined\/transitioned to their current role in the last 2 years The numbers reveal AI is a relatively emerging technology and a large number of professionals are gradually gravitating toward it Education 48% of AI professionals have a Master’s\/Post Graduation degree 5% of AI professionals in India hold a Ph.D. or a Doctorate degree AI Companies in India The 10 companies that employ the most AI professionals in India are Accenture, TCS, Cognizant, Infosys, Wipro, IBM, Microsoft, Amazon, Capgemini & HCL Technologies More than 1,000 companies in India claim to work on AI in some form. This includes a small number of companies into products (chatbots, AI-powered visual search and recommendation engine) and a larger chunk offering either offshore, recruitment or training services There is an increase of almost 30% year-over-year in the number of companies setting up dedicated AI teams in India Moreover, the number of vertical AI companies in India are still very few in number, compared to the strength of AI companies around the globe. Even in analytics, India accounts for just 8% of global Analytics companies Company Size On average, Indian AI companies have 87 employees on their payroll Almost 85% of Analytics companies in India have less than 50 employees AI Talent Divide Across Companies Almost 37% of AI professionals in India are employed with large-sized companies(with a total employee base of 10k+) Mid-size organizations (total employee base in the range of 200-10K) employ 29% of all AI professionals in India. Startups (less than 200 employee base) employees form 34% of AI professionals in India. A large percentage of AI professionals are absorbed by digitally mature big tech firms that are increasing their investment in AI in India. This is also a clear indication of the AI talent divide between enterprises and startups in India, which will only continue to widen further Cities Bengaluru leads the cities in terms of the size of the AI ecosystem. 32% of AI professionals in India are working in Bengaluru This is closely followed by Delhi NCR at 22% AI Salaries in India The median AI salary in India is INR 14.3 Lakhs across all experience level and skill sets. Around 40% of AI professionals in India have an entry-level salary of 6 Lakh onwards Almost 4% of AI professionals in India command a salary higher than 50 Lakhs Average Salary Trend Across Cities Mumbai is the highest paymaster in AI at almost 15.6 Lakhs per annum, followed by Bengaluru at 14.5 Lakhs Chennai is the lowest paymaster at 10.4 Lakhs AI Jobs in India While it is difficult to ascertain the exact number of open AI job openings, according to our estimates, close to 4,000 positions related to AI are currently available to be filled in India. Open AI jobs is a different metric than new jobs advertised per month Compared to worldwide estimates, India contributes 10% of open job openings currently. Growth in the number of AI jobs globally was much higher than in India 10 leading organizations with the most number of AI openings this year are – IBM, Accenture, Amazon, Fractal Analytics, Societe Generale, SAP Labs, 24\/ 7 Customer, Atos, Nvidia & Tech Mahindra The top skill sets that AI employers are looking for are Machine Learning, Natural Language Processing, Neural Networks, Analytics, Cloud Computing & Pattern Recognition Almost 92% of AI jobs advertised in India are for full-time roles; rest are part-time, internships or contract basis jobs AI Jobs By Cities In terms of cities, Bengaluru accounts for around 33% of AI jobs in India. This is down from 37% last year Delhi\/NCR comes second contributing 30% AI jobs in India. This is up from 23% last year Approximately 12% of AI jobs are from Mumbai, almost the same as last year Experience Requirement By AI Jobs Around 43% of AI jobs are looking for candidates with less than 5 years of experience 6% AI jobs are for freshers 57% AI job openings are for professionals with more than 5 years of job experience Here’s the complete Report: Download The report [attachments include=”31709″]","excerpt":"Everyone is talking about Artificial Intelligence — the new normal, which has entered almost every work process across industries. Enterprises are rethinking and strengthening their AI capabilities, using it as a tool to improve products and services. With AI becoming crucial to enterprise success, upskilling has become the new mantra among Indian IT professionals, who […]","categories":["AI Features"],"tags":["AI jobs in India"],"author_name":"Richa Bhatia","publish_date":"2018-12-18T08:51:52","publication_year":"2018","word_count":1024,"keywords":["AI jobs in India","machine learning","artificial intelligence","AI","neural network","chatbots","ML","cloud computing","RAG","Aim","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","neural network","analytics","Aim","RAG","chatbots","cloud computing"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/the-hitchhikers-guide-to-artificial-intelligence-2018-2019-by-aim-great-learning\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10110075,"title":"How Jarvislabs AI is Helping Developers Build LLMs","content":"“We believe in the Zoho School of Thought, inspired by Sridhar Vembu, that you don’t need to build a company from a city like Bengaluru,” said Vishnu Subramanian, founder and chief of Jarvislabs.ai in an exclusive interview with AIM. Situated on the outskirts of Coimbatore, Jarvislabs.ai is silently playing a crucial role in providing AI infrastructure (GPUs on rent) for budding entrepreneurs and developers at an affordable price, alongside helping them advance AI research. Jarvislabs.ai is one of the few companies in India that provides GPUs on rent for training, fine-tuning, and deploying AI models. Moreover, utilising services from hyperscalers like Microsoft Azure, Google Cloud, and AWS can be a costly affair for those who simply want to experiment. “If you compare with a hyper scalar, I would say we are like 60 to 70% cheaper” shared Subramanian. “We wanted to make it easy and affordable for a small company in Chennai or Bengaluru to deploy a model natively, which can be expensive, especially when considering platforms like AWS or Google Cloud Platform. Even for older GPUs, they often charge around 6.7 cents, whereas we provide a modern GPU at a comparable price.” added Subramanian. Unlocking Affordable GPU Prices Subramanian clarified that their strategy to keep GPU prices low involves not pursuing NVIDIA’s H100s. Instead, they provide alternatives that offer similar capabilities. “The reason we are able to keep the costs low is we are not going behind the fancy H100s which are super expensive today. But you can do a lot of really amazing stuff with the Quadro or the next generation GPUs.” said Subramanian. Read: NVIDIA H100 Vs A100 “We have A100s right now. We are not directly investing in H100; instead, we plan to partner with different hyperscalers or similar platforms to incorporate H100s into our platform. Currently, the GPUs we host include Quadro RTX 5000 and 6000. These are the latest generations and are as good as the older generation A100, but they come at a much cheaper price.” he added. “For example, RTX 6000 Ada comes at around $1.14, providing similar performance to A100, for which you would end up spending around $3,” he explained. While Subramanian didn’t disclose the exact number of GPUs acquired by Jarvislabs.ai, he did shed light on the strategy behind securing them. “There is a national distributor named RP Tech in India. They are the sole company responsible for importing GPUs into India. Fortunately, we have partnered with them, and from the start, they have been generous enough to assist us in obtaining GPUs at a very competitive price compared to the open market,” he said, saying that the prices are usually higher on ecommerce websites like Amazon, compared to buying it directly from distributors. Subramanian also mentioned that they’ve barely spent a penny on marketing. Word of mouth has always helped them, as they have been offering a decent and good product. “People have been kind enough to introduce us to their friends, and word of mouth continues,” he said. This is quite true as we came across Jarvislabs.ai when AIM was talking to one of the local developers, where he, in passing, mentioned about this AI lab, and how it gives GPUs-on-rent for building AI models and advancing AI research. Current Customers Subramanian said that their target customers are anyone ‘who is not a Fortune 500 company.’ Some of the notable customers of Jarvislabs.ai include Zoho and upGrad.”Some of the top companies in India, like Zoho, have been using LLMs for different workloads, probably to help students learn programming code, like a copilot or something similar. I’m not sure exactly what they’re doing, but I would assume that they could be fine-tuning some of the models.” said Subramanian. Universities like CMU (Carnegie Mellon University) from the US have been using Jarvislabs.ai’s services to train large language models. Additionally, Weights and Biases are also leveraging its services. “Our primary emphasis is on the global market, and until at least mid-2023, there wasn’t much usage from Indian customers. It was predominantly outside India, in the US and Europe. However, we see a slow change now; Indian customers are also starting to adopt more and more AI,” said Subramanian. Not Just a Hardware Company Subramanian emphasised that Jarvislabs.ai is not merely a hardware company but rather a comprehensive solution provider saying “How to train your model and how to deploy it are much more challenging engineering problems to solve. We are packaging this entire thing for our customers.” Additionally, as a bootstrapped company, Jarvislabs.ai cannot afford to offer substantial discounts to its customers, unlike VC-backed firms that often provide significant discounts as they aim to capture a larger market share. To compensate for this, Jarvislabs.ai is providing additional value to its customers at the same price. “We’re not just focusing on renting GPUs as raw material. We are trying to help. Let’s say, for example, you are a small company with 40-50 people. You may not have the DevOps team or the AI team to create an optimsed environment. So, we also simplify things for them. We have automated all these steps. With just a few clicks, all of this is done for you.” He shared that, at the end of the day, as a company, it also has to make money, and it’s not sustainable to buy GPUs at a very high price. “NVIDIA has been increasing prices year-on-year with every release. For example, the RTX 6000 used to cost us around three lakhs. The current generation costs around six to 6.5 lakhs per GPU, but the cost at which we give it to the customer has not increased a lot,” he said. Subrmanian said that he intends Jarvislabs.ai to be like an app store which will host multiple frameworks. “We are working on a new version of the product, and we hope to release it this month. This version will incorporate a lot more frameworks. When we started, there were only three popular frameworks: PyTorch, TensorFlow, and Fast.ai,” he concluded.","excerpt":"“If you compare with a hyper scalar, I would say we are like 60 to 70% cheaper” said Vishnu Subramanian.","categories":["AI Features"],"tags":["LLMs"],"author_name":"Siddharth Jindal","publish_date":"2024-01-05T16:52:00","publication_year":"2024","word_count":1001,"keywords":["Go","AWS","PyTorch","LLMs","AI","R","Scala","RAG","Aim","TensorFlow","Azure"],"extracted_tech_keywords":["AI","Aim","TensorFlow","PyTorch","RAG","AWS","Azure","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-this-coimbatore-based-ai-lab-is-helping-developers-build-llms\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":49061,"title":"Despite Much Demand, Why Is It So Difficult To Land A Job In Data Science","content":"The demand for data science across the industries is real. Over the past few years, the domain has been experiencing a rapid rise in jobs across the world and India is one such country that is experiencing a data explosion. According to one of our reports, the total number of analytics and data science job positions available are 97,000. Out of these, 97% of job openings in India are on a full-time basis while 3% are part-time or contractual. However, despite this skyrocketing need for data science professionals, there is a question that is keeping many aspirants awake at night — if there’s such a great demand for data scientists in India, then why is it so hard to get a job in this field? Four Key Reasons While many people have said that companies don’t understand what they need to do with data science, there are others who have said, “Most aspirants are not competent enough to be Data Scientists”. Either option is true to an extent. In this article, we will try to give you an idea about why it is hard to land data science jobs. 1. Candidates Don’t Possess The Much Needed Skills And Knowledge One of the most brutal truths is that applicants tend to blame the industry for them not being able to land a job. However, most of the time it is not true. Data science is that field of work which is responsible for solving some of the most complex business problems and it requires an immense amount of knowledge and skill set. Blurred Definitions: The demand for a data scientist is so much that companies face more challenges for hiring a data scientist than a graduate trying to land a job. While everybody nowadays is calling themselves data scientists, they are missing out on even some of the basic analytical, numerical aptitude. It is important to understand the fact that engineering is not the only thing that makes a data scientist — its statistics, mathematics, economics, and a lot of folks also are from the business background. Unrecognised Courses: Also, one of the major reasons is the cheap marketing strategies that data science institutes use. They grab as many students as possible and don’t train them properly and the same students end up facing rejection after rejection and blame the industry. To make it clear, data science is complex, and one has to be exceptionally good to in the field. 2. Companies Lack The Understanding While the lack of skills and knowledge among candidates is definitely a challenge, companies also tend to misunderstand data science (common with most of the startups). There are companies who want to be seen on the data science map and for that, they look to hire data scientists. But the question here is — to solve what problem? Convoluted Job Descriptions: Companies don’t even know what is that major problem they want to solve, what is the knowledge and skillset requirement to solve that problem, all they do is have a bunch of tools as a requirement without even knowing how, where or why it is used.  And without any understanding of these, they post jobs with jargons, that not only confuses the candidate but also doesn’t anyone to excel their career. Hiring a wrong candidate would not only throw challenges to a company but also to the candidate. Problem Statement: The first and foremost thing a company needs to do is understand the challenges they are facing, then prioritise the challenges and then go to hire a data science professional to solve them Also, never jargonise job descriptions — be clear about the problem statement and what you are looking for in a candidate. 3. Competition Is Increasing The data science jobs are in demand and witnessing that, a tremendous amount of folks are starting out with data science; many people are even making the transition to data science completely different domain. And when we are saying it is hard to land a job in the domain, this might be probably because of the increasing competition in the sector. The one who is saying that it is hard to land a data science job might be the one who is not getting hired because of the lack of skill sets and knowledge. There is always another perspective to a situation, while people are facing rejection, there are people who are managing to land jobs in this domain. The competition is real and cannot be denied. If you want to become a data scientist that stands out, do not follow the herd — take a different route. For example, if you are learning through MOOCs — it’s fine but just like every other aspirant, do not spend all your life in the MOOC spiral. 4. Companies Relying On Service Providers Data science is expensive, period. The paycheck of a data scientist says it all, and when you have an entire team, the cost is definitely going to get higher. In order to feel light on the capital and expenses, companies nowadays are turning to data science service providers. It eliminates the need for an in-house data science team and pays them high on a monthly basis. Companies now get their problems solved by data science service providers. However, this also has a drawback and that is most of the service providers doesn’t deliver value and end fooling companies. So, it goes back to the second point “Companies don’t know what they want.”","excerpt":"The demand for data science across the industries is real. Over the past few years, the domain has been experiencing a rapid rise in jobs across the world and India is one such country that is experiencing a data explosion. According to one of our reports, the total number of analytics and data science job […]","categories":["AI Hirings"],"tags":["Data Science Career","Data Science Jobs","Talent Crunch"],"author_name":"Harshajit Sarmah","publish_date":"2019-10-30T17:00:22","publication_year":"2019","word_count":916,"keywords":["data science","Go","API","programming_languages:R","AI","programming_languages:Go","Data Science Jobs","Talent Crunch","Data Science Career","analytics","R","startup"],"extracted_tech_keywords":["AI","data science","analytics","R","Go","API","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-hiring\/despite-much-demand-why-is-it-so-difficult-to-land-a-job-in-data-science\/","complexity_score":2,"technical_depth":9,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10120777,"title":"ChatGPT Brings Data to Life With Interactive Charts and Tables Directly from Google Drive and Microsoft OneDrive","content":"OpenAI is rolling out significant improvements to data analysis in ChatGPT, making the tool more robust and user-friendly for ChatGPT Plus, Team, and Enterprise users. The new features, available with the flagship model GPT-4o, allow users to add files directly from Google Drive and Microsoft OneDrive and interact with tables and charts. The ability to add files directly from Google Drive and Microsoft OneDrive eliminates the need to download files to a user’s desktop before uploading them to ChatGPT. This feature allows ChatGPT to quickly understand Google Sheets, Docs, and Slides files as well as Microsoft Office Suite files like Excel, Word and PowerPoint. Interactive tables allow users to expand and explore data in real time, with the ability to ask follow-up questions and receive suggested prompts to deepen their analysis. For example, users can ask ChatGPT to combine spreadsheets of monthly expenses and create a pivot table categorized by expense type. Customisable charts enable users to interact with various chart types, including bar, line, pie, and scatter plots. Users can hover over chart elements, ask additional questions, and select colours before downloading the charts for presentations or documents. For instance, users can select a Google Sheet with their company’s latest user data directly from Google Drive and ask ChatGPT to create a chart showing retention rates by cohort. These updates build on ChatGPT’s existing ability to understand and analyse datasets through natural language. Users can upload data files and have ChatGPT write and run Python code to perform tasks like merging datasets, creating charts, and uncovering insights. This capability makes it easier for beginners to conduct in-depth analyses and saves experts time on routine tasks. Carlyle Group VP David Vaughn highlighted the tool’s impact, stating, “ChatGPT is part of my toolkit for analyzing customer data, which has become too large and complex for Excel. It helps me sift through massive datasets, allowing me to conduct more data exploration on my own and reduce the time it takes to reach valuable insights.” OpenAI emphasizes comprehensive security and privacy in these updates. ChatGPT does not train on data from Team and Enterprise customers, and Plus users can opt out of training through Data Controls. OpenAI’s privacy and security policies include SAML SSO, compliance, and data encryption for ChatGPT Enterprise. Afterpay marketing manager Lauren Nowak praised the new features, saying, “ChatGPT walks me through data analysis and helps me better understand insights. It makes my job more fulfilling, helps me learn, and frees up my time to focus on more strategic parts of my job.” Similarly, Google also announced new features of Gemini at Google I\/O 2024. Gemini not only analyses the data from the sheet but also creates a nice visual to help users see the complete breakdown by category. This extends to various use cases in your inbox, like travel expenses, shopping, and remodelling projects. All of that information in Gmail can be put to good use, helping you work, plan, and play better. This ability to organise your attachments in Drive, generate a sheet, and do data analysis via Q&A will be rolling out to Labs users this September.","excerpt":"Users can now upload files directly from Google Drive and Microsoft OneDrive, eliminating the need to download files to their desktop before uploading them to ChatGPT.","categories":["AI News"],"tags":["ChatGPT","OpenAI"],"author_name":"Siddharth Jindal","publish_date":"2024-05-17T08:39:30","publication_year":"2024","word_count":520,"keywords":["Go","ChatGPT","OpenAI","AI","GPT-4o","ML","Python","GPT","GAN","R"],"extracted_tech_keywords":["AI","ML","GPT-4o","ChatGPT","OpenAI","Python","R","Go","GPT","GAN"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/chatgpt-brings-data-to-life-with-interactive-charts-and-tables-directly-from-google-drive-and-microsoft-onedrive\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":52272,"title":"How India Dealt With Cyberattacks In 2019","content":"Cyberattacks are rife in India, only the US and China are placed higher on this list. Bangalore, Mumbai, Delhi are among states which receives the highest traffic. In the first quarter of 2019, 48% of the total attack was recorded in Chennai. With companies suffering heavy damages (monetary as well as in data theft), cybersecurity insurance purchases are selling like hotcakes, 50% increase in 2018 alone. “Globally, the trend is to have overarching agencies for better and command and control. It is time we also have a similar structure. India’s capabilities in the cyber world have expanded and there are a large number of agencies — but sharing real-time information is always not enough. Cyber defence capabilities is a critical strategic requirement. I think this a very positive move,” commented former telecom secretary Aruna Sundararajan. The Cyberattacks Kundankulam Nuclear Power Plant (KNPP), came under a cyberattack in the last week of October. The damage was in the form of data theft, the same data can be used to enhance future attacks on the power plant. Even ISRO faced attacks prior to the Chandrayan 2 mission. India-based healthcare website was attacked in 2019, medical records and information of 68 million patients and doctors were stolen. On average, it takes foreign agencies 7 days to respond and contain cyberattacks, cybersecurity agencies in India take 9 days to do the same. The additional 2-day window is a lot of time, especially for hackers. In 2017, Indian servers were beached once every 10 minutes. What’s discouraging to know is that not all instances of cyber-related crimes are reported, hence, a breach every 10 minutes could very well be a breach every minute. Cyberattacks exploit reported vulnerabilities in the system and of the users, keeping oneself and your operating systems updated comes highly recommended. How The Authorities Handled The Attacks The authorities in India have also not been up to the task. Their reluctance to divulge information is worrisome, it took NPCIL (National Power Corporation of India Ltd.) 55 days to own their mistakes following Kundankulum incident. They have put their faith in systems and strategies which have been proven, in the past in other countries, to be inefficient and prone to attacks. A study by CrowdStrike, looked into the problem-solving skills and aptitudes of people in charge to measure the ability (or inability) of Indian institutions in tackling cyberattacks from well-equipped nations and organisations. Out of the ten people who respond to such attacks, nine said that the 1-10-60 (1 minute to detect the threat, 10 minutes to assess and quarantine and 60 minutes to resolve the issue) approach laid down by CrowdStrike was difficult to adhere to. Data Safety Indian IT sector growth, in recent years, has prompted the rise of digital data inventory. The vast amount of data accumulation has led to cyberattacks becoming frequent and the government and cybersecurity authorities to implement strict rules and regulations. Digital India Mission and increasing cybersecurity concerns have transformed this area into a multi-billion-dollar industry, currently valued at $4.5 billion, expected to reach $35 billion by 2030. The three sectors which are heavily invested in cybersecurity efforts are the Government, Information and Technology Services and Banking. Perils Of Going Digital People preferences are shifting from traditional to digital modes of transactions, this has led to an increase in the frequency of cyberattacks and each cyberattack is more sophisticated than the other. India was among the top nations which came under ransomware attacks in 2019. Stop, Ryuk and Purga were the most prominent ones. Demands were marked to be as high as five million dollars but usually ranged between $1-5 million on an average. Records revealed that the money spent on resolving the issue and paying for the damages exceeded the ransom amount. Kaspersky, who undertook this study, said that once the systems were breached, the best solution would be to rebuild, initiate an internal investigation and commence audits. PM Narendra Modi’s ‘Digital India’ initiative, though flamboyant in outlook but necessary for the country’s economic prosperity, has been more harmful than helpful. India does not have security personnel, as Taiwan does, who are specifically trained to counter cybersecurity threats. The infrastructure is not up to par yet with vacancies outnumbering qualified individuals. India can better equip themselves by providing vocational cybersecurity guidelines to its professionals. India can also participate in cybersecurity exchanges and programs with other Asian countries, it does run stimulations domestically but the experience gained interacting with other countries can expose India to new opportunities, software and technology and an understanding of other country’s cybersecurity models and platforms. Personal Data Protection Bill On 11th Dec 2019, Minister of Electronics and Information Technology, Mr Ravi Shankar Prasad, introduced the Personal Data Protection Bill in the Lok Sabha. The bill seeks to secure digital collection of data of individuals through the establishment of the Data Protection Authority which will supervise and authorise companies and institutions (domestic and global) from accessing personal information of the citizens of India. National Cybersecurity It has been reported that the Indian army is subject to recurring cyberattacks to the tune of twice a month on an average. India also makes it to the list of top 15 least cyber-secure countries in the world. It has led to the creation of tri-service command which will administer and oversee cybersecurity and Space operations. The Ministry of Defence has appointed two-star officers from the Indian Army and the Indian Navy to lead the Armed Forces Special Operations Division (AFSOD) and the Defense Cyber Agency (DCA) respectively; two-star officers from the Indian Airforce were appointed to head the Defense Space Agency (DSA). The DCA, under Rear Admiral Mohit Gupta, has been assigned two important functions – to fight cyber-crimes and to define and set guidelines to tackle cyber warfare. This initiative may lead to sharing of intel among the three wings which may bolster better communication and friendly relations. Lt. Col. Rajesh Pant, National Cybersecurity Coordinator, announced that the Government of India plans on releasing new cybersecurity Policy by early 2020. This plan will take into consideration incoming technologies (like 5G) too, the last update was in 2013.","excerpt":"Cyberattacks are rife in India, only the US and China are placed higher on this list. Bangalore, Mumbai, Delhi are among states which receives the highest traffic. In the first quarter of 2019, 48% of the total attack was recorded in Chennai. With companies suffering heavy damages (monetary as well as in data theft), cybersecurity […]","categories":["AI Features"],"tags":["Cyberattacks","Cybersecurity India","data protection","Digital India","government of india","Ransomware"],"author_name":"Yeshey Rabzyor Yolmo","publish_date":"2019-12-19T11:00:00","publication_year":"2019","word_count":1023,"keywords":["government of india","Cyberattacks","Digital India","Go","programming_languages:R","AI","Ransomware","R","programming_languages:Go","Git","RAG","Ray","Cybersecurity India","GAN","data protection"],"extracted_tech_keywords":["AI","Ray","RAG","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/cybersecurity-in-india\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":46821,"title":"Not Everyone Can Become A Data Scientist, Says Prof. Charanpreet Singh Of Praxis Business School","content":"Once considered a niche domain, Data Science has now emerged as a well established and essential functional area that helps businesses operate effectively and competitively. The demand for trained data science professionals has increased tremendously. Analytics and data science are a regular part of boardroom discussions across sectors and industries. Naturally, there is a huge surge in the excitement about getting into the data science field. Professionals from different backgrounds and with varied skill-sets are attempting to upskill themselves through formal training programs in full-time, hybrid and online modes to become ‘data scientists’. However, there’s a lot of conflicting narrative out there which is causing a good deal of confusion. On the one end, we hear claims that ‘anyone can become a data scientist’ and then there  industry experts and employers who debate vehemently that data science, like any other science, is for a select few. To get a clarity around these issues, we decided to speak to Prof. Charanpreet Singh from Praxis Business School. Praxis Business School has been the pioneers in Analytics education in India and has the longest history of selecting, teaching and placing students in this domain. 1) Analytics India Magazine: Since inception, Praxis has followed a clear and strict selection process — is that a message that Data Science is not for everyone? Charanpreet Singh: We have always had a selection process for admitting students in our program. I feel this is the responsible thing to do, because we have a full-time program and people leave existing jobs to pursue it. This kind of makes us feel responsible for their lives. As educators, we have to  be sure that the candidate has, in our opinion, a reasonably good chance of success in the field. It is a fact that data science is a red hot career option with high demand, salary and respect, so everyone wants a piece of this action. We believe that Data Science is a complex domain and demands an adequate investment of time and effort to strengthen the knowledge and skill levels from which one can chart career growth. And so, for the other question that if anyone can be a data scientist, my answer is NO. Here is a list of people in our experience who would find it hard to be data scientist: People who aren’t comfortable with numbers, do not enjoy the math of things. People who dreaded higher math and avoided it all through their prior education but now suddenly want to embrace it and absorb it People who think coding is for geeks and they are either averse to it or think they are well above it and would now wish to ‘manage’ things rather than ‘do’ things! People who do not have a penchant for analyzing things or what is popularly called a ‘problem solving attitude’. If you are not obsessed with solving a problem, irrespective of how tough and frustrating it is, and how long it takes, you may be chasing a wrong career. The ‘quick-fix-guys’ who believe be it the course, or the subsequent work  there is always a way around hard work 2) AIM: Praxis boasts of a large and very successful alumni base in analytics. In your experience, what kind of students actually make it big? What are the necessary traits for success in data science? CS: We have been in this for 8+ years now and our alumni has done exceedingly well – in most cases way ahead of our expectations. Our understanding of what works and what does not has also evolved over the years. Academic background certainly helps — but it is not the decisive factor. In order to make good use of this program, one has to have a fundamental comfort with numbers and technology — irrespective of the academic and\/ or professional background one hails from. Secondly, one has to have the courage, curiosity and commitment to learn new things — sounds trivial but the program is complex and rigorous and demands a sustained effort. One needs to look beyond curriculum-driven learning and participate on platforms like Kaggle, AIM and other global hackathons. All this means a lot of hard work. This commitment to learning needs to stay post the program and across ones career to keep pace with the rapidly evolving domain. And finally, a trait that is crucial for success is that one must enjoy the whole process of problem solving — all the hours of finding and cleaning data, trying out different techniques and then articulating the solution. So it’s more about attitude and compatibility with the domain than just about a specific educational qualification. 3) AIM: It seems like educational background or a particular type of experience isn’t a decider for you. What message do you have for someone aspiring to enter this domain? Image Source: Twitter CS: For someone wanting to enter the domain, I would urge them to do a lot of work home before deciding. This can shape up to be the most crucial career decision one has made. Research the domain, understand what kind of work is done in data science, how does a typical day look like in a data scientist’s life – establish that you would like to experience that life. Look beyond articles that talk about demand, and try and understand why is there such a demand. One would always be successful in a field one is good at — do a lot of retrospection and make sure that you have the capability to learn and to perform in this domain. Finally, research the quality of the programs you are applying to, and engage with the institute alumni before you decide. At Praxis we have had a diverse set of students — different educational backgrounds, varying years of experience across multiple sectors and functions; and these students have been able to do well during the program and carve out successful careers in data science. To illustrate the point, here is a small sample of students from Praxis Business School with a summary of their career progression: Name Profile prior to joining Praxis Current Profile Jayesh Baldania Class of 2011 BPharma (B.K. Mody Pharmacy College) Fresher Walmart Canada – Sr. Manager, Pricing – Strategy, Technology and Data Science Sharath Ghosh Class of 2012 BSc Biotech (Durgapur College of Commerce & Science) 1 Year, Mainframe Engineer with IBM Kellogg Company, Dublin – Lead Data Scientist Subhasish Das Class of 2012 BSc Statistics (Ramakrishna Mission, Narendrapur), MBA (CSREM) 1 year, Operations Executive in a Startup CITI – AVP – CCAR Modellling Bishnu Panda Class of 2012 BTech Electrical (WBUT) Fresher Fidelity Investments – Institutional Asset Management – Lead Ritesh Mohan Srivastava Class of 2013 BTech Biotech (Amity) 3 years, Business Analyst with Wipro Novartis – Advanced Analytics Leader Roma Agarwal Class of 2014 BSc Computers (IT BHU) MCA (IT BHU) 4 years, Product Analyst with Cognizant Deloitte – Consultant -Analytics Karthic Krishnan Class of 2016 BCom Loyola, Chennai MBA LIBA, Chennai 9 years, Assistant Manager with IFFCO Tokio IBM – Data Scientist Anirban Mukherjee Class of 2016 BSc Eco (St. Joseph’s Bangalore) 7 Months, Analyst with Tesco EY – Associate Consultant Kirtimaan Gopanayak Class of 2016 BSc Math ( Institute of Mathematics & Application), MSc Stat ( Utkal University) 6 months, Faculty (Central Univ of Orissa) Worxogo – Chief Statistician and Analytics Lead Safayet karim Class of 2016 BSc Stat (Vishwa Bharati Univ) MSc Stat (Aligarh Muslim Univ) 4 years, Research Associate with Centre De Sciences Humaines IBM – Data Scientist Chandrakanth Bajoria Class of 2016 BTech Electrical (Jadavpur) 4 years, Associate Manager with Jindal Steel Microsoft – Business Analytics Specialist Sakshi Singh Class of 2017 BTech Computer (Jayoti Vidyapeeth Women’s University, Jaipur) 2.5 years, Application Developer with IBM Tredence – Sr Business Analyst Nitesh Chowdhury Class of 2017 BBA Finance (St Xavier’s Kolkata) Fresher HSBC – Analyst Tejaswi Nadella Class of 2018 BTech AGFE (IIT KGP) PGP – ABM (IIM A) 1.5 years, Manager with Yes Bank EY – Senior Associate Sravanya Guru Tayi Class of 2019 BTech – Mech (Vignana Bharati Institute of Technology, Hyderabad) 2 year, System Engineer with Infosys Kotak Mahindra – Deputy Manager – Analytics Shweta Mayekar Class of 2019 BTech – CSE (KJ Somaiya) 9 Months, Author with Analytics Insight Bridgei2i Analytics Solutions – Business Analyst","excerpt":"Once considered a niche domain, Data Science has now emerged as a well established and essential functional area that helps businesses operate effectively and competitively. The demand for trained data science professionals has increased tremendously. Analytics and data science are a regular part of boardroom discussions across sectors and industries.  Naturally, there is a huge […]","categories":["AI Trends"],"tags":["how to become a data scientist","MSc data science","msc data science and analytics","MSc. in data science","pgp program in data science","Praxis Business School","praxis data science programme","praxis data scientist programme"],"author_name":"Richa Bhatia","publish_date":"2019-10-04T15:00:16","publication_year":"2019","word_count":1384,"keywords":["MSc. in data science","data science","Go","how to become a data scientist","Rust","API","AI","R","RAG","praxis data science programme","msc data science and analytics","pgp program in data science","Aim","analytics","MSc data science","praxis data scientist programme","Praxis Business School","startup"],"extracted_tech_keywords":["AI","data science","analytics","Aim","RAG","R","Go","Rust","API","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/not-everyone-can-become-a-data-scientist-says-prof-charanpreet-singh-of-praxis-business-school\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10076846,"title":"Finally, Microsoft Experiments with GPT-3 in Healthcare","content":"Two years ago, when Microsoft partnered with OpenAI, owning the exclusive licence of the GPT-3 language model, little did we know its full potential. A lot has changed since then. The company has been aggressively looking at utilising GPT-3, working alongside OpenAI teams to figure out possible solutions and impacts across sectors – and healthcare and life science is definitely on top of that list. “Medical knowledge doubles every 73 days,” said Junaid Bajwa, chief medical scientist at Microsoft, in an exclusive interaction with Analytics India Magazine, at Global AI Summit, Riyadh. He said it is truly about the richness of information – the data coming from publications on medical research, particularly related to ailments and treatments for various diseases and medical conditions across the globe. Looking at the rate at which research papers are published, he estimates that it has the potential to double every three days in a few years. In this backdrop, Bajwa believes that GPT-3 could be the answer to solving various challenges in healthcare, offering more specific, personalised, and result-backed healthcare solutions, treatments, and consultations to people across the globe. Bajwa is a chief medical scientist at Microsoft Research and a practising physician in the NHS (the UK’s National Health Service). Prior to this, he was the global lead for strategic alliance and solutions for the Global Digital of Excellence at Merck Sharp & Dohme. He is also a clinical associate professor at University College London. Bigger picture Bajwa thinks that GPT-3-like models could be used in creating a medical consultation\/treatment recommendation platform for medical practitioners across departments. Giving an example, he said, “If a loved one of ours is in consultation, trying to manage cancer, they will have a specific history, social demographic, ethnicity, etc. But, the treatment for that person is currently based on competitive paths that happened over time.” “What if at that moment in time, you could translate all the knowledge in the universe and say, for this person, who is 55-years old, from a South Asian background, who lives in this particular postcode, etc., this is the treatment regimen for them, these are the side effects, and understanding what works for them or not, and this is the best plan that we can go ahead with,” shared Bajwa, saying that taking that kind of information to deliver amazing insight will be truly revolutionary. “We are not there today, but are on the journey to make that happen,” he added. Microsoft GPT-3 play Last year, Microsoft unveiled its first features in a customer product powered by GPT-3, which helps users build apps without knowing how to write computer code or formulas. The same year, in July 2021, GitHub-owned by Microsoft, in collaboration with OpenAI, also launched a revolutionary new feature called Copilot, an AI-based pair programmer that plugs into the editor and offers real-time coding suggestions. One year later, in June 2022, the company made it available to all developers. These are just a few instances of how Microsoft uses GPT-3 to unleash new possibilities. Microsoft CTO Kevin Scott said that the scope of commercial and creative potential that can be unlocked through the GPT-3 model is profound. Further, he said that in aiding human creativity and ingenuity in areas such as writing and composition, describing and summarising large blocks of long-form data (including code), and converting natural language to another language – the possibilities are endless. Bets Big on Biomedical NLP Bajwa said that there are a few fascinating research happening in the area of Biomedical NLP, led by Hoifung Poon, senior director of Biomedical NLP at Microsoft, where the team is looking at how they apply that to real-world evidence, different datasets, and sources to understand better what is happening in the life sciences, healthcare, etc. Microsoft is betting big on Biomedical NLP, where the team is looking at accelerating progress in precision medicine. Leveraging deep research assets in deep learning and biomedical machine reading, the team focuses on advancing self-supervised learning, like conducting task-agnostic biomedical language model pretraining and proposing a general framework for task-specific self-supervision. Over the years, Microsoft has made some exciting new progress in deep collaboration with Microsoft partners such as JAX, Providence, and now OpenAI. Some of their work in the past includes the creation of BLURB, a comprehensive benchmark and leaderboard for biomedical NLP, and PubMedBERT models. Meanwhile, Google is also working on creating new machine-learning tools and discovering opportunities to increase the availability and accuracy of healthcare globally. Some of its research work, particularly in NLP, includes Underspecification in Scene Description to Depiction Tasks, Retrieval-guided Counterfactual Generation for QA, Source-summary Entity Aggregation in Abstractive Summarisation, PaLI: A Scalable Approach to Joint Modeling of Language and Vision, and InnerMonologue others. Protein structure prediction models Citing one of the MIT articles on ‘Analysing the potential of AlphaFold in drug discovery,’ Bajwa said that AlphaFold has given us new knowledge, a new paradigm shift. “But, the journey from what the initial discovery is, to translating into real molecules, and taking those molecules into the real world, there is still a massive journey for us to undertake,” shared Bajwa. He said, at Microsoft they have been working on similar things. However, they are working on an end-to-end pipeline, beyond the drug discovery aspect, where they have partnered with Novartis, Novo Nordisk, and others to apply to the real-world scientific research breakthrough and, ultimately, put it to real patients at the end of it. “Not just a paper publication, but something that delivers robust impact,” he concluded.","excerpt":"“We are not there today, but are on the journey to make that happen,” says Junaid Bajwa, chief medical scientist at Microsoft","categories":["Deep Tech"],"tags":[],"author_name":"Amit Naik","publish_date":"2022-10-10T12:00:00","publication_year":"2022","word_count":920,"keywords":["Go","OpenAI","AI","Scala","RAG","NLP","deep learning","analytics","JAX","R"],"extracted_tech_keywords":["AI","deep learning","NLP","analytics","OpenAI","JAX","RAG","R","Go","Scala"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/finally-microsoft-experiments-with-gpt-3-in-healthcare\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":30319,"title":"RPA Startup Automation Anywhere Gets $300 Million From SoftBank Vision Fund","content":"Automation Anywhere, a Robotic Process Automation startup, announced that they have raised $300 million from the SoftBank Vision Fund. The funding is subject to regulatory approvals and satisfaction of other customary closing conditions and is an expansion to the company’s Series A round announced in July, bringing the financing to over $500 million. RPA enables companies to automate business processes. By automating routine tasks in the workplace, RPA is increasing productivity and enabling people to focus on more challenging and fulfilling work. Mihir Shukla, CEO and Co-Founder at Automation Anywhere, said in a statement, “RPA is the most pervasive and frictionless path to delivering AI technologies across the enterprise, and is revolutionizing the way people work… With this additional capital, we are in a position to do far more than any other provider. We will not only continue to deliver the most advanced RPA to the market, but we will help bring AI to millions. As the introduction of the PC, we see a world where every office employee will work alongside digital workers, amplifying human contributions. Today, employees must know how to use a PC and very soon employees will have to know how to build a bot.” Praveen Akkiraju, managing partner at Softbank Investment Advisers, said in a statement, “Enterprises of all sizes are in the midst of a major transition to digital platforms and we are excited to partner with Automation Anywhere to accelerate this transformation… We believe that Mihir and team bring a clear vision, a strong technology platform and a passion for delivering real value and cost optimizations for customers.” Automation Anywhere is the only RPA provider that offers a complete suite of products capable of automating business processes at scale. Together, its foundational Enterprise RPA platform, its AI product, IQ Bot™, and Bot Insight™, a predictive analytics capability embedded into bots, are the industry’s most robust and scalable platform for delivering an Intelligent Digital Workforce. Automation Anywhere opened its Bot Store, the industry’s largest app store-like marketplace of downloadable bot applications that are designed by the company and its partner ecosystem to meet specific automation requirements for common business processes. The Bot Store increases the speed and simplicity in deploying bots at enterprise scale, for any size organization. It has been visited by more than 65,000 users since its launch in March 2018.","excerpt":"Automation Anywhere, a Robotic Process Automation startup, announced that they have raised $300 million from the SoftBank Vision Fund. The funding is subject to regulatory approvals and satisfaction of other customary closing conditions and is an expansion to the company’s Series A round announced in July, bringing the financing to over $500 million. RPA enables […]","categories":["AI News"],"tags":["Automation Anywhere","Robotic Process Automation","RPA","Softbank"],"author_name":"Prajakta Hebbar","publish_date":"2018-11-16T08:05:23","publication_year":"2018","word_count":389,"keywords":["API","AI","RPA","R","Scala","Git","automation","Automation Anywhere","Robotic Process Automation","ViT","analytics","GAN","predictive analytics","Softbank"],"extracted_tech_keywords":["AI","analytics","predictive analytics","R","Scala","Git","API","GAN","ViT","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/rpa-startup-automation-anywhere-gets-300-million-from-softbank\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10125105,"title":"Meet the Indian Who Created an Open Source Perplexity Over One Weekend","content":"When AIM asked why Perplexity AI couldn’t be built in India, little did we know that it was actually possible. The best part is that a guy did it over the weekend, and made it open source. Bishal Saha, a dropout from Lovely Professional University, created Omniplex. Speaking to AIM, Saha said that despite initial job rejections from Perplexity AI, it’s his tenacity that led him to creating an open-source alternative that is slowly gaining traction. https:\/\/twitter.com\/Mr_BishalSaha\/status\/1773439390808969630 “I saw Perplexity AI gaining huge traction and thought I could contribute, but they rejected me twice,” Saha recalled, that he had applied for a job at Perplexity, but they were only hiring people from Google. Undeterred, he decided to build his own version. “I built the Omniplex prototype in one weekend. It was just a single-page app, but it started gaining interest after I shared on Reddit,” he explained. Saha has made Omniplex open-source, sharing his code with the world. “I’m not asking for any money. The source code is openly available,” he stated. His commitment to open-source reflects a broader vision of democratising AI. For Saha, the cost per search query is minimal, but scaling remains an issue. Saha leverages credits from OpenAI and Bing APIs, managing costs effectively. “Per search query, it costs me about $0.007, and I’ve got $2,500 in credits from OpenAI. I haven’t spent half of it yet,” he adds. Omniplex is built using the Vercel AI SDK, Firebase, NextJs, and Bing, and currently runs on OpenAI’s GPT model. Saha states that his goal is to eventually support all LLMs such as Claude, Mistral, and Gemma. Perplexity is Nothing But a Hype Saha said that it is actually quite easy to build something like Perplexity. But most alternatives to Perplexity are not gaining much traction. “Even in our discussions, people are not excited about it. But I believe in the power of open-source,” he emphasised. Saha’s insights into the AI industry are sharp and critical. He described Perplexity AI as a “hype to the extent that it cannot be a billion-dollar business”. According to Saha, Perplexity’s user numbers and revenue figures often don’t add up. “I’ve done much more research, and you can also find that they use Bing for their searches and then scrape the top results. It’s not as revolutionary as it’s made out to be,” he asserted. “It’s not a big business; they’re paying a lot to acquire traffic.” In his own words, “They underestimated the power and potential of dedicated developers. I’m proving them wrong, one line of code at a time.” The Indian startup ecosystem is too focused on Indic LLMs Saha said that since he is not from an IIT or NIIT, the bias is evident in his own experiences and those of his peers. “YC looks for the tech pedigree. If you’re not from an IIT, you’re not worth an interview,” he shared, pointing to the broader issues of meritocracy and elitism in the startup world in India, and globally. “Most Indian investors are not into AI; they don’t understand it. They prefer startups from IITs or Ivy League colleges,” he said, while sharing that investors are often reluctant to speak to founders from other universities. Despite this, Saha’s journey is a testament to resilience and innovation. “I’ve worked for two YC-backed companies and did a lot of startups, but being a dropout from LPU, investors don’t like me much,” Saha shares candidly. He initially ventured into fintech, building account aggregators and financial advice apps such as Gullak Money and Pocket Money. However, due to stringent regulations and lack of investment, he pivoted towards AI. Now, Saha is shifting his focus again. “I’m moving out of the AI race. It’s absurd how many startups are entering YC without a clear plan for the next three years,” he said. Instead, he’s exploring augmented reality with a former Google employee in his stealth startup. “When AR goes mainstream, there will be a gap. Web and app developers lack 3D design experience. I want to build a platform like Figma for augmented reality,” he revealed.","excerpt":"“They underestimated the power and potential of dedicated developers. I’m proving them wrong, one line of code at a time.”","categories":["AI Features"],"tags":["Interviews and Discussions","Open Source AI","Perplexity AI"],"author_name":"Mohit Pandey","publish_date":"2024-06-28T10:00:00","publication_year":"2024","word_count":682,"keywords":["Go","API","startup","OpenAI","AI","innovation","Perplexity AI","RAG","Open Source AI","GPT","Aim","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","OpenAI","Aim","RAG","R","Go","API","GPT","innovation","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/meet-the-indian-who-created-an-open-source-perplexity-over-one-weekend\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":41838,"title":"A Day In The Life Of: An Expert In Loyalty Campaigns For Retail Industry","content":"In our column ‘A Day In The Life Of’, we are trying to step into the shoes of the awesome techies from various organisations and sectors who are working in areas like big data, data analytics, artificial intelligence and the internet of things, among others. This week Analytics India Magazine got in touch with Bhavik Shah, the Head CRM and Loyalty at Metro Brands, to talk about his enthusiasm for meticulous data prepping, new technologies and wanderlust. Shah begins his day at 6.30 am, choosing to reach his office at 8.30 am to avoid traffic. He says these early hours in the office are the key to his success, as he uses the peace and quiet in the office to his best advantage. “I try to finish all my task without any disturbance from my team or other internal stakeholders. That is why this is the most productive hour of my day.” As a health enthusiast, Shah tries to plan his days well so that he can squeeze in a workout even as he battles with the daily Mumbai traffic. “I don’t get enough time to pursue a full-fledged hobby, as travel time takes a toll thanks to Mumbai traffic. But I do manage to go for a walk or run in the evening,” he says. He, however, is of the firm belief that healthy work practice goes hand-in-hand with an active and refreshed mind. “It is very important to have an active social life to ensure you have a good work-life balance. I enjoy evening hangouts with friends over coffee — be it at home or at our favourite joints. We also plan weekend getaways every three months. These breaks really help me rejuvenate and there’s nothing better than going out with my life partner and a couple of friends,” he adds. Talking about his work process, Shah says, “My maximum time goes in designing loyalty campaigns, implementing and executing IoT projects and platforms, and building various analytical and statistical reports. I like to deep dive into the data to mine more meaningful analysis and portray this analysis in a graphical manner to make it easier to understand and absorb the data. Being more innovative and creative in presenting the data is my forte.” Right now, Shah is working on: Loyalty solutions CRM implementation Footfall tracker and analysis Repairs automation process Streamlining analytics hub and infrastructure “These projects over the long run will help our business to consolidate the data into one and make it a bigger, smarter data-pool. Every department can access their share of information from this pool, which, in turn, will help them plan better,” explains Shah. When asked about the problem he’s currently working on, Shah says candidly, “Currently, every organisation (in my sector) is facing one of two challenges: Customer-centric problems Store or inventory-related retail solutions “However, none of the big players have a single view platform which could make their life simpler. That is why I have been custom designing the requirement on my own to suit our needs,” he says. Shah is a lucky employee who can happily say that his organisation is making the best use of his talent. “My inputs and ideas are fairly accepted and even executed. So far, two of my projects are executed and three more are in the pipeline.” However, he feels that there’s a lot more to be achieved in this area. “Loyalty and analytics is a never-ending subject. Once you start digging, it will keep leading you to something or the other. The day you feel you have achieved the result, it is actually the time to open new doors and newer ways of looking at the data and designing the programmes and campaigns,” Shah says with a smile.","excerpt":"In our column ‘A Day In The Life Of’, we are trying to step into the shoes of the awesome techies from various organisations and sectors who are working in areas like big data, data analytics, artificial intelligence and the internet of things, among others. This week Analytics India Magazine got in touch with Bhavik […]","categories":["AI Features"],"tags":["Data Science Career","Interviews and Discussions"],"author_name":"Prajakta Hebbar","publish_date":"2019-07-05T10:40:56","publication_year":"2019","word_count":625,"keywords":["big data","Go","artificial intelligence","AI","ML","automation","Data Science Career","Ray","analytics","GAN","R","Interviews and Discussions"],"extracted_tech_keywords":["AI","artificial intelligence","ML","analytics","Ray","R","Go","big data","GAN","automation"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/a-day-in-the-life-of-an-expert-in-loyalty-campaigns-for-retail-industry\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10043091,"title":"Comprehensive Guide To Deseasonalizing Time Series","content":"Time series data is a collection of data points obtained in a sequence with time values. These time values can be regular periods or irregular. We use time-series data to predict the future data responses, which are based on past data. Generally, in a time series, some unusual effect of seasonality or trends and noise makes the prediction wrong. For better forecasting with time series, we need a stationary time series data set in which the effect of trends or seasons is negligible. In the article, we will discuss the seasonality of time series data and remove it. Types of Time-Series There are two types of time series: Additive and multiplicative. To understand both better, we need to know trends, seasonality and noise in time series data. More formally, we can describe those three as: Trend: The trend component makes changes in the overall time series. Seasonality: Change in time series value within a given time. Noise\/Random: Abrupt change in time series excluding the change produced by seasonality and trends. Fig1. Graphs for seasonality, trend and random. Image source In the above image, we can see how seasonality, trend and noise affect the whole observation of an additive time series data set. Interaction of those three in a dataset determines the type of time series data. Additive Time Series: In a time series, trend, seasonality, and noise make the additive time series. Time-Series = trend + seasonality + noise Soruce: https:\/\/www.slideshare.net\/ankit_ppt\/lesson-1-introductiontotimeseries Multiplicative Time Series: Multiplication of trend, seasonality, and noise make the time series multiplicative. Time-series = trend × seasonality × noise Source: https:\/\/www.slideshare.net\/ankit_ppt\/lesson-1-introductiontotimeseries Here in the above, we have seen the basics of a time series data set. We will discuss the seasonality of a data set and how to deseasonalize time series in the next step. Seasonality In a time series, seasonality is a component that tells us the changes or fluctuations are occurring in a repeated way for similar periods. For example, sales of umbrellas increase in the rainy season; it increases because rain can happen only once a year but will happen every year; hence we can say that there is a seasonality effect in the sales of umbrellas. A cyclic structure in a data set can be seasonality if the frequency of the trend graph is increasing or decreasing repeatedly but for a particular time. Understanding seasonality can improve the forecasting results. However, to make a clear relationship between the input and output some time we need to remove the seasonality. Removal of seasonality is called deseasonalizing time series. Many types of seasonality depend on the time series and frequency of fluctuations. Like Time of the day Daily WeeklyMonthlyyearly After removal of seasonality from time series, we can consider it as a seasonal stationary time series. For learning about deseasonalizing I am using airline passenger data set. In the data set, we have the records for passenger count of every month from 1949 to 1959. The data is having both trend and seasonality. We are going to remove the seasonality in the next steps. Code Implementation of Deseasonalizing Time Series Setting up the environment in google colab. Requirements : Python 3.6 or above, Importing the basic libraries : import pandas as pd import numpy as np import matplotlib.pyplot as plt %matplotlib inline Reading the data set: Input: data = pd.read_csv(\"\/content\/drive\/MyDrive\/Yugesh\/deseasonalizing time series\/AirPassengers.csv\", index_col=0, parse_dates=True) data.head() Output: Here we can see that the data set has the month column as index column and count of passengers in column 0. Let’s check for the trend graph of the dataset. Input : data.plot() Output: The dataset trend shows that it is a kind of additive time series, but this feature dataset also has a slight seasonality and trend. Let’s check for the two consecutive years trend to know the similarity more accurately. I am choosing the years 1957 and 1958 for the test. Input : data[\"1957\"].plot(kind='bar') data[\"1958\"].plot(kind='bar') Output: Here for the years 1957 and 1958, we can see that the amplitude of the trends is quite similar. Just a small amount of passenger count over the year has increased, but we would know that there is a seasonality effect in passenger count if we draw a trend. Knowing it better, we can decompose the data set into its components(seasonality, trend and noise) to decompose the data ‘statsmodels’ package has provided a function ‘seasonal_decompose’ under ‘statsmodels.tsa.seasonal’ module. Importing seasonal_decompose : from statsmodels.tsa.seasonal import seasonal_decompose Let’s check for the components : Input: decompose_data = seasonal_decompose(data, model=\"additive\") decompose_data.plot(); Output: Here in the above chart, we can see the decomposed structure of data and the structure of the components in the data set which were affecting it. Let’s make a graph for available seasonality. Input : seasonality=decompose_data.seasonal seasonality.plot(color='green') Output: In the seasonality graph, we can see the seasonality structure for every year, which is cyclic and repeatedly providing the same value. To check for the stationarity of the time series, statsmodels provides a plot_acf method to plot an autocorrelation plot. Input : from statsmodels.graphics.tsaplots import plot_acf plot_acf(data); Output: Here the blue area is the confidence interval, and the candles started coming inside after the 13th candle. This can be due to the seasonality of 12-13 months. We can cross-check it by dicky-fuller method. For more information about the test, you can visit this link. The statsmodels provides a function to perform the test. Importing function to perform the test: from statsmodels.tsa.stattools import adfuller Testing the data set by dicky-fuller method: Input : dftest = adfuller(data.Passengers, autolag = 'AIC') print(\"1. ADF : \",dftest[0]) print(\"2. P-Value : \", dftest[1]) print(\"3. Num Of Lags : \", dftest[2]) print(\"4. Num Of Observations Used For ADF Regression and Critical Values Calculation :\", dftest[3]) print(\"5. Critical Values :\") for key, val in dftest[4].items(): print(\"\\t\",key, \": \", val) Output: Here in the output, we can see that the p-value of the data set is more than 0.05. Because of this reason, only we can interpret the data as non-stationary. As we have seen that data is non-stationary, we can apply deseasonalization to the data set to make it more stable or stationary. Let’s perform the deseasonalization on the data set. Differencing over log-transformed time-series We try to normalize the seasonality value by the difference of log to passenger count and shifted the log value of passenger count to one step. Input : log_passengers = pd.DataFrame(data.Passengers.apply(lambda x : np.log(x))) log_diff = log_passengers - log_passengers.shift() ax1 = plt.subplot() log_diff.plot(title='after log transformed & differencing'); ax2 = plt.subplot() data.plot(title='original'); Output : In the output, we can compare the trend of the graph after deseasonalizing the data. Let’s check for the p-value of the new time series. Input : test = adfuller(log_diff.dropna().Passengers) print(\"p-value :\", test[1]) Output: The p-value is again greater than 0.05, so we can interpret the data as still non-stationary. Differencing over power-transformed time series We have first power transformed the data and then made a difference between power transformed data and one shift. Input: powered_transform = data.Passengers.apply(lambda x : x ** 0.5) powered_transform_diff = powered_transform - powered_transform.shift() ax1 = plt.subplot() powered_transform_diff.plot(title='after power transformed & differencing'); ax2 = plt.subplot() data.plot(title='original'); Output: After this, we can check the p-value using dicky – fuller test. Input: test = adfuller(powered_transform_diff.dropna().Passengers) print(\"p-value :\", test[1]) Output: Here in differencing overpower transformed time series, we have got a good p-value near about 0.02 and lower than 0.05 in that we can consider over data is stationary. Still, there are some more methods let’s just check for the result on those methods also. Differencing over rolling mean taken for 12 months: Input: rolling_mean = data.rolling(window = 12).mean() rolling_mean_diff = rolling_mean - rolling_mean.shift() ax1 = plt.subplot() powered_transform_diff.plot(title='after rolling mean & differencing'); ax2 = plt.subplot() data.plot(title='original'); Output: Let’s check for the p-value using the dicky-fuller method. Input test = adfuller(rolling_mean_diff.dropna().Passengers) print(\"p-value :\", test[1]) Output: Here we can see that the p_value is again less than 0.05. It means by the different methods; we are improving the stationarity of the dataset. Differencing over log-transformed & mean rolled time series: In this, we have applied the difference between the log transformation of the rolling mean and its shifted value by one step. Let’s check for the results. Input: logged_transform = pd.DataFrame(data.Passengers.apply(lambda x : np.log(x))) rolling_mean = logged_transform.rolling(window = 12).mean() diff = rolling_mean - rolling_mean.shift(1) ax1 = plt.subplot() diff.plot(title='after log transformed rolling mean & differencing'); ax2 = plt.subplot() data.plot(title='original'); Output: We can see that it has distorted the seasonality; it can be interpreted as this method is not as good as the other methods were. Let’s check for the p-value. Input: test = adfuller(diff.dropna().Passengers) print(\"p-value :\", test[1]) Output: As assumed, the p-value of the dataset is greater than 0.05. The dataset is not stationary. Differencing over power transformed & rolling mean time series This method will try to adjust seasonality using the difference between power transformed rolling mean and shifted by one step of power transformed rolling mean of data. Let’s check for the results. Input : powered_transform = pd.DataFrame(data.Passengers.apply(lambda x :  x ** 0.5)) rolling_mean = powered_transform.rolling(window = 12).mean() diff = rolling_mean - rolling_mean.shift(1) ax1 = plt.subplot() diff.plot(title='after power transformed rolling mean & differencing'); ax2 = plt.subplot() data.plot(title='original'); Output: In the output, we again see the distortion in the seasonality. let’s check for the p-value Input : test = adfuller(diff.dropna().Passengers) print(\"p-value :\", test[1]) Output: The p-value is one of the best we are having, but graph seasonality was not good; we interpreted the trend component as highly available after data after seasonality. So to improve it more we can go for detrending also. We have seen our data set was pretty clean and was in an ideal condition, but talking about the real world problem, the datasets do not behave like this generally. So we need to perform more and more tasks on the data set to make our predictions more accurate. In this article, we discussed the time series, had a basic overview of components of a time series, and performed differencing methods for deseasonalizing the time series data to obtain accuracy in our further modeling process. References All the information in this post is gathered from: Pandas timestamp data basics Statsmodels introduction and modules Google colab for python codesdataset","excerpt":"Time series data is a collection of data points obtained in a sequence with time values. These time values can be regular periods or irregular. We use time-series data to predict the future data responses, which are based on past data. Generally, in a time series, some unusual effect of seasonality or trends and noise […]","categories":["Deep Tech"],"tags":["Guide","Time Series","Time Series Analysis"],"author_name":"Yugesh Verma","publish_date":"2021-07-08T11:00:00","publication_year":"2021","word_count":1688,"keywords":["Time Series Analysis","Go","NumPy","TPU","AI","RPA","Colab","Python","Time Series","Matplotlib","R","Guide","Pandas"],"extracted_tech_keywords":["AI","Colab","Pandas","NumPy","Matplotlib","TPU","Python","R","Go","RPA"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/comprehensive-guide-to-deseasonalizing-time-series\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":64965,"title":"Case Study: How Intelligent Automation Helped This Indian Travel Provider To Streamline Their Business Process During The Crisis","content":"COVID-19 pandemic brought disruption to many industries; the travel industry is one of the most hit during this crisis. With people being mandated to stay at home, hotels, airlines, as well as online travel agencies, are struggling to keep up their “Business As Usual.” However, Thomas Cook India has been one such company that has leveraged the benefits of emerging technologies to run its business process effectively amid this crisis. Thomas Cook India, an integrated travel services company, with operations across 29 countries in five continents offers a broad spectrum of services including foreign exchange, corporate travel, visa and passport related services, as well as e-business. With a team of over 9,700 individuals and combined revenue of over $ 0.96 billion for the financial year ended March 31, 2019, Thomas Cook India has been a large travel service provider, and therefore managing business processes during this crisis has been a top priority for the company. Thomas Cook India realised that traditional initiatives were no longer sufficient for the transformation of the business in this fast-evolving technology landscape. Not only the company required centralisation of their business processes but also realised that traditional cost-reduction processes were no longer sufficient for the company. The back office of the company has also been going under major disruptive revolution as the process was changing the way it functions. And, at this juncture, the company decided to leverage the benefits of intelligent automation and virtual robotic workforce in partnership with Automation Anywhere to operate and engage in the Shared Services Area amid this crisis. The company is also exploring some machine learning solutions for a few of its processes. When asked, Geeta Degaonkar, Sr. VP of Shared Services at Thomas Cook India said that “In terms of ‘business continuity plan’, considering we had to get into the lockdown in a day’s time frame, we were still able to run our bots remotely, which has been extremely helpful for us.” Alongside, the implementation partner, Automation Anywhere, has continuously provided support to the company in order to have a smoother process amid this lockdown. “In fact, if a bot goes down because of any technical issue, it gets difficult to connect with it, the technicians of Automation Anywhere have been helpful and effective in solving the issues. Our operations team are able to review and monitor the bots remotely and are also able to gain the necessary output.” Onboarding the right vendor was indeed a concern for Thomas Cook India, where they evaluated a few companies before bringing in Automation Anywhere. The company assessed factors like successful case studies, product offerings network on the ground, available ecosystem, and the right skilled people who can support them, and understand the Indian industry and the importance of cost-efficiency for Indian companies. The company found Automation Anywhere to be flexible and nimble on the ground with the right attitude to understand business problems to provide appropriate solutions and decided to introduce intelligent automation, which delivered business value in a short period and leveraged robotic process automation to remove any obstacles in the path of their automation journey. To start their journey, the company deployed bots withing their Shared Services which supports various departments across, finance, accounting and business operations. The company providing support to their Shared Services, which, in turn, supported different entities across Thomas Group with different IT landscape, infrastructure and processes, therefore required intelligent automation to bring in all the operation of the entities under one umbrella for integrating core platforms in each of these entities. “RPAs fit really well for us wherein we could bring in multiple entity related processes, which were standardised to a certain extent but were still not producing enough efficiency; and there was a manual intervention, which was then dealt with required RPA solution,” said Degaonkar. Degaonkar also believes that digitisation is the only way forward for businesses to deal with this crisis and other potential ones in the future. Businesses need to digitise as much as possible in terms of enabling people to work remotely, and the critical process can go on operating without human intervention. “The businesses that are sitting on the fence will start jumping into leveraging new-age technologies like bots, machine learning that has enough case studies to substantiate that it is going only to support and help businesses to work very efficiently and smoothly given the situations that might come up in future,” said Degaonkar. Considering all the industries are currently struggling with their overall revenue, the cost has been a massive concern for businesses, “these new-age technologies like bots will also help in improving their bottom line by reducing their cost,” said Degaonkar. By building a next-gen shared service operation, Thomas Cook India managed to lower their cost at the SSC by 30 – 40% and also managed to create a scalable platform for driving future growth and improving customer experience. Alongside, the SSC was also able to strengthen internal controls, minimise risks related to errors due to human intervention, frauds, non-compliance, and delays, which, in turn, enhanced customer experience. By deploying these bots, the SSC managed to completely omit human intervention in several of its departments and processes, and therefore, received 100% accuracy, even during the crisis. These bots also helped the SSC to save thousands of human work hours, which would have been lost in repetitive and mundane activities. The company aimed to transform their processes with the help of robotics and automation, and currently running with 15 bots deployed at various depths of the company, including finance and operations. The company has also deployed automation in critical bank reconciliations along with credit noting, tracking of refunds from airlines and receipt clearances, to name a few. When asked, Milan Sheth, EVP – IMEA, Automation Anywhere, he said, “Our plans remain to stay by the side of the customers and align our technologies to meet their business objectives. Currently, we are helping Thomas Cook India, to formulate their business continuity plans by enabling the required technology shift to meet the current requirement.”","excerpt":"COVID-19 pandemic brought disruption to many industries; the travel industry is one of the most hit during this crisis. With people being mandated to stay at home, hotels, airlines, as well as online travel agencies, are struggling to keep up their “Business As Usual.” However, Thomas Cook India has been one such company that has […]","categories":["AI Features"],"tags":["Intelligent Agent","Robotic Process Automation"],"author_name":"Sejuti Das","publish_date":"2020-05-11T15:00:00","publication_year":"2020","word_count":1006,"keywords":["Go","machine learning","TPU","AI","Scala","Git","RAG","Aim","Robotic Process Automation","Intelligent Agent","ViT","R"],"extracted_tech_keywords":["AI","machine learning","Aim","RAG","TPU","R","Go","Scala","Git","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/case-study-how-intelligent-automation-helped-this-indian-travel-provider-to-streamline-their-business-process-during-the-crisis\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10073580,"title":"There’s a Whole ‘Learn X in Y Minutes’ Industry","content":"“Hello World, it’s Siraj! I’m a technologist on a mission to spread data literacy,” developer educator and YouTuber Siraj Raval’s bio on his YouTube channel declares. Raval, according to his LinkedIn profile, is the head of developer education at soaring Web3 startup Polygon Technology. He has a 43-minute Udemy course that explains what the Mumbai-based startup does, in a video titled, ‘Blockchain Development Guide: Polygon’. In 2019, San Francisco-based Raval was found to have plagiarised his quantum research paper by more than 90%. Andrew M Webb, a research associate from the University of Manchester, first noticed the glaring similarities between Raval’s paper ‘The Neural Qubit’ and Nathan Killoran’s ‘Continuous-variable quantum neural networks‘ during Raval’s livestream. He was forced to post an apology video on Twitter saying, “I made the video and paper in a week to align with my “2 videos per week” schedule. I hoped to inspire others to research. Moving forward, I’ll slow down and be more thoughtful about my output.” Raval’s statement signals the pressure most content creators face to churn out quick yet credible material online. This is especially tricky in serious disciplines like academia or training in professional courses where credibility is everything. Unperturbed, Raval has since moved on from the controversy to greener pastures. Until a month ago, he had been releasing videos focused on Bitcoin and Web3 ventures. Notably, his earlier instructional video ‘How to build a healthcare\/AI\/Bitcoin startup’ has a run time of less than an hour! Quick deadlines, shorter videos Things have only gathered speed with each day. A number of channels claim to cover basic programming languages like Python, Java, C++ and HTML within a matter of minutes. The unrealistic timelines beg the question — Why is everybody in such a rush? Learning a skill or a programming language isn’t the invention of the internet. The earliest version of these crash courses were first found in books — the vastly popular ‘Teach Yourself C in 21 days’ have been used by developers since the early 90s. The proliferation of online courses has however rushed the timelines to less than 10 minutes. Traditionally, experts have rallied against the depth of these courses. German-American professor and author of ‘How to Design Programs’ Matthias Felleisen discussed the trend in his book saying, “Bad programming is easy. Idiots can learn it in 21 days, even if they are dummies.” More accessible and practical A large portion of the developer community have contested this as an elitist mentality. Heated debates around the subject on forums like Y Combinator had a number of developers hailing the accessibility of these courses. In the words of a developer, “Not everyone is Ivy League educated and earns millions working for Google”. Others faulted the current educational system, which is geared towards passing standardised tests rather than learning itself. “Someone going through such a system might be excused for thinking that the way you learn something is by cramming facts until you can repeat them once, and then forgetting the whole thing because that test isn’t coming back,” a member commented. Developers have simply been forced to adapt to the current fast-paced environment. For job applicants who are about to attend full stack interviews, crash courses can be a quick refresher. On LinkedIn, a number of developers who have cracked interviews recommend online courses like Codehelp, Pepcoding and CodeWithHarry as resources for quick revisions before the Data Structure and Algorithms or DSA interviews. On the flipside, crash courses are also often the introductory step to coding. Indian content creators in the space use their understanding of the local context and create a personality to attract more viewers. “He (CodeWithHarry) uses desi hindi to explain concepts with real life examples and real scenarios,” one said. While another commented saying, “He keeps me motivated. I started my coding journey with him.” Another practical use for crash courses could be if developers want to compare different aspects of languages. A number of developers have also learned front-end development and Reactjs from these channels to create their own websites. For better or worse, these video tutorials have propagated the idea that programming could and should be much easier than it was.","excerpt":"For job applicants about to attend full-stack interviews, crash courses can be a quick refresher.","categories":["AI Trends"],"tags":[],"author_name":"Poulomi Chatterjee","publish_date":"2022-08-25T15:12:18","publication_year":"2022","word_count":697,"keywords":["Go","TPU","AI","neural network","ML","Python","Aim","C++","R","Java"],"extracted_tech_keywords":["AI","ML","neural network","Aim","TPU","Python","R","Go","Java","C++"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/theres-a-whole-learn-x-in-y-minutes-industry\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10170817,"title":"Microsoft Launches Magentic-UI to Automate Web Tasks with Human Oversight","content":"Microsoft has open-sourced Magentic-UI, a research prototype that allows users to automate web-based tasks while retaining control through a transparent, interactive interface. The tool uses a multi-agent system capable of browsing websites, executing code, and analysing files. The company said Magentic-UI is “especially useful for web tasks that require actions on the web, deep navigation through websites not indexed by search engines or tasks that need web navigation and code execution.” Magentic-UI is powered by AutoGen’s Magentic-One system and includes five specialised agents, namely, Orchestrator, WebSurfer, Coder, FileSurfer, and UserProxy. The Orchestrator acts as the lead agent and coordinates the workflow, while WebSurfer can interact with live websites and perform actions like clicking, typing, and uploading files. Coder and FileSurfer handle the execution of Python or shell commands and file conversion tasks, respectively. The UserProxy enables collaboration with the human operator. One of the system’s key features is its co-planning interface, where users and agents collaborate to define a step-by-step plan before execution. “Collaboratively create and approve step-by-step plans using chat and the plan editor,” the company said. Users can also edit the plan, “add, delete, edit, regenerate steps, and write follow-up messages to iterate.” Magentic-UI introduces additional controls, including “Action Guards,” where “sensitive actions are only executed with explicit user approvals,” and session indicators that signal when input is needed or a task is complete. The platform also supports “parallel task execution,” letting users run multiple workflows simultaneously. Another significant feature is plan learning and retrieval. The system can learn from previous runs to improve future task automation and automatically or manually retrieve saved plans in future tasks. Magentic-UI is built with Docker and can be installed on macOS, Linux, or Windows (with WSL2). Users can install it using pip and access the interface via a local port. Additional dependencies support integration with Azure and Ollama models. The interface has dual panels with a session navigator and a session workspace. The session workspace displays both the task plan and a live browser view. The system updates progress in real time and lets users pause or intervene during task execution. Microsoft describes Magentic-UI as “a platform to study human-agent interaction and experiment with web agents.” The system is intended not just for automation but also for research into how users interact with intelligent agents while maintaining oversight.","excerpt":"Magentic-UI is powered by AutoGen’s Magentic-One system and includes five specialised agents, namely, Orchestrator, WebSurfer, Coder, FileSurfer, and UserProxy.","categories":["AI News"],"tags":["Microsoft"],"author_name":"Siddharth Jindal","publish_date":"2025-05-27T13:54:23","publication_year":"2025","word_count":387,"keywords":["cloud_platforms:Azure","AI","R","docker","automation","Python","AutoGen","programming_languages:Python","llm_models:Llama","Azure","Microsoft"],"extracted_tech_keywords":["AI","AutoGen","Azure","docker","Python","R","automation","llm_models:Llama","cloud_platforms:Azure","programming_languages:Python"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/microsoft-launches-magentic-ui-to-automate-web-tasks-with-human-oversight\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10043358,"title":"EleutherAI’s GPT-J vs OpenAI’s GPT-3","content":"OpenAI’s not so open GPT-3 has an open-source cousin GPT-J, from the house of EleutherAI. Check out the source code on Colab notebook and a free web demo here. EleutherAI, founded by Connor Leahy, Leo Gao, and Sid Black, is a research group focused on AI alignment, scaling and open-source AI research. In March 2021, the company released two GPT-Neo models with 1.3 billion and 2.7 billion parameters respectively. Microsoft has the exclusive access to GPT-3’s source code as part of a larger agreement between the two companies. Microsoft has invested $1 billion. Interestingly, OpenAI’s GPT-1 and GPT-2 are open-source projects. EleutherAI EleutherAI project began on July 3, 2020, with the quest to replicate OpenAI GPT-family models. The north star of the research group is to replicate GPT-3 175 billion parameters and ‘break OpenAI-Microsoft monopoly’ on transformer-based language models. However, to build such powerful models, you need a massive amount of computing power. EleutherAI is currently supported by Google and CoreWeave (cloud computing providers). CoreWeave has offered high-performance GPU compute to develop future models with GPT-NeoX. GPT-NeoX is an in-development codebase based upon Megatron-LM and DeepSpeed and is designed for GPUs. Its GPT-Neo, on the other hand, is a codebase built on Mesh Tensorflow, designed for training on TPUs. Besides this, the research group has built 825 gigabytes (GB) of language modelling dataset called The Pile, curated from a set of datasets including arXiv, GitHub, Wikipedia, StackExchange, HackerNews, etc. Now, it has launched GPT-J, one of the largest models that EleutherAI has released till date. GPT-J is a 6 billion parameters model trained on The Pile, comparable in performance to the GPT-3 version of similar size — 6.7 billion parameters. “Because GPT-J was trained on GitHub (7 percent) and StackExchange (5 percent) data, it is better than GPT3 175B at writing code. However, in other tasks, it is significantly worse,” wrote artificial intelligence expert Alberto Romero, in his blog. GPT-J: JAX-based (Mesh) Transformer LM The name GPT-J comes from its use of JAX-based (Mesh) Transformer LM, developed by EleutherAI’s volunteer researchers Ben Wang and Aran Komatsuzaki. JAX is a Python library used extensively in machine learning experiments. GPT-J is the best performing publicly available Transformer LM in terms of zero-shot performance on various down-streaming tasks. Komatsuzaki said it allows more flexible and faster inference than TensorFlow and TPU counterparts. More than anything, the project required a substantially smaller amount of time than other large-scale models. JAX, xmap and TPUs are the right set of tools for the quick development of large scale models, he added. Our model design and hyperparameter choice closely follow those of 6.7B GPT-3 with some differences, including: The model was trained on 400 billion tokens from The Pile dataset with 800 GB text.Efficient attention (like linear, local or sliding window, etc.) was not used for simplicity, as it would not have significantly improved ‘throughput’ at this scale.The dimension of each ‘attention head’ was set to 256, which is more than that of GPT-3 of comparable size. “This noticeably improved the ‘throughput’ with minimal performance degradation,” said Komatsuzaki. The team made two minor architectural improvements for GPT-J–Rotary embedding for slightly better performance, and placed the attention layer and the feedforward layer in parallel for decreased communication. Performance As shown in the below table, the zero-shot performance is on par with GPT-3 of comparable size, and the performance gap from GPT-3 is closer than the GPT-Neo models. Performance across GPT-family of models (Source: Aran Komatsuzaki) Also, the throughput of the 6 billion GPT-for training (151K tokens\/s) is faster than the 2.7 billion GPT-Neo (148k tokens\/s) on the same hardware (TPU v3-256 pod), showcasing nearly 125 percent improvement in efficiency. The hardware has a ‘theoretical maximum’ of 13.4 PFLOPs (Peta floating-point operations per second), and GPT-J achieved 5.4 PFLOPs as measured in the GPT-3 paper (excluding attention computation and ignoring compute-memory tradeoffs like gradient checkpointing). “When taking these additional factors into account, approximately 60% of the theoretical maximum is utilised,” mentioned Komatsuzaki, saying GPT-J took roughly five weeks with TPU v3-256 pod. GPT-J vs GPT-3 Max Woolf, a data scientist at BuzzFeed, recently tested GPT-J’s coding abilities. He said he ran GPT-J against the test prompts he had used to test GPT-3 a year ago. “The exception is code generation, where GPT-J performed very well, and GPT-3 had performed very poorly,” he wrote, in his blog post, showcasing multiple examples and use cases. I got the 6B parameter GPT-J-6B running, and am testing it with my GPT-3 experimental prompts, in bold (Thread)First, Revenge of the Sith. pic.twitter.com\/KWe21Xozio— Max Woolf (@minimaxir) June 9, 2021 Wrapping up Romero said the results are impressive. He said it is just another GPT model. But on closer look, clear differences emerge. For instance, GPT-J is 30 times smaller than GPT-3 with 175 billion parameters. “Despite the large difference, ‘GPT-J’ produces better code, just because it was slightly more optimised to do the task,” he added. Further, he said further optimisation could give rise to platforms way better than GPT-3 (and not limited to coding). GPT-3 would become a jack of all trades, whereas the specialised systems would be the true masters, added Romero. Recently, the Chinese government-backed BAAI introduced Wu Dao 2.0, the largest language model to date, with 1.75 trillion parameters. It has surpassed Google’s Switch Transformer and OpenAI’s GPT-3 in size.","excerpt":"OpenAI’s not so open GPT-3 has an open-source cousin GPT-J, from the house of EleutherAI. Check out the source code on Colab notebook and a free web demo here.  EleutherAI, founded by Connor Leahy, Leo Gao, and Sid Black, is a research group focused on AI alignment, scaling and open-source AI research. In March 2021, […]","categories":["Global Tech"],"tags":["EleutherAI","GPT-3","GPT-J","Machine Learning","Machine Learning Latest","Machine Learning New","open source data science projects"],"author_name":"Amit Naik","publish_date":"2021-07-12T11:00:00","publication_year":"2021","word_count":889,"keywords":["GPT-3","open source data science projects","Machine Learning New","artificial intelligence","AI","machine learning","OpenAI","cloud computing","TPU","Machine Learning","Machine Learning Latest","GPT-J","Colab","Python","JAX","EleutherAI","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","OpenAI","TensorFlow","JAX","Colab","cloud computing","TPU","Python"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/eleutherais-gpt-j-vs-openais-gpt-3\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10040393,"title":"Comprehensive Guide To Ensemble Methods For Data Scientists","content":"A machine learning model is a mathematically designed algorithm that is trained with a collection of data. The nature of data defines the problem, such as regression, classification and language generation. Whatever the model’s performance during training and evaluation is, the model will surely struggle if it encounters some unforeseen data, i.e. data with a different pattern. Since the future is not in our hands, one can not think about unforeseen patterns while developing a model. Many algorithms can be used to model a specific problem. A particular algorithm can also be hyper-tuned to arrive at different models. Each model has its own ability to generalize on the data. In other words, each model can grasp certain key patterns that other models may miss. Thus, a model may fail to perform well on unforeseen data, but another model may perform better. When yet another new pattern arrives in data, the former model may perform better than the latter. Therefore, one can never assure that a specific model or algorithm will perform great for a specific kind of problem. In this scenario, how can one face the issue of unforeseen patterns in the future data? Ensemble Learning gives a robust solution to this issue. Ensemble Learning is the process of gathering more than one machine learning model for a task in a mathematical way to obtain better performance. If a few different models are developed, each for a unique expected pattern, their ensemble can tackle any pattern. Therefore, an ensemble of many models can outperform any individual machine learning model that it contains. The overall objective of ensemble learning (image source) Different Ensemble Approaches Ensemble methods are classified based on the mathematical way they are formulated. The most popular ensemble approaches are: BaggingBoostingStacking Bagging is the generalized approach of training a couple of individual weak models on subsets of training data and obtaining the final prediction from the individual predictions. The performance of one model is independent of the other. Hence bagging is also called parallel ensemble learning. Each model is training on a subset of training data sampled with replacement. The selection of the data subset is completely randomized. An example data in the training set may either be selected for training or be left. In other words, an example data may be selected for training none of the models, one, or more than one model. This random sampling helps reduce variance in the data. Thus bagging gives a robust ensemble for data with high variance. The sampling approach is called bootstrap sampling. This leads to the name Bootstrap Aggregating, or in short, Bagging. The individual models are independent of each other. It is a custom that individual models are formed from the same algorithm with different configurations. The final predictions of individual models are either voted for by the majority or averaged to arrive at the global prediction. Decision trees are famous for modeling on high-variance data. A randomized ensemble of a number of decision trees is called a Random Forest. Random Forests are well known for their performance in both classification and regression problems when the data has high variance. Boosting is a very systematic approach compared to bagging. In boosting, the training data is sampled without replacement such that each data example is used exactly once. In other words, training data is split into subsets whose count is equal to the number of individual models used. Boosting a sequential ensemble learning approach, in which a model is first trained on one subset. Another model is trained on another training subset to reduce the error caused by the first model. The successor model is trained on yet another training subset to reduce the error caused by its predecessor model. Thus each individual model is dependent on other models. The individual models may be formed from the same algorithm or different algorithms. However, it should be ensured that the individual models are weak enough not to overfit on its subset so that at the end of the training, the ensemble converges to a solution. Boosting is a good approach when the data has high bias. Since the entire training process intends to reduce error at each step, the subset for each model should be representative of the overall training data. Though the boosting approach is preferred for data with high bias, it also helps mitigate the effects of variance. Gradient Boosting, Adaboost, Extreme Gradient Boost are the most popular boosting algorithms in use. Stacking is an ensemble learning approach that stands in between bagging and boosting. It encourages a set of individual models, each trained individually on the entire training dataset. It employs a final model that predicts the output modelled on the outputs of individual models. Thus, one final model is used to boost the outputs of independent models. In contrast to bagging and boosting approaches which train their individual models on subsets of data, the individual models in stacking train up on all the training data. This approach is quite useful if the data is limited. Read more here. Ensemble Methods in Practical Usage Decisions in finance, medical diagnosis, e-commerce, weather forecasting, space science, remote sensing, computer security and anomaly detection are very hard to make because of the high bias and\/or variance in the data and variations in patterns with respect to time. Ensemble methods help find models that are generalized on the data and yield better results. Boosting or bagging are the mostly preferred approaches in such scenarios. Find a loan approval example here. Bagging vs Boosting We might always come across debates on the best ensemble methods. Tree-based methods are good at data with high variance. Bagging and boosting methods that incorporate tree-based individual models become quite popular in numerous applications. The top-preferred bagging method that employs tree-based individual models is Random Forest, and the top-preferred boosting method that employs tree-based individual models is Extreme Boost Gradient. Both the methods can be applied to regression problems and classification problems. However, the selection of a method is purely based on the training data. For regression: RandomForestRegressor, XGBRegressor For classification: RandomForestClassifier, XGBClassifier Read more about Bagging and Boosting comparisons here. AdaBoost method The oldest boosting method is Adaptive Boosting, technically known as AdaBoost. The concept behind the implementation of AdaBoost helped develop numerous boosting methods, including the Extreme Gradient Boosting method. Similar to most other boosting methods, AdaBoost also supports both regression and classification problems. By default, AdaBoost employs Decision Trees (classifier or regressor) as its individual methods. However, users can opt for any suitable algorithms as individual methods. Read more about AdaBoost implementation here. How to use Different Algorithms in Ensemble? A typical hybrid ensemble learning approach for a classification problem (image source) We have discussed that ensemble methods such as Random Forests, AdaBoost, Extreme Gradient Boosting employ decision trees as individual models by default. But, we wish to attempt different algorithms and choose the best decisions later. Voting is a suitable Ensemble approach that follows the bagging strategy and selects the final output by performing majority voting. For instance, we may build a couple of logistic regression models, decision tree classifiers, support vector machines, K-nearest neighbours classifier and Naive Bayes Classifier for a classification task. Each classifier is independent of each other in training and making predictions. The final classes can be determined by performing majority voting among the individual results. This approach is generally called hybrid ensemble learning. Read more about hybrid ensemble learning here. Handling Imbalanced Data via Ensemble Learning Class imbalance is a major problem in classification applications. When most of the data belong to one specific class, and a few of the data belong to another specific class, a classifier struggles to generalize classes. Because it sees a lot of data and learns well in one class and fails to formulate patterns belonging to the minority class. Ensemble Learning has a collection of weak classifiers that can not learn the majority class pattern like a strong individual classifier. Thus a strong individual classifier gives a poor performance on imbalanced classes, while an ensemble of weak classifiers performs greatly. Read more here. Ensemble of Deep Neural Networks The use of ensemble methods began with regression and classification problems with machine learning models. However, with improved compute power nowadays, ensemble methods become a gift to deep learning. There is a general complaint on deep learning architectures’ explainability. It is hard to understand and explain what exactly happens at hidden layers. Without a clear understanding, it becomes difficult to hyper-tune or alter parameters in a deep learning model. But an ensemble of deep learning models with average performances (i.e., not overfitting) gives extraordinary results in many cases than a carefully curated single model. Read more here. Further Reading Understanding XGBoost Algorithm In DetailDo Ensemble Methods Always Work? Ensemble Methods in SciKit-LearnRead on Wikipedia","excerpt":"Ensemble Learning is the process of gathering more than one machine learning model in a mathematical way to obtain better performance.","categories":["Deep Tech"],"tags":["bagging","Boosting","Ensemble Learning","ensemble method","gradient boosting","Guide","random forest","XGBoost"],"author_name":"Rajkumar Lakshmanamoorthy","publish_date":"2021-05-21T10:00:00","publication_year":"2021","word_count":1470,"keywords":["scikit-learn","machine learning","bagging","TPU","AI","neural network","Boosting","Ensemble Learning","RAG","random forest","gradient boosting","XGBoost","deep learning","ensemble method","anomaly detection","R","Guide"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","scikit-learn","XGBoost","RAG","anomaly detection","TPU","R"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/comprehensive-guide-to-ensemble-methods\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":34464,"title":"10 Machine Learning Projects Every Tech Aficionado Must Work On In 2019","content":"Machine learning and artificial intelligence have had a high impact on the evolving future of technology as well as human lives. Numerous innovations in the past few years were witnessed in the field of technology. The most important part of machine learning is to deal with real-world problems. In this article, we list 10 most useful machine learning projects for 2019 that will help you in tasks related to real-world problems. These projects equip ML enthusiasts with codes and datasets to learn and apply to real-life issues. (The list is in alphabetical order) 1| AutoKeras This is an open-source software library for AutoML (Automated Machine Learning) which is developed by DATA Lab at Texas A&M University. This open source project provides functions to automatically search for architecture and hyperparameters of deep learning models. To install the package, the pip installation is to be used as follows: pip install autokeras 2| AirSim This is an opensource, cross-platform project by Microsoft with the purpose to serve AirSim as a platform for AI research to experiment with deep learning, computer vision and reinforcement learning algorithms for autonomous vehicles. It supports hardware-in-loop with popular flight controllers such as PX4 for physically and visually realistic simulations. AirSim is mainly a simulator for drones, cars and more, that is built on Unreal Engine. It is developed as an Unreal plugin that can simply be dropped into any Unreal environment. It exposes APIs so you can interact with the vehicle in the simulation programmatically. 3| Detectron This is the Facebook AI Research’s software system that implements state-of-the-art object detection algorithms, including Mask R-CNN. It is written in Python and powered by the Caffe2 deep learning framework. The goal of Detectron is to provide a high-quality, high-performance codebase for object detection research and is designed to be flexible in order to support rapid implementation and evaluation of novel research. It includes the implementations of object detection algorithm such as Mask R-CNN, RetinaNet, Faster R-CNN, RPN, Fast R-CNN and R-FCN. 4| Deep-Photo-Styletransfer This open-source project is a deep learning approach to photographic style transfer that handles a large variety of image content while transferring the reference style. It has been developed by researchers at Cornell University and scores 8750 stars in GitHub in the last year. It helps in suppressing the distortions and yields satisfying photorealistic style transfers which are done by semantic segmentation and Matting Laplacian is used to constrain the transformation from the input to the output to be locally affine in colour-space. 5| Dopamine Dopamine is a research framework for fast prototyping of reinforcement learning algorithms. It aims to fill the need for a small, easily grokked codebase in which users can freely experiment with wild ideas (speculative research). The design principles are: Easy experimentation: Make it easy for new users to run benchmark experiments. Flexible development: Make it easy for new users to try out research ideas. Compact and reliable: Provide implementations for a few, battle-tested algorithms. Reproducible: Facilitate reproducibility in results. 6| Fast.ai The fast.ai library works to simplify training fast and accurate neural nets using modern best practices. The library is based on research into deep learning best practices undertaken at fast.ai and includes “out of the box” support for vision, text, tabular, and collab (collaborative filtering) models. Here is the complete documentation of fast.ai to take a deep dive into it. To install or update fastai, conda is highly recommended: conda install -c pytorch -c fastai fastai pytorch 7| FastText It is an open-source, free, lightweight library that allows users to learn text representations and text classifiers. FastText basically works on standard, generic hardware. It is a fast and accurate tool for text-based language identification that can recognize more than 170 languages. It takes less than 1MB of memory and can classify thousands of documents per second. In 2017, the Facebook AI Research (FAIR) team released pre-trained vectors in 294 languages, accompanied by two quick-start tutorials, to increase fastText’s accessibility to the large community of students, software developers, and researchers interested in machine learning. 8| Magenta Magenta is an open-source research project started by the researchers and engineers from Google Brain Team that helps in exploring the role of machine learning in the process of creating art and music. It involves developing new deep learning and reinforcement learning algorithms for generating songs, images, drawings, and other materials. Magenta is distributed as an open source Python library, powered by TensorFlow. This library includes utilities for manipulating source data (primarily music and images), using this data to train machine learning models, and finally generating new content from these models. It maintains a pip package for easy installation and Anaconda is highly recommendable. 9| Sonnet Sonnet is a library built on top of TensorFlow for building complex neural networks. The main principle of Sonnet is to first construct Python objects which represent some part of a neural network, and then separately connect these objects into the TensorFlow computation graph. To install sonnet, run: pip install dm-sonnet The recommended way to import Sonnet is to alias it to a variable named “snt”: import sonnet as snt 10| vid2vid This project is an implementation of PyTorch for high-resolution photorealistic video-to-video translation. It can be used for turning semantic label maps into photo-realistic videos, synthesising people talking from edge maps, or generating human motions from poses. Here is the link to the paper of full implementation of this project.","excerpt":"Machine learning and artificial intelligence have had a high impact on the evolving future of technology as well as human lives. Numerous innovations in the past few years were witnessed in the field of technology. The most important part of machine learning is to deal with real-world problems. In this article, we list 10 most […]","categories":["AI Trends"],"tags":["ML projects"],"author_name":"Ambika Choudhury","publish_date":"2019-02-05T11:50:58","publication_year":"2019","word_count":901,"keywords":["machine learning","artificial intelligence","AI","neural network","PyTorch","ML","computer vision","Aim","deep learning","ML projects","TensorFlow"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","neural network","computer vision","Aim","TensorFlow","PyTorch"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/10-machine-learning-projects-every-tech-aficionado-must-work-on-in-2019\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":58861,"title":"Essential Ways To Handle Data Cleaning","content":"Data cleaning isn’t the most attractive part when it comes to data science or machine learning, but it is one of the most important ones. There are no tricks nor any shortcuts for data cleaning, if one needs to have the best model possible, they need a better quality of data and a clean one. Machine learning and data scientists spend a lot of time in data cleaning because of a common belief among them that whatever data they put into the algorithm, the results solely depend upon it. Below are some tips when it comes to data cleaning: Better Data Quality It is a common notion around developers, where they chase perfecting the algorithm and making it look fancy, often ignoring one of the major factors that contribute to the success of an algorithm, the data quality. Data cleaning is a lot more important than it sounds, no matter how good one’s algorithm is or no matter how fancy it is, untidy data will give you abysmal results. Poor quality data also results in biased outcomes, which can afflict the businesses if firms fail to identify the potential flaws in it. Filtering Unnecessary Outliners Outliers can cause problems with specific models like linear regression models (reducing their robustness). But, removing an outlier just because it is big and not because it is uninformative might make your model miss out on information. Have a legitimate reason when you are thinking about removing an outlier. Removing Duplicate Observations This is one of the basic steps of data cleaning in data science. Duplicated observations frequently occur during data collection. They might occur during, combining datasets from multiple places, receiving data from other parties and scraping data. And a few are irrelevant observations, which are those ones that don’t actually fit into a specific problem, which are under consideration. These observations, if spotted correctly, will enhance one’s model. It is recommended to check for these observations before the engineering features come into play. Syntax Errors Making sure the data types are stored correctly can save a lot of time and help in creating a better model. All the values must be stored in relevant data types. There are some types of errors that need to be kept in mind: About Pad strings: Strings can be padded with spaces, and other characters to a certain width like some of the numerical codes are represented with inserting zeros to ensure they always have the same number of digits. 401 => 000401 (6 digits) Removing white spaces: Simply means removing extra white spaces at the beginning or the ending of the strings. “  hello world “ => “hello world” Fixing Structural Errors Structural errors are those that come into existence while processes like measurement, data transfer, etc. For example, one can check for typos, inconsistent capitalisation. Another way is to try to merge or include mislabelled classes into one. Standardising the Values Standardising, say, for Strings means, making sure all values are either in lower case or the upper case. Same way, the numerical values can be standardised to a certain measurement unit. For example, the length can be in meters and feet. The difference of one meter is considered the same as the difference of one foot, so one has to convert the height to one single unit. Missing Data Most algorithms do not accept missing values, so handling missing data becomes all the more crucial when it comes to algorithms and making one’s data cleaner. The missing data can be handled in two ways: Dropping observations with missing values: Dropping the missing values is not the most optimal way for the reason being that, when one drops an observation, it means dropping some information. Imputing missing values based on other observations: Imputing missing values is also not that optimal either. Imputing missing value means the value was originally missing, but when someone filled it in, which eventually leads to a loss in information no matter what imputation method one uses. Something missing can be informative as well; one can then add these missing values to the algorithm after they realise them. Imputing is like trying to fit a missing part of the puzzle back in after you have taken it out. The models built with missing values might not add any real information and keep reinforcing the patterns already provided by other features. A possible solution? Just tell the algorithm that something is missing. Handle missing categorical data: for missing categorical feature data, one can label them as ‘Missing’. It’s like adding a new class for the feature. Handling missing numerical data: To process numerical data, one should flag and fill the values. First, flag the observation with an indicator variable of the missingness, then fill the original value with 0 to meet the technical requirement of no missing values. Flagging and filling essentially allow the algorithm to estimate the optimal constant for missingness instead of filling it.","excerpt":"Data cleaning isn’t the most attractive part when it comes to data science or machine learning, but it is one of the most important ones. There are no tricks nor any shortcuts for data cleaning, if one needs to have the best model possible, they need a better quality of data and a clean one. […]","categories":["AI Features"],"tags":["Data Cleaning","data cleansing"],"author_name":"Sameer Balaganur","publish_date":"2020-03-17T16:00:00","publication_year":"2020","word_count":824,"keywords":["data science","Go","API","machine learning","AWS","AI","cloud_platforms:AWS","Git","data cleansing","Data Cleaning","data quality","R"],"extracted_tech_keywords":["AI","machine learning","data science","AWS","R","Go","Git","API","data quality","cloud_platforms:AWS"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/essential-ways-to-handle-data-cleaning\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10042452,"title":"All You Need To Know About Google’s Visual Inspection AI","content":"In 2019, Google Cloud identified six sectors as vital components of its growth: public, healthcare, financial services, retail, media, and manufacturing. Within manufacturing, the cost of quality control and inspection continues to be among the highest. The American Society for Quality estimates that the price of quality may be as high as 15 to 20 percent of annual sales revenues for many organisations. For larger manufacturers, this translates into billions of dollars every year. Additionally, the rapid increase in production volumes makes it difficult for humans to manually inspect defects in computer chips and other products. To combat this, Google Cloud has recently announced an approach, backed by artificial intelligence (AI), for visual inspection. The newly launched Visual Inspection AI is a purpose-built tool to help manufacturers and related workers and businesses to inspect and reduce product defects and decrease quality control costs. Powered by Google Cloud Platform’s computer vision technology, Visual Inspection AI goes beyond the traditional methods of supporting manufacturing quality control through its general-purpose AI product, AutoML. According to Kevin Prouty, Group Vice President of Energy and Manufacturing at IDC, “Google Cloud’s approach to visual inspection is the roadmap most manufacturing companies are looking for.” Visual Inspection AI aims to automate quality assurance workflows, thus allowing companies to identify and correct defects before shipping products. Through this, the new AI tool automates visual inspection using a set of AI and computer vision to improve production by increasing yields, reducing re-work, and cutting back on return-and-repair costs. Previous methods COVID-19 has increasingly driven manufacturers to adopt AI into their production processes. According to a Google Cloud survey, 76 percent of executives say they have embraced digital enablers such as AI, data analytics and cloud computing. Additionally, 66 percent of manufacturers who use AI in their daily operations have stated that their reliance on the technology is increasing. With this advancement, traditional methods to quality control inspections fall short. Traditionally, manufacturers include one or more steps to inspect products for defects visually. The visual inspection process is typically highly manual, making it vulnerable to human error and highly time-consuming. Moreover, traditional machinery used in machines are not flexible enough to adapt to product changes and can only detect a handful of defects at any time. Source: Google Cloud Artificial intelligence then is an agent that manufacturers are hopeful will bring in a more significant wave of innovation. Google Cloud listed multiple benefits of utilising AI, ranging from the reduced cognitive load for operators, fewer missed defects, no programming required (making it more flexible than previous machines), and the ability to detect hundreds of areas of interest on a product in seconds. Google’s new solution As per Kyocera Communications Systems, a major manufacturer of mobile phones for wireless service providers, Visual Inspection AI is an innovative service that non-AI engineers can use. Google Cloud says that its new Visual Inspection AI meets the needs of quality, testing, manufacturing, and process engineers who might not be well-versed in AI despite being experts in their respective fields. Thus, the new tool paves the way to many substantial benefits compared to general-purpose machine learning (ML) models, such as superior computer vision technology, shorter time-to-value and high scalability. Through this, customers can deploy solutions within weeks, and an interactive user interface guides them through the steps. Visual Inspection AI has also improved accuracy by up to 10 times from general ML approaches. Finally, Visual Inspection AI deep goes beyond simple anomaly detection. Instead, it allows customers to train models that detect, classify and locate multiple defect types in a single image—doing so provides follow-up tasks on production lines to be automated. Source: Google Cloud There are multitudes of ways in which businesses can use Google Cloud’s Visual Inspection AI in manufacturing. Automotive manufacturing, for one, can use it for paint shop surface inspection or press shop inspection—to look for scratches, dents, cracks or staining. On the other hand, electronics manufacturing could employ the tool for defects in printed circuit board components, and general-purpose manufacturing could improve upon procedures like packaging and label inspection, fabric inspection, metal welding seam inspections—to name a few. As per the above mentioned Google Cloud survey on manufacturing trends, the most common roadblock to AI integration is the lack of talent to leverage AI properly. Given this, Google Cloud’s new Visual Inspection AI appears as a brilliant step towards the proper deployment of artificial intelligence in the manufacturing industry.","excerpt":"In 2019, Google Cloud identified six sectors as vital components of its growth: public, healthcare, financial services, retail, media, and manufacturing. Within manufacturing, the cost of quality control and inspection continues to be among the highest. The American Society for Quality estimates that the price of quality may be as high as 15 to 20 […]","categories":["Global Tech"],"tags":[],"author_name":"Mita Chaturvedi","publish_date":"2021-06-28T11:00:00","publication_year":"2021","word_count":739,"keywords":["machine learning","artificial intelligence","AI","cloud computing","ML","computer vision","RAG","Aim","anomaly detection","analytics"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","computer vision","analytics","Aim","RAG","anomaly detection","cloud computing"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/all-you-need-to-know-about-googles-visual-inspection-ai\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":37308,"title":"Indian Data Localisation Norms Have Pushed WhatsApp Pay Into A Limbo","content":"India has been at the forefront of a digital revolution ever since the launch of Reliance Jio. This telecom provider caused a tariff war that has reduced India’s mobile Internet prices to some of the lowest in the world. This, coupled with accessible Internet feature phones and cheap smartphones, has driven the increased adoption of the Internet in the subcontinent. A natural companion to the growth of Internet connectivity in India has been WhatsApp, which Indians have used to stay connected. The company reports about 200 million users in India, but the penetration of the app in the country is much more. WhatsApp as a company has also run into multiple problems with regulators in India as the app was being used to spread misinformation heavily. This led to them creating features such as a “forwarded” label and restricting forwards to 5 people or groups at a time. Keeping this in mind, WhatsApp has treated India as one of its most important markets, testing new features before releasing them to the wild. The forward-thinking nature of WhatsApp to service the Indian population have not stopped there. They are also taking solid efforts to set up a payment system in the subcontinent. WhatsApp’s History With India The app has long been used to spread misinformation among the Indian population. Early WhatsApp users will remember the rumour of “Infected Mazaa” that was spread on the platform at around 2015. However, with the rapid rise of the Internet, over 72% of Indians have difficulties distinguishing real news from fake news. This has led to the rise of viral forwarding among the population. This culminated in 50 casualties over 2017 and ’18 due to WhatsApp forwards of child abduction mobs. This caused the company to take notice of the situation at the same time the government caught a hold of the situation. After a PR campaign and implementation of features like restricting the number of forwards and adding a ‘forwarded’ label to messages that had been sent directly from other groups. The damage seemed to be done, as the Indian government requested that the application offers traceability for messages. This could not be done without breaking the end-to-end encryption of the application, which was the reason WhatsApp failed to comply. However, the application later employed a variety of algorithms to curb the spreading of viral messages. This left a black mark on WhatsApp and a bitter taste in the mouth of regulators, as the app was directly responsible for the killings in their eyes. Even though there was no concrete action taken against the application by the government, they have since given the company a difficult time in the subcontinent. Indian Data Regulations Stifle Innovation The subcontinent has adopted stringent rules for the localization of data for companies that function within its borders. All user data is required to be stored in the same country, which the Indian government says is to promote the privacy of citizens data. However, this has been a major pain point for MNCs such as Facebook while rolling out features to do with WhatsApp. Financial data, specifically, has to be stored without exception on Indian soil. This was seen when the localisation needs for adopting UPI were expressed, leading to all data being stored in the country. The Reserve Bank of India, which regulates the banking environment in the company, had also made it clear that the data regarding the payments would be stored only in India, with no exceptions. Many payment companies attempted to move the court over this, which was promptly denied by the RBI. Moreover, companies such as PayTM and PhonePe were also for the localization of data, citing the need for it. This was the problem faced by WhatsApp, who wished to move into the FinTech space in India. WhatsApp’s Move Into FinTech Blocked The app was looking towards launching an UPI based payment feature, so as to capitalise on its huge consumer base. Its competitor, Google, had done so effectively with the launch of Google Tez, later rebranded to Google Pay. This feature, called WhatsApp Pay, would allow users to send money between their contacts using the UPI system. However, the RBI revealed that this would not be the case for WhatsApp. The application was found to be in violation of the norms described by the government, leading to them not approving their request for implementing WhatsApp Pay. This left the application in a resting state for now, as there is no regulatory approval to move forward. This was due to the fact that the data was still being stored overseas, and that the information being stored in the country was simply a copy. This was in strict violation of the norms, even as WhatsApp was compliant with the other norms. Even as WhatsApp scurries to gain approval and capitalize on the market, the bad impression left on regulators seems to have stayed until now. This has created an environment wherein the app is neither here, nor there, and continues to operate within the purview of its current features.","excerpt":"India has been at the forefront of a digital revolution ever since the launch of Reliance Jio. This telecom provider caused a tariff war that has reduced India’s mobile Internet prices to some of the lowest in the world. This, coupled with accessible Internet feature phones and cheap smartphones, has driven the increased adoption of […]","categories":["AI Features"],"tags":["UPI"],"author_name":"Anirudh VK","publish_date":"2019-04-04T11:23:41","publication_year":"2019","word_count":845,"keywords":["Go","API","programming_languages:R","AI","innovation","programming_languages:Go","Git","UPI","ViT","R"],"extracted_tech_keywords":["AI","R","Go","Git","API","ViT","innovation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/indian-data-localisation-norms-have-pushed-whatsapp-pay-into-a-limbo\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":43061,"title":"Is China&#8217;s Expertise In Artificial Intelligence Over-Hyped?","content":"China has been often touted as the fastest emerging hub for AI development, even surpassing the superpowers such as the USA in the emerging tech. Chinese companies and government are taking the analytics and AI play quite seriously, bringing newer and favourable policies around its adoption. Numbers suggest that in 2018, 60 per cent of total global AI investments poured into China with investments from VCs, private equity and the Chinese government. Not just the companies but educational institutes are taking AI seriously as many schools are teaching AI courses to make its citizens AI-ready. There is no doubt that China has been serious about its AI strategy, but is its power and supremacy in artificial intelligence real or exaggerated? Is China’s AI Power Exaggerated? There are many experts who believe that China’s AI play is exaggerated to a great extent. Jeff Ding, a researcher at the University of Oxford who studies China’s AI development, recently shared that many people have a hyped-up view of China’s AI capabilities. He strongly believes that it can be largely attributed to the media and government who wants to put the country in a good light. For many of the claims that China makes, there are no substantial proves to make a point. While numbers suggest otherwise, there is a lack of genuine understanding of the technical skills and capacity of Chinese companies. Many believe that most of China’s AI giants are less impressive than they seem, barring a few such as Baidu and Huawei. Some of the areas that need re-tinkering for China to claim its supremacy in AI are: Research papers: There are reports which suggest that China leads the US in the number of AI-related patents and research papers. However, when it comes to citation of these papers by other researchers, they hardly cite Chinese research papers and prefer papers released by US researchers. This puts up a question of the quality of papers generated by China. While there may be quantity, quality still needs some reassurance. AI workforce: While China’s large population may provide a larger percentage of AI specialists, quantifying the size of that workforce is difficult as there is no standard in doing so. Reports suggest that many Chinese AI expertise may only have an associate or technical certificates, unlike more advanced engineering degrees in other countries. Technologies like facial recognition are questionable: In technologies like facial recognition, China may be far behind the US in the compilation of database and execution of technology.  While China has trained its facial recognition models by training over one billion facial images, FBI for instance, has its own database of over 640 million faces, according to a Government Accountability Office report. They also lag in other technologies such as smart glasses. Data Privacy and ethics: Another point of concern is China’s data privacy policies, as there is no clear regulations or guidelines overlooking it. Ethics around AI are very much a concern in China just as it is in other countries. Accessible data and computing speed: AI strength and dominance largely depends on the amount of accessible data. While China has a larger population, the amount of data collected and accessible by other countries such as the US is quite high. Computing power also plays a large role, in which China is playing up quite efficiently compared to other factors. In a nutshell, Some of the advantages that China has over countries are: Extensive support from the government In terms of hardware, it now posses more than half of the world’s computers. Disadvantages: It lags in AI talent. Google alone employs a large chunk of top AI talent It lacks in terms of global networking. For instance, Google, Facebook, and Amazon possess far more total active users than Chinese counterparts Data privacy laws Outlook While China’s AI advances are witnessing a remarkable high, it cannot be claimed to be the superpower in AI yet. It may be doing exceedingly well but other countries such as the US have a huge advantage of talent and hardware when it comes to fields of analytics and artificial intelligence. China is, however, taking steps to grow at a faster rate. For instance, China’s Microsoft Research Asia alone has trained over 5,000 AI professionals, which shows that it is taking the training of AI talent quite seriously. There is no doubt that China is growing at a faster rate and it might not be far away when China becomes an economic power, but there is still a long way to go.","excerpt":"China has been often touted as the fastest emerging hub for AI development, even surpassing the superpowers such as the USA in the emerging tech. Chinese companies and government are taking the analytics and AI play quite seriously, bringing newer and favourable policies around its adoption.  Numbers suggest that in 2018, 60 per cent of […]","categories":["AI Features"],"tags":[],"author_name":"Srishti Deoras","publish_date":"2019-07-21T08:23:21","publication_year":"2019","word_count":755,"keywords":["Go","artificial intelligence","AWS","AI","cloud_platforms:AWS","RPA","programming_languages:R","Aim","analytics","R"],"extracted_tech_keywords":["AI","artificial intelligence","analytics","Aim","AWS","R","Go","RPA","cloud_platforms:AWS","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-chinas-expertise-in-artificial-intelligence-over-hyped\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":10005722,"title":"Detecting Anomalies in Wafer Manufacturing: Weekend Hackathon #18","content":"Weekend Hackathons are becoming more competitive, so we are back with a tougher one this time. Another exciting weekend hackathon to flex your machine learning classification skills by building an anomaly detection model to separate the good and anomalous products for one of India’s leading wafer manufacturers into 2 different classes. Detecting Anomalies can be a difficult task and especially in the case of labeled datasets due to some level of human bias introduced while labeling the final product as anomalous or good. These giant manufacturing systems need to be monitored every 10 milliseconds to capture their behavior which brings in lots of information and what we call the Industrial IoT (IIOT). Also, hardly a manufacturer wants to create an anomalous product. Hence, the anomalies are like a needle in a haystack which renders the dataset that is significantly Imbalanced and has a very less number of rows. The challenge will start on 28th Aug Friday at 6 pm IST. Click here to participate Problem Statement & Description Capturing such a dataset using a machine learning model and making the model generalize can be fun. In this competition, we bring such a use-case from one of India’s leading manufacturers of wafers(semiconductors). The dataset collected was anonymized to hide the feature names, also there are 1558 features that would require some serious domain knowledge to understand them. However, In the era of Deep Learning, we are challenging the data science community to come up with an anomaly detection model that can generalize well on the unseen set of data(Test data). In this hackathon, you will be creating a machine learning\/ deep learning model to classify the anomalies correctly using Area under the curve(AUC) as a metric. This dataset also provides huge scope to feature engineering\/dimensionality reduction and we are looking forward to some serious competition this time. Given are 1558 distinguishing factors that can predict the right class of a product. Your objective as a data scientist is to build a machine learning model that can accurately classify the class of good products as well as anomalous products as close as possible. Dataset Description: The unzipped folder will have the following files. Train.csv – 1763 rows x 1559 columns (includes Class as target column)Test.csv – 756 rows x 1558 columnsSample Submission.csv – sample format for submission file. Attribute Description: Feature_1 – Feature_1558 – Represents the various attributes that were collected from the manufacturing machineClass – (0 or 1) – Represents Good\/Anaomalous class labels for the products Skills: High Dimensionality Data, Overfitting-vs-UnderfittingAdvanced Classification Techniques, Gradient Boosting, Neural Nets, etcFeature engineering, Feature Selection TechniquesOptimizing Area under the curve(AUC) to generalize well on unseen data The datasets will be made available for download on Aug 28th, Friday at 6 pm IST. This hackathon and the bounty will expire on Aug 31st, Monday at 7 am IST. Click here to participate Bounties We have introduced a new set of prizes going forward. Continous 3 finishes In Weekend Hackathons Top-3 participants on the private leaderboard will be interviewed for #HackeroftheMonth.Stand a Chance to get an exclusive interview for your Data Science\/Machine Learning journey by Analytics India Magazine Who is the #hackerofthemonth ?? Any participant can become #hackerofthemonth by proving their mettle in the weekend hackathon leaderboards. We will award the #hackerofthemonth community recognition to participants who are in Top-3 for 3-consecutive weekend hackathons in a row. Yes, you got it right, it’s a hattrick!! Stand a chance to get Interviewed by the biggest AL\/ML media-house in the country for your Data Science and Machine Learning journey. Please note this PRIZE is only for the Weekend Hackathon series of competitions. Click here to participate Rules One account per participant. Submissions from multiple accounts will lead to disqualificationThe submission limit for the hackathon is 10 per day after which the submission will not be evaluatedAll registered participants are eligible to compete in the hackathonThis competition counts towards your overall ranking pointsWe ask that you respect the spirit of the competition and do not cheatThis hackathon will expire on 03rd August, Monday at 7 am ISTUse of any external dataset is prohibited and doing so will lead to disqualification Evaluation The submissions will be evaluated using the ROC-AUC score (Reciever Operating Characteristics – Area Under the Curve) metric. One can use roc_auc_score(actual, predicted)This hackathon supports private and public leaderboardsThe public leaderboard is evaluated on 30% of Test dataThe private leaderboard will be made available at the end of the hackathon which will be evaluated on 100% Test data Click here to participate","excerpt":"Weekend Hackathons are becoming more competitive, so we are back with a tougher one this time. Another exciting weekend hackathon to flex your machine learning classification skills by building an anomaly detection model to separate the good and anomalous products for one of India’s leading wafer manufacturers into 2 different classes.  Detecting Anomalies can be […]","categories":["Deep Tech"],"tags":["Hackathons","Hackathons India","Machinehack Hackathon","Weekend Hackathon"],"author_name":"Anurag Upadhyaya","publish_date":"2020-08-28T09:54:29","publication_year":"2020","word_count":753,"keywords":["data science","Go","machine learning","Weekend Hackathon","AI","R","ML","Hackathons","feature engineering","deep learning","anomaly detection","analytics","Machinehack Hackathon","Hackathons India"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","anomaly detection","R","Go","feature engineering"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/detecting-anomalies-in-wafer-manufacturing-weekend-hackathon-18\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":59526,"title":"Qure.ai Develops AI-Powered Solutions To Flatten The Curve On Covid-19","content":"As more and more AI companies reorient their focus to develop technology to fight Covid-19, Mumbai-based AI startup Qure.ai is on its way to launching two critical offerings. The team has been directing its R&D efforts and using past experience in the domain of AI and healthcare to help medical professionals tackle Covid-19. Operating in the domain of radiology, Qure.ai has been refining medical imaging accuracy with the assistance of machine-supported tools trained using millions of images. Aimed at medical imaging professionals, this AI solution developer taps deep learning technologies to provide automated interpretation of radiology exams, including X-rays, CTs and MRI scans. Qure.ai’s solutions for Covid-19 include: A tool that can interpret an X-ray to detect Covid-19, as well as quantify the proportion of lungs affected. This can be used to monitor the progression of patients, as well as to screen patients who need to undergo further testing.An app-based solution which can be used for contact tracing of patients. It can also enable radiologists to effectively triage medical cases. Thus, by enabling faster diagnosis and cost-effective solutions, it can greatly help healthcare providers who are working in the frontline amid the coronavirus pandemic. Targeting not just medical professionals, but also high-risk populations – especially the elderly and those with pre-existing conditions – Qure.ai is available to chat with users 24\/7 across time zones to offer help in this critical time. By scheduling an appointment, you can also share solutions and explore a partnership opportunity. In fact, the startup recently announced its strategic partnership with UK market leader Medica Group to develop AI tools to improve efficiency of radiology scan workload. Tuberculosis to Covid-19 Qure.ai’s experience with tackling tuberculosis (TB) in more than 10 countries has greatly helped them develop this solution. This tool was originally designed to interpret chest X-rays as a way to screen for tuberculosis and other lung diseases. What sets this solution apart is not just its affordability, but also because it significantly reduces the time taken to diagnosis. Furthermore, in a big boost to the potential of AI in healthcare, Qure.ai recently secured $16 million from Sequoia Capital and MassMutual Ventures to further invest into R&D.","excerpt":"As more and more AI companies reorient their focus to develop technology to fight Covid-19, Mumbai-based AI startup Qure.ai is on its way to launching two critical offerings. The team has been directing its R&D efforts and using past experience in the domain of AI and healthcare to help medical professionals tackle Covid-19. Operating in […]","categories":["Global Tech"],"tags":["Coronavirus","covid-19","qure.ai"],"author_name":"Anu Thomas","publish_date":"2020-03-23T16:20:02","publication_year":"2020","word_count":361,"keywords":["Go","API","covid-19","AI","programming_languages:R","programming_languages:Go","Coronavirus","Ray","Aim","deep learning","qure.ai","R","startup"],"extracted_tech_keywords":["AI","deep learning","Aim","Ray","R","Go","API","startup","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/qure-ai-develops-ai-powered-solutions-to-flatten-the-curve-on-covid-19\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10135695,"title":"Rahul Dravid Gets a New Job at Salesforce","content":"Rahul Dravid, the all-round Indian cricketer, has joined Salesforce as the first ever-brand ambassador to help Indian businesses grow using Salesforce CRM platform. This is part of the Growing With India campaign, which is Salesforce’s easy to use and set up all in one CRM platform for taking businesses of India to the next level. Customers like Royal Enfield, clear, Snapdeal, Listenlights, and several others use this platform for boosting customer experience. Salesforce has been at the forefront of building businesses globally, as well as in India. Recently, the company released its Agentforce Partner Network with AWS, Google, IBM, and Workday. The company also recently collaborated with NVIDIA for launching advanced autonomous AI agents for enterprise. These collaborations are set to optimise predictive and generative AI workflows, enabling enhanced insights and productivity for sales, service, marketing and IT teams, utilising Salesforce’s CRM platform. “Agentforce represents a monumental shift in the industry, enabling humans with AI to redefine how businesses achieve customer success,” said Brian Millham, president and chief operating officer at Salesforce. In other news, Fintech giant Klarna CEO Sebastian Siemiatkowski announced that the company will end its service provider relationships with Salesforce and Workday as part of a major internal overhaul driven by AI initiatives. This decision aligns with Klarna’s broader strategy to streamline its technology infrastructure by leveraging AI to standardise and simplify its processes.","excerpt":"This is part of the Growing With India campaign, which is Salesforce’s easy to use and set up all in one CRM platform for taking businesses of India to the next level.","categories":["AI News"],"tags":["Salesforce"],"author_name":"Mohit Pandey","publish_date":"2024-09-18T16:09:42","publication_year":"2024","word_count":227,"keywords":["Go","AWS","AI","ML","autonomous AI","RAG","AI agents","ViT","generative AI","Salesforce","R"],"extracted_tech_keywords":["AI","ML","generative AI","RAG","AWS","R","Go","ViT","autonomous AI","AI agents"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/rahul-dravid-gets-a-new-job-at-salesforce\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":41085,"title":"Is Polymorphic Phishing The Next Big Threat In Cyber Attacks?","content":"In the data-driven business world everyone in an organisation — from employees to CEO, use email on a day to day basis. The use of email has reached such a level that it has now even become a top cyberattack vector, and cybercriminals are persistently targeting high-value individuals through phishing emails who have privileged access or handle sensitive data within an organization. Today, people might say that phishing is old school and is not effective. But they are not aware that hackers are also leveraging the advanced technologies and using sophisticated methods to pwn. And one of those methods is Polymorphic Phishing Attacks. The Rise Of Polymorphic Phishing According to a source, IRONSCALES, an Israel-based cybersecurity firm was doing research on an automatic anti-phishing platform. And to the surprise, the firm has discovered that 42% of the phishing attempts they examined were “polymorphic” in nature. They identified 11,733 email phishing attacks that underwent at least one permutation over 12 months. And with 52,825 permutations, these attacks successfully made their way into 209,807 inboxes around the world. Polymorphic email phishing is a phishing email sent to multiple users where an attacker implements slight but significant and often random changes to an emails’ artefacts — at least one of the following is being changed either randomly or manually\/intentionally depending on the attack: Sender name, Sender address, Subject Greeting, Email body or signature. This strategic approach of hacking has been seen in the wild since at least 2016. At the very beginning, this attack was all about generating “polymorphic URLs”, thousands of different URLs that would lead to phishing or malware delivery pages. However, gone are those days — today, even this attack has even extremely sophisticated, enabling attackers to quickly develop phishing attacks that can easily trick\/manipulate and bypass most of the anti-phishing tools. Example Of Polymorphic Phishing Suppose you get an email and there is an attached that appears to be a pdf file. And when you fail to identify the email as phishing and open the file, it prompts an update message claiming the tool is not updated. And when you click the update link, it redirects you to a cloned web page that asks you to enter your credentials. Furthermore, the email spreads like a virus across the company. But the catch here is that every employee would receive the email with a slight yet significant change, which wouldn’t let the employees detect whether its that same email. The Role Of Dark Web Behind this sophisticated phishing email attack, the darknet is playing a major role. There are a huge number of tools available on the dark web that help hackers carry out polymorphic phishing attacks on big companies. According to a source, Over a five-month period, Dr Mike McGuire, Senior Lecturer in Criminology at the University of Surrey and his team analyzed over 70,000 Darknet websites in order to find out what type of tools and services are there. What they find out through their research is quite surprising as the target companies are big enough. Around 40% of dark web offerings are targeted hacking services that specifically designed to exploit Fortune 500 and FTSE 100 companies. Even though there were successful takedowns of dark web websites in the past, there are still numerous websites hosted on the dark web that are offering some of the most notorious services. And phishing tools are one among all those services that are gaining significant traction. Bottom Line Over the past couple of years, the world has witnessed both — technologies to mitigate cyber threats and technologies to power up cyber threats. Meaning, advancement in technology is not only for the better side, but even the wrongdoers are also making the best use of technology to advance their attacking strategy. Phishing is a great example — a type of attack that was once termed as not effective and easily detectable, today it has become a major threat. Targeted phishing attacks are increasingly bypassing gateway security controls and are landing right into employees’ mailboxes in every organisation around the globe. And the best way to cope with this is to have a strong and reliable security infrastructure. Also, sessions that train employees on different cyber-attacks.","excerpt":"In the data-driven business world everyone in an organisation — from employees to CEO, use email on a day to day basis. The use of email has reached such a level that it has now even become a top cyberattack vector, and cybercriminals are persistently targeting high-value individuals through phishing emails who have privileged access […]","categories":["AI Features"],"tags":["Cyber Attack","hacking","Phishing"],"author_name":"Harshajit Sarmah","publish_date":"2019-06-21T05:28:57","publication_year":"2019","word_count":703,"keywords":["Go","hacking","programming_languages:R","AI","data-driven","Phishing","ML","programming_languages:Go","RAG","Cyber Attack","Aim","GAN","R"],"extracted_tech_keywords":["AI","ML","Aim","RAG","R","Go","GAN","data-driven","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/is-polymorphic-phishing-the-next-big-threat-in-cyber-attacks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10057118,"title":"All You Need to Know About Unsupervised Reinforcement Learning","content":"Unsupervised learning can be considered as the approach to learning from the huge amount of unannotated data and reinforcement learning can be considered as the approach to learning from the very low amount of data. A combination of these learning methods can be considered as unsupervised reinforcement learning which is basically a betterment in reinforcement learning. In this article, we are going to discuss unsupervised Reinforcement learning in detail along with special features and application areas. The major points that we will discuss here are listed below. Table of Contents About Unsupervised Reinforcement LearningLeveraging Unsupervised Learning for Reinforcement LearningUnsupervised Learning to Speed-Up Reinforcement LearningExamples of Unsupervised Reinforcement Learning Why Unsupervised Reinforcement Learning? Unsupervised reinforcement learning is a combination of unsupervised learning and reinforcement learning. We can divide machine learning approaches into three subfields – supervised learning, unsupervised learning, and reinforcement learning. The basic definition of these subfields are as follows: Unsupervised Learning: It is a process of learning from a huge amount of unannotated data.Supervised Learning: It is a process of learning from a medium amount of data with annotated values. Reinforcement learning: It is a process of learning from reward signals. These rewards can be given by either the environment or humans in the form of a small amount of data. The below image depicts a comparison of these subfields in the context of data size, using a picture of cake where the bulk of the cake can be considered as unsupervised learning, the icing on the cake(medium on size) can be considered as supervised learning and the cherry on the top of the cake can be considered reinforcement learning. As in the last section, we have discussed the basics of these three forms of learning, we can also divide these forms of learning on the basis of active and passive learning as: With Teacher Without TeacherActive Reinforcement learningIntrinsic reward optimizationpassiveSupervised learning Unsupervised learning From the above table, we can say that, If the learning is active with the teacher, it can be considered as reinforcement learning that is based on extrinsic reward optimization.If the learning is active without a teacher, we can call it learning based on intrinsic reward optimization.If the learning is passive with a teacher,  It can be considered as supervised learning.If the learning is passive without a teacher,  It can be considered as unsupervised learning. When we talk about the basic process followed by unsupervised learning, we define objective functions on it such that the process can be capable of performing categorization of the unannotated data or unlabeled data. There are various problems that can be dealt with using unsupervised learning. Some of them are as follows: Label creation, annotation, and maintenance is a changing discipline that also requires a lot of time and effort.Many domains require expertise in annotation like law, medicine, ethics, etc. In reinforcement learning, reward annotation is also confusing. There are many questions like which kind of reward can be used, a continuous-valued reward or categorical valued reward,  sparse rewards, or dense rewards. Collecting human behavioural data for behaviour cloning is also challenging because there is no annotated information for such data. The above-given problems can be solved using unsupervised learning and here we find the answer of why there is always a point where unsupervised learning and reinforcement learning will cross each other. Most of the time, reinforcement learning is required when we try to perform some human tasks using robots. Unsupervised learning is inspired by how human infants learn. So we can understand now why we require unsupervised learning in reinforcement learning. There will always be a possibility that we will be able to make a robot learn skills, not tasks that can be done using unsupervised learning and reinforcement learning together. Leveraging Unsupervised Learning for Reinforcement Learning In the above section, we have seen what and why unsupervised learning. When we talk about the divisions of unsupervised learning, it can be performed in the following way. We can see that we have two basic divisions of unsupervised learning, generative and non-generative. In the context of the generative models, these models are used for learning about the probability distribution from a large amount of data of how the world behaves if we perform certain actions in the environment. These models can be used to generate fake data or for planning according to the behaviour. Using them for reinforcement learning can be considered as the process of performing reinforcement learning when we know about the reactions which can help in earning rewards for agents. On the other hand, when performing reinforcement learning, we can use non-generative unsupervised learning as external learning to speed up the task performed by reinforcement learning. The following chart is the sum-up of how to use UL in RL. Here in the above representation, we have seen that in two ways we can best use UL for RL. Using the generative models of unsupervised learning in reinforcement learning is a separate area of study. In this article, we are just generating an overview about “why it is important to use non-generative UL for RL”. Let’s go through some points to know “Why to use UL as an auxiliary task to speed up RL?”. Why use UL as an Auxiliary Task to Speed-up RL? The below-given points are the answer to the above question. It is the simplest way of combining UL and RL.We can maximize rewards and learn about the environment using the same shared network.Representations learned using the UL are very helpful for the RL tasks. The success of divisions from unsupervised learning for label efficient supervised learning.Challenge: What is the right UL objective that can work well with RL? How to capture the useful aspects from very high dimensional data? In a generalization of these points, we can say using unsupervised learning, we can make Label creation, annotation, and maintenance processes well-performing for a specific kind of data from the whole data. This helps RL to perform at high speed by making RL concentrate on that specific kind of data. Examples of Unsupervised Reinforcement Learning In this section of the article, we are going to discuss examples where different forms of unsupervised learning have been applied to the RL. Let’s take a look at the forms of unsupervised learning with their belonging divisions. Generative Unsupervised Learning Autoencoder Variational Autoencoder Non-generative Unsupervised Learning Contrastive learningSiamese networksData-augmentations Some of the work related to the generative unsupervised learning for reinforcement learning are as follows: UREAL Architecture from DeepMind: Here we can see reinforcement learning agents have achieved high-level performance in various games using the variational autoencoder.DARLA (DisentAngled Representation Learning Agent): DARLA significantly outperforms conventional baselines in zero-shot domain adaptation scenarios where interest data is hard to obtain, for which they have basically used autoencoder which extracts required data from a high dimensional unannotated data.World Models: These models can be trained quickly in an unsupervised manner using the variational autoencoder to learn a compressed spatial and temporal representation of the environment. Let’s see some of the works from the non-generative UL in RL. CURL (Contrastive Unsupervised Representations for Reinforcement Learning): This work is capable of extracting high-level features from the raw information using contrastive learning and performing reinforcement learning on extracted features.DADS (Unsupervised Reinforcement Learning for Skill Discovery): DADS designs an intrinsic reward function that encourages the discovery of “predictable” and “diverse” skills. The intrinsic reward function is high if the changes in the environment are different for different skills (encouraging diversity) and changes in the environment for a given skill are predictable.Reinforcement Learning with Augmented Data: This is a simple plug-and-play module that can enhance most RL algorithms by performing an extensive study of general data augmentations for RL on both pixel-based and state-based inputs, and introducing two new data augmentations. Final Words In this article, we have seen how by combining unsupervised learning and reinforcement learning we can make it unsupervised reinforcement learning. By the points given in the article, we know why unsupervised learning is a player in the process and we have seen how to best use UL for RL. Along with this, we have discussed some of the examples based on unsupervised reinforcement learning.","excerpt":"Unsupervised learning and reinforcement learning are two major type of learning methods. A combination of these learning methods can provide a betterment in reinforcement learning","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Data Science","Data Scientist","Deep Learning","Machine Learning","Python"],"author_name":"Yugesh Verma","publish_date":"2021-12-26T12:00:00","publication_year":"2021","word_count":1361,"keywords":["Go","data augmentation","machine learning","programming_languages:R","AI","Machine Learning","programming_languages:Go","RAG","Python","Deep Learning","Data Science","Data Scientist","R","contrastive learning","AI (Artificial Intelligence)"],"extracted_tech_keywords":["AI","machine learning","RAG","R","Go","data augmentation","contrastive learning","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/all-you-need-to-know-about-unsupervised-reinforcement-learning\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10070946,"title":"TCS FY23 Q1 results – touches 6 lakh employees mark, attrition at all time high, and more","content":"Last week, Tata Consultancy Services (TCS) announced its consolidated financial results for the quarter ending June 30, 2022 (FY-23, Q1). At the earnings call, the company reported a continuing growth momentum, where it has been steadily seeing progress across both markets and customers. TCS also revealed that the demand for technology is very robust, primarily driven by cloud transition, investments in consumer experience and operating model transformation initiatives focused on growth and transmission and customer optimisation aspects. The company says that it has a steady demand, reflected in its pipeline and deal closures. It has complete contract signings of $8.2 billion this quarter, with a couple of deals in the range of about 400+ million. In terms of employee headcount, TCS stands at 606,331, i.e., up 97,270 employees YoY basis, or a growth of 19.7 per cent for this quarter. It aims to continue adding more people to build up the capacity for its long-term growth while also balancing its demand for talent. The same is reflected in their attrition numbers as well. Revenue-wise, it has generated a growth of $6.8 billion, growing 16.2 per cent in rupee terms, 15.5 per cent in constant currency terms and 10.2 per cent year-on-year (YoY) in dollar terms. From a vertical market perspective, TCS recorded continuous growth momentum in its BFSI with 13.9 per cent YoY growth and clocking in at $2.18 billion of quarterly revenue. On the other hand, retail and CPG, which pretty much bore the brunt of the pandemic, have been recovering very well, according to TCS—who further stated that recovery continues with a 25 per cent YoY growth, clocking $1.08 billion in revenue. Manufacturing, which has been slightly volatile, continues to grow at 16.4 per cent on constant currency terms at $670 million for the quarter. India business stalls Assessing segment performance, North America, one of the largest markets for TCS, grew at 19.1 per cent YoY. “Continuing the very strong growth that we have been experiencing over the last many quarters, we are now at $3.6 billion on a quarterly basis in North America”, shares Rajesh Gopinathan, CEO and Managing Director at TCS. Further, he added that the UK and Europe grew slightly above 12 per cent YoY, each contributing moderately above a billion. However, TCS’ India business seems on a decline since the last financial quarter. Gopinathan reveals that India continues to be very volatile and further explains that the public sector and small and medium businesses seem unstable in the Indian context compared to the corporate sector that is doing relatively well. “We are also selective about what projects we are taking given the overall volatility in the market,” comments Gopinathan. Technology patents filed At the earnings call, TCS reported that it would continue investing in IP development. The company has filed a complete patent application of about 6,750 patents, i.e., 760 patents YoY. In terms of actual patents granted, TCS has 2,400 patents, up to 429 patents YoY basis. “This continues to be a key focus for us, not just from our production perspective, but also reflecting the kind of intellectual strength that the company has, across its solutions and various other products and research areas”, says Rajesh Gopinathan. TCS’s IP portfolio includes a comprehensive set of platforms and products that addresses enterprises’ business and technology needs across industries. This includes the TCS BaNCS™ suite for financial institutions; Algo Retail™ suite consisting of TCS Optumera™ and TCS Omnistore™; ignio™ cognitive automation suite; TCS ADD platform for the life sciences industry; TCS HOBS™ and the TwinX™ platforms for communications and media companies; Jile™ enterprise Agile planning and delivery tool; TCS MasterCraft™ suite of intelligent automation products and others. This is backed by many accelerators, frameworks and toolsets in areas like data analytics and insights, cloud migration, cloud management, IoT, enterprise application transformation and others. With these patents, how is TCS planning to position itself as a technology company going forward?—asked Tom Easton, Finance Editor at The Economist, at the earnings call. “It’s like asking which is our favourite child”, quips Gopinathan. He then elaborated further that they are equally important to TCS’ researchers who worked on them. However, he also added that the way to think about TCS’ patent portfolio and its approach to patents is not designed to unearth one item that is likely to pivot completely. Citing pharma, Gopinathan said that they have a much more broad-based research agenda across a wide variety of spectrums, and their approach is quite different from what one would observe in other industries. Using a verticalisation approach, TCS plans to consolidate and incorporate these patents into many of its product solutions. He adds, “Some of them could be game-changing on their own, but our approach to it is very different from a typical firm or a technology approach.” TCS on Metaverse TCS said they are currently working on various metaverse projects as proof of capability and individual solutions. The company said that it is currently working on the retail aspect, along with a few banking and financial services aspects. TCS chief Gopinathan said that the projects are exploratory. “But definitely it is something that will happen—how fast and how, [and] what the uptake will be difficult for us to predict,” he adds. He said their approach, however, is to remain engaged. Recently, TCS partnered with Tata group companies like Tanishq, Tata Motors and Croma to deploy metaverse solutions. Besides these partnerships, TCS revealed that it is currently working with a telecom vendor that is helping them develop an eCommerce platform. In addition, it is also working on building a metaverse solution in the wealth management space focused on the younger demographic. Rising attrition Last quarter, the attrition number of TCS was around 17.4 per cent. However, this quarter the attrition rate stands at 19.7 per cent and is increasing. “We think it will rise further in Q2, after which it should start tapering”, says Samir Seksaria, Chief Financial Officer at TCS. So, how is TCS addressing the attrition rate? Rajesh Gopinathan said that there was nothing different from what they saw at the beginning of the quarter while addressing TCS’ attrition rate. He further said that one of the rates is wage increases while the other is the continuing demand environment and the attrition environment leading to the increased operating costs, especially on the employee side. “I think it will take another few more months before it will start to come down. So, till then, the margin pressures will continue, but we hope to sequentially improve from where we have taken that and will hit completely”, says Gopinathan. Further, he revealed that they are keeping a close watch on the rates and will react in accordance to how the demand-supply situation eventually plays out. However, another perspective that can be considered is that the utilisation will increase once the attrition rate comes down. “That is a reasonable assumption. That is our technical operating model also when in a high-demand scenario, we first prioritise demand, and attrition is also linked into that”, says Gopinathan in response to Ravi Menon, Analyst at Macquarie. TCS also announced a salary hike of 5–8 per cent and higher for top performance with effect from April 1, 2022. This decision had a 1.5 per cent impact on operating margins while continued supply-side challenges entailed additional expenses such as backfilling expenses and higher subcontractor usage. This, along with normalising travel expenses, negated various operational efficiencies resulting in an operating margin of 23.1 per cent, a sequential contraction of 1.9 per cent. Compared to its peers, TCS has the lowest attrition rate. However, in terms of the trajectory of controlling attrition, the company seems to be lagging behind the curve compared to others. So, what is TCS doing differently to flatten the attrition rate in the coming quarters?—asked Debashish Mazumdar, Analyst at B&K Securities. In response, Gopinathan said its employee reward programme is much more holistic than its peers. “I believe our employee retention numbers also reflect this holistic employee engagement and total rewards programme. So, we are quite confident about where we are”, he says in reference to the pandemic times when the company had announced that it would not be laying off anybody. He added that TCS was the first company to announce salary hikes in October 2020. Further, he explained that their programme is not entirely based on the short-term impacts, and that TCS is quite confident with its current executive model. He believes that the numbers reflect that confidence as well—“We expect attrition to start tapering off in the next few months.” Additionally, TCS also reported that their workflow expenditure crossed the 600K mark this quarter, ending this quarter with 6,06,331 employees. “We continue to hire talent from across the world with a net addition of 14,136. It is a very diverse workforce with 153 nationalities represented and women making up 35.5 per cent of the base”, says Samir Seksaria. Seksaria further emphasised that TCS remains committed to investing in organic talent development to build the next-generation GNT workforce. He revealed the company clocked 12 million learning hours resulting in the acquisition of 1.7 million competencies. Overall, a good quarter TCS believes FY-23 Q1 has been a quarter of strong performance, continuing its journey from the last few quarters. However, in the coming quarter, the company said that it would continue to chase demand and keep a close eye on how the recession plays out in the near future. “With all our industry verticals showing good growth, our order book and pipeline are also very strong, giving us good visibility for the next few months. Our margin dipped this quarter due to salary increase and other supply-side related costs, but we are confident in our ability to bring it back to our preferred range over time”, says Rajesh Gopinathan. On the people front, TCS states that it would continue to hire across all its markets and add over 14,000 employees on a net basis. On the attrition, the company revealed that it has been elevated at 19.7 per cent in IT services on an LTM basis. “This will probably peak next quarter and then start tapering off”, concludes Rajesh Gopinathan.","excerpt":"TCS is currently working on various metaverse projects as proof of capability and individual solutions","categories":["IT Services"],"tags":["Quarterly Earnings","Tata Consultancy Services","TCS"],"author_name":"Amit Naik","publish_date":"2022-07-14T18:00:00","publication_year":"2022","word_count":1698,"keywords":["Tata Consultancy Services","Go","intelligent automation","programming_languages:R","AI","TCS","R","programming_languages:Go","automation","Aim","analytics","GAN","Quarterly Earnings"],"extracted_tech_keywords":["AI","analytics","Aim","R","Go","GAN","automation","intelligent automation","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/it-services\/tcs-fy23-q1-results-touches-6-lakh-employees-mark-attrition-at-all-time-high-and-more\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":5,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10103565,"title":"OpenAI Secretly Works on Q*, Inches Closer Towards AGI","content":"OpenAI is reportedly working on a project Q* (pronounced Q-Star), capable of solving unfamiliar math problems. A few people at OpenAI believe that Q* could be a big step towards achieving artificial general intelligence (AGI). At the same time, this new model is raising concerns among some AI safety researchers due to the accelerated advancements, particularly after watching the demo of the model circulated within OpenAI in recent weeks, as per The Information. The model is created by OpenAI’s chief scientist Ilya Sutskevar and other top researchers Jakub Pachocki and Szymon Sidor. Interestingly, this new development comes in the background of Andrej Karpathy – who also happened to be building JARVIS at OpenAI – recently posted  on X, saying that he has been thinking of centralisation and decentralisation lately. Karpathy is mostly talking about building an AI system where it involves a trade-off between centralisation and decentralisation of decision-making and information. In order to achieve optimal results, you have to balance these two aspects, and Q-Learning seems to be fitting perfectly in the equation to enable all of this. Damn it, why does @karpathy have to blow my mind at 9 pm the night before Thanksgiving pic.twitter.com\/D6y6ps5kQA— Peter Yang (@petergyang) November 23, 2023 What is Q-Learning? Experts believe that Q* is built on the principles of Q-learning which is a foundational concept in the field of AI, specifically in the area of reinforcement learning. Q-learning’s algorithm is categorised as model-free reinforcement learning, and is designed to understand the value of an action within a specific state. The ultimate goal of Q-learning is to find an optimal policy that defines the best action to take in each state, maximising the cumulative reward over time. Q-learning is based on the notion of a Q-function, aka the state-action value function. This function operates with two inputs: a state and an action. It returns an estimate of the total reward expected, starting from that state, alongside taking that action, and thereafter following the optimal policy. In simple instances, Q-learning maintains a table (known as the Q-table) where each row represents a state and each column represents an action. The entries in this table are the Q-values, which are updated as the agent learns through exploration and exploitation. This is how it works: A key aspect of Q-learning is balancing exploration (trying new things) and exploitation (using known information). This is often managed by strategies like ε-greedy, where the agent explores randomly with probability ε and exploits the best-known action with probability 1-ε.","excerpt":"This new development comes in the background of Andrej Karpathy thinking of centralisation and decentralisation lately.","categories":["Global Tech"],"tags":[],"author_name":"Siddharth Jindal","publish_date":"2023-11-23T12:50:33","publication_year":"2023","word_count":418,"keywords":["Go","OpenAI","AI","programming_languages:R","RPA","ML","programming_languages:Go","AI safety","R"],"extracted_tech_keywords":["AI","ML","OpenAI","R","Go","AI safety","RPA","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/openai-secretly-works-on-q-inches-closer-towards-agi\/","complexity_score":3,"technical_depth":9,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":true},{"article_id":10141736,"title":"Why Everyone Wants to Partner with ServiceNow","content":"Founded two decades ago, Santa Clara-based software giant ServiceNow is a leader in enterprise workflow technology, changing how businesses manage tasks and streamline operations. Fred Luddy, the company’s founder, identified inefficiencies in enterprise task management and sought to create a unified platform that connects the silos of people, processes, and systems. In September, the company released Xanadu, a range of new features in its latest Now Platform update. Among these innovations are ServiceNow AI agents, designed to enhance productivity round the clock across departments, including IT, customer service, HR, and procurement, which the company claims are grounded in reasoning abilities based on enterprise data. These are now released in partnership with Microsoft, which enables them to work on Microsoft 365 Copilot and all the existing Microsoft applications. Speaking with AIM, Sumeet Mathur, SVP and MD of ServiceNow India Technology & Business Centre, explained why everyone wants to partner with ServiceNow instead of competing with them. For context, most Indian IT companies, the big techs of the world, and even companies like Databricks and Snowflake have partnered with the company in recent years. “Our one-platform approach gives us a significant edge over competitors who rely on acquisitions to piece together solutions,” Mathur explained. ServiceNow stands out with its commitment to providing a single platform, data model, and architecture. This strategy ensures an integrated and seamless user experience across an enterprise’s various functions. ServiceNow’s biannual platform releases, named after cities, symbolise the company’s steady innovation. “We’ve gone from Aspen to Xanadu, with Yokohama and Zurich on the horizon,” Mathur added. The AI Edge ServiceNow has been actively working in AI for over a decade, making the transition to GenAI an incremental step rather than a leap. Three years ago, ServiceNow acquired Element AI, which laid the foundation for its advancements in LLMs. “Our GenAI offerings are persona-focused,” Mathur explained, emphasising tailored solutions for employees, agents, and developers. For example, employee-centric tools like AI Search and chatbots empower users to solve issues themselves, reducing the need to open tickets. On the agent side, GenAI-driven tools enhance productivity by summarising case notes and providing next-step recommendations. Developers, too, benefit from innovations like low-code and image-to-flow capabilities. “In Xanadu, we introduced mechanisms that allow customers and partners to build net-new GenAI capabilities on our platform, enabling them to solve use cases beyond our out-of-the-box offerings,” Mathur noted. This is after ServiceNow, in partnership with Hugging Face, released StarCoder 2, its own AI coding platform, trained on a larger dataset (7.5 terabytes) than its predecessors and on 619 programming languages. StarCoder 2 comes in three sizes – 3-billion,  7-billion and 15-billion-parameter models. The text-to-code LLM is trained on ServiceNow’s proprietary source code and thus generates output that is significantly more efficient, scalable, and suited for real-world environments compared to other open-source LLMs. “This is how we tailor LLMs for ServiceNow’s work. Additionally, we provide a low-code, no-code platform that empowers customers and partners to build their own applications. Since we serve diverse enterprises, it’s impractical to anticipate every use case for every industry or domain,” said Mathur. The Competitive Edge ServiceNow’s platform has evolved to cater to industry-specific needs, with solutions tailored to sectors such as retail, manufacturing, healthcare, and telecommunications. When asked about competition, Mathur provided a nuanced perspective. Companies like Snowflake and Salesforce, often perceived as competitors, are in fact collaborators in some areas. “We announced a partnership with Snowflake that allows zero-copy integration of data,” Mathur revealed. This enables customers to use Snowflake’s data lake capabilities seamlessly within ServiceNow workflows, enhancing efficiency and preserving data integrity. However, direct competitors like Atlassian and Zendesk are tackled differently. “Because we operate at a platform layer with products across IT, asset management, security, and more, our breadth and depth give us a unique positioning,” he explained. ServiceNow’s integrated approach contrasts with point solutions, making it a preferred choice for enterprises seeking end-to-end workflow management. Mathur said that ServiceNow is focused on where the enterprise understands the value and large-scale enterprise – the Fortune 500s of the world. “We play in the upper echelon of the enterprise segment and not in an SME space. So, we won’t compete there; we are not going to go after a $5,000 deal.” Indian AI Adoption India plays a pivotal role in the company’s global operations. Having completed 10 years, the centre is instrumental in driving innovation and addressing the specific needs of regional and global customers. Mathur highlighted the significance of India’s talent pool in enabling ServiceNow’s rapid development cycles. The company is committed to fostering AI advancements within India. “India’s vibrant technology landscape is ripe for GenAI adoption,” Mathur observed, emphasising the role of local enterprises in driving global shifts. Furthermore, ServiceNow’s AI Accelerator Program was launched primarily to empower startups and developers in India. It seeks to nurture innovative use cases and scalable AI-driven applications, further solidifying ServiceNow’s position as a catalyst for tech revamp in the region. As ServiceNow moves forward, its focus remains on empowering enterprises with transformative technologies. With the upcoming Yokohama and Zurich releases, the company is poised to deliver even more innovations. Mathur’s parting thought encapsulates ServiceNow’s ethos: “We’re not just building tools; we’re shaping the future of work by connecting people, processes, and technology in ways that truly matter.”","excerpt":"“Our one-platform approach gives us a significant edge over competitors who rely on acquisitions to piece together solutions,” said Sumeet Mathur, SVP and MD of ServiceNow India Technology & Business Centre.","categories":["AI Features"],"tags":["ServiceNow"],"author_name":"Mohit Pandey","publish_date":"2024-11-27T08:00:00","publication_year":"2024","word_count":877,"keywords":["ServiceNow","Hugging Face","GenAI","TPU","AI","chatbots","ML","Aim","Databricks","R","Snowflake"],"extracted_tech_keywords":["AI","ML","GenAI","Aim","Hugging Face","chatbots","TPU","Snowflake","Databricks","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-everyone-wants-to-partner-with-servicenow\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10161819,"title":"How Yulu Manages the Traffic Menace it Creates in Bengaluru","content":"In Bengaluru, when using a quick commerce service, one must have witnessed delivery representatives riding small blue electric scooters to deliver goods. Sometimes, these scooters can be spotted at parking spots on the sidewalk or lying around near the curb, which often creates a menace for the traffic. These electric scooters were built by Yulu, a smart shared mobility startup founded in 2017 by Amit Gupta, Naveen Dachuri, RK Misra and Anuj Tewari, to solve the simple problem of short-distance mobility. Dachuri, the CTO of the company, told AIM that he is aware of the inconvenience that these scooters sometimes create in the city and assured that the company handles it as much as possible from its end. While Dachuri delved a little into how the company plans to integrate chatbots in Indian languages on scooters, much of the plans are still in the ideation stage. Real-Time Predictive Models “If you have to run smart mobility, the most important thing is…to have some kind of intelligence built in the vehicle so that the manual interferences can stop,” Dachuri said. According to him, this can only be achieved if there is a device built inside the vehicle which can give information in real time about the whereabouts and the condition of the vehicle. Off-the-shelf products did not work for Yulu, as the changes needed for their business took months. To address this challenge, the company developed its proprietary Internet of Things (IoT) system, Yulu Connect. This system gathers vehicle data and performs edge computing, which enables real-time decision-making. It can do this both while the ride is on and the vehicle is parked. This ensured no delay and no compromise of the customer experience. Real-time data is primarily used to enhance customer experience, such as guiding users to the nearest battery-swapping station based on availability. MySQL is used for real-time data storage, while telemetry and sensor data flow into TiDB, a distributed SQL database. Additionally, streaming engines process live data to enable actions such as dispatching field executives for maintenance and directing users to optimal parking spots. Apart from using the data in real time, which is purely based on the current state of the vehicle, Yulu also uses the data to build predictive models to increase the uptime of the vehicle and improve the efficiency. This also helps in real-time prediction based on historical information. “We match these two pieces of information and tell users to go to a certain location so that your probability of getting a battery or a vehicle will be higher,” Dachuri explained. Reducing Repair and Maintenance The predictive models made by Yulu are beneficial for scheduling vehicle repairs and maintenance. The problem with shared mobility is that when original equipment manufacturers (OEMs) provide the vehicles, the learnings about the possible wear and tear are based on personal vehicles and not shared usage, where multiple people are using the same vehicle. This is where Yulu’s data comes into the picture. Dachuri and his team have collected data over the years about how the different types of braking, throttling, riding over potholes, and even sitting posture affect the long-term condition of the vehicle. “Everyone wants to ride in a smooth fashion. But unfortunately, the conditions make you take those routes. And that is how I at least see with respect to Bengaluru roads.” Addressing the issue of parking and vehicle mismanagement, Dachuri said that his team built a geofencing framework which encourages users to park at designated stations. However, geolocation accuracy remains a challenge due to factors like satellite visibility and urban infrastructure interference. To counter this, Yulu combines IoT device data with users’ smartphone GPS to optimise parking suggestions. However, this also results in significant errors and drainage of the user’s phone battery when the location is optimised. “We suggest operators to use the most optimal location accuracy, but oftentimes they end up choosing something else,” Dachuri said. He added that despite technological solutions, user education remains vital in ensuring compliance with parking guidelines. The Safety Issue By the time all these things are addressed, a lot of menace is already created. Yulu sometimes, which is very rare, also suffers from regulatory problems from the government, such as the seizure of vehicles or hefty fines because of the created menace by the gig workers, including driving on the wrong side, among several other violations. To address a lot of these road safety issues, Yulu has developed a system to detect wrong-way driving. Initially relying on Google APIs, the company found them insufficient due to frequent road direction changes. Consequently, they built an in-house solution using simple mathematical models to detect wrong-way movements and issue warnings and penalties to violators. Dachuri, however, said that many of these gig workers do it to deliver multiple orders from dark stores at the same time, which forces them to travel on the wrong side, which is also not an isolated case for Yulu users. “People opt for more deliveries than safety.” Dachuri added that the team tries to advise them as much as they can, but it is up to the users. Yulu also employs strategic field operations to maintain order. “During low-traffic hours, field executives reposition vehicles and ensure cleanliness. Signages and designated parking zones are regularly monitored, though evolving city infrastructure poses challenges in keeping them updated,” Dachuri explained. Yulu also offers ISA-certified helmets for drivers, which they can buy at a low price to ensure their safety. The only problem that still remains, according to Dachuri, is the education and awareness of Yulu users; the tech part is now mostly fixed.","excerpt":"Apart from using data in real time, Yulu also uses the data to build predictive models to increase the uptime of the vehicle and improve the efficiency.","categories":["AI Features"],"tags":["Bengaluru","Yulu"],"author_name":"Mohit Pandey","publish_date":"2025-01-20T16:10:55","publication_year":"2025","word_count":934,"keywords":["Go","API","AI","chatbots","RAG","Aim","ViT","SQL","Yulu","edge computing","Bengaluru","R"],"extracted_tech_keywords":["AI","Aim","RAG","chatbots","edge computing","R","SQL","Go","API","ViT"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-yulu-manages-the-traffic-menace-it-creates-in-bengaluru\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10088022,"title":"Google Introduces Offline Reinforcement Learning to Train AI Agents","content":"Researchers from Google have developed a pre-trained model built on the CQL algorithm for scaled offline Reinforcement Learning (RL) called Scaled Q-Learning, to efficiently train RL agents for decision-making tasks such as playing games or picking up objects. Contrary to the traditional approach, Scaled Q-Learning uses diverse data to learn representations for quick transfer to new tasks. Moreover, it outperforms Transformer-based methods and others which use larger models. The researchers evaluated the approach on a suite of Atari games, where the goal is to train a single RL agent using data from low-quality players and then learn new variations in pre-training or completely new games. This method made initial progress towards enabling more practical real-world training of RL agents as an alternative to costly and complex simulation-based pipelines or large-scale experiments. The team tested the method using two types of data: near-optimal data and low-quality data. They compared it to other methods and found that only Scaled Q-Learning improved on the offline data, reaching about 80% of human performance. The study shows that pre-training RL agents with multi-task offline learning can significantly improve their performance on different tasks, even for challenging ones like Atari games with different appearances and dynamics. The results demonstrate that the method can significantly boost the performance of RL, both in offline and online modes. In online RL, Scaled Q-Learning can provide improvements while methods like MAE yield little improvement. It can incorporate prior knowledge from pre-training games to improve the final score after 20k online interactions. In conclusion, Scaled Q-learning suggests that it is learning the game dynamics, not just improving the visual features like other techniques. As per the blog, the work could develop generally capable pre-trained RL agents that can develop applicable interaction skills from large-scale offline pre-training. Future research will involve validating these results on a broader range of more realistic tasks, in domains such as robotics and NLP.","excerpt":"Scaled Q-Learning can efficiently train RL agents to play Atari or pick up objects.","categories":["AI News"],"tags":["AI Research","Google","Google Research","Reinforcement Learning"],"author_name":"Tasmia Ansari","publish_date":"2023-02-24T12:10:41","publication_year":"2023","word_count":317,"keywords":["Go","Reinforcement Learning","programming_languages:R","AI","programming_languages:Go","AI Research","NLP","Google","ai_applications:robotics","ai_applications:NLP","R","Google Research"],"extracted_tech_keywords":["AI","NLP","R","Go","programming_languages:R","programming_languages:Go","ai_applications:NLP","ai_applications:robotics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/google-introduces-offline-reinforcement-learning-to-train-ai-agents\/","complexity_score":3,"technical_depth":8,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10049801,"title":"IIT Delhi Establishes Quantum Technologies Center Of Excellence (CoE)","content":"The Indian Institute of Technology (IIT) Delhi has created a Centre of Excellence (CoE) on Quantum Technologies to consolidate IIT Delhi’s research activity in diverse disciplines of Quantum Technologies. The CoE will promote synergy and coherence among the institute’s operations and assist principal investigators in securing additional funding from the DST and other funding bodies. Rajendra Singh, Head, School of Interdisciplinary Research (SIRe) and lead PI, IIT Delhi’s CoE on Quantum Technologies, stated, “Over the last century or more, quantum physics has had an unparalleled impact on civilisation. Researchers invented both the laser and the transistor by utilising quantum physics properties of light and materials. These inventions serve as the foundation for information technology as a whole – computers, the internet, and much more – which has altered our society to a considerable extent. That was the start of the first quantum revolution. The second quantum revolution is rapidly approaching, owing to advances in our ability to detect and manipulate single quantum objects such as photons and atoms.” “The focus areas would include quantum computing, quantum communication, quantum sensing and metrology, and quantum materials and devices,” Prof Singh stated when asked about the precise target areas for quantum technology research at IIT Delhi. Along with the design and development of new quantum materials, the centre conducts research on quantum processors and cryogenic controllers, semiconductor cubics such as CMOS and 2D materials, quantum sensing and measurement, and quantum biophotonics. Additionally, single-photon detectors and light sources based on semiconductors (two-dimensional materials, III-V), superconductors, and others will be developed.","excerpt":"The Indian Institute of Technology (IIT) Delhi launches a Centre of Excellence (CoE) on Quantum Technologies to consolidate IIT Delhi’s research activity in diverse disciplines of Quantum Technologies.","categories":["AI News"],"tags":["Quantum Computing","quantum technology"],"author_name":"Dr. Nivash Jeevanandam","publish_date":"2021-09-27T13:06:30","publication_year":"2021","word_count":256,"keywords":["Quantum Computing","API","funding","programming_languages:R","ViT","R","emerging_tech:quantum computing","quantum technology"],"extracted_tech_keywords":["R","API","ViT","funding","programming_languages:R","emerging_tech:quantum computing"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/quantum-technologies-center-of-excellence\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10162271,"title":"MapmyIndia Acquires 9.37% Stake in AI Startup SimDaaS Autonomy","content":"Digital mapping and geospatial services company MapmyIndia has acquired a 9.37% equity stake in Kanpur-based startup SimDaaS Autonomy, according to reports. The deal, valued at approximately INR 3 crore, involves both equity funding and preference shares. The partnership with SimDaaS aligns with MapmyIndia’s strategy to stay ahead in the geo-technology space. With this investment, the company intends to integrate advanced AI and simulation tools to enhance its offerings, particularly in the automotive sector. The collaboration is expected to support the creation of more advanced solutions for autonomous driving and safety technologies. MapmyIndia, founded in 1995 by Rakesh Verma and Rashmi Verma, is known for its comprehensive suite of navigation and location-based services. It caters to industries such as logistics, automotive, and consumer services. SimDaaS Autonomy, established in 2023 and based in Kanpur, focuses on developing simulation technologies for autonomous systems and advanced driver-assistance systems (ADAS). The startup is led by Bharat Lohani, a director with academic ties to IIT Kanpur. Its simulation-driven development platforms aim to enhance the design, testing, and validation of autonomous vehicles. Alongside this strategic acquisition, MapmyIndia recently witnessed significant organisational changes. Rohan Verma, CEO of MapmyIndia and son of the company’s founders, announced his resignation to focus on a new business-to-consumer (B2C) venture. Initially, MapmyIndia had planned to invest in Verma’s new project, but the board decided against it after receiving feedback from investors. This decision led to a 16% surge in MapmyIndia’s stock, indicating strong investor confidence in the company’s future direction. The acquisition of SimDaaS comes at a time when demand for autonomous systems and AI-driven solutions is accelerating. By leveraging SimDaaS’s expertise in simulation technologies, MapmyIndia is poised to expand its capabilities in the growing ADAS and autonomous vehicle markets. MapmyIndia x Automobiles Interestingly, MapmyIndia, entered into a strategic collaboration with Qualcomm Technologies to advance ‘Make in India’ solutions for the automotive sector. The partnership aims to integrate Qualcomm’s Snapdragon Digital Chassis solutions with MapmyIndia’s expertise to deliver connectivity solutions for both four-wheelers and two-wheelers. Similarly, MapmyIndia also partnered with Bengaluru-based startup KOGO AI to launch India’s first universal voice assistant for the automotive industry. KOGO AI is building a customisable AI operating system, KOGO OS, which can build, manage, and deploy various AI agents across industries. “Unlike Siri and Alexa, these agents not only give you access but can also operate apps and services on your behalf,” said Raj K Gopalakrishnan, co-founder and CEO of Kogo Tech Labs, in an exclusive interaction with AIM. Interestingly, in a strategic move to attract developers and businesses, Google recently announced a 70% reduction in the pricing of its Maps APIs. This significant price cut intensifies competition in the digital mapping sector, challenging companies such as MapmyIndia. Concurrently, MapmyIndia’s CEO, Verma, has openly criticised Ola’s mapping initiatives, referring to them as a “gimmick.” He expressed concerns about the quality and reliability of Ola’s mapping services and cautioned users to be wary of potential inaccuracies.","excerpt":"With this, MapmyIndia aims to integrate AI-driven simulation and autonomy solutions to enhance its geospatial and automotive technologies.","categories":["AI News"],"tags":["Google Map","KOGO","mapmyindia","Mergers and Acquisitions","Ola"],"author_name":"Vandana Nair","publish_date":"2025-01-27T18:08:09","publication_year":"2025","word_count":488,"keywords":["Go","API","Ola","AI","Git","Google Map","RAG","Aim","ViT","GAN","KOGO","Mergers and Acquisitions","R","mapmyindia","startup"],"extracted_tech_keywords":["AI","Aim","RAG","R","Go","Git","API","GAN","ViT","startup"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/mapmyindia-acquires-9-37-stake-in-ai-startup-simdaas-autonomy\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":60090,"title":"Hands-On Guide To LSTM Recurrent Neural Network For Stock Market Prediction","content":"Predicting stock prices is an uncertain task which is modelled using machine learning to predict the return on stocks. There are a lot of methods and tools used for the purpose of stock market prediction. The stock market is considered to be very dynamic and complex in nature. An accurate prediction of future prices may lead to a higher yield of profit for investors through stock investments. As per the predictions, investors will be able to pick the stocks that may give a higher return. Over the years, various machine learning techniques have been used in stock market prediction, but with the increased amount of data and expectation of more accurate prediction, the deep learning models are being used nowadays which have proven their advantage over traditional machine learning methods in terms of accuracy and speed of prediction. In this article, we will discuss the Long-Short-Term Memory (LSTM) Recurrent Neural Network, one of the popular deep learning models, used in stock market prediction. In this task, we will fetch the historical data of stock automatically using python libraries and fit the LSTM model on this data to predict the future prices of the stock. LSTM Recurrent Neural Network Long-Short-Term Memory Recurrent Neural Network belongs to the family of deep learning algorithms. It is a recurrent network because of the feedback connections in its architecture. It has an advantage over traditional neural networks due to its capability to process the entire sequence of data. Its architecture comprises the cell, input gate, output gate and forget gate. The cell remembers values over arbitrary time intervals, and the three gates regulate the flow of information into and out of the cell. The cell of the model is responsible for keeping track of the dependencies between the elements in the input sequence. The input gate controls the extent to which a new value flows into the cell, the forget gate controls the extent to which a value remains in the cell, and the output gate controls the extent to which the value in the cell is used to compute the output activation of the LSTM unit. However, there are some variants of the LSTM model such as Gated Recurrent Units (GRUs) that do not have the output gate. LSTM Networks are popularly used on time-series data for classification, processing, and making predictions. The reason for its popularity in time-series application is that there can be several lags of unknown duration between important events in a time series. (For more details on LSTM, please read this post: How to Code Your First LSTM Network in Keras) Stock Prediction In this task, the future stock prices of State Bank of India (SBIN) are predicted using the LSTM Recurrent Neural Network. Our task is to predict stock prices for a few days, which is a time series problem. The LSTM model is very popular in time-series forecasting, and this is the reason why this model is chosen in this task. The historical prices of SBIN are collected automatically using the nsepy library of python. We have used 6 years of historical price data, from 01.01.2013 to 31.12.2018. This data set contains 1483 observations with 12 attributes. After preprocessing, only dates and OHLC (Open, High, Low, Close) columns, a total of 5 columns, are taken as these columns have main significance in the dataset. The LSTM model is trained on this entire dataset, and for the testing purpose, a new dataset is fetched for the duration between 01.01.2019 to 18.09.2019. The stock prices for this new duration will be predicted by the already trained LSTM model, and the predicted prices will be plotted against the original prices to visualise the model’s accuracy. Implementation Before proceeding further, make sure that you have installed TensorFlow and nsepy libraries. TensorFlow will be used as a backend for LSTM model, and nsepy will be used to fetch the historical stock data. Once installed, follow the below steps: 1.Import the required libraries from nsepy import get_history as gh import datetime as dt from matplotlib import pyplot as plt from sklearn import model_selection from sklearn.metrics import confusion_matrix from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split import numpy as np import pandas as pd from sklearn.preprocessing import MinMaxScaler from keras.models import Sequential from keras.layers import Dense from keras.layers import LSTM from keras.layers import Dropout 2. We will fetch 6 years of historical prices of SBIN from 01.01.2013 to 31.12.2018. So we need to set the start and end dates and pass these parameters to the function for fetching the data. start = dt.datetime(2013,1,1) end = dt.datetime(2018,12,31) stk_data = gh(symbol='SBIN',start=start,end=end) 3. We can visualise the fetched data in the above step. For simplicity, only the day-wise closing prices are visualised. plt.figure(figsize=(14,14)) plt.plot(stk_data['Close']) plt.title('Historical Stock Value') plt.xlabel('Date') plt.ylabel('Stock Price') plt.show() 4. There are 12 columns in the fetched data. Many of the columns are not of our interest so only significant columns are selected to create the main dataset. stk_data['Date'] = stk_data.index data2 = pd.DataFrame(columns = ['Date', 'Open', 'High', 'Low', 'Close']) data2['Date'] = stk_data['Date'] data2['Open'] = stk_data['Open'] data2['High'] = stk_data['High'] data2['Low'] = stk_data['Low'] data2['Close'] = stk_data['Close'] 5. Preprocess the data in order to prepare it for the LSTM model. The data fetched in step one is used for training purpose only. For testing purpose, different data will be fetched later. train_set = data2.iloc[:, 1:2].values sc = MinMaxScaler(feature_range = (0, 1)) training_set_scaled = sc.fit_transform(train_set) X_train = [] y_train = [] for i in range(60, 1482): X_train.append(training_set_scaled[i-60:i, 0]) y_train.append(training_set_scaled[i, 0]) X_train, y_train = np.array(X_train), np.array(y_train) X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1)) 6. Define the LSTM Recurrent Neural Network. Here, you can add more LSTM layers and adjust the dropout in order to improve the accuracy of the model. regressor = Sequential() regressor.add(LSTM(units = 50, return_sequences = True, input_shape = (X_train.shape[1], 1))) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50, return_sequences = True)) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50, return_sequences = True)) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50)) regressor.add(Dropout(0.2)) regressor.add(Dense(units = 1)) 7. Compile and train the model defined in the above step. Iteratively, you can increase or decrease the epochs and batch size to get more accuracy. regressor.compile(optimizer = 'adam', loss = 'mean_squared_error') regressor.fit(X_train, y_train, epochs = 15, batch_size = 32) 8. Now, our model is trained and needs to be tested on the testing data. For this purpose, fetch the new data for a different period. Preprocessing steps are similar as we have done with training data. testdataframe= gh(symbol='SBIN',start=dt.datetime(2019,1,1),end=dt.datetime(2019,9,18)) testdataframe['Date'] = testdataframe.index testdata = pd.DataFrame(columns = ['Date', 'Open', 'High', 'Low', 'Close']) testdata['Date'] = testdataframe['Date'] testdata['Open'] = testdataframe['Open'] testdata['High'] = testdataframe['High'] testdata['Low'] = testdataframe['Low'] testdata['Close'] = testdataframe['Close'] real_stock_price = testdata.iloc[:, 1:2].values dataset_total = pd.concat((data2['Open'], testdata['Open']), axis = 0) inputs = dataset_total[len(dataset_total) - len(testdata) - 60:].values inputs = inputs.reshape(-1,1) inputs = sc.transform(inputs) X_test = [] for i in range(60, 235): X_test.append(inputs[i-60:i, 0]) X_test = np.array(X_test) X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1)) 9. Test the LSTM model on the new dataset. predicted_stock_price = regressor.predict(X_test) predicted_stock_price = sc.inverse_transform(predicted_stock_price) 10. Visualize the predicted stock prices with original stock prices. plt.figure(figsize=(20,10)) plt.plot(real_stock_price, color = 'green', label = 'SBI Stock Price') plt.plot(predicted_stock_price, color = 'red', label = 'Predicted SBI Stock Price') plt.title('SBI Stock Price Prediction') plt.xlabel('Trading Day') plt.ylabel('SBI Stock Price') plt.legend() plt.show() The plot is shown in the below image. Binding all the above steps together. #Importing the libraries from nsepy import get_history as gh import datetime as dt from matplotlib import pyplot as plt from sklearn import model_selection from sklearn.metrics import confusion_matrix from sklearn.preprocessing import StandardScaler from sklearn.model_selection import train_test_split import numpy as np import pandas as pd from sklearn.preprocessing import MinMaxScaler from keras.models import Sequential from keras.layers import Dense from keras.layers import LSTM from keras.layers import Dropout #Setting start and end dates and fetching the historical data start = dt.datetime(2013,1,1) end = dt.datetime(2018,12,31) stk_data = gh(symbol='SBIN',start=start,end=end) #Visualizing the fetched data plt.figure(figsize=(14,14)) plt.plot(stk_data['Close']) plt.title('Historical Stock Value') plt.xlabel('Date') plt.ylabel('Stock Price') plt.show() #Data Preprocessing stk_data['Date'] = stk_data.index data2 = pd.DataFrame(columns = ['Date', 'Open', 'High', 'Low', 'Close']) data2['Date'] = stk_data['Date'] data2['Open'] = stk_data['Open'] data2['High'] = stk_data['High'] data2['Low'] = stk_data['Low'] data2['Close'] = stk_data['Close'] train_set = data2.iloc[:, 1:2].values sc = MinMaxScaler(feature_range = (0, 1)) training_set_scaled = sc.fit_transform(train_set) X_train = [] y_train = [] for i in range(60, 1482): X_train.append(training_set_scaled[i-60:i, 0]) y_train.append(training_set_scaled[i, 0]) X_train, y_train = np.array(X_train), np.array(y_train) X_train = np.reshape(X_train, (X_train.shape[0], X_train.shape[1], 1)) #Defining the LSTM Recurrent Model regressor = Sequential() regressor.add(LSTM(units = 50, return_sequences = True, input_shape = (X_train.shape[1], 1))) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50, return_sequences = True)) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50, return_sequences = True)) regressor.add(Dropout(0.2)) regressor.add(LSTM(units = 50)) regressor.add(Dropout(0.2)) regressor.add(Dense(units = 1)) #Compiling and fitting the model regressor.compile(optimizer = 'adam', loss = 'mean_squared_error') regressor.fit(X_train, y_train, epochs = 15, batch_size = 32) #Fetching the test data and preprocessing testdataframe = gh(symbol='SBIN',start=dt.datetime(2019,1,1),end=dt.datetime(2019,9,18)) testdataframe['Date'] = testdataframe.index testdata = pd.DataFrame(columns = ['Date', 'Open', 'High', 'Low', 'Close']) testdata['Date'] = testdataframe['Date'] testdata['Open'] = testdataframe['Open'] testdata['High'] = testdataframe['High'] testdata['Low'] = testdataframe['Low'] testdata['Close'] = testdataframe['Close'] real_stock_price = testdata.iloc[:, 1:2].values dataset_total = pd.concat((data2['Open'], testdata['Open']), axis = 0) inputs = dataset_total[len(dataset_total) - len(testdata) - 60:].values inputs = inputs.reshape(-1,1) inputs = sc.transform(inputs) X_test = [] for i in range(60, 235): X_test.append(inputs[i-60:i, 0]) X_test = np.array(X_test) X_test = np.reshape(X_test, (X_test.shape[0], X_test.shape[1], 1)) #Making predictions on the test data predicted_stock_price = regressor.predict(X_test) predicted_stock_price = sc.inverse_transform(predicted_stock_price) #Visualizing the prediction plt.figure(figsize=(20,10)) plt.plot(real_stock_price, color = 'green', label = 'SBI Stock Price') plt.plot(predicted_stock_price, color = 'red', label = 'Predicted SBI Stock Price') plt.title('SBI Stock Price Prediction') plt.xlabel('Trading Day') plt.ylabel('SBI Stock Price') plt.legend() plt.show()","excerpt":"In this article, we will discuss the Long-Short-Term Memory (LSTM) Recurrent Neural Network, one of the popular deep learning models, used in stock market prediction. In this task, we will fetch the historical data of stock automatically using python libraries and fit the LSTM model on this data to predict the future prices of the stock.","categories":["Deep Tech"],"tags":["LSTM Network","lstm recurrent neural network"],"author_name":"Dr. Vaibhav Kumar","publish_date":"2020-03-27T10:00:00","publication_year":"2020","word_count":1550,"keywords":["NumPy","machine learning","Keras","AI","neural network","Ray","Matplotlib","deep learning","lstm recurrent neural network","LSTM Network","TensorFlow","Pandas"],"extracted_tech_keywords":["AI","machine learning","deep learning","neural network","Ray","TensorFlow","Keras","Pandas","NumPy","Matplotlib"],"url":"https:\/\/analyticsindiamag.com\/deep-tech\/hands-on-guide-to-lstm-recurrent-neural-network-for-stock-market-prediction\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":false},{"article_id":10068074,"title":"How to build machine learning models with Databricks?","content":"Databricks is a business software startup that provides Data Engineering tools for processing and transforming massive amounts of data to develop machine learning models. Traditional Big Data procedures are not only slow to complete jobs but also take more time to build up Hadoop clusters. However, Databricks is built on top of distributed Cloud computing infrastructures like Azure, AWS, or Google Cloud, which allow programmes to execute on CPUs or GPUs according to analytical needs. In this article, we will be learning about building a machine learning model in Databricks. Following are the topics to be covered. Table of contents Data uploadingCreating a clusterData preprocessingBuilding the ML model In this article, we will be building a multivariate linear regression model for predicting the charges on insurance offered by the company based on different features. While logging in there are two options one is for business use with the cloud services and the other is for community usage. You can select anyone based on needs. In this article, we will use the community edition. Data uploading Once you have registered and logged on to the Databricks web page as a community edition user. To upload the data click on “Browse file” to upload the data from the local machine in the Data import section. The webpage looks like this. Analytics India Magazine Once selected the file that is to be uploaded, the page would be redirected to the create a new table page. You can import and store the data using different sources. In this article, we are directly uploading the data and creating a new table in the notebook. Analytics India Magazine As in the above image, in the upload file section the file is being uploaded and now select the “Create Table in Notebook”. To analyze the data we need to create a “Cluster”. Let’s see how to create a cluster. Creating a cluster A Databricks Cluster is a set of computing resources and settings that may be used to run tasks and notebooks. Streaming Analytics, ETL Pipelines, Machine Learning, and Ad-hoc analytics are some of the workloads that can be executed on a Databricks Cluster. To create a cluster go to compute from the sidebar options and click on create a cluster. Analytics India Magazine Analytics India Magazine Define the cluster name and select the runtime version. In this article, we are going to use the Spark 3.2.1 version. So in the unpaid version, you can only use one cluster at a time to use multiple needs to upgrade to paid. Analytics India Magazine Data preprocessing Once the data is uploaded and the cluster is created the data could be analyzed. A notebook would be created where you need to store the uploaded data into a table so that it could be used for further analysis. Analytics India Magazine Now let’s see the process to store and read the data from the table. file_location = \"\/FileStore\/tables\/insurance.csv\" file_type = \"csv\" data=spark.read.csv(file_location,header=True,inferSchema=True) The schema is stored in the variable data. Let’s view the schema. data.printSchema() Analytics India Magazine The data could be viewed by using the “show” function. Let’s view the data. data.show() Analytics India Magazine Now convert the categorical data to numerical by using the encoder. For this article, we are using the ordinal encoder. from pyspark.ml.feature import StringIndexer indexer=StringIndexer(inputCols=['sex','smoker','region'], outputCols=['sex_enc','smoker_enc','region_enc']) df=indexer.fit(data).transform(data) df.show() The “StringIndexer” is an ordinal encoder which will encode the categorical columns. The original data need to be fit and transformed into the encoded data. Analytics India Magazine To use the data for the training and testing purpose the independent variables need to be assembled in one column. The “VectorAssembler” will transform the features into a vector column. So, that the entire dataset could be used at once. from pyspark.ml.feature import VectorAssembler feature_assembler=VectorAssembler(inputCols=['sex_enc', 'smoker_enc', 'region_enc','bmi', 'children','age'],outputCol=\"Independent features\") output=feature_assembler.transform(df) output.select('Independent features').show() Analytics India Magazine Let’s save the independent variable and dependent variable into another variable. So to split it into train and test. The target variable is “Charges” df_utils=output.select('Independent features','charges') df_utils.show() Analytics India Magazine Building the ML model The data has been processed and it is ready to be utilized for training and testing purposes. The data will be split into standard 70:30 ratio train and test respectively. from pyspark.ml.regression import LinearRegression train,test=df_utils.randomSplit([0.70,0.30]) lr=LinearRegression(featuresCol='Independent features',labelCol='charges') lr=lr.fit(train) The linear regression model is trained. Let’s check the coefficients and intercept of the linear relationship. To check the coefficients of the independent variable use “.coefficients” lr.coefficients Analytics India Magazine As there are six independent variables so there are six different coefficients. To check the intercept of the linear line use “.intercept” lr.intercept Analytics India Magazine Let’s evaluate the model performance. To evaluate the performance use “.evaluate” and store the results in a variable for further analysis. pred=lr.evaluate(test) pred.r2,pred.meanAbsoluteError,np.sqrt(pred.meanSquaredError) Analytics India Magazine The r_squared value is 0.74, the MAE is 4127.94 and the RMSE is 6063.94. We could say that the relationship between the dependent variable and independent variable is explained pretty well. To check the predictions use “.predictions” pred.predictions.show() Analytics India Magazine To deploy the model, save the model and then register. The model needs to be registered in the ML flow Model Registry. Once the model is registered it could simply reference this model within Databricks. It could be registered programmatically, as well as by UI in cloud services AWS, Azure, and GCP. Conclusions For data engineers, data scientists, data analysts, and business analysts, Databricks offers a Unified Data Analytics Platform. It is very adaptable across several ecosystems, including AWS, GCP, and Azure. Databricks ensure data reliability and scalability through delta lake. With this article, we have learned to build machine learning with Databrick in Apache Spark. References Read more about DatabricksRead more about the deployment in Databricks","excerpt":"Databricks offers a Unified Data Analytics Platform.","categories":["AI Trends"],"tags":["Cloud Computing","Databricks","spark","SQL"],"author_name":"Sourabh Mehta","publish_date":"2022-05-31T11:00:00","publication_year":"2022","word_count":949,"keywords":["machine learning","GCP","AWS","AI","cloud computing","TPU","spark","ML","Cloud Computing","analytics","SQL","model registry","Azure","Databricks"],"extracted_tech_keywords":["AI","machine learning","ML","analytics","model registry","cloud computing","AWS","Azure","GCP","TPU"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/how-to-build-machine-learning-models-with-databricks\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10169829,"title":"Why Developers are Quietly Quitting Golang","content":"When a developer Yash Batra chose Golang for an early-stage fintech startup, it felt like the right move. Golang, also called Go, was fast, modern, and battle-tested at Google, Kubernetes, and Docker. Six months later, however, he rewrote the backend entirely with Kotlin. His recent blog, titled ‘I Picked Golang for My Startup—Biggest Mistake of My Life’, has since resonated with dozens of developers who once believed in Go’s promise, only to find that it didn’t live up to the hype. “Everything we built took twice as long. We weren’t creating value—we were building tools just to build tools,” Batra wrote. “Go slowed us down, sabotaged our productivity, and nearly derailed our product vision…Go is your friend if you’re building the next Prometheus. But for startups? You’ll thank yourself for picking expressive tech,” Batra concluded. Batra’s message is clear—Go is not the developer’s favourite language anymore. And there’s even more concerning evidence to back this up. Ian Lance Taylor, one of the long-time maintainers of Go at Google, recently bid farewell to Google. But why? “Go has reached the status of being just another programming language,” Taylor said in his farewell note after leaving Google. “That is far beyond what any of us expected.” Taylor emphasised that Go is far from finished and will continue to evolve. However, he also acknowledged that the programming environment has changed, and Go must adapt accordingly. In today’s fast-paced product development environment, Go may be too low-level and rigid. Startups trying to move quickly often find Go’s minimalism more of a hurdle than a feature. Despite powering giants like Docker and Kubernetes, Go is no longer the rising star it once was. It Gets a Lot Deeper Go’s origin story is steeped in engineering excellence. It was developed in 2007 by Rob Pike, Ken Thompson, and Robert Griesemer at Google to simplify systems programming with strong concurrency support and a lean syntax. There was a time when Go was hailed as “the C of the cloud”, a go-to language for infrastructure tools, networking applications, and cloud-native microservices. Its creators imagined a language that would do less but do it well. And it did—for a particular set of problems. However, since 2022, as evident in a blog by fasterthanlime titled ‘Lies we tell ourselves to keep using Golang’, developers and companies have begun rethinking their relationship with the language, considering it a prototyping language or a good entry-level language. The blog notes that many early adopters of Go, who championed its simplicity and concurrency, are now stepping away due to ecosystem limitations, hiring challenges, and poor ergonomics in AI workflows. Batra echoed similar concerns in another blog post, explaining why he will never use Go for machine learning workloads. Why Developers are Moving On The recent departure of Taylor from Google, along with a wave of blog posts like Batra’s, has sparked questions about Go’s future. An increasing number of developers are turning to languages like Kotlin, Rust, and even emerging ones like Zig, which offer more flexibility or modern features. That said, it is important to note that not all of the criticism paints a complete picture. Go still has a dedicated fanbase, with supporters coming out in support of the language in comment sections. This clearly shows that the language is far from being dead. However, some common frustrations persist. Go’s verbose error handling, requiring explicit error checks, often leads to repetitive and brittle code. Some complain about the poor ergonomics for asynchronous programming. While goroutines are powerful, they can introduce race conditions and silent failures. Testing such conditions remains a challenge, as many developers highlighted in a HackerNews discussion. Go’s ‘no frameworks’ philosophy leaves developers spending cycles reinventing the wheel. Hiring is also a common issue. According to recruiters, the talent pool for Go is smaller than that of Java, Python, or JavaScript. Onboarding new developers can also take longer due to missing abstractions, making it not a language that developers are choosing to not learn. Companies can’t find enough experienced Go developers, and mid-career engineers hesitate to invest time in a language with limited job opportunities. On top of that, AI tools don’t generate idiomatic Go code well, making languages like Python even more attractive in an age of AI coding. Enter Zig, Kotlin, and Rust Go’s simplicity once made it attractive to Python and JavaScript developers. Yet, many are now moving to Zig, which promises clean syntax and blazing performance without Go’s baggage. Zig fans point out that the programming language compiles to native code, avoids garbage collection, and includes async\/await support. Tools like Bun, a JavaScript runtime, and TigerBeetle, a distributed financial database, are already showcasing Zig’s potential. Kotlin, on the other hand, wins with tooling. Its rich ecosystem—complete with powerful object-relational mapping (ORM), test suites, and Spring Boot integration—makes it easy to onboard and iterate. Rust continues to dominate in areas where performance and safety are non-negotiable, albeit with a steeper learning curve. Niche, Not Dead Despite the growing criticism, Go is thriving, particularly in the infrastructure and systems engineering space. Go remains the language of choice for building cloud-native tools, lightweight APIs and internal tooling at large organisations. Many engineers argue that Go was never meant to be a rapid product iteration language like JavaScript or Ruby—it was meant to be boring, stable, and scalable. “Go’s simplicity is its power, but also its curse,” one user noted on Hacker News. “It’s not trying to be Python. It’s trying to be reliable.” As Taylor aptly said, “I don’t think that any programming language is ever done—the programming environment changes all the time, and languages must evolve or die.”","excerpt":"“Go’s simplicity is its power, but also its curse.”","categories":["AI Features"],"tags":[],"author_name":"Mohit Pandey","publish_date":"2025-05-13T11:51:27","publication_year":"2025","word_count":940,"keywords":["Go","machine learning","Rust","AI","docker","microservices","Python","JavaScript","R","kubernetes"],"extracted_tech_keywords":["AI","machine learning","kubernetes","docker","microservices","Python","R","JavaScript","Go","Rust"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/why-developers-are-quietly-quitting-golang\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10172353,"title":"Knolskape Launches AI Academy to Foster AI Fluency in Firms","content":"Knolskape, a global experiential learning technology company, has launched AI Academy to build AI fluency across India’s workforce through AI-powered experiential and immersive role-based learning.. Knolskape’s AI Academy addresses the growing demand for AI skills through three structured learning paths: AI Ignite for individual contributors, AI Catalyst for mid-level managers, and AI Architect for senior leaders. The company’s proprietary 4E methodology powers each program—Evaluate, Educate, Experience, and Enable—which provides real-world, hands-on learning through immersive simulations and role-specific coaching. In an exclusive interaction with AIM, Kalyan Teja Maganti, co-founder and chief product officer at Knolskape, said that while 82% of businesses believe AI will transform their industry, only 59% are deploying it, and just 6% of employees feel confident using AI in their daily work. This highlights the “AI fluency crisis”, where employees lack the context and confidence to use accessible AI tools effectively, emphasising the need for both digital and human readiness in the workplace. The company has also collaborated with IT giant TCS, where many employees were currently on the bench undergoing training and not ready for client interviews, resulting in a significant workforce of tens of thousands. Knolskape trained approximately 7,000 to 8,000 individuals in the last two months as part of a transformation program. This included classroom sessions that covered three essential training components, while the fourth component was lacking. He added that the trainers provided valuable knowledge and developed a customised AI-enabled communication simulation using Genie, Knolskape’s proprietary simulation building platform “This allowed us to effectively address a key skills gap in just two months and execute the program without any [scaling] issues,” Maganti said. Organisations can select from ready-made modules, customised learning paths, or simulation licenses, all supported by real-time analytics. Knolskape, with a presence in over 75 countries and clients like BMW, HSBC, and Deloitte, is equipped to help Indian companies enhance their AI maturity. The AI Academy focuses on key areas such as AI tool usage, human-AI collaboration, ethical leadership, and strategic governance. Strong early demand is noted in cities like Bengaluru, Mumbai, and Delhi NCR, where AI hiring is rapidly increasing. The HR-tech startup has an instructional design team of content experts with extensive industry experience. They leverage human insights and AI technology to gather primary and secondary data from the industry. In developing customised solutions, the team follows a structured process that begins with online and offline activities, including interviews, surveys, and focus group discussions. Maganti added that this approach gathers valuable insights directly from organisations and industry stakeholders. Using this raw context, the instructional design team utilises generative AI content creation to enhance the platform, ensuring the content is well-informed and aligned with industry standards. Organisations are categorised into three layers in relation to AI: AI native companies are deeply involved in AI deployment but struggle with human capability. At the same time, AI-transforming organisations understand AI and can achieve outcomes, but lack workforce transformation at scale. Lastly, the emerging category includes sectors like banking that are just starting to adopt AI technologies. As per Maganti, this classification helps the company address their unique challenges in the AI journey.","excerpt":"Knolskape has launched a role-specific, simulation-based learning platform to build AI fluency across India.","categories":["AI News"],"tags":["AI Academy India","skills gap"],"author_name":"Smruthi Nadig","publish_date":"2025-06-25T12:00:00","publication_year":"2025","word_count":519,"keywords":["skills gap","Go","API","AI","Git","RAG","Aim","AI Academy India","generative AI","analytics","real-time analytics","R"],"extracted_tech_keywords":["AI","analytics","generative AI","Aim","RAG","R","Go","Git","API","real-time analytics"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/knolskape-launches-ai-academy-to-foster-ai-fluency-in-tech-firms\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10041576,"title":"How Did Airbnb Achieve Model Metric Consistency At Scale","content":"“As opposed to tables and columns, Airbnb’s platform concentrates on metrics and dimensions.” For the past couple of years, Airbnb has exemplified what a successful data-driven company should look like. It switched from Chronos to Apache Airflow to orchestrate workflow and invested in developing a collection of very crucial data tables, known as ‘core_data’. Airbnb analytics blossomed with `core data` serving as the cornerstone. Airbnb built and scaled its experimentation platform so that it could introduce A\/B testing into the company. Furthermore, a Data portal was established to help organise and record the company’s data as they prepared to launch Apache Superset is an open-source project. To evaluate and optimise and serve the internal users’ end-to-end needs, the company needed a comprehensive data platform. This need led to the construction of Minerva, Airbnb’s metric platform. Building on expertise in managing an experimental metric repository, the development team at Airbnb chose to employ six design principles for Minerva. The specifications for designing Minerva stated that it should be Scalable, Declarative, Consistent, Highly Available, and Well Tested. How was Minerva conceived (Source : Airbnb) As part of the company’s commitment to producing a single source of truth metric platform, Airbnb made significant investments in the development of Minerva, a platform that standardised the ways metrics are developed, computed, provided, and consumed. While ‘core data’ introduced multiple-step modifications to the data capabilities of Airbnb, it wasn’t done without substantial costs. By adopting the Core Data standard, users could discover which tables were needed, resulting in uniform table usage. But on the other side, it imposed a massive challenge on the centralised data engineering team because it required them to control and onboard an endless supply of fresh datasets, even when all the tables in an application were already in place. Such pipelines developed downstream of core data which had a significant impact on metric proliferation and divergence. Because of this, data scientists and engineers spend endless hours fixing errors. This is when the platform Minerva, created by Airbnb, came into effect. As opposed to tables and columns, Minerva concentrates on metrics and dimensions. By unifying the data from upstream data engineering teams and data from downstream consumption teams, the Minerva API allows enterprises to update core tables more flexibly while supporting many downstream customers. For example, necessary metadata must be submitted by the author when a metric is defined in Minerva. The configuration files all require information such as ownership, lineage and metric description. In the past, such metadata were present only in charts marked up as either internal corporate information or used in generic definitions by a variety of business intelligence (BI) tools. Every definition in Minerva handled a version-controlled piece of code. Advantages of Minerva Minerva is developed on top of open-source projects using Airflow, a workflow management framework, under the Apache licence. It is used for data pipeline-building, monitoring, and adjustment stages. All workflows on the platform are written in Python. It computes with Apache Hive and Apache Spark and consumes with Presto and Apache Druid. Minerva manages the entire life cycle, from conception to computation, serving, consumption, and final deprecation. It centralises critical business KPIs, dimensions, and other metadata in a centralised Github repository that is accessible to anybody within the company. Its development path ensures that the best data engineering standards such as code review, static validation, and test runs are followed. Its advanced computing pipeline features built-in checks to assure data quality which can automatically self-heal following job failures. Minerva effectively denormalized data by maximising the reuse of data and intermediate joined outcomes. It provides a uniform data API that allows for the on-demand delivery of aggregated and basic information. Minerva version control is also used to manage data definitions, data management, and data retention. Through Minerva’s API, Minerva data is made available to Airbnb’s proprietary R and Python clients. The availability of a notebook environment simplifies the data analysis process. The notebook environment produces and exposes the same data as Superset and Metric Explorer. What makes this data API noteworthy is that it allows the development of internal tools that can subsequently be made publicly available for the entire company. As COVID-19 froze global travel in March of 2020, Airbnb suffered considerably. In no time, reservations on Airbnb fell, and cancellations increased. Airbnb was able to drastically cut the time from data curation to insight finding through Minerva. An analysis of Minerva data also helped the company to anticipate and plan for the rapidly shifting market landscape confidently. Through this, they were able to change the shift in demand for local and long term stays.","excerpt":"As part of the company’s commitment to producing a single source of truth metric platform, Airbnb made major investments in the development of Minerva, a platform that standardised the ways metrics are developed, computed, provided, and consumed.","categories":["AI Features"],"tags":["Airbnb","covid-19"],"author_name":"Ritika Sagar","publish_date":"2021-06-10T11:00:00","publication_year":"2021","word_count":769,"keywords":["API","covid-19","AI","Apache Spark","Scala","Git","Python","data engineering","analytics","Airbnb","GitHub","R"],"extracted_tech_keywords":["AI","analytics","Apache Spark","Python","R","Scala","Git","GitHub","API","data engineering"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/how-did-airbnb-achieve-model-metric-consistency-at-scale\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":true,"is_research_content":true},{"article_id":10004539,"title":"Best Foreign Universities To Apply For Data Science Distance Learning Course Amid COVID","content":"While distance learning has always been a preferred choice for students to attain a degree from the best foreign universities, the COVID pandemic and the rising online learning has given it a whole new impetus. That being said, for data science professionals, having a foreign university degree can bring separate significance. Not only it provides an international perspective of working with people from across the world but also helps in differentiating the skills from others in the country. With online learning is gaining its traction amid this lockdown, the concept of distance learning has been diluted with many foreign universities providing online data science courses for learners across the world. Distance learning also provides the opportunity for students, amateurs as well as professionals to learn the complex subjects of data science at their own pace and also provide an opportunity to opt for more than one course at a time. So, if decided to pursue a degree course in data science from foreign universities, this article can provide a list of seven universities that offer distance data science courses. Also Read: How COVID-19 Enhanced The Scope Of Distance Learning For Data Scientists Master of Science in Business Analytics From: University Of Louisville About: Master of Science in Business Analytics by the UofL is an accelerated 13-month program which provides a team-based structured learning environment with diverse knowledge gaining possibilities. In this course, the learners can work both in teams or independently to complete the course. Not only this course provides a capstone project for better learning but also a competitive paid internship with elective opportunities. Starting with subjects like an introduction to linear algebra and data analytics, this course also covers marketing strategy, marketing research, consumer insights as well as an introduction to operations, supply chain and decision models. Additionally, with this course, the learners will graduate with advanced technical skills, a SAS analytics certificate, and eleven-months of related work experience. Application Deadline: Open Now Duration: 13-months Check out the course here. Fees: $30,000 bundled tuition and fees – regardless of residency Also Read: A Step-By-Step Guide To Studying Robotics In A Foreign University Post-Graduate Program in Data Science From: Amity Future Academy with eCornell, online vertical of Cornell University About: Postgraduate program in data science is a comprehensive course provided by Amity Future Academy in collaboration with eCornell, online vertical of Cornell University. This course is a 12 months degree course along with preparatory sessions, which will provide career guidance, along with 24×7 access to study material and video lectures. Moreover, the course comes with industry-driven comprehensive curriculum eight modules to follow with real-world projects and case studies to get hands-on. Also, it provides an opportunity to do a face-to-face meet up with industry experts as well as live interaction with potential employers. Starting from introduction to data science to data modelling and deep learning, it covers a vast range of topics for data scientists. Application Deadline: 14th August 2020 Duration: 12 months + preparatory sessions Check out the course here. Fees: ₹1,85,000 Also Read: 5 Skills That Universities Look For In A Candidate For Data Science Masters Online MSc in Business Analytics From: University of Bath Online About: MSc in Business Analytics is an online course from the University of Bath accredited by Bath School of Management. With access to the latest data insights, management principles and industry expertise, this course provides learners with relevant skills to lead a change in today’s data-driven industries. Led by experts of the sector, including specialists from IBM, this course encompasses a comprehensive industry-relevant curriculum which guarantees 85% of candidates to receive job offers within three months of graduating. The course covers subjects like business intelligence, databases, optimisation, business statistics, data mining, as well as machine learning. Application Deadline: Open Now Start Date: September 2020 Duration: 2 years, three months Check out the course here. Fees: £833* per 10 credit unit (10% alumni discount) Also Read: How This Indian Professor From University Of Buffalo Will ‘Teach’ Physics To AI MSc in Data Science From: University of London About: MSc in Data Science is a comprehensive online course from the University of London. Led by a faculty of industry experts, this course teaches how to apply data science for real-world problems and will provide in-depth knowledge of emerging technologies, statistical analysis and computational techniques. This course offers a flexible approach which fits around the learners’ working life, to keep a balance. MSc in data science covers a four-core module with subjects like mathematics and statistics for data science; machine learning; data programming in python; and big data analysis. Along with it, the course comes with two compulsory modules of data visualisation and data science research topics plus four optional modules to choose from. Application Deadline: 7th September 2020 Duration: 1-5 years, depending on module availability Check out the course here. Fees: £12000 + additional charges Also Read: Limitations Of Online Learning For Data Scientists Master of Data Science From: HSE University in collaboration with Coursera About: Masters of Data Science is a full online data science course provided by Higher School Of Economics, Russia, in collaboration with Coursera. The university has been recently ranked first place in 100 Best Russian Universities, according to Forbes. First fully English-taught online data science masters program, this course replicates the practitioner’s work environment with subjects like Python, Github, etc. Led by experts of Yandex and HSE, this degree urges students to complete applied industry-based projects designed by the university in collaboration with its industrial partners. The curriculum covers extensive topics relevant to analysts, data and computer scientists, software engineers, academic researchers, machine learning engineers. Application Deadline: Open Now Duration: 20 – 24 months Check out the course here. Fees: 577,500 – 1,155,000 RUB (Depending upon geographic location and currency) Also Read: Amid Surge In Online Courses, Here’s How To Choose The Right One Master’s Degree in Data Science From: The University of Texas at Austin About: Master’s Degree in Data Science is an online master’s program provided by the University of Texas in collaboration with edx. The course is specifically offered by the department of statistics and data sciences of the university. Led by experts of computer science and data science, this course’s curriculum has been designed to offer comprehensive knowledge in statistical theory and computer science application. The course covers subjects like simulation, data visualisation, data mining, data analysis, big data, along with topics of machine learning, algorithmic techniques, and optimisation to tackle real-world issues. Application Deadline: 15th September 2020 Duration: 1.5-3 years (Depends on the learners’ pace) Check out the course here. Fees: $10,000 Also Read: Top 8 Productivity Apps For Learners To Make The Most Of Online Courses Online MSc in Data Science From: Anglia Ruskin University, United Kingdom About: Online MSc in Data Science is a distance learning course provided by Anglia Ruskin University in the United Kingdom. This course not only advances your data science and ML career but also allows you to develop skills to apply in a wide range of industries. This course provides a hackathon-style boot camp that will help learners simulating in a real environment. Led by experts of the sector, this course has been co-designed and is co-delivered by Cambridge Spark, on an AI-powered learning platform. The course covers subjects like exploratory data analysis, machine learning technique, data engineering and big data, and deep learning application. Application Deadline: 10th August 2020 Duration: 2 years Check out the course here.Fees: £18,300","excerpt":"While distance learning has always been a preferred choice for students to attain a degree from the best foreign universities, the COVID pandemic and the rising online learning has given it a whole new impetus. That being said, for data science professionals, having a foreign university degree can bring separate significance. Not only it provides […]","categories":["AI Trends"],"tags":["Courses","Data Science","data science degree","latest technology in machine learning","MSc data science","msc data science and analytics","MSc. in data science","project topics for computer science"],"author_name":"Sejuti Das","publish_date":"2020-08-10T13:00:00","publication_year":"2020","word_count":1243,"keywords":["MSc. in data science","project topics for computer science","data science","Go","machine learning","data science degree","AI","ML","Git","Python","msc data science and analytics","deep learning","latest technology in machine learning","analytics","Courses","MSc data science","Data Science","R"],"extracted_tech_keywords":["AI","machine learning","ML","deep learning","data science","analytics","Python","R","Go","Git"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/best-foreign-universities-to-apply-for-data-science-distance-learning-course-amid-covid\/","complexity_score":4,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":true},{"article_id":463,"title":"How are Enterprises exploring the potential of the Space?","content":"The subject of space exploration has become a priority for many scientists and governments. The fact that we are a mere speck on the universe, pushes us even more to gain knowledge about things lying unknown in the vastness of the space. A lot of money is spent by the government in undertaking space explorations, in the hope of finding something new and useful, or even alien life forms. The present decade has seen many private firms and startups emerge with innovative offerings and projects dedicated to the space. The privatization of this industry is has upped the game for ‘space quest’, and this only implicates that the day is not far when me and you will probably be stepping on the lunar surface, or the Mars for that case. These private firms are making fruitful strides into space, which compels me to present you with some of the most interesting ways enterprises are exploring the potential of space. Launching satellites for commercial application SpaceX- leading the private players in Space Elon Musk founded space organization, SpaceX makes its major bulk of money by engaging in just one activity, launching commercial satellites into space. The agency undertakes both commercial and government project, charging $72 million on an average per launch. For government, SpaceX undertakes complex space missions such as resupplying the International Space Station, or placing a science satellite into orbit for National Oceanic and Atmospheric Administration (NOAA). Co-founded by Narayan Prasad, Dhruva Space is yet another example of an Indian startup created in 2012, with a vision to lead privatization of satellite industry in India. The company is working on developing an indigenous small satellite platform. However, the firm stresses its primary focus on assembly, integration, testing, and operation of satellites. Promoting space tourism Rocket-powered test flights have been conducted for SpaceShip Two, which is intended to take passengers to the brink of space. The spaceship has been designed by Virgin Galactic, and despite blown deadlines to the initiation of space tourism, the firm is pushing hard with its effort of making the dream come true. Turning asteroid mining into a reality Asteroid Mining Planetary Resources had stated its unique goal of mining asteroids within proximity using robotic spacecraft. The startup had earlier used Kickstarter to fund its own project involving successful launch of low-cost, publicly accessible satellite telescopes into orbit, back in 2014. The private firm is supported by director James Cameron and Google’s Larry Page and Eric Schmidt. Geospatial Analysis Earth2Orbit’s unique value proposition lies in its GeoIntel Analytics practice. Composite data sets, sophisticated algorithms, and advances in Deep Learning and AI are utilized in the process to deliver “actionable intelligence.” This helps private and government organizations alike, to make prudent decisions. E2O’s GeoIntel practice today, has transcended across various verticals climate change, agriculture, shipping, oil and gas, mining, retail, and smart city projects. Another interesting story is about Orbital Insight’s geoanalytics platform. The firm analyzes millions of satellite images at a time to equip industry leaders with advanced and unbiased knowledge of socio-economic trends. Earth Imagery Earth Imagery Planet.com uniquely leverages space to better life on earth. The firm offers its planet monitoring and planet basement services across a plethora of industries such as agriculture, civil government, defense and intelligence, energy and infrastructure, and forestry. Planet strives in the “space race” with a mission to image the entire Earth every day. This helps their clients perceive and understand global changes, while acquiring actionable insights. Innovative space-based manufacturing Made in Space brings something new and innovative to the table with its idea of taking 3-D printing into the orbit. The disruptive idea has been endorsed by NASA, and the startup is working on a project to build a 3-D printer for the International Space Station. The potential commercial applications of this futuristic idea is several-fold, and it will soon be a reality.","excerpt":"The subject of space exploration has become a priority for many scientists and governments. The fact that we are a mere speck on the universe, pushes us even more to gain knowledge about things lying unknown in the vastness of the space. A lot of money is spent by the government in undertaking space explorations, […]","categories":["IT Services"],"tags":["NASA"],"author_name":"Amit Paul Chowdhury","publish_date":"2017-02-02T06:20:40","publication_year":"2017","word_count":645,"keywords":["Go","NASA","AI","RAG","Ray","deep learning","ViT","analytics","GAN","R","analytics platform"],"extracted_tech_keywords":["AI","deep learning","analytics","Ray","RAG","R","Go","GAN","ViT","analytics platform"],"url":"https:\/\/analyticsindiamag.com\/it-services\/enterprises-exploring-potential-space\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":4,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":60202,"title":"Cognizant To Pay Additional 25 Percent In Base Pay To Indian Employees","content":"American tech giant, Cognizant announced that the company will pay an additional 25 percent of the base pay to the employees in India and the Philippines for the month of April due to the current pandemic. Currently, the tech giant has around 2,00,000 employees in India. According to sources, the additional pay will be paid to the employees up to the Associate level which will cover more than two-thirds of the workforce in India. CEO of Cognizant, Brian Humphries stated, “Our global company has a long history of operating virtually and remotely, and rallying and responding around the clock, wherever needed. As you’ve seen, India’s Prime Minister Mr. Narendra Modi has imposed a nationwide 21-day lockdown in an attempt to contain the spread of the coronavirus. The Philippines is under a state of national emergency. We support these and other efforts by governments across the world to “flatten the curve” and contain the spread of COVID-19.” According to Humphries, the tech giant has already enabled safety measures like working from home for the majority of its offshore teams to ensure employee health and safety, and to maintain continuity of service for its clients, while safeguarding their data and protecting access to their systems. The work from home measures includes provisioning new laptops and encrypting desktops and moving them to associate homes, enabling the use of BYOD (Bring Your Own Device), providing additional bandwidth connectivity and air cards — all with the appropriate client permissions and security protocols.","excerpt":"American tech giant, Cognizant announced that the company will pay an additional 25 percent of the base pay to the employees in India and the Philippines for the month of April due to the current pandemic.  Currently, the tech giant has around 2,00,000 employees in India. According to sources, the additional pay will be paid […]","categories":["Global Tech"],"tags":["Cognizant"],"author_name":"Ambika Choudhury","publish_date":"2020-03-27T15:03:01","publication_year":"2020","word_count":247,"keywords":["Go","programming_languages:R","AI","Cognizant","programming_languages:Go","ViT","R"],"extracted_tech_keywords":["AI","R","Go","ViT","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/global-tech\/cognizant-to-pay-additional-25-percent-in-base-pay-to-indian-employees\/","complexity_score":2,"technical_depth":6,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":56475,"title":"Top 7 AI-Powered Apps To Help You Get Fit","content":"As the world grows more fitness-conscious, the demand for tech-based solutions to satiate the appetite of this burgeoning population is sprouting as well. This has given rise to a slew of startups which are leveraging predictive analytics, artificial intelligence and other deep learning technologies to develop apps that will help its users get back in shape. Whether it measures calorie intake, making food recommendations, showing workouts, or an alarm to inform if you are off your fitness course — these features have been automated through AI by copying human intelligence. Here is a list of apps that will help you meet your fitness goals and stick to your new year’s resolutions: Aaptiv This audio-based app offers unlimited access to thousands of workouts, all led by top-level trainers. It adds as many as 30 new classes each week across 12 popular categories, including strength training, outdoor running and yoga. The app also took a significant-tech turn last year with its new service called ‘Aaptiv Coach’. This AI-based assistant builds personalised programs tailored to users’ fitness goals. It takes into account their current fitness levels and eating habits to develop a profile of the person based on which recommendations are shared. This includes multi-week plans for weight loss, flexibility, race training, and more. Cure.fit The abundance of data has allowed this company to leverage technology not just to ensure timely execution of its services but also achieving customer satisfaction. Cure.fit has an entire health ecosystem built into it which caters to overall fitness, nutrition as well as mental well-being. And AI has become the key for the company to work on these solutions. This includes scheduling classes, programming food menu and predicting order, allocating trainers and managing deliveries with all these processes driven by machine learning. Furthermore, the company extensively relies on customer feedback to filter out unpopular routines. GOQii Without limiting the platform to dole out fitness recommendations merely, this app has covered some ground in highlighting lifestyle conditions as well, including diabetes and hypertension. Its ecosystem of wearables can currently track blood pressure and heart rate. The app mines this data and uses AI to make predictions about a possible fatal health situation that should serve as a warning for people. The app, which has evolved into an all-inclusive healthcare platform, has several other features that could be a draw for many. This includes live and interactive sessions with fitness coaches, at-home lab test facilities, a health store, unlimited cloud storage to organise health records, and the opportunity to donate to a cause by merely walking. FitnessAI This app uses artificial intelligence to create personalised weight lifting plans for its users. How does it work? The app has tools built into its system that optimises sets, reps and weight for each exercise, each time. The algorithm tells users how long they should rest for and makes recommendations around how to adjust the difficulty level for subsequent times. The app also gives clear insights by tracking your progress over time. Strava Considered the most popular social network for athletes, Strava is your go-to app if you are an avid runner or cycle enthusiast. It provides comprehensive performance tracking and stores with a wealth of information around your activity. The app also got into AI athlete personalisation recently with its membership service Strava Summit. Offered as three packs — Training, Analysis and Safety, it is a step by the company to redirect its focus on building tools that will leverage artificial intelligence and machine learning to enhance personalisation and build rich athlete experiences. Some of the features include race analysis, power meter data, goal-setting and training plans. https:\/\/www.youtube.com\/watch?v=JAMLGx4jUoE HealthifyMe A leading health and fitness app in India, HealthifyMe’s artificial intelligence-based virtual assistant ‘Ria’ attends to its users’ queries around fitness and nutrition, in as many as ten languages. The bot also makes dietary recommendations after factoring in the users’ data and other health parameters. Being compatible with popular fitness wearables, the app uses data mined from these devices to build an algorithm that can easily be accessible by its users. The company made a deliberate push towards AI-powered services last year. Until then, it was popularly used to log and track calorie intake, get access to healthy recipes and tips for overall well being. Asana Rebel This Berlin-based yoga startup has attempted to recreate yoga by combining the discipline’s traditional practices – like Vinyasa and Ashtanga – with other forms of higher intensity exercises and fitness routines like burpees, planks, and mountain climbers. Compatible with Apple Watch, it makes it easy for users to personalise their fitness journey and log their progress. The app also has a slew of other features, including meditation sessions as well as quizzes to keep the mind sharp and healthy.","excerpt":"As the world grows more fitness-conscious, the demand for tech-based solutions to satiate the appetite of this burgeoning population is sprouting as well. This has given rise to a slew of startups which are leveraging predictive analytics, artificial intelligence and other deep learning technologies to develop apps that will help its users get back in […]","categories":["AI Trends"],"tags":["AI (Artificial Intelligence)","Deep Learning","fitness","meditation","Predictive AI","predictive analytics"],"author_name":"Anu Thomas","publish_date":"2020-02-11T19:00:00","publication_year":"2020","word_count":793,"keywords":["meditation","Go","artificial intelligence","machine learning","fitness","AI","R","ML","RAG","deep learning","analytics","Deep Learning","AI (Artificial Intelligence)","predictive analytics","Predictive AI"],"extracted_tech_keywords":["AI","artificial intelligence","machine learning","ML","deep learning","analytics","RAG","predictive analytics","R","Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-ai-powered-apps-to-help-you-get-fit-in-2020\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":1,"has_code_examples":false,"has_tutorial_content":true,"is_research_content":false},{"article_id":10122921,"title":"KissanAI accepted for NVIDIA Inception Program","content":"KissanAI, a pioneering startup in agricultural artificial intelligence, has been accepted into the prestigious NVIDIA Inception program. The program supports startups revolutionizing industries with advancements in AI and data science. As the first company to introduce a vernacular voice-based copilot application for farmers, KissanAI has been at the forefront of transforming agriculture through its Generative AI platform, AgriCopilot, and domain-specific Agri LLMs called Dhenu. By joining NVIDIA Inception, KissanAI gains access to cutting-edge technology, expert mentorship, and collaboration opportunities to further enhance its intelligent farming solutions. “We are thrilled to be part of NVIDIA Inception and leverage their resources to make our AI capabilities faster, more accurate, and impactful for farmers,” said Pratik Desai, CEO of KissanAI. “This partnership will accelerate our mission to drive agricultural productivity and sustainability.” Through the program, KissanAI will have early access to NVIDIA’s latest hardware and software innovations. The company will also benefit from NVIDIA’s global network of AI experts and marketing support to expand its reach. Founded in 2020, KissanAI’s AgriCopilot platform deploys AI workflows for use cases such as advanced crop advisory, conversational commerce, and sales assistance. Its Dhenu LLMs leverage a curated agricultural knowledge base to provide nuanced insights to farmers. The startup has already outpaced industry giants like Google and OpenAI in the agri-tech space. The acceptance into NVIDIA Inception marks a significant milestone for KissanAI as it continues to push the boundaries of AI in agriculture. The company remains dedicated to developing innovative solutions that meet the evolving needs of farmers worldwide.","excerpt":"KissanAI, the first to launch a voice-based AI copilot for farmers, is transforming agriculture with its set of domain-specific Agri LLM.","categories":["AI News"],"tags":["kissanai"],"author_name":"Shyam Nandan Upadhyay","publish_date":"2024-06-09T01:12:36","publication_year":"2024","word_count":253,"keywords":["kissanai","data science","Go","artificial intelligence","OpenAI","AI","innovation","RAG","ViT","generative AI","R"],"extracted_tech_keywords":["AI","artificial intelligence","data science","generative AI","OpenAI","RAG","R","Go","ViT","innovation"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/kissanai-accepted-for-nvidia-inception-program\/","complexity_score":2,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101832,"title":"Israel’s GenAI-Backed End Game","content":"While everyone now knows about Israel’s Iron Dome and its military prowess, little did we know about its secret weapon – the “Cyber Dome.” In 1955, David Ben-Gurion, Israel’s founding father, foresaw a destiny unfolding in the expanse of the Negev Desert, near the metropolis of Tel Aviv. Little did he know that his prophecy would later become a 1,800-acre cyber security complex in Beer Sheva, Negev’s capital and Israel’s cybersecurity hub. Israel is a revolting force in modern warfare. As the world watches the traditional battle evolve into a digital theatre, Israel has marshalled its finest minds to build the ‘Cyber Dome,’ designed to wage warfare in cyberspace. The dome is inspired from its Iron Dome air defence system, which was deployed in 2011. Gaby Portnoy, the  Director General of the Israel National Cyber Directorate (INCD) calls it the “secret sauce” noting the X factor is generative AI platforms. The development of this tech dome is fueling the blood bath between Gaza and Israel. The stakes are high for Israel as the world is watching the concept of warfare undergo a shift. Every three years, a new generation joins the cyber defence once they complete the military in Israel. Hence, the space between man and machine is equally divided in these war rooms, noted The Week’s Namrata Biji Ahuja who recently visited the facility in the middle of camel land. One half is nerdy youth monitoring real-time data on multiple screens. The other half is completely digital, running on AI\/ML — gathering, reading and interpreting data. Intelligence services use platforms akin to ChatGPT to filter important threats from the unlimited content flowing into the systems. These AI enabled military programmes are run by nearly 14,000 tech personnel at the Beer-Shava complex, she further stated. AI in the Warzone The Israel Defence was one of the world’s first to use AI to thwart threats. In 2021, the Middle Eastern country used AI broadly across an operation to identify targets. The technology was used to determine rocket trajectories based on radar information, intercepting those headed for densely populated areas. With experience on the warfield, the state now confronts a new challenge – reigning in AI. On the current global AI leaderboard, Israel claims the fourth position, with a projected market size of $1.54 billion in 2023. Since April 2023, generative AI-related startups in Israel have more than doubled from 67 to 144. The total amount raised by these startups so far has reached $2.3 billion, a 2.5x increase compared to the total figure from six months ago. The largest round so far in the country’s generative AI space belongs to natural language company, AI21 Labs, which raised a $155 million Series C. The country is also home to Runway ML which was one of the first editing tools companies to integrate machine learning in 2018. In a conversation with the tech industry’s wannabe saviour Elon Musk, the Israeli PM had mentioned about pushing forth the country to third rank globally in AI soon. Other government personalities have also aimed to parlayed Israel’s technological prowess to become an AI “superpower”. (Source: Statista.com) The defence forces have also relocated intelligence groups, including the 8200 Intelligence Unit, to Beersheva’s high-tech campus adjacent to a university that was arid fields of dirt and sand a few years ago. Despite the other developments, the digital rendition of the Iron Dome stands out.  Four months ago, Ronen Bar, the head of the Israeli security agency Shin Bet spoke about Tel Aviv developing the AI-powered “global cyber iron dome” system with a number of countries to identify and tackle threats. During an annual cyber conference, Bar declared that Israel had come to terms with the fact that “winning this war with sticks and stones” was simply not feasible.","excerpt":"The space between man and machine is equally divided in these war rooms noted The Week.","categories":["AI Highlights"],"tags":[],"author_name":"Tasmia Ansari","publish_date":"2023-10-20T18:30:00","publication_year":"2023","word_count":630,"keywords":["Go","ChatGPT","API","machine learning","AI","ML","Git","Aim","generative AI","R"],"extracted_tech_keywords":["AI","machine learning","ML","generative AI","ChatGPT","Aim","R","Go","Git","API"],"url":"https:\/\/analyticsindiamag.com\/ai-highlights\/israels-genai-backed-end-game\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10101188,"title":"Reinforcement Learning Craves Less Human, More AI","content":"Human feedback has proven essential in training machine learning algorithms. Back in 2019, the director of research at IBM emphasised that while reinforcement learning holds promise, relying solely on this approach can be challenging due to its heavy dependence on trial and error. Apart from being Google DeepMind’s most preferred technique to train its popular algorithm like AlphaGo and its latest iteration AlphaStar, reinforcement learning from human feedback (RLHF) has been particularly effective in aligning large language models (LLMs) with human preferences. Furthermore, a majority of the large language models have been trained using the reward approach including OpenAI’s ChatGPT. Even Meta AI’s framework PyTorch recently upgraded its RLHF elements for developers to easily build an RLHF training loop with limited RL knowledge. However, a prime hurdle lies in gathering high-quality human preference labels. This is where reinforcement learning from human feedback with AI feedback (RLAIF) comes into the picture, a novel framework by Google Research to train models with reduced reliance on human intervention. Researchers found that both exhibit similar performance, with RLHF having a slight edge, albeit not significantly so. To Human or Not One of the areas where ChatGPT has not been able to thrive yet is summarising documents. Since the chatbot debuted late last year, researchers have been exploring new methods for generating concise summaries via its more potent GPT-4-based paid version. Interestingly, in the latest Google study it was discovered that human evaluators preferred RLAIF 71% of the time and RLHF 73% of the time. However, in a direct comparison, the two were equally favoured, with a 50% win rate. Furthermore, summaries from both the methods were preferred over human-written reference summaries, with RLAIF at 79% preference and RLHF at 80%. Meanwhile, one important factor to consider is that both the recipes tend to produce longer summaries than the supervised fine-tuning (SFT) policy. This could contribute to the perceived quality improvements. The researchers conducted further analysis, which revealed that even after adjusting for length, both approaches, with and without AI feedback, outperformed the SFT policy by a similar margin. While the community continues to grapple with the summarising issues, RLAIF appears to be a viable alternative to RLHF without the need of human annotation. However, the researchers acknowledge the need for conducting further experiments on a broader spectrum of natural language processing (NLP) tasks to validate these findings, a path they intend to explore in their future research. Reinforced in Secret Just a few weeks ago, Google DeepMind proposed another new algorithm called reinforced self-training (ReST) for language modelling. It follows a similar process of removing humans from the loop by letting language models build their own policy with a single initial command. While ReST finds application in various generative learning layouts, its expertise lies in machine translation. In comparing, for ReST (Reward Estimation from Suboptimal Trajectories) with Online RL, a tactic frequently used in RLHF, the results indicate that the latter performs at par with the former when only one “Grow” step is used. However, when ReST incorporates multiple “Improve” steps, it highly surpasses Online RL in terms of higher rewards. Additionally, the study observed that Online RL exhibited an 8-point drop in BLEU score on the validation set, hinting at potential issues related to reward manipulation. In contrast, ReST demonstrated an ability to improve reward model scores without adversely affecting other performance metrics, suggesting it may impose a lower “alignment tax” compared to Online RL methods. These research developments have come in the wake of Project Gemini, which is ready to unseat ChatGPT as the premier generative AI tool, as per DeepMind boss Demis Hassabis. The model anticipated for almost a year now is reportedly being trained via the company’s pioneering technique — reinforcement learning. Even though much official details have reached the media yet, reinforcement learning with AI is expected to play a huge role in the training process. With the recent studies pointing towards the company’s interest in incorporating AI with reinforcement learning, we can’t wait to see what’s cooking in the research lab.","excerpt":"Google Research has proposed RLHF with AI feedback to train models better","categories":["AI Features"],"tags":["RLHF"],"author_name":"Tasmia Ansari","publish_date":"2023-10-06T12:00:00","publication_year":"2023","word_count":673,"keywords":["ChatGPT","Meta AI","machine learning","RLHF","OpenAI","AI","PyTorch","NLP","generative AI","R"],"extracted_tech_keywords":["AI","machine learning","NLP","generative AI","ChatGPT","OpenAI","Meta AI","PyTorch","RLHF","R"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/reinforcement-learning-craves-for-less-human-more-ai\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true},{"article_id":10120903,"title":"Indian Companies are Good at Copying Ideas Generated Elsewhere","content":"Infosys co-founder Narayana Murthy, who has been quite vocal about the need for the youth of the country to work 70-hours a week, recently said that Indians are good at applying ideas generated elsewhere for the betterment of the nation. He also added that it would take time for the country to invent new things. He may be right. The most recent Ola’s Krutrim model is just a replica of OpenAI’s model, simply moulded for Indian use cases. This is similar to Flipkart copying Amazon, Ola copying Uber, and Zomato and Swiggy copying DoorDash. Meanwhile, there are other quick commerce companies such as Zepto and Dunzo, which are completely new concepts and are thriving in India. Recently, Kunal Shah, the founder of CRED, posted on X saying that early stage startups should be easy to iterate and late stage startups should be hard to distract. This also highlights the mentality of Indian startups that are not iterating and not innovative as well. “Startups should be easy to unlearn,” replied Vivian. The sentiment that India has been lagging when it comes to building technology and leveraging AI has grown stronger. “The reality of India is that at this point of time, by and large, we have upgraded ourselves to the orbit of applying ideas and concepts that are invented outside India and do some innovation and become experts,” Murthy said, while also adding that it is the first step towards catching up with the West. Murthy is positive that India will be able to invent new things, but it would take time, and the youth has to be enthusiastic about it. “Because my own belief is that a youngster of today is at least 10-20 times smarter than what I was at their age.” A lot practical though Though this may sound like a negative take on India’s innovation, applying global ideas locally is a smart and practical approach. “It’s a strategy that accelerates innovation and progress, leveraging collective human knowledge to solve problems faster,” said a user on X. One of the reasons that India’s AI is a little bleak is that there is a lack of investment when compared to the West, along with a lack of industry and academia partnership. If this happens, it would create a pathway for more research and development within the country, which still seems to be lacking. A similar conversation recently occurred when experts from the Indian tech community said that there was nothing foundational being built out of India. Most of the research was just copy and paste from the West. “Who has challenged the original algorithm? While Transformers are a great piece, they have flaws in terms of compute and carbon,” said Nikhil Malhotra, global head-Makers Lab, Tech Mahindra, reiterating that most of the research in India is done on fine-tuned models. “Training something from scratch and turning it into the 10th best foundation model that no one will use in production is the wealth only a few companies with deep pockets can afford… even spending millions on failed training runs,” said Pratik Desai, adding that India has so many unique use cases that don’t need foundational model research and using models such as Phi, Orca, or Llama is enough. “India has never led any fundamental research, but we have a golden opportunity as AI can be a levelling field,” added Desai. “However, this requires a fundamental shift from coaching and academia to a change in mindset from parents, and founders to investors.” On the other hand, “If we don’t work on our own AI infrastructure, in the next 5-10 years, like we import oil, we will have to import AI,” said Gaurav Aggarwal, who is currently leading an AI initiative at Jio. He added that it pains him to see that India is not producing AI experts, just “slightly glorified engineers” who have no clue about what they are building. Similar thoughts were recently shared by Dharmesh BA, founder of a stealth startup. He said that though a lot of people are building AI, it is most likely a wrapper of GPT. “We live in an era where it’s easy to build but difficult to figure out what to build,” he added. What is the problem? The possible reasons behind India not inventing new things, as pointed out by several users on X, is the lack of risk-taking attitude, and our over-reliance on the traditional system of education. Meanwhile, some say that there are merits in being a second-mover, but India definitely needs to up its game. There have been several theories floating around the internet claiming that India has arrived late to the AI party. Even with the alleged late arrival, most of the AI development in India is focused around building AI  use cases by adopting AI models developed by the western countries, rather than the core technology. Sourav Das, researcher at IIIT Kalyani, had said the same thing earlier. “How many of them have made an algorithm, theory, or model from scratch,” he questioned, saying that everything is available on the internet and the researchers are just exploiting the resources. “There is no invention in India, just reusing the things that are already there,” he said, adding that all the fine-tuning is just getting “honourable mentions”. A lot of the current AI development is being driven by young developers building AI models on top of existing ones such as LLaMA and Mistral, but nothing concrete has come up yet. Though there are initiatives such as Ola’s Krutrim, Sarvam AI, Tech Mahindra’s Project Indus, and BharatGPT that are focused on building models from scratch, a lot of work still needs to be done.","excerpt":"The sentiment that India has been lagging when it comes to building technology and leveraging AI has grown stronger.","categories":["IT Services"],"tags":["Startups"],"author_name":"Mohit Pandey","publish_date":"2024-05-17T15:57:52","publication_year":"2024","word_count":945,"keywords":["Go","OpenAI","AI","AWS","Transformers","RAG","GPT","Ray","Aim","Startups","R"],"extracted_tech_keywords":["AI","OpenAI","Aim","Ray","Transformers","RAG","AWS","R","Go","GPT"],"url":"https:\/\/analyticsindiamag.com\/it-services\/indian-companies-are-good-at-copying-ideas-generated-elsewhere\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10094889,"title":"Top 7 VR Headsets For You","content":"At Apple WWDC conference 2023, the ios maker made a suite of announcements including the grand unveiling of its first ever mixed reality headset called Vision Pro. Set to be launched in early 2024, the exorbitantly priced headset, valued at $3,499, is the most expensive option for now, rendering it unaffordable for the majority of individuals So we have curated a list of alternatives that offer comparable functionalities while remaining within reasonable fiscal boundaries. Read more: Can Apple Save Meta? Oculus Meta Quest 3 Just when Apple introduced Vision Pro, Mark Zuckerberg, founder and CEO of Meta, also announced the release of its VR headset Quest 3 in the coming fall. Priced at $499, which is $100 higher than the previous model, Quest 2. Although the official prices for the UK and Australia haven’t been announced, Quest 3 boasts enhanced VR technology, using a next-generation Qualcomm Snapdragon chip for improved performance and graphics, offering double the graphical capabilities of its predecessor. It is expected to have a base storage capacity of 128GB, potentially with larger options available. The design has been revamped to be 40% slimmer and more comfortable, featuring an improved head strap and adjustable inter-pupillary distance (IPD). Additionally, the Touch Plus controllers have been ergonomically improved and may incorporate TruTouch haptic devices. Pimax 5k Super Released two years ago, the Pimax 5K Super stands as a premium VR headset that boasts an expansive field of view, superb resolution, and a speedy refresh rate at $725, making it pretty expensive. Users have given the 5K Super positive feedback, commending its immersive encounters and crisp visuals. This headset is also compatible with a broad array of games and applications. The Pimax 5K Super showcases several noteworthy specifications, including a resolution of 2560×1440 per eye, a field of view spanning 200 degrees, and a refresh rate of up to 180 Hz. It leverages SteamVR 1.0\/2.0 tracking and weighs a total of 470 grams. Additionally, it features a comfortable and adjustable design. However, it comes with a high price tag, lacks portability compared to some other VR headsets, can be challenging to set up, and demands a powerful PC to run games at high settings. HP Reverb G2 At $399, HP Reverb G2 aims to combine the excellent image clarity of its predecessor G1 with the comfort and audio capabilities found in the Valve Index. It has a high-resolution display, an impressive 2,160 x 2,160 pixels per eye and enhanced image quality when compared to its previous model. This results in a sharp and highly detailed visual experience with minimal distortions. Nevertheless, the field of view is narrower compared to VR headsets such as the Valve Index, and the headset’s comfort and fit may vary depending on the user’s head shape. Moreover, the Reverb G2 includes upgraded VR controllers and tracking, offering improved ergonomics and functionality. Although the headset has some limitations in terms of field of view and tracking accuracy, it delivers an overall good VR experience of remarkable quality according to users. HTC Vive XR Elite Unveiled at CES 2023, the HTC Vive XR Elite is a premium headset priced at $1,099 and offers versatility, a compact design, and weighs 625g (with battery). The Elite features a 4K display, 110-degree field of view, and 90 Hz refresh rate. It includes four cameras and a depth sensor for mixed reality and precise tracking. The XR Elite sets a new standard in portability and design, making it a compelling choice for extended VR sessions. HTC Vive Pro 2 The HTC Vive Pro 2 is compatible with the SteamVR platform and shares the same compatibility as the Valve Index. With a resolution of 2,440 by 2,440 pixels per eye and a refresh rate of 120 Hz, it delivers exceptional image quality. However, it comes at a steep price of $799 for the headset alone and requires additional purchases of base stations and controllers. Furthermore, it lacks essential base stations and controllers, which must be bought separately. Although it offers excellent visuals, its high cost and additional expenses make it less appealing compared to alternatives like the Valve Index or Oculus Quest 2. Valve Index The Valve Index is a high-priced VR system for PC users. It consists of a headset, controllers, and base stations. The headset has refresh rate of 120Hz, but it lacks distinctive features compared to its competitors. However, the controllers are its USP, offering individual finger tracking and a more immersive experience. The Valve Index comes at $999 and is recommended for avid VR enthusiasts and PC gamers, whereas the more affordable Oculus Quest 2 serves as a good entry-level alternative. Setting up the Index is easy, and it requires relatively modest hardware. PlayStation VR2 Known for easy setup and comfortable design, PlayStation VR2 has impressed users with stunning visuals and reliable tracking, enhancing the overall sense of presence. However, at $479, its cost can be prohibitive for those without a PS5 console, and the game library is somewhat limited. The controllers and headphones possess unique features but also exhibit some quirks. Nevertheless, the PSVR2 boasts an elegant and user-friendly design, featuring a lightweight feel and a single cable for ease of use. Setting it up is a breeze, and it offers solid performance. While it offers a curated collection of high-quality games, it lacks compatibility with the original PSVR library. The PSVR2’s price is relatively steep, particularly when factoring in the additional expense of a PS5 console. All in all, it stands as a compelling choice for mid-tier VR enthusiasts who already own a PS5. Read more: Apple is Turning Everyone into Digital Zombies","excerpt":"Apple’s exorbitantly priced Vision Pro is not everyone’s cup of tea, so we have found substitutes with similar features for you.","categories":["AI Trends"],"tags":["Top Trend"],"author_name":"Shritama Saha","publish_date":"2023-06-11T10:01:00","publication_year":"2023","word_count":935,"keywords":["Top Trend","Go","programming_languages:R","AI","programming_languages:Go","Git","RAG","Ray","Aim","GAN","R"],"extracted_tech_keywords":["AI","Aim","Ray","RAG","R","Go","Git","GAN","programming_languages:R","programming_languages:Go"],"url":"https:\/\/analyticsindiamag.com\/ai-trends\/top-7-vr-headsets-for-you\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":0,"has_code_examples":true,"has_tutorial_content":false,"is_research_content":false},{"article_id":10169816,"title":"Happiest Minds Q4 Profit Drops 53% Despite Soaring Revenue, Reports 2% Revenue from GenAI","content":"Happiest Minds reported a 30.5% year-on-year jump in revenue to ₹544.57 crore in the fourth quarter of FY25, powered by strong deal momentum in verticals like healthcare and BFSI and growing investments in GenAI. Despite the robust revenue growth, net profit for the quarter fell 52.76% to ₹34 crore due to a one-time bad debt and continued investments in growth areas. For the full fiscal year, revenues climbed 26.85% to ₹2,060.84 crore, while net profit declined 25.66% to ₹184.66 crore compared to the previous year. Ashok Soota, executive chairman, said, “The ten strategic transformational changes that we rolled out are shaping Happiest Minds’ future. Our strategic initiatives, along with the continued commitment of our teams, have us well-positioned for strong double-digit organic growth in FY26 and beyond.” The company’s GenAI business unit and its newly established new logo (NN) hunting team have begun contributing to the pipeline, reflecting management’s confidence in long-term revenue acceleration. In the investors release, the company reported 2.1% revenue from its newly formed GenAI Business Unit. “Our move to a vertical structure has resulted in accelerated growth in several verticals like Healthcare and BFSI,” said Joseph Anantharaju, co-chairman & CEO. During the earnings call, Anantharaju said that the revenues that Happiest Minds is getting from GenAI are industry-leading. “Could we do more? Absolutely. And that’s a goal that we’ve taken for ourselves,” he said. He explained that there are three distinct sources of revenue. The first is direct revenue—which the firm has been quite strict about and which we’ve been sharing. Then there is other AI and data elements, and that is what the firm’s Analytics and AI CoE provides. “We continue to see an increase in the share of the Healthcare vertical, which saw large new deals totalling $20 Mn from 4 customers. The transformations around GenAI and the NN hunting team have seen a good buildup in the pipeline that should result in revenue growth.” EBITDA for the full year stood at ₹462.24 crore, representing a margin of 21.4%, in line with guidance. “Adjusted for a one-time bad debt and continued investments in Gen AI and Sales teams, operating margin and EBITDA continue to be industry leading,” said Venkatraman Narayanan, MD & CFO. During Q4, the company added 14 new clients, bringing the total to 281. The quarter also saw significant project wins, including an AI-powered end-user chat platform for a global market research agency, connected product development for a US manufacturer, and risk consulting for a Middle Eastern bank. Attrition stood at 16.6% on a trailing 12-month basis, and employee count reached 6,632.","excerpt":"Happiest Minds continues to see an increase in the share of the Healthcare vertical, which saw large new deals totalling $20 million from 4 customers.","categories":["AI News"],"tags":["happiest minds","Quarterly Earnings"],"author_name":"Mohit Pandey","publish_date":"2025-05-13T09:39:34","publication_year":"2025","word_count":430,"keywords":["Go","API","GenAI","programming_languages:R","AI","happiest minds","R","Git","Ray","analytics","GAN","Quarterly Earnings"],"extracted_tech_keywords":["AI","analytics","GenAI","Ray","R","Go","Git","API","GAN","programming_languages:R"],"url":"https:\/\/analyticsindiamag.com\/ai-news-updates\/happiest-minds-q4-profit-drops-53-despite-soaring-revenue\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":3,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":false},{"article_id":10121982,"title":"Yann LeCun Puts a Brake on Musk&#8217;s AGI Timeline","content":"xAI chief Elon Musk and Meta’s chief AI scientist  Yann LeCun, recently engaged in friendly banter on X, and it looks like there is no end in sight. And LeCun is clearly winning the argument, or more accurately, is successfully derailing Musk’s plans to build a new supercomputer with 100,000 NVIDIA GPUs and achieve AGI next year. Interestingly, the drama is unfolding on X rather than on Meta’s Threads, which was launched about a year ago. Meta and xAI are competitors, with both social media platforms building open-source LLMs that can provide real-time information. Not sure if this is a deliberate move by Mark Zuckerburg or LeCun – who understands LLMs like no other – to pollute X with pessimist views, so Grok fails. “My relationship with X\/Twitter is love\/hate,” quipped LeCunn, further fuelling speculation. This development comes in the backdrop of xAI announcing a Series B funding round of $6 billion to expand its team, making xAI the second-most valuable AI startup at a $24 billion valuation, outperformed only by OpenAI, which is at an $86 billion valuation. It is also impressive to see it surpassed Anthropic in less than a year, which now stands as the third-most valuable AI startup at an $18 billion valuation. Advises People to Not Join xAI LeCun believes that LLMs that power generative AI products such as ChatGPT will never achieve the ability to reason and plan like humans, or achieve AGI. LeCun is of the opinion that animals are more intelligent than AI. “General intelligence, artificial or natural, does not exist. Cats, dogs, humans and all animals have specialised intelligence,” said LeCun recently. The banter between LeCun and Musk started after the latter invited people to join xAI’s mission. “Join xAI if you can stand a boss who claims that what you are working on will be solved next year (no pressure),” responded LeCun, advising interested candidates against joining Musk’s company. Further, he said that he likes Musk’s cars, rockets, solar panels, and satellite network but dislikes his vengeful politics, conspiracy theories, and hype. LeCun believes he is politically correct because he is a “scientist, not a business or product person” unlike Musk. When Musk questioned his contribution, as to how much research he conducted “in the last five years,” LeCun candidly replied saying, “Over 80 technical papers published since January 2022.” A few days ago, LeCun clarified that FAIR has roughly 500 scientists and engineers, and he doesn’t run FAIR; Joelle Pineau does. “In fact, I don’t run anything, I’m nobody’s boss. I’m the Chief AI Scientist: I provide ideas and advice to teams,” he said. xAI vs Meta Musk recently told his investors that his goal is to have the supercomputer operational by fall 2025. Once completed, this supercomputer — comprising NVIDIA’s flagship H100 GPUs — will be at least four times larger than the biggest existing GPU clusters, including those built by Meta Platforms for training AI models, he told investors. Currently, xAI is reportedly set to spend $10 billion on Oracle Cloud servers. In an interview with Norway Wealth Fund CEO Nicolai Tangen on X Spaces, Musk revealed that training the Grok 2 model requires approximately 20,000 Nvidia H100 GPUs. He added that training the Grok 3 model and future versions will necessitate 100,000 Nvidia H100 GPUs. In April, xAI introduced Grok-1.5V, a first-generation multimodal model. In addition to its strong text capabilities, Grok can process a wide variety of visual information, including documents, diagrams, charts, screenshots, and photographs. Meanwhile, Musk’s Tesla, now headquartered in Austin, is also developing a Dojo supercomputer. NVIDIA provided 35,000 H100 GPUs to Tesla, paving the way for the breakthrough performance of FSD Version 12, their latest autonomous driving software based on Vision. As far as Meta is concerned, LeCun recently confirmed that the company has obtained $30 billion worth of NVIDIA GPUs to train their AI models. Enough to run a small nation or even put a man on the moon in 1969. Earlier this year, Zuckerberg announced that Meta is building massive compute infrastructure to train Llama 3 and will acquire 350,000 H100s by the end of this year, aiming for a total of almost 600,000 H100s worth of compute. During the latest NVIDIA earnings call, the company said that the big highlight of the quarter was Meta’s announcement of Llama 3, their latest LLM, which was trained on a cluster of 24,000 H100 GPUs. Llama 3 powers Meta AI, a new AI assistant available on Facebook, Instagram, WhatsApp, and Messenger. Using Meta AI, users can access real-time information from across the web without having to bounce between apps. Interestingly, Meta AI Assistant is quite similar to xAI’s Grok, as both are generative AI applications on social media platforms and generate real-time information. Recently, Grok has faced criticism for generating hallucinated content on several occasions. Meanwhile, Meta is yet to release the Llama 3 400B models, which are still in training. The company said that over the coming months, they will release multiple models with new capabilities, including multimodality, the ability to converse in multiple languages, a much longer context window, and stronger overall capabilities. OpenAI is Likely to Reach AGI Before xAI and Meta According to reports, Microsoft and OpenAI are also working on plans for a data centre project that could cost as much as $100 billion and include an AI supercomputer called Stargate, set to launch in 2028. OpenAI chief Sam Altman recently proposed the concept of Universal Basic Compute, in which everyone would have access to a portion of GPT-7’s computing resources. “I wonder if the future looks something more like Universal Basic Compute than Universal Basic Income, and everybody gets like a slice of GPT-7 compute,” said Altman, in the recent episode of the All-In Podcast. Most recently, OpenAI announced that it had started training its next frontier model, GPT-5, and anticipates that it will bring the next level of capabilities on their path to achieve AGI. Sadly, there is no convincing LeCun. A few days ago, he took a dig at OpenAI, sarcastically saying: “Come work at ClosedAI. With AGI just around the corner, your shares will be worth 42 sextillionnollars. We can claw back your vested shares if you quit, unless you sign a non-disparagement agreement. Oh wait, sorry, we didn’t realise our contract was this harsh until someone published an article on it. You can keep them.”","excerpt":"LeCun believes he is politically correct because he is a “scientist, not a business or product person” unlike Musk.","categories":["AI Features"],"tags":["Meta","XAI","Yann LeCun"],"author_name":"Siddharth Jindal","publish_date":"2024-05-29T16:00:05","publication_year":"2024","word_count":1064,"keywords":["Anthropic","ChatGPT","Meta","Yann LeCun","Meta AI","GPT-5","AI","OpenAI","Grok 3","XAI","Aim","generative AI","xAI"],"extracted_tech_keywords":["AI","generative AI","GPT-5","ChatGPT","OpenAI","Anthropic","Meta AI","Grok 3","xAI","Aim"],"url":"https:\/\/analyticsindiamag.com\/ai-features\/yann-lecun-puts-a-brake-on-musks-agi-timeline\/","complexity_score":3,"technical_depth":10,"industry_relevance_score":2,"has_code_examples":false,"has_tutorial_content":false,"is_research_content":true}]